• We are sorry, but NCBI web applications do not support your browser and may not function properly. More information
Logo of clinexpimmunolLink to Publisher's site
Clin Exp Immunol. Dec 2006; 146(3): 390–399.
PMCID: PMC1810413

The influence of previous exposure to environmental mycobacteria on the interferon-gamma response to bacille Calmette–Guérin vaccination in southern England and northern Malawi


We report a large study of the effect of BCG vaccination on the in vitro 6-day whole blood interferon-gamma (IFN-γ) response to antigens from eight species of mycobacteria among schoolchildren in south-eastern England, where bacille Calmette–Guérin (BCG) vaccination is highly protective against pulmonary tuberculosis, and among young adults in northern Malawi, where BCG vaccination is not protective. In the UK children, BCG induced an appreciable increase in IFN-γ response to antigens from most species of mycobacteria. The degree of change was linked to the relatedness of the species to Mycobacterium bovis BCG, and provides further evidence of the cross-reactivity of mycobacterial species in priming of the immune system. IFN-γ responses to purified protein derivatives (PPDs) from M. tuberculosis and environmental mycobacteria were more prevalent in the Malawian than the UK group prior to vaccination; BCG vaccination increased the prevalence of responses to these PPDs in the UK group to a level similar to that in Malawi. There was no evidence that the vaccine-induced change in IFN-γ response was dependent upon the magnitude of the initial response of the individual to environmental mycobacteria in the United Kingdom or in Malawi. These observations should assist the development and interpretation of human clinical trials of new vaccines against M. tuberculosis in areas of both low and high exposure to environmental mycobacteria.

Keywords: BCG, English schoolchildren, environmental mycobacteria, IFN-γ, Malawi, tuberculosis vaccine


Although widely practised, bacille Calmette–Guérin (BCG) vaccination remains controversial because of the widely varying efficacy observed in different populations [1]. One explanation for these differences, supported by several studies in humans and experimental animals [28], is that they reflect population differences in previous exposure to cross-reacting environmental mycobacteria. Two mechanisms have been proposed: ‘blocking’, which implies that the previous immunity induced by exposure to environmental mycobacteria restricts the growth of the BCG; or ‘masking’, which implies that the BCG is unable to confer any additional immunity to that already induced by the natural mycobacterial exposure. Continued debate over these studies and hypotheses point to the need for immunological investigations of the role of environmental mycobacterial exposure on subsequent response to BCG vaccination in humans. Our current studies of BCG vaccination have provided such an opportunity.

We have reported that an appreciable proportion of schoolchildren in south-east England make an interferon (IFN)-γ response to purified protein derivatives (PPDs) of environmental species of mycobacteria, especially those of the MAIS (Mycobacterium avium, M. intracellulare and M. scrofulaceum) complex and M. marinum, prior to receiving their routine BCG vaccination [9], but these responses are mainly of low magnitude. We have also reported that this group showed a considerable increase in IFN-γ response and in Heaf skin test response, to M. tuberculosis PPD 1 year after receiving (Glaxo strain) BCG vaccination [10]. This contrasted with findings from a similiar study in a group of young adults in Malawi, where the same BCG vaccine induced only a small increase in the already high IFN-γ responses to M. tuberculosis PPD, and to a range of other mycobacterial proteins, including PPDs from a range of environmental mycobacteria [10,11]. The observed change in in vitro response to M. tuberculosis PPD mirrored the known efficacy of the BCG vaccine against pulmonary tuberculosis (TB) in each population. The lack of BCG-vaccine-induced increase in in vitro IFN-γ response to M. tuberculosis PPD in Malawi was ascribed to the previous high exposure of the Malawian adolescents and young adults to environmental mycobacteria. Both the prevalence and magnitude of sensitivity to the PPDs from environmental mycobacteria were higher in the Malawi group than in the UK group prior to BCG vaccination [9].

In this study we investigate the effect of BCG vaccination on responses to M. bovis, M. avium, M. intracellulare, M. scrofulaceum, M. marinum, M. fortuitum, M. kansasii and M. vaccae, in comparison with responses to M. tuberculosis PPD, in UK schoolchildren. This enables us to address cross-reactivity between different species of mycobacteria, and data are analysed in the context of relatedness to M. bovis BCG, as determined by 16S RNA typing [12]. We also investigate to what extent prior exposure to different species of mycobacteria interferes with BCG vaccine-induced IFN-γ response to antigens of M. tuberculosis and of these other mycobacteria in UK schoolchildren. Results from the United Kingdom are compared with analogous data from the parallel study of young adult BCG recipients in Malawi [11].

