 Let us resume and have our next invited speaker, Richard Sharp. And let me begin with a couple of things. Firstly, to acknowledge that the data that I'm going to present to you is the work of the people listed here on the slide, particularly Rod Mitchell, who's a pediatrician who's done the xenograft work, the majority of the xenograft work, and Sander van den Dris, who's done the mechanistic studies that I'm going to start with and that will probably be big news or very, very news to you that it's been two years in the developing. And perhaps more importantly than that, I need to actually start off by telling you how I'm coming at Thalates because my perspective is very different to a lot of you here, that I'm not a toxicologist. I am not interested in necessarily as the primary reason for risk assessing Thalates. I'm interested them as a tool to aid us in our understanding. Because the way that we come at this is by, we're interested in what Kim has already referred to is these disorders, which are sort of grouped together as testicular disgenesis syndrome and related factors, which are either increasing in incidence or already very common. And I think the evidence that they have prenatal origins or at least for some of these disorders, important prenatal origins for a high proportion, is growing increasingly. And it surprises even me. So we're interested in trying to understand what causes those disorders, the mechanisms that lead to them, what sort of factors might hit them. And I think the evidence is, as I've said, growing stronger, that they have a common fetal origin and that probably involves subtle, and I mean subtle, disorders of testosterone production or androgen action in fetal life that then predispose to development of one or more of these disorders. So we want to know what those mechanisms are, what sort of factors might hit them. And that's where Thalates, such as DPP, come in, because if we expose to these active Thalates, then they induce a spectrum of disorders, which is, if not identical, at least includes quite a number of those disorders of the human. And therefore what we've been doing is using DPP as a tool to actually try and explore the mechanisms, not necessarily with the view of proving or disproving that Thalates might cause those disorders in humans, but to identify the mechanisms and then to ask, well, if Thalates don't target them, what other factors might, including things like endogenous hormones. So that's how we're coming at it. So I think bear that in mind when you're asking me questions, because if you ask me a lot of the questions, the tough questions that you ask Kim, you're going to get a blank stare and a shrug of the shoulders, and I'll refer the question to somebody in the audience. So I want to start you back, again, conceptually with something which is a pretty picture, I think you'll agree. This is the moment when the testis is forming in a fetal rat, E13.5. So in red are the cittoly cells, that's Sox9. The green are the germ cells, which have actually migrated into the genphal ridge, and you'll see that there's a few stragglers here still making their way in. And you'll see that the seminiferous cords haven't yet formed. They're in the process of forming the cittoly cells, enveloping the germ cells. At this stage, there are no latic cells. They're not going to develop for another few days. But even at this stage and age, DBP is already exerting effects. Exerting effects, well, I'm not going to go into that story, but it's also having effects on the germ cells. I will go into that. It's having generalized effects, which are of considerable interest from a cell biology and lineage specification point of view. But the latic cell effects are not going to emerge for some time. So traditionally, these are the sorts of effects that you see, and others have seen, as Kim has already shown you, these are some of our own results, that when you treat with a high dose of DBP starting, we're starting on E13.5 dosing daily. And this is then looking at different ages. And you'll see that as you get towards the end of gestation, you've got about an 80% suppression of testosterone levels. So an impressive decrease. But really, as I've already sort of hinted at this morning, where the action really takes place is back here. Because it's only when you suppress testosterone or androgen action back here that you induce those disorders that we're interested in. So we're not so interested in the effects here, only in what's happening here. And what happens back there is that at E15.5, there's absolutely no effect of DBP exposure on testosterone when it's first starting. You do see suppression towards the end of the programming window, but it's certainly not as profound as here, and it doesn't even always reach this sort of level. It may only be 20%, 30%. And this explains why you don't see very severe hypospadias or very uncommonly in DBP exposed animals. So I want you to pay particular attention to this age-dependent pattern because I'm now going to get you to think in a very different way about the DBP effects. No effect at E15.5, effect emerging here becoming more pronounced as you go through gestation. Because through traditional view would be that in that programming window, there is no LH. So the drive to the fetal latex cells to make the testosterone that's bringing about, that's programming masculinization occurs via unidentified paracrine factors, of which there are potential candidates but which nothing proven physiologically. And that DBP is somehow affecting the production or action of those. So that would be how I would have solved this two years ago. Turns out that it's completely and utterly the wrong way around. But let me explain to you how we've arrived at that. So let's look now at what happens to latex cells and latex cell function as you go through this period. Firstly latex cell number, protestus. So you see that in controls, latex cell number increases as you go through this period. It's not a huge increase, it's about a four-fold increase. And despite what your eye might tell you when I show you some of the pictures later on and I've already heard latex cell hyperplasia mentioned today you don't get latex cell hyperplasia in DBP exposed animals. There's no effect on latex cell number. They just get distributed differently throughout the testis and they tend to aggregate. So there's no effect of DBP on latex cell number. So you can't explain the reduction in testosterone in that way. But things look very different when you start looking at the size of the latex cells. So I'm just showing you cytoplasmic volume, but it looks very similar for nuclear volume as well. But again there's about a four-fold increase in the size of individual latex cells. And for those of you who don't know the volume of cytoplasm is directly proportional to the amount of testosterone that they can make because it harbors all the steroidogenic organelles. And what you see in the DBP is it just never takes off. They remain the same size as when they first differentiate. So that when you look at testosterone production per million latex cells, remember latex cell number hasn't changed, again it increases progressively in the controls but in the DBP exposed animals it flatlines. So in fact what DBP is doing is not inhibiting steroidogenesis, it's preventing a normal increase, which is a completely different way of looking at it and as you'll see there's a completely different explanation from the one that we would have started with. So the traditional view is wrong. So what is the explanation? So let's go back and look in a little more detail at what some of the things that we know that DBP does and you've already seen some of this data from Kim as well, similar data, looking at expression of steroidogenic enzymes. One of the things we know is that DBP exposure