Multidrug-Resistant Salmonella enterica Serotype Typhi, Gulf of Guinea Region, Africa

We identified 3 lineages among multidrug-resistant (MDR) Salmonella enterica serotype Typhi isolates in the Gulf of Guinea region in Africa during the 2000s. However, the MDR H58 haplotype, which predominates in southern Asia and Kenya, was not identified. MDR quinolone-susceptible isolates contained a 190-kb incHI1 pST2 plasmid or a 50-kb incN pST3 plasmid.

T yphoid fever, which is caused by Salmonella enterica serotype Typhi, is endemic to the developing world; there were an estimated 26.7 million cases in 2010 (1). The incidence of typhoid fever in sub-Saharan Africa was an estimated 725 cases/100,000 persons in 2010, despite a lack of incidence studies conducted in West and central Africa (1). Antimicrobial susceptibility data are also scarce for this part of Africa. This issue is problematic because treatment with appropriate antimicrobial drugs is essential for recovery in the context of the global emergence of multidrug resistance.
In the Indian subcontinent and Southeast Asia, the multidrug-resistant (MDR) Salmonella Typhi H58 clone, which was named after its haplotype (a combination of defined chromosomal single-nucleotide polymorphisms [SNPs]) (2,3), has spread rapidly and become endemic and predominant. During the 1990s, this clone acquired a large conjugative incHI1 pST6 plasmid encoding resistance to ampicillin, chloramphenicol, and co-trimoxazole (4,5); also in the 1990s, this MDR clone became resistant to quinolones and showed decreased susceptibility to ciprofloxacin because of point mutations in the chromosomal gyrA gene (2). The H58 clone has also spread to eastern Africa, where it has been the most prevalent haplotype (87%) in Kenya since the early 2000s (6).
During 1997-2011, high incidence of MDR Salmonella Typhi was reported in some countries near the Gulf of Guinea in Africa, including Nigeria (7), Ghana (8,9), Togo (10), and the Democratic Republic of the Congo (11). During 1999-2003, a British surveillance system reported a prevalence of 19% (49/421) for MDR Salmonella Typhi isolates among imported cases of typhoid fever acquired in Africa, particularly in Ghana (12). However, nothing is known about the genotypes of these isolates, including whether they belong to the spreading MDR H58 clone.
We report data for the occurrence, genotypes, and characterization of the resistance mechanisms of MDR Salmonella Typhi isolates. These isolates were obtained from the French National Reference Center for Salmonella (FNRC-Salm), Institut Pasteur (Paris, France), and Centre Pasteur du Cameroun (Yaoundé, Cameroon).

The Study
Almost all Salmonella Typhi strains isolated in France are referred to the FNRC-Salm. Most isolates were obtained from travelers or immigrants, most of whom were infected in Africa and Asia. In Cameroon, the Centre Pasteur du Cameroun collects Salmonella Typhi isolates from several hospitals in.
Antimicrobial susceptibility testing was performed according to the guidelines of the antibiogram committee of the French Society for Microbiology (http://www.sfm.asso. fr/nouv/general.php?pa = 2). Isolates were considered to be MDR if they were resistant to ≥2 of the following antimicrobial drugs: amoxicillin, co-trimoxazole (trimethoprim/ sulfamethoxazole), chloramphenicol, or tetracyclines.
During 1996-2013, a total of 1,746 Salmonella Typhi isolates were collected through the French national surveillance system and subjected to antimicrobial susceptibility testing; 408 were acquired in sub-Saharan Africa (n = 237) and northern Africa (n = 171), and 55 (13.5%) of those acquired in Africa were MDR (Table). All but 1 of the MDR isolates were acquired in sub-Saharan Africa (Table). The proportion of MDR isolates increased from 0% during 1996-1999 to 22.3% (n = 23) during 2010-2013. Only 4 isolates from Africa were resistant to nalidixic acid, including 1 isolate resistant to ciprofloxacin. Because these isolates acquired after 2010 were not MDR isolates, they were not studied further.
In Yaoundé, the proportion of MDR isolates was high Mechanisms of antimicrobial drug resistance were determined as described (14). Genetic diversity and phylogenetic relationships were studied by using standardized XbaI-pulsed-field gel electrophoresis (PFGE) (http://www. cdc.gov/pulsenet/pathogens/index.html), haplotyping (5), and clustered regularly interspaced short palindromic repeats (CRISPR) typing (15). Haplotyping was based on identification of SNPs at 1,487 defined chromosomal loci, and CRISPR typing was based on detection of 32-bp sequences (spacers) within 1 or both CRISPR regions.
With the exception of the isolate from Morocco (ISP-03-07467) (2), none of the isolates had the H58 haplotype or contained the associated incHI1 pST6 MDR plasmid. We found 3 other lineages with different geographic distributions and MDR plasmids (Figures 1, 2).
Lineage A consisted mostly of haplotype H56 isolates and more rarely H42 (which differs from H56 by 2 SNPs) and was found only in the western part of the Gulf of Guinea region. Lineage B consisted of haplotype H55 isolates and was found in the eastern and southern parts of the Gulf of Guinea region. Lineage C consisted of haplotype H77 isolates and was found only in Cameroon. All 3 lineages had distinctive CRISPR1 spacer contents. XbaI-PFGE, which used a similarity value of ≥90% as a cutoff, correctly grouped (i.e., concordant with haplotyping and CRISPR results) all but 2 of the MDR isolates from Africa (Figure 1; online Technical Appendix).
The 3 lineages contained a large (≈190 kb) conjugative MDR incHI1 pST2 plasmid that differed among lineages. Resistance to trimethoprim was encoded by different class 1 integron gene cassettes: dfrA15, dfrA7, and dfrA1 for in-cHI1 plasmids of lineages A, B, and C, respectively. All in-cHI1 plasmids from lineage A encoded resistance to chloramphenicol, and none of those from lineage C encoded such resistance. A second smaller (50-kb) MDR plasmid belonging to the incN incompatibility group (pST3 by plasmid multilocus sequence typing), was present mostly in lineage C isolates, but was also found in 1 lineage A isolate (02-1739) (online Technical Appendix).

