Original Article

RECQL4 Regulates p53 Function in vivo During Skeletogenesis**

Linchao Lu1, Karine Harutyunyan1†, Weidong Jin1, Jianhong Wu1††, Tao Yang2†††, Yuqing Chen2, 3, Kyu Sang Joeng2, Yangjin Bae2, Jianning Tao2, Brian C. Dawson2, 3, Ming-Ming Jiang2, 3, Brendan Lee2, 3, Lisa L. Wang1*

1

Texas Children’s Cancer Center, Department of Pediatrics, 2Department of Molecular and Human Genetics,

Baylor College of Medicine, 3Howard Hughes Medical Institute, Houston, TX 77030 †

Current address: Department of Leukemia, University of Texas M. D. Anderson Cancer Center, Houston, TX

77030 ††

Current address: Syngenta, Research Triangle Park, NC 27709

†††

Current address: Center for Skeletal Diseases and Tumor Metastasis, Van Andel Research Institute, Grand

Rapids, MI 49503

*

Correspondence should be addressed to L.L.W. ([email protected]) Tel: 832-824-4822

**

This article has been accepted for publication and undergone full peer review but has not been through the copyediting, typesetting, pagination and proofreading process, which may lead to differences between this version and the Version of Record. Please cite this article as doi: [10.1002/jbmr.2436] Additional Supporting Information may be found in the online version of this article. Initial Date Submitted August 15, 2014; Date Revision Submitted November 30, 2014; Date Final Disposition Set December 12, 2014

Journal of Bone and Mineral Research This article is protected by copyright. All rights reserved DOI 10.1002/jbmr.2436

This article is protected by copyright. All rights reserved

1

Abstract RECQ DNA helicases play critical roles in maintaining genomic stability, but their role in development has been less well studied. Rothmund-Thomson syndrome, RAPADILINO, and Baller-Gerold syndrome are rare genetic disorders caused by mutations in the RECQL4 gene. These patients have significant skeletal developmental abnormalities including radial ray, limb and craniofacial defects. To investigate the role of Recql4 in the developing skeletal system, we generated Recql4 conditional knockout mice targeting the skeletal lineage. Inactivation of Recql4 using the Prx1-Cre transgene led to limb abnormalities and craniosynostosis mimicking the major bone findings in human RECQL4 patients. These Prx1-Cre+;Recql4fl/fl mice as well as Col2a1Cre+;Recql4fl/fl mice exhibited growth plate defects and an increased p53 response in affected tissues. Inactivation of Trp53 in these Recql4 mutants resulted in genetic rescue of the skeletal phenotypes, indicating an in vivo interaction between Recql4 and Trp53, and p53 activation as an underlying mechanism for the developmental bone abnormalities in RECQL4 disorders. Our findings show that RECQL4 is critical for skeletal development by modulating p53 activity in vivo. This article is protected by copyright. All rights reserved

Keywords: Rothmund-Thomson syndrome, RECQL4, skeletal development, cartilage, genetic animal models

This article is protected by copyright. All rights reserved

2

Introduction The RECQ genes encode DNA helicases which are important for preserving genomic stability.(1) In humans, there are five RECQ DNA helicases--RECQL, WRN, BLM, RECQL4, and RECQL5. Human mutations in the WRN, BLM, and RECQL4 genes cause three distinct clinical syndromes, Werner syndrome,(2) Bloom syndrome,(3) and Rothmund-Thomson Syndrome (RTS),(4) respectively. While the link between genomic instability and cancer has been well studied, the role of genes important for maintaining genomic integrity has been less well studied in the context of developmental defects. RTS (OMIM #268400) is a rare autosomal recessive condition associated with a wide variety of clinical manifestations. These include a characteristic skin rash (poikiloderma), small stature, sparse scalp hair, eyebrows and eyelashes, hyperkeratosis, dental and nail abnormalities, radial ray and other bone defects, gastrointestinal disturbances, cataracts, and predisposition to cancer, particularly osteosarcoma, a form of bone cancer, as well as skin cancer (primarily squamous cell and basal cell carcinomas).(5,6) Because all patients display poikiloderma of the skin, RTS has generally been classified as a genodermatosis. However, approximately 75% of patients have significant skeletal abnormalities, including radial ray defects, aplastic or hypoplastic bones, synostoses, brachymesophalangy, and metaphyseal trabecular defects.(7) At least 25% of RTS patients have osteopenia on skeletal surveys with accompanying histories of fractures as young children.(7) Approximately two-thirds of patients diagnosed with RTS have mutations in the RECQL4 gene (designated Type II RTS), while the gene defect for the other one-third of RTS patients (Type I RTS) is not currently known.(8) RECQL4 has also been shown to be mutated in two other recessive conditions, RAPADILINO syndrome(9,10) (OMIM #266280) and Baller-Gerold syndrome(11) (BGS, OMIM #218600), both of which have prominent skeletal defects, which also include patellar defects and craniosynostosis. Therefore the RECQL4 disorders can also be considered within the spectrum of human skeletal dysplasias. RECQL4 belongs to a family of RECQ DNA helicases that as a whole are important for maintenance of genomic stability and are linked to cancer and aging.(1,12,13) RECQL4 has been shown to play a role primarily in the initiation of DNA replication,(14-16) as well as in DNA repair(17-20) and telomere(21) and mitochondrial This article is protected by copyright. All rights reserved