affects SF1-dependent genes involved in cholesterol, transport and steroidogenesis and it's in the studies that we did with Simon Plummer then we show that there were actually seven different SF1-dependent genes in latex cells that are affected by DBP but only from E17.5 onwards, not at E15.5. And remember at E15.5 there's no effect on steroidogenesis. To get suppression, this is looking now at E21.5 so it's at the end of gestation when you see the biggest effects on steroidogenesis. CYP11's reduced star, CYP17, 3-beta HSD isn't. This is an SF1-dependent gene the same as these. Some people do see effects on 3-beta HSD. Maybe sometimes we see smaller facts but I think they're meaningless anyway because this is constitutively over-expressed. You could probably knock it down by 80% and there'd be no effect on steroidogenesis. And you see these changes at the protein level as well, CYP11 versus 3-beta no change. And if you look at SF1-dependent genes elsewhere in the semineferrous quartz such as anti-malarion hormone again there's no effect. An SF1 expression itself is unaffected. So this is sort of a curious thing from a mechanistic point of view. How are you affecting so many SF1-dependent genes but not all of them and in this particular fashion? So here's where the explanation comes from left field and probably for a factor that none of you have ever heard about. Certainly I hadn't. Two years ago, chicken, ovulbumin, upstream promoter, transcription factor 2, COOPTF2. If you haven't heard of it, you probably ought to have because it's a very important molecule and belies its name. It is, I think, still the most highly conserved protein that's actually been discovered yet. So 96% identity across species. And it's involved in lineage specification in a whole variety of tissues, fundamentally important roles. But it's also been shown in a variety of transfection systems that it will inhibit steroidogenesis by steroidogenic cells, and it does this by competing with SF1 for binding to response elements in the promoters of genes. When we've looked in rat promoters, there are overlapping SF1 and COOPTF2 binding sites in the promoters of STAR, CIP17, AMH, but not 3-beta HSD. Well, AMH is in there, but what you should know, or what you will see, is that COOPTF2 is not expressed in satoli cells, so it's not going to compete. But it is expressed in latex cells. So the hypothesis is, and this may sound to you that this is great that we were great in our intellectual thinking in coming up with this hypothesis, the hypothesis comes after the results. So we came up with a hypothesis after we got the data and tried to interpret it. And we stumbled across this completely by accident. So here's the STAR gene. There are four overlapping sites where COOPTF2 and SF1 can compete. So depending on the relative levels of expression of these two, you could get antagonism of SF1 action if you've got high expression of COOPTF2. So let's go on and actually look at COOPTF2 expression. Throughout the period that I've been talking about. So COOPTF2 is in green, latex cells, 3-beta HST are in red, and smooth muscle actin, which will pick out the blood vessels and the peritubular myoid cells that the seminiferous cords are in blue. So these are the seminiferous cords. You can't see the satoli cells or the germ cells because they don't express any of these factors. So you see that COOPTF2 is widely expressed in non-latex interstitial cells, and I don't know how clearly you can see it. Probably clear on the screens is that virtually all latex cells here, when they first differentiate, are expressing COOPTF2 in their nucleus. But as you move through gestation, then it switches off progressively. You probably can't see it very clearly here, but I'll show you the quantification in a minute. What you will be able to see very clearly, otherwise I should point out that you can probably see how much the latex cell cytoplasm increases in size as they go through here. And you'll see that in DPP-exposed animals, it's quite obvious that COOPTF2 remains on throughout. These are these big latex cell aggregates, and you can see also that latex cells are actually very small in size. So COOPTF2 is switching off progressively as you go through gestation and controls. It's not switching off at all in DPP-exposed animals. So we've quantified that. And this is in big numbers, and this is every latex cell in one cross-section, and I'll show you those images in a minute. So in the controls, you start off, you've got about 90% of latex cells are positive for COOPTF2, going down to something like 10% by the end of gestation. And then you switch off. It just never happens in DPP-exposed animals. And that's the mirror image of what happens to testosterone. Testosterone goes up as COOPTF2 goes down. It just doesn't happen in the DPP-exposed. And it's a dose-dependent effect. Increase the dose of DPP. You increase the proportion of latex cells that remain COOPTF2 positive, and you see the reverse with intratesticular testosterone. And in case I forget to mention it, I should also say that DPP is not the only thing that affects this mechanism. So how we think it really works is not like the original idea, but that when fetal latex cells first differentiate, they're held under repression from a stereogenic point of view via the expression of COOPTF2 in their nucleus that competes with SF1. And that what you have to do to actually get masculinisation underway is to remove that repression to allow testosterone, the capacity for testosterone to expand. Now, what causes COOPTF2 to be switched off is unknown. It would be very interesting to find out and suggest that maybe that's what DPP is affecting. Okay, later in gestation things change and you've got LH as a player as well, but I'm not going to be getting into that except maybe in question time. So the issue is, what I've shown you is a very nice association data. So what we were seeking to do is to take it a little bit beyond that and to say, how can we actually prove that this association is causal? So one of the ways that we've done it is to say, okay, we're normally treating to write this period and looking here when COOPTF2 in controls is switching off progressively. So let's wait until E19.5 when the majority of latex cells have expanded in size and switched off COOPTF2. What happens if we treat with DPP here? Because we know it will suppress genesis when we look here. What happens to COOPTF2 expression? And the answer is, you switch it back on. And there's the quantification and there's the suppression of testosterone. The suppression is not as great as if you've been treating all the way through, presumably because, firstly, the latex cells are larger and therefore more stereogenically active. Plus, we don't get back up to the levels of sort of 90% positive. These are the images we're working from. These are high-resolution confocal tiled images that take about 20 minutes to generate. They're horrendously expensive with our imaging costs at the moment. But you can magnify these, hit the magnification button five times without losing resolution. You can have huge individual cells which you can then score. So we do the analysis by scoring every latex cell in here as positive or negative. So what about the mouse? So in the mouse, the mechanism is there, but DPP does not affect it. There's no effect on testosterone. There's no maintenance of COOPTF2 expression in the mouse. So again, at least based on this association, there's a logical explanation for why you're seeing no inhibition of stereogenesis in the mouse. There are other things that will switch COOPTF2 on in the mouse and will reduce testosterone production. So based on this, then, we've got T suppression, COOPTF2 maintenance, and neither of these in the mouse. Of course, the obvious question is, well, what about the human? Firstly, let