Comparison of Four Different Humanized Hematopoietic Niche Xenotransplantation Methods to Engraft Myelodysplastic Syndromes (MDS)

Introduction: Next generation sequencing techniques have identified a large number of MDS associated acquired molecular lesions. However, translation of these possible molecular targets into new therapeutic strategies has been lagging behind. This is also due to a lack of functional experimental models of MDS, in which new hypotheses can be evaluated pre-clinically. Xenograft models in NSG mice have emerged as versatile preclinical platforms for investigation of functional pathomechanisms in MDS ([1] Medyouf et al., 2014, [2] Rouault-Pierre et al., 2017). The limiting factor of these models is the low engraftment of patient-derived CD34⁺ hematopoietic stem cells (HSCs). Efficient humanized 3D scaffolds in immune-compromised mouse models have been established, enabling to increase engraftment rates of normal and malignant hematopoiesis ([3] Reinisch et al., 2016, [4] Abarrategi et al., 2017). Therefore, we evaluated engraftment ability of IPSS low-risk, int-1 and high-risk-patient samples, in four different 3D scaffolds.

Methods: Currently we transplanted samples from 10 MDS patients in parallel into NSG mice testing the following conditions: A) Intrafemoral co-injection of CD34⁺ HSCs and MSCs according to [1]. Subcutaneous implantation of 3D scaffolds. Gelfoam (B) and Bio-OSS (C) [4], Matrigel ossicles (D) [3] and primary human bone isolated after hip replacement, inserted with Gelfoam, preseeded in vitro with MSCs and mononuclear cells (MNCs) and injected in vivo with CD34⁺ HSCs 8 weeks after implantation (human bone ossicles) (E). Ossicles, bone marrow (BM), peripheral blood and spleens were analyzed 12 weeks after implantation of hematopoietic cells.

Results: Gelfoam and human bone ossicles showed significantly higher hCD45⁺cell numbers compared to intrafemoral injection analyzed by flow cytometry. Engraftment in those two conditions was similarly robust. However, Gelfoam scaffolds showed higher percentual engraftment levels ranging up to 70% as compared to human bone ossicles ranging from 0.2% to 27%. Interestingly, we found systemic engraftment of hCD45⁺cells outside the injected bone fragment in the BM, peripheral blood and spleen solely in mice, which received human bone ossicles. In all other methods, hCD45⁺ cells could only be detected within the ossicles themselves. This result could possibly be explained due to transplantation of MNCs in this condition. That hypothesis was supported by another set of experiments using human bone ossicles (n=10), which showed that colonization of the scaffold was similar when transplanting either CD34⁺ cells + MSCs, MNCs+MSCs or MNCs only but systemic engraftment could only be seen in MNC transplanted mice.

Conclusion: Our data show that hCD45⁺cells and MSCs from MDS BM were able to colonize humanized ossicle scaffolds. Gelfoam and human bone ossicles were the most promising novel methods to improve MDS xenograft models. For systemic engraftment, application of MNCs seems to be necessary.