Selective Insensitivity of ZR-75-1 Human Breast Cancer Cells to 2-Methoxyestradiol: Evidence for Type II 17β-Hydroxysteroid Dehydrogenase as the Underlying Cause

Many studies have shown that 2-methoxyestradiol (2-MeO-E₂), a well-known nonpolar metabolite of the endogenous parent hormone 17β-estradiol (E₂) (reviewed in refs. 1, 2), has strong antiproliferative and apoptotic actions in a number of cultured cancer cell lines when it was present at pharmacologic concentrations (3–12). Among many of the human cancer cell lines tested thus far, several human breast cancer cell lines were found to be highly sensitive to the antiproliferative actions of 2-MeO-E₂ in vitro (3, 4, 12). Additional studies also showed that 2-MeO-E₂ inhibited the growth of human breast cancer cells when they were transplanted into immune-suppressed mice (10). Moreover, 2-MeO-E₂ has strong antiangiogenic effects both in vitro and in vivo at pharmacologic concentrations (9–11, 13). An earlier study revealed that some of the unique biological actions of 2-MeO-E₂ were not shared by several of its close structural analogues (such as 2-methoxy-E₁ and 4-hydroxy-E₂; ref. 10), which suggested a high degree of specificity for the actions of 2-MeO-E₂.

Because of the intriguing antiproliferative and antiangiogenic actions of 2-MeO-E₂, and also because of its presumed low toxicity, a great deal of research efforts have been initiated in the past few years to explore the usefulness of 2-MeO-E₂ as an effective, low-toxicity chemotherapeutic agent for human breast cancer as well as for other human cancers (14). In the present study, we compared the growth inhibitory effect of 2-MeO-E₂ in four human breast cancer cell lines, including three ER-positive cell lines (MCF-7, T-47D, and ZR-75-1) and one ER-negative cell line (MDA-MB-435s), to determine whether different human breast cancer cell lines have a differential sensitivity to the anticancer actions of 2-MeO-E₂. For the first time, we noted that the ZR-75-1 human breast cancer cells were highly insensitive to the antiproliferative actions of 2-MeO-E₂ (particularly when these cells were grown at a higher density), whereas the other three cell lines were all highly sensitive to 2-MeO-E₂. Additional experiments showed that the ZR-75-1 cells have a similar sensitivity to a number of commonly used anticancer agents as the other three cancer cell lines. These data suggested the existence of a selective insensitivity of the ZR-75-1 cells to 2-MeO-E₂, but not to several other commonly used anticancer agents.

During our search for the mechanism(s) underlying the selective insensitivity of the ZR-75-1 cancer cells to 2-MeO-E₂, we found that these cells contained very high activity of the type II 17β-hydroxysteroid dehydrogenase (17β-HSD), a steroid-metabolizing enzyme that catalyzes the conversion of E₂ to E₁ as well as the conversion of 2-MeO-E₂ to the biologically-inactive 2-methoxy-E₁. We believe that the high levels of the oxidative 17β-HSD present in ZR-75-1 cells are largely responsible for the selective insensitivity of these cells to 2-MeO-E₂ (the concept is depicted in Fig. 1). The present study provided detailed experimental data in support of this mechanistic explanation.

2-MeO-E₂, 2-methoxy-E₁, E₂, and E₁ were purchased from Steraloids (Newport, RI). Crystal violet, 50% glutaradehyde, Triton X-100, trypsin-EDTA solution (containing 0.25% trypsin and 0.02% EDTA), bovine serum albumin (BSA), fetal bovine serum (FBS), paclitaxel, doxorubicin, mitomycin C, and 5-fluorouracil were all obtained from the Sigma Chemical, Co. (St. Louis, MO). Vinorelbine tartrate (vinorelbine) was provided by GlaxoSmithKline (Research Triangle Park, NC). RPMI 1640 and Iscove's cell culture media and the antibiotics solution (containing 10,000 units/mL penicillin and 10 mg/mL streptomycin) were obtained from Life Technology (Rockville, MD). Spiro-δ-lactone was a generous gift from Dr. Donald Poirier at the Centre Hospitalier Universitaire de Québec (Canada).

[2,4,6,7,16,17-³H]E₂ and [2,4,6,7-³H]E₁ (specific activity = 118 and 100 Ci/mmol, respectively) were purchased from DuPont Life Science Products (Boston, MA). Before use in each experiment, each of the radioactive chemicals was repurified by using our high-performance liquid chromatography (HPLC) method (described later). ³H-Labeled 2-MeO-E₂ was biosynthetically prepared by incubating [³H]2-OH-E₂ (obtained from Moravek Biochemicals, Inc., Brea, CA) with porcine liver catechol-O-methyltransferase (Sigma) and S-adenosyl-l-methionine at 37°C for 2 hours. The incubation mixture was extracted twice with 7 mL of ethyl acetate and the extracts were dried under a stream of N₂. The radioactive 2-MeO-E₂ was isolated by using our HPLC system as described earlier (15, 16).