Subjects, materials and methods

Recruitment of subjects

In the United Kingdom, recruitment was through the schools’ BCG programme run in Redbridge and Waltham Forest Health Authority in East London and Essex, England. Children born between 1984 and 1986 were invited by letter to participate in the study at the time of their routine BCG vaccination, and recruitment proceeded following written consent from parent or guardian and verbal consent from the child. Exclusion criteria were evidence of previous BCG vaccination (BCG scar or vaccination records) or serious illness. Children were asked if they took any medication, either currently or within the last month. In Malawi, recruitment was through the Karonga Prevention Study. Candidates born between 1970 and 1988, and selected from the project database on the basis of having no BCG scar, not having been included in the Karonga vaccine trial, having no record of a positive HIV test and never having been confirmed or suspected as having either tuberculosis or leprosy, were visited in their homes by field staff who explained the study and obtained written consent, including for an HIV test. Exclusion criteria were evidence of a BCG scar, signs of tuberculosis, leprosy or other severe illness, generalized rashes, pregnancy or HIV positivity.

Ethical approval for these studies was given by the Local Research Ethics Committee of Redbridge and Waltham Forest Health Authority (R&WFHA), by the National Health Sciences Research Committee of Malawi and by the Ethics Committee of the London School of Hygiene & Tropical Medicine. All subjects gave informed consent.

Study protocol

In both locations, a blood sample was obtained and a skin test carried out at recruitment. A 5-ml intravenous blood sample was transferred immediately into a sterile tube (Greiner Labortechnik Ltd, Stonehouse, Gloucestershire, UK) containing 50 U of preservative-free sodium heparin (Monoparin; CP Pharmaceuticals Ltd, Wrexham, UK), and set up in the whole blood assay as soon as possible on the same day as venesection.

In the United Kingdom, skin testing was carried out on the volar surface of the forearm using the Heaf technique with tuberculin PPD (100 000 U/ml, Medeva, Leatherhead, UK) following standard procedures [13]. The induration was inspected after 7 days and graded by experienced nurses. Children with a Heaf grade of 2 or above were excluded from BCG vaccination, and hence from the second phase of the study. Children with a Heaf grade of 3 or above were referred for active investigation for tuberculosis, following standard procedures [13]. The remaining children were randomized in blocks of six to receive BCG vaccine (two-thirds, vaccine group) or to have vaccination delayed for 1 year (one-third, control group). In Malawi, skin testing was carried out on the volar surface of the forearm using the Mantoux technique, with 2 tuberculin units (TU) of M. tuberculosis PPD batch RT23 [Statens Serum Institut (SSI), Copenhagen, Denmark] and read after 48–72 h. Induration was measured across and along the arm and the mean value used in analyses. Individuals with induration to PPD RT23 of more than 10 mm were excluded from vaccination and referred for examination for tuberculosis. Individuals with 10 mm or less induration to PPD RT23 were randomized to receive BCG vaccine (two-thirds, vaccine group) or placebo (one-third, control group). Randomization was in blocks of six and was blind both to participants and to field staff. The vaccine (in both locations) was Glaxo strain (1077) (Evans Medical, Liverpool, UK). The placebo (used in Malawi) was the dextran matrix of the BCG vaccine.

In both the United Kingdom and Malawi, 12 months after vaccination the participants were recalled/revisited for follow-up testing. Procedures were repeated as for the recruitment phase. In the United Kingdom, individuals with less than a grade 2 Heaf result whose BCG vaccination had been delayed for a year were offered BCG vaccination. In Malawi, HIV-negative individuals with 10 mm or less skin-test induration to RT23 and who had received placebo in the recruitment phase were offered BCG vaccination.

Whole blood assay

The whole blood assays and enzyme-linked immunosorbent assays (ELISAs) were performed in the project laboratories in London and Chilumba, using identical protocols as described previously [10]. Heparinized whole blood was diluted 1 in 5 with serum-free medium (RPMI-1640 supplemented with 20 IU/ml penicillin and 20 µg/ml streptomycin plus 2 mM L-glutamine; Gibco brl, Paisley, UK) and plated in 96-well, round-bottomed tissue culture plates (Nunc, Roskilde, Denmark) at 100 µl/well. Cells were stimulated in quadruplicate with antigen, mitogen or with serum-free medium in a volume of 100 µl/well, giving a final volume of 200 µl/well and blood dilution of 1 in 10. Cell cultures were incubated at 37°C with 5% CO2. Supernatants were harvested on day 6 and stored at −70°C prior to ELISA.


