Monocytic Cytokines in Autoimmune Polyglandular Syndrome Type 2 Are Modulated by Vitamin D and HLA-DQ

Context Autoimmune polyglandular syndrome (APS-2: autoimmune Addison’s disease or type 1 diabetes) is conferred by predisposing HLA molecules, vitamin D deficiency, and heritable susceptibility. Organ destruction is accompanied by cytokine alterations. We addressed the monocytic cytokines of two distinct APS-2 cohorts, effects of vitamin D and HLA DQ risk. Methods APS-2 patients (n = 30) and healthy controls (n = 30) were genotyped for HLA DQA1/DQB1 and their CD14+ monocytes stimulated with IL1β and/or 1,25(OH)2D3 for 24 h. Immune regulatory molecules (IL-6, IL-10, IL-23A, IL-15, CCL-2, PD-L1), vitamin D pathway gene transcripts (CYP24A1, CYP27B1, VDR), and CD14 were analyzed by enzyme-linked immunosorbent assay and RTqPCR. Results Pro-inflammatory CCL-2 was higher in APS-2 patients than in controls (p = 0.001), whereas IL-6 showed a trend – (p = 0.1). In vitro treatment with 1,25(OH)2D3 reduced proinflammatory cytokines (IL-6, CCL-2, IL-23A, IL-15) whereas anti-inflammatory cytokines (IL-10 and PD-L1) rose both in APS-type 1 diabetes and APS-Addison´s disease. Patients with adrenal autoimmunity showed a stronger response to vitamin D. Expression of IL-23A and vitamin D pathway genes VDR and CYP27B1 varied by HLA genotype and was lower in healthy individuals with high-risk HLA (p = 0.0025; p = 0.04), while healthy controls with low-risk HLA showed a stronger IL-10 and CD14 expression (p = 0.01; p = 0.03). Conclusion 1,25(OH)2D3 regulates the monocytic response in APS-2 disorders type 1 diabetes or Addison´s disease. The monocytic cytokine profile of individuals carrying HLA high-risk alleles is proinflammatory, enhances polyglandular autoimmunity and can be targeted by vitamin D.

We investigated cytokine profiles of monocytes as precursors of macrophages and dendritic cells (DC) (14,15). Besides inhibiting inflammatory cytokines IL-2 and TNF-a (15), vitamin D suppresses monocytes' differentiation into mature DC, thereby reducing the number of antigen-presenting cells and T cell activating cells, thus inducing both innate and adaptive immune tolerance (14,16,17). These immunomodulatory effects have been successfully applied to experimental T1D models by protecting pancreatic b-cells from cytokine-induced inflammation and destruction (18)(19)(20).
While 1,25(OH) 2 D 3 is catalyzed by mitochondrial 1ahydroxylase (CYP27B1), degradation, and inactivation of both 25 (OH)D 3 and 1,25(OH) 2 D 3 is catalyzed by 24-hydroxylase (CYP24A1). By hydroxylation of the side chains CYP24A1 promotes the degradation into the water-soluble calcitroic acid and therefore preventing vitamin D toxicity. In the last decades, several studies illustrated the impact of HLA haplotypes and also of vitamin D on immunity independently from each other. Vitamin D was found to modulate excessive MHC II-and antigen expression (21) and healthy subjects homozygous for the HLA high-risk haplotypes showed increased surface expression of HLA-DR and -DQ, and peripheral blood mononuclear cells with increased IL1b and IFN-g (22). However, genetic susceptibility to APS-2 is conferred by HLA risk alleles, its interaction with the vitamin D system ill-defined. Only one Finnish study finds low vitamin D associated with the HLA B44 supertype (23), but not with HLA-DRB1 or HLA-DQB1 alleles. We therefore addressed this interaction by investigating in-vitro effects of active vitamin D 1,25 (OH) 2 D 3 (also referred to as calcitriol) on vitamin D pathway gene transcripts (CYP24A1, CYP27B1, VDR) and monocytic cytokines in patients with APS-2 and healthy controls in correlation with presence or absence of HLA risk heterodimers DQ2 and DQ8.

Subjects
A total of 15 T1D patients (9 female/6 male) and 15 AD patients (12 female/3 male) were recruited from the endocrine outpatient clinic of the University Hospital Frankfurt am Main, Germany. All patients were additionally affected by autoimmune thyroiditis (AIT) forming two variants of the APS-2. In AD/AIT patients mean age (± SD) was 57.5 years (± 13.7) and mean age of T1D/ AIT patients was 51.5 years (± 13.6). Healthy controls (HC, n = 30, 15 female/15 male) were volunteer blood donors without personal or a family history of autoimmune diseases and were randomly recruited from staff personnel or medical students from the University Hospital. Mean age (± SD) of HC was 41.1 years (± 12.6).

