Introduction

Protein ubiquitination is a reversible post-translational modification that regulates a multitude of biological functions including cell cycle regulation, kinase signaling, protein degradation, and DNA repair1. Deubiquitination by deubiquitinating enzymes (DUBs) counter-regulates this event and replenishes the ubiquitin pool. The importance of DUBs has been highlighted in various fields. Several studies have provided insight on how to define the DUBs interaction landscape, their regulation, and physiological roles2,3. The human genome encodes around 97 DUBs, of which 79 are predicted to be catalytically active4. DUBs cluster in five main families: ubiquitin-C-terminal hydrolases (UCHs), ubiquitin-specific peptidases (USPs), ovarian tumor proteases (OTUs), Machado-Joseph disease protein domain proteases (MJDs), and JAB1/MPN/MOV34 metalloenzymes (JAMMs)4.

The largest number of DUBs clusters under the USP umbrella4. The genomic landscape encompasses about 58 different USPs, few of which have a clear assigned function and substrate. Based on sequence analysis, six USPs (USP39, USP50, USP52, USP53, USP54, and USPL1) are predicted to be inactive, yet experimental evidence is still scarce5. One study has linked Usp53 to progressive hearing loss in mice and confirmed that USP53 exhibits a catalytically inactive domain, suggesting a mechanism of protein–protein interactions involving USP536. Despite the progress achieved so far in understanding the physiological role of USP53, nothing is known about the transcriptional regulation and function of this gene in bone biology.

Parathyroid Hormone (PTH) stands as a master regulator of mineral homeostasis. Incessant infusion of PTH promotes calcium and phosphorous release from bone through activation of osteoclast-mediated resorption7,8. However, PTH functions within its anabolic window through intermittent (iPTH) treatment that promotes bone formation through pleiotropic effects on osteoblasts and osteocytes9,10,11. In the cell, numerous studies have uncovered a multitude of signaling cascades that function downstream of PTH, parallel or synergistically to achieve the full anabolic response of iPTH (reviewed in12,13). The work of our group has characterized one of these signal transduction pathways: the PTH-Gαs-PKA-NACA pathway14. We have demonstrated the physiological relevance of this pathway in vivo and dissected the cascade that initiates with PTH binding to its receptor, PKA-dependent phosphorylation of the transcriptional coregulator NACA on residue Serine 99, and nuclear translocation of NACA14. NACA (alpha chain of the nascent polypeptide-associated complex, αNAC) is a physiologically relevant transcriptional coregulator of osteogenic targets in bone tissue14,15. NACA has been found to specifically bind DNA and assemble with transcription factors such as JUN and JUND homodimers to enhance the transcription of the osteocalcin (Bglap2) and Lrp6 genes15,16,17,18. More recently, we have shown that NACA is required for the CREB-mediated activation of the Nfil3 gene in osteoblasts19.

Here, we report the identification of Usp53 as a downstream target of the PTH-NACA pathway in osteoblasts, using Chromatin Immunoprecipitation with deep sequencing (ChIP-Seq) against NACA and RNA-Seq in MC3T3-E1 cells treated with PTH(1–34). We describe a mechanism downstream from activated PKA involving the JUN-CREB complex along with NACA regulating transcription from the Usp53 promoter in osteoblasts. Additionally, we describe a role for USP53 in mesenchymal cell differentiation. We report that the silencing of Usp53 in ST2 stromal cells enhanced osteoblastogenesis and inhibited adipogenesis in vitro. In addition, Usp53-deficient bone marrow stromal cells (BMSCs, also known as bone-marrow derived mesenchymal stem cells) exhibited more bone and less fat when implanted in immunocompromised mice.

Results

The PTH-NACA pathway regulates Usp53 in osteoblasts

Our lab has characterized the binding of NACA to multiple gene promoters in osteoblasts. In one study, the PTH (1–34) stimulation of osteoblasts induces the phosphorylation of NACA at S99, its subsequent nuclear translocation and binding to the Bglap2 promoter14. Another study has highlighted the role of NACA along with JUND in Lrp6 transcriptional regulation in MC3T3-E1 cells18.

To further our understanding of the significance of the PTH-activated NACA pathway in bone, we performed anti-NACA Chromatin Immunoprecipitation with deep sequencing (ChIP-Seq) in PTH-treated MC3T3-E1 osteoblast-like cells to identify genome-wide DNA-binding sites of NACA. This strategy identified a limited number of peaks (Supplementary Fig. S1A) when compared to the ChIP-Seq performed using naive IgG suggesting that PTH-activated NACA transcriptional targets are not abundant (Gene Expression Omnibus Accession number: GSE147070). These peaks mapped close to the transcription start site (TSS) within the proximal promoter region and included a match to the NACA binding consensus sequence 5′-g/cCAg/cA-3′ (Supplementary Fig. S1A). One candidate gene target for NACA in osteoblasts identified by this strategy was Ubiquitin-Specific Peptidase 53 (Usp53) (Fig. 1A). Binding of NACA to the proximal promoter of Usp53 was confirmed by ChIP-PCR on MC3T3-E1 cells stimulated with PTH (1–34) or vehicle under the same conditions (Fig. 1B). We amplified the Usp53 promoter region corresponding to the peak identified by ChIP-Seq. Using an anti-phosphoS99-NACA-specific antibody, results show a significant enrichment of the Usp53 promoter by NACA in vehicle-treated cells as compared to naive IgG. This enrichment was further enhanced by fourfold following 100 nM PTH (1–34) treatment for 30 min. Similar results were obtained using a pan-specific anti-NACA antibody (Fig. 1B). In parallel, PTH-responsive genes were identified by RNA-Seq gene expression profiling in PTH-treated MC3T3-E1 cells. The RNA-Seq analysis identified a total of 37,479 gene targets, of which the expression of 34,623 were unchanged while 2856 were either significantly up (695) or down-regulated (1392) by twofold or more (95% confidence; moderated t-test) (Supplementary Fig. S1B). GO term analysis of regulated gene targets revealed multiple functional groups associated with skeletal biology. The top GO terms from top 3 annotation clusters for upregulated genes from DAVID are: positive regulation of transcription (P = 3.6E−12), glucocorticoid receptor binding (P = 7.5E−5), and MAP Kinase tyrosine/serine/threonine phosphatase activity (P = 1.3E−5) (Supplementary Fig. S1B). Moreover, the functional analysis of down-regulated genes uncovered candidates associated with transcription and DNA binding (P = 9.1E−20), cell cycle (P = 2.5E−12), and chromosome segregation (P = 3.0E−7) (Supplementary Fig. S1B). A further evaluation of regulated candidate genes highlighted the enrichment of signaling pathways associated with major cellular skeletal functions (Supplementary Fig. S1C). PTH(1–34) treatment upregulated genes that are closely-related to osteoblast and osteoclast differentiation and downregulated genes associated with cell cycle and cellular survival (Supplementary Fig. S1C).

