Pharmacokinetic Data Are Predictive of In Vivo Efficacy for Cefixime and Ceftriaxone against Susceptible and Resistant Neisseria gonorrhoeae Strains in the Gonorrhea Mouse Model

Abstract

There is a pressing need for drug development for gonorrhea. Here we describe a pharmacokinetic (PK)/pharmacodynamic (PD) analysis of extended-spectrum cephalosporins (ESC) against drug-susceptible and drug-resistant gonococcal strains in a murine genital tract infection model. The PK determined in uninfected mice displayed a clear dose-response in plasma levels following single doses of ceftriaxone (CRO) (intraperitoneal) or cefixime (CFM) (oral). The observed doses required for efficacy against ESC-susceptible (ESC^s) strain FA1090 were 5 mg/kg of body weight (CRO) and 12 mg/kg (CFM); these doses had estimated therapeutic times (the time that the free drug concentration remains above the MIC [fT_MIC]) of 24 h and 37 h, respectively. No single dose of CRO or CFM was effective against ESC-resistant (ESC^r) strain H041. However, fractionation (three times a day every 8 h [TIDq8h]) of a 120-mg/kg dose of CRO resulted in estimated therapeutic times in the range of 23 h and cleared H041 infection in a majority (90%) of mice, comparable to the findings for gentamicin. In contrast, multiple CFM doses of 120 or 300 mg/kg administered TIDq8h cleared infection in ≤50% of mice, with the therapeutic times estimated from single-dose PK data being 13 and 27 h, respectively. This study reveals a clear relationship between plasma ESC levels and bacterial clearance rates in the gonorrhea mouse model. The PK/PD relationships observed in mice reflected those observed in humans, with in vivo efficacy against an ESC^s strain requiring doses that yielded an fT_MIC in excess of 20 to 24 h. PK data also accurately predicted the failure of single doses of ESCs against an ESC^r strain and were useful in designing effective dosing regimens.

Keywords: antibiotic resistance; cefixime; ceftriaxone; clearance; gonorrhea; mouse model; pharmacokinetics.

FIG 1

Timelines for blood collection and in vivo efficacy testing of ESCs in the gonorrhea mouse model. Female BALB/c mice received a 17β-estradiol pellet implanted subcutaneously 4 days prior to administration of test and control antibiotics or PBS. Streptomycin (STR) and trimethoprim (TMP) were administered for the duration of the experiments. (A) For PK studies, plasma samples were collected from uninfected mice on day 0 at 0.08, 0.25, 0.5, 1, 2, 4, 8, and 24 h posttreatment (n = 4 mice/dose/time point). (B) For in vivo efficacy studies, mice were inoculated vaginally 2 days following implantation of a pellet with the lowest dose of N. gonorrhoeae that infects 80% to 100% of mice for at least 10 days (ID_80–100). Vaginal swab specimens were cultured for 2 days (days −1 and 0) after bacterial inoculation to confirm infection. Decreasing doses of CRO or CFM were administered following collection of specimens for culture on day 0; PBS was administered as a negative control in all experiments, and GEN (i.p., 48 mg/kg, 5 doses, QD) was administered as a positive control for H041 infections. The colonization load was quantitated for 8 days posttreatment by daily culture of vaginal swab specimens. Gc, gonococcus.

FIG 2

Pharmacokinetic analysis of ESCs in the gonorrhea mouse model. (A and C) The concentrations of free CRO (A) or CFM (C) in plasma samples collected from uninfected mice after a single dose of test antibiotic are shown as the raw data (dots) overlaid with the predicted values (lines) for each dose over a 24-h period. Each data point represents the mean concentration for 4 mice per dose. Modeling of each antibiotic dose was determined from analysis of the plasma drug levels using Phoenix WinNonlin software. Plasma concentrations were adjusted to account for plasma protein binding for CRO (60% bound) and CFM (72.5% bound). (B and D) Simulated PK profiles of the free serum concentration for all single doses of CRO (B) or CFM (D). The MICs of the ESC^s strain FA1090 are indicated (solid black line) for CRO (0.0075 µg/ml) and CFM (0.0075 µg/ml). The MICs of the ESC^r strain H041 are also indicated (dotted line) for CRO (2 µg/ml) and CFM (8 µg/ml) for comparison. MICs were determined by agar dilution (52).

