BMP signaling-driven osteogenesis is critically dependent on Prdx-1 expression-mediated maintenance of chondrocyte prehypetrophy

Author’s Accepted Manuscript BMP signalling-driven osteogenesis is critically dependent on Prdx-1 expression- mediated maintenance of chondrocyte prehypetrophy Yogesh Kumar, Tathagata Biswas, Gatha Thacker, Jitendra Kumar Kanaujiya, Sandeep Kumar, Anukampa Shukla, Kainat Khan, Sabyasachi Sanyal, Naibedya Chattopadhyay, Amitabha Bandyopadhyay, Arun Kumar Trivedi

PII: DOI: Reference:

www.elsevier.com

S0891-5849(18)30071-6 https://doi.org/10.1016/j.freeradbiomed.2018.02.016 FRB13625

To appear in: Free Radical Biology and Medicine Received date: 3 November 2017 Revised date: 29 January 2018 Accepted date: 10 February 2018 Cite this article as: Yogesh Kumar, Tathagata Biswas, Gatha Thacker, Jitendra Kumar Kanaujiya, Sandeep Kumar, Anukampa Shukla, Kainat Khan, Sabyasachi Sanyal, Naibedya Chattopadhyay, Amitabha Bandyopadhyay and Arun Kumar Trivedi, BMP signalling-driven osteogenesis is critically dependent on Prdx-1 expression- mediated maintenance of chondrocyte prehypetrophy, Free Radical Biology and Medicine, https://doi.org/10.1016/j.freeradbiomed.2018.02.016 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting galley proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

BMP signalling-driven osteogenesis is critically dependent on Prdx-1 expressionmediated maintenance of chondrocyte prehypetrophy Yogesh Kumar11, Tathagata Biswas31, Gatha Thacker1, Jitendra Kumar Kanaujiya1, Sandeep Kumar3, Anukampa Shukla1, Kainat Khan2, Sabyasachi Sanyal1, Naibedya Chattopadhyay2, Amitabha Bandyopadhyay3*, Arun Kumar Trivedi1* 1

Biochemistry Division, CSIR-Central Drug Research Institute (CSIR-CDRI), Sector-10,

Jankipuram Extension, Lucknow, 226031, UP, India 2

Division of Endocrinology and Center for Research in Anabolic Skeletal Targets in Health

and Illness (ASTHI), CSIR-Central Drug Research Institute (CSIR-CDRI), Sector-10, Jankipuram Extension, Lucknow, 226031, UP, India 3

Department of Biological Sciences and Bioengineering, Indian Institute of Technology

Kanpur 208016, India [email protected] [email protected] #

Correspondence to.

Abstract During endochondral ossification, cartilage template is eventually replaced by bone. This process involves several well characterised, stereotypic, molecular and cellular changes in the cartilage primordia. These steps involve transition from resting to proliferative and then prehypertrophic to finally hypertrophic cartilage. BMP signaling is necessary and sufficient for

1

These authors contributed equally. 1

osteogenesis. However, the specific step(s) of endochondral ossification in which BMP signaling plays an essential role is not yet known. In this study we have identified Prdx1, a known scavenger of ROS, to be expressed in pre-hypertrophic chondrocytes in a BMP signalling-dependent manner. We demonstrate that BMP signaling inhibition increases ROS levels in osteogenic cells. Further, Prdx1 regulates osteogenesis in vivo by helping maintenance of Ihh expressing pre-hypertrophic cells, in turn regulating these cells’ transition into hypertrophy. Therefore, our data suggests that one of the key roles of BMP signaling in endochondral ossification is to maintain pre-hypertrophic state. Graphical abstract

Key Words: Peroxiredoxin-1, BMP Signaling Pathway, Endochondral Ossification, ROS, Osteogenesis

2

Introduction

Condensation of loose mesenchymal cells in the limb bud marks the first event of endochondral ossification during appendicular skeletal morphogenesis. The condensed mesenchymal cells undergo chondrogenesis. These chondrocytes express type II collagen. At the centre of this cartilage anlagen, chondrocytes undergo hypertrophic differentiation. The hypertrophic cells secrete type X collagen rich extra cellular matrix (ECM). Concurrent to hypertrophic differentiation the anlagen gets vascularized, subsequently the cartilage template is replaced by bone, which has its own type I collagen rich ECM (Karsenty et al., 2009; Karsenty and Wagner, 2002). PTHrP and Ihh are two critical regulators of the rate of hypertrophic differentiation. PTHrP is secreted by the periarticular perichondrium. As long as chondrocytes are within the zone of diffusion of PTHrP they do not express Ihh. The cells beyond the zone of PTHrP diffusion turn-on the expression of Ihh, exit proliferation and are termed pre-hypertrophic. These prehypertrophic cells eventually turn down expression of type II collagen and Ihh and turn-on type X collagen to mark the onset of hypertrophic differentiation. This transition from prehypertrophy to hypertrophy is a critical step in endochondral ossification. PTHrP and Ihh are linked through a negative feedback loop (Kronenberg, 2003; Lanske et al., 1996). Any molecular manipulation of PTHrP or Ihh leads to severe skeletal defects. Over-expression of either PTHrP (Weir et al., 1996) or Ihh (Vortkamp et al., 1996) leads to delayed hypertrophy and thereby delayed ossification. On the other hand, knockout of either PTHrP (Kronenberg, 2006) or Ihh (St-Jacques et al., 1999) though initially causes a delay in the onset of hypertrophy, but eventually results in massively accelerated hypertrophy thereby depleting the pool of proliferative cells and a reduction in the size of skeletal elements. Taken together, these data suggest chondrocyte hypertrophy is intricately linked to osteogenesis. While it is clear that PTHrP negatively regulates Ihh expression in the proliferating chondrocytes, the 3

molecular mechanism responsible for maintenance of Ihh expression in the pre-hypertrophic cells and/or downregulation of the same in hypertrophic cells is not understood. Recently, several studies implicated the role of reactive oxygen species (ROS) in osteogenic differentiation. Osteogenic induction in human bone marrow stromal cells (hMSCs) was associated with marked fall in intracellular ROS levels and an increase in the expression of few anti-oxidative enzymes. In fact, higher levels of ROS inhibited osteogenic differentiation of hMSC cells (Atashi et al., 2015; Chen et al., 2008). Further, increasing oxidative stress in osteoblasts led to reduced expression of osteogenic differentiation markers (Atashi et al., 2015; Mody et al., 2001). Interestingly, in contrast to osteoblast differentiation, higher ROS levels stimulated chondrocyte hypertrophy. Not only did the hypertrophic chondrocytes have high levels of ROS, but also reducing oxidative stress led to inhibition of hypertrophy (Morita et al., 2007). It is yet unknown which of the major signaling pathways regulating osteogenesis impact ROS levels during endochondral ossification. BMP signaling pathway is a major regulator of skeletal development (Bandyopadhyay et al., 2013). It is known that BMP signaling is both necessary (Bandyopadhyay et al., 2006) and sufficient (Urist, 1965) for osteogenesis. The role of BMP signaling in regulating ROS levels during endochondral ossification is yet to be established. In fact, despite the critical importance of this signaling pathway in osteogenesis, only a few genes have been identified (Prashar et al., 2014) that are expressed in a BMP signaling dependent manner during osteogenesis. Here we identified Prdx1, a critical ROS scavenger, as a gene expressed in a BMP signaling dependent manner in pre-hypertrophic chondrocytes. Expression of Prdx1 in several osteogenic cell lines, in vivo in rats and developing chick embryos co-relates with osteogenesis. Our data suggest that BMP signaling regulates transcription of Prdx1. Our invivo data suggests that during endochondral ossification in chick embryos removal of ROS by 4

