Synthesis of Melatonin-Tamoxifen Drug Conjugates

The synthetic approach toward the drug conjugates C2, C4, C5, C9, and C15 is shown in Supplemental Fig. 1. Briefly, N-desmethyltamoxifen (Gryder et al., 2013) was N-alkylated with Br(CH₂)_nCO₂Me (n = 2, 4, 5, 9, and 15). The products were subjected to ester hydrolysis, and the resulting acids were coupled with N-desacetylmelatonin (Chatterjie et al., 2001) to give the final drug conjugates. Detailed experimental procedures for the synthesis of C4, including full analytical characterization by ¹H NMR, ¹³C NMR, and mass spectrometry (MS), and the NMR spectra of all five drug conjugates are shown in Supplemental Figs. 1–5.

Competition Binding for Melatonin and ERs

Binding affinities of each of the drug conjugates (C2, C4, C5, C9, and C15) or melatonin, tamoxifen, or 4-OH-tamoxifen alone were measured using 2-[¹²⁵I]-iodomelatonin, [¹²⁵I]-estradiol, and [³H]-estradiol, respectively, as described previously with modification (Witt-Enderby and Dubocovich, 1996; Kirker et al., 2013; Suofu et al., 2017). In brief, experiments were conducted on the mouse uterus (Witt-Enderby et al., 2014) and in whole cell lysates prepared from human MT1 Chinese hamster ovary cells or MCF-7 BC cells grown to confluence on 10-cm plates. Cell lysates were prepared by first washing cells with 5 ml PBS and then lifting them into the buffer (10 mM KPO₄ and 1 mM EDTA, pH 7.4). Cells were then pelleted by centrifugation (277g for 5 minutes) and then resuspended in Tris-HCl buffer (50 mM, pH 7.4). Next, cells were added to tubes containing 115 pM 2-[¹²⁵I]-iodomelatonin (2200 Ci/mmol; PerkinElmer, Waltham, MA) in the presence of melatonin (1 pM to 100 μM) or drug conjugate (1 pM to 100 μM) or the absence of melatonin or drug conjugate (total binding). For 4-OH-tamoxifen or drug conjugate binding in MCF-7 cells, 2 nM [³H]-estradiol was used in the presence of 4-OH-tamoxifen (1 pM to 100 μM) or drug conjugate (1 pM to 100 μM). The incubation volume was 260 µl. Next, cells were incubated for 1 hour at room temperature and then terminated by the addition of ice-cold 50 mM Tris-HCl solution and rapid filtration over glass-fiber filters (0.22 μm; Schleicher & Schuell, Keene, NH) presoaked in 0.5% polyethyleneimine solution (v/v) (Sigma-Aldrich, St. Louis, MO). Each filter was washed twice with 5 ml cold buffer. Radioactivity of [¹²⁵I]-iodomelatonin and [³H]-estradiol was determined in a gamma counter and a scintillation counter, respectively.

Total Binding Assay

G protein–coupled estrogen receptor (GPR30) total binding to C4 and C5 in MCF-7 cells was assessed using [³H]-estradiol as described (Kirker et al., 2013 with modification) on MCF-7 cell lysates. Briefly, 200 μl cell lysate was added to tubes containing 3 nM [³H]-estradiol (PerkinElmer) in the absence (total binding) or presence of the following: 10 μM 4-OH-tamoxifen, 1 μM C4, 1 μM C5, and/or 1 μM G1 GPR30 agonist (Cayman Chemical Company, Ann Arbor, MI). The whole incubation volume was 260 µl. Binding data, initially expressed as moles per milligram protein, were repeated three times and subjected to one-way ANOVA.

