William Coleman, PhD

William B. Coleman, PhD


Office: 919-966-2699

Email: william_coleman@med.unc.edu

Research Interests

The research in our laboratory involves several major projects, including investigations related to the molecular pathogenesis of cancer and investigations related to the biology of liver stem-like progenitor cells. Four projects address the molecular pathogenesis of human cancers, including (i) identification, isolation, and characterization of human liver tumor suppressor genes, (ii) investigation of the methylation-dependent epigenetic regulation of BRCA1 in sporadic and hereditary breast cancers, (iii) investigations of mechanisms governing aberrant DNA methylation in human breast cancer, and (iv) investigations into molecular discrimination of multiple synchronous primary lung cancers. Two projects are focused on liver stem-like progenitor cells, including (i) liver progenitor cell responses to toxic liver injury, and (ii) transplantation of liver stem-like progenitor cells for correction of genetic liver disease. These projects are described in greater detail below.

Identification, Isolation, and Characterization of Human Liver Tumor Suppressor Genes. Hepatocellular carcinogenesis is a multi-step process that requires altered expression of multiple genes, effected through genetic and epigenetic mechanisms, resulting in a population of cells that has evolved autonomous growth control and is not sensitive to the endogenous negative control mechanisms that are present in the liver parenchyma. Several lines of evidence suggest that molecular mechanisms involving common genetic loci or inactivation of common tumor suppressor genes may govern the development of liver tumors in humans and rodents. This idea is based upon the observation that several chromosomal regions that are frequently altered in human or rat liver neoplasms contain syntenic clusters of genes that are shared between humans, rats, and other mammals. These syntenic regions may contain orthologous tumor suppressor genes that are inactivated in hepatocarcinogenesis in both species. We have proposed that chromosome transfer studies utilizing human chromosomes and rat liver tumor cell lines may facilitate the identification and localization of human genes responsible for tumor suppression in the liver. To this end, we have established and characterized a model for the functional identification of human tumor suppressor genes via transfer into rat liver tumor cell lines. Human chromosome 11 has been implicated in the pathogenesis of several human tumors, including hepatoblastoma and hepatocellular carcinoma, and is syntenic to rat chromosome 1, which is structurally altered in a significant fraction of rat liver tumor cell lines, suggesting that these chromosomes may contain a common liver tumor suppressor gene. In our previous studies, we directly tested this possibility by transferring an intact copy of human chromosome 11 (containing a selectable marker gene) into highly aggressive rat liver tumor cell lines (that display characteristic chromosomal abnormalities) through microcell-mediated chromosome transfer. Using this system, we demonstrated that human chromosome 11 suppresses the tumorigenic potential of some rat liver tumor cell lines, providing direct evidence for the presence of a liver tumor suppressor gene on this human chromosome. Molecular characterization of suppressed microcell hybrid cell lines facilitated localization of the tumor suppressor locus to a small region of human 11p11.2, forming the basis for positional cloning of candidate genes from this chromosomal region. In early studies we sought to identify, isolate, and characterize candidate genes corresponding to the 11p11.2 liver tumor suppressor gene. Rapid advances in the Human Genome Project during the same period of time greatly accelerated progress towards identification of candidate liver tumor suppressor genes from 11p11.2. To date, we have made significant progress towards identification and characterization of the human 11p11.2 liver tumor suppressor gene: (i) we have extensively mapped the human 11p11.2 liver tumor suppressor region, constructed a comprehensive STS-based map of this chromosomal region, and assembled of a BAC/PAC contig spanning the entire liver tumor suppressor region; (ii) we have employed a candidate gene approach and transcription mapping of expressed sequence tags (ESTs) to identify a number of candidate liver tumor suppressor genes based upon their chromosomal localization and expression pattern among suppressed MCH cell lines; (iii) we have completed an extensive analysis of human HCCs that demonstrates LOH or large-scale deletion involving the 11p11.2 liver tumor suppressor region; (iv) we have shown diminished expression or loss of expression of individual candidate genes through expression analysis of human HCC cell lines and have accumulated preliminary results suggesting that candidate gene expression may be subject to epigenetic regulation involving methylation and/or chromatin remodeling in HCC cell lines; and (v) we have shown that the molecular mechanism of tumor suppression by human 11p11.2 may involve induction of expression of other tumor suppressor genes. The valuable reagents generated and the significant observations made during the initial phases of this project form the basis for continued investigations aimed at characterization of the human 11p11.2 liver tumor suppressor gene and determination of its role in the molecular pathogenesis of human hepatocellular carcinoma.

