Incidence and molecular characterization of multidrug resistance in Gram‐negative bacteria of clinical importance from pharmaceutical wastewaters in South‐western Nigeria

Correspondence Avemaria Ifeoma Obasi, Department of Microbiology, University of Lagos, Lagos, Nigeria. Email: avemariaobasi@yahoo.com Abstract Background: The occurrence of antimicrobial resistant bacteria in the environment presents a major threat to public health because it reduces the effectiveness of an‐ timicrobial treatment. Aims: The study was set out to molecularly characterize Gram‐negative bacteria with multidrug resistance and resistance determinants from pharmaceutical wastewaters in Nigeria. Materials and Methods: Susceptibility of the bacterial isolates to 25 antibiotics be‐ longing to 10 categories was tested using the disc diffusion method and Vitek 2. Screening for AmpC, Extended Spectrum Beta‐lactamase and carbapenemase pro‐ duction was done by Polymerase Chain Reaction and sequencing. Results: Ninety‐seven Gram‐negative bacteria, comprising 27 Enterobacteria and 70 nonfermenter bacterial isolates were detected. Antibiotic resistance observed was highest (70.1%) for sulfamethoxazole / trimethoprim and multidrug resistance was revealed in 17 bacterial strains (Klebsiella pneumoniae [7], Enterobacter gergoviae [3], Sphingomonas paucimobilis [1], Empedobacter brevis [1], Chryseobacterium indologenes [1], Pseudomonas aeruginosa [1], Burkholderia cenocepacia [1], Burkholderia cepacia and Stenotrophomonas maltophilia [1]). Extended Spectrum Beta‐lactamase (CTX‐M‐15, SHV‐12, SHV‐2) was positive for 6 K. pneumoniae strains; there were neither AmpC detected nor the production of carbapenamase in all isolates tested. Discussion: The study confirmed the presence of multidrug resistant Gram‐negative bacteria with resistance determinants in wastewaters from pharmaceutical indus‐ tries in Nigeria. Compounds of the wastewater may directly select or co‐select these multidrug resistance strains. Conclusion: The output of drug resistant bacteria into the environment is a potential risk to public health and may facilitate the spread of resistant genes.


