The Resistance Mechanisms and Clinical Impact of Resistance to the Third Generation Cephalosporins in Species of Enterobacter cloacae Complex in Taiwan

Enterobacter cloacae complex (ECC) is ubiquitous in the environment and is an important pathogen causing nosocomial infections. Because routine methods used in clinical laboratories cannot identify species within ECC, the clinical significance of each species within ECC is less known. We applied hsp60 gene sequencing to identify the species/clusters of ECC and detected β-lactamase genes and class 1 integrons with PCR for 184 clinical ECC isolates in Taiwan from 2013 to 2014 to investigate the clinical impact of species within ECC. The four most common clusters were E. hormaechei subsp. steigerwaltii (cluster VIII) (29.9%), E. hormaechei subsp. oharae (cluster VI) (20.1%), E. cloacae subsp. cloacae (cluster XI) (12%), and E. kobei (cluster II) (10.3%). E. hormaechei, which consisted of four clusters (clusters III, VI, VII, and VIII), is the predominant species and accounted for 57.1% of the isolates. The ceftazidime resistance rate was 27.2%, and the ceftriaxone resistance rate was 29.3%. Resistance to third generation cephalosporin was associated with a higher 30-day mortality rate. In total, 5 (2.7%), 24 (13.0%), and 1 (0.5%) isolates carried ESBL, AmpC, and carbapenemase genes, respectively. Class 1 integrons were present in 24.5% of the isolates, and most of the cassettes pertain to antibiotic resistance. Resistance to third generation cephalosporins, multidrug resistance, and class 1 integrons were significantly more in E. hormaechei (clusters III, VI, VII, and VIII) than in the other species. The 30-day mortality rate and 100-day mortality did not differ significantly between patients with E. hormaechei and those with infections with the other species. In conclusion, the distribution of third generation cephalosporin resistance, multidrug resistance, and class 1 integrons were uneven among Enterobacter species. The resistance to third generation cephalosporins possessed significant impact on patient outcome.


Species (n)
Cluster intI1 (+) n Gene Cassette Array of Class 1 Integrons (n) a These four isolates carry two class 1 integrons; b this isolate carries three class 1 integrons.

Clinical Features of Patients Infected with Enterobacter
The above results showed that resistance-associated characteristics such as third generation cephalosporin resistance and class 1 integrons were mostly present in E. hormaechei (clusters III, VI, VII, and VIII). Therefore, we further examined if there was any difference in clinical characteristics between infection with E. hormaechei (clusters III, VI, VII, and VIII) and other species/clusters of Enterobacter. Table S1 (Supplementary File S3) revealed that the main differences between these two groups were antimicrobial resistance-related factors, such as third generation cephalosporin resistance and class 1 integrons. The proportion of patients infected with E. hormaechei was lower than those infected with other clusters of Enterobacter for healthcare-associated infection and related to surgery. There were no statistically significant differences in the other demographic data, comorbidities, therapeutic devices and procedures, and clinical outcomes (30-day and 100-day mortality).
The clinical features and significance of susceptibility to third generation cephalosporins in Enterobacter are summarized in Table 5. Patients with Enterobacter resistant to third generation cephalosporins were significantly associated with higher percentages of underlying diseases of kidney disease, indwelling devices use, ICU admission, and class 1 integrons. Moreover, patients with Enterobacter resistant to the third generation cephalosporins were more likely to have a significantly higher 30-day mortality (OR: 6; 95% CI: 2.24-16.06) and 100-day mortality (OR: 5.74; 95% CI: 2.24-14.70) than those infected with Enterobacter susceptible/intermediate to third generation cephalosporins. Furthermore, we analyzed the clinical characteristics of Enterobacter infection caused by the four most common species/clusters in this study (E. hormaechei subsp. steigerwaltii, E. hormaechei subsp. oharae, E. cloacae subsp. cloacae, and E. kobei; clusters VIII, VI, XI, and II). Table S2 (Supplementary File S3) shows that there were significant differences in the clinical characteristics of the four clusters, which included gastrointestinal disease (p = 0.042), hemodialysis (p = 0.020), site of acquisition (hospital-acquired and community-acquired, p = 0.018), class 1 integrons (p = 0.010), and outcomes (30-day mortality, p = 0.016; 100-day mortality, p = 0.014). E. cloacae subsp. cloacae (cluster XI) occurred more frequently than the other three species in community-acquired infections (38.9%). In addition, the proportion of third generation cephalosporin resistant E. hormaechei subsp. oharae (cluster VI) strains was significantly higher than that of E. kobei (cluster II) (cluster VI vs. II, p = 0.028). A higher proportion of patients infected with E. cloacae subsp. cloacae (cluster XI) had poor outcomes in terms of 30-day mortality (XI, 33.3%; II, 0; VI, 18.2%; VIII, 8.5%) and 100-day mortality (XI, 33.3%; II, 0; VI, 21.2%; VIII, 8.5%) than those infected with the other three species.

