Prevalence of pks genotoxin among hospital-acquired Klebsiella pneumoniae.

The pks genotoxic K. pneumoniae has recently triggered a widespread alarm. DNA damage and higher virulence have been linked to colibactin, a genotoxin expressed by the pks genomic island. Little is known about its molecular epidemiology in clinical isolates from Egypt. Therefore, this study was conducted to determine the prevalence and the microbiological and clinical features of pks harboring hospital-acquired K. pneumoniae isolates from Egypt. Eighty-seven hospital-acquired K. pneumoniae isolates from various specimen types were screened for pks colibactin island markers clbB, clbQ, clbA, and clbN by PCR. The pks-positive hvKp isolates were classified to one of the capsular types K1 and K2 using multiplex-PCR targeting K-serotype wzi and rmpA genes. The prevalence of pks+ strains was 27.6% (24/87). K1 capsular type, phenotypic, and genotypic hypervirulent isolates were significantly higher among pks+ strains than pks- strains (P < 0.001), while pks+ K. pneumoniae strains were found to be significantly less resistant to 8 of the antibiotic compounds tested than pks- strains. Carriage of K1 capsular type and mucoviscosity-associated rmp A gene and diabetes mellitus were identified to remain independent risk factors having a substantial association to pks-positivity by multivariate regression analysis. In conclusion, Hospital-acquired K. pneumoniae isolates in Egypt had an increased prevalence of the pks colibactin genotoxin. The significant occurrence of hypervirulent determinants in pks+ K. pneumoniae highlighted the genotoxin's possible pathogenicity combined with its distribution in several specimen types, which necessitates clinical attention and epidemic tracking.

Introduction

Klebsiella pneumoniae (K. pneumoniae) is a potential nosocomial superbug that is generating clinical concern [1]. Hypervirulent K. pneumoniae (hvKp), which includes K1 and K2 strains, frequently shows hypermucoviscous phenotypes and carries a variety of hypervirulence genes [2],[3]. Worryingly, hvKp has spread worldwide, causing serious metastatic infection, particularly in immunocompetent persons, besides the emergence of multidrug-resistant (MDR) hypervirulent strains, all creating a major challenge in the clinical field [4].

The pks genomic island, detected in the Escherichia coli meningitis strain IHE3034 in 2006, transcodes for a hybrid of nonribosomal peptide synthetase-polyketide synthase enzymes of colibactin, a genotoxin that has been linked to DNA double-strand breaks, chromosome abnormalities, cell cycle arrest, and immune cell death. The pks island contains 19 colibactin synthesis genes (clbA-clbR); only the presence of all pks genes is required for complete colibactin production [5]. Bacteremia, meningitis, and colorectal cancer have all been correlated to the carriage of the pks gene in E. coli [6],[7]. Other Enterobacteriaceae members, such as Citrobacter koseri, Klebsiella pneumoniae, and Enterobacter aerogenes, have also been reported to contain the colibactin island [8]. Lately, the pks genes were also detected on sequencing of emerging hvKp and were linked to increased virulence [9].

While colibactin is considered a virulent factor in K. pneumoniae, little is documented about its molecular epidemiology in clinical isolates from Egypt. Therefore, this research aimed to determine the prevalence of pks island and investigate the microbiological and clinical features of pks harboring hospital-acquired K. pneumoniae isolates from Egypt.

Materials and methods

Bacterial isolates

Eighty-seven isolates of K. pneumoniae recovered by the Medical Microbiology and Infection Control Unit (MDICU), Mansoura University, Egypt, between May 2020 and August 2021 were attained for this observational cross-sectional study from clinical samples from 87 patients. Among them, 81.6% (71/87) were male, and the median age was 58 years ranging from (43–65 years). Our study included 58 isolates collected from urine and blood (29 each), 20 isolates from respiratory samples, and nine from other clinical samples. These isolates were identified, handled, and preserved using standard microbiological laboratory procedures [10]. A positive “string test” was used to define the hypermucoviscous traits. The string test was affirmative when a bacteriology inoculation loop can stretch bacterial colonies on an agar plate to form a viscous string > 5 mm in length [11].

