Shifts in the Clonal Distribution of Methicillin-Resistant Staphylococcus aureus in Kuwait Hospitals: 1992-2010

Background As the epidemiology of methicillin-resistant Staphylococcus aureus (MRSA) is constantly changing globally, determining the prevailing MRSA clones in a local healthcare facility is important for better management of infections. This study investigated clonal composition and distribution of MRSA isolates in Kuwait’s hospitals using a combination of molecular typing methods. Materials and Methods In total, 400 non-repeat MRSA isolates were obtained between 1992 and 2010 in 13 public hospitals and were characterized using antibiogram, SCCmec typing, spa typing, and multilocus-sequence typing. Clonal assignment and detection of virulence factors and antibiotic resistance genes were performed by DNA microarray. Results The isolates were resistant to kanamycin (74.2%), erythromycin (69.5%), tetracycline (66.7%), gentamicin (61%), ciprofloxacin, (61%), fusidic acid (53.5%), clindamycin (41.5%), high-level mupirocin resistance (5.2%) and carried aphA3, aacA-aphD, ermA, ermC, mupA, tetK, tetM, fusC and far1. Molecular typing revealed 31 different MRSA clones consisting of ST239-MRSA-III (52.2%), ST22-MRSA-IV (9.2%), ST80-MRSA-IV (7.5%), ST5-MRSA-II/IV/V/VI (6.5%), ST30-MRSA-IV (3.5%), ST241-MRSA-III (2.7%), ST6-MRSA-IV (2.2%), ST36-MRSA-II (2%) and ST772-MRSA-V (1.75%). The isolates differed in the carriage of genes for enterotoxins, Panton–Valentine leukocidin (PVL), toxic shock syndrome toxin (tst-1), arginine catabolic mobile element (ACME) and exfoliative toxins. The number of clones increased from one (ST239-III-t037) in 1992 to 30 in 2010 including ST8-IV-t008 [PVL+] [ACME+] (USA300), ST772-V (Bengal Bay clone) and ST2816 identified for the first time in Kuwait. Conclusion The study revealed that the MRSA isolates belonged to diverse clones that changed in numbers and diversity overtime. Although ST239-MRSA-III, a healthcare-associated clone remained the dominant MRSA clone overtime, the newly emerged clones consisted mostly of community-associated.


Introduction
The number of MRSA isolates obtained in Kuwait hospitals has increased over the years [16]. Previous studies revealed that MRSA constituted 32% of S. aureus isolated from clinical samples in Kuwait and consisted of both healthcare-associated and community-associated strains [16,17]. Despite these studies, there has been no systematic characterization of MRSA isolated in Kuwait hospitals overtime to catalogue changes in their clonal distribution. The purpose of this study was to investigate the number and type of MRSA clones circulating in major Kuwait hospitals and determine their genotypic characteristics. The strains were investigated using a combination of molecular typing techniques including Staphylococcal chromosome cassette mec (SCCmec) typing, spa typing and multilocus-sequence typing (MLST). Additionally, the study aimed to determine if there were changes in the distribution of the MRSA clones in Kuwait hospitals over time. Tracking the emergence and spread of MRSA and the introduction of new clones will provide information needed to guide policy for appropriate therapy and infection control. In addition, it will provide information about the composition of MRSA clones in Kuwait, which is necessary for epidemiological purposes.

Collection of MRSA strains
In total 400 non-duplicate MRSA, isolates representing isolates obtained from different clinical ( Table 1) samples in 13 government Kuwait hospitals between 1992 and 2010 were  investigated. The isolates were selected from a collection of MRSA isolates archived at the MRSA Reference Laboratory in Kuwait. The isolates were previously characterized by pulsedfield gel electrophoresis (PFGE) and interpreted according to the criteria prescribed by Tenover et al., [18]. Those selected for this study represented different PFGE patterns obtained in 1992-2010. During the 90's and early 2000's, the number of PFGE patterns were small due to the presence of an endemic clone [19,20] which is reflected in the small numbers of isolates chosen for this study from these years.  [22]. Etest was used to determine the MIC for vancomycin and teicoplanin. The interpretation of the MIC values was based on the antibiotic breakpoint concentration recommended by the CLSI; [21].

