Rectal and Tracheal Carriage of Carbapenemase Genes and Class 1 and 2 Integrons in Patients in Neurosurgery Intensive Care Unit

The spread of multidrug-resistant Gram-negative bacteria, which is associated with the distribution of beta-lactamase genes and class 1 and 2 integrons, is a global problem. In this study, in the Moscow neurosurgery intensive care unit (neuro-ICU), the high prevalence of the above-stated genes was found to be associated with intestinal and tracheal carriage. Seven-point prevalence surveys, which included 60 patients in the neuro-ICU, were conducted weekly in the period from Oct. to Nov. 2019. A total of 293 clinical samples were analyzed, including 146 rectal and 147 tracheal swabs; 344 Gram-negative bacteria isolates were collected. Beta-lactamase genes (n = 837) were detected in the isolates, including beta-lactamase blaTEM (n = 162), blaSHV (n = 145), cephalosporinase blaCTX–M (n = 228), carbapenemase blaNDM (n = 44), blaKPC (n = 25), blaOXA–48 (n = 126), blaOXA–51–like (n = 54), blaOXA–40-like (n = 43), blaOXA–23-like (n = 8), and blaVIM (n = 2), as well as class 1 (n = 189) and class 2 (n = 12) integrons. One extensively drug-resistant Klebsiella pneumoniae strain (sequence type ST39 and capsular type K23), simultaneously carried beta-lactamase genes, blaSHV–40 and blaTEM–1B, three carbapenemase genes, blaNDM, blaKPC, and blaOXA–48, the cephalosporinase gene blaCTX–M, and two class 1 integrons. Before this study, such heavily armed strains have not been reported, suggesting the ongoing evolution of antibiotic resistance.


Introduction
Antimicrobial resistance is one of the most serious threats to global health care [1,2]. In the last two decades, an increased number of infections caused by multidrug-resistant Gram-negative bacteria (MDR-GNB) have been reported [3,4]. As a result, increases in morbidity and mortality rates have been observed, as well as a rise in health care costs [5]. Carbapenems were the most effective antibiotic therapy for infections caused by MDR-GNB until the increased prevalence of carbapenem resistance was described in clinical settings [6]. The widespread dissemination of carbapenem-resistant Gram-negative bacteria (CR-GNB) 2 patients in 5, 1 patient in 6, and 5 patients in 7 surveys. The median age of the patients was 53.8 years (ranging from 14 to 73 years old), and the sex ratio (male/female) of the patients was 1.07 (31/29). Most patients (n = 28) received carbapenem therapy before/during the surveys, few patients (n = 15) received cephalosporins with/no β-lactamase inhibitors followed by the patients, and the remaining patients (n = 17) were not treated with betalactams. Half of the patients (n = 30) had symptoms of gastrointestinal dysfunction (n = 13) and infection of the respiratory system (n = 21), and both symptoms were identified in four patients. Six patients died during this period; the mortality rate was estimated to be 10%. A total of 34 patients were discharged from the neuro-ICU before the end of the study (Figure 1).
Fourteen patients carried carbapenemase genes throughout the study period. Three patients did not have carbapenemase genes at the beginning of the study, but later, they became positive for bla OXA-48 +bla NDM +bla KPC , bla OXA-48 +bla VIM , and bla VIM +bla NDM genes, respectively. In contrast, three patients lost carbapenemase genes during their time in the ICU, and they were not detected at the end of the study. Interestingly, eleven clinical samples of seven patients simultaneously carried three carbapenemase genes, bla NDM , bla KPC , and bla OXA-48 , the cephalosporinase gene, bla CTX-M , and class 1 integrons. A combination of CRGs were also detected: bla NDM +bla KPC was detected in two patients; bla KPC +bla OXA-48 in two patients; bla VIM +bla OXA-48 in one patient; and bla VIM +bla NDM in one patient. The CRG combination bla OXA-48 +bla NDM +bla KPC was detected in seven patients. The bla OXA-48 +bla NDM +bla KPC CRGs were detected in six patients. Thus, 33 (11%) of the 293 samples were positive for two or more CRGs (Table 2).  In total, 65 (22%) of the 293 clinical samples were positive for CRGs. The blaOXA-48 gene was detected in 18 (53%) of the 34 CRG-positive r.s. and in 19 (61%) of 31 positive t.s.; the blaNDM gene was detected in 22 (64%) rectal and 17 (55%) tracheal samples; the blaKPC gene was detected in 17 (50%) rectal and 15 (48%) tracheal samples; and the blaVIM gene was detected in 4 (12%) rectal and 6 (19%) tracheal samples. It should be noted that the number of CRGs in the r.s. was two times higher compared to the number in the t.s. The asymptomatic rectal carriage of CRGs was estimated to be present in 17% of the patients, and the asymptomatic tracheal carriage of CRGs was estimated to be present in 18% of the patients ( Figure 2).r. In total, 65 (22%) of the 293 clinical samples were positive for CRGs. The bla OXA-48 gene was detected in 18 (53%) of the 34 CRG-positive r.s. and in 19 (61%) of 31 positive t.s.; the bla NDM gene was detected in 22 (64%) rectal and 17 (55%) tracheal samples; the bla KPC gene was detected in 17 (50%) rectal and 15 (48%) tracheal samples; and the bla VIM gene was detected in 4 (12%) rectal and 6 (19%) tracheal samples. It should be noted that the number of CRGs in the r.s. was two times higher compared to the number in the t.s. The asymptomatic rectal carriage of CRGs was estimated to be present in 17% of the patients, and the asymptomatic tracheal carriage of CRGs was estimated to be present in 18% of the patients (Figure 2).r. Table 2. Trends in the content of resistance genes in patients.

