Cell cycle exit during bortezomib‐induced osteogenic differentiation of mesenchymal stem cells was mediated by Xbp1s‐upregulated p21Cip1 and p27Kip1

Abstract Mesenchymal stem cells (MSCs) are multipotent cells capable of differentiating into a variety of cell types. Bortezomib, the first approved proteasome inhibitor used for the treatment of multiple myeloma (MM), has been shown to induce osteoblast differentiation, making it beneficial for myeloma bone disease. In the present study, we aimed to investigate the effects and underlying mechanisms of bortezomib on the cell cycle during osteogenic differentiation. We confirmed that low doses of bortezomib can induce MSCs towards osteogenic differentiation, but high doses are toxic. In the course of bortezomib‐induced osteogenic differentiation, we observed cell cycle exit characterized by G0/G1 phase cell cycle arrest with a significant reduction in cell proliferation. Additionally, we found that the cell cycle exit was tightly related to the induction of the cyclin‐dependent kinase inhibitors p21Cip1 and p27Kip1. Notably, we further demonstrated that the up‐regulation of p21Cip1 and p27Kip1 is transcriptionally dependent on the bortezomib‐activated ER stress signalling branch Ire1α/Xbp1s. Taken together, these findings reveal an intracellular pathway that integrates proteasome inhibition, osteogenic differentiation and the cell cycle through activation of the ER stress signalling branch Ire1α/Xbp1s.

medicine and immune diseases. 3,4 Nonetheless, our understanding of the mechanisms by which MSCs impact clinical and immunological abnormalities in these diseases remains incomplete. For instance, MSCs have been suggested to be attracted to primary tumours, thereby contributing to tumour metastasis as well as drug resistance. [5][6][7][8][9][10] On the other hand, chemotherapeutic drug treatments have been shown to alter the phenotype and differentiation potential of MSCs, and even render them more chemoprotective of the tumour cells. [11][12][13][14] Accordingly, further therapeutic efforts to target MSCs may help to prevent chemoresistance and disease relapse in tumours.
The proteasome is a central component of the protein degradation machinery in eukaryotic cells. Inhibition of the proteasome has emerged as a powerful approach for the treatment of multiple myeloma (MM), a haematologic cancer characterized by the accumulation of malignant plasma cells in the bone marrow (BM). 15 Bortezomib, as the first approved proteasome inhibitor, has been used as a first-line drug for the treatment of MM. 16 In addition to its direct antitumour activity, bortezomib also exerts bone protection effects in MM patients. Of note, the effect of bortezomib on bone formation has been suggested to be related to the enhanced differentiation of MSCs towards osteoblasts. 13,17,18 Although the fate determination and terminal differentiation of MSCs are known to be tightly controlled by diverse transcription factors and signalling pathways, many observations have identified important connections between cell fate decisions and the cell cycle machinery in pluripotent stem cells. [19][20][21] For example, terminal differentiation is usually associated with cell cycle exit, and the transition through mitosis and G 1 phase plays an essential role in establishing a window of opportunity for pluripotency exit and the initiation of differentiation. 19 The purpose of this study was to determine the mechanisms by which the bortezomib-induced differentiation of MSCs towards osteoblasts affects the cell cycle machinery. The antibodies against p21 Cip1 , p27 Kip1 , X-box-binding protein 1 (Xbp1s), activating transcription factor 6 (Atf6), 78 kDa glucose-regulated protein (Grp78), C/EBP homologous protein (Chop), cyclin D3, cyclin E1, cyclin-dependent kinase 2 (CDK2), cyclin-dependent kinase 4 (CDK4) and β-actin were obtained from Proteintech (Wuhan, Hubei, China), and the antibody against activating transcription factor 4 (Atf4) was obtained from Santa Cruz Biotechnology (Dallas, TX, USA). All other chemicals were obtained from Sigma-Aldrich (Burlington, MA, USA) unless otherwise specified.

| mBM-MSC isolation and expansion
Inbred male C57BL/6 mice aged 4-6 weeks were purchased from the animal centre of Xi'an Jiaotong University, housed and treated according to conditions approved by the Ethical Committee for Animal Experiments of the Xi'an Jiaotong University Health Science Center (No. 2015-123). In brief, individual mice were killed by cervical dislocation, and the whole body was soaked thoroughly with 70% ethanol solution for 2 min. Following dissection of the hind legs and vertebrae, all tissues were removed from around the bones and the bones placed in a Petri dish with 5 mL of Dulbecco's modified Eagle's medium (DMEM; HyClone, Logan, Utah, USA). The ligaments between femur and hip were cut, and the bone was cut below the ankle joint. The tibia was separated from the femur by bending slightly at the knee joint.
Holding the femur/tibia with a sterile forceps, both epiphyses were then removed with a sterile scissors. The contents of the bones were then flushed with a 1-mL syringe with a needle, into a Petri dish with 5 mL of medium. The medium was then aspirated and flushed several times to disperse the bone marrow cells. The vertebrae were then crushed with the backside of a 5-mL syringe in 5 mL of medium. The cell suspensions were then filtered through a nylon filter (70 µM mesh diameter) into a 50-mL tube.

