Advances in potential treatment options for myeloproliferative neoplasm associated myelofibrosis

Abstract

Introduction:

The Janus kinase (JAK)1/2 inhibitor ruxolitinib provides rapid, sustained and often dramatic benefits to patients with myelofibrosis, inducing spleen shrinkage and ameliorating symptoms, and improves survival. However, the drug has little effect on the underlying bone marrow fibrosis or on mutant allele burden, and clinical resistance eventually develops. Furthermore, ruxolitinib-induced cytopenias can be challenging in everyday practice.

Areas covered:

The developmental therapeutics landscape in MF is discussed. This includes potential partners for ruxolitinib being developed with an aim to improve cytopenias, or to enhance its disease-modifying effects. The development of other JAK inhibitors with efficacy post-ruxolitinib or other unique attributes is being pursued in earnest. Agents with novel mechanisms of action are being studied in patients whose disease responds sub-optimally to, is refractory to or progresses after ruxolitinib.

Expert opinion:

The JAK inhibitors fedratinib, pacritinib and momelotinib are clearly active, and it is expected that one or more of these will become licensed in the future. The activin receptor ligand traps are promising as treatments for anemia. Imetelstat has shown interesting activity post-ruxolitinib, and azactidine may be a useful partner for ruxolitinib in some patients. Appropriately, multiple pre-clinical and clinical leads are being pursued in this difficult therapeutic area.

1. Introduction

Medical management of MPN-associated myelofibrosis, referred to hereafter as myelofibrosis (MF), revolves around the use of the Janus kinase (JAK) 1/2 inhibitor ruxolitinib. Ruxolitinib is a potent anti-inflammatory drug, and broadly suppresses cytokines in patients with MF, leading to robust reductions in splenomegaly and symptom burden.[1] Long term follow-up of the pivotal COMFORT trials has shown that ruxolitinib prolongs the survival of persons with intermediate-2 or high risk MF.[2] Importantly, ruxolitinib is effective regardless of driver mutation status; indeed, no clinical parameter predictive of response to ruxolitinib or that might aid in selecting patients for this therapy has been identified.[3] A variant allele fraction (VAF) of mutant JAK2 >50% may correlate with a greater likelihood of response to ruxolitinib,[4] while the presence of ≥3 myeloid malignancy-associated mutations may predict for lower odds of a spleen response and worse survival.[5] On-target anemia and thrombocytopenia from JAK2 inhibition are important clinical concerns with ruxolitinib, and the drug is not indicated at present in individuals with platelets <50 × 10⁹/L. While anemia due to ruxolitinib is not prognostically adverse,[6] and ruxolitinib may actually overcome the negative prognostic impact of disease-associated anemia,[7] in practice, anemia represents the most common reason for interruption, dose reduction and discontinuation of ruxolitinib. Dose optimization of ruxolitinib is important for the depth of spleen response achieved, which correlates with survival.[8, 9] Ruxolitinib has only modest impacts on bone marrow fibrosis and the allelic burden of mutant JAK2,[10, 11] leading some to call its disease-modifying ability into question.[12, 13] Others have proposed that prolongation of survival by ruxolitinib may result from the drug’s positive effects on appetite, energy, weight, mobility and overall well-being.[14] Despite the well-established benefits of ruxolitinib, the limitations discussed here, along with eventual loss of response to the drug and the poor outcomes following ruxolitinib discontinuation,[15, 16] leave plenty of room for the development of new and innovative therapies for patients with MF.

2. Novel strategies to improve cytopenias

Anemia and thrombocytopenia represent significant challenges in the management of patients with MF and, are worsened by ruxolitinib, particularly early on in therapy. Both anemia (hemoglobin <10 g/dL) and thrombocytopenia (platelets <100 × 10⁹/L, worse if <50 × 10⁹/L) are also adverse prognostic factors for survival.[17–20] Red blood cell (RBC) transfusion need is an additional risk factor.[19] Current management of anemia in MF relies on erythroid stimulating agents (ESAs), danazol and immunomodulatory drugs (IMiDs®), with or without prednisone; response rates range from 20% to 60%.[21] Treatment of thrombocytopenia is particularly difficult; danazol and in particular, low dose thalidomide (50 mg/d) have some efficacy.[22–25]

2.1. Activin receptor ligand traps

Sotatercept and luspatercept belong to a novel class of compounds termed activin receptor ligand traps, which improve anemia by sequestering transforming growth factor beta (TGF-ß) superfamily ligands such as growth and differentiation factor 11 (GDF-11) that are secreted by bone marrow stromal cells and inhibit terminal erythropoiesis via Smad signaling, thus preventing the interaction of these ligands with activin receptors.[26] However, recent preclinical work showing that deletion of Gdf11 does not improve anemia in transgenic mouse models of β-thalassemia has cast doubts over the mechanism of action of this class of agents.[27] Both drugs have displayed significant efficacy against the anemia of lower risk myelodysplastic syndromes.[28, 29] In an investigator-initiated study (n = 45) performed at our institution in anemic subjects with MF, sotatercept led to a 35% response rate when used alone, and a 23% response rate when used in patients on a stable dose of ruxolitinib.[30] The responses included both transfusion independence (TI) and anemia responses as defined by international consensus criteria. The drug, given subcutaneously every 3 weeks, was very well tolerated, and several patients required drug repeated drug holidays for elevation of hemoglobin values above protocol-defined limits. Luspatercept, both alone and in conjunction with ruxolitinib, has been studied in a multi-center trial (NCT03194542) in both transfusion-dependent and –independent anemic subjects with MF; results from this trial are eagerly awaited.

