Double- and monofunctional CD4+ and CD8+ T-cell responses to Mycobacterium tuberculosis DosR antigens and peptides in long-term latently infected individuals

Authors

  • Susanna Commandeur,

    1. Department of Infectious Diseases, Leiden University Medical Centre, Leiden, The Netherlands
    Search for more papers by this author
    • These three authors cotributed equally to this work.

  • May Y. Lin,

    1. Department of Infectious Diseases, Leiden University Medical Centre, Leiden, The Netherlands
    Current affiliation:
    1. Department of Infectious Diseases and Immunology, University of Utrecht, Utrecht, The Netherlands
    Search for more papers by this author
    • These three authors cotributed equally to this work.

  • Krista E. van Meijgaarden,

    1. Department of Infectious Diseases, Leiden University Medical Centre, Leiden, The Netherlands
    Search for more papers by this author
    • These three authors cotributed equally to this work.

  • Annemieke H. Friggen,

    1. Department of Infectious Diseases, Leiden University Medical Centre, Leiden, The Netherlands
    Search for more papers by this author
  • Kees L. M. C. Franken,

    1. Department of Infectious Diseases, Leiden University Medical Centre, Leiden, The Netherlands
    Search for more papers by this author
  • Jan W. Drijfhout,

    1. Department of Immunohematology and Blood Transfusion, Leiden University Medical Centre, Leiden, The Netherlands
    Search for more papers by this author
  • Gro E. Korsvold,

    1. Department of Bacteriology and Immunology, Division of Infectious Disease Control, Norwegian Institute of Public Health, Oslo, Norway
    Search for more papers by this author
  • Fredrik Oftung,

    1. Department of Bacteriology and Immunology, Division of Infectious Disease Control, Norwegian Institute of Public Health, Oslo, Norway
    Search for more papers by this author
  • Annemieke Geluk,

    1. Department of Infectious Diseases, Leiden University Medical Centre, Leiden, The Netherlands
    Search for more papers by this author
  • Tom H. M. Ottenhoff

    Corresponding author
    1. Department of Infectious Diseases, Leiden University Medical Centre, Leiden, The Netherlands
    • Department of Infectious Diseases, Leiden University Medical Centre, Leiden, The Netherlands Fax: +31-71-526-6758
    Search for more papers by this author

Abstract

More than 2 billion individuals are latently infected with Mycobacterium tuberculosis (Mtb). Knowledge of the key Mtb antigens and responding T-cell subsets mediating protection against Mtb is critical for developing improved tuberculosis (TB) vaccines. We previously reported that Mtb DosR-regulon-encoded antigens are recognized well by human T cells in association with control of Mtb infection. The characteristics of the responding T-cell subsets, however, remained unidentified. We have therefore studied the cytokine production and memory phenotypes of Mtb DosR-regulon-encoded antigen-specific T cells from individuals who had been infected with Mtb decades ago, yet never developed TB (long-term latent Mtb-infected individuals). Using multi-parameter flow cytometry and intracellular cytokine staining for IFN-γ, TNF-α and IL-2, we found double and single cytokine-producing CD4+ as well as CD8+ T cells to be the most prominent subsets, particularly IFN-γ+ TNF-α+ CD8+ T cells. The majority of these T cells comprised effector memory and effector T cells. Furthermore, CFSE labeling revealed strong CD4+ and CD8+ T-cell proliferative responses induced by several “immunodominant” Mtb DosR antigens and their specific peptide epitopes. These findings demonstrate the prominent presence of double- and monofunctional CD4+ and CD8+ T-cell responses in naturally protected individuals and support the possibility of designing Mtb DosR antigen-based TB vaccines.

Introduction

Host defense against mycobacteria critically depends on effective innate and adaptive immunity, culminating in the activity of Mycobacterium tuberculosis (Mtb)-specific T cells and in the formation of granulomas that contain Mtb bacilli. Both CD4+ and CD8+ T-cell responses are involved, and it is undisputed that Th1- and Th17-like cytokines (IL-12, IFN-γ, TNF-α and IL-17) are crucial for optimal host immunity 1, 2. Tuberculosis (TB) continues to claim almost 2 million lives each year, and causes active (infectious) TB disease in over 9 million new cases per annum. Control of TB is further impeded by the strong increase in TB morbidity and mortality due to HIV co-infection, and the rise of multi-drug resistant and extensively drug-resistant Mtb strains 3. At least 2 billion people are latently infected with Mtb, representing a huge reservoir of latently infected individuals from which most new TB cases arise. While 90–98% of all Mtb-infected individuals are able to contain infection asymptomatically in a latent state, 2–10% of these Mtb-infected individuals will progress towards developing TB during their lifetime.

