Review of molecular subtyping methodologies used to investigate outbreaks due to multidrug-resistant enteric bacterial pathogens in sub-Saharan Africa

Background In sub-Saharan Africa, molecular epidemiological investigation of outbreaks caused by antimicrobial-resistant enteric bacterial pathogens have mostly been described for Salmonella species, Vibrio cholerae, Shigella species and Escherichia coli. For these organisms, I reviewed all publications describing the use of molecular subtyping methodologies to investigate outbreaks caused by multidrug-resistant (MDR) enteric bacterial infections. Objectives To describe the use of molecular subtyping methodologies to investigate outbreaks caused by MDR enteric bacterial pathogens in sub-Saharan Africa and to describe the current status of molecular subtyping capabilities in the region. Methods A PubMed database literature search (English language only) was performed using the search strings: ‘Africa outbreak MDR’, ‘Africa outbreak multi’, ‘Africa outbreak multidrug’, ‘Africa outbreak multi drug’, ‘Africa outbreak resistance’, ‘Africa outbreak resistant’, ‘Africa outbreak drug’, ‘Africa outbreak antibiotic’, ‘Africa outbreak antimicrobial’. These search strings were used in combination with genus and species names of the organisms listed above. All results were included in the review. Results The year 1991 saw one of the first reports describing the use of molecular subtyping methodologies in sub-Saharan Africa; this included the use of plasmid profiling to characterise Salmonella Enteritidis. To date, several methodologies have been used; pulsed-field gel electrophoresis analysis and multilocus sequence typing have been the most commonly used methodologies. Investigations have particularly highlighted the emergence and spread of MDR clones; these include Salmonella Typhi H58 and Salmonella Typhimurium ST313 clones. In recent times, whole-genome sequencing (WGS) analysis approaches have increasingly been used. Conclusion Traditional molecular subtyping methodologies are still commonly used and still have their place in investigations; however, WGS approaches have increasingly been used and are slowly gaining a stronghold. African laboratories need to start adapting their molecular surveillance methodologies to include WGS, as it is foreseen that WGS analysis will eventually replace all traditional methodologies.


Introduction
For molecular epidemiological investigation of outbreaks in sub-Saharan Africa caused by antimicrobial-resistant (including multidrug-resistant [MDR]) enteric bacterial pathogens, most published data describe pathogens belonging to the genus or species of Salmonella, Vibrio cholerae, Shigella and Escherichia coli. By 'molecular epidemiological investigation', I refer to the use of molecular subtyping techniques that analyse bacterial strains at the level of their nucleic acid and so gives an indication of genetic similarity of strains. 1 Molecular subtyping is vital for accurate epidemiological investigations of bacterial infections. 2 Molecular subtyping allows one to segregate unlike strains of the same species or serotype of bacteria and identify clones or clusters of bacteria (genetically related strains). Molecular subtyping allows one to track the spread of strains or clones and determine how local strains compare to those circulating worldwide. 3 Molecular subtyping data is critical for successful epidemiological investigation of outbreaks of disease; in particular, outbreak-related cases of disease can be differentiated from sporadic cases of disease. 4 Knowledge of the molecular epidemiology of bacterial infections provides information of major circulating (infecting) clones, so that in times of antimicrobial treatment and vaccine interventions, well-informed and educated decisions can be made to combat the disease. 5 Background: In sub-Saharan Africa, molecular epidemiological investigation of outbreaks caused by antimicrobial-resistant enteric bacterial pathogens have mostly been described for Salmonella species, Vibrio cholerae, Shigella species and Escherichia coli. For these organisms, I reviewed all publications describing the use of molecular subtyping methodologies to investigate outbreaks caused by multidrug-resistant (MDR) enteric bacterial infections.
