Wild-type cutoff for Apramycin against Escherichia coli

Background Apramycin is used exclusively for the treatment of Escherichia coli (E.coli) infections in swine around the world since the early 1980s. Recently, many research papers have demonstrated that apramycin has significant in vitro activity against multidrug-resistant E.coli isolated in hospitals. Therefore, ensuring the proper use of apramycin in veterinary clinics is of great significance of public health. The objectives of this study were to develop a wild-type cutoff for apramycin against E.coli using a statistical method recommended by Clinical and Laboratory Standards Institute (CLSI) and to investigate the prevalence of resistance genes that confer resistance to apramycin in E. coli. Results Apramycin susceptibility testing of 1230 E.coli clinical isolates from swine were determinded by broth microdilution testing according to the CLSI document M07-A9. A total number of 310 E.coli strains from different minimum inhibitory concentration (MIC) subsets (0.5–256 μg/mL) were selected for the detection of resistance genes (aac(3)-IV; npmA; apmA) in E. coli by PCR. The percentage of E. coli isolates at each MIC (0.5, 1, 2, 4, 8, 16, 32, 64, 128, and 256 μg/mL) was 0.08, 0.08, 0.16, 2.93, 31.14, 38.86, 12.85, 2.03, 1.46, and 10.41%. The MIC50 and MIC90 were 16 and 64 μg/mL. All the 310 E.coli isolates were negative for npmA and apmA gene, and only the aac(3)-IV gene was detected in this study. Conclusions The wild-type cutoff for apramycin against E.coli was defined as 32 μg/mL. The prevelance of aac(3)-IV gene mainly concentrated in these MIC subsets ‘MIC ≥ 64 μg/ mL’, which indicates that the wild-type cutoff established in our study is reliable. The wild-type cutoff offers interpretion criteria of apramycin susceptibility testing of E.coli.

infections in swine, cattle, sheep, poultry, and rabbits around the world since the early 1980s and was approved for use in China in 1999 [4]. Recently, many research papers have demonstrated that apramycin has significant in vitro activity against multidrug-, carbapenem-and aminoglycoside-resistant E.coli isolated in hospitals. And its excellent breadth of activity renders apramycin a promising drug candidate for the treatment of systemic Gram-negative infections [5][6][7][8][9][10][11]. The first resistant E. coli strain was detectable in nature shortly after the application of APR [12]. It has been determined to date that two resistance genes confer resistance to APR in E. coli. One is AAC (3)-IV, which encodes an aminoglycoside 3-N-acetyltransferase type IV enzyme [13]. The other is NpmA, which was identified in a clinical E. coli strain and encodes a 16S rRNA m1A1408 methyltransferase [14]. Moreover, another APR resistance gene, apmA, was detected in bovine methicillinresistant Staphylococcus aureus (MRSA) of sequence type 398 in 2011 and encodes for a protein of 274 amino acids [15]. APR resistance has been also detected in E.coli clinical isolates of hospitalized patients despite it has not been used in human medicine [16]. The horizontal transfer of the APR resistance gene aac(3)-IV results in the dissemination of APR-resistance E. coli isolates between animals and humans [17]. In addition, cross-resistance between APR and other aminoglycosides such as gentamicin (GEN) and tobramycin for the treatment of severe infections in humans has been well documented [18,19]. Previous study reported that pigs may have been an important reservoir for GENresistance bacteria transfer to humans [20]. Considering the importance of GEN in human medicine, improper use of APR in animals contributing to increased resistance is of great concern.
Wild-type cutoff values (CO WT ) are the useful tools available to laboratories performing susceptibility testing and to clinicians treating infections. In addition, the tools also provide alternative means for monitoring the emergence of drug resistance in any given bacterial species [21]. A statistical method was a more scientific method which has been adopted by the Clinical and Laboratory Standards Institute (CLSI) as a standard method for CO WT establishment [22,23]. The purposes of the present study were (i) to develop CO WT of APR against E.coli using a statistical method recommended by CLSI and (ii) to investigate the prevalence of genes that confer resistance to APR in E. coli.

