Slide 1

Thanks very much indeed. I’ve been asked to talk about antibody blockade of TNF-α in experimental crescentic glomerulonephritis.

Slide 2

As you know, crescentic glomerulonephritis can be caused by a number of different immunological mediators. For example, anti-GBM antibody and Good Pasture’s disease, immune complex deposition as in systemic lupus and the so-called pauci-immune glomerulonephritis where very few deposits are seen but over 80% of cases are associated with anti-neutrophil cytoplasm antibodies which of course, lead on to leucocyte activation.

Slide 3

Now, whichever of these pathways underlies the form of glomerulonephritis there are common inflammatory mechanisms involved. One of the most potent inflammatory mechanisms is that induced by TNF-α. TNF-α binds to two main receptors but most of its biological effects are mediated through the TNFR1 receptor and this forms a trimeric complex which then activates the TNF receptor associated death domain and through a number of intermediates activates the transcription factors NF-kB and AP-1. As you know, these transcription factors turn on genes responsible both for inflammation and cell survival. If for any reason this pathway is blocked, there’s an alternative pathway whereby this receptor is internalised and then acting through other intermediates including the caspases can lead to cell apoptosis.

Slide 4

What I’m going to be talking about really is inhibition of this pro-inflammatory TNF-α pathway. Now, we showed many years ago that TNF was involved in glomerular inflammation and this is a short-term study, a 24 hour study where animals were injected with a nephrotoxic globulin and then we exposed them to injections of an increasing amount of recombinant TNF. As you can see, the more TNF they have, the higher their albumin urea and they also had an increased number of capillary thrombi.

Slide 5

We then went on to see if we could antagonise the effects of LPS on these glomeruli using a soluble TNF receptor. Here you can see there was not much albuminuria with the nephrotoxic globulin alone, we chose a low dose. If you combine that with LPS, there was significant albuminuria and when we combined that with a soluble TNF receptor, again there was a significant improvement both in albuminuria and in glomerular thrombi. So we’ve proved in principal that blocking TNF prevents the enhancing effects of LPS.

Slide 6

However, this was only a very short-term model in which there were no crescents formed. So we went on to develop a longer-term model of nephrotoxic nephritis in the WKY rat which is very susceptible to this type of disease. Here when we inject the anti-GBM antibody, we see a crescentic glomerulonephritis by about 2 weeks. You can see this is deposition of a rabbit anti-rat GBM antibody and this is the associated crescent.

Slide 7

Again, if we use the soluble TNF receptor in this model over a 10 day period starting at day 4 continuing to day 10 which is the time of maximum crescent formation, you can see there was a significant reduction in albuminuria here with the soluble receptor and also almost a complete abolishment of crescent in this same model.

Slide 8

We wanted to look at the longer term effects of inhibiting TNF. For this we used a monoclonal anti-TNF antibody and here I show you proteinuria and renal function. In the left hand panel you can see that the anti-TNF antibody significantly reduced proteinuria in this model throughout the course up to 28 days. The control animals here have increasing amounts of proteinuria until 3 weeks and that then falls. This is not because they are getting better, this is because they are going into renal failure as shown here by their serum creatinine. The control animals show a continuing rise in creatinine, go into renal failure within another week or two, whereas the anti-TNF treated animals are completely protected from that.

Slide 9

This is shown here in this slide the controls showing glomerular scarring and a lot of tubular interstitial scarring and infiltration. I’ve deliberately shown you low powers, so you can have a look at the architecture of the kidney. With the anti-TNF the glomeruli are preserved with only minor abnormalities and the tubular interstitium is also normal.

So, we’ve had a marked effect on the architecture of the disease at 28 days. We also measured a number of measures of scarring including type IV collagen. You can see in the control animal these glomeruli and tubules are getting very severely scarred with deposition of type  IV collagen, anti-TNF completely prevents collagen formation and deposition.
So you can see in this model of anti-GBM disease in which the animals actually progressed to renal failure, treatment with anti-TNF both prevents the acute crescent formation and also prevents subsequent development of scarring and renal failure.

Slide 10

We then wanted to move on to study a model of vasculitis since vasculitis is a much commoner form of crescentic glomerulonephritis. We did this by immunising the WKY rats with myeloperoxidase in Freund’s adjuvant, so they then make their own autoimmune response against MPO and effectively they make MPO-ANCA. They develop crescentic GN rather slowly by 6 weeks. These are the ANCA, this is EAV serum placed on rat neutrophils, you can see a typical perinuclear ANCA pattern. This is a crescent which only forms in perhaps 20%, 10-20% of glomeruli but it’s very consistent in all animals.

