CME Slides Forum - F.F. Hou

A joint Congress by ERA-EDTA and ISN

THE COMPLEXITY OF THE FILTRATION BARRIER

Dontscho Kerjaschki, Vienna, Austria

Chair: Richard Glassock, Los Angeles, USA

Thierry Hannedouche, Strasbourg, France

kerjaschki

Prof D. Kerjaschki
Department of Pathology
Medical University of Vienna
Allgemeines Krankenhaus
Vienna, Austria

Mr Chairmen, I’m glad that we found after all two chairmen for this lecture here and this is not complete anarchy. I was given the topic, the complexity of the filtration barrier which in a way is astounding because the morphology of the filtration barrier is as simple as it could basically be.

You’re all aware that there are basically only two types of cells, the podocytes on the urinary space or on the outside, as we call it and on the inside endothelial cells with huge pores. I’m not going to talk about these fellows here because I guess Peter Mathieson will afterwards. These two cells are separated by a basement membrane which may or may not have highly organised or more or less organised structure and I will come back to this.

You get a little more information about the morphological disposition with tricks like this one where you get some sort of a negative stain of the filtration barrier and where you can visualise the surface coat, for example here on the podocytes. These are mainly sialoglycoproteins that form a coating but also the endothelial cell is covered with some sort of surface molecules and some -- may sometimes even occlude the little gaps here in the endothelium.

Now I’m not going to go through this, you’re familiar with this, this is just an inventory which is not even complete showing the molecules that participate in stabilising this entire structure as far as the glomerular epithelial cells and the podocytes are concerned there is just the slit diaphragm I’m not going into this too. There are many or several adhesion proteins that are known, there may be others we are not aware of. A common denominator that holds together the entire structure seems to be the actin cytoskeleton with its additional molecules like actinin and so on and so forth. You’re also aware of the literature that several of these proteins here when somehow faultily made and not functioning will cause severe disease and disruption of the function of the barrier.

as far as we have limited ourselves right now, are actually really the basement membrane and the podocytes and the podocytes you can clearly see the slit diaphragms, you’re aware of this but I’ll draw your attention also to the basement membrane here on the sole of the foot processes that somehow is engulfed on the inner side of the cell with a this dark material which seems to be actin and many other things. There are these fibres that somehow attach, it appears, these cell membranes to the matrix.

Now you all know also that the slit diaphragms are really interesting structures, they are porous as you can see here on these grazing sections and many years ago Karnovsky and his associates calculated that these lots that you find here in the pores that constitute the pores have about the size of albumin and they thought that this may be an ultimate barrier for albumin and hold back albumin. Now the basement membrane is a much more complicated structure because fixation and morphological observation somehow disturbs invariably its molecular disposition.

Now, you know these pictures from the textbooks, this is a highly ordered chickenwire model and there may be some thing to that, chickenwire consisting basically of regularly spaced type IV collagen molecules with additional molecules like laminin, proteoglycans negatively charged proteins that are thought to harbour negative charges in the GBM.

Now can we visualise some periodicity in the GBM? The answer is yes but again this is a highly artificial system based on fixation which means cross-linking of proteins and so on and so forth. Now here we used the trick many years ago by cationising a molecule, cytochrome C which can be seen very easily under the electron microscope

and on this grazing sections you will appreciate that there may be some periodicity in these little dots of cationised cytochrome C that actually sticks to negatively charged groups suggesting that there may be some periodicity in the system.

Now when we take this all together and I’ll rush through this now, what are the functions or which structures have been assigned filtering functions?

Well, I’ll just show you two possibilities of something which is established. If you take a large molecule like ferritin which has a molecular weight of half a million, which is far beyond the cut-off of albumin and it’s negatively charged on its surface, then it appears that at the level of the lamina rara interna there will be a barrier for this molecule.

However, when you cationise this molecule on the surface, it will behave differently, it will penetrate through the GBM, reach everything here and will decorate the membranes of the podocytes on the base and accumulate below the slit diaphragms.

What I think as a general rule we can learn from this is, if a molecule is able to interact one way or the other be it by charge or by receptor ligand type of interactions, with any component of the matrix here, then it may as well punch its private hole and the cut-off as we know it and the size of the molecules may not become relevant for this particular molecule, which makes the story much more complicated and much more individualised.

