GENETIC AND FUNCTIONAL ANALYSIS OF SUSCEPTIBILITY
TO LUPUS IN B6.NZMc7 CONGENIC AND SUBCONGENIC MICE
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
This dissertation is dedicated
to my daughters
I would like to thank the chairman of my committee, Dr. Edward Wakeland, for
his valuable instruction and encouragement in helping me finish my project. I am also
grateful to my committee members, Drs. Mark Atkinson, Phil Laipis, Joel Schiffenbauer,
and Jin Xiong She, for their helpful discussion and guidance.
I also thank Drs. Laurence Morel and Chandra Mohan, for their initiation of the
project and experiments and helpful advice. I thank Dr. Mohan for help with my
dissertation and discussion. I also appreciate help from Dr. Byron Croker with histology.
I thank Jocelyn Tulsian for daily mouse care. I thank Karen Achey for help with
sequencing TGF-B. I would like to extend my special thanks to Drs. sally Litherland and
Christy Myrick for help with my dissertation.
I have enjoyed my friendship with other members of the Wakeland lab, especially
Nana Tian, Christy Myrick, Alicia Deng, Ping Yang, Guangling Wang, Eric Liang and
Kim Blenman. They have enriched my life immeasurably for the past six years. I want to
thank my mother for her encouragement and support. Lastly, I am grateful to my
husband, Fangshu Yu, for his understanding and encouraging, and for his help and love
throughout the years.
TABLE OF CONTENTS
ACKNOWLEDGMENTS ...................................................................... iii
ABSTRACT ....................................................................................................................... vi
1 INTRODUCTION ............................................................................................... 1
Systemic Lupus Erythematosus (SLE) Overview ................................................ 1
The Causes of Human Systemic Lupus Erythematosus ................................. 2
New Zealand Black (NZB), New Zealand White (NZW), and New
Zealand Mixed (NZM) Mice .................................................................... 5
Genetic Analysis of Susceptibility to SLE in Murine Models ....................... 6
The Contribution of MHC and Non-MHC Genes to Murine Lupus
Susceptibility .......................................................................................... 10
Inheritance of SLE Susceptibility as a Threshold Liability ............................... 11
B Cell Activation and Tolerance in Lupus ................................................... 12
The Role of CD4 T Cells in the Pathogenesis of SLE ................................. 14
Autoantibody Production and Lupus Nephritis ............................................ 16
Rationale for This Study ............................................................................... 17
2 MATERIALS AND METHODS ................................................................. 20
M ice ................................................................................................................... 20
Genotypic Analysis ...................................................................................... 21
Measurement of Antibodies .......................................................................... 22
Flow Cytometric Assay .................................................................................. 23
T Cell Proliferation Assay ............................................................................ 24
Immunization ............................................................................................... 25
Histology ...................................................................................................... 26
Sequencing TGF-fl Coding Region ............................................................... 26
Statistics ........................................................................................................ 27
3 GENETIC DISSECTION OF PATHOGENESIS OF SLE: SLE3 ON
MURINE CHROMOSOME 7 LEADS TO A BROAD, LOW-
GRAD E AUTOIM M UNITY ................................................................. 29
Introduction .................................................................................................... 29
Results ................................................................................................................ 32
Discussion ...................................................................................................... 39
4 GENETIC DISSECTION OF SLE PATHOGENESIS:
SLE3 HAS TWO SEPARATE LOCI THAT INTERACT
SYNERGISTICALLY ............................................................................. 71
Introduction .................................................................................................... 71
Results ................................................................................................................ 74
Discussion ...................................................................................................... 78
5 SUM M ARY ..................................................................................................... 107
LIST OF REFEREN CES ................................................................................................. 112
BIOGRAPHICAL SKETCH ........................................................................................... 127
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
GENETIC AND FUNCTIONAL ANALYSIS OF SUSCEPTIBILITY
TO LUPUS IN B6.NZMc7 CONGENIC AND SUBCONGENIC MICE
Chairman: Edward K. Wakeland, Ph. D.
Major Department: Pathology, Immunology and Laboratory Medicine
This study dissects the complex traits of lupus by constructing and analyzing
B6.NZMc7 congenic and subcongenic mice that bear the whole or part of NZM2410-
derived lupus susceptibility interval Sle3.
B6.NZMc7 mice had high serum IgG and IgM levels and generated spontaneous
IgM antibodies against a variety of antigens, suggesting polyclonal B cell activation.
These mice produced IgG anti-nuclear antibodies (ANA) starting at an early age. The
autoantibody titers were lower than parental NZM2410 and penetrance of this phenotype
was approximately 70%. There was no difference between male and female mice in IgG
ANA production. The humoral autoimmune phenotypes were also expressed in
heterozygous mice with delayed onset and decreased penetrance, suggesting that SLE
susceptibility is inherited in a dominant and allele-dose dependent fashion.
One important finding is that Sle3, by itself, can lead to the development of
glomerulonephritis (GN). B6.NZMc7 mice developed chronic GN at a late age with about
40% penetrance. The presence of anti-dsDNA and GN was not strongly correlated. Two
out of nine GN positive animals did not produce IgG anti-DNA or anti-chromatin
autoantibodies. Other specific autoantibodies may be related to GN, and this needs further
The underlying immunological abnormalities were also examined in these mice.
Young B6.NZMc7 mice (1-3 month) were used with age- and sex-matched C57BL/6J
(B6) mice as controls. B6.NZMc7 mice exhibited a high percentage of CD4 T cells and
an increased CD4:CD8 T cell ratio in both the spleen and lymph node. A high proportion
of CD4 T cells were activated and expressed the activation marker CD69. These CD4 T
cells expanded prominently in vitro when stimulated with anti-CD3. B6.NZMc7 mice
also showed heightened lymphocyte proliferative responses to anti-CD3 and IL-2 and
exhibited reduced activation-induced T cell death. Consistent with these results in vitro,
B6.NZMc7 mice exhibited high immune responsiveness in vivo when they were
challenged with the T-dependent antigen DNP-KLH. These results suggest that Sle3
mainly impacts T cell activation, proliferation, and apoptosis.
Two subcogenic mouse strains were also generated, one carrying the centromeric
end (B6.NZMc7c) and the other carrying the telomeric end (B6.NZMc7t) of Sle3. The
mice from both the subcongenic strains had polyclonal B cell activation, however, they
produced very low levels of IgG ANA. Penetrance of IgG ANA production is about 20%
in the subcongenic mice, compared with 70% in B6.NZMc7. These results indicate both
loci are required in IgG ANA production. Only B6.NZMc7t mice showed T cell
hyperactivity, which suggests that the gene affecting T cell function is located on the
telomeric end of Sle3. Both strains developed chronic GN, but severity is less compared
to B6.NZMc7. These results suggest that at least two susceptibility genes are present in
the Sle3 interval and that these two genes interact synergistically to cause the
autoimmune phenotypes expressed in B6.NZMc7.
Systemic Lupus Erythematosus (SLE) Overview
SLE is a common, chronic autoimmune disease. It is estimated that over half to
three quarters of a million people suffer from it in the United States. The clinical features
of SLE are very unique. It is a multisystemic disease with the potential to involve
multiple organs, including the skin, joints, serosal surfaces, kidney, heart, lungs, and
central nervous system. The clinical manifestations are extremely diverse depending
largely on which organ system, or systems, is affected. It is variable ranging from skin
rash and joint pain with a mild systemic illness to a fulminant presentation with life-
threatening involvement of the important organs such as kidney (reviewed in refs.1, 2).
SLE occurs mostly in females, and the female to male ratio is about 9:1.
The production of autoantibodies against a spectrum of nuclear antigens is the
hallmark of SLE both in humans and mice. The deposition of circulating antigen-
antibody complexes along the vascular basement membranes of various target organs is
one of the notable pathologic features. However, the mere presence of the autoantibody
does not necessarily lead to autoimmune disease, other immunologic abnormalities are
needed to develop full-blown clinical autoimmune disease.
No single cause has yet been identified, although genetic, environmental and
stochastic factors contribute to the altered immune state (1). The major focus of SLE
research is to understand its cause. It is hoped that the basic and clinical research will
elucidate a plausible etiology, will give direction to disease prevention, and will provide
more specific treatment of the disease.
The Causes of Human Systemic Lupus Ervthematosus
It is well documented that genetic factors are key elements involved in
susceptibility to SLE (3-7, reviewed in Refs 8-9). SLE is concordant in 57% of
monozygotic twin and 5% of dizygotic twins (10) which strongly supports the hypothesis
of a multifactorial etiology with incomplete penetrance, but the characteristics of the
inherited susceptibility factors are not known. Available data indicate the involvement of
multiple genes (10-18). The expression of such genes could involve variables influencing
immune responses, immune regulation, complement, and/or immunoglobulin levels.
Studies of SLE in humans have found several genetic markers associated with
SLE: major histocompatibility complex (MHC) class I allele, B8; MHC class II alleles,
DR2, DR3, DQwl; MHC class III alleles, C4A null phenotype, C2 deficiency.
A significant association of SLE with a particular MHC allele has been
determined. Increased frequencies of MHC class II DR2 or DR3 alleles are associated
with SLE in Caucasian patients (15-18). Further studies indicated the extended haplotype
Al-B8-DR3 is associated with SLE susceptibility (19). Griffing et al. demonstrated that
67% of SLE patients with anti-dsDNA antibodies were DR3 (20). The role of MHC class
II antigens in the pathogenesis of SLE is unknown. Two potential mechanisms are of
particular interest: (a) selective presentation of foreign or self-antigenic peptides by the
polymorphic regions of MHC class II molecules; (b) negative and positive selection of T
cells by MHC class II molecules during maturation in the thymus.
Inherited deficiencies of several complement components have been associated
with lupus-like illnesses. The most common is a deficiency of the second and fourth
components of complement (C2, C4) (11-14). The class III, or complement, region of the
human MHC is located between the class I (B, C, A) and class II (DP, DQ, DR) regions
on the short arm of chromosome 6. The four complement structural genes in this cluster
are C4A, C4B, factor B, and C2. These four components are required for activation of the
complement pathways, C4 and C2 for the classical and factor B for the alternative
pathway. One of the major biological functions of these early components of the classical
pathway, C4 and C2, is to process immune complexes. Complement system dysfunction
may contribute to impaired clearance of immune complexes (21-23).
Non-MHC linked genes were also suggested from studies of two very interesting
families with multiple instances of undifferentiated autoimmune disease and no relation
with any HLA haplotype (24). A comprehensive linkage analysis has not yet been
reported in humans to show which specific chromosomes and regions are involved. The
nature of non-MHC genes in SLE is unknown.
Lupus is a polygenic disease with added contributions from environmental and
stochastic variance. The fact that the concordance rate in identical twins is not 100% can
be used to estimate environmental and stochastic contribution (25-27). Inheritance and
the expression of susceptibility genes predispose individuals to the development of
phenotypes, but phenotypic expression requires triggering from environmental and
stochastic events. Several environmental factors including ultraviolet light, bacterial and
viral infections, and drugs are capable of inducing or exacerbating lupus. Random events
in the development of the immune system, such as the generation of T cell receptor
(TCR) repertoire, also play a role in the generation of SLE-like syndrome. It is speculated
that these factors may act through different mechanisms to alter immune function.
Susceptibility to SLE is inherited as a complex trait characterized by incomplete
concordance of phenotype and genotype. Several features of the disease make genetic
analysis of SLE susceptibility in humans very difficult. First, in a family, not all
individuals who have the same inherited susceptibility alleles manifest disease
(incomplete penetrance). Second, in different families, different sets of susceptibility
alleles may produce identical phenotypes (genetic heterogeneity). Third, several
susceptibility alleles may be required for fully expressing certain phenotypes (polygenic
Development of animal models for this disease can be advantagous for elucidating
the pathogenic mechanisms of the disease and will help studies in human SLE. Animals
can be bred in large quantities and manipulated for genetic analysis. They can also be
maintained in the same facility, thereby minimizing environmental factors.
