Molecular epidemiology of Pseudomonas aeruginosa isolated from lower respiratory tract of ICU patients

May 31, 2021 (With 3 figures) Abstract Lower respiratory tract infections (LRTIs) caused by Pseudomonas aeruginosa are the most common infection among hospitalized patients, associated with increased levels of morbidity, mortality and attributable health care costs. Increased resistant Pseudomonas worldwide has been quite meaningful to patients, especially in intensive care unit (ICUs). Different species of Pseudomonas exhibit different genetic profile and varied drug resistance. The present study determines the molecular epidemiology through DNA fingerprinting method and drug resistance of P. aeruginosa isolated from patients with LTRIs admitted in ICU. A total of 79 P. aeruginosa isolated from patients with LRTIs admitted in ICU were characterized by Restriction Fragment Length Polymorphism (RFLP), Random Amplified Polymorphic DNA (RAPD) and Repetitive Extrapalindromic PCR (REP-PCR). Antibiotic resistance was determined by minimum inhibitory concentration (MIC) assay while MDR genes, viz, blaTEM, blaOXA, blaVIM, blaCTX-M-15 were detected by polymerase chain reaction (PCR). Of the 137 Pseudomonas sp isolated from ICU patients, 57.7% of the isolates were reported to be P. aeruginosa . The overall prevalence of P. aeruginosa among the all included patients was 34.5%. The RAPD analysis yielded 45 different patterns with 72 clusters with 57% to 100% similarity level. The RFLP analysis yielded 8 different patterns with 14 clusters with 76% to 100% similarity level. The REP PCR analysis yielded 37 different patterns with 65 clusters with 56% to 100% similarity level. There was no correlation among the different DNA patterns observed between the three different methods. Predominant of the isolates (46.8%) were resistant to amikacin. Of the 79 isolates, 60.8% were positive for blaTEM gene and 39.2% were positive for blaOXA gene. P. aeruginosa was predominantly isolated from patients with LRTIs admitted in ICU. The difference in the similarity level observed between the three DNA fingerprinting methods indicates that there is high inter-strain variability. The high genetic variability and resistance patterns indicates that we should continuously monitor the trend in the prevalence and antibiotic resistance of P. aeruginosa especially in


