Carbapenem-Resistant Klebsiella Pneumoniae  in Southwest China: Molecular Characterizations and Risk Factors Caused by NDM and KPC Producers


 Background: Klebsiella pneumoniae is one of the most common Enterobacteriaceae. In recent years, carbapenem-resistant Klebsiella pneumoniae (CRKP) has become one of the most important carbapenem-resistant Enterobacteriaceae. CRKP are usually resistant to antibiotics. Up to this day, the emergence of carbapenemase-producing K. pneumoniae has been a challenge for treatment of clinical infection.Methods: (i) 66 non-repetitive clinical CRKP isolates were identified by matrix-assisted laser analytical ionization time-of-flight mass spectrometer (MALDI-TOF-MS) and drug sensitivity analysis was performed by Vitek2 Compact. EDTA-synergy test and mCIM / eCIM test were used to detect drug-resistant phenotypes. (ii) Carbapenemase genes, extended-spectrum β-lactamase genes (ESBLs), cephalosporinase gene (AmpC), virulence genes, integron and resistance gene cassettes were amplified by PCR. (iii) Plasmid typing was performed by plasmid conjugation assay and PCR-based replicon typing (PBRT) method. (iv) The genetic environments of KPC-2 and NDM-1 were analyzed by using overlapping PCR. (v) MLST was used to analyze the molecular epidemiological characteristics of CRKP. (vi) Risk factors of CRKP infection by logistic regression model.Results: Our study revealed that 42 of the 66 CRKP isolates obtained from patients were identified as blaKPC-2, 24 blaNDM-1-positive strains were identified (20 blaNDM-1 and 4 blaNDM-5), of which 18 were from the neonatal departments. And CRKP strains were ESBL (extended-spectrum β-lactamases) and AmpC enzymes producer, Notably, we found two CR-hvKp (carbapenem-resistant hypervirulent klebsiella pneumoniae) strains, which contains blaKPC-2 gene and other resistant genes. Two of the 42 KPC-2-producing CRKP strains were positive for transconjugants, and the plasmid typing was the IncFII type. And two NDM-producing CRKP strains tested positive for transconjugants, which belonged to the lncX3 plasmid. Analysis of the genetic environment of these two genes has revealed that the highly conserved regions (tnpA-tnpR-ISkpn8-blaKPC-2) and conserved regions (blaNDM−1-bleMBL-trpF-tat) are associated with the dissemination of KPC-2 and NDM-1. Intl1 carrying drug resistance gene cassettes were widely distributed in CRKP. According to the MLST results, a total of 13 ST types were measured in 66 CRKP strains, ST11 and ST4495 were the main ST types, and the latter was the newly discovered ST type. Hematological disease, tracheal cannula and prior use of β-lactams and β-lactamase inhibitor combination were identified as independent risk factors for CRKP infections.Conclusion: These findings manifested the need for intensive surveillance and precautions to monitor the further spread of KPC and NDM in southwest China.


Introductin
Klebsiella pneumoniae (K. pneumoniae) is a common cause of opportunistic infections in hospitalized patients with antimicrobial resistance (1). it can often lead to aseptic site infections such as blood and other respiratory tract infections (2). Over the past decades, due to increasing prevalence of healthcare-associated infections caused by multidrug-resistant strains producing extended-spectrum β-lactamases and / or carbapenemases, K. pneumoniae has emerged as a major clinical and public health threat. A parallel phenomenon of severe community-acquired infections caused by 'hypervirulent' K. pneumoniae has also emerged owing to phenomenon of strains acquired virulence factors (1). Carbapenem-resistant Klebsiella pneumoniae (CRKP) has become a particularly important problem worldwide due to the rapid increase in drug resistance and subsequent high mortality (3) (4). In China, the prevalence of CRKP infection has rapidly increased from 0.7% in 2004 to 13.4% in 2014 (5), and the mortality rate of CRKP infection was approximately 40-50% since poor prognosis.