The continuing long-term goal of this research project is to determine the role of the human 11p11.2 liver tumor suppressor gene in the molecular pathogenesis of hepatocellular carcinoma, and to determine the mechanisms that govern the loss of function of this tumor suppressor in multi-step hepatocarcinogenesis in humans. Our current investigations focus on a small group of candidate liver tumor suppressor genes that were identified in our previous studies. The goals of these studies are to (i) characterize the involvement of candidate liver tumor suppressor genes in the suppression of the neoplastic phenotype of rat liver tumor cell lines using RNAi in vitro and in vivo, (ii) determine the ability of candidate genes to express tumor suppressor activity in vivo using transfected cell lines, (iii) evaluate the possible contributions of epigenetic mechanisms to the regulation of candidate liver tumor suppressor gene expression, (iv) examine the role of genetic alterations (LOH and/or mutation) in the inactivation of candidate liver tumor suppressor gene expression, (v) determine if alterations in candidate liver tumor suppressor gene expression represent early or later molecular alterations in multi-step hepatocarcinogenesis, and (vi) identify molecular targets and pathways in liver tumor cell lines that are subject to direct or indirect modification in response to candidate gene expression.

Investigation of the Methylation-dependent Epigenetic Regulation of BRCA1 in Sporadic and Hereditary Breast Cancers. Sporadic breast cancers account for about 90% of all cases, with hereditary breast cancer accounting for the balance. Large numbers of studies have examined the molecular pathogenesis of sporadic and hereditary breast cancer, but very few have examined the epigenetic contributions to this process. The possible contributions of methylation-dependent epigenetic regulation of BRCA1 in sporadic and hereditary breast cancer are currently under investigation in our laboratory.

While numerous molecular alterations have been identified in sporadic breast cancer, a definite role for the BRCA1 tumor suppressor gene has not been elucidated, although some evidence suggests the possibility that epigenetic silencing of BRCA1 expression may play a role in the molecular pathogenesis of this cancer. In the studies currently underway in our laboratory, we are examining the very basic question of whether methylation of the BRCA1 promoter results in reduced BRCA1 protein expression in sporadic breast cancers. We have assembled a group of 110 primary invasive ductal carcinomas of the breast from the archives of UNC Hospitals. Immunostaining of these tumors for BRCA1 protein expression reveals a significant subset of tumors with reduced protein expression or that lack detectable BRCA1 protein expression. Having identified tumors with normal, reduced, or no expression of BRCA1 protein, we have embarked upon determination of the methylation pattern of the BRCA1 promoter in the tumors. Our approach involves bisulfite DNA sequencing, which allows us to examine all CpG dinucleotides contained in the promoter region. These studies will (i) determine the prevalence of reduced BRCA1 protein expression among sporadic breast tumors, and (ii) determine if reduced BRCA1 expression correlates with modification of the BRCA1 promoter by hypermethylation. In future studies, we will perform correlative analysis to determine if statistically significant relationships exist between BRCA1 promoter methylation/silencing and clinical characteristics of tumors and/or patient outcome.

Hereditary breast cancers account for 10-15% of all breast cancer cases with ~50% associated with the susceptibility genes BRCA1 and BRCA2. Inheritance of a mutated copy of the BRCA1/2 genes increase the lifetime risk of breast cancer 5-8 fold and ovarian cancer 20-40 fold. Genetically-predisposed individuals typically present with cancer at an earlier age, with >50% of BRCA1 mutant carriers developing cancer by age 50. Members of breast cancer families often seek genetic counseling to assess their relative risk for cancer development. High-risk patients are evaluated to identify germline mutations in the BRCA1 (and/or the BRCA2) gene. The majority of these individuals carry a nucleotide sequence alteration in the BRCA1 gene (~80%) that results in a frameshift or missense mutation, whereas a subset of patients (~10%) possess chromosomal rearrangements affecting the gene. A third subset of patients, ~10% of high-risk patients, are found to lack discernable mutations in either BRCA1 or BRCA2, despite a calculated high probability for mutation based upon analysis of family history. We have suggested that methylation-dependent epigenetic silencing of the BRCA1 gene may account for a significant percentage of the cases in this subset of patients. Recent studies show that methylation-dependent epigenetic silencing of BRCA1 can contribute to the development of breast cancer that is indistinguishable from that of patients with BRCA1 mutation. Current studies in our laboratory are aimed at determining whether methylation-dependent epigenetic silencing of the BRCA1 gene confers breast cancer susceptibility in high-risk patients with a strong family history of the disease, but no discernible mutation of the BRCA1 gene. We have identified a group of patients that lack mutations in BRCA1/2 but that have early onset breast cancer. Using tumor samples form these patients, we are (i) examining the methylation status of the BRCA1 promoter in constitutional DNA of these patients, (ii) evaluating the expression of the BRCA1 protein in breast tumors from these patients, and (iii) determining if hypermethylation of the BRCA1 promoter influences the expression of BRCA1 in breast cancers from these patients. Further, we will identify the specific regions of the BRCA1 promoter that mediate methylation-dependent gene silencing, enabling the development of a practical molecular assay for constitutional methylation of the BRCA1 promoter, complementing standard mutation analyses of the BRCA1 gene, and providing useful information to the genetic counselor advising high-risk patients.

