The Genotypic and Phenotypic Characteristics Contributing to Flomoxef Sensitivity in Clinical Isolates of ESBL-Producing E. coli Strains from Urinary Tract Infections

We carried out a molecular biological analysis of extended-spectrum β-lactamase (ESBL)-producing E. coli strains and their sensitivity to flomoxef (FMOX). Sequence type (ST) analysis by multilocus sequence typing (MLST) and classification of ESBL genotypes by multiplex PCR were performed on ESBL-producing E. coli strains isolated from urine samples collected from patients treated at our institution between 2008 and 2018. These sequences were compared with results for antimicrobial drug susceptibility determined using a micro-liquid dilution method. We also analyzed cases treated with FMOX at our institution to examine its clinical efficacy. Of the 911 E. coli strains identified, 158 (17.3%) were ESBL-producing. Of these, 67.7% (107/158) were strain ST-131 in ST analysis. Nearly all (154/158; 97.5%) were CTX-M genotypes, with M-14 and M-27 predominating. The isolated strains were sensitive to FMOX in drug susceptibility tests. Among the patient samples, 33 cases received FMOX, and of these, 5 had ESBL-producing E. coli. Among these five cases, three received FMOX for surgical prophylaxis as urinary carriers of ESBL-producing E. coli, and postoperative infections were prevented in all three patients. The other two patients received FMOX treatment for urinary tract infections. FMOX treatment was successful for one, and the other was switched to carbapenem. Our results suggest that FMOX has efficacy for perioperative prophylactic administration in urologic surgery involving carriers of ESBL-producing bacteria and for therapeutic administration for urinary tract infections. Use of FMOX avoids over-reliance on carbapenems or β-lactamase inhibitors and thus is an effective antimicrobial countermeasure.


Introduction
Escherichia coli (E.coli) is the most frequently detected pathogen in urinary tract infections (UTIs), accounting for 50% to 85% of cases [1,2].Furthermore, the frequency of bacteria producing extended-spectrum β-lactamase (ESBL) has recently been increasing worldwide, including in Japan, and antibiotic resistance is becoming a larger problem [3,4].The abuse of broad-spectrum antibiotics is one of the major causes of the development of antimicrobial-resistant bacteria.The problem of antimicrobial resistance has become a public threat.The frequency of occurrence of UTIs caused by ESBL-producing Enterobacterales has been increasing globally.Among ESBL-producing Enterobacterales, ESBL-producing E. coli is considered the greatest concern [4].
ESBL-producing bacteria are drug-resistant organisms that carry genetic mutations in the enzyme that degrades β-lactams by hydrolyzing them into cephalosporins and monobactams.The ESBL gene is encoded on a transmissible plasmid and is widely transmitted horizontally to mycobacteria of different Enterobacterales, leading to the spread of

E. coli Isolates and Antibiotic Sensitivity
Among the urine samples collected between 2008 and 2018, 911 E. coli strains were isolated from urine, of which 158 (17.3%) were ESBL-producing (Figure 1).Of these, nearly three-quarters (116/158; 73%) were from outpatients, and the remainder (42/158; 27%) were from inpatients.All 158 ESBL-producing E. coli strains were resistant to CEZ, all but one to CFDN, and all but two to CFPM (Table 1).On the other hand, all strains were susceptible to IPM and MEPM, all but one to FOM, all but two to FMOX, all but four to CMZ, and all but ten to PIPC/TAZ.Other strains that showed sensitivity to each drug were 133 (84.2%) for GM, 79 (50%) for CAZ, 66 (41.8%) for ST, and 25 (15.8%) for LVFX.

Figure 1.
Trends of E. coli isolated in urine from urinary tract infections.Among the urine samples collected between 2008 and 2018, E. coli strains were isolated from urine, of which 17.3% were ESBLproducing strains.Of these, nearly 73% were from outpatients, and the remainder, 27%, were from inpatients.Trends of E. coli isolated in urine from urinary tract infections.Among the urine samples collected between 2008 and 2018, E. coli strains were isolated from urine, of which 17.3% were ESBL-producing strains.Of these, nearly 73% were from outpatients, and the remainder, 27%, were from inpatients.(8):517-522.)[18]; ‡ : fBreakpoint is not defined about faropenem, sitafloxacin.

Multilocus Sequence Typing and β-Lactamase Gene PCR
Results of MLST analysis of the 158 ESBL-producing E.coli strains are shown in Figure 2. ST131 was the most commonly identified strain, with 107 out of 158 (67.7%), and the frequency increased over time.The next most common was ST38 with eight strains (5.1%), and the other strains each had fewer than eight within the panel.