Despite strong international efforts in TB vaccine development, Mycobacterium bovis Bacillus Calmette-Guérin (BCG) continues to be the only available TB vaccine. BCG vaccination induces effective protection against severe TB in young children and protects against leprosy, but does not provide sufficient protection against the severe and contagious form of TB; pulmonary TB in adults 4, 5. Moreover, BCG does not protect against TB reactivation later in life. Ideally, not only improved preventive vaccines with pre-exposure activity but also therapeutic vaccines with post-exposure activity during late-phase infection are urgently required 2, 6. Such vaccines should prevent reactivation of TB from latency by inducing and maintaining robust immunity to Mtb antigens that are expressed by persisting Mtb bacilli during latent infection. Such immune responses may not only help controlling but perhaps also eradicating persisting bacilli.

Work from our group has shown that several genes from the recently identified Mtb DosR (Rv3133c) regulon encode particular antigens that induce significant T-cell responses in Mtb-infected individuals (latently infected or active TB disease) 7. This 48-gene DosR-regulon is expressed by Mtb during in vitro exposure to hypoxia, low-dose nitric oxide and carbon monoxide, conditions thought to be encountered by Mtb in vivo when persisting in immunocompetent hosts 8. Approximately half of the Mtb dosR-regulon genes are also expressed over prolonged periods of time in a related stress model, the enduring hypoxia response model 9. Of note, immunity to Mtb DosR-regulon-encoded antigens is associated with control of latent Mtb infection, as several DosR-regulon-encoded antigens are preferentially recognized by individuals with latent Mtb infection 7, 10, 11. Thus, enhancing immune responses to these antigens might contribute towards controlling persistent Mtb infection with the potential to help prevent reactivating TB disease.

The precise nature of the human T-cell response to Mtb DosR-regulon-encoded antigens has not been studied in detail thus far. Most studies have documented IFN-γ production in response to Mtb DosR antigens, but the major cellular source(s) of the produced IFN-γ have not been identified, neither was concomitant production of other cytokines assessed 7, 12–14.

In this study, we show that Mtb DosR-regulon-encoded antigens induce antigen and peptide specific, double and single cytokine-producing CD4+ and CD8+ T cells in elderly persons who had been infected with Mtb decades ago in the pre-antibiotic era, yet never developed TB (designated here as long-term latent Mtb-infected individuals (ltLTBIs)). Among the responding cells, IFN-γ+TNF-α+ CD8+ T cells were highly prevalent, the majority being effector memory (CCR7CD45RA) or effector (CCR7CD45RA+) T cells. Furthermore, a register of peptide epitopes recognized by both CD4+ and CD8+ T cells was identified for several Mtb DosR-regulon-encoded antigens, which are potently recognized in humans 7. Collectively, these results underscore the importance of Mtb DosR antigens and their association with control of latent Mtb infection.

Results

Selection of recombinant Mtb DosR antigens

Our previous work showed that Mtb DosR-regulon-encoded antigens are efficiently recognized by Mtb-exposed individuals, particularly asymptomatic tuberculin skin test-positive individuals 7, 12, 13. To study the nature of the response against these antigens in more detail, we selected Rv1733c and Rv2029c as two Mtb DosR proteins consistently ranking among the top ten most frequently recognized Mtb DosR antigens in Mtb-exposed individuals across different ethnic populations 7, 12, 13. The secreted protein Ag85B and the Mtb DosR antigen Rv2031c (HspX, hsp16, α-crystallin) were included as control antigens 15–17. Besides recombinant proteins (Table 1), overlapping sets of synthetic peptides of all four antigens were produced and tested as well (Supporting Information Table S1A–D).

Table 1. Selected Mycobacterium tuberculosis antigens tested in present study
 Rv numberGene namea)Molecular mass (kDa)a)Producta)References
DosR genesRv1733c 22,4Conserved transmembrane protein7, 11–13
 Rv2029cpfkB35,4Phosphofructokinase PfkB7, 11, 12
 Rv2031chspX16,3Heat shock protein HspX (α-crystallin)7, 10, 15–17
Reference geneRv1886cfbpB34,6Secreted antigen 85-B fbpB (mycolyltransferase 85B, Ag85B)31

Single, double and polyfunctional T-cell responses against Mtb DosR antigens in ltLTBIs

Mtb DosR antigen-specific T-cell recognition described previously in Mtb-exposed donors was based mostly on the production of IFN-γ. Previously, polyfunctional T cells producing IFN-γ, TNF-α and IL-2 have been suggested as possible markers of protective immunity, based on observations that vaccine-induced triple positive T cells correlated well with protection 18–24. However, other studies reported that such T cells were associated with active TB disease 25–28.

