Mutations in JAK2, CALR, and MPL and the effect on response to IFNα in MPN

Classical BCR-ABL1-negative myeloproliferative neoplasms (MPN), including primary myelofibrosis (PMF), polycythemia vera (PV) and essential thrombocythemia (ET), are typically characterized by stem-cell-derived clonal proliferation and driver mutations in Janus kinase 2 (JAK2^V617F), calreticulin (CALR^mut; includes CALR^del52 and CALR^ins5), or myeloproliferative leukemia virus oncogene (MPL^mut) genes.

Interferon alpha (IFNα), particularly pegylated IFNα, has been shown to produce hematologic responses in ET, PV, and early PMF. Unlike cytoreductive and JAK inhibitor therapy, IFNα has been shown to decrease the variant allele frequency (VAF) of JAK2^V617F in the blood cells of approximately 60% of patients, though its impact on CALR^mut is uncertain.¹

Mosca, et al. sought to identify factors that influence the molecular response rate to IFNα and to determine its molecular efficacy when faced with cells carrying driver mutations. Their work has recently been published in Blood and is summarized here.¹

Study design

The team performed a prospective, longitudinal, observational cohort study of 48 patients being treated with pegylated-IFNα, including 21 patients with PV, 22 with ET, and five with PMF.

• Blood and bone marrow samples were collected before and during, or only during, IFNα treatment for 3 months following treatment initiation.
• Samples were collected several times per year.
• CD34⁺ progenitor cells (at varying stages of maturity; HSC-enriched, immature, committed), granulocytes and platelets were isolated and seeded into plates with one cell/well.
• Colonies were isolated 14 days later, each corresponding to the clonal expansion of one progenitor cell.
• DNA was extracted from each colony and genotyped for JAK2, CALR and MPL status.

Two evaluations were performed:

• The effect of genotype and IFNα dosage on progenitor and mature cells from all patients.
• IFNα-induced dynamics of mutated hematopoietic stem cells (HSC) using mathematical modelling.

Model calibration was performed using 40 patients with data from at least three time points and hypothesis testing was performed on all patients with at least five progenitor measurements for JAK2^V617F

A compartmental mathematical model used:

• Compartment 1: inactive HSC,
• Compartment 2: active HSC, which can generate:
• Compartment 3: differentiated progenitors, which in turn will give rise to:
• Compartment 4: mature cells.

Statistical inferences were used to distinguish three populations of patients based on disease mutations. Within each population, patient- specific parameters were estimated based on Bayesian hierarchical framework.

Characteristics of study participants and their hematologic responses to IFNα are detailed in Table 1.  

Table 1. Patient characteristics and hematological response to IFNα*

Results

A total of 84,845 genotypes from progenitor derived colonies were analyzed. At a given time point the JAK2^V617F VAF in platelets and granulocytes was correlated and was similar among all progenitor types. From the VAF data for platelets and granulocytes, Mosca, et al. assessed the dynamics of JAK2^V617F, CALR^mut and MPL^mut cells during IFNα treatment and found that while CALR^mut VAF remained stable, JAK2^V617F VAF decreased over time in both the mature hematopoietic cells and the progenitors. CALR^mut and JAK2^V617F VAF became substantially different from each other following 600 days of treatment and the two MPL^mut cases showed a visible molecular response over time.

Effect of IFNα dosage on treatment response

For patients treated with higher IFNα doses, JAK2^V617F VAF decreased substantially in progenitor cells, following an initial increase in some cases, with greatest responses seen from Day 1,000 after treatment initiation.

In contrast, a progressive and continuous decrease in JAK2^V617F VAF was seen for patients treated with lower doses of IFNα. From around 600 days from treatment initiation, progenitors with homozygous JAK2^V617F were significantly more sensitive to IFNα treatment than those with heterozygous JAK2^V617F. CALR^mut progenitors were not associated with a significant IFNα dosage effect.

Modelling to infer IFNα-induced dynamics of mutated HSC and IFNα mechanism of action

The following inferences were made based on the mathematical modeling approach:

• Rapid overall depletion of homozygous JAK2^V617F HSC and of heterozygous MPL^mut HSC.
• Slow and simultaneous depletion of HSC, granulocytes and progenitors among heterozygous JAK2^V617F cells.
• Slow depletion of HSC in patients with CALR^mut cells who had a molecular response (most CALR^mut cells were heterozygous).

Inferred HSC dynamics supports the hypothesis that IFNα acts on mutated HSC. Also designed to provide insights into the IFNα mechanism of action, the model inferred that:

• IFNα promotes differentiating divisions of HSCs harboring JAK2^V617F and CALR^ins5 suggesting that it reduces their self-renewal capabilities.
• IFNα promotes an exit from quiescence more readily in homozygous than heterozygous JAK2^V617F HSCs.

Differential targeting of HSCs depending on zygosity, mutation type and IFNα dosage

• Despite few cases of CALR^del52 mutations, results suggested that HSCs with this mutation were less effectively targeted than those with CALR^ins5 mutations.
• In HSCs harboring CALR^mut, a higher dose of IFNα was associated with a poorer HSC response.
• A molecular response was detected in most HSCs with JAK2^V617F, with a response factor (R factor) that was significantly better in homozygous than heterozygous HSC.
• Heterozygous JAK2^V617F HSCs had an improved response with high dose IFNα (Mann-Whitney U, p = 0.0745; test of nullity of linear regression coefficient, p = 0.0498).
• An IFNα dosage effect was not observed for homozygous JAK2^V617F HSCs.
• Homozygous JAK2^V617F HSCs were depleted more rapidly than heterozygous JAK2^V617F HSC in responding patients (50% reduction in R factor was 350 days vs 920 days respectively).
• MPL^mut and CALR^ins5 achieved a 50% reduction in R factor at ~600 days.

Conclusion

Mosca, et al. concluded that the efficacy of IFNα in exhausting mutated HSC is dependent on the driver mutation type, with varying degrees of response relating to the mutations harbored by HSCs. IFNα is more efficient at targeting homozygous than heterozygous JAK2^V617F, and preferentially targets CALR^ins5 than CALR^del52. Dosage assessments in this study found that titration to a maximum tolerated dose was more likely to achieve a reduction in heterozygous JAK2^V617F HSC, thereby achieving a hematologic response, suggesting that dose intensity should be maintained throughout treatment. Conversely there is no evidence to support high-dose IFNα treatment in patients with homozygous JAK2^V617F HSC or in patients with CALR^mut HSC.

Currently, the use of IFNα is guided by hematologic response and tolerability, rather than causal mutation, but the authors hope that this study will highlight the importance of understanding the differential effects of these mutations in MPN management.