Phosphorylation of PLCγ1 by EphA2 Receptor Tyrosine Kinase Promotes Tumor Growth in Lung Cancer

This article is featured in Highlights of This Issue, p. 1613

Receptor tyrosine kinases (RTK) regulate signal transduction pathways that control cell proliferation, survival, and motility. Dysregulation of RTKs by mutations, amplifications, or overexpression can lead to oncogenic transformation and malignant progression (1). A number of RTKs have been identified as potential drivers of non–small cell lung cancer (NSCLC), one of which is EphA2 (2). The EphA2 RTK belongs to the EPH family, the largest family of RTKs, and is commonly overexpressed in NSCLC and associated with poor clinical outcomes (3). Targeted disruption of EphA2 impairs tumor growth in KRAS-mutant mouse models and in human NSCLC xenografts (4). Furthermore, EphA2 is overexpressed in EGFR tyrosine kinase inhibitor–resistant tumor cells (5). Loss of EphA2 reduced viability of erlotinib-resistant tumor cells harboring EGFR^T790M mutations in vitro and inhibited tumor growth in an inducible EGFR^L858R+T790M-mutant lung cancer model in vivo (5). Several EphA2 inhibitors including an antibody, a peptide, and a small-molecule inhibitor have been developed (6). An EphA2-targeting DOPC-encapsulated siRNA is currently in phase I clinical trials for advanced or recurrent solid tumors (NCT01591356). However, despite the interest in EphA2 as a therapeutic target, molecular mechanisms mediating EphA2 function, particularly its proximal downstream signals, are not well characterized.

Phospholipase C gamma (PLCγ) is a lipase activated by receptors in the cellular membrane, including RTKs and adhesion receptors. Once activated, PLCγ hydrolyzes phosphatidylinositol 4,5-bisphosphate to form diacylglycerol and inositol 1,4,5-trisphosphate, the latter promoting the transient release of intracellular Ca²⁺, another important signaling molecule. PLCγ is ubiquitously expressed and exists in two isoforms, PLCG1 and PLCG2, each with distinct functions in a variety of cell types and disease states (7, 8). PLCG1 plays a role in vasculogenesis and erythrogenesis as well as T-cell development and activity (9). Importantly, loss of PLCG1 is embryonic lethal in mice (10). PLCG2, meanwhile, is critical for B-cell development and maturation (8, 11). Both PLCγ isoforms are enriched and mutated in many cancers (8). Elevated PLCγ1 has been shown to drive metastasis and progression of breast cancer (12, 13), and its phosphorylation status is prognostic for metastatic risk (14). PLCγ has also been implicated in resistance to cancer treatment. In glioblastoma, PLCγ/HIF1α mediated FGFR1-induced radioresistance (15) while in head and neck and esophageal squamous cell carcinoma, the AXL-EGFR-PLCγ1 axis mediated resistance to PI3K inhibition (16). An acquired PLCG2 mutation also caused resistance to ibrutinib in chronic lymphocytic leukemia (17). While important roles for PLCγ have been identified in several cancer types, PLCγ role in lung cancer has yet to be elucidated.

In this report, we show that PLCγ is a novel target of the EphA2 RTK in lung cancer. We show that EphA2 interacts with and directly phosphorylates PLCγ for activation. In addition, knockdown of PLCG1 significantly reduces the growth of KRAS-mutant lung cancer cells in vitro and inhibits lung tumor growth in an orthotopic Kras-p53-Lkb1–mutant mouse model in vivo. Collectively, these studies identify the EphA2-PLCγ1 axis as a potential therapeutic target for KRAS-mutant lung cancer.

293FT, COS-7, and mouse Kras, p53, and Lkb1 (KPL) lines were cultured in DMEM supplemented with penicillin/streptomycin and10% FBS. Human lung cancer cell lines (A549, H23, H358, H2030, H2009, and HCC44) and BEAS2B cells were cultured in RPMI1640 supplemented with penicillin/streptomycin and 10% FBS. All cell lines were purchased from ATCC, except the murine KPL line which was generated in our lab as shown in Fig. 5. Cell lines were used between passages 1 and 50 after thaw and authenticated using short tandem repeat profiling at ATCC, most recently in June 2019. Mycoplasma was routinely tested approximately every 6 months to exclude possible contamination, most recently in November 2019, using the PlasmoTest Kit from InvivoGen.