Slide 11

In this model we have the advantage we can study some of the leucocyte, endothelial cell interactions in vivo using intravital microscopy. This is best done here in the mesenteric vessels. You can see here at a gross level little microvascular haemorrhages in the small mesenteric vessels.

Slide 12

At a microscopic level you can see these vessels with haemorrhage surrounding them. Now, if we take the exteriorised mesentery of the rats with EAV, the experimental vasculitis and a pleiotopical chemokine, we can then actually see the adherence transmigration and production of haemorrhage. This is a larger vessel, this is a small vessel down there, another large vessel and I just hope this is going to work, it’s not going to work, it worked earlier. What I’ll do rather than waste time, I won’t show you it, which is a great pity because it was a nice video. I might come back to it at the end but what you can see is with the presence of ANCA you can see leucocytes actually adhering to this vessel transmigrating and you can see an area of microvascular haemorrhage developing in real time, if the video worked.

Slide 13

So what of anti-TNF therapy in this model? Here you can see anti-TNF antibodies produce a significant reduction in tubulointerstitial nephritis and also in glomerulonephritis where they almost completely prevent crescent formation here.

Slide 14

Again, this is shown by the light microscopy with crescents in the control animals treated with a control antibody, normal glomeruli with the anti-TNF antibody and similarly the controls show a large infiltration with macrophages by 6 weeks which is a predominant inflammatory cell in this model and that’s almost completely prevented by anti-TNF antibody.

Slide 15

The animals also develop lung haemorrhage as part of their systemic vasculitis and this is significantly prevented by the anti-TNF antibody, this is a lung --- count and this just shows the gross appearance of the lungs with clear lungs with no --- in the treated animal and multiple small --- in the lungs of the control animals.

Slide 16

Now, we can actually quantify the amount of cells that are transmigrating using intravital microscopy and you can see here this shows adherent cells in the top panel and this shows transmigrated cells. In fact, the monoclonal antibody shows an increased number of adherent cells and there are more adherent cells because there’s a block to their transmigration. So, if you look at this lower graph, you’ll see there are far fewer transmigrating cells in the anti-TNF antibody treated animals. So, we can say at least in vasculitis that one mechanism of reducing inflammation is it prevents the adhesion and the extravasation of leucocytes across the endothelium.

Slide 17

Now, the other model of vasculitis is that developed by Charles Jennette’s group in Chapel Hill in which they used the mouse and they induced an anti-MPO antibody by immunising MPO knockout mice with MPO, human MPO and since the knockout mice regard this as a foreign antigen they make an extremely strong antibody response. They have high levels of anti-MPO antibodies these can then be transferred to naïve wild type mice and this gives the rapid crescentic GN. This shows the immunofluorescence of the glomeruli that’s pauci-immune so it’s not an immune complex disease and this shows crescent formation at 1 week in these mice.

Slide 18

Now, this disease is also greatly enhanced by either LPS or TNF as we previously found in our anti-GBM model. This shows anti-MPO antibodies alone causing segmental glomerular lesions. If LPS is added, you can see there’s far more both crescent formation and interstitial inflammation. This is reflected here in the stain for the macrophages. Anti-MPO antibodies alone, you have a few macrophages, add LPS and there are far more macrophages accumulating.

Slide 19

If these animals are then given an anti-MPO antibody, you can see there’s a significant reduction in the crescent formation in those receiving the antibody compared to the control monoclonal.

Slide 20

So in two different models of vasculitis here we’ve seen a significant effect of blockade of TNF. Now, although I was asked to talk about experimental models I will just show a very quick slide of our clinical study. This was an open-label pilot study it wasn’t a controlled study of 32 patients with vasculitis either Wegener’s or microscopic polyangiitis, 16 of whom presented with acute disease and 16 had persistent or grumbling disease activity, they all received infliximab which is the anti-TNF antibody widely used in rheumatoid arthritis and Crohn’s disease and it was used at 0, 2, 6 and 10 weeks in a rheumatoid arthritis type schedule and there was a very rapid remission of about 90% of patients in both groups. Now in the acute disease group we did treat them with conventional steroids and immunosuppression, so it’s hard to interpret that although we think they did probably respond more quickly, they certainly felt better almost immediately. But in study two we left them on their background maintenance treatment and the antibody was the only change in their therapy. So, in group 2 for 90% to go into remission I think suggests that that was a biological effect of the anti-TNF antibody. There was, of course, a 20% relapse after stopping the TNF therapy, so it certainly didn’t completely control the disease and there was 20% infection but it was significantly steroid sparing which might indeed be a good role for this approach.

