Molecular characterization of carbapenemase-producing Enterobacterales in a tertiary hospital in Lima, Peru

ABSTRACT Carbapenemase-producing Enterobacterales (CPE) are a growing threat to global health and the economy. Understanding the interactions between resistance and virulence mechanisms of CPE is crucial for managing difficult-to-treat infections and informing outbreak prevention and control programs. Here, we report the characterization of 21 consecutive, unique clinical isolates of CPE collected in 2018 at a tertiary hospital in Lima, Peru. Isolates were characterized by phenotypic antimicrobial susceptibility testing and whole-genome sequencing to identify resistance determinants and virulence factors. Seven Klebsiella pneumoniae isolates were classified as extensively drug-resistant. The remaining Klebsiella, Enterobacter hormaechei, and Escherichia coli isolates were multidrug-resistant. Eighteen strains carried the metallo-β-lactamase NDM-1, two the serine-carbapenemase KPC-2, and one isolate had both carbapenemases. The blaNDM-1 gene was located in the truncated ΔISAba125 element, and the blaKPC-2 gene was in the Tn4401a transposon. ST147 was the most frequent sequence type among K. pneumoniae isolates. Our findings highlight the urgent need to address the emergence of CPE and strengthen control measures and antibiotic stewardship programs in low- and middle-income settings. IMPORTANCE Genomic surveillance of antimicrobial resistance contributes to monitoring the spread of resistance and informs treatment and prevention strategies. We characterized 21 carbapenemase-producing Enterobacterales collected at a Peruvian tertiary hospital in 2018, which exhibited very high levels of resistance and carried numerous resistance genes. We detected the coexistence of carbapenemase-encoding genes (blaNDM-1 and blaKPC-2) in a Klebsiella pneumoniae isolate that also had the PmrB(R256G) mutation associated with colistin resistance. The blaKPC-2 genes were located in Tn4401a transposons, while the blaNDM-1 genes were in the genetic structure Tn125 (ΔISAba125). The presence of high-risk clones among Klebsiella pneumoniae (ST11 and ST147) and Escherichia coli (ST410) isolates is also reported. The study reveals the emergence of highly resistant bacteria in a Peruvian hospital, which could compromise the effectiveness of current treatments and control.

difficult to treat due to the limited effectiveness of available drugs, forcing clinicians to revert to older medications such as colistin (2,3).
Carbapenemases, as categorized by Ambler's classification, include Class A Klebsiella pneumoniae carbapenemase (KPC), Class B metallo-β-lactamases (MBLs) like New Delhi MBL (NDM), and Class D oxacillinases (OXA) such as OXA-48-like enzymes found in Enterobacterales (2).Klebsiella pneumoniae and Escherichia coli, among the CPE, have emerged as significant global health threats due to their high mortality rates (4).The widespread dissemination of CPE is primarily attributed to the horizontal transfer of antibiotic-resistance genes via mobile genetic elements, including plasmids and transposons (5).Various virulence factors enhance the pathogenicity of CPE.While there is extensive knowledge about the interplay between these factors and AMR in many bacterial pathogens (6), the specific genetic determinants that drive virulence and their interaction with resistance mechanisms are not fully understood (7).
In Peru, the first CPE infections were identified in late 2013 as a carbapenem-resistant K. pneumoniae, sequence type (ST) 340 of the high-risk clonal complex (CC) 258 (8,9).Subsequent reports indicate the spread of various carbapenemases in clinical settings and non-traditional hosts such as Providencia stuartii (10,11).By 2020, the spread of the metallo-β-lactamase NDM-1 in K. pneumoniae was evident, involving newer lineages like ST348 and ST147 (12).Despite numerous CPE reports in Peru and Latin America, comprehensive genomic data are needed to conduct in-depth evolutionary, molecular, and epidemiological studies.Understanding the relationships between different CPE strains, both locally and globally, is essential.To address this gap, we sequenced the genomes and analyzed the resistance and virulence determinants of 21 CPE samples from a tertiary hospital in Lima collected in 2018.

Genetic context of carbapenemase genes
The bla KPC-2 gene was found in K. pneumoniae KpCR01 and KpCR02, and E. coli EcCR01, located in a Tn4401a transposon of the Tn3 family.The gene was flanked by ISKpn6 and ISkpn7 elements near the tnpA and tnpR transposase genes (Fig. 3A).The blaNDM-1 gene in K. pneumoniae strain KpCR09, E. hormaechei strain EnCCR02, and E. coli strain EcCR02 was found within the ΔTn125 element (Fig. 3B).The bla NDM-1 gene was flanked by ISA-ba125 and the blaMBL gene (Fig. 3B) in both reference strains and the three isolates described here.In K. pneumoniae strain KpCR09, the blaTEM-26 and aac (3)-lle genes, which encode resistance to third-and fourth-generation cephalosporin and aminoglyco side-modifying enzyme (AME), respectively, were identified upstream of ΔTn125.