Prdx1 is critical for maintaining the pre-hypertrophic state and proper transition to hypertrophy requires downregulation of Prdx1. Materials and Methods Cell culture and Transfection BRITER and MC3T3E-1 cells were cultured in α-MEM (Sigma) supplemented with 10% FBS, 1% Anti-Anti and 1% MEM non-essential amino acid solution. Transfection in BRITER and MC3T3-E1 was performed by Lipofectamine LTX and Plus reagent (Invitrogen) as per manufacturer protocol. Chemical and antibodies 4-Hydroxytamoxifen (4-OHT) was purchased from Sigma (H6278). Antibodies used for immunoblotting; Prdx1 (sc-7381), Prdx2 (sc-23967), Prdx3 (sc-59661), Prdx4 (sc-23974), β-Tubulin (sc-9104) were procured from Santa cruz Biotechnologies while Anti-Runx2 (ab23981) and antiBmp2 (ab6285) was purchased from Abcam. β-actin (A-3854) was purchased from Sigma and pSmad1/5 (9516S) from Cell Signalling Technologies. 2-D gel electrophoresis 5 X106 BRITER cells were plated in 150 mm tissue culture plate. Next day (70-80 % confluency) cells were treated with 1µM of 4-hydoxytamoxifen for 24 hrs. Post treatment, old media removed and fresh α-MEM with 1% FBS were added and treated with 1µM of 4-hydoxy-tamoxifen for 12 hrs for knockdown of BMP (Bmp2 and Bmp4) and untreated cells. Cells were harvested by trypsinization followed by washing with 1X PBS twice. Cell pellets were dissolved in 10 mM Tris (pH=7.6) buffer supplemented with protease and phosphatase inhibitors, and incubated in ice for 30 min with occasional tapping. Cell lysates were centrifuged at 4 0C for 30 min at 16000 rpm and supernatant were collected (Tris Lysate). Remaining pellets were dissolved in Urea lysis buffer (7M Urea, 2M Thiourea, 4% CHAPS and 65mM DTT) supplemented with protease and phosphatase inhibitors. Cells were incubated for 1 hr at room temperature followed by centrifugation at 21 0C for 1 hr at 20000 rpm and supernatant were collected (Urea lysate). Total protein concentration was determined by Bradford 5

reagent (Bio Rad). Isoelectric focusing (IEF) was performed using 1.5 mg of protein lysate on Immobiline DryStrip, strips (18 cm, pH 3–10, GE healthcare). For the first dimension, the lysates were mixed with rehydration solution (7 M Urea, 2 M Thiourea, 4% CHAPS 65 mM DTT) in addition to 0.5% IPG buffer (Amersham Pharmacia Biotech). After IEF IPG strips were reduced with 2% DTT in SDS-equilibration buffer and alkylated with 2.5% IAA in SDS-equilibration buffer, followed by 2nd dimension electrophoresis on 12% polyacrylamide gel. Gels were stained with Coomassie Brilliant Blue (R250) to visualize the differentially expressed spots. The gels were scanned with Image scanner III (GE Healthcare) and quantitatively analyzed with Image Master 2D Platinum 7.0 software (GE Healthcare). Spots were manually selected from each gel and then analyzed in gel analysis software. For each protein spot, both from respective treated and untreated gel, % volume was calculated and then fold % volume was calculated by dividing % volume of treated spot with untreated. Fold volume of >1.0 are upregulated, whereas <1.0 are downregulated. Gel-to-gel and staining technique variations were overcome by comparing each protein spot to a number of other protein spots in the same gel whose expression did not change under a given experimental condition. Subsequently, differentially appearing protein spots from 4-OHT treated gel as compared with untreated condition were detected and analyzed. Protein identification and MALDI-TOF/TOF analysis Selected spots were carefully excised from the gels, destained with 50mM ammonium bicarbonateacetonitrile (1:1) for 15 min followed by 2 alternating washing steps with 50% acetonitrile and 50mM ammonium bicarbonate. The gel pieces were dehydrated at room-temperature with 100% acetonitrile and covered with 10μl of trypsin (working solution 5ng/μl prepared from 100ng/μl stock solution in 50mM ammonium bicarbonate) overnight at 37°C. The peptides were eluted in 10μl of 70% acetonitrile. The eluates were dried using a vacuum centrifuge and stored at −80°C. The peptides were resuspended in 5 μl of 20% acetonitrile and 0.1% TFA and sonicated for 3 min before processing for mass spectrometry. MALDI-TOF/TOF data was analyzed with ProteinPilotTM Software (AB Sciex) using MASCOT search engine. The peptide mass data were analyzed for corresponding protein 6

matching in the Swiss-Prot database with settings of peptide mass tolerance: ±150ppm, fragment mass tolerance: ±0.5Da and MAX missed cleavages set as 1 in MS/MS ion search using MASCOT search engine. Proteins that could be identified with more than 2 peptides had a % confidence index >99%, while the one with one peptide had a % confidence index >95%. Probability-based MOWSE scores were estimated by comparison of search results against estimated random match population and were reported as 10*LOG10 (p), where p is the absolute probability. Scores greater than 40 were considered significant (p < 0.05). All protein identifications were in the expected size range based on position in the gel. Osteoblast culture Mice calvarial osteoblasts (mOB) were obtained by using sequential digestion method as previously described (Khan et al., 2015). Calvaria of 1 to 2 days-old mice pups (both sexes) were pooled and were cleaned by removing the sutures and adherent mesenchymal tissues. Further, calvaria were subjected to four or five sequential (10 min each) digestions at 37C in α-MEM with 0.1% dispase and 0.1% collagenase P. Cells collected from the second to fifth digestions were plated in T-25 cm2 culture flask in α-MEM containing 10% FBS and 1% penicillin/streptomycin (complete growth medium). Cultures of mOB with 70%-80% confluence were trypsinized (Sigma) and plated as per experiment. Mineralized nodule formation and Alizarin staining 60%-70% confluent MC3T3-E1, a mouse calvarial cell line was plated in osteoblast growth media and transfected either with siPrdx1 or siControl (siCtrl). After 72 hours of transfection, cells were cultured in complete osteoblast growth media (OBM) and differentiation induction media (DIM) containing α-MEM with 10mM β–glycerophosphate, dexamethasone (10-7 M) and 50µg/mL ascorbic acid supplemented with 10% FBS and cultured for 14 days with change of medium every alternative day. When the experiment was terminated, cells were fixed with 4% paraformaldehyde. Alizarin red-S was used for staining of mineralized nodules followed by extraction of the stain using 10% CPC (cetylpyridium chloride) for colorimetric determination of the dye at 550 nm. 7

Luciferase reporter assay Luciferase assay was performed as described previously (Kumar et al., 2015; Thacker et al., 2016). Briefly TVA-BMSc or MC3T3-E1 cells were plated one day before transfection at the density of 1 × 105 cells/well. Next day, cells were transfected with expression plasmid for Prdx-luc promoter. 24 h post transfection, cells were treated with BMP2 for indicated time points, lysed in mammalian cell lysis buffer (200 mM Tris, pH 8.0 and 0.1% Triton X-100) and assayed for luciferase activity using luciferase assay reagent (Promega, Madison, WI). Data are presented as means of triplicate values obtained from representative experiments. ROS level detection BRITER cells were plated overnight then treated with 1µM 4-OHT and cultured at 37°C for 48 h in the presence or absence of 100ng Bmp2 as indicated. On completion of indicated treatment, cells were washed with 1X PBS and loaded with 10μM 2′, 7′-dichlorofluorescein diacetate (DCFH-DA) at 37°C for 30 min in complete darkness. After washing with PBS for three times, ROS levels were determined by measuring the fluorescence intensity at excitation wavelength 485 nm and emission wavelength 530 nm. Chicken embryos Fertilized White Leghorn Chicken eggs were procured from Ganesh Enterprises, Nankari, Kanpur, UP. The eggs were incubated in a humidified chamber maintained at 38C. Eggs were electroporated and subsequently harvested at specific stages of development assessed by Hamburger and Hamilton staging criteria (Hamburger and Hamilton, 1992). Tissue processing and immunohistochemistry After harvesting the chicken embryo, limbs were dissected and fixed overnight in 4% paraformaldehyde at 4C. Dissected limbs were then embedded in paraffin, cut into 5-7 µm sections and finally stained with mouse anti-collagen type X antibody (X-AC9; DSHB) following the protocol described previously (Schmid and Linsenmayer, 1985). Following another previously described 8

