Carbapenemases in Enterobacteriaceae – overview and importance

Introduction

In February, the World Health Organization published a list of bacteria for which there is an urgent need of new antibiotics(1). The critical list includes Enterobacteriaceae that are carbapenem-resistant or producers of extended-spectrum beta-lactamases (ESBL). Carbapenems are frequently the last antimicrobials that can be used in multidrug-resistant infections, especially in invasive ones or those associated with immunosuppression. Like all beta-lactams, their action is due to the ability of “permanently” acylate multiple different penicillin-binding proteins (PBP) at the same time, which results in the inhibition of cell wall formation, activation of autolytic cell wall hydrolases and destruction due to the osmotic pressure(2-4).
Carbapenem-resistant Enterobacteriaceae infections occur mainly in the hospital setting, but there have been reports of community spreading as well(5-7). The purpose of this short review is to highlight the carbapenemase diversity and have a close-up at the most frequently and clinically important sub-types of carbapenemases.

Carbapenemase classification

Carbapenem resistance may be caused by multiple mechanisms. The most frequent mechanism, that is also the most important both clinically and epidemiologically, is the production of one or multiple enzymes that have the ability of degrading carbapenems. These enzymes are called carbapenemases and they belong to the large group of beta-lactamases. There are different classifications of beta-lactamases based on their structure or function, but one of the most widely used is the revised Ambler classification(8). According to it, there are three molecular classes of carbapenemases identified in Enterobacteriaceae: A, B and D (enzymes in class C are known as “Amp-C” beta-lactamases, a type of cephalosporinases that hydrolyze carbapenems very poorly). Genes encoding these different types of carbapenemases are located either chromosomally, plasmidic or both.
Others mechanisms include the association of impermeability and another type of enzyme (usually an ESBL or a cephalosporinase) and they may be the subject of a future article. 

Class A

Chemically speaking, this class hydrolyses molecules using an active site serine at position 70(3). Along with classes C and D that have similar active-site serine β-lactamases, they form intermediates of acyl-enzyme, thus differentiating them from class B carbapenemases(3). This property allows them to degrade beta-lactams at various degrees, from penicillins (for which they have a high affinity as substrate) to carbapenems. Their activity is inhibited by beta-lactamase inhibitors (clavulanate, sulbactam, tazobactam, avibactam) and in the case of Klebsiella pneumoniae Carbapenemase (KPC), by boronic acid and EDTA(4).
Genetically speaking, they may be chromosomally encoded (SME, NmcA, SFC-1, BIC-1, PenA, FPH-1, SSME, NmcA, SFC-1, BIC-1, PenA, FPH-1, SHV-38), plasmid encoded (KPC, GES, FRI-1) or both (IMI)(3). From a clinical point of view, the most frequently found class A carbapenemases are KPC and to a lesser extent, GES and IMI. From the 22 KPC described to date(3), KPC-1/2 in Klebsiella pneumoniae is the most important (soon after their discovery, resequencing of KPC 1 and 2 showed that they actually have the same structure)(9).
The carbapenemase gene blaKPC is located on a Tn3  transposon (Tn4410) and is associated with high dissemination in the environment. The worldwide spread of KPC in Klebsiella pneumoniae is currently connected to a single clone called ST-258 although within a certain location, the clones may present different characteristics regarding, for instance, the Multi Locus Sequence Typing (MLST) type or the structure of plasmids(3).
A recent study based on literature review showed Romania as KPC-endemic(10), although smaller studies in our country did not show similar results(11,34). 
GES-type carbapenemases are enzymes that usually have an ESBL-like spectrum of activity, but changes in the active site confers them the ability to hydrolyze carbapenems (substitution of Glycine at position 170 by an Asparagine or a 37 Serine)(12,13). There are 31 GES enzymes described to date out of which 10 are carbapenemases (GES-2, -4, -5, -6, -14, -14, -15, -16 -18, -20 and - 36). However, only GES-5 and GES-6 have been identified in Enterobacteriaceae(3,14). The ­blaGES genes have been described as gene cassettes associated with class 1 or class 3 integrons on plasmids with different types of replicases(3,14). 
IMI carbapenemases may be either chromosomally, or plasmid-encoded. There are 9 variants described to date(3). The gene blaIMI-1 and its variant blaIMI-3 are located chromosomally and were found in E. cloacae or E. asburiae with increasing reports of their presence in clinical samples(3). IMI-2 was firstly found in E. asburiae and it’s the only known plasmid-encoded variant of IMI. A strain of E. coli carrying the blaIMI-2 gene was observed in Spain, suggesting the possibility of transfer to other enterobacteriaceae and also the probable future association with nosocomial infections(3,15).

