Exploring the Role of ATM and BRCA2 Mutations in Metastatic Prostate Cancer Outcomes, Journal Club - Rashid Sayyid & Zachary Klaassen
June 23, 2023
Rashid Sayyid and Zach Klaassen discuss a publication examining metastatic prostate cancer outcomes in relation to BRCA2 and ATM mutations. Sayyid highlights that around 20-25% of metastatic prostate cancer patients have homologous recombination repair (HRR) gene mutations which are not uniform and can significantly affect treatment responses. He explains that FDA-approved PARP inhibitors (olaparib and rucaparib) have differential responses across HRR gene mutations, indicating the need for further research to better define subtypes driven by ATM and BRCA2 mutations. Klaassen discusses the results, including the identification of molecular subgroups within metastatic prostate cancer, which can predict treatment responses and strategies. He also emphasizes the need for targeted therapies against ATM or BRCA2 mutations, potentially including PARP inhibitors or other therapies specific to ATM-mutated tumors.
Biographies:
Rashid Sayyid, MD, MSc, Urologic Oncology Fellow, Division of Urology, University of Toronto, Toronto, Ontario
Zachary Klaassen, MD, MSc, Urologic Oncologist, Assistant Professor Surgery/Urology at the Medical College of Georgia at Augusta University, Georgia Cancer Center
Biographies:
Rashid Sayyid, MD, MSc, Urologic Oncology Fellow, Division of Urology, University of Toronto, Toronto, Ontario
Zachary Klaassen, MD, MSc, Urologic Oncologist, Assistant Professor Surgery/Urology at the Medical College of Georgia at Augusta University, Georgia Cancer Center
Read the Full Video Transcript
Rashid Sayyid: Hello everyone, this is Rashid Sayyid, I'm a urologic oncology fellow at the University of Toronto, and along with Zach Klaassen, associate professor and program director at the Medical College of Georgia. We'll be going over the recent publication in Clinical Cancer Research looking at Metastatic Prostate Cancer Outcomes with BRCA2 versus ATM Mutations, which Exhibit Divergent Molecular Features and Clinical Outcomes.
We know that homologous recombination repair, or HRR gene mutations, occur in about 20-25% of metastatic prostate cancer patients. These genes, they regulate critical DNA repair processes which are essential for tumor cell survival, and now they're reported genomic tests to, first, help risk stratify patients, then, importantly, inform clinical decisions for metastatic prostate cancer patients. It's important to highlight that HRR gene mutations are not uniform, so distinct mutations such as ATM or BRCA2 may govern distinct clinical responses in metastatic prostate cancer patients.
Currently, we have two PARP inhibitors, olaparib and rucaparib, which are FDA approved for metastatic prostate cancer patients with HRR gene mutations. But it's important to highlight that we have differential responses to PARP inhibitors that are currently observed across the HRR gene mutations. So if we take the example of PROfound, which looked at olaparib versus abiraterone or ENZA in mCRPC patients with progression on prior ABI or ENZA, if we look at the forest plot below, we see that we have differential responses based on the HRR gene mutation. We see a much better response with BRCA2, and to a lesser extent BRCA1. On the other hand, we don't see much of our response with the ATM, although all three mutations are HRR gene alterations.
Similarly, the TRITON2 trial, which looked at rucaparib in mCRPC patient with one or two prior lines of ARSIs and one taxane for mCRPC, if we look at the biochemical response with PSA, we see that the response was much higher in the BRCA1/2-mutated cohort as opposed to ATM. Further evidence in this space comes from preclinical models, which employ gene depletion methods to induce loss of BRCA2 or ATM in prostate cancer cell lines. And this demonstrated that reduction in BRCA2 expression increases the PARP inhibitor response, whereas ATM reduction does not, which is similar to what we saw with the PROfound and TRITON2 trial.
