Regulation of microRNA-221, -222, -21 and -27 in articular cartilage subjected to abnormal compressive forces

Abstract

Key points: microRNAs (miRs) are small non-coding molecules that regulate post-transcriptional target gene expression. miRs are involved in regulating cellular activities in response to mechanical loading in all physiological systems, although it is largely unknown whether this response differs with increasing magnitudes of load. miR-221, miR-222, miR-21-5p and miR-27a-5p were significantly increased in ex vivo cartilage explants subjected to increasing load magnitude and in in vivo joint cartilage exposed to abnormal loading. TIMP3 and CPEB3 are putative miR targets in chondrocytes Identification of mechanically regulated miRs that have potential to impact on tissue homeostasis provides a mechanism by which load-induced tissue behaviour is regulated, in both health and pathology, in all physiological systems.

Abstract: MicroRNAs (miRs) are small non-coding molecules that regulate post-transcriptional target gene expression and are involved in mechano-regulation of cellular activities in all physiological systems. It is unknown whether such epigenetic mechanisms are regulated in response to increasing magnitudes of load. The present study investigated mechano-regulation of miRs in articular cartilage subjected to 'physiological' and 'non-physiological' compressive loads in vitro as a model system and validated findings in an in vivo model of abnormal joint loading. Bovine full-depth articular cartilage explants were loaded to 2.5 MPa (physiological) or 7 MPa (non-physiological) (1 Hz, 15 min) and mechanically-regulated miRs identified using next generation sequencing and verified using a quantitative PCR. Downstream targets were verified using miR-specific mimics or inhibitors in conjunction with 3'-UTR luciferase activity assays. A subset of miRs were mechanically-regulated in ex vivo cartilage explants and in vivo joint cartilage. miR-221, miR-222, miR-21-5p and miR-27a-5p were increased and miR-483 levels decreased with increasing load magnitude. Tissue inhibitor of metalloproteinase 3 (TIMP3) and cytoplasmic polyadenylation element binding protein 3 (CPEB3) were identified as putative downstream targets. Our data confirm miR-221 and -222 mechano-regulation and demonstrates novel mechano-regulation of miR-21-5p and miR-27a-5p in ex vivo and in vivo cartilage loading models. TIMP3 and CPEB3 are putative miR targets in chondrocytes. Identification of specific miRs that are regulated by increasing load magnitude, as well as their potential to impact on tissue homeostasis, has direct relevance to other mechano-sensitive physiological systems and provides a mechanism by which load-induced tissue behaviour is regulated, in both health and pathology.

Keywords: CPEB3; TIMP3; articular cartilage; mechanical load; miRNA.

Figure 1. Validation of mechanically‐regulated miRs in cartilage explants using qPCR

qPCR validation of mechanically‐regulated miRs, identified by NGS, in cartilage explants subjected to loads of 2.5 or 7 MPa (1 Hz, 15 min) and analysed 24 h post‐cessation of load for (A) miR‐21‐5p, (B) miR‐27a‐5p, (C) miR‐221, (D) miR‐222, (E) miR‐451, (F) miR‐483 and (G) miR‐453; unloaded explants served as controls. miR levels were normalized to the geometric mean of two reference genes (SDHA, YWHAZ) and further normalized relative to the unloaded control cDNAs. Data are presented as box plots depicting the mean ± 95% CI (n = 6 explants) and are representative of three independent experiments. Statistical analysis was performed using one‐way ANOVA with Tukey's post hoc test.

Figure 2. Validation of mechanically regulated miRNAs in a murine in vivo model of load‐induced joint degeneration

A, toluidine blue staining of a representative mouse knee joint at days 3 and 21 after ACL rupture to induce joint instability/joint degeneration. MTP, medial tibial plateau; MFC, medial femoral condyle; LTP, tibial plateau; LFC, lateral femoral condyle; ACL, anterior cruciate ligament. Yellow indicates inflammatory cell infiltrate. Validation of differential expression of (B) miR‐221, (C) miR‐222, (D) miR‐21‐5p and (E) miR‐27‐5p in articular cartilage after normalization to the geometric mean of the reference genes U6, β‐actin and 18s and further normalization to the uninjured knee cartilage. Data are presented as box plots depicting the mean ± 95% CI (n = 3 animals per experimental time point). Statistical analysis was performed using one‐way ANOVA with Tukey's post hoc test. [Color figure can be viewed at wileyonlinelibrary.com]

Figure 3. Validation of TIMP3 and CPEB3, putative target genes of miR‐221, miR‐222 or miR‐21, using TaqMan qPCR

Primary bovine chondrocytes were treated with either 50 nm miR mimic, 50 nm inhibitor or negative control siRNAs for each respective miR for 48 h, prior to analysis of the effect of overexpression and knockdown of miR‐221 on TIMP3 transcription (A and B), miR‐222 on TIMP3 transcription (C and D), and miR‐21 on TIMP3 (E and F) and CPEB3 (G and H) transcript levels after normalization to the geometric mean of the reference genes HPRT and YWHAZ and further normalization to respective negative control siRNAs. Data are presented as the mean ± 95% CI (n = 3 wells) and are representative of three independent experiments. Statistical analysis was performed using one‐way ANOVA with Tukey's post hoc test.

Figure 4. Verification of 3′‐UTR activation of target mRNAs containing the predicted miR seed sites using a luciferase promoter assay

SW1353 chondrosarcoma cells were co‐transfected with reporter plasmids containing either (A) TIMP3 or (B) CPEB3 3′‐UTRs and 50 nm miR‐221, miR‐222 or miR‐21‐5p mimics, or the negative control siRNA, for 24 h and luciferase levels were determined. Data are presented as the mean ± 95% CI (n = 3 wells) and are representative of three independent experiments. C, TIMP3 mRNA levels, as assessed using qPCR, in cartilage explants subjected to loads of 2.5 or 7 MPa (1 Hz, 15 min) and analysed 24 h post‐cessation of load; unloaded explants served as controls. mRNA levels were normalized to the geometric mean of two reference genes (SDHA, YWHAZ) and further normalized relative to the unloaded control cDNAs. Data are presented as box plots depicting the mean ± 95% CI (n = 6 explants) and are representative of three independent experiments. Statistical analysis was performed using one‐way ANOVA with Tukey's post hoc test.