Mutation-specific responses to cytoreductive therapy in patients with MPN

Myeloproliferative neoplasms (MPNs) are characterized by excessive proliferation of hematopoietic cells and comprise essential thrombocythemia (ET), polycythemia vera (PV), and primary myelofibrosis (PMF), including prefibrotic myelofibrosis (pre-MF). MPNs are associated with an increased risk of thrombohemorrhagic events, reduced life expectancy, and may transform into secondary acute myeloid leukemia (sAML). The majority of MPNs are driven by somatic mutations in JAK2, CALR, or MPL, and thus, increasing knowledge on the complex molecular landscape in MPN has led to more personalized prediction of outcomes and improved clinical decision-making. However, the predictive role of somatic mutations regarding response and resistance to cytoreductive therapy remains uncertain.

The MPN Hub is pleased to summarize the key findings from a recent study by Knudsen and colleagues¹ published in Blood Advances, investigating the role of somatic mutations in patients with MPN enrolled in the DALIAH trial (NCT01387763), treated with cytoreductive therapy (interferon-α [IFN-α] vs hydroxyurea [HU]).

Study design

The DALIAH trial was an investigator-initiated, open label, randomized controlled, phase III trial. Eligible patients were aged ≥18 years with a diagnosis of ET, PV, pre-MF or PMF based on the World Health Organization (WHO) 2008 criteria. Patients aged >60 years were randomly allocated (1:1:1) to receive HU, IFNα-2a, or IFNα-2b, whereas patients aged ≤60 years were randomly allocated (1:1) to receive IFNα-2a, or IFNα-2b. Treatment dose was modified based on efficacy and toxicity according to predefined dose levels.

Aims and method

• The primary aim of the study by Knudsen et al.¹ was to evaluate the association between complete clinicohematologic response (CHR) at 24 months and molecular response to IFNα and HU through sequential assessment of 120 genes.
• Next-generation sequencing (NGS) was performed in 202 and 135 pre- and 24 months post-treatment samples, respectively, from patients enrolled in the DALIAH trial. Amongst the 135 post-treatment patients, 121 were eligible for CHR assessment at 24 months.

Results

Clinical characteristics and somatic mutations at baseline

The median age was 62 years (range, 20–88), and 55% of the enrolled patients were male. The clinical baseline characteristics are summarized in Table 1.

Table 1. Clinical baseline characteristics by treatment*

Somatic mutations in 34 genes were detected in 191 patients (95%) at baseline and MPN phenotypic driver mutations were present in 92% of patients (Table 2).

• Depending on the MPN subtype, the number of mutations were significantly different (mean number of mutations: ET = 1.6; PV = 2.0; Pre-MF = 2.1; and PMF = 1.9).

• Although phenotypic driver mutations are considered mutually exclusive, JAK2^V617F and MPL mutations co-existed in three (1%) patients while 16 patients (8%) were triple-negative for JAK2, CALR, and MPL mutations.
• Patients with JAK2 uniparental disomy (JAK2-UPD) showed a significantly higher median JAK2 variant allele frequency (VAF) (0.48; interquartile range (IQR), 0.35–0.68) compared with those without JAK2-UPD (0.15; IQR, 0.09–0.26; p < 0.0001).
• Most frequent concomitant mutations included TET2, DNMT3A and ASXL1.
• Four percent of patients showed spliceosome gene mutations (SF3B1, SRSF2, U2AF1, ZRSR2) while 6% showed RAS/MAPK signaling (CBL, KRAS, NRAS, NF1, PTPN11, and RIT1).

Table 2. Somatic mutations at baseline by disease subtype*

ET, essential thrombocythemia; PV, polycythemia vera; PMF, primary myelofibrosis; Pre-MF, prefibrotic myelofibrosis. *Adapted from Knudsen et al.1 †Values in bold are statistically significant.
Mutations, %	ET (n = 72)	PV (n = 89)	Pre-MF (n = 16)	PMF (n = 25)	Total (n = 202)	p value†
Patients with no mutations	13	2	0	0	5	0.026
Patients with mutations
Triple negative	18	2	6	4	8	0.007
Driver mutations
JAK2	50	97	69	56	74	<0.001
JAK2-UPD	0	54	6	28	28	<0.001
CALR	22	0	25	36	14	<0.001
MPL	11	0	6	4	5	0.004
Concomitant mutations
TET2	19	26	31	28	24	0.61
DNMT3A	15	17	25	12	16	0.75
ASXL1	6	10	19	16	10	0.20

