Recent progress of JAK inhibitors for hematological disorders

Abstract JAK inhibitors are important therapeutic options for hematological disorders, especially myeloproliferative neoplasms. Ruxolitinib, the first JAK inhibitor approved for clinical use, improves splenomegaly and ameliorates constitutional symptoms in both myelofibrosis and polycythemia vera patients. Ruxolitinib is also useful for controlling hematocrit levels in polycythemia vera patients who were inadequately controlled by conventional therapies. Furthermore, pretransplantation use of ruxolitinib may improve the outcome of allo-hematopoietic stem cell transplantation in myelofibrosis. In contrast to these clinical merits, evidence of the disease-modifying action of ruxolitinib, i.e., reduction of malignant clones or improvement of bone marrow pathological findings, is limited, and many myelofibrosis patients discontinued ruxolitinib due to adverse events or disease progression. To overcome these limitations of ruxolitinib, several new types of JAK inhibitors have been developed. Among them, fedratinib was proven to provide clinical merits even in patients who were resistant or intolerant to ruxolitinib. Pacritinib and momelotinib have shown merits for myelofibrosis patients with thrombocytopenia or anemia, respectively. In addition to treatment for myeloproliferative neoplasms, recent studies have demonstrated that JAK inhibitors are novel and attractive therapeutic options for corticosteroid-refractory acute as well as chronic graft versus host disease.


Introduction
Aberrant activation of JAK kinase and downstream STAT proteins is one of the most important molecular hallmarks of hematological malignancies, especially in myeloproliferative neoplasms (MPNs) [1]. Driver mutations, JAK2V617F, calreticulin exon 9 mutations and mutations in MPL are central to the pathobiology of MPNs, including primary myelofibrosis (PMF), polycythemia vera (PV) and essential thrombocythemia (ET). In addition, activation mutation of colony stimulating factor 3 receptor (CSF3R) contributes to the development of chronic neutrophilic leukemia (CNL), another form of MPN [2]. All these driver mutations induce constitutive activation of JAK/STAT pathways.
The JAK/STAT pathway has been considered an attractive molecular target for the treatment of hematological malignancies, and several chemical compounds that have potent inhibitory action against the JAK/STAT pathway have been tested in preclinical and clinical settings [1]. To date, ruxolitinib is the only JAK inhibitor approved in Japan for higher risk myelofibrosis (MF) and PV. In contrast, fedratinib and pacritinib were also approved for MF in the United States in 2019 and 2022, respectively [3].
In addition to the pathophysiology of hematological malignancies, JAK/STAT pathways play crucial roles in the immunological actions of both acute and chronic graft vs. host disease (GVHD) in allo-hematopoietic stem cell transplantation. Very recently, phase 3 clinical trials revealed that ruxolitinib was superior to the best available therapy in controlling steroid-refractory acute and chronic GVHD [4]. As a result, the FDA approved ruxolitinib for the treatment of steroid-resistant acute and chronic GVHD.
This review discusses recent progress in JAK inhibitor treatment for both hematological malignancies, especially MPN, and GVHD. them, JAK inhibitors are approved for the treatment of primary myelofibrosis (PMF) and polycythemia vera (PV) in many countries, including Japan. Clinical trials are ongoing for essential thrombocythemia (ET) and chronic neutrophilic leukemia (CNL).

