Impaired chondrocyte U3 snoRNA expression in osteoarthritis impacts the chondrocyte protein translation apparatus

Although pathways controlling ribosome activity have been described to regulate chondrocyte homeostasis in osteoarthritis, ribosome biogenesis in osteoarthritis is unexplored. We hypothesized that U3 snoRNA, a non-coding RNA involved in ribosomal RNA maturation, is critical for chondrocyte protein translation capacity in osteoarthritis. U3 snoRNA was one of a number of snoRNAs with decreased expression in osteoarthritic cartilage and osteoarthritic chondrocytes. OA synovial fluid impacted U3 snoRNA expression by affecting U3 snoRNA gene promoter activity, while BMP7 was able to increase its expression. Altering U3 snoRNA expression resulted in changes in chondrocyte phenotype. Interference with U3 snoRNA expression led to reduction of rRNA levels and translational capacity, whilst induced expression of U3 snoRNA was accompanied by increased 18S and 28S rRNA levels and elevated protein translation. Whole proteome analysis revealed a global impact of reduced U3 snoRNA expression on protein translational processes and inflammatory pathways. For the first time we demonstrate implications of a snoRNA in osteoarthritis chondrocyte biology and investigated its role in the chondrocyte differentiation status, rRNA levels and protein translational capacity.

DMM. Mice (C57BL/6) were group housed in individually ventilated cages at a 12 h light/dark cycle, with ad libitum access to food and water. Under anesthesia, a 3 mm skin incision was made over the medial aspect of the patellar ligament through the joint capsule into the femorotibial joint of the left knee. The medial meniscotibial ligament was transected to destabilize the cranial pole of the medial meniscus from the anterior tibial plateau. In sham-operated mice (n = 3) the medial meniscotibial ligament was visualized but not transected. Mice were sacrificed 8 weeks post-surgery and stifles were fixated in phosphate-buffered 3.7% formalin. Ethical review was conducted by the University of Liverpool Animal Welfare and Ethical Review Body and ethical approval was obtained from the University of Liverpool (project license P74DC0667). All experimental protocols were performed in compliance with the UK Animals (Scientific Procedures) Act 1986 regulations. www.nature.com/scientificreports/ Knockdown and ectopic expression of U3 snoRNA. Transfection of non-OA HACs or SW1353 cells (seeded at 30,000 or 20,000 cells/cm 2 , respectively) with 100 nM of U3 snoRNA antisense oligonucleotide (ASO) or a scrambled version thereof (SCR) was performed using HiPerfect (Qiagen) according to the manufacturers' protocol. Non-OA HACs were transfected after 8 h of prior serum starvation, while SW1353 cells were transfected without prior serum starvation. Sequences of U3 snoRNA ASO and SCR (Eurogentec, Liège, Belgium) are shown in Supplementary Table 2. Transfection of non-OA HACs with a U3 snoRNA mini-gene was performed using Fugene6, following the manufacturers' protocol (Promega, Madison, Wisconsin, USA). The U3 snoRNA mini-gene was synthesized (Genecust, Boynes, France) and cloned into the pUC57 vector. The mini-gene consisted of the endogenous 500 nucleotide sequence upstream of the U3 snoRNA transcription start site, the pre-U3 snoRNA sequence and followed by 250 nucleotides downstream of the transcription termination sequence. The U3 mini-gene construct was transfected at 10 ng plasmid/cm 2 and because of the low amount of U3 snoRNA mini-gene supplemented with pGluc-basic-2-CMV (NEB, Ipswich, Massachusetts, USA) as carrier plasmid. In control conditions, only carrier pGluc-basic-2-CMV was transfected at equal total plasmid content as compared to the U3 mini-gene transfection.
U3 snoRNA promoter-reporter assay. Gene expression. Cells were rinsed with 0.9% NaCl and cell lysis, RNA isolation and subsequent cDNA synthesis were performed using the Cells-to-Ct kit (ThermoFisher Scientific, Waltham, Massachusetts, USA). Alternatively, cells were disrupted with TRIzol reagent (Life Technologies), followed by RNA isolation and quantification and cDNA synthesis as described previously 20 . The real-time quantitative polymerase chain reaction was performed as described previously 5 and using validated primer sequences described in Supplementary  Table 3. Ct-values were analyzed with the standard curve method and RNA (including non-coding-and messenger-RNA) expression was normalized to cyclophilin expression or the relative DNA-content between treatments. DNA-content for normalization was measured in parallel wells receiving the identical treatment. DNA-content was measured using a DAPI and HOECHST staining further explained in "SUNsET-assay". Differential gene expression was determined as a fold change compared to control conditions.
