Osteoblast-Specific Overexpression of Nucleolar Protein NO66/RIOX1 in Mouse Embryos Leads to Osteoporosis in Adult Mice

In previous study, we showed that nucleolar protein 66 (NO66) is a chromatin modifier and negatively regulates Osterix activity as well as mesenchymal progenitor differentiation. Genetic ablation of the NO66 (RIOX1) gene in cells of the Prx1-expressing mesenchymal lineage leads to acceleration of osteochondrogenic differentiation and a larger skeleton in adult mice, whereas mesenchyme-specific overexpression of NO66 inhibits osteochondrogenesis resulting in dwarfism and osteopenia. However, the impact of NO66 overexpression in cells of the osteoblast lineage in vivo remains largely undefined. Here, we generated osteoblast-specific transgenic mice overexpressing a FLAG-tagged NO66 transgene driven by the 2.3 kB alpha-1type I collagen (Col1a1) promoter. We found that overexpression of NO66 in cells of the osteoblast lineage did not cause overt defects in developmental bones but led to osteoporosis in the long bones of adult mice. This includes decreased bone volume (BV), bone volume density (bone volume/total volume, BV/TV), and bone mineral density (BMD) in cancellous compartment of long bones, along with the accumulation of fatty droplets in bone marrow. Ex vivo culture of the bone marrow mesenchymal stem/stromal cells (BMSCs) from adult Col1a1-NO66 transgenic mice showed an increase in adipogenesis and a decrease in osteogenesis. Taken together, these data demonstrate a crucial role for NO66 in adult bone formation and homeostasis. Our Col1a1-NO66 transgenic mice provide a novel animal model for the mechanistic and therapeutic study of NO66 in osteoporosis.