The nature of Mtb DosR antigen-responsive CD4+ and CD8+ T-cell subsets in untreated Mtb-exposed donors who had been infected several decades ago, yet never developed any signs or symptoms of active TB (ltLTBIs), was studied here. In vitro purified protein derivative of Mtb (PPD) negative (PPD) donors were included as uninfected controls. PBMCs of ltLTBIs and PPD donors were stimulated with Mtb DosR-regulon-encoded antigens or corresponding peptide pools and the responses were analyzed using multi-parameter flow cytometry (Supporting Information Fig. S1A and S1B). Donors were considered positive when the frequency of a double or poly functional T-cell subset population was ≥0.2%, which is equivalent to ≥200 events.

In ltLTBIs high percentages of IFN-γ, TNF-α and/or IL-2 cytokine-producing CD4+ and CD8+ T cells were found in response to PPD (0.23–7.91% and 0.25–7.55%, respectively), Rv2031c protein (0.21–19.71% and 0.25–20.35%, respectively) and the Rv2031c peptide pool (0.2–16.28% and 0.23–32.92%, respectively), whereas no such responses were observed in PPD controls (Fig. 1A). The highest frequencies were consistently found within the single cytokine-producing CD4+ and CD8+ T-cell populations. Interestingly, many double producing T cells were identified within the CD8+ T-cell population, as shown by Fig. 1B, which depicts the proportions of polyfunctional as well as double and single cytokine-producing T cells. For Mtb DosR antigen Rv1733c, two peptide pools were tested (Fig. 1C). Again high CD4+ and CD8+ T-cell responses were observed (0.43–14.41% and 0.2–14.25%, respectively), with single positive cells being the most frequent. In addition, substantial numbers of double cytokine-producing CD4+ and CD8+ T cells were present in both peptide pool responsive CD4+ and CD8+ T-cell populations, IFN-γ+TNF-α+ CD8+ T cells being the most frequent (Fig. 1D). Low to no Rv1733c-specific responses were identified within the PPD controls (Fig. 1C). A comparable pattern was observed for Rv2029c (0.29–8.41% CD4+ T cells and 0.36–9.55% CD8+ T cells). Unlike Rv1733c, the Rv2029c protein induced a considerable fraction of IFN-γ+TNF-α+ CD8+ T cells. Some responses to Rv2029c peptide pool 1 were also observed in the PPD group, but no responses were seen to peptide pools 2 and 3 (Fig. 1E and F). Of note, stimulation of PBMCs with Staphylococcus enterotoxin B induced high percentages of CD4+ and CD8+ T cells producing single (0.3–26.44% CD4+ T cells and 0.29–12.6% CD8+ T cells), double (0.23–22.26% CD4+ T cells and 0.24–20.17% CD8+ T cells) and triple (0.29–5.37% CD4+ T cells and 0.54–6.91% CD8+ T cells) cytokines in both ltLTBIs and PPD donors (data not shown).

Figure 1.

Frequency of antigen-specific polyfunctional T cells in ltLTBIs. Frequency of antigen-specific CD4+ and CD8+ T cells in ltLTBIs (upper panel) and PPD control donors (lower panel), producing combinations of IFN-γ, TNF-α or IL-2 after stimulation for 16 h with control antigens as determined by flow cytometry. (A) Stimulation with PPD, Rv2031c protein and Rv2031c peptide pool 1, n=13 and n=11, (C) stimulation with Rv1733c protein and corresponding peptide pools 1 and 2, n=11 and (E) stimulation with Rv2029c protein and corresponding peptide pools 1–3, n=11. CD4+ T cells are indicated as closed circles (•) and CD8+ T cells as open circles (○). Horizontal bars represent the median frequency of antigen-specific CD4+ and CD8+ T cells. (B, D and F) Pie chart representation of the proportion of single-, double- or triple-positive CD4+ and CD8+ T cells for each antigen. Only CD4+ and CD8+ populations of ≥2×105 events were analyzed. Donors were considered positive when the frequency of a double or polyfunctional T-cell subset population was ≥0.2%, which corresponds to ≥200 events. Range total CD4+ T cells: 0–3694 events and total CD8+ T cells: 0–3937 events.

Figure 1.

Continued.

