Epidermal growth factor receptor lacking C‐terminal autophosphorylation sites retains signal transduction and high sensitivity to epidermal growth factor receptor tyrosine kinase inhibitor

Abstract

Constitutively active mutations of epidermal growth factor receptor (EGFR) (delE746_A750) activate downstream signals, such as ERK and Akt, through the phosphorylation of tyrosine residues in the C‐terminal region of EGFR. These pathways are thought to be important for cellular sensitivity to EGFR tyrosine kinase inhibitors (TKI). To examine the correlation between phosphorylation of the tyrosine residues in the C‐terminal region of EGFR and cellular sensitivity to EGFR TKI, we used wild‐type (wt) EGFR, as well as the following constructs: delE746_A750 EGFR; delE746_A750 EGFR with substitution of seven tyrosine residues to phenylalanine in the C‐terminal region; and delE746_A750 EGFR with a C‐terminal truncation at amino acid 980. These constructs were transfected stably into HEK293 cells and designated HEK293/Wt, HEK293/D, HEK293/D7F, and HEK293/D‐Tr, respectively. The HEK293/D cells were found to be 100‐fold more sensitive to EGFR TKI (AG1478) than HEK293/Wt. Surprisingly, the HEK293/D7F and HEK293/D‐Tr cells, transfected with EGFR lacking the C‐terminal autophosphorylation sites, retained high sensitivity to EGFR TKI. In these three high‐sensitivity cells, the ERK pathway was activated without ligand stimulation, which was inhibited by EGFR TKI. In addition, although EGFR in the HEK293/D7F and HEK293/D‐Tr cells lacked significant tyrosine residues for EGFR signal transduction, phosphorylation of Src homology and collagen homology (Shc) was spontaneously activated in these cells. Our results indicate that tyrosine residues in the C‐terminal region of EGFR are not required for cellular sensitivity to EGFR TKI, and that an as‐yet‐unknown signaling pathway of EGFR may exist that is independent of the C‐terminal region of EGFR. (Cancer Sci 2009; 100: 552–557)

Epidermal growth factor receptor (EGFR), also termed HER1/ErbB‐1, is overexpressed and activated in many cancers.^{( 1} , ² , ^{3 )} Small‐molecule inhibitors of EGFR tyrosine kinase and antibodies have been shown to exhibit antitumor activity in several tumors.^{( 4} , ⁵ , ^{6 )} Somatic mutations of EGFR tyrosine kinase in non‐small cell lung cancer have been shown to be associated with hyperresponsiveness to gefitinib, a selective EGFR tyrosine kinase inhibitor (TKI).^{( 7} , ^{8 )} Many investigators have subsequently reported that EGFR mutations are strong determinants of the tumor response to EGFR TKI.^{( 9} , ^{10 )} Approximately 90% of non‐small cell lung cancer‐associated EGFR mutations in two reports consisted of two major EGFR mutations, namely, delE746_A750 in exon 19 and L858R in exon 21.^{( 11 )} We previously reported hypersensitivity to EGFR TKI of a PC‐9 cell line with delE746_A750 in exon 19, one of the commonly encountered mutations mentioned above, and this deletion mutant of EGFR was constitutively active and activated the ERK and Akt pathway.^{( 12} , ¹³ , ¹⁴ , ¹⁵ , ^{16 )} Binding of the receptor with its ligand leads to homodimerization and heterodimerization of the receptor tyrosine kinase.^{( 17} , ^{18 )} Thus, EGFR is a ligand‐activated tyrosine kinase that ultimately delivers cellular growth signals.