Virulence factors
K. pneumoniae isolate KpCR01 presented the yersiniabactin ybt9 gene located on the integrative conjugative element (ICE) Kp3, along with the genes encoding for the wzi50 capsular type and antigenic determinants K15 and O4.In contrast, isolate KpCR11 exhibited the ybt0 gene within the ICEKp12 element, accompanied by the wzi94 capsular type gene and antigenic determinants K62 and O1.Notably, all K. pneumoniae ST147 isolates in our study had the aerobactin (iuc5) gene, genes encoding for the wzi64 capsular type, and antigenic determinants KL64 and O2a.Additionally, in our two E. coli isolates, we identified several virulence genes, including fimH, associated with type 1 fimbriae; iucC, related to aerobactin synthetase; iutA, encoding the ferric aerobactin receptor; lpfA, indicative of long polar fimbriae; and sitA, associated with iron acquisition (Table S2).

DISCUSSION
The rapid spread of CPE poses a significant threat to global public health.Our study of CPE from one Peruvian hospital in 2018 confirms the increasing prevalence of bla NDM-1 over bla KPC-2 in Peru since 2013, alongside the emergence of OXA-48-like carbapenema ses, particularly OXA-181 (3)(4)(5)13).
Our findings confirm high resistance levels in CPE, particularly in Klebsiella pneumo niae.Notably, the KpCR02 isolate exhibited up to 21 AMR genes, including NDM-1 and KPC-2.All NDM-1-producing CPE showed resistance to β-lactams, including aztreonam, often due to AmpC and ESBL genes, with the blaCTX-M-15 gene being particularly prevalent in our data set (14,15).The co-production of carbapenemases NDM-1 and KPC-2 in K. pneumoniae has been previously described in South America, including Brazil, Argentina, Uruguay, Ecuador, and Paraguay (16).Our report provides further evidence of this occurrence and regional dispersion of CPE co-producing NDM-1 and KPC-2 carbapenemases.
The aminoglycoside resistance observed in our isolates could be attributed to genes that encode AMEs.The rmtG gene involved in 16S rRNA methylation confers high-level resistance to aminoglycosides (minimum inhibitory concentration, MIC ≥ 256 µg/mL) and is widely spread in K. pneumoniae found in South America (17).Among the AMEs, the acetyltransferases aac(3)-lle and aac(6')Ib-cr5 were the most frequently detected.The aac (3)-IIa has been associated with resistance to gentamicin, tobramycin, and netilmicin, as described elsewhere (18).The aac(6')Ib-cr is characterized by W102R and D179Y/G substitutions [both substitutions being present in aac(6')Ib-cr5], which, in addition to its capacity to inactivate several aminoglycosides, also result in the ability to inactivate several fluoroquinolones such as ciprofloxacin and norfloxacin (19).Nine K. pneumoniae and all E. coli isolates had amino acid substitutions at GyrA and ParC, the most common quinolone resistance mechanism (20).The remaining CPE did not present mutations in the established quinolone targets.However, all presented at least two transferable mechanisms of quinolone resistance and alterations in efflux pump regulators in one case.While highly unusual several years ago, this scenario is increasingly being described (19,20).
Resistance to colistin observed in K. pneumoniae isolates could be explained by the amino acid substitution detected in the PmrB (R256G) of 10 isolates, one of which also had the PhoQ(L348Q) mutation.Remarkably, both mutations are associated with lipid A modification (21).PmrB amino acid substitution (Y358N) was detected in two E. coli isolates; however, both were susceptible to colistin.Although the polymorphism may not be ruled out, this amino acid substitution has been previously described in E. coli isolates with high levels of colistin resistance (MIC >32 µg/mL) in the absence of mcr genes or other alteration in PmrAB, PhoPQ, or MgrB.Therefore, this single substitution may play a role in the development of colistin resistance (22).
Our genomic analysis found ST147 to be the dominant clone in K. pneumoniae, known for its high drug resistance and global dissemination.This includes the KpCR02 isolate, exhibiting both KPC-2 and NDM-1 and colistin resistance through pmrB mutations.The global emergence of KPC has been influenced by the wide dispersion of K. pneumoniae CC258, including ST11 and ST340 (23).ST11 is endemic in Brazil and several European countries (24,25), and ST340 has been previously reported in Peru (7).Several K. pneumoniae ST11, ST147, and ST348 outbreaks have also been reported in Peru (10,26).Our report of K. pneumoniae KPC-producing ST11 adds evidence of a high-risk international clone in Peru.In contrast to other reports, K. pneumoniae ST348 isolate KpCR11 described in this study significantly differs from other ST348 strains characterized by CTX-M-15 and KPC-3 production ( 27).The study reports K. pneumoniae ST4872, K. quasipneumoniae ST4952, and K. aerogenes ST215 for the first time in Peru.Among E. coli strains, ST471 and ST88 were identified, known for their resistance and virulence, alongside the first reported cases of E. hormaechei ST2054 in the region.
Analysis of the genetic context of the bla KPC-2 gene revealed that it was located in the Tn4401a transposon, a member of the Tn3 transposon family (28,29).To our knowledge, this is one of the few reports on Tn4401a in E. coli (30).Of note, current data differ from the previous genetic environment of bla KPC-2 detected in Peru, highlighting the presence of different genetic structures carrying this gene in the country (7).Furthermore, the analysis of the genetic context of the bla NDM-1 gene revealed that it was located in a truncated structure of Tn125.The architecture of the characterized ∆Tn was identical to those previously reported in Peru and other Latin American countries (31,32), and no differences were observed among the other CPE described here.
Hypervirulence is associated with the presence of additional siderophores, yersinia bactin (ybt), aerobactin (iuc), salmochelin (iro), and specific capsular serotypes (K1, K2, and K5) (33).Interestingly, the aerobactin iuc5 gene was detected in all K. pneumoniae ST147 samples.Besides potentially playing a critical role in vitro and in vivo virulence, it is a marker to identify highly virulent strains (33).This finding is compatible with that described in other CC258 isolates (34); however, in our isolate, neither characteris tic virulence determinants such as salmochelin or colibactin nor genetic determinants associated with hypermucoviscosity (rmpA) were detected.To date, 12 distinct O loci have been identified, with O1 and O2 as the most common antigens (35), while K-type variation has been associated with six conserved genes (galF, orf2, wzi, wza, wzb, and wzc).The wzi gene best predicts K antigen-related virulence (36).
The study's limitations include a need for gene expression data and a limited sampling period from one healthcare center in 2018, which restricts the broader application of our findings.Instead, the results described here should be considered reference data and a starting point for further studies using more complex designs and sampling methods.
Overall, our comprehensive analysis of 21 CPE isolates underscores the significant challenge posed by drug-resistant CPE strains, emphasizing the need for expan ded genomic surveillance, effective infection control, and antimicrobial stewardship programs to curb the spread of CPE and maintain the efficacy of antimicrobial agents, particularly in the understudied Latin American region.

