Association between fluoroquinolone resistance and MRSA genotype in Alexandria, Egypt

Antimicrobial stewardship isn’t strictly observed in most Egyptian hospitals, raising antibiotic resistance. Epidemiology of Egyptian MRSA isolates, or associations with resistance to other antibiotics remain largely unknown. We identified MRSA genotypes in Alexandria Main University Hospital (AMUH) and investigated rates of moxifloxacin resistance, an alternative MRSA treatment, among different genotypes. Antibiotic susceptibility of 72 MRSA clinical isolates collected in 2015 from AMUH was determined by disc diffusion and broth microdilution. spa- and Staphylococcal Cassette Chromosome mec (SCCmec) typing were performed; with multi-locus sequence typing conducted on isolates representing major genotypes. Resistance to moxifloxacin, levofloxacin and ciprofloxacin were 69%, 78% and 96%, respectively. spa type t037 (57%) was commonest, followed by t127 (12.5%), t267 (8%) and t688 (6%). SCCmec III predominated (57%), all of these were moxifloxacin resistant and 97.6% t037 (ST241). SCCmec IV, IV E and V represented 15%, 7% and 11% of the isolates, respectively, 79% of these were moxifloxacin susceptible and of different spa types. t127 (ST-1) was associated with SCCmec V in 56% of the isolates, mostly moxifloxacin susceptible. Moxifloxacin resistance was high, most resistant isolates belonged to t037 and SCCmec III, suggesting local dissemination and antibiotic pressure. We recommend caution in treating MRSA infections with moxifloxacin.


Results
Antibiotic susceptibility among the bacterial isolates. The majority of the isolates were from male patients (n = 43, 60%) and were isolated from non-ICU locations (n = 56, 78%), the median age was 35 years. We were able to categorize 56 of the isolates as being either CA or HA, of these 80% (n = 45) were HA and 20% (n = 11) were CA (Table 1).
Isolates belonging to spa type t127 were predominantly moxifloxacin susceptible (78%). Sixty seven percent of spa type t267 isolates had SCCmec type IV, and were moxifloxacin susceptible. Fifty six percent (n = 5) of   www.nature.com/scientificreports/ isolates belonging to spa type t127 also belonged to SCCmec type V and all but one isolate were moxifloxacin susceptible. Moxifloxacin resistance was variable among isolates belonging to spa types t688, t223, t044 and t304. One isolate of spa type t688 was identified as SCCmec IVE and the remaining isolates were of unknown SCCmec types. All isolates of spa type t223 and spa type t044 carried SCCmec IVE and SCCmec IV, respectively. Two isolates belonged to spa type t304; one isolate was moxifloxacin resistant and SCCmec III while the other isolate was moxifloxacin susceptible and SCCmec IV. spa types t416, t6978, t786 and t16221 were identified only once and in moxifloxacin susceptible isolates (Supplementary Table S1   Phylogenetic tree of the spa types detected using the neighbour-joining method. The isolates were grouped in three clades, clade 1 containing isolates belonging to spa types t044, t267, t786 and t127, clade 2 containing isolates of spa t304, t037 and t1622 and clade 3 containing isolates from spa t688, t223 and t6978.

