 For those of you who are not interested or expert in P53 biology, HDM2 and HDMX are both proteins, components of a large, multi-molecular complex that exists in association with P53, primarily to limit P53 function and stability. HDM2 has E3 ligase function, and it's capable of attaching ubiquitin to P53 and diverting it to the proteasome for destruction. This relationship between HDM2 and P53 is not entirely monogamous. HDM2 does have other substrates, and P53 can be degraded by HDM2 independent mechanisms. But for all practical purposes, the interaction that I have on this slide is the dominant regulator of P53 stability in a cell. This is the reason why P53 levels are critically low, almost undetectable, and unstressed cells, and the disruption of this interaction between HDM2 and P53 is the reason why P53 levels increase in the setting of cellular stress, such as in the case of hypoxia or DNA damage. The idea of targeting the interaction between HDM2 and P53 is not novel. In fact, I think it's actually two decades or more old. The HDM2-P53 interaction is actually one of the first non-kinase targets selected by the pharmaceutical industry for drug development. And we are just now beginning to reap the advantages of all of this research that has preceded. And we now have as many, I think, as six HDM2 antagonists that are in Phase 1 testing. Okay, why are we interested in this specifically in renal cancer? And there are actually several reasons. The most glib is the fact that in renal cancer, the P53 gene is almost always intact, neither mutated or deleted. Well, that simply means that if you had a drug that functions primarily through P53, the effect of the drug is not likely to be undermined by something as simple as the absence of the relevant target. But the real reason we're interested in this has to do with the biology of the molecule. I think most of everyone in this audience is familiar with the canonical P53 functions, which is to induce cell cycle arrest, to kill the cell through the induction of apoptosis or necrosis, or in some cases to induce cellular senescence. However, there are all sorts of less well advertised functions of P53 that have to do with its role in tissue remodeling. P53, for example, is intensely anti-angiogenic through a variety of mechanisms that I'll show later on. It's anti-inflammatory. It interferes with the recruitment of leukocytes into tumor tissue. It modifies the extracellular matrix. It has effects on cellular metabolism, generally through the diversion of cells from glycolysis to oxidative phosphorylation. Now, the reason this is interesting is that across the board, these P53 effects on hypoxic tissue remodeling are almost exactly counter to those of hypoxia itself through HIF. We got interested in this primarily through our earlier work on the mechanisms by which tumors, especially RCC develop resistance to VEGF antagonists. These tumor cells experience drugs that target the vasculature primarily as hypoxia. So it occurred to us if you had a drug that could limit the ability of tumor cells to adjust to the hypoxia triggered by angiogenesis disruption, you might have something that would prevent the early appearance of resistance and other clinical problems associated with the use of TKIs. So to test this idea, we took a renal cell line and introduced a SHRNA to P53 under the control of tetracycline. I'm just going to draw your attention on this rather busy figure. The bottom growth curve, which is what you get when you implant these cells, allow them to grow to a fairly good size and then treat the mouse with sous-tent. What you see is a very brief period, maybe seven, ten days, where the tumor doesn't grow and then it begins to grow. I can assure you that if you were to study these cells under the microscope, you would see that in that compressed period of time of about a week, they do everything that you would expect to see in a conventional tumor that responds initially in a patient and then later develops resistance. The tumors develop necrosis, they lose their vasculature, and then rapidly reacquire it within that rather narrow time frame. But I also want to draw your attention to this curve, which is what happens when you delete P53 from the cells. The point is that the curve overlaps that of our untreated control. In other words, if you do not allow P53 to come up in the setting of hypoxia that is induced by angiogenesis disruption, sous-tent has no effect at all on the growth of these cells. But I also want to draw your attention back to the original curve. If the activation of P53 is such a big deal and so critical to the effectiveness of VEGF targeted therapies, why is it that, at least in this cell line, the drugs work only briefly? Why is the effects so limited? And the answer is that because P53 activation and the consequences downstream are in fact rather limited in time, it's very easy to show the activation of P53 with tumor-associated hypoxia triggered by sous-tent. And you see most of the canonical genes that you would expect expressed here. Naxa, the gene that's associated with apoptosis, P21, and a lot of the others. When these tumors become resistant, which they tend to do rather easily, P53 levels are still maintained, almost at the same level. But what you see is the disappearance of many of the genes, in particular P21 and Naxa. So P53 function becomes progressively compromised as the treatment proceeds and in particular as the tumor develops resistance. Now part of this, we think, is due to the expression of this protein here. HDMX, a known P53 antagonist. Now I won't show you any data of this, but we have done a similar experiment with an HDMX SHRNA, and it has the same effects as an HDM2 antagonist. It prevents the emergence of resistance. We asked the question then if we could use an HDM2 antagonist that would raise P53 levels to extremely high levels. Would we be able to avoid the development of resistance? In other words, if we could just simply sustain P53 function, would that be sufficient to avoid the emergence of resistance? So these are just two xenografts, 786 and A498. And you can see in both cases, when you treat the cells with both sous-tents and an HDM2 antagonist, in this case the