Introduction: The progression of polycythemia vera (PV) to myelofibrosis (MF) is associated with significant morbidity and mortality. Interferon alfa (rIFNa), a disease-modifying agent, has potential to delay or prevent post-PV MF and improve overall survival but supporting data are required. We present results of the largest study demonstrating improved myelofibrosis-free and overall survival (MFS and OS) of rIFNa treated PV patients (pts) compared to other PV pts. Objectives: To estimate the MFS and OS of treated PV pts and determine the relative risk for post-PV MF and mortality of those treated with rIFNa compared to those treated with other standard therapy. Methods and Study Design: To ensure sufficient follow-up for analysis of long-term outcomes, this IRB approved study identified all adult pts treated at our center from 1974-2019 according to PVSG criteria (1974-2007), our published criteria (2008-2016) and WHO criteria (2016-2019) using a standardized query of electronic medical records. Demographic data, clinical characteristics, treatment history and outcomes were collected. The extended follow-up of this large PV cohort permitted us to evaluate the effectiveness of PV therapy using both intention-to-treat (ITT) and treatment duration (on-treatment) analyses. In the ITT analysis, pts were assigned to rIFNa or hydroxyurea (HU) arms according to the first cytoreductive treatment they received for at least one year or to phlebotomy only (PHL-O) if no cytoreductive treatment was given. On-treatment analysis was performed to account for cross-over and assess how duration of a given therapy influenced outcomes. The onset of post-PV MF was defined by IWG-MRT criteria. MFS and OS were estimated using Kaplan-Meier methods and the log-rank test compared survival between treatment arms in ITT analysis. Multivariate analysis of post-PV MF risk and mortality was performed using a Cox proportional hazards model. The model accounts for age at diagnosis and is stratified by treatment arms (ITT) or by treatment as a time-dependent covariate (on-treatment). Results: We identified 306 PV pts whose median age at diagnosis was 54 years (yr) (range 20-91) and of whom 151 (49%) were women. The median follow-up was 11 yr (range 1-45). The first line treatment was rIFNa in 75 (25%), HU in 134 (44%) or other cytoreductive regimens in 37 (12%). PHL-O was instituted in 60 pts (20%). Treatment cross-over occurred in 82 pts (27%), with the least from rIFNa arm (22%) (Table2). Treatment arms differed by age at diagnosis with a median of 50, 59 and 52 years for rIFNa, HU and PHL-O (p <0.01) (Table 3). The median MFS and OS was 19.5 and 26.3 yr for the entire group; 27 and 28 yr for rIFNa arm; 18 and 26 yr for HU arm; and 14 and 25 yr for PHL-O (log-rank p<0.01 for MFS and p=0.01 for OS) (Figure 1). In multivariate analysis that included age, rIFNa arm had a lower risk of post-PV MF or death compared to HU arm (HR 0.43, p=0.03 and HR 0.44, p=0.04 respectively) and to PHL-O arm (HR 0.22, p<0.01 and HR 0.35, p=0.03 respectively) (Figure 2). The PHL-O arm had a higher risk of post-PV MF compared to HU as well. Older age at diagnosis was a risk factor for post-PV MF and death. Accounting for cross-over, 138 pts received rIFNa at any time for a cumulative of 980 patient-years (median: 5.3, range 1-25 yr). On-treatment analysis associated rIFNa with an 8% and 7% relative risk reduction of post-PV MF and all-cause mortality respectively (age-adjusted HR of 0.92, p<0.01 and 0.93, p=0.01). Discussion: This is the largest study with the longest follow-up of rIFNa treated PV pts and the first to demonstrate that rIFNa yields superior MFS and OS compared to HU or PHL-O. This study addresses the critical issue that randomized controlled trials to date failed to answer owing to limited follow-up duration and lack of surrogate endpoints for survival. Although the median age of the entire group is younger than the reported median age at PV diagnosis, multivariate analysis showed that both the survival benefit of rIFNa and the reduced risk of fibrosis are independent of age. This study supports early use of rIFNa for PV, especially in younger patients who should not be deprived of a disease-modifying therapy for being "low risk" by consensus criteria. Conclusions: rIFNa yields improved MFS and OS of PV patients, independent of age, in this large study with extended follow up. Early use of rIFNa should be considered routinely in the management of PV. Disclosures Ritchie: agios: Other: Advisory board; Tolero: Other: Advisory board; Genentech: Other: Advisory board; Celgene, Incyte, Novartis, Pfizer: Consultancy; Ariad, Celgene, Incyte, Novartis: Speakers Bureau; Jazz Pharmaceuticals: Research Funding; Celgene, Novartis: Other: travel support; AStella, Bristol-Myers Squibb, Novartis, NS Pharma, Pfizer: Research Funding; Celgene: Other: Advisory board; Pfizer: Other: Advisory board, travel support. Silver:PharmEssentia: Consultancy, Honoraria, Membership on an entity's Board of Directors or advisory committees.