Myelofibrosis
2.1.1. Evidence of ruxolitinib Two randomized control phase 3 studies for highor intermediate-2-risk myelofibrosis (either PMF or post-PV/ET myelofibrosis) patients defined by the Dynamic International Prognostic Scoring System (DIPSS), COMFORT-1 [6] and COMFORT-2 [7] trials demonstrated that ruxolitinib was superior to improve splenomegaly and constitutional symptoms compared to placebo or best available therapy. A recently published open-label phase 3b JUMP study also confirmed that ruxolitinib provided a meaningful reduction in spleen size and constitutional symptoms [8]. Importantly, patients with intermediate-1 risk or with low platelet counts (<100 Â 10 9 /L) who did not meet the inclusion criteria for COMFORT1/ 2 studies were included in the JUMP study [8]. Improvement of spleen volume and constitutional symptoms by ruxolitinib were also confirmed by a study with a Japanese cohort [9]. Although the effects on survival were not the primary endpoint in either the COMFORT-1 or COMFORT-2 studies, long-term follow-up analysis of these studies suggested that ruxolitinib may improve the survival of MF patients [10,11]. Pooled analysis of five-year data of COMFOR-1 and COMFORT-2 study revealed that the risk of death was reduced by 30% in originally assigned to ruxolitinib compared to control [12]. The survival benefit of ruxolitinib in MF patients was also demonstrated by a recently published real-world data analysis [13]. In addition, a very recently published retrospective study with real-world data demonstrated that the survival of older MF patients (>65) was improved after the approval of ruxolitinib, especially in patients treated with ruxolitinib [14]. On the other hand, some investigators questioned the notion that ruxolitinib prolongs the survival of MF patients [15,16]. From the point of view of molecular and pathological responses, the effects of ruxolitinib were limited. In the COMFORT-1 study, the median decrease in the allele burden of JAK2V617F was 21.5% at week 48 [17]. After long-term follow-up, only 6 of 236 patients who received ruxolitinib either originally assigned to the ruxolitinib arm or crossover from the control arm achieved a complete molecular response [17]. In the COMFORT-2 study, approximately one-third of JAK2V617F-harboring patients reached a more than 20% reduction in the allele burden of JAK2V617F [11]. According to the changes in bone fibrosis grading, 15.8% of patients treated with ruxolitinib in the COMFORT-2 study achieved a reduction in fibrosis levels. Improvement of bone marrow fibrosis required long-term treatment with ruxolitinib [11]. The analysis of patients in a phase 1/2 study of ruxolitinib demonstrated that 15% (10/68) of the patients showed a decrease in bone marrow fibrosis at 24 months; in contrast, 36% (9/25) achieved improvement in bone marrow fibrosis at 60 months [18]. The effects of ruxolitinib on MF patients are summarized in Table 1.

Predictors of response to ruxolitinib
Many studies have investigated the clinical factors that determine the response to ruxolitinib, as summarized in Table 2. Analysis of the data from JUMP, a multicenter prospective phase 3b study, indicated that a lower DIPSS score, early treatment with ruxolitinib and a dose of ruxolitinib were critical for spleen response [19]. A study out of Italy reported that a higher DIPSS score, transfusion dependency, lower platelet counts, large spleen size and a lower ruxolitinib dose negatively impacted the spleen response [20]. Regarding  Several studies were also conducted to find a correlation between molecular mutation status and ruxolitinib response. A single-center study indicated that MF patients with a JAK2-V617F allele burden greater than 50% tended to have a spleen response [23].Using the data from a phase 1/2 study of ruxolitinib for myelofibrosis, Patel et al. analyzed the mutation profiles associated with the clinical benefits of ruxolitinib [24]. They found that the number of gene mutations was inversely correlated with spleen response. In addition, patients with more than three mutations had a higher discontinuation ratio of ruxolitinib and inferior survival [24]. Spigel et al. reported that the presence of ASXL1 or EZH2 mutations was associated with a shorter time to treatment failure (defined as the time from the start of JAK inhibitor treatment to treatment discontinuation, progression to the accelerated phase or blast phase, spleen progression or death) in MF patients treated with a JAK inhibitor (ruxolitinib or memolotinib) [25].

