Read the Full Video Transcript

Andrea Miyahira: Hi, everyone. I'm Andrea Miyahira at the Prostate Cancer Foundation. I'm pleased to welcome Dr. Laura Sena from Johns Hopkins University, who will discuss her recent paper, "Intratumoral heterogeneity drives acquired therapy resistance in a patient with metastatic prostate cancer," published in NPJ Precision Oncology. Dr. Sena, thank you for coming on today.

Laura Sena: Thanks so much for having me, Andrea. All right, so I'm really excited to present our paper with everybody today. So to introduce the paper, this is a diagram that's often used to summarize the course of prostate cancer in which you're looking at tumor volume over time. So patients with prostate cancer either present at this time point 0.1 with local disease or at this time point 0.2 with metastatic disease.

And for patients with localized prostate cancer, we treat them with either surgery or radiation with the intention to cure. However, some of these patients unfortunately will eventually relapse with metastatic disease. And so when this occurs, the medical oncologist has some good and bad news to discuss with the patient.

So the good news is that we have very highly effective systemic therapies for prostate cancer, and greater than 90% of patients will respond and enter a period of remission. But the bad news is that eventually, nearly all of these patients will relapse as their cancer learns to grow despite treatment. And this is called acquired therapy resistance. And it is arguably the greatest challenge that patients with metastatic prostate cancer and their medical oncologist face because these patients, in the absence of other competing health problems, will eventually die from their disease.

So in order to develop methods to combat acquired therapy resistance, we have to understand it better. So how does it occur? There's really two models to consider, which are not mutually exclusive. The first is called adaptation. And in this model, previously sensitive cancer cells change to become resistant to therapy. And the second model is called selection, in which a cancer cell in pink here, which was never sensitive to therapy, persists and then grows out to become the dominant population of cancer cells to drive the resistance.

So adaptation is a very well-accepted mechanism of resistance to treatment of prostate cancer, with the most infamous example being acquisition of androgen receptor amplifications or activating mutations, which drive resistance to hormone therapy. But there's much less clinical evidence that selection is a driver of resistance. So the value of this case is really that it provides proof of principle that selection can drive acquired therapy resistance in a patient.

So let's jump into the case. This patient initially presented with widely metastatic disease and a very bulky primary tumor and a PSA of greater than 6,000. He was first treated with leuprolide, which resulted in a marked decline in his PSA. And then he was treated with docetaxel and abiraterone without much of a PSA response.

At this point, the primary tumor sample underwent sequencing, and we determined that he had a very rare subtype of prostate cancer with mismatch repair deficiency that enables response to immunotherapy. So he was treated with nivolumab, and this resulted in a greater than 99% decrease in his PSA and a response that persisted over several months.

And when his PSA started to rise slightly, he was continued on nivolumab, and the AR inhibitor enzalutamide was added. And this controlled his disease for many years, greater than three years. But unfortunately, he developed resistance and afterwards was treated with cabazitaxel, bipolar androgen therapy, enzalutamide, nivolumab, and ipilimumab, none of which resulted in much of a response.

So at this point, we had really run out of most standard of care options for treatment. So he underwent a biopsy of a lymph node in his neck, a cervical lymph node. And what was very interesting was that this sample showed no evidence of mismatch repair deficiency. And in fact, it showed that it lacked BRCA2 expression.

So given this—and BRCA2 loss is known to predict response to PARP inhibition—so given this, we treated him with the PARP inhibitor olaparib, which resulted in a greater than 50% decline in his PSA, shrinkage of the tumors on his scans, and also an improvement in his physical function.

So this case raised the question really of what is the relationship between the cancer at diagnosis with mismatch repair deficiency and the cancer that was sampled greater than seven years later from a metastatic site with BRCA2 loss? So to assess this, we went back to the primary tumor and carefully assessed multiple sites by immunohistochemistry for the four mismatch repair proteins. And what we found was there were two distinct areas found within the primary tumor.

And this was from a TURP sample. And the first area showed very high-grade disease with a Gleason of 9 and showed evidence of MSH2 loss as well as loss of its binding partner, MSH6. But now we found this second area within the primary tumor that had a lower grade—so it's Gleason 7 disease—and showed intact expression of all four mismatch repair proteins. And this was similar to the expression pattern that we saw in the late metastatic tumor.