Interestingly, the IFN-γ+TNF-α+ CD8+ T-cell population consistently was the most frequent multiple cytokine-producing T-cell subset identified (Fig. 1B, D and F). To assess the memory phenotype of these cells, we measured expression of memory markers CCR7 and CD45RA by Mtb antigen or peptide responsive cells from the ltLTBI population (Fig. 2A and B). T-cell subsets were classified according to the model described by Seder et al. 29. Only a minor fraction of the IFN-γ+TNF-α+ CD8+ T cells appeared to be “naïve” (CCR7+CD45RA+) or central memory T cells (CCR7+CD45RA), while most were found to be effector memory (CCR7CD45RA) T cells, followed by effector (CCR7CD45RA+) T cells (percentages ranged between 36 and 62% (SD±0–35) for effector memory T cells and 22–51% (SD±2.8–32) for effector T cells).

Figure 2.

T-cell memory subset distribution of IFN-γ+TNF-α+ CD8+ T cells. Expression of T-cell memory markers CCR7 and CD45RA was analyzed by flow cytometry in IFN-γ+ TNF-α+ CD8+ T cells of (A) PPD responders (n=4) and (B) Mtb antigen or peptide pool responders from the ltLTBI donors. Rv2031c protein (n=3), Rv2031 peptide pool 1 (n=2), Rv1733c protein (n=2), Rv1733c peptide pool 1 (n=2), Rv1733c peptide pool 2 (n=3), Rv2029c protein (n=3), Rv2029c peptide pool 1 (n=4), Rv2029c peptide pool 2 (n=3) and Rv2029c peptide pool 3 (n=3). CD8+ populations with ≥2×105 events were analyzed. Effector memory T cells are CCR7 and CD45RA, central memory T cells are CCR7+ and CD45RA, naive T cells are CCR7+ and CD45RA+ and effector T cells are CCR7 and CD45RA+.

Taken together, our results show the presence of Mtb DosR-regulon-encoded antigen-specific double- and monofunctional CD4+ and CD8+ T-cell responses in ltLTBIs. IFN-γ+TNF-α+ CD8+ T cells were the most prominently present multiple cytokine-producing T cells, and comprised mainly effector memory and effector T cells.

Peptide epitope mapping of Mtb DosR antigens by CFSE proliferation and flow cytometry

Next, we analyzed single peptide-induced responses in PPD positive (PPD+) individuals in order to identify immunogenic Mtb DosR antigen epitopes. In view of the number of cells required for these analyses, we used buffy coat-derived PBMCs. PBMCs of PPD+ individuals were incubated with each single peptide of Mtb DosR Rv1733c, Rv2029c and Rv2031c and the control protein Ag85B. Proliferative responses were measured using CFSE labeling, an assay that we have described previously 27, 30. Figure 3 demonstrates typical proliferation profiles of CD4+ and CD8+ T cells in response to Mtb antigens and control conditions in one PPD+ donor. Following stimulation of PBMCs with PPD, Rv1733c or its corresponding peptides, significant CD4+ and to a lesser extent CD8+ T-cell proliferation were observed (Fig. 3A and B, respectively). No proliferation was observed to the irrelevant control peptide HIV-gag77–85 or for medium only (data not shown). A relative proliferation (see Materials and methods for calculation) of 10% was considered positive in this assay, in line with previous studies 27, 30. Responses to previously published HLA class I and class II restricted epitopes of Ag85B 31 and Rv2031c 17, 28, 32–34 could be confirmed, validating this approach (Fig. 3A and B).

Figure 3.

T-cell proliferation in response to Mtb antigens, peptides and control conditions. PBMCs from PPD responsive donor 9 were stimulated with PHA or HIVgag77–85 as positive and negative controls, respectively, or with recombinant Mtb antigens. After 6 days, cellular proliferation as measured by CFSE dye dilution was measured in (A) CD4+ or (B) CD8+ T cells. CD4+ and CD8+ T cells were gated from a CD3+ T-cell gate together with a live lymphocyte gate. Individual histogram plots show the percentage of relative proliferation. PBMCs were stimulated with (A) Rv1733c protein and corresponding peptides and (B) peptides of Ag85B and Rv2031c, containing HLA-A2 restricted epitopes.

Results for CFSE-labeled PBMCs from all 15 PPD+ donors in response to PPD, Mtb DosR-regulon-encoded proteins Rv1733c, Rv2029c and Rv2031c and Ag85B protein and all respective single peptides from each of the four antigens are given in Fig. 4A and B, showing comprehensive epitope maps for CD4+ (Fig. 4A) or CD8+ (Fig. 4B) T cells. Notwithstanding the expected inter-individual differences in antigen and peptide recognition patterns, all 15 PPD+ donors responded to at least one of the recombinant proteins and/or peptides investigated. Recombinant antigens Rv1733c, Rv2029c and Rv1886c (Ag85B) were recognized efficiently: 7/15 PPD+ donors recognized Rv2029c (CD4+: 15–97.2%, CD8+: 10.6–66.6%), 5/15 recognized Rv1733c (CD4+: 20.3–40%, CD8+: 12.2–31.1%) and 4/15 recognized Ag85B (CD4+: 13.8–53.4%, CD8+: 12.6–97.7%). Corresponding to our previous observations, Rv2031c/hspX/acr was recognized by a minority of the donors (CD4+: 10.9–16.4%, CD8+: 42.7%) 7, 12.

