Citation
The Role of Genetics in Vitiligo Susceptibility

Material Information

Title:
The Role of Genetics in Vitiligo Susceptibility
Creator:
Herbstman, Deborah
Place of Publication:
[Gainesville, Fla.]
Florida
Publisher:
University of Florida
Publication Date:
Language:
english
Physical Description:
1 online resource (111 p.)

Thesis/Dissertation Information

Degree:
Doctorate ( Ph.D.)
Degree Grantor:
University of Florida
Degree Disciplines:
Medical Sciences
Genetics (IDP)
Committee Chair:
Wallace, Margaret R.
Committee Co-Chair:
McCormack, Wayne T.
Committee Members:
Driscoll, Daniel J.
Garvan, Cynthia W.
Graduation Date:
8/8/2009

Subjects

Subjects / Keywords:
Alleles ( jstor )
Diseases ( jstor )
Genetics ( jstor )
Haplotypes ( jstor )
Introns ( jstor )
Melanin ( jstor )
Melanocytes ( jstor )
P values ( jstor )
Skin ( jstor )
Vitiligo ( jstor )
Genetics (IDP) -- Dissertations, Academic -- UF
aire, autoimmune, comt, dct, genetics, tyrosinase, tyrp1, vitiligo
Genre:
bibliography ( marcgt )
theses ( marcgt )
government publication (state, provincial, terriorial, dependent) ( marcgt )
born-digital ( sobekcm )
Electronic Thesis or Dissertation
Medical Sciences thesis, Ph.D.

Notes

Abstract:
Vitiligo is an autoimmune pigment disorder of the skin, which is seen on patients as depigmented areas that may gradually enlarge. Vitiligo is a common condition, affecting about 0.5-1% of all ethnic groups worldwide. It is also associated with an increased risk for other autoimmune diseases. The current treatments for vitiligo are difficult, expensive, and often yield marginal results. The cause of vitiligo is unknown, but is thought to involve both genetic and environmental factors. The goal of my study was to test the hypothesis that vitiligo pathogenesis is caused in part by genetic susceptibility to both autoimmune and autotoxic events due to genetic differences in genes involved in the regulation of the immune response, melanin production, and oxidative stress. To identify vitiligo susceptibility genes, human genomic DNA samples from patients with vitiligo, their family members, and healthy controls with no known autoimmune diseases were genotyped for a number of different single nucleotide polymorphisms (SNPs) in candidate genes. SNPs in COMT, TYR, TYRP1, DCT, and PAH, genes involved in melanin biosynthesis, were examined. TYR, TYRP1, and DCT are also melanocyte autoantigens, which further implicates them as vitiligo susceptibility genes. The immunoregulatory gene AIRE was tested. Genes involved in the regulation of oxidative stress and DNA repair, specifically, FXBO11 and MSH6, also were examined. Using case-control and family-based genetic association studies, as well as haplotype analysis, several putative susceptibility genes for vitiligo were identified. Susceptibility to vitiligo was linked to the AIRE gene. The AIRE protein is responsible for T-cell processing in the thymus. Significant results were also found in the melanin biosynthesis genes COMT, PAH, TYR, TYRP1 and DCT. These results demonstrate a possible role for genes involved in immune system regulation, as well as for genes involved in melanin synthesis in vitiligo susceptibility. The aim of my study was to identify genes involved in vitiligo susceptibility so that in the future, novel therapies that might prevent or ameliorate vitiligo may be developed based on my understanding of the genetic causes of vitiligo. ( en )
General Note:
In the series University of Florida Digital Collections.
General Note:
Includes vita.
Bibliography:
Includes bibliographical references.
Source of Description:
Description based on online resource; title from PDF title page.
Source of Description:
This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Thesis:
Thesis (Ph.D.)--University of Florida, 2009.
Local:
Adviser: Wallace, Margaret R.
Local:
Co-adviser: McCormack, Wayne T.
Electronic Access:
RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2010-08-31
Statement of Responsibility:
by Deborah Herbstman.

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
Copyright Herbstman, Deborah. Permission granted to the University of Florida to digitize, archive and distribute this item for non-profit research and educational purposes. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder.
Embargo Date:
8/31/2010
Classification:
LD1780 2009 ( lcc )

Downloads

This item has the following downloads:


Full Text





Mehers, K.L. and Gillespie, K.M.: The genetic basis for type 1 diabetes. Br Med Bull 88 (2008)
115-29.

Merimsky, O., Shoenfeld, Y., Chaitchik, S., Yecheskel, G. and Fishman, P.: Antigens and
antibodies in malignant melanoma. Tumour Biol 15 (1994) 188-202.

Millington, G.W. and Levell, N.J.: Vitiligo: the historical curse of depigmentation. Int J
Dermatol 46 (2007) 990-5.

Mizuta, I., Satake, W., Nakabayashi, Y., Ito, C., Suzuki, S., Momose, Y., Nagai, Y., Oka, A.,
Inoko, H., Fukae, J., Saito, Y., Sawabe, M., Murayama, S., Yamamoto, M., Hattori, N.,
Murata, M. and Toda, T.: Multiple candidate gene analysis identifies alpha-synuclein as a
susceptibility gene for sporadic Parkinson's disease. Hum Mol Genet 15 (2006) 1151-8.

Morris, G.A., Lowe, C.E., Cooper, J.D., Payne, F., Vella, A., Godfrey, L., Hulme, J.S., Walker,
N.M., Healy, B.C., Lam, A.C., Lyons, P.A. and Todd, J.A.: Polymorphism discovery and
association analyses of the interferon genes in type 1 diabetes. BMC Genet 7 (2006) 12.

Mozzanica, N., Frigerio, U., Finzi, A.F., Cattaneo, A., Negri, M., Scaglione, F., Fraschini, F. and
Foppa, S.: T cell subpopulations in vitiligo: a chronobiologic study. J Am Acad Dermatol
22 (1990) 223-30.

Mulekar, S.V.: Melanocyte-keratinocyte cell transplantation for stable vitiligo. Int J Dermatol 42
(2003) 132-6.

Naftzger, C., Takechi, Y., Kohda, H., Hara, I., Vijayasaradhi, S. and Houghton, A.N.: Immune
response to a differentiation antigen induced by altered antigen: a study of tumor
rejection and autoimmunity. Proc Natl Acad Sci U S A 93 (1996) 14809-14.

Nagamine, K., Peterson, P., Scott, H.S., Kudoh, J., Minoshima, S., Heino, M., Krohn, K.J.,
Lalioti, M.D., Mullis, P.E., Antonarakis, S.E., Kawasaki, K., Asakawa, S., Ito, F. and
Shimizu, N.: Positional cloning of the APECED gene. Nat Genet 17 (1997) 393-8.

Nancy, A.L. and Yehuda, S.: Prediction and prevention of autoimmune skin disorders. Arch
Dermatol Res 301 (2009) 57-64.

Nath, S.K., Majumder, P.P. and Nordlund, J.J.: Genetic epidemiology of vitiligo: multilocus
recessivity cross-validated. Am J Hum Genet 55 (1994) 981-90.

Naughton, G.K., Eisinger, M. and Bystryn, J.C.: Detection of antibodies to melanocytes in
vitiligo by specific immunoprecipitation. J Invest Dermatol 81 (1983) 540-2.

Nicolaidou, E., Antoniou, C., Stratigos, A. and Katsambas, A.D.: Narrowband ultraviolet B
phototherapy and 308-nm excimer laser in the treatment of vitiligo: a review. J Am Acad
Dermatol 60 (2009) 470-7.









Das, P.K., van den Wijngaard, R.M., Wankowicz-Kalinska, A. and Le Poole, I.C.: A symbiotic
concept of autoimmunity and tumour immunity: lessons from vitiligo. Trends Immunol
22 (2001) 130-6.

Davis, M.D., Ribeiro, P., Tipper, J. and Kaufman, S.: "7-tetrahydrobiopterin," a naturally
occurring analogue of tetrahydrobiopterin, is a cofactor for and a potential inhibitor of the
aromatic amino acid hydroxylases. Proc Natl Acad Sci U S A 89 (1992) 10109-13.

De La Vega, F.M., Isaac, H.I. and Scafe, C.R.: A tool for selecting SNPs for association studies
based on observed linkage disequilibrium patterns. Pac Symp Biocomput (2006) 487-98.

Drago, A., De Ronchi, D. and Serretti, A.: Incomplete coverage of candidate genes: a poorly
considered bias. Curr Genomics 8 (2007) 476-83.

Dunston, G.M. and Halder, R.M.: Vitiligo is associated with HLA-DR4 in black patients. A
preliminary report. Arch Dermatol 126 (1990) 56-60.

Em, S., Laddha, N.C., Chatterjee, S., Gani, A.R., Malek, R.A., Shah, B.J. and Begum, R.:
Association of catalase T/C exon 9 and glutathione peroxidase codon 200 polymorphisms
in relation to their activities and oxidative stress with vitiligo susceptibility in Gujarat
population. Pigment Cell Res 20 (2007) 405-7.

Fain, P.R., Babu, S.R., Bennett, D.C. and Spritz, R.A.: HLA class II haplotype DRB1*04-
DQB 1*0301 contributes to risk of familial generalized vitiligo and early disease onset.
Pigment Cell Res 19 (2006) 51-7.

Fain, P.R., Gowan, K., LaBerge, G.S., Alkhateeb, A., Stetler, G.L., Talbert, J., Bennett, D.C. and
Spritz, R.A.: A genomewide screen for generalized vitiligo: confirmation of AIS 1 on
chromosome lp31 and evidence for additional susceptibility loci. Am J Hum Genet 72
(2003) 1560-4.

Falabella, R.: Surgical therapies for vitiligo. Clin Dermatol 15 (1997) 927-39.

Finco, O., Cuccia, M., Martinetti, M., Ruberto, G., Orecchia, G. and Rabbiosi, G.: Age of onset
in vitiligo: relationship with HLA supratypes. Clin Genet 39 (1991) 48-54.

Fishman, P., Azizi, E., Shoenfeld, Y., Sredni, B., Yecheskel, G., Ferrone, S., Zigelman, R.,
Chaitchik, S., Floro, S. and Djaldetti, M.: Vitiligo autoantibodies are effective against
melanoma. Cancer 72 (1993) 2365-9.

Gajko-Galicka, A.: Mutations in type I collagen genes resulting in osteogenesis imperfecta in
humans. Acta Biochim Pol 49 (2002) 433-41.

Gibson, T.J., Ramu, C., Gemund, C. and Aasland, R.: The APECED polyglandular autoimmune
syndrome protein, AIRE-1, contains the SAND domain and is probably a transcription
factor. Trends Biochem Sci 23 (1998) 242-4.











Table 4-2. Linkage disequilibrium
Gene SNP ID
DCT rs9516418
DCT rs9524493
DCT rs9516413
DCT rs11618471
DCT rs11618471
DCT rs7987802
DCT rs7991232
DCT rs9524493
DCT rs4318084
DCT rs7987802
DCT rs7987802
DCT rs7991232
DCT rs11618471
DCT rs9524493
DCT rs11618471
DCT rs9524493
DCT rs4318084
DCT rs7987802
DCT rs7991232
DCT rs7991232
DCT rs11618471
DCT rs7991232
DCT rs9516418
DCT rs11618471
DCT rs11618471
DCT rs9524493
DCT rs4318084
DCT rs7991232
PAH rs1722381
TYR rs1042602
TYR rs1042602
TYR rs1042602
TYR rs1042602
TYR rs10765197
TYR rs12791412
TYR rs1042602
TYR rs12791412
TYR rs12791412
TYR rs12791412


(D') values for DCT, PAH, TYR, TYRP1, and COMT
SNP ID D'
rs9516413 0.02
rs4318084 0.25
rs1028805 0.38
rs7987802 1.00
rs9516413 1.00
rs4318084 1.00
rs9516413 0.59
rs1028805 0.41
rs9516413 1.00
rs1028805 0.97
rs9516418 0.90
rs7987802 0.94
rs7991232 0.97
rs9516413 0.79
rs9524493 0.98
rs7987802 1.00
rs9516418 0.54
rs9516413 0.61
rs4318084 0.91
rs1028805 0.88
rs9516418 1.00
rs9516418 0.83
rs1028805 0.62
rs1028805 1.00
rs4318084 0.97
rs9516418 0.95
rs1028805 0.99
rs9524493 0.88
rs1522307 0.48
rs12791412 0.95
rs1827430 0.87
rs2000554 0.85
rs10830250 0.96
rs1042602 0.99
rs621313 0.93
rs594647 0.94
rs594647 0.93
rs2000554 0.95
rs1827430 0.88










Table 4-19. Significant Hardy-Weinberg equilibrium analysis for TYRP1
Gene SNP p-value cases p-value controls
TYRP1 rs2762462 0.022* 0.66
TYRP1 rs2733833 0.011* 0.23
TYRP1 rs683 0.024* 0.61
TYRP1 rs2762464 0.019* 0.35
* denotes a significant finding


Table 4-20. Pairwise haplotype analysis results for TYRP1 rs683 and TYRP1 rs2733831
Frequency Controls Vitiligo patients p-values
0 [CA] 107(43.5%) 80(37.6%)
1 [CA] 104 (42.5%) 115(54.4%)
2 [CA] 34 (14%) 17 (8%)
Total (n) 245 212
X2 0.0203*
Odds ratio for 1 [CA] vs. 0 [CA] = 1.4950 95% CI=1.0083 to 2.2168 0.0454*
Odds ratio for 2 [CA] vs. 0 [CA] = n.s. n.s.
Odds ratio for 2 [CA] vs. 1 [CA] = 0.4572 95% CI=0.2409 to 0.8674 0.0166*
n.s. = not significant; denotes a significant finding; number in frequency column indicates the
number of copies of this haplotype


Table 4-21. Pairwise haplotype analysis results for TYRP1 rs683 and TYRP1 rs2733833
Frequency Controls Vitiligo patients p-values
0 [CT] 104 (42.5%) 79 (37%)
1 [CT] 107 (43.5%) 114(54%)
2 [CT] 34(14%) 18(9%)
Total (n) 245 212
X2 0.0434*
Odds ratio for 1 [CT] vs. 0 [CT] = n.s. n.s.
Odds ratio for 2 [CT] vs. 0 [CT] = n.s. n.s.
Odds ratio for 2 [CT] vs. 1 [CT] = 0.4979 95% CI=0.2652 to 0.9349 0.0300*
n.s. = not significant; denotes a significant finding; number in frequency column indicates the
number of copies of this haplotype









THE ROLE OF GENETICS IN VITILIGO SUSCEPTIBILITY


By

DEBORAH MARSHA HERBSTMAN


















A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY

UNIVERSITY OF FLORIDA

2009









CHAPTER 1
INTRODUCTION

Vitiligo is a common dermatological disorder of the epidermis and hair follicles, affecting

both genders equally and -0.5-1% of the population in all ethnic groups worldwide (Kovacs,

1998; Nordlund and Ortonne, 1998; Spritz, 2007). The average age of onset is 22 + 16 years

(Nordlund and Ortonne, 1998). The disease is defined clinically by expanding areas of

hypopigmentation on the skin surface due to the destruction or inactivation of epidermal

melanocytes (Badri et al., 1993; Majumder et al., 1993; Tobin et al., 2000; Spritz, 2007). Vitiligo

pathology is limited to the depigmentation of the epidermis, but the illness is often comorbid

with other autoimmune disorders such as alopecia areata and Hashimoto's thyroiditis (Badri et

al., 1993; Kemp et al., 1999a). Loss of pigmentation can occur anywhere on the body. The

striking appearance of depigmented areas surrounded by normally pigmented skin can cause

social stigmatism and physiological distress (Millington and Levell, 2007; Sampogna et al.,

2008). Because patients are not usually born with the disease and many report coinciding events

at or just prior to onset, it has been suggested that vitiligo might be "triggered" by environmental

factors (Manolache et al., 2009). Reported triggers include skin wounds, eczema, pregnancy,

psychological stressors, certain drugs, and medical procedures (e.g., bone marrow transplant)

(Brown-Harrell et al., 1996; Papadavid et al., 1996). When vitiligo lesions develop at sites of

trauma, such as abrasions, surgical scars, radiography treatments, resolving psoriasis or eczema,

contact dermatitis, or severe sunburn, this is known as the Koebner phenomenon (Powell and

Dicken, 1983; Levine and Ribeiro, 1994). It has been shown that the numbers of melanocytes in

depigmented vitiligo lesions are vastly reduced or absent; however, the mechanisms of

destruction have been widely debated (Nordlund and Ortonne, 1998). Treatment for vitiligo is

often both expensive and time and labor intensive for the patient, with insurance companies often











APPENDIX A
SINGLE NUCLEOTIDE POLYMORPHISM DETAILS


Table A-1. All SNPs genotyped
Base pair
Gene change & Chromosome: Allelic Genotypic TDT
Symbol rs# location base p-value p-value p-value
AIRE rs2776377 A/G promoter 21:44529072 0.86 0.46 1.00
C/G exon 7,
AIRE rs1800520 S278R 21:44534334 9.90E-09* 2.00E-07* 0.16
COMT rs2020917 T/C 5'UTR 22:18308884 0.12 0.18 0.56
COMT rs6269 G/A intron 2 22:18329952 0.90 0.39 0.10
DCT rs9516413 C/T 3'UTR 13:93889348 0.0047* 0.027* 0.0052*
DCT rs7987802 C/T intron 6 13:93895052 0.48 0.57 0.9
DCT rs4318084 C/T intron 6 13:93896999 0.26 0.13 6.2E-06*
DCT rs9524493 C/G intron 6 13:93899790 0.32 0.55 0.49
DCT rs9516418 C/T intron 6 13:93909510 0.88 0.98 0.0097*
DCT rs1028805 G/T intron 2 13:93917270 0.86 0.87 0.0025*
DCT rs11618471 A/G intron 2 13:93918325 0.23 0.35 3.7E-06*
DCT rs7991232 A/G intron 1 13:93928296 0.33 0.46 0.81
FBX011 rs960106 C/T intron 1 2:47932342 0.95 0.98 0.66
C/G 3'UTR
(FBXO11),
FBXO11, intron 8
MSH6 rs3136367 (MSH6) 2:47887055 0.13 0.21 0.16
PAH rsl722381 A/G intron 6 12:101771966 0.19 0.44 0.65
PAH rs1522307 A/G intron 2 12:101822647 0.38 0.39 0.79


rs1042602
rs621313
rs594647
rs10765197
rs12791412
rs2000554
rs10830250
rs1827430
rs2762462
rs2733831
rs2733833
rs683
rs2762464
rs1063380


A/C exon 1,
S192Y
G/A intron 1
T/C intron 1
A/C intron 2
A/G intron 2
G/A intron 2
C/G intron 3
G/A intron 4
C/T intron 4
A/G intron 5
G/T intron 6
A/C 3'UTR
A/T 3'UTR
C/T 3'UTR


11:88551344
11:88553311
11:88561205
11:88564976
11:88570229
11:88575589
11:88617255
11:88658088
9:12689776
9:12693484
9:12695095
9:12699305
9:12699586
9:12700090


* denotes a significant finding; all values are uncorrected p-values


TYR
TYR
TYR
TYR
TYR
TYR
TYR
TYR
TYRP1
TYRP1
TYRP1
TYRP1
TYRP1
TYRP1


0.49
0.81
0.67
0.52
0.05*
0.34
0.25
0.09
0.81
0.72
0.84
0.58
0.72
0.57


0.31
0.45
0.34
0.32
0.13
0.24
0.31
0.12
0.12
0.41
0.0254*
0.11
0.06
0.10


0.31
0.51
0.30
0.12
0.02*
0.29
0.26
0.17
1.0
0.81
0.43
0.81
0.90
0.39









Tyrosinase

The role of tyrosinase in pigment production is clear as this enzyme represents the rate-

limiting step of melanogenesis. Severe disruptions of tyrosinase function are known to cause

albinism, though I hypothesize that more subtle alterations of this gene may be linked to vitiligo.

Of the eight SNPs genotyped in TYR, only one yielded significant results in independent

analyses. The rs12791412 SNP, located in intron two, was significant in allele frequency

analysis between white cases and controls and in family-based testing. This significance was not

seen in genotype analysis. In my entire study, this was the only SNP that yielded a significant

finding for mean age of onset by genotype association. Patients homozygous for the minor allele

(G) had a significantly later mean age of onset. However, because I had a limited sample size for

my within-case analysis (n=209), and my finding was linked to the minor allele, my results must

be interpreted with caution. Further, the average age of onset for heterozygotes was not

intermediate to both homozygotes.

Six haplotype findings were significant and five of these contained rs1042602, a SNP in

exon 1 that causes the S192Y amino acid change. This SNP has been found to be significantly

linked to skin pigment variation in a South Asian population (Stokowski et al., 2007), and

several studies found it to be highly polymorphic in European populations contributing to normal

variation in skin color (Shriver et al., 2003; 2005; Norton et al., 2007). One group found

rs1042602 to be associated with freckling in an Icelandic population (Sulem et al., 2007).

Additional investigation by Sulem et al. (2007) found that the ancestral C allele of rsl042602

was fixed in the East Asian and Nigerian HapMap samples, whereas the A allele was found at a

frequency of approximately 35% in European populations. They found strong evidence that the

rs1042602 A allele (which was associated with the absence of freckles in Icelanders) was subject

to positive selection in European populations. Interestingly, only 1.7% of HapMap SNPs show a










Table 4-22. Pairwise haplotype analysis results for TYRP1 rs273383 land TYRP1 rs2762464
Frequency Controls Vitiligo patients p-values
0 [AA] 104 (42.5%) 76 (36%)
1 [AA] 104 (42.5%) 114(54%)
2 [AA] 37(15%) 22(10%)
Total (n) 245 212
X2 0.0434*
Odds ratio for 1 [AA] vs. 0 [AA] = 1.5171 95% CI=1.0181 to 2.2607 0.0405*
Odds ratio for 2 [AA] vs. 0 [AA] = n.s. n.s.
Odds ratio for 2 [AA] vs. 1 [AA] = 0.5475 95% CI=0.3030 to 0.9893 0.0460*
n.s. = not significant; denotes a significant finding; number in frequency column indicates the
number of copies of this haplotype




Table 4-23. Pairwise haplotype analysis results for TYRP1 rs2733831 and TYRP1 rs2733833
Frequency Controls Vitiligo patients p-values
0 [AT] 106 (43%) 80 (38%)
1 [AT] 103 (42%) 114(54%)
2 [AT] 35 (14%) 18 (8%)
Total (n) 245 212
X2 0.0275*
Odds ratio for 1 [AT] vs. 0 [AT] = n.s. n.s.
Odds ratio for 2 [AT] vs. 0 [AT] = n.s. n.s.
Odds ratio for 2 [AT] vs. 1 [AT] = 0.4763 95% CI=0.2539 to 0.8933 0.0208*
n.s. = not significant; denotes a significant finding; number in frequency column indicates the
number of copies of this haplotype




Table 4-24. Pairwise haplotype analysis results for TYRP1 rs2762464 and TYRP1 rs2733833
Frequency Controls Vitiligo patients p-values
0 [AT] 105 (43%) 78 (37%)
1 [AT] 104 (42%) 115 (54%)
2 [AT] 36(15%) 19(9%)
Total (n) 245 212
X2 0.0242*
Odds ratio for 1 [AT] vs. 0 [AT] = 1.4985 95% CI=1.0087 to 2.2263 0.0452*
Odds ratio for 2 [AT] vs. 0 [AT] = n.s. n.s.
Odds ratio for 2 [AT] vs. 1 [AT] = 0.4849 95% CI=0.2616 to 0.8988 0.0215*
n.s. = not significant; denotes a significant finding; number in frequency column indicates the
number of copies of this haplotype










CHAPTER 3
AUTOIMMUNE REGULATOR GENE

Introduction

As in other autoimmune diseases, one hypothesis for the genetic pathogenesis of vitiligo

is that immunoregulatory genes may contribute to the autoimmune response in vitiligo patients.

One gene implicated in immune function that is examined in the current study is the autoimmune

regulator, or AIRE, gene. The AIRE gene is located on chromosome 21q22.3 and consists of 14

exons coding for a 2445-base pair mRNA transcript; the translated product has 546 amino acids

with a molecular mass of approximately 57.5 kDa (Nagamine et al., 1997). At the subcellular

level, AIRE can be found in the cell nucleus in a speckled pattern also known as nuclear dots

(Bjorses et al., 1999).

Mutations in the AIRE gene are known to cause autoimmune polyendocrinopathy-

candidiasis-ectodermal dystrophy syndrome (APECED), otherwise known as autoimmune

polyglandular syndrome (APS). The most common features of APECED are parathyroid gland

failure, chronic susceptibility to candida yeast infection, and Addison's disease. Other

autoimmune diseases associated with APECED include alopecia, vitiligo, ovarian failure,

testicular atrophy, hypothyroidism, gastric parietal cell atrophy, intestinal malabsorption, and

insulin-dependent diabetes mellitus. It is possible that subtle changes in this gene may confer

vitiligo susceptibility, in the absence of APECED.

AIRE is an important DNA binding molecule involved in immune regulation (Kumar et al.,

2001; Peterson et al., 2008). The AIRE protein is predominantly expressed in thymic epithelial

cells (TECs), but also in some monocyte-derived cells of the thymus, in a subset of cells in

lymph nodes, in the spleen, and in the fetal liver. AIRE mRNA is weakly detected in peripheral

blood leukocytes but thymic AIRE expression is several fold higher than in other tissues (Su and









warranted to rule out other factors contributing to this association, which is still highly

significant even with the conservative Bonferroni correction.

Several groups examining autoimmune diseases and the AIRE gene have studied the same

SNP that I found to be significantly associated with vitiligo; however, the other studies yielded

mixed results. The rs1800520 SNP showed no evidence of association with type 1 diabetes in a

Finnish population (p=0.43) (Turunen et al., 2006), and genotype frequencies were nearly equal

among cases (0.015) and controls (0.016) for the homozygous minor allele that was found to be

very highly significant in my study. This SNP also showed no association to autoimmune

Addison's disease (AAD) in a Norwegian population (Boe Wolff et al., 2008). However, a

British group found rs1800520 to be significantly associated with alopecia areata (Tazi-Ahnini et

al., 2002).

Prior studies testing vitiligo association with the AIRE gene had mixed results. The same

British group that found AIRE to have strong association with alopecia areata also found very

strong (p=0.000414) associations with vitiligo and SNPs in the AIRE gene. Further, their

haplotype analysis of this gene also proved to be significant (Tazi-Ahnini et al., 2008). In

contrast, Jin and colleagues did not find significant association with AIRE and vitiligo in tests of

eight AIRE SNPs (2007a).

The immune-related gene AIRE was investigated because of its role in the modulation of

immune responses. Statistical analysis for cases and controls included only whites with vitiligo,

due to a lack of sufficient numbers of subjects of other ethnic backgrounds. An autoimmune

component in vitiligo susceptibility may be supported by the results of my study. The lack of a

significant TDT to support the association of AIRE with vitiligo suggested by my white case-

control analysis may indicate that larger sample sizes are needed to confirm this finding in









whereas MSH6 expression is increased in vitiligo patients compared to healthy controls,

FBXO11 expression is reduced (Le Poole et al., 2001).

Treatment

There is no cure for vitiligo; treatment methods function by stopping current progression of

the disease and aiding in the repigmentation of hypopigmented lesions. The treatment modalities

for vitiligo are difficult, expensive, and often disappointing, as they do not prevent the

reappearance of new lesions. Many treatments are not covered by health insurance. The early

options for vitiligo treatment were limited mainly to psoralens with UVA radiation (315-400

nm), also known as PUVA therapy, and the use of topical corticosteroids. With this approach,

repigmentation of the lesions is inconsistent and slow, and the lesions may worsen during

therapy (Njoo et al., 1998; Bethea et al., 1999; Buckley and du Vivier, 1999).

Current treatment options have expanded and now include a variety of choices such as

epidermal melanocyte autografts, UV-B therapy (between 280 and 315 nm), oral and topical

phenylalanine, and the 308-nm excimer laser (Falabella, 1997; Camacho and Mazuecos, 1999;

Njoo and Westerhof, 2001; Nicolaidou et al., 2009). Another treatment for vitiligo is the topical

application of UVB-activated pseudocatalase. Pseudocatalase functions to mimic the human

enzyme catalase, which removes H202 and other free radicals from the epidermis. In one study,

treatment of vitiligo patients with topical applications of the pseudocatalase cream, along with

frequent exposure to UVB radiation at a lower than normally therapeutic doses, allowed for the

halt of depigmentation in 95% of patients, and repigmentation of 60% of individuals

(Schallreuter, 1999; Schallreuter et al., 1999). However, a group from Australia was unable to

replicate the success of pseudocatalase creams seen by Schallreuter and colleagues, though they

did confirm the therapeutic benefit of higher UVB exposure and may have used a different

pseudocatalase formulation (Bakis-Petsoglou et al., 2009). Among the more unique modalities,









Park, H.H., Ha, E., Uhm, Y.K., Jin, S.Y., Kim, Y.J., Chung, J.H. and Lee, M.H.: Association
study between catalase gene polymorphisms and the susceptibility to vitiligo in Korean
population. Exp Dermatol 15 (2006) 377-80.

Passeron, T. and Ortonne, J.P.: Physiopathology and genetics of vitiligo. J Autoimmun 25 Suppl
(2005) 63-8.

Passi, S., Grandinetti, M., Maggio, F., Stancato, A. and De Luca, C.: Epidermal oxidative stress
in vitiligo. Pigment Cell Res 11 (1998) 81-5.

Pavel, S., Muskiet, F.A., de Ley, L., The, T.H. and van der Slik, W.: Identification of three
indolic compounds in a pigmented-melanoma cell-culture supernatant by gas
chromatography-mass spectrometry. J Cancer Res Clin Oncol 105 (1983) 275-9.

Peterson, P., Org, T. and Rebane, A.: Transcriptional regulation by AIRE: molecular
mechanisms of central tolerance. Nat Rev Immunol 8 (2008) 948-57.

Phillips, J., Gawkrodger, D.J., Caddy, C.M., Hedley, S., Dawson, R.A., Smith-Thomas, L.,
Freedlander, E. and Mac Neil, S.: Keratinocytes suppress TRP-1 expression and reduce
cell number of co-cultured melanocytes implications for grafting of patients with
vitiligo. Pigment Cell Res 14 (2001) 116-25.

Pitkanen, J. and Peterson, P.: Autoimmune regulator: from loss of function to autoimmunity.
Genes Immun 4 (2003) 12-21.

Powell, F.C. and Dicken, C.H.: Vitiligo and psoriasis. J Am Acad Dermatol 8 (1983) 136-7.

Purohit, S., Kumar, P.G., Laloraya, M. and She, J.X.: Mapping DNA-binding domains of the
autoimmune regulator protein. Biochem Biophys Res Commun 327 (2005) 939-44.

Ram, M. and Shoenfeld, Y.: Harnessing autoimmunity (vitiligo) to treat melanoma: a myth or
reality? Ann N Y Acad Sci 1110 (2007) 410-25.

Redondo, M.J., Fain, P.R. and Eisenbarth, G.S.: Genetics of type 1A diabetes. Recent Prog Horm
Res 56 (2001) 69-89.

Rice, T.K., Schork, N.J. and Rao, D.C.: Methods for handling multiple testing. Adv Genet 60
(2008) 293-308.

Risch, N. and Merikangas, K.: The future of genetic studies of complex human diseases. Science
273 (1996) 1516-7.

Riva, A. and Kohane, I.S.: SNPper: retrieval and analysis of human SNPs. Bioinformatics 18
(2002) 1681-5.

Rizzi, M., Ferrera, F., Filaci, G. and Indiveri, F.: Disruption of immunological tolerance: role of
AIRE gene in autoimmunity. Autoimmun Rev 5 (2006) 145-7.









Table 2-2. List of Applied Biosystems SNP Assay IDs and description of FBXO11 and MSH6
SNPs genotyped
Gene(s) SNP rs number Applied Biosystems ID Base pair change & location
FBXO11 rs960106 C 448948 10 C/T 5'UTR
FBXO11 & MSH6 rs3136367 C 22271588_10 C/G 3'UTR (FBXO11),
intron 8 (MSH6)


Table 2-3. Linkage disequilibrium
Gene SNP ID
FBXO11/MSH6 rs3136367


(D') values for FBXO11 and MSH6
SNP ID D'
rs960106 0.81









Materials and Methods

DNA collection, SNP genotyping, and statistical analyses for genetic association were

performed as described in Chapter 2. A total of twenty-six SNPs were selected from five genes

(Table 4-1). Two SNPs on PAH, eight on DCT (Figure 4-1A), eight on TYR, six on TYRP1

(Figure 4-1B), and two on COMTwere chosen to allow for haplotype analysis for each gene

(Table 4-1); only significant values are reported. Pairwise linkage disequilibrium (D') values

were calculated for all SNPs within a gene (Table 4-2). Hardy-Weinberg values were calculated

for all SNPs; only significant values are reported (Table 4-19).

Results

Phenylalanine Hydroxylase

Case-control and family-based analysis of the two individual SNPs in PAH yielded no

significant association with vitiligo (Table 4-3, 4-4, and 4-5). Hardy-Weinberg analysis was not

significant. The two SNPs were strongly linked, with D'=0.48 (Table 4-2). Pairwise haplotype

analysis revealed that the [GG] haplotype was significant for cases (11 out of 203) compared to

controls (0 out of 241), p=0.0007 (Table 4-6). Because zero controls had the [GG] haplotype,

odds ratios could not be calculated for this finding.

Dopachrome Tautamerase

Analysis of the DCT gene yielded several positive findings. Most notably, SNP

rs9516413, a C/T polymorphism located in the 3'UTR, was positively associated with vitiligo in

both the allelic (p=0.0047) and genotypic frequency analyses (p=0.0270) (Table 4-3 and 4-4).

Specifically, the C allele was overrepresented in cases and had a relative risk (RR) =1.43

(CI=1.1145 to 1.8397); TDT analysis also showed this allele to be significant (p=0.0052) (Table

4-5). In pairwise haploblock analysis, four of five significant findings for this gene (Table 4-7,

4-8, 4-10, and 4-11) involved the C allele from rs9516413. Three of the other SNPs significant









his excellent sense of humor. I thank Dr. Cynthia W. "Cyndi" Garvan for being an exceptional

teacher, an enthusiastic collaborator, a role model, and my committee member. I would like to

acknowledge Dr. Sally Litherland, whose advice and guidance helped contribute to this project.

I am grateful to my statistical collaborator Dr. Wei Hou; it is a pleasure to work with him. I

would like to thank the UF Center for Pharmacogenomics, especially the Director of the

Genotyping Core Laboratory, Dr. Taimour Langaee, and his talented technician, Lynda Stauffer.

Pandora Cowart has been both an expert database consultant and a friend. I thank my colleague

and friend Amy Non for helpful comments and advice. I am grateful to Dr. Roger Fillingim for

his collaboration and his support of my research. I acknowledge Dr. Steve Blackband for

computer support. I would like to thank the skilled and friendly staff of the Molecular Genetics

and Microbiology Department, especially Joyce Conners, Debbie Burgess, Steve Howard,

Michele Ramsey, Julie Dillard, and Dave Brumbaugh. Dr. Henry Baker, my Department

Chairman, is gratefully acknowledged for his scientific and career advice.

My graduate education has been made all the richer and saner thanks to a dearly loved

group of friends. In my heart, I thank you all. I am grateful for Maisara and Jim Bledsoe and

Sasha and Bo Yurke, as they are friends that have grown to be family. I thank Dr. Adam Rivers

for many late-night phone calls offering PhD support and humorous anecdotes about life. I thank

soon-to-be-a-doctor Andrea "Dr. Dre" Rivers for listening to my crazy stories and for sharing a

few of her own. Dan "BFw/oB" Capellupo's kindness and patience is truly appreciated; he is a

great dance partner and a cherished friend. I thank Stacey Simon for her love of stories and for

going along with a small sample set of my silly suggestions. I appreciate Pamela "Ella" Dubyak

for sharing my love of the ridiculous. I thank my friend and trivia buff extraordinaire Dr. Dimitri

Veras, who helped to copy edit this manuscript. Michelle Edwards and David Edmeades offered









UC) (Table 3-6). The odds ratio analysis showed that controls carrying the haplotype [AG] had

significantly decreased risk of developing signs of vitiligo by 0.1643 times, when the vitiligo

patient group was compared with the control group (OR = 0.1643, 95% CI = 0.0663 to 0.4070).

Discussion

The SNP that I found to be significantly associated with vitiligo, rs1800520, is located in

the AIRE gene SAND DNA-binding domain (Gibson et al., 1998). The SAND domain (named

after Spl00, AIRE-1, NucP41/75, DEAF-1) is a conserved 80-residue amino acid sequence found

in a number of nuclear proteins, many of which function in chromatin-dependent transcriptional

control (Bottomley et al., 2001). Purohit and colleagues found evidence that the SAND domain

of AIRE recognizes a TTATTA DNA motif (Purohit et al., 2005). Analysis of deletion

constructs of recombinant proteins revealed that the major activity of the AIRE SAND domain

resides in the area of amino acids 189-196. The SNP I found to be associated with vitiligo in my

study causes a change in amino acid 278 from a serine to an arginine; changes at this site have an

unknown effect on SAND activity though the substitution of a neutral amino acid serinee) for a

positively charged one (arginine) may disrupt DNA binding as DNA is negatively charged. It

may also affect protein folding or trafficking, or potentially affect regulation by phosphorylation.

The significant [AG] protective haplotype I observed is likely driven by the very strong

association with vitiligo susceptibility found with the C allele on rs1800520. Although the exact

mechanism of this haplotype's influence on AIRE function is unknown, it could alter AIRE

protein activity as rs1800520 does cause an amino acid change. Additionally, irrespective of the

effects of rs1800520, this haplotype may be in linkage disequilibrium with a gene alteration that

is actually responsible for the pathogenic effect. Although I cannot rule out a direct effect of this

haplotype on AIRE gene function and its role in vitiligo susceptibility, more investigation is









denying coverage. The goal of my work was to identify potential vitiligo susceptibility genes

with the hope that this knowledge may lead to better understanding of disease pathogenesis and

more targeted disease treatment.

Melanocytes and Pigmentation

Because vitiligo is a skin disease, a brief overview of the structure and physiology of this

tissue is included. Human skin consists of two main layers: the epidermis, a stratified squamous

epithelium mainly consisting of keratinocytes, and the dermis, an underlying layer of

vascularized connective tissue (Nordlund and Ortonne, 1998). The epidermis is composed of

four cell types: keratinocytes, melanocytes, and two types of non-pigmented granular

dendrocytes, Langerhans cells, and Granstein cells. Both Langerhans cells and Granstein cells

are antigen-presenting cells (APCs) that interact with T-cells to modulate the immunologic

response in the skin. For the purposes of better understanding vitiligo, keratinocytes and

melanocytes are described in further detail.

Keratinocytes develop from a basal layer of cells at the epidermal-dermal junction. These

cells are in a constant state of mitosis. As more cells form, the mature keratinocytes are forced

upward through the epidermis. Keratinocytes are involved in the constant renewal of the skin as

millions of the surface cells are sloughed off daily and new keratinocytes migrate toward the skin

surface.

Melanocytes are highly dendritic cells that contact all keratinocytes; they reside in the

basal layer of the epidermis and produce skin pigment called melanin. The production of

melanin is a highly complex process involving many enzymes and cofactors (Figure 1-1) (Casp,

2003). There are two types of melanins, black-brown eumelanin and red-yellow pheomelanin,

though both have the amino acid tyrosine as their initial substrate. The enzyme tyrosinase is

responsible for the first two steps in melanin production and has been identified as the rate-









the serine residue) was observed more often in cases compared to controls; odds ratios were

significant for carrying two risk alleles and for heterozygosity for the risk allele compared to

carrying no risk alleles. This would imply that carrying the C allele at rs1042602 is protective

for vitiligo while carrying the A allele confers risk. The other significant haplotype findings

were with SNPs located in intron 2 (rs10765197, rs12791412, rs2000554), intron 3

(rs10830250), and in intron 4 (rsl827430). The only haplotype not involving rs1042602 was one

that yielded a protective effect. Homozygosity for the G allele of rs12791412 and the A allele of

rs1827430 was observed more often in controls than in cases (OR=0.3761) when compared to

homozygosity for the other haplotype (p=0.0119).

Tyrosinase-Related Protein 1

Six TYRP1 SNPs were genotyped in my study. Hardy-Weinberg equilibrium was intact

for controls but was significantly deviated for patients for four TYRP1 SNPs (Table 4-19). One

of these SNPs in H-W disequilibrium, rs2733833, was the only TYRP1 SNP to have a significant

finding in my case-control genotypic analysis (Table 4-4). This G/T polymorphism, located in

intron 6, had a significant association at the genotype level (p=.0254), where the TT genotype

was observed less often in cases than controls. Significance for this SNP was not replicated in

the TDT test (Table 4-5). Pairwise haplotype analysis of the TYRP1 gene yielded six significant

pairwise associations (Tables 4-20, 4-21, 4-22, 4-23, 4-24, and 4-25); of the 6 TYRP1 SNPs that I

genotyped, 5 had at least one significant haplotype. Four of these involved rs2733833. The

other SNPs were located in intron four (rs2762462), intron 5 (rs2733831), and three in the

3'UTR (rs683, rs2762464, and rs1063380).

Catechol-O-Methyltransferase

Two COMT SNPs, one in the 5'UTR and one in intron two, were examined in my study.

Case-control and family-based analysis of the two individual SNPs in COMTyielded no









Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy

THE ROLE OF GENETICS IN VITILIGO SUSCEPTIBILITY

By

Deborah Marsha Herbstman

August 2009

Chair: Margaret R. Wallace
Co-chair: Wayne T. McCormack
Major: Medical Sciences-Genetics

Vitiligo is an autoimmune pigment disorder of the skin, which is seen on patients as

depigmented areas that may gradually enlarge. Vitiligo is a common condition, affecting about

0.5-1% of all ethnic groups worldwide. It is also associated with an increased risk for other

autoimmune diseases. The current treatments for vitiligo are difficult, expensive, and often yield

marginal results. The cause of vitiligo is unknown, but is thought to involve both genetic and

environmental factors. The goal of my study was to test the hypothesis that vitiligo pathogenesis

is caused in part by genetic susceptibility to both autoimmune and autotoxic events due to

genetic differences in genes involved in the regulation of the immune response, melanin

production, and oxidative stress.

To identify vitiligo susceptibility genes, human genomic DNA samples from patients with

vitiligo, their family members, and healthy controls with no known autoimmune diseases were

genotyped for a number of different single nucleotide polymorphisms (SNPs) in candidate genes.

SNPs in COMT, TYR, TYRP1, DCT, and PAH, genes involved in melanin biosynthesis, were

examined. TYR, TYRP1, and DCT are also melanocyte autoantigens, which further implicates

them as vitiligo susceptibility genes. The immunoregulatory gene AIRE was tested. Genes









BIOGRAPHICAL SKETCH

Deborah Marsha Herbstman was born in Champaign-Urbana, Illinois to Drs. Joe I. and

Barbara H. D. Herbstman. She grew up in the beautiful island paradise of Gainesville, Florida

with her older brother, Joshua Tobias Herbstman. A graduate ofF.W. Buchholz High School,

she received her Associate of Arts with High Honors from the University of Florida. In 2001,

she graduated from New College, the Honors College of the State of Florida, with a degree in

biology/chemistry.









mutation or migration, are rarely met (Shoemaker et al., 1998), which has led some to question

the relevance of testing HWE in disease association studies (Zou and Donner, 2006). One group

has even suggested that the design of case-control disease association studies will cause HWD

because the populations sampled are not random but rather selected because of their disease

status, causing alleles in the study populations to deviate from their frequency in the general

population (Li and Li, 2008).

Case-Control Analysis

The case-control approach was chosen due to the very large numbers of vitiligo patients

and families required to identify disease susceptibility genes by other methods (e.g., linkage

analysis and affected sibpair analysis) (Risch and Merikangas, 1996). Case-control studies

compare the frequencies of a genotype at a locus in a patient group versus a control group. The

two groups must be matched for ethnicity and geographical origins due to variations in gene

frequencies among different populations. Although DNA samples were collected from vitiligo

patients and their family members from all racial and ethnic groups, only the group of whites

(212 independent cases and 245 unrelated controls), which are almost exclusively from the

United States, was large enough to run a case-control analysis on the genotypic data.

Allelic and genotype frequencies were calculated and analyzed for each genetic marker for

the patient and control sets. In theory, if particular alleles or genotypes have no association with

vitiligo, the frequency of that allele or genotype in the patient population should be very similar

to the frequency in the control population. An allele or genotype that confers vitiligo

susceptibility or resistance (or one that is closely linked to such an effect) would hypothetically

differ in frequency between the affected and non-affected populations. To compare the

frequency of alleles and genotypes between the case and control populations, chi-squared (X2)









Additionally, patients and family members were asked to fill out personal information surveys

requesting information such as age, sex, race, and ethnic background, age of onset, family

history, presence of other autoimmune diseases, and pattern of depigmentation (Table 2-1). A

total of 714 individuals returned blood, data, and consent for this and future studies.

Blood Processing and DNA Extraction

DNA was extracted from blood samples leukocytess) using standard phenol/chloroform

extraction methods after red blood cell lysis (Casp et al., 2002). Whole genome amplification

was performed on some samples later as needed, using a commercially available kit per

manufacturer's instructions (GE Healthcare). HIPAA de-identified DNAs from unrelated

controls with no known autoimmune diseases (n= 333) were supplied by the DNA bank of the

University of Florida's General Clinical Research Center (GCRC).

Single Nucleotide Polymorphism (SNP) Selection

Once a candidate gene was chosen based on its potential involvement in vitiligo, genetic

markers within each gene were selected. The present study used single nucleotide

polymorphisms (SNPs). All of the SNPs used were two-allele polymorphisms. It is important to

note that these SNPs might encode a functional alteration in the protein, or more likely, they may

have no obvious functional relevance to protein function or stability. In candidate gene

association analyses, SNPs function solely as genetic markers within the gene. If a

polymorphism is inherited with a disorder at a rate that exceeds random chance, this may

indicate a potential susceptibility locus. If a polymorphism is inherited by patients at a rate

below random chance, this may indicate a potential disease resistance (protective) locus. In

either case, the polymorphism genotyped may be the actual change causing the disease

susceptibility or resistance, or the SNP could be linked to the mutation(s) of interest (also known

as being in linkage disequilibrium or LD).









has been shown to be differentially expressed in vitiligo patients. In one study, COMTwas

found to be expressed at higher levels in epidermal homogenates from vitiligo patients than

homogenates from healthy controls (Le Poole et al., 1994). However, a second study found

COMT activity levels to be lower in patients with acrofacial vitiligo (Tursen et al., 2002).

