Tracking characterization of Mycobacterium strains in Tanzania and some sub-Saharan African Countries: An overview on genotyping studies, implication and trends in advancing technology

Tuberculosis (TB) is a devastating chronic debilitating infection that imposes a considerable negative impact on human health. The disease also affects the quality of livestock, wildlife and their products. Mycobacterium tuberculosis (M. tuberculosis) is the causative agent for human TB, while Mycobacterium bovis (M. bovis) predominantly causes TB in animals, though is also zoonotic. In Tanzania and other developing countries timely diagnosis of disease is hampered by poor access to appropriate capacity and technology to characterize the pathogen. This review explores the diverse methods available for the diagnosis and characterization of Mycobacterial species and strains isolated from both humans and animals in sub Saharan Africa. The review will identify knowledge gaps and highlight direction for future investigation of the interface TB risk, which could lead to a better regional control strategy. A systematic search of PubMed, Google and Google Scholar retrieved 716 published articles on TB and methods used with the aim of tracking the advancement in technology and to reflect where we are and what can best be done to improve the strategy for best control of the disease. The search terms included but not limited to “(Whole Genome Sequencing AND sub-Saharan Africa); (Diverse TB strains + Human +Animals) AND Tanzania + Africa + Sub-Saharan Africa; (Diverse TB strains + human+ animals) OR Tanzania + Africa + Sub-Saharan Africa)”. The review illustrates an advancement of technology from 1950s to 2000s with only 17.7% studies having been done using DNA-based methods and 81.3% being studies that used conventional methods. Most of the molecular studies cluster in the midto late 2000s which could be due to lack of expertise, slow adoption of technology or the high cost of running these valuable molecular tests. This overview on genotyping studies and trend in molecular studies highlights the need for more investment in this region of the world. An increased use of molecular methods will help in increasing the chances of detecting new TB strains in circulation and identifying potential risks for cross-transmission at humans-livestock-wildlife interface. A strengthened ability to detect and characterizes disease will better support country and regional control strategies.

INTRODUCTION uberculosis (TB) is a devastating chronic debilitating infection which negatively impacts on the health and wellbeing of humans, and livestock and wildlife and their products. With potential for cross-transmission of the disease amongst species, an inter-sectoral participatory approach that involve multiple disciplines (One Health approach) could be relevant as a control and disease management strategy, particularly in developing countries. Whilst Mycobacterium Tuberculosis is predominantly the causative agent for human TB, Mycobacterium bovis predominantly causes TB in animals, though is also zoonotic. Zoonotic TB risk factors include consumption of animal products, particularly raw milk and uncooked meat. Transmission of bovine tuberculosis (bTB) by M. bovis from cattle to humans is well known but little is known of transmission of bTB between animals in Africa and risk of transmission to from humans. Many societies in Africa share shelters with animals, particularly cattle and goats. It is therefore important to explore the disease dynamics in these interactions. Understanding the habitual practices, and being able to identify the prevalent TB strains in circulation in sub-Saharan Africa are important to design suitable interventions to improve health in both humans and animals. In this review, we explore the diversity of Mycobacterial species and strains isolated from both humans and animals in sub-Saharan Africa in order to identify knowledge gaps to drive further research and support an evidence based regional control strategy.
TB occurs worldwide with about 10 million cases and 1.6 million deaths in 2017 (WHO). TB is amongst the top causes of deaths in women aged 15 to 44 years 1 and is even more prevalent in males. According to the World Health Organization (WHO) reports the largest number of new TB cases occurred in Asia, ( 62% of new cases globally in the last 2 years) and then sub-Saharan African (25% of all new cases in 2017) 2 . The global impact of TB is exacerbated by co-infection with HIV 8 and drug resistance [3][4][5][6][7][8][9][10] . Extensive drug resistant TB (XDR-TB) has been widely reported in many TB endemic countries 8,[11][12][13][14][15][16][17] . Tanzania is among the African countries with highest burden of TB in the world; having approximately 295 TB cases per 100,000 adults. In addition, Tanzania has a national HIV prevalence of 5%, which increases risk of TB infection by over 30 times 1 .
Existing economic and social obstacles challenge global TB elimination 5 . Treatment and prevention strategies are hampered by socio-economic problems and limited political will. Such strategies rely on better diagnostics to increase case detection 5 , and to reduce the risks of exposure. The implementation of disease control policies, including case management, require adequate and consistent funding 5 . Part of the strategy should bring in the concept of 'One Health' with medical, veterinary and other stakeholders such as environmentalists which could be a key control strategy 18 .