The three ER-positive human breast cancer cell lines (MCF-7, T-47D, and ZR-75-1) and one ER-negative human breast cancer cell line (MDA-MB-435s) were all obtained from the American Type Culture Collection (Manassas, VA). The methods for the in vitro culture of the ER-positive MCF-7 and T-47D cells as well as the ER-negative MDA-MB-435s cells were described in our recent study (12). For culturing the ER-positive ZR-75-1 cells, we used RPMI 1640 medium supplemented with 10% FBS and the same amount of antibiotics.

The human breast cancer cells were first propagated in the 75 cm² flasks under 37°C air with 5% CO₂ and 95% humidity to ∼80% confluence. They were then detached from the flask by treatment with 3 mL of the trypsin-EDTA solution for ∼5 minutes. Cell suspensions were centrifuged and the cell sediments were resuspended in the culture medium at the desired 10⁵ cells/mL density. A 0.1 mL aliquot of the cell suspension was then added to each well of the 96-well microplate usually at a final density of 2 × 10⁴ cells per well (unless otherwise indicated). After the cells were allowed to attach and grow for 48 hours, the cell culture medium was changed and different drug treatments were introduced at that time. In most experiments, the drug treatment lasted for 6 days with one medium change on the fourth day following the initial drug treatment.

Due to their high lipophilicity, the stock solutions of 2-MeO-E₂ (10 mmol/L) and paclitaxel (0.2 mmol/L) were prepared in pure ethanol (200 proof). The stock solutions for doxorubincin (5 mmol/L), mitomycin C (0.1 mg/mL), and vinorelbine (1 mmol/L) were prepared in phosphate buffer (pH 7.4). The stock solution containing 10 mmol/L 5-fluorouracil was prepared by first making a 50 mmol/L drug concentration in 1.0 N potassium hydroxide and followed by dilution with phosphate buffer (pH 7.4) to a 10 mmol/L drug concentration. All these stock solutions were filtered with a Millex syringe filter (0.22 μm acetatecellulose membrane) and the filtrates were stored at −20°C in tightly sealed sterile tubes. Shortly before introducing the anticancer agents to the cultured cancer cells, each chemical was freshly diluted with a buffer to the desired concentrations and an aliquot (usually 10 μL) of the drug-containing solution was added to each well. Usually, <0.1% of the original solvent of the stock solution was present in the final cell culture medium.

The cell density in the 96-well microplates was determined by using the crystal violet staining method (12, 17). Briefly, the culture medium in the microplates was first removed by aspiration, and then the cells in each well were fixed with 1% glutaraldehyde for 15 minutes. After removing the fixation solution, each well was rinsed with PBS buffer and allowed to dry at room temperature. The cells in each well were then stained with 50 μL of 0.5% crystal violet (dissolved in 20% methanol and 80% deionized water) for 15 minutes at room temperature, and the plates were rinsed carefully with tap water to remove residual crystal violet. The stained dye was then dissolved in 100 μL of 0.5% Triton X-100 overnight. After addition of 50 μL of 200-proof ethanol, the absorbance values of each well were measured at 560 and 405 nm with a UV_max microplate reader (Molecular Device, Palo Alto, CA), and the difference in the absorbance values at these two wavelengths were used to represent the cell density.

Confluent cells growing in the 96-well plates were used to determine the rate of metabolic interconversion between E₂ and E₁. Cells were incubated with 100 μL serum-free culture medium (supplemented with 100 IU/mL penicillin and 100 μg/mL streptomycin) and desired concentrations of E₂ or E₁ (containing ∼0.2 μCi of ³H-labeled E₂ or E₁). At different time points of the incubation, the medium in each culture well was removed and transferred to a microcentrifuge tube and 50 μL of methanol was added to each tube. After a brief vortex and centrifugation for 10 minutes at 14,000 rpm, an aliquot (50 μL) was injected into HPLC for analysis of the radioactive composition of E₂, E₁, and their metabolites.

The HPLC system used for this purpose consisted of a Waters 2690 separation module, a β-RAM radioactivity detector (IN/US Systems, Inc., Tampa, FL), a Waters UV detector (model 484), and a Ultracarb-5 octadecyl silane column (150 × 4.60 mm, Phenomenex, Torrance, CA). The solvent system for the separation of E₂, E₁, and their metabolites consisted of acetonitrile (solvent A), water (solvent B), and methanol (solvent C); the solvent gradient (A/B/C) was as follows: 3 minutes of isocratic at 16:68:16, 0.5 minutes of gradient to 45:45:10, 10 minutes of isocratic at 45:45:10, and then 0.5 minutes of gradient to 16:68:16. The gradient was then held for 6 minutes before analysis of the next sample. The quantification of the ³H-labeled E₂, E₁, and its metabolites was based on radioactivity measurements, and their identification was based on their retention times on the HPLC. Gas chromatography–mass spectrometry (GC-MS) analysis of the metabolically formed 2-methoxy-E₁ was used for the unequivocal identification of its structure.