Investigations of Mechanisms Governing Aberrant DNA Methylation in Human Breast Cancer. Breast carcinogenesis is known to be associated with both genetic and epigenetic events. Whereas a number of epigenetically-silenced genes have been identified in breast cancer and suggested to be causally related to neoplastic transformation of breast epithelia, no studies have emerged that survey alterations in gene expression in response to changes in DNA methylation in a breast cancer model system. Therefore, to identify genes that are epigenetically-regulated in human breast cancer, we treated MCF-7 breast cancer cells with the demethylating agent 5-aza-2’-deoxycytidine (5-aza) and the histone deacetylase inhibitor trichostatin A (TSA), and gene expression patterns were examined by microarray analysis. MCF-7 cells were treated for 3 weeks with 250 nM 5-aza or 5-aza + 50 nM TSA, and then allowed to recover for 5 weeks after treatment withdrawal. Through analysis of the microarray data, we identified 37 genes that were associated with a >2-fold increase in 5-aza-treated MCF-7 cells, but returned to control levels after withdrawal of treatment. Similarly, 70 genes were identified in 5-aza + TSA treated MCF-7 cells, that returned to control levels after treatment withdrawal. Comparative analysis revealed 20 genes represented in both groups of increased genes. In addition, 21 genes demonstrated a >2-fold decrease in expression level in 5-aza-treated MCF-7 cells, but returned to control levels after withdrawal of treatment, and 79 genes were associated with a >2-fold decrease in 5-aza + TSA treated MCF-7 cells, but returned to control levels after treatment withdrawal. Comparative analysis revealed 7 genes represented in both groups of decreased genes. DNA sequence analysis of the promoter and 5’-upstream sequences of the increased and decreased gene sets identified some interesting features that may influence the epigenetic regulation of these genes. Collectively, the genes identified in this study will be valuable for the continued investigation of epigenetic mechanisms in breast carcinogenesis, as well as molecular mechanisms responsible for aberrant DNA methylation in breast cancer. In addition, these genes will be useful in studies aimed at identification of DNA sequence features that govern or direct sequence-specific methylation and epigenetic regulation of gene expression. Analysis of gene promoters with classic CpG islands may enable the identification of novel control sequences that regulate methylation in breast cancer, or may confirm regulation by previously identified sequence elements. In addition, further characterization of the promoter sequences of genes that lack easily recognizable CpG islands or other CpG-containing sequence features may identify novel methylation-sensitive regulatory sequences, or sequence-specific methylation events that abrogate the binding of transcription factors or other regulatory proteins that are essential for gene expression. Likewise, additional investigation of the decreased gene set may identify new methylation-dependent mechanisms of gene regulation, or indirect pathways for down-regulation of genes in cancer cells.

