Seroreactivity against Specific L5P Antigen from Mycobacterium avium subsp. paratuberculosis in Children at Risk for T1D

Abstract

Aims/hypothesis: Although numerous environmental agents have been investigated over the years as possible triggers of type 1 diabetes (T1D), its causes remain unclear. We have already demonstrated an increased prevalence of antibodies against peptides derived from Mycobacterium avuim subsp. paratuberculosis (MAP) homologous to human zinc transporter 8 protein (ZnT8) and proinsulin in Italian subjects at risk for or affected by T1D. In this study, we compared titers of the previously detected antibodies with seroreactivity to MAP lipopentapetide (L5P) that recently emerged as a strong immunogenic component able to specifically distinguish MAP from other mycobacteria.

Methods: Plasma of 32 children and youth at risk for T1D including follow-up samples and 42 age-matched healthy controls (HC) recruited at the Tor Vergata University Hospital in Rome was analyzed by indirect ELISA for the presence of antibodies against MAP-derived epitopes MAP3865c133-141, MAP3865c125-133, MAP2404c70-85 and MAP1,4αgbp157-173 along with their ZnT8 and proinsulin homologs. The data were analyzed through two-tailed Mann-Whitney U test and relation between variables was determined by principal component analysis.

Results: Responses to L5P were not detectable in subjects whose initial seroreactivity to MAP peptides and their human homologs was lost in follow-up samples, whereas anti-L5P antibodies appeared constantly in individuals with a stable immunity against MAP antigens. The overall coincidence in positivity to L5P and the four MAP epitopes both in children at risk for T1D and HC exceeded 90%.

Conclusions: MAP-derived homologs may cross-react with ZnT8 and proinsulin peptides inducing immune responses at a young age in subjects predisposed for T1D. Thus, L5P may have a diagnostic value to immediately indicate the presence of anti-MAP seroreactivity when evaluation of a more complex antibody status is not required. Almost complete coincidence in responses to both types of antigens lends support to the involvement of MAP in T1D.

Fig 1. Prevalence of Abs against L5P and MAP-derived ZnT8 homolog in T1D at-risk subjects and HC.

Plasma samples were analyzed in duplicate for Abs directed against L5P (A) and MAP3865c_133-141 (B) epitopes. Distributions are relative to the sample sets including only the first or the last time-point collection in comparison to HC. Horizontal bars indicate means. The dotted line corresponds to the Abs positivity threshold. AUC and p-values relative to T1D at-risk subjects vs. HC test are given in the upper part. Specific percentage of reactivity is indicated above each distribution.

Fig 2. Correlation between Abs recognizing MAP-derived epitope and L5P antigen in children at risk for T1D and HC.

The distributions represent correlations between Abs against MAP3865c_133–141 and L5P in T1D at-risk subjects including samples of the first time-point (A) or the last time-point (B) and HC (C). Each circle corresponds to Abs detected in one sample. The dotted lines indicate cut-off values used in each assay to discriminate between positive and negative samples. R² coefficients are given for each distribution.

Fig 3. Principal component analysis of variables describing relationship with positivity to L5P in samples of children at risk for T1D.

Bi-plot illustrates correlation between levels of anti-L5P Abs and variables relative to T1D genetic predisposition, available demographics data and titers of Abs against MAP-derived homologous peptides: MAP3865c_133–141 (Msh1); MAP3865c_125-133 (Msh2); MAP2404c_70-85 (M1); MAP1,4αgbp_157-173 (M2); ZnT8_186-194 (Z1); ZnT8_178-186 (Z2); PI_64-80 (P1) and PI_46-61 (P2). Children reactive to L5P antigen are indicated by squares whereas circles correspond to negative samples. All variables are described by labels and their position on the plot is indicated by triangles. The distribution shows relationship between PC1 and PC2 explaining 67.67% of the total variation. Only samples of children with known HLA genotype are included.