Discussion
We aimed to investigate the clinical and microbiological characteristics of the species within E. cloacae complex (ECC) in this research. Our bacteria material was ECC isolates which were routinely identified from the clinical laboratory. Of the 184 isolates, 97.8% were classified into species and clusters based on hsp60 sequencing. However, three species other than previously defined ECC species and clusters were identified. It revealed the limitation of Enterobacter species identification with hsp60 sequencing. It was reported that determining taxonomic assignment using a single-gene-based approach may miss valuable information available from the rest of the genome and potentially lead to unreliable conclusions about taxonomic positions [4]. Given that the taxonomy of Enterobacter is complicated, we have added the nomenclature information with an updated classification and nomenclature of the genus Enterobacter using genome sequence-based analysis [4] for our isolates in Table 1.
The molecular epidemiology via PFGE revealed no large outbreak of Enterobacter due to specific clones in the Taiwan medical center. Under this background, we identified that the most common identified species/clusters in Taiwan are E. hormaechei subsp. steigerwaltii (cluster VIII) (29.9%), followed by E. hormaechei subsp. oharae (cluster VI) (20.1%), E. cloacae subsp. cloacae (cluster XI) (12%), and E. kobei (cluster II) (10.3%). Most other data for species distribution in ECC are from European countries. We summarize the distribution of different ECC species in different countries in Table S3 (Supplementary File S3). E. hormaechei subsp. steigerwaltii (cluster VIII) and E. hormaechei subsp. hoffmannii (cluster III) were the two most common clusters in Europe [5,6,26]. However, clusters VI and VIII accounted for most of the Taiwan isolates (50%), whereas E. hormaechei subsp. hoffmannii (cluster III) only accounted for 5.4%. This cluster distribution is similar to that in Guadeloupe where clusters VI and VIII accounted for 56.1% and cluster III was rare [11]. Though E. hormaechei subsp. hoffmannii (cluster III) was not common in clinical Enterobacter isolates in Taiwan, it was noteworthy that it is the most commonly identified species among carbapenem-nonsusceptible E. cloacae complex in Taiwan and in Southeast China [15,25]. Furthermore, clusters VI and VIII belong to a species named E. xiangfangensis as recommended by Wu et al. [4]. Most (63.5%) Enterobacter strains from human bloodstream infection in China are E. xiangfangensis [27]. According to the new nomenclature system, this species was also the most common Enterobacter species in our clinical isolates.
The lower rates of antibiotic resistance were observed among clinical Enterobacter isolates from Taiwan compared to those reported from Poland and Guadeloupe, including: amikacin (0% vs. 56 [11,28]. More than 70% of Enterobacter isolates resistant to third generation cephalosporins belonged to the four E. hormaechei clusters (III, VI, VII, and VIII). We observed similar findings with data from France and Guadeloupe in that E. hormaechei carried higher resistance rates to third generation cephalosporins when compared with other Enterobacter clusters [11,16].
In the study, only 2.7%, 13.0%%, and 0.5% of the Enterobacter isolates carried bla ESBL , bla AmpC , and carbapenemase genes, respectively ( Table 3). The percentages of bla ESBL and carbapenemase genes among Enterobacter strains from Taiwan (2.7% and 0.5%, respectively) were slightly lower than those from the United States (3% and 3%) [29], but far lower than those from Nepal (80.3% and 59.6%) [30]. That the rates with β-lactamase genes were lower than the resistance rate to third generation cephalosporins signifies that β-lactamase production partially contributed to the resistance to third generation cephalosporins and there may be other mechanisms of resistance to third generation cephalosporins such as efflux pumps, reduced permeability, and altered transpeptidases [31].