protocol (Chen and Cepko, 2002), the extent of the viral infection was detected using 3C2 monoclonal antibody immunoreactivity. In vivo electroporation constructs The avian retroviral vector RCASBP(A) (Hughes et al., 1987) was used to deliver both knockdown and overexpression constructs for Prdx1. The miRNAs targeting the chicken Prdx1 mRNA (Accession no. NM_001271932) were designed using the BLOCK-iT RNAi Designer (Invitrogen). Two pairs of oligos containing miRNA target sites were cloned into the pRmiR shuttle vector (Smith et al., 2009) separately (Table S2). These constructs were named as pRmiR-gga-mirPrdx1-1 and pRmiR-gga-mirPrdx1-2, the pRmiR vector expresses GFP protein. The efficacy of the miRNAs against chicken Prdx1 was verified by cloning the miRNA target regions for chicken Prdx1 in pCAGmCherry expression vector downstream of the coding sequence of mCherry (pCAG-mCherry-ggaPrdx1 sensor) and were co-transfected with pRmiR-gga-mirPrdx1-1 and pRmiR-gga-mirPrdx1-2 into HEK293T cells. pRmiR-gga-mir-Lacz, targeting Lacz mRNA, was used as control pRmiR. The expression of mCherry to GFP was analyzed for knockdown efficiency and the miRNA with maximum knockdown efficiency (Fig. S2) was chosen for downstream applications. The construct targeting chicken Prdx1 relatively better (pRmiR-gga-mirPrdx1-2) was named as pRmiR-ggamirPrdx1 (Fig. S2 c,d). This cassette of hairpin was then sub-cloned into the ClaI site of RCASBP(A) avian retroviral vector, henceforth referred to as RCASBP(A)-gga-mirPrdx1. For overexpression, nterminal FLAG tagged hPrdx1 (Jung et al., 2001) was cloned into RCASBP(A) and was named RCASBP(A)-gga-FLAG-hPrdx1. RNA in situ hybridization RNA in situ hybridization was performed as described previously (Singh et al., 2016). cDNA clones used to make digoxigenin labelled antisense probes generated by in vitro transcription are detailed in Table S2. RNA in situ hybridization signal for a given probe on the control and the test sections were developed on the same slide. In ovo electroporation 9

A window was made upon the egg shell for visualizing the embryo after lowering the embryos by removing 2-3ml of albumin. At stage HH14, the vitelline membrane was removed near the presumptive hind limb region of the embryo and bathed in 80-100µl sterile PBS + Pen Strep solution (ThermoFisher Scienific, Cat. no: 10378016). Thereafter, RCAS constructs mixed with 0.5µg/µl pCAGGS-GFP and 0.1% fast green was injected into the embryonic space between the somatic LPM and splanchnic LPM at a concentration of 2µg/µl using a microinjector. Electroporation was performed as described (Suzuki and Ogura, 2008). DHE staining Dihydroethidium (D7008 from Sigma Aldrich) was used for staining TVA-BMSC cells to detect presence of ROS. Live cells were stained with 25µM of DHE (dissolved initially to 10mM with chloroform) without fixation. After 30mins incubation at 37C, the cells were fixed for 5-minutes in paraformaldehyde and imaged immediately under fluorescent microscope. Skeletal prep of chicken embryo Harvested chicken embryos were eviscerated and thereafter fixed in 95% ethanol for 48hrs followed by overnight 100% acetone treatment. Then, the embryos were stained for 2-3 days in a solution constituting 1 volume of 0.3% alcian blue 8GX (Sigma-Aldrich): 1 volume of 0.1% alizarin red in 70% ethanol: 1 volume of glacial acetic acid: 17 volumes of ethanol. Post staining, 1% potassium hydroxide was used to clear the embryos and then imaged. Data Analysis and Statistics All results were expressed as mean±SEM. All data were analyzed using GraphPad Prism 5.0 (GraphPad, San Diego, CA). One-way ANOVA followed by Tukey’s multiple comparison tests was used to analyze data involving more than two groups. Results Proteomic analysis identifies Prdx1 as a candidate BMP signalling-dependent protein expressed in osteoblasts 10

To identify proteins that are expressed in osteoblast cells in a BMP signaling-dependent manner we took advantage of BRITER (BMP Responsive Immortalized Reporter) cells. BRITER is a stable cell line generated from immortalized calvarial osteoblast cells from a tamoxifen inducible Bmp2; Bmp4 double conditional knockout mouse (Yadav et al., 2012). BRITER cells can be depleted of endogenous Bmp2 and Bmp4 proteins by treating it with 4-OHT in vitro. As described before (Yadav et al., 2012), treatment of BRITER cells with 4-OHT dramatically reduced the abundance of Bmp2 and Bmp4 mRNAs (Fig. 1 a) as well as Bmp2 protein level (Fig. 1 b). We also observed absence of pSmad 1/5 in Bmp2/4 knockdown condition (Fig. 1 b). To assess global expression changes and identify targets of BMP signaling in osteoblasts, we applied 2-DE based proteomic approach as described in the “Materials and methods” section. Tris and urea lysates of 4-OHT treated and untreated BRITER cells were prepared and resolved on 2-DE. Gels were then stained with Coomassie Brilliant Blue to visualize the protein spots (Fig. 1 c [Tris lysate], Fig. 1 d [Urea Lysate]) followed by analysis of the scanned gel picture in Image scanner III (GE Healthcare) for fold volume changes. Differentially expressed proteins from both the conditions were subjected to in-gel trypsin digestion and subsequent mass spectrometry analysis. We identified 15 spots from trislysate and 14 spots from urea-lysate (Table 1) gels that showed highly significant MOWSE score (score greater than 40 were considered significant, p <0.05) and were more than 30% up or downregulated. We also identified other proteins which showed marginal changes in their expression upon Bmp2/4 depletion (Supplementary Fig. S1; Table S1).

11

Figure 1. Tamoxifen induced Bmp2/4 knockdown leads to down regulation of Prdx1: a-c) BRITER cells treated with 1 µM 4-OHT for 36 h (treated for 24 h followed by an additional 12 h after changing the media). a) mRNA levels of Bmp2 and Bmp4 at 24 h were analysed by real time PCR. b) Corresponding protein lysates were immunoblotted with anti-Bmp2 and anti-pSmad 1/5 antibodies. βactin level served as loading control. WCE of HEK293T and MC3T3-E1 were loaded as negative and 12

positive controls, respectively. c,d) 2DE of tris lysate (upper panel) and urea lysate (lower panel) of BRITER cells treated with 1 µM 4-OHT and differentially expressed proteins in treated (RHS) vs. untreated condition (LHS) were marked with arrow and identified. Image in inset shows magnified area from the respective gel. We observed Bmp2/4 depletion led to peroxiredoxins downregulation, particularly peroxiredoxin-1 (Prdx1) which showed maximum downregulation (almost absent) in 4-OHT treated BRITER cells (Spot no. 1 in Fig. 1 c). Prdx1 is expressed in a BMP signalling-dependent manner in osteoblasts and chondrocytes The fact that Prdx1 was absent in 4-OHT treated BRITER cells suggested that this gene was expressed in a BMP signaling-dependent manner. To test this possibility, we used different cell culture systems as well as chick embryonic skeletal system. As expected, treatment of BRITER cells with 4-OHT resulted in downregulation of Bmp2 and Runx2 (Fig. 2 a). Prdxs are a family of small (22-27 kDa) ubiquitous antioxidant enzymes involved in sensing and detoxifying ROS. At present, there are six known mammalian isoforms of Prdx proteins that can be divided into 3 subclasses of which Prdx1-4 belong to typical 2-cysteine (2-cys) Prdx group (Hall et al., 2011; Neumann et al., 2009). As we also identified Prdx2 and Prdx4 along with Prdx1 to be differentially regulated in our 2DE screen (Table S1), we examined changes in the expression of peroxiredoxins of the 2-cys Prdx group proteins through western blot analysis. Interestingly, while the abundance of Prdx1 and Prdx2 diminished following 4-OHT treatment, whereas Prdx4 remained unchanged and Prdx3 increased marginally (Fig. 2 a). Since Prdx1 seems to be the most widely expressed member of the Prdx family, displays the highest abundance in various tissues (Immenschuh and Baumgart-Vogt, 2005) and undergoes maximum downregulation upon depletion of Bmp2/4, in all subsequent experiments we have only investigated the expression level of Prdx1. Earlier, we have described establishment of a trilineage competent immortalized mouse BMSC cell line, TVA-BMSC (Yadav et al., 2016). In this cell line the level of BMP signaling can be attenuated by treating the cells with 4-OHT. Within 24 h of 4OHT treatment the abundance of Runx2, the osteoblast transcription factor and Prdx1 proteins were significantly reduced and the expression levels didn’t recover during the 72 h observation window (Fig. 2 b). In addition, time dependent upregulation of Runx2 and Prdx1 expression was observed in 13

