The Janus kinases — JAK1, JAK2, JAK3, and TYK2 — sit at one of the busiest intersections in immunology. Dozens of cytokines, from interferons to interleukins to growth factors, converge on receptors that have no enzymatic activity of their own. The signal only moves forward because a JAK is waiting on the receptor’s tail. That single fact explains both the therapeutic reach of JAK inhibitors and the central challenge of using them well: a drug that quiets one pathway rarely leaves the others fully untouched.

The short animation below walks through the family, the signaling cascade they drive, and the two distinct ways a small molecule can shut them down. The rest of this article unpacks each step.

01Four kinases, one family

All four JAKs are intracellular enzymes. None binds a cytokine directly. Instead, each is tethered to the cytoplasmic tail of a cytokine receptor, idle until its receptor is engaged. Three of them — JAK1, JAK2, and TYK2 — are broadly expressed. The fourth, JAK3, is restricted largely to hematopoietic cells and, uniquely, partners with only one receptor element: the common gamma chain. That restriction is not a footnote. It is why JAK3-selective inhibition is attractive for a narrowly immune effect, and why JAK3 biology reads so differently from its siblings.

The contrast in expression has direct clinical meaning. JAK2 is indispensable for erythropoietin and thrombopoietin signaling, so inhibition that reaches JAK2 carries a predictable cost in red cells and platelets. JAK1 sits at the center of inflammatory and interferon signaling, making it a frequent therapeutic target but also a broad one. TYK2 is more confined to interferon and the IL-12/23 axis, which is precisely why it has become a focus for agents seeking a narrower footprint. Reading a JAK inhibitor’s likely effects — both wanted and unwanted — begins with knowing which of these four it touches.

02One shared architecture

Every JAK is built from the same modular plan, read from the membrane inward. The FERM domain anchors the kinase to its receptor. The JH2 domain is a pseudokinase: it has no catalytic output of its own, but it regulates the enzyme, holding it in check until activation. The JH1 domain is the true catalytic engine — the part that phosphorylates downstream targets and propagates the signal.

Two of these domains matter enormously for drug design. JH1 is where classical inhibitors bind. JH2 is where a newer, more selective class of inhibitor works. We return to both below.

03Pairing defines identity

A JAK on its own does very little. What gives each kinase its functional identity is which partner it pairs with on a given receptor — and that pairing dictates which STAT transcription factors are ultimately activated.

Two pairings deserve emphasis because they are easy to conflate. The common gamma chain receptors run on JAK1 with JAK3. The IL-12 and IL-23 receptors run on JAK2 with TYK2 — not JAK2 alone, and not TYK2 alone. The therapeutic consequences of inhibiting one versus the other follow directly from these assignments.

04From cytokine to gene

When a cytokine binds, the receptor changes shape and draws its two tails together. That single motion brings the paired JAKs close enough to act on one another. They trans-phosphorylate — each activating its partner — and then phosphorylate tyrosine residues on the receptor tails themselves.

Those phosphorylated sites become docking platforms. STAT proteins are recruited to them, phosphorylated in turn, and then pair into dimers. The STAT dimer is the messenger: it travels from the membrane to the nucleus, binds DNA, and switches on transcription. Different receptor pairings recruit different STATs — STAT5 for many blood-cell signals, STAT4 and STAT3 for the IL-12/23 axis — so the identity of the cytokine is ultimately written into which genes are turned on.

The specificity of that final step is striking. The same handful of STAT proteins serve dozens of cytokines, yet the response stays faithful to the input because the receptor, the JAK pairing, and the docking sites together select which STATs are engaged and where they act. This is also why a perturbation high in the cascade — at the kinase — propagates so completely: remove the phosphorylation step and no docking sites form, no STAT dimer assembles, and the nuclear instruction is never written.

05Two ways to inhibit

Because the JAKs share one architecture, there is more than one place to interrupt them — and the choice of site is what determines selectivity.

Classical JAK inhibitors occupy the JH1 catalytic pocket and compete at the ATP-binding site. Because that pocket is highly conserved across the four kinases, achieving clean isoform selectivity is difficult — part of why pan-JAK and first-generation agents affect multiple pathways at once. A selective TYK2 inhibitor takes a different route entirely: it binds the JH2 pseudokinase domain, an allosteric site that holds the kinase off. Because JH2 is less conserved, this approach can target TYK2-dependent signaling — interferon type I, IL-12, and IL-23 — with greater precision.

It is worth being exact about what that TYK2 inhibitor does and does not spare. Blocking TYK2 dampens the pathways that depend on it, IL-12 and IL-23 included. The selective sparing of the IL-12/STAT4 axis is a property of a different drug class — oral IL-23-receptor antagonists that act at the receptor level — and should not be attributed to TYK2 inhibition. Keeping those mechanisms distinct matters when comparing agents.

06Why this matters at the bedside

The lesson of the JAK family is selectivity. Which kinase a drug inhibits, and where on that kinase it binds, together determine which cytokine pathways fall silent and which keep working. A broad inhibitor at the conserved catalytic site casts a wide net; an allosteric agent at a regulatory domain aims narrower. Neither is simply better — the right choice depends on which pathways a given disease is driven by, and which a given patient can least afford to lose. Read in that light, the shared architecture of these four kinases is not a complication. It is the map that makes targeted therapy legible.