Dominant serotype distribution and antimicrobial resistance profile of Shigella spp. in Xinjiang, China

Shigella represents one of the major diarrhea-inducing pathogens threatening public health, but its prevalence and antimicrobial resistance profile in Xinjiang Uygur Autonomous region, China, remains unclear. We conducted comprehensive investigation of Shigella serotype distribution and antimicrobial resistance pattern in Xinjiang, identifying 458 Shigella isolates between 2008 to 2014. Shigella flexneri was identified as predominant species, and several S. flexneri serotypes were isolated, including atypical serotypes 1c, 2c, and 4s. Dominant S. flexneri serotypes were 2a, 1b, 2b, and Xv, different from those generally dominant in China. A hybrid serotype pattern was observed, which included the major Chinese serotypes (2a, Xv) and those predominant in Pakistan (1b, 2b). Shigella sonnei was shown to have a lower frequency compared with that generally observed in China, but an increasing trend of infections associated with this pathogen was observed. Furthermore, a high frequency of drug resistance and different Shigella antimicrobial resistance patterns were demonstrated as well, including very severe resistance phenotypes, such as multidrug resistance and resistance to frontline antibiotics. Seventy-five cephalosporin-resistant Shigella isolates were frequently identified with the resistance determinants that can undergo horizontal transfer, such as blaOXA, blaTEM, blaCTX-M, and integrons, facilitating the development of cephalosporin resistance among Shigella subtypes. Additionally, genetic analyses demonstrated that all 86 quinolone-resistant S. flexneri isolates possess 3–4 mutation sites in quinolone resistance-determining regions, primarily contributing to their resistance to quinolone. However, S. sonnei isolates were not shown to be quinolone resistant. Co-resistance to cephalosporins and quinolones was detected in 17 S. flexneri isolates, and these isolates were additionally multidrug resistant and carried β-lactamase genes and quinolone-resistance determinants. As is demonstrated in this study, dominant serotypes of Shigella were distributed in unique trend with dangerous drug resistance patterns. Novel strategies are urgently required to prevent the development of drug resistance among diarrhea-inducing pathogens.

Introduction second incubation, Shigella colonies were picked and streaked on Luria-Bertani agar plates, followed by incubation at 37˚C for [16][17][18][19][20][21][22] h, after which these isolates were identified with API 20E test strips (bioMérieux SA, Marcy l'Etoile, France), according to the manufacturer's instructions. The serotyping of Shigella isolates was performed using Shigella Antisera (Denka Seiken, Tokyo, Japan) and monoclonal antibody reagents (MASF IV-1 and MASF IV-2, Reagensia AB, Stockholm, Sweden). Written informed consents were obtained from patients or their guardians. All experiments were approved and authorized by the Ethics Committees of the Institute of Disease Prevention and Control, People's Liberation Army, China.

Statistical analysis
Differences in the AMR rates and the frequency of AMR determinants between Shigella species were analyzed using χ 2 test. Variations in Shigella prevalence with time were analyzed using linear regression. All statistical analyses were performed using Graph Pad 7.0 software, and P < 0.05 was considered statistically significant.

Identification of isolates and distribution of Shigella spp
All Shigella isolates in our study were collected from sentinel hospitals in 12 cities or prefectures ( (Table 1). Of these, 365 (79.7%) were identified as S. flexneri isolates, and 93 (20.3%) were identified as S. sonnei isolates. S. boydii and S. dysenteriae presence was not detected. All collected S. flexneri isolates belonged to one of at least 16 serotypes (Fig 1), while serotype 2a (134, 36.7%) was identified as the major S. flexneri serotype detected during 7 years of surveillance, followed by 1b (47, 12.9%), 2b (46, 12.6%), and Xv (36, 9.9%) (Fig 1). Some atypical serotypes were identified, including three 1c isolates, four 2c isolates, and seven 4s isolates. Consistently, S. flexneri was shown to be the dominant Shigella species every year, while S. sonnei isolates accounted for less than 30% of samples (Fig 2). However, we observed a change in this trend, with the percentage of S. sonnei among the isolates showing the tendency to increase with time (linear regression analysis, P < 0.05) (S1 Fig). S. sonnei species, together with the dominant S. flexneri serotypes (2a, 1b, 2b, and Xv), were shown to constitute the majority of Shigella isolates in Xinjiang, with the frequency of above 70% each year (Table 1).

