Systemic Lupus Erythematosus (SLE) is an autoimmune disease predominantly affecting women of child-bearing age at a rate of 20 to 150 cases per 100,000 population [1
]. Genetic, environmental, and hormonal factors all contribute to SLE pathogenesis, and no single factor elicits disease if encountered alone. Different combinations of risk factors lead to variable and unique patient presentations, including photosensitivity, rash, arthritis, cytopenias, serositis, nephritis, fatigue, and psychosis [1
]. While the pathophysiology of SLE is incompletely understood, it is characterized by aberrant T and B cell activity, elevated autoantibody titers, and subsequent organ damage. The characteristic presence of autoantibodies (e.g., anti-dsDNA IgG, anti-chromatin IgG, anti-Sm IgG), support a pathologic role for B cells in SLE. Additionally, autoantibody titers have been shown to increase with disease activity, and newer immunotherapies (e.g., rituximab, belimumab) that target B cells have demonstrated some benefit [2
Vitamin D deficiency is a common laboratory finding in many autoimmune diseases including SLE [4
]. Vitamin D deficiency is inversely correlated with lupus disease activity, and mutations in the vitamin D receptor (VDR) have been identified in SLE populations [5
]. A couple of studies have investigated the relationship between vitamin D3 supplementation and SLE in animal models, demonstrating an overall protective effect ([6
] and reviewed in [9
]). Although evidence from these studies suggests that vitamin D3 plays a protective role in lupus, clinical supplementation studies in SLE patients have demonstrated inconclusive therapeutic benefits [10
]. Thus, whether supplementation or deficiency of vitamin D3 contributes to disease prevention or susceptibility, respectively, and the mechanism(s) by which such regulation would occur remains unknown.
Immune cells (monocytes, dendritic cells, macrophages, activated lymphocytes) express VDR and 1α-hydroxylase, the key enzyme that produces the biologically active form of vitamin D3 (1,25(OH)2
D3) from its precursor 25(OH)D3. Vitamin D3 has immune-modulating properties, such as an ability to inhibit Th1 and Th17 differentiation, while promoting Treg development [14
]. Additionally, vitamin D3 has been shown to inhibit B cell proliferation [17
], and differentiation into memory B cells and antibody-secreting plasma cells [20
]. There is also a reduction in immunoglobulin production [21
], suggesting a role for vitamin D3 in regulating differentiation and/or activity of downstream memory B cells or plasma cells. The specific role of vitamin D3 in B cell differentiation and autoantibody secretion in animal models of SLE, however, has not been determined.
The Act1 (TRAF3IP2 or CIKS) knockout mouse is a model for SLE and Sjogren’s syndrome [23
]. Act1 is a key adaptor protein in IL-17R signaling, as well as a negative regulator of BAFF-BAFFR and CD40-CD40L signaling in B cells [23
]. The Act1-/-
phenotype is associated with increased Th17 cells and Th17-related cytokines (IL-17A, IL-21, IL-22), expanded peripheral B cell populations, hypergammaglobulinemia and autoantibody production, as well as splenomegaly, lymphadenopathy, and mild nephritis. Given the known inhibitory activity of vitamin D3 on both Th17 cells and B cell differentiation and immunoglobulin production, we hypothesized that the absence of dietary vitamin D3 would promote disease development while vitamin D3 supplementation would suppress disease development. To test this hypothesis, we assessed whether 9 weeks of vitamin D3 restriction or supplementation was sufficient to alter the Act1-/-
autoimmune phenotype; specifically, the development of SLE-like characteristics. We found that dietary vitamin D3 restriction was associated with increased autoantibody and immunoglobulin production, as well as increased percentages of splenic memory B cells, while vitamin D3 supplementation had no significant effect on autoantibody levels and B cell differentiation patterns. Further studies in SLE patients confirmed a negative correlation between levels of memory B cells and vitamin D3, supporting the pathogenicity of vitamin D3 deficiency.