Vitiligo patients show an increased epidermal de novo synthesis and recycling of 6(R)-L-

erythro-5,6,7,8-tetrahydrobiopterin (6BH4). Vitiligo patients also show very low levels of 4a-

OH-tetrahydrobiopterin dehydratase (DH) activity, which is involved in the recycling pathway of

6BH4. Due to decreased DH activity, there is an increased buildup of the 7-isomer form of

6BH4, (7BH4), which functions as an effective competitor of 6BH4 (Davis et al., 1992;

Schallreuter et al., 2001). It is the accumulation of these oxidized pterins (6- and 7-biopterin)

that results in the fluorescence of vitiligo skin under a Woods UV (351 nm) lamp, which is used

for the clinical diagnosis of vitiligo in a patient (Schallreuter et al., 2001). A buildup of 7BH4

can alter phenylalanine hydroxylase (PAH) activity. PAH activity is often decreased in vitiligo

patients, which causes a buildup of epidermal L-phenylalanine levels (Schallreuter et al., 1994b;

Schallreuter et al., 1994c). Abnormal calcium homeostasis was observed in vitiligo

keratinocytes and melanocytes consistent with oxidative stress (Schallreuter et al., 2007). This

calcium disregulation likely also contributes to the disregulation of L-phenylalanine observed in

vitiligo patients, as L-phenylalanine is coupled to the release/uptake of calcium into the cytosol

of melanocytes to initiate melanogenesis (Schallreuter and Wood, 1999).

All of these abnormal biochemical events in the vitiligo epidermis may contribute to

increased levels of hydrogen peroxide (H202). High concentrations of H202 can deactivate

catalase, an enzyme normally involved in breaking down H202 and other free oxygen radicals.

Low catalase activities have previously been reported in the epidermis of vitiligo patients









Corrections for Multiple Testing

The Bonferroni correction stringent is a statistical method of adjusting and correcting for

multiple numbers of chi-square tests. Bonferroni's correction (ap) is represented by dividing the

test-wise significance level by the number of tests, and is represented by the following equation:

ap = a / k

where a = the testwise significance level (0.05) and k = the total number of tests performed (Rice

et al., 2008). A Bonferroni corrected p-value was calculated for all findings using the SAS

software package.

More recently, the false discovery rate (FDR) has been proposed as a less conservative

correction for multiple hypothesis testing (Hochberg and Benjamini, 1990). Instead of

controlling for any false positives (as in the Bonferroni method), FDR controls the expected

proportion of false positives. The FDR threshold is determined from the observed distribution of

p-values in a given data set. FDR adjusted p-values, also known as q-values, were calculated for

all analyses using the SAS software package. Given that my sample sizes are relatively small,

positive results may not be very robust. Thus, as indicated in my tables, I report the uncorrected

and Bonferroni corrected p-values, and the FDR corrected q-values for most findings. None of

the pairwise haplotype findings showed positive associations after corrections for multiple

testing were applied because there were a total of 900 tests for this method; therefore, only

uncorrected p-values are reported. The 900 pairwise haplotype tests are from 75 pairs of genes

compared x 4 possible haplotypes for each pair x 3 comparisons of the number of risk alleles (2

vs. 1, 2 vs. 0, and 1 vs. 0).

Results

A total of 30 SNPs were genotyped from eight different genes in 332 healthy controls, 355

vitiligo patients and 359 of their unaffected family members. All races were included for TDT









Dopachrome Tautamerase

Examination of the DCT gene yielded a number of interesting results. The strongest

association with vitiligo was observed with the C allele in rs9516413. This was seen in case-

control analysis at the allelic and genotypic levels, in TDT analysis, and in pairwise haploblock

analysis with five other DCT SNPs. As this SNP was observed in all of the significant pairwise

associations, it is likely that it drove the significance for all DCThaplotype findings. The

rs9516413 SNP is located in the 3'UTR and thus could potentially affect DCTRNA stability,

polyadenylation site usage, and/or regulate transcription, perhaps by affecting an enhancer or a

miRNA binding site. Additionally, this SNP may be in linkage disequilibrium with another gene

polymorphism not examined in the present study.

Four additional intronic SNPs had significant TDT results, but did not reach significance in

case-control analysis. With a larger sample size, or with patients and controls from more diverse

ethnic backgrounds, these SNPs may prove to be significant in subsequent studies. SNPs located

in introns may affect RNA splicing and/or stability, or they may be linked to other gene changes

that directly alter gene expression or protein function.

The only SNPs in my study significantly associated with presence of co-morbid

autoimmune diseases were observed in DCT. All three of these SNPs were located in intron 6.

As with all intronic SNPs, it is difficult to determine the role that these polymorphisms play in

enzyme activity. They may be linked to an epitope encoded by DCTthat is also common to

other self-proteins. Exons five and six would be a logical place to start a search for a linked

protein-level effect such as a missense polymorphism. It is possible that vitiligo patients with

and without co-morbid autoimmune diseases represent two different sub-sets of disease

pathogenesis, though the differential role of DCT in these subgroups needs to be explored.









Su, M.A. and Anderson, M.S.: Aire: an update. Curr Opin Immunol 16 (2004) 746-52.

Sulem, P., Gudbjartsson, D.F., Stacey, S.N., Helgason, A., Rafnar, T., Magnusson, K.P.,
Manolescu, A., Karason, A., Palsson, A., Thorleifsson, G., Jakobsdottir, M., Steinberg,
S., Palsson, S., Jonasson, F., Sigurgeirsson, B., Thorisdottir, K., Ragnarsson, R.,
Benediktsdottir, K.R., Aben, K.K., Kiemeney, L.A., Olafsson, J.H., Gulcher, J., Kong,
A., Thorsteinsdottir, U. and Stefansson, K.: Genetic determinants of hair, eye and skin
pigmentation in Europeans. Nat Genet 39 (2007) 1443-52.

Tazi-Ahnini, R., Cork, M.J., Gawkrodger, D.J., Birch, M.P., Wengraf, D., McDonagh, A.J. and
Messenger, A.G.: Role of the autoimmune regulator (AIRE) gene in alopecia areata:
strong association of a potentially functional AIRE polymorphism with alopecia
universalis. Tissue Antigens 60 (2002) 489-95.

Tazi-Ahnini, R., McDonagh, A.J., Wengraf, D.A., Lovewell, T.R., Vasilopoulos, Y., Messenger,
A.G., Cork, M.J. and Gawkrodger, D.J.: The autoimmune regulator gene (AIRE) is
strongly associated with vitiligo. Br J Dermatol 159 (2008) 591-6.

Teo, Y.Y., Fry, A.E., Clark, T.G., Tai, E.S. and Seielstad, M.: On the usage of HWE for
identifying genotyping errors. Ann Hum Genet 71 (2007) 701-3; author reply 704.

The International HapMap Project. Nature 426 (2003) 789-96.

Thomas, S., Thomas, M., Wincker, P., Babarit, C., Xu, P., Speer, M.C., Munnich, A., Lyonnet,
S., Vekemans, M. and Etchevers, H.C.: Human neural crest cells display molecular and
phenotypic hallmarks of stem cells. Hum Mol Genet 17 (2008) 3411-25.

Tobin, D.J., Swanson, N.N., Pittelkow, M.R., Peters, E.M. and Schallreuter, K.U.: Melanocytes
are not absent in lesional skin of long duration vitiligo. J Pathol 191 (2000) 407-16.

Tursen, U., Kaya, T.I., Erdal, M.E., Derici, E., Gunduz, O. and Ikizoglu, G.: Association
between catechol-O-methyltransferase polymorphism and vitiligo. Arch Dermatol Res
294 (2002) 143-6.

Turunen, J.A., Wessman, M., Forsblom, C., Kilpikari, R., Parkkonen, M., Pontynen, N.,
Ilmarinen, T., Ulmanen, I., Peltonen, L. and Groop, P.H.: Association analysis of the
AIRE and insulin genes in Finnish type 1 diabetic patients. Immunogenetics 58 (2006)
331-8.

Uhm, Y.K., Yoon, S.H., Kang, I.J., Chung, J.H., Yim, S.V. and Lee, M.H.: Association of
glutathione S-transferase gene polymorphisms (GSTM1 and GSTT1) of vitiligo in
Korean population. Life Sci 81 (2007) 223-7.

van den Wijngaard, R., Wankowicz-Kalinska, A., Pals, S., Weening, J. and Das, P.: Autoimmune
melanocyte destruction in vitiligo. Lab Invest 81 (2001) 1061-7.









LIST OF REFERENCES


Abdel-Naser, M.B., Kruger-Krasagakes, S., Krasagakis, K., Gollnick, H., Abdel-Fattah, A. and
Orfanos, C.E.: Further evidence for involvement of both cell mediated and humoral
immunity in generalized vitiligo. Pigment Cell Res 7 (1994) 1-8.

Abdel-Naser, M.B., Ludwig, W.D., Gollnick, H. and Orfanos, C.E.: Nonsegmental vitiligo:
decrease of the CD45RA+ T-cell subset and evidence for peripheral T-cell activation. Int
JDermatol 31 (1992)321-6.

A haplotype map of the human genome. Nature 437 (2005) 1299-320.

Akhtar, S., Gavalas, N.G., Gawkrodger, D.J., Watson, P.F., Weetman, A.P. and Kemp, E.H.: An
insertion/deletion polymorphism in the gene encoding angiotensin converting enzyme is
not associated with generalised vitiligo in an English population. Arch Dermatol Res 297
(2005) 94-8.

al Badri, A.M., Foulis, A.K., Todd, P.M., Gariouch, J.J., Gudgeon, J.E., Stewart, D.G., Gracie,
J.A. and Goudie, R.B.: Abnormal expression of MHC class II and ICAM-1 by
melanocytes in vitiligo. J Pathol 169 (1993) 203-6.

Alkhateeb, A., Fain, P.R. and Spritz, R.A.: Candidate functional promoter variant in the FOXD3
melanoblast developmental regulator gene in autosomal dominant vitiligo. J Invest
Dermatol 125 (2005) 388-91.

Alkhateeb, A., Stetler, G.L., Old, W., Talbert, J., Uhlhorn, C., Taylor, M., Fox, A., Miller, C.,
Dills, D.G., Ridgway, E.C., Bennett, D.C., Fain, P.R. and Spritz, R.A.: Mapping of an
autoimmunity susceptibility locus (AIS1) to chromosome lp31.3-p32.2. Hum Mol Genet
11 (2002) 661-7.

An autoimmune disease, APECED, caused by mutations in a novel gene featuring two PHD-type
zinc-finger domains. The Finnish-German APECED Consortium. Autoimmune
Polyendocrinopathy-Candidiasis-Ectodermal Dystrophy. Nat Genet 17 (1997) 399-403.

Anderson, M.S., Venanzi, E.S., Klein, L., Chen, Z., Berzins, S.P., Turley, S.J., von Boehmer, H.,
Bronson, R., Dierich, A., Benoist, C. and Mathis, D.: Projection of an immunological self
shadow within the thymus by the AIRE protein. Science 298 (2002) 1395-401.

Arcos-Burgos, M., Parodi, E., Salgar, M., Bedoya, E., Builes, J., Jaramillo, D., Ceballos, G.,
Uribe, A., Rivera, N., Rivera, D., Fonseca, I., Camargo, M. and Palacio, G.: Vitiligo:
complex segregation and linkage disequilibrium analyses with respect to microsatellite
loci spanning the HLA. Hum Genet 110 (2002) 334-42.

Arrunategui, A., Arroyo, C., Garcia, L., Covelli, C., Escobar, C., Carrascal, E. and Falabella, R.:
Melanocyte reservoir in vitiligo. Int J Dermatol 33 (1994) 484-7.









greater frequency difference between the European and African samples, and only 0.37% show a

greater frequency difference between the European and East Asian samples than rs1042602.

In five TYR pairwise haplotypes, I found that homozygosity for the C allele of rsl042602

(coding for serine) was significantly associated with vitiligo. It is unknown whether the amino

acid change of a serine to a tyrosine affects protein function, although it is a non-conservative

substitution and this residue could affect the active sites of this protein (although possible

structure changes are not known). Alternatively, these risk haplotypes may be in linkage

disequilibrium with another TYR change that is directly related to vitiligo pathogenesis. The

protective haplotype I observed in TYR involved two intronic SNPs.

Tyrosinase-Related Protein 1

Whereas the exact function of tyrosine-related protein 1 in melanogenesis is still debated, it

has been shown to enhance tyrosinase function (Kobayashi et al., 1994; Kobayashi et al., 1998).

Additionally, mutations severely disrupting TYRP1 enzymatic function are known to cause

human albinism (Oetting and King, 1999). I examined six TYRP1 SNPs in my study,

hypothesizing that subtle variation in the TYRP1 gene may contribute to vitiligo pathogenesis.

I observed four SNPs to be in Hardy-Weinberg disequilibrium (HWD) for cases but not

controls. Interpreting HWD remains controversial (Salanti et al., 2005; Zou and Donner, 2006;

Teo et al., 2007; Li and Li, 2008), but some groups suggest that a situation such as this is

supportive of genetic association because it implies that the allele frequencies are stable in

controls but changing in cases. Only one TYRP1 SNP had a positive finding in independent

analysis at the genotypic level. Located in intron 6, rs2733833, one of the SNPs in HWD, had a

significant association for genotype frequencies. Specifically, the TT genotype was observed

less often in cases than controls suggesting a possible protective effect. However, because this









limiting factor. Melanin is produced in membrane-bound granules called melanosomes.

Melanosomes migrate from the center of the melanocyte cell body to the end of the dendrites and

are deposited into keratinocytes (Nordlund et al., 1998). The melanosomes accumulate in the

keratinocytes and form a shield of melanin, which provides the skin with protection against

ultraviolet radiation from sunlight.

All humans have relatively the same quantity of melanocytes, so different skin

pigmentations are accounted for by variations in melanocyte activity or the rapidity of melanin

breakdown in keratinocytes. The melanocytes of dark brown- and black-skinned individuals

produce more melanin of a darker color than do the melanocytes of fair-skinned individuals.

Freckles and moles are localized areas of increased melanin production. Melanin production can

be stimulated by ultraviolet rays in sunlight, which results in the darkening of the skin; this is

commonly referred to as tanning. This increase in melanin augments the skin's protection

against the ultraviolet radiation damage that can cause skin cancer. The hypopigmented lesions

in vitiligo patients are a result of the destruction and/or inactivation of the melanin-producing

melanocytes. Keratinocytes still migrate to the surface of the epithelium, albeit without their

cargo of pigment. This results in patches of skin that look milky-white because they are devoid

of pigment.

Clinical Description of Vitiligo

There are two main types of vitiligo, segmental and non-segmental (also known as

generalized vitiligo), and classification relies on the distribution of hypopigmented lesions. In

segmental vitiligo, the areas of depigmentation are random and often occur on only one location

on the body (Nordlund and Ortonne, 1998). Segmental vitiligo sometimes involves the loss of

follicular melanocytes and a loss of hair pigment. Segmental vitiligo is less likely to repigment

than non-segmental vitiligo; however, it is also less likely to spread to other areas of the body.









Table 2-5. Family-based association (transmission disequilibrium test) results for FBXOll and
MSH6
Number of Not
informative Transmitted transmitted BonC FDR
Gene SNP parents (allele) (allele) p-value p-value q-value
FBX011 rs960106 75 35 (C) 40 (T) 0.56 1.00 0.84
FBXO011
FB6 rs3136367 45 17 (C) 28 (G) 0.10 1.00 0.43
SH6BonC denotes the Bonferroni corrected p-value
BonC denotes the Bonferroni corrected p-value









Jin, S.Y., Park, H.H., Li, G.Z., Lee, H.J., Hong, M.S., Hong, S.J., Park, H.K., Chung, J.H. and
Lee, M.H.: Association of angiotensin converting enzyme gene I/D polymorphism of
vitiligo in Korean population. Pigment Cell Res 17 (2004a) 84-6.

Jin, S.Y., Park, H.H., Li, G.Z., Lee, H.J., Hong, M.S., Park, H.J., Park, H.K., Seo, J.C., Yim,
S.V., Chung, J.H. and Lee, M.H.: Association of estrogen receptor 1 intron 1 C/T
polymorphism in Korean vitiligo patients. J Dermatol Sci 35 (2004b) 181-6.

Jin, Y., Bennett, D.C., Amadi-Myers, A., Holland, P., Riccardi, S.L., Gowan, K., Fain, P.R. and
Spritz, R.A.: Vitiligo-associated multiple autoimmune disease is not associated with
genetic variation in AIRE. Pigment Cell Res 20 (2007a) 402-4.

Jin, Y., Mailloux, C.M., Gowan, K., Riccardi, S.L., LaBerge, G., Bennett, D.C., Fain, P.R. and
Spritz, R.A.: NALP1 in vitiligo-associated multiple autoimmune disease. N Engl J Med
356 (2007b) 1216-25.

Kang, C.P., Lee, H.S., Ju, H., Cho, H., Kang, C. and Bae, S.C.: A functional haplotype of the
PADI4 gene associated with increased rheumatoid arthritis susceptibility in Koreans.
Arthritis Rheum 54 (2006) 90-6.

Kao, C.H. and Yu, H.S.: Depletion and repopulation of Langerhans cells in nonsegmental type
vitiligo. J Dermatol 17 (1990) 287-96.

Kemp, E.H., Ajjan, R.A., Waterman, E.A., Gawkrodger, D.J., Cork, M.J., Watson, P.F. and
Weetman, A.P.: Analysis of a microsatellite polymorphism of the cytotoxic T-
lymphocyte antigen-4 gene in patients with vitiligo. Br J Dermatol 140 (1999a) 73-8.

Kemp, E.H., Gawkrodger, D.J., MacNeil, S., Watson, P.F. and Weetman, A.P.: Detection of
tyrosinase autoantibodies in patients with vitiligo using 35S-labeled recombinant human
tyrosinase in a radioimmunoassay. J Invest Dermatol 109 (1997a) 69-73.

Kemp, E.H., Gawkrodger, D.J., Watson, P.F. and Weetman, A.P.: Immunoprecipitation of
melanogenic enzyme autoantigens with vitiligo sera: evidence for cross-reactive
autoantibodies to tyrosinase and tyrosinase-related protein-2 (TRP-2). Clin Exp Immunol
109 (1997b) 495-500.

Kemp, E.H., Waterman, E.A., Gawkrodger, D.J., Watson, P.F. and Weetman, A.P.:
Autoantibodies to tyrosinase-related protein-1 detected in the sera of vitiligo patients
using a quantitative radiobinding assay. Br J Dermatol 139 (1998) 798-805.

Kemp, E.H., Waterman, E.A., Gawkrodger, D.J., Watson, P.F. and Weetman, A.P.:
Identification of epitopes on tyrosinase which are recognized by autoantibodies from
patients with vitiligo. J Invest Dermatol 113 (1999b) 267-71.

Kemp, E.H., Waterman, E.A., Hawes, B.E., O'Neill, K., Gottumukkala, R.V., Gawkrodger, D.J.,
Weetman, A.P. and Watson, P.F.: The melanin-concentrating hormone receptor 1, a
novel target of autoantibody responses in vitiligo. J Clin Invest 109 (2002) 923-30.









Table 4-4. Case-control results for genotype analysis ofDCT, PAH, TYR, TYRP1, and COMT
BonC FDR
Gene SNP p-value p-value q-value
DCT rs9516413 0.0270* 0.81 0.27
DCT rs7987802 0.57 1.00 0.63
DCT rs4318084 0.13 1.00 0.39
DCT rs9524493 0.55 1.00 0.63
DCT rs9516418 0.98 1.00 0.98
DCT rs1028805 0.87 1.00 0.93
DCT rs11618471 0.35 1.00 0.55
DCT rs7991232 0.46 1.00 0.55
PAH rs1722381 0.44 1.00 0.55
PAH rs1522307 0.39 1.00 0.55
TYR rs1042602 0.31 1.00 0.55
TYR rs621313 0.45 1.00 0.55
TYR rs594647 0.34 1.00 0.55
TYR rs10765197 0.32 1.00 0.55
TYR rs12791412 0.13 1.00 0.39
TYR rs2000554 0.24 1.00 0.55
TYR rs10830250 0.31 1.00 0.55
TYR rs1827430 0.12 1.00 0.39
TYRP1 rs2762462 0.12 1.00 0.39
TYRP1 rs2733831 0.41 1.00 0.55
TYRP1 rs2733833 0.0254* 0.76 0.27
TYRP1 rs683 0.11 1.00 0.39
TYRP1 rs2762464 0.06 1.00 0.39
TYRP1 rs1063380 0.10 1.00 0.39
COMT rs2020917 0.18 1.00 0.49
COMT rs6269 0.39 1.00 0.55
*denotes a significant finding; BonC denotes the Bonferroni corrected p-value









CHAPTER 4
MELANIN BIOSYNTHESIS GENES

Introduction

Melanin biosynthesis is a complex process that takes place in specialized organelles called

melanosomes. Approximately 1500 proteins have been identified in melanosomes during of all

stages of their maturation, with about 600 expressed in any given stage. Approximately 100

proteins are shared by melanosomes from pigmented and nonpigmented melanocytes, and this

number likely represents the essential melanosome proteome (Chi et al., 2006). Several genes

involved in melanin biosynthesis were examined in the present study to determine their

association with vitiligo. All genes examined showed at least one positive result and are

discussed in this chapter. The involvement of these genes could be supportive of the autoimmune

and/or oxidative stress theories of vitiligo pathogenesis.

Whereas lack of phenylalanine hydroxylase is most commonly associated with the

autosomal recessive disorder phenylketonuria (PKU), the role of this enzyme in melanin

biosynthesis has been supported by a number of studies (Camacho and Mazuecos, 1999;

Schallreuter and Wood, 1999; Schallreuter et al., 2005). The PAH gene is located on

chromosome 12q22 and has 13 exons that encode the 453 amino acid enzyme phenylalanine

hydroxylase. This enzyme converts the essential amino acid phenylalanine into tyrosine, the

precursor of melanin. A disruption or a complete loss of PAH activity in the liver results in a

dramatic increase in serum concentrations of phenylalanine in PKU patients. PKU patients must

eat a special diet lacking phenylalanine because excess accumulation of this amino acid has

detrimental neurological effects. In addition to neurological impairment, untreated PKU patients

have poor pigmentation, adding support to the hypothesis that PAH protein function is vital to

melanogenesis. The role of PAH in producing tyrosine is consistent with the notion that









Buc, M., Fazekasova, H., Cechova, E., Hegyi, E., Kolibasova, K. and Ferencik, S.: Occurrence
rates of HLA-DRB 1, HLA-DQB 1, and HLA-DPB 1 alleles in patients suffering from
vitiligo. Eur J Dermatol 8 (1998) 13-5.

Buckley, D.A. and du Vivier, A.W.: Topical immunotherapy in dermatology. Int J Clin Pract 53
(1999) 130-7.

Bystryn, J.C.: Immune mechanisms in vitiligo. Clin Dermatol 15 (1997) 853-61.

Camacho, F. and Mazuecos, J.: Treatment of vitiligo with oral and topical phenylalanine: 6 years
of experience. Arch Dermatol 135 (1999) 216-7.

Cardon, L.R. and Bell, J.I.: Association study designs for complex diseases. Nat Rev Genet 2
(2001) 91-9.

Casp, C.B.: Genetic association of catalase and antigen processing genes with vitiligo
susceptibility, Pathology, Immunology, and Laboratory Medicine. University of Florida,
Gainesville, (2003). 3.

Casp, C.B., She, J.X. and McCormack, W.T.: Genetic association of the catalase gene (CAT)
with vitiligo susceptibility. Pigment Cell Res 15 (2002) 62-6.

Casp, C.B., She, J.X. and McCormack, W.T.: Genes of the LMP/TAP cluster are associated with
the human autoimmune disease vitiligo. Genes Immun 4 (2003) 492-9.

Cetani, F., Barbesino, G., Borsari, S., Pardi, E., Cianferotti, L., Pinchera, A. and Marcocci, C.: A
novel mutation of the autoimmune regulator gene in an Italian kindred with autoimmune
polyendocrinopathy-candidiasis-ectodermal dystrophy, acting in a dominant fashion and
strongly cosegregating with hypothyroid autoimmune thyroiditis. J Clin Endocrinol
Metab 86 (2001) 4747-52.

Chen, Y.F., Yang, P.Y., Hu, D.N., Kuo, F.S., Hung, C.S. and Hung, C.M.: Treatment of vitiligo
by transplantation of cultured pure melanocyte suspension: analysis of 120 cases. J Am
Acad Dermatol 51 (2004) 68-74.

Chi, A., Valencia, J.C., Hu, Z.Z., Watabe, H., Yamaguchi, H., Mangini, N.J., Huang, H.,
Canfield, V.A., Cheng, K.C., Yang, F., Abe, R., Yamagishi, S., Shabanowitz, J., Hearing,
V.J., Wu, C., Appella, E. and Hunt, D.F.: Proteomic and bioinformatic characterization of
the biogenesis and function of melanosomes. J Proteome Res 5 (2006) 3135-44.

Cui, J., Arita, Y. and Bystryn, J.C.: Cytolytic antibodies to melanocytes in vitiligo. J Invest
Dermatol 100 (1993) 812-5.

Cui, J. and Bystryn, J.C.: Melanoma and vitiligo are associated with antibody responses to
similar antigens on pigment cells. Arch Dermatol 131 (1995) 314-8.










Table 4-27. Summary of all findings for DCT, PAH, TYR, TYRP1, and COMT
Base pair Co- Age
change & Case/ morbid of Haplo-
Gene SNP ID location Control TDT HWE AAD onset type
DCT rs9516413 C/T 3'UTR A, G + 5
DCT rs7987802 C/T intron 6 + 1
DCT rs4318084 C/T intron 6 + + 1
DCT rs9524493 C/G intron 6 + 1
DCT rs9516418 C/T intron 6 +
DCT rs1028805 G/T intron 2 +
DCT rs 11618471 A/G intron 2 + 1
DCT rs7991232 A/G intron 1 1
PAH rs1722381 A/G intron 6 1
PAH rs1522307 A/G intron 2 1
TYR rs 1042602 C/A exon 1, 5
(S192Y)
TYR rs621313 G/A intron 1
TYR rs594647 T/C intron 1
TYR rs10765197 A/C intron 2 1
TYR rs12791412 A/G intron 2 A + + 2
TYR rs2000554 G/A intron 2 1
TYR rs10830250 C/G intron 3 1
TYR rs1827430 G/A intron 4 2
TYRP1 rs2762462 C/T intron 4 D 1
TYRP1 rs2733831 A/G intron 5 3
TYRP1 rs2733833 G/T intron 6 G D 4
TYRP1 rs683 A/C 3'UTR D 2
TYRP1 rs2762464 A/T 3'UTR D 2
TYRP1 rs1063380 C/T 3'UTR
COMT rs2020917 T/C 5'UTR 1
COMT rs6269 G/A intron 2 1
A, significant findings on the allelic level; G, significant findings on the genotypic level; AAD,
autoimmune disease; +, significant result; D, in Hardy-Weinberg disequilibrium; Pairwise
haplotype, number of significant pairwise associations observed









family-based association studies. Alternatively, it is possible that this is a spurious finding;

notably, our control allele frequency for rs1800520 differed from the ABI reported frequency.

Pairwise haplotype analysis of this gene was also significant, though the addition of other SNPs

on this gene may help to understand the mechanism underlying this finding.









Corrections for M multiple Testing ......................................................... ............... 41
R e su lts ............... ....... ......... ....................................................................................................4 1
G enotype by Sex A analyses ................................................ ...................... ............... 42
FBXO and MSH6 ........................ ............. .............. 42
D isc u ssio n ....................... ... ............. ... ...........................................................4 3

3 AUTOIMMUNE REGULATOR GENE.......................... ............................ 49

Intro du action ....................... ....... .............. ....49..........
M materials and M methods ...................................... .. ......... ......... .....50
R e su lts ........................... ............. ... ........................................................... 5 1
D isc u ssio n ....................... ... .. ......... ... ...........................................................5 2

4 M ELAN IN BIO SYN THESIS GENES............................................................................. ...58

In tro d u ctio n ................... ...................5...................8..........
M materials and M methods ...................................... .. ......... ......... .....61
R e su lts ................................ ....................................................................................................6 1
Phenylalanine H ydroxylase ...................... .......... ..... ..................... ...............61
D opachrom e Tautam erase ........................................................................... 61
T y ro sin a se .............. ....... .......... .....................................................................6 2
T yrosinase-R elated P protein 1 .......................................................... ..........................63
C atechol-O -M ethyltransferase ............................................................. .....................63
S u m m a ry ................................................................................6 4
D isc u s sio n ...............................................................................................................................6 4
Phenylalanine H ydroxylase ............... ........... ............ ..................... ...............64
D opachrom e Tautam erase ............................ ..................................... ...............65
T y ro sin a se .............. ....... .......... .....................................................................6 6
T yrosinase-R elated P protein 1 .......................................................... ..........................67
C atechol-O -M ethyltransferase ............................................................. .....................68

5 CONCLUSIONS AND FUTURE DIRECTIONS ...................................... ............... 85

F B X O an d M SH 6 ............. ...... ........ ...... ...................................... .......... .. .. ..85
A utoim m une R egulator (A IR E ) ............................................... ......................................... 86
M elanin B iosynthesis G enes.......................................................................... ................... 87
Final Thoughts and a N ote of Caution ..................................................... ...................90
F future D directions .............................................................................92

APPENDIX

A SINGLE NUCLEOTIDE POLYMORPHISM DETAILS ............................................. 95

L IST O F R E F E R E N C E S ........................ ...... ...........................................................................96

B IO G R A PH IC A L SK E TCH ................................... ............. .................................................111




8









encouragement and lots of good cheese. I am grateful for Tarek Saab, whose unique charm and

good humor left a mark on me. If I have seen a little further it is by standing on the shoulders of

astronomers because they look at things that are really, really far away. It has been a pleasure to

be a groupie to UF's Astronomy department; a nicer and smarter group of people it would be

difficult to find.

Saving the best for last, I thank my family. I appreciate their love, their patience, their

exquisite humor, and their exceptional DNA. My Father was my biggest fan and always

encouraged me to excel in academics. I am grateful to Joshua Tobias Herbstman; I could not ask

for a more supportive big brother. My Mother is my role model, my best friend, and the one who

always thought that I would make a good scientist.









Sequencing AIRE in vitiligo patients may yield additional information about linked variants and

this gene's role in vitiligo etiology. RNA analysis could reveal whether the susceptibility

haplotype has altered mRNA levels.

Melanin Biosynthesis Genes

I tested five genes involved in melanin biosynthesis to examine their role in vitiligo

pathogenesis: tyrosinase (TYR), tyrosinase-related protein 1 (TYRP1), dopachrome-tautomerase

(DCT), phenylalanine hydroxylase (PAH), and catechol-O-methyltransferase (COMT). Three of

these genes (DCT, TYR, and TYRP1) code for proteins that have been observed to be the targets

of autoantibodies in the sera of vitiligo patients. There may be a genetic component underlying

this observed autoimmunity, as seen in other autoimmune diseases like type I diabetes (Redondo

et al., 2001). A change in the DCT, TYR, and/or TYRP1 genes that results in an altered protein

may create an epitope, or reveal a cryptic epitope by affecting protein structure, which may lead

to inappropriate targeting of melanocytes by the immune system.

Reactive oxygen species (ROS) are generated and carefully regulated during each step of

melanin biosynthesis by proteins such as catalase and catechol-O-methyltransferase. A change

in enzymatic function during melanogenesis may cause an imbalance in the melanin

pigmentation cycle, leading to a build-up of toxic intermediates that would cause free radical

damage to the melanocyte. It is with this in mind that I tested the PAH, DCT, TYR, and TYRP1

genes because they play a vital role in melanin production. Furthermore, I tested COMTbecause

its protein acts to prevent toxic build-up of intermediate o-quinones in the melanocyte during

melanogenesis, and it may be involved in the regulation of melanin biosynthesis.

The only significant results observed in the PAH gene were in the haplotype analysis of

two intronic SNPs. The risk haplotype I observed was found exclusively in vitiligo cases but not

in controls. This risk haplotype may be indicate splice site errors in the introns that I tested or it


















phenylalanine tyrosine
phenylalanine hydrxylase


6-BH4


7-BH4


4a-OH-BH4


SNAD

NADH H20


-- q-BH2 *
vlase I


TRP-2 TRP-1
'_ dopaquinone >. eumelanins
tyrosinase DHI-2C DHI (brown/black)


\ 02 \-
\6-BH4


pheomelanins
(yellow/red)


L-dopa



dopamine


HNF1/4


norepinephrine


SPNPMT

epinephrine


Figure 1-1. Melanin biosynthesis pathway (Casp, 2003).


P35


+ /
G /d
GTP-cyclohydi



A









analyses. This would suggest that, at least in my study population, they are not vitiligo

susceptibility genes.

It is important to note that vitiligo is most likely a disease with a complex etiology.

Therefore, susceptibility genes that are uncommon in a population, or genes that have only a

subtle contribution to the vitiligo phenotype, may not show strong associations in case-control

analyses. It is with this in mind that I dismiss these genes as possessing no currently

demonstrable genetic association with vitiligo in my sample population.









Vitiligo spreads either through the expansion of existing macules or with the onset of new areas

of depigmentation (Nordlund and Ortonne, 1998).

Non-segmental vitiligo can appear anywhere on the body, but it most frequently appears

symmetrically affecting the hands, mouth, eyes, elbows, and/or knees (Lerner and Nordlund,

1978). It is unknown why an instigating event on one side of the body can trigger vitiligo not

only at the site of injury but also at the same site on the opposite side of the body. In contrast to

segmental vitiligo, non-segmental disease often does not affect the hair follicles in the areas of

depigmentation. Non-segmental vitiligo often gets progressively worse, with lesions enlarging

and more areas of the body being affected over time. Some patients with non-segmental vitiligo

almost completely depigment over the years; this condition is referred to as universal vitiligo. In

contrast to segmental disease, non-segmental vitiligo sometimes shows spontaneous

repigmentation, or partial repigmentation with treatment; however, the patient rarely fully returns

to pre-disease pigmentation. Because segmental and non-segmental vitiligo present differently

clinically, they may have different pathophysiologies (Nordlund and Ortonne, 1998; Bos, 2005).

Clinicians and researchers sometimes distinguish between "active" and "stable" vitiligo,

though these definitions are not universally applied (Yu et al., 1993; Mulekar, 2003; Chen et al.,

2004; Lotti et al., 2008). Generally, patients described as having active vitiligo are those whose

lesions have spread within the last three months. Patients are said to have stable vitiligo when

their lesions have not changed (depigmenting or repigmenting) within the past three-month to

two-year period.

Vitiligo Pathogenesis

Two of the most popular theories for the pathogenesis of vitiligo include: 1) the

autoimmune hypothesis and 2) and the autocytotoxic hypothesis. These theories are not









Schallreuter, K.U., Wood, J.M., Pittelkow, M.R., Buttner, G., Swanson, N., Korner, C. and
Ehrke, C.: Increased monoamine oxidase A activity in the epidermis of patients with
vitiligo. Arch Dermatol Res 288 (1996) 14-8.

Schallreuter, K.U., Wood, J.M., Pittelkow, M.R., Gutlich, M., Lemke, K.R., Rodl, W., Swanson,
N.N., Hitzemann, K. and Ziegler, I.: Regulation of melanin biosynthesis in the human
epidermis by tetrahydrobiopterin. Science 263 (1994b) 1444-6.

Schallreuter, K.U., Wood, J.M., Ziegler, I., Lemke, K.R., Pittelkow, M.R., Lindsey, N.J. and
Gutlich, M.: Defective tetrahydrobiopterin and catecholamine biosynthesis in the
depigmentation disorder vitiligo. Biochim Biophys Acta 1226 (1994c) 181-92.

Schallreuter, K.U., Zschiesche, M., Moore, J., Panske, A., Hibberts, N.A., Herrmann, F.H.,
Metelmann, H.R. and Sawatzki, J.: In vivo evidence for compromised phenylalanine
metabolism in vitiligo. Biochem Biophys Res Commun 243 (1998) 395-9.

Shoemaker, J., Painter, I. and Weir, B.S.: A Bayesian characterization of Hardy-Weinberg
disequilibrium. Genetics 149 (1998) 2079-88.

Shriver, M.D., Parra, E.J., Dios, S., Bonilla, C., Norton, H., Jovel, C., Pfaff, C., Jones, C.,
Massac, A., Cameron, N., Baron, A., Jackson, T., Argyropoulos, G., Jin, L., Hoggart,
C.J., McKeigue, P.M. and Kittles, R.A.: Skin pigmentation, biogeographical ancestry and
admixture mapping. Hum Genet 112 (2003) 387-99.

Smyth, D.J., Howson, J.M., Payne, F., Maier, L.M., Bailey, R., Holland, K., Lowe, C.E., Cooper,
J.D., Hulme, J.S., Vella, A., Dalhman, I., Lam, A.C., Nutland, S., Walker, N.M., Twells,
R.C. and Todd, J.A.: Analysis of polymorphisms in 16 genes in type 1 diabetes that have
been associated with other immune-mediated diseases. BMC Med Genet 7 (2006) 20.

Song, Y.H., Connor, E., Li, Y., Zorovich, B., Balducci, P. and Maclaren, N.: The role of
tyrosinase in autoimmune vitiligo. Lancet 344 (1994) 1049-52.

Spielman, R.S. and Ewens, W.J.: The TDT and other family-based tests for linkage
disequilibrium and association. Am J Hum Genet 59 (1996) 983-9.

Spielman, R.S., McGinnis, R.E. and Ewens, W.J.: Transmission test for linkage disequilibrium:
the insulin gene region and insulin-dependent diabetes mellitus (IDDM). Am J Hum
Genet 52 (1993) 506-16.

Spritz, R.A.: The genetics of generalized vitiligo and associated autoimmune diseases. J
Dermatol Sci 41 (2006) 3-10.

Spritz, R.A.: The genetics of generalized vitiligo and associated autoimmune diseases. Pigment
Cell Res 20 (2007) 271-8.

Stokowski, R.P., Pant, P.V., Dadd, T., Fereday, A., Hinds, D.A., Jarman, C., Filsell, W., Ginger,
R.S., Green, M.R., van der Ouderaa, F.J. and Cox, D.R.: A genomewide association study
of skin pigmentation in a South Asian population. Am J Hum Genet 81 (2007) 1119-32.









Because vitiligo is one of the autoimmune diseases associated with APECED, it was chosen as a

candidate gene. Further discussion of this gene and the results from the present study are found

in Chapter 3.

Melanin Biosynthesis Genes

Five genes critical to the melanin biosynthesis pathway were selected for this study:

tyrosinase (TYR), tyrosinase-related protein 1 (TYRP1), dopachrome-tautomerase (DCT),

phenylalanine hydroxylase, (PAH), and catechol-O-methyltransferase (COMT). These genes,

several of whose proteins have been observed to be melanocyte autoantigens, may be vitiligo

susceptibility genes. Further discussion of this gene set and the results from the present study are

found in Chapter 4.

DNA Repair Genes

Previous studies of vitiligo patients' melanocytes have shown subtle changes in the

expression levels of genes involved in DNA repair. Problems with DNA repair mechanisms may

lead to cell damage resulting in immune cell targeting of melanocytes. Genes examined in the

present study that are linked to DNA repair include F-box only protein 11 or FBXO11 and the

DNA G/T repair protein MutS, E. coli, homolog 6 gene orMSH6. Further discussion of these

genes and the results from the present study are found later in this chapter.

Materials and Methods

Subjects

Previous work in the McCormack lab, which formed the foundation of the current project,

involved establishing an internet website for both the University of Florida and at the National

Vitiligo Foundation to recruit subjects for my study (Casp et al., 2002; Casp et al., 2003). Kits

were mailed to interested patients and family members containing three 10 mL EDTA blood

collection tubes and IRB consent forms (project numbers 130-1998 and 416-1998).









Anderson, 2004; Mathis and Benoist, 2009). Due to this expression pattern, it is thought that the

AIRE protein may be involved in the maintenance of thymic tolerance, where AIRE up-regulates

the expression of organ-specific antigens to facilitate the negative selection of autoreactive T-

cells, thus enhancing self-tolerance and limiting autoimmunity (Anderson et al., 2002). It is

possible that detrimental mutations in AIRE may be responsible for incomplete negative selection

of self-antigens, resulting in the eventual development of multiple autoimmune disorders

(Pitkanen and Peterson, 2003; Liston et al., 2004; Liston and Goodnow, 2005). However,

several studies searching for mutations in the AIRE gene leading to non-APECED autoimmune

disease have not yielded significant findings (Turunen et al., 2006; Jin et al., 2007a; Boe Wolff et

al., 2008). SNPs within the AIRE gene were examined in my study to test if subtle alterations in

this gene's function may contribute to the autoimmune pathogenesis of vitiligo.

Materials and Methods

DNA collection and SNP genotyping were performed as described in Chapter 2. Statistical

analyses were performed as detailed in Chapter 2. Two SNPs on the AIRE gene were chosen to

test the hypothesis that this immunoregulatory gene is a vitiligo susceptibility gene (Table 3-1).

The rs2776377 SNP, located 5' upstream of AIRE but on the same haplotype block as the 5'

coding sequence, with high MAF in Caucasians in the Applied Biosystems validation set, was

examined. Additionally, rs1800520, a SNP in exon five on the same haplotype block as

rs2776377, was chosen because it causes a nonsynonymous change of a polar amino acid

serinee) to a charged residue (arginine) (S278R) in the SAND domain of the AIRE protein. This

may have an effect on the postulated AIRE DNA-binding activity (Tazi-Ahnini et al., 2002;

Purohit et al., 2005).









Table 2-4. Minor allele frequencies (MAF) for FBXOll and MSH6 SNPs from the HapMap project Caucasians from European
Ancestry (HapMap CEPH), Applied Biosystems Caucasian cohort (ABI Caucasian), and values for the white case and
control groups from the present study; X2 p-values and q-values for allelic comparisons are given
HapMap ABI # of White # of
Gene CEPH Caucasian White case white control white BonC FDR
Symbol SNP ID MAF MAF MAF cases MAF controls p-value p-value q-value
FBXOll rs960106 0.38 (C) 0.4 (C) 0.42 (C) 208 0.42 (C) 239 0.95 1.00 0.95
MSH6 rs3136367 0.22 (C) 0.27 (C) 0.26 (C) 188 0.30 (C) 238 0.13 1.00 0.65
BonC denotes the Bonferroni corrected p-value









(Schallreuter et al., 1991), though whether this defect is due to an increase in H202 from the

defective 6BH4 pathway or due to a separate problem in the catalase enzyme is unknown.

Additional studies did not demonstrate a reduction of CATmRNA in vitiligo patients. What has

been shown is that treatment with pseudocatalase, a bis-manganese III-EDTA- (HC03)2 synthetic

catalase substitute, has promoted repigmentation in some vitiligo patients, along with restoring

DH enzyme activity and a return to normal 7BH4 levels in the epidermis (Schallreuter et al.,

2001). Additionally, polymorphisms in the CATgene have been shown to be associated to

vitiligo susceptibility in several studies (Casp et al., 2002; Park et al., 2006; Em et al., 2007).

Another possible contributing factor to onset of vitiligo may be an inherent cytological

defect reported for melanocytes of vitiligo patients. Vitiligo melanocytes have been shown to

have ultrastructural defects including an abnormal dilation of the rough endoplasmic reticulum

(RER), circular RER profiles, and membrane-bound melanosome compartments seen on

transmission electron microscope examination (Boissy et al., 1991; Im et al., 1994). Further

observation of this phenomenon is seen in an immortalized vitiligo cell line where this abnormal

dilation is thought to be caused by a retention of a variety of proteins in the RER (Le Poole et al.,

2000). Although not cytotoxic to melanocytes in vitro, such defects might contribute to

melanocyte destruction in vivo. Morphological changes in melanocyte RER, as seen by electron

microscopy, are suggestive of retention of peptides in the RER (Boissy et al., 1991; Im et al.,

1994).

Alterations affecting the functional integrity in the protein tyrosinase, the rate-limiting

enzyme in melanin synthesis, have been hypothesized to play an important part in the

autocytotoxic model of vitiligo. If tyrosinase is damaged or misfolded, it is retained in the

endoplasmic reticulum. These malformed proteins are then ubiquitin-tagged and targeted for









degradation via the proteasome (Halaban et al., 2002). If oxidative stress causes necrosis of these

cells retaining misfolded tyrosinase, a large amount of the protein may be released into the

inflammatory microenvironment potentially to be picked up by antigen-presenting cells (APCs).

It is also possible that ROS are modifying these retained proteins, allowing the presentation of

cryptic epitopes by either APCs or by the melanocytes themselves, which have been shown to

phagocytize and present antigen in times of stress through MHC class II (Le Poole et al., 1993a;

1993b). Recruitment and involvement of other cells of the immune system can function to

increase oxidative stress. Secretion of chemokines and cytokines into the local environment can

increase inflammation, and potentially target cells for further destruction by necrosis and

apoptosis. Dead and dying cells may accumulate in the epidermis, drawing the attention of

circulating monocytes. This may further permit an increased presentation of self-peptides to

cells of the immune system. Thus, the autocytotoxicity of melanocytes may lead to

autoimmunity. My work has focused on several genes in this pathway including catechol-O-

methyl transferase (COMT), tyrosinase (TYR), tyrosinase- related protein 1 (TYRP1), and

dopachrome tautamerase (DCT).

Additionally, a number of studies have found altered expression for DNA repair and

antioxidant genes. The MutS E. coli Homolog 6 orMSH6 [formerly known as G/T mismatch-

binding protein (GTBP)] gene is involved in G/T DNA repair, and its upregulation in vitiligo

patients may be indicative of increased DNA damage (Le Poole et al., 2001). Although the exact

function of the F-Box Only Protein 11 (FBXO11) gene [formerly known as vitiligo-associated

protein-1 (VIT1)] is unknown, it does share homology to the MSH6 gene, and it may be involved

in the oxidative stress pathway either indirectly, or by regulation of MSH6. Interestingly,









CHAPTER 2
EXPERIMENTAL DESIGN:
CASE-CONTROL AND FAMILY-BASED ASSOCIATION STUDIES

Introduction

Genetic association studies examine whether a genetic variant (like a SNP) is associated

with a disease or trait on a population scale (Cardon and Bell, 2001). This approach has already

been successfully used to further our understanding of a number of diseases, such as diabetes

(Morris et al., 2006; Smyth et al., 2006), Parkinson's disease (Mizuta et al., 2006), and

rheumatoid arthritis (Hughes et al., 2006; Ikari et al., 2006; Kang et al., 2006) to name a few.

Due to budgetary limitations in my study, I had to restrict the number of candidate genes

examined to eight and the number of SNPs genotyped to thirty. These were chosen with the

following criteria in mind: evidence of the gene being involved (or likely to be involved) in

vitiligo pathogenesis, allele frequencies in my sample population, haploblock data, and the

likelihood that the SNP may have a functional effect (Drago et al., 2007).

