Mapping the Response of Human Osteocytes in Native Matrix to Mechanical Loading Using RNA Sequencing

Abstract

Osteocytes sense mechanical loads and transduce mechanical signals into a chemical response. They are the most abundant bone cells deeply embedded in mineralized bone matrix, which affects their regulatory activity in the mechanical adaptation of bone. The specific location in the calcified bone matrix hinders studies on osteocytes in the in vivo setting. Recently, we developed a three-dimensional mechanical loading model of human osteocytes in their native matrix, allowing to study osteocyte mechanoresponsive target gene expression in vitro. Here we aimed to identify differentially expressed genes by mapping the response of human primary osteocytes in their native matrix to mechanical loading using RNA sequencing. Human fibular bone was retrieved from 10 donors (age: 32-82 years, 5 female, 5 male). Cortical bone explants (8.0 × 3.0 × 1.5 mm; length × width × height) were either not loaded or mechanically loaded by 2000 or 8000 μɛ for 5 minutes, followed by 0, 6, or 24 hours post-culture without loading. High-quality RNA was isolated, and differential gene expression analysis performed by R2 platform. Real-time PCR was used to confirm differentially expressed genes. Twenty-eight genes were differentially expressed between unloaded and loaded (2000 or 8000 μɛ) bone at 6 hours post-culture, and 19 genes at 24 hours post-culture. Eleven of these genes were related to bone metabolism, ie, EGR1, FAF1, H3F3B, PAN2, RNF213, SAMD4A, and TBC1D24 at 6 hours post-culture, and EGFEM1P, HOXD4, SNORD91B, and SNX9 at 24 hours post-culture. Mechanical loading significantly decreased RNF213 gene expression, which was confirmed by real-time PCR. In conclusion, mechanically loaded osteocytes differentially expressed 47 genes, of which 11 genes were related to bone metabolism. RNF213 might play a role in mechanical adaptation of bone by regulating angiogenesis, which is a prerequisite for successful bone formation. The functional aspects of the differentially expressed genes in bone mechanical adaptation requires future investigation. © 2023 The Authors. JBMR Plus published by Wiley Periodicals LLC on behalf of American Society for Bone and Mineral Research.

Keywords: 3D; HUMAN BONE; MECHANICAL LOADING; NATIVE MATRIX; OSTEOCYTES; RNA SEQUENCING.

Fig. 1

Schematic diagram of the experimental setup. (A) Bone explants were pre‐cultured for 1 or 2 days, and then were either not loaded or mechanically loaded at a magnitude of 2000 or 8000 μɛ for 5 minutes. Immediately after mechanical loading, explants were post‐cultured (without mechanical loading) for 0, 6, or 24 hours. RNA was isolated from osteocytes in bone and gene expression was measures by RNA‐seq and/or real‐time PCR. (B) Osteocytes expressing sclerostin in unloaded bone, and bone loaded at 2000 and 8000 μɛ at 24 hours post‐culture. Arrows: sclerostin‐positive osteocytes. Scale bar = 20 μm. (C) Donor number and indication of RNA‐seq and/or real‐time PCR at 0, 6, and 24 hours post‐culture. Not underlined: RNA‐seq or real‐time PCR only; Underlined: RNA‐seq + real‐time PCR. PC = post‐culture; RNA‐seq = RNA‐sequencing.

Fig. 2

Differential expression analysis between unloaded (0 μɛ), 2000 μɛ, and 8000 μɛ loaded human cortical bone from four donors without post‐culture using RNA‐seq. (A) Heat map: all DEGs. Score: relative gene expression. (B) Volcano plot: DEGs between unloaded (0 μɛ) and 2000 μɛ loaded bone. Red dots: numbers corresponding to genes listed in E. (C) Volcano plot: DEGs between unloaded (0 μɛ) and 8000 μɛ loaded bone. Red dots: numbers corresponding to genes listed in E. (D). Venn diagram: 14 shared DEGs between unloaded (0 μɛ) and 2000 μɛ loaded bone, and between unloaded (0 μɛ) and 8000 μɛ loaded bone. (E) List of 14 upregulated or downregulated shared DEGs between unloaded (0 μɛ) and 2000 μɛ loaded bone, and between unloaded (0 μɛ) and 8000 μɛ loaded bone. n = 4.

Fig. 3

Gene expression of two of the shared 14 DEGs, related to bone metabolism, in osteocytes in unloaded (0 μɛ), 2000 μɛ, and 8000 μɛ loaded human cortical bone without post‐culture. (A) DLK1 gene expression (RNA‐seq; n = 4). (B) ADRA2B gene expression (RNA‐seq; n = 4). (C) ADRA2B gene expression (real‐time PCR; 0 μɛ, n = 5; 2000 μɛ, n = 7; 8000 μɛ, n = 5). **p < 0.01, ****p < 0.0001.

