Role of aminoglycoside-modifying enzymes and 16S rRNA methylase (ArmA) in resistance of Acinetobacter baumannii clinical isolates against aminoglycosides

Abstract INTRODUCTION: This study aimed to determine the role of genes encoding aminoglycoside-modifying enzymes (AMEs) and 16S rRNA methylase (ArmA) in Acinetobacter baumannii clinical isolates. METHODS: We collected 100 clinical isolates of A. baumannii and identified and confirmed them using microbiological tests and assessment of the OXA-51 gene. Antibiotic susceptibility testing was carried out using disk agar diffusion and micro-broth dilution methods. The presence of AME genes and ArmA was detected by PCR and multiplex PCR. RESULTS: The most and least effective antibiotics in this study were netilmicin and ciprofloxacin with 68% and 100% resistance rates, respectively. According to the minimum inhibitory concentration test, 94% of the isolates were resistant to gentamicin, tobramycin, and streptomycin, while the highest susceptibility (20%) was observed against netilmicin. The proportion of strains harboring the aminoglycoside resistance genes was as follows: APH(3′)-VIa (aphA6) (77%), ANT(2”)-Ia (aadB) (73%), ANT(3”)-Ia (aadA1) (33%), AAC(6′)-Ib (aacA4) (33%), ArmA (22%), and AAC(3)-IIa (aacC2) (19%). Among the 22 gene profiles detected in this study, the most prevalent profiles included APH(3′)-VIa + ANT(2”)-Ia (39 isolates, 100% of which were kanamycin-resistant), and AAC(3)-IIa + AAC(6′)-Ib + ANT(3”)-Ia + APH(3′)-VIa + ANT(2”)-Ia (14 isolates, all of which were resistant to gentamicin, kanamycin, and streptomycin). CONCLUSIONS: High minimum inhibitory concentration of aminoglycosides in isolates with the simultaneous presence of AME- and ArmA-encoding genes indicated the importance of these genes in resistance to aminoglycosides. However, control of their spread could be effective in the treatment of infections caused by A. baumannii.


INTRODUCTION
Acinetobacter baumannii, living in the soil, the water, and different hospital environments, is an important opportunistic pathogen that causes nosocomial infections such as pneumonia, urinary tract infections, intravenous catheter-associated infections, and ventilation-associated infections, particularly in intensive care units [1][2][3][4] . The ability of this microorganism to remain in the hospital environment and to spread among the patients, along with their resistance to several antibiotics, are the main driving forces behind large-scale recurrent events in different countries 5 .
The major antibiotics used for the treatment of infections caused by this organism are beta-lactams, aminoglycosides, fluoroquinolones, and carbapenems; however, A. baumannii has shown different rates of resistance against these antimicrobial agents [6][7][8] . These infections are difficult, costly, and sometimes impossible to treat owing to the high ability of A. baumannii to acquire antibiotic resistance genes and the development of multidrug-resistant (MDR) strains 9,10 . Aminoglycosides are one of the main drugs used for the treatment of Acinetobacter infections 11 ; however, recently, the resistance of A. baumannii to these antibiotics has also increased. Two main mechanisms of resistance to aminoglycosides are the alteration of the ribosome structure caused by mutations in the ribosomal 16S rRNA and the enzymatic resistance mechanism 12 . The enzymatic alteration of the aminoglycoside molecule at -OH or -NH 2 groups by aminoglycoside-modifying enzymes (AMEs) is the most important resistance mechanism [12][13][14] . AMEs are classified into three major groups: aminoglycoside phosphotransferase (APH), aminoglycoside acetyltransferase (AAC), aminoglycoside nucleotidyltransferase (ANT), and aminoglycoside adenylyltransferase (AAD) 5,13 . Aminoglycoside acetyltransferases cause acetylation of the -NH 2 groups of aminoglycosides at the 1, 3, 2', and 6' positions using acetyl coenzyme A as a donor substrate 15 . Aminoglycoside phosphotransferases phosphorylate the hydroxyl groups present in the structure of aminoglycosides at the 4, 6, 9, 3', 2'', 3'', and 7'' positions (seven different groups) with the help of ATP; the largest enzymatic group in this family is the APH(3′)-I group 16 . The proportion of strains harboring the aphA6 gene in A. baumannii is widespread, and this enzyme is the cause of resistance to neomycin, amikacin, kanamycin, paromomycin, ribostamycin, butirosin, and isepamicin 17 . Aminoglycoside nucleotidyltransferases are classified into 5 groups, and the genes encoding these enzymes can be found in chromosomes or transferred by plasmids and transposons 12 . These enzymes transfer an AMP group from ATP to a hydroxyl group at the 2'', 3'', 4', 6, and 9 positions of the aminoglycoside molecule 13 . In addition to AMEs, 16S rRNA methylation by the ArmA enzyme is a novel mechanism that contributes to the high level of aminoglycoside resistance in A. baumannii, as reported in the Far East, Europe, and North America 5 . This enzyme can be transferred by class 1 integrons and is often detected in carbapenemresistant A. baumannii isolates 18 . This study aimed to investigate the role of some important aminoglycoside-modifying enzymes and 16S rRNA methylase (ArmA) in the resistance of A. baumannii clinical isolates to aminoglycosides in Sari, located north of Iran.

