Natural history of Acinetobacter baumannii infection in mice

Abstract

In 2017, the WHO identified Acinetobacter baumannii as the top priority for the development of new antibiotics. Despite the need for new antibiotics, there remains a lack of well validated preclinical tools for A. baumannii. Here, we characterize and validate a mouse model for A. baumannii translational research. Antibiotic sensitivity for meropenem, amikacin, and polymyxin b was determined by the broth microdilution MIC assay. LD100 inoculums, in both blood and lung infection models, were determined in male and female C3HeB/FeJ mice that were challenged with various A. baumannii clinical isolates. Blood (blood infection model) or blood and lung tissue (lung infection model) were collected from infected mice at 2 and 18 hours and the bacterial burden was determined by quantitative culture. Blood chemistry was analyzed using the iStat system. Cytokines (IL-1ß, TNF, IL-6, and IL-10) were measured in the blood and lung homogenate by ELISA assay. Lung sections (H&E stains) were scored by a pathologist. In the blood infection model, the cytokines and physiological data indicate that mice become moribund due to sepsis (low blood pH, falling bicarbonate, and a rising base deficit), whereas mice become moribund due to respiratory failure (low blood pH, rising bicarbonate, and a falling base deficit) in the oral aspiration pneumonia model. We also characterized the timing of changes in various clinical and biomarker endpoints, which can serve as a basis for future interventional studies. Susceptibility was generally similar across gender and infection route. However, we did observe that female mice were approximately 2-fold more sensitive to LAC-4 ColR in the blood infection model. We also observed that female mice were more than 10-fold more resistant to VA-AB41 in the oral aspiration pneumonia model. These results establish parameters to follow in order to assess efficacy of novel preventative and therapeutic approaches for these infections.

Fig 1. The effect of gender on mouse susceptibility, blood infection model.

Mice were infected with the LD100 inoculum listed in Table 1 unless otherwise specified. Variation in host susceptibility by gender was not consistently observed across all strains or infection routes. In the blood infection model, a gender effect (female mice were more sensitive as compared to males) was only observed with mice infected with LAC-4 Col^R.

Fig 2. The effect of gender on mouse susceptibility, OA infection model.

Mice were infected with the LD100 inoculum listed in Table 1 unless otherwise specified. Variation in host susceptibility by gender was not consistently observed across all strains or infection routes. In the OA infection model, a gender effect (males were more sensitive as compared to females) was only observed with mice infected with VA-AB41.

Fig 3. Blood and lung bacterial densities in male and female mice infected with A. baumannii strains.

Male and female C3H/FeJ mice were challenged with A. baumannii strains HUMC1, VA-AB41, or LAC-4 Col^R. A) Mice were infected IV with >LD100s for each strain: HUMC1 (2.0x10⁷ CFU), VA-AB41 (2.7x10⁷ CFU), LAC-4 Col^R (9.7x10⁷ CFU). Blood was collected at 2 and 18 hours and serial dilutions were plated to determine the bacterial burden. No significant difference was observed between gender. At 2 hours, a significant difference was observed between HUMC1 vs VA-AB41 and LAC-4 Col^R infected mice (P < .01). At 18 hours, a significant difference was observed between HUMC1 and LAC-4 Col^R infected mice (P < .01). B) Mice were challenged via the lung with >LD100s for each strain: HUMC1 (4.7x10⁸ CFU), VA-AB41 (6.2x10⁸ CFU), LAC-4 Col^R (7.6x10⁷ CFU). Blood (B) or lungs (C) were collected at 2 and 18 hours and quantitatively cultured. No significant difference was observed between gender for the blood CFUs. At 18 hours, VA-AB41 was significantly different compared to HUMC1 and LAC-4 Col^R (P<0.05). For the lung homogenate, a significant difference was observed at 2 hours for LAC-4 Col^R vs VA-AB41 and HUMC1 (P < 0.01). At 18 hrs, there was a significant difference between HUMC1 and LAC-4 Col^R infected mice (P<0.01).

