Introduction

Overexpression of epidermal growth factor receptor (EGFR) in autocrine growth-regulated carcinoma and malignant glioma cells is associated with the expression of mutated EGFR species as part of neoplastic progression (Nishikawa et al., 1994; Moscatello et al., 1995; Nagane et al., 1996; Huang et al., 1997; Frederick et al., 2000). Among these EGFR variants, the 140–155 kDa EGFRvIII represents a relatively well-characterized oncoprotein with a major in-frame deletion involving exons 2–7 of the EGFR gene. Thus, the mutation eliminates most amino acids comprising subdomains I and II including the cysteine-containing EGF binding site (Nishikawa et al., 1994; Batra et al., 1995). EGFRvIII is constitutively active, cannot bind EGF (Moscatello et al., 1996), and is only minimally activated by EGF (Wong et al., 1992). EGFRvIII is widely expressed in malignant glioma and carcinoma cells, including breast carcinoma, non-small-cell lung carcinoma, ovarian carcinoma and others (Moscatello et al., 1995; Frederick et al., 2000). Mechanistic studies to elucidate the unique functions of EGFRvIII have been limited by the fact that this receptor is only expressed in vivo and is rapidly downregulated in cultured tumor cells precluding detailed investigation (Bigner et al., 1990). To date, most studies on EGFRvIII have focused on the definition of its transforming activity and its effects on downstream signal transduction pathways (Montgomery et al., 1995; Huang et al., 1997). However, the role of EGFRvIII in cellular responses to genotoxic stresses, such as ionizing radiation or other cytotoxic cancer therapeutic agents, has been only minimally investigated.

We have recently demonstrated that ionizing radiation in the therapeutic dose range of 1–5 Gy leads to 2–5-fold increases of EGFR Tyr phosphorylation resulting in radiation-induced activation of the receptor. This stimulation of EGFR mediates, through the activation of the MAPK and PI3K pathways, a major cytoprotective response (Schmidt-Ullrich et al., 1996, 1997, 1999, 2000; Dent et al., 1999) that confers on cells enhanced radioresistance due to accelerated cell proliferation (Contessa et al., 1999), an antiapoptotic response (Contessa et al., 2002), and improved repair functions after radiation exposures (Lammering et al., 2001b, 2001c). Both of these responses represent probable underlying mechanisms that modulate the radiosensitivity of human tumor cells in vitro and in vivo upon single and repeated radiation exposures (Schmidt-Ullrich et al., 1996; Lammering et al., 2001a, 2001b).

In the current study, we have extended our investigation of EGFR wild-type (wt) function on cellular radiation responses to the potentially modulating effects by EGFRvIII. As EGFRvIII is not expressed in cultured cells, these mechanistic studies were performed in vitro in Chinese hamster ovary (CHO) cells transfected with EGFRwt and/or EGFRvIII expression plasmids using conditions that yielded similar expression levels for both receptors. In these cells, the effects of EGFRvIII expression on acute cellular radiation responses were examined including the modulation of cell survival and apoptosis in vitro. The studies yielded two novel findings expanding our understanding of EGFR responses to radiation and the interactions between EGFRvIII and other ERBB receptor Tyr kinases. Based on these results, EGFRvIII furnishes a stronger cytoprotective radiation response than EGFRwt, which is likely due to the greater activation of EGFRvIII after irradiation. This activation is also transmitted into more potent stimulation of MAPK and AKT, thus identifying EGFRvIII as a critical component in the ERBB signaling circuitry and in cellular radiation responses.

Results

EGFRvIII is unresponsive to EGF but is inhibited by AG1478

The experiments in this section were designed to define the level of EGFRvIII ligand independence (Moscatello et al., 1996) and the constitutive activity of the receptor for radiation response studies in consecutive sections. These studies were performed with CHO cells that were cotransfected with phAc.EGFRvIII and pCMV.EGFRwt to achieve similar protein levels, and the Tyr phosphorylation levels of EGFRwt and EGFRvIII were quantified after treatment with EGF and/or AG1478 for inhibition of EGFR Tyr kinase activity. Western blot analyses with an anti-EGFR mAb showed similar expression levels under these conditions (Figure 1a). Treatment of CHO cells, coexpressing EGFRwt and EGFRvIII, with concentrations of 10 and 100 ng EGF resulted in expected dose-dependent increases in Tyr phosphorylation of EGFRwt. In contrast, the Tyr phosphorylation levels of EGFRvIII were only minimally affected (Figure 1b, left panel). In examining further the functional consequences of EGFRvIII expression on signaling after irradiation of CHO.EGFRvIII cells, Western blot analyses of EGFRvIII Tyr phosphorylation confirmed its complete inhibition after AG1478 treatment (Figure 1c). These data demonstrate the EGF independence of EGFRvIII activity and its sensitivity to Tyr kinase inhibition by AG1478.

Figure 1.