Screening for pks genomic island, capsular types, and hvKp genes

The boiling method was employed to recover the genomic DNA template from a fresh culture of K. pneumoniae for PCR amplification [12]. PCR primers set for the four typical genes of the pks genomic island were used to develop two flanking (clbB and clbQ) and two internal (clbA and clbN) amplicons to document the existence of a full pks island in the isolates since they are clustered at the 3′ and 5′ ends of pks [13].

Based on multiplex-PCR amplification of K-serotype wzi and rmpA genes, the pks-positive hvKp isolates were classified to one of the capsular types K1 and K2 as documented previously [14]. Ethidium bromide staining and UV transillumination were used to visualize the PCR products on 1.5% agarose gel. The primers applied in the study and PCR programs are listed in table 1.

Antimicrobial susceptibility pattern

K. pneumoniae isolates were tested for their antimicrobial sensitivity using disk diffusion testing, and results were interpreted as per the guidelines of the Clinical Laboratory Standards Institute (CLSI) for the breakpoints [15]. MDR-resistant isolates were those that were resistant to at least one antimicrobial agent in three or more of the classes [15]. K. pneumoniae ATCC 700603 was employed as quality control.

Data statistical analysis

SPSS software was used to analyze all data (version 25.0). The chi-square or Fisher's exact test were used to assess the significance of categorical variables. For continuous variables, the Mann-Whitney U test was used. Predictors of pks+ K. pneumoniae infection were analyzed using logistic regression. A multivariable logistic regression analysis was used to determine the association between predictor variables. Statistical significance was defined as a p-value < 0.05.

Ethics approval of research

The Institutional Review Board (IRB), Mansoura Faculty of Medicine, Egypt, code number: R/21.12.1539 approved the revision of patients' medical records and collection of bacterial isolates with the application of informed consent.

Results

Screening for pks genomic islands among our 87 K. pneumoniae isolates results in detecting colibactin structure markers represented in clbA, clbB, clbN, and clbQ amplified flanking and internal genes in 27.6% (24/87) isolates, which were classified as pks strains. Eleven of them were isolated from blood, eight from sputum, four from urine, and one from pus sample.

A total of 66.7% (58/87) of isolates were serotyped as K1 and K2. Capsular type K1 was prevalent in 45.9% (40/87) of isolates, while 20.6% (18/87) were classified as K2 capsular type. Twenty-nine isolates felt into other K1/K2 serotypes. According to data analysis, pks strains had a considerably significantly high frequency of K1 capsular type than pks strains (p < 0.001).

Capsular serotypes pattern of pks-positive and pks-negative K. pneumoniae isolates.

Hypermucoviscous isolates were significantly higher among pks compared to pks isolates (P < 0.001) as detected phenotypically by positive string test in 54.2% (13/24) pks isolates and genotypically by the presence of rmpA gene in 58.3% (14/24) pks isolates. Data are shown in table 2.

Regarding antimicrobial resistance, the pks isolates were less resistant to all antimicrobial agents tested versus pks isolates. There was a statistically significant difference in susceptibility of pks isolates to piperacillin-tazobactam, cefoperazone-sulbactam, aztreonam, meropenem, imipenem, levofloxacin, ciprofloxacin, and trimethoprim-sulfamethoxazole. Although there was a tendency toward more MDR pks isolates, the difference was insignificant. Figure 2 provides an overview of the antimicrobial results.

Figure 2.

Antimicrobial resistance pattern of pks-positive and pks-negative K. pneumoniae isolates.

*Note: Bold values denote statistical significance at the p < 0.05 level.

Concerning demographic and clinical data, patients above the age of 55 account for 70.1% of all infections, which may be due to increased susceptibility of old age to hospital-acquired infections. pks infected patients were significantly younger than pks isolates infected patients (P = 0.028). There was no significant gender difference between the two groups. Diabetic patients were more vulnerable to pks isolates than pks isolates (P = 0.002) table 3.

Carriage of K1 capsular type [Odds ratio (OR) 7.6, 95% confidence interval (95% CI) 2.49–23.194] and mucoviscosity-associated rmp A gene [OR 6.618, 95% CI 2.339–18.725], and diabetes mellitus [OR 4.857, 95% CI 1.744–13.528] were identified to remain independent risk factors having a substantial association to pks-positivity by multivariate regression analysis.

