A new 3-dimensional osteocyte model to investigate mechanical pathological mechanisms in osteoarthritis.. Identifying the Osteocyte Mechanosome

Osteocytes function as critical regulators of bone homeostasis by sensing mechanical signals. The mechanisms underlying abnormal joint mechanics are poorly understood despite it being a major risk factor for developing musculoskeletal diseases such as osteoporosis and osteoarthritis. This study used a human, cell-based physiological, 3D in vitro model of bone to define the osteocyte response to mechanical load. Human Y201 MSC cells were embedded at a density of 0.05 x 106 cell/gel in type I collagen gels in silicone plates and differentiated to osteocytes in osteogenic media at 37°C in 5% CO2/95% air atmosphere for 7 days. One hour prior to loading, media were removed and 800μL osteogenic media (DMEM containing 50µg/mL ascorbate-2-phosphate, 5mM β-glycerophosphate, 1nM dexamethasone) added. An hour later, silicone plates were loaded using a BOSE ElectroForce® 3200 loading instrument (TE Instruments, UK) to stretch the plate causing cyclic compression in all wells (pathophysiological load 4300με induced by 0.7mm displacement, 10Hz, 3000 cycles; n=5). Control gels in the silicone plate were placed into the loading device but received no load (n=4). Loading was controlled using WinTest® Software 4.1 with TuneIQ control optimization (BOSE). RNA was extracted from loaded samples 1 hour post load (n=5) and unloaded controls (n=4). RNA was eluted in 30µl RNAse/DNAse free water and RNA quality, and concentration assessed by TapeStation Analysis (Agilent). An RNA sequencing library was prepared for the mechanically loaded and control samples, using the New England Biolabs Ultra II directional RNA library prep kit (Wales Gene Park). cDNA was synthesized using this RNA which, after undergoing fragmentation, had adaptors ligated to the ends. The MiSeq Nano system (Illumina) was used to complete a sequencing library quality control after which sequencing was performed using the NovaSeq 6000 system (Illumina) running a 2 x 100bp paired-end reads run on a NovaSeq S1 flow cell. Trimming to remove adapter sequencer and poor-quality ends of reads was performed by Trim Galore using default parameters in paired-end mode. Trimmed paired-end reads were aligned to the GRCh38 no_alt_plus_hs38d1 analysis set reference using STAR (v2.5.1b), an ultrafast universal RNAseq aligner, following the 2-pass method. QC metrics were generated using FastQC (v0.11.2), and summary statistics were generated using Samtools (v0.1.19) flagstat. Raw counts were calculated for all samples for both (i) exons and (ii) genes using Subread featureCounts Version 1.5.1. Counts were generated for paired end read fragments summarized at exon level and then aggregated at transcript level. Differentially expressed genes were identified using an DEseq2 analysis on normalised count data. The resultant p-values were corrected for multiple testing and false discovery issues using the FDR method.