Introduction
Lower respiratory tract infections (LRTIs) are the most common infection among patients admitted in intensive care unit (ICU) (Bhatta et al., 2019). Patients admitted in ICUs are highly susceptible to nosocomial infections due to prolonged hospitalization, decreased mobility and exposure to various medications (Sydnor and Perl, 2011). The incidence of nosocomial infections are higher among developing countries when compared to developed countries and it ranges from 2.3% to 49.2% across different clinical settings (Alberti et al., 2002). Pseudomonas sP. is the common cause of nosocomial infections and are highly associated with the LRTIs with increased mortality. The ICU is considered as established endemic settings favoring the growth of Pseudomonas sP. Its ability to colonize in-patients made it as highly critical organisms within ICUs settings (Erbay et al., 2003). Among Pseudomonas sp, P. aeuroginosa is considered as the most predominant species associated with severe LRTIs among patients admitted in ICUs and the second most common pathogen associated with ventilator-associated pneumonia (Weinstein et al., 2005;Sievert et al., 2013). According to the National Nosocomial Infection Surveillance System (NNISS) from China, P. aeruginosa was the predominant pathogen isolated from the lower respiratory tract, accounting for 12. 8%, 12.3% and 13.4% during 2002-2004, 1999-2001 and 2005-2007, respectively (Ding et al., 2016;Wen et al., 2011). A multicenter cross sectional analysis from the United States reported P. aeuroginosa (36.2%) as the most common Gram negative organisms isolated from patients with LRTIs acquired during ICU stay (Claeys et al., 2018).
P. aeruginosa with a high genetic variation is related to several infectious diseases. Due to its genetic diversity, the existing differences in the pathogenicity, and certain genotypes of P. aeruginosa were isolated from certain clinical settings. Understanding the relatedness and distribution of the pathogen is highly important to determine the epidemiology of P. aeruginosa infections and designing appropriate pathogen control measures (Nazik et al., 2007). Phenotypic methods including biotyping, serotyping, hemagglutination, cytotoxin activity assay, bacteriocin typing and immunoblotting were available. However, they lack reproducibility and the fact that P. aeruginosa is phenotypically unstable makes it difficult to rely on these methods for epidemiological studies. Genotypic methods are including pulse field gel electrophoresis, multiple-locus variable number tandem repeat analysis and multilocus sequence typing have high discriminatory power (Maatallah et al., 2013). However, they are require technical expertise, expensive, laborious and time-consuming. The PCR based molecular techniques gained importance in elucidating the epidemiology of P. aeruginosa. The restriction fragment length polymorphism (RFLP), randomly amplified polymorphic DNA (RAPD) and Repetitive Extrapalindromic PCR (REP-PCR) analyses have been developed for P. aeruginosa genotyping (Nanvazadeh et al., 2013;Joseph et al., 2010;Ratkai et al., 2010). These techniques are reported to be less laborious, rapid with high discriminatory power and could be employed in different clinical laboratory settings.
Anti-pseudomonal quinolone or an aminoglycoside along with β-lactams is considered as the primary treatment option for P. aeruginosa infections (Chinese Thoracic Society, 1999;Chinese Society of Critical Care, 2013). However, the treatment of P. aeruginosa infection is challenging due to increasing drug resistance. The increasing prevalence of multidrug resistant (MDR) strains is recognized as a global problem during the treatment of P. aeruginosa infections (Rossi Goncalves et al., 2017). The MDR P. aeruginosa strains are associated with increased morbidity and mortality, prolonging hospital stay and higher costs. The pan-drug resistance, the lack of effective antibiotics against MDR P. aeruginosa strains and continues the dissemination of such strains pose serious challenges in the infection control management (Murugan et al., 2018). Hence, understanding the epidemiology of pathogens and antibiotic resistance patterns of P. aeruginosa is highly warranted. The present study determined the prevalence, genetic heterogeneity and antibiotic resistance patterns of P. aeruginosa isolated from sputum samples of patients admitted with LRTIs in ICU by RFLP, RAPD, REP-PCR, drug resistance by MIC and resistance gene detection by PCR.

Sample collection
A total of 229 samples (sputum-167, broncho alveolar lavage-62) were collected from patients with LRTIs adimitted in the ICU of Ninth Hospital of XI'an, Xi'an, Shaanxi, China between April 2017 and December 2018. From all patients or their legal representatives a written informed consent was obtained before collecting the samples. The hospital review board approval was obtained for the study. Samples were aseptically transported and processed immediately for microbiology analysis.

Isolation and identification of P. aeruginosa
The samples were sub-cultured on to the MacConkey agar (HiMedia, India), incubated at 37 °C for 18-24 hours. Non-fermenting, irregular, green to brown color, catalase and oxidase positive colonies showing typical colony morphology of Pseudomonas sP. were selected. The API  20 NE microbial identification system for the non-fastidious, non-enteric gram negative bacilli (Biomerieux, USA) was used to identify the organisms. After identification, the isolates were sub-cultured on nutrient agar (HiMedia, India) and stored at -40 °C in 20% glycerol containing nutrient broth until used for further analysis.

DNA extraction
DNA extraction was performed by boiling lysis method. Briefly, a loop full of overnight culture was suspended in 200 μL of sterile water and boiled for 10 minutes at 95 °C. Then the suspension was spin for 10 minutes at 14,000 g and supernatant was collected. The DNA present in the supernatant was quantified by spectrophotometry (NanoDrop ND-1000, USA) and stored at -20 °C until used.

