The Janus kinase family comprises four intracellular tyrosine kinases JAK1, JAK2, JAK3, and TYK2 that transduce signals from a broad range of cytokine receptors to downstream transcription factors, principally the signal transducer and activator of transcription (STAT) proteins. JAK inhibitors as a drug class have been in clinical use since ruxolitinib (Jakavi; Novartis) received its first European authorisation in 2012, but the clinical story of the class has been substantially rewritten by the emergence of selectivity as a differentiating principle. Whether a compound inhibits JAK1 preferentially, JAK1 and JAK2 in combination, JAK3 in addition, or all four isoforms simultaneously is not a pharmacological curiosity, it determines which cytokine pathways are blocked, which immune functions are disrupted, and which adverse effects the prescriber must anticipate and manage.
This article explains what JAK isoform selectivity means at the molecular level, how it translates into the clinical profiles of the approved compounds, and what the MHRA and EMA class-level safety warnings mean in practice for the prescriber choosing between agents.
What the Four JAK Isoforms Do
Each JAK isoform associates with a specific subset of cytokine receptors and transduces a correspondingly specific subset of cytokine signals. Understanding which isoform does what is the foundation for understanding why selectivity matters clinically.
JAK1 partners with JAK2, JAK3, or TYK2 on heterodimeric receptor complexes that transduce signals from type I and type II interferons, and from interleukins including IL-2, IL-4, IL-6, IL-7, IL-9, IL-10, IL-13, IL-15, IL-21, and others. JAK1's breadth of cytokine coverage is the reason that JAK1-selective inhibition produces the most consistent anti-inflammatory effect across autoimmune diseases driven by multiple cytokine axes simultaneously.
JAK2 is required for signalling through the erythropoietin receptor, the thrombopoietin receptor, and the granulocyte colony-stimulating factor receptor, in addition to several cytokine receptors shared with JAK1. JAK2 inhibition consequently produces anaemia, thrombocytopenia, and neutropenia at clinically relevant exposures. This is why myelosuppression is the dominant dose-limiting toxicity of ruxolitinib (JAK1/2 inhibitor) in myelofibrosis, and why JAK2 co-inhibition in a compound intended for rheumatoid arthritis introduces a haematological risk profile that is mechanistically predictable and dose-dependent.
JAK3 pairs exclusively with the common gamma chain (γc) of cytokine receptors, transducing signals from IL-2, IL-4, IL-7, IL-9, IL-15, and IL-21, the interleukins essential to lymphocyte development, survival, and activation. JAK3 is expressed predominantly in haematopoietic cells. Its inhibition produces a degree of lymphocyte suppression that parallels, in mechanism if not in magnitude, the effects of calcineurin inhibitors. This is pharmacologically rational for transplant immunosuppression (where the JAK3 inhibitor ritlecitinib is in development) but introduces immunosuppressive risk in autoimmune populations who are already receiving background DMARDs.
TYK2 transduces IL-12, IL-23, and type I interferon signalling. Its relevance to the current approved small molecule portfolio is more limited the selective TYK2 inhibitor deucravacitinib (Sotyktu; BMS) received MHRA authorisation for moderate to severe plaque psoriasis in 2023, operating through a distinct allosteric mechanism at the TYK2 regulatory domain rather than at the ATP-binding active site shared by the other three isoforms.
The Selectivity Profiles of Approved Compounds
The term "selective" in the context of JAK inhibitors requires careful handling. All approved JAK inhibitors are kinase inhibitors that target the conserved ATP-binding site and none is absolutely monoselective at therapeutic concentrations in vivo. Published selectivity data are generated in biochemical assays using isolated enzyme preparations, and the resulting IC50 values establish a selectivity ratio rather than a binary selective/non-selective classification. At higher plasma concentrations which occur at the top end of the therapeutic dose range, in the presence of drug-drug interactions that increase exposure, or in patients with hepatic impairment compounds that are nominally JAK1-preferring will inhibit JAK2 at pharmacologically relevant levels.
