Association of IL-10 and CCL5 single nucleotide polymorphisms with tuberculosis in the Sudanese population

Authors


Abstract

Objective

To investigate the genetic determinants for developing tuberculosis in Sudan.

Methods

Case study of 232 patients with tuberculosis and 206 healthy matched controls from Sudan. In the study population, three single nucleotide polymorphisms (SNPs) in the promoter regions of CCL5 and two in the promoter region of IL-10 were genotyped using polymerase chain reaction and restriction fragment length polymorphism (PCR-RFLP). These five SNPs influence the expression of these genes.

Results

There were significant differences in allele distribution for CCL5 -28 C/G (rs 2280788) and IL-10 -592 A/C (rs1800872) in patients with tuberculosis compared with healthy controls.

Conclusion

This indicates that the genotypes obtained for CCL5 and IL-10 are associated with an increased risk of developing active TB.

Abstract

Objectif

Investiguer les déterminants génétiques pour le développement de la tuberculose au Soudan.

Méthodes

Etude de cas de 232 patients atteints de tuberculose et 206 témoins sains appariés provenant du Soudan. Dans l’étude de population, trois polymorphismes nucléotidiques (SNP) dans les régions du promoteur de CCL5 et deux dans la région du promoteur de l’IL-10 ont été génotypés en utilisant la réaction en chaîne de la polymérase et le polymorphisme des fragments de restriction (PCR-RFLP). Ces 5 SNP influencent l'expression de ces gènes.

Résultats

Il y avait des différences significatives dans la distribution des allèles de CCL5 −28 C/G (rs2280788) et de IL-10 −592 A/C (rs1800872) chez les patients tuberculeux comparés aux témoins sains.

Conclusion

Cela indique que les génotypes obtenus pour CCL5 et IL-10 sont associés à un risque accru de développer une tuberculose active.

Abstract

Objetivo

Investigar los determinantes genéticos de desarrollar tuberculosis en Sudán.

Métodos

Estudio de casos de 232 pacientes con tuberculosis y 206 controles sanos, pareados, realizado en Sudán. En la población de estudio, se genotiparon tres polimorfismos de nucleótido simple (SNPs) en las regiones del promotor de CCL5 y dos en la región del promotor de IL-10, utilizando las técnicas de la reacción en cadena de la polimerasa y el polimorfismo de longitud de fragmentos de restricción (PCR-RFLP). Estos 5 SNPs tienen influencia sobre la expresión de los genes estudiados.

Resultados

Había diferencias significativas en la distribución de alelos para CCL5 −28 C/G (rs 2280788) e IL-10 −592 A/C (rs1800872) en pacientes con tuberculosis, en comparación con los controles sanos.

Conclusión

Los resultados indican que los genotipos obtenidos para CCL5 e IL-10 están asociados con un riesgo aumentado de desarrollar una TB activa.

Introduction

Tuberculosis, caused by Mycobacterium tuberculosis, remains one of the leading infectious diseases with a high morbidity and mortality in humans (WHO 2007). In 2011, the global prevalence of tuberculosis was estimated to be 12 million cases, equivalent to 170 cases per 100 000 inhabitants (WHO 2012). Most of these cases were found in Asia (59%) and Africa (26%; WHO 2012). Although not particularly named in the global tuberculosis report of the World Health Organization of 2012, Sudan also has a relatively high prevalence for tuberculosis. In 2011, the prevalence of tuberculosis in Sudan was as high as 201 cases per 100 000 inhabitants (WHO 2011) and 22 of these patients died of this disease (WHO 2011). Although the prevalence of active tuberculosis is high, even more people are infected with M. tuberculosis; it is estimated that only 5–10% will develop clinical disease (Ates et al. 2008; Trajkov et al. 2009).

