Fig 2: OPN loss impairs growth of basal-like PDA.(A) CRISPR-Cas9-mediated knockout of SPP1 in KP4 cells leads to a decrease in mRNA expression and secreted Osteopontin in two independent clones. (B,C) Immunoblot of KRT14 and E-cadherin (B) and GATA6 (C) levels following KO of SPP1 in KP4 cells. (D) Brightfield images of KP4 cell morphology when grown in 2D monolayer following KO of SPP1. Note the switch to a more epithelial phenotype in the KO cells. Scale bars, 100 µm. (E) In vitro growth rate of KP4 control versus SPP1 KO cell lines. Error bars represent s.e.m. (F) Growth rate of control and SPP1 KO KP4 cells following subcutaneous injection into the flank of NOD/SCID mice. Error bars represent s.e.m. (G) Relative size of control and SPP1 KO KP4 xenografts resected at day 25 (left) and comparative tumor weight (graph at right). (H) Representative H and E stained sections of lungs at endpoint following tail vein injection (TVI) of 2 × 106 KP4 control cells (WT; top) or SPP1 KO (bottom) into NOD/SCID mice. Scale bars, 400 µm. (I) Quantification of metastatic tumor burden in the lungs as a percentage of total lung area per mouse (n = 5 CTRL; n = 10 SPP1 KO). (J) Kaplan-Meier graph depicting survival of NOD/SCID mice following TVI of KP4 WT (black) or SPP1 KO cells (red). (K) High expression of SPP1 in basal-like PDA patient tumors from TCGA predicts shorter overall survival (SPP1 high n = 17, SPP1 low n = 48). Data from 65 patients. (L) Model depicting a classical to basal-like switch mediated by a GLI2-OPN signaling axis and in response to KRASG12D ablation. The classical and basal-like subtypes are marked by GLIlo/Hhhi and GLIhi/Hhlo expression, respectively. p-Value calculated by Log-rank test and by two-tailed unpaired t test. *p<0.05; **p<0.01; ***p<0.001.10.7554/eLife.45313.022Figure 7—source data 1.OPN loss impairs growth of basal-like PDA.

Fig 3: Loss of epidermal homeostasis and skin barrier integrity in postnatal inducible whole-body Skiv2l knockout mice.(A) A schematic diagram showing experimental design for generation of postnatal tamoxifen-inducible whole-body Skiv2l knockout mice (iSkiv2l–/–) by intraperitoneal tamoxifen injection in Skiv2lfl/flUBC-cre/ERT2 mice. (B) Skin lesion and hair loss of iSkiv2l–/– mice at indicated ages. (C) H&E staining of iSkiv2l–/– mouse dorsal skin and Skiv2lfl/fl littermate controls (3 months old). White arrows, enlarged keratinocytes with loose chromatin. Quantification of epidermal thickness of iSkiv2l–/– and Skiv2lfl/fl is shown on the right bar graph. n = 5 mice per genotype. Unpaired 2-sided Student’s t test, ***P < 0.001. (D) A schematic diagram showing epidermal layers. (E) Fluorescence immunohistochemistry analysis of proliferation marker Ki67 (green signaling in the nucleus) of iSkiv2l–/– mouse dorsal skin and Skiv2lfl/fl littermate controls (3 months old). Nuclei were counterstained with DAPI (blue). White arrows, Ki67 positive cells. Asterisk indicates nonspecific staining of stratum corneum of epidermis. Isotype IgG was used as a negative control. Dashed line, epidermal-dermal junction. Scale bar: 50 µm. (F) Cell-cycle analysis of keratinocytes isolated from iSkiv2l–/– mice and Skiv2lfl/fl littermate controls (3 months old). Statistical analysis of cell-cycle distributions are shown on the right. n = 4 per genotype. Unpaired 2-sided Student’s t test, **P <0.01, ***P < 0.001. (G) Fluorescence immunohistochemistry analysis of keratinocyte differentiation markers K14, K5, and K10 in iSkiv2l–/– mouse dorsal skin and Skiv2lfl/fl littermate controls (3 months old). Nucleus was stained with DAPI (blue). K14, keratin 14; K5, keratin 5; K10, keratin 10. Dashed line, epidermal-dermal junction. Scale bar: 50 µm. (H) Representative Toluidine blue staining of depilated iSkiv2l–/– mice and Skiv2lfl/fl littermate controls. (I) TEWL of iSkiv2l–/– mice (n = 20) and Skiv2lfl/fl littermate controls (n = 11). Unpaired 2-sided Student’s t test, ***P < 0.001.