The Clinical Efficacy of FMOX
The breakdown of cases treated with FMOX in our study is shown in Figure 7.Among the 33 cases that received FMOX at our institution (Figure 7), 14 were for prophylactic administration prior to surgery and 19 were for therapeutic administration for UTI.Of the 14 cases with prophylactic administration, 3 were switched to other agents due to postoperative infection, and of the 19 cases that received therapeutic administration, 6 were switched due to inadequate efficacy.
The breakdown of cases treated with FMOX in our study is shown in Figure 7.Among the 33 cases that received FMOX at our institution (Figure 7), 14 were for prophylactic administration prior to surgery and 19 were for therapeutic administration for UTI.Of the 14 cases with prophylactic administration, 3 were switched to other agents due to postoperative infection, and of the 19 cases that received therapeutic administration, 6 were switched due to inadequate efficacy.Of the 33 cases treated with FMOX at our institution (Table 3), 16 had E. coli, three had Proteus mirabilis, and one each had Klebsiella oxytoca, Citrobacter koseri, Pseudomonas aeruginosa, and MRSA.Two had other Gram-negative rods, two had Enterococcus spp, one had Gram-positive cocci, and three were culture-negative.There were a total of eight ESBL-producing bacteria identified: five E. coli and one each of Klebsiella oxytoca, Proteus mirabilis, and Citrobacter koseri (Table 4).For the five cases with E. coli, three received FMOX for surgical prophylaxis against urinary ESBL-producing E. coli, and all treatments successfully prevented postoperative infections.On the other hand, two patients received FMOX to treat UTIs: one patient had an effective response, and the other had an inadequate response and was switched to carbapenems.Of the 33 cases treated with FMOX at our institution (Table 3), 16 had E. coli, three had Proteus mirabilis, and one each had Klebsiella oxytoca, Citrobacter koseri, Pseudomonas aeruginosa, and MRSA.Two had other Gram-negative rods, two had Enterococcus spp, one had Gram-positive cocci, and three were culture-negative.There were a total of eight ESBLproducing bacteria identified: five E. coli and one each of Klebsiella oxytoca, Proteus mirabilis, and Citrobacter koseri (Table 4).For the five cases with E. coli, three received FMOX for surgical prophylaxis against urinary ESBL-producing E. coli, and all treatments successfully prevented postoperative infections.On the other hand, two patients received FMOX to treat UTIs: one patient had an effective response, and the other had an inadequate response and was switched to carbapenems.

ESBL-Producing E. coli
E. coli isolated from the urine of patients with UTI who were treated at the outpatient clinic or in a ward of Okayama University Hospital, Okayama, Japan, between January 2008 and December 2018 were included in our study.Pyuria was diagnosed as the presence of white blood cells (WBC) ≥5/HPF in a urine sediment specimen, and bacteriuria was diagnosed as a bacterial count ≥10 3 CFU/mL in catheter urine or 10 4 CFU/mL in midstream urine.Isolates were counted as one infection episode per patient, and plural isolates were counted in polymicrobial infections.
ESBL production of E. coli was confirmed using the disc method carried out according to Clinical and Laboratory Standards Institute (CLSI) document M100-S22 [21] with 30 µg/mL CTX.A total of 30 ESBL-producing E. coli strains were identified when the inhibition circle had a diameter larger than 5 mm using 30 µg/mL CTX, 30 µg/mL CTX plus 10 µg/mL CVA, 30 µg/mL CAZ, or 30 µg/mL CAZ plus 10 µg/mL CVA.
Sequence type analysis by multilocus sequence typing (MLST) and classification of ESBL genotypes by Multiplex PCR were performed on ESBL-producing E. coli strains.MLST is a molecular genetic method that involves the sequencing of seven housekeeping genes (adk, fumC, gyrB, icd, mdh, purA, and recA) to determine allele numbers that are then compared with information in enterobase (https://enterobase.warwick.ac.uk accessed on 1 March 2018).The sequencing type is determined by combining the seven types of genes (Table 5).Classification of ESBL genotypes by Multiplex PCR was performed according to the used primers as described by Dallenne et al. [22].Rapid DNA preparation was carried out by first placing one colony in 100 µL distilled water and heating at 95 • C. Total DNA (2 µL) was subjected to multiplex PCR in a 50 µL reaction mixture containing 1× PCR buffer (10 mM Tris-HCl, pH 8.3/50 mM KCl/1.5 mM MgCl 2 ), 200 mM of each deoxynucleotide triphosphate, a variable concentration of group-specific primers (Table 6), and 1 U Taq polymerase (Sigma Aldrich, St Quentin Fallavier, France).DNA was amplified with initial denaturation at 94 • C for 10 min followed by 30 cycles of 94 • C for 40 s, 60 • C for 40 s and 72 • C for 1 min, and a final 7-min elongation step at 72 • C. A 55 • C annealing temperature was used to amplify the bla VIM , bla IMP , and bla KPC genes, and 57 • C was used to amplify the bla GES and bla OXA-48 genes.Amplicons were visualized on a 2% agarose gel containing ethidium bromide run at 100 V for 1 h.A 100-bp DNA ladder (New England Biolabs, Ipswich, MA, USA) was used as a size marker.To identify the β-lactamase genes detected in the multiplex PCR assays, DNA sequence analyses of the amplicons were performed.Amplified PCR products were purified using the ExoSap purification kit (ExoSap-it, GE Healthcare, Piscataway, NJ, USA), and bidirectional sequencing was performed.Each sequence is aligned by multiple-sequence alignment using the BLAST program.The minimum inhibitory concentrations (MIC) of various antimicrobials were analyzed using micro-liquid dilution and agar dilution methods, and drug susceptibility was measured.Antimicrobial agents for which MICs were measured were cefazolin (CEZ), cefdinir (CFDN), flomoxef (FMOX), cefmetazole (CMZ), ceftazidime (CAZ), cefepime (CFPM), piperacillin/tazobactam (PIPC/TAZ), imipenem (IPM), faropenem (FRPM), meropenem (MEPM), levofloxacin (LVFX), sitafloxacin (STFX), fosfomycin (FOM), gentamicin (GM), sulfamethoxazole/trimethoprim (ST).

