Janus kinase inhibitors for the treatment of rheumatoid arthritis demonstrate similar profiles of in vitro cytokine receptor inhibition

Abstract Janus kinase (JAK) inhibitors have emerged as an effective class of therapies for various inflammatory diseases such as rheumatoid arthritis (RA). JAK inhibitors function intracellularly by modulating the catalytic activity of JAKs and disrupting the receptor‐mediated signaling of multiple cytokines and growth factors, including those with pro‐inflammatory activity. Understanding the inhibition profiles of different JAK inhibitors, based on the associated cytokine receptors and downstream inflammatory pathways affected, is important to identify the potential mechanisms for observed differences in efficacy and safety. This study applied an integrated modeling approach, using in vitro whole blood cytokine inhibition potencies and plasma pharmacokinetics, to determine JAK‐dependent cytokine receptor inhibition profiles, in the context of doses estimated to provide a similar clinical response in RA clinical trials. The calculated profiles of cytokine receptor inhibition for the JAK inhibitors tofacitinib, baricitinib, upadacitinib, and filgotinib and its metabolite, were generally similar when clinically efficacious doses for RA were considered. Only minor numerical differences in percentage cytokine receptor inhibition were observed, suggesting limited differentiation of these inhibitors based on JAK pharmacology, with each showing a differential selectivity for JAK1 heterodimer inhibition. Nevertheless, only robust clinical testing involving head‐to‐head studies will ultimately determine whether there are clinically meaningful differences between these JAK inhibitors. Furthermore, ongoing and future research into inhibitors with alternative JAK selectivity remains of clinical importance. Thus, all JAK inhibitors should be characterized via thorough preclinical, metabolic and pharmacological evaluation, adequate long‐term clinical data, and when available, real‐world experience.


| INTRODUC TI ON
Inflammatory diseases, such as rheumatoid arthritis (RA), are chronic immune-mediated conditions, which collectively have an estimated prevalence of 5%-7% in Western society, 1 can be highly disabling 2 and painful, and cause reduced quality of life. 3 Inflammation is a complex immune response involving diverse cell types and feedback loops promoting the production of pro-inflammatory mediators. 1,4 Notably, cytokines are a group of structurally unrelated protein messengers that, upon binding to and activating their specific receptors on immune cells, transmit signals that regulate the expression of numerous inflammatory genes. Under dysregulated conditions, cytokines play a key role in the pathologic inflammatory response characteristic of immune-mediated diseases. 1,5,6 Janus kinases (JAKs) are enzymes that are essential in the signaling pathways of type I and type II cell-surface cytokine receptors which lack intrinsic kinase catalytic activity. 5,6 There are four different JAK isoforms in humans (JAK1, JAK2, JAK3, and TYK2) which function in pairs to transmit intracellular signals from cytokine-activated receptors. 5 JAK1 pairs with JAK3 to mediate the signaling pathways of common gamma chain (γc) cytokines. JAK1 also pairs with JAK2 and/or TYK2 for signaling through receptors of the IL-6, IL-10, and interferon cytokine families. Additionally, JAK2 pairs with TYK2 for signaling through IL-12 and IL-23 cytokine receptors, and with itself for signaling from receptors for hormone-like cytokines such as erythropoietin (EPO). 5 Indeed, there are over 50 cytokines that signal through JAKmediated type I and II receptors, 6 many implicated in inflammatory disease pathophysiology. 1,4 Notably, IL-6 induces acute-phase proteins such as C-reactive protein (CRP) and may be involved in the autoimmune process through B-cell modulation and T-helper-17-cell differentiation. 7 Common γc cytokines play a key role in adaptive immune functions, for example, in T-cell and natural killer (NK)-cell differentiation. 5 JAKs are therefore an attractive therapeutic target for RA and other inflammatory diseases. 6 Biologic drugs for treating inflammatory diseases target extracellular elements of the inflammation pathway such as cytokines or their receptors. 1,4,5 In contrast, targeted synthetic JAK inhibitors reduce inflammation by directly binding to and modulating the intracellular catalytic activity of JAKs and disrupting the receptor-mediated signaling of multiple cytokines, including those of pro-inflammatory pathways. 4 Current JAK inhibitor drugs were designed to be selective for certain JAK isoforms. 8 Tofacitinib is an oral, small molecule JAK inhibitor for the treatment of RA, psoriatic arthritis, and ulcerative colitis.
Tofacitinib is reported to have functional selectivity for heterodimer pairs involving JAK1 and/or JAK3. 4,5 Other JAK inhibitors with ongoing or completed late development RA clinical trials include baricitinib, upadacitinib, and filgotinib. Baricitinib is approved for the treatment of RA, with reported preferential selectivity for JAK1 and JAK2. 9 Upadacitinib and filgotinib are under investigation for the treatment of inflammatory diseases including RA; both drugs, as well as filgotinib's active metabolite, are reported to selectively inhibit JAK1. [10][11][12] Understanding the different inhibition profiles of JAK inhibitors, based on the associated cytokine receptors and downstream inflammatory pathways inhibited, is important to better characterize the impact of JAK inhibition and the potential rationale for differences in clinical efficacy and safety profiles. Although JAK selectivity can be evaluated using enzymatic assays, the selectivity observed using biochemical assays may not necessarily be maintained when evaluated under physiologic cellular conditions. 13 Suggested reasons for this discrepancy include differences within the complex intracellular milieu, particularly when assessing activity in primary cells (ie, human whole blood), such as the difference in the adenosine triphosphate (ATP) Michaelis-Menten constant (K M ) of each kinase. 13 The objective of this study was to characterize cytokine receptor inhibition profiles of JAK inhibitors for the treatment of RA, in the context of drug doses that provided a similar response in patients in RA clinical trial settings. To achieve this, we applied an integrated modeling approach, using knowledge of both intracellular JAK-dependent cytokine signaling inhibition potencies and in vivo pharmacokinetics.  were obtained from R&D Systems. IL-21 (Catalog No. AF-200-21) was purchased from PeproTech.

