Vitamin D3 Status and the Association with Human Cathelicidin Expression in Patients with Different Clinical Forms of Active Tuberculosis

Low vitamin D (vitD3) is one of the most common nutritional deficiencies in the world known to be associated with numerous medical conditions including infections such as tuberculosis (TB). In this study, vitD3 status and its association with the antimicrobial peptide, human cathelicidin (LL-37), was investigated in Ethiopian patients with different clinical forms of TB. Patients with active TB (n = 77) and non-TB controls (n = 78) were enrolled in Ethiopia, while another group of non-TB controls (n = 62) was from Sweden. Active TB included pulmonary TB (n = 32), pleural TB (n = 20), and lymph node TB (n = 25). Concentrations of 25-hydroxyvitamin D3 (25(OH)D3) were assessed in plasma, while LL-37 mRNA was measured in peripheral blood and in samples obtained from the site of infection. Median 25(OH)D3 plasma levels in active TB patients were similar to Ethiopian non-TB controls (38.5 versus 35.0 nmol/L) and vitD3 deficiency (<50 nmol/L) was common in both groups (73%). Ethiopians (low latitude) had significantly lower 25(OH)D3 levels compared with Swedish non-TB controls (51.0 nmol/L, high latitude), but vitD3 status was not affected by tuberculin-positivity or HIV infection. Patients with local lymph node TB had significantly higher 25(OH)D3 levels compared with pulmonary TB patients (48.0 versus 29.0 nmol/L). Moreover, plasma 25(OH)D3 levels correlated with local LL-37 expression in granulomatous lesions in TB infected lymph nodes. Instead, systemic LL-37 mRNA expression in blood cells was elevated compared with the site of infection in pulmonary and pleural TB. Low vitD3 status may be associated with an enhanced peripheral expression of LL-37 in patients with intrathoracic TB that could result from chronic inflammation.


Introduction
Although tuberculosis (TB) is an ancient disease, it continues to be an enduring danger to society and remains the main cause of death from infectious diseases worldwide [1]. The susceptibility to Mycobacterium tuberculosis (Mtb), the causative agent of TB, as well as the course of TB disease is affected by both host-and bacteria-specific factors. Vitamin D 3 (vitD 3 ) status has been suggested to be associated with TB susceptibility and disease progression. The active form of vitD 3 is a steroid hormone with immunomodulatory properties and vitD 3 deficiency (VDD) has been associated with impaired control of Mtb infection [2,3]. VitD 3 is produced in the skin as a result of a direct exposure to sunlight, but can also be obtained via food intake or supplements [4]. Due to the importance of UVB light for the production of vitD 3 , VDD is subjected to seasonal variations [5,6]. It has been observed that VDD, defined as the concentration of 25-hydroxyvitamin D 3 (25(OH)D 3 ) below 50 nmol/L [7], was more common in dark-skinned people as strong skin pigmentation absorbs UVB sunlight and decreases the production of vitD 3 in the skin [8]. Accordingly, it has been reported that ethnicity was associated with a higher susceptibility to active TB [3,9]. VitD 3 has the ability to regulate multiple axes of the innate, as well as the adaptive, host immune response [10]. Induction of host defense peptides, such as human cathelicidin (LL-37), in phagocytes and epithelial cells is one of the specific functions ascribed to vitD 3 [11]. LL-37 has both antimicrobial and immunoregulatory activities, such as direct killing of bacteria via osmotic lysis, but also induction of autophagy, regulation of chemokine production and chemokine receptor expression, modulation of cytokine secretion and chemotactic effects on immune cells [12]. Several immune cell types, including neutrophils and alveolar macrophages, respond to Mtb infection by producing LL-37 in vitro [13,14], although it has also been shown that Mtb could suppress baseline LL-37 expression in human monocyte-derived macrophages [15]. Early during Mtb infection, LL-37 contributes to the control of bacterial growth and several studies have demonstrated that vitD 3 enhanced the expression of LL-37 and intracellular Mtb killing in macrophages in vitro [11,14,15]. Upon progression of TB disease, vitD 3 , as well as LL-37 itself, may contribute to dampening an excessive inflammatory response to Mtb by reducing the production of pro-inflammatory cytokines and promoting anti-inflammatory responses [16][17][18][19].
Despite a well-established role of vitD 3 in TB pathogenesis, it is still unknown how 25(OH)D 3 levels and LL-37 expression correlate with various forms of clinical TB. To address this question, we assessed vitD 3 status in TB patients and non-TB controls from an HIV-prevalent setting in Addis Ababa, Ethiopia, which is ranked seventh among high TB-burden countries. Plasma levels of 25(OH)D 3 in Ethiopian patients with different clinical forms of TB including pulmonary TB (PTB), pleural TB and lymph node TB (LNTB) were compared with non-TB controls from Ethiopia, as well as a non-TB control cohort from a low-endemic setting in Stockholm, Sweden. The expression of LL-37 mRNA was assessed in peripheral blood and in cells obtained from the sites of Mtb infection and related to patients' vitD 3 status.

