Andrea Miyahira: Hi, everyone. I'm Andrea Miyahira at the Prostate Cancer Foundation. I'm excited to be joined by Dr. Bill Isaacs of Johns Hopkins University and Dr. Jianfeng Xu of University of Chicago. They will discuss their recent paper, Discovery of Recurrent Frameshift Ashkenazi Jewish Founder Mutation in the PARP Inhibitor-Sensitive MMS22L Gene Associated with Higher Risk of Prostate Cancer. And this was published recently in European Urology Focus. Dr. Isaacs and Dr. Xu, thanks for joining us today.

Jianfeng Xu: Thank you.

William Isaacs: Yeah, thank you for having us. First of all, I just wanted to thank PCF for their support, continued support, and we were very fortunate to be supported by PCF over the years. So we're going to talk a little bit about some work that Jianfeng and I are very excited about. So we have been interested in looking for inherited aspects of prostate cancer and Dr. Patrick Walsh was very instrumental in inspiring us to work on this and providing support and collecting families over the years and as well as radical prostatectomy patients. And so we have a very nice collection of samples to work on. And so we've been very intrigued by when we first started, there were essentially no genes, high penetrance, low penetrance, anything known to increase risk for men in an inherited fashion for prostate cancer. And so we were very interested. And finally, I think we can just summarize and say that over the years there's been tremendous progress made.

And one of the, I think, key hallmarks of prostate cancer genetics is the importance of genes which were involved in DNA repair. Like almost all other solid tumors, things like BRCA2 and ATM, MSH2, et cetera, are really key. Then we think those are probably the most important single monogenic genes for inheritance of prostate cancer. And so we've been on the lookout for genes like that. We don't think there are very many that have the abilities of, for example, BRCA2 to increase risk and particularly for more aggressive disease, but those are the ones we want to make sure we find, if they're out there. And so that's where we think this reason why we're excited about MMS22L because we think it has some of the hallmarks of BRCA2 for risk of prostate cancer. So the study design was very simple to compare cases versus controls for gene sequences collected by whole exome sequencing.

And there was one gene that sort of stuck out and we didn't really know what to do with it. It was very rare. We only see it a couple of times. MMS22L had a loss of function mutation in it, but we only had about three examples of it. But we kept it in mind and brought it along with our other studies. And it turns out that when we went to look at this, it was not present in the random population. So we have men from non-Finnish European, African-American, and Ashkenazi-Jewish descent. So we looked at those three different populations. And the interesting thing was that it was only in this mutation and MMS22L was only in men of Ashkenazi descent. So that really knocked down our denominator of cases because we were looking at a denominator of 6,000 or 7,000 total prostate cancers, but only 400 or so of those were men of Ashkenazi descent.

And so when we looked at those, we actually found six guys who had the ... Which again, not a big number, but fortunately we could use gnomAD, which has many Ashkenazi men who we could use as controls, men not to have prostate cancer. And so the number, even though low, that we had in our population was higher and significantly higher than what we could see in the gnomAD study. And so we tried to accumulate as many Ashkenazi prostate cancers as possible and we basically were able to double that number to 11 or so cases of Ashkenazi men with prostate cancer. And this ended up being about one point a half percent, whereas the control data, which is thousands of Ashkenazi men, was around 0.3%. So almost a fivefold increase in risk with this one mutation, which we think is a founder mutation that's on the same haplotype in all these men.

So we think it's an interesting discovery of a founder mutation in a ... And I should point out that this MMS, not a whole lot is known about MMS22L, but what is known is that it's critical in homologous recombination DNA repair. And it interacts with a gene called TONSL, which it binds to when activated by double strand breaks. And you need this to undergo the homologous recombination repair of double strand breaks. So it's a clear DNA repair molecule or gene just like most of the more important genes for prostate cancer genetics. Basically, that is what the study comes down to. And the only other I think point that's really important to make here or a couple points, so at the same time, so I was scanning the literature, Jianfeng and I, looking for anything about MMS22L and there was essentially nothing in cancer genetics other than just some functional studies.

But then we saw a paper, actually this is a synthetic lethality paper. Yeah, here was this paper from Tsujino and Alan Kibel's group and I guess the GS group. I think Adam might've been a PCF fellow with me. I know he was a postdoc in my lab. It was fun to see this. So anyway, so they worked on, did a CRISPR screen to knock out genes which may increase sensitivity to PARP inhibition in a synthetic lethality method or phenomenon. And so they identified these 65 genes. And so sure enough, MMS22L was on this list. And not only was it on the list, but it was practically the most sensitive gene. So if you knock out this gene, the cells are incredibly sensitive to PARP inhibition. So we thought that, wow, this could be an important lead for some therapeutics. So anyway, so that was the other driving force behind this, was to look at these genes. And even as a general rule and just to look and see how many genes do actually have germline loss of function mutations that can potentially drive a synthetic lethality in a therapeutic way.

