Selection of an acceptable surfactant-free dissolution medium is common when developing analytical methods and dissolution tests for poorly soluble surfactant-free pharmaceutical substances. Pure and registered substances commonly utilize recommended non-aqueous solvents and assay methods. Aqueous media are however preferred in early drug development and pilot trials due to biopharmaceutical considerations or technological limitations.

Aprepitant is used as a model compound due to its low water solubility. According to literature, it exhibits a solubility of about 3–7 micrograms/ml between pH 2 and 10 and a sharp increase in solubility at pH 1 (about 130 micrograms/ml at pH 1.0). Based on the provided in vitro data, the substance demonstrates solubility values of 15.6 micrograms/ml (pH 1.2), 0.79 micrograms/ml (pH 4.5) and < 0.3 micrograms/ml (pH 6.8–7.4). Dissolution remains limited even in the presence of surfactants (PBS pH 6.6 + 0.1% SDS) [1–3]. These features make aprepitant an ideal model for studying hydrophobic substance behavior and the effects observed during their preparation for assay.

The goal of the study is to present practical criteria and recommendations when choosing an aqueous medium and sample preparation methods (including filtration) that minimize the systematic loss of the analytical signal.

MATERIALS AND METHODS

Materials. Aprepitant (99.6% purity, Fuxin Long Rui Pharmaceutical Co., Ltd., China). Purified water was produced by a laboratory water distiller. 95% ethanol. 10 ml disposable syringes. Syringe membrane filters of various materials and pores: PES 13 mm 0.45 microns, PVDF 25 mm 0.45 microns, Nylon (NY) 13 mm 0.22 microns, Polypropylene (PP) 13 mm 0.45 microns. All filters are hydrophilic.

Solution preparation. Aprepitant solutions were prepared as alcohol concentrates (e. g., aprepitant in 95% ethanol) that are diluted with purified water to target concentrations. The total organic phase content in solutions was less than 1%. Freshly prepared solutions were analyzed with and without syringe filtration. Sample stability was monitored by measuring sample parts with a spectrophotometer over time. The first 2 ml after filtration was discarded if it was not a fraction collection.

Instrumental analytical methods. The absorption spectra were recorded using the double-beam UV-visible spectrophotometer UV-2600 (Shimadzu Corp., Japan) with 1 cm quartz cuvettes. The measurements were carried out in the range of 185–500 nm in 0.2 nm increments (the corresponding dissolution medium was placed in the reference cell). Data were acquired and initially processed using UV-Probe software, version 2.42 (Shimadzu). Linearity and other key parameters were tested near operating concentrations.

Samples were weighed on Pioneer PA214 analytical scales (Ohaus Corporation, USA) with a 0.1 mg graduation (discreteness) and a standard deviation (reproducibility) of 0.1 mg.

Calibration and calculation of losses. The calibration dependence was determined using the straight-line method y = a·x + b, where y is the absorbance (Abs.) at 210 nm, and x is the concentration (μg/ml). The loss of substance during filtration was calculated as follows:

Formula1

Experimental observations. A series of measurements were carried out using different initial concentrations with sequential membrane filtration (fraction sampling through the same membrane), and acidic (0.1 M HCl) treatment to compare the solubility and stability of solutions.

RESEARCH RESULTS

Calibration curve

Data-based calibration dependence (tab. 1): y = 0,02983·x + 0,01318, R2 = 0,99541 (an alternative approximation of y = 0,0273·x + 0,0097, R2 = 0,9999 was obtained for a separate series of measurement, and the data were used to calculate concentrations of some control solutions).

Effect of filtration and membrane material

The research has shown that filtration of aqueous aprepitant solutions through membranes made of various materials (PES, PP, Nylon, PVDF) results in significant and hardly predictable losses of an analytical signal. The loss value varied in the wide 15–84% range depending on the initial concentration and polymer type. Summary of experimental series is provided in (tab. 2).

None of the investigational materials provided complete inertia with respect to aprepitant in an aqueous medium. The most critical losses were observed on PES membranes at high concentrations (up to 84%) and Nylon membranes (up to 59%). The high variability of the results (RSD of control solutions up to 7.5%) indicates that the process is not limited to simple adsorption.

