Hyunsoo Kim1, Young Ah Kim2, Young Hee Seo3, Hyukmin Lee3,4, Kyungwon Lee3,4
1Department of Laboratory Medicine, National Police Hospital, Seoul, 2Department of Laboratory Medicine, National Health Insurance Service Ilsan Hospital, Goyang, 3Research Institute of Bacterial Resistance and Department of Laboratory Medicine, Yonsei University College of Medicine, Seoul, 4Seoul Clinical Laboratories, Yongin, Korea
Corresponding to Young Ah Kim, E-mail: yakim@nhimc.or.kr
Ann Clin Microbiol 2022;25(4):119-124. https://doi.org/10.5145/ACM.2022.25.4.2
Received on 13 July 2022, Revised on 6 September 2022, Accepted on 8 September 2022, Published on 20 December 2022.
Copyright © Korean Society of Clinical Microbiology.
This is an Open Access article distributed under the terms of the Creative Commons Attribution Non-Commercial License (https://creativecommons.org/licenses/by-nc/4.0/) which permits unrestricted non-commercial use, distribution, and reproduction in any medium, provided the original work is properly cited.
Background: The application of genotypic antimicrobial sensitivity tests (ASTs) is dependent on the reliability of the predictions of phenotypic resistance. In this study, routine AST results and the presence of corresponding antimicrobial resistance genes were compared.
Methods: Eighty-four extended-spectrum-β-lactamase-producing Escherichia coli isolates from poultry-related samples were included in the study. The disk diffusion method was used to test for susceptibility to antimicrobial compounds, except colistin susceptibility, which was tested using the agar dilution method. Whole-genome sequencing (WGS) was performed using a NextSeq 550 instrument (Illumina, USA). Antimicrobial resistance genes were detected using ResFinder 4.1.
Results: Concordance rates between the genotype and phenotype ranged from 35.7% (ciprofloxacin) to 96.4% (tetracycline). The presence of tet was a good predictor of phenotypic resistance.
Conclusion: The genotype was a good predictor of tetracycline phenotypic resistance, but there was a gap in the prediction of phenotypic ASTs for trimethoprim-sulfamethoxazole, chloramphenicol, gentamicin, and ciprofloxacin. We concluded that WGS-based genotypic ASTs are inadequate to replace routine phenotypic ASTs.
Antimicrobial resistance, Phenotype, Genotype, Whole genome sequencing, Escherichia coli
Accurate antimicrobial susceptibility of pathogen is very important to apply effective antimicrobials to infected patients. Antimicrobial susceptibility testing (AST) is routinely based on the phenotypic method, which needs overnight incubation and tight control of experiments [1]. Recently, rapid genotypic AST has been proposed with introduction of various molecular methods [2].
Polymerase chain reaction (PCR)-based molecular AST assay usually detects well-selected specific antimicrobial resistance (AMR) genes and whole genome sequencing (WGS) detects a wide range of AMR genes, both of which can detect indirect surrogate markers for phenotypic AST [3]. However, genotypic AST might be influenced by the level of gene expression. There could be a discrepancy between genotype and phenotype AST. Functional ability is largely dependent on the AMR gene database in analytical programs [4]. Therefore, the reliability of predicting phenotypes should be evaluated before genotypic AST is applied.
In this study, we compared routine phenotypic AST with the presence of corresponding AMR genes or mutation detected by WGS. The purpose of this study was to provide an overview about the reliability of genotypic AST.