For transient knockdown, siRNAs were purchased from Dharmacon (smart pool siEphA2: catalog no. L-003116-00-0005; nontargeting pool: catalog no. D-001810-10-05; individual siPLCG1 #1–3: catalog no. J-003559-05, 07, 08). For stable short hairpin (shRNA) knockdown, lentiviral vector pLKO.1 was used (EphA2 shRNA #1 CGGACAGACATATGGGATATT; EphA2 shRNA#2 GCGTATCTTCATTGAGCTCAA; PLCG1 shRNA #1 ATGACAAAGCAATGTGACTGG; PLCG1 shRNA #2 ATGTAAACTTTGTTTCCCTGG; PLCG1 shRNA #3 AATTTCACGAATGTCAATGGC; PLCG1 shRNA #4 ATACCATTCGTGGTTCACAGG; GFP shRNA control GCAAGCTGACCCTGAAGTTCAT). For CRISPR/Cas9-mediated gene knockout, lentiviral vector LentiCRISPR v2 was used human PLCG1 gRNA #1 ATAGCGATCAAAGTCCCGTG; human PLCG1 gRNA #2 GTTCACTTCATCCTCAGATG; LacZ gRNA TGCGAATACGCCCACGCGAT; mouse PLCG1 gRNA #1 GCTAATGGAGGATACACTGC; mouse PLCG1 gRNA #2 CCGCGGCGCGGACAAAATCG). The PLCG1 full-length cDNA plasmid was purchased from Sino Biological Inc (catalog no. MG50804-G) and PLCG1 was subcloned into pCDH-puro vector with Flag-tag at its C-terminus. The EphA2 full-length cDNA plasmid (pCDH-puro EphA2-Myc) and its corresponding mutants (S897A, Y588F, Y594F, Y735F, Y930F, K646M, and D739N) were all from lab stocks. For adeno-associated virus (AAV) system, AAV9 and pΔF6 plasmids were purchased form Penn Vector Core at the University of Pennsylvania (Philadelphia, PA), and AAV-KPL plasmid was purchased from Addgene (catalog no. 60224). ALW-II-41-27 was purchased from MedChem Express.

To construct a comprehensive and high-quality human protein–protein interaction (PPI) network, we assembled 15 commonly used data sources with five types of experimental evidence: (i) binary PPIs tested by high-throughput yeast-two-hybrid (Y2H) systems; (ii) binary, physical PPIs from protein three-dimensional structures; (iii) kinase-substrate interactions from literature-derived low-throughput and high-throughput experiments; (iv) signaling networks derived from low-throughput experiments; (v) literature-derived identified by affinity purification followed by mass spectrometry and low-throughput experiments. In total, the updated human interactome consisted of 351,444 PPIs (edges or links) linked to 17,706 unique proteins (nodes). The detailed descriptions of building the human interactome are given in our recent studies (18–20). We then mapped the EPHA2, PLCG1, and PLCG2 into the PPIs network to construct EPHA2-PLCG1/PLCG2 subnetwork. Next, we performed Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analysis to identify the functional pathway related with PLCG1 and PLCG2.

MTT assay was used to evaluate the short-term proliferation of cells. A total of 2 × 10³ cells were plated into 96-well plates with six replicates in growth media. MTT reagent was added and the plates were read using plate reader (Synergy HT, BioTek) on days 1 to 6. Cell viability was normalized to day 1. Colony formation assays were used to evaluate the long-term proliferation of cells. A total of 400 cells were plated into 12-well plates with 3 to 4 replicates in growth media. For drug treatment experiments, drugs were added the day after cell attachment with an initial plating of 2 × 10⁴ cells. Cell colonies were visualized by crystal violet staining after 2 weeks for human cells or 1 week for mouse KPL cells.

Y2H screening was carried out by Hybrigenics Services (hybrigenics-services.com). The cytoplasmic EphA2 tail (AA 559-976) was cloned as a N-LexA-EPHA2-C fusion to be the bait against a lung cancer cDNA library (mix of A549, H1703, and H460) and 79 positive clones were selected on DO-3–selective medium plates. A confidence score (predicted biological score, PBS) was assigned to each interaction, then scores were stratified into categories based on the degree of confidence.

