Delivering Drug to the Test System

During the hit identification phase and the hit-to-lead selection process, the chemical diversity of the compound libraries evaluated can be very large, hence dictating the need for a generic vehicle to deliver compounds into the test system of interest. Polar aprotic solvents such as dimethyl sulfoxide (DMSO) have become the mainstay for delivering compounds into high throughput screening (HTS) assays of all types and in some cases actually storage of compound libraries. The value of these solvents derives from their ability to solubilize a very diverse set of chemical structures and yet be completely miscible with aqueous media, thus acting as a transfer agent for delivery of compound to the aqueous based assays.

The potential impact of these diluted aprotic solvents on the readout of the assay is generally built into validation of the assays, wherein tolerances and ranges of acceptable solvent levels are defined. However, it is not really possible to account for the highly variable propensity for compound precipitation upon dilution of the aprotic solvent into aqueous media. This problem continues to plague HTS and is generally accepted as part of the noise affiliated with HTS at the hit-to-lead stage, the tradeoff being the exceedingly large number of compounds screened and a general feeling that if the compound is too insoluble upon dilution it is good that it is discarded anyway. This thinking is fine when the number of hits is high and leads plentiful, but becomes more of a concern when the target is such that the number of hits and leads are very low, and any sort of false negative is undesirable.

The probability of precipitation of the compound is particularly high upon dilution of an aprotoic solvent into aqueous media, especially if the drug is near saturation solubility in the solvent itself. The solubility of the compound decreases

Precipitation

Soluble

Fraction Cosolvent

Figure 1. Typical compound solubilization curve in a cosolvent system such as DMSO / water (solid line). Samples below the curve would be soluble, while those above the curve would be prone to precipitation. Upon linear dilution from Point A (where compound is near saturation solubility in the DMSO) the probability for precipitation is high. Dilution from B may or may not result in precipitation depending on the dynamics of the dilution and exactly what the final dilution is. Dilution from point C should not precipitate. SDMSO and SH2O are saturation solubilities in neat solutions of DMSO and water, respectively.

Fraction Cosolvent

Figure 1. Typical compound solubilization curve in a cosolvent system such as DMSO / water (solid line). Samples below the curve would be soluble, while those above the curve would be prone to precipitation. Upon linear dilution from Point A (where compound is near saturation solubility in the DMSO) the probability for precipitation is high. Dilution from B may or may not result in precipitation depending on the dynamics of the dilution and exactly what the final dilution is. Dilution from point C should not precipitate. SDMSO and SH2O are saturation solubilities in neat solutions of DMSO and water, respectively.

exponentially as a function of decreasing cosolvent concentration, whereas the concentration of the drug and the aprotic solvent both decrease linearly upon dilution (Figure 1). Therefore, as shown in Figure 1, depending on the concentration initially in the DMSO solution (A vs B vs C on Figure 1) and the degree of dilution, the potential exists to be at concentrations exceeding the solubility of compound in that cosolvent mixture (curved bold line). In some cases, the final concentration may be at a point where it was soluble, but it was necessary to go through a region of coslovent composition where the compound could precipitate, but continue to a point where it would be soluble, leading to variable results depending on timing of transfers and readouts associated with the assay. The propensity for precipitation of a drug is governed by several processes; the actual concentration of the drug being a driver with respect to the statistical probability of molecules to come together for nucleation, the level of supersaturation, i.e. how high the concentration is above the thermodynamic saturation solubility, the inherent tendency for the solid to nucleate and then demonstrate crystal growth as dictated by solid energetics, and finally the factors which influence the diffusional rates of the molecules in the media.

As HTS or moderate throughput screening (MTS) assays are employed in the lead optimization stage, the level of chemical diversity is diminished, but the use of polar aprotic solvents often continues in this phase as well. However, since the goal is now one of optimization of structural properties, a readout free of ambiguity and false negatives becomes much more important to avoid miss-directing structural modification strategies. Unfortunately, once the optimization screening assays move toward in vivo evaluations, there is often a continued use of these aprotic solvents because they are known to solubilize the compound in previous in vitro studies. Continued use of DMSO often results from a lack of compound, resources or time to evaluate other vehicles or consider alternative delivery systems at this stage.

Lead optimization relies on a combination of HTS and MTS, along with a limited number of lower throughput assays, which often includes in vivo evaluation in animals. Irrespective of whether the assay involves ligand binding, ligand displacement, receptor occupation, or enzyme inhibition, the response is monitored as a function of the amount of compound added. However, the real driving force for a molecule to interact with a receptor is related to "free compound" in solution, or more correctly the chemical potential (|) of the compound in the respective media. The |i of the compound is a thermodynamic measure of the molecules escaping tendency from the solvating media surrounding it, hence the greater the | , the greater will be the tendency of the molecule to diffuse to a point of lower | or interact with another molecule or receptor, which in effect reduces it's Similarly, the rate at which a molecule traverses a membrane (d(mass)/d(time)) of given surface area (A) is the rate of flux per unit area (J).

Flux, J, is dictated by the effective permeability (Pe) and the concentration gradient of the compound from the apical (Capical) to the basolateral (Cbasolateral) side of the membrane, or more correctly, the difference in chemical potential between the two sides of the membrane, |apicaland |xbasolateral (Hilgers et al., 1990; Camenisch et al, 1996; Bohets et al., 2001).

In most cases where transport across a membrane is considered, removal and systemic distribution of the compound will result in sink conditions, making the flux dependent on the Pe and the | at the membrane surface. Pe is then directly proportional to the diffusion coefficient of the molecule through the membrane and the partition coefficient, P, between the membrane and the media (Pe oc p = (|imembrane / |imedia )) and inversely proportional to the thickness of the membrane. As reviewed by Camenisch et al., (1996), the form of the relationship between P and Pe can take many shapes depending on the data set chosen. Depending on the in vitro cell based assay, there may actually be an equilibrium reached if sink conditions are not present, wherein the transport rates will decrease with time until the | on both sides of the membrane is equal. In such situations, the final distribution of the compound will be related to the relative partitioning between two compartments and any binding or association that may occur in those compartments. In any event, it is clear that the actual flux across a membrane, or the equilibrium distribution, is related to the chemical potential of the compound in solution and not necessarily the concentration (Raub et al., 1993; Koeplinger et al., 1999).

Determination of drug concentration in solution is a simple concept, but is highly dependent on how separation and analytical methodology is employed to obtain the measurement. The analytical measurement of how much drug is "in solution" is often dependent on the intended use of the information. The two extremes range from the total amount of compound in solution, irrespective of the state of the drug (i.e. two phase dispersions, associated complexes, etc.) to the actual | of the drug in solution (indicative of the level of solvation at a molecular level.) The intended application of the "solubility" number will play a significant role in what type of number should be determined. If neither solubility nor dissolution are limiting, the total compound dispersed in the system is probably of greatest interest, as it all behaves as if were in solution. If a process is dissolution rate limited then the | and the total solubilized drug are important, whereas if solubility is limiting, the "free compound" or | is the primary measurement of interest. If the in vitro system involves receptor binding or occupation, the chemical potential is of greatest interest. It is important to always consider what is to be done with the solubility number to help define how to generate it.

Was this article helpful?

0 0

Post a comment