As previously described, two principal hypotheses concerning the etiology of vitiligo

include (1) the autoimmune model, which suggests that melanocyte death occurs through

inappropriate immune system destruction of pigment cells, and (2) the self-destruct model, which

suggests that biochemical and/or structural defects inherent to patient melanocytes contribute to

the initiation and/or progression of melanocyte cytolysis. To evaluate genetic factors that may

play a role in vitiligo pathogenesis, a candidate gene approach was designed. Genes involved in

immune system regulation, melanin biosynthesis, and DNA repair and oxidative stress regulation

were chosen based on their potential involvement in vitiligo pathogenesis. Case-control and

family-based analyses were performed on genotypic data garnered from vitiligo patients, their

family members, and non-affected controls. Pairwise haplotype analysis was also performed.


































2009 Deborah Marsha Herbstman









Badri, A.M., Todd, P.M., Garioch, J.J., Gudgeon, J.E., Stewart, D.G. and Goudie, R.B.: An
immunohistological study of cutaneous lymphocytes in vitiligo. J Pathol 170 (1993) 149-
55.

Baharav, E., Merimski, O., Shoenfeld, Y., Zigelman, R., Gilbrud, B., Yecheskel, G., Youinou, P.
and Fishman, P.: Tyrosinase as an autoantigen in patients with vitiligo. Clin Exp
Immunol 105 (1996) 84-8.

Bakis-Petsoglou, S., Le Guay, J.L. and Wittal, R.: A randomized, double-blinded, placebo-
controlled trial of pseudocatalase cream and narrowband ultraviolet B in the treatment of
vitiligo. Br J Dermatol (2009) [electronic publication, in press].

Bethea, D., Fullmer, B., Syed, S., Seltzer, G., Tiano, J., Rischko, C., Gillespie, L., Brown, D. and
Gasparro, F.P.: Psoralen photobiology and photochemotherapy: 50 years of science and
medicine. J Dermatol Sci 19 (1999) 78-88.

Bhatia, P.S., Mohan, L., Pandey, O.N., Singh, K.K., Arora, S.K. and Mukhija, R.D.: Genetic
nature of vitiligo. J Dermatol Sci 4 (1992) 180-4.

Birlea, S.A., Fain, P.R. and Spritz, R.A.: A Romanian population isolate with high frequency of
vitiligo and associated autoimmune diseases. Arch Dermatol 144 (2008) 310-6.

Bjorses, P., Pelto-Huikko, M., Kaukonen, J., Aaltonen, J., Peltonen, L. and Ulmanen, I.:
Localization of the APECED protein in distinct nuclear structures. Hum Mol Genet 8
(1999) 259-66.

Boe Wolff, A.S., Oftedal, B., Johansson, S., Bruland, O., Lovas, K., Meager, A., Pedersen, C.,
Husebye, E.S. and Knappskog, P.M.: AIRE variations in Addison's disease and
autoimmune polyendocrine syndromes (APS): partial gene deletions contribute to APS I.
Genes Immun 9 (2008) 130-6.

Boissy, R.E., Liu, Y.Y., Medrano, E.E. and Nordlund, J.J.: Structural aberration of the rough
endoplasmic reticulum and melanosome compartmentalization in long-term cultures of
melanocytes from vitiligo patients. J Invest Dermatol 97 (1991) 395-404.

Bos, J.D.: Skin immune system (SIS) : cutaneous immunology and clinical immunodermatology,
3rd ed. CRC Press, Boca Raton, 2005.

Bottomley, M.J., Collard, M.W., Huggenvik, J.I., Liu, Z., Gibson, T.J. and Sattler, M.: The
SAND domain structure defines a novel DNA-binding fold in transcriptional regulation.
Nat Struct Biol 8 (2001) 626-33.

Brown-Harrell, V., Nitta, A.T. and Goble, M.: Apparent exacerbation of vitiligo syndrome in a
patient with pulmonary Mycobacterium avium complex disease who received
clofazimine therapy. Clin Infect Dis 22 (1996) 581-2.









block information was obtained from the International HapMap Project (www.HapMap.org) and

was viewed using SNPbrowserTM Version 3.5 (Applied Biosystems) (De La Vega et al., 2006).

In order of priority, SNPs that caused an amino acid change were selected first, followed by

SNPs located in exons, those located at 5' or 3' UTR, SNPs located in introns, and lastly SNPs

located very close to a gene on the same haplotype block. For a few genes, more than two SNPs

were chosen. SNPs are listed in the chapter that discusses their association results.

Genotyping

SNP genotyping was performed at the University of Florida Center for Pharmacogenomics

using the Applied Biosystems 7900 HT SNP genotyping platform with TaqMan assays (Table

2-2). Five pL reactions in 384-well plates were prepared, and the assays were performed and

analyzed according to the manufacturer's recommendations.

Statistical Analysis

A number of statistical analyses were performed on the sample data, and results are

reported in this and subsequent chapters. These methods are described below.

Hardy-Weinberg Analysis

Hardy-Weinberg proportions were tested in cases and controls for all SNPs genotyped.

Results are reported only for those SNPs where deviations from HWE were observed.

Deviations from Hardy-Weinberg equilibrium (HWE) have been observed in many genetic

association studies (Salanti et al., 2005), though the interpretation of this finding remains

controversial (Zou and Donner, 2006; Teo et al., 2007; Li and Li, 2008). While some believe

that deviations from HWE, also known as Hardy-Weinberg disequilibrium (HWD), may indicate

genotyping errors (Gomes et al., 1999; Weiss et al., 2001; Teo et al., 2007), others refute the idea

of using HWE testing for this purpose (Zou and Donner, 2006). The assumptions underlying

HWE, including random mating, lack of selection according to genotype, and absence of










Table 4-1. List of Applied Biosystems SNP Assay IDs and description ofDCT, PAH, TYR,
TYRP1, and COMT SNPs genotyped


Gene(s)
DCT
DCT
DCT
DCT
DCT
DCT
DCT
DCT
PAH
PAH
TYR
TYR
TYR
TYR
TYR
TYR
TYR
TYR
TYRP1
TYRP1
TYRP1
TYRP1
TYRP1
TYRP1
COMT
COMT


SNP rs number
rs9516413
rs7987802
rs4318084
rs9524493
rs9516418
rs1028805
rs11618471
rs7991232
rs1722381
rs1522307
rs1042602
rs621313
rs594647
rs10765197
rs12791412
rs2000554
rs10830250
rs1827430
rs2762462
rs2733831
rs2733833
rs683
rs2762464
rs1063380
rs2020917
rs6269


Applied Biosystems ID
C 1872510 10
C 1872504 10
C 26557934 10
C 29504474 10
C 26557938 10
C 7593146 10
C 1872491 10
C 30496225 10
C 1402626 10
C 1402699 10
C 8362862_10
C 1054193 10
C 1054191 10
C 31959995 10
C 31959989 10
C 11665932 10
C 31959927 10
C 27097966 20
C 15931132 10
C 3119212 10
C 3119209 10
C 3119206 10
C 15931130 10
C 3119204 10
C 11731880 1
C 2538746 1


Base pair change & location
(amino acid change & location)
C/T 3'UTR
C/T intron 6
C/T intron 6
C/G intron 6
C/T intron 6
G/T intron 2
A/G intron 2
A/G intron 1
A/G intron 6
A/G intron 2
C/A exon 1, (S192Y)
G/A intron 1
T/C intron 1
A/C intron 2
A/G intron 2
G/A intron 2
C/G intron 3
G/A intron 4
C/T intron 4
A/G intron 5
G/T intron 6
A/C 3'UTR
A/T 3'UTR
C/T 3'UTR
T/C 5'UTR
G/A intron 2









Table 3-3. Minor allele frequencies (MAF) for AIRE SNPs from Applied Biosystems Caucasian cohort (ABI Caucasian) and values
for the white case and control groups from the present study; X2 p-values and q-values for allelic comparisons are given
ABI White # of White # of
Caucasian case white control white Control BonC FDR
SNP ID MAF MAF cases Case alleles MAF controls alleles p-value p-value q-value
A 191 (59%) A 276 (59%)
rs2776377 0.42 (G) 0.41 (G) 163 G 135 (41%) 0.41 (G) 233 G 190 (41%) 0.86 1.00 0.93
C 293 (90%) C 348 (73%)
rs1800520 0.13 (G) 0.10 (G) 163 G 33 (10%) 0.27 (G) 237 G 126 (27%) 9.90E-09* <0.0001* <0.0001*
denotes a significant finding; BonC denotes the Bonferroni corrected p-value









Table 3-4. Family-based association (transmission disequilibrium test) results for AIRE
Number of Not
informative Transmitted transmitted BonC FDR
SNP parents (allele) (allele) p-value p-value q-value
rs2776377 51 25 (A) 26 (G) 1.00 1.00 1.00
rs1800520 17 12 (C) 5 (G) 0.16 1.00 0.46


BonC denotes the Bonferroni corrected p-value


Table 3-5. Linkage disequilibrium
Gene SNP ID
AIRE rs2776377


(D') value for AIRE
SNP ID D'
rs1800520 0.47


Table 3-6. Pairwise haplotype analysis results for AIRE rs2776377 and rsl800520
Frequency Controls Vitiligo patients p-values


0 [AG] 188 (81.5%) 154 (96.5%)
1 [AG] 41(18%) 6(3.5%)
2 [AG] 1 (0.5%) 0 (0%)
Total (n) 230 160
X2 p<.0001*
Odds ratio for 1 [AG] vs. 0 [AG] = 0.1673 95% CI=0.0663 to 0.4070 p<.0001*
Odds ratio for 2 [AG] vs. 0 [AG] = n.d. n.d.
Odds ratio for 2 [AG] vs. 1 [AG] = n.d. n.d.


n.d. = not defined; denotes a significant finding; number in frequency column indicates the
number of copies of this haplotype









4-6 Pairwise haplotype analysis results for PAH rs1722381 and rs1522307........................76

4-7 Pairwise haplotype analysis results forDCTrs4318084 and rs9516413..........................77

4-8 Pairwise haplotype analysis results for DCTrs11618471 and rs9516413........................77

4-9 Pairwise haplotype analysis results for DCTrs7991232 and rs9516413..........................77

4-10 Pairwise haplotype analysis results forDCTrs9524493 and rs9516413..........................78

4-11 Pairwise haplotype analysis results forDCTrs7987802 and rs9516413..........................78

4-12 Significant x2 findings for DCT association for patients with co-morbid autoimmune
d ise a se s ......................................................... ..................................7 8

4-13 Pairwise haplotype analysis results for TYR rs10765197 and TYR rs1042602..................79

4-14 Pairwise haplotype analysis results for TYR rs1042602 and TYR rs12791412..................79

4-15 Pairwise haplotype analysis results for TYR rs1042602 and TYR rs10830250..................79

4-16 Pairwise haplotype analysis results for TYR rs1042602 and TYR rs2000554 ..............80

4-17 Pairwise haplotype analysis results for TYR rs1042602 and TYR rs1827430 ....................80

4-18 Pairwise haplotype analysis results for TYR rs12791412 and TYR rs1827430 ..................80

4-19 Significant Hardy-Weinberg equilibrium analysis for TYRP1 .......................................81

4-20 Pairwise haplotype analysis results for TYRP1 rs683 and TYRP1 rs2733831 ...................81

4-21 Pairwise haplotype analysis results for TYRP1 rs683 and TYRP1 rs2733833...................81

4-22 Pairwise haplotype analysis results for TYRP1 rs273383 land TYRP1 rs2762464 ............82

4-23 Pairwise haplotype analysis results for TYRP1 rs2733831 and TYRP1 rs2733833 ...........82

4-24 Pairwise haplotype analysis results for TYRP1 rs2762464 and TYRP1 rs2733833...........82

4-25 Pairwise haplotype analysis results for TYRP1 rs2762462 and TYRP1 rs2733833...........83

4-26 Pairwise haplotype analysis results for COMTrs2020917 and COMTrs6269 .................83

4-27 Summary of all findings for DCT, PAH, TYR, TYRP1, and COMT.............................84

A -1 A ll SN P s g en oty p ed ................................................................................. ................ .. 9 5









non-lesional vitiligo skin or skin of healthy controls, and upregulated in uninvolved vitiligo skin

compared to healthy control samples (Kingo et al., 2007). This may be a result of the lower

number of melanocytes present in lesional vitiligo skin. In studies using cell culture and the

chemotoxin 4-tert-butylphenol (4-TBP), a competitive inhibitor of tyrosinase, Manga et al.

observed increased expression of TYRP1 in vitiligo compared to controls (Manga et al., 2006).

This increased expression of TYRP1 significantly increased cytotoxic sensitivity of vitiligo

melanocytes to 4-TBP. Another group of investigators found that cultured keratinocytes from

vitiligo patients promote reduced TYRP1 protein expression in co-cultured melanocytes (Phillips

et al., 2001). One study found that a single nucleotide polymorphism or SNP (pronounced

"snip"), which causes altered promoter function in the gene FOXD3, was genetically associated

with vitiligo (Alkhateeb et al., 2005); this gene is a transcriptional regulator of TYRP1 (Thomas

et al., 2008). The sum of these findings support the hypothesis that alterations in the normal

functioning and/or expression patterns of tyrosinase-related protein 1 and dopachrome

tautamerase in melanocytes may contribute to oxidative stress observed in vitiligo.

Levels of the antioxidant enzymes superoxide dismutase, glutathione peroxidase and

malondialdehyde were significantly increased in tissue of patients with non-segmental vitiligo

compared to controls (Yildirim et al., 2003), presumably due to an increase of ROS in the skin of

vitiligo patients. Also, significant association to vitiligo was observed in polymorphisms of

glutathione S-transferase (GST) genes, which are important in protection against oxidative stress

(Uhm et al., 2007). Compared to healthy individuals, there is an increase in oxidative DNA

damage in the mononuclear component of peripheral blood leukocytes from vitiligo patients

(Giovannelli et al., 2004). The catechol-O-methyl transferase or COMTgene, which is involved

in catecholamine biosynthesis as well as melanin biosynthesis and oxidative stress regulation,









of my findings are false due to a mismatch of population subgroups. Whereas false negatives

cause researchers to miss truly informative findings, false positives lead them to chase genetic

wild geese. While obvious, it bears repeating that the greater the significance of a finding, the

more likely that finding represents a true association. Findings showing only weak significance

should be replicated in many independent studies to be validated.

Future Directions

I am currently working on trying "SNP chips" for my DNA samples, and have

collaboration with an NIH-sponsored VITGene consortium. Combined, these experiments will

allow us to examine tens of thousands of SNPs across the human genome. The results from

these new studies will include SNPs on the candidate genes discussed in my present study as

well as thousands of additional genes. This will allow us to understand the findings of the

present study and will likely suggest many more genes for follow-up analyses.

Additional studies need to be conducted to examine protein expression of AIRE and of

melanin biosynthesis genes to examine their role in vitiligo pathogenesis. This would involve

collection of skin samples from vitiligo patients. Obtaining thymus samples to better understand

AIRE protein expression would be much more difficult, though theoretically this could be done

be through tissue banking. However, the thymus atrophies with age so studies using tissues from

older individuals might not be informative.

The long-term goals of my research include determining whether "at-risk" individuals can

be identified in families with a history of vitiligo, which might influence the choice of strategies

for treatment or prevention of vitiligo. Future collaborations with clinicians who see vitiligo

patients would not only allow us greater access to a larger sample pool, but it would also enable

us to more accurately phenotype my subjects. Further, a prospective study testing for association

of a SNP or gene with a therapy outcome could yield very useful results. Working together,









Kim, S.M., Chung, H.S. and Hann, S.K.: The genetics of vitiligo in Korean patients. Int J
Dermatol 37 (1998) 908-10.

Kingo, K., Aunin, E., Karelson, M., Philips, M.A., Ratsep, R., Silm, H., Vasar, E., Soomets, U.
and Koks, S.: Gene expression analysis of melanocortin system in vitiligo. J Dermatol Sci
48 (2007) 113-22.

Kobayashi, T., Imokawa, G., Bennett, D.C. and Hearing, V.J.: Tyrosinase stabilization by Tyrpl
(the brown locus protein). J Biol Chem 273 (1998) 31801-5.

Kobayashi, T., Urabe, K., Winder, A., Jimenez-Cervantes, C., Imokawa, G., Brewington, T.,
Solano, F., Garcia-Borron, J.C. and Hearing, V.J.: Tyrosinase related protein 1 (TRP1)
functions as a DHICA oxidase in melanin biosynthesis. EMBO J 13 (1994) 5818-25.

Kovacs, S.O.: Vitiligo. J Am Acad Dermatol 38 (1998) 647-66; quiz 667-8.

Kumar, P.G., Laloraya, M., Wang, C.Y., Ruan, Q.G., Davoodi-Semiromi, A., Kao, K.J. and She,
J.X.: The autoimmune regulator (AIRE) is a DNA-binding protein. J Biol Chem 276
(2001) 41357-64.

Kwon, J.M. and Goate, A.M.: The candidate gene approach. Alcohol Res Health 24 (2000) 164-
8.

Le Poole, I.C., Boissy, R.E., Sarangarajan, R., Chen, J., Forristal, J.J., Sheth, P., Westerhof, W.,
Babcock, G., Das, P.K. and Saelinger, C.B.: PIG3V, an immortalized human vitiligo
melanocyte cell line, expresses dilated endoplasmic reticulum. In Vitro Cell Dev Biol
Anim 36 (2000) 309-19.

Le Poole, I.C., Das, P.K., van den Wijngaard, R.M., Bos, J.D. and Westerhof, W.: Review of the
etiopathomechanism ofvitiligo: a convergence theory. Exp Dermatol 2 (1993a) 145-53.

Le Poole, I.C. and Luiten, R.M.: Autoimmune etiology of generalized vitiligo. Curr Dir
Autoimmun 10 (2008) 227-43.

Le Poole, I.C., Sarangarajan, R., Zhao, Y., Stennett, L.S., Brown, T.L., Sheth, P., Miki, T. and
Boissy, R.E.: 'VIT1', a novel gene associated with vitiligo. Pigment Cell Res 14 (2001)
475-84.

Le Poole, I.C., van den Wijngaard, R.M., Smit, N.P., Oosting, J., Westerhof, W. and Pavel, S.:
Catechol-O-methyltransferase in vitiligo. Arch Dermatol Res 286 (1994) 81-6.

Le Poole, I.C., van den Wijngaard, R.M., Westerhof, W. and Das, P.K.: Presence of T cells and
macrophages in inflammatory vitiligo skin parallels melanocyte disappearance. Am J
Pathol 148 (1996) 1219-28.

Le Poole, I.C., van den Wijngaard, R.M., Westerhof, W., Dutrieux, R.P. and Das, P.K.: Presence
or absence of melanocytes in vitiligo lesions: an immunohistochemical investigation. J
Invest Dermatol 100 (1993b) 816-22.










Table 4-7. Pairwise haplotype analysis results for DCTrs4318084 and rs9516413
Frequency Controls Vitiligo patients p-values
0 [CC] 165 (67.35%) 118(55.7%)
1 [CC] 67 (27.35%) 71(33.5%)
2 [CC] 13 (5.3%) 23 (10.8%)
Total (n) 245 212
X2 0.0153*
Odds ratio for 1 [CC] vs. 0 [CC] = 1.5116 95% CI=1.0016 to 2.2814 0.0491*
Odds ratio for 2 [CC] vs. 0 [CC] = 2.4309 95% CI=1.1817 to 5.0005 0.0158*
Odds ratio for 2 [CC] vs. 1 [CC] = n.s. n.s.
n.s. = not significant; denotes a significant finding; number in frequency column indicates the
number of copies of this haplotype


Table 4-8. Pairwise haplotype analysis results for DCTrs 11618471 and rs9516413
Frequency Controls Vitiligo patients p-values
0 [AC] 165 (67.6%) 118(56%)
1 [AC] 67 (27.5%) 71(34%)
2 [AC] 12 (4.9%) 21 (10%)
Total (n) 244 210
X2 0.0195*
Odds ratio for 1 [AC] vs. 0 [AC] = 1.5116 95% CI=1.0015 to 2.2813 0.0491*
Odds ratio for 2 [AC] vs. 0 [AC] = 2.4152 95% CI=1.1425 to 5.1058 0.0210*
Odds ratio for 2 [CC] vs. 1 [CC] = n.s. n.s.
n.s. = not significant; denotes a significant finding; number in frequency column indicates the
number of copies of this haplotype




Table 4-9. Pairwise haplotype analysis results for DCTrs7991232 and rs9516413
Frequency Controls Vitiligo patients p-values
0 [GT] 176 (72%) 127 (60%)
1 [GT] 61 (25%) 70 (33%)
2 [GT] 7 (3%) 14 (7%)
Total (n) 244 211
X2 0.0129*
Odds ratio for 1 [GT] vs. 0 [GT] = 1.6164 95% CI=1.0714 to 2.4476 0.0222*
Odds ratio for 2 [GT] vs. 0 [GT] = 2.7193 95% CI=1.0652 to 6.9419 0.0364*
Odds ratio for 2 [CC] vs. 1 [CC] = n.s. n.s.
n.s. = not significant; denotes a significant finding; number in frequency column indicates the
number of copies of this haplotype









autoimmune diseases, though this hypothesis has not yet been tested. Alternatively, these

associations may help uncover pathogenic changes in DCTthat are yet to be determined.

I genotyped eight SNPs in the tyrosinase gene to investigate both the autoimmune and

autocytotoxic hypotheses underlying vitiligo pathogenesis, and my study yielded several

interesting results. One intronic SNP in the TYR was my only (uncorrected) significant finding

in my age of disease onset analysis. It is possible that patients with a later age of onset represent

a different sub-group of vitiligo, however, because my finding was associated with

homozygosity for the minor allele for an intron two SNP, this result must be interpreted with

caution. This same intron two SNP was also significant at the allelic but not the genotypic level

for my entire white case-control analysis. It is possible that the SNP itself causes an RNA

instability or leads to splicing problems, or this SNP may be a marker for another genetic

changed linked to vitiligo etiology.

Pairwise haplotype analysis for TYR showed six significant associations with vitiligo, and

five of these involved a non-synonymous SNP. It is not known how or if this change in exon 1

from a serine to a tyrosine alters tyrosinase function, though this SNP is thought to contribute to

normal pigment variation in individuals of European ancestry (Shriver et al., 2003; 2005; Norton

et al., 2007). The exact role oftyrosinase in vitiligo pathogenesis is unknown, but my findings

support further investigation of this gene.

I observed several significant findings in my study of the TYRP1 gene. TYRP1 protein is

thought to play an important role in enhancing tyrosinase function, and in its active state, it is

believed to be part of a multi-enzyme complex. I found SNPs in TYRP1 to be in Hardy-

Weinberg disequilibrium (HWD) for vitiligo cases but not controls, suggesting a possible linkage

disequilibrium difference between the two groups, which may be considered supportive of









Dead Sea climatotherapy has proven to be an effective vitiligo treatment. Although the specific

therapeutic mechanisms of this unique environment have yet to be elucidated, they are suggested

to be attributed to "pseudocatalase" activity of Dead Sea salts and natural UVB exposure via

sunlight (Schallreuter et al., 2002).

Genetics of Vitiligo

The importance of genetic factors in susceptibility to vitiligo is suggested by reports of

familial aggregation from several laboratories (Bhatia et al., 1992; Nordlund, 1997; Kim et al.,

1998; Passeron and Ortonne, 2005; Fain et al., 2006; Hu et al., 2006; Spritz, 2006). About 20%

of vitiligo patients have at least one affected first-degree relative, and the relative risk of vitiligo

for first-degree relatives of vitiligo patients is increased by at least seven- to ten-fold (Bhatia et

al., 1992; Nath et al., 1994). Multiple associations with human major histocompatibility

complex (MHC) genes have been claimed (Dunston and Halder, 1990; Venneker et al., 1992;

Buc et al., 1998; Casp et al., 2003) supporting the possible involvement of immunologically

relevant genes in the human leukocyte antigen (HLA) region. However, no common HLA

association has been observed. Studies have reported HLA associations in early onset vs. late

onset vitiligo cases. It is possible that there are different etiologies and different genetic factors

in early vs. late onset vitiligo, as well as in segmental vs. non-segmental vitiligo (Finco et al.,

1991; Orecchia et al., 1992; Arcos-Burgos et al., 2002). Other genes implicated in susceptibility

to vitiligo include ACE (Jin et al., 2004a; Akhtar et al., 2005), ESR1 (Jin et al., 2004b), CAT

(Casp et al., 2002; Park et al., 2006; Em et al., 2007), and NALP1 (Jin et al., 2007b) to name a

few. Based on all of these observations, vitiligo is hypothesized to be a polygenic disease, with

variant alleles in several unlinked loci possibly contributing to increased susceptibility to and/or

direct pathogenesis of vitiligo.









for haplotype findings were located in intron six (rs7987802, rs4318084, rs9524493) with the

other two located in intron two (rsl 1618471), and intron one (rs7991232). However, no other

DCT SNPs had significant associations in either the allelic or genotypic analyses (Table 4-3 and

4-4), nor was Hardy-Weinberg analysis significant for my white cases and controls. Including

rs9516413, five out of eight DCT SNPs tested had significant TDT findings (Table 4-5). The

other four SNPs were all intronic, with rs4318084 and rs9516418 both located in intron six, and

rs1028805 and rs 1618471 both located in intron two. Lastly, three out of four SNPs tested in

intron six met statistical significance in vitiligo patients who had co-morbid autoimmune

diseases: rs7987802 (p=0.038), rs4318084 (p=0.038), and rs9524493 (p=0.033) (Table 4-12).

Interestingly, rs9516413 was not among them despite its association in case-control analysis.

Tyrosinase

Of the eight SNPs genotyped in TYR, only one yielded significant results in independent

tests (Table 4-3 and 4-4). An A/G polymorphism in intron 2, rs12791412 was borderline

significant at the allelic level (p=0.05) (Table 4-3); the A allele had a RR= 1.08 (CI=1.0002 to

1.1662), though this SNP showed no significance upon genotypic analysis (p=0.13) (Table 4-4).

TDT analysis for rs12791412 was significant for excessive transmission of the A allele to cases

(p=0.02) (Table 4-5). Analysis of variance (ANOVA) indicated a significant association of this

SNP's genotype with age of disease onset; the GG cohort had a mean age of onset of 33.67

years, compared to the AA cohort mean age of onset at 19.58 years (p = 0.027) and the AG

group mean age of onset 14.56 years (p=0.0022). Hardy-Weinberg analysis of all TYR SNPs

showed no disequilibrium (data not shown). Six of the two-SNP haplotype analyses were

significant (Tables 4-13, 4-14, 4-15, 4-16, 4-17, and 4-18). Of these, five contain rs1042602, a

SNP in exon one that causes an amino acid change from serine to tyrosine at position 192

(S192Y). In haplotype analysis, homozygosity for the C major allele in rs1042602 (coding for









ACKNOWLEDGMENTS

I am grateful to all of the individuals who participated in this research study; their

generosity has made this work possible. I thank the American Vitiligo Research Foundation

(www.AVRF.org) for the opportunity to learn from their members and families, and for financial

support of this research. I would like to acknowledge AVRF's founder and president, Stella

Pavlides, for her encouragement and her energetic enthusiasm. I also wish to thank Marilyn

Rose Giordano; although she is no longer with us, I am inspired by the memory of her selfless

dedication to helping individuals, especially children, with vitiligo.

I would like to thank my mentor, Dr. Margaret R. "Peggy" Wallace for allowing me work

in her lab and grow as a scientist under her guidance. I am amazed at the depth and breath of her

scientific knowledge, and inspired by her commitment to outreach and community service

through her charitable work with patients and families. I am very lucky that Peggy appreciates

my sense of humor and doesn't mind hearing my laugh from down the hall. I owe most of those

laughs to my fantastic lab mates, past and present. Specifically, I would like to acknowledge

Michelle "Chello" Burch for being an amazing technician and a dear friend, Dr. Rebecca

"Boocca" Loda-Hutchinson for moral support and helpful instruction in both science and dance,

Dr. Hua Li for laboratory assistance and enjoyable conversations in which she allowed me to

encourage her to practice her English, and Dr. Angela Hadjipanayis for sharing her unique

perspective in science and in life. I thank Dr. Wayne T. McCormack for his dual roles as my co-

mentor and director of my PhD program. I am grateful for his help and advice, both in guiding

this project and in my scientific career. Dr. McCormack's dedication to his students and to

community service has truly been an inspiration; from one scout to another, I salute him.

I would like to thank my committee member Dr. Daniel C. Driscoll for being a mentor

both in the lab and in the clinic, for interesting conversations about genetics and politics, and for









Salanti, G., Amountza, G., Ntzani, E.E. and loannidis, J.P.: Hardy-Weinberg equilibrium in
genetic association studies: an empirical evaluation of reporting, deviations, and power.
Eur J Hum Genet 13 (2005) 840-8.

Sampogna, F., Raskovic, D., Guerra, L., Pedicelli, C., Tabolli, S., Leoni, L., Alessandroni, L. and
Abeni, D.: Identification of categories at risk for high quality of life impairment in
patients with vitiligo. Br J Dermatol 159 (2008) 351-9.

Schallreuter, K.U.: Successful treatment of oxidative stress in vitiligo. Skin Pharmacol Appl Skin
Physiol 12 (1999) 132-8.

Schallreuter, K.U., Chavan, B., Rokos, H., Hibberts, N., Panske, A. and Wood, J.M.: Decreased
phenylalanine uptake and turnover in patients with vitiligo. Mol Genet Metab 86 Suppl 1
(2005) S27-33.

Schallreuter, K.U., Gibbons, N.C., Zothner, C., Abou Elloof, M.M. and Wood, J.M.: Hydrogen
peroxide-mediated oxidative stress disrupts calcium binding on calmodulin: more
evidence for oxidative stress in vitiligo. Biochem Biophys Res Commun 360 (2007) 70-5.

Schallreuter, K.U., Lemke, R., Brandt, O., Schwartz, R., Westhofen, M., Montz, R. and Berger,
J.: Vitiligo and other diseases: coexistence or true association? Hamburg study on 321
patients. Dermatology 188 (1994a) 269-75.

Schallreuter, K.U., Moore, J., Behrens-Williams, S., Panske, A. and Harari, M.: Rapid initiation
of repigmentation in vitiligo with Dead Sea climatotherapy in combination with
pseudocatalase (PC-KUS). Int J Dermatol 41 (2002) 482-7.

Schallreuter, K.U., Moore, J., Wood, J.M., Beazley, W.D., Gaze, D.C., Tobin, D.J., Marshall,
H.S., Panske, A., Panzig, E. and Hibberts, N.A.: In vivo and in vitro evidence for
hydrogen peroxide (H202) accumulation in the epidermis of patients with vitiligo and its
successful removal by a UVB-activated pseudocatalase. J Investig Dermatol Symp Proc 4
(1999) 91-6.

Schallreuter, K.U., Moore, J., Wood, J.M., Beazley, W.D., Peters, E.M., Marles, L.K., Behrens-
Williams, S.C., Dummer, R., Blau, N. and Thony, B.: Epidermal H(2)0(2) accumulation
alters tetrahydrobiopterin (6BH4) recycling in vitiligo: identification of a general
mechanism in regulation of all 6BH4-dependent processes? J Invest Dermatol 116 (2001)
167-74.

Schallreuter, K.U. and Wood, J.M.: The importance of L-phenylalanine transport and its
autocrine turnover to L-tyrosine for melanogenesis in human epidermal melanocytes.
Biochem Biophys Res Commun 262 (1999) 423-8.

Schallreuter, K.U. and Wood, J.M.: Thioredoxin reductase its role in epidermal redox status. J
Photochem Photobiol B 64 (2001) 179-84.

Schallreuter, K.U., Wood, J.M. and Berger, J.: Low catalase levels in the epidermis of patients
with vitiligo. J Invest Dermatol 97 (1991) 1081-5.










LIST OF FIGURES


Figure


1-1 M elanin biosynthesis pathw ay ................................................. .............................. 30

4-1 Schematic representation of gene structure, SNP location, and HapMap CEPH
haploblocks ............................................... ..........................................70


page









Results

The rs1800520 SNP, which encodes the amino acid substitution, showed very highly

significant association to vitiligo, p=0.0000002 uncorrected (UC) in my case-control genotype

analysis (Table 3-2). The p-value after correcting for multiple tests using Bonferroni's method

(BonC) was <0.0001, and FDR q-value was <0.0001. Most notably, the GG genotype of the

minor allele (resulting in homozygosity for the arginine residue) was seen in 19 out of 237

controls (8%) and 0 out of 163 vitiligo patients (0%). The CG heterozygous genotype was

observed in 88 controls (37%) and 33 cases (20%). Allelic frequencies were calculated for cases

and controls for rs1800520, and the minor allele frequencies (MAF) from the Applied

Biosystems validation set (De La Vega et al., 2006) are given (Table 3-3); the x2 allelic

frequency analysis for my white case-control groups was significant at p=9.9038E-09

(uncorrected), p<0.0001 (BonC) and q<0.0001. The G allele was found to have a protective

effect: RR=0.38 (CI= 0.2667 to 0.5439). Conversely, the C allele for rs1800520 confers

susceptibility. It is noteworthy that my control MAF of 0.27 differs from the Appplied

Biosystems Incorporated (ABI) reported MAF of 0.13 for the rs 800520 SNP. ABI allele

frequencies were obtained from a set of 45 people, while my study included 163 controls

genotyped for rs1800520. In contrast, white case-control analysis of the rs2776377 SNP yielded

no significant findings at the allelic (Table 3-3) or genotype levels (p=0.46) (Table 3-2). TDT

analysis for both SNPs was not significant (Table 3-4).

Linkage disequilibrium was calculated for the two AIRE SNPs (Table 3-5), showing

moderate disequilibrium and a haplotype analysis yielded significant findings. The haplotype

with the highest log-likelihood for the AIRE gene was the "AG" haplotype (denoted as [AG]),

having the "A" allele for rs2776377 and the "G" allele for rs1800520. I found that the presence

of [AG] was significantly higher in the control group compared with vitiligo patients (p < .0001









analysis using the SAS software package was performed to examine whether there was a

statistical significance between male and female genotypes in the vitiligo patients compared to

the healthy controls.

Genotype by Age of Onset

The analysis of variance (ANOVA) method using the SAS software package was

performed, to examine whether there was a statistical significance between vitiligo patient

genotypes and age at disease onset.

Pairwise Haplotype Analysis

Lewontin's D'statistic for linkage disequilibrium was calculated for each pair of

polymorphisms for all whites genotyped; cases and controls were pooled for this analysis. The

values of Lewontin's D'can range from -1 to 1; I report the absolute value of D', with zero

indicating complete independence between two markers and one representing complete linkage

disequilibrium, where markers segregate together 100% of the time (Lewontin, 1964) (Table 2-

3).

Linkage disequilibrium (LD) and haplotype frequencies were calculated using a maximum-

likelihood model incorporating an iterative expectation-maximization (EM) algorithm previously

described by Liu et al. (2004). The y2 statistic was used to test the association between the

haplotypes and the case-control groups. Four 3 x2 x2 tests were calculated for each pair of SNPs

in a gene to ascertain the presence and number (2, 1, or 0) of risk reference haplotypes in cases

and controls (Liu et al., 2004). The risk reference haplotype can be determined by selecting the

highest log-likelihood value of the x2 tests. The risk associated with haplotypes was calculated

as the odds ratio with 95% confidence intervals. Analysis was conducted using only two SNPs

at a time (pairwise haplotypes) due to the limited number of samples in my study.









Table 2-1. Demographic information for vitiligo patients and their unaffected relatives
Vitiligo patients Unaffected relatives
Sex Race Sex Race
Male 128 Black 7 Male 164 Black 7


(36%)


Female 227
(64%)


Chinese
Hispanic
Indian (Asian)
Mixed
Native American
Vietnamese
White
Other
Not reported


2
28
8
16
1
1
279


Total


Female


(46%)

195
(54%)


Chinese
Hispanic
Indian (Asian)
Mixed
Native American
Vietnamese
White
Other
Not reported


Total


3
30
9
11
0
0
290
3
6
359










Table 4-16. Pairwise haplotype analysis results for TYR rs1042602 and TYR rs2000554
Frequency Controls Vitiligo patients p-values
0 [CA] 133 (54%) 109 (51%)
1 [CA] 102 (42%) 80 (38%)
2 [CA] 10(4%) 23 (11%)
Total (n) 245 212
X2 0.0203*
Odds ratio for 1 [CA] vs. 0 [CA] = n.s. n.s.
Odds ratio for 2 [CA] vs. 0 [CA] = 2.8065 95% CI=1.2801 to 6.1531 0.0100*
Odds ratio for 2 [CA] vs. 1 [CA] = 2.8924 95% CI=1.3015 to 6.4278 0.0091*
n.s. = not significant; denotes a significant finding; number in frequency column indicates the
number of copies of this haplotype


Table 4-17. Pairwise haplotype analysis results for TYR rs1042602 and TYR rs1827430
Frequency Controls Vitiligo patients p-values
0 [CA] 125 (51%) 98 (46%)
1 [CA] 101 (41%) 81 (38%)
2 [CA] 19 (8%) 33 (16%)
Total (n) 245 212
X2 0.0319*
Odds ratio for 1 [CA] vs. 0 [CA] = n.s. n.s.
Odds ratio for 2 [CA] vs. 0 [CA] = 2.2267 95% CI=1.1931 to 4.1557 0.0119*
Odds ratio for 2 [CA] vs. 1 [CA] = 2.1732 95% CI=1.1498 to 4.1073 0.0169*
n.s. = not significant; denotes a significant finding; number in frequency column indicates the
number of copies of this haplotype




Table 4-18. Pairwise haplotype analysis results for TYR rs12791412 and TYR rs1827430
Frequency Controls Vitiligo patients p-values
0 [GA] 130 (53%) 134 (63%)
1 [GA] 97 (40%) 71 (34%)
2 [GA] 18(7%) 7(3%)
Total (n) 245 212
X2 0.0395*
Odds ratio for 1 [GA] vs. 0 [GA] = n.s. n.s.
Odds ratio for 2 [GA] vs. 0 [GA] = 0.3761 95% CI=0.1519 to 0.9314 0.0119*
Odds ratio for 2 [GA] vs. 1 [GA] = n.s. n.s.
n.s. = not significant; denotes a significant finding; number in frequency column indicates the
number of copies of this haplotype









shed light on how environmental factors trigger vitiligo onset in some susceptible individuals but

not others.

I am establishing new statistical collaborations to use more sophisticated models to look at

both the combination of several genes simultaneously (to reveal gene-gene interactions), as well

as more complex family structures. Thus, the future may reveal additional discoveries, based on

new statistical analysis and additional clinical data. The goal of future work in vitiligo will be to

elucidate the complex pathogenesis ofvitiligo with a goal to provide targets for rational therapy

design.









Similar methods were employed for SNP selection for all of the genes tested. The

following software and search engine tools were used to identify SNPs: NCBI's SNP website

(http://www.ncbi.nlm.nih.gov/sites/entrez?db=snp), the SNPper website (http://snpper.chip.org/)

(Riva and Kohane, 2002), the HapMap website (http://www.hapmap.org/) (2003), and

SNPbrowserTM Version 3.5 software (Applied Biosystems) (De La Vega et al., 2006). A number

of factors were considered in choosing which SNPs to genotype, and these are discussed in their

order of selection priority. SNPs that had been reported as validated, either in the NCBI database

or by the HapMap project (2003), were given high selection priority. A SNP was considered

validated when it was seen to occur in multiple individuals, and the highest selection priority for

my study was given to SNPs that had been validated by several independent investigators.

Special consideration was given to SNPs that had been validated by Applied Biosystems using

their TaqMan SNP genotyping assays because all of the genotyping was conducted using

TaqMan due to cost considerations for the scale of my study.

Because the majority of the vitiligo patient samples used in my study was obtained from

American whites, SNPs with good minor allele frequencies in this group were chosen. The

higher the minor allele frequency, the more likely it is to observe sufficient numbers of that allele

to conduct meaningful analysis. Only SNPs with a minor allele frequency of >10% in whites

were used, to give us the greatest power to detect association based on my sample population.

An attempt was made to select at least two SNPs for each gene, one near the 5' end and one near

the 3' end, to obtain better coverage of the gene, as SNPs cluster in haplotype blocks. These

blocks are population-specific clusters where little recombination within the block is believed to

occur. Therefore, genotyping one "tagging" SNP on a haplotype block is hypothesized to allow

information to be inferred about the genotypes for all of the SNPs on the same block. Haplotype









testing for these SNPs also yielded no significant results (Table 2-5). No subsequent analysis

with these SNPs, including pairwise haploblock comparisons, yielded any significant findings

(data not shown).

Discussion

The FBXO11 gene is located on chromosome 2pl6 and consists of 23 exons that code for

an 844 amino acid protein. MSH6 is also located at the same region on chromosome 2pl6 and is

coded on the opposite strand from FBXO11; the 3' ends of both genes overlap by 5 kilobases

(kb) (Le Poole et al., 2001). MSH6 has 10 exons that code for a 1360 amino acid protein. A

previous study using semi-quantitative reverse transcription polymerase chain reaction (RT-

PCR) testing of MSH6 and FBXO11 mRNA from cells cultured from vitiligo patients and

healthy controls found that MSH6 mRNA was upregulated in vitiligo patients compared to

controls while FBX011 mRNA was downregulated in vitiligo patients (Le Poole et al., 2001).

Le Poole and colleagues hypothesized that the formation of RNA-RNA hybrids from the

overlapping regions of MSH6 and FBXO11 may interfere with the MSH6 protein's G/T

mismatch repair function through post-transcriptional gene silencing [also known as RNA

interference (RNAi)] (2001).

To my knowledge, no previous vitiligo association studies examining SNPs in MSH6 and

FBXO11 have been conducted. A genomewide screen of generalized vitiligo patients did not

find evidence to support genetic association for MSH6 and FBXO11; however, this study was

conducted in multiplex families and so may be less applicable to singleton cases, who comprise

the majority of vitiligo patients (Fain et al., 2003). Two review articles by Spritz dismiss the

pathogenicity of FBXO11, though no experimental findings are given to support this claim

(2006; 2007). In my study, the MSH6 and FBXO11 candidate genes did not demonstrate

statistically significant p-values (p<0.05, uncorrected) in any case-control and/or family-based









differences in vitiligo disease presentation and/or progression, and prior immune-suppressive

therapy (Ongenae et al., 2003).

Studies have implicated the AIRE gene, which causes autosomal recessive autoimmune

polyendocrinopathy-candidiasis-ectodermal dystrophy syndrome (APECED), in the autoimmune

pathogenesis of vitiligo (1997; Nagamine et al., 1997; Su and Anderson, 2004; Spritz, 2006).

AIRE produces a DNA-binding protein that regulates negative selection of autoreactive

thymocytes (Anderson et al., 2002) and may play a role in the cellular autoimmune pathogenesis

of vitiligo. Thus, this was one candidate gene I chose to investigate.

Vitiligo and melanoma

Finally, additional support for the autoimmune disease hypothesis comes from the

examination of melanoma, a malignancy of the melanocyte. The immune system can sometimes

recognize tumor cells and launch an immune response against these cells. In the case of

melanoma, immortalized melanocytes are the target cells. Many melanoma patients, especially

those being successfully treated, develop depigmentation resembling vitiligo (although the

distribution of affected regions is different) because healthy and diseased melanocytes share

many of the same surface antigens (Naftzger et al., 1996; Overwijk et al., 1999). This finding

suggests that the immune system can lose its ability to make the distinction between normal and

diseased melanocytes (Ram and Shoenfeld, 2007; Le Poole and Luiten, 2008). It has been

suggested that there is an improved prognosis for melanoma patients who develop vitiligo

(Overwijk et al., 1999; Wankowicz-Kalinska et al., 2003). Interestingly, vitiligo patient sera are

reported to have autoAb recognizing melanoma antigens and are able to inhibit melanoma cell

proliferation in vitro (Fishman et al., 1993) and cause regression of melanoma metastases in vivo

(Merimsky et al., 1994). This observation has led some investigators to study the abnormal

immune response in vitiligo as a basis for formulating new treatments for melanoma.









Venneker, G.T., Westerhof, W., de Vries, I.J., Drayer, N.M., Wolthers, B.G., de Waal, L.P., Bos,
J.D. and Asghar, S.S.: Molecular heterogeneity of the fourth component of complement
(C4) and its genes in vitiligo. J Invest Dermatol 99 (1992) 853-8.

Wakamatsu, K., Ito, S. and Fujita, K.: Production, circulation, and excretion of melanin-related
metabolites in B16 melanoma-bearing mice. Acta Derm Venereol 70 (1990) 367-72.

Wankowicz-Kalinska, A., Le Poole, C., van den Wijngaard, R., Storkus, W.J. and Das, P.K.:
Melanocyte-specific immune response in melanoma and vitiligo: two faces of the same
coin? Pigment Cell Res 16 (2003) 254-60.

Weiss, S.T., Silverman, E.K. and Palmer, L.J.: Case-control association studies in
pharmacogenetics. Pharmacogenomics J 1 (2001) 157-8.

Xie, Z., Chen, D., Jiao, D. and Bystryn, J.C.: Vitiligo antibodies are not directed to tyrosinase.
Arch Dermatol 135 (1999) 417-22.

Yildirim, M., Baysal, V., Inaloz, H.S., Kesici, D. and Delibas, N.: The role of oxidants and
antioxidants in generalized vitiligo. J Dermatol 30 (2003) 104-8.

Yu, H.S., Chang, K.L., Yu, C.L., Li, H.F., Wu, M.T. and Wu, C.S.: Alterations in IL-6, IL-8,
GM-CSF, TNF-alpha, and IFN-gamma release by peripheral mononuclear cells in
patients with active vitiligo. J Invest Dermatol 108 (1997) 527-9.

Yu, H.S., Kao, C.H. and Yu, C.L.: Coexistence and relationship of antikeratinocyte and
antimelanocyte antibodies in patients with non-segmental-type vitiligo. J Invest Dermatol
100 (1993) 823-8.

Zou, G.Y. and Donner, A.: The merits of testing Hardy-Weinberg equilibrium in the analysis of
unmatched case-control data: a cautionary note. Ann Hum Genet 70 (2006) 923-33.









basic scientists and physicians could develop and then test novel therapies that are more

precisely targeted to prevent or treat vitiligo based on the genetic findings.

Future studies could include epigenetic analysis of vitiligo candidate genes, such as the

ones discussed in my study, to examine if epigenetic changes contribute to vitiligo susceptibility.

With additionally funding, it would be vitally important to carefully phenotype all study subjects.

Careful attention to phenotypes may allow for a more detailed examination of patient

subpopulations, which may aid in understanding the complex genetic component of vitiligo. The

inclusion of ancestry informative markers (AIMs) in future studies may also aid in our genetic

examination of vitiligo by helping to more closely match cases and controls for genetic ancestry.

Because of the accessibility of skin and the relative ease of skin biopsies, vitiligo makes an

excellent disease model for other autoimmune diseases. It is possible that what we learn from

studying vitiligo genetic susceptibility may be applicable to the study of other autoimmune

diseases. Similarly, because vitiligo and melanoma both involve the melanocyte, knowledge

gained through studying vitiligo may be applicable to melanoma.

