Effectiveness and Safety of JAK Inhibitors in Autoinflammatory Diseases: A Systematic Review

Introduction Autoinflammatory diseases (AID) are rare diseases presenting with episodes of sterile inflammation. These involve multiple organs and can cause both acute organ damage and serious long-term effects, like amyloidosis. Disease-specific anti-inflammatory therapeutic strategies are established for some AID. However, their clinical course frequently includes relapsing, uncontrolled conditions. Therefore, new therapeutic approaches are needed. Janus Kinase inhibitors (JAKi) block key cytokines of AID pathogenesis and can be a potential option. Methods A systematic review of the literature in accordance with the PRISMA guidelines was conducted. Three databases (MEDLINE, Embase and Cochrane Central Register of Controlled Trials) were searched for publications regarding the use of JAKi for AID. Data from the included publications was extracted and a narrative synthesis was performed. Criteria for defining treatment response were defined and applied. Results We report data from 38 publications with a total of 101 patients describing the effects of JAKi in AID. Data on Type I Interferonopathies, Adult-Onset Still's Disease (AOSD), Systemic Juvenile Idiopathic Arthritis (sJIA), Familial Mediterranean Fever (FMF), and Behçet's Syndrome (BS) was identified. From a total of 52 patients with type I interferonopathies, in seven patients (7/52, 13.5%) a complete response was achieved, most (35/52, 67.3%) showed a partial response and a minority (10/52, 19.2%) showed no treatment response. For AOSD, a complete or a partial response was achieved by eleven (11/26, 42.3%) patients each. Two sJIA patients achieved complete response (2/4, 50%) and in two cases (2/4, 50%) a partial response was reported. Half of FMF patients showed a complete response and the other half had a partial one (3/6, 50.0%). Amongst BS patients most achieved a partial response (8/13, 61.5%). Five patients showed no response to therapy (5/13, 38.5%). Overall, the most frequent AEs were upper respiratory tract infections (17), pneumonia (10), BK virus viremia (10) and viruria (4), herpes zoster infection (5), viral gastroenteritis (2) and other infections (4). Conclusion The results from this systematic review show that JAKi can be beneficial in certain AID. The risk of AEs, especially viral infections, should be considered. To accurately assess the risk benefit ratio of JAKi for AID, clinical trials should be conducted.


INTRODUCTION
Autoinflammatory diseases (AID) are characterized by seemingly unprovoked inflammatory attacks in absence of pathogenic autoantibodies or antigen-specific T-cells. Defined as mono-and polygenic disorders of innate immunity, AID comprise a broad spectrum of rare diseases which may present with episodes of fever and sterile inflammation potentially causing severe morbidity and mortality. Due to advances in gene sequencing technology and the development of diagnostic criteria, new syndromes continue emerging (1,2).
Current management of AID includes targeted inhibition of specific cytokine signaling. For example, targeted IL-1 inhibition has been shown to be effective for some conditions such as cryopyrin-associated periodic syndromes (CAPS) (15) and AOSD (16). Unfortunately, some AID patients do not respond to targeted inhibition of specific cytokines and other treatment options are needed (17).
Janus Kinase inhibitors (JAK inhibitors, JAKi) interfere with signal transduction of the Janus Kinase-Signal transducer and activator of transcription (JAK-STAT) pathway causing effective suppression of downstream cytokine signaling. JAK-STAT signaling can be triggered by two types of cytokine receptors: type I receptors bind mainly cytokines (IL-2,−6,−9,−12,−15), hormones (growth hormone, GH) and colony stimulating factors, while type 2 receptors are activated mostly by interferon and IL-10 (18). They act as competitive antagonists at activation sites for Janus kinases and as such interrupt downstream signals along the JAK-STAT pathway, effectively leading to suppression of cytokine production. The JAK-STAT pathway includes several kinases and JAKi can be grouped by their kinase-specific effects: tofacitinib-JAK1, JAK2 and JAK3, baricitinib and ruxolitinib-selective inhibition of JAK1 and JAK2, upadacitinib and filgotinib-selective for JAK1. While JAKi are considered as "targeted therapies, " there is almost no other substance class that exerts an effect on such a large number of cytokines. The resulting immunomodulatory effects can be clinically illustrated by the fact that the drugs have already been approved for a number of rheumatic diseases such as rheumatoid arthritis (RA), psoriatic arthritis (19,20), polyarticular juvenile arthritis (tofacitinib) (21) and ankylosing spondylitis (upadacitinib, tofacitinib) (22,23).
First reports from an expanded access program study on the beneficial effects of JAKi in type I interferonopathies (24) have also been published. Due to their broad blockade of proinflammatory pathways, JAKi may ameliorate autoinflammatory processes and thus lead to clinical remission in otherwise refractory AID cases. The aim of this systematic literature review is to identify and analyze the available evidence on JAKi for the treatment of autoinflammatory diseases.

