Reviewing the safety of JAK inhibitors for the treatment of myelofibrosis

Over the past decade, seminal studies have led to the identification of specific driver mutations in the pathogenesis of myelofibrosis (MF) and related myeloproliferative neoplasms (MPN), including mutations in JAK2 V617F and JAK2 exon 12. These mutations are linked to dysregulation of the Janus kinase/signal transducer and activator of transcription (JAK/STAT) signaling pathway—the key mediator of more than 50 cytokine receptors involved in cell growth and survival, hematopoietic cell division and differentiation, and pro- and anti-inflammatory responses.

A new subclass of tyrosine kinase inhibitors was borne out of the discovery that MPN pathogenesis is dependent on dysregulation of the JAK/STAT pathway. JAK inhibitors (JAKi) are orally bioavailable small molecules that can be classified into types I–IV. Currently, all JAKi are type I inhibitors—targeting the ATP-binding pocket of the JAK protein—and can be active against both wild-type and mutant JAK, as well as other JAK family members and some non-JAK kinases, depending on the drug. This low selectivity results in adverse events (AEs) associated with both on-target and off-target activity, potentially leading to dose reduction and/or treatment discontinuation.

Giacomo Coltroa and MPN Hub Steering Committee member, Alessandro M. Vannucchi, recently reviewed the safety of JAKi and published their overview in Expert Opinion on Drug Safety.¹ The MPN Hub is happy to provide a summary.

Approved JAK kinase inhibitors¹

Ruxolitinib

The first medication approved for the treatment of MF was ruxolitinib, a selective inhibitor of JAK1, JAK2, and to a lesser extent, TYK2 and JAK3. Ruxolitinib was approved based on efficacy and safety data from the COMFORT-I and COMFORT-II trials. Results of these trials, as well as preliminary results from the phase IIIb single-arm JUMP trial, demonstrated marked and durable clinical benefits, including spleen volume reduction (SVR) and reduction of disease-related symptoms.

The main dose-limiting AEs in ruxolitinib-treated patients were anemia and thrombocytopenia, as shown in Table 1. These AEs were anticipated, as all three main hematopoietic cytokine receptors, Epo-R, Tpo-R/MPL, and GCSF-R/CSF3R, signal through JAK2.

Table 1.  Grade 3/4 hematological AEs and key non-hematological AEs reported in JAKi pivotal trials*

Anemia and thrombocytopenia were effectively managed with supportive care and/or dose titration, and treatment discontinuation was rare (1.0–2.0% and 3.4–3.7% of ruxolitinib-treated patients in COMFORT-II and JUMP, respectively). Notably, pooled analysis of the COMFORT trials suggested that on-treatment reductions in hemoglobin did not undermine the treatment-related survival benefit and that ruxolitinib may dilute the negative prognostic effect of MF-related anemia.

Across these trials, rates of non-hematological AEs were low, including rates of infection, though a recent meta-analysis showed a statistically significant increased risk of herpes zoster infection in ruxolitinib-treated patients compared to control groups.

Fedratinib

Fedratinib is a JAK2-selective inhibitor that has inferior immunosuppressive activity and exhibits greater inhibition of off-target tyrosine kinases compared to other JAKi. The U.S. Food and Drug Administration (FDA) recently approved fedratinib for the treatment of intermediate-2/high-risk primary or secondary MF based on favorable results from JAKARTA, a randomized, placebo-controlled phase III three-arm study of ruxolitinib-naïve patients, and JAKARTA-2, a single-arm, phase II study of patients refractory to or intolerant of ruxolitinib.

Anemia and thrombocytopenia were the most common dose-dependent hematological AEs in these studies. New-onset or worsening Grade 3/4 anemia occurred in 52% (42% in the 400 mg arm and 60% in the 500 mg arm) and 38% of fedratinib-treated patients in JAKARTA and JAKARTA-2, respectively. While new-onset or worsening Grade 3/4 thrombocytopenia was less frequent than anemia (22% across the JAKARTA studies), it was the leading cause of AE-related treatment discontinuation.

Nausea, vomiting, and diarrhea were the most common non-hematological AEs, and urinary tract infection was the most common infection. Of note, four cases of Wernicke’s encephalopathy were reported in the 500 mg arm of the JAKARTA trial and, despite permanent discontinuation of fedratinib and treatment with intravenous thiamine, cognitive deficits persisted in all patients, resulting in a clinical hold on drug development. The hold was lifted when no clinical evidence was found to support a causative role of fedratinib in thiamine deficiency or Wernicke’s encephalopathy, though fedratinib carries a boxed warning regarding low thiamine levels, and risk mitigation strategies will be evaluated in an ongoing phase III trial.

JAKi in advanced clinical development¹

Momelotinib

Momelotinib is a highly selective inhibitor of JAK1, TYK2, and JAK2, with limited off-target inhibition. Momelotinib was investigated in two phase III trials: SIMPLIFY-1 compared momelotinib 200 mg once daily to ruxolitinib in JAKi-naïve patients, and SIMPLIFY-2 randomized MF patients previously treated with ruxolitinib to momelotinib 200 mg once daily or best available therapy (BAT). In SIMPLIFY-1, momelotinib achieved non-superiority for SVR but not for symptoms response; in SIMPLIFY-2, momelotinib failed to demonstrate superiority for SVR.