The mid-1990s saw the start of noteworthy publications describing the use of molecular subtyping techniques to investigate the molecular epidemiology of antimicrobialresistant enteric bacterial pathogens in sub-Saharan Africa. 6 These molecular subtyping techniques included plasmid profiling, ribotyping (southern blotting of restricted genomic DNA and probing with ribosomal genes), random amplified polymorphic DNA analysis (also called arbitrarily primed polymerase chain reaction [PCR]), enterobacterial repetitive intergenic consensus elements PCR, pulsed-field gel electrophoresis (PFGE) analysis, multiple-locus variablenumber tandem-repeats analysis (MLVA), multilocus sequence typing (MLST), multilocus sequence analysis and whole-genome sequencing (WGS) analysis. Molecular subtyping data from sub-Saharan Africa is summarised in Table 1 and the countries involved are highlighted in Figure 1. To date, the most popular and well documented molecular subtyping techniques have included PFGE, MLVA, MLST and WGS analysis. Traditional molecular subtyping methodologies, such as PFGE, MLVA and MLST, have their advantages and disadvantages; these have been well reviewed elsewhere on numerous occasions. 1,2,7,8,9,10,11 Some molecular subtyping techniques (such as ribotyping, random amplified polymorphic DNA analysis and enterobacterial repetitive intergenic consensus elements PCR) may only allow a local comparison of bacterial strains within a single laboratory, because the workings of the methodology are difficult to standardise between laboratories, and so the data that is produced cannot be truly and accurately compared between laboratories. In contrast, a technique such as PFGE analysis, has been standardised by PulseNet International (http://www.pulsenetinternational.org/) and is successfully used for inter-laboratory comparison of subtyping data and global comparison of bacterial strains. 4 Newer approaches to molecular subtyping involve WGS analysis. 12,13,14 Analysis of WGS data for molecular epidemiological purposes can include multiple approaches; however, the more popular methods are whole-genome MLST and single nucleotide polymorphisms (SNP) analysis. 15 WGS data is electronically portable and can easily be shared between laboratories, allowing for a global inter-laboratory comparison of bacterial strains. Current challenges surrounding WGS include how to standardise the quality of WGS data generated and how to standardise the analysis of WGS data, in order to allow for a successful inter-laboratory comparison of analysed WGS data. PulseNet International has taken on this challenge and published their vision for implementation of WGS for global foodborne disease surveillance and global analysis of enteric pathogens. 15 The aim of the present manuscript is to review all publications that have described the use of molecular subtyping methodologies to investigate outbreaks due to multidrugresistant enteric bacterial pathogens in sub-Saharan Africa. The review focuses on enteric pathogens belonging to the genus and species of Salmonella, Vibrio cholerae, Shigella and Escherichia coli, because most molecular epidemiological investigations have been described for these microorganisms. The aim of this review is to inform the readers about the current status of molecular subtyping capabilities and activities in sub-Saharan Africa and finally suggest that the way forward for molecular subtyping in African laboratories is the implementation of WGS analysis as soon as possible.
A PubMed database literature search (English language only) was performed using the search strings: 'Africa outbreak MDR', 'Africa outbreak multi', 'Africa outbreak multidrug', 'Africa outbreak multi drug', 'Africa outbreak resistance', 'Africa outbreak resistant', 'Africa outbreak drug', 'Africa outbreak antibiotic', 'Africa outbreak antimicrobial'. These search strings were used in combination with genus and species names of the organisms listed above. All results were included in the review.

