Slide 1

Well, thank you very much Pierre for these very nice words and thank you for having invited me to this last plenary lecture session. I hope you will get something when you come home in terms of practical things to apply, if not today maybe hopefully tomorrow in your clinical practise.

Slide 2

Well I would like to discuss with you the issue of vasoactive peptides and the possibilities they may be involved in angiogenesis. Of course, as you know angiogenesis is a growth of new blood vessels from pre-existing ones. Angiogenesis and vasculogenesis are quite important during embryonic development, during exercise and also during the menstrual cycle. We need neoangiogenesis in order to heal our wounds and of course, angiogenesis is involved in several pathological conditions like retinopathies, inflammatory diseases and tumour growth and metastatic invasion.

Slide 3

Therefore, for a cardiovascular clinician and for a nephrologist I think it’s important to realise that neoangiogenesis is an important phenomenon in cardiovascular pathologies not only in acute and chronic ischemia but also in some other conditions like plaque rupture but in this case we need neoangiogenesis to elaborate collaterals or at least to elaborate new capillaries and in the case of plaque rupture it’s possible that an excess of neoangiogenesis is responsible to some extent of the rupture of the plaque. Of course, angiogenesis is undesired also as I said in tumours and retinopathies. Therefore, at the present time there’s a huge effort in the field of future therapies. At the present time a human VEGF antibody is used in colon carcinoma and it will be extended in other cancer conditions. The VEGF aptamer is used in some pathological diseases such as the degeneration linked to age in retinopathy and pro-angiogenic therapy has been used and proposed using VEGF and FGF and of course, there are a lot of studies going on with endothelial cell progenitors.

Slide 4

Now, let’s look at the basis of angiogenesis. Everything starts with a mesodermal precursor which leads to a bio potent cell called hemangioblast because this cell is about to give rise to both angioblast and the haematopoietic precursors and as you may remember VEGF is quite important here and gene inactivation of VEGF leads to two types of phenotypes first of all abnormal structure of the primary capillary plexus and second there’s a marked decrease in haematopoietic cell production. Then after there is the formation of the primary capillary plexus and this is the end of what is called vasculogenesis. Then after there is a maturation of the vascular network with the apposition of pericytes and vascular smooth muscle cells and we have the formation of this vascular network which will be remodelled later on during adult life by several actors like the two angiopoietins, angiopoietin 1 and 2 to the stage of the adult. It was supposed that the phase of vasculogenesis occurred only during embryonic life but in fact, it’s likely that progenitors of endothelial cells like hemangioblasts may circulate in adults and may contribute to the formation of neovessels in some conditions.

Slide 5

Now if we look at the major actors, which have been identified in angiogenesis through mainly gene inactivation, we have FGF, VEGF which is absolutely crucial with its two receptors, VEGFR1 and VEGFR2 to at the stage of primitive angiogenesis or vasculogenesis and then after the remodelling including several important steps and several proteins like angiopoietins-1 and 2 and its receptor and also PDGF, TGF-beta in order to elaborate the mature vessels through the formation of extracellular matrix. Some coagulation factors are also involved independently on the effect on coagulation like tissue factor and also like plasmin. So in this case we may wonder whether or not some peptides like the vasoactive peptides may also have an effect not on vasculogenesis but on angiogenesis. Why that?

Slide 6

Well, it was clearly shown that these vasoactive peptides can control vascular tone and blood flow delivery but in fact there are also growth promoting factors like angiotensin II that are involved in vascular remodelling and of course, exert many other functions. So the simplest and maybe too much naive question which we may ask is whether or not these peptides are also involved in angiogenesis/anti-angiogenesis balance.

Slide 7

Why that? Because after all you could think that by altering the blood flow by promoting a local and critical decrease in blood flow and therefore some degree of hypoxia they may contribute to the formation of neovessels. I’d like to discuss with you today at least 3 of these peptides and their receptors, adrenomedullin, endothelin, angiotensin II and I won’t have time to discuss the issue of bradykinin and its B2 receptor.

