Antibiotic resistance and typing of the methicillin-resistant Staphylococcus aureus clones in Kuwait hospitals, 2016–2017

Background Methicillin-resistant Staphylococcus aureus (MRSA) belong to diverse genetic backgrounds that differ in antibiotic resistance. Knowledge of the local clonal composition of MRSA strains is important for patients’ management and for designing effective control and eradication methods. The aim of this study was to compare the antibiotic resistance patterns and genotypic characteristics of MRSA isolates obtained in public hospitals in Kuwait in 2016 and 2017 for changes in their resistance patterns and clonal composition. Methods A total of 4726 MRSA isolates obtained in 2016–2017 from clinical specimens in Kuwait public hospitals were characterized using antibiogram, SCCmec typing, spa typing and DNA microarray. Results The isolates expressed resistance to fusidic acid (52.9%), kanamycin (41.6%), gentamicin (32.5%) and erythromycin (36.2%). The prevalence of high-level mupirocin resistance decreased from 3.7% in 2016 to 2.4% in 2017, while the proportion of resistance to other antibiotics remained relatively stable. A total of 382 spa types were detected with eight spa types, t688 (N = 547), t304 (N = 428), t860 (N = 394), t127 (N = 306), t044 (N = 230), t311 (N = 243), t223 (N = 184) and t002 (N = 181) constituting 53.1% of the MRSA isolates in 2016–2017. Of the 3004 MRSA isolates obtained in 2016 (N = 1327) and 2017 (N = 1677) selected for DNA microarray analysis, 26 clonal complexes (CCs) were identified. Most of the isolates belonged to CC1 (N = 248), CC5 (N = 833), CC6 (N = 241), CC8 (N = 292), CC22 (N = 421), CC30 (N = 177), CC80 (N = 177) and CC97 (N = 171). The prevalence of CC5 isolates has significantly (p ≤ 0.05) increased from 294 isolates in 2016 to 539 isolates in 2017. Although CC22 increased from 196 isolates in 2016 to 225 isolates in 2017, CC1 increased from 112 isolates in 2016 to 136 isolates in 2017, CC6 increased from 103 isolates in 2016 to 138 isolates in 2017, these changes were not significant (p ≥ 0.05). Conclusion These results revealed the diversity in the genetic backgrounds of MRSA isolates and the stable maintenance of the dominant MRSA clones in Kuwait hospitals in 2016 and 2017 suggesting an on-going transmission of these clones. Novel and creative infection prevention and control measures are required to curtail further transmission. Supplementary information Supplementary information accompanies this paper at 10.1186/s12866-020-02009-w.


