Aminoglycoside Resistance

INTRODUCTION

Antimicrobial resistance has been recognized as one of the most pressing public health and societal issues of our times. The problem is most acute in gram-negative bacteria, where strains resistant to multiple (multidrug-resistant [MDR]) or almost all (extensively drug-resistant [XDR]) available agents are emerging.¹ Of particular concern has been the spread of extended-spectrum β-lactamase (ESBL)-producing Enterobacteriaceae in the 1990s, which was followed closely by the emergence and rapid dissemination of carbapenemase-producing organisms.

The three key classes of antimicrobial agents with gram-negative activity include β-lactams (especially β-lactam-β-lactamase inhibitor combinations, later-generation cephalosporins, and carbapenems), fluoroquinolones, and aminoglycosides. Aminoglycosides were identified through systematic screening of soil Actinobacteria that started in the 1940s. The first aminoglycoside streptomycin was discovered from Streptomyces griseus and successfully used for the treatment of tuberculosis and then infections with gram-negative bacteria. A typical aminoglycoside possesses an amino-containing or non-amino-containing sugars linked to six-membered rings with amino group substituents, hence the name aminoglycoside. Numerous aminoglycosides have since been identified or semisynthesized and used in clinical practice.

Aminoglycosides are grouped into 4,6-disubstituted 2-deoxystreptamine (DOS), 4,5-disubstituted DOS, and 4-monosubstituted DOS based on their chemical structures (Figs. 1 and 2). Representative 4,6-disubstituted DOS agents include gentamicin, tobramycin, and amikacin, which are widely used as intravenous or nebulized formulations for the treatment of infections caused by gram-negative bacteria (usually in combination with a β-lactam agent), gram-positive bacteria (for synergistic activity with a β-lactam or peptidoglycan), and atypical mycobacteria (again in combination with other active agents). 4,5-Disubstituted DOS agents, represented by neomycin, are limited in their utility by toxicity and are administered either orally or topically but not intravenously. Monosubstituted DOS agents are represented by apramycin, which is used in veterinary medicine.

Fig. 1.

Aminoglycosides whose activities are compromised by methylation of nucleotide G1405 or A1408 of 16S rRNA. (From Yonezawa M, Ida T, Umemura E, et al. Antibiotics and chemotherapy (kagakuryohou no ryouiki), 31:(1476)67–73, 2015, a review article written in Japanese; with permission.)

Fig. 2.

Aminoglycosides whose activities are not affected by methylation of nucleotide G1405 or A1408 of 16S rRNA. (From Yonezawa M, Ida T, Umemura E, et al. Antibiotics and chemotherapy (kagakuryohou no ryouiki), 31:(1476)67–73, 2015, a review article written in Japanese; with permission.)

MECHANISMS OF AMINOGLYCOSIDE RESISTANCE

Aminoglycosides bind to the aminoacyl-tRNA recognition site (A-site) of the 16S rRNA that constitutes the 30S ribosomal subunit, leading to inhibition of polypeptide synthesis and subsequent cell death. Resistance to aminoglycosides may occur based on several mechanisms: (1) enzymatic modification and inactivation of the aminoglycosides, mediated by aminoglycoside acetyltransferases, nucleotidyltransferases, or phosphotransferases and commonly observed across gram-positive and -negative bacteria^2,3; (2) increased efflux; (3) decreased permeability; and (4) modifications of the 30S ribosomal subunit that interferes with binding of the aminoglycosides. For the latter, both mutations (nucleotide replacement) and posttranscriptional modifications have been associated with aminoglycoside resistance. Examples include point mutations in the 16S rRNA and the rpsL gene encoding the S12 protein in Mycobacterium tuberculosis leading to streptomycin resistance.⁴ However, mutations within the 30S ribosomal subunit do not seem to be a common aminoglycoside resistance mechanism among fast-growing pathogenic bacteria in general.

However, posttranscriptional modification of the 16S rRNA is commonly observed among aminoglycoside-producing Actinobacteria, including Streptomyces spp and Micromonospora spp, which are naturally resistant to these metabolites of their own. This process is mediated by posttranscriptional methylation of either the N-7 position of nucleotide G1405 or the N-1 position of nucleotide A1408 on the 16S rRNA by various 16S rRNA methyltransferases (16S-RMTases) (Fig. 3).⁵ Intrinsic N7-G1405 16S-RMTases are found in both Streptomyces spp and Micromonospora spp, whereas intrinsic N1-A1408 16S-RMTases have only been identified in Streptomyces spp.⁶ N7-G1405 16S-RMTases confer resistance to 4,6-disubstituted DOS agents, such as gentamicin, tobramycin, and amikacin, but not 4,5-disubstituted and monosubstituted DOS agents. In contrast, N1-A1408 16S-RMTases are capable of conferring resistance all these groups.

