TOLL-LIKE RECEPTORS AS MEDIATORS OF RENAL DISEASE

Detlef Schlöndorff, Munich, Germany

Chair: Detlef Schlöndorff, Munich, Germany

Andrew Neil Turner, Edinburgh, United Kingdom

Prof D. Schlöndorff Direktor Medizinische Poliklinik - Innenstadt Klinikum der Universität München Munich, Germany

---

Slide 1

Thank you very much. Well, it happens, as you get older when somebody chops off your head like age then you have young people in the lab working and you pretend that you’re part of them. So that’s how I speak about toll-like receptors. Most of this work was really done by Hans J. Anders in our lab in Munich.

Slide 2

Toll-like receptors. Toll was discovered in 1996 when in drosophila a developmental gene toll, which means crazy, was knocked out. When this gene was knocked out it was apparent that the animals or the fruit flies succumbed to massive fungal infection as shown here in this poor fly that has its eyes and its entire body replaced by fungal ifi. It was then recognised that in the mammalian system there are so-called toll-like receptors by molecular analysis.

Slide 3

Mean time we know about 10 human toll-like receptors and 11 in the murine system. Some of these are old friends of ours like the toll-like receptor 4, which is the classical LPS endotoxin receptor. Similar, TLR2 is a receptor for a bacterial membrane protein like lipoproteins or also lipo-proteoglycans and peptidoglycans. So these are all receptors on the outside that recognise patterns associated with molecular pathogens that are on the outside of bacterial fungi etc. and then signal to the inside through the innate immune system to this cell telling it there is danger. So these are also called danger-associated receptors. Some of the receptors are located on the endosomes, on the inside of the cell. These are TLR 3, 7, 8 and 9. These recognise nucleotides that are released from proliferating or phagocytised bacteria or viral antigens. So these are external, these are internal and are for nucleotides. They’re a very complicated system of adaptor proteins that result in a far more complicated signalling mechanism than is shown here that eventually results in 2 major signalling systems, one is through NF-kB resulting in the generation of proinflammatory cytokines obviously including chemokines and the other one resulting, at least in immune cells, in the generation of Interferon alpha and beta in a viral immune response trying to limit the viral proliferation.

Slide 4

Much more complicated than shown here none the less for the purpose of this talk this should suffice. You see here that these TLRs are distributed differentially on inflammatory cells where it could be said that really LPS receptor is pretty much present on all cells in the body and the TLR receptor is also very widely distributed but the others specifically the TLR 3, 7 and 9 are much more restricted to inflammatory cells.

Slide 5

Now how about in the kidney? I’m going to lead you through this and show you the evidence for this.

Slide 6

Let’s start with the ones that are widely distributed like TLR 2 and 4 for endotoxin and lipoprotein surface molecules of bacteria. They are distributed in the collecting tubular epithelial cells and in the proximal tubular epithelial cells and obviously, they could respond to ascending infection and they could respond to endotoxemia. Then there are the resident and infiltrating monocytes and dendritic cells that can exhibit TLR 7 and 9 and could respond to bacterial and viral infections. Finally, a novel observation could be the TLR 3 expression on mesangial cells and vascular smooth muscle cells that could have to do with infection associated immune complex glomerulonephritis and in the vasculature perhaps with infection associated vasculitis. So the obvious interaction would be between TLR 2 and 4 on tubular epithelial cells during infection and this is a work from our laboratory where we’ve taken advantage of the LPS resistant mouse strain that is defective in the receptor for TLR 4 when you create with a pathogenic E. coli strain an ascending infection, these animals don’t respond at all in their kidney, there is no immune response and they succumb to sepsis from this infection. The normal response is that you generate a micro abscesses in the kidney that allow the elimination of the infecting organism and this is shown here the normal animals have a marked pyuria, they respond to the infection, whereas the ones lacking TLR 4 don’t respond at all. You can see this here that this is associated with the generation of chemokine for polymorphonucleocytes LPS alone or here with the uropathogenic E. coli species a massive expression of this chemokine calling in the leucocytes forming the micro abscesses that is hardly present in the animals that lack the TLR 4 receptor.