Study Participants
TB patients and non-TB controls were recruited in December 2007 to December 2008 at the Chest Unit, Department of Internal Medicine, Black Lion University Hospital, Addis Ababa, Ethiopia, after providing signed informed consent [20,21]. Inclusion criteria: over 18-year old, HIV-negative or HIV-positive, sputum smear-negative patients with clinical symptoms and a clinical diagnosis of active TB. Exclusion criteria: history of previous TB, miliary TB, antimicrobial treatment longer than a week, antiretroviral treatment and a refusal to undergo HIV screening. Enrolled patients with active TB (n = 77) had either PTB (n = 32), pleural TB (n = 20) or LNTB (n = 25). PTB was confirmed by a positive Mtb culture of bronchoalveolar lavage (BAL), while pleural TB was confirmed by typical cyto-and histopathology of pleural fluid and biopsy samples and LNTB was similarly confirmed by typical histopathology of lymph node biopsy samples. Clinical TB diagnosis was based on: (1) assessment of the typical clinical TB symptoms lasting over four weeks, including persistent cough, general illness symptoms, such as fever, sweating, and weight loss; (2) pulmonary infiltrations/lesions or pleural effusions revealed by the chest X-ray examination; (3) occurrence of chronic non-tender cervical lymphadenopathy lasting over six weeks; (4) positive response to anti-TB chemotherapy. The Ethiopian non-TB control group (n = 78) consisted of both HIV-negative and HIV-positive individuals without any clinical symptoms of active TB. Additional non-TB controls used for mRNA expression analyses at the site of Mtb infection in the lung, pleura, and lymph nodes included patients with a minor hemoptysis (n = 10), patients with other non-TB pleural diseases (n = 5) and patients with non-specific lymphadenitis (n = 5). The Swedish non-TB control group (n = 62), consisted of HIV-negative individuals who were recruited at the Immunodeficiency Unit, Karolinska University Hospital, Huddinge, Sweden, as part of a previously described study [22]. Inclusion criteria: 18-75-year old and an increased susceptibility to respiratory tract infections defined as >42 days with symptoms from the respiratory tract during a 12-month period prior to study inclusion. Exclusion criteria: prophylactic antibiotic therapy, history of hypercalcemia or stones in the urinary tract, sarcoidosis, ongoing vitD 3 supplementation, HIV infection, and pregnancy.

Patient Materials
Plasma and peripheral blood mononuclear cells (PBMCs) were isolated from blood samples obtained from the Ethiopian study subjects at the time of diagnosis to assess 25(OH)D 3 levels in plasma and LL-37 mRNA expression in PBMCs. Part of the blood samples were used for HIV testing. Plasma was aliquoted and stored at −80 • C for vitD 3 analysis. Five million PBMCs were resuspended in 300 µL of RNAlater (Life Technologies, Invitrogen, Carlsbad, CA, USA) and stored at −80 • C for mRNA analysis. BAL fluid was collected from PTB patients with abnormal chest X-ray results using bronchoscopy as previously described [20]. Similarly, pleural fluid was collected from the pleura cavity of patients with pleural TB using thoracentesis [20]. BAL and pleural fluid aliquots were used for diagnostic purposes (Mtb-culture from BAL, and cytology from pleural fluid), while the remaining fluids were centrifuged at 1400 rpm for 25 min to pellet and resuspend the cells in RNAlater. Cell samples were stored at −80 • C for subsequent mRNA analysis. Lymph node biopsies were obtained through surgical removal of one lymph node in the cervical area from patients with a suspected LNTB. One part of the biopsy was used for diagnostic purposes (histopathology) and the remainder was snap frozen and stored at −80 • C for mRNA and immunohistochemical analysis.

