Distinct Bone Marrow Microenvironment Abnormalities in MDS and MPN with Fibrosis

Background: Bone marrow (BM) fibrosis is common in both myelodysplastic syndromes (MDS) and myeloproliferative neoplasms (MPN), its pathogenesis remains unclear. It is also unknown whether and how the pathophysiology of BM fibrosis in MDS and MPN are similar or different, while it is anticipated that altered BM microenvironment plays a key role.

Methods: To visualize various BM microenvironmental cells in MDS or MPN patients as compared to those in control subjects, we utilized multicolor immunostaining using a tyramide signal amplification system and a fluorescence spectrum analyzer. Paraffin-embedded BM biopsy samples from 47 patients with MDS (26 without and 21 with fibrosis), 10 with primary myelofibrosis or myelofibrosis secondary to essential thrombocythemia (MPN-F), and 17 with non-Hodgkin lymphoma without BM involvement (as control) were immunostained with antibodies for mesenchymal/neuronal cell markers. Gene mutations were analyzed by targeted sequencing.

Results: In control samples, C-X-C motif ligand 12 (CXCL12) was stained mainly in cells at perivascular areas and the periphery of a minor proportion of adipocytes. These patterns did not differ in MDS without fibrosis from control. However, CXCL12-stained non-adipose cells were significantly increased in MDS with fibrosis (MDS-F) [median (range), 5190 (2117-14299) μm²/mm² as compared to 628 (464-1331) μm²/mm² in control; p=0.00453]. These cells tended to be increased in also MPN-F though statistical significance was not reached [6096 (1098-24458) μm²/mm²; p=0.06304]. The localization of these cells was not confined to the perivascular areas in both diseases. There were no apparent differences in the degree or the pattern of staining for CXCL12 between MDS-F and MPN-F. We also focused on Schwann cells and Nestin-expressing stromal cells because these cells have been independently reported to function as an important hematopoietic stem cell niche. Neurofilament heavy chain (NFH) or neuron-specific class III beta-tubulin-positive nerve fibers (Tuj1) surrounded by glial fibrillary acidic protein (GFAP)-positive Schwann cells were rarely observed in control samples. This observation was true of MDS without fibrosis. However, GFAP-positive Schwann cells were significantly increased in MDS-F and MPN-F [median (range), 39457 (1325-60654) um²/mm² and 75951 (19879-326216) um²/mm², as compared to 6258 (1325-14842) um²/mm² in control; p=0.0017 and 0.0034, respectively]. When the section was stained for Nestin, Schwann cells in MPN-F did not show positive signals except for one case. In contrast, Nestin was frequently stained in the GFAP-positive Schwann cells of MDS-F (9 of 21 MDS-F, 42.9%); particularly 5 of 7 cases (71.4%) in which fibrosis was judged as severe (MF-3), showed positive signals. One exceptional case of MPN-F that showed Nestin-positivity in increased Schwann cells was a triple-negative case (no mutations in JAK2, CALR, or MPL); instead, mutations were found in NRAS, TET2, BCOR, and SMC1A, indicating that this case may have a feature of MDS.

Conclusion: Abnormal increases in CXCL12-expressing mesenchymal cells and nerve fiber-surrounding Schwann cells were visualized by multicolor fluorescence in MDS-F and MPN-F. Schwann cells in MDS-F and MPN-F showed a remarkable difference in the expression of Nestin. It is to be elucidated whether the increase in the Schwann cells has a role in the pathogenesis/pathophysiology, and if so, whether those Schwann cells with distinct phenotypes have differential roles in these two disease conditions.