Detection of Panton-Valentine leukocidin (PVL)
PVL was performed for all MRSA isolates using primers and PCR protocol published by Lina et al., [23].

SCCmec typing and Staphylococcus protein A (spa) typing
All MRSA isolates were characterized using SCCmec typing and spa typing. SCCmec typing was performed as described by Zhang et al., [24]. Detection of SCCmec-IV subtypes IVa, IVb, IVc, IVd, IVg and IVh was determined using multiplex PCR described by Zhang et al., [24] and Milheiriço et al., [25]. Spa typing was performed as described by Harmsen et al., [26].

Multilocus sequencing typing (MLST)
MLST was performed as described by Enright et al., [27] on 103 isolates representing different spa types in 1992-2010. Electronic based upon related sequence types (eBURST) (http://eburst. mlst.net) [28] analysis was performed and assigns each ST that shares at least five of seven identical alleles into a single clonal complex (CC).

DNA Microarray
DNA microarray was performed for 110 MRSA isolates representing each spa type and sequence type isolated from 1992 to 2010. Genes encoding virulence factors and antibiotic resistance were determined using the ArrayMate Reader DNA Microarray platform with Identibac S. aureus genotyping Kit 2.0 (Alere Technology, Jena, Germany) following protocols provided by the manufacturer.

Detection of genes for PVL among MRSA isolates
The result showed that 62 (15.5%) isolates were positive for PVL, while 338 isolates (84.5%) of the isolates was negative.
Clonal complex 1 (CC1). The seven isolates that belonged to CC1 were obtained in 2010. They consisted of two sequence types; ST1-V (three isolates) and ST772-V (four isolates). The ST1-V isolates corresponds to the ST1-MRSA-V SCCfus clone, while the ST772-V isolates correspond to the Bengal Bay/WA MRSA-60 clone. The ST1-V isolates consisted of three spa types, t127, t321, and t6811, which were positive for agrIII and cap8 but varied in their carriage of PVL and antibiotic resistance genes. ST1-V-t6811 isolate was negative for PVL but resistant to erythromycin and clindamycin mediated by ermC, whereas t127 and t321 isolates were positive for PVL, and were multiresistant to gentamicin, kanamycin, tetracycline, fusidic acid, ciprofloxacin, tobramycin and carried aacA-aphD, aphA3, fusC, sdrM and sat determinants.
All CC5 isolates harbored agrII, and cap5 and the genes for virulence determinants and antibiotic resistance summarized in Table 2. ST105-II-t002 isolate was PVL-negative, but was positive for tst gene in addition to genes, sea and egc gene cluster.
The ST5-IV clone consisted of two spa types, t306 and t2164, which were PVL-positive and carried similar enterotoxins genes, sea, sed, sej, ser and egc, and contained ermC mediated erythromycin and clindamycin resistance ( Table 2). The ST149-IV was PVL-negative but carried sea and egc. In addition, ST149-IV-t1154 isolate was positive for sec, sel and tst. All ST149-IV-t1154/t2164 isolates were resistant to erythromycin, clindamycin and fusidic acid and carried ermC, fusC, fosB and sdrM.
The ST5-V-t688 isolate carried sea, sed, sej, ser and egc and was sensitive to the majority of the antibiotics tested in this study except tetracycline mediated by tetK and tetM. In addition, the isolate harbored fexA, fosB and sdrM determinants. The ST1637-V-t5258 isolate was PVLnegative, but was positive for sea, sed, sej, ser and egc and was resistant to the aminoglycosides, gentamicin, kanamycin, amikacin and tobramycin and harbored aacA-aphD, fosB and sdrM determinants.
Clonal complex 6 (CC6). The two representative CC6 isolates selected for microarray analysis carried sea, agrI, cap8 and the virulence and antibiotic resistance genes summarized in  . Thirty-seven of the 110 isolates selected for microarray analysis belonged to CC8. The CC8 isolates belonged to six STs and 10 spa types. The STs consisted of ST239-III (19 isolates), ST241-III (10 isolates), ST8-IV (four isolates), ST113-IV (two isolates), ST8-V (one isolate), and ST1465-III (one isolate). All CC8 isolates were positive for agrI and cap8 but carried relatively few enterotoxins encoding genes (Table 3).