Background
Since its report in 1961 in England [1], methicillin-resistant Staphylococcus aureus (MRSA) has spread to many countries causing serious infections that are sometimes difficult to treat [2]. Initially, MRSA was a well-established pathogen among elderly patients with previous admission to healthcare settings and with history of previous antibiotic usage. This type of MRSA was known as healthcare-associated or healthcare-acquired MRSA (HA-MRSA) [3]. Then in the 1990's, a new lineage of MRSA emerged in people with no previous history of hospitalization or exposure to the healthcare system and previous antibiotic use which was designated as community-associated or communityacquired MRSA (CA-MRSA) [4].
Molecular epidemiological typing techniques, such as SCCmec typing, staphylococcal protein A (spa) typing, multilocus sequence typing (MLST), pulsed-filed gel electrophoresis (PFGE) and DNA microarray have been used to study the genetic background of MRSA. HA-MRSA isolates carry relatively large SCCmec genetic element belonging to type I, II, or III, and are usually resistant to multiple non-beta-lactam antibiotics [2]. In contrast, CA-MRSA isolates carry smaller sized SCCmec elements belonging to SCCmec type IV, V or VI and are usually sensitive to most non-beta-lactam antibiotics [2]. MRSA isolates carrying different SCCmec genetic elements have been further differentiated using MLST and eBURST which grouped them into sequence types and clonal complexes (CCs) [2].
Studies have shown that CA-MRSA have replaced HA-MRSA in many countries including Kuwait [5], Singapore [6], United Arab Emirates [7] and Portugal [8]. A previous study in Kuwait showed that the clonal composition of the MRSA has changed significantly from 1992 to 2010 [5] with the emergence of different CA-MRSA clones. The same report revealed that CC8/ ST239-III remained the most common clone in Kuwait hospitals from 1992 to 2010, although their prevalence decreased overtime [5]. During the same period, the prevalence of CA-MRSA clones including CC5/ST5-IV/ V, CC80/ST80-IV and ST1-IV/V were increasing [5]. To provide an update on the clonal composition of MRSA strains circulating in Kuwait hospitals, this study compared the antibiotic resistance patterns and genotypic characteristics of MRSA isolates obtained in public hospitals in Kuwait in 2016 and 2017 for changes in clonal composition and resistance profile.
The distribution of resistance phenotypes for all MRSA isolates obtained in 2016 and 2017 is shown in Table 1. All isolates were sensitive to vancomycin (MIC: ≤ 2 μg/ml), teicoplanin (MIC: ≤ 2 μg/ml) and linezolid. Besides betalactam resistance, most of the MRSA isolates obtained in Comparison of the distribution of antibiotic resistance in MRSA isolates obtained between 2016 and 2017 is shown in Table 1. During 2016-2017, there were significant (p ≤ 0.05) decrease in the prevalence of resistance to erythromycin, clindamycin and high-level mupirocin resistance, while resistance to chloramphenicol increased significantly ( Table 1).
Results of DNA microarray analysis revealed that the high-level mupirocin-resistant isolates were positive for mupA that encodes an alternative isoleucyl-tRNA synthetase (ileS2) which is unaffected by mupirocin. Similarly, resistance to gentamicin, erythromycin and clindamycin, and tetracycline corresponded with the presence of their respective determinants, aacA-aphD, erm(A)/erm(C), and tet(K)/tet(M). Fusidic acid resistance was mediated by fusC in most of the isolates and by fusB/faR1 in CC80 isolates.
There were no significant changes in the distribution of SCCmec types I, II, III and V in MRSA isolates between 2016 and 2017 (Table 2). MRSA isolates carrying type VI increased (p ≤ 0.05) from 7% in 2016 to 11.2% in 2017, while those carrying SCCmec type IV decreased (p ≤ 0.05) from 50% in 2016 to 44.5% in 2017 ( Table 2).

Prevalence of spa types among MRSA isolates in 2016-2017
In total, 382 spa types were identified among the MRSA isolates obtained in 2016-2017. The distribution of the common spa types among the isolates obtained in 2016 and 2017 is shown in Table 3. Spa type t688 (23%) was the dominant spa type detected in both years. This was followed by t304, t860, t127, t044, t311, t002 and t223 (Table 3). In addition, 354 spa types were detected in less than 10 isolates. Spa types could not be assigned for 118 isolates in both years.
A comparison of the distribution of spa types of MRSA isolates obtained in 2016-2017 presented in Table 3 revealed that the prevalence of spa type t688 increased (p ≤ 0.05) from 10.4% in 2016 to 12.6% in 2017, while the proportions of t002 and t044 were significantly decreased during the 2 years. No significant changes were observed among the other major spa types. Some sporadic spa types were observed in isolates obtained either in 2016 or 2017 as shown in the supplementary Table S1. The association of spa types with specific genotypes is presented in Table S1.

Distribution of MRSA clones in 2016-2017
The clonal complexes (CCs) of 3004 MRSA isolates obtained in 2016 (N = 1327) and 2017 (N = 1677), selected on the basis of spa types was determined using DNA microarray. The selection included all clinical samples from different hospitals with the same spa type.