Figure 4.

Proliferative CD4+ and CD8+ T-cell responses to Mtb DosR antigen epitopes among PPD+ individuals. PBMCs of 15 PPD+ donors were stimulated with either Mtb DosR antigen protein, corresponding single peptides or control antigen and peptides of Ag85B. Relative proliferation of (A) CD4+ and (B) CD8+ T cells is indicated in grey tones. Relative proliferation was calculated as follows; ((Δ geometric mean sample − Δ geometric mean control medium)/(Δ geometric mean PHA − Δ geometric mean control medium))×100%=% of maximal proliferation. Light grey; less than 3% proliferation, grey; between 3 and 10% proliferation, dark grey; over 10% proliferation. A relative proliferation of ≥10% was considered positive.

A substantial number of peptides was recognized by CD4+ and CD8+ T cells for Rv1733c (CD4+: 17/20 (10.1–76.9%) CD8+: 12/20 (10.4–100%)), Rv2029c (CD4+: 25/33 (10.4–100%) CD8+: 14/33 (10.3–66.6%)), Rv2031c (CD4+: 12/14 (10.2–53.8%) CD8+: 5/14 (11.3–42.7%)) and Ag85B (CD4+: 28/30 (10.1–75.3%) CD8+: 25/30 (10.9–97.7%)). Some peptides were recognized by CD4+ T cells from more than one-third of the donors (e.g. 6/15 donors in case of Ag85B peptides 9 and 13, and 5/15 for Ag85B peptides 5, 6), whereas other peptides were recognized by CD4+ T cells in 4/15 donors, such as Rv1733c peptide 2 and Ag85B peptides 10, 12, 16 and 22. CD8+ T-cell responses were particularly observed against Rv1733c and Ag85B; these responses were found in four to five donors; Rv1733c peptides 17 (5/15), 2 and 19 (4/15), and Ag85B peptides 5 and 13 (4/15). Notably, some peptides were recognized by both CD4+ and CD8+ T cells (Rv1733c peptide 2, Ag85B peptides 5 and 13). Table 2 shows the cumulative epitope recognition map for both CD4+ and CD8+ T cells in response to all tested proteins and peptides for all donors tested. Interestingly, the results suggest enrichment of epitopes in certain immunogenic regions, for example Rv1733c(1–40), Rv1733c(161–200) and Ag85B(81–180), which harbor Rv1733c peptides 1–3, 17–19 and Ag85B peptides 5–14.

Table 2. Cumulative epitope recognition of CD4+ and CD8+ T cells to Mtb antigens and corresponding peptides
inline image

The above-described Mtb DosR antigen-encoded peptide epitopes were recognized by donors with varying HLA genotypes. Many of the in vitro responses given in Fig. 4A and B matched with in silico epitope motif searches for the relevant HLA genotypes (data not shown) 35. This suggests that responses to Mtb dosR-regulon-encoded antigens occur in a wide range of HLA backgrounds. In order to better characterize the molecular interactions of Mtb DosR antigenic epitope presentation, we examined peptide recognition in the context of the highly frequent HLA-A*0201 genotype (New allele frequency database: http://www.allelefrequencies.net36) and found that Rv1733cp181–189 specific CD8+ T cells were able to lyse peptide loaded and endogenously processed Rv1733c-antigen loaded target cells in the context of HLA-A*0201 molecules (Supporting Information Fig. S2A and S2B).

Discussion

We have proposed that Mtb DosR-regulon-encoded antigens 7 that are expressed by Mtb during in vitro conditions mimicking intracellular infection represent rational targets for TB vaccination. Immune responses to Mtb DosR-regulon-encoded antigens are prominently found in latently infected individuals, and are associated with latent Mtb infection (LTBI) in several ethnically and geographically distinct populations 7, 11, 13. Additional work showed that the Mtb DosR-regulon-encoded antigen Rv2628 was strongly recognized by individuals with remote Mtb infection 13, 14. Thus far, the precise mechanisms and T-cell subsets responsible for the responses against Mtb DosR-regulon-encoded antigens have not been studied in detail; and virtually all studies have relied on measuring IFN-γ production by polyclonal cells. Here, we report peptide reactivity and memory phenotypes of Mtb DosR-regulon-encoded antigen-specific T cells in long-term LTBI, and moreover, document a large series of specific peptide epitopes recognized by specific CD4+ and CD8+ T cells.