If additional patients and/or different populations were examined with my putative positive

findings, my results may change. Also, additional phenotypic data may be analyzed so that I can

test how these may relate to genotype. Such data could be helpful in understanding observed

variability in phenotype (e.g., why some vitiligo patients respond to treatments and others do not,

or why some have unusual patterns of vitiligo). The identification of vitiligo susceptibility

alleles may reveal new pathways and potential targets for treatment. The involvement of

proteins expressed predominantly in melanocytes would also help specificity of targeted therapy;

for example, a skin cream may only affect melanocytes but not other cell types. This may also









significant association with vitiligo (Table 4-3, 4-4, and 4-5). Hardy-Weinberg analysis was not

significant. The two SNPs were very strongly linked, with D'equal to 0.79, and pairwise

haplotype analysis showed a protective haplotype of [CG] (p=0.0046), but only comparison of

individuals who carried one protective haplotype [CG] versus those carried zero was significant

(p=0.0011) (Table 4-26).

Summary

A summary table of all uncorrected findings for all genes in this chapter is included to aid

in the integration of this large amount of material (Table 4-27). As expected, given the p-values

and small sample size, multiple-test corrections resulted in p-values >0.05 for these analyses,

which might be true effects. For the purposes of the discussion, I will remark about the

significant uncorrected findings should they be validated in future studies.

Discussion

Phenylalanine Hydroxylase

Individually, the two PAH SNPs tested did not yield significant results; however, there was

a significant risk haplotype for individuals who were homozygous for the G allele for both

markers tested. Because both PAH SNPs tested are intronic, it is possible that one or both affect

splicing efficiency or isoform usage, if one or both are truly exerting the functional effect. This

could be further investigated in follow-up studies, likely using splice-site prediction software or

reverse-transcription PCR analyses. It is also possible that the [GG] haplotype observed only in

vitiligo patients is in linkage disequilibrium with some unidentified gene change that directly

contributes to vitiligo pathogenesis. Additional genotyping in more samples, replication in other

patient cohorts, and deep sequencing of this gene are logical steps to help to understand the

significance of this finding.









Autocytotoxic Hypothesis

The autocytotoxic hypothesis is based on the idea that the damage and/or loss of

melanocytes in vitiligo patients are a result of an inherent defect in the melanocyte, resulting in

the buildup of toxic intermediates or metabolites from the melanin biosynthesis pathway.

Melanogenesis produces large amounts of reactive oxygen species (ROS) like hydrogen

peroxide, so melanocytes are at risk of oxidative damage to proteins and DNA unless their

endogenous antioxidant systems are functional (Norris et al., 1996; Bystryn, 1997; Nordlund and

Ortonne, 1998; Passeron and Ortonne, 2005). A diverse body of evidence, including several

cellular abnormalities and biochemical defects that have been reported in vitiligo patients,

support this hypothesis (Schallreuter et al., 1996; Maresca et al., 1997; Passi et al., 1998;

Schallreuter, 1999). Upon cell death, melanocyte antigens are released which may elicit an

immune response.

Many abnormalities have been reported in the skin of vitiligo patients that implicate

reactive oxygen species in disease pathology. Because the melanin biosynthetic pathway

involves many enzymes and co-factors that regulate the creation and destruction of ROS,

alterations in this pathway (either increasing or decreasing enzymatic activity) may lead to the

buildup of toxic intermediates. Decreased levels of the antioxidant enzyme thioredoxin

reductase has been observed in vitiligo patient skin (Schallreuter and Wood, 2001). There is also

increased sensitivity to oxidative stress in melanocytes cultured from perilesional skin of vitiligo

patients (Jimbow et al., 2001). These cultured vitiligo melanocytes had reduced (but not absent)

levels of tyrosinase- related protein 1 (TYRP1) mRNA and their intracellular processing of

TYRP1 seemed abnormal. Using quantitative real-time PCR, Kingo and colleagues found

several aberrant gene expression levels in vitiligo patients. Notably, they observed that TYRP1

and dopachrome tautamerase (DCT) genes were downregulated in lesional skin compared to











Table 4-2. Continued
Gene SNP ID
TYR rs12791412
TYR rs1042602
TYR rs1827430
TYR rs10765197
TYR rs1827430
TYR rs10765197
TYR rs2000554
TYR rs10830250
TYR rs10830250
TYR rs10765197
TYR rs2000554
TYR rs10830250
TYR rs2000554
TYR rs10765197
TYR rs10830250
TYR rs10765197
TYR rs10765197
TYR rs621313
TYRP1 rs2733831
TYRP1 rs2762464
TYRP1 rs683
TYRP1 rs2762462
TYRP1 rs1063380
TYRP1 rs2733831
TYRP1 rs2733831
TYRP1 rs683
TYRP1 rs2733831
TYRP1 rs1063380
TYRP1 rs683
TYRP1 rs2762464
TYRP1 rs2762464
TYRP1 rs683
TYRP1 rs683
COMT rs2020917


SNP ID
rs10830250
rs621313
rs621313
rs12791412
rs594647
rs1827430
rs1827430
rs1827430
rs621313
rs621313
rs621313
rs594647
rs594647
rs594647
rs2000554
rs10830250
rs2000554
rs594647
rs2762462
rs2762462
rs2762462
rs2733833
rs2762462
rs1063380
rs2762464
rs2733831
rs2733833
rs2733833
rs1063380
rs1063380
rs2733833
rs2733833
rs2762464
rs6269


D'
0.92
0.94
0.88
0.99
0.87
0.85
0.86
0.85
0.96
0.96
0.94
0.96
0.93
0.97
0.96
0.98
0.98
1.00
0.98
1.00
1.00
1.00
1.00
0.91
0.88
0.91
0.96
0.92
0.98
1.00
0.98
0.99
0.99
0.79









tests, but only whites (212 independent cases and 245 unrelated controls) were used for all other

analyses due to power limitations. The positive associations are discussed in the following two

chapters. Described here are the completely negative results found. It is important to report

negative results in order to narrow the pool of possible vitiligo susceptibility genes and thereby

better focus limited research resources.

Genotype by Sex Analyses

Genotype by sex analysis both within cases and between cases and controls yielded no

significant results for any of the genes in my study (data not shown). Pairwise haplotype

analysis examining sex differences by genotype also yielded no significant results (data not

shown). This is not surprising considering that vitiligo appears to occur with equal frequency in

both men and women.

FBXO11 and MSH6

Allelic frequencies for FBXO11 and MSH6 were calculated for my white vitiligo case and

control populations, and the minor allele frequencies (MAF) from the HapMap project (2003)

and the Applied Biosystems validation set (De La Vega et al., 2006) are given, if available

(Table 2-4). Allelic frequencies of vitiligo patients and healthy controls did not differ

significantly for either rs3136367 [p=0.13, uncorrected(UC)] and rs960106 (p=0.95, UC). In

addition, the two MSH6 and FBX011 SNPs tested did not have significantly different genotype

frequencies for the vitiligo patients and controls. The uncorrected p-values for rs3136367 and

rs960106 for the genotype analysis were 0.21 and 0.98 respectively, indicating that there is no

significant difference between cases and controls. The test for Hardy-Weinberg showed that

there was disequilibrium for controls for the rs3136367 SNP (p=0.01, UC), though the

interpretation of this finding is not clear. It may be due to genotyping error; however, deviations

from HWE have been observed in many associations studies (Salanti et al., 2005). Family-based









CHAPTER 5
CONCLUSIONS AND FUTURE DIRECTIONS

This study examined the association of 30 SNPs across eight genes in relation to vitiligo

susceptibility. There are two popular theories underlying the pathogenesis of vitiligo:

autoimmunity and autocytotoxicity. The genes in my study were chosen because of their

possible involvement in one or both of these mechanisms, which are hypothesized to lead to the

melanocyte inactivation and/or destruction observed in vitiligo. It is plausible that both the

autoimmune and autocytotoxic hypotheses contribute to vitiligo pathogenesis. Different patient

subpopulations may have a more significant contribution from one or the other mechanism,

which may lead to a "dirty phenotype" and confound association studies such as ours. The role

of environmental triggers contributing to vitiligo etiology has been well documented. The

interplay of genetics and environment, although difficult to elucidate, is likely to be an important

component of the disease process.

My study provides strong evidence for genetic association with AIRE and suggests

possible genetic association with COMT, PAH, DCT, TYR, and/or TYRP1. If these associations

are confirmed in other patient sets, functional studies of vitiligo susceptibility genes may lead to

the design of novel strategies to prevent and/or treat vitiligo, other autoimmune diseases, and

melanoma.

FBX011 and MSH6

No evidence of association to vitiligo was found for FBX011 and MSH6. This finding

may be a 'true negative' if these genes do not play a significant role in vitiligo pathogenesis.

However, if these genes play a more subtle role in vitiligo etiology or if the markers I chose were

not closely linked to informative or causative changes on these genes, future studies may reveal

that FBXO11 and MSH6 do play a role in vitiligo susceptibility. This suggests that aberrant









SNP was not significant in allelic or family-based testing, additional studies are warranted to

replicate this finding and rule out the possibility of spurious association.

Pairwise haplotype analysis of the TYRP1 gene was particularly interesting. Of the six

SNPs genotyped, five were significant in at least one pairwise association. Three of these SNPs

were intronic and two were located in the 3'UTR. A total of six pairwise comparisons yielded

significant results. Four of these haplotypes were significant for one risk allele compared to

none, with an odds ratio >1, indicating a predisposition to disease. However, all six haplotypes

were significant for two risk alleles compared to one, with an odds ratio <1. This indicates that

homozygosity for these haplotypes is protective, but there is increased heterozygosity observed

in my vitiligo patient population (suggestive of a dominant-negative effect). A similar dominant-

negative finding was found in this same vitiligo cohort for a positively-associated SNP in the

catalase gene (Casp et al., 2002). Because tyrosinase-related protein 1 is believed to function as

part of a multienzyme complex in the melanosome, heterozygosity for changes in the TYRP1

gene may ultimately affect binding, stability, or function of this protein complex. Dominant-

negative effects are well understood in other systems such, as osteogenesis imperfecta involving

collagen multimers (Gajko-Galicka, 2002).

Catechol-O-Methyltransferase

Individually, the two COMT SNPs tested did not yield significant results; however, there

was a significant protective haplotype for individuals who were carried a C allele for a SNP in

the 5'UTR (rs2020917) and a G allele for a SNP in intron two (rs6269). Because neither of the

SNPs were significant in separate analyses, the [CG] haplotype may be in linkage disequilibrium

with a functional change that helps to prevent vitiligo, but this linkage may be weak so only the

combination of both alleles is strong enough to be detected. The homozygosity for the [CG]

haplotype was rare in both vitiligo patients (2%) and controls (3%), so only comparisons









Lerner, A.B. and Nordlund, J.J.: Vitiligo. What is it? Is it important? JAMA 239 (1978) 1183-7.

Levine, E.L. and Ribeiro, G.G.: Vitiligo and radiotherapy: the Koebner phenomenon
demonstrated in patients with vitiligo undergoing radiotherapy for carcinoma of the
breast. Clin Oncol (R Coll Radiol) 6 (1994) 133-4.

Lewontin, R.C.: The interaction of selection and linkage. I. General considerations; heterotic
models. Genetics 49 (1964) 49-67.

Li, K., Li, C., Gao, L., Yang, L., Li, M., Liu, L., Zhang, Z., Liu, Y. and Gao, T.: A functional
single-nucleotide polymorphism in the catechol-O-methyltransferase gene alter vitiligo
risk in a Chinese population. Arch Dermatol Res (2008) [Epub ahead of print].

Li, M. and Li, C.: Assessing departure from Hardy-Weinberg equilibrium in the presence of
disease association. Genet Epidemiol 32 (2008) 589-99.

Liston, A. and Goodnow, C.C.: Genetic lesions in thymic T cell clonal deletion and thresholds
for autoimmunity. Novartis Found Symp 267 (2005) 180-92; discussion 192-9.

Liston, A., Gray, D.H., Lesage, S., Fletcher, A.L., Wilson, J., Webster, K.E., Scott, H.S., Boyd,
R.L., Peltonen, L. and Goodnow, C.C.: Gene dosage--limiting role of AIRE in thymic
expression, clonal deletion, and organ-specific autoimmunity. J Exp Med 200 (2004)
1015-26.

Liu, T., Johnson, J.A., Casella, G. and Wu, R.: Sequencing complex diseases With HapMap.
Genetics 168 (2004) 503-11.

Lotti, T., Gori, A., Zanieri, F., Colucci, R. and Moretti, S.: Vitiligo: new and emerging
treatments. Dermatol Ther 21 (2008) 110-7.

Majumder, P.P., Nordlund, J.J. and Nath, S.K.: Pattern of familial aggregation of vitiligo. Arch
Dermatol 129 (1993) 994-8.

Manga, P., Sheyn, D., Yang, F., Sarangarajan, R. and Boissy, R.E.: A role for tyrosinase-related
protein 1 in 4-tert-butylphenol-induced toxicity in melanocytes: Implications for vitiligo.
Am J Pathol 169 (2006) 1652-62.

Manolache, L., Petrescu-Seceleanu, D. and Benea, V.: Correlation of stressful events with onset
of vitiligo in children. J Eur Acad Dermatol Venereol 23 (2009) 187-8.

Maresca, V., Roccella, M., Roccella, F., Camera, E., Del Porto, G., Passi, S., Grammatico, P. and
Picardo, M.: Increased sensitivity to peroxidative agents as a possible pathogenic factor
of melanocyte damage in vitiligo. J Invest Dermatol 109 (1997) 310-3.

Mathis, D. and Benoist, C.: Aire. Annu Rev Immunol 27 (2009) 287-312.

McGinnis, R.E., Ewens, W.J. and Spielman, R.S.: The TDT reveals linkage and linkage
disequilibrium in a rare disease. Genet Epidemiol 12 (1995) 637-40.










Table 4-25. Pairwise haplotype analysis results for TYRP1 rs2762462 and TYRP1 rs2733833
Frequency Controls Vitiligo patients p-values
0 [CG] 36(15%) 20(9.4%)
1 [CG] 105 (43%) 115(54.3%)
2 [CG] 104 (42%) 77 (36.3%)
Total (n) 245 212
X2 0.0350*
Odds ratio for 1 [CG] vs. 0 [CG] = 1.9425 95% CI=1.0570 to 3.5698 0.0325*
Odds ratio for 2 [CG] vs. 0 [CG] = n.s. n.s.
Odds ratio for 2 [CG] vs. 1 [CG] = 0.6720 95% CI=0.4520 to 0.9991 0.0495*
n.s. = not significant; denotes a significant finding; number in frequency column indicates the
number of copies of this haplotype


Table 4-26. Pairwise haplotype analysis results for COMTrs2020917 and COMTrs6269
Frequency Controls Vitiligo patients p-values
0 [CG] 162 (66%) 169 (80%)
1 [CG] 76 (31%) 38 (18%)
2 [CG] 7 (3%) 5 (2%)
Total (n) 245 212
X2 0.0046*
Odds ratio for 1 [CG] vs. 0 [CG] = 0.4750 95% CI=0.3040 to 0.7423 0.0011*
Odds ratio for 2 [CG] vs. 0 [CG] = n.s. n.s.
Odds ratio for 2 [CG] vs. 1 [CG] = n.s n.s.
n.s. = not significant; denotes a significant finding; number in frequency column indicates the
number of copies of this haplotype










Table 4-5. Family-based association (transmission disequilibrium test) results for DCT, PAH,


Gene
DCT
DCT
DCT
DCT
DCT
DCT
DCT
DCT
PAH
PAH
TYR
TYR
TYR
TYR
TYR
TYR
TYR
TYR
TYRP1
TYRP1
TYRP1
TYRP1
TYRP1
TYRP1
COMT
COMT


r
r
r
r
r
r
r
r
r
r
r
r
r
r
r
r
r
r
r
r
r
r
r
r
r
r


* denotes a significant finding; BonC denotes the Bonferroni corrected p-value


Table 4-6. Pairwise haplotype analysis results for PAHrs1722381 and rs1522307
Frequency Controls Vitiligo patients p-values
0 [GG] 215 (89%) 165 (81%)
1 [GG] 26 (11%) 27(13.5%)
2 [GG] 0 (0%) 11(5.5%)
Total (n) 241 203
X2 0.0007*
Odds ratio for 1 [GG] vs. 0 [GG] = n.s. n.s
Odds ratio for 2 [GG] vs. 0 [GG] = n.d. n.d.
Odds ratio for 2 [GG] vs. 1 [GG] = n.d. n.d.
n.s. = not significant; n.d. = not defined; denotes a significant finding; number in frequency
column indicates the number of copies of this haplotype


TYR, TYRP1, and COMT
Number of
informative
SNP parents
s9516413 62
s7987802 59
s4318084 67
-s9524493 74
s9516418 79
s1028805 74
s11618471 54
s7991232 70
s1722381 43
s1522307 54
s1042602 77
s621313 82
-s594647 75
s10765197 71
s12791412 62
s2000554 73
s10830250 80
s1827430 77
s2762462 60
s2733831 70
s2733833 58
-s683 68
-s2762464 67
s1063380 67
s2020917 47
-s6269 50


Transmitted
(allele)
42 (C)
29 (C)
52 (C)
34 (C)
28 (C)
50 (G)
44 (A)
36 (A)
23 (A)
28 (A)
34 (A)
38 (G)
33 (T)
42 (A)
40 (A)
32 (G)
45 (C)
33 (G)
30 (C)
36 (A)
32 (G)
35 (A)
34 (A)
30 (C)
22 (T)
20 (G)


Not
transmitted
(allele)
20 (T)
30 (T)
15 (T)
40 (G)
51(T)
24 (T)
10 (G)
34 (G)
20(G)
26(G)
43 (C)
44 (A)
42 (C)
29(A)
22 (G)
41(A)
35 (G)
45 (A)
30 (T)
34(G)
26 (T)
33 (C)
33 (T)
37 (T)
25 (C)
30(A)


p-value
0.0052*
0.90
6.20E-06*
0.49
0.0097*
0.0025*
3.70E-06
0.81
0.65
0.79
0.31
0.51
0.30
0.12
0.02*
0.29
0.26
0.17
1.00
0.81
0.43
0.81
0.90
0.39
0.66
0.16


BonC
p-value
0.16
1.00
0.0002*
1.00
0.29
0.08
0.0001*
1.00
1.00
1.00
1.00
1.00
1.00
1.00
0.60
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00
1.00


FDR
q-value
0.039*
0.96
<0.0001*
0.81
0.06
0.025*
<0.0001*
0.93
0.90
0.93
0.62
0.80
0.62
0.45
0.10
0.62
0.62
0.46
1.00
0.93
0.76
0.93
0.96
0.73
0.90
0.46









between individuals who were carried one protective haplotype versus those who carried zero

reached significance.

It is noteworthy that other SNPs in COMT have been found to be positively associated to

vitiligo. One group observed that homozygosity for a SNPs in COMT causing an amino acid

change (from valine to methionine at position 158), which lowers enzymatic activity, was

associated with acrofacial vitiligo in a Turkish population (Tursen et al., 2002). Another group

found this same SNP to be positively associated with an increase risk for generalized vitiligo in a

Chinese population, but not with segmental or acrofacial vitiligo (Li et al., 2008). Because

COMT enzyme may play a dual role in the melanocyte, both in the regulation of melanin

biosynthesis and in limiting damage from reactive oxygen species (ROS), the COMTgene

remains a possible vitiligo susceptibility gene. The results from my study and the work of others

have shown SNPs in COMTto be associated with vitiligo in three different populations, adding

support to this gene's contribution to disease etiology.









Gilhar, A., Zelickson, B., Ulman, Y. and Etzioni, A.: In vivo destruction of melanocytes by the
IgG fraction of serum from patients with vitiligo. J Invest Dermatol 105 (1995) 683-6.

Giovannelli, L., Bellandi, S., Pitozzi, V., Fabbri, P., Dolara, P. and Moretti, S.: Increased
oxidative DNA damage in mononuclear leukocytes in vitiligo. Mutat Res 556 (2004)
101-6.

Gomes, I., Collins, A., Lonjou, C., Thomas, N.S., Wilkinson, J., Watson, M. and Morton, N.:
Hardy-Weinberg quality control. Ann Hum Genet 63 (1999) 535-8.

Gottschalk, G.M. and Kidson, S.H.: Molecular analysis of vitiligo lesions reveals sporadic
melanocyte survival. Int J Dermatol 46 (2007) 268-72.

Halaban, R., Patton, R.S., Cheng, E., Svedine, S., Trombetta, E.S., Wahl, M.L., Ariyan, S. and
Hebert, D.N.: Abnormal acidification of melanoma cells induces tyrosinase retention in
the early secretary pathway. J Biol Chem 277 (2002) 14821-8.

Hann, S.K., Park, Y.K., Lee, K.G., Choi, E.H. and Im, S.: Epidermal changes in active vitiligo. J
Dermatol 19 (1992) 217-22.

Harning, R., Cui, J. and Bystryn, J.C.: Relation between the incidence and level of pigment cell
antibodies and disease activity in vitiligo. J Invest Dermatol 97 (1991) 1078-80.

Hochberg, Y. and Benjamini, Y.: More powerful procedures for multiple significance testing.
Stat Med 9 (1990) 811-8.

Hu, Z., Liu, J.B., Ma, S.S., Yang, S. and Zhang, X.J.: Profile of childhood vitiligo in China: an
analysis of 541 patients. Pediatr Dermatol 23 (2006) 114-6.

Hughes, L.B., Beasley, T.M., Patel, H., Tiwari, H.K., Morgan, S.L., Baggott, J.E., Saag, K.G.,
McNicholl, J., Moreland, L.W., Alarcon, G.S. and Bridges, S.L., Jr.: Racial or ethnic
differences in allele frequencies of single-nucleotide polymorphisms in the
methylenetetrahydrofolate reductase gene and their influence on response to methotrexate
in rheumatoid arthritis. Ann Rheum Dis 65 (2006) 1213-8.

Ikari, K., Momohara, S., Nakamura, T., Hara, M., Yamanaka, H., Tomatsu, T. and Kamatani, N.:
Supportive evidence for a genetic association of the FCRL3 promoter polymorphism with
rheumatoid arthritis. Ann Rheum Dis 65 (2006) 671-3.

Im, S., Hann, S.K., Kim, H.I., Kim, N.S. and Park, Y.K.: Biologic characteristics of cultured
human vitiligo melanocytes. Int J Dermatol 33 (1994) 556-62.

Jimbow, K., Chen, H., Park, J.S. and Thomas, P.D.: Increased sensitivity of melanocytes to
oxidative stress and abnormal expression of tyrosinase-related protein in vitiligo. Br J
Dermatol 144 (2001) 55-65.











Table 3-1. List of Applied Biosystems SNP Assay IDs and description ofAIRE SNPs genotyped
Base pair change & location
Gene SNP rs number Applied Biosystems ID (amino acid change & location)
AIRE rs2776377 C 2978271 1 A/G promoter
AIRE rs1800520 C 9480541 1 C/G exon 7 (S278R)


Table 3-2. White case-control results for genotype analysis ofAIRE
BonC FDR
SNP Case Control p-value p-value q-value
AA 54(33%) AA 86 (37%)
rs2776377 AG 83 (51%) AG 104 (45%)
GG 26(16%) GG 43 (18%) 0.46 1.00 0.55
CC 130(80%) CC 130(55%)
rs1800520 CG 33 (20%) CG 88 (37%)
GG 0 (0%) GG 19 (8%) 0.038* <0.0001* <0.0001*
* denotes a significant finding; BonC denotes the Bonferroni corrected p-value









1993b; Le Poole et al., 1996). IL-2 and IFN-y are also expressed in cellular infiltrates,

suggesting that activation of Thl cells may have occurred (Abdel-Naser et al., 1994). Other

work suggests that not all melanocytes are destroyed in the lesion; however, those remaining

have been rendered dysfunctional (Tobin et al., 2000; Gottschalk and Kidson, 2007).

Melanoblasts remain in hair follicles, however, as any repigmentation that does occur either

spontaneously or in response to treatments, begins in small spots around hair follicles, spreading

until repigmented areas coalesce (Arrunategui et al., 1994).

Further evidence of an abnormal immune response is a lowered CD4+/CD8+ ratio in

perilesional skin in vitiligo patients compared to controls (Mozzanica et al., 1990; Le Poole et

al., 1996). The increased numbers of CD8+ cells observed in vitiligo lesions has led LePoole and

colleagues to hypothesize that cytotoxic CD8+ cells are responsible for melanocyte destruction

(1996). A decrease in CD45RA T cells, which are naive T cells, has been observed in vitiligo

patient peripheral blood (Abdel-Naser et al., 1992). Disruptions in Langerhans cells in vitiligo

have also been reported in skin biopsies, with decreased numbers seen in active vitiligo and a

return to normal numbers in stable vitiligo (Kao and Yu, 1990). An unusual feature of vitiligo is

the abnormal expression of class II human leukocyte antigen (HLA) molecules by perilesional

melanocytes in about two-thirds of vitiligo patients, accompanied by a six-fold increase in the

expression of intercellular adhesion molecule-1 (ICAM-1) (al Badri et al., 1993). Peripheral

blood of active vitiligo patients has been shown to have increased concentrations of pro-

inflammatory cytokines IL-6 and IL-8 as well as a decreased production of GM-CSF, TNF-a,

and IFN-y (Yu et al., 1997). It is important to note that variations in published data on peripheral

cell imbalances may be due to a variety of factors such as differences in study populations,









DNA repair (at least via MSH6) is not a strong genetic component of vitiligo susceptibility.

However, FBX011 was observed to be downregulated in vitiligo patients compared to healthy

controls while MSH6 was upregulated (Le Poole et al., 1994). Because these two genes overlap

on their 3' ends, it is possible that double-stranded RNA intermediates are interfering in protein

expression via the RNAi mechanism. Alternatively, another mechanism that is yet to be

elucidated may be underlying this altered expression pattern.

Autoimmune Regulator (AIRE)

My study demonstrated significant genetic association of vitiligo with the autoimmune

regulator gene, or AIRE. There is a substantial body of research supporting immune system

dysfunction in vitiligo patients. Because the AIRE gene is vital to the negative selection of

autoreactive T cells, it is possible that subtle defects in AIRE allow for the escape of autoreactive

T cells from the thymus, contributing to the autoimmunity hypothesized to cause vitiligo. I

observed a significant association between vitiligo and a SNP causing an amino acid change in

the SAND DNA-binding domain of AIRE, even when the stringent Bonferroni correction was

applied. Pairwise haplotype analysis with this non-synonymous SNP and an intronic one also

yielded a significant finding. However, the family-based association studies were not

significant. This is not unusual in association studies (especially given our limited number of

families), so it does not negate the case-control result. Additional studies, using protein

modeling and/or functional analysis, are needed to help understand the significance of this serine

to arginine amino acid change and its influence on the SAND domain. For example, it is not

known whether serine is a site for phosphorylation, which could significantly affect protein

function. It is possible that the non-synonymous SNP I examined is directly related to vitiligo

genetic susceptibility. If not causative, this SNP and the significant AIRE haplotype I observed

may be linked to a polymorphism that reduces normal immune function in vitiligo patients.

























DCT
40,083bp


B

.oo t CD
CB CT Cp '0 C 1







TYRPI
16,02 bp


Figure 4-1. Schematic representation of gene structure, SNP location, and HapMap CEPH haploblocks. Exons are represented by
black boxes, introns are represented by white boxes, arrows show SNP location, and haploblocks are shown by shaded
boxes. Note: haploblocks extend off of the ends of genes. A) Dopachrome tautamerase (DCT), 40,083base pairs (bp).
B) Tyrosinase-related protein 1 (TYPRI), 16,802bp.









alterations in gene function could have an effect on melanogenesis by decreasing the availability

of tyrosine (Schallreuter et al., 1998).

Essential to melanogenesis are several genes in the tyrosinase gene family that are

expressed in melanocytes: tyrosinase (TYR), tyrosinase-related protein 1 (TYRP1) and

dopachrome tautamerase (DCT), (formerly known as tyrosinase-related protein 2 (TYRP2)).

Disruptive recessive mutations in TYR and TYRP1 are linked to different types of oculocutaneous

and ocular albinism; this supports their essential role in pigmentation and as candidates for

involvement in vitiligo (Oetting and King, 1999). No known human phenotypes have been

linked to changes in DCT. However, DCT along with TYR and TYRP1 proteins were observed

by several groups to be possible targets of autoantibodies found in the serum of vitiligo patients

(Song et al., 1994; Kemp et al., 1997a; Kemp et al., 1997b; Kemp et al., 1998; Okamoto et al.,

1998; Kemp et al., 1999b). This further supports the hypothesis of these genes contributing to

vitiligo susceptibility.

The five-exon TYR gene is located at chromosome 1 1q14-q21 and encodes the 529 amino

acid tyrosinase enzyme, with 50% of the coding region being in exon one. Tyrosinase has a dual

function in melanin production. It oxidizes tyrosine to dopa in the first and rate-limiting step in

the melanin biosynthetic pathway, and it oxidizes dopa to dopaquinone, an intermediate step in

the melanin pathway. Tyrosinase is hypothesized to be part of a multienzyme complex in the

melanosome that includes tyrosinase- related protein 1 and dopachrome tautamerase. The

TYRP1 and DCT enzymes are thought to be involved in the latter part of the melanin

biosynthetic pathway, though their exact role(s) in melanin biosynthesis have yet to be elucidated

(Oetting, 2000). However, it is known that TYRP1 significantly enhances tyrosinase catalytic

function (Kobayashi et al., 1994; Kobayashi et al., 1998). TYRP1 is located on chromosome




































To my Mother









analyses were performed using one degree of freedom for the allelic frequency test and two

degrees of freedom for the genotype frequency test. Chi-squared (X2) analysis is represented by

the following equation:

X2= X[(Expected Value Obtained Value)2]/Expected Value

The Microsoft Excel 2003 software package was used for X2 analysis for allele frequency values

for the white case-control analysis. The SAS software package was used for X2 analysis for

genotype frequency values for the white case-control groups.

The strength of allelic association can also be measured by calculating an allele's relative

risk (RR):

RR = [a(b+d)/b(a+c)],

where "a" and "b" are the numbers of patients and controls with a given allele respectively, and

"c" and "d" are the corresponding numbers without that allele. The greater the relative risk, the

more frequently the allele is found among vitiligo patients and the less frequently it is found

among controls. A predisposing effect is indicated by a RR that is significantly higher than 1; a

protective effect is indicated by a RR significantly smaller than 1. A neutral effect is indicated

by a RR that is not significantly different from 1. Significance for RR is given by a 95%

confidence interval (CI) that does not include 1.

Family-Based Testing

To further validate statistical significance determined by the case-control analysis, the

transmission disequilibrium test (TDT), a family-based association test, was performed

(Spielman et al., 1993; McGinnis et al., 1995; Spielman and Ewens, 1996). The TDT examines

whether alleles of the marker of interest are transmitted at a frequency greater than 50% from an

informative parent to an affected child. In the TDT, an informative parent is defined as a parent









Negative association results are discussed in this chapter, with positive results in chapters 3 and

4.

Vitiligo Candidate Gene Selection

Pubmed (http://www.ncbi.nlm.nih.gov/) was used to search current literature to identify

putative vitiligo genes including those linked to autoimmunity, melanin biosynthesis, and

oxidative stress. The candidate gene approach allows researchers to focus their investigation by

examining a limited number of genes that are selected based on the specific pathophysiology of a

disease. Candidate gene studies do not need large multi-generation families, but can be

performed using unrelated groups of patients (cases) and unaffected controls, or through small

families. Candidate gene studies also have the advantage of being better suited for determining

susceptibility genes in more complex disorders where the relative risk associated with any one

gene is relatively small (Kwon and Goate, 2000). Because it has been hypothesized that vitiligo

pathogenesis could be a direct result of autoimmunity, and/or autotoxicity due to biochemical

defects in the melanocytes, genes that regulate the immune system, melanocyte biochemistry,

and oxidative stress were deemed suitable choices for candidate genes.

Autoimmune Regulator (AIRE)

Disruptive mutations in the AIRE gene are responsible for autoimmune

polyendocrinopathy-candidiasis-ectodermal dystrophy syndrome (APECED). The most

common inheritance pattern for APECED is autosomal recessive (Rizzi et al., 2006), though an

autosomal dominant inheritance pattern has also been reported (Cetani et al., 2001). The most

common features of APECED are parathyroid gland failure, chronic susceptibility to candida

yeast infection, and Addison's disease. Other autoimmune diseases associated with APECED

may include alopecia, vitiligo, ovarian failure, testicular atrophy, hypothyroidism, gastric parietal

cell atrophy, hepatitis, intestinal malabsorption, and insulin-dependent diabetes mellitus.









The inconsistencies observed in my study concerning the significance associated with one

marker for a particular gene, but not confirmed by other markers for the same gene, or by TDT

analysis, may be a function of several factors. These incongruities do not necessarily exclude

any of the genes studied as candidates for involvement in vitiligo; however, they do necessitate

that the results of my study be considered with relative caution until the suggested associations

can be confirmed by replication or other methodologies. Lack of support by a significant TDT

for associations suggested by case-control analysis may also be a result of insufficient number of

informative families. The location of SNPs may also influence whether or not a marker reveals

an association. A marker at one end of a gene may not show significant association if a

pathogenic mutation is located at the other end of the gene. This may be one reason why

haplotype analysis of SNPs yielded different information for some of the genes examined, as

multiple markers may give different information about pathogenic changes in a gene.

If positive associations are found, additional investigation into the genetic factors

underlying vitiligo susceptibility should be conducted on an increased number of subjects, on

subjects of different ethnic backgrounds in case-control analysis, and on an increased number of

families for TDT analysis. Sequence analysis could then be conducted to elucidate the nature of

the genetic characteristics) suggested by case-control and TDT analysis to influence the

development of the vitiligo phenotype. Additional methodologies, like gene expression

experiments and protein structure studies, are needed to confirm results and better elucidate a

protein's function in disease pathogenesis.

Another potential confound is that association may be due to population stratification,

where the case and controls groups may be from different ethnic or geographic subgroups.

Although every effort was made to match cases and controls in my study, it is possible that some









association. It is noteworthy that in my haploblock analysis of TYRP1, I observed a greater

number of heterozygotes in my vitiligo patients compared to controls. This heterozygosity of the

TYRP1 gene may result in a destabilization of the enzyme complex similar to a dominant-

negative effect. An increase in heterozygosity in vitiligo patients was previously observed for a

SNP in the catalase gene, where a dominant-negative effect was also hypothesized for that

tetrameric protein complex (Casp et al., 2002).

The catechol-O-methyltransferase or COMT gene was shown to have a protective

haplotype in my white case-control analysis. This supports the idea that the COMT enzyme

functions to protect the melanocyte during melanogenesis. A reduction of COMT activity or an

alteration of COMT expression in vitiligo patients may contribute to the autocytotoxic model of

vitiligo (Le Poole et al., 1994; Tursen et al., 2002; Li et al., 2008). Based on my finding and the

work of others, further exploration of the COMT gene in additional genetic association studies is

warranted to uncover its role in genetic susceptibility to vitiligo. Validated association could be

followed by functional studies.

Final Thoughts and a Note of Caution

In any association study including ours, there is the possibility that spurious associations

may be found. This is especially true for complex diseases where multiple genes may have

subtle effects on disease pathogenesis. Because several of the p-values for suggested

associations were only moderate (0.01
reported are spurious. My study case-control analysis consisted of 212 white vitiligo patients

and 245 unrelated controls with no known autoimmune diseases. Investigations involving a

larger cohort may be able to yield more highly significant results and reveal spurious

associations.









TABLE OF CONTENTS

page

A C K N O W L E D G M E N T S ..............................................................................................................4

LIST O F TA B LE S ......... ..... .............. ................................................................. 9

LIST O F FIG U R E S ......... .... ............................. .... .. ...... ................ 11

ABSTRAC T ............................... ..................... 12

CHAPTER

1 INTRODUCTION ............... .......................................................... 14

M elanocytes and P igm entation ....................................................................... ..................15
Clinical Description of Vitiligo ..................................................................... 16
V itiligo Pathogenesis .................. ................... ................. ............ .. ............ 17
Autoim m une H hypothesis ........................................................... .. ............... 18
H um oral autoim m unity .................................................. .............................. 18
C cellular autoim m unity ........ .......................................................... ...... .... ....... 19
V itiligo and m elanom a .............................................................................21
A utocytotoxic H hypothesis ....................................................................... ..................22
T reatm en t ................... ...................2...................7..........
G genetics of V itiligo ............................................................................................... ....... 28

2 EXPERIMENTAL DESIGN: CASE-CONTROL AND FAMILY-BASED
A SSO C IA T IO N STU D IE S ........................................................................... ...................3 1

Introdu action ............................................................................................ 3 1
V itiligo C candidate G ene Selection .............................................................. .....................32
Autoimmune Regulator (AIRE) ........................... ........................ ...............32
M elanin B iosynthesis G enes ........................................ ............................................33
D N A R repair G enes .................................................... ............ ............33
M materials and M methods ................................... ... .. .......... ....... ...... 33
S u objects ..................................................................................................3 3
Blood Processing and DNA Extraction............................................... ........ ....... 34
Single Nucleotide Polymorphism (SNP) Selection ............. ........................................34
G en oty ping ..............................................................................36
Statistical A n aly sis ....................................................... ...................36
H ardy-W einberg A nalysis............................................ .................... ...............36
C ase-C control A naly sis........... .............................................................. ......... .......... 37
F am ily-B ased T testing ................ ..... .................................................................. .. .. ..38
Genotype by Co-Morbid Autoimmune Diseases .................................... ...............39
G enotype by Sex .........................................................................39
G enotype by A ge of O nset ........................................... .................. ............... 40
Pairw ise H aplotype A nalysis................................................. .............................. 40




Full Text

PAGE 1

THE ROLE OF GENETICS IN VITILIGO SUSCEPTIBILITY By DEBORAH MARSHA HERBSTMAN A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLOR IDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2009 1

PAGE 2

2009 Deborah Marsha Herbstman 2

PAGE 3

To my Mother 3

PAGE 4

ACKNOWLEDGMENTS I am grateful to all of the individuals who partic ipated in this research study; their generosity has made this work possible. I thank the American Vitiligo Research Foundation ( www.AVRF.org ) for the opportunity to learn from their members and families, and for financial support of this research. I would like to acknowledge AVRFs founder and president, Stella Pavlides, for her encouragement and her energetic enthusiasm. I also wish to thank Marilyn Rose Giordano; although she is no longer with us, I am inspired by the memory of her selfless dedication to helping individuals, especially chil dren, with vitiligo. I would like to thank my mentor, Dr. Margaret R. Peggy Wallace for allowing me work in her lab and grow as a scientist under her guid ance. I am amazed at the depth and breath of her scientific knowledge, and inspired by her co mmitment to outreach and community service through her charitable work with patients and fa milies. I am very luc ky that Peggy appreciates my sense of humor and doesnt mind hearing my laugh from down the hall. I owe most of those laughs to my fantastic lab mates, past and pres ent. Specifically, I would like to acknowledge Michelle Chello Burch for being an amazing technician and a dear friend, Dr. Rebecca Boocca Loda-Hutchinson for moral support and help ful instruction in both science and dance, Dr. Hua Li for laboratory assist ance and enjoyable conversations in which she allowed me to encourage her to practice her English, and Dr. Angela Hadjipanayis for sharing her unique perspective in science and in lif e. I thank Dr. Wayne T. McCormack for his dual roles as my comentor and director of my PhD program. I am grateful for his help a nd advice, both in guiding this project and in my scientific career. Dr McCormacks dedication to his students and to community service has truly been an inspiration; from one scout to another, I salute him. I would like to thank my committee member Dr Daniel C. Driscoll for being a mentor both in the lab and in the clinic, for interesting conversations about genetics and politics, and for 4

PAGE 5

his excellent sense of humor. I thank Dr. Cynt hia W. Cyndi Garvan for being an exceptional teacher, an enthusiastic collaborator, a role model, and my committee member. I would like to acknowledge Dr. Sally Litherland, whos e advice and guidance helped co ntribute to this project. I am grateful to my statistical collaborator Dr. Wei Hou; it is a pleasure to work with him. I would like to thank the UF Center for Pharm acogenomics, especially the Director of the Genotyping Core Laboratory, Dr. Taim our Langaee, and his talented technician, Lynda Stauffer. Pandora Cowart has been both an expert database consultant and a friend. I thank my colleague and friend Amy Non for helpful comments and advi ce. I am grateful to Dr. Roger Fillingim for his collaboration and his support of my res earch. I acknowledge Dr. Steve Blackband for computer support. I would like to thank the skilled and friendly staff of the Molecular Genetics and Microbiology Department, especially J oyce Conners, Debbie Burgess, Steve Howard, Michele Ramsey, Julie Dillard, and Dave Brumbaugh. Dr. Henry Baker, my Department Chairman, is gratefully acknowledged for hi s scientific and career advice. My graduate education has been made all the richer and saner thanks to a dearly loved group of friends. In my heart, I thank you all. I am grateful for Mais ara and Jim Bledsoe and Sasha and Bo Yurke, as they are friends that have grown to be family. I thank Dr. Adam Rivers for many late-night phone calls offering PhD support and humorous anecdotes about life. I thank soon-to-be-a-doctor Andrea Dr. Dre Rivers for listening to my crazy st ories and for sharing a few of her own. Dan BFw/oB Capellupos kindne ss and patience is truly appreciated; he is a great dance partner and a cherishe d friend. I thank Stacey Simon for her love of stories and for going along with a small sample set of my silly suggestions. I appreciate Pamela Ella Dubyak for sharing my love of the ridiculous. I thank my friend and trivia buff extraordinaire Dr. Dimitri Veras, who helped to copy edit this manuscript. Michelle Edwards and David Edmeades offered 5

PAGE 6

encouragement and lots of good cheese. I am gr ateful for Tarek Saab, whose unique charm and good humor left a mark on me. If I have seen a li ttle further it is by stan ding on the shoulders of astronomers because they look at things that are real ly, really far away. It has been a pleasure to be a groupie to UFs Astronomy department; a ni cer and smarter group of people it would be difficult to find. Saving the best for last, I thank my family. I appreciate their love, their patience, their exquisite humor, and their exceptional DNA. My Father was my bi ggest fan and always encouraged me to excel in academics. I am gratef ul to Joshua Tobias Herbstman; I could not ask for a more supportive big brother. My Mother is my role model, my best friend, and the one who always thought that I would make a good scientist. 6

PAGE 7

TABLE OF CONTENTS page ACKNOWLEDGMENTS...............................................................................................................4 LIST OF TABLES................................................................................................................. ..........9 LIST OF FIGURES.......................................................................................................................11 ABSTRACT...................................................................................................................................12 CHAPTER 1 INTRODUCTION................................................................................................................. .14 Melanocytes and Pigmentation...............................................................................................15 Clinical Description of Vitiligo..............................................................................................1 6 Vitiligo Pathogenesis.......................................................................................................... ....17 Autoimmune Hypothesis.................................................................................................18 Humoral autoimmunity............................................................................................18 Cellular autoimmunity..............................................................................................19 Vitiligo and melanoma.............................................................................................21 Autocytotoxic Hypothesis...............................................................................................22 Treatment................................................................................................................................27 Genetics of Vitiligo........................................................................................................... ......28 2 EXPERIMENTAL DESIGN: CASE-CONTROL AND FAMILY-BASED ASSOCIATION STUDIES....................................................................................................31 Introduction................................................................................................................... ..........31 Vitiligo Candidate Gene Selection.........................................................................................32 Autoimmune Regulator ( AIRE ).......................................................................................32 Melanin Biosynthesis Genes...........................................................................................33 DNA Repair Genes..........................................................................................................33 Materials and Methods...........................................................................................................33 Subjects............................................................................................................................33 Blood Processing and DNA Extraction...........................................................................34 Single Nucleotide Polymorphism (SNP) Selection.........................................................34 Genotyping......................................................................................................................36 Statistical Analysis........................................................................................................... .......36 Hardy-Weinberg Analysis...............................................................................................36 Case-Control Analysis.....................................................................................................37 Family-Based Testing......................................................................................................38 Genotype by Co-Morbid Autoimmune Diseases............................................................39 Genotype by Sex..............................................................................................................39 Genotype by Age of Onset..............................................................................................40 Pairwise Haplotype Analysis...........................................................................................40 7

PAGE 8

Corrections for Multiple Testing.....................................................................................41 Results.....................................................................................................................................41 Genotype by Sex Analyses..............................................................................................42 FBXO11 and MSH6 .........................................................................................................42 Discussion...............................................................................................................................43 3 AUTOIMMUNE REGULATOR GENE................................................................................49 Introduction................................................................................................................... ..........49 Materials and Methods...........................................................................................................50 Results.....................................................................................................................................51 Discussion...............................................................................................................................52 4 MELANIN BIOSYNTHESIS GENES...................................................................................58 Introduction................................................................................................................... ..........58 Materials and Methods...........................................................................................................61 Results.....................................................................................................................................61 Phenylalanine Hydroxylase.............................................................................................61 Dopachrome Tautamerase...............................................................................................61 Tyrosinase..................................................................................................................... ...62 Tyrosinase-Related Protein 1..........................................................................................63 Catechol-O-Methyltransferase........................................................................................63 Summary..........................................................................................................................64 Discussion...............................................................................................................................64 Phenylalanine Hydroxylase.............................................................................................64 Dopachrome Tautamerase...............................................................................................65 Tyrosinase..................................................................................................................... ...66 Tyrosinase-Related Protein 1..........................................................................................67 Catechol-O-Methyltransferase........................................................................................68 5 CONCLUSIONS AND FUTURE DIRECTIONS.................................................................85 FBXO11 and MSH6 ................................................................................................................85 Autoimmune Regulator ( AIRE )..............................................................................................86 Melanin Biosynthesis Genes...................................................................................................87 Final Thoughts and a Note of Caution....................................................................................90 Future Directions....................................................................................................................92 APPENDIX A SINGLE NUCLEOTIDE POLYMORPHISM DETAILS.....................................................95 LIST OF REFERENCES...............................................................................................................96 BIOGRAPHICAL SKETCH.......................................................................................................111 8

