Carbapenem Resistance Proles of Pathogenic Escherichia coli in Uganda

Background: Escherichia coli has been implicated as one of the main etiological agents of diarrhea, urinary tract infections, meningitis and septicemia worldwide. The ability to cause diseases is potentiated by presence of virulence factors. The virulence factors inuence the capacity of E. coli to infect and colonize different body systems. Thus, pathogenic E. coli are grouped into DEC strains that are mainly clustered in phylogenetic group B1 and A; ExPEC belonging to A, B2 and D. Coexistence of virulence and beta-lactamase encoding genes complicates treatment outcomes. Therefore, this study aimed at presenting the CR proles among pathogenic E. coli. Methods: This was a retrospective cross-sectional study involving use of archived E. coli clinical isolates collected in 2019 from four Ugandan tertiary hospitals. The isolates were subjected to antibiotics sensitivity assays to determine phenotypic resistance. Four sets of multiplex PCR were performed to detect CR genes, DEC pathotypes virulent genes, ExPEC PAI and the E. coli phylogenetic groups. Results: Antibiotic susceptibility revealed that all the 421 E. coli isolates used were MDR as they exhibited 100% resistance to more than one of the rst-line antibiotics. The study registered phenotypic and genotypic CR prevalence of 22.8% and 33.0% respectively. The most predominant gene was blaOXA-48 with genotypic frequency of 33.0%, then blaVIM(21.0%), blaIMP(16.5%), blaKPC(14.8%) and blaNDM(14.8%). Spearman’s correlation revealed that presence of CR genes was highly associated with phenotypic resistance. Furthermore, of 421 MDR E. coli isolates, 19.7% harboured DEC virulent genes, where EPEC recorded signicantly higher prevalence (10.8%) followed by S-ETEC(3.1%), STEC(2.9%), EIEC (2.0%) and L-ETEC(2.0%). Genetic analysis characterized 46.1% of the isolates as ExPEC and only PAI IV536(33.0%) and PAI IICFT073(13.1%) were detected. Phylogenetic group B2 was predominantly detected (41.1%), followed by A(30.2%), B1(21.6%), and D(7.1%). Furthermore, 38.6% and 23.1% of the DEC and ExPEC respectively expressed phenotypic resistance. Conclusion: Our results exhibited signicant level of CR carriage among the MDR DEC and ExPEC clinical isolates belonging to phylogenetic groups B1 and B2 respectively. Virulence and CR genetic factors are mainly located on mobile elements. Thus, constitutes a great threat to the healthcare system as it promotes horizontal gene transfer. studies observed that PAIs are mobile genetic elements (transposons) that are transferred from one E. coli strain to another through horizontal gene transfer mediated by bacteriophages, conjugative plasmids, conjugation and homologous DNA recombination [64, 81, 82] Our data indicate high level of carriage of carbapenem resistance among the DEC and ExPEC clinical isolates belonging to phylogenetic group B1 and B2 respectively. DEC and ExPEC pathogenicity and antimicrobial resistance are mediated by genetic factors such as chromosomal/plasmid borne virulence and antibiotic resistance genes as well as chromosomal PAIs virulent genes. Plasmid and PAIs are mobile genetic elements that facilitate horizontal gene transfer contributing to plasticity of the genome. In light of this, routine genetic analysis of E. coli clinical and environment isolates is important to better understand the level of pathogenicity and antimicrobial as this will inform the possible burden such isolates are likely to pose to the healthcare system.


Introduction
Escherichia coli is one of the most prevalent commensals of the human gastro-intestinal tract (GIT) microbiota. However, some E. coli are pathogenic. Pathogenic E. coli comprises of diarrheagenic E. coli (DEC) [1] and Extra-intestinal pathogenic E. coli (ExPEC) pathotypes [2]. Diarrheagenic pathotypes are responsible for all gastrointestinal tract E. coli infections most importantly diarrhea. Diarrhea is one of the principal causes of illness and death among children under 5 years in developing countries and DEC pathotypes account for the biggest percentage. Reaching protective immunity against DEC in children is hard as DEC is composed of a wide range of pathotypes, hence variant antigens. Extra-intestinal pathogenic E. coli is accountable for all E. coli associated infections outside the gastrointestinal tract, such as meningitis, urinary tract infections (UTI), pneumonia, septicemia, among others [3][4][5]. An alarming prevalence of bacterial UTI has been registered in primary healthcare. Escherichia coli has been implicated to be the chief etiology of both community and nosocomial acquired UTI worldwide.
