OAU: Obafemi Awolowo University Teaching Hospital

This work was supported by Official Development Assistance (ODA) funding from the National Institute of Health Research [16/136/111: NIHR Global Health Research Unit on Genomic Surveillance of Antimicrobial Resistance].

This research was commissioned by the National Institute of Health Research using Official Development Assistance (ODA) funding. INO is an African Research Leader supported by the UK Medical Research Council (MRC) and the UK Department for International Development (DFID) under the MRC/DFID Concordat agreement that is also part of the EDCTP2 program supported by the European Union. The funders had no role in the content, crafting or submission of this paper. The views expressed in this publication are those of the authors and not necessarily those of the funders or their affiliates.

The authors: No reported conflicts of interest. All authors have submitted the ICMJE Form for

Klebsiella is a ubiquitous gram-negative genus that can cause a variety of opportunistic infections [ 1, 2 ]. Klebsiella pneumoniae, the species most commonly associated with human disease, is frequently implicated in life-threatening bacteremia and sepsis arising from translocations from non-sterile niches or medical devices in hospitals [ 3 ]. Although the precise burden of infectious diseases resulting from Klebsiella spp. is unknown in Africa, there is an upward trend in reports of K. pneumoniaeassociated bloodstream infections, including from Nigeria, Côte d9Ivoire, Malawi, Gambia, and South Africa [4-8]. There have also been isolated case reports of non-pneumoniae infections in Nigeria [9]. Altogether, these infections have high mortality rates and incur high costs [ 10, 11 ]. Rigorous surveillance of Klebsiella clones circulating within Nigerian hospitals and communities is needed to inform treatment guidelines and prioritize AMR interventions. Whole genome sequencing (WGS) offers the attractive prospect of meeting this need within local resource constraints. WGS tools also provide taxonomic resolution below serotype levels, a feat impossible in resource-limited settings by classical typing methods [ 13 ]. Mapping the diversity of Klebsiella populations at high resolution, with spatiotemporal dynamics, also makes it possible to elucidate the origin and dissemination of AMR [ 12, 13 ].

In contrast to similarly constrained settings elsewhere in Africa [14316], few Nigeria reports include in-depth analysis of more than a handful of Klebsiella isolates. For example, as of the time of writing, Pathogenwatch, which sources its public genomes from ENA, includes only 109 Klebsiella genomes from Nigeria [17]. Nigeria recently initiated national AMR surveillance following commissioning of a 2017 National Action Plan, and the Nigeria Centre for Disease Control (NCDC) enrolled the country in the WHO9s Global Antimicrobial Resistance Surveillance System (GLASS) [18]. NCDC is building a network that currently consists of 9 AMR sentinel hospital laboratories and 2 national reference laboratories [19]. The very low number of Klebsiella genomes from Nigeria compelled the

Nigeria node of the Global Health Research Unit (GHRU) on Genomic Surveillance of AMR, which provides WGS services to the reference laboratories, to collect and sequence retrospective isolates. In addition to increasing the number of relevant genomes in publicly available databases that can support prospective surveillance, this study aimed to understand the population structure, evolution, pathogenicity, and transmission dynamics of this critical-priority pathogen circulating in southwestern

Retrospective Klebsiella isolates were collected from 3 hospitals that had archival isolates: Obafemi Awolowo University Teaching Hospital (OAU, Ile-Ife, Osun State, Nigeria; n = 92), University College Hospital (UCH, Ibadan, Oyo State, Nigeria; n = 30), and Ladoke Akintola University Teaching Hospital (LAU, Oshogbo, Osun State, Nigeria; n = 12). Of the 134 isolates, 87 were from blood, comprising 45 from OAU, 30 from UCH, and 12 from LAU. We additionally sequenced 47 non-blood isolates from OAU, recovered from urine (n = 37), ocular (n = 2), stool (1), throat (n = 2), and rectal (n = 5) swabs.

Of the 134 genomes confirmed as K. pneumoniae by WGS, 18 (13%) and 33 (25 %) were identified by reference laboratory VITEK and the sentinel laboratories, respectively, as either other members of the Enterobacteriaceae family, Acinetobacter baumannii, or Pseudomonas aeruginosa (Supplementary Table 1). All WGS-identified Klebsiella quasipneumoniae isolates were identified either as K. pneumoniae or Enterobacter complexes by VITEK and the sentinel laboratories, respectively (Supplementary Table 1). The phylogenetic tree, epidemiological data, and analyses data can be visualized in Microreact for retrospective Klebsiella pneumoniae (https://microreact.org/project/GHRUNigeriaKpneumoniae/2a856694), and OAU outbreak isolates (ST17: https://microreact.org/project/brfq17BXzwmqptfcrNRfZR/71b2f0f1; ST25: https://microreact.org/project/8FP4F1D5fSQMv6FDxbR39b/bd381f0b ; ST307: https://microreact.org/project/sV5NsJ8szcorvFAeFQgV4E/6498edf1 ; ST5241: https://microreact.org/project/u5RPX7CjyitjRYMQ9ByW 4W/1a3570e2).