Purified protein derivatives (PPDs) for in vitro use from M. tuberculosis (batch RT48 for in vitro use, lot 191), M. avium (batch RS10/2, lot 39), M. intracellulare (batch RS23, lot 28), M. scrofulaceum (batch RS 95, lot 18), M. marinum (batch RS 170, lot 11), M. kansasii (batch RS 30, lot 19) and M. fortuitum (batch RS 20, lot 17) were provided by SSI. PPD from M. bovis [prepared at the Central Veterinary Laboratory (CVL), Weybridge, Surrey, UK] was provided by NIBSC (Potters Bar, UK) and from M. vaccae (batch R877R) by Dr J. Stanford (University College, London, UK). Details of the preparation of these PPDs appear elsewhere [11]. All the mycobacterial antigens were used at a final concentration of 5 µg/ml. Controls were the mitogen phytohaemagglutinin (PHA) (Difco Laboratories/Becton Dickinson, Oxford, UK; final concentration 5 µg/ml); a non-mycobacterial antigen, streptokinase–streptodornase (SK/SD; Varidase, Wyeth Laboratories, Maidenhead, Berks, UK, final concentration 250 U/ml); and whole blood cultures incubated without stimulation in serum-free medium as the negative control (NC).

Measurement of cytokines

IFN-γ concentrations were measured in single 100 µl samples of supernatant by quantitative ELISA, as described previously [9]. To control for interplate and intraplate variation, a positive control supernatant was used. In the UK study, mean IFN-γ measurement in this positive control was 276 pg/ml for plates used to test recruitment samples (n = 166) and 296 pg/ml for plates used to test follow-up samples (n = 142). The coefficients of variation (CV) between plates were 14% and 24%, and the mean variability of duplicate measurements was 4% and 5% for recruitment and follow-up plates, respectively. In the Malawi study a different positive control sample was used: mean IFN-γ measurement of this sample was 815 pg/ml for plates used to test recruitment samples (n = 126), and 747 pg/ml for plates used to test follow-up samples (n = 97); the CVs between plates were 18% and 22% and the mean variability of duplicate measurements was 4% and 4%, for recruitment and follow-up plates, respectively. Positive control samples were exchanged between the United Kingdom and Malawi laboratories to ensure direct comparability of data.

Data analysis

Data text files were transferred from Revelation into Foxpro and analysed using stata 8·0. Negative control (NC) values were subtracted from all results. A ‘positive’ IFN-γ response was defined as > 62 pg/ml, twice the limit of detection of the assay. This threshold aims to reduce the inclusion of false positive responses, and frequency distribution data for prevaccination responses among these children indicate that this response threshold is appropriate (the frequency distributions were almost all bi-modal, with a ‘peak’ representing individuals responding below the detection limit, a very low proportion of individuals with a response 31–62 pg/ml and an approximately log-normal distribution of response among individuals with a response > 62 pg/ml) [9]. All values below the limit of detection of the IFN-γ assay (< 31 pg/ml) were set to 15 pg/ml.

To compare vaccine and control groups, χ2 tests were used for the proportion of individuals (‘responders’) with a ‘positive’ IFN-γ response (defined as > 62 pg/ml). For analysis of the change in IFN-γ response, a log-transformation (base 2) was used for both the pre- and the post-vaccination IFN-γ responses, the difference between these 2-values was calculated and linear regression was used to assess the effect of BCG vaccination on this difference, with and without stratifying on the prevaccination response to M. tuberculosis and environmental mycobacterial PPDs; t-tests were used in the context of linear regression to assess the effect of particular explanatory variables on the magnitude of the pre–post change in IFN-γ response. This was carried out for all individuals, and then restricted to those who made an IFN-γ response of ≤ 62 pg/ml prior to vaccination. To analyse the effect of previous sensitivity to MAIS antigens on subsequent response to BCG vaccine, individuals were categorized according to their prevaccination IFN-γ response to the MAIS antigens as described in the text. Regression analysis was used rather than χ2 or correlation analysis, in order to capture the size of the vaccine-induced change and to take account of factors other than vaccination as reflected in the control group.


Four hundred and twenty-four UK schoolchildren were recruited for the study between February 1999 and April 2000. Of these, 397 were eligible (Heaf grades 0 or 1), and 340 were followed-up a year later for repeat testing (follow-up). Of these, 240 received BCG vaccination (by random allocation) at recruitment (vaccine group); the remaining 100 had their BCG delayed for 1 year (control group). In Malawi, 633 young adults were recruited between January and November 1998; 562 were eligible for the study (Mantoux 10 mm or less) and 483 were followed-up a year later. Of these, 329 received BCG vaccination (vaccine group) and 154 received placebo (control group) by random allocation.

