J. Sep. Sci. 2008, 31, 1537 – 1549 N. S. Quiming et al. 1537 Noel S. Quiming 1, 2 Nerissa L. Denola 1, 3 Shahril Reza Bin Samsuri 1 Yoshihiro Saito 1 Kiyokatsu Jinno 1 1 School of Materials Science, Toyohashi University of Technology, Toyohashi, Japan 2 Department of Physical Sciences and Mathematics, College of Arts and Sciences, University of the Philippines Manila, Manila, Philippines 3 Department of Pharmaceutical Chemistry, College of Pharmacy, University of the Philippines Manila, Manila, Philippines Original Paper Development of retention prediction models for adrenoreceptor agonists and antagonists on a polyvinyl alcohol-bonded stationary phase in hydrophilic interaction chromatography Retention prediction models based on multiple linear regression (MLR) and artificial neural network (ANN) for adrenoreceptor agonists and antagonists chromato- graphed on a polyvinyl alcohol-bonded stationary phase under hydrophilic interac- tion chromatography were described. The models showed the combined effects of solute structure and mobile phase composition on the retention behavior of the analytes. Using stepwise MLR, the retentions of the studied compounds were satis- factorily described by a five-predictor model; the predictors being the %ACN, the log- arithm of the partition coefficient (log D), the number of hydrogen bond donors (HBD), the desolvation energy for octanol (FOct), and the total absolute atomic charge (TAAC). The inclusion of the solute-related descriptors suggested that hydro- philic interactions such as hydrogen bonding and also ionic interactions are possi- ble mechanisms by which analytes are retained on the studied system. ANN predic- tion models were also derived using the predictors derived from MLR as inputs and log k as outputs. The best network architectures were found to be 5-3-1 for the data- sets at pH 3.0 and 4.0, and 5-4-1 for the dataset at pH 5.0. The optimized ANNs showed better predictive properties than the MLR models for both training and test sets under all pH conditions. Keywords: Adrenoreceptor agonists / Artificial neural network / Multiple linear regression / Poly- vinyl alcohol-bonded phase / Quantitative structure – retention relationships / Received: November 19, 2007; revised: January 10, 2008; accepted: January 11, 2008 DOI 10.1002/jssc.200700598 1 Introduction The adrenoreceptor agonists and antagonists represent two groups of compounds that are widely used as thera- peutic agents. For example, the b-adrenoreceptor antago- nists, also known as the b-blockers, are used for the treat- ment of hypertension, angina pectoris, arrhythmia, and congestive heart failure [1, 2]. The a-, b 1 , and b 2 -agonists, on the other hand, are used illegally as performance- enhancing substances that can be used in human and animal sports [3, 4] (World Anti-Doping Agency (2003), the world anti-doping code: prohibited list, international standard, version 3.0. http://www.wada-ama.org; cited 12 June 2007). Other examples of b 2 -agonists such as salbuta- mol and bambuterol are employed for the treatment of asthma and chronic obstructive pulmonary diseases. HPLC methods using RP stationary phases have already been developed for the analyses of adrenoreceptor ago- nists and antagonists [5 – 10]. A general problem of such methods when applied to basic and polar substances is the severe peak broadening and tailing due to specific interactions of these analytes with the support. This problem may be overcome by using ion-pairing reagents [11 – 13] or by adding chaotropic anions in the mobile phase [14, 15]. Studies on the retention behavior of b- blockers on conventional RP-columns [14 – 17] as well as on monolithic columns [18] have already been reported. Fragkaki et al. [19] also reported the quantitative struc- ture – retention relationship of the GC-MS relative reten- Correspondence: Professor Kiyokatsu Jinno, School of Materials Science, Toyohashi University of Technology, Toyohashi 441- 8580, Japan E-mail: jinno@chrom.tutms.tut.ac.jp Fax: +81-532-48-5833 Abbreviations: ANN, artificial neural network; HBD, hydrogen bond donors; HILIC, hydrophilic interaction chromatography; LOO-CV, leave-one-out crossvalidation; MLR, multiple linear re- gression; PVA, polyvinyl alcohol; RMSEP, root mean square er- rors of prediction; TAAC, total absolute atomic charge i 2008 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.jss-journal.com