Randomly amplified polymorphic DNA
RAPD was performed as described by Nanvazadeh et al. (2013). The PCR reaction mixture contains 12.5 μL of ReadyMix™ Taq PCR Reaction Mix (Sigma-Aldrich, USA), 5 μL of the template, 10 pm of primer 3'-AGCGGGCCAA-5' and molecular grade PCR water to made up to 25 μL. The PCR was performed with the following cycling conditions: initial denaturation for 2 minutes at 96 °C, followed by 3 cycles for 1 minute at 94 °C, 2 minutes at 36 °C, 2 minutes at 72 °C. Then 29 cycles for 1 minute at 94 °C, 1 minute at 58 °C, 1 minute at 72 °C and a final extension for 5 minutes at 72 °C. After PCR, DNA patterns resolved in 1.5% (w/v) agarose gel were analyzed against 1 Kb DNA marker.

Restriction fragment length polymorphism
RFLP was performed as described by Joseph et al. (2010). A 50 μL reaction mixture contains 5 μL of the template, 10 pm of each primer (forward: 5'-GACGGGTGAGTAATGCCTA -3' and reverse: 5'-CACTGGTGTTCCTTCCTATA -3') and 25 μL of ReadyMix™ Taq PCR Reaction Mix (Sigma-Aldrich, USA). The PCR was performed with the following cycling conditions: initial denaturation for 2 min at 95 °C, followed by 30 cycles for 20 sec at 94 °C, 20 sec at 54 °C, 40 sec at 72 °C and a final extension for 1 minute at 72 °C. Restriction digestions was performed with 10 μL of the amplicons (expected size of 618 bp) by 2 U of HinfI (ThermoFisher Scientific, USA) restriction enzymes as per manufacturer's instruction for 16 h at 37 °C. DNA patterns resolved in 1.8% (w/v) agarose gel were analyzed against 100 bp DNA marker.

Repetitive extrapalindromic PCR
REP-PCR was performed as described by Sorkh et al. (2017). A 50 μL reaction mixture contains 5 μL of the template, 20 pm of each primer (forward: 5'-ICGICTTATCIGGCCTAC -3' and reverse: 5' -IIIICGICGICATCIGGC-3') and 25 μL of ReadyMix™ Taq PCR Reaction Mix (Sigma-Aldrich, USA). The PCR was performed with the following cycling conditions: initial denaturation for 2 minutes at 95 °C, followed by 45 cycles for 30 sec at 95 °C, 1 minute at 38 °C, 2 minutes at 72 °C and a final extension for 16 minutes at 72 °C. DNA patterns resolved in 1.8% (w/v) agarose gel were analyzed against 100 bp DNA marker.

Phylogenetic tree analysis
DNA fingerprints obtained for RAPD, RFLP and REP-PCR, respectively, were analyzed separately for phylogenetic tree construction. They were coded to 1 or 0 binary forms, respectively. Using the NTSYSpc 2.11p software (Exeter Software, Setauket, USA) clustering (NJ-UPGMA method) was performed and phylogenetic tree was constructed.

Minimum inhibitory concentration
Minimum inhibitory concentration (MIC) of various antibiotics including gentamicin, amikacin, piperacillin-tazobactam, ciprofloxacin, cefepime, ceftazidime, imipenem and meropenem (Sigma-Aldrich, USA) against P. aeruginosa was determined. Innoculum was prepared by adjusting the overnight culture to 0.5 McFarland's standard. The MIC assay was performed at a concentration of each antibiotics ranging from 0.03 µg/mL to 256 µg/mL by micro broth dilution method; using Muller-Hinton Broth (MHB, HiMedia, India) as described in Clinical Laboratory Standard Institute (CLSI, 2015) guidelines. Briefly, 10 µL of culture was inoculated into MHB containing different concentration of antibiotics and incubated at 37 °C for 24 h. The MIC was determined visually by the highest concentration showing the absence of growth.

Resistant gene detection
Antibiotic resistant genes including bla TEM , bla OXA , bla  , bla VIM were detected through PCR as described earlier (Murugan et al., 2018). PCR was performed using ReadyMix™ Taq PCR Reaction Mix (Sigma-Aldrich, USA). Briefly, 25 μL of ReadyMix™, 5 μL of template DNA, 5pmol of each primer described in Table 1, and molecular grade PCR water to make up to 50 μL was used. Mastercycler  X50s (Eppendorf, Germany) was used for all PCR assays (RAPD, RFLP, REP-PCR). The PCR cycling conditions for resistant gene detection included: initial denaturation for 5 min at 94 °C followed by 35 cycles for 30 sec at 94 °C, 30 sec at 56 °C, 1.5 min at 72 °C and a final extension for 7 min at 72 °C. After PCR, amplicons were resolved in 1.2% agarose gel electrophoresis.