The ST8-IV-t008 isolate, resembling the Lyon/UK EMRSA-2 clone was isolated in 2005. It was negative for PVL and ACME but positive for sea.
The ST8-V-t064 isolate yielded negative results for PVL and ACME but carried sea, seb, sek and seq.
Microarray analysis classified the ST22-IV isolates into three groups comprising (i) t032 Clonal complex 30 (CC30). Isolates belonging to CC30 were first isolated in 1996 and continued to be isolated until 2010. The 11 CC30 isolates selected for microarray analysis belonged to two sequence types, ST30-IV (seven isolates) and ST36-II (four isolates). The ST30-IV isolates belonged to spa types, t019, t318, t345 and t1130, whereas the ST36-II isolates belonged to spa types t018 and t605. Microarray analysis revealed that the ST30-IV isolates were similar to the CC30-IV [PVL+] Southwest Pacific clone, whereas the ST36-II isolates were similar to the UK EMRSA-16 clone.
The ST36-II isolates consisted of ST36-II-t018 (N = 3) and ST36-II-t605 (N = 1). All ST36-II were positive for agrIII, cap8 and egc gene cluster. All isolates were PVL-negative. Two isolates obtained in 1999 (t018 and t605) were tst-positive. The ST36-II isolates were resistant to kanamycin, erythromycin, clindamycin, spectinomycin and ciprofloxacin, and carried ermA, aadD, sdrM and fosB. Of the three ST36-II-t018 isolates, one isolate obtained in 2010 was resistant to gentamicin and carried additional aacA-aphD and additional erythromycinresistance determinant, ermC. Another isolate obtained in 1999 carried aphA3. Two isolates obtained in 1999 were resistant to tetracycline and carried tetK, while two isolates obtained in 1999 and 2010 were resistant to trimethoprim and carried dfrS1 (Table 4).
The PVL-positive ST80-IV isolates were similar to the CC80-IV [PVL+] European CA-MRSA clone, while the PVL-negative ST80-IV isolates defined a different CC80-MRSA variant. Genes for virulence and antibiotic resistance of CC80 are summarized in Table 4.
The ST88-IV isolates differed in their carriage of enterotoxin encoding genes. The ST88-IV-t4067/t5041 isolates carried sea, while the ST88-IV-t690 isolate carried sea, sek and seq. One t690 isolate carried lukF-PV and lacked lukS-PV. The isolates were also non-multiresistant to antibiotics and carried few antibiotic resistance genes. All the ST88-IV isolates were resistant to ** The t4067 isolate was sensitive to tetracycline and lacked tet gene. *** The t315 isolate also carried sec.
erythromycin and clindamycin and carried ermC. In addition, the two ST88-IV isolates belonging to t690 and t5041 were resistant to tetracycline and carried tetK. The ST1289-IV-t5562 (WA MRSA-2 clone) isolate lacked enterotoxin and PVL genes, and was only resistant to trimethoprim and carried dfrS1 (Table 5). Clonal complex 97 (CC97). All four CC97 isolates were obtained in 2010. The isolates belonged to ST97-V and spa types t1234 and t13204. All CC97 isolates were positive for agrI and cap5,sak, scn, but were negative for genes encoding staphylococcal enterotoxins, PVL, TSST-1, ET and ACME ( Table 5). The CC97 isolates were susceptible to non-beta-lactam antibiotics tested, although they carried sdrM (Table 5).
Clonal complex (CC913). CC913 isolates were all isolated in 2010 and were represented by four isolates that belonged to ST913-IV-t991. A representative isolate selected for microarray analysis was positive for agrII, cap8, sed, etA, etD and the resistance determinants ermC and sdrM (Table 5).

New sequence type (ST).
A new sequence type, ST2816, was identified during this study. The isolate carried SCCmec-V and belonged to spa type t605. The isolate was negative for enterotoxins, PVL, TSST-1, ET and ACME but was positive for ebh, fnbB, map, sak and scn (Table 5). It was resistant to tetracycline and ciprofloxacin and carried tetk.