Ruxolitinib failure and discontinuation
Although ruxolitinib provides clinical benefits, more than half of patients go on to develop ruxolitinib intolerance or resistance within two to three years [26]. In COMFORT studies, approximately 50% of patients discontinued ruxolitinib within three years and up to 75% at five years [6,7]. Other studies outside clinical trials also reported a high rate of ruxolitinib discontinuation in MF patients [27,28]. The pattern of intolerance or resistance to ruxolitinib is variable, including recurrence of splenomegaly, worsening of anemia or thrombocytopenia and progression to blastic phase or infectious complications. Recently, the Canadian Myeloproliferative Neoplasm Group recommended the use of ruxolitinib failure instead of intolerance or resistance and defined these situations [26]. The outcomes of patients who discontinued ruxolitinib were poor. Newberry et al. reported that the median survival of patients who discontinued ruxolitinib in a phase 1/2 study was 14 months [29]. Low platelet numbers at the start or end of ruxolitinib treatment were associated with shorter survival [29]. Approximately 35% of the ruxolitinib discontinuation patients acquired additional new mutations, such as ASXL1 [29]. The poor outcomes of patients who discontinued ruxolitinib were also confirmed by a multicenter, observational, retrospective study conducted in Europe [30]. At three years, 40.8% of the patients stopped ruxolitinib, and the median survival after ruxolitinib discontinuation was 13.2 months [30].

New generation JAK inhibitors
As discussed above, although ruxolitinib provided clinical merits in MF patients in several studies, the durability of the response was relatively short term. In addition, once patients progressed to ruxolitinib failure, the outcomes were quite poor. Therefore, it is an emergent problem to establish novel strategies to overcome ruxolitinib failure. One approach is the combination of ruxolitinib and other classes of drugs, including PI3 kinase inhibitors, Bcl2/Bcl-XL inhibitors and MDM2 inhibitors [31]. New generation JAK inhibitors, fedratinib, pacritinib and momelotinib, are also in progress (  enrolled in this study. Fedratinib was well tolerated and provided clinically meaningful spleen reduction [33]. For the next step, a phase 2 study was conducted, and 31 high-or intermediate-2-risk MF patients were recruited [34]. In this study, the patients received 300 mg, 400 mg or 500 mg fedratinib once daily. The spleen response rates at 12 weeks were 30%, 50% and 64% in each group. Common adverse events were anemia, thrombocytopenia and gastrointestinal symptoms (nausea, vomiting and diarrhea) [34]. A double-blind, randomized controlled phase 3 study, known as the JAKARTA study, was also performed [35]. In this study, intermediate-2 or high-risk MF patients were randomly assigned to three arms: 400 mg/day of fedratinib, 500 mg/day of fedratinib or placebo. The primary endpoint of this study was spleen response, defined as a more than 35% reduction in spleen volume from baseline at 24 weeks. The primary endpoint was achieved in 36%, 40% and 1% of patients in each group, respectively. Improvement of symptoms was found in 36%, 34% and 6% of patients, respectively. However, four patients who were treated with 500 mg of fedratinib developed Wernicke's encephalopathy. Therefore, the clinical trial for fedratinib was stopped in 2013. After careful evaluation, the clinical hold was lifted in 2017. The updated results of this study were reported in 2021, comparing the 400 mg/day fedratinib group and the placebo group [36]. To investigate whether fedratinib has clinical merits in patients who were resistant to or intolerant to ruxolitinib, an open-label phase 2 study named JAKARTA-2 was conducted [37]. The primary endpoint was spleen response, defined as a > 35% reduction in spleen volume. Initially, 97 patients were enrolled in this study, and 83 patients were assessable. In the initial report of this trial, per-protocol analysis was used, and 46 of 83 patients (55%) achieved spleen response. Recently, the results of this study were updated by employing intentionto-treat analysis [38]. The spleen response was 31% with the ITT method. In addition, the definition of ruxolitinib failure was refined. Among the patients who met the refined criteria, the spleen response was 30%. No Wernicke's encephalopathy was reported in the JAKARTA2 analysis [38]. Improvement of myelofibrosis-related symptoms and health-related quality of life in MF patients were also confirmed in the JAKARTA2 trial [39]. The common adverse events shown in the fedratinib study were gastrointestinal toxicities, i.e., nausea, vomiting and diarrhea. Although Wernicke's encephalopathy cases were not reported in later studies, it is recommended that thiamine levels be monitored before and during treatment with fedratinib [32].