Antibacterial susceptibility testing
The original MICs distributions and MICs cumulative distributions of APR are presented in Fig. 1

Establishment of CO WT
The MIC distributions (1-64 μg/mL) for APR were statistically consistent with a normal distribution (skewness = 0.194 and kurtosis = 0.386). Non-linear regression curve fitting of cumulative log 2 MIC data was selected as the preferred method for determining the means and standard deviations of MIC distributions owing to the normal (Gaussian) distribution is widely accepted. The process involves fitting an initial subset and generating estimates (in log2) of the strain number, mean and standard deviation in the subset. Repeat this process by reducing the previous subset in each successive column to create the next subset, and repeat the curve fitting until it is clear that there is a subset in which the absolute difference between the true and estimated separation numbers is the smallest. The optimum MIC range from 0.5 to 256 μg/ml was obtained from non-linear regression, the five subsets examined demonstrated that the subset 'MIC = 32 μg/ mL' gave the minimum difference (Table 1 and Fig. 2). The probability of an MIC at 32 μg/ml was 99.18%, which encompassed 95% of the WT isolates according to the NORMDIST function in Microsoft Excel ( Table 2). As a result, the CO WT was defined as 32 μg/mL.

The prevalence of APR resistance genes
A total number of 310 E.coli clinical isolates containing different MIC subsets (0.5-256 μg/mL) were conveniently selected for the detection of three resistance genes (aac(3)-IV; npmA; apmA) in E. coli by PCR. The prevalence of APR resistance genes presented in Table 3. All the 310 E.coli clinical isolates were negative for npmA and apmA gene by PCR. The only resistance gene in E.coli that confer resistance to APR is aac(3)-IV in this study. The prevalence of aac(3)-IV gene was 91.59% (98/ 107) in the subset 'MIC = 256 μg/ mL'; was 64.71% (11/ 17)

Discussion
APR, an aminoglycoside antibiotic, was used in veterinary therapy and animal husbandry in the early 1980's in several European countries and was approved to use in China since 1999 [4]. However, a recent study demonstrated that APR is a promising drug candidate for the treatment of systemic Gram-negative infections that are resistant to treatment with other aminoglycoside antibiotics by evaluating the in vitro activity of APR against multidrug-, carbapenem-and aminoglycoside resistant Enterobacteriaceae and Acinetobacter baumannii in patient from Europe, Asia, Africa and South America [6]. In this study, the results that 171 isolates among the 1230 E.coli clinical isolates had MICs ≥64 μg/ mL were similar to the previous study [24]. Resistant E. coli are generally isolated from diseased pigs in our study, and E. coli from pigs may be an important reservoir for transfer of APR-resistance genes or APR-resistant bacteria to humans [20]. Marshall and Levy, 2011 summarized the evidence from animal on farms to human transfer of resistant bacteria. One is to acquire resistance by direct contact with animals, and the other is the spread of antibiotic resistance through the food chain [25]. The effect of antimicrobial usage on the prevalence of resistant bacteria in animals is significant [26].
Phenotypic resistance is commonly interpreted according to the clinical standards and recommended breakpoints from the European Committee on Antimicrobial Susceptibility Testing (EUCAST) or the CLSI. For aminoglycosides, the MIC breakpoints of amikacin, GEN, netilmicin and tobramycin were established by EUCAST, and the MIC breakpoints of netilmicin, kanamycin, amikacin, tobramycin and GEN were established by CLSI. However, the MIC breakpoint of APR was not established by either EUCAST or CLSI. To set breakpoints required a combination of MIC values, pharmacokinetic/ pharmacodynamic relationship and clinical outcome data [21]. However, it is very difficult and expensive to generate this kind of data required for breakpoint determination. The CO WT is a useful tool for the interpretation of antimicrobial susceptibility testing results conducted in laboratories [21]. In this study, the CO WT was defined as ≤32 μg/mL by using a statistical method recommended by CLSI and was similar with that the epidemiological cut-off value (ECOFF) routinely used for APR was >16 μg/ mL by the Laboratory of Swine diseases, Kjellerup, Denmark and by the Danish Veterinary Institute, Frederiksberg, Denmark [27]. Different use of apramycin in pigs and chickens results in different susceptibility of clinical E. coli strains to apramycin. Tian To date, it has been determined that two resistance genes in E. coli (aac(3)-IV, npmA) confer resistance to APR [13,14]. The gene aac(3)-IV is the only identified gene causing enzymatic cross-resistance between APR and GEN [29]. GEN is a critically important drug and is generally combinated with beta-lactam as the first choice antimicrobial for severe human infections [19]. In this study, the high prevalence of aac(3)-IV gene was observed in the resistant E.coli isolates, which was consistant with other previous studies [4,19,[30][31][32]. The npmA gene, confers high resistance to many aminoglycoside types upon the host E. coli, was originally found in an E.coli strain isolated in 2003 from the urine of an inpatient in a general hospital in Japan [14] and did not appear in the scientific literature until August 2017 from China [33]. The npmA gene was not detected in any  Conclusion Given the lack of interpretion criteria of APR susceptibility testing, the CO WT (≤ 32 μg/mL) for APR against E.coli was established by using a statistical method recommended by CLSI in this study. The prevelance of APR resistance gene aac(3)-IV mainly concentrated in these MIC subsets "MIC ≥ 64 μg/ mL", which indicates that the CO WT established in our study is reliable. The CO WT offers guidance for APR susceptibility testing of E.coli isolated from animals.

