Standard Water Generating Vials for Lipophilic Compounds

Abstract

The study of non-polar compounds in aqueous environments has always been challenging due to their poor solubility in aqueous media. The low affinity of non-polar compounds toward polar solutions facilitates their attachment to glassware, which results in unstable sample concentrations. To address this challenge, and to enable the preparation of a stable mixture of hydrophobic compounds in an aquatic environment, we introduce an in-vial standard water generating system consisting of a vial containing appropriate aqueous solution and a polydimethylsiloxane thin film spiked with target compounds. In this system, a solution with a stable analyte concentration is attained once equilibrium between the thin-film and aqueous solution has been achieved. The developed standard water system was studied using endocannabinoids and phospholipids as model hydrophobic compounds of biological importance, with results indicating that the concentration of hydrophobic compounds in water can remain stable over multiple days. The results also showed that analytes released from the thin film can compensate for analyte loss due to extractions with solid-phase micro extraction fibers, thereby re-establishing equilibrium. Thus, the vial is suitable for the repeatable generation of non-polar standards for routine analysis and quality control. The results of this work show that the developed system is stable and reproducible and therefore appropriate for studies requiring the measurement of free concentrations and accurate quantification.

Graphical Abstract

Hydrophobic compounds are major components in foods, environmental pollutants, biological samples, and pharmaceuticals. Lipids are a particularly prevalent group of hydrophobic compounds that play an important role in biological functions such as energy storage, forming cellular membrane matrices, and cellular signaling. Lipids are also involved in a wide variety of biological processes due to their structural diversity and physicochemical properties. Glycerophospholipids (also referred to as phospholipids) are present in virtually all living organisms, typically as constituents of cellular membranes and in regulatory functions—directly and indirectly—through their metabolites. Dipalmitoylphosphatidylcholine (DPPC) is one of the most abundant phospholipids in mammalian pulmonary surfactants, and a staple in surfactant and model membrane studies.^1-3 Its metabolite, lyso-palmitoylphosphatidylcholine (LPC, LPC 16:0), is also quite abundant in biological samples, and it has recently been identified as a possible plasma biomarker (among other molecules) for obesity, hepatocellular carcinoma, and colorectal cancer in metabolomics studies.^4-6 Similarly, endocannabinoids (ECs) have been the focus of numerous recent studies due to their influence in food intake,⁷ energy balance, communication between cells, appetite and metabolism, memory, intestinal motility,⁸ and immune responses.⁹ The most well-known ECs are N-arachidonoylethanolamide (anandamide; AEA) and 2-arachidonoylglycerol (2-AG), which were discovered in 1992 and 1995, respectively.^10-12 1-Arachidonoyl glycerol (1-AG) is an isomer of 2-AG that occurs physiologically but is not active on cannabinoid receptors.¹³

The analysis of hydrophobic compounds in biological matrices is tedious due to their structural diversity, wide range of biological concentrations and isomeric species, and diverse physicochemical stabilities. The analysis of ECs is complicated by spontaneous isomerization between 1-AG and 2-AG, as well as trace concentration levels and volume limitations¹⁴ for biological samples. Although the literature contains a few studies showing that AEA is quite stable at 4 °C in distilled water in the presence of albumin and/or hydroxypropyl-β-cyclodextrine (HPCD),¹⁵ its stability in aqueous media remains largely unexamined.