EGFRvIII expression in CHO cells, response to EGF and inhibition of Tyr phosphorylation by AG1478. (a) Expression levels of EGFRvIII and EGFRwt proteins in CHO cells after cotransfection with 0.3 g of the phAc.EGFRvIII plasmid and varying amounts, 0.01, 0.03 and 0.1 g, of the pCMV.EGFRwt plasmid; 48 h after transfection, protein expression levels were quantified with anti-EGFR mAb. The ratio of 0.3 and 0.03 g of EGFRwt and EGFRvIII cDNA, respectively, was chosen for experiments under panel b. The observed bands for EGFRvIII showed the typical electrophoretic pattern of two bands of previously reported sizes between 140 and 155 kDa (Nishikawa et al., 1994), whereas pCMV.EGFRwt transfected CHO cells produced the expected 170 kDa EGFR protein. (b) Cells coexpressing EGFRwt and EGFRvIII (see panel a) were incubated with 10 or 100 ng/ml of EGF for 5 min; after coimmunoprecipitation of EGFRwt and EGFRvIII, Tyr phosphorylation levels were quantified by Western blotting with an anti-Tyr phosphorylation mAb. Western blots of immunoprecipitated EGFRwt and EGFRvIII protein verified equal expression levels. Lysate of A431 cells and CHO cells transfected with 0.5 g phAc.EGFRvIII plasmid (EGFRvIII; see panel c) served as controls to identify protein size of EGFRwt and EGFRvIII, respectively. Despite the EGF-induced activation of EGFRwt, EGF treatment failed to increase the Tyr phosphorylation of EGFRvIII. (c) Complete inhibition of EGFRvIII Tyr phosphorylation by AG1478 treatment in CHO cells transfected with 0.5 g phAc.EGFRvIII plasmid (right two lanes); protein controls probed with anti-EGFRvIII mAb ensured equal loading (left two lanes). Data are representative of three independent experiments

Full figure and legend (103K)

Ligand-independent radiation-induced activation of EGFRvIII

The functional consequences of EGFRvIII expression on cellular radiation responses were examined in CHO.EGFRvIII cells, and the effects on downstream signal transduction pathways after irradiation were compared to those in CHO cells expressing EGFRwt alone (CHO.EGFRwt). Western blot analyses showed similar protein expression levels of EGFRvIII and EGFRwt 48 h after transfection using conditions described above (Figure 2a). The EGFRvIII protein showed the typical electrophoretic pattern of two bands of previously reported sizes between 140 and 155 kDa (Figure 1) (Nishikawa et al., 1994), whereas pCMV.EGFRwt transfected CHO cells produced the expected 170 kDa EGFR protein. As was previously shown for different carcinoma and malignant glioma cell lines (Schmidt-Ullrich et al., 2000; Lammering et al., 2001a), radiation increased EGFRwt Tyr phosphorylation as did its ligand EGF. Quantification of these responses in the transiently transfected cells demonstrated that treatment with 10 ng/ml EGF induced a 3.3-fold (P<0.001) increase in EGFRwt Tyr phosphorylation, which was similar to a 2.8-fold increase after irradiation of cells with 2 Gy (P=0.0015) (Figure 2b). Despite lacking response of EGFRvIII to EGF, a single radiation exposure of 2 Gy caused an immediate and maximum 4.3-fold activation of EGFRvIII (P=0.001); this transient response returned to baseline phosphorylation by 30 min (Figure 2c). The response patterns of immediate radiation-induced activation of EGFRvIII or EGFRwt were very similar in CHO.EGFRvIII or CHO.EGFRwt cells, respectively, but the incremental maximum activation was significantly greater for EGFRvIII when compared to EGFRwt (P=0.003; Figure 2b,c). Western blotting with an anti-EGFRvIII mAb revealed that both molecular weight species, the p140 and p155 protein, of EGFRvIII responded with similar increases in Tyr phosphorylation when compared to untreated controls (Figure 2c).

Figure 2.

Radiation-induced stimulation of EGFRvIII Tyr phosphorylation in CHO cells. (a) Conditions of similar expression levels of EGFRwt and EGFRvIII in CHO cells after transient transfection with 0.5 g pCMV.EGFRwt or 0.5 g phAc.EGFRvIII were used throughout. These conditions were established by transfection with increasing amounts of pCMV.EGFRwt cDNA at 0.1, 0.2 and 0.5 g. Protein expression was quantified 48 h after transfection by Western blotting (see Materials and methods). (b) EGFRwt Tyr phosphorylation after treatment with EGF or radiation. CHO cells were transfected with 0.5 g pCMV.EGFRwt cDNA and, 48 h later, exposed to 10 ng/ml of EGF or 2 Gy of radiation (see Materials and methods). Tyr phosphorylation levels of EGFRwt as a function of time after radiation were quantified by Western blotting (upper lanes) assuring identical protein loading (lower lanes). Lysate of A431 cells served as control for the protein size of EGFRwt. (c) EGFRvIII Tyr phosphorylation after treatment with EGF or radiation. CHO cells, transfected with 0.5 g phAc.EGFRvIII cDNA, were treated as described under panel b. EGFRvIII Tyr phosphorylation was quantified for EGF after 5 min and for radiation between 1 and 30 min by Western blots (upper lanes) assuring identical protein loading (lower lanes). Data are representative of at least three independent experiments

