CME Slides Forum - D. Harris

A joint Congress by ERA-EDTA and ISN

IMMUNE MECHANISMS OF INTERSTITIAL INFLAMMATION IN MEMBRANOUS NEPHROPATHY

David Harris, Sydney, Australia

Chair: Dontscho Kerjaschki, Vienna, Austria

Claudio Ponticelli, Milan, Italy

harris

Prof. David Harris
Department of Renal Medicine and Transplantation
The University of Sydney at Westmead Hospital
Westmead, Australia

So in this talk I’m going to cover 4 questions about interstitial inflammation in membranous nephropathy and then go on a little bit to talk about ways of trying to influence the interstitial inflammation. So the first question in two parts is how common is it and how important?

Well it’s very common, it’s almost a sine qua non of membranous to see interstitial inflammation. How important is it is shown in a number of studies and this one is from Daniel Cattran Unit in Canada looking at survival in patients, in almost 400 patients with membranous and looking at the effect of tubulo-interstitial disease and this includes fibrosis and tubular change so it’s not just interstitial inflammation showing that the extent of tubulo-interstitial damage has a marked effect on survival. However, when they controlled for patient age, blood pressure and serum creatinine at the time of diagnosis, tubulo-interstitial change was no longer an independent predictor of survival. Interestingly, in this study they showed that these histological changes had no bearing on whether a patient would respond or not to therapy.

That’s shown further in this next study which is as you can see quite a small study, a retrospective one that was published 15 years ago and looked at 2 small groups of patients who had a normal serum creatinine and fairly mild changes on their renal biopsy, so minimal or minor tubular atrophy and interstitial fibrosis. They were matched for age and they were matched for glomerular changes. What the authors showed was that the patients that responded to steroid therapy with either a partial or a complete reduction in their proteinuria in fact had a greater number of inflammatory cells in their interstitium. So they had more T cells, more CD4 T cells, more CD8 cells and more monocyte macrophages. So the presence of interstitial inflammation doesn’t exclude response to therapy and may even be a predictor of response.

This study from 20 years ago from Bob Atkins’s lab David Hooke was the first author on this a paper, looked at interstitial inflammation in a whole range of human glomerular nephritides and what they showed was that the density of interstitial inflammation in membranous was the same as in other non-proliferative glomerulonephritides but not surprisingly less than in proliferative types of GN. They also looked at the types of cells that made up the interstitial infiltrate and showed that the makeup of the infiltrate was the same no matter what the type of GN, be that proliferative or non-proliferative, so that the majority of cells were T cells and there were also a large number of monocyte macrophages.

Now, that’s in humans and there are a number of studies that have looked at this in animal models, the animal model being Heymann nephritis and this study is from Mark Penny looking in active Heymann nephritis at glomerular and tubulo-interstitial inflammation and this study shows the discordance between what you see in the glomerulus and what you see in the interstitium. So at 4 weeks these animals had glomerular deposition of immunoglobulin but they weren’t yet proteinuric. Within their glomerulus they had an increase in CD4 cells and an increase in CD8 cells but not monocytes compared to control. When you looked at the interstitium, there were no interstitial inflammatory cells. When you looked at the rats 8 weeks later, the make up of inflammatory cells in the glomerulus was different. There was an increase in CD4 and CD8 cells but also at this time period in monocyte macrophages. In the interstitium the make up was slightly different, so there was a large increase in CD4 cells, not much change in CD 8 and a change but to a lesser extent in monocyte macrophages. So there’s a discordance between what happens in the glomerulus and what happens in the interstitium and this suggests that the pathogenesis in the two compartments is different.

This study was performed by Wu when she was still in our lab looking at one subgroup of cells amongst the T cells and these were the γδT cells. She showed that the γδT cells in the cortex of these kidneys contained a type of T cell receptor that was found in other models of proteinuric renal disease only in a relatively small number. These γδ cells expressed IL4 and IL5 and TGFβ and the suggestion from the study was that this was in fact a regulatory response to injury within the tubular interstitium. In another animal model in adriamycin nephrosis we’ve actually done the studies where we’ve inhibited these cells using a monoclonal antibody and shown that if you do that, the disease gets worse suggesting that these cells are in fact protective. We haven’t yet done that in membranous or in Heymann nephritis.