PAGE 9

LIST OF TABLES Table page 2-1 Demographic information for vitiligo patients and their unaffected relatives...................45 2-2 List of Applied Biosystems SNP Assay IDs and description of FBXO11 and MSH6 SNPs genotyped.................................................................................................................46 2-3 Linkage disequilibrium (D ) values for FBXO11 and MSH6 .............................................46 2-4 Minor allele frequencies (MAF) for FBXO11 and MSH6 SNPs from the HapMap project Caucasians from European Ances try (HapMap CEPH), Applied Biosystems Caucasian cohort (ABI Caucasian), and values for the white case and control groups from the present study; 2 p-values and q-values for a llelic comparisons are given........47 2-5 Family-based association (transmission disequilibrium test) results for FBXO11 and MSH6 .................................................................................................................................48 3-1 List of Applied Biosystems SNP Assay IDs and description of AIRE SNPs genotyped...........................................................................................................................55 3-2 White case-control result s for genotype analysis of AIRE .................................................55 3-3 Minor allele frequencies (MAF) for AIRE SNPs from Applied Biosystems Caucasian cohort (ABI Caucasian) and values for th e white case and control groups from the present study; 2 p-values and q-values for allelic comparisons are given.......................56 3-4 Family-based association (transmission disequilibrium test) results for AIRE .................57 3-5 Linkage disequilibrium (D ) value for AIRE ......................................................................57 3-6 Pairwise haplotype analysis results for AIRE rs2776377 and rs1800520..........................57 4-1 List of Applied Biosystems SNP Assay IDs and description of DCT, PAH, TYR, TYRP1, and COMT SNPs genotyped.................................................................................71 4-2 Linkage disequilibrium (D ) values for DCT PAH TYR TYRP1 and COMT .................72 4-3 Minor allele frequencies (MAF) for DCT PAH TYR TYRP1, and COMT SNPs from the HapMap project Caucasians from Eu ropean Ancestry (HapMap CEPH), Applied Biosystems Caucasian cohort (ABI Caucas ian), and values for the white case and control groups from the present study; 2 p-values and q-values for allelic comparisons are given........................................................................................................74 4-4 Case-control results for genotype analysis of DCT PAH TYR TYRP1, and COMT .......75 4-5 Family-based association (transmission disequilibrium test) results for DCT, PAH TYR, TYRP1 and COMT ...................................................................................................76 9

PAGE 10

4-6 Pairwise haplotype analysis results for PAH rs1722381 and rs1522307...........................76 4-7 Pairwise haplotype analysis results for DCT rs4318084 and rs9516413...........................77 4-8 Pairwise haplotype analysis results for DCT rs11618471 and rs9516413.........................77 4-9 Pairwise haplotype analysis results for DCT rs7991232 and rs9516413...........................77 4-10 Pairwise haplotype analysis results for DCT rs9524493 and rs9516413...........................78 4-11 Pairwise haplotype analysis results for DCT rs7987802 and rs9516413...........................78 4-12 Significant 2 findings for DCT association for patients with co-morbid autoimmune diseases..............................................................................................................................78 4-13 Pairwise haplotype analysis results for TYR rs10765197 and TYR rs1042602..................79 4-14 Pairwise haplotype analysis results for TYR rs1042602 and TYR rs12791412..................79 4-15 Pairwise haplotype analysis results for TYR rs1042602 and TYR rs10830250..................79 4-16 Pairwise haplotype analysis results for TYR rs1042602 and TYR rs2000554.................80 4-17 Pairwise haplotype analysis results for TYR rs1042602 and TYR rs1827430....................80 4-18 Pairwise haplotype analysis results for TYR rs12791412 and TYR rs1827430..................80 4-19 Significant Hardy-Weinberg equilibrium analysis for TYRP1 ..........................................81 4-20 Pairwise haplotype analysis results for TYRP1 rs683 and TYRP1 rs2733831...................81 4-21 Pairwise haplotype analysis results for TYRP1 rs683 and TYRP1 rs2733833...................81 4-22 Pairwise haplotype analysis results for TYRP1 rs2733831and TYRP1 rs2762464............82 4-23 Pairwise haplotype analysis results for TYRP1 rs2733831 and TYRP1 rs2733833...........82 4-24 Pairwise haplotype analysis results for TYRP1 rs2762464 and TYRP1 rs2733833...........82 4-25 Pairwise haplotype analysis results for TYRP1 rs2762462 and TYRP1 rs2733833...........83 4-26 Pairwise haplotype analysis results for COMT rs2020917 and COMT rs6269.................83 4-27 Summary of all findings for DCT PAH TYR TYRP1, and COMT ..................................84 A-1 All SNPs genotyped...........................................................................................................95 10

PAGE 11

LIST OF FIGURES Figure page 1-1 Melanin biosynthesis pathway...........................................................................................30 4-1 Schematic representation of gene structure, SNP location, and HapMap CEPH haploblocks.................................................................................................................... ....70 11

PAGE 12

Abstract of Dissertation Pres ented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy THE ROLE OF GENETICS IN VITILIGO SUSCEPTIBILITY By Deborah Marsha Herbstman August 2009 Chair: Margaret R. Wallace Co-chair: Wayne T. McCormack Major: Medical Sciences-Genetics Vitiligo is an autoimmune pigment disorder of the skin, which is seen on patients as depigmented areas that may gradually enlarge. Vitiligo is a common co ndition, affecting about 0.5-1% of all ethnic groups worldw ide. It is also associated w ith an increased risk for other autoimmune diseases. The current treatments for vitiligo are difficult, expensive, and often yield marginal results. The cause of vitiligo is unknow n, but is thought to involve both genetic and environmental factors. The goal of my study was to test the hypothesis th at vitiligo pathogenesis is caused in part by genetic susceptibility to both autoimmune and autotoxic events due to genetic differences in genes involved in th e regulation of the immune response, melanin production, and oxidative stress. To identify vitiligo susceptibility genes, hum an genomic DNA samples from patients with vitiligo, their family members, and healthy c ontrols with no known autoimmune diseases were genotyped for a number of different single nucleotid e polymorphisms (SNPs) in candidate genes. SNPs in COMT, TYR TYRP1 DCT and PAH genes involved in melanin biosynthesis, were examined. TYR TYRP1 and DCT are also melanocyte autoantigens, which further implicates them as vitiligo susceptibility genes. The immunoregulatory gene AIRE was tested. Genes 12

PAGE 13

involved in the regulation of oxidativ e stress and DNA repair, specifically, FXBO11 and MSH6 also were examined. Using case-control and family-based genetic association studies, as well as haplotype analysis, several putative susceptibility genes fo r vitiligo were identified. Susceptibility to vitiligo was linked to the AIRE gene. The AIRE protein is responsible for T-cell processing in the thymus. Significant results were also found in the melanin biosynthesis genes COMT, PAH TYR TYRP1 and DCT These results demonstrate a possibl e role for genes involved in immune system regulation, as well as fo r genes involved in melanin synthe sis in vitiligo susceptibility. The aim of my study was to identify genes involved in vitiligo suscep tibility so that in the future, novel therapies that might prev ent or ameliorate vitiligo ma y be developed based on my understanding of the genetic causes of vitiligo. 13

PAGE 14

CHAPTER 1 INTRODUCTION Vitiligo is a common dermatological disorder of the epidermis and hair follicles, affecting both genders equally and ~0.5-1% of the populat ion in all ethnic groups worldwide (Kovacs, 1998; Nordlund and Ortonne, 1998; Spritz, 2007). The average age of onset is 22 16 years (Nordlund and Ortonne, 1998). The disease is defined clinically by expanding areas of hypopigmentation on the skin surface due to the destruction or inactivation of epidermal melanocytes (Badri et al., 1993; Majumder et al ., 1993; Tobin et al., 2000; Spritz, 2007). Vitiligo pathology is limited to the depigmentation of the epidermis, but the illness is often comorbid with other autoimmune disorders such as alopecia areata and Hash imotos thyroiditis (Badri et al., 1993; Kemp et al., 1999a). Loss of pigm entation can occur anywhere on the body. The striking appearance of depigmented areas su rrounded by normally pigmented skin can cause social stigmatism and physiological distress (Millington and Levell, 2007; Sampogna et al., 2008). Because patients are not usually born with the disease and many report coinciding events at or just prior to onset, it ha s been suggested that vitiligo might be triggered by environmental factors (Manolache et al., 2009). Reported triggers include skin wounds, eczema, pregnancy, psychological stressors, certain drugs, and medical procedures (e.g., bone marrow transplant) (Brown-Harrell et al., 1996; Papa david et al., 1996). When vitili go lesions develop at sites of trauma, such as abrasions, surgical scars, radi ography treatments, resolving psoriasis or eczema, contact dermatitis, or severe sunburn, this is known as the Koebner phenomenon (Powell and Dicken, 1983; Levine and Ribeiro, 1994). It has been shown that the numbers of melanocytes in depigmented vitiligo lesions are vastly reduced or absent; however, the mechanisms of destruction have been widely debated (Nordlund and Ortonne, 1998). Treatment for vitiligo is often both expensive and time and labor intensive for the patient, with insurance companies often 14

PAGE 15

denying coverage. The goal of my work was to identify potential vitiligo susceptibility genes with the hope that this knowle dge may lead to better understandi ng of disease pathogenesis and more targeted disease treatment. Melanocytes and Pigmentation Because vitiligo is a skin disease, a brief overview of the structure and physiology of this tissue is included. Human skin consists of two main layers: the epidermis, a stratified squamous epithelium mainly consisting of keratinocyt es, and the dermis, an underlying layer of vascularized connective tissue (Nordlund and Or tonne, 1998). The epidermis is composed of four cell types: keratinocytes, melanocytes and two types of non-pigmented granular dendrocytes, Langerhans cells, and Granstein cells. Both Langerhans cells and Granstein cells are antigen-presenting cells (APCs) that interact with T-ce lls to modulate the immunologic response in the skin. For the purposes of be tter understanding viti ligo, keratinocytes and melanocytes are described in further detail. Keratinocytes develop from a basal layer of ce lls at the epidermal-de rmal junction. These cells are in a constant state of mitosis. As mo re cells form, the mature keratinocytes are forced upward through the epidermis. Keratinocytes are i nvolved in the constant renewal of the skin as millions of the surface cells are sloughed off daily and new keratinocytes migrate toward the skin surface. Melanocytes are highly dendritic cells that contact all keratin ocytes; they reside in the basal layer of the epidermis and produce skin pigment called melanin. The production of melanin is a highly complex process involving ma ny enzymes and cofactors (Figure 1-1) (Casp, 2003). There are two types of melanins, blac k-brown eumelanin and red-yellow pheomelanin, though both have the amino acid tyrosine as their initial substrate. The enzyme tyrosinase is responsible for the first two steps in melanin production and has been id entified as the rate15

PAGE 16

limiting factor. Melanin is produced in me mbrane-bound granules called melanosomes. Melanosomes migrate from the center of the mela nocyte cell body to the end of the dendrites and are deposited into keratinocytes (Nordlund et al ., 1998). The melanosomes accumulate in the keratinocytes and form a shield of melanin, wh ich provides the skin w ith protection against ultraviolet radiation from sunlight. All humans have relatively the same quant ity of melanocytes, so different skin pigmentations are accounted for by variations in melanocyte activity or the rapidity of melanin breakdown in keratinocytes. The melanocytes of dark brownand black-skinned individuals produce more melanin of a darker color than do the melanocytes of fair-skinned individuals. Freckles and moles are localized areas of incr eased melanin production. Melanin production can be stimulated by ultraviolet rays in sunlight, which results in the darkening of the skin; this is commonly referred to as tanning. This increase in melanin augments the skins protection against the ultraviolet radiation damage that can cause sk in cancer. The hypopigmented lesions in vitiligo patients are a result of the destruc tion and/or inactivation of the melanin-producing melanocytes. Keratinocytes still migrate to the surface of the epithelium, albeit without their cargo of pigment. This results in patches of sk in that look milky-white because they are devoid of pigment. Clinical Description of Vitiligo There are two main types of vitiligo, se gmental and non-segmental (also known as generalized vitiligo), and classi fication relies on the distributi on of hypopigmented lesions. In segmental vitiligo, the areas of depigmentation are random and often occur on only one location on the body (Nordlund and Ortonne, 1998). Segmenta l vitiligo sometimes involves the loss of follicular melanocytes and a loss of hair pigment. Segmental vitiligo is less likely to repigment than non-segmental vitiligo; however, it is also le ss likely to spread to other areas of the body. 16

PAGE 17

Vitiligo spreads either through the expansion of existing macules or with the onset of new areas of depigmentation (Nor dlund and Ortonne, 1998). Non-segmental vitiligo can a ppear anywhere on the body, but it most frequently appears symmetrically affecting the hands, mouth, eyes elbows, and/or knees (Lerner and Nordlund, 1978). It is unknown why an instig ating event on one side of the body can trigger vitiligo not only at the site of injury but also at the same site on the oppos ite side of the body. In contrast to segmental vitiligo, non-segmental disease often does not affect the hair fol licles in the areas of depigmentation. Non-segmental vitiligo often gets progressively worse, with lesions enlarging and more areas of the body being affected over time. Some patients w ith non-segmental vitiligo almost completely depigment over the years; this condition is referred to as universal vitiligo. In contrast to segmental disease, non-segmen tal vitiligo sometimes shows spontaneous repigmentation, or partial repigmentation with trea tment; however, the patien t rarely fully returns to pre-disease pigmentation. B ecause segmental and non-segmental vitiligo present differently clinically, they may have different pathophys iologies (Nordlund and Ortonne, 1998; Bos, 2005). Clinicians and researchers sometimes distingui sh between active and stable vitiligo, though these definitions are not universally applie d (Yu et al., 1993; Mulekar, 2003; Chen et al., 2004; Lotti et al., 2008). Genera lly, patients described as having active vitiligo are those whose lesions have spread within the last three months. Patients are said to have stable vitiligo when their lesions have not changed (depigmenting or repigmenting) within the past three-month to two-year period. Vitiligo Pathogenesis Two of the most popular theories for the pa thogenesis of vitili go include: 1) the autoimmune hypothesis and 2) and the autocyto toxic hypothesis. These theories are not 17

PAGE 18

mutually exclusive; the onset of disease may involve a combination of autoimmune as well as autotoxic events. Autoimmune Hypothesis The autoimmune hypothesis of vitiligo is the mo st widely held, wherei n the destruction of melanocytes results from an autoimmune res ponse launched against cellular components of the melanocyte. The apparent association between vi tiligo and several other autoimmune disorders, such as alopecia areata, Addisons disease, Hashim otos thyroiditis, and pernicious anemia, lends support to the autoimmune theory (Sch allreuter et al., 1994a; Yu et al., 1997) Humoral autoimmunity Studies of serum autoantibodies (autoAb) dire cted against melanocyt e antigens suggest a possible humoral component to the autoimmune response in vitiligo. For example, Bystryns group has reported that the incidenc e and level of anti-pigment cel l autoAb is correlated with vitiligo disease activity (Harning et al., 1991). There are several reports that the autoAb in vitiligo patient sera can be cytolytic to melanocytes by complement-mediated and antibodydependent cell-mediated cytotoxicity (ADCC) mechanisms (Harning et al., 1991; Cui et al., 1993; Gilhar et al., 1995). In one study, antibodies to melanocytes were present in 83% of vitiligo patients and 7% of controls (Cui and Bystryn, 1995). Several studi es have observed anti-tyrosinase autoAb in vitiligo patients (Baharav et al., 1996; Kemp et al., 1997a), although another report disputes the identification of the prin cipal autoantigen as tyrosinase (X ie et al., 1999). AutoAb against several melanocyte proteins including tyrosi nase-related protein 1 (TYRP1), dopachrome tautamerase (DCT) (formerly known as tyrosinase -related protein 2), and melanin-concentrating hormone receptor 1 (MCHR1) have been identif ied by several investig ators (Naughton et al., 1983; Cui et al., 1993; Song et al., 1994; Kemp et al., 1997a; 1997b; 1998; Okamoto et al., 1998; 18

PAGE 19

Kemp et al., 1999b; 2002). It has also been repo rted that antibodies ag ainst keratinocytes and degenerated keratinocytes are presen t in perilesional skin in vitiligo patients (Norris et al., 1996). Taken together, these observations may support a possible role for anti-melanocyte autoAb in the induction and/or progression of vitiligo. Although anti-melanocyte antibod ies have been discovered in the serum of vitiligo patients, some argue that the ex istence of autoantibodies may simp ly be a marker of the active disease process, as in diabetes (Mehers and Gillespie, 2008). While early evidence suggested that antibodies were directed toward antigens on the melanocyt e surface, many autoantibodies are against intracellular compone nts of the melanocyte in vi tiligo patients (Passeron and Ortonne, 2005). In vitro work with these autoantibodies has suggested that the presence of the anti-melanocytic autoantibodies may just be a mark er of active vitiligo a nd may have little or no active role in initiating or maintaining the di sease (Bos, 2005). This finding has lead to the questioning of the pathogenic role of autoantibodies. Cellular autoimmunity There is strong evidence supporting a possible role of cell-mediated immunity in the pathogenesis of vitiligo. The firs t evidence of cell-mediated immun ity playing a role came from the observation of invading lymphocytes in studies of inflammatory vitiligo, a disease characterized by a raised red ri m surrounding the depigmented lesi on (van den Wijngaard et al., 2001). Lesional skin of vitiligo pa tients has been shown to contai n significantly higher levels of T cells than found in skin of h ealthy controls and in skin outs ide lesions in the same patients (Badri et al., 1993; Le Poole et al., 1996). Some immunohistoc hemical studies suggest that melanocytes are almost completely destroyed, as opposed to being inactive or dormant, and this loss is accompanied by dermal and epidermal inf iltrates in the active lesion, which include an increase in CD3+, CD4+, and CD8+ T cells (Hann et al., 1992; Badr i et al., 1993; Le Poole et al., 19

PAGE 20

1993b; Le Poole et al., 1996). IL-2 and IFNare also expressed in cellular infiltrates, suggesting that activation of T h1 cells may have occurred (Abde l-Naser et al., 1994). Other work suggests that not all melanocytes are dest royed in the lesion; however, those remaining have been rendered dysfunctional (Tobin et al., 2000; Gottschalk and Kidson, 2007). Melanoblasts remain in hair follicles, however as any repigmentation that does occur either spontaneously or in response to treatments, begins in small spots around hair follicles, spreading until repigmented areas coalesce (Arrunategui et al., 1994). Further evidence of an abnormal i mmune response is a lowered CD4+/CD8+ ratio in perilesional skin in vitiligo patients compared to controls (Mozzanica et al., 1990; Le Poole et al., 1996). The increased numbers of CD8+ cells observed in vitiligo lesions has led LePoole and colleagues to hypothesize that cytotoxic CD8+ cells are responsible fo r melanocyte destruction (1996). A decrease in CD45RA T cells, which are nave T cells, has been observed in vitiligo patient peripheral blood (Abdel-Naser et al., 1992). Disruptions in Langerhans cells in vitiligo have also been reported in skin biopsies, with decreased numbers seen in active vitiligo and a return to normal numbers in stab le vitiligo (Kao and Yu, 1990). An unusual feature of vitiligo is the abnormal expression of class II human leukoc yte antigen (HLA) mol ecules by perilesional melanocytes in about two-thirds of vitiligo pati ents, accompanied by a six-fold increase in the expression of intercellular adhesion molecule-1 (ICAM-1) (al Badri et al., 1993). Peripheral blood of active vitiligo patients has been shown to have increased concentrations of proinflammatory cytokines IL-6 and IL-8 as well as a decreased pr oduction of GM-CSF, TNF, and IFN(Yu et al., 1997). It is important to note th at variations in publish ed data on peripheral cell imbalances may be due to a variety of f actors such as differences in study populations, 20

PAGE 21

differences in vitiligo disease presentation an d/or progression, and prior immune-suppressive therapy (Ongenae et al., 2003). Studies have implicated the AIRE gene, which causes autosomal recessive autoimmune polyendocrinopathy-candidiasis-ectodermal dystr ophy syndrome (APECED), in the autoimmune pathogenesis of vitiligo (1997; Nagamine et al., 1997; Su and Anderson, 2004; Spritz, 2006). AIRE produces a DNA-binding protein that regula tes negative selection of autoreactive thymocytes (Anderson et al., 2002) and may play a role in the cellular autoimmune pathogenesis of vitiligo. Thus, this was one candida te gene I chose to investigate. Vitiligo and melanoma Finally, additional support for the autoimmune disease hypothesis comes from the examination of melanoma, a malignancy of the me lanocyte. The immune system can sometimes recognize tumor cells and launch an immune response against th ese cells. In the case of melanoma, immortalized melanocytes are the targ et cells. Many melanoma patients, especially those being successfully trea ted, develop depigmentation re sembling vitiligo (although the distribution of affected regions is different) because healthy and diseased melanocytes share many of the same surface antigens (Naftzger et al., 1996; Overwijk et al., 1999). This finding suggests that the immune system can lose its abili ty to make the distinction between normal and diseased melanocytes (Ram and Shoenfeld, 2007; Le Poole and Luiten, 2008). It has been suggested that there is an improved prognosis for melanoma patient s who develop vitiligo (Overwijk et al., 1999; Wankowicz-Ka linska et al., 2003). Interestingl y, vitiligo patient sera are reported to have autoAb recogni zing melanoma antigens and are ab le to inhibit melanoma cell proliferation in vitro (Fishman et al., 1993) and cause re gression of melanoma metastases in vivo (Merimsky et al., 1994). This observation has led some investigators to study the abnormal immune response in vitiligo as a basis fo r formulating new treatments for melanoma. 21

PAGE 22

Autocytotoxic Hypothesis The autocytotoxic hypothesis is based on th e idea that the damage and/or loss of melanocytes in vitiligo patients are a result of an inherent defect in the melanocyte, resulting in the buildup of toxic intermediate s or metabolites from the me lanin biosynthesis pathway. Melanogenesis produces large amounts of reactive oxygen species (ROS) like hydrogen peroxide, so melanocytes are at risk of oxidative damage to proteins and DNA unless their endogenous antioxidant systems ar e functional (Norris et al., 1996; Bystryn, 1997; Nordlund and Ortonne, 1998; Passeron and Ortonne, 2005). A diverse body of eviden ce, including several cellular abnormalities and biochemical defects that have been reported in vitiligo patients, support this hypothesis (Schallreuter et al., 1996; Maresca et al., 1997; Passi et al., 1998; Schallreuter, 1999). Upon cell death, melanocyte antigens are released which may elicit an immune response. Many abnormalities have been reported in the skin of vitiligo patients that implicate reactive oxygen species in disease pathology. Because the melanin biosynthetic pathway involves many enzymes and co-factors that regu late the creation and destruction of ROS, alterations in this pathway (eit her increasing or decreasing enzy matic activity) may lead to the buildup of toxic intermediates. Decreased levels of the antioxidant enzyme thioredoxin reductase has been observed in vitiligo patient sk in (Schallreuter and Wood, 2001). There is also increased sensitivity to oxidative stress in melanoc ytes cultured from perilesional skin of vitiligo patients (Jimbow et al., 2001). These cultured v itiligo melanocytes had reduced (but not absent) levels of tyrosinaserelated protein 1 ( TYRP1 ) mRNA and their intracellular processing of TYRP1 seemed abnormal. Using quantitative real-time PCR, Kingo and colleagues found several aberrant gene expression levels in vitiligo patients. Notably, they observed that TYRP1 and dopachrome tautamerase ( DCT ) genes were downregulated in lesional skin compared to 22

PAGE 23

non-lesional vitiligo skin or skin of healthy controls, and upregul ated in uninvolved vitiligo skin compared to healthy control samples (Kingo et al., 2007). This may be a result of the lower number of melanocytes present in lesional vitili go skin. In studies using cell culture and the chemotoxin 4tert-butylphenol (4-TBP), a competitive inhibitor of tyrosinase, Manga et al. observed increased expression of TYRP1 in vitiligo compared to controls (Manga et al., 2006). This increased expression of TYRP1 significan tly increased cytotoxic sensitivity of vitiligo melanocytes to 4-TBP. Another group of invest igators found that cultu red keratinocytes from vitiligo patients promote reduced TYRP1 protein expression in co-cul tured melanocytes (Phillips et al., 2001). One study found that a single nucleotide polymorphism or SNP (pronounced snip), which causes altered promoter function in the gene FOXD3, was genetically associated with vitiligo (Alkhateeb et al., 2005); this gene is a transcriptional regulator of TYRP1 (Thomas et al., 2008). The sum of these findings suppor t the hypothesis that alte rations in the normal functioning and/or expression patterns of tyrosinase-rela ted protein 1 and dopachrome tautamerase in melanocytes may contribute to oxidative stress obser ved in vitiligo. Levels of the antioxidant enzymes superoxide dismutase, glutathione peroxidase and malondialdehyde were significantly increased in tissue of patients with non-segmental vitiligo compared to controls (Yildirim et al., 2003), presum ably due to an increase of ROS in the skin of vitiligo patients. Also, significant association to vitiligo was observed in polymorphisms of glutathione S-transferase ( GST) genes, which are important in protection against oxidative stress (Uhm et al., 2007). Compared to healthy indi viduals, there is an increase in oxidative DNA damage in the mononuclear component of peri pheral blood leukocytes from vitiligo patients (Giovannelli et al., 2004). The catechol-O-methyl transferase or COMT gene, which is involved in catecholamine biosynthesis as well as melanin biosynthesis and oxi dative stress regulation, 23

PAGE 24

has been shown to be differentially expr essed in vitiligo patients. In one study, COMT was found to be expressed at higher levels in ep idermal homogenates from vitiligo patients than homogenates from healthy controls (Le Pool e et al., 1994). Howeve r, a second study found COMT activity levels to be lower in patients wi th acrofacial vitiligo (Tursen et al., 2002). Vitiligo patients show an increased epidermal de novo synthesis and recycling of 6(R)-Lerythro-5,6,7,8-tetrahydrobiopterin (6BH4). Vitiligo patients also show very low levels of 4aOH-tetrahydrobiopterin dehydratase (DH) activity, which is involved in the recycling pathway of 6BH4. Due to decreased DH activity, there is an increased buildup of the 7-isomer form of 6BH4, (7BH4), which functions as an effective competitor of 6BH4 (Davis et al., 1992; Schallreuter et al., 2001). It is the accumulation of these oxidi zed pterins (6and 7-biopterin) that results in the fluorescence of vitiligo skin under a Woods UV (351 nm) lamp, which is used for the clinical diagnosis of v itiligo in a patient (Schallreuter et al., 2001). A buildup of 7BH4 can alter phenylalanine hydroxylase (PAH) activity. PAH activity is often decreased in vitiligo patients, which causes a buildup of epidermal Lphenylalanine levels (Schallreuter et al., 1994b; Schallreuter et al., 1994c). Abnormal calci um homeostasis was observed in vitiligo keratinocytes and melanocytes cons istent with oxidative stress (S challreuter et al., 2007). This calcium disregulation likely also contributes to the disregulation of L-phenylalanine observed in vitiligo patients, as L-phenylalan ine is coupled to the release/uptake of calc ium into the cytosol of melanocytes to initiate melanogenesis (Schallreuter and Wood, 1999). All of these abnormal biochemical events in the vitiligo epidermi s may contribute to increased levels of hydrogen peroxide (H2O2). High concentrations of H2O2 can deactivate catalase, an enzyme normally involved in breaking down H2O2 and other free oxygen radicals. Low catalase activities have previously been re ported in the epidermis of vitiligo patients 24

PAGE 25

(Schallreuter et al., 1991), though whether this defect is due to an increase in H2O2 from the defective 6BH4 pathway or due to a separate proble m in the catalase enzyme is unknown. Additional studies did not demonstrate a reduction of CAT mRNA in vitiligo patients. What has been shown is that treatment with pseudocatalase, a bis-manganese III-EDTA(HCO3)2 synthetic catalase substitute, has promoted repigmentation in some vitiligo patients, along with restoring DH enzyme activity and a return to normal 7BH4 levels in the epidermis (Schallreuter et al., 2001). Additionally, polymorphisms in the CAT gene have been shown to be associated to vitiligo susceptib ility in several studies (C asp et al., 2002; Park et al., 2006; Em et al., 2007). Another possible contributing fact or to onset of vitiligo may be an inherent cytological defect reported for melanocytes of vitiligo pati ents. Vitiligo melanocytes have been shown to have ultrastructural defects including an abnormal dilation of the rough endoplasmic reticulum (RER), circular RER profiles, and memb rane-bound melanosome compartments seen on transmission electron microscope examination (Boissy et al., 1991; Im et al., 1994). Further observation of this phenomenon is s een in an immortalized vitiligo cell line where this abnormal dilation is thought to be caused by a retention of a variety of protei ns in the RER (Le Poole et al., 2000). Although not cytotoxic to melanocytes in vitro such defects might contribute to melanocyte destruction in vivo Morphological changes in melanocyte RER, as seen by electron microscopy, are suggestive of retention of peptid es in the RER (Boissy et al., 1991; Im et al., 1994). Alterations affecting the functional integrity in the protein tyrosinase, the rate-limiting enzyme in melanin synthesis, have been hypot hesized to play an important part in the autocytotoxic model of vitiligo. If tyrosinase is damaged or misfolded, it is retained in the endoplasmic reticulum. These malformed proteins are then ubiquitin-tagged and targeted for 25

PAGE 26

degradation via the proteasome (H alaban et al., 2002). If oxidative stress causes necrosis of these cells retaining misfolded tyrosina se, a large amount of the prot ein may be released into the inflammatory microenvironment potentially to be picked up by antigen-presenting cells (APCs). It is also possible that ROS are modifying these retained proteins, allowing the presentation of cryptic epitopes by either APCs or by the melanocytes themselves, which have been shown to phagocytize and present antigen in times of stress through MHC class II (Le Poole et al., 1993a; 1993b). Recruitment and involvement of other ce lls of the immune system can function to increase oxidative stress. Secretion of chemokines and cytokines into the local environment can increase inflammation, and potentially target cells for further destru ction by necrosis and apoptosis. Dead and dying cells may accumulate in the epidermis, draw ing the attention of circulating monocytes. This may further permit an increased pres entation of self-peptides to cells of the immune system. Thus, the auto cytotoxicity of melanocytes may lead to autoimmunity. My work has focused on severa l genes in this pathway including catechol-Omethyl transferase (COMT ), tyrosinase ( TYR ), tyrosinaserelated protein 1 ( TYRP1), and dopachrome tautamerase ( DCT ). Additionally, a number of studies have f ound altered expression for DNA repair and antioxidant genes. The MutS E. coli Homolog 6 or MSH6 [formerly known as G/T mismatchbinding protein ( GTBP)] gene is involved in G/T DNA repa ir, and its upregulation in vitiligo patients may be indicative of increased DNA damage (Le Poole et al., 2001) Although the exact function of the F-Box Only Protein 11 (FBXO11 ) gene [formerly known as vitiligo-associated protein-1 ( VIT1 )] is unknown, it does share homology to the MSH6 gene, and it may be involved in the oxidative stress pathway eith er indirectly, or by regulation of MSH6 Interestingly, 26

PAGE 27

whereas MSH6 expression is increased in vitiligo pa tients compared to healthy controls, FBXO11 expression is reduced (Le Poole et al., 2001). Treatment There is no cure for vitiligo; treatment me thods function by stopping current progression of the disease and aiding in the repigmentation of hypopigmented lesions. The treatment modalities for vitiligo are difficult, expensive, and often disappointing, as they do not prevent the reappearance of new lesions. Many treatments ar e not covered by health insurance. The early options for vitiligo treatment were limited ma inly to psoralens with UVA radiation (315-400 nm), also known as PUVA thera py, and the use of topical corticos teroids. With this approach, repigmentation of the lesions is inconsistent and slow, and the lesions may worsen during therapy (Njoo et al., 1998; Bethea et al., 1999; Buckley and du Vivier, 1999). Current treatment options have expanded and now include a variety of choices such as epidermal melanocyte autografts, UV-B therapy (between 280 and 315 nm), oral and topical phenylalanine, and the 308-nm excimer laser (Falabella, 1997; Camacho and Mazuecos, 1999; Njoo and Westerhof, 2001; Nicolaidou et al., 2009). Another treatment for vitiligo is the topical application of UVB-activated pseudocatalase. Pseudocatalase functions to mimic the human enzyme catalase, which removes H2O2 and other free radicals from the epidermis. In one study, treatment of vitili go patients with topical applications of the pseudocatalase cream, along with frequent exposure to UVB radiation at a lower than normally therapeutic doses, allowed for the halt of depigmentation in 95% of patients, and repigmentation of 60% of individuals (Schallreuter, 1999; Schallreuter et al., 1999). However, a group from Australia was unable to replicate the success of pseudocatalase creams seen by Schallreuter and colleagues, though they did confirm the therapeutic be nefit of higher UVB exposure and may have used a different pseudocatalase formulation (Bakis-Petsoglou et al., 2009). Among the more unique modalities, 27

PAGE 28

Dead Sea climatotherapy has proven to be an eff ective vitiligo treatment. Although the specific therapeutic mechanisms of this unique environment have yet to be elucidat ed, they are suggested to be attributed to pseudo catalase activity of Dead Sea sa lts and natural UVB exposure via sunlight (Schallreuter et al., 2002). Genetics of Vitiligo The importance of genetic factors in susceptibility to vitiligo is s uggested by reports of familial aggregation from several laboratories (Bhatia et al., 1992; Nordlund, 1997; Kim et al., 1998; Passeron and Ortonne, 2005; Fain et al., 2006; Hu et al., 2006; Spritz, 2006). About 20% of vitiligo patients have at least one affected firs t-degree relative, and the relative risk of vitiligo for first-degree relatives of vitiligo patients is incr eased by at least sevento ten-fold (Bhatia et al., 1992; Nath et al., 1994). Multiple associ ations with human major histocompatibility complex (MHC) genes have been claimed (Duns ton and Halder, 1990; Venneker et al., 1992; Buc et al., 1998; Casp et al., 2003) supporting the possible involvement of immunologically relevant genes in the human leukocyte an tigen (HLA) region. However, no common HLA association has been observed. Studies have repo rted HLA associations in early onset vs. late onset vitiligo cases. It is possible that there are different etiologies and different genetic factors in early vs. late onset vitiligo, as well as in se gmental vs. non-segmental vitiligo (Finco et al., 1991; Orecchia et al., 1992; Ar cos-Burgos et al., 2002). Other genes implicated in susceptibility to vitiligo include ACE (Jin et al., 2004a; Akhtar et al., 2005), ESR1 (Jin et al., 2004b), CAT (Casp et al., 2002; Park et al., 2006; Em et al., 2007), and NALP1 (Jin et al., 2007b) to name a few. Based on all of these observations, vitiligo is hypothesized to be a polygenic disease, with variant alleles in several unlinke d loci possibly contribu ting to increased susceptibility to and/or direct pathogenesis of vitiligo. 28

PAGE 29

Vitiligo susceptibility rarely follows a simple Mendelian i nheritance pattern, though it has been observed (Alkhateeb et al., 2005; Birl ea et al., 2008). This non-Mendelian type of inheritance is common among othe r autoimmune diseases (Majumder et al., 1993; Nath et al., 1994; Alkhateeb et al., 2002). For example, in t ype 1 diabetes, the risk of an individual developing diabetes with a first degree relative affected is only between 1-9% (Redondo et al., 2001). As with many late onset diseases with an autoimmune component, genetics alone do not dictate disease onset. The role of environmenta l factors is evident in vitiligo, and the complex interactions between genetics and the environm ent is critically important to understanding disease pathogenesis (Nancy and Yehuda, 2009). However, the current research methodologies available make the gene-environment interaction one of the most difficult areas to elucidate. Based on what we currently understand about the genetic and environmental contributions to vitiligo pathogenesis, it may be considered a complex disease. My studies have been aimed at testing the candidate genes discus sed in this chapter for genetic association, to ascertain whether there is evidence that variants at these genes contribute to viti ligo susceptibility or account for some of the clinical variation seen in patients. 29

PAGE 30

DHI-2C DHI Figure 1-1. Melanin biosynthe sis pathway (Casp, 2003). 30

PAGE 31

CHAPTER 2 EXPERIMENTAL DESIGN: CASE-CONTROL AND FAMILY-B ASED ASSOCIATION STUDIES Introduction Genetic association studies examine whether a genetic variant (like a SNP) is associated with a disease or trait on a population scale (Cardon a nd Bell, 2001). This approach has already been successfully used to further our understandi ng of a number of diseases, such as diabetes (Morris et al., 2006; Smyth et al., 2006), Park insons disease (Mizuta et al., 2006), and rheumatoid arthritis (Hughes et al., 2006; Ikari et al., 2006; Kang et al., 2006) to name a few. Due to budgetary limitations in my study, I had to restrict the number of candidate genes examined to eight and the number of SNPs genotyped to thirty. These were chosen with the following criteria in mind: evidence of the gene be ing involved (or likely to be involved) in vitiligo pathogenesis, allele frequencies in my sample population, haploblock data, and the likelihood that the SNP may have a func tional effect (Drago et al., 2007). As previously described, two principal hypot heses concerning the etiology of vitiligo include (1) the autoimmune model, which s uggests that melanocyte death occurs through inappropriate immune system destruction of pigm ent cells, and (2) the self-destruct model, which suggests that biochemical and/or st ructural defects inherent to pa tient melanocytes contribute to the initiation and/or progression of melanocyte cytolysis. To eval uate genetic factors that may play a role in vitiligo pathogenesis, a candidate gene approach was design ed. Genes involved in immune system regulation, melani n biosynthesis, and DNA repair a nd oxidative stress regulation were chosen based on their pote ntial involvement in vitiligo pa thogenesis. Case-control and family-based analyses were performed on genotypi c data garnered from vitiligo patients, their family members, and non-affected controls. Pa irwise haplotype analysis was also performed. 31

PAGE 32

Negative association results are discussed in this chapter, with positive results in chapters 3 and 4. Vitiligo Candidate Gene Selection Pubmed ( http://www.ncbi.nlm.nih.gov/) was used to search current literature to identify putative vitiligo genes including those linked to autoimmunit y, melanin biosynthesis, and oxidative stress. The candidate gene approach al lows researchers to focus their investigation by examining a limited number of genes that are sele cted based on the specific pathophysiology of a disease. Candidate gene studies do not need large multi-generation families, but can be performed using unrelated groups of patients (cas es) and unaffected controls, or through small families. Candidate gene studies also have the advantage of being better suited for determining susceptibility genes in more complex disorders wh ere the relative risk as sociated with any one gene is relatively small (Kwon and Goate, 2000). Because it has been hypothesized that vitiligo pathogenesis could be a direct re sult of autoimmunity, and/or au totoxicity due to biochemical defects in the melanocytes, gene s that regulate the immune sy stem, melanocyte biochemistry, and oxidative stress were deemed suitable choices for candidate genes. Autoimmune Regulator ( AIRE) Disruptive mutations in the AIRE gene are responsible for autoimmune polyendocrinopathy-candidiasis-ectodermal dystrophy syndrome (APECED). The most common inheritance pattern for APECED is auto somal recessive (Rizzi et al., 2006), though an autosomal dominant inheritance pattern has also been reported (Cetani et al., 2001). The most common features of APECED are parathyroid gl and failure, chronic susc eptibility to candida yeast infection, and Addison's disease. Other autoimmune diseases associated with APECED may include alopecia, vitiligo, ovarian failure, te sticular atrophy, hypothyroidism, gastric parietal cell atrophy, hepatitis, intestinal malabsorpti on, and insulin-dependent diabetes mellitus. 32

PAGE 33

Because vitiligo is one of the autoimmune diseas es associated with APECED, it was chosen as a candidate gene. Further discussion of this gene and the results from the present study are found in Chapter 3. Melanin Biosynthesis Genes Five genes critical to the melanin biosynthe sis pathway were selected for this study: tyrosinase ( TYR) tyrosinase-related protein 1 ( TYRP1) dopachrome-tautomerase ( DCT), phenylalanine hydroxylase, ( PAH), and catechol-O-methyltransferase ( COMT) These genes, several of whose proteins have been observed to be melanocyte autoan tigens, may be vitiligo susceptibility genes. Further discussion of this gene set and the results from the present study are found in Chapter 4. DNA Repair Genes Previous studies of vitiligo patients melanocytes have shown subtle changes in the expression levels of genes involved in DNA repa ir. Problems with DNA repair mechanisms may lead to cell damage resulting in immune cell targeting of melanoc ytes. Genes examined in the present study that are linked to DNA repair include F-box only protein 11 or FBXO11 and the DNA G/T repair protein MutS, E. coli, homolog 6 gene or MSH6. Further discussion of these genes and the results from the present study are found later in this chapter. Materials and Methods Subjects Previous work in the McCormack lab, which formed the foundation of the current project, involved establishing an internet website for both the University of Florida and at the National Vitiligo Foundation to recruit subjects for my st udy (Casp et al., 2002; Casp et al., 2003). Kits were mailed to interested patients and family members containing three 10 mL EDTA blood collection tubes and IRB consent form s (project number s 130-1998 and 416-1998). 33

PAGE 34

Additionally, patients and family members were as ked to fill out personal information surveys requesting information such as age, sex, r ace, and ethnic background, age of onset, family history, presence of other autoi mmune diseases, and pattern of depigmentation (Table 2-1). A total of 714 individuals return ed blood, data, and consent for this and future studies. Blood Processing and DNA Extraction DNA was extracted from blood samples (leukocy tes) using standard phenol/chloroform extraction methods after red blood cell lysis (Cas p et al., 2002). Whole genome amplification was performed on some samples later as need ed, using a commercially available kit per manufacturers instru ctions (GE Healthcare). HIPAA de-identified DNAs from unrelated controls with no known autoimmune diseases (n= 333) were supplied by the DNA bank of the University of Floridas General C linical Research Center (GCRC). Single Nucleotide Polymorphism (SNP) Selection Once a candidate gene was chosen based on its potential involvement in vitiligo, genetic markers within each gene were selected. The present study used single nucleotide polymorphisms (SNPs). All of the SNPs used were two-allele polymorphisms. It is important to note that these SNPs might encode a functional al teration in the protein, or more likely, they may have no obvious functional relevance to protein function or stability. In candidate gene association analyses, SNPs function solely as genetic markers within the gene. If a polymorphism is inherited with a disorder at a rate that exceeds random chance, this may indicate a potential susceptibility locus. If a polymorphism is inherited by patients at a rate below random chance, this may indicate a potentia l disease resistance (pro tective) locus. In either case, the polymorphism genotyped may be the actual change causing the disease susceptibility or resistance, or the SNP could be linked to the mutation(s) of interest (also known as being in linkage disequilibrium or LD). 34

PAGE 35

Similar methods were employed for SNP sel ection for all of the genes tested. The following software and search engine tools were used to identify SNPs: NCBIs SNP website ( http://www.ncbi.nlm.nih.gov/sites/entrez?db=snp), the SNPper website (http://snpper.chip.org/ ) (Riva and Kohane, 2002), the HapMap website ( http://www.hapmap.org/ ) (2003), and SNPbrowser Version 3.5 software (Applied Biosys tems) (De La Vega et al., 2006). A number of factors were considered in choosing which SNPs to genotype, a nd these are disc ussed in their order of selection priority. SNPs that had been reported as validated, either in the NCBI database or by the HapMap project (2003), were given high selection priority. A SNP was considered validated when it was seen to occur in multiple i ndividuals, and the highest selection priority for my study was given to SNPs that had been validated by several independent investigators. Special consideration was given to SNPs that had been validated by Applied Biosystems using their TaqMan SNP genotyping assays because all of the genotyping was conducted using TaqMan due to cost consideratio ns for the scale of my study. Because the majority of the vitiligo patient samples used in my study was obtained from American whites, SNPs with good minor allele frequencies in this group were chosen. The higher the minor allele frequency, th e more likely it is to observe sufficient numbers of that allele to conduct meaningful analysis. Only SN Ps with a minor allele frequency of 10% in whites were used, to give us the greatest power to de tect association based on my sample population. An attempt was made to select at leas t two SNPs for each gene, one near the 5 end and one near the 3 end, to obtain better coverage of the gene, as SNPs cluste r in haplotype blocks. These blocks are population-specific clus ters where little recombination within the block is believed to occur. Therefore, genotyping one tagging SN P on a haplotype block is hypothesized to allow information to be inferred about the genotypes fo r all of the SNPs on the same block. Haplotype 35

PAGE 36

block information was obtained from the International HapMap Project (www.HapMap.org ) and was viewed using SNPbrowser Version 3.5 (Applie d Biosystems) (De La Vega et al., 2006). In order of priority, SNPs that caused an ami no acid change were selected first, followed by SNPs located in exons, those located at 5 or 3 UTR, SNPs located in introns, and lastly SNPs located very close to a gene on the same haplotype block. For a few genes, more than two SNPs were chosen. SNPs are listed in the chapte r that discusses their association results. Genotyping SNP genotyping was performed at the University of Florida Center for Pharmacogenomics using the Applied Biosystems 7900 HT SNP genot yping platform with TaqMan assays (Table 2-2). Five L reactions in 384-well plates were prepar ed, and the assays were performed and analyzed according to the ma nufacturers recommendations. Statistical Analysis A number of statistical analyses were pe rformed on the sample data, and results are reported in this and subsequent chapters These methods are described below. Hardy-Weinberg Analysis Hardy-Weinberg proportions were tested in cases and controls for all SNPs genotyped. Results are reported only for those SNPs wh ere deviations from HWE were observed. Deviations from Hardy-Weinberg equilibrium (HWE) have been observed in many genetic association studies (Salanti et al., 2005), though the interpretation of this finding remains controversial (Zou and Donner, 2006; Teo et al., 2007; Li and Li, 2008). While some believe that deviations from HWE, also known as Ha rdy-Weinberg disequilibrium (HWD), may indicate genotyping errors (Gomes et al., 1999; Weiss et al ., 2001; Teo et al., 2007), others refute the idea of using HWE testing for this purpose (Zou and Donner, 2006). The assumptions underlying HWE, including random mating, lack of select ion according to genotype, and absence of 36

PAGE 37

mutation or migration, are rarely met (Shoemaker et al., 1998), which has led some to question the relevance of testing HWE in disease asso ciation studies (Zou and Donner, 2006). One group has even suggested that the design of case-cont rol disease association studies will cause HWD because the populations sampled are not random but rather selected because of their disease status, causing alleles in the st udy populations to deviate from their frequency in the general population (Li and Li, 2008). Case-Control Analysis The case-control approach was c hosen due to the very large numbers of vitiligo patients and families required to identify disease susceptibility genes by other methods (e.g., linkage analysis and affected sibpair analysis) (Risch and Merikang as, 1996). Case-control studies compare the frequencies of a genotype at a locus in a patient group versus a control group. The two groups must be matched for ethnicity and ge ographical origins due to variations in gene frequencies among different populations. Althou gh DNA samples were collected from vitiligo patients and their family members from all ra cial and ethnic groups, only the group of whites (212 independent cases and 245 unr elated controls), which are almost exclusively from the United States, was large enough to run a casecontrol analysis on the genotypic data. Allelic and genotype frequencies were calculated and analyzed for each genetic marker for the patient and control sets. In theory, if particular alleles or genotypes have no association with vitiligo, the frequency of that a llele or genotype in the patient population should be very similar to the frequency in the control population. An allele or genotype that confers vitiligo susceptibility or resistance (or on e that is closely linked to such an effect) would hypothetically differ in frequency between the affected and non-affected populations. To compare the frequency of alleles and genotypes between th e case and control populations, chi-squared ( 2) 37

