Structure-dependent retention of steroid hormones by common laboratory materials

Abstract

The tendency of steroid molecules to adsorb to various materials, particularly plastics, has been known of for decades but has not received widespread attention in the scientific community, and a modern, systematic study is lacking. This adsorption is an important consideration for researchers working with steroid hormones as it could skew the results of various experiments. Here we show that steroids adsorb to various vessels used in experiments, including microcentrifuge tubes, glass vials, and cell culture plates, in a manner that depends on the steroid’s molecular structure and on the type of vessel. The lipophilicity of steroids is a strong predictor of the degree of adsorption, with nearly 50% of the most lipophilic steroid tested, pregnenolone, retained in a high-adsorbing microcentrifuge tube after one hour incubation of an aqueous pregnenolone solution followed by removal of the aqueous solvent. We also show the effects of other factors such as incubation time, centrifugation, and temperature on adsorption, and show that adsorption can be mostly prevented by the presence of serum proteins in steroid solutions and/or by the use of low-adsorbing tubes.

1. Introduction

The tendency of steroid hormone molecules to adsorb to laboratory materials is an important consideration for researchers working with these compounds, as such adsorption could potentially skew the results of various experiments. Several studies spanning the decades from the 1950s to the present era have reported on these tendencies [1–7], but generally not in a systematic manner, and these findings have not received widespread attention in the scientific community. Probably the best known study was published by Bruning et al. in 1981 [4] and has received approximately one citation per year since. We recently showed that steroids accumulate at heightened concentrations in cells and that the lipophilicity of a steroid, measured by its log octanol-water partition coefficient (log K_ow), predicts the degree of accumulation [8], raising the question of whether a similar relationship exists for steroid adsorption to surfaces. We therefore set out to test the effects of various factors, including steroid molecular structure, type of vessel containing the steroid solution, time of exposure, type of solution containing the steroid, and temperature on steroid adsorption, and to provide guidance on designing experiments to minimize potential errors due to this adsorption.

2. Materials and Methods

2.1. Steroids

All experiments utilized [³H]-labeled steroids (PerkinElmer, Waltham, MA) at ~100,000 cpm. The following steroids (with specific radioactivities) were used: [7-³H(N)]-pregnenolone (12.7 Ci/mmol), [1,2,6,7-³H(N)]-progesterone (93.4 Ci/mmol), [1,2,6,7-³H(N)]-dehydroepiandrosterone (76.1 Ci/mmol), [1β-³H(N)]-androst-4-ene-3,17-dione (26.5 Ci/mmol), [1,2,6,7-³H(N)]-testosterone (95.5 Ci/mmol), [1,2,4,5,6,7-³H(N)]-dihydrotestosterone (93.3 Ci/mmol), [2,4,6,7-³H(N)]-estradiol (94.0 Ci/mmol), [2,4,6,7-³H(N)]-estrone (84.4 Ci/mmol), and [1,2,6,7-³H(N)]-hydrocortisone (cortisol) (78.3 Ci/mmol). [³H]-labeled steroids were not purified. Because different [³H]-labeled steroids have different cpm/mol, molar concentrations ranged from approximately 2 to 20 nM; our results indicated that varying molar concentration well beyond this range did not substantially affect the fraction of steroid adsorbed (see Results section 3.4).

2.2. Cells

LNCaP prostate cancer cells were purchased from the American Type Culture Collection (Manassas, VA) and maintained in RPMI-1640 with 10% fetal bovine serum (Gemini, West Sacramento, CA). HMC-1–8 breast cancer cells were provided by M. Abazeed and maintained in RPMI-1640 with 10% FBS.

2.3. Steroid adsorption

Unless indicated otherwise, all experiments were performed with [³H]-labeled steroids in 200 μL Dulbecco’s PBS (D-PBS) incubated for one hour at room temperature (22°C) in 1.5 mL Fisherbrand Premium microcentrifuge tubes (catalog number 05–408-129, Thermo Fisher Scientific, Waltham, MA). After incubation, D-PBS was pipetted into scintillation vials containing 1.5 mL Liquiscint scintillation fluid (National Diagnostics, Atlanta, GA) and samples were vortexed. Caps of microcentrifuge tubes were then removed and empty microcentrifuge tubes were placed into scintillation vials and filled with 1.5 mL scintillation fluid. Total radioactive signal in each sample was assayed by LS 6000IC scintillation counter (Beckman Coulter, Brea, CA).