In vitro responses to the negative control (RPMI-1640) were below the level of detection in 97% of the UK study group at recruitment. This was not altered by BCG vaccination (vaccine group 97%, control group 99% at follow-up). In vitro IFN-γ responses to the positive control stimulus PHA were detected in 99% of vaccine and control individuals at recruitment, with a median response of > 2000 pg/ml. Responses to the non-mycobacterial SK/SD antigen were detected in 85% of the subjects at recruitment, with median responses in the control and vaccine groups of 728 and 531 pg/ml, respectively (P = 0·6). Vaccination with BCG did not alter the in vitro response to either of these stimuli − the post/prevaccination ratio in IFN-γ response to SK/SD was 0·9 (data not shown). Similar quality control data were recorded in the Malawi study [10].

A clear increase in IFN-γ response to M. tuberculosis PPD was observed 1 year after BCG vaccination (Fig. 1a), as reported previously for a subset of these UK individuals [10]. The average ratio of post/prevaccination response (‘size of change’) was 9·2 (95% CI = 6·8–12·5). As shown in Table 1, increases in IFN-γ response were also observed for PPDs from other mycobacteria, the size of change being correlated with the relatedness of the species to M. bovis BCG as defined by 16S RNA analysis [12] (with the exception that a greater change was seen to M. tuberculosis PPD than to M. bovis PPD). Detailed results are presented in Fig. 1b–d for M. bovis, M. avium and M. fortuitum. We have reported previously that these children showed varying sensitivity to the panel of PPDs prior to vaccination, which was interpreted as an indicator of previous exposure to these different (or to closely related) species − the most prevalent sensitivity being observed to M. avium PPD [9]. To exclude the possibility that the size of change was related to degree of previous sensitivity to a particular species the analysis was repeated, restricted to those individuals whose recruitment IFN-γ response was below 62 pg/ml to each species. The same order of size of change in response to each species, again in parallel with the relatedness of each species to M. bovis BCG, was observed following BCG vaccination (data not shown).

Fig. 1
Change in interferon (IFN)-γ response between recruitment and follow-up of control and vaccinated subjects in the United Kingdom to mycobacterial purified protein derivatives (PPDs). Subjects were bled prior to vaccination and 1 year later and ...
Table 1
Change in interferon (IFN)-γ response between recruitment and follow-up of subjects in the United Kingdom to purified protein derivatives (PPDs) from different mycobacterial species in order of decreasing relatedness to Mycobacterium bovis bacille ...

The response to each of the antigens at recruitment is presented as percentage of responders (> 62 pg/ml) in UK and Malawi groups in Fig. 2a. Data are combined for control and vaccine groups, as they showed very similar results. Before vaccination, responses to all species except M. vaccae were more prevalent in Malawi than in the United Kingdom. In Malawi, for several of the antigens, the proportion of subjects making a positive IFN-γ response increased over the year of the study in the control group as well as in the vaccinated group; this was not observed in the United Kingdom (Fig. 2b). Following vaccination, the prevalence of responses was, in general, similar between the two locations (Fig. 2c), with no evidence of a difference in response to M. tuberculosis PPD (82% versus 78%, P = 0·21). The prevalence of response was higher post-vaccination in the United Kingdom than in Malawi for two PPDs: M. marinum (89% versus 82%, P = 0·03) and M. bovis (72% versus 50%, P < 0·001). Figure 3a,b shows the median size of IFN-γ response among those who responded with at least 62 pg/ml in the United Kingdom and Malawi groups at recruitment and follow-up, respectively. For each PPD these responses were distributed approximately log-normally (data not shown). Prior to vaccination the median responses were similar in each location, with the exception of M. tuberculosis PPD, for which responses were higher in Malawi (P < 0·001). One year later, in the UK group there were significant vaccine-induced changes in median response among responders to M. tuberculosis, M. marinum, M. avium, M. intracellulare, M. scrofulaceum and M. bovis (P < 0·001, P < 0·001, P< 0·001, P < 0·001, P = 0·03 and P = 0·05, respectively), but not to M. kansasii, M. fortuitum and M. vaccae (the species most distantly related to BCG). There was no significant vaccine-induced increase in median IFN-γ response among responders to any of the species in Malawi. Following BCG vaccination the median response among responders was significantly higher in the UK group than in the Malawi group to M. bovis (P = 0·01), M. tuberculosis (P = 0·002), M. marinum (P = 0·01) and M. avium (P = 0·04).