Introduction
In mammals, fat and long bones form through intramembranous and endochondral ossifcation, respectively. In intramembranous ossifcation, mesenchymal stem cells diferentiate directly into cells of the osteoblast lineage, including preosteoblast, immature osteoblast, mature osteoblast, and osteocyte. In endochondral ossifcation, mesenchymal stem cells diferentiate into osteochondroprogenitors, which then segregate into chondrocytes that form a cartilage template or osteoprogenitors that diferentiate toward the osteoblast lineage to form bone tissue. Maintenance of healthy bone mass depends upon constant homeostasis, a balance between osteoblastmediated bone formation and hematopoietic osteoclast-mediated bone resorption. Imbalance between these two processes results in osteoporosis, which is a common bone disease in the aging population and especially in postmenopausal women [1]. Another mechanism of osteoporosis involves dysregulated osteogenesis and/or adipogenesis from bone marrow-derived mesenchymal stem/stromal cells (BMSCs), which have the capacity to diferentiate into both osteoblastic and adipocytic lineages [2,3]. In aging patients with osteoporosis, the decrease in bone formation often coincides with the increase in marrow adiposity [4,5]. However, it remains unclear how these two processes are coupled in the bone marrow.
Commitment and diferentiation of mesenchymal stem cells or BMSCs into the osteoblastic lineage are tightly controlled by specifc signaling molecules including bone morphogenetic protein 2 (BMP2) and insulin-like growth factor 1 (IGF1), as well as essential bone-forming transcription factors, such as runt-related transcription factor 2 (Runx2) and Osterix (Osx) [6][7][8]. In mice with a germline deletion of the Osx (Sp7) gene, formation of both endochondral and intramembranous bones was completely abolished [6]. Te activity of Osx can be regulated by the nucleolar protein 66 (NO66), also known as Ribosomal Oxygenase 1 (RIOX1). We previously showed that NO66 binds to Osx and inhibits the transactivation of its target genes bone sialoprotein (Bsp) (Ibsp) and Osteocalcin (Oc) (Bglap) [9]. In addition to inhibiting Osx activity, NO66 also negatively regulates osteogenic diferentiation of mesenchymal stem cells both in vitro and in vivo [9][10][11]. In the cultured pluripotent mesenchymal precursor C2C12 cell line, overexpression of NO66 inhibits osteogenic diferentiation and expression of the osteoblast-diferentiation marker gene [9]. In transgenic mice, overexpression of NO66 in cells of Prx1-expressing mesenchymal origin inhibits osteogenesis and reduces the number of Osxexpressing preosteoblasts and osteoblasts, concurrent with the reduction in IGF1R/Akt signaling and expression of bone formation marker genes, leading to both growth retardation and osteopenia phenotype [11]. Conversely, depletion of NO66 in cells of Prx1-expressing mesenchymal origin increases the number of preosteoblasts and osteoblasts upregulates expression of bone-forming signal molecules and marker genes, including Bmp2, Igf1, Osx (Sp7), type I collagen a1 (Col1a1), and Bsp (Ibsp). As a result, NO66-null mice exhibit an acceleration of skeletogenesis as well as bone formation [10], demonstrating an essential role for NO66 in commitment and osteogenic diferentiation of mesenchymal stem cells. Moreover, our previous in vitro studies showed that NO66 is a negative regulator of osteoblast diferentiation [9,12]. Knockdown of NO66 in mouse preosteoblast MC3T3-E cell line accelerates the BMP2-induced osteoblast diferentiation and maturation [9]. However, it remains unclear whether NO66 is important in osteoblast diferentiation and maturation in vivo.
Here, we generated transgenic mice overexpressing a FLAG-tagged NO66 transgene driven by the 2.3-kb Col1a1 promoter to study the in vivo role of NO66 in osteoblastogenesis. Previous studies indicated that this promoter is highly active in the osteoblast lineage of mouse embryos beginning around embryonic day 14.5 (E14.5) [13,14]. We found that osteoblast-specifc overexpression of NO66 shows minimal alteration in the formation of developmental bone. In contrast, it afects the homeostasis of mature bones, including the impaired formation of cancellous bone and increased accumulation of bone marrow fat (BMF), the predominant feature of osteoporosis. Generation and genotyping of the Col1a1-NO66 transgenic mice were performed, as previously described [11]. Briefy, a FLAG-tagged cDNA [9] was subcloned downstream of the 2.3 kB Col1a1-promoter in a pBluescript SK vector, to generate a new Col1a1-NO66 construct ( Figure  1(a)). Plasmid DNA of this construct was microinjected into fertilized B6D2 F1 eggs, which were then transferred into CD1 foster mothers. Transgenic ofspring were crossed three times with wild-type (WT) C57BL/6 mice to establish transgenic lines for the study. Te founders and ofspring of transgenic mice were genotyped by PCR assays using tail DNA and Hgh primers: 5′-TGTCTGACTAGGTGTCCTTC; 5′-GCAAGCAACTCAAATGTCC, as described previously [11].

RNA Extraction and RT-qPCR
Assay. RNA extraction and real-time quantitative polymerase chain reaction (RT-qPCR) were performed as previously described [10,11]. Briefy, total RNA of mouse calvaria at embryonic day 18.5 (E18.5) was isolated using TRIzol reagent (Invitrogen) according to the manufacturer's protocol. Total RNA was pretreated with TURBO DNase (Ambion) to remove genomic DNA contamination and then reverse-transcribed into cDNAs using a high-capacity reverse transcription kit (Applied Biosystems). For RT-qPCR assay, 50 ng cDNA and a gene-specifc TaqMan primer-probe, including mouse NO66, Bsp, Col1a1, and Alp (Applied Biosystems) were used in each PCR reaction (triplicate). Levels of mRNA expression were normalized relative to the Hprt housekeeping gene.