RNA Extraction, Reverse Transcription, and Quantitative PCR in Real Time (RT-qPCR)
RNA was extracted using the RNeasy Mini Kit (Cat. no. 74104; Qiagen, Hilden, Bayern), according to the manufacturer´s instructions. For reverse-transcription, 30 ng RNA was transcribed in accordance with instructions of Affinity Script QPCR Kit (Cat. no. 600559; Agilent Technologies) and the resulting cDNA was stored at -80°C. Specific Taqman assay primers (Thermo Fisher Scientific) and qPCR Rox Mix (Cat. no. AB-1138; Thermo Fisher Scientific, Schwerte, Germany) were used to perform the RT-qPCR assays. Primers for genes of the vitamin D pathway (CYP24A1 Hs00167999_m1, CYP27B1 Hs00168017_m1, and VDR Hs01045843_m1), genes encoding for immunological important cytokines and chemokines (IL-23A Hs00900828_g1, CCL-2 Hs00234140_m1, IL-6 Hs00985639_m1, IL-10 Hs00961622_m1, IL-15 Hs01003716_m1, CD274 (PD-L1) Hs01125301_m1) and the gene encoding the monocytic marker CD14 (Hs02621496_s1) were analyzed and compared to gene expression of the house-keeping reference control 18sRNA (Hs99999901_s1). Gene expression levels were quantified by ABI 7300 PCR system (Thermo Fisher Scientific) and the relative transcription levels were analyzed using the comparative cycle threshold (CT), as means of relative quantification, normalized to endogenous reference 18sRNA and expressed as 2 -DCT x10 6 .
The experimental part regarding IL-6 und CCL-2 mRNA expression, as well as CD14 expression in 13 AD/AIT patients, has been published (24) and is therefore not repeated here.

Enzyme-Linked Immunosorbent Assay
Enzyme-linked immunosorbent assays (ELISAs) were performed using the monocytic supernatants after cell culture. The supernatants were centrifuged for 7 min at 290 g using a benchtop centrifuge to remove cells and cell debris before storage at -80°C. The secreted proteins of IL-6 (Cat. no. 88-7066; RRID :

HLA DQ Genotyping
Genomic DNA was extracted from blood samples using the salting out method based upon reaction of negatively charged DNA with positively charged sodium. High-resolution sequence specific-primer (SSP) analysis was used for HLA-typing based on FIGURE 1 | Organs affected in autoimmune polyglandular syndrome type 2. The autoimmune polyglandular syndrome type 2 can affect a wide variety of organs in the body, triggering various autoimmune diseases. For reasons that are still unclear, APS-2 destroys particularly endocrine organs, leading to the development of type 1 diabetes, Addison's disease, Hashimoto's thyroiditis, and Graves' disease. However, non-endocrine organs such as the skin, the gastrointestinal tract, the brain, the liver, and also muscles and cartilage may be affected by the body's own destruction, which is often driven by antibodies and autoreactive lymphocytes. template's 3´end match or mismatch with PCR-based sequencespecific oligonucleotides. All subjects were genotyped for the MHC class II HLA DQA1/DQB1 alleles DQA1 *0101, *0102, *0103, *0104, *0201, *0301, *0401, *0501, *0601 and DQB1 *0201, *0301, *0302, *0303, *0401, *0402, *0501, *0502, *0503, *0601, *0602, *0603, *0604, using established primers (25). Therefore, all possible a-chain and b-chain variants of DQ were analyzed and subjects were typed according to variants of DQA1 and DQB1 alleles: DQ2 is encoded by the DQB1*0201 and DQA1*0501 alleles, while subjects carrying DQA1*0301 and DQB1*0302 are typed DQ8. Subjects carrying HLA-DQA1*0501 and DQB1*0201 on both loci or DQB1*0201 on both alleles and only one HLA-DQA1*0501 were identified as homozygous for DQ2 (DQ2/DQ2) and subjects with DQA1*0301 and DQB1*0302 on both loci as homozygous DQ8/DQ8. Since DQA1 is in linkage-disequilibrium with DQB1 the extended genotype was determined and categorized in three different risk groups depending on DQ2 and/or DQ8 presence or absence. According to our previous work (10,24) and congruent findings from others (9, 11) Figure 2D). Also resting monocytes showed an increased CD14 expression upon 1,25(OH) 2 D 3 (see Supplementary  Material Table A1).