| INTRODUC TI ON
Antibiotic-resistant bacteria have emerged due to the selective pressure of antimicrobial use in humans and animals. Water plays an important role in the dissemination of these organisms in the environment (Amaya et al., 2012). This is due to its dynamic nature as it serves as reservoirs, gathering and dispersing bacteria carrying antibiotic-resistance elements; these are mostly encountered in municipal and industrial wastewaters as well as reclaimed waters, thereby posing public health risks. Resistance to antimicrobial agents in clinically relevant Gram-negative bacteria is an increasingly important problem, which in the last few years has spread from the hospital setting to the community (Ghai & Ghai, 2018;Martínez-Martínez & Calvo, 2010). The emergence of antimicrobialresistant bacteria presents a major threat to public health because it reduces the effectiveness of antimicrobial treatment, leading to increased morbidity, mortality, and healthcare expenditure (Lyimo, Buza, Subbiah, Smith, & Call, 2016). Studies have shown the prevalence of antibiotic-resistant pathogens in wastewaters discharged into water bodies (Guardabassi, Petersen, Olsen, & Dalsgaard, 1998;Pruden et al., 2006;Quach-Cu, Herrera-Lynch, Marciniak, & Adams, 2018;Reinthaler et al., 2003), leading to the challenges in antimicrobial therapy. A few authors, in their antibiotic-resistance comparative review using wastewaters from a hospital and a pharmaceutical plant, disclosed that the wastewater from the pharmaceutical plant had increased prevalence of antibiotic resistance among Acinetobacter species (Guardabassi et al., 1998). Recently, a study in Nigeria (Obayiuwana et al., 2018) identified antibioticresistant strains, such as Acinetobacter sp., Klebsiella pneumoniae, Proteus mirabilis, Enterobacter sp., and Bacillus sp. with majority of isolates harboring resistance genes of the sulfonamide-resistant dihydropteroate synthase of Gram-negative bacteria (sul1, sul2) and intl1 genes from pharmaceutical wastewaters. Also Tahrani et al. (2017) identified different genera of bacteria from pharmaceutical wastewaters in Tunisia, and these include Citrobacter, Acinetobacter, Pseudomonas, Delftia, Shewanella, and Rheinheimera; however, they suggested that most of these can be utilized as good candidates for bioremediation, thus useful for biotechnological applications. Although previously isolated genus; Exiguobacterium, Dechloromonas, Haliscomenobacter and Morganella from pharmaceutical wastewaters were also recorded; with some of them having 100% antibiotic resistance to amoxicillin and 6 patterns of multidrug resistance among 11 antibiotics tested, and these could have potential in spreading multidrug resistance as well as the associated public health implications (Tahrani et al., 2015).
It is noteworthy to say that some of the environmental isolates from the wastewaters observed by these authors could have some relevance in bacterial pathogenesis as a result of the dissemination of resistance genes being transferred to their pathogenic counterparts. In a very recent study in press (University of Cambridge, England, 2019), drug-resistant Enterococcus faecium was isolated from a municipal wastewater plant and when a genomic comparison was done with E. faecium from bloodstream isolates of infected patients, results revealed two major lineages with ampicillin-resistant bacteria in clade A1 and A2 and vancomycin-resistant bacteria exclusive to clade A1, displaying a relatedness of environmental isolates to those that cause serious human diseases. There have been reports on the extensive detection of antibiotic-resistance bacteria (ARB) and antibiotic-resistance genes (ARG) in wastewaters (Bouki et al., 2013), highlighting the need for further research studies in this area. Therefore, the aim of the present study was to screen for antibiotic-resistant Gram-negative bacteria of clinical importance in pharmaceutical wastewaters in Nigeria and characterize multidrug-resistant (MDR) strains with resistance determinants.

| Wastewater sampling
Wastewater samples were collected from discharge points from 10 pharmaceutical industries located in South-western Nigeria. The Global Positioning System (GPS) readings were taken for every sample location. The industries are producers of antibiotics, such as ciprofloxacin and cotrimoxazole, analgesic agents, cough syrups, and a few other healthcare products. During and after the manufacturing process, the generated wastewaters are discharged directly (without pretreatment) into local rivers and lakes where human activities take place. The untreated wastewater samples were collected randomly (n = 30) from 10 pharmaceutical industries sampled from different sampling points along the wastewaters discharge channels from each pharmaceutical industry. These were preserved on ice packs contained in a flask and then taken immediately to the laboratory for routine microbiological analysis. Samples were processed immediately or stored at 4°C until use. Visit to each pharmaceutical industry for the collection of samples was done three times during the period of sampling. The sampling period was for 18 months, May 2011 to November 2012.

| Isolation and characterization of Gramnegative bacteria
Bacteria were isolated on nutrient agar plates (Talaro, 2009), and subsequent screening for Gram-negative bacteria was done by the Gram-staining technique (Kola et al., 2012). Further characterization was carried out with the lactose fermentation test, oxidase test, and catalase test (Forbes et al., 2015;John et al., 2009).

| Identification of Gram-negative bacteria
Gram-negative isolates were further identified using an automated method with Vitek 2 (Biomeriux Inc.). The Vitek card GN (Ref.) 21341 was used made up of 60 ingredients including sugars (Kola et al., 2012). In addition, specific PCRs for the identification of K. pneumoniae sensu stricto (Bialek-Davenet et al., 2014)

