Era of Hope 2011 Plenary Sessions

Session Title: Morning Session – Roads Less Traveled: New Perspectives in Breast Cancer Research

Title of Presentation: Clinging to Survive -- the Role of PVRL4 in Cancer: A Link Between Cell Aggregation and Malignant Transformation?

Stephen Elledge, PhD, Harvard Medical School: So I’m the moderator and also the first science speaker. My name is Steve Elledge. Can we have the first slide?

So the topic I’m going to talk to you about today, the title of my talk is Clinging to Survive, and this is about a new gene that we’ve began to study and we think has an important role in breast cancer. And let me just take you through the background on how—how we set up our science and—and what—what the basic information is that we need to understand it.

So we know that the survival of epithelial cells is determined by their interactions with the basement membrane. And it was shown as early as 1964 MacPherson and Montagnier that—that if you take normal cells and plate them in soft agars such that they can't attach to the basement membranes, normal cells die. But tumor cells can survive.

Now more recently, Schwartz and Frisch identified a similar phenomenon called anoikis and in that they showed that—that cells grown in monolayer when they’re detached from their matrix and grown in suspension that these cells actually die. But if you do the same experiment with tumor cells, they survive. So this is a very common property of tumor cells and it’s something we called anchorage independent growth; it’s also referred to as transformation.

So we believe that anchorage independence determines survival of cancer cells throughout the process of tumorigenesis. So initially cells start out attached to their basement membrane, but dysplastic growth allows them to grow beyond contact with their membrane. We think we have a role here for anchorage independent growth and then an invasive carcinoma when these cells begin to move into other parts of nearby tissues they encounter a foreign matrix. We think that—that can actually provide some survival advantage there and then during the process of intravasation where the cells enter into the bloodstream they’re now no longer able to associate with any sort of extracellular matrix. And as they travel through the body, the ability to survive under these conditions is heavily selected for, and finally of course if they metastasis you can see that they—they also now are in a new foreign environment.

So we wanted to take advantage of this to find new genes important in—in breast cancer, and to do so we took advantage of some observations made in Bob Weinberg’s lab and that is that the transformation of normal human mammary epithelial cells required a finite number of genetic perturbations and what they showed was that if you take immortalized HMECs, mammary epithelial cells immortalized with telomerase and introduced various oncogenic insults such as large T antigen which turns off RB and P53 and elevated Myc that these cells are teetering on the edge of transformation. And if you add certain potent oncogenes such as PI3 Kinase or RAS that now you can convert these into transformed cells that grow in soft agar.

So we took advantage of this to do genetic screens to look for genes that could take these cells that are teetering on the brink and force them to go into anchorage independent growth. And we did this by introducing ORF libraries or cDNA libraries which over-produce proteins or by turning off proteins using RNA interference. And so we’ve done both of these types of transformation screens, and I just want to summarize one that was recently published and then move onto talk about the main topic PVRL4.

So one gene that we found that if you turned off allows these cells to be transformed was a protein tyrosine phosphatase, PTPN12 and what we learned from this—this is work that was done—initiated in my lab by [Trey Westbrook] when he was a post doc and carried on in his own lab at Baylor, we found that PTPN12 actually acts to—to turn off a variety of tyrosine kinases that could drive proliferation. And this exists in many cells, and what we find is—we think is that there are—the ways that you can become transformed, one of them of course that’s very familiar is the amplification of Her2, leads to malignant transformation and the other is we find that a PTPN12 is mutated in breast cancers in particular non-Her2 and it’s even—it’s mostly a mutated and a triple negative disease.

So to get rid of PTPN12 now these—these basal levels of these kinases can—can drive tyrosine signaling in—in the absence of amplification of Her2 so the two important things from this work I’ll just summarize are that we think that a number of triple negative cancers are actually driven by tyrosine phosphorylation. And up to 60% don’t express PTPN12, which we think is interesting and potentially a biomarker. And secondly if these cancers are driven by basal levels of kinases since we have inhibitors to a lot of these kinases, using them in combination could be used to treat this disease. And we’ve shown this as xenograft models and we’re hoping to get a clinical trial together to—to try to test this in triple negative disease.

Now the rest of my talk I’d like to talk about a gene PVRL4 which induces transformation. If you over-express it it’s called the Poliovirus Receptor Like-4; it’s also called Nectin-4 and you can see if you introduce this into cells you can see that they become highly transformed.