Changes in the prevalence of MRSA clones from 1992 to 2010 Table 6 summarizes the changes in the numbers and types of MRSA clones from 1992 to 2010. The results showed that the number of MRSA clones changed from one major clone in 1992 to 30 different clones in 2010.
The 15 representative isolates obtained in 1992 belonged to the same clone, ST239-III-t037. However, the number of clones increased to three in 1996 and five in 1999. In 1996, three clones, ST239-III-t037/t421, ST30-IV-t019 and ST241-III-t1902 were detected. Therefore, 1996 marked the introduction of the ST30-IV, the Southwest Pacific CA-MRSA clone, into Kuwait hospitals. Similarly, 1997 marked the introduction of ST80-IV, a European CA-MRSA clone. The types and numbers of MRSA clones detected in 1998 remained the same as the previous years; but by 1999, the number of MRSA clones increased to five and marked the introduction of ST36-II, the Epidemic-MRSA-16 (EMRSA-16) clone.
Thirty MRSA clones were detected among the 205 representative MRSA isolates obtained in 2010. These consisted of 12 clones that were identified in previous years and 18 new clones ( Table 6). Additionally, it was observed that clones detected in previous years were associated with different spa types in 2010. For example, the ST5-IV clone previously detected in 2001-2002 belonged to spa type t306, but ST5-IV that were isolated in 2010, were associated with spa types, t002, t688 and t2164. Similarly, ST5-II clone that was detected in 2001-2002 was associated with spa type t105, but in 2010, the ST5-II was detected in four isolates that were associated with spa types, t002, t003 and t242. Similarly, the 33 ST22 isolates obtained in 2010 contained new spa types, t852, t790, t3107, t309, t3935, t5708 and t5983 in addition to spa types, t223 and t032 that were detected in 2005. The ST239-III-t037/t421/t945 clone was found in 69 isolates in 2010 including spa type, t860 that was not detected in previous years.

Discussion
This study revealed a striking diversity in the numbers and types of MRSA clones obtained in Kuwait hospitals from 1992 to 2010. Although, several MRSA clones were identified, the majority (64.2%) corresponded to the well-established healthcare-associated ST239-III and ST22-IV MRSA clones [5].
Microarray analysis distinguished two types of ST239-III isolates in this study. One type, consisting of spa types, t037/t421/t860/t945, resembled the well characterized pandemic clone also known as the Hungarian/Brazilian clone [8,9] which is one of the most successful MRSA lineages reported in many parts of the world [29,30,31,32] including the Middle East [33,34,35]. The isolates in this group were positive for sea, sek and seq but lacked PVL and ACME. These were similar to ST239 strains reported in Saudi Arabia [33], Turkey [34], Russia [36] and Germany [37]. The second type consisted of ST239-III-t030 and ST1465-III-t459, characterized by their carriage of ACME genes. ACME has been widely reported among ST8-IV (USA300) [38] clone but less commonly in other MRSA clones [39,40]. However, ACME-positive ST239-III MRSA isolates have been reported in Australia [10], Singapore [41], Malaysia [42] and strains belonging to other genetic backgrounds such as ST764-SCCmec-IIa obtained from cases of acute otitis media in Japan [43] indicating that ACME is spreading among MRSA with different genetic backgrounds. This report of ACME in MRSA isolates obtained in Kuwait hospitals signifies the emergence of ACME producing MRSA clone in our region.
Microarray analysis of antibiotic resistance genes in these isolates yielded two interesting results. First, it showed that none of the fusidic acid resistant ST239-MRSA-III isolates carried  fusB or fusC determinants included in the microarray panel suggesting the involvement of an additional mechanism of fusidic acid resistance such as fusA [44]. A previous study in Kuwait showed that fusidic acid resistant MRSA isolates expressed chromosomally-mediated high levels of fusidic acid (MIC: ! 256 μg/L) suggesting the presence of fusA [45]. Therefore, the isolates in this study may harbor fusA. Secondly, the microarray results revealed the presence of fosB encoded fosfomycin resistance that was not previously reported in MRSA isolates in Kuwait. Fosfomycin susceptibility testing is not performed routinely in Kuwait, which explains why the resistance was missed even though a substantial number of the isolates were resistant. The detection of this resistance by microarray highlights the benefits of this technology over traditional antibiotic susceptibility testing in detecting novel resistance determinants.