me just remind you of something which is rather fundamentally important, and that is the way human fetal latex cells operate is rather different to those in rodents, because we know that they are expressed in electroceptors at least from relatively early ingestation, if not from the moment they develop, and that HCG is produced from the placenta and that that is an important driver to testosterone during the period of masculinisation. And we know that because if you have mutations inactivating mutations of the electroceptor in humans, then you end up with masculinisation disorders, whereas if you have that in the mouse or the rat, you don't. There's no effect whatsoever, because fetal latex cell testosterone production during that critical period in rodents is, I think, driven by that COOPTF2 lifting of repression, not via the electroceptor. And again, LH is a player in the second half of gestation. So in the human, and I would say also in the marmoset, the COOPTF2 mechanism is potentially there in that if we look at the percentage of latex cells that are positive for COOPTF2, it drops off as you go through gestation. But it's not as dramatic as I've shown you in the rat, and we're only starting at about 35%, because what we're missing are the earlier first trimester ones, which would be back at the equivalent of E15.5 in the rat. So we can presume that there might be more positive cells back there, but we can't prove that with any certainty. So I think the earliest we've got here is about 10 weeks, 10 or 11 weeks. And of course, mechanism being there doesn't mean that DBP is going to affect it, because I've just shown you that doesn't happen in the mouse. So on to the xenograph system. So Kim has already introduced you to this, but there are differences in the way that we've approached it. Firstly, we're dealing with elective abortions. They're all fully consented. I'm only going to be showing you second trimester. We've done some first trimester, but not really, I don't think we've done any with DBP exposure. We're grafting under the skin of the back, not under the kidney capsule, putting in about four to six grafts per mouse. They're adult castrate males, so that we know that if we measure testosterone in the host or semloviescal weight, it's a measure of testosterone production by the grafts, not by the host. We're treating with DBP or MBP. Measuring these, the important thing, something again different from what Kim was telling you, is that we're treating with HCG, as would be the case in situ. Why are we treating with HCG? Because we're doing our grafts over six weeks, and we're treating for the end of the period, starting working our way backwards from that six-week period. And if you don't treat with HCG, you don't have any measurable testosterone, and you don't get semloviescal weight above castrate control levels. So despite the fact that the mouse has got elevated LH levels, because it's got no testes to feed back to the hypothalamus pituitary, that mouse LH is completely incapable of stimulating the human fetal latex cells. As is, a dose of 5RU8CG, which we're administering three times a week. When we switch, we sort of switched, actually it occurred by accident, from 5 to 20, towards for the last two weeks, you've got stimulation, and now we treat solely with 20RU. For most of our studies, that I'm going to show you. So if we treat in that way, we are reproducing what's occurring in situ, although, of course, this isn't necessarily the same level of HCG whose exposure might occur in vivo. But we've got to do this anyway, but if we want to test phthalates, this is just really telling you what it's like. Here's the mouse being castrated. Actually, this doesn't occur at the same time. We castrate the mice and then do the graft in two weeks later. Here's the size of the pieces. And when you collect them six weeks later, this is what they look like on the underside of the skin. And they usually grow in size by about three-fold on average over that six-week period. Here's what one of them looks like when you cover it. Probably a very nice specimen. Looks almost like a completely normal, little testis on its own. If you showed this to a pathologist, they probably wouldn't be able to tell it from normal, although I think this is one which actually hasn't been HCG treated. But on all the criteria that we've looked at so far, these continue to grow and develop normally. I can't really believe that it's going to be completely normal, let's say, based on the parameters we've looked at so far. It's the same as it should be. There's no effect of DVP treatment on graft survival. You get about 75% to 80% survival of the graft, so it's not 100%, or on recovered graft weight. They don't all grow to the same extent. They're not exactly the same size when they go in. And you need to actually think about graft weight because if you get differences between hosts, then if there's more overall latex cells, then you'll get more testosterone and more effect on semoviescal weight, and you'll see that there is variation and that you need to show that there's no effect on graft weight because it might give you a spurious result. And the same is true for testosterone, although it's not so convincing. Is there an effect on COOPTF2 expression when we're treating with DBP for the last three weeks of grafting? Well, we haven't looked at this systematically. What we've done is just looked at about 16, I think, different samples, and we can't tell the vehicle from the DBP. There's quite a lot of variation in the percentage of latex cells that we're expressing. I think in this one here, there are actually quite a few positive cells in this sample and not quite so many in the DBP exposed, but there's other ones that are the other way around. Overall, there doesn't seem to be any effect, but we've not done the systematic quantification yet, so that's a tentative result. But it might be what we would expect because there's no effect on testosterone levels, either measured directly in the animal at the time of sacrifice or the readout provided by semoviescal weight. So what this actually shows is data from eight individual second trimester fetuses of varying ages, which have all been grafted for six weeks, all been HCG treated from one week after grafting. We don't treat during the first week when they're establishing a blood supply. And then for the last four days, they've been treated with DBP or the last 21 days, so that all the grafts are the same age, if you like, when we complete the experiment and controls get the vehicle for 21 days, so we can directly compare all of these. And remember that there are also two or three replicates for each of these, for each fetus. So we're doing two or three hosts per fetus. And so we've analyzed the results in various ways, ignoring the fetus, or in this case, ignoring the host or taking account of the fact that you've got replicates. It doesn't matter which way we do the analysis. There's absolutely not a hint of any difference in testosterone production or action, as measured by the seminal vesicles. And one of the questions that it does get raised is, well, are the mice... Well, we've had it said to us that mice are not so good at metabolizing TBP to MVP, although from Kevin Guido's paper, it looks, if anything, the other way around, that mice are rather better. But nevertheless, we decided that we should... We don't know for the nude mice what the situation is. So we have, for a smaller number, this is only for four individual fetuses, looked at the effect of equivalent dose of MVP, treated for 21 days, but again, there's no evidence of any effect. So based on this, it doesn't look... We're finding the same as Kim found in his system. This one is slightly different in that we're talking about graphs that have been allowed to develop and grow, that are HCG