The Journal of Immunology

Modulatory Effects of 1,25-Dihydroxyvitamin D3 on Human Differentiation

Sheng Chen,*† Gary P. Sims,1* Xiao Xiang Chen,† Yue Ying Gu,† Shunle Chen,† and Peter E. Lipsky2*

1,25-Dihydroxyvitamin D3 (1,25(OH)2D3) can modulate immune responses, but whether it directly affects B cell function is unknown. Patients with systemic lupus erythematosus, especially those with antinuclear Abs and increased disease activity, had decreased 1,25(OH)2D3 levels, suggesting that vitamin D might play a role in regulating autoantibody production. To address this, we examined the effects of 1,25(OH)2D3 on B cell responses and found that it inhibited the ongoing proliferation of activated B cells and induced their apoptosis, whereas initial cell division was unimpeded. The generation of plasma cells and postswitch memory B cells was significantly inhibited by 1,25(OH)2D3, although the up-regulation of genetic programs involved in B cell differentiation was only modestly affected. B cells expressed mRNAs for involved in vitamin D ␣ activity, including 1 -hydroxylase, 24-hydroxylase, and the vitamin D , each of which was regulated by 1,25(OH)2D3 and/or activation. Importantly, 1,25(OH)2D3 up-regulated the expression of p27, but not of p18 and p21, which may be important in regulating the proliferation of activated B cells and their subsequent differentiation. These results indicate that

1,25(OH)2D3 may play an important role in the maintenance of B cell homeostasis and that the correction of vitamin D deficiency may be useful in the treatment of B cell-mediated autoimmune disorders. The Journal of Immunology, 2007, 179: 1634–1647.

he biologically active form of vitamin D3, 1,25-dihy- dendritic cells (3, 4), but their contribution to systemic vitamin D 3 droxyvitamin D3 (1,25(OH)2D3), which was originally metabolism is unknown. 1,25(OH)2D3 also induces the expression described as an essential hormone for bone and mineral of the major 1,25(OH) D -degrading enzyme 24-hydroxylase T 2 3 homeostasis, has also been shown to have immunomodulatory ef- (CYP24A1) (2). Most of the known biological effects of fects (1). Vitamin D3 can be derived from dietary sources, but the 1,25(OH)2D3 are mediated through the vitamin D3 receptor main source of vitamin D3 is obtained through photosynthesis in (VDR), a member of the superfamily of nuclear hormone receptors the skin. Exposure of skin to UV light (270–300 nm) catalyzes the (5) that is expressed by different cells of the immune system and first step in vitamin D3 biosynthesis, converting 7-dehydroxycho- elsewhere (5, 6). lesterol into previtamin D , which is followed by a spontaneous 3 The immunoregulatory effects of 1,25(OH)2D3 are mainly and temperature-dependent isomerization into vitamin D3.Toob- thought to be mediated through its action on APCs, with the most tain the biologically active metabolite 1,25(OH)2D3, vitamin D3 potent reported effects on dendritic cells (DCs) (7, 8). The in vitro must first be hydroxylated by D3-25-hydroxylase (CYP2D25) differentiation of DCs from monocytes or murine bone marrow- in the liver into 25-hydroxyvitamin D3 (25(OH)D3). Further derived precursors is inhibited by 1,25(OH) D . Moreover, the ␣ 2 3 hydroxylation of 25(OH)D3 by 25(OH)D3-1- -hydroxylase Ag-presenting function of monocytes and DCs is profoundly in- (CYP27B1) occurs mainly in the proximal convoluted tubule cells hibited, as is the surface expression of costimulatory molecules of the kidney, resulting in the biologically active 1,25(OH)2D3 (2). and also IL-12 production (9). In addition, expression of the im- In addition, CYP27B1 is expressed by activated macrophages and munosuppressive cytokine IL-10 by DCs, which opposes the Th1-

inducing effects of IL-12, is increased by 1,25(OH)2D3 (7, 8).

Direct effects of 1,25(OH)2D3 on T lymphocytes have also been *Autoimmunity Branch, National Institute of Arthritis and Musculoskeletal and Skin reported. T cell proliferation and cell cycle progression from G Diseases, National Institutes of Health, Bethesda, MD 20892; and †Shanghai Clinical 1a Center for Rheumatology, Renji Hospital, Shanghai Jiaotong University School of to G1b are inhibited in vitro by 1,25(OH)2D3. Cytokine production Medicine, Shanghai, China by T cells is also modulated, with inhibition of the Th1 cytokine Received for publication April 24, 2007. Accepted for publication May 21, 2007. IFN-␥ and an increase of the Th2 cytokines IL-4, IL-5, and IL-10

The costs of publication of this article were defrayed in part by the payment of page (10). Thus, 1,25(OH)2D3 is thought to polarize activated T cells charges. This article must therefore be hereby marked advertisement in accordance toward a Th2 phenotype. In addition, it has been shown that ex- with 18 U.S.C. Section 1734 solely to indicate this fact. pression of the Fas ligand by activated T cells could be repressed 1 Current address: Medimmune, One Medimmune Way, Gaithersburg, MD 20878. by 1,25(OH)2D3 (11). 2 Address correspondence and reprint requests to Dr. Peter E. Lipsky, Intramural Research Program Autoimmunity Branch Chief, National Institute of Arthritis and The effects of l,25(OH)2D3 on B cell function has not been Musculoskeletal and Skin Diseases, 9000 Rockville Pike, Building 10, Room 6D47C, examined in detail. Published reports have yielded conflicting ob- Bethesda, MD 20892. E-mail address: [email protected] servations. It has been claimed that 1,25(OH)2D3 may indirectly 3 Abbreviations used in this paper: 1,25(OH)2D3, 1,25-dihydroxyvitamin D3; AID, ␤ ␤ inhibit B cell function as a consequence of the impairment of CD4 activation-induced cytidine deaminase; ANA, antinuclear Ab; 2M, 2-microglobu- lin; BCL6, B cell 6; BLIMP1, B lymphocyte-induced maturation T cell responses or the inhibition of cytokine production by mono- 1; DC, dendritic cell; ERN1, endoplasmic reticulum to nucleus signaling; IRF4, IFN cytes/macrophages (12). Alternatively, a direct inhibitory effect of regulatory factor 4; PI, propidium iodide; RA, rheumatoid arthritis; SLE, systemic 1,25(OH)2D3 on IgE production by human B cells has also been lupus erythematosus; SLEDAI, SLE disease activity index; VDR, vitamin D3 recep- tor; XBP1, X box-binding protein. reported (13). www.jimmunol.org The Journal of Immunology 1635