Fig. 2
Effect of BCG vaccination on prevalence of interferon (IFN)-γ responses to mycobacterial purified protein derivatives (PPDs) in the United Kingdom and Malawi. Subjects were bled prior to vaccination and 1 year later and IFN-γ responses ...
Fig. 3
Effect of bacille Calmette–Guérin (BCG) vaccination on magnitude of interferon (IFN)-γ response to purified protein derivatives (PPDs) from Mycobacterium tuberculosis and environmental mycobacteria in the United Kingdom and Malawi. ...

In order to examine the effect of previous IFN-γ response to environmental mycobacterial PPDs on the observed response to M. tuberculosis PPD following BCG vaccination, we considered that any effect of a pre-existing IFN-γ response to MAIS PPDs (to which were observed the most frequent responses in each location) should be observed most clearly by comparing subjects with the highest and lowest IFN-γ responses at recruitment. The extremes were grouped as follows: ‘non-responders’ (defined as ≤ 62 pg/ml IFN-γ response to M. avium, M. intracellulare and M. scrofulaceum PPDs) and ‘high responders’ (defined as > 250 pg/ml to all three). Analysis was restricted to those individuals making a response of < 250 pg/ml IFN-γ to M. tuberculosis PPD at recruitment, in order to include only individuals in whom a clear increase in response could be observed (responses over 250 pg/ml are already approaching the maximum response that can be measured accurately). In the United Kingdom, no relationship was observed between the size of initial response to the MAIS PPDs (comparing non-responders versus high responders as defined above) and the induced change in IFN-γ response to M. tuberculosis PPD (P = 0·36). In Malawi, the vaccine-induced change was smaller among individuals with a high (> 250 pg/ml) response to the MAIS PPDs, but this was not statistically significant (P = 0·32). Figure 4 shows size of change by category of initial response for the UK and Malawi participants. This was still the case in both locations, when the stringency of difference between no/low and high initial response was reduced to permit larger group sizes for comparison (non-responder now defined as < 62 pg/ml IFN-γ to any two species of M. avium, M. intracellulare and M. scrofulaceum with response to the third being < 125 pg/ml, and a responder defined now as > 250 pg/ml to any of the three and > 62 pg/ml to the other two). The UK and Malawi results were also analysed by considering the prevaccination response to M. avium alone. Data were categorized as above, by no/low versus high initial response (≤ 62 pg/ml versus > 250 pg/ml IFN-γ response to M. avium PPD prior to vaccination); and additionally to include data from all vaccinees (n = 240, grouped by ≤ 62 pg/ml, 63–500 pg/ml, > 500 pg/ml IFN-γ response to M. avium PPD prior to vaccination; data not shown). Neither analysis found any evidence that the size of the vaccine-induced change in IFN-γ response to M. tuberculosis PPD was affected by the baseline response to M. avium, in either Malawi or the United Kingdom.

Fig. 4
Effect of previous interferon (IFN)-γ response to MAIS (Mycobacterium avium, M. intracellulare and M. scrofulaceum) complex purified protein derivatives (PPDs) on subsequent change in IFN-γ response to Mycobacterium tuberculosis PPD following ...


This study addresses the cross-reactivity between immune responses to mycobacterial antigens induced by natural exposure to environmental species of mycobacteria, and whether this influences subsequent response to BCG vaccination in humans, as has been reported from animal models. We have measured immune responses by IFN-γ production to in vitro stimulation with mycobacterial PPDs in diluted whole blood cultures, from volunteers tested before and 1 year after BCG vaccination or no BCG vaccination in controls. Our study groups were schoolchildren in the United Kingdom, in whom we would expect BCG vaccination to induce 70–80% protection against pulmonary tuberculosis [14], and young adults in Malawi, in whom we would expect the same vaccine to confer no protection [15]. We have reported previously that the group in Malawi showed much greater previous sensitivity to various species of environmental mycobacteria than the UK group [9,11].

BCG vaccination of UK schoolchildren dramatically increased their IFN-γ response to M. tuberculosis PPD [10] and to most of the other species tested. This cross-reactivity between immune responses to different species of mycobacteria is consistent with previous findings in animal models: for example, M. tuberculosis-specific T cells from mice infected with M. tuberculosis conferred protection against M. avium, M. simiae and M. kansasii infection when transferred passively to T cell-deficient mice [16]. In humans, a study of tuberculosis and leprosy incidence in Malawi as a function of sensitivity to a panel of skin test antigens derived from different species (and strains) of mycobacteria concluded that individuals with evidence of dominant exposure to ‘fast grower’ species (but not ‘slow grower’ species) were at reduced risk of contracting both of these diseases [17]. This may reflect cross-reactivity between the many antigens which different mycobacterial species share in common, but also differential responses to different species, which can be detected in in vitro studies [9].