Slide 7

When looking is it important that the TLR 4 be expressed on the leukocytes or on the kidney we generated chimeric animals? You see here the micro abscess formation as a defence mechanism in the wild-type animals associated here with a urinary leukocyturia. When you knockout only the TLR4 in the leukocytes, you get a reduction in the response but you still have some response. If you knock it out in the kidneys but you leave it intact in the leukocytes, you see that the tubular expression of TLR4 is very important and when you have no TLR4 on either tubular epithelial cells or on infiltrating leukocytes, you have no response at all and you just have massive replication of E. coli. So TLR4 is important in the defence against ascending urinary tract infection and it’s important that it be localised on the tubular cells but also on the infiltrating leukocytes. In addition, TLR4 is responsible for LPS induction of chemokines in the kidney.

Slide 8

Now how about TLR2, another surface receptor, another receptor for surfaces of bacteria and infectious agents?  This is a paper from the group of Alan Van Der Wal in Paris that showed that for interstitial leptospiral infection, a common disease in humans, that you absolutely need TLR 2 in order to set off a response and that this can be restricted to the leptospiral surface protein LipL32 a lipoprotein that binds to TRL2 and prevents and responds to the infection. When this is lacking, there is no response and overwhelming infection.

Slide 9

Now how else could these lipoproteins influence the kidney?

Slide 10

This is lipid A, a ligand for TLR2 that generates MCP-1 and RANTES, two of the major chemokines involved in the attraction of monocytes and T-cells and I had shown you before the data on TLR4. so during systemic infection with activation of TLR2 and TLR4 reaching the kidney lipoproteins and LPS reaching the kidney you have activation of these systems by proximal tubular epithelial cells and the result of LPS of sepsis on the kidney that is clinically all too well known to all of you.

Slide 11

Now some people have postulated that there is mimicry of endogenously generated proteins and components that can bind to TLR4 and TLR2 such as breakdown products of the matrix hyaluronan, biglycan that Tamm Horsefall protein, if it reaches the interstitium can activate TLRs fibrinogen, if it reaches there in high content and heat shock proteins. So these could supposedly activate TLR 2 and 4.

Slide 12

What is the evidence for this? The best one I think is from the group of Leehmans in the Amsterdam group who have shown here that in animals that are lacking TLR2 and that are submitted to ischemia-reperfusion there is a marked decrease in the tissue damage and in the generation of chemokines as shown here, chemokines and cytokines. KC is a chemokine in the mouse for polymorphonucleocytes, MCP-1 and here the proinflammatory cytokines much less generation and this generation occurs in the kidney and you see that there’s less tubular damage in the animals that lack the TLR 2 receptor on the tubular cells.

Slide 13

Again this group also did chimeric mice to show that really the TLR2 on the tubular epithelial cells through interaction with endogenous ligands for TLR 2 cause the inflammatory response and the tubular damage in this model of ischemia-reperfusion.

Slide 14

So non-bacterial ligands can also interfere and interact with these receptors. Now during infection TLR 2, 3, 4 and 7 could interact with pre-existent immunological disease in the kidney on immune cells.

Slide 15

This is shown here for TLR 4 with endotoxin in a model of anti-GBM nephritis non accelerated model, that if you inject either LPS or the antibody alone, you get very little effect.

Slide 16

However, if you inject both the antibody and LPS you get a marked glomerulonephritis and tubular interstitial nephritis as also shown here in the histology. So LPS can aggravate immune mediated glomerular diseases by activating the inflammatory cells.

Slide 17

I had indicated earlier that there are intracellular TLR receptors for double stranded viral RNA TLR3 and for bacterial and viral DNA that is of a specific structure in order to discriminate between its own DNA and the bacterial or viral DNA the body takes advantage of the fact that bacterial and viral DNA is rich in CpGs and this repeats of cytosine poly guanosines, which is not present in mammalian DNA and also that this is under methylated.

Slide 18

Now again Anders in our group, first in a model of an immune complex induced GN by repeated injection of HAF, you see you get an immune-complex GN as demonstrated by the albuminuria. When you use a mammalian type DNA, you get no aggravation. If you use a bacterial type DNA, you get an aggravation of the proteinuria.

