A Fluorometric CYP19A1 (Aromatase) Activity Assay In Live Cells

Abstract

Inhibition of estrogen synthesis is an integral component of the frontline pharmacologic therapy for the treatment of estrogen receptor positive cancers. However, there is currently no direct, high-throughput-ready assay for aromatase (also known as CYP19A1) that can be performed in live cells. Here we present a cell-based assay that allows for multiplexed assessment of enzyme activity, protein half-life, cell viability, and identification of inhibitors with slow off-rates.

Graphical Abstract

A fluorometric assay was developed for CYP19A1 (aromatase) for use in live cells. This allows for multiplexed evaluation of enzyme activity, cell viability, and protein production and degradation.

1. Introduction

Cytochrome P450 19A1, commonly referred to as aromatase, is a key regulator of estrogen homeostasis.^[1] The role of estrogens as proliferative and carcinogenic agents in estrogen-receptor positive (ER⁺) breast cancers was uncovered in the latter half of the 20^th century, and the blockade of this pathway became a first-line treatment for ER⁺ breast cancer patients.^[2] Currently there are three FDA approved CYP19A1 inhibitors (AIs): anastrozole, letrozole, and exemestane (Scheme 1). These are designated into two mechanistic classes, with the first two being non-steroidal and the third a steroidal type inhibitor.^[3] While CYP19A1 inhibitors are essential drugs for postmenopausal breast cancer survivors, each compound has side-effects, and loss of efficacy is common in advanced stage disease due to various drivers of acquired resistance.^[4] New CYP19A1 inhibitors could be beneficial, but there are currently no direct, convenient, high-throughput methods to assess enzyme activity that are amenable to screening compound libraries.^[5] This has slowed research in this important area.

Scheme 1.

Aromatase substrates and inhibitors. A) Reaction of DBF to produce emissive products. B) Structures of 19A1’s substrate testosterone, and FDA approved aromatase inhibitors.

The few assays that are currently available to screen for CYP19A1 inhibitors have significant drawbacks. The human placental microsomal assay, developed between 1974 and 1980,^[6] utilizes microsomes generated from ex vivo tissues that exhibit a high degree of aromatase activity. Microsomes provide a membrane environment to assess enzymatic turnover, which models the cellular setting, where an N-terminal sequence anchors CYP19A1 in the ER membrane. A system for regenerating NADPH is commonly used to supply the reducing equivalents needed for activity, along with radiolabelled substrates such as [1β−³H]-androstenedione or [1β−³H]-16α-hydroxyandrostenedione.^[7] While microsomes provide a membrane, the lack of the cellular context is suboptimal for small molecule screening, as there is no information gained on cellular uptake of compounds, or off-target effects, such as binding to intracellular proteins that might either inactivate the inhibitor or cause toxicity. Radiometric substrates have also been used in live-cell assay formats,^[8] but the drawbacks include regulatory and safety concerns, and consequently, are not commonly utilized by most research laboratories. Other assays rely on ELISA^[9] or analytical equipment such as GC- or LC-MS^[10] to measure the production of aromatase metabolites, providing an activity measurement that is low-throughput and time consuming. To improve throughput, a cellular assay utilizing time-resolved fluorescence was developed, though the assay requires specialized donor-acceptor pairs.^[11] A more general cell based high-throughput assay was developed to identify inhibitors in the estrogen metabolic pathway. For this, MCF7 cells were engineered with luciferase expression under the control of a promoter regulated by an estrogen response element (ERE; three repeats were used to increase hormone induction sensitivity). The assay provided a convenient luminescent readout, but was indirect, as it reported on binding of estrogen and endocrine disrupting chemicals to the ERE, rather than aromatase activity.^[12]

Given the need for a simple cell-based assay format to measure CYP19A1 activity, we undertook its development. The result is a readily accessible assay for rapid and cost-effective quantification of enzyme activity under endogenous cellular conditions. The same assay format can be used to determine cytotoxicity, identify slow off-rate inhibitors, and the effect of the inhibitor on CYP19A1 production and degradation (Figure 1).

Figure 1.

Multiple functional readouts provided by this genetically encoded cell-based aromatase assay. Fluorescent signals quantifying enzyme activity and cytotoxicity can be determined in the same plate. Wash-out experiments provide insights into off-rates, and the photoconvertible CYP19A1-Dendra2 fusion reports on protein production and degradation.

