Prevalence and Virulence Profiles of Klebsiella pneumoniae Isolated From Different Animals

Abstract

Background: Klebsiella pneumoniae liver abscess (KLA) is an invasive disease, and the occurrence of infection is related to its virulence factors and colonization of the host's gastrointestinal (GI) tract. Some animal-sourced isolates share virulence factors with human pathogens. However, the potential of K. pneumoniae as a zoonotic agent has not been confirmed in murine infection model.

Objectives: To identify the prevalence and virulence profiles of K. pneumoniae colonization in companion and wild animals and subsequently determine the pathogenicity of selected strains.

Methods: Forty-five K. pneumoniae isolates (45/302) were obtained from faeces of companion or wild animals. Virulence factors, gyrA polymerase chain reaction with the restriction fragment length polymorphism (PCR-RFLP) and pulsed field gel electrophoresis (PFGE) were detected and compared with our previous collection of 60 human pathogens. For KLA model and cytotoxicity test, three animal-sourced isolates, CHKP0009 (snake, K1, KpII), CHKP0021 (turtle, K2, pLVPK, KpI, cluster I) and CHKP1027 (dog, non-K1/K2, HV, KpI, cluster III), with similar genotype and/or phenotype to human pathogens were selected and evaluated for their virulence with human hypervirulent K. pneumoniae (hvKp) CG43S.

Results: The prevalence of K. pneumoniae was higher in companion than wild animals. K. pneumoniae was primarily isolated from dogs, turtles and snakes. Some animal-sourced isolates carried virulence factors and revealed phylogenetic relatedness with human pathogens. In KLA model, BALB/c mice infected with snake isolate CHKP0009 and dog isolate CHKP1027 survived for 14 days but showed significant bacterial loads in the liver and spleen. Notably, the pet turtle isolate CHKP0021 presented comparable virulence with human hvKp CG43S and induced liver abscess formation. All three selected animal-sourced isolates could colonize in the GI tract and possess cytotoxic ability. These findings demonstrated pathogenicity of the animal K. pneumoniae isolates. In addition, the high prevalence of K. pneumoniae in companion animals and some isolates with virulence profiles suggested animal-sourced K. pneumoniae has the zoonotic potential to cause human disease.

Conclusion: Animals are the natural hosts of zoonotic pathogens. Some animal-sourced K. pneumoniae isolates are not only pathogenic in vivo but also exhibit phylogenetic relatedness to human pathogens, suggesting the existence of a zoonotic risk for K. pneumoniae between these two populations.

Keywords: Hypervirulent; Klebsiella pneumoniae; liver abscess; zoonotic pathogen.

FIGURE 1

XbaI‐PFGE analysis of K. pneumoniae isolates from various sources. A dendrogram was generated using BioNumerics software for the XbaI‐PFGE profiles of K. pneumoniae. Isolates were collected from human patients with liver abscess or pneumonia infection, companion animals (including turtles, snacks, dogs and alpaca), wild animals (including sea otter and Sunda pangolin) and environmental soils. Based on approximately 70% intra‐group similarity of pulsotypes, clusters I, II and III were identified and red‐framed.

FIGURE 2

Virulence assessment of selected animal‐sourced K. pneumoniae. (A) Bacterial suspensions containing turtle isolate CHKP0021 at 1 × 10³ (solid circles) or 1 × 10⁴ CFU (solid squares) were IP injected into mice (n = 6 per group), and their survival rates were plotted. Open circles represent the survival curve of mice infected with 1 × 10⁴ CFU of human hvKp CG43S, a pathogenic K. pneumoniae from a liver abscess patient, and was used as a positive control. (B) Bacterial loads in the liver and spleen of snake isolate CHKP0009‐ and dog isolate CHKP1027‐infected mice were determined at 14 days post‐infection. (C–E) Liver sections were prepared from experimental mice sacrificed at 72 h post‐infection, stained with H/E and microscopically examined at 200×. The representative section from PBS‐treated mice was the control in (C), and the infecting agent employed was turtle isolate CHKP0021 in (D) or human hvKp CG43S in (E). A microabscess was labeled with a black arrow. The scale bar represented 50 µm. (F) Hepatic inflammation of human hvKp CG43S‐ and turtle isolate CHKP0021‐infected mice were graded by using the modified 18‐point Knodell HAI and illustrated by box‐whisker plot. (G) Bacterial loads in the colon were determined at 1, 2 and 3 days after oral inoculation with a suspension containing 1 × 10⁹ CFU of turtle isolate CHKP0021 (circles), snake isolates CHKP0009 (squares) or dog isolate CHKP1027 (triangles). Data were presented as the geometric mean 95% confidence interval. (H) For the in vitro cytotoxicity test, human hvKp CG43S, turtle isolate CHKP0021, snake isolate CHKP0009, dog isolate CHKP1027 or Escherichia coli DH5α were co‐cultured with BNL CL.2 cells at a MOI of 100 for 2, 4 or 6 h. Cytotoxicity was determined after measuring the activity of lactate dehydrogenase released from the damaged cells. The result of the individual experimental group was compared with that of the E. coli DH5α group (control). *p < 0.05 and was considered as significantly different. Error bars represent SEs from three experiments.