PAGE 38

analyses were performed using one degree of fr eedom for the allelic frequency test and two degrees of freedom for the genotype frequency test. Chi-squared (2) analysis is represented by the following equation: 2= [(Expected Value Obtained Value)2]/Expected Value The Microsoft Excel 2003 software package was used for 2 analysis for allele frequency values for the white case-control analysis. The SAS software package was used for 2 analysis for genotype frequency values for the white case-control groups. The strength of allelic associa tion can also be measured by cal culating an alleles relative risk (RR): RR = [a(b+d)/b(a+c)], where a and b are the numbers of patients and controls with a given allele respectively, and c and d are the corresponding numbers without that allele. The greater the relative risk, the more frequently the allele is found among vitil igo patients and the less frequently it is found among controls. A predisposing effect is indicated by a RR that is significantly higher than 1; a protective effect is indicated by a RR significantly smaller than 1. A neutral effect is indicated by a RR that is not significantly different fr om 1. Significance for RR is given by a 95% confidence interval (CI) that does not include 1. Family-Based Testing To further validate statistical significance determined by the case-control analysis, the transmission disequilibrium te st (TDT), a family-based association test, was performed (Spielman et al., 1993; McGinnis et al., 1995; Spielman and Ewens, 1996). The TDT examines whether alleles of the marker of interest are transmitted at a frequency greater than 50% from an informative parent to an affected child. In the TDT, an informative parent is defined as a parent 38

PAGE 39

who is heterozygous for the SNP of interest. Chisquare analysis is then performed to determine the statistical significance of the TDT, also known as the 2 td statistic. The equation for the TDT analysis of SNP data is: TDT = (b-c)2 / (b+c) where b is the number of times allele 1 of the SNP is transmitted from a heterozygous parent to an affected child and c is the number of times allele 2 is transmitted from a heterozygous parent to an affected child. The Microsoft Excel 2003 software package was used to conduct 2 td tests. My study included 145 families with at least one affected member and one informative parent. Because the TDT looks at direct tran smission of alleles from parents to children [examining identical by descent or (IBD) sharing], it is not affected by po pulation differences in allelic frequencies. It is for this reason that families of all races were included in my TDT analyses, however most of the TDT power in my study is from white families, as only 31 families (21%) had non-white members. Genotype by Co-Morbid Autoimmune Diseases Vitiligo patients with and without a history of other autoimmune diseases were analyzed separately and compared, to dete ct associations that might be related to other autoimmunity (Kemp et al., 1999a; Spritz, 2006). Analysis wa s performed on 212 white patients in SAS using a 3 2 table to look for significant ge notype associations related to individuals with co-morbid autoimmune diseases. Within my white pa tient population, 62 had co-morbid autoimmune diseases and 150 reported none. Genotype by Sex A 2 analysis using the SAS software package was performed to examine whether there was a statistical significance between male and female vitiligo patient genotypes. Another 2 39

PAGE 40

analysis using the SAS software package was performed to examine whether there was a statistical significance between male and female genotypes in the vitiligo patients compared to the healthy controls. Genotype by Age of Onset The analysis of variance (ANOVA) met hod using the SAS software package was performed, to examine whether there was a stat istical significance between vitiligo patient genotypes and age at disease onset. Pairwise Haplotype Analysis Lewontins D statistic for linkage disequilibrium was calculated for each pair of polymorphisms for all whites genotyped; cases and c ontrols were pooled for this analysis. The values of Lewontins D can range from to 1; I report the absolute value of D with zero indicating complete independence between two ma rkers and one representing complete linkage disequilibrium, where markers segregate togeth er 100% of the time (Lewontin, 1964) (Table 23). Linkage disequilibrium (LD) a nd haplotype frequencies were calculated using a maximumlikelihood model incorporating an iterative expectation-maximizat ion (EM) algorithm previously described by Liu et al. (2004). The 2 statistic was used to test the association between the haplotypes and the case-cont rol groups. Four 3 2 tests were calculated for each pair of SNPs in a gene to ascertain the presence and number ( 2, 1, or 0) of risk reference haplotypes in cases and controls (Liu et al., 2004). The risk reference haplotype can be determined by selecting the highest log-likelihood value of the 2 tests. The risk associated with haplotypes was calculated as the odds ratio with 95% confidence interval s. Analysis was conduc ted using only two SNPs at a time (pairwise haplotypes) due to th e limited number of samples in my study. 40

PAGE 41

Corrections for Multiple Testing The Bonferroni correction string ent is a statistical method of adjusting and correcting for multiple numbers of chi-square tests. Bonferronis correction ( ) is represented by dividing the test-wise significance level by the number of test s, and is represented by the following equation: = / k where = the testwise significance le vel (0.05) and k = the total number of tests performed (Rice et al., 2008). A Bonferroni co rrected p-value was calculated for all findings using the SAS software package. More recently, the false discove ry rate (FDR) has been proposed as a less conservative correction for multiple hypothesis testing (Hoc hberg and Benjamini, 1990). Instead of controlling for any false positives (as in the Bonferroni method), FDR controls the expected proportion of false positives. The FDR threshold is determined from the observed distribution of p-values in a given data set. FDR adjusted p-va lues, also known as q-valu es, were calculated for all analyses using the SAS software package. Gi ven that my sample sizes are relatively small, positive results may not be very robust. Thus, as indicated in my tables, I report the uncorrected and Bonferroni corrected p-values and the FDR corrected q-values for most findings. None of the pairwise haplotype findings showed positiv e associations after corrections for multiple testing were applied because th ere were a total of 900 tests for this method; therefore, only uncorrected p-values are reporte d. The 900 pairwise haplotype tests are from 75 pairs of genes compared 4 possible haplotypes for each pair 3 comparisons of the number of risk alleles (2 vs. 1, 2 vs. 0, and 1 vs. 0). Results A total of 30 SNPs were genotyped from eight different genes in 332 healthy controls, 355 vitiligo patients and 359 of their unaffected fami ly members. All races were included for TDT 41

PAGE 42

tests, but only whites (212 independent cases and 245 unrelated controls) were used for all other analyses due to power limitations. The positive associations are discu ssed in the following two chapters. Described here are the completely ne gative results found. It is important to report negative results in order to narro w the pool of possible vitiligo su sceptibility genes and thereby better focus limited research resources. Genotype by Sex Analyses Genotype by sex analysis both within cases a nd between cases and controls yielded no significant results for any of the genes in my study (data not shown). Pairwise haplotype analysis examining sex differences by genotype also yielded no significa nt results (data not shown). This is not surprising considering that vitiligo appears to occur with equal frequency in both men and women. FBXO11 and MSH6 Allelic frequencies for FBXO11 and MSH6 were calculated for my white vitiligo case and control populations, and the minor allele fre quencies (MAF) from the HapMap project (2003) and the Applied Biosystems validation set (De La Vega et al., 2006) are given, if available (Table 2-4). Allelic frequenc ies of vitiligo patients and healthy controls did not differ significantly for either rs3136367 [p=0.13, uncorr ected(UC)] and rs960106 (p=0.95, UC). In addition, the two MSH6 and FBXO11 SNPs tested did not have significantly different genotype frequencies for the vitiligo patients and contro ls. The uncorrected p-values for rs3136367 and rs960106 for the genotype analysis were 0.21 and 0.98 respectively, indicating that there is no significant difference between cases and controls. The test for Hardy-Weinberg showed that there was disequilibrium for controls for the rs3136367 SNP (p=0.01, UC), though the interpretation of this finding is not clear. It may be due to ge notyping error; however, deviations from HWE have been observed in many associati ons studies (Salanti et al., 2005). Family-based 42

PAGE 43

testing for these SNPs also yiel ded no significant results (Table 2-5). No subsequent analysis with these SNPs, including pairwise haplobloc k comparisons, yielded any significant findings (data not shown). Discussion The FBXO11 gene is located on chromosome 2p16 a nd consists of 23 exons that code for an 844 amino acid protein. MSH6 is also located at the same region on chromosome 2p16 and is coded on the opposite strand from FBXO11; the 3 ends of both genes overlap by 5 kilobases (kb) (Le Poole et al., 2001). MSH6 has 10 exons that code for a 1360 amino acid protein. A previous study using semi-quantitative revers e transcription polymerase chain reaction (RTPCR) testing of MSH6 and FBXO11 mRNA from cells cultured from vitiligo patients and healthy controls found that MSH6 mRNA was upregulated in viti ligo patients compared to controls while FBXO11 mRNA was downregulated in vitilig o patients (Le Poole et al., 2001). Le Poole and colleagues hypothe sized that the formation of RNARNA hybrids from the overlapping regions of MSH6 and FBXO11 may interfere with the MSH6 proteins G/T mismatch repair function through post-trans criptional gene silenc ing [also known as RNA interference (RNAi)] (2001). To my knowledge, no previous vitiligo a ssociation studies examining SNPs in MSH6 and FBXO11 have been conducted. A genomewide screen of generalized vitiligo patients did not find evidence to support genetic association for MSH6 and FBXO11; however, this study was conducted in multiplex families and so may be less applicable to singleton cases, who comprise the majority of vitiligo patients (Fain et al., 2003 ). Two review articles by Spritz dismiss the pathogenicity of FBXO11, though no experimental findings ar e given to support this claim (2006; 2007). In my study, the MSH6 and FBXO11 candidate genes did not demonstrate statistically significant p-values (p 0.05, uncorrected) in any case-c ontrol and/or family-based 43

PAGE 44

44 analyses. This would suggest that, at leas t in my study population, they are not vitiligo susceptibility genes. It is important to note that vitiligo is most likely a dis ease with a complex etiology. Therefore, susceptibility genes that are uncom mon in a population, or genes that have only a subtle contribution to the vitiligo phenotype, may not show strong associations in case-control analyses. It is with this in mind that I dismiss these ge nes as possessing no currently demonstrable genetic association with vitiligo in my sample population.

PAGE 45

45Table 2-1. Demographic information for viti ligo patients and their unaffected relatives Vitiligo patients Sex Race Unaffected relatives Sex Race Male 128 Black 7 Male 164 Black 7 (36%) Chinese 2 (46%) Chinese 3 Hispanic 28 Hispanic 30 Female 227 Indian (Asian) 8 Female 195 Indian (Asian) 9 (64%) Mixed 16 (54%) Mixed 11 Native American 1 Native American 0 Vietnamese 1 Vietnamese 0 White 279 White 290 Other 4 Other 3 Not reported 9 Not reported 6 Total 355 Total 359

PAGE 46

46 Table 2-2. List of Applied Biosyste ms SNP Assay IDs and description of FBXO11 and MSH6 SNPs genotyped Gene(s) SNP rs number Applied Biosyste ms ID Base pair change & location FBXO11 rs960106 C____448948_10 C/T 5 UTR FBXO11 & MSH6 rs3136367 C__22271588_10 C/G 3 UTR (FBXO11), intron 8 (MSH6) Table 2-3. Linkage disequilibrium (D ) values for FBXO11 and MSH6 Gene SNP ID SNP ID D FBXO11/MSH6 rs3136367 rs960106 0.81

PAGE 47

47Table 2-4. Minor allele frequencies (MAF) for FBXO11 and MSH6 SNPs from the HapMap project Caucasians from European Ancestry (HapMap CEPH), Applied Biosystems Caucasian cohor t (ABI Caucasian), and valu es for the white case and control groups from the present study; 2 p-values and q-values for allelic comparisons are given Gene Symbol SNP ID HapMap CEPH MAF ABI Caucasian MAF White case MAF # of white cases White control MAF # of white controls p-value BonC p-value FDR q-value FBXO11 rs960106 0.38 (C) 0.4 (C) 0.42 (C) 208 0.42 (C) 239 0.95 1.00 0.95 MSH6 rs3136367 0.22 (C) 0.27 (C) 0.26 (C) 188 0.30 (C) 238 0.13 1.00 0.65 BonC denotes the Bonfe rroni corrected p-value

PAGE 48

Table 2-5. Family-based association (tra nsmission disequilibrium test) results for FBXO11 and MSH6 Gene SNP Number of informative parents Transmitted (allele) Not transmitted (allele) p-value BonC p-value FDR q-value FBXO11 rs960106 75 35 (C) 40 (T) 0.56 1.00 0.84 FBXO11 MSH6 rs3136367 45 17 (C) 28 (G) 0.10 1.00 0.43 BonC denotes the Bonfe rroni corrected p-value 48

PAGE 49

CHAPTER 3 AUTOIMMUNE REGULATOR GENE Introduction As in other autoimmune diseases, one hypothesi s for the genetic pathogenesis of vitiligo is that immunoregulatory genes may contribute to the autoimmune response in vitiligo patients. One gene implicated in immune function that is examined in the current study is the autoimmune regulator, or AIRE gene. The AIRE gene is located on chromosome 21q22.3 and consists of 14 exons coding for a 2445-base pair mRNA transcri pt; the translated produ ct has 546 amino acids with a molecular mass of approximately 57.5 kDa (Nagamine et al., 1997). At the subcellular level, AIRE can be found in the cell nucleus in a speckled pattern also known as nuclear dots (Bjorses et al., 1999). Mutations in the AIRE gene are known to cause autoimmune polyendocrinopathycandidiasis-ectodermal dystrophy syndrome (A PECED), otherwise known as autoimmune polyglandular syndrome (APS). The most common features of APECED are parathyroid gland failure, chronic suscepti bility to candida yeast infecti on, and Addisons disease. Other autoimmune diseases associated with APECED include alopecia, viti ligo, ovarian failure, testicular atrophy, hypothyroidism, gastric parietal cell atrophy, intestinal malabsorption, and insulin-dependent diabetes mellitus. It is possibl e that subtle changes in this gene may confer vitiligo susceptibility, in the absence of APECED. AIRE is an important DNA binding molecule invol ved in immune regulation (Kumar et al., 2001; Peterson et al., 2008). The AIRE protein is predominantly expressed in thymic epithelial cells (TECs), but also in some monocyte-derived cells of the thym us, in a subset of cells in lymph nodes, in the spleen, and in the fetal liver. AIRE mRNA is weakly detected in peripheral blood leukocytes but thymic AIRE expression is several fold highe r than in other tissues (Su and 49

PAGE 50

Anderson, 2004; Mathis and Benoist, 2009). Due to th is expression pattern, it is thought that the AIRE protein may be involved in the maintenance of thymic tolerance, where AIRE up-regulates the expression of organ-specific an tigens to facilitate the negative selection of autoreactive Tcells, thus enhancing self-tolerance and limiting au toimmunity (Anderson et al., 2002). It is possible that detrimental mutations in AIRE may be responsible for incomplete negative selection of self-antigens, resulting in the eventual development of multiple autoimmune disorders (Pitkanen and Peterson, 2003; Li ston et al., 2004; Liston a nd Goodnow, 2005). However, several studies searching for mutations in the AIRE gene leading to non-APECED autoimmune disease have not yielded signifi cant findings (Turunen et al., 2006; Jin et al., 2007a; Boe Wolff et al., 2008). SNPs within the AIRE gene were examined in my study to test if subtle alterations in this genes function may contribute to the autoimmune pathogenesis of vitiligo. Materials and Methods DNA collection and SNP genotyping were performed as described in Chapter 2. Statistical analyses were performed as detailed in Chapter 2. Two SNPs on the AIRE gene were chosen to test the hypothesis that this imm unoregulatory gene is a vitiligo sus ceptibility gene (Table 3-1). The rs2776377 SNP, located 5 upstream of AIRE but on the same haplotype block as the 5 coding sequence, with high MAF in Caucasians in the Applied Biosystems validation set, was examined. Additionally, rs1800520, a SNP in e xon five on the same haplotype block as rs2776377, was chosen because it causes a nonsy nonymous change of a polar amino acid (serine) to a charged residue (a rginine) (S278R) in the SAND domain of the AIRE protein. This may have an effect on the postulated AIRE DN A-binding activity (Taz i-Ahnini et al., 2002; Purohit et al., 2005). 50

PAGE 51

Results The rs1800520 SNP, which encodes the amino acid substitution, showed very highly significant association to vitiligo, p=0.0000002 uncorre cted (UC) in my case-control genotype analysis (Table 3-2). The p-value after correc ting for multiple tests us ing Bonferronis method (BonC) was <0.0001, and FDR q-value was <0.0001. Most notably, the GG genotype of the minor allele (resulting in homozygosity for the arginine residue) was seen in 19 out of 237 controls (8%) and 0 out of 163 vitiligo patien ts (0%). The CG heterozygous genotype was observed in 88 controls (37%) and 33 cases (20%). Allelic frequencies were calculated for cases and controls for rs1800520, and the minor alle le frequencies (MAF) from the Applied Biosystems validation set (De La Vega et al., 2006) are given (Table 3-3); the 2 allelic frequency analysis for my white case-c ontrol groups was significant at p=9.9038E-09 (uncorrected), p<0.0001 (BonC) and q<0.0001. The G allele was found to have a protective effect: RR=0.38 (CI= 0.2667 to 0.5439). Convers ely, the C allele for rs1800520 confers susceptibility. It is noteworthy that my c ontrol MAF of 0.27 differs from the Appplied Biosystems Incorporated (ABI) reported MA F of 0.13 for the rs1800520 SNP. ABI allele frequencies were obtained from a set of 45 people, while my study included 163 controls genotyped for rs1800520. In contrast, white casecontrol analysis of the rs2776377 SNP yielded no significant findings at the alle lic (Table 3-3) or genotype leve ls (p=0.46) (Table 3-2). TDT analysis for both SNPs was not significant (Table 3-4). Linkage disequilibrium was calculated for the two AIRE SNPs (Table 3-5), showing moderate disequilibrium and a haplotype analysis yielded signif icant findings. The haplotype with the highest log-likelihood for the AIRE gene was the AG haplotype (denoted as [AG]), having the A allele for rs2776377 and the G a llele for rs1800520. I f ound that the presence of [AG] was significantly higher in the control group compared with vitiligo patients (p < .0001 51

PAGE 52

UC) (Table 3-6). The odds ratio analysis showed that controls carrying the haplotype [AG] had significantly decreased risk of developing si gns of vitiligo by 0.1643 times, when the vitiligo patient group was compared with the cont rol group (OR = 0.1643, 95% CI = 0.0663 to 0.4070). Discussion The SNP that I found to be significantly asso ciated with vitiligo, rs1800520, is located in the AIRE gene SAND DNA-binding domain (Gibson et al., 1998). The SAND domain (named after S p100, AIRE -1, N ucP41/75, D EAF-1) is a conserved 80-re sidue amino acid sequence found in a number of nuclear proteins, many of which function in chromatin-dep endent transcriptional control (Bottomley et al., 2001). Purohit a nd colleagues found evidence that the SAND domain of AIRE recognizes a TTATTA DNA motif (Pur ohit et al., 2005). Analysis of deletion constructs of recombinant proteins revealed th at the major activity of the AIRE SAND domain resides in the area of amino acids 189. The SNP I found to be associated with vitiligo in my study causes a change in amino acid 278 from a serine to an arginine; changes at this site have an unknown effect on SAND activity though the substitution of a neutral amino acid (serine) for a positively charged one (arginine) may disrupt DNA binding as DNA is negatively charged. It may also affect protein folding or trafficking, or potentially a ffect regulation by phosphorylation. The significant [AG] protectiv e haplotype I observed is likel y driven by the very strong association with vitiligo susceptibility found with the C alle le on rs1800520. Although the exact mechanism of this haplotypes influence on AIRE function is unknown, it could alter AIRE protein activity as rs1800520 does cause an amino acid change. Additionally, irrespective of the effects of rs1800520, this haplotype may be in linka ge disequilibrium with a gene alteration that is actually responsible for the pa thogenic effect. Although I cannot rule out a direct effect of this haplotype on AIRE gene function and its role in vitiligo susceptibility, more investigation is 52

PAGE 53

warranted to rule out other f actors contributi ng to this association, which is still highly significant even with the conser vative Bonferroni correction. Several groups examining autoimmune diseases and the AIRE gene have studied the same SNP that I found to be significantly associated w ith vitiligo; however, th e other studies yielded mixed results. The rs1800520 SNP showed no eviden ce of association with type 1 diabetes in a Finnish population (p=0.43) (Turune n et al., 2006), and genotype fr equencies were nearly equal among cases (0.015) and controls (0.016) for the homozygous minor allele that was found to be very highly significant in my study. This SN P also showed no association to autoimmune Addisons disease (AAD) in a Norwegian populat ion (Boe Wolff et al., 2008). However, a British group found rs1800520 to be significantly associated with alopecia area ta (Tazi-Ahnini et al., 2002). Prior studies testing vitiligo association with the AIRE gene had mixed results. The same British group that found AIRE to have strong association with alopecia areata also found very strong (p=0.000414) associations with vitiligo and SNPs in the AIRE gene. Further, their haplotype analysis of this gene also proved to be significant (Tazi-Ahnini et al., 2008). In contrast, Jin and colleagues did not find significant association with AIRE and vitiligo in tests of eight AIRE SNPs (2007a). The immune-related gene AIRE was investigated because of its role in the modulation of immune responses. Statistical analysis for cases and controls included on ly whites with vitiligo, due to a lack of sufficient nu mbers of subjects of other et hnic backgrounds. An autoimmune component in vitiligo susceptibility may be supp orted by the results of my study. The lack of a significant TDT to support the association of AIRE with vitiligo suggested by my white casecontrol analysis may indicate that larger sample sizes are needed to confirm this finding in 53

PAGE 54

family-based association studies. Alternatively, it is possible that this is a spurious finding; notably, our control allele fr equency for rs1800520 differed from the ABI reported frequency. Pairwise haplotype analysis of this gene was also significant, though the addition of other SNPs on this gene may help to understand th e mechanism underlying this finding. 54

PAGE 55

55 Table 3-1. List of Applied Biosystems SNP Assay IDs and description of AIRE SNPs genotyped Gene SNP rs number Applied Biosystems ID Base pair change & location (amino acid change & location) AIRE rs2776377 C___2978271_1_ A/G promoter AIRE rs1800520 C___9480541_1_ C/G exon 7 (S278R) Table 3-2. White case-control resu lts for genotype analysis of AIRE SNP Case Control p-value BonC p-value FDR q-value rs2776377 AA 54 (33%) AG 83 (51%) GG 26 (16%) AA 86 (37%) AG 104 (45%) GG 43 (18%) 0.46 1.00 0.55 rs1800520 CC 130 (80%) CG 33 (20%) GG 0 (0%) CC 130 (55%) CG 88 (37%) GG 19 (8%) 0.038* <0.0001* <0.0001* denotes a significant finding; BonC denot es the Bonferroni corrected p-value

PAGE 56

56Table 3-3. Minor allele frequencies (MAF) for AIRE SNPs from Applied Biosystems Caucasian cohort (ABI Caucasian) and values for the white case and control gr oups from the present study; 2 p-values and q-values for allelic comparisons are given SNP ID ABI Caucasian MAF White case MAF # of white cases Case alleles White control MAF # of white controls Control alleles p-value BonC p-value FDR q-value rs2776377 0.42 (G) 0.41 (G) 163 A 191 (59%) G 135 (41%) 0.41 (G) 233 A 276 (59%) G 190 (41%) 0.86 1.00 0.93 rs1800520 0.13 (G) 0.10 (G) 163 C 293 (90%) G 33 (10%) 0.27 (G) 237 C 348 (73%) G 126 (27%) 9.90E-09* <0.0001* <0.0001* denotes a significant finding; BonC denot es the Bonferroni corrected p-value

PAGE 57

Table 3-4. Family-based association (tra nsmission disequilibrium test) results for AIRE SNP Number of informative parents Transmitted (allele) Not transmitted (allele) p-value BonC p-value FDR q-value rs2776377 51 25 (A) 26 (G) 1.00 1.00 1.00 rs1800520 17 12 (C) 5 (G) 0.16 1.00 0.46 BonC denotes the Bonfe rroni corrected p-value Table 3-5. Linkage disequilibrium (D ) value for AIRE Gene SNP ID SNP ID D AIRE rs2776377 rs1800520 0.47 Table 3-6. Pairwise haplotype analysis results for AIRE rs2776377 and rs1800520 Frequency Controls Vitiligo patients p-values 0 [AG] 188 (81.5%) 154 (96.5%) 1 [AG] 41 (18%) 6 (3.5%) 2 [AG] 1 (0.5%) 0 (0%) Total (n) 230 160 2 p<.0001* Odds ratio for 1 [AG] vs. 0 [AG] = 0.1673 95% CI=0.0663 to 0.4070 p<.0001* Odds ratio for 2 [AG] vs. 0 [AG] = n.d. n.d. Odds ratio for 2 [AG] vs. 1 [AG] = n.d. n.d. n.d. = not defined; denotes a significant finding ; number in frequency column indicates the number of copies of this haplotype 57

PAGE 58

CHAPTER 4 MELANIN BIOSYNTHESIS GENES Introduction Melanin biosynthesis is a complex process that takes place in specializ ed organelles called melanosomes. Approximately 1500 proteins have been identified in melanosomes during of all stages of their maturation, with about 600 expr essed in any given stage. Approximately 100 proteins are shared by melanosomes from pigmented and nonpigmented melanocytes, and this number likely represents the e ssential melanosome proteome (Chi et al., 2006). Several genes involved in melanin biosynthesis were examin ed in the present study to determine their association with vitiligo. A ll genes examined showed at le ast one positive result and are discussed in this chapter. The involvement of these genes could be supportive of the autoimmune and/or oxidative stress theori es of vitiligo pathogenesis. Whereas lack of phenylalanine hydroxylase is most commonly associated with the autosomal recessive disorder phenylketonuria (PKU), the role of this enzyme in melanin biosynthesis has been supported by a number of studies (Camacho and Mazuecos, 1999; Schallreuter and Wood, 1999; Scha llreuter et al., 2005). The PAH gene is located on chromosome 12q22 and has 13 exons that encode the 453 amino acid enzyme phenylalanine hydroxylase. This enzyme converts the essentia l amino acid phenylalanine into tyrosine, the precursor of melanin. A disruption or a complete loss of PAH activity in the liver results in a dramatic increase in serum concen trations of phenylalanine in PKU patients. PKU patients must eat a special diet lacking phenylalanine because excess accumulation of this amino acid has detrimental neurological effects. In addition to neurological impairment, untreated PKU patients have poor pigmentation, adding supp ort to the hypothesis that PAH protein function is vital to melanogenesis. The role of PAH in producing ty rosine is consistent with the notion that 58

PAGE 59

alterations in gene function could have an eff ect on melanogenesis by decreasing the availability of tyrosine (Schallreu ter et al., 1998). Essential to melanogenesis are several genes in the tyrosinase gene family that are expressed in melanoc ytes: tyrosinase (TYR ), tyrosinase-related protein 1 ( TYRP1 ) and dopachrome tautamerase ( DCT ), (formerly known as tyrosinase-related protein 2 ( TYRP2 )). Disruptive recessive mutations in TYR and TYRP1 are linked to different types of oculocutaneous and ocular albinism; this suppor ts their essential role in pi gmentation and as candidates for involvement in vitiligo (Oe tting and King, 1999). No known human phenotypes have been linked to changes in DCT However, DCT along with TYR and TYRP1 proteins were observed by several groups to be possible targets of autoantibodies found in the serum of vitiligo patients (Song et al., 1994; Kemp et al., 1997a; Kemp et al., 1997b; Kemp et al., 1998; Okamoto et al., 1998; Kemp et al., 1999b). This further supports the hypothesis of these genes contributing to vitiligo susceptibility. The five-exon TYR gene is located at chromosome 11q14-q21 and encodes the 529 amino acid tyrosinase enzyme, with 50% of the coding region being in exon one. Tyrosinase has a dual function in melanin production. It oxidizes tyrosine to dopa in the first and rate-limiting step in the melanin biosynthetic pathway, and it oxidizes dopa to dopaquinone, an in termediate step in the melanin pathway. Tyrosinase is hypothesized to be part of a multienzyme complex in the melanosome that includes tyrosinaserelated pr otein 1 and dopachrome tautamerase. The TYRP1 and DCT enzymes are thoug ht to be involved in the latter part of the melanin biosynthetic pathway, though their exact role(s) in melanin biosynthesi s have yet to be elucidated (Oetting, 2000). However, it is known that TYRP 1 significantly enhances tyrosinase catalytic function (Kobayashi et al., 1994; Kobayashi et al., 1998). TYRP1 is located on chromosome 59

PAGE 60

9p23 and has eight exons encodi ng a 538 amino acid protein. DCT which is located on chromosome 13q32, contains eight exons that encode a protein of 520 amino acids. The catechol-O-methyltransferase or COMT gene is located on chromosome 22q11.21 and has six exons that code for a 272 amino acid protein. COMT protein is in volved in regulation of oxidative damage in the melanocyte by preven ting the formation of toxic o-quinones during melanin synthesis (Pavel et al., 1983). One group found that epidermal homogenates from vitiligo patients express higher levels of COMT ac tivity than homogenates from healthy controls (Le Poole et al., 1994). This increased level of COMT protein may indicate higher levels of oxidative stress in vitiligo patients, which supports the autocytotoxic mode l of disease etiology. It has been hypothesized that COMT may play a direct role in the regulation of melanin biosynthesis (Wakamatsu et al., 1990). The pr oposed mechanism of th is regulation is the methylation of the melanin precursor molecule dihydroxy indole-2-carboxylic acid (DHI-2C) by COMT, making it unavailable for incorporat ion into melanin (Das et al., 2001). Because reactive oxygen species are byproducts of their enzymatic activities, subtle alterations in genes involved in the melani n biosynthetic pathway may result in increased oxidative stress in the melanocyte, which would be supportive of the autocytotoxic hypothesis of vitiligo pathogenesis. Additionally, the TYR TYRP1 and DCT genes may also contribute to the autoimmune component of vitiligo pathogenesis. As in other autoimmune diseases, the genetic susceptibility in vitiligo patients may be due to proteins expressed in melanocytes acting as autoantigens that target melanoc ytes for destruction by the immune system. I examined SNPs in candidate genes PAH TYR TYRP1, DCT, and COMT 60

PAGE 61

Materials and Methods DNA collection, SNP genotyping, and statistical analyses for genetic association were performed as described in Chapter 2. A total of twenty-six SNPs were selected from five genes (Table 4-1). Two SNPs on PAH eight on DCT (Figure 4-1A), eight on TYR six on TYRP1 (Figure 4-1B) and two on COMT were chosen to allow for haplotype analysis for each gene (Table 4-1); only significant values are re ported. Pairwise linka ge disequilibrium (D ) values were calculated for all SNPs within a gene (Table 4-2). Hardy-Weinberg values were calculated for all SNPs; only significant va lues are reported (Table 4-19). Results Phenylalanine Hydroxylase Case-control and family-based analysis of the two individual SNPs in PAH yielded no significant association with viti ligo (Table 4-3, 4-4, and 4-5). Ha rdy-Weinberg analysis was not significant. The two SNPs were strongly linked, with D =0.48 (Table 4-2). Pairwise haplotype analysis revealed that the [GG] haplotype was significant for cases (11 out of 203) compared to controls (0 out of 241), p=0.0007 (Table 4-6). Be cause zero controls had the [GG] haplotype, odds ratios could not be cal culated for this finding. Dopachrome Tautamerase Analysis of the DCT gene yielded several positiv e findings. Most notably, SNP rs9516413, a C/T polymorphism located in the 3 UTR, was positively associated with vitiligo in both the allelic (p=0.0047) and genotypic frequenc y analyses (p=0.0270) (Table 4-3 and 4-4). Specifically, the C allele was overrepresented in cases and had a relative risk (RR) =1.43 (CI=1.1145 to 1.8397); TDT analysis also showed this allele to be signifi cant (p=0.0052) (Table 4-5). In pairwise haploblock an alysis, four of five significant fi ndings for this gene (Table 4-7, 4-8, 4-10, and 4-11) involved the C allele from rs9516413. Three of the other SNPs significant 61

PAGE 62

for haplotype findings were located in intron six (rs7987802, rs4318084, rs9524493) with the other two located in intron two (rs11618471), and intron one (rs7991232). However, no other DCT SNPs had significant associations in either th e allelic or genotypic an alyses (Table 4-3 and 4-4), nor was Hardy-Weinberg analysis signific ant for my white cases and controls. Including rs9516413, five out of eight DCT SNPs tested had significant TDT findings (Table 4-5). The other four SNPs were all intronic, with rs4318084 and rs9516418 both located in intron six, and rs1028805 and rs11618471 both located in intron two. Lastly, three ou t of four SNPs tested in intron six met statistical significance in viti ligo patients who had co-morbid autoimmune diseases: rs7987802 (p=0.038), rs4318084 (p=0. 038), and rs9524493 (p=0.033) (Table 4-12). Interestingly, rs9516413 was not among them despite its association in case-control analysis. Tyrosinase Of the eight SNPs genotyped in TYR, only one yielded significant results in independent tests (Table 4-3 and 4-4). An A/G poly morphism in intron 2, rs12791412 was borderline significant at the allelic level (p=0.05) (Table 4-3); the A allele had a RR= 1.08 (CI=1.0002 to 1.1662), though this SNP showed no significance upon genotypic analysis (p=0.13) (Table 4-4). TDT analysis for rs12791412 was significant for exce ssive transmission of the A allele to cases (p=0.02) (Table 4-5). Analysis of variance (ANO VA) indicated a significan t association of this SNPs genotype with age of disease onset; th e GG cohort had a mean age of onset of 33.67 years, compared to the AA cohort mean age of onset at 19.58 years (p = 0.027) and the AG group mean age of onset 14.56 years (p=0.0022) Hardy-Weinberg analysis of all TYR SNPs showed no disequilibrium (data not shown). Si x of the two-SNP haplotype analyses were significant (Tables 4-13, 4-14, 415, 4-16, 4-17, and 4-18). Of th ese, five contain rs1042602, a SNP in exon one that causes an amino acid chan ge from serine to tyrosine at position 192 (S192Y). In haplotype analysis, homozygosity for the C major allele in rs1042602 (coding for 62

PAGE 63

the serine residue) was observed more often in cases compared to cont rols; odds ratios were significant for carrying two risk alleles and for he terozygosity for the risk allele compared to carrying no risk alleles. This would imply that carrying the C allele at rs1042602 is protective for vitiligo while carrying the A allele confers risk. The other significant haplotype findings were with SNPs located in intron 2 (rs10765197, rs12791412, rs2000554), intron 3 (rs10830250), and in intron 4 (rs1827430). The onl y haplotype not involving rs1042602 was one that yielded a protective effect Homozygosity for the G allele of rs12791412 and the A allele of rs1827430 was observed more often in controls th an in cases (OR=0.3761) when compared to homozygosity for the other haplotype (p=0.0119). Tyrosinase-Related Protein 1 Six TYRP1 SNPs were genotyped in my study. Ha rdy-Weinberg equilibrium was intact for controls but was significantly deviated for patients for four TYRP1 SNPs (Table 4-19). One of these SNPs in H-W disequilibrium, rs2733833, was the only TYRP1 SNP to have a significant finding in my case-control genotypic analysis (T able 4-4). This G/T polymorphism, located in intron 6, had a significant asso ciation at the genotype level (p=.0254), where the TT genotype was observed less often in cases th an controls. Significance for th is SNP was not replicated in the TDT test (Table 4-5). Pairwise haplotype analysis of the TYRP1 gene yielded six significant pairwise associations (Tables 4-20, 4-21, 4-22, 4-23, 4-24, and 4-25); of the 6 TYRP1 SNPs that I genotyped, 5 had at least one si gnificant haplotype. Four of these involved rs2733833. The other SNPs were located in intron four (rs2762462), intron 5 (rs2733831), and three in the 3 UTR (rs683, rs2762464, and rs1063380). Catechol-O-Methyltransferase Two COMT SNPs, one in the 5 UTR and one in intron two, were examined in my study. Case-control and family-based analysis of the two individual SNPs in COMT yielded no 63

PAGE 64

significant association with vitiligo (Table 4-3, 4-4, and 4-5). Hardy-Weinberg analysis was not significant. The two SNPs were very strongly linked, with D equal to 0.79, and pairwise haplotype analysis showed a protective haplotype of [CG] (p=0.0046), but only comparison of individuals who carried one protec tive haplotype [CG] versus thos e carried zero was significant (p=0.0011) (Table 4-26). Summary A summary table of all uncorrected findings for all genes in this chapter is included to aid in the integration of this large amount of material (Table 4-27). As expected, given the p-values and small sample size, multiple-test corrections resulted in p-values >0.05 for these analyses, which might be true effects. For the purpos es of the discussion, I will remark about the significant uncorrected findings should they be validated in future studies. Discussion Phenylalanine Hydroxylase Individually, the two PAH SNPs tested did not yield signif icant results; however, there was a significant risk haplotype for individuals who were homozygous for the G allele for both markers tested. Because both PAH SNPs tested are intronic, it is possible that one or both affect splicing efficiency or isoform usage, if one or both are truly exerting the functional effect. This could be further investigated in follow-up studies likely using splice-site prediction software or reverse-transcription PCR analyses. It is also possible that the [GG] ha plotype observed only in vitiligo patients is in linkage di sequilibrium with some unidentif ied gene change that directly contributes to vitiligo pathogenesi s. Additional genotyping in more samples, replication in other patient cohorts, and deep sequenc ing of this gene are logical st eps to help to understand the significance of this finding. 64

PAGE 65

Dopachrome Tautamerase Examination of the DCT gene yielded a number of inte resting results. The strongest association with vitiligo was observed with th e C allele in rs9516413. This was seen in casecontrol analysis at the allelic and genotypic levels, in TDT anal ysis, and in pairwise haploblock analysis with five other DCT SNPs. As this SNP was observed in all of the significant pairwise associations, it is likely that it drove the significance for all DCT haplotype findings. The rs9516413 SNP is located in the 3 UTR and thus could potentially affect DCT RNA stability, polyadenylation site usage, and/or regulate transcripti on, perhaps by affecting an enhancer or a miRNA binding site. Additionally, this SNP may be in linkage disequilibrium with another gene polymorphism not examined in the present study. Four additional intronic SNPs had significant TDT results, but did not reach significance in case-control analysis. With a larger sample size, or with patients and controls from more diverse ethnic backgrounds, these SNPs may prove to be significant in subsequent studies. SNPs located in introns may affect RNA splicing and/or stabilit y, or they may be linked to other gene changes that directly alter gene ex pression or protein function. The only SNPs in my study significantly associated with presence of co-morbid autoimmune diseases were observed in DCT All three of these SNPs were located in intron 6. As with all intronic SNPs, it is difficult to dete rmine the role that these polymorphisms play in enzyme activity. They may be linked to an epitope encoded by DCT that is also common to other self-proteins. Exons five and six would be a logical place to star t a search for a linked protein-level effect such as a missense polymorphism It is possible that vitiligo patients with and without co-morbid autoimmune diseases re present two different sub-sets of disease pathogenesis, though the differential role of DCT in these subgroups needs to be explored. 65

PAGE 66

Tyrosinase The role of tyrosinase in pigment production is clear as this enzyme represents the ratelimiting step of melanogenesis. Severe disruptio ns of tyrosinase function are known to cause albinism, though I hypothesize that more subtle alterations of this ge ne may be linked to vitiligo. Of the eight SNPs genotyped in TYR, only one yielded significant results in independent analyses. The rs12791412 SNP, located in intr on two, was significant in allele frequency analysis between white cases and controls and in family-based te sting. This significance was not seen in genotype analysis. In my entire study, th is was the only SNP that yielded a significant finding for mean age of onset by genotype associ ation. Patients homozygous for the minor allele (G) had a significantly later mean age of onset. However, because I had a limited sample size for my within-case analysis (n=209), and my finding wa s linked to the minor allele, my results must be interpreted with caution. Fu rther, the average age of ons et for heterozygotes was not intermediate to both homozygotes. Six haplotype findings were significant and five of these contained rs1042602, a SNP in exon 1 that causes the S192Y amino acid change. This SNP has been found to be significantly linked to skin pigment variation in a Sout h Asian population (Stokow ski et al., 2007), and several studies found it to be highly polymorphic in European populations contributing to normal variation in skin color (Shriver et al., 2003; 2005; Norton et al., 2007). One group found rs1042602 to be associated with freckling in an Icelandic population (Sulem et al., 2007). Additional investigation by Sulem et al. (2007) found that the ancestral C allele of rs1042602 was fixed in the East Asian and Nigerian HapMap samples, whereas the A allele was found at a frequency of approximately 35% in European populations. They found strong evidence that the rs1042602 A allele (which was associated with the absence of freckles in Icelanders) was subject to positive selection in European populations. Interestingly, only 1.7% of HapMap SNPs show a 66

PAGE 67

greater frequency difference between the European and African samples, and only 0.37% show a greater frequency difference between the Europe an and East Asian samples than rs1042602. In five TYR pairwise haplotypes, I f ound that homozygosity for the C allele of rs1042602 (coding for serine) was significantly associated with vitiligo. It is unknown whether the amino acid change of a serine to a tyrosine affects protein functi on, although it is a non-conservative substitution and this residue could affect the active sites of this protein (although possible structure changes are not known) Alternatively, these risk haplotypes may be in linkage disequilibrium with another TYR change that is directly relate d to vitiligo pathogenesis. The protective haplotype I observed in TYR involved two intronic SNPs. Tyrosinase-Related Protein 1 Whereas the exact function of tyrosine-related protein 1 in melanogenesis is still debated, it has been shown to enhance tyrosinase function (Kobayashi et al., 1994; K obayashi et al., 1998). Additionally, mutations severe ly disrupting TYRP1 enzymatic function are known to cause human albinism (Oetting and King, 1999). I examined six TYRP1 SNPs in my study, hypothesizing that sub tle variation in the TYRP1 gene may contribute to vitiligo pathogenesis. I observed four SNPs to be in Hardy-Wei nberg disequilibrium (HWD) for cases but not controls. Interpreting HWD remains controvers ial (Salanti et al., 2005; Zou and Donner, 2006; Teo et al., 2007; Li and Li, 2008), but some groups suggest that a situation such as this is supportive of genetic association because it implie s that the allele frequencies are stable in controls but changing in cases. Only one TYRP1 SNP had a positive finding in independent analysis at the genotypic level. Located in in tron 6, rs2733833, one of the SNPs in HWD, had a significant association for genotype frequencies. Specifically, the TT genotype was observed less often in cases than controls suggesting a possible protective effect. However, because this 67

PAGE 68

SNP was not significant in allelic or family-bas ed testing, additional studies are warranted to replicate this finding and ru le out the possibility of spurious association. Pairwise haplotype analysis of the TYRP1 gene was particularly interesting. Of the six SNPs genotyped, five were significant in at least one pairwise association. Three of these SNPs were intronic and two were located in the 3 UTR. A total of six pair wise comparisons yielded significant results. Four of these haplotypes were significant for one risk allele compared to none, with an odds ratio >1, indicating a predisposition to disease. However, all six haplotypes were significant for two risk allele s compared to one, with an odds ratio <1. This indicates that homozygosity for these haplotypes is protective, but there is in creased heterozygosity observed in my vitiligo patient population (suggestive of a dominant-negative effect). A similar dominantnegative finding was found in this same vitiligo cohort for a positively-associated SNP in the catalase gene (Casp et al., 2002). Because tyrosina se-related protein 1 is believed to function as part of a multienzyme complex in the mela nosome, heterozygosity for changes in the TYRP1 gene may ultimately affect binding, stability, or function of this protein complex. Dominantnegative effects are well understood in other syst ems such, as osteogenesis imperfecta involving collagen multimers (Gajko-Galicka, 2002). Catechol-O-Methyltransferase Individually, the two COMT SNPs tested did not yield significant results; however, there was a significant protective haplot ype for individuals who were ca rried a C allele for a SNP in the 5 UTR (rs2020917) and a G allele for a SNP in intron two (rs6269). Because neither of the SNPs were significant in separate analyses, the [CG] haplotype may be in linkage disequilibrium with a functional change that he lps to prevent vitiligo, but this linkage may be weak so only the combination of both alleles is strong enough to be detected. The homozygosity for the [CG] haplotype was rare in both vitiligo patients (2%) and controls (3%), so only comparisons 68

PAGE 69

69 between individuals who were carried one protec tive haplotype versus those who carried zero reached significance. It is noteworthy that other SNPs in COMT have been found to be positively associated to vitiligo. One group observed that homozygosity for a SNPs in COMT causing an amino acid change (from valine to methionine at positi on 158), which lowers enzymatic activity, was associated with acrofacial vitiligo in a Turkis h population (Tursen et al., 2002). Another group found this same SNP to be positively associated with an increase ri sk for generalized vitiligo in a Chinese population, but not with segmental or ac rofacial vitiligo (Li et al., 2008). Because COMT enzyme may play a dual role in the me lanocyte, both in the regulation of melanin biosynthesis and in limiting damage fr om reactive oxygen sp ecies (ROS), the COMT gene remains a possible vitiligo susceptibility gene. The results from my study and the work of others have shown SNPs in COMT to be associated with vitiligo in three different populations, adding support to this genes cont ribution to di sease etiology.

