Immunomodulatory actions of vitamin D in mycobacterial infection

With the exception of a single report⁽ Reference Raab ^{5 )}, vitamin D and its metabolites have not been shown to possess antimycobacterial activity in the absence of cells. However, the active metabolite of vitamin D, 1,25-dihydoxyvitamin D (1,25(OH)₂D), has long been recognised to induce antimycobacterial activity in vitro in mononuclear phagocytes, the cells that control growth of MTB⁽ Reference Rook, Steele and Fraher ^{6 )}. Ligation of macrophage Toll-like receptor 2/1 heterodimers by mycobacterial antigens induces expression of the vitamin D receptor (VDR) and the 1-α hydroxylase enzyme CYP27B1⁽ Reference Liu, Stenger and Li ^{7 ,} Reference Krutzik, Hewison and Liu ^{8 )}, which synthesises 1,25(OH)₂D from the principal circulating vitamin D metabolite 25-hydroxyvitamin D (25(OH)D). Because extra-renal 1-α hydroxylase follows first-order kinetics, the rate at which it synthesises 1,25(OH)₂D depends on the availability of 25(OH)D substrate⁽ Reference Vieth, McCarten and Norwich ^{9 )}. Orally ingested vitamin D is freely converted to 25(OH)D⁽ Reference Vieth, Glorieux, Pike and Feldman ^{10 )}, and this provides the rationale for administering ‘parent’ vitamin D to induce antimycobacterial responses at the site of disease.

1,25(OH)₂D modulates immune responses by ligating membrane VDR to induce rapid effects (within minutes) or nuclear VDR to induce genomic effects (within hours)⁽ Reference Norman, Mizwicki and Norman ^{11 )}. Experiments using selective agonists and antagonists of these two receptors indicate that ligation of nuclear VDR is both necessary and sufficient for induction of antimycobacterial responses by 1,25(OH)₂D in vitro ⁽ Reference Martineau, Wilkinson and Newton ^{12 )}. 1,25(OH)₂D modulates the host response to mycobacterial infection by pleiotropic mechanisms including the induction of reactive nitrogen and oxygen intermediates⁽ Reference Rockett, Brookes and Udalova ^{13 ,} Reference Sly, Lopez and Nauseef ^{14 )}, down-regulation of the gene encoding tryptophan-aspartate containing coat protein⁽ Reference Anand and Kaul ^{15 )}, promotion of phagolysosome fusion⁽ Reference Hmama, Sendide and Talal ^{16 )}, suppression of matrix metalloproteinase enzymes implicated in the pathogenesis of pulmonary cavitation⁽ Reference Coussens, Timms and Boucher ^{17 )} and induction of antimicrobial peptides including cathelicidin LL-37⁽ Reference Liu, Stenger and Li ^{7 ,} Reference Martineau, Wilkinson and Newton ^{12 )} and human β-defensin 2⁽ Reference Liu, Schenk and Walker ^{18 )}. Cathelicidin LL-37 possesses antimycobacterial activity⁽ Reference Liu, Stenger and Li ^{7 ,} Reference Martineau, Newton and Wilkinson ^{19 )} and also induces autophagy⁽ Reference Yuk, Shin and Lee ^{20 ,} Reference Shin, Yuk and Lee ^{21 )}; 1,25(OH)₂D₃-induced antimycobacterial activity has been reported to be dependent on expression of the gene encoding cathelicidin LL-37⁽ Reference Liu, Stenger and Tang ^{22 )}.

Vitamin D and tuberculosis: historical aspects

The clinical features of vitamin D deficiency were first described in 1651, when Glisson, Bate and Regemorter published ‘A treatise of the rickets: being a disease common to children’⁽ Reference Glisson, Bate and Regemorter ^{23 )}. In addition to noting the classical musculoskeletal features of rickets, the authors made the following observation from an autopsy of an infant with the condition: ‘One amongst us doth attest, that he saw glandulous knobs and bunches so numerous that they seemed to equalise, if not exceed, the magnitude of the lungs themselves; they were situated between the lungs and the mediastinum … and were extended from the Canel bone to the Diaphragma’. TB is a well-recognised cause of mediastinal lymphadenopathy in children⁽ Reference De Ugarte, Shapiro and Williams ^{24 )} and it is interesting to speculate whether this represents the earliest case report of TB associated with vitamin D deficiency.