Statistical analysis
Mean and ranges were used to represent the continuous variables; number and percentages were used to represent categorical variable. Statistical significance (P<.05 was considered statistically significant) was determined through Chi-Square and ANOVA test using MINITAB statistical software (MINITAB version 13.1; MINITAB Inc, PA, USA).

Patients and isolates
Of the 229 patients, 145 (63.3%) were male and 84 (36.7%) were female (mean age 43.3±5.3 years). Among these 157 (68.6%) patients had the history of smoking and 82 (35.8%) patients were under mechanical ventilator support. A detailed patient's characteristics are presented in Table 2. Of the 229 sputum samples, a total of 137 (59.8%) samples isolated Pseudomonas sp.by culture method. All Pseudomonas sP. isolates were characterized using API  20 NE microbial identification system (Biomerieux, USA).
Among the Pseudomonas sp., 79 (57.7%) isolates were identified as P. aeruginosa. Of these, 47 (34.3%) isolates were from male patients and 32 (23.4%) were from female patients. The overall prevalence of P. aeruginosa among the all included patients was 34.5% (male 20.5%; female, 14.0%). The presence of P. aeruginosa was significantly higher than other Pseudomonas sp (P = .04). Among the 82 patients who were under mechanical ventilator support, 23 (28%) patients isolated P. aeruginosa.

