CME Slides Forum - S. Maynard

A joint Congress by ERA-EDTA and ISN

VEGF ANTAGONISM AS A CAUSE OF PREECLAMPSIA

Sharon Maynard, Washington, USA

Chair: Duk-Hee Kang, Seoul, Korea

Detlef Schöndorff, New York, USA

Maynard

Dr Sharon E. Maynard
School of Medicine
George Washington University
Washington, D.C., USA

I’d like to start by congratulating the conference organisers for really pulling together a beautiful set of speakers. I think these talks dovetail so beautifully and together make an absolutely compelling case for the role of VEGF locally in the glomerulus in endothelial health. So you’ve heard about animal studies, elegant animal studies showing that a knock down of VEGF produced endotheliosis, endothelial damage, proteinuria. You’ve heard about a human experiment where human beings who have cancer receive anti-VEGF treatments and develop hypertension and proteinuria. I’m going to be telling you about a spontaneous human disease where there is transient VEGF blockade and there is transient glomerular endotheliosis and proteinuria which resolves after delivery and after the VEGF blockade is alleviated once the placenta is gone. So let me bring up my talk.

So you see my disclosures there, VEGF antagonism is a cause of preeclampsia. So first just to tell you why in trying to figure out the cause of preeclampsia what got Ananth Karumanchi and me by his side interested in sFlt where soluble VEGF are one in particular. Well preeclampsia first of all for those of you who don’t know is characterised by the new onset of hypertension and proteinuria in a pregnant woman in the second half of her pregnancy. It can also affect other organs such as the brain where it can cause seizures and cerebral oedema. It can affect the liver where it can cause liver injury and the HELLP syndrome but really all these manifestations are unified by the presence of widespread internal vascular endothelial dysfunction.

But even though the target organ in preeclampsia is the maternal endothelium, the source of preeclampsia, the origin has long been thought to be in the placenta. There are a variety of reasons for that that I don’t have time to go into, but that fact has led to this paradigm for preeclampsia whereby the placenta is thought to secrete some sort of circulating factor which present in the maternal circulation causes the endothelial dysfunction and for many years people were trying to figure out what this mysterious preeclampsia factor might be.

Ananth Karumanchi’s idea was that if this was a secreted protein made by the placenta, it might be detectable as being upregulated based on the mRNA level by microarray analysis. So we performed what was really a very simple microarray experiment, we collected placenta from women with and without preeclampsia and we looked to see which genes were upregulated in preeclampsia and that’s where we found sFlt or VEGFR1. So here you just see the overexpression of sFlt, VEGFR1 in the placentas of women with preeclampsia and I’m going to show you this is a highly stylised cartoon of Flt. You know my nomenclature is kind of old fashioned so throughout my talk I’m going to be talking about Flt-1 but Flt-1 as Doctor Kurgan mentioned is synonymous with VEGFR-1 one of VEGF’s receptors. So Flt-1 has an extra cellular ligand domain a transmembrane domain and an intracellular signalling domain. As Doctor Kurgan mentioned Flt is thought to be a decoy receptor and VEGFR-2 probably mediates the signalling of VEGF but Flt binds both VEGF and placental growth factor sister molecule whose function has not been really been well described and of course, VEGFR-2 also binds it. Soluble Flt is different from Flt in that it’s a splice variant which has been truncated and it consists of the extracellular ligand binding domain but lacks the transmembrane domain and the intracellular domain that link it to the domain and allow it to function as a VEGF signal this signalling molecule. Because of that sFlt binds to VGEF in the circulation and acts as a sink preventing VEGF from interacting with it s endogenous receptors. So it acts as a circulating VEGF antagonist.

So we knew this about sFlt when we got back our initial microarray results. Now every great discovery I think has an ‘ah ah’ moment and the ‘ah ah’ moment for this one was when we saw right about the time we were getting our microarray results the first data was coming out on the effects of anti-VEGF therapies in cancer patients, which you just heard about from Doctor Alpers and Doctor Azizi. The observation that a molecule that inhibits VEGF and can cause hypertension and proteinuria in these cancer patients made us confident that we were really on to something here and that a VGEF inhibitor might cause the same symptoms in preeclampsia. So our hypothesis was that increased systemic sFlt from the placenta leads to antagonism of VEGF and/or placental growth factor in women who have preeclampsia thus leading to its widespread endothelial effects including hypertension and glomerular endothelial damage.

So we set off to test this hypothesis and the first thing we did was measure sFlt Levels in the circulation of women with preeclampsia. We found that sFlt levels were indeed high in women with mild and especially severe preeclampsia compared to normal pregnancies and that the levels fell dramatically after delivery. You can see here just prior to delivery and 48 hours after delivery sFlt levels in the circulation fell corresponding with resolution of the hypertension and proteinuria in these women.