Tyr‐1068, Tyr‐1148, and Tyr‐1173 in the C‐terminal region are the major autophosphorylation sites in human EGFR. These C‐terminal phosphorylation sites of EGFR interact with adaptor proteins.^{( 19} , ^{20 )} Phosphorylation of the C‐terminal autophosphorylation sites of EGFR, triggered by epidermal growth factor (EGF), in turn trigger an intracellular signal cascade involving proteins such as ERK, Akt, Janus kinase, and signal transducer and activator of transcription.^{( 15} , ²¹ , ^{22 )} Src homology and collagen homology (Shc) is a molecular adaptor protein that binds phosphorylated tyrosines within activated EGFR, and is itself phosphorylated on tyrosine residues upon stimulation of EGFR. The phosphorylated CH1 site of Shc then engages the binding site for the SH2 domain of growth factor receptor‐bound protein (Grb) 2. The SH3 domain of Grb2 directly interacts with the guanyl nucleotide exchange factor son of sevenless homolog (Sos).^{( 23} , ^{24 )} Sos catalyzes the conversion of GDP to GTP on Ras, resulting in Ras activation. Activated GTP‐Ras recruits Raf kinase to the plasma membrane, resulting in Raf activation and phosphorylation of its downstream target ERK kinase.^{( 25} , ^{26 )}

Phosphorylation of tyrosine residues at the C‐terminal region of EGFR is believed to be important in cell signaling triggered by wild‐type EGFR.^{( 27} , ^{28 )} However, the role of this region in an active mutant of EGFR (delE746_A750) has yet to be elucidated in detail. To clarify the biological functions of the tyrosine residues at the C‐terminal region of EGFR, we constructed several mutants with C‐terminal‐truncated or substitution of tyrosine residues to phenylalanine in the C‐terminal region. We showed that EGFR lacking C‐terminal autophosphorylation sites still generated signals, with retention of cellular hypersensitivity to EGFR TKI.

Materials and Methods

Expression constructs. The method used to obtain full‐length cDNA of wild‐type EGFR has been described previously.^{( 12 )} Wild‐type EGFR cDNA and 15 bp‐deletion EGFR (delE746_A750) were introduced into pcDNA3.1 (Invitrogen, Carlsbad, CA, USA) with a myc‐tag at its C‐terminus. The EGFR cDNA with substitution of seven tyrosine residues to phenylalanine in the C‐terminal region was amplified by mutagenesis; the QuikChangeα Site‐Directed Mutagenesis Kit (Stratagene, La Jolla, CA, USA) was used for the polymerase chain reaction and a primer set was synthesized (Supporting Information Table S1). The cDNA of the C‐terminal‐truncated EGFR with 15‐bp deletion (EGFR‐D‐Tr) was amplified using the following primer set: forward, CCT CCT CTT GCT GCT GGT GGT G; reverse, GAA CAAGCT TGA CAA GGT AGC GCT GGG GGT CTC. After the polymerase chain reaction products were cut with ClaI and HindIII, they were ligated to the ClaI and HindIII sites of the pcDNA3.1 expression vector containing EGFR‐D cDNA. The cDNA of wild‐type EGFR with the C‐terminal truncation at amino acid 980 (EGFR‐Wt‐Tr) was made from the ClaI and XhoI fragments of the pcDNA3.1 expression vector containing wild‐type EGFR and the ClaI and XhoI fragments of the pcDNA3.1 expression vector containing EGFR‐D‐Tr.

Epidermal growth factor receptor cDNA with the myc‐tag in pcDNA3.1 was cut and introduced into a pQCLIN retroviral vector (BD Biosciences Clontech, San Diego, CA, USA) together with enhanced green fluorescent protein (EGFP) followed by the internal ribosome entry sequence, to monitor the expression of the inserts indirectly. A pVSV‐G vector (Clontech, Palo Alto, CA, USA) for constitution of the viral envelope, pGP vector (Takara, Yotsukaichi, Japan), and the pQCXIX constructs were cotransfected into HEK293 cells using FuGENE6 transfection reagent (Roche Diagnostics, Basel, Switzerland). Briefly, 80% confluent cells cultured in a 10‐cm dish were transfected with 2 µg pVSV‐G vector plus 6 µg pQCXIX vector. Forty‐eight hours after the transfection, the culture medium was collected and the viral particles were concentrated by centrifugation at 15 000 g for 3 h at 4°C. The viral pellet was then resuspended in fresh Dulbecco's modified Eagle's medium (DMEM; Sigma, St Louis, MO, USA). The titer of the viral vector was calculated by counting the EGFP‐positive cells that were infected in serial dilutions of a virus‐containing medium and then determining the multiplicity of infection. HER2 and HER3 introduced retrovirally into HEK293 cells were used as positive controls in western blotting.