ST22-IV clone was the second most common clone isolated in Kuwait hospitals. It is a wellknown epidemic MRSA clone that emerged in the United Kingdom in the early 1990s [13] and has become prevalent in Europe since then [31,46,47,48]. The results of this study showed that ST22-IV MRSA corresponded to three groups including CC22-UK EMRSA-15/Barnim EMRSA, CC22-MRSA-IV-PVL+ and CC22-tst+ UK EMRSA-15/Middle Eastern variant clones. Interestingly, the majority of the isolates were related to the tst+ UK EMRSA-15/Middle Eastern variant [15]. The diversity of the ST22-IV variants circulating in Kuwait hospitals may reflect diverse sources of their acquisition.
The other major MRSA clones detected during this study included ST80-IV, ST30-IV, ST6-IV and ST772-V. ST80-IV-MRSA was previously reported as the most common CA-MRSA clone in Kuwait hospitals in 2005-2006 [16, 17] and was postulated to consist of European and some indigenous clones due to differences in their antibiotic resistance phenotypes [17]. The present results showed that the ST80-IV MRSA were heterogeneous, belonged to different spa types including t044, t042, t8154 and t376 which confirms previous suggestions of their diverse route of acquisitions [17]. Surprisingly only 2.9% ( ST6-IV-t304 MRSA was recently reported as the dominant MRSA clone in a hospital in Oman [35]. Spa type t304 has also been reported in MRSA isolates in UAE [49], MSSA obtained from patients in Lebanon [50] and in MSSA from Camels in UAE [5] suggesting that the strains may be transmitted between humans and Camels. It would be interesting to undertake further investigations into the prevalence of this clone in the region.
The ST772-V MRSA, known as the Bengal Bay clone because of its origin in Bangladesh and India [51], is becoming the dominant MRSA clone in India [52] but has also been reported in Australia where it is called WA60 [53] and in the United Kingdom [54]. Although ST772-V clone was reported previously from patients in Saudi Arabia [33], UAE [49], Qatar [55] and Oman [35], this is the first report in Kuwait hospitals. ST772 isolates were usually associated with spa type t657. However, other spa types including t10795, t345 and t12211 were detected in small numbers in this study signaling the evolution of variants of the Bengal Bay clone.
The ST8-IV-t008, also known as USA300, was also detected for the first time in Kuwait hospitals. The USA300 is a multiresistant, PVL + CA-MRSA clone that has been recognized as the leading cause of community-associated as well as healthcare-associated infections in North American hospitals [56,57,58]. The USA300 MRSA clone is also increasingly being reported in hospitals outside North America [34,43,59]. The detection of the USA300 CA-MRSA clone in Kuwait highlights its growing importance as an international epidemic MRSA clone.
Isolates carrying SCCmec II genetic element are classified as HA-MRSA whereas those carrying SCCmec IV, V and VI are classified as CA-MRSA. Therefore, the ST5 isolates in this study consisted of HA MRSA and CA-MRSA clones. The low prevalence of ST5-IV MRSA in this study (2.7%) mirrors the situation in Asian countries such as Japan, China, Taiwan [60] and Sri Lanka [32] where ST5-IV-MRSA have been rare. Other members of CC5 identified in this study included the ST5/ST1637-V, which was related to WA MRSA 11/34/35/90/108 [53], the ST5-II, related to the New York/Japan MRSA clone, ST149-IV, related to the Maltese clone, and the ST5-VI, which was related to the New Pediatric clone. All of the CC5 isolates carried the egc gene cluster and all CC5 isolates, except those related to the Maltese (ST149-IV) and [tst+] New York/Japan clones (ST5-II) carried an additional enterotoxin, ser [61]. The egc gene cluster was also common in CC5 isolates reported from Western Australia [53], Malta [61], Germany [37], Ireland [62] and USA [63]. In addition, this study and studies from Malta and Ireland reported the presence of sea and sej among the CC5 isolates [61,62]. Whereas, ST5-IV (Pediatric clone) were PVL-positive, tst was found in ST105-II (New York/Japan clone) and ST149-IV (Maltese clone) isolates. ST97-V isolates were detected in six MRSA isolates in this study. ST97-MRSA-V isolates were previously reported to cause an outbreak in a neonatal unit of a Kuwait hospital in 2007 [64]. However, the ST97-V isolates obtained in 2010 in this study were all susceptible to non-beta-lactam antibiotics; whereas the 2007 outbreak isolates were resistant to gentamicin, kanamycin and fusidic acid indicating that the two ST97-V isolates were unrelated. The ST97 isolates carried sak and scn encoding genes (immune evasion cluster/IEC type E) suggesting that they are of human origin [7].
