Serum indoleamine 2,3-dioxygenase activity is associated with reduced immunogenicity following vaccination with MVA85A

Abstract

Background: There is an urgent need for improved vaccines to protect against tuberculosis. The currently available vaccine Bacille Calmette-Guerin (BCG) has varying immunogenicity and efficacy across different populations for reasons not clearly understood. MVA85A is a modified vaccinia virus expressing antigen 85A from Mycobacterium tuberculosis which has been in clinical development since 2002 as a candidate vaccine to boost BCG-induced protection. A recent efficacy trial in South African infants failed to demonstrate enhancement of protection over BCG alone. The immunogenicity was lower than that seen in UK trials. The enzyme Indoleamine 2,3-dioxygenase (IDO) catalyses the first and rate-limiting step in the breakdown of the essential amino acid tryptophan. T cells are dependent on tryptophan and IDO activity suppresses T-cell proliferation and function.

Methods: Using samples collected during phase I trials with MVA85A across the UK and South Africa we have investigated the relationship between vaccine immunogenicity and IDO using IFN-γ ELISPOT, qPCR and liquid chromatography mass spectrometry.

Results: We demonstrate an IFN-γ dependent increase in IDO mRNA expression in peripheral blood mononuclear cells (PBMC) following MVA85A vaccination in UK subjects. IDO mRNA correlates positively with the IFN-γ ELISPOT response indicating that vaccine specific induction of IDO in PBMC is unlikely to limit the development of vaccine specific immunity. IDO activity in the serum of volunteers from the UK and South Africa was also assessed. There was no change in serum IDO activity following MVA85A vaccination. However, we observed higher baseline IDO activity in South African volunteers when compared to UK volunteers. In both UK and South African serum samples, baseline IDO activity negatively correlated with vaccine-specific IFN-γ responses, suggesting that IDO activity may impair the generation of a CD4+ T cell memory response.

Conclusions: Baseline IDO activity was higher in South African volunteers when compared to UK volunteers, which may represent a potential mechanism for the observed variation in vaccine immunogenicity in South African and UK populations and may have important implications for future vaccination strategies.

Trial registration: Trials are registered at ClinicalTrials.gov; UK cohort NCT00427830, UK LTBI cohort NCT00456183, South African cohort NCT00460590, South African LTBI cohort NCT00480558.

Figures

Figure 1

The immune response following vaccination with MVA85A is lower in South African than UK adults. Antigen specific IFN-γ ELISPOT responses to Ag85A overlapping peptide pools 0–24 weeks following vaccination with MVA85A are shown. A) MVA85A vaccinated adults (red = South African, blue = UK), n = 22 subjects per time point. *indicates a significant difference in response between UK and South Africa at the indicated time point (Mann Whitney Test *P < 0.05, **P < 0.005). B) The total immune response following vaccination with MVA85A is lower in South Africa when compared to the UK. Area under the curve (AUC) of the IFN-γ ELISPOT response to Ag85A peptides from 0–24 weeks was calculated for each volunteer. AUC values for each volunteer were then plotted by group (red = South African, blue = UK) n = 22 subjects per group. *indicates significant difference from week 0 (Mann Whitney Test P < 0.05).

Figure 2

IDO mRNA following MVA85A vaccination is produced by CD14+ monocytes and dependent on IFN-γ stimulation. A) PBMC from BCG vaccinated subjects boosted with MVA85A were cultured with 85A peptides or media only and the fold increase in IDO mRNA expression over time was determined for each subject (n = 8). Ag85A-induced IDO mRNA expression was significantly greater at weeks 1 and 4 when compared to week 0 (Wilcoxon *P < .05). B) The increase in IDO mRNA expression correlated with the IFN-γ ELISPOT response (data from week 1 following MVA85A) (Spearman’s correlation P < .05, r = 0.79). C) PBMC from BCG vaccinated subjects boosted 4 weeks previously with MVA85A were cultured with recombinant human (rh) IFN-γ, 85A peptides or 85A peptides and anti-IFN-γ antibodies and the fold change in expression in stimulated compared to unstimulated PBMC was determined (n = 5). Recombinant human IFN-γ and 85A peptide stimulation induced the expression of IDO mRNA. Co-culturing cells with 85A peptides and anti-IFN-γ antibodies resulted in significant reduction of 85A specific IDO mRNA expression (Wilcoxon *P < .05). D) CD14+ magnetic beads were used to deplete monocytes from total PBMC. IDO expression was enriched in the CD14+ fraction and depleted when CD14+ cells were removed (n = 5). Mann Whitney was used for comparison between groups (**P < .005).

Figure 3

Serum IDO activity is higher in South African compared to UK adults. A) Serum IDO activity was measured in serum using LC-MS to quantity L-Trp and L-Kyn in participant serum. The ratio of L-Trp/L-Kyn gives a measure of IDO enzyme activity in host serum. B) IDO activity was assessed in serum from UK and South African adults at baseline (pre-vaccination, week 0) and 1 week following immunisation with MVA85A. Mann Whitney was used for comparison between groups (*P < .05).

Figure 4

Serum IDO activity is negatively correlated with the 85A peptide specific IFN-γ ELISPOT response. IDO activity at baseline for both UK and South African volunteers was correlated with the 85A peptide specific IFN-γ ELISPOT response at A) week 4 and B) week 24 following vaccination with MVA85A (Spearman’s correlation P < .05).