Bacterial Isolates and Asymptomatic Carriage
A total of 344 Gram-negative bacterial isolates were collected from the patients. Overall, 183 isolates were collected from the rectal swabs (r.s.), and 161 isolates were collected from the tracheal swabs (t.s.). Klebsiella pneumoniae was the most prevalent bacterium and other species, B. gladioli, Citrobacter spp., Enterobacter spp., H. alvei, K. aerogenes, K. oxytoca, K. variicola, M. morganii, P. mirabilis, P. stuartii, S. maltophilia, and S. marcescens (12%). The isolates obtained from the patients with gastrointestinal dysfunction and/or respiratory infection were identified as K. pneumoniae (38%), A. baumannii (21%), E. coli (16%), and P. aeruginosa (8%) and other species, Citrobacter spp., Enterobacter spp., M. morganii, P. mirabilis, and S. maltophilia (17%). Interestingly, among the isolates obtained from the patients without clinical manifestation, A. baumannii was found twice as often in the tracheal swabs compared with the rectal swabs. Moreover, among isolates obtained from the patients without clinical manifestation, E. coli was more prevalent in the rectal swabs compared with tracheal swabs ( Figure 4). All of the bacterial isolates were collected from 55 patients, while no GNB organisms were found in samples from 5 patients (namely 14, 16, 27, 40, and 60). Bacterial isolates were only collected from the r.s. in six patients (namely 5, 17, 24, 25, 30, and 56), and they were only collected from the t.s. in three patients (namely 33, 51, and 54). The number of isolates collected from one patient varied significantly: 1-10 isolates were obtained from 46 patients, 11-20 isolates were obtained from 6 patients, and 21-30 isolates were obtained from 3 patients.
Interestingly, among the isolates obtained from the patients without clinical manifestation, A. baumannii was found twice as often in the tracheal swabs compared with the rectal swabs. Moreover, among isolates obtained from the patients without clinical manifestation, E. coli was more prevalent in the rectal swabs compared with tracheal swabs (Figure 4).

Resistomes of P. aeruginosa Clinical Isolates
P. aeruginosa isolates were obtained from 14 (23%) patients, including four patients (namely 1, 2, 3, and 32) who simultaneously carried P. aeruginosa in both r.s. and t.s. MDR phenotypes were identified for 100% of P. aeruginosa isolates. A total of five beta-lactamase genes, including three bla CTX-M genes and two bla VIM carbapenemase genes, and 32 class 1 integrons were detected in the P. aeruginosa genomes. The gene combination of bla CTX-M+ int1 was detected in three isolates collected from three patients (namely 2, 3, and 20); bla VIM +int1 was detected in two isolates collected from one patient (namely 34) (Figure 9). Antibiotics 2022, 11, x FOR PEER REVIEW 11 of 19 detected in two isolates from two patients (namely 1 and 46), and blaCTX-M+blaOXA-23-like+blaOXA-51-like was detected in one isolate from one patient (namely 1) (Figure 8).