In-solution tryptic digestion and mass spectrometry proteomics. Cell lysates following U3 snoRNA knockdown (or control) in SW1353 cells were used 21 . Cells were harvested from triplicate wells of a six well plate with parallel wells used for confirmation of U3 snoRNA knockdown. In-solution tryptic digestion on 10 μg protein was undertaken as previously described 22 . For the secretomes in-solution tryptic digestion was carried out on 10 µl of StrataClean resins (Agilent Genomics, 400,714) on 100 µg of protein for each sample as described previously 23 . Liquid chromatography-tandem mass spectrometry (LC-MS/MS) analysis was performed on trypsin digests using an Ultimate 3,000 Nanosystem (Dionex, ThermoFisher Scientific) online to a Q-Exactive Quadrupole-Orbitrap instrument (Thermo Scientific) 24 . Proteins were identified using an in-house Mascot server (Matrix Science, London, UK). Search parameters used were as follows: enzyme; trypsin, peptide mass tolerances 10 ppm, fragment mass tolerance of 0.01 Da, 1+, 2+, and 3+ ions, with carbamidomethyl cysteine as a fixed modification and methionine oxidation as a variable modification, searching against the UniHuman Reviewed database, with a false discovery rate (FDR) of 1%, a minimum of two unique peptides per protein.
Label-free quantification of mass spectrometry proteomics data. Progenesis QI software (version 4, Waters, Manchester, UK) was used to identify fold changes in protein abundance between U3 snoRNA knockdown and scrambled ASO control condition 25 . Only unique peptides were used for quantification, and with p-values < 0.05, were considered to be differentially expressed (DE). Proteomic data has been deposited in the PRIDE ProteomeXchange and can be accessed using the identifier PXD017253 26 .
Ingenuity pathway analysis. Functional analysis of differentially expressed proteins following U3 knockdown in SW1353 cells was undertaken to evaluate the differences in protein abundance. Ingenuity pathway analysis has been performed as previously described 27 .
Protein translation capacity. For a puromycilation assay 28 www.nature.com/scientificreports/ (Sigma-Aldrich, Dorset, UK). After washing with PBS-T, wells were incubated for 1 h at room temperature with goat anti-mouse Alexa488 (ThermoFisher Scientific). Following a final wash step, the fluorescence signal intensity was determined using a Tristar LB942 (Berthold, Bad Wildbad, Germany) equipped with excitation filter F485 and emission filter F353. Fluorescence data were normalized to DNA-content 30 . For that, the same wells were subsequently washed with HEPES-Buffered Saline (HBS), followed by 1 h incubation with 5 µg/mL DAPI (Invitrogen) plus 5 µg/mL HOECHST 33342 (Invitrogen) in HBS. After washing steps with HBS, fluorescence signal intensity was determined using a Tristar LB942 (Berthold), using the excitation filter F355 and emission filter F460.