| Alizarin Red S staining
mBM-MSCs were plated in 35-mm-diameter culture dishes and grown in DMEM containing 10% FBS, 100 U/mL penicillin, 100 μg/ mL streptomycin, and 2 mM L-glutamine at 37°C in a humidified incubator with 5% CO 2 in the air. When cell density reached 70%-

| RNA purification and Real-time PCR analysis
Total RNA from cells was extracted using Ultrapure RNA Kit  Table S1.

| Western blotting analysis
Western blotting was performed as described previously. 16

| Chromatin Immunoprecipitation
Chromatin immunoprecipitation (ChIP) was performed as described previously. 16 Briefly, mBM-MSCs treated with vehicle or 2.5 nM bortezomib for 16 h were cross-linked with 1% formaldehyde.  (Table S2) covering the putative regions of the p21 Cip1 and p27 Kip1 promoters.

| Statistical analysis
Results were statistically analysed in GraphPad Prism 5.0 (GraphPad Software Inc, San Diego, CA, USA) and presented as mean ± SEM. Statistically significant differences between two groups were assessed by two-tailed unpaired t test. P < .05 was considered statistically significant.

| Bortezomib decreases mBM-MSC cell viability
To assess the effects of bortezomib on cell viability, we performed MTT assay in mBM-MSCs grown in various concentrations of bortezomib for 24 h and 48 h. As shown in Figure 1A  much higher amounts of calcium phosphate crystals than the control cells. To further prove the regulatory role of bortezomib in osteogenesis, we measured the changes of the other bone formation markers and confirmed that bortezomib can induce the expression of Runx2, Sp7, Col1A1, alkaline phosphatase (ALP) and osteocalcin (OCN/BGLAP) ( Figure S1).

| Bortezomib inhibits mBM-MSC cell proliferation
Given the potential association between cell differentiation and proliferation, we further investigated the effects of bortezomib on cell proliferation during bortezomib-induced osteogenic differentiation. By using EdU incorporation assay, we found that bortezomib dose-dependently decreased the numbers of EdUpositive mBM-MSCs, which represent the proliferating population ( Figure 2A and B).

| Bortezomib induces G 0 /G 1 phase cell cycle arrest
Based on the finding above that bortezomib inhibits the proliferation of mBM-MSCs, we further analysed the effect on the cell cycle distribution. As shown in Figure 2C and 2D, bortezomib treatment for 24 h significantly induced G 0 /G 1 phase arrest in mBM-MSCs.
Compared with the control group, the percentages of G 0 /G 1 phase of cells treated with 2.5 nM and 5 nM of bortezomib were increased from 55.14 ± 5.132 to 67.36 ± 6.067 and 68.117 ± 2.743, respectively.
In contrast, the proportion of S phase cells were decreased from 32.017 ± 1.991 to 21.807 ± 2.844 and 19.940 ± 4.321. However, there was no significant change in the proportion of cells in the G 2 /M phase.

| Bortezomib triggers changes in cell cycle machinery
To further determine the molecular mechanism underlying G 0 /G 1 phase cell cycle arrest, we examined the effects of bortezomib on the expression of G 0 /G 1 phase-associated cyclins, cyclin-dependent kinases (CDKs) and cyclin-dependent kinase inhibitors (CKIs). As shown in Figure 3A, bortezomib treatment has no effects on the expression of cyclin D3 and cyclin E1. However, the expression of Cdk2 and Cdk4 was markedly decreased by bortezomib ( Figure 3B).
In contrast, the expression of p21 Cip1 and p27 Kip1 was significantly increased by bortezomib ( Figure 3C). In line with the increase in p21 Cip1 and p27 Kip1 at the protein level, we further observed that the mRNA levels of p21 Cip1 and p27 Kip1 were significantly up-regulated by bortezomib ( Figure 3D).