PAGE 70

70 Figure 4-1. Schematic representation of gene structure, SNP location, and HapMap CEPH haploblocks. Exons are represented by black boxes, introns are represented by white boxes, arrows show SNP location, and haploblocks are shown by shaded boxes. Note: haploblocks extend off of the e nds of genes. A) Dopachrome tautamerase (DCT ), 40,083base pairs (bp). B) Tyrosinase-related protein 1 ( TYPR1 ), 16,802bp. B A

PAGE 71

Table 4-1. List of Applied Biosyste ms SNP Assay IDs and description of DCT, PAH, TYR, TYRP1, and COMT SNPs genotyped Gene(s) SNP rs number Applied Biosystems ID Base pair change & location (amino acid change & location) DCT rs9516413 C___1872510_10 C/T 3 UTR DCT rs7987802 C___1872504_10 C/T intron 6 DCT rs4318084 C__26557934_10 C/T intron 6 DCT rs9524493 C__29504474_10 C/G intron 6 DCT rs9516418 C__26557938_10 C/T intron 6 DCT rs1028805 C___7593146_10 G/T intron 2 DCT rs11618471 C___1872491_10 A/G intron 2 DCT rs7991232 C__30496225_10 A/G intron 1 PAH rs1722381 C___1402626_10 A/G intron 6 PAH rs1522307 C___1402699_10 A/G intron 2 TYR rs1042602 C___8362862_10 C/A exon 1, (S192Y) TYR rs621313 C___1054193_10 G/A intron 1 TYR rs594647 C___1054191_10 T/C intron 1 TYR rs10765197 C__31959995_10 A/C intron 2 TYR rs12791412 C__31959989_10 A/G intron 2 TYR rs2000554 C__11665932_10 G/A intron 2 TYR rs10830250 C__31959927_10 C/G intron 3 TYR rs1827430 C__27097966_20 G/A intron 4 TYRP1 rs2762462 C__15931132_10 C/T intron 4 TYRP1 rs2733831 C___3119212_10 A/G intron 5 TYRP1 rs2733833 C___3119209_10 G/T intron 6 TYRP1 rs683 C___3119206_10 A/C 3 UTR TYRP1 rs2762464 C__15931130_10 A/T 3 UTR TYRP1 rs1063380 C___3119204_10 C/T 3 UTR COMT rs2020917 C__11731880_1_ T/C 5 UTR COMT rs6269 C___2538746_1_ G/A intron 2 71

PAGE 72

Table 4-2. Linkage disequilibrium (D ) values for DCT PAH TYR TYRP1 and COMT Gene SNP ID SNP ID D DCT rs9516418 rs9516413 0.02 DCT rs9524493 rs4318084 0.25 DCT rs9516413 rs1028805 0.38 DCT rs11618471 rs7987802 1.00 DCT rs11618471 rs9516413 1.00 DCT rs7987802 rs4318084 1.00 DCT rs7991232 rs9516413 0.59 DCT rs9524493 rs1028805 0.41 DCT rs4318084 rs9516413 1.00 DCT rs7987802 rs1028805 0.97 DCT rs7987802 rs9516418 0.90 DCT rs7991232 rs7987802 0.94 DCT rs11618471 rs7991232 0.97 DCT rs9524493 rs9516413 0.79 DCT rs11618471 rs9524493 0.98 DCT rs9524493 rs7987802 1.00 DCT rs4318084 rs9516418 0.54 DCT rs7987802 rs9516413 0.61 DCT rs7991232 rs4318084 0.91 DCT rs7991232 rs1028805 0.88 DCT rs11618471 rs9516418 1.00 DCT rs7991232 rs9516418 0.83 DCT rs9516418 rs1028805 0.62 DCT rs11618471 rs1028805 1.00 DCT rs11618471 rs4318084 0.97 DCT rs9524493 rs9516418 0.95 DCT rs4318084 rs1028805 0.99 DCT rs7991232 rs9524493 0.88 PAH rs1722381 rs1522307 0.48 TYR rs1042602 rs12791412 0.95 TYR rs1042602 rs1827430 0.87 TYR rs1042602 rs2000554 0.85 TYR rs1042602 rs10830250 0.96 TYR rs10765197 rs1042602 0.99 TYR rs12791412 rs621313 0.93 TYR rs1042602 rs594647 0.94 TYR rs12791412 rs594647 0.93 TYR rs12791412 rs2000554 0.95 TYR rs12791412 rs1827430 0.88 72

PAGE 73

73 Table 4-2. Continued Gene SNP ID SNP ID D TYR rs12791412 rs10830250 0.92 TYR rs1042602 rs621313 0.94 TYR rs1827430 rs621313 0.88 TYR rs10765197 rs12791412 0.99 TYR rs1827430 rs594647 0.87 TYR rs10765197 rs1827430 0.85 TYR rs2000554 rs1827430 0.86 TYR rs10830250 rs1827430 0.85 TYR rs10830250 rs621313 0.96 TYR rs10765197 rs621313 0.96 TYR rs2000554 rs621313 0.94 TYR rs10830250 rs594647 0.96 TYR rs2000554 rs594647 0.93 TYR rs10765197 rs594647 0.97 TYR rs10830250 rs2000554 0.96 TYR rs10765197 rs10830250 0.98 TYR rs10765197 rs2000554 0.98 TYR rs621313 rs594647 1.00 TYRP1 rs2733831 rs2762462 0.98 TYRP1 rs2762464 rs2762462 1.00 TYRP1 rs683 rs2762462 1.00 TYRP1 rs2762462 rs2733833 1.00 TYRP1 rs1063380 rs2762462 1.00 TYRP1 rs2733831 rs1063380 0.91 TYRP1 rs2733831 rs2762464 0.88 TYRP1 rs683 rs2733831 0.91 TYRP1 rs2733831 rs2733833 0.96 TYRP1 rs1063380 rs2733833 0.92 TYRP1 rs683 rs1063380 0.98 TYRP1 rs2762464 rs1063380 1.00 TYRP1 rs2762464 rs2733833 0.98 TYRP1 rs683 rs2733833 0.99 TYRP1 rs683 rs2762464 0.99 COMT rs2020917 rs6269 0.79

PAGE 74

74Table 4-3. Minor allele frequencies (MAF) for DCT PAH TYR TYRP1, and COMT SNPs from the HapMap project Caucasians from European Ancestry (HapMap CEPH), Applied Biosystems Cau casian cohort (ABI Caucasian), and values for the white case and control groups from the present study; 2 p-values and q-values for allelic comparisons are given Gene Symbol SNP ID HapMap CEPH MAF ABI Caucasian MAF White case MAF # of white cases White control MAF # of white controls p-value BonC p-value FDR q-value DCT rs9516413 0.21 (C) 0.26 (C) 207 0.18 (C) 241 0.0047* 0.14 0.07 DCT rs7987802 0.19 (T) 0.20 (T) 209 0.18 (T) 244 0.48 1.00 0.92 DCT rs4318084 0.26 (T) 0.24 (T) 209 0.28 (T) 238 0.26 1.00 0.78 DCT rs9524493 0.45 (G) 0.42 (G) 207 0.46 (G) 238 0.32 1.00 0.78 DCT rs9516418 0.41 (C) 0.36 (C) 207 0.36 (C) 238 0.88 1.00 0.93 DCT rs1028805 0.33 (T) 0.32 (T) 208 0.32 (T) 237 0.86 1.00 0.93 DCT rs11618471 0.16 (G) 0.17 (G) 210 0.20 (G) 242 0.23 1.00 0.78 DCT rs7991232 0.38 (A) 0.36 (A) 207 0.39 (A) 235 0.33 1.00 0.78 PAH rs1722381 0.35 (G) 0.43 (G) 0.43 (G) 164 0.38 (G) 240 0.19 1.00 0.78 PAH rs1522307 0.36 (G) 0.36 (G) 0.34 (G) 164 0.37 (G) 240 0.38 1.00 0.81 TYR rs1042602 0.42 (A) 0.35 (A) 208 0.38 (A) 237 0.49 1.00 0.92 TYR rs621313 0.42 (G) 0.44 (G) 207 0.45 (G) 238 0.81 1.00 0.93 TYR rs594647 0.34 (T) 0.40 (T) 207 0.41 (T) 239 0.67 1.00 0.93 TYR rs10765197 0.31 (C) 0.35 (C) 209 0.38 (C) 240 0.52 1.00 0.92 TYR rs12791412 0.20 (G) 0.23 (G) 209 0.29 (G) 239 0.05* 1.00 0.50 TYR rs2000554 0.33 (G) 0.35 (G) 206 0.38 (G) 239 0.34 1.00 0.78 TYR rs10830250 0.27 (G) 0.33 (G) 208 0.37 (G) 239 0.25 1.00 0.78 TYR rs1827430 0.33 (G) 0.30 (G) 207 0.36 (G) 238 0.09 1.00 0.65 TYRP1 rs2762462 0.23 (T) 0.29 (T) 205 0.28 (T) 241 0.81 1.00 0.93 TYRP1 rs2733831 0.35 (A) 0.43 (A) 0.44 (A) 200 0.43 (A) 242 0.72 1.00 0.93 TYRP1 rs2733833 0.29 (T) 0.36 (T) 0.37 (T) 205 0.36 (T) 238 0.84 1.00 0.93 TYRP1 rs683 0.31 (C) 0.39 (C) 206 0.37 (C) 238 0.58 1.00 0.92 TYRP1 rs2762464 0.3 (A) 0.39 (A) 204 0.38 (A) 242 0.72 1.00 0.93 TYRP1 rs1063380 0.28 (T) 0.37 (T) 207 0.35 (T) 238 0.57 1.00 0.92 COMT rs2020917 0.32 (T) 0.37 (T) 0.31 (T) 164 0.26 (T) 234 0.12 1.00 0.65 COMT rs6269 0.49 (G) 0.41 (G) 165 0.41 (G) 237 0.90 1.00 0.93 denotes a significant finding; BonC denot es the Bonferroni corrected p-value

PAGE 75

Table 4-4. Case-control results for genotype analysis of DCT PAH TYR TYRP1, and COMT Gene SNP p-value BonC p-value FDR q-value DCT rs9516413 0.0270* 0.81 0.27 DCT rs7987802 0.57 1.00 0.63 DCT rs4318084 0.13 1.00 0.39 DCT rs9524493 0.55 1.00 0.63 DCT rs9516418 0.98 1.00 0.98 DCT rs1028805 0.87 1.00 0.93 DCT rs11618471 0.35 1.00 0.55 DCT rs7991232 0.46 1.00 0.55 PAH rs1722381 0.44 1.00 0.55 PAH rs1522307 0.39 1.00 0.55 TYR rs1042602 0.31 1.00 0.55 TYR rs621313 0.45 1.00 0.55 TYR rs594647 0.34 1.00 0.55 TYR rs10765197 0.32 1.00 0.55 TYR rs12791412 0.13 1.00 0.39 TYR rs2000554 0.24 1.00 0.55 TYR rs10830250 0.31 1.00 0.55 TYR rs1827430 0.12 1.00 0.39 TYRP1 rs2762462 0.12 1.00 0.39 TYRP1 rs2733831 0.41 1.00 0.55 TYRP1 rs2733833 0.0254* 0.76 0.27 TYRP1 rs683 0.11 1.00 0.39 TYRP1 rs2762464 0.06 1.00 0.39 TYRP1 rs1063380 0.10 1.00 0.39 COMT rs2020917 0.18 1.00 0.49 COMT rs6269 0.39 1.00 0.55 *denotes a significant finding; BonC denot es the Bonferroni corrected p-value 75

PAGE 76

Table 4-5. Family-based association (tra nsmission disequilibrium test) results for DCT, PAH TYR, TYRP1 and COMT Gene SNP Number of informative parents Transmitted (allele) Not transmitted (allele) p-value BonC p-value FDR q-value DCT rs9516413 62 42 (C) 20 (T) 0.0052* 0.16 0.039* DCT rs7987802 59 29 (C) 30 (T) 0.90 1.00 0.96 DCT rs4318084 67 52 (C) 15 (T) 6.20E-06* 0.0002* <0.0001* DCT rs9524493 74 34 (C) 40 (G) 0.49 1.00 0.81 DCT rs9516418 79 28 (C) 51 (T) 0.0097* 0.29 0.06 DCT rs1028805 74 50 (G) 24 (T) 0.0025* 0.08 0.025* DCT rs11618471 54 44 (A) 10 (G) 3.70E-06 0.0001* <0.0001* DCT rs7991232 70 36 (A) 34 (G) 0.81 1.00 0.93 PAH rs1722381 43 23 (A) 20 (G) 0.65 1.00 0.90 PAH rs1522307 54 28 (A) 26 (G) 0.79 1.00 0.93 TYR rs1042602 77 34 (A) 43 (C) 0.31 1.00 0.62 TYR rs621313 82 38 (G) 44 (A) 0.51 1.00 0.80 TYR rs594647 75 33 (T) 42 (C) 0.30 1.00 0.62 TYR rs10765197 71 42 (A) 29 (A) 0.12 1.00 0.45 TYR rs12791412 62 40 (A) 22 (G) 0.02* 0.60 0.10 TYR rs2000554 73 32 (G) 41 (A) 0.29 1.00 0.62 TYR rs10830250 80 45 (C) 35 (G) 0.26 1.00 0.62 TYR rs1827430 77 33 (G) 45 (A) 0.17 1.00 0.46 TYRP1 rs2762462 60 30 (C) 30 (T) 1.00 1.00 1.00 TYRP1 rs2733831 70 36 (A) 34 (G) 0.81 1.00 0.93 TYRP1 rs2733833 58 32 (G) 26 (T) 0.43 1.00 0.76 TYRP1 rs683 68 35 (A) 33 (C) 0.81 1.00 0.93 TYRP1 rs2762464 67 34 (A) 33 (T) 0.90 1.00 0.96 TYRP1 rs1063380 67 30 (C) 37 (T) 0.39 1.00 0.73 COMT rs2020917 47 22 (T) 25 (C) 0.66 1.00 0.90 COMT rs6269 50 20 (G) 30 (A) 0.16 1.00 0.46 denotes a significant finding; BonC denot es the Bonferroni corrected p-value Table 4-6. Pairwise haplotype analysis results for PAH rs1722381 and rs1522307 Frequency Controls Vitiligo patients p-values 0 [GG] 215 (89%) 165 (81%) 1 [GG] 26 (11%) 27 (13.5%) 2 [GG] 0 (0%) 11 (5.5%) Total (n) 241 203 2 0.0007* Odds ratio for 1 [GG] vs. 0 [GG] = n.s. n.s Odds ratio for 2 [GG] vs. 0 [GG] = n.d. n.d. Odds ratio for 2 [GG] vs. 1 [GG] = n.d. n.d. n.s. = not significant; n.d. = not defined; denot es a significant finding; number in frequency column indicates the number of copies of this haplotype 76

PAGE 77

Table 4-7. Pairwise haplotype analysis results for DCT rs4318084 and rs9516413 Frequency Controls Vitiligo patients p-values 0 [CC] 165 (67.35%) 118 (55.7%) 1 [CC] 67 (27.35%) 71 (33.5%) 2 [CC] 13 (5.3%) 23 (10.8%) Total (n) 245 212 2 0.0153* Odds ratio for 1 [CC] vs. 0 [CC] = 1.5116 95% CI=1.0016 to 2.2814 0.0491* Odds ratio for 2 [CC] vs. 0 [CC] = 2.4309 95% CI=1.1817 to 5.0005 0.0158* Odds ratio for 2 [CC] vs. 1 [CC] = n.s. n.s. n.s. = not significant; denotes a significant finding; number in frequency column indicates the number of copies of this haplotype Table 4-8. Pairwise haplotype analysis results for DCT rs11618471 and rs9516413 Frequency Controls Vitiligo patients p-values 0 [AC] 165 (67.6%) 118 (56%) 1 [AC] 67 (27.5%) 71 (34%) 2 [AC] 12 (4.9%) 21 (10%) Total (n) 244 210 2 0.0195* Odds ratio for 1 [AC] vs. 0 [AC] = 1.5116 95% CI=1.0015 to 2.2813 0.0491* Odds ratio for 2 [AC] vs. 0 [AC] = 2.4152 95% CI=1.1425 to 5.1058 0.0210* Odds ratio for 2 [CC] vs. 1 [CC] = n.s. n.s. n.s. = not significant; denotes a significant finding; number in frequency column indicates the number of copies of this haplotype Table 4-9. Pairwise haplotype analysis results for DCT rs7991232 and rs9516413 Frequency Controls Vitiligo patients p-values 0 [GT] 176 (72%) 127 (60%) 1 [GT] 61 (25%) 70 (33%) 2 [GT] 7 (3%) 14 (7%) Total (n) 244 211 2 0.0129* Odds ratio for 1 [GT] vs. 0 [GT] = 1.6164 95% CI=1.0714 to 2.4476 0.0222* Odds ratio for 2 [GT] vs. 0 [GT] = 2.7193 95% CI=1.0652 to 6.9419 0.0364* Odds ratio for 2 [CC] vs. 1 [CC] = n.s. n.s. n.s. = not significant; denotes a significant finding; number in frequency column indicates the number of copies of this haplotype 77

PAGE 78

Table 4-10. Pairwise haplot ype analysis results for DCT rs9524493 and rs9516413 Frequency Controls Vitiligo patients p-values 0 [CC] 168 (69%) 123 (58%) 1 [CC] 68 (28%) 73 (35%) 2 [CC] 8 (3%) 15 (7%) Total (n) 244 211 2 0.0310* Odds ratio for 1 [CC] vs. 0 [CC] = n.s. n.s Odds ratio for 2 [CC] vs. 0 [CC] = 2.5442 95% CI=1.0652 to 6.9419 0.0364* Odds ratio for 2 [CC] vs. 1 [CC] = n.s. n.s. n.s. = not significant; denotes a significant finding; number in frequency column indicates the number of copies of this haplotype Table 4-11. Pairwise haplot ype analysis results for DCT rs7987802 and rs9516413 Frequency Controls Vitiligo patients p-values 0 [CC] 212 (86%) 167 (79%) 1 [CC] 31 (13%) 36 (17%) 2 [CC] 2 (1%) 9 (4%) Total (n) 245 212 2 0.0206* Odds ratio for 1 [CC] vs. 0 [CC] = n.s. n.s Odds ratio for 2 [CC] vs. 0 [CC] = 5.5370 95% CI=1.1784 to 26.0168 0.0302* Odds ratio for 2 [CC] vs. 1 [CC] = n.s. n.s. n.s. = not significant; denotes a significant finding; number in frequency column indicates the number of copies of this haplotype Table 4-12. Significant 2 findings for DCT association for patients with co-morbid autoimmune diseases Gene SNP No co-morbid disease co-morbid disease p-value DCT rs7987802 CC 100 (68%) CT 43 (29%) TT 4 (3%) CC 37 (60%) CT 18 (30%) TT 7 (11%) 0.038* DCT rs4318084 CC 84 (57%) CT 56 (38%) TT 8 (5%) CC 39 (64%) CT 14 (23%) TT 8 (13%) 0.038* DCT rs9524493 CC 41 (28%) CG 77 (52%) GG 29 (20%) CC 28 (47%) CG 24 (40%) GG 8 (13%) 0.033* denotes a significant finding 78

PAGE 79

Table 4-13. Pairwise haplot ype analysis results for TYR rs10765197 and TYR rs1042602 Frequency Controls Vitiligo patients p-values 0 [AC] 137 (56%) 110 (52%) 1 [AC] 98 (40%) 79 (37%) 2 [AC] 10 (4%) 23 (11%) Total (n) 245 212 2 0.0205* Odds ratio for 1 [AC] vs. 0 [AC] = n.s. n.s. Odds ratio for 2 [AC] vs. 0 [AC] = 2.8691 95% CI=1.3095 to 6.2864 0.0084* Odds ratio for 2 [AC] vs. 1 [AC] = 2.8468 95% CI=1.2793 to 6.3352 0.0104* n.s. = not significant; denotes a significant finding; number in frequency column indicates the number of copies of this haplotype Table 4-14. Pairwise haplot ype analysis results for TYR rs1042602 and TYR rs12791412 Frequency Controls Vitiligo patients p-values 0 [CA] 105 (43%) 80 (38%) 1 [CA] 117 (48%) 89 (42%) 2 [CA] 23 (9%) 43 (20%) Total (n) 245 212 2 0.0043* Odds ratio for 1 [CA] vs. 0 [CA] = n.s. n.s. Odds ratio for 2 [CA] vs. 0 [CA] = 2.4708 95% CI=1.3767 to 4.4346 0.0024* Odds ratio for 2 [CA] vs. 1 [CA] = 2.4761 95% CI=1.3900 to 4.4111 0.0021* n.s. = not significant; denotes a significant finding; number in frequency column indicates the number of copies of this haplotype Table 4-15. Pairwise haplot ype analysis results for TYR rs1042602 and TYR rs10830250 Frequency Controls Vitiligo patients p-values 0 [CA] 130 (53%) 106 (50%) 1 [CA] 103 (42%) 80 (38%) 2 [CA] 12 (5%) 26 (12%) Total (n) 245 212 2 0.0169* Odds ratio for 1 [CA] vs. 0 [CA] = n.s. n.s. Odds ratio for 2 [CA] vs. 0 [CA] = 2.6846 95% CI=1.2916 to 5.5801 0.0082* Odds ratio for 2 [CA] vs. 1 [CA] = 2.8385 95% CI=1.3467 to 5.9829 0.0061* n.s. = not significant; denotes a significant finding; number in frequency column indicates the number of copies of this haplotype 79

PAGE 80

Table 4-16. Pairwise haplotype analys is results for TYR rs1042602 and TYR rs2000554 Frequency Controls Vitiligo patients p-values 0 [CA] 133 (54%) 109 (51%) 1 [CA] 102 (42%) 80 (38%) 2 [CA] 10 (4%) 23 (11%) Total (n) 245 212 2 0.0203* Odds ratio for 1 [CA] vs. 0 [CA] = n.s. n.s. Odds ratio for 2 [CA] vs. 0 [CA] = 2.8065 95% CI=1.2801 to 6.1531 0.0100* Odds ratio for 2 [CA] vs. 1 [CA] = 2.8924 95% CI=1.3015 to 6.4278 0.0091* n.s. = not significant; denotes a significant finding; number in frequency column indicates the number of copies of this haplotype Table 4-17. Pairwise haplot ype analysis results for TYR rs1042602 and TYR rs1827430 Frequency Controls Vitiligo patients p-values 0 [CA] 125 (51%) 98 (46%) 1 [CA] 101 (41%) 81 (38%) 2 [CA] 19 (8%) 33 (16%) Total (n) 245 212 2 0.0319* Odds ratio for 1 [CA] vs. 0 [CA] = n.s. n.s. Odds ratio for 2 [CA] vs. 0 [CA] = 2.2267 95% CI=1.1931 to 4.1557 0.0119* Odds ratio for 2 [CA] vs. 1 [CA] = 2.1732 95% CI=1.1498 to 4.1073 0.0169* n.s. = not significant; denotes a significant finding; number in frequency column indicates the number of copies of this haplotype Table 4-18. Pairwise haplot ype analysis results for TYR rs12791412 and TYR rs1827430 Frequency Controls Vitiligo patients p-values 0 [GA] 130 (53%) 134 (63%) 1 [GA] 97 (40%) 71 (34%) 2 [GA] 18 (7%) 7 (3%) Total (n) 245 212 2 0.0395* Odds ratio for 1 [GA] vs. 0 [GA] = n.s. n.s. Odds ratio for 2 [GA] vs. 0 [GA] = 0.3761 95% CI=0.1519 to 0.9314 0.0119* Odds ratio for 2 [GA] vs. 1 [GA] = n.s. n.s. n.s. = not significant; denotes a significant finding; number in frequency column indicates the number of copies of this haplotype 80

PAGE 81

Table 4-19. Significant Hardy-Weinbe rg equilibrium analysis for TYRP1 Gene SNP p-value cases p-value controls TYRP1 rs2762462 0.022* 0.66 TYRP1 rs2733833 0.011* 0.23 TYRP1 rs683 0.024* 0.61 TYRP1 rs2762464 0.019* 0.35 denotes a significant finding Table 4-20. Pairwise haplot ype analysis results for TYRP1 rs683 and TYRP1 rs2733831 Frequency Controls Vitiligo patients p-values 0 [CA] 107 (43.5%) 80 (37.6%) 1 [CA] 104 (42.5%) 115 (54.4%) 2 [CA] 34 (14%) 17 (8%) Total (n) 245 212 2 0.0203* Odds ratio for 1 [CA] vs. 0 [CA] = 1.4950 95% CI=1.0083 to 2.2168 0.0454* Odds ratio for 2 [CA] vs. 0 [CA] = n.s. n.s. Odds ratio for 2 [CA] vs. 1 [CA] = 0.4572 95% CI=0.2409 to 0.8674 0.0166* n.s. = not significant; denotes a significant finding; number in frequency column indicates the number of copies of this haplotype Table 4-21. Pairwise haplot ype analysis results for TYRP1 rs683 and TYRP1 rs2733833 Frequency Controls Vitiligo patients p-values 0 [CT] 104 (42.5%) 79 (37%) 1 [CT] 107 (43.5%) 114 (54%) 2 [CT] 34 (14%) 18 (9%) Total (n) 245 212 2 0.0434* Odds ratio for 1 [CT] vs. 0 [CT] = n.s. n.s. Odds ratio for 2 [CT] vs. 0 [CT] = n.s. n.s. Odds ratio for 2 [CT] vs. 1 [CT] = 0.4979 95% CI=0.2652 to 0.9349 0.0300* n.s. = not significant; denotes a significant finding; number in frequency column indicates the number of copies of this haplotype 81

PAGE 82

Table 4-22. Pairwise haplot ype analysis results for TYRP1 rs2733831and TYRP1 rs2762464 Frequency Controls Vitiligo patients p-values 0 [AA] 104 (42.5%) 76 (36%) 1 [AA] 104 (42.5%) 114 (54%) 2 [AA] 37 (15%) 22 (10%) Total (n) 245 212 2 0.0434* Odds ratio for 1 [AA] vs. 0 [AA] = 1.5171 95% CI=1.0181 to 2.2607 0.0405* Odds ratio for 2 [AA] vs. 0 [AA] = n.s. n.s. Odds ratio for 2 [AA] vs. 1 [AA] = 0.5475 95% CI=0.3030 to 0.9893 0.0460* n.s. = not significant; denotes a significant finding; number in frequency column indicates the number of copies of this haplotype Table 4-23. Pairwise haplot ype analysis results for TYRP1 rs2733831 and TYRP1 rs2733833 Frequency Controls Vitiligo patients p-values 0 [AT] 106 (43%) 80 (38%) 1 [AT] 103 (42%) 114 (54%) 2 [AT] 35 (14%) 18 (8%) Total (n) 245 212 2 0.0275* Odds ratio for 1 [AT] vs. 0 [AT] = n.s. n.s. Odds ratio for 2 [AT] vs. 0 [AT] = n.s. n.s. Odds ratio for 2 [AT] vs. 1 [AT] = 0.4763 95% CI=0.2539 to 0.8933 0.0208* n.s. = not significant; denotes a significant finding; number in frequency column indicates the number of copies of this haplotype Table 4-24. Pairwise haplot ype analysis results for TYRP1 rs2762464 and TYRP1 rs2733833 Frequency Controls Vitiligo patients p-values 0 [AT] 105 (43%) 78 (37%) 1 [AT] 104 (42%) 115 (54%) 2 [AT] 36 (15%) 19 (9%) Total (n) 245 212 2 0.0242* Odds ratio for 1 [AT] vs. 0 [AT] = 1.4985 95% CI=1.0087 to 2.2263 0.0452* Odds ratio for 2 [AT] vs. 0 [AT] = n.s. n.s. Odds ratio for 2 [AT] vs. 1 [AT] = 0.4849 95% CI=0.2616 to 0.8988 0.0215* n.s. = not significant; denotes a significant finding; number in frequency column indicates the number of copies of this haplotype 82

PAGE 83

Table 4-25. Pairwise haplot ype analysis results for TYRP1 rs2762462 and TYRP1 rs2733833 Frequency Controls Vitiligo patients p-values 0 [CG] 36 (15%) 20 (9.4%) 1 [CG] 105 (43%) 115 (54.3%) 2 [CG] 104 (42%) 77 (36.3%) Total (n) 245 212 2 0.0350* Odds ratio for 1 [CG] vs. 0 [CG] = 1.9425 95% CI=1.0570 to 3.5698 0.0325* Odds ratio for 2 [CG] vs. 0 [CG] = n.s. n.s. Odds ratio for 2 [CG] vs. 1 [CG] = 0.6720 95% CI=0.4520 to 0.9991 0.0495* n.s. = not significant; denotes a significant finding; number in frequency column indicates the number of copies of this haplotype Table 4-26. Pairwise haplot ype analysis results for COMT rs2020917 and COMT rs6269 Frequency Controls Vitiligo patients p-values 0 [CG] 162 (66%) 169 (80%) 1 [CG] 76 (31%) 38 (18%) 2 [CG] 7 (3%) 5 (2%) Total (n) 245 212 2 0.0046* Odds ratio for 1 [CG] vs. 0 [CG] = 0.4750 95% CI=0.3040 to 0.7423 0.0011* Odds ratio for 2 [CG] vs. 0 [CG] = n.s. n.s. Odds ratio for 2 [CG] vs. 1 [CG] = n.s n.s. n.s. = not significant; denotes a significant finding; number in frequency column indicates the number of copies of this haplotype 83

PAGE 84

Table 4-27. Summary of all findings for DCT PAH TYR TYRP1, and COMT Gene SNP ID Base pair change & location Case/ Control TDT HWE Comorbid AAD Age of onset Haplotype DCT rs9516413 C/T 3 UTR A, G + 5 DCT rs7987802 C/T intron 6 + 1 DCT rs4318084 C/T intron 6 + + 1 DCT rs9524493 C/G intron 6 + 1 DCT rs9516418 C/T intron 6 + DCT rs1028805 G/T intron 2 + DCT rs11618471 A/G intron 2 + 1 DCT rs7991232 A/G intron 1 1 PAH rs1722381 A/G intron 6 1 PAH rs1522307 A/G intron 2 1 TYR rs1042602 C/A exon 1, (S192Y) 5 TYR rs621313 G/A intron 1 TYR rs594647 T/C intron 1 TYR rs10765197 A/C intron 2 1 TYR rs12791412 A/G intron 2 A + + 2 TYR rs2000554 G/A intron 2 1 TYR rs10830250 C/G intron 3 1 TYR rs1827430 G/A intron 4 2 TYRP1 rs2762462 C/T intron 4 D 1 TYRP1 rs2733831 A/G intron 5 3 TYRP1 rs2733833 G/T intron 6 G D 4 TYRP1 rs683 A/C 3 UTR D 2 TYRP1 rs2762464 A/T 3 UTR D 2 TYRP1 rs1063380 C/T 3 UTR COMT rs2020917 T/C 5 UTR 1 COMT rs6269 G/A intron 2 1 A, significant findings on the al lelic level; G, significant fi ndings on the genotypic level; AAD, autoimmune disease; +, significant result; D, in Hardy-Weinberg disequilibrium; Pairwise haplotype, number of significant pairwise associations observed 84

PAGE 85

CHAPTER 5 CONCLUSIONS AND FUTURE DIRECTIONS This study examined the association of 30 SNPs across eight genes in relation to vitiligo susceptibility. There are two popular theories underlying the pathogenesis of vitiligo: autoimmunity and autocytotoxicity. The gene s in my study were chos en because of their possible involvement in one or both of these mech anisms, which are hypothesized to lead to the melanocyte inactivation and/or dest ruction observed in vitiligo. It is plausible that both the autoimmune and autocytotoxic hypotheses contribut e to vitiligo pathogenesis. Different patient subpopulations may have a more significant cont ribution from one or the other mechanism, which may lead to a dirty phenotype and confound association studies such as ours. The role of environmental triggers contributing to vitiligo etiology ha s been well documented. The interplay of genetics and environm ent, although difficult to elucidate, is likely to be an important component of the disease process. My study provides strong evidence for genetic association with AIRE and suggests possible genetic association with COMT, PAH DCT TYR and/or TYRP1 If these associations are confirmed in other patient sets, functional stud ies of vitiligo susceptib ility genes may lead to the design of novel strategies to prevent and/or treat vitiligo, other autoimmune diseases, and melanoma. FBXO11 and MSH6 No evidence of association to vitiligo was found for FBXO11 and MSH6 This finding may be a true negative if these genes do not pl ay a significant role in vitiligo pathogenesis. However, if these genes play a more subtle role in vitiligo etiology or if the markers I chose were not closely linked to informative or causative ch anges on these genes, future studies may reveal that FBXO11 and MSH6 do play a role in vitiligo susceptibility. This suggests that aberrant 85

PAGE 86

DNA repair (at least via MSH6 ) is not a strong genetic componen t of vitiligo susceptibility. However, FBXO11 was observed to be downr egulated in vitiligo patients compared to healthy controls while MSH6 was upregulated (Le Poole et al., 1994). Because these two genes overlap on their 3 ends, it is possible that do uble-stranded RNA intermediates are interfering in protein expression via the RNAi mechanism. Alternatively, another mechanism that is yet to be elucidated may be underlying this altered expression pattern. Autoimmune Regulator ( AIRE) My study demonstrated significan t genetic association of vi tiligo with the autoimmune regulator gene, or AIRE There is a substantial body of research supporting immune system dysfunction in vitiligo patients. Because the AIRE gene is vital to the negative selection of autoreactive T cells, it is possible that subtle defects in AIRE allow for the escape of autoreactive T cells from the thymus, contributing to the autoimmunity hypothesized to cause vitiligo. I observed a significant associati on between vitiligo and a SNP cau sing an amino acid change in the SAND DNA-binding domain of AIRE, even when the stringent Bonferroni correction was applied. Pairwise haplotype analysis with th is non-synonymous SNP and an intronic one also yielded a significant finding. However, the family-based association studies were not significant. This is not unusua l in association studies (especially given our limited number of families), so it does not negate the case-contr ol result. Additional studies, using protein modeling and/or functional analysis are needed to help understand th e significance of this serine to arginine amino acid change and its influence on the SAND domain. For example, it is not known whether serine is a site for phosphorylation, which could significantly affect protein function. It is possible that the non-synonymous SNP I examined is directly related to vitiligo genetic susceptibility. If not causa tive, this SNP and the significant AIRE haplotype I observed may be linked to a polymorphism that reduces normal immune function in vitiligo patients. 86

PAGE 87

Sequencing AIRE in vitiligo patients may yield additional information about linked variants and this genes role in vitiligo e tiology. RNA analysis could reve al whether the susceptibility haplotype has altered mRNA levels. Melanin Biosynthesis Genes I tested five genes involved in melanin bi osynthesis to examine their role in vitiligo pathogenesis: tyrosinase ( TYR), tyrosinase-related protein 1 ( TYRP1), dopachrome-tautomerase ( DCT ), phenylalanine hydroxylase ( PAH), and catechol-O-methyltransferase ( COMT) Three of these genes ( DCT TYR and TYRP1 ) code for proteins that have b een observed to be the targets of autoantibodies in the sera of vitiligo patients. There may be a genetic component underlying this observed autoimmunity, as seen in other autoimmune diseases like type I di abetes (Redondo et al., 2001). A change in the DCT, TYR, and/or TYRP1 genes that results in an altered protein may create an epitope, or reveal a cryptic epitope by affecting protein structure, which may lead to inappropriate targeting of me lanocytes by the immune system. Reactive oxygen species (ROS) are generated an d carefully regulated during each step of melanin biosynthesis by proteins such as catalas e and catechol-O-methyltransferase. A change in enzymatic function during melanogenesis may cause an imbalance in the melanin pigmentation cycle, leading to a build-up of toxic in termediates that would cause free radical damage to the melanocyte. It is wi th this in mind that I tested the PAH DCT TYR and TYRP1 genes because they play a vital role in melanin production. Furthermore, I tested COMT because its protein acts to prevent toxic build-up of intermediate o-quinones in the melanocyte during melanogenesis, and it may be involved in th e regulation of melanin biosynthesis. The only significant results observed in the PAH gene were in the haplotype analysis of two intronic SNPs. The risk haplotype I observe d was found exclusively in vitiligo cases but not in controls. This risk haplotype may be indicate splice site errors in the introns that I tested or it 87

PAGE 88

may be in linkage disequilibrium with another mutation within the structure of the gene that could alter the PAH enzyme. This finding lost significance with the mu ltiple-tests correction, but given the relatively small samples sizes, th ese results are promising enough to pursue. Further studies are needed to validate and then comprehend exactly what genetic change is leading to this finding in vitiligo. PAH is re sponsible for the conversion of phenylalanine to tyrosine, the beginning substrate for melanogene sis. Therefore, subtle changes in the PAH gene may moderately alter the enzymatic activity a nd may contribute to viti ligo susceptibility. Vitiligo patients have been described as having d ecreased levels of epidermal PAH (Schallreuter et al., 1998; Schallreuter and Wood, 1999; Schallreuter et al., 2005) My significant haplotype finding supports the possi bility that altered PAH function may have an effect on the vitiligo disease process. Dopachrome-tautomerase has been observed to be a vitiligo autoantigen. Its exact role in the melanogenesis is yet to be elucidated, but it is believed to act in the eumelanin pathway and may function as part of a multienzyme complex. In my within-case analysis looking at vitiligo patients with and without co-morbid autoimmune diseases, the only significant findings were with three SNPs in intron six of DCT Because vitiligo patients with and without comorbid autoimmune diseases may represent different su b-groups of the disease, it is possible that changes in DCT are informative with these sub-groups. Additionally, several pairwise haplotypes involving a gene in the 3 UTR were positively associated with vitiligo susceptibility. The significance of these tests was lost upon multiple testing corrections, although there is debate in the field that in sma ll association studies, findings such as mine should be considered promising. The exact mechanisms of these associat ions are yet to be eluc idated. It is possible that epitopes on the DCT enzyme are similar to other self-antigens in patients with other 88

PAGE 89

autoimmune diseases, though this hypothesis has not yet been tested. Alternatively, these associations may help uncover pathogenic changes in DCT that are yet to be determined. I genotyped eight SNPs in the tyrosinase ge ne to investigate both the autoimmune and autocytotoxic hypotheses underlyi ng vitiligo pathogenesis, an d my study yielded several interesting results. One intronic SNP in the TYR was my only (uncorrected) significant finding in my age of disease onset analysis. It is possibl e that patients with a later age of onset represent a different sub-group of vitiligo, however, because my finding was associated with homozygosity for the minor allele for an intron tw o SNP, this result must be interpreted with caution. This same intron two SNP was also signif icant at the allelic but not the genotypic level for my entire white case-control analysis. It is possible that the SNP itself causes an RNA instability or leads to splicing problems, or this SNP may be a marker for another genetic changed linked to vitiligo etiology. Pairwise haplotype analysis for TYR showed six significant associ ations with vitiligo, and five of these involved a non-synonym ous SNP. It is not known how or if this change in exon 1 from a serine to a tyrosine alters tyrosinase function, though this SNP is thought to contribute to normal pigment variation in individuals of Eur opean ancestry (Shriver et al., 2003; 2005; Norton et al., 2007). The exact role of tyrosinase in vitiligo pathoge nesis is unknown, but my findings support further investiga tion of this gene. I observed several significant findings in my study of the TYRP1 gene. TYRP1 protein is thought to play an important role in enhancing ty rosinase function, and in its active state, it is believed to be part of a multi-en zyme complex. I found SNPs in TYRP1 to be in HardyWeinberg disequilibrium (HWD) for vitiligo cases but not controls, sugges ting a possible linkage disequilibrium difference between the two gr oups, which may be considered supportive of 89

PAGE 90

association. It is no teworthy that in my ha ploblock analysis of TYRP1, I observed a greater number of heterozygotes in my vitiligo patients comp ared to controls. This heterozygosity of the TYRP1 gene may result in a destabilization of th e enzyme complex similar to a dominantnegative effect. An increase in heterozygosity in vitiligo patients was previously observed for a SNP in the catalase gene, where a dominant-neg ative effect was also hypothesized for that tetrameric protein complex (Casp et al., 2002). The catechol-O-methyltransferase or COMT gene was shown to have a protective haplotype in my white case-cont rol analysis. This supports the idea that the COMT enzyme functions to protect the melanocyte during melanogenesis. A reduction of COMT activity or an alteration of COMT expression in vitiligo patients may contribute to the autocytotoxic model of vitiligo (Le Poole et al., 1994; Turs en et al., 2002; Li et al., 2 008). Based on my finding and the work of others, further exploration of the COMT gene in additional genetic association studies is warranted to uncover its role in genetic susceptibi lity to vitiligo. Validated association could be followed by functional studies. Final Thoughts and a Note of Caution In any association study including ours, there is the possibility that spurious associations may be found. This is especially true for complex diseases where multiple genes may have subtle effects on disease pat hogenesis. Because several of the p-values for suggested associations were only moderate (0.01
PAGE 91

The inconsistencies observed in my study conc erning the significance associated with one marker for a particular gene, but not confirmed by other markers for the same gene, or by TDT analysis, may be a function of se veral factors. These incongru ities do not necessarily exclude any of the genes studied as candidates for involve ment in vitiligo; however, they do necessitate that the results of my study be considered with relative caution until the suggested associations can be confirmed by replication or other met hodologies. Lack of support by a significant TDT for associations suggested by casecontrol analysis may also be a result of insufficient number of informative families. The location of SNPs may also influence whether or not a marker reveals an association. A marker at one end of a gene may not show significant association if a pathogenic mutation is located at the other e nd of the gene. This may be one reason why haplotype analysis of SNPs yiel ded different information for some of the genes examined, as multiple markers may give different information about pathogenic changes in a gene. If positive associations are found, additional investigation into the genetic factors underlying vitiligo susceptibility should be cond ucted on an increased number of subjects, on subjects of different ethnic backgrounds in casecontrol analysis, and on an increased number of families for TDT analysis. Sequence analysis coul d then be conducted to elucidate the nature of the genetic characteristic(s) suggested by case-control and TD T analysis to influence the development of the vitiligo phenotype. A dditional methodologies, like gene expression experiments and protein structure studies, are needed to confir m results and better elucidate a proteins function in disease pathogenesis. Another potential confound is th at association may be due to population st ratification, where the case and controls groups may be from different ethnic or geographic subgroups. Although every effort was made to match cases and controls in my study, it is possible that some 91

PAGE 92

of my findings are false due to a mismatch of population subgroups. Whereas false negatives cause researchers to miss truly informative findings, false positives lead them to chase genetic wild geese. While obvious, it bears repeating th at the greater the significance of a finding, the more likely that finding represents a true asso ciation. Findings showing only weak significance should be replicated in many independent studies to be validated. Future Directions I am currently working on trying SNP chips for my DNA samples, and have collaboration with an NIH-sponsored VITGene consortium. Combined, these experiments will allow us to examine tens of thousands of SN Ps across the human genome. The results from these new studies will include SNPs on the candidate genes discussed in my present study as well as thousands of additional genes. This will allow us to understand the findings of the present study and will likely suggest many more genes for follow-up analyses. Additional studies need to be conducted to examine protein expression of AIRE and of melanin biosynthesis genes to examine their role in vitiligo pathogenesis. This would involve collection of skin samples from vitiligo patients. Obtaining thymus sa mples to better understand AIRE protein expression would be much more di fficult, though theoretical ly this could be done be through tissue banking. However, the thymus at rophies with age so studies using tissues from older individuals might not be informative. The long-term goals of my research include determining whether at-risk individuals can be identified in families with a history of vitili go, which might influence the choice of strategies for treatment or prevention of vitiligo. Future collaborations with clinicians who see vitiligo patients would not only allow us gr eater access to a larg er sample pool, but it would also enable us to more accurately phenotype my subjects. Fu rther, a prospective study testing for association of a SNP or gene with a thera py outcome could yield very useful results. Working together, 92

PAGE 93

basic scientists and physicians could develop and then test nov el therapies that are more precisely targeted to prevent or treat vitiligo based on the genetic findings. Future studies could include epigenetic analysis of vitiligo candidate genes, such as the ones discussed in my study, to examine if epigenetic changes contribute to v itiligo susceptibility. With additionally funding, it would be vitally important to carefully phenotype all study subjects. Careful attention to phenotypes may allow fo r a more detailed examination of patient subpopulations, which may aid in understanding the complex genetic component of vitiligo. The inclusion of ancestry informative markers (AIMs) in future studies may also aid in our genetic examination of vitiligo by helping to more closel y match cases and controls for genetic ancestry. Because of the accessibility of skin and the relative ease of skin biopsies, vitiligo makes an excellent disease model for other autoimmune diseases. It is po ssible that what we learn from studying vitiligo genetic suscepti bility may be applicable to the study of other autoimmune diseases. Similarly, because vitiligo and me lanoma both involve the melanocyte, knowledge gained through studying vitiligo may be applicable to melanoma. If additional patients and/or different populati ons were examined with my putative positive findings, my results may change. Also, additional phenotypic data may be an alyzed so that I can test how these may relate to genotype. Such data could be helpful in understanding observed variability in phenotype (e.g., why some vitiligo patients respond to treatments and others do not, or why some have unusual patterns of vitiligo). The identification of vitiligo susceptibility alleles may reveal new pathways and potential targets for treatment. The involvement of proteins expressed predominantly in melanocytes w ould also help specificity of targeted therapy; for example, a skin cream may only affect melanoc ytes but not other cell t ypes. This may also 93

PAGE 94

shed light on how environmental factors trigger vitiligo onset in some susceptible individuals but not others. I am establishing new statistical collaborations to use more s ophisticated models to look at both the combination of several genes simultaneously (to reveal gene-gene in teractions), as well as more complex family structures. Thus, the future may reveal additional discoveries, based on new statistical analysis and additional clinical data. The goal of future work in vitiligo will be to elucidate the complex pathogenesis of vitiligo with a goal to provide targets for rational therapy design. 94

PAGE 95

APPENDIX A SINGLE NUCLEOTIDE POLYMORPHISM DETAILS Table A-1. All SNPs genotyped Gene Symbol rs# Base pair change & location Chromosome: base Allelic p-value Genotypic p-value TDT p-value AIRE rs2776377 A/G promoter 21:44529072 0.86 0.46 1.00 AIRE rs1800520 C/G exon 7, S278R 21:44534334 9.90E-09* 2.00E-07* 0.16 COMT rs2020917 T/C 5'UTR 22:18308884 0.12 0.18 0.56 COMT rs6269 G/A intron 2 22:18329952 0.90 0.39 0.10 DCT rs9516413 C/T 3'UTR 13:93889348 0.0047* 0.027* 0.0052* DCT rs7987802 C/T intron 6 13:93895052 0.48 0.57 0.9 DCT rs4318084 C/T intron 6 13:93896999 0.26 0.13 6.2E-06* DCT rs9524493 C/G intron 6 13:93899790 0.32 0.55 0.49 DCT rs9516418 C/T intron 6 13:93909510 0.88 0.98 0.0097* DCT rs1028805 G/T intron 2 13:93917270 0.86 0.87 0.0025* DCT rs11618471 A/G intron 2 13:93918325 0.23 0.35 3.7E-06* DCT rs7991232 A/G intron 1 13:93928296 0.33 0.46 0.81 FBXO11 rs960106 C/T intron 1 2:47932342 0.95 0.98 0.66 FBXO11, MSH6 rs3136367 C/G 3'UTR ( FBXO11), intron 8 ( MSH6 ) 2:47887055 0.13 0.21 0.16 PAH rs1722381 A/G intron 6 12:101771966 0.19 0.44 0.65 PAH rs1522307 A/G intron 2 12:101822647 0.38 0.39 0.79 TYR rs1042602 A/C exon 1, S192Y 11:88551344 0.49 0.31 0.31 TYR rs621313 G/A intron 1 11:88553311 0.81 0.45 0.51 TYR rs594647 T/C intron 1 11:88561205 0.67 0.34 0.30 TYR rs10765197 A/C intron 2 11:88564976 0.52 0.32 0.12 TYR rs12791412 A/G intron 2 11:88570229 0.05* 0.13 0.02* TYR rs2000554 G/A intron 2 11:88575589 0.34 0.24 0.29 TYR rs10830250 C/G intron 3 11:88617255 0.25 0.31 0.26 TYR rs1827430 G/A intron 4 11:88658088 0.09 0.12 0.17 TYRP1 rs2762462 C/T intron 4 9:12689776 0.81 0.12 1.0 TYRP1 rs2733831 A/G intron 5 9:12693484 0.72 0.41 0.81 TYRP1 rs2733833 G/T intron 6 9:12695095 0.84 0.0254* 0.43 TYRP1 rs683 A/C 3'UTR 9:12699305 0.58 0.11 0.81 TYRP1 rs2762464 A/T 3'UTR 9:12699586 0.72 0.06 0.90 TYRP1 rs1063380 C/T 3'UTR 9:12700090 0.57 0.10 0.39 denotes a significant finding; all values are uncorrected p-values 95