Upadacitinib (Rinvoq; AbbVie) is licensed by the MHRA for rheumatoid arthritis, psoriatic arthritis, ankylosing spondylitis, atopic dermatitis, Crohn's disease, and ulcerative colitis. Its biochemical selectivity profile demonstrates approximately 60-fold selectivity for JAK1 over JAK2 in enzyme inhibition assays, placing it at the more selective end of the clinical portfolio. In whole-blood assays at clinically relevant concentrations, upadacitinib inhibits JAK1-dependent cytokine signalling (IL-6, IL-15) more completely than JAK2-dependent signalling (erythropoietin). The clinical consequence is that dose-dependent anaemia and neutropenia, while observed, are less prominent than with less selective agents at equivalent anti-inflammatory doses. Upadacitinib is available in immediate-release 15mg and extended-release 15mg, 30mg, and 45mg formulations; the higher doses used in atopic dermatitis and inflammatory bowel disease produce a broader cytokine inhibition profile that moves the compound further from its nominal selectivity ratio.
Baricitinib (Olumiant; Eli Lilly/Incyte) is licensed for rheumatoid arthritis and atopic dermatitis and has a JAK1/2 dual inhibition profile with approximately 5- to 6-fold selectivity for JAK1 over JAK2 in biochemical assays, substantially less JAK2-sparing than upadacitinib. It also has activity against TYK2 and, importantly, against AAK1 (adaptor-associated protein kinase 1), which has been proposed as a mechanism for its antiviral activity in hospitalised COVID-19 patients, the indication for which MHRA authorisation was granted in 2022 under an expedited assessment. The JAK2 co-inhibition profile means that baricitinib produces more consistent effects on haematopoietic signalling, and dose-dependent changes in haematological parameters require monitoring at the initiation and during treatment escalation.
Filgotinib (Jyseleca; Galapagos/Gilead) holds MHRA authorisation for rheumatoid arthritis and ulcerative colitis and is described as JAK1-selective, with a biochemical selectivity ratio for JAK1 over JAK2 of approximately 30-fold. It is the only approved JAK inhibitor in the UK for which an active metabolite, GS-829845, with similar JAK1 selectivity but lower maximum concentration than the parent compound, contributes meaningfully to the pharmacological profile. This metabolite pharmacology is relevant to drug-drug interaction assessment: inhibitors or inducers of CYP3A4 affect the parent compound's exposure, and the active metabolite contribution must be considered when interpreting interaction data.
Tofacitinib (Xeljanz; Pfizer) is licensed for rheumatoid arthritis, psoriatic arthritis, and ulcerative colitis. Its selectivity profile is the broadest of the approved autoimmune indications portfolio, with co-inhibition of JAK1, JAK2, and JAK3 at therapeutic concentrations. This pan-JAK-including-JAK3 profile was associated in the ORAL Surveillance cardiovascular safety study (Ytterberg et al., New England Journal of Medicine, 2022; 9,321 patients; median follow-up 4.0 years; funded by Pfizer) with statistically significant increases in major adverse cardiovascular events (MACE) and malignancy relative to tumour necrosis factor inhibitors in patients aged 50 years or older with at least one additional cardiovascular risk factor. This finding was the direct driver of the EMA and MHRA class-level safety review that followed and produced the label changes and prescribing restrictions that now apply across the entire JAK inhibitor class.
Real fact: The ORAL Surveillance trial enrolled 9,321 patients with rheumatoid arthritis aged 50 or older with at least one cardiovascular risk factor, making it the largest prospective cardiovascular safety study conducted for any JAK inhibitor and the primary basis for current MHRA and EMA class-level safety warnings.