Diarrheagenic E. coli are grouped into eight pathotypes basing on virulent factors responsible for their pathogenicity. These include Enteropathogenic E. coli (EPEC), Enteroinvasive E. coli (EIEC), Enteroaggregative E. coli (EAEC), Enterotoxigenic E. coli (ETEC) and Diffusely Adherent E. coli (DAEC), Shiga toxinproducing E. coli (STEC) also commonly known as enterohemorrhagic E. coli (EHEC) or Verotoxigenic E. coli (VTEC), the newly identi ed adherent invasive E. coli (AIEC) which is alleged to been associated with Crohn's disease but not with any diarrheagenic infections and a hybrid pathotype, enteroaggregative hemorrhagic E. coli (EAHEC) carrying STEC and EAEC virulence genetic determinants [3]. Thus, pathogenic DEC encompasses a genetically heterogeneous family of E. coli with a plastic genome. Several research articles suggest that each pathotype possesses and codes for distinctive virulence and colonization determinants harboured in their genomes distinguishing them from other pathotypes and non-virulent strains. These virulence factors for each pathotype are encoded for by conserved genes and are restricted within geographical boundaries [6,7]. Therefore, molecular typing of Escherichia coli to identify the different DEC pathotypes can be achieved by targeting virulent genes. These virulent genes include; eae for typing of EPEC; stx for STEC/EHEC; est/ for st-ETEC; elt for lt-ETEC aggR for EAEC; ipaH for EIEC. eae gene is translated into Intimin polypeptide which is the key factor for attaching and effacing lesions; stx gene encodes for the Shiga-like toxin; elt and est genes are translated into Thermolabile and Thermostable toxins respectively; ipaH gene accounts for invasion capacity and aggR gene is translated into a transcriptional activator protein of aggregative adherence mbriae [8].
Furthermore, PCR analysis clusters E. coli strains into A, B1, B2, and D phylogenetic groups due to the presence of the chuA and yjaA genes as well as TSPE4.C2 DNA fragment [13]. The intestinal pathogenic E. coli strains belong to groups A, B1 and D, extraintestinal pathogenic E. coli strains generally follow under groups B2 and D, while commensal E. coli strains to groups A and B1 [13,14] High levels antibiotic resistance in Enterobacteriaceae is of great concern to the healthcare system [15,16]. Escherichia coli like other Enterobacteriaceae has evolved to acquire different mechanisms of antibiotic resistance which confer protection to lethal doses of different classes of antibiotics. Carbapenems are the most suitable antibiotics used in the treatment of multidrug resistant (MDR) gram-negative bacteria infections. Studies have documented high prevalence of carbapenem resistant Enterobacteriaceae (CRE) in Uganda [17,18]. However, the carbapenem resistance pro les of DEC and ExPEC human isolates have not been investigated, yet for meaningful treatment outcomes and prescription decisions, knowledge about pathogen susceptibility patterns to antibiotics in question is very important. Thus, this study was aimed at pro ling the carbapenem resistance pro les of intestinal and extraintestinal human pathogenic E. coli isolates for genetic markers allied with DEC and ExPEC strains. The study relied on the screening for DEC genetic markers, PAI associated sequences for ExPEC and determination of phylogenetic group and genetic determinants of carbapenem resistance (CR).