Discussion
Lower respiratory tract infections are one of the important infectious diseases in developing countries and are often associated with high mortality (Agrawal and Kombade, 2014). Patients who are under ventilator support in the ICU settings are at high risk of having LRTIs with P. aeruginosa (Bobik and Siemiątkowski, 2014). The microbiological diagnosis of lower respiratory infection is often challenging due to possible contamination of the specimen by upper respiratory tract microorganisms (Ahmed et al., 2013). Thus, accurate diagnosis through the appropriate clinical specimen and appropriate techniques plays an important role. In this study, sputum and broncho alveolar lavage samples were aseptically collected and processed immediately. In our study, 57.7% of the Pseudomonas strains were identified as P. aeruginosa accounting up to 34.5% of the overall samples collected. Our study reported a higher (34.5%) prevalence rate of P. aeruginosa than that reported from Nepal (13.9%), Pakistan (24%) and India (26.8%) in patients with LRTIs (Bhatta et al., 2019;Fatima et al., 2012;Ahmed et al., 2013). The NNISS from China reported lower prevalence of P. aeruginosa ranging from 12. 3-13.4% during 1999-2007 in patients with LRTIs (Ding et al., 2016;Wen et al., 2011). A multicenter cross sectional analysis from the United States reported that 36.2% of the Gram negative organisms isolated from patients with LRTIs admitted in ICU were identified as P. aeruginosa, which is comparable (34.5%) to that reported in our study (Claeys et al., 2018). While a study from Brazil, when compared to our results, reported a higher percentage (51.9%) of P. aeruginosa isolates from ICU patients (Matos et al., 2016). Although the prevalence of P. aeruginosa varies in a different region, our study reported a higher rate compared to other studies from Asian countries including China. Thus, there is a need to continuously monitor the prevalence of P. aeruginosa especially in patients admitted in ICU with LRTIs in this region. Among the 79 patients who were isolated with P. aeruginosa, predominant (52, 65.8%) of them were from male patients. Similarly, a study from Pakistan reported that predominant (70.8%) of the P. aeruginosa were isolated from male patients with LRTIs; the reported rate was slightly higher than our report (Fatima et al., 2012). In our study, P. aeruginosa infection was reported to be significantly higher in male patients (P<.05). Our result indicates that P. aeruginosa isolated from patients admitted in ICU with lower respiratory infection showed varied prevalence rates and have epidemiological diversity, compared to other studies from different geographies.
Genetic complexity and the presence of large genome allows P. aeruginosa to thrive in different ecological conditions. Genomic heterogeneity forms the basis of several molecular techniques and is considered a valuable tool in epidemiological studies. Although several molecular techniques are available, PCR based molecular techniques, which are less laborious and cost-effective has gained importance in elucidating the epidemiology of P. aeruginosa. The RAPD, RFLP and REP-PCR are widely used in several epidemiological studies of P. aeruginosa (Nanvazadeh et al., 2013;Joseph et al., 2010;Ratkai et al.,  2010). The RAPD analysis in our study yielded 45 different patterns, RFLP yielded 8 different patterns and REP-PCR yielded 37 different patterns. A study from Iran, which included 50 P. aeruginosa from burn patients for RAPD, reported only 9 different genotypes which are lower than our report (Nanvazadeh et al., 2013). A study from Czech Republic which included 47 P. aeruginosa isolated from various clinical samples of patients admitted in ICU reported 22 different genotypes when analyzed by RAPD (Vaněrková et al., 2017). A study from Switzerland which included 55 (54-clinical, 4-environmental) P. aeruginosa isolated from burnt unit reported 8, 12 and 12 different patterns when digested the amplicons with BamHI, EcoRI and PstI, respectively (Blanc et al., 1993). A study from Iran, which analyzes 75 P. aeruginosa isolated from burn patients by REP-PCR reported 40 different REP types, a comparable number of REP-PCR pattern was obtained from our study (Sorkh et al., 2017).Although we did not include different isolates from the same patients, we observed that 20.3% of our isolates showed 100% intra-strain similarity with 8 different clusters when analyzed by RAPD. The RFLP analysis revealed 13 different cluster with 100% intra-strain similarity. While in REP-PCR 22.8% of the isolates 100% intra-strain similarity with 9 different cluster. We found that 12 isolates had common patterns across three different techniques. Overall, all three methods revealed high heterogeneity among our isolates. However, RAPD and REP-PCR had good discriminatory power and yielded different patterns with high strain variability when compared to RFLP. Apart from the isolates, which showed 100% intra-strain similarity majority of our stains, yielded several different DNA fingerprints and showed high genetic variability. The genetic variation among our isolates suggests continues monitoring of genotypes with a geographical specific approach is required for effective identification of different types of P. aeruginosa strains. The treatment of P. aeruginosa infection is often challenging due to its inherent resistance to a broad range of antibiotics, its ability to develop MDR and acquired resistance through chromosomal mutations (Gales et al., 2001;Zahraa et al., 2014). In our study, 46.8% (MIC range 0.06-128 µg/mL) of our isolates were found to be