New Zealand Black (NZB). New Zealand White (NZW), and New Zealand Mixed
Several inbred mouse strains that spontaneously develop a disease similar to
human SLE have been important models (1). Over the past three decades, extensive
studies of these murine models, particularly in the area of molecular genetics, have
provided a great deal of information to help our understanding of the disease. They have
also lead us to a better understanding of human SLE and brought us closer to the ultimate
goal of unraveling the mystery of autoimmune diseases.
The mouse strains that have been extensively studied include (NZB X NZW)F1,
(NZB X SWR)F1, BXSB, and MRL-lpr/lpr (28). These strains produce high levels of
antinuclear antibodies (ANAs) and develop lupus-like glomerulonephritis (GN). The New
Zealand black (NZB) (H-2d) strain was initially bred for coat color by M. Bielschowsky
at the University of Otago Medical School in New Zealand (29). The genetic background
of this strain is unknown. NZB mice develop a disease state characterized by an
autoimmune hemolytic anemia (28, 30-31). The lupus-like nephritis is less severe in NZB
mice than in its hybrids. NZB mice produce ANAs, but not high levels of IgG antibodies
to double stranded DNA (dsDNA) or histones. The New Zealand White (NZW) (H-2z)
strain was developed by Hall at the Otago Medical School Animal Breeding Station (31),
and its genetic background is also unknown. NZW mice are, for the most part,
phenotypically normal. (NZB X NZW)F1 hybrid mice, first described by Helyer and
Howie (32), develope florid renal dysfunction and failure similar to human lupus
nephritis. The hybrid mice also develop autoantibodies to a number of nuclear
components including ssDNA, dsDNA, histone and chromatin. The F 1 hybrids of NZB
and NZW provided the first experimental model of human SLE and have been studied
extensively by a large number of investigators. Other two murine strains that also
spontaneously develop a SLE-like syndrome are MRL and BXSB. They were developed
by Murphy and Roths at the Jackson Laboratory (33-34). Addition of these strains has
greatly enhanced the potential value of murine SLE as an experimental model and has
broadened our understanding of autoimmunity.
The New Zealand Mixed (NZM)(H-2z) strains were produced and characterized
by Rudofsky (35). They are a collection of inbred strains derived from selective
inbreeding of progeny from a cross between NZB and NZW mice. These strains have
inherited various genomic segments from the two parental strains, express a board
spectrum of SLE-related autoimmune phenotypes and show different susceptibilities to
lupus. One strain, NZM2410, produces high levels of IgG autoantibodies to nuclear
antigens as early as 4-6 months of age and develops severe lupus-like renal disease in
Genetic Analysis of Susceptibility to SLE in Murine Models
Analysis of autoimmune phenotypes and lupus-like renal disease in (NZB X
NZW) Fl, F2 and backcrosses between NZB and NZW mice illustrates the genetic
complexity and polygenic inheritance of the disease. Earlier genetic studies had
limitations and were not able to assign chromosomal locations of susceptibility genes,
however, these classic studies still provide valuable data. Braverman (36) crossed NZB
and NZW and observed varying severities of renal histopathological changes in the F2
mice, which suggested the involvement of multiple genes from both NZB and NZW.
Knight & Adams (37-38) performed (NZB X NZW)F1 X NZW and (NZB X NZW)F1 X
NZW backcrosses and concluded that NZB contributes a single dominant gene and NZW
contributes two genes to the development of severe renal disease and death. These
investigators also concluded that one of the two NZW genes might be linked to the MHC.
Shirai and colleagues studied the genetics of (NZB X NZW)F1 traits using backcross
analyses and suggested that two dominant genes from NZB and one dominant gene from
NZW contribute to the formation of gp70-anti-gp7O immune complexes (39). Kotzin and
colleagues reported one dominant gene from NZW contributes to renal disease and
suggested a strong correlation between the NZW MHC haplotype, H-2z, and the
phenotype (40-41). Hirose and coworkers extended these observations by constructing
NZB (H-2d) and NZW (H-2z) congenic strains with the reciprocal H-2 and comparing
(NZB.H-2z X NZW)F 1 and (NZB X NZW.H-2d) F I to original (NZB X NZW)F 1. Their
results suggested that heterozygosity at the MHC might be required for disease
development (42). From these studies, two to three genes from NZB and one to two genes
from NZW were estimated to contribute to the disease. The number of genes involved
could be dependant on the particular trait or traits examined as well as the genetic
Recent advances in molecular biology techniques have provided new means to
screen the whole genome for susceptibility loci and to assess the involvement of available
candidate genes. The identification of large numbers of highly polymorphic simple
sequence repeats spaced throughout the genome provide the basis for mapping studies
(43-45). The polymerase chain reaction (PCR) method makes it possible to genotype
large numbers of animals (46), and the development of sophisticated computer programs
makes data analysis quick and accurate. These techniques permit the identification of
multiple loci contributing to a given trait without requiring preexisting knowledge of
candidate loci or their positions. Several groups have used these tools to perform genome-
wide searches and to identify susceptibility loci responsible for lupus nephritis and a
variety of lupus-associated autoimmune phenotypes.
Studies of the (MRL-Ipr X CAST/EI)F1 X MRL-lpr backcross mice identified
two loci associated with renal disease, Lrdml on chromosome 7 and Lrdm2 on
chromosome 12 (47). The study of (NZB X NZW)F2 intercross mice by Kono et al. (48)
identified eight susceptibility loci (Lbwl to -8 on chromosomes 17, 4-7, 18, 1, and
11 respectively) associated with anti-chromatin autoantibody production (Lbwl, Lbw7,
Lbw8), GN (Lbw], Lbw2, Lbw6), and/or mortality (Lbwl-6). Drake et al. investigated
NZB contribution to lupus-like renal disease. Their work with (NZB X NZW)Fl X NZB
backcross showed that a single NZB locus (nbal) on chromosome 4 was associated with
renal disease and death (49). This group also performed genetic analysis in (NZB X
SM/J)F1 X NZW mice and found six NZB loci (on chromosomes 1, 4, 7, 10, 13, 19)
contributing to renal disease (50). They confirmed two loci previously identified (nbal on
chromosome 4, and nba2 on chromosome 1) and detected two new loci (on chromosomes
11 and 14) in a (BALB.H2Z X NZB)F1 X NZB cross (51).
Morel et al. (52) identified three susceptibility intervals and the MHC locus
associated with GN in the NZB/W-derived NZM2410 lupus-prone strain backcrossed to
B6. The three susceptibility loci are Slel on the telomeric end of chromosome 1, Sle2 on
the centromeric end of chromosome 4, and Sle3 on the centromeric end of chromosome 7.
Genetic analysis indicated several ways in which susceptibility genes could influence
overall disease threshold liability in lupus. First, each of these susceptibility alleles
independently contributed to the overall GN susceptibility in an additive fashion, which is
commonly described as a threshold of genetic liability. Second, several susceptibility
alleles are needed to express high frequency of GN, but none of these alleles are
absolutely required. This is an important feature of a polygenic disease. Third, although
the mice express an identical array of susceptibility genes, the proportion of NZM2410
mice developing GN and penetrance of GN increased with age, indicating that the
expression of complex traits requires "triggering events" (53).
In summary, genetic mapping studies have identified 22 susceptibility loci that are
linked to autoimmune phenotypes of lupus. These studies illustrate the complexity of the
genetics of SLE. Within each cross, some loci were associated with a single distinct
phenotype, suggesting they may act on different pathways and contribute specific
component phenotypes to disease susceptibility. Some loci contributed to multiple
phenotypes, suggesting they play roles in a common pathway linked to multiple
phenotypes. Interestingly, susceptibility loci on chromosome 1, 4, and 7 have been
reported by three groups in different crosses, indicating these susceptibility genes may be
common to the lupus pathogenesis in NZB and NZW strains. The number and location of
genes detected are different depending on which affected strains and resistant strains are
used and how they are crossed. This difference may represent the epistatic interactions of
susceptibility genes, resistance genes, and background genes. Clear demonstration of the
nature of each locus and the relationship between loci would require the production of
congenic strains carrying different single loci and polycongenic strains.
The Contribution of MHC and Non-MHC Genes to Murine Lupus Susceptibility
The contribution of certain MHC alleles to lupus susceptibility has been well
documented both in human and mouse models. Such an association is reasonable because
autoimmune diseases are T cell dependent and all T cell mediated responses are MHC
restricted. The MHC may affect predisposition to autoimmune disease by shaping the T
cell receptor repertoire, selecting and presenting peptide to T cells. In studies with crosses
of NZB X NZW or NZB X SWR mice, heterozygosity at the MHC (H-2/ or H-2/q)
conferred greater susceptibility to lupus than homozygous parental haplotypes (54-57).
Antibodies to class II molecules can prevent the development of the disease (58). Further
support for MHC class II as a disease predisposing molecule derives from a report in
which a well-studied I-AmI2 3 chain mutation was bred onto the NZB background (59).
NZB mice congenic for this class II mutation (NZB.H-2m'2) developed anti-dsDNA
antibodies and lupus-like renal disease, similar to (NZB X NZW)F1 mice, whereas NZB
mice congenic for H-2b (NZB.H-2b) did not express any signs of autoimmune disease.
The I-AP sequences of H-2"2 and H-2b differ only three amino acids in the peptide-
binding groove at positions 67, 70, and 71. This three amino acid substitution within the
MHC class II molecule is capable of altering the NZB phenotype to one of severe lupus-
like renal disease. It has been hypothesized that the novel hybrid MHC class II molecules
generated by a and P chains from different haplotype to pair contribute to autoimmunity
Although the MHC has a major impact on lupus susceptibility, it is not required
and insufficient by itself for the development of the disease. Association of H-2Z with the
disease is imperfect: 14% of heterozygous H-2z backcross mice did not develop the
disease, whereas 12% of homozygous H-2d backcross mice did (61). NZB mice crossed
with normal background BALB/c mice congenic for H-2' do not develop disease (62).
This work illustrates that both MHC and non-MHC genes are required. Genome-wide
searches using microsatellite-defined chromosomal maps have identified multiple non-
MHC loci responsible for predisposing to lupus.
Inheritance of SLE Susceptibility as a Threshold Liability
The SLE is a polygenic disease requiring multiple genes. How multiple genes
interact to cause a disease susceptibility phenotype is not known. Data from a (NZM2410
X B6)F1 X NZM2410 backcross showed that the increase in GN frequency correlated
with an increase in the number of susceptibility loci carried by individual progeny (52).
These results suggest that the disease is inherited as a threshold of genetic liability. This
conclusion is supported by work done by other groups using different backcrosses to
study lupus (50), diabetes (63), and experimental allergic encephalomyelitis (64-65). The
threshold liability hypothesis is based on the concept that several disease susceptibility
genes contribute equally to the overall liability of developing disease and that the
probability and onset of an individual developing disease is proportional to the number of
susceptibility genes inherited. We further demonstrated this relationship by analysis of
three congenic strains that contain only one susceptibility locus from NZM2410 (66).
Each congenic strain expressed distinctly different component phenotypes and multiple
genes were necessary to obtain full expression of SLE.
B Cell Activation and Tolerance in Lupus
The primary function of B lymphocytes is to make antibodies against components
of foreign organisms. Self-reactive B cells are eliminated at multiple checkpoints along
the B cell lineage (reviewed in ref. 67). Each of these checkpoints has triggering
thresholds that impose limits on the extent of tolerance. This is a kind of balance system.
On the one side, such limits ensure that a sufficient fraction of the repertoire is preserved
to mount rapid and effective antibody responses to foreign antigens. On the other side, the
same limits increase the risk of autoimmune disease. A variety of events, either
exogenous or endogenous, could tip the delicate balance so that harmful autoreactivity is
the outcome. If this imbalance is permanent, chronic autoimmune phenomena such as
SLE may occur.
The fundamental event in the pathogenesis of lupus is B cell hyperactivity, which
is responsible for the hypergammaglobulinemia, spontaneous polyclonal antibody
production, and secretion of various autoantibodies (68). Polyclonal B cell activation is
the earliest and most common immunological abnormality of NZB, (NZB X NZW)F1,
and other lupus-prone strains preceding the pathogenic autoimmune response (69-71).