Full figure and legend (118K)

Radiation-induced activation of EGFRvIII mediates enhanced MAPK activation

Having identified EGFRvIII as a target of radiation-induced activation, we next explored the ability of EGFRvIII to transmit radiation signals into the MAPK cascade and compared the findings with the radiation-induced MAPK activation that was identified as an important downstream target of EGFRwt in previous studies from our laboratory (Contessa et al., 1999; Reardon et al., 1999; Schmidt-Ullrich et al., 2000). Using conditions similar to those described for experiments shown in Figure 2, CHO cells were transfected with phAc.EGFRvIII or pCMV.EGFRwt or subjected to mock (CHO.mock) transfection and irradiated 48 h later. Western blot analyses, as shown in Figure 2a, confirmed similar expression levels of EGFRvIII and EGFRwt. In cells expressing EGFRvIII, we found that 2 Gy induced a transient maximum 8.5-fold (P=0.009) activation of MAPK peaking at 2 min and returning to baseline levels within 30 min (Figure 3a, left panel), thus paralleling the phosphorylation response of the receptor (Figure 2c). Pretreatment of cells with AG1478 reduced this response >75% to a twofold radiation-induced MAPK activation substantiating its dependence upon EGFRvIII function. By comparison, activation of EGFRwt after 2 Gy mediated a significant maximum 3.5-fold (P=0.045) activation of MAPK within 5 min (Figure 3a, middle panel), a response that was also blunted by AG1478. The modest twofold activation of MAPK in CHO.mock cells was interpreted as a background response by these cells as it was not affected by treatment with AG1478 (Figure 3a, right panel) and, therefore, was likely independent of EGFRvIII or EGFRwt expression. The different responses of CHO.EGFRvIII and CHO-EGFRwt cells to EGF treatment were impressively corroborated by the MAPK analyses. In CHO.EGFRvIII cells, EGF induced a modest but significant, 1.4–1.6-fold (P=0.02), activation of MAPK whereas CHO.EGFRwt cells yielded a fivefold increase (P=0.004; Figure 3b). The radiation-induced activation of MAPK in CHO cells expressing EGFRvIII was greater than the EGF-induced activation of MAPK in cells expressing EGFRwt (Figure 3b). These data demonstrate that EGFRvIII expression has a profound effect of cellular radiation responses leading to potent receptor activation and transmitting amplified signals along the MAPK pathway.

Figure 3.

EGFRvIII- and EGFRwt-dependent activation of MAPK after exposure of transiently transfected CHO cells to EGF or radiation. (a) Radiation (2 Gy)-induced MAPK activation in CHO cells expressing EGFRvIII, EGFRwt or no EGFR protein. The effects of EGFRvIII or EGFRwt proteins were assessed through controls from cells pretreated with AG1478 to eliminate receptor Tyr phosphorylation completely. Cells were transfected with 0.5 g cDNA each of phAc.EGFRvIII (CHO.EGFRvIII), pCMV.EGFRwt (CHO.EGFRwt) or an empty vector (CHO.mock) as described in Materials and methods (see Figure 2a). Irradiation and MAPK activity time course experiments were initiated 48 h after transfection. For EGFR inhibition, cells were pretreated with 5 M AG1478 for 30 min. Activity is expressed as fold increase of ³²P incorporation relative to basal activity without treatment. Data are expressed as means.d. of three independent experiments. Western blot analyses of representative cell samples transfected in parallel under identical conditions verified constant expression levels of the corresponding proteins and no EGFR expression in CHO.mock cells. (b) MAPK activity, normalized relative to basal levels without treatment in CHO.EGFRvIII or CHO.EGFRwt cells upon AG1478 treatment or EGF exposure at 10 or 100 ng/ml, compared to the radiation response at 2 Gy. Data for the fold activation of MAPK activity at 2 or 5 min after 2 Gy for CHO.EGFRvIII and CHO.EGFRwt, respectively, were taken from panel a. Data are presented as means.d. of three independent experiments

Full figure and legend (103K)

Radiation-induced activation of EGFRvIII is also transmitted through the AKT/PI3K pathway

In other cell systems, the PI3K signal transduction pathway has been identified as another important downstream target of EGFRvIII (Moscatello et al., 1998). We therefore investigated the ability of radiation to induce PI3K activation by studying the response of protein kinase B (PKB)/AKT proto-oncogene in CHO.EGFRvIII cells and compared the findings with radiation-induced AKT responses in CHO.EGFRwt cells. Radiation transiently activated AKT 3.2-fold (P=0.009) within 15 min (Figure 4a, left panel). By comparison, activation of EGFRwt after 2 Gy mediated a maximum 1.4-fold (P=NS (nonsignificant)) activation of AKT within 15 min (Figure 4a, middle panel), a response that was essentially unaffected by AG1478. The complete inhibition of the radiation-induced activation of AKT in CHO.EGFRvIII cells by AG1478 strongly suggests that the response was initiated at the level of EGFRvIII, a conclusion that is further substantiated by the finding that CHO.mock cells furnished no radiation-induced AKT activation (Figure 4a, right panel). However, AG1478 treatment of CHO.EGFRvIII cells resulted in a minimum upregulation of basal AKT activity without any further stimulation by radiation (Figure 4a,b). The radiation-induced activation of AKT in CHO cells expressing EGFRvIII was greater than the radiation-induced activation of AKT in cells expressing EGFRwt (Figure 4b). These data show that radiation-induced activation of EGFRvIII is, at least in part, transmitted through the PI3K/AKT pathway.

