Analysis of Metformin Effects on Bone Marrow Fibrosis and Disease Progression in Primary Myelofibrosis Patients: Preliminary Results of an Open Label Phase II Trial (FIBROMET)

Background: Primary Myelofibrosis (PMF) is a chronic myeloproliferative neoplasm (MPN) characterized by increased myeloid proliferation and associated with mutations that induce tyrosine-kinase activation mainly via JAK-STAT pathway, culminating in extensive bone marrow (BM) fibrosis in the course of disease progression. In contrast to the monoclonal origin of hematopoietic cells, fibroblasts proliferation is polyclonal, and mediators involved in fibrosis, neoangiogenesis and osteosclerosis seem to be involved in disease progression. Metformin (MTF) is a biguanide that exerts selective antineoplastic activity in a variety of malignancies, through its action on nutrients privation and hypoxia, leading to apoptosis. In JAK2-mutated cell lines, MTF reduced cell viability, proliferation and clonogenicity, while in Jak2^V617F knock-in-induced mice, MTF reduced Ba/F3 JAK2^V617F tumor burden and splenomegaly. These data suggest that MTF could have a therapeutic effect in PMF patients.

Aims: To conduct an open label phase II study to evaluate MTF effects on BM fibrosis, inflammation mediators, JAK-STAT pathway activation and disease progression in PMF patients.

Methods: PMF non-diabetic adults were eligible. Subjects with severe renal function impairment were not included. Patients received MTF (Glifage XR®) in rising doses until a maximum of 2500mg PO daily, according to tolerance. Primary endpoint was BM fibrosis reversion. Secondary endpoints included reduction of inflammation and downregulation of the JAK-STAT pathway. Clinical data was systematically compiled. Blood and BM samples were collected at the time points: pretreatment (0), 3 mo and 6 mo. Collagen was evaluated in BM biopsy specimens by Masson's trichrome stain: three representative areas from each slide were analyzed and the collagen/sample area was quantified using Image J software; the mean percentage of each slide was used for statistics. IL-6, IL-8 and TNF-α levels were analyzed in BM samples using multiplex assay. Phosphorylation status of intracellular proteins STAT3 and STAT5 was analyzed by flow cytometry and the percentage of cells was recorded using FlowJo software. In order to evaluate gene modulation following MTF exposure, samples at time points 0 and 6 mo were analyzed by PCR array for insulin signaling genes (PAHS-030Z, Qiagen). Genes with ±1.5 fold-change in both directions were selected for validation. For each experiment, statistical analysis was performed and a p value <0.05 was considered statistically significant. Results were expressed as medians (min-max). This trial was approved by the Institutional and National Review Board; written informed consent was obtained from all subjects.

Results: 11 patients (aged 40-84y) were included. Two subjects had early treatment discontinuation due to non-related causes. The median exposure to MTF was 10 mo (5-11) and the median dose was 2000mg/day (1500-2500mg). The most frequent adverse event was diarrhea (n=3). No life threatening event occurred. A reduction in BM collagen area percentage was observed comparing pretreatment biopsies (26.9% (14.8-53.1%)) versus 3 months (3.8% (2.3-4.0%), p=0.062) and versus 6 months of MTF use (0.84% (0.12-17.1%), p=0.125), however, this result was not statistically significant probably due to the low number of patients analyzed (n=5). IL-6, IL-8 and TNF-α levels did not differ between time points. Flow cytometry analysis demonstrated a trend in STAT3 phosphorylation decrease when comparing pretreatment samples versus 6 months of MTF use, though this result was not statistically significant (p=0.06). Mean fluorescence intensity for pSTAT3 was: pretreatment 10.53 ± 5.75, 3 mo 7.34 ± 2.4, 6 mo 5.41 ± 1.14; and for pSTAT5: pretreatment 14.03 ± 7.41; 3 mo 10.71 ± 7.74; 6 mo 6.03 ± 1.41. PCR array for insulin signaling genes showed 21 genes downregulated after 6 months of MTF treatment, including genes previously associated with MPN phenotype: INS (0 and 6 mo treatment fold-decrease: 0.18), NOS2 (0.24), VEGFA (0.34), LEP (0.34), IGFBP1 (0.38) and IRS2 (0.62).

Conclusions: In this study, metformin showed to be a safe and well-tolerated drug. Our preliminary results demonstrated a trend in BM collagen reduction in PMF patients following metformin treatment. Downregulation of important genes associated with MPN phenotype was also noted. The trial is ongoing and these results will be validated at other time points for all subjects.