Discussion
MRSA remains a major causative agent of infections on a global scale, with worse outcomes relative to methicillin susceptible infections 27,28 . The prevalence of MRSA in Egypt is high (52%) compared to other North African and Mediterranean countries such as Morocco (19%), Libya (31%), Algeria and Tunisia (45%) 6 . In this scenario, the treatment of infections caused by MRSA requires the use of an alternative agent such as moxifloxacin 12 , that is readily available on the Egyptian market. Few reports have described the levels of moxifloxacin resistance among Egyptian S. aureus [18][19][20] . Tackling moxifloxacin resistance in Egypt requires a better understanding of the molecular lineages of Egyptian MRSA clones 21 . In this study, moxifloxacin resistance among 72 Egyptian MRSA clinical isolates was determined relative to their genotype.
Rates of resistance to fluoroquinolones were higher than previously reported 29,30 , and although moxifloxacin had the highest level of in-vitro susceptibility, sixty nine percent of the isolates were moxifloxacin resistant. SCCmec type III was the most common SCCmec type and almost always associated with spa type t037 (ST-241) and moxifloxacin resistance. This is not surprising as the spa t037/SCCmec III clone is a predominant clone in Africa 25 . The association between spa t037/SCCmec III/ST-241 and moxifloxacin resistance seen among the studied isolates is in agreement with previous findings that SCCmec III isolates showed 89% resistance to ciprofloxacin 31 . Moreover, a previous study described a spa t037/SCCmec III/ST-241 clone that was moxifloxacin resistant among isolates from Nigeria 32 . Almost 70% of the SCCmec type III isolates in the current study were HA-MRSA, in accordance with previous reports that HA-MRSA strains mainly belong to SCCmec types I, II and  www.nature.com/scientificreports/ III, while CA-MRSA strains are mainly of SCCmec types IV and V [33][34][35] . In the current study, isolates belonging to SCCmec types IV, IVE and V were found among HA-MRSA as well as CA MRSA infections. Upon isolate collection, data on prior hospitalization was not obtained which could have changed the classification as HA or CA and might explain the discordance. However the association between SCCmec type and community-or hospital onset of infection is also less clear cut, possibly due to spread of hospital strains into the community and vice versa. In addition, the limited number of CA MRSA isolates in the current study makes it harder to draw solid conclusions. The moxifloxacin resistant ST-241-III/t037 clone might represent the majority of the tested isolates as spa type t037 and SCCmec III were the most predominant types. This agrees with a study that showed that the major clonal complex causing HA-MRSA in Africa was ST-239/ST-241-III 36 . However, to the best of the authors' knowledge this is the first report of the moxifloxacin resistant ST-239-III/t304 clone.
The majority of the isolates belonging to SCCmec types IV, IVE and V were susceptible to moxifloxacin with resistance rates of 18%, 20% and 25%, respectively. This concurs with Kilic et al. who showed levofloxacin resistance of 16.8% among SCCmec type IV isolates 37 . The moxifloxacin susceptible ST-1-V/t127 MRSA clone has been reported among human isolates in a European study, yet the porcine MRSA isolates from the same study were 95% resistant to ciprofloxacin 38  In conclusion, there were five new MRSA clones identified in this study; ST-239-III/t304, ST-1502-IV/t044, ST-4808-IV/t267, ST-22-IVE/t223 and ST-22-IVE/t6978 in addition to ST-97-V/t267 clone that was reported for the first time among human MRSA isolates. Resistance to fluoroquinolones was common (69% resistance to moxifloxacin), and appears to be driven by the predominant spa t037/SCCmec III clone which is moxifloxacin resistant. This may point to a potential clonal dissemination of this strain within hospitals. spa type t16221 and MLST ST-4808 type were newly identified in the present study, and were both moxifloxacin susceptible. The high rates of moxifloxacin resistance detected among the isolates calls for stricter implementation of antimicrobial stewardship guidelines and infection control practices among Egyptian hospitals.

Materials and methods
Sample collection and identification. Seventy two MRSA isolates collected from the Medical Microbiology laboratory at Alexandria Main University Hospital (AMUH) between September and December 2015 were included in the study. AMUH is the largest teaching hospital in northern Egypt with four satellite hospitals and a total capacity of 3500 beds. The isolates represented all non-duplicate MRSA isolates obtained from different clinical specimens, including pus, blood, sputum, urine, tissue, aspirate and broncho-alveolar lavage (BAL). The identity of the isolates was confirmed using Gram staining, bacterial growth and fermentation of mannitol salt agar, growth on DNase agar and slide coagulase testing using Dryspot Staphytect Plus (Oxoid Ltd, England) 40 . The isolates were classified as obtained from HA versus CA infections based on a 48 h window between admission and specimen collection.
Antibiotic susceptibility testing. Methicillin resistance was confirmed by cefoxitin disc diffusion testing (Oxoid Ltd, England), and susceptibility of the isolates to moxifloxacin, ciprofloxacin and levofloxacin was determined using disc diffusion and confirmed by Minimum Inhibitory Concentration (MIC) determination  spa typing. The polymorphic X region of spa gene was amplified according to Harmsen et al. 47 and the PCR products were Sanger sequenced using the ABI 3130XL Genetic analyzer (Inqaba Biotechnologies, South Africa;). Chromatograph sequence files were processed using the BioEdit Sequence Alignment software for creation of the consensus sequence. Sequence analysis was performed using spatyper online (http://spaty per.forti nbras .us) and/or the Ridom StaphType software (Munster, Germany) and spa clonal complexes (spa-CC) were allocated using the 'based upon repeat pattern' (BURP) algorithm that is implemented within the software 47,48 . spa types that were too short to presume ancestry (< 5 repeats within the hypervariable Xr region of the spa gene) were excluded from the BURP analysis. A phylogenetic tree of the identified spa types was constructed using the Molecular Evolutionary Genetic Analysis X software (MEGA X) 49 using the neighbour-joining method 50 . The new spa type was submitted to the Ridom spaServer under accession number 174278 51 .
MLST. Eleven isolates representing the most common spa and SCCmec types were selected for MLST typing according to Enright et al. 52 . Briefly, seven housekeeping genes were targeted using their specific primers. The PCR products were Sanger sequenced and the consensus sequence of each gene as generated by the BioEdit Sequence Alignment software was identified as a specific allele type. The loci were combined and identified as ST types using PubMLST online (https ://pubml st.org/saure us/) 53 . A phylogenetic tree was constructed using PHY-LOViZ software (http://www.phylo viz.net/) by UPGMA and eBURST 54 to investigate the evolutionary history of the identified MLST types and the degree of relatedness. The eBURST algorithm method was applied using eBURST v3 online tool (http://eburs t.mlst.net/defau lt.asp).

Data availability
The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.