Nontyphoidal Salmonella
For nontyphoidal Salmonella (NTS), molecular epidemiological analysis of Salmonella Typhimurium and Salmonella Enteritidis are most commonly described in sub-Saharan Africa. One of the first reports using molecular subtyping techniques to investigate enteric bacterial pathogens isolated in sub-Saharan Africa was a 1991 study that used plasmid profiling (following restriction endonuclease digestion) to distinguish African isolates of Salmonella Enteritidis from United States (US) isolates. 16 Molecular subtyping of NTS then progressed to PFGE analysis of MDR Salmonella Typhimurium isolated in Kenya, where 64 isolates were grouped into eight PFGE clusters and were described as multiclonal. 17 Salmonella Typhimurium accounts for a significant proportion of reported cases of NTS-associated invasive disease in sub-Saharan Africa. NTS-associated invasive disease has developed to become a leading public health challenge in sub-Saharan Africa. 18,19 A predominant type of invasive Salmonella Typhimurium in sub-Saharan Africa is a MDR strain that has been designated ST313 (based on its MLST profile). Clonal spread of this MDR ST313 strain was first reported from Kenya and Malawi in 2009, with the strain showing resistance to ampicillin, chloramphenicol, sulphamethoxazole, trimethoprim, streptomycin and kanamycin. 20 In Kenya, a more recent analysis of Salmonella Typhimurium outbreak isolates, using WGS-SNP analysis, identified a single clade of MDR ST313 strains showing resistance to ampicillin, ceftriaxone, chloramphenicol, sulphamethoxazole and trimethoprim; extended-spectrum beta lactamase (ESBL) genes bla CTX-M-15 , bla TEM-1 and bla OXA-1 were also harboured by the isolates. 21 The MDR ST313 has further been reported from Malawi, Nigeria, Democratic Republic of the Congo and South Africa. 22,23,24 Leekitcharoenphon and coworkers 22 used MLST, PFGE and WGS-SNP analysis to describe a close relationship between MDR ST313 strains from Nigeria and Democratic Republic of the Congo; strains were isolated from invasive (blood) and non-invasive (stool) specimens; all isolates harboured bla TEM1b , catA1, strA/B, sul1, and dfrA1 genes coding for resistance to various classes of antimicrobial agents. Keddy and coworkers 23 used MLST to investigate Salmonella Typhimurium isolates from South Africa and reported that NTS meningitis in South Africa was highly associated with the MDR ST313 strain.   were also shown to be highly related to isolates associated with a similar waterborne outbreak in the same area in 1993. Besides PFGE analysis, MLVA is often described as a useful molecular subtyping method for discrimination of strains belonging to the Salmonella species. Tau and coworkers 38 described the development and evaluation of a MLVA assay for molecular subtyping of Salmonella Typhi in sub-Saharan Africa. They evaluated the MLVA assay on a panel of African isolates and showed that it had higher discriminatory power as compared to PFGE; the MLVA assay was able to differentiate outbreak isolates from sporadic isolates.
Whole-genome sequencing -single nucleotide polymorphisms analysis has been increasingly used to investigate the molecular epidemiology of Salmonella Typhi in sub-Saharan Africa; a common theme has been the description of the MDR Salmonella Typhi H58 strain. 5, 39 Wong and coworkers 5 used WGS-SNP analysis to investigate 1832 Salmonella Typhi isolates from 63 global countries, to identify one major MDR lineage (named H58) that has arisen and spread across Asia and Africa over the past 30 years. Multiple H58 transfers have occurred, moving from Asia into Africa. An MDR H58 epidemic has been described in sub-Saharan Africa where H58 lineages are displacing antimicrobial-susceptible strains. MDR H58 lineages are associated with resistance to ampicillin, trimethoprim, sulphonamides, chloramphenicol, streptomycin and tetracycline, including reduced susceptibility to fluoroquinolones. 5,39 This MDR H58 lineage has now been described in the eastern and southern regions of sub-Saharan Africa including Kenya, Tanzania, Malawi and South Africa. 5,38,39,40 Hendriksen and coworkers 41 also used WGS-SNP analysis to investigate an outbreak of Salmonella Typhi in Zambia; most isolates showed resistance to ampicillin, chloramphenicol, streptomycin, sulphamethoxazole and trimethoprim, while some isolates also had reduced susceptibility to fluoroquinolones. Isolates belonged to a new variant of the H58 haplotype, namely haplotype H58B. Interestingly, investigation of MDR Salmonella Typhi from Nigeria -western Africa, using WGS-SNP analysis, has not detected the H58 lineage; instead the majority of isolates belonged to a different lineage -H56 lineage, which carries genes coding for resistance to ampicillin, tetracycline, chloramphenicol and sulphamethoxazole (bla TEM1 , catA1, tetB, dfrA1, sul1). This H56 lineage is relatively common across Africa, predominantly in western and central regions of sub-Saharan Africa. 42