Class 1 integrons were found in 24.5% of the Enterobacter isolates, whereas 55% of the Enterobacter isolates in Poland carried class 1 integrons [28], which might be associated with the difference of antimicrobial resistance rates of Enterobacter in the two countries. In our surveillance, class 1 integrons are mostly distributed in three E. hormaechei clusters (clusters VI, III, and VIII). In Poland, class 1 integrons were found mostly in E. hormaechei subsp. steigerwaltii (cluster VIII), accounting for 81.6% of class 1 integron-positive strains [28]. In this study, resistant gene cassettes carried on class 1 integrons, such as dfrA, aadA, and aadB, were widespread in class 1 integrons, which agrees with previous studies [32][33][34][35].
Many studies have indicated that patients infected with third generation or broadspectrum cephalosporin resistant/nonsusceptible isolates, including Enterobacter spp. [19,40,41] had a worse clinical response, more days in hospital, a poorer outcome, and a higher mortality rate [19,[42][43][44][45] than those infected with susceptible isolates. Our study found that patients infected with third generation cephalosporin-resistant Enterobacter had higher 30-day and 100-day mortality rates than those infected with third generation cephalosporin susceptible/intermediate Enterobacter, though patients with Enterobacter resistant to third generation cephalosporins also had higher rates of kidney disease, indwelling devices use, and ICU admission.
In this study, we did not observe a significant difference between E. hormaechei and the other species in terms of demographic data, comorbidities, therapeutic devices and procedures, and clinical outcomes (30-day and 100-day mortality). With regard to the virulence of specific clusters, Liu et al. reported the virulence of cluster I strains was significantly higher than that of the other cluster strains according to the results of the Galleria mellonella infection model [46]. Cluster II (E. kobei) has strong biofilm formation ability under nutrientdeficient conditions but is associated with low virulence and pathogenicity [46]. However, the case number of cluster I in our study is too small to obtain enough clinical finding. Interestingly, we found the mortality rate to be zero for 18 cluster II patients in the study. Patients with E. cloacae subsp. Cloacae (cluster XI) had poor outcomes and had significantly higher 30-day mortality and 100-day mortality rates. The above suggests the Enterobacter species/cluster may have different clinical significance. However, the resistance to third generation cephalosporins clearly impacts the clinical outcome for Enterobacter infection.
Limitations of the research included (1) the fact that the bla ACT gene was not detected in some species, which might be due to the variations of nucleotides at the primer sequences for intrinsic and plasmid-mediated AmpC β-lactamase genes, subsequently leading to missed detection using PCR. Further research is needed. (2) We aimed to investigate the clinical and microbiological characteristics of the species within E. cloacae complex (ECC) in this research. Our bacteria material was ECC isolates which were identified using an automated system in a clinical laboratory, but this did not include all Enterobacter species. Therefore, the research findings apply to species in ECC but not all Enterobacter species.

