1, 25-dihydroxy-vitamin D3 with tumor necrosis factor-alpha protects against rheumatoid arthritis by promoting p53 acetylation-mediated apoptosis via Sirt1 in synoviocytes

Impaired apoptosis of fibroblast-like synoviocytes (FLSs) causes synovial hyperplasia, facilitating destruction of cartilage and bone in rheumatoid arthritis (RA). Tumor necrosis factor (TNF)-α, a dominant inflammatory mediator in RA pathogenesis, promotes progression of RA symptoms. Prevalence of 1, 25-dihydroxy-vitamin D3 (hereafter termed VD) deficiency is 30–63% in patients with RA. Whether VD leads to apoptosis or enhances TNF-α-mediated apoptosis in FLSs to ameliorate RA is unclear. To determine this, 10-week-old CYP27B1-deficient (CYP27B1−/−) mice with collagen-induced arthritis (CIA) were intraperitoneally treated with 1 μg/kg VD every other day for 9 weeks. RA phenotypes were compared between vehicle-treated CYP27B1−/− and wild-type CIA mice. Human rheumatoid FLS-MH7A cells were treated with Dulbecco's modified Eagle's medium (DMEM) without fetal bovine serum (FBS) for 24 h, then with different concentrations of VD and TNF-α, human vitamin D receptor (VDR) siRNA or the p53 pro-apoptotic inhibitor pifithrin-α. Apoptosis and p53 pro-apoptotic signaling were analyzed. The 19-week-old vehicle-treated CYP27B1−/− CIA mice had increased cumulative arthritis scores and levels of serous rheumatoid factors and C-reactive protein. They had exacerbated articular cartilage and bone destruction, joint space narrowing, joint stiffness, deformity and dysfunction, synovitis and TNF-α secretion, FLS hyperplasia with increased proliferation and decreased apoptosis compared to CIA mice. These RA phenotypes that were aggravated in CIA mice by CYP27B1 deficiency were largely rescued by VD treatment. In vitro, VD with TNF-α treatment upregulated p53 acetylation-mediated apoptosis in MH7A cells by promoting Sirt1 translocation from the nucleus to the cytoplasm. These findings indicated that VD with TNF-α protected against RA by promoting apoptosis of FLSs. The results indicated that clinical administration of VD could be a specific therapy to promote FLS apoptosis and prevent RA progression.

autoimmune diseases. [13][14][15] Emerging clinical epidemiology evidence indicates that the prevalence of VD deficiency is 30-63% in people with RA, serum VD levels negatively correlate with RA activity, and VD supplements are a beneficial RA treatment. 2,14 Data show that VD upregulates TNF-α-mediated pro-apoptotic effects in prostate tumor cells and osteoblasts and down-regulates pro-inflammatory effects in RA macrophages and T cells. 12,16 However, whether VD leads to apoptosis or enhances TNF-α-mediated apoptosis in FLSs to ameliorate RA is unclear. The levels of C-RP (mg L ) -1 The pathological features of collagen-induced arthritis (CIA) in mice are consistent with typical pathological alterations in RA patients. Therefore, CIA is the most widely studied RA model. 9 We induced CIA in CYP27B1-deficient (CYP27B1 −/− ) mice, and intraperitoneally injected them with 1 μg/kg VD every other day. Their RA phenotypes were compared with vehicle-treated CYP27B1 −/− and wild-type (WT) CIA mice. We treated human rheumatoid FLS-MH7A cells with different concentrations of VD, TNF-α and human VDR siRNA and the p53 pro-apoptotic inhibitor pifithrin-α. Apoptosis and p53 pro-apoptotic signaling were analyzed.

Results
Levels of serum VD in VD-treated CYP27B1 −/− CIA mice restored to normal. Radioimmunoassay showed that serum VD levels in 10-week-old CYP27B1 −/− mice or 19-week-old vehicle-treated CYP27B1 −/− CIA mice were significantly decreased compared to those in 10-week-old WT or 19-week-old WT CIA mice, respectively. We confirmed that serum VD levels of 19-week-old VD-treated CYP27B1 −/− CIA mice were restored to normal (Supplementary Figure S1A).