Interestingly, decreased 1,25(OH)2D3 serum concentrations Table I. Characteristics of the patients with SLE have been reported in many autoimmune diseases, including sys- temic lupus erythematosus (SLE) (14, 15). SLE is an autoimmune Characteristic SLE Patients (n ϭ 112) disease characterized by immune dysregulation resulting in over- Age in years, mean Ϯ SEM (range in 33.8 Ϯ 1.57 (15–66) production of autoantibodies (16). Although the exact cause of years) SLE remains unknown, recent studies, including a demonstration Female (%) 90.2% (101/112) of the effectiveness of therapeutic B cell depletion, strongly im- Disease duration in years, mean Ϯ SEM 4.5 Ϯ 0.5 (0.01–20) plicate B cells as central players in the pathogenesis of this disease (range in years) Ϯ Ϯ (17–20). We therefore, hypothesized that 1,25(OH) D may be SLEDAI score, mean SEM (range) 6.5 0.9 (0–22) 2 3 Arthritis percentage (no./total) 31.3% (35/112) important in maintaining B cell homeostasis and that deficiency of Rash percentage (no./total) 23.2% (26/112) 1,25(OH)2D3 might contribute to B cell hyperactivity in SLE. Oral ulceration percentage (no./total) 17.9% (20/112) Renal involvement percentage (no./total) 37.5% (42/112) In this study, we measured 1,25(OH)2D3 and 25(OH)D3 levels in patients with SLE and correlated them with disease activity and Vasculitis percentage (no./total) 7.1% (8/112) Hematological abnormalities percentage 25% (28/112) the presence of antinuclear Abs (ANA). Subsequently, we inves- (no./total) tigated the direct effects of 1,25(OH)2D3 on primary human B cells CNS involvement percentage (no./total) 3.6% (4/112) to gain more insight into the potential role of vitamin D on auto- Glucocorticoid use, mean Ϯ SEM 23.2 Ϯ 2.9 (0–160) immune disease pathogenesis. Our results provide evidence that mg/day (range) vitamin D might be a useful alternative therapy for the B cell hyperactivity characteristic of SLE. cytokines and stimuli: 1 ␮g/ml anti-CD40 (R&D Systems), 3 ␮g/ml goat Materials and Methods Ј F(ab )2 anti-IgM (Jackson ImmunoReseach Laboratories), 250 ng/ml IL-21 Clinical samples (Cell Sciences), 50 ng/ml IL-4 (R&D Systems), and 1,25(OH)2D3 (usually 10 nM) or 25(OH) D3 (both from Sigma-Aldrich). For the analysis of serological levels of 25(OH)D and 1,25(OH)2D3,pe- ripheral blood samples from patients with SLE, patients with rheumatoid Proliferation assay arthritis (RA), and demographically comparable healthy controls were ob- tained from the Shanghai Clinical Center for Rheumatology, Renji Hospi- Cells were cultured for 3, 4, or 5 days as described above and pulsed for tal, Shanghai, China after informed consent. The diagnosis of SLE relied an additional 16 h with [3H]thymidine (1 ␮Ci, 37 kBq) and then harvested. on American College of Rheumatology revised criteria (21). Clinical charts The [3H]thymidine incorporation was measured with a Top Count micro- of all patients were reviewed and disease activity was scored according to plate scintillation counter (Packard Instruments). Alternatively, prolifera- the SLE disease activity index (SLEDAI) (22). For the analysis of B cell tion was assessed by CFSE (Invitrogen Life Technologies) dilution. function in vitro, blood samples from normal healthy adult donors were Briefly, 1 ϫ 107 B cells were incubated with 1 ␮M of CFSE for 15 min at obtained from the Warren G. Magnuson Clinical Center Blood Bank (Be- room temperature and then quenched with an equal volume of 100% FCS thesda, MD). The collection of samples was approved by the National Institute for 2 min and washed. CFSE-labeled cells were cultured for 3 to 6 days as of Arthritis and Musculoskeletal and Skin Diseases/National Institute of Dia- mentioned above and data were collected by FACS analysis. betes & Digestive & Kidney Diseases (NIAMS/NIDDK) Institutional Review Board (Bethesda, MD), and informed consent was obtained according to the Ig ELISA Declaration of Helsinki. Culture supernatants were incubated on goat antihuman IgG-and IgM- coated (Bethyl Laboratories) Maxisorp plates (Nalge Nunc International) Determination of serum 25(OH)D and 1,25(OH)2D3 levels and bound Ig was detected using alkaline phosphatase-conjugated goat Serum total 25(OH)D and 1,25(OH)2D3 levels were measured in SLE pa- anti-human IgG or IgM (Bethyl Laboratories) with Sigma Fast pNPP al- tients, patients with RA, and healthy donors by RIA or ELISA (both from kaline phosphate substrate (Sigma-Aldrich). ALPCO Diagnostics) according to manufacturer’s instructions. Both as- says yielded similar results. ELISPOT assay B cell enrichment, flow cytometry, and cell sorting The number of IgG- and IgM-secreting cells was determined with the ELISPOT assay. MultiScreenHTS plates (polyvinylidene difluoride mem- B cells were enriched by negative selection from buffy coats or leuka- brane; Millipore) were coated overnight at 4°C with 50 ␮l/well goat anti- pheresis samples using RosetteSep or StemSep B cell purification Ab human IgG or IgM (Bethyl Laboratories), respectively, at a concentration Ͼ mixtures (StemCell Technologies). Enriched B cells were 90 and 95% of 5 ␮g/ml diluted in 0.05 M carbonate PBS (Sigma-Aldrich). After wash- pure, respectively. Purified B cells were stained with various mAb com- ing, unbound surfaces of wells were blocked with culture medium for 1 h binations for 20 min on ice in staining buffer (1% BSA and 5% FCS in at 25°C. Stimulated cells from 7-day cultures were washed three times and PBS). The directly conjugated mAb used were anti-IgD-FITC, anti-CD27-PE, resuspended with medium containing 10% Ultra Low IgG FCS (Invitrogen anti-CD40-PE, anti-CD86-FITC, anti-IgG-PE, anti-CD19-allophycocyanin, Life Technologies). Serial dilutions of cells were added to wells in tripli- anti-HLA-DR-PE, (BD Pharmingen), anti-CD19-PerCpCy5.5, and anti-CD38- cate and incubated at 37°C overnight. Medium containing 10% Ultra Low allophycocyanin (clone HB7) (BD Immunocytometry). Stained cells were IgG FCS was included in all experiments as a negative control. Thereafter, washed and data were collected immediately using a four-color FACScalibur plates were washed with TBS plus 0.05% Tween 20 four times and incu- (BD Biosciences) or fixed in 1% paraformaldehyde and analyzed within 24 h. bated with either biotinylated goat-anti-human IgM or IgG (Invitrogen) Data were analyzed using FlowJo software (Tree Star). B cell populations diluted 1/2000 in TBS plus 0.05% Tween 20 at 25°C for 3 h. Plates were were sorted using the DakoCytomation MoFlo. Sorted B cell populations were then washed in TBS plus 0.05% Tween 20 and incubated with 0.5 ␮g/ml Ͼ 98% pure. alkaline phosphatase-conjugated streptavidin (Sigma-Aldrich) in TBS plus Cell apoptosis and viability assay 0.05% Tween 20 for1hat25°C. All plates were developed with the alkaline phosphatase substrate kit III (Vector Laboratories) after washing To quantify apoptosis, cultured B cells were double-stained with annexin in distilled water. Ig-secreting cells appeared as blue spots, which were V-FITC conjugate and propidium iodide (PI) using TACS annexin V kits counted using a CTL Series 3B analyzer (Cellular Technology). from Oncogene according to the manufacturer’s directions. The stained B cells were immediately analyzed using a FACSCalibur flow cytometer (BD RNA preparation and quantitative RT-PCR Biosciences). Total RNA from B cells was purified by using the RNeasy mini kit B cell stimulation (Qiagen) according to the manufacturer’s manual. Reverse reactions were prepared using the SuperScript one-step PCR system with B cells were cultured in 96-well plates at 1 ϫ 106 cells/ml in 100-␮l platinum Taq polymerase and ROX reference dye (Invitrogen Life Tech- volumes of culture medium (10% FCS in RPMI 1640 medium with L- nologies). Fifty nanograms of isolated RNA was added per reaction with glutamine and penicillin-streptomycin) alone or supplemented with various 1.2 mM MgSO4. TaqMan Assays-on-Demand expression primer/ 1636 1,25-DIHYDROXYVITAMIN D3 MODULATES B CELLS

FIGURE 1. Patients with SLE have decreased 25(OH)D and 1,25(OH)2D3 levels. Serum levels of 25(OH)D and 1,25(OH)2D3 were measured by RIA and ELISA. A, 25(OH)D levels in SLE patients (n ϭ 57) compared with normal controls (NC; n ϭ 28) and RA patients (n ϭ 29). Mean levels in SLE patients were significantly different from those in NC and RA patients (p Ͻ 0.001). B, 25(OH)D levels in patients with established SLE with a mean duration since diagnosis of 5.11 Ϯ 0.88 years (n ϭ 45) compared with newly diagnosed untreated patients with a mean duration of 25.8 Ϯ 9.2 days (n ϭ 12). There Ͼ was no significant difference in mean 25(OH)D levels in newly diagnosed or established SLE patients (p 0.05). C, 1,25(OH)2D3 levels in SLE patients ϭ ϭ Ͻ (n 87) compared with normal controls (NC) (n 30). The 1,25(OH)2D3 levels were significantly lower in SLE patients (p 0.001). D, 1,25(OH)2D3 levels in nonactive SLE patients (SLEDAI Յ 4, n ϭ 35) compared with active patients (SLEDAI Ͼ 4, n ϭ 52). Active SLE patients had significantly lower Ͻ ϭ ϭ levels of 1,25(OH)2D3 (p 0.001). E, 1,25(OH)2D3 levels in SLE patients with positive ANA (n 71) compared with those without ANA (n 11). Ͻ ϭ Patients without ANA had significantly higher levels of1,25(OH)2D3 (p 0.05). F, 1,25(OH)2D3 levels in SLE patients with anti-dsDNA (n 39) ϭ Ϯ compared with those without anti-dsDNA (n 43). 1,25(OH)2D3 levels did not differ between these groups. Data are shown as mean SEM, and .(p Ͻ 0.05 ,ء ;p Ͻ 0.01 ,ءء ;p Ͻ 0.001 ,ءءء) significant differences between the groups are shown

␤ ␤ 2 probe sets for 2-microglobulin ( 2M; assay identifier Hs99999907_ were examined using human cell cycle RT Profiler PCR array and human m1), VDR (Hs01045840_m1), CYP27B1 (Hs00168017_m1), CYP24A1 NF␬B signaling pathway RT2 Profiler PCR array, respectively (SuperArray (Hs00167999_m1), Pax5 (paired box gene 5; Hs00277134_m1), Bcl6 (B Bioscience) according to the manufacturer’s instructions. In brief, cDNA cell CLL/lymphoma 6, zinc finger protein 51; Hs00153368_m1), was prepared from 1 ␮g total RNA by using a RT2 PCR array first strand BLIMP1(B-lymphocyte-induced maturation protein 1; Hs00153357_m1), kit. A total volume of 25 ␮l of PCR mixture, which included 12.5 ␮lof IFN regulatory factor 4 (IRF4; Hs00180031_m1), X box-binding protein 1 RT2 Real-Time SYBR Green/ROX PCR master mix from SuperArray Bio- (XBP1; Hs00231936_m1), endoplasmic reticulum to nucleus signaling 1 science (containing HotStart DNA polymerase, SYBR Green dye, and the ␮ ␮ (ERN1; Hs00176385_m1), activation-induced cytidine deaminase (AID; ROX reference dye), 11.5 l of double-distilled H2O, and 1 l of template Hs00221068_m1), p27 (cyclin-dependent kinase (CDK) inhibitor 1B; cDNA, was loaded in each well of the PCR array. PCR amplification was Hs00153277_m1), p21 (CDK inhibitor 1A; Hs00355782_m1), and p18 conducted with an initial 10-min step at 95°C followed by 40 cycles of (CDK inhibitor 2C; Hs00176227_m1) were purchased from Applied Bio- 95°C for 15 s and 60°C for 1 min). The fluorescent signal from SYBR systems, and quantitative RT-PCR was performed according to the man- Green was detected immediately after the extension step of each cycle, and ufacturer’s instructions. mRNA expression for each gene was calculated in the cycle at which the product was first detectable was recorded as the triplicate using the comparative cycle threshold method with efficiency cycle threshold. Data were imported into an Excel database and analyzed ␤ calculations, and all mRNA levels were normalized to 2M. using the comparative cycle threshold method with normalization of the ␤ The effects of 1,25(OH)2D3 on the expression of 84 cell cycle regulated raw data to housekeeping including 2M, hypoxanthine phosphori- genes and 84 genes associated with NF-␬B-mediated signal transduction bosyltransferase 1, ribosomal protein L13a, GAPDH, and ACTB (␤-actin). The Journal of Immunology 1637

FIGURE 2. Vitamin D (VitD) inhibits the proliferation of human peripheral B cells. A, Purified B cells were cultured with (VitD) or without (nil) Ϯ 1,25(OH)2D3 (10 nM) and stained with annexin V/PI. Percentages of viable cells (double negative for annexin V and PI) are shown as mean SEM of triplicate cultures and are representative of three independent experiments. B, Purified B cells were stimulated with combinations of anti-IgM (␣IgM), ␣ anti-CD40 ( CD40), and IL-21 as indicated in the presence (VitD) or absence (nil) of 1,25(OH)2D3 (10 nM). After 3 days, cells were pulsed with [3H]thymidine and incorporation was measured 16 h later. Data are the mean Ϯ SEM of triplicate cultures and are representative of three independent p Ͻ 0.01). C, Purified B cells labeled with CFSE were ,ءء ;0.05 ,ء) experiments. Significant differences between vitamin D and control responses are shown ϩ cultured with anti-IgM, anti-CD40, and IL-21 in the presence (VitD) or absence (nil) of 1,25(OH)2D3 (10 nM). At the time points indicated, CD19 B cells were analyzed for CFSE dilution. Data shown are representative of results from three experiments.