The pathogenic autoantibody-producing B cells are driven by self-antigens (reviewed in
ref. 72). This subset of B cells are clonally expanded and their immunoglobulin genes are
modified by somatic mutation. IgG autoantibody production in lupus is selective for
certain self-antigens and there is strong evidence that B cells are intrinsically defective in
murine lupus-prone mice. SCID mice populated with long-term pre-B cells from (NZB X
NZW)F1 mice, but not those populated with (B6 X DBA/2)Fl pre-B cells, developed
hypergammaglobulinemia and increased IgM and IgG autoantibodies to nuclear antigens,
and 20% of these animals also developed mild nephritis (73). Further studies using pre-B
cell lines derived from the parental NZB and NZW strains suggest that B-lineage cells
from both NZB and NZW parental strains manifest abnormalities associated with the
development of lupus-like disease (74).
It is understandable that errors in central or peripheral tolerance at the T or B cell
level can cause systemic autoimmune disease. However there is scant evidence for
defects in B cell tolerance in New Zealand mice (75). Studies indicate that potentially
pathogenic B cells are part of the peripheral repertoire and they can be expanded by graft-
versus-host reactions (76-77). Normal animals can generate autoantibody responses after
immunization with self-peptides (78-80). Thus, defects in central B cell tolerance are
unlikely to be necessary to allow for lupus-like autoantibody production. On the other
hand, defects in the regulation of B cell apoptosis have the potential to lead to
development of autoreactive B cells and increase autoantibody production. There are
some reports of a defect in B cell apoptosis induction in New Zealand mice (81-82). The
importance of this factor in pathogenesis of the disease needs further study.
The Role of CD4 T Cells in the Pathogenesis of SLE
The association of SLE with particular MHC class I alleles, class switching, and
affinity maturation of autoantibody production in this disease strongly suggest that CD4
T cells are important in pathogenesis of the disease. Evidence for the involvement of CD4
T cells includes: the autoinimune condition of (NZB X NZW)Fl mice is improved
following treatment with anti-CD4 antibodies (83); both T and B cells are needed in order
to induce anti-dsDNA IgG synthesis (84); the incidence of lupus nephritis is delayed and
reduced after blocking T cell-B cell interaction (85). Certain CD4 T cells have been
shown to drive pathogenic B cells to selectively secrete anti-DNA autoantibodies in
murine models and in humans (86-89). The major subclass of pathogenic anti-DNA
autoantibodies is IgG2a and IgG2b which is also consistent with CD4 T cell help.
Although an anti-DNA response is the hallmark of lupus, efforts to induce such
autoantibodies by immunization with native (n) DNA have generally been unsuccessful.
The determinants of autoreactive CD4 T cells may reside on different molecules. CD4 T
cell clones specific for anti-nDNA antibody induction were shown to proliferate in the
presence of nucleosomes but not their constituent nDNA or histones (90). These CD4 T
cells may be specific for nucleosomes. Thus, for anti-nDNA antibody induction, CD4 T
cells may provide intermolecular help to anti-DNA specific B cells via the binding of
nucleosomes or other protein-DNA complexes.
Loss of central or peripheral T cell tolerance will cause systemic autoimmune
disease. Studies in lupus mice have suggested that high affinity responses to self-antigens
are tolerated in the thymus and normal thymic deletion of superantigen reactive V3-
bearing T cells have been observed (91-92). T cells specific for dominant determinants of
self-antigens are deleted efficiently in thymus, but T cells directed to cryptic self-
determinants may escape tolerance induction and be part of the normal peripheral T cell
repertoire (reviewed in ref. 93). Therefore, abnormal presentation of cryptic self-
determinants in the periphery or a defect in peripheral T cell tolerance in lupus could
activate autoimmune T cells.
Autoantibody Production and Lupus Nephritis
The hallmark of SLE in human and murine models is elevated serum levels of
antibodies to nuclear antigens. There is substantial evidence that anti-dsDNA antibodies
play an important role in the pathogenesis of renal disease in lupus mice. First, anti-DNA
autoantibodies are found concentrated in kidney elutes. Second, monoclonal anti-DNA
antibodies derived from lupus prone mice form glomerular immune deposits and induce
different structural and functional abnormalities in vivo after transfer into normal or
preautoimmune mice (94). Third, mouse IgG anti-DNA autoantibodies can bind to
isolated perfused kidney and induce kidney dysfunction (95). Fourth, genetic studies of
lupus mice have shown a direct relationship between anti-DNA antibody level, GN, and
mortality (96). However, not all anti-DNA autoantibodies are pathogenic. The cationic
charge, isotype, and cryoprecipitability will determine the pathogenic potential of anti-
DNA antibodies (97).
The pathogenic anti-DNA antibodies induce end-organ damage by tw
mechanisms (98). One is that anti-DNA antobodies directly or indirectly bind to the
glomerulus via heparan sulfates, the major glycosaminoglycan constituent of the
glomerular basement membrane (99-10 1). Another is that anti-DNA antobodies first form
immune complexes and then deposit onto anionic sites on glomerular basement
Although most studies have focused only on anti-DNA antibodies in the cause of
lupus nephritis, considerable evidence indicates that autoantibodies to non-DNA antigens
are also important. Genetic studies have suggested a pathogenic role for autoantibodies to
the endogenous retroviral glycoprotein (gp70) and gp70-anti-gp7O immune complexes in
the (NZB X NZW)Fl model (105). Relatively large amounts of the retroviral
glycoprotein gp70 circulate in the blood of both lupus and non-lupus mice, but only lupus
strains spontaneously produce autoantibodies against this molecule and contain gp70-
anti-gp70 complexes in their serum (106). However, studies on sublines of MRL/lpr mice
specifically bred for low gp70 levels indicated that autoimmune disease, including GN, is
essentially unchanged (107). Thus, although gp70-anti-gp7O deposits may be involved in
the pathogenesis of lupus nephritis, they appear not to be necessary.
Rationale for This Study
SLE is a common, chronic disease. It is estimated that over half to three quarters
of a million people suffer from it in the United States alone, while in other parts of the
globe the numbers are even higher (108). The cause of SLE is unknown. The central
thrust of SLE research is to understand its cause, especially genetic predisposition. If
genes contributing to the susceptibility of SLE can been identified: (a) people with high
risk could be screened and those with susceptibility genes could be started on
preventative therapies, (b) diagnosis will be more easy and accurate, (c) treatment will be
specific according to individual genetic defects, and (d) it will help us understand the
fundamental immunology of self and non-self discrimination.
Because the inheritance of SLE is complex, identification of genes involved in
SLE susceptibility in outbred human populations will be very difficult. Animal models,
such as (NZB X NZW)F1, readily allow experimental crosses to identify susceptibility
genes. Genetic studies in murine models will help to identify SLE susceptibility genes in
humans by comparing the syntenic regions of human and mouse chromosomes and by
providing information about pathogenic pathways.
Success has been made in the identification of genes responsible for simple
Mendelian traits wherein identification is fairly straightforward: linkage analysis,
positional cloning, and detection of mutation (109). Yet many of the most important
medical conditions including heart disease, hypertension, diabetes, asthma, SLE, and
cancer are inherited as complex polygenic traits. The geneticist's challenge is now to
tease apart the multifactorial causes of these diseases.
Our lab focuses on studies of SLE using the lupus-prone murine model
NZM2410. The approaches and strategy used are reviewed by Wakeland (53). We have
identified three intervals (Slel on chromosome 1, Sle2 on chromosome 4, and Sle3 on
chromosome 7) and MHC locus, which confer strong susceptibility to GN (52).
Questions to be answered include what phenotypes each locus contributes, what is the
nature of each susceptibility gene, and how different loci interact with each other. We
analyzed congenic mice to address these questions. This research focuses on Sle3, the
lupus susceptibility interval on murine chromosome 7. B6.NZMc7 is one of four
congenic mouse strains. It carries the Sle3 interval on a resistant B6 background. Genetic
and phenotypic analysis was performed in B6.NZMc7 mice, including serologic studies,
cellular immunology studies, and histology studies (Chapter 3) to define the contribution
of Sle3 by itself and to explore the underlying mechanism by which Sle3 contributes to
lupus. Sle3 is a large interval (about 38 cM) and linkage studies suggest two susceptibility
loci may present in this interval. To explore this possibility, we produced two
subcongenic mouse strains bearing the centromeric end (B6.NZMc7c) or the telomeric
end (B6.NZMc7t) of Sle3 on a resistant B6 background (Chapter 4). These subcongenic
lines were characterized serologically, immunologically, and histologically for the
MATERIALS AND METHODS
All mice used in this study were bred and maintained under conventional
conditions in our colony in the University of Florida Department of Animal Resources.
Breeding pairs of C57BL/6J (B6) were originally obtained from the Jackson Laboratory
and NZM2410 breeders were obtained from the Wadsworth Center for Laboratory and
Research, New York State Department of Health (Albany, New York). Both male and
female mice were used for all experiments. Animals were age- and sex-matched for all
The B6.NZMc7 congenic strain bearing the Sle3 susceptibility interval on
chromosome 7 was one of four congenic strains generated (110). Figure 2-1 is a
schematic picture showing the breeding strategy used for the production of congenic
mice. Briefly, B6 was chosen as a resistant strain. NZM2410 was backcrossed six
successive times to B6. At each generation, progeny were selected using three simple
sequence repeats (SSR) markers (Ckmm, D7Mit69, D7Mit7O) on chromosome 7 to
maintain the Sle3 interval and two to three other markers per chromosome to minimize
background genes contamination. Finally, the mice were intercrossed to obtain the
homozygous Sle3 interval.
The B6.NZMc7c and B6.NZMc7t subcongenic mouse strains were produced by
crossing B6.NZMc7 heterozygotes to B6 to generate a series of recombinants.
Genotyping was performed in these recombinant progeny using SSR markers covering
the Sle3 interval. Two subintervals were chosed, one contains the centromeric end of Sle3
(from ckmm to D7Mit1 54), and other contains the telomeric end of Sle3 (from D7Mit69
to D7Mit 147). Progeny bearing these two subintervals were crossed to B6 again to
generate more animals with these specific recombinations. Finally, mice bearing these
heterozygous intervals were intercrossed to produce mice that were homozygous for these
Genomic DNA was isolated from mouse tail biopsies by standard
phenol:chloroform extraction. Primers flanking SSR were purchased from Research
Genetics (Huntsville, Ala.) or synthesized at the University of Florida. Polymerase chain
reaction (PCR) technique was used for genotyping. Genomic DNA (100ng) was
amplified in 20 ul reactions with 200 nM paired primers, 0.2 mM dNTPs, and 0.75 U of
Taq DNA polymerase in a standard incubation buffer containing 50 mM MgCI2.
Amplifications were carried out in a Perkin-Elmer/Cetus 9600 thermal cycler under the
following conditions: I cycle at 940 C for 2 min; 35 cycles at 94o for 30 sec, 50-620 C for
30 sec, and 720 for 30 sec; 1 cycle at 72 0 C for 7 min. The PCR products were detected
using 5% agarose gels containing ethidium bromide. Optimal annealing temperature was
determined for each primer pair. Each mouse was scored as NZM2410 homozygous, B6
homozygous, or heterozygous for each locus genotyped.
Measurement of Antibodies
Serum was collected bimonthly starting at two mo. of age by tail bleeding. Serum
was tested against a panel of antigens including dsDNA, ssDNA, chromatin or
histone/DNA complex, histone, DNP-KLH, and thyroglobulin using ELISA as previously
described (66,90). Briefly, Immulon II plates (Dynatech, Chantilly, VA) were coated with
each of following antigens: calf thymus dsDNA (50ug/ml); ssDNA (50ug/ml); chromatin
(lug/ml) (prepared as previous described in ref. 111); total histone (1Oug/ml); bovine
thyroglobulin (10ug/ml); DNP-KLH (lug/ml (Sigma Chemical Co., St. Louis, MO). The
plates were blocked with blocking buffer (3%BSA, 0.1%gelatin, 3mM EDTA in PBS).
Serum from test animals was diluted 1:100 and incubated on these plates for two hours.
