Final Report Summary - CDELP (CHD1 Deletion: Implications to Outcome and Treatment in Prostate Cancer)
This fellowship project aimed to characterize a subclass of prostate cancer with deletion in the gene chromodomain helicase DNA binding protein 1 (CHD1). Tumors with deletions of CHD1 have a distinct molecular phenotype characterized by increased genome instability, specific DNA methylation pattern, increased AR activity and often mutations in the gene SPOP. Although the genomic loss of CHD1 compared to rarely occurring somatic mutations suggest that CHD1 might function as tumor suppressor in prostate cancer, its function during tumor progression is not well understood. In order to implement the findings from recent prostate cancer genome studies into a more precise patient care, we sought to analyze whether CHD1 status predicts outcome or response to currently available therapies. Furthermore, as CHD1 deleted prostate cancers are genomically unstable, we aimed to evaluate whether these cancers would benefit from DNA damaging therapy. Finally, in order to unravel the molecular mechanism altered due to CHD1 deletion, we performed in vitro studies in newly established models of CHD1 deleted prostate cancer.
CHD1 expression and clinical outcome
We initially established a variety of readouts to analyze CHD1 expression and copy number status in patient samples including formalin-fixed and paraffin embedded (FFPE) tumor biopsies as well as circulating cell-free tumor DNA. These readouts include CHD1 immunohistochemistry (IHC), fluorescence in situ hybridization (FISH) of FFPE tumor samples as well as digital droplet PCR (ddPCR) on cell- free tumor DNA extracted from patients plasma. We then established a clinically well-annotated patient cohort of men (n=68), who developed metastatic castration resistant prostate cancer (mCRPC). This cohort reflects a population of men with highly aggressive prostate cancer with the majority having progressed to docetaxel (88%) and abiraterone (94%). Patient-matched biopsies from primary hormone sensitive prostate cancer (HSPC) and mCRPC were included. We evaluated CHD1 protein expression and copy number status during disease progression and correlated this with outcome and response to docetaxel and abiraterone. We found that patients bearing tumors with above median expression of CHD1 (CHD1high) at time of diagnosis had a significantly worse overall survival than patients with lower than median expression of CHD1 (CHD1low). When comparing the patient-matched diagnostic and CRPC biopsies we found an increase of CHD1 expression during disease progression. Although not reaching statistical significance, there was a trend towards shorter time on abiraterone and docetaxel in CHD1high tumors suggesting a faster development of resistance to these drugs. Thus, increased CHD1 expression but not deletion marks a subset of patients with more aggressive disease and poorer prognosis.
Function of CHD1 in DNA double strand break repair
In order to functionally characterize the role of CHD1 deletion in prostate cancer cells, we created novel isogenic benign prostate epithelial and prostate cancer cell line models expressing or lacking CHD1. We used the CRISPR/ CAS9 approach to introduce insertions and deletions (indels) into the cancer cell genome leading to truncated CHD1 protein isoforms. We performed single cells sorting of CHD1-CRISPR/CAS9 transfected cells and expanded single cell derived isogenic clones. The isogenic cell clones were than characterized for CHD1 status, proliferation, cell cycle, tumor formation in mouse xenograft models, transcriptional profiling, synthetic lethality siRNA screens, anticancer drug screens and in depth molecular analysis of DNA double strand break (DSB) repair competence. To mechanistically study the role of CHD1 in DSB repair we established immunoprecipitation assays for endogenous protein and analyzed its prostate cancer-specific interactome by mass spectrometry. When determining the CHD1 interactome in prostate cancer cells using immunoprecipitation we found that CHD1 interacts with components of the DNA double strand break (DSB) repair machinery. We then analyzed whether loss of CHD1 alters the competence CRPC cells to repair DNA double strand breaks and found that although the overall DSB resolution was similar, CHD1 deleted cells had an increased error-prone DSB repair mode (NHEJ) activity compared to CHD1 wildtype parental cells. Mechanistically, this can be explained by an increased expression of NHEJ pathway genes in CHD1 deleted CRPC cells. We performed drug screens using approved anticancer drugs to more precisely target the CHD1 deleted subclass of prostate cancer and found and increased sensitivity for DNA-PK inhibition. DNA-PK functions as an early mediator of the error-prone NHEJ DSB repair response, which is elevated in CHD1 deleted cells. In order to identify further molecular targets to specifically target CHD1 deleted cancer cells we performed a siRNA screen looking for synthetic lethal interactions. This analysis identified genes, which are involved in DSB repair, replication and cell cycle checkpoint control further suggesting that the CHD1 deleted subclass of prostate cancer might benefit from DNA damaging therapy.
CHD1 deletion and response to DNA damaging therapy
To evaluate whether CHD1 deleted cancers would benefit from DNA damaging therapy we analyzed the CHD1 copy number status in a cohort of mCRPC patients, which have been treated with the PARP inhibitor olaparib (n=49) and correlated CHD1 status with response. We did not identify cases with homozygous CHD1 deletion in the first part of the trial. CHD1 heterozygous deletions and copy number gains were equally common in the responding and non-responding patients. We will expand this analysis in the next parts of the trial to determine whether homozygous loss of CHD1 sensitizes to olaparib. Interestingly, in preclinical studies using a CRPC patient-derived organoid (PDO) model with a homozygous loss of CHD1, we found an increased sensitivity to olaparib compared to PDOs with normal CHD1 copy number. In addition, we started to collect and process biopsies from men with mCRPC participating in an observational clinical study, which aims to evaluate the short and long term safety profile of Radium-223 (Reassure, NCT02141438). We will correlate CHD1 expression with response to DNA damaging Radium-223 treatment.