So anyway, so that's what this study is. So the bottom line is that here are the numbers. So you can see they're small, but they're highly significant. So in our discovery set, had about 1.3% versus 0.3% or about a four-fold increase in risk. And the validation, it was a little higher on the same sort of numbers. And so a combined total of about 1.5% in the cases versus 0.3% in the controls. So you get a nice significant P-value. And probably the most important thing with it really caught our eye was that with only several exceptions and mainly because we had missing data. All these guys with this mutation had clinically significant disease and many had a metastatic disease at the time of diagnosis. And so they were not localized low-grade, low-risk prostate cancers. They were just the opposite. So we thought this, again, sort of was like a BRCA2 mechanism that had the ability to reassociate with the more aggressive phenotype.

And so the only other point probably to raise here is that so this mutation is only found in Ashkenazi. At least that's our experience. There is another mutation that is rare. It is found in non-Finnish Europeans. And we do see in our population it's a C coding sequence 340G>A, so it abolishes a splice site. So it's a deleterious mutation. And in the non-Finnish European, tremendous numbers that are in the gnomAD, we do see an association, a significant association with men who carry that and prostate cancer risk, but that's really the only other mutation that we see. So I will stop there and allow time for questions. I'm sure there are many things I left out. I don't know. Jianfeng, do you spot anything that I should mention here?

Jianfeng Xu: Yeah, so I think great introduction and history about PCF support as well as your history of research. So I do want to emphasize two unique aspects of this study and this finding. In our previous study to identify gene for prostate cancer is using candidate gene, like DNA repair gene or brute force looked in entire genome. I think both actually are powerful, allow us to identify DNA repair gene BRCA2 to ATM. And the brute force allow us to identify how to be searching, which discovered by Dale as a group, but I think those two approaches reach a limit. And really I do not think that we can continue to find more gene for prostate cancer using those two approach. So this time we use two kinds of novel approaches. One is considering the clinical aspect of many patients respond well to PARP inhibitor. So try to identify patients with the mutation in PARP inhibitor sensitive gene.

Probably not only lead to the discovery of gene, but also more clinical meaningful, so that's number one. Number two is really from a technical aspect. I say brute force and we compare thousands cases, 100,000 controls, but we were able to identify anything like the mutation or human mutation in that gene over 1%, but those genes are very rare in the genome. So in order to find those kind of 1%, you have to go to a specific population. We call the founder population because founder population, so because the founder population is kind of elevate the proportion of certain mutation in that particular population. So the reason that we found this thing is because this mutation, even though it's rare in the European or in the general population, less than point some percent, but it is one point some percent in Ashkenazi positive cancer cases.

This is the reason that we are able to zoom in to indicate this thing versus everything else. So those two things actually are quite unique. And the second approach actually are being used by many other group these days and they go to founder mutation like Icelandic founder mutation and the Northern European founder mutation or founder population to identify. So it's a trend because we just cannot use brute force things anymore. I mean, anything that are important or common enough have been identified. We have to use some novel approach. So that's, I would say, a really the key reason or impact of this particular finding.

Andrea Miyahira: Thank you so much, Dr. Isaacs and Dr. Xu, for sharing this information with us. So you talked a little bit about the normal function of MMS22L. Does it have any known role in prostate or other cancers or do we know if it interacts with any other prostate cancer drivers?

William Isaacs: I can answer that. There's a lot known about the interactions that MMS22L occurs during in its role in homologous recombination repair. It interacts with a gene product called TONSL. And actually we do see some mutations in TONSL. We haven't really followed them up very well. They're very rare and to the point where we can't really make a story yet with that, but they may certainly that may be involved. And so it does. It's certainly not known as any sort of prostate cancer driver. Actually, there may be one or two papers out there that even look at that in prostate. So again, it is seen in prostate, it's expressed in prostate, this TONSL gene. And actually, there was a very interesting paper that I forgot to mention. But anyway, so there is some data out there that's of interest, but in general we don't think it's offering, although we're hopeful this will change, mechanistic insight into carcinogenesis for prostate cancer specific way.

Andrea Miyahira: So I think you investigated the other genes that were identified that caused sensitivity to PARP inhibitors from that panel of 65. Did you find any alterations in any of those genes in prostate cancer patients?

Jianfeng Xu: We actually looked all 65 genes that implicated in that paper and we looked at general mutation in those prostate cancer patients compared to control. So the finding is, number one, most mutations, including MMS22L, are rare. Number two is when you compare case and control, there are three genes actually have some evidence of association, including MMS22L. But if you look at them when you go deeper into specific population, then we see it's much stronger in Ashkenazi Jewish. So the answer to your question, yeah, there are some mutations, but they are rare at this stage. We really do not have evidence that is power to detect the difference. However, this mutation in MMS22L among Ashkenazi Jews actually stood out and that we have a status of power to detect that.

Andrea Miyahira: Okay, thank you. And have any patients with prostate cancer with the MMS22L received PARP inhibitors? And if so, do we know how they responded?

