Female mice placed on study were at least 6 weeks of age at the time of arrival, weighed between 16 and 20.1 g on the day of exposure, and were free from clinical signs of disease at the time of exposure. Mice were randomized to study groups by body weight. Patterned assignment was utilized to select animals in Groups 1, 11, 12, and 13 to subgroups and to select all study animals to a challenge run order. At least one animal from each study group was selected for each challenge exposure run. One hundred ninety female BALB/c mice were exposed to aerosolized B. pseudomallei (strain K96243) at a targeted 8100 cfu/animal, equivalent to 50 × LD50 (actual mean inhaled dose of 11,100 cfu; 69 × LD50) on a single day. Ten aerosol exposure runs of 19 mice per run were conducted. Aerosol concentrations of all exposure runs ranged from 7450 cfu/animal to 18,300 cfu/animal (standard deviation 4119 cfu/animal). The mass median aerodynamic diameter of inhaled particles ranged from 1.26 to 1.31 µm (geometric standard deviation 1.47–1.56 µm), allowing for lower lung deposition of the inhaled particles33,34. Following aerosol exposure, mice were treated with test (ceftazidime) or control (sterile saline) article. Two treatment cohorts were included in the study. The first cohort examined the efficacy of a 14 day ceftazidime regimen through assessment of bacterial tissue burden and survival 60 days post-cessation of treatment. Eight distinct treatment regimens were assessed (Table 1) which were comprised of two treatment frequencies: twice daily (BID) and four times daily (QID); two initiation time points: 24 h post-exposure (24 h) and 48 h post-exposure (48 h); and two dosages: 600 mg/kg/day and 1200 mg/kg/day. The 600 mg/kg/day dosage was achieved by administering 300 mg/kg BID or 150 mg/kg QID while the 1200 mg/kg/day dosage was achieved by administering 600 mg/kg BID or 300 mg/kg QID. Doses were administered approximately every 12 h for BID dosing and approximately every 6 h for QID dosing. Tissue burden was examined in the kidney, liver, lung, and spleen in all treated mice, as well as in a cohort of untreated mice at 24, 36, 48, and 60 h post-exposure to assess disease progression. A satellite cohort of mice was used for pharmacokinetic analysis of three dosages of ceftazidime following a single dose administered approximately 24 h post-exposure. No exposed animals were excluded from survival analysis. No blinding or masking was utilized during the conduct of the study.
Female BALB/c mice, six to eight weeks of age at arrival, were procured from Charles River Laboratories (Stone Ridge, NY). General procedures for animal care and housing met AAALAC International recommendations, current requirements stated in the “Guide for the Care and Use of Laboratory Animals” (National Research Council, Current Edition), current requirements as stated by the U.S. Department of Agriculture through the Animal Welfare Act, as amended, and conformed to testing facility standard operating procedures. The animal use protocol was approved by the Institutional Animal Care and Use Committee (IACUC) and by the United States Army’s Animal Care and Use Review Office (ACURO). Study findings have been reported in accordance with ARRIVE guidelines.
Test and control articles
Ceftazidime for injection, USP (Hospira, Lake Forest, IL) was diluted to 95 mg/mL in sterile water for injection, USP. Dosing solutions were further diluted in sterile saline, USP to 15, 30, or 60 mg/mL to achieve dosages of 150 mg/kg/dose, 300 mg/kg/dose, and 600 mg/kg/dose, respectively. Sterile saline was administered to animals in the control group. Doses were administered intraperitoneally (IP) at a target volume of 10 mL/kg. Injections were administered at contralateral sites at each administration to minimize localized tissue damage.
Challenge agent and aerosol exposure
Burkholderia pseudomallei strain K92643 was utilized for animal exposures. Strain selection was based upon previous internal characterization. Internally conducted minimum inhibitory concentration (MIC) testing of the K96243 strain yielded a MIC of 1 µg/mL for ceftazidime. The lot used for the aerosol exposures was propagated and characterized by the Battelle Biomedical Research Center. Fresh B. pseudomallei suspensions were prepared by inoculating stock material with LB broth with 4% glycerol (LBG) and incubating at 37 °C (± 2 °C) and shaking at 250 rpm for 18–24 h. The resulting starter culture was diluted in sterile LBG to reach an OD600 of 0.200 (± 0.05). This culture was then incubated at 37 °C (± 2 °C) and 250 rpm for 18–20 h. The culture was examined for purity following gram-staining. The culture was then centrifuged at 10,000 rcf for ten minutes. The resulting pellet was washed and resuspended in phosphate buffered saline with gelatin and trehalose (BSGT). The wash and resuspension procedure was repeated a total of two times. The suspension was removed and adjusted with buffer to reach an OD600 of 2.1 (± 0.1). Aerosol exposures were conducted as previously described35. An inhaled dose of 8100 cfu equating to 50 × LD50, was targeted for consistency with previous internal model development experiments.