It is believed that the turning point of developing active tuberculosis is early in infection. Once inhaled into the lung, M. tuberculosis is phagocytised by alveolar macrophages, wherein it can replicate (Ramakrishnan 2012). As it is currently not known what happens in the human host, most assumptions are based on animal models. From animal models, it was learned that in latent tuberculosis infections, the infected macrophages migrate into tissues and aggregate into granulomas, while in active tuberculosis, some defect in granuloma formation is usually present (Flynn & Chan 2001; Ramakrishnan 2012). The C-C motif chemokine 5 (CCL5) and the cytokine interleukin 10 (IL-10) both are important in this process. CCL5 is known to play a significant role in the tuberculous granuloma formation by inhibiting the intracellular growth of M. tuberculosis in macrophages (Chensue et al. 1999; Saukkonen et al. 2002) and stimulation of phagocytosis and nitric oxide production in these macrophages (Lima et al. 1997). Thus, CCL5 potentially activates and regulates macrophage responses against M. tuberculosis infection.

IL-10 is an anti-inflammatory cytokine, which down-regulates the interferon-gamma (IFN-γ) production of T cells and the secretion of tumour necrosis factor (TNF-α) and nitric oxide (Tso et al. 2005). From IL-10 knockout mice and IL-10 over-expressing mice, it was learned that loss of IL-10 results only in a minor increase in resistance against tuberculosis, while overexpression of IL-10 shows no increase in susceptibility to tuberculosis but confers a reactivation susceptible phenotype (Tso et al. 2005). Therefore, IL-10 could play a role in the reactivation of latent tuberculosis (Tso et al. 2005).

As these assumptions are mostly based on animal models, and it is not always possible to transfer data from animal models directly to the human situations, we used a genetic study to determine whether there is an association between CCL-5 and IL-10 in humans. Many studies have shown that genetic factors such as single nucleotide polymorphisms (SNPs) can affect the activities of various chemokines and cytokines and are therefore found to be associated with disease susceptibility (Turner et al. 1997; Shin et al. 2005; Tian et al. 2009; Selvaraj et al. 2011). The most common SNPs described for CCL5 were -403 G/A (rs2107538), -28 C/G (rs2280788) and In1.1 T/C (rs2280789), where the -403A, -28G alleles were associated with high CCL5 production (Liu et al. 1999; Tian et al. 2009). The SNPs described for IL-10 are -819 T/C (rs1800871) and -592 A/C (rs1800872), where the C alleles for both polymorphisms were correlated with high IL-10 production (Turner et al. 1997). The association of these five SNPs with the development for tuberculosis has been studied in different populations with varying results. The association of the -403 G/A and -28 C/G SNPs in the CCL5 gene with tuberculosis was determined with PCR-RFLP in Spanish (Sanchez-Castanon et al. 2009), Tunisian (Ben-Selma et al. 2011a), Indian (Selvaraj et al. 2011) and Chinese (Chu et al. 2007) populations. It appeared that in the Spanish and Tunisian populations, the -403A and the -28G alleles were associated with tuberculosis, but not in Indian and Chinese populations (Chu et al. 2007; Sanchez-Castanon et al. 2009; Ben-Selma et al. 2011a; Selvaraj et al. 2011). For the Indian and Chinese populations, the In1.1 T/C SNP was also investigated (Chu et al. 2007; Selvaraj et al. 2011), but no association with this SNP was found for tuberculosis. The association of the -819 T/C SNP in the IL10 gene with tuberculosis was determined using various typing methods in Tunisian (Ben-Selma et al. 2011b), Turkish (Ates et al. 2008), Indian (Selvaraj et al. 2008) and Chinese (Ma et al. 2010) populations, but no association was found. The association of the -592 A/C SNP in the IL-10 gene with tuberculosis was determined using various typing methods in Tunisian (Ben-Selma et al. 2011b), Turkish (Ates et al. 2008), Chinese (Tso et al. 2005) and Korean (Shin et al. 2005) populations. Only in the Korean population, an association between the -592 A/C SNP and tuberculosis was found: in the population with active tuberculosis, the -592A allele was more prevalent (Shin et al. 2005). In summary, in previous population studies, the CCL5 -403A, the CCL5 -28G and the IL-10 -592A alleles have been associated with tuberculosis in certain populations but not in all. To determine the true role of these SNPs in the genetic predisposition in the development of tuberculosis, association studies in other populations need to be carried out. In this study, we describe the first association of these five expression SNPs in a sub-Saharan Sudanese population.