Vibrio cholerae
Ribotyping and enterobacterial repetitive intergenic consensus elements PCR were among the first molecular subtyping techniques used to characterise V. cholerae O1 in Africa. 43,44 In 1996, Dalsgaard and coworkers 43 documented the first molecular epidemiological investigations of MDR V. cholerae O1 in Africa. They used ribotyping to investigate outbreak isolates in Guinea-Bissau; strains isolated in 1994-1995 showed a ribotype pattern which was different fro strains isolated in 1987, suggesting that the 1994-1995 outbreak was as a result of the introduction of a novel strain into the country. Later, ribotyping analysis of V. cholerae O1 outbreak isolates from Senegal showed the same ribotype pattern as the 1994-1995 Guinea-Bissau outbreak strain, suggesting that Senegal acquired their outbreak strain from Guinea-Bissau. 45 In 2001, Dalsgaard and coworkers 46 reported on the clonal relationship (using ribotyping) of cholera isolates from Mozambican migrant workers associated with a 1998 outbreak that occurred in South African provinces bordering Mozambique. MDR V. cholerae O1 isolates were associated with the outbreak among the migrant workers and the isolates showed resistance to furazolidone, streptomycin, sulphamethoxazole, trimethoprim and tetracycline; isolates also showed the presence of class 1 integrons and the SXT element. Another analysis of Mozambican MDR isolates associated with the 1998 Mozambique/South Africa cholera outbreak confirmed the presence of the SXT element among isolates, with enterobacterial repetitive intergenic consensus elements PCR dividing isolates into two different molecular subtypes. 47 Mobile genetic elements are mostly responsible for the molecular basis of MDR V. cholerae O1 in sub-Saharan Africa. Mobile genetic elements include transposable elements (SXT elements), integrons and conjugative plasmids. The SXT element is a self-transmissible element that integrates into the chromosome and carries genes encoding resistance to several antimicrobial agents, including chloramphenicol, sulfamethoxazole, streptomycin, trimethoprim and furazolidone. The SXT element has been reported from numerous countries in sub-Saharan Africa. 46,48,49,50,51 In 2006, Scrascia and coworkers 52 investigated V. cholerae O1 isolates from multiple outbreaks that occurred in Kenya during 1998 and 1999. Most isolates showed an identical ribotype profile and similar random amplified polymorphic DNA analysis profile suggesting a clonal origin for the outbreaks; isolates were resistant to chloramphenicol, spectinomycin, streptomycin, sulphamethoxazole and trimethoprim. Over the period 2008-2009, published data from Ghana, Cameroon, Ethiopia and Somalia reported similar trends (as described above) for MDR V. cholerae O1 isolates following analysis using ribotyping and random amplified polymorphic DNA analysis -that of each country reporting a clonal origin for outbreak isolates within their respective country. 50,53,54,55 From 2007 onwards, trends in molecular subtyping of V. cholerae, started shifting to an increased use of PFGE analysis. In 2007, Keddy and coworkers 56  isolates showed similar PFGE patterns: isolates showed the presence of the SXT element encoding multidrug resistance; some isolates also showed ESBL activity due to the presence of the bla TEM gene encoding TEM-63 β-lactamase. 58 In October 2010, the Haiti cholera outbreak started, with involvement of MDR V. cholerae O1. 59,60 PFGE analysis of Haiti outbreak strains isolated over the period October 2010 to February 2011 showed that a single PFGE profile predominated. This predominant profile was also shown in MDR strains from Cameroon, South Africa, India, Nepal, Pakistan, Afghanistan and Oman 61 ; this global comparison of V. cholerae O1 PFGE patterns was facilitated via PulseNet International laboratories. Further investigation with WGS-SNP analysis showed that the Haiti outbreak strain was a hybrid-type of El Tor strain encoding a classical-type of cholera toxin and that the outbreak strain was most closely related to strains originating from India and Cameroon. 