One major issue in studies of hydrophobic compounds is their low solubility and stability in aqueous media, which has hampered their accurate quantification with a wide range of techniques, including microextraction. In addition, hydrophobic chemicals easily adsorb onto commonly used plastic and glassware, with adsorption rates being higher for plastic surfaces compared to glass surfaces.^15,16 Therefore, the adsorption of these compounds onto plastic or glassware in the laboratory could alter the solution concentration and decrease recovery during experiments. For example, due to the comprehensive adhesion of hydrophobic compounds to plastic and glass surfaces in water, only a small percentage of the spiked compounds is measurable. Therefore, the use of silanized or siliconized vials is recommended for the transfer, extraction, and storage of hydrophobic compounds, including lipids, proteins, and ECs.^16-18 Ideally, lab equipment should be made from stable materials that minimize compound degradation and interactions between the molecule and the inner walls of the vial. In this respect, silica (SiO₂)-coated plastic vials are especially useful for minimizing adsorption events.¹⁹ Indeed, the use of silicones to siliconize container surfaces in the production of pharmaceuticals has been practiced for many years, with this technique first being patented by Goldman in 1949.²⁰ Since then, siliconization and thin polymer films have been investigated as coatings for the walls of glass vials to minimize non-specific adsorption for certain drug groups.¹⁶ Despite these efforts, the adsorption of hydrophobic compounds to the walls of glass vials has still not been completely solved, resulting in the persistence of low reproducibility in sample concentration.^16,19-21

This work presents a new standard water generating system that uses carbon mesh-supported thin-film membranes to produce stable aqueous solutions of hydrophobic compounds with reasonable concentrations. The thin film used in this system consists of a carbon mesh coated with a layer of hydrophobic glue-like polymers, such as polydimethylsiloxane (PDMS). Submerging the PDMS-coated thin film in the aquatic sample allows the non-polar compounds to diffuse from the film to the aqueous solution. Although a similar idea has been proposed for the preparation of gas mixtures,^22,23 the application of thin films for the preparation of water generating systems has not been reported to date. The proposed thin-film-based water vial system could solve the stability and sticking problems of hydrophobic compounds in water, as it is capable of producing stable, reproducible, and reusable standard mixtures of lipophilic compounds in water. Three ECs and two phospholipids were analyzed to demonstrate the applicability of the proposed system. However, because the phospholipids and ECs required different experimental conditions, they were analyzed via separate studies.

MATERIALS AND METHODS

A complete list of material, methods, and instruments is provided in the Supporting Information.

Silanization Procedure for Plastic and Glassware.

To prevent the adsorption of hydrophobic compounds, all lab equipment that could potentially come into contact with the analytes was silanized. To this end, plastic pipette tips, caps, and glass insert tubes were immersed in Aquasil solution (diluted 1% with water) for 30 min, rinsed with methanol and distilled water, respectively, and dried prior to use.

Preparation of Water Vial System.

First, the PDMS thin films was prepared according to the procedure described in the Supporting Information. Next, the thin film was manually cut into different sizes and the pieces were placed inside the vials to cover the glass walls.

The analytes were homogenously spiked onto the thin-film side that would become the inner leaflet and sample solution ([i.e., liquid chromatography–mass spectrometry (LC–MS-grade water or PBS buffer] was then added to the vial. Phospholipids were spiked on the thin film before placing it inside the vial, whereas ECs were spiked directly to the thin film while already inside the vial. Once the solution had been added, the vial was sealed with a silanized cap, briefly sonicated for 20 s, and then agitated at 1000 rpm until equilibrium between the thin film and sample was established. The equilibrated sample solution was subsequently used for the stability investigations. The size of the thin film used for each vial was dependent on the vial’s dimensions. For example, a 3.5 × 3.5 cm thin film was used for the 4 mL silanized vial and a 4 × 7 cm thin film was used for the 20 mL vial. To study the sample composition, solid-phase microextraction (SPME) was used in each of the studies. Details regarding the procedure used to prepare the SPME fibers and the subsequent extraction procedure can be found in the Supporting Information.

Equilibrium Time between Thin-Film and Sample.