| Antibodies
Antibodies specific to phosphorylated signal transducer and ac-  Human whole blood was collected from 13 healthy donors (seven males and six females) via venipuncture into Vacutainer collection tubes containing sodium heparin, in accordance with Pfizer protocols (Protocol No. GOHW RDP-01) approved by the Shulman Institutional Review Board. Blood was warmed to 37°C prior to use.

| Enzymatic potency of JAK inhibitors
The potency of tofacitinib, upadacitinib, baricitinib, and filgotinib and its metabolite against the four JAK isoforms, JAK1, JAK2, JAK3, and TYK2, was measured in terms of half-maximal inhibitory concentration (IC 50 ). Human JAK activity was determined using a microfluidic assay to monitor phosphorylation of a synthetic peptide by the recombinant human kinase domain of each of the JAK isoforms.
Test compounds were solubilized in DMSO to a stock concen-

| Whole blood potency
Inhibition curves and IC 50 values were determined for cytokine signaling of tofacitinib, baricitinib, upadacitinib, and filgotinib and its metabolite. For each cytokine assay, all five compounds were tested side-by-side in quadruplicate (ie, using blood from four donors).

| Plasma protein binding
An equilibrium dialysis method was used to determine the plasma fraction unbound (f u ) values, as described previously. 14

| Blood-to-plasma ratio
Human blood-to-plasma ratio was measured by Unilabs York Bioanalytical Solutions. Test compounds were incubated in quadruplicate with fresh human blood (mixed gender, at least 1 sample per gender, Clinical Trials Laboratory Services Ltd, London, UK) at 1 μmol/L in a humidified incubator (95% relative humidity; 5% CO 2 /95% air) for 1 and 3 hours at 37°C with shaking at 450 RPM.
Following incubation, plasma samples were obtained by centrifuging blood samples at 3000g for 7 minutes. Both plasma and blood samples were matrix-matched with each other and quenched with acetonitrile containing internal standard. The solutions were centrifuged, and the supernatant was analyzed by LC-MS/MS. Peak area ratios were used to calculate blood-to-plasma ratio.

