A New Approach for the Enrichment of Lung Cancer Initiating Cells by Aldefluor Staining

Material Information

A New Approach for the Enrichment of Lung Cancer Initiating Cells by Aldefluor Staining
Ucar, Deniz A
Place of Publication:
[Gainesville, Fla.]
University of Florida
Publication Date:
Physical Description:
1 online resource (80 p.)

Thesis/Dissertation Information

Master's ( M.S.)
Degree Grantor:
University of Florida
Degree Disciplines:
Medical Sciences
Committee Chair:
Scott, Edward W.
Committee Members:
Petersen, Bryon E.
Rowe, Thomas C.
Graduation Date:


Subjects / Keywords:
Aldehydes ( jstor )
Cancer ( jstor )
Cell growth ( jstor )
Cell lines ( jstor )
Cells ( jstor )
Enzyme activity ( jstor )
Enzymes ( jstor )
Lung neoplasms ( jstor )
Stem cells ( jstor )
Tumors ( jstor )
Medicine -- Dissertations, Academic -- UF
bibliography ( marcgt )
theses ( marcgt )
government publication (state, provincial, terriorial, dependent) ( marcgt )
born-digital ( sobekcm )
Electronic Thesis or Dissertation
Medical Sciences thesis, M.S.


The cancer stem cell hypothesis proposes that only a subpopulation of cells within the tumor referred to ?cancer initiating cells? (CICs) can generate tumors. The CICs have the properties of adult stem cells, in particular the ability to self-renew and differentiate into multiple cell types. Understanding the molecular biology of the cancer stem cells is crucial for the development of more effective cancer treatments. At present, several cell surface markers have been identified that are useful for enrichment of CICs. Aldehyde dehydrogenase, ALDH class1A1 and class3A1 (ALDH1A1 and ALDH3A1) expression in lung cancer might become potential markers for the characterization of lung CICs. Using ALDEFLUOR flow cytometry based assay, the laboratory of Dr. Moreb was able for the first time to measure ALDH activity in solid tumor cells and document heterogeneity in expression among lung cancer cell lines. The ALDEFLUOR staining system allows for the identification and isolation of stem/progenitor cells based on their ALDH enzymatic activity, rather than their cell surface phenotype. The laboratory of Dr. Jan Moreb has successfully fractionated human lung cancer cell lines, using the ALDEFLUOR staining method for subsequent in vitro studies. In culture, ALDHbr cells display stem cells-like behavior including slow growth rates and resistance to chemotherapeutic agents. We hypothesize that sorting for ALDHbr cells will enrich for CICs that are capable of generating tumors. This study examines the in vivo functional differences between the ALDHbr and ALDHlo sorted cells from the human lung cancer cell line H522 to establish a useful experimental system for studying CICs. This hypothesis has been functionally tested using a xenograft non-obese diabetic/severe combined immunodeficient (NOD/SCID) mouse model to assay for tumor formation. The preliminary results show evidence that higher levels of ALDH altered the in vivo tumor growth rate and was consistent with their in vitro behavior. Tumors derived from ALDHbr cells grew slower than tumors initiated from ALDHlo cells. The same phenomenon was observed at all cell doses that were tested and including a range from 500?105 cells per dose. The serial transplantation experiments demonstrated that both ALDHbr and ALDHlo cells, which were isolated from primary tumors, could generate secondary tumors. Moreover, as predicted from the in vitro studies, ALDHbr cells generated secondary tumors containing ALDHlo cells. These in vivo results established a correlation between low levels of ALDH activity and enhanced tumor growth. This study has recently been extended to the human pancreatic cancer cell line MIA PaCa-2 to determine whether the ALDH-based enrichment method can be used for other types of human cancers. The in vitro experimental results were consistent with the observations using the human lung cancer cell line H522. ALDHbr sorted cells grew slower in culture compared to ALDHlo cells. These results were confirmed by quantifying their ALDH enzyme activity and determining the cell cycle profile of the sorted ALDHbr and ALDHlo populations. Overall, the preliminary results indicate that ALDEFLUOR based sorting has the potential to enrich for cancer initiating cell activity from multiple human tumor-derived cell lines. ( en )
General Note:
In the series University of Florida Digital Collections.
General Note:
Includes vita.
Includes bibliographical references.
Source of Description:
Description based on online resource; title from PDF title page.
Source of Description:
This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Thesis (M.S.)--University of Florida, 2007.
Adviser: Scott, Edward W.
Electronic Access:
Statement of Responsibility:
by Deniz A Ucar.

Record Information

Source Institution:
Rights Management:
Copyright Ucar, Deniz A. Permission granted to the University of Florida to digitize, archive and distribute this item for non-profit research and educational purposes. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder.
Embargo Date:
LD1780 2007 ( lcc )


This item has the following downloads:

Full Text




2 2007 Deniz A. Ucar


3 To all the people who supported me during my masters program


4 ACKNOWLEDGMENTS I would first like to thank my mentors, Dr. Edward Scott and Dr. Christopher Cogle, for the excellent guidance and opportunities they have provided me during my stay in this lab. I gratefully thank Dr. Jan Moreb and Blanca Ostmark, who provided the cells and their data for this project. I would also like to thank my third committee member, Dr Thomas Rowe, for his time, positive energy, and guidance. I thank Dr. Br yon Petersen also for being my knight in shining armor at the last minute by becoming a committee member when I needed the most. My deepest thanks are extended to the members of the Scott lab, especially Marda Jorgensen for her invaluable guidance and time for immunohistochemistry analysis ; Dr. Robert Fisher, also my good friend and leading motivator, for his cons tant nourishments and support even beyond science; Gary Brown, a Scottish master, for his gr eat assistance and expertise in animal work and expanding my vocabulary in English language; Doug Smith, for his time, patience and efforts to help me for FACS analysis and confocal imag ing; Li Lin and Dustin Hart for their endless help and friendship; and Dr. Ann Fu, Dr. Koji Hosaka, Dr. Jeff Harris, Niclas Bengtsson, Sam Kim, and Erin Wilmer. Their help was invaluable a nd their presence made work a pleasant place. I also appreciate Joyce Conners for her cont inual support towards the completion of this degree and being such a wonderful caring Ameri can Mom. I would also include my previous professors; Dr. Lung-Ji Chang and Dr. Philip Laip is, and their lab members in my list of the people who I deeply appreciate for their guidance and continuing friendships. I thank the flow cytometry core members, especially Neal Benson for giving me technical advice on many occasions and for their help in maintain the core. I thank Linda Kephart Fallon from the IRB office who helped me with the approval as we ll as being a good friend. I would like to thank animal care service members, es pecially Kristina Stei nfeldt, who has been very friendly and supportive since the day we met.


5 Outside UF, and most importantly, I would lik e to thank my parents, Mine and Arslan Ucar, my sister, Derya and Kaan Sehri, my br other, Yigit Ucar and his family. Although they live in Turkey, whenever I need th em, they did not hesitate to trav el thousand miles to be beside me. I do appreciate them for their support and car ing. I would like to thank my friends, Brian Motyer, Shuhong Han, Qing Yang, Jennifer Barrel l, and Yeliz Gedik, who have provided strength, wisdom, motivation, and love. Finally, I thank my parrot, Murphy, who has brought me so much joy, and endured listening practice of my final thesis defense. I DO LOVE ALL OF THEM.


6 TABLE OF CONTENTS page ACKNOWLEDGMENTS...............................................................................................................4 LIST OF TABLES................................................................................................................. ..........8 LIST OF FIGURES................................................................................................................ .........9 ABSTRACT....................................................................................................................... ............10 CHAPTER 1 INTRODUCTION TO CANCER INITIATING CELLS.......................................................12 Cancer Stem Cell Hypothesis.................................................................................................12 Origin of Cancer Initiating Cells............................................................................................14 Evidence of Cancer Initiating Cells........................................................................................15 Characterized CICs of Solid Tumors......................................................................................16 2 ALDEHYDE DEHYDROGENASE (ALDH) AND CANCER.............................................20 Aldehydes...................................................................................................................... .........20 ALDH........................................................................................................................... ..........22 Enzymatic Activities of ALDHs.....................................................................................23 Role of ALDHs in Resistance to Drugs..........................................................................24 Deficiencies of ALDH.....................................................................................................25 ALDH Expression in Stem Cells.....................................................................................25 Effects of ALDH on Cell Prolif eration and Differentiation............................................28 ALDH Expression in Cancer Cells..................................................................................31 ALDH as a Selective Marker for CICs............................................................................32 3 MATERIALS AND METHODS...........................................................................................33 Cell Lines..................................................................................................................... ...........33 Analysis of ALDH Activity....................................................................................................35 ALDEFLUOR Assay.......................................................................................................35 ALDH Enzyme Activity Assay.......................................................................................36 FACS Analysis and Sorting of ALDEFLUOR Stained Cells................................................38 FACS Analysis................................................................................................................39 Cell Sorting................................................................................................................... ...39 Non-Obese Diabetic/Severe Combined Imm unodeficient (NOD/SCID) Xenotransplant.....40 Animals........................................................................................................................ ....40 Irraditation of Mice..........................................................................................................41 Transplanting Cells into NOD/SCID Mice..................................................................... 42 Tumor Evaluation............................................................................................................42 Immunohystochemistry..........................................................................................................43


7 Sectioning and Preparation of Paraffin Embedded Tissues............................................43 Sectioning and Preparation of OCT Blocks....................................................................45 CD133 Staining and FACS Analysis of Cells........................................................................47 Cell Cycle Analysis............................................................................................................ ....47 Reagents....................................................................................................................... ...........48 RPMI Complete Media....................................................................................................48 DMEM Complete Media.................................................................................................49 ALDEFLUOR Kit...........................................................................................................49 Matrigel High Concentration...........................................................................................50 Tumor Enzyme Digestion Buffer....................................................................................50 4% Paraformaldehyde (100L ).......................................................................................50 Immunocytochemistry.....................................................................................................51 Flow Cytometry Staining Buffer.....................................................................................52 Flow Cytometry Running Buffer.....................................................................................52 Vindelov Method Stock Solutions...................................................................................52 4 RESULTS AND DISCUSSION.............................................................................................54 Results........................................................................................................................ .............54 In Vitro ALDH Expression of Huma n Lung Cancer Cell Line H522.............................54 ALDHbr Cells Give Rise to Both ALDHbr and ALDHlo Cells........................................56 ALDH Levels Altered th e Tumor Growth Rate..............................................................56 ALDH Enzyme Activity Decreased in vivo.................................................................... 57 ALDHbr Primitive CICs Were More Tumorigenic..........................................................58 ALDH Expression Pattern of CICs Within Tumor Bulk.................................................58 Staining of H522 Cells for Other CIC Marker CD133....................................................59 MIA PaCa-2 Human Pancreatic Can cer Cell Line ALDH Expression...........................59 Discussion..................................................................................................................... ..........60 LIST OF REFERENCES............................................................................................................. ..75 BIOGRAPHICAL SKETCH.........................................................................................................80


8 LIST OF TABLES Table page 4-1 Measurement of ALDH activity and pr oteins in human l ung cancer cell lines...................63 4-2 Limiting dilution assay tumor growth ra tes: The dose groups and the weekly tumor volume measurements [length x (width) 2 x 0.52] (mm3). *(Standard Deviation).............63 4-3 Serial transplantation assay tumor grow th rates: The dose groups and the weekly tumor volume measurements [length x (width) 2 x 0.52] (mm3). *(Standard Deviation)..................................................................................................................... ........64


9 LIST OF FIGURES Figure page 4-1 Western blot analysis of ALDH1A1 and ALDH3A1 prot eins expressions in human lung cancer cell lines......................................................................................................... ...64 4-2 ALDH enzyme activity assay of first and second sets of sorted H522 cells........................64 4-3 Representative subcutaneous tumors [due to injection of (1x105) ALDHbr (right eight weeks post transplantation; tu mor volume= 3594mm3) and (1x105) ALDHlo (leftfive-weeks post transplantation; tumor volume= 4159mm3) H522 cells. ]..............65 4-4 H&E staining............................................................................................................... .........65 4-5 The FACS analysis of AL DEFLUOR stained tumor cells..................................................67 4-7 Tumor growth chart of first transplanted animals................................................................68 4-8 Tumor growth chart of second serially transplanted animals..............................................68 4-9 Tumor growth chart of third serially transpla nted animals..................................................69 4-10 ALDH1A1 and ALDH3A1 st aining of tumor samples........................................................69 4-11 ALDEFLUOR based FACS analysis of the MIA PaCa-2 (20%) and BEAS-2B (80%) cells.................................................................................................................... ........72 4-13 Cell cycle analysis of fr actionated MIA PaCa-2 cells..........................................................73


10 Abstract of Thesis Pres ented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Master of Science A NEW APPROACH FOR THE ENRICHMENT OF LUNG CANCER INITIATING CELLS BY ALDEFLUOR STAINING By Deniz A. Ucar December 2007 Chair: Edward Scott Major: Medical Sciences The cancer stem cell hypothesis proposes that only a subpopulation of cells within the tumor referred to cancer initiating cells (CICs) can generate tumors. The CICs have the properties of adult stem cells, in particular the ab ility to self-renew and di fferentiate into multiple cell types. Understanding the molecular biology of the cancer stem cells is crucial for the development of more effective cancer treatments. At present, several cell surface markers have been identified that are useful for enrichment of CICs. Aldehyde dehydroge nase, ALDH class1A1 and cla ss3A1 (ALDH1A1 and ALDH3A1) expression in lung cancer might become potential markers for the characterization of lung CICs. Using ALDEFLUOR flow cytometry based assay, th e laboratory of Dr. Moreb was able for the first time to measure ALDH activity in solid tumor cells and document heterogeneity in expression among lung cancer cell lines. The ALDEFLUOR staining system allows fo r the identification and isolation of stem/progenitor cells based on their ALDH enzy matic activity, rather than their cell surface phenotype. The laboratory of Dr. Jan Moreb ha s successfully fractionated human lung cancer cell lines, using the ALDEFLUOR staining method fo r subsequent in vitr o studies. In culture,


11 ALDHbr cells display stem cells-lik e behavior including slow grow th rates and resistance to chemotherapeutic agents. We hypothesize that sorting for ALDHbr cells will enrich for CICs that are capable of generating tumors This study examines the in vi vo functional differences between the ALDHbr and ALDHlo sorted cells from the human lung cancer cell line H522 to establish a useful experimental system for studying CICs. This hypothesis has been functionally tested using a xenograft non-obese diabetic/severe combined immunodeficient (NOD/SCID) mouse model to assay for tumor formation. The prelimin ary results show evidence that higher levels of ALDH altered the in vivo tumor growth rate and was consistent with their in vitro behavior. Tumors derived from ALDHbr cells grew slower than tumors initiated from ALDHlo cells. The same phenomenon was observed at all cell doses th at were tested and including a range from 5005 cells per dose. The serial transplantatio n experiments demonstrated that both ALDHbr and ALDHlo cells, which were isolated from primary tumors, could generate secondary tumors. Moreover, as predicted from the in vitro studies, ALDHbr cells generated secondary tumors containing ALDHlo cells. These in vivo results establishe d a correlation between low levels of ALDH activity and enhanced tumor growth. This study has recently been extended to the human pancreatic cancer cell line MIA PaCa2 to determine whether the ALDH-based enrichme nt method can be used for other types of human cancers. The in vitro experi mental results were consistent with the observations using the human lung cancer cell line H522. ALDHbr sorted cells grew slower in culture compared to ALDHlo cells. These results were confirmed by quantifying their ALDH enzyme activity and determining the cell cycle pr ofile of the sorted ALDHbr and ALDHlo populations. Overall, the preliminary results indicate th at ALDEFLUOR based sor ting has the potential to enrich for cancer initiating cell activity fr om multiple human tumor-derived cell lines.


