Genetic diversity and distribution of Mycobacterium tuberculosis genotypes in Limpopo, South Africa

Background Tuberculosis remains a major health problem and knowledge of the diversity of Mycobacterium tuberculosis strains in specific geographical regions can contribute to the control of the disease. This study describes the genetic profile of M. tuberculosis in five districts of Limpopo Province. Methods A total 487 isolates were collected from the National Health Laboratory Services from all regions/districts of Limpopo Province. Only 215 isolates were confirmed to be M. tuberculosis by Bactec Mycobacterium Growth Indicator Tube 960® and Rhodamine-Auramine staining. Isolates were subcultured on Löwenstein-Jensen medium agar slants to validate purity. They were spoligotyped and data analysed using the international spoligotyping database 4 (SpolDB4). Results Of the 215 isolates, 134 (62.3%) were genotyped into 21 genotype families while 81 (37.7%) were orphans. The 81 orphans were further subjected to resolution employing SpolDB3/RIM. Overall, the study revealed a high diversity of strains of 32 predominantly the non-Beijing lineages: the LAM- LAM3 (9.8%), LAM9 (4.7%) and LAM11- ZWE (3.3%), the T-T1(15.0%), T2 (0.9%), T2-T3 (1.4%), the CAS-CAS1-Delhi 5 (1.9%) and CAS1-KILI (1.4%) the MANU2 (1.4%), U (0.5%), X-X1(1.4%), X3 (1.9%), S (9.8%), CAS (1.4%), LAM7(0.9%), T3(0.5%), LAM8(4.7%), T4(1.4%), X2(0.4%), AI5(1.9%), LAM1(0.5%), FAMILY33 (1.9%), EAI4(1.4%), M. microti (1.9%). The Beijing and Beijing-like families were (14.9%) and (0.9%), respectively. A total of 28(13%) clusters and 77(36%) unique cases were identified. Beijing strain (SIT 1) formed the biggest cluster constituting 14%, followed by LAM3 (SIT 33), T1 (SIT 53) and LAM4 (SIT 811) with 7%, 5.1% and 2.8%, respectively. The Beijing family was the only genotype found in all the five districts and was predominant in Mopani (18.8%), Sekhukhune (23.7%) and Vhembe (23.3%). Dominant genotypes in Capricorn and Waterberg were LAM3 (11.9%) and T1 (13.3%), respectively. Conclusion A wide diversity of lineages was demonstrated at district level. A high number of clusters per district provided evidence of on-going transmission in this Province.


Background
Tuberculosis (TB) remains one of the major health challenges worldwide, despite the availability of a vaccine and treatment options [1]. South Africa is disproportionately affected by TB, with an estimated incidence of 940/100000 persons per year, probably driven by the high HIV frequency in South Africa [2,3]. The transmission dynamics of MTB lineages may depend on whether they are clustered to indicate an ongoing transmission or unique for reactivated TB [4]. The discovery of polymorphic DNA in M. tuberculosis strain has led to molecular typing of M. tuberculosis strains using techniques such as IS6110 restriction fragment length polymorphism (RFLP), spoligotyping and mycobacterial interspersed repetitive units of variable-number tandem repeats (MIRU-VNTR) [5,6]. Spoligotyping is a rapid reliable and cost-effective method based on analysis of polymorphism in the direct repeat (DR) chromosomal region with identical 36-bp DRs alternating with 35 to 41 bp unique spacers specific for studying the genetic diversity of M. tuberculosis complex members [7]. Spoligotyping has been used in molecular epidemiological studies conducted in South Africa and other countries [8][9][10][11][12]. Nine superfamilies or clades of M. tuberculosis complex have been identified, namely, M. africanum, Beijing, M. bovis, East African-Indian (EAI), central Asian (CAS), T group of families, Haarlem, X and Latin American Mediterranean (LAM) family [13]. Multiple mycobacterial lineages have been reported in several provinces of South Africa comprising of the Beijing and non-Beijing genotypes [8,[14][15][16]. However, the studies on South African genetic diversity of M. tuberculosis have been pitched at provincial levels, and most researchers have not paid any attention to such diversity within the provinces except for the study in three areas of the Free State Province [17]. Furthermore, Limpopo Province of South Africa shares its borders with Zimbabwe to the north, Mozambique to the east and Botswana to the west. Genotyping and continuous epidemiological monitoring of circulating strains of M. tuberculosis is crucial for TB control program and tracing of transmission. The province is comprised of rural and urban settings and has mining companies in rural areas and cities/towns. There remains possibilities or chances of divers M. tuberculosis complex strains due to increased cross migration from one area to another. A study by Stavrum et al. (2009) on diversity of MTB genotypes targeted 8 provinces in South Africa excluding the Limpopo Province [15]. One study by Sekati, Molepo and Nchabeleng (2015) on molecular characterisation of MTB strains tested only 5 isolates from the Limpopo Province [9]. Therefore, the purpose of this study was to determine the M. tuberculosis genotypes, their distribution and diversity around the districts of the Limpopo Province.