What about chemotherapy? Again, if we look at taxanes, we see a superior response in the ATM-mutated cohort. Conversely, when we look at platinum-based chemotherapy, a better response is seen in the BRCA2 mutated patients. As such, it becomes clear that this is quite a heterogeneous patient population and we need to better define the metastatic prostate cancer subtypes that are driven by ATM and BRCA2 mutations.
To this end, the Caris Precision Oncology Alliance Consortium aggregated a cohort of 1,827 microsatellite stable metastatic prostate cancer collected from metastatic tissue sites from the bones, lymph nodes, and liver. The Caris Life Sciences database was queried to assess the molecular alterations, gene expression, and survival outcomes of ATM, 88 patients, or BRCA2-mutated metastatic prostate cancer patients. The Caris Life Sciences database includes factors that are available in insurance claims such as gender, age of diagnosis, therapeutic agents that were given with the dates of administration, the time of last contact, and the tissue site from where the biopsy sample was obtained.
The authors included four different sub cohorts. They included the BRCA2-mutated cohort, which was 98 patients, the ATM-mutated, which was 88 patients, they included an HRR-deficient cohort, which is 193, and this HRR deficient cohort included patients with HRR gene mutations other than BRCA2 or ATM. The list of the mutated genes is summarized here. They also include a control group of homologous recombination proficient patients, meaning those were the patients that did not have any HRR gene mutations. For these patients, the investigators perform whole exome sequencing, targeted next-generation sequencing, and whole tissue sampling for all biopsy samples.
Whole exome sequencing and next generation sequencing was performed on genomic DNA that was isolated from FFPE tumor samples after microdissection with the nextSequence platform. The tumor mutational burden was derived by counting all non-synonymous missense, meaning these are the mutations that cause an alteration in the amino acid, nonsense, in-frame insertion deletion, and frameshift mutation that were not previously described as germline alterations. They used a cutoff of at least 10 mutations used in this setting, and the reason they used this cutoff was that tumor mutational burden of at least 10 mutations is known to have a higher response rate compared to patients with a tumor mutational burden of less than 10. The gene profiles for ATM or BRCA2-mutated tumors were generated by considering the differential expression values based on Limma analyses, and a differential gene expression was considered significant on the basis of an adjusted p-value of less than 0.01.
Overall survival outcomes were obtained from insurance claims database and patient death was assumed for any patient without a claim for at least 10 days. This algorithm has been shown to have 95% validity based on validation studies from the National Death Index. Survival was considered from the time of biopsy till death or last contact, and Cox proportional hazards ratios were calculated for each comparison group. Finally, statistical significance was evaluated using Fisher's-Exact, Mann Whitney, and the Chi-squared test with Benjamini-Hochberg correction to adjust or account for inflated type one errors with multiple comparisons.
At this point, I'll turn it over to Zach to go over the results and discussion of this paper.
Zach Klaassen: Rashid, thanks so much for that great introduction. This table looks at the patient characteristics of metastatic prostate cancer and the prevalence of mutations, as well as including the biopsy site. If you look at the top of this table, this has a broken down by ATM mutation, BRCA2, HRD other, as well as the homologous recombination proficient patients, as Rashid described. So, total cases, 88 ATM mutation representing four 7.41% of the population, BRCA2 98 cases, 8.26%, HRD other 193 cases, 16.26%, and HRR-proficient, 808, essentially a control group, 68% of this population.
Looking at the bottom, and we'll look at this again on the next slide in a graphic form, we see that the lymph node biopsy was the site for 389 patients, second most common was bone at 279, liver at 169. So for those that are a little more attuned to looking at this from a graphical representation, we see that ATM is at the bottom, BRCA2, followed by HRDother, HRR proficient specimens. Again, most commonly the site for tissue was lymph node in roughly 35-38% of these patients. Second most common was bone, followed by liver, and you can see the other less common sites of tissue procurement.
