Fig 2: Deprivation of NO in non-malignant mammary epithelial cells in 3D cultures induces disorganized colonies, accompanied by induction of TGFβ, ERBB2 and senescence markers. (A) Representative images of MCF10A cells cultivated in 3D ECM under treatment of control (PBS), L-arginine or L-NAME for 3 weeks. Top 2 rows: phase images. Middle 3 rows: staining for the basal marker, integrin α6 (ITGA6) or apical marker, GM130. Bottom 2 rows: staining for lumen marker, cleaved caspase 3 (CC-3). Nuclei were counterstained with DAPI (blue). Scale bars: 20 μm. (See the quantification of colony size, and plot profiles of ITGA6, GM130 and CC-3 signals in Supplementary Fig. 5A,B). (B) Representative images of phospho-SMAD3 (top 2 rows), ERBB2 (rows 3–4), p21 (rows 5–6) or β-Gal (bottom 2 rows) staining of drug-treated MCF10A cells in 3D cultures. Nuclei were counterstained with DAPI (blue). Scale bars: 20 μm. (See the quantification in Supplementary Fig. 5C).

Fig 3: Effects of TGFßR1, SMAD3, and MMP13 inhibition on TRAP staining and effects of MMP13 overexpression (OE) on jaw length.Quail lower jaws harvested at HH35 and placed in culture with control beads (left side) and treatment beads (right side) soaked in (A) TGFßR1 inhibitor (TGFßi, n=5), (B) SMAD3 inhibitor (SMAD3i, n=5), and (C) MMP13 inhibitor (MMP13i, n=5). The effects of inhibitor treatments can be seen on TRAP staining (red) after 5 days of culture (red dashed lines). Quantification of the % area of TRAP-positive staining for (D) TGFßR1 inhibitor (medium gray), (E) SMAD3 inhibitor (white), and (F) MMP13 inhibitor (dark gray) in comparison to the contralateral control side. (G) In ovo electroporations of the pPIDNB-Mmp13 OE construct were performed at HH8.5 and embryos were allowed to develop until HH35 when they were treated with a single dose of dox. Electroporation efficiency and extent of OE were evaluated during embryo collection at HH40 by detecting RFP (red). Microcomputed tomography (µCT) analysis of duck specimens electroporated with (H) empty vector or (I) Mmp13 OE construct. A 3D model was annotated with landmarks from which (J) lower jaw distances were calculated. Embryos overexpressing Mmp13 showed significantly shorter jaws compared to empty vector controls. * denotes significance from control within each group p=0.05, ** denotes significance from control within each group p=0.01, and *** denotes significance from control within each group p=0.001. Figure 5—source data 1.Effects of TGFßR1, SMAD3, and MMP13 inhibition on TRAP staining in quail embryos. Figure 5—source data 2.Effects of MMP13 overexpression on jaw length in duck embryos as visualized by µCT.

Fig 4: Sensitivity of chick and duck cells to rTGFß1 and effects of TGFßR1 and SMAD3 inhibition on RUNX2 and MMP13.(A) pSMAD3 protein levels in cells treated for 2 hr with 5 ng/ml rTGFß1 (dark gray) in chick (blue) and duck (yellow) cells show a significant induction in chick (n=8). (B) Chick and duck cells treated with rTGFß1 for 1–24 hr. Chick Runx2 mRNA increases with rTGFß1 treatment at 3 and 6 hr and at 24 hr, while duck Runx2 does not increase until 24 hr. In rTGFß1 treated cells, duck have significantly lower Runx2 at every time point compared to chick (n=8). (C) Chick Mmp13 mRNA increases with rTGFß1 treatment at 6 hr and at 24 hr, while duck Runx2 does not increase until 24 hr. In rTGFß1 treated cells, duck have lower Mmp13 at every time point compared to chick (n=8). (D) MMP13 protein in cells treated for 24 hr with rTGFß1 shows an induction in chick but no response in duck (n=8). (E) RUNX2 protein levels in chick cells treated for 24 hr with rTGFß1 (dark gray), TGFßR1 inhibitor (medium gray), a combination of both rTGFß1 and TGFßR1 inhibitor (black), SMAD3 inhibitor (white), and a combination of both rTGFß1 and SMAD3 inhibitor (light gray). RUNX2 protein increases with rTGFß1, but when rTGFß1 is combined with either a TGFßr1 or a SMAD3 inhibitor, there is a significant decrease compared to rTGFß1 alone (n=12). (F) MMP13 protein increases with rTGFß1 treatment, but when rTGFß1 is combined with either a TGFßR1 or a SMAD3 inhibitor, there is a significant decrease compared to rTGFß1 alone (n=12). * denotes significance from control p=0.05, ** denotes significance from control p=0.01, *** denotes significance from control p=0.001, # denotes significance from rTGFß1, † denotes significance between chick and quail, and ‡ denotes significance between chick and duck.

Fig 5: Deprivation of NO in developing mouse mammary glands stiffens the ECM and induces precancerous lesions, accompanied by upregulation of TGFβ and ERBB2 and induction of senescence. (A) Whole mounts of carmine-stained 4th mammary glands of mice after i.p. treatment with control (PBS), L-arginine or L-NAME for 6 weeks (from 4 weeks to 10 weeks old). Scale bars: top, 10 mm; bottom, 200 μm. (B) Representative images of S-NOC staining (top), SHG (collagen) (second row) and phospho-SMAD3 (TGFβ signal) staining (bottom 2 rows) of drug-treated mammary glands. Nuclei were counterstained with DAPI (blue). Scale bars: 50 μm. (See the original micrographs and quantification of S-SNOC, SHG and phospho-SMAD3 levels in Supplementary Fig. 4A–C). (C) H&E staining of drug-treated mammary glands. Note that intraductal papillomas are indicated with pink asterisks (*). Scale bars: 50 μm. (D) Representative images of staining for ERBB2 (top), HP1 (second row), β-Gal (third row), p21 (fourth row) and p27 (bottom row) of drug-treated mammary glands. Nuclei were counterstained with DAPI (blue). Scale bars: 50 μm (See the quantification of ERBB2, HP1, β-Gal, p21 and p27 staining in Supplementary Fig. 4D–H).