A Novel in-Vivo Model of Human Myelofibrosis Based on Transplantation of Thrombopoietin Overexpressing CD34+ Cells into Humanized Bone Marrows in NSG Mice Recapitulates Key Disease Hallmarks

Myelofibrosis (MF) is the most severe form of the myeloproliferative neoplasms (MPN). Current standard MF therapies - except for allogeneic stem cell transplantation - are not curative and have only a limited effect on fibrosis, leaving patients with an urgent need for more effective and targeted treatments. While murine models have successfully captured some aspects of MF, the true challenge lies in creating an in vivo model that faithfully mirrors human MF pathophysiology. We therefore aimed to establish the first humanized MF model based on transplantation of thrombopoietin (THPO)-overexpressing human CD34+ cells into humanized bone marrow tissues established in immunodeficient NSG mice.

We first generated humanized bone marrow (BM) tissues (ossicles) by subcutaneous implantation of 2×10⁶ human BM-derived stromal cells (BMSCs) into immunodeficient NSG using a modified matrix preparation. This resulted in the formation of subcutaneous ossicles containing functional BM after 8 weeks (8-week ossicles).

To mimic human MF development in vivo, we adopted a strategy to overexpress THPO in hematopoietic stem and progenitors (HSPCs). We generated THPO-overexpressing (THPO-oe) CD34+ HSPCs by lentiviral transduction of human CD34-enriched (MACS) bone marrow cells using THPO-GFP lentivirus. Control cells were transduced with control lentivirus vector (GFP).

THPO-oe CD34+ cells and controls, respectively, were transplanted into 8-week ossicles by direct intraossicle injection (3×10⁵ cells/ossicle). FACS analysis of ossicles showed robust engraftment (up to > 80%) of viable human hematopoietic cells as early as 4 weeks after transplantation of THPO-oe CD34+ cells. The cellularity of THPO-oe CD34+ injected ossicles was 1.4 ± 0.6 ×10⁷ cells/ossicle which was comparable to controls. Engrafted human CD45+ cells demonstrated a significantly higher myeloid to lymphoid ratio of 0.14 in the THPO group (n=17) compared with 0.08 for GFP controls (n=14) (p = 0.0147). THPO overexpression in human BM CD34+ cells furthermore resulted in a notable increase in human ossicle megakaryocytes as indicated by a 1.67-fold increase (n=11) of CD41a+/CD42b+ megakaryocytes compared with controls. An increase in megakaryocyte production from THPO-oe CD34+ cells was furthermore confirmed by in vitro culture experiments. Additionally and importantly, histological analysis of ossicles showed marked reticulin fibrosis in 8-week THPO-oe CD34+ cell transplanted ossicles compared to controls, which corresponded to grade 2 MF. Time course experiments are in progress to evaluate the kinetics of fibrosis development in this model.

Interestingly, human hematopoietic cells were clearly detected in murine spleens and femurs as early as 4 weeks following intra-ossicle injection, indicating that transplanted human cells migrated to murine hematopoietic tissues. However, spleen size was not increased in mice transplanted with THPO-oe CD34+ cells and preliminary data showed no increased formation of reticulin fibers in murine spleens and bone marrows.

In summary, our study demonstrates that THPO overexpression in human BM CD34+ cells is sufficient to induce BM myelofibrosis in a humanized xenotransplantation model (MF ossicles). This novel model, which is the first of its kind, corresponds to one of the established murine MF models and faithfully recapitulates important MF disease hallmarks. The MF ossicle model will therefore not only provide a platform for comprehensive mechanistic studies in human MF but will also serve as a valuable tool for drug testing and identification of novel targets for therapy.