RNA electrophoresis and northern blotting. Two micrograms total RNA was separated on a 6% PAAgel (6% PAA (Acrylamide/Bis-acrylamide 30:1); 1 × TBE (Tris-borate-EDTA; Merck, Darmstadt, Germany); 8 M urea), electro-transferred to a Hybond-N membrane (GE Healthcare; Chicago, Illinois, USA) and immobilized by UV-crosslinking (CL-1000, UVP). Membrane was pre-hybridized for 1 h at 62 °C under agitation in hybridization buffer (2 × SSC, 5 × Denhardts solution, 200 μg/ml fish sperm DNA, 0.1% SDS). Membrane was subsequently hybridized overnight with the U3 snoRNA probe (5 nM; Supplementary Table 4) at 62 °C under agitation. Non-hybridized probe was rinsed from the membrane using a series of SSC-SDS wash buffers. Membrane was then incubated in blocking buffer [1 × PBS, 0.5% SDS, 0.1% I-Block (ThermoFisher Scientific)] for 30 min, followed by 30 min incubation in blocking buffer containing 0.2 μg/ml Streptavidin-AP (Life Technologies). After washing with PBS the membrane was equilibrated with detection buffer (100 mM Tris-HCl pH 9.5; 100 mM NaCl; 1 mM MgCl 2 ) and detection was carried out using CDP-Star (Sigma-Aldrich, Dorset, UK) and visualized/quantified using a Bio-Rad Chemidoc MP imaging system. U6 snRNA was detected as a reference RNA on the same membrane by reprobing and repeating the procedure in-full while using the U6 snRNA probe. Full length images of northern blots are shown in Supplementary Fig. 4. The U3 snoRNA signal was normalized by the signal for U6 snRNA and relative differences between conditions were calculated.
Statistical analyses. Statistical significance has been calculated using GraphPad Prism software version 5.0 (La Jolla, California, USA) using paired or unpaired (depending on experiment) Student's t-test . Depending on the experiment this was done 1-tailed or 2-tailed. Details per experiment are indicated in the corresponding figure legends. Conditions were compared to control and statistical significance was set at P < 0.05. To test for normal distribution of the input data, D' Agostino-Pearson omnibus normality tests were performed. All quantitative data sets presented here passed the normality tests. Bars in graphs represent mean ± standard deviation (SD).

Results
U3 snoRNA expression is decreased in cartilage and chondrocytes as function of OA. In order to find clues whether ribosome biogenesis is disturbed in osteoarthritic cartilage at the snoRNA level, a previously performed miRNA microarray experiment on young non-OA cartilage and old OA cartilage 19 , on which probe sets for snoRNAs were also present, was reanalyzed ( Supplementary Fig. 1). A number of snoRNAs was found to be differentially expressed between non-OA and OA cartilage and amongst these U3 snoRNA expression was significantly lower in OA cartilage (Fig. 1A). Since U3 is one of the most abundant snoRNAs playing a key role in ribosome biogenesis, we investigated its involvement in chondrocyte homeostasis in more detail. To verify whether isolated OA chondrocytes have decreased U3 snoRNA levels, we measured U3 snoRNA expression in chondrocytes isolated from young non-OA and from old OA cartilage. Expression of U3 snoRNA in chondrocytes from OA cartilage was reduced ( Fig. 1B) compared to non-OA chondrocytes. Reduced U3 snoRNA expression in OA chondrocytes was accompanied with a distinctive cellular phenotype associated with (pre-)hypertrophic OA chondrocytes 5,31 , with significantly reduced mRNA expression of chondrogenic genes COL2A1; ACAN; SOX9 and NKX3-2 and increased expression of chondrocyte hypertrophy-associated genes COL10A1; RUNX2; ALPL; MMP13; ADAMTS5; COX2 and IL6. Since there was an age difference between the non-OA cartilage/chondrocytes and the OA cartilage/chondrocytes used to determine differences in U3 snoRNA expression (Fig. 1A/B), we next measured OA-dependent alterations in chondrocyte U3 snoRNA levels in an experimental mouse model for traumatic OA (destabilization of medial meniscus; DMM 32 ). Eight weeks post-DMM surgery knees were processed for U3 snoRNA in situ hybridization (ISH). A general reduction of U3 snoRNA ISH signal was observed in articular chondrocytes within the weight-bearing area of the joint in DMM-operated compared to sham-operated knees ( Fig. 1C; red arrows). This was also observed in the menisci ( Fig. 1C; blue arrowheads). U3 snoRNA levels were unaltered in chondrocytes in non-weight-bearing areas of the knee joints in sham-operated versus DMM-operated-mice ( Fig. 1C; black arrows). Next, we asked whether OA-like conditions are capable of reducing U3 snoRNA expression levels in chondrocytes. To this end, non-OA human articular chondrocytes (HACs) from different donors were exposed to 20% (v/v) OA synovial fluid. U3 snoRNA expression was significantly diminished in HACs exposed to OA synovial fluid (Fig. 1D). Exposing non-OA HACs to non-OA synovial fluid caused a significant increase in the expression of U3 snoRNA (Fig. 1E).