Specific antibodies were detected using alkaline phosphatase-conjugated goat anti-mouse
IgG or IgM (Boehringer Mannheim) and pNPP substrate. Optical density (OD) values
were obtained with an automated spectrophotometer at 405 nM. All samples were run in
duplicate. To allow for interassay comparisons, OD values were converted to a relative
unit scale (U/ml) in reference to a standard curve obtained with serial dilutions of a
NZM2410 positive serum added in each ELISA plate. Alternatively, the results are
expressed as individual percentage (%OD) (i.e., the ratio of the absorbance from the test
sample divided by the absorbance of 3 mo B6 pooled serum x 100) (112). Positive
sample serum was defined as having antibody titers above the mean titers plus two
standard deviation of five 10-12 mo-old B6 samples on the same plate.
Serum total IgG and IgM were measured using a sandwich ELISA. Briefly,
Immulon I plates (Dynatech, Chantilly, VA) were coated with capture antibodies, anti-
mouse IgM and IgG (Boehringer Mannheim), and blocked with blocking buffer described
above. IgG or IgM standards (Cappel Research) and 1:100 dilution of serum were added
for two hours. Detection of antibodies was as described above.
Flow Cytometric Assay
Flow cytometric analysis (FACS) was performed as described previously (85). All
primary antibodies were obtained from Pharmingen (San Diego, CA) and used at pre-
titrated dilution. Single cell suspensions of the spleen and lymph nodes were prepared by
gently mincing the tissues in a serum-free DMEM using the frosted ends of two sterile
glass slide. Lysis of erythrocytes was made in hypotonic, buffered ammonium chloride
solution. Bone marrow cells were obtained by washing the cells out of the femur with a
needle mounted on a 1-ml syringe. Peritoneal cells were obtained by lavage of the
peritoneal cavity with the serum-free DMEM using a 3-ml syringe. Cell surface Fc
receptors were first blocked with staining medium (PBS, 5% horse serum, 0.05% azide)
containing 10% normal rabbit serum. Cells were then stained on ice with optimized
amounts of FITC, phycoerythrin, or biotin-conjugated primary monoclonal antibodies
diluted in staining medium for 30 min. Following two washes, biotin conjugated
antibodies were revealed using streptavidin-PE (Gibco BRL, Grand Island, NY) or
streptavidin-Quantum Red (Sigma Chemicals, St.Louis, MO). Cell staining was analyzed
using a FACScan (Becton-Dickinson, San Jose, CA). Dead cells were excluded on the
basis of scatter characteristics and propidium iodide uptake and 10,000 events were
acquired per sample. Data were analyzed using Lysis II software (Becton-Dickinson).
Anti-IgM-PE was obtained from Catag (South San Francisco, CA). PE-coupled
anti-I-A (M5/114) was purchased from Boehringer Mannheim (Indianapolis, IN). The
following dye- or -biotin-coupled antibodies were obtained from PharMingen (San
Diego, CA): CD4 (RM4-5), CD5 (53-7.3), CD8 (Ly-2), CD25 (7D4), CD80/B7-1 (16-
T Cell Proliferation Assay
1 x 106 erythrocytes-depleted spleen cells/well were cultured in 200 pl
Dulbecco's modified Eagle's medium (DMEM), supplemented with 10% heat-inactivated
horse serum, 0.1mM nonessential amino acids, lmM sodium pyruvate, 10 mM Hepes,
0.05 mM 2-ME, lOug/ml gentamicin, and anti-bacterial/anti-mycotic (ABAM) in flat-
bottomed 96-well tissue culture plates (Costar Corp.). Cells were stimulated with anti-
CD3 (1Oug/ml, lug/ml, or 0.1ug/ml) or rIL-2 (20U/ml, 1OU/ml, or 5Umi) for 24h, 72h,
120h. 1 tiCi of [3H] thymidine in 25.d culture medium was added to each well at 18h
before harvesting. Incorporated thymidine was measured by using standard liquid
scintillation counting techniques. Mean standard error of mean (SEM) of triplicate
samples was presented.
To assess CD4 and CD8 T cell proliferation in vitro, splenocytes were cultured at
1 x 106 cells/ 200ul DMEM as described above with anti-CD3 (lug/ml) for 24h, 48h, 72h.
The cells were counted, stained with FITC-anti-CD4 and PE-anti-CD8. FACS analysis
was performed as described above. The number of CD4 and CD8 T cells was determined
by total counted cells /ml multiplied by the percentage of CD4 and CD8 T cells.
B6 and B6.NZMc7 mice (Four to six per group, 2mo old, sex matched) were
injected intraperitoneally with DNP-KLH (100 ug/mouse; Calbiochem, La Jolla, CA) in
complete Freund's adjuvant (CFA) and boosted two weeks later with same dose of DNP-
KLH in incomplete Freund's adjuvant. Serum was collected before immunization (dO),
14 days after first immunization (d14), and 7 and 30 days after the booster (d21, d51).
Immunized mice were assessed for serum anti-hapten titers using ELISA. Serum was
diluted 500, 2500, 12500, 62500 fold in PBS containing 2% BSA and incubated on DNP-
BSA (Calbiochem, La Jolla, CA) coated Immulon II plates. Detection of antibodies was
as described above. Antibody activity is expressed as OD.
Mice were sacrificed at 12mo of age. Thymus, lung, heart, liver, lymph node,
spleen, and kidney were fixed in 10% neutral buffered formalin and embedded in
paraffin. Sections were cut, placed on slides, and stained with hemotoxylin, eosin, and
periodic acid-Schiff. Multiple sections from control and experimental strains were
examined by light microscopy for indication of inflammation and tissue damage in a
blind fashion. Glomerulonephritis (GN) was scored from 0 to 4, in which 1 = < 10% of
glomeruli affected; 2 = 11-25% of glomeruli affected; 3 = 26 50% of glomeruli affected;
and 4 = >50% of glomeruli affected. Control B6 mice had scores of 0 and 1. Glomerular
lesions were classified as mesangiopathic (M) and capillary hyaline (H). The H lesion
was considered more severe than the M lesion.
Sequencing TGF-A3 Coding Region
Total RNA was extracted from liver of B6.NZMc7 and B6 mice using a total
RNA isolation Kit (Ambion, Austin, Texas). cDNA was obtained by reverse-
transcriptase-PCR using random primers and total RNA. For sequencing the TGF-fi gene,
five pairs of overlapping primers were designed according to published mouse sequence
from Genebank to amplify 1.6kb of cDNA. Two mice were used from each strain and
PCR was performed as described above. Five segments of PCR products from each
mouse were cloned using a TA cloning kit (Invitrogen Corp., San Diego, CA).
Sequencing of two independent clones from each segment was performed in both
directions on an AB 1373 automated sequencer (Applied Biosystems, Foster City, CA).
A nonpaired Student's t-test was used to determine statistical significance of
quantitative data and a chi-square test was used for qualitative data.
Production of Congenic Strains
N1 (50% B6)
N2 (75% B6)
N3 (87.5% B6)
N4 (93.8% B6)
nb bb C57BI/6
b bb C57BI/6
N7 (99.2% B6)
I! I ''
nn nn nb nb bb
Figure 2-1. Illustration of the breeding strategy used in the construction of congenic
mouse strains (N=breeding generation; m=NZM2410 allele; b=B6 allele). In
parentheses is the average percentage of the mouse genome that should be B6-derived.
GENETIC DISSECTION OF PATHOGENESIS OF SLE: SLE3 ON MURINE
CHROMOSOME 7 LEADS TO A BROAD, LOW-GRADE AUTOIMMUNITY
SLE is a prototypic autoimmune disease characterized by the involvement of
multiple organ systems and females preponderance. The disease is strikingly
heterogeneous with regard to organ patterns, severity, clinical course, genetic variable,
and immunologic expressions (reviewed in refs. 1, 2). Although the etiology of SLE is
unclear, a combination of genetic factors, environmental factors, and stochastic events are
required to produce disease (reviewed in refs. 8, 9). The disease is a consequence of
processes involving loss of immune tolerance, pathogenic autoantibody production, and
tissue destruction mediated by direct autoantibody binding and/or the deposition of
immune complexes. The production of autoantibodies against nuclear antigens is the
hallmark of SLE.
The analysis of etiopathogenesis of SLE in humans is very difficult due to
polygenic control of the disease, low penetrance, heterogeneity, and non-organ specific
nature of this disease. Several inbred mouse strains that spontaneously develop disease
similar to human SLE have been important models for elucidating the pathogenesis of the
disease (28). (NZB X NZW)F l is the first and most intensively studied model. More than
90% of female (NZB X NZW)F1 mice develop severe lupus nephritis and die of renal
failure (30). These mice also produce high level autoantibodies against a number of
nuclear components. The New Zealand Mixed (NZM) strains are newer inbred strains
derived from selective inbreeding of progeny from crosses between NZB and NZW mice
(35). A total of 27 NZM strains were produced. These strains have inherited various
genomic segments from the two parental strains and they are all homozygous for the
NZW histocompatibility (H-2) haplotype (Ku, Au, Sz, DM). The genetic differences among
these strains lead to phenotypic differences in disease expression. The NZM2410 strain
produces high levels of IgG autoantibodies to nuclear antigens as early as four to six
months of age and develops severe lupus-like renal disease in both sexes.
Dr. Morel performed interval mapping in a (NZM2410 X B6) X NZM2410
backcross using highly polymorphic simple sequence repeats (SSRs) as genetic markers.
Three susceptibility intervals and MHC locus on chromosome 17 were identified as
conferring susceptibility to acute GN. The three intervals are Slel, on the telomeric end of
chromosome 1; Sle2, on the centromeric end of chromosome 4; and Sie3, on the
centromeric end of chromosome 7 (52). Similar regions have also been found to confer
SLE susceptibility in (NZB X NZW)F1 (48,50). Genetic analysis suggested that the
disease susceptibility is inherited as a multifactorial threshold liability. First, none of
these four loci are required for the development of acute GN. Second, each susceptibility
locus independently contributes to the overall incidence of GN with a relative small
effect. Third, at least three susceptibility alleles are required in order to express a high
level of GN.
Although interval mapping has provided information about the numbers and
chromosomal locations of the genes responsible for autoimmunity in NZM2410, it has
provided little about the nature of the component phenotypes contributed by each
susceptibility gene and it is not productive for finely mapping genes. The approach we
used following interval mapping was to move the four susceptibility intervals from
NZM2410 to the resistant strain B6 background using a speed congenic technique
(110,113). This congenic approach, which originated from the Nobel Prize-winning
studies of histocompatibility by George Snell (114-115), is designed to dissect a complex
trait into its component genetic elements and allow for the fine mapping,, and
identification and characterization of each gene. This approach has been used in the
genetic analysis of insulin-dependent diabetes in the NOD mouse, where either resistance
intervals were bred to a susceptible background (116-118) or susceptible intervals were
bred to a resistant background (119).
We recently produced a collection of B6 congenic strains bearing the individual
lupus susceptibility intervals from NZM24 10. Here we report the detailed analysis of the
B6.NZMc7 congenic strain and the component phenotypes expressed by Sle3. Sle3 is
likely a major susceptibility locus leading to pathogenesis of lupus because this locus
leads to two major phenotypes, ANA production and GN, and because this locus was
found in multiple SLE-related crosses (reviewed in ref. 120). To understand the role of
Sle3 in the pathogenesis of SLE, we examined spontaneous IgM and IgG responses to
self-nuclear antigens and foreign antigens in these mice from 2 to 12 months of age, and
examined in vivo immune responses to antigenic challenge in 2 month old B6.NZMc7.
We performed the surface phenotypic and functional analyses of lymphoid cells from the
spleen, peritoneal cavity, bone marrow, and lymph node in these mice to determine the
nature of possible lymphocyte defects leading to autoantibody production. The results
indicate that Sle3 profoundly affects T cell activation, proliferation, and subset
distribution at an early age, which may lead to production of polyclonal IgM and IgG and
anti-nuclear autoantibodies. T cell changes in B6.NZMc7 mice compared to B6 are
heightened CD4:CD8 T cell ratio, increased CD69+CD4 T cells in both the spleen and
lymph node, increased lymphocyte proliferation and CD4 T cell expansion in response to
anti-CD3 stimulation in vitro, and increased in vivo immune response to T-dependent
antigens. B cells from the spleen and lymph node of B6.NZMc7 mice did not differ from
those of B6 in number, B cells subset (B 1 vs. B2 cells), or function (B cell proliferation
in vitro). B6.NZMc7 mice showed some characteristics of B cell hyperactivity including
elevated serum IgM levels, polyclonal IgM, and high expression of some B cell surface
markers at a late age. Finally, Sle3 leads to autoantibody-mediated GN.