12 CHAPTER 1 INTRODUCTION TO CANCER INITIATING CELLS Cancer Stem Cell Hypothesis There is increasing evidence that malignant tu mors are initiated and maintained by cancer initiating cells (CICs). Normal adult stem cells ar e defined by three main properties namely; selfrenewal, differentiation, and hom eostatic control (Sanders et al. 2006). Like normal tissue stem cells, CICs have potential to proliferate and mainta in their stemness. They can also give rise to aberrantly differentiated cells that cause the he terogeneity observed tumor cell populations. They have the capacity to modulate their prolifera tion and differentiation acco rding to environmental stimuli and genetic restrains (Dalerba et al. 2007). CICs, like normal stem cells, retain the essential property of self-protection through the activity of multiple drug resistance transporters and anti-apoptotic mechanism (Dalerba et al. 2007). Within a tumor, cells with stem-like properties have been reported to exist in leukemia (Bonnet and Dick 1997), brain cancer (Singh et al. 2003), breast cancer (Al-Ha jj et al. 2003), colon cancer (OB rien et al. 2007), pancreatic cancer (Li et al. 2007), and liver can cer (Shengyong et al. 2007). CIC is a subset of the tumor that can give rise to heterogeneous tumor cell populations similar to the originating tumor. According to this concept, not every cancer cell can produce a tumor. These cells have indefinite proliferative po tential to control the fo rmation and growth of a tumor. The CIC theory was founded on the premis e that tumor formation, growth and metastasis are governed by a distinct sub-popu lation of cancer cells. In essen ce, the theory states that tumorogenesis and organogenesis are similar in many respects. After all, normal stem cells enable organ generation in the earliest stages of development and allow regeneration of tissues throughout the life span of an indi vidual (Sanders et al. 2006). In addition to maintaining their


13 de-differentiated state by means of self-renewal, stem cells also gi ve rise to progenitors that upon successive divisions differentiate to form cells of multiple lineages. Indeed, stem cells are among the longest live d cells present in the organism and are therefore more likely to accumulate the necessary mutations to render them tumorigenic (Buzzeo et al. 2007). The concept that cancer/leukemia stem cells exist and contribute to the relapses and lack of cure has inflamed this area of research and the importance of CI C has been demonstrated for several tumor types (Bonnet and Dick 1997, Si ngh et al. 2003, Al-Hajj et al. 2003, OBrien et al. 2007, Li et al. 2007, Shengyong et al. 2007). The isolation, ch aracterization and functional analysis of these tumorigenic en tities have been facilitated by advances in exogenous culture, cell sorting and mouse xenografting techniques. One way to facilitate the proper identification of stem cells is the establishment of common stem markers. Attempts to do that have relied on taking molecular and bi ological markers of well defined stem cell populations, such as embryoni c stem cells, hematopoietic stem cells, and mesenchymal stem cells and applying those marker s to candidate stem cells from the different tissues (Cogle et al. 2003). The term stemness refers to all these common markers as well as the critical biological functions of multipotentiality in terms of the self renewal potential and the ability to differentiate to the various downstream mature progenitors. Attempts to establish a stem cell molecular signature using gene profil ing have been hampered by the difficulty of obtaining a pure population of stem cells. It will be important for alde hyde dehydrogenase, ALDH1A1 and/or ALDH3A1 to be a marker of a progenitor other th an the cancer stem (Moreb et al. 2007). If indeed ALDH activity through ALDH1A1 and/or ALDH3A1 expression is pr oven to be one of the stemness markers,


14 then that would be a great leap towards iden tifying CIC in other tumors and designing new treatments targeting these CICs. The presence of a stem cell population in a tumor has implications for the diagnosis, prognosis and even more profound implications for the treatment of cancer, as it is these CICs that must be targeted to achieve cure. Origin of Cancer Initiating Cells There are currently two hypotheses for the origin of the cancer stem cells (Dalerba et al. 2007, Tan et al. 2006). The first pers pective suggests that tumors arise from normal adult stem cells. Long-term tissue derived stem cells have a higher potential to ac cumulate mutations and transform into tumor generating stem cells (T an et al. 2006). A normal stem cell may be transformed into a cancer stem cell through dysfunction of the normal proliferation and differentiation pathways. The most convincing ev idence supporting this scenario is leukemias (Dalerba et al. 2007). The second scenario is the gene ration of tissue-specif ic progenitors with stem cell-like properties. Genetic alterations in regul atory components of th e signaling pathways controlling self-renewal and resi stance to apoptosis may bestow normal progenitor cells to behave like CICs and be able to form tumors (Tan et al. 200 6). Upon exposure to a carcinogen, quiescent cells present in a tissue may alter its ge ne expression pattern, turn on growth promoting genes, and subsequently the ability to form can cer. These events have been observed in blast crisis chronic myelogenous leukemia (CML), a committed granulocyte-macrophage progenitor, to gain self-renewal capacity a nd thus reacquire stem-like pr operties due to the effects of mutations (Reya et al. 2001). Each of these hypothe ses has its own rationale yet to be definitively proven.


15 Evidence of Cancer Initiating Cells Indeed the concept of cancer stem cells was proposed many years ago, and CIC has gained significance in the last decade. In the 1960s, drug-resistant cancer stem cells were initially isolated from blood cancers. Tritium-labeling studi es allowed scientist to observe that primitive cancer cells capable of generati ng post mitotic progeny, evidenced the existence of a leukemic stem cell (Clarke et al. 2006). The functional existence of CICs is demons trated by in vivo self-renewal assay (Clarke 2005). This assay is based on rec overing single-cell suspensions from the cancer of interest. In some cases, the recovered cell population is furthe r fractionated or enriched using cell surface markers. The recovered cancer cells are inje cted into immunodeficient mouse models, and the subsets are compared with respect to their tumo r forming capacity. Accord ing to the CIC model, only a specific subset of the cancer cell population can sustain in vivo tumor growth, whereas all other subsets have a very limite d proliferative potential and do not form tumors or cancer. Recently, this supposal has been confirmed in several tumor systems (Dalerba et al. 2007). Three key components define the existence of a CIC population: 1. Only a small subset of cancer cells within each tumor is capable of initiating tumors when transplanted into immunodeficient mice. 2. These subset of cells differ from nontumorge nic cells by a distinctiv e profile of surface markers or selective properties, and thereby allo wing their enrichment by flow sorting or other immunoselection procedures. 3. Cancer initiating cells yi eld a mixed cell populations c onsist of self-renewing cancer initiating cells and nontumorgenic cells. Bonnet and Dick reported the first evidence of CICs in 1997 (Bonnet and Dick 1997). A subpopulation of the CD34+/CD38leukemic cells initiated can cer in non-obese diabetic/severe combined immunodeficient (NOD/SCID) mice that wa s similar to the patients. The frequency of the CIC in human acute myeloid leukemia bei ng used in this study was detected to be less


16 than 1 in 10,000 (Bonnet and Dick 1997). The exis tence of leukaemic CICs incited further research into other types of can cers. Today, CICs have been iden tified in several solid tumors, including brain cancer (Singh et al. 2003), breas t cancer (Al-Hajj et al 2003), colon cancer (OBrien et al. 2007), pancrea tic cancer (Li et al. 2007), a nd liver cancer (Shengyong et al. 2007). Characterized CICs of Solid Tumors Following the concept of the leukaemic cancer stem cells, investigators pursued the characterization of cancer initiating cells in solid tumors. In solid tumors, CICs were first defined in the brain tumors (Ignatova et al. 2006). Using an in vitro ne urosphere assay system, single cell suspensions prepared from clinical biopsies of human brain cancer were observed to form spheres, culturing single cell su spensions of the tumor cells in a media to observe formation of spheres, performed using clinical biopsies of brain cancers. They found that approximately 0.1% of the tumor cells, which shared the expression patte rn of neural stem cells, were able to form neurospheres (Singh et al. 2003, Igna tova et al. 2006). Subsequent in vivo studies functionally characterized the CICs of brain tumors using th e neural stem cell marker, CD133, for enrichment (Singh et al. 2004, Sheila et al. 2004). Both CD133+ and CD133sorted populations were transplanted into the fr ontal lobes of NOD/SCID mice. As few as 100 CD133+ cells were able to initiate tumor, whereas CD133yield no tumors. Breast cancer initiating cells have been defi ned by Al-Hajj et al.. (Al-Hajj et al. 2003). Metastatic breast carcinomas were examined for their cell surface marker expressions and fractionated a subpopulat ion with the ESA+/CD44+/CD24-/low profile was shown to be enriched for the breast cancer initiating cell s. In vivo, limiting dilution assay re sults revealed that as little as 100 of fractionated cells were capable of recapitulating the primary tumor phenotype when infused into NOD/SCID mice, yet the unfractur ed or CIC counterparts did not grow de novo


17 tumor (Al-Hajj et al. 2003). To characterize the si gnaling pathways that are functioning as breast CICs, the researchers applied in vitro sphere formation assay to breast CICs. Mammospheres grown under different culture conditions were comp ared for the expression of genes involved in stem cell self renewal (Liu et al. 2006). Hedge hog (Hh), which functions in self-renewal, was increased in the CIC population compared to thei r counterparts, indicating that Hh signaling may be a mechanism of self-renewal of mammary CICs (Liu et al. 2006). Besides brain and breast tumors, interesting resu lts have been collected on prostate cancer, where CIC subpopulation appears characterized by the expression of CD44 (Collins et al. 2005). In prostate cancer and benign prostatic hyperplasia, a small population of human prostate basal cells that express CD133 and the marker of stem cells in prostate epithelia alpha2beta1, could reconstitute prostatic-l ike acini in nude mice (Richards on et al. 2004). Later on, the group reported that the CICs from human prosta te tumors were characterized by their CD44+/A2B1hi/CD133+ expression pattern (Patrawala et al. 2 006). They showed that in vitro this cell population possesses a significant capacity for extensive prolifer ation, self-renewal, differentiation, and invasion. In melanoma, in vitro culture of primary tu mor cells with embryonic stem cell media was used for the enrichment of melanoma spheroids forming cells that are capable of long-term self-renewal, multi-lineage differentiation and in vivo tumorigenicity (Fang et al. 2005). John Dick and Ruggero De Maria, in two diffe rent studies, have also shown that primary human colon carcinoma cancer in itiating cells can be enrich ed based on CD133 expression (OBrien et al. 2007, Ricci-Vitaiani et al. 2007). The in vitro long term culturing capacity of the CD133+ cancer initiating cells was confirmed by primary xenograft experiments into NOD/SCID mice and serial transplantat ion of the engrafted CD133+ colon cancer cells. Subcutaneous and


18 subrenal capsule implantation of the CD133+ colon cancer resulted in faster tumor growth, and formed tumors were similar in morphol ogy to the original tumors, whereas CD133cell fraction was either unable to initiate tumors or could not be serially transplanted into the secondary host. Tumor-initiating frequency was estimated at one in 3000 CD133+ colon cancer cells. Among the three tested cell lines for huma n hepatocellular carcinomas that were characterized, Huh-7 cells was found to be expressed CD133 (Suetsugu et al. 2006). The CD133+ cell population of this cell line has higher in vitro proliferative potential and forms tumors in immunodeficient mice; however, CD133cells did not induce tumor formation. Similarly, the in vitro and vivo studies done by another group, using primary human hepatocellular carcinoma biopsies, suggested that as few as 100 CD133+ cells inoculated into NOD/SCID mice initiated tumors, whereas CD133cells could not form tumor (Yin et al. 2007). Thus, the identif ication of CD133+ cells may be a potential marker to enrich for cancer initiating cells in hepatocellular carcinomas. At present, pancreatic CICs have been charac terized using a similar sc hema as used for the enrichment of breast CIC population (Olemp ska et al. 2007). Using the NOD/SCID mice xenotransplant functional assay, th e investigators found that CD44+CD24+ESA+ fraction had100fold higher tumorigenicity when compared to non-CIC fraction (Olempska et al. 2007). Furthermore, this pancreatic CIC population be stowed self-renewal capacity and expressed the stem cell signaling molecule, sonic Hh. Further study of this pancreatic CIC population may lead to improved treatment regiments by combining ch emotherapy, radiation, and use of therapeutics which block the Hh signaling pathway, thereby inducing CIC differentiation. More recently, cells have been isolated from a human lung cancer ce ll line which had been sorted for Hoechst dyeexcluding side population ( SP) cells and were enriched for CICs (Ho et


19 al. 2007). Modeling the hematopoietic stem cells (HSCs) isolation by th eir ability to efflux Hoechst 33342 dye, researchers used flow cy tometry and Hoechst 33342 dye efflux assay to isolate and characterize SP cells from six human lung cancer cell lines (H460, H23, HTB-58, A549, H441, and H2170). To test the tumorigenicity of the isolated SP, they compared the in vivo and in vitro behavioral pr operties of SP and non-SP cells. According to Matrigel invasion assay, SP cells had higher potential for invasive ness. Further characterization of this SP phenotype displayed several stem cell properties including self-renewal, elevated expression of ABCG2 as well as other ATP-binding cassett e transporters, resistance to multiple chemotherapeutic drugs, and elevated telomerase reverse transcriptase expression. In addition, the SP was found to be composed of cells in the G0 quiescent state based on their lower mRNA levels of minichromosome maintenance (MCM) 7, a member of the MCM family of proteins which is critical to the DNA replication complex. The tested sixteen clin ical lung cancer samples were also displayed existence of a smaller SP population. Their results evidenced that SP is an enriched source of lung CICs with stem cell properties.


20 CHAPTER 2 ALDEHYDE DEHYDROGENASE (ALDH) AND CANCER This chapter is devoted to aldehydes and their metabolizing enzyme ALDH, and their respective role in tumorogenesis. Aldehydes ar e highly reactive molecules that have both positive and negative effects on biological systems. For example, retinoids, vitamin A chemical metabolites, are converted to essential aldehydes that benefit vision, correct the function of epithelial cells, glycoprotein synthesis, and erythropoiesis (formati on of red blood cells) (Pijnappel et al. 1993). However, some aldehydes, as in the case of metabolite of nitrosamine that causes urinary bladder cancer, which may have cytotoxic, mutagenic, and carcinogenic effects. In this chapter, the expre ssion pattern of ALDHs in normal tissue, stem cells, and cancer cells will also be discussed. The principle focus in this chapter will be ALDH1A1 and ALDH3A1, which are cytosolic enzymes expressed in both normal and cancer cells. Further investigation in establishing an intimate correlation betw een the ALHD expression and tumorogenesis will allow further insight in dete rmining the feasibility of using ALDH activity as a parameter for the enrichment and the targeting of the CICs. Aldehydes Aldehydes are organic compounds that c ontain a terminal carbonyl group. These compounds are very rich in th e environment. Aldehydes comprise a major portion of the carbonyl products. They exist in cigarette smoke, many foods, fruits and vegetables giving them the flavors and odors (Lindahl 1992). However, th e majority of aldehydes are produced in the body as a result of the metabolism of other co mpounds as in the case of endogenous aldehydes derived from the metabolism of amino acids, ca rbohydrates, vitamins, ster oids, and lipids and include acetaldehyde from threonine catabolism, 21-dehydrocorticosteroids from corticosteroid