Cell Culture

MCF-7, MMC, MDA-MB-231, and BT-549 cells were obtained from American Type Culture Collection (Manassas, VA) and were cultured in Dulbecco’s modified Eagle’s medium (DMEM)/F12 containing 10% fetal bovine serum (FBS) and 1% penicillin/streptomycin at 37°C in a humidified atmosphere with 5% CO₂. BT-549 cells were cultured in the same condition (10% FBS and 1% penicillin/streptomycin at 37°C in a humidified atmosphere with 5% CO₂) using RPMI-1640 medium.

Development and Characterization of TamR MCF-7 Cells

The characterization of TamR MCF-7 cells is described in Supplemental Fig. 6. In brief, MCF-7 cells, obtained from American Type Culture Collection, were cultured at 37°C in a humidified atmosphere with 5% CO₂ for 6 months in DMEM (phenol red-free)/F-12 containing 10% charcoal-stripped FBS and 1% penicillin/streptomycin and 100 nM 4-OH-tamoxifen; 6 months was the amount of time required for MCF-7 cells to become resistant to tamoxifen. For growth assays, wild-type (WT) and TamR MCF-7 cells were counted using the trypan blue exclusion assay after exposure to 100 nM 4-OH-tamoxifen for 1–7 days. For immunocytochemistry images, MCF-7 and TamR cells were treated with either vehicle or 100 nM 4-hydroxytamoxifen for 24 hours. After treatment, cells were fixed in 4% paraformaldehyde for 15 minutes. Cells were then permeabilized with 0.3% Triton-X followed by the addition of rabbit Ki-67 (1:1000, catalog no. 9129S; Cell Signaling Technology, Danvers, MA) and mouse ∝-tubulin (1:1000, catalog no. 3873S; Cell Signaling Technology) primary antibodies. Goat anti-mouse Alexa Fluor 488 nm (1:1000, catalog no. A-11001; Life Technologies, Carlsbad, CA) and goat anti-rabbit Alexa Fluor 555 nm (1:1000, catalog no. A-21429; Life Technologies) were used as secondary antibodies. A Hoechst (Fisher) stain was used to visualize the nucleus. Images were obtained with the EVOS fluorescent inverted microscope (Life Technologies) under 20× objective (scale bar = 200 μm). Basal expression of pERK5, pERK1/2, ER-α, NF-κB, and pAKT in TamR MCF-7 and MCF-7 cells was measured by Western blot. Bands were quantified using Image Studio Lite software (LI-COR Biosciences, Lincoln, NE) and normalized against β-actin (see the “Western Blot” section for further details).

Ex Vivo Treatment of PDX Tumors

The TU-BcX-4QAN PDX tumor was established from a mastectomy specimen of an African American woman and represents a drug-resistant TNBC tumor. The PDX tumor was propagated and maintained in SCID/Beige immunodeficient mice (CB17.cg-PrkdcscidLystbg/Crl) obtained from Charles River (Wilmington, MA). After two serial transplantation passages in mice, the tumor was removed and small tumor pieces were dissected (∼3 × 3 mm²). The tumor pieces were kept intact and placed in individual wells of a 24-well plate. Drug solutions of DMSO/PBS vehicle control, melatonin (10 µM), tamoxifen (10 µM), melatonin plus tamoxifen (10 µM of each), C4 (10 µM), and C5 (10 µM) were made and added to the intact tumor pieces (1 ml/well) in triplicate. After 24 hours of incubation at 37°C in 5% CO₂, tumor pieces were removed and placed in 1.5-ml centrifuge tubes and stored at −20°C.