Bacterial Isolates
A total of 184 consecutive Enterobacter isolates identified as E. cloacae complex with a VITEK 2 system were collected from Kaohsiung Medical University Hospital (KMUH), a 1720-bed medical center in Kaohsiung, Taiwan, from December 1, 2013, to June 14, 2014. The identification of bacterial isolates was performed using the VITEK 2 microbial identification system (bioMérieux, Hazelwood, MO, USA). Isolates were stored at −80 • C in GermBank stocks (CMP TM Culture Media, New Taipei City, Taiwan) until processing.

Antimicrobial Susceptibility Testing
Antimicrobial susceptibility was tested using the broth dilution method according to the guidelines of the Clinical and Laboratory Standards Institute (CLSI) [47]. The following antimicrobial agents were tested: ampicillin, amikacin, ceftazidime, cefmetazole, ceftriaxone, cefazolin, ertapenem, cefepime, gentamicin, levofloxacin, meropenem, ampicillin/sulbactam, sulfamethoxazole/trimethoprim, tigecycline, and piperacillin/tazobactam. Isolates resistant to at least one antimicrobial agent in three or more antimicrobial classes are defined as multidrug resistant isolates.

Species Identification of ECC Based on hsp60 Sequencing
Polymerase chain reaction (PCR) analysis for partial sequencing of the hsp60 gene was performed using primers; the conditions and protocol were as described previously [8]. A 341-bp fragment of the hsp60 gene was amplified and sequenced. A 272-bp fragment of the hsp60 gene was obtained for the 184 strains, and its sequence was analyzed using BLAST searches on the NCBI website against nucleotide databases. Sequences were analyzed using MEGA 11 software (version 11.0.13). The sequence of the fragment was compared to reference sequences from type strains previously described in taxonomic studies [3,8] using the ClustalW algorithm. The type strains were described previously [4][5][6]8,10]. The phylogenetic tree was constructed using neighbor-joining analysis. Thus, each isolate was assigned to its respective species and cluster.

Analysis of Class 1 Integrons and Gene Cassettes
PCR was used to detect the presence of class 1 integrons and to amplify class 1 integron cassettes as previously described [55,56]. Gene cassettes within the class 1 integrons were identified using nucleotide sequencing, and similarity searches of each gene with nucleotide sequences in the GenBank database were performed with the BLASTN program (https://blast.ncbi.nlm.nih.gov/Blast.cgi, accessed on 17 March 2022).

Pulsed Field Gel Electrophoresis (PFGE)
Clonal relatedness of Enterobacter isolates was determined using PFGE, which was performed according to a previously described protocol [57]. The restriction enzyme XbaI (New England Biolabs Inc., MA, USA) was used at the temperature suggested by the manufacturer. Restriction fragments were analyzed using GelCompar II software 6.5 (Applied Maths, Austin, TX, USA), and dendrograms of the patterns were constructed using the unweighted pair group method with the arithmetic mean based on the Dice similarity index. PFGE patterns were interpreted in accordance with the criteria of Tenover et al. [58]. Isolates with >85% similarity in PFGE banding patterns were designated as a pulsotype.

Analysis of Clinical Features of Patients Infected with ECC
This was a retrospective, observational study of patients with positive cultures of ECC from 1 December 2013, to 14 June 2014, at KUMH. Patients who underwent repeated sampling within 2 months, those infected with microorganisms other than Enterobacter, and those with incomplete medical records were excluded. A total of 161 patients were analyzed. Patient information was retrospectively retrieved from electronic medical records. The parameters included demographic data, comorbidities, therapeutic devices, and procedures (such as indwelling devices, hemodialysis, mechanical ventilation, and surgeries), exposure to drugs prior to isolation (steroids within 3 months, antimicrobials within 3 months and 2 weeks), sites of acquisition, and clinical outcomes. Sites of acquisition included hospital-acquired, community-acquired and healthcare-associated infections. Hospitalacquired infection was defined as an infection that occurred >48 h after admission to the hospital [19]. Community-acquired infection was defined as infection in patients undergoing outpatient treatment who had not been hospitalized or had not resided in a healthcare facility in the previous 3 months [44,59]. Healthcare-associated infection was defined as patients undergoing outpatient treatment who had been hospitalized or had resided in a healthcare facility in the previous 3 months. Clinical outcomes were assessed based on 30-day mortality or 100-day mortality from specimen collection.

Statistical Analyses
The chi-square test or Fisher exact test was used to compare categorical variables. Statistical significance was set at p < 0.05. All statistical analyses were performed using IBM SPSS AMOS 20.0 software.

Conclusions
In conclusion, third generation cephalosporin resistance, multidrug resistance and class 1 integrons are significantly higher in E. hormaechei (clusters III, VI, VII, and VIII), compared to the other species/clusters. Patients infected with third generation cephalosporinresistant Enterobacter have significantly higher 30-day mortality and 100-day mortality rates than those infected with Enterobacter susceptible/intermediate to third generation cephalosporins. Our findings on the unequal distribution of drug resistance profiles and class 1 integrons among Enterobacter species/cluster and the significant clinical impact of some species further emphasize the need for a larger scale investigation of the species of Enterobacter.
Supplementary Materials: The following are available online at https://www.mdpi.com/article/ 10.3390/antibiotics11091153/s1, File S1: Figure S1. The phylogenetic tree resulting from analysis of the hsp60 gene sequences of 184 Enterobacter isolates and previously reported sequences of type strains. File S2: The partial hsp60 sequences of 184 Enterobacter isolates. File S3: Table S1. Clinical characteristics and outcomes of cases infected with E. hormaechei (clusters III, VI, VII, and VIII) and the other Enterobacter species, Table S2. Clinical characteristics and outcomes of cases infected with the four most common species/clusters in this study, Table S3. Comparison of the distribution of Enterobacter species among ECC in different countries [4][5][6]8,10,26].