21 catabolism, and retinal from vitamin A me tabolism (Lindahl 1992). Among the endogenous aldehydes, the lipid peroxidation products have become importan t due to their carcinogenic effects (Lindahl 1992). In this group the mo st common aldehyde formed is malondialdehyde (MDA) which comprises 70%, the hexanal (HA) comprising 15%, and the 4-hydroxy-2-nonenal (HNE) comprising 5% of the total aldehydes produ ced (Townsend et al. 2001). The serum levels of aldehydes derived from lipid peroxidation ha ve been documented to increase in cancer patients, and the pattern of aldehydes appeared to be unique for each cancer type (Yazdanpanah et al. 1997). Peroxly radicals are the primary free radicals interm ediate formed during the lipid peroxidation during the oxidative stress. This ch ain reaction results in the production of reactive aldehydes that can react with the DNA and may introduce mutations via base alkylation (Yazdanpanah et al. 1997). Methylation, for exam ple, is the most common type of alkylation. Neoplasias are characterized by "methylation im balance" where hypomethylation and localized hypermethylation appear at the same time in the genome (Corley et al. 1994). In this scenario, hypomethylation may lead to chromosomal instabili ty and increased mutati on rates. Therefore, the methylation state of genome can be a promis ing biomarker for tumorogenesis for prostate cancer; hypermethylation of the pi-c lass glutathione S-transferase ge ne (GSTP1) appears to be a promising diagnostic indicator of prosta te neoplasias (Baylin et al. 1998). On the other hand, majority of the exogenous aldehydes are xenobiotics. Xenobiotics are chemical compounds that are found in an organi sm, but they are not normally produced or expected to be present. This term can also m ean presence of an excessive amount of a compound than its usual concentration. In most cases, xenobiotics are drugs or environmental pollutants that are found at low levels. Xenobio tic aldehydes can be produced via biotransformation of a large number of drugs and other xenobiotics. In this group, one of the most im portant xenobiotics is


22 antitumor prodrugs. For instance, metabolism of the antitumor agent cyclophosphamide (CP) produces the active form of the alkylating drug, 4-hydroxycyclophosphamide/ aldophosphamide (Nakayama et al. 2004). Aldehydes are in genera l long lived and highly reactive. Thereby, the transformation of a prodrug to an active aldehyde fo rm may take place in the liver, and than the active aldehyde product can diffuse or be transported to the target site. Although most aldehydes have harmful biological effects in the body such as cytotoxicity, mutagenicity, and carcinogenicit y, some aldehydes are known to have carcinostatic effects (Sovic et al. 2001). These carcinos tatic aldehydes appear to be byproducts of lipid peroxidation. For example, the lipid peroxidation product 4-hydroxynonenal (HNE) acts as a cell growth modulator if used at low concentrations; how ever, it is also strongl y cytotoxic at higher concentrations for a number of cell types (Sovic et al. 2001). Ho wever, these aldehydes have a short half life and severe si de effects that prevent th eir clinical usefulness. In the body, aldehydes can be metabolized by aldehyde oxidase, aldo-keto reductase, and ALDHs (Lindahl 1992). An aldehyde oxidoreductase catalyzes the oxidat ion of a variety of organic aldehydes and N-heterocyclic compound s to carboxylic acids, and also oxidizes quinoline and pyridine derivatives. Similar to ALDH, aldo-keto reductase is a cytosolic enzyme that uses NADH as a cofactor, and reduce a va riety of aldehydes and ketones to their corresponding alcohols (Lindahl 19 92). In contrast to aldehyde oxidase and aldo-keto reductase, ALDH are ubiquitous in every cellular compartm ent of the body and have a broad substrate specificity that will be discu ssed further in this chapter. ALDH ALDHs are a group of NAD(P)+-dependent enzyme s involved in the metabolism of a wide variety of aliphatic and aromatic aldehydes ge nerated from various endogenous and exogenous precursors to their corresponding carboxylic acids (Sldek et al. 2002).


23 ALDHs are the major aldehyde metabolizing enzymes. Although their primary role isdetoxifying the cells, through cat alyzing oxidative reactions of some xenobiotics, they can produce harmful metabolites. This group of enzy mes, in particular ALDH1A1, also plays a significant role in chemotherapeutic drug resist ance via converting the active form of the drug (4-hydroxycyclophosphamide/ aldophosphamide) into harmless carboxyphosphamide. Many allelic variants within the ALDH gene family ha ve been identified, resulting in pharmacogenetic heterogeneity between individuals which, in mo st cases, results in distinct phenotypes as discussed below. Many disparate aldehydes are ubiquitous in nature and are toxic at lo w levels because of their chemical reactivity. Thus levels of metabolic-intermedi ate aldehydes must be carefully regulated which explains the existence of several distinct ALDH families in most studied organisms with wide constitutive tissue distribution. All ALDHs require either NAD or NADP as a cofactor. Changes in ALDH activity have been observed during experimental liver and urinary bladder carcinogenesis and in a number of human tumors. Currently, 19 genes in the mouse and rat, and 18 human putative pr otein coding sequences, that have a conserved ALDH "signature sequence, have been identified (Sldek et al. 2002). These enzymes were classified into three gr oups. Cytosolic class 1, m itochondrial class 2, and class 3, which is generally associated with tumors and other isozymes. They can be constitutively expressed as well as their expression can be induced by xenobiotic factors. Enzymatic Activities of ALDHs ALDHs have broad substrate specificity. Ox idation of aldehydes is classified as a detoxification reaction. ALDHs are the main responsible enzymes in the removal of acetaldehydes (Tabeke et al. 2001). The most we ll studied function of these enzymes takes place in the alcohol metabolism. Following ethanol ingestion, alcohol dehydr ogenase (ADH) oxidizes


24 ethanol to acetaldehyde, and ALDH oxidizes this substrate to acet ic acid, which can be further metabolized in the body as an energy sour ce. Similarly, ALDH detoxifies acrolein, the hepatotoxic metabolite formed from allyl alcohol. ALDH family enzymes are also responsible for the vitamin A metabolism in which this family of enzymes irreversibly converts reti naldehydes to the retinoi c acids (RA). As we mentioned earlier, retinoic acid s are crucial compounds in the retinoic acid signaling pathway. The RA signaling is known to have a significant ro le in hematopoietic stem cell self-renewal and differentiation which will be discussed in more details in a separate section. ALDHs can also behave like esterases, whic h hydrolyze esters such as para-nitrophenyl acetate (Lindahl et al. 1992). As mentioned earlie r, cancer patients have an increase in the serum levels of aldehydes such as hexanal (HA), tran s-2-octenal (t2OE), tran s-2-nonenal (t2NE), and malondialdehyde (MDA) have been shown to be oxidized by either cl ass 1 or class 3 ALDH (Townsend et al. 2001). In the case of 2-but oxyethanol and alcohol, aldehyde dehydrogenases sequentially catalyze the formation of the hemato -toxic metabolite, 2-butoxyacetic acid (Pereira et al. 1991). Role of ALDHs in Resistance to Drugs More importantly, ALDHs are one of the major responsible enzymes in drug resistance especially against chemotherapeutic agents namely oxazaphosphorines (Moreb et al. 2007b). Cellular sensitivity to cyclophosphamide and other oxazaphosphorines, including 4hydroperoxycyclophosphamide (4HC), mafosfamide and ifosfamide, is closely related to ALDH1A1 and ALDH3A1. These enzymes are show n to catalyze the detoxification of these agents (Sldek et al. 2002). In a different study, cytosolic ALDH1 has been shown to be associated with cyclophosphamide (CTX) and me thotrexate (MTX) resistance in hematopoietic cells (Takebe et al. 2001).


25 In addition to their catalytic properties, so me ALDH proteins functi on as binding proteins that have non-catalytic inte ractions with chemically diverse endogenous or exogenous compounds. ALDH1A1 has been identified as an androgen-binding protein in human genital fibroblasts (Pereira et al. 1991), and a fla vopiridol-binding protei n in non-small cell lung carcinomas (Schnier et al. 1999). Flavopiridol induces reversible G1 and G2 phase cell cycle arrest via inhibiting cyclin kinase D phos phorilation independently from p53 and pRB regulations. Therefore, the synthetic flavopiridol has a poten tial to be used as an antitumor agent, yet the presence of ALDH1 has shown to prevent its function via binding to flavopiridol rather than catalyzing its reducti on (Schnier et al. 1999). Deficiencies of ALDH Deficiency of these enzymes or due to ge ne polymorphisms, the coding region may cause inactivation of the enzymes, resulting in differe nt defects including intolerance to alcohol and increased risk of ethanol-induced cancers (A LDH2 and ALDH1A1) (Yokoyama et al. 2003). Examples of ALDH deficiencies related to gene polymorphisms include: Sjgren-Larsson syndrome (ALDH3A1) (Carney et al. 2004), type II hyperprolinemia (ALDH4A1) (Geraghty et al. 1998), 4-hydroxybutyric aciduria (ALDH5 A1) (Gordon 2004), developmental delay (ALDH6A1) (Vasiliou and Nebert 2005), hypera mmonemia (ALDH18A1), and late onset of Alzheimers disease (ALDH2) (Marchitti et al. 2007). ALDH Expression in Stem Cells The intracellular level of ALDHs been shown to be higher in primitive hematopoietic cells. In 1990, a group of investigators evaluated th e relative levels of cytoplasmic ALDH in fractionated normal human bone marrow cells (Kasta n et al. 1990). They found that there is a direct correlation between the maturation stage and the level of ALDH protein. They have shown that the most primitive hematopoietic stem ce lls express the highest level of ALDH, while


26 immature erythroid cells express ALDH at inte rmediate levels and lymphocytes express the lowest level. (Kastan et al. 1990) The hematopoietic system is one of the mo st studied stem cell research area that continually progress. One important aspect in th is field is the isolation of the most primitive HSCs. In the course of time, a variety of purif ication strategies have been developed. Most commonly, cells have been fractionated based on the expression of cell surface markers such as Sca-1, thy-1, CD34, CD38, HLA-DR, and c-kit (K iel et al. 2005). La ter on, the signaling lymphocyte activation molecular (SLAM) fam ily of cell markers (CD150, CD244, CD48) has been shown to be expressed on hematopoietic st em and progenitor cells. SLAM family receptors distinguish hematopoietic stem and progenitor cells and reveal endothelial niches for stem cells. Specifically, long term HSCs identif ied by their expression of a CD150+CD244-CD48phenotype (Kiel et al. 2005). Ot her groups have fractionated th e bone marrow cell compartment to enrich for stem cells based on size and cell de nsity, efflux of metabolic dyes, or resistance to cytotoxic agents. In addition to these methods, investigators developed a simpler method for enriching for primitive human hematopoietic ce lls namely ALDEFLUOR staining. This system is based on conversion of a chemical compound to fluorescent substrate by an enzyme that is expressed at relatively high levels in stem cells compared with more differentiated cells and detecting using fluorescence-activ ated cell sorting (FACS). It has been known that HSCs are rich in ALDH activity (Storms et al. 1999). Th is group developed a fluorescent substrate for ALDH, called BODIPY aminoacetaldehyde (BAAA) which could be used to characterize primitive human hematopoietic cells. A successf ul strategy was developed by combining cell morphology and low complexity cells possess ing low light scatte ring properties (SSClo) to enrich for SSCloALDHbr from human umbilical cord blood cells This phenotype was depleted of


27 lineage-committed cells, 40% surface positive for CD34+CD38lo/cells, and enriched 50 to 100fold for primitive hematopoietic progenitors detected in shortand long-term culture analyses. Because stem cells, in general, share severa l common characteristics su ch as self-renewal, giving rise to different cell populations has prompted resear chers to examine stem cell populations for conserved patterns of ALDH expression. With respect to using ALDH expression as marker for stem cell activit y, ALDEFLUOR staining has become a useful technique for isolating stem cells from other tissues. For exam ple, brain derived neural stem cells (NSCs) have shown to contain high levels of ALDH activity besides the presence of ATPbinding cassette transporter G2 and high telomerase (Corti et al. 2006). In this study, an experimental system was established to enrich for primitive neural stem cells expressing high levels of ALDH from murine em bryonic and adult neurospheres. As demonstrated for HSCs, single-cell suspensions of neural cells can be stained with AL DEFLUOR and analyzed by flow cytometry (Storm et al. 1999). This study describe d a population of cells with low side scatters (SSClo) and bright ALDH (ALDHbr) activity. Comparative cell culture assays containing the fractionated cell populations evidenced that ALDHbr /SSClo cells were able to generate new neurospheres, thereby demonstr ating self-renewal, and the presence of both neurons and microglias confirmed the presence of neural stem -like cells. Subsequent immunocytochemistry (IHC) and real-time reverse transcriptionpolymerase chain reaction (RT-PCR) analysis substantiated their results. To eval uate the in vivo functionality of SSCloALDHbr cell population, the fractionated cells were injected into m ouse brain. The in vivo results demonstrated engraftment of donor-derived ne urons with mature morphology in the cortex and subcortical areas, thereby confirming the in vi vo differentiation capacity of the SSCloALDHbr cells. The


28 authors concluded that the ALDEFLUOR staining can be an effective method for the enrichment of NSCs, and may improve the development of a cell-mediated therapeutic strategy for neurodegenerative diseases. In a more recent study, ALDEFLUOR based fr actionation of umbilical cord blood (UCB) allowed researchers to isolate mo re functional endothelial progenito r cells (EPCs) (Nagano et al. 2007). EPCs are known to be involved in the repa ir of ischemic vascular injuries and may constitute an effective cell-base d therapy. UCB is a rich source of many stem cell types; however it is important to develop an e ffective method to isolate EPCs to prevent complications due to using mix cell populations for the cell therapy. Therefore, a combined strategy of negative immunoselection and cell-based cultu re expansion was used for th e enrichment of EPCs from UCB. In addition, the enrich ed cell population was furthe r fractionated based on ALDH expression levels. Unlike HSCs and NS Cs, this study indicated that the ALDHlo fraction of EPCs bestowed a higher abilit y to proliferate and migrate compared to ALDHbr cells. Moreover, low ALDH expression enhanced the response of th e cells to hypoxic conditions by up-regulating hypoxia-inducible factor proteins, VEGF, CX CR4, and GLUT-1, whereas high ALDH lacked this response. The in vivo testing in a mouse isch emia flap model also verified the capacity of ALDHlo EPCs to participate in the repair of damage d vessels. Thus, they have suggested that the isolation of ALDHlo EPCs appears to have a great potentia l for inducing rapid neovascularization and subsequent repair of ischemic tissues. Effects of ALDH on Cell Proliferation and Differentiation As mentioned previously, ALDH plays a pivotal role in retin oic acid (RA) biosynthesis and mediated RA signaling pathways. RAs bind to two nuclear receptors, RA-receptor (RAR) and the retinoid-X receptor (RXR) (Alvarez et al. 2000). Active isomers of RA, all-trans RA and 9-cis RA, induce transcription of genes through dis tinct mechanisms. All-trans RA activates gene


29 expression mainly through RAR/RXR heterodimers, whereas 9-cis RA activat es transcription of various genes via binding to RXR homodimers or heterodimers of RXR and other nuclear receptors (Alvarez et al. 2000). ALDH1A1 is one of the enzymes involved in the production of RAs. Dr. Moreb et al. has shown that all-trans RA (ATRA) inhi bit the enzyme activity of both ALDH1A1 and ALDH3A1 proteins (Moreb et al. 2005). The Northern and Western blotting analysis results revealed that the mRNA levels of ATRA treated cells were similar to untreated cells, yet protein levels were 50% reduced. Therefore, the aut hors have suggested that the inhibition of enzyme activity occurs at posttra nslational level and may involve the ubiquitinproteasome pathway. Reversible inhibition of ALDH enzyme activ ity lead investigators to successfully manipulate the differentiation of HSCs. Inhibitio n of ALDH and retinoid signaling promoted the expansion of human hematopoietic stem cells due to blocking di fferentiation and promoting self renewal (Marchitti et al. 2007). This study was the first to de monstrate that ALDH is a key regulator of HSC differen tiation. Blockage of ALDH enzymatic activity with diethylaminobenzaldehyde (DEAB) blocked the differentiation of human HSCs despite the stimulations with cytokines. This culture syst em resulted in a 3.4fold expansion of the most primitive human HSCs during short-term culture. To experimentally address the direct effects of the DEAB inhibition on the ALDH-dependent RA signaling pathway, the authors tested for reactivation of the RA signaling pathway utilizing the RAR agonist all trans-retin oic acid that resulted in successful engraftment of non-obe se diabetic/severe combined immunodeficient (NOD/SCID) mice. Their results suggested that the retinoic acids generated by ALDH is one component regulating HSC fate. Modulation of ALDH activity and retinoid signaling may become an effective approach to amplify human HSCs.