Fig. 3.

Three-dimensional structure of the A-site of 16S rRNA bound to gentamicin.

ACQUIRED 16S RIBOSOMAL RNA METHYLTRANSFERASES IN GRAM-NEGATIVE BACTERIA

The first 16S-RMTases mediating aminoglycoside resistance and reported from outside aminoglycoside-producing Actinobacteria were ArmA in Klebsiella pneumoniae and RmtA in Pseudomonas aeruginosa.^7,8 ArmA was first identified in an isolate of K pneumoniae from the urine of a patient admitted to a hospital in Paris in 2000,⁷ but its nucleotide sequence could be found on a plasmid carried by a Citrobacter freundii clinical strain identified in Poland in 1996.^9,10 ArmA was subsequently confirmed to function as N7-G1405 16S-RMTase.¹¹ RmtA was identified in a sputum isolate of P aeruginosa in Japan in 1997.⁸

With the subsequent discovery of additional enzymes, a total of 9 N7-G1405 16S-RMTases, some with proven function (ArmA, RmtB, and RmtC) and others with putative function based on amino acid sequence similarity and resistance phenotype (ie, resistance to gentamicin, tobramycin, and amikacin but susceptibility to neomycin and apramycin), have been identified (Table 1). They include ArmA, RmtA, RmtB (which includes RmtB1 and RmtB2 alleles), RmtC, RmtD (which includes RmtD1 and RmtD2 alleles), RmtE, RmtF, RmtG, and RmtH (Table 2). These acquired N7-G1405 16S-RMTases share modest to high amino acid sequence similarities with each other.

Table 1.

Resistance phenotype of aminoglycosides conferred by acquired N7-G1405 and N1-A1408 16S-RMTases

Aminoglycosides	N7-G1405 16S-RMTase	N1-A1408 16S-RMTase
	RmtA through RmtH ArmA	NpmA
4,6-Disubstituted DOS (gentamicin, tobramycin, amikacin)	R	R
4,5-Disubstituted DOS (neomycin)	S	R
Monosubstituted DOS (apramycin)	S	R
No DOS ring (streptomycin)	S	S

Table 2.

Overview of acquired N7-G1405 and N1-A1408 16S-RMTases

16S-RMTase	Common Species	Common Coresistance	Prevalence	Distribution
ArmA	Klebsiella pneumoniae Acinetobacter baumannii	CTX-M ESBL NDM carbapenemase OXA-23 carbapenemase	Very high in A baumannii High among NDM producers	Worldwide
RmtA	Pseudomonas aeruginosa	—	Low	Japan, Korea
RmtB	Escherichia coli K pneumoniae	CTX-M ESBL NDM carbapenemase	High in China High among NDM producers	Worldwide
RmtC	K pneumoniae Proteus mirabilis	NDM carbapenemase	High among NDM producers	India, United Kingdom
RmtD	P aeruginosa K pneumoniae	CTX-M ESBL KPC carbapenemase	Low	South America
RmtE	E coli	CMY-2 AmpC	Very low	United States
RmtF	K pneumoniae	NDM carbapenemase	High among NDM producers	India, United Kingdom
RmtG	K pneumoniae	CTX-M ESBL KPC carbapenemase	Low	South America
RmtH	K pneumoniae	CTX-M ESBL	Very low	Iraq
NpmA	E coli K pneumoniae Enterobacter spp	—	Very low	Japan, Saudi Arabia

Abbreviation: NDM, New Delhi Metallo-β-lactamase.

In contrast, only a single acquired N1-A1408 16S-RMTase has been discovered to date. This enzyme, named NpmA, was identified from an Escherichia coli clinical strain in Japan in 2007.¹² NpmA confers a broader spectrum of aminoglycoside resistance that includes neomycin and apramycin in addition to gentamicin, tobramycin, and amikacin, and possesses an amino acid sequence that is distinct from those of the N7-G1405 16S-RMTases (see Table 1). NpmA has subsequently been shown to catalyze modification of nucleotide A1408.¹³

Reviewed next are the genetic context and epidemiology of the acquired 16S-RMTases that have been identified to date.