Fig 4: PD-L1+ basal cells can produce basal, luminal, and alveolar cells during mammary development. (A) Illustration of the lineage-tracing strategy. (B) Control, uninduced mice. (C) Whole-mount confocal imaging analysis showing that no GFP+ cells were present in the mammary ducts of uninduced PD-L1CreERT2–2A–tdTomato/+; Rosa26mTmG/+ mice. Scale bars, 20 µm. (D) FACS analysis indicating that no GFP+ cells were present in the mammary epithelium of uninduced mice. (E) Experimental design for the short-term tracing experiments; tamoxifen (TAM) was administered at the pubertal stage (5 weeks old). (F) Whole-mount confocal microscope images showing an elongated GFP+ basal cell. Scale bars, 10 µm. (G) Section imaging indicating the basal location of a GFP+ cell. Scale bars, 10 µm. (H) FACS analysis indicating that GFP+ cells were mainly basal cells at 48 h after tamoxifen administration. (I–L) Experimental design for the long-term tracing experiments (I), imaging (J,K), and FACS analysis (L) of the distribution of the GFP+ cells at 3 weeks after TAM administration. Scale bars, 10 µm. (M–P) Experimental design for the long-term tracing experiments (M), Imaging (N,O), and FACS analysis (P) of the GFP+ cell distribution at pregnancy day 14.5. Scale bars, 10 µm. K14+ cells are indicated by arrows. K8+ cells are indicated by arrowheads; n = 5 mice. FACS, fluorescence-activated cell sorting; PD-L1, programmed cell death ligand 1; GFP, green fluorescent protein.

Fig 5: Dysregulation of epidermal development in in vitro cultured Stat3 ?K5 keratinocytes. (A) Time-dependent changes in in vivo mRNA expression of Lor, Flg, Krt1, Krt10, Ivl, Krt5, and Krt14 are shown as lines for K5-cre (blue), and KO-D (green) mice. (B) Histological distribution of keratin 1 and 14, and filaggrin protein in ear skin of WT and KO-D mice. Green and blue colors show Ki-67 expression and nuclei, respectively. (C) Transepidermal water loss (TEWL) from the ear of mixed background Stat3f/f (n=4) or Stat3 ?K5 (n=5) mice was measured at 16 weeks of age. (D) Hematoxylin and eosin staining of 3D-cultured keratinocytes (day 14). Keratinocytes were isolated from mixed background Stat3-sufficient (K5-cre) or deficient (KO) newborn mice. After 3 days of culture in liquid medium (2D-culture), keratinocytes were induced to undergo 3D skin structure formation by exposing the apical layer to air, and cells were cultured for another 7 or 14 days. (E) Heatmap depicting in vitro keratin gene expression in 2D (day 0) and 3D (day 7 and 14) cultures. (F) Heatmap depicting in vitro cell proliferation-related gene expression in 2D (day 0) and 3D (day 7 and 14) cultures. (G) Time-dependent changes in gene expression of Dnajc8, Eif3l, Endog, Entpd2, Park2, Rpl13a, Rpl22, and Rpl26 in the ear skin of mixed background (left) or B6 background (right) Stat3-sufficient (K5-cre for mixed background, Stat3 f/f for B6 background) and deficient (KO) mice are shown as lines for Stat3-sufficient (blue, n=4) and Stat3-deficient (KO, green, n=4) mice. Scale bar = 100 µm. *p<0.05, **p<0.01.