The next several slides will look similar to this one, and this is basically Kaplan-Meier for clinical outcomes on the top. And then essentially forest plots at the bottom, which are correlated with the Kaplan-Meier curves. And so this is clinical outcomes of ATM, BRCA2, and HRDother versus HRR proficient for non-castrate metastatic prostate cancer patients after ADT. And so you can see, looking at ATM versus HRR proficient, absolutely no difference in outcomes, and we see in the HRDother versus HRR proficient, we see a splitting of the curves here, not quite statistically significant. We can see a p-value of 0.079. But we can see here BRCA2 in blue versus HRR proficient, for those patients that are non-castrate after ADT, these BRCA2-mutated patients do statistically significantly worse than the HRR proficient patients.
This is the same looking slide looking at castrate metastatic prostate cancer patients after NHTs, and when we look at these Kaplan-Meier curves and the corresponding plot at the bottom, there is statistical significance for all of these. So ATM versus HRR proficient did statistically worse in the castrate setting after NHT. BRCA2 versus HRR proficient, again, statistically significant worse, p-value of 0.017, early and consistent splitting of the curves. And finally, for HRDother versus HRR proficient, again, statistically significant worse outcomes compared to those that are HRR proficient in the castrate setting after NHTs.
The next couple slides will look at taxane therapy, and this is the plot for non-castrate metastatic prostate cancer patients. We can see here for ATM and BRCA2 versus HRR proficient, no difference whatsoever. We do see a slight splitting of the curves for HRDother versus HRR proficient. However, when we look specifically in a little bit more detail, as you can see in the bottom right, when we look at BRCA2 mutated versus HRDother, we see that the BRCA2-mutated patients do statistically significantly better than the HRD other patients in the non-castrate setting after taxane therapy.
This looks at the taxane therapy in the castrate metastatic prostate cancer setting, and we do see splitting of these curves. None of these combinations of ATM, BRCA2, HRD other versus HRR proficient are statistically significant, but we do see a trend towards worse outcomes for these mutated patients compared to the HRR proficient patients.
Let's look a little bit closer to the genomic landscape of ATM and BRCA2-mutated patients, and so this is sort of broken down into five parts of this figure. On the top left, this is looking at metastatic prostate cancer patients with BRCA2 and HRDother genes having increased genomic instability scores compared to HRR proficient patients, whereas ATM patients have less genomic instability scores compared to these HRR proficient patients. At the top, looking at part B, this is looking at AR activating mutations. A couple of points here. There was no difference in AR activating mutations in the BRCA2 or ATM-mutant metastatic prostate cancer. And secondly, HRDother tumors had depletion of the TMPRSS2 fusion, which you can see here in the green where the arrow was located.
Bottom left is tumor suppressor mutations. ATM-mutated tumors had lower rates of TP53 alterations and low rates of RB1 or PTEN mutations versus BRCA2 and HRR proficient mutated prostate cancer. When we look at part D, which is copy number alterations, we find that ATM-mutated tumors had significant co-application with genes on chromosome 11q13. Finally, in the bottom right, this is looking at immune-related markers. Metastatic prostate cancer with BRCA2 had greater tumor mutation burden, high status versus HRR proficient patients.
This next slide looks at transcriptome analysis of ATM and BRCA2-mutated metastatic prostate cancer. Part C and D is looking at ATM and part E and F is looking at BRCA2. The summary for C and D is that ATM-mutated tumors had decreased levels of HES4 and SNHG6, whereas, on the right side of the screen, looking at BRCA2, these mutated tumors had robust correlations with increased genes that regulate the cell cycle, which includes E2F7 and CBX2.
By way of discussion relative to prostate cancer samples that lacked HRR mutations, mutations in ATM or BRCA2 were associated with trends towards worse clinical outcomes with taxane chemotherapy as well as inferior outcomes to ADT or NHTs. These findings support the need for investigating targeted therapies against ATM or BRCA2 mutations, and this includes, as we know, PARP inhibitors, but also potentially other targeted therapies specific to the ATM mutated tumors. These exhibit significantly less features of DNA damage with lower genomic instability scores, fewer TP53 mutations, and lower tumor mutation burden. Thus, it is unlikely that additional treatment targeting the functional loss of DNA repair will be beneficial.