So albumin may be holds back here whilst cationic proteins may migrate through. In addition LDL particles, for example, may interact with the LDL receptor like domains in the proteoglycans here and for those also the transition through the GBM may be simple and easy.

Now this is a static picture but what happens when you really think about the living machinery of the cell? There will be products, metabolic products released from the endothelial cells but also from the glomerular epithelial cells, from the podocytes all the time. These will actually incubate so to say the surface of the podocytes and the matrix of the basement membrane continuously. Now let me say a few words about reactive oxygen species because we think they are very crucial for the function of this entire system. I’m not going to give you a course on oxygen radicals, I just want to impress on you that there are several compounds generated either in mitochondria or in other sites in the podocytes that release the hydroxyl anion which is a very reactive compound that will do a lot of damage to different molecules as I will show you in a minute.
As a matter of fact, the podocytes even when not challenged by inflammation or by immune complexes or by complement are generating as it appears oxygen radicals in the mitochondrial metabolism all the time and they may be released in small dosages as I will show in a minute and then may be involved in the maintaining of the filter. But when challenged for example by complement in immune complex disease all of a sudden the podocytes will start to express the same enzymes or similar enzymes as neutrophils that generate the respiratory burst and there will be a lot of oxygen radicals that are released into the matrix and flushes here everything.

Now what is the evidence that oxygen radicals are really produced by podocytes? Well I’ll show you a simple experiment what we did was a simple proteomic approach in this series of pictures. We looked at early stages of puromycin nephrosis where we intoxicate rats and you get flattening of the foot processes and enhanced permeability of the glomerular filter. Now, what we have done here is something very simple, we have isolated glomeruli, we have labelled the proteins of the normal glomeruli green and we have also made disease and labelled all the proteins of the isolated glomeruli in red. When you superimpose this, you can pick with an appropriate machinery those proteins that are specific for the diseased and for the healthy system only and when you look at what happens in the disease and I’m not going to guide you through this I just want to show you these 2 red underlined things, you will find a lot of molecules, a lot of proteins that are expressed by the podocytes and by the podocytes’ mitochondria that are involved in lipid import and in oxygen radical metabolism. In the context the punch line is there will be much more oxygen radical production under this situation.

Now the radicals are actually nasty molecules, they will break up matrix proteins like laminin here, they will aggregate proteins, they will clip collagen type IV at different points and there may even be some pecking or -- for example, intactin is very sensitive to oxygen radicals while it takes a lot of radicals in the long time of treatment to dissociate and clip type IV collagen.

But there is an additional facet of toxicity to oxygen radicals. This is lipid peroxidation.

Whenever radicals, especially the hydroxyl anion meets unsaturated fatty acids, a sort of an explosive chain reaction will happen that generates the presumably most toxic compounds in our bodies and in cells in general. I’m not going to go through all of this I just wanted to show you here that the aldehydes generated in this process are very much the same as we use for fixation in histological tissues. These aldehydes will cross-link proteins by forming adducts with available amino groups on the surface of the proteins and they can be detected.

Now I’ll show you very, very quickly something which was published some time ago that this has a strong bearing on glomerular damage in immune complex disease especially in Heymann’s nephritis.

Heymann’s nephritis as you know is a model of membranous nephropathy and one of the causes or may be the main cause is the expression of a protein which we call megalin which is sitting on the base of the foot processes and which is seen here by circulating antibodies that percolate here through and bind forming immune complexes that are shed and then the cycle is repeated and this is how the deposits grow.

Now intriguingly this huge molecule which has the size of more than 500.000 with three times or 4 times the size almost of IgG binds to several proteins, including glycoproteins and ApoE but it uses only just a very few points on its surface. As a matter of fact, there’s one major point in this area here, we have recently obtained the crystal structure of this and it seems that there is a motif here in megalin that is very similar to the major binding region in the LDL receptor.

Now why am I showing you this? I’m showing you this because the antibodies that form immune complexes have chosen for a reason we don’t understand to be specifically directed just against this binding region. In other words, the antibodies that form immune complexes in these particular diseases are actually made from competing antibodies when it comes to the binding of lipoproteins. They block the binding of lipoproteins and the metabolism of lipoproteins by megalin and the result is an accumulation.

