Detection of extended spectrum beta-lactamase genes in Pseudomonas aeruginosa isolated from patients in rural Eastern Cape Province, South Africa

The proliferation of extended spectrum beta-lactamase (ESBL) producing Pseudomonas aeruginosa represent a major public health threat. In this study, we evaluated the antimicrobial resistance patterns of P. aeruginosa strains and characterized the ESBLs and Metallo- β-lactamases (MBL) produced. Strains of P. aeruginosa cultured from patients who attended Nelson Mandela Academic Hospital and other clinics in the four district municipalities of the Eastern Cape between August 2017 and May 2019 were identified; antimicrobial susceptibility testing was carried out against thirteen clinically relevant antibiotics using the BioMérieux VITEK 2 and confirmed by Beckman autoSCAN-4 System. Real-time PCR was done using Roche Light Cycler 2.0 to detect the presence of ESBLs; blaSHV, blaTEM and blaCTX-M genes; and MBLs; blaIMP, blaVIM. Strains of P. aeruginosa demonstrated resistance to wide-ranging clinically relevant antibiotics including piperacillin (64.2%), followed by aztreonam (57.8%), cefepime (51.5%), ceftazidime (51.0%), piperacillin/tazobactam (50.5%), and imipenem (46.6%). A total of 75 (36.8%) multidrug-resistant (MDR) strains were observed of the total pool of isolates. The blaTEM, blaSHV and blaCTX-M was detected in 79.3%, 69.5% and 31.7% isolates (n = 82), respectively. The blaIMP was detected in 1.25% while no blaVIM was detected in any of the strains tested. The study showed a high rate of MDR P. aeruginosa in our setting. The vast majority of these resistant strains carried blaTEM and blaSHV genes. Continuous monitoring of antimicrobial resistance and strict compliance towards infection prevention and control practices are the best defence against spread of MDR P. aeruginosa.

www.nature.com/scientificreports/ acquisition of Extended-Spectrum β-Lactamases (ESBLs) and Metallo-β-Lactamases (MBLs) 9 . ESBLs are a cluster of β-lactamases that inactivates β-lactams especially oxymino-β-lactams and monobactams, and are repressed by β-lactamase inhibitors, such as clavulanic acid. They are encoded on plasmids and can easily be conveyed from one organism to another 10 . ESBL enzymes according to Ambler classification are categorized into two, A and D. The most prevalent enzymes in class A include bla TEM , bla CTX-M and bla SHV , and has been described in P. aeruginosa strains 10,11 . The emergence of beta-lactamase enzymes is majorly due to chromosomal mutation and procurement of resistance genes which are moved about on various mobile genetic elements (MGEs) such as-bacteriophages, genomic islands, integrons, plasmids and insertion sequences 12 . The production of these enzymes is a going concern for infection control supervision because it restricts therapeutic choices. Continuous monitoring and timely detection of ESBL and MBL producing organisms is critical to establish suitable antimicrobial therapy and to thwart their spread 13 . Polymerase chain reaction (PCR)-based methods are critical to establish the prevalence and characterization of beta lactamases due to the presence of multiple resistance genes in some microorganisms 14 . Real-time PCR (rPCR) detection of ESBL enhances faster diagnosis and timely management of epidemiological information for monitoring outbreak situations 15 . Studies on ESBL-producing P. aeruginosa in South Africa have been documented from other provinces [16][17][18] but scarce data exist in the Eastern Cape particularly in the former Transkei region on the molecular detection of ESBLs and MBLs in P. aeruginosa. Antibiotic surveillance studies are important for the design of control strategies for preventing bacterial resistance and establishing therapeutic guidelines as well as for a better understanding of bacterial epidemiology. The first reported National antimicrobial resistance (AMR) surveillance in South Africa 19 reported ESKAPE organisms causing bacteremia hence not much data for -comparative analysis. The data from the comprehensive view of AMR in blood cultures for ESKAPE pathogens revealed that 20% and 25% of P. aeruginosa bloodstream isolates were resistant to piperacillin/tazobactam and carbapenems respectively. To the best of our knowledge, there are few reports on surveillance of antimicrobial resistance (AMR) in clinical isolates of P. aeruginosa from all samples obtainable particularly from the Eastern Cape. The aim of this study was to examine the antimicrobial susceptibility profiles of clinical strains of P. aeruginosa obtained from patients attending healthcare facilities in the four district municipalities in Eastern Cape and to investigate their ESBL and MBL resistance mechanisms.