Slide 19

This is associated with a marked increase in the activity and chronicity index of this glomerular disease. In other words, both the active GN as well as the chronic changes are enhanced when you have this bacterial DNA injected at the same time.

Slide 20

When you look at where it is located we took advantage that we can label these bacterial DNAs and inject them and then co-stain them with either markers of monocyte macrophages or markers of dendritic cells and you can see here this is the injected CpG oligonucleotide, this is a staining for F4/80, for the monocytes, they co-localise and similarly they co-localise with the dendritic cells, they do not co-localise with any endogenous intrinsic renal cells. So this is through interaction with monocytes, B cells and dendritic cells.

Slide 21

Now, how about when we use a viral infection like double stranded RNA and for this you can use poly IC RNA and you see here this is in a model of lupus nephritis in the LPR mouse that when we co-inject this virus like RNA, that at 20 weeks we have a massive increase in the glomerular expression of TLR3 and we get a mesangial lysis-like aggravation of the kidney disease associated also with a crescent formation and a really massive acceleration of the glomerular disease.

Slide 22

Now here when we re-inject this poly I:C labelled RNA in a labelled form, you see that it localises to the glomerulus and that here is the localisation in the glomerulus and that it localises to infiltrating interstitial inflammatory cells. When we stain for TLR3, the staining is in the infiltrating interstitial cells and in the glomerulus it’s there in a pattern that resembles mostly a mesangial distribution.

Slide 23

Here is again shown the massive aggravation of the disease as associated with multiple measurements here for monocyte infiltration, here for T-cell infiltration in the interstitium and here for fibrosis and progression. This is a control RNA that does not have the viral similarity and that causes no worsening of the disease compared to the normal lupus nephritis in these animals.

Slide 24

Here, this is just a quantification of the expression of MCP-1 and RANTES message in these kidneys and you see that the viral RNA injection causes massive local this is in situ and immune fluorescence for the above chemokines, that locally there’s a massive increase in the chemokine expression and hence infiltration of inflammatory cells and worsening of disease.

Slide 25

We then looked across the border at TLR 1 through 9 in macrophages and in cultured murine mesangial cells and you see that macrophages express multiple TLR receptors, especially TLR9. This is responsible for the aggravation with the CpG DNA that I showed you previously and the localisation to the macrophages in the kidney. By contrast TLR 3 is highly expressed on cultured mesangial cells and by immune histology I had already shown you in situ that it was there.

Slide 26

So viral and bacterial nucleic acids can enhance renal disease by at least 3 different mechanisms. They can bind through TLR9 on B cells and macrophages, dendritic cells, enhance auto-antibody formation and macrophage activation. They can bind to in situ localised that means in the kidney localised macrophages and dendritic cells generating more proinflammatory cytokines including TNF-α and IL-6 and thereby aggravating the disease. As a novel mechanism we propose that viral RNA can bind to TLR3 on mesangial cells and vascular smooth muscle cells, as well as on macrophages and dendritic cells generating proinflammatory cytokines, chemokines and leading to an aggravation of the disease or perhaps even to initiation of de novo disease.

Slide 27

Now how could that be? How could TLR3 influence a primary glomerular disease?

Slide 28

We thought of the model of the cryoglobulinemic induced membranoproliferative GN secondary to chronic hepatitis C viral infection as demonstrated here in these pictures from Rick Johnson published many years ago in the New England Journal that there is immune complex deposition with a membranous and mesangial distribution and here between the membrane and the mesangial cell.

Slide 29

This is an immune histology for TLR3 in human kidney and you see there is a wonderful distribution of the mesangium that’s positive for TLR3, was an intracellular granular pattern and the same can be shown on small vessels on the vascular smooth muscles of small vessels.

Slide 30

In cultured human mesangial cells the poly I:C RNA causes a massive induction, no a combination of proinflammatory cytokines, causes a massive induction of TLR3 expression of the receptor for the viral RNA on mesangial cells. When we then incubate these pre-stimulated cells with poly I:C itself and look at the expression of chemokines or proinflammatory cytokines, you see that this causes a massive induction

Slide 31

Now, how could we show that this might at least correlate with the hepatitis C associated GN? We took both laser micro-dissected glomeruli from biopsy specimens of GN that were positive for Hepatitis C virus and for those that had primary membranoproliferative GN and were hepatitis C negative, both of these groups had cryoglobulins in their serum.