The Class-Level Safety Warning and What It Means
Following the ORAL Surveillance findings, the EMA's Pharmacovigilance Risk Assessment Committee (PRAC) conducted a class review in 2022 and concluded that the cardiovascular and malignancy risks identified with tofacitinib should be treated as a class effect for all JAK inhibitors pending further data. The MHRA subsequently adopted equivalent labelling changes across the entire approved JAK inhibitor portfolio in the UK. The class warning identifies increased risks of serious cardiovascular events, malignancies (including lymphoma), venous thromboembolism, and serious infections in patients with relevant risk factors.
The mechanistic basis for applying the tofacitinib findings to JAK1-selective agents such as upadacitinib and filgotinib is pharmacologically contested. The JAK3 co-inhibition and the broader cytokine blockade of tofacitinib are not reproduced by JAK1-selective compounds at therapeutic doses, and the immune surveillance mechanisms affected differ between the two selectivity profiles. The class warning is a precautionary regulatory position rather than a conclusion derived from equivalent safety data across all agents; prospective long-term safety data for upadacitinib and filgotinib at the population scale of ORAL Surveillance do not yet exist.
In practice, the MHRA prescribing guidance requires that JAK inhibitors should be used with caution in patients aged 65 years or older, those who are current or former long-term smokers, those with other cardiovascular risk factors, those with other malignancy risk factors, and those with a history of venous thromboembolism. Patients who do not fit these risk profiles are not subject to the same degree of restriction, but the prescriber must assess the individual patient's risk-benefit balance and document that assessment.
CYP450 Interactions and Pharmacokinetic Considerations
JAK inhibitors as a class are metabolised primarily through CYP3A4, with variable contributions from CYP2C19 and other CYP isoforms depending on the compound. Tofacitinib is a CYP3A4 substrate; its exposure increases substantially with strong CYP3A4 inhibitors (ketoconazole, clarithromycin) and decreases with strong inducers (rifampicin). Upadacitinib is similarly CYP3A4-dependent. Baricitinib is primarily cleared by organic anion transporter 3 (OAT3)-mediated renal excretion with limited CYP involvement, making it a mechanistically distinct interaction profile OAT3 inhibitors such as probenecid increase baricitinib exposure, while the CYP3A4 pathway is less critical.
For the prescriber managing patients on background immunosuppressive therapy, combination DMARDs, or common comorbidity medications, JAK inhibitor interaction screening at initiation and at the time of any concomitant medication change is a clinical obligation rather than a pharmacological theoretical exercise. The combination of a JAK inhibitor with azathioprine or ciclosporin is contraindicated for most approved compounds due to additive immunosuppression risk; the combination with methotrexate is standard and covered within the licensed indications.
Selecting Between Compounds in Practice
The selectivity data, the adverse event profiles from clinical trials, and the class-level safety warning converge on a practical decision framework. In patients with rheumatoid arthritis who require a JAK inhibitor after inadequate response to biologic DMARDs, the prescriber should consider: the patient's cardiovascular risk profile (higher risk favours a JAK1-preferring agent and the lowest effective dose); the presence of significant anaemia or thrombocytopenia at baseline (which argues against compounds with greater JAK2 co-inhibition); the patient's age, smoking history, and malignancy risk (which trigger the class-level cautions regardless of selectivity); and the concomitant medication landscape (which determines interaction risk by mechanism, not by compound class alone).
The MHRA and EMA selectivity data are publicly available in the European Public Assessment Reports for each compound. The prescriber who reads the EPAR for upadacitinib and the EPAR for tofacitinib will find the biochemical selectivity ratios, the whole-blood assay data, and the haematological monitoring requirements presented in comparable formats that allow direct comparison. This is the level of mechanistic engagement the prescribing decision requires.
The JAK inhibitors are a class in which the molecular detail is not academic background, it is the clinical foreground. The selectivity ratio between JAK1 and JAK2 is the number that determines whether this patient's haemoglobin will fall, whether that patient's lymphocyte count will drop below the threshold that triggers dose interruption, and whether the choice between two nominally similar compounds is justified by something more substantive than habit or market share.