Figure 4.

EGFRvIII- and EGFRwt-dependent activation of AKT after exposure of transiently transfected CHO cells to EGF or radiation. (a) Radiation-induced (2 Gy) AKT activation in CHO cells expressing EGFRvIII, EGFRwt or no EGFR protein (see the legend of Figure 3). The effects of EGFRvIII or EGFRwt proteins were assessed through controls from cells pretreated with AG1478 to eliminate receptor Tyr phosphorylation completely. Treatments were 48 h after transfection using conditions described under Figure 3. AKT activity is expressed as fold increase of [³²P] phosphate incorporation relative to basal activity without exposure to radiation. Data are expressed as means.d. of three independent experiments. Western blot analyses of representative cell samples transfected in parallel under identical conditions verified constant expression levels of the corresponding proteins and no EGFRvIII or EGFRwt expression in CHO.mock cells. (b) AKT activity, normalized relative to basal levels without treatment, in CHO.EGFRwt and CHO.EGFRvIII cells upon AG1478 treatment or EGF exposure at 100 ng/ml for 15 min, compared to the radiation response at 2 Gy. Data for the fold activation of AKT at 15 min after 2 Gy were taken from panel a. Data are presented as means.d. of three independent experiments

Full figure and legend (135K)

Expression of EGFRvIII confers an enhanced cytoprotective response to ionizing radiation

To correlate radiation-induced activation of EGFRvIII with the radiosensitivity of CHO cells, we examined cell survival after single radiation doses of 2, 4 and 8 Gy using colony formation assay and comparing cells expressing EGFRvIII with cells expressing EGFRwt or mock-transfected cells. In addition, the contribution of EGFRvIII was examined through quantitative inhibition of EGFRvIII function using AG1478. As assessed by colony formation assays, CHO.EGFRvIII cells were less sensitive to varying single radiation exposures than CHO.EGFRwt or mock-transfected CHO cells (P<0.001; Figure 5a). EGFRvIII mediated a significantly stronger cytoprotective response to radiation than EGFRwt (P<0.001), as reflected by a dose-enhancement ratio (DER) of 1.61 relative to 1.32 for control relative to CHO.EGFRvIII and CHO.EGFRwt cells, respectively.

Figure 5.

EGFRwt- and EGFRvIII-mediated modulation of radiosensitivity after exposure of transiently transfected CHO cells to varying single doses of ionizing radiation. (a) Clonogenic survival curves for CHO.EGFRwt, CHO.EGFRvIII or CHO.mock cells after radiation with single doses of 2, 4 or 8 Gy. For experimental conditions, see Materials and methods. The surviving fraction after 12 days was determined from the number of colonies with 50 cells relative to the number of cells plated after correction for plating efficiency. Data are provided as means.d. from a minimum of three experiments. (b) Conversion of the EGFRvIII- mediated modulation of radiosensitivity through treatment with AG1478. The effects of AG1478 treatment on the cellular radiosensitivity of CHO.mock and CHO.EGFRvIII were quantified by colony formation after irradiation of 2 and 4 Gy; pretreatment of cells with AG1478 was as described under Materials and methods. The surviving fraction was determined as described under panel 5a. Data are provided as means.d. from a minimum of three experiments

Full figure and legend (91K)

As demonstrated in Figure 5b, expression of EGFRvIII conferred an almost twofold increase in clonogenic survival after exposure to 4 Gy, indicating a doubling of radioresistance in this assay conferred by EGFRvIII expression compared to CHO.mock cells (Figure 5b, right panel). AG1478 treatment almost completely eliminated this increase in radioresistance of CHO.EGFRvIII cells (Figure 5b), again indicating that the cytoprotective response is initiated at the level of EGFRvIII.

Expression of EGFRvIII inhibits radiation-induced apoptosis

We further investigated the effects of EGFRvIII and EGFRwt expression on the radiation-induced apoptosis in CHO cells 24 h following ionizing radiation exposure (Kennedy et al., 1999; Contessa et al., 2002). By comparison, irradiation of CHO.EGFRvIII cells with a single dose of 4 Gy induced only a minimal 1.1-fold increase (P=NS) in apoptosis relative to unirradiated controls. In contrast, exposure of both CHO.EGFRwt and CHO.mock cells to 4 Gy significantly enhanced apoptosis by an average of 1.9-fold (P=0.015; Figure 6). These results demonstrate that the presence of EGFRvIII, but not EGFRwt, counteracts radiation-induced apoptosis in CHO cells, and suggest that the antiapoptotic response is mediated through radiation-induced EGFRvIII activation.