William Isaacs: Yeah, so fortunately we have some investigators, some colleagues that we have at Hopkins and at North Shore that are interested in following this up, particularly the therapeutic at and that we're very excited that they are interested in this. And unfortunately, I don't think there is very much data. Again, there's not that many patients that have been determined to have this mutation, so we're very excited about people, first of all, confirming this finding in the population and then see what that means for treatment. One thing that we are aware of is the cases at Hopkins, Dr. Jun Luo and a postdoc worker with him, Mayuko Yakanayama, have been looking at this. And so I think they actually have some very interesting findings that they are working on for submission now.

So I don't want to give anything away, but I think that there's definitely some interesting and actually maybe some opposite activity that ... So what we don't have is really any specific PARP inhibitor treated patients that have been followed up, but it does appear that some of the more conventional treatments, androgen deprivation that may be helpful. So anyway, we're excited to hear about what's going on there. So there is activity there and I think we should hear something about that quite soon.

Andrea Miyahira: Okay, thank you. And how much of prostate cancer heritability can now be accounted for with known risk genes and other tools such as PRS scores?

Jianfeng Xu: We all know prostate cancer has the highest heritability among all cancer, except now melanoma skin cancer. So it's a 58% of our risk to prostate cancer due to inherited genetic balance. So among inherited genetic balance, we now know about 50% of variants, those genetic thing, can be accounted for by either polygenic risk score or by known monogenic gene like HOXB13, BRCA2, ATM. But when I lump those things together, the reality is that those genes like BRCA2, HOXB13, even though they are quite important, but they are rare. So they do not contribute to a lot of heritability in the population. On the other hand, polygenic risk score does. So 44%, actually we call the heritability of familiar aggregation, can be accounted by those 451 genes that across the genome that we measure those using polygenic risk score versus 7% actually can be accounted for by those monogenic gene like BRCA2, ATM.

But still, I think we have much, much better position right now to understand the genetic basis. And so we can use this information to understand why you develop prostate cancer and how we actually assessing the risk of individual men in the population who are more likely to develop prostate cancer. So especially we can use this information to counsel men with prostate cancer for their family members. It is very powerful right now because we know so much compared to just 10 years ago.

Andrea Miyahira: Okay, thank you. And what are your next steps in these studies, including plans to apply these findings in the clinic?

Jianfeng Xu: Yeah, so Bill probably already mentioned that actually there's a follow up of clinical presentation and also in term of how those carriers respond to PARP inhibitor, so that's very exciting. Paper is coming. I think also particularly Hopkins group is doing additional functional study, try to better understanding the gene, the mutation in relationship with prostate cancer. But on my side, in Devon North Shore or University of Chicago side, we focus more on clinical translation of the finding. And so we already developed genetic testing that include this MMS22L in the panel. And our urologist already use this in information. For example, he identify a carrier in patient with active surveillance right now, but based on our understanding and those patients more likely have aggressive disease. So he used this information together with other clinical information and actually operate this patient two weeks ago. So that is a very clear action taken by implement the genetic testing, including this gene in the clinical setting.

And also, I think, like I say, we focus more on the translation. So Bill and us have a new paper published which systematically examine all the monogenic genes that actually recommended by NCCN guidelines for germline testing. We provide evidence which gene actually is truly, with status evidence, associative with risk, associated with progression, and associated with response to PARP inhibitor. We believe this is very helpful for urologists across the country to take the test and to interpret the result. And finally, I think that's something is related to this, but beyond this thing, since you are asking what's the next question, we're very excited about how we use genetic information, particularly germline information in the prostate cancer survivorship, right? So this paper, the ProtecT paper in the New England Journal, last two years, almost two years ago. So the reality, if you diagnose with the localized prostate cancer, after 15 years follow up 3% people patient died of prostate cancer, regardless which PARP, active surveillance surgery and the radiation.

However, the overall is about 30% of those patients die. So only 3% died of prostate cancer. 27% of people died of other disease, including other cancer, including cardiovascular disease, including the deaths due to fracture osteoporosis. So what we are trying to do right now, actually it's active research here, we focus on not just treating cancer, which everyone is doing particularly in this field, doing very well, but we focus on treating a patient as a whole because very few actually died of prostate cancer. They died of other disease. Our approach is using genetic testing, germline testing to identify individuals at high risk for those complications and then use this information to certify which patient is more appropriate to take particular kind of treatment, like ADT because ADT will affect hormone, will affect the things. So I think this will play a major role in term of prostate cancer survivorship. I just don't think this is well appreciated right now and I hope we kind of bring enough attention to this key question.

William Isaacs: The most important thing that we can do is to affect treatment of prostate cancer. And so I think all these words are very important and we need to just confirm the findings that we think are true and build on them. And certainly if they're not, then we need to go on to other ones. But this whole idea of looking at the patient as a whole obviously is extremely important and I think we need to spend more time thinking about that. And again, I appreciate the ability to work with folks and we look forward to interacting with our colleagues both at Hopkins and also another USC and England and various folks to see if we can confirm this result and hopefully build on it.

Andrea Miyahira: Okay. Well, thank you so much, Dr. Xu and Dr. Isaacs, for sharing this with us today.

Jianfeng Xu: Thank you.

William Isaacs: Thank you very much.