A novel culture method that sustains ERα signaling in human breast cancer tissue microstructures

BC samples were collected at the Lisbon Oncology Hospital (Instituto Português de Oncologia de Lisboa Francisco Gentil – IPOLFG). The use of patient material was approved by the IPOFLG ethics committee and all patients have signed an informed consent form to agree to donate the material for research purposes. All tissues were anonymized before transfer to the laboratory for further processing.

MDA-MB-231 cell line was obtained from the American Type Culture Collection (ATCC). MDA-MB-231 cells were cultured in Dulbecco’s Modified Eagle Medium (DMEM) high glucose and pyruvate medium (Gibco), supplemented with 10% (v/v) fetal bovine serum (FBS, Gibco) and 1% (v/v) Penicillin/Streptomycin (P/S, Gibco) at 37 °C in 5% CO₂. Mycoplasma contamination was routinely checked.

This study was elaborated on treatment-naïve patient-derived BC tissue. The method for processing and culture was successfully applied to 63 female breast tumors (Table 1). Tumor samples were collected during surgery and immediately submerged in phenol red-free DMEM/F-12 (Gibco), supplemented with 1% (v/v) P/S (Gibco) and 10% (v/v) FBS (Gibco). Samples were kept at 4 °C and transported to the laboratory within 1 to 3 h after surgery (Fig. 1a). Sixty-three BC samples were collected, with an average weight of 315 ± 225 mg (Figure S1).

Tissue samples were mechanically dissociated with two surgical scalpels to obtain pieces of 1 to 2 mm of diameter. Subsequently, the minced tissue was resuspended in phenol red-free DMEM/F-12, HEPES medium (Gibco) containing 0.09 U/mL of Collagenase A (Roche), 30 U/mL of Benzonase (Merck Millipore), 10% (v/v) FBS (Gibco) and 1% (v/v) P/S (Gibco). Digestion was performed in an incubator at 37 °C, in a humidified atmosphere containing 5% CO₂. After 12–15 h of enzymatic digestion, tumor fragments (tissue microstructures, average of 1 mm³) were sedimented by centrifugation at 100x g for 5 min at 4 °C and washed with Phosphate-Buffered Saline (PBS; Life Technologies) (Fig. 1a).

Tissue microstructures were entrapped in alginate, employing protocols previously developed by our team [13]. Briefly, tissue microstructures were dispersed in 1 mL of 2% (w/v) of Ultrapure Ca²⁺ MVG alginate (UP MVG NovaMatrix, Pronova Biomedical, Oslo, Norway) dissolved in NaCl 0.9% (w/v). Encapsulation was performed using an electrostatic bead generator (Nisco VarV1, Zurich, Switzerland), with an air flow rate of 10 mL/h, at 5.3 V under air pressure of 1 bar, using a 1.1 mm nozzle. The resulting alginate droplets containing tissue micro fragments (1–2 fragments/droplet) were cross-linked in a 100 mM CaCl₂/10 mM HEPES (pH 7.4) solution for 10 min, washed three times in a 0.9% (w/v) NaCl solution and finally equilibrated in culture medium. Encapsulated tissue microstructures were then transferred into 6-well plates and placed under orbital shaking (100 rpm), in a humidified incubator, with 5% CO₂. Encapsulated tissue microstructures cultures were maintained up to 30 days, with 50% medium exchange every 3–4 days (Fig. 1a). Cultures were maintained in human mammary epithelial cell (HMEC) culture medium: DMEM/F12 phenol red free with 1% P/S (v/v) solution (both from Life Technologies), 5 ng/mL Epidermal Growth Factor (EGF), 10 μg/mL Insulin, 0.5 μg/mL Hydrocortisone, 0.5 μg/mL Transferrin, 0.1 mM Isoprotenol, 0.1 mM Ethanolamine, 0.1 mM O-Phosphoethanolamine, 70 μg/mL Bovine Pituitary Extract (all reagents are from Sigma-Aldrich) and 100 μg/mL Primocin (InvivoGen Europe). Non-encapsulated tissue microstructures were maintained under the same culture conditions. Encapsulated tissue microstructures were assessed for cell viability, architecture, cell populations, ECM deposition, ERα presence and signaling, as described below; the extent of assessment performed for each sample was determined by the initial sample size.