Some 200 years later, Chapman reported results of administering cod liver oil to patients with ‘consumption’, and made the following observation: ‘the beneficial effects of the oil were manifested most speedily and most decisively, in the improvement of the appetite, aspect of the countenance, strength and spirits’. He concluded that cod liver oil was ‘probably the only remedial agent by which the vital powers may be enabled to struggle successfully against that malady’⁽ Reference Chapman ^{25 )}. This report represents the first circumstantial evidence that administration of a preparation containing vitamin D improved clinical outcome in patients with TB, although it should be noted that no control group was studied, and that any beneficial effects of cod liver oil may have been attributable to its content of vitamin A rather than vitamin D⁽ Reference Karyadi, West and Schultink ^{26 )}. The first TB sanatorium was opened in Gorbersdorf, Germany (today Sokolowsko, Poland) in 1859, and heliotherapy (exposure of TB patients to sunlight, which induces cutaneous vitamin D synthesis) subsequently became common practice, and was credited with improvements in clinical outcome in many cases⁽ Reference Mayer ^{27 )}. In 1903, Niels Finsen was awarded the Nobel Prize in Physiology or Medicine for his discovery that shortwave UV light was effective in the treatment of cutaneous TB⁽ Reference Roelandts ^{28 )}. Vitamin D₂ was purified and crystallised in 1931⁽ Reference Askew, Bruce and Callow ^{29 )} and Charpy subsequently pioneered the use of pharmacologic doses (≥1·25 mg daily) of vitamin D₂ to treat cutaneous TB⁽ Reference Charpy ^{30 )}. Vitamin D₂ was also used to treat pulmonary TB, both as a single agent and, following the introduction of effective anti-TB chemotherapy, as an adjunct to antibiotic treatment⁽ Reference Martineau, Honecker and Wilkinson ^{31 )}.

Association between vitamin D deficiency and susceptibility to tuberculosis

In 1985, Davies observed that people migrating to the United Kingdom from countries with a high incidence of latent MTB infection experienced rates of active TB that exceeded rates in their countries of origin, and that this increased risk coincided with the development of vitamin D deficiency, probably arising as a result of decreased sun exposure⁽ Reference Davies ^{32 )}. He suggested that vitamin D deficiency may predispose to reactivation of latent MTB infection in this setting, a hypothesis supported by his observation that vitamin D deficiency associated with susceptibility to active TB⁽ Reference Davies, Brown and Woodhead ^{33 )}. Since then, eleven case–control studies investigating the association between vitamin D status and susceptibility to active TB have been published. Of these, seven have reported a statistically significant association between vitamin D deficiency and susceptibility to active TB⁽ Reference Davies, Church and Brown ^{34 –} Reference Ho-Pham, Nguyen and Nguyen ^{40 )}, three have reported a non-statistically significant trend towards such an association⁽ Reference Grange, Davies and Brown ^{41 –} Reference Martineau, Leandro and Anderson ^{43 )} and one⁽ Reference Nielsen, Skifte and Andersson ^{44 )} has reported that active TB was associated with both ‘high’ and ‘low’ serum 25(OH)D concentrations (>140 and <75 nmol/l, respectively). Potential explanations for an association between vitamin D deficiency and active TB include both causality (i.e. vitamin D deficiency impairs host immune response to MTB and causes susceptibility) and reverse causality (i.e. active TB causes vitamin D deficiency, due to anorexia, decreased exposure to sunlight in debilitated patients, or MTB-induced dysregulation of vitamin D metabolism⁽ Reference Sita-Lumsden, Lapthorn and Swaminathan ^{38 )}).

Association between susceptibility to tuberculosis and polymorphisms in the vitamin D receptor and vitamin D binding protein

Human VDR is encoded by the VDR gene located on chromosome 12q. This gene is polymorphic, and numerous SNP have been described. The hypothesis that VDR variants might associate with susceptibility to active TB was first investigated by Bellamy et al., who reported an association between carriage of the T allele of the TaqI VDR polymorphism and susceptibility to active TB in a case–control study conducted in Gambian adults⁽ Reference Bellamy, Ruwende and Corrah ^{45 )}. Wilkinson et al. subsequently reported that associations between susceptibility to TB and carriage of the T allele of the TaqI VDR polymorphism and the ff genotype of the FokI VDR polymorphism in Gujarati Asians living in London were restricted to vitamin D-deficient individuals⁽ Reference Wilkinson, Llewelyn and Toossi ^{36 )}; this study is the first to report that gene–environment interactions may operate to influence susceptibility to active TB. Numerous case–control studies investigating the association between VDR variants and susceptibility to active TB have since been published; a recent meta-analysis of twenty-three such studies reported that in Asian populations, the FokI ff genotype associated with susceptibility to active TB (pooled OR 2·0, 95% CI 1·3, 3·2), and the BsmI bb genotype (defined by the presence of two restriction sites for the Bsm1 endonuclease) was associated with protection against active TB (pooled OR 0·5, 95% CI 0·4, 0·8); no associations were seen in African or Latin American populations, however⁽ Reference Gao, Tao and Zhang ^{46 )}.