Figure 6.

EGFRvIII-mediated modulation of radiation-induced apoptosis in CHO cells. Radiation-induced apoptosis in CHO.EGFRvIII, CHO.EGFRwt and CHO.mock cells after exposure of cells to 4 Gy. The cells were irradiated with a dose of 4 Gy and harvested 24 h following irradiation. Radiation significantly increased apoptosis in CHO.EGFRwt and CHO.mock compared to CHO.EGFRvIII cells. Fold increase was calculated as fold increase in the per cent apoptosis in irradiated compared to control cells. The results are the average of two independent experiments, provided as means.e.m (s.e. of the mean)

Full figure and legend (52K)

Discussion

The data presented in this study provide new insights into the biological functions of the constitutively active EGFRvIII. To date, mechanistic studies on this naturally occurring truncated receptor have been limited since its spontaneous expression is only seen in vivo in naturally occurring tumors or xenografts of human tumor cells (Lammering et al., 2001a) and is rapidly lost in vitro (Bigner et al., 1990). In extension of previous studies on the role of EGFRvIII as an oncoprotein (Ekstrand et al., 1994; Nishikawa et al., 1994; Batra et al., 1995; Nagane et al., 1996; Huang et al., 1997), have started to define its involvement in cellular responses to genotoxic stress, such as ionizing radiation. Our results define EGFRvIII as an important response molecule that is unresponsive to EGF (Huang et al., 1997) but is activated by radiation to a greater extent than EGFRwt. Since intensified responses of EGFRvIII are also transmitted as amplified signals at the level of MAPK (Montgomery et al., 1995) and AKT, EGFRvIII can be expected to mediate a more powerful cytoprotective response after irradiation than we have previously demonstrated for EGFRwt (Contessa et al., 1999; Schmidt-Ullrich et al., 2000; Lammering et al., 2001a, 2001b).

Studies in CHO cells were performed because of the high transfection efficiencies (Ekstrand et al., 1994), and the ability to control expression levels of both EGFRvIII and EGFRwt. The independent expression of both EGFR molecules was required to define their relative roles in generating activation responses, as quantified by receptor Tyr phosphorylation, after exposure of cells to growth factor ligands or ionizing radiation. While we have characterized the responses for EGFRwt in other cells, it was important to quantify the potential contribution of the constitutively active EGFRvIII to cytoprotective responses in the same cell system. In addition to the controlled expression of EGFRwt and EGFRvIII, we also confirmed their effects on cellular signaling and survival responses using the EGFR-specific tyrphostin kinase inhibitor AG1478 (Han et al., 1996).

Irradiation of CHO cells, expressing similar levels of EGFRwt or EGFRvIII, demonstrated independent activation of each receptor (Schmidt-Ullrich et al., 1996, 1997). However, the 4.3-fold stimulation of EGFRvIII was consistently higher, yielding a significant 1.8-fold (P=0.03) greater activation relative to EGFRwt. The fundamental difference in response of the receptors was demonstrated upon exposure of cells to EGF that induced a 3.3-fold activation of EGFRwt but no significant enhancement in Tyr phosphorylation of EGFRvIII most likely due to deletions in the COOH-terminal EGF-binding domain (Wong et al., 1992). A minor activation response through interaction with other ERBB receptors cannot be excluded (Chu et al., 1997; O'Rourke et al., 1998b). The radiation response of EGFRvIII may be of considerable importance since we have demonstrated in previous studies that EGFR furnishes a significant cytoprotective response conferring cellular radioresistance that is mediated by radiation-induced proliferation and antiapoptosis (Contessa et al., 1999), and by direct radiosensitization (Lammering et al., 2001a, 2001b) most likely due to transcription-controlled increased expression of repair proteins, including PCNA (Amorino et al., 2002; Amorino et al., unpublished results). In line with previous experiences, the relative importance of radiation-induced activation of both EGFRwt and EGFRvIII is adequately studied by exposure of cells to AG1478 since this treatment almost completely inhibited activation of different EGFR species and, therefore, can be expected to block downstream signaling (Han et al., 1996; Schmidt-Ullrich et al., 1997; Figure 1).

The radiation-induced activation signals from EGFR are transmitted through the MAPK pathway (Contessa et al., 1999; Reardon et al., 1999; Lammering et al., 2001b). This is affirmed by our finding that cells expressing EGFRvIII produced a minimum 2.5-fold greater activation of MAPK than cells expressing EGFRwt (Figure 3). The close functional link between EGFR Tyr phosphorylation and MAPK activation in transfected CHO cells was corroborated by the almost complete inhibition of MAPK by AG1478, confirming observations in NIH3T3 and U87MG cells that were transfected with EGFRvIII (Montgomery et al., 1995; Moscatello et al., 1995; Han et al., 1996). Since we have previously identified MAPK as a downstream effector of the EGFR-mediated cytoprotective response after irradiation, it can be expected that EGFRvIII expression in cells enhances this response relative to EGFRwt. These conclusions may apply to PI3K, which demonstrates higher activity levels in EGFRvIII expressing cells (Moscatello et al., 1998) and acts in concert with MAPK in transmitting EGFR signals after irradiation to enhance antiapoptotic responses (Contessa et al., 2002).