Cell culture and transfection.  The human embryonic kidney HEK293 cell line was obtained from the American Type Culture Collection (Manassas, VA, USA) and cultured in DMEM supplemented with 10% fetal bovine serum, penicillin, and streptomycin (Sigma) in a humidified atmosphere of 5% CO₂ at 37°C. The HEK293 cells were transfected with the viral vectors.

In vitro growth‐inhibition assay.  The growth‐inhibitory effects of AG1478 (Biomol International, Plymouth Meeting, PA, USA) in HEK293/Wt, HEK293/Wt‐Tr, HEK293/D, HEK293/D7F, and HEK293/D‐Tr cells were examined using a 3, 4, 5‐dimethyl‐2H‐tetrazolium bromide (MTT) assay as described previously.^{( 29 )}

Immunoprecipitation.  The culture cells were washed twice with ice‐cold phosphate‐buffered saline (PBS) (–), and lysed with a lysis buffer containing 20 mM Tris‐HCl (pH 7.0), 50 mM NaCl, 5 mM ethylenediaminetetraacetic acid, 10 mM Na pyrophosphate, 50 mM NaF, 1 mM Na orthovanadate, 1% TritonX‐100, and the Complete Mini protease inhibitor mix (Roche Diagnostics). The lysates were cleared by centrifugation at 15 000 g for 10 min and the protein concentrations of the supernatants were measured using a bicinchoninic acid protein assay (Pierce Biotechnology, Rockford, IL, USA).

The cell lysates (500 µg) were immunoprecipitated by overnight incubation with 3 µg anti‐EGFR antibody, anti‐HER3 antibody (Upstate Biotechnology, Lake Placid, NY, USA), anti‐HER2 antibody (Santa Cruz Biotechnology, Santa Cruz, CA), or anti‐c‐Myc (Roche Diagnostics), followed by further incubation with protein‐G agarose (Santa Cruz Biotechnology) for 1 h. Bound proteins were washed three times with lysis buffer and eluted in Laemmli sample buffer containing 2‐mercaptoethanol. The eluted proteins were subjected to 2–15% gradient sodium dodecylsulfate–polyacrylamide gel electrophoresis (SDS‐PAGE) and immmunoblotted as described above.

Immunoblotting.  Whole‐cell lysates and the immunoprecipitates were separated using 2–15% gradient SDS‐PAGE and blotted on to a polyvinylidene fluoride membrane. The membrane was probed with anti‐EGFR, anti‐HER3 (Upstate Biotechnology), anti‐phospho(Tyr845)‐EGFR, anti‐phospho(Tyr1068)‐EGFR, anti‐phospho(Tyr1173)‐EGFR, anti‐HER2, anti‐phospho‐tyrosine, anti‐p44/42 mitogen‐activated protein (MAP) kinase, anti‐phospho‐p44/42 MAP kinase, anti‐Shc, anti‐phospho‐Shc (Cell Signaling, Beverly, MA), anti‐Sos (Santa Cruz), anti‐Grb2 (BD Biosciences, San Jose, CA), and anti‐c‐Myc (Roche Diagnostics) antibodies by incubation for 2 h at room temperature and then with horseradish peroxidase‐conjugated anti‐rabbit IgG antibody or anti‐mouse IgG antibody for 1 h at room temperature. Finally, the proteins were visualized with an enhanced chemiluminescence western blotting detection system (GE Healthcare, Piscataway, NJ, USA).