The resistance mechanism of Enterobacteriaceae to carbapenem antibiotics is very complex. However, phenotypic drug resistance of carbapenems is usually caused by two major mechanisms: (i) β-lactamase activity-binding structural mutation and (ii) carbapenemases production (6). The former mechanism is combination of extended-spectrum β-lactamases (ESBLs) and AmpC enzyme with porin loss or reduced expression, resulting the decrease of outer membrane permeability and the change of penicillin binding protein in carbapenems (7,8). As for later mechanism, carbapenemases are classi ed by molecular structure and belong to Ambler classi cation system A, B, D three β-lactamases, class A and D carbapenemases require serine hydrolysis of β-lactam, while class B metallo-β-lactamases (MBLs) require zinc at their hydrolysis active sites (6). In addition, KPCs is the most common class A gene in Enterobacteriaceae worldwide (9). KPC2 carbapenemases are widely disseminated, because its encoding gene bla KPC−2 is usually located on the plasmid of bacteria, and the plasmid where it is located can often be widely transferred in different bacteria, especially in Enterobacteriaceae. Most KPC−2-producing strains belonged to clone group (CG)258, ST258 and ST11 were the dominant ST types. ST258 has spread worldwide since its emergence in the early 21st century, particularly in North America, Latin America and several European countries (10). However, in Asia, the dominant CRKP is ST11, which accounts for 60 % in China (11). NDM-1 carbapenemases in class B have been reported in China for the rst time in 2013, and the NDM-1-producing clinical strains have spread rapidly in China, even causing explosive epidemic in some regions (12,13).
Recently, not only the outbreaks of CRKP have become increasingly common in China (14), but also the prevalence of hvKp has created new challenges for the clinician. Genetic factors that confer high virulence to hvKp existing in a large virulence plasmid, perhaps an integrated binding element, therefore, we focus on predictive biomarkers located on virulence plasmid (e.g. rmpA and iucA) (15). Importantly, MDR-hvKp (multiple drug resistance hypervirulent k. pneumoniae) strains are gradually reported worldwide due to horizontal transfer of mobile genetic elements (16). Therefore, the objectives of this study were the following: (i) To describe the prevalence of clinical CRKP isolates collected for approximately 4 years. (ii) To identify the antimicrobial resistance pro les, resistance genes, integrons, drug resistance gene cassettes, and virulence genes among these CRKP strains. (iii) To report plasmid analysis (iv) To detect genetic environment of KPC-2 and NDM-1. (v) To examine the clonal relatedness of these CRKP strains. (vi) To evaluate the risk factors CRKP infections in hospitalized patients.

Materials And Methods
Study Design and Data Collect.
During the 4 years study period (September 2016 to August 2019), this retrospective study was performed in the A liated Hospital of Southwest Medical University (Luzhou, China), where is a 3200-bed large teaching hospital with 43 wards and approximately 120,000 annual admissions. 565 clinical K. pneumoniae isolates were collected from this hospital (WHONET.2020). We selected 66 CRKP isolates which were from the patients for the rst time and excluded duplicate isolates from the same patient. All CRKP isolates were con rmed by matrix-assisted laser desorption ionization time-of-ight mass spectrometry (MALDI-TOF MS, Bruker Daltonics, Bremen, Germany).
We used a stepwise matching technique to identify appropriate control cases from patients infected with carbapenem-susceptible Klebsiella pneumoniae (CSKP) infection, in order to study the risk factors of CRKP infections. For each patient with CRKP infection, we selected a matched control patient from the pool of patients with CSKP infections. Patient matched to a case for the site of infection, gender, age ± 2 years and year of hospital admission.
Screening of CRKP and phenotypic detection of carbapenemase.
The CIM-test utilizes antibiotic susceptibility-testing disks as previously described (17) (19), 2-ml aliquots of Trypticase soy broth (TSB) (Haibo, Qingdao, China) were directly inoculated with a 1-µl loopful of K. pneumoniae colonies, a 2-mL tube of TSB supplemented with EDTA at a nal concentration of 5mM. A 10-µg meropenem disk (OXOID, Thermo Fisher Scienti c, Massachusetts, USA) was placed in each inoculated tube, which were incubated for 4 h (± 15 min), then meropenem disks placed on a Mueller-Hinton agar plate after lawn inoculation of 0.5 McFarland bacterial suspension of meropenem-susceptible Escherichia coli ATCC 25922, no carbapenemase activity will imply there will be a zone, whereas enzymatic inactivation will produce no zone. The mCIM is considered positive if the zone size is 6 to 15 mm or pinpoint colonies are present within a 16-to 18mm zone (20). Only when the isolate is positive for mCIM, eCIM results are recorded. When the zone size increases by ≥ 5 mm compared to the zone size observed by mCIM, the test isolation of MBL production is positive, and if the area size increases by ≤ 4 mm, the test isolation of MBL is negative.