Sample collection and bacterial isolates
This study was performed on A. baumannii isolated from patients admitted to different educational hospitals in Sari, north of Iran, for 6 months (April 2019 to September 2019). The clinical specimens included blood, urine, respiratory secretions (bronchial lavage and tracheal secretions), CSF, and ulcer (surgical and burn wound). The clinical isolates were identified using conventional microbiological tests 19 and confirmed by polymerase chain reaction (PCR) amplification of the blaOXA-51 gene using specific primers 20 ; the reaction conditions are shown in Table 1.

DNA extraction, PCR, and multiplex-PCR
DNA was extracted from all A. baumannii isolates grown for 24 h using an alkaline lysis method with sodium dodecyl sulphate (SDS) and NaOH, as previously published 23 , with few modifications. In brief, first, we prepared a lysis buffer by dissolving 0.5 g of SDS and 0.4 g of NaOH in 200 µL of distilled water. Next, 4-6 colonies of the bacteria were suspended in 60 µL of lysis buffer and subsequently heated at 95 °C for 10 min. In the next step, the suspension was centrifuged at 13000 rpm for 5 min, and 180 µL of distilled water was added to the microtubes. The obtained supernatant was frozen at −20 °C until use as the extracted DNA in PCR.
Two sets of multiplex-PCR were used to detect AME-encoding genes in A. baumannii isolates using the specific primers shown in Table 1. APH(3′)-VIa (aphA6), ANT(2")-Ia (aadB), and ArmA genes were detected in the same set; AAC(6′)-Ib (aacA4) and AAC(3)-IIa (aacC2) were identified in the second set; and the ANT(3")-Ia (aadA1) gene was detected by PCR alone. The PCR and multiplex-PCR were performed in 25 µL of final volume containing 12.5 µL of the master mix (Ampliqon, Denmark), 10 pmol of each primer (Bioneer, South Korea), and 500 ng of template DNA; the reaction solutions were brought to the desired volume through the addition of distilled water. The genes were amplified under standard conditions using a thermocycler machine (Bio-Rad, USA). All reactions were performed in 34 cycles, and the conditions are shown in Table 1.

Statistical analysis
The data were analyzed using SPSS (version 21). Categorical data were analyzed using the Fisher's exact test, and a P-value less than 0.05 was considered statistically significant. In addition, an independent t-test was used to examine the mean age of the subjects.

Patients, samples, and bacterial isolates
In this study, 100 non-duplicated A. baumannii clinical isolates were collected from 100 patients admitted to the teaching and educational hospitals of Sari, north of Iran. All isolates identified using the phenotypic method contained the blaOXA-51 gene according to the PCR results. The mean age of the patients was 42.08±25.08 years (minimum age: 6 months; highest age: 88 years), and 50% of the patients were male. There was no significant difference between men and women in terms of mean age (p=0.64). Most of the bacterial isolates (34%) were obtained from patients admitted to the burn wards, while 29%, 21%, and 16% of the isolates were collected from the ICU, surgery, and pediatric wards, respectively. The most common type of specimen (73%) for isolation of the bacteria was the wound samples; however, 15% and 12% of other clinical isolates were obtained from urine and blood cultures, respectively.