2.2. IMiDs®

Because of the ability of IMiDs®, both as single agents and in combination with prednisone, to improve cytopenias, particularly anemia, in patients with MF (reviewed in ref.[21]), there has been interest in combining these agents with ruxolitinib. Concurrent administration of lenalidomide with ruxolitinib was found to be difficult because of excessive myelosuppression.[31] However, both low dose thalidomide and pomalidomide have been explored in combination with ruxolitinib. The ongoing pomalidomide trial requires patients to be anemic (hemoglobin <10 g/dL) or transfusion-dependent, and platelets have to be >100 × 10⁹/L.[32] The dose of pomalidomide was 0.5 mg/day in cohort 1 (n = 40), together with 10 mg twice daily of ruxolitinib, but dose modification of the latter was allowed. Eighteen percent of evaluable (i.e., treated for 12 cycles) patients had International Working Group for MPN Research and Treatment (IWG-MRT) responses,[33] mostly clinical improvement (CI) in anemia, and an additional 27% derived “clinical benefit”, defined as improvement of MF-associated symptoms and/or doubling of RBC transfusion intervals/≥1 g/dL increase in hemoglobin without transfusions. Cohort 2 (n = 27) of the study allows intra-patient dose escalation of pomalidomide up to 2 mg/day based on safety considerations. In a recent presentation of the data from this trial, of 17 evaluable patients in cohort 2, 2 had an IWG-MRT response (CI) and 8 clinical benefit (improvement in quality of life). The thalidomide trial uses 50 mg/day in all patients, and enrolls both ruxolitinib-naïve patients, who receive ruxolitinib alone for 12 weeks before thalidomide is introduced, and patients on ruxolitinib for at least 3 months with a stable dose for the preceding 4 weeks but not in a complete or partial remission (PR) per IWG-MRT criteria.[34] A minimum baseline platelet count of 50 × 10⁹/L is required. The dose of ruxolitinib cannot be increased until after 6 cycles of combination therapy. Early results from this ongoing phase 2 investigator-initiated trial in 12 response-evaluable patients (i.e., received at least six cycles of combination therapy) were recently presented: 5 patients experienced CI, 4 of whom also had a platelet response. A platelet response was seen in 2 additional patients. Although early, these results strengthen the evidence for low dose thalidomide as a treatment strategy for thrombocytopenic patients with MF, who have no good therapeutic options, and may even allow dose optimization of ruxolitinib over time.

3. Targeting bone marrow fibrosis

Improvements in bone marrow fibrosis have been particularly difficult to achieve in patients with MF; indeed, only 15.8% of patients receiving ruxolitinib on COMFORT-2 had improvements in bone marrow fibrosis grade after 5 years of follow-up.[11] No clinical benefit was observed at 24 weeks in a phase 2 study of simtuzumab, a monoclonal antibody directed against the extracellular matrix enzyme lysyl oxidase-like-2, either alone or in combination with ruxolitinib.[35] The clinical hypothesis is that improvement in bone marrow fibrosis will lead to improvements in cytopenias. Preclinical studies support targeting Gli1⁺ and leptin-receptor-expressing mesenchymal stromal cells (MSCs) that become myofibroblasts in the bone marrow.[36, 37] Other investigators have shown that SLAMF7^high-fibrocytes from patients with MF, particularly those harboring JAK2 V671F, can be targeted by elotuzumab, a monoclonal antibody against SLAMF7 approved for the treatment of multiple myeloma.[38] TGF-ß has been implicated in the pathogenesis of bone marrow fibrosis in MF[39, 40] but efforts to target this with a monoclonal antibody (fresolimumab) were thwarted by cessation of production of the drug by the manufacturer.[41] However, AVID200 is a TGF-ß “trap” that appears promising in preclinical studies.[42]

The discovery that bone marrow fibrosis-driving fibrocytes in primary myelofibrosis (PMF) are clonal (neoplastic) and monocyte-derived, and that their differentiation can be suppressed by pentraxin-2 (serum amyloid P), an endogenous protein that localizes to sites of tissue injury and is involved in repair, ushered in a new way of thinking about bone marrow fibrosis in MPN-associated MF, previously considered a reactive process.[43] That fibrocytes are monocyte-derived and that their differentiation leads to bone marrow fibrosis in MF was later confirmed by other investigators.[44] PRM-151 is a recombinant form of pentraxin-2 that has led to durable responses in splenomegaly, symptoms and blood counts, as well as improved both reticulin and collagen fibrosis in patients with MF receiving the drug either alone or in combination with ruxolitinib.[45] This agent, administered intravenously (IV) once every 4 weeks, has an excellent tolerability profile.