Results
Identification of P. aeruginosa and demographics. During the study period, a total of 204 P. aeruginosa isolates were identified from a range of clinical specimens of patients who attended various healthcare facilities in the OR Tambo district municipality, Alfred Nzo, Joe Gqabi and Amathole districts in the Eastern Cape Province. The strains were identified by Vitek 2 system (bioMérieux, Inc., USA), and confirmed by both Microscan autoscan-4 system (Beckman Coulter, Inc. USA) and rPCR using specific primer and probes targeting gyrB. The majority of the strains were from male patients (60%) while 40% belonged to female patients. The strains were predominantly from pus and wound swabs (80.4%), with surgical wounds constituting 43.3%, burn wounds 3.7% and others accounted for 53.0%. These samples originated from Surgical (33.3%), General (18.1%) and Paediatrics (11.3%) wards.
The mean age of patients was 32.8 years ranging from 6 days to 84 years. The male population have a mean age of 30.5 years ranging from 6 days to 83 years while the female mean age was 36.2 years with age ranging from 22 days to 84 years. These patients were drawn from four district municipalities with OR Tambo having the most at 80.4% with the least patient drawn from Joe Gqabi at 0.5% (Table 1).

Molecular Detection of ESBL-and MBL-encoding genes.
Of 204 P. aeruginosa isolates, 82 were tested by singleplex rPCR for detection of ESBL and MBL. These data revealed that ESBL-genotypic resistance is driven by bla TEM (79.3%) followed by bla SHV (69.5%) and lastly bla CTX-M (31.7%). MBL-genotypic resistance, bla VIM , was not detected in all strains tested while only one bla IMP (1.25%) was detected ( Table 2). The most common ESBL-genotype combination among the P. aeruginosa was a combination of bla TEM + bla SHV (40.5%).