The Detection of Resistance Gene, Virulence Gene and Integrase-associated Gene.
Conjugation and plasmid analysis.
To evaluate whether the carbapenemase genes are located on the plasmid and assess whether these genes can be horizontal transferred. The receptor strain was sodium azide-resistant E. coli strain J53. Implant donor and recipient strains in Mueller-Hinton plate and incubated overnight at 37°C. Then, an appropriate amount of strains was inoculated in a glass tube containing 5 mL LB medium and cultured at 37°C for 16-18h.
Subsequently, 400 µL donor strain and 200 µL recipient strain were added to the glass tube containing 800 µL LB broth medium and cultured at 37°C for 16-18 h. Transconjugants were selected using LB plates containing sodium azide (180 µg/mL) and imipenem (0.5 µg/mL), meanwhile, the LB plate was used as blank control (21). Con rmation that conjugation had taken place in E. coli J53 was carried out by MALDI-TOF MS system for the presence of resistance gene by PCR analysis. In addition, plasmid analysis was performed for strains with successful conjugation which was determined by PCR-based replicon typing as described previously (22).
To investigate the genetic environments of the NDM-1 and KPC-2, we use overlapping PCR and sequencing were applied to analyze. (23) (24). The PCR amplicon was sequenced, the DNA sequences obtained was spliced and compared to those in the NCBI gene bank database (25) (Table S3, S4).

Statistical analysis.
All analyses and graphs were performed using SPSS v.24.0 software (SPSS Inc., Chicago, USA) The chi-square test or Fisher's exact test was used to analyze categorical variables. Continuous variables were presented as means ± standard deviation (SD), and were evaluated by a more appropriate student's t-tests or Mann-Whitney U-test. Multivariable logistic regression analysis was performed to identify independent risk factors of CRKP infection. All biologically plausible variables with a value of P < 0.1 within univariate analysis were included in the following multiple logistic regression model. P < 0.05 was considered as statistically signi cant, and all probability values were two-tailed distribution.

Results
Bacterial Isolates.
The results of EDTA-synergy test, mCIM / eCIM test and drug susceptibility test were consistent. All 66 CRKPs showed positive mCIM test, namely carbapenemase. Among them, 24 strains showed eCIM positive, that is, Metallo Beta Lactamases (MBL) positive.
Plasmid and Integrase-associated Analysis.
Likewise, kpn473 and kpn475 containing bla NDM -containing plasmids were positive for transconjugants and these plasmids belonged to lncX3.
Characterization of Genetic Environment of KPC-2 and NDM-1.
For KPC-2, the genetic structure compared with pKP048 (GenBank Accession No.FJ628167). 42 KPC-2-producing strains could be divided into three different types (A-C) based on the analysis of genetic structures, among which type C was the most principal (n = 32), followed by type A (n = 8), type B (n = 3). Type A harbored same structure as pKP048. In type B, Tn1721-tnpR and Tn1721-tnpA deletions downstream of ISKpn6-like. Compared with type B, type C lacks ISkpn6-ike, and its gene structure is tnpA-tnpR-ISkpn8-bla KPC-2 ( Fig.2A).
For NDM-1, compared with pNDM-BJ01 (GenBank accession no. JQ001791), genetic structures surrounding the bla NDM-1 gene demonstrated there are three types A, B and C. As for type A(n=12), a remnant of insertion sequence ISAba125 was present upstream of the bla NDM-1 gene, and loss of downstream genes of tat. Compared with type A, the retention sequence ISAba125 was lacked in type B. In type C, the upstream of bla NDM-1 was IS30, and the downstream of bla NDM-1 was ble MBL -trpF-tat-cutA-groES-groEL. (Fig.2B) Detection of ST types.

Risk Factors and Multivariate Analysis of CRKP Infection.