Detection of AMR determinants and integrons
A total of 57 (15.6%) S. flexneri and 18 (19.4%) S. sonnei isolates showing high-level resistance to cephalosporin were selected to detect the presence of antibiotic-resistance determinant genes and integrons ( Table 4). Rates of bla OXA-1 , bla TEM-1, and bla CTX-M expression among the selected S. flexneri isolates were 82.5%, 61.4%, and 89.5%, respectively. Moreover, 48 S. flexneri isolates contained the CTX-M gene, including 23 (40.4%) isolates containing CTX-M-1 group genes, 29 (50.9%) isolates with CTX-M-9 group genes, and 4 (7.0%) isolates containing both group genes. No expression of bla VIM and bla NDM was observed in the tested isolates. Furthermore, the majority of tested S. flexneri isolates was shown to harbor integrons, including 49 (86.0%) isolates that were shown to harbor class 1 integrons and 51 (89.4%) that harbored class 2 integrons (Table 4). In contrast to the S. flexneri isolates, all selected S. sonnei isolates were negative for bla OXA-1 expression, while only 38.9% expressed bla TEM ( Table 4). The expression rate of class 1 integrons detected in S. sonnei isolates (5.6%) was considerably lower than that in the S. flexneri isolates (86.0%), while the rate of isolates expressing class 2 integrons reached 83.3%. All 18 tested cephalosporin-resistant S. sonnei isolates were demonstrated to harbor bla CTX-M genes, including 14 (77.8%) isolates with CTX-M-1 group genes, nine (50%) isolates containing CTX-M-9 group genes, and six (33.3%) isolates carrying genes from both of these groups (Table 4). Among cephalosporin-resistant isolates, a higher frequency of isolates carrying both CTX-M-1 and CTX-M-9 group genes was detected among S. sonnei isolates, compared with that observed among S. flexneri isolates (χ 2 analysis, P < 0.05).
Moreover, in this study, we identified 86 S. flexneri isolates resistant to quinolones (levofloxacin or norfloxacin) ( Table 4). QRDR mutations were detected in three amino acids of the gyrA gene (Ser83Leu, Asp87Asn/Gly, His211Tyr) and one amino acid of the parC gene (Ser80-Ile). All 86 isolates were shown to harbor Ser83Leu mutations in the gyrA and Ser80Ile in the parC genes. Additionally, 62 isolates were shown to carry Asp87Asn mutation in the gyrA gene, 16 isolates had Asp87Gly mutation in the gyrA and 82 isolates had His211Tyr mutation in the gyrA gene. No point mutations in the gyrB and parE genes were observed. All 86 quinolone-resistant isolates were identified to carry at least three types of QRDR mutations, and 74 (86.0%) isolates were shown to harbor four types of QRDR mutations. In contrast, only three isolates were shown to harbor PMQR gene mutations (one in qnrB, two in qnrS), and no isolates were positive for acc(6 0 )-Ib-cr.
Notably, 17 Shigella isolates were resistant to cephalosporins and quinolones simultaneously (Table 5), and these were the MDR S. flexneri strains. Other than the resistance to cephalosporins and quinolones, the resistance to penicillins (17, 100%) was one of the most frequent types of resistance, followed by the resistance to tetracyclines (16,94

Discussion
Here, we demonstrated that the most frequent species of Shigella genus in Xinjiang in all years of our study was S. flexneri. Recently, the S. sonnei has shown increasing prevalence in China and even become the dominant species in southeast and central parts of China [36]. Furthermore, together with the economic growth in Xinjiang, an obvious rising trend in S. sonnei infections can be observed. The frequency of identified S. sonnei isolates altered from 6.7% in 2008 to 27.1% in 2014, which was, even considering the increase, shown to be below the average frequency of S. sonnei infections registered in China, where it was the dominant Shigella species (58.2% in 2011-2013) [36]. Various studies have concluded that S. sonnei prevalence increase with economic development [37]. As Xinjiang will continue to enhance its level of economy and sanitation, it is likely that S. sonnei species would become even more of a local public health concern. The design of vaccine candidate and shigellosis prevention strategy should consider this change of Shigella epidemics.  Considering serotypes, 2a and Xv were reported to be dominant S. flexneri serotypes in China [36], and 2a was reported to be dominant S. flexneri serotype in Xinjiang [27]. However, the results of our study further demonstrated that 2a serotype was still the most prevalent in Xinjiang but Xv serotype accounted for a smaller proportion of S. flexneri isolates, while 2a, 1b, 2b, and Xv represent the dominant serotypes detected between 2008 and 2014. Serotypes 1b and 2b are less frequent in China [36], but common in Pakistan [38], a country bordering Xinjiang. Therefore, the dominant S. flexneri serotypes distributed in Xinjiang likely represent the mixture of the major serotypes in China and adjacent countries.