Cancer cells in culture medium were plated onto coverslips coated with 0.5% gelatin in DPBS. Cells were washed with PBS and fixed with 4% paraformaldehyde after 24 hours growth. 5% goat serum plus 0.3% triton X-100 in DPBS was used to permeabilize cells. Anti-PLCG1 rabbit polyclonal antibody (Santa Cruz Biotechnology, catalog no. SC-81, 1:200) and anti-EphA2 mouse mAb (EMD Millipore, catalog no. 05-480, 1:400) diluted in blocking buffer were applied to cells and incubated for overnight at 4°C. The DuoLink Proximity Ligation Kit (Sigma-Aldrich, catalog no. DUO92102) was used according to manufacturer's instructions.

For Western blotting, 10 to 30 μg of total protein from cell lysates were separated by SDS-PAGE, transferred to a nitrocellulose membrane and probed with indicated antibodies. Primary antibodies used in this study were as follows: rabbit anti-EphA2 (Santa Cruz Biotechnology, catalog no. SC-924, 1:1,000), mouse anti-EphA2 (EMD Millipore, catalog no. 05-480, 1:1,000), mouse anti-PLCG1 (Santa Cruz Biotechnology, catalog no. SC-7290, 1:500), rabbit anti-PLCG2 (Santa Cruz Biotechnology, catalog no. SC-407, 1:500), rabbit anti-EphA2 Y588 (Cell Signaling Technology, catalog no. 12677, 1:500), rabbit anti-PLCG1 Y783 (Cell Signaling Technology, catalog no.14008, 1:500), mouse anti-β-actin (Santa Cruz Biotechnology, catalog no. SC-47778, 1:1,000). Secondary antibodies used were as follows: anti-rabbit IgG HRP (Promega, catalog no. W4011, 1:5,000), anti-mouse IgG HRP (Promega, catalog no. W4021, 1:5,000), anti-rabbit IgG IRDye 800CW (LI-COR, catalog no. 926-32211), and anti-mouse IgG IRDye 680LT (LI-COR, catalog no. 926-68020). Antibodies were diluted in PBST/5% nonfat milk. Signal was detected using enhanced chemiluminescence substrate (West Femto or West Pico, Thermo Fisher Scientific) or by LI-COR Odyssey Infrared Imaging System.

For immunoprecipitation (IP), cells were lysed in IP buffer (10 mmol/L Tris-HCl pH = 7.5, 150 mmol/L NaCl, 2 mmol/L EDTA, 0.5-1% Triton X-100). A total of 1 μg of total protein was incubated with anti-Myc tag or anti-Flag agarose beads or indicated antibody then Protein G Dynabeads (Invitrogen) overnight at 4°C. Beads were washed with lysis buffer and boiled with 20 μL SDS loading buffer. The soluble fraction was loaded for immunoblot analysis.

Lung tumor sections were stained with hematoxylin and eosin by Vanderbilt University Translational Pathology Shared Resource. Proliferating cell nuclear antigen (PCNA) staining (Cell Signaling Technology, rabbit anti PCNA, catalog no. 13110) for proliferation or terminal deoxynucleotidyl transferase–mediated dUTP nick end labeling (TUNEL) assay (Millipore, ApopTag Red In Situ Apoptosis Detection Kit, catalog no. S7165) for apoptosis was performed as described previously (21, 22). Tumor area of 20× images was analyzed by ImageJ or CellSens software.

Formalin-fixed, paraffin-embedded human lung adenocarcinoma tissue arrays (catalog no. LC641) were purchased from Biomax. Following rehydration, antigen retrieval was accomplished using Retrievagen A (BD Biosciences) as per manufacturer's instructions. Following blockade of endogenous peroxidases using 3% H₂O₂, tissues were first permeabilized for 5 minutes with 0.3% Triton X-100 in PBS and then washed in PBS. Tissues were blocked with 2.5% goat serum and then probed with antibodies against EphA2 (10 μg/mL; catalog no. 34-7400, Zymed Laboratories) or phospho-PLCγ1 (Tyr783; 1:50; catalog no. 14008S, Cell Signaling Technology) overnight at 4°C. Samples were incubated with biotinylated anti-rabbit IgG (1:250; catalog no. 550338, BD Biosciences) and subsequently incubated with horseradish peroxidase (HRP) streptavidin (catalog no. SA-5704, Vector Laboratories), both for 1 hour at room temperature, with thorough washing before and after each step. Staining was performed using liquid diaminobenzidine (catalog no. SK-4103, Vector Laboratories) and hematoxylin (catalog no. H-3401, Vector Laboratories). Following dehydration, stained tissues were mounted using Cytoseal XYL. Slides were imaged on a Leica SCN400 Slide Scanner at 40× magnification at a resolution of 0.25 μm/pixel. The diaminobenzidine stained area and total tissue area were determined from deconvoluted high-resolution images of each tissue core in the Leica Digital Image Hub. The percentage of tissue area positive for diaminobenzidine staining is shown.

