Mice

129 Sv female mice transgenic for the activated rat neuT oncogene, which spontaneously develop multiple mammary tumors, were generated as previously described [25, 27]. The BALB-neuT strain originally used for the backcrosses originated from a transgenic CD1 random-bred breeder male mouse (no. 1330) carrying the mutated rat HER-2/neu oncogene driven by the MMTV promoter [24, 25]. All mice were housed in the animal facility at Istituto Superiore di Sanità in accordance with the European Community Regulation (https://eur-lex.europa.eu/eli/reg/2019/1010/oj). At each generation, the presence of the rat HER-2 transgene was routinely checked by polymerase chain reaction (PCR) on tail DNA using primers hybridizing to vector (5-ATCGGTGATGTCGGCGATAT-3) and to MMTV sequences (5-GTAACACAGGCAGATGTAGG-3). The mammary glands of all transgenic virgin female mice were inspected once a week for tumor appearance. Progressively growing masses with a mean diameter > 1 mm were regarded as tumors. Individual neoplastic masses were measured with a caliper in two perpendicular diameters and tumor volume was calculated according to the formula: V = 1/6 π x D x d x [(D + d)/2] [28], where d and D represent shorter and longer tumor measurements, respectively. Tumor multiplicity was calculated as the cumulative number of incident individual tumors/total number of mice and reported as mean ± standard error. Bio-nFeR freshly dissolved in sterile water was administered by oral gavage 100 mg/kg, 5 days/week, from week 14 to week 32 post-birth (PB), or to euthanasia. Mice were examined twice a week for signs of distress and were euthanized when 10/10 mammary glands developed tumors. All surviving mice were euthanized at week 32 PB for ethical reasons related to repeated gavage procedures. At the end of the experiments, mice were sacrificed by cervical dislocation, and tumors harvested ex vivo were measured and weighed. Fragments were collected from each tumor for subsequent analyses or storage. All the procedures were approved by the Ethics Committee for Animal Experimentation of the Istituto Superiore di Sanità, according to the Italian regulation (DL 4.3.2014 N. 26). For the second transplantation assay, cells were dissociated from individual tumors and pooled for each mouse. Cell pools generated from single mice were then transplanted subcutaneously into secondary recipient NOD.Cg-Prkdc scid Il2rg tm1Wjl /SzJ (NSG) mice. Each treatment group included cell pools generated from six individual mice. For each cell pool, different cell doses (1, 10, 100 and 1000) were tested. Mice were recorded negative when no graft was observed after 24 weeks from the inoculation. CSC frequency was calculated by the extreme limiting dilution analysis software ELDA [29].

Plasma fenretinide determination

The quantification of fenretinide (4-HPR) and its main metabolites, O-methylated (4-MPR), 4-oxo-substituted β-ionone ring (4-oxo-4HPR ), dehydrogenated 4-HPR (DH-4HPR ), was performed according to [30].

Shortly, 30 µL aliquot of plasma was added with 3 ng of deuterated internal standard (2H4 4-HPR), deproteinized by 90 µL of acetonitrile and centrifuged 5 min at 13,200 rpm at 4 °C. The supernatant was recovered and 8 µL injected into a HPLC system (1200 series pump and auto sampler Agilent Technologies, Santa Clara, CA, USA). Chromatographic separation was achieved on a Gemini-C18 column (50 mm × 2.0 mm, 5 μm particle size; Phenomenex Inc., Torrance, CA, USA) at 35 °C, protected with a Security Guard™ ULTRA cartridges C18. The detection, obtained via high-resolution mass spectrometry (HRMS), was carried out on high-resolution LTQ-Orbitrap XL mass spectrometer (Thermo Scientific Inc., Waltham, MA, USA), equipped with an electrospray source (ESI) operating in positive ion mode.

Cell lines

MCF7, generated from a human invasive breast ductal carcinoma, ER-positive [31] and MDA-MB-231, generated from human breast adenocarcinoma, ER- and PR-negative [32] cell lines were purchased from ATCC (Manassas, VA, USA). TUBO, a murine mammary tumor cell line cloned from a BALB-neuT mouse mammary carcinoma [33] was from Sigma Aldrich (St. Louis, MO, USA). Cells were cultured according to the manufacturers’ instruction, and used within the 10th passage.

