Myelofibrosis (MF) is characterized by hematopoiesis occurring not only in the marrow but also in extramedullary sites such as the spleen. Since the spleen is not normally a site of hematopoiesis in man, we hypothesized that the MF hematopoietic cells (HC) condition the splenic stroma to create a “favorable” pathologic microenvironment which differentially influences MF-hematopoietic stem cell (HSC)/hematopoietic progenitor cell (HPC) proliferation and trafficking that ultimately determines disease phenotype and progression. In order to test this hypothesis, we phenotypically and functionally characterized fibroblasts present within the spleens of patients with MF.

Single cell suspensions of splenic cells were prepared from surgically removed MF and normal spleens. Splenic cells were then cultured in DMEM supplemented with 10% FBS without additional cytokines until a substantial outgrowth of fibroblasts was observed, at which point cells were trypsinized and subsequently subcultured until morphologically uniform fibroblast monolayers formed. MF splenic fibroblasts (MSF) were isolated and purified from 2 out of 8 MF spleens studied (SP3 and SP11), but fibroblasts were isolated and purified from all 3 normal spleens available. MSFs and normal splenic fibroblasts (NSF) were phenotypically characterized by monoclonal antibody staining and flow cytometry. MSFs and NSFs both were negative for CD45 and CD34, indicating that they are distinct from HCs. MSFs and NSFs expressed similar levels of CD90, CD105 and CD29, as normal bone marrow mesenchymal stem cells as well as the integrins, CD44 and VLA-4. However, higher levels of α-smooth muscle actin (α-SMA) and vimentin, characteristics of cancer activated fibroblasts (CAF), were expressed by MSFs as compared with NSFs (% α-SMA: MSFs: 81.3±1.6%; NSFs: 65.8 ±12.2%; % vimentin: MSFs: 94.1±0.6%; NSFs: 31.9 ±16.5). Moreover, NSFs cultured with medium conditioned with MF splenic mononuclear cells (MNCs) nearly doubled the expression of α-SMA and vimentin by NSFs. Unlike the HCs from JAK2V617F positive patients (JAK2V617F allele burden of SP3 and SP11 PB granulocytes was 85% and 90%, respectively), MSFs from these 2 patients lacked JAK2V617F. Moreover, none of SP11 MSFs had a marker chromosomal abnormality, del(20q), although 58% of splenic MNCs had del(20q). These findings suggest that MF spleen microenvironmental cells are not affected by genetic alterations observed in MF HCs.

CAFs isolated from a variety of malignancies have been reported to promote the growth of cancer cells, increase angiogenesis to a greater extent, and have a distinct gene expression pattern compared to normal fibroblasts (Allinen M, et al. Cancer Cell 2004; 6:17–32). We therefore assessed the capacity of SP11 MSFs to support MF HSCs/HPCs from the same patient using an in vitro cobblestone area forming assay and an assay for long term culture initiating cells. SP11 MF CD34+ cells were inoculated on a SP11 MSF or NSF monolayer (1:5 ratio with each fibroblasts) and incubated for 4 weeks in IMDM supplemented with 10% FBS without exogenous cytokines. Two weeks following the culture, the number of cobblestone areas formed was greater than that observed when MF CD34+ cells were cultured with NSFs (P<0.01). Four weeks following the culture, MF CD34+ cells cultured on the paired MSF generated more total cells, CD34+, CD34+CXCR4+, CD34+CD90+ cells than that observed in cultures using NSFs (Total, CD34+, CD34+CXCR4+ cells, P all <0.05; CD34+CD90+ cells, P=0.12). A fraction of total cells was also assayed for HPCs. The total number of colonies formed was counted and genotyped for the presence of JAK2V617F and del(20q). A greater number of colonies were generated in cultures of MF CD34+ cells on the paired MSFs than that achieved on the NSFs. Moreover, 55% and 9% of the colonies generated in cultures on MSFs were JAK2V617F+ and had del(20q), respectively. By contrast, only 10% of colonies harbored JAK2V617F and none of the colonies had del(20q) in cultures with NSF. These findings suggest that CAFs derived from the spleen of MF patients facilitate the proliferation of malignant MF splenic HSCs/HPCs and likely play a role in determining disease phenotype. Further studies on the mechanism underlying these effects will facilitate our understanding of the role of MF splenic CAFs plays in determining disease progression.


No relevant conflicts of interest to declare.

Author notes


Asterisk with author names denotes non-ASH members.