2. Materials and Methods
2.1. Patient Enrollment
SLE patients seen by a rheumatologist (H.S.) between July 2017 and June 2018 at the Cleveland Clinic Department of Rheumatologic and Immunologic Disease (ages 18–80) were invited to volunteer for an ongoing Lupus Registry. Patients were eligible for inclusion if ACR criteria were met. There were no exclusions based on disease activity, flares, or type of therapy. Demographic information, medical history, and relevant clinical data were collected and managed using REDCap electronic data capture tools hosted at the Cleveland Clinic [27
]. At this visit, patients provided blood samples that were processed for serum and peripheral blood mononuclear cells (PBMCs) and frozen at −80 °C, until later processing of all samples, concurrently. Fourteen random patient samples were selected for PBMC B cell analysis as described below. All samples were obtained after patients provided written informed consent and after approval of the study by the Cleveland Clinic Institutional Review Board.
2.2. Vitamin D3 Dietary Intervention and Animal Care
Act1-/- mice on the Balb/c background were bred at the University of North Carolina Gnotobiotic center, and transferred to specific pathogen-free housing at Lerner Research Institute at 5–7 weeks of age. All mice were born within 3 weeks of each other. Immediately upon arrival at the Lerner Research Institute, the mice were divided into 3 dietary treatment groups—0 IU/kg (low), 2 IU/kg (normal), or 10 IU/kg (high) of vitamin D3/kg chow. The three treatment groups were matched for age and sex to limit potential biases. Mice (n = 15) were maintained on their assigned diet for the duration of this 9-week study. Mice were bled for serum at 0, 3, 6, and 9 weeks post-transfer. Due to the immunodeficiency status of Act1-/- mice, cages were changed twice weekly and Vetropolycin gel was applied as needed throughout the experiment. Mice had access to hydrogel to prevent dehydration if necessary. All animal procedures were approved by the Cleveland Clinic Institutional Animal Care and Use Committee.
2.3. Organ Harvest/Preparation
Tissue samples were both frozen in OCT™ and prepared for paraffin embedding by 24 h fixation in 10% formalin, followed by 80% ethanol. Spleen, submaxillary gland, and cervical lymph nodes were weighed prior to preservation. Kidneys were cut in half longitudinally prior to preservation. Paraffin embedding, sectioning (5 µm), and hematoxylin and eosin staining were performed by the Lerner Research Institute Histology Core. Periodic Acid Schiff (PAS) staining was performed with the Richard–Allan scientific PAS stain kit (Thermo Scientific, Waltham, MA, USA).
Histology measurements were performed on H&E and PAS-stained paraffin-embedded kidney sections. To quantify the mean glomerular area, 8–15 glomeruli per mouse were traced and measured for area using the Keyence BZ-X700 All-in-one microscope, then averaged for each mouse. A renal pathologist (J.N.), blinded to the treatment groups, scored H&E and PAS-stained kidneys for the presence of endocapillary hypercellularity, extracapillary proliferation, immune deposits, tubular atrophy, tubular casts, tubular dilation, and interstitial fibrosis and inflammation. A scale of 0–5 was used for each feature analyzed (8), in which 0 is absent, 1 is 1–5%, 2 is 6–13%, 3 is 11–20%, 4 is 21–50%, and 5 is greater than 50% of the glomeruli/tubules/area of interest, summing to a maximum possible score of 40.
2.4. Immunofluorescence Staining of Kidney Tissue
Half kidneys were immediately frozen in OCT™ and subsequently sectioned at 5 µm. Following a brief acetone fixation, sections were stained with anti-IgG or anti-IgM-TexasRed (Southern Biotech, Birmingham, AL, USA) and anti-C’3-FITC antibodies (Immunology Consultants Laboratory, Inc, Portland, OR, USA). Images were obtained using the Keyence BZ-X700 Series at a fixed fluorescence intensity across slides, which was selected to minimize background fluorescence. Mean integrated fluorescence area for IgG- or IgM-TexasRed was calculated per glomerulus and averaged for each mouse. A minimum of 5 glomeruli was evaluated per mouse.
2.5. Serum Cytokine Flow Cytometry
Serum was obtained after 9 weeks of treatment and assessed for selected cytokines by a Cytokine Bead Array (BD Bioscience). Briefly, four Flex Sets containing fluorescent beads with pre-conjugated antibodies to each cytokine of interest (IL-1α, IL-6, IL-10, IL-17A) and a fluorescent detection agent were combined and incubated with 50 µl of serum according to the manufacturer’s protocol. Beads with bound cytokines were run on an LSR flow cytometer (BD Biosciences, San Jose, CA, USA) with the assistance of the Lerner Research Institute Flow Cytometry Core. Data were analyzed using FlowJo v10 software.
