Mechanisms that clear mutations drive field cancerization in mammary tissue

Mice

All mice used for experiments were adult females from a mixed background, housed under standard laboratory conditions and receiving food and water ad libitum. All experiments were performed in accordance with the guidelines of the Animal Welfare Committees of the Netherlands Cancer Institute and KU Leuven. Sample size was determined using a resource equation approach, mice were randomly assigned to experimental groups and blinding was performed during data analysis. R26R-Confetti^het (JAX stock no. 013731)^61,62; R26-CreERT2^het (JAX stock no. 008463)⁶³ mice were injected intraperitoneally with tamoxifen (Sigma-Aldrich), diluted in sunflower oil, to activate Cre recombinase. To achieve clonal density labelling (fewer than one MaSC per duct on average), R26R-Confetti^het;R26-CreERT2^het mice were injected with 1 mg of tamoxifen per 25 g of body weight between 10 and 15 weeks of age. Ovariectomies were performed between 10 and 15 weeks of age, at least 7 days before lineage tracing initiation. The third, fourth and fifth mammary glands of R26R-Confetti;Brca1^fl/fl;Trp53^fl/fl (refs. ^31,64) or R26R-Confetti mice were intraductally injected with recombinant TAT-Cre protein (20 units per gland diluted in 20 µl of PBS, Sigma-Aldrich) between 10 and 15 weeks of age. This TAT-Cre injection resulted in roughly one labelled cell for every 100–200 cells (Extended Data Fig. 4e). As the confetti construct comprises four distinct colours, there is, on average, one cell labelled with a confetti colour per 400 cells. Considering that a MaSC-progeny unit consists of roughly five to ten cells, a single confetti-labelled cell is induced in one out of 40–80 units. Over time, many clones become extinct, leading to a dilution in the number of clones and making collisions even less likely. For each of the experiments, mice were analysed at different time points after lineage tracing initiation as indicated in Fig. 1b. The injected mammary glands of R26R-Confetti;Brca1^fl/fl;Trp53^fl/fl mice at the latest time point (225 days) were analysed when one of the injected glands developed a palpable tumour of at least 5 × 5 mm, which was between 200 and 250 days after recombination. Tumour sizes did not exceed 1,500 mm³ in accordance with the guidelines of the Animal Welfare Committee of the Royal Netherlands Academy of Arts and Sciences, the Netherlands Cancer Institute and KU Leuven. Samples were randomly allocated to the experimental groups, sample size was not determined a priori and investigators were not blinded to experimental conditions, except where indicated. For clonal analysis of the R26R-Confetti;Brca1^fl/fl;Trp53^fl/fl model, we analysed n = 4 mice (14 days), n = 4 mice (64 days), n = 5 mice (120 days) and n = 6 mice (225 days). For clonal analysis of the R26R-Confetti model, we analysed n = 3 mice (14 days), n = 3 mice (64 days), n = 5 mice (120 days) and n = 6 mice (225 days). Clonal analysis of the ovariectomized R26R-Confetti;Brca1^fl/fl;Trp53^fl/fl and R26R-Confetti^het models was performed on at least n = 3 mice per time point. Adult CAG;;KikGR female mice⁶⁵ (RIKEN no. CLSTCDB0201T-117830853340) were used to visualize the short-term dynamics of the mammary gland by repeated imaging through a mammary imaging window as described below. R26-mTmG female mice (JAX no. 007676)⁶⁶ were used to visualize the long-term stability of the mammary gland through a repeated skin-flap procedure as described below.