FIG 3

A single 5-mg/kg dose of CRO showed 100% efficacy against an ESC^s strain and produced a therapeutic time of 23.6 h. (A) Percentage of mice infected with FA1090 following administration of a single dose (arrow) of CRO, given at concentrations of 30 to 0.06 mg/kg or PBS over 8 days posttreatment (n = 8 to 10 mice/dose/experiment). Changes in the percentage of mice colonized were analyzed with a Kaplan-Meier survival curve and a log-rank (Mantel-Cox) test; P was <0.02 when CRO doses ranging from 30 to 0.25 mg/kg were compared to the PBS control. Results represent those from four combined independent experiments (total n = 9 to 35 mice/dose total). (B) Percentage of mice that cleared infection within 48 h posttreatment with each dose of CRO. (C) Vaginal swab specimens were quantitatively cultured to determine the bacterial burden (number of CFU per milliliter) before (day 0) and following (arrow) administration of CRO. The mean bacterial burden is shown for each treatment group; error bars indicate the standard error of the mean (SEM). The limit of detection was 20 CFU/ml (dotted line). Differences in the mean number of CFU per milliliter recovered over the course of infection were determined by repeated-measures two-way analysis of variance (ANOVA) with Bonferroni’s post hoc analysis comparing the treatment groups to the PBS control group (P < 0.005 for all doses of CRO tested). (D) Fold change in the number of CFU per milliliter between pretreatment cultures (day 0) and cultures of specimens collected 48 h after CRO administration for each individual mouse. The dotted line represents no change, and solid lines indicate the median fold change for each treatment group. Significance was calculated using a Kruskal-Wallis test with Dunn’s multiple comparisons to compare CRO treatment groups to the PBS control (*, P < 0.0001). (E) The predicted therapeutic time (fT_MIC) that the free serum concentration of CRO remained above the MIC of FA1090 (0.0075 µg/ml) is shown for each dose of CRO.

FIG 4

A single 12-mg/kg dose of CFM showed 100% efficacy against an ESC^s strain and produced a therapeutic time of 36.8 h. (A) Percentage of mice infected with FA1090 following administration of a single dose (arrow) of CFM, given at concentrations of 60 to 0.19 mg/kg, or PBS over 8 days posttreatment (n = 9 to 12 mice/dose/experiment). Changes in the percentage of mice colonized were analyzed with a Kaplan-Meier survival curve and a log-rank (Mantel-Cox) test; P was ≤0.0003 when CFM doses ranging from 60 to 0.75 mg/kg were compared to the PBS control. Results represent those from two combined independent experiments (n = 9 to 22 mice/dose total). (B) Percentage of mice that cleared infection within 48 h posttreatment with each dose of CFM that was administered. (C) Vaginal swab specimens were collected daily and quantitatively cultured to determine the bacterial burden (number of CFU per milliliter) before (day 0) and following (arrow) administration of CFM. The mean bacterial burden is shown for each treatment group; error bars indicate the SEM. The limit of detection was 20 CFU/ml (bottom dotted line). Differences in the mean number of CFU per milliliter recovered over the course of infection were determined by repeated-measures two-way ANOVA with Bonferroni’s post hoc analysis comparing the treatment groups to the PBS control group (P < 0.0001 for CFM doses ranging from 60 to 0.75 mg/kg). (D) Fold change in the number of CFU per milliliter between pretreatment cultures (day 0) and cultures of specimens collected 48 h after CFM administration for each individual mouse. The dotted line represents no change, and solid lines indicate the median fold change of each treatment group. Significance was calculated using a Kruskal-Wallis test with Dunn’s multiple comparisons to compare CFM treatment groups to the PBS control group (*, P ≤ 0.02; **, P < 0.0001). (E) The predicted therapeutic time that the free serum concentration of CFM remained above the MIC of FA1090 (0.0075 µg/ml) (fT_MIC) is shown for each dose of CFM administered.