2. Results and Discussion

A number of CYPs act on a wide variety of fluorometric substrates. These molecules are non-emissive until metabolism by the enzyme induces a chemical modification that radically increases the emission. Dibenzylfluorescein (DBF) has been previously used as a simple fluorometric reporter with purified CYP19A1.^[13] However, its application in cells was not explored, possibly due to the need to induce two chemical transformations for optimal signal (Scheme 1A). The second step requires the hydrolysis of an ester by the addition of strong base to convert fluorescein benzyl ester (1) to fluorescein (2). In addition, it was not clear if the DBF would be cell penetrant.

To test the suitability of DBF for optical detection of CYP19A1 activity in live cells, an HEK293 T-REx cell line expressing an inducible CYP19A1 was generated. This allowed for control over the production of the enzyme by addition of tetracycline. Moreover, as HEK cells do not endogenously express appreciable levels of CYPs, background signals were minimized. Following induction of transcription and addition of DBF, initial studies demonstrated an increase in emission at 530 nm in cells at 24 hrs without the need for the addition of strong base (Figure 2A). Based on this finding, a dual expression plasmid was generated to improve the signal to noise ratio (S/N) by co-expressing CYP19A1 with cytochrome P450 oxidoreductase (POR). This reductase partner is essential for CYP activity, but POR is present in limiting quantities in most cellular systems. Co-expression of CYP19A1 and POR resulted in a further increase in signal, from 2.9 to 4.4 fold over background.

Figure 2.

Aromatase in live cells. A) Activity of aromatase determined by emission of fluorescein benzyl ester in stable cell lines containing the specified genes. B) Localization of aromatase to the ER. Nuclei were stained with Hoechst 33342 (blue), the ER with ER-ID (red) and CYP19A1-Dendra2 is green. C) Dose responses of aromatase inhibitors. D) Washout experiments differentiate slow off-rate inhibitors.

In order to easily monitor CYP19A1 expression and cellular localization, a C-terminal fusion with the photoactivatable protein, Dendra2, was generated. Notably, this fusion resulted in a higher observed production of the emission signal from 1. This resulted in 2.4, 4.5, 4.7 and 5.4-fold increases in signal over background for the CYP19A1, CYP19A1 + POR, CYP19A1-Dendra2, and CYP19A1-Dendra2 + POR cell lines (Figure 2A). In all cases, emission was further increased by a factor of ~ 2x upon adding base to the cells (Figure 3A).

Figure 3.

Evaluation of CYP19A1 cell lines. A) Comparison in fold-change of emission for different cell lines under indicated conditions. CYP19A1 = 19A1; Dendra2=d2. B) Comparison of enzyme activity values and cell viability measurements allows for identification of false positives. Cisplatin = CP, doxorubicin = Dox, mitoxantrone = Mito, and H89 is a kinase inhibitor. All compounds (10 μM) were incubated with the cells in the presence of DBF for 24 hrs, followed by the addition of resazurin.

Immunoblots for CYP19A1 showed a substantial increase in the amount of CYP19A1-Dendra2, explaining the higher emission values in the activity assay seen with these cell lines (Figure S1A). The localization of CYP19A1-Dendra to the ER of the cells was confirmed by assessment of co-localization with an ER selective dye (ER-ID; Figure 2B). The Pearson’s correlation coefficient was 0.72 for cells expressing CYP19A1-Dendra (Figure S7). As this cell line is not clonal, variable expression levels were observed in individual cells.

Since the fold-change in the emission for the DBF turnover was low, the z-factor (also referred to as z’) was calculated to assess the robustness of the assay. The emission signal from the CYP19A1 cell lines, without addition of NaOH, resulted in z-factors that ranged from 0.4 to 0.75 for time points less than 8 hrs, which is suboptimal. However, the z-factor was between 0.7 and 0.9 at time points between 8 and 24 hrs, which provides for an “excellent” assay.^[14] The addition of NaOH gave z-factors between 0.71 and 0.9 for time points between 6 and 24 hrs. All other cell lines gave a z-factor between 0.66 and 0.92 over a time period of 4 to 24 hrs without the addition of NaOH, allowing for the assay to be run after a convenient 4 hr incubation with DBF.