Results from a randomized phase 2 study comparing 3 different doses (10 mg/kg, 3 mg/kg and 0.3 mg/kg) of PRM-151 in 97 anemic and/or thrombocytopenic patients with MF previously treated with (76%) or not candidates for ruxolitinib were recently presented.[46] There were significant numbers of discontinuations in all 3 arms, and only 51 patients (53%) completed the main study (9 cycles), of whom 48 continued on into an open-label extension phase. The primary endpoint of a ≥1 grade decrease in bone marrow fibrosis at any time point was achieved by 27.8% of patients overall, and there were no significant differences between the 3 dosing arms. However, the 10 mg/kg dosing cohort did have worse baseline clinical characteristics overall. Hemoglobin improvement (HI), defined as a ≥50% improvement in RBC transfusion requirements in RBC transfusion-dependent patients or a hemoglobin increase of ≥1 g/dL lasting ≥12 weeks in RBC transfusion-independent patients, occurred in 15 of 69 patients (21.7%). Platelet improvement (PI), defined as a ≥50% reduction in platelet transfusions, platelet counts of ≥25 or ≥50 × 10⁹/L in thrombocytopenic patients not requiring platelet transfusions, or platelet count doubling without transfusions, all lasting ≥12 weeks, occurred in 16 of 45 patients (35.6%). Thirty four percent of patients had a ≥50% reduction in their MPN symptom assessment form (SAF) total symptom score (TSS)[47] at any time on the study, and 31 of 76 evaluable patients (41%) had some degree of spleen volume reduction (SVR).

4. Rational ruxolitinib-based combinations

Given the well-established benefits as well as shortcomings of ruxolitinib discussed earlier, and its status as the only drug specifically approved for MPN-associated MF, there has been enormous interest in developing combination strategies, some empiric and some mechanism-based, involving ruxolitinib. Despite demonstration of synergism in the laboratory between ruxolitinib and histone deacetylase inhibitors (HDACis)[48] and hedgehog pathway (smoothened) antagonists,[49] the results of clinical trials were disappointing,[50, 51] leading to some of them being terminated early.[52, 53]

4.1. PI3K inhibition

Interruption of signaling via the phosphatidylinositol-3-kinase (PI3K) /mammalian target of rapamycin (mTOR) pathway has been shown to synergize in vitro with JAK2 inhibition in MF.[54, 55] However, although interesting clinical activity was reported in a trial of single agent everolimus conducted in the pre-ruxolitinib era,[56] a frontline trial of ruxolitinib in combination with the pan-PI3K inhibitor buparlisib did not suggest clear benefit over what might be expected with ruxolitinib alone, and was halted early.[57] Interest in this area has since shifted to the delta isoform-specific PI3K inhibitors, and an “add-on” approach to ruxolitinib has been adopted.

The PI3Kδ inhibitor parsaclisib is being studied in add-on fashion in patients who have a “sub-optimal” response to ruxolitinib, defined in the protocol as the presence of a palpable spleen extending >10 cm below the left costal margin or 5–10 cm plus active MF symptoms (1 symptom score ≥5 or 2 symptom scores ≥3 each) in patients on ruxolitinib for ≥6 months with a stable dose for ≥8 weeks.[58] Early results from this trial (n = 31) showed median reductions of spleen volume of 10.9% and 8.8% at 12 (27 patients evaluable) and 24 (19 patients evaluable) weeks, respectively. The median reductions in the MPN-SAF TSS among evaluable patients at weeks 12 and 24, respectively, were 14.4% and 35.9%. Of special interest with regards to this class of drugs given the experience in chronic lymphocytic leukemia, grade 3 transaminitis was reported in one patient, grade 1–2 rashes that resolved on study in several patients, and colitis not at all. The PI3Kδ inhibitor umbralisib was also studied in add-on fashion in 23 patients on a stable dose of ruxolitinib for ≥8 weeks; however, the determination of lost, sub-optimal or no response to ruxolitinib was left up to investigator discretion in this trial.[59] Two patients achieved IWG-MRT-defined complete remission (CR) on this study, 11 CI and 8 stable disease (SD). Both complete responders had post-PV/ET MF and grade 1–2 bone marrow fibrosis. Median SVR on study was 13% and the median reduction in the MPN-SAF TSS was 35%. Five patients achieved ≥2 g/dL increases in hemoglobin. Grade 3 colitis and dyspnea were reported in 1 patient each, while grade 3 diarrhea and asymptomatic grade 3 elevations in amylase and lipase occurred in 2 patients each.