Fig 4. Temperature and clinical scores in infected mice.

Male and female C3H/FeJ mice were challenged with with A. baumannii strains HUMC1, VA-AB41, or LAC-4 Col^R. Mice were infected using the bacteremia and oral aspiration pneumonia models. Unless otherwise noted, no significant difference was observed between gender. A) At 2 hours, there was a significant difference between LAC-4 Col^R and. HUMC1 (P<0.05). At 18 hours, there was a significant difference between VA-AB41 and LAC-4 Col^R and HUMC1 (P<0.05). B) At both 2 and 18 hours, there was a significant difference between VA-AB41 vs. LAC-4 Col^R and HUMC1 (P<0.05). C) At 2 hours, there was a significant difference between gender for LAC-4 Col^R infected mice (P<0.05). (P<0.05). At 18 hours, a significant decrease was observed for all infecting strains (P<0.05). D) At both 2 and 18 hours, a significant decrease in activity was observed for all infecting strains (P<0.01).

Fig 5. Sepsis biomarkers.

Blood samples were collected from healthy and infected mice at 2 and 18 hours post-infection. BUN = blood urea nitrogen, a marker of renal function. Base excess = difference between measured and predicted serum bicarbonate. Falling base excess is consistent with a metabolic acidosis due to sepsis, while rising base deficit is consistent with respiratory failure, trapping CO2 in the blood, resulting in increased bicarbonate and carbonic acid levels. BUN was significantly increased at 18 hours for all infecting strains in the bacteremia model but only HUMC1 and VA-AB41 in the OA model (P<0.05). Bicarbonate was significantly increased in the OA infection morel for all infecting strains (P<0.05). Glucose was significantly decreased for all infecting strains in both the bacteremia and OA models at 18 hours post infection (P <0.05). The pH was significantly decreased in both infection models for only mice infected with LAC-4 Col^R (P < .001).

Fig 6. Host immune response.

Blood samples were collected from healthy and infected mice at 2 and 18 hours post-infection. TNF was significantly upregulated at 2 hours and 18 hours for all infecting strains in both the bacteremia and OA infection models (P<0.05) but only at 18 hours for IL-6 (P<0.05).

Fig 7. Host immune response.

The expression of the pro-inflammatory cytokine Il-1β was more variable and was not significantly upregulated across all conditions. The mice infected with LAC-4 Col^R or VA-AB41, the anti-inflammatory cytokine IL-10 was upregulated at 18 hours in the bacteremia model (P<0.05), however there was no significant difference in the OA infection model. This could be due to the lesser disease severity in the OA model as compared to the bacteremia model. Il-6 was significantly upregulated at 18 hours for all infection conditions (P<0.05).

Fig 8. Histopathology of the lungs at 2 hr post infection in the pneumonia model.

A) Baseline (0 hr) histology of uninfected lung (100X magnification). At 2 hours post-infection mice infected with HUMC1 (B), VA-AB41 (C), and LAC-4 Col^R (D) demonstrated interstitial edema and neutrophilic inflammation adjacent to the bronchovasculature, with clear alveolae. H&E stain.

Fig 9. Histopathology of the lungs at 18 hr post infection in the pneumonia model.

Severe hemorrhagic pneumonia is seen at 18 hours. A) Representative image of HUMC1-infected lung showing relatively mild acute insterstitial inflammation (400X magnification). B) VA-AB41 infection resulted in a heavy interstitial neutrophilic response (400X magnification). C) In LAC-4 Col^R infected lung, heavy organism load is accompanied by an exuberant neutrophilic infiltrate adjacent to bronchovascular structures (400X magnification). D) HUMC1-infected lung: low power examination illustrates expansion of interstitium by acute inflammation and distortion of normal air space architecture (50X magnification). E) Lung infected by VA-AB41 on low scanning power examination shows central distribution of the infectious process in this mouse model (25X magnification). F) LAC-4 Col^R lung infection showed areas of tissue necrosis associated with high bacterial burden and acute inflammation (400X magnification).