Class B

This class possesses the highest carbapenemase activity(16). Their spectrum of activity includes all beta-lactams (penicillins, cephalosporins, carbapenems), sparing the monobactam aztreonam. Carbapenemases belonging to this class are called metallo-beta-lactamases (MBL) and, chemically speaking, they need one or two Zn2+ in the active site in order to exert their hydrolyzing ability(17). Chromosomal genes coding them were first described in environmental bacteria such as Bacillus cereus, Aeromonas spp. and Stenotrophomonas maltophilia(4), but after 1990 they were more and more associated with transferable mechanisms. MBLs are inhibited by the metal ion chelator EDTA and sodium mercaptoacetate (SMA) but, due to their toxicity, they cannot be used as treatment(18).
The most important MBLs are IMP (imipenemase), VIM (Verona Integron-encoded Metallo-β-lactamase) and NDM (New Delhi Metallo-beta-lactamase). IMP and VIM  are most frequently transmitted by class 1 integrons through cassettes which allows them to integrate resistance genes that encode for other classes of antimicrobials as well(4).  
There are 60 IMP variants described in Enterobacteriaceae, Acinetobacter spp., Pseudomonas spp. to date. The first IMP reported was in Serratia marcescens isolate in Japan(4). Most of blaIMP genes are usually embedded in transposons and/or plasmids which help them spread horizontally and usually co-exist with other resistance genes, such as aacA, catB, and blaOXA, making such infections extremely difficult to treat(18).
The first VIM variant described in Enterobacteriaceae was in 2002 in Athens(10). At the moment, VIM-2 is the most reported MBL worldwide and it’s endemic in Southern Europe and Southeast Asia(4). A study showed that all-cause 14-day mortality in VIM-1 K. pneumoniae ranged from 18.9% in carbapenem-susceptible (MIC <=4 µg/mL) infections to 42.9% in those resistant to carbapenems(19). There have also been described pandrug-resistant VIM strains, rising concerns regarding future spreading of this type of strains(20).  
Since they were first discovered in 2008, NDM-1 producing bacteria have been mostly associated with the Indian subcontinent as it is considered the main reservoir for blaNDM-1 genes(4). However, there have been reports worldwide(21-23). From the 16 currently described variants in infected or colonized patients, most are somehow linked to India, Pakistan, Sri Lanka or Bangladesh (by travel, hospitalization, colonization due to contaminated water etc.).
The blaNDM-1 gene is associated with different plasmid types that vary in size and associated resistance genes(4). These associated resistance genes give a large number of resistance patterns as they might encode other carbapenemases, AmpC cephalosporinases, ESBLs along with aminoglycoside (armA, rmtA and rmtC), macrolide (ereC), rifampicin (arr-2), sulfamethoxazole (sul2), quinolones (qnrA6 and qnrB1) and chloramphenicol (cmlA) resistance(4,24). Also, blaNDM-1 is located in the same operon as the gene encoding for a functional bleomycin resistance protein, bleMBL, both being controlled by the same promoter, ISAba125, located upstream of blaNDM-1(4,17). 

