Richard Carpenter, Ph.D.
308 Jordan Hall
1001 E. 3rd St
Bloomington, IN 47405
Phone: (815) 855-8412
Research Program Membership
Department of Biochemistry and Molecular Biology
Indiana University at Bloomington
Metastasis is the process whereby cells from a primary tumor migrate to a distant tissue and colonize to form secondary tumors. The process of metastasis is a complex, stepwise process that remains to be fully characterized. Metastasis is of the utmost importance clinically as it accounts for >90% of breast cancer deaths. The primary focus of my laboratory will be to study mechanisms of cancer progression, metastasis, and metastatic tumor treatment resistance and use the greater understanding of these mechanisms to initiate translational studies with the long-term hope of having significant clinical impact. Specifically, my research program will study the novel AKT-HSF1 signaling pathway that promotes breast cancer EMT and on truncated GLI1 (TGLI1), a novel alternative splice variant of the transcription factor GLI1 that promotes robust tumor angiogenesis. Discovery of the novel AKT-HSF1 Pathway in Breast Cancer Heat shock factor 1 (HSF1) is a transcription factor considered to be the master regulator of the heat shock response 1. Inactive HSF1 is relegated to the cytosol and under conditions of cell stress, such as heat shock, HSF1 is released from an inhibitory complex allowing HSF1 trimerization, nuclear localization, S326 phosphorylation, and upregulation of target genes. The most well-known HSF1 target genes include many of the heat shock proteins, which are expressed upon cell stress to preserve protein integrity, cell function, and cell survival ultimately. The relevance of HSF1 to cancer has become increasingly recognized in recent years with evidence that HSF1 is overexpressed in several cancer types, including breast cancer 2. It is also evident that HSF1 can regulate many genes independent of the heat shock program that support cancer phenotypes 3. In 2015, I published data indicating HSF1 may play a role in the early stages of breast cancer metastasis 4. HER2 was seen to promote epithelial-to-mesenchymal transition (EMT), a phenotypic transition of epithelial-like cancer cells to a more migratory mesenchymal-like phenotype that is considered an early step in the process of metastasis. Evidence indicated the EMT transcription factor Slug was a prominent factor in HER2-mediated EMT in multiple breast cancer cell lines. TFSearch, a tool to predict transcription factor binding, indicated the presence of HSF1 consensus binding sites within the Slug promoter. Further examination revealed HSF1 could directly bind to the Slug gene promoter at these consensus sites and upregulate Slug gene expression. This novel finding indicated a downstream mechanism for HER2-mediated EMT and prompted the question of how HER2 mediates HSF1 activity. HSF1 can be post-translationally modified at several residues but phosphorylation at S326 is a key modification necessary for HSF1 activity 5. When a panel of serine/threonine kinases known to be downstream of HER2 were screened, AKT activation was strongly correlated with HSF1 activation. Further experiments revealed a physical interaction between AKT and HSF1 and a kinase assay indicated AKT directly phosphorylated S326 of HSF1. Interruption of either PI3K/AKT or HSF1 via small molecular inhibitors or gene knockdown prevented HER2-induced Slug expression and EMT. While this pathway was discovered in HER2-positive breast cancer, the activity of this pathway is not limited to the HER2-enriched breast cancer subtype. Activation of HSF1 also occurs at similar rates in Luminal A/B and Basal subtypes as it does HER2-enriched tumors (40-50% tumors). The involvement of AKT-HSF1 signaling in EMT suggests this pathway may promote breast cancer metastasis. Taken together, these and other data suggest the AKT-HSF1 pathway promotes EMT of breast cancer and I hypothesize AKT-HSF1 contributes to breast cancer progression and metastasis. Future Directions for the AKT-HSF1 Pathway 1) Determine the effects of AKT-mediated HSF1 activation breast cancer progression and metastasis. It is unknown whether AKT-HSF1 signaling promotes metastasis of breast cancer as this pathway has not been investigated in metastasis of any cancer type. My observations indicate HSF1 is strongly activated in several breast cancer brain metastasis tumors. Furthermore, AKT and HSF1 are concurrently activated in metastatic variants of breast cancer cells and knockdown or inhibition of HSF1 significantly reduces the viability of these cells. Therefore, I hypothesize that AKT-mediated HSF1 activation promotes breast cancer progression and metastasis. To fully investigate this hypothesis I will utilize dominant negative (S326A) and constitutively active (S326D) forms of HSF1 in addition to CRISPR-mediated knockdown of AKT and HSF1 to a) determine the effects of S326 phosphorylation on HSF1 biologic activity including on the protein half-life of HSF1, subcellular localization of HSF1, and DNA binding activity of HSF1 to known target genes and their expression; b) determine the relationship between HSF1 activation and the three AKT isoforms by assessing which isoforms are co-activated with HSF1, which isoforms interact with and phosphorylated HSF1, and whether HSF1 phosphorylation and activity are altered with loss of each AKT isoform; c) determine the role of AKT-HSF1 signaling on malignant phenotypes in vitro including cell proliferation, apoptosis, cell migration, and cell invasion; d) determine the role of AKT-HSF1 signaling on breast cancer growth and metastasis in vivo using the mammary fat pad model and experimental metastasis models with intracardiac/intravenous injection of cells. 