Discussion

Colibactin genotoxin, encoded by the pks island enzymes, has been linked to host DNA damage and higher virulence among E. coli and K. pneumoniae, resulting in severe infection sequelae [9],[17]. However, limited research data about the prevalence and microbiological variables related to pks-positivity in K. pneumoniae isolates.

In 24 (27.6%) out of the 87 clinically collected K. pneumoniae strains from patients with primary K. pneumoniae infections at a tertiary university hospital in Egypt, the pks colibactin island markers clbB, clbQ, clbA, and clbN were simultaneously detected. As a result, the prevalence rate was 27.6%, representing the first epidemiological study on a pks island hosting K. pneumoniae in Egypt.

Our findings backed up Lan et al. results, which documented a high pks prevalence in K. pneumoniae with 26.8% [18]. Positive frequencies of pks among K. pneumoniae recovered from different body sites were listed to be 25.6 % and 16.7 %, respectively, in two prior studies conducted in Taiwan [3],[19]. The pks gene was also highly prevalent in E. coli isolated from bacteremia patients, varying from 31.5 % to 58 % [20].

Hypervirulent variants are classified according to specific capsular serotypes, particularly K1 and K2 strains, which have been documented to be the most virulent of the 79 capsular types [21]. In this study, PCR screening-detected K1 type in 79.2% (19/24) of the colibactin-positive K. pneumoniae. The co-carriage of K1 type and colibactin genes was significantly correlated [OR 7.6, 95% CI 2.49–23.194, P < 0.001]. A close association of pks genes to K1 type was previously reported by Lai et al., who found that 66% (35/53) of the pks were K1 [19]. K1 or K2 capsular serotype-specific genes are located on the same virulence plasmid and other virulence encoding genes, which contributes to resistance to phagocytosis and intracellular killing by macrophages and neutrophils. This may partially explain the absence of non K1/K2 serotypes among pks isolates [22],[23].

On the other side, the existence of genotoxic K. pneumoniae was unrelated to the type of infection since colibactin-positive isolates were insignificantly distributed among bacteremia (11/29; 37.9%), respiratory infection (8/29; 27.6%), urinary tract infection (4/20; 20%), and other cases (1/9; 11.1%).

Despite the widespread use rmp A gene for genotypic hvKp detection, the hypermucoviscous phenotype is still thought to be one of the most relevant virulence factors of hypervirulent strains [24]. More than half of our genotoxic K. pneumoniae clinical isolates were positive for hypermucoviscous phenotype and rmp A gene (54.2% and 58.3%, respectively); this was significantly higher than in the non-genotoxic group (P < 0.001). The growing evidence stated by preceding publications [3],[18], indicating that the colibactin genotype is linked to hypervirulent strains, is also supported by our results of significant correlation of pks genes and mucoviscosity-associated rmp A gene [OR 6.618, 95% CI 2.339–18.725]. More research is needed to determine if the pks genomic island contributes directly to virulence or acts as a signal tangled in pathogenesis.

The absence of a positive regulator required for full rmpA expression, or the presence of a negative regulator of Lon protease at the posttranscriptional level, could clarify the positivity for rmpA in the four phenotypically non-hypermucoviscous K. pneumoniae strains, which was previously detailed by Yu et al. [25].

The present study detected minimal antimicrobial resistance in pks isolates, as Shi et al. [26] reported. According to statistical analysis, colibactin-positive isolates were significantly less resistant to 8 of the 15 antimicrobial compounds tested than colibactin-negative isolates. This was probably due to the high numbers of hypervirulent serovars and virulence genes found in pks isolates, as virulence is typically associated with lower drug resistance. Alternatively, maybe partially attributable to the fact that all of these isolates were K1 and K2 capsular types, which are often less resistant to the antimicrobial agents [27],[28].