In Vitro Treatment of PDX-Derived Cells for Western Blot

TU-BcX-4IC cells were derived from a mastectomy specimen of a Caucasian woman. The TU-BcX-4IC tumor was a metaplastic breast carcinoma that had a TNBC PAM50 subtype. The cells were maintained in DMEM supplemented with 10% FBS, insulin, nonessential amino acids, minimal essential amino acids, antibiotics/antimycotics, and sodium pyruvate at 37°C in 5% CO₂. TU-BcX-4IC cells were plated in T-25 flasks and exposed to serum-free media for 24 hours before treatment. DMSO/PBS vehicle control, melatonin (10 µM), tamoxifen (10 µM), melatonin plus tamoxifen (10 µM of each), C4 hybrid (10 µM), and C5 hybrid (10 µM) were added to the cells in triplicate. After 36 hours of incubation at 37°C in 5% CO₂, images were obtained using brightfield microscopy (40× magnification) and cells were harvested. Adherent cells were harvested from the DMSO and melatonin treatments. Detached cells were harvested in the media from the tamoxifen, melatonin plus tamoxifen, C4, and C5 treatments.

Cell Viability

The MTT [3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide] assay method was used to determine cellular viability after exposure to each of the drug conjugates (C2, C4, C5, C9, and C15) per the manufacturer’s instructions. Subconfluent monolayer cell lines grown on 10-cm culture plates were trypsinized and plated equally into each well of a 48-well plate. The next day, cells were exposed to concentrations (1 pM, 1 nM, and 1, 10, and 100 μM) of melatonin, each of the drug conjugates, or the controls tamoxifen (Millipore Sigma, Burlington, MA) or 4-OH-tamoxifen (Millipore Sigma) alone or in combination with the inhibitors for 24 hours. For the inhibitor studies assessing the involvement of mitogen-activated protein kinase kinase MEK1/2, MEK5, or PI3K, PD98059 (Santa Cruz Biotechnology, Dallas, TX), Bix02189 (Santa Cruz Biotechnology), and pictilisib (PI3K inhibitor; SellerChem, Houston, TX) were all added at a final concentration of 10 μM. After 24 hours, 25 μl MTT (dissolved in sterile water; 5 mg/ml; Sigma-Aldrich) was added to each well (0.5 mg/ml final concentration), and the plates were placed in the incubator for 3 hours (5% CO₂ and 37°C). Next, the plates were centrifuged at 50g (37°C for 5 minutes). The entire medium was aspirated, and 250 µl DMSO (Fisher Scientific, Pittsburgh, PA) was added to each well to stop the reaction. The plate was wrapped in aluminum foil and incubated at room temperature for 15 minutes to dissolve the MTT-formazan crystals. Absorbance was measured at 570 nm (VICTOR3 1420 multilabel counter; PerkinElmer).

Cell Migration

Wound/scratch assays were performed in 24-well cell plates to determine the invasiveness of the cells after exposure to each of the drug conjugates (C2, C4, C5, C9, and C15). Twenty-four hours before the assay, cells were plated and allowed to settle. Next, a border was created using a 200-μl pipette tip that was dragged across the bottom of each well to make a visible and uniform space between the cells. The floating cells were aspirated, and the cells were refed fresh media containing the various drug treatments described for the MTT assays above. The border width was measured at baseline (time 0 hours) and then 24 hours after exposure to the treatments (time 24 hours) added alone or in combination with the inhibitors (where indicated) using an EVOS digital inverted fluorescence microscope. Any width changes were expressed as percentages. For the inhibitor studies assessing the involvement of MEK1/2, MEK5, or PI3K, PD98059, Bix02189, and pictilisib were all added at a final concentration of 10 μM. Border widths were calculated as (24 − 0 hours)/0 hours × 100, where a more positive number indicates inhibition of migration and a more negative number indicates stimulation of migration. To calculate the maximum inhibition of TamR MCF-7 cell migration, the percent change in border width was normalized against the minimum inhibition of migration, whereas the minimum inhibition of migration was considered as 0% and no border change was considered as 100%.