Animal tuberculosis, zoonotic and epizoonotic implications
Bovine TB (bTB) in humans is a well-documented zoonosis and is thought to be acquired mainly from cattle through the consumption of infected meat or milk, or via the respiratory route. However, humans can also be a source of infection to livestock, for example there are reports of the isolation of M. tuberculosis in goats and sheep in Ethiopia 19 and even a wild elephants in South Africa 20 . The epizoonotic and zoonotic nature of TB would benefit from a One Health approach to control. For example, the use of genotyping and whole genome sequencing methods, can help to identify sources and transmission routes within and across species and geographic regions 21 , which may identify points for intervention. These advances in technology, however, need committed and coordinated efforts from both within and outside the health sector 22 . In addition, it will require willingness from governments in endemic areas to invest in TB disease control programs 18 and not entirely depend on donor-funded programs.
There is insufficient information on the distribution, epidemiology patterns and zoonotic potential of bTB in all environments, but perhaps most pertinently in traditional rural environments 23 . Putting in place diagnostic programs focusing on isolation, culture and pattern discrimination of mycobacteria from human and animal TB cases is important for pinpointing the mechanisms and routes of disease transmission, spread and zoonotic impact. Undiagnosed cases contribute towards the force of infection and disease recurrence 24 . This is particularly so in resource-poor countries (mostly disease endemic African countries) where rare resource sharing could be critically important. In such resource sharing fashion, mycobacterial culture as well as isolation facilities could be done at reduced cost, consequently, enhancing effective and efficient diagnosis 23 . However, diagnosis is only the first step towards control and unfortunately, the test and slaughter strategy for TB positive animals as a control measure is rarely used in African countries, for a variety of reasons including cultural backlash and the lack of funds for adequate compensation. Thus, without change, developing countries (and particularly African countries) will remain focal points for a number of infections in both humans and animals, adversely affecting their economy. The eradication of TB will not be possible without good cooperation and collaboration between medical and veterinary personnel for effective and costefficient investigation, prevention, control and management of the disease 23 . This is well supported by the East African Community which has so far conducted a series of curriculum validation meetings to advocate regional pandemic preparedness with a 'One Health' approach. This is an important step because, although it targets pandemics other than TB, it builds the capacity of practitioners for the development of a common approach for the control of zoonotic infections. A missing link previously identified by Mbugi et al 25 .
The epidemiology of tuberculosis has been extensively studied for decades, and considerable technological advancements have increased our ability to explore the disease and its causative agent. In more recent times pathogens such as M. tuberculosis and other bacteria have been characterized using their nucleic acids (DNA and RNA) and proteins to study factors such as lineage or strain diversity, drug resistance as well as other physico-chemical properties important for pathogen survival in the host. These methods have provided a particularly suitable tool to investigate the epidemiology of bacterial diseases, including sub-Saharan Africa 26 . Several molecular tools, including polymerase chain reaction (PCR), restriction fragment length polymorphisms (RFLP) analysis, spacer oligonucleotide typing (spoligotyping), mycobacterium interspaced repetitive units variable number tandem repeats (MIRU-VNTR) and sequencing have gradually been adopted in many countries to study the epidemiology of tuberculosis. The advantage of these molecular methods in tuberculosis studies rests on their ability to not only discriminate between isolates and identify strains but also to distinguish antibiotic sensitive from resistant strains in M. tuberculosis complex infection. These methods have been used successfully to identify chains of transmission and to trace the origins of infection (or pathogen) 27,28 .
Although various genotyping tools have been instrumental in advancing our understanding of TB dynamics, it is becoming clear that whole genome sequencing is superior to conventional genotyping for M. tuberculosis studies, particularly to provide a better understanding of M. tuberculosis genome evolution over time in its natural host context. Thus evaluation of the trend of advancement in using these new and modern technologies over time is necessary to get a clear picture on stage of our diagnostic capacity. Diagnostic approaches range from conventional genotyping to sequencing, allowing for diverse species and strains of M. tuberculosis complex (MTBC) to be discerned. This approach has been useful in tracing the evolution of strains over time 29 , the geospatial distribution of strains 30 as well as the strains of M. tuberculosis that seem to be independently acquiring drug resistance 31,32 . Proper diagnosis is important, since complimentary studies on nontuberculous mycobacteria (NTM) have shown the potential of diverse NTM species for causing disease in both animals and humans 33 . Despite the availability of these genotyping methods, very few studies tracing TB strains in human and animals in sub-Saharan Africa have been conducted.