Materials And Methods
Study design, site and source of bacteria isolates This was a cross sectional-laboratory-based study conducted at the Microbiology Laboratory and Molecular Biology Laboratory, College of Veterinary Medicine Animal Resources and Biosecurity (CoVAB) Makerere University. The study involved use of archived MDR Escherichia coli samples isolated between January and December, 2019 from clinical specimens in the Microbiology Laboratories of Mulago National Referral Hospital (MNRH), Mbale Regional Referral Hospital (MRRH), Mbarara Regional Referral Hospital (MBRRH) and Kampala International University Teaching Hospital (KIU-TH). The samples were transported in peptone water to the Microbiology Laboratory, CoVAB. Overnight cultures of E. coli were prepared by pipetting 1 ml of peptone water containing each isolate into 49 ml of Luria-Bertani (LB) broth. Glycerol stocks of each different isolate were made by adding 500 µl of the overnight LB culture to 500µL of 50% glycerol in a 2 ml screw top tube and mixed gently mix. The screw tubes were stored at -80 0 C until further use.
Biochemical assays to con rm the identity of E. coli To con rm the identity of each isolate, Microgen (Micro-biology International) kits for biochemical assays were employed using procedures described by the manufacturer (www.microgenbioproducts.com).

DNA extraction
Pure colonies of E. coli from different samples were selected and each sub-cultured in 5 ml of Luria-Bertani broth using sterile inoculating loop. The bacterial suspension was incubated in shaker incubator at 37ºC for 24hrs. Then, 1 ml of bacterial suspension was transferred into a 1.5 ml eppendorf tube, centrifuged at 10,000 rpm for 10 minutes. The supernatant was discarded and the pellet was re-suspended in 200 µl of Gram-negative bacteria lysis buffer provided in the Qiagen DNA extraction. Bacterial total genomic DNA was extracted following the Qiagen DNA extraction protocol and stored at -20 o C until further use.  [21] were used as positive controls.  GAT ATT TTT GTT GCC ATT GGT TAC C Phylogenetic Classi cation.
Phylogenetic classi cation exhibited that the E. coli strains belonged to four groups (A, B1, B2, or D) based on the presence of the chuA and yjaA genes and the DNA fragment (TSPE4.C2). Thus, a multiplex PCR was run to determine the phylogenetic classes of the E. coli strains using primers targeting chuA, yjaA and TSPE4.C2 DNA sequences, Table 4. The PCR ampli cation was conducted by adapting [13] methods. Brie y, the PCR contained 2.5 µl of template DNA, 1U Taq DNA polymerase (Biomatik, USA) in 1x PCR buffer (Biomatik), 200 µM dNTP, 2.5 mM MgCl 2 , and 0.8 M of each primer, Table 1. Ampli cation was conducted using the following PCR conditions; initial denaturation at 94 °C for 5 minutes, then 30 cycles performed at 94 °C for 5 seconds, 54 °C for 10 seconds, 72 °C for 30 second with a nal extension step at 72 °C for 5 minutes. Phylogenic groups and subgroups were assigned depending on chuA, yjaA, and TspE4.C2 gene combinations [13,14], Table 5. CGC GCC AAC AAA GTA TTA CG   Table 8. Four isolates were found to co-harbour more than one carbapenemase encoding genes, with blaOXA and blaNDM co-existing in two isolates, blaOXA-48 and blaKPC in one isolate and blaOXA-48, blaKPC and blaNDM in one isolate but exhibited no phenotypic resistance,  Table 9. urine, virginal swabs, blood, wound/pus swabs was substantially higher than ExPEC prevalence obtained from tracheal aspirate, sputum and anal swabs, Table 9.
Distribution of the Escherichia coli Phylogenetic groups E. coli (421) isolated from several clinical specimens were characterized into four phylogenetic groups (PG) and six phylogenetic subgroups based on the triplex PCR. E. coli belonging to phylogenetic group B2 was predominantly detected and scored a prevalence of 41.1%. This was trailed by phylogenetic group A (30.2%), phylogenetic group B1 (21.6%) and phylogenetic group D (7.1%). E. coli belonging to Phylogenetic group A, B2 and D were majorly isolated from urine samples whereas phylogenetic group B1 isolates were mainly obtained from anal swabs, Table 10.     [29]. Contrary, this frequency is higher than carbapenem resistance levels reported in countries like Ghana (7.2%) [30], Morocco 5.99% [31], and Ethiopia 2.73%, [32] with similar healthcare settings but lower than the incidences above 50% reported in South Africa, Egypt and Tunisia [33][34][35][36][37].
