Costimulatory Effects of an Immunodominant Parasite Antigen Paradoxically Prevent Induction of Optimal CD8 T Cell Protective Immunity

Abstract

Trypanosoma cruzi infection is controlled but not eliminated by host immunity. The T. cruzi trans-sialidase (TS) gene superfamily encodes immunodominant protective antigens, but expression of altered peptide ligands by different TS genes has been hypothesized to promote immunoevasion. We molecularly defined TS epitopes to determine their importance for protection versus parasite persistence. Peptide-pulsed dendritic cell vaccination experiments demonstrated that one pair of immunodominant CD4+ and CD8+ TS peptides alone can induce protective immunity (100% survival post-lethal parasite challenge). TS DNA vaccines have been shown by us (and others) to protect BALB/c mice against T. cruzi challenge. We generated a new TS vaccine in which the immunodominant TS CD8+ epitope MHC anchoring positions were mutated, rendering the mutant TS vaccine incapable of inducing immunity to the immunodominant CD8 epitope. Immunization of mice with wild type (WT) and mutant TS vaccines demonstrated that vaccines encoding enzymatically active protein and the immunodominant CD8+ T cell epitope enhance subdominant pathogen-specific CD8+ T cell responses. More specifically, CD8+ T cells from WT TS DNA vaccinated mice were responsive to 14 predicted CD8+ TS epitopes, while T cells from mutant TS DNA vaccinated mice were responsive to just one of these 14 predicted TS epitopes. Molecular and structural biology studies revealed that this novel costimulatory mechanism involves CD45 signaling triggered by enzymatically active TS. This enhancing effect on subdominant T cells negatively regulates protective immunity. Using peptide-pulsed DC vaccination experiments, we have shown that vaccines inducing both immunodominant and subdominant epitope responses were significantly less protective than vaccines inducing only immunodominant-specific responses. These results have important implications for T. cruzi vaccine development. Of broader significance, we demonstrate that increasing breadth of T cell epitope responses induced by vaccination is not always advantageous for host immunity.

Fig 1. CD4⁺ TSaa57-74 (p7)/IA^d- and CD8⁺ TSaa359-367 (TSKd1)/K^d-specific T cell responses are minimally sufficient for induction of protective T. cruzi immunity.

Panel A shows a schematic of the TS consensus protein (all 12–15 active TS subfamily members have at least 90% homology within their catalytic domains), and immunodominant TS CD4 and CD8 epitopes. In panels B and C, BALB/c mice were vaccinated with dendritic cells (DC) pulsed (or not) with this pair of CD4 and CD8 epitopes and later challenged with T. cruzi. BALB/c CD11c⁺ splenic DC were purified from BALB/c mice 2 weeks after injection of 5x10⁶ B16-Flt3L-producing cells and 1 day after i.v. injection of 1μg LPS. 1x10⁶ DC pulsed (or left unpulsed) with 50 μg/ml of the indicated peptides were injected i.v. into groups of naïve BALB/c mice 3 times, 2 weeks apart. Mice were challenged 1 month later with T. cruzi (N = 5 in each of the two control groups and N = 10 in DC+TS peptide group). Both parasitemia (B; 2 weeks post-infection) and mortality (C) were significantly reduced in mice given DC pulsed with both TSaa57-74 (p7) and TSaa359-367 (TSKd1) [*p<0.001 by Mann-Whitney U test (B) and *p<0.01 by 2-tailed Fisher exact and Log-Rank [Mantel-Cox] tests(C)]. Survival results are representative of 3 independent experiments.

Fig 2. TSKd1-specific CD8⁺ T cell responses provide immunodominant protective activity.

To understand whether or not the TSaa359-367 peptide (TSKd1) is immunodominant and absolutely required for induction of protective T. cruzi immunity, a site-directed mutagenesis strategy was utilized (A) to completely remove predicted binding of TSKd1 to H2-K^d. Groups of BALB/c mice were vaccinated twice, 2 weeks apart with negative control DNA, the WT TS DNA (WT TS DNA) or the new TSKd1 null TS DNA (TSKd1 null DNA). One month after the final vaccination, pooled splenic CD8⁺ T cells were obtained from representative vaccinated mice and stimulated with control APC (NC A20), APC transfected with TS aa1-678 (TS A20), or APC pulsed with TSKd1 (A20 +TSKd1) in overnight IFN-γ ELISPOT assays (B). Other groups of vaccinated mice (N = 16–17 each) were challenged with 5,000 T. cruzi BFT s.c. 4 weeks following the final immunization and survival monitored (C). *, p<0.05 compared with NC DNA and CD8 Null TS DNA groups by 2-tailed Fisher exact tests and Log-Rank [Mantel-Cox] analyses. Results are representative of at least 3 experiments.

Fig 3. Identification of trans-sialidase MHC class I-restricted T cell epitope(s).