| Determination of minimum inhibitory concentrations (MIC) and detection of extended spectrum beta-lactamase (ESBL)
Minimum inhibitory concentrations (MICs) are considered the "gold standard" for determining the antimicrobial susceptibilities of bacteria.
Minimum inhibitory concentration is defined as the lowest concentration (mg/L) of an antibiotic that showed 100% inhibition of bacterial growth. confirmation, the proportional reduction of growth in wells containing a cephalosporin combined with clavulanic acid was then compared with that achieved by the cephalosporin alone. The result was interpreted as ESBL-positive or ESBL-negative through a computerized automated expert system (AES). Three control strains (Escherichia coli ATCC 25922, K. pneumoniae ATCC BAA-1705, and Pseudomonas aeruginosa ATCC 27853) were used to validate the measurement. Overnight cultures were introduced into tubes containing 2.5 ml inhalation solution (0.45% sodium chloride) with the aid of an inoculating loop and then further procedure applies as previously described (Kola et al., 2012).

| Disk diffusion method (Kirby Bauer method)
Bacterial isolates were subjected to the disk diffusion method (John et al., 2009)

| Multidrug-resistant (MDR) Gram-negative bacteria (MRGN)
Multidrug resistance was defined as the nonsusceptibility of an isolate to at least 1 agent in ≥3 antimicrobial categories; however, lists of antimicrobial agents were developed for each organism or organism group, as proposed harmonized templates that could be used by clinical, reference, and public health microbiology laboratories (Magiorakos et al., 2012).

| Modified Hodge test (MHT) for the phenotypic detection of carbapenemases
Phenotypic detection of the production of carbapenemase was performed using the modified Hodge test (MHT). Bacterial isolates with resistance to carbapenem (ertapenem, imipenem, and/or meropenem) were tested for carbapenemase production. This method was done by establishing a McFarland standard of 0.5 of bacterial solution of a carbapenem-sensitive E. coli (ATCC 25922) that has already been introduced into 2.5 ml sterile NaCl with a sterile cotton swab, and measuring with the densiChek. Then, 0.5 ml of the suspension was further dissolved in 4.5 ml of NaCl to attain a ratio of 1:10.
The suspension was further seeded on 2 plates of Mueller-Hinton agar. Sterile disks of ertapenem, imipenem, and meropenem (from BD sensidisc) antibiotics were carefully placed each in a triangular pattern on each plate, and this was done 15 min after seeding the plate. One control strain (carbapenemase-producer) was introduced into the seeded plates by making lines 2 mm away from the points where the 3 antibiotic disks were placed, making sure the lines do not intercept one another. The same procedure was carried out for the test strain on the second plate incubating the plates aerobically at 37°C for 18-24 hr. Positive results for carbapenemase producers were shown by the heart shape deformation of the inhibition zone, while negative results retained the usual round shape (Kothari et al., 2013).

| Preparation of DNA
The preparation of DNA for PCR gene amplification was carried out by introducing colonies of cell materials of overnight bacterial culture of ESBL-positive, resistance to third-generation cephalosporins and carbapenem resistance, plated on Luria Bertani (LB) agar in 200 µl of phosphate-buffered saline (PBS). Further procedure applies as previously described (Kola et al., 2012).

| Detection of ESBL encoding genes by PCR using a multiplex primer
The mixture for ESBL-MP-PCR reaction (25 µl) contained: final extension for 4 min at 72°C. Then, 5 µl of each PCR product was visualized by 1.4% (w/v) agarose gel electrophoresis as previously described (Pfeifer et al., 2009). All identified bla TEM genes were sequenced directly from the ESBL-MP-PCR reaction;

µl
bla SHV , bla CTX-M-group1/2 , and bla CTX-M-group9 gene fragments were sequenced after PCR amplification of the whole gene using further primers (Table S1) (Kola et al., 2012;Pfeifer et al., 2009). For the sequencing of PCR-amplified genes, 14 µl of each PCR product was transferred into 15 µl Eppendorf tubes and 1 µl of either the forward or reverse primers were introduced separately into 2 different tubes containing the PCR product as described (Pfeifer et al., 2009). Then, further sequencing procedure was performed by a company (Mwg eurofins, Germany). Analysis of the resulting gene sequences was done with the use of basic local alignment search tool (BLAST; www.nbci.nlm.nih.gov) to identify and confirm the amplified genes.