Mammary imaging window implantation, repeated skin-flap procedure and intravital imaging

Mice were anaesthetized using isoflurane (Isovet) inhalation (1.5/2% isoflurane/air mixture). The fourth mammary gland of adult CAG;;KikGR female mice was imaged repeatedly through a mammary imaging window as previously described^47,48. The fourth mammary gland of adult R26-mTmG mice was imaged repeatedly with a skin flap as previously described⁴⁷. To visualize the mammary gland, mice were placed in a facemask within a custom designed imaging box. Isoflurane was introduced through the facemask and ventilated by an outlet on the other side of the box. The imaging box and microscope were kept at 34 °C by a climate chamber surrounding the entire stage of the microscope including the objectives. Imaging was performed on an inverted Leica SP8 Dive system (Leica Microsystems) equipped with four tuneable hybrid detectors, a MaiTai eHP DeepSee laser (Spectra-Physics) and an InSight X3 laser (Spectra-Physics) using the Leica Application Suite X (LAS X) software. All images were collected at 8 bit and acquired with a ×25 water immersion objective with a free working distance of 2.40 mm (HC FLUOTAR L ×25/0.95W VISIR 0.17). For the CAG;;KikGR model, Kikume Green was excited at 960 nm and detected at 490–550 nm. Each imaging session, all visible ducts through the imaging window were imaged using a tiled z scan with ×1–2 zoom and a z-step size of 5–10 μm. For the skin-flap imaging, TdTomato was excited at 1,040 nm and detected at 540–730 nm. All visible ducts were imaged together in one tile z scan with a ×0.75 zoom, a z-step size of 10–20 μm. These parameters allowed to scan large regions of up to 2 cm² in less than 3 h. At the end of the first skin-flap imaging session, the skin was closed with a continuous, non-resorbable suture. After 3 months, the skin flap was re-opened for the second imaging session, and the same imaging fields were retraced using the nipple and collagen I structures of the first imaging session as landmarks.

Staging of the mice

To determine the oestrous cycle stage of the mice, a vaginal swab was collected as described⁶⁷. In short, the vagina was flushed using a plastic pipette filled with 50 µl PBS, and the liquid was transferred to a dry glass slide. After air drying, the slide was stained with Crystal Violet and the cell cytology was examined using a light microscope.

Clone isolation and CNA sequencing

Here, 225 days after Cre mediated recombination, fourth mammary glands were extracted, fixed overnight in 1% paraformaldehyde (PFA), incubated in sucrose overnight and stored in optimal cutting temperature (OCT) at −80 °C. For microdissection of individual clones, OCT blocks were thawed at room temperature in the dark for 30 min and mammary glands removed from OCT and washed in 50 ml of PBS on ice. Mammary glands were dissected under a benchtop fluorescent macroscope (Zeiss) using Dumont forceps and fine scissors, using the clone morphology to distinguish between transformed and untransformed clones. Each dissected clone was washed in 1 ml of PBS on ice for 2–5 h (until the end of the dissection procedure). Per mammary gland, one piece of non-fluorescent tissue from the inguinal lymph node was dissected as internal sequencing control. After washing, pieces were lysed in 70 µl of Arcturus lysis buffer following the instructions of the ThermoFisher Scientific Arcturus PicoPure kit KIT0103. Lysis was carried out in a PCR cycler for 18 h at 65 °C followed by 30 min at 75 °C and holding at 4 °C. Samples were then purified using the Roche FFPE DNA extraction kit 06650767001 50-588-384 following the manufacturer’s instructions with elution in 25 µl of PCR grade water. For DNA sequencing, library preparation was carried out with a KAPA Hyper kit (Roche; KK8504) according to the manufacturer protocol with four PCR cycles, before the samples were sequenced by low-coverage whole genome sequencing. The copy number alteration (CNA) analysis was conducted in R, using QDNAseq with 50 kb bins and the mm10 mouse reference genome. This methodology yielded copy number values from both normal and transformed clones, along with an internal control sample. Normalization was achieved by first converting copy number values to log₂, then subtracting the internal control sample’s values from those of the normal and transformed tumours. We averaged these adjusted copy number values for each replicate across both clone types (13 early normal clones and 13 early transformed clones). Data visualization was executed using the ggplot2 package in R, with specific emphasis on certain chromosomes. Regarding the late-stage tumours published in ref. ⁴⁰, copy number profiling data corresponding to ten Wap-Cre;Brca1^fl/fl;Trp53^fl/fl (WB1P) female mice harbouring a WB1P late-stage mammary tumour, along with internal control samples (spleen) was used. The CNA sequence analysis included the use of cutadapt for adaptor sequence removal and BWA for sequence alignment (using bwa aln, bwa mem) to the mm10 mouse genome. This procedure mirrored the earlier steps up to plotting with ggplot2, repeated for ten WB1P replicates.