exposed, but we're coming up with the same results as Kim has described. But I guess that some of the caveats that were raised this morning would apply equally in this situation. They're not first trimester fetuses. The other problem we have, of course, is what we want to know is that we really like a positive control. We can suppress testosterone production with ketoconazole in these graphs, but that's really not... It's not too surprising, and it's also not particularly helpful. And the problem we have with doing the positive control of putting the rat fetal testis in is that we can't do the same experiment as we've done with the human. So if we're going to put them in and leave them for six weeks, we'll be able to see all the testis would, and they really will be going through puberty because they develop very well as xenographs and you tend to get somewhat accelerated development. So if we then treated with DBP in that situation, irrespective of whether we did or didn't see an effect, I think you could easily argue that we're not comparing an equivalent with what we have in the human situation. So what we've done is to sort of... I guess you could say that we've just done the best that we possibly could, and I think it's a bit of a fudge. And that is that we've grafted E17.5 fetal testis under the skin for four days. And we've treated with DBP. So this is going to mean they're not going to have a blood supply, they're not going to have a patent blood supply. They would do if they went under the kidney capsule, as Kim said, but I think you just have to accept that these will not have a patent blood supply. And therefore, we don't know how much DBP or MBP is getting through to the tissue, and we also don't know... Well, we know that the tissue looks reasonable, if not fantastic. I wouldn't say it looked fantastic. But we do see effects. Testosterone is nearly significantly down, but seminal viscal weight is definitely down, and star and CIP-11 expression are significantly suppressed. So that's about the best that we could do with coming up with a positive control, but it is not a direct comparison with the human. I think it's apples and pears. So how does this fit into the overall scheme of things? Well, if we go back to our publication from 2007 by Nino Hormark, then what we had shown then is that if we... we're doing the equivalent experiment, but this solely in vitro with fetal testis explants, of taking them and culturing them for two days without HCG, with HCG or with hydroxycholesterol to stimulate steroidogenesis, and it doesn't actually matter. MBP has no effect under basal or stimulated conditions. And René Haber has found exactly the same for first trimester fetal testis explants. So I think based on that, you would say that... and what Kim has shown this morning, that everything looks negative, but none of it is in situ. But we have got one piece of evidence of our own for in situ, and that's where we take the marmoset, which is an extremely good model for the human in terms of testis development. And what we've done is to expose them to monobutyl phthalate, 500 milligrams per kilogram per day, for seven weeks, ministered orally every day. And we've started... we don't know exactly when the programming window is in the marmoset, but we know that we've covered it by starting at this point in time and continuing for seven weeks. We've looked at a total of nine. In fact, I think we've now done about 13 marmosets in total. And then we've looked at the offspring at either two to four days or as adults. These are small numbers. These are expensive and difficult studies. And we've also actually looked at about another three or four animals at 14 days, and I'm not going to show you any data for those. And what we're looking for is an effect on their male reproductive phenotype, because obviously the marmoset pregnancy is close to 150 days. So these animals have long been off treatment when they're actually giving birth, so measuring testosterone is not necessarily going to be informative and it might have happened back here. But what should be is if they've got male reproductive disorders and if there's effects like there is in the rat, then we would expect to see these sort of numbers versus these disorders. And what we actually see is nothing. And that's true at birth. It's true at two weeks. It's true in adulthood. So in total, something like about 13, 14 animals, I think. And I should remind you that in this situation, testosterone production by the fetal laying cells in the marmoset, like the human, is being driven by CG. So this... If you want to look at it this way, this is a surrogate for the human. It's that final piece of clinching data, if you like, that says what we're seeing in the other systems, which have their imperfections. It's the same in this system, which is probably the closest we're going to get to directly doing it in the human. So going back to our slide, the rat is the only one where we've got these things operating. It doesn't happen in the mouse. We don't get T-suppression in the human, but we need to work this up to Peter, but it doesn't look as though there's anything there. In the marmoset, we're indirectly concluding that there hasn't been no T-suppression during that programming window, and we don't know about coop maintenance because we haven't looked at that stage during pregnancy, but the mechanism is there. So, no effects on the fetal testis would be the conclusion. Nothing in life, in science, is ever quite that simple. I think that that conclusion is correct based on what there is at the moment, but then there's always a complication. It actually goes back to Neenah Hallmark's study in the marmoset because what we showed then was that if you take a newborn marmoset, remember marmosets are like humans. They have this mini puberty after birth. Testosterone levels are elevated into the low adult range in the marmoset for about four weeks. So what we've done is to take marmosets a couple of days after birth and to treat either with vehicle or with a oral dose of M.B.P. and we've looked five hours later and there's about a halving in testosterone levels. So, contrary to what you find in the fetus, there's clear evidence you're getting effects and yet my presumption all along has been that the leading cells that are present in the neonate and testosterone then are fetal-like. We also did a co-twin study. Marmosets have a high proportion of twins and triplets. They're non-identical, but they share a common percentage so they are very, very similar. They're a perfect experiment. If you get male co-twins, you can use one as a control for the other. It cuts down on costs and really gets rid of a lot of the variability because marmosets are extremely variable. Another co-twin study, five co-twins, treated for 14 days continuously with M.B.P. and looked at the end of this period, there's no effect on testosterone. So, does this disagree with what I just showed you? No, it doesn't because when you look within the testes, things look completely different. They look as though there's more latex cells present. When you do the analyses and this was in Nina's paper, if you look at latex cell number, these are the co-twins, so the control and the blue is the M.B.P. exposed. Four of the co-twins increase their latex cell number and the one that doesn't, in true variability fashion, what that does is hypertrophies its latex cells. Don't ask me why one marmoset does it different to the others. The bottom line is that when you look at latex cell cytoplasmic volume protestus, it's significantly increased. It's increased in every M.B.P. exposed animal. So the interpretation of this would be that M.B.P. suppresses testosterone production. I've shown you that it does on its first dose. That will cause reduced testosterone negative feedback so you get a compensatory