62 The Haiti cholera outbreak identified the need for higher resolution molecular subtyping methodologies for comparison of V. cholerae O1 isolates; this sparked the increased use of nucleotide sequencing approaches including WGS, to generate nucleotide sequence data for comparison of isolates. In 2011, Thompson and coworkers 63 characterised V. cholerae O1 outbreak isolates from Ghana using multilocus sequence analysis of housekeeping genes and identified two major clusters in Ghana. In 2011, Mutreja and coworkers 64 provided evidence for multiple waves of global transmission (from the Bay of Bengal) within the seventh cholera pandemic, which included multiple waves of transmission into sub-Saharan Africa; these data were made possible by analysis of WGS data (WGS-SNP analysis approach) from a global collection of V. cholerae O1 isolates, including African outbreak isolates. In 2013, Kiiru and coworkers, 65 reported a WGS-SNP analysis on Kenyan V. cholerae isolates which included environmental isolates (serogroups O1 and non-O1) and V. cholerae O1 clinical isolates associated with outbreaks; they found that a clade comprising clinical isolates and some environmental isolates mapped back onto wave three of the monophyletic seventh cholera pandemic phylogeny, while some other environmental isolates were phylogenetically very different from the monophyletic seventh pandemic lineage of V. cholerae O1. The continued presence of MDR associated STX elements were documented on the genomes of MDR clinical and MDR environmental isolates. Recent isolates of V. cholerae O1 from Tanzania have also been found to map back onto wave three of the monophyletic seventh cholera pandemic phylogeny, as determined by WGS-SNP analysis. 66 Interestingly, WGS-SNP analysis of MDR V. cholerae O1 outbreak isolates recently sourced from Cameroon has found that these Cameroonian isolates do not map onto the phylogeny of isolates from other African countries such as Kenya, Tanzania, Zambia, Zimbabwe and Mozambique; instead, Cameroonian isolates are of their own distinct clonal cluster (clade), forming part of an isolated reservoir of MDR V. cholerae O1 in the Lake Chad basin. 67 Although WGS approaches for comparison of V. cholerae isolates have gained a stronghold in sub-Saharan Africa, the traditional molecular subtyping methodologies of PFGE, MLVA and MLST are still commonly used to investigate the molecular epidemiology of V. cholerae in sub-Saharan Africa, as these technologies still remain more accessible and more affordable for most laboratories 66,68,69,70,71,72,73 . For these recent reports, the central themes remain the same as previous studies. The major conclusions for these studies are summarised as follows. The most commonly reported cholera-causing organism is characterised as V. cholerae O1 biotype El Tor, the so-called 'atypical' El Tor variant housing genetic determinants coding for a 'classical-type' of cholera toxin. The majority of V. cholerae O1 isolates show the presence of SXT elements encoding a large part of the MDR phenotype. The majority of isolates show MDR. A common profile includes resistance to ampicillin, chloramphenicol, sulphamethoxazole, trimethoprim, spectinomycin and streptomycin, including reduced susceptibility to ciprofloxacin. Molecular subtyping of V. cholerae O1 isolates shows examples of geographical clustering, as already described above in the mention of Cameroonian isolates characteristic of the Lake Chad basin region; Moore and coworkers 72 found that isolates from western Africa (Togo and Guinea) form a closely related group which is separated from a distinctive cluster associated with the African Great Lakes region (Democratic Republic of the Congo and Zambia). Smith and coworkers 71 found that a Mozambican cluster of isolates was distinctly different (including a different antimicrobial resistance profile) as compared to isolates from western Africa (Togo, Guinea and Côte d'Ivoire) and central Africa (Democratic Republic of the Congo).

Shigella species
Shigella dysenteriae was the first Shigella species in sub-Saharan Africa that was characterised by molecular subtyping techniques; these details were first documented in the years 1996 and 1997. 6,74 In 1996, Kariuki and coworkers 6 documented an outbreak of dysentery due to MDR S. dysenteriae serotype 1 in Kenya: all 22 outbreak isolates were resistant to ampicillin, chloramphenicol, sulphamethoxazole, trimethoprim, streptomycin and tetracycline; PFGE and plasmid profiling showed clonality among the isolates. In 1997, Pillay and coworkers 74 documented a nosocomial outbreak of dysentery caused by MDR S. dysenteriae serotype 1 which transpired in a psychiatric institution in South Africa. Isolates were resistant to ampicillin, chloramphenicol, tetracycline and cotrimoxazole; ribotyping showed clonality among the isolates. Later in 1996, a report documented how PFGE analysis was used to investigate MDR S. dysenteriae serotype 1 isolates associated with two separate dysentery outbreaks in the Central African Republic; PFGE analysis provided evidence to show that each outbreak was associated with a different clone. 75 In 2009, Smith and coworkers 76 reported a cluster of 29 watery diarrhoea cases associated with MDR Shigella boydii serotype 2 isolates in South Africa. Isolates were resistant to ampicillin, trimethoprim, sulphamethoxazole and streptomycin; PFGE analysis showed clonality among the isolates.