TSaa359-367 (TSKd1) has been observed to be an immunodominant epitope during both TS vaccination and T. cruzi infection. We utilized a consensus immunoinformatic approach using several MHC prediction tools to identify other TS sequences predicted to bind BALB/c MHC (H2-K^d, H2-D^d, and H2-L^d, shown in S1 Table). APC (A20) were pulsed with these synthetic peptides and used to stimulate purified splenic CD8⁺ T cells from naïve (A), WT TS DNA-vaccinated (B), and T. cruzi “Infection Memory” mice (C) in overnight IFN-γ ELISPOT assays. Infection Memory (C) was induced by multiple, virulent Tulahuen strain T. cruzi challenges, resulting in mice with potently protective T cell immunity directed against a variety of parasite antigens and epitopes. Nearly all of the predicted peptides elicited memory T cell IFN-γ responses in mice previously infected with T. cruzi and mice vaccinated with WT TS DNA. Results are representative of 3 independent experiments.

Fig 4. The WT DNA vaccine induces CD8⁺ T cell responses directed against both immunodominant and subdominant T cell epitopes.

BALB/c mice were vaccinated i.m. twice, two weeks apart with 100μg of WT TS DNA or TSKd1 null TS DNA. Four weeks later, CD8⁺ splenic T cells from these vaccinated mice were stimulated overnight in IFN-γ ELISPOT assays with APC (A20 cells) pulsed with TS peptides predicted to bind BALB/c MHC [H2-K^d (A), H2- D^d (B) and H2-L^d(C). Results are representative of 3 experiments. As expected, vaccination of mice with WT but not TSKd1 null TS DNA constructs elicited T cell responses to TSKd1. CD8⁺ T cells from WT TS DNA vaccinated mice also responded to TS peptides Kd2, Kd5, Kd6, Kd7, Kd8, Kd9, Dd2, Dd6, Ld1 and Ld2, while cells from TSKd1 null vaccinate mice responded to only TS Kd8. Mutation of the 2 binding residues of TSKd1 to H2-Kd1 resulted in a marked immunofocusing of CD8⁺ T cell responses. Panel D further shows that WT TS DNA vaccination induced higher levels of TS-specific antibody, compared with TSKd1 null DNA vaccination.

Fig 5. TSKd1 tolerization during WT TS DNA vaccination reduces CD8⁺ T cell responses directed against both immunodominant and subdominant TS epitopes.

BALB/c mice were injected with tERK-1 control or TSKd1 peptide i.v. on indicated days before and after i.m. vaccination with WT TS DNA followed by T. cruzi challenge as shown in panel A. One month following the final immunization, CD8⁺ splenic T cells from representative mice were stimulated with APC (A20) pulsed with TS peptides in overnight IFN-γ ELISPOT assays (B). Results are representative of 2 independent experiments. Other groups of tolerized TS DNA vaccinated mice were challenged with 5,000 T. cruzi BFT and survival monitored (C; 10–12 mice/group; p<0.05 comparing TS DNA vaccinated tolerized with TSKd1 or the tERK-1 peptide by Fisher exact tests).

Fig 6. TS enzymatic activity is associated with costimulatory effects.

In panel A, enzymatic activities of WT and TSKd1 mutant proteins were determined in lysates and supernatants of 293T cells transfected with WT TS and TSKd1 null DNA using fluorometric TS enzymatic activity assays. Panel B shows the co-stimulatory properties of WT rTS and TSKd1 null rTS on naïve CD8⁺ T cells. CD8⁺ T cells from naïve 4–5 week old BALB/c mice were purified by positive magnetic bead selection and incubated with suboptimal doses of PMA (12.5ng/ml) ± WT or TSKd1 null rTS. After 3 days, proliferation was measured by ³H-Thymidine incorporation. Shown are incorporated ³H-Thymidine counts per minute (CPM) above suboptimal PMA treatment alone. Similar TS costimulation assays using naïve CD8⁺ T cells were conducted with CD45 inhibitor PTP or Src-family kinase inhibitor PP2 added (C). In panel D, increasing concentrations of TSKd1 peptide were added to constant amounts of WT TS in TS enzyme assays. Results shown are representative of 2–4 independent experiments.

Fig 7. The WT TS vaccine-induced subdominant epitope responses are counterproductive for host immunity.

B16-Flt3L-induced and in vivo LPS matured CD11c⁺ dendritic cells were pulsed with TS peptides and injected i.v. into groups of naïve BALB/c mice (1x10⁶ DC, twice, 2 weeks apart). The peptide pool was comprised of the panel of K^d, D^d, and L^d-restricted epitopes differentially induced by WT TS and TSKd1 null TS vaccination (TS peptides Kd2, Kd5, Kd6, Kd7, Kd9, Ld1, and Dd2). Groups of control DC (no peptide) and TS peptide-pulsed DC vaccinated mice were challenged 4 weeks later with 5,000 T. cruzi BFT and survival monitored (A). Mice vaccinated with p7 + TSKd1 alone were significantly more protected than mice vaccinated with p7 + pool (+ TSKd1) by Log-Rank [Mantel-Cox] Test, p<0.0001. Shown are cumulative results from multiple independent experiments. In addition, CD8⁺ T cells from representative vaccinated mice harvested pre-challenge (one month following the final DC immunization) were stimulated in IFN-γ ELISPOT assays with APC (A20) pulsed with the immunodominant TSKd1 peptide (B).