No abstract

SummaryErythropoiesis is a multiweek program coupling massive proliferation with progressive cellular differentiation ultimately enabling a limited number of hematopoietic stem cells (HSCs) to yield millions of erythrocytes per second1. Erythropoietin (Epo) is essential for red blood cell (RBC) production but this cytokine acts well after irreversible commitment of hematopoietic progenitor cells (HPCs) to an erythroid fate. It is not known if terminal erythropoiesis is tethered to the pool of available immature hematopoietic stem and progenitor cells (HSPCs). We now report that megakaryocyte-derived TGFβ1 compartmentalizes hematopoiesis by coupling HPC numbers to production of mature erythrocytes. Genetic deletion of TGFβ1 specifically in megakaryocytes (TGFβ1ΔMk/ΔMk) increased functional HSPCs including committed erythroid progenitors, yet total bone marrow and spleen cellularity and peripheral blood cell counts were entirely normal. Instead, excess erythroid precursors underwent apoptosis, predominantly those erythroblasts expressing the Epo receptor (Epor) but not Kit. Despite there being no deficiency of plasma Epo inTGFβ1ΔMk/ΔMkmice, exogenous Epo rescued survival of excess erythroid precursors and triggered exuberant erythropoiesis. In contrast, exogenous TGFβ1 caused anemia and failed to rescue erythroid apoptosis despite its ability to restore downstream TGFβ-mediated Smad2/3 phosphorylation in HSPCs. Thus, megakaryocytic TGFβ1 regulates the size of the pool of immature HSPCs and in so doing, improves the efficiency of erythropoiesis by governing the feed of lineage-committed erythroid progenitors whose fate is decided by extramedullary renal Epo-producing cells sensing the need for new RBCs. Independent manipulation of distinct immature Epo-unresponsive HSPCs within the hematopoietic compartments offers a new strategy to overcome chronic anemias or possibly other cytopenias.

Background: Therapy for chronic anemias is limited to RBC transfusions and Erythropoiesis Stimulating Agents (ESA) which often work on only transiently or not at all. New approaches to treat chronic anemia are needed but development has been limited by our incomplete understanding of erythropoiesis, most of which relates to the terminal maturation of erythroid precursors. Erythropoietin (Epo) acts during a very narrow window of erythropoiesis, well after progenitor commitment to an exclusively erythroid fate. It is not known if the final steps of RBC maturation are coupled to the earlier stages of hematopoietic stem and progenitor cell (HSPC) differentiation; a process that begins almost three weeks earlier when an HSC starts its march towards committed RBC precursors via a series of branching cell fate decisions.We searched for independent control and compartmentalization of erythropoiesis that could couple early hematopoiesis to terminal erythroid commitment and maturation. Results: We deleted TGFβ1 in megakaryocytes (TGFβ1ΔMk/ΔMk) and found that peripheral blood counts were normal in TGFβ1ΔMk/ΔMkmice compared to TGFβ1FL/FLcontrols despite the pool of primitive hematopoietic cells being expanded (Fig. 1a). Similarly total bone marrow cellularity was