4.2. Bromodomain inhibition

Inhibition of bromodomain and extra-terminal (BET) proteins has the potential to transcriptionally down-regulate numerous oncoproteins, including c-Myc, nuclear factor kappa B (NF-κB), cyclin D1, cyclin dependent kinases 4 and 6 (CDK4/6), the anti-apoptotic protein Bcl-xL and PIM1 kinase.[60] Combined JAK/BET inhibition has been shown to markedly suppress cytokine production, reduce leukocytosis, splenomegaly, extramedullary hematopoiesis and mutant allele burden and reverse bone marrow fibrosis in vivo in a MPL W515L⁺ mouse model.[61] Inhibition of NF-κB signaling by BET inhibition was the major driver of synergy. These findings served as the basis of a clinical trial (MANIFEST) of the BET inhibitor CPI-0610, either alone in MF patients refractory to, intolerant of or ineligible for ruxolitinib, or added on to ruxolitinib in those having a suboptimal response or disease progression.[62] The trial has separate cohorts for patients who are and are not transfusion-dependent within both the CPI-0610 monotherapy and add-on arms. For patients who are not transfusion-dependent, a ≥5 cm palpable spleen is required for eligibility. A 4-week washout from prior JAK inhibitor therapy is required in the monotherapy arm, and patients enrolled to the combination arm must have been on ruxolitinib for at least 24 weeks with a stable dose over the preceding ≥8 weeks. Preliminary results from this ongoing trial show a median best SVR of 19.2% (16 evaluable patients), ≥50% improvement in TSS in 6 of 11 patients on treatment for >12 weeks, achievement of RBC TI in 2 patients, some evidence of improvement of bone marrow fibrosis, mean increases in hemoglobin of ≈1.5 g/dL in the combination cohort and of about 2 g/dL in the monotherapy cohort and reduction over time in the levels of pro-inflammatory cytokines (C-reactive protein and interleukins 8 and 18).[63] Grade 3 thrombocytopenia occurred in 4 of 34 patients receiving CPI-0610 in combination with ruxolitinib (11%), one of whom required dose reduction of ruxolitinib, but not with CPI-0610 monotherapy.

4.3. Azacitidine

A more empiric combination is that of ruxolitinib with azacitidine, as the two agents may target distinct clinical aspects of the disease. In an investigator-initiated trial conducted in the upfront setting, azacitidine was introduced after 12 weeks of ruxolitinib monotherapy at a dose of 25 mg/m²/day on days 1–5 of a 4–6 week cycle, which could be increased up to 75 mg/m²/day as tolerated, in an effort to minimize myelosuppression from the combination.[64] In the most recent update from this ongoing trial (n = 54), the objective response rate (ORR) per IWG-MRT 2013 criteria was 72%, with 2 (4%) partial responses (PR) and the rest CI.[65] Forty two patients had a ≥5 cm palpable spleen at baseline; of these patients, 64% and 57% experienced a >50% reduction in palpable spleen length at any time point on the study and at 24 weeks, respectively. CI in symptoms was observed in 54% of patients. Overall, 23% of responses occurred after the addition of azacitidine. Interestingly, 60% of evaluable patients (n = 35) had bone marrow fibrosis (reticulin/collagen/osteosclerosis) responses with a median time to response of 12 months. Four (7.4%) patients discontinued due to hematologic toxicity, and 7 patients (13%) developed grade 3–4 infection. New grade 3/4 anemia, neutropenia and thrombocytopenia occurred in 35%, 20% and 26% of patients, respectively.

4.4. Other combinatorial strategies

A number of other ruxolitinib-based rational combinations based on laboratory evidence of synergism[66–69] are being evaluated in clinical trials, data from which are eagerly awaited. Examples include the add-on (to ruxolitinib) studies of the Bcl-2/-xL inhibitor navitoclax (NCT03222609), the heat shock protein 90 (HSP90) antagonist PU-H71 (NCT03373877), and the PIM kinase inhibitor INCB053914 (NCT02587598). The first-in-class NEDD8-activating enzyme inhibitor pevonedistat is also being studied in combination with ruxolitinib in an investigator-initiated trial: patients must have been on ruxolitinib for at least 3 months on a stable dose for ≥8 weeks and have not achieved a CR by IWG-MRT criteria (NCT03386214). A phase 1 trial (NCT02370706) combining ruxolitinib with both the CDK4/6 inhibitor ribociclib and the PIM kinase inhibitor PIM447 was terminated prematurely due to poor accrual despite in vitro and in vivo data supporting the combination.[70] Yet other concepts, e.g., ruxolitinib plus poly (ADP-ribose) polymerase (PARP) inhibitors[71] or mitogen activated protein kinase kinase (MEK) inhibitors[72] await translation.

5. Newer JAK inhibitors

A number of investigational JAK inhibitors have entered the clinic over the years, but the development of most has been discontinued, generally because of toxicity.[73] The 3 agents farthest along in clinical development are discussed in detail below. NS-018 is a JAK2 inhibitor that may be somewhat selective for JAK2 V617F[74] and appeared promising in a phase 1 trial (n = 48, 23 (48%) previously treated with a JAK inhibitor);[75] however, in the last reported results from the phase 2 portion of this study in 29 JAK inhibitor-exposed patients with MF (26 evaluable), the rate of ≥35% SVR was only 12%, while the rate of reduction of ≥50% in TSS was 35%.[76] Itacitinib is a JAK1 inhibitor that similarly appears substantially less effective for SVR than for symptom improvement, but is less myelosuppressive by virtue of sparing JAK2;[77] this agent is currently being studied either alone or added on to low dose ruxolitinib (<20 mg daily) in patients who initially had a spleen response to ruxolitinib but later lost it or discontinued ruxolitinib due to hematologic toxicity (NCT03144687).