Discussion
The current study revealed that antibiotic resistance was observed in piperacillin (64.2%), aztreonam (57.8%), Cephalosporins (cefepime 51.5% and ceftazidime 51.0%), antipseudomonal penicillins + β-lactamase inhibitor (piperacillin/tazobactam 50.5%) and imipenem (46.6%). (Fig. 1). Others included gentamicin (35.3%), meropenem (24.0%) and amikacin (20.1%). Tobramycin was found to be the most potent antibiotic with a susceptibility of 91.7% followed by both doripenem and ciprofloxacin (88.7%) and levofloxacin (80.9%). Data from surveillance on isolates of P. aeruginosa in the South African public sector is not in agreement with the present study 20 . They reported much lower resistance rates of 15%, 16%, 24%, 23% and 19% in cefepime, ceftazidime, imipenem, meropenem and piperacillin/tazobactam respectively. The isolates were recovered from blood cultures only, possibly this might account for the variations in resistance rate, alternatively this might be due to regional variations in the empirical use of these antimicrobials 20 . This also underscores the importance of continuous www.nature.com/scientificreports/ local antimicrobial resistance surveillance for appropriate antibiotic treatment recommendations at the local and facility level since the data only accounted for bacteremic isolates, as there is no national surveillance data to facilitate comparison. In this study, the percentage of resistance of 11.3% to ciprofloxacin was within the same range of 13.4% described by Ramsamy et al. 18 . The data obtained were from nine public sector hospitals in KwaZulu-Natal Province. Additionally, gentamicin resistance of 17% and imipenem resistance of 13% as reported in the study was lower in comparison with resistance reported in the current study at 35.3% and 46.6% respectively. The susceptibility ranges of 75%-92% of P. aeruginosa isolates in this study to some routine   www.nature.com/scientificreports/ antibiotics considered for therapy is encouraging but the increase in resistance exhibited to cephalosporins and imipenem is concerning. This might be due to selective pressure to those antibiotics and it will be important to monitor the prescription of these antibiotics. Owing to endless alteration, resistance exhibited to range of β-lactam antibiotics is challenging, thus making β-lactamase production the commonest cause of drug resistance and antimicrobial treatment failure 7,21 . This study detected an average resistance of 51.3% to the cephalosporins (ceftazidime and cefepime). Piperacillin and gentamicin resistance was 64.2% and 35.3% respectively similar to the findings of Uc-Cachon et al. 22 . The emerging level of resistance displayed to the cephalosporins highlight the development of cephalosporinases among resistant strains of these organisms. The cephalosporins due to their wide spectrum of activity are a significant class of antimicrobials used in controlling several infections however; the emergence of cephalosporinases can in effect hamper their clinical usefulness 23 . The reported increasing penicillinase-producing β-lactamases strains among these organisms validates the noticeably observable high rate resistance of our isolates to piperacillin 23 . Piperacillin is a penicillin beta-lactam antibiotic with in-vitro activity against Gram-positive and Gram-negative aerobic and anaerobic bacteria but because it is prone to hydrolysis by β-lactamase enzymes, its combination with tazobactam, a β-lactamase inhibitor, enhances the in-vitro activity of piperacillin to bacterial cells. This was noticed in the differences in the resistance to the two antibiotics with piperacillin having a higher resistance of 64.2% as compared to 50.5% in piperacillin/tazobactam. Antibiotic resistance is a public health menace with an alarming proportion that is receiving collective attention more so that several studies have found a correlation between level of antibiotic prescription with the prevalence of antibiotic resistance 24,25 . Patients with resistant P. aeruginosa infections have a poor prognosis hence it is imperative that P. aeruginosa strains presenting severe drug resistance is monitored 26 . The swift spread and the emergence of MBL-and ESBL-producing P. aeruginosa of clinical origin is distressing and of great threat. Furthermore, level of antibiotic usage, horizontal gene transfer (HGT) event and environmental factors may account for variations in resistance patterns among strains isolated from diverse countries and regions. In the present study, 36.8% of the strains were MDR (defined as non-susceptibility to at least one agent in three or more antimicrobial categories). Studies have indicated that multidrug resistance often results into limited treatment options and adverse clinical and economic outcomes 27,28 .
Antimicrobial treatment is further hampered by the production of extended spectrum beta-lactamases and metallo beta-lactamases. The emergence of ESBL-producing P. aeruginosa is increasingly reported as a major cause of health-care associated infections. In the hospital locale, infections resulting from these resistant organisms are increasingly challenging to treat due to the intensity of resistance exhibited to the most commonly recommended antibiotics 14,29 . This study found out that the most prevalent genotype for ESBL production was bla TEM, which was detected in 65 (79.3%) strains followed by bla SHV (69.5%) and bla CTX-M (31.7%) ( Table 2). It has been reported that ESBL genes show variation depending on the geographical location, the findings of Erhlers et al. 30 , Chen et al. 31 and Miranda et al. 32 , from South Africa, China and Brazil respectively corroborated our results of the prevalent genotype as bla TEM while in contrast Jamali et al. 33 reported the prevalent gene to be bla SHV . The predominant ESBL and MBL genes detected in a study conducted in Durban on MDR P. aeruginosa isolates were GES-2, OXA-21, and VIM-2 34 . The least detected ESBL genotype from this study was bla CTX-M (31.7%) similar to Miranda et al. 32 . Although no bla VIM was detected in our study, MDR P. aeruginosa encoding bla VIM-2 gene have been reported in a tertiary hospital in Cape Town, which was responsible for an outbreak, and in a public hospital in Port Elizabeth 35,36 . The phenotypic resistance displayed to the carbapenems particularly imipenem which is not validated by the genotypic MBL result may be due to other resistance mechanism such as efflux over expression or forfeiture of exterior membrane protein 37 .
Several researchers have reported on the concurrence of different β-lactamase genes found in the same strains 38,39 . The most common ESBL combination in this study was a combination of bla TEM + bla SHV (40.5%) contrary to Chen et al. 31 , who reported the commonest to be bla SHV + bla CTX-M . The second most common genotype combination was bla SHV + bla CTX-M. This study showed the most predominant ESBL gene was bla TEM , which www.nature.com/scientificreports/ is corroborated by other studies. Prior to now, bla TEM used to be the most prevalent but recent reports suggest that the CTX-M-type group of ESBLs may now be the most predominant type globally 40 . These discrepancies may be due in part to varied geographic location, different levels of healthcare facilities involved, varied levels of exposure to healthcare settings, antibiotic use and antibiotic stewardship practices.