The size of change of response to a particular PPD induced by BCG vaccination paralleled the relatedness (as measured by 16S RNA analysis [12]) of that species to M. bovis BCG, with the exception that a greater change was seen to M. tuberculosis PPD than to M. bovis PPD. Least change was seen to the two ‘fast-growing’ species, M. fortuitum and M. vaccae, which are related most distantly to M. bovis and M. tuberculosis. The rank order of size of change remained when analysis was restricted to those individuals who made no measurable response prior to vaccination.

Studies in animal models have suggested that the influence of previous infection with an environmental species on subsequent BCG vaccination may be related to the ability of that species to persist in vivo following inoculation [3]. One study compared the growth and immunizing potential of several strains of M. avium and M. fortuitum (isolated from northern Malawi) when inoculated into mice [2]. The M. avium grew readily in the murine host, whereas M. fortuitum did not. After recovery from these infections, the mice were challenged with live BCG: replication of BCG was inhibited in the mice which had initially received M. avium, but not in those which had received M. fortuitum. As data on in vivo survival and growth of different species of environmental mycobacteria in the human are not available, we are unable to include this factor in our analysis of size of immune response induced by BCG vaccination subsequent to exposure to each of these species in our study groups. However, it is likely that the different species included in our study have different effects upon the human immune system in ways related to their survival and growth characteristics, as well as to their ubiquity in the environment and phylogenetic relatedness to M. tuberculosis and to BCG vaccine. It is also known that none of these crude antigen preparations is entirely specific, and it is likely that immune recognition of any particular antigen also reflects exposure to related organisms, not necessarily the species from which the antigen was derived.

When results from Malawi were compared to those from the United Kingdom, the differences between the two locations were striking. The prevalence of T cell recognition of the different species of mycobacteria prior to vaccination was higher in Malawi than in the United Kingdom, except for M. marinum and M. kansasii for which the prevalences were similar, and M. vaccae, for which a slightly greater prevalence was observed among UK children (17% compared to 14% in Malawi, among control subjects). Following vaccination, the prevalence of response was similar between locations for five of the nine mycobacterial species tested. The exceptions to this were IFN-γ responses to PPDs from M. scrofulaceum and M. fortuitum, which remained higher in Malawi, M. kansasii responses which were now more prevalent in the Malawi subjects, and M. bovis responses which were now more prevalent in the UK subjects. These differences should not be over-interpreted, but may indicate that as well as inducing IFN-γ responses to antigens common to all the different species of mycobacteria, there was also induction of more restricted responses to antigens shared between BCG and certain other species.

Among control (placebo-inoculated) subjects in Malawi, an increase in prevalence of responses to the various species of mycobacteria was observed over the course of the year. This was greatest for M. vaccae, M. kansasii, M. marinum and M. scrofulaceum. This is likely to reflect continued exposure to these (or closely related) species among these subjects, although prevalence of prior T cell recognition of these species ranged from low (14% to M. vaccae) to high (86% to M. scrofulaceum). In the United Kingdom there was no obvious increase in the proportion of responders to any species of environmental mycobacteria among control subjects over the course of the year, reflecting the much lower intensity of natural exposure to mycobacteria in the UK environment.

Full frequency distributions of prevaccination data for each study group have already been published [9,11]. Among those classified as responders, the magnitude of IFN-γ response was similar in the United Kingdom and Malawi prior to vaccination. One year after vaccination, the median response in the vaccinated group was higher in the United Kingdom than in Malawi to PPDs from species most closely related to BCG, including M. tuberculosis. In Malawi, the magnitude of response increased in both control and vaccinated groups to all species except M. bovis, M. fortuitum and M. vaccae. This may be further evidence of repeated exposures to environmental species in Malawi. Direct comparison of the magnitude of the IFN-γ response between locations may not be meaningful, as in Malawi the BCG vaccination event may be the latest in a series of repeated exposures to mycobacterial antigens whereas in the United Kingdom a mycobacterial exposure is likely to be relatively infrequent.