Slide 8

Now adrenomedullin has been discovered, as you know in pheochromocytomas. It’s a very potent vasodilating peptide, which is markedly expressed in the fetoplacental unit. So gene inactivation of adrenomedullin leads to a lethality during embryonic life at day 13.5 – 14 and lethality is due to severe haemorrhages and there is a detachment of endothelial cells from the basement membrane. The heterozygous mice survive, they have an increased blood pressure which shows actually the tonic effect of adrenomedullin at least in mice due to a decrease in NO production.

Slide 9

As you can see here from this work from Shindo in the adrenomedullin knockout mice at E14 you have severe haemorrhages in the embryo. You can see the haemorrhages here in the lung compared to the wild type mouse and also compared to a capillary lumen of a wild type mouse you can see here the thinning of the basement membrane and the detachment of an endothelial cell from the capillary lumen.

Slide 10

Now I’d like to discuss another peptide which we believe has also in some conditions an angiogenic effect and this is endothelin. Now like adrenomedullin, endothelin is not only a vasoactive peptide it’s also a very important peptide during embryonic development. As you may remember endothelin gene knockout, either the ETA or the endothelin converting enzyme leads to cardiac defects as well as  aortic ulcers defect and vascular wall abnormalities. Endothelin has been discovered by culturing endothelial cells under hypoxic conditions and therefore, the regulation of the endothelin gene is under the control of hypoxia inducible factor 1 alpha as well as other transcription factors. I’ll show that the endothelin system is markedly expressed in neovessels during carcinogenesis. We will see that endothelin has an effect, which is a typical effect on endothelial cells of an angiogenic compound, it induces cellular proliferation, cellular migration and this is NO dependant and it’s likely that the main receptor involved is ETB.

Slide 11

For example here is shown in all vascular endothelial cells the presence of the endothelin converting enzyme which is key for converting big endothelin into the active peptide endothelin and as you can see by in situ hybridisation in arteries or in arterioles in either elastic or muscular arteries you can see the expression of the endothelin converting enzyme.

Slide 12

And this is true also for the protein which is labelled here, not in the vascular smooth muscle cells but in the endothelial cells and we found it in several types of cancers.

Slide 13

Now we were interested in looking at the endothelin gene system expression in another clinical setting, which are pheochromocytomas, as you know pheochromocytomas are rare tumours of chromaffin tissues and 10% of the pheos are malignant. However, it’s always difficult to predict whether or not a pheochromocytoma is malignant and this was based mostly on clinical observations like recurrency, lymph node or distant metastases, which were, which are still of course good indicators of the malignancy. But in fact, malignant pheochromocytomas can be due to mutations affecting the state of oxygenation of the cell, for example, mutations affecting the VHL protein and also genes involved in the respiratory mitochondrial chain like the SDH, subunit B or sometimes D. There’s a marked angiogenesis in malignant pheochromocytomas.

Slide 14

And therefore, we looked at the presence of some of the angiogenic genes in malignant pheochromocytomas compared to benign pheochromocytomas and looking systematically at a series of pro-angiogenic genes you can see here in boxes which ones were found to be markedly expressed. EPAS 1 which is  HIF 2 alpha, HIF 2 alpha, VEGF and its receptor and the  endothelin system interestingly enough was also quite markedly expressed.

Slide 15

And as you can see here the vascular architecture of a normal adrenal gland versus a benign pheocromocytoma, there’s not much difference. In fact you have small capillaries indicated here but in malignant pheochromocytomas you can see a destruction of the normal capillary structure of the tissue as well as late formation, rupture of capillary vessels.

Slide 16

As you can see here again by  in situ hybridisation you can see here the marked expression of VEGFR-1 and VEGF is also quite markedly expressed not in the endothelial cells but in the tumoral cells but interestingly enough ETB is markedly expressed as you can see here on the endothelial cell of this malignant pheocromocytoma and you can see here a few endothelial cells over expressing ET1.

