Epidemiological Features and Impact of High Glucose Level on Virulence Gene Expression and Serum Resistance of Klebsiella pneumoniae Causing Liver Abscess in Diabetic Patients

Introduction

Klebsiella pneumoniae after colonization (in its virulent state) crosses the intestinal-mucosal barrier and targets the liver through the portal vein system, subsequently causing K. pneumoniae-caused liver abscesses (KLA). K. pneumoniae is bacteria that predominantly contributes to pyogenic liver abscesses worldwide.¹ KLA is generically cryptogenic and frequently becomes complicated because of invasive infection to other organs due to K. pneumoniae virulence factors, which include, capsular serotype, rmpA, and aerobactin.² Because of invasive systemic infections, the KLA mortality rate equates to 10% within 30 days of hospitalization.³ Unfortunately, the dramatically increased multi-drug resistance (MDR) of hyper-virulent K. pneumoniae (hvKP), especially resistance to carbapenems and third-generation cephalosporins, presents a great challenge for clinical treatment.

KLA is accompanied by many chronic diseases. Studies have shown that diabetes mellitus is a concomitant condition in approximately 29.3–44.3% of liver abscess cases.^4,5 Hyperglycemia weakens the immune barrier and increases the susceptibility to hvKP infection by inhibiting phagocytosis⁶ and neutrophil extracellular traps (NETs) generation,⁷ reducing macrophages’ function,⁸ and impairing the bacteriolysis of antibodies and complement.⁹ Poor glycemic control, also affects the pathogenicity of K. pneumoniae. Lee et al found that high glucose levels stimulated the biosynthesis of polysaccharide and gene expression of hvKP’s cps, thereby increasing phagocytosis resistance and development of invasive syndrome.¹⁰ Exogenous glucose can also activate the production of K. pneumoniae’s type 3 fimbriae. These functions are under the control of global modulator cyclic AMP (cAMP) receptor protein (CRP) and cAMP-CRP signaling pathway.¹¹ However, it is unknown if the expressions of other known pathogenic factors, for instance, the rmpA, ompA, and colibactin are also affected.

The K. pneumoniae pathogenic factors help its survival and colonization outside the gastrointestinal tract. RmpA is present at the large virulence plasmid and promotes capsular production.¹² Outer membrane protein A (OmpA) is a unique protein, abundantly located in the outer membrane of Gram-negative bacteria, and it is highly conserved throughout its evolution.¹³ Since ompA expression is under the control of two stress-stimulated ribonucleolytic mechanisms, many environmental triggers can modulate ompA expression, for example, acid challenge,¹⁴ antimicrobial peptide,¹⁵ cAMP,¹⁶ and polyamines.¹⁷ Moreover, in Gram-negative bacteria, ompA performs many functions such as toxicity, invasion, serum resistance, adhesion, and biofilm formation.¹⁸

Enhanced polysaccharide biosynthesis stimulated by glucose further increased the K. pneumoniae resistance to phagocytosis of neutrophils and destruction of blood leukocytes. However, it has not been studied whether K. pneumoniae cultured with high-glucose increases its resistance to serum killing. Serum comprises >30 complement system proteins, this system is associated with the host’s innate immune response.¹⁹ The structural integrity of the cellular envelope is crucial for serum survival. Capsule and OmpA have been implicated in serum resistance.^18,20

In this investigation, the clinical manifestations associated with KLA patients in the presence or absence of diabetes and the association between glycemic control and the invasive syndrome were determined. Then, the present study aimed to explore the impact of high exogenous glucose levels on rmpA, ompA, and colibactin clbB gene expression and serum resistance of K. pneumoniae in in-vitro assays. It is expected that this will provide a new and reasonable explanation for the high incidences of sepsis and invasive infections in KLA patients with diabetes.

Materials and Methods

Study Population

This retrospective cohort investigation analyzed the KLA patients who were also diabetic and were enrolled at The First Affiliated Hospital of Anhui Medical University, from January 2020 to November 2022 in a tertiary medical center with a 3000-bed capacity. And was duly authorized by the human ethics council of First Affiliated Hospital of Anhui Medical University. The KLA patients were confirmed after the culture reports from the Department of Microbiology were reviewed and the diagnosis was based on the typical clinical symptoms, liver abscess cavity imaging examinations, and culture identification via K. pneumoniae blood isolate or puncture fluid samples. A total of 24 K. pneumoniae isolates were collected in non-diabetic group, including 14 strains from puncture fluid and 10 strains from blood. In diabetic group, 33 K. pneumoniae isolates were collected, including 23 strains from puncture fluid and 10 strains from blood. Diabetic patients were elaborated as those who were previously diagnosed with either type 2 or type 1 diabetes and/or those who were either taking insulin and/or oral hypoglycemic drugs. Participants’ laboratory reports, sex, underlying diseases, age, clinical manifestations, and management were gathered. Sepsis-3.0 criteria were used for defining sepsis.²¹ And the infection was deemed metastasized when the infection spread to a distant site by the same pathogen as the pyogenic liver abscess (K. pneumoniae).