Analysis and statistics lower in patients with ANA compared with those without ANA Ϯ ϭ Ϯ ϭ Ͻ Statistical significance of the data was analyzed with a two-tailed, non- (14.4 1.3 pg/ml, n 71 vs 22.0 4.2 pg/ml, n 11, p 0.05), paired or paired Student’s t test, using Microsoft Excel Windows software. but no significant difference was observed in patients with or with- ,denotes p Ͻ 0.05, out anti-dsDNA (15.0 Ϯ 1.8 pg/ml, n ϭ 39 vs 13.5 Ϯ 1.7 pg/ml (ء) Data are plotted as mean Ϯ SEM. An asterisk denotes p Ͻ 0.001. n ϭ 43) ( p Ͼ 0.05) (Fig. 1, E and F). The data indicated that ءءء denotes p Ͻ 0.01, and ءء patients with SLE had decreased levels of 25(OH) D and Results 2 1,25(OH)2D3, and the latter was related to disease activity and SLE patients have significantly diminished levels of 25(OH) D ANA production. These results prompted us to examine the impact and 1,25(OH)2D3 of vitamin D on B cell function. A total of 112 SLE patients who fulfilled American College of Rheumatology revised criteria (21) were assessed. Their charac- 1,25(OH)2D3 inhibits proliferation of activated B cells teristics are shown in Table I. The serum 25(OH)D level was sig- Initially, the effect of 1,25(OH)2D3 on B cell survival was examined. Ϯ ϭ nificantly lower in SLE patients (11.5 1.5 ng/ml, n 57) than 1,25(OH)2D3 did not affect the survival of unstimulated B cells in Ϯ in demographically comparable healthy controls (59.2 6.5 ng/ culture (Fig. 2A). We next assessed the effect of 1,25(OH)2D3 on B ϭ Ϯ ϭ Ͻ ml, n 28) and RA patients (54.6 5.2 ng/ml, n 29) ( p cell proliferation. As shown in Fig. 2B, 1,25(OH)2D3 exerted a sig- 0.001) (Fig. 1A). Notably, the level of 25(OH)D in newly diag- nificant inhibitory effect on B cell proliferation, especially in cultures nosed SLE patients without any treatment was 11.6 Ϯ 2.1 ng/ml activated with a combination of stimuli, e.g., anti-IgM/anti-CD40 (n ϭ 12), which was not statistically different from that in SLE ( p Ͻ 0.05), anti-CD40/IL-21, or anti-IgM/anti-CD40/IL-21 ( p Ͻ patients with established disease (11.8 Ϯ 1.8 ng/ml, n ϭ 45) ( p Ͼ 0.01). B cells stimulated with either anti-CD40 or anti-IgM alone had 0.05) (Fig. 1B). No correlation was found between the level of a limited proliferative capacity and, as a consequence, the effect of

25(OH)D and various clinical parameters including renal involve- 1,25(OH)2D3 was less consistent. ment, disease activity assessed by SLEDAI (22), and glucocorti- CFSE labeling was also used to examine the effect of Ϯ coid use. The serum level of 1,25(OH)2D3 in SLE patients (14.5 1,25(OH)2D3 on the proliferation of B cells. After 3 days in culture 1.2 pg/ml, n ϭ 87) was also significantly lower than in normal some CFSE dilution was observed. Notably, cultures with and Ϯ ϭ Ͻ controls (29.8 1.5 pg/ml, n 30, p 0.001) (Fig. 1C). without 1,25(OH)2D3 exhibited a similar frequency of cells with 1,25(OH)2D3 levels were significantly lower in patients with active modestly diluted CFSE (Fig. 2C). As the culture period was pro- SLE as indicated by a SLEDAI score of Ͼ4 (12.2 Ϯ 1.6 pg/ml, longed, cells with increasing dilution of CFSE were observed. n ϭ 52) compared with those with a SLEDAI score of Յ4 (19.4 Ϯ However, the extent of CFSE dilution was strikingly diminished ϭ Ͻ 1.9 pg/ml, n 36, p 0.001) (Fig. 1D). Additional analysis in 1,25(OH)2D3-containing cultures compared with control cul- demonstrated that the level of 1,25(OH)2D3 was significantly tures. These results were consistent with the conclusion that 1638 1,25-DIHYDROXYVITAMIN D3 MODULATES B CELLS

FIGURE 3. Vitamin D (VitD) induces apoptosis of activated human peripheral B cells. Purified B cells were cultured with no stimulus (nil), anti-IgM ␣ ␣ ( IgM), and anti-CD40 ( CD40) in the presence (VitD) or absence (Nil) of 1,25(OH)2D3 (10 nM). A, After 4 days in culture, cells were washed and stained with annexin V and PI and analyzed by flow cytometry. The numbers in the upper left quadrants represent the percentage of apoptotic cells in culture. B, Percentages of apoptotic cells in 3- and 4-day cultures with no stimulus (nil) or anti-IgM (␣IgM) and anti-CD40 (␣CD40) and with or without the addition Ϯ of 1,25(OH)2D3 (VitD; 10 nM). Data are mean SEM of triplicate determinations from one of three independent experiments. Asterisks indicate the .(p Յ 0.01 ,ءء) statistical significance of the difference between vitamin D and untreated B cells

1,25(OH)2D3 had little effect on initial cell division, whereas VDR and 24-hydroxylase can be induced by 1,25(OH)2D3 ongoing proliferation of stimulated B cells was suppressed. on B cells We next examined the expression of the VDR and the known 1,25(OH)2D3 induces apoptosis of activated B cells vitamin D responsive gene 24-hydroxylase (CYP24A1) by quan- To determine whether the inhibitory effect of 1,25(OH)2D3 on pro- liferation was associated with the induction of apoptosis, we ex- titative PCR. Freshly isolated B cells constitutively expressed very amined the percentage of apoptotic cells by annexin V/PI staining. low levels of VDR mRNA. Compared with day 0 expression, the mean fold increase of VDR expression by B cells after 3 days of During 4 days in culture without stimulation 1,25(OH)2D3 treat- ment had no significant effect on survival or apoptosis (Fig. 3A). culture with anti-IgM/anti-CD40, anti-CD40/IL-21, anti-IgM/anti- Stimulation with anti-IgM/anti-CD40 decreased the frequency of CD40/IL-21, or without stimulation was 1.45, 2.88, 4.44, and 1.46- apoptotic cells, but in cultures of activated B cells,1,25(OH)2D3 fold respectively. 1,25(OH)2D3 increased VDR expression in both significantly increased the percentage of early apoptotic cells at activated and unstimulated B cells (Fig. 4A), but not earlier than day 4 (27.5% vs 37.5%, p Ͻ 0.01). This was time dependent, as day 3 (data not shown). In contrast, CYP24A1 expression was this difference was not apparent at earlier time points (Fig. 3B). not detectable in either resting or activated B cells. However, it

FIGURE 4. Vitamin D (VitD)-inducible genes are up-regulated on activated B cells. Purified B cells were stimulated in various conditions in the presence (VitD) or absence (nil) of 1,25(OH)2D3 (10 nM) as indicated. Total RNA was extracted from cells at day 0 and after 3 days in culture. was detected by quantitative RT-PCR in triplicate. A, Expression of VDR relative to ␤2M. Left panel shows results from stimulated vs fresh B cells, and the right panel shows results from B cells culture without stimulation (␣IgM, anti-IgM; ␣CD40, anti-CD40). B, Expression of CYP24A1 relative to ␤2M. Left panel shows results from stimulated vs fresh B cells, and the right panel shows results from B cells cultured without stimulation. Data are the mean Ϯ SEM from triplicates from one of two representative experiments. Significant differences between vitamin D (VitD) and control (nil) .(p Ͻ 0.001 ,ءءء ;p Ͻ 0.01 ,ءء ;p Ͻ 0.05 ,ء) responses are shown The Journal of Immunology 1639

FIGURE 5. Vitamin D (VitD) increases CD38 expression on activated B cells. A, Purified B cells were cultured with various stimuli in the presence or absence (nil) of 1,25(OH)2D3 (10 nM). After 4 or 5 days in culture, cells were washed and cell surface markers were analyzed by flow cytometry. The expression of IgD, CD27, CD38, CD21, CD23, HLA-DR, CD86, and CD40 was compared between vitamin D (solid line) and nil (dotted line) cultures. (␣IgM, anti-IgM; ␣CD40, anti-CD40). B, The plasma cells population based on the phenotype, CD19ϩ/Ϫ, CD27high, CD38high (upper right quadrant), was ␣ ␣ compared between cultures with or without 1,25(OH)2D3 (10 nM) after 7 days of culture with the stimuli indicated ( IgM, anti-IgM; CD40, anti-CD40). C, Concentration-dependent effect of 1,25(OH)2D3 (VitD; 10 nM) on CD38 expression and plasma cell generation from B cells cultured with anti-CD40 and IL-21 for 7 days. Geom, Geometric.

was significantly induced by 1,25(OH)2D3(Fig. 4B). The results CD38 expression (data not shown), but CD40 expression was not demonstrate that the VDR is expressed and inducible in primary B altered by 1,25(OH)2D3. These results show that 1,25(OH)2D3 cells. Activation signals or 1,25(OH)2D3 up-regulated VDR ex- specifically increases CD38 expression by activated B cells but has pression, whereas CYP24A1 was only induced by 1,25(OH)2D3 no effect on the expression of a variety of other lineage, differen- but not by stimulation. tiation, or activation markers. Increased expression of CD38 is commonly used as a pheno- 1,25(OH)2D3 specifically increases CD38 expression on B cells typic marker of plasma cells (25). To determine whether the

It was previously reported that 1,25(OH)2D3 could increase CD38 1,25(OH)2D3-mediated increase in CD38 expression was associ- expression by lymphocytes (23, 24). To confirm this and to exam- ated with plasma cell differentiation, we used a method of in vitro ine the effect of 1,25(OH)2D3 on other B cell markers, we activated stimulation known to induce plasma cells from peripheral B cells purified B cells with various stimuli with and without (anti-CD40 and IL-21 with or without anti-IgM) and examined the