Experiments testing adsorption in different vessels additionally utilized three other types of 1.5 mL microcentrifuge tubes: Costar low binding microcentrifuge tubes (catalog number 3207, Corning Inc., Corning, NY), Eppendorf DNA LoBind tubes (catalog number 022431021, Eppendorf, Hamburg, Germany), and Axygen Maxymum Recovery tubes (catalog number MCT-150-L-C, Corning Inc.). Two other types of vessel were also used: Eppendorf 24-well cell culture plates (catalog number 0030722116, Eppendorf) and Xpertek 2 mL clear glass vials (33 expansion borosilicate glass, catalog number 954002, P.J. Cobert Associates, St. Louis, MO). In experiments with cell culture plates, steroids retained in plates after removal of D-PBS were collected using 200 μL 1:1 ethyl acetate:isooctane (in tests with microcentrifuge tubes, extraction of steroids from empty tubes with 1:1 ethyl acetate:isooctane resulted in no steroids left behind in the tubes), which was pipetted into scintillation vials containing scintillation fluid.

For time course experiments, 500 μL D-PBS with [³H]-labeled DHEA or pregnenolone were incubated in 1.5 mL Fisherbrand microcentrifuge tubes or in wells of Eppendorf 24-well cell culture plates, with separate tubes/wells having samples collected after 5 minute, 30 minute, 2 hour, 7 hour, and 24 hour incubations. A one minute, 6000 g centrifugation of tubes on a 5424R centrifuge (Eppendorf) was separately performed. For the cell culture plates, steroids retained in plates after removal of D-PBS were collected using 500 μL 1:1 ethyl acetate:isooctane.

For experiments testing adsorption from different solutions, [³H]-labeled DHEA or pregnenolone in 200 μL UltraPure distilled water (Invitrogen, Carlsbad, CA), D-PBS, RPMI-1640 media, or RPMI 1640 with 10% charcoal-stripped fetal bovine serum (Gemini) was incubated in 1.5 mL Fisherbrand microcentrifuge tubes.

For experiments testing effects of concentration on adsorption, [³H]-labeled DHEA or pregnenolone to total concentrations of 2, 20, and 200 nM was added in 500 μL D-PBS along with additional ethanol so the total ethanol content was the same across the different concentrations of DHEA or pregnenolone.

For the experiment testing adsorption in the presence of cells, LNCaP cells or HMC-1–8 (~1,500,000 cells per tube) were aliquoted into 1.5 mL Fisherbrand microcentrifuge tubes in 1 mL serum-free RPMI-1640 and treated with [³H]-labeled pregnenolone, then incubated for 1 hour at 37°C on a tube rotator. Cells were then pelleted by 5 min. 250 g centrifugation in an Allegra 6R centrifuge (Beckman-Coulter). Media were collected, cell pellets were collected in 800 μL D-PBS, and total counts of [³H]-labeled pregnenolone for the media, cell pellet samples, and empty tubes were assayed in the same way as in other experiments. In parallel, tubes containing media without cells underwent the same procedure, including a simulated cell pellet collection. For the cell pellet samples, scintillation counting was performed on a cell suspension. Lysis of the cells was not performed in the experiments from which data were used, but in a test experiment in which cell samples obtained in parallel were added to 1% Triton X-100 (Sigma-Aldrich, St. Louis, MO) to lyse the cells we observed that this did not affect the results.

2.4. Chromatographic lipophilicity parameter determination

[³H]-labeled steroids were added to 50% methanol and injected on a Waters 1525 HPLC system (Waters Corp., Milford, MA). Steroids were separated on a Luna 150 × 4.6 mm, 3.0 μm particle size C₁₈ reverse-phase column (Phenomenex, Torrance, CA) with a methanol/water gradient at 50°C. Column effluent was mixed with Liquiscint scintillation cocktail and analyzed using a β-RAM model 4 in-line radioactivity detector (LabLogic, Brandon, FL).

2.5. Quenching corrections

To determine whether quenching of signals during scintillation counting could skew results, we measured the total signal from pregnenolone and DHEA standards (each in triplicate for each condition) added directly to scintillation cocktail, to which was also added either no additional solvent or 200 μL distilled water, D-PBS, serum-free RPMI-1640, RPMI-1640 with 10% CSS, or 1:1 isooctane-ethyl acetate. The results indicated that 1:1 isooctance-ethyl acetate did not cause quenching but each of the other solvents did by averaged factors of 0.89 for water, 0.80 for PBS, 0.83 for serum-free media, and 0.74 for 10% CSS media. Therefore, all results were adjusted using the appropriate correction factors to account for this quenching. We also tested whether the presence of microcentrifuge tubes or glass vials caused quenching during scintillation counting and found that it did not.

