Molecular epidemiology of Mycobacterium africanum in Ghana

Background Mycobacterium africanum comprises two phylogenetic lineages within the M. tuberculosis complex (MTBC) and is an important cause of human tuberculosis (TB) in West Africa. The reasons for this geographic restriction of M. africanum remain unclear. Here, we performed a prospective study to explore associations between the characteristics of TB patients and the MTBC lineages circulating in Ghana. Method We genotyped 1,211 MTBC isolates recovered from pulmonary TB patients recruited between 2012 and 2014 using single nucleotide polymorphism typing and spoligotyping. Associations between patient and pathogen variables were assessed using univariate and multivariate logistic regression. Results Of the 1,211 MTBC isolates analysed, 71.9 % (871) belonged to Lineage 4; 12.6 % (152) to Lineage 5 (also known as M. africanum West-Africa 1), 9.2 % (112) to Lineage 6 (also known as M. africanum West-Africa 2) and 0.6 % (7) to Mycobacterium bovis. Univariate analysis revealed that Lineage 6 strains were less likely to be isoniazid resistant compared to other strains (odds ratio = 0.25, 95 % confidence interval (CI): 0.05–0.77, P < 0.01). Multivariate analysis showed that Lineage 5 was significantly more common in patients from the Ewe ethnic group (adjusted odds ratio (adjOR): 2.79; 95 % CI: 1.47–5.29, P < 0.001) and Lineage 6 more likely to be found among HIV-co-infected TB patients (adjOR = 2.2; 95 % confidence interval (CI: 1.32–3.7, P < 0.001). Conclusion Our findings confirm the importance of M. africanum in Ghana and highlight the need to differentiate between Lineage 5 and Lineage 6, as these lineages differ in associated patient variables. Electronic supplementary material The online version of this article (doi:10.1186/s12879-016-1725-6) contains supplementary material, which is available to authorized users.


Background
Tuberculosis (TB) remains one of the main global public health problems, particularly in resource-limited settings [1]. With an estimated 9 million new cases annually, and a pool of two billion latently infected individuals, control efforts are struggling in most parts of the world. The TB problem is further exacerbated by a strong synergy with Human immunodeficiency virus infection and acquired immune deficiency syndrome (HIV/AIDS), which is a particularly big challenge in sub-Saharan Africa. Moreover, the global emergence of drug resistance increasingly complicates TB control. The Word Health Organization (WHO) estimates that in 2013, of the 1.1 million TB cases co-infected with HIV, 80 % occurred in Africa, making Africa the hardest hit of the two epidemics [1].
Human TB is mainly caused by Mycobacterium africanum (MAF) and Mycobacterium tuberculosis sensu stricto (MTB), both members of Mycobacterium tuberculosis complex (MTBC), which also includes several sub-species adapted to a variety of wild and domestic animals [2,3]. MAF first identified in 1968 in Senegal, was initially described biochemically as an intermediary between MTB and M. bovis [4]. Like MTB, MAF strains were found to be sensitive to pyrazinamide; like M. bovis, they tended to be a weak producer of niacin, microaerophilic, and unable to reduce nitrate to nitrite [5]. Furthermore, similar to M. bovis they are unable to use glycerol as a sole carbon source due to the lack of functional pyruvate kinase [6]. Initial biochemical features subdivided MAF into two separate groups; the West-African sub-species and East-African also referred to as MAF type 1 and 2 [7]. However, based on genomic deletion analyses [2,8,9], we now know that MAF type 2 variant is a sub-lineage of MTB Lineage 4, which has been reclassified as MTB "Uganda sublineage" [10]. MAF type 1 or the "true" M. africanum consists of two phylogenetically distinct lineages that differ in their geographic distribution: Lineage 5 (also known as MAF West African 1) is found in the eastern part of West-Africa, and Lineage 6 (also known as West-Africa genotype II) is mainly found in the western part of West-Africa. A few countries like Ghana and Cote d'Ivoire harbor both Lineages 5 and 6 [9,11].
The reason why MAF is largely restricted to the West African region remains unknown, but one possibility might be that MAF is adapted to distinct human population (s) in West Africa [9]. In support of this hypothesis, using a limited retrospective collection of MTBC isolates from South-western Ghana, we recently reported an association between MAF and the Ewe ethnic group [12]. One of the limitations of that previous study was that due to lack of relevant data, we could not control for possible confounding by HIV co-infection. Hence, we followed-up on this initial observation with a larger population-based prospective study in which we included more detailed clinical data and TB cases from both the southern and northern part of Ghana.

