Fig 1: Co-suppression of EPHA2 and EGFR enhances elimination of cancer cells. A, MDA-MB-231 were transduced with the CAS9-encoding pLv5 lentiviral vector (MDA-Cas9), subjected to selection with blasticidin, and CAS9 expression was assessed by Western blot (left). MDA-Cas9 cells were transduced with LV04 lentiviral vectors from Sanger human CRISPR library encoding EphA2-targeting sgRNAs, ID 2400992 and ID 2400602 (MDA-EphA2-KO) or control, non-targeting sgRNA Lenti CRISPR Universal Non-Target Control #2 Plasmid (LV04 vector; MDA-NTC) and subjected to puromycin selection. EPHA2 knockout was confirmed by Western blot (right). B, MDA-EphA2-KO and control cells were seeded at 1.6×103 cells per well (three wells per condition) in the MamoCult Basal Medium (Stemcell Technologies, Cat # 05621) into ultralow attachment 24-well plates and allowed to form tumorspheres for 8 days in the presence of 10 μmol/L of erlotinib or a matching concentration of DMSO. Tumorspheres in each well were trypsinized and individual cells counted. The graph represents abundance of the erlotinib-treated cells as percentages relative to matching DMSO controls. C, The indicated cells were seeded and allowed to form tumorspheres as in (B). Images of tumorspheres in each well were taken using an EVOS m5000 imager, and overall tumorsphere area was summarized using ImageJ software. The graph represents tumorsphere area of erlotinib-treated cells as a percentage of relative to the area in a matching DMSO control. D, Representative images of erlotinib-treated and DMSO-treated tumorspheres formed by MDA-EphA2-KO and MDA-NTC cells; scale bar, 1,000 μm. E, The indicated cells were seeded in DMEM medium with 1% FBS at 4.5×103 cells per well (five wells per condition) into 96-well plates to form monolayer cultures. Cells were treated for 72 hours with the indicated concentrations of erlotinib or DMSO concentration matching DMSO volume loaded with the highest erlotinib dose. Cell survival was quantified using the resazurin assay (R&D Systems, Cat# AR002). The graph represents survival of erlotinib-treated cells as percentages relative to matching DMSO controls. F, CAS9 was expressed in MIA PaCa-2 cells (MiaPaCa-Cas9) and EPHA2 knocked out (MiaPaCa-EphA2-KO) as in (A). Nontargeting sgRNA was used as a control (MiaPaCa-NTC), and CAS9 expression and EPHA2 knockout were confirmed by Western blot. G, The effect of erlotinib on tumorspheres formed by MiaPaCa-EphA2-KO and MiaPaCa-NTC cells was examined as in (C); cell seeding was done at 1×103 cells/well into ultralow attachment 24-well plates. Tumorsphere cell counting (as in B) was not performed, as we could not get single-cell suspensions from these tumorspheres without damaging cells. H, Representative images of tumorspheres formed by the indicated cells in the presence of 10 μmol/L of erlotinib or matching DMSO control; scale bar, 1,000 μm. I, MiaPaCa-EphA2-KO and MiaPaCa-NTC cells were treated with erlotinib or DMSO in monolayer cultures, and cell survival was analyzed as in (E). J, Analysis of the AUC of PDX models representing non–small cell lung carcinoma and colorectal cancer. PDX models treated or not with cetuximab were classified on the basis of EPHA2 expression levels (top and low quartiles). The analysis reveals significantly stronger reduction in tumor size in response to cetuximab treatment in models with low EPHA2 levels compared with tumors with high EPHA2 expression. *, P < 0.05, Kolmogorov–Smirnov test. Data from monolayer (E and I) and tumorsphere (B, C, and G) experiments were analyzed using the Student t test, *, P < 0.01. Western blot images were optimized using PowerPoint software where required.