3. Results

3.1. Effects of steroid lipophilicity and molecular structure and of type of vessel on steroid adsorption to vessels

To test the effects of steroid molecular structure and lipophilicity on adsorption, we used nine different [³H]-labeled steroids: pregnenolone, progesterone, dehydroepiandrosterone (DHEA), Δ⁴-androstenedione (AD), testosterone, dihydrotestosterone (DHT), estradiol, estrone, and cortisol, covering a broad range of lipophilicities and all having log K_ows that were experimentally determined in a single study [9] (Fig. 1A). Because steroid adsorption to microcentrifuge tubes, both when steroid solutions are prepared for treatment of cells and when samples containing steroids are collected, could be a confounding factor in numerous experimental designs, we tested adsorption to four different types of polypropylene microcentrifuge tubes (Fisher, Costar, Eppendorf, and Axygen; see Materials and Methods for specific product details), three of which (Costar, Eppendorf, and Axygen) are specifically marketed as providing low binding or low retention of samples. We also tested adsorption to glass vials and to tissue culture (TC) treated polystyrene plates. After one hour incubations of steroids dissolved in Dulbecco’s phosphate-buffered saline (D-PBS), each of the steroids displayed some degree of adsorption to each of the vessels, but the fraction of steroid adsorbed varied enormously based on both steroid structure and type of vessel. The most striking result was that for the Fisher microcentrifuge tubes, pregnenolone, the most lipophilic of the tested steroids, displayed almost 50% adsorption – that is, after completely removing the aqueous solution of pregnenolone from the tubes almost 50% of the pregnenolone remained in the tubes. In each of the different vessels, pregnenolone and progesterone, the two pregnanes and the two most lipophilic steroids, were adsorbed more than the other steroids. The other three types of microcentrifuge tubes, the glass vials, and the TC plates all displayed less of a tendency to adsorb steroids, but all still adsorbed a very considerable fraction of pregnane molecules with the exception of the Costar microcentrifuge tubes, which limited adsorption to about 9% (for pregnenolone) or less (Fig. 1B). Plotting the logs of the steroid fractions retained by the vessels against the log K_ows of the steroids revealed that for each of the different vessels there was a positive correlation between lipophilicity and adsorption to vessels, although the strength of the correlation varied across the different vessels, ranging from an R² of 0.51 for the Eppendorf microcentrifuge tubes to an R² of 0.80 for the Costar microcentrifuge tubes (Fig. 1C). Although lipophilicity was generally a strong predictor of adsorption, there were discrepancies, such as, for example, progesterone adsorbing more than the higher log K_ow pregnenolone in Axygen tubes and estradiol adsorbing less than the lower log K_ow AD in several different vessels, suggesting that other specific aspects of molecular structure beyond lipophilicity could affect tendencies to adsorb to certain surfaces.

Figure 1. Steroids adsorb to different vessels in a manner that depends on type of vessel and on steroid lipophilicity.

(A) List of nine steroids used in these experiments along with experimentally determined log octanol-water partition coefficients (log K_ows) obtained from Leszczynski and Schafer [9]. (B) Fraction of steroid retained after one hour incubation in D-PBS solution in four different types of microcentrifuge tubes as well as tissue culture plates and glass vials (see Materials and Methods for details of different vessels) for the nine steroids. All experiments were performed twice in duplicate; plots show each individual value with lines representing mean values. (C) Graphs of logs of fraction of steroid retained vs. log K_ows for the nine steroids and six different vessels. Mean values ± standard deviations are shown. For all graphs, slopes of trendlines are different from zero (p < 0.0001).

An alternate method for determining lipophilicity is the use of reverse-phase high performance liquid chromatography (RP-HPLC) [10, 11], in which more lipophilic compounds tend to have larger retention times. The retention factor (k) of a compound is given by k = (t_r − t₀)/t₀ where t_r is the retention time of the compound and t₀ is the retention time of an unretained compound, or “dead time.” Using RP-HPLC, we measured the retention times of the nine steroids and calculated the retention factors (Fig. 2A), then plotted the logs of the steroid fractions retained by the different vessels against the logs of the retention factors (Fig. 2B). This confirmed the trend of more lipophilic steroids being retained more by each of the different vessels, and the chromatographically determined lipophilicity parameters in fact tended to be better predictors of steroid adsorption than the octanol-water partition coefficients, with R² values ranging from 0.76 to 0.91 for the chromatographic parameters vs. from 0.51 to 0.80 for the octanol-water coefficients.

