Molecular typing of Mycobacterium tuberculosis complex isolated from pulmonary tuberculosis patients in central Ethiopia

This study was performed at three different sites in central Ethiopia. These sites were Woliso and Atat towns and their surroundings in the southwest of Addis Ababa at a distance of 114 km and 187 km, respectively. The third site was Fiche town and its surrounding in the northwest of Addis Ababa at 115km. Sample collection was performed at hospitals located at these three sites, namely, St. Lukas, Atat and Fitche hospitals located at Woliso, Atat and Fitche tows, respectively. Sputum samples were collected from smear positive TB cases visiting these three hospitals.

A health institution-based cross-sectional study was conducted on 338 smear positive TB patients visiting three hospitals, between October 2012 and September 2013. Three consecutive sputum samples (spot, early morning and spot) were collected from each of these 338 smear positive TB patient and pooled together for culturing. Sample collection was performed prior to the beginning of TB treatment. TB patients less than 18 years old excluded from this study. Socio-demographic data of the patients were obtained from the medical records of all patients.

Those samples positive for acid fast bacilli (AFB) by Ziehl-Neelsen (ZN) staining technique were collected labeled and samples from the individual patient pooled together. These sputum samples for culture were stored at −20°C and then transported in a cold box (at +4°C) to Aklilu Lemma Institute of Pathobiology (ALIPB), Addis Ababa, within a week for culture.

Morning and spot sputum samples were collected and processed for culture following the WHO Guideline [14]. Briefly, equal volume of 4% NaOH was mixedwith sputum sample and the mixture was centrifuged at 3000 rpm for 15 min at room temperature. After decanting the supernatant the sediment was neutralized with 2 N HClusing phenol red as an indicator. Neutralization was achieved when the color of the solution was changed from purple to yellow. Thereafter, 100 μl of the suspension was inoculated ontotwo sterile LJ medium slopes (which were enriched with either pyruvate or glycerol). The inoculated media were then incubated at 37°C in slanted positionfor 1 week and upright position for 4–5 weeks. The growth of the bacteria was read every week until the 8^thweek of culture.

Colonies were removed from the surface of LJ medium and suspended in 200 μl of sterile double distilled water. Thereafter, the colonies and water were mixed thoroughly and then, the mixture was heated at 80°C for 1 h in water bath. This is followed by centrifugation after which the supernatant was collected and used for amplification [15].

Identification of M. tuberculosis from the other members of M. tuberculosis complexspecies was done using RD9-based PCR. RD9-PCRwas performed on heat-killed cells to confirm the presence or absence of RD9 using three primers namely, RD9flankF, RD9 IntR, and RD9flankR. Amplification was done by standard thermo cycler (VWR Thermo cycler, UK). The PCR amplification mixture used consisted of 10 μlHotStarTaqMaster Mix (Qiagen, United Kingdom), 7.1μl distilled water, 0.3 μl of each three primers and 2 μl of DNA template(heat killed cells), giving a total volume of 20 μl. The PCR reaction was heatedat 95°C for 15 min after which it was subjected to 35 cycles consisting of 95 °C for one min, 55 °C for one minutes, and 72 °C for one minute. Thereafter, the reaction mixture was maintained at 72 °C for 10 min following which the product was removed from the thermocycler and run on agarose gel electrophoresis. For gel electrophoresis, 8 μl PCR products was mixed with 2 μl loading dye, loaded onto 1.5% agarose gel and electophoresed at 100 V and 500 mA for 45 min. The gel was then visualized using a computerized Multi- Image Light Cabinet (VWR). M. tuberculosis H37Rv, M. bovisbacilleCalmette-Guérin, and water were included as positive and negative controls. Interpretation of the result was based on bands of different sizes, as previously described by Parsons et al. [16].