Materials and methods

Study population

A prospective, cross-sectional, observational, descriptive case study was conducted from 2008 to 2010 in Soba University Hospital, Abu-Anjaa Chest Hospital at Omdurman, Omdurman Teaching Hospital, Elshab Hospital, Alacademy, Baharri, Ebrahim Malik, Mayoo Hospital, Sudan. Blood samples were taken from all patients and healthy controls. All patients with tuberculosis had microbiological (by culture and/or smear) evidence of M. tuberculosis disease (Table 1). The healthy controls had no evidence of tuberculosis disease and were matched on age, gender and BCG status (Table 1). All patients with tuberculosis were new patients and had no previous anti-tuberculosis treatment. When the patients were diagnosed, they started for 2 months on a combination of isoniazid, rifampicin, pyrazinamide and ethambutol and continued for another 4 months on isoniazid and rifampicin only. The collected blood samples were further tested for other infectious diseases, such as hepatitis B (HBsAg, InTec products, INC, China), hepatitis C (Rapid Anti-HCV Test, InTec products, INC, China), syphilis (RAPIDAN TESTER, product code: RTTP01, Turkey) and HIV (HIV1, 2 Cassete test, Clinotech Diagnostics & Pharmaceuticals, Canada). Furthermore, all participants were subjected to a questionnaire interview to elicit age, sex, occupation, home, habits, past illness, history of having tuberculosis, X-ray and other clinical information. Of the 191 patients, 22 were alcoholic and 27 were smokers. The study was approved by the Ethics Committee of Soba University Hospital, Khartoum, Sudan. Written informed consent was obtained from all participants.

Table 1. Characteristics of the Study population
  PatientsControlsP-value
  1. a

    Unpaired t-test.

  2. b

    Fisher's exact test.

Total number 191206 
Mean age (range) 36.9 (15–89)31.2 (17–85)= 1.0a
Gender (male/female) 133/58130/76= 0.20b
OccupationGovernmental employee6 (3.1%)2 (0.9%)= 0.16b
Workers65 (34.0%)158 (76.7%)< 0.0001b
Other job4 (2.1%)2 (0.9%)= 0.43b
Jobless75 (39.3%)6 (2.9%)< 0.0001b
Housewife29 (15.2%)17 (8.3%)= 0.04b
Student12 (6.3%)21 (10.2%)= 0.21b
BCG vaccination 105 (55.0%)98 (47.6%)= 0.16b
Definite tuberculosisPresence of MTB in sputum based on both smear and culture22 (11.5%)0 (0%) 
Presence of MTB in sputum specimen only by smear97 (50.8%)0 (0%) 
Presence of MTB in sputum specimen only by culture72 (37.7%)0 (0%) 
HIV1,2 All negativeAll negative 
Hepatitis B All negativeAll negative 
Hepatitis C All negativeAll negative 

Culture

For each of the 191 patients, three sputum specimens were collected in sterile disposable containers. The patients were asked to cough up the sputum after taking a deep breath. Each sputum sample was decontaminated by adding an equal volume of 4% sodium hydroxide to 4 ml of sputum in an appropriate container. The sample was homogenised by shaking in a shaker or by hand for 15 min at 37 °C. The supernatant was removed from the pellet and added to a splash-roof discard container with 5% phenol. A few drops of hydrochloric acid with phenol red were added to the pellet to neutralise the alkaline reaction, and the neutralised pellet was inoculated on Löwenstein–Jensen medium and transferred on a slide for Ziehl–Neelsen (ZN) staining. The leftover pellet was stored for at least 3 days until it was confirmed that the inoculated medium was not contaminated. When the medium appeared to be contaminated, the stored pellet was treated with a small amount of 5% oxalic acid for 20–30 min to remove alkaline-resistant contaminants. It was diluted with sterile distilled water, centrifuged and inoculated onto media. From each specimen, three media were inoculated: two tubes with Löwenstein–Jensen medium containing glycerol, to improve the growth of M. tuberculosis but not M. bovis, and one tube with Löwenstein–Jensen medium containing sodium pyruvate to improve the growth of M. bovis. Then, the seeded solid media were incubated in a horizontal position for 3 days at 37 °C to prevent the inoculum from sliding to the bottom of the tube, and the subsequent incubation was placed in an upright position for a period of 8 weeks. The inoculated media were examined weekly for growth. If no growth had occurred, the plates were further incubated for up to 8 weeks before being discarded as negative. Bacterial suspensions from all positive cultures were placed in cryo vials (2 ml) containing glycerol and stored at −70 °C. M. tuberculosis was identified based on culture morphology and microscopic observation of cording formation on ZN stain. Confirmation was performed by subculturing on Löwenstein–Jensen medium containing 10 mg/l thiophene-2-carboxylic acid (TCH) and Löwenstein–Jensen medium containing 500 mg/l para-nitrobenzoic acid (PNB) and culturing for 4 weeks at 37 °C. If a strain was sensitive to PNB and resistant to TCH, it was considered to belong to the M. tuberculosis complex (Giampaglia et al. 2005).