Dynamics in sequential filtration (fraction filtration)

During sequential filtration of a sample in 10 mL fractions, fundamental differences in material behavior were found (figure).

PES membranes have a traditional saturation profile of the adsorption centers: their losses decreased to 38% for the first fraction and to 10% for the third fraction. An abnormal concentration growth (+20% to the initial level) was seen for the fourth fraction. It proves there is desorption of the accumulated substance or failure of the adsorption layer when the stream hydrodynamics is changed.

In contrast to this, PVDF filtration demonstrated a chaotic variability of losses (23–39%) without any pronounced saturation trend accompanied with a lower solution stability in time.

The role of mechanically induced nucleation

Experiments confirm that substance losses result from both adsorption and transitions of dissolved substances into a solid phase triggered by mechanical stress.

1. The mechanical action effect: passing the solution through a syringe without a filter led to a background signal in the 500 nm range. It proves that nucleation (crystal nuclei) from the oversaturated solution is provoked by pressure changes and turbulence.
2. Spectrum shift: in supersaturated solutions (36.8 micrograms/ml), a simultaneous drop in the absorption peak (210 nm) and an increase in light scattering in the long-wavelength region (300–500 nm) were recorded, which confirms that suspended particles were formed.
3. The filter as a crystallization center: when trying to filter supersaturated solutions, the membrane initiated heterogeneous nucleation, completely removing matter from the filtrate. That was not the case with transparent solutions.

Experiments in an acidic medium (0.1 M HCl) have shown that the protonated molecule of aprepitant formed a stable true solution, without pronounced effects of crystallization and adsorption (filtration losses for solutions of 17.9 micrograms/ml and 91.12 micrograms/ml were 1.6% and 0.7%, respectively).

DISCUSSION OF RESULTS

Our data correlate with studies indicating that membrane filtration is a critical stage of sample preparation, which can significantly distort the results of quantitative determination. In particular, when studying the equilibrium solubility of pharmaceutical substances, the concept of the “distortion effect of filter” (DEF) was introduced, showing that adsorption on the membrane surface can lead to unpredictable errors [4]. Our results confirm this thesis for aprepitant.

The analysis of the obtained data has, apparently, revealed two processes that reduce the analytical signal.

1. Polymer matrix adsorption. Like in similar papers [5], we observed the dependence of losses on material hydrophobicity. High PES and Nylon losses are in line with data [6] that take the materials as problematic for hydrophobic analytes and recommend using PTFE as a more inert alternative. The membrane saturation effect recorded in fraction filtration is also described for other classes of compounds.
2. Induced heterogenous nucleation. This appears to be the dominant factor for oversaturated aprepitant solutions. In contrast to the pure adsorption described for true solutions, we observed an optical density increase in the non-analytical region (300–500 nm) and a sharp drop in concentration after a mechanical action. We assume that the microporous structure of the filter and the shear stress during passage through the pores serve as triggers of the phase transition. This explains why losses can exceed 80%. It occurs because the filter does not just resorb the molecules, but also initiates avalanche crystallization of the entire sample.

The use of an acidic medium (0.1 M HCl) eliminated the problem, converting aprepitant into an ionized and more soluble form. This is consistent with the principles described in [7], where the pH of the medium and ionic strength is referred to as key factors in sorption control. It can also be noted that in many studies and in the regulatory documentation for aprepitant, an acidified mobile phase in HPLC is used.

CONCLUSIONS

The purpose of the study has been achieved as critical factors of sample preparation affecting the validity of aprepitant analysis were identified.
The main mechanism of aprepitant loss in neutral aqueous media is not only adsorption, but also mechanically and superficially induced membrane nucleation.
As far as the hypothesis for future research goes, we suggest that filtration stress testing (comparing centrifugation and filtration under different conditions) can be a component in validating analytical techniques for low-solubility drugs.
Practical use: In practice, priority should be given to centrifugation. If filtration is unavoidable, it is necessary to validate not only the filter material, but also the pre-saturation volume, considering the filter as a consumable with a certain “distortion effect” (DEF). The most effective method of sample stabilization is to transfer the substance to a true solution by protonation (using an acidic medium), which eliminates the influence of the filter material.