2016 IMPaCT Research Vignette

Title: Hypermutation and Genomic Testing for Precision Therapy in Advanced Prostate Cancer

Investigator: Colin Pritchard, M.D., Ph.D., University of Washington

My research is about developing biomarkers in prostate cancer —genomic biomarkers that can predict therapy. I’m specifically interested in a class of genes called DNA repair genes.

I’m a physician, and I think that having, you know, both basic researchers and physicians involved in the research team is critical; but having people that—that have their hands in both worlds—is even better, I’d say. It’s also really critical, I think, to have different types of physicians who have different areas of expertise involved in the research team.

This project is focused on a sub-type of prostate cancer that has many mutations, so we call it the hypermutated sub-type. This was discovered not too long ago, about 5 years ago. I got interested in that, and I said, gosh, you know, there’s 10-times, 100-times more mutations in these xenograft lines compared to the other ones. What’s the mechanism? What’s causing that? And, not surprisingly, the mechanism looks like it’s going to be mismatch repair mutations and—and the reason I say that’s not surprising is because hypermutation in other cancer types—colorectal cancer, endometrial cancer, other cancers that are better studied in terms of hypermutation—that’s already a known mechanism.

So we did focus on mismatched repair deficiency as a mechanism early on and published a paper a couple years go in Nature Communications describing the mechanism of hypermutated prostate cancer, and it was mismatch repair deficiency; but, unlike those cancers where the dominant sporadic mechanism is epigenetic silencing of a gene called MLH1, in prostate cancer it was largely a structural rearrangement, so like complex genomic rearrangements in genes MSH2 and MSH6.

The other thing that was really interesting is that, in the autopsy series we looked at, we were looking at the University of Washington autopsy series in collaboration with Bob Vesella and Colm Morrissey, Eva Corey and others in the GU-Cancer Research Group at the University of Washington and, among that series, we were able to find 103 unique patients in that series, and almost 10 percent had hypermutations. So that was a lot higher than we expected and, honestly, I think that’s, in this series, probably going to be higher than in some other series because these are all advanced prostate cancers by definition. These are—these are men who have—have died from prostate cancer, so they’re the most advanced. So, at least in very advanced disease, we know the prevalence can be quite high in this series, up to 10 percent.

We found that all of the hypermutated cases have underlying mismatch repair gene mutations. And, of those, 9 out of 10 were MSH2 and MSH6. The 10th one was a MLH1, which is another mismatch repair gene, homozygous deletion.

This figure here shows the 10 unique patients and what their mutations were; and so, what you can see, is they were all hypermutated. They also had the phenotypic correlate to a mismatch repair deficiency, which is called micro-satellite instability, which I’m abbreviating MSI here. And that’s a hallmark of mismatch repair deficiency. It’s sort of a phenotypic correlate, a genomic signature, and micro-satellite instability is very, very well studied in colorectal cancer and endometrial cancer, especially in colorectal cancer, where it’s done clinically, and our lab happens to do this clinically too.

So it’s nice how the sort of our clinical work and our research really inform each other because I’m—I was quite familiar with micro-satellite instability from my clinical work, primarily in colorectal and endometrial cancer, but then able to apply that knowledge to prostate cancer and, really, you know learn—learn about what micro-satellite instability is like in prostate cancer.

The next thing we wanted to know was: could we develop clinical diagnostics to accurately detect hypermutation and—and micro-satellite instability, which is sort of the phenotypic correlate of that of mismatch repair deficiency in prostate cancer? And so we developed a method we call mSINGS, which just stands for MSI by next generation sequencing—MSI by NGS, mSINGS—so that’s what we call it. So we were among the first in the country to develop this bioinformatics technique—to do this in general from—from the next generation sequencing data. So we then validated that clinically on our assays, applied the mSINGS method to our prostate cancer samples, and then were able to assess micro-satellite instability in a more fine-tuned way like that.

Unlike traditional methods of micro-satellite instability—which look at like 5 loci in the genome —sometimes 10. With the mSINGS method, we can look it up to 3,000 loci, for example, with exome data or more. And so we get a much, much more granular, fine-tuned picture of the—of the sort of genomic landscape of micro-satellite instability specifically in prostate cancer.

But what we’re really excited about is the possibility that hypermutation and micro-satellite instability in prostate cancer will predict a more favorable outcome, particularly in a class of immunotherapy called checkpoint inhibitors. So those are PD-1 and PDL-1 inhibitors. We did have at least one patient who was able to get on a checkpoint immunotherapy drug and did have a PSA response, at least preliminarily, so you know this—this is [inaudible], and we don’t know yet if—if this is going to be predictive in prostate cancer or not, but I think it's a very exciting area to study.