Tuberculin Skin Test (TST)
The tuberculin skin test was applied to assess delayed type hypersensitivity towards Purified Protein Derivative (PPD). A total of 0.1 mL of PPD (5TU, SSI, Copenhagen, Denmark) was injected intradermally in the volar aspect of the forearm. The transverse diameter of induration in the injection site was measured by trained nurses 48-72 h later. A positive TST result was recorded when the diameter exceeded 10 mm in HIV-negative individuals, and 5 mm in HIV-positive individuals.

25(OH)D 3 Concentrations in Plasma
Plasma samples were collected from the study subjects and vitD 3 status was determined by assessment of 25(OH)D 3 concentrations in plasma using the DiaSorin assay conducted at the Chemical Laboratory, Karolinska University Hospital, Solna, Sweden. To demonstrate seasonality of vitD 3 status, the 25(OH)D 3 concentrations in TB patients and Swedish non-TB controls, were subdivided into two seasons: June-November and December-May. Ethiopian TB patients were sampled at the time of diagnosis, while longitudinal samples were obtained from the Swedish non-TB controls. Thus, matched samples were obtained from the Swedish non-TB controls in June-November (average 25(OH)D 3 concentration from two samples per individual) and December-May (average 25(OH)D 3 concentration from three samples per individual).

mRNA Extraction and Quantitative Real-Time PCR
RNA was extracted from cell samples (5 × 10 6 ) or frozen lymph node tissue sections (2 × 50 µm) using the Ambion RiboPure extraction kit (Invitrogen) according to the manufacturer's instructions. RNA was reverse transcribed to cDNA using superscript reverse transcriptase (SuperScript VILO cDNA master mix, Invitrogen). Amplification of target genes from cDNA was performed using ABI PRISM 7700 sequence detection system, and primer and probe sequences targeting the LL-37 gene were commercially purchased from Applied Biosystems (HS 00189030_m1) (Foster City, CA, USA). 18S was tested together with a panel of commonly used house-keeping genes and was selected as the calibrator as its expression was constitutive and stable in both test and control samples. Hence, Ct (cycle threshold) values were normalized to 18S to provide the delta Ct values. The relative mRNA expression was determined using the Livak method [23]. The data was analysed using 7500 software v2.0.6 (Applied Biosystems, Foster City, CA, USA) and presented as fold change of mRNA in the TB-positive groups (TB and TB/HIV co-infected) in relation to average Ct value recorded for non-TB controls.

Immunohistochemistry and Quantitative Image Analysis of Frozen Tissue Sections
Cryopreserved lymph node tissue biopsies were embedded in OCT-compound (Tissue-TEK, Sakura, Torrance, CA, USA) and cut into 8 µm-thick sections, mounted on HTC microscope slides (Histolabs, Gothenburg, Sweden) and fixed in 4% formaldehyde (Sigma, Stockholm, Sweden) for 15 min. The primary antibody used was polyclonal rabbit anti-human LL-37 (PA-LL37100; Innovagen, Lund, Sweden) and the biotinylated secondary antibody was swine anti-rabbit F(ab ' ) 2 (Dako, Glostrup, Denmark). Immunohistochemistry (IHC) was performed according to the ABC-method as previously described [24]. Positive staining was developed using a diaminobenzidine substrate (Vector Laboratories, Burlingame, CA, USA) while haematoxylin was used for nuclear counterstaining. Acquired computerized image analysis was applied to quantify IHC stainings in situ using light microscopy (Leica Microsystems, Germany) as previously described [25]. Positive immunostaining was quantified at the single-cell level in 25 to 50 high-power fields using a Qwin 550 software program (Leica Imaging Systems, Wetzlar, Germany). Protein expression was determined as the percent positive area of the total relevant cell area. Tissue sections stained with secondary antibodies only were used as negative controls.

Statistical Analysis
The data were analyzed using a non-parametric Kruskal-Wallis test with Dunn's post-test (comparing more than two unmatched groups) or the Mann-Whitney test (comparing two unmatched groups). A p-value < 0.05 was considered statistically significant. Correlation analyses were conducted using the Spearman's correlation test. Statistical analysis and preparation of graphs were performed using the GraphPad Prism software, v.7.03 (GraphPad Software, La Jolla, CA, USA).

