Necessity of Bumped Kinase Inhibitor Gastrointestinal Exposure in Treating Cryptosporidium Infection

Abstract

There is a substantial need for novel therapeutics to combat the widespread impact caused by Crytosporidium infection. However, there is a lack of knowledge as to which drug pharmacokinetic (PK) characteristics are key to generate an in vivo response, specifically whether systemic drug exposure is crucial for in vivo efficacy. To identify which PK properties are correlated with in vivo efficacy, we generated physiologically based PK models to simulate systemic and gastrointestinal drug concentrations for a series of bumped kinase inhibitors (BKIs) that have nearly identical in vitro potency against Cryptosporidium but display divergent PK properties. When BKI concentrations were used to predict in vivo efficacy with a neonatal model of Cryptosporidium infection, these concentrations in the large intestine were the sole predictors of the observed in vivo efficacy. The significance of large intestinal BKI exposure for predicting in vivo efficacy was further supported with an adult mouse model of Cryptosporidium infection. This study suggests that drug exposure in the large intestine is essential for generating a superior in vivo response, and that physiologically based PK models can assist in the prioritization of leading preclinical drug candidates for in vivo testing.

Keywords: Cryptosporidium; drug development; gastrointestinal; physiologically based pharmacokinetic model.

Figure 1.

Predicting in vivo efficacy with dose-normalized bumped kinase inhibitor (BKI) maximal concentration (C_max) and area under the curve (AUC). A, Observed reduction in a neonatal mouse model of Cryptosporidium parvum infection was plotted as a function of the dose-normalized C_max/EC₅₀ ratio for each BKI where EC₅₀ is the BKI concentration that inhibits 50% of Cryptosporidium growth in vitro. There was not a significant association between the 2 values. B, Similar to the dose-normalized C_max/EC₅₀ ratio, the dose-normalized AUC was not a predictor of in vivo efficacy. C, Finally, sigmoidal E_max model was used to predict the in vivo efficacy, but neither the total nor the unbound dose-normalized C_max predicted the in vivo efficacy.

Figure 2.

The full systemic and gastrointestinal (GI) bumped kinase inhibitor (BKI) concentration versus time curve was simulated for each BKI concentration over the duration of the neonatal mouse study. Gastroplus software was used to generate a physiologically based pharmacokinetic model to simulate drug dissolution, absorption, distribution, and elimination for each BKI; BKI 1294 (A, C) and 1553 (B, D) are shown as examples. A, B, Simulated total (solid lines) and unbound (dashed lines) BKI plasma concentrations. C, D, Exposure profiles in the lumen of the GI tract. Each line symbolizes a section of the Gastroplus advanced compartmental absorption and transit model. Because the plasma half-life of BKI 1553 is longer than that of BKI 1294, the systemic accumulation is greater for BKI 1553 over the duration of the study. In the GI lumen, whereas BKI 1294 exposure is predicted to be maintained between doses in the cecum and the ascending colon, BKI 1553 levels are predicted to rapidly rise and fall in each section of the GI tract after each dose, producing long periods with little BKI 1553 exposure between dosing.

Figure 3.

Simulated systemic and gastrointestinal (GI) concentrations of bumped kinase inhibitor (BKI) were used to predict the neonatal in vivo efficacy. A, B, As expected from the dose-normalized predictions, the model described with Equation 1 did not predict the in vivo efficacy when simulated plasma maximal concentration (C_max) (A) or average concentration (C_avg) (B) was used for the predictions. C, However, a significant (P < .05) association between predicted and observed efficacy was obtained when the ascending colon enterocyte C_avg (R² = 0.60) or C_max (R² = 0.47) levels were used with the sigmoidal E_max model. D, The ascending colon lumen concentration of BKI did not predict the in vivo efficacy.

Figure 4.

Simulating Cryptosporidium parvum growth in vivo and predicting in vivo efficacy with an adult mouse model of cryptosporidiosis. Adult interferon γ knockout mice were infected with nanoluciferase-expressing C. parvum oocytes. Therefore, oocyst excretion can be quantified by measuring luminescence after lysis of the fecal content. A, In vivo growth rate of C. parvum was simulated by log-transforming the observed relative luminescence in the feces of 3 vehicle-treated mice and fitting the data with a linear regression. B, C, The resulting growth rate was incorporated into the in vivo efficacy studies with bumped kinase inhibitor (BKI) 1553 (B) and 1294 (C). Ascending colon lumen or enterocyte levels were used to predict the inhibition of C. parvum growth with 5 daily oral doses of BKI 1553 (10 mg/kg) (B) or BKI 1294 (60 mg/kg) (C). The slope of the predicted relative luminescence unit (RLU) values was compared with that observed in the mice treated with BKI by analysis of covariance, and if the slopes differed significantly (P ≤ .05), the prediction was considered inaccurate. In vivo efficacy was accurately predicted with BKI 1553 enterocyte values for the average and maximal concentrations (C_avg), but the lumen values overpredicted the in vivo efficacy (B). Similarly, the BKI 1294 efficacy was accurately predicted with the enterocyte C_max and C_avg, but the lumen BKI 1294 levels over predicted the in vivo efficacy (C). For each study, 3 mice were used in both the BKI-treated and the vehicle-treated groups. Open circles represent results generated from the pooled feces of the 3 mice.

Figure 5.

Simulated enterocyte bumped kinase inhibitor (BKI) concentrations in the ascending colon accurately predicted an in vivo dose response. Based on the accuracy of predicted efficacy with ascending colon enterocyte levels of BKI 1294 and 1553, the enterocyte maximal and average concentrations (C_max and C_avg) for BKI 1534 were used to predict in vivo efficacy with 5 daily oral doses of BKI 1534 at 6 (A), 20 (B), or 60 (C) mg/kg. Predicted efficacy was compared with observed efficacy in an experiment where dosing was initiated on day 3 after infection (open circles). The slope of the predicted relative luminescence unit (RLU) values was compared with that observed in the BKI-treated mice by analysis of covariance, and if the slopes differed significantly (P ≤ .05) the prediction was considered inaccurate. Although the BKI 1534 enterocyte C_avg and C_max predicted the in vivo efficacy with daily dosing of 6 mg/kg (A), only the C_max predicted the in vivo efficacy at daily doses of 20 mg/kg (B) or 60 mg/kg (C). For each study, there were 3 mice in both the BKI-treated and the vehicle-treated groups. Open circles represent results generated from the pooled feces of the 3 mice.

Figure 6.

Plasma and gastrointestinal (GI) levels of bumped kinase inhibitor (BKI) observed after a single, oral dose. BKI plasma (A, B) and GI (C, D) concentrations were quantified 0.5, 2, and 4 hours after a single oral dose of BKI 1553 (10 mg/kg) or BKI 1294 (60 mg/kg). Despite a dose that was 6-fold greater, BKI 1294 plasma concentrations were much lower than BKI 1553 concentrations, in good agreement with previous results. However, systemic levels of BKI did not represent GI concentrations, and BKI 1294 levels were significantly greater in each of the 4 GI sections measured in this study.