Quantification of the distribution of proliferation

Oestrous-cycling mice and mice that had undergone ovariectomy with a 2-week recovery period (all above 8 weeks of age) received 0.5 mg ml^−1 EdU in drinking water (refreshed every second day) for 1 week. 3D imaging was performed on three cycling mice and five ovariectomized mice. Per mouse, one-quarter of the mammary gland was taken for subsequent analysis. Samples were fixed in 4% PFA overnight and stained using the FLASH protocol with FLASH Reagent 2 (ref. ⁶⁸). Before adding the primary antibodies, samples were stained for EdU as follows. Tissues were incubated in 5 ml of 3% bovine serum albumin for 1 h, followed by three washes in PBS for 20 min each. EdU was detected with an Alexa-647 azide. The reaction cocktail for EdU fluorescent labelling was prepared according to the manufacturer’s guidelines using the Click-It EdU imaging kit (ThermoFisher Scientific). Per gland, 0.5 ml of reaction cocktail was added for incubation for 4 h at room temperature with gentle agitation on a nutator. The cocktail was removed, samples washed once in 3% bovine serum albumin in PBS for 20 min, followed by three washes in FLASH blocking buffer for 20 min each. Subsequently, samples were stained with primary and secondary antibodies overnight each. Primary antibodies used were KRT8 (rat, Troma-I, Merck Millipore, 1:800) and αSMA (mouse IgG2a, clone 1A4, ThermoFisher Scientific, 1:600). Secondary antibodies used were donkey antirat Alexa-488 and donkey antimouse Alexa546 (ThermoFisher Scientific, catalogue nos. A21208 and A10036, respectively, 1:400), combined with Hoechst 33342 for nucleus detection. Samples were imaged on an Andor Dragonfly spinning disc system, installed on an inverted Leica DMI8 microscope with an Andor Zyla 4+sCMOS camera using a ×10, 0.45 NA Fluo objective (Leica). Imaging was carried out with a 40 μm disc using 405 nm excitation, a 561 nm optically pumped semiconductor laser and 637 nm diode lasers. Images were visualized with Imaris Viewer using gamma correction, ortho slicers and cutting planes to depict deeper tissue layers. For each mammary gland, distribution of proliferation was quantified in five regions. Ripley analysis using QuPath⁶⁹ was performed with a custom-made script (available at https://github.com/BioImaging-NKI/qupath_ripley). The image was opened in QuPath and a freehand line was drawn by hand to outline the duct for analysis. A multipoint annotation was drawn by hand to mark the positions of proliferating cells along the duct. The script calculated Ripleys K function and normalized it to an unclustered distribution resulting in Ripley’s L function. Data were plotted in GraphPad Prism v.10. For simulations, we have generated clustered and unclustered data in Python.