To compare the serum VD levels between VD-treated WT mice and vehicle-treated WT mice, 10-week-old male WT mice were intraperitoneally injected with 1 μg/kg VD or vehicle every other day for 3 weeks. The levels of serum VD did not differ between 13-week-old VD-treated and vehicle-treated WT mice (Supplementary Figure S1A). It demonstrated that VD supplement did not increase serum level of VD in healthy individuals with sufficient VD.
Clinical indexes of rheumatoid arthritis were ameliorated in CYP27B1 −/− CIA mice by VD. To assess if VD treatment ameliorated clinical indexes of RA in CYP27B1 −/− CIA mice, arthritis severity, swelling degree, rheumatoid serum biochemical measurements, and bone and joint erosion of arthritic hind palms were subjectively evaluated in 19-weekold, VD-treated CYP27B1 −/− CIA, vehicle-treated CYP27B1 −/− and WT CIA mice. Compared with WT CIA mice, body size was decreased; however, cumulative arthritis scores from 1 to 9 weeks after primary immunization, paw thickness, levels of rheumatoid factors (RFs) and C-reactive protein (C-RP) from serum increased significantly in vehicle-treated CYP27B1 −/− CIA mice (Figures 1a-f). X-ray and threedimensional (3D) reconstructed graphs of hind palms showed that metatarsophalangeal joints were characterized by blurred articular surface, cartilage erosion and narrowed articular space on radiographs; intertarsal joints showed deformity and phalanxes were eroded on 3D-reconstructed graphs (Figures 1b and g). Compared with vehicle-treated CYP27B1 −/− CIA mice, body size was increased; however, cumulative arthritis scores from 1 to 9 weeks after primary immunization, paw thickness, and RFs and C-RP serum levels decreased significantly in VD-treated CYP27B1 −/− mice (Figures 1a-f). The VD treatment ameliorated the arthritic characteristics of metatarsophalangeal joints and intertarsal joints, and decreased bony erosion of phalanxes in VD-treated CYP27B1 −/− CIA mice (Figures 1b and g).
Histological bone destruction, cartilage erosion and synovial inflammation of rheumatoid arthritis were ameliorated in CYP27B1 −/− CIA mice by VD. To investigate if VD treatment ameliorated histological bone destruction, cartilage erosion and synovial inflammation of RA in CYP27B1 −/− CIA mice, HE, total collagen, and Safranin O and fast green double dyeing (SO-FG) histochemical staining were performed to analyze trabecular and cartilage volume. TNF-α, IL-1β, CD3 and F4/80 immunohistochemical staining for synovial inflammation were performed. Compared with WT CIA mice, trabecular and cartilage bone volume decreased significantly; however, percentages of TNF-α, IL-1β, CD3 and F4/80 positive inflammatory cells in FLSs increased in vehicle-treated CYP27B1 −/− CIA mice ( Figure 2). Compared with vehicle-treated CYP27B1 −/− CIA mice, trabecular and cartilage bone volume increased significantly; however, percentages of TNF-α, IL-1β, CD3 and F4/80-positive inflammatory cells in synovial membranes decreased in VD-treated CYP27B1 −/− CIA mice ( Figure 2). Synovial hyperplasia caused by increased FLS proliferation and decreased apoptosis were ameliorated in CYP27B1 −/− CIA mice by VD. To determine if VD treatment ameliorated synovial hyperplasia in CYP27B1 −/− CIA mice, HE staining was used for comparisons with WT CIA mice. The number of synovial cells increased in vehicletreated CYP27B1 −/− mice. Compared with vehicle-treated CYP27B1 −/− CIA mice, the number of synovial cells decreased in VD-treated CYP27B1 −/− CIA mice (Figures 3a  and e).