increase in LH. We can't prove that because there's no assay for measuring LH in the marmoset, but we know that LH will stimulate latex cell proliferation, or that should be hypertrophy, not hyperplasia. And this actually fits with what is also some association data in the human during this period at three months of age. And I've heard about other data that might actually support that as well. So what this is suggesting is that there's something different between the fetal testis and the neonatal testis, but would that also be true for the human? I think there's certainly a possibility that it is. We don't know why there's that difference. And we've been going back and asking what might happen to that coop mechanism, for example. We haven't looked at that yet in the marmoset, but we can do that. So we've got HCG from the presenter. That coop repression might be a factor to a greater or lesser extent. We don't really know that. We can't tease that apart in the human. LH becomes a player, but certainly after birth, the only player is LH. And the only difference that I can tell you between these two systems is that HCG is the stimulus here and LH here, they're both working through the same receptor. And I can't imagine why they should be any different, but maybe they are. We'll look at this in our xenograph system. For example, we might be able to treat with LH, not CG, and see if DPP would have effects then. So let me switch gears and move on to germ cell effects. Well, I'm not sure they are germ cell effects, but thinking about effects that might be manifest in the germ cells. This is, you've heard about multi-nucleated germ cells from Kim. I'm talking about this aggregation which we think is, this is Cytoplasm stained in brown, which is, we think it's the result of withdrawal of Cytoplasm, so that it's probably a Cytoplasm effect, but we still don't know the mechanism for that. But it's nothing to do with CUPTF2. CUPTF2 isn't expressed in any of these cells at any stage. So we've looked in the xenographs and asked, do we see germ cell aggregation in the human? And the answer is, we do, but it's really quite sporadic. So it's not something which is hugely consistent. It's not like what Kim was showing you for the multi-nucleated gonocytes. We see it in occasional xenographs, in occasional hosts. We never see it in the controls, that's for sure. But it's not something that's robust enough that you could hang on. But it would suggest that there are effects. But what were... In fact, the reason we developed the xenograph system to begin with wasn't for any of the things that I've been showing you. It was to actually look at how we might get a handle on the origins of testicular germ cell cancer in the human. And what this is showing here is what happens to the germ cells as you go through gestation in the human. So that when you start at nine weeks, this is fetal life the F stands for, and this is postnatally, that all the germ cells are pluripotent and expressing op4, and none of them are expressing what we would call markers of germ cell differentiation, VASA or MAJ4. And as you go through gestation and out into the postnatal period, so these gradually transform into these. And if you look at where we've taken a fetus at nine weeks of gestation, a first trimester one, and grafted that and then looked six weeks later, and this is what you're seeing here, then we would expect that the proportion of op4 positive cells would have dropped, and we should see at least some of these cells beginning to emerge. So at nine weeks, this is the pre-graph control. All the germ cells are op4 positive. There's none expressing the markers, and you probably can't see it very well, but you're beginning to see some cells emerging here. And to cut a long story short, if you look in later gestation, et cetera, you see this progressive differentiation of germ cells occurring as far as we can tell, normally in the xenografts. So the question that we wanted to ask is, well, what could make that go wrong, because that would take you down the direction of testis cancer. And for example, does DBP have an effect? So here's the sorts of staining that you end up with. So this is a xenograft now, six weeks after collection, after grafting, and the host has been treated with vehicle for the last three weeks. And we're now looking at two germ cell markers, red, op4, so these are the undifferentiated germ cells. Blue is MAJ4, which are the differentiated, and then we're also looking at proliferation marker Ka67. And you can see, this is a seminiferous cord, but another one here, another one here, that what you see is absolutely typical of what you see in the human and also in the marmoset, in that within the same cord and indeed, right next to each other, you can have a differentiated germ cell and an undifferentiated germ cell sitting there quite happily. You won't see that in the rodent. Everything is sort of synchronized in terms of their development. The other thing that you'll see is that when you look at proliferation, it's only the op4 positive cells that are highly proliferative. About 60 to 80% of them are proliferating, whereas the differentiated germ cells are quiescent. That is the same as you see in rodents. So what happens if you're exposed to DbP? If you're going down the cancer route, then what we would expect to see is that proportion of op4 positive cells is going to be increased because some of these cells here would have started off as op4 positive at the time of profiting. So if we prevented that, we would expect proportions to change. I'm only showing you one example, and quite honestly we're still to do all the detailed quantification we're in the process of doing it. I don't see that. In fact, you could interpret this as saying that when you look at the proportions, there are, the proportions are this far more of the differentiated germ cells versus the undifferentiated. And that we've actually advanced differentiation. Well, I don't think that is the case. I think what has actually happened, and I can't show you the data directly, is that we've actually lost op4 positive cells. It's not effective differentiation. You've selectively depleted the op4 positive cells. Certainly if you look at the ratio, the ratio changes significantly, but if you look at the absolute numbers, it's not that the numbers of these have gone up relative to these. It's the fact that the numbers of these have actually gone down relative to these. And the reason that makes some sort of sense is that's exactly what you see in the rat. Even though the rat develops its germ cells differently, what it does do, it has a period when all the germ cells are op4 expressing. And this then switches off, it switches off synchronously, and these germ cells then cease proliferation, and they don't start up again until they're spomatogonia. And it's during this period here, when you're exposing to TBP, that you see a big reduction in germ cell numbers. You see a bit of a climb back here because you also actually delay switching this off. So in fact, that you partly compensate for this by prolonging proliferation. So this may be the same or similar effect that we're seeing in the human. And we haven't really got the equivalent data for the marmoset, because we would need to be looking in fetal life. And although we've exposed in fetal life, as I've shown you, we were only looking at the animals after birth. But it does remind me that we did find something odd in two out of, was it six marmosets or five? I can't remember how many we looked at at birth. Let me orientate you by showing you this here. This is an MbP exposed animal. It's not a control. But here's a seminiferous cord. Everything that you see where there's in cytoplasm is a germ cell. And then here, here's our oct-4, our pluripotency marker. And you'll see that the majority