bone marrow stromal cells isolated from rat femurs when cultured in osteogenic media for 21 days (Fig. 2 c), suggesting the involvement of Prdx1 with osteoblast differentiation. Similarly, treating MC3T3-E1 (Fig. 2 d) and mOB (Fig. 2 e) with recombinant Bmp2 resulted in upregulation of Prdx1 and Runx2. Endochondral ossification is a culmination of multiple steps involving obligatory chondrogenic differentiation and eventual hypertrophy. Because in vitro osteogenesis of progenitor or stromal cells does not mimic endochondral ossification, therefore to determine the stage of endochondral ossification at which the role of Prdx1 is required, we decided to investigate the skeletal cell type expressing Prdx1 in vivo. For this purpose, we conducted RNA in-situ hybridisation on sections of chick embryonic limb skeleton. We observed a basal level of Prdx1 mRNA expression in the proliferating chondrocytes of HH34 chicken tibia with specifically higher level of expression in the pre-hypertrophic chondrocytes (Fig. 2 f). Interestingly, no Prdx1 mRNA expression could be detected in the hypertrophic chondrocytes (Fig. 2 g). To investigate the dependence of Prdx1 mRNA expression in vivo on the level of BMP signaling, we used a previously reported miRNA construct, RCASBP(A)-gga-mirBmp2/mirBmp4 (Prashar et al., 2014), that targets chick Bmp2 and Bmp4 simultaneously. This retroviral construct was electroporated into HH14 chicken embryo and infection was detected by immunohistochemistry against one of the viral coat proteins using 3C2 antibody. We observed that a patch of pre-hypertrophic cells within HH34 chick tibia that is infected by RCASBP(A)-gga-mirBmp2/mirBmp4 (Fig. 2 h, box h´) has reduced the level of Prdx1 expression (Fig. 2 i, box i´) while pre-hypertrophic cells that were not infected by the virus (Fig. 2 h, box h´´) continued to express Prdx1 (compare boxes, i´ and i´´ in Fig. 2 i). Though we originally discovered Prdx1 to be expressed in an osteoblast cell line in a Bmp signalling-dependent manner, our data suggests that it may actually function in the terminal differentiation of chondrocytes which is essential for subsequent osteogenesis to occur.

14

Figure 2. Prdx1 expression is BMP signaling dependent: a) Lysates of BRITER cells treated with 1 µM 4-OHT were resolved and immunoblotted with antibodies against Bmp2, Runx2, Prdx1, Prdx2, Prdx3 and Prdx4. b) Knockdown of Bmp2/4 in TVA-BMSCs were performed by addition of 4-OHT. Lysates for indicated time points (post-4-OHT addition) were immunoprobed with Runx2 and Prdx1 antibodies; V = vehicle treated. c) Lysates of rat bone marrow stromal cells (isolated from femur bone marrow) cultured in DIM for indicated time points were resolved and immunoprobed with Runx2 and Prdx1 antibodies. d) MC3T3-E1 cells were treated with 100ng/ml Bmp2 for indicated time points and protein lysates were resolved and immunoblotted with Runx2 and Prdx1 antibodies. β-actin level served as loading control for (a-d). e) Lysates of mOB treated with 100ng/ml Bmp2 for 48 h were resolved on 10% SDS-PAGE and immunoprobed with Runx2 and Prdx1 antibodies. β-Tubulin level served as loading control. f) Endogenous expression of Prdx1 mRNA in developing chicken bone. g) Marked region from (f), showing strong expression of Prdx1 mRNA in the pre-hypertrophic region h) 3C2 antibody staining demarcating the regions of RCASBP(A)-gga-mirBmp2/mirBmp4 infection. Box h´ marks the infected patch of pre-hypertrophic cells, while box h´´ marks the uninfected patch of pre-hypertrophic region. i) Prdx1 mRNA expression upon downregulation of Bmp2/4 in the developing chicken bone, box i´ shows Prdx1 mRNA expression in the region corresponding to box h´, while box i´´ shows the same for the corresponding uninfected region of box h´´. Scale bars represents 100µm.

Prdx1 transcription is regulated by BMP signaling 15

We next investigated whether BMP signaling stimulated transcription of Prdx1. For this purpose, we used a well characterized Prdx1 reporter construct (Shiota et al., 2008). We introduced this Prdx1-Luc construct in TVA-BMSC or MC3T3-E1 cells. In both cell types, recombinant Bmp2 (25 nM) doubled the firefly luciferase activity. Also, the stimulation by Bmp2 was more pronounced at an earlier (2 h) rather than the later time point (6 h) (Fig. 3 a and b).

Figure 3. Bmp2 transcriptionally activates Prdx1: a) TVA-BMSC and b) MC3T3-E1cells were transfected with Prdx1-luc reporter construct and treated with increasing doses of Bmp2 for indicated time points. Cells were lysed and luciferase activity was measured. Results are given as standard error of mean (±S.E.M.); *P<0.05, **P<0.001, ***P<0.0001. One-way ANOVA with Turkey’s multiple comparison test was performed using GraphPad Prism Version 5.00. The bars marked, TVABMSC/MC3T3-E1 – untransfected cells, pcDNA3 – empty vector control and Prdx1-luc - TVABMSC/MC3T3-E1 cells transfected with Prdx1-luc.

16

Prdx1 expression is positively associated with osteogenesis Runx2, a key transcription factor for skeletal differentiation is primarily expressed in the hypertrophic chondrocytes in vivo and is essential for endochondral ossification. As shown in Fig. 2 c, bone marrow stromal cells isolated from femurs undergoing differentiation displayed time-dependent upregulation of Prdx1 as well as Runx2, suggesting that Prdx1 expression was positively correlated with osteoblast differentiation. In keeping, dexamethasone treatment suppressed expression of both Runx2 and Prdx1 in MC3T3-E1 (Fig. 4 a) and TVA-BMSC (Fig. 4 b) cell lines. Notably, dexamethasone is known to inhibit osteoblast differentiation. Bone formation is significantly impaired under bone loss conditions including estrogen (E2) deficiency and lactation due to reduced osteoblast function. We studied the expression of Prdx1 under osteopenic (OVX and lactation) conditions in rats. As expected, Runx2 expression was reduced in both osteopenic conditions and so was that of Prdx1. E2 supplementation to OVX rats (to counter osteopenia) restored the inhibitory effect OVX had over the expression of Prdx1 (Fig. 4c).

17

Figure 4. Prdx1 is down regulated during conditions of bone loss: a) Immunoblot evaluation of Runx2 and Prdx1 in lysates of MC3T3-E1 cells treated with 0.1 µM dexamethasone for indicated time points. b) Immunoblot evaluation of Runx2 and Prdx1 in lysates of TVA-BMSC cells treated with 0.1 µM dexamethasone for indicated time points. c) Protein extracts isolated from femurs (devoid of bone marrow) of various groups of SD rats (sham, ovary intact; OVX, ovariectomized and maintained for 3 months; OVX+E2, OVX supplemented with 17-β estradiol (5 g/kg/d s.c.) and lactating females (10 days after parturition) were immunoblotted with Runx2 and Prdx1 antibodies. d) WCE of MC3T3-E1 transfected with either 25nM, 50nM of siPrdx1 or 50nM siControl were immunoblotted with Prdx1 and β-actin antibodies to verify specific knockdown. a-d) β-actin level served as loading control. e) MC3T3-E1 cells were transfected with siPrdx1 or siControl. 72 h post-transfection, media was changed and cells were allowed to differentiate for 14 days in DIM or osteoblastic medium (OBM). Alizarin red staining was performed and visualized under bright field microscope (20X; Leica) for analysis. Mineralization was measured using 10% CTC. Absorbance was measured at 595nm. Values represent mean ± SD (n=3).