A U3 snoRNA gene promoter-reporter assay in SW1353 cells (Fig. 1F) or in HACs (Fig. 1G) revealed that OA synovial fluid reduced U3 snoRNA promoter transcriptional activity in SW1353 cells and two out of three HAC donors. In addition we observed that U3 snoRNA promotor transcriptional activity is reduced in HACs by exposure to the two katabolic cytokines IL1β (Fig. 1H) or TNFα (Fig. 1I) www.nature.com/scientificreports/ scriptomic level. Therefore we altered U3 snoRNA expression levels in non-OA primary chondrocytes using a U3 snoRNA-specific antisense oligonucleotide (ASO). Following ASO transfections in five individual non-OA chondrocyte cultures, reduced U3 snoRNA expression was confirmed ( Fig. 2A). As a result, expression of COL2A1, SOX9, NKX3-2, RUNX2, MMP13 and IL6 was significantly reduced in all five chondrocyte isolates ( Fig. 2B-G), but with variable levels of change. Reciprocally, we ectopically increased expression of U3 snoRNA moderately by transfecting primary chondrocytes with a U3 mini-gene (Fig. 3). The activity of the U3 mini-gene was confirmed by northern blot (Fig. 3A) and significantly elevated U3 snoRNA expression was confirmed in primary chondrocytes (Fig. 3B). The elevated U3 snoRNA expression levels resulted in upregulated levels of mRNAs coding for COL2A1 and NKX3-2 ( snoRNA is rate-limiting in the generation of mature rRNAs 12,33 and 18S rRNA in particular. To determine whether rRNA levels are altered in primary chondrocytes in OA conditions, we measured 18S, 5.8S and 28S rRNA expression. In concert with reduced U3 snoRNA levels in old OA chondrocytes (Fig. 1A/B), we observed reduced expression levels of 18S and 5.8S rRNAs in old OA chondrocytes, while 28S rRNA levels remained unaltered (Fig. 4A). Following the exposure of HACs to OA synovial fluid we observed a reduction of 18S and 5.8S transcript levels (Fig. 4B), while 28S rRNA expression was not significantly altered. OA synovial fluid also reduced expression of 18S and 5.8S rRNAs in HACs (Fig. 4B), while an opposite response of 18S and 5.8S rRNA levels was detected when HACs were exposed to non-OA synovial fluid (Fig. 4B). In both conditions 28S rRNA expression was not significantly changed. Alteration of U3 snoRNA expression in primary chondrocytes (Figs. 2 and 3) led to changes in chondrocyte rRNA levels. Reduction of U3 led to reduced 18S and 28S rRNA levels, with a sharp decrease of 5.8S rRNA levels in four out of five chondrocyte donors (Fig. 4C). Ectopic U3 expression led to increased 18S and 28S rRNA levels (Fig. 4D), while not changing 5.8S rRNA levels in three out of four chondrocyte donors. Overall, data demonstrate that 18S and 5.8S rRNA levels in chondrocytes are susceptible to OA conditions and rRNA transcript levels respond to U3 snoRNA expression.
U3 snoRNA levels influence the activity of the chondrocyte protein translation apparatus. Our data demonstrate alterations in chondrocyte rRNA levels in OA conditions and following knockdown or ectopic expression of U3 snoRNA. Since rRNAs are the catalytically active subunits of the translating ribosome, we next investigated the consequences for overall protein translational capacity. Protein translational capacity was determined in chondrocytes from young non-OA and old OA cartilage and a significantly reduced translational capacity was observed in osteoarthritic chondrocytes (Fig. 5A). Translational capacity of chondrocyte cultures with diminished U3 snoRNA expression was significantly reduced (Fig. 5B), while ectopic induction of U3 snoRNA expression resulted in a mild, but significantly increased translational capacity (Fig. 5C). These data show that U3 snoRNA is able to change the chondrocyte's translational capacity. To determine the full impact of U3 snoRNA knockdown and the concomitant reduction in rRNA levels on the chondrocyte translational apparatus, we subsequently conducted whole proteome analysis on SW1353 cells in which U3 snoRNA expression was reduced. A reduction of U3 snoRNA and concomitant reduction in 18S and 5.8S rRNA levels IL6 mRNA levels were determined by RT-qPCR analysis. Data from U3 ASO samples were calculated relative to controls transfected with the SCR ASO. Results were normalized to relative total DNA content in parallel wells. Statistical significance was determined using 2-tailed paired Student's t-tests. Bars show the mean (± SD). *P < 0.05, **P < 0.01, ***P < 0.001 versus control conditions.