To study the time required to establish an equilibrium between the sample containing ECs and the thin film, 3 separate vials were prepared by manually cutting thin films into 3.5 × 3.5 cm pieces and placing them into the silanized vials (4 mL) such that they covered the entire surface area of the vial’s glass wall. To obtain a final concentration of 10 μg/mL, appropriate amounts of 2-AG were spiked directly onto the thin film and 4 mL water was added to the vials. The vials were then sealed with a silanized cap and agitated at 1000 rpm until an equilibrium had been achieved. To determine the equilibrium time and to study the sample concentration, several samples were obtained over multiple days for direct injection into the LC instrument, with SPME (SPME fiber) being applied to analyze the stability of the vials. The repeatability and reproducibility of the different vials were assessed by repeating 3 sequential extractions under agitation at 1000 rpm for 2 h. After each extraction, the SPME-HLB fibers were washed for 5 s in 300 μL MS-grade water, followed by desorption with 300 μL ACN/MeOH for 1 h with agitation at 1000 rpm. The same procedure was repeated multiple times over 7 days.

To determine the equilibrium time for the aqueous media and the phospholipid-enriched PDMS thin films, three independent vials were prepared by spiking either 2.5 μg of LPC or 125 ng of DPPC onto 4 × 7 cm PDMS thin films to obtain total concentrations of 125 and 6.25 ng/mL, respectively. After the addition of 20 mL of water, the vials were briefly sonicated and left to equilibrate under magnetic agitation at 1000 rpm. A small aliquot was taken daily over a five-day period, treated, and stored at −80 °C prior to liquid chromatography tandem mass spectrometry (LC-MS/MS) analysis.²⁴ For vials containing PBS, a liquid–liquid extraction step was performed to avoid the introduction of non-volatile phosphate salts into the MS. Briefly, this step consisted of adding 100 μL of MeOH/methyl-tert-butyl ether (30:70) to the aqueous aliquot, followed by agitation for 30 min at 1000 rpm and centrifugation for 10 min at 10,000 rpm. After performing a second liquid–liquid extraction step, both upper phases were removed, pooled together, vacuum evaporated for 10 min at 30 °C, and then reconstituted in mobile phase solution.

Relationship between Spiked Concentration on Thin-Film and Sample Concentration (Total vs Free Concentration).

The relationship between the amount of analyte spiked onto the thin film and the sample concentration depends on the analyte’s physiochemical properties and its tendency to absorb into the PDMS thin-film polymer. To study the relationship between the concentration of analytes spiked onto the thin-film and the actual concentration in the water sample, standard water generating vials were prepared. Here, 20 mL vials were spiked with 1, 5, and 10 μg/mL of AEA (C18:1), 10–1000 ng/mL of LPC, and 6.25–500 ng/mL of DPPC, with three vial replicates for each concentration being used in the phospholipid studies (i.e., LPC and DPPC). After spiking the analyte onto the thin film, the vials were filled with water and mixed under 1000 rpm agitation for multiple days until equilibrium had been achieved. Once equilibrium had been attained, an aliquot sample (water or PBS) was taken from each vial at different time intervals and injected directly into the LC–MS to calculate the sample concentration after equilibrium.

Free Concentration Recovery.

During SPME extraction, analytes adsorb onto the SPME fiber, which reduces the concentration of analytes in the sample and disrupts the equilibrium between the samples and thin films. To assess the analyte recovery rate under disrupted system equilibria, multiple aliquots were withdrawn at different time points, while a miniaturized C18-coated SPME fiber was exposed to aqueous portion. The aliquots were then subjected to liquid–liquid extraction (as explained before) and stored at −80 °C until LC–MS/MS analysis.

Reusability after Multiple Extractions.

One goal of this study was to prepare a water generating vial with high reusability over time, including after extractions. To this end, three different 20 mL water generating vials were prepared (5 μg/mL spiked concentration for AEA and 2-AG, 100 ng/mL for LPC, and 6.25 ng/mL for DPPC) and mixed under 1000 rpm agitation for 2 days until an equilibrium had been achieved. Each vial was analyzed using an SPME fiber, with an HLB fiber being applied for 10 min to extract the ECs and a C18 fiber being applied for 60 min to extract the phospholipids. In the case of the ECs, the sample concentrations before extraction and 5 min after extraction were studied via direct injection into a LC/MS instrument, whereas just the SPME extracts were monitored for the lipids.

