In vitro Activity of Apramycin Against Carbapenem-Resistant and Hypervirulent Klebsiella pneumoniae Isolates

Objective The emergence of carbapenem-resistant and hypervirulent Klebsiella pneumoniae (CR-hvKp) strains poses a significant public threat, and effective antimicrobial therapy is urgently needed. Recent studies indicated that apramycin is a potent antibiotic with good activity against a range of multi-drug resistant pathogens. In this study, we evaluated the in vitro activity of apramycin against clinical CR-hvKp along with carbapenem-resistant non-hvKp (CR-non-hvKp) isolates. Methods Broth microdilution method was used to evaluate the in vitro activities of apramycin, gentamicin, amikacin, imipenem, meropenem, doripenem, ertapenem and other comparator “last-resort” antimicrobial agents, including ceftazidime-avibactam, colistin and tigecycline, against eighty-four CR-hvKp and forty CR-non-hvKp isolates collected from three Chinese hospitals. Multilocus Sequence typing (MLST), molecular capsule typing (wzi sequencing) and antimicrobial resistance genes were examined by PCR and Sanger sequencing. Pulsed-field gel electrophoresis and next generation sequencing were conducted on selected isolates. Results Among the 84 CR-hvKp isolates, 97.6, 100, 97.6, and 100% were resistant to imipenem, meropenem, doripenem and ertapenem, respectively. Apramycin demonstrated an MIC50/MIC90 of 4/8 μg/mL against the CR-hvKp isolates. In contrast, the MIC50/MIC90 for amikacin and gentamicin were >64/>64 μg/mL. All CR-hvKp isolates were susceptible to ceftazidime-avibactam, colistin and tigecycline with the MIC50/MIC90 values of 0.5/1, 0.25/0.5, 1/1, respectively. For CR-non-hvKp, The MIC50/90 values for apramycin, gentamicin and amikacin were 2/8, >64/>64, and >64/>64 μg/mL, respectively. There were no statistical significance in the resistance rates of antimicrobial agents between CR-hvKp and CR-non-hvKp groups (p > 0.05). Genetic analysis revealed that all CR-hvKp isolates harbored blaKPC–2, and 94% (n = 79) belong to the ST11 high-risk clone. 93.6% (44/47) of amikacin or gentamicin resistant strains carried 16S rRNA methyltransferases gene rmtB. Conclusion Apramycin demonstrated potent in vitro activity against CR-hvKp isolates, including those were resistant to amikacin or gentamicin. Further studies are needed to evaluate the applicability of apramycin to be used as a therapeutic antibiotic against CR-hvKp infections.


