Abstract
Primary myelofibrosis (PMF) is a disease characterized by bone marrow fibrosis, extramedullary hematopoiesis, risk of transformation to acute myeloid leukemia, and a substantial symptom burden with diminished quality of life. Allogeneic hematopoietic cell transplantation (HCT) is the only curative option; however, disease relapse and graft versus host disease (GVHD) are significant barriers to long-term survival. The discovery of the JAK2 V617F mutation, and subsequent development of JAK inhibitors, resulted in improved survival and significant improvements in spleen volumes and symptom scores. Though the effect of JAK inhibition on transplant outcome is poorly understood, using JAK inhibition to achieve maximal response prior to HCT is standard practice at major centers. After allogeneic HCT, a significant proportion of patients with steroid-refractory GVHD have clinical responses to JAK inhibition. Targeting this pathway is a key component in the management of patients with PMF before and after allogeneic HCT.
Introduction
Primary myelofibrosis (PMF) is an uncommon disease that arises from clonal hematopoiesis and a relative excess of cytokines (i.e. basic fibroblast, platelet-derived and vascular endothelial growth factors) that promote the constitutional symptoms, fibrosis and angiogenesis frequently seen in the disease.1–3 Clinically, these patients manifest with peripheral blood cytopenias, splenomegaly secondary to extramedullary hematopoiesis, and the transformation to acute myeloid leukemia (AML). 4 PMF has an age- and sex-adjusted incidence of 1–1.46 per 100,000 persons annually, and median age at diagnosis of 67 years (range 43–84).5–8 PMF is a heterogeneous disease with a complex pathobiology and often aggressive disease course with a median overall survival (OS) of approximately 5 years, ranging from 1 to 13 years. 4 Although generally regarded as an acquired disorder, kindred with myeloproliferative neoplasms (MPNs) have been widely reported, suggesting a genetic predisposition for the disease may exist.9,10
Allogeneic hematopoietic cell transplantation (HCT) is the only curative therapy for patients with PMF. In recent years, post-HCT outcomes have improved as a result of better pre-HCT disease control, advances in supportive care, safer conditioning regimens, and enhanced long-term care. 11 Despite these improvements, relapse of the underlying disease and graft-versus-host disease (GVHD) are important barriers to long-term survival. Across all diseases, relapse leads to the death of 37–48% of patients and GVHD is the cause of death for 18–20% patients after allogeneic HCT. A prospective study of 103 patients with PMF, post-essential thrombocythemia myelofibrosis (PET-MF) and post-polycythemia vera-myelofibrosis (PPV-MF) that underwent allogeneic HCT demonstrated a 5-year event-free survival (EFS) of 51% and OS of 67%. 12 In PMF, the 3-year probability of survival with HCT is 38–58%; however, patients with high-risk features (hemoglobin ⩽10g/dl, grade 3 marrow fibrosis and/or >1% peripheral blasts) have an inferior anticipated 3-year survival of less than 20%.13–15 Thus, reducing these risk factors pre-HCT is an important management strategy.
The discovery of the Janus kinase/signal transducers and activators of transcription (JAK-STAT) signaling pathway, and the development of JAK inhibitors, are important advances in the care of patients with PMF and other MPNs. JAK inhibitors improve the symptom burden and OS of patients with PMF independent of allogeneic HCT.16–18 In the pre-transplant setting, the use of these agents to achieve the best possible response prior to the start of conditioning is standard practice at many transplant programs; however, there are limited data to support this practice. After HCT, JAK inhibition has emerged as a new tool in the management of patients with acute and chronic GVHD, with response rates that exceed 80%. 19 Further, a small but growing body of data suggests JAK inhibitors may also serve as steroid-sparing agents. 20 Both before and after HCT, JAK inhibition represents an important modality for improving the outcomes of patients with PMF.