While activation of MAPK and PI3K could predict the decreased radiosensitivity of cells, the contribution of EGFRvIII or EGFRwt to cellular radioresistance was more directly quantified by clonogenic survival of CHO.EGFRvIII and CHO.EGFRwt cells. The expression of EGFRvIII raised the surviving fraction in a radiation dose-dependent manner, demonstrating that CHO cells expressing EGFRvIII were significantly more resistant than their counterparts transfected under identical conditions with EGFRwt or control plasmids (P< 0.001). After 2 Gy the surviving fraction increased from approximately 50 to 85% (1.7-fold), an effect completely reversed upon inhibition of the receptor by AG1478. This difference changed to a 2.5-fold increase in relative radioresistance of EGFRvIII expressing cells after 4 Gy. These data represent the most direct demonstration that radiation-induced activation of EGFR species provides a strong cytoprotective signal with increased radioresistance through enhanced proliferation and repair (Contessa et al., 1999; Reardon et al., 1999; Lammering et al., 2001a). From studies on p185neu (ERBB2), the impact of ERBB receptors on cellular radiosensitivity may not be limited to EGFR and its mutants (O'Rourke et al., 1997, 1998a, 1998b), but likely depend upon the entire ERBB expression profile of a given tumor cell (Bowers et al., 2001; Lammering et al., 2001a; Contessa et al., 2002).

In summary, this study demonstrates the importance of EGFRvIII, likely expressed in many human tumors in vivo, in tumor cell radiation responses. Possibly due to its state of constitutive activation, EGFRvIII produces a greater activation signal after irradiation of cells than previously described for EGFRwt. Therefore, EGFRvIII can be expected to also produce a more pronounced cytoprotective response mediated by amplified activation of downstream effectors such as MAPK and PI3K. These findings therefore imply that in some carcinoma and glioma cell types, therapeutic strategies targeting EGFR need to assure activity against EGFRvIII in order to achieve most effective radiosensitization of human carcinoma and malignant gliomas.

Materials and methods

Reagents and cell lines

Unless otherwise specified, all reagents were from Sigma Chemical Co. (St Louis, MO, USA). All electrophoresis reagents were from BioRad (Hercules, CA, USA); the EGFR tyrosine kinase inhibitor tyrphostin AG1478 was from Calbiochem (San Diego, CA, USA); EGF was purchased from Sigma Chemical Co. (St Louis, MO, USA). Media and antibiotics were from GIBCO-BRL (Rockville, MD, USA), and fetal bovine serum (FBS) was from Intergen (Purchase, NY, USA). The following immunological reagents were used: the immunoprecipitating anti-EGFR monoclonal antibody (mAb) from Transduction Laboratories (Lexington, KY, USA), the anti-Tyr phosphorylation mAb, Ab2, from Oncogene Science (Cambridge, MA, USA), the anti-EGFR mAb for Western blotting, Ab14, from Neomarkers (Fremont, CA, USA), and anti-EGFRvIII mAb, DH8.3 from Abcam (Cambridge, UK). The secondary Ab, an anti-mouse alkaline phosphatase conjugate, was from Promega (Madison, WI, USA). Other reagents included Protein G Plus/Protein A Agarose from Oncogene Science (Cambridge, MA, USA), mAb to p42/44 MAPK (ERK2), C-14, and to AKT (sc-1619) both as agarose conjugates from Santa Cruz (Santa Cruz, CA, USA). Myelin basic protein (MBP) as substrate for MAPK (Sigma Chemical Co., St Louis, MO, USA) or AKT (Santa Cruz, Santa Cruz, CA, USA) was used as stocks of 10 and 5 mg/ml in H₂O, respectively. The pCL6 plasmid was used as an empty vector control. The EGFRwt cDNA was kindly provided by A Ullrich (Max-Planck-Institute for Biochemistry, Martinsried, Germany), and the phAc.EGFRvIII expression plasmid was obtained from D Bigner (Department of Pathology, Duke University, Durham, NC, USA) (Batra et al., 1995).

The CHO cell line and the A431 squamous carcinoma cell line were obtained from the American Type Tissue Collection (ATTC; Rockville, MD, USA). All cells were maintained at 37°C in 95% air/5% CO₂ except for irradiation, which was performed at 20°C in air. The CHO and the A431 cell lines were maintained in RPMI1640 medium containing 5% FBS (RPMI/5FBS) and antibiotics (penicillin/streptomycin). Only mycoplasma-free cultures were employed as assured by monthly testing.