We next investigated whether, Prdx1 was necessary for osteoblast differentiation. For this purpose, we performed siRNA-mediated knockdown of Prdx1 expression in MC3T3-E1 cells and validated the same using western blotting as indicated in Fig. 4 d. Osteoblast differentiation assessed by nodule 18

formation (stained with alizarin red) was significantly reduced in cells transfected with Prdx1 siRNA (siPrdx1, Fig. 4e). Inhibition of BMP signaling leads to increased ROS generation Prdx1 is a known scavenger of ROS. However, to the best of our knowledge, the role of BMP signaling in regulating ROS level during osteogenesis has never been investigated. We therefore assessed ROS level upon depletion of Bmp2 and Bmp4 in the BRITER cells. 4-OHT treatmentinduced depletion of Bmp2 and Bmp4 in BRITER cells led to 2.5 fold increases in the ROS level, which was more than the ROS level attained by H2O2 treatment (Fig. 5 a). This dramatic increase in ROS can be reversed by addition of recombinant Bmp2 to Bmp2/4 depleted TVA-BMSCs (Fig. 5 a). As expected, upon 4-OHT treatment, level of Prdx1 decreased in BRITER cells and was rescued upon recombinant Bmp2 treatment (Fig. 5 b). We have also investigated the level of ROS in live TVABMSC cells using DHE ROS detection assay. Clearly, upon 4-OHT treatment, bright red fluorescence was visible in the nucleus exhibiting the presence of ROS (Fig. 5 c) – in absence of which DHE failed to localize in the nucleus and hence we observed the absence of red puncta in the nuclei of 4-OHT untreated TVA-BMSC cells (Fig. 5 d).

19

Figure 5. Reduction of BMP signaling level causes increase in ROS level and concomitant decrease in Prdx1 abundance: a) ROS level in BRITER cells were measured either upon treatment with 1µM 4-OHT or upon treatment with 100ng Bmp2 for 48 h post 4-OHT treatment. Control = untreated; vehicle = ethanol; H2O2 = positive control. b) Western blotting of protein lysates of mOB treated with vehicle or 1µM 4-OHT with Prdx1 antibody. Similar blot for BRITER cells treated with Vehicle or 1µM 4-OHT or 1µM 4-OHT+100ng Bmp2. β-actin probed as loading control. c) Untreated or d) 4-OHT treated TVA-BMSC stained with dihydroethidium. Scale bars represents 50µm. Removal of ROS is critical for maintenance of pre-hypertrophic state while mis-expression of Prdx1 blocks transition from pre-hypertrophic to hypertrophic state Next, we investigated the role of Prdx1 using developing chicken skeletal system. Prdx1 is expressed at a basal level in the proliferating chondrocytes and at a particularly high level in the prehypertrophic cells of the developing chicken cartilage, while it was absent in the hypertrophic cells (Fig. 2 f). To assess if this specific expression was indeed important for the pre-hypertrophic state of the cells, we knocked down Prdx1 with a viral miRNA knockdown construct, RCASBP(A)-ggamirPrdx1 (see Materials and Methods). The construct was electroporated into HH14 embryonic chick hind limb bud to knockdown Prdx1 in the developing limb skeleton. As before, the virus infected region was identified by immunohistochemistry using 3C2 antibody. Infected patches of pre20

hypertrophic cells (Fig. 6 a) displayed attenuated Prdx1 mRNA expression (Fig. 6 b, compare the regions marked with white arrowheads with the surrounding). Interestingly, within an infected patch of pre-hypertrophic cells (Fig. 6 c, box c´), Ihh mRNA expression was significantly reduced (Fig. 6 d, box d´ and Fig. S2 g) when compared to an uninfected pre-hypertrophic patch (Fig. 6 c, box c´´) within the same skeletal element (compare, Ihh mRNA expression between box d´ and box d´´ of Fig. 6 d). The cells in the infected patch continued to express type II collagen (Fig. S2 i). Further, type X collagen expression was also downregulated in RCASBP(A)-gga-mirPrdx1 infected patches of hypertrophic region (Fig. 6 f), as compared to the hypertrophic region of an uninfected contralateral limb skeletal element (Fig. 6 e). We next investigated the phenotypic consequences of mis-expression of Prdx1. For this purpose, HH14 hind limb bud was electroporated with RCASBP(A)-gga-FLAG-hPrdx1 (mis-expression construct). Again, infection was detected by 3C2 antibody (Fig. 6 i, white asterisk). Within the infected patch, type X collagen expression was markedly abrogated while the immediately adjacent uninfected portion continued to express type X collagen (compare the regions marked with yellow asterisk [infected] and yellow arrowhead [uninfected] in Fig. 6 h).

21

Figure 6. Prdx1 is critical for maintenance of pre-hypertrophy: a) 3C2 antibody staining (in green) demarcating the regions of pre-hypertrophic chondrocytes in HH34 femur infected with RCASBP(A)-gga-mirPrdx1 (knock-down construct) and Prdx1 mRNA expression is pseudocolored in red. b) Same section as (a) only showing mRNA expression of Prdx1 pseudocolored in red. White arrows indicating infected regions with attenuated Prdx1 mRNA expression. c) RCASBP(A)-ggamirPrdx1 (knock-down construct) infected HH34 tibia. Box-c´ marks infected pre-hypertrophic region while box-c´´ marks the other pre-hypertrophic region which is sparsely infected. d) Ihh mRNA expression in a section adjacent to the one used in panel (c). Box-d´ is the pre-hypertrophic region corresponding to box-c´ and box-d´´ is the same for box-c´´. e) Type X collagen expression in the hypertrophic region of an uninfected HH34 skeletal element. f) Type X collagen expression in the hypertrophic region of a HH34 skeletal element, contralateral to the limb used in panel (e), infected 22

with RCASBP(A)-gga-mirPrdx1 (knock-down construct). 3C2 (green) marks infection. g) Type X collagen expression in the hypertrophic region of an uninfected HH34 tibia. h) Type X collagen expression in the hypertrophic region of a HH34 skeletal element, contralateral to the limb used in panel (g), infected with RCASBP(A)-gga-FLAG-hPrdx1 (mis-expression construct). Yellow asterisk marks the region of putative hypertrophic cells infected by the virus while yellow arrowhead points towards the region which is uninfected. i) RCASBP(A)-gga-FLAG-hPrdx1 (mis-expression construct) infection in a serial section of (h). White asterisk denoting 3C2 marked infected domain. j) Ihh expression in the pre-hypertrophic cells of an uninfected HH34 tibia. k) Ihh expression in the RCASBP(A)-gga-FLAG-hPrdx1 (mis-expression construct) infected HH34 tibia, contralateral to the limb used in panel (j). l) 3C2 (green) staining depicting the extent of RCASBP(A)-gga-FLAG-hPrdx1 (mis-expression construct) infection in the same section as (k). m) Merged image of panels (k) and (l). Ihh mRNA expression is pseudocolored in red and 3C2 in green. n) Marked region from (m), white asterisk marks infected proliferating chondrocytes. o) Marked region from (m) where many of the putative hypertrophic chondrocytes are infected. p) Alizarin red and alcian blue stained HH34 chick skeleton. The right limb (boxed region) is infected with RCASBP(A)-gga-FLAG-hPrdx1 (misexpression construct). q) Boxed region from (p), asterisk marks infected limb with shorter and broader tibia devoid of ossification. Scale bars represents 100µm. The uninfected contralateral tibia of the same embryo had normal type X collagen expression pattern in the hypertrophic cells (Fig. 6 g). Upon Prdx1 mis-expression, Ihh mRNA expression was sustained even in the putative hypertrophic cells (Fig. 6 k,m and o) in comparison to an uninfected skeletal element where it is absent in the middle of the element i.e. the hypertrophic zone (Fig. 6 j). Interestingly, ectopic expression of Ihh mRNA was not observed in the proliferating chondrocytes, which were yet to enter pre-hypertrophy, even if infected with Prdx1 mis-expression construct (Fig. 6 n, white asterisk). Mis-expression of Prdx1 in limb skeletal element (Fig. 6 p) also resulted in shorter but broader tibial elements (Fig. 6 q). As expected, we observed, loss of alizarin red staining in the infected limb in alcian blue/alizarin red-stained HH34 chicken embryonic skeleton.

23

Figure 7. Different zones in developing cartilage anlagen and graphical model depicting role of Prdx1: (a-e) RNA in situ hybridization for markers of different domains of developing cartilage - a) PCNA (proliferation marker) marks the distal proliferative zone; b) PTHrP; c) Ihh (prehypertrophic chondrocyte marker) marks the prehypertrophic chondrocytes flanking either side of the hypertrophic zone; d) Prdx1 and e) immunohistochemistry for ColX (hypertrophic cell marker) on comparable sections from HH34 chick hindlimb. Black arrow heads point towards the expression domains of respective markers. (f-h) Graphical model summarising the major findings of this work. f) f. schematically summarizes the expression domains presented in panels a-e and a model of Prdx1 action. g) Model depicting knockdown of Prdx1 in developing cartilage. The resulting loss of Ihh signifies block in prehypertrophy which in turn does not allow the chondrocytes to go into hypertrophy. h) Model depicting overexpression of Prdx1. In this case, the expanded/ectopic Ihh expression domain does not allow expression of ColX, in turn blocking hypertrophic differentiation of the chondrocytes as Ihh negatively regulates hypertrophy. The numbers in panel “f” indicates the literature source of the relationships depicted. 1 = this study; 2 = (Vortkamp et al., 1996); 3 = (Vortkamp et al., 1996); 4 = (Ray et al., 2015).