Driver mutations in myeloproliferative neoplasms
The JAK-STAT signaling pathway facilitates a wide array of cytokines and growth factors in animals and humans that drive proliferation, differentiation and other cellular functions. In mammals, this pathway consists of four JAK proteins (JAK1, JAK2, JAK3 and Tyk2) and seven STAT proteins (STAT1, STAT2, STAT3, STAT4, STAT5a, STAT5b, STAT6) which drive the signaling of >50 cytokines and growth factors.21,22 JAK-STAT signaling is initiated when cytokines and the corresponding transmembrane receptors interact, leading to JAK activation, cytoplasmic tail phosphorylation, dimerization of STAT, and subsequent gene transcription. 23 In murine models, constitutive JAK-STAT phosphorylation drives cytokine hypersensitivity, inflammatory disease, peripheral blood abnormalities, and myeloproliferation, whereas JAK2 knockouts die during embryogenesis due to a failure to develop meaningful hematopoiesis.24–27
In 2005, the single base-pair substitution of phenylalanine for valine at residue 617 (V617F) within JAK2 was linked with clonal expansion in myeloproliferative neoplasms, and most dramatically, polycythemia vera (PV).28–31 This gain-of-function mutation enhances JAK2 kinase activity and increases erythropoietin-induced cell signaling and the sensitivity of progenitor cells to growth factors and cytokines, all of which represent key biologic features of MPNs. 28 The incidence of the JAK2V617F mutation is highest in patients with PV (~90%), and occurs less frequently in essential thrombocythemia (ET) (~50%) or PMF (~50%).29–31 Although JAK2V617F is the most common mutation, alternate driver mutations exist in the PMF. Gain-of-function mutations in exon 12 of JAK2 are observed in 4–13.7% of patients with MPNs, are associated with a PV phenotype, and tend to be younger at diagnosis than JAK2V617F-mutated patients.32,33 Several studies, including a pooled analysis, report that 5% of PMF and 3–5% of ET patients carry MPL mutations and while usually mutually exclusive, a small percentage of patients have both JAK2V617F and MPLW515L mutations.34,35 Mutations in calreticulin (CALR) are present in 17% of patients, most commonly in patients with ET and PMF. 36 More than 50 CALR mutations have been identified to date; however, the 52 base -pair deletion (type 1) and a five base-pair insertion (type 2) are the most common. Type 1 mutations are associated with a more aggressive phenotype and a higher risk of progression to myelofibrosis. Type 2 CALR mutations are more frequently associated with an ET phenotype and a more indolent clinical course. 37 Whereas PMF represents diagnosis made de novo, PET-MF and PPV-MF have nearly identical phenotypes, yet are defined by a progression from ET or PV, respectively.
JAK inhibition: a means to improve survival in myelofibrosis
Allogeneic HCT is the only known curative therapy for patients with PMF, yet long-term disease control and transplant related mortality (TRM) are persistent barriers to survival. Several poor-risk features such as high DIPSS score, adverse cytogenetic abnormalities, advanced age, and severe marrow fibrosis are independently associated with inferior outcomes, and are common in higher-risk PMF patients.14,15,38 Aggressive pre- and post-HCT management may enhance long-term survival after HCT and improve outcomes, particularly in high-risk patients. Ruxolitinib, the first commercially available JAK2 inhibitor, has improved the outcomes of these patients, potentially increases the proportion of myelofibrosis patients eligible for HCT, and, thus increases the number of patients that may eventually be cured of their disease. Further advancements in regulating the JAK-STAT pathway and/or other novel therapies may enhance this effect.
Ruxolitinib, a selective and potent inhibitor of both JAK1 and JAK2, was one of the earliest JAK inhibitors evaluated in hematologic malignancies. In early studies, ruxolitinib was shown to be well-tolerated with rapid, widespread reduction of extramedullary hematopoiesis and vasoactive symptoms in most patients.39,40 These findings were confirmed in the COMFORT-I study in which 41.9% of patients treated with ruxolitinib reached the primary endpoint of ⩾35% reduction in splenic volume and nearly half experienced a ⩾50% reduction in their symptom burden on the Total Symptom Score Symptom Assessment Form (TSS-SAF).16,41 After 5 years of follow up, a limited number of patients maintained a durable reduction in their spleen size as well as improved OS. 16 Similarly, in the COMFORT-II study, a higher percentage of ruxolitinib patients experienced a reduction in spleen volume compared to controls. Again, ruxolitinib was associated with a relative reduction in the risk of death and a trend toward improved OS.17,42
The survival benefits seen in the COMFORT-I and COMFORT-II studies are thought to be secondary to reductions in splenomegaly, enhanced performance status and improvements in constitutional symptoms.16,17 Consistent with this, health-related quality-of-life scores improved with ruxolitinib use, specifically improvements in physical and role functioning, fatigue and appetite loss. 43 Importantly, those patients who benefited most were those with higher-risk disease (i.e. intermediate-2 or high-risk disease).18,44 Compared with other hematologic malignancies, patients with myelofibrosis are unique in that they begin conditioning chemotherapy with their primary disease intact, leading to potential delays in engraftment, and graft failure. Citing improvements in survival, quality of life, and MF-related symptoms in higher-risk patients, pre-HCT treatment with a JAK inhibitor should be considered, although data supporting this approach are limited, and peripheral blood cytopenias and/or intolerance may complicate the use of these agents. At many transplant centers, these patients are treated until maximum response prior to the start of conditioning chemotherapy. Patients who are not considered to be candidates for HCT may receive therapy indefinitely.