Direct microscopy

From each sputum sample, two smears were prepared: one was prepared before decontamination and homogenisation and the other was prepared after decontamination and homogenisation. Both smears were stained with ZN stain and were examined microscopically for acid fast bacilli. For every side, more than 50 fields were examined (Lamden et al. 2000).

DNA isolation

Genomic DNA was isolated from blood samples with the large volume kit for the MagNA Pure system (Roche, Almere, the Netherlands) according to the manufacturer's descriptions. The isolated DNA was stored at −20 °C.

Genotyping

Genomic variants of CCL5 and IL-10 genes were detected by PCR followed by restriction enzyme fragment analysis (PCR-RFLP). All PCR primers and restriction enzymes are stated in Table 2. Each of the PCRs consisted of a pre-denaturation step of 4 min at 94 °C and 40 cycles each of 30 s denaturation at 94 °C, 30 s annealing at 55 °C and 30 s elongation at 72 °C. This was followed by a post-elongation step of 7 min at 72 °C. Restriction endonucleases were obtained from Fermentas (st. Leon-rot, Germany) and Roche (Penzberg, Germany) and were used as described by the manufacturer. Restriction fragments were visualised by electrophoresis on 2% agarose gels (Hispanagar, Sphaero Q, Leiden, the Netherlands).

Table 2. PCR primers and restriction enzymes for genotyping the different single nucleotide polymorphisms
Gene(s)Rs numberPrimer sequence(s) (5′→-3′)Restriction endonucleaseAlleleLength (bpa)Reference
  1. a

    bp: number of basepairs.

CCL5
-403G/A2107538R403fw CACAAGAGGACTCATTCCAACTCARsaIG180 + 26Bozzi et al. (2006)
R403rv GTTCCTGCTTATTCATTACAGATCGTAA206
-28C/G2280788R28fw ACTCCCCTTAGGGGATGCCCGTHincIIC152 + 23Bozzi et al. (2006)
R28rv GCGCAGAGGGCAGTAGCAATG175
Int1.1 T/C2280789RIfw CCTGGTCTTGACCACCACAMboIIC225 + 118Balloy & Chignard (2009)
RIrv GCTGACAGGCATGAGTCAGAT343
IL-10
-819T/C1800871IL-10F1 ATCCAAGACAACACTACTAAMaeIIIT292 + 217 + 79Costa et al. (2007)
IL-10R1 TAAATATCCTCAAAGTTCCC509 + 79
-592A/C1800872IL-10F1 ATCCAAGACAACACTACTAARsaIA306 + 232 + 42Costa et al. (2007)
IL-10R1 TAAATATCCTCAAAGTTCCC240 + 232 + 66

Statistical analysis

The mean age of the patient population and the control population were compared by the unpaired t-test. Gender, occupation and BCG-vaccination status between the patient and control population were compared with the Fisher's exact test. Verification of Hardy–Weinberg equilibrium (HWE) was performed in the control population with Pearson's chi-square test. The effect of the CCL5 and IL-10 polymorphisms on susceptibility to tuberculosis was assessed with the Fisher's exact test. P-value of < 0.05 was deemed statistically significant. All statistical analyses were performed using spss for Windows v11.0 statistical analysis software.