For kinetic analysis of the enzyme-mediated interconversion between the 17β-hydroxyl form of an estrogen and its 17-keto form, whole cell homogenates were used. To prepare whole cell homogenates, large-scale culture of each cell line was first made, and the harvested cells were then sonicated and homogenized in 50 mmol/L sodium phosphate buffer (pH 7.4, containing 1 mmol/L EDTA). The whole cell homogenates were stored at −80°C in 200 μL aliquots, and the protein content of the homogenates was determined by using the Bio-Rad protein assay with BSA as standard.

Enzymatic reactions consisted of 0.5 mg/mL protein of the whole cell homogenates, desired concentrations of the nonradiolabeled substrate (containing 0.2-0.5 μCi of ³H-labeled estrogen substrate), 2 mmol/L of a supporting cofactor in a final volume of 200 μL of 50 mmol/L sodium phosphate buffer [containing 1 mmol/L EDTA (pH 7.4)]. The incubation was carried out at 37°C for 1 hour (unless otherwise indicated). The amount of the radioactive estrogen metabolites formed and the amount of the substrates remained after incubation were determined using our HPLC analytic method as aforementioned.

Primers. Three known oxidative 17β HSD human isozymes (namely, types II, IV, and VIII) were analyzed in the present study. For type II 17β-HSD, the forward primer was designed as 5-GCT GGT CTT GGT ATT TGC-3, which corresponds to its mRNA sequence 479 to 499, and the backward primer was 5-CTT GTC ACT GGT GCC TGC GAT-3, complementary with its mRNA sequence 972 to 912 (namely, 514 bases of the amplified sequence). For type IV 17β-HSD, the forward primer was 5-GCT CTG GAG GCT TTG GTG GAA-3, corresponding to its mRNA sequence 1,442 to 1,462, and the backward primer was 5-GGC GGC GTC CTA TTT CCT CAA-3, complementary with its mRNA sequence 1,931 to 1,951 (namely, 510 bases of the amplified sequence). For type VIII 17β-HSD, the forward primer was 5-TCT CGC CCA CCA TCT GTC GTT-3, corresponding to its mRNA sequence 307 to 327, and the backward primer was 5-CCA AGA ATG CGA CCA CAT CTG CC-3, complementary with its mRNA sequence 726 to 748 (namely, 442 bases of the amplified sequence).

The forward primer for human UDPGT_h-2 (also called type IV UDPGT) was designed as 5-ACC TGC CAA ACC CCT GCC TAA G-3, corresponding to its mRNA sequence 863 to 884, and the backward primer was 5-CAC ACA GAC CAG CAG GAA CCC AA-3, complementary with its mRNA sequence 1,495 to 1,517. The amplified sequence is a 655 base fragment within the translated region.

Reverse transcription-PCR. For the isolation of the total RNAs, 1 mL Tri reagent (containing guanidine isothiocanide, phenol, and sarkosyl) was added to the cultured cells. After repetitive pipetting to dissolve the cells, the mixture was transferred to a 1.5 mL Eppendoff tube and followed by addition of 200 μL chloroform, and then the mixture was vortexed until it appeared milky. The mixture was centrifuged at 14,000 rpm for 20 minutes. The supernatant was transferred to a new tube containing 500 μL isopropanol. After gentle reversing twice to thrice, the tube was centrifuged at 14,000 rpm for another 20 minutes. After removal of the supernatant, the pellet was washed with 75% ethanol thrice and then it was dissolved in TE buffer.

For RT-PCR, 5 μg of the total RNA in 8 μL water was mixed with 2 μL oligo-dT and 2 μL deoxynucleotide triphosphate (dNTP) in a 1.5 mL Eppendoff tube. After incubation at 65°C for 15 minutes, the tube was chilled on ice. Each tube was then supplemented with 4 μL of the buffer, 1 μL RNase-out, and 2 μL DTT. After preheating at 37°C for 2 minutes, 1 μL reverse transcriptase was added. The tube was incubated at 37°C for 60 minutes. The reaction was terminated by incubation at 70°C for 15 minutes.

In the in vitro growth inhibition experiments, the cell growth rate of the control and drug-treated groups were expressed as mean ± SE of the values obtained from six to eight replicate wells. The IC₅₀ values were calculated according to the equation for the sigmoidal dose-response curves (with variable slopes) using the nonlinear regression curve-fitting model of the Prism software. Unless otherwise indicated, one-way ANOVA was used for multiple comparisons and the Tukey's test was used for the pairwise comparisons. The Michaelis-Menten curves and the Eadie-Hofstee plots for the enzyme kinetics were drawn by using the Prism software. P < 0.05 was considered to be statistically significant and P < 0.01 was considered statistically very significant.