To further determine if VD treatment decreased proliferation and increased apoptosis of FLSs in CYP27B1 −/− CIA mice, proliferating cell nuclear antigen (PCNA) immunohistochemistry staining was used to detect proliferation, and Caspase3 staining and TUNEL assays were used to analyze apoptosis. Compared to WT CIA mice, the percentage of PCNA-positive FLSs decreased and percentages of Caspase3-positive and TUNEL-positive FLSs significantly increased in CYP27B1 −/− CIA mice; the percentage of PCNA-positive FLSs increased with a significant increase in percentages of Representative radiographs of (Ba) underside and (Bb) side of arthritic hind palms. Blue arrowhead: interphalangeal joints of arthritic hind palms; red arrow: hind palms. (c) Cumulative arthritis scores for subjective evaluation of arthritis severity of arthritic hind palms from 1 to 9 weeks after primary immunization. (d) Paw thickness relative to WT mice. (e) Levels of rheumatoid factors (RFs) and (f) C-reactive protein (C-RP) in serum determined by spectrophotometry. Values are mean ± S.E.M. of six determinations per group. *Po0.05; **Po0.01; ***Po0.001 compared with WT group at the same age. # Po0.05; ## Po0.01; ### Po0.001 compared with KO group at the same age. (g) Representative 3D-reconstructed images of (Ga) intertarsal joint and (Gb) interphalangeal joints of arthritic hind palms. Blue arrow: deformity of intertarsal joints; red arrow: bone erosion of phalanxes  Apoptosis of human rheumatoid FLSs increased by VD with TNF-α. To investigate if VD with TNF-α promoted apoptosis of rheumatoid FLSs, human rheumatoid FLS-MH7A cells were treated with different concentrations of VD and/or TNF-α. Apoptosis was detected by AV and PI double staining and flow cytometry. Results are shown in detail in Table 1. These results demonstrated that with TNF-α treatment at the corresponding concentration, VD supplementation significantly increased the apoptosis of rheumatoid FLSs. Moreover, the pro-apoptotic effect of VD was increased with elevated concentrations of TNF-α (Figures 4a and b).
To detect further expression of pro-apoptotic and antiapoptotic genes, real-time RT-PCR were performed for Bcl-2 binding component 3 (also known as p53 upregulated modulator of apoptosis; PUMA), Bcl-2-associated X (Bax) protein and Bcl-2 (Table 1). These results demonstrated that with TNF-α treatment at the corresponding concentration, VD supplementation significantly increased expression of proapoptotic genes and decreased expression of anti-apoptotic genes in rheumatoid FLSs. Moreover, under VD treatment at the corresponding concentration, expression of pro-apoptotic genes was increased with TNF-α concentration. Expression of anti-apoptotic genes was decreased with increased TNF-α concentration (Figures 4c-e).
Human rheumatoid FLS apoptosis after VD with TNF-α was mediated by VDR and p53 pro-apoptotic signaling.
To further investigate if apoptosis of rheumatoid FLSs induced by VD with TNF-α treatment was mediated by VDR and p53 pro-apoptotic signaling, human rheumatoid FLS-MH7A cells were knocked down with VDR siRNA.
Compared to negative control (NC) siRNA, VDR gene expression was downregulated to 17.87% in cells with VDR siRNA1, 52.52% in cells with VDR siRNA2 and 30.24% in cells with siRNA3 (Supplementary Figure S1C). β-actin gene expression was downregulated to 36.4% in cells with β-actin siRNA, but was not altered in cells transfected with VDR  Cells were treated with VD and/or TNF-α and p53 proapoptotic inhibitor PFT-α, and apoptosis detected by AV and PI double staining by flow cytometry. Detailed results are shown in Table 2. The pro-apoptotic effect induced by VD was Table 1 VD with TNF-α promoted apoptosis of rheumatoid FLSs 1, 25-dihydroxy-vitamin D 3 promotes apoptosis of synoviocytes X Gu et al attenuated when VDR was knocked down or p53 was inhibited. Moreover, attenuation of the pro-apoptotic effect was more obvious in the p53-inhibited group than in VDRknocked down group (Figures 5a and b).
To detect further expression of pro-apoptotic and antiapoptotic genes, real-time RT-PCR was performed for PUMA, Bax and Bcl-2 (Table 2). When VDR was knocked down or p53 was inhibited, expression of the upregulated pro-apoptotic genes induced by VD was decreased, and expression of the downregulated anti-apoptotic gene induced by VD was increased. Moreover, the above changes were more obvious in the p53-inhibited group than in VDR-knocked down group (Figures 5c-e).