A total of 84 extended-spectrum-β-lactamase producing Escherichia coli (ESBL-EC) were included in this study, isolated from poultry, poultry farm environment, or workers from January to August 2019 during the project collaborated with the Korea Disease Control and Prevention Agency [5]. Species identification was performed by a MALDI Biotyper (Bruker Daltonik, Bremen, Germany). The disk diffusion method was used for antimicrobial susceptibility of cefotaxime, ceftazidime, cefepime, cefoxitin, aztreonam, imipenem, meropenem, ertapenem, amikacin, gentamicin, ciprofloxacin, trimethoprim-sulfamethoxazole, tetracycline, tigecycline, chloramphenicol, and nitrofurantoin. The diameter of inhibition zone was interpreted according to Clinical and Laboratory Standards Institute (CLSI) criteria [6]. To detect colistin-resistant isolates, test organisms were screened on Mueller-Hinton agar (Oxoid, Basingstoke, UK) containing colistin (0, 1, 2, and 4 μg/mL) using E. coli ATCC25922 strain as an internal control. If minimal inhibition concentration was > 2 µg/mL, the isolate was regarded as colistin-resistant organism according to CLSI breakpoints for Pseudomonas aeruginosa and Acinetobacter spp. because there were no CLSI breakpoints for Enterobacteriaceae [6].
ESBL production was confirmed by PCR and sequencing of ESBL genes (blaTEM, blaSHV, and blaCTX-M) for any isolate showing resistance to cefotaxime or ceftazidime as described in a previous study [7]. For WGS, DNAs of freshly sub-cultured isolates were extracted using a GenElute™ Bacterial Genomic DNA Kit (Sigma-Aldrich, St. Louis, MO, USA) and 8 μg of input genomic DNA was used. Entire genomes of ESBL-EC isolates were sequenced using a NextSeq 550 instrument (Illumina, San Diego, CA, USA). Sequences were assembled with Spades (version 3.11.1) and annotated with Prokka (version 1.13.7). Data of antimicrobial resistance genes were obtained from the website of Center for Genomic Epidemiology [8], including ResFinder 4.1 with 90% ID threshold and 60% minimal length [9].
In this analysis, carbapenem, amikacin, tigecycline, nitrofurantoin, and colistin were excluded because nearly all isolates were susceptible to these antimicrobials except that one isolate showed resistance to nitrofurantoin. The presence of resistance genes was compared with results of phenotypic antimicrobial susceptibility test. When a related resistance gene or mutation was present in a phenotypic resistant isolate, the isolate was defined as a concordant isolate. When multiple resistance genes were involved in phenotypic resistance, any of resistance genes was regarded as a possible gene.
Resistance genes for corresponding antimicrobial phenotypes were detected, including β-lactam (blaOXA-1, blaTEM-1, blaCMY-2, blaCTX-M-1, blaCTX-M-14, blaCTX-M-15, blaCTX-M-27, blaCTX-M-55, and blaCTX-M-65), aminoglycoside (aadA1, aadA2, aadA5, aadA12, aac(6’)Ib-cr, aac(3’)-IIa, aac(3’)-IId, aac(3’)-IIe, aac(3’)-IVa, aph(3’)Ia , aph(4)-Ia), quinolone (aac(6’)Ib-cr, qnrB19, qnrS1, qnrS2), trimethoprim-sulfomethoxazole (dfrA1, dfrA12, dfrA14, dfrA17, sul1, sul2, sul3), tetracycline (tet(A), tet(B)), and chloramphenicol (catA1, catB3). Chromosomal mutations of parC, parE, gyrA, and gyrB for ciprofloxacin resistance were also searched.
Concordance rates ranged from 35.7% (ciprofloxacin) to 52.4% (trimethoprim-sulfamethoxazole), 51.2% (chloramphenicol and gentamicin), and 96.4% (tetracycline) when any related resistance gene was considered to be responsible for the resistance phenotype.
The effect of individual resistance gene was summarized as concordance rate (Table 1). The presence of tet, sul plus dfr, cat, qnrB19, and qnrS2 well predicted phenotypic resistance.
All ESBL-EC isolates were resistant to cefotaxime but susceptible to cefoxitin. However, susceptibilities to ceftazidime, cefepime, and aztreonam were different according to CTX-M type. CTX-M-55 producers showed high resistance rates to these three lactams (Table 2).