While DMSO has been reported to interfere with aromatase activity,^[11] the cellular assay maintained a consistent emission signal up to a concentration of 1% DMSO. At higher concentrations the emission signal decreased by 5% at concentrations up to 2% DMSO (Figure S2A). The assay was compatible with up to 5% fetal bovine serum and could be performed in Leibovitz’s L-15 medium, allowing the flexibility to carry out experiments without the presence of CO₂ (Figure S2B).

Several inhibitors of CYP19A1 were evaluated and their EC₅₀ values in the cell lines compared to those reported in the literature (Figure 2C, Table 1). All cell lines gave a similar EC₅₀ value, regardless of POR expression or the presence of the Dendra2 fusion tag, and were in good agreement with a commercial in vitro assay (Table 1).

Table 1.

Comparison of EC₅₀ values from in-cell assay vs. in vitro assays.

Compound	IC50 (nM) In cell assay[a]	IC50 (nM) In vitro assay[b]	IC50 (nM) Literature[c]	Mechanism
Anastrozole	9.6 (7.6–12)	3.3 (1.7–6.5)	0.8 – 600[15]	Heme binding
Letrozole	0.53 (0.36–0.77)	1.2 (0.3–47)	0.07– 20[15]	Heme binding
Exemestane	32 (27.1–37.7)	34.5(22.7–52.4)	42[16]	Irreversible inhibitor

^[a]Values in parentheses reflect the 95% confidence interval. Results were consistent across all cell lines.

^[b]BioVision in vitro assay utilizing purified recombinant aromatase.

While assessment of enzyme inhibitory concentrations has great value, it would also be helpful to be able to use the assay to obtain insights into mechanistic features of CYP19A1 inhibitors. Some agents, such as exemestane, are irreversible inhibitors, while others, such as anastrozole, are competitive inhibitors. To determine if the assay could differentiate inhibitors that have slow off-rates or utilize irreversible mechanism of inactivation, exemestane and anastrozole were incubated with the cells for 90 min in the presence of the translation inhibitor, cycloheximide (CHX). The cells were then extensively washed, followed by the addition of media containing DBF and CHX. Emission readings were taken over 8 hrs (Figure 2D). Cells treated with anastrozole and then washed exhibited greater activity, with 65% of the signal intensity observed for cells that had not been treated with an inhibitor. In contrast, treatment with exemestane, followed by washout, resulted in only 20% of the reference signal. Based on this data, a simple washout protocol could be used to identify inhibitors with slow off-rates or irreversible mechanisms of enzyme inactivation. Alternatively, washout experiments with competitors present in the media used in the washes, such as the CYP19A1 substrate testosterone, could also be useful to gain information about binding modes.

During screening, it is possible that some compounds could be misidentified as inhibitors if they induce cytotoxicity. These “false positives” would result in a reduction in signal intensity due to cell death, rather than enzyme inhibition. To address this, the assay was expanded to include the subsequent addition of resazurin for evaluation of cell viability. First, the activity assay was read for fluorescein benzyl ester emission. Resazurin was then added to the cells in the presence of DBF, and incubated for one hr. The conversion of resazurin to resorufin was measured with an excitation of 535 nm and emission of 595 nm, which is suitably separated from the fluorescein emission at 530 nm.

To validate the approach, CYP19A1 expressing cells were incubated with known compounds that induce cytotoxicity (Figure 3B). Following a 24 hr incubation with the compounds, resazurin was added. The resulting resorufin emission was reduced by 50%, which correlated well with the reduction in dibenzylfluorescein turnover. Thus, chemical scaffolds that could be misidentified as hits for CYP19A1 inhibition could be triaged based on off-target cytotoxicity using one assay.

We have previously reported that photoconversion of Dendra2 can be used as a cell-based, high-throughput screen to identify small molecules that alter protein translation or degradation.^[17] Several reports have shown that inhibitors of CYP19A1 can slow the degradation of this enzyme, as determined by ³⁵S pulse chase.^[18] To determine if the Dendra2 tag on CYP19A1 could recapitulate these results with a more convenient optical signal, the CYP19A1-Dendra2 cell line was irradiated with 405 nm light for 1 min to convert ~ 50% of the population of CYP19A1-Dendra2 from the green (530 nm) to the red emissive (595 nm) form. This stable pool of red-emissive CYP19A1-Dendra2 was then monitored to evaluate protein degradation, while the increase in signal of the green-emissive pool of protein reflected continued translation of the protein.