VitD 3 Levels in Plasma from Ethiopian TB Patients as Well as Non-TB Controls were Significantly Lower Compared to Swedish Non-TB Controls
The demographics of the study participants are shown in Table 1. We performed a post hoc analysis of previous studies including Ethiopian patients with active TB (n = 77) and non-TB controls from Ethiopia (n = 78) [20,21] or non-TB controls from Sweden (n = 62) [22]. Gender, age, and numbers of HIV-infected subjects were similar between the Ethiopian groups. In the Swedish non-TB control group, we observed a relatively higher proportion of females and a higher median age compared with the Ethiopian cohort. Fewer TST-positive cases were also found among Ethiopian non-TB controls compared with active TB, which was expected. Interestingly, weight loss was relatively lower in the LNTB group compared with PTB and pleural TB. Furthermore, the occupational level was generally low among active TB patients, between 30% and 53%. The study subjects were tested for 25(OH)D 3 levels in plasma, showing that most of the Ethiopian TB patients as well as the non-TB controls were VDD (73%), defined as 25(OH)D 3 below 50 nmol/L ( Table 1 and Figure 1a). Interestingly, we found no difference in vitD 3 levels between TB patients and Ethiopian non-TB controls, while 25(OH)D 3 levels in the Swedish controls were significantly higher compared to both active TB (p = 0.008) and Ethiopian controls (p < 0.0001) (Figure 1a). VDD was also common in the Swedish cohort (50%), although most deficient cases were sampled during winter time (December-May) (Figure 1b). Accordingly, 25(OH)D 3 levels assessed from June to November were significantly higher (p = 0.001) in the Swedish controls (Figure 1b). A similar, but not significant, trend of relatively lower 25(OH)D 3 levels around the rainy season (June-November) was also observed in the Ethiopian TB patients (Figure 1b).

VitD 3 Status in Ethiopian TB Patients and Non-TB Controls was not Affected by Latent TB or HIV Infection but was Associated with the Clinical Form of TB Disease
Next, we compared vitD 3 status in active TB patients to non-TB controls depending on their TST or HIV status (Figure 2a). All TST-positive individuals had a positive Quantiferon-TB Gold test result, measuring Mtb-specific production of IFN-γ, which suggested that these subjects had latent TB (data not shown). Overall, no differences in vitD 3 levels were observed comparing these groups (Figure 2a). Neither could we detect any significant differences in vitD 3 comparing active TB with non-TB controls grouped into HIV-negative and HIV-positive individuals (Figure 2b). However, there was a tendency towards higher vitD 3 levels in TB/HIV co-infected patients as well as in HIV-positive patients without TB as compared to HIV-negative patients (Figure 2b). Finally, we also grouped the TB patients into the different clinical forms of TB disease: PTB (n = 32), pleural (n = 20) and LNTB (n = 25). Interestingly, patients with PTB had significantly lower (p = 0.034) vitD 3 levels than patients with extra-pulmonary LNTB, while pleural TB patients had intermediate vitD 3 levels (Figure 2c).

LL-37 mRNA Expression was Elevated in the Peripheral Blood of TB Patients Compared to non-TB Controls and also Systemically in PTB and Pleural TB Compared to the Site of Mtb Infection
LL-37 mRNA expression was assessed in PBMCs from TB patients and non-TB controls and also in cells obtained from the site of Mtb infection. There was a significant increase (p < 0.001) in LL-37 mRNA in PBMCs from active TB compared with non-TB controls (Figure 3a). Moreover, comparing LL-37 mRNA expression in peripheral blood to the site of infection, revealed significantly reduced (p < 0.001) levels at the site of Mtb infection in both PTB (BAL cells) and pleural TB (pleural fluid cells) (Figure 3b,c), while there was no difference in LL-37 mRNA comparing PBMCs to lymph node tissue from patients with LNTB (Figure 3d).

LL-37 Expression in Mtb-Infected Lymph Nodes Correlated to Plasma levels of 25(OH)D 3
We observed a positive correlation between LL-37 mRNA levels assessed in Mtb-infected lymph node samples and 25(OH)D 3 levels in plasma obtained from LNTB patients (Figure 4a). There was no correlation in the corresponding PBMC samples from these patients (Figure 4b). No LL-37 mRNA/25(OH)D 3 correlations were detected in PTB or pleural TB (data not shown). These results may imply that plasma levels of vitD 3 in patients with LNTB (Figure 2c) are associated with LL-37 levels at the local site of Mtb infection, but not in peripheral blood. Accordingly, we quantified protein expression of LL-37 in lymph node tissues from LNTB patients and non-TB controls using microscopy and computerized image analysis (Figure 4c,d). Local expression of LL-37 in LNTB was focused to granulomatous areas rich in neutrophils (cells with segmented nuclei [26]) while the expression in non-TB controls was lower and more randomly scattered (Figure 4c). Image analysis confirmed that protein expression of LL-37 was relatively higher in LNTB compared to the controls (Figure 4d), although there were large individual variations in LL-37 expression in the tissue. Altogether, these results may suggest that LL-37 expression is predominantly localized in the granulomas and, to a lesser extent, in the surrounding tissue.