Table 1. Concordance rates between phenotypic antimicrobial susceptibility and genotype with the presence of resistance genes or mutation
Concordance rates: % (n/n)* | ||||
---|---|---|---|---|
GM-R | CP-R | SXT-R | TET-R | CIP-R |
aac(3′)-IIa: 100 (1/1) | catA1: 100% (2/2) | sul only: 3 (1/35) | tet(A): 99 (82/83) | parC: 50 (1/2) |
aac(3)-IId: 100 (2/2) | catB3: 100% (3/3) | sul2: 0 (0/32) | tet(A)+tet(B): 100 (1/1) | qnrB19: 100 (2/2) |
aac(3)-IId+aadA2: 67 (2/3) | sul3: 50 (1/2) | qnrS1: 0 (0/3) | ||
aac(3)-IId+aadA5: 100 (1/1) | sul1+sul2: 0 (0/1) | qnrS2: 100 (1/1) | ||
aac(3)-IIe: 50 (1/2) | dfr only: 33 (1/3) | qnrS2+aac(6′)Ib-cr: 100 (3/3) | ||
aac(3)-Iva+aadA1+aadA2 | dfrA12: 50 (1/2) | aac(6′)Ib-cr: 0 (0/1) | ||
+aph(4)-Ia: 100 (1/1) | dfrA14: 0 (0/1) | |||
aac(3)-Iva+aadA2+aph(4)-Ia; 33 (1/3) | sul+dfr: 88 (29/33) | |||
aac(3)-Iva+aph(4)-Iva: 27 (4/15) | sul1+dfrA17: 100 (1/1) | |||
aac(3)-VIa+aadA1:100 (1/1) | sul2+dfrA1: 100 (1/1) | |||
aac(6′)Ib-cr: 0 (0/1) | sul2+dfrA12: 100 (1/1) | |||
aac(6′)Ib-cr+aadA5: 0 (0/3) | sul2+dfrA14: 100 (5/5) | |||
aadA1: 0 (0/3) | sul2+dfrA17: 100 (5/5) | |||
aadA1+aadA5: 0 (0/1) | sul3+dfrA17: 100 (1/1) | |||
aadA2: 14 (1/7) | sul1+sul2+dfrA12: 75 (9/12) | |||
aadA5: 0 (0/7) | sul1+sul2+dfrA14+drfA17: 50 (1/2) | |||
aadA5+aph(3′)-Ia: 0 (0/1) | sul1+sul2+dfrA17: 100 (5/5) | |||
aadA12: 0 (0/2) | ||||
aph(3′)-Ia: 0 (0/1) |
*The number of isolates with phenotypical resistance/the number of isolates with resistance genes or mutations. TET-R tet(A): 99 (82/83) tet(A)+tet(B): 100 (1/1) CIP-R parC: 50 (1/2) qnrB19: 100 (2/2) qnrS1: 0 (0/3) qnrS2: 100 (1/1) qnrS2+aac(6′)Ib-cr: 100 (3/3) aac(6′)Ib-cr: 0 (0/1) Abbreviations: GM-R, phenotypic resistance to gentamicin; CP-R, phenotypic resistance to chrolamphenicol; SXT-R, phenotypic resistance to trimethoprimsulfamethoxazole; TET-R, phenotypic resistance to tetracycline; CIP-R, phenotypic resistance to ciprofloxacin.