Flomoxef Efficacy
To investigate the efficacy of flomoxef (FMOX) against ESBL-producing E. coli, samples from 33 patients with UTI who were treated with FMOX at our institution between 2008 and 2018 were collected.The rationale for the choice of FMOX, urinary isolates, drug sensitivity, administration method, and outcome were retrospectively reviewed from patient charts to determine the clinical positioning of FMOX.

Statistical Analyses
Statistical analyses were performed using EZR software (version 1.71; Saitama Medical Center, Jichi Medical University, Saitama, Japan) [23].The antimicrobial susceptibilities among the strains were compared with Fisher's exact test.Results with p values < 0.05 were considered statistically significant.

Discussion
In our study, the susceptibility of ESBL-producing E. coli to FMOX was comparable to that seen for MEPM and TAZ/PIPC.Among the E. coli strains, ST131 was the most common sequence type (67.7%), and in our study population, the percentage increased annually.Consistent with previous reports, CTX-M was the most common genotype, and the frequency of CTX-M-27 increased over that of CTX-M-14 after 2013; CTX-M-15 appeared in 2012.The susceptibility rate of CTX-M-15 to FMOX, TAZ/PIPC, CMZ, CAZ, and GM was lower than that for other genotypes.
ESBL-producing E. coli were first reported by Knothe et al. in 1983 [24], and ESBLproducing E. coli were isolated in Japan by Ishii et al. in 1993 [25].TEM and SHV types of ESBL-producing E. coli were prevalent in Europe and the United States in the 1980s, and CTX-M types spread worldwide beginning in 2000; a pandemic of these types was declared in 2006 [5].Since 2010, CTX-M-27 and CTX-M-15 have been detected in Japan and abroad [26,27].These phenotypes tend to be more resistant to drugs than CTX-M-14, which was the mainstay of AMR [28].In our study, we also observed increased resistance rates to LVFX, CAZ, and GM by CTX-M-27 and CTX-M-15 compared to CTX-M-14, which is comparable to earlier findings.In addition, the frequencies of TEM and SHV were similar to those reported by Aihe et al. [29], but for TEM, the frequency was lower than that seen by Jena et al. [30].There has been almost no genotype analysis of UTIs in Japan, and such studies would be useful for better understanding the prevalence of multidrugresistant E. coli in the country.Moreover, since 2000, ST131 has become the predominant E. coli isolate worldwide [8], and our results were consistent with this trend.The ST131 group had a higher rate of insensitivity to LVFX and CAZ compared to non-ST131 strains (Figure 3).Among antimicrobial agents, fluoroquinolone resistance is increasing [31], and of the 158 isolates in this study, the majority (133/158; 84.2%) were resistant to LVFX.The most promising antimicrobial agents are carbapenems [32], and other antimicrobial agents, including TAZ/PIPC, CMZ, and FMOX, continue to have good activity [33,34].FOM is reported to be effective in the US and Europe [35], and the new drug TAZ/CTLZ is also expected to be effective [36].We observed a similar trend in our study.However, due to the risk of AMR, the use of carbapenems and TAZ/PIPC should be limited whenever possible, and CMZ and FMOX should be used with preference.
We also examined the clinical status of FMOX, and our findings suggest that FMOX has some efficacy for perioperative prophylactic administration in urologic surgery for carriers of ESBL-producing bacteria and for therapeutic administration for UTIs.FMOX is an oxacefem-derived antimicrobial agent that was first released in 1988 and is effective against Gram-negative rods, MSSA, and anaerobes, including E. coli.FMOX has also been reported to be effective against ESBL-producing bacteria [24] and ESBL-producing E. coli [34].In addition, FMOX has efficacy against CTX-M-type ESBL-producing bacteria with urinary excretion [16,37] and for UTIs caused by ESBL-producing E. coli in children [18].The MIC data for both IPN and MEPM look