Transient transfection of CHO cells

CHO cells were routinely seeded at 1.5 10⁵ cells per 60-mm dish. Cells were cultured for a total of 5 days in RPMI/5FBS. On day 3, transient transfections were performed. The conditions for optimal transfection efficiency were determined by Western analysis of EGFRvIII and EGFRwt protein expression levels 48 h after transfecting the CHO cells with varying amounts of the expression plasmids. In parallel transfection, experiments with identical amounts of a pK7.GFP (green fluorescence protein) plasmid, provided by I. Macarra (University of Virginia, Charlottesville, VA, USA), re performed for microscopic quantification of transfection efficiency of GFP-positive cells after 48 h. Optimal conditions yielding >90% transfection efficiencies were established with 0.5 g of the phAc.EGFRvIII or pCMV.EGFRwt expression plasmid and 1.5 g of the empty vector using Lipofectamine Plus™ as a carrier (GIBCO-BRL, Gaithersburg, MD, USA). For mock transfections, referred to as CHO.mock, 2 g of the empty vector was transfected under identical conditions. For the development of CHO cells coexpressing EGFRvIII and EGFRwt, varying amounts of the pCMV.EGFRwt DNA were cotransfected with 0.3 g of the phAc.EGFRvIII DNA and varying amounts of the empty vector keeping the total amount of DNA at 2 g.

Cell treatments and irradiation

Cell treatments and irradiation experiments were performed on day 5, 48 h after transfection. Pretreatment of cells was with 5 M AG1478 for 30 min. EGF exposures of cells was at 10 or 100 ng/ml for 5 min except for EGF treatment for AKT assays, which was for 15 min. For all irradiation experiments, cells were exposed to single doses of ionizing radiation at a dose rate of 1.8 Gy/min using a ⁶⁰Co source. In time course experiments, cells were irradiated and incubated at 37°C for the times specified. Thereafter, media were removed and cells were washed once in ice-cold PBS, rapidly frozen on dry ice and stored at -80°C until further processing.

Immunoprecipitation and Western blot analysis

Immunoprecipitation and Western blotting experiments were performed as previously described (Schmidt-Ullrich et al., 1997; Contessa et al., 1999; Reardon et al., 1999). Briefly, frozen cells were lysed in ice-cold lysis buffer, and cell lysis was facilitated by repeated passage through a 20-gauge needle. The protein levels in supernatants obtained after centrifugation (16 000 r.p.m., microcentrifuge) were quantified with a Bio-Rad protein assay kit (Bio-Rad Labs, Richmond, CA, USA). Equal amounts of cellular protein were employed for Western blotting of EGFRwt, EGFRvIII and Tyr phosphorylation or for immunoprecipitation using anti-EGFR mAb at 4°C for 90 min followed by incubation with Protein G Plus/Protein A Agarose for 45 min. Immunoprecipitates were washed once in lysis buffer and twice in PBS. After size fractionation in 6% SDS–polyacrylamide gels, proteins were transferred electrophoretically onto nitrocellulose membranes (BioRad, Hercules, CA, USA). The membranes were then incubated with primary mAb followed by incubation with peroxidase-conjugated secondary mAb. For visualization, bands were developed with the CDP-Star kit (Tropix, Inc., Bedford, MA, USA). For analysis of Tyr phosphorylation levels autoradiograms were quantified using Sigma Scan software (Jandel Scientific, San Rafael, CA, USA) (Schmidt-Ullrich et al., 1996).

Immune complex kinase assays

Identical protein aliquots from cell lysates were incubated with anti-ERK2 or anti-AKT mAbs at 4°C for 3 h or overnight, respectively. After washing once with lysis buffer and twice with 1 kinase buffer, 25 mM -glycerophosphate, pH 7.4, 15 mM MgCl₂, immunoprecipitates were suspended in 40 l H₂O containing the appropriate substrates, consisting of 20 g MBP for MAPK or 1 g AKT substrate for AKT. The kinase reaction was initiated by the addition of 10 l of 5 kinase buffer, 500 M ATP, and [^-32P]ATP (5000 c.p.m./pmol) (ICN Biomedicals, Irvine, CA, USA) and maintained at 37°C for 30 min. Aliquots of the assay mixture were placed on P81 paper and nonspecific binding of [^-32P]ATP was removed by extensive washing in 180 mM phosphoric acid. Incorporation of ³²P into bound MBP or AKT substrate was quantified by absorption to P81 followed by liquid scintillation spectroscopy (Beckman Instruments Inc., Schaumburg, IL, USA) (Reardon et al., 1999).

Colony formation assay

Irradiation of cells was 48 h after transfection at doses of 2, 4 and 8 Gy. Cells were incubated for an additional 24 h, harvested and plated for clonogenic survival. The number of cells was adjusted to generate 50–300 colonies per dish at each radiation dose and plated into 60-mm culture dishes. Cells were maintained at 37°C with 5% CO₂ for 12 days, stained with crystal violet, and colonies containing 50 cells were counted to determine surviving fractions (Lammering et al., 2001a, 2001b).

