The long non-coding RNA ROCR contributes to SOX9 expression and chondrogenic differentiation of human mesenchymal stem cells

Abstract

Long non-coding RNAs (lncRNAs) are expressed in a highly tissue-specific manner and function in various aspects of cell biology, often as key regulators of gene expression. In this study, we established a role for lncRNAs in chondrocyte differentiation. Using RNA sequencing we identified a human articular chondrocyte repertoire of lncRNAs from normal hip cartilage donated by neck of femur fracture patients. Of particular interest are lncRNAs upstream of the master chondrocyte transcription factor SOX9 locus. SOX9 is an HMG-box transcription factor that plays an essential role in chondrocyte development by directing the expression of chondrocyte-specific genes. Two of these lncRNAs are upregulated during chondrogenic differentiation of mesenchymal stem cells (MSCs). Depletion of one of these lncRNAs, LOC102723505, which we termed ROCR (regulator of chondrogenesis RNA), by RNA interference disrupted MSC chondrogenesis, concomitant with reduced cartilage-specific gene expression and incomplete matrix component production, indicating an important role in chondrocyte biology. Specifically, SOX9 induction was significantly ablated in the absence of ROCR, and overexpression of SOX9 rescued the differentiation of MSCs into chondrocytes. Our work sheds further light on chondrocyte-specific SOX9 expression and highlights a novel method of chondrocyte gene regulation involving a lncRNA.

Keywords: Cartilage; Chondrogenesis; Differentiation; Epigenetics; LncRNA; MSC; SOX9.

Fig. 1.

Expression of lncRNAs from the SOX9 locus. (A) UCSC genome browser schematic of cartilage RNA-Seq reads aligned to the human genome with RefSeq gene annotations, Ensembl gene predictions and active H3K27Ac chromatin marks. Reads are pooled from six neck of femur (NOF) fracture cartilage donors. Primer locations are indicated by red arrowheads. (B) Expression of SOX9 locus lncRNAs and SOX9 in RNA extracted from OA cartilage measured by real-time RT-PCR normalised to 18S. Values are mean±s.e.m. of data pooled from five separate donors. (C) Subcellular localisation of SOX9-AS1 and ROCR in comparison with small nuclear RNA U2 and SOX9 mRNA pooled from two OA HAC donors. Values are mean±s.e.m. of ΔCT between an equal fraction of nuclear and cytoplasmic RNA. (D) Expression of SOX9-AS1 and ROCR in RNA extracted from OA synovium and joint fat pad. Values are mean±s.e.m. of data pooled from eight separate synovium and fat pad donors. (E) Expression of SOX9-AS1 and ROCR in an RNA tissue panel. Values are the technical mean of data pooled from three donors per tissue.

Fig. 2.

Expression of lncRNAs during MSC differentiation. (A) Expression of the indicated genes in RNA extracted from MSCs undergoing chondrogenic differentiation at the indicated time points between Day 0 and Day 7. (B-D) Expression of SOX9-AS1 (B), ROCR (C) and LINC01152 (D) during MSC chondrogenic differentiation. (E) Expression of the indicated genes in RNA extracted from MSCs undergoing osteoblastogenic and adipogenic differentiation. (F-H) Expression of SOX9 (F), SOX9-AS1 (G) and ROCR (H) during MSC osteoblastogenic (Osteo.) and adipogenic (Adipo.) differentiation. Values are mean± s.e.m. of data pooled from three or four MSC donors.

Fig. 3.

Effect of lncRNA depletion on MSC chondrogenic differentiation. (A-D) MSCs were transfected for 2 days with SOX9-AS1 or ROCR-targeting or non-targeting control (siCon) siRNA prior to chondrogenic differentiation in hanging transwell inserts. (A) SOX9-AS1 and ROCR expression in RNA extracted from Day 3 and Day 7 chondrogenic discs. Expression is presented as a percentage of non-targeting control levels. (B) Representative Day 3 and 7 chondrogenic discs. (C) Wet mass of Day 7 chondrogenic discs. (D) Representative Safranin O staining of Day 7 chondrogenic discs. (E-G) MSCs were transfected for 1 day with ROCR-targeting or non-targeting control siRNA prior to chondrogenic differentiation in a V-bottom 96-well plate. (E) Representative Day 7 chondrogenic pellets. (F) GAG levels assayed by DMB assay in Day 7 chondrogenic pellets. (G) DNA quantification by PicoGreen assay in Day 7 chondrogenic pellets. Values are mean±s.e.m. of data pooled from three (A-D) or four (E-G) MSC donors. *P<0.05; ***P<0.001 for lncRNA siRNA versus non-targeting siRNA. Significant differences between sample groups were assessed by one-way analysis of variance followed by the Bonferroni post-hoc test for multiple comparisons or a two-tailed Student's t-test was performed for single comparisons.

