TAX (cat. no. T1912), COL (cat. no. C-9754), DOX (cat. no. PHR1789) and CIS (cat. no. 479306) were purchased from Sigma-Aldrich (Darmstadt, Germany). VCR sulfate (cat. no. SC-201434) was obtained from Santa Cruz Biotechnology (Dallas, TX, USA) and ABT-263 (Navitoclax) (cat. no. A3007) from ApexBio (Houston, TX, USA). SAL sodium salt was obtained from commercially available veterinary premix SACOX^® following acidic extraction using the procedure described previously (28).

Breast tumor tissue was provided by the Cooperative Human Tissue Network (CHTN, http://www.chtn.org/), a National Cancer Institute supported resource. Other investigators may have received samples from these same tissue specimens. Fresh breast cancer tissue was collected from a 33 year old African American female patient diagnosed with 16.3 cm malignant phyllodes tumor of the breast. Immediately after surgical resection, a portion of the specimen was transported and stored in fresh RPMI 1640 medium on ice for 24 h prior to use. The day after surgical removal, slices were prepared from the specimen according to our previously established methodology (27) based on the studies of Van der Kuip et al (29). Briefly, 200 µm slices were cut in sterile, cold PBS (cat. no. 21-030-CM, Corning, Manassas, VA, USA) supplemented with 1% antibiotic/antimycotic solution (cat. no. A5955, Sigma-Aldrich, Deisenhofen, Germany) using a microtome with vibrating blade (Leica Biosystems VT1200, Nussloch, Germany). Blades were steam sterilized before use. Several individual slices from different parts of the specimen were immediately placed into embedding cassettes (cat. no. 27158-2B and cat. no. 27154-1, Ted Pella, Redding, CA, USA) and fixed in 10% neutral-buffered formalin for 24 h and further stored in 70% ethanol prior to processing. The remaining slices were distributed in separate wells of a 24 well plate in 1 ml of Mammary Epithelial Cell Basal Medium (cat. no. C-21215, Promo Cell, Heidelberg, Germany) supplemented with Mammary Epithelial Cell Supplement Pack (cat. no. C-39110, Promo Cell, Heidelberg, Germany), 100 µg/ml gentamicin (cat. no. G1397, Sigma-Aldrich, Deisenhofen, Germany) and 0.05 µg/ml amphotericin B (cat. no. A2942, Sigma-Aldrich, Deisenhofen, Germany). The plate was incubated at 37°C in a constant atmosphere of 5% CO₂ on a shaking platform at 150 rpm (Orbi-Shaker Jr, Benchmark Scientific, Sayreville NJ, USA). After 24 h to allow tissue equilibration, treatment with 0.2 % DMSO or ABT-263, SAL, DOX, TAX, VCR, COL and CIS, at concentrations of 2, 8 and 16 µM was initiated for 72 h. Concentrations were selected based on previous findings of van der Kuip et al (29). The medium was changed every 24 h. Tissue slices were then fixed and stored as described above.

Immunohistochemical staining of untreated, vehicle- and drug-treated samples was performed by UAMS Experimental Pathology Core. For histopathological examination, paraffin embedded sections (4 µm) were stained with hematoxylin and eosin (H&E) (cat. no. 7231, Richard-Allan Scientific, Thermo Fisher Scientific, Waltham, MA, USA). Dako DAB+ (cat. no. K3468, Dako Liquid, Carpenteria, CA, USA) was used for 3 min prior to counterstaining with hematoxylin. Immunohistochemical staining for Ki-67 (1:100, rabbit monoclonal anti-Ki67, cat. no. ab16667, Abcam, Cambridge, MA, USA) was performed with biotinylated goat anti-rabbit second antibody (1:400) (cat. no. BA-1000, Vector, Burlingame, CA, USA) followed by detection using the Vectastain ABC Elite detection system for 30 min (cat. no. PK-6100, Vector, Burlingame, CA, USA). Epitope retrieval (cat. no. S1699, Target Retrival Solution, Dako, Carpenteria, CA, USA) was achieved prior to the staining in a decloaking chamber for 20 min (Biocare medical). Dead or dying tumor cells were quantified manually based on changes in cell and nuclear morphology. Quantification was performed employing a Nikon Eclipse E600 microscope at 4×, 10× and 20×. Specifically, tumor cells which were markedly reduced in size and/or showed nuclear fragmentation were scored. Effects of drugs on normal ductal epithelial cells were similarly assessed by evaluating cellular and nuclear morphology of individual cells as well as the extent to which epithelial organization was disrupted. A total of at least 50 cells were examined per condition each of which was conducted in triplicate.