1,25(OH)2D3, and then stained the cells for various surface mark- effect of 1,25(OH)2D3 (26). The induction of CD38 resulting from ers. The expression of CD38 on activated B cells was significantly 1,25(OH)2D3 was not associated with an increase in CD27 expres- high high increased by 1,25(OH)2D3. This difference was not observed with sion and, indeed, fewer CD38 CD27 plasma cells were gen- unstimulated B cells or cultures activated with anti-CD40 or anti- erated in the presence of 1,25(OH)2D3 (Fig. 5B). Both the increase IgM alone, but it was apparent in B cells stimulated with the com- of CD38 expression and the decrease of plasma cell generation bination of anti-IgM/anti-CD40, anti-CD40/IL-21, or anti-IgM/anti- occurred in a concentration-dependent manner, with the maximum CD40/IL-21. In contrast, the expressions of IgD, CD21, CD23, effect noted reproducibly at 10 nM or less (Fig. 5C). CD27, CD86, and HLA-DR were altered by activating stimuli 1,25(OH)2D3 inhibits plasma cell differentiation and Ig secretion but not further affected by 1,25(OH)2D3 (Fig. 5A). Similarly, 1,25(OH)2D3 had no effect on the expression of CD19 or CD20 A more detailed analysis showed that 1,25(OH)2D3 inhibited (data not shown). To examine the expression of CD40, we used plasma cell differentiation and the secretion of Ig (Fig. 6). The anti-IgM and IL-4 to activate B cells and up-regulate CD40 ex- generation of plasma cells was inhibited throughout the culture pression. Using these conditions, 1,25(OH)2D3 again increased period after day 5 and 1,25(OH)2D3 caused significant decreases in 1640 1,25-DIHYDROXYVITAMIN D3 MODULATES B CELLS

FIGURE 6. The inhibitory effect of vitamin D (VitD) on plasma cell differentiation and Ig production. Purified B cells were cultured with anti-CD40 ␣ ␣ ( CD40) and IL-21 with or without anti-IgM I ( IgM) in the presence (VitD) or absence (nil) of 1,25(OH)2D3 (10 nM). A, The generation of plasma cells (CD19ϩ/Ϫ CD27highCD38high) was assessed after 5, 7, and 9 days in culture (left panel). The frequencies of plasma cells in vitamin D vs nil cultures after a 7-day incubation of anti-CD40/IL-21 (n ϭ 12) and anti-IgM/anti-CD40/IL-21 stimulated B cells (n ϭ 8) were determined and shown as the mean percentage of the no vitamin D control (right panel). B, IgG and IgM production determined by ELISA. The results are the mean Ϯ SEM and are p Ͻ 0.001). C, IgG- (left) and IgM-secreting cells ,ءءء ;p Ͻ 0.01 ,ءء) representative of five independent experiments. Significant differences are shown (right) were determined at day 7 by ELISPOT. Data are the mean Ϯ SEM from triplicate determinations and are representative of one of three independent ␣ experiments. D, Purified B cells were cultured with anti-CD40 ( CD40) and IL-21 for 5 days, after which 1,25(OH)2D3 (10 nM) was added to the culture. The effects of vitamin D (VitD) on the persistence of plasma cells and Ig production over the next 5 days were assessed and compared with the no vitamin .(p Ͻ 0.05 ,ء) D controls (nil). Data are the mean Ϯ SEM from triplicate determinants. Significant differences are shown as both the percentage (Fig. 6A, left panel) and absolute number of percent inhibition progressively increased as the cultures were plasma cells (data not shown). In multiple experiments, the gen- prolonged with the maximal inhibition noted at day 9 ( p Ͻ eration of phenotypically defined CD19ϩ/ϪCD38highCD27high 0.001). A significant albeit smaller reduction in IgM production plasma cells was inhibited by 23–34% (Fig. 6A, right panel). Sim- in 1,25(OH)2D3 cultures was only apparent at day 9. The effect ilar results were noted when B cells purified by cell sorting and of 1,25(OH)2D3 on the number of Ig-secreting cells generated Ͼ98% pure were examined (data not shown). in these cultures was determined by ELISPOT assay. In agree-

Analysis of culture supernatants revealed that 1,25(OH)2D3 ment with the phenotype and Ig production data, 1,25(OH)2D3 also significantly decreased the secretion of Ig. Stimulation with significantly reduced the number of IgG-secreting cells in cul- anti-CD40 and IL-21 with or without anti-IgM induced consid- tures of activated B cells. The number of IgM secreting cells erable Ig secretion by day 5 that increased through day 9 (Fig. was also significantly decreased by 1,25(OH)2D3 in anti-CD40/ 6B). 1,25(OH)2D3 significantly inhibited IgG secretion. The IL-21 cultures (Fig. 6C). The Journal of Immunology 1641

FIGURE 7. Vitamin D (VitD) in- hibits the generation of postswitch memory B cells from naive B cells. A, Preswitched naive B cells (CD19ϩ IgGϪCD27Ϫ) were sorted from en- riched peripheral B cells from three healthy donors. Postsort analysis of the naive B cells is shown. The purity of sorted cells was Ͼ99%. B, Naive B cells were stimulated as indicated with (VitD) or without 1,25(OH)2D3 (10 nM). After 5 days in culture, the B cells were stained for expression of CD19 and IgG or IgM (upper right quadrant)(␣IgM, anti-IgM; ␣CD40, anti-CD40). C, The percentages of postswitched memory B cells were compared between culture with (VitD) or without (nil) vitamin D. D, After 7 days in culture, the numbers of plasma cells and IgG in the super- natants were analyzed. Data are the mean Ϯ SEM from triplicate determi- nations of one of three independent experiments. Asterisks indicate the sta- tistical significance between vitamin D ;p Յ 0.05 ,ء) treated and control B cells .(p Ͻ 0.001 ,ءءء ;p Ͻ 0.01 ,ءء

Notably, when 1,25(OH)2D3 was added at day 5 of culture, no whereas there were more B cells expressing surface IgM (Fig. 7B). effect on the maintenance of plasma cells or on Ig secretion was This reduction in class switching correlated with reduced plasma noted over the next 5 days (Fig. 6D). This result indicates that cell numbers and IgG production in parallel 7-day cultures (Fig.

1,25(OH)2D3 inhibits the generation of plasma cells but not their 7D). These data demonstrate that 1,25(OH)2D3 inhibits generation subsequent persistence or Ig secretion. of both postswitch memory cells and plasma cells.

1,25(OH)2D3 inhibits the generation of postswitched memory B 1,25(OH)2D3 down-regulates the expression of XBP1 and ERN1 cells from naive B cells mRNA but not that of PAX5, B cell lymphoma 6 (BCL6), AID, B lymphocyte-induced maturation protein 1 (BLIMP1), MTA3, To determine whether the inhibitory effect of 1,25(OH)2D3 was specific for plasma cell generation, we also examined its role in and IRF4 regulating the generation of postswitch memory B cells. To ad- Thus far, the results clearly demonstrated that 1,25(OH)2D3 re- dress this, naive B cells (CD19ϩIgGϪCD27Ϫ) were purified by duced B cell proliferation, the generation of postswitched memory sorting (Ͼ99% purity) as shown in Fig. 7A and stimulated in a B cells, plasma cell differentiation, and Ig secretion. However, the manner known to induce class switch recombination (26). As stage at which 1,25(OH)2D3 exerts its effects remained unclear. To shown in Fig. 7B, the combination of anti-CD40 and IL-21 with or gain further insight into the nature of B cell inhibition by without anti-IgM was able to induce postswitched cells expressing 1,25(OH)2D3, we examined the expression of key transcription surface IgG. In the absence of IL-21, few cells expressed surface factors involved in B cell maturation and plasma cell differentia- IgG. Notably, the frequency of IgG positive cells was significantly tion. We compared the expression of genes involved in germinal lower in the 1,25(OH)2D3-containing cultures (Fig. 7, B and C), center reactions, including PAX5, BCL6, and AID as well as those 1642 1,25-DIHYDROXYVITAMIN D3 MODULATES B CELLS

␬ Table II. Effect of 1,25(OH)2D3 on NF- B regulatory gene expression

Mean of Fold p mRNA Change Value

Angiotensinogen (serpin peptidase inhibitor, 0.55 0.09 clade A, member 8) v-akt murine thymoma viral oncogene 1.10 0.34 homolog 1 Activating 1 1.19 0.11 B cell CLL/lymphoma 10 1.16 0.11 B cell CLL/lymphoma 3 2.12 0.16 B factor, properdin 0.68 0.05 Baculoviral IAP repeat-containing 2 0.73 0.12 Caspase recruitment domain family, member 4 0.79 0.07 Caspase 1, apoptosis-related cysteine peptidase 0.85 0.29 (interleukin 1␤, convertase) Caspase 8, apoptosis-related cysteine peptidase 0.87 0.35 Chemokine (C-C motif) ligand 2 0.09 0.12 CD40 antigen (TNF receptor superfamily 0.85 0.20 member 5) CASP8 and FADD-like apoptosis regulator 1.29 0.12 Conserved helix-loop-helix ubiquitous kinase 1.08 0.42 Colony-stimulating factor 2 (granulocyte- 0.51 0.27 macrophage) Colony stimulating factor 3 (granulocyte) 0.63 0.16 Solute carrier family 44, member 2 1.45 0.00 EDAR-associated death domain 0.48 0.04 Endothelial differentiation, lysophosphatidic 0.89 0.39 acid G-protein-coupled receptor, 2 FIGURE 8. Vitamin D (VitD) inhibits XBP1 and ERN1 expression by Early growth response 1 1.37 0.22 activated B cells. RNAs were extracted from freshly isolated B cells and B ELK1, member of ETS oncogene family 0.88 0.30 Coagulation factor II (thrombin) receptor 0.59 0.10 cells cultured with anti-IgM, anti-CD40, and IL-21 with (VitD) or without Fas (TNFRSF6)-associated via death domain 0.88 0.25 (nil) 1,25(OH)2D3 (10 nM) for 3 days. The mRNA expression of various Fas ligand (TNF superfamily, member 6) 0.84 0.37 gene products was determined by RT-PCR and the fold change was com- v-fos FBJ murine osteosarcoma viral oncogene 0.83 0.32 pared with freshly isolated B cells. A, Relative expression of PAX, BCL6, homolog and AID. B, Relative expression of BLIMP1, IRF4, MAT3, XBP1, and Gap junction protein, ␣1, 43 kDa (connexin 43) 0.16 0.20 ERN1. Data represent the mean Ϯ SEM of three independent experiments Heme oxygenase (decycling) 1 1.36 0.21 conducted in triplicate from different donors. Significant differences be- 5-Hydroxytryptamine (serotonin) receptor 2B 0.64 0.04 p Ͻ 0.05). Intercellular adhesion molecule 1 (CD54), 0.91 0.40 ,ء) tween vitamin D and control responses are shown human rhinovirus receptor Interferon, ␣1 0.61 0.16 Interferon, ␤1, fibroblast 0.55 0.14 involved in plasma cell differentiation, including BLIMP1, IRF4, Interferon, ␥ 0.80 0.39 MTA3, XBP1, and ERN1 (27). All of the analyses were conducted Inhibitor of ␬ light polypeptide gene enhancer 1.11 0.39 ␤ at day 3 before significant effects on the frequency of memory B in B cells, kinase Inhibitor of ␬ light polypeptide gene enhancer 0.86 0.33 cells and plasma cells or Ig secretion were evident. Interestingly, in B cells, kinase ␧ 1,25(OH)2D3 had no significant effect on the expression of PAX5, Inhibitor of ␬ light polypeptide gene enhancer 0.81 0.04 BCL6, AID, BLIMP1, MAT3, or IRF4. However, even at this in B cells, kinase ␥ relatively early time point the expressions of XBP1 and ERN1 Interleukin 10 0.81 0.27 Interleukin 1, ␣ 0.70 0.01 were significantly reduced but not eliminated by 1,25(OH)2D3 Interleukin 1, ␤ 0.38 0.05 (Fig. 8). The lack of a consistent effect of 1,25(OH)2D3 on the Interleukin 1 receptor, type I 0.71 0.14 genetic programs regulating the differentiation of memory cells or Interleukin 6 (interferon, ␤2) 3.27 0.01 plasma cells suggests that it may inhibit their maturation by other Interleukin 8 0.76 0.22 means. Interleukin-1 receptor-associated kinase 1 0.66 0.19 Interleukin-1 receptor-associated kinase 2 3.09 0.09 v-jun sarcoma virus 17 oncogene homolog 0.93 0.36 1,25(OH)2D3 increases the expression of p27 mRNA but not that (avian) of p21 and p18 Lymphotoxin ␣ (TNF superfamily, member 1) 0.60 0.15 ␤ To determine the mechanism involved in the inhibitory effect of Lymphotoxin receptor (TNFR superfamily, 0.96 0.47 ␬ member 3) 1,25(OH)2D3 on B cells, we initially screened an array of NF- B- Mucosa-associated lymphoid tissue lymphoma 0.88 0.24 mediated signal transduction pathway genes. As shown in Table II, translocation gene 1 Mitogen-activated protein kinase kinase kinase 1 1.25 0.23 a few mRNAs were up- or down-regulated by 1,25(OH)2D3, but no general pattern of gene regulation was observed consistent with Myeloid differentiation primary response gene 1.01 0.48 ␬ (88) a major effect on the NF- B signaling pathway. We next examined NACHT, leucine rich repeat and PYD 0.26 0.19 the effect of 1,25(OH)2D3 on cell cycle-related gene expression by containing 12 B cells (shown in Table III). We found a number of genes, includ- Nuclear factor of ␬ light polypeptide gene 0.95 0.40 ing cyclin D1, cyclin D2, cyclin T1, cyclin T2, CDK4, and CDK6, enhancer in B-cells 1 (p105) ␬ were significantly down-regulated by 1,25(OH) D . In contrast, Nuclear factor of light polypeptide gene 1.00 0.50 2 3 enhancer in B-cells 2 (p49/p100) the CDK inhibitor p27 was significantly up-regulated by (Table continues) 1,25(OH)2D3, whereas most genes, including CDK inhibitor p21, The Journal of Immunology 1643