Discussion
VDD is common in dark-skinned individuals living in Africa and is associated with many disease conditions, including an elevated risk to develop active TB [4]. In the present study, we demonstrate that Ethiopian patients with active TB were mostly VDD, and that VDD was not more common in TB patients compared with non-TB controls. Overall, 25(OH)D 3 levels were significantly higher in non-TB controls from Sweden, but only in plasma samples collected during summer time (June-November). Accordingly, there was a pronounced seasonal variability in vitD 3 status in this group. VitD 3 status in the Ethiopian cohort was not affected by either TST-positivity or HIV status. Patients with LNTB had higher 25(OH)D 3 levels in plasma compared to PTB and pleural TB, which correlated with increased expression of LL-37 mRNA at the site of Mtb infection in the lymph nodes. In contrast, patients with PTB and pleural TB showed elevated mRNA levels of LL-37 in peripheral blood compared to the site of infection, which may suggest that LL-37 in the circulation is rather a biomarker of active disease than representing an effective antimicrobial response at the local site.
A meta-analysis of observational studies from 1980-2006, concluded that there was a 70% probability that a healthy individual would have higher 25(OH)D 3 serum level than a patient with untreated TB [27]. A later meta-analysis found that the risk of active TB is dependent on the degree of VDD, as vitD 3 levels ≤25 nmol/L was significantly associated with TB, the range 26-50 nmol/L was associated with an intermediate risk for TB and the range 51-75 nmol/L showed no such association [28]. A systematic review mapping the extent of VDD in Africa, confirmed that VDD and vitD 3 insufficiency is prevalent in African TB patients [29]. However, a recently performed meta-analysis failed to show any association between VDD and an increased risk of TB in either HIV-negative or HIV-positive Africans [30]. Apparently, vitD 3 status can vary considerably between different latitudes, populations, ethnic groups, melanin pigmentation and season of sampling. Accordingly, case-control studies from the African continent have shown that TB patients from Ethiopia [31,32] and South Africa [31] were mostly VDD (<50 nmol/L) and also had lower vitD 3 levels than non-TB controls. Similarly, VDD in individuals with latent TB enhanced the risk of TST conversion [30], whereas sufficient 25(OH)D 3 levels (>75 nmol/L) might protect against TST conversion [33]. In addition, sputum-conversion after two months of standard anti-TB treatment was less likely in children with intrathoracic TB (PTB and/or pleural TB) who were VDD or vitD 3 insufficient compared to vitD 3 sufficient children [34]. In contrast, studies from Zimbabwe [35], Gambia [36] and Tanzania [37] demonstrated that PTB patients with a relatively low proportion of VDD, had higher vitD 3 levels than non-TB controls. Another study from Guinea-Bissau [38] found that VDD was not more frequent in PTB patients than in healthy controls, but the mean 25(OH)D 3 concentration remained lower. Importantly, mean 25(OH)D 3 concentrations varied significantly in these studies from 30-83 nmol/L and severe VDD (<25 nmol/L) was rare in most of these populations [35,36,38]. Combined, these studies support the assumption that active TB is associated with vitD 3 status and that individuals with VDD are at higher risk to develop the disease [28]. VitD 3 status is much dependent on seasonal variations, which is characteristic for high latitudes where both the quantity and quality of sunlight (UVB) differs considerably between winter and summer time and also compared to countries at low latitudes [39]. Accordingly, several studies have reported a seasonal pattern of TB incidence that correlated to low vitD 3 levels [39][40][41][42]. Although our study showed no difference in vitD 3 status between Ethiopian TB patients and non-TB controls, there was a trend of seasonality in this cohort with lower vitD 3 levels around the rainy season (June-November). As expected, the seasonality of vitD 3 status was more pronounced in the Swedish non-TB control group with significantly improved vitD 3 status around summer time (June-November). Overall, vitD 3 status in individuals from Stockholm, Sweden, located at the high latitude (59 • N) was significantly higher compared to both active TB patients and non-TB controls in Addis Ababa, Ethiopia, at low latitude (9 • N). This is consistent with results from a small cross-sectional study, revealing that 25(OH)D 3 concentrations were significantly higher in both non-pregnant and pregnant Norwegian females compared to Ethiopian females in samples obtained just after the corresponding winter season [43]. Thus, despite sunshine throughout the year in Ethiopia, several studies have shown that VDD is very common [31,32,43]. While variations in sun exposure and skin pigmentation cannot explain differences in low vitD 3 status in Ethiopians compared to other African populations, other factors could be of importance including diet (no food supplements) or behavior, such as avoiding direct sunlight and primarily indoor activities, especially during the rainy season [29].