Further case–control studies have investigated associations between polymorphisms in the vitamin D binding protein (DBP) and susceptibility to active TB. DBP is a highly expressed multifunctional 58 kDa serum glycoprotein encoded on chromosome 4. Two common polymorphisms at codons 416 and 420 of exon 11 of the DBP gene give rise to the three major electrophoretic variants of DBP, termed group-specific component 1 fast, group-specific component 1 slow and group-specific component 2 (Gc2). These variants differ in their functional characteristics: the group-specific component 1 fast and group-specific component 1 slow variants have been reported to have greater affinity for 25(OH)D than the Gc2 variant⁽ Reference Arnaud and Constans ^{47 )}, potentially leading to more efficient delivery of 25(OH)D to the target tissues, while the Gc2 variant is associated with decreased circulating concentrations of 25(OH)D, 1,25(OH)₂D and DBP⁽ Reference Lauridsen, Vestergaard and Hermann ^{48 ,} Reference Abbas, Linseisen and Slanger ^{49 )}. Case–control studies conducted in India, Russia and Kuwait have not reported any association between DBP genotype and susceptibility to TB⁽ Reference Papiha, Agarwal and White ^{50 –} Reference Bahr, Eales and Nye ^{52 )}, but a more recent study reported an association between the Gc2 allele of DBP and susceptibility to active TB among Gujarati Asians living in London. This association was preserved if serum 25(OH)D concentration was <20 nmol/l, but not if serum 25(OH)D was ≥20 nmol/l, suggesting that profound vitamin D deficiency and Gc2 genotype may interact to increase susceptibility to TB⁽ Reference Martineau, Leandro and Anderson ^{43 )}.

Prospective observational studies

In contrast to the numbers of published cross-sectional studies, relatively few cohort studies investigating associations between vitamin D status or VDR genotype and TB have been conducted. Two studies have examined the influence of VDR genotype on response to antimicrobial therapy: Roth et al. reported that the FF genotype of the FokI VDR polymorphism and the Tt genotype of the TaqI VDR polymorphism associated with faster sputum culture conversion in a cohort of pulmonary TB patients in Peru⁽ Reference Roth, Soto and Arenas ^{53 )}, while Babb et al. reported no difference in time to sputum culture conversion according to TaqI or FokI VDR genotype among South African TB patients⁽ Reference Babb, van der Merwe and Beyers ^{54 )}. Recently, a cohort study conducted in Pakistan⁽ Reference Talat, Perry and Parsonnet ^{55 )} reported that profound vitamin D deficiency among healthy household TB contacts at baseline associated with increased risk of development of active TB over the subsequent 4 years: seven out of thirty contacts with baseline plasma 25(OH)D <17·5 nmol/l developed active TB during follow-up, compared with one of thirty-two with plasma 25(OH)D 17·5–33·5 nmol/l and none of thirty with plasma 25(OH)D >33·5 nmol/l. This association retained significance after adjustment for age and sex, although other potential confounders were not taken into account in the analysis. The observation that increased risk of TB reactivation was almost exclusively confined to individuals with profound vitamin D deficiency is interesting, particularly when taken together with reports from case–control studies that profound vitamin D deficiency is most strongly associated with susceptibility to TB⁽ Reference Wilkinson, Llewelyn and Toossi ^{36 )}: the implication is that, if vitamin D deficiency does indeed predispose to active TB, then relatively modest elevations of serum 25(OH)D might be effective for the prevention of active disease.