PAGE 96

LIST OF REFERENCES Abdel-Naser, M.B., Kruger-Krasagakes, S., Kras agakis, K., Gollnick, H., Abdel-Fattah, A. and Orfanos, C.E.: Further evidence for involve ment of both cell mediated and humoral immunity in generalized vitili go. Pigment Cell Res 7 (1994) 1-8. Abdel-Naser, M.B., Ludwig, W.D., Gollnick, H. and Orfanos, C.E.: Nonsegmental vitiligo: decrease of the CD45RA+ T-cell subset and evidence for peripheral T-cell activation. Int J Dermatol 31 (1992) 321-6. A haplotype map of the human genome. Nature 437 (2005) 1299-320. Akhtar, S., Gavalas, N.G., Gawkrodger, D.J., Watson, P.F., Weetman, A.P. and Kemp, E.H.: An insertion/deletion polymorphism in the gene encoding angiotensin converting enzyme is not associated with generalised vitiligo in an English population. Arch Dermatol Res 297 (2005) 94-8. al Badri, A.M., Foulis, A.K., Todd, P.M., Gariou ch, J.J., Gudgeon, J.E., Stewart, D.G., Gracie, J.A. and Goudie, R.B.: Abnormal expression of MHC class II and ICAM-1 by melanocytes in vitiligo. J Pathol 169 (1993) 203-6. Alkhateeb, A., Fain, P.R. and Spritz, R.A.: Candi date functional promoter variant in the FOXD3 melanoblast developmental regulator gene in autosomal dominant vitiligo. J Invest Dermatol 125 (2005) 388-91. Alkhateeb, A., Stetler, G.L., Ol d, W., Talbert, J., Uhlhorn, C., Taylor, M., Fox, A., Miller, C., Dills, D.G., Ridgway, E.C., Bennett, D.C., Fain, P.R. and Spritz, R.A.: Mapping of an autoimmunity susceptibility locus (AIS 1) to chromosome 1p31.3-p32.2. Hum Mol Genet 11 (2002) 661-7. An autoimmune disease, APECED, caused by mutations in a novel gene featuring two PHD-type zinc-finger domains. The Finnish-Ger man APECED Consortium. Autoimmune Polyendocrinopathy-Candidiasis-Ectodermal Dystrophy. Nat Genet 17 (1997) 399-403. Anderson, M.S., Venanzi, E.S., Klein, L., Chen, Z., Berzins, S.P., Turley, S.J., von Boehmer, H., Bronson, R., Dierich, A., Benoist, C. and Mathis, D.: Projectio n of an immunological self shadow within the thymus by the AIRE protein. Science 298 (2002) 1395-401. Arcos-Burgos, M., Parodi, E., Salgar, M., Bedoya, E., Builes, J., Jaramillo, D., Ceballos, G., Uribe, A., Rivera, N., Rivera, D., Fonseca, I., Camargo, M. and Palacio, G.: Vitiligo: complex segregation and linkage disequilibrium analyses with respect to microsatellite loci spanning the HLA. Hum Genet 110 (2002) 334-42. Arrunategui, A., Arroyo, C., Garcia L., Covelli, C., Escobar, C., Ca rrascal, E. and Falabella, R.: Melanocyte reservoir in vitiligo. Int J Dermatol 33 (1994) 484-7. 96

PAGE 97

Badri, A.M., Todd, P.M., Garioch, J.J., Gudgeon, J.E., Stewart, D.G. and Goudie, R.B.: An immunohistological study of cutaneous lym phocytes in vitiligo. J Pathol 170 (1993) 14955. Baharav, E., Merimski, O., Shoenfeld, Y., Zigelm an, R., Gilbrud, B., Yecheskel, G., Youinou, P. and Fishman, P.: Tyrosinase as an autoan tigen in patients with vitiligo. Clin Exp Immunol 105 (1996) 84-8. Bakis-Petsoglou, S., Le Guay, J.L. and W ittal, R.: A randomized, double-blinded, placebocontrolled trial of pseudocatalase cream and na rrowband ultraviolet B in the treatment of vitiligo. Br J Dermatol (2009) [e lectronic publication, in press]. Bethea, D., Fullmer, B., Syed, S., Seltzer, G., Ti ano, J., Rischko, C., Gillespie, L., Brown, D. and Gasparro, F.P.: Psoralen phot obiology and photochemotherapy: 50 years of science and medicine. J Dermatol Sci 19 (1999) 78-88. Bhatia, P.S., Mohan, L., Pandey, O.N., Singh, K.K ., Arora, S.K. and Mukhija, R.D.: Genetic nature of vitiligo. J Dermatol Sci 4 (1992) 180-4. Birlea, S.A., Fain, P.R. and Spritz, R.A.: A Ro manian population isolate with high frequency of vitiligo and associated autoimmune di seases. Arch Dermatol 144 (2008) 310-6. Bjorses, P., Pelto-Huikko, M., Kaukonen, J., Aaltonen, J., Peltonen, L. and Ulmanen, I.: Localization of the APECED protein in distinct nuclear structures. Hum Mol Genet 8 (1999) 259-66. Boe Wolff, A.S., Oftedal, B., Johansson, S., Br uland, O., Lovas, K., Meager, A., Pedersen, C., Husebye, E.S. and Knappskog, P.M.: AIRE variations in Addison's disease and autoimmune polyendocrine syndromes (APS): part ial gene deletions contribute to APS I. Genes Immun 9 (2008) 130-6. Boissy, R.E., Liu, Y.Y., Medrano, E.E. and Nord lund, J.J.: Structural aberration of the rough endoplasmic reticulum and melanosome compartmentalization in long-term cultures of melanocytes from vitiligo patients. J Invest Dermatol 97 (1991) 395-404. Bos, J.D.: Skin immune system (SIS) : cu taneous immunology and clinical immunodermatology, 3rd ed. CRC Press, Boca Raton, 2005. Bottomley, M.J., Collard, M.W., Huggenvik, J.I. Liu, Z., Gibson, T.J. and Sattler, M.: The SAND domain structure defines a novel DNAbinding fold in transcriptional regulation. Nat Struct Biol 8 (2001) 626-33. Brown-Harrell, V., Nitta, A.T. and Goble, M.: Apparent exacerbation of vitiligo syndrome in a patient with pulmonary Mycobacterium avium complex disease who received clofazimine therapy. Clin Infect Dis 22 (1996) 581-2. 97

PAGE 98

Buc, M., Fazekasova, H., Cechova, E., Hegyi, E., Kolibasova, K. and Ferencik, S.: Occurrence rates of HLA-DRB1, HLA-DQB1, and HLA-DP B1 alleles in patients suffering from vitiligo. Eur J Dermatol 8 (1998) 13-5. Buckley, D.A. and du Vivier, A. W.: Topical immunotherapy in dermatology. Int J Clin Pract 53 (1999) 130-7. Bystryn, J.C.: Immune mechanisms in vitiligo. Clin Dermatol 15 (1997) 853-61. Camacho, F. and Mazuecos, J.: Treatment of vitili go with oral and topical phenylalanine: 6 years of experience. Arch Dermatol 135 (1999) 216-7. Cardon, L.R. and Bell, J.I.: Association study de signs for complex diseases. Nat Rev Genet 2 (2001) 91-9. Casp, C.B.: Genetic associati on of catalase and antigen pr ocessing genes with vitiligo susceptibility, Pathology, Immunology, and Laborat ory Medicine. University of Florida, Gainesville, (2003). 3. Casp, C.B., She, J.X. and McCormack, W.T.: Ge netic association of the catalase gene (CAT) with vitiligo susceptibility. Pigment Cell Res 15 (2002) 62-6. Casp, C.B., She, J.X. and McCormack, W.T.: Gene s of the LMP/TAP cluster are associated with the human autoimmune disease vi tiligo. Genes Immun 4 (2003) 492-9. Cetani, F., Barbesino, G., Borsari, S., Pardi, E., Cianferotti, L., Pinchera, A. and Marcocci, C.: A novel mutation of the autoimmune regulator gene in an Italian kindred with autoimmune polyendocrinopathy-candidiasis-ectodermal dystrophy, acting in a dominant fashion and strongly cosegregating with hypothyroid au toimmune thyroiditis. J Clin Endocrinol Metab 86 (2001) 4747-52. Chen, Y.F., Yang, P.Y., Hu, D.N., Kuo, F.S., Hu ng, C.S. and Hung, C.M.: Treatment of vitiligo by transplantation of cultured pure melanocyte suspension: analysis of 120 cases. J Am Acad Dermatol 51 (2004) 68-74. Chi, A., Valencia, J.C., Hu, Z.Z., Watabe, H., Yamaguchi, H., Mangini, N.J., Huang, H., Canfield, V.A., Cheng, K.C., Yang, F., Abe, R ., Yamagishi, S., Shabanowitz, J., Hearing, V.J., Wu, C., Appella, E. and Hunt, D.F.: Prot eomic and bioinformatic characterization of the biogenesis and function of melanos omes. J Proteome Res 5 (2006) 3135-44. Cui, J., Arita, Y. and Bystryn, J.C.: Cytolytic antibodies to melanocytes in vitiligo. J Invest Dermatol 100 (1993) 812-5. Cui, J. and Bystryn, J.C.: Melanoma and vitili go are associated with antibody responses to similar antigens on pigment cells Arch Dermatol 131 (1995) 314-8. 98

PAGE 99

Das, P.K., van den Wijngaard, R.M., Wankowicz-Ka linska, A. and Le Poole, I.C.: A symbiotic concept of autoimmunity and tumour immun ity: lessons from vitiligo. Trends Immunol 22 (2001) 130-6. Davis, M.D., Ribeiro, P., Tipper, J. and Kauf man, S.: "7-tetrahydrobiopterin," a naturally occurring analogue of tetrahydrobi opterin, is a cofactor for and a potential inhibitor of the aromatic amino acid hydroxylases. Proc Natl Acad Sci U S A 89 (1992) 10109-13. De La Vega, F.M., Isaac, H.I. and Scafe, C.R.: A tool for selecting SNPs for association studies based on observed linkage disequilibrium patterns. Pac Symp Biocomput (2006) 487-98. Drago, A., De Ronchi, D. and Serretti, A.: Incomp lete coverage of candidate genes: a poorly considered bias. Curr Genomics 8 (2007) 476-83. Dunston, G.M. and Halder, R.M.: Vitiligo is a ssociated with HLA-DR4 in black patients. A preliminary report. Arch Dermatol 126 (1990) 56-60. Em, S., Laddha, N.C., Chatterjee, S., Gani, A.R., Malek, R.A., Shah, B.J. and Begum, R.: Association of catalase T/C exon 9 and glutathione per oxidase codon 200 polymorphisms in relation to their activities and oxidative st ress with vitiligo susceptibility in Gujarat population. Pigment Cell Res 20 (2007) 405-7. Fain, P.R., Babu, S.R., Bennett, D.C. and Sp ritz, R.A.: HLA class II haplotype DRB1*04DQB1*0301 contributes to risk of familial gene ralized vitiligo and early disease onset. Pigment Cell Res 19 (2006) 51-7. Fain, P.R., Gowan, K., LaBerge, G.S., Alkhateeb, A., Stetler, G.L., Talbert, J., Bennett, D.C. and Spritz, R.A.: A genomewide screen for gene ralized vitiligo: confirmation of AIS1 on chromosome 1p31 and evidence for additional susceptibility loci Am J Hum Genet 72 (2003) 1560-4. Falabella, R.: Surgical therapies fo r vitiligo. Clin Dermatol 15 (1997) 927-39. Finco, O., Cuccia, M., Martinetti, M., Ruberto, G., Orecchia, G. and Rabbiosi, G.: Age of onset in vitiligo: relationship with HLA s upratypes. Clin Genet 39 (1991) 48-54. Fishman, P., Azizi, E., Shoenfeld, Y., Sredni, B., Yecheskel, G., Ferrone, S., Zigelman, R., Chaitchik, S., Floro, S. and Djaldetti, M.: Vitiligo autoantibodies are effective against melanoma. Cancer 72 (1993) 2365-9. Gajko-Galicka, A.: Mutations in type I collagen genes resulting in osteogenesis imperfecta in humans. Acta Biochim Pol 49 (2002) 433-41. Gibson, T.J., Ramu, C., Gemund, C. and Aasland, R.: The APECED polyglandular autoimmune syndrome protein, AIRE-1, contains the SAND domain and is probably a transcription factor. Trends Biochem Sci 23 (1998) 242-4. 99

PAGE 100

Gilhar, A., Zelickson, B., Ulman, Y. and Etzioni, A.: In vivo destruction of melanocytes by the IgG fraction of serum from patients with vi tiligo. J Invest Dermatol 105 (1995) 683-6. Giovannelli, L., Bellandi, S., P itozzi, V., Fabbri, P., Dolara, P. and Moretti, S.: Increased oxidative DNA damage in mononuclear leukoc ytes in vitiligo. Mutat Res 556 (2004) 101-6. Gomes, I., Collins, A., Lonjou, C., Thomas, N.S., Wilkinson, J., Watson, M. and Morton, N.: Hardy-Weinberg quality control. Ann Hum Genet 63 (1999) 535-8. Gottschalk, G.M. and Kidson, S.H.: Molecular anal ysis of vitiligo lesions reveals sporadic melanocyte survival. Int J Dermatol 46 (2007) 268-72. Halaban, R., Patton, R.S., Cheng, E., Svedine, S ., Trombetta, E.S., Wahl, M.L., Ariyan, S. and Hebert, D.N.: Abnormal acidification of melano ma cells induces tyrosinase retention in the early secretory pathway. J Biol Chem 277 (2002) 14821-8. Hann, S.K., Park, Y.K., Lee, K.G., Choi, E.H. a nd Im, S.: Epidermal changes in active vitiligo. J Dermatol 19 (1992) 217-22. Harning, R., Cui, J. and Bystryn, J.C.: Relation between the incidence an d level of pigment cell antibodies and disease activity in vit iligo. J Invest Dermatol 97 (1991) 1078-80. Hochberg, Y. and Benjamini, Y.: More powerful procedures for multiple significance testing. Stat Med 9 (1990) 811-8. Hu, Z., Liu, J.B., Ma, S.S., Yang, S. and Zhang, X. J.: Profile of childhood vitiligo in China: an analysis of 541 patients. Pe diatr Dermatol 23 (2006) 114-6. Hughes, L.B., Beasley, T.M., Patel, H., Tiwari, H.K., Morgan, S.L., Baggott, J.E., Saag, K.G., McNicholl, J., Moreland, L.W ., Alarcon, G.S. and Bridges, S.L., Jr.: Racial or ethnic differences in allele frequencies of single-nucleotide polymorphisms in the methylenetetrahydrofolate reduct ase gene and their influence on response to methotrexate in rheumatoid arthritis. Ann Rheum Dis 65 (2006) 1213-8. Ikari, K., Momohara, S., Nakamura, T., Hara, M ., Yamanaka, H., Tomatsu, T. and Kamatani, N.: Supportive evidence for a genetic association of the FCRL3 promoter polymorphism with rheumatoid arthritis. An n Rheum Dis 65 (2006) 671-3. Im, S., Hann, S.K., Kim, H.I., Kim, N.S. and Park, Y.K.: Biologic characteristics of cultured human vitiligo melanocytes. In t J Dermatol 33 (1994) 556-62. Jimbow, K., Chen, H., Park, J.S. and Thomas, P. D.: Increased sensitivity of melanocytes to oxidative stress and abnormal expression of ty rosinase-related protein in vitiligo. Br J Dermatol 144 (2001) 55-65. 100

PAGE 101

Jin, S.Y., Park, H.H., Li, G.Z., Lee, H.J., H ong, M.S., Hong, S.J., Park, H.K., Chung, J.H. and Lee, M.H.: Association of angiotensin c onverting enzyme gene I/D polymorphism of vitiligo in Korean population. Pi gment Cell Res 17 (2004a) 84-6. Jin, S.Y., Park, H.H., Li, G.Z., Lee, H.J., Hong, M.S., Park, H.J., Park, H.K., Seo, J.C., Yim, S.V., Chung, J.H. and Lee, M.H.: Associat ion of estrogen recept or 1 intron 1 C/T polymorphism in Korean vitiligo patients. J Dermatol Sci 35 (2004b) 181-6. Jin, Y., Bennett, D.C., Amadi-Myers, A., Holland, P., Riccardi, S.L., Gowan, K., Fain, P.R. and Spritz, R.A.: Vitiligo-associated multiple au toimmune disease is not associated with genetic variation in AIRE. Pi gment Cell Res 20 (2007a) 402-4. Jin, Y., Mailloux, C.M., Gowan, K., Riccardi, S.L., LaBerge, G., Bennett, D.C., Fain, P.R. and Spritz, R.A.: NALP1 in vitiligo-associated multiple autoimmune disease. N Engl J Med 356 (2007b) 1216-25. Kang, C.P., Lee, H.S., Ju, H., Cho, H., Kang, C. and Bae, S.C.: A functional haplotype of the PADI4 gene associated with increased rheuma toid arthritis susceptibility in Koreans. Arthritis Rheum 54 (2006) 90-6. Kao, C.H. and Yu, H.S.: Depletion and repopulati on of Langerhans cells in nonsegmental type vitiligo. J Dermatol 17 (1990) 287-96. Kemp, E.H., Ajjan, R.A., Waterman, E.A., Gawkrodger, D.J., Cork, M.J., Watson, P.F. and Weetman, A.P.: Analysis of a microsatel lite polymorphism of the cytotoxic Tlymphocyte antigen-4 gene in patients w ith vitiligo. Br J Dermatol 140 (1999a) 73-8. Kemp, E.H., Gawkrodger, D.J., MacNeil, S., Watson, P.F. and Weetman, A.P.: Detection of tyrosinase autoantibodies in patients with vitiligo using 35S-labeled recombinant human tyrosinase in a radioimmunoassay. J Invest Dermatol 109 (1997a) 69-73. Kemp, E.H., Gawkrodger, D.J., Watson, P.F. and Weetman, A.P.: Immunoprecipitation of melanogenic enzyme autoantigens with vitiligo sera: evidence for cross-reactive autoantibodies to tyrosinase and tyrosinase-relate d protein-2 (TRP-2). Clin Exp Immunol 109 (1997b) 495-500. Kemp, E.H., Waterman, E.A., Gawkrodger, D.J., Watson, P.F. and Weetman, A.P.: Autoantibodies to tyrosinase-related protein1 detected in the sera of vitiligo patients using a quantitative radiobinding as say. Br J Dermatol 139 (1998) 798-805. Kemp, E.H., Waterman, E.A., Gawkrodger, D.J., Watson, P.F. and Weetman, A.P.: Identification of epitopes on tyrosinase wh ich are recognized by autoantibodies from patients with vitiligo. J Invest Dermatol 113 (1999b) 267-71. Kemp, E.H., Waterman, E.A., Hawes, B.E., O'Ne ill, K., Gottumukkala, R.V., Gawkrodger, D.J., Weetman, A.P. and Watson, P.F.: The melani n-concentrating hormone receptor 1, a novel target of autoantibody responses in vitiligo. J Clin Invest 109 (2002) 923-30. 101

PAGE 102

Kim, S.M., Chung, H.S. and Hann, S.K.: The gene tics of vitiligo in Ko rean patients. Int J Dermatol 37 (1998) 908-10. Kingo, K., Aunin, E., Karelson, M., Philips, M.A., Ra tsep, R., Silm, H., Vasar, E., Soomets, U. and Koks, S.: Gene expression analysis of mela nocortin system in vitiligo. J Dermatol Sci 48 (2007) 113-22. Kobayashi, T., Imokawa, G., Bennett, D.C. and H earing, V.J.: Tyrosinase stabilization by Tyrp1 (the brown locus protein). J Biol Chem 273 (1998) 31801-5. Kobayashi, T., Urabe, K., Winder, A., Jimene z-Cervantes, C., Imokawa, G., Brewington, T., Solano, F., Garcia-Borron, J.C. and Hearing, V. J.: Tyrosinase relate d protein 1 (TRP1) functions as a DHICA oxidase in melani n biosynthesis. EMBO J 13 (1994) 5818-25. Kovacs, S.O.: Vitiligo. J Am Acad Dermatol 38 (1998) 647-66; quiz 667-8. Kumar, P.G., Laloraya, M., Wang, C.Y., Ruan, Q.G., Davoodi-Semiromi, A., Kao, K.J. and She, J.X.: The autoimmune regulator (AIRE) is a DNA-binding protein. J Biol Chem 276 (2001) 41357-64. Kwon, J.M. and Goate, A.M.: The candidate ge ne approach. Alcohol Res Health 24 (2000) 1648. Le Poole, I.C., Boissy, R.E., Sarangarajan, R., Ch en, J., Forristal, J.J., Sheth, P., Westerhof, W., Babcock, G., Das, P.K. and Saelinger, C.B. : PIG3V, an immortalized human vitiligo melanocyte cell line, expresse s dilated endoplasmic reticulum In Vitro Cell Dev Biol Anim 36 (2000) 309-19. Le Poole, I.C., Das, P.K., van den Wijngaard, R. M., Bos, J.D. and Westerhof, W.: Review of the etiopathomechanism of vitiligo: a conve rgence theory. Exp Dermatol 2 (1993a) 145-53. Le Poole, I.C. and Luiten, R.M.: Autoimmune etiology of generalized vitiligo. Curr Dir Autoimmun 10 (2008) 227-43. Le Poole, I.C., Sarangarajan, R., Zhao, Y., Stennett, L.S., Brown, T.L., Sheth, P., Miki, T. and Boissy, R.E.: 'VIT1', a novel gene associat ed with vitiligo. Pigment Cell Res 14 (2001) 475-84. Le Poole, I.C., van den Wijngaard, R.M., Smit, N.P., Oosting, J., Westerhof, W. and Pavel, S.: Catechol-O-methyltransferase in vitil igo. Arch Dermatol Res 286 (1994) 81-6. Le Poole, I.C., van den Wijngaard, R.M., Westerhof, W. and Das, P.K.: Presence of T cells and macrophages in inflammatory vitiligo skin parallels melanocyte disappearance. Am J Pathol 148 (1996) 1219-28. Le Poole, I.C., van den Wijngaard, R.M., Westerhof, W., Dutrieux, R.P. and Das, P.K.: Presence or absence of melanocytes in vitiligo le sions: an immunohistochemi cal investigation. J Invest Dermatol 100 (1993b) 816-22. 102

PAGE 103

Lerner, A.B. and Nordlund, J.J.: Vitiligo. What is it? Is it important? JAMA 239 (1978) 1183-7. Levine, E.L. and Ribeiro, G.G.: Vitiligo and radiotherapy: the Koebner phenomenon demonstrated in patients with vitiligo undergoing radiotherapy for carcinoma of the breast. Clin Oncol (R Coll Radiol) 6 (1994) 133-4. Lewontin, R.C.: The interaction of selection and linkage. I. Ge neral considerations; heterotic models. Genetics 49 (1964) 49-67. Li, K., Li, C., Gao, L., Yang, L., Li, M., Liu, L., Zhang, Z., Liu, Y. and Gao, T.: A functional single-nucleotide polymorphism in the catec hol-O-methyltransferase gene alter vitiligo risk in a Chinese population. Arch Derm atol Res (2008) [Epub ahead of print]. Li, M. and Li, C.: Assessing departure from Ha rdy-Weinberg equilibrium in the presence of disease association. Genet Epidemiol 32 (2008) 589-99. Liston, A. and Goodnow, C.C.: Genetic lesions in thymic T cell clonal deletion and thresholds for autoimmunity. Novartis Found Sy mp 267 (2005) 180-92; discussion 192-9. Liston, A., Gray, D.H., Lesage, S ., Fletcher, A.L., Wilson, J., Webster, K.E., Scott, H.S., Boyd, R.L., Peltonen, L. and Goodnow, C.C.: Gene dosage--limiting role of AIRE in thymic expression, clonal deletion, and organ-speci fic autoimmunity. J Exp Med 200 (2004) 1015-26. Liu, T., Johnson, J.A., Casella, G. and Wu, R.: Sequencing complex diseases With HapMap. Genetics 168 (2004) 503-11. Lotti, T., Gori, A., Zanieri, F., Colucci, R. and Moretti, S.: Vitiligo: new and emerging treatments. Dermatol Ther 21 (2008) 110-7. Majumder, P.P., Nordlund, J.J. and Nath, S.K.: Pa ttern of familial aggregation of vitiligo. Arch Dermatol 129 (1993) 994-8. Manga, P., Sheyn, D., Yang, F., Sarangarajan, R. and Boissy, R.E.: A role for tyrosinase-related protein 1 in 4-tert-butylphenol -induced toxicity in melanocyt es: Implications for vitiligo. Am J Pathol 169 (2006) 1652-62. Manolache, L., Petrescu-Seceleanu, D. and Benea, V. : Correlation of stressful events with onset of vitiligo in children. J Eur Acad Dermatol Venereol 23 (2009) 187-8. Maresca, V., Roccella, M., Roccella, F., Camera, E ., Del Porto, G., Passi, S., Grammatico, P. and Picardo, M.: Increased sensitivit y to peroxidative agents as a possible pathogenic factor of melanocyte damage in vitiligo. J Invest Dermatol 109 (1997) 310-3. Mathis, D. and Benoist, C.: Aire. Annu Rev Immunol 27 (2009) 287-312. McGinnis, R.E., Ewens, W.J. and Spielman, R.S.: The TDT reveals linkage and linkage disequilibrium in a rare disease. Genet Epidemiol 12 (1995) 637-40. 103

PAGE 104

Mehers, K.L. and Gillespie, K.M.: The genetic ba sis for type 1 diabetes. Br Med Bull 88 (2008) 115-29. Merimsky, O., Shoenfeld, Y., Chaitchik, S., Yech eskel, G. and Fishman, P.: Antigens and antibodies in malignant melano ma. Tumour Biol 15 (1994) 188-202. Millington, G.W. and Levell, N.J.: Vitiligo: th e historical curse of depigmentation. Int J Dermatol 46 (2007) 990-5. Mizuta, I., Satake, W., Nakabayashi, Y., Ito, C ., Suzuki, S., Momose, Y., Nagai, Y., Oka, A., Inoko, H., Fukae, J., Saito, Y., Sawabe, M., Murayama, S., Yamamoto, M., Hattori, N., Murata, M. and Toda, T.: Multiple candidate gene analysis identifies alpha-synuclein as a susceptibility gene for sporadic Parkins on's disease. Hum Mol Genet 15 (2006) 1151-8. Morris, G.A., Lowe, C.E., Cooper, J.D., Payne, F ., Vella, A., Godfrey, L., Hulme, J.S., Walker, N.M., Healy, B.C., Lam, A.C., Lyons, P.A. and Todd, J.A.: Polymorphism discovery and association analyses of the interferon gene s in type 1 diabetes BMC Genet 7 (2006) 12. Mozzanica, N., Frigerio, U., Finzi, A.F., Cattaneo, A., Negri, M., Scaglione, F., Fraschini, F. and Foppa, S.: T cell subpopulations in vitiligo: a chronobiologi c study. J Am Acad Dermatol 22 (1990) 223-30. Mulekar, S.V.: Melanocyte-keratino cyte cell transplantation for st able vitiligo. Int J Dermatol 42 (2003) 132-6. Naftzger, C., Takechi, Y., Kohda, H., Hara, I ., Vijayasaradhi, S. and Houghton, A.N.: Immune response to a differentiation antigen indu ced by altered antigen: a study of tumor rejection and autoimmunity. Proc Natl Acad Sci U S A 93 (1996) 14809-14. Nagamine, K., Peterson, P., Scott, H.S., Kudoh, J., Minoshima, S., Heino, M., Krohn, K.J., Lalioti, M.D., Mullis, P.E., Antonarakis, S.E., Kawasaki, K., Asakawa, S., Ito, F. and Shimizu, N.: Positional cloning of the APECED gene. Nat Genet 17 (1997) 393-8. Nancy, A.L. and Yehuda, S.: Prediction and preven tion of autoimmune skin disorders. Arch Dermatol Res 301 (2009) 57-64. Nath, S.K., Majumder, P.P. and Nordlund, J.J.: Genetic epidemiology of vitiligo: multilocus recessivity cross-validated. Am J Hum Genet 55 (1994) 981-90. Naughton, G.K., Eisinger, M. and Bystryn, J.C.: Detection of antibodie s to melanocytes in vitiligo by specific immunoprecipitation. J Invest Dermatol 81 (1983) 540-2. Nicolaidou, E., Antoniou, C., Stratigos, A. and Katsambas, A.D.: Narrowband ultraviolet B phototherapy and 308-nm excimer laser in the tr eatment of vitiligo: a review. J Am Acad Dermatol 60 (2009) 470-7. 104

PAGE 105

Njoo, M.D., Spuls, P.I., Bos, J.D., West erhof, W. and Bossuyt, P.M.: Nonsurgical repigmentation therapies in vitiligo. Meta-ana lysis of the literatu re. Arch Dermatol 134 (1998) 1532-40. Njoo, M.D. and Westerhof, W.: Vitiligo. Pathogene sis and treatment. Am J Clin Dermatol 2 (2001) 167-81. Nordlund, J.J.: The epidemiology and genetics of vitiligo. Clin Dermatol 15 (1997) 875-8. Nordlund, J.J. and Ortonne, J.P.: Vitiligo vulgari s. In: Nordlund, J.J. and et al. (Eds.), The Pigmentary System Physiology and Pa thophysiology. Oxford Univ. Press, 1998, pp. 513-551. Norris, A., Todd, C., Graham, A., Quinn, A.G. and Thody, A.J.: The expression of the c-kit receptor by epidermal melanocytes may be re duced in vitiligo. Br J Dermatol 134 (1996) 299-306. Norton, H.L., Kittles, R.A., Parra, E., McKei gue, P., Mao, X., Cheng, K., Canfield, V.A., Bradley, D.G., McEvoy, B. and Shriver, M.D.: Genetic evidence for the convergent evolution of light skin in Europeans and Ea st Asians. Mol Biol Evol 24 (2007) 710-22. Oetting, W.S.: The tyrosinase gene and oculocut aneous albinism type 1 (OCA1): A model for understanding the molecular bi ology of melanin formation. Pigment Cell Res 13 (2000) 320-5. Oetting, W.S. and King, R.A.: Molecular basis of albinism: mutations and polymorphisms of pigmentation genes associated with albinism. Hum Mutat 13 (1999) 99-115. Okamoto, T., Irie, R.F., Fujii, S., Huang, S.K., Nizze, A.J., Morton, D.L. and Hoon, D.S.: Antityrosinase-related protein-2 immune response in vitiligo patients and melanoma patients receiving active-specific immunotherapy. J Invest Dermatol 111 (1998) 1034-9. Ongenae, K., Van Geel, N. and Naeyaert, J.M. : Evidence for an autoimmune pathogenesis of vitiligo. Pigment Cell Res 16 (2003) 90-100. Orecchia, G., Perfetti, L., Malagoli, P., Borghini F. and Kipervarg, Y.: Vitiligo is associated with a significant in crease in HLA-A30, Cw6 and DQw3 and a decrease in C4AQ0 in northern Italian patients Dermatology 185 (1992) 123-7. Overwijk, W.W., Lee, D.S., Surman, D.R., Irvin e, K.R., Touloukian, C.E., Chan, C.C., Carroll, M.W., Moss, B., Rosenberg, S.A. and Res tifo, N.P.: Vaccination with a recombinant vaccinia virus encoding a "self" antigen i nduces autoimmune vitiligo and tumor cell destruction in mice: requirem ent for CD4(+) T lymphocytes. Proc Natl Acad Sci U S A 96 (1999) 2982-7. Papadavid, E., Yu, R.C., Munn, S. and Chu, A.C.: Strict anatomical coexis tence of vitiligo and psoriasis vulgaris--a Koebner phenomenon? Clin Exp Dermatol 21 (1996) 138-40. 105

PAGE 106

Park, H.H., Ha, E., Uhm, Y.K., Jin, S.Y., Kim, Y.J., Chung, J.H. and Lee, M.H.: Association study between catalase gene polymorphisms and th e susceptibility to vitiligo in Korean population. Exp Dermatol 15 (2006) 377-80. Passeron, T. and Ortonne, J.P.: Physiopathology and genetics of vitiligo. J Autoimmun 25 Suppl (2005) 63-8. Passi, S., Grandinetti, M., Maggio, F., Stancato, A. and De Luca, C.: Epidermal oxidative stress in vitiligo. Pigment Cell Res 11 (1998) 81-5. Pavel, S., Muskiet, F.A., de Ley, L., The, T.H. and van der Slik, W.: Identification of three indolic compounds in a pigmented-melano ma cell-culture supernatant by gas chromatography-mass spectrometry. J Ca ncer Res Clin Oncol 105 (1983) 275-9. Peterson, P., Org, T. and Rebane, A.: Transcriptional regulation by AIRE: molecular mechanisms of central tolerance. Nat Rev Immunol 8 (2008) 948-57. Phillips, J., Gawkrodger, D.J., Caddy, C.M., Hedley, S., Dawson, R.A., Smith-Thomas, L., Freedlander, E. and Mac Neil, S.: Keratinocytes suppress TRP-1 expression and reduce cell number of co-cultured melanocytes imp lications for grafting of patients with vitiligo. Pigment Cell Res 14 (2001) 116-25. Pitkanen, J. and Peterson, P.: Autoimmune regul ator: from loss of function to autoimmunity. Genes Immun 4 (2003) 12-21. Powell, F.C. and Dicken, C.H.: Vitiligo and ps oriasis. J Am Acad Dermatol 8 (1983) 136-7. Purohit, S., Kumar, P.G., Laloraya, M. a nd She, J.X.: Mapping DNA-binding domains of the autoimmune regulator protein. Bioc hem Biophys Res Commun 327 (2005) 939-44. Ram, M. and Shoenfeld, Y.: Harnessing autoimmunity (vitiligo) to treat melanoma: a myth or reality? Ann N Y Acad Sci 1110 (2007) 410-25. Redondo, M.J., Fain, P.R. and Eisenbarth, G.S.: Genetics of type 1A diabetes. Recent Prog Horm Res 56 (2001) 69-89. Rice, T.K., Schork, N.J. and Rao, D.C.: Methods for handling multiple testing. Adv Genet 60 (2008) 293-308. Risch, N. and Merikangas, K.: The future of gene tic studies of complex human diseases. Science 273 (1996) 1516-7. Riva, A. and Kohane, I.S.: SNPper: retrieval and analysis of human SNPs. Bioinformatics 18 (2002) 1681-5. Rizzi, M., Ferrera, F., Filaci, G. and Indiveri, F.: Disruption of im munological tolera nce: role of AIRE gene in autoimmunity. Autoimmun Rev 5 (2006) 145-7. 106

PAGE 107

Salanti, G., Amountza, G., Ntzani, E.E. and Ioa nnidis, J.P.: Hardy-Weinberg equilibrium in genetic association studies: an empirical evaluation of reporting, deviations, and power. Eur J Hum Genet 13 (2005) 840-8. Sampogna, F., Raskovic, D., Guerra, L., Pedicelli, C., Tabolli, S., Leoni, L., Alessandroni, L. and Abeni, D.: Identification of categories at ri sk for high quality of life impairment in patients with vitiligo. Br J Dermatol 159 (2008) 351-9. Schallreuter, K.U.: Successful treatment of oxidati ve stress in vitiligo. Skin Pharmacol Appl Skin Physiol 12 (1999) 132-8. Schallreuter, K.U., Chavan, B., Rokos, H., Hibberts, N., Panske, A. and Wood, J.M.: Decreased phenylalanine uptake and turnover in patients with vitiligo. Mol Genet Metab 86 Suppl 1 (2005) S27-33. Schallreuter, K.U., Gibbons, N.C., Zothner, C., Abou Elloof, M.M. and Wood, J.M.: Hydrogen peroxide-mediated oxidativ e stress disrupts calcium binding on calmodulin: more evidence for oxidative stress in vitiligo. Biochem Biophys Res Co mmun 360 (2007) 70-5. Schallreuter, K.U., Lemke, R., Brandt, O., Schw artz, R., Westhofen, M., Montz, R. and Berger, J.: Vitiligo and other diseases : coexistence or true association? Hamburg study on 321 patients. Dermatology 188 (1994a) 269-75. Schallreuter, K.U., Moore, J., Behrens-Williams, S., Panske, A. and Harari, M.: Rapid initiation of repigmentation in vitiligo with Dead Sea climatotherapy in combination with pseudocatalase (PC-KUS). Int J Dermatol 41 (2002) 482-7. Schallreuter, K.U., Moore, J., Wood, J.M., Beazl ey, W.D., Gaze, D.C., Tobin, D.J., Marshall, H.S., Panske, A., Panzig, E. and Hibberts, N.A.: In vivo and in vitro evidence for hydrogen peroxide (H2O2) accumulation in the ep idermis of patients with vitiligo and its successful removal by a UVB-activated pseudocat alase. J Investig Dermatol Symp Proc 4 (1999) 91-6. Schallreuter, K.U., Moore, J., Wood, J.M., Beazl ey, W.D., Peters, E.M., Marles, L.K., BehrensWilliams, S.C., Dummer, R., Blau, N. and Thony, B.: Epidermal H(2)O(2) accumulation alters tetrahydrobiopterin (6BH 4) recycling in vitiligo: identification of a general mechanism in regulation of all 6BH4-dependent processes? J Invest Dermatol 116 (2001) 167-74. Schallreuter, K.U. and Wood, J.M.: The importance of L-phenylalanine transport and its autocrine turnover to L-tyro sine for melanogenesis in human epidermal melanocytes. Biochem Biophys Res Commun 262 (1999) 423-8. Schallreuter, K.U. and Wood, J.M.: Thioredoxin re ductase its role in epidermal redox status. J Photochem Photobiol B 64 (2001) 179-84. Schallreuter, K.U., Wood, J.M. a nd Berger, J.: Low catalase levels in the epidermis of patients with vitiligo. J Invest Dermatol 97 (1991) 1081-5. 107

PAGE 108

Schallreuter, K.U., Wood, J.M ., Pittelkow, M.R., Buttner, G., Swanson, N., Korner, C. and Ehrke, C.: Increased monoamine oxidase A act ivity in the epidermis of patients with vitiligo. Arch Dermatol Res 288 (1996) 14-8. Schallreuter, K.U., Wood, J.M., Pittelkow, M.R ., Gutlich, M., Lemke, K.R., Rodl, W., Swanson, N.N., Hitzemann, K. and Ziegler, I.: Regula tion of melanin biosynt hesis in the human epidermis by tetrahydrobiopt erin. Science 263 (1994b) 1444-6. Schallreuter, K.U., Wood, J.M., Ziegler, I., Lemke, K.R., Pittelkow, M.R., Lindsey, N.J. and Gutlich, M.: Defective tetrahydrobiopterin and catecholamine biosynthesis in the depigmentation disorder vitiligo. Bioc him Biophys Acta 1226 (1994c) 181-92. Schallreuter, K.U., Zschiesche, M., Moore, J ., Panske, A., Hibberts, N.A., Herrmann, F.H., Metelmann, H.R. and Sawatzki, J.: In vivo evidence for compromised phenylalanine metabolism in vitiligo. Biochem Bi ophys Res Commun 243 (1998) 395-9. Shoemaker, J., Painter, I. and Weir, B.S.: A Bayesian characterization of Hardy-Weinberg disequilibrium. Genetics 149 (1998) 2079-88. Shriver, M.D., Parra, E.J., Dios, S., Bonilla, C ., Norton, H., Jovel, C., Pfaff, C., Jones, C., Massac, A., Cameron, N., Baron, A., Jack son, T., Argyropoulos, G., Jin, L., Hoggart, C.J., McKeigue, P.M. and Kittles, R.A.: Skin pigmentation, biogeographical ancestry and admixture mapping. Hum Genet 112 (2003) 387-99. Smyth, D.J., Howson, J.M., Payne, F., Maier, L. M., Bailey, R., Holland, K., Lowe, C.E., Cooper, J.D., Hulme, J.S., Vella, A., Dalhman, I., La m, A.C., Nutland, S., Walker, N.M., Twells, R.C. and Todd, J.A.: Analysis of polymorphisms in 16 genes in type 1 diabetes that have been associated with other immune-media ted diseases. BMC Med Genet 7 (2006) 20. Song, Y.H., Connor, E., Li, Y., Zorovich, B., Ba lducci, P. and Maclaren, N.: The role of tyrosinase in autoimmune vi tiligo. Lancet 344 (1994) 1049-52. Spielman, R.S. and Ewens, W.J.: The TDT and other family-based tests for linkage disequilibrium and association. Am J Hum Genet 59 (1996) 983-9. Spielman, R.S., McGinnis, R.E. and Ewens, W.J.: Transmission test for linkage disequilibrium: the insulin gene region and insulin-dependent diabetes mellitus (IDDM). Am J Hum Genet 52 (1993) 506-16. Spritz, R.A.: The genetics of generalized vi tiligo and associated au toimmune diseases. J Dermatol Sci 41 (2006) 3-10. Spritz, R.A.: The genetics of generalized vitili go and associated autoimmune diseases. Pigment Cell Res 20 (2007) 271-8. Stokowski, R.P., Pant, P.V., Dadd, T., Fereday, A., Hinds, D.A., Jarman, C., Filsell, W., Ginger, R.S., Green, M.R., van der Ouderaa, F.J. a nd Cox, D.R.: A genomewide association study of skin pigmentation in a South Asian population. Am J Hum Genet 81 (2007) 1119-32. 108

PAGE 109

Su, M.A. and Anderson, M.S.: Aire: an update. Curr Opin Immunol 16 (2004) 746-52. Sulem, P., Gudbjartsson, D.F., Stacey, S.N ., Helgason, A., Rafnar, T., Magnusson, K.P., Manolescu, A., Karason, A., Palsson, A., Thorleifsson, G., Jakobsdottir, M., Steinberg, S., Palsson, S., Jonasson, F., Sigurgeirss on, B., Thorisdottir, K., Ragnarsson, R., Benediktsdottir, K.R., Aben, K.K., Kiemen ey, L.A., Olafsson, J.H., Gulcher, J., Kong, A., Thorsteinsdottir, U. and Stefansson, K.: Ge netic determinants of hair, eye and skin pigmentation in Europeans. Nat Genet 39 (2007) 1443-52. Tazi-Ahnini, R., Cork, M.J., Gawkrodger, D.J ., Birch, M.P., Wengraf, D., McDonagh, A.J. and Messenger, A.G.: Role of the autoimmune re gulator (AIRE) gene in alopecia areata: strong association of a potentially func tional AIRE polymorphism with alopecia universalis. Tissue Antigens 60 (2002) 489-95. Tazi-Ahnini, R., McDonagh, A.J., Wengraf, D.A., Lovewell, T.R ., Vasilopoulos, Y., Messenger, A.G., Cork, M.J. and Gawkrodger, D.J.: Th e autoimmune regulator gene (AIRE) is strongly associated with vitili go. Br J Dermatol 159 (2008) 591-6. Teo, Y.Y., Fry, A.E., Clark, T.G., Tai, E.S. and Seielstad, M.: On the usage of HWE for identifying genotyping errors. Ann Hum Ge net 71 (2007) 701-3; author reply 704. The International HapMap Proj ect. Nature 426 (2003) 789-96. Thomas, S., Thomas, M., Wincker, P., Babarit, C., Xu, P., Speer, M.C., Munnich, A., Lyonnet, S., Vekemans, M. and Etchevers, H.C.: Huma n neural crest cells display molecular and phenotypic hallmarks of stem cells Hum Mol Genet 17 (2008) 3411-25. Tobin, D.J., Swanson, N.N., Pittelkow, M.R., Peters, E.M. and Schallreuter, K.U.: Melanocytes are not absent in lesional skin of long duration vitiligo. J Pathol 191 (2000) 407-16. Tursen, U., Kaya, T.I., Erdal, M.E., Derici, E., Gunduz, O. and Ikiz oglu, G.: Association between catechol-O-methyltr ansferase polymorphism and vitiligo. Arch Dermatol Res 294 (2002) 143-6. Turunen, J.A., Wessman, M., Forsblom, C., Kilpikari, R., Parkkonen, M., Pontynen, N., Ilmarinen, T., Ulmanen, I., Peltonen, L. and Groop, P.H.: Association analysis of the AIRE and insulin genes in Finnish type 1 diabetic patients. Immunogenetics 58 (2006) 331-8. Uhm, Y.K., Yoon, S.H., Kang, I.J ., Chung, J.H., Yim, S.V. and Lee, M.H.: Association of glutathione S-transferase gene polymorphisms (GSTM1 and GSTT1) of vitiligo in Korean population. Life Sci 81 (2007) 223-7. van den Wijngaard, R., Wankowicz-Kalinska, A., Pals, S., Weening, J. and Das, P.: Autoimmune melanocyte destruction in vitili go. Lab Invest 81 (2001) 1061-7. 109

PAGE 110

Venneker, G.T., Westerhof, W., de Vries, I.J., Dr ayer, N.M., Wolthers, B.G., de Waal, L.P., Bos, J.D. and Asghar, S.S.: Molecular heterogeneity of the fourth component of complement (C4) and its genes in vitiligo. J Invest Dermatol 99 (1992) 853-8. Wakamatsu, K., Ito, S. and Fujita, K.: Production, ci rculation, and excretion of melanin-related metabolites in B16 melanoma-bearing mice. Acta Derm Venereol 70 (1990) 367-72. Wankowicz-Kalinska, A., Le Poole, C., van den Wijngaard, R., Storkus, W.J. and Das, P.K.: Melanocyte-specific immune response in melano ma and vitiligo: two faces of the same coin? Pigment Cell Res 16 (2003) 254-60. Weiss, S.T., Silverman, E.K. and Palmer, L.J.: Case-control association studies in pharmacogenetics. Pharmacogenomics J 1 (2001) 157-8. Xie, Z., Chen, D., Jiao, D. and Bystryn, J.C.: Vi tiligo antibodies are not directed to tyrosinase. Arch Dermatol 135 (1999) 417-22. Yildirim, M., Baysal, V., Inaloz, H.S., Kesici, D. and Delibas, N.: The role of oxidants and antioxidants in generalized vit iligo. J Dermatol 30 (2003) 104-8. Yu, H.S., Chang, K.L., Yu, C.L., Li, H.F., Wu, M. T. and Wu, C.S.: Alterations in IL-6, IL-8, GM-CSF, TNF-alpha, and IFN-gamma releas e by peripheral mononuclear cells in patients with active vitiligo. J I nvest Dermatol 108 (1997) 527-9. Yu, H.S., Kao, C.H. and Yu, C.L.: Coexistence and relationship of antikeratinocyte and antimelanocyte antibodies in patients with nonsegmental-type vitiligo. J Invest Dermatol 100 (1993) 823-8. Zou, G.Y. and Donner, A.: The merits of testing Hardy-Weinberg equilibrium in the analysis of unmatched case-control data: a cautiona ry note. Ann Hum Genet 70 (2006) 923-33. 110

PAGE 111

BIOGRAPHICAL SKETCH Deborah Marsha Herbstman was born in Champa ign-Urbana, Illinois to Drs. Joe I. and Barbara H. D. Herbstman. She grew up in the beautiful island paradise of Gainesville, Florida with her older brother, Joshua Tobias Herbst man. A graduate of F. W. Buchholz High School, she received her Associate of Arts with High H onors from the University of Florida. In 2001, she graduated from New College, the Honors College of the State of Florida, with a degree in biology/chemistry. 111