Slide 21

If we look at the Birmingham Vasculitis Activity Score, you can see here in study 1 the acute patients, in study 2 the grumbling patients there was a very rapid fall in vasculitis activity which was sustained throughout the period of treatment.

Slide 22

So, if I could conclude, we’ve shown that glomerular inflammation in many experimental models of crescentic GN is dependent on TNF. We’ve shown that anti-TNF therapy is effective both in models of anti-GBM disease and of systemic vasculitis. I think the initial clinical studies of anti- TNF antibody in systemic vasculitis are encouraging but clearly randomised controlled trials are required before we can conclude that for sure.
Finally I’d like to thank my many colleagues at Imperial College and the Hammersmith hospital for the work with the rat models that I showed you and many collaborators with all of the other studies. Thanks very much for you attention.

Slide 23

Chairman: Thank you very much Charles for this very interesting conference. Any comments, any questions from the audiences? Charles in the last years there have been some reports….

Dr Pusey: Look bottom left there are the leucocytes, they’re rolling there adhering and in the bottom left you see this little vessel can you see? It’s developing haemorrhage as you watch due to the ANCA. There are more and more adhering as the video runs and this area of haemorrhage gets progressively larger and the cells there look they’re stopping, that little vessel’s becoming completely blocked and these changes can be inhibited in this model with the anti-TNF antibody. I’m very sorry it didn’t work before, obviously my technical ability was not very good. Thank you. Sorry. I’m glad you saw that because that’s really nice. Ok back to the question.

Question: --- Poland. I wish to congratulate you on this presentation but I’d just like to hear a comment about a more particular issue. You have such a remarkably high rate of remission with 88% or almost 90% rate of remission with your infliximab treatment. Given the complexity of pathogenicity of this disease in humans, isn’t it astonishing in a sense that such a high, I think it’s quite unique, isn’t it?

Dr Pusey: Most of the more recent studies using conventional immunosuppression initial remission about 80% have been achieved in some of our European vasculitis studies. So I’m not saying that our remission rate is much, much greater than that. Obviously in study one they did also receive conventional drugs. They didn’t receive an adjuvant therapy like plasma exchange or methylprednisolone, they received infliximab instead of that but it’s hard to interpret but I agree with you in the second, the chronic grumbling disease patients who often had disease outside of the kidneys perhaps in the upper respiratory tract or the lungs, they did seem to show a very rapid and a very good response and this really reminds you of the sort of response that’s seen in rheumatoid arthritis where a great many patients at least show an initial significant response. So you’re right it is fairly remarkable but I think perhaps less so than you might think. It’s rather like rheumatoid arthritis and it’s in the chronic phases and I think can I also add there is a trial out there which I didn’t comment on because it wasn’t really in my remit called the Wegener’s granulomatosis etanercept trial and this is a trial run in America using a soluble P75 TNF receptor in patients with largely non-renal vasculitis and they didn’t manage to show a significant benefit. It’s not quite clear whether perhaps they weren’t using sufficient of the drug or perhaps there’s a biological difference between the soluble receptors and the TNF antibodies. So I think that trial, the WGET trial we call it is out there and does suggest that in the chronic Wegener’s population a course of the soluble receptor did not seem to be significantly beneficial but I think that does not mean that the anti-TNF antibodies are not beneficial because they work in a different way, they bind to membrane bound TNF and they can cause cellular apoptosis. They have different effects to the soluble receptor.

Question: Charles, there are some reports about patients with rheumatoid arthritis that develop crescentic glomerulonephritis paradoxically after the treatment with anti-TNF antibodies. Do you have any comments about this?

Dr Pusey: Yes, I’m aware of that and also sometimes immune phenomenon like anti-DNA antibodies. I haven’t ever seen that happening myself but I guess it maybe because blocking TNF is inhibiting various parts of the immune system and if, for example, it’s inhibiting induction of apoptosis of abnormal cells, it could lead to immune phenomena but in the doses we used I never saw that. But you’re right we have to be careful when we’re modifying such an important mediator, we must be very careful of other effects on the immune system. But I think the outstanding safety and efficacy of anti-TNF antibody in rheumatoid and in Crohn’s disease suggest that it will be a safe approach to use and I think probably most of you will know that the anti-TNF antibody is very effective in Crohn’s disease but the soluble receptor is in fact, not very effective. So again, Crohn’s disease and vasculitis they might be similar maybe that explains the WGET trial.

Chairman: Ok thank you very much.