Effect of Sample Matrix on the Distribution of Compounds.

The distribution constant, K, of the analytes between the matrix and the thin film depends on the nature of the compounds, the sample, and the thin film, as well as other parameters, such as the sample-to-thin-film volume ratio. To study whether variations in the sample composition can affect the distribution of analytes between phases (i.e., thin film and sample), water generating vials were prepared by spiking 8 μg/mL of AEA and 20 μg/mL of 2 AG onto PDMS thin-films that had been cut (3.5 × 3.5 cm) to fit a 4 mL vial. Once the thin films had been prepared, the vials were filled with MS-grade water or PBS (pH 7.4) and mixed under 1000 rpm agitation for 2 days until equilibrium had been achieved. Then, aliquots were injected into a LC instrument to determine the concentration of analytes in the prepared vials. To avoid injecting non-volatile phosphates into the ESI source, the first 2 min of the eluent from the column was directed to waste using the LC-instrument’s 6-port valve.

RESULTS AND DISCUSSION

Equilibrium Time/Repeatability of Different Vials.

To check the stability of each vial and the time required to establish an equilibrium between the analyte-loaded thin film and the sample, multiple samples were obtained from 3 vials spiked with 2-AG over a 7 day span, followed by direct injection into the instrument. The results are shown in Figures 1 and 2. The results indicate that the system achieved an equilibrium within 2 days for each vial containing ECs and within 3 days for the vials containing phospholipids. Despite the differences in the pathways to equilibrium concentration among the different vials, the equilibrium times, and concentrations (~30 ng/mL) are similar. The results of this test confirm the procedure’s repeatability and reveal that the equilibrium condition depends on thermodynamic constants (i.e., the K value of the analyte for distribution between the thin film and sample), regardless of the pathway to equilibrium.

Figure 1.

Concentration of water generating vial over 7 days obtained by direct injection of solution before and after extractions for vial #1–3 (Day# refers to the number of days after preparing the vials and spiking analytes onto the thin-films).

Figure 2.

Aqueous “free” concentration of phospholipids at different stages of equilibration after preparing the water vial for (A) LPC prepared in water at 125 ng/mL and (B) DPPC prepared in water at 6.25 ng/mL.

The other key finding of this test is that the concentration of ECs in aqueous media remains stable over multiple days. As the results show, the sample concentration remains stable for at least 7 days (and potentially more) after preparation, which makes it an appropriate tool for method development. Additionally, extractions were performed using SPME fibers to monitor the extracted amount and sample concentration (before and after extraction). As both the desorption signal (Figure S1) and sample concentration (Figure 1) show, the sample concentration remained stable, despite performing multiple extractions over a series of days.

Similarly, the findings showed that the sample concentration for phospholipids stabilized by the third day of daily monitoring for extractions carried out over a five-day span (Figure 2). Each marker in Figure 2 represents the average of three aliquots taken and analyzed from each of three independent vials. The highest variation was observed for the aliquots taken on the first day, which was likely because equilibrium conditions had yet to be achieved. Phosphatidylcholines and zwitterionic surfactants are known to form monolayers on water–air interfaces, which would result in hot spots of concentration within the vials.² Thus, a sonication step was implemented to avoid hot spots, with 15–20 s being enough to remove visible bubbles. Longer times were avoided to prevent cavitation damage on the PDMS thin film and excessive self-aggregation of the phospholipids.

Relationship between Spiked Concentration on Thin-Film and Aqueous Sample Concentration.