Literature Search Strategy
A literature search was conducted using PubMed with search teams "Whole Genome Sequencing AND sub-Saharan Africa". To ensure that the retrieved articles are relevant, PRISM guidelines (http://www.prisma-statement.org/) were used, and the search included sufficient terms to extract maximum number of informative publications. The search, initially included only studies done in Tanzania, then some sub-Saharan Africa. The initial search retrieved 40 published article (Table 1), three [34][35][36] done in Tanzania and the rest (37 articles) done in other areas. Most of these articles were from South Africa and Ethiopia with little coming from Djibouti (1), Mali (1), Nigeria (1), Equatorial Guinea (1), Malawi (1) and Uganda (1). The value of whole genome sequencing has been important in tracking evolutional changes in mycobacterial genome that gradually resulted into development of TB drug resistance.

Search results and implication
A total of 716 publications from search were eligible for inclusion in this review 442 (61.7%) of which, originated from Tanzanian studies and 274 (30.3%) publications originated from other countries ( Table 2). The geographic spread of these latter publications were variably and regionally distributed along South Africa, the Horn of Africa, West Africa, East and Central African countries. In the general research results, both conventional and molecular studies were unevenly distributed. The uneven distribution of these studies is a reflection of country and regional variability in adoption of modern technology in disease screening and diagnosis.
In Tanzania there were just 16 studies from 2001 to 2018 which have characterized TB in both humans and animals to genetic level ( Table 1). All of those studies used genotyping methods, of which 6 focused on characterization of the infective bacteria (M. bovis) in cattle [37][38][39][40][41] and rodents 42 , only nine 34,35,[43][44][45][46][47][48][49][50] dealt with infection in humans while the rest were mainly studies that involved conventional methods both in humans and animals. The 16 genetic studies include those carried out at human-animal interface. One study (Katale et al) 33 explored species diversity of non-tuberculous mycobacteria isolated from humans, livestock and wildlife at their interface.
In Tanzania between 2014 and 2018, there was a gradual increase of genotyping studies resulting in 66 publications (retrieved from PubMed). Similarly, 15 publications have been published that use the rapid GeneXpert test (Cepheid, USA), a molecular test for TB detecting the presence of TB bacteria, as well as testing for rifampicin resistance ( Table  2). Despite the approximately 4-fold increase in genetic based studies over the previous 4 years compared to those prior to 2014, only four (4/66, 6.1%) of the studies 33,39,48,49 , have focused on the animal-human interface in tuberculosis. The rest focused on human TB alone or animal TB alone or were simply diagnostic, leaving a large gap in our understanding of the dynamics of tuberculosis at the interface.
Genotyping and phylogeographical analyses done in Tanzania have highlighted the predominance of the CAS, T, EAI, and LAM MTBC lineages in Tanzania with the most frequent Spoligotype International Types (SITs) being SIT21/CAS1-Kili, SIT59/LAM11-ZWE and SIT126/EAI5. Other circulating lineages include Haalem, Beijing, LAM, S, T, and X and Manu. Studies done in the past four years while mapping the MTBC genetic diversity in Tanzania, neighbouring East African and other several African countries 49 , have described a new lineage (designated EAI3-TZA) that seems to be specific to Tanzania 48,49 . The studies collectively highlighted the absence of clear evidence for recent cross-species transmission of either M. tuberculosis or M. bovis between humans, livestock and wild animals 33,39,48,49 as well as identifying a novel M. bovis strain which have not been previously reported in the Serengeti ecosystem 48 . This information is important when evaluating prevailing strains in an area where humans and animals are in intense contact with potential for cross-transmission.
Those studies focusing exclusively on animal to animal transmission have also highlighted that, despite wild animals being at risk of acquiring M. bovis infection from livestock due to occasional interactions in sharing of pasture and water sources, no M. bovis was isolated from hunted wild animals. It should be noted that such studies may have been biased to low endemic areas and have not sampled large numbers of animals, nor targeted animals that were visibly unhealthy and therefore the presence of M. bovis cannot be ruled out and this does not rule out a call for integrated efforts by all stakeholders for effective control of spread of tuberculosis 41 .
In Uganda both M. tuberculosis and M. bovis have been studied in humans and animals. Characterization of modern M. tuberculosis strains has indicated spoligotypes and drug susceptibility patterns of isolates from tuberculosis patients to consist of strains mainly belonging to the Uganda genotype and with a low anti-TB drug resistance rate 51 . A wide diversity of strains had previously been reported 52 with the majority of the TB cases supposedly due to reactivation rather than re-infection. Studies have also shown that despite TB epidemic genotypes being predominantly localized 53,54 , strain types were not associated with drug resistance nor HIV sero-status 53 . Despite having little impact on the clinical course for individual patients, infection with multiple MTBC strains has been shown to occur in patients with a first episode of pulmonary disease, in settings with high TB incidence 55 .
Characterization of M. bovis isolates in Uganda have indicated some spoligotype patterns that had not been previously reported 56 , indicating the lack of comprehensive studies in this country. It was reported that infected carcasses, even with multiple sites of infection, are not routinely condemned as unfit for human consumption, and even if this should be done, people may illegally obtain meat from the carcasses for consumption, particularly in rural areas 56 . This situation occurs in most of the traditional African communities and societies where cattle are kept for prestige rather than as a source of wealth and food. This cultural difference between so-called developed and developing countries necessitates the need for specific and relevant approaches in control strategies 57 .  (Table 1)

