Multidrug-resistant (MDR) gram-negative bacteria present a serious threat to global public health, especially, carbapenem-resistant Klebsiella pneumoniae (CRKP). Colistin has been considered one of the last resorts in treating severe infections that retain activity against K. pneumoniae carbapenemase (KPC)-producing K. pneumoniae. However, the emergence and global dissemination of colistin resistance compromise the efficacy of colistin.

Colistin is a cationic peptide synthesized by Gram-positive species, and it exhibits antimicrobial activity through multiple pathways, including direct antibacterial activity, antiendotoxin colistin activity, vesicle–vesicle contact pathway, hydroxyl radical death pathway, and respiratory enzyme inhibition (Sun et al., 2018). With antimicrobials, bacteria develop a series of resistance systems, including chromosomal gene mutations and plasmid-associated resistance genes to adapt to the environment. Colistin resistance in K. pneumoniae is mediated by several mechanisms. Mutations in genes encoding PhoP/PhoQ and PmrA/PmrB two-component regulatory systems regulate the expression of pmrC and pmrHFIJKLM operons and are responsible for the synthesis and transfer of 4-amino-4-deoxy-L-arabinose (L-Ara4N) cationic groups to lipid A. This modification increases the positive charge on lipopolysaccharides (LPSs) and therefore, decreases colistin binding, leading to colistin resistance (Poirel et al., 2017). Alternatively, MgrB is a small transmembrane protein produced upon the activation of the PhoP/PhoQ signaling system and acts as a negative regulator of this system. It regulates the expression of etpB, which is related to the modification of LPS. Mutations in mgrB gene upregulate the expression of the PhoP/PhoQ system, thereby leading to colistin resistance. Generally, mgrB mutations are caused by insertion sequences (ISs; IS5-like, IS1F, ISKpn13, ISKpn14, and IS10R) or point mutations (Cannatelli et al., 2014). In addition, hyperproduction of capsular polysaccharide reduces the interactions of colistin with the bacterial surface, leading to colistin resistance (Poirel et al., 2017). Besides, horizontal transfer of plasmid-borne mcr gene play a significant role in the dissemination of colistin resistance among various bacteria. Since its first identification in the late 2015 (Liu et al., 2016), the determinants of transferable colistin resistance have extended further away from mcr-1 to include various novel alleles (Sun et al., 2018). Until now, nine novel mcr variants have been reported, namely, mcr-2, mcr-3, mcr-4, mcr-5, mcr-6, mcr-7, mcr-8, mcr-9, and the recently identified mcr-10 (Wang et al., 2018; Xiang et al., 2018; Yang et al., 2018; Kieffer et al., 2019; Wang et al., 2020). Moreover, the mcr family has been increasingly reported in various genera of Enterobacteriaceae worldwide, isolated from food animals, meat and vegetables, the environment, infected patients, and asymptomatic human carriers (Skov and Monnet, 2016). All MCR proteins are characterized as phosphoethanolamine (PEtN) transferases. They catalyze the attachment of PEtN to lipid A and lead to a reduction of the negative charge of LPS through structural alterations of lipid A and a decrease in the binding of colistin, thus resulting in colistin resistance (Sun et al., 2018). However, although these alleles belong to the group of PEtN transferases, the similarities in amino acid sequences vary, reflecting different genetic origins of these mcr alleles. Similarly, previous studies have demonstrated the complex dissemination of these alleles across the diversified species of Enterobacteriaceae and plasmid reservoirs and genetic environment for mcr-like genes (Wang Q. et al., 2017; El-Sayed Ahmed et al., 2020). Nevertheless, limited reports have indicated the presence of mcr-1, mcr-3, mcr-7, and mcr-8 in K. pneumoniae at a relatively low prevalence. Although several studies have elucidated colistin-resistance mechanisms from different aspects, the underlying mechanism of colistin resistance in clinical K. pneumoniae isolates in our hospital was unclear.