2) Assess the efficacy of combined inhibition of AKT and HSF1 in primary and metastatic breast cancer. Although the AKT-HSF1 pathway was discovered in HER2-positive breast cancer cells, the pathway is active in other subtypes as well prompting the question of whether inhibition of this pathway is efficacious in breast cancer. Tumor initiating cells (or cancer stem cells) are important to treatment resistance, tumor recurrence, and metastasis. HSF1 expression is correlated with expression of multiple markers for tumor initiating cells including CD44 and ESA (EpCAM). Both AKT and HSF1 have higher activation in mammospheres, an in vitro model to enrich the stem cell population. Mammosphere formation was enhanced by ectopic expression of HSF1 and reduced with interruption of HSF1. Small molecule inhibitors to AKT (MK-2206) and HSF1 (KRIBB11) reduced mammosphere formation as well as the CD44highCD24lowESAhigh subpopulation, a known tumor initiating subpopulation. Additionally, combination of these inhibitors results in a synergistic killing of breast cancer cells. Therefore, I hypothesize AKT-HSF1 signaling promotes a tumor initiating cell phenotype and combined inhibition of AKT and HSF1 can suppress breast cancer progression and metastasis. To fully investigate this hypothesis, I will use inhibitors to AKT and HSF1 as well as knockdown approaches to determine the ability of dual inhibition of AKT and HSF1 to suppress a) the tumor initiating cell population in vitro using the mammosphere model as well as known tumor initiating subpopulations, b) tumor growth from tumor initiating cells, c) primary tumor growth and spontaneous metastasis in vivo, d) metastasis using experimental metastasis models. 3) Identify an activation signature for the AKT-HSF1 pathway to predict patient outcomes. HSF1 expression level is not a predictors for HSF1 activity. Inactive HSF1 is suppressed in a protein complex with Hsp90 and RalBP1, both of which are released upon cell stress allowing HSF1 activation. Analysis of a breast cancer cohort revealed that HSF1 expression alone is a weak predictor for metastasis-free survival. However, patients with high expression of HSF1 along with high Slug, Hsp90, or Hsp70 (known HSF1 target genes) had significantly poorer metastasis-free survival suggesting HSF1 activity may be a good predictor for breast cancer patient outcome. Therefore, I hypothesize HSF1 activity can predict breast cancer patient outcomes. To fully address this hypothesis, I will a) generate a gene expression signature indicative of HSF1 activity with RNA-seq using a cell culture model with the pathway “turned on” or “turned off” and validate the signature using cell lines and patient tumors. The generated HSF1 activity signature will be used in conjunction with publicly available databases to determine the association of HSF1 activity with: b) tumor progression, c) metastasis-free survival, d) breast cancer subtype, e) genetic drivers of breast cancer, and f) sensitivity to breast cancer chemotherapy. Truncated GLI1 (TGLI1) is Novel Promoter of GBM Angiogenesis In a secondary project, I also plan to focus attention on truncated GLI1 (TGLI1), a gain-of-function alternative splice variant of the transcription factor GLI1, which is the terminal effector of the hedgehog signaling pathway. TGLI1 results from an alternative splicing whereby exon 3 and a portion of exon 4 are removed resulting in a loss of 123 bp (41 aa) 6,7. TGLI1 is expressed exclusively in cancer cells as TGLI1 has not been detected in any healthy tissues whereas TGLI1 has been observed in several tumor types including glioblastoma (GBM) and breast cancer. Compared to GLI1-expressing cells, cells expressing TGLI1 are more migratory, invasive, and grow larger tumors 7,8. Interestingly, TGLI1 tumors were more vascularized than GLI1 tumors and a subsequent angiogenesis PCR array uncovered several pro-angiogenic genes, such as VEGF-A and HPSE, which were upregulated in TGLI1 tumors 8-10. TGLI1 was found to directly bind and regulate the expression of several pro-angiogenic genes in this GBM xenograft model. These results were confirmed with a strong relationship between TGLI1 levels and expression of these identified target genes in a cohort of GBM patient tumors. Furthermore, a gene expression signature indicative of TGLI1 activity is a strong predictor of survival in GBM patients and metastasis-free