Thirty-six B6.NZMc7 congenic mice were initially obtained. These mice differed
genetically from B6 at the three selecting loci for the Sle3 interval, an unknown number
of unselected flanking loci, and unlinked contaminating loci (probability 0.78%). To
define the NZM2410 interval and to establish a congenic line, thirteen marker loci on
chromosome 7 were used to map the termini of the congenic interval and to identify
recombination. The mice were grouped according to their genotype (Figure 3-1). Fifteen
mice from group 1 that carry about 30 cM of chromosome 7 from NZM2410 were chosen
as the founders of the congenic strain and used to continue breeding. These mice and their
progeny were used for phenotypic studies. The autoimmune phenotypes expressed in both
homozygous and heterozygous B6.NZMc7 mice were identified by comparison with age-
matched B6 controls. All animals were housed in identical conditions and bled bimonthly
beginning at 2 months of age for autoantibody analysis..
B6.NZMc7 mice have hypergammaglobulinemia and spontaneous polyclonal B
cell activation. As the most common immunological abnormalities in lupus mice,
hypergammaglobulinemia and spontaneous polyclonal B cell activation were first
examined using total serum IgM and IgG levels and spontaneous IgM antibody
production to four antigens: histone, ss DNA, bovine thyroglobulin, DNP-KLH.
B6.NZMc7 mice had significantly high levels of total serum IgM (Figure 3-2, panel A)
and IgG (Figure 3-3) compared with B6 at 9 months of age (p< 0.0001). Elevated IgM
levels started at 5 months of age and were found in both homozygous and heterozygous
mice (Figure 3-2, panel B). These mice also spontaneously produced IgM
polyclonal/polyreactive antibodies against the four antigens at 5 months and this
increased with age (Table 3-1). The penetrance of polyclonal IgM is 85% in 9 months
B6.NZMc7 (percentage of positive mice who produced IgM antibodies against two or
more of these four antigens). These mice were also tested for spontaneous IgG antibodies
to the same four antigens, and penetrance was 64% at 9 months of age.
B6.NZMc7 mice spontaneously produce IgG anti-nuclear autoantibodies (ANA).
The presence of IgG antibodies against a variety of nuclear antigens is the characteristic
of SLE in both humans and mouse models. NZM2410 produce high levels of ANA at an
early age. To determine if Sle3 by itself can lead to this phenotype and to determine the
antibody characteristics of B6.NZMc7 compared with NZM2410, we quantified IgG
autoantibodies to four nuclear antigens (ssDNA, dsDNA, histone, chromatin or
DNA/histone complex) bimonthly from 2 to 12 month old. B6.NZMc7 produced IgG
autoantibodies against these four nuclear antigens starting at 2 months of age (p<0.01),
especially anti-DNA and anti-chromatin antibodies (Figure 3-4). The titers were much
lower in the congenic mice compared to NZM2410 (data not shown) and the average
titers did not increase with age. This phenotype is dominant with slightly decreased levels
and delayed onset in heterozygous B6.NZMc7. Penetrance of IgG ANA, as shown in
Table 3-2, is about 60-80% and does not typically increase with age unlike what is seen
in B6.NZMcl mice (L Morel, personal communication). After analyzing individual mice,
we found IgG ANA titers in some B6.NZMc7 mice to wax and wane from 2 to 12
months (Figure 3-5). Autoantibody levels did not stay positive for long in these mice and
quickly dropped, which may be due to tissue deposition of antibodies. Waxing and
Waning of IgG ANA has also been reported in many SLE patients (121).
The fact that SLE is preponderant among women during their reproductive life
implies that sex hormones have a role in the pathogenesis of autoinmunity in humans.
Sex hormone manipulation experiments in lupus-prone murine strains also show that
female sex hormones facilitate the development of autoimmunity (122-123). After
detailed analysis of serology data comparing male and female B6.NZMc7 congenic mice,
we found that there was no gender bias in titer or frequency of autoantibodies against any
autoantigens from 5 to 12 months (p>0.05).
B6.NZMc7 mice exhibit a high immune response to DNP-KLH. To test immune
system function before disease onset or in the very early stage of the disease, 2 months
old B6.NZMc7 mice and B6 control mice were challenged in vivo with a T-dependent
antigen, DNP-KLH. IgM and IgG anti-DNP antibodies were measured before
immunization (dO), 14 days after first immunization (d14, primary immune response), 7
days after the booster (d21, secondary immune response), and at d51. B6.NZMc7 mice
showed high primary and secondary IgM anti-DNP immune responses (p < 0.05) (Figure
3-6), high secondary IgG anti-DNP immune responses (p < 0.01) (Figure 3-7), and anti-
DNP antibodies were remained longer compared with B6. This experiment indicates that
the genetic defects in B6.NZMc7 mice lead to a hyper-immune activation status and may
involve both B and T cell compartments.
To define abnormalities in the B and T cell compartments, cell surface molecule
expression, cell function and subset distributions of B and T cell were investigated.
Sle3 affects T cell subopulations. To explore the impact of Sle3 on T cells, cell
subpopulations were first investigated via surface marker expression. In order to define
primary defects, our studies focused on young B6.NZMc7 mice (1-3 month). Summary of
FACS analysis results is presented in Table 3. The results indicate that the CD4:CD8 T
cell ratio was significantly increased in both the spleen and lymph node of B6.NZMc7
mice compared to B6 (Figure 3-8). There was no difference between males and females
(p > 0.05). High proportions of spleen CD4 T cells expressed the activation marker CD69
(p < 0.01) indicating the activation of CD4 T cell in the early stages of disease. This
effect was not seen in the lymph nodes. There was no difference in the expression of L-
selectin between B6.NZMc7 and B6 T cells.
Hyper responsiveness of CD4 T cells in vitro. To study CD4 and CD8 cell
response to stimulation in vitro, spleen cells were cultured with anti-CD3 (1 ug/ml) for
various times, stained with anti-CD4 and anti-CD8, and analyzed using flow cytometry.
The results are shown in Figure 3-9. Before culture (data not shown) and at 24 hours post
culture, CD4 and CD8 cell numbers were not significantly different between B6.NZMc7
and B6. However, the absolute numbers of CD4 T cells were higher in B6.NZMc7 mice
than in B6 at 48 and 72 hours of culture. In contrast to CD4 cells, the numbers of CD8
cells were less in B6.NZMc7 mice. The decrease of CD4 cell numbers both in
B6.NZMc7 and B6 mice at 72 hours of culture was observed. This may be due to the low
concentration of anti-CD3 used, since there is no such decrease of CD4 cell numbers in
experiments using 2.5 ug/ml anti-CD3 stimulation combined with anti-CD28 (C Mohan,
personal communication). Figure 3-10 shows 2D dot plots of flow cytometry of this
experiment. Percentage of CD4 cells was higher in B6.NZMc7 than in B6 after 48 hours
of culture (49% vs. 33%). This increased expansion of CD4 T cells in B6.NZMc7 may be
due to a high proportion of activated CD4 T cells, high responsiveness of CD4 T cells, or
reduced apoptosis of CD4 T cells. Indeed, B6.NZMc7 T cells also exhibited reduced
activation-induced T cell death (data not shown).
Sle3 impacts T cell proliferation. The extent of lymphocyte proliferation in 2
month B6.NZMc7 mice was also examined using 3H-thymidine incorporation assay in
spleen cell culture with graded doses of anti-CD3 or rIL-2 for 72 hours compared with
B6. As shown in Figure 3-11, B6.NZMc7 mice exhibited heightened proliferative
responses to anti-CD3 and rIL-2. The data suggest that T cells from B6.NZMc7 have an
early CD3-mediated T-cell activation aberration.
Sle3 and B cell compartment. There were no significant differences between
B6.NZMc7 and B6 in B cell numbers and B cell subsets (BI vs. B2 cells). In 9 to12
months old B6.NZMc7 mice, I-A (MHC class II molecular), B7-1 (costimulatory
molecular) expressions on B cell surface were high (data not shown), which indicated that
high proportions of B6.NZMc7 B cells were activated in the late stage of disease. There
was also no significant difference in vitro B cell proliferation responses to LPS and anti-
IgM between B6 and the congenic mice (data not shown).
Sle3 leads to low incidence of glomerulonephritis (GN). B6.NZMc7 mice
survive for at least 12 months with no obvious proteinuria. The kidney histology was
examined in one year old B6.NZMc7 and B6 mice. The total frequency of GN was about
40% in B6.NZMc7 homozygous mice and 7 % in heterozygous mice (scored 2 and up)
(Figure 3-12). Severe GN (scored 3 to 4, more than 25-50% of glomeruli being affected)
was only seen in 20% of the homozygous mice. GN was seen in both males and females
with no sex bias. Two distinct types of GN were present in these animals. One was
hypercellular and the other showed large intracapillary hyaline deposits with minimal
cellularity. As a control, one year old B6 had few pathological changes (scored 1 or less
Detailed data analysis was performed to define the role of anti-dsDNA
autoantibodies in the pathogenesis of GN. The results showed no strong relationship
between titers or onset of the antibodies and GN. Two out of nine GN positive animals
were negative for both anti-dsDNA and anti-chromatin autoantibodies.
Histological analysis of the spleen, lymph node, lung, and liver showed lymphoid
hyperplasias with one out of 33 samples having a lymphoma.
Sequence analysis of candidate gene: transforming growth factor-beta (TGF-,) in
B6.NZMc7. TGF-/3 is a potent immunoregulatory molecule. Although it can both
suppress and enhance immune cell functions in vitro (124-126), most evidence suggests
that it is a natural immunosuppressor (127-130). TGF-/3 knockout mice exhibited a
wasting syndrome accompanied by multifocal lymphoid infiltrates in the heart, lungs,
salivary glands, and other organs (131-132). They also showed SLE-like autoimmune
serology (133). TGF-,/is located within the Sle3 interval of murine chromosome 7, which
makes it a good candidate gene. Primers were designed to amplify 1.5kb previously
published mRNA of TGF-/3 coding region (134). RT-PCR showed that TGF-f3 was
expressed in the spleen and kidney (data not shown). The sequence of TGF-fl was not
polymorphic between B6.NZMc7 and B6 (data not shown). No conclusive data was
obtained in expression assays. Further studies are needed to establish if TGF-/3 is
involved in SLE pathogenesis.
The genetic basis of lupus is very complex, involving contributions from multiple
genes. To identify these genes, interval mapping was performed, and three susceptibility
intervals and one locus were identified: Slel on chromosome 1, Sle2 on chromosome 4,
Sle3 on chromosome 7, and the MHC locus on chromosome 17 in (NZM2410 X B6)F I X
NZM2410 backcross (52). The phenotypes in the progeny reflect the epistatic interaction
of these multiple genes. We believe that each susceptibility interval plays a unique role in
the initiation of pathogenesis. Separating each susceptibility interval was achieved using
a congenic breeding approach in which each interval bearing individual susceptibility
gene or genes from NZM2410 was moved to the resistant B6 background by repeated
backcross and selection (113). B6 and B6 congenic for Slel, Sle2, or Sle3 have same
MHC and background genes except for the selected genomic segment derived from
NZM2410. We successfully produced four congenic strains and genetically dissected the
phenotypes associated with each susceptibility interval. Slel on murine chromosome 1
leads the loss of tolerance to the H2A/H2B/DNA sub-nucleosome with little or no kidney
damage (135). Sle2 on murine chromosome 4 leads to B cell hyperactivity with no IgG
autoantibodies or nephritis (136). Here we present detailed phenotypic analysis of Sle3 on
murine chromosome 7 and explore the nature of possible lymphoid cell defects causing
the appearance of autoantibodies.