Three Mtb DosR antigens, Rv1733c, Rv2029c and Rv2031c (HspX, α-crystallin) were tested in this study. Strong Mtb DosR antigen-specific CD4+ and CD8+ polyfunctional T-cell responses were detected in ltLTBIs. The highest responses were observed among single cytokine-producing CD4+ and CD8+ T-cell subsets (either TNF-α+, IL-2+ or IFN-γ+, depending on the stimulus) followed by double producing CD4+ and particularly CD8+ T cells. Of interest, the most frequent multiple cytokine-producing T cells were IFN-γ+TNF-α+ CD8+ T cells. These cells were further characterized as effector memory (CCR7 and CD45RA) or effector (CCR7 and CD45RA+) T cells, which have the ability to perform immediate effector functions. This is compatible with an important role for CD8+ T cells in Mtb infection 37, 38.

Mtb antigen-specific polyfunctional T cells have been studied intensely the last few years, both in vaccination and in observational studies in Mtb-infected individuals 18–29, 39. There is currently no consensus whether polyfunctional T cells represent a marker of protective immunity or of disease activity. The vaccine MVA85A (recombinant replication-deficient vaccinia Ankara, expressing Ag85A) induced polyfunctional CD4+ and CD8+ T cells producing IFN-γ, TNF-α and IL-2 as well as IFN-γ and TNF-α in mice, which correlated with TB protection 19. This vaccine also induced increased CD4+ T cells expressing IFN-γ, TNF-α and IL-2 in humans when given as a booster to previous BCG vaccination 20, 21. Similar results were reported following human vaccination with the BCG booster AERAS-402 (recombinant replication-deficient Adenovirus (Ad35) virus, expressing a polyprotein of Ag85A, Ag85B and TB10.4) 22. Finally, mice vaccinated with hybrid subunit vaccines H1 (Ag85-ESAT6) and H56 (H1+Rv2660) also had high numbers of triple cytokine-producing CD4+ T cells 23, 24. However, observational studies in humans have associated polyfunctional CD4+ T cells with TB disease 25, 26. Although limited information is available on Mtb-specific polyfunctional CD8+ T cells, two of our recent studies showed that double- and single-producing CD8+ T cells are prominent in latently Mtb-infected individuals and cured TB patients and that IFN-γ+/IL-2+ CD8+ T cells are a biomarker of protective host defense against Mtb27, 28. Moreover, no difference in polyfunctional CD4+ T-cell profiles could be identified between BCG-vaccinated children from a high endemic area that either developed TB or did not, indicating that polyfunctional T cells are not a biomarker of BCG-induced protection against TB 39. In our study here we show the presence of mostly single and double positive T cells, the latter mainly present in CD8+ T cells, supporting previous findings that single and double positive T cells are prominent in LTBI 25, 28. This suggests that these double and single cytokine-producing T cells play a significant role in Mtb immunity, although their precise nature and mechanisms of action requires more detailed studies.

While most studies on polyfunctional T cells have focused on highly expressed Mtb early phase proteins such as ESAT6 and Ag85B, instead, we here have analyzed Mtb antigens that are expressed during dormancy. It remains possible that antigens expressed during different phases of infection may preferentially induce different patterns of single, double and polyfunctional T cells.

A striking observation was the wealth of epitopes that could be identified in Mtb DosR-regulon-encoded antigens, in accordance with the significant immunogenicity of Mtb DosR-regulon-encoded antigens in a wide variety of HLA backgrounds 40. The donors used to detect single peptide responses were anonymous Dutch blood bank donors. Although we have no precise information about their mycobacterial exposure status, we have shown previously that over 50% of blood bank donors respond to PPD; furthermore, responses to Mtb DosR antigens were also observed in nontuberculous mycobacteria-exposed donors, probably due to the high conservation of these antigens 41.

Within several Mtb DosR-regulon-encoded antigens highly immunogenic regions could be identified and a substantial number of peptides elicited both CD4+ and CD8+ T-cell responses. Although HLA-class I presented peptides are typically 8–11 amino acids long, whereas HLA-class II ligands can be between 10 and 25 amino acids 35, 42, 43, we nevertheless found efficient CD8+ T-cell responses using 20-mers and confirmed Rv1733c-specific lysis of target cells by Rv1733cp181–189 specific CD8+ T cells. It has been suggested that apoptosis, induced by the cytotoxic activity of CTL, can inhibit Mtb growth or even kill Mtb bacteria 44–46. Granulysin may play a role in this mycobactericidal activity 47. In addition, vaccine-induced CD8+ T cells in mice indeed can reduce bacterial load in vivo 48. Again, this suggests a protective role for CD8+ T cells in Mtb infection.