| Data analyses
An integrated modeling approach was applied to determine cytokine receptor inhibition profiles. Whole blood IC 50 values were converted to unbound values (IC 50,u ) using measured blood-to-plasma ratios (2) IC 50,u = IC 50 blood:plasma ratio × f u plasma concentration was used as it has been shown to be a predictive exposure metric of the efficacy of tofacitinib and filgotinib, rather than C max or C min . 19   and using blood from four donors per cytokine. Additionally, Figure   S1 presents IC 50 curves (one of the four obtained) for representative cytokine receptors from different receptor classes (IFNγ, IFNα, IL-6, IL-15, IL-12, and EPO) for illustrative purposes.

| Blood-to-plasma ratios
From quadruplicate measurements, mean blood-to-plasma ratios were determined to be 1.20 for tofacitinib, 1.32 for baricitinib, 1.16 for upadacitinib, 1.22 for filgotinib, and 1.09 for filgotinib metabolite.       The IC 50 values determined in this study were consistent with the literature, with each JAK inhibitor demonstrating greater selectivity for JAK1 vs other isoforms. 4,5,[9][10][11] JAK1 is presumed to be a key target for RA and other inflammatory diseases since it associates with receptors for γc cytokines, interferons, type II cytokine receptors (eg, IL-6), and other interleukins. 5,22 Also per the literature, tofacitinib showed comparative selectivity for JAK3, and baricitinib showed comparative selectivity for JAK2. 4,5,9 Cytokine receptor inhibitory concentration (IC xx ) profiles were generally similar among the JAK inhibitors, with small numerical differences in percentage cytokine receptor inhibition, suggesting

TA B L E 2 (Continued)
3 RA studies of upadacitinib and filgotinib. 17,18 As such, the occurrence of HZ is likely a "class effect" of inhibiting at least JAK1, 34 although viral reactivation could also be dependent on the overall impact of JAK inhibition.
Notably, IC 50 values for filgotinib and its metabolite were not and EPO activity and production. 6,35,36 In a pooled analysis of two LTE studies of tofacitinib 5 or 10 mg BID for RA, platelet counts initially decreased from baseline then stabilized over time, and hemoglobin levels increased from baseline over time. 37 In the tofacitinib LTE study ORAL Sequel, increases in hemoglobin levels were observed from baseline to month 24, which then remained stable to month 96. 38 In contrast, in an integrated analysis of phase 1b, phase 2, phase 3, and LTE RA studies, baricitinib 4 mg QD was associated with an initial increase in platelet counts, which then returned toward baseline; hemoglobin levels decreased from baseline to week 20, then returned to baseline or higher. 33  TA B L E 3 Actual and unbound C av values for clinically meaningful doses of tofacitinib, baricitinib, upadacitinib, and filgotinib and its metabolite robust settings is required. For example, BMS-986165 has not been associated with laboratory changes such as increased lipid levels, 44 which can occur with IL-6 inhibition. 46 In conclusion, by applying an integrated modeling approach, the JAK inhibitors tofacitinib, baricitinib, upadacitinib, and filgotinib, at doses conveying reasonably equivalent clinical efficacy for RA, exhibited generally similar cytokine receptor inhibition profiles. Although some small numerical differences were observed, these do not appear to translate to significant differences in the JAK inhibitors' clinical profiles. At the same time, we appreciate that only robust clinical testing involving head-to-head studies may determine whether there are clinically meaningful differences between these JAK inhibitors. All JAK inhibitors, including novel second-generation compounds with minimal JAK1 effects, need to be characterized via thorough preclinical, F I G U R E 1 Cytokine receptor inhibitory concentrations for modeled exposures of tofacitinib 5 mg BID, baricitinib 4 mg QD, upadacitinib 15 mg QD, and filgotinib/metabolite 200 mg QD. BID, twice daily; CD, cluster of differentiation; EPO, erythropoietin; G-CSF, granulocyte colony-stimulating factor; GM-CSF, granulocytemacrophage colony-stimulating factor; ICxx, proportion of JAK inhibitory concentration; IFN, interferon; IL, interleukin; JAK, Janus kinase; pSTAT, phosphorylated signal transducer and activator of transcription protein; QD, once daily; TPO, thrombopoietin; TYK, tyrosine kinase metabolic and pharmacological evaluation, adequate long-term clinical data, and when available, real-world experience.