Table 1.
PE/PPE proteins selected for this study.
Figure 1.
IFN-γ responses to PE/PPE peptide pools in experimentally infected cattle.
Whole blood from cattle experimentally infected with 3×103 CFU M. bovis AF2122/97 was stimulated with pools of peptides representing 36 PE/PPE proteins at 12–14 weeks post-infection. IFN-γ production was assessed by the ELISA-based Bovigam assay. (A) IFN-γ responses to 101 PE/PPE peptide pools in experimentally infected cattle (n = 17) expressed as OD450– Nil antigen well reading (ΔOD450). Grey box highlights responses to PE/PPE peptides derived from PE/PPE proteins encoded by partial esx regions. Peptides derived from PE/PPE proteins from complete ESX regions and those encoded as isolates PE/PPE pairs are also indicated. Responses to an EsxA/B peptide cocktail are also shown for comparison. (B) Twenty-five PE/PPE pools that elicited positive responses in ≥40% cattle tested, ranked according to proportion responders.
Table 2.
Bovine and human responses to selected PE/PPE peptide pools.
Figure 2.
Comparison of proportion responders in human and bovine hosts with active TB.
Responses to 28 PE/PPE peptide pools were assessed in both cattle and humans with active TB, by Bovigam or ELISPOT assays. Individual cattle samples were scored as positive if ΔOD450 values were ≥0.1, and human samples were scored as positive if ≥20 SFC were counted per 106 PBMC. Linear regression analysis produced the line shown. A Pearson correlation test revealed a statistically significant correlation between the proportion human responders and the proportion bovine responders; r(26) = 0.539, 2-tailed p value = 0.0031.
Figure 3.
Immune responses to PE/PPE proteins in humans with active and latent infection.
PBMC from M. tuberculosis-sensitized individuals were stimulated with PE/PPE peptide pools, and IFN-γ producing cells were enumerated by ELISPOT. (A) Proportion responders to all PE/PPE pools analyzed. Grey bars, individuals with latent infection; Black bars, individuals with active infection. (B) Number of IFN-γ producing cells measured after stimulation with PE/PPE peptide pools. White symbols, individuals with latent infection; Black symbols, individuals with active infection. No statistically significant differences were observed between responses in individuals with latent and active infection for any of the peptide pools analyzed (as determined by Mann Whitney U test). Results for a representative set of pools shown, horizontal lines indicate median responses.
Figure 4.
Temporal profile of bovine IFN-γ responses to PE/PPE peptide pools.
IFN-γ responses were measured following stimulation of whole bovine blood with PE/PPE peptides and control antigens before, and at multiple time points following, experimental infection of cattle with M. bovis AF2122/97. (A) Responses to controls antigens; avian PPD (PPD-A), bovine PPD (PPD-B) and an EsxA/B peptide cocktail. (B) to (D) Representative responses to sets of PE/PPE peptide pools. Open grey symbols, PE pools; Solid black symbols, PPE pools; ▪, ▴, ▾, ♦, •, pools A-E, respectively (B) IFN-γ responses to PE18/PPE26 peptide pools. (C) IFN-γ responses to PE31/PPE60 peptide pools. (D) IFN-γ responses to PE8/PPE15 peptide pools. Mean responses for a single representative animal are shown.
Figure 5.
Immune recognition correlates with sequence conservation.
(A) IFN-γ responses were measured following stimulation of whole bovine blood with PE/PPE peptide pools, and scored as positive if ΔOD450 values were ≥0.1. The proportion of responders was plotted along the length of the PE/PPE proteins, where A corresponds to the peptide pool closest to the N-terminal. *, 2-tailed p value <0.05 (unpaired t-test). The letters A to F denote the position of the peptide pool across the length of the protein, with A nearest the N-terminal. PE_A represents the conserved 110 aa region, while PPE_A and PPE_B encompasses the conserved 180 aa region of the PPE proteins. (B) The proportion responders to PE/PPE peptide pools was plotted against % identity, which corresponds to the % identity of the amino acid sequence corresponding to that pool to its closest match in the M. tuberculosis genome sequence. Linear regression analysis produced the line shown. A Spearman correlation test indicated revealed a statistically significant correlation between % identity and proportion bovine responders; rho(98) = 0.401, 2-tailed p value <0.0001.
Figure 6.
Whole blood IFN-Υ responses to individual and pooled PE/PPE peptides.
Peptide pools eliciting positive responses were further analysed by assessing the contribution of each constituent peptide at 3, 5 or 9 weeks post-infection. Representative data are shown for peptides corresponding to (A) One pool derived from PE8, and (B) One pool derived from PPE15. The solid lines represent the temporal response to a peptide pool, while individual symbols correspond to individual peptides. (C) and (D) BLASTP homology searches were performed with peptides which elicited positive responses, and closest matches are shown in a CLUSTALW alignment.
Figure 7.
IFN-γ responses of short-term T cell lines to potentially cross-reactive PPE peptides.
Short-term T cell lines were raised by primary stimulation of bovine PBMC with individual PE or PPE peptides in the presence of IL-2. Following initial expansion with (A) PE31.10, (B) PPE18.14 or (C) PPE60.14, cells were re-stimulated with selected peptides, predicted to be immunologically cross-reactive. IFN-γ production was then measured using the ELISA-based Bovigam assay. The amino acid sequence of the initial stimulation peptide is shown in alignment with restimulation peptides, with amino acid substitutions indicated. Responses are shown as ΔOD450 values, and the dashed line indicates the positivity cut-off value of 0.1.