Our 6–10 day incubation period may have allowed internalization and processing of peptides for HLA-class I presentation or allowed cross-presentation via alternative antigen presentation pathways 49, 50. Peptides appeared to be more sensitive tools for detecting CD4+ or CD8+ T-cell responses compared to recombinant proteins; this is likely due to the preprocessed nature of peptides, facilitating highly efficient antigen presentation. Second, peptides were present at much higher molar concentrations since proteins and peptides were tested at 10 μg/mL, regardless of their molecular mass. The lack of competition for processing, with otherwise dominant epitopes in recombinant proteins, may also have permitted identification of subdominant epitopes using peptides. Thus, peptide-based epitope mapping also offers the potential to elucidate subdominant epitopes, which might be exploited in designing improved vaccines by inducing immunity to a broader epitope repertoire than would be seen following natural infection or protein vaccination 51, 52. Of note, previous work has shown the efficacy of vaccines containing subdominant epitopes in protection against infection with Mtb53.

In conclusion, we report the presence of Mtb DosR-regulon-encoded peptide antigen-specific single and double functional CD4+ and CD8+ T-cell responses in ltLTBIs. We show that the majority of multiple cytokine-producing T cells comprise IFN-γ+TNF-α+ CD8+ T cells; these cells were characterized as mainly effector memory or effector T cells. Furthermore, we describe a large series of new peptide epitopes expressed by Mtb DosR-regulon-encoded antigens, which are recognized by CD4+ and/or CD8+ T cells of PPD+ donors. These results significantly enhance our understanding of the human immune response to Mtb phase-dependent antigens in long-term control of infection, and pave the way for designing Mtb DosR antigen and/or peptide-based vaccination approaches to TB.

Materials and methods

Study subjects

We studied PBMCs derived from a Norwegian group that had been exposed to Mtb decades ago, but had never developed TB despite lack of any treatment. This population was designed as long-term LTBI (ltLTBIs) (n=13). Their ages ranged from 62 to 74 years (average 70 years) with tuberculin skin test indurations ranging from 12 to 60 mm (average 18 mm). About 77% (10/13) of the Norwegian donors tested positive for Quantiferon® TB Gold (Cellestis Carnegie, Victoria, Australia). PBMCs of healthy PPD negative (PPD) blood bank donors were used as negative controls. Donors were considered PPD negative when IFN-γ responses to PPD was <100 pg/mL. For the second study, buffy coats from 21 in vitro PPD responsive (PPD+) healthy anonymous, HLA-typed blood bank donors were included. PPD responding donors were considered positive when IFN-γ responses (corrected for background values) to PPD exceeded 100 pg/mL, in line with our previous studies 7, 54, 55. Buffy coats were used since the number of cells derived from that source allowed us to perform experiments in which the Mtb DosR antigen and all single peptides could be tested simultaneously. All donors were HIV-negative and written informed consent was obtained prior to venipuncture. The study protocol (P027/99) was approved by the Regional Committees for Medical and Health Research Ethics in Norway and the Institutional Review Board of the Leiden University Medical Center. PBMCs were isolated by standard Ficoll density gradient centrifugation using Leucosep® tubes (Greiner, Bio-one, Alphen aan den Rijn, The Netherlands). PBMCs were collected and stored in liquid nitrogen until use.

Mtb recombinant antigens and synthetic peptides

Recombinant proteins were produced as described previously 56. In short, PCR was used to amplify the selected Mtb H37Rv genes from genomic H37Rv DNA. The PCR products were cloned using Gateway Technology (Invitrogen, San Diego, CA, USA) and were subsequently sequenced. Escherichia coli strain BL21 (DE3) was used to over-express Mtb proteins. Recombinant proteins were further purified as described previously 56. All recombinant proteins were tested in quality control assays including, size and purity check, determination of residual endotoxin levels as well as non-specific T-cell stimulation and cellular toxicity in lymphocyte stimulation assays 55. PPD (batch RT49) was purchased from Statens Serum Institute (Copenhagen, Denmark).

Synthetic peptides were synthesized as previously described 57. Peptides from Mtb DosR antigens Rv1733c, Rv2029c, Rv2031c and control antigen Ag85B were 20-mers peptides with 10 aa overlap, except peptides 20–22 of Ag85B which were 15-mers with 10 aa overlap (Table S1A–D). The 20-mer peptides of Rv1733c and Rv2029c were elongated with two lysine (K) residues at the C-terminal to improve solubility. The HLA-A*0201-restricted, HIV-1 p17 Gag77–85 epitope (SLYNTVATL) was used as control peptide 58.