Figure 1
figure 1

ChIP-Seq and RNA-Seq identify Usp53 as a target of the PTH-NACA axis in osteoblasts. (A) Annotated screenshot given by the Integrative Genomics Viewer (IGV) showing NACA-ChIP-Seq peak (highlighted in blue) detected in the proximal region of the Usp53 promoter. ChIP-Seq was performed with anti-NACA-pS99 antibody or naïve IgG on MC3T3-E1 cells induced or not with 100 nM PTH(1–34) for 30 min. (B) NACA binds the proximal Usp53 promoter. Quantitative chromatin immunoprecipitation (ChIP) was performed with anti-NACA-pS99-specific antibody, pan-specific anti-NACA antibody, or naive IgG on MC3T3-E1 cells induced or not with 100 nM PTH(1–34) for 30 min. Inputs and bound DNA sequences were amplified by Real Time PCR (RT-PCR) using SYBR green technologies and specific primers flanking the NACA binding site within the Usp53 promoter. Relative promoter occupancy was calculated as enrichment over vehicle treated cells and results are means ± SD of three independent experiments (n = 3). (C) Induction of Usp53 mRNA expression post treatment by PTH(1–34) for 1 h as determined by RNA-sequencing. Results are shown as fragments per kilobase of exon per million fragments mapped (FPKM), as given by RNA-sequencing. (D) RT-qPCR analysis of mRNA expression for Usp53 in primary osteoblast cells treated with vehicle or 100 nM PTH(1–34) for the indicated time points. (E) Immunoblot of whole cells extracts from MC3T3-E1 cells stably expressing shRNAs targeting Naca (shNACA.1 and shNACA.2) and scrambled shRNA (shScr) as control. Membranes were probed with anti-NACA and anti-GAPDH antibodies (as loading controls). Specific binding to NACA proteins is denoted by the arrow and NS indicates non-specific binding of the NACA-antibody. Full-length blots are presented in Supplementary Figure S4. (F) RT-qPCR analysis of mRNA expression of Usp53 in stably transfected MC3T3-E1 cells (shScr, shαNAC.1, or shαNAC.2) treated with vehicle or 100 nM PTH(1–34) for 2 h. Data are expressed relative to levels measured in the shScr expressing cells treated with vehicle. Results are means ± SD of three independent experiments (n = 3). #P < 0.0001; ANOVA with Bonferroni’s post-hoc test for panels B, D, and E; unpaired t test was used for panel C.

RNA-Seq performed on MC3T3-E1 cells stimulated for 60 min with 100 nM PTH (1–34) showed a threefold increase in Usp53 transcript levels as compared to vehicle-treated cells (Fig. 1C). This increase in Usp53 transcript levels was further confirmed using quantitative RT-qPCR in primary calvarial osteoblasts treated with vehicle or 100 nM PTH (1–34) at different time points. The expression of Usp53 increased after 1 h of treatment, peaked at 2 h of treatment and then returned to basal levels by 6 h (Fig. 1D).

The ChIP-Seq data indicated the binding of NACA to the proximal region of the Usp53 promoter. In order to investigate the role of NACA in Usp53 gene regulation, we used a short hairpin RNAs (shRNAs)-mediated knockdown approach to target Naca in MC3T3-E1 cells. The efficiency of NACA knockdown (using shNACA.1 or shNACA.2) was demonstrated by immunoblotting compared to MC3T3-E1 cells stably expressing a scrambled shRNA (shScr) (Fig. 1E). We then checked the effect of NACA knockdown on Usp53 mRNA expression levels. The knockdown of NACA did not impact the expression levels of Usp53 at the basal levels (Fig. 1F). PTH(1–34) stimulation of shScr-expressing MC3T3-E1 cells increased the expression levels of Usp53 by threefold. The PTH(1–34) response was completely blunted in NACA knockdown MC3T3-E1 cells (Fig. 1F).

Together, our data show that Usp53 is a PTH target in osteoblasts, where NACA can be involved directly or indirectly in the regulation of its expression following PTH treatment. The data also indicate that NACA contributes to the PTH-mediated transcriptional induction of Usp53.

Usp53 promoter hosts a functional binding site for NACA

Our data show that NACA has an impact on the transcriptional regulation of Usp53 following PTH(1–34) stimulation. We have previously characterized the DNA binding domain of the NACA protein to a loose consensus sequence centered around 5′-g/cCAg/cA-3′20. In silico analysis of the Usp53 promoter region between nucleotides − 1000 and + 94 using MatInspector software21 identified one potential binding site for NACA and one binding site for the cAMP response element binding protein (CREB) (Fig. 2A). The predicted proximal binding site of NACA to the promoter (5′-GGCTCAGATCCCC-3′, between nucleotides − 333 and − 345) lies within the peak identified by ChIP-Seq analysis (Fig. 2A).

Figure 2
figure 2

The proximal promoter of Usp53 contains binding sites for NACA and CREB. (A) Mouse Usp53 promoter sequence from nucleotide − 406 to nucleotide + 94 relative to the transcriptional initiation site (arrow). Nucleotides corresponding to the peak identified by ChIP-Seq are bracketed and in bold. Putative binding sites for NACA and CREB (highlighted) were called using Mat Inspector (Genomatix software suite). (B) Electrophoretic mobility shift assay (EMSA) analysis showing the binding of NACA to its binding site in the Usp53 promoter. EMSA was performed using probes harboring the NACAbds sequence present in the Myoglobin (lanes 1–4) or the Usp53 promoter (lanes 5–8), or a probe with a mutated NACAbds sequence present in the Usp53 promoter (Usp53 mut; lanes 9 and 10). Nuclear extracts (NE) proteins from HEK293 cells overexpressing NACA and treated with 6Bnz-cAMP to induce nuclear translocation were incubated with the probes. Negative binding was assessed by incubating the probes with the binding buffer (lanes 1 and 5). The black arrow (Shift; bottom arrow) indicates the generation of a slower migrating EMSA product, the resultant of NACA binding to NACAbds sequences in both the Myoglobin and Usp53 probes. Addition of anti-NACA antibody led to an additional supershift complex (SS; top arrow), not detected when adding naive non-specific (NS) IgG to the binding reactions.

In order to further validate the binding of NACA to its potential binding site (bds) within the Usp53 promoter, we performed an electrophoretic mobility shift assay (EMSA) using probes corresponding to NACAbds. The binding of NACA to the Myoglobin promoter was used as a positive control in these experiments20. The formation of mobility shift complexes (Fig. 2B, lanes 2 and 6) confirmed the binding of proteins from NACA-enriched nuclear extracts to the probes harboring NACAbds. We further confirmed the presence of NACA in the shifted DNA–protein complexes. The addition of anti-NACA antibody to the reactions super shifted (SS) the formed complexes (Fig. 2B, lanes 4 and 8). This supershift was not observed when adding naïve IgG (Fig. 2B, lanes 3 and 7). In order to further validate the fidelity of NACA binding, we mutated its potential binding site within the Usp53 promoter. The 5′-GGCTCAGATCCCC-3′ NACA binding site was mutated to 5′-GGCTCgGgTCCCC-3′. The formation of a DNA–protein complex was abrogated when incubating NACA protein extracts with mutated NACAbds probes (Fig. 2B, lane 10). These experiments confirm the binding of NACA to its cognate site within the Usp53 proximal promoter.

NACA is implicated in the regulation of the Usp53 promoter

We cloned a fragment of the Usp53 promoter corresponding to nucleotides − 2325 to + 238 bp (− 2325/+ 238 bp) relative to the annotated transcription start site into the pGL4.10[LUC2] reporter vector. In order to check the functionality of the NACA binding site, we implemented site-directed mutagenesis and shRNA-mediated knockdown techniques. The 5′-GGCTCAGATCCCC-3′ NACA binding site was mutated to 5′-GGCTCgGgTCCCC-3′ in the context of the − 2325/+ 238 bp promoter fragment. The cloned promoter fragment (− 2325/+ 238 bp) carrying a wild-type response element for NACA (Wt-NACAbds) showed a tenfold increase in the transcriptional activity of the luciferase reporter as compared to cells transfected with pGL4 vector (Fig. 3A). The Usp53 promoter activity was reduced by 40% when we transfected the same fragment carrying a mutated sequence of the NACA response element (Mut-NACAbds) (Fig. 3A, black bars). The mutation abrogated the PTH-induced activation of the cloned Usp53 fragment, resulting in > 50% reduction in the promoter activity (Fig. 3A, grey bars). Interestingly, the knockdown of NACA using shαNAC.1 or shαNAC.2 shRNAs resulted in > 80% reduction in Usp53 promoter activity as compared to reporter gene expression of MC3T3-E1 cells transfected with shScr (Fig.3B, black bars). This impact on basal levels of Usp53 expression from the cloned promoter (Fig. 3B, black bars) as compared to what was observed on basal level expression of the endogenous gene (Fig. 1E) most probably reflects differences between naked plasmid DNA vs. endogenous gene in its native chromatin environment. The absence of NACA blunted the induction of the Usp53 promoter fragment by PTH (Fig.3B, grey bars). Taken together, these results show that the presence of NACA and its proper binding to its response element is critical for promoter activity.