To establish primary PTB cells, a procedure for conditional reprogramming of cells (CRC) was followed (21). Briefly, the remaining tumor specimen was cut into 1-mm thick slices which were dissociated by incubating in a mixture of 0.25× Collagenase/Hyaluronidase and 0.25× Dispase (cat. no. 7919, cat. no. 7913, Stem Cell Technologies, Vancouver, Canada) diluted in F-medium 3:1 (v/v) F-12 Nutrient Mixture (Ham)/Dulbecco's modified Eagle's medium (cat. no. 11765-054, Gibco, Grand Island, NY, USA), 5% fetal bovine serum (cat. no. FP-0500-A, Atlas Biologicals, Fort Collins, CO, USA), 24 µg/ml adenine (cat. no. A2786-5G, Sigma-Aldrich, Deisenhofen, Germany), 8.4 ng/ml cholera toxin (cat. no. C8052-.5MG, Sigma-Aldrich, Deisenhofen, Germany), 10 ng/ml epidermal growth factor (cat. no. PHG0311, Life Technologies, Waltham, MA, USA), 1× antibiotic/antimycotic solution (cat. no. A5955, Sigma-Aldrich, Deisenhofen, Germany), 0.4 µg/ml hydrocortisone (cat. no. H4001, Sigma-Aldrich, Deisenhofen, Germany), 5 µg/ml insulin (cat. no. 128-100, Cell Applications, Inc., San Diego, CA, USA) for 2 h. Mouse fibroblast conditioned medium was prepared from Swiss 3T3-J2 mouse fibroblasts (cat. no. EF3003, Kerafast, Inc., Boston MA, USA) cultured in DMEM supplemented with 10% bovine calf serum (iron supplemented without gamma-irradiation or heat-inactivation from GE Healthcare, Little Chalfont, England) following manufacturer instructions. Dissociated cells were then seeded in 12-well tissue culture plates in a mixture of F-medium and mouse fibroblast culture supernatant at a ratio of 4:1 (v/v) supplemented with 5 µM Rho kinase inhibitor (cat. no. ALX-270-333-M005, Enzo Life Sciences, Farmingdale, NY, USA). Cells were maintained at 37°C in a humidified incubator with 5% CO₂. Colonies became visible after 5 to 7 days and cells were passaged after two weeks (approximately 10⁵ cells) and expanded after reaching 80 to 90% confluence. An image of cells at passage 6 was recorded using phase contrast microscopy using an EVOS FL Auto Cell Imaging System (Thermo Fisher Scientific).

Human MCF-7 mammary gland adenocarcinoma cells originally isolated from a 69 year old Caucasian woman with several characteristics of differentiated mammary epithelium were cultured in Eagle's Minimum Essential Medium (EMEM) (cat. no. 30-2003, ATCC, Manassas, VA, USA) supplemented with 10% (v/v) heat-inactivated fetal bovine serum (FBS) (cat. no. FP-0500-A, Atlas Biologicals, Fort Collins, CO, USA), and 1% Penicillin/Streptomycin Solution 100× (cat. no. 30-002-Cl, Corning, Manassas, VA, USA). Human MDA-MB-231 mammary gland adenocarcinoma cells isolated as one of a series of breast tumor lines from pleural effusions of a 47 year old Caucasian female were cultured in DMEM/Ham's Nutrient Mixture F12 1:1 (cat. no. 51445C, Sigma-Aldrich, Deisenhofen, Germany) supplemented with 5% (v/v) heat-inactivated fetal bovine serum (FBS) (cat. no. FP-0500-A, Atlas Biologicals, Fort Collins, CO, USA), 1% Penicillin/Streptomycin Solution 100× (cat. no. 30-002-Cl, Corning, Manassas, VA, USA) and 1 mM L-Glutamine (cat. no. 25005-Cl, Corning, Manassas, VA, USA). Both cell lines were tested via short tandem repeat profiling in July 2018 by Genetica DNA Laboratories (Burlington, NC, USA) and verified as authentic, giving a 100% match when compared to the known reference profile (30). Human MCF 10A mammary epithelial cells originally isolated from 36 year old Caucasian women were purchased from ATCC, Manassas, VA, USA (cat. no. CRL-10317) and cultured in Mammary Epithelial Cell Basal Medium (cat. no. C-21215, Promo Cell, Heidelberg, Germany) supplemented with Mammary Epithelial Cell Supplement Pack (cat. no. C-39110, Promo Cell, Heidelberg, Germany), 100 µg/ml gentamicin (cat. no. G1397, Sigma-Aldrich, Deisenhofen, Germany) and 0.05 µg/ml amphotericin B (cat. no. A2942, Sigma-Aldrich, Deisenhofen, Germany).