Antimicrobial susceptibility pattern
According to the results of the disk agar diffusion method, the most and least effective antibiotics in the present study were imipenem and ciprofloxacin, with resistance rates of 75% and 100%, respectively ( Table 2). Moreover, 94% of the isolates were detected as multi-drug resistant (MDR), and the most MDR isolates were collected from wound samples. Table 2 presents the antibiotic resistance patterns of all A. baumannii clinical isolates in this study based on hospital wards, as well as sample types. Resistance to the tested antibiotics was not significantly correlated with the sample types and hospital wards where the samples were collected.
Moreover, according to the MIC results, the resistance rate against gentamicin, kanamycin, tobramycin, and streptomycin was 94%, while the highest susceptibility (20%) of A. baumannii isolates was observed against netilmicin. In contrast, 74%, 68%, and 78% of our clinical isolates were resistant to amikacin, netilmicin, and spectinomycin, respectively. The MIC ranges of aminoglycosides and their relationship with the presence of AMEs-encoding genes are shown in Table 3.

Gene profiles of the isolates
The frequency of each aminoglycoside resistance gene and its relation with the MIC ranges are shown in Table 3. In total, the proportions of aminoglycoside resistance genes among our clinical isolates of A. baumannii were as follows: The relationship between the presence of aminoglycoside resistance genes and the aminoglycoside susceptibility pattern of the isolates is shown in Table 4. There was a significant association between the presence of all resistance genes and the non-susceptibility (resistance or intermediate resistance) to all aminoglycosides, except armA and resistance to netilmicin. Important data from this table indicates that in some groups, such as gentamicin-and tobramycin-resistant groups, all resistant isolates contained some AMEs-encoding genes such as aacC2, aacA4, and aadA1.