2.6. Enzyme-Linked Immunosorbent Assay (ELISA)
Serum from mice treated for 3–9 weeks was used for all ELISA experiments. Mouse 1,25(OH)2
D3 and 25(OH)D3 levels were measured by ELISA (MyBioSource and Eagle Bioscience, USA, respectively). Anti-dsDNA IgG and anti-SSB IgG ELISAs were run according to the manufacturer’s guidelines (Alpha Diagnostic International, San Antonio, TX, USA). Immunoglobulin ELISAs were performed as described previously [28
]. Briefly, microtiter plates were coated with 2.5µg/mL of unlabeled goat anti-mouse Ig (Southern Biotech) in PBS and blocked with PBS-gelatin. Mouse serum was diluted in assay buffer (5 mg/mL bovine γ-globulin, 5% gelatin, 0.05% Tween-20 in PBS) as follows: IgG (1:200,000), IgG1 (1:100,000), IgG2a (1:100,000), IgG2b (1:200,000), IgG3 (1:75,000), IgA (1:50,000), IgE (1:200,000), IgM (1:50,000). Samples were added to plates and incubated at room temperature for 90 min, then washed with PBS; secondary HRP-conjugated goat anti-mouse antibodies specific for IgG, IgG1, IgG2a, IgG2b, IgG3, IgA, IgE, IgM (all Southern Biotech) were added, followed by incubation for 1 h. Assays were visualized using the TMB substrate kit (Thermo Scientific, Waltham, MA, USA) and read on a Victor 3 plate reader (Perkin Elmer, Waltham, MA, USA) at 450 nm.
2.7. Flow Cytometry
At the time of sacrifice (9 weeks of treatment), spleens were isolated. Approximately 50% of each spleen was mashed and red blood cells lysed in ACK buffer (0.15 M NH4Cl, 0.01 M KHCO3, 0.2 mM EDTA) for 5 min. Cells were washed and resuspended in PBS for surface staining against CD3 (145-2C11, PE-Cy7), CD4 (GK1.5, APC), CD8 (53-6.7, PerCP-Cy5.5), B220 (RA3-6B2, PerCP-Cy5.5), CD93/AA4.1 (AA4.1, PE), CD38 (90, FITC), GL7 (GL-7, Biotin), Streptavidin (PE), and/or CD138 (281-2, PE) (all antibodies from eBioscience, Waltham, MA, USA) for 30 min. Samples stained only for surface markers were fixed in 1% paraformaldehyde before analysis. A subset of splenocytes was analyzed for cytokine production after subsequent stimulation with PMA (2 μg/mL) and ionomycin (1 μg/mL) for 4 h at 37 °C, 5% CO2, the last 2 h in the presence of 0.4 μL/100 μL cells GolgistopTM (BD Bioscience). Intracellular staining was performed with fluorescence-conjugated antibodies against RORγt, IL-17A, IL-21, and IL-22 (all from eBioscience). All intracellular stains were performed using the Foxp3 transcription factor staining buffer kit (eBioscience) and run within 15 h of permeabilization on the BD LSR Fortessa (BD Biosciences, CA, USA).
Human PBMCs were isolated from SLE patients by Ficoll gradient and stored at −80 °C. On the day of analysis, PBMCs were thawed, washed and resuspended in human Fc Block (BD Bioscience), followed by surface staining against CD19 (HIB19, FITC), CD27 (0323, APC), CD38 (HIT2, Pe-Cy7), and IgD (IA6-2, PE) (all antibodies from eBioscience) in PBS for 30 min. PBMCs were fixed in 1% paraformaldehyde before being run on the BD LSR Fortessa. All data from human and mouse experiments were analyzed using Flowjo v10 software.