EpCAM expression was assessed by binding of FITC-conjugated anti-EpCAM monoclonal mouse anti-human antibody (1:10 dilution) (cat. no. 324203, Biolegend, San Diego, CA, USA). MCF-7, MDA-MB-231 and human primary phyllodes cells (about 5×10⁴ cells) were harvested with GIBCO^® enzyme-free cell-dissociation buffer (cat. no. 13151-014, Life Technologies, Grand Island, NY, USA), washed with flow-cytometry buffer (PBS containing 1% BSA) and incubated with FITC-conjugated anti-EpCAM antibody (1:10 dilution) in the same buffer for 30 min followed by washing twice with flow-cytometry buffer. Cells were fixed with 1% (final concentration) PFA. Acquisition and analysis of data was performed by UAMS Flow Cytometry Core using a LSRFORTESSA flow cytometer (BD Biosciences, San Jose, CA, USA) and FlowJo^® software (FlowJo LLC, Ashland, OR, USA).

A 3-(4,5-dimethylthiazol-2-yl)-2,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT)-based assay (31,32), was used to evaluate the effect of drugs on the viability of primary phyllodes cells and MCF 10A cells. Phyllodes cells (0.35×10⁴/well) in 100 µl of F-medium were seeded in 96-well plates (TPP, Trasadingen, Switzerland). After 24 h cells were treated with DOX in the following concentrations: 1, 3, 10, 30, 100, 300, 1,000, 3,000 and 10,000 nM, as well as SAL and ABT-263 at a single concentration of 10 µM for 96 h with control cells receiving vehicle (0.1% DMSO) alone. The experiment was performed in triplicate. MCF 10 A cells (1×10⁴/well) in 100 µl of MEGM were seeded in 96-well plates (TPP, Trasadingen, Switzerland) and after 24 h treated with DOX, SAL and ABT-263 at the following concentrations: 1, 3, 10, 30, 100, 300, 1,000, 3,000 and 10,000 nM (of DOX and SAL) and 300, 1,000, 1,200, 1,400, 1,600, 1,800, 2,200, 3,000 nM of ABT-263 for 96 h with control cells receiving vehicle (0.1% DMSO) alone. The experiment was performed in quadruplicate. After treatment, 10 µl of MTT solution (5 mg/ml, cat. no. M2128, Sigma-Aldrich, Deisenhofen, Germany) was added to each well, and the plate was incubated at 37°C for 4 h in a humidified 5% CO₂ incubator. Medium was then aspirated and 150 µl of DMSO was added to each well and the plate agitated on a shaking platform (150 rpm) for 10 min. Absorbance was recorded at 540 nm using a BioTek Plate Reader. Inhibition of formation of colored MTT formazan was taken as an index of cytotoxicity activity. IC₅₀ values were determined by non-linear regression analysis using GraphPad Prism 6 for Windows (GraphPad Software).

1-way ANOVA with Tukey's multiple comparisons test was used to determine statistically significant differences between means in experimental groups. Data are presented as a mean ± SD. For ex vivo experiments three biological replicates were performed for every drug concentration. Cell viability assay employing primary phyllodes cells was performed in triplicate, whereas MCF 10 A cells in quadruplicate. EpCAM staining was performed in duplicate. Statistical analysis was performed using GraphPad Prism 6 for Windows (GraphPad Software, Inc., San Diego, CA, USA).

A single specimen of fresh, unfrozen MPTB was obtained, sliced, incubated with select drugs, and subjected to H&E staining, as described in Materials and methods. A representative image of an untreated MPTB section fixed and stained immediately upon receipt is shown in Fig. 1 (left panel). The presence of a biphasic neoplasm made of neoplastic epithelial and mesenchymal (spindle cell) elements is consistent with PT. This was confirmed by the CHTN pathology report which subsequently became available. The report indicated that the tumor showed biphasic fibroepithelial growth with pushy lobulated margins, hypercellular stroma with leaf-like architecture, stromal outgrowth, moderate to severe nuclear atypia, high mitotic rate, and focal necrosis. Importantly, the overall morphological features and tumor cell viability remained unchanged in vehicle treated samples at 72 h (Fig. 1, right panel) compared to the untreated original tumor sample (Fig. 1, left panel), indicating that incubation had no deleterious effects. Next, explant cultures were treated with drugs including ABT-263, SAL, DOX, TAX, VCR, COL, and CIS. Representative images of H&E stained slices at 20× magnification are presented in Fig. 2 for the most active compounds ABT-263, SAL, DOX, and in Fig. S1 for the other compounds. Healthy versus dead or drug-affected tumor or normal epithelial cells were distinguished and quantified. As evident from representative images in Fig. 2 and the corresponding quantitation in Fig. 3, ABT-263, SAL and DOX treatment caused marked tumor cell death in a dose-dependent manner compared to vehicle treated slices (Fig. 3). Of additional interest, treatment with ABT-263 had very little to no toxic effect on normal epithelial cells regardless of the dose (Fig. 2). This is evident from examination of the images in Fig. 2, where ducts retained their normal architecture and cellularity even with the highest tested ABT-263 concentration (indicated by arrow 1), and from quantitative assessment of toxicity (Table I). In contrast, the anti-tumor activities of SAL and DOX were accompanied by significant toxic effects on normal epithelial cells, particularly after DOX treatment (arrow 2 in Fig. 2; Table I).