3

maintenance.(22,23) Presence of RECQL4 mutations correlates significantly with presence of skeletal abnormalities(7) as well as risk for developing osteosarcoma.(8) However, despite the high prevalence of bone developmental abnormalities and the exceedingly high rate of osteosarcoma in Type II RTS patients, the specific role of RECQL4 in the development and maintenance of the skeletal system has not been delineated. Therefore we sought to investigate the effect of deleting Recql4 during development of the skeletal system by generating a tissue specific conditional Recql4 knockout (KO) mouse model. Previously there were three attempts to generate global Recql4 KO mouse models. The first model by Ichikawa et al. deleted Recql4 exons 5-8 and displayed early embryonic lethality.(24) The other two models, which differed in the deleted regions, gave rise to variable and milder phenotypes that mimicked some aspects of the human disease.(25,26) In order to better understand the in vivo mechanisms contributing to the developmental skeletal abnormalities seen in patients with RECQL4-associated human diseases and to circumvent early embryonic lethality, we generated a mouse Recql4 conditional allele (hereafter called Recql4fl/fl) based on the first global mouse model by flanking exons 5-8 with loxP sites (Fig. S1). We crossed Recql4fl/fl mice with transgenic mice expressing Cre recombinase under the control of the regulatory elements of the paired-related homeobox gene 1 (Prx1) (hereafter called Prx1-Cre)(27) to delete Recql4 in early skeletal progenitor cells and also crossed Recql4fl/fl mice with transgenic mice expressing Cre recombinase driven by the regulatory elements of the proα1(II) collagen gene (Col2a1) (hereafter called Col2a1Cre)(28) to delete Recql4 in chondrocytes and in a subset of osteo-chondroprogenitor cells.(29,30)

Materials and Methods Mice All animal procedures were approved by the Baylor College of Medicine Institutional Animal Care and Use Committee. Prx1-Cre,(27) Col2a1-Cre(28) and Trp53fl/fl (31) mice have been described previously. PCR genotyping was performed using published primers. For generation of the Recql4 conditional allele, please see Supporting Information. To generate Recql4 conditional mutants, Prx1- or Col2a1-Cre male transgenic mice were crossed with Recql4fl/fl females to generate Prx1- or Col2a1-Cre+;Recql4fl/+ males, which were subsequently mated with Recql4fl/fl or Recql4fl/+ females. In our studies, Prx1- or Col2a1-Cre+;Recql4fl/+, Recql4fl/+, Recql4fl/fl, and Prx1- or This article is protected by copyright. All rights reserved

4

Col2a1-Cre+ mice were used as littermate controls (Ctrl) as they were phenotypically similar to wild type (Recql4+/+) mice. Recql4fl/+ and Recql4fl/fl mice were used as wild type littermate controls (WT Ctrl). To generate Recql4 and Trp53 double conditional mutants, Prx1- or Col2a1-Cre+;Recql4fl/+ males were crossed with Trp53fl/+ females to generate Prx1- or Col2a1-Cre+;Recql4fl/+;Trp53fl/+ males, which were then crossed with either Recql4fl/+;Trp53fl/+ or Recql4fl/fl;Trp53fl/fl females. Vaginal plugs in female mice were used to determine embryo age. Skeletal analysis Alcian blue and Alizarin red staining of E18.5, P0 or P21 mouse skeletons was performed using Alcian blue 8GX (Sigma) for cartilage and Alizarin red S (Sigma) for bone as described previously.(32) Micro-computed tomography (μCT) scanning of P21 mouse skulls was performed using a μCT-40 system (Scanco Medical). Histology Tissues were fixed in 4% paraformaldehyde. Hematoxylin and eosin (H&E) and von Kossa staining were performed on paraffin embedded sections using standard protocols. For embryos and mice younger than seven days, undecalcified tissues were used. Decalcification was performed for mice older than seven days using 14% EDTA (Sigma) solution. For cell density quantification in E18.5 distal femur growth plates, the number of cells in the area of 104 μm2 was counted in two different H&E stained sections for each sample. BrdU incorporation assay Pregnant female mice were intraperitoneally injected with BrdU solution (Life Technologies, 10 μl/g body weight) and sacrificed 2 h later to collect E17.5 or E18.5 embryos. BrdU staining was performed on paraffin embedded sections using BrdU In-Situ Detection Kit (BD Pharmingen). For BrdU incorporation rate, 100 proliferating chondrocytes in distal femur growth plates from each section were counted to determine the percentage of BrdU positive cells. Two different sections were counted for each sample. Terminal deoxynucleotidyl transferase dUTP nick end labeling (TUNEL) assay TUNEL assay was performed on paraffin embedded sections of E17.5 or E18.5 distal femur growth plates using In Situ Cell Death Detection Kit (Roche Diagnostics). DAPI was used for nuclear counterstaining. For quantification, the number of positive cells was counted from a section of the entire distal femur growth plate for each sample. This article is protected by copyright. All rights reserved