normal in TGFβ1ΔMk/ΔMkmice (Fig. 1b). Excess HSCs in TGFβ1ΔMk/ΔMkmice appeared capable of robust differentiation because the number of immature lineage-negative (Linneg) hematopoietic progenitor cells was increased in the marrows of TGFβ1ΔMk/ΔMkmice (Fig. 1c). Thus, it remained unexplained why the expanded number of HSPCs (Fig. 1d) do not increase blood counts and marrow cellularity. We hypothesized that the excess progenitors observed in the TGFβ1ΔMk/ΔMkmice failed to increase blood counts because their progeny were unneeded, and inadequately supported by homeostatic levels of late-acting cytokines. Indeed, bone marrow apoptosis was increased in the TGFβ1ΔMk/ΔMkmice compared to controls, as reported by AnnexinV (AV) binding (Fig. 1e-f). Apoptosis of lineage-marker negative (Linneg), Kit+Sca1neg(LKSneg) HPCs and LKS+HSPCs was rare in both TGFβ1ΔMk/ΔMkmice and littermate controls (Fig. 1g). These results suggest that excess, hematopoietic precursors present in the TGFβ1ΔMk/ΔMkmice are pruned by apoptosis during hematopoietic differentiation. We found 10-fold apoptosis in TGFβ1ΔMk/ΔMkprecursors populations BM (Fig. 1h). Epo levels were normal in the serum of these mice, we reasoned that the excess, unneeded cells were not supported physiologic Epo levels. To test this, we treated mice with exogenous Epo. Indeed, we found that the excess erythroid apoptosis could be rescued by administration of very low doses of Epo (300U/kg)(Fig. 1i-j). Whereas TGFβ1Flox/Floxmice showed minimal reticulocytosis and no change in blood counts, TGFβ1ΔMk/ΔMkmice responded with reticulocytosis and erythrocytosis within 6 days (Fig. 1k-l). In contrast, treatment of mice with TGFβ1 worsened the erythroid apoptosis observed in TGFβ1ΔMk/ΔMkmice and caused mild anemia. These results suggest that erythropoiesis is subject to modular regulation with megakaryocytic TGFβ1 constraining the pool of erythroid committed progenitors that are then licensed to mature via Epo signaling. We thought that blockade of TGFβ signaling could phenocopy these effects by inducing overproduction of erythroid committed precursors. To test this, we pre-treated B6 mice with a TGFβ1 neutralizing antibody (1D11) or non-targeting, isotype control antibody (13C4) and then either PBS or low-dose Epo (Fig. 1m). TGFβ neutralization by 1D11 reduced pSmad2/3 MFI in HSPCs in wild-type mice whereas the 13C4 control had no effect, demonstrating on-target activity (Fig. 1n). Low-dose Epo triggered a brisk erythropoietic response in mice treated with 1D11 but not those treated with the 13C4 control (Fig. 1o). Exogenous Epo rescued the erythroid precursor dropout observed in B6 mice treated with 1D11 but did not affect the low apoptosis observed in mice treated with the 13C4 control (Fig. 1p-r). Therefore, the boundary of megakaryocytic TGFβ1 activity is compartmentalized within the marrow with predominant effects on immature HSPCs while excluding their progeny (Fig. 4s). Conclusion: This work also promises new therapies for chronic anemias by combining TGFβ inhibitors to increase the outflow of immature progenitors with ESAs to support erythroid maturation. Disclosures No relevant conflicts of interest to declare.