METHODS
The Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidelines were followed for preparing the manuscript (25).

Protocol and Registration
A study protocol was registered at PROSPERO (CRD42021270369) prior to the systematic search (26).

Data Sources and Searches
The following databases were systematically searched for publications investigating the role of JAKi in AID treatment: MEDLINE via PubMed, EMBASE via Ovid, Cochrane Central Register of Controlled Trials (via Cochrane Library). The search was conducted on 30 June 2021 and updated on 16 October 2021. The results were supplemented by a backwards search of relevant publications (reference screening).
The search strings were built based on two components using the Boolean operator and (AID and JAKi). Within those components, multiple terms were linked by or. For each syndrome, the full and the abbreviated terms were used including at least one synonym for each condition. For MEDLINE both Medical Subject Headings (MeSH) terms and free-text words were used. All keywords were used to search within titles and abstracts of publications.
Details of the complete search strategy for all searched databases can be found in the Supplementary Materials.

Study Selection
Criteria for inclusion were developed using the Patient, Intervention, Comparator, Outcome (PICO) scheme (27). Of interest were following diseases/syndromes:  Defined as Intervention was the usage of JAKi (tofacitinib, upadacitinib, baricitinib, filgotinib or ruxolitinib). As Comparator we accepted any other treatment. For Outcome we analyzed treatment response (see below) and safety (considered were reports on any adverse events).
No restrictions were applied concerning publication date, age, and number of recruited patients. Only studies published in English were included. Considered for inclusion were both retrospective (e.g., case reports, case-series, case-control studies) and prospective studies (e.g., randomized controlled trials, nonrandomized controlled trials, prospective observational studies).
Assessment for eligibility was performed by two independent reviewers (AP and ZB), following inclusion and exclusion criteria ( Table 1). First, only title and abstract were screened. Suitable publications were then assessed in full text. Where there were discrepancies in the evaluation of the eligibility of a publication by the two reviewers, a third reviewer acted as an arbiter (MK).

Data Collection Process and Data Items
Data extraction and management was performed with Microsoft Excel 2016. A standardized data extraction sheet was designed and used for extraction of study characteristics and outcome data, which was carried out by one of the reviewers (ZB). Data was extracted from each publication on: (1) study characteristics; (2) patient characteristics at baseline; (3) patient characteristics after intervention.

Summary Measures, Synthesis
Due to the lack of randomized controlled trials and the heterogeneity of data, a narrative synthesis was carried out. Results were reported based on the Synthesis Without Metaanalysis (SWiM) guideline (28).
Here, the treatment response of each patient was classified as complete, partial or none based on the available data on clinical symptoms and laboratory parameters prior/post intervention. A complete response was defined as resolution of all clinical symptoms and normalization of inflammatory parameters (Erythrocyte Sedimentation Rate, ESR, and/or C-Reactive Protein, CRP); as partial when either clinical symptoms resolved or laboratory markers normalized, and as none when both remained unchanged or worsened.

RESULTS
The first database search identified 582 records of which 70 were removed (duplicate records). The 512 records were screened. Reference screening of included publications additionally identified 4 suitable publications. The updated search (June to October 2021) identified further 80 publications, of which 75 were screened.
Overall, 38 original publications were included for data extraction and analysis. A total of 101 AID patients treated with a JAKi could be identified. Figure 1 provides details on the selection process of included studies.

