Model-based personalized medicine in transplantation
Pierre Marquet
INSERM U850, Limoges, France; Univ. Limoges, France; CHU Limoges, France.
Personalized medicine encompasses at least precision diagnostics, targeted therapies (when available) or personalized treatment, as well as personalized patient care.
Anti-rejection therapy in organ transplantation has long been a privileged application field for the development of medicine personalization, owing to the narrow therapeutic margin of most immunosuppressants (IS), the unavoidable pharmacokinetic (PK) interactions between the drugs associated in the usual therapeutic regimens and the unescapable long term toxicity of most IS.
The variability in the therapeutic and toxic effects of immunosuppressants obviously has pharmacokinetic and pharmacodynamic (PD) origins, both with genetic and environmental causes. Over the last two decades, patient care has been improved by using PK monitoring in order to identify the sources of, and compensate as much as possible for, the variability of IS drugs therapeutic and side effects. This has involved population pharmacokinetic (popPK) analysis and modelling, Bayesian estimators of exposure and dose adjustment tools (some of which have been made accessible through websites).
At the same time, the IS pharmacogenetic-pharmacokinetic relationships have been explored. Among the many polymorphisms in genes coding for drug metabolizing enzymes or efflux transporters explored, only a few have shown a strong enough impact on PK to be taken into account for a priori (before the first dose) or a posteriori (based on measured concentrations) dose adjustment, and/or have been proposed as covariates in popPK models. The influence of pharmacogenetics on drug targets has been studied less so far, although it holds promise for a better understanding of the sources of pharmacodynamic variability.
The pharmacodynamics of these drugs, such as the measurement of the calcineurin pathway activity, IMPDH activity, leucocyte proliferation or activation in patients’ blood, have been explored so as to estimate their variability, investigate their sources and evaluate their potential use for treatment personalization. This is also a prerequisite for useful PK/PD modelling.
Finally, biomarkers are being searched to anticipate and improve the diagnosis of renal graft injuries and of graft failure. Ultimately, they may be usefully implemented in predictive models of graft function or survival.
In summary, a more global, pharmacogenetic-pharmacokinetic-pharmacodynamic-clinical approach might help to significantly improve individual treatment strategies in the long term, including informed and iterative selection of IS drugs and their dose adjustment, i.e. true treatment personalization. This will obviously require sophisticated models able to integrate many risk factors of different types and predict patient and graft outcome.