In addition to revealing the diversity of MRSA clones that circulated in Kuwait hospitals from 1992 to 2010, the study also demonstrated changes in the composition of the MRSA clones during the same period. The distribution of the MRSA clones in Kuwait hospitals remained largely unchanged from 1992 to 1999, as the clonal composition of MRSA was mostly those of healthcare-acquired MRSA represented by ST239-III, the Brazilian/Hungarian MRSA clone. Despite the increase in the number of MRSA clones from 1996 onwards, the ST239-III clone continued to be the dominant clone in Kuwait hospitals. Similar to the result of this study, ST239-III has remained the predominant MRSA clone in Saudi Arabia [33] and Malaysia [42]. Other studies have reported the replacement of ST239-III MRSA as the dominant MRSA clone by CA-MRSA clones in UAE [49], Singapore [65] and Portugal [66], India, ST239-III was replaced by CC22-MRSA-IV and ST772-MRSA-V [52]. Similarly, displacement of ST239-III MRSA by ST22-MRSA-IV has been reported in Germany [67] and Czech Republic [47]. These results highlight the differences in the distribution of MRSA clones in different countries.
The emergence of CA-MRSA in the mid 90's changed the epidemiology of MRSA globally [68]. The present study has shown that CA-MRSA appeared in Kuwait hospitals in 1996 with the detection of ST30-MRSA-IV, the Southwest Pacific clone (SWP), followed by ST80-IV in 1997. The other CA-MRSA clones appeared in 2000 and beyond. It is interesting that the ST30-IV MRSA isolates appeared in Kuwait hospital in 1996 a year after it was reported in New Zealand [69]. It is possible that a South West Pacific national or a returning Kuwaiti tourist to that region introduced it to Kuwait.
The prevalence of ST80-IV, the European CA-MRSA clone, increased in from its introduction in 1997 until 2010. Curiously, ST80-IV-MRSA has not been reported as extensively in Asian countries as in European countries. A surveillance conducted by Song et al., [32], in Asian countries including Korea, Taiwan, Hong Kong, Thailand, Vietnam, India, Sri Lanka and the Philippines) did not detect ST80-IV-MRSA, and only few studies in Malaysia [70] and Singapore [71] reported the presence of ST80 in small numbers of their MRSA isolates. In contrast, ST80-IV has become the most common CA-MRSA clones in the Middle East and North Africa [16,17,33,49,72,73]. The presence of ST80 isolates in the Middle East and not in other parts of Asia highlights the yet unexplained observations whereby certain clones flourish in certain places but not in others.
In conclusion, the results of this study has shown that MRSA isolated in Kuwait hospitals belonged to diverse genetic backgrounds suggesting multiple routes of their acquisition. The majority of the isolates were related to the established healthcare-associated MRSA clones, ST239-III and ST22-IV. However, spa typing revealed that the isolates were heterogeneous representing imported and local clones. The study also revealed a shift in the clonal composition of MRSA isolates in Kuwait hospitals overtime. The number of MRSA clones increased from one in 1992 to 30 in 2010. In the 90's and early 2000's, the majority of the isolates belonged to ST239-III. Although ST239-III remained the dominant clone in 2010, other clones have appeared with the majority identified as CA-MRSA.