Pacritinib
Pacritinib blocks both JAK2 and FMS-like tyrosine kinase-3 but not JAK1. In a phase 2 study, pacritinib provided a favorable spleen response irrespective of baseline platelet numbers [40]. After this study, two phase 3 randomized trials, PERSIST-1 and PERSIST-2, were performed. In the PERSIST-1 study, intermediate/high-risk MF patients who were naïve to JAK inhibitors were randomly assigned to pacritinib or best available therapy (not including JAK inhibitors) [41]. It should be noted that there were no exclusion criteria for baseline anemia or thrombocytopenia. The primary endpoint was spleen volume reduction of 35% or more from baseline at 24 weeks, which was achieved by 19% of patients in the pacritinib arm versus 5% in the best available therapy group (p ¼ .0003) [41]. Importantly, spleen volume reduction was equally found in low platelet number group patients [41]. In the PERSIST-2 trial, MF patients with low platelet numbers (<100 Â 10 9 /L) were enrolled [42]. Approximately half of the enrolled patients had a history of treatment with ruxolitinib. The patients were randomly assigned to receive pacritinib (either 400 mg once daily or 200 mg twice daily) or the best available therapy, including ruxolitinib [42]. Pacritinib was superior to the best available therapy in spleen volume reduction and symptom improvement [42].

Momelotinib
Momelotinib is a JAK1/JAK2 inhibitor initially introduced as CYT387. A phase 1/2 study was conducted for high/intermediate risk MF patients. Fifty percent of patients achieved a spleen response [43]. Notably, among the 33 patients who were red blood cell transfusion dependent, 70% achieved transfusion independence for at least 12 weeks [43]. After that, two phase 3 randomized trials of momelotinib, SIMPLIFY-1 and SIMPLIFY-2, were conducted [44,45]. The SIMPLIFY-1 study enrolled JAK inhibitor naïve intermediate/high-risk patients, and the patients were randomly assigned to receive either momelotinib or ruxolitinib [44]. Regarding the spleen response, momelotinib was not inferior to ruxolitinib but not when it came to symptom improvement [44]. In contrast, momelotinib improved transfusion rate and transfusion independence [44]. In the SIMPLIFY-2 study, momelotinib was evaluated in intermediate/high-risk MF patients who had a history of ruxolitinib treatment and an inadequate response to ruxolitinib or required blood cell transfusion during ruxolitinib treatment. The patients received momelotinib or the best available therapy, which was usually ruxolitinib. This study indicated that momelotinib was not superior to best available therapy in spleen response [45]. Peripheral neuropathy occurred in approximately 10% of patients treated with momelotinib [44]. Regarding the mechanisms of improvement of anemia, an in vitro study by Asshoff et al. suggested that momelotinib decreases the production of hepcidin, leading to improvement of iron homeostasis and stimulation of erythropoiesis [46].

Polycythemia vera
PV is one of the forms of myeloproliferative neoplasms characterized by increased red cell production. At present, the treatment goal of PV is to prevent thromboembolic complications and ameliorate constitutional symptoms. Patients over 60 years old or with a history of previous thromboembolic complications are considered to have a high risk for thrombosis and should be treated with cytoreductive therapy [47,48]. For cytoreductive therapy for PV, hydroxycarbamide (HU) has been widely used. However, approximately 10 to 15% of PV patients develop refractoriness or intolerance to HU [49]. Two randomized phase 3 studies, RESPONSE and RESPONSE-2, revealed that ruxolitinib was superior to the best available therapy to control hematocrit and other hematological parameters in HU-resistant/intolerant PV patients [50,51]. Ruxolitinib also improved constitutional symptoms and splenomegaly. A five-year follow-up analysis of these studies indicated the long-term usefulness and safety of ruxolitinib [52,53]. The efficacy and safety of ruxolitinib in HU resistance/intolerance were also confirmed in PV patients from Japan [54]. Although ruxolitinib effectively reduces hematocrit levels, it is still controversial whether ruxolitinib reduces thromboembolic risk in PV patients. A recently published retrospective real-world analysis demonstrated that ruxolitinib could reduce the incidence of arterial thrombosis [55]. Conversely, a meta-analysis of randomized control studies did not show significant differences in the risk of thrombotic events in ruxolitinib-treated PV patients compared to control groups [56].