resistance to amikacin while only 43% (MIC range 0.5-64 µg/mL) of isolates were found to be resistance to gentamicin. In contrast, a meta-analysis from China reported amikacin resistance in 22.5% of P. aeruginosa isolated from patients with VAP, which is lower than our result (Ding et al., 2016). While the study reported higher (51.1%) gentamicin resistance (Ding et al., 2016). A study from Nepal that included 19 P. aeruginosa isolated from patients admitted in ICU with LRTIs reported higher resistance to cefepime (68.4%), ciprofloxacin (68.4%), amikacin (52.6%), gentamicin (52.6%) and lower resistance to imipenem (10.5%) when compared to our results (Bhatta et al., 2019). Similarly, a study from Brazil reported higher rates of antibiotic resistance for meropenem (64.8%), imipenem (64.8%), ceftazidime (63%), cefepime (48.1%) and piperacillin/tazobactam (46.3%) (Matos et al., 2016). While a study from Pakistan when compared to our results, reported lower resistance to amikacin (35%), imipenem (24%) and higher resistance to cefepime (40%) among P. aeruginosa isolated from patients with LRTIs (Fatima et al., 2012). A study from India that included P. aeruginosa isolates from patients admitted in ICU with LRTIs reported varied results; the resistance rate of ceftazidime (30.4%) and ciprofloxacin (34.8%) corroborates with that reported in our study (Agrawal and Kombade, 2014). While resistance to imipenem (86.9%), cefepime (34.8%) was higher and gentamicin (8.7%) was lower than our result (Agrawal and Kombade, 2014). A study from China that included P. aeruginosa from patients with VAP in intensive care units reported that 50% of their isolates showed resistance to ceftazidime and all (100%) the isolates showed resistance to ciprofloxacin, which is much higher than our result (Xiao et al., 2013). In our study, the highest antibiotics resistance rate was observed for amikacin (46.8%) while the lowest was observed for piperacillin-tazobactam (3.8%).
Due to the stability against ESBLs, carbapenems (imipenem and meropenem) are considered as the most potent agents for MDR P. aeruginosa (Zavascki et al., 2010). In our study, meropenem and imipenem resistance were comparable, where 26.5% of our isolates were resistant to imipenem and 29.1% of isolates were resistant to meropenem. Our result on meropenem resistance (29.1%) was slightly higher (25%) than that reported from India (Murugan et al., 2018). A study from Brazil reported much higher rates of resistance (64.8%) for meropenem and imipenem among P. aeruginosa isolated from ICU patients (Matos et al., 2016). Majority of our isolates found to be resistant to amikacin and gentamicin, however, there was no significant difference in resistance rates was found among the isolates (P = 1.00). In our study, 24 (30.4%) of isolates were found to be MDR strains which were lower than that reported from Brazil (37%) and Asia (42.8%) (Chung et al., 2011;Matos et al., 2016). Of the 24 MDR isolates 9 (11.4%) were isolated from patients who were under mechanical ventilator support. All our MDR isolates were found to be resistance to gentamicin which is in agreement with the findings of a study from Qatar (AbdulWahab et al., 2017).
The family of genes, TEM, SHV, CTX-M, PER, VEB, GES, and IBC encoding the ESBL enzymes was prevalent among P. aeruginosa (Rawat and Nair, 2010). Among several ESBL genes present, bla TEM gene was predominantly carried by P. aeruginosa that confers resistance towards penicillin group of antibiotics (Dumas et al., 2006). In our study bla TEM was predominant gene (60.8%) present among our isolates. Our findings are in agreement with a study from India, which reported that bla TEM was predominant gene present among the P. aeruginosa isolate, however, with higher rates (72.5%) to that our result (Murugan et al., 2018). The study also reported that the presence of bla OXA (33.5%), bla VIM (11.5%) and bla CTX-M gene (5%) genes were lower than our result (Murugan et al., 2018). In our study 26.6% of isolates were positive for bla OXA gene, in contrast a study from Iraq reported that 93% of their isolates were positive of bla OXA gene which was much higher than out report (Zahraa et al., 2014). The bla OXA gene is considered as a naturally occurring gene, the high prevalence indicate a potential horizontal transfer in which Class D β-lactamases might get introduced by other co-inhabiting bacterial species (Girlich et al., 2004;Kong et al., 2005). We report that 15.2% of our isolates possessed bla VIM gene lower (1.6%) or absence of bla VIM gene among their P. aeruginosa isolates (Fallah et al., 2013;Moosavian and Rahimzadeh, 2015;Amini and Mobasseri, 2017). When compared to the antibiotic resistance and the presence of resistance genes with the 3 different molecular typing, we did not find any specific patterns or genotypes associated with antibiotic resistance or resistance genes.

Conclusion
P. aeruginosa was the predominant isolate with an overall prevalence of 34.5% isolated from patients (male 20.5%; female, 14.0%) with LRTIs admitted in ICU. Majority of the isolates were resistant to amikacin and bla TEM was the predominant gene detected among our isolates. The difference in the similarity level observed between the three DNA fingerprinting methods indicates that there is high inter-strain variability. The differences in resistance patterns emphasize the need for personalized antibiotic treatment. The high genetic variability and resistance patterns indicate that we should continuously monitor the trend in the prevalence and antibiotic resistance of P. aeruginosa especially in patients with LRTIs admitted in ICU.