Similarly to other reports, we found that vitD 3 levels in Ethiopians were not influenced by HIV status [6,35,41,44]. It has previously been described that a high proportion of TB/HIV co-infected South African patients were VDD and severe VDD was more common compared with active TB in HIV-negative patients, although VDD was associated with an enhanced TB susceptibility in both HIV-negative and HIV-positive individuals [31]. In line with our results, a small case-control study in Botswana showed that 78% of HIV-infected cases were VDD, however, no difference in vitD 3 status was detected comparing HIV patients with or without active PTB, suggesting that low vitD 3 levels are not a major risk factor of HIV-infected as compared with HIV-uninfected individuals [45]. Contrarily, the prevalence of optimal 25(OH)D 3 levels (>75 nmol/L) was relatively high among HIV-infected Ugandan patients with and without TB [46]. While many studies have been performed to investigate vitD 3 status in PTB patients and controls, few studies have addressed differences in 25(OH)D 3 levels in PTB to patients with extra-pulmonary forms of TB. In this study, we found that vitD 3 levels were relatively lower in patients with PTB, and also pleural TB, compared with LNTB. Local TB lymphadenitis is a common and relatively mild form of TB compared with PTB. Instead, significantly lower vitD 3 levels were previously shown in plasma and particularly in pleural fluid from patients with pleural TB compared with healthy controls [47]. It has been described that low vitD 3 levels and severe VDD were more common in extra-pulmonary TB compared with PTB, suggesting that VDD was primarily a risk factor for extra-pulmonary disease [3]. However, data from local LNTB only, was not presented in this study [3]. We have previously shown that PTB patients with severe VDD have down-modulated LL-37 mRNA and protein expression in granulomatous TB lesions in the lung [48]. Similarly, results from this study suggested that LL-37 was elevated in the peripheral circulation in patients with PTB and pleural TB, likely as a consequence of an acute phase response, rather than the induction of antimicrobial LL-37 at the site of Mtb infection in the lung. Instead, local LL-37 expression in LNTB correlated with relatively higher 25(OH)D 3 levels in plasma and LL-37 was confined to the granulomatous lesions in the lymph nodes. Interestingly, in vitro vitD 3 stimulation of human macrophages from PTB patients showed that LL-37 expression was more significantly induced in patients with non-cavitary TB compared to patients with more severe cavitary TB [49]. Thus, it is tempting to speculate that advanced or poorly controlled intrathoracic TB is more often associated with a low vitD 3 status.
Apart from vitD 3 status, it is also likely that genetic variations in the intracellular vitamin D receptor (VDR) may affect the susceptibility to TB infection and that VDR polymorphisms are affected by ethnicity [50]. Some polymorphisms may result in longer, more inactive VDR proteins, while other genetic variants could affect VDR mRNA stability [50]. Meta-analyses suggest that the VDR FokI polymorphism is associated with an increased risk of pulmonary TB in East Asians [51], while the t allele of the VDR TaqI polymorphism is significantly associated with an increased TB risk in South and West Asians [52]. In addition, certain VDR polymorphisms also contribute to TB susceptibility when observed in combination with VDD [53]. Therefore, although we did not determine the VDR genotype in this study cohort, genetic variations in the VDR and other proteins involved in the vitD 3 signaling pathway, such as the vitD 3 binding protein, may affect the susceptibility to TB.

Conclusions
In conclusion, this study underlines that the extent and degree of VDD can vary significantly between different study populations based on several independent factors that may influence the susceptibility to TB. While it is likely that VDD constitutes a risk to develop active TB, VDD may also affect, or be affected by, the severity of TB disease. Thus, more severe forms of intrathoracic TB may be associated with lower vitD 3 status and lower expression of antimicrobial LL-37 at the site of Mtb infection. Instead, higher systemic levels of LL-37 may be a consequence of reduced TB control and enhanced pathological inflammation.

Conflicts of Interest:
The authors declare no conflict of interest.