Evidence on Effectiveness and Safety for Chronic Atypical Neutrophilic Dermatosis With Lipodystrophy and Elevated Temperature (CANDLE) Syndrome
The database search identified four case reports (full text n = 3, conference abstracts n = 1) (29-32) and two articles reporting results of the same compassionate use study (24,33). A total of fourteen patients were treated with a JAKi. Median age of JAKi initiation was 8.5 years (1.5-17 years, reported for 4 patients). Eleven patients received baricitinib (11/13, 84.6%), and three received tofacitinib (3/14, 21.4%). Mean treatment duration was 92.4 months (reported for 13 patients). All but one patient received glucocorticoids (GC) in addition to a JAKi. Data on baseline characteristics, treatment and response are shown in Supplementary Table 1.
One patient discontinued therapy after 67.5 months because of acute kidney injury following a series of infections (pneumocystis jirovecii pneumonia, clostridium difficile, influenza, and rotavirus). Details on AEs are summarized in Table 2.
Data was extracted and analyzed for a total of twentyeight patients. Median age of JAKi initiation was 7.  (2), BK viruria (1) and gastroenteritis (1) all without the need for therapy discontinuation or reports of dose reduction. JAKi dose reduction was required in two cases: after recurring respiratory infections and in one case of BK viremia. Treatment Types of adverse events

Evidence on Effectiveness and Safety for Aicardi Goutières Syndrome (AGS)
The systematic searches identified three case reports on AGS (full text n = 2, letters n = 1) (47-49). One additional letter reports preliminary results of an open-label single center study involving 35 patients with AGS (50). The publication is discussed in the Discussion section since no individual patient data was available.
Here, we report the available data on two pediatric patients and one adult. The median age of JAKi initiation was 11 years (1.5-22 years). The patients were treated for a mean duration of 28.3 months (18-43 months

Evidence on Effectiveness and Safety for Other Type I Interferonopathies
Three case reports (51-53) (conference abstracts n = 1, full text n = 2) and two articles reporting results of the same compassionate use study (24,33) regarding type I interferonopathies were identified.
A total of seven patients were treated. One patient was diagnosed with DNase II deficiency. For the rest either only "other type I interferonopathy" was reported as diagnosis or a novel mutation was described ( Table 1). Median age of JAKi initiation was 4 years (1 month −17 years, reported for 3 patients). Five patients (5/7, 71.4%) received baricitinib, and one patient each received tofacitinib or baricitinib (1/7, 14.3% each). Mean treatment duration was 33.4 months. GC were used in five patients (5/7, 71.4%), one patient received concomitant cyclosporine therapy and one patient received a combination of mepacrine and hydroxychloroquine. A summary of baseline characteristics, treatment and response is shown in Supplementary Table 1. A complete response was achieved by only one patient (1/7, 14.3%) under combination of baricitinib and cyclosporine A (5 mg/kg/d). Four patients (4/7, 57.1%) had a partial response and two (2/7, 28.6%) showed no response to therapy. For three patients (3/5, 60%) GC dose reduction was possible and one (1/5, 20%) successfully tapered GC.

Evidence on Effectiveness and Safety for Adult-Onset Still's Disease (AOSD)
For AOSD three case reports (54)(55)(56) and three case series (57)(58)(59) were identified (conference abstracts n = 2, letters to the editor n = 2, publications in full text n = 2). Data on AEs was available for 24 patients (24/26, 92.3%) and were overall rare: pneumonia (3/24, 12.5%) and menometrorrhagia (1/24, 4.2%); the latter required therapy discontinuation in one patient. One of the patients with bacterial pneumonia died after a 217-day long hospital stay. Otherwise no AEs required hospitalization. Details on AEs are summarized in Table 2.

Evidence on Effectiveness and Safety for Systemic Juvenile Idiopathic Arthritis (SJIA)
Two case reports (60, 61) and one case series (57) (published as conference abstracts n = 1, letters n = 1, in full text n = 1) were identified.