Essential thrombocythemia
In contrast to MF or PV, only a few clinical studies have been performed for essential thrombocythemia (ET). In an open-label phase 2 study of ET patients who were refractory or intolerant to HU, ruxolitinib effectively decreased platelet numbers [57].In addition, ruxolitinib improved ET-related symptoms. including pruritus, bone pain and night sweats [57]. In contrast, a randomized phase 2 study of patients resistant/intolerant to HU demonstrated that ruxolitinib was not superior to the best available therapy in terms of hematological response [58]. The rates of thrombosis, hemorrhage and disease transformation also did not differ between the groups [58].

Chronic neutrophilic leukemia
Chronic neutrophilic leukemia is a rare form of myeloproliferative neoplasms characterized by the marked elevation of mature neutrophils. Mutation of the CSF3R gene is a hallmark of CNL and is incorporated into diagnostic criteria [2]. In an in vitro analysis, ruxolitinib effectively suppressed the proliferation of primary cells from CNL patients with CSF3R mutations [2]. To date, only a few case reports have indicated clinical and molecular responses to ruxolitinib in CNL patients [59,60]. Recently, a phase 2 study of ruxolitinib for CNL was reported. According to this report, one-third of CNL patients showed a clinical response [61].

Pretransplantation use of ruxolitinib for myelofibrosis patients
At present, allogenic hematopoietic stem cell transplantation (allo-HSCT) is the only curative therapeutic option for MF; however, the high nonrelapse mortality rate associated with allo-HSCT is a concern. The presence of splenomegaly negatively affects outcomes after allo-HSCT in MF patients [62], and constitutional symptoms are one of the major risk factors for nonrelapse mortality [63]. Therefore, it is reasonable to hypothesize that the use of ruxolitinib in pretransplantation settings may ameliorate both splenomegaly and constitutional symptoms, leading to improved transplantation outcomes. In contrast, there have been several concerns about using ruxolitinib before allo-HSCT, including increased risk of infection, negative effects on the engraftment of stem cells or induced discontinuation syndrome. However, retrospective studies revealed that the response to ruxolitinib prior to HSCT is an important predictive factor for favorable survival [64,65]. Prior use of ruxolitinib did not increase the ratio of graft failure or serious infections [64,65]. The impact of ruxolitinib before allo-HSCT in myelofibrosis has also been investigated in several prospective settings [66][67][68]. Although some cases developed ruxolitinib discontinuation syndrome [68], these studies suggested that pretransplantation use of ruxolitinib was well tolerated. The schedule of ruxolitinib tapering before transplantation has not been established. In 2015, the European Leukemia Net (ELN) and European Blood and Marrow Transplantation Group (EBMT) recommended that the tapering of ruxolitinib should be started five to seven days prior to conditioning and stopped the day before conditioning [69]. In a report from Gupta et al., the researchers tapered ruxolitinib over four days and stopped one to twodays before initiation of conditioning therapy for transplantation [66]. No ruxolitinib discontinuation syndrome was documented in this study [66]. In contrast, a retrospective analysis indicated that approximately 15% (10/66) of MF patients who received ruxolitinib prior to transplantation developed ruxolitinib discontinuation syndrome [64]. In this study, the schedule of ruxolitinib tapering was variable. Further studies are required to investigate whether the pretransplantation use of ruxolitinib can improve the outcome of transplantation. It is also important to establish an ideal schedule for tapering of ruxolitinib prior to transplantation. To improve the outcomes of GVHD, novel therapeutic approaches, including JAK inhibitors, have been investigated. Spoerl et al. reported that inhibition of JAK1/2 using ruxolitinib reduces GVHD in a murine model [70]. Based on these preclinical findings, they also treated six patients with steroidrefractory GVHD and found improvement in GVHD [70]. After this report, the role of ruxolitinib in GVHD treatment was evaluated in several clinical studies. In the open-label phase 2 study, ruxolitinib was administered in patients with grade II to IV steroid-refractory acute GVHD [71]. Seventy-one patients were enrolled in this study. On days 28 and 39 (54.9%) patients showed a response, and the median duration of response was 345 days [71]. The efficacy of ruxolitinib in steroid-refractory acute GVHD was also evaluated in a randomized phase 3 trial [72]. In this study, the patients were randomly assigned to ruxolitinib or control therapy. The overall response rate at day 28 was significantly higher in the ruxolitinib group than in the control group (62% vs. 39%, odds ratio 2.64). In phase 3 randomized studies on the treatment of glucocorticoid refractory chronic GVHD, ruxolitinib also demonstrated a greater response than control therapy [73]. The most common adverse events associated with ruxolitinib were anemia and thrombocytopenia. In addition to ruxolitinib, there are current clinical studies on the effects of the JAK1 inhibitor itacitinib and the JAK1/JAK2 inhibitor baricitinib on GVHD [74].

