Although the clinical course of patients with myelofibrosis (MF) is punctuated by the development of extramedullary hematopoiesis leading to symptomatic splenomegaly, the spleen is not a hematopoietic organ in normal adults. Spleens from patients with MF are characterized by increased numbers of hematopoietic cells and increased micro vessel density within the red pulp. Previously, we reported that the siderophore, lipocalin-2 (LCN2), was elaborated by MF myeloid cells which leads to increased MF marrow fibrosis, osteoblastogenesis,and depletion of the pool of normal hematopoietic cells but promotion of the proliferation of MF cells (Lu M, Blood, 2015). Using nanostring nCounter technology, we have now demonstrated that LCN2 transcript levels were significantly higher in MF spleen (4 fold) than in normal spleen samples (p=0.008) and that LCN2 transcript levels were only modestly increased in MF marrow (p=0.48). The data identified the spleen as a major source of LCN2 in MF patients. These findings led us to further examine the effects of LCN2 on MF spleen CD34+ cells and micro environmental cells. Normal spleen adherent cells (AC) were treated with LCN2 and conditioned media (CM) were collected. CD34+ cells isolated from normal BM or MF spleen were cultured with this CM for an additional 5 days. LCN2 treated AC CM increased MF hematopoietic cell numbers by 2 fold and increased MF CFU-GM and BFU-E numbers by 40% (p<0.05) and 90% (p<0.01), respectively, but decreased the proliferation of normal HPC. These effects were far greater than that observed with the addition of LCN2 alone, suggesting that cytokines beyond LCN2 are elaborated by splenic AC that are capable of affecting MF hematopoiesis. We next demonstrated that protein levels of IL-8 and CXCL1, but not CXCL2 were dramatically elevated in AC CM following treatment with LCN2 (p=0.037 and p=0.046). By contrast, LCN2 treatment of marrow AC led to the generation of CM with increased levels of CXCL2 but not CXCL1, demonstrating that a differential responses of splenic and marrow AC to LCN2. Furthermore, LCN2 increased VEGF, IL-8 and CXCL1 mRNA levels in splenic stromal cells. Splenic stromal cell IL-8 transcript levels were knocked down using siRNA resulting in a reduction in VEGF mRNA levels, which could be overcome by addition of greater amounts of IL-8. Furthermore, the effect of the additional IL-8 was blocked by adding an anti-IL8 antibody, thereby confirming the regulation of VEGF by IL-8. These results indicate that LCN2 promotes the elaboration of a cascade of cytokines by splenic AC which are capable of affecting splenic endothelial cell proliferation as well as malignant hematopoiesis.
We, then, cultured MF splenic and normal BM CD34+ cells in media alone and in a co-culture system with MF splenic endothelial cells in the absence of exogenous cytokines. The total MF and normal BM cell number was dramatically increased when they were co-cultured with EC as compared to the CD34+ cells cultured alone. The absolute MF CD34+ cell number was increased more than 50 fold (p<0.001), while the numbers of normal BM CD34+ cells were increased to a lesser degree (18 fold). LCN2 treatment of the endothelial cells layers did not further increase their ability to promote the proliferation of MF CD34+ cells. These data indicate that splenic endothelial cells serve as a niche which supports hematopoiesis within the spleens of MF patients and that splenic fibroblasts elaborate a number of cytokines which affect splenic micro vessel density. We attempted to target this MF hematopoietic / endothelial cell interaction by adding a CXCR1/2 receptor antagonist, reparixin, to the endothelial cell based cultures which decreased MF CD34+ cell numbers by 35 % (p=0.04). These data suggest that CXCR1/2 receptor antagonists might be used to disrupt the endothelial cell niche within the MF spleen. In summary, our data indicate that LCN2 promotes the elaboration of cytokines such as CXCL1, IL-8 and VEGF within the splenic microenvironment which leads to an increase in splenic endothelial cell proliferation which is supportive of MF CD34+ cell proliferation. Furthermore this cascade of events can be interrupted by potential treatments which reduce LCN2 levels as well as reparixin which disrupts the hematopoietic cell niche supporting activity of splenic endothelial cells, thereby providing a means of correcting the dysregulated tumor microenvironment present within the MF spleen.
No relevant conflicts of interest to declare.
Asterisk with author names denotes non-ASH members.