Functional T-cell analysis

T-cell phenotype analysis was performed as previously described 59. In brief, PBMCs were stimulated for 16 h with protein (10 μg/mL) or peptide pools (5 μg/mL) in the presence of co-stimulatory antibodies anti-CD28/anti-CD49d (Sanquin, The Netherlands and BD Biosciences respectively). After 4–6 h, Brefeldin A (3 μg/mL; Sigma) was added to the culture. Cell surface staining was performed for the following markers; CD3-PB, CD4-PercP/Cy5.5, CD8-AmCyan, CD45RA-PE/Cy5, CD25-APC/Cy7 and CCR7-PE/Cy7. Subsequently, intracellular markers were stained with IFN-γ-Alexa700, TNF-α-APC, IL-2-PE and CD69-FITC (BD Biosciences) using Intrastain kit (Dako Cytomation, Denmark). Samples were acquired on an LSRII. CD4+ and CD8+ populations of ≥2×105 events were analyzed using FlowJo (Treestar, Ashland, OR, USA) and SPICE software (provided by Dr. Mario Roederer, Vaccine Research Center, NIAID, NIH, USA). Boolean gate analysis was used to study the different single, double and polyfunctional CD4+ and CD8+ T cells.

T-cell proliferation

Proliferation was measured using carboxy-fluorescein diacetate, succinimidyl ester (CFSE) dilution and flow cytometry. PBMCs from study subjects were thawed, washed and labeled with CFSE (Molecular Probes, Leiden, The Netherlands) at a final concentration of 5 μM for 10 min at 37°C. Washed, counted and viable cells were seeded in triplicates in 96-well round-bottom plates at a concentration of 1.5×105 cells/well in the presence of control antigens (PPD 5 μg/mL, PHA 2 μg/mL (Remel, Oxoid, Haarlem, The Netherlands)) or test antigens and peptides (all at final concentrations of 10 μg/mL). Cells were cultured in IMDM supplemented with glutamax, 100 U/mL penicillin, 100 μg/mL streptomycin (Invitrogen, Breda, The Netherlands) and 10% human serum at 37°C and 5% CO2. After 7 days of incubation, cells were stained for further analysis on the flow-cytometer.

Cells were stained for the following surface markers; CD8-APC (DakoCytomation, Heverlee, Belgium), CD3-PerCP and CD4-PE (BD Biosciences), washed in PBS 0.1% BSA (Sigma Aldrich, Zwijndrecht, The Netherlands), fixed in 1% paraformaldehyde (Pharmacy LUMC, The Netherlands) and acquired on an LSRII with HTS plate loader (BD Biosciences). Analysis was performed using FACS DIVA software (BD Biosciences). Live lymphocyte gated cells combined with gating of CD3+ CD4+ and CD3+ CD8+ T cells were analyzed for proliferation using CFSE dye dilution. The Δ geometric mean was used as a measure of proliferation and calculated as follows: Δ geometric mean=geometric mean (non-proliferated cells) - geometric mean (total cells). The Δ geometric mean was then used to calculate the “relative proliferation”, which is the percentage of maximal proliferation (PHA) corrected for spontaneous proliferation (HIV-1 p17 Gag77–85) ((Δ geometric mean sample − Δ geometric mean control medium)/(Δ geometric mean PHA − Δ geometric mean control medium))×100%=% of maximal proliferation. The cut-off value for a positive proliferative response was arbitrarily set at 10% relative proliferation in order to limit the number of candidate epitopes to be evaluated in subsequent experiments 30.

IFN-γ ELISA

IFN-γ concentration in cellular supernatants was detected using ELISA (U-CyTech, Utrecht, The Netherlands) as previously described 59.

Acknowledgements

This work was supported by a grant from the Foundation Microbiology Leiden, the European Commission within the sixth Framework Program (FP6), the Bill and Melinda Gates Foundation, TI Pharma (project D-101-1), Grand Challenges in Global Health (GC6♯74, GC12♯82), ISA Pharmaceuticals and TBVAC contract no. LSHP-CT-2003-503367 (the text represents the authors' views and does not necessarily represent a position of the Commission who will not be liable for the use made of such information). We thank Corine Prins, Sandra Arend, Michèl R. Klein, Willem Verduijn and his colleagues for their support.

Conflict of interest: The authors declare no financial or commercial conflict of interest.

Ancillary