Antibodies and reagents

Fenretinide (code 65646-68-6) was purchased from Olon (Milan, Italy). L-α-phosphatidylcholine from egg yolk, glyceryl tributyrate, and all the other chemicals were purchased from Sigma-Aldrich. Anti-β-tubulin and anti-β-actin antibodies were purchased from Sigma-Aldrich; Phospho-S6 Ribosomal Protein (Ser240/244) #2215 and S6 Ribosomal Protein (5G10) Rabbit Ab #2217, Caspase-3 Rabbit Ab #9662, Caspase-7 (D2Q3L) Rabbit Ab #12,827, Phospho-p38 MAPK (Thr180/Tyr182) (D3F9) XP® Rabbit Ab #4511, p38 MAPK (D13E1) XP® Rabbit Ab #8690, mTOR Rabbit Ab #2972, Phospho-mTOR (Ser2448) Rabbit Ab #2971, 4E-BP1 (53H11) Rabbit Ab #9644, Phospho-4E-BP1 (Ser65) Rabbit Ab #9451, Bcl-2 (D17C4) Rabbit Ab #3498, ALDH1 (D4R9V) Rabbit Ab #12,035, and Ki67 (D3B5) Rabbit Ab (IHC Formulated) #12,202, Ki67 (D3B5) Rabbit Ab #9129, PCNA (D3H8P) Rabbit Ab #13,110 were from Cell Signaling Technology (Danvers, MA, USA). Anti-CDKN2A (p16, 15C10C30) was from Biolegend (London, United Kingdom). Cyclin D1 (A-12) Mouse mAb #sc-8396, ERK-1 (K-23) Rabbit pAb #sc-94, p-ERK (E-4) Mouse mAb #sc-7383 were from Santa Cruz Biotechnology (Texas, USA). Secondary anti-mouse and anti-rabbit antibodies coupled to horseradish peroxidase were from Bio-Rad (Hercules, CA, USA). Alexa Fluor 647 conjugated anti-Ki67 (cloneB56), FITC conjugated anti CD24 (clone M1/69), PE conjugated anti-CD44 (clone IM7) were from BD Pharmingen (San Diego, CA).

Bio-nFeR preparation

Bio-nFeR lot preparation and quality control are described in detail in [9]. Briefly, FeR was homogeneously mixed with L-α-phosphatidylcholine and glyceryl tributyrate dispersed in alkaline ethanol to a final weight ratio of 1:9:1 w: w:w, respectively. Ethanol was then removed from the mixture by a rotary evaporator and the dry residue was stored at -20 °C until use. Reconstitution of the dry residue to Bio-nFeR nanomicelles was accomplished by dissolving the residue in water at 30 °C in an ultrasound bath with a wave frequency of 40 kHz. The dissolved phase (100 mg/mL) was subsequently filtered through 0.4 μm pore filters to obtain homogeneously dispersed nanomicelles of controlled mean size. Characterization of the Bio-nFeR preparation was performed by multiple assays including fluorescence microscopy (to detect auto-fluorescent nanomicelles containing FeR), dynamic light scattering, drug loading, solubilization, drug release from the nanomicelles over time [9]. Bio-nFeR lot was divided in batches, which were aliquoted and stored at -20 °C. Before use, each Bio-nFeR frozen batch was thawed and diluted with sterile water to the desired concentrations (see Results). As for our routine procedure, the batch was tested after thawing by assessing its biological activity on two freshly thawed lung spheroid lines, in comparison with published reference data [9] (Additional file: Fig. S1).

Viability assay

Cell viability was determined by CellTiter-Glo luminescent cell viability assay (Promega, Madison, WI, USA) according to the manufacturer’s directions. Briefly, cells were detached from flasks by trypsin incubation at 37 °C (Thermo Fisher Scientific) and seeded in 96-well plates (3 × 103/well, three replicates per experimental point), in culture medium. Dishes were incubated in a humidified atmosphere at 37 °C, 5% CO2. Cells were treated with Bio-nFeR at different concentrations (from 5µM to 100µM) as indicated in the Results, and then analyzed after 72 h. Luminescence was detected by a DTX880 multimode microplate reader (Beckman Coulter, Brea, CA, USA).

Flow cytometry and cell cycle

For flow cytometry, cells (either cell lines detached from flasks by short treatment TrypLE Express (Thermo Fisher Scientific), or cells dissociated from primary tumors by mechanical/enzymatic treatment with TrypLE Express), were resuspended in PBS (5 × 105/ml), 0.4% BSA/0.5 M EDTA, and labeled with antibodies (PE-anti CD44, FITC-anti CD24 and Alexa Fluor 647-anti Ki67, see Antibodies and Reagents section in Methods), for 1 h on ice.

For CD44/CD24 detection, the analysis was preceded by depletion of mouse hematopoietic cells with the Mouse Lineage Cell Depletion Kit (Miltenyi Biotec, Bergisch Gladbach, Germany), to obtain lineage-negative (LINneg) cell populations. Marker analyses (CD44/CD24 and Ki67) were performed by a FACSCanto flow cytometer equipped with a DIVA software (Beckton Dickinson). Cell population was gated based on FS (forward scatter) and SC (side scatter) properties, to exclude debris. SSC-A vs. SSC-W were then plotted to exclude doublets, and dead cells were excluded by staining with the viability dye 7AAD (7-amino-actinomycin D, Sigma-Aldrich). The cells within the viability gate were further analysed for marker(s) expression. At least 20.000 single cells were collected per sample.