Evidence on Effectiveness and Safety for FMF
Two case series (62,63) and one case report (64) (full text n = 2, letters n = 1) were identified.
A total of six patients with FMF were treated with a JAKi. Median age at JAKi initiation was 35.5 years (16-64 years). All patients were treated with tofacitinib 10 mg/d. Mean treatment duration was 4.5 months (2-12 months). One patient received no supportive treatment (1/6, 16.7%), one (1/6, 16.7%) received a combination of GC, colchicine, and sulfasalazine. Four patients (4/6, 66.7%) were prescribed colchicine. A summary of baseline characteristics, treatment and response is shown in Supplementary Table 1. A complete response was shown by half of the patients (3/6, 50%). The other three patients (3/6, 50%) developed no further flares, but acute phase reactants remained elevated, thus only a partial response was achieved.

Evidence on Effectiveness and Safety for Behçet's Syndrome
Only one publication on BS could be identified (65). Thirteen patients were treated with a JAKi. Median age at JAKi initiation was 42 years (22-73 years). All patients received tofacitinib 10 mg/d. Patients were treated for a mean duration of 10.8 months (5-21 months). All but one patient received concomitant GC therapy (12/13, 92.3%). Tofacitinib was administered as additional therapy to other drugs such as azathioprine, thalidomide, leflunomide, colchicine, salazosulfapyridin. A summary of baseline characteristics, treatment and response is shown in Supplementary Table 1.
According to the criteria used in this systematic review, most patients achieved a partial response (8/13, 61.5%). Five patients showed no response to therapy (5/13, 38.5%), of whom one patient's condition worsened (1/13, 7.7%) and tofacitinib was withdrawn after 9 months of treatment. Two cases of herpes zoster reactivation were observed, both of which led to discontinuation of tofacitinib. No other AEs were reported ( Table 2).

Evidence on Effectiveness and Safety for Other Syndromes
No articles regarding CAPS, TRAPS, PFAPA, PAPA or MKD could be identified.

DISCUSSION
To our knowledge, this is the first systematic review analyzing the safety and effectiveness of JAKi in AID. The overview of the available clinical evidence is based on observational studies such as case reports and case series. Effectiveness was evaluated based on clinical response, defined by the authors of this systematic review as complete, partial or no response depending on (complete) symptom resolution and/or normalization of laboratory parameters for inflammation. Furthermore, AEs were described.