| ER stress signalling Xbp1s is involved in the transcriptional regulation of bortezomib-induced p21 Cip1 and p27 Kip1
To further investigate whether ER stress is involved in bortezomibinduced G 0 /G 1 phase arrest, we analysed the expression of key ER stress signalling-related proteins, including the ER stress markers Grp78 and Chop, as well as three major regulators Xbp1s, Atf4 and Atf6, in response to bortezomib treatment. As shown in Figure 4A To validate the regulatory relationship between the activation of ER stress signalling and the induction of p21 Cip1 and p27 Kip1 , we used MKC3946 (an inhibitor of inositol-requiring enzyme 1α (IRE1α)) and GSK2606414 (an inhibitor of double-stranded RNA-activated protein kinase [PKR]-like ER kinase (PERK)) to block the bortezomib-activated ER stress signalling pathways accordingly. As shown in Figure 4B, when the bortezomib-induced Xbp1s was aborted by MKC3946, we also found a decrease in the expression of p21 Cip1 and p27 Kip1 .
However, when using GSK2606414 to block PERK-Atf4 signalling, although we observed a marked decrease in Atf4, it had no significant effects on the expression of p21 Cip1 and p27 Kip1 ( Figure 4C). More importantly, we further confirmed that the MKC3946-aborted up-regulation of p21 Cip1 and p27 Kip1 happened at the mRNA level, validated by real-time PCR ( Figure 4D). Given the potential effects of MKC3946 on the cells, we further analysed the changes of cell cycle and found that the combination of MKC3946 with bortezomib significantly decreased the percentage of S phase, but MKC3946 alone had no effects on the cell cycle distribution ( Figure S2). These results strongly suggest that the activation of Xbp1s may be tightly associated with the expression of p21 Cip1 and p27 Kip1 .

| Enforced expression of XBP1s up-regulates p21 Cip1 and p27 Kip1 and induces G 0 /G 1 cell cycle arrest in mBM-MSCs
To further investigate the role of Xbp1s in cell cycle arrest, we used a Tet-On lentiviral system to overexpress human spliced XBP1 in mBM-MSCs. We found that enforced expression of XBP1s inhibited

| Transcriptional regulation of p21 Cip1 and p27 Kip1 by Xbp1s
To elucidate the potential transcriptional regulation of Xbp1s, we sought to determine whether Xbp1s binds to the p21 Cip1