So—so what—what brought us—made us interested in this was a paper that we saw in 2007 that showed that PVRL4 is over-expressed in about 62% of ductal carcinomas. So that got us to—to look at it more carefully to see if there were any genetic alterations in breast cancer. And our analysis that showed PVRL4 is focally amplified in breast tumors. Now it’s on Chromosome 1Q and that arms is often—often lost but they’re also focal deletions in about 14% of breast cancers.

In addition, a paper that just came out 2—2 months ago showed that patients that have—when—this is a Kaplan Meier Survival Plot—patients that have low levels of expression of PVRL4 survive much longer than patients that—that have stained positive for PVRL4 and this is—this is a 15-month window here which is really quite impressive.

So let me tell you about this family of proteins. These are cell surface proteins; they have immunoglobulin-like repeats and they’re thought to interact and trans with similar proteins of the family on other cells. And from one that’s been studied most, PVRL3, it—they think that it promotes adherence junctions via stabilization of—of E-Cadherin. It also hooks up through Afadin to the actin cytoskeleton to promote filopodia and lamellipodia.

However we don’t think that PVRL4 is doing this because when we overexpress it in our mammary epithelial cells we do not see stabilization of E-Cadherin. Instead we’ve been focusing on a—on another property that we noticed about PVRL4 which is that induces these—these cells expressing it to aggregate in suspension. So you can see these cells are suspended. These are normal cells. They—they exist as single cells in the suspension but in PVRL4 overexpressing cells they form these aggregates. And that’s—that’s quantified on the right; it’s quite a bit of aggregation there and we think this may actually be important because this activity tracks with all the other important activities I’ll tell you about.

So we wanted to know more about this aggregation and as I said they—they tend to act in trans and a partner for PVRL4 that’s been suggested as PVRL1, a related protein. And so we decided to ask whether or not it was important for this—the activities that we see with PVRL4 so we made short hairpins to PVRL1. This just shows the RNA is knocked down and we did the experiment shown in the next slide.

So we took mammary epithelial cells and made them either green or red and then the green cells which have normal levels of PVRL1 we overexpressed PVRL4 and you can see these clumps of green cells. They begin to clump. And then we mixed in normal cells that don’t overproduce PVRL4. And these either can—you can see that—that these cells now can still enter into these clumps. They don’t clump by themselves, but they aggregate in the presence of the green cells.

Now if we turn off PVRL4 in the red cells by adding our short hairpins, you can see that the green cells still aggregate but now the red cells don’t. So we take this as evidence that—that these cells can aggregate and trans and that they require PVRL1 for this activity.

So just to summarize some of the other findings we’ve found, we—we can remove the cytoplasmic domain and the transmembrane domain and replace it with a transmembrane domain of another unrelated protein and just have these on the surface. And if you do that and generate cells that only have the extracellular domains, the PVRL1 and 4 they still aggregate and they still transform. So we don’t think it’s acting through the signal transduction through the cell.

So we went on to look at cancer lines that overproduce this protein, and we found one called SUM190; it’s a focal amplification and this amplification contains only about four or five genes. So it’s really very specific. So what we found—if we take PVRL4 away from these breast cancer cells using short hairpins, and this is the assay for the knockdown, the green and yellow bars are controls, and these four bars to the right are different hairpins for PVRL4, you can see that these hairpins have a deleterious effect on colony formation in the clonogenic assays which is when you plate these cells at low density you can see that this is what it looks like in controls. And if you remove PVRL4 they can't survive. Now this is even in the presence of extracellular matrix.

If you go and ask about transformation you see a similar—a similar observation; these cells can no longer form colonies in soft agar.

So we wanted to see whether or not the aggregation phenotype is correlating with this. And so some 190 cells we can see form aggregates when placed in suspension and—and in fact if you then add a monoclonal antibody to PVRL4 you can see that these—these cells now no longer aggregate. And the same is true as if you knocked down PVRL4 so the aggregation phenotype correlates with all the positive growth properties of PVRL4.

So another thing that we observed and has been seen for some cancer cells is that—that some 190 cells can interact with human lung microvascular endothelial cells. And these are the cells that—that line the vasculature and so we’ve done this—this labeling trick again. Now the green cells here are some 190s and the red cells are the microvascular endothelial cells and you can see that some 190 actually can recruit these endothelial cells into aggregates.

However, if you add an antibody to PVRL4 or knock it down with an shRNA these aggregates no longer form, so PVRL4 is required for the ability of these cells to interact.

So we were interested in the possibility that it’s the aggregation itself that might cause this transformation and to test this we looked for alternative ways to induce aggregation. And one—one we found in the literature was an antibody, the CD44 caused cells to—to form aggregates when you added just the antibody and this happens within an hour. And so we—we looked at whether or not this could actually affect the transformation and now these normal mammary epithelial cells that we’ve been using, and sure enough if you add just this antibody to the cells during plating, not before, but during plating—so they actually don’t have a chance to form the aggregates before they’re plated, these—these antibodies, which we think link cells together, actually dramatically increase transformation.