Therefore, in this study, we identified 53 colistin-resistant K. pneumoniae (COLR-KP) isolates between 2017 and 2019 to investigate the primary mechanisms of colistin resistance and evaluate the potential prevalence of clinical K. pneumoniae isolates in our hospital.

During the coronavirus disease 2019 pandemic, the problem of bacterial resistance cannot be ignored. The World Health Organization celebrated the “World Antimicrobial Awareness Week” in 2020, aiming to raise awareness regarding the problems of antimicrobial resistance and establish a united front to preserve antimicrobials and curb the development of antimicrobial resistance. Colistin is an ultimate line of refuge against lethal infections by MDR gram-negative pathogens. Unfortunately, this last-resort antibiotic has been challenged by the mobilized colistin resistance determinant mcr family, resulting in a rapid spread worldwide with unexpected diversity.

The overall prevalence of colistin resistance in clinical K. pneumoniae isolates was 4.5% (53/1,171), which is higher than that in national surveillance from the CHINET report (CHINET data, 2020. http://www.chinets.com/Data/AntibioticDrugFast). In addition, the rising prevalence trend of COLR-KP was observed with each passing year, with a significant increase from 2.8% (17/603) in 2018 to 6.3% (36/568) in 2019. However, of all patients without colistin exposure, none had a history of recent overseas travel. It is possible that the spread of colistin resistance can occur in hospitals in the absence of colistin use, which is consistent with previous study (Huang et al., 2018). The fact that the prevalence of COLR-KP increased in 2019 coincided well with the consumption of colistin in our hospital, suggesting that selective pressure from colistin use is a factor that intensifies resistance rather than a definite correlation to colistin use.

Among the COLR-KP isolates in our study, >50% were MDR, which further limited the current treatment options. Four of these isolates were KPC-producing K. pneumoniae ST11 clones, indicating that colistin resistance emerged among the KPC-2-producing K. pneumoniae pandemic ST11 clonal lineage isolates in our region. Recent clinical studies have reported that mortality rates among patients with infections caused by COLR-KPC K. pneumoniae isolates were higher than those among patients infected by colistin-susceptible KPC K. pneumoniae strains. This observation is attributable to further narrowing of available therapeutic options (Giacobbe et al., 2015).

Of particular concern were two COLR-KPC K. pneumoniae isolates, KP67 and KP69, which exhibited pandrug resistance (PDR) and were recovered from patients with BSI; these were resistant to all tested antimicrobials, including tigecycline. These findings suggest that colistin resistance is imparted to KPC-producing K. pneumoniae with minimal impact on its ability to either spread or cause invasive infection. This finding agrees with those of an earlier report (Arena et al., 2016).

The genomic features of isolates were consistent with the previously described data for other K. pneumoniae genomes (Wyres and Holt, 2018). MLST data showed that 13 COLR-KP isolates had diverse genetic backgrounds and five of these isolates belonged to the epidemic ST11 clone. Phylogenetic analysis also showed a genetic diversity, including those ST11 isolates. Although isolated in the same period (within a week) from the same department, the mechanisms of colistin resistance in ST11 KP67 and KP69 strains were different, indicating that the colistin resistance of these strains proceeded in a distinctly evolutionary pathway.

Of the 13 colistin-resistant KP isolates, four carried mcr genes, and among these, three carried mcr-1 and one carried mcr-8 genes. The genetic characterization of mcr-1-bearing plasmids showed that mcr-1 in KP4 and KP30 was located in the transferable approximately 33-kb IncX4 plasmid, the most prevalent version of the mcr-1-bearing plasmid reservoirs (Fernandes et al., 2016; Zurfluh et al., 2016). Comparative analysis revealed that the mcr-1-bearing plasmid pMCR_KP4 (KP4) identified in this study displayed >95% nucleotide identity to various plasmids, including pMCR_WCHEC1618 (GenBank accession no. KY463454) isolated from E. coli in Sichuan, located 1,490 km away from Nanchang, China, and pIBMCmcr1 (GenBank accession no. MF449287) isolated in Italy. Interestingly, KP30 bore a different mcr-1 plasmid, which showed a large fragment inversion. It indicated that KP30 acquires the mcr-1 gene from a different approach. The mcr-8.2-bearing fragment from KP23 (pMCR_KP23), which showed a >99.9% identity to plasmid pMCR8_095845 (GenBank accession no. CP031883), was also recovered from clinical K. pneumoniae strain in 2016 in Sichuan, China (Ma et al., 2020).