5

Immunoblot analysis Mouse E13.5 forelimbs and E18.5 costal cartilage were isolated and extracted with RIPA buffer using a T25 homogenizer (IKA). Lysates were loaded in SDS-PAGE gels under denaturing conditions and transferred to nitrocellulose membranes. Enhanced chemiluminescence (Pierce, SuperSignal Pico or Femto kit) or infrared imaging system (LI-COR) was used to visualize proteins on membranes. Primary antibodies were purchased from Cell Signaling Technology (anti-p53, #2524; anti-Phospho-p53 (Ser15), #9284), BD Pharmingen (anti-CDKN1A, #556431), and Sigma (anti-β-actin, #A2228). Secondary antibodies were purchased from Life Technologies and LI-COR. Real time RT-PCR For extraction of E13.5 mouse forelimb total RNA, GenElute Mammalian Total RNA Kit (Sigma) was used with on-column DNase digestion after tissue homogenization. For extraction of E18.5 costal cartilage total RNA, Trizol (Life Technologies) was first used followed by GenElute Mammalian Total RNA Kit with on-column DNase digestion. First-strand cDNA was generated from 1 μg RNA using SuperScript III First-Strand Synthesis System (Life Technologies). Real time PCR was performed using StepOnePlus system (Life Technologies) and Fast SYBR Green Master Mix (Life Technologies). The relative gene expression levels were calculated using 2−ΔΔCT method.(33) Please see Supporting Information for primer sequences. Statistical analyses Unpaired, two-tailed Student’s t-test was used to calculate statistical significance (p value) between two groups. One-way analysis of variance (ANOVA) was used to calculate statistical significance (p value) for multiple groups. P values less than 0.05 were considered statistically significant. Data are expressed as mean ± s.d. throughout this paper.

Results Inactivation of Recql4 using the Prx1-Cre transgene recapitulates the skeletal defects seen in patients with RECQL4 disorders.

This article is protected by copyright. All rights reserved

6

The Prx1-Cre transgene is highly expressed in the mesenchyme of developing limbs and in some non-neural cranial mesenchyme.(27) We first confirmed the deletion efficiency of Recql4 by the Prx1-Cre transgene in E13.5 forelimb tissues of Prx1-Cre+;Recql4fl/fl mice by real-time RT-PCR (Fig. S2A). Prx1-Cre+;Recql4fl/fl mice were born with a normal Mendelian ratio (data not shown). However, they displayed severe limb developmental defects at birth (Fig. 1A), and about one-third died before the weaning age most likely due to difficulty feeding as a result of their limb defects (Fig. S3A). These mutant mice started to show limb defects at E12.5 compared to littermate controls (Fig. 1B, S4). Whole skeleton preparations of P0 mice showed that the forelimbs of Prx1-Cre+;Recql4fl/fl mice were severely foreshortened and deformed (Fig. 1C). Hindlimbs of these mutants were less affected but were also smaller and shorter compared to littermate controls (Fig. 1D). In these mutants, missing digits or decreased mineralization in the remaining phalanges were commonly observed (arrows in Fig. 1C, D). At three weeks of age, mutant mice that survived exhibited significantly smaller limbs, as well as abnormal ossification in the joint region of the hindlimbs (Fig. 1E). Prx1-Cre+;Recql4fl/fl mutants also displayed growth retardation (Fig. S3B). Because the Prx1-Cre transgene is also expressed in cranial tissues of mesenchymal origin,(27,34) we examined the skulls of Prx1-Cre+;Recql4fl/fl mice. Whole skeleton preparations of P0 skulls showed bilateral coronal suture synostoses in these mutant mice at birth (Fig. 2A). The coronal sutures in these mutants appeared to already be fused and mineralized at E16.5 (Fig. 2B). Micro-CT analysis of three week old skulls confirmed the synostoses of the coronal sutures as well as the squamosal sutures in these mutants (Fig. 2C-E). Craniosynostosis is one of characteristic findings in BGS patients carrying RECQL4 mutations. The skeletal defects in Prx1Cre+;Recql4fl/fl mice, particularly missing digits, foreshortened long bones, and synostoses, recapitulate many of the major skeletal abnormalities seen in the RECQL4-associated disorders.

Recql4 is required for cartilage development. Because of the abnormal joint formation seen in our Prx1-Cre+;Recql4fl/fl mice (Fig. 1E) as well as some human patients with RECQL4 mutations, we examined the growth plates of our mutant mice. We examined the formation of primary ossification centers (POC) of long bone in Prx1-Cre+;Recql4fl/fl embryos at E13.5. Alcian blue and nuclear fast red staining of tibia POC demonstrated that there were no apparent abnormalities in mutant embryos This article is protected by copyright. All rights reserved

7

compared to littermate controls (Fig. S5). However, at E18.5, H&E staining of distal femur growth plates revealed that chondrocytes from Prx1-Cre+;Recql4fl/fl mice had significantly reduced cell density and increased cell size in the resting zone (RZ), proliferating zone (PZ) and hypertrophic zone (HZ) compared to littermate controls (Fig. 3A, B). At three weeks of age, mutant mice showed disorganized distal femur growth plates and abnormal chondrocyte morphology demonstrating increased cell size (Fig. 3C). It appears that inactivation of Recql4 in mesenchymal progenitor cells primarily affects the growth plate chondrocytes. These cartilage phenotypes correlate with the known expression of Recql4 in the developing cartilage(9) and may contribute to the skeletal defects in the human RECQL4 disorders at least in the appendicular skeleton. To more specifically investigate the physiological role of Recql4 in cartilage, we crossed Recql4fl/fl mice with Col2a1-Cre transgenic mice(28) to inactivate Recql4 in chondrocytes. We first confirmed the deletion efficiency of Recql4 by this transgene (Fig. S2B) and also showed that the expression levels of Col2a1 gene were unchanged (Fig. S6) in E18.5 costal cartilage of Col2a1-Cre+;Recql4fl/fl mutants. These mutant mice were born with a normal Mendelian ratio (data not shown) and appeared to have normal limb development at birth. However, all mutant pups exhibited severe rib cage and vertebral defects and died in the perinatal period (Fig. 3D). H&E staining of E18.5 distal femur growth plates of Col2a1-Cre+;Recql4fl/fl mutants showed similar findings as Prx1Cre+;Recql4fl/fl mutants; i.e., significantly decreased chondrocyte density and increased cell size in the RZ, PZ and HZ of growth plates (Fig. 3E, F). Furthermore, when we isolated primary chondrocytes from E18.5 costal cartilage, we observed that chondrocytes from Col2a1-Cre+;Recql4fl/fl mutants also displayed increased cell size compared to littermate controls (Fig. S7A). Overall, the skeletal defects displayed in these mouse models indicate that Recql4 has a crucial role in cartilage development and homeostasis.