The antiproliferative effect of 2-MeO-E₂ in the ER-negative MDA-MB-435s human breast cancer cells is shown in Fig. 2A. 2-MeO-E₂ inhibited the growth of these cells in a concentration-dependent manner and the average IC₅₀ value was 0.8 μmol/L. The data comparing the antiproliferative effect of 2-MeO-E₂ in three ER-positive human breast cancer cell lines, MCF-7, T-47D, and ZR-75-1, are shown in Fig. 2B-D. Different cell lines had very different sensitivities to the antiproliferative action of 2-MeO-E₂. 2-MeO-E₂ inhibited the growth of MCF-7 cells in a concentration-dependent manner, with IC₅₀ of 0.75 to 1.0 μmol/L, and it also inhibited the growth of T-47D cells with IC₅₀ of 1.2 μmol/L. A near-complete inhibition of the growth of these two cell lines was reached when 2 μmol/L 2-MeO-E₂ was present. In contrast, increasing the concentrations of 2-MeO-E₂ in the culture medium from 0.1 to 2 μmol/L only exerted a very weak antiproliferative effect in the ZR-75-1 cells (Fig. 2D). Notably, the growth inhibitory effect in the ZR-75-1 cells was dependent on the density of the cells present in each culture well. When the ZR-75-1 cells were initially seeded in the 96 wells at a higher concentration (twice as usual, i.e., at 4 × 10⁴ cells/well), the insensitivity of the cells to the antiproliferative actions of 2-MeO-E₂ became more pronounced. In several separate experiments when the cells were seeded at this higher density, the maximal inhibition of the growth of the ZR-75-1 cells was <20% at up to a 10 μmol/L concentration of 2-MeO-E₂ (Fig. 2E). In one of the experiments where cells were seeded at a high density, the maximal inhibition of the ZR-75-1 cell growth was ∼40% at a 20 μmol/L concentration (data not shown).

To provide clues as to whether the observed low sensitivity of the ZR-75-1 cells to 2-MeO-E₂ was due to a nonspecific resistance of these cells to multiple anticancer drugs or due to a selective resistance only to 2-MeO-E₂, we compared the antiproliferative actions of several commonly used anticancer agents, including 5-fluorouracil, doxorubicin, mitomycin C, colchicine, paclitaxel, vinorelbine, vinblastine, and vincristine in all these four human breast cancer cell lines. We found that the ZR-75-1 cells had a similar overall sensitivity as the other three human breast cancer cell lines to the antiproliferative actions of these agents (Table 1). These results clearly suggest that the marked insensitivity of the ZR-75-1 cells to 2-MeO-E₂ was selective only to this anticancer steroid, which likely was mediated by a 2-MeO-E₂–specific mechanism that is different from the mechanism(s) of the usual multidrug resistance. In addition, we also observed that the ER-positive ZR-75-1 cells had a considerably low sensitivity to the antiestrogenic action of ICI-182,780 (a pure ER antagonist) compared with the ER-positive MCF-7 and T-47D cells (Table 1).

To determine the underlying mechanism(s) for the observed selective insensitivity of ZR-75-1 cells to 2-MeO-E₂, we set out to explore the possibility that this insensitivity may be caused by the presence of high levels of steroid-selective metabolizing enzyme(s) in these cells, which provides rapid metabolic inactivation of 2-MeO-E₂. Our initial experiments showed that when [³H]E₂ or [³H]E₁ were introduced to each of the four cell lines in culture, the distinct difference in their metabolism was the drastically different rates of interconversion between E₂ and E₁. Following this observation, we conducted a detailed comparison of the rates of interconversion between E₂ and E₁ (and between 2-MeO-E₂ and 2-methoxy-E₁) in these four cell lines and some of the data are described below.

Conversion of E₂ to E₁ and 2-MeO-E₂ to 2-methoxy-E₁. To determine the rate of metabolic conversion of E₂ to E₁ in different human breast cancer cell lines, cells were first grown to near confluence in the 96-well plates. The cells were then incubated for a desired length of time in the presence of 100 μL serum-free culture medium supplemented with 1 μmol/L E₂ (containing ∼0.5 μCi ³H-labeled E₂). The profile of E₂ metabolites present in the culture medium was determined by HPLC analysis.

When the ER-positive MCF-7 and T-47D cells were cultured under these experimental conditions for up to 24 hours (Fig. 3A and B), little or no E₁ was detected in the culture medium, although a small fraction (20-30%) of [³H]E₂ was found to be converted to other radioactive metabolites (mostly water-soluble conjugates) during this period. Similarly, little or no E₁ was detected as a metabolite 24 hours after introduction of 1 μmol/L [³H]E₂ to the ER-negative MDA-MB-435s cells (Fig. 3D), but ∼20% of the original [³H]E₂ was metabolically converted to a water-soluble conjugate.

When the ZR-75-1 cells were tested, the disappearance of 1 μmol/L [³H]E₂ in the culture medium occurred very rapidly and in a time-dependent manner (Fig. 3C). The apparent half-life (T_1/2) of E₂ in the culture medium was only ∼2 hours, and the concentrations of E₂ remaining after 12 and 24 hours of incubation were only 8.0% and 7.6%, respectively, of its initially introduced concentration (Fig. 3C). Our HPLC analysis showed that the major metabolite formed from E₂ was E₁, which accounted for ∼90% of all radioactive E₂ metabolites detected. Notably, the rate of E₁ formation increased in a manner that is almost exactly opposite to the rate of E₂ disappearance, suggesting that the conversion of E₂ to E₁ was a predominant pathway of E₂ metabolism in this cell line.