Type I Interferonopathies
Most reports (n = 25) included in this systematic review investigated the use of JAKi for type I interferonopathies.
Interferonopathies represent a group of rare monogenic AID, characterized by a disturbed control of interferon-mediated immune responses, especially of type I interferons. JAKi are potent inhibitors of the JAK-STAT pathway, involved in interferon signaling (66). Thus, the application of JAKi in interferonopathies seems rational. In this analysis, reports on SAVI, CANDLE, AGS and other interferonopathies were included.
For AGS, one publication (50) was identified but individual patient data was not available. This article reported preliminary results of an open-label single center study involving 35 patients with molecularly confirmed AGS. This is to our knowledge the largest AGS cohort treated with JAKi. All patients received baricitinib. The authors report overall improvement of daily diary scores within 1 month of therapy initiation. Neurological function was evaluated based on key developmental milestones, with 20 patients (20/35, 57.1%) meeting new milestones and 12 (12/35, 34.3%) gaining two to seven new skills. In the three cases presented in this systematic review, neurological symptoms were leading in just one case (49) showing improvement under ruxolitinib. Regarding the safety profile of JAKi, in this AGS cohort (50) only one case of BK viremia was described-a relatively frequent event in interferonopathy patients discussed here (10/52, 19.2%).
Notably, amongst interferonopathy patients, JAKi was most efficient for CANDLE patients: complete remission was achieved by six patients (6/14, 42.6%). Additionally, a GC sparing effect in this group is suggested by the available data, since seven patients (7/11, 63.6%) were able to discontinue GC and in four patients (4/11, 36.4%) GC dose reduction was possible.
One hypothesis for the difference in treatment outcome between interferonopathies is that the better treatment response is owed to a higher JAKi dosage. A direct comparison is difficult since mostly pediatric patients were treated and JAKi dosage was reported as mg per kilogram without documenting weight for each individual patient. Mean baricitinib and tofacitinib dosages in CANDLE cases were 6.8 and 5 mg/d, respectively. Mean baricitinib doses for SAVI were 6 mg/d (reported for 6/7, 85.7%) up to a mean 8 mg/d in other interferonopathies. Most SAVI patients were treated with ruxolitinib at a mean dose 10.8 mg/d (reported for 13/18, 72.2% patients treated).
A head-to-head comparison of effectiveness is difficult due to (1) the small number of patients treated (2) the different choice of JAKi and the respective dosage used and (3) partial missing individual data.
Regarding safety it should be mentioned that most infectious AEs in this analysis occurred in patients with type I interferonopathies: seven cases of pneumonia (7/10, 70%), all UTIs (17/17, 100%) and all cases of BK viremia and viruria (10/10 and 4/4, respectively; 100%) ( Table 2). The proportion of patients in this group who experienced any AE is also greater compared to other groups: around 17% of AOSD patients (4/24, 16.7%), even less in BS patients (2/13, 15.4%) and none of the FMF and sJIA patients.
Of overall 59 AEs reported, 52 (88.1%) were due to infections ( Table 2). In general JAKi show a heterogenous risk of infectious complications. For example, a known class effect for JAKi is an elevated risk of herpes zoster (67)(68)(69)(70). In one recent study, serious infections were more frequent with tofacitinib at a dose of 10 mg twice daily compared to TNF inhibition (71), which contradicts some available evidence pointing to a similar risk of serious infections under JAKi compared to other biological disease-modifying antirheumatic drugs (bDMARDs) (72,73). The risk for opportunistic infections (herpes zoster, tuberculosis) under tofacitinib in this study was higher compared to a TNFinhibitor, and even more so when tofacitinib dose was 10 mg compared to 5 mg. In this systematic review two cases of herpes infections occurred under tofacitinib and baricitinib each. One case occurred under ruxolitinib.
A controversially discussed adverse effect of JAKi are thromboembolic events. This risk has been shown to be elevated relative to TNF inhibitors, along with a clinically meaningful risk of serious heart-related AEs, cancers, blood clots and death in older patients with RA (71). However, several (meta-) analyses did not provide evidence supporting an increased risk of thromboembolism with JAKi (67,74,75). In this systematic review, thromboembolic events were not reported.
The discrepancy in the frequency of AEs in the different disease groups here could reflect the inconsistency in reporting of AEs, commented further below. It should be considered that for FMF, sJIA and Behçet's syndrome only a few reports were available for analysis. However, one reason for the higher incidence of infections amongst interferonopathy patients might be due to a dose dependent effect. All FMF and Behçet's syndrome patients, and most AOSD patients received a "standard dose" of tofacitinib (5 mg/d) or baricitinib (4 mg/d). As mentioned above, JAKi doses varied amongst interferonopathy patients. Nevertheless, baricitinib was often administered at doses higher than 4 mg/d-up to a mean 6.8 mg/d in CANDLE patients and a mean 6 mg/d for SAVI patients. Notably, most interferonopathy patients were pediatric patients, suggesting a higher dose pro kilogram body weight.
Another hypothesis is that interferonopathy patients generally have a higher risk for infections. While infections can be considered potential triggers for disease onset or flares (5), a predisposition for infections in interferonopathy patients is currently not proven. However, for other diseases with a prominent interferon signature, such as systemic lupus erythematosus (SLE) and dermatomyositis, a susceptibility for infections has been reported (76,77). Thus, a possible explanation for the elevated incidence of infectious AEs in this subgroup could be an intrinsically dysfunctional immune system leaving patients exposed to an increased risk of infection.