Cells were imaged over a 24 hr period, with hundreds of cells captured per image. The intensity of CYP19A1 signal with emission at 595 nm decreased by 5x in the images, while the emission at 530 nm increased 2.3x (Figure 4A). Quantifying the emission intensities demonstrated a loss in post-translated CYP19A1, with a steady increase in the population of newly translated CYP19A1 (Figure 4A, B). The rate of decay was similar to that seen with untagged CYP19A1 (Figure S3).

Figure 4.

Use of Dendra2 reporter to monitor aromatase protein lifetime. A) The green emissive form of Dendra2 (top) reports on existing and newly translated protein, while the red emission (bottom) reports solely on the post-translated protein. Scale bar of 20 μm indicated at bottom right. B) Emission intensity from microscopy experiments of untreated CYP19A1-Dendra2 expressing cells over a 24 hr period. C) Ratio of the emission intensity for green emissive Dendra2 (emission intensity at 530 nm at T=24 hr / Intensity at T=0 hr) for cells treated without compound (NC), Exemestane (Exe), Anastrozole (Ana), and Letrozole (Let); D) Ratio of the emission intensity for red emissive Dendra2 (emission intensity at 595 nm).

To assess the ability of small molecules that impact CYP19A1 levels in cells by altering the protein half-life, cells expressing CYP19A1-Dendra2 were incubated with different inhibitors. Treatment with anastrozole resulted in only a 15% reduction in 595 nm emission at 24 hrs, compared to 80% for the untreated cells, indicating that the protein was not being degraded at the rate seen in the untreated cells. Meanwhile, a 250% increase in intensity was observed at 530 nm compared to untreated cells, demonstrating that more of the newly synthesized protein was preserved in the cells (Figure 4C, D). These effects were not seen with exemestane, where the changes were comparable to untreated cells. Letrozole was similar to anastrozole, but not as extreme, with a 64% reduction in the 595 nm emission intensity and a 13% increase in 530 nm emission intensity. Similar results for the degradation of CYP19A1 were observed by immunoblot in the CYP19A1 cell line treated with cycloheximide followed by incubation with inhibitors over 24 hrs (Figure S4, S5).

These results demonstrate the advantages of an engineered cell line, coupled with the appropriate fluorogenic reporter, for analysis of CYP activity and lifecycle. Given the interest in CYP inhibitors and targeted protein degraders, we anticipate this system will allow many groups to easily evaluate new chemical entities for regulation of CYP19A1 activity and protein levels.

Conclusion

While aromatase inhibitors are essential medicines and resistance to these drugs is a significant concern, there are currently no direct, high-throughput-ready assays for CYP19A1. The assay detailed here provides information on multiple features of interest for medicinal chemists working on aromatase. It has not escaped our notice that this system could also be used to better understand endogenous conditions that impact CYP19A1 function and stability.

Experimental Section

CYP19A1 cloning:

To generate the CYP19A1 construct, the gene for CYP19A1 variant 2 was purchased from Origene, and PCR amplified with primers containing the restriction sites for HindIII and EcoRV. The PCR product and pcDNA4 T/O were purified, digested with HindIII and EcoRV followed by agarose gel purification. The purified plasmid and CYP19A1 were ligated at room temperature, transformed into Top10 competent cells, and selected on LB agar plates containing 50 μg/ml carbenicillin. Colonies were screened for CYP19A1 by PCR and confirmed by sequence analysis.

To incorporate Dendra2 at the 3’ end of CYP19A1, the gene for Dendra2 was amplified using a primer set containing the sequence for the EcoRV restriction site followed by nucleotides to create a 6x glycine linker with the other primer containing the sequence for the XbaI restriction site.

Cytochrome P450 oxidoreductase (POR) was placed under the control of the TetO₂ inducible CMV promoter, restriction sites for AgeI and BsiWI were incorporated into the CYP19A1 plasmids using Quikchange site directed mutagenesis. The gene for POR was purchased from Origene and amplified by PCR with primers containing the restriction sites for KpnI and XhoI. The gene for POR was then ligated into pcDNA4 T/O, and plasmids containing POR verified by sequence analysis. Primers were then used to amplify pcDNA4 T/O-POR from the region upstream of the CMV enhancer sequence and downstream of the polyA signal with primers containing AgeI and BsiWI restriction sites. The resulting fragment was ligated into the CYP19A1 pcDNA4 T/O plasmids and the incorporation of POR confirmed by sequence analysis.

