Mannose-binding lectin and susceptibility to tuberculosis: a meta-analysis

Abstract

It has been proposed that mannose-binding lectin (MBL) levels may impact upon host susceptibility to tuberculosis (TB) infection; however, evidence to date has been conflicting. We performed a literature review and meta-analysis of 17 human trials considering the effect of MBL2 genotype and/or MBL levels and TB infection. No significant association was demonstrated between MBL2 genotype and pulmonary TB infection. However, the majority of studies did not report MBL2 haplotype inclusive of promoter polymorphisms. Serum MBL levels were shown to be consistently elevated in the setting of TB infection. While this may indicate that high MBL levels protect against infection with TB, the increase was also of a degree consistent with the acute-phase reaction. This analysis suggests that the relatively poorly characterized MBL2 genotypes reported are not associated significantly with susceptibility to pulmonary TB infection, but high MBL serum levels may be.

Introduction

Balanced polymorphisms are the result of beneficial effects of resistance to prevalent infections due to physiological changes consequent on genetic variation. Well-characterized examples in human biology include haemoglobinopathies (sickle-cell and alpha-thalassaemia) and Plasmodium falciparum[1]. One of the most common polymorphisms on a global scale is that involving mannose-binding lectin (MBL), a pattern recognition receptor of the innate immune system. This liver-derived, acute-phase reactant recognizes pathogen-associated molecular patterns, killing microorganisms via activation of the lectin complement pathway and opsonophagocytosis [2,3]. MBL is also involved in modulation of other inflammatory pathways contributing to autoimmune disease, apoptosis and vascular disease [4]. Despite its manifold effects in innate immune system pathways, there is a high frequency of MBL deficiency that arises due to polymorphisms of the MBL2 gene. The evolutionary advantage of MBL deficiency is unclear.

MBL production is controlled by the MBL2 gene, and polymorphisms of the structural regions of the gene or its promoter are associated with relative or absolute serum MBL deficiencies [5]. The presence of key structural and promoter polymorphisms in a detailed MBL2 haplotype is reasonably well correlated with reduced serum MBL levels, and genotypic analyses are used frequently as surrogates for MBL serum levels. The MBL2 structural gene variants, B, C and D, are referred to collectively as O while A is the wild-type. Prior to recognition of the importance of MBL2 promoter polymorphism, MBL deficiency was defined on the basis of structural gene polymorphism alone and variably as the presence of any variant allele, [AO or OO] or compound heterozygosity for variant alleles [OO]. While there is considerable variation in MBL levels within all genotype groupings other than O/O, the wild-type is generally associated with high MBL levels, A/O with intermediate and O/O with low or absent MBL.

MBL has been shown to be involved in the control of many microorganisms, including bacteria, fungi, parasites and viruses [6–9], and MBL deficiency has been associated with an increased frequency of various infections, including sepsis, aspergillosis, meningococcal disease and invasive pneumococcal infections [8,10–13]. Intracellular pathogens, including Mycobacterium tuberculosis, co-opt macrophage phagocytosis to assist with establishing and disseminating infection [14]. Therefore, it has been proposed that high MBL serum levels may lead to increased tuberculosis infections (TB) through promotion of M. tuberculosis opsonization [15]. This has been strengthened by studies demonstrating that MBL enhances phagocytic activity against other mycobacteria and demonstration of a protective effect of MBL deficiency against at least some forms of M. leprae infection [15–18]. A number of clinical and genetic studies have been performed to consider the impact of MBL levels or MBL polymorphisms on the development of TB. Results from these studies have been conflicting or contradictory, and it has been unclear whether MBL deficiency states result in increased susceptibility to tuberculosis infection. To attempt to clarify this situation, therefore, we carried out a meta-analysis of studies considering the association between MBL deficiency and tuberculosis infection.

Materials and methods

Identification of studies

For the meta-analysis, we included all published studies that considered the association between tuberculosis and MBL2 polymorphisms. A literature search for the MeSH terms ‘tuberculosis OR TB OR mycobacteria’ and ‘MBL OR mannose-binding lectin OR mannose-binding protein’ was performed using Medline and PubMed and abstracts were reviewed for relevance. No language restrictions were applied to the search strategy. References of articles were also reviewed for additional relevant citations not included in the original search protocol. Two of the authors (J.T.D. and D.P.E.) independently reviewed the full text of all articles to ensure that they met preset criteria for inclusion.