DISCUSSION
Autoimmune disorders, including endocrinopathies are associated with genetic variants of the vitamin D system (26)(27)(28)(29)(30)(31). The multiplicity of endocrinopathies in APS-2 suggests a more severe immune defect, involving a stronger genetic susceptibility. HLA DQ risk carriers may have a distinct vitamin D effect on monocytic cytokines since vitamin D response elements (VDRE) are located upstream from promoter regions of HLA class II genes (32). CD14 + monocytes of HC and APS-2 patients demonstrated a 1,25(OH) 2 D 3 -induced negative feedback regulation of VDR and CYP27B1 and increased CYP24A1 expression, revealing a natural regulatory vitamin D circuit in both patients and HC. Vitamin D upregulated CD14 in both T1D/AIT patients and HC, in agreement with earlier reports (33,34). Thereby vitamin D suppresses DC formation of monocytes (34,35).
Patients' monocytes express and secrete higher levels of inflammatory cytokines IL-6 and CCL-2. In vitro vitamin D treatment reduces IL-6, CCL-2, IL-23A, and IL-15 and enhances pro-inflammatory IL-10 and PD-L1 expression and secretion. AD/ AIT patients displayed a lower expression of the pro-inflammatory cytokines and an increased production of anti-inflammatory IL-10, which can be attributed to their GC substitution. Both steroids vitamin D and GC act synergistically mediated by a co-activator complex (MED14) which enhances both VDR and glucocorticoid receptor action (36). This suggests, that a combined administration of vitamin D and GC could resolve inflammatory conditions and could potentially also reduce the need for higher GC doses to avoid GC side effects. For diseases like asthma and psoriasis this combined therapy has already been approved (37)(38)(39)(40). The combination therapy with vitamin D and glucocorticoids therefore offers a new anti-inflammatory therapeutic approach for APS-2 patients. The immune response to both vitamin D and to GC is variable and depends on genetic variants (27,41,42). We have identified specific epitopes of risk HLA alleles associated with APS-2 with the HLA-DQB1 position 57 defining disease susceptibility (25). We found HLA DQA1*0301:DQB1*0302 conferring a significantly increased risk for APS-type 1 diabetes and less for APS-Addison´s disease. Here we show that healthy carriers of ab-HLA DQ heterodimers conferring protection, differ for IL-10, CD14 and vitamin D pathway gene expression. Protective HLA alleles (neither DQ2 nor DQ8) associate with higher levels of antiinflammatory IL-10 and CD14, corresponding to a tolerogenic immune profile. In contrast high-risk HLA carriers showed a lower expression of VDR and CYP27B1 which implies less responsiveness to vitamin D. This corresponds to the finding that monocytes from subjects homozygous for HLA high-risk show a stronger surface expression of HLA-DR and -DQ, and also more IL-1b and IFN-g (22). Furthermore newborns with high-risk HLA DR4-DQ8 have less potential for Th2 differentiation with reduced CCR4-, IL-13 and Th2-inducing transcription factor GATA-3 levels (43). This could suggest that tailored vitamin D supplementation based on the HLA genotype aimed to modulate monocytic cytokine responses in such individuals. Large scale genetic and epigenetic fine mapping in 21 autoimmune diseases identified causal variants primarily in regulatory elements thatamongst others-enhance antigen processing for MHC presentation (44). Thereby autoimmune susceptibility is twotiered. An altered cytokine profile can enhance the expression of HLA molecules leading to enhanced autoantigen presentation in carriers of high-risk HLA. The combined genetic and epigenetic risk can be targeted by vitamin D that suppresses HLA-DR and CD4 antigen expression with autocrine downregulation of T cells (21). This effect can be relevant for primary prophylaxis as shown in a controlled trial of Influenza vaccination of elderly individuals where adjuvant vitamin D shifts the TH1-cytokine response towards a more immunosuppressive state with increased levels of TH2 cytokines IL-4, IL-5, IL-10 (45).
Due to the limited numbers in our investigation these findings need to be confirmed in larger cohorts. The molecular mechanisms for HLA DQ allele specific vitamin D effects need to be addressed by upstream sequencing and functional studies for an individualized vitamin D therapy. Nevertheless, our study shows that individuals at high genetic risk develop an amplified monocytic cytokine response that can be modulated by vitamin D.

DATA AVAILABILITY STATEMENT
The datasets presented in this study can be found in online repositories. The names of the repository/repositories and accession number(s) can be found in the article/Supplementary Material.