Western Blot

To identify potential intracellular targets, C4- or C5-mediated effects on pERK1/2, pERK5, p-AKT, NF-κB, RUNX2, and β1-INTEGRIN proteins were analyzed by Western blot analysis. BC cells were treated with each drug conjugate (1 nM and 10 μM) in the absence or presence of 10 μM PD98059, BIX02189, or pictilisib for 0 minutes (baseline) or 15 minutes, after which the cells were scraped into Laemmli sample buffer containing β-mercaptoethanol, heated for 5 minutes at 95°C, cooled, and then stored at −20°C until use. Western blot analysis was conducted using the Odyssey Western Blotting Kit IV RD (catalog no. 926-31084; LI-COR Biosciences). Equal amounts of each sample (20 μl) and 5 μl molecular weight marker (Precision Plus Protein, catalog no. 161-0373; Bio-Rad, Hercules, CA) were loaded onto 10% gels. Proteins, separated by SDS-PAGE, were transferred to nitrocellulose membranes. Membranes were incubated in blocking buffer (LI-COR Biosciences) for 30 minutes and then incubated with the following respective primary antibodies along with mouse anti–β-actin (catalog no. 926-42212; LI-COR Biosciences) overnight at 4°C: rabbit anti-pERK1/2 and anti-pERK5 (T218/Y220, 1:1000 dilution, catalog no. 3371; Cell Signaling Technology), rabbit anti–phospho-AKT (1:750 dilution, catalog no. 9271; Cell Signaling Technology), rabbit anti–NF-κB (p52, 1:1000 dilution, catalog no. sc-298; Santa Cruz Biotechnology), rabbit anti-Runx2 (1:1000 dilution, catalog no. sc-10758; Santa Cruz Biotechnology), rat ER-α (1:750 dilution, catalog no. sc-53493; Santa Cruz Biotechnology), and rabbit anti–β1-integrin (M-106, 1:1000 dilution, catalog no. sc-8978; Santa Cruz Biotechnology). Next, the blots were washed with PBS containing Tween 20 and then incubated with the following secondary antibodies for 30 minutes at room temperature: goat anti-rabbit (IRDye 800CW, 1:20,000 dilution, catalog no. 925-32211; LI-COR Biosciences) and goat anti-mouse (IRDye 680RD, 1:20,000 dilution, catalog no. 925-68070; LI-COR Biosciences). For the ER-α primary antibody, goat anti-rat (IRDye 800CW, 1:20,000 dilution, catalog no. 925-32219; LI-COR Biosciences) was used for 30 minutes at room temperature. Bands were quantified using Image Studio Lite software (LI-COR Biosciences) and normalized against β-actin to control for differences in protein loading between treatment groups.

Microsomal Incubation Method

Human liver microsomes (Ultrapool HLM 150, catalog no. 452117; Corning, Woburn, MA) and mouse liver microsomes (pooled female MLMs, catalog no. 452702; Corning) were used to assess susceptibility to hepatic metabolism of C4 and C5. Specifically, microsomes were incubated with 5 μM of either C4, C5, or tamoxifen in the presence of an NADPH regenerating system (0.5 mM NADPH, 1 U/ml glucose-6-phosphate dehydrogenase, and 10 mM glucose 6-phosphate) at 37°C for 0, 5, and 10 minutes in mouse microsomes and for 0, 10, and 20 minutes in human microsomes. The longer incubation time was required for human microsomes compared with mouse microsomes because of slow metabolism. Reactions were stopped and precipitated using 50% acetonitrile. Reaction vials were centrifuged at 9500g (4°C for 15 minutes). The rate of drug loss was measured by high-performance liquid chromatography (HPLC)-UV detection. After optimization of the HPLC conditions [column prewash 2 minutes with 70% solvent A (10% methanol in phosphate buffer) and 30% solvent B (100% methanol) for 2 minutes; sample runs = 25 minutes at a gradient flow rate of 0.4 ml/min, where solvent B was changed from 30% to 90%], retention times for C4 (20.46 minutes) and C5 (20.93 minutes) were similar to that of tamoxifen (20.46 minutes). Peak absorbance of each sample was normalized against the internal standard, celecoxib. Celecoxib was used as an internal standard for this study because it gave a retention time (11.51 minutes) distinguishable from the retention times for tamoxifen, C4, and C5 melatonin-tamoxifen drug conjugates.