Multiplex PCR screening identi ed carbapenemase encoding genes in 33.0% of the isolates. This genotypic carbapenem resistance prevalence corroborates with earlier studies conducted in the East African region [18,25,38] and elsewhere [27,39] that reported levels ranging from 25-40%. Contrary, this frequency is signi cantly lower than carbapenem genotypic levels reported by studies in Tunisia (76.7%) [40], South Africa (68% and 86%) [33,34], Egypt (89.6%) [37], Turkey (49.5%) [41]. KPC, VIM, NDM, OXA-48 and IMP are the commonest carbapenemases worldwide [42]. Findings of this study revealed the existence of all those carbapenemases encoding genes in Uganda and OXA-48 was the most predominant gene in contrast with previous studies in the region [18,25] but in agreement with recent studies in carried out in Africa ( [31,33,36,40]. OXA-48 carbapenemase was rst detected in Turkey and it became epidemic in the Middle East and Mediterranean countries [41]. This indicates that OXA-48 habouring Enterobacteriaceae have spread widely in sub-Saharan Africa to become to most prevalent. This study found considerable variation between phenotypic and genotypic resistance. Among the E. coli isolates that harboured blaVIM gene, 97.3% exhibited phenotypic resistance while for blaOXA-48, only 37.9% expressed phenotypic resistance. It is important noting that four isolates coharboured more than one gene each but susceptible to carbapenems. Carbapenemases expressed by OXA-48 and its variant genes possess low carbapenems hydrolyzing activity [43][44][45]. This provided an insight into why 62.1% of the isolates which possessed OXA-48-like genes did not exhibit phenotypic resistance. Alteration and reduced expression of the outer membrane proteins that act as drug channels complement enzymes expressed by the resistant genes and this mechanism is highly effective against Ertapenem [45,46]. Thus, carbapenem resistance is not exclusively due to expression carbapenemases. This explains why not all the isolates that harbored carbapenemase genes were carbapenem insusceptible and why resistance to ertapenem was signi cantly higher. Despite of absence of carbapenem resistance genes, a total of eight sample displayed phenotypic resistance. Thus, resistance in these isolates may be attributed to ( University Hospital metallo-lavtamase (KHM) [49].  [58,59].
In this study, multiplex PCR was used to target Pathogenicity Islands (PAI). PAIs harbour virulent genes in ExPEC that are responsible for pathogenicity [60][61][62]. The overall prevalence of ExPEC as revealed by molecular typing of PAI in our study was 46.1% (194/421). Of the two PAIs detected, PAI IV536 also known as high pathogenicity Island (HPI) was substantially dominant with a genotypic frequency of 71.7% and PAI IICFT073 had a frequency of 28.3%. This is in agreement with previous studies [24,60] which reported PAI IV536 as the most prevalent PAI. The main virulence genes residing in the PAI IV536 and PAI IICFT073 are yersiniabactin siderophore iron-uptake system and P. mbriae as well as iron regulated proteins respectively [61,63,64]. A previous study in Uganda reported high prevalence of E. coli with P. mbriae virulent factor encoded for by the pap gene in UPEC [65]indicating high prevalence of PAI IICFT073 pathotypes. However, this study never attempted to detect genes encoding the yersiniabactin siderophore iron-uptake system in PAI IV536 UPEC; thus, there is no available data about the prevalence of PAI IV536 for comparison purposes. As anticipated, ExPEC that harboured PAIs were majorly isolated from urine and virginal swabs. However, a total of 55 isolates obtained from blood (17) [13,14]. In our study, phylogenetic analysis predominantly clustered E. coli clinical isolates into B2 followed by A, B1 and D and this corroborates with ndings from previous studies [67][68][69]. However, contradicting results have been reported worldwide where A is the most abundantly isolated phylogroup [70][71][72][73]. Distribution of E. coli phylogroups among different ecological zones is in uenced by environment factor; thus, this accounts for variability in prevalence of the phylogenetic groups in different countries [74]. In this study, statistically similar (P value 0.9998) distribution of phylogenetic groups A, B1, B2 and D among regions was observed. This pattern of distributes indicates inter-region transmission of UPEC, DEC and commensals. It was observed that phylogenetic group A, B2 and D strains were majorly isolated from urine and this is in a rmative with all studies that conducted phylogenetic analysis of E. coli clinical isolates [67,[75][76][77] whereas B1 strains were predominantly isolated from anal/fecal swabs, this does not corroborate with previous studies which found phylogroup A strains as the most dominant fecal isolates [72,73,78].