Genomic DNA was extracted using QuickExtract DNA Extraction Solution (Lucigen) according to the manufacturer's protocol. Target regions were amplified with indicated primers (Supplementary Fig. S2) using the Q5 polymerase (NEB) and T7E1 mismatch detection assay was performed and analyzed according to previously published protocols (23).

AAV was produced in 293FT cells, chemically purified and concentrated according to published methods (24). AAV titer was determined using AAVpro Titration Kita (Takara Bio Inc., catalog no. 6233). A titer of 2 × 10¹¹ viral genome copies in 75 μL DPBS was intratracheally injected into 7- to 8-week-old Rosa26-LSL-Cas9-EGFP mice (stock no.: 024858, The Jackson Laboratory). Tumor development was monitored weekly by MRI and by GFP imaging after sacrifice. For orthotopic lung tumor growth, KPL lines were transfected with luciferase expression plasmid and intravenously injected via tail vein. Lung tumor growth was measured once a week by bioluminescence imaging.

All animal experiments were preapproved by the Vanderbilt Institutional Animal Care and Use Committee and followed all state and federal rules and regulations.

Data were presented as mean ± SD or SEM and statistically analyzed by two-tailed Student t test or two-way ANOVA with Tukey post hoc correction. All experiments were performed at least two independent times and P < 0.05 was treated as statistically significant. GraphPad Prism 8 was used for statistical analysis.

To identify EphA2-interacting partners in human lung cancer cells, the intracellular domain (AA 559-976) of EphA2 was used as bait for a Y2H screen against the cDNA library of lung cancer cells (www.hybrigenics-services.com). A total of 79 positive clones were screened, resulting in 15 total candidates. The candidates were stratified into three categories based on confidence of interaction: very high confidence (3 proteins, orange), high confidence (4 proteins, blue), and moderate confidence (8 proteins, green; Fig. 1A). One to 2 candidates were identified in each category that have been previously reported to interact with EphA2, validating the efficacy and accuracy of the Y2H screen (25, 26). We chose to explore PLCG1 and PLCG2 further because both are PLCγ family members and novel interactors with EphA2 identified in this screen with very high confidence and high confidence, respectively.

To verify the interaction between EphA2 and PLCγ (PLCγ1 and PLCγ2) identified in the Y2H screen, EphA2 and PLCG1 or PLCG2 were expressed in COS-7 cells. Both PLCγ isoforms were immunoprecipitated with EphA2; however, in the reverse direction, only IP of PLCγ1 was able to coimmunoprecipitate EphA2 (Fig. 1B), suggesting a weaker interaction between EphA2 and PLCγ2 compared with PLCγ1.

PLCγ1 and PLCγ2 are important mediators of intracellular signal transduction and are activated by phosphorylation by upstream kinases (27). To determine whether EphA2 is required for PLCγ activation, we assessed phosphorylation by immunoblotting. Expression of EphA2 increased the overall phosphorylation levels of both PLCγ1 and PLCγ2 including two different tyrosine sites on each protein (Y783 and Y1253 for PLCγ1; Y759 and Y1217 for PLCγ2; Fig. 1C). We also observed enhanced phosphorylation of Y783 on endogenous PLCγ1 in EphA2-overexpressing samples. Conversely, overexpression of PLCγ had no effect on the phosphorylation of EphA2 (Fig. 1C), suggesting signaling proceeds from EphA2 to PLCγ.