5.1. Fedratinib

The JAK2 inhibitor fedratinib (400 mg daily) was superior to placebo for both ≥35% SVR and ≥50% TSS reduction at 24 weeks (36% versus 1% and 36% versus 7%, respectively; p <0.001 for both comparisons) in a double-blind, randomized, phase 3 trial (JAKARTA),[78] but the development of this agent was subsequently discontinued because of concerns regarding Wernicke’s encephalopathy (WE), putatively due to inhibition of neuronal thiamine uptake by fedratinib,[79, 80] although this is controversial.[81] A recent analysis of eight potential cases of WE that occurred on clinical trials of fedratinib found only one to be a confirmed case of WE.[82] Two other cases were felt to likely have been WE, but recovered despite continuing to receive fedratinib. All potential WE patients had protracted nausea and vomiting that may have contributed to malnutrition and thiamine deficiency. JAKARTA-2 was a single-arm, open-label, phase 2 study of fedratinib, 400 mg/day, in 97 ruxolitinib-exposed (minimum 14 days) patients with MF.[83] No formal definition of ruxolitinib resistance or intolerance was employed in this trial, and a minimum 14-day washout from ruxolitinib was required. Among 83 evaluable patients, the rate of ≥35% SVR at 24 weeks was 55%. Twenty three of 90 evaluable patients (26%) had a symptom response using the MF SAF.[84] This trial was terminated when the fedratinib development program was shut down owing to concerns over WE. Clinical development of fedratinib has now resumed in earnest. In a recent re-analysis of JAKARTA-2 using stringent criteria (Table 1) to define “ruxolitinib failure” (n = 79), the ≥35% SVR rate at 24 weeks was found to be 30% and the symptom response rate relatively unchanged at 27%.[85] “FREEDOM” (NCT03755518) is a phase 3b, single-arm, open-label study of fedratinib, 400 mg daily, in patients with MF who have failed ruxolitinib as defined by the aforementioned criteria. “FREEDOM2” (NCT03952039) is a phase 3 trial comparing fedratinib to best available therapy (BAT) in a similar patient population.

Table 1.

Criteria for “ruxolitinib failure” used in recently presented studies in myelofibrosis.

Trial	Investigational drug	Criteria
JAKARTA2 (re-analysis)*	Fedratinib*	Relapsed: Ruxolitinib treatment for ≥ 3 months with regrowth, defined as < 10% SVR or < 30% decrease in spleen size from baseline, following an initial response
		Refractory: Ruxolitinib treatment for ≥ 3 months with < 10% SVR or < 30% decrease in spleen size from baseline
		Intolerant: Ruxolitinib treatment for ≥28 days complicated by development of RBC transfusion requirement (≥2 units per month for 2 months); or grade ≥3 thrombocytopenia, anemia, hematoma and/or hemorrhage while receiving ruxolitinib
IMBARK™	Imetelstat	Worsening of splenomegaly-related abdominal pain at any time after the start of JAK inhibitor therapy and EITHER: • No reduction in spleen volume or size after 12 weeks of JAK inhibitor therapy, OR • Worsening splenomegaly at any time after the start of JAK inhibitor therapy documented by: ✓ Increase in spleen volume from nadir by ≥25% measured by MRI or CT, or ✓ Increase in spleen size by palpation

^*Same or very similar criteria used in FREEDOM trials of fedratinib (NCT03755518, NCT03952039) and PAC203 trial of pacritinib (NCT03165734).

5.2. Momelotinib

Momelotinib is a JAK1/2 inhibitor with the unique benefit of improving anemia in patients with MF,[86, 87] an effect that may be mediated, at least in part, via inhibition of the type 1 activin A receptor and suppression of hepatic hepcidin production.[88] This agent has recently been resurrected after having been shelved following mixed results in two phase 3, randomized clinical trials (RCTs). Momelotinib was compared head to head against ruxolitinib in 432 JAK inhibitor-naïve patients with MF in the SIMPLIFY-1 trial that had a non-inferiority design.[89] Non-inferiority at 24 weeks was demonstrated for the primary endpoint, ≥35% SVR (26.5% for momelotinib and 29% for ruxolitinib, p = 0.011) but not for the key secondary endpoint of ≥50% reduction in TSS (28.4% for momelotinib and 42.2% for ruxolitinib, p = 0.98). RBC transfusion rate, TI, and transfusion dependence were all improved with momelotinib (all with nominal p ≤ 0.019). The SIMPLIFY-2 trial compared momelotinib to BAT (that could be ruxolitinib and was in 89% of patients) 2:1 in 156 patients with MF who had been on ruxolitinib for at least 28 days and either required RBC transfusions while on ruxolitinib or ruxolitinib dose reduction to <20 mg twice a day because of grade ≥3 thrombocytopenia, anemia, or bleeding.[90] This trial did not meet its primary endpoint of superiority of momelotinib for ≥35% SVR at 24 weeks (7% for momelotinib and 6% for BAT, P = 0.9). Because of this, statistical significance could not be claimed for the secondary endpoints, but 26% of patients receiving momelotinib had a ≥50% reduction in TSS compared to 6% of those receiving BAT (nominal p = 0.0006). Similarly, the analyses of RBC transfusion rates, TI and transfusion dependence all favored momelotinib, like in SIMPLIFY-1. Momelotinib will now be studied in a phase 3 RCT (“MOMENTUM”) versus danazol (2:1 randomization) in 180 patients with MF with the primary endpoint being ≥50% TSS reduction at week 24 and the key secondary endpoint being the RBC TI rate at week 24.[91] Low grade peripheral neuropathy that is usually not reversible is a concern with momelotinib.[92]

