Burn, Timothy C.

Abstract Background: Current evidence links some of the disease manifestations in myelofibrosis (MF) to abnormal cytokines, likely produced by clonally involved megakaryocytes and monocytes. Furthermore, the recent discovery of JAK2/MPL mutations in MF suggests the contribution of abnormal JAK-STAT signaling to both clonal myeloproliferation and cytokine-driven debility. In order to gain additional pathogenetic insight regarding cytokine-phenotype correlations in MF, we looked into the plasma cytokine profile of MF before and after treatment with INCB018424, a selective JAK1/2 inhibitor. Methods: The current study includes subjects with MF enrolled in an ongoing phase 1–2 study of oral INCB018424 (doses ranging from 25 mg/day to 50 mg BID). Plasma samples were obtained prior to treatment and at intervals of 2 weeks, 1 month and 2 months following INCB018424 dosing. Samples were submitted to Rules Based Medicine Human MAP multiplexed immunoassay system to obtain plasma levels on a range of protein analytes. Results: Plasma cytokine levels in MF patients (n=53) vs. normal healthy volunteers (n=15): Compared to normal controls, plasma levels of pro-inflammatory cytokines and markers of inflammation were significantly increased in MF patients (Table; mean +/− SD values). Furthermore, the observed inflammatory cytokine levels in MF were often higher than those seen in patients with rheumatoid arthritis or cancer-associated cachexia (data to be presented at the meeting). Correlation of plasma cytokine levels in MF with JAK2 V617F mutational status, MF subtype and/or constitutional symptoms/cachexia: Comparison of JAK2V617F positive (n=40) and negative (n=13) MF cases suggested significantly (p<0.01) higher IL-1RA (mean +/− SD = 5575 +/− 917 vs. 1291 +/− 359 pg/ml) and CRP (17.4 +/− 1.6 vs. 6.7 +/− 1.9 μg/ml) levels for the former whereas the other cytokines were elevated to a similar extent. Plasma cytokine levels in PMF (n=30) vs. post-PV MF (n=15) vs. post-ET MF (n=8) were not significantly different. The presence of prior splenectomy did not appear to alter the specific MF cytokine profile (Table 1); the abnormal cytokine profile in MF is, therefore, not necessarily a consequence of marked splenomegaly. There was a direct correlation between the levels of pro-inflammatory cytokines and the presence or absence of constitutional symptoms/cachexia (Figure). Similarly, increased inflammatory cytokine levels in MF were accompanied with significantly decreased leptin levels, a surrogate for nutritional status (Table). Post-INCB018424 treatment cytokine levels: Treatment with INCB018424 induced a rapid decrease in MF-associated inflammatory cytokine levels, in parallel with the observed clinical benefit of both reduced splenomegaly and improvement in constitutional symptoms/cachexia (data to be presented at the meeting). Conclusion: The plasma cytokine profile of MF is reminiscent of a chronic inflammatory state with levels of pro-inflammatory cytokines that are possibly higher than those seen in other inflammatory/malignant conditions. Furthermore, the current study suggests a fundamental link between these cytokines and MF-associated constitutional symptoms/cachexia. Cytokine modulation through JAK-STAT inhibition appears to be a mechanism of action for INCB018424 in MF. Cytokine Normal volunteer (N = 15) MF with splenectomy (N = 6) MF without splenectomy (N = 47) IL-1b (pg/ml) 0.6 +/− 0.04 44 +/− 36 41 +/− 25 IL-1RA (pg/ml) 103 +/− 10 7759 +/− 3939 4111 +/− 763 IL-6 (pg/ml) 0 9.7 +/− 3.3 53.6 +/− 22.5 IL-8 (pg/ml) 7.6 +/− 1.17 1618 +/− 1165 2376 +/− 451 TNFa (pg/ml) 2.6 +/− 0.21 38 +/− 6.8 45 +/− 8.8 TNFRII (ng/ml) 3.1 +/− 0.12 27.3 +/− 7.4 24.7 +/− 2.5 CRP (mg/ml) 1.5 +/− 0.49 21.7 +/− 4.5 13.9 +/− 1.6 Leptin (ng/ml) 10.8 +/− 3.5 3.8 +/− 0.9 2.74 +/− 0.6

Abstract Background: Activating mutations in the pseudokinase domain of JAK2 occur at a high frequency in Philadelphia chromosome-negative myeloproliferative disorders (MPDs). Increasing JAK2 V617F allele burden has been shown to correlate with disease severity (bone marrow dysfunction, organomegaly and constitutional symptoms), which is consistent with exaggerated JAK2 signaling playing a central role in MPDs. INCB018424 is a potent and selective JAK1/2 inhibitor currently being evaluated in the clinic for the treatment of MPDs including primary myelofibrosis (PMF), post-polycythemia vera and essential thrombocythemia myelofibrosis (Post-PV/ET MF). Methods: The percentage of mutant JAK2 V617F relative to wild-type JAK2 was determined in peripheral blood using a quantitative single-nucleotide extension assay. The assay is based on the single nucleotide extension of a DNA primer that anneals adjacent to the V617F mutation. The percentage of JAK2 V617F in bone marrow was determined using pyrosequencing as previously described. The relationship between JAK2 V617F allele burden and clinical phenotype was assessed with samples obtained at study entry. JAK2 V617F allele burden was assessed in nineteen PMF and Post-PV/ET MF patients treated with 25 mg BID INCB018424 for a minimum of three months using peripheral blood and bone marrow. Results: There is an excellent correlation in the JAK2 V617F allele burden between samples obtained from bone marrow and peripheral blood, suggesting that the fraction of cells bearing the mutant clone remains stable during hematopoiesis in this patient group. The mean %V617F values were 59%, 80% and 86% for Post-ET MF (n=2), Post-PV MF (n=8) and PMF (n=9) patients. V617F allele burden correlated with spleen size, absolute neutrophil count (ANC) and Body Mass Index (BMI) in this patient population (R2 values if 0.61, 0.17 and 0.44, respectively), with a higher allele burden predicting larger spleen size, a higher ANC and a lower BMI. Eighteen of the nineteen evaluable JAK2 V617F-positive patients had baseline JAK2 V617F values above 50%, with the mean %V617F value for all evaluable patients being 81%, implying that the majority of hematopoietic stem cells contain the mutant allele in this patient population. During INCB018424 therapy (at least 3 months of treatment) in these patients, profound improvements in clinical endpoints including reduction or resolution of splenomegaly and constitutional symptoms are observed, yet only a modest decrease in JAK2 V617F allele burden is seen in both bone marrow and peripheral blood (13% and 9 % decrease from baseline, respectively, at 3 months). Conclusions: While the patient group is small and heterogeneous in nature, clinical correlates of V617F allele burden observed in myelofibrosis patients in this study are similar to those reported in PV. In addition, despite profound clinical improvements in these patients, V617F allele burdens do not change dramatically suggesting that the clinical activity of INCB018424 might involve inhibition of downstream signaling in cells harboring activating JAK2 mutations rather than changing the allele burden.