Cell viability was correlated with resazurin reduction capacity (PrestoBlue™ Cell Viability Reagent, ThermoFischer Scientific), according to manufacturer’s instructions. Encapsulated and non-encapsulated samples were incubated for 1 h with PrestoBlue reagent in culture medium, at 37 °C, in a humidified atmosphere incubator, containing 5% CO₂. Medium was sampled in quadruplicate and resazurin reduction evaluated by fluorescence detection (ext/em 560/590 nm) in a fluorimeter (Infinite®200 PRO NanoQuant, Tecan Trading AG). Resazurin reduction was evaluated for 1 month, once a week. Data is represented as fold-change in resazurin reduction relative to the first week of the assay.

Samples were collected after 1 month of culture and alginate capsules were de-polymerized with 50 mM EDTA for 5 min at RT. De-encapsulated tissue microstructures were centrifuged at 300x g, 5 min at 4 °C, washed with PBS, fixed with formol overnight at RT. For paraffin cell-block preparation, the cellular suspension was centrifuged for 5 min, at 1270x g, resuspended in 10% (v/v) buffered formalin (VWR BDH Chemicals, ref. 9713.9010) to which a drop of haematoxylin was added for specimen counterstain, and stored in a 1.5 mL microtube. The remaining supernatants were subjected to a second centrifugation, for 5 min, at 1990x g. The supernatant was discarded and four drops of liquefied HistoGel (Thermo Scientific, ref. HG-4000-012) were added to the pellet. After gentle homogenization with a Pasteur pipette and centrifugation for 2 min, at 1990x g, the sample was placed at − 20 °C for 5 min to solidify. The cone shape solidified sample was removed from the microtube, cut along the meridional section and placed in a biopsy cassette, which was then immerged in a container with buffered formalin to be included in paraffin. After processing, the samples were sectioned and stained with hematoxylin and eosin (H&E) (Dako CoverStainer for H&E equipment, Agilent, Santa Clara, CA, USA). Paraffin blocks were sectioned (3 μm) for H&E and immunohistochemical staining. Immunohistochemistry (IHC) was carried out using standard procedures implemented at IPOLFG; antigen retrieval was done using Cell Conditioning 1 (CC1, Ventana) and tissue staining was performed using an automated IHC/ISH slide staining Ventana BenchMark Ultra (all from Ventana Medical Systems, Inc). Antibodies and details on the protocol used are indicated in Table S1. Histologic analysis was performed by an expert breast pathologist. IHC analysis was performed for cultures derived from BC samples of 18 patients. Due to primary material limitations, E-cadherin, CD45, ki-67, ER and p63 levels were assessed in 8 different samples; vimentin was assessed in 9 and CD31 in 2.

Fibrillar Collagen was assessed by multi-photon microscopy. After 1 month in culture, encapsulated tissue microstructures were collected, fixed in PFA 4% (w/v) in PBS for 30 min, washed thrice with PBS and kept at 4 °C until further analysis. Samples were imaged with two-photon-excited fluorescence (TPEF), second harmonic generation (SHG) and infrared (IR) absorption in a home-made multiphoton microscope [14]. The excitation laser was a Ti:Sapphire at 810 nm and the laser power, at entrance of the microscope, was of 40 mW. Initial tests performed with 100 mW resulted in no observable sample damage. The Illumination objective was an Olympus 25 × 1.05 W. The TPEF signal was collected through a photomultiplier tube (PMT) in backward direction (using a LP410 filter) while IR absorption and SHG (405/25 filter) were collected in forward direction through a Nikon 25 × 1.10 W objective, using a photodetector and a PMT respectively. During acquisition, 3–4 images were averaged to reduce noise.

At day 28–30 of culture, encapsulated BC tissue microstructures were stimulated with 10 nM 17-β-estradiol (Sigma-Aldrich). Three days before 17-β-estradiol challenge, encapsulated tissue microstructures were washed thrice with PBS and were then kept in phenol red-free HMEC medium without insulin, hydrocortisone and EGF, which may trigger activation or phosphorylation of ER [15‐20]. Alternatively, a 50% culture medium exchange was performed by the time of 17-β-estradiol challenge. Control wells were also included, in which only ethanol (17-β-estradiol vehicle) was added to a final concentration of 0.001% (v/v). After 24 h of exposure, encapsulated tissue microstructures were collected and alginate dissolved (as described in section 6 of Materials and Methods). Challenge with 17-β-estradiol was performed in encapsulated microstructures derived from 16 different patients, of which 9 in depleted medium and 7 in complete medium.