interesting and are as good as those for FMOX in the present study.In fact, in a clinical situation, a previous study showed that no significant differences were observed in the febrile period and recurrence rate between the FMOX-initiated group and the other antibiotic groups, suggesting that FMOX may be a useful alternative carbapenem antibiotic for UTIs caused by ESBL-producing E. coli [38].The cause of the change in the percentage of ESBLs within the observation period is unknown, but the effect of changes in therapeutic agents is less likely, and no severe cases were observed in the current reports.We think that FMOX would be a suitable alternative for perioperative prophylactic administration because of its narrow-spectrum antimicrobial feature, therapeutic administration for UTIs caused by ESBL-producing bacteria of moderate or lower severity, and de-escalation from empiric therapy against ESBL-producing bacteria.This approach would allow for more limited use of carbapenem and combination β-lactamase inhibitors as part of AMR countermeasures.Uropathogenic E. coli strains have a variety of flexible and adaptive genomic pools.These genomic pools are the arsenal of a wide range of virulence and drug resistance genes.Therefore, the armed uropathogenic E. coli pathotypes with a powerful virulome enable them to survive against the human body's immune system and the deathly effects of antimicrobial agents.However, there is neither a regular pattern for drug resistance properties nor ESBL production in reported isolated pathotypes of Uropathogenic E. coli worldwide.For example, the presence or absence of mrk genes has no correlation with antimicrobial susceptibility in uropathogenic E. coli pathotypes [39].On the other hand, the presence of hlyA and cnf1 virulence genes may lead to increased bacterial sensitivity to fluoroquinolones [40].The gene profiles may be helpful for choosing the appropriate antibiotic for determined uropathogenic E. coli strains in the procedure for the treatment of UTIs.Further study is needed to determine the sensitivity of FOMX from the perspective of gene profiles.
The present study had some limitations.First, this study was done in a single hospital in Japan, and due to budget limitations, a multicenter study was not possible.Access to medical data of the patients, such as underlying patients characteristics, diseases, and recent medications, to correlate their links with colonization of E. coli strains and their antimicrobial resistance phenotypes was not possible, because of the lack of registry system.In addition, there was no follow-up program at the time of the study to understand differences in the success or complications of UTIs caused by E. coli identified in our hospital.Although our strain pools could not represent the entire ESBL-producing E. coli population in this region, both temporally and geographically, the data generated in the present study provided an insight into the extent of β-lactam resistance among ESBL producers associated with UTIs in Japan.Moreover, the susceptibility data obtained in this study support further consideration of the potential use of FMOX as an alternative option for the treatment of UTIs caused by ESBL-producing E. coli.Second, the major drawback of the present study was that the analyzed isolates were collected from 2008 to 2018 and did not reflect the current situation well.However, it is meaningful work to provide the basic molecular characteristics of a common causative pathogen for community-based UTIs.Further studies on changes in microbiological characteristics in Japan are necessary to clarify this issue.
Despite these limitations, our results showed ESBL-producing E. coli isolates with common rep-types presented diversity in their clone types and antibiotic-resistance patterns.Constant monitoring would be done to control their spread in the community because of the high prevalence of the strains and their involvement in UTIs.