As shown in Table 4, we present the results of the univariable analysis of matched data regarding risk factors associated with CRKP infections. Compared with the CRKP group, statistically signi cant differences were observed for history of respiratory disease (P = 0.034), renal disease (P = 0.039) and hematological disease (P=0.02), need for tracheal cannula (P<0.001), prior use of β-lactams and β-lactamase inhibitor combination (P=0.009

Discussion
In light of carbapenem-resistant Klebsiella pneumoniae being widely distributed in China, the increasing prevalence and global spread of these clinically resistant strains posed a serious threat to public health. At the same time, due to the acquisition of virulence plasmids, 'hypervirulent' Klebsiella pneumoniae was also emerged (1,27). In this work, we have found some particularly noteworthy ndings, rst, our results revealed that the main mechanism of CRKP collected in this study could be attributed to carbapenemase producer. And, CRKP strains were ESBL and AmpC enzymes producer. bla KPC-2 and bla NDM-1 are two major carbapenemase genes in our study. Second, the result indicated almost CRKP strains carry virulence-associated genes (81.8%, entB+ mH+mrkD), otherwise, the genes peg-344, iroB, iucA, plasmid-borne rmpA gene ( p rmpA), and p rmpA2 gene are regarded as regulator of hypervirulent phenotype. At present, in the study of CRKP it is found that it can be produced by hvKp through obtaining the plasmid encoding carbapenemases (15). Notably, we found two CR-hvKp strains, which contains bla KPC-2 gene and other resistant genes. This conclusion proved that CR-hvKp began to appear and gradually became popular worldwide, with rapid spread (28). Third. The bla NDM-1 gene has been rapidly prevalent worldwide and it often co-exists with other resistant genes in bacteria, leading to multidrug resistance, which further challenges the selection of clinical antibiotics (29). In this study, bla NDM-1 and bla IMP-4 genes co-exist in two strains, especially, we found a strain carry bla NDM-5 and bla IMP-4 genes. The rst reported drug-resistant bacteria carrying bla IMP-4 and bla NDM-1 simultaneously were K. pneumoniae isolated from 2012 (30). In recent years, strains harboring bla NDM gene and another type of carbapenemase genes have emerged worldwide (31) (32,33). The location of these carbapenemase genes is variable, which may be located on chromosome genes and plasmids. Simultaneously, they harbor more resistant genes, making the possibility of horizontal transmission of resistant genes between strains greater, and even can produce higher resistance, which poses great challenges to clinical treatment (34).
By studying the molecular detection and genetic characteristics of CRKP. In China, K. pneumoniae ST11 were always associated with bla KPC-2 (35).
In the forty-two KPC-2-producing CRKP strains, there were 7 ST types, of which 35 belonged to ST11. The rest were ST318, ST467, ST147, ST405, ST23, ST17. In recent years, more and more studies have pointed out that mobile genetic elements may promote the transmission of bla KPC-2 gene.
Almost all bla KPC-2 genes are located on plasmids, and resistance genes spread rapidly through plasmid horizontal transfer (36). Previous studies have reported KPC-2 located on diversity of plasmids IncN, IncL / M and IncR except IncFII. In our study, KPC-2 often binds to IncF plasmid, consistent with the results of this study (37). bla KPC-2 is often located on Tn4401 in European and American countries. The tnpA sequences of Tn4401 and Tn3 were about 22 % identical and 39 % similar, while the tnpR nucleotide sequences were completely different. In China, the immediate environment surrounding bla KPC-2 in pKP048 was considered to result from the integration of a Tn3-based transposon and a partial Tn4401 structure, with the ORFs ordered as follows: Tn3-tnpA, Tn3-tnpR, ISKpn8, the bla KPC-2 gene, the ISKpn6-like element, Tn1721-tnpR, Tn1721-tnpA (38). Our results demonstrated that the genetic environment of KPC-2 was diversity, which can be divided into three types. Twenty-four NDMproducing strains were divided into 7 ST types. Interestingly, so it is considered to be the outbreak that 18 of the 24 NDM-producing strains were from the neonatal department. In recent years, the outbreak of NDM-producing CRKP has been mostly in the neonatal department, but ST types were different. First reported outbreak of NDM-1 K. pneumoniae ST76 and ST37 in Shanghai in 2016 (39). Subsequently, NDM-1-producing K. pneumoniae broke out in Yunnan, China, and all of them belonged to ST105, meanwhile, its bla NDM-1 gene was located on IncFI plasmid. NDM-1 producing strains present in the hospital environment also posed a potential risk and the incubator water acted as a diffusion reservoir of NDM-1producing bacteria (28). The outbreak of ST1419, ST234 and ST1412 followed (40,41). In 2019, the outbreak of NDM-5-Producing K. pneumoniae in China began, and all strains belonged to ST337, carrying ESBLs resistant genes, all bla NDM-5 genes were located on IncX3 plasmid (42). In our study, the 14 NDM-1-producing strains, 13 belonged to ST4495 (4-1-99-1-9-5-5), ST4495 is the newly discovered ST sequence type in this study. Among the 4 NDM-5-producing strains, Kpn494, Kpn499 and Kpn502 belonged to ST2407 (2-1-99-1-9-5-5), while Kpn504 belonged to ST11. It is worth noting that the main molecular sequence type of NDM-1 is ST4495(4-1-99-1-9-5-5) and the main molecular sequence type of NDM-5 is ST2407 (2-1-99-1-9-5-5), the difference between them was only gapA gene. The reasons for this phenomenon are speculated to following possibilities:(i) NDM-producing CRKP may undergo gene evolution and transformation in a short time, including NDM-1 evolution into its variant NDM-5, while housekeeping gene gapA changes. (ii) The NDM-5 is not evolved from NDM-1, but a new carbapenemase gene carried by the patient. It is necessary to further study its epidemiology to clarify the causes and mechanisms. Among CRKP in neonatal department, bla NDM gene is more common than bla KPC , and its drug resistance is also signi cantly stronger than the latter. Due to the particularity of newborns, it is more restricted in the use of antibiotics compared with adults (43). The outbreak of NDM-producing CRKP poses greater challenges and threats treatment.