ArmA

The armA gene was initially identified on plasmid pCTX-M3 in C freundii and plasmid pIP1204 in K pneumoniae, both of which also carried an ESBL gene bla_CTX-M-3.^10,14 In these plasmids, armA is located downstream of insertion sequence ISCR1, which follows a class 1 integron comprising dhfrA12 (trimethoprim resistance), aadA2 (streptomycin resistance), and sul1 (sulfonamide resistance). Typically, mel and mph2 (macrolide resistance) are located downstream of armA. Although the incompatibility groups of the plasmids and the gene cassette contents of the class 1 integron may be variable, this overall genetic context of armA seems to be well conserved in the family Enterobacteriaceae and Acinetobacter baumannii.

armA is one of the most frequently encountered 16S-RMTase genes along with rmtB, and is widely distributed in Enterobacteriaceae and A baumannii. In Enterobacteriaceae, it has been found mostly in species, such as K pneumoniae, involved in health care–associated infections, but there are also reports of armA found in species implicated in food-borne and diarrheal illnesses, including Salmonella enterica^14,15 and Shigella flexneri.¹⁴ In A baumannii, armA is often identified in carbapenem-resistant strains that also carry the acquired carbapenemase gene bla_OXA-23, although these two genes are located on separate plasmids.¹⁶ In a study of carbapenem-nonsusceptible A baumannii isolates collected from hospitals across the United States, 49% of the isolates carried armA, suggesting a high prevalence in this species (Doi Y, 2011, unpublished data).¹⁷

Another specific group of bacteria where armA seems to have high prevalence is Enterobacteriaceae producing NDM (New Delhi Metallo-β-lactamase)-type carbapenemase. NDM-1, the first NDM-type carbapenemase, was initially reported in K pneumoniae and E coli strains that were isolated from a patient who had traveled from New Delhi, India to Sweden in 2009.¹⁸ NDM-producing Enterobacteriaceae has since spread worldwide rapidly causing many outbreaks, including recent ones in Denver and Chicago.^19,20 Soon after the discovery of NDM-1 it became clear that many Enterobacteriaceae strains producing NDM-1 or its variants were also highly resistant to 4,6-disubstituted DOS agents including gentamicin, tobramycin, and amikacin.²¹ Investigations into the bla_NDM-carrying plasmids have revealed that bla_NDM is frequently colocated with armA, or other 16S-RMTase genes (especially rmtB, rmtC, and rmtF), on the same plasmids.²² Some of these plasmids additionally carry plasmid-mediated fluoroquinolone-resistance genes, such as qnrB1.²³ Acquisition of these MDR plasmids would therefore simultaneously confer resistance to most β-lactams including carbapenems, aminoglycosides, and fluoroquinolones, the three key groups of agents with activity against gram-negative bacteria. armA has also been found in P aeruginosa carrying another metallo-β-lactamase gene bla_IMP-1 in Korea,²⁴ and K pneumoniae or other Enterobacteriaceae carrying a Klebsiella pneumoniae carbapenemase (KPC)-type carbapenemase gene bla_KPC-2 in Italy and China.^25,26

Finally, there seems to be a reservoir of armA in food animals. For instance, armA has been reported in E coli from chickens in China,²⁷ and S enterica in chicken meat at a supermarket in La Réunion Island.²⁸ However, the extent of spread of armA in food animals is unclear at this point.

RmtA

RmtA was first identified in Japan in a P aeruginosa clinical isolate that had been isolated in 1997 and showed high-level resistance to the 4,6-disubstituted DOS aminoglycosides.⁸ The 6.2-kb genetic region including rmtA of P aeruginosa is located between two copies of a kappa-gamma element, a 262-bp possible mobile element.²⁹

Compared with ArmA, the occurrence of RmtA has been sporadic so far, with a limited number of reports of rmtA-carrying P aeruginosa coming from Japan and Korea.^30,31 However, a plasmid carrying bla_NDM-1 and rmtA was reported in a K pneumoniae clinical strain that was isolated from a patient who was hospitalized in India and later sought care in Switzerland.³² This association suggests that there may be an unrecognized reservoir of rmtA in India, where other 16S-RMTase genes are frequently found with bla_NDM-1.

RmtB

RmtB was first reported from a Serratia marcescens clinical strain that was isolated in Japan in 2002.³³ rmtB is located on a plasmid and downstream of a Tn3-like transposon, and bla_TEM-1. Although the sequences downstream of rmtB are variable, a fluoroquinolone efflux gene qepA1 or its variants can sometimes be found adjacent to rmtB.

rmtB has revealed a worldwide distribution among Enterobacteriaceae, with reports of its identification coming from Asia, the Americas, Europe, the Middle East, Africa, and Oceania.^6,34–36 Like armA, rmtB is frequently associated with bla_NDM-1 on the same plasmids.³⁷ Therefore, the ongoing spread of bla_NDM-1-carrying Enterobacteriaceae likely contributes significantly to further dissemination of rmtB.