In SIMPLIFY-1 and SIMPLIFY-2, Grade 3/4 anemia occurred in 6% and 14% of momelotinib-treated patients, respectively, compared to 23.1% of ruxolitinib-treated patients and 14% of BAT-treated patients. Both trials showed higher rates of anemia improvement associated with momelotinib: 67% and 43% of patients achieved transfusion independence at 24 weeks compared to 50% and 21% of patients in the control arms of SIMPLIFY-1 and SIMPLIFY-2. In both trials, Grade 3/4 thrombocytopenia occurred in 7% of momelotinib-treated patients. First-dose AEs, including dizziness, hypotension, flushing, nausea, and headache, were reported in 7% of patients in SIMPLIFY-1; other non-hematological AEs from both SIMPLIFY trials are reported in Table 1.

Pacritinib

Pacritinib was designed to target JAK2, though it has been shown to target a range of tyrosine and non-tyrosine kinases and has a potent effect on the proliferation of kinase-driven myeloid and lymphoid lines. It was evaluated in two global, phase III randomized controlled trials: PERSIST-1 (400 mg once daily vs non-ruxolitinib BAT in JAK-naïve patients) and PERSIST-2 (400 mg once daily, 200 mg twice daily, or BAT in patients with platelet count ≤100 x 10⁹/L). Both studies found pacritinib to be effective in SVR, and results suggested that the 200 mg twice daily dose may be more effective than a single daily 400 mg dose.

Two phase I dose-escalation studies of pacritinib were conducted and showed that dose-limiting toxicities were mostly gastrointestinal events and were reversible at lower doses. Similar results were seen in a pooled safety analysis, and there were no clinically significant declines in mean hemoglobin or platelet count from baseline. No dose reductions were reported due to thrombocytopenia in patients with baseline platelet count ≤100 × 10⁹/L. Diarrhea was the most common AE in the PERSIST trials, and new-onset or worsening Grade 3/4 anemia and thrombocytopenia were reported in 17–24% and 12–31% of pacritinib-treated patients.

The FDA placed a clinical hold on pacritinib development in February 2016 due to concern over life-threatening cardiovascular and hemorrhagic events in PERSIST-1, though a re-analysis of the data in both PERSIST studies showed that bleeding events occurred at a similar rate in the pacritinib and BAT arms, and cardiovascular events also occurred at similar rates in all arms. The hold was removed in January 2017, and a dose-finding study has been designed to identify the lowest effective dose of pacritinib.

Special AEs associated with JAKi therapy

Reports of lymphoid malignancies occurring at an increased rate in patients treated with JAKi, specifically ruxolitinib, led investigators to surmise that JAKi-related immunosuppression contributed to the development of lymphoma, particularly because the B-cell lymphomas that developed in these patients resembled those occurring in immunocompromised patients. Further studies, however, have failed to demonstrate a significant increase of aggressive lymphomas in JAKi-treated patients. Further research into this potential association is needed.

JAKi in early clinical development: Ilginatinib and itacitinib¹

Ilginatinib is a selective JAK2 inhibitor. Promising preliminary results from a phase I/II study show the drug to be effective in SVR and constitutional symptom reduction. Common hematological AEs included anemia and thrombocytopenia: 6% and 17% of patients, respectively, reported Grade 3/4 events. Common non-hematologic AEs were mainly neurological and Grade 1/2.

Itacitinib is a potent and selective inhibitor of JAK1 that, despite its low affinity for JAK2, has reported rates of symptoms response similar to other JAKi, including modest SVR. New-onset or worsening Grade 3/4 anemia, thrombocytopenia, and neutropenia were reported in 33%, 24%, and 4% of patients, respectively, and non-hematological AEs were generally mild or moderate.

JAKi with halted clinical development: Lestaurtinib, gandotinib, AZD1480, and XL019¹

Lestaurtinib is a multi-kinase inhibitor that was initially developed as an FLT3 inhibitor for the treatment of patients with acute myeloid leukemia; its potency as a JAK2 inhibitor led to a phase II trial investigating its use in treating patients with JAK2 V617F-mutated MF. Results were modest, and Grade 3/4 anemia and thrombocytopenia were reported in 14% and 36% of patients, respectively. The most common non-hematological AEs were gastrointestinal.

Gandotinib is a potent JAK2 inhibitor that has shown a dose-dependent selectivity for JAK2 V617F. Results from a single-arm phase II study of patients with MPN demonstrated an overall response rate of 27% among patients with MF, suggesting that blocking mutant JAK2 isn’t as effective as blocking both wild-type and mutant forms. Grade 3/4 hematological AEs included anemia (12%) and thrombocytopenia (2%).

AZD1480 is a potent and selective inhibitor of both JAK1 and JAK2; the drug was administered to 35 patients with MF in a phase I study that reported clinical efficacy in preliminary results, though the study was terminated due to neurological AEs.

XL019 is another selective inhibitor of JAK2 that elicited neurotoxicity in all 30 patients with MF treated with the drug at all dose levels and demonstrated low response rates, halting further development.

Combinations¹

There are currently a number of molecules being investigated in combination with JAKi as a way to improve the efficacy and safety of MF treatment. Those that have shown benefit in combination with ruxolitinib are listed in Table 2.

Table 2. Molecules that have demonstrated benefit in combination with ruxolitinib*

Conclusion

JAKi have revolutionized the treatment of MF, though there are important limitations to treatment with this class of drugs. While most patients experience durable SVR and symptom response, the curative potential of JAK inhibition is limited, as these drugs do not induce meaningful molecular responses. The off-target effects of current JAKi pose another challenge regarding the management of hematological and non-hematological AEs, as well as the other rare but potentially fatal drug-specific AEs discussed here. The management of treatment with JAKi requires an expert knowledge of these risks and the ability to manage dose modifications to limit serious AEs without compromising the benefits of treatment.