Table II. (Continued) Table III. Effect of 1,25(OH)2D3 on cell cycle regulatory gene expression Mean of Fold p mRNA Change Value Mean of Fold p Nuclear factor of ␬ light polypeptide gene 1.56 0.03 mRNA Change Value enhancer in B-cells inhibitor, ␣ Protein phosphatase 1A (formerly 2C), 0.91 0.28 Cyclin-dependent kinase 6 0.83 0.12 magnesium-dependent, ␣ isoform Cyclin-dependent kinase 7 (MO15 homolog, 0.99 0.43 v-raf-1 murine leukemia viral oncogene 0.92 0.38 Xenopus laevis, cdk-activating kinase) homolog 1 Cyclin-dependent kinase 8 0.79 0.13 v- reticuloendotheliosis viral oncogene 0.70 0.21 Cyclin-dependent kinase inhibitor 1A (p21, Cip1) 0.74 0.08 homolog (avian) Cyclin-dependent kinase inhibitor 1B (p27, Kip1) 2.44 0.03 v-rel reticuloendotheliosis viral oncogene 1.10 0.36 Cyclin-dependent kinase inhibitor 2A (melanoma, 1.65 0.38 homolog A, nuclear factor of ␬ light p16, inhibits CDK4) polypeptide ger Cyclin-dependent kinase inhibitor 2B (p15, 0.60 0.11 v-rel reticuloendotheliosis viral oncogene 0.84 0.23 inhibits CDK4) homolog B, nuclear factor of ␬ light Cyclin-dependent kinase inhibitor 3 (CDK2- 0.96 0.35 polypeptide ger associated dual specificity phosphatase) Ret finger protein 2 0.92 0.24 CHK1 checkpoint homolog 0.72 0.19 Ras homolog gene family, member A 1.08 0.36 (Schizosaccharomyces pombe) Receptor (TNFRSF)-interacting serine- 1.19 0.25 CHK2 checkpoint homolog (S. pombe) 0.83 0.17 threonine kinase 1 CDC28 protein kinase regulatory subunit 1B 0.76 0.08 Solute carrier family 20 (phosphate 0.80 0.20 CDC28 protein kinase regulatory subunit 2 0.86 0.04 transporter), member 1 Cullin 1 0.98 0.36 Signal transducer and activator of transcription 1.01 0.49 Cullin 2 0.74 0.14 1, 91 kDa Cullin 3 0.71 0.10 TANK-binding kinase 1 0.98 0.45 DEAD/H (Asp-Glu-Ala-Asp/His) box 1.13 0.23 Toll-like receptor adaptor molecule 2 0.55 0.02 polypeptide 11 (CHL1-like helicase homolog, Toll-like receptor 1 0.75 0.16 Saccharomyces cerevisiae) Toll-like receptor 2 0.30 0.10 Dynamin 2 1.13 0.31 Toll-like receptor 3 1.47 0.26 transcription factor 4, p107/p130 binding 0.82 0.23 Toll-like receptor 4 1.04 0.46 Growth arrest and DNA-damage inducible, ␣ 1.85 0.22 Toll-like receptor 6 0.83 0.07 General transcription factor IIH, polypeptide 1, 0.96 0.26 Toll-like receptor 7 0.65 0.01 62 kDa Toll-like receptor 8 0.72 0.34 G2 and S phase expressed 1 0.93 0.15 Toll-like receptor 9 0.69 0.22 Hect domain and RLD 5 0.84 0.13 Transmembrane emp24 protein transport 0.99 0.48 HUS1 checkpoint homolog (S. pombe) 0.85 0.18 domain containing 4 Kinetochore associated 1 0.79 0.01 Tumor necrosis factor (TNF superfamily, 0.48 0.07 Karyopherin ␣2 (RAG cohort 1, importin ␣1) 0.74 0.05 member 2) MAD2 mitotic arrest deficient-like 1 (yeast) 0.92 0.27 Tumor necrosis factor, ␣-induced protein 3 0.74 0.24 MAD2 mitotic arrest deficient-like 2 (yeast) 0.81 0.05 Tumor necrosis factor receptor superfamily, 0.56 0.05 MCM2 minichromosome maintenance deficient 0.90 0.14 member 10a 2, mitotin (S. cerevisiae) Tumor necrosis factor receptor superfamily, 0.82 0.24 MCM3 minichromosome maintenance deficient 3 1.19 0.42 member 10b (S. cerevisiae) Tumor necrosis factor receptor superfamily, 0.53 0.01 MCM4 minichromosome maintenance deficient 4 0.79 0.13 member 1A (S. cerevisiae) Tumor necrosis factor receptor superfamily, 0.95 0.44 MCM5 minichromosome maintenance deficient 1.21 0.32 member 7 5, cell division cycle 46 (S. cerevisiae) Tumor necrosis factor (ligand) superfamily, 1.16 0.28 Antigen identified by monoclonal antibody Ki-67 0.65 0.04 member 10 Ménage-à-trois 1 (CAK assembly factor) 0.72 0.05 Tumor necrosis factor (ligand) superfamily, 1.82 0.23 MRE11 meiotic recombination 11 homolog A 0.73 0.02 member 14 (S. cerevisiae) TNFRSF1A-associated via death domain 1.28 0.10 Nibrin 0.87 0.14 Toll-like receptor adaptor molecule 1 1.39 0.10 Proliferating cell nuclear antigen 0.83 0.09 RAD1 homolog (S. pombe) 1.01 0.48 RAD17 homolog (S. pombe) 1.07 0.12 RAD51 homolog (RecA homolog, Escherichia 1.56 0.22 remained unchanged. We verified these findings using quantitative coli) (S. cerevisiae) PCR and found that p27, but not p21 or p18 mRNA, was signif- RAD9 homolog A (S. pombe) 0.96 0.44 icantly up-regulated by 1,25(OH) D in activated B cells (Fig. 9). Retinoblastoma 1 (including osteosarcoma) 0.94 0.27 2 3 Retinoblastoma binding protein 8 0.85 0.14 B cells can directly metabolize 25(OH)D into a biologically Retinoblastoma-like 1 (p107) 1.31 0.27 3 Retinoblastoma-like 2 (p130) 0.97 0.37 active molecule A3, 14 kDa 1.36 0.22 SERTA domain containing 1 1.35 0.18 It has been shown that the precursor of 1,25(OH)2D3, 25(OH)D3, ␣ S-phase kinase-associated protein 2 (p45) 0.70 0.18 is present in serum and is converted to the active form by 1 - SMT3 suppressor of mif two 3 homolog 1 0.72 0.02 hydroxylase (CYP27B1), which is mainly expressed in kidney and (yeast) also can be found in monocytes/macrophages (3, 4). The expres- Transcription factor Dp-1 0.85 0.11 sion of 1␣-hydroxylase (CYP27B1) mRNA by human peripheral Transcription factor Dp-2 (E2F dimerization 0.91 0.26 B cells was therefore investigated by RT-PCR. As shown in Fig. partner 2) ␣ Tumor protein (Li-Fraumeni syndrome) 0.97 0.37 10A, B cells constitutively expressed 1 -hydroxylase mRNA, and Ubiquitin-activating enzyme E1 (A1S9T and 2.86 0.21 it was up-regulated following stimulation but not further induced BN75 temperature sensitivity complementing) by 1,25(OH)2D3. To determine whether this up-regulation had a 1644 1,25-DIHYDROXYVITAMIN D3 MODULATES B CELLS

mRNA is constitutively expressed in human primary B cells at low levels and is up-regulated following stimulation and enhanced in

the presence of 1,25(OH)2D3 in a time-dependent manner. Previ- ously, the expression of VDR on B cells has been controversial. It has been claimed that resting B cells do not contain detectable amounts of VDR (13). However, it has also been reported that VDR is constitutively expressed on human tonsil B cells and can be further up-regulated by activation (13, 24) and that VDR is expressed on the B cell lymphoma cell lines SUDHL4 and SUDHL5 (28). Our data analyzing the expression of VDR mRNA

and also the functional activity of 1,25(OH)2D3 clearly show that the VDR is expressed by human peripheral B cells. Moreover, our data clearly demonstrate that VDR expression by B cells is regu-

lated both by 1,25(OH)2D3, as it is on other cells (5), but also by activation signals. These results indicate that vitamin D may exert differential effects on activated vs resting B cells and also may have different effects in individuals with different levels of serum

