Isolation and Genetic Characterization of Antibiotic Resistant E. coli from Sewage Water, Islamabad, Pakistan

Abstract

Carbapenemases and extended spectrum beta lactamases (ESBLs) producing Enterobacteria including Escherichia coli (E. coli) have been listed as a critical health threat. Antibiotic resistant E. coli alone caused almost 200,000 human deaths in 2019 (Murray, Ikuta et al. 2022). If same trends continued uninterrupted, antimicrobial resistant (AMR) infections will be the leading cause of human deaths by 2050. There is a dire need for surveillance to see what possible AMR threat is present in our environment. Sewage water is considered a good source of surveillance because it is a kind of reactor which possesses microbes from all sources. Sewage water has human fecal contamination thus can also provide an insight into community gut carriage. Seasonal factors can affect the composition of microbial communities in a certain environment of a geographic region and can provide an insight into controlling/driving factors of microbes. The current study was designed to analyze three clinically important AMR E. coli in our environment to have insight into their prevalence, resistance markers, seasonal/temporal variations, and prevalent sequence types along with their resistance protentional. There were 240 sewage samples (60 samples/ season) collected from same 60 sites in four seasons (fall, winter, spring, summer) of a year starting from October-2016 to September-2017. Sewage samples were subjected onto screening media, which were optimized to directly select antibiotic resistant E. coli from sewage water. These included meropenem resistant E. coli (MEM-R E. coli), ciprofloxacin resistant E. coli (CIP-R E. coli) and cefotaxime resistant (CTX-R E. coli). All colonies on antibiotic containing selective MacConkey agar which appeared as E. coli based on colony morphology were further confirmed through standard biochemical tests and detection of uid gene through PCR. All confirmed E. coli isolates were analyzed for their phenotypic resistance, antibiotic resistance genes, and phylogenetic groups. All study isolates were subjected to 15 different antibiotic discs through Kirby Baur disc diffusion test. Clermont quadruplex PCR was used for analysis of phylogenetic groups. For screening of antibiotic resistance markers in study isolates, 4 classes of antibiotic resistance genes (ARGs) were screened including carbapenemases, ESBLs, plasmid mediated quinolones resistance genes (PMQRs), and plasmid encoded colistin resistance genes. A subset of study isolates was analyzed for their sequence type by 2-locus/CH sequence typing. Another subset of study isolates was whole genome sequenced to analyze their multi locus sequence typing (MLST), ARGs, plasmid profile, and virulence genes profile. x There were 93% (n=56) sites which had at least one type of antibiotic resistant E. coli present during at least once in the sampling year. All 240 sewage samples processed for MEM-R E. coli isolation. For CIP-R and CTX-R E. coli 120 sewage samples (spring and summer season) were processed. MacConkey agar supplemented with 3μg/ml meropenem was used to select MEM-R E. coli from sewage water. For CIP-R and CTX-R E. coli 4μg/ml, and 1μg/ml antibiotic concentrations were used respectively. There was 27% prevalence of MEM-R E. coli followed by 58% CIP-R E. coli and 70% CTX-R E. coli in sewage water. Temporal/ seasonal variation of MEM-R E. coli showed higher prevalence in summer (50%) followed by spring (22%), fall (18%) and lowest in winter (17%). The p value was below 0.05 which showed the temporal variation of MEM-R was significant. The prevalence of CIP-R E. coli was 68% and 48% in summer and spring season respectively. The p value was <0.05 indicating a significant temporal variation of CIP-R E. coli. The CTX-R E. coli was also higher in summer at 75% followed by 65% in spring season. There was total 566 confirmed E. coli isolates including 124 MEM-R, 224 CIP-R, and 218 CTXR E. coli which were further analyzed. All study isolates were MDR including 40% XDR isolates and there was no PDR isolate. 51% of MEM-R E. coli were XDR followed by CIP-R (42%) and CTX-R (32%) XDR isolates. Among all spring isolates 50% were XDR while summer season isolates were 34% XDR. Spring season had significantly higher XDR prevalence than summer season. Antibiogram of study isolates was as followed: ampicillin (AMP) 98%, ceftazidime (CAZ) 67%, ampicillin-clavulanic acid (AMC) 72%, cefotaxime (CTX) 94%, cefepime (FEP) 72%, aztreonam (ATM) 80%, meropenem (MEM) 19%, ciprofloxacin (CIP) 84%, chloramphenicol (C) 43%, nalidixic acid (NA) 91%, gentamicin (CN) 48%, doxycycline (DO) 75%, nitrofurantoin (F) 15%, fosfomycin (FOT) 1% and sulfamethoxazole-trimethoprim (SXT) 93%. Isolates were most susceptible to fosfomycin (97%). Phylogenetic group analysis result in 32% isolates remained unknown while a total of 68% isolates were typeable including B1 (1%), B2 (6%), A (1%), A/C (15%), D/E (2%), E/cladI/II (7%), F (<1%) and Clad I/II (36%). The pathogenic groups B2 was more prevalent in spring season (11%) while isolates of summer season mostly belonged to cladI/II (45%). Detection of ARGs showed, among carbapenemases, blaOXA-48 (33%) was most prevalent, followed by blaNDM (8%), and blaKPC (<1%). Among ESBLs highest prevalence was of blaCTX-M xi (61%), followed by blaTEM (58%), and blaSHV (24%). Among PMQR markers the most prevalent was qnrS (75%), followed by qnrB (33%) and qnrA (4%). Among mcr group of genes (mcr-1 to mcr-5) only mcr-1 (37%) was found. MEM-R E. coli had 23% isolates which had ≥1 gene of carbapenemases, ESBLs and PMQRs. Among CIP-R and CTXR E. coli 46% and 39% isolates had co-carriage of ESBLs and PMQRs. A subset of study isolates was analyzed for their sequence type by CH/2-locus typing. There were seven isolates with successfully assigned ST, five of these isolates belonged to ST405 and two isolates belonged to ST216 and ST3285 each. The whole genome sequence analysis of a subset of eight isolates representative of three sets of study isolates of spring and summer season was done. Their MLST analysis showed two of isolates belonged to ST-405 and single isolate belonged to each of ST-48, ST-361, ST-424, ST-1139, ST- 7366 and ST-7401. Phylogenetic relatedness of ST showed ST-7366, ST-361, ST-48 belonged to Clad1 and Clade II had only two STs, including ST-405 and ST-7401. ST-1139 and ST-424 were single STs. A plethora of 27 different ARGs were found among different STs. The most prevalent resistance marker was blaTEM-1B found in six isolates. Only ST405 had blaNDM-5. ST361 carried blaOXA-181. Various incompatibility groups of plasmids were found and IncF was the most common type among all. There were 38 types of virulent genes found in different STs and ST405 was carrying (n=31) higher number of virulence factor in comparison to others. Comparison of carbapenemases plasmid showed blaNDM-5 of ST405 was similar to previously reported E. coli isolate of chicken meat from Pakistan. While blaOXA-181 plasmid of ST361 was similar to K. pneumoniea human isolate plasmid isolated from China. Current study findings clearly suggest that wastewater is contaminated with antibiotics resistance E. coli. There were nine different STs and all were potentially human pathogens. Presence of highly virulent and extensive drug resistant ST-405 was predominant. There is need to focus on strategies to specifically control the clinically important AMR bacteria in the environment to avoid occupational hazard and potentially community hazard in future.