The first identification of MCR-1-producing K. pneumoniae in the study hospital was in 2018, highlighting a much later emergence than that of E. coli in 2014 as reported previously (Quan et al., 2017). Notably, the mcr-1-bearing plasmid in the colistin-resistant E. coli was IncX4, which was identical to those of KP4 and KP30, implying that the IncX4-bearing-mcr-1 was successfully disseminated and circulated among various species of Enterobacteriaceae in this hospital. Accordingly, there were questions about whether mcr-1 was an emerging resistance trait in other countries as mcr-1-positive isolates were collected as early as in 2005 in France (Haenni et al., 2016).

Of a particular concern is the transfer of a plasmid with mcr-1 into carbapenemase-producing isolates intra-hospital, which leads to nosocomial infections with PDR strains, leaving no beneficial treatment options available. In our study, we did not detect mcr-1 gene in all KPC-2-producing COLR-CRKP isolates. Nevertheless, we detected two PDR strains resistant to all antibiotics, including tigecycline, with the colistin resistance associated with mgrB inactivation. These patients had no colistin exposure history before the specimen collection; consequently, tracing the origin of selective pressure for these isolates was difficult.

Plasmid pMCR_KP4 and pMCR_KP30 successfully transferred the mcr-1 gene to E. coli J53AZ^R, which harbored an intact set of conjugative transfer genes to facilitate the transfer of plasmids among various isolates, consistent with its ability to conjugate into E. coli J53AZ^R, as described above. In contrast, the mcr-8.2 could not be transferred from the KP23 isolate to the recipient strain by conjugation despite several attempts. Moreover, no genes encoding conjugation-related proteins were found in the sequence of pMCR_KP23; therefore, it is suggested that the mcr-8.2-bearing plasmid in this strain was non-self-transmissible, which is consistent with the results of a previous report (Ma et al., 2020). Although the IS3-like family was detected in this plasmid, it could mobilize the intervening genetic components of mcr-8.2 or acquire colistin resistance.

Interestingly, KP18 was an mcr-1-ESBL-producing isolate, which also carried a novel plasmid-borne RND-type tigecycline-resistance determinant tmexCD1-toprJ1 (Lv et al., 2020). More recently, a research group first identified the tmexCD1-toprJ1 located on the conjugative IncFIA plasmid (pHNAH8I-1) in a PDR K. pneumoniae isolate obtained from a chicken in Sichuan. This strain also carried both carbapenem-resistance gene bla _NDM-1 and colistin- resistance genes mcr-1 and mcr-8 (Lv et al., 2020). Although tmexCD1-toprJ1 mainly spreads among K. pneumoniae isolates from food animals, the simultaneous appearance of mobilized colistin- and tigecycline-resistance genes in clinical isolates is an alarming evolutionary trend, which may lead to rapid horizontal transfer in hospital settings; this requires attention for a continuous and meticulous surveillance.