Conclusions
This study is the first surveillance report on antimicrobial susceptibility testing and molecular detection of resistant genes of P. aeruginosa strains from clinical samples of patients attending healthcare facilities in four district municipalities of Eastern Cape Province, South Africa. The study showed a high rate of MDR P. aeruginosa in our setting. The vast majority of these resistant strains carried bla TEM and bla SHV genes. Early detection and characterization of ESBLs is critical to contain their dissemination, prevent outbreak and optimise therapy. Continuous monitoring of antimicrobial resistance and strict compliance towards infection control practices are the best defence against continuous spread of MDR P. aeruginosa. The limitation of this study is the fact that we could not screen P. aeruginosa isolates for the presence of all reported genes (GES-2 and OXA-21 genes) from other South African provinces due to funding constraints. Specimen collection and analysis. Non-duplicate P. aeruginosa isolates were collected from clinics and hospitals from the four district municipalities. Specimens included throat swabs, wound swabs, swabs from abscesses, sputum, urine, blood culture and catheter tips. Demographic characteristics of patients and medical histories were collected from medical records including date of specimen collection, gender and age. All samples were routinely cultured on MacConkey and Blood agar plates. Blood and sputum were also cultured on chocolate agar. Suspected colonies were plated on Cetrimide agar and identified by gram staining, colony characteristics, motility, pyocyanin production and characteristics grape-like odour 41 . Strains were identified to the species level with Vitek 2 GN (bioMérieux, Inc. USA) ID cards and confirmed by Microscan NID 2 panels (Beckman Coulter, Inc. USA). Specific primers and probes targeting gyrB were amplified by singleplex rPCR and were also used to confirm identity of the isolates. Antimicrobial susceptibility. Antimicrobial susceptibility was obtained by determining MIC using Microscan dehydrated broth microdilution method with negative MIC Panel Type 44 (NM44) (Beckman Coulter, Inc. USA) following the manufacturer's guidelines 42 MICs were interpreted following CLSI guidelines (M100-S27 breakpoints) 43 . The following antibiotics were tested in the panels: amikacin, aztreonam, cefepime, ceftazidime, ciprofloxacin, doripenem, gentamicin, imipenem, levofloxacin, meropenem, piperacillin/tazobactam, piperacillin and tobramycin.

Methods
Criterion for multidrug resistance. The classification of MDR was performed according to Magiorakos et al. 44 . (MDR was defined as non-susceptibility to at least one agent in three or more antimicrobial categories).

Molecular ESBL and MBL detection by singleplex rPCR. Genomic DNA was extracted using Roche
MagNA Pure Bacteria lysis buffer, MagNA Pure Compact Nucleic Acid Isolation kit and PCR grade water (Roche Applied Science, Indianapolis), following manufacturer's instructions. The DNA was used as a template in the rPCR analysis. Real time PCR for bla CTX-M , bla SHV , bla TEM , bla IMP and bla VIM was carried out in the Light Cycler 2.0 instrument (Roche Applied Science, Germany) using Fast start Light Cycler 480 Hybridization probes Master kit (Roche Diagnostics, USA). The choice of testing for these genes was the result of scarcity of data in our setting. Specific primers and probes (Table 3) targeting the genes bla CTX-M , bla SHV , bla TEM , bla IMP and bla VIM were amplified by singleplex rPCR. Primers were designed by TIB-Molbiol (Berlin, Germany). rPCR assay was performed in a 32 carousels using 20 µL capillaries with each capillary containing a total volume of 20 µL including 2 µL of Light Cycler FastStart DNA Master Hybridization Probe, 2 µL of primers (0.5 mM for each forward and reverse), 2.4 µL of MgCl 2 , 2 µL of extracted DNA, and water to make up the volume of 20 µL. DNA amplification was carried out with the following run conditions: Pre-incubation for 5 min at 95 °C, followed by 45 cycles of amplification with denaturation at 95 °C for 30 s, annealing and extension for 1 min at 60 °C, and then a single cycle of cooling for 30 s at 40 °C 11 . Absolute quantification was carried out using the Light Cycler software 4.05.