30 The role played by retinoic acid receptor alpha (RAR ) to mediate the si gnaling of retinoic acid (RA) to coordinate the prol iferation to differentiation trans ition has been of great interest. Further research to address th is question has targeted a mnage trois 1 (MAT1) subunit of RAR (Kastan et al. 1990). It has been proposed that MAT1 regulat es G1 exit, a cell cycle stage in which cells commonly will commit to proliferati on or to differentiation. Previous studies have shown that cyclin-dependent kinase-activating kinase (CAK) complex phosphorylates RAR In myeloid leukemia cells, the lack of RA-induced RAR -CAK dissociation and MAT1 degradation suppresses cell differentiation by i nhibiting CAK-dependent G1 exit and sustaining CAK hyperphosphorylation of RAR In contrast with the my eloid leukemia cells, normal hematopoietic progenitor cells have been found that MAT1 degradation occurs during granulocytic development due to dynamic expr ession of ALDH1A1 and 1B1, which catalyze RA synthesis. Inhibition of ALDH activity in terrupts MAT1 degradation and hence the differentiation of the cells. Meanwhile loss of RAR phosphorylation by CAK induces RA-target gene expression and diffe rentiation. These studies suggest that the RAR -CAK signaling is a crucial factor rendering the fate of cells dur ing granulopoiesis and determining if a normal myeloid or leukomogenic path is chosen. Apparently, ALDHs also play a pivotal role in cell proliferation and differentiation (Chute et al. 2006). Consistent with othe r investigators, this study has obs erved that there is an inverse correlation between the ALDH level of cells a nd the cell growth rate. High levels of ALDH reduce the growth rate of cells. Interestingly, in an experimental setting in which human corneal epithelial cell lines are stably expressing ALDH3A1 at a high level, the tr ansfected cells exhibit an elongated cell cycle, decrea sed plating efficiency, and re duced DNA synthesis. This group also observed that overexpression of ALDH3A1 is localized to both cytoplasm and the nucleus


31 of the cells. Closer examination of the transfect ed epithelial cells resulte d in reduced levels of cyclin A and cyclin B-dependent kinase activit ies, phosphorylation of th e retinoblastoma protein (pRb) the protein levels of cyclins A, B, and E, the transcription factor E2F1, and the cyclindependent kinase inhibitor p21 Th ese ALDH3A1 expressing cells e xhibited increased resistance to the cytotoxic effects of the DNA-damaging agen ts mitomycin C and Vp-16. It is important to notice that the corneal epithelium is a self-r enewing stratified epithelial tissue. ALDH Expression in Cancer Cells Alcohol dehydrogenase (ADH) and ALDH play a significant role in the metabolism of many biological substances. ADH part icipates in the metabolism of ethanol, retinoic acid, lipid peroxidation products, leukotrien e and glutathione metabolism. ALDH is responsible for oxidation of acetaldehyde and othe r aldehydes and metabolism of histamine and retinoic acid. The activity of total ADH and ALDH was also not significantly lower in the cancer cells (Jelski et al. 2004). The decrease of activity of class I ADH isoenz yme may be a factor of some disorders in metabolic pathways with particip ation of these isoenzymes that can lead to carcinogenesis. Acetaldehyde can be produced by co lonic bacteria and, to a much lower degree, by mucosal ADH. The human colon mucosa contains three classes of ADH isozymes: I, III, and IV. The second enzyme responsible for alc ohol metabolism in humans is ALDH, which catalyzes the oxidation of acetaldehyde to acetic acid in colorectal cancer. ALDH can be divided into two groups according to their Michaelis constant values for acetaldehyde ALDH1andALDH2 isoenzymes belong to the lo w-Km (3 M) forms, whereas ALDH3 and ALDH4 are the high-Km (5 mM) forms. Changes in expression of ALDH isomers have been reported in many cancer types (Jelski et al. 2006). Even within a certain cancer type there was variability in the ALDH levels. As


32 discussed in the Role of ALDHs in Resistance to Drugs, this group of enzyme has been a valuable parameter for the choosing effective treatment approaches. ALDH as a Selective Marker for CICs As previously discussed in the ALDH Expr ession in Stem Cells section, ALDEFLUOR staining method has become an important selectiv e marker for normal stem cells. Considering all the similarities between the adult stem cells and cancer initiating cells, we hypothesized that ALDH activity might be a useful selective marker for the CIC. To establish an in vivo system to test our hypothesis, we used ALDEFLUOR staining to fractionate the lung cancer cell line H522 into ALDHbr and ALDHlo group. The preliminary resu lts of this study will be discussed in the next chapter. In the well characterized he matopoietic system, ALDH fractionation has been applied to acute myeloid leukemias (Cheung et al .). Fifty-eight bone marrow (BM) samples were collected from AML, acute lymphoblastic leuke mia (ALL) and normal cases. In 14 AML cases, a high SSCloALDHbr 14.89% cell population was id entified (ALDH+AML), with the majority of the SSCloALDHbr cells coexpressing CD34+. In some cases, there was undetectable or rare SSCloALDHbr population (ALDH-AML). Consider ing other clinicopath ologic variables, ALDH+ AML was significantly associated with adverse cytogenetic abnormalities. CD34+ BM cells from ALDH+AML engrafted significantly better in NOD/SCID mice with the engrafting cells showing molecular and cytoge netic aberrations identical to the original clones. Normal BM contained a small 2.92% SSCloALDHbr population, but none of the ALL cases showed this fraction.


33 CHAPTER 3 MATERIALS AND METHODS This chapter covers the general methods and materials that were used during the course of this study. It is important to note that ther e were no human subjects used in this study that required IRB approval. Prior to animal usage in our experiments, all pr otocols were approved by the Institutional Animal Care and Use Committee of the University of Florida. All animal procedures were reviewed and approved by the University of Florida Animal Care and Use Committee, performed in an AALAC approved facili ty, and treated accordi ng to the regulations. Cell Lines The human non-small lung cancer cell (NSCLC) line H522 and the immortalized normal lung epithelial cell line BEAS-2B were generous ly provided by Dr. Jan Moreb (University of Florida, Gainesville, FL). The human pancreatic cancer cell line MIA Pa Ca-2 was obtained from Dr. William G. Cance (University of Florida, Ga inesville, FL). The H522 and MIA PaCa-2 cell lines were used as representative tumor-derived cancer cell lines to enrich for CICs. The BEAS2B cell line is known to express high leve ls of aALDH ans was used to optimize the ALDEFLUOR substrate and inhibitor levels. The H522 cell line (Ref: ATCC-CRL-5810) was derived from a 60 year old Caucasian male cancer patient prior to therapy. These cells carry the K-ras 12 mutation and a mutation in codon 191 of the p53 gene. This cell line is a hypotriploid cell lin e with the modal chromosome number occurring in 68% of cells counted. The polyploid cells occurred at 3.0%. The BEAS-2B epithelial cell line was isolat ed from normal human bronchial epithelium obtained from an autopsy of a non-cancerous indi vidual. The epithelial ce lls were transformed with an adenovirus 12-SV40 hybrid virus (Ad12SV40) for generation of the cell line. The cells retain the ability to unde rgo squamous differentiation in respons e to serum, and can be used to


34 screen chemical and biological agents for abil ity to induce or affect differentiation and/or carcinogenesis. The cells stain positively for keratins and SV 40 T antigen (Ref: ATCC-CRL9609). The lung epithelial and H522 cancer cell lin es were cultured in RPMI-1640 medium (Gibco Invitrogen) with 10% FBS (Gibco Invitr ogen) in 5% CO2 cell culture incubator at 37oC, and used within 2 passages when in the log phase of growth. The MIA PaCa-2 cell line (Ref: ATCC-CRL-142 0) was derived from pancreatic tumor tissue obtained from a 65-year-old Caucasian male in 1975. This is a hypotriploid human cell line. The modal chromosome number is 61. Si xteen to 20 marker chromosomes are commonly found in a cell. A few normal chromosomes are ab sent. The established ce ll line reportedly has a doubling time of about 40 hours and a colony-forming efficiency in soft agar of approximately 19%. The cell line is composed of two morphologically different cell populations; one is round shaped, and the other is spindlelike. The ratio of these cell po pulations were almost 1:1, and when the cells were sorted based on their AL DH activity, there was no si gnificant morphological difference between ALDHbr and ALDHlo cells. The only distinction between these cells that could be observed was their localiz ation during growth in a culture flask. Once the cells reached confluency, the spindle-like ce lls composed the adherent laye r, while the round shaped cells grew on top of the spi ndle-like cell layer. MIA PaCa-2 cells were cultured as reco mmended by ATCC in Dulbeccos modified Eagles medium (DMEM) with 4mM L-glutamin e adjusted to contain 1.5 g/L of sodium bicarbonate and 4.5 g/L of gl ucose, 87.5% with 10% FBS (Invitrogen) and 2.5% GIBCO Horse Serum (Invitrogen).


35 Analysis of ALDH Activity ALDEFLUOR Assay The ALDEFLUOR reagent system, which included the substrate Bodipyaminoacetaldehyde diethylacetal (BAAA-DA), wa s used the ALDEFLUOR assay. Cells were washed in 1xPBS for 5 minutes, and 1x106 cells were resuspended in 1 l of ALDEFLUOR assay buffer. For the negative control tube 5l of the specific ALDH inhibitor, diethylaminobenzaldehyde (DEAB), was added at a fi nal concentration of 50 mmol/l. Five l of ALDEFLUOR substrate was added to 1l of cells and immediately mixed, 0.5 l was transferred to the control tube. Therefore, each sample started with 1x106 cell concentration would be divided into two groups that one 0.5x106 cells with DEAB and substrate used for set up a background gate and the other half is the sample Samples were incubated for 45-60 minutes at 37oC. BAAA-DA is converted to Bodipy-ami noacetaldehyde (BAAA) by cytoplasmic ALDH. BAAA is a fluorescent substrate for ALDH. This substrate is uncharged and can freely diffuse across the plasma membrane of intact viab le cells. Cytoplasmic ALDH converts BAAA into Bodipy-aminoacetate (BAA), which is retained intrac ellularly because of its net negative charge, and therefore prevents free diffusion. The a ssay buffer supplied with the ALDEFLUOR kit contains a transport inhibitor, wh ich blocks efflux of the BAA from the cells. As a result, viable cells expressing high levels of ALDH retain BAA and thus fluoresce. After the incubation at 37oC, samples must be transferred to ice or 4oC refrigerator and protected from light by covering w ith aluminum foil until ready fo r FACS analysis or flow cell sorting. For FACS analysis, ALDEFLUO R stained samples are stable at 4oC for 24 hours, and there is no need to fix the cells to stabilize the fluorescent signal.


36 It is important to keep in mind that ALDEFLUOR staining method is based on the enzymatic activity of cells, theref ore there is a direct correlation between the total enzyme level of the sample and the amount of the substrate a nd the inhibitor used for the assay. According to Dr. Morebs observations, when the cell line ex presses high levels of ALDH, the ALDEFLUOR assay was not accurate (Moreb et al. 2007). It was most likely due to relative shortage of substrate and/or DEAB versus th e amount of enzyme activity in the tested cells. As shown in table 1, BEAS-2B cells do not express ALDH. Th ereby, the BEAS-2B cell line was utilized for optimization of ALDEFLUOR substrate and inhib itor levels for the high ALDH expressing cells. Because if the enzyme level of a sample is high suggested amount of the inhibitor might not be enough to evenly prevent the enzy matic activity of every cell. Ho wever, substituting a proportion of sample cells with BEAS-2B cells can reduce the total enzyme level of a sample while keeping the number of the total cells at the optimal leve ls. This approach may provide a better separation of ALDHbr and ALDHlo cells. Therefore, to double ch eck the accuracy of ALDEFLUOR staining, samples were analyzed by with or with out mixing 20:80 ratio of BEAS-2B cell line. ALDH Enzyme Activity Assay Measurement of ALDH enzyme activity was performed using a method that has been previously optimized by the laboratory of Dr. Ja n Moreb (University of Florida) (Moreb et al. 2007). Before starting the assay, the multi-channe l Beckman DLC 64 cuvette spectrophotometer (Beckman Coulter, Inc., Fullerton, CA) was turn ed on for at least 2 hours prior measuring activity to allow a sufficient amount of time for the light source and temperature controller to warm up. The reaction cuvettes we re also placed into the spec trophotometer during the warming step.


37 While the spectrophotometer was warming up, we started preparing our samples. Cells were grown to 70% confluence in a T-75 (Fis her Scientific) flask and removed from the incubator (37C, 5% CO2) 1 days after plating. The media from the flask was transferred to a 50 l falcon tube (Fisher Scientific) leaving th e pipette in the tube, 2.5 l of 0.25% TrypsinEDTA (Gibco, Invitrogen) added to each flask and incubated at 37oC for 5 minutes. Detached cells were transferred to 50 l tube and s pun down for 6 minutes at 1200rpm. The supernatant poured off and cells were resuspend in 3 l of 1X PBS. The number of viable cells was counted using a hemacytometer and the viability dye, trypan blue. In this essay, it is very important to quantify the number of viab le cells. A minimum of 5x105 cells were lysed in 0.5 l of buffer containing 50 mM Tris (p H 8), 25 mM EDTA, 5 mM -mercaptoethanol, 1 mM phenylmethylsulfonyl fluoride, and 0.1% sarcos yl. The suspension was mixed by vortex and centrifuged at 10,000 rpm for 10 min at 4C. Af ter the centrifugation, the supernatant was carefully transferred in to a clean 1.5 l microcentrifuge tube (Sigma-Aldrich) using a P1000 pipette and placed on ice. The rema ining lysis buffer was kept at 37oC for use in the assay. For the accuracy and consistency of enzyme activity assay, it is very important to be careful at the following steps. We used thre e cuvettes for each sample: one blank and two duplicates of reaction, and assayed each sample individually. Four pipetmen were set the following volumes: 552l, 30l, 12l, and 6l were required to quickly add the samples and reaction components to cuve tte. In the following sequential orde r, 552 l of lysis buffer, 30l of the protein extract, and 12l of NAD+ (0.068 grams of NAD in 370l water) added to each cuvette, followed by mixing well with a 200 l pipetman. At this step, we set the spectrophotometer for blank (zero) (Note: the blank was the mixt ure of lysis buffer, sample protein extract, and NAD+.) The two duplicates of reacti on cuvettes were added 6l of


38 propionaldehyde substrate (100l 500mM propi onaldehyde=3.6l of 0.5M substrate stock dissolved in 96.4l water), followed by mixing we ll with a 200l pipetman. The reduction of NAD+ to NAD measured the change in absorban ce at 340 nm over 5 minutes. A control reaction (blank), in which propionaldehyde was not added, monitored as the background endogenous rate of NAD+ reduction. To determine the protein concentrations of th e cell extracts, the BioR ad protein assay kit with bovine serum albumin as a protein sta ndard (BioRad, Hercules, CA) was used. Using plastic disposable cuvettes, 500l of BioRad prot ein assay buffer and 25l of lysis buffer and 1l of ptotein lysate was added to each cuvette. Positive samples turned blue. Samples were incubated for 10 minutes at room temperature protected from light by covering with foil. Following the incubation at room temperature, fixed 500nm wavelength OD readings of the samples were measured by the Beckman DLC 64 cuvette spectrophotometer. FACS Analysis and Sorting of ALDEFLUOR Stained Cells Flow cytometry has become a powerful techni que for the characterization of individual cells. As a cell is passing through a laser li ght beam, cellular morphology and expression of fluorescent labeled markers for each individual cell can be detected by this method. In most cases, flow cytometry is used for detection of fluorescent protein expressing cells and/or fluorescent conjugated antibody bound to cell surface proteins. The advanced machines can detect different non-overlapping fl uorescent dyes and fluorochromes which allows scientist to simultaneously use multiple markers for the multicolor analysis. The development of the ALDEFLUOR stai ning method made possible to detect cytoplasmic aldehyde dehydrogenase (ALDH) enzyme activity in viable cells by flow cytometry. As discussed in ALDH expression in Stem Cells section of the previous chapter, this method