5.3. Pacritinib

Pacritinib is a relatively non-myelosuppressive JAK2 inhibitor that has been studied in clinical trials in patients with MF that did not specify a minimum platelet count.[93, 94] In the PERSIST-1 phase 3 RCT conducted in JAK inhibitor-naïve patients with MF (n = 327, randomized 2:1 to pacritinib, 400 mg daily or BAT, which could not include ruxolitinib), pacritinib was superior to BAT for the primary endpoint, ≥35% SVR, at 24 weeks (19% for pacritinib versus 5% for BAT, p = 0.0003).[95] For ≥50% TSS reduction, there was no significant difference between the two groups at 24 weeks when considering the intention to treat (ITT) population, but 36% of pacritinib patients as opposed to 14% of BAT patients achieved this endpoint in the evaluable population (p = 0.029). The differences became statistically significant in the ITT population as well at week 48 (15% for pacritinib versus 0% for BAT, p = 0.0027). The PERSIST-2 phase 3 RCT compared 2 doses of pacritinib, 200 mg twice daily and 400 mg once daily, against BAT in 311 thrombocytopenic (platelets ≤100 × 10⁹/L) patients with MF.[96] Prior JAK inhibitor therapy was permitted (48% of the patients had received prior ruxolitinib), and ruxolitinib could be used as BAT (and was in 45% of patients). Placement of a “full clinical hold” by the US Food and Drug Administration (FDA) on the pacritinib development program that was subsequently lifted[97, 98] affected the analyses of this trial, with the result that only 226 patients (approximately equal numbers in each of the 3 arms) were able to be included in the ITT efficacy population. This trial met one of its co-primary endpoints (≥35% SVR at week 24, 18% for the combined pacritinib arms versus 3% for BAT, p = 0.001) but not the other (≥50% TSS reduction at week 24, 25% for the combined pacritinib arms versus 14% for BAT, p = 0.08). However, twice daily pacritinib was significantly superior to BAT for both ≥35% SVR (22% versus 3%, p = 0.001) and ≥50% TSS reduction (32% versus 14%, p = 0.01) at 24 weeks. Gastrointestinal side effects, particularly diarrhea, are common with pacritinib, especially early on, and have been attributed to its fms-like tyrosine kinase 3 (FLT3) inhibitory activity. Pacritinib was subsequently studied in a dose-ranging phase 2 trial (PAC203) that evaluated 100 mg daily, 100 mg twice daily and 200 mg twice daily in 150 patients who had “failed” ruxolitinib as defined by the criteria in Table 1. This trial (NCT03165734), whose results are still awaited, will be amended into a registration-directed trial (“PACIFICA”) of pacritinib, 200 mg twice daily, versus physician’s choice (including low dose ruxolitinib), in severely thrombocytopenic (<50 × 10⁹/L) patients with MF who have had no more than 3 months of ruxolitinib.[99]

6. Beyond JAK inhibition: new drug classes in myelofibrosis

6.1. Imetelstat

The telomerase inhibitor imetelstat led to CR or PR in 7 of 33 patients (21%) with MF, 48% of whom had received prior JAK inhibitor therapy, in a pilot study, with all 4 patients who achieved a CR having reversal of bone marrow fibrosis.[100] Responses were confined to JAK2 mutated and ASXL1 wild type patients, but did not correlate with baseline telomere length. CRs were enriched among patients with mutations in genes encoding spliceosome components. Toxicity in terms of cytopenias and liver enzyme abnormalities was substantial. These findings led to a multi-center, randomized study (n = 107) evaluating 2 doses of imetelstat, 9.4 mg/kg and 4.7 mg/kg, administered IV every 3 weeks, in patients who had failed JAK inhibitor therapy as defined by the criteria set forth in Table 1.[101] Following an interim analysis conducted after 20 patients in each dosing arm had reached the 12-week response assessment time point, the lower dose arm was closed and the patients permitted to dose-escalate. Enrollment was also suspended to the higher dose arm. At week 24, 10% of patients in the 9.4 mg/kg dosing arm achieved ≥35% SVR and 32% experienced a ≥50% reduction in their TSS. Nineteen of the 107 patients (18%) had improvement in bone marrow fibrosis. Interestingly, after a median follow-up of 27.4 months, the median survival times in the 4.7 and 9.4 mg/kg dosing arms were 19.9 and 29.9 months, respectively.