Resistomes of P. aeruginosa Clinical Isolates
P. aeruginosa isolates were obtained from 14 (23%) patients, including four patients (namely 1, 2, 3, and 32) who simultaneously carried P. aeruginosa in both r.s. and t.s. MDR phenotypes were identified for 100% of P. aeruginosa isolates. A total of five beta-lactamase genes, including three blaCTX-M genes and two blaVIM carbapenemase genes, and 32 class 1 integrons were detected in the P. aeruginosa genomes. The gene combination of blaCTX-M+int1 was detected in three isolates collected from three patients (namely 2, 3, and 20); blaVIM+int1 was detected in two isolates collected from one patient (namely 34) (Figure 9).
In our study, the rate of class 1 and 2 integrons' carriage in MDR-GNB isolated from patients in the neuro-ICU was 76%. These data were somewhat different from reports from France and Uganda (53 and 81%, respectively) [16,48]. Class 1 integrons were more prevalent (n = 248) than class 2 integrons (n = 14) in the isolates in this study. Similar proportions were described in previously published studies from Spain, Iran, and Tunisia [48][49][50]. In this study, the rate of class 1 integrons among CR isolates was 73%, while in reports from Iran, this index was higher (86-95%) [51,52]. In our study, class 1 integrons were detected in 76% of MDR P. aeruginosa, 70% of MDR K. pneumoniae, 68% of MDR A. baumannii, and 34% MDR of E. coli. These data are consistent with the findings of reports from Iran and Australia [52][53][54].
In total, the asymptomatic rectal and tracheal carriage of CRGs was found in 17 and 18% of patients, respectively; the rectal and tracheal carriage of class 1 and 2 integrons was found in 23 and 15% of patients, respectively. Such patients may be considered as a potential source of AMR gene transmission in the ICU. Moreover, according to some reports, colonization with MDR potential pathogens and CRGs' carriage may be a prerequisite for the development of nosocomial infections [34].
Thus, this study highlighted the asymptomatic carriage of carbapenemase genes and the prevalence of potential nosocomial pathogens in the intestine and in the trachea of patients in the neuro-ICU. This is important for clinicians, because it will help them to improve the strategies used for hospital infection control and choose optimal antimicrobial therapies. The novelty of this study is the description of CR-GNB strains that simultaneously carry three carbapenemase genes, bla OXA-48 +bla NDM +bla KPC . The limitation of our study was that its single-center nature meant it was impossible to generalize the results. Although the study provided information about the patients regarding prior antibiotic therapy and their length of hospitalization before the study, it did not reveal the importance of these aspects in explaining the reasons for the persistence of resistance genes.

Bioethical Requirements
In this study, we anonymized the data of patients in the ICU. According to the Requirements of the Russian Federation Bioethical Committee, each patient signed informed consent to treatment and laboratory examination. The study was a retrospective observational study. The Burdenko National Medical Research Center of Neurosurgery Review Board approved the study and granted a consent waiver status. Approval Code: #11/2018. Approval Date: 1 November 2018.

Study Design
Seven point-prevalence surveys were conducted weekly among all of the neuro-ICU patients in the period from 18 October 2019 to 29 November 2019. Two clinical samples (rectal and tracheal swabs) were collected from each patient on each survey date. The presence of antimicrobial resistance genes was detected in every clinical sample using realtime PCR (RT-PCR), Gram-negative bacteria were isolated from the samples (Figure 10).

DNA Extraction and AMR Gene Detection in Clinical Samples
Total DNA from clinical samples was extracted using an AmpliPrime DNA-sorb-B reagent kit (InterLabService, Moscow, Russia) in accordance with the manual from the manufacturer. Real-time PCR with specific primers was performed with the obtained DNA preparations to detect beta-lactamase genes-bla TEM , bla SHV , bla CTX-M , bla OXA-48 , bla NDM , bla VIM , and bla KPC -and class 1 and 2 integrons, as described previously [55].

Isolation and Identification of Gram-Negative Bacteria
Gram-negative bacteria were collected from clinical samples through growth in Luria-Bertani (LB) broth (Difco, Sparks, MD, USA) at 37 • C for 18 h and the isolation of single bacterial colonies on Lactose TTC agar with Tergitol-7 (SRCAMB, Obolensk, Russia) and 50 mg/L ampicillin (Thermo Fisher Scientific, Waltham, MA, USA). Bacterial identification was conducted using a MALDI-TOF Biotyper instrument (Bruker, Karlsruhe, Germany). The following ATCC reference strains were used as controls: Escherichia coli ATCC 25922, Pseudomonas aeruginosa ATCC 27853, and Klebsiella pneumoniae ATCC 700603. Bacterial isolates were stored in 15% glycerol at −80 • C.