Chemical crosslinking assay.  After treatment or no treatment with EGF (R&D Systems, Minneapolis, MN, USA) the chemical crosslinking assay was carried out in intact cells as described previously.^{( 13 )} The transfected cells were washed with ice‐cold PBS (+) and incubated for 30 min at room temperature in PBS (+) containing 2 mM crosslinker bis(sulfosuccinimidyl)suberate (Pierce Biotechnology). The reaction was terminated with 20 mM Tris (pH 7.5) for 15 min at room temperature. The cells were washed with PBS (+), and 15 µg protein was resolved by 2–15% gradient SDS‐PAGE and then immunoblotted with anti‐EGFR and anti‐phospho‐EGFR antibodies.

Results

Epidermal growth factor receptor lacking C‐terminal autophosphorylation sites (EGFR‐D‐Tr and EGFR‐D7F) retains signal transduction.  To examine the role of the tyrosine residues in the C‐terminal region of EGFR in signal transduction, we constructed vectors containing wild‐type EGFR, a deletion mutant (delE746_A750 EGFR) with C‐terminal truncation, or a mutant with substitution of seven tyrosine residues in the C‐terminal region (Fig. 1a), and transfected these vectors into HEK293 cells with rather low expression levels of endogenous EGFR. The expression of exogenous EGFR in the transfectants was confirmed by immunoblotting with anti‐EGFR antibodies (Fig. 1b).

Figure 1.

Epidermal growth factor receptor (EGFR) constructs and their expression. (a) Structures of the various EGFR mutants. EGFR‐Wt, wild‐type human EGFR; EGFR‐Wt‐Tr, wild‐type kinase domain of EGFR with C‐terminal truncation at amino acid 980; EGFR‐D, EGFR with a 15‐bp deletion from the tyrosine kinase domain (delE746_A750); EGFR‐D7F, 15‐bp deletion of EGFR (delE746_A750) and substitution of seven tyrosine residues to phenylalanine (Y992F, Y1068F, Y1045F, Y1068F, Y1086F, Y1148F, Y1173F); and EGFR‐D‐Tr, 15‐bp deletion of EGFR (delE746_A750) with C‐terminal truncation at amino acid 980. EGFR‐Wt, EGFR‐D, EGFR‐D7F, and EGFR‐D‐Tr contained a myc‐tag. EGFR‐Wt‐Tr contained a flag‐tag. ECD, extracellular domain; TK, tyrosine kinase; TM, transmembrane. (b) Stable transfectants were lysed and cell lysates containing equal amounts of protein were immunoblotted with anti‐EGFR antibody recognizing the extracellular domain of EGFR. A band with a molecular weight of ~170 kDa was detected in the HEK293/Wt, HEK293/D, and HEK293/D7F cells, and a band of lower molecular weight was detected in the HEK293/D‐Tr and HEK293/Wt‐Tr cells. Mock, HEK293/Mock; Wt, HEK293/Wt; Wt‐Tr, HEK293/Wt‐Tr; D, HEK293/D; D7F, HEK293/D7F; and D‐Tr, HEK293/D‐Tr.

In order to examine the signal transduction of EGFR in the transfectants, we analyzed the phosphorylation status of EGFR and its downstream molecules. Phosphorylation of EGFR at the Y845 and Y1173 tyrosine residues was detected in HEK293/Wt and HEK293/D cells cultivated in medium containing 10% fetal bovine serum (Fig. 2a). Enhanced phosphorylation of the Y1068 tyrosine residue was observed specifically in the HEK293/D cells, suggesting that Y1068 is constitutively active in delE746_A750 EGFR. This phenomenon is consistent with our previous reports.^{( 29} , ^{30 )} On the other hand, no significant phosphorylation of Y845, Y1068, or Y1173 was observed in the HEK293/D7F and HEK293/D‐Tr cells. ERK and Akt are major downstream pathways of EGFR. We examined the phosphorylation of ERK and Akt in the transfectants. Increased phosphorylation of ERK was observed in the HEK293/D7F, HEK293/D‐Tr, and HEK293/D cells, even though HEK293/D7F and HEK293/D‐Tr cells were transfected with EGFR lacking the C‐terminal autophosphorylation sites.