39 was originally developed to enrich for human hematopoietic stem and progenitor cells (HSC/ HPC) based on their high ALDH levels (Sto rms et al. 1999). Using the ALDEFLUOR method primitive HSC were enriched from human umbilical cord blood (UCB) by lineage depletion (Lin-) followed by selection of cells with high aldehyde dehydrogenase (ALDH) activity. ALDHbrLincells constituted 22.6.0% of the Linpopulation, and highly co-expressed primitive HSC phenotypes (CD34+CD38and CD34+CD133+). Taking advantage of this system, the laboratory of Dr. Moreb modifi ed this technique to enrich fo r the CICs of human lung cancer cell lines as described in the next section. FACS Analysis ALDEFLUOR fluorescence was detected using the green fluorescence channel (Ex 488nm, Em Filter = 530/30 of the FACS caliber). The cultured cells were approximately 99% viable; therefore we did not need to use a viab ility dye to exclude death cells. However, cells recovered from tumor samples had many dead cell s. Therefore, Via Probe (BD Bioscience) was employed to exclude the dead cells, which retain the Via-Probe reagent. Via-Probe secreted fluorescence was detected usi ng the red fluorescence channe l (Ex 488nm, Em Filter = 670LP). Data for 10,000-20,000 cells was collected and anal yzed using Cell Quest software, version 3.3 (BD Biosciences). Cell Sorting Similar to FACS analysis, cell sorting process detect fluorescent labeled cells passing through a single-file stream of sheath fluid that ar e excited by a laser. As the cells travel through the excitation laser, reflections of the light beams are detected by different emission filters. The targeted cells are gated accord ing to phenotypic characteristic using an electric field which allows cell separation based on change and eventually distribution into collec tion tubes. In this study, the ALDEFLUOR-based sorting of the H522 cell line was performed by the laboratory of


40 Dr. Moreb. Sorting of the MIA PaCa-2 cell line was executed in the laboratory of Dr. Edward Scott. When the ALDEFLUOR staining based sorting was performed, in order to collect the brightest (ALDHbr) cells, which express the highest leve ls of ALDH, the gate was set for collecting only the 2% of the brightest. In the same concept, only the 2% of the dimmest cells were gated to collect ALDHlo cells. (It is very important to be cautious about keeping the samples in a cooler during the sorting because at room temperature, ALDH enzymatic activity can increase the intensity of the fluorescence). All cell sorting was carried out by the flow cytometry core members using the FACS Vantage SE (Becton Dickinson) instrument and Cell Quest software (BD Biosciences) at the University of Florida, ICBR Flow Cytometry facility located in the Academic Research Building. Non-Obese Diabetic/Severe Combined Immuno deficient (NOD/SCID) Xenotransplant The NOD/SCID tumor xenograft mouse model ha s been used for many years in cancer research. The mouse strain was generated by mating the SCID mutati on onto the non-obese diabetic background. Homozygous animals have impaired T and B lymphocyte development (Ref: Charles River Laboratories-strain: 394). Recently, the function of cancer initiating cell has been monitored by an in vivo se lf-renewal assay (Cla rke 2005). This assay is based on observing the tumor growth ability of fractionated cance r cell subsets following transplantation into immunodeficient mouse models. Animals Six to eight weeks old NOD/SCID mice were obtained from Charles River Laboratories (Wilmington, Massachusetts). Animals were used and maintained under specific pathogen-free conditions. All food, water and litte r were sterilized prior to us e. Temperature (20-21C) and


41 humidity (50-60%) were controlled. Daily light cycles were 12 hours lig ht and 12 hours dark. Cages were changed once or twic e a week. Animals were manipul ated under sterile conditions according to defined conditions approved by the Institutional Animal Care and Use Committee of the University of Florida. Preparation of Tumor Single Cell Suspension Excised tumors were suspended in steril e serum-free 1XPBS on ice and mechanically dissociated using scissors, followed by further mi ncing with a sterile sc alpel to yield 2 mm pieces. The tumor pieces were transferred to a conical tube and washed once with PBS. The washed tumor pieces were transf erred to a scintillation vial tube, followed by the addition of 2l of digesting enzyme solution to each sample. Th e samples were digested at 37C for 60 minutes, and mixed every 20 minutes with a P1000 pipette (t he end of the pipette tip was cut to prevent clogging during the mixing step). At the end of th e 37C incubation step, 1 l of cold PBS was added to the digested slurry, and the mixture wa s filtered through a 70 micrometer cell filter into a 50 l conical tube on ice to recover a single-ce ll suspension of tumor cells. During this step, it is important to keep the tissue samples on ice. The recovered cell suspension was washed with 5 l of cold 1X PBS. The supern atant was decanted and the cell pe llet was resuspenced in 1X PBS with 10% FBS, and a cell count was performed using a hematocytometer. Irraditation of Mice Although the NOD/SCID mouse exhi bits multiple defects in i nnate immunity, it still displays low NK cell activity that might cause reje ction of the injected cancer cells. Therefore, prior to cell injections, animals were sublet hally irradiated with 200 centiGray of gamma irradiation. Initially, we used 350 centiGray irradiation as de scribed by John Dick group (19).


42 However, all the irradiated animals died within two weeks. All subsequent irradiations for NOD/SCID mice were with 200centiGray. Transplanting Cells into NOD/SCID Mice To calculate the cell concentra tion for injection doses, viabil ity and number of cells were determined by mixing 50l of cells with 50l of 0.4% trypan blue stain (Gibco Invitrogen) and counting the bright cells using a hemacytome ter under an inverted microscope. After determining cell density, the cells were suspe nded in 100l 1x PBS per injection dose. One hundred l of cells were mixed well with 0.4 l of Matrigel HC and the mixtures were drawn into 1-l syringes. For each sorted cell fracti on, two to three 6-8 week-old NOD/SCID mice were anesthetized with 1.5% concentration of isoflurane (Baxte r; Deerfield, Illinois) using an 18gauge needle. The cell:Matrigel HC was injected under the skin near the scapula. Animals were observed daily to monitor tumor growth and ev aluated for overall clin ical condition including weight loss and indications of pain, distress, and abnormal behavior and physiology. Tumor growth at the sites of injecti on was measured weekly with calip ers as described below. As a prophylactic treatment, animals were given a re giment of antibiotics (2.5 l of 100mg/l antibiotic solution in 500 l water bottle) incl uding enrofloxacin (Bayer HealthCare; Shawnee Mission, Kansas), for two weeks. Tumor Evaluation Caliper measurements of the longest (L) and shortest (W) tumor diameters (mm) were obtained weekly and recorded on the animal proce dure sheet. The formula for an ellipsoid sphere (0.52 x L x W2) was used to calculate the tumor volume. The volume was converted to tumor weight assuming unit density (i.e., 1 mm3 = 1 mg). For humane reasons, animals were sacrificed when the implanted tumor volume became >4,000 mm3 (25. Ho MM, Ng AV, Lam et al. 2007)


43 Immunohystochemistry Immunohistochemistry was used to characte rize the ALDH expression pattern of tumor samples analyzed during this project. In or der to perform reproducib le accurate staining, optimization of the antibody concentrations a nd retrieval techniques was performed. More importantly, the specificity of the antibodies was verified by using both positive and negative controls. The immunohistochemistry protocols that we used for this study are described in this section. Sectioning and Preparation of Paraffin Embedded Tissues (Note: During this procedure the slides are never allowed to dry. If the tissues on the slides dry, there is an increased potential for nonspecific binding of antibody that may increase background). For our immunohystochemistry st aining, the concentration of ALDH1A1 and ALDH3A1 antibodies were not know n. Therefore, we had to first determine the optimal dilution of our antibodies by titration. We chose to us e normal human lung tissue because ALDH1A1 and ALDH3A1 are known to be expresse d in human alveolar epitheli al cells. We compared 1:100, 1:200, 1:300, and 1:400 dilutions of each primary antibody with 1:200 and 1:400 dilutions of the secondary antibodies. The expected staining patterns of the sequent ially sectioned slides allowed the optimal conditions to be defined. The optimal concentrations are as follows: 1:200 for the ALDH1A1 antibody and 1:300 for the ALDH3A1 an tibody. A 1:200 dilution of secondary goatanti chicken IgG antibody resulted in background staining on the chicken Ig Y control slide, which necessitated a 1:400 dilution of the secondary antibody concentration being used as the optimal concentration. Tumors excised from the animals were im mediately placed in 4% PFA and underwent overnight immersion fixed at 4oC. Samples were then submitted to the Department of Pathology Molecular Pathology Core Lab for processing an d embedding. Paraffin embedded blocks were


44 sectioned using a Microm microtom e (Heidelberg) at a thickness of 5 microns and transferred to plus charged microscope slides (Fisher Scientific). Slides were left to air dry overnight at room temperature. The following day, the samples underwent deparaffinization and rehydration as follows: Xylene 2X for 5 minutes 100 % Ethanol 2X for 2 minutes 95 % Ethanol for 2 minutes 3% H2O2 in Methanol for 10 minutes (To quench endogenous peroxidase activity) 70% Ethanol for 1 minute H2O twice for 1 minute After the deparaffinization and re hydration steps, the slides we re transferred to a plastic coplin jar containing citrate buffe r (0.1M Citrate Buffer, pH 6.0). The coplin jar with slides was transferred to a beaker and filled with water. The beaker containing the coplin jar setup was transferred to a GE microwave oven and set to 50% power for seven minutes. Once seven minutes had passed, the slides were kept in th e microwave for an additional 18 minutes, for a total of 25 minutes. Following the antigen retr ieval of ALDH1A1 and ALDH3A1, slides were washed with Tris Buffered Saline contai ning Tween 20 at pH 7.6 buffer (TBS-T, DAKO Cytomation) to equilibrate the slides. The exce ss buffer was removed with a gauze pad (Fisher Scientific), and the slides we re blocked for 20 minutes at r oom temperature with 1.5% goat serum (Vector Laboratories) diluted in TBS-T containing avidin (4 drops/l; Avidin/Biotin Blocking kit, Vector Laboratories). This step decreases the potential for nonspecific binding by the secondary antibody. Slides were then rinsed with TBS-T and the excess buffer was removed with a gauze pad. Primary antibodies of chicke n anti-ALDH1A1 IgY (1:200) and chicken antiALDH3A1 IgY (1:300) were diluted in Zymed diluent containing 4 dr ops/l biotin block (Vector Labs). The diluted anti human ALDH antibodies were incubated for 2 hours at room temperature, followed by a 5 minute washing st ep with 1X TBS-T. The secondary antibody,


45 biotinylated goat anti-chicken IgG (1:400) (V ector Laboratories) wa s incubated at room temperature for 30 minutes followed by a 5 minut e washing step. Detection of the secondary antibody was performed with the ABC Elite Standard kit (Vector Laboratories) and subsequently visualized with the substrate, Diaminobenzid ene (DAB) (Vector Laboratories). Tissue samples were lightly counterstained by dipping the slides 15-20 times in Gill's 2 hematoxylin (RichardAllan Scientific). Following the counterstaining step, the slides were sequentially processed through the following solutions to dehydrate the tissue sections: H2O 2X for 30 seconds 70% Ethanol for 30 seconds 95 % Ethanol for 30 seconds 100 % Ethanol 2X for 30 seconds Xylene 2X for 1 minutes Following the dehydration step in ethanol and xy lene, the slides were mounted in Xylseal (Richard-Allan Scientific). Representative tissue sections were stained with a hematoxylin-eosin (H&E) to access the morphology of the cells within the tumors. Sectioning and Preparation of OCT Blocks Freshly isolated tumor samples were rinsed with 1X PBS and positioned in a standard freezing cassette containing O.C.T. Freezing co mpound (Tissue-Tek) and transferred to a dry ice: 2-methylbutane bath (Sigme-Aldrich). During th is step it is important to prevent bubbling of the OCT medium by positioning the cassettes in the bath to allow no direct contact between the OCT medium and the methyl butane. Once the tissue OCT blocks were frozen, samples were removed, wrapped in aluminum foil, and placed in a -80 C freezer. The optimal concentration of the anti -human CD133 mouse monoclonal antibody was determined using fresh frozen normal human br ain tissue. The optimal staining pattern was observed at a concentration of 5 l/l for the primary antibody and 1:200 dilution for the


46 secondary antibody. Before sectioning, a new cryos tat blade was installed and the OCT tissue blocks were acclimated to -20 C within cryostat apparatus for at least 15 minutes. Acclimation of OCT block is very important fo r consistency of the sectioning of tissues. Serial sections were cut (3 sections per slide) at 5 microns. Suffi cient sections were generated to check for CD133 expression, monitor background stai ning with the IgG negative control, and H&E staining to access morphology. After the sectioning, slides were dried at room temperature for 10 minutes and fixed in 20 C acetone for 5 minutes. The slides were rem oved and allowed to air dry for 20 minutes, transferred to a coplin ja r containing acetone at -20 C for 5 minutes, and washed with 1X Tris buffered saline (TBS) at room temperature for 3 minutes. The 1X TBS maintains integrity of the tissue better than 1X TBS-T buffer. After th e 3 minute equilibration step, excess buffer was blotted away with a gauze pad. During this procedure, the slides were placed in an opaque humidified chamber during the incubation ste p, which protects the fluorescent-conjugated secondary antibody from photo bleaching. The slid es were blocked for 20 minutes at room temperature with (15 g/l diluted in 1X TBS) horse serum (Vector Labora tories), washed in 1X TBS for 5 minutes and blo tted with a gauze pad. Following the serum blocking, excess serum was blotted from each slide using a gauze pad, and slides were in cubated overnight at 4 C with 5 g/l (1:10) mouse-anti human CD133 diluted in Zymed dilutent (Mab) (Miltenyi Biotsrhech clone: Clone AC133) or 5 g/ l (1:20) mouse IgG1 Mab (Vector Laboratories). For overn ight incubations, the slides were overlayed with a plastic slide cover to prevent evaporati on. The following day, the slides were washed 3X for 5 minutes with 1X TBS at room temperature. Excess buffer was blot ted away with a gauze pad. Slides were then incubated with a 5 g/ l (1:400) of Donkey anti-Mouse AF 488 (green)


47 secondary antibody for 60 minutes at room temp erature Following the incubation step, slides were washed 3X for 3 minutes each using TBS at room temperature. Excess buffer was removed and one drop of Vectashield with DAPI mounti ng media (Vector Labora tories) was added to each tissue section, followed by gentle placement of a glass coverslip to prevent the formation of air bubbles. The processed slides were stored in a slide folder (Fischer Sc ientific) and stored at 4 C until visualized by fluorescence microscopy. Th e optimal fluorescent signal is stable for approximately two weeks when properly stored at 4 C in the dark. CD133 Staining and FACS Analysis of Cells The H522 lung cancer and MIA PaCa-2 pancreatic cancer cell lines were further analyzed for their CD133 expression by flow cytometry. Cu ltured cells were harv ested in a tube as described previously, and counted with a hemacytometer. Following the cell counting, 1X107cells were resuspended and incubated with 5 g (10 l of 0.5 g/ l) of Phycoerythrin (PE) conjugated mouse anti-human CD133 (M ilteny Biotec, Clone AC133) in 100 l of staining buffer (see reagents section). The antibody-cell mixture was incubated at 4 C for 20 minutes in the dark. The samples were then centrifuged to pellet the cells at 1200rpm at 4oC for 5 minutes and then resuspended in 1 l of FACS running buffer (see reagents section). The FACS analysis of the samples was performed using a FACS Calibur (BD Biosciences, San Jose, CA). Phycoerythrin (PE) fluorescence labeled cells were detected using the red fluorescence channel (Ex 488nm, Em Filter = 585/42). Cell Cycle Analysis Cells were grown to 70% confluence in a T-75 (Fisher Scientific) flask and removed from the incubator (37C, 5% CO2) 1 days after plating. The media from the flask was transferred to a 50 l falcon tube (Fisher Scientific) l eaving the pipette in the tube, 2.5 l of


48 0.25% Trypsin-EDTA (Gibco, Invitrogen) was added to each flask and incubated at 37oC for 5 minutes. Detached cells were transferred to 50 l tube and centrifuged for 6 minutes at 1200rpm. The supernatant decanted and cells were resuspend in 3 l of 1X PBS. The number of viable cells was counted using a hemacytometer with the viability dye, trypan blue. For cell cycle analysis by flow cytometry, 5x105 cells were suspended in 36 0l of solution A containing trypsin, gently mixed and incubated at r oom temperature for 10 minutes. Following the incubation, 300l of solution B containing tryps in inhibitor and RNAse, was added, mixed gently and incubated at room temperature for 10 minutes. To stain the genomic DNA content of the cell the addition, gently mixing and incubation of s, 300l of solution C containing PI and spermine tetrahydrochloride was added, mixed gen tly and incubated at room temperature for 10 minutes. The samples were placed in ice until the FACS analysis step (samples must be analyzed within three hours). DNA content wa s analyzed using a Becton Dickinson FACScan flow cytometer (Becton Dickinson, San Jose, CA). The distribution of cells in the different phases of the cell cycle was anal yzed from DNA histograms using the ModFit cell cycle analysis software version 3.1 (Verity Inc, Topsham, ME). Cell cycle analysis was done on unsorted, ALDHbr and ALDHlo sorted H522 human lung cancer and the MIA PaCa-2 human pancreatic cancer cells. Reagents RPMI Complete Media Materials: RPMI-1640 medium with high glucos e (Gibco Invitrogen), fetal bovine serum (FBS, Gibco Invitrogen), Penicillin-S treptomycin (Gibco Invitrogen), 500 L Stericup disposable vacuum filtration system (0.22 m pore, Millipore).


