Fig 1.
MCF10A lumen clearance and acini size are EGF triggered.
Representative confocal images through the equatorial plane of acini illustrate the influence of EGF on acinar morphology. A. Acini grown for 33 days were stimulated with EGF for the first 9 days. B. A comparative sample with continuous EGF supplementation for 33 days. Scale bars = 20 μm. C. Plot illustrates the relation between lumen formation frequency and temporal EGF stimulation. D. Box plot analysis shows the relation of acini diameter to temporal EGF supplementation. At least 10 spheres were analyzed per sample. Horizontal lines: group medians. Boxes: 25–75% quartiles. Vertical lines: range, peak and minimum. Hatched bars indicate best combinatory effect size. ***: p < 0.0001.
Fig 2.
Capturing the differentiation process of MCF10A acini.
A. Equatorial cross sections through MCF10A acini demonstrate the differentiation process leading to apical and basal polarization. Basal polarization markers: laminin-5, collagen IV (BM formation). Apical polarization markers: GM-130 (Golgi apparatus). Proliferating-nuclear-antigen PCNA and active caspase-3 marked s-phase and apoptotic cells, respectively. DRAQ5 was used to counterstain cell nuclei, cytoskeletal F-actin was stained with actin-phalloidin. Note that stains were performed on different spheres for low (1–12 days)-, semi (13–24 days)-, highly (25–35 days)-matured states. Scale bars = 20 μm. B. Immunofluorescence data were used to infer the development of acinar structures depending on temporal EGF withdrawal. Temporal progression is highlighted with arrows. Acini were grouped according to their differential grades.
Fig 3.
Tight junction formation in MCF10A cells depends on culture conditions.
A. The tight junction (TJ) protein ZO-1 is visible as characteristic kissing points in 2D MCF10A monolayer cluster (white arrow head). B. Representative immunofluorescence stains for low- and highly-matured acini. Cortical F-actin signals (green) indicate acinar cell boundaries. ZO-1 protein (red) was only present as diffuse signal within the cytoplasm, irrespective of acinar-maturation states. The merged ZO-1 and F-actin stains demonstrate loss of colocalization and the absence of TJ-complexes. C. β-catenin signals indicate homogeneously distributed cell-cell contacts (adherence junctions) in both 2D MCF10A monolayer and in 3D MCF10A acini. Scale bars = 20 μm.
Fig 4.
The basement membrane retards molecule influx into MCF10A acini.
A. Colocalization of the BM specific laminin-5 protein with dextran molecules demonstrated the accumulation of dextran within the BM scaffold. The white arrowhead indicates a cytoplasmic localization (after 2 days) of dextran tracer (40 kDa). B. Representative image sequence of in situ dextran permeation in MCF10A acini with highly-developed BM (EHS-matrix embedded). Black arrows indicate first temporal appearance and spatial localization of dextran (10 kDa) (contrast inverted). C. Principal time course of molecule influx through the BM into MCF10A acini. Red: dextran, grey cytoplasm, blue: nuclei. Scale bars = 20 μm. Plots illustrate a semi-quantitative analysis of dextran permeation. Average fluorescence intensity profiles are shown for (D.) acini with low-developed BM and with (E.) highly-developed BM. Data are normalized to the background plateau for each molecular weight (see S2 Fig). Mean values with s.d. are shown. Full lines represent fits to a semi-empirical law (see S1 Protocol). Please note that absolute signal intensities of low- and semi-matured acini are not directly comparable because increased autofluorescence imposed the use of an adapted masking algorithm for the semi-matured sample group.
Fig 5.
Intact collagen IV meshwork is essential for size-dependent molecule retardation.
A. Control stain of BM specific collagen IV protein after enzymatic collagenase IV treatment. Comparative detection of laminin-5 protein demonstrated the colocalization of residual collagen IV protein within the remaining BM structure. B. The time-lapse of dextran influx in collagenase IV treated MCF10A acini with highly-developed BM (semi-matured group) is independent of the dextran tracer size. Left panel (0 min): bright field images. Other panels: representative dextran signal appearance (contrast inverted images) at indicated locations was independent from dextran molecular weights. IC: Intercellular cleft signal. Scale bars = 20 μm.
Fig 6.
The basement membrane sustains the mechanical integrity of breast acini.
The basement membrane sustains structural integrity without cellular networks. A-A”. Representative MCF10A acini embedded in EHS matrix were treated with OGP detergent. Acinar swelling and cell-BM separation are indicated by newly formed cell-free clefts during the first seven minutes (see inset). A”‘. Finally, acinar structures shrank to initial size. B. Only sparse contact points were visible between BM (red) and cell debris (green) after OGP-treatment. Nuclei (blue) were counterstained with DRAQ5. Scale bars = 20 μm. C. Schematics of the experimental AFM set up. D. AFM approach-retraction cycles as performed on the same 17-days old MCF10A acinus before (black) and after (red) OGP-treatment. E. Force values needed to reach a certain indentation depth into the acinus as directly read from the raw data in plot B. F. Plot illustrates the force distributions for increased indentation depths in native acini with low-developed BM (n = 40), native acini with highly-developed BM (n = 40) and OGP-treated acini with highly-developed BM (n = 31). Mean values with s.d. are shown. Differences between native and OGP-treated samples are highly significant (p < 0.001).