So we took this to animals, in this case pregnant Sprague-Dawley rats. We administered adenovirus expressing sFlt to pregnant rats in their early second trimester and then we measured their blood pressure and their proteinuria 8 days later in the mid third trimester and we harvested their kidneys. So what we found was that sFlt induced significant hypertension in both pregnant and non-pregnant rats who received sFlt, as compared to control rats and they also developed heavy proteinuria both pregnant and non-pregnant rats that were administered sFlt developed heavy proteinuria compared to controls.
Pregnancy wasn’t required for this to be observed because we were in fact bypassing the placenta. We were giving the sFlt directly. This tells me that the placenta mostly is important as a source of sFlt in preeclampsia disease.

So looking at the kidneys of these rats I just have time to show you the electron microscopy but these images will be familiar to you because you’ve seen a lot of glomerular endotheliosis in the last hour or two. So this is the control rat, this shows you normal open capillary loops with nice thin fenestrated endothelium

and in the rats that were treated with sFlt we saw endotheliosis. We saw marked endothelial cell swelling with obliteration of the capillary lumen, loss of the fenestrated endothelium and very key here was that podocyte processes were preserved in most diseases, human diseases with nephrotic range proteinuria there’s effacement of the podocytes. So this TMA this is glomerular endotheliosis, the characteristic lesion of preeclampsia.

So we wanted to see whether sFlt was up prior to the development of preeclampsia because there are a lot of derangements in circulating molecules, cytokines, oxidative stress, once preeclampsia is established all sorts of things are altered but if we could show the differences before the onset of disease, then that would be at least a little more evidence in favour of our hypothesis. So we collaborated with Richard Levine who had stored serum samples from women, some of which had gone on to develop preeclampsia. We measured sFlt levels in these women. So the blue line is sFlt levels in normal women, you can see that it rises towards the end of pregnancy. These black dots are serum samples that were drawn from women after the onset of hypertension in proteinuria, so women with full blown preeclampsia, you can see marked elevations in sFlt. The red line are those same women who went on to develop preeclampsia but these blood samples were drawn prior to the onset of hypertension or proteinuria. You can see as early as 25-28 weeks gestation there is a significant increase in sFlt levels but you know this wasn’t present from the beginning of the pregnancy, we saw this change maybe 5-8 weeks before the onset of symptoms.

The other kind of circumstantial evidence that supports the sFlt hypothesis comes from looking at a few different risk factors for preeclampsia. Most preeclampsia occurs in healthy women without any risk factors but there are several risk factors that are strongly associated with preeclampsia and I like to group them into 2 different categories actually. The lower half women with hypertension, with diabetes, with obesity probably have something that leads to increased maternal susceptibility to preeclampsia.

But there are some others that don’t really have anything to do with the mum herself or may not. First pregnancies in women with multiple gestations like twins and triplets are at increased risk for preeclampsia. Mothers who are carrying pregnancies with trisomy 13 are at increase risk for preeclampsia and these are really foetal placental risk factors. I’m going to talk about these first 3 because you might expect with these placental risk factors that women who have these conditions might have alterations in their circulating sFlt levels and in fact they do.

This first study I’m going to show you is from Ravi Thadani and Myeles Wolf and they simply looked at sFlt levels in serum samples drawn from a group of women in their first trimester of pregnancy, from their first pregnancy and then from their second pregnancy. So the same group of women gave blood samples twice during their first pregnancy and their second pregnancy. He found higher sFlt levels in first compared to second pregnancies which may be part of the reason why women who are having first pregnancies are at much higher risk for having preeclampsia. We don’t why sFlt levels are up in these women but it’s one thread and it certainly supports the case for sFlt.

Women who carry babies with trisomy 13 are also at increased risk for preeclampsia and Ananth Karumanchi and some colleagues noticed that sFlt, the gene encoding sFlt and Flt happened to be on chromosome 13. So the foetus and the placenta having 3 copies of chromosome 13 may have an increased sFlt burden which carries over into the maternal circulation. That’s exactly what they found. Here they are showing actually the sFlt/PLGF ratio which I don’t have time to talk about PLGF much but it falls when sFlt levels are up and it’s also a very accurate predictor of preeclampsia and you can see the ratio is much higher in women with foetuses affected by trisomy 13. Perhaps this is the reason why they’re at increased risk for preeclampsia.

Finally women with twins and triplets are at increased risk for preeclampsia. Maybe it’s just that they have more placental tissue and more placental tissue means more generation of sFlt. These are circulating sFlt levels drawn in the late second and third trimester from women who had singleton twin and triplet gestations and clearly you can see that women with twins and triplets had much higher circulating sFlt levels. None of these women developed preeclampsia, we excluded women with pre-eclampsia from this analysis, but it shows you why they may be at increased risk, just the increased sFlt burden.

Finally you know smoking has kind of a counterintuitive association with preeclampsia. We try not to make it advertise it too widely but women who smoke actually have a reduced risk for preeclampsia. We don’t want patients to know this because we don’t want to give smokers any reason to keep smoking during pregnancy, it has a lot of other adverse effects on the developing foetus. But what we found this is a sub-analysis from Richard Levine’s data was that women who were speakers from the various time points during pregnancy consistently had lower sFlt levels than women who were non-smokers. This may somehow mediate the protective effect of smoking. Somehow smoking decreases endogenous sFlt production.