Specimen CT (SP-CT) Scanning.
Femurs of mice at three-month (3 m) old were isolated and fxed in 4% paraformaldehyde overnight at room temperature and then stored in 70% ethanol for SP CT scanning (GE Medical Systems, London, Ontario). Scanning and analyses of bone images were conducted in the Small Animal Imaging Facility at the University of Texas MD Anderson Cancer Center, as described previously [10,15]. (BMSC). Primary culture of BMSC was performed following the procedures described previously [16] with minor modifcations. Briefy, femoral and tibia bones were isolated from 3 m old nontransgenic and Col1a1-NO66 transgenic mice. After removing connective tissues, bones were kept in ice-cold PFA solution containing 1 × PBS (without Ca 2+ and Mg 2+ )/2% FBS (v/v)/1 × Pen/Strep antibiotics. Bone marrow was fushed out using 4-5 ml of alpha-MEM supplemented with 5% FBS and 1 × Pen/Strep antibiotics (using a 25 G needle) and transferred into a 50 ml tube on ice. Te bone marrow plugs were homogenized by repeated passage through a 21 G needle, mixed 1 : 1 with distilled sterile water for 2 min to lyse erythrocytes, and then centrifuged at 380 × g for 5 min. Te cell pellet was resuspended in complete culture medium (alpha-MEM, 20% FBS, 2 mM L-glutamine, 100 U/ml of Pen/Strep), and cultured in a 37°C humidifed incubator containing 3% O 2 , 5% CO 2 , and 90% N 2 . Te culture medium was changed every other day. Upon reaching 80% confuence, cells were trypsinized and replated for further analysis.

Isolation and Culture of Bone Marrow Stromal Cells
For osteogenic diferentiation, cells were seeded at a density of 15 × 10 3 cells/cm 2 in normal culture medium. On day six, the culture medium was replaced with osteogenic medium containing the complete alpha-MEM supplemented with 10 mM sodium β-glycerophosphate, 10 nM dexamethasone, and 50 μg/ml ascorbic acid. To initiate adipogenic diferentiation, the cells were cultured in complete alpha-MEM supplemented with 5 μg/ml insulin, 1 μM rosiglitazone, and 1 μM dexamethasone [17].

Statistical Analysis.
All statistical results are presented as the mean ± standard error (SE). Te diferences between groups were calculated using the two-tailed Student t-test and p values <0.05 were considered statistically signifcant.

Generation of Osteoblast-Specifc NO66 Transgenic Mice.
To examine the role of NO66 in mature osteoblasts in vivo, we generated transgenic mice overexpressing the FLAG-tagged NO66 in cells of the osteoblastogenic lineage driven by a 2.3kb type I collagen promoter (Col1a1) (Figure 1(a)). In mice, the 2.3 kB Col1a1 promoter is active around embryonic day 14.5 (E14.5) and specifcally drives transgene expression at high levels in osteoblasts and odontoblasts [14,18]. We performed pronuclear injections of the Col1a1-NO66 construct and generated two stable transgenic mouse lines, named Col1a1-NO66-1 and Col1a1-NO66-2. To determine expression of NO66 transgene in membranous bone tissue, we performed RT-qPCR assays using total RNA of calvaria from E18.5 wild type (WT) control and transgenic embryos. We observed the high-level mRNA expression of NO66 in both transgenic lines compared with WT control embryos (Figure 1

Formation of Intramembranous and Endochondral Bones
in Col1a1-NO66 Mice during Development. Next, we sought to examine whether NO66 overexpression in cells of the osteoblast lineage could afect intramembranous bone formation during early development. We performed RT-qPCR assays using total RNA of calvaria from E18.5 mouse embryos for changes in expression of bone-formation marker genes. We did not observe signifcant changes in the mRNA expression of bone sialoprotein (Bsp), Col1a1, and alkaline phosphatase (Alp) in NO66 transgenic embryos from bone lines when comparing with those of WT controls (Figure 3(a); p > 0.05). Tis suggests that NO66 overexpression in cells of the osteoblast lineage may have minimal infuence in the formation of intramembranous bones during embryogenesis.
We then examined the phenotypic changes in endochondral bones of mouse embryos and newborn mice. We performed histological examination of long bones of nontransgenic and Col1a1-NO66 transgenic mice from both lines at diferent development stages. In femur sections of E15.5 NO66 transgenic embryos, we did not observe any reduction in mineralized bone tissues marked by either von Kossa or Alizarin red staining when comparing with WTcontrol embryos (S Figures 2A-2D). H&E staining of femur sections of newborn WT and transgenic mice also showed similar morphological appearance (S Figures 2E  and 2F). When we stained the femoral sections of newborn WT and transgenic mice with Alcian blue (for cartilage) and von Kossa (for mineralized bone tissue), we also observed the similar staining pattern between the WT and  transgenic mice (Figures 3(b)-3(e)). Tese data indicate that NO66 overexpression in cells of the osteoblast lineage has minimal inhibitory efect on the formation of endochondral bones during embryonic and early mouse development.  We then performed the histological examination of the long bones from 3 m to 4 m-old male and female mice. H&E staining of the femur sections of 4 m-old male transgenic mice revealed more empty-spaces and less marrow tissues than those in WT controls (S Figure 4). When we stained the femur sections of 3 m-old female WT and transgenic (from both lines) mice, we observed fewer trabeculae underneath their growth plates than those in the WT control mice (S Figures 5A-5C and Figures 5(a) and 5(b)). Interestingly, we also observed a markedly increased number of airflledballoon-like cells (or empty cells), which implicate adipocytes (identifed by yellow arrows) [19], in the bone marrow of transgenic femur sections when compared with WT controls (S Figures 5D-5F and Figures 5(a) and 5(b)).

