Ruxolitinib Therapy Effectively Modulates CD34⁺ Hematopoietic Progenitors and Bone Marrow Angiogenesis in Patients with Myelofibrosis

Background: Myelofibrosis (MF) associated with myeloproliferative neoplasms (MPN) is a multifactorial process resulting from profound modifications of the bone marrow (BM) and spleen stroma, as demonstrated by the presence of reticulin and collagen fibrosis, osteosclerosis and neo-angiogenesis. These stromal accumulations suggest a deregulation of BM stem cell niches, in which hematopoietic progenitor cells (HPCs) are engaged in a constant crosstalk with their stromal environment. Recently, there is strengthening evidence that the BM microenvironment which might contribute to the clinical outcomes in MF can be effectively modulated by JAK inhibitor therapy.

Methods: A total of 68 patients (pts) with primary, post polycythemia vera or essential thrombocythemia MF, presenting at baseline (BL) with various degrees of BM fibrosis according to the WHO grading system, were selected from a phase 1/2 study (NCT00509899). All cases had both a BL and a sequential BM biopsy taken at 24 months (mo) following therapy with the JAK1/2 inhibitor ruxolitinib (RUX). Analysis included immunohistochemical and morphometric assessment of CD34+ HPCs (overall frequency, clustering index) by consensus after independent blinded review by 3 pathologists and computer assisted evaluation of BM angiogenesis. Quantification of microvascular density (MVD), luminal dilatation or microvessel area (MVA), and additional parameters, such as shape (form factor), tortuosity, and branching (maximal vessel length), were performed by whole slide digital imaging of the trephine biopsy and subsequent image processing using the ImageJ system. For 16 pts, additional analyses for vascular endothelial growth factor (VEGF) expression at BL and 24 mo were performed.

Results: At BL, 79.5% of MF pts presented with a BM fibrosis grade of 2 or 3 according to the WHO definition. Following RUX therapy, 73.5% of cases showed an improvement or stabilization and in 6 pts (8.8%) a further regression to an earlier disease stage was observed. Increase in BM fibers was conspicuously accompanied by higher MVD and MVA at BL, which was also associated with elevated levels of angiogenic cytokines like VEGF. The structure of vessels at BL was significantly altered with an enhanced irregularity of shape and tortuosity. Post-RUX BM samples revealed a reduction in MVD -and in particular MVA- independent of BM fibrosis response. However, cases with improvement in fibrosis post-RUX showed a more profound change in microvessel architecture. Normalization of vessel structure was accompanied by a significant reduction of VEGF expression levels at 24 mo (p < 0.001). Overall frequency and clustering of CD34+ HPCs was reduced by RUX therapy at 24 mo. Improvement or stabilization of BM fibrosis post-RUX was in most cases associated with a corresponding reduction of CD34+ HPCs (88.9%) and a decreased clustering index (77.7%). In this regard, modulation of CD34+ HPCs was associated with a greater spleen size reduction at 24 mo (-10.9 vs -9.7 cm, p=0.061), whereas changes in hemoglobin values and platelet counts showed no correlation to this parameter. Similar effects were observed for MVD, MVA, and additional morphometric alterations of BM microvessel architecure.

Conclusions: Our results support previous reports that therapy with the JAK inhibitor RUX effectively modulates the abnormal microenvironment that contributes to the maintenance of the neoplastic clone. In this regard, it is tempting to speculate that RUX therapy might modify vascular niches and therefore could impact the abnormal trafficking and homing of CD34+ HPCs and endothelial precursors that characterize MF.