INTRODUCTION
Klebsiella pneumoniae is a clinically important pathogen, causing a wide range of diseases including pneumonia, urinary tract infections (UTIs), bloodstream infections, in hospital settings among neonates, elderly and immunocompromised individuals (Paczosa and Mecsas, 2016;Bengoechea and Sa Pessoa, 2019). Hypervirulent K. pneumoniae (hvKp) is an increasing recognized pathotype that is more virulent than classical K. pneumoniae (cKp), which showed the propensity to cause serious, lifethreatening infections in otherwise healthy individuals in the community (Shon et al., 2013;Gu et al., 2018).
Historically, hvKp isolates are susceptible to most antibiotics, however, multidrug-resistant hvKp, especially carbapenemresistant strains, are increasingly reported in the past few years (Cejas et al., 2014;Li et al., 2014;Siu et al., 2014). The emergence of carbapenem-resistant and hypervirulent K. pneumoniae (CR-hvKp) was either due to the acquisition of a pLVPK-like virulence plasmid by classic carbapenem-resistant K. pneumoniae strains (Gu et al., 2018) or through acquisition of a carbapenemaseencoding plasmid by hvKp isolates (Shon et al., 2013). The rmpA, rmpA2, iutA, iucABCD and iroBCDN genes harbored by pLVPK-like virulence plasmid is responsible for the virulence phenotype (Yang et al., 2019). A recent study from China reported a fatal outbreak of ventilator associated pneumonia caused by ST11 CR-hvKp strains (Gu et al., 2018). More recently, one hypermucoviscos isolate with KPC-3 carbapenemase-encoding plasmid was firstly described in the United States, which also exhibited colistin heteroresistance (Wozniak et al., 2019). The rapid dissemination of CR-hvKp strains in different global regions underscores the urgent need of appropriate antibiotic therapy against these life-threating pathogens.
Aminoglycosides including amikacin, gentamicin and tobramycin, which belong to the 4,6-disubstituted deoxystreptamine (DOS) subclass, are among the few drugs that retain certain in vitro activity against CRE (Livermore et al., 2011). However, the acquisition of 16S rRNA methylases confer high level resistance to the 4,6-disubstituted DOS aminoglycosides, which imposes serious challenge for clinical management (Hu et al., 2017). Apramycin is of the 4-monosubstituted DOS subclass, markedly different in its chemical structure from other clinically used aminoglycoside antibiotics. Its unique structure makes apramycin molecules inherently resilient to almost all resistance determinants commonly found in multi-drug resistant strains (Juhas et al., 2019). Indeed, recent studies indicated that apramycin is a potent antibiotic with good activity against a range of clinical pathogens, including multidrugresistant Mycobacterium tuberculosis, carbapenem-resistant Enterobacteriaceae, spectinomycin-resistant Neisseria gonorrhoeae and Staphylococcus aureus (Smith and Kirby, 2016;Hu et al., 2017;Truelson et al., 2018;Galani et al., 2019;Riedel et al., 2019). In order to evaluate whether apramycin can be served as an alternative antibiotic to treat the CR-hvKp infections, we examined the in vitro activity of apramycin against a collection of CR-hvKp isolates collected from three different hospitals in China. Its in vitro activity was compared with the results from other aminoglycosides (amikacin and gentamicin) and some "last resort" antibiotics, including colistin, tigecycline and ceftazidime-avibactam.

Antibiotic Susceptibility Assay
The antimicrobial susceptibility was assessed via a minimum inhibitory concentration (MIC) broth microdilution method. Results were interpreted using Clinical Laboratory Standards Institute (CLSI) breakpoints (Clinical and Laboratory Standards Institute, 2019) with the exception of tigecycline and polymyxin B, which were interpreted using EUCAST breakpoints (The European Committee on Antimicrobial Susceptibility Testing, 2019). National Antimicrobial Resistance Monitoring System (NARMS) breakpoints for enteric bacteria were used for the interpretation of apramycin susceptibility, in which strains were classified as susceptible (MIC ≤ 8 µg/mL), intermediate (MIC = 16∼32 µg/mL), or resistant (MIC ≥ 64 µg/mL) (National Antibiotic Resistance Monitoring System [NARMS], 2002). MICs were performed in duplicate on two separate days. Escherichia coli ATCC 25922 and Pseudomonas aeruginosa ATCC 27853 were used as the quality control strains in each experiment. K. pneumoniae ATCC 700603 was used as quality control strain for ceftazidime-avibactam which was tested against ceftazidime-avibactam and ceftazidime alone to confirm the activity of avibactam in the combination.
The apramycin resistant isolate, KpApr172, was subject to next generation sequencing using the Illumina HiSeq system (Illumina, San Diego, CA, United States). Genomic DNA was isolated using a Wizard R Genomic DNA Purification Kit (Promega, Madison, WI, United States). Sequencing reads were de novo assembled using Spades 3.12.0 (Bankevich et al., 2012). In silico multi-locus sequence typing was performed using MLST 2.0 (Larsen et al., 2012) while the acquired antimicrobial resistance genes were identified using ResFinder 3.0 (Zankari et al., 2012). The plasmid replicons in the sequenced isolates were identified using PlasmidFinder 2.0 (Carattoli et al., 2014). The whole-genome sequence of KpApr172 was deposited in the GenBank database under accession number WUJI00000000.