Overall, the K. pneumoniae isolates sequenced in this study belonged to 39 di fferent sequence types (STs), with ST307, ST5241, ST15, and ST25 being most common (Figure 1). Only 4 STs (ST15, ST101, ST147, and ST307) were found across the 3 sentinel sites (Figure 1). Klebsiella quasipneumoniae genomes (n = 5) belonged to 4 STs: ST1602, ST2133, ST5250 (n = 1 each), and ST5249 (n = 2).

We identified few virulence genes associated with invasiveness, including yersiniabactin (ybt genes and fyuA; n = 42, 31.34%), aerobactin (iuc; n = 2), salmochelin (iro; n = 2), and capsule expression upregulators rmpADC and rmpA2 (n = 2). Only 2 ST86 isolates had a virulence score of 3, the highest recorded in our collection. No acquired K pneumoniae virulence gene was detected in novel ST5241 (Figure 2), nor in the non-pneumoniae Klebsiella species.

We detected 33 different phenotypically-defined K (capsular) loci from the invasive K. pneumoniae isolates sequenced, with KL102, KL123, KL2, KL62, and KL27 representing the 5 most common K loci (Supplementary Table 2). Eleven different O loci were detected among the invasive K. pneumoniae isolates recovered from bloodstream infections. Collectively, the 3 O loci 3 O1v1, O2v2, and O1v2 3 accounted for more than 65% of the K. pneumoniae strains, and O5 was solely detected in ST5241 genomes.

Genes conferring resistance or reduced susceptibility to at least 5 antibiotic classes were detected in 116 (86.6%) of K. pneumoniae genomes. Aside from core blaSHV, these included8 ³-lactamase genes, of which the extended-spectrum ³-lactamase (ESBL) blaCTX-M-15 was by far the most common, present in 71.6% (n = 96) of the K. pneumoniae isolates belonging to 33 STs (Figure 2). As seen in other studies, we found this ESBL in strains with plasmids belonging to IncFIB_K (n = 70/96), IncFII_K (n = 51/96), and IncR (n = 43/96) incompatibility grou ps [26, 27]. The blaNDM-1 (n = 8/134), blaNDM-5 (n = 2/134), and blaOXA-48 (n = 1/134) genes were the only carbapenemases detected, and these occurred in 8 STs, including ST392 ( blaNDM-1; n = 2), ST147/147-1-locus-variant (blaNDM-1; n = 2), ST530 (blaNDM5; n = 2), ST43, ST15, ST307, ST716 (blaNDM-1; n = 1 each), and ST219 ( blaOXA-48; n = 1) (Figure 2). OmpK35 truncations were observed together with blaCTX-M-15 in 11 isolates and with blaNDM-1 in 5 isolates. Altogether, 18 (13.5%) K. pneumoniae strains carried carbapenemase genes, and/or ESBL genes plus porin defects known to confer carbapenem resistance [28]. Phenotypic antimicrobial susceptibility testing found 11 of 128 isolates carbapenem non-susceptible, and blaNDM-1 (n = 6), blaNDM-5 (n = 2), and/or blaCTX-M-15 and OmpK35 truncations (n = 1) could account for nonsusceptibility in 9 of them (Figure 2). Concordance analysis showed that phenotypic AST and genotypic AMR gene prediction results agree substantially for carbapenems (concordance: 92.2%; sensitivity: 82% [48.22% - 97.72%]; specificity: 93 .2% [86.97% - 97%]).

Reduced susceptibility to nalidixic acid (n = 70/125) or ciprofloxacin (n = 114/125) could be explained by the presence of one or more combinations of mutations in the quinolone resistance determining region (QRDR) of gyrA and parC (n = 60), plasmid-mediated quinolone resistance genes (qnrS, qnrA, qnrB; n = 82), and/or efflux-mediating mutations in acrR gene (n = 98) (Figure 2). The 5 K. quasipneumoniae were resistant to phenicols, tetracyclines, sulphonamides, trimethoprim, and ³-lactams (https://microreact.org/project/4f6Y56ECE979wkj8oDBh k2/dee16dac). One LAU isolate, collected 6 months after an outbreak of K. quasipneumoniae strains carrying blaNDM-5 in Abuja, Nigeria, was remotely related to those outbreak strains (SNP distance g 235), but, like the other 4 K. quasipneumoniae from the current study, it lacked carbapenemase genes [9].