of germ cells are actually oct-4 negative. I haven't got the vasa on here, but they're vasa positive. They're differentiated germ cells. But there's these bunches of cells that are pluripotent. We've not seen this in any controls. But it's true in all this sort of thing, the number of animals that we've looked at, we've probably only ever looked at maybe about probably under 20 control animals in this period. And we've never found these. They express all the pluripotency markers. So is it abnormal or is it an extreme variation of normal in a particular animal? Well, this is the problem when you've moved down to these species where you're actually limited in numbers. We don't know. We don't know if that's any significance at all. All we know is that the animals that we have followed through to adulthood, they don't have testis cancer. They don't have oct-4 positive cells. Certainly when we look at the distribution, the percentage of germ cells which are still undifferentiated at birth, or just after birth in the marmoset, you see that it ranges from none up to about 20, just shy of 30% in controls in a similar range of the MVP exposed, apart from this one animal here, which is up at 60%. And this is what this actually wasn't the animal that I showed you in the previous slide. So that's all the data that I'm going to show you. Just to round up at least the major conclusions that I would reach, that we have no evidence that steroidogenesis is suppressed in the fetal primate testis that's exposed to HCG in situ in the marmoset or in the xenograph system in the human. No effects of DbP or MbP. That we've identified a likely mechanism, a plausible mechanism, I should say, the persistence of Coup Tf2 that looks as though it explains the effects in the rat and why you don't see those effects in the mouse. And it doesn't look like that system is being affected in the human, but we need to work that up completely. Conversely, I've shown you that steroidogenesis by neonatal marmoset testis is affected by MbP exposure. Therefore, will it be the same for the human? And that it doesn't prevent germ cell differentiation, which would be the biggest concern, but it does induce loss of onc4 positive germ cells in xenographs as occurs in vivo in the rat. So I think one of the consistent messages that's coming over from my talk and Kim's talk is there are germ cell effects in the human xenograph situation. What the consequences of this are, I can't tell you. I'm not sure that there would necessarily be consequences other than you're going to have a shortfall of germ cells for a while, but they could easily get compensated for. And I think that's it. And I'm probably not ready for the questions, but I know I'm going to get them. Thank you, Richard. Questions? Richard, thank you very much. Richard, what do you think would happen if you stimulated rat cells, put them into your xenograph model, if you stimulated them with CG, co-ionic gonadotrophin? Well, the answer is I don't know. I presume that you would stimulate steroidogenesis. Whether you would then get the effects of DPP, I don't know. I think my concern is that as I sort of try to emphasise, when we're putting in fetal testis xenographs, to me it's a suboptimal system. We've got some things which haven't got a patent blood supply. They may be, if you like, teetering on the brink of survival, come along and stimulate the lating cells to actually go about their business. You may tip them over the brink. So I don't know how well you would control it. The ideal thing would be to be doing that in situ in the rat to actually be treating with HCG. But of course, HCG won't cross the placenta, so you can't get it into the fetus. But I guess behind your question is not the concern, is the explanation for why there's no effect in the human and the marmoset in fetal life because they're exposed to HCG. And I think the answer is it could be the case. But that is the physiological situation. It looks like you got all the tough questions, Kim. I think maybe we're food deprived. ATP levels have dropped precipitously. Just a clarifying question. Did you tell us the doses you were using? I don't remember seeing. Or for the majority of the studies, it was 500 milligrams per kilogram per day. Maybe the DPP or MVP. Carry on. The similar question to you as I asked Kim. Your human fetal test material is also fairly old and outside the programming window. You have no concerns that this might unduly skew your results? Well, I think as the same was said when you asked a question to Kim, I don't think we can exclude the possibilities that there could be effects on first trimester during that critical programming window. I think what we'd have to say is that it would run counter to all the evidence we've got in front of us because it would have to be via presumably a separate mechanism to what we're currently talking about. It would be inconsistent with the explant data. It would be inconsistent with the marmoset in vivo data. And therefore, it seems like clutching at straws to explain it. But the bottom line is you can't rule it out now. And as we've already said, science makes fools of us all. So there could be a completely different mechanism. I found it highly interesting what you told us about COOP TF2 and how DBP influences this. It does have the effect in the rat but not in the human or... Can you offer or you touched on it yourself a little? Can you speculate on a mechanism? Why would it be that the effects of DBP in these two species on COOP TF2 are so different? I have no idea. I mean, if you search the literature for what regulates COOP TF2 then you're probably going to get nothing. It's got a huge array of things in its promoter so that there are lots of things that are potentially important. One of which is GATA4. And GATA4 is involved in, is critical for specifying both development of satelis cells and lating cells. And I can tell you that GATA4 expression is affected by DBP in a time-specific way. I don't know that that offers an explanation for the COOP and it's not so easy to actually do those sorts of experiments. To actually get handles on these sorts of things involves making transgenics or doing transfections and a lot of those are quite messy. Because what I found really interesting was that in the rat if you treat with DBP even outside the male programming window you can actually switch COOP TF2 back on. That I think is the most provocative of your observations but for some reason that doesn't seem to happen in the mouse or in your human fetal xenographs. But, well, I think it would be absolutely crucial to know why that is. Yeah, one of the things that I did say is that there are other things that will keep COOP TF2 switched on in the lating cells. And one of those will do it in both the rat and the mouse and with associated reductions in testosterone. So, so far we haven't found a situation where there's a discordance between COOP TF2 expression and the maintenance of stereogenesis. The big problem with that whole thing is the only way you can look at it is the way that we've looked at it. Because COOP TF2 is expressed at such high levels in non-ladic interstitial cells that if you look at the message level or the protein level protestus, you don't see any change whatsoever. Interesting difference that, well, it wasn't a difference exactly, but you did see COOP TF2 in your human fetal testes, samples. It was lower, but then you were using fetal testes, samples that were not from second trimester, not first trimester. Interesting to see if you got some early first trimester samples and you could look at that to see if there again you have very high levels like you do in the rat. That would be interesting. Because if that's true, then you begin to wonder why you're not seeing similar results in the human. And then you're replicating the rat