Results

Characteristics of patients with tuberculosis and healthy control subjects

The study population consisted of 191 tuberculosis patients identified based on a positive ZN smear of a sputum specimen and/or by positive culture (Osborne 1995). The control population comprised 206 healthy unrelated people from the same endemic area in Sudan, and they were matched on age, gender and BCG status (Table 1) and showed no signs of any lung disease. Unfortunately, the occupation of the control population differed from that of the patient population.

Distribution of CCL5 and IL-10 gene polymorphisms

To elucidate the possible deficiencies in CCL5 and IL10 production among patients with tuberculosis, genotype (Table 3) and allele frequencies (Table 4) in the promoters of the genes encoding for CCL5 and IL-10 were determined. To determine whether the SNPs reached HWE, the Pearson's chi-square test was performed. It appeared that in the control population, the genotype distributions for the CCL5 -28 C/G and CCL5 In1.1 T/C SNPs reached HWE, while the SNPs CCL5 -403 G/A, IL-10 -592 A/C and IL-10 -819 T/C were in disequilibrium (Table 3). The disequilibrium was caused by an excess of homozygosity in the control population. In the patient population, all genotype distributions were in Hardy–Weinberg disequilibrium.

Table 3. Genotype distributions and Hardy–Weinberg equilibrium in Sudanese patients with tuberculosis and healthy controls
GenotypePatients with tuberculosis = 191(%)HWE of patient population*Control = 206(%)HWE of control populationa
  1. a

    Hardy Weinberg Equilibrium (HWE) of population as calculated by the Pearson's chi-square test. A P-value <0.05 was considered with a deviation of HWE.

CCL5 -403 G/A
GG84 (43.97)<0.0168 (33.01)<0.01
AG7 (3.66)47 (22.81)
AA100 (52.36)91 (44.17)
CCL5 -28 C/G
CC183 (95.29)<0.01202 (98.05)0.88
CG1 (0.52)4 (1.94)
GG7 (3.66)0 (0.00)
CCL5 In1.1 T/C
TT182 (95.29)<0.01183 (88.83)0.40
TC5 (2.62)23 (11.16)
CC4 (2.09)0 (0.00)
IL-10-592A/C
AA127 (66.49)<0.01100 (48.54)<0.01
CA47 (24.60)68 (33.01)
CC17 (8.90)38 (18.44)
IL-10 - 819T/C
TT42 (21.99)<0.0170 (33.98)<0.01
TC126 (65.97)73 (35.44)
CC23 (12.04)63 (30.58)
Table 4. Allele frequencies of patients with tuberculosis in comparison with a matching healthy control population
CCL5 genotypePatients with tuberculosis patients = 191(%)Control = 206(%)P-value for AssociationOR (95% CI)
  1. P-values are calculated with the Fisher's exact test. Significant P-values are highlighted in bold.

CCL5 -403 G/A
G-allele175 (45.8)183 (44.4)0.721.06 (0.80–1.40)
A-allele207 (54.2)229 (55.6)
CCL5 -28 C/G
C-allele367 (96.1)408 (99.0) 0.009 0.24 (0.08–0.73)
G-allele15 (3.9)4 (1.0)
CCL5 In1.1 T/C
T-allele369 (96.6)389 (94.4)0.171.68 (0.84–3.36)
C-allele13 (3.4)23 (5.6)
IL-10 -592A/C
A-allele301 (78.8)268 (65.0) <0.0001 2.00 (1.45–2.75)
C-allele81 (21.2)144 (35.0)
IL-10 -819T/C
C-allele210 (55.0)213 (51.7)0.391.14 (0.86–1.51)
T-allele172 (45.0)199 (48.3)

To determine whether there was an association between any of the studied SNPs and tuberculosis, the allele frequencies between the control population and the patient population were compared with the Fisher's exact test. It appeared that in the tuberculosis population, the CCL5 -28G was significantly more often encountered (Fisher's exact test, = 0.03) than in the control population (Table 4). Furthermore, the IL-10 -592A allele was more commonly found in the tuberculosis population (Fisher's exact test, < 0.0001; Table 4). For the other polymorphisms tested, no association with tuberculosis was found in the population tested (Table 4). Therefore, only the CCL5 -28G and the IL-10 -592A alleles were associated with tuberculosis in the Sudanese population.

