Desarrollo de herramientas combinadas para la evaluación experimental de la absorción transdérmica de fármacos con diferente actividad farmacológica
Fecha
2015-06-25
Autores
Ancede Gallardo, Evys
Título de la revista
ISSN de la revista
Título del volumen
Editor
Universidad Central “Marta Abreu” de Las Villas. Centro de Bioactivos Químicos.
Resumen
Disponer de datos de permeabilidad cutánea in vivo resulta de vital importancia en la evaluación de riesgo químico en la industrias farmacéutica y cosmética. Las agencias regulatorias exigen determinar estos datos de permeabilidad para cualquier compuesto que vaya a estar en contacto con la piel de alguna forma. Los ensayos in vivo son difíciles de realizar, presentan un alto costo y son criticados debido a consideraciones éticas. Para lidiar con este problema nosotros proponemos una metodología que combina los ensayos in vitro con los métodos in silico. Se construyó una base de datos formada por 583 compuestos de diferente
naturaleza estructural y con diferente actividad farmacológica, con la mayor homogeneidad posible y se separó en dos series, una de entrenamiento y otra de predicción. Se construyeron dos modelos predictivos usando árboles de clasificación y regresión (C&RT) con descriptores calculados con los programas computacionales Dragon y Modeslab respectivamente. Se obtuvieron dos árboles de desición que clasifican los compuestos en alta o baja permeabilidad con una especificidad de 83.48 y 86.16 para los descriptores del Modeslab y del Dragon
respectivamente. Se propuso un nuevo algoritmo para la estimación de Kp y tL a partir de la optimización no lineal de la ecuación de Scheuplein a partir de los datos experimentales de los ensayos in vitro. Este trabajo propone el uso de estas nuevas herramientas para el manejo de los datos experimentales y nuevos modelos in silico con una alta capacidad de predicción de la permeabilidad cutánea.
To have in vivo cutaneous permeability data is of vital importance in chemical risk assessment in the pharmaceutic and cosmetic industries. The regulatory agencies demand to determine these permeability data for any compound that is in contact with the skin in some way. The in vivo tests are difficult to perform and it have a high cost and are criticized due to ethical considerations. To deal with problem we propose a methodology that combines the in vitro assays with in silico methods. A database consists of 583 compounds of different structural nature and different pharmacological activity was constructed, with the greater homogeneity possible and separated into two series, one for training the models and another for prediction. Two predictive models were built using classification and regression trees (C&RT) with descriptors calculated with Dragon and Modeslab computational softwares respectively. Two decision trees that classify the compounds in high or low permeability with a global classification of 78.26% and 80% from Modeslab and Dragon descriptors respectively was obtained.We intended a new algorithm for estimating Kp and tL starting of the non lineal optimization of the Scheuplein’s equation from experimental data of the in vitro assays. This work proposes the use of these new tools for management of the experimental in vitro data and new in silico models with a high predictive power of skin permeability.
To have in vivo cutaneous permeability data is of vital importance in chemical risk assessment in the pharmaceutic and cosmetic industries. The regulatory agencies demand to determine these permeability data for any compound that is in contact with the skin in some way. The in vivo tests are difficult to perform and it have a high cost and are criticized due to ethical considerations. To deal with problem we propose a methodology that combines the in vitro assays with in silico methods. A database consists of 583 compounds of different structural nature and different pharmacological activity was constructed, with the greater homogeneity possible and separated into two series, one for training the models and another for prediction. Two predictive models were built using classification and regression trees (C&RT) with descriptors calculated with Dragon and Modeslab computational softwares respectively. Two decision trees that classify the compounds in high or low permeability with a global classification of 78.26% and 80% from Modeslab and Dragon descriptors respectively was obtained.We intended a new algorithm for estimating Kp and tL starting of the non lineal optimization of the Scheuplein’s equation from experimental data of the in vitro assays. This work proposes the use of these new tools for management of the experimental in vitro data and new in silico models with a high predictive power of skin permeability.
Descripción
Palabras clave
Industria Farmacéutica, Drug Industry, Absorción Cutánea, Skin Absorption, Piel, Skin, Dermis, Epidermis, Técnicas In Vitro, In Vitro Techniques