Additional analysis of the PLCγ protein–protein interactome identified 57 interacting proteins for PLCγ1, far more than the 14 interacting proteins of PLCγ2 (Fig. 1D). Interestingly, KEGG enrichment analysis of the PLCγ1 interactome identified 22 proteins that were significantly enriched in the Ras signaling pathway (P_adjusted = 2.1e-22; Fig. 1E), a pathway known to be important in driving a significant portion of NSCLC.

Because PLCγ1 had a stronger interaction with EphA2 (Fig. 1B) and PPI analysis of PLCG1 showed a broader functional network than PLCγ2, we focused on PLCγ1′s interaction with EphA2 in KRAS-mutant lung cancer for the remainder of the study.

PLCγ is a well-known downstream effector of several RTKs, including EGFR, platlet-derived growth factor receptor, VEGFR, and TrkB (9). Phospho-tyrosine sites on the RTKs interact with the SH2 domain of PLCγ for activation (9). Consistent with other RTKs, our results from the Y2H screen indicated that the SH2 domain of PLCγ1 was involved in the interaction with the EphA2 intracellular domain. To characterize the interaction between EphA2 and PLCγ1, a series of EphA2 mutants were made, including two kinase dead mutants (K646M, D739N), four tyrosine phosphorylation site mutants (Y588F, Y594F, Y735F, Y930F), and one serine phosphorylation site mutant (S897A; Fig. 2A). These phosphorylation sites were chosen on the basis of the literature and analysis of EphA2 in the PhosphoSitePlus PTM Resource (https://www.phosphosite.org/homeAction.action) as putative sites most likely to be involved in EphA2 kinase activity or its cellular function.

In EphA2 and PLCG1 coexpressing COS-7 cells (Fig. 2B and C), wild-type EphA2 increased the level of phospho-PLCγ1, whereas the ability to phosphorylate PLCγ1 was impaired in cells expressing kinase dead or the two tyrosine mutants (Y588F and Y594F) that are situated within the juxtamembrane domain of EphA2 and are known to be required for proper kinase activity (28). Mutation of the other two tyrosine sites in the kinase domain and SAM domain as well as the serine site had little effect on PLCγ1 phosphorylation. Furthermore, the level of phospho-PLCγ1 was dependent on the level of EphA2 expression in the cells (Fig. 2D). Consistent with Fig. 2C, expression of the EphA2 kinase dead mutant (K646M) had no effect on phosphorylation of PLCγ1 even at very high doses. The Y588F mutant, which exhibited reduced kinase activity, only weakly affected phosphorylation of PLCγ1 at high expression levels.

We also expressed EphA2 in BEAS2B cells, a immortalized bronchial epithelial cell line with low endogenous EphA2 expression and analyzed the phosphorylation levels of endogenous PLCγ1 (Fig. 2E). The result here showed that only wild-type EphA2, but not K646M or Y588F mutants, could phosphorylate PLCγ1, demonstrating that PLCγ1 phosphorylation is dependent on the kinase activity of EphA2.

EphA2 is highly expressed in many KRAS-mutant lung cancer cells and has been shown to regulate tumor malignancy (4), prompting us to investigate if PLCγ1 is regulated by EphA2 in these cells. Expression of wild-type EphA2 in H23 cells led to phosphorylation of PLCγ1, while kinase-dead EphA2 had no effect on p-PLCγ1 levels (Fig. 3A). Corresponding experiments knocking down EphA2 by siRNA in a panel of KRAS-mutant lung cancer cell lines (H23, H2009, A549, HCC44, H2030, H358) led to a decrease of p-PLCγ1 (Fig. 3B and C). Furthermore, loss of EphA2 by shRNA or siRNA in either H23 or H2009 cells diminished p-PLCγ1 in serum-starved, stimulated, or normal growth conditions (Fig. 3B and D). Stimulation with ephrin-A1-Fc, a surrogate for EphA2 canonical ligand ephrin-A1, also increased p-PLCγ1 in H23 and H2009 cells (Fig. 3E). This signal was depleted when EphA2 was knocked down by shRNA (Fig. 3F). Pharmacologic inhibition of EphA2 kinase activity using the small-molecule inhibitor ALW-II-41-27(4; ALW) also inhibited phosphorylation of PLCγ1, with increasing doses of ALW inhibiting both p-EphA2 and p-PLCγ1 in H23 and H2009 cells (Fig. 3G; Supplementary Fig. S1A). IHC analysis of EphA2 and p-PLCG1 in a human lung adenocarcinoma tumor microarray showed a positive correlation between EphA2 and p-PLCγ1 staining (Supplementary Fig. S1B and S1C), supporting the relevance of the EphA2-PLCG1 axis in human lung cancer.

