Myeloablation Triggers Bone Marrow Niche Remodeling Resulting in Transient Collagenopathy and Impaired Platelet Function

Allogeneic hematopoietic stem cell transplantation (HSCT) requires myeloablative conditioning, which is achieved by chemotherapy and total body irradiation (TBI), leading to rapid clearance of all peripheral blood cells. We recently demonstrated that HSCT patients exhibit an additional severe platelet dysfunction. While it is overall well understood how platelets are cleared in response to TBI and then reappear, the underlying modifications within the bone marrow (BM) causing defects in platelet number and function remain elusive. Here, we aimed to decipher BM remodeling and subsequent changes in platelet biology using a mouse model undergoing TBI.

C57BL/6J mice were subjected to 5 Gy TBI and sacrificed within the first week after irradiation for BM immunofluorescence staining and confocal microscopy. We detected sinusoidal vasodilation as early as 6 hours post TBI, peaking after 3 days, accompanied by massive loss in cellularity including a decline in megakaryocytes (MK). BM pericyte processes, surrounding and shielding all sinusoids, retracted into clusters, which was accompanied by markedly reduced collagen IV expression (Col IV) and extensive ectopic release of platelet-like particles (PLPs) into the BM cavity. These observations imply that the blood-BM barrier is disrupted in response to Col IV degradation, as confirmed by the de novo detection of a cleaved collagen peptide, providing evidence for a specific collagen degradation mode. Expression and activity of the BM collagenase matrix metalloproteinase 9 (MMP9) was assessed by gelatin-zymography in wildtype and MMP9-null mice. MMP9 activity increased in wildtypes post TBI, while no gelatinolytic activity was found in BM of mutant mice. We established an in situ zymography assay using unfixed femur sections and dye-quenched (DQ) Col IV, allowing us to map MMP9 activity by detecting the fluorescence signal, which is generated upon cleavage. We measured elevated DQ fluorescence in the intercellular BM space, indicative of increased MMP9 activity, which was largely abrogated in mice lacking MMP9. Intriguingly, mutant mice exhibited less Col IV degradation upon irradiation and only 45% of PLPs within the BM cavity, implying that MMP9-mediated Col IV cleavage causes the blood-BM barrier to lose its proplatelet guiding properties. To test the modulatory effect of collagenolysis, BM-derived MKs were seeded onto collagenase A-degraded gelatin, Col I or Col IV. As expected, proplatelet formation (PPF) was inhibited on untreated Col I, but normalized after collagenase A treatment. In contrast, degradation of Col IV did not have any effect on PPF. Released PLPs of these MK cultures were stimulated with high dose collagen-related peptide (CRP-X_L) and their activation was monitored by CD62P surface expression. Col I-derived PLPs were less reactive than those generated on gelatin or Col IV. Collagenase A-mediated gelatin, Col I or IV cleavage resulted in a 1.1 to 1.4-fold increase of CD62P⁺ PLPs, suggesting increased responsiveness to CRP-X_L. Our results provide unprecedented evidence that collagen subtypes and cleavage products do not only influence platelet production, but also their reactivity. In order to corroborate our findings in vivo, we studied peripheral platelet function after TBI. Mice developed a transient thrombocytopenia and newly formed platelets had an elevated mean platelet volume. Expression of GPIb/V/IX, integrin α_IIbβ₃ and GPVI was reduced to 40% of control levels, implying a highly dysregulated surface receptor profile. Platelet stimulation with thrombin, CRP-X_L or convulxin led to diminished CD62P exposition and absent integrin activation. Light transmission aggregometry with adjusted platelet counts revealed massively impaired aggregation after thrombin stimulation and virtually no aggregation upon activation with collagen or CRP-X_L.

Our data demonstrate for the first time that PPF and platelet function can be modulated not only by the composition of key BM matrix proteins, but also by their degradation and turnover products. We provide several lines of experimental evidence why peripheral platelet counts and function are markedly impaired after TBI. We hypothesize that our conclusions might help to explain hemostatic abnormalities in certain collagenopathies, in which collagen production or assembly are affected, ultimately leading to blood-BM barrier deregulation and aberrant thrombopoiesis.