Statistical Analysis
Chi-square and Fisher's exact analysis were employed to compare resistance rate of antimicrobial agents between CR-hvKp and CR-non-hvKp groups. Statistical analyses were performed using GraphPad Prism 6.0 (GraphPad Software, La Jolla, CA, United States). P value of <0.05 was considered to indicate statistically significant differences.
Notably, one CR-non-hvKp isolate (KpApr172) was resistant to apramycin. PCR showed that it contains the aminoglycoside acetyltransferase 3-IV gene aac(3)-IV, which encodes resistance to apramycin, gentamicin, netilmicin and tobramycin. Further next generation sequencing revealed that KpApr172 was a strain ST11 and harbored the KL47 capsular type. Resistance gene mining indicated that KpApr172 harbored 17 antimicrobial resistance genes encoding resistance Aminoglycoside modifying enzyme (AME) Aac(6 )-Ib 6 (7.14%) 3 (7.5%) to β-lactams (bla KPC-2 , bla NDM-5 , bla CTX-M-3 , and bla SHV-12 ), aminoglycosides [armA, aadA2, aac(3)-IId, aac(3)-IV and aph(4)-Ia], macrolid [msr(E) and mph(E)], bleomycin (ble), florfenicol (floR), quaternary ammonium compound (qacE 1) and sulfonamide-trimethoprim (sul1, sul2 and dfrA12). The quinolone resistance-determining region (QRDR) genes gyrA, gyrB, parC and parE were further examined, and we observed mutations encoding amino acid substitutions at Ser83-Ile and Asp87-Gly within the QRDR regions of GyrA, and Ser80-Ile in ParC. Examination of the outer membrane protein OmpK35 and OmpK36 genes revealed a premature stop codon at AA63 in OmpK35, due to an adenine (A) deletion mutation at nt 82, and the glycine and aspartic acid insertion at AA134. In silico plasmid replicon typing identified five plasmid groups in KpApr172, belonging to incompatibility groups X3, R, FII and ColE-like (n = 2), respectively. In order to further characterize the transferability of apramycin resistance gene aac(3)-IV, we attempted to transfer them via conjugation to a recipient E. coli J53 AzR strain (data not shown), however, we were not able to transfer the aac(3)-IV by conjugation, suggesting the gene may be harbored by a non-conjugative plasmid or on the chromosome. Examination of the aac(3)-IV-containing contig from the genome assemblies revealed that aac(3)-IV was located on a 13,397 bp segment, along with the resistance genes, floR, sul2 and aph(4)-Ia (Figure 2). BLASTn analysis showed that the sequences of this contig were identical (100% query coverage and 100% identities) to the corresponding regions in some plasmids, including pHNAH67 (KX246266, from E. coli of chicken), pCFSA122-1 (CP033224, from Salmonella enterica of pork), pSH16G4466 (MK477617, FIGURE 2 | Sequence structure of aac(3)-IV-containing element in pHNAH67 (KX246266) and KpApr172 (this study). Colored arrows indicate open reading frames, with red, yellow and orange arrows representing resistance genes, mobile elements and additional genes, respectively. Light blue shading denotes regions of shared 100% homology between different elements.

