In silico prediction of peptides binding to multiple HLA-DR molecules accurately identifies immunodominant epitopes from gp43 of Paracoccidioides brasiliensis frequently recognized in primary peripheral blood mononuclear cell responses from sensitized individuals

Abstract

One of the major drawbacks limiting the use of synthetic peptide vaccines in genetically distinct populations is the fact that different epitopes are recognized by T cells from individuals displaying distinct major histocompatibility complex molecules. Immunization of mice with peptide (181-195) from the immunodominant 43 kDa glycoprotein of Paracoccidioides brasiliensis (gp43), the causative agent of Paracoccidioidomycosis (PCM), conferred protection against infectious challenge by the fungus. To identify immunodominant and potentially protective human T-cell epitopes in gp43, we used the TEPITOPE algorithm to select peptide sequences that would most likely bind multiple HLA-DR molecules and tested their recognition by T cells from sensitized individuals. The 5 most promiscuous peptides were selected from the gp43 sequence and the actual promiscuity of HLA binding was assessed by direct binding assays to 9 prevalent HLA-DR molecules. Synthetic peptides were tested in proliferation assays with peripheral blood mononuclear cells (PBMC) from PCM patients after chemotherapy and healthy controls. PBMC from 14 of 19 patients recognized at least one of the promiscuous peptides, whereas none of the healthy controls recognized the gp43 promiscuous peptides. Peptide gp43(180-194) was recognized by 53% of patients, whereas the other promiscuous gp43 peptides were recognized by 32% to 47% of patients. The frequency of peptide binding and peptide recognition correlated with the promiscuity of HLA-DR binding, as determined by TEPITOPE analysis. In silico prediction of promiscuous epitopes led to the identification of naturally immunodominant epitopes recognized by PBMC from a significant proportion of a genetically heterogeneous patient population exposed to P. brasiliensis. The combination of several such epitopes may increase the frequency of positive responses and allow the immunization of genetically distinct populations.

Figure 1

Western blotting of PCM patients’ sera and a healthy individual against gp43 of P. brasiliensis. Serum dilution 1:40; gp43 concentration used: 2.5 μg/mL.

Figure 2

Scanning of gp43 of P. brasiliensis by TEPITOPE algorithm at 3% threshold. Five peptides marked with arrows, predicted to bind to at least 10 of the 25 HLA-DR molecules included in the algorithm, were selected as promiscuous epitopes. Peptides marked with asterisks were predicted not to bind significantly to any HLA-DR molecules and were used as peptide controls.

Figure 3

Predicted promiscuity of peptides for binding with high affinity to different HLA-DR molecules at 3% threshold (▪); frequency of peptides binding to 9 different HLA-DR molecules (□); frequency of responders to peptides tested in proliferation assay with PBMC of 19 treated and healed nonanergic PCM patients (▧).

Figure 4

A: Frequency of responders to P10 peptide (gp43 [181-195]) and its neighboring peptides gp43(183-197), gp43(180-194), and gp43(179-199). B–D: Average SI of responders to peptides P10 and its neighboring peptides at 0.1 μM (B), 1.0 μM (C), and 10.0 μM (D)