To determine whether the presence of different E₂ substrate concentrations had any effect on the overall patterns of the metabolic conversion of E₂ to E₁ in cultured ZR-75-1 cells, three different concentrations of E₂ covering a wide range (10-1,000 nmol/L) were tested (one representative data set was shown in Fig. 4, top). The percent conversion of E₂ to E₁ and the apparent half-lives (T_1/2) were found to be independent of the [³H]E₂ substrate concentrations present, suggesting that the metabolic conversion of E₂ to E₁ in intact ZR-75-1 cells primarily followed the first-order kinetics, with an apparent T_1/2 of ∼2 hours.

For comparison, we also determined the metabolism of [³H]2-MeO-E₂ at two subtoxic concentrations (0.1 and 0.5 μmol/L) in cultured ZR-75-1 cells under the same experimental conditions. The radioactive [³H]2-MeO-E₂ used in these experiments was biosynthetically prepared and purified with HPLC a day before its use in the experiments (described in Materials and Methods). The metabolic patterns for the conversion of [³H]2-MeO-E₂ to 2-methoxy-E₁ was almost the same as the pattern for the conversion of E₂ to E₁. The apparent T_1/2 was ∼2.5 hours (Fig. 4, bottom).

Conversion of E₁ to E₂. We also determined the rate of metabolic conversion of E₁ to E₂ in these cell lines under the same conditions. When the ER-positive MCF-7 and T-47D cells were incubated with a medium containing 1.0 μmol/L [³H]E₁, these two cell lines showed a similar overall pattern for the conversion of E₁ to E₂, with ∼50% of E₁ converted to E₂ in MCF-7 cells and ∼65% converted in T-47D cells after 24 hours (Fig. 3A, and B′). In comparison, when the ER-positive ZR-75-1 cells were cultured in a medium containing 1.0 μmol/L [³H]E₁, the majority (∼80%) of the E₁ was not metabolized and only <20% of the E₁ was converted to other non-E₂ metabolites after 24 hours (Fig. 3C ′).

In the ER-negative MDA-MB-435s cells, the rate of conversion of 1.0 μmol/L [³H]E₁ to E₂ was slow (Fig. 3D ′). Notably, although E₁ disappeared in cultured MDA-MB-435s cells in a time-dependent manner, most of the E₁ was converted to other metabolites (not E₂), including a glucuronidated metabolite of E₁ (based on its HPLC retention time).

In summary, for the oxidative conversion of the 17β-hydroxysteroids to 17-ketosteroids, the ZR-75-1 cells had the highest metabolic activity, whereas the other cell lines (MCF-7, T-47D, and MDA-MB-435s) had little or no detectable activity. For the reductive conversion of the 17-ketosteroids back to 17β-hydroxysteroids, the T-47D and MCF-7 cells had a high activity, whereas the ZR-75-1 and MDA-MB-435s cells had a very low activity. On the basis of these findings, it is apparent that the ZR-75-1 cells in culture have the highest net activity among the four cell lines tested that favors the metabolic accumulation of 2-methoxy-E₁, but not 2-MeO-E₂.

The ZR-75-1 cells expressed type II 17β-HSD, whereas the MCF-7, T-47D, and MDA-MB-435s cells had no detectable levels of expression of the type II 17β-HSD (Fig. 5). In comparison, all four cell lines expressed types IV and VIII 17β-HSD, and no significant difference in the amount of their RNA levels was observed (Fig. 5). The expression level was not affected by treatment of these cells with 10 nmol/L E₂ for 24 or 48 hours.

The human UDPGT_h-2 is a major form of the glucuronsyltransferase that has specific activity for the glucuronidation of estrogens and is expressed in human breast cancer cells (18). The expression of this representative conjugation enzyme was also measured in all four cell lines. Whereas all four cell lines were found to express detectable levels of UDPGT_h-2, the level of expression was highest in the MDA-MB-435 cells, and only very low levels of this conjugating enzyme were expressed in the three ER-positive cell lines (Fig. 5). Notably, the high level of expression of the UDPGT_h-2 in MDA-MB-435s cells perfectly agreed with our observation that a substantial portion of the radioactive estrogen substrate was converted in a time-dependent manner to a glucuronide in these cells.

Using crude whole cell homogenates prepared from cultured ZR-75-1 cells, we have also systematically characterized the biochemical properties of the 17β-HSD activity in this cell line. In addition, we have also compared the catalytic activity of 17β-HSD for the conversion of E₂ to E₁ and 2-MeO-E₂ to 2-methoxy-E₁ in all four cell lines. These biochemical studies were designed to validate our following suggestions: (a) the type II 17β-HSD in the ZR-75-1 cells is largely responsible for the rapid metabolic conversion of E₂ to E₁ and the conversion of 2-MeO-E₂ to 2-methoxy-E₁ in these cells and (b) the difference in the overall oxidative 17β-HSD activity in these four cell lines determines their differential sensitivity to 2-MeO-E₂. Some of the related data from this part of the study are summarized below.