| D ISCUSS I ON
The development of multicellular organisms relies on the temporal and spatial control of cell proliferation and differentiation. 19,[22][23][24] Developmental signals not only direct cell cycle progression but also set the frame for cell cycle regulation by determining cell typespecific cell cycle modes. 25,26 Usually, inhibition of the cell cycle is a requisite for terminal differentiation. 23,25,27,28 However, the precise cell cycle mechanisms for growth/differentiation transition remain unclear. In this study, we found that there exists a cell cycle exit that is mediated by the accumulation of CKIs p21 Cip1 and p27 Kip1 during bortezomib-induced osteogenic differentiation of MSCs and Bortezomib is a proteasome inhibitor of the 26S proteasome that plays a central role in protein degradation. The introduction of bortezomib has been a major breakthrough in the treatment of MM. 29 Besides the anti-MM activity, both preclinical and clinical data also substantiate that bortezomib plays a significantly beneficial role in bone formation. 30 The increased osteoblast differentiation in BM hypothesizes one possible mechanism behind bone protection. 17,[31][32][33][34][35] In the current study, by using mBM-MSCs as an in vitro model, we demonstrated that bortezomib can induce osteogenic differentiation, validated by the markedly enhanced ARS staining. Our findings in mBM-MSCs confirmed the previous report in human MSCs. 13,36 Meanwhile, EdU incorporation assay demonstrated that cell proliferation was almost entirely blocked by bortezomib. Cell cycle analysis further indicated that a G 0 /G 1 phase arrest was induced by bortezomib in mBM-MSCs. These findings strongly indicate a link between G 0 /G 1 phase arrest and bortezomib-induced differentiation.
Cell cycle progression is tightly governed by CDKs, which are the major regulators of the cell division cycle, activated by cyclin binding and inhibited by CKIs. 36,37 The close cooperation between this trio is necessary for ensuring orderly progression through or exit from the cell cycle. For this reason, we further studied the changes of cyclins, CDKs and CKIs in response to bortezomib treatment and found that the expression of G 0 /G 1 phase-related CDKs such as Cdk2 and Cdk4 was decreased by bortezomib. More importantly, the expression of p21 Cip1 and p27 Kip1 was observed to be increased significantly by bortezomib. Considering that p21 Cip1 and p27 Kip1 were extensively characterized as negative regulators of progression through G 1 to S phase in mammalian cells, and several lines of evidence have suggested that p21 Cip1 and p27 Kip1 exert similar effects on cell cycle progression by mediating the inhibition of Cdk2 and/or Cdk4 activities, 38,39 it is reasonable to speculate that the induction of p21 Cip1 and p27 Kip1 may play an important role in the cell cycle exit induced by bortezomib.
It is known that p21 Cip1 and p27 Kip1 can inhibit cell cycle progression in response to numerous stimuli, but little is known IRE1α-XBP1, PERK-Atf4 and Atf6. 44 Focusing on the mechanisms of inducing p21 Cip1 and p27 Kip1 , we further investigated whether the up-regulation of p21 Cip1 and p27 Kip1 is related to the ER stress signalling activated by bortezomib. Firstly, we found that the up-regulation of p21 Cip1 and p27 Kip1 occurred at the mRNA level.
Next, we confirmed that bortezomib can activate both PERK-Atf4 and IRE1α-Xbp1s signalling pathways in mBM-MSCs. Thirdly, we confirmed that Xbp1s other than Atf4 plays a major role in regulating p21 Cip1 and p27 Kip1 expression. More importantly, by perform- ing ChIP assay, we demonstrated the direct interaction between Xbp1s and the promoter of p21 Cip1 and p27 Kip1 , further supporting the role of Xbp1s in transactivating the transcriptional activity of the p21 Cip1 and p27 Kip1 .
Xbp1 is a bZIP (basic-region leucine zipper) transcription factor that interacts specifically with the conserved X2 boxes of major histocompatibility complex class II gene promoters. 45 Xbp1 can yield two isoforms: unspliced Xbp1 (Xbp1u) and spliced Xbp1 (Xbp1s). In response to ER stress, the mRNA of Xbp1u is spliced to generate Xbp1s, which is considered as the active form, playing a pivotal role in ER stress signalling. Nonetheless, Xbp1u has also been shown F I G U R E 6 Xbp1s binds to the promoter of p21 Cip1 and p27 Kip1 . (A) Graphic representation of the putative Xbp1s binding sites in p21 Cip1 and p27 Kip1 promoter. Two putative Xbp1s binding sites were identified in the promoter of the p21 Cip1 and p27 Kip1 by searching for the Eukaryotic Promoter Database. (B-C) Chromatin immunoprecipitation followed by real-time PCR assay of Xbp1s binding in the p21 Cip1 and p27 Kip1 promoters in response to 0 and 2.5 nM bortezomib treatment for 16 h. Results are expressed as percentage of input. *P < 0.05 compared with control (n = 3) F I G U R E 7 Diagrammatic presentation of the potential mechanism of cell cycle exit during bortezomib-induced osteogenic differentiation of mBM-MSCs to inhibit Xbp1s-mediated effects. For example, Xbp1u has been demonstrated to down-regulate the expression of p21 Cip1 by negatively inhibiting the p53/p21 axis. 46 Moreover, it has been indicated that Xbp1s is essential for bone morphogenic protein 2-induced osteoblast differentiation through up-regulating the transcription of Osterix, which is an osteoblast-specific transcription factor. 47 We further demonstrated that Xbp1s plays central roles in regulating several osteogenic differentiation-related genes in response to bortezomib stimuli (data not shown). Focusing on the effect of Xbp1s in the cell cycle, we further showed that forced expression of XBP1s in mBM-MSCs can directly trigger the accumulation of p21 Cip1 and p27 Kip1 . Meanwhile, we observed that forced expression of Xbp1s can drive mBM-MSC differentiation into osteoblasts (data not shown).
In addition to the well-known function of CKIs in cell cycle control, it is becoming increasingly apparent that CKIs also play indispensable roles in processes such as transcription and epigenetic regulation. Both p21 Cip1 and p27 Kip1 are known to interact with a range of transcription factors involved in modulating the expression of numerous genes in various biological processes. 48 In this regard, one limitation of this study is that we cannot conclude whether the up-regulated p21 Cip1 and p27 Kip1 directly stimulate the expression of osteogenic-related genes. Secondly, we cannot conclude whether p21 Cip1 and p27 Kip1 play redundant roles in this process. For example, although both p21 Cip1 and p27 Kip1 proteins were induced during erythroid differentiation, only p27 Kip1 is associated with the inactivation of Cdk2, and p21 Cip1 may have a function independent of growth arrest during erythroid differentiation. 49 In myeloid leukaemia cells, p21 Cip1 and p27 Kip1 have been demonstrated to induce distinct cell cycle effects and differentiation programmes. 39

| CON CLUS IONS
In this study, we demonstrated that bortezomib-induced p21 Cip1 and p27 Kip1 are required for cell cycle exit during osteogenic differentiation and that induction of p21 Cip1 and p27 Kip1 by bortezomib is transcriptionally regulated by activation of the ER stress signalling pathway Ire1α/Xbp1s. These findings may provide valuable information enabling a better understanding of the mechanisms underlying proteasome inhibitor-induced osteogenic differentiation of MSCs.

ACK N OWLED G EM ENTS
This research was supported by the National Natural Science

CO N FLI C T O F I NTE R E S T
The authors report no conflict of interest.

AUTH O R CO NTR I B UTI O N S
JH and DZ designed the experiments, analysed and interpreted the experimental results and wrote the manuscript. RF, LL, LL, YM and NL performed most of the experiments and analysed the experimental data. PC and RAW carried out Western blotting and real-time PCR analysis. BW made substantial contributions to the conception and design of the study and revised the manuscript. All authors read and approved the manuscript.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data used to support the findings of this study are available from the corresponding author upon request.