Figure 2.

Epidermal growth factor receptor (EGFR) lacking C‐terminal autophosphorylation sites retains EGFR signal transduction. (a) The HEK293/Mock, HEK293/Wt, HEK293/D, HEK293/D7F, and HEK293/D‐Tr cells were lysed, and the cell lysates were immunoblotted with anti‐phospho‐EGFR (p‐EGFR Y845, Y1068, Y1173), anti‐EGFR (recognizing the extracellular domain), anti‐phospho‐ERK, anti‐ERK, anti‐phospho‐Akt, and anti‐Akt antibodies. (b) The HEK293/Wt, HEK293/D, HEK293/D7F, and HEK293/D‐Tr cells were incubated in 1% serum starve medium for 12 h, followed by treatment with 10 ng/mL epidermal growth factor for 10 min at 37°C. The cell lysates were immunoblotted. Mock, HEK293/Mock; Wt, HEK293/Wt; D, HEK293/D; D7F, HEK293/D7F; and D‐Tr, HEK293/D‐Tr.

We also examined ligand‐dependent signals in these cells under the 1% serum starve medium (Fig. 2b). Ligand‐stimulated phosphorylation of EGFR was observed in the HEK293/Wt cells transfected with wild‐type EGFR. Constitutive phosphorylation of EGFR and a further increase in the EGFR phosphorylation response to EGF were observed in the HEK293/D cells. On the other hand, no significant phosphorylation in response to EGF binding was observed in the HEK293/D7F and HEK293/D‐Tr cells. Downstream of EGFR, increased phosphorylation of ERK and Akt was observed in response to EGF in the HEK293/D7F and HEK293/D‐Tr cells, as well as the HEK293/Wt and HEK293/D cells. These results indicate that EGFR lacking the C‐terminal autophosphorylation sites (EGFR‐D‐Tr and EGFR‐D7F) retained signal transduction ability.

Transfectants with EGFR lacking C‐terminal autophosphorylation sites retain their hypersensitivity to EGFR TKI.  EGF stimulation increased the growth of HEK293/Wt cells significantly but did not affect their sensitivity to AG1478 (data not shown). To examine the role of the C‐terminal region of EGFR in cellular sensitivity to EGFR TKI, the sensitivity of these transfectants was examined by growth‐inhibition assay (Fig. 3a). HEK293/Wt and HEK293/Wt‐Tr cells with normal EGFR in relation to the kinase domain were relatively resistant to EGFR TKI, with IC₅₀ values of 3.0 ± 0.97 and 8.1 ± 0.99 µM. On the other hand, HEK293/D (0.028 ± 0.018 µM), HEK293/D7F (0.047 ± 0.030 µM), and HEK293/D‐Tr (0.017 ± 0.017 µM) cells were ~100 times more sensitive to AG1478 compared to HER293/Wt cells (Fig. 3a), suggesting that the cells transfected with EGFR lacking C‐terminal phosphorylation sites retained hypersensitivity to EGFR TKI. There were no differences in the proliferation rates of these cell lines under the absence of drug exposure (data not shown).

Figure 3.

Sensitivity of cell growth and downstream epidermal growth factor receptor (EGFR) signaling to AG1478 in the mutant EGFR transfectants. (a) The growth‐inhibitory effect of AG1478 in HEK293/Wt, HEK293/Wt‐Tr, HEK293/D, HEK293/D7F, and HEK293/D‐Tr cells. The seeded cells were exposed to AG1478 for 72 h and the cellular proliferative activity was determined by MTT assay. (b) The HEK293/Wt, HEK293/D, HEK293/D7F, and HEK293/D‐Tr cells were incubated in 1% serum starve medium for 12 h, followed by exposure to 20 or 200 nM AG1478 for 3 h at 37°C. The cell lysates were immunoblotted with anti‐phospho‐EGFR (p‐EGFR Y1068), anti‐EGFR (recognizing the extracellular domain), anti‐phospho‐ERK, or anti‐ERK antibodies. Mock, HEK293/Mock; Wt, HEK293/Wt; Wt‐Tr, HEK293/Wt‐Tr; D, HEK293/D; D7F, HEK293/D7F; D‐Tr, HEK293/D‐Tr.

To elucidate the effect of EGFR TKI on the EGFR‐triggered signal cascade, the phosphorylation status of EGFR and ERK was examined in the transfectants treated with AG1478 under the 1% serum starve medium (Fig. 3b). AG1478 at a concentration of 20 nM inhibited the phosphorylation of EGFR in HEK293/D cells, but not in the other cell lines. The increased phosphorylation of ERK observed in the HEK293/D, HEK293/D7F, and HEK293/D‐Tr cells was inhibited by AG1478 at 20 nM. These results suggest that signal transduction from C‐terminal‐truncated EGFR to downstream molecules allows sensitivity to EGFR TKI to be retained, just like the deletion mutant of EGFR (delE746_A750).

Endogenous HER families are not involved in the dimerization of EGFR‐D‐Tr and EGFR‐D7F.  We hypothesized that the signals from EGFR lacking the C‐terminal autophosphorylation sites were transduced through heterodimerization with endogenous EGFR, HER2, or HER3. No significant endogenous EGFR expression or its phosphorylation was observed in the HEK293/Mock cells (Fig. 4a). Very low levels of intrinsic HER2 or HER3 expression were detected in the HEK293 cells, and the expression levels seemed not to be involved in significant drug sensitivity nor increased signal transduction (Fig. 4b,c). Therefore, it is not likely that heterodimerization of EGFR lacking C‐terminal autophosphorylation sites with endogenous HER receptors contributes to the signal transduction. It is thus speculated that homodimerization of EGFR lacking C‐terminal autophosphorylation sites transduces the signals to downstream molecules. Indeed, the results of the chemical crosslinking assay revealed clear homodimerized bands in the HEK293/Wt, HEK293/D, and HEK293/D7F cells (Fig. 5a). In the HEK293/D‐Tr cells, homodimerized bands with lower molecular weights (indicated by the black arrow) were detected (Fig. 5a) and these dimers were not phosphorylated (Fig. 5b). Taken together, we speculate that EGFR lacking C‐terminal autophosphorylation sites form homodimers.

Figure 4.

Heterodimerization of mutant epidermal growth factor receptor (EGFR) with endogenous receptors of the HER family in mutant EGFR transfectants. Expression of endogenous EGFR and response to epidermal growth factor stimulation. HEK293/Mock, HEK293/Wt, HEK293/D, HEK293/D7F, and HEK293/D‐Tr cells were incubated in 1% serum starve medium for 12 h followed by the addition of 10 ng/mL epidermal growth factor for 10 min at 37°C. (a) The whole‐cell lysates of HEK293/Mock and HEK293/W+ cells containing equal amounts of protein were immunoblotted with anti‐phospho‐EGFR (p‐EGFR Y1068) and anti‐EGFR (recognized extra‐cellular domain). (b,c) The lysates were immunoprecipitated with anti‐EGFR, anti‐HER2, or anti‐HER3 antibodies, and immunoblotted with anti‐EGFR, anti‐HER2, anti‐HER3, or anti‐phosphotyrosine antibodies to detect the dimerization and phosphorylation of EGFR and endogenous HER2 or HER3. HER2, HER2‐introduced HEK293 cells as a positive control; HER3, HER3‐introduced HEK293 cells as a positive control.

