JAK Inhibitors: Can't Help It 'Bout the Shape I'm In

 BY DEREK LOWE

Messing around with cytokine signaling is a very large business indeed (because it's a very large area of cell biology as well). We've heard a lot about it during the pandemic, for example, with the too-vigorous immune response that gets many severely infected people into trouble. And that illustrates the balance in modulating such signaling: it's absolutely necessary for the immune pathways to function at all, so you certainly don't want to turn everything down too hard. But having these pathways roaring full blast for an extended period is bad news as well (thus the two-phase treatment with Covid patients - you don't want to start turning down cytokine signaling too early, nor do you want to wait too long if someone has progressed to the point where it's a problem).

Autoimmune conditions are a big reason that this is such a large market for the drug industry, of course: rheumatoid arthritis, psoraisis, and other conditions can indeed be improved if you just reach and in turn down an inappropriate or overactive immune response. There are a lot of ways to do this, which makes sense, considering the amount of biology involved. Autoimmune patients generally get methotrexate first, although the details of its effects on the immune system are complex and frankly not understood as thoroughly as you'd imagine for a drug that's been around since the Trumann administration. It can also be given with some of the newer agents, such as antibodies or fusion proteins that disrupt TNF-alpha signaling.

Those biologics are among the best-selling drugs in the world, and there have been many proposals to try to break into that market with small molecules rather than injected therapies. One class that's had massive amounts of work put into are the Janus kinase (JAK) inhibitors. I'm impressed that that Wikipedia link mentions that "JAK" originally stood for "Just Another Kinase" (it's true), but it turns out that they have two very similar phosphorylation domains (one regulatory, one acting on substrates), so retconning them with the two-faced Roman god worked out fine. (Not all of our benchtop nomenclature holds up so well. I wonder how many people who think about hERG assays realize that the acronym goes back to the Drosophila zanies and their mutant flies who looped around in ether vapors: the human Ether-a-Go-Go receptor. . .)

JAK (there are several subtypes) is a middleman in a lot of cytokine signaling, with the JAK-STAT pathway being a mighty bridge between extracellular signaling and downstream effects on gene transcription. And since we know an awful lot about making kinase inhibitors it was recognized early on as a promising drug target. Several of them have made it to market: Xeljanz (tolfacitinib), whose name is a leading candidate for the drug that sounds most like a planet in an old science fiction story and which hits both JAK1 and JAK3,  Olumiant (baricitinib, JAK1 and JAK2), Jyseleca (filgotinib, more JAK1 selective), Rinvoq (upadacitinib, more JAK1 selective), Jakafi (ruxolitinib, JAK1 and JAK2), and Inrebic (fedratinib, more JAK2 selective). Those first four are all approved as anti-inflammatories for autoimmune disease; the last two are used in myelofibrosis and some cancers. And those have shown good efficacy; they're the first oral drugs whose effects on rheumatoid arthritis can match those of the anti-TNF biologics. 

You'll note the various subtype selectivities mentioned, but those are, believe me, only a very rough guide. The differential selectivities of these compounds is a very knotty subject, especially in its implications for therapeutic efficacy and for incidence of side effects. That last part has been the subject of some news just recently. The FDA required Pfizer to do a post-approval study of Xeljanz to further characterize those effects versus the biologic therapies, and that one read out earlier this year. It wasn't reassuring: Xeljanz did not meet the non-inferiority endpoints for cardiovascular and oncology side effects. This was in an older population with known cardiovascular risk factors (and oncology risk factors such as smoking), but that's the real-world population for a lot of those RA prescriptions, of course. 

As the old Peter Green/Fleetwood Mac song had it, "Don't ask me what I think of you - I might not give the answer that you want me to". So it is with clinical trial data, and the question has been what the FDA would do with this information. Today they required warning labels on Xeljanz, Olumiant, and Rinvoq. Jyseleca isn't approved yet in the US, and Jakafi and Inrebic aren't approved for indications like RA at all. No one was particularly surprised at this for the first two, but Rinvoq (as mentioned above) is a more selective second-generation compound, and Abbvie (and its investors) weren't expecting it to be lumped in on the warning label requirement. As you'll see from that link, it looks like the entire JAK class is already slipping behind the biologics with physicians and insurors, and all this is going to run right into plans to extend their use to less severe autoimmune conditions like atopic dermatitis. It may still be that some JAK inhibitor types are cleaner than others, or show better efficacy in particular conditions that will raise their risk/benefit above the rest. But it's not going to easy to sort this out - if Abbvie or any other company would like to try to prove it, they're welcome to go for it in a controlled trial. And that's probably the only thing that would really convince everyone, but you're looking at several years of work and expense. If you enjoy making such calls, then a career as a pharma executive might be for you!

https://www.science.org/content/blog-post/jak-inhibitors-can-t-help-it-bout-shape-i-m