Slide 17

If you compare the amount of mRNA according to these different genes you can see a clear difference between benign and malignant pheochromocytomas for VEGF, ETB and for the HIF-2 alpha gene. Therefore, it’s possible to make a scoring and to try to predict whether or not you are dealing with a malignant versus benign  pheochromocytoma. Of course, this doesn’t mean that endothelin is angiogenic so we had to use another model to show it.

Slide 18

The model I’m going to describe briefly is a model which we like in the lab because it’s easy to do, it’s quantitative, it’s a very well established model for the study of cardiovascular development as well as haematopoiesis and it’s important to have always in mind that angiogenesis and vasculogenesis go together during the embryonic phase and this is a chick embryo system. You can see here the chick embryo at day 5 and then at day 12 after incubation. You can study the chick embryo within the shell or you can make a shell-less culture and this allows you to study many compounds applied directly on the embryo. You can also, microinject within a vessel a compound of interest, for example, an avian retrovirus according to a gene of interest or an SI and you can quantify the images obtained from the vessels after the injection of FITC strand.

Slide 19

Now doing this we looked at the effect of endothelin. We could not apply directly endothelin in the chick embryo because there is a strong vasospasm which occurs and which precludes the study of neoangiogenesis.

Slide 20

So what we did was to manufacture a cell, a CHO cell where all the ingredients for making endothelins are present, the endothelin 1, the EC1 gene and we culture those cells just above the chick embryo. There is no immune reaction in the chick embryo as you may know and therefore the cells are well tolerated during the days of incubation. After 5-6 days one can study by light microscopy or in situ hybridisation or of course biochemistry the effect of the application of the recombinant cell producing endothelin.

Slide 21

As you can see here compared to CHO wild type you can see the CHO producing endothelin is able to provoke a marked angiogenesis in a few days. It’s a typical pattern called a spoke wheel pattern of neoangiogenesis.

Slide 22

Now we looked a little bit more carefully and as you can see here compared to the wild type, the CHO over expressing endothelin leads to a nodule formation just under the chorioallantoid membrane ectoderm and you can see a marked increase in the vascular density as well as a marked increase in the hypoxia inducible factor 2 alpha and there’s a strong endothelial cell proliferation as shown by BrDU immunostaining.

Slide 23

This is really due to endothelin because you can block neoangiogenesis by an ETA receptor antagonist but also by an ETA/ETB mixed receptor antagonist or even by inhibiting the ECE by phosphoramidon.

Slide 24

And this is probably mediated by VEGF because you can also inhibit the ET-induced neoangiogenesis by applying a VEGF-inhibitor like this one. There is no more development of neoangiogenesis.

Slide 25

So if we want to summarise the endothelin story, first of all we can say that endothelin and all the components of the endothelin system are expressed in various tumours.  It maybe that the endothelin system is involved in experimental cancers. It’s quite likely that besides an effect on neoangiogenesis there is a direct growth promoting effect of endothelin in cancer cells. There is a direct and indirect (via VEGF) angiogenic effect of endothelin at least in the system I’ve described and therefore, it will be quite interesting to see what will happen with the trials with ETa or ETb or mixed antagonists in human cancer and as you may know there is much interest in this system especially in prostrate cancer.

Slide 26

Now what I’ve said for endothelin is more or less the same for angiotensin II at least in pharmacological doses angiotensin II is angiogenic in simple models like rabbit cornea or the CAM, I’m not going to show you that again. It’s a regulator of microvessel growth. Angiotensin II is able to induce growth factors like FGF, which is involved in angiogenesis, VEGF, endothelin by the way and PEGF. It’s also involved likely in the development of renal vasculature as you may remember the absence of the angiotensin receptor or angiotensinogen or the renin gene development leads to a rarefaction of the renal arterioles. This effect is likely to be mediated by the AT1 and not the AT2 receptor. 