2.8. Real-Time Reverse-Transcriptase PCR
Splenocytes were isolated from whole spleen as described above and immediately frozen at −80 °C. RNA was isolated using the RNAeasy Plus Micro Kit (Qiagen, Valencia, CA) and converted into cDNA using the qScript cDNA SuperMix (Quanta BioSciences, Gaithersburg, MD, USA). PCR was subsequently performed using 100 ng cDNA and the PerfeCTa® SYBR® Green FastMix® ROX (Quanta BioSciences) in the 7300 Real-Time PCR System (Applied Biosystems, Foster City, CA, USA). β-actin was used as the internal control for all RT-PCR experiments. All primers were created by Integrated DNA Technologies (Skokie, IL). PCR analyses were performed in triplicate, or duplicate for β-actin. Primer sequences were as follows: Aicda 5′-GCGGACATTTTTGAAATGGTA-3′ and 3′-GGGAGTTTCAGAATCCGGTT-5′; Rgs13 5′-ATCTACATCCAGCCACAGTCTC-3′ and 3′- TCAGGAGTTGTTGGTACATTTCAG-5′; Il6ra 5′-ACACACTGGTTCTGAGGGAC-3′ and 3′-GTGGACGGTTGGAACACCAT-5′; Itga4 5′-TGTGCAAATGTACACTCTCTTCCA-3′ and 3′-CTCCCTCAAGATGATAAGTTGTTCAA-5′; Actb 5′-TGGGAATGGGTCAGAAGGAC-3′ and 3′-GGGTCTAGTACAAACTCTGG-5′.
2.9. Statistical Analyses
Comparisons between treatment groups were analyzed by Student’s t-test with Welch’s correction for non-normally distributed samples. Two-way ANOVA was used to test for statistical differences in vitamin D3 levels over time in study animals. Associations between vitamin D3 and PBMC isolated cells were analyzed by linear regression, with reported r2 values after checking for linearity and error of variance. Statistical analyses and graphs were performed in GraphPad Prism (La Jolla, CA, USA) or JMP Pro version 14 (SAS Institute, Cary, NC, USA). Significance level was set at alpha = 0.05 for all analyses.
Persistent in the autoimmune literature is the lack of clarity regarding the role of vitamin D3 in autoimmune disease. Despite pervasive vitamin D deficiency across autoimmune diseases, there has been little consensus on a mechanism explaining how vitamin D deficiency relates to disease features, and few studies have explored these mechanisms in animal models of SLE [6
]. In this study, we explored the relationship between low, normal and high vitamin D3 exposure and lupus-like disease in the Act1-/-
mouse and in a cohort of well-controlled SLE patients. Interestingly, we found that low vitamin D3 levels specifically affected memory B cells and, at least in the mouse model, antibody production.
Spontaneous autoimmune models often require time to develop features of autoimmunity. Despite a short experimental timeframe of 9 weeks, we detected significant elevations in anti-dsDNA IgG antibody levels and total immunoglobulin levels (total IgG and IgG3) in mice fed a vitamin D3-restricted diet compared to mice fed diets containing normal or high vitamin D3. Levels of IgA, IgG1 and IgG2b did not reach statistical significance, but nevertheless trended in a similar fashion, while serum IgM levels showed a reciprocal pattern. Interestingly, we observed a significant decrease in serum IL-10 in the high vitamin D3 group, and IL-10 has been shown to promote IgG1 and IgG3 production by human B cells [47
]. Higher levels of IL-10 in the low and normal vitamin D3 groups might therefore contribute to the observed elevations in IgG1 and IgG3. If so, it is interesting that vitamin D3 treatment in this mouse model led to a reduction in IL-10, thus suggesting that IL-10 plays a pathogenic rather than anti-inflammatory role in the Act1-/-
mouse. A similar relationship has been reported in both mouse models of lupus and SLE patients [48
]. Additionally, the low number of mice in each of the treatment groups must be considered when interpreting these data.
Anti-dsDNA IgG is a strong marker of lupus-like disease and was highly elevated in Act1-/-
mice fed a low vitamin D3 diet. However, no differences were observed between the normal and high vitamin D3 groups. This finding suggests that a lack of vitamin D3 has a greater impact on disease than high vitamin D3 supplementation. In fact, the majority of our data suggest that high amounts of vitamin D3 fail to provide additional benefit compared to an adequate amount of vitamin D3. This observation is somewhat contradictory to the available data from vitamin D3-manipulated MRL-lpr/lpr lupus-prone mice [6
]. One explanation may be that these studies tested the efficacy of active vitamin D3 analogs (i.e., 1,25(OH)2
D3), as an intervention, while our study was aimed at evaluating the effect of 25(OH)D3, as this is the form frequently used in clinical studies [53
]. Although vitamin D3 levels were significantly different and stable already after 3 weeks of treatment, providing the animals with steady-state levels of vitamin D3 for six weeks, it is also possible that constitutive high vitamin D3 levels are required beyond the limited disease course of 9 weeks. As such, a longer interventional period might have also elicited measurable differences between the normal and high vitamin D3 groups. The fact that the Act1-/-
mice studied here did not display the overt splenomegaly or lymphadenopathy commonly seen in Act1-/-
] also supports this explanation (data not shown).