Deletion of Recql4 leads to increased apoptosis and decreased cell proliferation. To determine the cellular basis of reduced chondrocyte density in Recql4 mutant growth plates, we performed apoptosis and proliferation assays in distal femur growth plates of both Prx1-Cre+;Recql4fl/fl and Col2a1Cre+;Recql4fl/fl mutant embryos. TUNEL assay showed significantly increased numbers of apoptotic cells in E18.5 growth plates of Prx1-Cre+;Recql4fl/fl mutants compared to littermate controls (Fig. 4A). BrdU incorporation assay This article is protected by copyright. All rights reserved

8

was used to determine chondrocyte proliferation in growth plates. We observed a significantly reduced number of BrdU positive cells in the PZ of Prx1-Cre+;Recql4fl/fl growth plates compared to littermate controls (Fig. 4B). We performed similar assays on Col2a1-Cre+;Recql4fl/fl mutant growth plates and obtained similar findings (Fig. 4C, D). Thus the abnormal cell density phenotype observed in our Prx1-Cre+;Recql4fl/fl and Col2a1-Cre+;Recql4fl/fl mice can be explained by a combination of proliferation and apoptotic defects.

Inactivation of Recql4 induces p53 activation. Because RECQL4 participates in multiple aspects of DNA metabolism,(1,13) and our findings indicate that inactivation of Recql4 in the skeletal system causes increased apoptosis, we performed molecular analyses to understand the basis of this cellular defect. We hypothesized that inactivation of Recql4 could result in increased/unresolved DNA damage in mutant cells. Therefore, we examined Recql4 mutant growth plates to determine if there was increased DNA damage. Immunohistochemical staining of DNA damage marker γH2AX(35) showed that mutant growth plate chondrocytes had strong γH2AX staining in both Prx1-Cre+;Recql4fl/fl and Col2a1-Cre+;Recql4fl/fl embryos at E18.5 (Fig. S8). It has been shown that upon DNA damage, p53 can be phosphorylated at Ser15 by ATR and ATM,(36-38) which abolishes the interaction of p53 with MDM2 and prevents p53 from being targeted to the proteasomal degradation pathway.(39) To determine if p53 signaling could be activated in Recql4 mutant mice, we first examined total p53 (also known as Trp53) protein levels in E13.5 forelimb tissues of Prx1-Cre+;Recql4fl/fl mutants. Immunoblotting showed that these mutants appeared to have similar total p53 protein levels compared to littermate controls (Fig. 5A). However, the amount of phosphorylated p53 at Ser15 was significantly increased in mutant mice (Fig. 5A). These results suggest that the increased/unresolved DNA damage in Recql4 mutant cells could induce p53 activation through phosphorylation of p53 at Ser15. In addition, previous in vitro studies by De et al demonstrated that RECQL4 physically interacts with and directs p53 to the mitochondria in unstressed conditions in normal cells, while in RECQL4 deficient cells, p53 accumulates in the nucleus to activate downstream targets of p53.(23) We analyzed p53 cellular localization by immunofluorescence microscopy in primary chondrocytes from Col2a1-Cre+;Recql4fl/fl mutant embryos at E18.5,

This article is protected by copyright. All rights reserved

9

and found that there was no apparent change in p53 cellular localization in mutant cells compared to WT cells (Fig. S9). To further investigate whether this increased p53 phosphorylation enhances p53 activity, we examined the expression of downstream target genes of p53 in tissues from both Recql4 mutant models. Indeed, the expression of CDKN1A (also known as p21) and Bax were up-regulated 2-3 fold in E13.5 forelimb tissues of Prx1Cre+;Recql4fl/fl mutants compared to littermate controls as shown by real-time RT-PCR (Fig. 5B), while the Trp53 transcripts were unchanged (Fig. S10A). Similar results were observed in Col2a1-Cre+;Recql4fl/fl mutants; i.e., the levels of transcripts of CDKN1A, Bax and puma were significantly increased in E18.5 costal cartilage compared to littermate controls (Fig. 5C), while the expression of Trp53 was unchanged (Fig. S10B). The protein level of CDKN1A was also increased in E18.5 costal cartilage from Col2a1-Cre+;Recql4fl/fl mutants (Fig. 5D). Because p53 signaling can induce cellular senescence, and mice with a Recql4 targeted mutation exhibit increased senescence,(40) we performed senescence-associated beta-galactosidase (SA-β-gal) staining in E18.5 primary chondrocytes from costal cartilage at passage 0. We observed increased levels of senescence in Col2a1Cre+;Recql4fl/fl mutant cells (more than 20%) compared to littermate control cells (less than 5%) (Fig. S7A, B). Our findings suggest that deletion of Recql4 in the skeletal progenitor cells and chondrocytes causes strong activation of p53, which is primarily mediated through the DNA damage response.