Disease dynamics and risk of resurgence
Major factors potentiating TB resurgence include poverty, consequently poor health facilities, which lead to failures in the treatment systems as well as immigration 80 . In addition, a more recent risk is co-morbidity with HIV infection, which increases risk of progression to active disease and latent TB reactivation [81][82][83] . There is also an alarming rise in drug resistant TB cases. In most localities, it is not known whether the drug resistant cases result from treatment failures, i.e. acquisition, or are from transmission. Genotyping and sequencing studies are ideal tools to establish the major driver of these epidemics and are crucial for design of appropriate intervention and treatment [84][85][86][87][88][89] . However, this is only possible where there is adequate funding and health care facilities. Therefore, studies using advanced tools have been sporadic and limited in scope while their implementation for routine screening has been impossible in most countries. For example only three published studies 35,36,47 investigated M. tuberculosis at the whole genome sequencing level in Tanzania, while another one (Katale et al, Bacterial Zoonozes Community of Practice under SACIDS, Katale, personal communication) has been done but is not yet published. However, evidence suggests regional TB epidemics in Africa, characterized by genetically distinct lineages of M. tuberculosis. M. tuberculosis in these regions may have been introduced from either Europe or Asia and has spread through pastoralism, mining and war 30 .
Emergence of migrating, highly virulent strains of M. tuberculosis, sometimes in association with multidrug resistance, is a warning sign of a serious threat to TB control 90 . This necessitates the identification of major driving forces for the transmission dynamics within specific populations, which in turn may have significant impact on disease control and vaccine development strategies 30,91 . Such goal cannot be achieved without use of advanced molecular methods for screening, diagnosis and comparative molecular epidemiological studies. Advancement in genotyping methods for examples, can provide sufficient information that can determine the approach for planning control strategies. In African settings where there is close interaction and intense contact between humans and animals (including wild animals), the risk of possible crosstransmission of antibiotic resistant TB strains between species is high. This may be an emerging problem, as suggested by the increasing isolation of M. tuberculosis from animals, including pigs 92 . In such circumstances, the increase in cross-transmission would consequently mean cross-transmission of resistant TB (M. bovis or M. tuberculosis) strains, identification of which, needs advanced diagnostic tools. Improved diagnostic facilities as well as the use of isoniazid prophylactic therapy in endemic areas 5 may be the best global TB control strategy particularly in this era where the focus can as well be on better management of HIV and TB co-infections. All these require early screening and diagnosis at point of care, which can be ensued via advancement in technology.