So, I hope I’ve given you some data and certainly the data presented to you by the previous speakers I usually include in my talk as supporting evidence for the role of sFlt as linking the placental dysfunction with the maternal endothelial dysfunction. sFlt is certainly not the only player here we know for example, that soluble endoglin another molecule secreted by the placenta that Doctor Karumanchi has studied is also part of this pathway. There probably are many other molecules as well but you see most of the question marks here are upstream. What is it that’s causing, that’s inducing increased sFlt production in some women and not in others? I’d like just to take 3 more minutes to talk to you a little bit about some exciting data that’s been published within the last year that’s give us insight related to angiotensin-1 receptor autoantibodies and their role in preeclampsia.

So angiotensin receptor agonistic autoantibodies were first described as being increased in the circulation of women with preeclampsia about 10 years ago and multiple subsequent studies have validated it. So it was hypothesised that they may be part of the pathogenic pathway of preeclampsia. So in this absolutely brilliant study by Yang Zhou and colleagues published less than a year ago they decided to test this by isolating IgG from women with preeclampsia and from women who have normotensive pregnancies and then injecting that IgG into pregnant rats to see what would happen. What they found was I’m just showing the hypertension result here that the mice who received IgG from women with preeclampsia developed severe hypertension compared to the mice who received IgG from women who did not have preeclampsia and that this effect was blocked by losartan, an angiotensin receptor blocker, these agonistic antibodies work through the AT1 receptor and hence would be blocked by losartan and they were also blocked by administration of a 7 aminoacid epitope that specifically will block the AT1 autoantibodies showing that it is these autoantibodies that mediate this effect.
In the kidney pathology in these mice was glomerular endotheliosis, the same lesion that I showed you in my sFlt treated rats and the same lesion that you saw in the previous speakers’ slides.
So you might think that at first this is a competing story from the sFlt story. Inject sFlt into rats and they get preeclampsia like syndrome. Inject agonistic antibodies into rats and into mice and they get preeclampsia syndrome but there was one key difference in the experiments that they did and that is when they treated non-pregnant mice, the non-pregnant mice got hypertension but they did not get the renal lesion, they did not get proteinuria and when they looked at their kidneys they did not get glomerular endotheliosis.

So they went on to look at sFlt in these animals that were treated with AT1 receptor autoantibodies. They found in fact that pregnant rats that were treated with AT1 receptor autoantibodies, the IgG from women with preeclampsia had an upregulation of sFlt. Their sFlt levels were increased in mice that were treated compared to the control mice. These mice obviously developed hypertension and proteinuria but when non-pregnant mice were treated, they did not have any increase in sFlt production because they didn’t have any placenta.
So this is why these mice developed hypertension but no proteinuria and the authors hypothesised that it is the sFlt that mediates the glomerular lesion and that the AT1 receptor autoantibodies increase sFlt production by the placenta and that it’s this endogenous sFlt that then leads to the glomerular disease.

So in conclusion VEGF antagonism and preeclampsia, excess placental sFlt mediates maternal endothelial disease in preeclampsia. VEGF blockade by sFlt1 in the glomerulus results in endothelial damage and AT1 receptors autoantibodies may contribute to pathologic upregulation of placental sFlt.

Chairman: Thank you very much for that very nice and very lively talk. Can we have questions or comments from the audience?

Dr. Maynard: During the pause I’d like to acknowledge Ananth who’s the mastermind of this work and also my other collaborators. Thanks.

Question: Can I ask you, is there any evidence that these antibodies that are generated directly trigger message for sFlt and not for membrane bound Flt?

Dr. Maynard: The antibodies in particular? Well, first let me say that in preeclampsia it appears that both sFlt and membrane bound Flt are upregulated. But it’s only the sFlt that is secreted into the maternal circulation and has the end organ effects on the kidney. In terms of the mechanism by which the angiotensin 1 receptor autoantibodies induce sFlt, it hasn’t been studied at all yet but it’s been hypothesised that by stimulating the AT1 receptor in the placenta perhaps they produce placental ischemia and it’s known that hypoxia and activation of HIF, hypoxia inducible factor will lead to sFlt upregulation but there’s no experimental studies so far looking at that.

Question: Thank you for a nice presentation. I have a question. An elevation of uric acid was known as one these phenomena in preeclampsia way before the development of the proteinuria and the hypertension. Do you have any comment about the clinical or experimental evidence linking the hyperuricemia and sFlt1, the production from the placenta?

Dr. Maynard: I don’t have a lot to say on that. You know obviously outside of the preeclampsia literature there are some people who believe that increased uric acid may actually play a role in human hypertension generally in the adult. Certainly uric acid levels are high and sometimes clinicians use them but they’re not part of the diagnostic criteria because they do tend to be variable. But I don’t know of any specific studies that tried to tie together sFlt and uric acid.