Isolates that were positive for M. tuberculosis by RD9 PCR were further characterized by spoligotyping following the procedure described by Kamerbeeket al [17] and by observing the instructions of the spoligotype kit supplier (Ocimum Biosolutions Company, Iisselstein, and the Netherlands). The direct repeat (DR) region of the isolate was amplified by PCR using oligonucleotide primers (DRa and DRb) derived from the DR sequence [17]. The amplified biotinylated products were hybridized to a set of 43 oligonucleotides covalently bound to a membrane (Animal and Plant Health Agency, Great Britain). Bound fragments were incubated with streptavidinperoxidase conjugate and hybridizing DNA was detected by the enhanced chemiluminescence method, by exposure to X-ray film (Hyperfilm ECL, Amersham) as specified by the manufacturer’s instruction. The presence and absence of spacers was visualized on the film as black and white squares, respectively. Characterized strains of M. bovisand M. tuberculosis H37Rv were used as positive controls, whereas Qiagen distilled water (Qiagen company, Germany) was used as a negative control.

The results of spoligotyping were converted into octal and binary formats. These binary and octal formats of the strains were entered into query box so that the name of the strains are retrieved from the database if the spoligotype pattern of the strain in question fits the pattern that has already been registered in the SPolDB4 database [18] and at http://​www.​pasteur-guadeloupe.​fr:​8081/​SITVITDemo/​ [10]. If the pattern of the strain in question has not been registered in SPolDB prior to this study, the strain was considered as an orphan. In this study, an isolate is referred to as a colony that was grown on LJ media and found to be AFB-positive after staining with Ziehl Neelsen staining, whereas a strain is an isolate (s) with specific spoligotype pattern. Thus, a strain can consist of a single isolate or several isolates. A strain with a single isolate is termed as a unique strain while a strain with more than one isolate is considered as clustered strain. An online tool Run TB-Lineage http://​tbinsight.​cs.​rpi.​edu/​run_​tb_​lineage.​html was also used to predict the major lineages using a conformal Bayesian network (CBN) analysis and sub lineage using knowledge based Bayesian network (KBBN).

Among 338 smear positive sputa samples297 (87.9%) samples were confirmed as culture positive. A total of 297M. tuberculosis isolates were utilized to carry out RD9-PCRand spoligotyping analysis of which 281 gave valid spolygotyping data while the remaining 16 isolates did not give any pattern up on spolygotyping. One hundred thirty (46.2%) of 281 isolates obtained from female while 151 (53.7%) were isolated from males. Classification of the study participated on the basis of age showed that 99 (35.3%) were between 18 and 28 years of age, majority (47.7%; 134/281) were originated from Woliso and its surroundings (Table 1).

Region of difference (RD) 9-based polymerase chain reaction (PCR). The 281 isolates were analyzed using RD9 PCR and the result indicated that all isolates had intact RD9 implying that all the isolates were M. tuberculosis (Fig. 1).

Spoligotyping of 281 isolates yielded 90 different spolygotype patterns. Of these, 32 strains were clustered containing of 223 (79.3%) isolates and 58 (20.6%) were unique strains (Tables 2 and 4). The overall diversity of the isolates was 32.4%. Out of the 90 spolygtype pattern, 45 were registered in the international data base (Table 2) and the remaining 45 were not found in the database (Table 3). The dominantly identified strains were SIT53, SIT149, and SIT 54 consisting of 43 isolates, 37 isolates, 34 isolates, respectively (Table 2). Classification of the spoligotype patterns using TB-insight RUN TB-Lineage showed that 86.8, 6.4, 5.3, 1.4% of the isolates belonged to the Euro-American lineage, East-African-Indian, Indo-oceanic and M. africanum, respectively (Table 2).

Of the 281 isolates typed, 134(47.7) were originated from Woliso and its surroundings, with clustering rate of 83.6% (112/134), 97(34.5) were originated from Fiche with clustering rate of 72.2% (70/97). The remaining isolates were from Atat town with clustering rate of 14.6% (41/281). This finding did not show statistically significant difference in the proportion of clustering across the three source sites of the isolates (p-value = 0.163). Similarly, there was not significant association of the source site of the isolate with major lineage identified by CBBN (p-value = 0.877) as well as the type of dominant isolate (p-value = 0.109). The proportions of occurrence of the dominant lineage (Euro-American) at Woliso, Fitche and Atat towns were 85.5% (112/131), 89.1%(90/101) and 85.7% (42/49), respectively (Table 4).