Table 2. Concordance rates between resistance phenotype to cephalosporins and genotypes of CTX-M and PACBL
CTX-M type | CTX-R % (n) | CAZ-R % (n) | CEF-R % (n) | FOX-R % (n) | AZT-R % (n) |
---|---|---|---|---|---|
CTX-M-1 (n=11) | 100 (11) | 0 | 18 (2) | 0 | 18 (2) |
CTX-M-14 (n=29) | 100 (29) | 0 | 4 (1) | 0 | 0 |
CTX-M-15 (n=8) | 100 (8) | 25 (2) | 88 (7) | 0 | 100 (8) |
CTX-M-27 (n=3) | 100 (3) | 0 | 67 (2) | 0 | 67 (2) |
CTX-M-55 (n=24) | 100 (24) | 42 (10) | 75 (18) | 0 | 79 (19) |
CMY-2 plus CTX-M-55 (n=4) | 100 (4) | 100 (4) | 0 | 100 (4) | 100 (4) |
CTX-M-65 (n=5) | 100 (5) | 0 | 0 | 0 | 20 (1) |
Abbreviations: PACBL, plasmid-mediated AmpC-like β-lactamase; CTX-R, phenotypic resistance to cefotaxime; CAZ-R, phenotypic resistance to cefotazidime; CEF-R, phenotypic resistance to cefepime; FOX-R, phenotypic resistance to cefoxitin; AZT-R, phenotypic resistance to aztreonam.
To choose effective antimicrobials for patients with bacterial infection and survey resistant organisms, AST is very important [1]. Genotypic AST can be used to detect corresponding genes for antimicrobial resistance. It is applicable in a rapid manner without overnight incubation, which is usually needed in a phenotypic AST.
In the present study, the concordance between genotype and phenotype was not very good except for resistance to tetracycline. Low concordance rates were noted for resistance to trimethoprim-sulfamethoxazole, chloramphenicol, gentamicin, and ciprofloxacin, which seemed to be mostly due to small number of cases. Others might be following reasons. First, detected multidrug efflux pump genes such as floR, oqxA, and oqxB were not included in comparison because their substrate specificity was not fully understood yet [10,11]. Second, the existence of genes was not equal to activity because gene expression levels might be different. Finally, there are many unknown resistance mechanisms and resistance genes. In this study, the possibility of unknown resistance genes was high for chloramphenicol-resistant or ciprofloxacin-resistant isolates because verified resistance genes were not detected in many of these isolates.
Bortolaia et al.[12] have reported that the concordance from 1,520 observations including 16 antimicrobials is 97%, ranging from 71.6% for cefepime and 100% for most antimicrobials in E. coli. Tyson et al.[13] have reported a specificity of 97.8% and a sensitivity of 99.6% for over 30 resistance genes and a number of resistance mutations in E. coli. Recently, Golden et al.[14] have reported high categorical agreements for ciprofloxacin, gentamicin, ceftriaxone, and trimethoprim/sulfamethoxazole in the evaluation of a total of 671 E. coli isolates. This study used a different definition, in which genotype and phenotype were determined to agree when an isolate was phenotypically non-susceptible and possessed known resistance genes or mutations, or when the same resistance genes or mutations were absent in a phenotypically susceptible isolate [14].
Our result showed that the presence of tet, sul plus dfr, cat, qnrB19, and qnrS2 seemed to well predict phenotypic resistance, although the number of cases was limited. For aminoglycoside modifying enzymes, it was hard to find a tendency due to a small number of type-specific isolates. In this study, phenotypically resistance rather than non-susceptibility was used, which could contribute to the difference of the results in comparison of previous studies. The definition was used to evaluate the clear correlation between antimicrobial resistance genes and resistance phenotype to simply rule out the treatment options. E. coli is a major etiologic agents of urinary tract infection and effective antimicrobial therapy is possible with the ʻintermediate’ in susceptibility test [15].
The concordance rates between genotypes of CTX-M and cefotaxime resistance was very high as wellknown [7], ceftazidime resistance varied according to the CTX-M types. CTX-M-55-producing isolates showed higher ceftazidime resistance rates than other types (Table 2).
In this study, genotype well predicted tetracycline phenotypic resistance, but there was a gap in prediction of phenotypic AST of trimethoprim-sulfamethoxazole, chloramphenicol, gentamicin, and ciprofloxacin. We concluded that WGS-based genotypic AST is inadequate in replacement of routine phenotypic AST. Further study is needed to determine the effectiveness of WGS in identifying resistance genotypes of multidrugresistant E. coli and whether these genotypes correlate with observed phenotypes.
