Fig 2: Gemcitabine induces tumor fibrosis through increased PlGF/VEGF-A expression, which activates cancer-associated fibroblasts (CAFs).On day 10 when tumor volume reached 50–100 mm3, tumor-bearing mice were intraperitoneally treated with either PBS (control, n = 6) or gemcitabine (50 mg/kg every 3 days for five times, n = 6). a Tumor volumes of control (Con) and gemcitabine (Gem) treated mice. ns, no significance. Two-way ANOVA with Sidak’s multiple comparisons test was performed. b Tumor weights of control (Con) and gemcitabine (Gem) treated mice (n = 6/group) compared at the end of the experimental period. No statistical difference was observed. c Representative images showing second-harmonic generation (SHG) signals, which indicate the distribution of collagen. SHG signals were obtained at 840 nm and pseudo-colored with green. Scale bar, 100 µm. d Average percentages of collagen+ area out of the total area in tumors treated with PBS (control) or gemcitabine (n = 25/group). For quantification, four to five SHG images of 512 µm in width and height were obtained per tumor. P = 0.013. e Relative Plgf and Vegfa mRNA expression levels of Pan02 tumor tissue homogenates measured by qRT-PCR. Data from four biological repeats for control group and five biological repeats for gemcitabine group were analyzed. Plgf: P = 0.022, Vegfa: P = 0.509. ns, no significance. f Relative Plgf and Vegfa protein levels of Pan02 tumor tissue homogenates measured by ELISA. Data from seven biological repeats (n = 7) were analyzed. PlGF: P = 0.008, VEGF-A: P = 0.133. ns, no significance. g Expression of PlGF/VEGF-related receptors (NRP1; VR1, VEGFR1; VR2, VEGFR2) in CAFs. h RT-PCR detection of myCAF markers, Col1a1, Acta2, and Fn1 transcripts, in CAFs in response to PlGF, or VEGF-A. After sequential serum deprivation for sorted CAFs, the indicated proteins (PlGF, VEGFA) and atezolizumab or Ate-Grab were added and co-incubated for 6 h. Data from three biological repeats (n = 3) were analyzed. Col1a1: P = 0.005, Acta2: P = 0.030, Fn1: P < 0.001. All Data are presented as the mean ± SEM. *P < 0.05, **P < 0.01, ***P < 0.001 vs. control; two-tailed Student’s t-test (b, d, e, f, h). Source data are provided as a Source Data file.

Fig 3: Activation of PSCs impaired upon CM5480 treatment in vitro.(A) Representative scheme of in vitro culturing of mouse PSC culture marking the CM5480 treatments during the 1-week time course. (B) Immunofluorescence staining and relative intensity values of VIM, Orai1, GFAP, and a-SMA in cultured PSCs. Scale bar: 100 µm. (C) Representative images and ratio of Oil Red O+ cells in cultured PSCs. n = 4/group, 5–7 images/culture, data points in the bars indicate individual field of views (B and C). Scale bar: 20 µm. (D) mRNA expression of Orai1, Acta2, Tgfb1, Saraf, and Fn1 in cultured PSCs (n = 4/group from 4 animals). (E) Representative images and relative intensity values of Orai1 and percentage of PSCs with spontaneous Orai1 puncta formation in the control culture (n = 4/group, 5–7 images/culture). Scale bar: 10 µm. A P value less than 0.05 was considered significant by unpaired t test (B and D [Orai1, Fn1], and E) and Mann-Whitney U test (C and D [Saraf, Tgfb1, Acta2]). Data represent mean ± SEM. Explanatory images were created with BioRender.

Fig 4: Orai1 inhibition impairs the activation of PSCs in CP.(A) Immunofluorescence staining and relative intensity values of a-SMA and VIM in mouse pancreatic tissue (n = 3/group, 5–7 images/animal). Scale bar: 20 µm. (B) Acta2 gene expression in pancreatic tissue (n = 3/group). (C) IHC images of a-SMA (top) and GFAP (bottom) in mouse pancreatic tissues (n = 3/group, 5–7 images/animal). Scale bar: 50 µm. Bar charts summarize the ratio of a-SMA+ and GFAP+ and GFAP/a-SMA–double-positive cells (activated PSCs) (orange bars) in percentage of GFAP+ total cell number (bars with dashed line). A P value less than 0.05 was considered significant by unpaired t test (A) and Mann-Whitney U test (B and C). Data represent mean ± SEM. Explanatory images were created with BioRender. cer., cerulein; p.s., physiological saline.

Fig 5: Ate-Grab is a multi-paratopic decoy receptor that can target both PlGF/VEGF-A and the PD-1/PD-L1 axis.a Schematic diagram of the Ate-Grab design. The single-chain variable fragment of atezolizumab with a [G4S]3 linker was fused to the N-terminus of VEGF-Grab. b Representative image of SDS-PAGE analysis on VEGF-Grab and Ate-Grab. Both reducing and non-reducing conditions were used for 10% acrylamide gel electrophoresis, and products were visualized through Coomassie blue staining. R, Reducing condition; NR, Non-reducing condition. c Target binding affinities of Ate-Grab and template drugs assessed via ELISA. d Western blot analysis for phospho-VR2, VR2, phospho-ERK, ERK, and ß-actin after treatment with IgG, VEGF-Grab (VG), Ate-Grab (ATG), or without any treatment (control) in HUVECs incubated with VEGF-A. Total and phosphorylated levels of VEGFR2 (VR2) and ERK were detected. The experiment was repeated for three times independently with similar results. Relative quantities of e phospho-VEGFR2/VEGFR2 and f phospho-ERK/ERK across control, IgG, VG, and ATG groups. Data from three biological repeats were analyzed (n = 3/group). g Representative images from wound-healing assay depicts tube formation of HUVECs treated with each indicated protein combination (25 nM) in the presence of VEGF-A (1 nM). Scale bar, 500 µm. h Relative quantities of tube-formed area from the tube formation assay. Data from three biological repeats were analyzed (n = 3/group). i PD-1/PD-L1 blockade bioassay for Ate-Grab and atezolizumab. j Relative expression levels of myCAF (Col1a1, Acta2) and apCAF (Cd74, H2-Ab1) markers in CAFs from Pan02 tumors. CAFs were isolated from Pan02 tumors via fluorescence-activated cell sorting (FACS) and treated with PlGF or the indicated drug combinations with PlGF for 6 h after the sequential serum deprivation. Data from three biological repeats were analyzed (n = 3/group). Data are presented as the mean ± SEM. *P < 0.05 (Col1a1-Control vs Ate-Grab+PlGF in j: P = 0.029), ***P < 0.001 versus control (e, f, h, j). Statistical significance was accessed by One-way ANOVA with Tukey’s multiple comparisons (e, f, h, j). Source data are provided as a Source Data file.