Protein concentrations and incubation time. When 10 μmol/L E₂ was used as substrate and 2 mmol/L NAD as cofactor, the rate of conversion of E₂ to E₁ at different concentrations of cell homogenates (0, 0.25, 0.5, 0.75, and 1.0 mg protein per milliliter) correlated linearly with increasing protein concentrations (Fig. 6A). Under the same reaction conditions and at a homogenate concentration of 0.5 mg protein per milliliter, increasing the incubation time from 15 to 90 minutes resulted in an almost linear increase in the conversion of E₂ to E₁ (Fig. 6B). Similar patterns were also observed when 10 μmol/L E₁ was used as substrate and 2 mmol/L NADH as cofactor (Fig. 6A ′ and B′). Based on these assays, an optimized protein concentration of 0.5 mg/mL and an incubation time of 30 minutes were devised for most of the enzymatic assays.

Cofactor preference. The ability of NAD, NADH, NADP, or NADPH to support the interconversion between E₂ and E₁ was compared when 10 μmol/L of [³H]E₂ or [³H]E₁ was incubated with ZR-75-1 whole cell homogenates (at 0.5 mg/mL protein) in the presence of each of the cofactors (at 2 mmol/L). For the conversion of E₂ to E₁, NAD was found to be the optimal cofactor and no appreciable conversion was detected when NADH, NADP, or NADPH was used as a cofactor (Fig. 6C). For the reverse conversion of E₁ to E₂, NADH was the best cofactor, followed by NADPH, but NAD and NADP could not support the reaction (Fig. 6C ′).

Reaction pH. At pH 4 to 5, no detectable conversion of E₂ to E₁ was observed. The rate of conversion of E₂ to E₁ increased continuously from pH 5 to 9 roughly in a linear manner, and the conversion was nearly plateaued when the pH reached 9 (Fig. 6D). For the conversion of E₁ to E₂, no appreciable activity was detected at pH 4, but the rate increased rapidly when the pH increased from 4 to 5, and the rate remained high when the pH varied from 5 to 8. At pH > 8, the rate of conversion of E₁ to E₂ started to decrease (Fig. 6D ′).

Incubation temperature. For the conversion of E₂ to E₁ (at pH 7.4), the enzymatic activity at 45°C was ∼32% higher than that at 37°C, but this activity was completely destroyed when the temperature was increased to 60°C. At pH 9.0, the enzymatic activity was not significantly different when the temperature was increased from 37°C to 45°C, but again the activity was almost completely lost when the temperature was increased to 60°C (Fig. 7A).

For the conversion of E₁ to E₂, the enzymatic activity (at pH 7 and 9) was increased by 84% or 82%, respectively, when the incubation temperature was increased from 37°C to 45°C, but this activity was mostly lost at 60°C, regardless of the pH differences (Fig. 7B).

Inhibition by spiro-δ-lactone, 17α-estradiol, and progesterone. Spiro-δ-lactone is a potent, highly specific inhibitor of the human type II 17β-HSD (19, 20). The presence of spiro-δ-lactone in the reaction mixture caused a concentration-dependent inhibition of the enzymatic conversion of [³H]E₂ to E₁ and also the conversion of [³H]2-MeO-E₂ to 2-methoxy-E₁ (Fig. 7A and B), with highly similar IC₅₀ values. It is of note that the IC₅₀ values of spiro-δ-lactone measured in the present study were comparable with the earlier measurements when the recombinant human type II 17β-HSD was assayed as the enzyme for the conversion of E₂ to E₁ (19, 20). In addition, the presence of 17α-estradiol or progesterone at 2 or 10 μmol/L (two nonspecific inhibitors) also showed similar degrees of inhibition of the conversion of [³H]E₂ to E₁ or [³H]2-MeO-E₂ to 2-methoxy-E₁ (Fig. 7C and D).

Determination of enzyme kinetic parameters (K_m and V_max). The enzyme kinetics for the conversion of E₂ to E₁ were determined under the optimized in vitro reaction conditions. The incubation mixture consisted of 0.5 mg/mL of cell homogenate protein, 2 mmol/L of NAD, and different concentrations of E₂ in a final volume of 0.2 mL of 50 mmol/L potassium phosphate buffer [containing 1 mmol/L EDTA (pH 7.4)]. The Michaelis-Menten curve and the Eadie-Hofstee plot for the 17β-HSD–mediated conversion of E₂ to E₁ were shown in Fig. 8A. The apparent K_m and V_max values (calculated according to curve regression analysis) were 1.8 μmol/L and 37.0 pmol/mg protein/min, respectively.