FIG 5

Multiple doses of CRO with a therapeutic time of 22.9 h were able to eradicate an ESC^r strain. (A) Simulated PK profiles of CRO available in the plasma of uninfected mice after administration of a 120-mg/kg dose delivered once (QD), twice (BID), or three times (TID) over a 24-h period. The MIC of CRO for H041 is indicated (black dotted line, 2 µg/ml). (B) The predicted therapeutic time that the free serum concentration of CRO remained above the MIC is shown for each dosing regimen. (C) Percentage of mice that cleared infection within 48 h posttreatment with each dosing regimen. Doses are arranged as decreasing therapeutic time. (D) Percentage of mice infected with H041 following i.p. administration of 120 mg/kg CRO (arrow) given QD, BID, or TID (over 24 h) over 8 days posttreatment (n = 8 to 10 mice per dose/experiment). GEN treatment was used as a positive control (n = 19), and PBS was used as a negative control (n = 19). For fractionated doses, the arrow indicates the time point at which the first dose was given. Results represent those from two combined independent experiments (n = 10 to 19 mice/dose total). The arrow indicates the time point at which the first dose was given. Changes in the percentage of mice colonized were analyzed with a Kaplan-Meier survival curve and a log-rank (Mantel-Cox) test (P ≤ 0.005 for 120 mg/kg CRO given BID or TID versus PBS). Significance was also calculated by comparing treatment groups to the GEN positive-control group, with TID treatment being more effective than GEN (P = 0.006) and QD treatment being less effective than GEN (P ≤ 0.001). (E) Vaginal swab specimens were collected daily and quantitatively cultured to determine the bacterial burden (number of CFU per milliliter) before (day 0) and following (arrow) CRO administration. The mean bacterial burden is shown for each treatment group; error bars indicate the SEM. The limit of detection was 20 CFU/ml (bottom dotted line). Differences in the mean number of CFU per milliliter recovered over the course of infection were determined by repeated-measures two-way ANOVA with Bonferroni’s post hoc analysis comparing treatment groups to the PBS control group (P ≤ 0.001 for GEN and 120 mg/kg for CRO TID and BID regimens). Changes in the recoverable bacterial burden over the course of the experiment were also compared to those achieved with GEN (P = 0.002 for 120 mg/kg CRO QD). (F) Fold change in number of CFU per milliliter between pretreatment cultures (day 0) and cultures of specimens collected 48 h after CRO administration for each individual mouse. The CRO dosing regimens are indicated on the x axis, and treatment groups are organized from the highest (left) to the lowest (right) therapeutic time. The dotted line represents no change, and solid lines indicate the median fold change for each treatment group. Significance was calculated using the Kruskal-Wallis test with Dunn’s multiple comparisons to compare CRO treatment groups to the PBS control group (*, P ≤ 0.01; **, P ≤ 0.0002).

FIG 6

No CFM treatment regimen was 100% efficacious against an ESC^r strain. (A) Simulated PK profiles of CFM available in the plasma of uninfected mice after administration of a 300-mg/kg dose delivered once (QD), twice (BID, 12 h apart), or three times (TID, 8 h apart) over a 48-h period of time. The MIC of CFM for H041 is indicated (black dotted line, 8 µg/ml). (B) The predicted therapeutic time that the free serum concentration of CFM remained above the MIC is shown for each concentration (and dosing regimen) of CFM tested. (C) Percentage of mice that cleared infection within 48 h posttreatment with each concentration and dosing regimen of CFM that was administered. Doses are arranged as decreasing therapeutic time. (D) Percentage of mice that were infected with H041 following p.o. administration of 300 mg/kg CFM (arrow) as QD, BID, or TID dosing regimens (given over 24 h) and from which specimens were cultured for 8 days posttreatment (n = 9 to 10 mice per dose/experiment). GEN treatment was used as a positive control (n = 19), and PBS was used as a negative control (n = 24); results represent those from two combined independent experiments. The arrow indicates the time point at which the first dose was given. Changes in the percentage of mice colonized were analyzed with a Kaplan-Meier survival curve and a log-rank (Mantel-Cox) test; P was ≤0.008 when all 3 dosing regimens of 300 mg/kg CFM (and GEN) were compared to the PBS control; however, GEN was significantly more effective than any of the CFM dosing regimens (P ≤ 0.005). (E) Vaginal swab specimens were quantitatively cultured to determine the bacterial burden (number of CFU per milliliter) on the day before (day 0) and following (arrow) administration of CRO. The mean bacterial burden is shown for each treatment group; error bars indicate the SEM. The limit of detection was 20 CFU/ml (bottom dotted line). Differences in the mean number of CFU per milliliter recovered over the course of infection were determined by repeated-measures two-way ANOVA with Bonferroni’s post hoc analysis comparing treatment groups to the PBS control group (P ≤ 0.01 for the TID treatment regimen of 300 mg/kg CFM and GEN). Changes in the recoverable bacterial burden over the course of the experiment were comparable to those achieved with GEN for all regimens except the 300-mg/kg BID treatment regimen, which was less effective (P = 0.03). (F) Fold change in the number of CFU per milliliter between pretreatment cultures (day 0) and cultures of specimens collected 48 h after CRO administration for each individual mouse. The concentration of CFM and the dosing regimen are indicated on the x axis; treatment groups are organized from the highest (left) to the lowest (right) therapeutic time. The dotted line represents no change, and solid lines indicate the median fold change for each treatment group. Significance was calculated using a Kruskal-Wallis test with Dunn’s multiple comparisons to compare CFM treatment groups to the PBS control group (*, P ≤ 0.0002).