Class D

Carbapenemases in this class are called oxacillinases with over 350 variants described to date(25). Their enzymatic activity is inhibited in vitro by NaCl. Oxacillinase-48 (OXA-48) and its variants were identified only in Enterobacteriaceae, not in Acinetobacter spp., and it’s the most clinically important class D carbapenemase as it has the most efficient imipenem-hydrolyzing activity from the class(26). OXA-48 producers hydrolyze penicillins at a high level but may hydrolyze carbapenems poorly, thus presenting high inhibition diameters on the antibiogram resulting in errors of interpretation and, therefore, in treatment(27). They resist the action of beta-lactamase inhibitors, but they remain susceptible to third generation cephalosporins (in the absence of other associated resistance mechanisms) which results in the probable underestimation of OXA-type producers(4). 
Usually, OXA-48 strains don’t have associated resistance to other classes of antibiotics because blaOXA-48 is linked to a single IncL/M type self-conjugating plasmid(4,26,28). However, its plasmidic location makes it easy to spread in the hospital settings and in community. Also, the association of different resistance mechanisms is increasing, a multiyear, multicenter study describing 3 strains of K. pneumoniae that co-carried NDM-1, OXA-48, and CTX-M-genes isolated in Romania(17). 
There are a series of OXA-like enzymes that have an identical or a very similar spectrum of hydrolysis with OXA-48, but they may be identified only by using PCR, not phenotypically on the antibiogram (OXA-181, OXA-162, OXA-204, OXA-232(28)). It is believed that OXA-type precursors belong to Shewanella ­species(29). Out of these variants, blaOXA-181, the gene coding for the production of OXA-181, has been associated with blaNDM-1 and blaVIM-5 genes and with the Indian subcontinent(28).

Discussions and importance

CPE infections are a growing problem worldwide. The rapid dissemination in the hospital settings and in the community along with the limited therapeutic options are causes of great concern and call for immediate measures. 
Regarding carbapenem-resistance through production of carbapenemases, it is important to know that not all carbapenemases are necessarily worrisome. If the genes carrying the carbapenemases production are encoded chromosomally, they will not spread horizontally; they will only carry on the gene to the next generation of bacteria. This means that from a clinical and epidemiological point of view, plasmid-encoded carbapenemases are those worth paying more attention to, because they can “pass” their resistance to other bacteria and cause epidemics. 
Also, it is very important to differentiate between cases of infection and carriage. CPE carriers excrete bacteria inconstantly for a long period of time and treatment is need only in selected cases. 
In cases of CPE infections, the specter of in vitro antibiotic hydrolysis has a great variability ­therefore every antibiogram must be carefully interpreted by the clinical microbiologist who must efficiently communicate with the clinician in order of taking the best therapeutic decision for the patient.
What will the future bring? The high and continuingly increasing incidence of ESBLs will most likely make carbapenems consumption high as well. High utilization of carbapenems will force bacteria to find ways of survival thus giving rise to pan-drug resistance. Because antibiotic development is a long process, the first thing to do is limiting the spread of these bacteria which can be done by taking rigorous hygiene measures. 
Regarding the treatment, there are articles recommending certain molecules or antibiotic associations, depending on the carbapenemase class, but further in vivo studies are necessary in order to make specific treatment protocols for CPE infections.
For class A carbapenemases, it appears that the last remaining treatment options are polymyxins, tigecycline and, if susceptible, aminoglycosides(30). ST-258 clones usually remain susceptible to gentamicin(30). However, not all infections may be treated with these molecules.
As class B carbapenemases are not able of hydrolyzing aztreonam, there have been studies regarding its activity, both alone and in combination with beta-lactamase inhibitors. A study comparing in vitro activity of aztreonam versus aztreonam-avibactam combination(31) showed increased activity of the combination against isolates producing class A, B or D carbapenemaes. Other studies assessed the activity of ceftazidime-avibactam (Avycaz) and showed that this combination has better activity on KPC and OXA-48 enzymes but lacks activity on MBLs(32). 
Fecal transplant is another promising method of intestinal decolonization(33). In the mean time, the development of new antibiotics should be a priority for the pharmaceutical companies, universities and other research groups.
 Faced with a post-antibiotic era, there is an important need for action. Educating doctors regarding good antibiotic prescription, as well as educating the population concerning the existence of CPE and what they can do about it (good hygiene, use of antibiotics only when necessary etc.) are simple methods that can limit the spread of these bacteria and potentially save lives.   n

Conflict of interests: The authors declare no conflict of interests.