Figure 3
figure 3

NACA, JUN, and CREB are implicated in Usp53 promoter regulation. (A) Firefly luciferase (Fluc) reporter assay was performed using MC3T3-E1 cells overexpressing wildtype or mutated Usp53-Luc vector (-2325/+ 238 bp). The mutated region lies in the proximal NACA binding site of the Usp53-Luc vector. (B) Firefly luciferase (Fluc) reporter assay were performed using MC3T3-E1 cells stably expressing shRNAs targeting Naca (shαNAC.1 and shαNAC.2) or scrambled shRNA (shScr) as a control by lentiviral infection. When indicated, cells were treated with vehicle or 100 nM PTH (1–34) for 24 h before performing the assay. (C) Firefly luciferase (Fluc) reporter assay in MC3T3-E1 cells transfected with Usp53-Luc vector (− 2325/+ 238 bp) alone or along with JUN and/or CREB expression vectors treated with vehicle or 100 nM PTH(1–34) for 24 h. (D) Firefly luciferase (Fluc) reporter assay in MC3T3-E1 cells stably expressing shRNAs against Naca (shNACA.1 and shNACA.2) or scrambled shRNA (shScr), transfected with Usp53-Luc vector (-2325/+ 238 bp) alone or along with JUN and/or CREB expression vectors. Cells were treated with 100 nM PTH (1–34) for 24 h before performing the assay. (E) Firefly luciferase (Fluc) reporter assay in MC3T3-E1 cells transfected with Usp53-Luc vector (-2325/+ 238 bp) alone or along with NACA-S99D or NACA mutant deltaDBD (69–80), JUN and/or CREB expression vectors. Cells were treated with 100 nM PTH (1–34) for 24 h before performing the assay. (F) Firefly luciferase (Fluc) reporter assay in MC3T3-E1 cells transfected with Usp53-Luc vector (-2325/+ 238 bp) alone or along with 50 ng of NACA, JUN and/or CREB expression vectors, with or without 100 ng of dominant negative DN-PKA expression vector. Cells were treated with 100 nM PTH(1–34) for 24 h before performing the assay. Empty pGL4 vector was used as a negative control. Firefly luciferase counts were normalized to Renilla counts. Results are presented as the mean fold induction ± SD. *P ≤ 0.05; **P ≤ 0.01; ***P ≤ 0.001; #P < 0.0001; ANOVA with Bonferroni’s post-hoc test for all panels.

NACA is required for the JUN/CREB-mediated synergistic regulation of the Usp53 promoter

A potential binding site for CREB lies between nucleotides − 5 and − 14 relative to the annotated transcription initiation site, within the Usp53 proximal promoter (Fig. 2A). The in silico predictions suggest a potential role for CREB along with NACA in the transcriptional regulation of the promoter. We have recently observed that both NACA and CREB can regulate the transcription of the Nfil3 promoter in MC3T3-E1 cells19. Besides, previous studies have highlighted the need for a bona fide transcription factor partner such as the basic domain-leucine zipper (bZIP) homodimeric AP-1 family member JUN to regulate transcription, as NACA lacks a transcriptional activation domain17,22. Considering this, we tested whether CREB along with JUN activate the transcription of the cloned promoter fragment alone or in combination with NACA.

We first assessed the JUN-CREB regulation of the Usp53 cloned fragment at basal levels and in response to PTH(1–34) stimulation in MC3T3-E1 cells. At basal levels, the overexpression of 50 ng of JUN expression vector did not induce a statistically significant increase in the promoter activity. However, the transcriptional activity of the cloned promoter fragment (-2325/+ 238 bp) was significantly enhanced following the overexpression of 50 ng of CREB and further potentiated when co-expressing CREB and JUN together (Fig. 3C, black bars). The CREB-mediated and the dual activation by JUN and CREB were dramatically increased following PTH(1–34) treatment for 24 h (Fig. 3C, grey bars). This data highlighted the combinatorial actions of the JUN/CREB complex in regulating the Usp53 promoter activity at basal levels and following PTH(1–34) treatments.

We then assessed the JUN-CREB activation of the Usp53 cloned fragment in Naca-knockdown cells. The transcriptional activity of the cloned promoter fragment (-2325/+ 238 bp) was higher than the empty pGL4 vector (Fig. 3D, black bars). As previously demonstrated in Fig. 3C, similar trends of promoter activation were obtained following CREB overexpression or co-expression of CREB and JUN together in MC3T3-E1 cells transfected with shScr (Fig. 3D, black bars). Interestingly, this additive activation was completely abolished in MC3T3-E1 cells transfected with NACA shRNAs (Fig. 3D, light and dark grey bars). It is important to note that the absence of NACA did not affect the promoter activation when either JUN or CREB were transfected alone (Fig. 3D, light and dark grey bars).

We then investigated whether the binding of NACA to DNA is needed for the optimal promoter activation carried by JUN and CREB. To mimic the PTH(1–34) effect on NACA translocation, we overexpressed NACA-S99D mutant; a phosphomimetic form of NACA that shows enhanced nuclear localization14. As described in Fig. 3D, similar patterns of promoter activation were seen when overexpressing JUN and CREB, alone or combined (Fig. 3E, black bars). The overexpression of NACA-S99D mutant or a NACA mutant lacking its DNA binding domain (delta DBD(69–80)) had no effect on the JUN or CREB-mediated activation of the promoter fragment (Fig. 3E, light and dark grey bars). However, overexpressing NACA-S99D along with JUN and CREB in MC3T3-E1 cells boosted the promoter activity as compared to conditions with JUN and CREB alone (Fig. 3E, grey bar). This enhanced activity was not seen when overexpressing the delta DBD (69–80) NACA mutant in the same condition (Fig. 3E, dark grey bar).

We interpret this data to mean that NACA binding to its cognate site is important for the combined actions of JUN and CREB in the regulation of Usp53 promoter activity.

The NACA/JUN/CREB activation of the Usp53 promoter is mediated by PKA

Our data (Fig. 3C–E) have demonstrated the PTH-induced CREB activation of the Usp53 promoter. One interesting aspect of this regulatory mechanism is the involvement of protein kinase A (PKA) in this activation. To test this possibility, we have blocked endogenous PKA signaling by expressing a dominant-negative form of PKA (DN-PKA)23 and assessed the integrity of the NACA/JUN/CREB activation. The transcriptional activation of the Usp53 promoter by CREB alone, CREB/JUN together or NACA/JUN/CREB combined was completely abolished following the overexpression of 100 ng of DN-PKA expression vector (Fig. 3F). This data indicates that PKA is implicated in the regulation of the Usp53 promoter by the NACA/JUN/CREB trio in MC3T3-E1 cells.

Usp53 affects mesenchymal cell lineage selection and differentiation in vitro

Two main isoforms of Usp53 have previously been described: a full-length isoform comprising 15 coding exons and a short isoform containing exons 1–7 of the Usp53 gene6. Little is known about the expression of Usp53 in different lineages. We analyzed the expression of the two isoforms by RT-qPCR in long and flat bones as well as in fat compartments of wild-type mice at 3-weeks of age. Usp53 was expressed in calvaria and tibia and in white (gonadal) and brown (interscapular) fat. In all tissues tested, the expression of the long, full-length isoform was predominant (Supplementary Fig. S2).

To examine the putative role of Usp53 in mesenchymal cell differentiation, we targeted its expression in ST2 stromal cells. Short hairpin RNAs (shRNAs) targeting Usp53 (shUsp53) or a control scrambled shRNA (shScr) were transfected into ST2 cells and stable pools expressing the distinct shRNAs were established. The stably transfected cells were differentiated in either osteogenic or adipogenic medium for 6 days. The expression levels of Usp53 showed a significant 50% reduction following knockdown and remained stable through the osteoblast differentiation regimen (Figs. 4A, 5A and 6A). Knockdown of Usp53 in ST2 and MC3T3-E1 cells using two different shUsp53 vectors resulted in significant reduction in USP53 protein levels (Figs. 4C, 5C and 6B).