Isolates
The rectal swabs collected on each visit from the target animals were pooled and tested as one analytical sample. A total of 1230 E.coli isolates were used in the study, which including 858 isolates identified from rectal swabs of pigs in different province in China by using the biochemical identification and PCR method according to   Laboratory of Veterinary Biotechnology, Harbin Veterinary Research Institute, Chinese Academy of Agricultural Sciences (n = 108), Husbandry and Veterinary College, Jilin University (n = 112), and College of Animal Husbandry and Veterinary Science, Henan Agricultural University (n = 152). All of the bacterial isolates were confirmed by polymerase chain reaction (PCR) [36].

Chemicals and reagents
Pure powder of APR was purchased from Zhejiang Hisun Pharmaceutical Co., Ltd., Taizhou [37]. Only those results, for which the QC MICs were within the established reference range (4-8 μg/mL), were used in the study [38]. All MICs determinations were performed in triplicate.

Data analysis
The definitions of the subsets, lognormal distribution, skewness, kurtosis, and CO WT are presented in Table 4. The MICs were transformed into log 2 values in order to analyze the MIC distributions. The kurtosis and skewness of each MIC distribution were tested. To confirm the presence of more than one MIC distribution, frequency distributions of MIC data were analyzed by nonlinear least squares regression analysis based on the following Cumulative Gaussian Counts equation: Z = ((X-Mean))/SD, Y=N*zdist(z) according to the previous study [41], in which the Mean is the average of the original distribution, from which the frequency distribution was created; SD is the standard deviation of the original distribution (calculated by Graphpad prism 6.0 software, San Diego, CA). Three parameters, the total number (N) in the presumed unimodal distribution, the mean, and SD (both log 2 ) were estimated. N was estimated rather than taken as a constant in the regression, because of Lognormal Distribution A frequency (probability) distribution where the data are distributed in a Gaussian (normal) manner after the data points have been converted to logarithms. [22] Skewness Lack of symmetry in a frequency distribution. [22] Kurtosis Excessive peaking or flattening of a frequency distribution when compared with the normal distribution. [22] CO WT CO WT also known as the epidemiological cutoff (ECV), defined as the highest susceptibility endpoint of the wild-type (WT) population MIC, has been shown to detect the emergence of in vitro resistance or to separate WT isolates (without known mechanisms of resistance) from non-WT isolates (with mechanisms of resistance and reduced susceptibilities to the antibacterial agent being evaluated). CO WT are calculated by taking into account the MIC distribution, the modal MIC of each distribution, and the inherent variability of the test (usually within one doubling dilution) and should encompass ≥95% of isolates. [22,39,40] [15] the desire to fit the data to the distribution without assuming that N truly contained only wild-type isolates [22]. The NORMINV and NORDIST functions in Microsoft Excel were used to set the WT distribution cutoffs which were used to determine the MIC that encompass at least 95% of that distribution [22,42].

Molecular characterisation of mechanisms of resistance to APR
A total number of 310 E.coli strains from different MIC subsets (0.5-256 μg/mL) were conveniently selected for the detection of resistance genes in E. coli that confer resistance to APR by PCR. The primers used in this study are presented in Table 5. Genomic DNA was extracted using a TIANamp Bacteria DNA Kit (TIANGEN BIO-TECH (BEIJING) CO., LTD.) according to the manufacturer's instructions. Then, 2 μL (400 ng/μL) was added to a reaction mixture containing 25 μL Premix TaqTM Version 2.0 plus dye, 13 μL sterile ddH 2 O, 5 μL 10 μM primer F and 5 μL 10 μM primer R. Amplification conditions were 94°C for 5 min, followed by 30 cycles of 94°C for 30 s, 55°C for 30 s (52°C for apmA) and 72°C for 1 min, and a final elongation at 72°C for 10 min. PCR products were analysed on 1.5% (w/v) agarose gelsstained with ethidium bromide. The amplified products were sequenced by the Sangon Biotech Co., Ltd., Shanghai, China. E. coli ATCC 25922 strains was used as negative controls.