Slide 27

Now, of course, this is quite interesting because we have angiotensin II receptor antagonist, ACE inhibitors, anti tumoral we will have inhibitors and therefore it’s important to know what will be the effect or what are the effects of inhibitors also in the angiotensin system in angiogenesis. There are some animal data, which are interesting to consider. For example, there has been described an anti-tumoral effect of ACE inhibitor and angiotensin II receptor blockers in some animal models at least.  There’s also argument for saying that there’s a pro-angiogenic effect of RAS blockers in acute leg ischemia in the mouse model. However, in humans this is not clear at all. First of all, there is a controversy about an anti-tumoral effect of ACE inhibitors. Tony Levey in this city described in Lancet probably more than 10 years ago that there was an anti-tumoral effect of ACE inhibitors in hypertensive patients treated with an ACE inhibitor compared to patients treated with a non-ACE inhibitor. However, this has not been confirmed but will require more study. There is no clean data about the effect of ACE inhibitor in diabetic retinopathy although there is some indication that they might have some effect on the proliferative diabetic retinopathy. To my knowledge there is no study reported in chronic leg ischemia. However, it will not be easy to conclude why because there are several issues to be considered. First of all, there is an angiogenic effect of bradykinin and therefore, an ACE inhibitor will block as a pro-angiogenic effect of angiotensin II but at the same time might potentiate the pro- angiogenic effect of bradykinin because bradykinin is pro-angiogenic. Second I would like to discuss that with you there’s an anti-angiogenic effect of angiotensinogen.

Slide 28

So inside the renin angiotensin system you have 2 components which act in an opposite fashion, angiotensin II which is pro- angiogenic and again at pharmacological doses and I’ll show that angiotensinogen, the renin substrate is anti-angiogenic. Well I’ll show that first of all angiotensinogen is anti-angiogenic and this is new because a single non function of angiotensinogen was to serve as a renin substrate that has been implicated in high blood pressure as you know. But angiotensinogen is a non-inhibitory serpin, a serin protease inhibitor and we will see that some serpins like anti-thrombin, PEDF, maspin, calistatin by the way also are anti-angiogenic and therefore, we hypothesised that angiotensinogen and its derivatives could be anti-angiogenic molecules.

Slide 29

Here you can see the schematic structure of AT III, PEDF and maspin, which have been reported during the late 90s and the early 2000 years to be anti-angiogenic. PEDF is quite interesting because PEDF is over expressed in the retina at birth and PEDF contains at  its end terminus a growth factor, a nervous factor activity but the rest of the molecule, which is serpin, is probably inhibiting neoangiogenesis at this crucial step of retinal vascolarization. Angiotensinogen also contains at its end terminus the decapeptide, angiotensin I and the rest of the molecule is a serpin also, it can be cleaved specifically in this site like some other serpins.

Slide 30

Therefore, we are dealing with 3 molecules the intact angiotensinogen, the Des (AngI)AGT, which doesn’t have anymore the angiotensin I at its end terminus and the cleaved angiotensinogen at this very specific site of hydrolysis. As I said angiotensinogen is a serpin by sequence homology with serpins by the fact that this gene is organised in a similar way to that of alpha 1- antitrypsin and modelling of angiotensinogen reveals a typical 3D serpin structure.

Slide 31

Now, angiotensinogen is able to inhibit endothelial cell proliferation like HUVEC or human microvascular endothelial cells and you can see that increasing doses of angiotensinogen inhibits the cell proliferation whether it’s induced by VEGF or by FGF. 

Slide 32

And in a similar way angiotensinogen is able to inhibit in vitro capillary tube formation on matrigel and this is true not only for angiotensinogen but for Des (AngI)AGT and for the RCL, the reactive centre loop-cleaved angiotensinogen. So angiotensinogen at least in vitro is anti-angiogenic it is typical of an anti-angiogenic compound.

Slide 33

But in ovo it’s also anti-angiogenic using again the chick chorioallantoid membrane assay you can see here compared to controls that angiotensinogen applied on to the CAM is clearly anti-angiogenic.

Slide 34

And this anti-angiogenic effect of angiotensinogen is due to a decrease in cell proliferation as shown by the marked decrease in BrDU incorporation in angiotensinogen CAM.

Slide 35

And also an increased apoptosis as shown here by the CAM space immunostaining in angiotensinogen treated CAMs.