Loss of Bone and Accumulation of Bone Marrow Fat in
In comparison, the thickness of cortical bones around the middle shaft of those transgenic mice showed minimal alteration when compared with WT controls (S Figures 5G-5I), which was consistent with the SP-CT scanning results (Figure 4  exhibit an inverse relationship in osteoporotic patients as well as animal models [20][21][22][23]. Terefore, we speculated that the inhibition of osteogenesis by osteoblast-specifc overexpression of NO66 might increase adipogenesis from the BMSCs of adult Col1a1-NO66 mice. To test this hypothesis, we performed ex vivo culture of BMSCs from the 3 m-old WT and transgenic mice, and then induced the primary BMSCs to undergo adipogenic/ osteogenic diferentiation. After 7 days of adipogenic induction, Oil Red O staining showed more fatty droplets formed in BMSCs from NO66 transgenic mice compared to that in WT controls (Figures 5(d)-5(f ); p < 0.001). In contrast, osteogenic induction showed signifcant less mineralization in cultured transgenic BMSCs (Figures 5(g)-5(i); p < 0.001). To further confrm these observations, we examined mRNA expression of NO66, Osx, Bsp, and Fabp4 in the primary cultured BMSCs after induction of osteogenesis or adipogenesis by RT-qPCR. In osteogenic-diferentiated BMSCs derived from those Co1a1-NO66 transgenic mice, we observed a high-level increase in expression of NO66 and a marked decrease in expression of the boneforming marker Bsp; however, we only observed a minimal alteration in expression of Osx (Figure 6), consistent with our previous study [9]. By comparison, in adipogenic-diferentiated BMSCs expression of Fabp4 (a marker of fat formation) was highly increased in the NO66 transgenic mice when compared with nontransgenic controls ( Figure 6). Tese results demonstrated an inverse correlation between BMSCs-derived adipogenic and osteogenic diferentiation in transgenic mice.

