Histone modifications critically contribute to the epigenetic orchestration of bone development - in part by modifying accessibility of genes to transcription factors. Based on the previous finding that histone H2A deubiquitinase 2A-DUB/Mysm1 interacts with the p53-axis in hematopoiesis and tissue development, we here analyzed the molecular and cellular mechanisms of Mysm1-p53 interplay in bone development.

The bone phenotype of 4–5 week-old Mysm1-/- (MKO), Mysm1-/-p53-/- (DKO) and corresponding wildtype (WT) mice was determined using µCT and histology. Primary osteoblasts, mesenchymal stem cells (MSCs) and osteoclasts were isolated from long bones to assess cell proliferation, differentiation, apoptosis and activity. Statistics: one-way ANOVA, p<0.05.

MKO mice displayed an osteopenic bone phenotype compared to WT (BV/TV: 5.7±2.9 vs. 12.5±4.2, TbN: 1.3±0.6 vs. 2.7±0.7 1/mm, respectively), and these effects were abolished in DKO mice (BV/TV: 17.8±2.6, TbN: 3.7±0.4 1/mm). MKO mice compared to WT also showed both in vitro and in vivo disturbed osteoclast formation (in vitro: 1.5±1.2 vs. 9.9±1.8 OcN/mm2, in vivo OcN/BPm: 1.4±1.0 vs. 3.0±0.7 cells/mm, respectively) accompanied by increased apoptosis and DNA damage; additional p53 knockout attenuated these effects (7.8±1.8 OcN/mm2 and OcN/BPm: 2.2±1.0 cells/mm). Primary osteoblasts from both MKO and DKO mice showed decreased expression of the transcription factor Runx2 and of the osteogenic markers. ChIP-Seq analysis revealed direct binding of Mysm1 to Runx2 promoter regions in osteoblasts, implying that Mysm1 here regulates osteogenic differentiation. In contrast, MKO-MSCs differentiation did not differ from WT, but DKO-MSCs displayed a significantly increased expression of Alpl, Bglap and Runx2. The different effects of Mysm1-/- in MSCs and osteoblasts presumably resulted from the lower expression level of Mysm1 in MSCs in comparison to mature osteoblasts.

Thus, our data demonstrate that H2A deubiquitinase Mysm1 is essential for the epigenetic control of bone development via distinct mechanisms: 1) In osteoclasts, Mysm1 is involved in maturation of osteoclast progenitors and osteoclast survival. 2) In osteoblasts, Mysm1 directly controls Runx2 expression, thereby explaining osteopenic phenotype of MKO mice. 3) In MSCs, Mysm1 may play an inferior role due to low expression level. However, loss of p53 increases Runx2 expression during MSC differentiation, leading to normal bone formation in DKO mice.

Confirming clinical evidence, we recently demonstrated in a rodent model that a severe trauma which induces an acute systemic inflammation considerably impairs fracture healing. Interleukin-6 (IL-6) is a key cytokine in posttraumatic inflammation as its serum level correlates with injury severity and mortality. IL-6 signals are transmitted by the transmembrane glycoprotein 130 (gp130) via two distinct mechanisms: firstly, through classic signalling via the membrane-anchored IL-6 receptor and secondly, through trans-signalling using a soluble IL-6 receptor. Whereas IL-6 trans-signalling is considered a danger signal driving inflammation, classic signalling may mediate anti-inflammatory, pro-regenerative processes. The role of the two distinct pathways in bone healing has not yet been elucidated. Here, we studied the function of IL-6 in the pathophysiology of compromised bone healing induced by severe trauma.

Male C57BL/6J mice received an osteotomy of the right femur stabilized with an external fixator. Systemic inflammation was induced by additional blunt chest trauma (TxT) applied immediately after the osteotomy. Mice were injected with either fusion protein sgp130Fc, which selectively inhibits IL-6 trans-signalling, or a neutralizing anti-IL-6 antibody (IL-6 Ab), blocking both signalling pathways. Control mice received vehicle solution. Animals were euthanised 21 days after surgery. Fracture healing was analysed by biomechanical testing, μCT, and histomorphometry (n= 6–9; p=0.05; ANOVA/Fisher LSD post hoc).

Thoracic trauma significantly impaired fracture healing [bending stiffness (EI) −57%, p<0.00]. Treatment with sgp130Fc significantly attenuated bone regeneration as demonstrated by an increased EI (+110%, p<0.00) and a trend of augmented apparent Young”s modulus (+69%, p=0.13) compared to TxT control. Histomorphometric analysis could not detect differences in the amount of bone, confirming µCT results, but revealed a significantly decreased cartilage area after treatment with sgp130Fc (−76%, p=0.01). Inhibition of both signalling pathways with IL-6 Ab, however, did not have any effects.

In conclusion, severe trauma significantly impaired fracture healing, confirming previous studies. Treatment with sgp130Fc ameliorated the negative effects providing evidence that IL-6 trans-signalling triggers the excessive immune response after trauma impairing bone regeneration. Injection of IL-6 Ab did not improve fracture healing thereby implying that classic signalling may rather have beneficial effects.