So we were interested in this property also because of the new emerging phenomenon of clusters, of circulating tumor cells, which are clusters of cancer cells. They’re heterogeneous clusters of circulating tumor cells of about four to twelve cells each and they’ve been found in a number of different cancers including prostate cancer, mouse models that are driven by KRAS and p53. They also have found that these CTCs can adhere to the microvasculature in—in different experiments and it’s also been found that in mice, lung metastases start out intravascularly from the proliferation of endothelium attached tumor cells.

So we think that—that this property of allowing cells to adhere to one another may give you advantages during the—the process. We call this clinging and that it can enhance the survival of cancer cells throughout the process of tumorigenesis. This is—we feel that at this stage when cells are now moving into new tissues you can see that there is—there may be a growth advantage by the ability to adhere to blood vessels or to each other and we think that one possibility is that if these cells adhere to one another and form a common surface, one way that they could take advantage of that is if they secrete survival factors into that patch. That—those factors will not diffuse away and they can give more signaling. This is all hypotheses but this is what we’re thinking about this and we know we have a little ways to go to prove this.

So we’d just like to summarize here and say that PVRL4 we think could be a new target in breast cancer. The reason is—is that we think it’s important for the survival of breast cancer cells and it’s also on the cell surface. So it could potentially be treated like Her2 with Herceptin or other antibody reagents. One of the good things about it is it’s not very highly expressed in any of the tissues. There are just a very few tissues that express it so that’s an important aspect.

So I’d just like to thank the person who did this. This was all done by a single person, a Graduate Student in my lab, Natasha Pavlova. She’s an excellent student and thank you for your attention. I’ll take any questions now.

Question: So you have that concept of the tumor cells feeding each other. I think it’s a very attractive one. Have you done in vivo studies where you use an antibody and use some cytotoxic agent or anti-Her2 or something like that to—?

Steve Elledge: Yeah those—those are—that’s a great suggestion and that’s the experiments we’re hoping to do now. That—that’s exactly the right way to go I think.

Question: In terms of the observation in mice about the clustering of cells in the lung endothelium, do you think that maybe the receptor that you’re looking at might play a role in a seed cell binding to the blood vessel wall and then other circulating cells sort of gathering on top of it?

Steve Elledge: Yes; that’s exactly what we’re thinking that it—it could also you know produce more cells but it can gather other cells and allow them to survive on their way through the bloodstream.

Question: So are you looking—using whole animal imaging or anything, the ability to do that early trapping stuff?

Steve Elledge: No; we haven't started those experiments but we are interested in looking at the role of metastases and another point I meant to—to mention was that in the same paper where they looked at survival they also analyzed the occurrence, the frequency with which you see PVRL4 in lymph nodes. And—and so if you look at patients—it’s only—it’s only present 60% of breast cancer patients, but if you look at patients with three or more positive nodes it’s 100%. So we think it’s probably going to have a role in metastases and that’s just a correlation. Until we get a—a model that works like that we can't really test it.

Question: You—you showed a paper where they showed PVRL4 over-expression in cancers but—and the way you sort of have done this in vitro experiments is—is to look at PVRL4 and one interaction. So did you sort of find these two proteins co-localized in breast cancers?

Steve Elledge: PVRL1 is expressed widely throughout the body so it’s on a lot of cells. So—so the only one that’s actually different in breast cancer versus normal tissue is PVRL4.

Question: So have—have you thought about the heterogeneity enough? I mean is—is it—is it possible that one of the rules of this could be in keeping the cells together in a normal setting and there could be some cells which where you don’t have enough of the protein then—disseminate and you know they could have other ways of surviving and so what do you think about—about those kinds of issues early on as you think about these studies and start thinking about it as a potential target?

Steve Elledge: Well I think there are a lot of possibilities out there including the one you mentioned. I think that the fact that these only show up in cancer cells, breast cancer cells—suggests that it’s promoting some property. It could be survival and tumor growth but you also see it in metastases and—and other places that beyond the normal tumor, so if they had to turn it off and were to move they certainly would turn it back on. That’s all I can say.

Okay; I think we should move onto the next—. Oh one more question; okay.

Question: Yeah; sorry. Are PVRL4 overexpressing cells resistant to apoptotic stimuli?

Steve Elledge: Well if you call anoikis apoptotic stimuli I would say yes, and we haven't tested others.