Microbiologic Data

For recognizing all the isolates, Matrix-Assisted Laser Desorption/Ionization-Time of Flight mass spectrometry (MALDI-TOF MS, Vitek MS, bioMérieux, France) was utilized. In accordance with the 2019 Clinical and Laboratory Standards Institute’s recommendations, 28 antibiotics were tested for their susceptibility via VITEK 2 (Card number: AST-GN13) system or Kirby-Bauer Disk Diffusion (Oxoid, UK) (CLSI) method. K. pneumoniae (ATCC700603) and Escherichia coli (ATCC25922) isolates were kept as standards to ensure quality control. Agar dilution assessment with the help of ceftazidime and cefotaxime combined with clavulanate was performed to confirm ESBL. Resistance to carbapenem (meropenem, imipenem, and ertapenem) was confirmed via disk diffusion protocol.

Detection of Capsular Serotypes and Virulent Genes

Columbia blood agar plates were prepared for inoculating the strains which were then incubated for 18 hours at 37°C in 5% CO2 chamber. K. pneumoniae isolates DNA was collected via a DNA extraction kit (Sangon Biotech, Shanghai, China) by following the instructions provided by the manufacturer. K1, K2, K5, K20, K54, and K57 capsular serotypes and virulence-associated genes, including wcaG, magA, rmpA, alls, iucb, aerobactin, iroN, kfu, entB, irp1, clbA, clbB, and clbN, were determined by PCR. 25μL reactions were prepared with 2× Spark Taq PCR Master Mix (Solely Bio, Shandong, China) (12.5μL), forward and reverse primer (10 pmol/μL primer stock) (both 1μL), genomic DNA (1μL), and water (9.5μL). PCR was run; for 4 min initial activation at 95°C, then 30s cycles 35 times at 94 °C, 60 °C for 30s, and 72 °C for 60s, and lastly, a 10 min final extension at 72 °C. 1% agarose gel electrophoresis was performed for separating PCR products. Table S1 enlists all the primers employed.

Quantitative RT-PCR

3 K. pneumoniae (serotype K1) strains isolated from the participants and classic K. pneumoniae ATCC700603 were considered the representing strains in the subsequent experiments. The three clinical isolates were named KP133, KP164 and KP188. Among these isolates, KP133 was isolated from the non-diabetic group and the remaining two were from the diabetic group. K. pneumoniae were cultured in glucose-free LB broth, 0.5% glucose LB broth, or 0.5% glucose containing 1mM cAMP LB broth, respectively. After incubated at 37°C for 6h, total RNA was acquired from exponential growth phase bacterial cells via the Bacterial RNA Kit (YEASEN, China) according to the developer’s protocol. RNA was reverse-transcribed with the RT Master Mix (Takara, Japan) using random primers. qRT-PCR was carried out in a Light-Cycler 480 (Roche, Basel, Switzerland) instrument using SYBR Premix Ex TaqII (Takara, Japan). The cycling parameters included; 30s at 95°C, 5s of 40 cycles at 95°C, and 60°C for 20s. Relative gene expression was measured by the comparative threshold cycle 2^−ΔΔCTmethod. 23S rRNA was selected as an endogenous reference. All samples were analyzed thrice. The primer’s sequences are listed in Table S1.

Serum Killing Assays

The bacterial susceptibility to human serum was assayed using a modified method by Yeh et al.²² KP133, KP164 and KP188 strains were isolated from the participants and were considered the representing strains in the subsequent experiments. The serotypes and virulence-associated genes of the three clinical hvKP strains are presented in Table S2. Briefly, human serum from 8 healthy donors was incubated with LB broth cultured hvKP strains in the presence and absence of 0.5% glucose. Bacteria were diluted to 4×10⁶ colony-forming units (c.f.u.)/mL in physiological saline. The bacterial suspensions (25μL) and the acquired nonimmune human serum (75μL) were dispensed in 96 well-plate, mixed, and incubated at 37 °C. Viability was determined immediately after 1h, 2h, and 3 h incubation. After mixing, samples were taken and serial dilutions were plated on MH agar, then incubated at 37°C for 18h for colony counts.

