Multi-parameter drug

Appendix 17 Multi-parameter Drug Development/Identification Software ... 

Klys M, Rojek S, Kulikowska J, Bozek E, Scislowski M (2007) Usefulness of multi-parameter opiates-amphetamines-cocainics analysis in hair of drug users for the evaluation of an abuse profile by means of LC-APCI-MS-MS. J Chromatogr B Analyt Technol Biomed Life Sci 854(1—2) 299—307. doi S 1570-0232(07)00337-6 [pii] 10.1016/j.jchromb.2007.04.040... 

This multi-parameter approach provides reliable information to drug discovery teams on the relative toxicity of new compounds in a class, toxicity relative to similar drugs already on the market, STRs, subcellular targets, identification of the mechanism of adverse effects, and plasma concentrations where toxicity would be expected to occur in vivo. 

Clements, M., Thomas, N. (2014). High-throughput multi-parameter profiling of electrophysiological drug effects in human embryonic stem cell derived cardiomyocytes using multi-electrode arrays. 

Segall, M. (2011) Multi-parameter optimisation in drug discovery. 

Lusher, S.J., McGuire, R., Azevedo, R., Boiten, J.W., van Schaik, R.C., and de Vlieg, J. (2011) A molecular informatics view on best practice in multi-parameter compound optimization. Drug Discovery Today, 16, 555-568. 

Figure 9.2 The optimization of active compounds towards drugs takes place in a multi parameter space.

Drugs and toxicants with multi-exponential behavior depicted in Figure 6.14 require calculation of the various micro constants. An alternative method involves using model-independent pharmacokinetics to arrive at relevant parameters. Very briefly, it involves determination of the area under the curve (AUC) of the concentration-time profiles. The emergence of microcomputers in recent years has greatly facilitated this approach. 

Combinatorial libraries are often synthesized using simple molecules with only a few synthetic steps, whereas natural products are produced in complex multi-step reactions with complex molecules. Also, the parameter space between natural products and chemical libraries seems to be different Synthesized drugs occupy an intermediate space between natural and chemical libraries, also a hint that the space of the drug libraries should be adjusted. 

The process of drug discovery amounts to the search for optimality in a hyper-dimensional, multi-response surface area, and thus is a complex process. As discussed earlier, log P and the other R05 parameters, together with other properties such as PSA, and the number of flexible bonds, are significant contributors to a large number of models for different ADME/T properties. The open question is whether the use of such models provides an added value, compared to the simple and easily interpretable R05 criteria or the recent trends for the drug-likeness formulated by Gleeson. But can these models provide added value ... 

The equivalent sink constraint is illustrated in Figure 8.8. In Figure 8.8A, the constraint holds and hence the parameters estimated from either the noncompartmental model (left) or the multicompart-mental model (right) will be equal. If the multi-compartmental model is a model of the system, then, of course, the information about the drug s disposition will be much richer, since many more specific parameters can be estimated to describe each compartment. 

Yui et al. found that a-cyclodextrin formed inclusion complexes with poly(e-lysine) in aqueous solution [151]. The inclusion complexation is controllable by changing several parameters, such as the molar composition between poly(e-lysine) and a-cyclodextrin, the pH and the ionic strength of aqueous media [152]. Yui s group is actively developing new polyrotaxane drugs utilizing multi-valent interactions [153-155]. 

Numerical identifiability also becomes a problem with a poorly or inadequately designed experiment. For example, a drug may exhibit multi-exponential kinetics but due to analytical assay limitations or a sampling schedule that stops sampling too early, one or more later phases may not be identifiable. Alternatively, if sampling is started too late, a rapid distribution phase may be missed after bolus administration. In these cases, the model is identifiable but the data are such that all the model components cannot be estimated. Attempting to fit the more complex model to data that do not support such a model may result in optimization problems that either do not truly optimize or result in parameter estimates that are unstable and highly variable. 

---