We compared vaccine-induced changes in response to M. tuberculosis PPD in individuals with no or low previous response to PPDs of the MAIS complex with the change in individuals with high previous response to PPDs of the MAIS complex. This analysis has been presented previously for Malawi, where a lesser change was seen in initial MAIS high than in initial MAIS low responders [11], although this did not achieve statistical significance. No difference at all was observed between the two groups in the United Kingdom. This was still the case when previous sensitivity to M. avium PPD alone was considered. This finding is inconsistent with results obtained in animal models, in which previous exposure to certain environmental mycobacteria led to inhibition of the growth and subsequent protective effect of the BCG vaccination, which led in turn to hypotheses of ‘masking’ or ‘blocking’ mechanisms, or a combination of both, to explain the failure of BCG vaccine in areas with high environmental mycobacterial exposure [1]. A possible explanation for this disparity may be that an environmental exposure to mycobacteria for a UK child may lead to a transient colonization sufficient to engender a memory T cell response (detectable later by response of whole blood cultures to PPD from the particular or closely related species) but not a sufficiently established infection to induce other immune mechanisms necessary for an inhibitory effect against subsequent infection with M. tuberculosis, or vaccination with BCG. The animal model studies referred to used a sufficient dose and route of mycobacteria to ensure that infection took place, which may not be directly comparable with the human experience of exposure to environmental mycobacteria in locations such as the United Kingdom, where this is a less frequent and less intense event. It may be more comparable with the situation in Malawi; however, previous exposure to mycobacteria is so prevalent in Malawi that even in the large group studied there were relatively small numbers of non-responders to environmental mycobacteria PPDs, and of high responders who were not already high responders to M. tuberculosis PPD, and statistical significance between these two groups was not achieved. This indicates that it is probably not possible to test directly the masking or blocking hypotheses in humans without much larger cohort studies, which may need to include protection against tuberculosis as an outcome. A further implication of our findings is that an IFN-γ response to a mycobacterial antigen per se can be a feature of either a ‘protective’ or ‘non-protective’ immune response to TB − although this does not contradict our proposal that an increase in IFN-γ response to M. tuberculosis antigen in the whole blood assay is a valuable marker of vaccine-induced protection.

The protective efficacy of BCG in UK teenagers is known to be high (70–80%), and to be maintained for at least 10 years after vaccination [14,18]. Why the vaccine ‘fails’ in the other 20–30% is likely to be of great significance in understanding the key elements of the immune response required for protection. Genetics may play a role and we are currently looking for genetic differences between the children in our UK study who did and did not show an increase in blood IFN-γ response and/or skin test response to M. tuberculosis PPD following BCG vaccination. However, what is likely to be of greater importance to the majority of vaccinees is that full protection against TB may rely on an immune mechanism of which IFN-γ production by mainly CD4+ T cells (as is induced by stimulation with the soluble antigenic components of PPD), is not itself sufficient, but requires the presence of one or more additional components, or perhaps the absence of down-regulatory cytokines such as interkeukin (IL)-4, IL-10 or transforming growth factor (TGF)-β. Until a fuller understanding of what constitutes protective immunity is achieved, the contribution of masking or blocking to the poor protection induced by BCG vaccination in Malawi and other tropical countries is likely to remain uncertain. We are currently investigating alternative immune pathways in our continuing studies of immune response to BCG in the United Kingdom and Malawi. The findings of these ongoing immuno-epidemiological studies should inform those developing new vaccines against tuberculosis, which ideally will induce protection against TB well above the current ‘ceiling’ of 70–80% towards protection for all individuals in all locations. Further understanding of how particular environmental species survive and engender immune responses in the human host, together with the knowledge of prevalent species in a particular locality, may permit a location-specific approach to this conundrum. At the same time, a greater understanding of which immune mechanisms confer protective immunity against TB, in addition to those mechanisms of which an increase in IFN-γ response to certain mycobacterial antigens is a key part, should facilitate the identification of which of the new candidate TB vaccines is most likely to succeed.


The UK study was supported by LEPRA, with additional funds from WHO; the Malawi study was supported by the Wellcome Trust, with additional funds from LEPRA. We thank Ann Berry, Kathryn Brady, Sally Edwards, Anna Hadassi, Makki Hameed, Mary Hayde, Mary Heath, Freda Lock, Ann Marsden, Marie Murphy, Shakuntala Patel, Christine Sloczynska, Agnes Udom and Margaret Walsh in Redbridge and Waltham Forest Health Authority for help with the UK school study; the staff and students of Heathcote School, Highams Park School, Wanstead High School, Hainault Forest High School, Woodbridge High School, Trinity Catholic High School and Woodford County High School for their co-operation and participation in this project; Elizabeth King for laboratory assistance at LSH & TM; and Anna Randall for data entry. We thank Michael Brennan, Kaare Haslov, Glyn Hewinson, John Stanford and the Central Veterinary Laboratories, Weybridge, for providing the mycobacterial antigen preparations used in this study; and Evans Medical for donating the bacille Calmette–Guérin vaccine and placebo preparations used in Malawi. We thank the people of Karonga for participating in these studies and National Health Sciences Research Committee of Malawi for permission to publish this paper. This publication is dedicated to the memory of Barbara Holland.


