Reduced somatosensory innervation alters the skeletal transcriptome at a single cell level in a mouse model of type 2 diabetes

Small peripheral neuropathies (SPN) are a common complication in diabetes, affecting around 50% of the diabetic population. Co-occurrence of diabetic peripheral neuropathy (DPN) and diabetic bone disease has led to the hypothesis that DPN influences bone metabolism, although little experimental evidence has yet supported this premise. To investigate, high fat diet (HFD) feeding in mice was performed followed by phenotyping of skeletal-innervating neurons and bone architectural parameters. Results showed that HFD feeding conditions resulted in a marked decrease in skeletal innervation (69-76% reduction in Beta-III-Tubulin-stained nerves, 50-66% reduction in CGRP-stained nerves in long bone periosteum). These changes in skeletal innervation were associated with alterations in bone mass and in cortical and trabecular bone microarchitecture. Single cell RNA-sequencing of sensory neurons and bone tissue was next utilized to reconstruct potential nerve-to-bone signaling interactions, including implication of sensory nerve derived neurotrophins (Bdnf), neuropeptides (Gal, Calca and Calcb) and other morphogens (Vegfa, Pdgfa, and Angpt2). Moreover, scRNA-Seq identified marked shifts in periosteal cell transcriptional changes within HFD fed conditions, including a reduction in proliferation, osteogenic differentiation potency and reductions in WNT, TGFb and MAPK signaling activity. When isolated, periosteal cells from HFD fed mice showed deficits in proliferative and osteogenic differentiation potential. Indeed, these cellular changes in proliferation and differentiation capacity were restored by treatment of HFD exposed periosteal cells to sensory neuron conditioned medium. In sum, HFD modeling of Type 2 diabetes results in a skeletal polyneuropathy. Moreover, the combination of multi-tissue scRNA-Seq and isolated in vitro studies strengthen the case for altered nerve-to-bone signaling in diabetic bone disease. Overall design: Four left femurs and tibias were dissected from ND and HFD groups at the experimental endpoint for scRNA-Seq. The bones were digested in a pre-warmed dissociation solution (3?mg?ml-1 collagenase I (Worthington) and 4?mg?ml-1 dispase II (Roche) in PBS) for 1.5?h?min at 180?r.p.m. at 37?°C, run through a 40?µm filter (VWR) and rinsed with 3?ml medium (aMEM (Gibco) with 20% FBS, 1% penicillin/streptomycin). Periosteal cells were centrifuged at 300g for 30?min at 4?°C. The cell pellet was washed twice with PBS and spun for 10?min at 300g at 4?°C. Periosteal cells were then resuspended in HBSS at a concentration of ~1,000 cells/µl. Cell viability was assessed with Trypan blue exclusion on a Countess II (ThermoFisher Scientific) automated counter and showed >85% viability. Cells were sent to the JHMI Transcriptomics and Deep Sequencing Core. The library was generated using the 10X Genomics Chromium controller following the manufacturer's protocol