Nomenclature and data extraction

The primary outcome considered in the meta-analysis was the association between pulmonary tuberculosis infection and the presence of MBL2 polymorphisms in patients without human immunodeficiency virus (HIV). For the primary analysis, and to allow appropriate comparison of all studies, cases and controls were classified as AA (wild-type MBL2 genotype), AO (structural gene polymorphism heterozygous MBL2 genotype) or OO (compound heterozygote MBL2 genotype). Subsequent analyses were also performed for the association between pulmonary tuberculosis and MBL2 polymorphisms in HIV-positive patients, and of the association between tuberculosis and serum MBL levels. While studies may have also analysed the potential influence of MBL on the site or severity of tuberculous disease, to meet inclusion criteria all required a direct comparison between subjects with active pulmonary TB and healthy, non-TB diseased controls.

Statistical analysis

Primary and secondary analyses were performed using stata version 10·1 (Stata Corporation, College Station, TX, USA). Odds ratios (OR) and 95% confidence intervals (95% CI) were calculated for the association between tuberculosis infection and MBL2 polymorphism in each study.

To consider evidence of publication bias, we prepared funnel plots of the studies included in the final analysis. Chi-squared tests were performed to assess the degree of heterogeneity between trials, and both fixed and random-effects metaregression models were used.

Results

Studies included in the primary meta-analysis

Seventeen publications relating to MBL and tuberculosis infection in human subjects were identified [19–35]. Two were excluded as they provided data only on MBL serum levels and not MBL2 polymorphisms [19,20]. One study was excluded, as it considered only population prevalence of tuberculosis and polymorphisms without individual data [21]. One study was excluded as it did not provide sufficient individual raw data for analysis [22]. One study was excluded as data from patients with pulmonary and extrapulmonary disease could not be separated for MBL2 polymorphism analysis [27]. Data from the remaining 12 studies were included in the primary analysis of MBL2 genotype frequency in HIV-negative patients with pulmonary TB versus healthy controls, containing a total of 1815 patients and 2666 controls [23–26,28–35]. Summary data from the included studies are shown in Table 1.

Table 1.

Summary of papers describing mannose-binding lectin (MBL) and tuberculosis susceptibility.

Setting	Patient population	Controls	Case subjects	Controls	Polymorphisms included	Serum MBL levels
India [33]	HIV+/− adults with TB	Healthy controls	148 HIV-; 109 HIV+	146	Codon 52, 54, 57; X/Y promoter region	Yes
The Gambia [29]	HIV- adults with SS+ PTB	Blood donors	397	422	Codon 54, 57	No
Poland [20]	Active TB	Non-TB lung disease and healthy controls	68	94	n.a.	Yes
Italy [25]	Adults with PTB	Household contacts	277	288	Codon 52, 54, 57; X/Y promoter region	No
Turkey [34]	HIV- children with culture+ TB	Healthy age-matched children	44	99	Codon 54, 57	Yes
Poland [23]	Adults with PTB	Healthy controls	108	92	Codon 52, 54, 57	Yes
USA [22]	Adults HIV- active TB	Healthy controls exposed to TB but no active disease	487	232	Codon 52, 54, 57	No
Malawi [24]	Adults with active TB	Unrelated heathy controls	168	546	Codon 57	No
Spain [31]	HIV+/− adults with TB	HIV+ without TB and healthy controls	106 HIV-; 414 HIV+	344	Codon 52, 54, 57; X/Y promoter genotype	No
Tanzania [19]	Adults HIV+/−, TB+/−	HIV- school students and trauma patients	94 HIV-; 150 HIV+	113	n.a.	Yes
South Africa [27]	Active and past TB infection	Healthy controls from same community	155	79	Codon 52, 54, 57	Yes
Chinese Han [26]	Adult males with pulmonary TB	Healthy army servicemen	152	293	Codon 52, 54, 57	No
Sub-Saharan Africa [21]	Various SSA countries with differing TB prevalence	n.a.	626	n.a.	Codon 54, 57	No
Turkey [32]	Adults with PTB	Healthy adults	49	100	Codon 54, 57	No
Denmark [28]	Adult HIV- TB+	Blood donors and laboratory staff	109	250	Codon 52, 54, 57; X/Y promoter region	Yes
India [35]	Adults with past PTB or Hx of PTB	Exposure but no past history of TB	202	109	Codon 52, 54, 57	No
India [30]	HIV- adults with PTB	Healthy adults	48	58	Codon 52, 54, 57	Yes

HIV: human immunodeficiency virus; n.a.: not applicable; PTB: pulmonary tuberculosis; SSA: sub-Saharan Africa; TB: tuberculosis.