Other AID
Part of the systematic database searches about JAKi for treating monogenic AID were CAPS, TRAPS MKD and FMF. Of those, publications were identified only for FMF. In FMF patients, JAKi resulted in a complete response in half of the patients and a partial response for the rest (3/6, 50% each). Eleven AOSD patients had a complete response and the same number of patients a partial response (11/26, 42.3% each). Two sJIA patients completely responded to JAKi therapy (2/4, 50%) and for the other two (2/4, 50%) a partial response was reported. Amongst BS a partial response was achieved by most (8/13, 61.5%), and five (5/13, 38.5%) showed no response to therapy.
Although JAKi did not lead to complete remission in all AOSD patients, the majority of them were able to taper or withdraw GC-used as supportive treatment in most patients (

Limitations and Considerations for the Future
Due to the rare nature of AID and the relatively recent availability of JAKi, there are currently no RCTs available, and most publications included were case reports or series. Therefore, conducting a risk of bias assessment was not possible. Many authors were contacted to complete missing data, however in some instances despite our best effort complete information could not be obtained. In the analyzed publications a relatively low systematization in conduct and reporting was observed, which in part resulted in limited details on individual patient characteristics at baseline and post JAKi treatment. In order to include as much information as possible on the topic, congress abstracts were also included in the analysis. Abstracts are generally considered to potentially lower the overall evidence level in a systematic review. Therefore, only abstracts providing sufficient clinical data were included in the final analysis (78).
To present the results of this systematic review, a classification based on clinical symptoms and laboratory parameters was performed. Accordingly, treatment response was classified as complete, partial or none. Although this approach has not been validated, it has been previously used by other investigators (57)(58)(59) and serves as base for objectifying and summarizing the available evidence. The body of evidence found did not suffice for quantitative analysis due to its heterogeneity. Instead, an extensive narrative synthesis was conducted.
To this date, no universal criteria for reporting outcomes in AID patients exist. To improve and standardize reporting on treatment strategies in AID we suggest the following type of reporting ( Table 3). clinical symptoms, inflammatory parameters, concomitant diseases, previous therapies; for the use of JAKi-exact dose, as well as information on any supportive treatment, including dosage; for a precise evaluation therapeutic response statements on dynamics of clinical symptoms, as well as inflammatory parameters should be noted. AEs especially infections should be closely monitored. In the publications included in this systematic review reports on AEs were sometimes insufficient-those were not documented in around 15% of cases (83/101, 82.2%). Furthermore, disease (specific) activity scores and response criteria to compare AID studies are urgently needed. For monogenetic inflammasomopathies (FMF, CAPS, TRAPS, MKD) the Auto-Inflammatory Diseases Activity Index (AIDAI) (79) is a validated score but was only reported in one study concerning JAKi use in FMF (63). An EULAR task force is currently preparing specific criteria for AOSD which should be applied for future reporting (80). Regarding type I interferonopathies Frémond, M. et al., 2016 suggested a disease activity score for SAVI patients: the Disease Activity Rating Scale of TMEM173-mutated patients (42). This score for SAVI needs to be validated and scores for the other interferonopathies need to be developed. Overall, given their rare nature, a considerable number of AID patients treated with a JAKi (101) could be identified. The available evidence showed most patients did respond to JAKi therapy. This review was conducted to summarize the available evidence on new therapeutic possibilities for AID patients and to highlight the need for well-designed clinical trials investigating JAKi in AID. Currently, one phase 3 clinical trial investigating baricitinib in CANDLE, SAVI and AGS is being conducted (81). Research is actively underway in the direction of sJIA with two ongoing phase 3 randomized double-blind, placebo-controlled studies on baricitinib and tofacitinib (82,83).

CONCLUSION
This systematic review provides results from observational studies showing first pieces of evidence on treatment effectiveness of JAKi for AID. To validate these results and confirm efficacy and safety of JAKi for specific AID, clinical trials need to be initiated.

DATA AVAILABILITY STATEMENT
The original contributions presented in the study are included in the article/Supplementary Material, further inquiries can be directed to the corresponding author/s.

AUTHOR CONTRIBUTIONS
ZB, GB, and MK: systematic review concept and design. ZB, NR, and MK: systematic review protocol. ZB: database search. ZB and AP: study selection. MK: arbiter during publications screening. ZB: data extraction and synthesis. ZB and MK: manuscript drafting. ZB, MK, NR, GB, and AP: manuscript revision and final review. All authors approved the final version.