24

Discussion ROS has long been implicated in different pathological developments (Juranek et al., 2013). In comparison, very little has been reported about the role of ROS in embryonic development. In this study, we demonstrate that BMP signaling, a major signaling pathway for skeletal development, transcriptionally regulates a ROS scavenger molecule Prdx1 and in turn the abundance of ROS in skeletal progenitor cells. Moreover, our data suggests that regulation of ROS abundance is critically important for the transition of pre-hypertrophic chondrocytes into hypertrophic chondrocytes in developing chick embryonic limb skeleton. In this study, we identified Prdx1 as a target of BMP signaling by comparing the proteome of an osteoblast cell line (BRITER) before and after the depletion of Bmp2 and Bmp4 ligands. We verified this finding in BRITER cell line by western blot analysis. Subsequently, we have demonstrated that abundance of Prdx1 went down in an osteogenic progenitor cell line, TVA-BMSC, upon depletion of Bmp2 and Bmp4 while it increased when primary mouse osteoblast cells or immortalized MC3T3E-1 osteoblast cells were exogenously treated with Bmp2. Further, through the use of a well characterized Prdx1 reporter construct, we demonstrated that the reporter activity is stimulated in either TVABMSC or MC3T3E-1 cells when treated with exogenous Bmp2. Finally, through the use of a Bmp2/4 double knockdown miRNA construct we demonstrated BMP signaling dependent expression of Prdx1 in vivo as well. Next, we demonstrated, in vitro, using dexamethasone treated TVA-BMSC and MC3T3E-1 cells and in vivo using ovariectomized rat femoral protein extracts that expression of Prdx1 correlated with osteogenesis. We further demonstrated that knockdown of Prdx1 in osteoblast cells blocked osteogenesis in vitro. Taken together, our data thus far clearly demonstrates that BMP signaling regulates Prdx1 expression during skeletal differentiation and Prdx1 expression correlates with osteogenesis. Before this study, role of BMP signaling in regulating ROS abundance during osteogenesis was never investigated. Biochemical assays demonstrated that depletion of Bmp2 and Bmp4 in BRITER cells resulted in dramatic upregulation of ROS abundance. In fact, ROS level in 4-OHT treated BRITER cells was 25

even higher than H2O2 treated BRITER cells (Fig. 5 a). Abundance of ROS and its relation to Bmp2 and Bmp4 depletion was also verified cytologically in TVA-BMSCs (Fig. 5 c and d). This increase in ROS abundance upon depletion of BMP ligands correlates well with decrease in Prdx1 mRNA abundance. It has been demonstrated that while increasing the ROS level stimulated hypertrophic differentiation decreasing it inhibited hypertrophic differentiation (Morita et al., 2007). This finding taken together with the data obtained in this study suggested that BMP signaling-dependent Prdx1 expression in developing cartilage regulates the ROS level which in turn is critically important for hypertrophic differentiation. A study by Morita et. al. observed maximum ROS level in the hypertrophic cells which is dramatically reduced in the pre-hypertrophic cells (Fig. 1 a of Morita et al., 2007). The in situ hybridisation pattern of Prdx1 mRNA that we observed in developing chick limb skeleton perfectly complements this observed distribution of ROS abundance. We observed very high Prdx1 mRNA expression in the pre-hypertrophic cells, while the same couldn’t not be detected in the hypertrophic cells. In order to investigate the role of Prdx1 in skeletal development in vivo, we conducted both lossand gain-of-function experiments. The miRNA-mediated knockdown of Prdx1 blocked hypertrophic differentiation. Paradoxically, mis-expression of Prdx1 also blocked hypertrophic differentiation and subsequent mineralization. In keeping, RCASBP(A)-gga-FLAG-hPrdx1 infected hypertrophic cells failed to express type-X collagen (Fig. 6 h, yellow asterisk), while cells within the same section which were uninfected, expressed type-X collagen (Fig. 6 h, yellow arrow head). Further, chick embryonic skeleton infected with RCASBP(A)-gga-FLAG-hPrdx1 (Fig. 6 q) closely resembles the skeletal phenotype observed upon Ihh mis-expression (Vortkamp et al., 1996). In both these cases, the skeleton appeared shorter, broader and didn’t undergo mineralization in contrast to uninfected contralateral limb. On the other hand, miRNA-mediated knockdown of Prdx1 downregulated expression of Ihh in pre-hypertrophic cells but did not result in ectopic hypertrophic differentiation, rather these cells continued to express type II collagen. Thus BMP signaling-mediated induction of Prdx1 and consequent lowering of ROS level appears to be critically important for the expression of Ihh and therefore transition from proliferating chondrocytes to pre-hypertrophic chondrocytes. It was 26

interesting to observe, however, that upon mis-expression of Prdx1 in embryonic chick skeletal element, Ihh expression persisted in the infected putative hypertrophic cells (Fig. 6 o) but no such Ihh expression was induced in the proliferating chondrocytes (Fig. 6 n). Thus, mis-expression of Prdx1 and consequent lowering of ROS level is not sufficient for induction of Ihh expression. However, it appears that Prdx1 expression and maintenance of low level of ROS is necessary for maintenance of Ihh expression. These observations, taken together with existing literature, suggest that increase in ROS, presumably achieved through downregulation of Prdx1 expression, might play a critical role in turning-off Ihh expression and thus, facilitating the transition of pre-hypertrophy to hypertrophy in cartilage element. In agreement, recently it has been demonstrated that hyperoxia inhibited expression of Shh and Ptc1 (Dang et al., 2017) while hypoxia promoted the expression of Shh (Al Ghouleh et al., 2017), suggesting that ROS regulation of hedgehog ligand expression may be a general feature. Further, ROS has been previously demonstrated to affect cell proliferation and thus we investigated whether overexpression of Prdx1 affects proliferation of cartilage cells (Han et al., 2003). However, we could not detect significant change in the proliferation profile of cartilage cells upon Prdx1 overexpression (data not shown). Prior to this study, the only embryonic context in which Prdx1 has been studied is the vasculature. Recently, Huang et. al., reported the role of Prdx1 in zebrafish vascular development. In this study, the authors suggested that the expression of Prdx1 depends on BMP signaling (Huang et al., 2017). Another study demonstrated that in the vasculature ROS abundance increased with the expression of a BMP inhibitor, Gremlin (Al Ghouleh et al., 2017). In this study, we have established a link between BMP signaling, Prdx1 expression and ROS abundance in the critical step of transition of prehypertrophic to hypertrophic chondrocytes during endochondral ossification. Our data taken together with published literature allow us to propose a model for the role of BMP signalling in regulating hypertrophic differentiation of cartilage cells (Fig. 7). According to our model, BMP signalling promotes expression of Prdx1 in the pre-hypertrophic chondrocytes. This zone of low ROS abundance is necessary for maintenance of pre-hypertrophic state of the cartilage cells. Downregulation of Prdx1 expression by a yet to be identified mechanism prompts downregulation of Ihh. 27

Ihh which normally inhibits onset of hypertrophy. Over-expression of Prdx1 expands the domain of Ihh expression thereby blocking onset of hypertrophy. On the other hand, knockdown of Prdx1 presumably causes loss of the low ROS region and thereby prehypertrophic differentiation. In absence of pre-hypertrophic differentiation, cartilage cells could not proceed to hypertrophy. It should be noted that we discovered and initially characterized expression of Prdx1 as well as its role in osteogenesis in cell culture systems, where overt signs of hypertrophy were not apparent. Nonetheless, it has been observed that blocking Hh signaling, through administration of a known inhibitor, cyclopamine, attenuates osteogenesis in BMSCs (Li et al., 2015). Our in vivo experiments conducted in chick embryos relate well to the data earlier obtained in developing mouse skeleton. Thus it is likely that the mechanism uncovered in this study is conserved across vertebrate species. Author Contributions Conception and Design: A.K.T. and A.B.; Acquisition of data: Y.K., T.B., G.T., J.K., S.K., A.S. and K.K.; Analysis and interpretation of data: A.K.T., S.S., N.C., A.B.; Writing and review of manuscript: A.K.T., A.B., Y.K. and T.B. Study supervision: A.K.T., N.C. and A.B.; Approving final version of manuscript: Y.K., T.B., A.K.T. and A.B.; A.K.T., A.B., Y.K. and T.B. take responsibility for the integrity of data. Acknowledgments Funding from the Council of Scientific and Industrial Research (BSC0201 and YSA0002) to A.K.T. and grant BT/PR14857/BRB/10/891/201 from Department of Biotechnology, Govt. of India to A.B. is acknowledged. Authors also acknowledge technical support provided by Mass spectroscopy Unit of Sophisticated and Analytical Instrument Facility of CSIR-CDRI. CDRI manuscript number for this article is 238/2017/AKT. Conflict of Interest The authors have no conflicts of interest with the contents of this article. References 28