Whole-mount immunofluorescence staining of mammary glands

The third, fourth and fifth mammary glands were dissected and incubated in a mixture of collagenase I (1 mg ml^−1, Roche Diagnostics) and hyaluronidase (50 μg ml^−1, Sigma-Aldrich) at 37 °C for optical clearance, fixed in periodate–lysine–PFA buffer (1% PFA; Electron Microscopy Science), 0.01 M sodium periodate, 0.075 M l-lysine and 0.0375 M P-buffer (0.081 M Na₂HPO₄ and 0.019 M NaH₂PO₄; pH 7.4) for 2 h at room temperature, and incubated for at least 3 h in blocking buffer containing 1% bovine serum albumin (Roche Diagnostics), 5% normal goat serum (Monosan) and 0.8% Triton X-100 (Sigma-Aldrich) in PBS. Primary antibodies were diluted in blocking buffer and incubated overnight at room temperature. Secondary antibodies diluted in blocking buffer were incubated for at least 6 h. Nuclei were stained with 4,6-diamidino-2-phenylindole (DAPI) (0.1 μg ml^−1; Sigma-Aldrich) in PBS. Glands were washed with PBS and mounted on a microscopy slide with Vectashield hard set (H-1400, Vector Laboratories). Primary antibodies used were: anti-KRT14 (rabbit, Covance, PRB155P, 1:700), anti-ECAD (rat, eBioscience, 14-3249-82, 1:700), anti-oestrogen receptor (rabbit, no. 13258, Cell Signaling, 1:100), anti-progesterone receptor (rabbit, Clone SP2, MA5-14505, ThermoFisher Scientific, 1:200) and anti-SMA (mouse IgG2a, clone 1A4, Sigma-Aldrich, 1:600). Alexa Fluor 647 and Alexa Fluor 488 Phalloidin were used 1:500 (A-22287 and A-12379, ThermoFisher Scientific) and incubated together with the secondary antibodies. Secondary antibodies used were: goat antirabbit, goat antirat or goat antimouse IgG2a, all conjugated to Alexa-647 (ThermoFisher Scientific, catalogue nos. A21245, A21247 and A21241, respectively, 1:400).

Whole-mount imaging of mammary glands

Imaging of whole-mount mammary glands was performed using an inverted Leica TCS SP8 confocal microscope, equipped with a 405 nm laser, an argon laser, a diode-pumped solid-state laser 561 nm laser and a HeNe 633 nm laser. Different fluorophores were excited as follows: DAPI at 405 nm, cyan fluorescent protein (CFP) at 458 nm, green fluorescent protein (GFP) at 488 nm, yellow fluorescent protein (YFP) at 514 nm, red fluorescent protein (RFP) at 561 nm and Alexa-647 at 633 nm. DAPI was collected at 440–470 nm, CFP at 470–485 nm, GFP at 495–510 nm, YFP at 540–570 nm, RFP at 610–640 nm and Alexa-647 at 650–700 nm. All images were acquired with a ×20 (HCX IRAPO N.A. 0.70 WD 0.5 mm) dry objective using a Z-step size of 1–5 μm (total Z-stack around 200 μm). 3D overview tile scans of the mammary glands were acquired by scanning large tile-scan areas (xyz). Next, detailed images were obtained of the individual clones. All images were stitched and processed in the true 3D real-time Rendering LAS X 3D Visualization module (Leica Microsystems) and further processed using ImageJ software (https://imagej.nih.gov/ij/).

Clonal analysis on whole-mount glands

Three-dimensional tile-scan images of whole-mount and fully intact mammary glands were used to manually reconstruct the ductal network by outlining the ducts based on the labelling by ECAD, SMA or KRT14 (between 400 and 600 tiles, ×10 objective, Z-step size of 5–10 µm). After localization of the confetti clones in these 3D overview scans, each clone was imaged in detail with a ×25 water objective using confocal imaging by taking a Z-stack with step size between 1 and 3 µm. On the basis of the overlap with the luminal- or basal-cell-specific labelling and cellular morphology (that is, a cuboidal shape for luminal cells and an elongated shape for basal cell), the labelled confetti cells were identified and annotated in the schematic outline of the mammary tree, including information on their confetti colour (GFP, green; YFP, yellow; RFP, red and CFP, cyan) and their identity: that is, luminal or basal. Regions in which, for technical reasons, the gland could not be visualized well were omitted from analysis (in three out of 160 glands). Clone sizes, referring to the number of cells within each clone, were determined through manual visual inspection of tissue samples, with the quantification performed by eye using detailed Z-stack images and 3D rendering of each individual clone. Using custom-made.NET software (available on request from J.v.R.), the coordinates of the branch points, and the position of the labelled cells in ducts and in ductal ends were scored. To calculate the surviving clone fraction, the total number of clones was determined for each of the indicated lineage tracing time points by analysing the entire mammary gland in three dimensions using our whole-gland imaging approach (n = 3 glands per time point of two individual mice). Next, the average numbers of clones identified at 64, 120 and 225 days after lineage tracing initiation were divided by the average number of clones identified 14 days after lineage tracing initiation resulting in the surviving clone fraction as depicted in Figs. 2d and 5g.