TB strain diversity in animals
A variety of molecular methods ranging from simple techniques (Polymerase Chain Reaction, PCR) to Next Generation Sequencing, have variably been adopted to characterize TB strains in diverse clonal complexes both in humans and animals. Similar to TB strain distribution in humans, studies carried in bTB, have found a degree of regional and geographic distribution. While a clonal complex of M. bovis (Af2) was isolated at high frequency from cattle in Uganda, Burundi, Tanzania, and Ethiopia and was identified as an East African strain 37 , another clonal complex (Af1) has been found to be a predominantly West African strain 93 with high frequencies of isolation from Mali, Nigeria, Cameroon, and Chad. Similar to clonal complexes that have been identified and destined in different regional foci in Africa are the European M. bovis clone complexes Eu1 21 and Eu2 94 which, within Europe, are of high frequency in certain countries and low in others, but have also been identified in Africa, most likely imported with cattle originating from Europe. Within the clonal complexes, there are specific families and strains 95 , allowing further discrimination and which will allow contact tracing, infection route identification and design of interventions. All these groupings of TB strains and their distribution countrywide and regionally, has been made possible via advancement in molecular-based mycobacterial studies.

Contribution of MOTT in TB infection
Mycobacteria other than M. tuberculosis (MOTT), commonly known as NTMs (non-tuberculous mycobacteria), are mostly ubiquitous environmental mycobacteria found in soil, water, dust as well as in food [96][97][98][99] . They do not usually cause infection or illness in healthy animals or humans, however many can cause life threatening illness under situations when the immune response becomes weakened, for example [100][101][102] . Included in this group are Mycobacterium avium complex (MAC), Mycobacterium intracellulare, Mycobacterium kansasii, Mycobacterium marinum, Mycobacterium ulcerans which all habitually cause opportunistic infections. People with critically compromised immune response such as those with HIV/AIDS or chronic lung diseases are very vulnerable to infection by these organisms and symptomatically the disease manifests similarly to TB infection (primarily in lungs, disseminating to other body tissues), however antibiotic susceptibility is often quite different from M tuberculosis 103,104 . As a result, these opportunistic infections may often be confused with TB infection. As such, these mycobacteria have also been referred to as atypical mycobacteria to differentiate them from the typical mycobacterium causing tuberculosis. Contradicting views are available on the spread of atypical mycobacteria in man. Despite atypical mycobacteria living on human skin or in the nose, there is no or little evidence of the mechanism of dissemination. Reports propose the idea of atypical mycobacterial infections as being not spread from person to person but through direct contact with the bacteria in the environment 111 including drinking water 105 . Direct transmission from animals to humans has again been questioned and considered unimportant for human infection 106 and infected patients are not necessarily isolated as a protective measure to limit human to human transmission 103 . Regardless of their mode of transmission, the impact of MOTT should receive more attention. At present there are no reliable clinical predictors for the identification of MOTT infection that have been identified for adults and in children, or animals 107 . Furthermore, identification of MOTT infection in resource-limited countries is sorely lacking 98 . A recent study in South Africa 108 has revealed the diversity of NTM species in cattle and African buffaloes ascribed to at least eight clusters, with a possible eight different NTM species. This report is particularly important as it suggests that inclusion of NTM in investigative and control strategies incorporated in various TB control programs could be important. In so doing any chances of spread of infection both in humans and animals might, to a greater extent, be minimized. The presence of NTM species in all likelihood negatively affects the host immune response to tuberculosis and thus progression to TB disease and will possibly complicate immune based diagnostics 108 .