Specific to the BRCA2 mutated tumors, these tumors exhibit genomic instability with significantly elevated scores, which is a potential marker for PARP inhibition responsiveness. Finally, these tumors also have higher tumor mutation burden, which may potentially exhibit the greatest immunotherapy response.
In conclusion, this work establishes differential genotypes and phenotypes of metastatic prostate cancer with tumoral mutations in either BRCA2 or ATM. These findings indicate that these two common mutational events can be used to stratify metastatic prostate cancer into molecular subgroups to predict future treatment strategies or response to standard of care treatments. Finally, while further mechanistic studies are required, this analysis identify specific forms of cell signal events in metastatic prostate cancer.
We thank you very much for your attention. We hope you enjoyed this UroToday Journal Club discussion.
Rashid Sayyid: Hello everyone, this is Rashid Sayyid, I'm a urologic oncology fellow at the University of Toronto, and along with Zach Klaassen, associate professor and program director at the Medical College of Georgia. We'll be going over the recent publication in Clinical Cancer Research looking at Metastatic Prostate Cancer Outcomes with BRCA2 versus ATM Mutations, which Exhibit Divergent Molecular Features and Clinical Outcomes.
We know that homologous recombination repair, or HRR gene mutations, occur in about 20-25% of metastatic prostate cancer patients. These genes, they regulate critical DNA repair processes which are essential for tumor cell survival, and now they're reported genomic tests to, first, help risk stratify patients, then, importantly, inform clinical decisions for metastatic prostate cancer patients. It's important to highlight that HRR gene mutations are not uniform, so distinct mutations such as ATM or BRCA2 may govern distinct clinical responses in metastatic prostate cancer patients.
Currently, we have two PARP inhibitors, olaparib and rucaparib, which are FDA approved for metastatic prostate cancer patients with HRR gene mutations. But it's important to highlight that we have differential responses to PARP inhibitors that are currently observed across the HRR gene mutations. So if we take the example of PROfound, which looked at olaparib versus abiraterone or ENZA in mCRPC patients with progression on prior ABI or ENZA, if we look at the forest plot below, we see that we have differential responses based on the HRR gene mutation. We see a much better response with BRCA2, and to a lesser extent BRCA1. On the other hand, we don't see much of our response with the ATM, although all three mutations are HRR gene alterations.
Similarly, the TRITON2 trial, which looked at rucaparib in mCRPC patient with one or two prior lines of ARSIs and one taxane for mCRPC, if we look at the biochemical response with PSA, we see that the response was much higher in the BRCA1/2-mutated cohort as opposed to ATM. Further evidence in this space comes from preclinical models, which employ gene depletion methods to induce loss of BRCA2 or ATM in prostate cancer cell lines. And this demonstrated that reduction in BRCA2 expression increases the PARP inhibitor response, whereas ATM reduction does not, which is similar to what we saw with the PROfound and TRITON2 trial.
What about chemotherapy? Again, if we look at taxanes, we see a superior response in the ATM-mutated cohort. Conversely, when we look at platinum-based chemotherapy, a better response is seen in the BRCA2 mutated patients. As such, it becomes clear that this is quite a heterogeneous patient population and we need to better define the metastatic prostate cancer subtypes that are driven by ATM and BRCA2 mutations.
To this end, the Caris Precision Oncology Alliance Consortium aggregated a cohort of 1,827 microsatellite stable metastatic prostate cancer collected from metastatic tissue sites from the bones, lymph nodes, and liver. The Caris Life Sciences database was queried to assess the molecular alterations, gene expression, and survival outcomes of ATM, 88 patients, or BRCA2-mutated metastatic prostate cancer patients. The Caris Life Sciences database includes factors that are available in insurance claims such as gender, age of diagnosis, therapeutic agents that were given with the dates of administration, the time of last contact, and the tissue site from where the biopsy sample was obtained.