Mammary epithelial cell sorting and real-time qPCR

The third, fourth and fifth mammary glands of R26R-Confetti;Brca1^fl/fl;Trp53^fl/fl mice were intraductally injected with recombinant TAT-Cre protein (20 units per gland diluted in 20 µl PBS, produced in-house) between 10 and 13 weeks of age. Then 120 to 180 days after injection, mammary glands were harvested, minced and digested at 37 °C for 30 min in a mixture of collagenase A (2 mg ml^−1, Roche Diagnostics), hyaluronidase (300 μg ml^−1, Sigma-Aldrich) and DNase (1 mg ml^−1) in DMEM/F12 (Gibco). After 10 min incubation with TripLE (Gibco) at 37 °C cells were strained through a 100 μm cell strainer (Fisher Scientific) to obtain single cells. Cells were spun down for 10 min at 550 RCF (relative centrifugal force) at 4 °C followed by blocking for 15 min on ice in 5 mM EDTA/PBS with 2% sterile filtered normal goat serum (Gibco). CD45-Alexa-647 (clone 30-F11, 03123, Biolegend, 1:200) and EpCAM-APC/Cy7 (clone G8.8, 118218, Biolegend, 1:200) were diluted in 5 mM EDTA/PBS with 2% normal goat serum and incubated for 30–45 min on ice to label the immune population (CD45) and the epithelial population (epithelial cell adhesion molecule). Cells were centrifuged for 5 min at 800 RCF at 4 °C and pushed through a 35 μm cell strainer. The FACS Aria III Special Ordered Research Product (BD Biosciences) was used to sort confetti⁺ and confetti cells, by applying a broad FSC/SSC gate, followed by gates excluding doublets (for the gating strategy, see Extended Data Fig. 1d). Afterwards, non-immune (AF647^–; 670/30) confetti-positive (RFP⁺ (YG610/20), GFP⁺/YFP⁺ (BL530/30), CFP⁺ (V450/50)) and, separately, confetti-negative ((RFP^– (YG610/20), GFP^–/YFP^– (BL530/30), CFP^– (V450/50)) epithelial cells (APC/Cy7⁺; 780/60) were collected. Similarly, non-immune (AF647^–; 670/30) epithelial cells (APC/Cy7⁺; 780/60) were collected from three R26R-Confetti;Brca1^fl/fl;Trp53^fl/fl mice that had not received TAT-Cre intraductally as a negative control. Data were analysed in FlowJo v.10 for the gating strategy (Extended Data Fig. 1d). Cells were spun down for 10 min at 800 RCF at 4 °C and DNA was isolated using the PicoPure DNA extraction kit (Applied Biosciences; KIT0103) according to the manufacturer’s instructions. The same method was applied to isolate genomic DNA (gDNA) from K14-Cre;Brca1^fl/fl;Trp53^fl/fl mammary tumour organoids, representing fully recombined samples as a positive control. gDNA concentration was measured using the DeNovix DS-11 spectrophotometer. DNA was diluted to 50–75 ng ml^−1 and used for real-time qPCR using the SYBR Green Master Mix (ThermoFisher Scientific, catalogue no. 4309155) in a QuantStudio 6 Flex Real-Time PCR system using the primers listed in the table below. Reactions contained roughly 75 ng of template gDNA and 1 µM of both forward and reverse primers in 20 µl reaction volume. Expression values were calculated by transforming delta–delta Ct values (2^-ΔΔCt). Ribosomal Protein L38 (Rpl38) was used as a housekeeping gene. To confirm correct qPCR product amplification, 25 μl of qPCR product of the control samples (R26R-Confetti;Brca1^fl/fl;Trp53^fl/fl mice that had not received TAT-Cre intraductally) was loaded on an 2% agarose gel with loading buffer (Bioxline, catalogue no. BIO-37045) and a DNA ladder (Meridian Bioscience, catalogue no. BIO-33056) and run at 80 V for 1.5 h, after which the qPCR product was cut out of the gel and purified using the NucleoSpin Gel and PCR Clean-up kit (Macherey-Nagel, catalogue no. 740609.50) according to the manufacturer’s instructions, and was confirmed by sequencing using the qPCR primers.