Sirt1 nuclei to cytoplasm translocation is promoted in human rheumatoid FLSs by VD with TNF-α. To determine if VD with TNF-α treatment promoted Sirt1 translocation from nuclei to cytoplasm in rheumatoid FLSs, Sirt1 immunofluorescence was performed with sections of paraffin-embedded interphalangeal joints from CIA mice or MH7A cells. Sirt1 expression in nuclear or cytoplasm fractions of MH7A cells was detected by western blots.
Compared with the serum-free control group, the percentage of Sirt1-positive nuclei relative to DAPI nuclei and Sirt1 protein in the nuclear fraction of MH7A cells increased in the 30 ng/ml TNF-α treatment group; both parameters significantly decreased with 10 -7 M VD, or 10 -7 M VD with 30 ng/ml TNF-α treatment. Compared with the 30 ng/ml TNF-α treatment, the percentage of Sirt1 positive nuclei relative to DAPI nuclei and Sirt1 protein in the nuclear fraction of MH7A cells significantly decreased with 10 -7 M VD, or 10 -7 M VD with 30 ng/ml TNF-α treatments. Compared with the 10 -7 M VD-treatment group, the percentage of Sirt1-positive nuclei relative to DAPI nuclei and Sirt1 protein in the nuclear fraction of MH7A cells decreased with 10 -7 M VD with 30 ng/ml TNF-α treatment (Figures 7c-f).
Compared with the serum-free control group, Sirt1 protein in the cytoplasmic fraction of MH7A cells decreased in the 30 ng/ ml TNF-α-treatment group; it significantly increased with 10 -7 M VD, or 10 -7 M VD with 30 ng/ml TNF-α treatment. Compared with the group treated with 30 ng/ml TNF-α, Sirt1 protein in the cytoplasmic fraction of MH7A cells significantly increased with groups treated with 10 -7 M VD, and 10 -7 M VD with 30 ng/ml TNF-α. Compared with the 10 -7 M VD treatment group, Sirt1 protein in the cytoplasmic fraction of MH7A cells increased in the group of 10 -7 M VD and 30 ng/ml TNF-α treatments (Figures 7c-f).
These results demonstrated that VD with TNF-α promoted Sirt1 translocation from the nucleus to the cytoplasm in rheumatoid FLSs more than VD alone, and reversed the inhibitory effect of TNF-α on Sirt1 translocation from the nucleus to the cytoplasm.
Apoptosis of normal FLSs were not induced by VD with TNF-α. To determine whether apoptosis of normal FLSs was induced by VD with TNF-α, normal FLSs were treated with the same treatment as MH7A cells. The percentages of AV-positive and PI-negative (AV+PI − ), AV and PI doublepositive (AV+PI+) and total AV-positive (AV+) normal FLSs in the serum-free control group were 34.87 ± 1.62%, 6.60 ± 1.48% and 41.47 ± 1.69%, respectively (Supplementary Figures S2A and S2B). However, the percentages of AV +PI − , AV+PI+ and total AV+FLSs in the serum-free control group of MH7A cells were 5.99 ± 0.65%, 3.66 ± 0.47% and 9.65 ± 0.73% respectively (Figures 4a and b). These results demonstrated that the resistance of normal FLSs to serumfree culture is poorer than MH7A cells, and easier to be induced into apoptosis. Previous studies support that normal FLSs appear less resistant to apoptosis than RA-FLSs. 4 Thus, normal FLSs were treated with DMEM or DMEM and 10 -7 M VD and/or 10 or 30 ng/ml TNF-α for 24 h. The percentages of AV+PI − , AV+PI+ and total AV+FLSs in the serum-free control group of normal FLSs were 5.0 ± 0.75%, 6.78 ± 0.71% and 11.78 ± 0.25%, respectively (Supplementary Figures S2C and S2D). Compared with the serum-free control group, the percentages of AV+PI − , AV+PI+ or total AV+FLSs in the 10 -7 M VD group, 10 ng/ml TNF-α group, 10 ng/ml TNF-α with 10 -7 M VD group, 30 ng/ml TNF-α group, and 30 ng/ml TNF-α with 10 -7 M VD group were no different ( Supplementary Figures S2C and S2D).