World over, an increase in pathogenic and commensal E. coli strains harbouring antibiotic resistance determinants has been observed. The situation has been complicated by acquisition of antibiotic resistance by other Enterobacteriaceae as several studies have reported that infections caused by resistant bacteria are hard to treat, lead to increase in treatment costs, morbidity and mortality [79]. Antibiotic resistance in Enterobacteriaceae is mainly mediated by betalactamase enzymes that inactivate beta-lactam antibiotics by hydrolyzing the peptide bond of the beta-lactam ring. Among the beta-lactamases, carbapenemases are the most important because acquisition of carbapenem resistance genes confer resistance to all beta lactam antibiotics. Furthermore, carbapenem are the most suitable choice antibiotics for treatment of MDR Gram-Negative bacterial infections; [43] thus, infection with carbapenem resistant bacteria signi cantly prolong the period of stay in hospital and responsible for 10% mortality [80]. Thus, in this study we assessed the carriage of carbapenem resistance and virulence genetic factors among E. coli phylogroups. We observed that among the 83 isolates that harboured virulence genetic determinants for DEC, 98.8% (82) and 1.2% (1) belonged to phylogenetic group B1 and A respectively and 38.6% (32) expressed phenotypic resistance. Whereas 86.1% (167), 12.4% (24) and 1.6% (3) of the isolates that had PAIs were characterized as phylogroups B2, D and A respectively and 24.1% (47) were carbapenem resistant.
Coexistence of virulence factors and carbapenem resistance was observed in 18.8% (79/421) of the total isolates. Our ndings show that carbapenemases production was signi cantly higher in B1 and B2 (P < 0.0001). This is extremely scaring as DEC and ExPEC mainly fall under phylogenetic groups B1 and B2 respectively. Furthermore, existence of virulence genes and genetic determinants of resistance in phylogenetic groups A and D where commensal mainly fall should be treated as a major threat as they are considered to reservoirs of genetic determinants of virulence and antibiotic resistance and they donate these traits to the pathogenic strains of phylogroups B1 and B2 through horizontal gene transfer, arbitrated regularly by plasmids and transposons [44]. Indeed, previous studies observed that PAIs are mobile genetic elements (transposons) that are transferred from one E. coli strain to another through horizontal gene transfer mediated by bacteriophages, conjugative plasmids, conjugation and homologous DNA recombination [64,81,82]

Conclusion
Our data indicate high level of carriage of carbapenem resistance among the DEC and ExPEC clinical isolates belonging to phylogenetic group B1 and B2 respectively. DEC and ExPEC pathogenicity and antimicrobial resistance are mediated by genetic factors such as chromosomal/plasmid borne virulence and antibiotic resistance genes as well as chromosomal PAIs virulent genes. Plasmid and PAIs are mobile genetic elements that facilitate horizontal gene transfer contributing to plasticity of the genome. In light of this, routine genetic analysis of E. coli clinical and environment isolates is important to better understand the level of pathogenicity and antimicrobial as this will inform the possible burden such isolates are likely to pose to the healthcare system.