Statistical Analysis

SPSS (25.0 version) was utilized for assessing the statistics. Normally distributed continuous variables were demonstrated as the mean ± standard deviation and for their comparison, Student’s t-test was applied. Categorical variable data were disclosed as n (%) and Fisher’s exact test was performed for their comparative analysis. Statistical analysis was performed using GraphPad Prism 6.01. All histograms related quantitative data were expressed as mean ± standard deviation (SD). Comparisons between two groups were made using an unpaired t-test. P<0.05 was considered significant.

Results

Comparison of Clinical Characteristics of KLA with Diabetes and without Diabetes

The 57 K. pneumoniae liver abscess diagnosed patient’s demographic manifestations, underlying disorder, and complications associated with diabetes (n=33, 57.9%) and without diabetes (n=24, 42.1%) are summarized in Table 1. More than half of the patients were male (42/57, 73.7%) and the average age of the patients at diagnosis was 58.5±14.7 years (range 30 to 99). Fever and abdominal pain were the predominant manifestations in these two groups. Diabetes followed by hypertension (29.8%, 17/57) and biliary tract diseases (28.1%, 16/57) was the most common underlying disease. Poor glycemic control patients had the trend of higher sepsis rate (51.5% vs 20.8%, P=0.028) and invasive infections (45.5% vs 12.5%, P=0.01), and therefore longer hospital stays (23.3±7.4 vs 17.3±8.6, P=0.01) than those with controlled glycaemia. Laboratory investigations recorded on admission are also shown in Table 1. The levels of CRP were positively correlated with the degree of infection, and >100 mg/L CRP indicated a serious infection. In this research, the proportion of CRP>100 mg/L in the diabetic group was significantly higher than that in the non-diabetic group.

Table 1 Baseline Characteristics, Clinical Presentation, and Outcome of KLA Patients with Diabetes and without Diabetes

Microbiological Characteristics and Antimicrobial Susceptibility

In addition to its natural ampicillin resistance, the isolated K. pneumoniae was highly sensitive to other antibiotics. A few strains were resistant to levofloxacin, ciprofloxacin, ampicillin/sulbactam, and cotrimoxazole, but there was no difference between the two groups (Table 2). Unfortunately, compared to the high antibiotic susceptibility of K. pneumoniae in the diabetic group, the non-diabetic group had two multi-drug resistant strains, including an extended-spectrum β-lactamase-producing K. pneumoniae (ESBL-KP) and a carbapenem-resistant K. pneumoniae (CR-KP) isolate.

Table 2 Antimicrobial Susceptibility of K. pneumoniae Isolates from KLA Patients with Diabetes and without Diabetes

Four serotypes were detected in K. pneumonia patients, these were K1 (57.9%, 33/57), K2 (19.3%, 11/57), K5 (1.8%, 1/57), and K57 (7%, 4/57), while K20 and K54 were not detected in any tested isolates (Table 3). Eight isolates (14.0%, 8/57) were not classified successfully (referred to as; K-non-typable isolates). The capsular serotypes prevalence among diabetic and non-diabetic isolates were similar. The expression of virulence genes including wcaG, magA, rmpA, kfu, iro, entB, irp1, aero, and alls showed no marked change between the diabetic and non-diabetic K. pneumoniae isolates. However, all diabetic isolates harbored rmpA and irp1, while 87.5% of non-diabetic isolates expressed these important virulence-associated genes (P=0.068). Colibactin encoded by the pks gene cluster (clbA, clbB, and clbN), damages DNA and enhances virulence.²³ In this research, the colibactin system markers clbA, clbB, and clbN were simultaneously identified in 42.1% (24/57) isolates, which were considered as pks⁺ K. pneumoniae.²⁴ The positive rates of pks⁺ K. pneumoniae among diabetic isolates (51.5%, 17/33) were higher than those among non-diabetic isolates (29.2%, 7/24). However, no notable difference (p=0.11) in pks expression between the two cohorts was noticed.