| Screening for plasmid-mediated AmpC
All isolates with resistance to third-generation cephalosporins and/or carbapenems were screened for presence of plasmid-encoded AmpC genes using previously described PCR primers (Table S1) and resolution on 1.4% (w/v) agarose gel electrophoresis (Kola et al., 2012). PCR conditions were 30 cycles of initial denaturation for 2 min at 95°C, denaturation for 30 s at 95°C, annealing for 60 s at 70°C, extension at 30 s at 72°C, and final extension for 4 min at 72°C.

| Standard laboratory strains for susceptibility test
Bacteria used as standard controls were E. coli (ATCC 25922), K. pneumoniae (ATCC BAA-1705), and P. aeruginosa (ATCC 27853) obtained from the diagnostic unit of the Fredrich Loeffler Institute, University of Greifswald, Germany. All standards were based on the EUCAST (European Committee on Antimicrobial Susceptibility Testing, 2013) and CLSI (Clinical and Laboratory Standards Institute, 2012) for clinical breakpoint in susceptibility testing.

| Statistical analysis
Descriptive Charts from Microsoft Excel package (2010 version) from windows 8.1 operating system were used to analyze the data on susceptibility profile.

| Isolation and characterization of Gramnegative bacteria
Gram-negative bacteria identified from wastewater samples were 97, and cellular morphology revealed that they were all bacilli (rodshaped bacteria).

| Antimicrobial susceptibility testing
The analyses of the profile of resistance to 25 antibiotics were observed in the Gram-negative bacterial isolates. The profile showed that overall percentage resistance ranges from 0% (ertapenem) to 70.1% (sulfamethoxazole/trimethoprim). Susceptibility was low for tigecycline (16.7%) and high for meropenem (94.8%; Table 2).
B. cepacia group (n = 2), S. maltophilia (n = 1), S. paucimobilis (n = 1), and E. brevis (n = 1) isolates were all resistant/intermediate resistant to the carbapenems (imipenem and meropenem). The MIC values for the Vitek 2 breakpoint for resistance in these bacterial isolates range from ≤1 for gentamicin in E. cloacae strains to ≥320 for trimethoprim/sulfamethoxazole for all bacterial isolates; also for susceptibility, the range was from ≤0.2 for Amikacin in E. gergoviae strains to ≥20 trimethoprim/sulfamethoxazole in P. mirabilis strains.

| Detection of ESBL production by Vitek 2
The screening for ESBL by Vitek 2 was positive for six K. pneumoniae isolates located at two geographical points, one from Ogun state and five from Lagos state as detected by the computerized AES.

| MDR Gram-negative bacteria (MRGN)
MRGN was observed in isolates (n = 17) from 6 pharmaceutical industries wastewater samples, at the 2 geographical locations sampled. These include the following: K. pneumoniae (n = 7), E. gergoviae (n = 3), E. brevis (n = 1), S. paucimobilis (n = 1), P. aeruginosa (n = 1), C. indologenes (n = 1), B. cenocepacia, B. cepacia (n = 1), and S. maltophilia (n = 1; Tables 1 and 3). All other isolates were either resistant to one single or more antibiotics of the same or different class, and some could also be referred to as multiple antibiotic resistances (MAR).  (Table 3); however, they were all negative for the production of carbapenemases by MHT, as observed by the inhibition zones that retained the round shape while the control strains indicated positive result (heart-shaped inhibition zones).

