Introduction

Rheumatoid arthritis (RA) is a systemic autoimmune disease which has a prevalence of approximately 0.5–1% of the population of the industrialized world [1]. It is characterized by cartilage destruction and bone erosion, eventually leading to a high risk of morbidity and mortality. The etiopathogenesis of RA remains unknown, with the involvement of more than one mechanism in disease development. Emerging evidence suggests that vitamin D plays an important role in immune regulation, including RA. Beyond its endocrine role in bone metabolism, vitamin D is endowed with remarkable immunomodulatory properties, acting as a hormone precursor [2]. The importance of the immune cell as a target of vitamin D in RA was illustrated by Wei et al. [3]. Vitamin D could modulate innate immunity by promoting monocytes’ transformation into macrophages and affecting the release of related chemokines and cytokines [2]. In the adaptive immune system, vitamin D could inhibit B cells’ and T cells’ proliferation and differentiation, and downregulates the expression of immunoglobulins and autoantibodies to regulate humoral and cellular immunity [4, 5]. Despite evidence supporting the inverse association between the levels of vitamin D and disease activity [6, 7], results of previous studies are contradictory [8] regarding the level of vitamin D and affliction with disease. Besides, there are few studies targeting the association between the absolute numbers of immunocytes and vitamin D in RA. Therefore, there still needs to be a comprehensive study to assess the issue. Given this gap in the literature, this study was performed to determine if patients with RA are at higher risk of vitamin D deficiency and to explore a potential association between vitamin D and clinical/experimental characteristics of RA.

Methods

Patients and enrolment criteria

In a cross-sectional case–control study, we evaluated the charts of over 6000 RA patients. A total of 280 consecutive treatment-naïve RA patients from rheumatologic clinic of the Second Hospital of Shanxi Medical University were enrolled from September 2015 to November 2016. All patients fulfilled the American College of Rheumatology (ACR) 1987 revised criteria for RA [9]. The 140 age- and sex-matched healthy volunteers were mostly recruited from patients’ acquaintances that lived with them (patients’ relatives), to minimize the influence of lifestyles such as dressing and nutritional habits on vitamin D status. Patients receiving or who had ever received vitamin D, corticosteroids, disease-modifying anti-rheumatic drugs (DMARDs), or tumor necrosis factor antagonists, and those who had hepatic or renal insufficiency were excluded. Additional exclusion criteria were as follows: older than 85 years, malabsorption, pregnancy, lactation, systemic lupus erythematosus (sunlight exposure limited), malignancy, diabetes mellitus, hyperthyroidism, celiac disease, inflammatory bowel diseases, current and/or long-term usage of tuberculosis or fungal medications.

Study design

Demographic data like gender, age, weight, height, body mass index (BMI = kg/m2), and disease duration (time elapsed since symptom onset) were included. Stratified by BMI, RA individuals were classified into four groups: underweight (BMI < 18.5 kg/m2), normal (BMI 18.5–23.9 kg/m2), overweight (BMI ≥ 24 kg/m2), and obese (BMI ≥ 28 kg/m2) [10]. DAS28-CRP scores were calculated using following equations for patients with complete data [11]:

DAS28-CRP = [0.56 * sqrt (tender joint count) + 0.28 * sqrt (swollen joint count) + 0.36 *ln (CRP +)] * 1.10 + 1.15.

Patients were categorized into three groups by DAS28: high disease activity (DAS28: >5.1), moderate disease activity (DAS28: 3.2–5.1), and low disease activity (DAS28: <3.2) [12, 13].

Vitamin D deficiency was defined as 1,25-dihydroxycholecalciferol (1,25(OH)2D3), the active form of vitamin D, levels <25 ng/ml, as measured by enzyme-linked immunosorbent assay (ELISA) in accordance with expert consensus [14]. Erythrocyte sedimentation rate (ESR) was analyzed by the Westergren method. Immunophenotypes were determined by a FACSCalibur flow cytometer (Becton Dickinson Bio, USA). The cells were stained with following antibodies: total T cell (CD3+CD19−; normal range [NR]: 955 ~ 2860/μl), total B cell (CD3−CD19+; NR: 90 ~ 560/μl), natural killer (NK) cell (CD3−/CD16+CD56+; NR: 150 ~ 1100/μl), T helper 1 (Th1) cell (IFN-γ; NR: 5.52 ~ 182/μl), T helper 2 (Th2) cell (IL-4; NR: 4.04 ~ 21/μl), T helper 17 (Th17) cell (IL-17; NR: 3.07 ~ 14/μl), and regulatory T (Treg) cell (CD4+CD25+Foxp3; NR:17.7 ~ 54.2/μl). The ratio of Th17 cell versus Treg cell (Th17/Tregs) was calculated (NR: 0.09 ~ 0.47). Serum autoantibodies including rheumatoid factor (RF), antiperinuclear factor (APF), anti-keratin antibody (AKA), anti-mutated citrullinated vimentin (anti-MCV), and anti-cyclic-citrullinated peptides (anti-CCP) antibodies were examined by the chemiluminescence microparticle immunoassay. Quality control protocols were included, running known standards each day before testing samples. In addition, the laboratory was enrolled in external quality assurance testing programs with the College of American Pathologists and the United Kingdom National External Quality Assurance Service.