Clinical isolates
From January to December 2018, 21 consecutive, non-replicated, and unique suspected CPE isolates were chosen based on their phenotypic resistance to meropenem at an MIC of ≥1 µg/mL using disk-diffusion assays.Clinical isolates were obtained from blood, urine, and other biological samples from inpatients and outpatients admitted to the tertiary public hospital, Hospital Nacional Arzobispo Loayza in Lima, Peru.The hospital's laboratory characterized all isolates at the species level using a standard biochemical test panel (37) and sent them to Universidad Peruana Cayetano Heredia (UPCH) for phenotypic and genomic characterization.

Carbapenemase screening and identification
The phenotypic screening of carbapenemase production was done with the Triton Hodge Test and the carbapenem inactivation method as previously described (38).Inhibition tests using EDTA and phenylboronic acid were used to distinguish classes of carbapenemases (39).

Antimicrobial susceptibility testing
Resistance phenotypes were determined using the VITEK2 System with AST-N249 cards (bioMérieux, France) following the manufacturer's instructions.MIC values were interpreted using clinical breakpoints established by the Clinical and Laboratory Standards Institute (CLSI) M100 guidelines (40).The European Committee for Antimicro bial Susceptibility Testing breakpoints were used for tigecycline interpretation since CLSI breakpoints were not available at the time of the study (41).Colistin susceptibility was assessed by the colistin broth disk elution test as previously described (42).Intermediate resistance results were analyzed and reported as resistant, and the criteria suggested by Magiorakos et al. were used to classify bacteria as MDR, XDR, or PDR (43).

DNA extraction and whole-genome sequencing
Genomic DNA was extracted using the GeneJET DNA purification kit (Thermo Fisher Scientific, USA) from single colonies that were incubated in 1 mL of tryptic soy broth (Becton & Dickinson, USA) at 37°C and shaking at 300 rpm for 8 hours.The extracted gDNA was quantified with a Qubit 4.0 fluorometer, and Illumina libraries were construc ted using 1 ng of gDNA with the Nextera XT kit (Illumina, USA).Paired-end libraries were sequenced on an Illumina MiSeq instrument generating 250-nt reads resulting in a mean per-base coverage of 75× (minimum 51× and maximum 108×).

FIG 3 (
FIG 3 (A) Genetic context of the structures of the Tn4401a transposon carrying bla KPC-2 (GenBank accession number KT378596) and (B) genetic context of Tn125 (GenBank reference number JN872328) in our Enterobacterales isolates.Shading indicates 100% sequence similarity.