Table 3 Capsular Serotypes and Virulence Factors of K. pneumoniae Isolated from Patients with Diabetes and without Diabetes

Effect of Exogenous Glucose on the rmpA, ompA, and clbB Transcription of K. pneumoniae

The impact of exogenous high glucose level on the transcriptions of K. pneumoniae virulent genes was analyzed. With the help of classic K. pneumoniae ATCC700603 and 3 serotype-K1 hvKP clinical isolates, the effect of exogenous high glucose level on rmpA, ompA, and clbB genes expression were assessed. Virulence genes expression of the three isolates was shown in Table S2. The clinical isolates were cultured in LB broth (in the presence or absence of 0.5% glucose). The mRNA levels of rmpA, ompA, and clbB were quantified by qRT-PCR. As Figure 1B–D depicts, adding exogenous glucose (0.5%) in LB broth increased the mRNA of rmpA, ompA, and clbB genes expression. The intracellular cAMP levels were reduced in LB broth with glucose^25,26 and exogenous 1mM cAMP was added to the glucose-augmented broth to observe the mRNA levels of these genes. The exogenous cAMP addition suppressed the glucose effect on the transcription of rmpA and ompA, but did not affect clbB (Figure 1B–D). K. pneumoniae ATCC700603 did not express virulence factors rmpA and clbB. Our results showed high glucose levels also increased the transcription of ompA in K. pneumoniae ATCC700603, and this up-regulation was controlled by cAMP (Figure 1A). These results suggested that high glucose concentration up-regulated the transcription of rmpA and ompA in K. pneumoniae by the cAMP signaling pathway.

Figure 1 Exogenous glucose up-regulated rmpA and ompA transcription through the cAMP-signaling pathway. Quantitative reverse-transcription polymerase-chain-reaction (qRT-PCR) assays were performed to investigate the expression of ompA, rmpA, and clbB genes in K. pneumoniae ATCC700603 (A), clinical isolates KP133 (B), KP164 (C), and KP188 (D) cultured in just LB broth, 0.5% glucose LB broth, and 0.5% glucose containing 1mM cAMP LB broth at 37°C for 6h. *P<0.05, **P<0.01, ***P<0.001, ****P<0.0001. Abbreviation: n.s., not significant.

Environmental Glucose Stimuli Increased K. pneumoniae Resistance to Serum Killing

High glucose levels could enhance cps genes expression of hvKP, which increases resistance to phagocytosis.⁶ However, the literature suggesting whether K. pneumoniae cultured with high-glucose increases its resistance to serum killing is still missing. To evaluate the serum’s sensitivity to the K. pneumoniae pathogenic effect, grown in high glucose, serum killing assays were performed. Bacterial survival in normal human serum was recorded at different time points. As shown in Figure 2A, the survival rate of K. pneumoniae ATCC700603 strain decreased over time. However, after co-cultured with serum for 2 and 3 hours, the survival rate of the strain cultured in LB with high glucose level (0.5%) was increased compared with that in LB broth. Compared with classic K. pneumoniae ATCC700603, clinical isolates (KP133, KP164 and KP188) expressed virulence factors and showed a hypermucoviscous phenotype. The rate of hvKP strains survival was also markedly enhanced when grown in LB with 0.5% glucose than that in LB broth lacked glucose (Figure 2B–D).

Figure 2 K. pneumoniae cultured with high concentration of glucose increased its resistance to serum killing. Human serum from 8 healthy donors was collected and incubated with K. pneumoniae ATCC700603 (A), clinical isolates KP133 (B), KP164 (C), and KP188 (D) that had been previously cultured in LB and LB with 0.5% glucose. Viability was determined immediately and after 1h, 2h, and 3 h of incubation. *P<0.05. Abbreviation: n.s., not significant.

Discussion

Diabetes is a risk factor for infectious diseases, such as KLA, tuberculosis, melioidosis, and COVID-19.^27–30 Previous studies have shown that diabetes impairs the innate and adaptive immune system, impairing macrophages, T cells, neutrophils, and NK cells.³¹ This immune dysfunction increases the risk of infection in diabetic patients. Lin et al found that diabetic control was crucial for the clinical KLA characteristics, especially when dealing with metastatic complications associated with KLA.³⁰ Consistent with the above study, we observed that substandard control of glucose in patients leads to a higher rate of invasive infections than in those with controlled glycaemia. The major abnormalities of KLA in the laboratory were increased CRP and liver dysfunction.^32,33 It was also demonstrated in this investigation that the levels of CRP in diabetic-KLA patients were higher, which suggested a more severe infection. Sepsis is a lethal organ dysfunction caused by a dysregulated host response to an infection. Septic-KLA patients had an increased risk of chronic metastatic complications and prolonged hospital stays.³³ The results from this research indicated that the incidence of sepsis was higher and the hospital stays were longer in KLA with diabetes.