Whole blood collection and pharmacokinetic analysis
Cardiac blood collection was performed as a terminal procedure. Prior to collection, mice were anesthetized with a mixture of ketamine (80–100 mg/kg) and xylazine (5–10 mg/kg). Bacteremia specimens were collected into K3EDTA tubes and maintained at room temperature until processing. Pharmacokinetic (PK) samples were collected into K3EDTA tubes and stored on wet ice until processing. Plasma was harvested and sterile filtered using a 0.2 micron PES filter. Samples were stored in a freezer set to maintain − 80 °C until analysis.
Ceftazidime was extracted from plasma and quantitated using a high performance liquid chromatography tandem mass spectrometry (LC–MS/MS) method developed internally. It is a sensitive and high throughput method utilizing protein precipitation extraction followed by strong cation exchange (SCX) LC–MS/MS. Detection was accomplished using a Shimadzu Prominence XR Series HPLC and AB Sciex Triple Quad 5500 mass spectrometer in positive ionization mode (ESI+). Chromatographic separation was obtained using an Agilent Zorbax 300-SCX 5µ, 2.1 × 150 mm column, a gradient mobile phase consisting of 95:5 25 mM ammonium formate:acetonitrile (v:v) (A) and 70:30 25 mM ammonium formate:acetonitrile with addition of 500 mM ammonium formate (B) at a flow rate of 0.5 mL/min., and a column oven temperature of 30 °C. Ceftazidime pentahydrate obtained from US Pharmacopeia (USP) was used in the preparation of the solutions for the standards and quality control samples prepared in mouse plasma. A chemical analog, cefepime hydrochloride obtained from USP, was prepared as an internal standard. The addition of the internal standard allowed for reproducible quantification as the ceftazidime and cefepime were monitored using multiple reaction monitoring (MRM) at 547.0/468.0 and 481.0/86.0 respectively. Regression analysis was evaluated by analyzing a set of calibration standards at eight concentrations. The lower limit of quantitation for ceftazidime was determined to be 50 ng/mL in plasma using a 100 µL sample size.
Free plasma concentrations were adjusted using an estimated 26% and 15% protein binding for mice and humans, respectively36. PK and exposure parameters were estimated using non-compartmental analysis with Phoenix WinNonlin software (Certara L.P., Princeton, NJ). The following acceptance criteria were used to evaluate the concentration–time profiles: (1) the coefficient of determination (r2) for the terminal linear phase was greater than or equal to 0.85, (2) the time of the last observed concentration was greater than three times the half-life, and (3) AUC∞ had less than 20 percent of the area extrapolated. The mouse data were then fit to a compartmental model in WinNonlin which was used to simulate concentrations for PK/PD target attainment.
Tissue collection and bacterial enumeration
At the time of necropsy, a portion of the kidney, liver, lung, and spleen were aseptically collected from each mouse, weighed, and individually homogenized in 1 mL of Dulbecco’s phosphate buffered saline with 0.01% gelatin (BSG) using a gentleMACS Dissociator (Miltenyi Biotec, Bergisch Gladbach, Germany). Tissue homogenates were serially diluted in triplicate and spread on LB agar with 4% glycerol (LBGA) or Columbia blood agar. Spread plates were placed in an incubator set to maintain 37 °C for an average of 58 h (41.4–93.1 h). Plates were enumerated and the cfu/gram of tissue calculated for statistical analysis.
Statistical analyses were performed using SAS (version 9.4) and R (version 3.6.3) on the 64-bit platform. For determination of group sizes, power analysis based upon a single one-sided Boschloo’s Test conducted at a Type I error rate of 5% was conducted to detect statistically significant differences in survival rates between a single treated group and a control group. When assuming a control group has a 5% survival rate, ten animals per treatment group allows for 88% power when assuming the probability of survival in the treated group is at least 60%. Two additional animals included in the 48 h post-exposure groups allowed statistical significance to be maintained in the event of early deaths. The proportion of surviving animals and exact 95% confidence interval were calculated for each group in the efficacy cohort. One-sided Boschloo’s tests were used to compare survival proportions between each of the treated groups and the saline control group (Group 2) at end of study, end of treatment, and 30 days post cessation of treatment. Two-sided Boschloo’s tests were used to compare survival proportions between each pair of timepoints within each group. The Bonferroni-Holm multiple comparison procedure was used to control the overall Type I error rate across each set of tests at 5%. The log-rank test was used to test for significant differences in time to death between each of the treated groups and the saline control group. The Bonferroni-Holm multiple comparison procedure was used to control the overall Type I error rate across all the tests at 5%. Kaplan–Meier curves were plotted for each group to represent survival and time to death data visually. For quantitative assessment of bacterial loads in tissues, geometric means and 95% confidence intervals were calculated for each tissue type by group. Analysis of variance (ANOVA) models were fitted to the base-10 log-transformed bacterial loads to evaluate the effects of treatment on tissue bacterial loads. The mean bacterial load for each of the treatment groups were statistically compared to that of the saline control group (Group 2) within the model for each tissue type. Dunnett’s multiple comparison procedure was used to control the overall Type I error rate across all the tests per tissue at 5%.