Encapsulated tissue microstructures were also challenged with fulvestrant (ICI 182,780), an ER antagonist and degrader [20‐22]. For these experiments, 3–5 days after encapsulation HMEC medium was supplemented with 1 μM fulvestrant (Tocris Bioscience). Twice a week, half volume of culture medium was changed and fulvestrant was replenished to keep a constant concentration. After 2 weeks, samples were centrifuged at 300x g, 5 min at 4 °C, washed with PBS and processed for IHC (as detailed above) or RT-qPCR analysis. Samples for RT-qPCR were stored in RNAlater Stabilization Solution (Roche), according with the manufacturer’s instructions, until further analysis; samples for western blot were snap frozen at − 80 °C. Challenge with fulvestrant was performed in encapsulated microstructures derived from 8 different patients, of which 7 were evaluated by RT-qPCR and 3 by Western Blot.

Tissue microstructures were thawed, total RNA was extracted in a tissue lyser (Precellys Evolution Homogenizer, Bertin Instruments) and purified using the RNAeasy Kit (Qiagen), according to the manufacturer’s instructions. Reverse transcription was performed using Sensiscript RT kit (Qiagen), also according with the manufacturer’s instructions. qPCR was performed in triplicates, using the SYBR green I Master kit (Roche), in a LightCycler 480 II (Roche). We evaluated expression of ERα (ESR1) and its downstream target genes, pS2, AREG and PGR [23], and of two housekeeping genes, RPL22 [12] and 36B4 [23]. Primer sequences are provided in Table S2. Due to the scarcity of ERα negative BC samples, a ERα and PR negative BC cell line, MDA-MB-231, was employed as basal expressing control [24]. Results are shown as fold change in mRNA amount compared to the vehicle control (CTRL), calculated according to the 2^-ΔΔCt method [25], considering a geometric mean of the 2 housekeeping genes used.

Samples were thawed, resuspended in Laemmli Buffer (20% Glycerol, 4% SDS in 100 mM Tris Buffer, pH 6.8) and lysed in a Tissue homogenizer (Precellys Evolution, Bertin Instruments). BC Microstructure lysates were recovered, sedimented to remove cell debris, sonicated and stored at − 80 °C until use.

Protein quantification was performed in a Nanodrop ND-2000C (Thermo Scientific). Proteins were denatured and loaded in a electrophoresis gel (NuPAGE 4–12% Bis-Tris Gel) under reducing conditions for 50 min (200 V) and then electrophoretically transferred using a wet transfer system (Bio-Rad, 30 V, 18 h, 4 °C) into nitrocellulose membranes. Membranes were blocked for 1 h in TBS with 0.1% (w/v) Tween 20, 5% (w/v) non-fat dried milk and further incubated with the primary antibodies (Mouse anti-Human ERα, 1D5 Clone, Dako, final dilution 1:500; Rabbit anti-β tubulin, H-235, SC-9104, SantaCruz, final dilution 1:1000, used as loading control) and respective secondary HRP-conjugated secondary antibodies (Sheep anti Mouse IgG NA931; Donkey anti Rabbit IgG NA934; GE Healthcare, final dilution 1:20000). Membranes were developed using Amersham ECL Select Western Blot Detection Reagent (GE Healthcare) and visualized using a ChemiDoc System (BioRad).

Statistical analysis was performed using GraphPad Prism version 6.0 (GraphPad Software). Data were analyzed as indicated in the figure legends. The Mann-Whitney test was performed to evaluate statistical difference between conditions. Data are presented as mean ± SD, unless otherwise specified.

To establish an ER+ BC ex vivo model, we investigated the possibility of retaining the TME and consequently ERα signaling of patient-derived tissue microstructures immobilized within alginate capsules and cultured under agitation (Fig. 1a). Encapsulated tissue microstructures were cultured for up to 30 days, showing high cell viability, as indicated by maintenance of resazurin reduction capacity along culture time (97 ± 28% by the end of week 4, relatively to the beginning of the culture, Figure S2a). Moreover, detection of extracellular lactate in culture medium (data not shown), as an indicator of high metabolic activity [26], corroborated the high cell viability within the encapsulated tissue microstructures.