Recql4 interacts genetically with Trp53 in vivo during skeletogenesis. To understand to what extent activated p53 is responsible for the skeletal phenotypes in these mouse models with Recql4 loss of function, we performed genetic rescue experiments. First, we used Prx1-Cre to inactivate both Recql4 and Trp53 in mesenchymal progenitor cells. The total p53 protein levels were significantly reduced in E13.5 forelimb tissues of Prx1-Cre+;Recql4fl/fl;Trp53fl/fl double mutants (Fig. 6A, B). Residual p53 may be due to the fact that Prx1-Cre transgene is not expressed in the limb bud ectoderm,(27) or it may also be due to incomplete deletion of the floxed Trp53 gene. Nevertheless, the phosphorylation of p53 at Ser15 was almost completely abolished in E13.5 forelimb tissues of these double mutants compared to Prx1-Cre+;Recql4fl/fl and Prx1Cre+;Recql4fl/fl;Trp53fl/+ mutants (Fig. 6A, C). The striking forelimb developmental defects seen in Prx1This article is protected by copyright. All rights reserved

10

Cre+;Recql4fl/fl mice were partially rescued in the Prx1-Cre+;Recql4fl/fl;Trp53fl/fl mice (Fig. 6E). In addition, CDKN1A protein levels in E13.5 forelimb tissues of double mutants were significantly reduced compared to that in the Prx1-Cre+;Recql4fl/fl mice (Fig. 6A, D). Furthermore, the postnatal survival rate was also increased to about 92% (22 out of 24) in these double mutants before the weaning age. When the Col2a1-Cre transgene was used to delete Recql4 and Trp53 in the cartilage, the rib cage defects in Col2a1-Cre+;Recql4fl/fl mutants were also partially rescued (Fig. 7A). We also observed decreased perinatal lethality (4 out of 10) in Col2a1-Cre+;Recql4fl/fl;Trp53fl/fl newborn (P0) pups compared to Col2a1-Cre+;Recql4fl/fl mutants; however, none of the double mutants was able to survive past the weaning age. Next, we investigated if the chondrocyte defects in growth plates were rescued in double mutants. H&E staining of distal femur growth plates showed that the chondrocyte cell density and cell size were similar between Col2a1-Cre+;Recql4fl/fl;Trp53fl/fl and Col2a1-Cre+;Recql4fl/fl embryos at E18.5 (Fig. S11). Similar observation was obtained for Prx1Cre+;Recql4fl/fl;Trp53fl/fl mice (data not shown). Although the chondrocyte morphology was not rescued, TUNEL assay of E18.5 distal femur growth plates demonstrated that the number of apoptotic cells in Col2a1Cre+;Recql4fl/fl;Trp53fl/fl mice was significantly reduced compared to Col2a1-Cre+;Recql4fl/fl mice (Fig. 7B). Collectively, our results show that Recql4 is required for skeletal development via its modulation of p53 function in vivo.

Discussion Mutations in RECQL4 cause three distinct human syndromes , all of which display prominent skeletal developmental abnormalities.(10) Previous attempts to model RECQL4-associated human diseases using Recql4 global knockout mouse models did not focus specifically on examination of the skeletal system. The purpose of this study was to study the molecular mechanisms of RECQL4 in skeletal development by inactivating Recql4 specifically in mesenchymal progenitor cells and chondrocytes. Here we show that using a Cre-loxP recombination system, inactivation of mouse Recql4 in mesenchymal progenitor cells of developing limbs and some craniofacial tissues causes skeletal developmental defects including foreshortened limbs, digit defects, and craniosynostosis, recapitulating the major skeletal findings in RECQL4 associated human disorders. Furthermore, deletion of Recql4 This article is protected by copyright. All rights reserved

11

in cartilage leads to perinatal lethality, rib cage defects, and abnormal growth plate morphology as well as reduced chondrocyte density and increased chondrocyte cell size. The perinatal lethality seen in Col2a1-Cre+;Recql4fl/fl mice was not observed in Prx1-Cre+;Recql4fl/fl mice even though Prx1-Cre transgene is expressed earlier than Col2a1-cre transgene during the differentiation of skeletal progenitor cells. This is likely due to the fact that Prx1Cre transgene is not expressed in the axial skeleton in contrast to the Col2a1-Cre transgene. The cartilage phenotypes in our Recql4 conditional mutants could explain the finding of symmetrical short stature in some patients carrying RECQL4 mutations. Therefore, our genetic mouse model can provide a useful tool to understand the skeletal characteristics of patients with RECQL4 mutations. Interestingly, inactivation of Recql4 in the skeletal system causes increased p53 response leading to increased apoptosis and reduced cell proliferation. Genetic inactivation of Trp53 in both of our Recql4 mutant models can partially rescue the skeletal developmental abnormalities, increased cell death, and postnatal/perinatal lethality in Recql4 single mutants. Our in vivo data suggest that p53 response contributes partially to the molecular mechanisms of skeletal abnormalities in RECQL4 associated human disorders. RECQL4 has been shown to participate in many DNA metabolic processes including initiation of DNA replication, DNA damage repair, and maintenance of telomeres and mitochondrial DNA integrity.(13) As a group, RECQ DNA helicases are felt to be important for maintaining genomic stability.(1) Therefore, the increased p53 activation in our Recql4 mutants could be caused by the replication defects and/or accumulation of DNA damage in our Recql4 mutant mice. We show that deletion of Recql4 in mouse skeletal progenitor cells and chondrocytes leads to significantly increased DNA damage and elevated phosphorylation of p53 at Ser15 which can be phosphorylated by ATM and ATR in response to increased DNA damage in cells.(36-38) Although genetic inactivation of Trp53 rescued the p53 response in Recql4 single mutants, the chondrocyte cell size did not appeared to be rescued in double mutants. This could be due to persistent DNA damage in chondrocytes lacking Recql4, which could then result in a senescence phenotype with increased cell size. In addition, Recql4 could also directly participate in chondrocyte development independent of p53 function. Nevertheless, the in vivo genetic interaction between Recql4 and Trp53 demonstrated by our double conditional knockout mice suggests that RECQL4 can function in skeletal development by modulating p53 activity. This article is protected by copyright. All rights reserved