B6.NZMc7 showed generalized immune dysfunction. These animals exhibited
hypergammaglobulinemia and polyclonal B cell activation. This phenotype is a most
common immunological abnormality among murine lupus models (137-139). Intrinsic B
cell defects were suggested by the finding of an enhanced in vitro responsiveness of these
B cells to LPS and conventional antigens, and by several cell transfer experiments (73-74,
140-143). The mechanisms underlying polyclonal B cell activation in B6.NZMc7 are not
clear. These mice did not show differences in B cell number, B cell activation or
proliferation in vitro responses to LPS and anti-IgM. Polyclonal B cell activation is
known to accelerate autoimmune disease. This is supported by the strong association of
this phenotype with disease development in the various murine lupus models (70), by the
report that it can accelerate mortality, autoantibody level, and by GN in lupus-prone but
not in normal mice (144).
IgG anti-nuclear autoantibody production is a hallmark of SLE both in humans
and mouse models. Sle3 by itself leads to IgG anti-nuclear autoantibody production
starting at an early age. Autoantibody titers in B6.NZMc7 were lower than in the parental
NZM2410 indicating the requirement of other genes for high level IgG ANA production.
IgG ANA level in B6.NZMc7 was not constant and uniquely waxed and waned from 2 to
12 months of age. The causes of this are unknown. A possible explanation is that the
antibodies and antigens form immune-complexes and deposit in tissues including kidney.
One of the most interesting features is that Sle3 leads to development of GN. In
contrast to acute GN in parental NZM2410, B6.NZMc7 developed chronic GN with
relatively low penetrance, and no apparent proteinuria. Two types of GN were observed,
one was hypercellular and the other was large intracapillary hyaline deposit with minimal
cellularity, whereas in NZM2410 GN showed a mixture of these two types. These
observations suggest that the gene(s) in the Sle3 interval may play a key role in mediating
development of GN in NZM2410 and the other gene(s) may act to accelerate GN. This
conclusion has been confirmed by the study of double congenic mice containing Slel and
Sle3 (C Mohan, L Morel, unpublished observations). B6.NZMclc7 double congenic mice
developed acute GN with high penetrance, as seen in NZM2410. The genes causing
kidney damage can potentially act on different points of the pathogenic pathways. For
example, genes may enhance the production of pathogenic autoantibodies, impact
clearance of immune complexes, facilitate binding of autoantibodies or immune
complexes to kidney, and increased the sensitivity of the kidney to damage.
It has been reported that anti-DNA antibodies have pathogenic potential (94-97),
however, our study does not find a strong correlation between IgG anti-dsDNA and GN
in B6.NZMc7. Two out of nine GN positive mice fail to produce anti-dsDNA or anti-
chromatin autoantibodies. One possibility is that GN is caused by autoantibodies against
other antigens. Further studies are in progress to search for other autoantibodies such as
antibodies to the endogenous viral protein, gp70 and to small nuclear ribonucleoproteins.
Another possibility is that titers of autoantibodies in these two mice are low and that
tissue deposition of the autoantibodies makes it serum negative in these two mice.
IgG ANA production can be triggered by specific autoantigens and can be the
result of polyclonal B cell activation. There is a report that polyclonal B cell activation
alone does not lead to IgG ANA production (136). Increasing evidence suggests that
autoantibodies appear to be the products of autoantigen-driven immune responses (145-
149) characterized by IgG class switching, somatic mutation, and clonal expansion of
lupus B cells. This pathogenic autoantibody production in both spontaneous and
experimentally induced SLE is T helper (Th) cell dependent and Th cells are
predominantly CD4 positive (84,87-89). The studies of using anti-CD4 monoclonal
antibody to deplete or block the activation of CD4 T cells further demonstrate a
requirement for this subset in lupus development in mouse models (150-153). These T
cells may be polyclonally activated (154 -155) or autoantigen-specific (156-157).
Both B and T cell defects can lead to the overproduction of pathogenic
autoantibodies. This study indicates that Sle3 mainly impacts T cell function. First,
development of autoinmune disease in young B6.NZMc7 mice was accompanied by an
increased percentage of CD4 T cells and an increased CD4:CD8 T cell ratio in both the
spleen and lymph node. However there was no significant difference found in total T cell
numbers between young B6.NZMc7 and B6 mice (data not shown). Second, a high
percentage of CD4 T cells in B6.NZMc7 was activated and expressed activation marker
CD69. When cultured with anti-CD3 in vitro, CD4 T cells expanded preferentially. Total
lymphocytes also showed heightened proliferative responses to anti-CD3 and IL-2
stimulation. Hyper responsiveness was further demonstrated by in vivo T-dependent
antigen DNP-KLH challenge assays. B6.NZMc7 showed heightened primary and
secondary immune responses to DNP. Third, B6.NZMc7 CD4 T cells exhibited reduced
activation-induced T cell death. In addition, preliminary data from bone marrow transfer
assays showed expression of a high CD4:CD8 T cell ratio can be transferred by
B6.NZMc7 bone marrow to B6 recipients (E. Sobel, unpublished observations). Taken
together, these results suggest intrinsic T cell defects are present in B6.NZMc7 congenic
CD4 T cells can augment IgG ANA production. They can provide cognate help,
which augments antigen-specific B cells via gp39 and CD40 transducing a second signal
for B cell growth and differentiation (85), and non-antigen-specific help, which augment
any bystander B cells through providing B cell growth and differentiation cytokines (87).
The autoantigen-nonspecific T cell help may promote generalized humoral autoinimunity,
such as hypergammaglobulinemia and autoantibody production, but autoantigen-specific
T cell help is required to maturing such autoimmune responses to causing end-organ
disease (158). T cell hyperactivity in B6.NZMc7 may provide both autoantigen-specific
and autoantigen-nonspecific help for autoantibody production and immune complex-
associated end-organ disease.
We can not make a conclusion about B cell abnormalities in this study. Polyclonal
B cell activation, IgG ANA production, and high primary and secondary immune
responses in vivo may indicate possible B cell abnormalities and lead to the loss of
tolerance to nuclear antigens in B6.NZMc7. It is unclear whether these observations are
due to intrinsic B cell defects or normal B cell responses to abnormal environment factors
including T cell hyperactivity, or both. Pre-B cell transfer experiments are in progress and
will help to address this question.
The humoral phenotypes and nephritis were found in both B6.NZMc7
homozygous and heterozygous mice, with delayed onset and less severity as well as
decreased penetrance in heterozygotes. This result indicates the mode of inheritance is
dominant and in an allele dose-dependent fashion. Further studies are needed to
determine the mode of inheritance of cellular defects identified here.
This study demonstrates that B6.NZMc7 bearing SIe3 exhibit complex and
multifaceted phenotypes which are different from B6.NZMcl and B6.NZMc4. Sle3
contributes to lupus susceptibility in NZM2410 predominantly by impacting T cell
activation and proliferation leading to the low grade IgG ANA production and lupus
nephritis. The phenotypes in B6.NZMc7 are less pronounced than these in NZM2410.
The relative complex phenotypes may reflect the presence of two or more SLE-
susceptibility genes within this interval. This possibility has been suggested by the
presence of two peaks in genetic linkage studies on backcross mice (52) and further
supported by results from linkage studies in intercross mice (159). To explore this
possibility, we separated the telomeric and centromeric peaks within the Sle3 interval by
generating two subcongenic strains. Detailed analysis of these subcongenic mice is
presented in Chapter 4.
TGF-f is a potential candidate gene located in the SIe3 interval. The coding region
of this gene has been sequenced and it was found that there is no consistent sequence
polymorphism between B6.NZMc7 and B6. Further studies are needed to determine if
there are any differences in expression levels of this gene.
1 2 3 4 5 6 7 8 9
D7MIT178 2.3 0 0 11 )1 1"0 10 EJ 00 0 0 0
CKMM 4 0j N Nm 11 U m1 C ] E C 0
D7MIT76 4.6 ** ** 1)1) 10i:] 101 ImDJI U U 0
D7MIT154 5 U mu mi 11) E 0 13 N
D7MITl14 9.8 mm so N 111 11 ) D 1m ) U N ID EJ
D7MIT155 11.2 NNE E* Ern m U ujj Em* I 10 10 i]
D7MIT25 14.3 0 N E E No mE on *0 Ir 10 1 E i
D7MIT69 17.8 N* mm No EnE U N U E
D7MIT158 18.9 N E N 0 E E E N N N E MEN U N U
P 23.4 mmm mm mm mm m mmm N E
D7MIT70 23.8 E** N 0 N M mom 0 U U
D7MIT31 34.1 *** 100 ml 0 N lU 3)3) N)E3) U[ I 0 I
D7MIT68 50.6 [ [ [ 00 00 00 00 0) OD E r 3I 0
numbers of mice 2 7 6 3 3 1 1 1 1 1 2 2 1 2 1 1 1
Fig 3-1. Genotypes of the 36 original B6.NZMc7 congenic mice. Thirteen markers coving the Sle3 interval,
(indicated on the left) were used to genotype each animal. The distance of each marker to the centrimeric end is
shown in the second column and is expressed in units of centiMorgan (cM). Black boxes represent mice typed as
NZM24 10 homozygotes, white boxes, mice typed as B6 homozygotes, half black boxes represent heterozygous
mice. Mice from group 1 were used as founders for breeding all B6.NZMc7 mice used in this study.
Marker loci cM
Figure 3-2. Total serum IgG and IgM in 9 month old mice. High levels of serum IgG and IgM were
observed in B6.NZMc7 homozygous mice (black dot) compared with B6 (white triangle). Each dot
represents an individual mouse. Horizontal line is mean value.
B6 B6.NZMc7hom. B6 B6.NZMcThom.
n=10 a=21 n=1O n-21
Figure 3-3. Kinetics of total serum IgM levels in B6.NZMc7 homozygous mice. There is no significant
difference between B6.NZMc7 homozygotes (black dot, all groups n=23, 2 mo group, n=13) and
heterozygotes (white dot, n=5-7). Horizontal lines are mean values. Each dot represents an individual
Table 3-1. Spontaneous production of IgM antibodies against a panel of four antigens in B6.NZMc7 homozygotes
compared with age-matched B6.
5 mo 7 mo 9 mo 11-12 mo
g B6 B6.NZMc7 B6 B6.NZMc7 B6 B6.NZMc7 B6 B6.NZMc7
ssDNA 10825a 342-+114*b 14232 359104* 141+0.32 356120" 16564 473-157"
Histone 131-34 357116* 14524 26547* 15135 294075* 14358 394136*
Thyroglobulin 170089 605318* 29198 1179749* 417251 1012-684* 397184 1545991"
DNP-KLH 105019 284095* 18341 32084* 21763 374145* 22390 433163*
a All data expressed as mean % OD standard error of mean (SEM). % OD is the ratio of the sample OD divided by the OD
of normal B6 pooled serum x 100 (112). The number of mice tested is 14 B6, and 34 B6.NZMc7.
b represents student's t test significance, p < 0.001.
Figure 3-4. IgG anti-dsDNA and anti-chromatin autoantibody production in 2 to 12 month old B6.NZMc7
mice. B6.NZMc7 homozygous mice (black dot, all groups n=23, 2 mo group, n=13) produce both anti-
dsDNA and anti-chromatin antibodies starting at 2 mo of age (p<0.001). The phenotypes are also found in
B6.NZMc7 heterozygous mice (white dot, n=8) with slightly low titers. Significant difference was
detected in B6.NZMc7 homozygous and heterozygous mice compared with B6 (white triangle, n=8)
Horizontal lines represent mean values.
I I I I I I -- I
0 % I: to
12 2 5 7 9 11
12I I I 7I
12 2 5 7 9 11
Table 3-2. Penetrance of IgG anti-nuclear autoantibody production in B6.NZMc7.
2 mo 5 mo 7 mo 9 mo 11-12 mo
Antigen B6.NZMc7 B6.NZMc7 B6.NZMc7 B6.NZMc7 B6.NZMc7
hom.b het. hom. het. hom. het. hom. het. hom. het.
dsDNA 17%c NDd 37% 10% 45% 17% 39% 28% 74% 65%
ssDNA 33% ND 40% 5% 64% 21% 73% 22% 82% 65%
Chromatin 17% ND 26% 5% 61% 50% 45% 44% 41% 65%
Histone 33% ND 40% 10% 24% 17% 55% 17% 44% 50%
N_ 12 35 20 33 24 33 18 34 20
a Number of mice tested.
b Homozygous (hom.)B6.NZMc7, heterozygous (het.) mice.
c Percentage of positive mice. The mouse is counted as positive if its antibody titers are above the levels of mean plus 2 standard
deviation found in six 12 mo old B6 mice assayed on the same plate.
d ND: not determined.