In the present study, the most common chromosomal mutation associated with colistin resistance was the T246A amino acid substitution in pmrB. This substitution was identified in most COLR-KP isolates and reported in colistin-resistant K. pneumoniae in a Tunisian Teaching Hospital (Jaidane et al., 2018). However, the T246A was predicted as neutral mutation that was most likely unrelated to colistin resistance, consistent with of the findings of a previous report (Longo et al., 2019). All ST11 isolates had R256G and T246A amino acid substitutions in pmrB, similar to that previously described in ST258 isolates (Aires et al., 2016), and the ST11 is a close variant of the international pandemic clone ST258. Furthermore, some isolates presented several new mutations as amino acid substitutions: S203P, T240M, and V257A in PmrB; E35A and M66I in PmrA; and V24G in PhoQ. These mutations were predicted to be neutral, with no effect on the biological functions of these proteins. The underlying mechanisms of colistin resistance in GNB are complex and not completely understood yet (El-Sayed Ahmed et al., 2020). According to our results, there still exists some unknown molecular mechanism leading to colistin resistance, which requires further investigations. However, this prediction is not definitive, and further investigation is needed to confirm the relationship between these mutations and colistin resistance.

MgrB mutation was the most frequently reported molecular mechanism of colistin resistance in K. pneumoniae (Cannatelli et al., 2014), whereas only three isolates possessed mgrB mutations in our study. The ST105 KP12 and ST11 KP69 isolates had complete deletion; the ST11 KP67 isolate had a truncated gene due to a promoter deletion. Insertional inactivation caused by diverse ISs at different locations within the mgrB gene was often responsible for acquired colistin resistance in K. pneumoniae (Nordmann et al., 2016). Some complete deletions of the mgrB locus had been reported as well. Of note, isolate KP67 was significantly more virulent and exhibited resistance to tigecycline and carbapenem, as well as to colistin, causing BSI.

Various resistance genes were carried by these COLR-KP isolates, which conferred resistance to carbapenems, cephalosporins, fluoroquinolones, aminoglycosides, macrolides, sulfonamides, tetracyclines, and trimethoprims. Four carbapenem-resistant COLR-KP isolates produced KPC-2 as the predominant carbapenemase in the world and China (Munoz-Price et al., 2013; Zong et al., 2020). In addition, critical virulence factors associated with the hypervirulent K. pneumoniae, including aerobactin (iucABCD and iutA), salmochelin (iron and iroBCD), and yersiniabactin (ybt), were found in the isolates KP4, KP11, KP19, and KP67. The combination of hypervirulence and high drug resistance also resulted in veritable “superbug,” thus making infection control intractable.

COLR-KP emerged in our hospital and was increasing with each passing year. Although the molecular determinants of colistin resistance in K. pneumoniae clinical isolates in our region are still chromosomal mutations, mcr-1 and mcr-8 have been detected in MDR K. pneumoniae isolates without colistin use. Moreover, the coexistence of plasmid-borne mcr-1 and tmexCD1-toprJ1 in K. pneumoniae further threatens public health. Of a particular concern is the simultaneous emergence of hypervirulence and colistin–tigecycline–carbapenem resistance in an epidemic clone ST11 K. pneumoniae isolate, causing a lethal and untreatable infection.

The data presented in the study are deposited in the NCBI repository, accession number PRJNA718703 can be found below:https://www.ncbi.nlm.nih.gov/bioproject/?term=PRJNA718703.

The studies involving human participants were reviewed and approved by Ethics Committee of the First Affiliated Hospital of Nanchang University. The patients/participants provided their written informed consent to participate in this study.

XH, JG, and LZ designed the study. YLL, YQL, and XL wrote the manuscript. LZ reviewed the manuscript. NH and QL managed the strains and data collection. ZW, WZ, XL, and YQL prepared the figures and tables. All authors contributed to the article and approved the submitted version.

This work was supported by the National Natural Science Foundation of China (32060040 and 31760261), the Natural Science Foundation of Jiangxi Province (20202BABL216084, 20192ACBL21042, 20202BAB216045, 20192BBG, and 20204BCJL23054), the Science and Technology Research Project of Education Department of Jiangxi Province (GJJ180130), and the Major Science and Technology Project of Jiangxi Province (20181BBG70030).

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

The reviewer QL declared a past co-authorship with several of the authors, LZ and YZ, to the handling editor.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.