Figure 4
figure 4

Usp53 controls osteogenic and adipogenic fate determination in vitro. ST2 cells expressing scrambled shRNAs (shScr) targeting control and shRNAs targeting Usp53 were grown under osteogenic conditions (A) or adipogenic conditions (B) for 6 days. (C) Immunoblot of whole cells extracts from ST2 cells stably expressing shRNA targeting Usp53 (shUsp53) and scrambled shRNA (shScr) as control. Membranes were probed with anti-USP53 and anti-GAPDH antibodies (as loading controls). Relative signal intensity quantified using ImageJ is shown below each lane. Full-length blots are presented in Supplementary Figure S5. Gene expression analysis of osteogenic (D) or adipogenic (F) differentiation markers. Quantitative RT-qPCR using TaqMan probes against osteoblast or adipocyte differentiation markers, normalized to Gapdh, was performed on RNA isolated from shScr-stably transfected (dashed line) or shUsp53-stably transfected (solid line) pools of ST2 cells grown for 6 days in osteogenic or adipogenic medium. ST2 cultures were examined for ALP staining at day 12 (E) or oil red O at day 6 (G). Results are presented as the mean fold change \(\pm\) SD. *P ≤ 0.05; **P ≤ 0.01; ***P ≤ 0.001; #P < 0.0001; ANOVA with Bonferroni’s post-hoc test for all panels.

Figure 5
figure 5

Usp53 knockdown with a different shRNA confirms the impact on osteogenic and adipogenic fate determination in vitro. ST2 cells expressing scrambled shRNAs (shScr) targeting control and shRNAs targeting Usp53 were grown under osteogenic conditions (A) or adipogenic conditions (B) for 6 days. (C) Immunoblot of whole cells extracts from ST2 cells stably expressing shRNA targeting Usp53 (shUsp53.1) and scrambled shRNA (shScr) as control. Membranes were probed with anti-USP53 and anti-GAPDH antibodies (as loading controls). Relative signal intensity quantified using ImageJ is shown below each lane. Full-length blots are presented in Supplementary Figure S6. Gene expression analysis of osteogenic (D) or adipogenic (F) differentiation markers. Quantitative RT-qPCR using TaqMan probes against osteoblast or adipocyte differentiation markers, normalized to Gapdh, was performed on RNA isolated from shScr-stably transfected (dashed line) or shUsp53-stably transfected (solid line) pools of ST2 cells grown for 6 days in osteogenic or adipogenic medium. ST2 cultures were examined for von Kossa staining at day 8 (E) or oil red O at day 6 (G). Results are presented as the mean fold change ± SD. **P ≤ 0.01; ***P ≤ 0.001; #P < 0.0001; ANOVA with Bonferroni’s post-hoc test for all panels.

Figure 6
figure 6

Usp53 knockdown enhances the differentiation of committed osteoblasts. (A, C, D) Knockdown of Usp53 in MC3T3-E1 cells and gene expression analysis of osteogenic differentiation markers. Quantitative RT-qPCR using probes against Usp53, Alpl, and Bglap2, normalized to Gapdh, was performed on RNA isolated from shScr-stably transfected (dashed line) or shUsp53-stably transfected (solid line) pools of MC3T3-E1 cells differentiated for 21 days in osteogenic medium. (B) Western blot analysis of Usp53 expression. Total protein isolated from MC3T3-E1 cells stably transfected with shScr or shUsp53 at D1, confirming the knockdown of Usp53. GAPDH was used as a loading control. Full-length blots are presented in Supplementary Figure S7. (E) MC3T3-E1 cultures were examined for calcium deposition using alizarin red staining at D21. Results are presented as the mean fold change ± SD. *P ≤ 0.05; **P ≤ 0.01; ***P ≤ 0.001; #P < 0.0001; ANOVA with Bonferroni’s post-hoc test for all panels.

We then monitored the level of transcripts expressed during early- and late-stage osteoblastic differentiation. The knockdown of Usp53 enhanced the expression of early-stage osteoblast markers, such as Osterix (Sp7) and Dlx5 as well as late-stage osteoblast markers such as osteocalcin (Bglap2) and alkaline phosphatase (Alpl) (Fig. 4D). Alkaline phosphatase activity was markedly increased in the absence of Usp53 (from 34.4 ± 4.2% in controls to 74.7 ± 4.1% in Usp53 knockdown cells; P = 0.0001; t-test), assessed by staining at Day 6 (Fig. 4E). These results show that the absence of Usp53 enhanced osteoblast differentiation. This data was further validated using another shRNA against Usp53 to eliminate the probability of any off-target effects associated with shRNA use (Fig. 5A, C–E).

Similar results were obtained upon analysis of MC3T3-E1 osteoblast-like cells. Usp53-knockdown MC3T3-E1 cells exhibited a drastic increase in osteocalcin (Bglap2) and Alpl expression levels throughout differentiation (Fig. 6C, D) together with increased calcium deposition (from 13.1 ± 2.5% in controls to 40.2 ± 2.4% in Usp53 knockdown cells; P = 0.0001; t-test) as assessed by Alizarin Red Staining (Fig. 6E). We interpret these results to mean that Usp53 inhibits osteoblastogenesis of both bipotential stromal cells and committed osteoblasts.

We also cultured the Usp53 knockdown ST2 cells in adipogenic medium for 6 days. Again, the expression levels of Usp53 were stable throughout adipocyte differentiation and significant knockdown was measured (Fig. 4B). Interestingly, Usp53-deficient ST2 cells cultured in adipogenic media showed decreased induction of adipogenic activators such as Fabp4, Cebpα/β, and Pparγ (Fig. 4F). This translated into decreased lipid accumulation (optical density of staining decreased from 1.2 ± 0.1% in shScr-transfected control cells to 0.4 ± 0.04% in Usp53 knockdown cells; P < 0.0001; t-test) as assessed by Oil Red O Staining at Day 6 (Fig. 4G). Thus, the absence of Usp53 caused a substantial reduction in adipocytic differentiation capacity. Again, these observations were validated using another shRNA against Usp53 to avoid off-target effects associated with shRNA use (Fig. 5B, F, G).

Taken together, our data suggest that Usp53 is a regulator of the lineage-making decisions of mesenchymal cells, altering their differentiation into osteoblasts or adipocytes reciprocally.

PTH responses in Usp53 knockdown osteoblasts and adipocytes

We have shown that Usp53 expression is induced following PTH(1–34) treatment of MC3T3-E1 cells. To determine whether Usp53 alters PTH responses in cultured cells, we treated Usp53-knockdown ST2 and MC3T3-E1 cultures with PTH(1–34) at different timepoints through differentiation. The expression levels of Usp53 showed a significant 70% reduction following knockdown and remained stable through the osteoblast (osteo) and adipocyte (adipo) differentiation regimens (Fig. 7A, solid grey bars). PTH(1–34) stimulation for 2 h did not induce any significant changes in Usp53 expression levels of shScr (Fig. 7A, dashed black bars) or shUsp53-stably expressing ST2 cells (Fig. 7A, dashed grey bars), cultured in osteogenic or adipogenic medium for 6 days (Fig. 7A).

Figure 7
figure 7

PTH responses in Usp53 knockdown osteoblasts and adipocytes. ST2 cells expressing scrambled shRNAs (shScr) targeting control and shRNAs targeting Usp53 were grown under osteogenic or adipogenic conditions for 6 days. Gene expression analysis of Usp53 (A) osteogenic (B) or adipogenic (C) differentiation markers. Quantitative RT-qPCR using TaqMan probes against osteoblast or adipocyte differentiation markers, normalized to Gapdh, was performed on RNA isolated from shScr-stably transfected or shUsp53-stably transfected pools of ST2 cells grown for 6 days in osteogenic or adipogenic medium treated with vehicle or 100 nM PTH(1–34) for 2 h prior to analysis. Gene expression analysis of Usp53 (D) or osteogenic (E) differentiation markers of MC3T3-E1 cells. Quantitative RT-qPCR using TaqMan probes against osteoblast differentiation markers, normalized to Gapdh, was performed on RNA isolated from shScr-stably transfected or shUsp53-stably transfected pools of MC3T3-E1 cells grown for 10 days in osteogenic medium treated with vehicle or 100 nM PTH(1–34) for 2 h prior to analysis. Results are presented as the mean fold change ± SD. *P ≤ 0.05; **P ≤ 0.01; ***P ≤ 0.001; #P < 0.0001; ANOVA with Bonferroni’s post-hoc test for all panels.

