Reversing resistance to anticancer treatment supported by pharmacokinetic/pharmacodynamic modeling of tumor growth kinetics in xenograft mice
Glenn Gauderat, , Sylvain Fouliard and Marylore Chenel
Clinical Pharmacokinetics and Pharmacometrics division, Servier, France
Objectives:
The development of drug combination in oncology is a key domain for answering therapeutic need. For example, combination therapy can be used to overcome drug resistance. In this context, the standard of care drug A is a targeting drug approved for the treatment of Non-Small Cell Lung Cancer treatment (NSCLC). Unfortunately, about half of patients develop drug resistance within 10 months. Resistance-lifting drug S is a candidate to be administered in combination with the standard of care in order to reverse tumor resistance and drug B is a competitor drug targeting the same pathway. Using a Pharmacokinetic/Pharmacodynamic (PK/PD) modeling approach, the aim of the present study was to compare our candidate drug to the competitor, when administered in combination with the standard of care on a resistant NSCLC tumor model.
Methods:
Tumor growth kinetics were determined on SCID mice bearing human NSCLC xenografts resistant to the standard of care drug. Six treatment groups of 8 mice were used to compare drug combinations potencies: control, drug S in monotherapy, drug A in monotherapy, drug B in monotherapy, drug S + drug A and drug B + drug A. Two to 3 weeks daily oral administrations were initiated 18 days after subcutaneous inoculation of tumor cells. Both resistance-lifting drugs were measured in blood at two time points after two weeks of treatment. Extensive PK data from previous studies were used to determine the PK model structures and population parameters for both resistance-lifting drugs, while concentrations measured in the present study were used to estimate individual exposures through Bayesian estimation. Tumor volumes were measured daily during 60 days (including tumor relapse) and modeled using the Simeoni [1] tumor growth inhibition (TGI) model. Parameter estimation was performed using Phoenix® NLME™. The standard of care drug concentrations were not modeled since it was assumed that it did not exert any effect on its own. Resistance-lifting drugs potencies were estimated separately for each treatment group and compared both in monotherapy and in combination with the standard of care drug.
Results:
Although a high inter-individual variability was observed, individual blood concentrations of drug S and drug B were correctly described by the PK models developed previously. The PD model that best described the tumor volume data was a Simeoni TGI model with a linear effect model. Transit compartments and first-order rate constant of transit K1 were removed from the original model since they did not improve the model performance.
As expected, Drug S, A and B used as monotherapy had very little if any activity on tumor growth kinetics compared to controls. Monotherapy groups were fitted using tumor growth parameters estimates from the control group in order to get drug potency estimates in reasonable computation time. Drug S was more potent than drug A and drug B in monotherapy.
In combination, our candidate drug was two times more potent than the competitor, with tumor volumes predicted lower at the end of the treatment period. Interestingly, the tumor relapse data after combination therapies allowed estimating tumor growth parameters under combination treatment conditions that were considerably different from those estimated from the control group, with tumor growth being slower after relapse than for control group. Indeed, to estimate tumor growth parameters was mandatory to properly model the tumor volume in the combination treatment groups, with tumor growth rate being about 1.4 times lower for drug B + drug A and about 3.3 times lower for drug S + drug A compared to control.
Conclusions:
A PK/PD analysis allowed comparing two drug potencies when administered in combination with a standard of care drug. Besides being more potent that drug B, another unexpected asset of drug S was a reduced tumor growth rate after relapse. Now, it is unclear whether this difference in tumor growth parameters is due to the drug S itself or to the fact that a smaller tumor cell subpopulation remained at the end of the treatment period compared to drug B. The present study supports the benefit of tumor volume measurement after relapse to better characterize a drug activity.
References:
[1] Simeoni M, Magni P, Cammia C, De Nicolao G, Croci V, Pesenti E, Germani M, Poggesi I, Rocchetti M (2004) Predictive pharmacokinetic-pharmacodynamic modeling of tumor growth kinetics in xenograft models after administration of anticancer agents. Cancer Res 64:1094–1101