Trends in Molecular based Mycobacterial studies
We found published work on tuberculosis in Africa ranging from 7 articles in the 1950s to 600 publications in the 2000s as retrieved from PubMed which is over 80 folds higher in the 2000s than in the 1950s (Search strategy: (Diverse TB strains + Human +Animals) AND Tanzania + Africa + Sub-Saharan Africa; (Diverse TB strains + human+ animals) OR Tanzania + Africa + Sub-Saharan Africa). The search was also complemented by Google and Google Scholar search to widen the chance of capturing more articles. Microsoft Excel sheet was used to add up the number of articles in each category and for calculating proportions where needed to. Where specific tests were needed, we included in the search term e.g. DNA sequencing, GeneXpert accordingly. A total of 716 articles (Table 2) were obtained and manually categorized by EVM into different categories namely molecular, GeneXpert or others (all other approaches than molecular-based). The contribution of published molecular work in TB to the general research work is 17.7% of the total retrieved published work from 1950-2018. Among these, 15.1% are typical genotyping emanating publications and about 2.6% from GeneXpert (Table 2). In the Table we also note that the move from conventional to genotyping molecular methods has been increasing with time with larger proportions of molecular work clustering in the years 2000 -2009. The trend however is not uniform which could be a reflection of donor dependence for funding such projects and thus periodicity. Too much dependence on donors may have a detrimental effect (boom and bust cycles) and since donors may not prioritize on developing country problems, instead being rather biased to donor priorities. Abuzeid 109 highlights problems with foreign aid flows from developed to developing countries. This paper further emphasizes the need for examination of such aids for potential detrimental effects, instead of providing solutions to poverty as envisaged. The study also uncovers that most of GeneXpertbased studies falls in the years 2010 onwards, an indication that the technique is relatively new in terms of implementation in Africa.  1  39  2018  11  3  25  20  19   2  79  2017  21  5  53  42  37   3  50  2016  13  3  34  22  28   4  71  2015  12  2  57  25  46 East Africa Science 2020 | Volume 2 | Issue 1  16   5  56  2014  9  2  45  38  18   6  46  2013  5  2  39  34  12   7  53  2012  5  1  47  42  11   8  33  2011  1  1  31  24  9   9  32  2010  4  0  28  24  8   10  26  2009  6  0  20  14  12   11  22  2008  2  0  20  16  6   12  22  2007  4  0  18  19  3   13  19  2006  3  0  16

CONCLUSION AND RECOMMENDATIONS
In Africa, the occurrence of zoonosis is common, in the main part from the close proximity of humans and animals. Exploring the disease dynamics at these interfaces could provide the key for developing successful interventions to improve health in both humans and animals. We now have molecular tools which can examine the transmission dynamics of Mycobacterial disease in these settings. In order to do this, a strategy that involves medical and veterinary professions as well as other stakeholders will be needed. The Southern African Centre for Infectious Disease Surveillance (SACIDS), as well as other consortia working under the One Health concept including CORDS (Coordinating Organizations for Regional Disease Surveillance) may be a good starting point for such an approach. From this review we found an increase in published work in retrievable papers of over 80-fold in late years compared to earlier years, which is an indicator of trend-wise development in research. Most of molecular studies, however, cluster in the mid-to late 2000s, with little developments since, suggesting a slow adoption of technology and a high cost of running molecular tests. This overview of genotyping studies and trend in molecular studies should therefore provide an insight at regional level for more dedicated efforts to invest in this area. This review lays the foundation for a more robust investigation of TB particularly at human-animal interface to reduce the potential risk of cross-species transmission. This can be achieved by using molecular screening methods to assess the disease dynamics at this interface for early detection to provoke deployment of best control strategies.

Limitations
Possibility of missing of some papers and reports that were not published is potential limitation that we think might have been the case for our review. However, the 716 published articles in the area we focused are believably, sufficient to reflect the trend in which technological advancement has been growing. We encourage researchers to publish their research work so that can be kept tracked. Our future plans are to further do a systematic review and meta-analysis that could cover the whole of Africa with an improved search strategy.