Ethical Statement and Patient inclusion criteria
The Institutional Review Board of the Noguchi Memorial Institute for Medical Research (NMIMR) and the Ethikkommission Beider Basel (EKBB) in Basel, Switzerland approved the study and its protocols. Following informed consent, consecutive sputum smear-positive TB cases were recruited from September 2012 to April 2014 from all TB diagnostic centres in the Accra Metropolitan (Southern) and the Mamprusi East (Northern Region) Health Administrations. The standard procedure for sputum sample collection as outlined by the National Tuberculosis Program (NTP) for routine diagnosis of TB in Ghana was used in this study. Participants provided written consent unless the participant was illiterate; in which case witnessed oral consent was used. In accordance with ethical review board regulation in Ghana consent was sought from guardians of children below the age of 18 before enrolment into the study and in some cases child assent was also sought. Samples were taken only after a detailed explanation of the study aims and written or thumb-printed consent obtained for participation. Only newly diagnosed smear-positive, pulmonary TB patients before initiation of treatment or less than 4 weeks of treatment were enrolled into the study. All eligible TB patients were encouraged to undergo HIV testing before initiation of anti-TB treatment according to national guidelines. Data collected from enrolled patients included age, sex, immunosuppressive co-morbidity with HIV, place of work, previous episode of TB, housing type, place of residence, ethnicity, smoking, level of education, income, and presence of Bacillus Calmette-Guerin (BCG) scar. All patients and staff involved with the study were blinded to the final data obtained and none of the authors have a conflict of interest in the study.

Laboratory tests
The sputum samples were re-examined for the presence of acid-fast bacilli (AFBs) at NMIMR and cultured on solid agar, after which isolate deoxyribonucleic acid (DNA) was extracted for genotyping analysis. Briefly, we took a 10 μL loop full from a Lowenstein-Jensen medium slope and, after heat killing, extracted DNA first by digestion with lysozyme and proteinase K, solubilized by detergents sodium dodecyl sulphate and cetrimonium bromide, followed by chloroform isopropanol extraction [13].

Genotyping of MTBC isolates
Classification of mycobacterial isolates into the main phylogenetic lineages within the MTBC were by TaqMan real-time PCR (TaqMan, Applied Bio systems, USA) using probes targeting lineage-specific SNPs as reported by Stucki et al. [14]. Spoligotyping was performed to define the sub-lineages and strain families within each of the main lineages circulating in Ghana [15]. Spoligotyping patterns were defined according to SITVITWEB database [16] (http://www.pasteur-guadeloupe.fr: 8081/ SITVIT_ONLINE). SITVITWEB assigned shared types numbers were used whenever a Spoligotyping pattern was found in the database while families and subfamilies were assigned based on the MIRU-VNTRplus database (http: //www. miru-vntrplus.org) [17].

Anti-TB Drug Susceptibility Testing
All MTBC isolates were screened for their susceptibility to isoniazid (INH) and Rifampicin (RIF) using the Genotype MTBDRplus (Hain lifescience), according to the manufacturer's protocol [18]. Briefly, drug resistance was expressed as the absence of wild-type band, presence of mutation band or both.

Data entry, management and analysis
Information from the structured questionnaire was double entered using Microsoft Access and validated to remove duplicates and data entry inconsistencies. Spoligotype patterns were entered in a binary format. A series of univariate and multivariable logistic regression models were fitted to assess the relationship between MTBC lineage (s) (primary independent variable) and host variables. For the MTBDRplus assay, definition of specific drug resistance was based on mutations within the katG gene and inhA promoter for INH and rifampicin resistance determining region of the rpoB gene for RIF. This was captured as either absence of specific wild-type band and/or presence of mutation band corresponding to specific mutations.

Patient characteristics
From September 2012 to April 2014, 1,330 smearpositive TB cases recruited from various diagnostic laboratories in the Greater Accra Region and the Northern Region of Ghana were enrolled into the study. We excluded 80 cases from further analysis because no MTBC cultures were obtained from their sputa. A further 30 strains identified as non-tuberculous mycobacteria and 9 strains which produced ambiguous genotyping results were excluded, leaving 1,211 unique MTBC patient isolates for further analyses.
A summary of characteristics of the study population is provided in    Spoligotyping of the 264 MAF (Lineage 5 and 6 combined) isolates revealed 92 distinct spoligotypes patterns. Fifty-two unique patterns (singletons) were observed with the remaining 40 patterns grouped into clusters of 2-26 isolates respectively (Additional file 1: Table S1). Within the 940 MTB isolates, we had 101 patterns, consisting of 60 patterns grouped into clusters of 2-349 isolates respectively, 19 singletons and remaining 11 isolates identified as orphans. We compared the 193 spoligotypes found in this study with those contained in an international spoligotype database, 130 of our spoligotypes were already described in SITVIT database. The other 63 spoligotypes were novel.