CYP19A1 cell lines:

To ensure attachment and uniform distribution of the HEK293 T-REx cells, 5 ml of Geltrex was thawed and added to 500 ml of DMEM and this solution was used to coat surfaces for cellular attachment. Cells were plated in DMEM media supplemented with 10% FBS and 100 U penicillin and 100 μg/ml streptomycin at 450,000 cells per well on Geltrex coated 6 well dishes, and transfected the following day using Lipofectamine 2000. In brief, 32 μl of Opti-MEM was added to 1 μg of plasmid while 32 μl of Opti-MEM was added to 2 μls of Lipofectamine 2000. The solutions were mixed after a 5 min incubation at room temperature and added to the cells after 25 min. The cells were transfected for 5 hrs, followed by removal of media and the addition of DMEM supplemented with 10% FBS and Pen-Strep. After 48 hrs the cells were transferred to Geltrex coated T75 flasks and grown in the presence of 500 μg/ml zeocin and 7.5 μg/ml blasticidin. Cells were maintained as a stable pool in DMEM media supplemented with 10% FBS, 100 U penicillin, 100 μg/ml streptomycin, 500 μg/ml zeocin, and 7.5 μg/ml blasticidin.

Immunoblotting:

To verify inducible expression of CYP19A1, cell lysates were prepared and immunoblotted. Cells were seeded at 500,000 cells/well in 6 well dishes and grown for 24 hrs in the presence or absence of 1 μg/ml tetracycline. Media was removed, followed by the addition of 1 ml ice cold PBS. The cells were then scraped, and pelleted at 3500 rpm for 5 min at 4 °C. Following detergent lysis, the protein concentration for each sample was determined by BCA. Protein lysates aliquots of 8 μg were loaded per lane on a 4–12% bis-tris gel and separated by electrophoresis, followed by a 1 hr transfer to nitrocellulose at 100 V. The membrane was blocked for 1 hr with PBST (PBS with 0.1% Tween 20) containing 5% non-fat milk. Primary antibodies for CYP19A1, POR, and GAPDH were purchased from Santa Cruz Biotechnology and used at a 1:1000 dilution. Following an overnight incubation of the membrane with primary antibody at 4 °C, the membrane was washed with PBST and incubated with HRP conjugated secondary antibody (Jackson Labs) at a 1:5000 dilution in PBST containing 5% nonfat milk. After a 1 hr incubation at room temperature, the membrane was washed with PBST and developed with Clarity ECL (Bio-Rad laboratories). The immunoblots were imaged with a ChemiDoc (Bio-Rad laboratories).

19A1 in cell activity assay:

Cell lines were seeded onto Geltrex coated 96 well plates at 40,000 cells/well, and grown overnight in DMEM media containing 1 μg/ml tetracycline. The media was then replaced with Opti-MEM supplemented with 2% FBS and 1 μg/ml tetracycline. Compounds were serially diluted in Opti-MEM supplemented with 2% FBS and 1 μg/ml tetracycline and an equal volume of compound added to the cells. Following a 1 hr incubation, dibenzylfluorescein was added to the cells to give a 1 μM concentration. Time points were taken over a period of 24 hrs using a Spectrafluor Plus plate reader (Tecan) with an excitation wavelength of 480 nm and emission wavelength of 530 nm. For systems where base was used in a second step, sodium hydroxide was added to the 96 well plates to bring the solution to a 2 M concentration. The plates were then incubated at room temperature for 5 min and fluorescence measured.

Based on the comparison of CYP19A1-expressing cells to their parental cell lines, a signal-to-noise ratio of 7.8:1 was calculated with 1 μM dibenzylfluorescein with the plate reader set to a gain of 85. To determine the limit of blank and limit of detection, the plate reader gain was lowered to capture the full range of signal as the concentration of dibenzylfluorescein was increased. From these data, a limit of blank of 1115 RFU and a limit of detection of 1348 RFU were found based on previously reported calculations.^[19]

In order to evaluate an assay, it is desirable to have a dimensionless parameter to assess the quality of the response, which is defined by the size of the signal window. The z factor (or z’; the screening window coefficient)^[14] provides this by quantifying the separation band, which is the difference between the negative control (background) and positive control (signal). This band is defined as the separation between the third standard deviation of the average signals for each condition. Using the third standard deviation results in the difference between each signal at the 99.73% confidence limit. The equation used to calculate the z factor for the assay is:

$$Z\text{}factor=1-\frac{(3{\sigma }_T-3{\sigma }_{NT})}{\left|{\mu }_T-{\mu }_{NT}\right|}$$

Where ${\sigma }_T$ = standard deviation for the positive control (CYP19A1 induced with tetracycline); ${\sigma }_{NT}$ = standard deviation for the background signal (no CYP19A1 expression, as no tetracycline was added); ${\mu }_T$ = mean with tetracycline; ${\mu }_{NT}$ = mean in the absence of tetracycline.

Washout experiment:

Cells were seeded in 96 well plates as described above and allowed to grow overnight. The media was then removed and replaced with L-15 media supplemented with 2% FBS. Anastrazole and exemestane were incubated with the cells for 90 min. The media was removed and the cells washed 3 times with L-15 media, followed by the addition of L15 containing 12 μM cycloheximide. Dibenzylfluorescein was added and turnover measured over 18 hrs.

CYP19A1 imaging:

The HEK293 T-REx CYP19A1-Dendra2 + POR cells were plated at 80,000 cells per 35 mm Geltrex coated glass bottom dish. The cells were grown overnight in DMEM containing 10% FBS with 1 μg/mL tetracycline at 37 °C with 5% CO₂. Cells were washed and stained with an endoplasmic reticulum-selective dye (ER-ID) and the nuclear dye Hoechst 33342, using the manufacturers recommended protocol for staining live adherent cells (Enzo Life Sciences). Images were captured on a Nikon A1R confocal microscope with a 60x, 1.40 NA, infinity corrected oil immersion objective using excitation wavelengths of 405 nm, 488 nm, and 561 nm for Hoechst 33342, Dendra2, and ER-ID respectively. Each image was captured with a 1024 ×1024 pixel resolution. Hoechst 33342 was visualized on the 405 nm excitation channel with 20% laser power, 152% gain, and 0% offset. Dendra2-green was visualized on the 488 nm excitation channel with 1% laser power, 78% gain, and 0% offset. The ER-ID was visualized on the 561 nm excitation channel with 1% laser power, 33% gain, and 0% offset.

CYP19A1 half-life measurements:

The HEK293 T-REx CYP19A1-Dendra2 cells were seeded at 100,000 cells per 35 mm Geltrex coated glass bottom dish. The cells were grown overnight in DMEM containing 10% FBS with 1 μg/ml tetracycline at 37 °C with 5% CO₂. Media was exchanged for extracellular solution^[20] followed irradiation of the cells with 405 nm light for 1 min using a 405 nm Loctite Flood Array with 144 individually reflectorized LEDs (item No. 1167593; regulated by an LED Flood System Controller, item No. 1359255). Cells were then incubated with 0.3 μM anastrozole, 0.3 μM letrozole, or 3 μM exemestane diluted in opti-MEM supplemented with 2% FBS and 1 μg/ml tetracycline. Imaging was carried out on a Nikon A1R Confocal microscope equipped with a Galvano scanner and GaAsP detectors. Images were captured using a 10x, 0.45 NA, infinity corrected air objective and a 60x, 1.40 NA, infinity corrected oil immersion objective. Dendra2-green was visualized on the 488 nm excitation channel with 24% laser power, 20% gain, and 0% offset. Dendra2-red was visualized on the 561 nm excitation channel with 75% laser power, 60% gain, and 0% offset. Images were captured after 0, 8, 12, 20, and 24 hours. For each field of view 1024 × 1024 pixel images with 2.4 μs pixel dwell were captured. Images were converted into separate 8-bit greyscale images for quantitation of mean fluorescent intensity using ImageJ. Plating density allowed for the imaging of thousands of cells for each timepoint. Data was plotted in Prism GraphPad 9.0.0.

CYP19A1 and CYP19A1-Dendra2 cells were seeded at 500,000 cells/well in 6 well dishes and grown for 24 hrs in the presence of 1 μg/ml tetracycline. Media was removed, followed by the addition of opti-MEM supplemented with 2% FBS and 1 μg/ml tetracycline containing 10 μM Cycloheximide. After 10 min, CYP19A1 inhibitors were added to the wells with a final concentration of 0.3 μM for anastrozole and letrozole and 3 μM for exemestane. The cells were harvested after incubation periods of 8, 16 and 24 hrs and processed as described for immunoblotting.