Slide 32

You can see here that TLR3 in the micro-dissected glomeruli only of the biopsies that were from patients with hepatitis C-associated GN had increased expression of TLR3, that this was associated with glomerular increased expression of the chemokine RANTES and as not shown for other proinflammatory cytokines.

Slide 33

So we’re hypothesising that mesangial activation of TLR3 by viral double stranded RNA containing immune complexes could through binding to phagocytosis of the immune complex, binding of the immune viral RNA to the endosomal TLR3 activate the mesangial cell and thereby cause a local GN.
We’re hypothesising further, that TLR3 could also play a role in viral disease-associated and exacerbated GN, for example, IgA nephritis when associated with viral disease, perhaps even HIV nephropathy.
Finally it is intriguing to speculate that the presence of TLR3 on vascular smooth muscle could have something to do with the role of the vascular smooth muscle during viral disease-associated vasculitis that also occurs, for example, in hepatitis C.

Slide 34

Now this work was done by a lot of people, mostly the group of H. J Anders in our group together with Clemens Cohen, with Pawar, Patole, Schmid, Segerer and Woernle and  most important also in collaboration with Gröne in the pathology in Heidelberg and the biopsy study was made possible through the European Renal cDNA Biopsy Bank. Thank you very much.

Slide 35

 

Chairman: Detlef, thank you very much for your masterly walk through an area, which is changing very fast, and I’m sure has further to go. Any questions?

Question: Do you consider that renal immunopathology is in relation with renal the renal compartment glomerular interstitial and juxta glomerular? I consider that compartment as an immune compartment like immune organs, digestive and respiratory organs. Which is your opinion? If Toll receptors at their disposition are in relation with these compartments?

Prof Schlöndorff: Well, I think the mesangial compartment is a very special compartment because of the fact that the glomerular endothelium is a fenestrated one and therefore, macromolecules like immune-complexes or lipoproteins etc. have pretty free access to the mesangial cell. Now, do I consider the mesangial cell a primary cell of the immune system? I don’t think so. I think it’s more like a pericyte, a cell that is exposed routinely through to immune-complexes and whose role is essentially to clear all the junk that it gets exposed to, so that it doesn’t build up and I think it’s a matter, its quantity becomes quality. If there is too much of this, then it becomes an immune response but I think the normal mesangial cell will make sure that this gets removed without causing inflammation. If too much of it comes in, then complement gets activated, then TNF-α gets activated and then the cell becomes a proinflammatory cell but I don’t think it’s a real cell of the immune system in the stricter sense but I think it’s a question of definition definitely.

Question: Sandrine Florquin from Amsterdam. Thank you very much for this nice overview. I have a question about the function of TLR3 in mesangial cells in the last part of your talk because you really nicely showed that there is an increase in TLR3 in hepatitis C-related MPGN but if you have cryos related MPGN you have less and significantly less TLR3 expression on mesangial cells but these patients still develop really severe MPGN related to the cryos so my question is what’s the chicken and the egg in this story?

Prof Schlöndorff: Well, I think there are 2 different mechanisms. Let me make it clear, I said we hypothesise ok? Because so far this is an association, we have not proven that this is mediated through TLR3. Our hypothesis is that immune complexes containing double stranded viral RNA, as they do in chronic hepatitis C, those immune complexes do contain double stranded viral RNA, could mediate even at low titres an activation of the mesangial cell and thereby, cause GN. Cryoglobulins from primary cryoglobulinemic diseases do not contain double stranded viral RNA. Their mechanism of action to cause GN would have to be a different one and I guess there are multiple ways of how cryoglobulins affect the glomerulus. Number one, there is simply the precipitation of cryoglobulins in the capillary loops that you can very frequently observe, you simply have the capillary loop like a sausage filled with cryoglobulins and you can have a MP type of disease and there you have complement activation and I think there the mechanism of activation would be through complement and not through TLR3. Again let me pre-phrase it, this is a hypothesis.

Chairman: Thank you very much. We’re going to need to move on now Detlef thank you very much indeed.