Fig 2: Effects of CM from WT or Pdia3−/− MEFs on attachment and spreading of breast cancer cell lines(A,C,E) Cell areas for each cell line (HCC1937 (A), MDA-MB-231 (C) or MCF-7 (E) under the same conditions as in Figure 5. (B,D,F) Cell attachment scored from nuclei per 63x microscope frame. In each dot plot, the horizontal line shows the median. (G) Data from time courses as change in % RWD for MDA-MB-231 cells plated for 24 h under the indicated conditions. For each condition, the mean (solid line) SD (dotted line) are shown. Data are from two experiments. (H,I) Analyses of the time course data, as box plots of ‘AUC’ (H), or ‘Rate of Initial Wound Closure’ (I). In (H,I), each dot represents the value calculated for each metric from each well in each independent experiment. Within each box plot, each vertical column represents a single independent experiment. All data analysed by two-way ANOVA with Tukey’s post hoc analysis in R and plotted using ggplot2.

Fig 3: PDIA3 modulates cell adhesion through CCN2. A: representative immunoblots showing higher amounts of CCN2 protein in total cell lysate (TCL) and heparin-agarose pulldowns of CM (CM + Hep) from WT-MEFs than in equivalent samples from Pdia3-/- MEFs. GAPDH was used as a loading control for TCL. B: quantified analysis of CCN2 levels in TCL or heparin-agarose pulldowns of CM (CM + Hep) as in A. CCN2 bands (asterisks) from TCL were normalized against GAPDH band intensity. For each pair of samples, the WT sample was set as 1 and the Pdia3-/- sample normalized against this. Each column represents the mean and error bars the standard deviation. Analysis by paired t test. From n = 3 experiments. C and D: recombinant CCN2 confers adhesion-promoting activity on CM from Pdia3-/- MEFs. C: representative immunofluorescence images of F-actin and vinculin (the merged images also include DAPI stain) in Pdia3-/- MEFs (KO) plated for 2 h with CM from Pdia3-/- MEFs (KOCM), or KOCM plus the indicated concentrations of recombinant CCN2, examined by confocal microscopy and presented as XYZ stacks. Scale bar = 20 µm. D: quantified analysis of cell areas of Pdia3-/- MEFs (KO) under the conditions as indicated, from n = 3 experiments. E: immunoprecipitation of CCN2 or fibronectin (FN) from 48 h WTCM, followed by immunoblotting for CCN2 (top) or fibronectin (bottom). Nonimmune rabbit IgG (Con) was used as a negative control. F and G: immunodepletion of CCN2 from WTCM [WTCM(-CCN2_IP)] impairs its adhesion-promoting activity on Pdia3-/- MEFs. F: representative confocal immunofluorescence images for actin and vinculin (the merged images also include DAPI stain) in Pdia3-/- MEFs (KO) incubated with WTCM either after immunodepletion with a control IgG [WTCM(-IgG_IP)] or with anti-CCN2 [WTCM(-CCN2_IP)], presented as XYZ stacks. Scale bar = 20 µm. G: quantified analysis of cell areas of Pdia3-/- MEFs plated as in F under the indicated conditions for 2 h. From n = 3 experiments. H: add-back of CCN2 to immunodepleted WTCM [WTCM(-CCN2_IP+CCN2)] restores cell spreading of Pdia3-/- MEFs. Quantified analysis of cell areas of Pdia3-/- MEFs plated under the indicated conditions for 2 h. From n = 3 experiments. In D, G, and H, each datapoint represents one cell and red horizontal bars indicate the mean. Statistical analysis by ANOVA with Welch’s test. I: In vitro binding of TSP1 to surface-bound CCN2. Datapoints show means and error bars show standard deviation. J: quantitative analysis of fluorescent area per HDF cell by in situ proximity ligation with antibodies to CCN2 or TSP1 (C9) as indicated. Each datapoint is one cell and red horizontal bars indicate the mean. Analysis by Kruskal–Wallis test. From n = 3 experiments. CCN2, cell communication network 2; CM, conditioned medium; Con, control; KO, knockout; MEFs, mouse embryonic fibroblasts; WT, wild type; WTCM, conditioned medium from wild-type MEFs.

Fig 4: Correlations of identified PDIA3-dependent, heparin-binding, secreted proteins with ECM-related signatures and protein networks. A: top ten significant overlaps of gene signature sets of the Molecular Signatures database for the proteins at least twofold decreased (gene names in plain text) or increased (gene names in bolded text) in CM of Pdia3-/- MEFs relative to the CM of WT-MEFs, as identified by TMT proteomics. Details of the TMT proteomics data set are in Supplemental Table S1. Hhlp1 is not included in the figure because it is not mapped in the Molecular Signatures database. B: STRING functional protein interaction network based on the above proteins plus TSP1 and fibronectin. Asterisk marks CCN2 (identified by the alternative gene name ctgf in the network output). CCN2, cell communication network 2; CM, conditioned medium; ECM, extracellular matrix; MEFs, mouse embryonic fibroblasts; TMT, tandem mass tag; WT, wild type.

Fig 5: Conditioned media or extracellular matrix from WT-MEFs promotes adhesion of Pdia3-/- MEFs. A: representative immunofluorescence confocal images for F-actin and vinculin in Pdia3-/- MEFs (KO) fixed after 2 h adhesion in presence of either CM or isolated ECM from 48 h cultures of WT-MEFs (labeled as +WTCM or on WTECM, respectively) or equivalent samples from Pdia3-/- MEFs (labeled as KO + KOCM or KO on KOECM, respectively), presented as XYZ stacks. Scale bar = 20 µm (inset scale = 5 µm). B: quantified analysis of cell area from immunofluorescence images of Pdia3-/- MEFs under the same conditions as in A (n = 4 experiments). C: quantified analysis of focal adhesions per cell from immunofluorescence images for vinculin in Pdia3-/- MEFs under conditions as in A (n = 4 experiments). B and C: statistical analysis by Kruskal-Wallis test. D and E: inhibition of PDIA3 in WT-MEFs reduces the adhesion-promoting activity of WTCM on Pdia3-/- MEFs. D: representative confocal immunofluorescence images for F-actin and vinculin in Pdia3-/- MEFs after 2 h adhesion with CM from WT-MEFs grown for 48 h in the presence (+16 F) or absence of 1 µM 16F16, presented as XYZ stacks. Scale bar = 20 µm. E: quantified analysis of cell areas of Pdia3-/- MEFs from immunofluorescence images as in D and n = 3 experiments. F and G: inhibition of PDIA3 in WT-MEFs reduces the adhesion of WT-MEFs. F: representative confocal immunofluorescence images for F-actin and vinculin in WT-MEFs plated in the presence (WT + 16 F) or absence of 1 µM 16F16 for 2 h with CM from WT-MEFs grown for 48 h either in the presence (WTCM +16F16) or absence (WTCM) of 1 µM 16F16, presented as XYZ stacks. Scale bar = 20 µm. G: quantified analysis of cell area from immunofluorescence images as in F and n = 3 experiments. E and G: statistical analysis by unpaired t test with Welch’s test. In all graphs, red horizontal bars show the mean. CM, conditioned medium; ECM, extracellular matrix; KO, knockout; MEFs, mouse embryonic fibroblasts; WT, wild type; WTCM, conditioned medium from wild-type mouse embryonic fibroblasts; WTECM, extracellular matrix from wild-type mouse embryonic fibroblasts.