Analysis of Antibiotic Resistance and Virulence Traits (Genetic and Phenotypic) in Klebsiella pneumoniae Clinical Isolates from Pakistan: Identification of Significant Levels of Carbapenem and Colistin Resistance

Introduction

Klebsiella pneumoniae (KP) is often associated with hospital and community acquired infections including pneumonia, septicemia, and urinary tract infections (UTIs). Emergence of multidrug-resistant (MDR) KP reduces treatment options and is associated with higher mortality rates (up to 40–50%), particularly in organ transplant patients.¹ In the last decade, increased resistance against third-generation cephalosporins resulted in the greater use of carbapenems which subsequently contributed to the emergence of carbapenem-resistant Enterobacteriaceae (CRE) strains.² Resistance against carbapenems is mainly conferred by plasmids and transposons encoding carbapenemases although other factors such as AmpC β-lactamase over-production, porin defects, and efflux pumps also contribute significantly to resistance.³ In Europe, the high frequency of sporadic CRE infections has resulted in CRE emergence being considered an epidemic.⁴ Carbapenem-resistant K. pneumoniae (CRKP) is a prominent member of CRE strain set and it has become a substantial threat in both hospital and community health settings. Its combined resistance to almost all frontline antibiotics, including penicillins, β-lactams, carbapenems, fluoroquinolones, and aminoglycosides poses an unprecedented challenge. CRKP strains carry class A K. pneumoniae carbapenemase (KPC), and class B New-Delhi (NDM-1) and Verona integron-encoded metallo-β-lactamases (VIM).⁵ MDR-CRKP strains have been implicated in transfer of resistance to other closely related pathogens such as Enterobacter spp. and Escherichia coli. Because of severe and life-threatening infections in healthy hosts, the recent emergence of hyper-virulent (hypermucoviscous) K. pneumoniae (hvKP) in the Asian Pacific region has attracted much attention. Such strains cause critical pneumonia, endophthalmitis, liver abscesses, and meningitis, and they have the unusual and concerning characteristic of metastatic spread to different organs in otherwise healthy adults.⁶

The first reported emergence of CRKP from Pakistan was in 2010.⁷ Subsequently, the presence of bla_NDM-1 (conferring carbapenem resistance) was confirmed in clinical strains.⁸ Recent developments suggest that CRKP strains display increased resistance to colistin and fosfomycin (some of the few antibiotics still effective against NDM-1 producing strains).^9,10 The current study determines the MDR profile, antibiotic resistance genes, virulence profile, and hypermucoviscous genotypes of clinical K. pneumoniae strains in Pakistan.

Materials and Methods

Sample Collection and Antibiotic Susceptibility Profiling

Two hundred clinical isolates of K. pneumoniae were collected from two tertiary-care hospitals located in the twin cities of Rawalpindi and Islamabad. The study protocol was approved by the Institutional Review Board of the Armed Forces Institute of Pathology and informed consent of patients was acquired. Ethical approval was granted by Bio-Ethical Committee (BEC) of Quaid-i-Azam University Islamabad. Experimental work was conducted at Quaid-i-Azam University (Islamabad) in compliance with the recommendations of the Institutional Ethical Committee. The isolates were identified using Gram-staining, colony morphology, and biochemical testing using the API 10S kit (Biomerieux). Antibiotic susceptibility to ampicillin (30 μg), ceftriaxone (30 μg), aztreonam (30 μg), gentamicin (30 μg), ciprofloxacin (5 μg), levofloxacin (5 μg), fosfomycin (50 μg), polymyxin B (300 U), colistin (10 μg), imipenem (10 μg), and meropenem (10 μg) was determined by disc diffusion assays. Data were interpreted according to the guidelines provided by the Clinical and Laboratory Standards Institute (CLSI, 2017). E. coli ATCC 25922 was used as a control strain throughout the experimental procedures.

Phenotypic Detection of ESBLs and Carbapenemases

The ESBL phenotypes were confirmed via double-disc synergy testing, using discs of amoxicillin-clavulanic acid (30 μg), cefotaxime (30 μg), ceftazidime (30 μg), and aztreonam (30 μg). ESBL production was inferred by any distortion or ≥5 mm augmentation of an inhibition zone of a cephalosporin disc towards the amoxicillin-clavulanate disc.¹¹ Carbapenemase production was assessed by the Modified Hodge test (MHT). E. coli ATCC 25922 and K. pneumoniae ATCC 700603 were used as control strains. The formation of cloverleaf indentations on a lawn of E. coli ATCC 25922 characterized the tested isolates as carbapenemase producers (CLSI M100, 2017).