Discussion

Association between CCL5 and IL-10 gene polymorphisms and tuberculosis

As in many association studies, it is important to define the patient and control population as clearly as possible. In the present study, the patients with tuberculosis were identified based on ZN staining and/or a positive culture. As seen in Table 1, a large proportion of our patients was identified on ZN staining only, as is customary in resource-limited countries. Unfortunately, ZN staining is not specific for M. tuberculosis, other mycobacteria can be detected by this method as well. Therefore, it is possible that some of these patients had non-M. tuberculosis bacilli in their sputum. We still considered these patients to be likely to be suffering from tuberculosis, because they also had clinical signs and radiography indicative for tuberculosis. The identification of M. tuberculosis by culture was more reliable for the identification of M. tuberculosis, but unfortunately, we were only able to culture 72 isolates. By also using PNB and TCH selective media, the identification of M. tuberculosis complex was more reliable. The accuracy of this method to differentiate between non-M. tuberculosis strains and M. tuberculosis strains is 99.4% (Giampaglia et al. 2005).

A large number of studies have shown that polymorphisms in the CCL5 and IL-10 genes were implicated in susceptibility to tuberculosis in several populations (Shin et al. 2005; Tso et al. 2005; Chu et al. 2007; Ates et al. 2008; Selvaraj et al. 2008, 2011; Sanchez-Castanon et al. 2009; Ma et al. 2010; Ben-Selma et al. 2011a,b). In our study, we genotyped the five expression SNPs CCL5 -403 G/A, CCL5 -28 C/G, CCL5 In1.1, IL-10 -592 A/C and IL-10 -819 T/C in the Sudanese tuberculosis population. We chose these polymorphisms because these play a role in modulating the transcription of the genes (Liu et al. 1999).

In the Sudanese population, three of five SNPs tested here were not in HWE. Lack of HWE can be caused by genotyping errors or population stratification (Zintzaras 2010). Population stratification can include differences between groups of ethnic origin or differences between groups of similar ethnic origin but with limited admixture (Zintzaras 2010). The assays used here were also used by others to determine the allele frequencies of these selected SNPs, and no deviation was found in these studies, which makes genotyping errors less plausible (Forte et al. 2009; Andersen et al. 2010; Ben-Selma et al. 2011a; Qin et al. 2011). In the Sudanese population studied here, deviation of the HWE was caused by a notable excess of homozygosity. This excess of homozygosity in the Sudanese population for other SNPs has been noted before by Eid et al. (2010). Eid et al. (2010) determined a total of 26 informative SNPs in three different Sudanese Populations: Hausa, Massalit and Sinnar. For the Hausa and Massalit populations, the SNPs tested were in HWE; however, in the Sinnar population, eight of these SNPs deviated (Eid et al. 2010). In contrast to the Hausa and Massalit populations, the Sinnar population originates from Central Sudan. Most of our patients belonged to the Gawama'a tribe, which is the dominant tribe in the Umm Ruwaba District in the west of the Sudan. People from this tribe originate from sedentary farmers, who descended from both Nubians and Arabs from the North. When the Arabs settled in Kordofan during the 16th and 17th century, they mixed with the Nuba tribe and gradually replaced them (U.S. StCF 1988; Jakubaschk 2002). Around 6% of the population are nomadic (U.S. StCF 1988; Jakubaschk 2002). As there was no evidence of departure from random mating, the deviation of HWE may be due this mixed Arabic and African origin from the Gawama'a tribe with varying allele frequencies.

