Identification of human T cell antigens for the development of vaccines against Mycobacterium tuberculosis

Abstract

Development of a subunit vaccine for Mycobacterium tuberculosis (Mtb) depends on the identification of Ags that induce appropriate T cell responses. Using bioinformatics, we selected a panel of 94 Mtb genes based on criteria that included growth in macrophages, up- or down-regulation under hypoxic conditions, secretion, membrane association, or because they were members of the PE/PPE or EsX families. Recombinant proteins encoded by these genes were evaluated for IFN-gamma recall responses using PBMCs from healthy subjects previously exposed to Mtb. From this screen, dominant human T cell Ags were identified and 49 of these proteins, formulated in CpG, were evaluated as vaccine candidates in a mouse model of tuberculosis. Eighteen of the individual Ags conferred partial protection against challenge with virulent Mtb. A combination of three of these Ags further increased protection against Mtb to levels comparable to those achieved with bacillus Calmette-Guérin vaccination. Vaccine candidates that led to reduction in lung bacterial burden following challenge-induced pluripotent CD4 and CD8 T cells, including Th1 cell responses characterized by elevated levels of Ag-specific IgG2c, IFN-gamma, and TNF. Priority vaccine Ags elicited pluripotent CD4 and CD8 T responses in purified protein derivative-positive donor PBMCs. This study identified numerous novel human T cell Ags suitable to be included in subunit vaccines against tuberculosis.

Figure 1

Levels of IFN-γ released by antigen-stimulated human PBMC were measured in vitro. PPD⁺ and PPD^- PBMC were incubated for 72 h in medium, PHA (10 μg/ml), Mtb lysate (10 μg/ml), or Mtb recombinant proteins (50 μg/ml). Antigens are grouped according to the selection criteria: m, membrane-associated; S, secreted proteins; P, PE/PPE; E, EsX; M, macrophage growth required; H, hypoxic response; D, database searches. (A) IFN-γ in supernatants was measured by sandwich ELISA. Mean (Mean_Ag - Mean_Medium) ± SEM are shown for PPD⁺ (n=18) and PPD^- (n=7) PBMC. (B) Percent of PPD⁺ positive responders.

Figure 1

Levels of IFN-γ released by antigen-stimulated human PBMC were measured in vitro. PPD⁺ and PPD^- PBMC were incubated for 72 h in medium, PHA (10 μg/ml), Mtb lysate (10 μg/ml), or Mtb recombinant proteins (50 μg/ml). Antigens are grouped according to the selection criteria: m, membrane-associated; S, secreted proteins; P, PE/PPE; E, EsX; M, macrophage growth required; H, hypoxic response; D, database searches. (A) IFN-γ in supernatants was measured by sandwich ELISA. Mean (Mean_Ag - Mean_Medium) ± SEM are shown for PPD⁺ (n=18) and PPD^- (n=7) PBMC. (B) Percent of PPD⁺ positive responders.

Figure 2

Antibody responses to Mtb recombinant antigens were measured in sera isolated from mice vaccinated with Rv0577, Rv2875, Rv1886, Rv2608, Rv3478, Rv3619, Rv3620, Rv1813, and Rv3044 + CpG. Sera were tested for antigen-specific IgG1 and IgG2c antibody. (A) Mean of reciprocal dilution + SEM, and (B) IgG2c:IgG1 ratios are shown. *P < 0.05, Student’s t-Test. Data shown are representative of two independent experiments.

Figure 3

IFN-γ responses to Mtb recombinant antigens were determined in vaccinated mice. Splenocytes were prepared from animals immunized with Rv0577, Rv1626, Rv1813, Rv1886 (85B), Rv2875, Rv2608, Rv3619, Rv3620, Rv3044, and Rv3478 + CpG or saline, 3 wks after the last injection, and tested for IFN-γ in vitro recall responses. Cells were plated in duplicate at 2 × 10⁵ cells/well and cultured with medium, Con A (3 μg/ml), PPD (10 μg/ml), or each recombinant antigen (10 μg/ml) for 48 h. Frequencies of IFN-γ-secreting cells (spot-forming unit, SFU) were determined by ELISPOT. Mean + SD (n = 5 mice) shown are representative of two independent experiments.