Determination of MICs of K. pneumoniae Isolates

The broth micro-dilution method was used to quantify antibacterial resistance against gentamicin, levofloxacin, and colistin. Briefly, adjusted cultures (to give a final CFU/mL of ~5 x 10⁵) of the test organisms were inoculated into Muller-Hinton broth (Oxoid, UK) in microtiter plates containing two-fold serial dilutions of an initial antibiotic stock solution giving final concentrations ranging from 0.125 to 128 μg/mL. The bacterial cultures were incubated for 18 h at 37°C before growth was assessed; the lowest concentrations of antibiotic where inhibition occurred were designated as the MIC, as per CLSI guidelines. The reference breakpoints for the interpretation of MICs against gentamicin, levofloxacin, and colistin were set as ≥16, and ≥2 μg/mL, respectively.¹²

DNA Extraction and Amplification of Genes

DNA was extracted by the phenol-chloroform method.¹³ PCR amplifications were performed for the genes encoding carbapenemases (bla_NDM, bla_OXA-48, and bla_KPC), ESBLs (bla_TEM, bla_SHV, bla_CTXM-14, bla_CTXM-15, bla_OXA), and efflux pump proteins (acrAB, tolC). Plasmid-encoded genes specifying colistin resistance (mcr-1, mcr-2), the virulence gene for hypermucoviscosity (rmpA), iron acquisition genes (iucA, kfu), biofilm regulator (bss), adhesion genes (fimH, mrkD, ycfM), the serum resistance factor gene (traT), and porins genes (ompK35, ompK36) were amplified as described elsewhere.¹⁴ Plasmid DNA was extracted using a Thermo Scientific Gene JET plasmid miniprep kit.

Detection of Virulence Phenotypes in K. pneumoniae Isolates

Hypermucoviscous phenotypes were tested by using the string test through the formation of a ≥5 mm mucoviscous string from a single bacterial colony. Nigrosine staining was performed for visualization of K. pneumoniae capsule under light microscopy. The isolates were tested for haemagglutination, and for gelatinase and lipase production. The biofilm potential of the isolates was quantified using a microtiter plate assay by comparing the “cut-off” optical densities (ODc) with negative controls at 492 nm.¹¹

Results

Sample Distributions and Antibiotic Resistance

Two hundred clinical K. pneumoniae isolates were collected from nine different sample sites (Table 1; see Table S1 for full data set). The overall resistance level for carbapenems (imipenem and meropenem) was 38%. The greatest frequency of resistant was to ampicillin (94%), followed by ceftriaxone (71%). The frequencies of resistance against gentamicin (55%), ciprofloxacin (53%), colistin (15%), aztreonam (59%), and fosfomycin (22%) were above the threshold levels at which these antibiotics should no longer be used in empirical treatment of K. pneumoniae infections¹⁵ (Table 2). Resistance against polymyxins B and E was 16 and 15%, respectively. The prevalence of MDR within the 200 isolates was 63% (125), of which 58% (72) were carbapenem-resistant (MDR-CRKP) (Table 1). Of the 72 MDR-CRKP strains, 28% (20) were ESBL producers. The prevalence of MDR-KP within K. pneumoniae isolates was similar in female and male patients, although there was a more than twofold higher prevalence of the combined MDR with HMV traits in isolates from males than from females. Also, a high proportion of the MDR (38%) and combined MDR/HMV (60%) strains were isolated from urine whereas much lower proportions were found in tissue, bile, swab, and catheter tip samples (Table 1).

Table 1 Genetic and Phenotypic Traits of MDR Klebsiella pneumoniae Isolates

Table 2 Genotypic and Phenotypic Comparison Between Hypermucoviscous and Non-Hypermucoviscous K. pneumoniae Isolates

MIC Values for Gentamicin, Levofloxacin and Colistin

MICs for gentamicin, levofloxacin, and colistin were determined for the 63 carbapenemase producers and 137 non-producers. For gentamicin, 38% of the carbapenemase producers exhibited MICs >256 µg/mL whereas only 9% of the non-producers showed resistance to such gentamicin concentrations. For levofloxacin, the MIC observed for the carbapenemase producers was twofold higher (128 µg/mL) than the non-producers (64 µg/mL). For the majority of the carbapenem-sensitive isolates, colistin MIC values were between 4 and 8 µg/mL. The differences in MICs for levofloxacin, gentamicin, and colistin between carbapenem-resistant and -sensitive isolates were statistically significant (p < 0.01).