Sanofi compound MI319, the development of resistance is at minimum delayed, if not in fact prevented. This is what happens when you look at the Western blots. You can see that in this case, you not only activate P53, which you can do with sous-tent alone, but you restore the function of P53. You can see P21 and NOXA, and many of these P53-dependent genes restored. One of the most dominant effects that we see in these cells that are treated both with a TKI and with an HDM2 antagonist is the profound lack of vasculature. I mentioned previously that P53 has anti-angiogenic effects. Well, this is just a simplified version or view of all of the anti-angiogenic effects of P53. Some of them are mediated through microRNAs. Some of them are through the metabolism of collagen, in particular the generation of angiostatic peptides such as endostatin or canstatin. But some of them are mediated through the suppression of HIF1 and HIF2. Now, the suppression of HIF1 is in the literature. That was actually reasonably well known. But there was no literature that suggested that P53 activation could suppress HIF2. So we actually looked at that. This is an IHE, an immunohistochemical slide of renal cancer xenographs that are treated with nothing with the P53 agonist or HDM2 antagonist MI319 sutent or the combination of the two. You can see the marked reduction in HIF staining in these slides. And it's better illustrated here with this bar graph and in particular with this western block where you can see that HIF2 is markedly down modulated in cells treated with MI319 alone or in combination with sutent. Now, to investigate the mechanism by which this occurred, we did quite a bit of work that I won't show you, but the mechanism is in fact P53 dependent. If you knock out P53, excuse me, and then treat cells with this HDM2 antagonist, you can see P53 activate here, not in the presence of the SHRNA. HIF goes down here, but not here. So it's certainly P53 dependent. Won't go into the details of it, but we showed that the loss of HIF2 here is mediated primarily at the level of protein stability. So we ask, is there a E3 ligase that could go after HIF2 that might be P53 inducible? And there is, actually. There's this protein called FBW7. It's a known P53 target. And if you look through the amino acid sequence of HIF2, there's two recognitions, motifs, for this E3 ligase. So what we did here, I can get it back if the reverse is not working. Can someone reverse this slide? It's just going forward. Okay. What we did here is introduce an SHRNA to FBW7, and we showed that when you treat the cells with the HDM2 antagonist MI319, FBW3 is induced, FBW7 is induced, but not here with the SHRNA. And we lose this ability to now modulate HIF. Now, the ability of FBW7 to ubiquitinate and degrade HIF2 is dependent upon its prior state of phosphorylation. Most of the substrates of FBW7 require prior phosphorylation by this enzyme Gsk3-beta. So if you knock out Gsk3-beta, the effect of FBW7 induction is lost completely. So here's with the SHRNA. You treat with MI319, P53 is induced, Gsk is not affected, but here the levels are low. Here, HIF2 is downmodulated, here it's not. So this basically says that the downmodulation that we induce in vitro, at least, by activating P53 is not only FBW7 dependent, but critically dependent upon the activity of this enzyme Gsk3-beta. I'm just going to review one other angiogenic mechanism if I have time, and that's the ability to suppress certain chemokines that attract inflammatory leukocytes. SDF1, for example, is known to be a P53 target, and it's readily suppressed by using an HDM2 antagonist. Here's what happens when you treat renal cancer xenografts with Sutent. This is readily induced in the stroma primarily, primarily through hypoxia, and you can completely prevent that. If you look at the cells that they regulate, these inflammatory myeloid cells, you can see a few of these cells in the control. In our hands, the recruitment of these cells is actually increased with submit-nib, presumably because of tissue hypoxia, but you can almost completely prevent their influx into tumors by using a P53 drug like MI319. This is just a biograph illustrating the same principle. We've looked at these myeloid-derived cells that we isolated from the spleen of mice that were treated either with Sutent or with MI319 or with both drugs, and one of the things that's characteristic of these cells is that they make massive amounts of TGF beta if they are taken from a mouse that got Sutent, but not from an untreated mouse and not from a mouse that got both an HDM to antagonist and a submit-nib. So these drugs not only regulate the trafficking of the cells, they also regulate their function. As you might surmise, the TGF beta that these cells make is extremely clonogenic. If you incubate 786O cells with the supernatants from these cells, you get tons of colonies which otherwise are hard to find. This is just an experiment to show that the effect on the ability to promote colony formation is limited to the cells that have both GR1 and CD11B expression. In other words, it's limited to this myeloid-derived suppressor cell phenotype. This again just examines the effect of treatment on the ability of these myeloid cells to promote clonogenicity. And as you can see here, I just want to draw your attention to the final two bar graphs. The effect is lost if the mice get treated with an HDM to antagonist. So in conclusion, I'd just like to sum up one of the reasons why or the reasons why we think HDM to antagonist have potential specifically as adjuncts to VEGF targeted therapies. One is that concurrent HDM to blockade does prevent the development of resistance to synitinid. Part of this effect is through the additive anti-angiogenic effects in which the HDM to antagonist adds to the anti-angiogenic effects of synitinid. The use of these drugs, in particular the Sanofi compound MI319, is associated with a diminution in hip to alpha, possibly mediated through this E3 ligase, FBW7. And we've shown that the inclusion of an HDM to antagonist interferes with the recruitment of myeloid cells to the tumor. I'd just like to thank the AACR curate and our kidney cancer spore for supporting this work. Thank you.