Looking into the osteogenic differentiation of Usp53-knockdown ST2 cultures, we assessed the changes in early- and late-stage osteogenic differentiation markers at basal levels and following PTH(1–34) stimulation. At D1 of differentiation, the knockdown of Usp53 enhanced the expression of Sp7, Dlx5, Bglap2, and Alpl significantly in both vehicle- (Fig. 7B, solid grey bars) and PTH(1–34)-treated (Fig. 7B, dashed grey bars) cultures. At D6 of differentiation, the increase in Sp7 expression levels in vehicle-treated cultures (Fig. 7B, solid grey line) was abolished following PTH(1–34) treatment for 2 h (Fig. 7B, dashed grey line). On the other hand, the knockdown of Usp53 did not alter the expression of Dlx5 at D6 of differentiation in both vehicle-(Fig. 7B, solid grey bar) and PTH(1–34)-treated (Fig. 7B, dashed grey bar) cultures. Moreover, the enhanced expression of Bglap2 and Alpl in Usp53-knockdown cultures at basal levels (Fig. 7B, solid grey bars) was unchanged following PTH(1–34) treatment (Fig. 7B, dashed grey bars) at D6 of differentiation.

Assessing the adipogenic differentiation of Usp53-knockdown ST2 cultures, we found that loss of Usp53 inhibits the expression of Fabp4, AdipoQ, and Pparg in Vehicle-treated ST2 cultures through differentiation (Fig. 7C, solid bars). The decreased expression of adipogenic markers was not altered by PTH(1–34) stimulation (Fig. 7C, dashed bars). Cebpα was exclusively downregulated at D6 of adipogenic differentiation in Usp53-knockdown ST2 cultures (Fig. 7C, solid grey bar) and PTH(1–34) treatment of those cultures had no effect on the decreased expression of the gene (Fig. 7C, dashed grey bar).

The same approach was used to characterize the impact of Usp53 knockdown on PTH responses in committed osteoblast cells. The expression levels of Usp53 showed a significant 60% reduction following knockdown and remained stable through MC3T3-E1 differentiation regimen (Fig. 7D, solid grey bars). PTH(1–34) stimulation for 2 h increased the expression levels of Usp53 at D1 and D10 of osteoblastic differentiation, but the levels remained significantly reduced in knockdown cells (Fig. 7D, dashed bars). When changes in the expression levels of osteoblast differentiation markers were monitored, Usp53-knockdown MC3T3-E1 cultures exhibited increased expression levels of Sp7 in vehicle-treated (Fig. 7E, solid grey bars) cultures at D1. PTH(1–34) treatment for 2 h blocked the expression of Sp7 to a similar magnitude in both shScr- and shUsp53-stably expressing MC3T3-E1 cultures at D1 and D10 (Fig. 7E, dashed black and grey bars). Moreover, the knockdown of Usp53 increased the expression levels of Bglap2 throughout differentiation (Fig. 7E, solid bars) and PTH(1–34) treatment had no impact on this increase (Fig. 7E, dashed bars). Expression levels of Alpl were enhanced by Usp53-knockdown (Fig. 7E, solid bars) exclusively at D1, with no significant modulation post PTH(1–34) stimulation (Fig. 7E, dashed bars). Finally, the knockdown of Usp53 had no impact on Dlx5 expression levels through differentiation under any condition (Fig. 7E, solid and dashed bars).

Our data suggest that Usp53 expression levels impact PTH responses differentially across the osteoblastic differentiation sequence. It also indicates that the knockdown of Usp53 does not affect PTH responses during adipogenic differentiation.

Usp53 affects mesenchymal cell lineage-making decisions and differentiation in vivo

To ensure that the observed effects were not restricted to established cell lines, we performed an in vivo differentiation assay with murine bone marrow stromal cells (BMSCs)24. BMSCs were isolated from femoral and tibial bone of C57BL/6 wild-type male mice and infected with shUsp53 or the control shScr and their ability to form bone or fat in vivo was assessed (Fig. 8A).

Figure 8
figure 8

Usp53 knockdown enhances osteoblastogenesis and inhibits adipogenesis in vivo. (A) Scheme showing the workflow of the in vivo osteogenesis assay performed on immunocompromised mice. The mouse drawing is based on a free download from flyclipart.com which was adapted for our purposes. The software used to create the diagram was Adobe Illustrator CS5 (https://www.adobe.com/ca/products/illustrator.html) (B) Knockdown of Usp53 in bone marrow stromal cells (BMSCs). Quantitative RT-qPCR using TaqMan probe against Usp53, normalized to Gapdh, was performed on RNA isolated from stable BMSCs cultures transfected with shScr or shUsp53. (C) Immunoblot of whole cells extracts from BMSCs stably expressing shRNA targeting Usp53 (shUsp53) and scrambled shRNA (shScr) as control. Membranes were probed with anti-USP53 and anti-GAPDH antibodies (as loading controls). Relative signal intensity quantified using ImageJ is shown below. Full-length blots are presented in Supplementary Figure S8. (D) H & E staining of murine implants (shScr and shUsp53) after 4-weeks of BMSCs transplantation embedded in collagen sponge. (E) Immunostaining of the implants (shScr and shUsp53) showing Osterix staining for osteoblasts (red) and DAPI staining for nuclei (blue). (F) Immunostaining of the implants (shScr and shUsp53) showing Perilipin staining for adipocytes (red). (G) Quantification of the newly formed bone in the implants using Micro-CT (upper panel); quantification of Osterix-positive osteoblast cells (middle panel); quantification of perilipin-positive adipocytes (lower panel). BV/TV, bone volume/tissue volume. *P < 0.05, **P < 0.01, ***P < 0.001; t test was used as statistical test in all panels.

The expression level of Usp53 showed a 50% reduction following knockdown in BMSCs (Fig. 8B). The efficiency of USP53 knockdown (using shUsp53) was demonstrated by immunoblotting compared to BMSCs stably expressing scrambled shRNA (shScr) (Fig. 8C). Cells were implanted on collagen scaffolds and inserted subcutaneously into female immunocompromised mice (Fig. 8A). Four (4) weeks after surgery, the mice were sacrificed and the implants were processed for microcomputed tomography (micro-CT) and histological analysis. Initially, sections were stained with H&E, which stains the newly deposited bone matrix. BMSCs expressing shUsp53 showed increased H&E staining (Fig. 8D). Bone formation was also assessed by micro-CT. The relative bone volume versus tissue volume (BV/TV) was higher in Usp53 knockdown implants as compared to control, but the experimental variation prevented to reach statistical significance (Fig. 8G, top panel). To further confirm the presence of bone cells, we performed immunostaining using an anti-SP7 (Osterix) antibody to identify osteoblasts (Fig. 8E) and quantified the results by counting Osterix-positive cells in different histological sections. Usp53 knockdown implants showed a significant increase in Osterix-positive cells (Fig. 8G, middle panel). Adipocytes were detected by immunostaining with an anti-perilipin antibody, which detects surface-associated lipid droplets present in adipocytes. Usp53 knockdown implants showed reduced perilipin staining as compared to control (Fig. 8F), and this decrease was statistically significant (Fig. 8G, lower panel). We conclude that a decrease in Usp53 expression enhanced osteoblastogenesis and suppressed adipogenesis in vivo.