49 Protocol : In sterile laminar flow hood, measured ou t 10 % less than desired total volume of media in a Millipore vacuum 0.22 m filter top, added 10% fetal bovine serum (FBS), and 1g/ l penicillin/ streptomycin antibiotic. Then, filtere d the media using vacuum and stored the media at 4C in the dark for up to 12 months. DMEM Complete Media Materials: Dulbeccos Modified Eagles Medium High Glucose, 1x (DMEM, Gibco Invitrogen), fetal bovine serum (FBS, Gibco Invitrogen), GIBCO Horse Serum (Gibco Invitrogen)Penicillin-Strepto mycin (Gibco Invitrogen), 500 L Stericup disposable vacuum filtration system (0.22 m pore, Millipore). Protocol: In sterile laminar flow hood, measur ed out 12.5 % less than desired total volume of media in a Millipore vacuum 0.22 m filter top, added 10% of FBS, 2.5% of horse serum (HS), and 1g/ l penicillin/ streptomycin antibiotic. Then, filtered the media using vacuum and stored the media at 4C in the dark for up to 12 months. ALDEFLUOR Kit Materials: ALDEFLUOR Kit contai ns: Dry ALDEFLUOR reagent, 50 g; Diethylaminobenzaldehyde (DEAB) 1.5 mM in 95% ethanol, 1 l; Hydrochloric Acid (HCl), 2N, 1.5 l; Dimethylsulphoxide (DMSO), 1.5 l; ALDEFLUOR Assay Buffer, 4 bottles of 25 l each; ALDEFLUOR Quick Reference Guide (StemCell Technologies). Protocol: All the necessary supplies were asse mbled and allowed kit reagents to come to room temperature (RT), 18C before use. To activate the ALDEFLUOR reagent: 25 l of DMSO was added to the vial of dry ALDEFLUOR reagent, mixed well, and let stand for 1 min at room temperature. Then, 25 l of 2N HCl was added, mixed, and incubated for 15 min at room temperature. Following the incubation, 360 l of ALDEFLUOR Assay Buffer was added


50 to the vial and mixed. Activated ALDEFLUOR subs trate was dispensed into aliquots and stored frozen at or below -20C. Matrigel High Concentration Materials: BD Matrigel Basement Membrane Matrix High Concentration (HC) Phenol Red Free (BD Bioscience). Protocol: One day before the injecti on, the Matrigel-HC was thawed at 4oC overnight. Tumor Enzyme Digestion Buffer Materials: 10% stock Collagena se A (100x) in dH2O (Roche), Dispase II (Roche), Stock CaCl2 (2.5M stock; 1000x). Protocol: In cold room, 10 l of Dispase II transferred into a 15 ml tube, added 100 l of 10% Collagenase A and 10 l of 2.5mM CaCl2, and aliquot of 2 l screw cap brown tubes storde at -20C in the dark for up to 12 months. 4% Paraformaldehyde (100 L) Materials: Paraformaldehyde (Sigma), 250 L glass beaker, 1N sodium hydroxide (NaOH), 1N hydrochloric acid (H Cl), 1X PBS, glass Pasteur pipettes, aluminum foil, 100 l glass bottle, filter paper, sma ll funnel, stir bar/magnetic stirrer hot plate, pH meter. Protocol: In chemical fume hood, 4 g (4%) paraformaldehyde mixed with 60 l of 1X PBS, and heated to approximately 55C in a gla ss beaker and covered with aluminum foil while mixing with a stir bar for 30 minutes. Af ter the solution became semi-clear, 30 l of 1X PBS added, and the pH was adjust to 7.2 using 1N HCl and 1N NaOH. Then, the final volume adjusted to 100 l with 1X PBS, filtered through Whatma n paper into clean glass bottle and stored at 4C up to 1 week.


51 Immunocytochemistry 0.1M Citrate Buffer Antigen Retrieval Solution (pH 6.0) Avidin-Biotin Blocking Kit (Avidin A and Avidin B solutions) Ready to Use (Vector Laboratories) Antibody Diluent Zymed (Invitrogen) DAB plus kit In 2.5 l water added 1 drop of the buffer, 2drops of DAB chromogen, and 1 drop of hydrogen peroxide, and mixed thoroughly (m ust be prepared freshly). (Dako) 3%Hydrogen Peroxide 45 l of methanol added 5 l of 30% Hydrogen peroxide stock (Sigma) Tris Buffered Saline with Tween 20 at pH 7.6 (TBS-T): 180 l of water added 20 l of 10X Tris-Buffered NaCl Solution with Tween 20, pH 7.6 (DAKO) Antibodies Chicken anti-human ALDH-1A1 and ALDH-3A1 polyclonal antibodies were provided by Dr. Jan Moreb who received these antibodies fr om Dr. L. Sreerama (St. Cloud University, Minneapolis, MN) and Dr. N.E. Sldek (Uni versity of Minnesota, Minneapolis, MN). Biotinylated goat anti-chicken IgG, avidin-bioti nylated horseradish peroxidase conjugate, an avidin/biotin blocking kit, a Vectastain ABC kit and a peroxidase substrate kit were purchased from Vector Laboratorie s (Burlingame, Calif.). Mouse anti human CD133PE conjugated antibody (Miltenyi Biotech. Clone CD133) (For flow cytometry analysis).


52 Mouse anti human CD 133unconjugated antibody (Milteny i Biotech. Clone CD133) (For IHC analysis). Flow Cytometry Staining Buffer MaterialsPhosphate buffered saline (D-PBS, Gibco), sodium azide (Sigma), bovine serum albumin (BSA, Sigma), 500 l Stericup disposable vacuum filtration system (0.22 m pore, Millipore), stirring plate and magnetic stir bar. Protocol On a stirring plate, mix 500 l of 1X D-PBS with 10 grams (2%) of BSA and 0.5 grams (0.1%) of sodium azide until dissolved. Sodium azide is extremely toxic, so take proper precautions. Filter the resulting solution through a 0.22 m filter cup, discard the filter cup and close the container with the provided steril e cap. The solution is stable for 6 months if stored at 4 C. Flow Cytometry Running Buffer Materials Phosphate buffered saline (1X DPBS, Gibco), sodium azide (Sigma), 500 l Stericup disposable vacuum filtration system (0.22 m pore, Millipore), stirring plate and magnetic stir bar. Protocol On a st irring plate, mix 500 l of 1X D-PBS with 0.5 grams (0.1%) of sodium azide until solution is clear. Sodium azide is extr emely toxic, so take proper precautions. Filter the resulting solution through a 0.22 m filter cup, discard the filter cup and close the container with the provided sterile cap. The solution is stable for 6 months if stored at 4 C. Vindelov Method Stock Solutions Materials Trisodium Citrate-2 H2O (Fisher Scientific), NP-40 detergent (Sigma), Spermine tetrahydrochloride (Sigma), Tris (Sigma ), Trypsin (Sigma), Trypsin inhibitor (Sigma), RNAse (Sigma), Propidium Iodide (Calbiochem).


53 Protocol Stock Solution In a 2000 l glass beaker, 1000 mg Trisodi um Citrate-2 H2O, 1000l NP40 detergent, 522 mg Spermine te trahydrochloride, and were a dded and the final volume was adjusted to 1000 l with a pH of 7.6. Solution AIn a 500 l glass beaker, 250 l Stock solution (pH 7.6) and 7.5 mg Trypsin was mixed and stored at -70oC in 2 l plastic vials. Solution B In a 500 l glass beaker, 250 l Stock solution (pH 7.6), 125 mg Trypsin inhibitor, and 25 mg RNAse mixed and stored at -70oC in 2 ml plastic vials. Solution C In a 500 l glass beaker, 250 l Stock solution (pH 7.6), 104 mg Propidium iodide, and 290 mg Spermine tetrahydro chloride mixed and stored at -70oC in 2 l plastic vials protecting from light within a aluminum foil.


54 CHAPTER 4 RESULTS AND DISCUSSION Chapter 4 is devoted to the resu lts and discussion of this st udy. In this study, the primary goal was to establish an in vivo system to evaluate the ALDEFLUOR-based CIC subpopulation enrichment procedure as applied to the of human non-small lung cancer cell line H522. Through the collaboration of laboratory of Dr. Edward Scott and Dr. Jan Moreb this study could be constituted. At the start of this study, the in vitro experiment characterizing a number of human lung cancer cell lines including H522 had been co mpleted by the laboratory of Dr. Moreb. The laboratory of Dr. Scott previously set up a ssays to monitor tumor vasculogenesis and angiogenesis using immunodefi cient mice. For this study, in vivo tumor methodology was modified to perform in vivo limiting dilution and se rial transplantation assays as explained in detail in Materials and Methods. More importantly, because this study was a collaborative effort, it was necessary to describe the results and observations from previ ous in vitro experiments that were performed by the laboratory of Dr. Moreb. Overall the main objective was to unde rstand the biological significance of ALDH class1A1 and class 3A1 (ALDH1A1 and ALDH3 A1, respectively) expression in lung cancer. These enzymes have been reported to be highly expressed in some lung cancer cell lines as well as in patient lung cancer samples (Moreb et al. 2007a). This st udy also examined whether these enzymes could be used as marker s for lung cancer initiating cells. Results In Vitro ALDH Expression of Hu man Lung Cancer Cell Line H522 As we mentioned previously, the in vitro studies, performed in the laboratory of Dr. Moreb, profiled the ALDH expression levels of the 12 different human lung cancer cell lines


55 including H522. According to the ALDEFLUOR st aining results, 71% of H522 cells were found to be ALDH positive. The H522 cell line had 82 (nmol/107 cells per min) of enzyme activity and contained 0.44nmo l ALDH1A1 and 0.6nmol ALDH3A1 according to western blot analysis (Table 4-1, Figure 4-1. Moreb et al 2007.). The ALDEFLUOR based sorting of the H522 cell line and expansion of the resulting AL DHbr and ALDHlo cells in culture were also executed by the Moreb lab. For the sorting regiment based on ALDEFLUOR, the highest (ALDHbr) cells, which express the highest levels of ALDH, the gates were set to collect the brightest 2% as the ALDHbr population and the dimmest 2% as ALDHlo cells. Sorting cells using this gating strategy resu lted in dramatically low yiel d from both the br ight and low populations, which necessitated that the cells undergo an in vitro e xpansion before use. For this study, ALDHbr and ALDHlo H522 cells from two diffe rent sorting experiments were used. The first set of sorted cells was used in transplant ation experiments using a cell dose of 105 cells. The resulting sorted ALDHbr cells had 168 (n mol/107 cells per min) enzyme activity, and ALDHlo cells had 66.58 (nmol/107 cells per min) enzyme activity (Moreb et al., unpublished data) (Figure 4-2). According to their observation, ALDHbr cells had a 3-4 fold slower doubling rate than the ALDHlo cells. The second set of sorted cells was used for the limiting dilution assay using 500 to 104 cells per dose. The second attempt at sorting the H522 cells resulted in very low yields, which required the sorted cell s to be cultured for a longer period of time. During the expansion of the cells from the s econd set, enzyme activity was measured and determined to be 57.88 (nmol/107 cells per mi n) for the ALDHbr cells, and 23.15 (nmol/107 cells per min) for the ALDHlo cells (Moreb et al., unpublished data) (Figure 4-1). The ALDHbr cells from the second set were gr owing at a 2 fold slower doubli ng rate than the ALDHlo cells. However, they did not notice any growth rate di fferences between the ALDHlo cells of the first


56 and the second sorting attempts These observations and enzyme activity assay results were consistent with the proposed model about a corr elation between the ALDH enzyme activity and the growth rate of cells as discussed in Effects of ALDH on Cell Proliferation and Differentiation section in the second chapter. At high levels of ALDH enzyme activity, the growth rate of the cells is reduced. ALDHbr Cells Give Rise to Both ALDHbr and ALDHlo Cells To compare the growth characte ristics of the fractionated H522 cells, the laboratory of Dr. Moreb cultured the ALDHbr and ALDHlo cell populations under th e same culture condition. ALDEFLUOR FACS analysis and immunochemical an alysis of the cultured cells demonstrated that the ALDHbr cells were capable of generating both ALDHbr and ALDHlo cells, whereas ALDHlo cells maintained their low ALDH activity levels (data not shown). ALDH Levels Altered the Tumor Growth Rate It has been known that the human lung cancer cell line H522 can generate tumors in vivo (Desai et al. 2006). To investigat e possible differences in tumor formation potential between the ALDHbr and ALDHlo sorted cells, varying ce ll doses were injected under the skin of NOD/SCID mice and monitored for tumor developmen t. As shown in Table 4-2, both cell groups yielded tumors (two to three animals for each gro up) at every cell doses tested and ranged from 500 to 1x105 cells. Representative tumors along with morphologi cal analysis by HematoxylinEosin (HE) staining for ALDHbr and ALDHlo cells showed no signi ficant differences (Figures 4-3, 4-4). However, tumor growth rates were dramatically altered based on the ALDH expression levels of the transplant ed cells. As expected from the re sults of in vitro experiments, the ALDHlo cells rapidly formed tumors, whereas the tumor growth rate of ALDHbr cells were significantly slower. After 5 weeks post-transplantation ALDHlo tumors from 1x105 cell dose grew to a volume of 3904*Standard Dev. mm3 size, which was 33 fold larger than the


57 ALDHbr tumor at the same time point (Table 42). Interestingly, an a dditional three weeks was required for the ALDHbr cell-generated tumors to reach an equivalent size (3493 mm3). Furthermore, as mentioned above, for the limiting dilution in vivo assay, the ALDEFLUOR sorting of the H522 cell line was performed on two spare occasions. The cells from the first cell set acquired higher ALDH activity than the se cond time sorted ALDHbr and ALDHlo cells (Figure 4-2). Although, the first set of sorted ce ll used for the 1x105 cell do se transplantation, the tumors formed from the second sorting 1x104 cells grew faster and sized larger than the 1x105 dose. Strikingly, the second time sorted ALDHb r and ALDHlo cells, which had significantly lower ALDH activity compared to ALDHbr and ALDHl o cells of the first set, displayed a more tumorigenic characteristic (see Figure 4-2 for enzyme activity levels and Table 4-2 for tumor sizes). Indeed, this result further confirm that the higher ALDH levels delay the tumor progression as explained in mo re details in discussion. ALDH Enzyme Activity Decreased in vivo To detect the ALDH enzymatic activity changes that occur in vivo during the growth of the tumor, the tumor samples were assayed for ALDH activity using both the ALDEFLUOR flow cytometry and ALDH enzyme activity assays. Unexpectedly, the FACS analysis of ALDEFLUOR stained ALDHbr tumor cells displayed no shift of brightly stained cells compared to the DEAB inhibited control cells (Figure 4-5). These results coul d be due to either insufficient inhibition of enzyme activity in negative contro l, or the cells may have reduced ALDH activity. The enzyme activity assay further confirmed the reduction in the enzymatic activity of cells. The tumor cells from ALDHbr cells injected animals had three fold reduction in enzymatic activity compared to the initial sorted cell population (Figure 4-6). However, ALDHlo cells also had reduction in ALDH activity.