were confirmed ( Supplementary Fig. 2). Ingenuity Pathway Analysis (IPA) analysis of the differential proteomes revealed that IPA networks "synthesis of protein", "translation of protein", "metabolism of protein" and "initiation of translation of mRNA" were top deregulated networks (Fig. 5D) and confirm a global impact of the reduction of U3 snoRNA expression on protein translational processes. Also networks involving transcription and expression of RNA were prominently represented in the IPA analysis. The differential proteome that is observed following U3 snoRNA reduction concerns a particular global upregulation of differentially expressed proteins ( Supplementary Fig. 3), rather than a downregulation and contains a great number of protein species involved in the process of protein translation and ribosomal stress responses. An additional proteomic analysis of the secretome of these U3 snoRNA knockdown cells was conducted. Relevant to inflammatory responses in OA cartilage, we could demonstrate that secreted proteins involved in antagonizing inflammatory processes were present in the secretomes of U3 knockdown cells in lower abundance (Fig. 5E). Together these data show alterations in chondrocyte protein translational activity following alterations in U3 snoRNA expression levels, with U3 snoRNA impacting protein translation molecular routes.

U3 snoRNA expression is increased by BMP7.
We previously demonstrated that bone morphogenetic protein 7 (BMP7) is capable of ameliorating the OA chondrocyte phenotype 5 . Considering this anabolic action of BMP7 we evaluated whether U3 snoRNA and its subsequent rRNA targets can be induced in chondrocytes by BMP7. A U3 snoRNA promoter-reporter experiment in SW1353 cells or in non-OA HACs demonstrated that U3 snoRNA promoter transcriptional activity was induced following stimulation with BMP7 ( Fig. 6A/B). In both chondrocyte cell models BMP7 induced U3 gene expression and increased expression of 18S and 5.8S rRNAs (Fig. 6C/D) www.nature.com/scientificreports/ BMP7 had an increased translational capacity (Fig. 6E), supporting the notion that the BMP7-mediated induction of U3 snoRNA and rRNA expression increases the chondrocyte's translational capacity.

Discussion
U3 snoRNA is an abundant snoRNA with a pivotal function in the generation of mature 18S rRNA in eukaryotes 14 . It's snoRNP biochemistry 34 and mechanism-of-action 35 have been extensively studied in the past decades. However, it was only recently described to be univocal involved in 18S rRNA generation in human cells 12 . In contrast with its decade-long thorough molecular characterization, insight into its involvement in human disease is limited to a role in cell proliferation and tumorigenesis 12,36 , while it remained unknown whether U3 snoRNA levels can be regulated via extracellular cues. We demonstrate that in different osteoarthritic conditions the expression level of U3 snoRNA in articular chondrocytes is diminished. Reduced expression of U3 snoRNA was previously reported in serum of ageing mice 37 , and as a result of nucleus pulposus ageing in mice 38 . As a general study limitation in this field, we could not investigate the potential influence of age difference on U3 snoRNA expression between non-OA and OA cartilage/chondrocytes. To circumvent potential U3 expression interpretation issues, we reinforced our analyses by performing additional U3 snoRNA expression analyses in a murine model for OA and used synovial fluid from OA and non-OA individuals (no statistically significant age difference) on chondrocytes. Both as models in which age was not a potential confounder. Data from these models confirmed the OA-dependent reduction of chondrocyte U3 snoRNA expression. We could also conclude that reduced OA chondrocyte U3 snoRNA expression is, at least in part, mediated via diminished U3 gene transcription activity. This is the first data demonstrating that the extracellular environment is capable of controlling cellular U3 snoRNA levels, thereby tuning the cell's capacity to generate mature rRNA species. U3 snoRNA is transcribed from a dedicated transcriptional unit under the control of an RNA polymerase II-driven (RNAPII) promoter. Little is known about the control www.nature.com/scientificreports/ of snoRNA transcription from dedicated transcriptional units, but our group previously reported that transcriptional activity of another dedicated snoRNA gene (RMRP) can also be influenced via major chondrocyte signaling pathways during chondrogenesis 39 . OA synovial fluid is a complex body fluid containing a plethora of catabolic morphogens and cytokines/chemokines 40,41 . Therefore we expect that the signaling molecule composition of OA synovial fluid acts on major chondrocyte signaling pathways, leading to the observed decreased