1
Inhibitor Manufacturer Target
BI-5521 Boehringer Ingelheim GSK3
Dasatinib hydrochloride Selleck Chemicals SRC family kinases (SFKs)
Defactinib Selleck Chemicals FAK
FRAX597 Selleck Checmicals Class I PAKs (PAK1, PAK2, PAK3)
LY294002 Calbiochem PI3K
LY364947 Calbiochem TGFβ Type I Receptor
ML-7 hydrochloride Calbiochem MLCK
NSC23766 Calbiochem RAC
Rhosin Calbiochem RHO
SB203580 Calbiochem p38
U-0126 Calbiochem MEK
Y27632 Calbiochem ROCK
Supplementary Table 1: Inhibitors used in this study. The inhibitors used in this study are listed together with the manufacturer and the protein(s) targeted by the inhibitor.
2
Antigen Antibody Clone RMSD
E-Cadherin HPA004812 19.72%
E-Cadherin CAB000087 23.06%
E-Cadherin CAB028364 13.42%
E-Cadherin CAB072855 12.67%
E-Cadherin CAB072856 24.98%
E-Cadherin CAB072857 13.87%
Keratin 19 CAB000031 10.73%
Keratin 19 HPA002465 16.44%
Supplementary Table 2: Intra-patient variation of the proportion of fibrotic elements.
The root mean square deviation (RMSD) was calculated between biopsies for the same patient across all antibodies for which data is presented in Figure 1B, Supplementary Figure 1A, and/or Supplementary Figure 1C.
3 Supplementary Figure 1: Variation of the stromal proportion is comparable across antibodies and antigens used. (A) Quantification of the percentage of tumor area occupied by stromal elements for biopsy specimens stained by all anti-E-cadherin antibodies in the Human Protein Atlas. Note that data for the antibody CAB072856 shown in Figure 1B is again shown here to facilitate comparison between the various antibodies for which data were available. (B) Representative Keratin 19 staining (brown) images of human PDAC biopsy specimens demonstrating varying proportion of tumor area occupied by stromal elements. Images were obtained from the Human Protein Atlas (www.proteinatlas.org) [5,6].
(C) Quantification of the percentage of tumor area occupied by stromal elements for biopsy specimens stained by all anti-Keratain 19 antibodies in the Human Protein Atlas. In (A) and (C), data are shown as a box-and-whisker plot demarcated by quartiles overlaid on individual measurements.
4 Supplementary Figure 2: Expression of epithelial markers in various PDAC cell-lines.
(A-D) The PDAC cell-lines Capan-2 (A), BxPC-3 (B), SUIT-2 (C), and MiaPaCa-2 (D) cultured in 2D on glass-bottom dishes were stained for the epithelial markers E-cadherin (green) and Cytokeratin (red). Nuclei (Hoechst 33342) are shown in blue. Scale bars = 25 μm.
5 Supplementary Figure 3: Cancer cell organization within 3D PDAC tissues correlates with epithelial differentiation status of PDAC cell-lines. (A and B) Human PDAC cell- lines, SUIT-2 (A) and MiaPaCa-2 (B), were used to create heterotypic 3D PDAC tissues by co-culturing with normal human dermal fibroblasts (NHDFs). PDAC cells were seeded at various ratios ranging from 1/10 (5x104 cells) to 1/5000 (1x102 cells) against a fixed number (5x105 cells) for NHDFs. The 3D PDAC tissues were stained for Cytokerain (pan-CK; green), α-SMA (red), and nuclei (Hoechst 33342; blue). Scale bars = 100 μm.
6 Supplementary Figure 4: E-cadherin expression in PDAC cells is necessary for tumor nest formation in 3D PDAC tissues. (A) RT-qPCR was performed to analyze the knockdown efficiency in Capan-2 cells treated with siRNA targeting CDH1 (siCDH1). Data shown as mean ± S.D. of n = 3 biological replicates. (B) Capan-2 cells with or without siCDH1 treatment were stained for E-cadherin (green). Nuclei (Hoechst 33342) are shown in blue.
Scale bars = 25 μm. (C) 3D PDAC tissues were generated by co-culturing Capan-2 cells treated with or without siCDH1 treatment with fibroblasts. 3D PDAC tissues were collected on day 8 of culture and stained for E-cadherin (green), Cytokeratin (red), and nuclei (Hoechst 33342; blue). Scale bars = 100 μm.
7 Supplementary Figure 5: Generation of heterotypic 3D PDAC tissues using human PDAC patient-derived pancreatic stellate cells (PSCs). (A-C) Heterotypic 3D PDAC tissues were generated by co-culturing Capan-2 cells with primary PSCs derived from three different human PDAC patients. PDAC cells were seeded at various ratios ranging from 1/10 (2.5×104 cells) to 1/5000 (0.5×102 cells) against a fixed number (2.5×105 cells) for PSCs. The 3D PDAC tissues were stained for E-cadherin (green), α-SMA (red), and nuclei (Hoechst 33342; blue). Scale bars = 100 μm. (D-F) Quantification of area occupied by stromal elements, shown as percentage of total area for 3D PDAC tissues created using PSCs derived from different patients as shown in (A-C). Data shown as mean ± S.D. of measurements on n = 4 image fields for each experimental condition (seeding ratio).
