Abstract Interferon-alpha (IFNα) remains the only cytoreductive therapy for myeloproliferative neoplasms (MPNs) capable of inducing molecular remission through reduction of the JAK2V617F variant allele frequency (VAF). However, clinical responses to IFNα are variable, and molecular remissions often require prolonged treatment. Predictive biomarkers of response are lacking, and early indicators of molecular efficacy could significantly inform treatment decisions. Here, we report findings from a retrospective analysis of MPN patients treated with IFNα and a prospective longitudinal study employing high-frequency sampling to assess early VAF kinetics across multiple biospecimen types, including saliva (a non-invasive source which contains abundant white blood cells), whole blood, and plasma. In a retrospective cohort of 62 MPN patients receiving various therapies, including IFNα, we observed discordance between hematologic and molecular responses. While 87% of IFNα-treated patients achieved hematologic remission by European LeukemiaNet criteria, only 23% of all patients with hematologic remission showed 50% reduction in JAK2V617F VAF. Among IFNα responders, a reproducible pattern emerged, an early rise in JAK2V617F VAF followed by a gradual decline. This supports the hypothesis that IFNα may induce cycling and subsequent exhaustion of mutant hematopoietic stem cells. To evaluate this observation prospectively, we enrolled 12 MPN patients initiating or reinitiating IFNα therapy in a year-long study involving weekly saliva collection and monthly peripheral blood draws. Participants self-collected saliva samples remotely, allowing high-frequency, non-invasive monitoring. Paired blood samples were used to quantify VAF in whole blood, WBCs, plasma (cell-free DNA), and colony-forming myeloid progenitors. Salivary DNA proved consistent in yield and quality, and JAK2V617F VAF trends tracked closely with those in blood when adjusted for epithelial contamination via epigenetic cell counting (ECC). Among evaluable JAK2V617F patients, two early VAF patterns were identified: a transient rise followed by decline, and a steady decline from baseline. While both groups had similar trends in complete blood counts, spleen size, and symptom scores, those with an initial VAF rise exhibited lower neutrophil-to-lymphocyte ratios (NLR), a biomarker previously associated with favorable IFNα responses. These findings suggest that early VAF elevation may reflect biologically productive clonal cycling, potentially preceding molecular remission. We also examined additional biological correlates of IFNα response, including myeloid progenitor genotyping and whole-blood RNA sequencing. Colony-forming assays showed shifts in the ratio of wild-type to heterozygous and homozygous JAK2 mutant colonies over time, providing insight into the clonal architecture of the response. However, bulk RNA-seq analysis showed strong inter-patient variability and no consistent gene expression signature of response. This likely reflects both sample heterogeneity and limitations of whole-blood profiling; future work incorporating single-cell RNA-seq or purified cell populations may yield more informative biomarkers. In conclusion, this study demonstrates the feasibility of intensive, non-invasive longitudinal monitoring of JAK2V617F VAF in MPN patients on IFNα therapy. Saliva, when corrected with ECC, provides a robust alternative to blood sampling. Our data suggest that early VAF kinetics, particularly a transient rise followed by a decline, may serve as a predictive biomarker of IFNα response and should be further evaluated in larger cohorts. Identification of such biomarkers could enable early adaptation of treatment strategies and facilitate precision therapy for patients with MPN.
Sworder et al. (Mon,) studied this question.