Statistical analysis

Nominal variables results were expressed as mean ± standard deviation (SD). Categorical variables were compared by chi-square test, and continuous variables were compared by one-way analysis of variance. Linear regression was performed to examine associations and estimate the contributing factors affecting serum 1,25(OH)2D3 levels. P value of less than 0.05 was considered statistically significant. Statistical analyses were performed by SPSS version 23.0 (IBM Corp., Armonk, NY, USA).

Results

Clinical information and serum 1,25(OH)2D3 levels of RA individuals and healthy volunteers

There was no significant difference in age, gender, and BMI between RA individuals and healthy volunteers (p > 0.05; Table 1). The average age of RA patients was 53.55 ± 13.87 years, ranging from 22 years to 85 years. Among all the subjects, 208 (72.7%) were female. The mean serum 1,25(OH)2D3 levels of the RA patients were significantly lower than those of healthy controls (12.24 ± 6.68 ng/ml versus 21.08 ± 7.14 ng/ml; p < 0.05; Table 2). The main RA-related findings with vitamin D insufficiency or with vitamin D sufficiency are shown in Table 2, which also includes the main characteristics of the healthy controls. Using recommended cut-off points shown in Table 3, 132 (46.2%) RA patients had severe vitamin D deficiency with 1,25(OH)2D3 levels <10 ng/ml, 118 (41.3%) of the subjects had below 20 ng/ml, and only 15 (5.2%) had over 25 ng/ml. Using the suggested threshold (≤25 ng/ml) [14], the overall prevalence of vitamin D insufficiency for RA was 94.8% (265 out of 280) compared with 52.1% (73 out of 140) for healthy controls (Table 3).

Table 1 The RA individuals were compared with the healthy volunteers in general (\(\bar{\mathrm{x}}\) ± s)
Full size table
Table 2 Main characteristics of RA patients and healthy volunteers, and the association between RA patients with vitamin D insufficiency or with vitamin D sufficiency
Full size table
Table 3 Distribution of serum 1,25(OH)2D3 levels in the studied cohort
Full size table

1,25(OH)2D3 and BMI in RA individuals

When stratified using subtypes of BMI, a significant difference of serum 1,25(OH)2D3 levels was found among obese (BMI ≥28 kg/m2; 5.21 ± 2.08 ng/ml), overweight (BMI ≥ 24 kg/m2; 11.6 ± 6.29 ng/ml), normal-weight (BMI: 18.5–23.9 kg/m2; 16.80 ± 4.12 ng/ml), and underweight (BMI < 18.5 kg/m2; 27.22 ± 3.18 ng/ml) in RA individuals (F = 31.41, p = 0.001; Fig. 1a). An inverse association was found between 1,25(OH)2D3 and ESR in obese and overweight RA individuals (βobese = −0.385, βoverweight = −0.395, both p < 0.05), but not in normal- and underweight (Table 4).

Fig. 1
figure 1

The association between 1,25-dihydroxycholecalciferol (1,25(OH)2D3) and body mass index (BMI) score, disease activity, and anti-cyclic-citrullinated peptides (anti-CCP) in rheumatoid arthritis (RA) individuals. a A significant difference of serum 1,25(OH)2D3 levels was found among obese (BMI ≥ 28 kg/m2), overweight (BMI ≥ 24 kg/m2), normal-weight (BMI:18.5–23.9 kg/m2), and underweight (BMI < 18.5 kg/m2) RA individuals (F = 31.41, p = 0.001). b The scores of DAS28 for high, moderate, and low disease activity. c Significantly lower 1,25(OH)2D3 values were found in high disease activity group compared to those having moderate and low disease activity (p < 0.05). d The serum 1,25(OH)2D3 levels for RA patients in the presence of anti-CCP were lower than that in the absence of anti-CCP (t = –3.08, p < 0.01)