Pharmacokinetic Analysis of C4 and C5

All original studies using animals were carried out in accordance with the Guide for the Care and Use of Laboratory Animals as adopted and promulgated by the U.S. National Institutes of Health and were approved by the Duquesne University Animal Care and Use Committee or local equivalent (IACUC approval no. 1607-07M1). C57BL/6J female mice (2 months) were used to assess the pharmacokinetic properties of C4 and C5. The drugs were initially dissolved in ethanol [5% (v/v)] and then added to hydroxypropyl methylcellulose (5 mg/ml in ddH₂0) to obtain a final concentration of 0.5% (w/v). C4 and C5 were each administered separately by two routes of administration—subcutaneous and oral. The outcomes were compared with the subcutaneous administration of tamoxifen, which served as the control group. One group of mice (n = 8) received a subcutaneous dose of 1 mg/kg body weight, whereas another group received an oral dose of 10 mg/kg body weight. Different doses for the two routes were given to achieve similar exposures for the two routes based on experience with tamoxifen (Reid et al., 2014). For both routes of administration, blood was collected at similar time points (0, 30, 60, 120, 180, 360, 480, and 1440 minutes) except that one earlier time of 15 minutes was also taken for the subcutaneous route. One C56BL/6J female mouse aged 2 months was used for each time point. Blood was collected in a K₂-EDTA–coated tube by decapitation after CO₂ euthanasia. Blood was then centrifuged at 1166g (4°C for 5 minutes) to separate the plasma. The plasma was then stored at −80°C until liquid chromatography (LC)–MS was performed.

LC–Tandem MS Method for Quantification of Tamoxifen, C4, and C5 in Mouse Plasma

Deuterated tamoxifen (TAM-d₅) was purchased from Toronto Research Chemicals (Toronto, ON, Canada). Mouse plasma in Li- Heparin was purchased from Lampire Biologic Laboratories. Acetonitrile (HPLC grade), isopropanol, hexane, and formic acid (MS grade) were purchased from Fisher Chemicals. Individual stock solutions (20 μg/ml) of tamoxifen, C4, and C5 were prepared in methanol and stored at −20°C until use. The standards for the calibration curve were prepared by performing serial dilution of working stock solutions (64 ng/ml) of tamoxifen, C4, and C5 in methanol and then dividing them into 125-μl aliquots. TAM-d₅ was used as the internal standard. Working stock solutions (50 ng/ml) of TAM-d₅ were prepared in methanol, divided into 20 μl aliquots, and stored at −20°C until use.

Calibration standards (0.25, 0.5, 1, 2, 4, 8, 16, and 32 ng/ml) of tamoxifen, C4, or C5 were prepared in 115 μl mouse plasma and then spiked with TAM-d₅ (2 ng/ml final concentration) to quantify all of the compounds during the extraction process. Quality control (QC) samples of 0.1 ng/ml (the lower limit of quantification, LLQ), 2 ng/ml (QC low), and 32 ng/ml (QC high) were run in parallel and in triplicate for each experimental run of tamoxifen, C4, and C5. Before the extraction process, 10 μl of 1 M NaOH was added to 200 μl of each sample followed by vortexing for 30 seconds. Next, 1000 μl of extraction solvent [hexane/isopropanol, 95:5 (v/v)] was added to each tube and then placed on a rotary shaker for 5 minutes. Samples were then centrifuged for 5 minutes at 9500g to separate the aqueous phase (bottom layer) from the organic phase (top layer). Next, 900 μl supernatant was transferred to a clean test tube. The remaining aqueous layer in the original tube was subjected to the same extraction process by adding another 1000 μl extraction solvent. The total transferred supernatant (organic phase) was air dried under N₂ flux. The dry residue was dissolved in 200 μl methanol and subjected to LC with MS detection.

