Antibodies recognizing Mycobacterium avium paratuberculosis epitopes cross-react with the beta-cell antigen ZnT8 in Sardinian type 1 diabetic patients

Abstract

The environmental factors at play in the pathogenesis of type 1 diabetes (T1D) remain enigmatic. Mycobacterium avium subspecies paratuberculosis (MAP) is transmitted from dairy herds to humans through food contamination. MAP causes an asymptomatic infection that is highly prevalent in Sardinian T1D patients compared with type 2 diabetes (T2D) and healthy controls. Moreover, MAP elicits humoral responses against several mycobacterial proteins. We asked whether antibodies (Abs) against one of these proteins, namely MAP3865c, which displays a sequence homology with the β-cell protein zinc transporter 8 (ZnT8) could be cross-reactive with ZnT8 epitopes. To this end, Ab responses against MAP3865c were analyzed in Sardinian T1D, T2D and healthy subjects using an enzymatic immunoassay. Abs against MAP3865c recognized two immunodominant transmembrane epitopes in 52-65% of T1D patients, but only in 5-7% of T2D and 3-5% of healthy controls. There was a linear correlation between titers of anti-MAP3865c and anti-ZnT8 Abs targeting these two homologous epitopes, and pre-incubation of sera with ZnT8 epitope peptides blocked binding to the corresponding MAP3865c peptides. These results demonstrate that Abs recognizing MAP3865c epitopes cross-react with ZnT8, possibly underlying a molecular mimicry mechanism, which may precipitate T1D in MAP-infected individuals.

Figure 1. Prevalence of anti-MAP3865c Abs in Sardinian T1D and T2D patients.

Sera were tested for their reactivity against plate-coated MAP3865c-MBP fusion protein. Ab distribution is shown for T1D (A) and T2D (B) patients compared to healthy controls. Dotted lines indicate the cut-off for positivity used in each assay, as calculated by ROC analysis. The percent fraction of Ab+ sera is indicated on top of each distribution, while bars indicate the corresponding median ± interquartile range. AUC and p values are given in the top right corner. Figures show representative experiments out of three performed.

Figure 2. Validation of anti-MAP3865c Ab assays.

(A) Reactivity against the MAP-MBP fusion protein is MAP-specific. Ab+ and Ab-negative sera from T1D and healthy donors were challenged either with the MAP-MBP fusion protein (as in Fig. 1) or with a LacZ-MBP control protein. The dotted line indicates the cutoff for positivity. (B) Intra- and inter-assay variability of MAP3865c ELISA Ab assays. For intra-assay variability (white bars), the same serum was tested in 20 replicate wells; bars show readouts of each single well. CV is 2.8%. For inter-assay variability (grey bars), the same serum was tested in 4 separate experiments; bars show mean ± SEM of triplicate wells from each experiment. CV is 7.4%.

Figure 3. Homology between MAP3865c and ZnT8 proteins.

(A) Aminoacid sequence alignment of ZnT8 (Slc30A8) and MAP3865c proteins. Conserved aminoacid residues are highlighted in yellow within the MAP3865c sequence and listed in bold below the two sequence alignment rows. The other color codes refer to the ZnT8 protein structure shown in (B): sequences highlighted in green belong to the 3 intra-luminal loops; the sequence in red belongs to the fourth transmembrane domain, while those in blue belong to the other transmembrane regions; sequences not highlighted fall within the 4 extra-luminal fragments. Dotted rectangles show the MAP3865c_125–133/ZnT8_178–186 and MAP3865c_133–141/ZnT8_186–194 peptides studied in subsequent experiments. The topology of the ZnT8 protein is also shown in panel (B), where the 3 intra-luminal loops (in green) become extracellularly exposed once the insulin granule is released. Conversely, the 4 extra-luminal domains (in black) are exposed to the cytosol and remain intracellular upon insulin exocytosis.

Figure 4. Ab reactivities against the MAP3865c protein are inhibited by MAP3865c_125–133 and MAP3865c_133–141 peptides.

Ab+ and Ab-negative sera from T1D patients were pre-incubated overnight with saturating concentrations (5.5 µM) of MAP3865c_125–133, MAP3865c_133–141, the two peptides in combination, MAP3865c-MBP fusion protein and control or no peptide. Their reactivity on MAP3865c-MBP-coated ELISA plates was subsequently tested. Bars depict means ± SEM of triplicate wells and results are representative of three separate experiments.

Figure 5. Prevalence of Abs against homologous MAP3865c and ZnT8 epitopes in T1D patients.

Prevalence of Abs against MAP3865c_125–133 (A) and its homologous ZnT8_178–186 (B); and against MAP3865c_133–141 (C) and its homologous ZnT8_186–194 (D) in T1D and healthy subjects. Data representation is the same as in Fig. 1.

Figure 6. Prevalence of Abs against homologous MAP3865c and ZnT8 epitopes in T2D patients.

Prevalence of Abs against MAP3865c_125–133 (A) and its homologous ZnT8_178–186 (B); and against MAP3865c_133–141 (C) and its homologous ZnT8_186–194 (D) in T2D and healthy subjects. Data representation is the same as in Fig. 1.

Figure 7. Correlation between titers of MAP3865c- and ZnT8-reactive Abs recognizing different epitopes.

Correlations are shown between titers of Abs recognizing (A) MAP3865c_125–133 and its homologous ZnT8_178–186 epitope; (B) MAP3865c_133–141 and its homologous ZnT8_186–194 epitope; (C) MAP3865c_125–133 and its consecutive MAP3865c_133–141 epitope; (D) ZnT8_178–186 and its consecutive ZnT8_186–194 epitope. Each circle represents the titers of one T1D or healthy donor.

Figure 8. Ab reactivities against MAP3865c epitopes are inhibited by the homologous ZnT8 epitopes.

(A) Ab+ and Ab-negative sera from T1D patients were pre-incubated overnight with saturating concentrations of MAP3865c_125–133 (white bars), ZnT8_178–186 (hatched bars), control (grey bars) or no peptide (black bars) and their reactivity on MAP3865c_125–133-coated ELISA plates subsequently tested. (B) The same sera were preincubated with MAP3865c_133–141 (white bars), ZnT8_186–194 (hatched bars), control (grey bars) or no peptide (black bars) and their reactivity on MAP3865c_133–141-coated ELISA plates subsequently tested. Bars depict means ± SEM of triplicate wells and results are representative of two separate experiments.