More recent molecular epidemiological investigations of Shigella species in sub-Saharan Africa has involved WGS-SNP analysis of isolates. Connor and coworkers 77 used WGS-SNP analysis to investigate 351 global (including Africa) isolates of Shigella flexneri, to show that S. flexneri is composed of seven distinct phylogenetic groups. Each phylogenetic group contains geographically restricted sub-lineages that seem to have constantly inhabited areas for several decades; there is limited evidence of intercontinental transmission. Antimicrobial resistance determinants have been acquired and maintained locally on several episodes -to summarise, S. flexneri from sub-Saharan Africa cluster together and form a distinct lineage as compared to lineages from other parts of the world. Njamkepo 79 reported on the etiology of acute diarrhoea in adults in Nigeria: shiga-toxin producing E. coli and enteroaggregative E. coli were significantly associated with the diarrhoea. PFGE analysis of shiga-toxin producing E. coli isolates provided evidence to suggest that an outbreak had occurred. In 2012, Tau and coworkers 80 reported on the characterisation of E. coli O104 isolates from South Africa for the years 2004-2011. This study was undertaken in response to the 2011 outbreak of bloody diarrhoea and haemolytic uraemic syndrome which occurred in Germany, caused by an enteroaggregative shiga-toxin producing E. coli O104:H4 81 . The South African investigation found two antimicrobial-susceptible enteropathogenic E. coli O104:non-H4 isolates and five enteroaggregative E. coli O104:H4 isolates (showing resistance to ampicillin, sulphamethoxazole and trimethoprim). Pulsed-field gel electrophoresis analysis showed that these South African isolates were unrelated to the Germany outbreak strain.

Conclusion
Whole-genome sequencing is undoubtedly the way forward for investigation of enteric pathogens in diagnostic and public health laboratories. Numerous publications have analysed data sets to compare results obtained using old-fashioned molecular subtyping methodologies versus results obtained using newer WGS analysis methodologies. 14,82,83,84 Conclusions from these studies are unanimous and convincing, in that WGS trumps all other methodologies. Old-fashioned methodologies offer limited genetic resolution and limited discriminatory ability and there can sometimes be a lack of accuracy; all these hinder epidemiologic investigations. For example, data from oldfashioned methodologies can incorrectly conclude that unrelated isolates are indistinguishable, resulting in false alerts and unnecessary investigations. Whole-genome sequencing provides a supreme unparalleled discriminatory ability, can reflect phylogeny and can provide an evolutionary context. In addition, WGS can provide valuable information related to genetic determinants conferring antimicrobial resistance. Numerous publications have reported how analysis of WGS data can be used to infer antimicrobial resistance in bacterial pathogens; 85,86,87,88,89 these studies have reported high concordance between WGS resistance predictions and conventional phenotypic antimicrobial susceptibility testing methodologies. Even so, conventional phenotypic susceptibility testing methodologies still have their place in the laboratory and must continue to be used for the foreseeable future. For now, WGS antimicrobial resistance prediction is best used as an adjunct to conventional phenotypic susceptibility testing.
In summary, public health laboratories in Africa need to urgently reconsider and adapt their molecular surveillance methodologies to include WGS. It is foreseen that WGS analysis will eventually replace all traditional methods currently used in the microbiology laboratory, both traditional phenotypic methods and traditional genotypic methods including molecular subtyping methods such as PFGE, MLVA and MLST. 90,91,92,93 Within the next year, the PulseNet USA programme is planning to have WGS (and analysis of WGS data) fully implemented as its primary method for molecular epidemiological investigation of enteric pathogens. 94 The next challenge is how to effectively implement WGS analysis globally in public health laboratories, how to standardise the quality of WGS data generated and how to standardise the analysis of WGS data, all in order to allow for a successful inter-laboratory comparison of analysed WGS data. PulseNet International has taken this challenge and published their vision for implementation of WGS for global foodborne disease surveillance and global analysis of enteric pathogens. 15 PulseNet Africa (a regional network of PulseNet International) believes and supports this vision for implementation of WGS for global analysis of enteric pathogens.