Sample size, collection and culturing
This was a cross-sectional study conducted between November 2012 and November 2013 and hence the sample size was calculated using the formular (Z 1-α/2 ) 2 P(1-P)/d 2 , with d of 0.05 according to Charan and Biswas (2013) [18]. With a proportion of 11.4% TB cases reported by NHLS during the period of the study in Limpopo Province, the required minimum TB positive sample size was 170 with an attrition of 10%. The samples were collected from archived isolates confirmed to be positive for tuberculosis (TB) from the National Laboratory Services (NHLS) in Polokwane. The NHLS is a referral laboratory for regional TB hospitals and clinics in Limpopo Province which provides laboratory and related public health services to more than 80% of the population through a National network of laboratories. Fig. 1 shows the map of South Africa and the location of Limpopo Province and its districts or regions where samples were collected [19,20].
Convenient sampling was used and data were extracted from NHLS, a major referral facility in the Limpopo Province, including demographic, data were collected from data registry at NHLS in Polokwane hospital covering age range of 1 year to 65 years and above consisting of 254 (52%) and 233 (48%) males and females, respectively. A total of 215 M. tuberculosis isolates identified by a positive rhodamine-auramine stain and a positive Mycobacterium Growth Indicator Tube 960 (MGIT) [NHLS Polokwane] were included in the study. In order to obtain a pure culture of M. tuberculosis strains, a loopful of inoculum from the MGIT tube was sub-cultured onto Löwenstein-Jensen agar slants (Diagnostic Media Products, Johannesburg, South Africa). The L-J cultures were incubated at 37°C for up to 6 weeks or until there was visible confluent growth.

Deoxyribonucleic acid (DNA) extraction and spoligotyping
The extraction of DNA from M. tuberculosis culture isolates was carried out by suspending a loopful of bacterial colony collected from LJ medium agar slants in 50 μl of distilled water in a screw capped micro centrifuge tube. The bacterial suspension was boiled for 20 min and then centrifuged for 5 min at 14000 g to sediment cell debris. The supernatant containing genomic DNA was recovered and stored at -20°C until further testing. Spoligotyping was carried out using the commercially available from Ocimum Biosolutions, India, by the standard method described before.7 The direct repeat region (DR) of genomic DNA was amplified using the primers DRa (biotinylated at the 5′ end) and DRb. Chromosomal DNA of M. tuberculosis strain H37Rv and M. bovis BCG P3 were included as positive controls. The amplified DNA was subsequently hybridized to a set of 43 oligonucleotide probes by reverse line blotting. The presence of spacers was visualized on film as black squares after incubation with streptavidin-peroxidase and detected with ECL (Enhanced Chemoluminescence System Detection Liquid) (Isogen Life Science).

Statistical analysis
The data were analyzed using Statistical Package for Social Sciences (SPSS) version 23. The Chi-square test was used to calculate the frequencies of lineages in each district. The results obtained from spoligotyping were entered in a binary format as excel spreadsheets (Microsoft). Spoligotyping patterns of the isolates were matched to the patterns in the Spoligotyping international database (SpolDB4) to identify the lineages. The strains with Spoligotype similar to any pattern of strain already existing in the database were labelled with already deefined 'shared type' number.
Isolates with no matching pattern in SpolDB4 were regarded as orphans. Isolates with same genetic pattern were regarded as a 'cluster' , indicating recent transmission of M. tuberculosis, while isolates with different genetic pattern were regarded as 'unique' , arising from distantly acquired or reactivation of TB infection.

