Mixed Bone Marrow Chimerism after Murine Stem Cell Transplantation Results in Distinct Platelet Subpopulations with Opposing Responsiveness

Megakaryocytes (MKs) produce platelets by transendothelial proplatelet formation (PPF), but detailed mechanisms regulating PPF in vivo remain elusive. Under homeostatic conditions, PPF is exclusively directed into the vessel lumen and is controlled by MK polarization toward sinusoids and by localized expression of extracellular matrix (ECM) proteins at the vessel wall. CXCL12-abundant reticular (CAR) cells are possible new regulators of PPF, as MKs form pores preferably through SECs at CAR cell-free sites at the vessel wall. Total body irradiation (TBI) before hematopoietic stem cell transplantation (HSCT) is an example for therapeutic use of myeloablation. HSCT-related complications include prolonged thrombocytopenia and an increased bleeding risk due to poor MK engraftment. Thromboembolic events represent additional causes for morbidity and mortality in patients. The events leading to delayed MK engraftment and the role of platelets in HSCT-associated thromboembolism are yet unknown. We aimed to decipher the regulation of PPF in the native BM and after HSCT.

We studied the contribution of CAR cells to PPF in an in vitro CAR cell/MK co-culture system, using ST-2 and MS-5 as murine reticular cell lines. MK differentiation and PPF were assessed by FACS and confocal laser scanning microscopy (CLSM). Our results suggest enhanced MK differentiation indicated by an increased number of CD41⁺/CD42a⁺ cells upon CAR cell co-culture, while PPF was reduced, indicating that CAR cells act as inhibitors of PPF. Sublethal (5Gy) TBI led to depletion of MKs/platelets and to massive vasodilation already 6h post TBI. This was accompanied by retraction and clustering of CAR cells and an increased number of platelet-like particles in the BM cavity, implying that altered CAR cell morphology resulted in ectopic platelet release.

We studied ECM proteins after TBI by CLSM and detected sinusoid-specific degradation of collagen IV and laminin α5. Matrix degradation was linked to an increased matrix metalloproteinase (MMP) 9 activity using complementary zymography techniques, and was confirmed by immunoblotting of a cleaved collagen peptide. We used conditional MMP9^-/- and vessel-specific laminin α5-deficient mice for TBI and antibody-mediated platelet depletion and found no impact of MMP9 or laminin α5 loss on PPF or MK numbers and size. Vessel recovery was markedly delayed in MMP9-null mice, suggesting that MMP9 and laminin α5 are dispensable for thrombopoiesis, but MMP9 plays an essential role in vascular remodeling after TBI.

We next established a murine HSCT model, using mice that ubiquitously express dsRed in all cells as BM donors for lethally (10Gy) irradiated C57B6/J mice. Donor-derived dsRed⁺ cells were readily detectable in the recipient BM already one week after HSCT and were unexpectedly found within clusters close to the endosteum. MKs and Ly6G⁺ cells also displayed a clustered distribution in the recipient BM, which we did not observe for other blood cell lineages. This implies a similar engraftment pattern for MKs and granulocytes. We used a multiplex chemokine assay to measure chemokine levels and found peak expression of CXCL1, CXCL9 and CCL7 at several distinct time points after HSCT, suggesting a selective and consecutive role of these chemokines for MK engraftment and cluster formation.

Platelet count dropped massively after TBI, in conjunction with an increased mean platelet volume (MPV). Surface receptor expression levels were decreased by approx. 50%. Concomitantly, platelet activity was severely impaired upon thrombin stimulation and virtually absent after GPVI activation with CRP-X_L or convulxin. Platelet aggregation and in vitro thrombus formation were markedly impaired. When we differentiated between donor- and recipient-derived platelet subpopulations, we found dsRed⁺ platelets to be hyperreactive, while dsRed^- platelets were less responsive, even when compared to wildtype controls. These distinct reactivity levels were masked when we analyzed the whole platelet population.

In summary, we provide first evidence for a specific MK engraftment pattern after HSCT, which is associated with a mixed platelet chimerism comprising opposing responsiveness. Our data will help to understand platelet-related thrombotic and bleeding complications after HSCT.

No relevant conflicts of interest to declare.