Exploration of a new tool for antineoplastic agents dosage
4 October 2019Antoine Dowek1,2, Laetitia Lê1,3, Tom Rohmer3, Hynd Remita4, Francois-Xavier Legrand5, Marc Lavielle3 et Eric Caudron1,2
1 Service de Pharmacie, Hôpital européen Georges Pompidou, Assistance Publique-Hôpitaux de Pais, 20 rue Leblanc, 75015 Paris
2 Lip(Sys)2 Chimie Analytique Pharmaceutique, UFR de Pharmacie, 5 rue Jean Baptiste Clément, 92290 Châtenay-Malabry, Université Paris-Sud, Université Paris-Saclay
3 Inria Saclay Ile-de-France, CMAP, école polytechnique Route de Saclay 91128 Palaiseau, France
4 Laboratoire de Chimie Physique, CNRS-UMR 8000, Université Paris-Sud, Université Paris-Saclay, Bâtiment 349, 91405 Orsay, France
5 Institut Galien Paris-Sud, Univ. Paris-Sud, CNRS, Université Paris-Saclay, 92290 Châtenay-Malabry, France
Antineoplastic agents are for most part very toxic medicines with a high risk of complications. The need for powerful analytical tools capable of identifying and quantifying them accurately is essential. The aim of this work is to evaluate the quantitative capabilities of an innovative analytical technique: surface enhanced Raman spectroscopy (SERS) through the the determination of the 5-fluorouracil (5FU). The technique is capable of enhancing the intensity of the signal of the analyte by using metallic nanostructures, called SERS substrates for the quantification of low concentrations with high specificity.
Materials and methods
Spherical silver nanoparticules (AgNPs) were synthesized by reducing silver nitrate by sodium citrate. These were brought into contact with the 5FU under optimal conditions in order to obtain the most intense SERS signal. A portable Raman spectrophotometer Mira® was used to measure 432 spectras (3 measures, 144 samples) corresponding to 8 concentrations (from 1 to 50 µg/mL). Analysis was repeated during six days using fresh synthesis of AgNPs. Because of the complexity of spectral data, multivariate analyses were developed to treat data with R ®software. The best model was chosen for the lowest predictive relative errors.
The most intense enhancement was performed for the following experimental conditions : 100 µL of 5FU with 500 µL of AgNPs, then 50 µL of acid citric buffer at pH 6, for a time of contact between AgNPs / 5FU of 8 minutes and an acquisition time of 3 seconds. Data analysis revealed an enhancement factor by 104, but also an important variability, especially between different syntheses. This variability was controlled by the development of appropriate mathematical tools. Concentrations from 1,0 to 2.5 µg/mL were removed because spectral signal was too low. The best results were obtained for a non-linear regression model with normalization by a Raman scattering from citrate allowing a relative prediction error of 9% at the 4µg/mL concentration. Spectral analysis in SERS combined with data optimization succeeded in reaching concentration of µg/mL which is 1000 times lower than conventional Raman spectroscopy.
The feasibility of quantifying 5FU in SERS for concentrations well below those measured in Raman has been demonstrated with good predictive capabilities. SERS is a specific and sensible analytical technique with a potential for many applications, from analitycal control to surface contaminations by antineoplastic agents.