| Screening for presence of beta-lactamase encoding genes
The presence of β-lactamase gene in K. pneumoniae strains was demonstrated by PCR amplification with ESBL multiplex primers.
ESBL genes detected in six K. pneuomoniae isolates were bla CTX-M-15 , TA B L E 2 Number and percentage of antimicrobial susceptibility of gram-negative bacterial isolates from pharmaceutical wastewater samples Antimicrobial agent

| Screening for plasmid-mediated AmpC in MDR isolates
Screening for AmpC by resolution of PCR product on gel electrophoresis showed that there was no indication of CITM, ACCM, MOZM, FOZM, and DHAM enzymes when the primer pairs were used.
However, there was a weak signal from two strains of E. gergoviae with the EBCM primer. DNA sequencing reaction and blasting revealed that they were outer protein membranes.

| Associated resistance in ESBL K. pneumoniae
Six K. pneumoniae strains harboring the ESBL genes have associated resistance in other classes of antibiotics. These ESBL-producing strains with resistance to the 3rd-and 4th-generation cephalosporins producing the enzymes TEM-1, SHV-1, SHV-2, SHV-11, SHV-12, SHV-28, and CTX-M-15 were also resistant to the fluoroquinolones (ciprofloxacin, moxifloxacin), the aminoglycosides (gentamicin, tobramycin) as well as the folate pathway inhibitors (trimethoprim/ sulfamethoxazole).  for cefpodoxime (55.3%). These facts were also reported in a study by Chagas et al. (2011) from hospital wastewater in Brazil. They observed that the microorganisms isolated showed higher antibioticresistance rates to amikacin and sulfamethoxazole/trimethoprim among the non-β-lactam antibiotics. Their findings also confirmed that most of the strains were susceptible to carbapenems, and among antibiotics of the β-lactam group, the highest resistance rate was found for cefalotin, a 1st-generation cephalosporin.

| D ISCUSS I ON
Multidrug-resistant Gram-negative bacteria were observed in 17 isolates of the present study (Tables 1 and 2 bacterial isolates from drinking water sampled; another single isolate of this bacterium with CTX-M-10 ESBL was found in a bronchial aspirate from a Spanish hospital during a 12-year period with resistance to piperacillin, cefazolin, cefotaxime, cefuroxime, ceftazidime, cefepime, aztreonam, and nalidixic acid (Canton et al., 2002), which also conforms to that from our study, although ESBL was not detected. To the best of our knowledge, this is the first documented detection of MDR E. gergoviae from pharmaceutical wastewaters in Nigeria.
Sphingomonas paucimobilis is another opportunistic pathogen that rarely causes infection in humans because of its low virulence. A recent study reported by Pratama et al. (2016); and Saeb, David, and Al-Brahim (2014) identified several virulence genes in the genome of S. paucimobilis that have similarity to those of P. aeruginosa. They are usually susceptible to carbapenems, aminoglycosides, trimethoprim/ sulfamethoxazole, and piperacillin/tazobactam and resistant to penicillins and first-generation cephalosporins (Pratama et al., 2016). Its resistance to penicillins and first-generation cephalosporins is due to the production of chromosomally encoded beta-lactamase production (Corkill et al., 1991). Our study detected that this bacterium was resistant to piperacillin, piperacillin/tazobactam, cefotaxime, imipenem, ciprofloxacin, moxifloxacin, and aztreonam, although it has been reported that the susceptibility to third-generation cephalosporins and fluoroquinolones is variable (Al-Halawani et al., 2016;Hsueh et al., 1998). Other studies reported that S. paucimobilis was resistant to amikacin, ceftazidime, and fluoroquinolones (Toh et al., 2011), resistant to cefoxitin and ceftazidime (Özdemir et al., 2011),