To verify that EphA2 phosphorylates PLCγ1 via a direct interaction between the two proteins in lung cancer cells, we performed DuoLink proximity ligation assay (PLA). We showed that loss of either EphA2 or PLCγ1 by shRNA reduced the number of EphA2-PLCγ1 interactions compared with control shGFP (Fig. 3H and I). In addition, EphA2 coimmunoprecipitated with PLCγ1 in H23 shGFP control cells, but not in shEphA2 cells, providing evidence of an endogenous EphA2-PLCγ1 interaction (Fig. 3J). Together, these data show that EphA2 interacts with and phosphorylates PLCγ1 in human KRAS-mutant lung cancer cell lines.

Despite PLCγ1′s well-known roles in regulating T-cell development and homeostasis (29) and breast cancer cell migration and invasion (30), a role for PLCγ1 in lung cancer remains unclear. We used three independent siRNAs to knockdown PLCG1 in H23 and H2009 cells (Fig. 4A). In both cell lines, transient loss of PLCγ1 significantly reduced cell viability compared with control (Fig. 4B). Long-term effects of PLCγ1 loss on cell proliferation were assessed by stable knockdown or knockout of PLCG1 by four independent shRNAs or CRISPR-Cas9–mediated genome editing followed by MTT and colony formation assays (Fig. 4C–E; Supplementary Fig. S2). shPLCG1 cells showed a much slower growth rate and rarely formed colonies even after 2 weeks in culture compared with shGFP control. Similarly, PLCγ1 loss significantly hindered colony formation in sgPLCG1 cells compared with sgLacZ control, with the reduction in colony formation correlating with the efficiency of knockout (Fig. 4F and G). Thus, we demonstrate that PLCG1 promotes the growth of human lung cancer cells in vitro.

To evaluate the in vivo role of PLCγ1 in tumor growth within a competent immune environment, a mouse KPL lung tumor model was established (Fig. 5A–C) based on the report from Platt and colleagues (24). An adeno-associated virus (AAV) carrying three adjacent sgRNAs targeting Kras, p53, and Lkb1 genes (KPL), together with a Cre expression cassette and a mutant Kras^G12D genomic template, was purified and delivered into the lungs of Rosa26-LSL-Cas9-EGFP knock-in recipient mice via intratracheal instillation (Fig. 5A). Cre-mediated recombination allows for EGFP and Cas9 expression in target cells, which leads to mutation of the three target genes via nonhomologous end joining (p53 or Lkb1) or homology-directed repair (Kras G12D). The mice developed abundant lung nodules approximately 2–3 months after viral instillation, as monitored by MRI (Fig. 5B). Tumor nodules were also visible by EGFP after lung dissection (Fig. 5C). EGFP-positive tumor cell populations were then isolated from the tumor mass and single cell clones were established (Fig. 5C). Western blots were used to confirm KPL mutations in the clones (Fig. 5D), and clone KPL-C2 was selected for use in the following experiments.

In addition to the targeted KPL mutations, KPL-C2 cells had high EphA2 and p-EphA2 expression. Pharmacologic inhibition of EphA2 by ALW effectively blocked both p-EphA2 and p-PLCγ1 (Fig. 5E, arrow) and colony growth of KPL cells in a dose-dependent manner (Fig. 5F), indicating EphA2 might modulate PLCγ1 activity to regulate KPL tumor cell growth. CRISPR-Cas9–mediated genome editing was used to generate PLCγ1 knockout KPL cells (Fig. 5G). PLCγ1-deficient KPL cells showed a significant decrease in colony growth in vitro (Fig. 5H). sgLacZ control and sgPLCG1 knockout cells were subsequently injected into the tail veins of immune competent mice (Rosa26-LSL-Cas9-EGFP). sgLacZ cells developed a significant number of tumors in the lungs at week 3, while PLCG1-deficient cells formed a very limited number of tumors compared with the control cells (Fig. 5I and J). In some mice, loss of PLCγ1 completely inhibited tumor formation. PCNA and cleaved caspase-3 IHC staining of the tumors showed that PLCγ1 deficiency in these KPL cells inhibited tumor cell proliferation but had little effect on apoptosis (Fig. 5K–M), respectively, in agreement with our previous findings that EphA2 knockdown affected tumor cell proliferation but not apoptosis (31, 32). Collectively, these data show that PLCγ1 promotes KPL lung tumor growth.