Discussion
In this study, we demonstrated that CYP27B1 deficiency increased cumulative arthritis scores and levels of serous RFs and C-RP. CYP27B1 deficiency exacerbated articular cartilage and bone destruction, joint space narrowing, joint stiffness, deformity and dysfunction, synovitis and TNF-α 1, 25-dihydroxy-vitamin D 3 promotes apoptosis of synoviocytes X Gu et al secretion. CYP27B1 deficiency caused FLSs hyperplasia with increased proliferation and decreased apoptosis in CIA mice. These typical RA phenotypes in CYP27B1 deficiency were largely rescued by VD treatment. In vitro, results demonstrated that VD with TNF-α treatment upregulated p53 acetylationmediated apoptosis in human rheumatoid FLS-MH7A cells by promoting Sirt1 translocation from nuclei to cytoplasm. These findings indicated that VD with TNF-α protected against RA by 10  Previous studies suggest that FLS hyperplasia is a key to both initiation and the perpetuation of RA. FLS hyperplasia has been linked most prominently to the progressive destruction of articular structures, particularly cartilage. 7,17 The ability of hyperplastic FLSs to erode articular cartilage is a multistep process that includes attachment to cartilage and synthesis of proteolytic enzymes that degrade the extracellular matrix. 4 A broad array of inflammatory mediators and cytokines are secreted by hyperplastic FLSs to induce apoptosis of articular cartilage. 4 Hyperplastic FLS results in joint stiffness, deformity and dysfunction. 4 Therefore, inhibiting inflammation and proliferation and promoting apoptosis of FLS is critical for RA treatment.
Previous observations demonstrate that TNF-α induces proinflammatory responses and upregulates TNF-α and IL-1β through activating the NF-κB signaling pathway that stimulates FLS overgrowth to form a mass of synovial tissue. 11,16,18 Our results revealed that CYP27B1 deficiency upregulated expression of TNF-α and IL-1β, increased CD3-positive and F4/80-positive inflammatory cells and PCNA-positive FLSs. We found that VD downregulated expression of TNF-α and IL-1β, and decreased CD3-positive and F4/80-positive inflammatory cells and PCNA-positive FLSs in CYP27B1-deficient CIA mice. Results from a previous study support that NF-κBmediated pro-inflammatory gene transcription is repressed by VD, resulting in suppressed production of inflammatory cytokines such as IL-1β and TNF-α. 13 Our study further demonstrated that VD and TNF-α treatment significantly inhibited NF-κB-p65 expression in human rheumatoid FLSs. Thus, VD could be used for RA treatment as an antiinflammatory and anti-proliferative molecule for hyperplastic FLSs.
Several lines of evidence suggest that impaired and inhibited apoptosis contributes to the accumulation of RA-FLSs, and is pivotal to RA-FLS-mediated disease progression. 17 Therefore, finding pro-apoptotic molecules of hyperplastic FLSs is important for RA treatment. Our results revealed that CYP27B1 deficiency decreased Caspase3positive and TUNEL-positive FLSs. We also found that VD increased the percentages of Caspase3-positive and TUNELpositive FLSs in CYP27B1-deficient CIA mice compared to WT CIA mice. The percentage of TNF-α-positive cells in synovial membranes increased in VD-treated CYP27B1deficient CIA mice compared to WT CIA mice. Consistent with these results, VD might cooperate with TNF-α to promote apoptosis of RA-FLSs in vivo. Consistently, VD with TNF-α treatment significantly increased AV+PI-, AV+PI+ and AV +cells and Caspase3 protein in human rheumatoid FLSs in vitro. Thus, VD could also be a pro-apoptotic molecule in hyperplastic FLSs, in cooperation with TNF-α, to promote apoptosis for RA treatment.
