Decreased HLA-DQ expression on peripheral blood cells in children with varying number of beta cell autoantibodies

Abstract

The risk for type 1 diabetes is strongly associated with HLA-DQ and the appearance of beta cell autoantibodies against either insulin, glutamate decarboxylase (GAD65), insulinoma-associated protein-2 (IA-2), or zinc transporter 8 (ZnT8). Prolonged exposure to autoantibodies may be related to T cell exhaustion known to occur in chronic infections or autoimmune disorders. It was hypothesized that autoantibody exposure may affect HLA-DQ expression on peripheral blood cells and thereby contribute to T cell exhaustion thought to be associated with the pathogenesis of type 1 diabetes. The aim of this study was to determine whether autoantibody exposure as an expression of autoimmunity burden was related to peripheral blood cell HLA-DQ cell surface expression in either 1) a cross-sectional analysis or 2) cumulative as area under the trajectory of autoantibodies during long term follow-up in the Diabetes Prediction in Skåne (DiPiS) study. Children (n = 67), aged 10-15 years were analyzed for complete blood count, HLA-DQ cell surface median fluorescence intensity (MFI), autoantibody frequency, and HLA genotypes by Next Generation Sequencing. Decreased HLA-DQ cell surface MFI with an increasing number of autoantibodies was observed in CD16⁺, CD14⁺CD16^-, CD4⁺ and CD8⁺ cells but not in CD19⁺ cells and neutrophils. HLA-DQ cell surface MFI was associated with HLA-DQ2/8 in CD4⁺ T cells, marginally in CD14⁺CD16^- monocytes and CD8⁺ T cells. These associations appeared to be related to autoimmunity burden. The results suggest that HLA-DQ cell surface expression was related to HLA and autoimmunity burden.

Keywords: Autoantibodies; Autoimmunity; Cell surface imunofluorescence; Human Leukocyte antigen; Type 1 diabetes.

Fig. 1

Schematic of the Diabetes Prediction in Skåne (DiPiS) study timeline and time of sampling into our cross-sectional study. A timeline (panel A) of the key events of enrolment and follow-up in DiPiS and sampling in the present study. Screening of high risk for type 1 diabetes was carried out from 2000 until 2004. Children with increased risk of type 1 diabetes were enrolled at 2 years of age between 2002 and 2004 and followed either annually (negative or single autoantibody) or every three months (if multiple autoantibodies were detected at any earlier visit) until the age of 15 or diagnosis of type 1 diabetes. The sampling into our study was divided in two parts, part 1 where n = 21 subjects were sampled and part 2 where n = 46 subjects were sampled. Autoantibody profiles (panel B) of the n = 67 children in our study during follow-up as part of the DiPiS study and at time of sampling into our study. The timeline plot shows the visits (circles for visits as part of DiPiS follow-up, stars for time of sampling into our study) and autoantibody count (negative = green, single = yellow, multiple = red). Autoantibodies measured were GADA, IA2A, IAA and any of the three variants of ZnT8A against arginine, tryptophan or glutamine at position 325 (R/W/Q, respectively). Autoimmunity burden calculated as area under the trajectory of autoantibodies over time is presented to the right.

Fig. 2

Complete blood count of whole blood in n = 67 children, stratified by the autoimmunity burden measured as the number of autoantibodies detected at the time of sampling (sAB) (panel A) or cumulative autoimmunity burden (cAB) measured as area under the trajectory of autoantibodies over time (panel B). The HLA-DQ2/8 status is indicated by the color and shape (blue triangles: DQ2/8, red circles: non-DQ2/8). P-value (p) shown is the nominal p-value obtained using likelihood ratio tests to examine the association between complete blood count and sAB or cAB, respectively, for each cell type. None of the p-values remained significant after adjustment for multiple comparisons (corrected using the Benjamini-Hochberg procedure assuming 16 comparisons and a 5% false discovery rate).

Fig. 3

HLA-DQ cell surface median fluorescence intensity (MFI) on isolated peripheral blood cells in n = 67 children, stratified by autoimmunity burden measured as the number of autoantibodies detected at the time of sampling (sAB) (panel A) or cumulative autoimmunity burden (cAB) measured as area under the trajectory of autoantibodies over time (panel B). The HLA-DQ2/8 status is indicated by the color and shape (blue triangles: DQ2/8, red circles: non-DQ2/8). P-value (p) shown is the nominal p-value obtained using likelihood ratio tests to examine the association between HLA-DQ cell surface MFI and sAB or cAB, respectively, for each cell type. The p-values that remain significant after adjustment for multiple comparisons are indicated with a ∗ (corrected using the Benjamini-Hochberg procedure assuming 12 comparisons and a 5% false discovery rate).

Fig. 4

Estimates and 95% confidence intervals (Est, 95% CI)) of the association between HLA-DQ cell surface median fluorescence intensity (MFI) on isolated peripheral blood cells and HLA-DQ2/8 (panel A) and with autoimmunity burden measured as the number of autoantibodies detected at the time of sampling (sAB) (panel B) or cumulative (cAB) measured as area under the trajectory of autoantibodies over time (panel C). The models for HLA-DQ2/8 were adjusted for age and sex (Model 1) and complete blood count (CBC) of peripheral blood cells, red blood cells and platelets in addition to the parameters in Model 1 (Model 2). The models were fit using linear regression with robust standard errors. (See the Supplemental Table 2 panel A–C, respectively, for detailed results corresponding to these plots).