Another unique feature of RmtB, as with ArmA, is its association with food animals, but RmtB seems to be more prevalent than ArmA in this ecological niche. Most reports come from China, with rmtB identified in high rates in E coli from pigs, farm workers and their environment,^38,39 chickens,²⁷ and pets.⁴⁰

RmtC

RmtC was first reported in a Proteus mirabilis clinical strain that was isolated from a hospitalized patient in Japan in 2003.⁴¹ Located on a nonconjugative plasmid, rmtC is adjacent to an ISEcp1-like element, which also provides the promoter sequence for the expression rmtC.⁴²

The second RmtC-producing isolate was reported from Australia in P mirabilis, which was isolated from the urine of a patient who had recently returned from India.⁴³ Then, 13 clonally related S enterica Virchow strains among the 2004 to 2008 culture collection at the Health Protection Agency in the United Kingdom were found to possess rmtC.⁴⁴ Remarkably, 4 of the 12 patients affected by these S enterica Virchow isolates reported recent travel to India. Soon thereafter, rmtC began to appear in conjunction with bla_NDM-1, just as had been observed with armA and rmtB. rmtC was found in 12 of 18 bla_NDM-1-carrying E coli isolates from the United Kingdom, India, and Pakistan.⁴⁵ In New Zealand, all 5 bla_NDM-1-carrying Enterobacteriaceae isolates referred to a national reference laboratory possessed rmtC, with all cases associated with recent health care contact in India.⁴⁶ In a survey at a hospital in India, 3.7% of Enterobacteriaceae isolates had rmtC, often along with bla_NDM.⁴⁷ rmtC has also been found among K pneumoniae clinical isolates in Nepal.⁴⁸ These data suggest that rmtC likely originated in the Indian subcontinent and rmtC is being incorporated by MDR/XDR Enterobacteriaceae, in particular those producing NDM-type carbapenemase.

RmtD

RmtD is unique in that it has only been reported from South America. First identified in P aeruginosa producing SPM-1 metallo-β-lactamase in Brazil,⁴⁹ rmtD (now also termed rmtD1 because of the identification of its variant rmtD2) seems to have been mobilized by putative transposase ISCR14 (formerly Orf494).⁵⁰ rmtD1 and rmtD2 have since been identified in various Enterobacteriaceae species in Brazil, Chile, and Argentina.^51,52 More recently, cocarriage of rmtD1 or rmtD2 with bla_KPC-2 has been reported in K pneumoniae from Brazil.⁵³ Unlike with NDM-1-producing Enterobacteriaceae, rmtD1 or rmtD2 and bla_KPC are located on separate plasmids.

RmtE

Only two RmtE-producing strains have been reported to date, both E coli, one from a cow and another from a patient in the United States.^54,55 rmtE is located on a class 1 integron, but as an independent gene and not an integron gene cassette, on a large plasmid that also carries bla_CMY-2 encoding an AmpC β-lactamase (cephalosporinase).⁵⁶

RmtF

More so than any other acquired 16S-RMTases that have been identified to date, RmtF is very closely associated with NDM. The first RmtF-producing strain identified was K pneumoniae coproducing NDM-1 that was isolated from a patient in La Réunion Island.⁵⁷ rmtF has so far been found on class 1 integrons downsteam of bla_NDM.⁵⁸

In a single hospital surveillance study in India, 3.4% of Enterobacteriaceae carried rmtF, 59% of them along with bla_NDM.⁴⁷ Isolates carrying rmtF have since been reported from Nepal, Australia, and Minnesota.^16,36,48,59

RmtG

Like RmtD, RmtG is largely unique to South America. First identified in KPC-producing K pneumoniae clinical isolates in Brazil,⁵³ it has been found in Chile and Miami, all in K pneumoniae.^60,61 rmtG is flanked downstream by ISVsa3, but this insertion sequence likely belongs to the site of insertion of the module carrying rmtG rather than the rmtG-containing module itself.⁶² There are currently no data on the epidemiology of rmtG.

RmtH

RmtH was recently identified in an ESBL-producing K pneumoniae strain that was isolated from a US soldier who suffered wound infection from an explosion during deployment in Iraq.⁶³ rmtH is bracketed by two copies of ISCR2, which likely played a role in the initial mobilization of this unusual 16S-RMTase gene.