situation, but you're not seeing the other effects. That would be interesting. Well, as I was telling Kim earlier as well, that it certainly isn't for adult-ladic cells. It's not been looked at for fetal-ladic cells. And COOP TF2 is a repressor of LH receptor expression. So another reason for removing repression, and it works time-wise for the rat, is to actually allow LH receptors to be expressed. So how that works for the human, I don't know. And maybe once CG starts working through the LH receptor, if it is expressed, maybe that helps turn COOP TF2 off. But I mean that's pure speculation. With the results you've shown and the experiences you've made, do you think we are too narrow-minded to just look at the male programming window in the prenatal phase? Do you think we might also have to take into consideration the possibilities of influence in the postnatal phase? Yeah, I think that would be a very real concern. I think the problem is that we can't sort of frame that concern very accurately because we don't have a very clear picture about what testosterone is doing in that neonatal period. We know that it's probably helping drive proliferation of Satoli cells. We know that it's also playing a role in penis growth, certainly in the human, but I think it's also true in the rat. But I think my concern would be also whether or not there might be, for example, central effects on brain development. And it's some sort of way rather left-field thinking in that, in that, for example, in preterm babies, you get premature activation of that process. So therefore, you could say that what you're getting is exposure of the brain to abnormally high levels of testosterone for fetal age in those. And they are at greater risk of autism, which is people often saying that that's associated with overexposure to androgens, although I've never been completely convinced by it. It just raises that possibility. And I'm not saying, I wouldn't say any more than that it's the fact that it's a possibility, that people haven't really been able to tie down what androgens are doing at that time because you really don't have any access to collecting samples or doing studies in children. You haven't got a rodent model that you can actually look at. The only thing you could look, for example, would be in a marmoset. And then, how are you going to then, what aspects of behavior, et cetera, are you going to look at and when? So you're starting really with no information. But I would be potentially concerned if our marmoset studies are directly translatable to human and that there are effects of phthalates potentially in that period in the human, then I think there is that concern. It's an unspecified concern and it may be without foundation, but... Can I elaborate a little? So the suppression of testosterone synthesis in neonatal life in marmosets by MVP, used MVP, can you offer a mechanism for that effect? Why would that be? No, I can't at the moment. As I've said, if you'd asked me to predict based on fetal life, I'd have said there, categorically there would be no effect because I would think that they were fetal type-lady cells. Actually, I've just realized if I can just go back to the previous question because I'm sort of thinking on my feet, actually, maybe that concern that I just raised is not realistic because what I also show you from the marmoset is what the limited human data also suggests is that if you do suppress duogenesis in that period, you get compensation so that you don't actually end up with a shortfall of androgens because the feedback loop is working, so you correct it. So maybe you don't get, so then you have to then say, could there be consequences for that, of that compensation? Maybe within the testers and I think then we're into the realms of complete speculation and unknowns. I have fairly general questions now which concern me in particular in the light of the evidence both of you presented today. We have to go back a little. If you look at the literature, we have this fairly well-established structural activity relationship. First, I think proposed by Paul Foster about salads and which of them are active and which aren't and the result is that side-chain length of carbon four, between four and six is important. And if I'm not totally wrong, that idea was actually established on the basis of studies in the rat with what you, Richard, would describe as the old paradigm or the old thinking. Now you throw a new light and a new perspective on all this. Would that not throw into doubt the validity of this structural activity relationship in the human as well? The reason I'm asking this is there's no clear evidence, but there are signs, there are warning signs in epidemiological studies that a phthalate like DEP may actually induce adverse effect in the male offspring. Cryptocardisms have been described in one study by Ormond et al. conducted in London. There was this association among male babies from hairdressers, and if you look at what hairdressers are mostly exposed to, it's none of the phthalates you are studying, Richard or Kim, it is mostly DEP. There are other, the evidence is indirect, but we know in the rat DEP doesn't do anything. So do we have to rethink everything in the light of both your observations and invite some comments from the two of you? Who's going to do DEP? That would be an obvious thing to do, wouldn't it? It has been, I think there were actually earlier studies done by Tim Gray that were not in fetal exposures, but were germ cell responses in sort of pubertal testis where there was an effect related to phthalate chain length where that first came out, and that was in the 80s, and there was a consistent observation of effect there. That's been sort of one of the dictums of faith in phthalate literature is around that structure activity relationship, but I think it's a reasonable question. We haven't looked. Everything becomes possible. I guess I'd just add, I'd be surprised. I'd add that. Because it has been very consistent in terms of the structure activity has been very consistent in what we have seen. Well, rat, and I think in, well, I have to think about that some more in other species. I think there was some work in ferrets, but in the fetal period, more in adult studies, there's been other species. Did people ever look for germ cell effects with DEP, do you know? Well, again, the sort of early Tim Gray studies were, he looked across. I think he looked at DEP, but he looked at a number of different phthalates. And there was a very clear structure activity relationship. But did that include looking at germ cells? Well, there wasn't multi-nucleated germ cell, pop off from co-cultures. That was his essay. So, on a quantitative basis, we see effects in rats. But we also see effects in mice and marmosets later in life. Do we see effects in marmosets that are related to the endocrine potency of some phthalates? Not in fetal life. No. At any life stage. Well, I showed you the results for M.B.P. in the neonatal period. So there are effects. We've never looked in adults. I don't know, maybe you'll know about the literature as people look at other ages. Right, so in adults, you basically don't see lighting cell effects to a significant extent in adult rodents. So that effect seems to be quite strong in fetus. Richard was making the point earlier, though, that the marmoset effect appears to be transient because of compensatory mechanisms. Right? So if you, and you could argue that for anything that inhibits genesis, that what you're not going to see is an avert fall in testosterone levels, you should see a change in the LH to testosterone ratio that tells you you've got what we call compensated latic cell failure. That's what I'm trying to get at. So we see effects in marmosets, in the postnatal life. Yeah, in that postnatal period. In the postnatal period. And there are