The authors included four different sub cohorts. They included the BRCA2-mutated cohort, which was 98 patients, the ATM-mutated, which was 88 patients, they included an HRR-deficient cohort, which is 193, and this HRR deficient cohort included patients with HRR gene mutations other than BRCA2 or ATM. The list of the mutated genes is summarized here. They also include a control group of homologous recombination proficient patients, meaning those were the patients that did not have any HRR gene mutations. For these patients, the investigators perform whole exome sequencing, targeted next-generation sequencing, and whole tissue sampling for all biopsy samples.
Whole exome sequencing and next generation sequencing was performed on genomic DNA that was isolated from FFPE tumor samples after microdissection with the nextSequence platform. The tumor mutational burden was derived by counting all non-synonymous missense, meaning these are the mutations that cause an alteration in the amino acid, nonsense, in-frame insertion deletion, and frameshift mutation that were not previously described as germline alterations. They used a cutoff of at least 10 mutations used in this setting, and the reason they used this cutoff was that tumor mutational burden of at least 10 mutations is known to have a higher response rate compared to patients with a tumor mutational burden of less than 10. The gene profiles for ATM or BRCA2-mutated tumors were generated by considering the differential expression values based on Limma analyses, and a differential gene expression was considered significant on the basis of an adjusted p-value of less than 0.01.
Overall survival outcomes were obtained from insurance claims database and patient death was assumed for any patient without a claim for at least 10 days. This algorithm has been shown to have 95% validity based on validation studies from the National Death Index. Survival was considered from the time of biopsy till death or last contact, and Cox proportional hazards ratios were calculated for each comparison group. Finally, statistical significance was evaluated using Fisher's-Exact, Mann Whitney, and the Chi-squared test with Benjamini-Hochberg correction to adjust or account for inflated type one errors with multiple comparisons.
At this point, I'll turn it over to Zach to go over the results and discussion of this paper.
Zach Klaassen: Rashid, thanks so much for that great introduction. This table looks at the patient characteristics of metastatic prostate cancer and the prevalence of mutations, as well as including the biopsy site. If you look at the top of this table, this has a broken down by ATM mutation, BRCA2, HRD other, as well as the homologous recombination proficient patients, as Rashid described. So, total cases, 88 ATM mutation representing four 7.41% of the population, BRCA2 98 cases, 8.26%, HRD other 193 cases, 16.26%, and HRR-proficient, 808, essentially a control group, 68% of this population.
Looking at the bottom, and we'll look at this again on the next slide in a graphic form, we see that the lymph node biopsy was the site for 389 patients, second most common was bone at 279, liver at 169. So for those that are a little more attuned to looking at this from a graphical representation, we see that ATM is at the bottom, BRCA2, followed by HRDother, HRR proficient specimens. Again, most commonly the site for tissue was lymph node in roughly 35-38% of these patients. Second most common was bone, followed by liver, and you can see the other less common sites of tissue procurement.
The next several slides will look similar to this one, and this is basically Kaplan-Meier for clinical outcomes on the top. And then essentially forest plots at the bottom, which are correlated with the Kaplan-Meier curves. And so this is clinical outcomes of ATM, BRCA2, and HRDother versus HRR proficient for non-castrate metastatic prostate cancer patients after ADT. And so you can see, looking at ATM versus HRR proficient, absolutely no difference in outcomes, and we see in the HRDother versus HRR proficient, we see a splitting of the curves here, not quite statistically significant. We can see a p-value of 0.079. But we can see here BRCA2 in blue versus HRR proficient, for those patients that are non-castrate after ADT, these BRCA2-mutated patients do statistically significantly worse than the HRR proficient patients.
This is the same looking slide looking at castrate metastatic prostate cancer patients after NHTs, and when we look at these Kaplan-Meier curves and the corresponding plot at the bottom, there is statistical significance for all of these. So ATM versus HRR proficient did statistically worse in the castrate setting after NHT. BRCA2 versus HRR proficient, again, statistically significant worse, p-value of 0.017, early and consistent splitting of the curves. And finally, for HRDother versus HRR proficient, again, statistically significant worse outcomes compared to those that are HRR proficient in the castrate setting after NHTs.