Determinants of cephalosporin resistance
Furthermore, we detected atypical serotypes 1c, 2c, and 4s in Xinjiang. Serotype-based vaccines are currently under development, as a promising strategy against Shigella epidemics [39].  A future Shigella vaccine potentially applied in Xinjiang should be developed in accordance considering the characteristic Shigella serotypes in this region, and future surveillance studies in Xinjiang should pay close attention not only to the newly emerged serotypes but the predominant subgroups in surrounding regions as well, in order to prevent potential Shigella epidemics caused by bacteria with novel or imported O-antigen types. Shigella isolates demonstrated high levels of resistance to antibiotics. Both S. flexneri and S. sonnei revealed high AMR rate to some common used antibiotics like penicillins and tetracyclines. Thus, these older-generation drugs should not be contained in empirical therapy of shigellosis. Notably, 330 of 365 S. flexneri isolates and 59 of 93 S. sonnei isolates collected from Xinjiang showed MDR profiles. The high-level MDR frequency further restrict the choice of antibiotics in the clinical treatment of bacterial infections. Since the efficacy of older-generation antibiotics decreased due to the development of resistant strains, in a previous study, quinolones and third-generation cephalosporins were recommended as frontline antimicrobials for the empiric treatment of diarrhea-inducing pathogens [40]. However, in our study, 75 (16.38%) isolates were resistant to cephalosporin and 86 (18.77%) were resistant to norfloxacin. Additionally, 17 (3.71%) isolates were resistant to both cephalosporins and quinolones. Considering the currently used frontline antimicrobials, these resistance phenotypes threaten the effectiveness of therapy [41,42]. The mobility and dissemination of the resistant strains may increase in Xinjiang as it becomes an increasingly crucial region connecting China and other countries. Surveillance to Shigella in China revealed its high MDR frequency [7, 16,17,26,43,44]. The high-level resistant pathogen could migrate along with increasing trans-regional human activities. Therefore, novel preventive strategies are urgently required to prevent the spreading of AMR among Shigella strains. Clinicians prescribing anti-infective therapies should be more cautious, since the unreasonable use of antibiotics may further accelerate the accumulation and spread of AMR [45][46][47].
The AMR profile differed between S. flexneri and S. sonnei strains. Specifically, S. flexneri revealed higher MDR levels, demonstrating also some specific and important MDR phenotypes. A proportion of S. flexneri strains (37.3%) was shown to be resistant to the combination of commonly used antibiotics, including ampicillin, chloramphenicol, and trimethoprim/sulfamethoxazole, which was not observed among S. sonnei isolates. Quinolone resistance and even co-resistance to cephalosporin and quinolone emerged in S. flexneri strains, while S. sonnei demonstrated the sensitivity to quinolone. As a long-term predominant species in China [48], the traditional antibiotic-associated selection of S. flexneri has been underway for decades, and by frontline antibiotics in recent years, which led to the development of the MDR strains and the resistance to cephalosporin and quinolone. Therefore, our results imply that the antibiotic therapy of choice may differ between two Shigella species. Treatment of S. flexneri infections may be more complicated than that of S. sonnei, and a drug susceptibility test should be performed immediately after diagnosing a patient with S. flexneri infection. Antibiotic abuse should be more controlled, in order to reduce the selection pressure on S. flexneri strains, but the quinolone treatment may represent a safer antibiotic choice if the infective pathogens are identified as S. sonnei.
In this study, we further elucidated the genetic background and mechanisms underlying Shigella resistance to cephalosporins and quinolones. Cephalosporin-resistant S. flexneri were shown to frequently express bla CTX-M , bla OXA , and bla TEM , three main genes conferring the resistance to cephalosporin [49,50]. All cephalosporin-resistant S. sonnei expressed bla CTX-M as well. These genes were reported to be frequently encoded by plasmids [51][52][53][54][55], which facilitates the horizontal transfer of resistance to β-lactamase antibiotics [49,56]. Previous studies reported that class 1 and class 2 integrons may provide the resistance to other types of drugs and be responsible for the dissemination of AMR [35,57,58]. Almost 90% of the isolated S. flexneri strains were shown to harbor two integrons. The intl2 was also frequently identified in the S. sonnei isolates (83.3%). These results indicate that the cephalosporin resistance determinants can actively disseminate among Shigella cells or transferr within microflora. The PMQR genes were reported to be always located in mobile genetic elements such as plasmids [59]. Here, the presence of qnrS and qnrB genes was detected at a very low level in quinolone-resistant S. flexneri in Xinjiang, while a number of QRDR mutations were identified, indicating that the mutations in QRDRs primarily underlie the resistance to quinolone in Shigella isolates investigated here. Different QRDR mutations confer various levels of resistance [60][61][62]. Notably, all quinolone-resistant isolates were shown to harbor at least three QRDR mutations, showing that their simultaneous presence may underlie the observed increase in the resistance to quinolones.
We analyzed and presented here the prevalence of Shigella species and serotypes in Xinjiang, China. S. flexneri was shown to be the dominant Shigella species, with a unique dominant serotype pattern (2a, 1b, 2b, Xv), which represents a hybrid pattern comprising serotypes prevalent in adjacent regions. High levels of AMR were observed, especially by S. flexneri isolates. Emergence of frequently observed MDR and resistance to frontline antibiotics can severely restrict the choice of antibiotic therapy used for the treatment of Shigella infections. Since unsafe sanitation conditions remain present in this region, food-borne or water-borne shigellosis epidemic will remain a significant public health concern in future [63,64]. Therefore, the prevalence, trends, and AMR patterns of Shigella species and serotypes in Xinjiang should be closely monitored, and novel strategies are urgently required to prevent the spreading of the AMR among Shigella strains.