8 Supplementary Figure 6: Visualization of extracellular matrix (ECM) structures in 3D PDAC tissues. 3D PDAC tissues generated from co-culturing Capan-2 cells with NHDFs (first row) or PSCs derived from three different PDAC patients (second to fourth rows) were stained for the ECM components Collagen (A-D), Fibronectin (E-H), or hyaluronic acid (I-L).
In addition, E-cadherin (green) and nuclei (Hoechst 33342; blue) were stained. The seeding ratio adopted for the data presented in this figure is 1/1000, 1/50, 1/500, and 1/500 for NHDF, PSC #1, PSC #2, and PSC #3, respectively. Scale bars = 100 μm.
9 Supplementary Figure 7: RT-qPCR confirmation of knockdown efficiency for siRNAs used against fibroblasts in this study. (A, C, D, and F) RT-qPCR analyses were
performed to confirm the knockdown efficiency in NHDFs treated with siYAP (A), siSMAD2 (C), siSMAD3 (D), and siACTA2 (F). Data shown as mean ± S.D. of n = 3 biological replicates. (B, E, and G) Western blots were performed to confirm the successful
knockdown of YAP, SMAD2, SMAD3, or α-SMA in NHDFs treated with siYAP (B), siSMAD2 or siSMAD3 (E), and siACTA2 (G).
10 Supplementary Figure 8: YAP nuclear localization in fibroblasts is not sufficient for acquisition of a robust myofibroblastic phenotype. (A) 3D tissues consisting of only NHDFs collected after 4, 6, or 8 days of culture were stained for α-SMA (red), YAP (green), and nuclei (Hoechst 33342; blue). Scale bars = 100 μm. (B and C) Quantification of the normalized fluorescence intensity of α-SMA (B) and nuclear YAP (C) in (A).
11 Supplementary Figure 9: YAP-dependent TGF-β signaling through SMAD2/3 induces α-SMA expression in fibroblasts cultured in 3D. (A) 3D tissues were generated from NHDFs with or without siRNAs targeting either SMAD2 or SMAD3. The resulting 3D tissues were treated with TGF-β3 starting from day 4 and collected after 8 days of culture. The 3D tissues were stained for α-SMA (red) and nuclei (Hoechst 33342; blue). (B) Quantification of the normalized fluorescence intensity of α-SMA in (A). (C) 3D tissues were generated from NHDFs with or without treatment with siYAP. The resulting 3D tissues were treated with TGF-β3 starting from day 4 and collected after 8 days of culture. The 3D tissues were stained for α-SMA (red) and nuclei (Hoechst 33342; blue). (D) Quantification of the
normalized fluorescence intensity of α-SMA in (C). Scale bars = 100 μm.
12 Supplementary Figure 10: TGF-β signaling is necessary for the acquisition of a myofibroblastic phenotype and is dependent on fibroblast YAP activity in 3D PDAC tissues. (A) 3D PDAC tissues were generated by co-culturing Capan-2 cells with NHDFs with or without treatment with LY364947 (inhibitor of TGF-β signaling) starting from day 4 of culture. 3D PDAC tissues were collected on day 8 of culture and stained for α-SMA (red), SMAD2/3 (green), and nuclei (Hoechst 33342; blue). (B) Quantification of the normalized fluorescence intensity of α-SMA in non-tumor areas of (A). (C) 3D PDAC tissues were generated by co-culturing Capan-2 cells with NHDFs with or without treatment with siYAP.
The resulting 3D PDAC tissues were treated with TGF-β3 starting from day 4 and collected after 8 days of culture. The 3D tissues were stained for α-SMA (red), SMAD2/3 (green), and nuclei (Hoechst 33342; blue). (D) Quantification of the normalized fluorescence intensity of α-SMA in non-tumor areas of (C). Regions encircled by dotted lines indicate tumor areas in (A) and (C). Scale bars = 100 μm.
13 Supplementary Figure 11: SMAD2/3 fails to localize in the nucleus of fibroblasts in 3D fibroblast monoculture. (A) 3D tissues consisting of only NHDFs were collected after 4, 6, or 8 days of culture. The 3D tissues were stained for α-SMA (red), SMAD2/3 (green), and nuclei (Hoechst 33342; blue). Scale bars = 100 μm. (B) Quantification of the mean fluorescence intensity of α-SMA in (A).
14 Supplementary Figure 12: Pharmacological inhibition of intracellular signaling
effectors reveals differential regulation of nuclear localization of YAP versus SMAD2/3 in 3D PDAC tissues. (A and B) 3D PDAC tissues treated with various
pharmacological inhibitors starting from day 4 of culture were collected on day 8. Inhibitors used were U-0126 (inhibitor of MEK), BI-5521 (GSK3), SB203580 (p38), and LY294002 (PI3K). DMSO was used as control. 3D PDAC tissues were stained for YAP or SMAD2/3 (green) in (A) and (B), respectively. In addition, α-SMA (red) and nuclei (Hoechst 33342;
blue) were stained. Scale bars = 100 μm. (C-E) Quantification of nuclear YAP (C), nuclear SMAD2/3 (D), and the normalized fluorescence intensity of α-SMA (E) in non-tumor areas.
Regions encircled by dotted lines indicate tumor areas in (A) and (B). Scale bars = 100 μm.