Figure 3-5. Waxing and waning nature of anti-ANA production in B6.NZMc7 mice. Each line represents
an individual mouse. A total of 8 mice were selected from 23 mice tested from 2 to 12 months of age for
anti-dsDNA at the same ELISA plate. The horizontal line represents mean plus 2 standard deviation serum
antibody levels present in ten 12 month old B6 mice. Levels of antibody above this line are considered
iEli UUUMMi EM I
Figure 3-6. Immune responses to T-dependent antigen DNP-KLH in vivo. A) IgM anti-DNP antibody
production before immunization, 14 days after primary immunization, 7 and 30 days after boosted
injection (d21, and d5 1) at serum dilutions of 1:500 in B6.NZmc7 mice (white dots, dashed line) and age-
sex-matched B6 (black dots, solid line). B) IgM anti-DNP antibodies at day 51 at different serum dilutions.
Data are from one of three independent experiments, and each dot is expressed as mean OD plus standard
deviation from five B6.NZMc7 or B6 mice. *, p <0.05; **, p<0.01.
o'2 -" _""
0 14 21 51 -
I in 500 1 in 2500 1 in 12500 1 in 62500
Figure 3-7. Immune responses in vivo to T-dependent antigen DNP-KLH. A) IgG anti-DNP antibody
production before immunization, 14 days after primary immunization, 7 and 30 days after boosted
injection (d21,d51) at serum dilution of 1:12500 in B6.NZMc7 mice (white dots, dashed line) and age-
sex-matched B6 (black dots). B) IgG anti-DNP antibodies at day 51 at different serum dilutions. Data are
from one of three independent experiments, and each dot is expressed as mean OD plus standard deviation
from five B6.NZMc7 or B6 mice. *, p < 0.05; **, p < 0.01
14 21 51 0 I I
I i 500 1 in 2500 1in 12500 1 in 62500
Time (Days) Serum dilution
Table 3-3. Summary of FACS analysis of T cells in B6.NZMc7 mice compared with age matched B6.
1-3 mo mice 1-3 mo mice
Spleen Lymph node
B6 B6.NZMc7 B6 B6.NZMc7
n-26 n-24 n-16 n-9
%CD4 T cells 17.452.97a 20.16+2.7**b 26.104.71 35.676.23***
%CD8 T cells 11.732.48 11.192.22 24.575.89 24.415.83
CD4:CD8 T cell ratio 1.510.16 1.860.42*** 1.12+0.33 1.540.45*
CD69+CD4 T cells 16.49.3 238** 2416 1513
" All data expressed as mean standard deviation (SD).
b represent student t test significance, p < 0.05; **, p < 0.01; ***, p <0.001.
Figure 3-8. Spleen and lymph node CD4:CD8 T cell ratio in 1 to 3 month B6.NZMc7 mice. There was
no difference between males and females.Data is pooled from 10 independent experiments, for each test,
age- and sex- matched mice were used and B6 mice included. The results are presented as mean plus
standard deviation. *, p < 0.05, ***, p < 0.001.
Figure 3-9 B6.NZMc7 spleen CD4 and CD8 T cells expansion after anti-CD3 (1 ug/ml) stimulation in
vitro. Black bars are B6, white bars are B6.NZMc7. The mean value of duplicates is presented. Data
represent one of three independent experiments. All mice are age- and sex-matched. Cell numbers were
calculated using FACS data and cell counting.
Culture Time (hours)
Figure 3-10. Plots show increased percentage CD4 cells in splenic lymphocytes cultured with anti-CD3
(1 ug/ml) for 24h, 48h, and 72h from B6.NZMc7 compared with age-matched B6. This experiment is
representative of three independent experiments.
Oh 24h 48h 72h
11 70 42.9- 5f
S 17.93 33.46 .32.44
C D 8 .32.6 *. 346
~19.75 204 44694
:.7.2._.': .'. : !'' .0 ",' '- i3 .6 =.
Figure 3-11. Lymphocyte proliferation assay. Splenocytes 1 x 106 were culture with graded doses of anti-
CD3 (panel A) or rIL-2 (panel B) for 72h. Lymphocyte proliferation was measured by 3H-thymidine
incorporation which added in the last 18 h of culture. Dashed line represents B6.NZMc7, solid line age-
and sex-matched B6. Each bar represents mean cpm (counts per minute)standard error of mean (SEM)
incorporated by triplicate cultures. This experiment is representative of 3 independent experiments.
* p <0.05, ** p < 0.001
Figure 3-12. The frequency of GN in B6.NZMc7 homozygous (hom) and heterozygous (het) for the Sle3
interval. GN was graded according to percent of glomeruli affected (grade 1 to 4) and the morphological
pattern of nephritis (H= hyaline, M= mesangiopathic).
1 3+4H GN
B6NZMc7 horn. B6.NZMc7 het.
GENETIC DISSECTION OF SLE PATHOGENESIS:
SLE3 HAS TWO SEPARATE LOCI THAT INTERACT SYNERGISTICALLY
SLE is a prototypic systemic autoimmune disease characterized by production of
autoantibodies against a spectrum of nuclear antigens and autoantibody-mediated organ
damages especially kidney damage. The cause of disease is unclear. The genetic factors
and environmental factors as well as stochastic factors contribute to the altered immune
state (1). Identification of underlying gene mutations or allelic variants that affect the
function of genes and an understanding of their functional consequences will impact the
diagnosis, treatment, and prevention of SLE.
Several inbred mouse strains that spontaneously develop a disease similar to
human SLE are important models (1). The most-studied is the F1 hybrid of NZB and
NZW. New Zealand Mixed (NZM) are inbred strains derived from selective inbreeding of
progeny from a cross between NZB and NZW (35). A total of 27 NZM strains were
produced. These strains have inherited various genomic segments of the two parental
strains and are homozygous at the histocompatibility (H-2) locus for the NZW haplotype
(Ku, Au, SZ, Dz). The genetic differences among these strains lead to phenotypic
differences in disease expression. NZM2410 is a strain that produces high levels of IgG
autoantibodies to nuclear antigens as early as 4 to 6 months of age and develops severe
lupus-like renal disease in both sexes. A genome scan was performed in the backcross
progeny of NZM2410 and B6 to identify the chromosomal locations of susceptibility
genes (52). Three susceptibility intervals and the MHC locus were found to confer acute
GN susceptibility. The three intervals were: Slel, in the telomeric end of chromosome 1;
Sle2, in the centromeric end of chromosome 4; Sle3, in the centromeric end of
chromosome 7. Similar regions have also been identified in NZB X NZW crosses (48,50,
reviewed in ref. 160). Although these linkage studies have provided important
information about the chromosomal locations of susceptibility genes and the inheritance
of the disease, they tell us little about the role of each gene in the pathogenesis of lupus
and they are not productive for fine mapping. The construction of congenic strains is a
good approach to address these issues. The susceptibility intervals from the susceptible
donor strain are introduced into the genome of the resistant recipient strain by repetitive
backcross of the donor strain to the recipient (113). We recently generated a collection of
B6 congenic strains bearing the individual lupus susceptibility intervals from NZM2410.
Thus, by comparing B6 and the B6 congenic strains differing only for the specific
chromosomal segments, we are able to characterize the phenotypes contributed to disease
pathogenesis by each susceptibility interval. Interestingly, these congenic strains exhibit
different serological and immunological abnormalities and varying susceptibility to
Detailed analysis of B6.NZMc7 bearing the Sle3 interval is presented in Chapter
3. In summary, our results indicate that Sle3 profoundly affects T cell activation,
proliferation, differentiation, and apoptosis. B6.NZMc7 mice had increased CD4:CD8 T
cell ratios and increased CD69+CD4+ T cells in both the spleen and lymph node. They
showed heightened lymphocyte proliferation and CD4 T cell expansion in response to
anti-CD3 stimulation in vitro and increased in vivo immune responses to T-dependent
antigens. CD4 T cells exhibited reduced activation-induced cell death. B6.NZMc7 spleen
and lymph node B cells did not differ from B6 in number, B cells subset (BI vs. B2
cells), or functional analysis (B cell proliferation in vitro). However, B6.NZMc7 mice
showed high serum IgM and IgG levels, high polyclonal IgM antibodies, and high
expressions of some B cell surface markers in a late age. Finally, Sle3 lead to
autoantibody-mediated GN. The relatively complex phenotypes in B6.NZMc7 suggest
the presence of two or more SLE-susceptibility loci within this interval. This possibility
has also been suggested by the presence of two peaks in our genetic linkage studies on
backcross (52) and intercross mice (159).
Here we explore this possibility by generating two subcongenic strains bearing the
telomeric or the centromeric peaks of Sle3. The construction of B6.NZMc7c and
B6.NZMc7t subcongenic strains was described in Chapter 2. B6.NZMc7c contains the
centromeric end of Sle3 and B6.NZMc7t contains the telomeric end of Sic3. These strains
were analyzed for their serological phenotypes, evidence of GN, and T cell phenotypes
compared to B6 and the parental B6.NZMc7 strain. Results suggest that Sle3 has at least
two separate susceptibility loci that interact in a synergistic way to cause polyclonal B
cell activation, IgG ANA production, and development of GN. However, increased
CD4:CD8 T cell ratios and increased CD4 T cell expansions in vitro as well as the
heightened T cell proliferations in vitro were seen only in the mice carrying the telomeric
end of Sle3.
Dense linkage map of SSR markers on murine chromosome 7 covering the Sle3
interval. In order to produce the subcongenic mouse strains and finely map susceptibility
loci, a precise map of SSR markers is very important. Ninty-six (NZM2410 X B6)Fl X
NZM2410 backcross and 144 F2 intercross progeny representing approximately 400
meiosis were genotyped using 25 markers covering the Sle3 interval on murine
chromosome 7. The map of SSR markers was constructed using MAPMARKER
computer software (161). As shown in Figure 4-1, the order and genetic distances
between markers that we obtained were mostly in agreement with published data (162).
However, one marker order difference was observed. The marker D7Mit69 was proximal
to D7Mit8l and D7Mitl58 in our map, but distal to these markers in the published map.
We also defined the order of 12 markers (underlined in Figure 4-1), which could not be
separated due to the small number of mice genotyped in the construction of published
map. This SSR marker map is helpful in determining recombination sites during
production of the subcongenic strains.
Generating the subcongenic strains was described in Chapter 2. As shown in
Figure 4-1, B6.NZMc7c contains the centromeric end of Sle3 with NZM2410-derived
alleles from marker ckmnm (4 cM) to D7Mit154 (5 cM). B6.NZMc7t contains the
telomeric end of Sle3 with NZM2410-derived alleles from marker D7Mit69 (17.8 cM) to
D7Mitl47 (30.6 cM).
Both loci lead to h)pergammaglobulinemia and spontaneous Dolyclonal B cell
activation. The mice from both subcongenic strains showed generalized immune
hyperactivity like parental B6.NZMc7 mice. They had high total serum IgG and IgM
levels compared with B6 at 9 months of age (Figure 4-2). They showed high polyclonal
IgM antibodies against ssDNA, histone, DNP-KLH, thyroglobulin at 5 months of age
(Figure 4-3) as well as at 7, 9, and 12 months of age (data not shown). However, titers of
polyclonal IgM antibodies were lower than B6.NZMc7. Consistent with low antibody
titer, the penetrance of polyclonal IgM was lower also and there was little polyclonal IgG
antibody production in either of the subcongenic strains (Figure 4-4). These results
suggest that both loci can independently cause a low-grade polyclonal B cell activation,
but they need to interact synergistically in generating the pronounced phenotypes seen in
Both loci are required for IgG ANA production. B6.NZMc7 mice spontaneously
produce IgG antibodies against a broad spectrum of antigens, including nuclear and non-
nuclear antigens. This phenotype was also examined in mice bearing the subintervals. As
shown in Table 4-1, the frequency of IgG autoantibody production to a panel of four
nuclear antigens was very low. For example, the frequency of anti-dsDNA at 12 months
is about 13% in B6.NZMc7c, and 32% in B6.NZMc7t compared to 74% in B6.NZMc7.
