INTRODUCTION:There are significant limitations with the use of cell lines and primary cells from mice/rats in the study of human bone tissue.  More importantly, current published studies are limited to bulk samples, which are unable to detect underlying cellular heterogeneity. To begin to investigate differences in osteocyte populations, we sought to devise a reliable method of isolation of primary human bone cells using juvenile bone.  Subsequently, the isolated primary cells are harvested for transcriptome analysis.  We used samples from both trabecular and cortical bone, in ambulatory and nonambulatory patients to tease out differences in gene expression for these cell populations and clinical scenarios.  We hypothesize that weightbearing will upregulate genes related to bone formation in trabecular bone more than cortical bone.
METHODS: Deidentified discarded bone samples from pediatric patients (under age 18) were obtained from routine orthopaedic surgeries.  Samples were initially debrided of all periosteal soft tissue.  Trabecular/marrow tissue was separated from cortical bone and both were morselized.  Samples were then rinsed in several washes of saline and weighed, followed by sequential collagenase digests to remove any remaining marrow elements and soft tissue.  Trabecular samples underwent an additional collagenase digest.  Samples were then placed in alternating EDTA and collagenase with saline rinses in between in order to remove mineral and cells.  Cells collected during digestion were kept for single cell transcription analyses. After digest, the bone chips were  pulverized and RNA harvested using a Trizolextraction. Samples were prepared for sequencing using the Illumina Total Stranded RiboZero Gold library prep kit , followed by sequencing on a Novaseq 6000 S4 flowcell (2x150 base pair reads, 20Gb of data/sample).
RESULTS: A total of 20 samples were evaluated.  Five samples (two cortical, three trabecular) had low quality QC metrics, and were not used in analysis.  RNAseq from bulk RNA demonstrated similar expression patterns between trabecular and cortical bone in terms of the effect of loading.  In evaluating expression with p-values <0.01, there were 43genes that overlapped in both cell types that had a difference in expression between weightbearing and nonweightbearing(table 1).  SOST, MEPE, PHEX, DMP1, PDPN, COL1A1 and RUNX2  were all detected, consistent with an osteocytic profile, but showed no differences with weightbearing.  FGF23 demonstrated significantly higher expression in nonambulatory trabecular bone compared to ambulatory bone, but no difference in cortical bone between ambulatory states was detected.  Several genes that have been traditionally associated with brain/nerve cells demonstrated increased expression in bone samples, especially nonambulatory. (table 2) .
DISCUSSION: The study of primary human bone cells has been limited by the availability of cells to study. Herein we have developed a method for the isolation of primary cells from juvenile bone, using both trabecular and cortical samples. Initial characterization of these cells has demonstrated there are population differences looking at bone types, and that weightbearing status is associated with upregulation of some novel factors that have not been described before. Additional studies are ongoing evaluating differences at the single cell level. SIGNIFICANCE/CLINICAL RELEVANCE: Studies of normal developing juvenile bone are needed in order to fully characterize cellular heterogeneity.  Next generation sequencing techniques of primary bone cells will allow for identification of cell subtypes that may be more responsive to weightbearing, and will likely elucidate novel genes that could be leveraged into new strategies for optimizing bone health

2055 - The Effects of Physiologic Loading on Bone: An Analysis of Primary Osteocytes using Next Generation Sequencing