This report on the population structure of Klebsiella in 3 tertiary hospitals addresses critical knowledge gaps regarding the characteristics ofK. pneumoniae in Nigeria and presents 5 K. quasipneumoniae genomes, which without WGS cannot be differentiated from K. pneumoniae in our setting, even at reference laboratory level. Ours and other data reveal that multiple K. quasipneumoniae lineages, including resistant ones, circulate in Nigeria [9]. We also corroborate previous reports on the diversity of K. pneumoniae, which pose challenges as well as opportunities for curtailing and combating this pathogen [29, 30].

The high prevalence genes conferring resistance to fluoroquinolones and extended-spectrum ³lactams, last-line options for Nigeria9s least affluent patients, in the most common lineages, reported here is of concern in spite of relatively low rates of carbapenem resistance in our setting, or elsewhere in Africa, compared to other countries [ 15, 31, 32, 33 ]. Moreover, while uncommon in this study, ST147 and ST392 strains bearing blaNDM-1, and ST101 strains carrying blaOXA-48were documented [37]. All these lineages deserve close and careful monitoring in national prospective surveillance, and further work is necessary to identify and track mobile elements they carry. The preponderance of plasmids in the Klebsiella strains sequenced reflects the known propensity of this genus to harbor and diversify mobile elements in the face of high selective pressure and has implications for resistance in the many species to which K. pneumoniae transmits DNA [ 2, 34 ].

Clonal group (CG) 307 is believed to be displacing the notorious CG258 as a primary lineage in South Africa, Italy, Colombia, and the USA, and it may be more virulent [ 31, 36-40 ]. The high ST307 frequency here, and in Malawiand South Africa, may explain the low overall prevalence of carbapenem resistance we observed compared to other locations [14, 17, 32, 35]. We recorded 21 ST307 isolates, but no ST258 isolates, and only 7 CG258 isolates 335 ST340 and 2 ST11. Clusters of likely clonal isolates within clonal complexes 15, 25, 17, ST5241, 392, and 101 were identified among retrospective isolates from OAU, the most sampled sentinel site, within clonal complexes, 15, 25, and 17, as well as ST307, ST5241, 392, and 101. The close genetic distances observed within 4 of these clones (< 4 pairwise SNP differences in each case) and the similar or identical resistance, virulence, and plasmid replicon profiles, and temporal clustering of their isolation strongly suggest that each of these clusters represents a retrospectively identified outbreak. As fewer retrospective isolates were stored at the other two institutions, we cannot rule out similar occurrences that are below our limits of detection. Outbreaks were caused by highly resistant lineages such as the ST17 cluster, but also more sensitive ones like the ST5241 and ST307 K. pneumoniae clusters. While outbreaks can sometimes be detected by phenotypic testing in diagnostic labs alone, the minimal biochemical and narrow disc diffusion test repertoires employed at sentinel labs, which underlie many species9 misidentifications (Supplementary Table 1), make detection unlikely when the number of outbreak isolates is small. Our data show that, as in countries with more intensive surveillance, outbreaks likely commonly occur within health facilities in Nigeria and elsewhere in West Africa and will typically be missed without genomic support [ 7, 12, 42, 43 ]. OAU and UCH currently provide subsidized blood culture for many patients, but most Nigerian facilities with blood culture do not, and health insurance coverage is low. Therefore, identifying outbreaks is likely to be curtailed by patients9 inability to pay. Our data reveal that recent efforts to strengthen Nigeria9s AMR response and boost infection prevention and control could be mutually enhancing if blood culture of routine febrile patient blood culture is facility-supported, and if prospective genomic surveillance is used. This study has some limitations. First, very few archival isolates from only 3 facilities were available, and epidemiological data were incomplete. Furthermore, the sequence types observed here were sampled only from invasive infections. Nonetheless, this study represents an important starting point for documenting Klebsiella lineages, and we will build on these findings with prospective surveillance at more hospitals in southwestern Nigeria, which have now been enrolled in the national surveillance system.