This study was approved by the Institutional Review Board of National Health Insurance Service Ilsan Hospital, Goyang, Korea as required by hospital policy (IRB No. NHIMC 2022-07-018).
No potential conflicts of interest relevant to this article were reported.
None.
1. Hendriksen RS, Seyfarth AM, Jensen AB, Whichard J, Karlsmose S, Joyce K, et al. Results of use of WHO Global Salm-Surv external quality assurance system for antimicrobial susceptibility testing of Salmonella isolates from 2000 to 2007. J Clin Microbiol 2009;47:79-85.
2. Fluit AC, Visser MR, Schmitz F. Molecular detection of antimicrobial resistance. Clin Microbiol Rev 2001;14:836–71.
3. Boolchandani M, D’Souza AW, Dantas G. Sequencing-based methods and resources to study antimicrobial resistance. Nat Rev Genet 2019;20:356-70.
4. Feldgarden M, Brover V, Gonzalez-Escalona N, Frye JG, Haendiges J, Haft DH, et al. AMRFinderPlus and the Reference Gene Catalog facilitate examination of the genomic links among antimicrobial resistance, stress response, and virulence. Sci Rep 2021;11:12728.
5. Kim H, Kim YA, Seo YH, Lee H, Lee K. Prevalence and molecular epidemiology of extendedspectrum-β-lactamase (ESBL)-producing Escherichia coli from multiple sectors of poultry industry in Korea. Antibiotics 2021;10:1050.
6. Clinical and Laboratory Standards Institute (CLSI). Performance standards for antimicrobial susceptibility testing; Supplement M100. Wayne; PA: 2020.
7. Sidjabat HE, Paterson DL, Adams-Haduch JM, Ewan L, Pasculle AW, Muto CA, et al. Molecular epidemiology of CTX-M-producing Escherichia coli isolates at a tertiary medical center in western Pennsylvania. Antimicrob Agents Chemother 2009;53:4733–9.
8. Center for Genomic Epidemiology. www.genomicepidemiology.org [Online] (last visited on 10 February 2022).
9. Florensa AF, Kaas RS, Clausen PTLC, Aytan-Aktug D, Aarestrup FM. ResFinder – an open online resource for identification of antimicrobial resistance genes in next-generation sequencing data and prediction of phenotypes from genotypes. Microb Genom 2022;8:000748.
10. Braibant M, Chevalier J, Chaslus-Dancla E, Pagès JM, Cloeckaert A. Structural and functional study of the phenicol-specific efflux pump FloR belonging to the major facilitator superfamily. Antimicrob Agents Chemother 2005:49:2965–71.
11. Li J, Zhang H, Ning J, Sajid A, Cheng G, Yuan Z, et al. The nature and epidemiology of OqxAB, a multidrug efflux pump. Antimicrob Resist Infect Control 2019;8:44
12. Bortolaia V, Kaas RS, Ruppe E, Roberts MC, Schwarz S, Cattoir V, et al. ResFinder 4.0 for predictions of phenotypes from genotypes. J Antimicrob Chemother 2020;75:3491-500.
13. Tyson GH, McDermott PF, Li C, Chen Y, Tadesse DA, Mukherjee S, et al. WGS accurately predicts antimicrobial resistance in Escherichia coli. J Antimicrob Chemother 2015;70:2763-9.
14. Golden AR, Karlowsky JA, Walkty A, Baxter MR, Denisuik AJ, McCracken M, et al. Comparison of phenotypic antimicrobial susceptibility testing results and WGS-derived genotypic resistance profiles for a cohort of ESBL-producing Escherichia coli collected from Canadian hospitals: CANWARD 2007–18. J Antimicrob Chemother 2021;76:2825–32.
15. Bunnell KL, Wenzler E, Harrington AT, Danziger LH. Impact of Clinical and Laboratory Standards Institute breakpoint changes on susceptibility rates of cephalosporins in uncomplicated urinary tract infections caused by Enterobacteriaceae. Diagn Microbiol Infect Dis 2018;90:335-6.