DISCUSSION
Aminoglycosides resistance is mostly due to the chemical modification by aminoglycoside-modifying enzymes (AMEs), including acetyltransferases (AACs), phosphotransferases (APHs) and nucleotidyltransferases (ANTs) (Ramirez and Tolmasky, 2010). Some studies had shown AAC(6 )-Ib is one of the most prevalent and clinically relevant AMEs, which confers the resistance to amikacin and other aminoglycosides (Ramirez et al., 2013). A Greek study showed more than 70% of carbapenemases-producing Enterobacteriaceae strains carry aac(6 )-Ib gene (Galani et al., 2019). However, only 7.14% of CR-hvKp and 7.5% of CR-non-hvKp isolates were found to carry this gene in our study. By contrast, we found a high frequency of ARM/RMT methyltransferases gene in our study. ARM/RMT methyltransferases were found to confer high levels of resistance to most aminoglycosides other than apramycin (Galimand et al., 2003). ARM/RMT genes armA and rmt had been frequently described in clinical isolates throughout America, Europe and India (Fritsche et al., 2008). Our study showed 93.6% (44/47) CR-hvKp and 95.6% (22/23) CR-non-hvKp amikacin or gentamycin resistant strains contained the 16S rRNA methylase gene rmtB. These results suggested that gentamycin and amikacin have limited activities against clinical CRKp isolates, including CR-hvKp, in Chinese hospitals.
In comparison, our results showed that all CR-hvKp isolates were susceptible to apramycin, whereas ∼ 50% of them were resistant to amikacin or gentamycin. Similarly, almost all CR-non-hvKp strains (39/40) were susceptible to apramycin irrespective of the presence of KPC-2 or NDM carbapenemase genes. In addition, our results were in accordance with previous studies that apramycin is not be inactivated by general AMEs and methylases (Livermore et al., 2011), making it an potential antibiotic against clinical amikacin or gentamycin resistant strains.
Nevertheless, apramycin resistance has also been described in clinical isolates worldwide (Threlfall et al., 1986;Chaslus-Dancla et al., 1989, 1991Hunter et al., 1993;Johnson et al., 1995;Mathew et al., 2003;Zhang et al., 2009). Production of the acetylase aac(3 )-IV was the major mechanism underlying apramycin resistance in clinical isolates. The aac(3 )-IV gene encoding apramycin resistance was usually harbored by a mobile plasmid (Chaslus-Dancla et al., 1989;Zhang et al., 2009). In this study, we identified one aac(3 )-IV-harboring apramycin resistant isolate KpApr172 in CR-non-hvKp with a high MIC value of >128 µg/ml. Genomic analysis showed that KpApr172 belonged to the epidemic CRKp ST11 clone, while the aac(3 )-IV-harboring contigs showed highly identity to plasmid or chromosome sequences of isolates from animal sources. Since apramycin is only licensed for use in farm animals in China, we suspected that horizontal transfer of apramycin resistant genes from animals to human might be responsible for the dissemination of apramycin resistance (Yates et al., 2004). Therefore, aac(3 )-IV gene should be preferably screened for apramycin resistant isolates in clinical settings.
In this study no CR-hvKp isolates were found to be resistant to the "last resort" antimicrobial agents, including ceftazidimeavibactam, colistin and tigecycline, however, the treatment arsenal for CR-hvKp could be easily compromised by the spread of mobile resistant genes. Plasmid-mediated colistin and tigecycline resistance mechanisms, MCR-1 and Tet(x), have been described in China (Liu et al., 2016;He et al., 2019;Sun et al., 2019). In addition, the FDA approved cefetazidime-avibactam is ineffective against metallo-β-lactamase (e.g., NDM) producers. Furthermore, the original KPC-2 ST11 CR-hvKp outbreak study showed that despite being in vitro susceptible to tigecycline and colistin, long-term treatment with colistin or tigecycline (alone or in combination with several other antibiotics) was not able to eradicate KPC-2-producing ST11 CR-hvKp isolates, resulting in fatal outcomes (Gu et al., 2018). This convergence of increased resistance and hypervirulence of CR-hvKp underscores the need for the improvement of the antimicrobial treatment choices.
Our study showed apramycin had a potent in vitro activity to CR-hvKp irrespective of most commonly occurring carbapenemases and aminoglycoside modifying enzymes. In addition, recent studies had showed that apramycin was associated with fewer ototoxic and nephrotoxic side effects in comparison with other human use aminoglycosides antibiotics (Perzynski et al., 1979;Matt et al., 2012;Ishikawa et al., 2019). These findings suggested apramycin and apramycin-derivative compounds could be a promising antibiotic for the treatment of CR-hvKp infections.
To our knowledge, this is the first study to examine the activities of apramycin against CR-hvKp clinical isolates. Our results indicate that apramycin remain highly active against CR-hvKp. These data support a possible role for apramycin in the treatment of infections due to this hypervirulent and multidrugresistant strains. A further clinical evaluation of apramycin treatment efficacy is warranted.

DATA AVAILABILITY STATEMENT
The whole-genome sequence of KpApr172 was deposited in the GenBank database under accession number WUJI00000000.

AUTHOR CONTRIBUTIONS
MH, XS, and JL contributed to work, data analysis, and manuscript preparation. SN, SC, HD, and FY analyzed the data. Y-WT and BK prepared the manuscript. HZ and LC contributed to study design, data analysis, and manuscript preparation.

SUPPLEMENTARY MATERIAL
The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fmicb. 2020.00425/full#supplementary-material FIGURE S1 | Dendrogram of PFGE patterns of 15 CR-hvKp isolates from hospitals in Shanghai (a), Suzhou (b), and Jinan (c), respectively. PFGE typing revealed most of the CR-hvKp strains not belong to the same clone.
TABLE S1 | Primers used in this study.