Discussion
In this study, we generated an osteoblast-specifc NO66 transgenic mouse model to investigate the role of NO66 in mammalian bone formation and homeostasis in vivo. Our data showed that the overexpression of NO66 from the 2.3 kB Col1a1 promoter in cells of the osteoblast lineage has minimal impacts on the formation of both endochondral and intramembranous bones during the early developmental stages but afects bone formation and homeostasis in adult mice. Tis includes the decreased formation of cancellous bone tissues and fat accumulation in the bone marrow of long bones-an osteoporotic phenotype in adult mice. Te imbalance between adipogenesis and osteogenesis in the bone marrow is the predominant feature of osteoporosis [24,25]. Numerous studies have demonstrated an inverse association between the accumulation of adipocytes and the loss of bone tissues in the marrow of aging patients with osteoporosis [4,[20][21][22][23]26]. In our present study, the loss of cancellous bone tissues and the accumulation of BMF, which we observed in the long bones of adult Col1a1-NO66 transgenic mice (Figures 4 and 5), resemble the features of age-related osteoporosis in humans. Terefore, our Col1a1-NO66 transgenic mouse model constitutes a novel animal model for age-related osteoporosis.
Te commitment and lineage diferentiation of mesenchymal stem cells are controlled by specifc transcription factors and signaling molecules [25,[27][28][29]. In our previous studies, we showed that deletion or overexpression of NO66 in the Prx1-expressing mesenchymal lineage afects diferentiation of both osteogenic and chondrogenic progenitors, resulting in an abnormal skeleton and bone mass in adult animals [10,11]. Here, we observed that transgenic mice with overexpression of NO66 in osteoblast lineage have minimal alteration in bone length (S Figure 3) or cartilage morphology (Figure 3(c)). Tis suggests that the osteoblastspecifc overexpression of NO66 has little impact on chondrogenesis. Te trabecular bone loss observed in the bone marrow of adult Col1a1-NO66 transgenic mice is consistent with our previous in vitro study showing that NO66 is a negative regulator of osteoblast diferentiation [9]. Given that the accumulation of BMF is often inversely correlated with bone loss in osteoporosis [4], it is likely that the increased fatty tissue observed in transgenic bone marrow is a secondary efect induced by an imbalance between osteogenesis and adipogenesis from BMSCs. However, we cannot rule out that certain signaling molecules, which are important for BMSC lineage commitment and diferentiation, may also be afected by NO66 overexpression. Te precise molecular mechanism underlying this process requires further investigation.
NO66 is a negative regulator of Osx activity [9][10][11][12]. NO66 knockdown in cultured preosteoblasts markedly increases osteoblast diferentiation and expression of Osx target genes [9]. Similarly, genetic ablation of NO66 in cells of the Prx1-expressing mesenchymal lineage accelerates osteoblast diferentiation, expression of Osx-target genes Bsp, and Oc [10]. In contrast, mesenchymal overexpression of NO66 inhibits osteoblast diferentiation, Osx-target gene expression, and bone formation [10,11], highlighting the inverse relationship between the level of NO66 and the activity of Osx. However, a previous study showed that osteoblast-specifc deletion of Osx driven by this 2.3 kB Col1a1 promoter does not cause clear bone defects in embryos and newborn pups but leads to osteopenia phenotypes in growing mice, including decreased BMD in the lumbar vertebra, thinner cortex of long bones, and immature trabecular bones [15]. Since NO66 acts as an inhibitor of Osx activity, in the transgenic mice overexpressing NO66 in the osteoblast lineage driven by the same 2.3 kB Col1a1 promoter, one would predict that the phenotypes of Col1a1-NO66 mice should be similar to those of the osteoblast-specifc Osx mutant mice. In agreement with this prediction, we did not observe overt defects in those Col1a1-NO66 embryos and newborn mice when comparing with the nontransgenic controls, phenotypically and molecularly (Figure 3 and S Figure 2). Instead, we observed a signifcant decrease in bone volume fraction and BMD in the distal femoral bones of adult Col1a1-NO66 mice ( Figure 4). However, the cortical bones of our adult Col1a1-NO66 mice did not appear thinner or more porous, unlike those observed in the osteoblastspecifc Osx mutant mice [15], implicating a partial inhibition of Osx by NO66 overexpression. In addition, the predominant feature of our Col1a1-NO66 mice is increased adipogenesis in the bone marrow of long bones, which was not shown in the osteoblast-specifc Osx mutant mice [15]. Tis suggests that the phenotypic change observed in the long bones of the Col1a1-NO66 mice may not simply be

Conclusions
Our studies using a Col1a1-NO66 transgenic mouse model provide the frst in vivo evidence that the chromatin modifer NO66 plays a crucial role in the regulation of aging bone formation and homeostasis. Our Col1a1-NO66 transgenic mice constitute a new epigenetic animal model to study the chromatin alterations in the context of age-related osteoporosis, as well as a potential preclinical animal model for the evaluation of select novel drugs for the prevention and treatment of human osteoporosis.

Data Availability
All data are fully available without restriction. All relevant data are within the manuscript and its supporting information fles.

Conflicts of Interest
Te authors declare that they have no conficts of interest.