DISCUSSION
Overuse and misuse of antibiotics in the treatment of infections caused by A. baumannii has led to the emergence of MDR isolates in hospitals and health centers 24 . The spread of AME-encoding genes among the clinical isolates of A. baumannii is an important concern in the prescription of these traditional and effective antibiotics, as 94% of our isolates were resistant to kanamycin, gentamicin, streptomycin, and tobramycin. However, we found that netilmicin was the most effective aminoglycoside, as this antibiotic is not commonly used in the treatment of bacterial infections. This finding was similar to that of another Iranian study 1 . However, their isolates revealed an MIC 50 ≤8 µg/mL, while in the present study it ranged from 128 µg/mL, indicating an increased resistance rate in our region. However, the MIC ranges of other clinically important aminoglycosides such as amikacin, gentamicin, and tobramycin in the present study were significantly higher than those reported in previous studies in Iran and other countries 1,5,25 , while these ranges were almost similar to those reported by Yoo Jin Cho et al. in 2009 22 .
The molecular analysis of AME-encoding genes in the present study showed a high frequency of aphA6, aadB, aadA1, aacA4, aacC2, and armA genes, consistent with the previous studies from Iran 1,5 . Given that AME-encoding genes can spread by transferable genetic elements 18 , this high proportion would be justified. Possibly, these resistance genes can spread between different gram-negative bacteria such as Pseudomonas aeruginosa and Enterobacteriaceae. Further confirmation of this hypothesis can be found in another study conducted in Iran on the clinical isolates of P. aeruginosa, according to which aadB and aacA4 were the most prevalent AME genes 26 . In a study performed by Lee et al. in Korea, the highest frequency was reported for aphA6 (71%), aacC1 (56%), and aadB (48%) 27 .
In addition, a high proportion of aphA6 and aadB was reported by Akers et al. in USA in agreement with the present study; almost 42% of their isolates were collected from the burn ward and ICU. However, the resistance rates toward gentamicin and amikacin in their isolates were 96.6% and 57.1%, respectively 25 . However, aphA6 confers resistance to amikacin and kanamycin 17 . Interestingly, 74% and 88.3% of our isolates containing aphA6 exhibited MIC values of ≥256 µg/mL for amikacin and kanamycin, respectively. Moreover, a study carried out in Poland revealed that aphA6 was the second most prevalent AME gene (78.7%) among 61 A. baumannii isolates 28 . However, aadB, the second most prevalent AME gene in the current study, confers resistance to gentamicin, tobramycin, and kanamycin in gram-negative bacteria 26 , while 95.8%, 93.1%, and 87.6% of our isolates containing aadB showed an MIC range of ≥256 µg/mL for these antibiotics, respectively.
In addition, we found that 33% of our isolates contained the aacA4 gene. Other research performed in the USA detected only one isolate carrying this gene from blood and wound infections that were resistant to gentamicin, tobramycin, and amikacin 25  Additionally, the armA gene, which is an effective factor in the development of resistance to aminoglycosides in A. baumannii, can be placed on plasmids and frequently recognized in carbapenem-resistant isolates 18 . This gene encodes a 16S rRNA methylase, resulting in limited access of aminoglycosides to the ribosome of the bacteria and causing high-level aminoglycoside resistance (HLAR) against gentamicin, tobramycin, amikacin, and kanamycin 1 . Surprisingly, among 22 armA-positive A. baumannii isolates in this study, 21 (95.4%), 16 (72.7%), 18 (81.8%), and 21 (95.4%) isolates showed high-level resistance (MIC≥256 µg/mL) to gentamicin, tobramycin, amikacin, and kanamycin, respectively, with an MIC 50 ≥128 μg/mL. Considering that most isolates in the present study were MDR, and a high proportion of strains harboring the AME genes was detected, the simultaneous presence of carbapenem-resistance genes and AME genes in A. baumannii has been proven 18,28 ; this assumption may also be true for our isolates. However, 75-97% of our isolates were resistant to carbapenems and other β-lactams. Other studies from South Korea, Iran, and North America have reported armA production by A. baumannii 1,27,29 . Additionally, other researchers have revealed the role of the armA gene in high-level resistance to amikacin and gentamicin 22,30 .
In addition to the material presented, the most important problem observed in our study was the simultaneous presence of aminoglycoside resistance genes. We detected 22 gene profiles, while Nowak et al. detected only 3 combinations of AME genes from 61 carbapenem-resistant and aminoglycoside non-susceptible A. baumannii isolates 28 . Our most prevalent combinations were APH(3')-VIa+ANT(2")-Ia (39 isolates) with 95-100% resistance rates against aminoglycosides and AAC(3)-IIa+AAC(6')-Ib+ANT(3")-Ia+APH(3')-VIa+ANT(2")-Ia (14 isolates) of which 93-100% were resistant to aminoglycosides. The common point between our study and the study by Nowak et al. was the presence of aphA6 among most of the isolates. However, Akers et al. detected 16 AME gene profiles, of which 12 (75%) isolates had a combination of these genes 25 . The most prevalent (38/107 isolates) combination of their study included APH(3')-Ia+ANT(2'')-Ia, and 35 (92.1%) were concurrently resistant to gentamicin, tobramycin, and 9/10 amikacin. Nevertheless, 85% of our A. baumannii isolates carried more than one AME gene, of which 52 (61.1%) contained 2 AME genes concurrently and most of them were resistant to all tested aminoglycosides. Moreover, we found that as the number of AME genes increased, the likelihood of resistance to aminoglycosides, especially gentamicin, tobramycin, streptomycin, and kanamycin, increased. Due to the higher proportion of strains harboring the AME genes, especially aph, it may be better to use phosphotransferases and acetyltransferase inhibitors such as the bovine antimicrobial peptide indolicidin, as previously reported 31 , in combination with aminoglycosides in our region, Iran.

CONCLUSIONS
High-level aminoglycoside MIC ranges in isolates with the simultaneous presence of AME and ArmA-encoding genes indicated the importance of these genes in resistance to aminoglycosides in A. baumannii. However, it seems that the selection of the appropriate antibiotic based on antimicrobial susceptibility testing and the use of combination therapy would be effective in overcoming this problem in such countries. Therefore, it is necessary to collect data from monitoring studies for the prevention, treatment, and control of the infections caused by this microorganism.