Genetic diversity by spoligotyping
In this study, a total of 21 genotype families were identi-  Table 1 and Fig. 2, respectively.

Geographical distribution of Spoligotypes
The geographical distribution of M. tuberculosis Spoligotypes is shown in Table 1. A comparison of lineages among the five districts shows that the Beijing family was the only genotype found in all the five districts. The dominant genotype in the Capricorn District was LAM -16.2%) followed by T (14.0%) and Beijing (10.9%). In Mopani district, LAM (20.2%) was the most common followed by the Beijing (18.8). The Sekhukhune district was dominated by the Beijing family (26.3%) followed by T family (5.8%). In the Vhembe district, the Beijing family was the most common (23.3%) followed by the LAM11-ZWE (13.3%). The Waterberg district had a high frequency of Beijing genotype (10.0%). Of the 215 isolates, 134 isolates with identified lineages were grouped into clusters. At Province level, 15 clusters were identified. The highest number of clusters and genotype families was observed in the Capricorn district, while Mopani district had the least number of clusters and genotype families (Fig. 3).   The Vhembe district with Zimbabwe as a neighbouring country has the third highest frequency of TB with T1 (2.3%), T2 (3.3%), LAM11-ZWE (13.3%), LAM3 (0.9%), LAM4 (10.0%), LAM8 (1.4%) and MANU2 (3.3%) of the non-Beijing strains and 23.3%0f the Beijing strains. The situation in this district that has Mozambique and Zimbabwe as neighbouring countries is consistent with earlier reports indicating LAM Spoligotype described as the predominant strain amongst Zimbabweans [10,[21][22][23]. Mopani district with Mozambique as a neighbouring country has M. tuberculosis strains of Beijing (18.8%), LAM3 (6.3%), LAM4 (12.5%), U (6.3%), X2 (0.5%). Family 33 (0.9%) was also found in this district, which also have mines in this province. The Waterberg district with Botswana as its neighbour was found to have a wide strain family diversity, predominantly the Beijing family (10,0%) followed by CAS1 DEHLI (6.7%) and M. Microti (0.5%).
At Province level, 134 isolates out of the 215 genotyped isolates were grouped into 15 clusters and showed a 60% clustering percentage. Majority of the isolates were clustered in the Beijing (25%), T1 (18.5%), LAM3 (11.5%) and S (9.2%) lineages. Capricorn had the highest number of clusters followed by Waterberg, Vhembe, Sekhukhune and Mopani. Within Capricorn district, the highest number of isolates clustered in the LAM3 (11.9%), while in Mopani, Sekhukhune and Vhembe districts the highest number of isolates (18.8%), (23.7%) and (23.3%), respectively were clustered in the Beijing family. Waterberg had a high number of isolates (13.3%) clustered in T1 lineage.
A limitation of this study includes the use of only the spoligotyping technique, with a low discriminatory power, and thus may have overestimated the clustering results and limited further resolution of genotypes There is an inherent limitation within the Spoligotyping international database (SpolDB4) resulting in a high percentage of orphans (37.7%) in our study similar to the results of 43% orphans (17). However, the study provides valuable and useful information for comparison with other published studies from both within and outside of South Africa. Furthermore, the study has provided information regarding the diversity of M. tuberculosis strains in the five districts of Limpopo Province and areas where transmission of M. tuberculosis is high, suggesting the need for early diagnosis and timeous treatment.

Conclusion
The study demonstrated that the TB epidemic in the Limpopo Province is caused by a wide diversity of lineages. A high diversity of M. tuberculosis strains was also demonstrated at district level. The non-Beijing family was the most frequent with predominant LAM and T1 genotype lineages. A high number of clusters per district provided evidence of on-going transmission in this province. Further studies employing other techniques such as