Under exactly the same in vitro reaction conditions, we also determined the kinetic parameters for the conversion of radioactive 2-MeO-E₂ to 2-methoxy-E₁. The Michaelis-Menten curve and the Eadie-Hofstee plot were shown in Fig. 8B. The apparent K_m and V_max values (calculated according to curve regression) were 1.6 μmol/L and 35.3 pmol/mg protein/min, respectively. Notably, additional GC-MS analysis of the metabolite peak collected off the HPLC confirmed that the major metabolite formed from 2-MeO-E₂ was 2-methoxy-E₁ (data not shown).

Comparison of the reductive 17β-hydroxysteroid dehydrogenase activity in all four cell lines. The whole cell homogenates from the ZR-75-1 cells contained the highest catalytic activity for the conversion of E₂ to E₁ and 2-MeO-E₂ to 2-methoxy-E₁. In comparison, the activity in the MCF-7 and T-47D cells was very low, only one fifth of the activity in the ZR-75-1 cells, and homogenates from the MDA-MB-435s cells had the lowest levels of the reductive 17β-HSD activity, approximately one tenth of that in the ZR-75-1 cells (Fig. 8C). Notably, because RT-PCR analysis showed that the three sensitive cell lines (MCF-7, T-47D, and MAD-MB-435s) expressed no detectable levels of type II 17β-HSD mRNA (Fig. 5), the very low catalytic activity detected in the homogenates of these cell lines is likely contributed by other 17β-HSD isozymes (such as types IV and VIII).

The results of our present study showed that the ER-positive ZR-75-1 human breast cancer cells were insensitive to the antiproliferative actions of 2-MeO-E₂. After comparing the sensitivity of the four cell lines used in this study to several other anticancer agents, including 5-fluorouracil, mitomycin C, doxorubicin, colchicine, vinorelbine, and paclitaxel, we noted that the ZR-75-1 cells were selectively insensitive to the anticancer actions of 2-MeO-E₂, which was not due to the development of a general multiple drug resistance in this cancer cell line. Also, we noted that the ER-positive ZR-75-1 cells had a lower sensitivity to the antiproliferative actions of ICI-182,780 (a pure ER antagonist) compared with the ER-positive MCF-7 and T-47D cells. Although the mechanisms for the antiproliferative actions of 2-MeO-E₂ and ICI-182,780 are completely different, both of these agents share the same steroid core structure with a free 17β-hydroxyl group. This similarity has led us to speculate that the selective insensitivity of the ZR-75-1 cells to 2-MeO-E₂ is likely attributable to a mechanism that is associated with the steroid chemical structure, such as due to the presence of high levels of certain steroid-specific metabolizing enzyme(s).

To probe the underlying mechanism(s) for this selective insensitivity, we compared all four cell lines for their ability to metabolize E₂ and 2-MeO-E₂. We found that when E₂ or 2-MeO-E₂ was added to the culture medium (at up to 1 μmol/L), they were almost completely converted to E₁ or 2-methoxy-E₁ in cultured ZR-75-1 cells within the first ∼24 hours, whereas little conversion was detected in the sensitive MCF-7, T-47D, and MDA-MB-435s cells. Similarly, when the whole cell homogenates prepared from these four cell lines were assayed in vitro for the enzymatic activity to convert 2-MeO-E₂ to 2-methoxy-E₁ (or E₂ to E₁), we confirmed that the ZR-75-1 cells had the highest overall catalytic activity, whereas the activity in other three cell lines was much lower. Therefore, the ability of each cell line to convert 2-MeO-E₂ to 2-methoxy-E₁ (a metabolite with little or no anticancer activity) was, to a large extent, directly correlated with the insensitivity to 2-MeO-E₂.

The 17β-HSDs is a family of NAD(H)- and/or NADP(H)-dependent enzymes that catalyze the interconversion between 17β-hydroxysteroids and 17-ketosteroids (21–26). These enzymes have a widespread distribution in gonadal and extragonadal tissues. At least eight isozymes of 17β-HSD (types I-VIII) have already been identified (21–26). Each of these isozymes generally is classified either as oxidative or reductive 17β-HSD on the basis of the dominant reaction it catalyzes. Whereas the oxidative isozymes (including types II, IV, and VIII) catalyze the conversion of 17β-hydroxysteroids to 17-ketosteroids, the reductive isozymes (types I, III, V, VI, and VII) predominantly catalyze the reduction of 17-ketosteroids to 17β-hydroxysteroids.

Our data on the interconversion between E₂ and E₁ in cultured MCF-7 and T-47D cells were largely in agreement with those reported earlier (27, 28). Notably, a few earlier studies have suggested that the oxidative activity of 17β-HSD present in ZR-75-1 cells was due to the presence of the type II isozyme (25, 29–31), a 42.8 kDa enzyme (mRNA = 1.4 kb) consisting of 387 amino acid residues (21, 32, 33). Using RT-PCR, we determined the mRNA levels of various oxidative 17β-HSD isozymes (namely, types II, IV, and VIII) in all four cell lines used in this study. The ZR-75-1 cells expressed high levels of type II 17β-HSD, whereas the MCF-7, T-47D, and MDA-MB-435s cells expressed no detectable levels of type II 17β-HSD (Fig. 5). Because our data showed that all four cell lines expressed similar levels of type IV and VIII 17β-HSD, and also because the overall reductive 17β-HSD activity in the whole cell homogenates from these three sensitive cell lines were much lower than the activity in ZR-75-1 cells (Fig. 8C), it is clear that the relative contribution of the types IV and VIII 17β-HSDs to the overall conversion of 2-MeO-E₂ to 2-methoxy-E₁ (or E₂ to E₁) is of far lesser importance compared with the type II 17β-HSD.