Discussion and conclusions
Our molecular epidemiological analysis revealed that i) MAF remains an important pathogen in Ghana, causing one fifth of all human TB cases in Ghana, ii) TB patients of the Ewe ethnicity were more likely to be infected with Lineage 5, and iii) Lineage 6 isolates were less likely to be INH-resistant and more frequent in TB patients coinfected with HIV.
The observation that MAF remains an important cause of pulmonary TB in Ghana is in contrast to the decline of MAF that has been reported in Cameroon [19]. In that country, MAF might gradually being outcompeted by more virulent MTB. In support for this hypothesis, MAF has been associated with reduced virulence in animal models [4], and a longer latency and a slower rate of progression to active disease in humans [9]. The reason for the persistence of MAF in Ghana over time is unclear but might be partially related to our previous finding that MAF is associated with the Ewe ethnic group [12]. Lineage 5, a finding that we replicated in the present study, drove this association. While we were not able to control for HIV infection in our previous study, in the present study, the patient HIV status was available and included in the analysis. Yet despite this potential confounding factor, the association between Lineage 5 and Ewe ethnicity remained significant. Based on these findings, we hypothesize that Lineage 5 is particularly well adapted to infect and cause TB in the Ewe population [12].
This hypothesis follows on an increasing body of evidence indicating that different MTBC genotypes might be adapted to different human populations [20][21][22][23]. Moreover, several studies have reported interactions between host and pathogen genotypes in human TB [22][23][24]. For example, a study in Vietnam found an association between the T597C allele of the Toll-like receptor 2 gene (TLR2) and infection by Lineage 2 [25]. Similarly, studies performed in Ghana have reported human polymorphism in 5-lipoxygenase (ALOX5) [26] and Mannose Binding Lectin (MBL) which gives protection to TB caused by either MAF or MTB [27]. Most recently, a study in South Africa reported on the association between different HLA class I types and disease caused by different MTBC strain   All variables with an OR above 1.5 or below 2/3 were included in the multivariable model, *P < 0.05, **P < 0.001, MAF = Mycobacterium africanum MTB mycobacterium tuberculosis sensu stricto families in a South African indigenous population [28]. More studies are needed to confirm the hypothesis that the genetic background of Ewe populations is particularly conducive to infection and disease caused by Lineage 5 and to identify the biological processes involved in this interaction.
Even if the association between Lineage 5 and Ewe ethnicity is contributing to the stable prevalence of MAF in Ghana over time, about half of all MAF cases are due to Lineage 6, which is phylogenetically distinct from Lineage 5 [29]. The findings reported here indicate that Lineage 5 and Lineage 6 also differ phenotypically with respect to their associations with particular patient variables. We found that Lineage 6 was less likely to show any form of INH resistance. This is consistent with reports from Mali that found Lineage 6 less likely to be multidrug-resistant compared to other MTBC lineages [30]. Drug resistance in many bacteria, including MTBC is often associated with a fitness cost [31]. Hence, the fact that Lineage 6 is already attenuated compared to other MTBC lineages [9,32] might limit its potential for drug resistance acquisition. In support of this idea, Lineage 6 has been associated with inhA promoter mutations conferring low levels of resistance to isoniazid that have no known impact on bacterial fitness [33]. In contrast, most other MTBC strains tend to harbour mutations in katG conferring high levels of resistance, some of which are associated with a high fitness cost [34].
The relative attenuation of Lineage 6 has led to the hypothesis that this lineage might be an opportunistic pathogen [35]. This view is supported by the findings reported here, as well as previous work by others [9] showing an association between Lineage 6 and HIV coinfection [34]. A previous study from Ghana found no such association, perhaps reflecting differences in study design [36]. Moreover, de Jong et al. reported that Lineage 6 was significantly more frequent in elderly TB patients, consistent with both slower progression to active disease [37] and the role of age-related immunosuppression [34]. Recent advances in phylogenomics of the MTBC show that even though Lineage 6 has primarily been isolated from humans, it is phylogenetically most closely related to several animal-adapted variants of the MTBC [24] including the recently discovered chimpanzee bacillus [38] as well as M. suricattae and M. mungi which cause TB-like disease in meerkats and banded mangooses, respectively [39]. The potential for an unknown animal reservoir for Lineage 6 has been discussed for some time but so far, none has been discovered [9]. Moreover, unlike other animal-adapted members of the MTBC that rarely cause TB in humans and cannot maintain a full life cycle in human populations [40], Lineage 6 transmits readily among humans [9]. So far, the role of HIV in driving Lineage 6 remains unclear.
Even though HIV had a dramatic impact on the TB epidemic in Sub-Saharan Africa, Lineage 6 remains restricted to the Western part of the continent, which shows much lower rates of HIV infection compared to Southern Africa. Hence, other factors are likely to determine to epidemiology of Lineage 6.
In conclusion, our study confirms that MAF remains an important cause of human TB in Ghana. Our findings also emphasize the need to differentiate between Lineage 5 and 6, and show that the differences between these lineages are relevant for understanding the phylogeography of MAF; while human genetic factors may drive Lineage 5, unknown environmental factors seem to influence the epidemiology of Lineage 6. More generally, these results also highlight the current problems with the species nomenclature in the MTBC [39]. Indeed, the phylogenetic, phenotypic, and ecological differences between Lineage 5 and Lineage 6 argue against them being collectively referred to as MAF, as opposed to just separate lineages within MTB. Finally, given the current efforts in the development of new TB vaccines, our findings also highlight the need to consider the MTBC variability in the development of new tools and strategies to control TB especially in area where Lineages 5 and 6 are prevalent.

Additional files
Additional file 1: