Fig 1.
Cells derived from several tissue types accumulate steroids at high concentrations and uniformly have a preference for 3β-OH, Δ5-steroids over 3-keto, Δ4-steroids and progestogens over androgens.
(A) Schematic representation of cellular steroid uptake. In a model of uptake driven solely by passive diffusion and disregarding lipophilicity-driven bioaccumulation (top panel), concentrations of free steroids inside and outside cells would be equivalent. When lipophilicity-driven bioaccumulation is taken into account, steroid concentrations in cells or tissue can become greater than those in media or blood, and to a much greater extent for more lipophilic than for less lipophilic steroids. (B) A specific example of the concept schematically represented in (A): degree of steroid uptake depends on molecular structure; for example, progestogens (e.g. pregnenolone and progesterone) are more lipophilic and are preferentially taken up more than corresponding androgens (e.g. DHEA and AD) and 3β-OH, Δ5-steroids (e.g. pregnenolone and DHEA) are more lipophilic and are preferentially taken up more than corresponding 3-keto, Δ4-steroids (e.g. progesterone and AD). For reference, carbons 3, 4, 5, and 6 are labeled on the pregnenolone structure. Note the hydroxyl group at position 3 and double bond between carbons 5 and 6 on the 3β-OH, Δ5-steroids and the carbonyl group at position 3 and double bond between carbons 4 and 5 on the 3-keto, Δ4-steroids. Log Kow values, a measure of lipophilicity [25], are shown for the four steroids. (C-E) Ratios of cellular concentrations to original treatment concentrations in culture media for nine different steroids in (C) prostate cancer (LNCaP) cells, (D) placental choriocarcinoma (JEG-3) cells, and (E) breast cancer (HMC-1-8) cells. (F-H) Ratios of cellular concentrations to original treatment concentrations in culture media for four different steroids in (F) prostate cancer (LNCaP) cells, (G) placental choriocarcinoma (JEG-3) cells, and (H) breast cancer (HMC-1-8) cells. All graphs show mean ± SD from one representative experiment with biological duplicates and all experiments were performed at least twice. For all graphs, the uptake of pregnenolone was greater than all other steroids (p < 0.001, Tukey’s multiple comparison test after one-way ANOVA). Additionally, uptake of progesterone was greater than AD (p < 0.001) and uptake of DHEA was greater than AD (C: p = 0.01; D-E: p < 0.001).
Fig 2.
Cellular preferences for progestogens and 3β-OH, Δ5-steroids are independent of cell viability.
(A-B) Live (A) and heat-killed (B) LNCaP cells exposed to trypan blue stain. Note the staining of the heat-killed cells, and note also that the live and dead cells are similar from a gross morphological standpoint. (C-D) Ratios of cellular concentrations to original treatment concentrations in culture media for six different steroids in live (C) and dead (D) LNCaP cells incubated in tubes. (E-F) Ratios of cellular concentrations to original treatment concentrations in culture media for six different steroids in live (E) and dead (F) JEG-3 cells incubated in tubes. All graphs show mean ± SD from one representative experiment with biological duplicates and all experiments were performed at least twice.
Fig 3.
Live vs. dead cell uptake comparisons and uptake saturation both reveal contributions of active and passive accumulation.
Fractions of total steroids (media plus cells) collected in the cell samples when live and dead LNCaP cells were incubated with 2 nM [3H]-labeled steroids alone and with 1 μM unlabeled steroids. For pregnenolone, progesterone, and testosterone, uptake of 2 nM [3H]-labeled steroid with 0 unlabeled steroid in live cells was greater than uptake in any of the other three conditions (Tukey’s multiple comparison test after one-way ANOVA, pregnenolone: p = 0.004 for blue vs. orange; p < 0.001 for blue vs. gray and blue vs. yellow, progesterone: p = 0.01 for blue vs. orange; p = 0.02 for blue vs. gray; p = 0.02 for blue vs. yellow, testosterone: p < 0.001 for all three comparisons). For DHT, p = 0.07 from one-way ANOVA.
Fig 4.
Half maximal free steroid uptake occurs within minutes with highest uptake favoring 3β-OH, Δ5-steroids and progestogens.
Concentrations (mean ± SD) of pregnenolone (A), progesterone (B), DHEA (C), and AD (D) in prostate cancer (LNCaP) cell samples at time points ranging from 0 to 30 min. Uptake is expressed in units of μL/mg (see Methods). One-phase association curves were fit to data using GraphPad Prism 5. Half-maximum times with 95% confidence intervals: pregnenolone 4.1 min (3.2–5.8 min), progesterone 0.91 min (0.75–1.18 min), DHEA 0.43 min (0.36–0.56 min), AD 0.16 min (0.10–0.35 min). The best-fit curve for each steroid was different from each other steroid’s best-fit curve (extra sum of squares F test, p < 0.001). All graphs represent three assays with duplicate samples; error bars represent standard deviations.
Fig 5.
Uptake of pregnenolone and conversion to progesterone are faster than uptake of DHEA and conversion to AD.