Providing further support to the suggestion that the type II 17β-HSD present in the ZR-75-1 cells is the dominant isozyme that is largely responsible for the conversion of 2-MeO-E₂ to 2-methoxy-E₁, we also showed that the kinetic characteristics of the 17β-HSD activity in the ZR-75-1 whole cell homogenates were highly similar to those of purified type II 17β-HSD in many ways. (a) The effect of pH (pH 7 and 9) on the two enzyme activities was similar. (b) The 17β-HSD from the ZR-75-1 cells was almost unchanged when the temperature was increased from 37°C to 45°C (at pH 9), but the activity was almost completely lost when it was further increased to 60°C. A very similar pattern in response to temperature change was also observed with the purified human type II 17β-HSD isozyme but not with other isozymes, such as types I and III (19). (c) NAD was the optimal cofactor for the enzymatic conversion of E₂ to E₁, whereas NADH was the optimal cofactor for the reversed conversion of E₁ to E₂. These cofactor requirements are the same as observed earlier with the purified human type II 17β-HSD (32). (d) The apparent K_m values for the 17β-HSD in the ZR-75-1 cells were similar to those determined for the purified type II 17β-HSD isozyme (20, 32, 33). Taken together, these data clearly suggest that the main oxidative activity of 17β-HSD present in the ZR-75-1 cells came from the type II isozyme.

In this study, we also found that the biochemical properties for the metabolic conversions of E₂ to E₁ and 2-MeO-E₂ to 2-methoxy-E₁ are almost identical: (a) very similar K_m and V_max values, (b) very similar response to temperature changes, and (c) very similar inhibition patterns by spiro-δ-lactone (a specific type II 17β-HSD inhibitor refs. 19, 20), 17α-estradiol, and progesterone. Collectively, these data showed that the same type II 17β-HSD that catalyzes the conversion of E₂ to E₁ in the ZR-75-1 cells also catalyzes the conversion of 2-MeO-E₂ to 2-methoxy-E₁.

It is of note that the type II 17β-HSD is broadly expressed in endometrial hyperplasia and carcinoma as well as certain forms of breast cancer (21, 33–35). In addition, a high activity of the oxidative 17β-HSD is present in the liver and kidney (21, 33). The high activity of the hepatic oxidative 17β-HSD may contribute importantly to the extensive metabolic inactivation when 2-MeO-E₂ is given to patients through oral administration for treatment of the cancer. In line with this suggestion, 2-methoxy-E₁ (not 2-MeO-E₂) was found to be the major metabolite present in the circulation of pregnant women (2). In addition, recent pharmacokinetic studies in humans and animals also showed that large amounts of 2-methoxy-E₁ (instead of 2-MeO-E₂) were present in circulation after oral administration of 2-MeO-E₂. Because 2-methoxy-E₁ has little antiproliferative activity compared with 2-MeO-E₂ in cultured breast cancer cells (data not shown), the presence of high levels of the oxidative 17β-HSD that catalyze rapid conversion of 2-MeO-E₂ to 2-methoxy-E₁ is believed to significantly reduce the sensitivity of cancer cells to 2-MeO-E₂. In this context, it is of note that novel derivatives of 2-MeO-E₂ with modifications at the 17β position to make them less susceptible to type II 17β-HSD–mediated metabolic inactivation may show greater anticancer activity. Similarly, the results of our study also suggest that the use of selective type II 17β-HSD inhibitors may effectively enhance the anticancer activity of 2-MeO-E₂.

In summary, the results of our present study showed that the ZR-75-1 human breast cancer cells are selectively insensitive to the antiproliferative actions of 2-MeO-E₂. Mechanistically, the insensitivity to 2-MeO-E₂ is associated with the rapid metabolic conversion of 2-MeO-E₂ to 2-methoxy-E₁. Overall, our enzyme kinetic studies of the conversion of 2-MeO-E₂ to 2-methoxy-E₁ and E₂ to E₁ along with our studies of the 17β-HSD gene expression clearly shown that the type II 17β-HSD is predominantly responsible for the oxidative inactivation of 2-MeO-E₂. Our results also suggest that derivatives of 2-MeO-E₂ with modifications at the 17β position to have a reduced susceptibility to the 17β-HSD may have stronger anticancer activity, and a concomitant use of a potent selective 17β-HSD inhibitor may enhance the anticancer activity of 2-MeO-E₂ in breast cancer patients.

Grant support: NIH grant RO1 CA97109.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.