To examine the effect of the degree of MBL deficiency, pooled data were considered according to genotype in two different manners. Twelve studies [23–26,28–35] contained sufficient data for primary analysis of wild-type versus any MBL2 variant allele (OA/OO) genotype, representing a wide range of intermediate and extremely low MBL levels. Ten studies [23–26,28–31,33,35] contained sufficient information for wild-type versus compound heterozygote (OO) genotype frequency in cases and controls, representing a comparison between normal and extremely low MBL levels alone.

Primary analysis

Chi-squared testing of the included studies demonstrated a high degree of heterogeneity (P < 0·001). Due to the high degree of heterogeneity, a random-effects meta-regression model was considered to be most appropriate and was applied throughout.

Figure 1 shows the odds ratios (OR) for tuberculosis infection between subjects with wild-type MBL2 genotypes (AA) and those with either single (AO) or compound heterozygous (OO) MBL2 mutations. ORs from individual studies ranged from 0·18 to 3·94, with a combined OR of 0·87 (95% CI 0·59–1·28). Figure 2 shows the OR for tuberculosis infection comparing subjects with AA genotypes and those with OO MBL2 variants. OR from individual studies ranged from 0·14 to 2·30, with a combined OR of 0·55 (95% CI 0·22–1·34). Neither analysis, then, demonstrates a significant difference between groups, suggesting that MBL2 genotype relying on structural gene variant allele alone does not significantly influence tuberculosis susceptibility.

Fig. 1.

Meta-analysis of influence of mannose-binding lectin 2 genotype on tuberculosis susceptibility: AA versus OA/OO forest plot.

Fig. 2.

Meta-analysis of influence of mannose-binding lectin 2 genotype on tuberculosis susceptibility: AA versus OO forest plot.

An additional ad hoc meta-analysis was performed on studies that reported a complete MBL2 genotypic profile inclusive of promoter polymorphisms. Although only a minority of studies reported such data, this group was chosen as such genotype profiles are associated considerably more strongly with MBL serum levels than structural genotypes alone. Using this subset, patients and controls were reanalysed based on the frequency of high or low MBL-producing genotype. O/O and XA/O were considered low MBL-producing genotypes in this analysis, with other genotypes considered to be high MBL-producing. This analysis, shown in Fig. 3, did not demonstrate a significant effect of MBL2 genotype on likelihood of pulmonary TB infection [25,28,31,33], with results influenced significantly by a single outlying study.

Fig. 3.

Meta-analysis of influence of mannose-binding lectin 2 genotype with promoter regions on tuberculosis susceptibility.

Secondary analyses

Genotypes in HIV-positive patients

Two studies [31,33] contained sufficient data to allow comparison of MBL2 wild-type versus MBL2 variant compound heterozygote genotype frequency in HIV-positive patients with and without tuberculosis infection versus healthy control. These studies included a total of 173 cases and 393 controls, and summary data are presented in Table 2.

Table 2.

Summary of studies of mannose-binding lectin (MBL2) genotype on tuberculosis susceptibility in human immunodeficiency virus-positive patients versus controls.

Reference	TB cases AA	TB cases OO	Healthy controls AA	Healthy controls OO	OR (95% CI)
[33]	31 (46·3)	5 (7·5)	29 (59·2)	3 (6·1)	0·64 (0·14–2·93)
[31]	39 (36·8)	3 (2·8)	183 (53·2)	27 (7·8)	1·9 (0·56–6·64)
Total	70 (40·5)	8 (4·6)	212 (53·9)	30 (7·6)	1·2 (0·54–2·82)

TB: tuberculosis.

The two studies analysed conflict directly, with one suggesting a protective effect of wild-type MBL2 genotypes and the other suggesting an increased susceptibility to TB infection. Neither study achieved statistical significance independently. When considered together, these results do not show a significant association between deficiency-associated MBL2 genotypes and TB susceptibility (OR 1·2, 95% CI 0·54–2·82).

Serum MBL levels in HIV-negative patients

Eight studies reported collection of serum MBL levels from at least some subjects [19,20,23,27,28,33–35]. One study was excluded because it reported MBL levels in subjects with TB but not controls [28]. One study presented MBL levels only according to subject genotype, and the data did not permit overall comparison of subjects and controls [23]. One study was available only in abstract form in English and did not contain sufficient detail for inclusion [20]. One study contained data only on HIV-positive subjects [33]. In total, four studies contained sufficient data to allow comparison of serum MBL levels in HIV-negative patients with and without tuberculosis [19,27,33–35]. The included studies contained a total of 341 patients with active tuberculosis and 349 controls.