Fig. 4

Differential expression analysis between unloaded (0 μɛ), 2000 μɛ, and 8000 μɛ loaded human cortical bone from five donors with 6 hours post‐culture using RNA‐seq. (A) Heat map: all DEGs. Score: relative gene expression. (B) Volcano plot: DEGs between unloaded (0 μɛ) and 2000 μɛ loaded bone. Red dots: numbers corresponding to genes listed in E. (C) Volcano plot: DEGs between unloaded (0 μɛ) and 8000 μɛ loaded bone. Red dots: numbers corresponding to genes listed in E. (D) Venn diagram: 28 shared DEGs between unloaded (0 μɛ) and 2000 μɛ loaded bone, and between unloaded (0 μɛ) and 8000 μɛ loaded bone. (E) List of 28 upregulated or downregulated shared DEGs between unloaded (0 μɛ) and 2000 μɛ loaded bone, and between unloaded (0 μɛ) and 8000 μɛ loaded bone. *TMEM183B gene expression was upregulated in 2000 μɛ loaded bone, but downregulated in 8000 μɛ loaded bone compared to unloaded bone. n = 4.

Fig. 5

RNA‐seq‐obtained gene expression of seven of shared 28 DEGs, related to bone metabolism, in osteocytes in unloaded (0 μɛ), 2000 μɛ, and 8000 μɛ loaded human cortical bone with 6 hours post‐culture. (A) EGR1 gene expression. (B) FAF1 gene expression. (C) H3F3B gene expression. (D) PAN2 gene expression. (E). RNF213 gene expression. (F) SAMD4A gene expression. (G) TBC1D24 gene expression. n = 4. *p < 0.05, **p < 0.01.

Fig. 6

Real‐time PCR‐obtained gene expression of seven of the shared 28 DEGs, related to bone metabolism, in osteocytes in unloaded (0 μɛ), 2000 μɛ, and 8000 μɛ loaded human cortical bone with 6 hours post‐culture. (A) EGR1 gene expression. (B) FAF1 gene expression. (C) H3F3B gene expression. (D) PAN2 gene expression. (E) RNF213 gene expression. (F) SAMD4A gene expression. (G) TBC1D24 gene expression. 0 μɛ, n = 5; 2000 μɛ, n = 7; 8000 μɛ, n = 7. *p < 0.05.

Fig. 7

Differential expression analysis between unloaded (0 μɛ), 2000 μɛ, and 8000 μɛ loaded human cortical bone from five donors with 24 hours post‐culture using RNA‐seq. (A) Heat map: all DEGs. Score: relative gene expression. (B) Volcano plot: DEGs between unloaded (0 μɛ) and 2000 μɛ loaded bone. Red dots: numbers corresponding to genes listed in E. (C) Volcano plot: DEGs between unloaded (0 μɛ) and 8000 μɛ loaded bone. Red dots: numbers corresponding to genes listed in E. (D) Venn diagram: 19 shared DEGs between unloaded (0 μɛ) and 2000 μɛ loaded bone, and between unloaded (0 μɛ) and 8000 μɛ loaded bone. (E) List of 19 upregulated or downregulated shared DEGs between unloaded (0 μɛ) and 2000 μɛ loaded bone, and between unloaded (0 μɛ) and 8000 μɛ loaded bone. n = 4.

Fig. 8

Gene expression of four of 19 DEGs, related to bone metabolism, in osteocytes in unloaded (0 μɛ), 2000 μɛ, and 8000 μɛ loaded human cortical bone with 24 hours post‐culture. (A) EGFEM1P gene expression (RNA‐seq; n = 4). (B) HOXD4 gene expression (RNA‐seq; n = 4). (C) SNORD91B gene expression (RNA‐seq; n = 4). (D) SNX9 gene expression (RNA‐seq; n = 4). (E) EGFEM1P gene expression (real‐time PCR; 0 μɛ, n = 5; 2000 μɛ, n = 6; 8000 μɛ, n = 7). (F) HOXD4 gene expression (real‐time PCR; 0 μɛ, n = 5; 2000 μɛ, n = 6; 8000 μɛ, n = 7). (G) SNORD91B gene expression (real‐time PCR; 0 μɛ, n = 5; 2000 μɛ, n = 6; 8000 μɛ, n = 7). (H) SNX9 gene expression (real‐time PCR; 0 μɛ, n = 5; 2000 μɛ, n = 6; 8000 μɛ, n = 7). *p < 0.05, **p < 0.01, ***p < 0.005.

Fig. 9

RNA‐seq‐obtained gene expression of DEGs in osteocytes in their native bone matrix in response to mechanical loading at 2000 and 8000 μɛ at different post‐culture times (0, 6, and 24 hours). (A) FAF1 gene expression. (B) PAN2 gene expression. (C) SAMD4A gene expression. (D) RNF213 gene expression. (E) H3F3B gene expression. (F) SNX9 gene expression. (G) TBC1D24 gene expression. 0 μɛ, n = 6; 2000 μɛ, n = 4; 8000 μɛ, n = 4. *p < 0.05, **p < 0.01.