1. Hendriksen RS, Seyfarth AM, Jensen AB, Whichard J, Karlsmose S, Joyce K, et al. Results of use of WHO Global Salm-Surv external quality assurance system for antimicrobial susceptibility testing of Salmonella isolates from 2000 to 2007. J Clin Microbiol 2009;47:79-85.
2. Fluit AC, Visser MR, Schmitz F. Molecular detection of antimicrobial resistance. Clin Microbiol Rev 2001;14:836–71.
3. Boolchandani M, D’Souza AW, Dantas G. Sequencing-based methods and resources to study antimicrobial resistance. Nat Rev Genet 2019;20:356-70.
4. Feldgarden M, Brover V, Gonzalez-Escalona N, Frye JG, Haendiges J, Haft DH, et al. AMRFinderPlus and the Reference Gene Catalog facilitate examination of the genomic links among antimicrobial resistance, stress response, and virulence. Sci Rep 2021;11:12728.
5. Kim H, Kim YA, Seo YH, Lee H, Lee K. Prevalence and molecular epidemiology of extendedspectrum-β-lactamase (ESBL)-producing Escherichia coli from multiple sectors of poultry industry in Korea. Antibiotics 2021;10:1050.
6. Clinical and Laboratory Standards Institute (CLSI). Performance standards for antimicrobial susceptibility testing; Supplement M100. Wayne; PA: 2020.
7. Sidjabat HE, Paterson DL, Adams-Haduch JM, Ewan L, Pasculle AW, Muto CA, et al. Molecular epidemiology of CTX-M-producing Escherichia coli isolates at a tertiary medical center in western Pennsylvania. Antimicrob Agents Chemother 2009;53:4733–9.
8. Center for Genomic Epidemiology. www.genomicepidemiology.org [Online] (last visited on 10 February 2022).
9. Florensa AF, Kaas RS, Clausen PTLC, Aytan-Aktug D, Aarestrup FM. ResFinder – an open online resource for identification of antimicrobial resistance genes in next-generation sequencing data and prediction of phenotypes from genotypes. Microb Genom 2022;8:000748.
10. Braibant M, Chevalier J, Chaslus-Dancla E, Pagès JM, Cloeckaert A. Structural and functional study of the phenicol-specific efflux pump FloR belonging to the major facilitator superfamily. Antimicrob Agents Chemother 2005:49:2965–71.
11. Li J, Zhang H, Ning J, Sajid A, Cheng G, Yuan Z, et al. The nature and epidemiology of OqxAB, a multidrug efflux pump. Antimicrob Resist Infect Control 2019;8:44
12. Bortolaia V, Kaas RS, Ruppe E, Roberts MC, Schwarz S, Cattoir V, et al. ResFinder 4.0 for predictions of phenotypes from genotypes. J Antimicrob Chemother 2020;75:3491-500.
13. Tyson GH, McDermott PF, Li C, Chen Y, Tadesse DA, Mukherjee S, et al. WGS accurately predicts antimicrobial resistance in Escherichia coli. J Antimicrob Chemother 2015;70:2763-9.
14. Golden AR, Karlowsky JA, Walkty A, Baxter MR, Denisuik AJ, McCracken M, et al. Comparison of phenotypic antimicrobial susceptibility testing results and WGS-derived genotypic resistance profiles for a cohort of ESBL-producing Escherichia coli collected from Canadian hospitals: CANWARD 2007–18. J Antimicrob Chemother 2021;76:2825–32.
15. Bunnell KL, Wenzler E, Harrington AT, Danziger LH. Impact of Clinical and Laboratory Standards Institute breakpoint changes on susceptibility rates of cephalosporins in uncomplicated urinary tract infections caused by Enterobacteriaceae. Diagn Microbiol Infect Dis 2018;90:335-6.