After yesterday’s post on mRNA vaccines and RNA mechanisms, I wanted to highlight a completely non-coronavirus application that has recently appeared in the literature. It’s from BioNTech, the German company who’s been working with Pfizer on one of the two mRNA coronavirus vaccines, of course, but they have other therapeutic interests as well.
This latest paper is on an interesting approach to autoimmune disease. As it stands, the great majority of therapies available for such conditions are things that lower the immune response in general. And while that works (up to a point, and certainly not in everyone affected), it also has obvious problems, because you are increasing a person’s vulnerability to infection as well. There’s a long list of diseases that involve inappropriate inflammation of the body’s own tissues (multiple sclerosis, arthritis, type I diabetes, and lupus – just to take some big ones right off the top) and for many years now people have been trying to find a way to turn down those inflammatory pathways without turning down the whole immune response. Autoimmune disorders are generally a mixture of a person’s genetic background, the life history of their own particular immune systems and what it’s responded to over the years, and various environmental factors. The phrase “ill-defined” shows up a lot when you try to talk about the details of how those interact, though – it’s clear that all three of them are important to different degrees in different people, but it’s generally impossible to say exactly what’s going on, who in particular is at greater risk, and what in particular they might be able be able to do about it.
One goal has been to try to selectively affect autoreactive T cells, but that’s a lot easier said than done. The regulatory T cells and regulatory B cells are key players in immune tolerance, the “friend or foe” recognition system that keeps our own immune systems from attacking everything in sight. If you could present some of the antigens involved to those cells in a way that they accepted them as normal human proteins rather than as an external threat, you could presumably turn down their response. A lot of work over the years has gone into trying to do that – Figure 1 of that paper will show you over a dozen different approaches that people have tried. Among them are DNA vaccines, infusion of various blood cell types expressing or carrying these autoantigens, treatment with various polymers and nanoparticles carrying the autoantigens, and more. The autoantigen peptides themselves have been processed and modified in a number of ways for these different delivery systems as well. It’s fair to say that while some of these have shown some of the desired effects on immune cell populations, none of them have had dramatic effects on the clinical course of the autoimmune diseases themselves, Those have generally been MS or type 1 diabetes, not least because rodent models of both of them are available (not perfect models, to be sure, but definitely better than in many other therapeutic areas, and ones that have been used to validate other therapies in those diseases in the past).
This latest work looks at the mouse model for multiple sclerosis, experimental autoimmune encephalomyelitis (EAE). BioNTech and others have been trying to target the population of lymphoid antigen-presenting cells, known to be very important in immune tolerance mechanisms, but without setting off any of the general inflammation pathways. They have a liposomal formulation (see yesterday’s post) that when injected into the muscle tissue seems to end up almost entirely in the lymphatic system, and they’ve been doing all sorts of modifications to the RNA payload (such as replacement of uridine with methylpseudouridine) to make it as non-immunogenic by itself as possible. The liposome lipids themselves are also chosen to be as non-immunogenic as possible, too – the coronavirus mRNA vaccines actually get an adjuvant boost from such properties, but you don’t want that in this case.
The payload in this case codes for a known epitope of myelin oligodendrocyte glycoprotein (MOG), amino acids 35 to 55. Treating the mice with the modified RNA had a much different effect than with native uridine-containing RNA, as they had hoped. They also did not see signs of inflammatory cytokine signaling with the modified payload, but the treated mice were able to mount a normal inflammatory response afterwards, so it’s not that the treatment impaired their immune systems in general.
How about therapeutic effects on the disease? The modified MOG-epitope mRNA treatment “was capable of blocking all clinical signs of EAE in mice”, the paper says. At all points in the EAE progression in mice, treatment stopped the deterioration and restored various degrees of motor function. These effects lined up with a long list of blood and cellular markers, and were confirmed in a separate form of the EAE model.
A very important effect was the induction of “bystander tolerance”. All these antigen-presenting ideas have a shot at working if you know which antigens are causing trouble. But often we don’t, or there’s an obviously incomplete list of them. If one could generate regulatory T cells (Tregs)with the right activity, though, those cells could damp down the autoimmune response in general, without having to know what antigens were setting it off. The team saw strong evidence for such bystander effects in several models.
The mRNA platform presents some very interesting clinical opportunities. For example, there are antigens (such as MOG 35-55) that are thought to be important in many patients, but there are surely others that are peculiar to smaller subsets (or indeed, even to particular patients). If these can be identified, there’s a real possibility for individualized therapy, similar to what’s being doing with chimeric antigen receptor T-cells in cancer. The paper raises this possibility at the end, and it appears that BioNTech is thinking about making a serious run at this in the clinic.
I’m really interested in seeing how that works out. Effective immune therapy is probably going to have to end up rather targeted to individuals, given how different everyone’s immune system is, so a platform that allows such specificity is very welcome indeed. This also has application to the lesser-known autoimmune diseases (there are dozens and dozens of them), for which there are often very poor treatment options. This work looks like a real advance, and I hope to see it in human trials as soon as feasible.