Generation of Brca1;Trp53 mutant and WT organoids followed by Ki-67/CC3 staining

The fourth and fifth mammary glands of R26R-Confetti;Brca1^fl/fl;Trp53^fl/fl mice were intraductally injected with recombinant TAT-Cre protein (20 units per gland diluted in 20 µl PBS, Sigma-Aldrich) between 10 and 15 weeks of age. Then, 64 or 225 days post-induction, mammary glands were harvested and prepared for fluorescence-activated cell sorting (FACS) as described before. Both confetti-positive Brca1^{−/−};Trp53^{−/−} and non-recombined control cells were seeded in a 24-well plate, 10,000 cells per drop of Cultrex Basement Membrane Extract (Type 2, 3532-010-02, R&D Systems) and cultured in the DMEM/F12 (Gibco) supplemented with iInsulin-transferrin-selenium (100×, catalogue no. 41400045, Gibco), B-27 Supplement (50×, catalogue no. 17504044, Gibco), NAC 1.25 mM (N-acetyl-l-cysteine, 0.125 M in PBS, catalogue no. 6169116, Biogems), mFGF2 2.5 nM (Fibroblast Growth Factor 2, catalogue no. 100-18B, PeproTech) and mEGF 2 nM (Epidermal Growth Factor, catalogue no. 3165-09, PeproTech). After 2 weeks of culture, organoids were fixed with 4% PFA (catalogue no. 47347, AlfaAesar) for 10–15 min at room temperature inside the Basement Membrane Extract droplet on an orbital shaker at 25 rpm. Afterwards, organoids were washed three times for 10 min with PBS, followed by incubation in permeabilization buffer (5% normal goat serum (catalogue no. 16210072, Gibco) and 0.5% Triton X-100 ((Sigma-Aldrich) in PBS) for 3 h. To stain for cell proliferation and cell death, primary antibodies Ki-67 (rat, clone SolA15, 14-5698-82, eBioscience, 1:100) and Cleaved Caspase-3 (rabbit, Asp175, no. 9661, Cell Signaling Technology, 1:400), respectively, were added in the blocking buffer (5% normal goat serum (Gibco) in PBS), and incubated overnight at 4 °C. Organoids were washed three times for 15 min with PBS and secondary antibodies goat antirabbit IgG Antibody, Alexa Fluor 647 (catalogue no. A21244, Thermo Scientific, 1:400), goat antirat IgG Antibody, Alexa Fluor 647 (catalogue no. A21247, Thermo Scientific, 1:400) were added and incubated for more than 5 h at room temperature covered in aluminium foil on an orbital shaker at 25 rpm. Organoids were washed three times for 15 min with PBS and stained organoids were mounted by adding 200 µl of Vectashield mounting medium (VECTASHIELD HardSet Antifade Mounting Medium, H-1400, Vector Laboratories). Organoid imaging was performed on an inverted Leica SP8 Dive system (Leica Microsystems), in which Alexa-647 secondary antibodies were excited at 635 nm and detected between 660 and 700 nm, and organoids were imaged using brightfield. The Ki-67/CC3 ratio was derived by first calculating the organoid area and Ki-67⁺ or CC3⁺ areas using ImageJ software (https://imagej.nih.gov/ij/), then calculating the percentage of Ki-67 or CC3 expressing cells per organoid, followed by calculation of the Ki-67/CC3 ratio for every organoid.

Statistics

P values and statistical tests performed are included in the figure legends or Supplementary Information 1. The longitudinal data of the clone fractions (Figs. 2d and 5g) was analysed using a regression model with a time effect, for which the interaction between time and group was tested. For full details, see Supplementary Information 2. Details on statistics concerning the mathematical modelling can be found in the Supplementary Information 4.

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.