Genetic loci of Staphylococcus aureus associated with anti-neutrophil cytoplasmic autoantibody (ANCA)-associated vasculitides

Abstract

The proteinase 3 (PR3)-positive anti-neutrophil cytoplasmic autoantibody (ANCA)-associated vasculitis (AAV) granulomatosis with polyangiitis (GPA) has been associated with chronic nasal S. aureus carriage, which is a risk factor for disease relapse. The present study was aimed at comparing the genetic make-up of S. aureus isolates from PR3-ANCA-positive GPA patients with that of isolates from patients suffering from myeloperoxidase (MPO)-ANCA-positive AAV, and isolates from healthy controls. Based on a DNA microarray-based approach, we show that not only PR3-ANCA-positive GPA patients, but also MPO-ANCA-positive AAV patients mainly carried S. aureus types that are prevalent in the general population. Nonetheless, our data suggests that MPO-ANCA-associated S. aureus isolates may be distinct from healthy control- and PR3-ANCA-associated isolates. Furthermore, several genetic loci of S. aureus are associated with either PR3-ANCA- or MPO-ANCA-positive AAV, indicating a possible role for pore-forming toxins, such as leukocidins, in PR3-ANCA-positive GPA. Contrary to previous studies, no association between AAV and superantigens was detected. Our findings also show that a lowered humoral immune response to S. aureus is common for PR3-ANCA- and MPO-ANCA-positive AAV. Altogether, our observations imply that the presence or absence of particular virulence genes of S. aureus isolates from AAV patients contributes to disease progression and/or relapse.

Figure 1

PCA of the ClonDiag microarray data. The x-axis represents PC1 and the y-axis represents PC2, which describes 45% and 17% of all of the variation in the data, respectively. Note that the circle indicating CC7 also includes five isolates that belong to the different low-abundance CC variants. The same is true for 2 isolates in the CC5 circle. The blue and red boxes display single genes or groups of genes that were associated either positively with PC1 and negatively with PC2 (red box), or the other way round (blue box). The associations with PC1 and PC2 are strongly correlated with the different CCs, of which the abundances differ between PR3-ANCA, MPO-ANCA and HC. The localization within each CC circle however also shows a clear intra-CC-specific association of PC1 with either MPO-ANCA (left) or PR3-ANCA and HC (right).

Figure 2

Visualization of the differences in variation (Δ) within each of the six main CCs, as described by ΔPC1 (A) and ΔPC2 (B), of PR3-ANCA, MPO-ANCA and HC isolates. Dashed lines represent median values.

Figure 3

Prevalence of selected genes that were identified before or after correlation analysis of ΔPC1 in PR3-ANCA, MPO-ANCA and HC S. aureus isolates. (A) The genes cap-5, cap-8, can, sasG, Q2YUB3, lukX, isaB, mprF, set4/ssl10 and set7/ssl2, (B) superantigen-encoding genes, and (C) leukocidin, immune evasion and protease genes.

Figure 4

Hierarchical clustering analysis on the variation within each of the 6 main clonal complexes, as described by the first 6 principal components (ΔPC1 – ΔPC6). A complete linkage clustering on the weighted results of ΔPC1 – ΔPC6, which together described 83% of the variation present within the entire data set, separate nearly all of the S. aureus isolates up into three unique groups. The small group indicated by *, represents two S. aureus isolates from CC8 which differ from the rest of CC8 as they are ST72 MRSA isolates, and hence form an outlier. ΔPC3 (left) does not strongly distinguish PR3-ANCA (red), MPO-ANCA (blue) and HC isolates (green) from one another and hence is represented as the top bar being the least relevant.

Figure 5

IgG responses of PR3-ANCA and MPO-ANCA patients or HC to staphylococcal antigens. Serum IgG levels against (A) secreted superantigens and superantigen-like proteins and (B) selected surface and secreted proteins. Depicted are the median with boxes (25% and 75%) and whiskers (minimum to maximum) of all sera per group. Statistical significances were tested using the Kruskal-Wallis test (with post-hoc Dunn’s test). *p < 0.05, **p < 0.01, ***p < 0.001 versus HC and ^#p < 0.05, ^##p < 0.01, ^###p < 0.001, ^####p < 0.0001 versus HC.