Figure 5.

The cells under 1% serum starved medium were allowed to react with 2 mM of the chemical crosslinking regent BS₃ before the crosslinking reaction was quenched. The cell lysates were immunoblotted with (a) anti‐epidermal growth factor receptor (EGFR) (recognizing the extracellular domain) and (b) anti‐phosphoEGFR (p‐EGFR Y1173) antibodies to detect the dimerization and phosphorylation of wild‐type and mutant EGFR. Black arrow, EGFR dimer; open arrow, EGFR monomer. Mock, HEK293/Mock; Wt, HEK293/Wt; D, HEK 293/D; D7F, HEK293/D7F; D‐Tr, HEK293/D‐Tr. EGF, epidermal growth factor.

Despite a lack of C‐terminal autophosphorylation sites, transfected cells retain their capacity for EGFR‐dependent Shc phosphorylation.  Binding of adaptor proteins to the C‐terminal region of EGFR is essential for EGFR signal transduction. It is widely recognized that tyrosines 1068 and 1086 are most important for Sos and Grb2 activation; EGFR‐D7F and EGFR‐D‐Tr lack these tyrosine residues. Sos and Grb2 were coprecipitated with EGFR in the HEK293/Wt and HEK293/D cells, but not in the HEK293/D7F or HEK293/D‐Tr cells (Fig. 6a). The bands were confirmed by reblotting of the membranes used for immunoblotting (data not shown). Another adaptor protein, Shc, also binds to the C‐terminal region of EGFR, and phosphorylation of Shc activates the ERK pathway. An increase in phosphorylated p46 and p52 Shc was observed in the HEK293/D, HEK293/D7F, and HEK293/D‐Tr cells compared with the HEK293/Mock and HEK/Wt cells (Fig. 6b). The phosphorylation of Shc observed in the HEK293/D, HEK293/D7F, and HEK293/D‐Tr cells was completely inhibited by 20 nM AG1478 (Fig. 6c). These results suggest that EGFR lacking C‐terminal autophosphorylation sites activates Shc in a C‐terminal‐independent manner, and that Shc‐mediated signals may be involved in the hypersensitivity to EGFR TKI of HEK293 cells expressing EGFR lacking C‐terminal autophosphorylation sites.

Figure 6.

Interaction between mutant epidermal growth factor receptor (EGFR) and adaptor proteins. The cells were cultured under normal conditions. (a) The lysates of HEK293/Wt, HEK293/D, HEK293/D7F, and HEK293/D‐Tr cells were immunoprecipitated with anti‐myc tag antibody; the precipitates were immunoblotted with anti‐son of sevenless homolog (Sos) and anti‐growth factor receptor‐bound protein (Grb) 2 antibodies. (b) Whole‐cell lysates containing equal amounts of protein were immunoblotted with anti‐phospho‐Src homology and collagen homology (Shc) and anti‐Shc antibodies. (c) The cells incubated in 1% serum starve medium for 12 h were treated with 20 or 200 nM AG1478 for 3 h and then the lysates were immunoblotted with anti‐phospho‐Shc or anti‐Shc antibodies. Mock, HEK293/Mock; Wt, HEK293/Wt; D, HEK293/D; D7F, HEK293/D7F; D‐Tr, HEK293/D‐Tr. PY, anti‐phospho‐tyrosine.

Discussion

In the present study, we investigated the relationship between phosphorylation of tyrosine residues in the C‐terminal region of EGFR and cellular sensitivity to EGFR TKI. Increased phosphorylation of Shc and ERK was observed in HEK293/D7F and HEK293/D‐Tr cells, which expressed EGFR lacking autophosphorylation sites in the C‐terminal region. Previous reports have demonstrated that cells expressing EGFR lacking C‐terminal autophospholylation sites retain EGF‐induced mitogenic and transforming activity.^{( 27} , ^{28 )} Our data and these previous reports suggest that there exist other EGFR signaling pathways besides those mediated by the C‐terminal tyrosine residues. In addition, these signaling pathways are operative in the active EGFR mutant (delE746_A750) as well as wild‐type EGFR.^{( 28 )} The results of our growth‐inhibition assay demonstrated the hypersensitivity of HEK293/D7F and HEK293/D‐Tr cells to EGFR TKI, and phosphorylation of ERK and Shc in these cells was also inhibited. These results suggest that this EGFR signaling pathway contributes to tumor cell growth.