Discussion

In addition to Lrp618, Bglap214, and Nfil319, we have added Usp53 to the list of target genes regulated by the PTH-NACA pathway in osteoblasts. This finding is significant as the literature lacks any characterization of Usp53 regulation or function in any tissue. The identification of Usp53 as a target of the PTH-NACA pathway led us to investigate the transcriptional regulation of Usp53 and assess whether Usp53 may have a function in mesenchymal cell lineage specification.

Little is known about the upstream signaling regulating Usp53 expression. One study has reported the induction of USP53 in human bone marrow MSCs when treated with BMP-2 and Wnt3a together25. This is the first report demonstrating the regulation of Usp53 by PTH signaling. The ChIP-Seq data reported herein show that PTH-activated NACA was recruited to the proximal promoter region of Usp53. The same regulatory model has been previously reported for the Bglap2, Lrp6, and Nfil3 promoters in osteoblasts. Our work highlighted their functions in fine-tuning PTH signaling (in the case of Lrp6), modulating osteoblast function (for Bglap2), and differentially regulating osteoblast vs. osteocyte differentiation (Nfil3)14,18,19. It is yet to be determined whether the PTH-induction of Usp53 in osteoblasts or osteocytes dictates an osteoanabolic or osteocatabolic effect on bone.

Lessons learnt from different promoters regulated by the PTH-NACA pathway helped uncover the missing pieces in the Usp53 regulatory scheme. We have previously shown that NACA engages with bZIP family transcription factors and potentiates their activity. JUN and JUND are two partners of NACA regulating the expression of the Bglap2 and Lrp6 promoters, respectively16,18. In addition, we have recently characterized the transcriptional regulation of the Nfil3 promoter by both NACA and CREB in osteoblasts19. Our data herein suggest that NACA is indispensable for Usp53 promoter activity, its response to PTH, and its full activation by the JUN-CREB complex. We have previously demonstrated that NACA does not interact with CREB at basal levels, nor following cAMP activation or PTH(1–34) induction19. In this respect, it is plausible to hypothesize that JUN and CREB heterodimerize26 and interact with NACA to achieve the optimal potentiation of the Usp53 promoter. We hypothesize that the JUN partner of the heterodimer mediates the interaction of the dimer with NACA. One exciting aspect is the contribution of PKA to the regulatory circuit of Usp53. We envisage this involvement to not only induce CREB activation, but also induce the nuclear translocation and activation of NACA via phosphorylation. This scenario is highly probable, as our previous work has dissected the PTH-Gαs-mediated regulation and activation of NACA by PKA in osteoblasts14.

USP53 has been reported to be expressed in skeletal muscle, hair follicles, and cardiac muscle27. We have further detected its expression in bone tissue (calvaria and tibia) and adipose tissue (white fat and brown fat). The prevalence of the full-length isoform is most likely related to its function in those tissues.

Little is known about the biological function of USP53. Sequence comparisons with other members of the USP gene family revealed that USP53 lacks a functionally essential histidine residue within the His-box involved in the recognition of the C terminus of ubiquitin28, which has led authors to suggest that the protein is devoid of enzymatic activity5,6. The lack of proteolysis activity suggests that USP53 may function as a protein scaffold supporting the assembly of complexes implicated in mesenchymal differentiation. This is supported by the reported interaction of USP53 with ZO-1/TJP1 and ZO-2/TJP2 tight junction scaffolding proteins in polarized epithelial cells of the ear in mice, where it was suggested that USP53 is part of the tight junction complex6. Alternatively, USP53 can act as a decoy protein protecting complexes from other deubiquitinases, as is the case with USP18 regulating TAK1-dependent signaling in a protease-independent manner29.

Cantu syndrome (OMIM#:239850) is a rare condition characterized clinically by hypertrichosis (excessive hair growth), cardiomegaly, and bone abnormalities30. A 375 kb duplication on chromosome 4q26-27 has been identified in a patient with severe Cantu syndrome combined with obesity (BMI > 40)27. The duplication region encompasses the three genes MYOZ2, FABP2, and USP5327. Since increased copy number (and presumably increased expression) was reported in the Cantu syndrome patient with multiple bone abnormalities27, we first attempted overexpression strategies by transient transfection or infection with constitutive or inducible Usp53 expression vectors. In all cases, mesenchymal cells did not survive the addition of extraneous Usp53 coding sequence (data not shown). When Usp53 expression was decreased using shRNA-mediated knockdown, this affected mesenchymal cell lineage-making decisions.

In the context of mesenchymal differentiation, the depletion of either Usp53 (Figs. 4, 5, and 8) or Naca (Supplementary Fig. S3) in ST2 cells enhanced osteoblastogenesis and inhibited adipogenesis. This similar biological outcome provides a link between the transcriptional regulation of Usp53 and its functional characterization in regards to mesenchymal lineage selection. The data provide strong circumstantial evidence that NACA and USP53 form part of a common signaling pathway regulating the differentiation of mesenchymal progenitors. Another approach that could be used to support this interpretation is to alter the gene dosage of Usp53 and Naca in compound heterozygous mutant mice in vivo to confirm the contribution of the pathway to mesenchymal commitment and differentiation.

We show that reducing Usp53 expression modulates lineage selection in BMSCs and established bipotential ST2 stromal cells, favoring osteoblastogenesis. This suggests that USP53 normally acts to block osteoblast differentiation. It appears counter-intuitive that the induction of the expression of an anti-osteoblastogenic gene could contribute to the mechanisms propagating the anabolic effect of iPTH. This paradox may be due to a dosage effect, as neither of the shRNAs directed at Usp53 were able to completely abrogate USP53 expression (causing between 50 and 70% inhibition), although this seems unlikely. Another possibility is that the effect of USP53 on lineage selection may be influenced in vivo due to the niche environment of the stem/stromal cells31,32. During osteoblastogenesis, the impact of Usp53 knockdown on PTH responses was different depending on the maturation stage of the cells (compare Fig. 7B, E). One interpretation of this result is that the effect of USP53 downstream from iPTH is mediated by a differentiated cell type instead of through stem cells or early lineage cells. Osteocytes, which are critical for maximal iPTH anabolic responses13, represent a probable target. Finally, it remains possible that USP53 is required for the endocrine role of PTH in regulating mineral homeostasis and not for its osteoanabolic effect. Answers to these questions will be provided by phenotype characterization following inactivation of Usp53 using a floxed Usp53 allele and differentiation stage-specific Cre drivers.

This is the first report characterizing the regulation and function of Usp53 in vitro. In summary, we show that Usp53 expression is induced by the PTH-activated NACA axis in osteoblasts. We also offer a mechanism involving the NACA-JUN-CREB trio regulating Usp53 transcription. Our data strongly suggest that Usp53 regulates the lineage fate of mesenchymal cells by impacting osteoblastogenesis and adipogenesis. At the pathological level, increased marrow fat content is observed in bone loss conditions such as age-related osteoporosis33,34,35. It will prove interesting to inactivate Usp53 in specific cell types in vivo and uncover the molecular mechanism by which USP53 alters mesenchymal differentiation and function to mediate PTH responses.

Materials and methods

Plasmids, antibodies, and reagents

The wildtype-NACA, NACA-S99D mutant14, and NACA delta DBD (69–80) mutant16 expressing vectors were previously described. The PKA-dominant-negative (DN-PKA) (cAMP-unresponsive mutant of R1α regulatory subunit) expression vector pcDNA3.1-DN-PKA is a gift provided by N. O. Dulin (University of Chicago, Chicago, IL)23. The affinity-purified chicken anti-NACA15, the rabbit pan-specific anti-NACA antibody18,19, and the rabbit polyclonal anti-NACA-phosphoS99 (anti-pS99)14 antibodies have also been previously reported. Rabbit anti-USP53 (HPA035844, Prestige antibodies) was purchased from Sigma-Aldrich. Rabbit anti-GAPDH (HRP conjugate, catalog no. 8884) was purchased from Cell Signaling. PTH(1–34) was purchased from Bachem (Torrance, CA) and N6-benzoyladenosine cAMP (6Bnz-cAMP; catalog no. B009) was from Biolog (Bremen, Germany). Ascorbic acid, β-glycerophosphate, dexamethasone (Dex), 3-isobutyl-1-methylxanthine (IBMX), insulin, and hexadimethrine bromide were obtained from Sigma-Aldrich.