As a critical tumor suppressor gene and apoptotic regulator implicated in RA, p53 induces cell apoptosis through posttranslational modification to upregulate downstream proapoptotic genes Bax, PUMA and Noxa. P53 also combines with the anti-apoptotic gene Bcl-2 to inhibit anti-apoptotic effects. 4,19 Loss or inactivation of p53 caused by functional mutations has been universally demonstrated in RA-FLSs, most extensively in patients with advanced, destructive disease. 4,20,21 P53 mutations increase apoptosis resistance and the level of inflammatory secretion in RA-FLSs. 4,21 We found that VD with TNF-α treatment increased Bax and PUMA and decreased Bcl-2 mRNA and protein in human rheumatoid FLSs. Moreover, PFT-α, a p53 pro-apoptotic inhibitor, decreased apoptosis of human rheumatoid FLSs induced by VD and TNF-α. Our results demonstrated that VD with TNF-α promoted apoptosis of human rheumatoid FLSs through upregulating p53 pro-apoptotic signaling and ameliorated the development of RA. Recent evidence suggests that PUMA induces apoptosis not only through a p53-dependent mechanism, but also through p53-independent signaling. 4,22 Consistently, our results suggested that PUMA was still expressed, even if PFT-α was supplied to inhibit p53 apoptotic signaling in human rheumatoid FLSs. Therefore, apoptosis of RA-FLSs induced by VD and TNF-α supplement might be partly mediated by p53-independent pro-apoptotic signaling; this hypothesis remains to be investigated.
P53 acetylation is a post-translational modification that enhances the ability of p53 to regulate downstream genes. 23 Equilibrium of p53 acetylation is partly maintained by Sirt1. Sirt1 preferentially de-acetylates p53 at K382 to have a profound negative impact on the capacity of p53 to induce expression of target genes involved in apoptosis such as PUMA and Bax. 19 The p53 de-acetylation caused by Sirt1 inhibits p53-dependent pro-apoptosis by increasing ubiquitination-mediated degradation of p53. 4,24 Evidence suggests that Sirt1 is highly expressed in RA-FLSs, which is positively related to RA severity. 24,25 Our study found that VD with TNF-α inhibited p53 de-acetylation, increased p53 acetylation and promoted p53-dependent pro-apoptosis.
NF-κB is constitutively activated in RA. 26 The antiapoptogenic function of NF-κB is well recognized, and is achieved through the gene expression of several antiapoptogenic molecules, including Bcl-xL, XIAP, cIAP1, cIAP2 and TRAFs induced by NF-κB. 6 Previous studies have reported that NF-κB-p65 phosphorylation at Ser-536 contributes to the inhibition of p53 transcriptional activity and subsequent anti-apoptotic effects. 27,28 Our results demonstrated that VD with TNF-α treatment significantly   1, 25-dihydroxy-vitamin D 3 promotes apoptosis of synoviocytes X Gu et al promoted apoptotic resistance and inflammatory secretion in RA-FLSs; 25 Sirt1 gene silencing promotes apoptosis of RA-FLSs. 29 In contrast, evidence suggests that Sirt1 is important in anti-inflammation through downregulating expression of the cysteine-rich protein 61 in RA-FLSs; 30 resveratrol, a Sirt1 activator, induces apoptosis of RA-FLSs. 31 Recent studies report that subcellular localization of Sirt1 is associated with cell fate. Nuclear-localized Sirt1 is involved in anti-apoptosis, whereas cytoplasmically localized Sirt1 promote apoptosis that is independent of de-acetylation but dependent on Caspase signaling. 32,33 Moreover, overexpression and nuclear-localization of Sirt1 significantly promotes apoptotic resistance in RA-FLSs. 25 Thus, the subcellular localization of Sirt1 in FLSs may be pivotal to RA progression. Our results demonstrated that TNF-α treatment inhibited translocation of Sirt1 from nuclei to cytoplasm; VD with TNF-α treatments promoted Sirt1 translocation from nuclei to cytoplasm in human rheumatoid FLSs. However, further investigation is needed to determine if cytoplasm-localized Sirt1 is necessarily associated with pro-apoptosis in RA-FLSs and the exact regulating mechanism.