6.2. Alisertib

Another therapeutic approach involves targeting aurora kinase A, an intervention that has been demonstrated to promote polyploidization and differentiation of driver mutation-carrying megakaryocytes in PMF, shown to be central to MPN pathogenesis,[102] resulting in potent anti-fibrotic and anti-tumor effects in vivo.[103] In a phase 1 study of the aurora kinase A inhibitor alisertib in 24 patients that were intolerant (2 patients) or refractory (13 patients) to ruxolitinib or unlikely to benefit from it (9 patients), the rate of ≥50% reduction in palpable spleen length was 29% (4 of 14 patients with baseline palpable splenomegaly ≥5 cm) and that of ≥50% reduction in the MPN-SAF TSS (7 of 22 evaluable patients) was 32%, both sustained for at least 12 weeks.[104] There was not a consistent effect of alisertib on plasma cytokine levels. Two of 19 evaluable patients (11%) had an anemia response. Of seven evaluable patients, 5 (71%) had a one grade reduction in bone marrow fibrosis, and 6 (86%) showed restoration of GATA1 staining, deficiency of which due to defective ribosomal function has been implicated in the impaired megakaryocyte maturation typical of PMF.[105]

6.3. LCL-161

Preclinical studies suggest that JAK2 V617F leads to both increased levels of tumor necrosis factor alpha (TNFα) in the MF bone marrow microenvironment and confers resistance to TNFα via down-regulation of X-linked inhibitor of apoptosis (XIAP).[106, 107] However, the expression of cellular inhibitor of apoptosis (cIAP) protein is increased in MF CD34⁺ cells.[107] These findings formed the basis of an ongoing investigator-initiated trial of the IAP antagonist LCL-161 in patients with MF.[108] Data on 44 patients, 25 (57%) of whom had previously received a JAK inhibitor, were included in the most recent presentation of results from this study. Notably, the median baseline platelet count was 50 × 10⁹/L, and all but 2 patients were intermediate-2 or high risk by the International Prognostic Scoring System (IPSS).[17] The overall response rate (ORR) was 32% by IWG-MRT criteria; most responses were CI in symptoms or anemia. Responses correlated with on-treatment declines in cIAP1 levels, while high baseline levels and on-treatment increases of XIAP were implicated as a possible resistance mechanism.

6.4. Tagraxofusp

Tagraxofusp is a novel, CD123-targeted fusion protein consisting of human interleukin-3 conjugated to truncated diphtheria toxin that was recently approved for the treatment of patients with blastic plasmacytoid dendritic cell neoplasm (BPDCN).[109] This agent has been studied in JAK inhibitor-refractory or –intolerant patients with MF, including those with baseline platelets <50 × 10⁹/L, in an attempt to target the CD123⁺ plasmacytoid dendritic cells (pDCs) in the bone marrow microenvironment.[110] While the drug was clearly active, spleen response rates using conventional measures (≥50% reduction by palpation or ≥35% SVR by imaging) were not reported. Activity was noted in patients with monocytosis (monocytes share a common precursor with CD123⁺ pDCs), which has been recognized as a poor prognostic feature in PMF and may be indicative on accelerated phase disease.[111, 112]

6.5. IMG-7289

Another novel epigenetic target in MF besides BET proteins is lysine-specific histone demethylase 1 (LSD1, also called KDM1A).[113] The first results of an ongoing multicenter phase 1/2 study of the orally administered LSD1 inhibitor IMG-7289 in 16 patients with MF who were refractory/resistant to, intolerant of or ineligible for standard therapy (including ruxolitinib) were recently presented.[114] Baseline platelets ≥100 × 10⁹/L were required. No dose-limiting toxicity (DLT) occurred. Of 9 efficacy-evaluable patients, none experienced a ≥35% SVR after 12 weeks, although six had some reduction or stabilization of spleen volume, while all patients reported improvement in symptoms, with 5 having a ≥50% reduction in their MPN-SAF TSS. Two patients had ≥1 grade improvement in bone marrow fibrosis.

6.6. Other novel agents

A number of other novel targeted agents have been or are being studied as monotherapy in the post-JAK inhibitor setting. The hedgehog inhibitor glasdegib, approved for the treatment of older patients with acute myeloid leukemia (AML), was disappointing in a phase 1/2 trial,[115] while data are eagerly awaited on the murine double minute 2 (MDM2) inhibitor KRT-232 (NCT03662126). JAK2 V617F- and TGF-ß-induced overexpression of MDM2 make this an attractive therapeutic strategy in TP53-wild type MF,[116, 117] and early clinical results of MDM2 inhibition in PV are promising.[118] The selective inhibitor of nuclear transport (SINE) selinexor, recently approved for heavily pre-treated patients with multiple myeloma, is also being studied in patients with MF who have had an inadequate response to or are intolerant of JAK inhibitor therapy in the ESSENTIAL trial (NCT03627403). This trial, based on selective inhibition by selinexor of viability of and colony formation by MF CD34⁺ cells,[119] uses a very similar definition of inadequate response to JAK inhibitor therapy as used in the parsaclisib “add-on” trial.[58]

7. Conclusions

It is an exciting time in drug development for MPN-associated MF. Although many agents have been tried and failed in the clinic, the success of ruxolitinib has generated immense enthusiasm for research in the field and this, in turn, has uncovered important leads, many of which have emerged as “drugable” targets, as discussed in this review. Indeed, it appears that eight years after the approval of ruxolitinib, we may again be entering an era of new drug approvals for MPN-associated MF.