Conclusions
Analysis of older isolates and previously published data (2) showed that susceptible Salmonella Typhi strains of haplotypes H42, H56, and H77 had already been identified in Senegal in 1962, Tunisia in 1969, and Cameroon in 1958, respectively. This finding suggests that the MDR isolates from lineages A and C are derived from local Salmonella Typhi populations in Africa, rather than being recently imported from other regions to which this bacterium is endemic. Haplotype H55 was previously restricted largely to the Indian subcontinent and eastern Africa (2) (1) Cameroon (7), detected in association with an incHI1 pST2 plasmid in India during the mid-1970s (5). Therefore, lineage B might have been was imported into central Africa from eastern Africa/southern Asia. A previous study also reported isolation of an MDR clone in the Democratic Republic of the Congo in 2004 that was resistant to quinolones, showed decreased susceptibility to ciprofloxacin, and belonged to the Asian H58 lineage (2). Because only a limited number of isolates from central Africa were tested in our study, studies of a larger collection of isolates might provide more information about bacterial genotypes/MDR plasmids circulating in central Africa.
Despite intrinsic limitations of a laboratory surveillance system for typhoid fever that is used mostly for travelers and immigrants and has the bias of preferential links caused by colonial history and choices of tourist destinations, we documented emergence of 3 MDR Salmonella Typhi lineages in the Gulf of Guinea area. Two lineages found in Guinea and Cameroon were local lineages that acquired MDR conjugative plasmids, either a large incHI1 pST2 plasmid or a smaller incN pST3 plasmid. The H58 lineage, which is currently predominant in Asia and eastern Africa, was not detected among MDR isolates from West and central Africa. Figure 1. Characteristics of 50 Salmonella enterica serotype Typhi isolates. The dendrogram was generated by using BioNumerics version 6.6 software (Applied Maths, Sint-Martens-Latem, Belgium) and shows results of cluster analysis on the basis of XbaIpulsed-field gel electrophoresis (PFGE) fingerprinting. Similarity analysis was performed by using the Dice coefficient, and clustering analysis was performed by using UPGMA. CRISPR1, clustered regularly interspaced short palindromic repeats 1; MDR, multidrug resistant; AMX, amoxicillin; STR, streptomycin; SXT, trimethoprim/sulfamethoxazole; CHL, chloramphenicol; TET, tetracycline; SPT, spectinomycin; SSS, sulfamethoxazole. Zaire is the former name of the Democratic Republic of the Congo. Plasmid multilocus sequence typing (pMLST) was performed as described for incN http://pubmlst.org/plasmid/) or incHI1 (4). Plasmid size was estimated by using S1 nuclease PFGE (14). For isolates from Africa, red indicates linage A, blue indicates lineage B, and green indicates lineage C.
This study was supported by the Institut Pasteur, the Réseau International des Instituts Pasteur, the Institut de Veille Sanitaire, and the French Government Investissement d'Avenir Program (Integrative Biology of Emerging Infectious Diseases, Laboratory of Excellence, grant ANR-10-LABX-62-IBEID).
At the time of this study, Ms. Baltazar was a predoctoral student at the FNRC-Salm, Paris, France. She is currently a doctoral candidate at the University of Limoges, Limoges, France. Her primary research interest is mechanisms of antimicrobial drug resistance. Emerging Infectious Diseases • www.cdc.gov/eid • Vol. 21, No. 4, April 2015