In conclusion, the 'in vivo' and 'in vitro' data presented in this paper strongly suggest that VD with TNF-α protected against RA by promoting apoptosis of RA-FLSs. Our findings implied that clinical administration of VD could be as a specific therapy to prevent RA progression.

Materials and Methods
Mice and genotyping. Adult CYP27B1 heterozygote (CYP27B1 +/ − ) mice (129sv/J hybrid background) were backcrossed 10-12 times to the BALB/c background and mated to generate CYP27B1 homozygotes (CYP27B1 −/− ) and their wild-type (WT) littermates. Mice were genotyped by PCR as described previously. 34,35 Male CYP27B1 −/− and WT mice were used at 10 weeks of age. Mice were weaned at 3 weeks and fed rescue diet (TD96348 Teklad, Madison, WI, USA) containing 2% calcium, 1.25% phosphorus and 20% lactose until they were 19 weeks old. We confirmed that serum calcium and phosphorus levels were normalized in CYP27B1 −/− mice and WT littermates fed with rescue diet (Supplementary Table S1).
This study was carried out in strict accordance with the guidelines of the Institute for Laboratory Animal Research of Nanjing Medical University in Nanjing, China. The protocol was approved by the Committee on the Ethics of Animal Experiments of Nanjing Medical University (Approval ID 20111201).
Administration of VD in vivo. The 10-week-old mice were divided into three groups of VD-treated CYP27B1 −/− CIA, vehicle-treated CYP27B1 −/− CIA or WT CIA mice. For 9 weeks, animals received intraperitoneal treatment of vehicle or 1 μg/kg VD (Sigma-Aldrich, St. Louis, MO, USA) every other day.
The 10-week-old male WT mice derived from BALB/c background were divided into two groups of VD-treated and vehicle-treated mice. For 3 weeks, animals received intraperitoneal treatment of vehicle or 1 μg/kg VD every other day.
Measurements of serum calcium, phosphorus and VD. Mice were anesthetized with 3% pentobarbital sodium (40 mg/kg) and the abdomen depilated. Blood was taken by suction from the heart with a 1 ml syringe.
Induction and assessment of collagen-induced arthritis. CIA was induced as previously described. 9,14,36 Immunization-grade chicken type II collagen (Chondrex Inc., Redmond, WA, USA) was dissolved in 0.1 mM acetic acid (2 mg/ml) at 4°C overnight and emulsified with an equal volume of complete Freund's adjuvant ((Sigma-Aldrich) on ice. Male CYP27B1 −/− and WT littermates at 10 weeks of age were injected subcutaneously at the tail base with 100 μl emulsion, which was considered the primary immunization on Day 1. Secondary immunization on Day 22 was similar to primary immunization on Day 22. Mice were analyzed by three independent, blinded examiners every sixth day from Day 7 and monitored for signs of arthritis onset using clinical parameters of cumulative arthritis scores for subjective evaluation of arthritis severity. Clinical arthritis was scored as previously described: 9 0, no evidence of erythema and swelling; 1, erythema and mild swelling confined to the tarsals or ankle joint; 2, erythema and mild swelling extending from the ankle to the tarsals; 3, erythema and moderate swelling extending from the ankle to metatarsal joints; and 4, erythema and severe swelling encompassing the ankle, foot and digits, or ankylosis of the limb. Paw swelling was assessed by measuring the thickness of affected hind paws with 0-10 mm calipers. 36 Rheumatoid serum biochemical measurements. At 19 weeks, mice were anesthetized with 3% pentobarbital sodium (40 mg/kg) and the abdomen depilated. Blood was taken by suction from the heart with a 1 ml syringe. Serum was isolated for measurements of RFs (E035 RF detection kit) and C-RP (E023 C-RP detection kit) according to the manufacturers' instructions (Nanjing Jiancheng Bioengineering Institute, Nanjing, China).
For histochemistry or immunohistochemistry, samples were decalcified and dehydrated in a series of graded ethanol solutions, embedded in paraffin and cut into 5-μm sections with rotary microtome (Leica Microsystems Nussloch GmbH, Nubloch, Germany). 38 Skeletal radiography and micro-computed tomography. Hind palms were removed and dissected free of soft tissue. Contact radiographs were