Treatment of PTB explant cultures with TAX, VCR, COL, and CIS was also performed under similar conditions (Fig. S1A). However, none of these drugs produced significant effects on the tumor cells thus quantification and statistical analysis of dead tumor cells was not performed.

Additionally, sections from paraffin embedded tissue were analyzed by immunohistochemical staining for Ki-67, which is an established marker of cell proliferation (33). Representative images at 20× magnification are presented in Fig. 4 and Fig. S1B. In untreated tumor sections, the Ki-67 proliferation index was very low (<5%), in a good agreement with a previous study (34), whereas strong Ki-67 immunolabeling was observed in some of the ductal epithelial cells (Fig. S1B). Ductal epithelial Ki-67 immunolabeling was largely maintained after ABT-263 treatment but was not observed after treatment with SAL or DOX (Fig. 4), consistent with the differential effects of the drugs on normal cells.

In order to further assess the selectivity of ABT-263 towards tumor versus normal cells compared to SAL and DOX, both of which appeared less discriminate in the ex vivo model, we performed MTT viability assay on the MCF 10A cell line. These are non-tumorigenic mammary epithelial cells commonly used as a surrogate for normal mammary cells. As shown in Fig. 5, IC₅₀ values were ABT-263, 1.60±0.03 µM; SAL, 0.11±0.03 µM; and DOX, 0.02±0.01 µM. Thus MCF 10A cells are much more sensitive to SAL (20×) and DOX (120×) compared to ABT-263, in good agreement with observations made in the ex vivo model (Fig. 2 and Table I).

To further investigate the response of PTB cells to anticancer drugs we sought to generate primary cells from the MPTB we obtained. This was facilitated by recent advances in the preparation of primary cells through Rho kinase inhibition and the use of fibroblast feeder cells or conditioned medium derived therefrom (21,22). An image from a phase contrast microscope of the cells derived from the MPTB tissue using this procedure is shown in Fig. 6. It is evident that the population of cells was mixed, with the majority showing an elongated, triangular morphology. In addition, cells with a more spherical morphology were present, as were infrequent cells with large intracellular vacuoles or vesicles. In order to confirm the presence of PT cells in this population, the expression of EpCAM (CD326), a transmembrane glycoprotein present on the surface of epithelial cells, was evaluated (35). MCF-7 and MDA-MB-231 cell lines, characterized by high and low EpCAM expression, respectively (36), were used as positive and negative controls. As shown in Fig. 7, the two breast cancer cell lines expressed EpCAM consistent with expectations, and were used to set a gate denoting high versus low expression. When the phyllodes primary cells were examined, a range of EpCAM expression was observed, with about 30% of the population in the range defined as high, indicating the presence of epithelial derived cells (Fig. 8). Once primary cells were established, they proliferated with a doubling time of 3–5 days initially, and viable cells up to passage 6 were obtained before they showed signs of slowed growth. Thus the overall number of cells obtained was limited and precluded an in-depth characterization of their properties and response to the drugs. Nonetheless, viability assays after select drug treatment were conducted using the colorimetric MTT assay (Figs. 9 and 10) (31,32). This method is based on the conversion of MTT to blue MTT-formazan crystals by mitochondrial enzymes present in viable cells. Data for DOX using primary phyllodes cells at passage 6 are shown in Fig. 9. A concentration-dependent decrease in viability was observed with an IC₅₀ value of 0.40±0.07 µM. Because of limited numbers of cells, other drugs could not be tested with a full range of concentrations, but preliminary experiments indicated that the primary PTB cells were also sensitive to ABT-263 and to SAL used at 10 µM, which was the highest concentration employed in MTT assay (Fig. 10).