To further develop the proposed vial, it was important to figure out the relationship between the spiked amounts and the final aqueous sample concentration. Because the thin film and sample were in equilibrium, we expected to observe a linear relationship between the spiked amount and the sample concentration. However, increasing the sample concentration caused this line to plateau due to various reasons, such as sample saturation. For this study, 3 water generating vials were prepared by spiking the thin films with different amounts of AEA (C18:1). Initially, samples were taken from each vial for over 3 days. The goal in this study was to re-confirm the equilibrium time and determine whether varying the spiked amount changes the equilibrium time. The results of this experiment are provided in Figure S2 (the results are provided as the signal relative to the equilibrium signal for easier visualization). As the data suggests, while the vial spiked with higher amounts of analyte appears to require a few more hours to reach equilibrium, it can be assumed that 2 full days is enough for all the vials to reach equilibrium.

To find the relationship between the spiked amount and the sample concentration, the same vials described in the previous section were directly spiked and their concentrations were calculated. The equation describing from the relationship between the spiked amounts, and the final sample concentration is shown in Figure 3. As can be seen, in the studied concentration range, the relationship is linear with R² = 0.9979. This linear relationship is critical, as the preparation of a sample with a “known” and “stable” concentration is vital in analytical studies. This stability has been illustrated in a previous section, and this linear relationship helps the user to find the spiking amount required to obtain the final sample solution concentration.

Figure 3.

Concentration of AEA (C18:1) in a standard water generating vial measured by direct injection (the error bars are not visible due to their small range).

Unlike ECs, a linear relationship was not observed between the total amount spiked to the vial and the final aqueous concentration for phospholipids over the studied concentration range (Figure 4). Linear behavior is only attained in small spiking ranges for both LPC and DPPC with plateaus occurring in water samples at 8 and 0.3 ng/mL, respectively. The lack of linearity is likely influenced by the inherent surfactant properties of phospholipids, with self-association into micelles reducing the free concentration. Critical micelle concentration (CMC) for phospholipids with biological relevance falls in the 10⁻¹⁰–10⁻⁸ M range (CMC, 4.5 × 10⁻¹⁰ M)²⁵ with a clear dependence on the length and saturation of their hydrophobic tails; thus, virtually all phospholipids are above their respective CMCs at physiological concentrations. As expected, the presence of electrolytes in the solution caused an average decline of 35% in the solubility of LPC for samples containing PBS, with an ionic strength of 0.16 M, which agrees with other/previous similar works done on hydrophobic polycyclic aromatic hydrocarbons²⁶.

Figure 4.

Aqueous concentration of phospholipids at different total concentrations for LPC in water and buffered solutions.

Reusability after Multiple Extractions.

Based on the previous step, it was understood that the system requires 2–3 days to reach equilibrium. Therefore, the prepared vials were agitated for 2 days before being used in studies. The goal of this experiment was to study the reusability of the water vials after extractions. Once equilibrium had been achieved, the sample was obtained from the vials before performing extractions and 5 min after extractions with the SPME fiber. According to the results in Figures 5 and 6, the water vial system’s concentration remained stable, even after multiple extractions, and was able to re-establish equilibrium quickly after each extraction for the ECs. These results confirmed the reusability of the prepared vials for multiple extractions.

Figure 5.

Concentration of water generating vial by direct injection of i-th sample taken before and 5 min after extractions for (A) AEA and (B) 2-AG.

Figure 6.

Sequential SPME extractions to assess the repeatability of extractions from the prepared standard aqueous systems prepared using water and PBS for LPC (100 ng/mL total concentration) and DPPC (6.25 ng/mL total concentration). The intersampling interval was fixed to 5 min for extractions 1–7 and increased to 15 h between the 7th and 8th extractions.