Fig. 4.

Effect of lncRNA depletion on MSC chondrogenic gene expression. (A) MSCs were transfected for 2 days with SOX9-AS1 or ROCR-targeting or non-targeting control (siCon) siRNA prior to chondrogenic differentiation in hanging transwell inserts. RNA was extracted and expression of the indicated genes measured by real-time RT-PCR. (B-D) MSCs were transfected for 1 day with SOX9-AS1 or ROCR-targeting or non-targeting control siRNA prior to chondrogenic differentiation in a V-bottom 96-well plate for up to 24 h. RNA and protein was extracted and expression of SOX9 mRNA (B) or protein (C) measured by real-time RT-PCR or immunoblotting, respectively. (D) Expression of L-SOX5a and SOX6. Values are mean±s.e.m. of data pooled from three (A) or four (B-D) MSC donors. *P<0.05; **P<0.01; ***P<0.001 for lncRNA siRNA versus non-targeting siRNA. Significant differences between sample groups were assessed by one-way analysis of variance followed by the Bonferroni post-hoc test for multiple comparisons or a two-tailed Student's t-test was performed for single comparisons.

Fig. 5.

Effect of lncRNA depletion on MSC osteoblastogenic and adipogenic differentiation. (A-C) MSCs were transfected for 2 days with SOX9-AS1 or ROCR-targeting or non-targeting control (siCon) siRNA prior to osteoblastogenic differentiation. (A) Representative matrix mineralisation assayed by Alizarin Red staining after 21 days. (B) Quantification of Alizarin Red staining. (C) RNA was extracted and expression of the indicated genes at Day 7 measured by real-time RT-PCR. (D-F) MSCs were transfected for 2 days with SOX9-AS1 or ROCR-targeting or non-targeting control siRNA prior to adipogenic differentiation. (D) Representative fat droplet generation assayed by Oil Red O staining after 14 days. (E) Quantification of Oil Red O staining. (F) RNA was extracted and expression of the indicated genes at Day 7 measured by real-time RT-PCR. Values are mean±s.e.m. of data pooled from four MSC donors. *P<0.05; ***P<0.001 for lncRNA siRNA versus non-targeting siRNA. Significant differences between sample groups were assessed by one-way analysis of variance followed by the Bonferroni post-hoc test for multiple comparisons or a two-tailed Student's t-test was performed for single comparisons.

Fig. 6.

Effect of SOX9 overexpression in combination with ROCR depletion on MSC chondrogenic differentiation and gene expression. MSCs were transduced with SOX9 or control lentivirus (pCDH) for 1 day then transfected for 1 day with ROCR-targeting or non-targeting control (siCon) siRNA prior to chondrogenic differentiation in a V-bottom 96-well plate. (A) Expression of SOX9 in only non-targeting control siRNA pellets at Day 3. (B) GAG levels assayed by DMB assay in only non-targeting control siRNA pellets at Day 7. (C) Expression of SOX9 at Day 3. (D) GAG levels assayed by DMB assay at Day 7. (E,F) Expression of the indicated genes at Day 3. (C-F) Expression is presented as a percentage of non-targeting control levels for cells transduced with each virus. Values are mean±s.e.m. of data pooled from four MSC donors. *P<0.05; **P<0.01; ***P<0.001 for lncRNA siRNA versus non-targeting siRNA, or for SOX9 versus control lentivirus where indicated above the chart. Significant differences between sample groups were assessed by one-way analysis of variance followed by the Bonferroni post-hoc test for multiple comparisons or a two-tailed Student's t-test was performed for single comparisons.