Now if you have accumulation of lipoproteins with lipids here in the immune deposits, and you have oxygen radicals that are released in relatively high concentrations as it appears from the podocytes, you will have lipid peroxidation going on. I’ll show you this as one of my last slides because I want to impress on you the following, there is a trick where you can visualise the proteins that have been attacked by lipid peroxidation adducts. With a trick you can introduce --- into these compounds and when you then look at normal glomeruli and passive Heymann glomeruli that are proteinuric and ask the question, which proteins can be found in the GBM that have been attacked by lipid peroxidation? You will find that in the disease there are two major protein -- that have been identified as type IV collagen.

But look also in the normal glomerulus that is non-proteinuric you will find that there is something like this going on. So what we think is that it could be that oxygen radical generation, lipid peroxidation goes on all the time in our body, it may be a regulatory mechanism because these molecules are actually tagging matrix proteins and other proteins for metalloproteases and maybe this is the kiss of death for a matrix protein because it signals to the next metalloproteases clip me and this is may be how the turnover of the matrix could function altogether.

Now in excess however, it’s conceivable that the lipid peroxidation products will either cross-link the membrane proteins in the matrix and open pores, disturb the structure if there is some structure or they will cleave and clip the proteins and by doing so open other structures. So this is what we believe happens.

Now finally, there is an enigma. Very frequently almost invariably proteinuria is associated with flattening of the foot processes, retraction, fusion whatever you want to call it.

Now, this is a highly permeable glomerulus and the question is, how come? Where does the protein penetrate? Through a sealed monolayer of epithelial cells. Well there are several concepts. One concept is that there’s an epithelial transport system that somehow throws out the albumin on the urinary side. Actually there’s very little evidence for this, we always find in proteinuria accumulation of albumin and IgG in the podocytes but actually this exocytosis mechanism is not established. It must be very efficient to be operative and account for proteinuria.

Alternatively however, this also comes from Morris Karnovsky’s lab this old idea is that there is detachment of epithelial cells and now there seems to be a lot of evidence that this really happens and this is associated with proteinuria

and you can do this in such a primitive way when you perfuse kidneys with protamine sulfate when you compensate for the negative charges and you do many other things on the way, you may even detach sometimes the podocytes from the matrix. Now it’s very important I think to understand the molecular biology and pathology of this mechanism because this could be a really crucial event in how proteinuria forms.

Finally, I wanted to show you this, I can’t resist doing this because this is how the situation I think appears to us at this point. That’s an old painting from an Austrian church, a medieval painting showing Jesus Christ and here you see the emperor washing his hands in innocence. I’m not showing you this because of this motive, I’m showing it because of this. The artist was commissioned to paint a lion but he had never seen a lion in his life, he relied on the literature and his imagination and may be this is very similar to what we think glomerular permeability and complexity of the filter could be. Thank you very much.

Chairman: Thank you Professor Kerjaschki for this comprehensive review. Any comments or questions from the floor? There is an ongoing debate on the respective contribution of the layers of the glomerular filtration barrier can you comment on that?

Prof. Kerjaschki: Well, this is a déjà-vu when you go back to the literature you will see that in the 70s there was exactly the same argument. I think the real problem is that nobody has come up with a non-traumatic way of following up tracer molecules. Whatever you do, you have somehow to interfere with the function of the filter either you have traces that become visible only when you have fixed the tissue that means cross-linked the tissue on the one side. You have the problem of the collapse of the filter, the release of pressure in the last fractions of seconds which in this micro milieu will blur the picture and let the traces diffuse around. There are possibilities that you look at surface glomeruli with two photon techniques they also have limitations so I think we are back to square one where we were in the 70s and I wonder if anyone will come up with something reasonable.

Question: Dontscho I have a question for you, it’s something that’s puzzled me for a number of years. It has to do with whether there is an entity of the subepithelial space. A few years ago I think that concept was reintroduced and I wanted to know do you think it’s an artefact of fixation or is it a real space?

Prof. Kerjaschki: Actually, I would pass on this question. I don’t know as much as you don’t.

Question: As for human membranous nephropathy, since the debate about Heymann’s nephritis was that megalin was absent from human glomeruli, do you think given the identity of this domain with the LDL receptor, that maybe the LDL receptor could be involved in the pathogenesis of human membranous nephropathy?

Prof. Kerjaschki: Nobody knows, you know there are people looking now at the lipoprotein phospholipase A2 receptor which could be the analogous receptor and target antigen in humans when compared to megalin so nobody really knows at this point. What we see however, is that there is accumulation of lipoproteins apparently also in human membranous nephropathy.