Intervention studies

Despite the compelling results from the laboratory and observational studies reviewed above, no randomised controlled trials of vitamin D supplementation for the prevention of active TB have been published to date. The absence of such studies reflects the very considerable methodological and logistic challenges of conducting them. Because the annual risk of reactivation of latent TB is low in immunocompetent individuals (<1% even in individuals with a strongly positive and newly converted tuberculin skin test⁽ Reference Horsburgh ^{2 )}), very large sample sizes and prolonged follow-up will be needed to detect all but the largest effects of vitamin D supplementation on TB incidence in such populations. One study has attempted to circumvent this problem by investigating the effect of vitamin D supplementation on a surrogate outcome measure of antimycobacterial response: the BCG-lux assay, which measures the ability of whole blood to restrict bioluminescence of a reporter mycobacterium⁽ Reference Kampmann, Gaora and Snewin ^{56 )}. This investigation found that a single dose of 2·5 mg vitamin D enhanced the ability of TB contacts’ whole blood to restrict mycobacterial bioluminescence at 24 h post-inoculation⁽ Reference Martineau, Wilkinson and Wilkinson ^{57 )}, providing further evidence that trials of vitamin D supplementation for the prevention of TB are justified.

In contrast to prevention studies, randomised controlled trials to determine whether adjunctive vitamin D enhances response to antimicrobial therapy can be powered with more modest numbers of participants, because the majority of TB patients respond rapidly to antimicrobial therapy. Seven such studies have been published to date (summarised in Table 1). Three of these trials had biochemical primary outcomes: two reported no hypercalcaemia in TB patients receiving either 125 μg vitamin D daily⁽ Reference Gwinup, Randazzo and Elias ^{58 )} or a single oral dose of 2·5 mg vitamin D⁽ Reference Martineau, Nanzer and Satkunam ^{59 )}, and one reported hypercalcaemia occurring in nineteen of thirty TB patients receiving daily doses of 10–95 μg vitamin D⁽ Reference Narang, Gupta and Jain ^{60 )}. However, this third study, by Narang et al., also reported that a daily dose of 60 μg vitamin D elevated mean serum Ca in healthy controls; a finding that contrasts with other studies which demonstrate that identical⁽ Reference Tjellesen, Hummer and Christiansen ^{61 )} or considerably higher⁽ Reference Stern, Taylor and Bell ^{62 )} doses of vitamin D do not induce hypercalcaemia in healthy people. It is possible, therefore, that the actual doses of vitamin D administered in Narang's study were considerably higher than reported. The remaining four clinical trials listed in the table had clinical primary outcomes. Morcos et al. investigated the effects of 25 μg vitamin D daily on twenty-four children in Egypt receiving antimicrobial therapy for TB, and showed no effect on body weight or resolution of symptoms⁽ Reference Morcos, Gabr and Samuel ^{63 )}. Nursyam et al. subsequently conducted a trial of a daily dose of 250 μg vitamin D in sixty-seven pulmonary TB patients in Indonesia⁽ Reference Nursyam, Amin and Rumende ^{64 )}. In this study, adjunctive vitamin D enhanced sputum smear conversion at 6 weeks after initiation of antimicrobial therapy (thirty-four out of thirty-four v. twenty-five out of thirty-three smear-converted in intervention v. control arm at 6 weeks, P=0·002); no effect of the intervention was seen at 8 weeks. The vitamin D status of participants was not assessed at either baseline or follow-up in this study, and details of safety monitoring, including monitoring of serum Ca concentrations, were not reported. In the largest treatment trial published to date, Wejse et al. randomised 365 adult TB patients in Guinea–Bissau to receive three doses of 2·5 mg vitamin D₃ or placebo at initiation of antimicrobial therapy, and again at 5 and 8 months⁽ Reference Wejse, Gomes and Rabna ^{65 )}. The intervention had no effect on the primary outcome measure (a specially designed TB score) or on serum 25(OH)D concentration. Mean serum 25(OH)D concentrations at baseline were 78 nmol/l v. 79 nmol/l in intervention v. control groups. Most recently, another trial investigated the effect of a 2-weekly dose of 2·5 mg vitamin D on time to sputum culture conversion in 146 patients with smear-positive pulmonary TB in the UK⁽ Reference Martineau, Timms and Bothamley ^{66 )}. A 79 nmol/l increase in serum 25(OH)D was seen among participants in the intervention arm of the study, which was associated with a non-statistically significant trend towards faster sputum culture conversion (P=0·14). A pre-planned subgroup analysis revealed that adjunctive vitamin D significantly hastened sputum culture conversion by more than 17 d in participants with the tt genotype of the TaqI VDR polymorphism (hazard ratio 8·09, 95% CI 1·36, 48·01; P=0·02).

Table 1. Summary of randomised controlled trials investigating effects of adjunctive vitamin D in patients with tuberculosis (TB)

* Type of vitamin D (D₂ v. D₃) not reported.