Investigations into Molecular Discrimination of Multiple Synchronous Primary Lung Cancers. Cancer of the lung and bronchus accounted for an estimated 158,900 deaths in 1999, representing 28% of all cancer deaths, and the leading cause of cancer deaths among men (31% of deaths) and women (25% of deaths). The majority of lung cancers are attributable to exposure to known carcinogenic agents, particularly cigarette smoke. Therapy for lung cancer varies depending upon the tumor type and other clinical variables (stage, grade, location, size). Surgery is the preferred treatment choice for some tumor types (squamous cell carcinoma, adenocarcinoma), whereas other types are generally treated with chemotherapy (small cell lung carcinoma). Despite the variety of treatment modalities that can be applied to lung cancer management, the overall survival rates for affected individuals are not good. The average five year survival rates for all patients and all stages of disease is only 13-14%. The survival rate increases to 49% if the disease is detected early (localized), but only 15% of tumors are discovered this early. The majority of lung cancer cases (48%) are not detected until after the development of distant metastases, leading to a dismal 5-year survival rate (2%). The overall poor probability of surviving lung cancer probably reflects (i) the difficulty with early detection (in the absence of active monitoring) of this tumor, (ii) the failure to detect tumors while localized, and (iii) the ineffectiveness of currently available non-surgical therapies (radiation and chemotherapy). Thus, increasing lung cancer patient survival depends upon early detection (when tumors are localized) and surgical intervention. However, the decision to perform curative surgery for lung cancer can be complicated by the presence of multiple tumor nodules. It is recognized that multiple primary tumors can arise in the lung due to the generalized exposure of this epithelium to the multiplicity of carcinogens contained in cigarette smoke. Estimates of the incidence of synchronous lung cancers range from 1% to 14% of cases, depending upon the diagnostic criteria employed. These synchronous tumors can arise in the same lobe of the lung, or in different lobes, and may be histologically distinct or identical. Patients with multiple (synchronous) tumors are treated surgically, if it can be shown that the tumors represent multiple primary (independent) tumors versus local/regional metastasis from one primary tumor. However, this determination cannot typically be made using conventional histologic/staging criteria.

Due to the diagnostic difficulty of differentiating multiple primary (synchronous) lung tumors from locally invasive (metastatic) lung tumors, a percentage of patients receive inappropriate or less effective clinical treatment for their disease. We have suggested that the application of molecular diagnostic techniques would enable the discrimination of synchronous and metastatic lung cancers, based upon genetic differences in these tumors. The project currently underway in our laboratory is investigating the feasibility of using PCR-base allelotyping techniques to distinguish between multiple primary lung tumors and locally metastatic lung tumors. We have collected DNA samples from patients with multiple synchronous primary tumors, as well as from patients with known metastatic lesions. Microsatellite DNA fingerprints of these tumors are generated and compared, and differences in allelotype are scored. This analysis facilitates discrimination of tumors based upon differences in the microsatellite fingerprint. Current efforts are aimed at identification particularly sensitive and useful markers for this analysis, and development of subsets of markers that is sensitive and specific in this type of analysis.

Liver Progenitor Cell Responses to Toxic Liver Injury. We are investigating liver regeneration in rats that have been treated with retrorsine, which is an hepatotoxic pyrrolizidine alkaloid. In retrorsine-treated rats, the capacity for liver regeneration through the proliferation of hepatocytes is impaired, and replacement of liver mass lost to surgical partial hepatectomy is accomplished through the proliferation of a novel small hepatocyte-like progenitor cell population. Retrorsine-injured hepatocytes enter the cell cycle in response to partial hepatectomy, but do not undergo cell division. Instead, these cells arrest as megalocytes, which are ultimately replaced by the progeny of the small hepatocyte-like cells. We have characterized the temporal appearance of the small hepatocyte-like progenitor cell population, and have characterized their phenotype. These cells share phenotypic traits with fetal hepatoblasts, oval cells, and fully differentiated hepatocytes, but are morphologically and/or phenotypically distinct from each. Small hepatocyte-like cells emerge early following partial hepatectomy in retrorsine-treated rats, proliferate rapidly to form expanding cellular aggregates that replace megalocytic hepatocytes, and concurrently acquire the panoply of differentiated features typical of mature hepatocytes. Replacement of lost hepatocytes and complete hepatic regeneration from small hepatocyte-like progenitor cells has not been observed in other models of liver injury in which replication of residual hepatocytes is impaired, suggesting that the small hepatocyte-like cells represent a previously unrecognized progenitor cell compartment of the adult liver. Current efforts seek to (i) examine the cytokine-dependence of small hepatocyte-like cell activation after surgical partial hepatectomy (PH), (ii) evaluate the growth factor requirements for proliferation of small hepatocyte-like cells during liver regeneration, (iii) characterize the expression of protein components of cytokine and growth factor-mediated signaling pathways in small hepatocyte-like cells, (iv) analyze changes in gene expression that accompany activation and proliferation of small hepatocyte-like cells in response to partial hepatectomy.