58 ALDHbr Primitive CICs Were More Tumorigenic Interestingly, the serial transplantation of ALDHbr and ALDHlo tumor cells demonstrated that ALDHbr cells were more tumorigenic th an the progenitor ALDHlo cells. Although, ALDH levels of both cell fractions were decreased in vivo, ALDHbr cells were ab le to give rise to serially transplantable tumors with higher tumor growth rate, whereas ALDHlo tumor growth rates were subsequently decrease d with each serial transplanta tions (Table 4-3). Assuming that high ALDH expressing primitive CIC population produ ced both more CICs and progenitor cells that resulted in the tumor volume of 3373 mm3 following the s econdary transplant, and volume of 2392* mm3 of the third transplant (Table 43). In this scenario, ALDHlo cells primarily composed of differentiated progenitors with a fini te proliferative potential (Figures 4-7,4-8,4-9). ALDH Expression Pattern of CICs Within Tumor Bulk To observe the in vivo distribution of CICs those highly express ALDH isomers: ALDH1A1 and ALDH3A1, tumor sections were stained with anti human ALDH isomers specific antibodies (see material s and methods). The immunohysto chemistry (IHC) staining of tumor sections demonstrated presence of ALDH positive cells that were localized in a close proximity to the small blood vessels. A recent study has revealed that small blood vessels associated with brain tumors orchestrate a di stinct microenvironment that is critical for maintaining CICs (Calabrese et al. 2007). The in vestigators examined wh ether the perivascular niche secreted factors promote maintenance of CICs and generation of tumors in vivo. To compare the proposed effects of vascular endothelial niche on the CIC population, the researchers transplanted human br ain tumors into mice with or without human vascular cells to initiate the formation of niches for the CIC population. The mice transplanted with the human vascular endothelial cells presented an increase in CICs as well as enhanced initiation and


59 proliferation of tumors. In addition, antiangiogeni c treatments that disrupt the CIC niche reduced the number of CICs and arrest tumor growth. However, this phenomenon was not observed for the non-CIC tumor cells. Although, the ALDEFLUOR based flow cytometric analysis and enzyme activity assays suggested an overall reduction in ALDH e xpression in tumor cells, IHC staining using ALDH1A1 and ALDH3A1 exhibited the presence of clustered posit ively stained cells appeared around the vessels. Similar to brain CICs, H522 lung CICs were preferably located close to the vascular niche, which further supporting the ex istence of the ALDH expressing lung CICs. More importantly, both ALDHbr and ALDHlo tumors contained ALDH1A1 and ALDH3A1 positive cells that displayed a similar staining pattern as an indicato r of CIC markers (Figure 4-10). Staining of H522 Cells for Other CIC Marker CD133 To elucidate on the possible association of CD133 to ALDEFLUOR-based enriched lung CICs, the ALDHbr cells additionally stained by CD133. Ho wever, in three different attempts, flow cytometric and immunochemical analysis exhibited no CD133 positive staining of the ALDHbr and ALDHlo cells. The laboratory of Dr. Moreb has also further verified by FACS analysis that H522 ce lls are CD133 negative. MIA PaCa-2 Human Pancreatic Cancer Cell Line ALDH Expression To test whether the ALDEFLUOR based sorting can be applied for the enrichment of CICs from other solid tumors, MIA PaCa-2 cell line wa s utilized. ALDEFLUOR st aining of pancreatic cancer cells demonstrated that 27% of th e cells were ALDH positive. To evaluate the ALDEFLUOR sorting efficiency, the fractiona ted cell populations were mixed 2:10 with immortalized lung epithelial ce ll line (Beas-2B), which is know n to lack ALDH activity and analyzed by flow cytometry. The FACS an alysis revealed that 16% of the ALDHbr (20%) BEAS2B (80%) mixed cells, 3.4% of the ALDHlo (20%) cell mixture, and 8% of the unfractionated


60 MIA PaCa-2 cell (20%) mixture were ALDEFL UOR positive (Figure 4-11). The ALDH enzyme activity assay results also confirmed that ALDE FLUOR based sorting appr opriately fractionated pancreatic cancer cells according to their ALDH le vels (Figure 4-12). For the in vitro evaluation of the MIA PaCa-2 cell line, unsor ted cells were compared to ALDHbr and ALDHlo populations. The short and long term cu ltural behaviors of these cells were similar to the in vitro growth pattern of the H522 l ung cancer cell line. ALDHbr cells had a five fold slower doubling rate than the cells with the low ALDH activity. However, the Vindelov method of DNA staining for the cell cycle analysis of the cells exploited no signi ficant difference in G2 and S phase proliferating cell percentages between the ALDHbr (G2: 6.23%, S: 7.04% ) and ALDHlo (G2: 9.71%, S:6.92%) cell populations (Figure 4-13). However, the unfractionated panc reatic cancer cells appeared to be the most actively prol iferating group (G2: 4.71%, S: 18.78%). Discussion Evidence suggests that CICs exist in the H 522 human lung cancer cell line, which can be enriched by ALDEFLUOR based so rting. According to CIC theo ry, only a subpopulation of a heterogeneous tumor cells is capable of regene rating tumors. This CIC population is defined by their stemness properties, more specifically self-renewal and asym metrical cell division. Stemming from the well established hematopoietic stem cell (HSC) system, in this study, CICs of the H522 human lung cancer cell line was examin ed for a prospective correlation between the CICs and the ALDH activity. Storm et al. were able to isolate primitive HSCs utilizing ALDEFLUOR based sorting of human umbilical co rd blood cells (Storm et al. 1999). Following, acute myeloid leukemias ALDH fractionation, Dr. Ja n Moreb has been applied to this sorting method to human lung cancer cell lines (Cheung et al., Moreb et al. 2007). However, the in vivo tumor generating ability of the ALDH positive lung cancer cell lines were yet to be investigated. This study provided that as low as 500 ALDH expressing H522 cells were able to give rise to


61 tumors. More importantly, the ALDH levels of th e cells were found to be a determinant of the tumor propagation. Although, both ALDHbr and ALDHlo sorted cells could develop tumors, in vitro and in vivo the doubling rate of tumor cells were inversely correlated to the ALDH levels of the cells; as the ALDH levels decrease, tumors grow faster. Currently, the CICs are considered to be the tumorigenic cells that initiate tu mor when transplanted into immunodeficient mouse model, yet the tumor growth rate has not been c onsidered as a valuable parameter to measure. The trend for the characterization of CIC is ev aluating the minimal cell dose and serially transplantable cells to initiate tumor. Indeed, if we were to narrow down the criteria to define CICs based on the lowest number of cells and se rially transplanting capacity; the in vivo study results may appear suggesting that the ALDHlo population carried the CICs due to its fast tumorigenicity and serially engraftment ability. Du ring the first set of experiments, (see results section for details) by the time the ALDHlo tumors grew to their maximum allowed size, the ALDHbr tumors were almost undetectable. If we were to terminate the ALDHbr animal study at the same time with the maximized ALDHlo tumors, we would not be able to observe the very slow growing ALDHbr tumors. However, this study results may indicate a furt her fractionation of CICs into the long term cells for ALDHbr and cells for the short term ALDHlo CICs. This phenomenon can be explained using the HSC syst em, in which the stem cells were further fractionated into long-term (LTHSC) and short-term (ST-HSC)/ premature progenitors that acquire limited proliferative capac ity (Kiel et al. 2005). Although the LT-HSCs are considered to be the most primitive HSCs that can regenerate a whole hematopoietic system from a single cell transplantation, the reconstitution progress takes longer period of time due to the steps in generating first progenitor cells and then the maturation period of the progenitor cells (Kiel et al. 2005). In this study what we obser ved in tumor progression of ALDHbr tumors might be similar


62 coincidence to the LT-HSC engraftment. The small size, high ALDH expressing (SSCloALDHbr) human umbilical cord blood cells were also displayed the most primitive subpopulation (Storm et al. 1999). As proposed the ra nking of ALDH levels in the he matopoietic system cells in a hierarchal order, the highest ALDH expressing cells are the most primitive HSCs, which also could be the case in this study as the highest AL DH levels indicating the primitive CICs that can self-renew and generate progenitor CICs; thereby developing tumors slower than the ALDHlo tumors. The serially transpla ntation in vivo assays performed in this study also revealed that serially tumor generating ability of AL DHlo tumors was reduced by each subsequent engraftment, whereas tumor progressi on of ALDHbr cells was increasing. In addition to evidence the stemness prope rties of the high ALDH expressing CICs, IHC staining for ALDH1A1 and ALDH3A1 isomers displayed a localization of ALDH positive cells in a close proximity to the perivascular niches, wh ich are proposed to be the CICs niche for brain tumor initiating cells (Calabrese et al. 2007). For primitive cells, niche is a home from where they receive the necessary signaling to maintain their stemness. If this suggested endothelial vascular cells are also niche for the H522 CICs, our IHC analysis results perfectly fit in this scenario. However, ALDH enzymes functions in the detoxification of the cells and their expression can also be regulat ed by xenobiotic factors (Slde k et al. 2002). The high ALDH expression of the perivasvular cells may also i ndicate the up regulation of these isomers due to external stimulants diffusing from the blood vess els. Therefore, the underlying mechanism and significance of this localization incidence must be investigat ed before stating a certain explanation. Overall results of this study indicated that the ALDEFLUO R based sorting of H522 lung cancer and MIA PaCa-2 pancreatic cancer cell lines has the potential to enrich for the CICs.


63 This study also suggested a possible role for ALDH1A1 and ALDH3A1 proteins in the maintenance of CICs. A proposed mechanism re gulating cell proliferation and differentiation may involve the retinoic acid receptor si gnaling pathway (Alvarez et al. 2000) Further research by using clini cal primary human tumor materials are needed to investigate this important new mechanism associated with ALDH. It seems likely that effective cancer treatments must target both the bulk of rapi dly proliferating tumor cells and the smaller population of self-renewing CICs. Table 4-1. Measurement of ALDH activity and prot eins in human lung cancer cell lines (Moreb et al. 2007) Table 4-2. Limiting Dilution Assay Tumor Growth Rates: The dose groups and the weekly tumor volume measurements [length x (width) 2 x 0.52] (mm3). (S tandard D eviation ) ( : First set of sorted cells that had higher ALDH levels) Dose Group Inj Day 1st wk 2nd wk 3rd wk 4th wk 5th wk 6th wk 7th wk 8th wk ALDHbr 105 52 0 107 93 63 117 339 2268 3493 310 ALDHlo 105 52 0 0 140 1081 3904 360 ALDHbr 104 52 0 95 148 438 1423 4 ALDHlo 104 52 0 107 377 1194 4235 384 ALDHbr 103 52 0 0 38 68 1133 3 ALDHlo 103 52 0 0 107 528 2984 3 ALDHbr 500 52 0 0 45 68 371 0 ALDHlo 500 52 0 0 59 363 1524 190


64 Table 4-3. Serial transplantation assay tumor growth rates: The dose groups and the weekly tumor volume measurements [length x (width) 2 x 0.52] (mm3). *(Standard Deviation) Inj Day 1st wk 2nd wk 3rd wk 4th wk 5th wk 6th wk 7th wk 8th wk 1st br 10e5 (n=3) 52 0 107 93 63 117 339 2268 3493 1st lo 10e5 (n=2) 52 0 0 140 1081 3904 2nd br 10e6 (n=2) 52 47 326 1423 3373 2ndlo 10e6 (n=1) 52 68 175 568 1582 3rd br 10e6 (n=2) 52 68 123 435 2392717 4405 3rd lo 10e6 (n=2) 52 31 60 211 762309 2512 Figure 4-1. Western blot analysis of ALDH1A1 and ALDH3A1 proteins expressions in human lung cancer cell lines (M oreb et al. 2007). 0 50 100 150 200 Activity in (nmol/107 cells. min) ALDH Enzyme Activity Assay 1st Set 16866.58 2nd Set 57.8823.15 ALDHbr H522ALDHlo H522 Figure 4-2. ALDH enzyme activity assay of first and second sets of sorted H522 cells.


65 Figure 4-3. Representative subcutaneous tumors [due to injection of (1x105) ALDHbr (right eight weeks post transplantation; tu mor volume= 3594mm3) and (1x105) ALDHlo (leftfive-weeks post transplantation; tumor volume= 4159mm3) H522 cells. ] ALDHbr Tumor (20X) Figure 4-4. H&E staining.


66 ALDHlo Tumor (20X) Figure 4-4. Continued


67 Figure 4-5. The FACS analysis of ALDEFLUOR stained tumor cells. ALDH Enzyme Activity Assay 119.36 168 40.15 66.58 44 0 50 100 150 200 Unsorted H522 ALDHbr H522 ALDHbr Tumor ALDHlo H522 ALDHlo TumorActivity in (nmol/107 cells.min)Figure 4-6. The comparison of ALDH enzyme activ ity of injected cells and the tumor cells.


68 First Transplanted Animals Tumor Growth 0 500 1000 1500 2000 2500 3000 3500 4000 4500 Inj Day 1st wk 2nd wk 3rd wk 4th wk 5th wk 6th wk 7th wk 8th wkTumor size in mm3 1st ALDHbr 10e5 1st ALDHlo 10e5 Figure 4-7. Tumor growth chart of first transplanted animals. Second Serially Transplanted Animals' Tumor Growth 0 1000 2000 3000 4000 Inj Day1st wk2nd wk3rd wk4th wkTumor size in mm3 2nd ALDHbr 10e6 2 n dA LDH lo 10e 6 Figure 4-8. Tumor growth chart of sec ond serially transplanted animals.