The decision to move forward with allogeneic HCT is based on several factors, including disease burden and symptoms, the presence of high-risk mutations such as ASXL1 and SRSF2, quality of life, anticipated life expectancy, performance status, availability of a suitable donor, and others. Several scoring systems, including the Dynamic International Prognostic Scoring System (DIPPS) and DIPPS-plus are routinely used to gauge disease risk and expected survival, and to guide discussions regarding HCT.45,46 Transplantation is not generally recommended for patients with low-risk disease due to a median survival of 15 years.45,47 Individuals with intermediate-1 disease, especially younger patients and persons with high-risk myeloid mutations, should be considered for HCT since the majority will ultimately die from their disease or its complications. 47 Patients with intermediate-2 and higher-grade forms of the disease who are suitable candidates for HCT should be allografted.
In recent years, the use of JAK inhibitors have improved disease outcomes, but they also carry unique challenges. Abruptly stopping these agents may precipitate rebound splenomegaly, worsening of peripheral blood cytopenias and shock-like symptoms. 48 Because of this, and because of concerns associated with the use of myelosuppressive therapy during the period of stem-cell engraftment, JAK inhibitors should be tapered prior to the start of conditioning chemotherapy. There is no universally agreed-upon approach and clinical practice is heterogeneous. One recent study successfully tapered ruxolitinib over 2 weeks and discontinued after the initiation of conditioning chemotherapy and prior to stem-cell infusion (at day 4). 49
A small fraction of PV patients will progress to PPV-MF (approximately 6–14% at 15 years) and be considered for HCT. 50 Younger PV patients may develop PPV-MF later in their lifetime and may be evaluated for allogeneic HCT. Consequently, maximizing vascular fitness, minimizing proliferative complications and thromboembolic disease, and careful monitoring for progression are key elements of a comprehensive care plan for patients with PV and may improve their candidacy. Ruxolitinib is beneficial in PV patients who are refractory to, or intolerant of, hydroxyurea. In the RESPONSE study, 20.9% of ruxolitinib patients met the study’s primary endpoint of improved hematocrit and ⩾35% reduction in spleen volume versus <1% of controls. 51 These responses were durable, with the majority of patients maintaining their response for ⩾80 weeks. 52
Whether the improvement in these patients’ outcomes are indicative of a more indolent disease biology or if these agents favorably alter the disease remains an area of investigation. Regardless, patients with PMF who are able to tolerate pre-HCT JAK inhibition should be treated to maximal response to maximize post-HCT engraftment and survival. Further studies are needed to more definitively describe the favorable post-HCT impact these agents have in PMF.
JAK inhibitors in development
Several JAK inhibitors are currently in clinical development, with a goal to improve upon the early success of ruxolitinib. In the PERSIST-1 study, pacritinib, a combined JAK2 and FLT3 inhibitor, led to significant reductions in splenomegaly by week 24 compared to the best available therapy (excluding other JAK2 inhibitors). 53 In the PERSIST-2 study, pacritinib reduced splenomegaly and improved total symptom scores when compared to the best available therapy (including ruxolitinib). 54 Importantly, pacritinib appears less myelosuppressive and may allow more aggressive therapy in patients with peripheral blood cytopenias. Momelotinib was compared to ruxolitinib in the SIMPLIFY-1 trial in patients with high-risk, treatment naïve PMF, PPV-MF or PET-MF. Momelotinib was non-inferior to ruxolitinib in reducing splenic volume and led to some improvements in hematologic parameters (specifically hemoglobin); however, it did not improve disease-associated symptoms. 55 Likewise, in the SIMPLIFY-2 study, momelotinib failed to improve spleen volumes but its use led to improvements in disease symptoms and transfusion dependence compared to the best available, second-line therapy (including ruxolitinib). 56 Development of momelotinib is stalled, but effective JAK inhibitors that can be safely given to anemic patients with MF are in considerable demand.