In this study, the virulence genes and capsular serotypes were detected by PCR. 94.7% (55/57) K. pneumoniae expressed rmpA and at least one siderophore. Moreover, hvKP strains isolated from KLA patients were sensitive to a range of antibiotics. So, empirical antibiotic therapy for liver abscesses (including three generation of cephalosporin, carbapenems and piperacillin-tazobactam) was generally effective. However, compared to the high antibiotic susceptibility of K. pneumoniae in the diabetic group, the non-diabetic group had two multi-drug resistant strains, including an ESBL-KP and a CR-KP isolate. The CR-KP isolate belonged to K2 serotype and expressed siderophores virulence genes (entB and irp1). Both the two multi-drug resistant K. pneumoniae isolates recovered from patients with hepatic diseases and repeated hospitalizations. Therefore, it is also necessary to pay attention to the occurrence of KLA after liver surgery, especially caused by multi-drug resistant K. pneumoniae.

Diabetes is an important risk factor for KLA.³⁴ Hyperglycaemia dysregulates the neutrophil phagocytosis of K. pneumoniae capsule serotypes K1 and K2.⁶ While, high glucose levels also affect K. pneumoniae pathogenicity, including enhancement of CPS and type 3 fimbriae.^10,11 The glucose levels used in our experiments is supposed to be similar to the human bloodstream concentrations. Literature suggests that healthy control would have 0.1% glucose blood concentration, while diabetic one would have 0.2–0.5% glucose blood concentrations.¹⁰ Our study showed that addition of exogenous glucose (0.5%) in LB broth enhanced mRNA of ompA and rmpA gene expression. RmpA is an important virulent factor in hvKP isolate, which could stimulate capsular production, resulting in the formation of hypermucoviscous phenotype.¹² In our study, 87.5% (21/24) of K. pneumoniae isolated from KLA without diabetes harbored the rmpA gene, while 100% of K. pneumoniae in KLA with diabetes were rmpA-positive. Furthermore, high glucose increased the mRNA of rmpA by regulating the cAMP-CRP signal pathway. As rmpA activates K. pneumoniae’s CPS biosynthesis, the results indicated that this is done not only by the direct cps genes regulation but also probably by increasing rmpA gene expression.

OmpA is an abundant predominant outer membrane protein of Enterobacteriaceae¹³ and is under the influence of many environmental stimuli. Gibert et al revealed that glucose augmented culture medium reduced the transcription of the ompA gene in Escherichia coli.¹⁶ Increased glucose inhibits intracellular second messenger cAMP production and inactivates the cAMP-CRP signaling pathway.^25,26 Liu et al demonstrated that ompA increased in protein levels in the crp deletion K. pneumoniae by proteomics analysis. However, there was no significant difference in the level of transcription in the high-glucose environment.³⁵ Therefore, the regulation of ompA by high glucose is still controversial. This investigation revealed that high glucose enhanced the mRNA of ompA and the addition of exogenous cAMP suppressed its enhancement.

Resistance to serum killing is associated with K. pneumoniae hyper-virulence. The CPS, O-antigen of LPS, and outer membrane proteins are important pathogenic factors that protect K. pneumoniae from being killed by the serum.¹⁹ K. pneumoniae was able to prevent complement activation by modifying sugar structures.³⁶ A thick outside polysaccharide layer creates a physical barrier in the bacterium that inhibits the penetration of the membrane attack complex (MAC), thereby preventing bacterial lysis.³⁷ Apart from the capsular polysaccharides and LPS structures, K. pneumoniae also utilizes outer membrane proteins to escape from the complement system detection.³⁸ OmpA directly restricts the complement cascade to mediate serum resistance.³⁹ Our results indicate that pre-incubation of K. pneumoniae strains in media containing high glucose enhanced rmpA and ompA gene expression, which might lead to increased bacterial resistance to serum killing.

Conclusion

In summary, KLA patients with diabetes showed a higher incidences of sepsis and invasive infections, and their length of hospital stay was also prolonged than those without diabetes. But there was no significant differences in antibiotic susceptibility and virulence genes between the two groups. Furthermore, high glucose levels increased the transcription of rmpA and ompA in hvKP by the cAMP signaling pathway and enhanced the resistance to serum killing, thus providing a new and reasonable explanation for the high incidence of sepsis and invasive infections in KLA patients with diabetes.