**3.2 Role of toxicokinetics in dog liver toxicity**

Even though this drug candidate was known to elicit a strong pharmacological response that could be capable of inducing the adverse effect observed in the dog toleration study, the potential of this drug candidate to form an acyl glucuronide (M2) in liver was evident and thus this metabolite may also be the cause of these adverse effects (Kenny et al., 2005). Furthermore, the potential preferential distribution of this drug candidate to the liver may also predispose its adverse effects. Therefore to investigate these hypotheses, the plasma and liver (also kidney and fat for comparison) were analyzed for drug candidate and its metabolites in the dog after 14 days of repeated daily oral doses of the drug candidate (i.e., parent).

After toxicokinetic evaluation of the tissues and plasma, the concentrations of parent in liver were consistently lower than plasma at 2, 6, and 24 hours postdose, suggesting no preferential distribution of the drug to the liver (Table 5). Furthermore, the acyl glucuronide metabolite (M2) along with other metabolites (M1, M3, and M4) were only observed in the plasma and not in the liver (Table 6), suggesting that these metabolites were not the cause of the observed liver toxicity. These results suggested that the observed liver toxicity in dog was caused by the strong pharmacological response of the drug candidate and probably not caused by an off target effect of M2 (or any other metabolites observed in plasma). Furthermore, the lack of preferential distribution of parent to the liver indicated that the toxicokinetic analysis of plasma exposure was correct in evaluating the risk for observed liver toxicity in the potential further development of this drug candidate.

### **3.3 Conclusion**

Toxicokinetic evaluation of tissue (where toxicity is observed) and plasma for drug and its metabolites will allow further mechanistic understanding of the cause of the observed tissue toxicity and will aid in the choice of the most relevant matrix for sampling in order for the correct evaluation of risk in further development of the drug candidate.

Knowledge of tissue toxicokinetics will increase the understanding about the potential mechanism of an organ-specific toxicity and can potentially assist in identifying a backup

This case example (described below) will highlight an investigation into liver toxicity where the mechanism of the liver toxicity was questioned. This drug candidate induced a strong pharmacological response; therefore, an investigation was launched to investigate whether the liver toxicity induced by this drug was a result of its strong pharmacology or an off target effect (i.e., independent of its targeted receptor pharmacology) from one of the

In a dog toleration study at the lowest dose tested (10 mg/kg), slight, acute central-lobular and portal inflammation with individual hepatocyte necrosis was observed. Therefore, no NOAEL could be assigned in this study which markedly complicated the further

Even though this drug candidate was known to elicit a strong pharmacological response that could be capable of inducing the adverse effect observed in the dog toleration study, the potential of this drug candidate to form an acyl glucuronide (M2) in liver was evident and thus this metabolite may also be the cause of these adverse effects (Kenny et al., 2005). Furthermore, the potential preferential distribution of this drug candidate to the liver may also predispose its adverse effects. Therefore to investigate these hypotheses, the plasma and liver (also kidney and fat for comparison) were analyzed for drug candidate and its metabolites in the dog after

After toxicokinetic evaluation of the tissues and plasma, the concentrations of parent in liver were consistently lower than plasma at 2, 6, and 24 hours postdose, suggesting no preferential distribution of the drug to the liver (Table 5). Furthermore, the acyl glucuronide metabolite (M2) along with other metabolites (M1, M3, and M4) were only observed in the plasma and not in the liver (Table 6), suggesting that these metabolites were not the cause of the observed liver toxicity. These results suggested that the observed liver toxicity in dog was caused by the strong pharmacological response of the drug candidate and probably not caused by an off target effect of M2 (or any other metabolites observed in plasma). Furthermore, the lack of preferential distribution of parent to the liver indicated that the toxicokinetic analysis of plasma exposure was correct in evaluating the risk for observed

Toxicokinetic evaluation of tissue (where toxicity is observed) and plasma for drug and its metabolites will allow further mechanistic understanding of the cause of the observed tissue toxicity and will aid in the choice of the most relevant matrix for sampling in order for the

drug candidate that has a markedly lower potential for this organ-specific toxicity.

**3. Case example: Toxicokinetics and liver toxicity** 

**2.4 Conclusion** 

metabolites of the drug.

**3.1 Liver toxicity in dog** 

**3.3 Conclusion** 

development of this drug candidate.

**3.2 Role of toxicokinetics in dog liver toxicity** 

14 days of repeated daily oral doses of the drug candidate (i.e., parent).

liver toxicity in the potential further development of this drug candidate.

correct evaluation of risk in further development of the drug candidate.


Table 5. Concentration-Time Profile of Parent in Dog Plasma, Liver, Kidney, and Fat.

Fed Beagle dogs (n=18) were administered a single oral dose or repeated daily oral doses for 14 days of 10 mg/kg drug. Liver, kidney, fat, and plasma were collected at 2, 6, and 24 hours post dose from three dogs at each time point with and without formic acid (formic acid was added to potentially increase the stability of the acyl glucuronide metabolite). Bioanalysis of liver, kidney, fat, and plasma for drug candidate was performed.


ND = not detected

Table 6. Peak Area Counts Versus Time Profile of Parent and its Metabolites in Dog Plasma, Liver, Kidney, and Fat.

properties for this chemical series to alter CNS distribution were not possible since these alterations markedly reduced potency for the pharmacological receptor. Interestingly, some of these molecules (in the same chemical series) were identified as substrates for Pgp. In the MDR1-MDCK cell model, the efflux ratio of the Pgp substrates was between 2 and 3. Since Pgp is known to reduce CNS distribution through efflux of drug candidate from the apical membrane of the endothelial cells in the blood brain barrier into the blood (Cordon-Cardo et al., 1989), the effect of Pgp on the CNS distribution of these potential backup molecules was determined in the mouse (as discussed previously, monkeys were not a practical model for this exploration). CNS concentrations were approximately 10-fold less for one of these backup drug candidates compared to the lead drug candidate (Figure 4). Therefore, this backup drug candidate was advanced into clinical trials and CNS toxicity was never

Fig. 3. Concentration-Time Profile of Lead Drug Candidate in Mouse Brain and Plasma after

0 20 40 60

Plasma

Brain

**Time (hr)**

Fasted CD1 mice (n=27) were administered a single oral dose (20 mg/kg) of the lead drug candidate. Brains and plasma were collected at 0.25, 0.5, 1, 2, 4, 6, 8, 24, and 48 hours post dose from three mice at each time point. Bioanalysis of brain and plasma of the lead drug

observed in monkey and human.

a Single Oral Dose (20 mg/kg)

10

100

1000

**Concentration (ng/mL or g)**

10000

candidate was performed.

Fed Beagle dogs (n=18) were administered a single oral dose or repeated daily oral doses for 14 days of 10 mg/kg drug. Liver, kidney, fat, and plasma were collected at 24 hours post dose from three dogs at each time point with and without formic acid (formic acid was added to potentially increase the stability of the acyl glucuronide metabolite). Bioanalysis of liver, kidney, fat, and plasma for drug (Parent) and it metabolites (M1, M2, M3, and M4) was performed. Peak areas were integrated for both parent and metabolites in each matrix. Data from kidney and fat are not shown.