Slide 36

Now if we want to summarise this part of the work, we can say that several serpins are anti-angiogenic. Whether or not these serpins are so-called inhibitory or non-inhibitory serpins, which is the case for angiotensinogen they saw an in vitro but also an in vivo effect, I’m going to show you that now in some other animal models. What is interesting in the case of angiotensinogen this is the case for angiotensinogen itself, not for angiotensin, it’s rarely due angiotensinogen, it’s not due to angiotensin I or II and also the endothelial cell proliferation which we observed occurs at physiological doses because angiotensinogen circulates in blood at the level of 1 microM and as you can see here in vitro at least this effect is observed at 0.1 microM.

Slide 37

Now we asked ourselves the following question, is it possible that angiotensinogen could affect the vascular wall remodelling in some specific vascular territories? Is it possible that angiotensinogen could exert an anti-tumoral effect like we can observe it with some other anti-angiogenic compounds and also with some other serpins and what is the mechanism of action? We don’t know yet but it is likely that there is an AGT receptor or a binding protein we don’t know yet about the signalling which is involved.

Slide 38

Now, in order to show whether or not in some conditions again experimental and artificial type of conditions angiotensinogen could effect vascular wall remodelling. We used transgenic mice over expressing human AGT in kidney this was provided by K. Sigmund, the interest of these mice is that they have a very high level of human ATG but they have no clear phenotype, I mean no increase in blood pressure apparently because human ATG is not cleaved by mouse renin. What we have shown is that these mice over expressing human ATG in their kidney have a marked reduction of the calibre of the small renal arterioles, a reduction of around 20% and this defect appears as soon as 5 days of age and seems to be due to a pro-apoptotic effect of AGT.

Slide 39

But here you can see, for example, how we can by computer calculate the vascular wall thickness by measuring the length of the artery and the surface which is indicated here and doing that we can see here, for example, in these male mice over expressing AGT in the kidney you can see this decrease in the renal artery thickness.

Slide 40

In the kidney of females, which are not over expressing human AGT, there is no effect and in other tissues in male where there is no angiotensinogen over production, there is no effect on the calibre of the arteries or arterioles. So this effect seems to be quite localised in a tissue over expressing angiotensinogen.

Slide 41

Now the next question we asked as I said was whether or not we could show an anti tumoral effect using angiotensinogen so what we did was, with Michele Perricaudet, to construct a human adeno angiotensinogen and to use this construct in several models to look at the possible anti-tumoral effect of angiotensinogen. So we used several models, we used a pre-established mammary tumour model in nude mice. We used also an in vitro pre-treatment of these mammary cancer cells by the adeno virus expressing AGT and then we transfected those cells in nude mice and looked at the effect on tumour growth and finally, we used a model of lung metastasis which is a model of melanoma in syngenic mice.

Slide 42

Now here is one of these experiments. You can see here the administration of human mammary cancer cells in nude mice. Then after the growth of the tumour to a limited extent we inject within the tumour either PBS or an anti adenovirus or the adenovirus according to angiotensinogen and then we followed up the tumour growth and the tumoral neoangiogenesis.

Slide 43

As you can see here, you have a marked inhibition of the tumoral progression in those mice treated with the adenovirus according to angiotensinogen compared to the two controls, the PBS control as well as the adenovirus, as the anti-adenovirus.

Slide 44

Now we used another model, which is the metastatic, pseudometastatic tumour model. In this model we take a B16F10 melanoma cells which are pre-treated with the anti-adenovirus or the adenovirus according to angiotensinogen. Those pre-treated cells are injected in the retro-orbital sinus and then 10 days later you can look at metastases in the lungs.

Slide 45

And actually you can count the metastases because they are black metastases as you can see here.

Slide 46

So here you can clearly see that there is a marked reduction of the number of metastases and the adenovirus treated mice compared to the 2 control animals.