Conclusions
Our tissue-based analysis of clinically aggressive prostate cancer suggests that high CHD1 expression rather then CHD1 loss is associated with a more aggressive disease. Although our current results do not answer, whether CHD1 deleted tumors are more sensitive to DNA damaging therapy, preclinical in vitro analysis suggest a role for CHD1 in DSB repair, which when lost, sensitizes cancer cells to olaparib and DNA-PK inhibitor treatment.
CHD1 expression and clinical outcome
We initially established a variety of readouts to analyze CHD1 expression and copy number status in patient samples including formalin-fixed and paraffin embedded (FFPE) tumor biopsies as well as circulating cell-free tumor DNA. These readouts include CHD1 immunohistochemistry (IHC), fluorescence in situ hybridization (FISH) of FFPE tumor samples as well as digital droplet PCR (ddPCR) on cell- free tumor DNA extracted from patients plasma. We then established a clinically well-annotated patient cohort of men (n=68), who developed metastatic castration resistant prostate cancer (mCRPC). This cohort reflects a population of men with highly aggressive prostate cancer with the majority having progressed to docetaxel (88%) and abiraterone (94%). Patient-matched biopsies from primary hormone sensitive prostate cancer (HSPC) and mCRPC were included. We evaluated CHD1 protein expression and copy number status during disease progression and correlated this with outcome and response to docetaxel and abiraterone. We found that patients bearing tumors with above median expression of CHD1 (CHD1high) at time of diagnosis had a significantly worse overall survival than patients with lower than median expression of CHD1 (CHD1low). When comparing the patient-matched diagnostic and CRPC biopsies we found an increase of CHD1 expression during disease progression. Although not reaching statistical significance, there was a trend towards shorter time on abiraterone and docetaxel in CHD1high tumors suggesting a faster development of resistance to these drugs. Thus, increased CHD1 expression but not deletion marks a subset of patients with more aggressive disease and poorer prognosis.
Function of CHD1 in DNA double strand break repair
In order to functionally characterize the role of CHD1 deletion in prostate cancer cells, we created novel isogenic benign prostate epithelial and prostate cancer cell line models expressing or lacking CHD1. We used the CRISPR/ CAS9 approach to introduce insertions and deletions (indels) into the cancer cell genome leading to truncated CHD1 protein isoforms. We performed single cells sorting of CHD1-CRISPR/CAS9 transfected cells and expanded single cell derived isogenic clones. The isogenic cell clones were than characterized for CHD1 status, proliferation, cell cycle, tumor formation in mouse xenograft models, transcriptional profiling, synthetic lethality siRNA screens, anticancer drug screens and in depth molecular analysis of DNA double strand break (DSB) repair competence. To mechanistically study the role of CHD1 in DSB repair we established immunoprecipitation assays for endogenous protein and analyzed its prostate cancer-specific interactome by mass spectrometry. When determining the CHD1 interactome in prostate cancer cells using immunoprecipitation we found that CHD1 interacts with components of the DNA double strand break (DSB) repair machinery. We then analyzed whether loss of CHD1 alters the competence CRPC cells to repair DNA double strand breaks and found that although the overall DSB resolution was similar, CHD1 deleted cells had an increased error-prone DSB repair mode (NHEJ) activity compared to CHD1 wildtype parental cells. Mechanistically, this can be explained by an increased expression of NHEJ pathway genes in CHD1 deleted CRPC cells. We performed drug screens using approved anticancer drugs to more precisely target the CHD1 deleted subclass of prostate cancer and found and increased sensitivity for DNA-PK inhibition. DNA-PK functions as an early mediator of the error-prone NHEJ DSB repair response, which is elevated in CHD1 deleted cells. In order to identify further molecular targets to specifically target CHD1 deleted cancer cells we performed a siRNA screen looking for synthetic lethal interactions. This analysis identified genes, which are involved in DSB repair, replication and cell cycle checkpoint control further suggesting that the CHD1 deleted subclass of prostate cancer might benefit from DNA damaging therapy.
CHD1 deletion and response to DNA damaging therapy
To evaluate whether CHD1 deleted cancers would benefit from DNA damaging therapy we analyzed the CHD1 copy number status in a cohort of mCRPC patients, which have been treated with the PARP inhibitor olaparib (n=49) and correlated CHD1 status with response. We did not identify cases with homozygous CHD1 deletion in the first part of the trial. CHD1 heterozygous deletions and copy number gains were equally common in the responding and non-responding patients. We will expand this analysis in the next parts of the trial to determine whether homozygous loss of CHD1 sensitizes to olaparib. Interestingly, in preclinical studies using a CRPC patient-derived organoid (PDO) model with a homozygous loss of CHD1, we found an increased sensitivity to olaparib compared to PDOs with normal CHD1 copy number. In addition, we started to collect and process biopsies from men with mCRPC participating in an observational clinical study, which aims to evaluate the short and long term safety profile of Radium-223 (Reassure, NCT02141438). We will correlate CHD1 expression with response to DNA damaging Radium-223 treatment.
Conclusions
Our tissue-based analysis of clinically aggressive prostate cancer suggests that high CHD1 expression rather then CHD1 loss is associated with a more aggressive disease. Although our current results do not answer, whether CHD1 deleted tumors are more sensitive to DNA damaging therapy, preclinical in vitro analysis suggest a role for CHD1 in DSB repair, which when lost, sensitizes cancer cells to olaparib and DNA-PK inhibitor treatment.