mouse has an intrinsic drive for Th17 development due to unregulated STAT3 activity leading to Th17 differentiation [34
]. Despite support in the literature for a role for vitamin D3 in reducing Th17 cell differentiation [14
], we did not observe any changes in the splenic Th17 cell population, serum IL-17A, or stimulated production of IL-17A. It is possible that the mechanism of vitamin D3-mediated inhibition of Th17 differentiation occurs upstream of STAT3, which might explain why no differences in Th17 cells were observed via staining for RORγt. Alternatively, the intrinsic drive towards Th17 development might have a greater influence on Th17 differentiation than vitamin D3. Further studies are needed to determine the exact target of vitamin D3 in Th17 differentiation and function.
Given the correlation between autoantibody titer and SLE disease activity, it is reasonable to infer that vitamin D3′s association with SLE may be attributable to its role in the regulation of splenic B cell differentiation. In fact, we found the B cell compartment to be the primary target of vitamin D3 restriction in the Act1-/-
mouse. Vitamin D3 restriction specifically promoted memory B cells, without affecting germinal center B cells or plasmablasts/plasma cells. Increased Aicda, Bcl6
, and Itga4
expression in the low vitamin D3 group further supported these data. Consistent with these data, vitamin D3 supplementation demonstrated transcriptional regulation of VDR in the germinal center and inhibition of Aicda
expression (encoding AID, activation-induced deaminase), a cytidine deaminase critical for class-switch recombination and somatic hypermutation [56
]. While expression of Bcl6
is not limited to B cells, it is interesting to note that the protein is also found in T follicular helper cells and memory T cells [42
], and thus reduced expression corresponds with reduced humoral immune activation in general. The relationship between memory B cells and vitamin D3 was further supported by a negative correlation between serum 25(OH)D3 and memory B cells, but not total B cells and plasmablasts, in PBMC samples from SLE patients. Overall, these observations are consistent with early studies on B cell development that showed inhibition of B cell activation and differentiation by vitamin D3 metabolites and analogs [17
It is puzzling that the observed difference in memory B cells is not mimicked in the populations of plasmablasts and plasma cells, as would be expected if the memory B cells were a result of T cell-dependent germinal center activity. T cell-independent memory B cells (B1b cells) have previously been identified as a population of B cells primarily located in the peritoneal and pleural cavities with the capacity to develop memory phenotypes and produce immunoglobulins without T cell interactions, thereby bypassing the germinal center [57
]. Therefore, the isolated effect observed on memory B cells could be a consequence of the development of T cell-independent memory B cells. This possibility is further supported by the high level of IgG3 identified in the low vitamin D3 group, as it has been shown that in particular IgG3 is produced in response to T cell-independent type 2 antigens [58
]. Unfortunately, there are surprisingly few data in the clinical literature regarding memory B cells and vitamin D3 levels in SLE patients, though evidence from multiple sclerosis (MS) studies has shown an association between lower intrathecal 25(OH)D3 and increased class-switched memory B cell accumulation in the CSF during MS relapses [19
Finally, the origin of vitamin D deficiency in SLE is unknown. It is possible that the vitamin D deficiency observed in SLE patients is a result of reduced sun exposure due to photosensitivity; however, such a correlation remains speculative. Regardless of the cause, vitamin D deficiency is widely believed to play a role in pathogenesis. In support thereof, the few animal studies available have demonstrated improvement in lupus features when supplemented with 1,25(OH)2
], yet no improvement was observed when animals were supplemented with 25(OH)D3 [46
]. Based on these studies, it is possible that calcitriol supplementation is needed to achieve clinical improvement. Such a requirement may explain the lack of clinical success, as calcitriol, but not 25(OH)D3, supplementation carries a risk of toxicity and is therefore not typically used for SLE patients without severe kidney disease. Investigations into whether there are therapeutic differences between the two and whether 1,25(OH)2
D3 can be therapeutic at non-toxic concentrations will be interesting.