Introduction: Polycythemia vera (PV) is characterized by an increased red cell mass resulting in whole blood hyperviscosity, a strong predictor for thrombosis which remains a significant cause of morbidity and mortality1. Leukocytosis, thrombocytosis, and phlebotomy (PHL) rates are reported additive risk factors for thrombosis but their relative significance has been debated. PHL and cytoreductive therapy mitigates the risk of thrombotic complications. However, the relevance of these parameters remain insufficiently studied2,3. Our primary objectives were to assess the significance of these risks and their associations with thrombosis. We also evaluated whether first-line interferon-α (rIFNα), hydroxyurea (HU), and phlebotomy-only (PHL-O) therapy is associated with reduced thrombotic risk. Methods: After IRB approval, 328 patients (pts) were evaluated after diagnosis according to PVSG criteria (1974-2007), published Weill Cornell criteria (2008-2016)4, or WHO 2016 criteria. Demographics, clinical history, laboratory values, bone marrow findings, and genetic mutations were collected by querying our platform containing aggregated clinical data, the Observational Medical Outcomes Partnership Common Data Model5. Using intention-to-treat analysis, pts were assigned to a first-line therapy defined as continuous cytoreductive therapy for ≥1 consecutive year or PHL-O. Covariate differences between the first-line therapy groups at diagnosis were determined using χ2 tests for categorical variables. Overall survival (OS) was derived by the Kaplan-Meier estimator and comparisons of thrombosis risk were performed using a Cox proportional hazards model adjusted for clinically significant covariates such as age. Results: The characteristics of 165 men (50.3%) and 163 women (49.7%) with PV are shown in Table 1. Followup extended up to 45.0 years (yrs) with a median of 10.3 yrs. Median OS was 32.5 yrs. Splanchnic vein thrombosis, stroke, and deep vein thrombosis were the most common events. Median age at first event was 59.1 yrs. Predisposing factors associated with thrombosis included uncontrolled HCT, increasing leukocytosis, and ≥5 PHL during the first year after diagnosis. Elevated HCT was the greatest contributor to thrombosis: males with HCT of 53.0% had thrombotic complications at 10 times the rate of those with HCT of 43.5% and females with HCT of 51.9% had a thrombosis at 6 times the rate of those with HCT 40.0% (Fig 1). Pts with leukocytosis (30.8x109/L) had a thrombosis at 2.5 times the rate of those with a white blood cell (WBC) value of 9.4x109/L. There was a weak association between platelet (PLT) count and thrombosis (data not shown). The mean number of first-year PHL was 5.3±4.3; those pts requiring ≥5 had a thrombosis at 2 times the rate of those that required less (p=0.011). The difference in cumulative incidence of thrombotic events within 10 yrs was statistically significant for pts requiring ≥5 first-year PHL (p=0.031, Fig 2). There was a significant difference in the cumulative incidence of thrombosis during the first 10 yrs of diagnosis dependent on first-line therapy (PHL-O: n=117, HU: n=84, rIFNα: n=40) (p=0.021, Fig 3). The apparent superiority of IFN over HU and PHL in this retrospective study is suggested by 10 year cumulative incidences of thrombotic events of 3.2, 18.0, and 30.6%, respectively. Discussion: There was a significant correlation between HCT, WBC count, and PHL requirements, but not PLT count, with thrombotic events. Elevated HCT was the most important risk factor. However, leukocytosis and first-year PHL rates also contribute as indicated by the higher hazard ratios. This suggests a need for cytoreductive therapy from disease onset. Our analysis shows that rIFNα reduces thrombotic risk when compared to HU and PHL. Conclusion: Multivariate analysis indicates that elevated HCT level is the most important parameter for correlation with thrombosis within the first 10 yrs of illness. However, WBC and PHL rates are also significant. Since thrombosis occurs from the time of diagnosis, our data suggest the need for cytoreductive intervention from onset. Analysis of the three most common first-year therapies shows that HU and PHL are inferior to rIFNα in reducing thrombotic risk. Disclosures Ritchie: Celgene: Other: Advisory board; Pfizer: Other: Advisory board, travel support; Celgene, Novartis: Other: travel support; Jazz Pharmaceuticals: Research Funding; Genentech: Other: Advisory board; agios: Other: Advisory board; Tolero: Other: Advisory board; AStella, Bristol-Myers Squibb, Novartis, NS Pharma, Pfizer: Research Funding; Ariad, Celgene, Incyte, Novartis: Speakers Bureau; Celgene, Incyte, Novartis, Pfizer: Consultancy. Silver:PharmEssentia: Consultancy, Honoraria, Membership on an entity's Board of Directors or advisory committees.