The next couple slides will look at taxane therapy, and this is the plot for non-castrate metastatic prostate cancer patients. We can see here for ATM and BRCA2 versus HRR proficient, no difference whatsoever. We do see a slight splitting of the curves for HRDother versus HRR proficient. However, when we look specifically in a little bit more detail, as you can see in the bottom right, when we look at BRCA2 mutated versus HRDother, we see that the BRCA2-mutated patients do statistically significantly better than the HRD other patients in the non-castrate setting after taxane therapy.
This looks at the taxane therapy in the castrate metastatic prostate cancer setting, and we do see splitting of these curves. None of these combinations of ATM, BRCA2, HRD other versus HRR proficient are statistically significant, but we do see a trend towards worse outcomes for these mutated patients compared to the HRR proficient patients.
Let's look a little bit closer to the genomic landscape of ATM and BRCA2-mutated patients, and so this is sort of broken down into five parts of this figure. On the top left, this is looking at metastatic prostate cancer patients with BRCA2 and HRDother genes having increased genomic instability scores compared to HRR proficient patients, whereas ATM patients have less genomic instability scores compared to these HRR proficient patients. At the top, looking at part B, this is looking at AR activating mutations. A couple of points here. There was no difference in AR activating mutations in the BRCA2 or ATM-mutant metastatic prostate cancer. And secondly, HRDother tumors had depletion of the TMPRSS2 fusion, which you can see here in the green where the arrow was located.
Bottom left is tumor suppressor mutations. ATM-mutated tumors had lower rates of TP53 alterations and low rates of RB1 or PTEN mutations versus BRCA2 and HRR proficient mutated prostate cancer. When we look at part D, which is copy number alterations, we find that ATM-mutated tumors had significant co-application with genes on chromosome 11q13. Finally, in the bottom right, this is looking at immune-related markers. Metastatic prostate cancer with BRCA2 had greater tumor mutation burden, high status versus HRR proficient patients.
This next slide looks at transcriptome analysis of ATM and BRCA2-mutated metastatic prostate cancer. Part C and D is looking at ATM and part E and F is looking at BRCA2. The summary for C and D is that ATM-mutated tumors had decreased levels of HES4 and SNHG6, whereas, on the right side of the screen, looking at BRCA2, these mutated tumors had robust correlations with increased genes that regulate the cell cycle, which includes E2F7 and CBX2.
By way of discussion relative to prostate cancer samples that lacked HRR mutations, mutations in ATM or BRCA2 were associated with trends towards worse clinical outcomes with taxane chemotherapy as well as inferior outcomes to ADT or NHTs. These findings support the need for investigating targeted therapies against ATM or BRCA2 mutations, and this includes, as we know, PARP inhibitors, but also potentially other targeted therapies specific to the ATM mutated tumors. These exhibit significantly less features of DNA damage with lower genomic instability scores, fewer TP53 mutations, and lower tumor mutation burden. Thus, it is unlikely that additional treatment targeting the functional loss of DNA repair will be beneficial.
Specific to the BRCA2 mutated tumors, these tumors exhibit genomic instability with significantly elevated scores, which is a potential marker for PARP inhibition responsiveness. Finally, these tumors also have higher tumor mutation burden, which may potentially exhibit the greatest immunotherapy response.
In conclusion, this work establishes differential genotypes and phenotypes of metastatic prostate cancer with tumoral mutations in either BRCA2 or ATM. These findings indicate that these two common mutational events can be used to stratify metastatic prostate cancer into molecular subgroups to predict future treatment strategies or response to standard of care treatments. Finally, while further mechanistic studies are required, this analysis identify specific forms of cell signal events in metastatic prostate cancer.
We thank you very much for your attention. We hope you enjoyed this UroToday Journal Club discussion.