Ceph CAZ
Carb IMP

TA B L E 3 (Continued)
and to cefotaxime and amikacin (Cheong et al., 2008). As there are no definitive guidelines for antimicrobial therapy for this bacterial infections, treatment is done with individualized antibiotic therapy according to the in vitro susceptibility profile of clinical isolates (Özdemir et al., 2011;Pratama et al., 2016;Toh et al., 2011). Some studies recommended fluoroquinolones, third-generation cephalosporins, and carbapenems as excellent drugs of choice as initial therapy for S. paucimobilis infection (Hsueh et al.,. 1998;Maragakis et al., 2009;Morrison & Shulman, 1986;Ryan & Adley, 2010), but unfortunately in our study, this bacterium was resistant to these classes of antibiotics recommended, which begs the question; can an individual infected with this bacterium be successfully treated in a case of an emergency? This susceptibility profile of multidrug resistance in this bacterium poses new threat to the antibiotic therapy of this infection; however, Ryan and Adley (2010) suggested that therapy should be adjusted when the susceptibility results are available, or if the patient fails to respond to initial therapy. They also observed that no antibiotic-resistant mechanisms have yet been elucidated for S. paucimobilis.
Empedobacter brevis was previously known as Flavobacterium breve and the pathogenicity of the bacteria is rare and is usually limited to healthcare workers. However, no specific virulence factors have been identified for these species (Sharma et al., 2016).
Although there is a first case report of neonatal bacteraemia and meningitis secondary to E. brevis and the infant responded well to obtained from the study by Araque-Caldeson et al. (2008) with MIC values of the same range. Thus, this close similarity would suggest that environmental B. cepacia complex may be a potential source of infection of immune-compromised or hospitalized patients (Nazik et al., 2018), at least in Nigeria. Burkholderia spp. from pharmaceutical wastewaters in this study was intrinsically resistant to ciprofloxacin which agrees with previous investigations (Walsh & Duffy, 2013), where it was also reported that clinically relevant Burkholderia spp. was intrinsically resistant to ciprofloxacin.
There are little or no data on spread of multidrug resistance S. maltophilia in the community (colonization) and environment in Nigeria. The detection of this potentially pathogenic microorganism with clinically relevant resistance in the wastewater sampled in our study is an indication of potential risk for the microbiological pollution of water resources. In addition, resistance to trimethoprim/ sulfamethoxazole, the drug of choice for treatment, is also burdensome (Ismail et al., 2017;Juhász, Pongrácz, Iván, & Kristóf, 2015).
This is the first confirmation of S. maltophilia isolate with multidrug resistance from a pharmaceutical wastewater in Nigeria.
Chryseobacterium indologenes is a healthcare-associated pathogen (HCAP) and has been widely reported to cause life-threatening infections in patients, especially ones on chronic immunosuppressant drugs (Mehta & Pathak, 2018). Intrinsic resistance to several antibiotic and multidrug resistances to major antibiotics like we observed in our study has also been reported by Mehta and Pathak (2018) in their case report. They emphasized concerns on how this affected the effective empirical treatment of this bacterial infection, more so with the scant availability of data from literature which has been challenging. Detection of C. indologenes in wastewaters from pharmaceutical industries is not adequately documented; this seems to be the first report of multidrug-resistant C. indologenes isolated from our study in Nigeria.
Among the antimicrobials evaluated, the carbapenems (imipenem and meropenem) are the most powerful antibiotics used as therapy for severe infections. However, the resistance mechanism of these bacteria to the carbapenems is not the production of the carbapenamase enzyme but could be due to intrinsic characteristics, mutation, porin loss, or efflux pump resulting to their ability to become resistant to the carbapenems. The resistance determinants in these bacteria from the pharmaceutical wastewaters correspond to that of their closely related clinically relevant species. Alouache et al. Koch Institute Wernigerode, Germany, for their excellent professional and technical assistance. We also thank Adebayo Shittu from Obafemi Awolowo University, Nigeria, for his mentorship and contributions.

CO N FLI C T O F I NTE R E S T
None declared.

AUTH O R CO NTR I B UTI O N
AIO has made major contribution to the research by acquisition and analysis of data, and writing the manuscript; EOU involved in the conception of research and writing the manuscript; and SUN was key to the conception and supervision of research.

DATA AVA I L A B I L I T Y S TAT E M E N T
The raw data underlying the main results of this study will be archived.