The cell cycle status of BC cell lines was assessed by staining freshly detached single cells with 50 mg/ml propidium iodide (PI) dissolved in 0.1% trisodium citrate buffer, 9.65 mM NaCl, 0.1% NP40, 200 mg/ml RNAse for 1 h at room temperature. The analysis was performed by a Cytoflex LX flow cytometer (Beckman Coulter), using yellow laser (561 nm). Acquisition was done at a low flow rate under 300 events/second. At least 20.000 single cells/sample were recorded. Cell population was gated by FS (forward scatter) and SC (side scatter) analysis. PI emission was then measured by the bandpass 610/20 nm filter. The gated population was plotted for PI area versus PI width to identify cell doublets and clumps, which were gated out. PI was plotted on a linear scale to clearly distinguish the cell cycle phases. The PI histogram graph of this gated population shows the three distinct phases of the proliferating cell population: G0/G1, S and G2 /M. The percentage of cells in each cell cycle phase was quantified by using markers set within the analysis, which was conducted using the CyExpert software (Beckman Coulter).

Western blotting

Fragments of frozen tissues (~ 50 µg) were lysed in lysis buffer (10 mM Tris pH8, 150 mM NaCl, 60 mM Octyl-β-Glucoside, supplemented with protease inhibitor/phosphatase inhibitor cocktails I and II from Sigma-Aldrich). Tissue homogenization was performed with Pro 200 Kema Keur (Pro Scientific Inc. Oxford) at max speed, at 4°C, for 30”. Lysate’s concentration was determined by Bradford assay (Bio-Rad Laboratories, Hercules). Equal amounts of proteins were run on a 4–12% precast gel (Thermo Fisher Scientific) and then transferred to nitrocellulose membranes (GE Healthcare Life sciences). Blots were blocked with TBST 5% non-fat dry milk (Bio-Rad Laboratories, Hercules, CA, USA) and incubated overnight at 4 °C with primary antibodies (described in the Antibodies and Reagents section) diluted in TBST/BSA 5%. After three washes in TBST, blots were incubated for 45 min with specific secondary HRP-conjugated antibodies dissolved in TBST, 5% BSA. Chemiluminescent signals were detected with Amersham ECL Prime or Select western blotting detection reagent (GE Healthcare Life Sciences, Barrington, IL, USA). Immunoblotting images were recorded and analyzed by Bio-Rad ChemiDoc Imagers (Bio-Rad Laboratories, Hercules). Immunoblot densitometry quantification was performed by ChemiDocMP (Bio-Rad Laboratories, Hercules) and signal intensity was quantified with the Image Lab software. Normalization was performed using antibodies against β-actin or β-tubulin (both from Sigma-Aldrich) as reference standards. The antibodies used for Western blotting are indicated in the antibodies and reagents section.

Migration/Invasion assay

Single cells (MCF7, TUBO, and MDA-MB-231, 1.5 × 104/experimental point) were suspended in 200 µl of serum-free medium and plated in Matrigel® into the upper wells of Boyden Chambers containing porous 8 μm diameter polycarbonate membranes (Costar Scientific Corporation). Lower wells contained 500 µl of complete medium (10% FBS). Bio-nFeR (20µM, 40µM and 80 µM) was added into the upper wells. After 48 h, upper wells were removed, and cells migrated to the lower wells were fixed in 4% paraformaldehyde (PFA), stained with DAPI in PBS 1% NP40 for 5 min, and counted under a fluorescence Zeiss Axio Scope.A1 microscope equipped with a 10x objective. The number of migrated cells was quantified by the ZEN 2.6 software (blue edition).

Scratch test

MDA-MB-231, TUBO and MCF7 cells were seeded into 6-well tissue culture plates, and incubated at a density optimized to reach confluency after growing overnight at 37 °C. The following day, confluent cell monolayers were scraped in a straight line with a p200 pipet tip to create the scratch. Debris were removed by washing the cells with PBS and then replicate wells were filled with 4 mL of fresh medium, in the presence or the absence of 20 µM Bio-nFeR. Cell migration was monitored by time-lapse imaging using a Zeiss LSM900 confocal microscope. For MDA-MB-231 cell line, images were taken at a 10 min time intervals up to 24 h. For MCF7 and TUBO cells, images were taken at a 20 min time intervals up to 48 h. Three images for each time point were then analyzed with ImageJ by using a specific plugin (https://github.com/AlejandraArnedo/Wound-healing-size-tool/wiki), which calculates the average distance (width) between the edges of the scratch on each image. Data represent the mean percentage of width change over time ± SD, on the three images taken for each time frame.