Primary cultures and cell lines

ST2 stromal cells (obtained from Riken Cell bank, Tsukuba, Japan) and MC3T3-E1 (subclone 4) osteoblastic cells (obtained from ATCC, Manassas, VA) were grown in minimum essential medium alpha (α-MEM, Gibco) supplemented with 10% fetal bovine serum (FBS, Hyclone, Logan, UT), 100 U/ml penicillin, and 100 µg/ml streptomycin sulfate at 37 °C in a humidified atmosphere with 5% CO2. The HEK-293 T cell line (ATCC) was cultured in Dulbecco’s modified Eagle’s medium (DMEM, Gibco) supplemented with 10% FBS under the same conditions.

Murine BMSCs were isolated from 6-week-old C57BL/6 male mice (obtained from Charles River Canada Laboratories) as previously described36. Briefly, tibia and femoral marrow compartments were used to isolate BMSCs. After cutting the epiphysis, the femurs and tibia were flushed with complete media using a 27-gauge needle. The flushed cells were then filtered using a 70 µm filter mesh to get rid of bone chips and cell clumps. Finally, the cells were resuspended in additional complete medium and cultured in DMEM supplemented with 20% FBS, 100 U/ml penicillin, and 100 µg/ml streptomycin sulfate for differentiation experiments. Neonatal murine calvarial cells were harvested from calvariae of 3-day-old WT C57BL/6 male mice. Primary osteoblasts were isolated as previously described14.

Chip Sequencing (ChIP-Seq), RNA-Sequencing (RNA-Seq), and Chromatin immunoprecipitation (ChIP)

Conventional ChIP assay, ChIP-Seq, and RNA-Seq were performed as previously described18. The ChIP-Seq against NACA has identified a peak in Chromosome 3 (3qG1); Chr3: 122,984,707-122,984,915. For ChIP-PCR, Usp53 promoter fragment was amplified using 5′-CTGAAACCCACAGAACCCCG-3′ forward and 5′-GGATTTCCACGGCAGACGC-3′ reverse primers. Reactions were performed using Power SYBR Green master mix following the manufacturer's instructions (ThermoFisher Scientific) and carried out in a QuantStudio 7 Flex Real-Time PCR System (ThermoFisher Scientific). For RNA-Seq, a cutoff of 1 in logscale was used to filter regulated genes by a twofold change. David Bioinformatics Resources 6.8 (NIAID,NIH) was used for gene ontology (GO) analysis and functional clustering 37.

The ChIP-Seq and RNA-Seq data were deposited in the Gene Expression Omnibus database of the National Institute for Biotechnology Information (Accession numbers: GSE147070 and GSE154355, respectively).

Electrophoresis mobility shift assay (EMSA)

EMSA assays were performed as described elsewhere18. HEK-293T cells were transfected with 10 µg of NACA-expressing vector14. Twenty-four hours (24 h) following transfection, cells were treated for 1 h with 100 μm 6Bnz-cAMP to induce phosphorylation of NACA at residue Serine 99 and nuclear translocation14. Oligonucleotide probes spanning the NACA binding site (NACAbds) within the Usp53 promoter region (5′-ACCCCGGCTCAGATCCCGCG-3′) or the positive control NACAbds from the Myoglobin promoter region (5′-AGGGCCAGAGAAAGACA-3′)20 were synthesized with a GG overhang. The probes were 32P-labeled using [α32P]-dCTP (PerkinElmer, Waltham, MA) and Klenow DNA Polymerase I (New England Biolabs, Ipswich, MA).

Luciferase assay

Transfections and luciferase assays were performed as previously described 19. MC3T3-E1 cells were seeded at a density of 20,000 cells/well, 14 h prior to transfection. In each well, 300 ng of Usp53 promoter reporter vector and 15 ng of Renilla luciferase internal control vector (pRL-null, Promega) were transfected using the lipofectamine LTX and PLUS Reagent kit following the manufacturer’s instructions (Invitrogen). When indicated, 100 ng of NACA-S99D mutant or NACA delta DBD (69–80) mutant and 50 ng of CREB-expressing vector (MR204788, Origene), 50 ng of JUN-expressing vector22 or 100 ng of DN-PKA expressing vector were co-transfected. Twenty-four (24) h post transfection, the Dual-Glo luciferase assay was performed according to the manufacturer’s protocol (Promega) and signal was measured using a VICTOR Nivo multimode microplate reader (Perkin Elmer, Waltham, MA). Results were calculated by normalizing Firefly relative light units to Renilla relative light units and are expressed as fold-change over the empty vector. When indicated, cells were treated with 100 nM PTH(1–34) for 24 h before luciferase measurement.

Vector cloning

Promoter cloning was performed as previously described19. Briefly, genomic DNA from MC3T3-E1 cells was used to amplify the Usp53 promoter fragment (-2325/+ 238 bp). The fragment was subcloned into the pGL4.10[LUC2] vector (Promega, Madison, WI) using KpnI and NheI restriction sites that were added during the PCR step. All generated vectors were verified by sequencing. Primer sequences and plasmids are available upon request.

Short hairpin RNA (shRNA) gene silencing

Vectors expressing the short hairpins RNAs (shRNAs) targeting Naca were generated as previously described18. Vectors expressing shRNAs targeting Usp53 (shUsp53, TRC no. 0000030914; shUsp53.1, TRC no. 0000030918) and the non-targeted control shRNA (shScr; #SHC016) were purchased from Sigma-Aldrich. For stable silencing of Naca and Usp53, lentiviral infections were performed as previously described18,19. Briefly, lentiviral particles were prepared by transfecting HEK-293 T cells with 10 µg of shRNA expressing vector, 7.5 µg psPAX2 (viral packaging vector, a gift from Dr. Didier Trono; Addgene plasmid #12260), and 2.5 µg pMD2.G (viral envelope encoding vector, a gift from Dr. Didier Trono; Addgene plasmid #12259), using the calcium/phosphate DNA-precipitation technique38. Lentivirus-enriched medium was collected 48 h post-transfection, filtered using a 0.45 µm filter (Millipore, Billerica, MA), and supplemented with 8 µg/ml hexadimethrine bromide (Polybrene, Sigma-Aldrich). The lentiviral medium was used to infect BMSCs, ST2 or MC3T3-E1 cells for 24 h. Infected cells were then cultured in the presence of 2.5 µg/ml (BMSCs and ST2) or 5 µg/ml (MC3T3-E1) of puromycin for selection. Selection was performed for 5 days.

Osteogenic differentiation

ST2 cells or MC3T3-E1 cells were plated at 50,000 cells/cm2 for differentiation assays. To induce osteogenic differentiation of ST2 cells, cells were incubated in α-MEM medium at 37 °C with the supplementation of 50 µg/ml of ascorbic acid and 100 ng/ml of BMP2 (Gibco). For osteogenic differentiation, MC3T3-E1 cells were cultured in α-MEM medium at 37 °C with the supplementation of 50 µg/ml of ascorbic acid and 10 mM of β-glycerophosphate. Medium was changed every 2 days for the duration of the experiment. When indicated, cultures were treated with 100 nM of PTH(1–34) for 2 h prior to analysis.

Adipogenic differentiation

ST2 cells were plated at 50,000 cells/cm2 for adipogenic differentiation. The cells were incubated in DMEM medium at 37 °C with IBMX (0.5 mM), dexamethasone (1 µM), insulin (1 µg/ml), and rosiglitazone (1 µM) for the first day. On day 2, differentiation medium was changed and supplemented with only insulin (1 µg/ml) and rosiglitazone (1 µM). Complete DMEM medium (without supplements) was used from day 4 until the end of the differentiation assay on day 6. When indicated, cultures were treated with 100 nM of PTH(1–34) for 2 h prior to analysis.