12

p53 was initially thought to be dispensable for skeletogenesis since Trp53 null mice appeared to have no obvious skeletal developmental defects.(41) However, Lengner et al. showed that osteoblastic inactivation of Mdm2, which encodes an E3 ubiquitin-protein ligase that targets p53 for proteasomal degradation, caused increased p53 activity and multiple skeletal developmental defects.(42) Furthermore, p53 signaling appears to be a negative regulator of osteogenesis by transcriptionally repressing essential transcription factors Osterix or Runx2.(42,43) Interestingly, increased p53 response has also been observed in other mouse models of human syndromes with skeletal abnormalities. Mutations in TCOF1 cause an autosomal dominant inherited disease Treacher Collins syndrome which is characterized by craniofacial developmental anomalies.(44) Mice heterozygous for Tcof1 exhibited increased stabilization of p53 protein and increased apoptosis in neuroepithelia, and the craniofacial defects in these mice could be rescued by inactivating p53 signaling genetically or biochemically.(45) In addition to skeletal findings, RTS patients carrying RECQL4 mutations have a very high risk of developing osteosarcoma.(8) RAPADILINO patients are also at risks for osteosarcoma and lymphoma.(10) Besides the increased genomic instability caused by RECQL4 mutations, our finding of an in vivo genetic interaction between Recql4 and Trp53 in our mouse models suggests that mutations in RECQL4 may lead to dysregulated p53 activity which could result in initiation of tumorigenesis. However, RECQL4 may have other undiscovered functions which are critical for tumorigenesis independent of p53 function. These functions could be masked by p53 activation in Recql4 single mutants and may be readily revealed in Recql4 and Trp53 double mutant mice. Further investigation of the functional role of RECQL4 in skeletal development and tumorigenesis may uncover potential therapeutic targets to treat these human diseases.

Acknowledgments We thank the following people for their contributions to this manuscript: L.A. Donehower, J.T. Yustein, H.C. Zeng, S. Chen, E.M. Munivez, P.M. Campeau, S.C. Nagamani, T.K. Bertin, A. Erez and A.K. Panigrahi for helpful discussions and advice; N. Parikh for SA-β-gal staining protocol; K. L. Moon and T.L. Rideau for technical support. This work was supported by the following grants from the National Institutes of Health: AR059063 (to L.L.W.) from the National Institute Of Arthritis And Musculoskeletal And Skin Diseases, HD070394 (to B.L.) and This article is protected by copyright. All rights reserved

13

HD024064 (to the BCM Intellectual and Developmental Disabilities Research Center) from the Eunice Kennedy Shriver National Institute Of Child Health & Human Development, AI036211, CA125123, and RR024574 (to the BCM Advanced Technology Cores). This work was also funded by: The Rolanette and Berdon Lawrence Bone Disease Program of Texas, BCM Center for Skeletal Medicine and Biology and BCM Curtis and Doris K. Hankamer Foundation Collaborative Research Grant (to B.L. and L.L.W.); Amschwand Sarcoma Cancer Foundation, Carousel, Young Friends of Texas Children’s Cancer Center, Kurt Groten Family Research Scholar’s Program, and Gillson Longenbaugh Foundation (to L.L.W). The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institute of Arthritis and Musculoskeletal and Skin Diseases, the National Institute Of Child Health & Human Development or the National Institutes of Health.