One of the genes studied that was in HWE in the control population was the CCL5 -28 C/G polymorphism. The allele frequencies of this polymorphism differed between the patients with tuberculosis and the control population. The CCL5 -28G allele was more frequently found in the patient population than in the healthy control population. The association of the CCL5 -28G allele with tuberculosis was also found in other populations originating from Tunisia (Ben-Selma et al. 2011a) and Spain (Sanchez-Castanon et al. 2009), but not in populations originating from India (Selvaraj et al. 2011; Mishra et al. 2012) and China (Chu et al. 2007). No association for the CCL5 -403 G/A and In1.1 T/C polymorphisms was found in our Sudanese population. The CCL5 -28G variant was previously reported to elevate promoter activity and increases CCL5 protein expression (Liu et al. 1999). Therefore, based on these population studies, it seems that the ability to generate a more elevated CCL5 level increases the risk of developing tuberculosis. This seems in contrast with the results obtained from an animal study published by Cardona (Cardona et al. 2003). Although the immune system from the mouse can differ from that from the human, most data are still generated from immune naïve animal models. In the study of Cardona et al. (2003) it was shown that lower CCL5 expression levels were found in the more susceptible DBA/2 mice than in C57BL/6 mice. But as both mice strains were able to cause tuberculosis and did express high levels of CCL5, this difference could still be an artefact (Cardona et al. 2003). The role of the high levels of CCL5 in the generation of tuberculosis in mice seems to be related to the granuloma formation. In vivo intraperitoneal CCL5 administration resulted in a trend to larger granuloma sizes, while anti-CCL5 treatment resulted in reduced granuloma sizes without altering the specific granuloma composition (Chensue et al. 1999). By using CCL5 KO mice, Vesosky demonstrated that the absence of CCL5 did not abolish granuloma formation, but it altered the early M. tubuerculosis granuloma development. At later time points, the lung granulomas in the CCL5 KO mice were significantly larger and more frequent than in the wild-type mice; furthermore, also the M. tuberculosis burden was increased (Vesosky et al. 2010). Based on both human and animal studies, it appears that CCL5 indeed plays a role in the development of tuberculosis, but the exact role in the human situation is still not clear and cannot be directly translated from mouse studies.

We also observed that in the Sudanese tuberculosis patient population, the IL-10 -592A allele was more often encountered than in the normal control population. This was also found for the tuberculosis cohorts in Korea (Shin et al. 2005) and China (Liang et al. 2011) but not for tuberculosis cohorts in Tunisia (Ben-Selma et al. 2011b), Columbia (Henao et al. 2006) and Turkey (Ates et al. 2008). The IL-10 -592A allele is associated with a lower IL-10 production. Therefore, an initial lower IL-10 production seems to be associated with the risk of developing tuberculosis. Again this seems in contrast with animal studies (Redford et al. 2010). Redford showed that infection with M. tuberculosis resulted in high expression of IL-10 and that IL-10 KO mice had a significant reduction in bacterial load in the lung as compared to WT control mice (Redford et al. 2010). The increased protection in IL-10 KO mice was associated with an accelerated and enhanced Th1 response in the lung (Redford et al. 2010). Expression of IL-10 was also related to therapy success, and the level of IL-10 was higher in early responders than in late responders (Su et al. 2010).

The genotypes associated with tuberculosis in our Sudanese cohort were associated with a higher CCL5 production and a lower IL-10 production. From the animal study, we learned that these phenotypes should diminish the infection instead of increasing the susceptibility. As the immune system of the mouse is different from that of the human, one should be careful by linking animal studies and human population studies. In addition to offering protection against the development of tuberculosis, the phenotypes found in this association study may not offer protection against the development of tuberculosis but prevent a bad outcome of the infection itself. Although many people succumb to tuberculosis, patients in our cohort survived their infection. In a prospective study performed on tuberculosis patients with culture-proven pulmonary tuberculosis and healthy subjects in a medical center in northern Taiwan, it was demonstrated that a high initial plasma level of anti-inflammatory cytokine IL-10 was associated with increased 6-month mortality (Wang et al. 2012); therefore, patients who produce lower levels of IL-10 were more likely to survive a tuberculosis infection. Clearly more studies into the role of CCL5 and IL-10 in the development of tuberculosis are needed to determine the exact role of these cytokines in this disease.

Acknowledgements

We thank Mr. Mohammed Noor for statistical support.

Ancillary