Therefore, the emergence of NDM-producing CRKP in neonatal department should be paid enough attention. As one of the most signi cant metalloβ-lactamases, bla NDM-1 has been rapidly spread worldwide recently. It was found that the rapid horizontal transmission of bla NDM-1 was closely related to mobile genetic elements. bla NDM-1 gene mostly located in transposon Tn125, the structure as following: ISAba125-bla NDM-1 -ble MBL -trpF-dsbC-cutA-groES-groEL-ISCR27-ISAba125. In this study, we detected the surrounding environment of bla NDM-1 gene and found that the surrounding environment of bla NDM-1 gene was mainly composed of A, B and C three types. Our results demonstrated that the genetic environment of these types is partly similar to transposon Tn125.
Integrons are genetic elements that allow e cient capture and expression of exogenous genes. It plays a critical role in the dissemination of antibiotic resistance, particularly among gram-negative bacterial pathogens. In this study, almost all isolates carried IntI1 gene, indicating that integron has become an important element of drug resistance in K. pneumoniae, and to some extent, it also makes the mechanism of drug resistance in K. pneumoniae more changeable and complex. There were six genes encoding drug resistance cassette structures in 46 strains with positive variable region, including dfrA12, addA2, dfrA12-addA2, dfrA27-arr3, gcuF-dfrA12 and OXA-10-addA1-aacA4.addA was the most common resistance gene, followed by dfrA. The high detection rate of these two drug resistance genes may be related to the widespread use of the two antibiotics in clinic and environment, thus causing the prevalence of related drug resistance gene cassettes. In this study, the drug resistance gene cassette bla OXA-10 -addA1-aacA4 was identi ed from a CRKP strain. OXA-10, as a class D carbapenemase, has weaker hydrolysis ability to carbapenem antibiotics. In recent years, many OXA-10-like enzymes have been found to be involved in bacterial resistance (44).
The main nding of our analysis is that hematological disease, tracheal cannula and exposure to β-lactams and β-lactamase inhibitor combination were independent risk factors associated with the development of CRKP infections. Most preview studies have demonstrated that the history of tracheal cannula was independent risk factor for CRKP infection (45-48). Meanwhile, exposure to antibiotic was also associated with CRKP infection, such as exposure to quinolones (49), β-lactams and β-lactamase inhibitor combination (50) and carbapenems (51), which was partly consistent with our study. Inappropriate treatment had the potential to increase the selection and transmission of CRKP. Notably, in the present study, hematological disease was rarely found to be a risk factor for CRKP infection. Possible explanation may be that part of the isolates in this study were from neonatology department, newborns could have poor functional status and they were susceptible to be infected. There are some limitations in our study. First, we should acknowledge that the number of patients included in this study is relatively small, although this is a common problem in studies assessing risk factors for multidrug-resistant microbial infections (49). Secondly, the study was performed in a singlecenter setting, so that some important risk factors may be missed. The data used and/or analyzed in this study are available from the corresponding author on reasonable request.

Competing interests
The authors declare that they have no con ict of interest.
Author contributions ZL, ZD and JY designed the work; YL collected data; ZL, ZD and XJ performed experiment; JX participated in the data analysis; ZL, ZD and TL drafted the initial manuscript; YD and ZZ participated in its design and coordination. All authors read and approved the nal manuscript.