HPLC followed by tandem MS was conducted using a system model 1200 and triple quadrupole mass spectrometer (model 6460) with a jet stream electrospray ion source (Agilent Technologies, Santa Clara, CA). Data acquisition and chromatographic peak integration were performed using Agilent MassHunter software (versions B02.01 and B02.00). All analyses were performed in positive ion mode and chromatograms were recorded in multiple reaction monitoring scan mode. MS collision parameters for tamoxifen, TAM-d₅, C4, and C5 are shown in Supplemental Table 1, whereas MS acquisition and source parameters for the instrument are summarized in Supplemental Table 2.

Chromatographic separation was achieved with a Hypersil GOLD reverse phase C18 column (50 × 4.6 mm, 3 μm packing; Thermo Scientific, Waltham, MA) thermostated at a temperature equal to 25°C. Deionized water and acetonitrile, both acidified with 0.1% formic acid, were used as mobile phase A and B, respectively, at the flow rate of 0.5 ml/min. A linear gradient separation was used where mobile phase B was initially set at 20% for 1 minute, then increased to 80% over 5 minutes, and then further increased to 95% over 0.1 minutes where it remained for 1.9 minutes. Next, mobile phase B was decreased to initial conditions over 0.1 minutes, and the system was re-equilibrated for 3.9 minutes before the following injection. The total run time was 12 minutes with an injection volume of 40 μl. The injection needle was washed in between analyses using a wash solution of acetonitrile containing 0.1% formic acid (v/v) for 20 seconds. The needle draw and ejection speed was set to 100 μl/min.

Calibration samples and QC samples (described above and performed in triplicate) were used for the determination of linearity, accuracy, and precision. Calibration curves were plotted by correlating the peak area ratio of each compound (relative to TAM-d₅ as the internal standard) as a function of the concentration of spiked standard solutions with a weight factor of 1/y² or 1/relative response². The concentration range (0.25–32 ng/ml with an LLQ of 0.1 ng/ml) for the calibration standards was chosen based on the level of concentration expected in the unknown samples. The concentration of the QC samples was set to 0.1 ng/ml (LLQ), 2 ng/ml (QC low), and 32 ng/ml (QC high) for tamoxifen, C4, and C5 based on the range of calibration concentrations for each compound.

Middle concentrations of the calibration curve were used to determine matrix effects and extraction recovery. The matrix effect was determined by comparing the peak area in postextracted spiked samples versus the peak area in a standard methanol solution. At least six independent plasma samples with different concentrations were tested for the matrix effect for each compound. Extraction recovery was calculated by comparing the peak area of each extracted sample (spiked standard in the blank matrix) with the postextracted spiked sample at the same concentration; this represented 100% recovery. Recovery was calculated for at least four independent plasma samples with different concentrations for each compound. Matrix effects and extraction recoveries for tamoxifen, C4, and C5 are shown in Supplemental Table 3.

Statistical Analysis

All statistical analyses were performed using GraphPad Prism software (version 6; GraphPad Software, La Jolla, CA) and determined in advance. Data represent the mean ± S.D. unless mentioned otherwise. Data points that were determined to be outliers by the Grubbs test were excluded. The mean IC₅₀ values of cell viability and cell migration were obtained by nonlinear regression analysis [log (inhibitor) versus response (three parameters)] and fit by least sum of squares. For the saturation binding analyses, B_max values comparing WT MCF-7 cells to TamR MCF-7 cells used the unpaired t test. For all biphasic curves, two sites (Fit Ki nonlinear regression model in GraphPad Prism) were used. The total cell number obtained via the MTT analysis comparing WT versus TamR MCF-7 cells was analyzed by two-way ANOVA followed by the Bonferroni post hoc test. For other statistical comparisons, one-way ANOVA followed by the Newman–Keuls multiple comparisons test was performed. Mean differences between treatment groups were considered significant at P < 0.05.