U3 snoRNA expression. Indeed our data demonstrated that the katabolic cytokines IL1β and TNFα attenuated U3 promoter activity. Reciprocally, we demonstrate that non-OA synovial fluid and BMP7 are capable of inducing U3 snoRNA transcription and subsequent U3 snoRNA levels in chondrocytes. A network of snoRNAs is involved in key ribosome biogenesis processes 16 , and U3 snoRNA is a factor in the generation of 18S rRNA 14 . Indeed our data show that manipulation of chondrocyte U3 snoRNA levels impacts 18S rRNA levels, but also alters 5.8S and 28S rRNA levels in many instances. Taking into consideration the 47S pre-rRNA multi-cistronic origin of 18S rRNA, we speculate that aberrations in 18S rRNA levels may impact 5.8S and 28S rRNA levels possibly due to a 40S over 60S ribosomal subunit imbalance 42 . Also other snoRNAs involved in 47S pre-rRNA processing could be impacted by OA conditions. Our unpublished data suggest deregulated chondrocyte U13 and RMRP snoRNA levels in OA conditions. This may provide additional explanations for the observed abberations in the levels of other rRNAs. In contrast with our current findings, other recent work demonstrated increased protein translation activity in OA cartilage [43][44][45] . The OA cartilage used in our study was derived from end-stage (K&L grade 3-4) knee OA patients, while the OA grade of the human OA-lesioned cartilage presented by others 45 appears moderate. Taking into account that the reported increased protein translation activity in a rat model for OA seems to fade with progression into late OA development 45 , we speculate that the contradictory observations may be explained by OA severity of the analyzed cartilage. This is further supported by the idea that an inflammatory component is much more dominant in early OA than in end-stage OA 46 , and it Puromycilation data (A-C) were calculated relative to corresponding controls and were normalized to DNA content. Statistical significance was determined using Student's t-tests; (A) 2-tailed unpaired, (B/C) 2-tailed paired. Bars show the mean (± SD). *P < 0.05, **P < 0.01, ***P < 0.001 versus control conditions. (D) SW1353 cells were transfected with a U3 ASO or SCR control and cultured for 24 h (knockdown was confirmed in Supplementary Fig. 2), after which LC-MS/MS and label-free quantification was conducted. Ingenuity pathway analysis was performed on the differentially expressed proteins between conditions. Green nodes; decreased expression following knockdown, red nodes; increased expression following knockdown. Intensity of colour is related to a higher fold-change. Key to the main features in the networks is shown. (E) Upstream analysis in Ingenuity Pathway Analysis on the differentially expressed secreted proteins between conditions [(D) U3 snoRNA knockdown in SW1353]. This analyses linkage to differentially expressed proteins through coordinated expression, identify potential upstream regulators that has been observed experimentally to affect gene expression. Green nodes; decreased expression following U3 snoRNA knockdown, red nodes; increased expression following knockdown. Intensity of colour is related to a higher fold-change. Key to the main features in the networks is shown. www.nature.com/scientificreports/ was shown that the increase in translation activity could be recapitulated by IL1β 45 . This suggests that an increase in chondrocyte translation activity in OA may be predominantly related to early OA, while impaired protein translation may be a hallmark of end-stage OA. While increased chondrocyte protein synthesis capacity in early OA may be related to mTOR signaling and imbalanced autophagy 47 , it is unknown what the specific consequence is of a diminished capacity of global protein synthesis in relation to development or progression of osteoarthritis. We propose that U3 snoRNA-related reduction of chondrocyte protein translation affects the chondrocyte's ability to maintain the molecular build-up of the cartilage proteinaceous ECM. Alterations in protein translation capacity may render articular cartilage vulnerable to any biomechanical, catabolic, traumatic or age-related changes in the joint that are associated with OA development and progression 2 . In addition, our gene expression data, which are a representation of the overall chondrocyte differentiation status 3,4 , show that alterations in chondrocyte U3 snoRNA levels affects its differentiation status. Data from the secretome proteomics analysis further show U3-dependent alterations in the levels of secreted proteins involved in balancing inflammatory pathways (A2M; AFP; HGF; AHSG). Together this shows that besides an influence on protein translation, alterations in chondrocyte U3 snoRNA levels additionally impact the chondrocyte's homeostatic integrity.