Figure 2. Chromatographic lipophilicity parameters confirm the relationship between steroid lipophilicity and adsorption.

(A) List of nine steroids with retention times (t_r) obtained using reverse phase high performance liquid chromatography (RP-HPLC; values are averages of two runs of all nine steroids), retention factors k = (t_r − t₀)/t₀, and logs of retention factors. (B) Graphs of logs of fraction of steroid retained vs. log k for the nine steroids and six different vessels. Mean values ± standard deviations are shown. For all graphs, slopes of trendlines are different from zero (p < 0.0001).

We also examined the strength of the correlations between five different computationally predicted versions of log K_ow each obtained from the EPA’s online Chemistry Dashboard [12] (Table 1). Of the five, the NICEATM predictions showed the strongest correlations for each of the different vessels. Interestingly, for most of the predictive models, as well as for the experimentally determined log K_ows (Fig. 1C), the correlations for the Eppendorf tubes were weaker than for any of the other vessels.

Table 1. Correlations between different predicted log K_ow values and steroid adsorption to different vessels.

R² values from correlations between five different computationally predicted forms of log K_ow and log of steroid retained for the six different vessels, along with the means ± standard deviations of the R² values for each predictive model.

	EPISUITE	NICEATM	ACD/Labs Consensus	ACD/Labs	OPERA
Fisher tubes	0.55	0.92	0.78	0.69	0.0
Costar tubes	0.62	0.77	0.76	0.71	0.71
Eppendorf tubes	0.24	0.79	0.50	0.37	0.39
Axygen tubes	0.53	0.78	0.64	0.61	0.63
cell culture plates	0.40	0.83	0.62	0.52	0.54
glass vials	0.42	0.88	.63	0.54	0.55
mean ± SD	0.46 ± 0.14	0.83 ± 0.06	0.65 ± 0.10	0.57 ± 0.12	0.59 ± 0.12

3.2. Effects of incubation time and centrifugation on adsorption

After determining these relationships using the set of nine steroids, we then set out to determine the effects of other factors on steroid adsorption in a set of experiments using two steroids, pregnenolone and DHEA. We first tested the time course of steroid adsorption to microcentrifuge tubes and to TC plates using incubation times ranging from 5 minutes to 24 hours, as well as the effect of centrifugation on adsorption. Fisher microcentrifuge tubes were used for these and all subsequent experiments because they were the highest adsorbing tubes and therefore would make it easier to see changes in adsorption resulting from changes in other experimental conditions. In microcentrifuge tubes, the amount of steroid adsorbed continued to increase, in a saturating manner, for the full 24 hour time course (Fig. 3A). A brief (one minute) centrifugation of the tubes resulted in a level of adsorption that would result from a longer incubation with no centrifugation (Fig. 3B). In TC plates, the amount of steroid adsorbed appeared to reach its maximal level more rapidly than in microcentrifuge tubes (Fig. 3C).

Figure 3. Steroid adsorption increases with incubation time and with centrifugation.

(A) Fraction of steroid retained in Fisher microcentrifuge tubes after incubation times of 5 minutes, 30 minutes, 2 hours, 7 hours, and 24 hours for two steroids, pregnenolone and DHEA. Experiment was performed twice in duplicate and points on graph show mean ± SD. (B) Fraction of steroid retained in Fisher microcentrifuge tubes after 1 min. 6000 g centrifugation. Experiment was performed twice in triplicate and graph shows mean ± SD. (C) Fraction of steroid retained in TC plates after the same incubation times as in (A). Experiment was performed twice in duplicate and points on graph show mean ± SD.

3.3. Effects of type of solution on adsorption

We next tested the effect of the type of solution in which the steroids were dissolved on the amount of adsorption to microcentrifuge tubes and TC plates. Whereas all other experiments used steroids dissolved in D-PBS, in this experiment we additionally tested solutions of distilled water, serum-free cell culture media, and cell culture media with 10% charcoal-stripped fetal bovine serum. The results showed that steroids dissolved in water adsorbed to a similar degree as steroids dissolved in D-PBS. Binding of steroids to serum proteins in the media with 10% serum eliminated a large majority of the adsorption, whereas the adsorption from serum-free media was at a level in between that from D-PBS or water and that from media with 10% serum (Fig. 4). When steroids were dissolved in solutions of 50% methanol, no adsorption was observed (data not shown).