Multidrug-resistant hvKp strains that produce extended-spectrum lactamase (ESBL) or carbapenemase have been labeled a challenging scenario in clinical cases [29],[30]. MDR was detected in 29.2% of our colibactin-positive isolates relative to 50.8% in the negative group characterized from similar situations of low association of virulence with drug resistance. Even though the MDR frequency among pks-positive isolates is currently low, the evolution of MDR paired with genotoxicity is concerning. So antimicrobial resistance should be carefully monitored in these genotoxic isolates.

In line with other research [18], underlying disorders such as diabetes mellitus, as well as the K1 capsular type, were all significant risk factors for genotoxic K pneumoniae infections.

Conclusions

Hospital-acquired K. pneumoniae isolates in Egypt had an increased prevalence of the pks colibactin genotoxin. The presence of hypervirulent determinants in pks highlighted the genotoxin potential pathogenicity. Its distribution in several specimen types aroused concerns about genotoxic K. pneumoniae traits in various pathological conditions, which necessitates clinical attention and epidemic tracking.

Table 1.

Applied primers in the study.

Primer	DNA sequence (5′–3′)	PCR program	Amplicon size (bp)
pks genomic island
clbA	F:CTAGATTATCCGTGGCGATTCR:CAGATACACAGATACCATTCA	10 min at 95 °C,30 cycles (45 sec at 94 °C, 45 sec at 54 °C, and 1 min at 72 °C); 10 min at 72 °C [12].	1311
clbB	F:GATTTGGATACTGGCGATAACCGR:CCATTTCCCGTTTGAGCACAC		579
clbN	F:GTTTTGCTCGCCAGATAGTCATTCR:CAGTTCGGGTATGTGTGGAAGG		733
clbQ	F:CTTGTATAGTTACACAACTATTTCR:TTATCCTGTTAGCTTTCGTTC		821
Capsular serotypes and Virulence gene
K1	F: GGTGCTCTTTACATCATTGCR:GCAATGGCCATTTGCGTTAG	15 min at 95 °C,30 cycles (30 sec at 94 °C, 90 sec at 60 °C, and 90 sec at 72 °C); 10 min at 72 °C [13]	1283
K2	F:GACCCGATATTCATACTTGACAGAGR:CCTGAAGTAAAATCGTAAATAGATGGC		641
rmpA	F:ACTGGGCTACCTCTGCTTCAR:CTTGCATGAGCCATCTTTCA		516

Table 2.

Factors associated with carriage of pks genomic island in hospital-acquired K. pneumoniae isolates.

	N (%) of pks + isolates (n = 24)	N (%) of pks −isolates (n = 63)
Capsular types			<0.001*
K1	19 (79.2%)	21 (33.3%)	<0.001*
K2	5 (20.8%)	13 (20.6%)	1.00
Non K1/K2	0 (0%)	29 (46%)	<0.001*
Virulence factor
Hypermucoviscous phenotype	13 (54.2%)	8 (12.7%)	<0.001*
Regulator of mucoid phenotype A (rmp A) gene	14 (58.3%)	11 (17.5%)	<0.001*
MDR	7 (29.2)	32 (50.8%)	0.07
Sample type			0.340
Blood	11 (45.8%)	18 (28.6%)	0.127
Respiratory	8 (33.3%)	21 (33.3%)	1.00
Urine	4 (16.7%)	16 (25.4%)	0.387
Others	1 (4.2%)	8 (12.7%)	0.227

*Note: A statistically significant p value < 0.05.

Table 3.

Demographic and Clinical characteristics of pks-positive and pks-negative K pneumoniae-infected patients.

	N (%) of pks + isolates (n = 24)	N (%) of pks −isolates (n = 63)
Age, median, range	56 (50–65)	60 (43–61)	0.028*
Male	18 (75%)	53 (84.1%)	0.361
Underlying disorder
Diabetes mellitus	17 (70.8%)	21 (33.3%)	0.002*
Hypertension	11 (45.8%)	17 (27%)	0.093
Pulmonary infection	10 (41.7%)	17 (27%)	0.186
Liver cirrhosis	5 (20.8%)	12 (19%)	1.00
Renal disorder	4 (16.7%)	9 (14.3)	0.747
Surgery within 30 days	2 (8.3%)	4 (6.3%)	0.666

*Note: A statistically significant p-value < 0.05