We demonstrated the hypersensitivity of the transfectants (HEK293/D7F and HEK293/D‐Tr cells) to AG1478. We previously reported the hypersensitivity of transfectants carrying mutant EGFR to AG1478 as well as gefitinib, ZD6474, and erlotinib.^{( 8} , ¹⁵ , ^{31 )} Therefore, it can be easily speculated that the HEK293/D7F and HEK293/D‐Tr cells would also be hypersensitive to the clinically available EGFR TKI and AG1478.

Somatic EGFR mutation in lung cancer has been reported, and over 20 types of mutations have been reported.^{( 10 )} The L858R point mutation in exon 21 of EGFR is a major point mutation (such as in delE746_A750) that contributes to EGFR TKI hypersensitivity.^{( 32 )} Interestingly, we constructed cells that overexpressed EGFR‐Wt‐Tr and EGFR‐D‐Tr, and a mutant truncated form of EGFR similar to EGFR‐Wt‐Tr was previously found in patients with glioblastoma.^{( 33 )} The mutant was truncated at amino acid 958 of EGFR and the frequency was relatively high in 7 of 48 patients. Therefore, it would be of interest to determine in future studies whether this C‐terminal‐truncated form of delE746_A750 EGFR, similar to EGFR‐D‐Tr, might be identifiable in human materials in the clinical setting.

We attempted to clarify the signaling pathway from the C‐terminal region of EGFR. We observed the phosphorylation of ERK and Shc in HEK293/D7F and HEK293/D‐Tr cells, and these phosphorylations were inhibited by exposure to AG1478. These phosphorylations were not observed in HEK293/Mock, HEK293/Wt, or HEK293/Wt‐Tr cells. Our results suggest that the constitutively active mutant EGFR lacking C‐terminal autophosphorylation sites is sufficient for activation of the downstream pathway. However, it remains unknown how signals are transduced from EGFR without a C‐terminal region to Shc, as no direct binding of Grb2 or Shc with EGFR lacking the C‐terminal region was detected in the HEK293/D7F and HEK293/D‐Tr cells (Fig. 6a). We attempted to identify the mediator molecules binding to EGFR‐D‐Tr and EGFR‐D7F by mass analysis of immunoprecipitates; however, no clear mediator molecules were identified. As a possible indirect mechanism, Sasaoka et al. postulated that ErbB2–Shc signals from EGFR lacking C‐terminal autophosphorylation sites.^{( 34 )} However, we consider this unlikely from the results of our experiments because no significant expression of Erb2 was detected in the HER293 cells.

The results of the crosslinking assay demonstrated that a complex of lower molecular weight was present in the HEK293/D‐Tr cells compared with the HER293/Wt cells, indicating that truncated EGFR forms homodimers in the HEK293/D‐Tr cells. Thus, it can be speculated that homodimerized truncated EGFR directly transduces signals downstream.

In conclusion, our results indicate that an as‐yet‐unknown signaling pathway of EGFR exists that is independent of the C‐terminal region of EGFR, and these regions are not required for cellular sensitivity to EGFR TKI.

Supporting information

Table S1. Primer set for epidermal growth factor receptor cDNA with substitution of seven tyrosine residues to phenylalanine in the C‐terminal region

Please note: Wiley‐Blackwell are not responsible for the content or functionality of any supporting materials supplied by the authors. Any queries (other than missing material) should be directed to the corresponding author for the article.

Supporting info item