Figure 4. Serum proteins prevent most of the steroid adsorption.

Fraction of pregnenolone (A) or DHEA (B) retained in Fisher microcentrifuge tubes, and fraction of pregnenolone (C) or DHEA (D) retained in TC plates, after one hour incubation in distilled water, D-PBS, serum-free culture media, or culture media with 10% charcoal-stripped fetal bovine serum. Experiments were performed twice in duplicate; plots show each individual value with lines representing mean values. For pregnenolone in tubes, retention from 10% CSS media was less than from all other solutions and retention from serum-free media was less than from distilled water or D-PBS (p < 0.001 for all comparisons, Tukey’s multiple comparison test after one-way ANOVA). For DHEA in tubes, retention from 10% CSS media was less than from all other solutions (p < 0.001 for 10% CSS media vs. distilled water or D-PBS; p = 0.004 for 10% CSS media vs. serum-free media) and retention from serum-free media was less than from distilled water (p = 0.005) or D-PBS (p = 0.002). For pregnenolone in TC plates, retention from 10% CSS media was less than from all other solutions and retention from serum-free media was less than from distilled water or D-PBS (p < 0.001 for all comparisons). For DHEA in TC plates, retention from 10% CSS media was less than from all other solutions (p < 0.001 for all comparisons) and retention from serum-free media was less than from distilled water (p = 0.01) or D-PBS (p = 0.003).

3.4. Effects of steroid concentration and temperature on adsorption

To examine whether the concentration of a steroid affects the fraction of that steroid that adsorbs to a vessel, we tested 2, 20, and 200 nM concentrations of [³H]-pregnenolone and [³H]-DHEA. The results (Fig. 5) were not indicative of an effect of steroid concentration on fractional adsorption.

Figure 5. Steroid concentration appears not to affect fraction adsorbed.

Fraction of pregnenolone or DHEA retained in Fisher microcentrifuge tubes after 1 hour incubation in D-PBS at concentrations of 2, 20, and 200 nM for each steroid. Experiment was performed twice in duplicate and graph shows mean ± SD. For pregnenolone across different concentrations, p = 0.43 from one-way ANOVA. For DHEA across different concentrations, p = 0.31 from one-way ANOVA.

We also tested the effect of temperature on adsorption to microcentrifuge tubes, using incubations at three temperatures: 4°C, 22°C (room temperature), and 37°C (all previous experiments having been performed at room temperature). The results indicated that the tendency of steroids to adsorb was increased by lower temperature – at 4°C, nearly two-thirds of the pregnenolone adsorbed to the tubes during a one hour incubation (Fig. 6). This may be due to decreased Brownian motion at lower temperatures; it does not appear to be due to changes in steroid solubility as no precipitate was observed.

Figure 6. More steroid adsorption occurs at lower temperatures.

Fraction of pregnenolone or DHEA retained in Fisher microcentrifuge tubes after 1 hour incubation in D-PBS at 4°C, 22°C, and 37°C. Experiment was performed twice in triplicate and graph shows mean ± SD. For both pregnenolone and DHEA, retention at 4°C was greater than at 22°C or 37°C (p < 0.001 for all comparisons, Tukey’s multiple comparison test after one-way ANOVA).

3.5. Competition between adsorption and cellular steroid uptake

The finding that various different laboratory materials can adsorb large amounts of lipophilic steroids raises an important question for the design and interpretation of experiments: if cells are incubated with a steroid in a vessel that can adsorb, for example, 40% of the molecules of that steroid, does that mean that the amount of steroid available to the cells is decreased by 40%, or does the cellular tendency to accumulate steroids[8] compete with the steroid adsorption to the vessel? To address this question, we separately incubated LNCaP prostate cancer cells and HMC-1–8 breast cancer cells in serum-free media with pregnenolone in microcentrifuge tubes, pelleted the cells, and measured the amount of pregnenolone in the media, cells, and empty tubes. In parallel, we performed the same procedure in tubes that contained media without cells, including a simulated pellet collection (the cell pellets were collected by quickly pipetting PBS down into the tubes and then pipetting up the PBS and cells). The results showed that the fraction of total pregnenolone remaining in the empty tubes was about 27% when no cells were present but only about 5% (LNCaP) or 9% (HMC-1–8) when cells were present (Fig. 7), indicating that the cellular affinity for steroids outcompetes most of the adsorption to plastic.