1,25(OH)2D3. Active 1,25(OH)2D3 is mainly degraded by 24-hydroxylase (CYP24A1) (2). Microarray analysis from human colon carcinoma and ovarian cancer cell lines revealed that CYP24A1 was the most

inducible gene responsive to 1,25(OH)2D3 (29, 30). In this report FIGURE 9. The effect of vitamin D (VitD) on cell cycle-regulated we show that 24-hydroxylase was significantly up-regulated fol- genes. RNAs were extracted from freshly isolated B cells and B cells lowing the incubation of human B cells with 1,25(OH)2D3.As cultured with anti-IgM, anti-CD40, and IL-21 with (VitD) or without (nil) opposed to the VDR, CYP24A1 was not altered by B cell activa-

1,25(OH)2D3 (10 nM) for 3 days. The mRNA expressions of p27, p21, and tion. These results further demonstrate that human B cells can p18 were determined by SuperArray and/or quantitative PCR. A, The rel- respond to 1,25(OH)2D3 directly. In addition, the results suggest ative expression of p21 and p27 determined by SuperArray. Data represent that the activity of vitamin D on B cells might be influenced not Ϯ the mean SEM of three independent experiments. B, The relative ex- only by VDR expression but also by the capacity to degrade the pression of p21, p27, and p18 as determined by quantitative PCR. Data active molecule. The increased susceptibility of activated B cells to represent the mean Ϯ SEM of three independent experiments. Significant many of the effects of vitamin D might therefore reflect the up- differences between vitamin D-treated and control B cells are shown .p Ͻ 0.05). regulation of VDR but not CYP24A1 by these cells ,ء) ␣ 25(OH)D3-1- -hydroxylase (CYP27B1), the enzyme responsi- ble for the final hydroxylation and activation of 25(OH)D3 into functional consequence, we added the precursor, 25(OH)D3,to 1,25(OH)2D3, is mainly found in the proximal convoluted tubule purified B cells stimulated with anti-CD40 and IL-21 and exam- cells of the kidney (2). Interestingly, we found that CYP27B1 ined the expression of CD38, plasma cell differentiation, and IgG mRNA was also expressed by resting B cells and could be further production. 25(OH)D3 increased CD38 expression and inhibited induced by stimulation, but not by 1,25(OH)2D3. Moreover, we plasma cell generation and IgG production in a concentration-de- found that the precursor, 25(OH)D3, had similar effects on purified pendent manner (Fig. 10B). The precursor had effects similar to B cells compared with the active form, albeit at higher concentra- those of the active form of vitamin D3, but at a 25-fold higher tions. Therefore, 25(OH)D3 might be metabolized to 1,25(OH)2D3 concentration. Because monocytes contaminating the peripheral B by B cells themselves and may represent a source for the extrarenal cell populations might contribute to these results, B cells that were synthesis of 1,25(OH)2D3, as is thought to be the case for macro- sorted from tonsils and lacked detectable monocytes/macrophage phages and dendritic cells (3, 4). In conditions such as SLE, in contamination were assessed and similar results were obtained which there is diffuse B cell activation (20), systemic vitamin D (data not shown). There results are consistent with the conclusion metabolism might therefore be significantly influenced. that human B cells can take up and metabolize 25(OH)D3 to active The capacity of 1,25(OH)2D3 to inhibit proliferation has been 1,25(OH)2D3 that can inhibit B cell function. reported in lymphocytes and in a variety of human cancer cell lines

(31). We showed that 1,25(OH)2D3 also exerted an inhibitory ef- Discussion fect on B cell proliferation. By using CFSE analysis, initial B cell

In this study we show that 1,25(OH)2D3 has potent direct effects division was not affected by 1,25(OH)2D3 whereas subsequent on- on B cell responses, inhibiting proliferation, generation of class- going proliferation was progressively inhibited. This effect coin- switched memory B cells, plasma cell differentiation, and Ig pro- cides with activation and 1,25(OH)2D3-mediated up-regulation of duction. Reduced levels of 1,25(OH)2D3 in SLE patients, partic- the VDR and suggests that a threshold level of VDR engagement ularly in those patients with high disease activity scores and might be required for the antiproliferative effect of the occupied

ANAs, suggest that vitamin D-dependent B cell regulation may VDR to become apparent. Notably, 1,25(OH)2D3-mediated inhi- play an important role in maintaining normal B cell homeostasis bition of proliferation was associated with apoptosis of the acti- and that decreased levels of 1,25(OH)2D3 may contribute to B cell vated and dividing B cells. These results suggest that differentia- hyperactivity in SLE. Finally, because activated B cells express tive events that require the initial expansion of B cells might be many of the molecules involved in vitamin D metabolism, their eliminated as a result of the 1,25(OH)2D3-mediated death of pro- characteristic hyperactivity in SLE could contribute to the de- liferating B cells. creased levels of 1,25(OH)2D3 found in these patients. The inhibition of proliferation and the induction of apoptosis are The varied functions of 1,25(OH)2D3 are mainly mediated likely to result in the significant reduction in plasma cell differen- through binding to the VDR. We clearly demonstrate that VDR tiation and Ig production that we observed. Previous studies have The Journal of Immunology 1645

FIGURE 10. B cells can convert inactive 25(OH)D3 into the active form of vitamin D. A, CYP27B1 is constitutively expressed by fresh B cells and is induced by stimulation. B cells were analyzed either immediately after isolation or were incubated for 3 days with the indicated stimuli with ␤ ϫ 6 (1,25(OH)2D3) or without (nil) 1,25(OH)2D3 (10 nM). The expression of CYP27B1 was determined by RT-PCR relative to 2M mRNA ( 10 ). The p Ͻ ,ء) means Ϯ SEM of one of three independent experiments are shown. Significant differences between vitamin D-treated and control B cells are shown p Ͻ 0.01) (␣IgM, anti-IgM; ␣CD40, anti-CD40). B, Purified B cells were ,ءء) Significant differences between day 0 and stimulated B cells are shown .(0.05

stimulated with anti-CD40 and IL-21 with either 10 nM 1,25(OH)2D3 or different concentrations of 25(OH)D3 ranging from 0 to 1000 nM. After 7 days in culture, the cells were washed and stained for expression of CD19, CD27, and CD38. The geometric (Geom.) mean fluorescence of CD38 and the high high frequency of plasma cells (CD27 CD38 ) were determined. The culture supernatants were analyzed for IgG production. The 25(OH)D3 dose response (solid line) is compared with 10 nM 1,25(OH)2D3 (dashed line).

shown that plasma cell differentiation requires initial B cell pro- that encodes a more stable and active protein (35), was also down-

liferation (32) and increases with the number of cell divisions (33). regulated by 1,25(OH)2D3. The inhibitory effect of 1,25(OH)2D3 We have previously shown that the combination of IL-21 and anti- on the expressions of XBP1 and ERN1 mRNA levels may explain CD40 stimulation with or without BCR crosslinking is a potent the greater inhibitory effect of vitamin D on Ig secretion and de- inducer of proliferation and plasma cell differentiation and that all tection of plasma cells by ELISPOT compared with its more mod- of the plasma cells generated had undergone extensive prolifera- est inhibition of the differentiation of phenotypically defined

tion (26). Using this system, we demonstrated that 1,25(OH)2D3 plasma cells. Although inhibition of the expressions of XBP1 and had inhibitory effects on plasma differentiation and Ig production. ERN1 mRNA by 1,25(OH)2D3 may contribute to a reduced Interestingly, this effect is not evident when the B cells are treated amount of Ig secretion per phenotypically defined plasma cell, it is

with 1,25(OH)2D3 after 5 days in culture, indicating that unlikely to explain the inhibition of the generation of both plasma 1,25(OH)2D3 inhibits the generation of plasma cells but not their cells and postswitched memory cells as well as the ongoing pro- subsequent persistence and lends support to the conclusion that the liferation of activated B cells by vitamin D.

inhibition of B cell proliferation by 1,25(OH)2D3 is responsible for Several important intracellular pathways have been reported to ␬ the reduction of Ig-secreting cells and Ig production. be inhibited by 1,25(OH)2D3, such as NF- B activation and cell Several transcription factors regulate different stages of B cell cycle progression. The suppressive effect of 1,25(OH)2D3 on the maturation and plasma cell differentiation (34). The most notable NF-␬B signaling pathway was previously observed in T cells, finding of the current study was that expression of most of these monocytes or macrophages (9, 36, 37), and could have affected the

transcriptional regulators was unaffected by 1,25(OH)2D3, includ- expression of various essential cell surface and secreted molecules. ing PAX5, BCL6, AID, BLIMP1, MTA3, and IRF4. These results The NF-␬B pathway, however, did not appear to be generally sup-

provide no support for the possibility that 1,25(OH)2D3 inhibits pressed by 1,25(OH)2D3 in activated B cells as assessed by the memory cell or plasma cell differentiation by directly affecting the analysis of expression of various genes influenced by this signaling

expression of any of these regulators of B cell maturation. How- pathway. Because 1,25(OH)2D3 directly inhibited the proliferation ever, the expression of XBP1 was modestly but significantly of activated B cells, we therefore examined whether vitamin D

down-regulated by 1,25(OH)2D3. In addition, mRNA for ERN1, might suppress the expression of a variety of cell cycle regulators which is required for processing XBP1 mRNA to a spliced form and found that 1,25(OH)2D3 increased the expression of p27 and 1646 1,25-DIHYDROXYVITAMIN D3 MODULATES B CELLS decreased the mRNA levels of CDK4 and 6 and cyclin D. Cell clusion that vitamin D is involved in maintaining normal B cell cycle progression is under the control of CDK, the activity of homeostasis and that vitamin D deficiency in SLE patients may which is dependent on the expression of specific CDK inhibitors. contribute to B cell hyperactivity, the breakdown of B cell toler-