Discussion
This study investigated antibiotic resistance and clonal composition of MRSA isolates obtained from patients in Kuwait public hospitals in 2016-2017. The results revealed some changes in the prevalence of resistance to some antibiotics over the 2 years. Significantly, the prevalence of high-level mupirocin resistance decreased from 3.7% in 2016 to 2.4% in 2017. This is consistent with previous report of low prevalence of high-level mupirocin resistance in Kuwait hospitals in recent years [9]. This is reassuring because it indicates that mupirocin can still be used successfully to control MRSA infections. On the other hand, the prevalence of fusidic acid resistance increased from 51% in 2016 to 55% in 2017 which was consistent with previous reports of the high prevalence of fusidic acid resistance in Kuwait hospitals [9]. High prevalence of fusidic acid resistance in MRSA isolated in Kuwait has been suggested to be due to several factors including the emergence or importation of many MRSA genotypes that are resistant to fusidic acid in Kuwait hospitals [10], and the extensive use of topical fusidic acid creams that are readily available over the counter without prescription in Kuwait.
Most of the MRSA isolates in this study carried SCCmec IV, V and VI indicating that the majority of the isolates belonged to the community-associated MRSA genotypes. The significant increase in the proportion of isolates carrying SCCmec VI from 2016 to 2017 could be explained by the increase in the number of CC5 isolates with SCCmec VI.
The majority of the isolates obtained in the 2 years belonged to spa types t688, t304, t860, t127, t044, t311, t223 and t002 that were also dominant previously in Kuwait hospitals [5]. Most of the CC5 isolates detected in this study were of spa type t688. In contrast, t002 was the dominant spa type of CC5-MRSA obtained in New Zealand [11], Switzerland [12] and Canada [13]. However, similar to the findings of this study, t688 is also a common spa type among MRSA isolates reported in Egypt [14], a country with a large population of expatriates in Kuwait.
CC5 is one of the dominant and widely spread MRSA clones reported worldwide [15][16][17][18][19]. In this study, the CC5-MRSA isolates significantly increased (p ≤ 0.05) from 294 isolates in 2016 to 539 isolates in 2017 which could be due to the introduction of new CC5 strains in 2017.
The study revealed a significant reduction in the proportion of CC8 and CC30 isolates in Kuwait hospitals during the study period. CC8 (ST239-MRSA-III) was the predominant clone among MRSA isolates obtained in Kuwait hospitals in the 90's but its prevalence has since reduced [5]. The low prevalence of ST239-MRSA-III isolates observed in both 2016 and 2017 confirms its decline as a major contributor to MRSA infections in Kuwait hospitals although it is still a major cause of health-associated infections elsewhere [23].
The USA300 (ST8-MRSA-IV [PVL + /ACME + ]) clone, the dominant CA-MRSA clone in North America [24] was detected for the first time in Kuwait in 2010 [5]. Although the number of USA300 isolates detected in 2017 (N = 9) remained remarkably low, it represents a significant increase on the single isolate obtained in 2010. As the USA300 is an important cause of infections, it is important to monitor its prevalence among patients in Kuwait hospitals.
The CC22-MRSA-IV is a well-known epidemic MRSA clone that emerged in the United Kingdom in the early 1990s [25] and soon became prevalent in other European countries [26][27][28]. Although CC22-MRSA was the second most common clonal complex in this study (421 isolates) similar to results of a previous study in Kuwait [5], their genotypes have increased from three genotypes  (Table S1).

Conclusion
The study revealed the diversity in the genetic backgrounds of MRSA isolates and the stable maintenance of the dominant MRSA clones in Kuwait hospitals in 2016-2017 suggesting an on-going transmission of these clones. It also demonstrated the emergence of new variants of known genotypes in Kuwait hospitals in 2016 and 2017. Novel and creative infection prevention and control measures are required to curtail further transmission. It is still not clear why some MRSA clones are able to persist while others fail to survive in the healthcare environment. This warrants further investigations to identify the factors that contribute to the spread and maintenance of the successful MRSA clones.