The third question I want to address is can the change that we see in the interstitium be a primary? In other words be a primary part of the disease process.

An interesting study from Alison Eddy looked at this question 15 years ago and using passive Heymann nephritis, so she was giving the anti-serum to the animals, she showed that the anti-serum actually reacted to epitopes on the basolateral membrane of the tubuloepithelial cells and also to the brush border membrane. When you look at what happened at the basolateral membrane, early on there were fine granular deposits along the membrane and these then clumped together and became rearranged forming larger focal deposits. Then subsequently there was a focal interstitial infiltrate consisting mainly of monocyte macrophages but also CD8 cells around those deposits. She showed that this was independent of proteinuria and independent of complement. On the other hand, looking at the antibody that reacted with the brush border membrane there were diffused deposits along the brush border membranes this is very early in disease during a heterologous phase and this led to sloughing of the microvilli and subsequently impairment of the tubule or regeneration of the tubule cell. This she showed was complement-dependent.

Now what about in humans? Well, I think we all know that it’s unusual to find significant tubulo-interstitial disease early in a case of primary membranous GN, in other words as part of the primary disease process. There are occasional case reports, so there’s a dozen or so in the literature of membranous GN associated with anti-tubular basement membrane nephritis at the same time. These children have an antibody against the 58 KD tubulo-interstitial antigen which was from the tubular basement membrane. There are also occasional case reports of immune deposits in other positions such as along Bowman’s capsule but that’s, as you can tell a very rare finding in membranous GN.

However, in secondary membranous nephritis it’s much more common to find involvement of the interstitium early on in the disease and probably part of the primary disease process. In fact, when you look at a renal biopsy and you look for hints which suggest to you that the case is one of secondary membranous nephritis rather than primary membranous nephritis, what you find is that if you see immunoglobulin along the tubular basement membrane or you find antibody reacting with a tubular basement membrane itself, then that suggests to you that in fact this is a case of secondary rather than primary membranous GN.

We’ve heard from Pierre about the antigens that are being identified for primary membranous GN and we all know that there are a large number of potential antigens that have been described with secondary membranous glomerulus nephritis. Whether these antigens, the antibodies and the immune complexes that derive from them lead to primary tubulo-interstitial involvement I think depends very much on the nature of the antigen itself and of the antibody and the immune complex.

The fourth question that I want to address then is how does secondary interstitial inflammation arise?

So I’ve just said that primary interstitial inflammation as part of the primary disease process is unusual but we know that interstitial inflammation is very common in membranous GN. So how does that arise? If we look at the possible causes, they’d arrive from factors that are found in the tubular lumen, which I’ll discuss. They may also be a spread of factors from the glomerulus probably against that early on in the disease is the absence of crescents and adhesions in membranous GN, so there’s no sort of link between the glomerulus and the surrounding interstitium. In fact, this may also spread along the peritubular capillary.

Now, I’ll say it at the outset that there is very little to suggest that there is anything special about the secondary tubulo-interstitial changes that occur in membranous GN as opposed to those that occur in any proteinuric renal disease.
This diagram I think is a good one to illustrate the role of the tubular cell in causing interstitial inflammation.

So this schema is one that is well known to all of you and it’s the sort of schema that you see in every discussion of tubulo-interstitial injury occurring in proteinuria renal disease so as I’ve said I don’t think there’s anything particularly special about membranous. You’ve heard about complement and the discussion was mainly to do with the glomerulus but complement is also important in the tubular injury. So that’s C5b-9 which is being filtered across the glomerulus and therefore, got into the tubular lumen or smaller complement components such as C3. C3 can also be produced by the tubular cell itself and can be activated by the tubular cell and it’s felt by some investigators to be the primary cause of tubulo-interstitial damage in membranous GN. I think the importance of these smaller complement factors as opposed to C5b-9 has been shown by the studies of Tim Spicer that showed that the disease still occurred in C6 deficient mice.