Titers of IgG autoantibodies in both the subcongenic strains were also much lower
compared with B6.NZMc7 at 7 months of age (Figure 4-5) as well as at 9 and 12 months
of age (data not shown). In summary, neither locus by itself on B6 background is
sufficient to drive spontaneous IgG ANA production.
Responses of subcongenic mice to DNP-KLH challenge. As described above,
both the subcongenic strains spontaneously produced IgM antibodies to nuclear and non-
nuclear antigens, but they spontaneously produced little IgG autoantibodies. To further
assess immune responses, we challenged these mice in vivo with DNP-KLH. IgM and
IgG anti-DNP antibodies were examined before immunization, primary immune
response, secondary immune response, and duration of the response. In contrast with
B6.NZMc7 mice that produced high IgM anti-DNP antibodies, there was no significant
difference in the IgM anti-DNP antibody levels between the subcongenic strains and B6
(p>0.05). However, the IgM anti-DNP antibodies lasted longer in B6.NZMc7t (p<0.01)
(Figure 4-6, panel A), which could be seen from serum dilution 1:500 to 1:62500 (Figure
4-6, panel B). A high IgG anti-DNP secondary response was observed in B6.NZMc7t
(p<0.01), which was also seen in B6.NZMc7. The IgG anti-DNP antibodies lasted longer
in both the subcongenic strains compared with B6 (p<0.01) (Figure 4-7).
B6.NZMc7t exhibited increased CD4:CD8 T cell ratios. To define T cell
abnormalities in the subcongenic strains, we first examined the T cell surface marker
expression using flow cytometry. Table 4-2 summarizes the results of FACS analysis.
B6.NZMc7t mice showed increased percentages of CD4 T cells and increased CD4:CD8
T cell ratios in both the spleen and lymph node (Figure 4-8), with similar findings in
B6.NZMc7 (data not shown). A high percentage of splenic CD4 T cells in B6.NZMc7t
expressed activation marker CD69 (p < 0.01).
CD4 T cells from B6.NZMc7t exhibit high expansion in vitro. We also
examined absolute CD4 and CD8 T cell numbers after splenocytes from both the
subcongenic strains were stimulated with anti-CD3 in vitro. Consistent with increased
CD4 T cells in vivo, the absolute numbers of CD4 T cells were also increased in vitro in
B6.NZMc7t (Figure 4-9, panel A). Although CD4 T cell numbers were higher in
B6.NZMc7t compared with B6 and B6.NZMc7c, CD4 T cell numbers were even more
significantly increased after 48h of culture. A small drop of cell numbers in 72h culture
was noticed in B6.NZMc7 (Chapter 3), and also seen in B6.NZMc7t, probably due to
reasons discussed in Chapter 3. In contrast, CD8 cells were highly expanded in both B6
and B6.NZMc7c compared to B6.NZMc7t (Figure 4-9, panel B).Figure 4-10 shows the
change of percentage CD4 and CD8 T cells in 2D-dot-plot from the same experiment.
These results further demonstrate that the telomeric end of Sle3 impacts T cell function.
The telomeric subinterval also impacts lymphocyte proliferation. T cell
hyperactivity is an important phenotype in B6.NZMc7 mice and is believed to play a
important role in driving B cell autoantibody production. We therefore examined T cell
proliferation in subcongenic mice using a 3H-thymidine incorporation assay. T cells from
B6.NZMc7t exhibited an increased stimulation index (counts per minute in stimulated
cultures divided counts per minute in unstimulated cultures) in response to anti-CD3 (10
ug/ml) and IL-2 (10 U/ml) in vitro (Figure 4-11). T cells had higher responses to anti-
CD3 than to IL-2. The heightened proliferative responses to these stimuli may be due to
increased numbers of activated T cells, increased signal transduction, or reduced
Both loci lead to a low incidence of glomerulonephritis (GN). The subcongenic
mice were sacrificed at 12 months for kidney histology. As shown in Figure 4-12,
surprisingly, the total frequency of GN (scored 2 and above) in both homozygous
subcongenic mice is as high as that in parental B6.NZMc7 (about 45%). However, the
overall severity of GN was less in the subcongenic strains compared with B6.NZMc7.
Both the subcongenic strains exhibited the mesangiopathic glomerular lesions that were
less severe than the hyaline lesions in B6.NZMc7. There was no sex bias in GN
frequency. GN was correlated with the production of autoantibodies against at least one
of four nuclear antigens, but not strongly correlated with the titers and onset of
autoantibody production (data not shown).
Although significant volumes of data describing many humoral and cellular
abnormalities in murine lupus mice have been generated in the past three decades, the
genetic defects causing these abnormalities have not been determined. Although genetic
linkage studies have identified at least nine different loci to confer lupus susceptibility
(163), contribution and nature of each locus in the pathogenesis of abnormalities need to
be defined. Our approach to address these questions is via the production and analysis of
congenic strains bearing individual loci. From studies of B6.NZMc7 (Chapter 3), we
found Sle3 on murine chromosome 7 impacted T cell function, led to polyclonal B cell
activation, IgG ANA production, and lupus nephritis. In this study, we further dissected
the complex traits of B6.NZMc7 by generating two subcongenic strains and analyzing
We found that two subintervals both loci within Sle3 could lead to spontaneous
polyclonal B cell activation by themselves. These phenotypes are also observed in B6
mice congenic for the Sle2 interval on chromosome 4 (B6.NZMc4) (136). These
observations suggest that polyclonal B cell activation is a basic abnormality in murine
lupus and contributes to the pathogenesis of IgG ANA and lupus nephritis. Since
B6.NZMc4 mice have polyclonal B cell activation and lack IgG ANAs and nephritis,
polyclonal B cell activation alone is not sufficient for their autoimmune phenotypes.
The penetrance of autoantibodies against the four nuclear antigens decreased
dramatically in both subcongenic strains compared with B6.NZMc7, which indicates
requirement of both loci. The total GN frequency in the subcongenic strains is same as in
the parental strain, but the severity is reduced. These results further confirme our
hypothesis that SLE is inherited as a threshold liability (53) requiring multiple
susceptibility genes to cause high penetrance of phenotypes. This feature also makes the
further narrowing of subintervals very difficult.
GN is associated with IgG autoantibody production against four nuclear antigens
in both the subcongenic strains. We believe the conclusion is also true in parental
B6.NZMc7 mice because immunoflorescence assas of B6.NZMc7 kidneys showed the
presence of antibodies and complements in kidney (data not shown). Our experiments can
not define the specificity of autoantibodies causing GN.
The susceptibility genes causing the humoral autoinimune responses in the
subcongenic strains could impact T cell function, B cell function, or both. In a previous
study, we demonstrated that Sle3 primarily impacts T cell function, which may drive
lupus B cell activation and production of autoantibodies. To map this gene(s), we
performed the same sets of experiments on the subcongenic strains. We found that
B6.NZMc7t mice exhibited increased CD4:CD8 T cell ratios, higher IgG immune
responsiveness to the T-dependent antigen, heightened T cell proliferative responses to
anti-CD3 and IL-2. These findings suggest this gene(s) may locate in the telomeric end of
Sle3. B6.NZMc7t mice contain the disease gene(s), but fail to produce the same level of
IgG ANA as B6.NZMc7 indicating requirement of the gene(s) from the centromeric end
of Sle3 to augment these phenotypes.
B6.NZMc7c mice showed no obvious T cell functional aberrations. The
underlying mechanism for polyclonal B cell activation and lower level IgG ANA
production is not clear. It could be due to B cell dysfunction, subtle T cell function
changes, or both. Further detailed studies are needed.
This study successfully dissects the complex traits associated with Sle3 and
identifies the existence of at least two loci in the Sle3 interval. The experimental findings
indicate that both loci are required to generate certain levels of humoral autoinimunity
and severe lupus nephritis, but only the telomeric locus impacts T cell activation,
proliferation, and differentiation.
Figure 4-1 A map of simple sequence repeat (SSR) markers on murine chromosome 7 covering the Sle3
interval. The map was constructed using MAPMAKER software according to 400 meiosis. NZM2410-
derived susceptibility intervals (white bars) in B6.NZMc7 and the two subcongenic strains are showed,
lines represent B6-derived alleles. The distances from each marker to the centromere were expressed in
--"'' D7Mit78 D7Mit79
7Mit158 I D7Mit83
Figure 4-2. Total serum IgG and IgM level in 9 month subcongenic strains compared with B6.NZMc7.
Each dot represents individual mouse. Horizontal line is mean value.
I. I I I
136 B6.NZIW7 IKNEkc D6T JNZbc7t
FI0 v=21 0=24 32
I I I
96 BNDOh? B6IQi. k7c B6.NDk7t
3=10 m721 v=24 1=25
Figure 4-3. Polyclonal IgM antibodies against a panel of four antigens in 5 month old subcongenic strains
compared with B6 and B6.NZMc7. A similar data pattern is also seen in 7, 9, and 12 month old mice from
these four strains with increased titers. Antibody titer is expressed as %OD (detailed definition in Methods
and Materials). Data is presented as mean and standard deviation. p < 0.05, ** p < 0.01, *** p < 0.001.
800 0 B6 n=14
U B6.NZMc7c n=35
EJ B6.NZMc7t n=30
C U B6.NZMc7 n=34
o 0 .....
ssDNA Histone DNP-KLH Thyroglobulin
Figure 4-4. Penetrance of IgM and IgG polyclonal antibody production in 9 month old subcongenic mice
compared with parental B6.NZMc7. The mouse was counted as positive if it produced antibodies above
the mean plus 2 SD often B6 to any two of four antigens (ssDNA, histone, DNP-KLH, thyroglobulin).
90% l B6.NZMc7c n=24
U B6.NZMc7 n=33
Polyclonal IgM Polyclonal IgG
Table 4-1. Penetrance of IgG anti-nuclear autoantibody production in B6.NZMc7c and B6.NZMc7t homozygous mice.
5 mo 7 mo 9 mo 11-12 mo
c7c c7t c7c c7t c7c c7t c7c c7t
dsDNA 3%b 9% 8% 24% 0% 21% 13% 32%
ssDNA 12% 12% 4% 8% 8% 11% 13% 27%
dsDNA/histone 3% 3% 15% 8% 25% 18% 8% 23%
Histone 3% 3% 15% 8% 4% 7% 17% 23%
Na 33 34 26 25 24 28 24 22
a Number of mice tested.
b Percentage of positive mice. A mouse is counted as positive if its antibody titers are above the levels of mean plus 2 standard
deviations in six 12 month old B6 mice assayed on same plate.
Figure 4-5. IgG autoantibody production against a panel of four nuclear antigens in 7 month homozygous
and heterozygous subcongenic mice compared with B6.NZMc7 and B6. The data is expressed as % OD
detailedl definition in Methods and Materials). Horizontal lines indicate the mean value of antibody titers.
dsDNA 450 dsDNA/histone
0 o S 0
0 0 030 0
ic o o 2@@s
100O 200- 0
0 0ft 0 0
so + -+-- rg ,,, 7 10o0- o
0 3 Histone
IS0 200 0
100. AA 1 100
50 ,0 _ __ _ _
B6 B6.NZMc7 B6.NZIc7c B6.NZMk7t
horn. ht. hon. heL hon, het B6 B6.NZMc7 B6.NZMc7c B6.NZMc7t
n-25 n=33 n--24 n=20 n=30 n=21 n=31 horn. het hom het hom. het
n=25 n=33 n-24 n=20 n=30 n=21 n=31
Figure 4-6. Immune responses to T-dependent antigen DNP-KLH in vivo. A) IgM anti-DNP antibody
production before immunization (dO), 14 days after primary immunization (d14), 7 and 30 days after boost
(d21 and d5 1) at Serum dilution of 1:500 in B6.NZmc7c mice (white dots with dashed line), B6.NZMc7t
mice (white triangles), and age- and sex-matched B6 (black dots). B) IgM anti-DNP antibodies on 5 Idays
at different serum dilutions. Data is from one of three independent experiments, and each dot is expressed
as mean OD and standard deviation of five animals. *, p < 0.05; **, p < 0.01.