Susceptibility to Antimicrobials
The minimal inhibitory concentrations (MICs) of antimicrobials belonging to six functional groups: beta-lactams (ampicillin, amoxicillin/clavulanic acid, cefotaxime, ceftazidime, and meropenem), tetracyclines (tigecycline), fluoroquinolones (ciprofloxacin), phenicols (chloramphenicol), aminoglycosides (gentamicin), and sulfonamides (trimethoprimsulfamethoxazole) were determined via the dilution in agar method using Mueller-Hinton agar (Merck, Darmstadt, Germany) according to the Clinical and Laboratory Standards In-stitute (CLSI) guidelines, performance standards for antimicrobial susceptibility testing [56]. The results were interpreted according to the European Committee on Antimicrobial Susceptibility Testing. Break-point tables were used for the interpretation of MICs and zone diameters, Version 12.0, 2022 (http://www.eucast.org, accessed on 8 June 2022). Escherichia coli strains ATCC 25922 and ATCC 35218 were used for quality control. The criterion for defining multi-drug resistant (MDR) isolates was non-susceptibility to ≥1 agent in ≥3 antimicrobial categories; extensively drug-resistant (XDR) isolates were non-susceptible to ≥1 agent in all but ≤2 categories [32]. real-time PCR (RT-PCR), Gram-negative bacteria were isolated from the samples ( Figure  10).

DNA Extraction and AMR Gene Detection in Clinical Samples
Total DNA from clinical samples was extracted using an AmpliPrime DNA-sorb-B reagent kit (InterLabService, Moscow, Russia) in accordance with the manual from the manufacturer. Real-time PCR with specific primers was performed with the obtained DNA preparations to detect beta-lactamase genes-blaTEM, blaSHV, blaCTX-M, blaOXA-48, blaNDM, blaVIM, and blaKPC-and class 1 and 2 integrons, as described previously [55].

Isolation and Identification of Gram-Negative Bacteria
Gram-negative bacteria were collected from clinical samples through growth in Luria-Bertani (LB) broth (Difco, Sparks, MD, USA) at 37 °C for 18 h and the isolation of single bacterial colonies on Lactose TTC agar with Tergitol-7 (SRCAMB, Obolensk, Russia) and 50 mg/L ampicillin (Thermo Fisher Scientific, Waltham, MA, USA). Bacterial identification The prevalence of bacterial species and antimicrobial resistance phenotypes was calculated using Microsoft Excel v. 1909.

Whole-Genome Sequencing
The whole-genome sequencing of the isolates was performed using a Nextera DNA Library Preparation Kit and MiSeq Reagent Kit v3 (300 cycles) on an Illumina MiSeq platform. Reads without quality filtering were de novo assembled using Unicycler v 0.4.7 [65] with default parameters. The annotation was performed using the NCBI Prokaryotic Genome Annotation Pipeline [66]. Multilocus sequence typing (MLST) and the identification of antibiotic resistance genes, virulence genes, plasmids, and restriction-modification systems were conducted using the web resource of the Center for Genomic Epidemiology (http://www.genomicepidemiology.org/, accessed on 22 April 2022) and the BIGSDB database (https://bigsdb.pasteur.fr/klebsiella/klebsiella.html, accessed on 22 April 2022).

Conclusions
The current study focused on the carriage of MDR and ESKAPE Gram-negative bacteria and antimicrobial resistance genes in patients in the neurosurgery-ICU in Moscow in October-November 2019. In total, 55 out of 60 patients harbored significant antimicrobial resistance mechanisms in the neuro-ICU, which implies that infection control measures (contact precautions, etc.) are of critical importance to avoid the spread of resistance. The inappropriate use of antimicrobials should be reduced through antimicrobial stewardship interventions to reduce the effects of antimicrobial resistance gene selection in patients. It was shown that the asymptomatic rectal and tracheal carriage of carbapenem resistance genes was estimated to occur in 17 and 18% of patients, respectively. These patients were a potential source of the transmission of these genes in the ICU. The obtained data indicate the importance of monitoring the asymptomatic carriage of antimicrobial resistance genes, especially carbapenem resistance genes and integrons, and preventing the transmission of such genes within ICUs.