RNA isolation and quantitative real-time PCR

Total RNA was extracted from cells or tissues using TRIzol reagent following the manufacturer’s instructions (Invitrogen). 1–2 µg of RNA were reverse-transcribed into cDNA using the high-capacity cDNA reverse transcription kit (Thermo Fisher Scientific), following the manufacturer’s protocol. Gene expression was assessed using the Taqman Fast Advanced Master Mix (Thermo Fisher Scientific) and the following TaqMan primers: Usp53 (Mm00476778-m1), Gapdh (Mm99999915-g1), Dlx5 (Mm00438430), Alpl (Mm00475834-m1), Sp7 (Mm00504574-m1), Bglap2 (AIWR1XJ), Fabp4 (Mm00445878-m1), Cebpβ (Mm00843434-s1), Cebpα (Mm00514283-s1), Pparg2 (Mm00440940-m1), and AdipoQ (Mm00456425-m1). Expression levels of Usp53 (long and short isoforms) were assessed using the Power SYBR Green PCR Master Mix (Thermo Fisher Scientific) and the following isoform-specific , oligonucleotide primers: Usp53 short isoform, forward, 5′-GAAGTGTCCTAGTAACTGTGGCC-3′ and reverse, 5′-GAATGAAAGCAACTGTGATACCCC-3′; Usp53 long isoform, forward, 5′-CGACACAGGGATTTGGTTGATG-3′, and reverse, 5′-CAGAGGTGTAGCTCTCATGGG-3′; and Gapdh, forward, 5′-CATCACTGCCACCCAGAAGACTG-3′, and reverse, 5′-ATGCCAGTGAGCTTCCCGTTCAG-3′. All reactions were carried out in a 7500 real-time PCR system (Thermo Fisher Scientific).

Staining

Alkaline phosphatase (ALP) staining was performed on ST2 cells on day 12 of osteogenic differentiation. Staining was performed using the TRACP and ALP double-stain kit (TAKARA) following the manufacturer’s instructions. Von Kossa staining was performed on ST2 cells on day 8 of osteogenic differentiation. Cells were fixed in 10% formalin at RT and stained for mineralized matrix using 5% silver nitrate. Alizarin red staining of MC3T3-E1 cells was performed as previously described 19. Alizarin red staining was performed on MC3T3-E1 cells on day 21 of osteogenic differentiation. Briefly, cells were fixed in 70% ethanol for 1 h at − 20 °C and then incubated with alizarin red solution (40 mM, pH 4.2) (Sigma-Aldrich) for 15 min at RT. Cells were then washed with dH2O to remove excess stain. Finally, cells were washed with PBS (1x) for 15 min and photographed. ALP, Von Kossa, and alizarin red staining were quantified using ImageJ (IJ1.46r, NIH) and presented as mean percentage (± SD) of stained area / total area of 3 independent experiments (n = 3). Oil Red O staining was performed on ST2 cells on day 6 of adipogenic differentiation. Cells were fixed in 10% formalin (Sigma-Aldrich) for 45 min at RT. Following fixation, cells were rinsed with dH2O and then 60% isopropanol for 2–5 min. Cells were then incubated with Oil Red O solution for 5 min. Finally, cells were rinsed 3–4 times with dH2O and photographed. For Oil Red O quantification, 150 µl of 100% 2-propanol was added to each well and incubated for 5 min at room temperature. Optical density (OD) was then measured at 520 nm using a VICTOR Nivo multimode microplate reader (Perkin Elmer, Waltham, MA). Data was presented as mean of OD (± SD) of 3 independent experiments (n = 3).

In vivo osteogenesis assay

Animal experimentation was carried out in compliance with the ARRIVE guidelines. All animal procedures were reviewed and approved by the McGill Institutional Animal Care and Use Committee and followed the guidelines of the Canadian Council on Animal Care. For the in vivo osteogenesis assay, 2 × 106 shScr or shUsp53-knockdown BMSCs were seeded in Gelfoam absorbable collagen sponges (Pfizer) and implanted subcutaneously on the backs of 6-wk-old athymic nude mice (Crl:NIH-Lystbg Foxn1nu Btkxid; 201; Charles River Laboratories). After 4 weeks, the transplants (n = 4 for both shScr and shUsp53) were harvested and further processed for micro-CT and histological analysis. For micro-CT, transplants were fixed in 4% PFA in PBS for 48 h at room temperature and washed with 1 × phosphate-buffered saline. Transplants were then scanned with a 6-µm pixel size using a SkyScan 1272 µCT system (Bruker, Belgium). The parameters were 5 µm pixel size, 50 kV, 194 µA, 0.5 mm Al filter, angular rotation step of 0.450, and an exposure time of 500 ms with a total scan duration of 45 min. Three-frame averaging was used to improve the signal-to-noise ratio. After scanning, 3-dimensional (3D) microstructural image data were reconstructed using the manufacturer’s software (SkyScan NRecon). For histology, transplants were then decalcified in EDTA and embedded in paraffin. Paraffin blocks were sectioned (6 µm) and stained with H&E or incubated with specific primary antibodies against Osterix (1/100; ab22552; Abcam) or Perilipin (1/200; ab3526; Abcam). Prior to immunostaining, sections were pretreated with citrate buffer (pH 6.0) and heated for 10 min at a temperature between 95 °C and 100 °C for antigen retrieval. Primary antibodies were detected using Alexa Fluor 594 conjugated goat anti-rabbit IgG (H + L) (1/250; Invitrogen). Nuclei were stained with ProLong Gold antifade reagent with DAPI (Thermo Fisher Scientific). Images were acquired using a Leica DMR fluorescence microscope (Leica Microsystems, Wetzlar, Germany) connected to a digital DP70 camera (Olympus, Center Valley, PA). Image processing included whole image channel filtering to remove noise and whole image adjustment of brightness, contrast, color balance, and sharpening using Adobe Photoshop v. 12.1. Photoshop images were then flattened and imported into Adobe Illustrator v. 15.1 to build the final montages.

Western blotting

Protein extraction and immunoblotting were performed as previously described19. Whole cell extracts were prepared from MC3T3-E1, ST2, and BMSC cells. Cells were washed twice with ice-cold PBS (1x) and scraped with lysis buffer (300 mM NaCl, 50 mM HEPES, pH 7.6, 1% Triton X-100, and protease/phosphatase inhibitor cocktail (Cell Signaling)). Cells were then kept on ice for 10 min and briefly sonicated with 1 or 2 pulses for 15 s using an Ultrasonic dismembrator model 500 (Thermo Fisher Scientific, Waltham, MA). Following sonication, cells were spun down at 4 °C for 15 min at maximum speed. Supernatant was aliquoted and stored at − 80 °C. For western blots, equal amounts of cell extracts (10 µg protein) were resuspended in 6X Laemmli buffer. The samples were boiled for 5 min and run on a denaturing SDS‐PAGE for 1.5 h then transferred to a PVDF membrane (Amersham, UK). The membrane was blocked for 45 min in 5% non-fat dry milk. After blocking, the membrane was incubated with the primary antibody diluted 1:1000 in 5% non-fat dry milk, overnight at 4 °C. The membrane was later incubated with the secondary antibody conjugated with horseradish‐peroxidase, anti‐mouse, or anti‐rabbit‐ HRP, diluted 1:20,000. Development was done using the Western Lightening Chemiluminescence Kit (Perkin Elmer, Akron, OH, USA) and exposure to X-Ray film. The protein bands were visualized on the developed film. Western blot analysis and quantification was performed with ImageJ (IJ1.46r, NIH). Uncropped original blots are determined to be too large to place in multi-panel figures but are presented in full-length in Supplementary Figures S4–S8.

Statistical significance

Data are presented as means ± one standard deviation (SD). Comparisons were made by analysis of variance (ANOVA) with Bonferroni’s post hoc test , or unpaired t tests. A probability (P) value lower than 0.05 was accepted as significant. For each panel, three independent experiments were performed, each in triplicates.