Full size image
Table 4 The associations of serum 1,25(OH)2D3 concentrations with ESR or DAS28 in RA individuals stratified on subtypes of BMI, disease activity, and anti-CCP
Full size table

Levels of 1,25(OH)2D3 and disease activity in RA individuals

A considerable proportion of RA patients (39.3%) had a mean DAS28 score of 6.17 ± 0.77, representing high disease activity (DAS28 scores >5.1). In addition, 126 (45.0%) patients (4.30 ± 0.52) had moderate disease activity (DAS28 scores 3.2–5.1), and 44 (15.7%) patients (2.56 ± 0.60) had low disease activity (DAS28 scores <3.2). A significant difference in serum 1,25(OH)2D3 levels was found among those three groups of different disease activity (F = 554.64, p < 0.05; Fig. 1b). Significantly lower values of 1,25(OH)2D3 (8.10 ± 3.64 ng/ml) were found in the high disease activity group compared to those of moderate and low disease activity (27.00 ± 3.54 ng/ml and 30.84 ± 2.67 ng/ml; p < 0.05 and p < 0.05, respectively; Fig. 1c). A significantly negative association between 1,25(OH)2D3 levels and DAS28 scores (β = −0.164, p = 0.018) was observed. In addition, serum 1,25(OH)2D3 levels were inversely associated with ESR and DAS28 scores (β = −0.387, −0.567, p < 0.01) in RA patients with high disease activity (Table 4).

1,25(OH)2D3 and anti-CCP in RA individuals

The serum levels of 1,25(OH)2D3 for RA patients in the presence of anti-CCP (10.86 ng/ml) were lower than those in the absence of anti-CCP (15.98 ng/ml; t = −3.08, p < 0.01; Fig. 1d). Serum 1,25(OH)2D3 levels in the anti-CCP-positive group were negatively associated with DAS28 scores (β = −0.464, p < 0.01). There was no obvious association between the 1,25(OH)2D3 and RF, APF, AKA, or anti-MCV (Table 4).

Levels of 1,25(OH)2D3 and Th 17 and Treg cell counts in RA individuals

The relationship between 1,25(OH)2D3 and immune cells is summarized in Fig. 2. No statistically substantial associations were observed between 1,25(OH)2D3 and certain variables (T cell, B cell, NK cell, Th1 cell, Th2 cell, Th17/Treg) in RA patients (Fig. 2a–f). However, the levels of 1,25(OH)2D3 were inversely associated with Th 17 cell counts (β = −0.158, p = 0.019) and positively associated with Treg cell counts (β = 0.146, p = 0.025; Fig. 2g, h).

Fig. 2
figure 2

The association between 1,25-dihydroxycholecalciferol (1,25(OH)2D3) and immune cells in rheumatoid arthritis (RA) patients. a–f No significant association was found between 1,25(OH)2D3 and T cell, B cell, natural killer (NK) cell, T helper 1(Th1)-cell, T helper 2(Th2)-cell, or T helper 17 (Th17)/regulatory T (Treg) (β = −0.087, −0.058, 0.085, −0.045, −0.104, −0.105, respectively, all p > 0.05). The levels of 1,25(OH)2D3 were inversely associated with Th 17 cell counts (β = −7 cell p = 0.019; g) and positively associated with Treg cell counts (β = 0.146, p = 0.025; h)

Full size image
Fig. 2
figure 3

(Continued)

Full size image

Discussion

In this study, concentrations of 1,25(OH)2D3 in the patients with RA were significantly lower than those in healthy controls, consistent with previous cross-sectional studies in white [15], Chinese [16], Indian [17], and Iranian [6] RA patients. With an increasing incidence of vitamin D deficiency in patients with established RA, nearly 95% of subjects had 1,25(OH)2D3 levels below the currently accepted thresholds, indicating that vitamin D deficiency is a general problem. Vitamin D insufficiency was more common in women, with a prevalence of 72.7%, which was in good agreement with our previous studies [18]. The finding in men as an adverse risk factor in low serum 1,25(OH)2D3 levels is intriguing, and possibly due to their higher levels of androgens [19, 20]. We also found an inverse correlation between 1,25(OH)2D3 and ESR in obese and overweight RA individuals (high BMI), consistent with Nikiphorou’s study [21]. A recent cross-sectional analysis of 120 men and women reported that a decrease in 1,25(OH)2D3 serum value was observed in parallel with an increase in BMI level (r = −0.266, p = 0.037) [22]. BMI is a well-established risk factors for vitamin D deficiency and these associations were confirmed in the recent study [23].

