Fig 2: Model for the anticancer actions of DDAs(A) DDAs function to suppress tumor cell division and survival through mechanisms involving suppression of Akt phosphorylation, downregulation of EGFR (HER1), HER2, and HER3, and activation of UPR. Overexpression of EGFR or HER2 potentiates each of these mechanisms. (B) At normal expression levels, EGFR and HER2 are folded efficiently and do not induce ER stress (upper panel). Overexpression of EGFR or HER2 cooperates with DDA treatment, which blocks disulfide bond formation and protein folding, to activate UPR (lower panel). Red asterisks denote ER stress markers examined experimentally in Figures 1-6.

Fig 3: Identification of variants and pathway enrichment in PDX models. a We detected 1325 variations, including 581 nonsynonymous single-nucleotide variations, 225 CNVs, 513 indels, and 6 translocation fusions. b Genomic alterations were analyzed with different clinicopathological features. Navy, non-synonymous SNV; dark red, CNVs; yellow-green, frameshift indel; purple; inframeshift indel; and pink, fusion. c Genomic landscape analysis of genes altered in the PDX and TCGA datasets. Red, amplification; blue, deletion; and green, mutation. d Several relevant pathways, including the MAPK, ErbB, cell cycle, mTOR, and VEGF, were found to be enriched. e Details for molecular alterations involved in the ErbB and cell cycle pathway. Blue, amplification; green, mutation

Fig 4: CRISPR screen-identified gene knockouts specifically toxic to D492 or D492M cells. (A) Schematic of the CRISPR loss-of-function screening strategy. (B, C) Heat maps of the knockouts significantly more toxic to D492 compared with D492M (B) or D492M compared with D492 (C) with the significance level logP = -3 in either two- or four-week samples, as calculated using RSA. Color intensity indicates logP values. Red dots indicate selective hits that were also found using the orthogonal CRISPR analysis tool MAGeCK-MLE. (D, E) STRING interaction network of the toxic knockouts specific for D492 (D; red and blue indicate proteins involved in ERBB and GnRH signaling, respectively) or D492M (E; red and blue indicate proteins involved in cell cycle/mitosis and RNA metabolism/spliceosome, respectively); the panels below indicate KEGG/Reactome pathways enriched in the D492- or D492M-specific hit lists; FDR, false discovery rate.

Fig 5: Pro-inflammatory activation of macrophages in vitro induces ErbB4 expression. ErbB family member RNA expression was determined after 6 h stimulation of (a) BMDMs classically activated with IFN?/LPS, (b) BMDMs alternatively activated with IL-4, (c) neutrophils activated with LPS, (d) neutrophils primed with GM-CSF and then stimulated with LPS, (e) BMDCs activated with IFN?/LPS. For (a–e), n=4–6 independent experiments per group. ND, not detectable. Error bars represent S.E.M. *P<0.05; **P<0.01; ***P<0.001. (f and g) Unstimulated (naive) or classically activated (M1; 24 h activation shown) BMDMs were subjected to (f) immunofluorescence staining (representative of three independent experiments) for ErbB4 (red), nuclei (blue) with or without blocking peptide to primary antibody, or (g) western blot analysis (representative blots from n=4 independent experiments) for EGFR, ErbB2, ErbB3, ErbB4 (full length, FL; intracellular domain, 4ICD)