Association between somatic mutations and clinical characteristics at baseline

• JAK2-UPD was most frequently detected in patients with PV and was significantly associated with higher hemoglobin (p = 0.0003), hematocrit (p < 0.0001), neutrophil count (p = 0.039), and lower platelet count (p < 0.0001) compared to patients with PV without JAK2-UPD.
• Patients with ET were more likely to present with triple-negative disease compared to other MPN subtypes, and significantly associated with younger age compared with patients harboring one of the three phenotypic driver mutations (median 44 vs 64 years; p = 0.006).
• CALR was significantly associated with higher platelet count (p = 0.004) or elevated lactate dehydrogenase (LDH) (p = 0.0008) in patients with ET, pre-MF and PMF compared to patients with JAK2 (+/− MPL)-mutated MPN or triple-negative MPN.
• Co-existence of ASXL1 was significantly associated with JAK2, detected in 13% of JAK2-mutated vs 2% of JAK2-WT patients (p = 0.029).
• Age ≥60 years (54% vs 26%; p < 0.0001) as well as a history of major thrombosis (age-adjusted odds ratio [OR] = 1.96; 95% CI, 0.94–4.12; p = 0.073) and prior stroke (age-adjusted OR = 5.29; 95% CI, 1.59–17.54; p = 0.007) were significantly associated with mutations in TET2, DNMT3A, or ASXL1.
• Prior stroke (age-adjusted OR = 3.03; 95% CI, 1.03–9.01; p = 0.044) was also significantly associated with TET2 alone.

CHR at 24 months and its association with somatic mutations

• Median time to CHR was 5.7 months (IQR, 1.8–10.5) for HU, 4.9 months (IQR, 2.1–8.9) for IFNα-2a, and 6.0 months (IQR, 1.8–10.1) for IFNα-2b.
• CHR was achieved in 21%, and 26% of patients treated with HU and IFNα, respectively (p = 0.68).
• A total of 40% of all patients discontinued treatment at 24 months, with treatment-related toxicity as the most common reason across all groups (HU, 8% vs IFNα-2a, 30% vs IFNα-2b, 38%)

NGS was performed in 135 patients at 24 months, and driver mutations were still detectable in all patients.

• Significantly more patients treated with IFNα showed decreased JAK2 VAF compared to those treated with HU (94% vs 75%, respectively; p = 0.01). The median absolute JAK2 VAF reduction was also significantly greater in patients treated with IFNα compared with HU (0.11 vs 0.05, respectively; p = 0.005).
• Patients treated with IFNα harboring JAK2-UPD had a greater absolute JAK2 VAF reduction compared with those without JAK2-UPD (p < 0.0001) while patients treated with HU harboring JAK-UPD showed no significant reduction in JAK2 VAF (p = 0.76).
• The most frequent treatment-emergent mutation was DNMT3A (39%), followed by TET2 (11%), ASXL1, PPM1D, and TP53 (8%, each).
  • The median VAF of these mutations was low (1.5%) and mainly occurred in patients with JAK2 mutation (97%).
  • Treatment-emergent mutations in DNMT3A were more common in patients treated with IFNα (61% vs 21% with HU; p = 0.046) while PPM1D or TP53 mutations were more common in patients treated with HU (36% vs 6% with IFNα; p = 0.06).

• CHR at 24 months was attained in 23% (34/150), and 37% (11/29) of patients with JAK2 or CALR, respectively, and 21% (18/84) in patients with DNMT3A, TET2, or ASXL1 mutations, respectively.
• Patients treated with HU achieving CHR showed a significant decline in JAK2 VAF (median 0.25–0.08; p = 0.03) compared with those not achieving CHR (median 0.30–0.26; p = 0.10).
• Similarly, JAK2-positive patients treated with IFNα, attaining CHR had significant reduction in the JAK2 VAF (median 0.29–0.07; p < 0.0001) compared with those who did not achieve CHR (median 0.27–0.14; p < 0.0001).
• On the contrary, patients treated with IFNα, harboring mutant CALR, VAF did not significantly reduce either in those achieving CHR (median 0.17–0.13; p = 0.078) or those not achieving CHR (median 0.21–0.17; p = 0.066).

CHR in patients with/without treatment-emergent mutations: Significantly more patients treated with IFNα and without treatment-emergent mutations, achieved CHR (51%, 35/68 vs 22%, 4/18; p = 0.034) compared to patients treated with HU without treatment-emergent mutations, respectively.

• Treatment-emergent DNMT3A-mutations were significantly increased among patients treated with IFNα failing to achieve CHR compared with non-DMNT3A-mutations, respectively (89%, 8/9 vs 20%, 1/5; p = 0.02).

Conclusion

This study elucidated the association between mutation-specific responses to IFNα and HU at 24 months in patients with MPN. Distinct molecular responses were observed with IFNα in JAK2- and CALR-mutated patients with MPN. The study also demonstrated that DNMT3A-mutated clones/subclones emerged during treatment with IFNα in patients who did not achieve CHR. The findings from this study show treatment- and mutation-specific patterns of responses in patients with MPN treated with IFNα and HU, which may have clinical implications in the future.