Genetic Factors Specifying Antibiotic Resistance

The presence of genes encoding ESBL factors (bla_TEM, bla_SHV, bla_CTXM, bla_OXA), carbapenemases (bla_KPC, bla_NDM-1, bla_OXA-48 type), efflux pump (acrAB, tolC) and colistin markers (mcr-1, mcr-2) was investigated. Among the carbapenemase genes, bla_NDM-1 was the most predominant (39%) and its occurrence was significantly higher in comparison to bla_KPC (5%). The frequency of occurrence for the carbapenemase-encoding bla_OXA-48 gene was 24%, and the colistin-resistance mcr-1 and mcr-2 genes were detected in 12 and 9% of isolates, respectively. Carriage rates for bla_TEM, bla_SHV, bla_CTXM-14, bla_CTXM-15, and bla_OXA were ~3 fold higher in MDR-ESBL strains than in MDR (non-ESBL) strains (Table 1).

Prevalence Virulence Factors and hvKP

For the fimbrial and non-fimbrial adhesins, the occurrence rates amongst the 200 isolates were as follows: fimH, 19%; mrkD, 18%; and ycfM, 22%. The polysaccharide regulator gene (bss) was detected in 16% of the isolates. Overall, 22% of the isolates displayed a mucoviscous phenotype (string-test positive), while a similar 24% of the MDR isolates displayed this phenotype (Table 1). The rmpA gene encoding the “regulator of mucoid phenotype” was found in 49% of the hypervirulent phenotypes, but was not found in any of the non-HMV strains (Table 2). Isolates demonstrating an HMV phenotype showed higher resistance against fosfomycin and colistin, but a lower resistance to polymyxin B (Table 2). The frequency of occurrence of the virulence determinants iucA, kfu, bss, fimH, and mrkD was higher in the hypermucoviscous MDR isolates (Table 1). The porin genes, ompK35 and ompK36, were found in 56 and 55% of isolates, respectively, and had slightly higher occurrence in the non-hypermucoviscous strains compared to the HMV strains (Table 2). Furthermore, haemagglutination (100%), capsule (99%), and biofilm formation (94%) were observed at high frequency in 200 K. pneumoniae isolates. Gelatinase production and blood hemolysis were observed in 12 and 8% of the isolates.

Discussion

Over the past decade, preferential use of carbapenems has generated significant selective pressure leading to the emergence of MDR-CRKP¹⁶ and such strains are increasingly associated with life-threatening infections.¹⁷ Likewise, emergence of hyper-virulent Klebsiella strains is causing an increasing challenge, particularly in countries lacking central surveillance systems for infectious diseases.¹⁸ However, little is known about the prevalence of hvKP strains in Pakistan. In addition, very limited data are available on colistin and fosfomycin resistant CRKP. In this study, we observed that 38% of the K. pneumoniae isolates collected from patients were CRKP, which indicates an alarming increase of CRKP frequency in Pakistan with respect to that reported in 2010.⁷ Similar increases in CRKP have been observed in India and China.¹⁹ The resistance levels recorded for aztreonam (59%), gentamicin (55%), ciprofloxacin (53%), and imipenem (38%) were higher than the recent antimicrobial surveillance study in South Korea reporting 13.5, 24.4, 24, and 3.1% resistance, respectively.²⁰ Our findings align with the resistance pattern of multidrug-resistance in K. pneumoniae isolates from Iran with 36% aztreonam, 15.9% gentamicin, 19.4% levofloxacin, and 11.8% imipenem resistance.²¹ The highest acquired resistance observed here (after the broad-spectrum cephalosporin, ceftriaxone) was for aztreonam, indicating the frequent usage of this drug against K. pneumoniae infections within South Asian hospital settings. This is concerning as aztreonam is the only mono-cyclic β-lactam antibiotic (monobactam) that is effective against IMP-, VIM-, NDM–type carbapenemase-producing K. pneumoniae.²² The resistance levels recorded for colistin (15%), polymyxin B (16%), and fosfomycin (22%) were comparatively low with respect to other antibiotics tested, possibly because of their limited administration due to their associated toxicity and/or to their stringent dose administration which is based on the principle of “highest safe” rather than “minimum effective dosage”; such prescription practices would limit the emergence and spread of resistance.²²