I now want to address in the last few minutes a couple of studies that have looked at altering the interstitial infiltrate in membranous GN. This is a study from Mark Penny again looking at active Heymann nephritis and looking at the effect of anti-CD8 treatment on cells in the interstitium or cells in the glomerulus. When the treatment was started early and the renal cortex was examined 12 weeks later, there was a marked reduction in inflammatory cells, CD4 cells, CD8 cells and also monocyte macrophages. However, if you looked at the glomerulus as well, you found that there was also a reduction in the same cell types in the glomerulus. So the changes that you saw in the interstitium may well have just reflected what happened in the glomerulus and in fact, there was a reduction with that treatment in the proteinuria. So it doesn’t really tell us anything special about the interstitium itself. However, when you used delayed treatment, so this treatment was started after 6 weeks when the disease was established the animals were already proteinuric and then you looked at them at 12 weeks, once again there was a significant reduction in the number of CD4 cells, CD8 cells and monocyte macrophages.
So this suggests that CD8 cells were involved in the ongoing injury seen in this model. Perhaps the second part of that study suggests that these factors were independent of what was going on in the glomerulus.

We’ve also looked at this in our laboratory in a number of different ways. This first study which I’ll describe also focused on CD8 cells. But what we did here was to take pathogenic CD8 cells out of the glomerulus and then looked at the sequence of the T cell receptor and formed a vaccine using the sequences in that T cell receptor. The animals were treated with that vaccine and after several weeks of treatment there was a significant reduction in the interstitial infiltrates and I’ll show which cells in the next slide. There was also a reduction in tubular damage. In addition, there was a reduction in glomerulosclerosis and proteinuria, so once again our study doesn’t differentiate what’s happening in the interstitium from what also is happening in the glomerulus.

This is just looking at the different components of the interstitial infiltrate. If you look at this side compared to control groups here, there’s a reduction in the in the macrophage infiltration and the infiltration with CD8 cells and CD4 cells.

Now there are many other factors that are involved in inducing interstitial inflammation. A group of effectors are the co-stimulatory molecules. One co-stimulatory molecule that we’ve looked at in membranous GN or in Heymann nephritis is OX40. We’ve used a monoclonal antibody against OX40 and shown a reduction with its treatment in the interstitial infiltrates and in tubular damage. Once again however there’s a protection of the glomerular injuries, so less glomerulosclerosis.

So it doesn’t differentiate what’s going on in the glomerulus from the interstitium. In this study we showed an increase in the number of regulatory cells in the animals treated with the OX40 antibody and our feeling is that perhaps this increase in regulatory cells can partially explain the protective effect of blocking OX40.

Now we haven’t looked at that in this particular model, we have looked at it in other models such as adriamycin nephrosis and showed that if you use regulatory cells, be they T lymphocytes or macrophages you can protect against injury in proteinuric renal disease. So it’s likely that that is also the case in membranous. This slide just shows the reduction in infiltrate.

The last study which I’ll just mention only very briefly using an antibody against IL-6 receptor shows a similar protection and once again, an increase in the number of regulatory cells in the kidney suggesting that at least partially, this was one of the protective mechanisms.

So to summarise Interstitial Inflammation in Membranous Nephropathy predicts poor renal prognosis, but not independently of baseline clinical variables. In data I didn’t show you it does not correlate with the rate of decline of renal function or baseline proteinuria, and it does not preclude response to treatment.

The interstitial inflammation follows glomerular Ig deposition and proteinuria, and may be discordant with or occur without glomerular infiltration. It consists mainly of T cells and also macrophages, and a few regulatory cells and despite their small number the regulatory cells may be quite important. Rarely the interstitial inflammation may be part of the primary disease process, but more usually it’s secondary. In those rare primary cases it depends on the nature of the Ag and the immune complex. So most often it’s secondary to glomerular factors including urinary proteins, complement and pathogenic leukocytes and it may be a target for specific therapy. Thank you.

Chairman: Thank you Professor Harris. We have time for one or two questions. Please?

Question: Seeing T Regs coming up is an indicator that maybe new lymphatic vessels are growing. Did you have a chance to look at new lymphatic new angiogenesis?

Prof. Harris: We haven’t in membranous GN, we have in other models and others have shown that there are increased numbers of lymphatic vessels but we haven’t looked at it.