We found that the worse the indices of disease activity, the lower the 1,25(OH)2D3 levels or the higher the proportion of patients with vitamin D deficiency. Wen H et al. [18], Cutolo et al. [24], and Patel et al. [25] all shared a similar outcome with our current analysis, and reported an inverse association between DAS28 and 1,25(OH)2D3 levels in their entire samples (n = 132, 118, and 206, respectively). However, it does have contradictory results to previous studies [26] in which 1,25(OH)2D3 failed to yield a significant association with disease activity, mainly regarding differences in exposure to the number of cases, measurement methods, environments, illumination time or geographical locations.

Interestingly, our study reported a relationship between low 1,25(OH)2D3 body supplies and positive anti-CCP in RA patients. Our study is in agreement with Kerr et al. [27]. Sahebari et al. [28] also suggested that in the early diagnosed RA patients, 1,25(OH)2D3 and anti-CCP serum values were negatively correlated (r = −0.5, p = 0.04). These findings may be in line with the hypothesis of 1,25(OH)2D3 deficiency as a predisposing factor for the initiation of autoimmune process. As noted, the documents in agreement with the bone-preserving role of 1,25(OH)2D3 on the specific facets of human immunity have been shown to be on the rise in recent years, but it seems that the associations between 1,25(OH)2D3 and specific autoantibodies are not clear yet. A large number of studies need to be investigated further.

Interestingly, we have shown that 1,25(OH)2D3 may act by restoring Th17 and Treg balance, thereby restoring immune homeostasis, which is interesting in light of the correlations between these cells and vitamin D in RA. Colin [5] also suggested that 1,25(OH)2D3 may contribute its bone-sparing effects in RA patients by modulating levels of Th17 and inhibiting Th17 cytokines, which is in good agreement with our data. Besides 1,25(OH)2D3 stimulating Treg activity, polarized Tregs express a higher level of Treg-associated markers such as CTLA4, PD1, and CD25, and their suppressive capacity is enhanced by 1,25(OH)2D3 [2]. Also, the important role of Th17 cells and the suppressive capacity of Treg cells were highly related to the beneficial effect of 1,25(OH)2D3 in multiple sclerosis [29], systemic lupus erythematosus [30], asthma [31], and renal transplant [32] patients. Therefore, it is hypothesized that 1,25(OH)2D3 suppresses autoimmunity at least partially via inhibition of Th17 activity and optimization of Treg function [33].

The existence of a worse response to treatment in patients with hypovitaminosis D compared to those with normal 1,25(OH)2D3 levels at RA onset was demonstrated by Di Franco [34]. Furthermore, Hajjaj-Hassouni found out that absence of supplementation of vitamin D is related to higher prevalence of vitamin D deficiency in a COMORA study from 15 countries [35]. Despite some controversies [2], the majority of reports reinforce the idea of the added value for the treatment of autoimmunity [36]. A possible explanation for this observation could be precisely the 1,25(OH)2D3 screening and supplementation strategies employed both at the clinic but also at the national level, minimizing the risk of 1,25(OH)2D3 deficiency, even though there is no consensus concerning the amount that should be indicated.

Several limitations of this study should be noted. First, the study population was selected from the Second Hospital of Shanxi Medical University. Thus, these findings may not be representative of patients in other areas. Second, since intake of 1,25(OH)2D3 was able to be affected by different seasons, geographical limits, outdoor exercise or not, sun exposure, and other factors, we did not strictly balance the factors upon selection of experimental patients. Therefore, a longer follow-up period of our cohort is required to validate the role of vitamin D (or analog) supplementation, beyond correction of deficiency, as a treatment modality, and to determine if higher vitamin D levels lead to standard clinical improvement in disease activity and immune regulation.

Conclusions

Our preliminary data showed a mutual influence of vitamin D on disease activity, BMI index, Th17 cell, Treg cell, and anti-CCP antibody, especially in female RA patients. These findings may be in line with the hypothesis that vitamin D deficiency might be a predisposing factor for the initiation and progression of RA. However, we cannot rule out the possible impact of the cross-sectional design on the study results and we also are not able to determine whether vitamin D deficiency directly impacts RA disease activity and immune regulation, or whether the reverse may be true. In spite of this, serum 1,25(OH)2D3 levels could be a marker to monitor disease activity in RA patients and vitamin D may be an alternative supplementary treatment for RA.