Sample collection
In total, 4726 single patients, MRSA isolates were obtained from different clinical samples submitted to the clinical Microbiology diagnostic laboratory in 11 Public hospitals in Kuwait in 2016 (N = 2305) and 2017 (N = 2421). The isolates were identified using biochemical tests and tube coagulase at the diagnostic microbiology laboratory. Once it was identified as MRSA in the diagnostic laboratories, the isolates were sent to the MRSA Reference Laboratory located in the Department of Microbiology, Faculty of Medicine, Kuwait University for molecular typing where they were retested and confirmed as MRSA. The isolates were sub-cultured twice on brain-heart infusion agar (BHIA) plates to obtain pure colonies and incubated at 35°C for 18 h. Pure cultures were preserved in beads and stored at − 20°C and − 80°C. They were recovered on brain-heart infusion agar (BHIA) and incubated at 35°C prior to further testing.

Antibiotic susceptibility testing
Susceptibility to penicillin G, gentamicin, kanamycin, erythromycin, clindamycin, tetracycline, fusidic acid, trimethoprim, mupirocin, ciprofloxacin, chloramphenicol, rifampicin, cefoxitin, linezolid, vancomycin and teicoplanin were tested using the disc diffusion method according to the Clinical Laboratory Standards Institute (CLSI) [32]. Susceptibility to cefoxitin, vancomycin, teicoplanin, and mupirocin were confirmed by minimum inhibitory concentration (MIC) determination with Etest strips (BioMerieux, Marcy l'Etoile, France) according to the manufacturer's instructions. S. aureus strain ATCC25923 and ATCC29213 were used as quality control strains for the disc diffusion and MIC determination, respectively. Susceptibility to fusidic acid was interpreted according to the British Society to Antimicrobial Chemotherapy (BSAC) [33].

Staphylococcal cassette chromosome mec (SCCmec) typing
SCCmec typing was performed using PCR for all MRSA isolates. Six types of SCCmec was determined by multiplex PCR using primers and protocols published previously [34]. Five μl of the PCR product was analyzed by 1.5% agarose gel electrophoresis to confirm amplification. Five S. aureus strains represented by COL (SCCmec I), XU642 (SCCmec II), WBG 525 (SCCmec III), WBG 9465 (SCCmec IV), WBG 8318 (SCCmec V) were used as quality control for each SCCmec type. The SCCmec types of the isolates were also derived from DNA microarray analysis.

Staphylococcal protein a (spa) typing
All MRSA isolates were investigated by spa typing. Amplification of spa gene was performed using synthetic primers previously published [35]. The PCR protocol consisted of an initial denaturation at 94°C for 4 min, followed by 25 cycles of denaturation at 94°C for 1 min, annealing at 56°C for 1 min, and extension for 3 min at 72°C, and a final cycle with a single extension for 5 min at 72°C. Five μl of the PCR product was analyzed by 1.5% agarose gel electrophoresis to confirm amplification. The amplified PCR product was purified using MicroElute Cycle-Pure Spin kit (Omega Bio-tek, Inc. USA) and the purified DNA was then used for sequencing PCR. The sequencing PCR product was then purified using Ultra-Sep Dye Terminator Removal kit (Omega Bio-tek, Inc. USA). The Purified DNA was sequenced in an automated 3130 × 1 genetic analyzer (Applied Biosystem, USA). The sequence of spa gene was analyzed using the Ridom Staph Type software (Ridom GmbH, Wurzburg, Germany). The software detected the spa repeat and assigned each isolate with spa type.

DNA microarray
Based on spa typing, representative MRSA isolates obtained in 2016 and 2017 were subjected to DNA microarray to determine their clonal complex (CC) using the S. aureus Genotyping kit 2.0 (Alere, GmbH, Germany) with a protocol provided by the manufacturer [36].

Statistical analysis
To determine if the difference in the distribution of the isolates obtained in 2016 and 2017 is statistically significant, 2-tailed Chi square and Fisher exact was performed using Graphpad (https://www.graphpad.com/quickcalcs/ catMenu/). Also, the significance was calculated by comparing the proportions between two sample sizes using Epicalc 2000 Version 1.02 (J & Myatt M, Brixton Books, Brixton, UK). P ≤ 0.05 was considered to be statistically significant.