Previous studies have shown that 1,25(OH)2D3 inhibits the prolif- ance, and the generation of some autoantibodies. eration of cancer cells by inducing gene transcription of the CDK Consistent with this, we also found that those SLE patients with inhibitors p21 and/or p27, which can inhibit cell cycle progression positive ANA titers had lower levels of 1,25(OH)2D3 than those at the G1/S transition, depending on the cell type (38–40). In vivo, without ANA. Interestingly, such a difference was not seen be- 1,25(OH)2D3 administration can effectively restore p27 accumu- tween the patients who were positive or negative for dsDNA Abs. lation in cancer cells as detected by and The persistent and therapy-resistant presence of ANA in SLE even reduce the tumor burden (41). Several mechanisms have been re- during the clinically quiescent phases has been explained by the ported to be involved, including increasing p27 transcription me- production of these autoantibodies from long-lived plasma cells diated by VDR-Sp1 interactions with the p27 promoter (42), sta- (52). In contrast, the presence of the dsDNA Ab is usually corre- bilization of the protein through VDR-induced reduction of CDK2 lated with SLE activity and decreases after immunosuppressive or activity and Skp2 abundance (the main factors responsible for p27 B cell-depletion therapy, suggesting that these autoantibodies orig- degradation), or induction of PTEN, a phosphatase that dephos- inate from short-lived dividing plasmablasts (20, 53). Thus, phorylates p27 (40). Other CDK inhibitors have been shown to 1,25(OH)2D3 may play a role in regulating the generation of long- play an essential role in B cell responses. In this regard, p18 has lived plasma cell and be less potent in suppressing the rapid pro- been shown to be required for B cells to terminate proliferation and duction of autoantibodies from dividing plasmablasts. The kinetics differentiate into functional plasma cells (38, 43, 44). Previous of the in vitro inhibition of B cell responses by 1,25(OH)2D3 is studies have confirmed that p18 and p27 were involved in regu- consistent with this possibility. lating different steps of B proliferation. Whereas p27 was thought Besides the inhibitory effect on plasma cell differentiation and Ig to regulate cell cycle entry by forming a ternary complex with production, our data also demonstrated that 1,25(OH)2D3 exerted CDKs and cyclins, p18 was directly involved in induction of the its inhibitory effects on memory B cell differentiation. Relative

G1 cell cycle arrest by blocking the association of CDK4 and increases in memory cells and plasma cells are both characteristic CDK6 with cyclin D (44). Interestingly, our data demonstrated that B cell abnormalities in lupus (54). In contrast to naive B cells and the mRNA level of p27, but not of p21 or p18, was up-regulated plasma cells, memory B cells are not targeted by conventional by 1,25(OH)2D3 in activated human B cells. These results suggest immunosuppressive therapies and may contribute to disease flares the possibility that 1,25(OH)2D3 could inhibit B cell proliferation (55). Thus, vitamin D therapy may offer a new opportunity to by up-regulating p27 and thereby inhibit the cell cycle entry of induce remission in SLE by preventing or decreasing the differ- previously cycling B cells. Because 1,25(OH)2D3 also decreased entiation of memory cells. This possibility is supported by murine mRNA levels of CDK4 and CDK6 as well as cyclin D, the effect studies in which administration of 1,25(OH)2D3 or its analog has of p27 on ongoing B cell proliferation might be more profound. been shown to decrease proteinuria or increase survival in a spon-

The 1,25(OH)2D3-mediated induction of p27 may limit ongoing B taneous mouse lupus model, MRL/lpr (56). cell proliferation and thereby play an important role in the control In conclusion, 1,25-dihydroxyvitamin D3 may play an impor- of B cell responses. These findings suggest the possibility that the tant role in the maintenance of B cell homeostasis. Correction of major effects of vitamin D on plasma cell and memory cell dif- vitamin D deficiency may be useful not only to prevent osteopenia ferentiation may result from the suppression of ongoing B cell but also in the suppression of B cell overactivity in autoimmune proliferation, which is required before these differentiation steps disorders such as SLE. can occur (32). Additional experiments will be necessary to test this hypothesis. Acknowledgments Decreased vitamin D concentrations have been reported in many We gratefully acknowledge the assistance of the Flow Cytometry Section autoimmune diseases such as multiple sclerosis, type 1 diabetes of the Office of Science and Technology, National Institute of Arthritis and mellitus, RA, fibromyalgia, and SLE (14, 15, 45, 46). We therefore Musculoskeletal and Skin Diseases (Bethesda, MD) for sorting B cell sub- examined the serum levels of 25(OH)D and 1,25(OH)2D3 in Chi- sets. We are also grateful to Dr. John A. Hardin for many stimulating nese lupus patients. Consistent with some previous findings, both discussions and to Dr. Nancy Longo for reviewing the manuscript.

25(OH)D and 1,25(OH)2D3 levels were significantly lower in pa- tients with SLE compared with healthy donors (14, 46). However, Disclosures another study reported significantly lower 25(OH)D3 but not The authors have no financial conflict of interest. 1,25(OH)2D3 levels in SLE patients (47). The explanation for the differences in these studies is not clear but could relate to ethnicity References or disease duration (46). Although previous studies suggested a 1. van Etten, E., and C. Mathieu. 2005. Immunoregulation by 1,25-dihydroxyvita- possible relationship between depressed levels of 25(OH)D3 and min D-3: Basic concepts. J. Steroid Biochem. Mol. Biol. 97: 93–101. the presence of lupus renal disease (46), we were able to demon- 2. Jones, G., S. A. Strugnell, and H. F. DeLuca. 1998. Current understanding of the strate that lower levels of 1,25(OH) D significantly correlated molecular actions of vitamin D. Physiol. Rev. 78: 1193–1231. 2 3 3. Overbergh, L., B. Decallonne, D. Valckx, A. Verstuyf, J. Depovere, J. Laureys, with increased disease activity. Moreover, significantly decreased O. Rutgeerts, R. Saint-Arnaud, R. Bouillon, and C. Mathieu. 2000. Identification vitamin D levels were detected in 12 newly diagnosed patients and immune regulation of 25-hydroxyvitamin D-1-␣-hydroxylase in murine mac- before any treatment was administered, indicating that the defi- rophages. Clin. Exp. Immunol. 120: 139–146. 4. Hewison, M., L. Freeman, S. V. Hughes, K. N. Evans, R. Bland, A. G. ciency may be present at the onset of lupus and possibly before. Eliopoulos, M. D. Kilby, P. A. H. Moss, and R. Chakraverty. 2003. Differential This could explain, in part, why African Americans who are often regulation of and its ligand in human monocyte-derived den- dritic cells. J. Immunol. 170: 5382–5390. chronically vitamin D deficient (46, 48, 49) also have a higher risk 5. Haussler, M. R., G. K. Whitfield, C. A. Haussler, J. C. Hsieh, P. D. Thompson, and severity of SLE (50). It is known that SLE may be preceded S. H. Selznick, C. E. Dominguez, and P. W. Jurutka. 1998. The nuclear vitamin by autoantibody production that may exist for many years before D receptor: Biological and molecular regulatory properties revealed. J. Bone Miner. Res. 13: 325–349. diagnosis (51), and it is possible that vitamin D deficiency may 6. Mathieu, C., and L. Adorini. 2002. The coming of age of 1,25-dihydroxyvitamin contribute to this. The current results are consistent with the con- D-3 analogs as immunomodulatory agents. Trends Mol. Med. 8: 174–179. The Journal of Immunology 1647