Figure 7. Cellular affinity for a steroid outcompetes steroid adsorption to plastic.

Fraction of pregnenolone collected in media, cell pellet or simulated pellet collection, and empty tube for one hour incubations of serum-free media with pregnenolone in Fisher microcentrifuge tubes with and without LNCaP prostate cancer cells (A) or HMC-1–8 breast cancer cells (B). Experiment was performed twice in triplicate (A) or three times in triplicate (B) and graphs shows mean ± SD. Retention in empty tubes was greater without cells than with cells for both cell lines (t-tests, p < 0.001).

4. Discussion

In this study, we have demonstrated that steroids have a strong tendency to adsorb to various common laboratory materials in a manner that depends on the steroid molecular structure, with more lipophilic steroids generally tending to adsorb more than less lipophilic steroids, thus both confirming and extending prior studies [1–7]. We have shown that the type of vessel used to contain a steroid solution can strongly affect the amount of adsorption, even among different types of polypropylene microcentrifuge tubes, with the highest adsorbing tube retaining nearly 50% of pregnenolone molecules and the lowest adsorbing tube retaining just under 10%. We have also examined other factors affecting the amount of adsorption, such as incubation time or centrifugation, type of solution, steroid concentration, and temperature. Notably, longer incubation times or centrifugation both increase adsorption, the presence of serum proteins in a solution eliminates most of the adsorption, changes in concentration over the 2 to 200 nM range we tested do not substantially affect the fractional adsorption, and when looking at the effect of temperature, lower temperatures result in more adsorption. Lastly, we showed that when cells are present in a vessel, steroid adsorption to the vessel is greatly decreased, suggesting that steroids that would adsorb to the surface of a vessel during a control, no-cell experiment would in fact still be available to the cells if cells were present.

Although our study is not the first to report on steroid adsorption to laboratory materials, it expands on previous studies in several key ways. Most previous studies employed much more limited sets of steroids. One previous study [4] examined a similarly broad set of steroids to ours, but reported on the adsorption of steroids to different types of plastic tubing. Our study, by assaying adsorption to different types of laboratory vessels such as microcentrifuge tubes, borosilicate glass vials, and cell culture plates, and by examining the effects of a range of solvents including D-PBS, serum-free culture media, and culture media with 10% serum, is more directly applicable to experimental conditions commonly encountered in laboratory research on steroid action. Additionally, whereas some previous studies commented in a descriptive sense that less polar steroids had greater adsorptive tendencies [3, 4], ours is the first, to our knowledge, to quantify the specific relationship between adsorption and lipophilicity (Fig. 1C, Fig. 2C). Our recent demonstration of a similar relationship between steroid lipophilicity and passive accumulation in cells [8] provides an important additional context to these studies, and our experiment showing that cellular affinity for a steroid outcompetes most of the adsorption to plastic (Fig. 7) brings together these two lines of research and is another important new insight from our study in comparison to previous studies on steroid adsorption.

Our findings have important implications for the design and interpretation of experiments involving steroids. If a researcher treating cells with a highly lipophilic steroid prepares, in a high-adsorbing microcentrifuge tube, an aqueous steroid solution from which to treat the cells, the actual steroid concentration to which the cells are exposed could be markedly lower than the intended concentration. To avoid this, the solution could be prepared with 10% serum, or if this is not appropriate for the experiment, the solution should be prepared in a low-adsorbing tube. Additionally, if samples are collected (for example, of serum-free culture media) of which the steroid content is to be analyzed, unnecessary additional transfers of samples from one vessel to another should be avoided, because some fraction of steroid molecules would be lost during each transfer. As a general rule, in any experimental scenario in which adsorption of steroids could skew results, it is a good idea to directly test the amount of adsorption that occurs, as well as to use internal standards to control for the potential effects of adsorption. Lastly, although our study focused on steroid hormones, the results are likely applicable to various other lipophilic substances, and are consistent with studies on adsorption to plasticware of lipophilic drugs [13] as well as of transport of environmental contaminants by plastics[14].

Highlights.

• Steroids have strong tendencies to adsorb to common laboratory materials.

• More lipophilic steroids tend to adsorb more strongly.

• This could skew the results of various experiments.

• Guidance is provided on ways to minimize this adsorption.