Notable nonhematological adverse events associated with JAK inhibitors
The JAK/STAT pathway plays important roles in the regulation of the immune system. Indeed, a wide variety of cytokines that control immune function activate the JAK and STAT families. Thus, there is the possibility that inhibitors for JAK may increase the risk of secondary malignancies or opportunistic infections. In addition, several case series indicated that respiratory distress syndrome is associated with rapid discontinuation of ruxolitinib. Some types of JAK inhibitors have neurological or cardiac toxicities. This section discusses JAK inhibitors and associated notable adverse events outside hematological toxicity.

Secondary cancer
It is still controversial whether JAK inhibitors increase the risk of secondary malignancies in MPN patients. In 2018, Porpaczy et al. reported the possible association between the use of JAK inhibitors and the development of aggressive lymphoma in MF patients [75]. Furthermore, using a mouse model, they demonstrated that inhibition of STAT1 activity by ruxolitinib might contribute to the development of lymphoma [75]. However, later large-scale studies did not support this notion [76,77]. Nonmelanoma skin cancer (NMSC) is another malignancy found in patients during treatment with ruxolitinib [78]. It should be noted that NMSC associated with ruxolitinib could show an uncommon aggressive phenotype [79][80][81]. The association between ruxolitinib use and the development of NMSC was confirmed in a large-scale retrospective case-control study on both polycythemia vera and MF [82,83]. The risk of second primary malignancies in MF patients treated with ruxolitinib was also analyzed as a post authorization safety study [84]. This study did not show any significant increased risk of secondary cancer, including lymphoma and NMSC [84].

Opportunistic infections
JAK inhibitors have multiple immunosuppressive actions [85]. Ruxolitinib blocks both the differentiation and function of dendritic cells. The function of T cells and NK cells is also abrogated by ruxolitinib. Indeed, clinical trials with ruxolitinib reported an increased risk of infections. In addition, a variety of opportunistic infection cases associated with ruxolitinib have been reported., including cryptococcosis [86], toxoplasmosis [87], pneumocystis pneumonia [88], cytomegalovirus retinitis [89] and disseminated molluscum contagiosum [90]. Among these opportunistic infections, reactivation of the varicella zoster virus (VZV), hepatitis B virus (HBV) and tuberculosis are especially important for clinicians [91].
In the RESPONSE trial, approximately 6% of patients treated with ruxolitinib suffered from reactivation of VZV [50]. The association between ruxolitinib use and VZV reactivation was also confirmed by a systematic review and meta-analysis of published data [92]. In addition, some patients developed a severe form of VZV infection, i.e., meningoencephalitis [93]. However, primary VZV prophylaxis for patients who will receive ruxolitinib is still a matter of debate [91]. A recently published position paper by the European Conference of Infections in Leukemia (ECIL) recommended careful monitoring for infection but not primary prophylaxis [94]. In contrast, guidelines from the German Society of Hematology and Medical Oncology recommend prophylaxis with acyclovir to prevent herpes zoster [95]. Hepatitis B virus reactivation is another important problem during treatment with ruxolitinib. Several case reports demonstrated elevation of HBV-DNA after treatment with ruxolitinib both in polycythemia vera [96] and primary myelofibrosis [97,98]. ECIL recommended screening of HBV infection, including screening for the Hubs-antigen, anti-HBs antibody and the anti-HBc antibody, before the start of ruxolitinib. In addition, if HBc-Ab is positive, HBV-DNA should be addressed. Prophylactic entecavir should be used in HBV-seropositive patients or in HBV-DNA positive patients [94]. Increases in mycobacterial infections, both mycobacterium tuberculosis and atypical mycobacterial infections, have also been reported in patients with MF or PV treated with ruxolitinib [99]. It should be noted that some patients developed life-threatening disseminated tuberculosis during treatment with ruxolitinib [100,101]. To avoid fatal mycobacterium infection, the medical history of potential tuberculosis should be checked, and if necessary, an IFN-g release assay should be conducted [94].