Three of the studies reported that serum was collected for MBL sampling prior to or shortly after the introduction of anti-TB therapy [19,27,35], while in the remaining study timing of sample collection was not reported [34]. One study also reported sampling an additional group of patients after completion of therapy [27]. In one study, MBL levels were not available in the published text, but were kindly provided for inclusion ([19]; P. Garred, personal communication).

Each of the four studies found higher median levels of serum MBL in patients with active TB disease compared with controls. The results from the studies were each independently significant, with P values ranging from <0·01 to 0·04. Because each study was reported as independently significant, we did not perform a formal meta-analysis on these data. Serum MBL-level data from active TB versus controls are shown graphically in Fig. 4. The consistent finding of higher MBL serum levels in patients with TB than uninfected controls suggests strongly that high MBL levels are associated with active TB disease.

Fig. 4.

Median serum mannose-binding lectin levels in human immunodeficiency virus-negative patients with tuberculosis compared with healthy controls.

Discussion

Our meta-analysis of accessible, published data has demonstrated no statistically significant association between MBL2 genotype and pulmonary TB infection. Review of studies considering MBL levels has, however, demonstrated a consistent increase in MBL levels in patients with tuberculosis. There are a number of mechanisms that could account for this discrepancy between MBL levels and MBL2 genotype and their associations with TB. First, MBL2 genotype involving structural gene polymorphisms alone is a poor predictor of serum MBL levels. Therefore the direct assessment of MBL phenotype through measurement of blood levels may be the best way to reveal an association between MBL sufficiency and predisposition to TB from the available data. Perhaps the most plausible explanation, however, is that MBL is elevated in active tuberculosis infection as part of an acute-phase reaction. If this is so, then MBL would not appear to be involved significantly in host susceptibility to tuberculosis infection.

In an attempt to investigate these alternative explanations, we performed additional ad hoc subgroup analysis on studies that reported complete MBL2 genotypic profile, including promoter regions. The small sample of this subgroup analysis does not permit significant conclusions to be drawn from the lack of association between complete MBL2 genotype and TB susceptibility; however, were such results repeated in larger studies it may provide additional support for the hypothesis that MBL is not involved in tuberculosis infection and elevated MBL levels seen in patients with TB represent its acute-phase response.

Although some studies have suggested that MBL may not have a significant overall acute-phase response, patients with wild-type MBL2 genotypes have been generally found to have raised MBL levels in this setting [30]. This is consistent with the study populations included in this meta-analysis, where the proportion of patients homozygous for wild-type MBL2 accounted for 92% of the populations where both MBL levels and genotypes were available. In these populations, therefore, the acute-phase properties of MBL are likely to be dominant. This contrasts with other studies of the acute-phase change seen in septic patients who had a higher frequency of MBL2 variant alleles [36]. Support for this conclusion can be seen in studies where MBL levels have been studied in the acute-phase reaction. Thiel et al. demonstrated elevations of serum MBL 1·5–3 times baseline following surgical procedures; although insufficient data were available for a formal comparison, this magnitude of increase appears consistent with the difference in each of the studies reviewed here [37]. In the single study which compared patients with active tuberculosis and those with a past history of infection, serum MBL levels were found to be higher in the acute phase, although this difference was small and not statistically significant (P = 0·38; [27]). No study, to our knowledge, has compared serial MBL levels in patients during and after active tuberculosis infection; this would be of interest in future research.

Overall, this meta-analysis is limited by the large degree of heterogeneity in the designs of the studies analysed, and conclusions drawn may be less applicable to specific subpopulations. It has also been suggested that the high degree of genetic heterogeneity in the populations studied may account for the conflict between results [25]. However, our meta-analysis employed a random effects model designed to counter these variations and found no overall effect of MBL2 genotype on TB susceptibility. Additional attempts at considering this hypothesis (for instance, meta-analysis according to geographic region; not shown) did not suggest a more significant impact of MBL2 genotype. Equally, when studies were ranked on the basis of methodological quality and reanalysed, no significant alteration to our primary analysis could be demonstrated (not shown).

If MBL deficiency does not confer protection against tuberculosis, it is challenging to propose another disease where MBL deficiency is known to be protective that may have promoted the observed high frequency of MBL2 polymorphisms. To lead to such widespread polymorphisms as observed in MBL, a condition must have had a substantial effect on reproductive fitness over many generations. Candidate non-infectious diseases such as vascular disease are unlikely to have had such an impact on MBL2 genetic polymorphism, as it is only in recent history (and in industrialized nations) that such diseases have accounted for high burden of mortality. Further research into the factors promoting diversity in MBL2 polymorphism will be awaited with interest.

Disclosure

All authors wish to declare that they have no conflict of interests in this study or its publication, financial or otherwise.