So in order to understand really the clonal evolution of this patient's cancer, we performed whole exome sequencing on both areas of the primary tumor, normal tissue, and paired up with Rachel Karchin's laboratory at Johns Hopkins, who has a lot of expertise with tumor evolution analysis. And so what they determined is that the loss of MSH2 and loss of BRCA2 was a very early branching event in the ancestry of this patient's cancer.

And interestingly, the lineage derived from MSH2 loss was not identified in the late metastatic samples—so both the cervical lymph node sample as well as the ctDNA sample obtained after progression on PARP inhibitor. And all of the subclones identified in the late metastatic disease appear to be related to the low-grade area with BRCA2 loss.

So the summary points for this case is that clonal dominance of prostate cancer can shift over time with selective pressure of therapy. Clonal heterogeneity can enable acquired therapy resistance due to outgrowth of a clone with primary resistance. And clones with genomic features that drive therapy resistance can ultimately threaten patient survival, even despite less aggressive histologic features at diagnosis. And repeat genomic evaluation of metastatic prostate cancer following development of acquired therapy resistance can change clinical decision-making.

So I'd like to acknowledge the co-first authors for this report, Dr. Dena Rhinehart, who's a medical oncology fellow here, and Dr. Jiaying Lai, who's a postdoc fellow in Rachel Karchin's lab; Rachel Karchin for her guidance on the clonal evolution analysis, as well as all of the co-authors on this paper, and our funding sources at NCI, and especially the Prostate Cancer Foundation, and Andrea. So thank you so much.

Andrea Miyahira: OK, thank you so much, Dr. Sena, for sharing this really interesting case there. So the MMR deficient clone—do you think it was cleared by immunotherapy? Or do you think it still remained at all?

Laura Sena: I think that's a great question. And basically, the question is, did we cure his mismatch repair deficient cancer? So, of course, we don't know because we're limited by the fact that we sampled only two areas of metastatic disease. But we didn't find any evidence of this clonal lineage of the mismatch repair deficient cancer in these two areas. We also know that immunotherapy can cure other types of metastatic cancers. So I think it's possible, and it's a really exciting aspect of this case.

Andrea Miyahira: Thank you. And how common do you think that primary tumor heterogeneity contributes to treatment resistance like this case? And if it's common, what would be your thoughts about whether a patient diagnosed with metastatic disease should have primary tumor treatment?

Laura Sena: So I think one thing that's clear is that prostate cancer is a very heterogeneous disease, so it's poised for this type of clonal selection by treatment. But it's hard to estimate how often it really is a driver of therapy resistance. So a highly unique aspect of this case was that his cancer initially was dominated by a clone with mismatch repair deficiency and then was dominated by a clone with homologous recombination repair deficiency.

And these just happen to be the two subtypes of prostate cancer for which we have targeted treatments, so we could really infer that those clones were dominant based on their response to the treatments. And so really, if we think about how to assess this shift in clonal dominance over time in other patients, we need to sample multiple tumor sites in order to really know which clone is dominant. And this is something I think that's really only feasible at autopsy, and then we're really limited to one time point of assessment. So one thing that's pretty exciting, though, is circulating tumor DNA analysis because it might be representative of multiple tumor sites and can be assessed over time.

Andrea Miyahira: Thank you. And do you think primary tumor heterogeneity will have different relevance based on the timing of diagnosis along the clinical history of disease?

Laura Sena: Yeah. So I think it's an interesting point to consider. So for example, clinical trials have shown us that for patients with low-volume metastatic disease, they have longer survival when the primary tumor is treated with radiation. And we don't fully understand why that is. But it's interesting to speculate that perhaps it's due to reduction of clonal diversity from the primary tumor in limiting potential for outgrowth of resistant clones in the future.

Andrea Miyahira: Thanks. And what is the takeaway message for clinicians and for researchers?

Laura Sena: So I think that the value of this case report—and really any case report—is just to show us what's possible. But the challenge is for us to understand really how does this apply to other patients. And so I think in this case, it was clear that repeat genomic evaluation was very helpful clinically. It changed clinical management.

Should we—I think that other clinicians should think about this for their patients. Cost effectiveness, I think, is the real barrier to just universally saying, with every chance for therapy resistance, we should sequence. But I think as cost of sequencing decreases over time, this might be more realistic.

And then I think the other takeaway for me is that cancer heterogeneity is very difficult to model in the lab, but it probably is a very important feature that we need to consider as we develop new strategies for treatment, especially if our goal is to cure metastatic cancer.

Andrea Miyahira: OK. Well, thank you so much for sharing this with us today.

Laura Sena: Thank you for having me.