Colony formation assay

Clonogenic units present in xenograft-dissociated cells were assessed by plating 500 cells/ml per well in triplicate in 24-well plates containing a soft agar bilayer (0.3% top and 0.4% bottom layer; SeaPlaque Agarose; Cambrex). Cultures were incubated in humidified atmosphere at 37 °C and 5% CO2 for 21 days. Colonies were stained with crystal violet (0.01% in 10:1 methanol to water), and counted under a light microscope. Data represent the percentage of colonies normalized to the number of single cells counted the day after plating.

Histology

Tumors and lungs ex vivo were fixed for 18–24 h in 10% buffered formalin immediately after removal. After fixation, samples were processed, and embedded in paraffin, sectioned and stained by H&E. To evaluate metastases, each lung was put in a separate tissue cassette, and then 3 sections, 2 μm thick at 2 mm intervals, were cut from every sample. H&E-stained slides were then scanned with Aperio C2 Pathology Scan (Leica) to obtain high quality digital slides. Metastases were then counted and measured. Digital pictures (1x) were obtained by using Aperio ImageScope program.

Immunohistochemistry

For immunohistochemistry, 5 μm sections were deparaffinized in xylene and rehydrated in a series of graded ethanol washes. Endogenous peroxidase activity was blocked with 3% H2O2 for 10 min. Antigen retrieval was performed by incubation with citrate buffer (pH 6.0) for 20 min at 100 °C followed by cooling at room temperature for 30 min. Slides were then incubated with anti-Ki67 (#9129, Cell Signaling Technology) and anti-PCNA (Proliferating Cell Nuclear Antigen) (#13110, Cell Signaling Technology) antibodies overnight at + 4 °C and for 1 h at 37 °C, respectively. Immunoreactions were detected with EnVision Detection SystemsPeroxidase/DAB, Rabbit/Mouse kit (#K5007, DAKO, Agilent Technologies), following manufacturer’s instructions. Slides were counterstained with Mayer’s hematoxylin (#MHS32, Sigma Aldrich), dehydrated, and mounted with Canada balsam mounting medium (#C1795, Sigma Aldrich). Images were acquired with a Zeiss Axio Scope.A1 Microscope equipped with 5x and 20x objectives. Image analysis was performed using the software ZEN 2.6 (blue edition). The proliferation index was calculated within the region of interest (ROI) by computer-assisted imaging with the ZEN 2.6 software (blue edition), as the ratio between the pixel count for Ki67- or PCNA-positive nuclei, over the pixel count for total nuclei staining (hematoxylin-positive).

Immunofluorescence and confocal analysis

Paraffin-embedded section of primary tumor tissue samples were deparaffinized, rehydrated and treated with citrate buffer as described in the previous paragraph. Slides were then quenched with 1 M glycine in PBS and incubated overnight at 4 °C with primary antibody anti-Ki67 (#9129, Cell Signaling Technology). After washing in PBS, sections were incubated with secondary antibodies (see Antibodies and Reagents section) for 45 min at room temperature in the dark. Terminal deoxynucleotidyl transferase dUTP nick end labeling (TUNEL) assay was performed using in situ cell death detection KIT fluorescein (Roche-12156792910) following the manufacturer’s instructions. Nuclei were counterstained with DAPI (Invitrogen) for 15 min at room temperature. Slides were permanently mounted with Prolong-Gold Antifade (Thermo Fisher) and analyzed using a Zeiss LSM900 Confocal microscope equipped with a 60x oil immersion objective.

Human MCF7, MDA-MB-231 and murine TUBO cells were seeded into multi-wells on poly-L-lysine-coated coverslips and treated for 24 h with Bio-nFenR at different concentrations for each line, as indicated in the Results. After washing with PBS, cells were fixed with PFA 4% for immunofluorescence analysis. Staining with anti-Ki67 antibody was performed as described above. Nuclei were counterstained with DAPI (Invitrogen) for 15 min at room temperature. Slides were permanently mounted with Prolong-Gold Antifade (Thermo Fisher) and analyzed using a Zeiss Axio Scope.A1 Microscope equipped with 20x objective.

The proliferation index was calculated within the region of interest (ROI) by computer-assisted imaging with the ZEN 2.6 software (blue edition), as the ratio between the pixel count for Ki67-positive nuclei, over the pixel count for total nuclei staining (DAPI-positive).

Statistical analysis

Statistical analyses were performed using GraphPad Prism version 4.0 for Windows (GraphPad Software). Statistical significance is expressed as *, p < 0.05, **, p < 0.01 and ***, p < 0.001. Results are presented as the mean ± SD or mean ± SEM where appropriate. Unpaired Student’s t-test was used for group comparison. IC50 was calculated according to the formula Y = Bottom + (Top-Bottom)/(1 + 10LogEC50X) by GraphPad. The survival test was analyzed by Long-rank (Mantel-Cox) test.