This article is protected by copyright. All rights reserved

14

References

1. Chu WK, Hickson ID. RecQ helicases: multifunctional genome caretakers. Nat Rev Cancer. 2009;9(9):64454. 2. Yu CE, Oshima J, Fu YH, et al. Positional cloning of the Werner's syndrome gene. Science. 1996;272(5259):258-62. 3. Ellis NA, Groden J, Ye TZ, et al. The Bloom's syndrome gene product is homologous to RecQ helicases. Cell. 1995;83(4):655-66. 4. Kitao S, Shimamoto A, Goto M, et al. Mutations in RECQL4 cause a subset of cases of Rothmund-Thomson syndrome. Nat Genet. 1999;22(1):82-4. 5. Larizza L, Roversi G, Volpi L. Rothmund-Thomson syndrome. Orphanet J Rare Dis. 2010;5:2. 6. Wang LL, Levy ML, Lewis RA, et al. Clinical manifestations in a cohort of 41 Rothmund-Thomson syndrome patients. Am J Med Genet. 2001;102(1):11-7. 7. Mehollin-Ray AR, Kozinetz CA, Schlesinger AE, Guillerman RP, Wang LL. Radiographic abnormalities in Rothmund-Thomson syndrome and genotype-phenotype correlation with RECQL4 mutation status. AJR Am J Roentgenol. 2008;191(2):W62-W66. 8. Wang LL, Gannavarapu A, Kozinetz CA, et al. Association between osteosarcoma and deleterious mutations in the RECQL4 gene in Rothmund-Thomson syndrome. J Natl Cancer Inst. 2003;95(9):669-74. 9. Siitonen HA, Kopra O, Kaariainen H, et al. Molecular defect of RAPADILINO syndrome expands the phenotype spectrum of RECQL diseases. Hum Mol Genet. 2003;12(21):2837-44.

This article is protected by copyright. All rights reserved

15

10. Siitonen HA, Sotkasiira J, Biervliet M, et al. The mutation spectrum in RECQL4 diseases. Eur J Hum Genet. 2008;17(2):151-8. 11. Van Maldergem L, Siitonen HA, Jalkh N, et al. Revisiting the craniosynostosis-radial ray hypoplasia association: Baller-Gerold syndrome caused by mutations in the RECQL4 gene. J Med Genet. 2006;43(2):148-52. 12. Bachrati CZ, Hickson ID. RecQ helicases: suppressors of tumorigenesis and premature aging. Biochem J. 2003;374(Pt 3):577-606. 13. Croteau DL, Singh DK, Hoh FL, Lu H, Bohr VA. RECQL4 in genomic instability and aging. Trends Genet. 2012;28(12):624-31. 14. Sangrithi MN, Bernal JA, Madine M, et al. Initiation of DNA replication requires the RECQL4 protein mutated in Rothmund-Thomson syndrome. Cell. 2005;121(6):887-98. 15. Thangavel S, Mendoza-Maldonado R, Tissino E, et al. Human RECQ1 and RECQ4 helicases play distinct roles in DNA replication initiation. Mol Cell Biol. 2010;30(6):1382-96. 16. Xu X, Rochette PJ, Feyissa EA, Su TV, Liu Y. MCM10 mediates RECQ4 association with MCM2-7 helicase complex during DNA replication. EMBO J. 2009;28(19):3005-14. 17. Fan W, Luo J. RecQ4 facilitates UV light-induced DNA damage repair through interaction with nucleotide excision repair factor xeroderma pigmentosum group A (XPA). J Biol Chem. 2008;283(43):29037-44. 18. Schurman SH, Hedayati M, Wang Z, et al. Direct and indirect roles of RECQL4 in modulating base excision repair capacity. Hum Mol Genet. 2009;18(18):3470-83. 19. Singh DK, Karmakar P, Aamann M, et al. The involvement of human RECQL4 in DNA double-strand break repair. Aging Cell. 2010;9(3):358-71.

This article is protected by copyright. All rights reserved

16

20. Crevel G, Vo N, Crevel I, et al. Drosophila RecQ4 is directly involved in both DNA replication and the response to UV damage in S2 cells. PLoS One. 2012;7(11):e49505. 21. Ghosh AK, Rossi ML, Singh DK, et al. RECQL4, the protein mutated in Rothmund-Thomson syndrome, functions in telomere maintenance. J Biol Chem. 2012;287(1):196-209. 22. Croteau DL, Rossi ML, Canugovi C, et al. RECQL4 localizes to mitochondria and preserves mitochondrial DNA integrity. Aging Cell. 2012;11(3):456-66. 23. De S, Kumari J, Mudgal R, et al. RECQL4 is essential for the transport of p53 to mitochondria in normal human cells in the absence of exogenous stress. J Cell Sci. 2012;125(Pt 10):2509-22. 24. Ichikawa K, Noda T, Furuichi Y. [Preparation of the gene targeted knockout mice for human premature aging diseases, Werner syndrome, and Rothmund-Thomson syndrome caused by the mutation of DNA helicases]. Nippon Yakurigaku Zasshi. 2002;119(4):219-26. 25. Hoki Y, Araki R, Fujimori A, et al. Growth retardation and skin abnormalities of the Recql4-deficient mouse. Hum Mol Genet. 2003;12(18):2293-9. 26. Mann MB, Hodges CA, Barnes E, et al. Defective sister-chromatid cohesion, aneuploidy and cancer predisposition in a mouse model of type II Rothmund-Thomson syndrome. Hum Mol Genet. 2005;14(6):81325. 27. Logan M, Martin JF, Nagy A, et al. Expression of Cre Recombinase in the developing mouse limb bud driven by a Prxl enhancer. Genesis. 2002;33(2):77-80. 28. Ovchinnikov DA, Deng JM, Ogunrinu G, Behringer RR. Col2a1-directed expression of Cre recombinase in differentiating chondrocytes in transgenic mice. Genesis. 2000;26(2):145-6. 29. Ford-Hutchinson AF, Ali Z, Lines SE, et al. Inactivation of Pten in osteo-chondroprogenitor cells leads to epiphyseal growth plate abnormalities and skeletal overgrowth. J Bone Miner Res. 2007;22(8):1245-59. This article is protected by copyright. All rights reserved