Most patients with MPNs have elevated levels of JAK-dependent pro-inflammatory cytokines (e.g. IL-6) consistent with JAK1 hyperactivation. Itacitinib is a selective JAK1 inhibitor that was evaluated in patients with intermediate or high-risk MF regardless of JAK2 mutation status. Patients received benefit from the medication across three separate dose levels. Treated patients experienced reductions in their symptom scores (28.6–35.7%), splenomegaly (14.2–17.4%) and packed red blood cell (PRBC) transfusion needs (53.8%). 57 Recently, the preliminary results of two early studies were presented that stand to further improve upon the success of ruxolitinib. A phase I study of ruxolitinib and the PI3Kδ inhibitor TGR-1202 was undertaken based on the observation that PI3Kδ is highly expressed in patient samples with MF and its ex-vivo inhibition leads to favorable effects in MF and other diseases.58,59 In this study, 83% of patients derived clinical benefit from the combination, with hematologic improvement, reduction in spleen size, and/or symptom improvement. 60 Second, the early results of a phase II study with sotatercept, an activin receptor type IIA (ActRIIA) ligand trap that binds and sequesters the ligands of transforming growth factor beta (TGF-B) superfamily, revealed 5 of 14 patients experiencing improvements in anemia, with reduced transfusion dependence. 61
Inhibiting JAK-STAT signaling in GVHD
The JAK-STAT signaling cascade is involved in the regulation of several inflammatory pathways and plays an integral role in interferon-induced chemokine signaling – an important pathway in acute and chronic GVHD.62–64 Ruxolitinib impairs the differentiation of CD4+ T cells into IFN-γ and IL-17A producing cells and promotes FoxP3+ regulatory T cells, both of which are key components of immune tolerance. 65 Murine studies of ruxolitinib prophylaxis during transplantation with major histocompatibility complex-mismatched grafts demonstrated reduced histologic evidence of GVHD and improved survival.65,66 Some of the earliest clinical data came from six patients with refractory acute GVHD who responded to ruxolitinib. Serum levels of IL-6 and the soluble IL-2 receptor decreased with treatment in all analyzed patients, highlighting its role in controlling the disease. 65 Patients with GVHD treated with ruxolitinib appear to maintain the desirable graft versus leukemia (GVL) effects while controlling GVHD. 66 These findings have led to the recognition that the JAK-STAT signaling cascade is an attractive target in the post-HCT setting. Further retrospective analysis of ruxolitinib in acute and chronic GVHD for 95 patients treated with ruxolitinib salvage therapy for steroid-refractory GVHD from 19 HCT centers revealed an overall response rate (ORR) of 81.5% in acute GVHD (46.3% complete response [CR]) and 85.4% in chronic GVHD. The primary side effects were peripheral blood cytopenias and cytomegalovirus reactivation in both groups. 19 Khoury and colleagues reported the results of 19 patients with severe steroid-dependent chronic GVHD treated with ruxolitinib salvage therapy. Responses were observed after a median of 14 days; 15/19 patients were able to discontinue prednisone and four remained on low doses. 67 In pediatric HCT, of 11 patients treated with ruxolitinib, 45% responded [1 CR, 4 partial response (PR)], and the most frequent toxicities were liver function abnormalities and peripheral blood cytopenias. 68
Ruxolitinib, itacitinib and the JAK1/2 inhibitor baricitinib are being explored in HCT. Preliminary results with itacitinib in patients with grade IIB-IVD acute GVHD revealed an ORR of 83.3% in first-line patients and 64.7% in steroid-refractory patients. 69 Several studies are being planned to evaluate the role of these agents in the maintenance setting after allogeneic HCT and active clinical trials involving JAK inhibitors are noted in Table 1.
Summary of active HCT clinical trials with JAK1/JAK2 inhibitors.
DFS, disease-free survival; ORR, overall response rate; OS, overall survival.
Conclusion
Allogeneic HCT remains the only available curative therapy for MF but is not widely available. The widespread use of JAK inhibitors may be increasing the number of eligible patients; however, post-HCT outcomes are historically poor. Both GVHD and disease relapse represent significant drivers of post-transplant mortality. The JAK-STAT signaling pathway has evolved into an attractive target both before and after transplant. In the pre-HCT setting, JAK inhibitors reduce symptoms and improve quality of life, as well as reduce spleen volumes. After HCT, JAK-STAT inhibition demonstrates significant response rates in both acute and chronic GVHD with preservation GVL-effects. Development of new JAK inhibitors and other novel treatments promises to further enhance the care of these patients.
Footnotes
Author contribution statement
Manuscript was written by MB, MRS and BS. All authors were responsible for reviewing the draft manuscript, and all authors provided approval of the final draft manuscript.
Funding
This research received no specific grant from any funding agency in the public, commercial or not-for-profit sectors.
Conflict of interest statement
The authors declare that there is no conflict of interest.