7. Penna, G., and L. Adorini. 2000. 1␣,25-dihydroxyvitamin D-3 inhibits differen- 32. Vernino, L., L. M. McAnally, J. Ramberg, and P. E. Lipsky. 1992. Generation of tiation, maturation, activation, and survival of dendritic cells leading to impaired nondividing high-rate Ig-secreting plasma-cells in cultures of human B-cells alloreactive T cell activation. J. Immunol. 164: 2405–2411. stimulated with anti-Cd3-activated T-cells. J. Immunol. 148: 404–410. 8. van Halteren, A. G. S., E. van Etten, E. C. de Jong, R. Bouillon, B. O. Roep, and 33. Tangye, S. G., and P. D. Hodgkin. 2004. Divide and conquer: the importance of C. Mathieu. 2002. Redirection of human autoreactive T-cells upon interaction cell division in regulating B-cell responses. Immunology 112: 509–520. with dendritic cells modulated by TX527, an analog of 1,25 dihydroxyvitamin 34. Shapiro-Shelef, M., and K. Calame. 2005. Regulation of plasma-cell develop- D-3. Diabetes 51: 2119–2125. ment. Nat. Rev. Immunol. 5: 230–242. 9. D’Ambrosio, D., M. Cippitelli, M. G. Cocciolo, D. Mazzeo, P. Di Lucia, R. Lang, 35. Calfon, M., H. Q. Zeng, F. Urano, J. H. Till, S. R. Hubbard, H. P. Harding, F. Sinigaglia, and P. Panina-Bordignon. 1998. Inhibition of IL-12 production by S. G. Clark, and D. Ron. 2002. IRE1 couples endoplasmic reticulum load to 1,25-dihydroxyvitamin D-3: involvement of NF-␬B downregulation in transcrip- secretory capacity by processing the XBP-1 mRNA. Nature 415: 92–96. tional repression of the p40 gene. J. Clin. Invest. 101: 252–262. 36. Cohen-Lahav, M., S. Shany, D. Tobvin, C. Chaimovitz, and A. Douvdevani. 10. Boonstra, A., F. J. Barrat, C. Crain, V. L. Heath, H. F. J. Savelkoul and A. 2006. Vitamin D decreases NF-␬B activity by increasing I ␬B ␣ levels. Nephrol. O’Garra. 2001. 1␣,25-dihydroxyvitamin D3 has a direct effect on naive CD4ϩ T Dial. Transplant. 21: 889–897. cells to enhance the development of Th2 cells. J. Immunol. 167: 4974–4980. 37. Giarratana, N., G. Penna, S. Amuchastegui, R. Mariani, K. C. Daniel, and 11. Cippitelli, M., C. Fionda, D. Di Bona, F. Di Rosa, A. Lupo, M. Piccoli, L. Frati, L. Adorini. 2004. A vitamin D analog down-regulates proinflammatory chemo- and A. Santoni. 2002. Negative regulation of CD95 ligand gene expression by kine production by pancreatic islets inhibiting T cell recruitment and type 1 vitamin D3 in T lymphocytes. J. Immunol. 168: 1154–1166. diabetes development. J. Immunol. 173: 2280–2287. 12. Muller, K., C. Heilmann, L. K. Poulsen, T. Barington, and K. Bendtzen. 1991. 38. Munker, R., T. Kobayashi, E. Elstner, A. W. Norman, M. Uskokovic, W. Zhang, The role of monocytes and T-cells in 1,25-dihydroxyvitamin-D3 mediated inhi- M. Andreeff, and H. P. Koeffler. 1996. A new series of vitamin D analogs is bition of B-cell function-in vitro. Immunopharmacology 21: 121–128. highly active for clonal inhibition, differentiation, and induction of WAF1 in myeloid leukemia. Blood 88: 2201–2209. 13. Heine, G., K. Anton, B. M. Henz, and M. Worm. 2002. 1␣,25-dihydroxyvitamin 39. Wang, Q. M., F. Chen, X. Luo, D. C. Moore, M. Flanagan, and G. P. Studzinski. D inhibits anti-CD40 plus IL-4-mediated IgE production in vitro. Eur. J. Im- 3 1998. Lowering of p27(Kip1) levels by its antisense or by development of re- munol. 32: 3395–3404. sistance to 1,25-dihydroxyvitamin D-3 reverses the G1 block but not differenti- 14. Huisman, A. M., K. P. White, A. Algra, M. Harth, R. Vieth, J. W. Jacobs, ation of HL60 cells. Leukemia 12: 1256–1265. J. W. Bijlsma, and D. A. Bell. 2001. Vitamin D levels in women with systemic 40. Liu, W., S. L. Asa, and S. Ezzat. 2002. Vitamin D and its analog EB1089 induce lupus erythematosus and fibromyalgia. J. Rheumatol. 28: 2535–2539. p27 accumulation and diminish association of p27 with Skp2 independent of 15. Becker, A., R. Fischer, and M. Schneider. 2001. Bone density and 25-OH-vitamin PTEN in pituitary corticotroph cells. Brain Pathol. 12: 412–419. D in systemic lupus erythematosus. Z. Rheumatol. 60: 352–358. 41. Dackiw, A. P. B., S. Ezzat, P. Huang, W. Liu, and S. L. Asa. 2004. Vitamin D3 16. Gladman, D. D., M. B. Urowitz, J. M. Esdaile, B. H. Hahn, J. Klippel, R. Lahita, administration induces nuclear p27 accumulation, restores differentiation, and M. H. Liang, P. Schur, M. Petri, and D. Wallace. 1999. Guidelines for referral and reduces tumor burden in a mouse model of metastatic follicular thyroid cancer. management of systemic lupus erythematosus in adults. Arthritis Rheum. 42: Endocrinology 145: 5840–5846. 1785–1796. 42. Li, P. F., C. R. Li, X. H. Zhao, X. H. Zhang, S. V. Nicosia, and W. L. Bai. 2004. 17. Anolik, J. H., and M. Aringer. 2005. New treatments for SLE: cell-depleting and P27(Kip1) stabilization and G(1) arrest by 1,25-dihydroxyvitamin D-3 in ovarian anti-cytokine therapies. Best Pract. Res. Clin. Rheumatol. 19: 859–878. cancer cells mediated through down-regulation of cyclin E/cyclin-dependent ki- 18. Looney, R. J., J. H. Anolik, D. Campbell, R. E. Felgar, F. Young, L. J. Arend, nase 2 and Skp1-Cullin-F-box Protein/Skp2 ubiquitin ligase. J. Biol. Chem. 279: J. A. Sloand, J. Rosenblatt, and I. Sanz. 2004. B cell depletion as a novel treat- 25260–25267. ment for systemic lupus erythematosus: a phase I/II dose-escalation trial of rit- 43. Morse, L., D. Q. Chen, D. Franklin, Y. Xiong, and S. ChenKiang. 1997. Induc- uximab. Arthritis Rheum. 50: 2580–2589. tion of cell cycle arrest and B cell terminal differentiation by CDK inhibitor 19. Van den Bergh, B., D. Selleslag, J. R. Boelaert, E. G. Matthys, M. Schurgers, p18(INK4c) and IL-6. Immunity 6: 47–56. S. Vandecasteele, and A. De Vriese. 2005. Management of therapy-resistant sys- 44. Schrantz, N., G. E. Beney, M. T. Auffredou, M. F. Bourgeade, G. Leca, G., and temic lupus erythematosus with rituximab: Report of a case and review of the A. Vazquez. 2000. The expression of p18(INK4) and p27(kip1) cyclin-dependent literature. Acta Clin. Belg. 60: 102–105. kinase inhibitors is regulated differently during human B cell differentiation. 20. Grammer, A. C., and P. E. Lipsky. 2003. B cell abnormalities in systemic lupus J. Immunol. 165: 4346–4352. erythematosus. Arthritis Res. Ther. 5: S22–S27. 45. Grant, W. B. 2006. Epidemiology of disease risks in relation to vitamin D in- 21. Hochberg, M. C. 1997. Updating the American College of Rheumatology revised sufficiency. Prog. Biophys. Mol. Biol. 92: 65–79. criteria for the classification of systemic lupus erythematosus. Arthritis Rheum. 46. Kamen, D. L., G. S. Cooper, H. Bouali, S. R. Shaftman, B. W. Hollis, and 40: 1725–1725. G. S. Gilkeson. 2006. Vitamin D deficiency in systemic lupus erythematosus. 22. Bombardier, C., D. D. Gladman, M. B. Urowitz, D. Caron, and C. H. Chang. Autoimmun. Rev. 5: 114–117. 1992. Derivation of the SLEDAI: a disease-activity index for lupus patients. 47. Muller, K., N. J. Kriegbaum, B. Baslund, O. H. Sorenson, M. Thymann, and Arthritis Rheum. 35: 630–640. K. Bentzen. 1995. Vitamin D metabolism in patients with rheumatic diseases: 23. Stoeckler, J. D., H. A. Stoeckler, N. Kouttab, and A. L. Maizel. 1996. 1 ␣,25- low serum levels of 25-hydroxyvitamin D3 in patients with systemic lupus ery- dihydroxyvitamin D-3 modulates CD38 expression on human lymphocytes. thematosus. Clin. Rheumatol. 14: 397–400. J. Immunol. 157: 4908–4917. 48. Harris, S. S. 2006. Vitamin D and African Americans. J. Nutr. 136: 1126–1129. 24. Morgan, J. W., N. Kouttab, D. Ford, and A. L. Maizel. 2000. Vitamin D-mediated 49. Nesby-O’Dell, S., K. S. Scanlon, M. E. Cogswell, C. Gillespie, B. W. Hollis, gene regulation in phenotypically defined human B cell subpopulations. Endo- A. C. Looker, C. Allen, C. Doughertly, E. W. Gunter, and B. A. Bowman. 2002. crinology 141: 3225–3234. Hypovitaminosis D prevalence and determinants among African American and white women of reproductive age: third National Health and Nutrition Exami- 25. Joshua, D., B. Pope, R. Brown, L. Barrow, A. Murray, and J. Gibson. 2002. nation Survey, 1988–1994. Am. J. Clin. Nutr. 76: 187–192. Phenotyping primitive plasma cells. Br. J. Haematol. 117: 252–253. 50. McCarty, D. J., S. Manzi, T. A. Medsger, R. Ramseygoldman, R. E. Laporte, and 26. Ettinger, R., G. P. Sims, A. M. Fairhurst, R. Robbins, Y. S. da Silva, R. Spolski, C. K. Kwoh. 1995. Incidence of systemic lupus-erythematosus: race and gender W. J. Leonard, and P. E. Lipsky. 2005. IL-21 induces differentiation of human differences. Arthritis Rheum. 38: 1260–1270. naive and memory B cells into antibody-secreting plasma cells. J. Immunol. 175: 51. James, J. A., M. T. McClain, M. R. Arbuckle, and J. B. Harley. 2004. Autoan- 7867–7879. tibodies before the clinical onset of systemic lupus erythematosus: the authors 27. Matthias, P., and A. G. Rolink. 2005. Transcriptional networks in developing and reply. New Engl. J. Med. 350: 305–305. mature B cells. Nat. Rev. Immunol. 5: 497–508. 52. Manz, R. A., S. Arce, G. Cassese, A. E. Hauser, F. Hiepe, and A. Radbruch. 2002. 28. Hickish, T., D. Cunningham, K. Colston, B. C. Millar, J. Sandle, A. G. Mackay, Humoral immunity and long-lived plasma cells. Curr. Opin. Immunol. 14: M. Soukop, and J. Sloane. 1993. The effect of 1,25-dihydroxyvitamin-D(3) on 517–521. lymphoma cell-lines and expression of vitamin-D receptor in lymphoma. Br. J. 53. Anolik, J. H., D. Campbell, R. Felgar, J. Rosenblatt, F. Young, and R. J. Looney. Cancer 68: 668–672. 2002. B lymphocyte depletion in the treatment of systemic lupus (SLE): phase 29. Zhang, X. H., P. F. Li, J. Y. Bao, S. V. Nicosia, H. G. Wang, S. A. Enkemann, I/II trial of Rituximab (Rituxan (R)) in SLE. Arthritis Rheum. 46: S289. and W. L. Bai. 2005. Suppression of death receptor-mediated apoptosis by 1,25- 54. Renaudineau, Y., J. O. Pers, B. Bendaoud, C. Jamin, and P. Youinou. 2004. dihydroxyvitamin D-3 revealed by microarray analysis. J. Biol. Chem. 280: Dysfunctional B cells in systemic lupus erythematosus. Autoimmun. Rev. 3: 35458–35468. 516–523. 30. Wood, R. J., L. Tchack, G. Angelo, R. E. Pratt, and L. A. Sonna. 2004. DNA 55. Dorner, T., and P. E. Lipsky. 2004. Correlation of circulating CD27high plasma microarray analysis of vitamin D-induced gene expression in a human colon cells and disease activity in systemic lupus erythematosus. Lupus 13: 283–289. carcinoma cell line. Physiol. Genomics 17: 122–129. 56. Vaisberg, M. W., R. Kaneno, M. F. Franco, and N. F. Mendes. 2000. Influence 31. Dusso, A. S., A. J. Brown, and E. Slatopolsky. 2005. Vitamin D. Am. J. Physiol. of cholecalciferol (vitamin D3) on the course of experimental systemic lupus ϫ 289: F8–F28. erythematosus in F1 (NZB W) mice. J. Clin. Lab. Anal. 14: 91–96.