The rate at which the PDMS reservoir replenishes the free concentration during an extraction with a miniaturized SPME fiber was assessed for the studied phospholipids (Figure 4). As can be seen, the concentration decreased rapidly within the first minute of extraction, which is consistent with the fast uptake typical of the extraction phase on the SPME fiber. However, the times required for the systems to replenish the initial concentration were rather different for each phospholipid—approximately 10 min for LPC and even longer for DPPC—despite having the same large surface area ratio between the lipid-loaded PDMS film and the miniaturized SPME fiber. The markedly slower recovery rate for DPPC was likely due to kinetically hindered desorption from the PDMS film, as lipid/lipid interactions in thin films heavily influence the desorption of hydrophobic compounds.²⁷ Regardless of the initial rate, both systems appeared to replenish the initial concentration after 60 min of extraction, which was exploited to assess the repeatability of the systems for either SPME extractions or direct measurement of the aqueous concentration (Figure 6).

In all cases, the effects of SPME extraction were negligible, accounting for only ~5% depletion of the free concentration. For example, SPME extraction accounted for only ~1.3% of the drop in phospholipids compared to the total amount spiked to the vial. The minimal depletion observed after eight extractions that up to 40 extractions can be performed before significant depletion of the system will occur. However, DPPC lacks good intraday repeatability, with significant depletion being observed after the fifth SPME extraction in a single day. This result is attributable to DPPC’s very high affinity toward the hydrophobic coating and its low water solubility, which results in a very slow desorption rate from the PDMS film and, consequently, long re-equilibration times. However, by adjusting the time between successive extractions (as observed by the eight-extraction step following overnight recovery), one could theoretically extract repeatable amounts of DPPC for up to 12 days prior to significant depletion of the total spiked amount. The markedly slower re-equilibration rate for DPPC is likely due to kinetically hindered desorption from the PDMS thin film, as lipid/lipid interactions in thin films heavily influence the desorption of hydrophobic compounds.

Effect of Sample Matrix on the Distribution of Compounds.

To examine how the sample matrix influences the distribution of analytes between the sample and thin film, two vials with similar conditions were prepared for water and PBS samples. The results from the direct injection of the PBS and water samples into a LC instrument are shown in Figure 7. As can be seen, the PBS samples had lower concentrations of both AEA and 2-AG compared to the water samples, which can be attributed to the salting-out effect. The presence of salts in PBS can reduce the solubility of hydrophobic compounds, which causes them to partition toward thin films to a higher degree. The matrix composition can affect the distribution of analytes differently depending on the nature of the sample and thin film and/or the physicochemical properties of the analytes. As the results of this study reveal, it is important to prepare water generating vials with a sample matrix that matches the conditions of the final experiment. If samples with different physiochemical properties are required, it is important to prepare new vials under new conditions to understand how they impact vial preparation.

Figure 7.

Direct injection from water and PBS for ECs.

CONCLUSIONS

The preparation of a stable sample solution in aquatic media using hydrophobic analytes has always been difficult due to the nature of these compounds, particularly challenging to maintain stable-free concentration of these analytes. In this study, a thin-film-based standard water-generating vial was designed and implemented for the preparation of a sample mixtures using EC and phospholipids, which are hydrophobic molecules with biological importance. The results from this study proved that the standard mixture prepared using this method remains stable at least up to 7 days. Furthermore, this work shows that the standard mixture can be reused multiple times due to the ability to re-establish the equilibrium between the thin-film and sample. The exception to this feature is cases wherein the sample depletion is so large that it cannot be replenished.

The linear relationship between the spiked amount and the concentration of ECs in the sample can be used to predict the concentration of the sample or to find the appropriate spiking amount to get the desired concentration. The findings also showed that the nature of the analyte, the sample, and how they interact with one another significantly influence the distribution of analytes between phases. Thus, the experimental conditions should be optimized to obtain accurate results.

In general, the proposed method is based on the equilibrium between the thin-film and the sample, which means the equilibrium time and the concentration can be adjusted by changing the analytes, the type of the thin film, or the sample matrix. The standard water-generating vial presented herein can be used with any other type of hydrophobic compound for the purpose of method development. The vial can be used as a reliable, stable, and reproducible source of hydrophobic compounds in aquatic media and can be applied as a tool for free concentration calculations in samples where the preparation of a stable solution is challenging.