We have detailed the characteristics of Klebsiella isolates from 3 southwestern hospitals in Nigeria. In addition to incorporating Nigeria-derived data into global phylogeographies of multidrug-resistantK. pneumoniae, our findings point to the significance of ST307 in Nigeria. Carbapenem-resistant STs 147 and 392, as well as novel ST5241 K. pneumoniae, may also represent future surveillance priorities for southwestern Nigeria. We have shown the benefits of public health genomic surveillance of pathogens 33 in particular, the potential for outbreak identification. Infection prevention and control is a key pillar of Nigeria9s AMR National Action Plan and the data from this paper suggest that genomic surveillance and IPC implementation could be mutually reinforcing in this collaboration among scientists, health institutions, and the public health authority in Nigeria.

Members of the NIHR Global Health Research Unit for the Genomic Surveillance of Antimicrobial Resistance: Khalil Abudahab, Harry Harste, Dawn Muddyman, Ben Taylor, and Nicole Wheeler of the Centre for Genomic Pathogen Surveillance, Big Data Institute, University of Oxford, Old Road Campus, Oxford, United Kingdom and Wellcome Genome Campus, Hinxton, UK; Pilar DonadoGodoy, Johan Fabian Bernal, Alejandra Arevalo, Maria Fernanda Valencia, and Erik C. D. Osma Castro of the Colombian Integrated Program for Antimicrobial Resistance Surveillance 3 Coipars, CI Tibaitatá, Corporación Colombiana de Investigación Agropecuaria (AGROSAVIA), Tibaitatá 3 Mosquera, Cundinamarca, Colombia; K. L. Ravikumar, Geetha Nagaraj, Varun Shamanna, Vandana Govindan, Akshata Prabhu, D. Sravani, M. R. Shincy, Steffimole Rose, and Ravishankar K.N of the Central Research Laboratory, Kempegowda Institute of Medical Sciences, Bengaluru, India; Jolaade J Ajiboye of the Department of Pharmaceutical Microbiology, Faculty of Pharmacy, University of Ibadan, Oyo State, Nigeria; Celia Carlos, Marietta L. Lagrada, Polle Krystle V. Macaranas, Agnettah M. Olorosa, June M. Gayeta, and Elmer M. Herrera of the Antimicrobial Resistance Surveillance Reference Laboratory, Research Institute for Tropical Medicine, Muntinlupa, the Philippines; Ali Molloy, alimolloy.com; John Stelling, The Brigham and Women's Hospital; and Carolin Vegvari, Imperial College London.

We sincerely appreciate technical assistance from sentinel site staff as well as Ifeoluwa Akintayo, Ifeoluwa Komolafe, and Faith Ifeoluwa Popoola. We also thank the team of curators of the Institut Pasteur MLST/cgMLST databases for curating the data and making them publicly available at https://bigsdb.pasteur.fr. We thank Lucy Goodchild for editorial assistance. drug resistance in invasive Klebsiella pneumoniae from South and Southeast Asia. Genome Med. 2020; 12(1): 11. 4. Olalekan A, Onwugamba F, Iwalokun B, Mellmann A, Becker K, Schaumburg F. High proportion of carbapenemase-producing Escherichia coli and Klebsiella pneumoniae among extended-spectrum ³-lactamase-producers in Nigerian hospitals. J Glob Antimicrob Resist 2020; 21: 8312. 5. Müller-Schulte E, Tuo MN, Akoua-Koffi C, Schaumburg F, Becker SL. High prevalence of ESBLproducing Klebsiella pneumoniae in clinical samples from central Côte d9Ivoire. Int J Infect Dis 2020; 91: 2073209. 6. Musicha P, Cornick JE, Bar-Zeev N, et al. Trends in antimicrobial resistance in bloodstream infection isolates at a large urban hospital in Malawi (199832016): a surveillance study. Lancet Infect 7. Okomo U, Senghore M, Darboe S, et al. Investigation of sequential outbreaks of Burkholderia cepacia and multidrug-resistant extended spectrum ³-lactamase producing Klebsiella species in a West African tertiary hospital neonatal unit: a retrospective genomic analysis. The Lancet Microbe 2020; 1: e1193e129. 8. Mbelle NM, Feldman C, Sekyere JO, Maningi NE, Modipane L, Essack SY. Pathogenomics and Evolutionary Epidemiology of Multi-Drug Resistant Clinical Klebsiella pneumoniae Isolated from

Key (LAU = Ladoke Akintola University Teaching Hospital, OAU = Obafemi Awolowo University Teaching Hospital, UCH = University College Hospital).