69 Third Serially Transplanted Animals' Tumor Growth 0 1000 2000 3000 4000 5000 6000 Inj Day1st wk2nd wk3rd wk4th wk5th wkTumor Size in mm3 3rd ALDHbr 10e6 3rd ALDHlo 10e6 Figure 4-9. Tumor growth chart of thir d serially transplanted animals. ALDH1A1 (40X) A Figure 4-10. ALDH1A1 and AL DH3A1 staining of tumor samples A) ALDH1A1 (40X); B) IgG Control (20X); C) ALDH3 A1 (40X); D) ALDH3A1 (10X)


70 IgG Control (20X) B ALDH3A1 (40X) C Figure 4-10. Continued


71 ALDH3A1 (10X) D Figure 4-10. Continued


72 Figure 4-11. ALDEFLUOR based FACS analysis of the MIA PaCa-2 (20%) and BEAS-2B (80%) cells.


73 Figure 4-13. Cell cycle analysis of fractionated MIA PaCa-2 cells.


74 Figure 4-13. Continued


75 LIST OF REFERENCES Al-Hajj M, Wicha MS, Benito-Hernandez A, Morrison SJ, Clarke MF. 2003. Prospective identification of tumorigenic breast can cer cells. Proc Natl Acad Sci USA 100:39833988. Baylin SB Herman JG Graff JR Vertino PM Issa JP. 1998. Alterations in DNA methylation: A fundamental aspect of neoplasia. Adva nces in Cancer Research 72:141-196. Bonnet D, Dick JE. 1997. Human acute myeloid leukemia is organized as a hierarchy that originates from a primitive hematopoietic cell. Nat Med 3:730-737. Buzzeo MP, Scott EW, Cogle CR. 2007. The hunt fo r cancer-initiating cells : A history stemming from leukemia. Leukemia 21(8):1619-1627. Calabrese C, Poppleton H, Kocak M, Hogg TL, Fuller C, Hamner B, Oh EY, Gaber MW, Finklestein D, Allen M, Frank A, Bayaz itov IT, Zakharenko SS, Gajjar A, Davidoff A, Gilbertson RJ. 2007. A perivascular niche fo r brain tumor stem cells. Cancer Cell 11(1):69-82. Carney G, Wei S, Rizzo WB. 2004. Sjgren-Lar sson syndrome: Seven novel mutations in the fatty aldehyde dehydrogenase ge ne ALDH3A2. Hum Mutat 24(2):186. Chute JP, Muramoto GG, Whitesides J, Colv in M, Safi R, Chao NJ, McDonnell DP. 2006. Inhibition of aldehyde dehydrogenase and re tinoid signaling induces the expansion of human hematopoietic stem cells. Pr oc Natl Acad Sci USA 103:11707-11712. Clarke MF, Dick JE, Dirks PB, Eaves CJ, Ja mieson C HM., Jones DL, Visvader J, Weissman IL, Wahl GM. 2006. Cancer stem cellsPersp ectives on current st atus and future directions: AACR Workshop on Cancer Stem Cells. Cancer Res. 66:(19). Clarke MF. 2005. A self-renewal assay for cance r stem cells. Cancer Chemother Pharmacol 56(Suppl. 1):64-68. Cogle CR, Guthrie SM, Sanders RC, Allen WL, Scott EW, Petersen BE. 2003. An overview of stem cell research and regulatory is sues. Mayo Clin Proc 78(8):993-1003. Collins AT, Berry PA, Hyde C, Stower, MJ, Ma itland, NJ. 2005. Prospective identification of tumorigenic prostate cancer st em cells. Cancer Res 65:10946-10951. Corley RA, Bormett GA, Ghanayem BI. 1994. P hysiologically based pharmacokinetics of 2butoxyethanol and its major metablite, 2-butoxya cetic acid, in rats and humans. Toxicol Appl Pharmacol 129,61-79. Dalerba P, Cho RW, Clarke FM. 2007. Cancer st em cells: Models and concepts. Annu Rev Med 58:18.1-18.18


76 Desai N, Trieu V, Yao Z, Louie L, Ci S, Yang A, Tao C, De T, Beals B, Dykes D, Noker P, Yao R, Labao E, Hawkins M, Soon-Shiong P. 2006. Increased antitumor activity, intratumor paclitaxel concentrations, and endothelial cell transport of cremophor-free, albuminbound paclitaxel, ABI-007, compared with crem ophor-based paclitaxel. Clin Cancer Res 12(12):3869 Fang D, Nguyen TK, Leishear K, Finko R, Kulp AN, Hotz S, Van Belle PA, Xu X, Elder DE, Herlyn M. 2005. A tumorigenic subpopulation with stem cell prope rties in melanomas. Cancer Res 65:9328-9337. Geraghty MT, Vaughn D, Nichols on AJ, Lin WW, Jimenez-Sanchez G, Obie C, Flynn MP, Valle D, Hu CA. 1998. Mutations in the De lta1-pyrroline 5-carboxylate dehydrogenase gene cause type II hyperprolinemia Hum Mol Genet. 7(9):1411-1415. Gordon N. 2004. Succinic semialdehyde dehydrogen ase deficiency (SSADH) (4-hydroxybutyric aciduria, gamma-hydroxybutyric aciduria).Eur J Paediatr Neurol 8(5):261-265. Ho MM, Ng AV, Lam S, Hung JY. 2007. Side population in human lung cancer cell lines and tumors is enriched with stem-lik e cancer cells. Cancer Res 67:10. Ignatova TN, Kukekov VG, Laywell ED, Susl ov ON, Vrionis FD, Steindler DA. 2002. Human cortical glial tumors contain neural stemlike cells expressing astroglial and neuronal markers in vitro. Glia 39:193-206. Jelski W, Zalewski B, Chrostek L, Szmitkowski M. 2004. The activity of cla ss I, II, III, and IV alcohol dehydrogenase isoenz ymes and aldehyde dehydrogenase in colorectal cancer. Digestive Diseases and Sciences 49(6):977-981. Jelski W, Chrostek L, Szmitkowski M, Markiewicz W. 2006. The activity of class I, II, III and IV alcohol dehydrogenase isoenzymes and alde hyde dehydrogenase in breast cancer. Clin Exp Med 6:89-93. Kastan MB, Schlaffer E, Russo JE, Colvin OM Civin CI, Hilton J. 1990. Direct demonstration of elevated aldehyde dehydrogenase in human hematopoietic progenitor cells. Blood 75(10):1947-1950. Lam S, Hung JY. 2007. Side population in human l ung cancer cell lines and tumors is enriched with stem-like cancer cells. Cancer Res 67:10. Li C, Heidt DG, Dalerba P, Burant CF, Zhang L, Adsay V, Wicha M, Clarke MF, Simeone DM. 2007. Identification of pancreatic cancer stem cells. Cancer Res 67(3):1030-1037. Lindahl R 1992. Aldehyde dehydrog enases and their role in ca rcinogenesis. Crit Rev Biochem Mol Biol 27(4-5):283-335. Liu S, Dontu G, Mantle ID, Patel S, Ahn NS, Jackson KW, Suri P, Wicha MS. 2006. Hedgehog signaling and Bmi-1 regulate self-renewal of normal and malignant human mammary stem cells. Cancer Res 66:6063-6071.


77 Luo P, Wang A, Payne KJ, Peng H, Wang JG, Parri sh YK, Rogerio JW, Triche TJ, He Q, Wu L. July 2007. Intrinsic RAR{alpha}-CAK signaling involves coordination of the restricted proliferation and granulocytic differentiati on of human hematopoietic stem cells. Stem Cells 2007-0264. (Stem Cells Express, published online July 12, 2007; doi:10.1634/stemcells.2007-0264) Marchitti SA, Deitrich RA, Vasiliou V. 2007. Neurotoxicity and metabolism of the catecholamine-derived 3,4-di hydroxyphenylacetaldehyde and 3,4dihydroxyphenylglycolaldehyde: The role of aldehyde dehydrogenase. Pharmacol Rev 59(2):125-150. Moreb JS, Zucali JR, Ostmark B, Benson NA. 20 07a. Heterogeneity of aldehyde dehydrogenase expression in lung cancer cell lines is reve aled by ALDEFLUOR flow cytometry-based assay. Cytometry Part B. Clinical Cytometry 72B:281-289. Moreb JS, Mohuczy D, Ostmark B, Zucali JR 2007b RNAi-mediated knockdown of aldehyde dehydrogenase class-1A1 and class-3A1 is sp ecific and reveals that each contributes equally to the resistance against 4-hydroperoxycyclophosphamide. Cancer Chemother Pharmacol 59(1):137 Nakayama M, Gonzalgo ML, Yegnasubraman ian S, Lin X, Marzo AMD, Nelson WG. 2004. GSTP1 CpG island hypermethylation as a mo lecular biomarker for prostate cancer. Journal of Cellular Bioc hemistry 91(3): 540-552. O'Brien CA, Pollett A, Gallinger S, Dick JE. 2007. A human colon cancer cell capable of initiating tumour growth in immuno deficient mice. Nature 445:106-110. Olempska M, Eisenach PA, Ammerpohl O, U ngefroren H, Fandrich F, Kalthoff H. 2007. Detection of tumor stem cell markers in pa ncreatic carcinoma cell lines. Hepatobiliary Pancreat Dis Int 6:92-97. Pappa A, Brown D, Koutalos Y, DeGregori J, White C, Vasiliou V. 2005. Human aldehyde dehydrogenase 3A1 inhibits proliferation a nd promotes survival of human corneal epithelial cells. The Journal of Biological Chem 280(30):27998-28006. Patrawala L, Calhoun T, Schneider-Broussard R, Li H, Bhatia B, Tang S, Reilly JG, Chandra D, Zhou J, Claypool K, Coghlan L, Tang DG. 2006. Highly purified CD44+ prostate cancer cells from xenograft human tumo rs are enriched in tumorige nic and metastatic progenitor cells. Oncogene 25:1696-1708. Pereira F, Rosenmann E, Nylen E, Kaufma n M, Pinsky L, Wrogemann K. 1991. The 56 kDa androgen binding protein is an aldehyde dehydrogenase. Bioche m Biophys Res Commun 175:831-838. Pijnappel WW, Hendriks HF, Folkers GE, van den Brink CE, Dekker EJ, Edelenbosch C, van der Saag PT, Durston AJ. 1993. The retinoid li gand 4-oxo-retinoic acid is a highly active modulator of positional specifica tion. Nature 366(6453):340-344.


78 Reya T, Morrison SJ, Clarke MF, Weissman IL 2001. Stem cells, cancer, and cancer stem cells. Nature 414:105-111. Ricci-Vitiani L, Lombardi DG, Pilozzi E, Biffoni M, Todaro M, Peschle C, De Maria R. 2007. Identification and expansion of human colon-cancer-initiating cells. Nature 445(7123):111-115. Richardson GD, Robson CN, Lang SH, Neal DE, Maitland NJ, Collins AT. 2004. CD133, a novel marker for human prostatic epithe lial stem cells. J Cell Sci 117:3539-3545. Sanders RC, Slayton WB, Cogle CR, Fisher RC, Scott EW. 2006. Stem cell research. Paediatric Respiratory Reviews 7:135-140. Schnier JB, Kaur G, Kaiser A, Stinson SF, Sa usville EA, Gardner J, Nishi K, Bradbury EM, Senderowicz AM. 1999. Identification of cyto solic aldehyde dehydroge nase 1 from nonsmall cell lung carcinomas as a flavopirid ol-binding protein. FE BS Lett. 454:100-104. Singh SK, Clarke ID, Hide T, Dirks PB. 2004. Cancer stem cells in nervous system tumors. Oncogene 23:7267-7273. Singh SK, Clarke ID, Terasaki M, Bonn VE, Hawkins C, Squire J, Dirks PB. 2003. Identification of a cancer stem cell in human br ain tumors. Cancer Res 63:5821-5828. Singh SK, Hawkins C, Clarke ID, Squire JA, Ba yani J, Hide T, Henkelman RM, Cusimano MD, Dirks PB. 2004. Identification of human brain tumour initiating cells. Nature 432:396401. Sldek NE, Kollander R, Sreerama L, Ki ang DT. 2002. Cellular levels of aldehyde dehydrogenases (ALDH1A1 and ALDH3A1) as pr edictors of therap eutic responses to cyclophosphamide-based chemotherapy of br east cancer: a retrospective study. Rational individualization of oxazaphos phorine-based cancer chemothe rapeutic regimens. Cancer Chemother Pharmacol 49(4):309-321. Sovic A, Borovic S, Loncaric I, Kreuzer T, Zarkovic K, Vukovic T, Wg G, Hrascan R, Wintersteiger R, Klinger R, Zurak N, Sc haub RJ, Zarkovic N. 2001. The carcinostatic and proapoptotic potential of 4-hydroxynonenal in HeLa cells is associated with its conjugation to cellular proteins Anticancer Res 21(3B):1997-2004. Storms, RW, Trujillo, AP, Springer, JB, Shah L, Colvin OM, Ludeman SM, and Smith C. 1999. Isolation of primitive human hematopoietic progenitors on the basis of aldehyde dehydrogenase activity. Proc. Natl. Acad. Sci. Cell Biology 96:9118-9123. Suetsugu A, Nagaki M, Aoki H, Motohashi T, Kunisada T, Moriwaki H. 2006. Characterization of CD133+ hepatocellu lar carcinoma cells as cancer stem/progenitor cells. Biochem Biophys Res Commun 29:351(4):820-824.


79 Takebe N, Zhao S, Adhikari D, Mineishi S, Sade lain M, Hilton J, Colvin M, Banerjee D, Bertino JR. 2001. Generation of dual resistance to 4-hydroperoxycyclophosphamide and methotrexate by retroviral transfer of th e human aldehyde dehydrogenase Class 1 gene and a mutated dihydrofolate reductase gene. Molecular Therapy 3:88-96. Tan BT, Park CY, Ailles LE, Weissman IL. 2006. The cancer stem cell hypothesis: A work in progress. Lab Investigation 86:1203-1207. Townsend AJ, Leone-Kabler S, Haynes RL, Wu Y, Szweda L, Bunting KD. 2001. Selective protection by stably transfected human AL DH3A1 (but not huma n ALDH1A1) against toxicity of aliphatic aldehydes in V79 cells Chem Biol Interact 130-132(1-3):261-273. Vasiliou V, Nebert DW. 2005. Analysis and update of the human aldehyde dehydrogenase (ALDH) gene family. Hu m Genomics 2(2):138-143. Wicha MS, Liu S, Dontu G. Cancer Stem Cells: An old ideaA paradigm shift. Cancer Res 66:(4):1883-1890. Yazdanpanah M, Luo X, Lau R, Greenberg M, Fisher LJ, Lehotay DC. 1997. Cytotoxic aldehydes as possible markers for childhood can cer. Free Radic Biol Med 23(6):870-878. Yin S, Li J, Hu C, Chen X, Yao M, Yan M, Jian g G, Ge C, Xie H, Wan D, Yang S, Zheng S, Gu J. 2007. CD133 positive hepatocellular carc inoma cells possess high capacity for tumorigenicity. Int J Cancer 120:1444-1450. Yokoyama T, Yokoyama A, Kato H, Tsujinaka T, Muto M, Omori T, Haneda T, Kumagai Y, Igaki H, Yokoyama M, Watanabe H, Yosh imizu H. 2003. Alcohol flushing, alcohol and aldehyde dehydrogenase genotypes, and risk for esophageal squamous cell carcinoma in Japanese men. Cancer Epidemiol Bi omarkers Prev. 12(11 Pt 1):1227-1233.


80 BIOGRAPHICAL SKETCH Deniz A. Ucar was born on February 22, 1974 in Ankara, Turkey. After graduating from Ozel Ari Kolleji High School in Ankara, Turkey, she attended University of Istanbul, where she earned her Doctor of Veterinary Medicine (DVM) degree in 1997. After four years of volunteer work and employment in various animal hospita ls, Deniz had her own veterinary practice in Ortakoy, Istanbul. Encouraged by her professors, sh e left her successful ve terinary practice to come to the US to do science. She came to the US speaking no English. While she was learning English at the English Language Institute, she wa s awarded a scholarship and then finished the premedical program at Santa Fe Community Collage in one year. In 2005, she became a volunteer in the laborato ry of Dr. Lung-Ji Chang to gain some experience while awaiting her masters program application result. Deniz started the University of Floridas College of Medicine masters program in fall 2005. After joining the laboratory of Dr. Edward Scott, she began working on different projects and completed her masters program with the project on establishing a system for the enrichment of cancer initiating cells. After graduation, she will continue on as a student to pursue a doctorate degree in the University of Floridas College of Medicine IDP Program.