Figure 1 shows the culture procedure we have developed.

Explant model schematic: Individual steps are highlighted from tissue collection to fixing for immunohistochemistry. i. shows example of tissue processing, ii. tissue can be seen within the Boyden chamber, encased in hydrogel and iii. shows an example of the tissue microarray (TMA) produced for each sample

Normal non-cancerous breast tissue was collected following risk-reduction surgery. Research samples were obtained from the Manchester Cancer Research Centre (MCRC) Biobank with fully informed consent. Ethical approval for the study was granted by the MCRC Biobank under authorisation number 18/NW/0092. Sample details are displayed in Table 1.

Tissue was placed immediately into collection medium consisting of DMEM High Glucose (SIGMA, D6546) with 100U/mL Penicillin/100 µg/mL Streptomycin (SIGMA, P0781) and stored for up to 24 h at 4°C. Excess adipose tissue was removed, and tissue was cut into 2–4 mm3 pieces before culture.

An animal-free hydrogel (VitroGel RGD, TebuBio, TWG003) was used to provide support to explants. The hydrogel was mixed with 0.5 × PBS to achieve the desired elastic modulus (which is representative of stiffness), according to the manufacturer’s instructions [27]. This was then mixed with medium to initiate hydrogel gelation and 100 µL was immediately pipetted into Boyden chambers suspended over a well containing 700 µL of medium. Chambers were incubated for 2 h at 37°C to allow the hydrogels to set. The same procedure was followed to overlay 150 µL of hydrogel on top of the explant and, once the hydrogel was set, 200 µL of medium was added to the top. During culture, 50% of the medium below and above the explant was refreshed every 2–3 days.

Explant medium (ExM): DMEM/F12 (Thermo, 11330032) containing B27 supplement (no vitamin A; Invitrogen, Paisley, UK, 12587010), 2 mM L-glutamine (SIGMA, G7513) and 100 U/mL Penicillin/100 µg/mL Streptomycin (SIGMA, P0781).

Clevers’ medium (CM[28]): DMEM/F12 containing 5% R-spondin conditioned medium, 5 nM neuregulin (Peprotech, 100–03), 5 ng/mL epidermal growth factor (Peprotech, AF-100–15), 100 ng/mL noggin (Peprotech, 120-10C), 500 nM A83-01 (Tocris, 2939), 5 µM Y27632 (Abmole, S1049), 500 nM SB202190 (Sigma, S7067), 1 × B27 (with vitamin A, Gibco, 1750444), 1.25 nM N-acetylcysteine (Sigma, A9165), 5 mM nicotinamide (Sigma, N0636), 1 × Glutamax (Invitrogen, 12634–034), 10 mM HEPES (Invitrogen, 15,630–056), 100 U/mL Penicillin/100 µg/mL Streptomycin and 50 ng/mL FGF2 (Thermo, 100-18B).

FCS medium: DMEM/F12 and 10% foetal calf serum (FCS, Thermo, 10270106), 100 U/mL Penicillin/100 µg/mL Streptomycin.

For activation and inhibition studies, 10 nM 17β-oestradiol (SIGMA, E2758) and/or 100 nM fulvestrant (SIGMA, I4409) was added to the medium following explant encapsulation in its hydrogel support and was refreshed with each medium change.

Tissue was removed from the hydrogel, formalin-fixed and paraffin-embedded, with 6 explants per block, and 4 µm slices were prepared for immunohistochemistry. Staining was performed using the Bond Max autostainer (Leica) and Ventana Discovery autostainer (Roche). TMAs were scanned using the Olympus VS120.

Staining on the Leica Bond Max was performed with 20 min of antigen retrieval at pH6 (Ki67, progesterone receptor (PR) and cleaved caspase 3) and pH9 (oestrogen receptor α (ERα)). Primary antibodies: mouse α-ERα (6F11, Life Technologies, MA513304) 1:200, mouse α-Ki67 (MIB-1, DAKO, M7240) 1:100, mouse α-PR (636, DAKO, M3569) 1:500, rabbit α-cleaved caspase 3, 1:200 (5A1E, New England Biolabs, 9664S) and EnVision + Single Reagents (HRP, Mouse, Agilent, K400111-2) used as secondary, following the manufacturer’s instructions.

Staining on the Ventana was performed for CD4 and CD8 using an ultraView Universal DAB Detection Kit (Roche, 760–500) and CD68 using an OptiView DAB IHC Detection Kit (Roche, 76–700). Slides were deparaffinised, antigens were retrieved using standard cell conditioning (CC1), primary antibody incubations were performed for 16 min, and bluing with haematoxylin II was performed for 4 min. The following primary antibodies were used according to the manufacturer's instructions: α-CD4 (SP35, Roche, 790–4423), α-CD8 (SP57, Roche, 790–4460) and α-CD68 (KP-1, Roche, 790–2931).

Scoring was performed, and percentage positive calculated, in https://imagej.net/ij/ at 10 × magnification. All epithelial cells in the explant (ERα, PR, Ki67, Caspase) were counted and immune cells (CD4, CD8, CD68) were treated as a single population whether inter- or intraductal. Fold change from control was calculated to demonstrate changes following culture.

To test the elastic modulus of the low, moderate and high stiffness gels, 500 µL samples were prepared using ExM in Boyden chamber hanging inserts, as described above. The inserts were incubated at 37°C for 24 h prior to rheological testing. Hydrogels were removed from the hanging insert and transferred to the rheometer. The 25 mm upper parallel plate of the rheometer was lowered to the desired trim gap size of 500 µm, and the gels were allowed to equilibrate for 3 min at room temperature. Single frequency (1Hz) amplitude sweeps were performed between 0.001% and 100% shear strain using an Anton Parr MCR 302E rheometer.

One-way ANOVA tests were performed, with pairing of samples, and comparisons made to day 0 or untreated sample as appropriate. A Dunnett’s correction for multiple testing was performed. For rheology, measurements were repeated four times, and a one-way ANOVA was performed on the linear viscoelastic regions of the gels. Significance is highlighted in each figure; *P < 0.05, **P < 0.01, ***P < 0.001.