survival in breast cancer patients. Therefore, I hypothesize TGLI1 is a novel promoter of tumor cell aggressiveness and tumor angiogenesis. Future Directions for TGLI1 1) Determine the role of TGLI1 in the therapeutic resistance of GBM TGLI1 expression is higher in GBM compared to lower grade gliomas and, specifically, in the mesenchymal subtype of GBM. GBM is more resistant to standard therapies compared to lower grade gliomas while the mesenchymal subtype is the most resistant to therapy among GBM subtypes 11,12. Temozolomide (TMZ) is the standard of care chemotherapy for GBM in conjunction with surgical resection and radiotherapy. Despite the aggressive nature this treatment strategy, 90% of GBMs recur at the primary site indicating these tumors have some therapeutic resistance. Additionally, TGLI1 is strongly pro-angiogenic and GBM is characteristically resistant to anti-angiogenic treatments (e.g. bevacizumab) with a high percentage of recurrence following this treatment. Furthermore, a gene expression signature for TGLI1 activity predicts patient response to bevacizumab. Therefore, I hypothesize TGLI1 may contribute to the therapeutic resistance characteristic of GBM. To fully address this hypothesis, I will investigate the role of TGLI1 in resistance to TMZ resistance and bevacizumab resistance. I will initially do a screening process to determine the contribution of TGLI1 to the resistance of TMZ and bevacizumab by assessing: a) the response of different GBM cell lines to these treatments with differing levels of TGLI1 expression, b) the expression of TGLI1 in tumors classified as resistant to these treatments, c) the expression of TGLI1 in a cell culture model of resistance and whether knockdown of TGLI1 alters the sensitivity to these treatments, and d) use publicly available gene expression databases to assess the ability of TGLI1 activity to predict resistance to these treatments (data available for both TMZ and bevacizumab resistance). If any further role of TGLI1 in GBM therapeutic resistance is found using these screening experiments, in addition to TGLI1 activity predicting bevacizumab response, the underlying mechanism will be investigated. TGLI1 is a transcription factor and therefore it is likely TGLI1 promotes therapeutic resistance by regulation of its target genes. Therefore, a cell culture model of bevacizumab resistance (and any other modes of resistance TGLI1 is found to have a role) along with a treatment-sensitive cell line will be used to: i) assess the differential expression of known TGLI1 target genes in resistant and sensitive lines, ii) determine whether knockdown of TGLI1 itself or its target genes affects treatment sensitivity, and iii) screen for other potential genes using a PCR cancer drug resistance array with and without TGLI1 knockdown. 2) Determine the role of TGLI1 in other Tumor Types TGLI1 expression has only been studied in GBM and breast cancer to date. However, a gene expression signature indicative of TGLI1 activity is predictive of patient survival in a cohort of lung adenocarcinoma. Lung adenocarcinoma is a subtype of non-small cell lung cancer (NSCLC) and accounts for approximately 40% of all lung cancers. Additionally, angiogenesis is an important contributor to development and progression of lung adenocarcinoma. Considering the pro-angiogenic nature of TGLI1 in other tumor types, I hypothesize TGLI1 contributes to angiogenesis lung adenocarcinoma. To fully address this hypothesis, I will 1) screen expression databases to determine whether the TGLI1 activity signature is differentially expressed in tumors compared with healthy controls and/or whether TGLI1 activity is predictive of patient outcomes in other solid tumors including colorectal, hepatocellular, ovarian, prostate, pancreatic, and melanoma cancers; 2) directly assess the percentage of lung adenocarcinomas (and any other tumor type wherein TGLI1 activity predicts patient outcome) that express TGLI1; and lastly 3) determine whether TGLI1 expression promotes migration, invasion, and angiogenesis in lung adenocarcinoma (and other tumor types identified from above) cells. References 1. Westerheide, S. D. et al. Curr Protein Pept Sci 13, 86-103 (2012). 2. Santagata, S. et al. PNAS 108, 18378-18383 (2011). 3. Mendillo, M. L. et al. Cell 150, 549-562 (2012). 4. Carpenter, R. L. et al. Oncogene 34, 546-557 (2015). 5. Guettouche, T. et al. BMC Biochemistry 6, 4 (2005). 6. Carpenter, R. L. & Lo, H. W. Vitamins and hormones 88, 115-140 (2012). 7. Lo, H. W. et al. Cancer research 69, 6790-6798 (2009). 8. Zhu, H. et al. Cancer letters 343, 51-61 (2014). 9. Cao, X. et al. Oncogene 31, 104-115 (2012). 10. Carpenter, R. L. et al. Oncotarget 6, 22653-22665 (2015). 11. Phillips, H. S. et al. Cancer cell 9, 157-173 (2006). 12. Verhaak, R. G. et al. Cancer cell 17, 98-110 (2010).
Post-doctoral Fellowship - Wake Forest University School of Medicine, Winston Salem, NC 06/2017
Post-doctoral Fellowship - Duke University School of Medicine, Durham, NC 06/2014
Ph.D. - Thomas Jefferson University, Philadelphia, PA 05/2012