The 22% fosfomycin and 15% colistin resistance observed here highlight a significant increase in the resistance against these antibiotics in recent years. The resistance level recorded for colistin (15%) is approximately five times that reported in 2019 and 2020, i.e. 2% and 2.9%, respectively.^23,24 The recent surge in colistin resistance is alarming²⁵ since polymyxins, particularly colistin methanesulfonate (CMS), are routinely used to treat CRKP infection.²⁶ Colistin resistance is mainly associated with genetic mutations in housekeeping genes such as mgr and phoAB; however, the plasmid-mediated mcr-1 and mcr-2 factors pose a threat for rapid global spread of such resistance.²⁷ Plasmid-mediated resistance determinants elevate the MICs of antibiotics, resulting in higher antibiotic doses for effective treatment, which in turn leads to an increased probability that chromosomal mutations will arise that confer a further increase in resistance.²⁸ Our data match previous findings²⁹ demonstrating that horizontal transfer of mcr-1 contributes to an elevated MIC value against colistin and that some mcr positive isolates remain within the susceptible range (Table 3). For majority of the mcr-1 and mcr-2 positive isolates, the MIC for colistin ranged from 4 to 8 μg/mL and can thus be defined as colistin resistant.³⁰ This is the first report from Pakistan that indicates an association of the plasmid-borne mcr genes with elevated colistin resistance as determined by the microbroth dilution method. Although a report in 2017 indicated colistin resistance in CRE stains from Pakistan, no attempt was made to determine the presence of the mcr genes.⁹ However, here we find a clear co-association of mcr and carbapenemase loci with 92% (23/25) of mcr isolates carrying at least one carbapenemase gene (Table 3). Although mcr-1 does not play a direct role in resistance to other antibiotics, the results reported here showing a coexistence of mcr-1 with ESBL and bla_NDM determinants reflect a major development in the emergence of resistance in K. pneumoniae, where efficacy of carbapenems and colistin is threatened.³¹

Table 3 Colistin Resistance in CRKP and HMV K. pneumoniae Strains

In this study, the MICs determined for fluoroquinolone, aminoglycoside, and polymyxin were notably higher for the carbapenemase-producing K. pneumoniae. CRKP frequently harbors a combination of resistance mechanisms including carbapenemases, drug efflux, and loss of porin function^32,33 leading to higher levels of resistance to antibiotics, as observed here in the carbapenemase-producing isolates. Molecular screening for the resistance markers in 200 K. pneumoniae isolates in this study confirmed bla_NDM-1 as the predominant (39% occurrence) carbapenemase-encoding gene followed by bla_OXA-48 (24%) and bla_KPC,(5%). Very high occurrence of bla_NDM-1 and bla_OXA-48 carbapenemases suggests that they have a predominant role in the recent emergence of CRKP phenotypes in Pakistan.^34,35

Hypervirulent K. pneumoniae strains are endemic in the Asian Pacific region,³⁶ but data are extremely scarce on prevalence of such strains in Pakistan.³⁷ In this study, 13% of the MDR isolates carried rmpA which is the first report for the presence of the rmpA gene in hvKP isolates, with resistance to frontline antibiotics, from Pakistan.³⁷ Overall, 22% of the isolates exhibited a positive string test and 11% of the isolates carried the rmpA gene. Up to 100% rmpA prevalence in hypermucoviscous isolates has been reported previously.³⁸ Another study showed 41% rmpA carriage in non-hypermucoviscous strains indicating additional underlying factors required for the hvKP phenotype.³⁹ Contrary to other studies that observed lower antimicrobial resistance in hvKP isolates,⁶ here a higher resistance against cephalosporins, aminoglycosides, fluoroquinolones, and carbapenems were found in hvKP strains. Until 2015, hypervirulent and the multidrug-resistant K. pneumoniae isolates were reported to belong to distinct clones with no reports of MDR and virulence convergence.⁴⁰ However, a recent combination of these traits is evident from a report of convergent multidrug-resistant hypervirulent strains from South and Southeast Asia in 2020.⁴¹

Conclusion

This study highlights the prevalence of CRKP clinical isolates with notable virulence potential. The results support previous studies indicating that the bla_NDM-1- and bla_OXA-48-encoded carbapenemases are the most prevalent K. pneumoniae carbapenemases in Pakistan. The rise in colistin and fosfomycin resistant strains globally points to significant treatment constraints in the near future. In general, the K. pneumoniae isolates studied here showed a high prevalence of resistance gene carriage but a relatively low virulence gene carriage; interestingly, the hypermucoviscous isolates exhibited increased resistance potential, a trait not frequently associated with the hypermucoviscous K. pneumoniae isolates.