The original tumors were very heterogeneous, not only between but also within patients (Fig. 1b): tissue architecture varied in epithelial versus stromal content, cell organization and on the presence/absence of immune cells (CD45+ cells). A complete mixture of malignant epithelial cells and stromal cells was rarely observed. Instead, there were islets of tumor cells surrounded by multiple stromal cells (Fig. 1b, upper panels). These histopathological characteristics were maintained in encapsulated tissue microstructures cultured for a month (Fig. 1b, lower panels). By day 30 of culture, E-cadherin, vimentin, CD31 and CD45 were immunohistochemically-detected (Fig. 2a). The detection of membranous E-cadherin indicated that carcinoma cells maintained the typical cell-cell adhesions and differentiated phenotype [27]. On the other hand, vimentin detection confirmed the presence of stromal cells. CD45, also known as leucocyte common antigen, is a transmembrane glycoprotein present in all nucleated cells of the hematopoietic lineage [28] and has been broadly used to assess immune cell population presence in breast tissue, such as tumor-infiltrating lymphocytes [29‐31]. CD45⁺ cells were detected in 5 out of the 8 cases which presented immune cells in the original tissue (Fig. 2a). In two analyzed tissue microstructures, CD31 positivity confirmed the presence of endothelial cells (Fig. 2a). Absence of cells positive for the basal/myoepithelial marker p63 was observed similarly to the original tumors (Figure S2b). Ki67-positive cells were also detected at different levels, indicating the presence of proliferating cells even after 1 month of culture (Fig. 2b). Although at low levels, this is consistent with the parental tissues, where the median of proliferating cells was 20% (Q1 = 15; Q3 = 30). Second harmonic generation analysis (SHG) of encapsulated BC tissue microstructures revealed dense and organized/fibrillar collagen fibers in peripheral regions of the samples analyzed, surrounding areas of cellularity (Fig. 3). As a culture control, non-encapsulated tissue microstructures were cultured in parallel. A significant decrease in resazurin reduction ability after 3–4 weeks of culture was observed, suggesting a reduced cell viability of these cultures. Remarkably, cell viability was increased in encapsulated versus non-encapsulated tissue microstructures (Figure S2c).

Altogether, we were able to extend the lifespan of BC explant cultures for up to 1 month whilst maintaining tissue architecture, the different cell types of the BC microenvironment, and cell viability.

After 1 month in culture, ER+ carcinoma cells were still detected in the encapsulated tissue microstructures by IHC analysis (Fig. 4a), typically in lesser extent that in the original sample. When sample material was not sufficient for IHC evaluation, mRNA was quantified, relatively to MDA-MB-231, a human cancer cell line which does not express ERα nor PR [24] (Figure S3a). All samples presented higher expression of the ERα gene (ESR1) than MDA-MB-231 cells, indicating ERα gene expression after 1 month of culture (Figure S3b and Figure S5).

To assess ERα function, encapsulated tissue microstructures derived from ER+ BC from 9 distinct patients were stimulated with 10 nM 17-β-estradiol for 24 h and the mRNA levels of the ER target genes evaluated: protein PS2 (also known as Trefoil Factor 1 -TFF1-, pS2), progesterone receptor (PGR) and amphiregulin (AREG) [23]. AREG and PGR were upregulated upon challenging with 17-β-estradiol compared to vehicle-controls (mean fold increase in AREG and PGR expression of 3.4 ± 5.6 and 6.3 ± 11 respectively, Fig. 4b, Figure S4a). Strikingly, we detected a generalized upregulation of pS2 (in 7 out of 9 tissue microstructures), with a mean fold increase in gene expression of 45 ± 45, compared to the vehicle-treated control (Fig. 4b, Figure S4a and Figure S5). In general, there was a trend for a positive correlation between ESR1 basal expression and the expression of the three ER target genes upon estrogen challenge (Figure S4b, R > 0 by Pearson Correlation), even though not significant, probably due to the intrinsic variability of primary tumors and the sample size. Tissue microstructure cultures derived from ER-negative BC tumors were also treated with 17-β-estradiol and no upregulation of ER downstream genes was observed (Figure S4c). This data further corroborates that the original phenotype was maintained in culture. A different set of encapsulated tissue microstructures, derived from 7 ER+ BC samples, were maintained in HMEC medium until 17- β-estradiol challenge. These showed a mild stimulation of ERα target genes (on average, 2.3-, 1.8- and 1.2-fold increase relatively to vehicle control for pS2, PGR and AREG, respectively, Figure S4d).

To further confirm intact ERα signaling in encapsulated tissue microstructures, cultures derived from additional ER+ BC samples were exposed to fulvestrant (or ICI 182,780), a ERα full antagonist [32] widely used in endocrine therapy [33]. After 2 weeks of exposure, a generalized down regulation of AREG compared to vehicle controls was observed (Fig. 4c). For two of the tumors, we also evaluated PGR and pS2 response and observed a strong reduction of mRNA levels compared to vehicle controls (Fig. 4d). Additionally, we assessed ERα protein levels in three of the tumors and observed a tendency for reduction compared to vehicle control conditions (Figure S6). Collectively, these results indicate that ERα is expressed in encapsulated tissue microstructures derived from ER+ BC samples and can respond to stimulation and inhibition.