NpmA

NpmA was discovered in an E coli clinical strain that was isolated from a patient in Japan in 2003 and is the only acquired N1-A1408 16S-RMTases known to date.¹² Because of its unique site of action, NpmA confers resistance to 4,5-disubstituted and monosubstituted DOS agents (eg, neomycin and apramycin, respectively). npmA and its flaking sequences are bracketed by the two copies of IS26, suggesting their involvement in the mobilization of npmA. K pneumoniae and Enterobacter spp carrying npmA were recently reported from Saudi Arabia.³⁵

PREVALENCE OF 16S RIBOSOMAL RNA METHYLTRANSFERASES

Prevalence data on acquired 16S-RMTases remain relatively scarce, and low prevalence rates of 1% or less have been reported among Enterobacteriaceae from Europe, Japan, and Argentina.^{30,52,64–67} However, some studies, most of them single-center, have found alarmingly high rates of 16S-RMTase genes among Enterobacteriaceae. In Korea, rates of 2.8% to 11.4% have been reported among Enterobacteriaceae.^68,69 In China, of 680 and 337 unique E coli and K pneumoniae clinical isolates collected at a hospital between 2006 and 2008, 5.4% and 6.2% were positive for rmtB or armA, respectively.^70,71

The prevalence rates may be even higher in India and Saudi Arabia. Of 1000 consecutive Enterobacteriaceae clinical isolates collected at a hospital in India between 2010 and 2011, a total of 14% carried at least one 16S-RMTase gene, including armA, rmtB, rmtC, and rmtF.⁴⁷ Of 330 unique Enterobacteriaceae clinical isolates collected at a hospital in Saudi Arabia in 2011, a total of 37% carried at least one 16S-RMTase gene.³⁵

CLINICAL IMPLICATIONS

To date, no data are available on the impact of acquired 16S-RMTase production and clinical outcome of patients when they are treated with aminoglycosides, largely because aminoglycosides by themselves are not considered the first-line agents for either empiric or definitive therapy for infections from gram-negative bacteria in most clinical settings, and also because prevalence rates of 16S-RMTases are relatively low in developed countries.

Nonetheless, the real threat of 16S-RMTases is the loss of a potential treatment option for the salvage therapy for MDR/XDR gram-negative bacterial infections where treatment options are already limited. 16S-RMTase genes are increasingly identified along with other significant resistance genes, especially carbapenemase genes, in the same isolates. Given the extremely high level of aminoglycoside resistance conferred by the acquired 16S-RMTases, this precludes the use of key aminoglycosides (gentamicin, tobramycin, and amikacin) even when carbapenems have already been excluded from the treatment option. For instance, plazomicin, a new aminoglycoside agent that is under clinical development specifically for use against carbapenem-resistant Enterobacteriaceae, is not active against isolates that produce acquired 16S-RMTase. This is most prominently observed for NDM-producing Enterobacteriaceae that frequently coproduces 16S-RMTase.⁷² Although most KPC-producing K pneumoniae remain susceptible to plazomicin at this point, increasing reports of KPC and 16S-RMTase coproduction are a cause for concern.^25,26,53,73

SUMMARY

Antimicrobial resistance in gram-negative pathogens has become one of the most challenging issues faced in daily clinical practice. Especially troublesome is the global spread of carbapenem-resistant Enterobacteriaceae, which has accelerated since the appearance of KPC and NDM-type carbapenemases. Aminoglycoside resistance mediated by acquired 16S-RMTase is a relatively new mechanism that was described in the early 2000s, but it now seems to be converging with the carbapenemase epidemic, thereby facilitating the emergence of XDR and, in some instances, pandrug-resistant organisms. Although not first-line in many clinical scenarios, aminoglycosides remain an important class of agents with excellent bactericidal activity when the organisms of interest are resistant to other classes, especially β-lactams and fluoroquinolones. Therefore, the ongoing dissemination of 16S-RMTases among already MDR organisms is an unwelcome event.

KEY POINTS.

• Aminoglycoside-producing Actinobacteria produce 16S ribosomal RNA methyltransferase (16S-RMTase) to protect themselves.

• High-level aminoglycoside resistance caused by production of acquired 16S-RMTase in pathogenic gram-negative bacteria was first reported in the early 2000s.

• Bacteria that produce 16S-RMTases frequently coproduce ESBL, and more recently, carbapenemase, especially NDM-1.

• Spread of 16S-RMTase-producing bacteria further compromises the already limited treatment options for infections caused by MDR/XDR pathogens.