1. Andersen P, Doherty TM. The success and failure of BCG − implications for a novel tuberculosis vaccine. Nat Rev Microbiol. 2005;3:656–62. [PubMed]
2. Brandt L, Cunha JF, Olsen A, et al. Failure of the Mycobacterium bovis BCG vaccine. some species of environmental mycobacteria block multiplication of BCG and induction of protective immunity to tuberculosis. Infect Immun. 2002;70:672–8. [PMC free article] [PubMed]
3. Buddle BM, Wards BJ, Aldwell FE, Collins DM, de Lisle GW. Influence of sensitization to environmental mycobacteria on subsequent vaccination against bovine tuberculosis. Vaccine. 2002;20:1126–33. [PubMed]
4. Collins FM. Immunogenicity of various mycobacteria and corresponding levels of cross-protection developed between species. Infect Immun. 1971;4:688–96. [PMC free article] [PubMed]
5. Edwards ML, Goodrich JM, Muller D, Pollack A, Ziegler JE, Smith DW. Infection with Mycobacterium avium-intracellulare and the protective effects of Bacille Calmette–Guerin. J Infect Dis. 1982;145:733–41. [PubMed]
6. Howard CJ, Kwong LS, Villarreal-Ramos B, Sopp P, Hope JC. Exposure to Mycobacterium avium primes the immune system of calvesfor vaccination with Mycobacterium bovis BCG. Clin Exp Immunol. 2002;130:190–5. [PMC free article] [PubMed]
7. Orme IM, Collins FM. Efficacy of Mycobacterium bovis BCG vaccination in mice undergoing prior pulmonary infection with atypical mycobacteria. Infect Immun. 1984;44:28–32. [PMC free article] [PubMed]
8. Palmer CE, Long MW. Effects of infection with atypical mycobacteria on BCG vaccination and tuberculosis. Am Rev Respir Dis. 1966;94:553–68. [PubMed]
9. Weir RE, Fine PEM, Nazareth B, et al. Interferon-γ and skin test responses of schoolchildren in southeast England to purified protein derivatives from Mycobacterium tuberculosis and other species of mycobacteria. Clin Exp Immunol. 2003;134:285–94. [PMC free article] [PubMed]
10. Black GF, Weir RE, Floyd S, et al. BCG-induced increase in interferon-gamma response to mycobacterial antigens and efficacy of BCG vaccination in Malawi and the UK: two randomised controlled studies. Lancet. 2002;359:1393–401. [PubMed]
11. Black GF, Dockrell HM, Crampin AC, et al. Patterns and implications of naturally acquired immune responses to environmental and tuberculous mycobacterial antigens in Northern Malawi. J Infect Dis. 2001;184:322–9. [PubMed]
12. Pitulle C, Dorsch M, Kazda J, Wolters J, Stackebrandt E. Phylogeny of rapidly growing members of the genus Mycobacterium. Int J Syst Bacteriol. 1992;42:337–43. [PubMed]
13. Department of Health. Tuberculosis: BCG immunization. In: Salisbury DM, Begg NT, editors. Immunisation against infectious disease. 2. London: HMSO; 1996. pp. 231–2.
14. Sutherland I, Springett VH. Effectiveness of BCG vaccination in England and Wales in 1983. Tubercle. 1987;68:81–92. [PubMed]
15. Karonga Prevention Trial Group. Randomized controlled trial of single BCG, repeated BCG, or combined BCG and killed Mycobacterium leprae vaccine for prevention of leprosy and tuberculosis in Malawi. Lancet. 1996;348:17–24. [PubMed]
16. Orme IM, Collins FM. Crossprotection against nontuberculous mycobacterial infections by Mycobacterium tuberculosis memory immune T lymphocytes. J Exp Med. 1986;163:203–8. [PMC free article] [PubMed]
17. Fine PEM, Floyd S, Stanford JL, et al. Environmental mycobacteria in northern Malawi: implications for the epidemiology of tuberculosis and leprosy. Epidemiol Infect. 2001;126:379–87. [PMC free article] [PubMed]
18. Hart PD, Sutherland I. BCG and vole bacillus vaccines in the prevention of tuberculosis in adolescence and early adult life. BMJ. 1977;22:293–5. [PMC free article] [PubMed]

Articles from Clinical and Experimental Immunology are provided here courtesy of British Society for Immunology
PubReader format: click here to try


Related citations in PubMed

See reviews...See all...

Cited by other articles in PMC

See all...


Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...