EphA2 has emerged as a promising target in several tumor types including lung, pancreatic, breast, and glioblastoma (4, 31–35), yet understanding of its proximal downstream signals is not comprehensive. In this report, we identify PLCγ as a novel downstream effector of EphA2 capable of promoting tumor progression in the context of KRAS-mutant lung cancer. In addition to PLCγ1, several other well-known signaling molecules, including S6K1-pBAD, JNK-c-JUN, mTOR, and ERK, are also known to function downstream of EphA2 in lung cancer (4, 31, 36, 37). How cells precisely regulate timing and localization of these interactions downstream of EphA2 remains unanswered. Along with PLCγ, our Y2H screen also identified other EphA2 interactors including Src family proteins, the PTPN3 phosphatase, and PIK3R1, a regulatory subunit of PI3K (Fig. 1A). While many of these hits have been implicated in lung tumorigenesis (38–42), additional studies are required to determine whether these interactors participate in EphA2 signaling during lung tumor progression.

The role of PLCγ1 in tumor cell proliferation is controversial. While PLCγ1 has been implicated in directing cell-cycle progression (43, 44), other reports suggest that PLCγ1 may negatively regulate cell proliferation (45). These conflicting reports suggest that PLCγ1 regulation of cell proliferation may be context dependent, perhaps varying based on the tumor type or activating growth factor. In this study, we show that loss of PLCG1 reduces cell viability of KRAS-mutant lung cancer cell lines in vitro and reduced PCNA staining of KPL lung tumors in vivo. Thus, our data suggest in the context of KRAS-mutant lung cancer, PLCγ1 facilitates tumor cell proliferation. Furthermore, PLCγ has been implicated in AXL-mediated resistance to PI3K inhibition in neck and esophageal squamous cell carcinomas (16). Because EphA2 also plays critical roles in tumor resistance to EGFR kinase inhibitors in lung cancer (5) and B-Raf inhibitors in melanoma (46), it will be interesting to investigate whether PLCγ mediates the EphA2 signaling pathway in drug-resistant cells.

In summary, our data reveal that PLCγ1 is a novel interactor of the EphA2 RTK in lung cancer cells, and our data support the idea that targeting this EphA2- PLCγ1 signaling axis could be a promising therapeutic option for treating lung cancer.

L.C. Kim reports grants from NCI during the conduct of the study. D.M. Brantley-Sieders reports grants from NIH/NCI during the conduct of the study. J. Chen reports grants from NCI and VA during the conduct of the study. No potential conflicts of interest were disclosed by the other authors.

W. Song: Conceptualization, data curation, formal analysis, investigation, visualization, methodology, writing-original draft, writing-review and editing. L.C. Kim: Data curation, formal analysis, investigation, writing-original draft, writing-review and editing. W. Han: Data curation, formal analysis. Y. Hou: Data curation, formal analysis. D.N. Edwards: Data curation, formal analysis. S. Wang: Data curation. T.S. Blackwell: Resources. F. Cheng: Data curation, formal analysis, supervision. D.M. Brantley-Sieders: Resources, data curation, supervision. J. Chen: Conceptualization, resources, formal analysis, supervision, project administration, writing-review and editing.

We would like to acknowledge the Vanderbilt Small Animal Image Core for assistance with live animal imaging and the Vanderbilt Translational Pathology Shared Resource for their help with immunohistochemistry staining of tumor sections. This work was supported by a VA Merit Award 5101BX000134 and a VA Research Career Scientist Award (to J. Chen), NIH grants R01 CA177681 (to J. Chen and D.M. Brantley-Sieders), R01 CA95004 (to J. Chen), T32 CA009592 (to L.C. Kim), and F31 CA2220804-01 (to L.C. Kim).

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.