Our whole proteome analysis uncovered that reduction of U3 snoRNA levels induces ribosome stress, which is indicated by aberrations in molecular pathways controlling key aspects of rRNA and ribosomal protein gene transcription (PGAM1; EDF1; MAPK1; AHCY; CARM1; FLNA), rRNA processing (FSCN1; RPS8; RAN; RPS24), ribosomal subunit transport (RAN), translation initiation (PHGHD; AKT1S1; EIF4H; ARPC5; TIAL1; EIF2S3; PABPC1; HSPB1; TPT1; FN1), chain elongation (DARS; EPRS; DDX3X) and protein folding/ chaperoning (AHSA1; HSPA8; HSP90; HSPB1) ( Supplementary Fig. 3). Ribosome stress following reduction of U3 snoRNA expression was previously reported and is accompanied by stabilization of p53 12 . Indeed we also observed increased expression of p53 after U3 snoRNA knockdown in SW1353 cells (data not shown). It was previously reported that p53 expression is also increased in OA chondrocytes 48 . Stabilisation of p53 feeds back to ribosome biogenesis and down regulates translational activity 49,50 . Besides the U3 snoRNA-dependent availability of rRNAs for ribosome biogenesis, this may provide a link between our observed OA-related impairment of U3 snoRNA expression levels in chondrocytes and their translational activity. In addition, IPA analyses of the U3 knock-down cellular proteome demonstrated that a great number of the differentially expressed proteins are targets of Myc (data not shown). We therefore speculate that stabilization of p53 may impact chondrocyte gene ]_Hs_U3_promoter plasmid. Subsequently, SW1353 cells (A) were exposed to 10 or 30 nM BMP7 and HACs (B) to 10 nM BMP7 for 20 h and Nanoluc luciferase levels were measured. Data were normalized and calculated relative to control conditions (RLU). (C) SW1353 cells and (D) non-OA primary chondrocytes (n = 4), were exposed to 1 nM BMP7 for 24 h. Expression of U3 snoRNA; 18S, 5.8S and 28S rRNA was determined relative to control conditions by RT-qPCR analysis. Results were normalized to cyclophilin expression. (E) SW1353 cells were exposed to 1 nM BMP7 for 24 h. Protein translation capacity was determined using a puromycilation assay and data were normalized for DNA content. Statistical significance was determined using Student's t-tests; (A/B) 1-tailed unpaired, (C/E) 2-tailed unpaired, (D) 2-tailed paired. Bars show the mean (± SD). *P < 0.05, **P < 0.01, ***P < 0.001 versus control conditions. Scientific RepoRtS | (2020) 10:13426 | https://doi.org/10.1038/s41598-020-70453-9 www.nature.com/scientificreports/ transcription processes via Myc, providing a potential explanation for the impact and discrepancies found in chondrocyte gene expression following alteration of U3 levels (Figs. 2, 3).
To comprehend the implications of the larger snoRNA network 16 on the function of the articular chondrocyte, other members of the non-canonical and canonical classes of snoRNAs should also be investigated. However, our study for the first time demonstrates the involvement of a snoRNA in articular chondrocyte biology. The OA-related impaired expression of U3 snoRNA has detectable consequences for articular chondrocyte protein translation capacity, rendering U3 snoRNA expression a potential therapeutic target in OA treatment. Indeed our data show that U3 snoRNA expression can be induced by the growth factor BMP7, providing a potential manner to target U3 snoRNA in OA chondrocytes.