8. Expert opinion

Although ruxolitinib has transformed the care of patients with MF, much remains to be done. Given that MF is a rare disease (US prevalence 4–6/100,000),[120] it is encouraging to see the enormous interest in and numerous initiatives for drug development in this field, mostly fueled by important advances in basic and translational biology that have enhanced our understanding of disease pathogenesis and progression. For example, recognition that mutant calreticulin acts as a “rogue chaperone” for the thrombopoietin receptor, transporting it to the cell surface and physically binding to its extracellular domain, and is immunogenic has opened up the possibility of using vaccine-based strategies to treat CALR-mutant MF.[121–123]

From a practical standpoint, improved strategies to counter the cytopenias, particularly anemia, caused or worsened by ruxolitinib are an important consideration. While data from the ongoing study of luspatercept in anemic subjects with MF are eagerly awaited, the possibility of this agent receiving regulatory approval for lower risk MDS based on the positive MEDALIST trial,[28] coupled with the available data on sotatercept in MF,[30] may make it an option for off-label use before a formal registration is obtained for MF. Another useful strategy to manage anemia early on in ruxolitinib treatment may be to maintain a dose of 10 mg twice daily for the first 12 weeks before escalating, as shown in the REALISE study.[124] However, there are a significant number of patients whose main clinical problem is anemia rather than splenomegaly or symptoms – this represents an unmet medical need that could be addressed to an extent by the activin receptor ligand traps. Low dose thalidomide appears promising as a strategy to combat thrombocytopenia, and the ongoing ruxolitinib plus thalidomide combination trial is an important one to watch in this regard.[34] The development of pacritinib going forward will focus on patients with platelets <50 × 10⁹/L, an extremely difficult population to treat.[99] If approved, momelotinib, given its effects on anemia, will be a very welcome addition to the JAK inhibitor arsenal, and it will be interesting to see if it is adopted for frontline use, given the long experience with ruxolitinib in this space.

The lack of uniform definitions for inadequate or suboptimal response to ruxolitinib or, for that matter, “ruxolitinib failure”, has created much confusion in the field. It is also unclear what rates of ≥35% SVR or ≥50% TSS reduction it is reasonable to expect in ruxolitinib-exposed patients who are receiving a second agent in add-on fashion or in the second-line setting. Some trials, e.g., the MANIFEST trial of CPI-0610[62, 63] and the trials of pevonedistat (NCT03386214) and umbralisib[59] are designed to be broadly inclusive, largely leaving the determination of an insufficient response to ruxolitinib up to the treating physician, while others, such as the parsaclisib[58] and selinexor (NCT03627403) trials have attempted to formalize the definitions of suboptimal response to or failure of ruxolitinib. The re-analysis of JAKARTA-2 using a fairly rigorous definition of ruxolitinib failure (Table 1) has shown reasonable efficacy of fedratinib in the second line.[85] Similarly, the survival data reported with the 9.4 mg/kg dose of imetelstat are particularly intriguing, given that multiple groups have reported the median survival of patients who discontinue ruxolitinib to be 13–14 months[15, 16, 125] and that this trial also used a very reasonable definition of JAK inhibitor failure for eligibility (Table 1).[101] Clinical resistance to ruxolitinib may be due to the “persistence” phenomenon observed in the laboratory and, until effective second-line options are available, brief withdrawal of ruxolitinib followed by re-challenge, akin to what has been seen in preclinical studies, may benefit some patients.[126, 127]

Improvement of bone marrow fibrosis has now been reported with a number of investigational agents/combinations; these include PRM-151,[46] ruxolitinib plus azacitidine,[65] imetelstat[101] and alisertib.[104] Although we now know that bone marrow fibrosis is not a “reactive” phenomenon[43, 44] and that the grade of bone marrow fibrosis does impact prognosis to an extent,[128, 129] it would be more clinically meaningful if a clear and consistent correlation between reduction in bone marrow fibrosis grade and improvement in cytopenias could be demonstrated.

Finally, despite ruxolitinib’s ability to improve survival of patients with advanced MF, there are no medical interventions at present that can prevent transformation to AML, a catastrophic complication. The few patients who achieve CR may be successfully salvaged by allogeneic hematopoietic cell transplantation, but overall, the prognosis remains dismal.[130] Novel strategies are perhaps most urgently needed for these patients, who have a distinct genetic profile from other patients with secondary AML[131] and were excluded from trials of recently approved agents.[132, 133]

Article highlights.

• Ruxolitinib remains the only approved therapy for patients with MPN-associated MF, although other Janus kinase (JAK) inhibitors, e.g., fedratinib, pacritinib and momelotinib, are being actively developed for this indication.

• Many rational ruxolitinib-based combinations are being pursued in clinical trials; some of these aim to counteract ruxolitinib-induced cytopenias, while many others are based on laboratory evidence of synergism and are being tested as “add on” therapy in patients with “insufficient” responses to ruxolitinib.

• Ruxolitinib “failure”, which has no uniformly accepted definition, nevertheless represents a major unmet need: a number of novel agents are being studied in this setting.