Figure 4

TNF responses to Mtb recombinant antigens were determined in vaccinated mice. Splenocytes were prepared from animals immunized with Rv0577, Rv1626, Rv1813, Rv1886 (85B), Rv2875, Rv2608, Rv3619, Rv3620, Rv3044, and Rv3478 + CpG or saline, 3 wks after the last injection, and tested for TNF in vitro recall responses. Cells were plated in duplicate at 2 × 10⁵ cells/well and cultured with medium, Con A (3 μg/ml), PPD (10 μg/ml), or each recombinant antigen (10 μg/ml) for 48 h. Frequencies of TNF-secreting cells (SFU) were determined by ELISPOT. Mean + SD (n = 5 mice) shown are representative of two independent experiments.

Figure 5

Antigen-specific CD4 and CD8 T cell cytokine responses were analyzed in vaccinated mice. Splenocytes, isolated from animals immunized with Rv1813, Rv3620, Rv2608, Rv3619, Rv3478, or Rv1886 + CpG, or saline, were plated at 2 × 10⁶ cells/well and cultured with medium, or each recombinant antigen (20 μg/ml) for 12 h in the presence of GolgiStop. (A) CD4 T cells were identified by ICS based on CD3 and CD4 expression, and further gated as CD44^high (antigen experienced) or CD44^low (naïve). The percent of cells expressing two cytokines or more of IFN-γ, TNF, and IL-2 were determined. Means (n = 3) are shown. (B) CD8 antigen specific T cells were identified by ICS as CD3/CD44^high/CD8 cells also positive for IFN-γ and/or TNF, and further analyzed for their content in GrB. Numbers of cytokine⁺ (IFN/TNF) and cytokine⁺GrB⁺ per million splenocytes were determined. Mean ± SD (n = 3 mice) shown are representative of two independent experiments.

Figure 6

The efficacy of a combination of three antigens adjuvanted with CpG was addressed. Mice were immunized s.c. three times, three wks apart with 25 μg CpG, 8 μg single antigen + CpG, 6 μg of each antigen + CpG (“3 Ags+CpG”), or once i.d. with 5×10⁴ BCG. Four weeks after the last boost, mice were challenged with Mtb H37Rv. (A) The number of viable bacteria (CFU Log₁₀) in the lungs was determined 4 wks post-challenge. One way ANOVA followed by Dunnett’s Multiple Comparison Test was used for statistical analysis;* P < 0.05, ** P < 0.01. (B) Mice were bled one week after the last boost and sera were tested for antigen-specific IgG2c antibody. Mean of reciprocal dilution + SD (n = 3 mice) are shown. (C - D) Splenocytes were isolated from the vaccinated animals three wks after the last boost, plated in duplicate at 2 × 10⁵ cells/well, and cultured with medium, Con A (3 μg/ml), PPD (10 μg/ml), or each recombinant antigen (10 μg/ml) for 48 - 72 h. (C) IFN-γ levels measured by ELISA in 72 h supernatants. (D) Frequencies of TNF-secreting cells (SFU) determined by ELISPOT. Mean + SD (n = 3 mice) shown are representative of two independent experiments.

Figure 7

Human PPD⁺ CD4 T cells respond to recombinant Mtb antigen stimulation. PBMC from 6 healthy PPD⁺ subjects with diverse HLA types were incubated for 12 h in medium, PPD (10 μg/ml), or recombinant antigens (20 μg/ml) in the presence of anti CD28/CD49d and GolgiStop. T cells were identified by ICS based on CD3 expression, and further gated as CD4/CD45RO memory T cells. Percent of CD4 T cells expressing IFN-γ, TNF, or the two cytokines in response to a panel of Mtb antigens are shown.

Figure 8

Human PPD⁺ CD8 T cells respond to recombinant Mtb antigen stimulation. PBMC from 6 healthy PPD⁺ subjects with diverse HLA types were incubated for 12 h in medium, PPD (10 μg/ml), or recombinant antigens (20 μg/ml) in the presence of anti CD28/CD49d and GolgiStop. T cells were identified by ICS based on CD3 expression, and further gated as CD8/CD45RO memory T cells. Percent of CD8 T cells expressing IFN-γ, TNF, or the two cytokines in response to a panel of Mtb antigens are shown.