Normalized concentrations (mean ± SD) over time of progesterone and AD after JEG-3 cells were treated in parallel with 100 nM pregnenolone or DHEA. Results from one representative experiment with biological triplicates are shown; similar results were obtained in a second experiment.
Fig 6.
Preferential accumulation of progestogens and 3β-OH, Δ5-steroids is observed in human prostate tissue.
Box and whisker plots of (A) blood concentrations, (B) normal prostate tissue concentrations, and (C) tissue to blood concentration ratios (tissue concentrations normalized to mean blood concentrations) for four steroids using samples from twenty prostate cancer patients. Prostatic tissue was collected at the same time as peripheral blood in patients undergoing radical prostatectomy. The ratio for pregnenolone was larger than the ratios for all other steroids; the ratio for progesterone was larger than the ratio for AD and the ratio for DHEA was larger than the ratio for AD (two-tailed Mann-Whitney U tests after Kruskal-Wallis test, p < 0.001 for all comparisons).
Fig 7.
Differences in octanol-water partition coefficients predict differences in cellular uptake of steroids.
(A) Experimentally determined log octanol-water partition coefficient (log Kow) values for nine steroids, obtained from Leszczynski and Schafer [30]. (B-D) Graphs of logs of average cell-to-media concentration ratios (from experiments described in Fig 1) vs. log Kow for the same nine steroids, in (B) prostate cancer (LNCaP) cells, (C) placental choriocarcinoma (JEG-3) cells, and (D) breast cancer (HMC-1-8) cells. For each cell line, all steroids were assayed in at least two experiments with biological duplicates; graphs represent mean ± SD of concentration ratios. (E) Graph of logs of average tissue-to-blood concentration ratios (from data described in Fig 4) vs. log Kow for pregnenolone, progesterone, DHEA, and AD. Data from twenty prostate cancer patients are included; graph shows mean ± SD of concentration ratios. For all graphs, slopes of trendlines are different from zero (p < 0.0001).
Fig 8.
More lipophilic steroids preferentially accumulate in membrane fractions of cells.
(A) Fraction of the total steroid content in the membrane plus cytosol fractions that was contained in the membrane fraction for four different steroids after fractionation of LNCaP cells that were treated with 1 μM unlabeled steroid. Graphs show mean ± SD from two independent experiments each performed in triplicate for each steroid. Pregnenolone content in membrane fractions was greater than any other steroid (Tukey’s multiple comparison test after one-way ANOVA, p < 0.001 for all three comparisons). (B) Graph of logs of the data from (A) vs. log Kow of the four steroids. Slope of trendline is different from zero (p < 0.0001).
Fig 9.
Modifying pregnenolone to decrease its lipophilicity results in decreased cellular uptake.
(A) Structures of pregnenolone (center) and four different hydroxypregnenolones. Note the addition of a single -OH group at a different location on the structure of each hydroxypregnenolone. XLOGP3 values (computational predictions of log Kows based on molecular structures) were obtained from PubChem. (B) Ratios of cellular concentrations to original treatment concentrations in culture media for the five steroids after treatment of LNCaP cells with 100 nM unlabeled steroid. Graph shows mean ± SD from one representative experiment with biological triplicates and the experiment was performed twice. The uptake of pregnenolone was greater than all other steroids (p < 0.001, Tukey’s multiple comparison test after one-way ANOVA). (C) Graph of logs of average cell-to-media concentration ratios vs. XLOGP3 for the five steroids. Graph shows mean ± SD from six total samples per steroid in two experiments. Slope of trendline is different from zero (p < 0.0001).
Fig 10.
Serum proteins prevent most of the passive cellular uptake of steroids.
(A) Box and whisker plot of concentration ratios for six steroids (2 nM [3H]-labeled steroid with 1 μM unlabeled steroid) in LNCaP cells in human serum. Experiments were performed in triplicate with at least two independent experiments for each steroid. Cellular steroid content was corrected by control experiments with no cells to account for effects of residual serum; note that this resulted in some values <0 for certain steroids, which is not physically possible, but these values were included to fully account for variability resulting from measurement uncertainty. Ratio for pregnenolone was greater than ratios for DHEA, AD, testosterone, and DHT (Tukey’s multiple comparison test after one-way ANOVA, p < 0.001 for all four). Ratio for progesterone was greater than ratios for DHEA (p = 0.003), testosterone (p < 0.001), and DHT (p = 0.002). For pregnenolone vs. progesterone, p = 0.17. (B) Fractions of total steroids (serum plus cells) collected in the cell samples when LNCaP cells were incubated in charcoal-stripped fetal bovine serum with 2 nM [3H]-labeled steroids alone and with 1 μM unlabeled steroids. Experiments were performed in duplicate with two independent experiments. Cellular steroid content was corrected by control experiments with no cells to account for effects of residual serum. For pregnenolone, progesterone, testosterone, and DHT, uptake of 2 nM [3H]-labeled steroid with 0 unlabeled steroid was greater than uptake with 1 μM unlabeled steroid (t-tests, p < 0.001 for pregnenolone, progesterone, and testosterone; p = 0.002 for DHT).