Al Ghouleh, I., S. Sahoo, D.N. Meijles, J.H. Amaral, D.S. de Jesus, J. Sembrat, M. Rojas, D.A. Goncharov, E.A. Goncharova, and P.J. Pagano. 2017. Endothelial Nox1 Oxidase Assembly in Human Pulmonary Arterial Hypertension; Driver of Gremlin1-Mediated Proliferation. Clin Sci (Lond). Atashi, F., A. Modarressi, and M.S. Pepper. 2015. The role of reactive oxygen species in mesenchymal stem cell adipogenic and osteogenic differentiation: a review. Stem cells and development. 24:1150-1163. Bandyopadhyay, A., K. Tsuji, K. Cox, B.D. Harfe, V. Rosen, and C.J. Tabin. 2006. Genetic analysis of the roles of BMP2, BMP4, and BMP7 in limb patterning and skeletogenesis. PLoS Genet. 2:e216. Bandyopadhyay, A., P.S. Yadav, and P. Prashar. 2013. BMP signaling in development and diseases: a pharmacological perspective. Biochemical pharmacology. 85:857-864. Chen, C.M., and C.L. Cepko. 2002. The chicken RaxL gene plays a role in the initiation of photoreceptor differentiation. Development. 129:5363-5375. Chen, C.T., Y.R. Shih, T.K. Kuo, O.K. Lee, and Y.H. Wei. 2008. Coordinated changes of mitochondrial biogenesis and antioxidant enzymes during osteogenic differentiation of human mesenchymal stem cells. Stem Cells. 26:960-968. Dang, H.X., J. Li, C. Liu, Y. Fu, F. Zhou, L. Tang, L. Li, and F. Xu. 2017. CGRP attenuates hyperoxiainduced oxidative stress-related injury to alveolar epithelial type II cells via the activation of the Sonic hedgehog pathway. International journal of molecular medicine. 40:209-216. Hall, A., K. Nelson, L.B. Poole, and P.A. Karplus. 2011. Structure-based insights into the catalytic power and conformational dexterity of peroxiredoxins. Antioxidants & redox signaling. 15:795-815. Hamburger, V., and H.L. Hamilton. 1992. A series of normal stages in the development of the chick embryo. 1951. Dev Dyn. 195:231-272. Han, M.J., B.Y. Kim, S.O. Yoon, and A.S. Chung. 2003. Cell proliferation induced by reactive oxygen species is mediated via mitogen-activated protein kinase in Chinese hamster lung fibroblast (V79) cells. Mol Cells. 15:94-101. Huang, P.C., C.C. Chiu, H.W. Chang, Y.S. Wang, H.H. Syue, Y.C. Song, Z.H. Weng, M.H. Tai, and C.Y. Wu. 2017. Prdx1-encoded peroxiredoxin is important for vascular development in zebrafish. FEBS letters. 591:889-902. Hughes, S.H., J.J. Greenhouse, C.J. Petropoulos, and P. Sutrave. 1987. Adaptor plasmids simplify the insertion of foreign DNA into helper-independent retroviral vectors. J Virol. 61:3004-3012. Immenschuh, S., and E. Baumgart-Vogt. 2005. Peroxiredoxins, oxidative stress, and cell proliferation. Antioxidants & redox signaling. 7:768-777. Jung, H., T. Kim, H.Z. Chae, K.T. Kim, and H. Ha. 2001. Regulation of macrophage migration inhibitory factor and thiol-specific antioxidant protein PAG by direct interaction. J Biol Chem. 276:15504-15510. Juranek, I., D. Nikitovic, D. Kouretas, A.W. Hayes, and A.M. Tsatsakis. 2013. Biological importance of reactive oxygen species in relation to difficulties of treating pathologies involving oxidative stress by exogenous antioxidants. Food and chemical toxicology : an international journal published for the British Industrial Biological Research Association. 61:240-247. Karsenty, G., H.M. Kronenberg, and C. Settembre. 2009. Genetic control of bone formation. Annu Rev Cell Dev Biol. 25:629-648. Karsenty, G., and E.F. Wagner. 2002. Reaching a genetic and molecular understanding of skeletal development. Dev Cell. 2:389-406. Khan, M.P., A.K. Singh, A.A. Joharapurkar, M. Yadav, S. Shree, H. Kumar, A. Gurjar, J.S. Mishra, M.C. Tiwari, G.K. Nagar, S. Kumar, R. Ramachandran, A. Sharan, M.R. Jain, A.K. Trivedi, R. Maurya, M.M. Godbole, J.R. Gayen, S. Sanyal, and N. Chattopadhyay. 2015. Pathophysiological Mechanism of Bone Loss in Type 2 Diabetes Involves Inverse Regulation of Osteoblast Function by PGC-1alpha and Skeletal Muscle Atrogenes: AdipoR1 as a Potential Target for Reversing Diabetes-Induced Osteopenia. Diabetes. 64:2609-2623. 29

Kronenberg, H.M. 2003. Developmental regulation of the growth plate. Nature. 423:332-336. Kronenberg, H.M. 2006. PTHrP and skeletal development. Ann N Y Acad Sci. 1068:1-13. Kumar, Y., I. Kapoor, K. Khan, G. Thacker, M.P. Khan, N. Shukla, J.K. Kanaujiya, S. Sanyal, N. Chattopadhyay, and A.K. Trivedi. 2015. E3 Ubiquitin Ligase Fbw7 Negatively Regulates Osteoblast Differentiation by Targeting Runx2 for Degradation. J Biol Chem. 290:3097530987. Lanske, B., A.C. Karaplis, K. Lee, A. Luz, A. Vortkamp, A. Pirro, M. Karperien, L.H. Defize, C. Ho, R.C. Mulligan, A.B. Abou-Samra, H. Juppner, G.V. Segre, and H.M. Kronenberg. 1996. PTH/PTHrP receptor in early development and Indian hedgehog-regulated bone growth. Science. 273:663-666. Li, L., Q. Dong, Y. Wang, Q. Feng, P. Zhou, X. Ou, Q. Meng, T. He, and J. Luo. 2015. Hedgehog signaling is involved in the BMP9-induced osteogenic differentiation of mesenchymal stem cells. International journal of molecular medicine. 35:1641-1650. Mody, N., F. Parhami, T.A. Sarafian, and L.L. Demer. 2001. Oxidative stress modulates osteoblastic differentiation of vascular and bone cells. Free radical biology & medicine. 31:509-519. Morita, K., T. Miyamoto, N. Fujita, Y. Kubota, K. Ito, K. Takubo, K. Miyamoto, K. Ninomiya, T. Suzuki, R. Iwasaki, M. Yagi, H. Takaishi, Y. Toyama, and T. Suda. 2007. Reactive oxygen species induce chondrocyte hypertrophy in endochondral ossification. Journal of Experimental Medicine. 204:1613-1623. Neumann, C.A., J. Cao, and Y. Manevich. 2009. Peroxiredoxin 1 and its role in cell signaling. Cell Cycle. 8:4072-4078. Prashar, P., P.S. Yadav, F. Samarjeet, and A. Bandyopadhyay. 2014. Microarray meta-analysis identifies evolutionarily conserved BMP signaling targets in developing long bones. Dev Biol. 389:192-207. Ray, A., P.N. Singh, M.L. Sohaskey, R.M. Harland, and A. Bandyopadhyay. 2015. Precise spatial restriction of BMP signaling is essential for articular cartilage differentiation. Development. 142:1169-1179. Schmid, T.M., and T.F. Linsenmayer. 1985. Immunohistochemical localization of short chain cartilage collagen (type X) in avian tissues. J Cell Biol. 100:598-605. Shiota, M., H. Izumi, N. Miyamoto, T. Onitsuka, E. Kashiwagi, A. Kidani, G. Hirano, M. Takahashi, M. Ono, M. Kuwano, S. Naito, Y. Sasaguri, and K. Kohno. 2008. Ets regulates peroxiredoxin1 and 5 expressions through their interaction with the high-mobility group protein B1. Cancer science. 99:1950-1959. Singh, P.N., A. Ray, K. Azad, and A. Bandyopadhyay. 2016. A comprehensive mRNA expression analysis of developing chicken articular cartilage. Gene Expr Patterns. 20:22-31. Smith, C.A., K.N. Roeszler, T. Ohnesorg, D.M. Cummins, P.G. Farlie, T.J. Doran, and A.H. Sinclair. 2009. The avian Z-linked gene DMRT1 is required for male sex determination in the chicken. Nature. 461:267-271. St-Jacques, B., M. Hammerschmidt, and A.P. McMahon. 1999. Indian hedgehog signaling regulates proliferation and differentiation of chondrocytes and is essential for bone formation. Genes Dev. 13:2072-2086. Suzuki, T., and T. Ogura. 2008. Congenic method in the chick limb buds by electroporation. Dev Growth Differ. 50:459-465. Thacker, G., Y. Kumar, M.P. Khan, N. Shukla, I. Kapoor, J.K. Kanaujiya, S. Lochab, S. Ahmed, S. Sanyal, N. Chattopadhyay, and A.K. Trivedi. 2016. Skp2 inhibits osteogenesis by promoting ubiquitinproteasome degradation of Runx2. Biochim Biophys Acta. 1863:510-519. Urist, M.R. 1965. Bone: formation by autoinduction. Science. 150:893-899. Vortkamp, A., K. Lee, B. Lanske, G.V. Segre, H.M. Kronenberg, and C.J. Tabin. 1996. Regulation of rate of cartilage differentiation by Indian hedgehog and PTH-related protein. Science. 273:613622. 30