Conclusions
The frequency of ESBL-producing E. coli is increasing annually, especially in outpatients.Although antimicrobial susceptibility has not significantly decreased, we have continued to observe E. coli strains like CTX-M-15 that have a multidrug-resistant genotype.Throughout the study period, favorable sensitivity to FMOX was observed, suggesting that FMOX continues to be a valid treatment option that can prevent AMR.We think that the continuous surveillance of ESBL-producing E. coli at both the institutional and national levels will preserve the clinical efficacy of broad-spectrum cephalosporin and prevent the emergence of carbapenem resistance in local E. coli strains.

Figure 1 .
Figure 1.Trends of E. coli isolated in urine from urinary tract infections.Among the urine samples collected between 2008 and 2018, E. coli strains were isolated from urine, of which 17.3% were ESBL-producing strains.Of these, nearly 73% were from outpatients, and the remainder, 27%, were from inpatients.

Figure 2 .
Figure 2. MLST analysis of ESBL-producing E. coli strains.ST131 was the most commonly identified strain with 67.7%, and the frequency increased over time.The next most common was ST38, with 5.1%.

Figure 2 .
Figure 2. MLST analysis of ESBL-producing E. coli strains.ST131 was the most commonly identified strain with 67.7%, and the frequency increased over time.The next most common was ST38, with 5.1%.

Figure 4 .
Figure 4. Genotype.In genotyping using multiplex PCR, the CTX-M type was the most common genotype, accounting for 97.5%, followed by the TEM type with 32.9%.However, only 1.3% were SHV types.

Figure 4 .
Figure 4. Genotype.In genotyping using multiplex PCR, the CTX-M type was the most common genotype, accounting for 97.5%, followed by the TEM type with 32.9%.However, only 1.3% were SHV types.

Figure 4 .
Figure 4. Genotype.In genotyping using multiplex PCR, the CTX-M type was the most common genotype, accounting for 97.5%, followed by the TEM type with 32.9%.However, only 1.3% were SHV types.The classification of the CTX-M type is shown in Figure 5.For the classification of the CTX-M type, the CTX-M-9 group members CTX-M-14 type and CTX-M-27 type predominated among the study samples.The CTX-M-14 type was predominant in 54 of 154 (35.1%) strains, and the CTX-M-27 type was seen in 72 of 154 (46.8%) strains.The CTX-M-14 type predominated through 2013, while the CTX-M-27 type predominated in 2014.The next most common type was CTX-M-15, which belongs to the CTX-M-1 group, with 12 of 154 (7.8%).

Figure 5 .
Figure 5.The classification of the CTM-M type.For the classification of the CTX-M type, the CTX-M-9 group members, CTX-M-14 type, and the CTX-M-27 type predominated among the study samples.CTX-M-14 type was predominant in 35.1% of strains, and CTX-M-27 type was seen in 46.8% of strains.The CTX-M-14 type predominated through 2013, and the CTX-M-27 type came to dominate in 2014.The next most common type was CTX-M-15, which belongs to the CTX-M-1 group, with 7.8%.

Figure 5 .
Figure 5.The classification of the CTM-M type.For the classification of the CTX-M type, the CTX-M-9 group members, CTX-M-14 type, and the CTX-M-27 type predominated among the study samples.CTX-M-14 type was predominant in 35.1% of strains, and CTX-M-27 type was seen in 46.8% of strains.The CTX-M-14 type predominated through 2013, and the CTX-M-27 type came to dominate in 2014.The next most common type was CTX-M-15, which belongs to the CTX-M-1 group, with 7.8%.

Figure 7 .
Figure 7.The breakdown of cases treated with FMOX.Fourteen were for prophylactic administration prior to surgery, and 19 were for therapeutic administration for UTI.Of the 14 cases with prophylactic administration, 3 were switched to other agents due to postoperative infection, and of the 19 cases that received therapeutic administration, 6 were switched due to inadequate efficacy.

Figure 7 .
Figure 7.The breakdown of cases treated with FMOX.Fourteen were for prophylactic administration prior to surgery, and 19 were for therapeutic administration for UTI.Of the 14 cases with prophylactic administration, 3 were switched to other agents due to postoperative infection, and of the 19 cases that received therapeutic administration, 6 were switched due to inadequate efficacy.

Table 1 .
Antibiotic susceptibility to different antibiotic agents.

Table 1 .
Antibiotic susceptibility to different antibiotic agents.

Table 2 .
Two ESBL-producing Escherichia coli strains that are resistant to FMOX.

Table 3 .
Bacteria types were identified in urine samples collected before administration of FMOX.

Table 4 .
Outcomes for treatment with FMOX against ESBL-producing bacteria.

Table 5 .
Primers and reaction conditions used for MLST.

Table 6 .
Group-specific primers used for the assays.