Ruxolitinib discontinuation syndrome and tumor lysis syndrome
Tefferi and Pardanani reported five cases who developed serious respiratory failure after discontinuation of ruxolitinib [102]. After this report, several groups reported the same sequelae, known as ruxolitinib discontinuation syndrome [103,104].Although the precise mechanisms of the syndrome are not yet fully understood, the rebound of inflammatory cytokines may play important functions. To prevent ruxolitinib discontinuation syndrome, slow tapering of ruxolitinib with concomitant use of corticosteroids is recommended [105]. Although the incident is rare, ruxolitinib-associated tumor lysis syndrome has also been reported [103,106].

Neurological toxicities
Wernicke's encephalopathy cases were reported in an early clinical trial of fedratinib in patients who received doses of 500 mg or higher [35]. In contrast, in the JAKARTA2 study, in which the dose of fedratinib was 400 mg, no Wernicke's encephalopathy case was reported [38].However, it is strongly recommended that thiamine levels should be evaluated before the start of fedratinib and carefully monitored during the treatment [32]. It should also be noted that approximately 10% of patients who were treated with momelotinib developed peripheral neuropathy [107]. At present, no risk factors have been established for the prediction of peripheral neuropathy associated with momelotinib [107].

Cardiac toxicities
Safety data of PERSIST-1 and PERSIST-2 suggested the potential cardiac toxicities of pacritinib, including prolongation of QTc time, atrial fibrillation or cardiac failure [41,42]. Therefore, electrocardiogram is recommended, with monitoring when necessary [108].

Summary and perspective
JAK inhibitors, especially ruxolitinib, provide several clinical merits, including reduction of spleen volume and mitigation of constitutional symptoms leading to improvement of the quality of life of patients with MF. In addition, some studies suggested that JAK inhibitors might improve the survival of MF patients. However, to date, there is no concrete evidence that JAK inhibitors can modify the nature of MF. Only a few patients achieved complete eradication of malignant clones after treatment with ruxolitinib. Pathological improvement of bone marrow fibrosis grade is also modest in ruxolitinib-treated patients. More than half of the MF patients who initially responded to ruxolitinib had discontinued the treatment due to loss of response, development to blastic phase or hematological/nonhematological toxicities, comprehensively expressed as ruxolitinib failure. In addition, the prognosis of the patients who discontinued ruxolitinib was usually poor. Therefore, the development of new strategies to overcome ruxolitinib failure is an emergent clinical problem. One approach is the development of other types of JAK inhibitors. The JAKARTA study revealed that fedratinib showed spleen volume reduction and improvement of constitutional symptoms in MF patients who were resistant to or intolerant to ruxolitinib. Pacritinib could be used for MF patients with thrombocytopenia who are not suitable for ruxolitinib treatment. For patients with severe anemia, momelotinib could be an alternative approach. Combination therapy with JAK inhibitors and other agents is another approach for overcoming ruxolitinib failure. Apart from MF, ruxolitinib is approved for the treatment of PV, and clinical trials for ET and CNL are ongoing.
Another important area of JAK inhibitor use in hematological disorders is the treatment of GVHD. Ruxolitinib is superior for controlling acute as well as chronic GVHD compared to conventional treatment. Studies with itacitinib or baricitinib are also in progress. The therapeutic role of JAK inhibitors in hematological disorders will be continuously expanded.

Disclosure statement
Personal fees received by the author as honoraria from Novartis.