effects in vitro in adult human testes on steroidogenesis. Exactly. Which is again, which is a bit different to rodents. Now I don't know about in your studies, Russ, what do you see in your guise? Do you see any evidence for compensated latic cell failure? Some associations mostly with testosterone and DEHP and I've also looked at the LHT ratio. Primarily in relation to MEHP that's what we've seen. But you see some evidence? Yes, some evidence. So there's maybe some consistency. And Hain's data may have looked at that as well. So what I'm trying to get at is in your point of view, do we have to regard the phthalates or some of the phthalates as being active in one or the other way you've shown in your presentations for humans? I don't think you could rule it out at the moment. I would say it's unlikely to occur in fetal life. I'm talking general. I'm talking effects in humans not only in fetal life. I would think based on the evidence we've got in front of us we'd say that probably there is the potential for effects, yes. I guess the difficulty always is when we're in our animal model situations we're using very large doses. So there's a very big step down for that to the human exposure studies. But if that's what the human association studies if they come up with similar data to our marmosets at environmental levels then I think you would have to, although you're not seeing effective testosterone suppression you're seeing evidence that it has happened is ongoing I think you would have to interpret that in the light of the animal model evidence and say that it's suggestive that there is an effect ongoing. The human data, also I'm just remembering there's another study from China in which it was occupational exposure to DEHP and lower testosterone but and then some studies that we've published with Shauna Swan where we put together our infertility clinic population with men from the general population and saw that association but the levels statistically were lower it's not that the men were in basically hypotestosterone or a clinical range of abnormal levels they were slightly decreased statistically. One of the questions is that you're going to have to grapple with is the certainty of the evidence at different for different scenarios and different time points so I look at the data that Richard and I have generated on the xeno transplants and around the fetal period because that was the focus of the major concern when we started doing our work and it's consistent and complementary really quite different approaches using the same kind of idea and yet the results are quite consistent so there in terms of human experience there's enough data there to have some certainty around the fetal period with the evidence that we have I think we have much less from an objective around adult responses although on general physiologic principles you might anticipate that there be much more capacity to compensate for any kinds of effects in adult. Richard can I come back to the germ cell aggregations which you see in the human fetal testis xenographs what I wasn't entirely clear about your view of the significance of this what do you think is the significance of those aggregations? Well I guess the only reason I was presenting it was that it's another little bit of evidence that there is exposure of the graphs and effects but beyond that I don't think we'd say because we really don't know what that aggregation we think we know what it might indicate but we don't know what the consequences will be because once exposure stops those germ cells disaggregate and as far as we know they don't apatose at least not in any major way so whether or not their developments is affected and compromised no idea I just couldn't comment on that it could just be something that's purely incidental although it might tell us something about a mechanism via which you would get effects on the Cetone cell at that stage in development And another sort of set of questions or maybe just one question for both of you all in terms of and maybe this address gets a little bit to what Andres was saying as well I mean we're after sets of phthalates not maybe just the sets that you all have looked at do you have any evidence that there are some sets that act differently than other sets are you speaking structural activity that put it in categories for you or are you lumped together phthalates we've never looked at anything other than dbp or mbp so I can't really comment in any informed way other than I'm a guest to say that what Kim said earlier that we're I think we've all been thinking in the one way that there's a consistency at least in the rat for the way which phthalates have effects which would sort of indicate that it's like to be a common mechanism but I guess another way of looking at that is to say that there that is a hell of a lot of presumptions in it and it's just because we haven't found any effects of the other phthalates that have been tested in the rat but as it's already been sort of indicated that does not preclude that there might be effects in other species via completely different mechanism yeah I don't really have anything add to that and what I said earlier the data in the rat is actually quite strong around structure activity, relationship and I'd say powerful and consistent honing in on certain chain lengths and combinations but that's in the rat and now we appreciate that there are quite significant species differences in some of the responses both if you had to make a decision about other sets a lack of data I mean how would you address that I think it's a dangerous move to make a decision about anything in the absence of data yeah I think it would be very nice in the Xenograph system if DEP was tested and found to be negative I think that would then it's not going to answer your question completely but at least it's going to say that it is negative in the human and in the rat I'd agree it hasn't been done with the magnitude of with the enormous amount of data you have now accumulated on this topic would you still consider that the delayed syndrome in animals is comparable to the human testicular syndrome in humans and is this possibly related to the late exposure I'm looking at your yes to the first question and I would say no to the second based on the evidence we've got I think that all along what we've been saying about using DBP in the rat is that we're using it as a tool to uncover mechanisms to then ask what factors might perturb those mechanisms other than DBP and we've identified a mechanism and we've identified also that other things do perturb including for example glucocorticoids so it could be a stress effect if you like that might work through it but there are other pesticides some pesticides look as though they may target that mechanism other hormones so those might be players in the human whereas DBP is not but I think you know it's without knowing what the sort of the pathway of a fact is that leads to COOPTF2 and to know in there for the sorts of things that you might be looking at what sorts of activities might impinge on those you're groping around in the darker bit but I think that's where even if the rat turns out to be not a good model in terms of health relevance or health relevant effects of phthalates in the human I think it's still got something to offer from the point of view of mechanisms which is the reason why we would keep looking at it unless there was something better came along okay I think we'll break for lunch Mike? Yeah wouldn't you come back it's one o'clock do we have constraints in terms of not for some time I haven't booked any means of getting to the airport yet I know that I'm not walking that's the only thing I mean it's at seven o'clock or something like that do you want it an hour or a little longer you want to do it come back at two reconvene at two o'clock