Weir, E.C., W.M. Philbrick, M. Amling, L.A. Neff, R. Baron, and A.E. Broadus. 1996. Targeted overexpression of parathyroid hormone-related peptide in chondrocytes causes chondrodysplasia and delayed endochondral bone formation. Proc Natl Acad Sci U S A. 93:10240-10245. Yadav, P.S., M.P. Khan, P. Prashar, S. Duggal, S.K. Rath, N. Chattopadhyay, and A. Bandyopadhyay. 2016. Characterization of BMP signaling dependent osteogenesis using a BMP depletable avianized bone marrow stromal cell line (TVA-BMSC). Bone. 91:39-52. Yadav, P.S., P. Prashar, and A. Bandyopadhyay. 2012. BRITER: a BMP responsive osteoblast reporter cell line. PLoS One. 7:e37134.

31

Table 1 List of differentially expressed proteins (>30% up or downregulated) Identified from Tris and Urea lysate fractions.

Identified differentially expressed proteins (>30% up or downregulated) from Tris lysate fraction Sam Sp Identified Ac. Sco Mw(Da Seque Expres Expres Status in KD/ Fold ple ot Protein No. re ) nce sion sion expression I. /pI cover value value D. age Uni KD (%) 1 1 Peroxired P357 126 22390/8 11 1.0040 Absent in KD oxin-1/ 00 .26 3 Osteoblas t-specific factor 3 2 3 Galectin-3 P161 204 27612 0.1212 0.1722 Up regulated 9 10 /8.46 18 67 /1.421134 3 11 Macropha P348 62 12667 9 0.2724 0.3544 Up regulated ge 84 /6.79 13 07 /1.300992 migration inhibitory factor 4 12 Myosin Q606 169 17090/4 29 0.8945 0.3963 Down regulated light 05 .56 61 9 /0.443111 polypeptid e6 5 18 Proliferati P179 212 29108 27 0.5504 0.8109 Up regulated ng cell 18 /4.66 33 62 /1.473316 nuclear antigen 6 20 F-actinP477 109 33090 13 0.1172 0.1741 Up regulated capping 53 /5.34 /1.485429 61 83 protein subunit alpha-1 7 22 60 kDa P630 862 61088 19 0.5448 0.7186 Up regulated heat shock 38 /5.91 65 98 /1.319039 protein, mitochond rial 8 23 TP803 171 60042 7 0.1233 0.1798 Up regulated complex 16 /5.72 98 13 /1.457179 protein 1 subunit epsilon 9 29 Putative O890 130 16595 31 0.2186 0.2928 Up regulated 32 RNA86 /6.84 25 75 /1.339623 binding protein 3 10 33 SH3 Q9JJ 65 12917/4 8 1.0134 0.6963 Down regulated domainU8 .87 5 88 /0.687146

11

36

12

40

13

42

14

43

15

44

binding glutamic acid-richlike protein Leukocyte elastase inhibitor A Calumeni n Tubulin alpha-1A chain Tubulin beta-5 chain Elongatio n factor 1alpha 1

Q9D 154

162

29108/5 .85

13

0.1725 1

O358 87 P683 69

250

37155/4 .49 50788/4 .94

9

1.0389 1 0.4173 1

0.2309 25

Down regulated /0.553366

P990 24

90

50095/4 .78

16

0.4173 1

0.2309 25

Down regulated /0.553366

P101 26

41

50424/9 .10

3

0.1176 85

0.1532 23

Up regulated /1.301975

186

7

0.2386 94

Up regulated /1.383653

Absent in KD

Identified differentially expressed proteins (>30% up or downregulated) from Urea lysate fraction Sam Sp Identified Ac. Sco Mw(Da Seque Expres Expres Status in KD/ Fold ple ot Protein No. re ) nce sion sion expression I. /pI cover value value D. age Uni KD (%) 1 4 Vimentin P201 570 53712/5 22 0.0216 0.0361 Up .06 52 63 33 regulated/1.667959 19309422 2

11

Histone H2B type 1-B

Q644 75

110

13944/1 0.31

23

0.0490 24

0.0197 416

Down regulated/0.402692 558746736

3

15

Galectin-1

P160 45

289

15198/5 .32

23

0.0647 058

0.0456 338

Down regulated/0.705250 533955225

4

19

14-3-3 protein epsilon

P622 59

62

29326/4 .63

4

0.0647 058

0.0356 338

Down regulated/0.550704 882715305

5

25

Protein disulfideisomerase A3

P277 73

565

57099/5 .88

20

0.0290 283

0.0132 207

40S ribosomal protein SA

P142 06

71

32931/4 .80

6

0.0957 593

0.1351 88

6

28

Down regulated/0.455441 75855975 Up regulated/1.411747 99732245 33

7

30

Annexin A5

P480 36

119

35787/4 .83

3

0.1235 43

0.1854 05

Up regulated/1.500732 53846839

8

32

Protein disulfideisomerase

P091 03

777

57507/4 .79

21

0.0246 346

0.0338 479

Up regulated/1.373998 3600302

9

33

Vimentin

P201 52

663

53712/5 .06

32

0.0368 118

0.0624 502

Up regulated/1.696472 32680825

10

34

Vimentin

P201 52

708

53712/5 .06

35

0.1650 63

0.2582 91

Up regulated/1.564802 52994311

11

36

Beta-2microglob ulin

P018 87

64

13928/7 .79

7

0.1917 63

0.1280 394

Down regulated/0.667696 062326935

12

38

Ubiquitin

P629 91

175

8560/6. 56

40

0.0255 152

0.0469 618

Up regulated/1.840542 10823352

13

55

Myristoyl ated alaninerich Ckinase substrate

P266 45

49

29701/4 .34

10

0.0361 906

0.0239 045

Down regulated/0.660516 819284566

14

56

Protein disulfideisomerase A3

P277 73

585

57099/5 .88

19

0.2535 67

0.3757 39

Up regulated/1.481813 48519326

Highlights   

Prdx1 is expressed in a BMP signaling dependent manner during osteogenesis. Prdx1 is transcribed in the pre-hypertrophic cells in a BMP signaling dependent manner Expression of Prdx1 is essential for maintenance of pre-hypertrophic state

34

 

Lowering ROS is necessary but not sufficient for Ihh expression and onset of prehypertrophy Low Prdx1 is key for the transition of pre-hypertrophic to hypertrophic chondrocytes

35