17

30. Wang W, Nyman JS, Ono K, et al. Mice lacking Nf1 in osteochondroprogenitor cells display skeletal dysplasia similar to patients with neurofibromatosis type I. Hum Mol Genet. 2011;20(20):3910-24. 31. Jonkers J, Meuwissen R, van Der GH, et al. Synergistic tumor suppressor activity of BRCA2 and p53 in a conditional mouse model for breast cancer. Nat Genet. 2001;29(4):418-25. 32. Engin F, Yao Z, Yang T, et al. Dimorphic effects of Notch signaling in bone homeostasis. Nat Med. 2008;14(3):299-305. 33. Livak KJ, Schmittgen TD. Analysis of relative gene expression data using real-time quantitative PCR and the 2(-Delta Delta C(T)) Method. Methods. 2001;25(4):402-8. 34. Seo HS, Serra R. Tgfbr2 is required for development of the skull vault. Dev Biol. 2009;334(2):481-90. 35. Mah LJ, El-Osta A, Karagiannis TC. gammaH2AX: a sensitive molecular marker of DNA damage and repair. Leukemia. 2010;24(4):679-86. 36. Canman CE, Lim DS, Cimprich KA, et al. Activation of the ATM kinase by ionizing radiation and phosphorylation of p53. Science. 1998;281(5383):1677-9. 37. Banin S, Moyal L, Shieh S, et al. Enhanced phosphorylation of p53 by ATM in response to DNA damage. Science. 1998;281(5383):1674-7. 38. Tibbetts RS, Brumbaugh KM, Williams JM, et al. A role for ATR in the DNA damage-induced phosphorylation of p53. Genes Dev. 1999;13(2):152-7. 39. Shieh SY, Ikeda M, Taya Y, Prives C. DNA damage-induced phosphorylation of p53 alleviates inhibition by MDM2. Cell. 1997;91(3):325-34. 40. Lu H, Fang EF, Sykora P, et al. Senescence induced by RECQL4 dysfunction contributes to RothmundThomson syndrome features in mice. Cell Death Dis. 2014;5:e1226. This article is protected by copyright. All rights reserved

18

41. Donehower LA, Harvey M, Slagle BL, et al. Mice deficient for p53 are developmentally normal but susceptible to spontaneous tumours. Nature. 1992;356(6366):215-21. 42. Lengner CJ, Steinman HA, Gagnon J, et al. Osteoblast differentiation and skeletal development are regulated by Mdm2-p53 signaling. J Cell Biol. 2006;172(6):909-21. 43. Wang X, Kua HY, Hu Y, et al. p53 functions as a negative regulator of osteoblastogenesis, osteoblastdependent osteoclastogenesis, and bone remodeling. J Cell Biol. 2006;172(1):115-25. 44. Wise CA, Chiang LC, Paznekas WA, et al. TCOF1 gene encodes a putative nucleolar phosphoprotein that exhibits mutations in Treacher Collins Syndrome throughout its coding region. Proc Natl Acad Sci U S A. 1997;94(7):3110-5. 45. Jones NC, Lynn ML, Gaudenz K, et al. Prevention of the neurocristopathy Treacher Collins syndrome through inhibition of p53 function. Nat Med. 2008;14(2):125-33.

This article is protected by copyright. All rights reserved

19

Figure Legends Fig. 1. Prx1-Cre+;Recql4fl/fl mice have limb developmental defects. (A, B) Abnormal forelimbs (arrows) of Prx1Cre+;Recql4fl/fl mice shown at P0 (A) and at E15.5 (B). (C-E) Representative Alizarin red and Alcian blue staining of forelimbs (C) and hindlimbs (D) of Prx1-Cre+;Recql4fl/fl and littermate control mice at P0 and at 3 weeks (E). Arrows point to missing or malformed digits (B-D); asterisk indicates abnormal joint ossification (E). Ctrl, littermate control. See also Fig. S4.

Fig. 2. Prx1-Cre+;Recql4fl/fl mice display craniosynostosis. (A) Representative Alizarin red and Alcian blue staining of P0 skulls from Prx1-Cre+;Recql4fl/fl and Ctrl mice. (B) Representative von Kossa and nuclear fast red staining of coronal sutures in skulls from E16.5 Prx1-Cre+;Recql4fl/fl and Ctrl mice. Scale bars: 50 μm. (C, D) Micro computed tomography (Micro-CT) scans of skulls from three week-old Prx1-Cre+;Recql4fl/fl and Ctrl mice showing coronal (C) and squamosal (D) sutures. Scale bars: 1.0 mm. (E) Single scan section of micro-CT displaying coronal and squamosal sutures in skulls from Prx1-Cre+;Recql4fl/fl and Ctrl mice. C, coronal suture; Sq, squamosal suture. Arrows point to the location of coronal and/or squamosal sutures (A-E).

Fig. 3. Prx1-Cre+;Recql4fl/fl and Col2a1-Cre+;Recql4fl/fl mice have growth plate abnormalities. (A) Representative hematoxylin and eosin (H&E) staining of E18.5 distal femur growth plates of Prx1-Cre+;Recql4fl/fl and Ctrl mice. Boxed areas are shown at higher magnification below. RZ, resting zone; PZ, proliferating zone; HZ, hypertrophic zone. Scale bars: 50 μm. (B) Quantification of cell density in different zones of E18.5 distal femur growth plates of Prx1-Cre+;Recql4fl/fl and Ctrl mice. n=3, **p