Slide 47

We can do the same type of experiment but now in mice over expressing the human AGT. The mice, I have just described before for the effect on the renal artery vascular wall on remodelling. Interestingly enough in these mice, over expressing human angiotensinogen with no adenovirus at all, you can see here that by administration of these melanoma cells you have a marked decreased metastasis expression or numbers in the lung compared to the control wild type mice showing that the circulating and probably the local high concentration of angiotensinogen is important for metastatic dissemination for preventing metastatic dissemination. In order to convince you even more that this molecule might play an important role in tumour formation I’ll show you here a recent experiment which we are doing with Francois Vincent in Paris. This is a model of genetically determined hepatocarcinoma in which the SV40 large-T antigen is under the control of the  AT III promoter. In this case you have  a hepatocarcinoma which develops with several phases, adenomatous, carcinomatous and huge metastatic invasion and we crossed these mice with the mice over expressing human ATG and we made the speculation that these host mice there will be a limited tumour progression due to the high expression of human AGT.

Slide 48

This is exactly the case as you can see here by the Kaplan-Meier survival curve of the hepatocarcinoma mice compared to the host mice with the human AGT transgene, you can see that there is a delay in the death of 28 days.

Slide 49

Interestingly enough also there is another phenotype to follow which is the velocity of the hepatic renal artery flow in which you can see in the hepatocarcinoma mice you have a marked increase in the hepatic blood flow compared to the wild type mice and in the mice over expressing human ATG but also bearing the transgenic large T antigen you have during the phase of hyperplasia and nodule formation, you have a reduction of the hepatic blood flow compared to the hepato cell carcinoma mice. So at the same time there is a marked reduction of neoangiogenesis in the mice over expressing ATG.

Slide 50

So if we want to conclude this part of the study, we could say that angiotensinogen as other serpins is anti-angiogenic, an effect which is observed at near physiological concentration. AGT exerts an anti-tumoral effect in different models as I showed you. An over expression of angiotensinogen effects vascular wall remodelling and all these effects, this is very important are independent of angiotensin generation.

Slide 51

Now there are some interesting new perspectives of course. We need to look at the effect of angiotensinogen during embryonic development in some other territories and renal arteries either retina and brain blood barrier might be interesting sites to study. We need to elucidate the receptor/binding sites as I said, the mechanism of action and of course, we can always speculate a gene therapy product but it’s pure speculation.

Slide 52

I’d like to conclude this talk by saying that first of all, it’s true that several vasoactive peptides exert a pro-angiogenic effect in vitro and in vivo but you have to remember this pro-angiogenic effect is observed at pharmacological doses, it has to be now clearly demonstrated that some of these pro-angiogenic effects could occur at more physiological doses but at least I think I’ve convinced you from the literature that adrenomedullin and from some of our own work that endothelin, angiotensin but also bradykinins are pro-angiogenic vasoactive peptides. Therefore, it’s interesting because this pro-angiogenic effect is unrelated to their effect on the vascular tone because bradykinin and adrenomedullin are vasodilators, whereas endothelin and angiotensins are vasoconstrictors and therefore, there’s no relationship between the effect on the vascular tone and the pro-angiogenic effect of these compound. Importantly this effect is mediated by the receptors, which are G-coupled receptors. This is important because we have developed in the last years agents able to block these receptors, not only angiotensin II receptor blockers but also endothelin receptor blockers and there is a lot of work going on on adrenomedullin and therefore, it will be interesting to see whether the clinical use of these pharmacological agents blocking these receptors has any effect on angiogenesis or anti-angiogenesis in various clinical settings. So this is just to stimulate your curiosity and I think it needs further investigation.

Slide 53

I’d like just to finish by saying that of course this work is a team work we were not the wining team unlike one of the presentations I heard yesterday with an Italian team but at least we are an interesting team working together especially with Jean-Marie Gasc in this work and Judith Favier played also a crucial role in this work as well as the others indicated here. This work has been done also in part in collaboration with Anne-Paule Gimenez and Xavier Jaunemaire in Pompidou hospital and the adenovirus work has been done in collaboration in the institute Gustave Roussy with Celine Bouquet and I thank you very much for you attention.

Chairman: Thank you very much Pierre for this beautiful lecture which provides new insights into the angio pro-angiogenic role of vasoactive peptides and which may have important therapeutic implications in the  new future, thanks again.