Transplantation of Liver Stem-like Progenitor Cells for Correction of Genetic Liver Disease. In previous studies we have shown that the propagable WB-F344 rat liver epithelial cell line (i) expresses an hepatocytic phenotype in vitro when cultured in the presence of sodium butyrate, and (ii) engraft and differentiate into hepatocytes when transplanted into the liver of a syngeneic rat host. These results clearly demonstrated that cultured liver stem-like cells can produce hepatocyte progeny in vitro and in vivo under the appropriate conditions. In current studies, we are investigating the use of stem cell transplantation for correction of liver insufficiency related to a genetic defect. We are employing the Nagase rat model of analbuminemia in these studies. Nagase rats do not express a functional albumin gene and exhibit serum albumin levels that are too low to detect. We are transplanting WB-F344 cells into these rats to examine whether the differentiated progeny of the transplanted stem cells will partially or completely restore normal serum albumin levels in these animals. In addition, we are examining gene augmentation strategies to determine if the use of high level expression vectors can accentuate the normalization of the disease phenotype in the absence of increased numbers of transplanted cells.

Selected Publications

Sandhu, R., Roll, J.D., Rivenbark, A.G., and Coleman, W.B. (2015) Dysregulation of the epigenome in human breast cancer: Contributions of gene-specific DNA hypermethylation to breast cancer pathobiology and targeting the breast cancer methylome for improved therapy. Am. J. Pathol. 185:282-292.

Sandhu, R., Rivenbark, A.G., Mackler, R.M., Livasy, C.A., and Coleman, W.B. (2014) Dysregulation of microRNA expression drives aberrant DNA hypermethylation in basal-like breast cancer. Int. J. Oncol. 44:563-572.

Roll, J.D., Rivenbark, A.G., Sandhu, R., Parker, J.S., Jones, W.D., Carey, L.A., Livasy, C.A., and Coleman, W.B. (2013) Dysregulation of the epigenome in triple-negative breast cancers: Basal-like and claudin-low breast cancers express aberrant DNA hypermethylation. Exp. Mol. Pathol. 95:276-287.

Rivenbark, A.G., O’Connor, S.M., and Coleman, W.B. (2013) Molecular and cellular heterogeneity in breast cancer – Challenges for personalized medicine. Am. J. Pathol. 183:1113-1124.

Coleman, W.B. (2013) Breast cancer personalized medicine: Challenges and opportunities. Am. J. Pathol. 183:1036-1037.

Rivenbark, A.G. and Coleman, W.B. (2012) Field cancerization in mammary carcinogenesis – Implications for prevention and treatment of breast cancer. Exp. Mol. Pathol. 93:391-398.

Sandhu, R. Rivenbark, A.G., and Coleman, W.B. (2012) Loss of post-transcriptional regulation of DNMT3b by microRNAs accounts for the hypermethylation defect observed in a subset of breast cancers. Int. J. Oncol. 41:721-732.

Sandhu, R., Rivenbark, A.G., and Coleman, W.B. (2012) Enhancement of chemotherapeutic efficacy in hypermethylator breast cancer cells through targeted and pharmacologic inhibition of DNMT3b. Breast Cancer Research and Treatment 131:385-399.

Sandhu, R., Parker, J.S., Jones, W.D., Livasy, C.A., and Coleman, W.B. (2010) Microarray-based gene expression profiling for molecular classification of breast cancer and identification of new targets for therapy: Possibilities for improvement in the management of basal-like breast cancer. Lab Medicine 41:364-372.

Roll, J.D., Rivenbark, A.G., Jones, W.D., and Coleman, W.B. (2008) Overexpression of DNA methyltransferase 3b contributes to a hypermethylator phenotype in human breast cancer cell lines. Molecular Cancer 7:15.

Rivenbark, A.G., Livasy, C.A., Boyd, C.E., Keppler, D., and Coleman, W.B. (2007) Methylation-dependent silencing of CST6 in primary human breast tumors and metastatic lesions. Exp. Mol. Pathology 83:188-197.

Rivenbark, A.G., Jones, W.D., and Coleman, W.B. (2006) Methylation-dependent epigenetic silencing of CST6 in human breast cancer cell lines. Lab. Invest. 86:1233-1242.

Rivenbark, A.G., Jones, W.D., Risher, J.D., and Coleman, W.B. (2006) DNA methylation-dependent epigenetic regulation of gene expression in MCF-7 breast cancer cells. Epigenetics 1:32-44.

View list of publications from PubMed