Measurement of apoptosis

Apoptosis rates were determined 24 h following exposure to 4 Gy. The culture medium and trypsinized cells were centrifuged and resuspended in methanol : acetic acid (3 : 1). The cells were then washed twice in methanol : acetic acid, pipetted onto microscopic slides and allowed to air dry. Following Giemsa staining, averages of 1500 cells were counted per slide to determine the percentage of apoptosis.

Statistical analysis

The effects of various treatments were compared using Student's t-test. Statistical comparisons between clonogenic survival curves were carried out by use of the F-test. The radiation DERs for mock control cells compared to conditions of EGFRvIII or EGFRwt expression were derived from the ratios of the dose for a survival of 37% (D₃₇) for the survival curves for EGFRvIII or EGFRwt corresponding cells and the mock control survival curves in single-dose clonogenic survival assays. Differences yielding a P-value of <0.05 were considered statistically significant. All values and means shown are s.d. (standard deviation), unless otherwise specified.

References

1. Amorino GP, Hamilton VM, Valeric K, Dent P, Lammering G and Schmidt-Ullrich RK. (2002). Mol. Biol. Cell., 13, 2233–2244.
2. Batra SK, Castelino-Prabhu S, Wikstrand CJ, Zhu X, Humphrey PA, Friedman HS and Bigner DD. (1995). Cell Growth Differ., 6, 1251–1259. | PubMed | ISI | ChemPort |
3. Bigner SH, Humphrey PA, Wong AJ, Vogelstein B, Mark J, Friedman HS and Bigner DD. (1990). Cancer Res., 50, 8017–8022. | PubMed | ChemPort |
4. Bowers G, Reardon D, Hewitt T, Dent P, Mikkelsen RB, Valerie K, Lammering G, Amir C and Schmidt-Ullrich RK. (2001). Oncogene, 20, 1388–1397. | Article | PubMed | ISI | ChemPort |
5. Chu CT, Everiss KD, Wikstrand CJ, Batra SK, Kung HJ and Bigner DD. (1997). Biochem. J., 324 (Part 3), 855–861. | PubMed |
6. Contessa JN, Hampton J, Lammering G, Mikkelsen RB, Dent P, Valerie K and Schmidt-Ullrich RK. (2002). Oncogene, 21, 4032–4041. | Article | PubMed | ChemPort |
7. Contessa JN, Reardon DB, Todd D, Dent P, Mikkelsen RB, Valerie K, Bowers GD and Schmidt-Ullrich RK. (1999). Clin. Cancer Res., 5, 405–411. | PubMed |
8. Dent P, Reardon DB, Park JS, Bowers G, Logsdon C, Valerie K and Schmidt-Ullrich R. (1999). Mol. Biol. Cell, 10, 2493–2506. | PubMed | ISI | ChemPort |
9. Ekstrand AJ, Longo N, Hamid ML, Olson JJ, Liu L, Collins VP and James CD. (1994). Oncogene, 9, 2313–2320. | PubMed | ISI | ChemPort |
10. Frederick L, Wang XY, Eley G and James CD. (2000). Cancer Res., 60, 1383–1387. | PubMed | ISI | ChemPort |
11. Han Y, Caday CG, Nanda A, Cavenee WK and Huang HJ. (1996). Cancer Res., 56, 3859–3861. | PubMed | ISI | ChemPort |
12. Huang HS, Nagane M, Klingbeil CK, Lin H, Nishikawa R, Ji XD, Huang CM, Gill GN, Wiley HS and Cavenee WK. (1997). J. Biol. Chem., 272, 2927–2935. | Article | PubMed | ISI | ChemPort |
13. Kennedy SG, Kandel ES, Cross TK and Hay N. (1999). Mol. Cell. Biol., 19, 5800–5810. | PubMed | ISI | ChemPort |
14. Lammering G, Hewit TH, Hawkins WT, Contessa JN, Reardon DB, Lin PS, Valerie K, Dent P, Mikkelsen RB and Schmidt-Ullrich RK. (2001a). J. Natl. Cancer Inst., 93, 921–929. | Article | PubMed | ChemPort |
15. Lammering G, Lin PS, Contessa JN, Hampton JL, Valerie K and Schmidt-Ullrich RK. (2001b). Int. J. Radiat. Oncol. Biol. Phys., 51, 775–784. | Article | PubMed | ISI | ChemPort |
16. Lammering G, Valerie K, Lin PS, Mikkelsen RB, Contessa JN, Feden JP, Farnsworth J, Dent P and Schmidt-Ullrich RK. (2001c). Clin. Cancer Res., 7, 682–690. | PubMed | ISI | ChemPort |
17. Montgomery RB, Moscatello DK, Wong AJ, Cooper JA and Stahl WL. (1995). J. Biol. Chem., 270, 30562–30566. | Article | PubMed | ISI | ChemPort |
18. Moscatello DK, Holgado-Madruga M, Emlet DR, Montgomery RB and Wong AJ. (1998). J. Biol. Chem., 273, 200–206. | Article | PubMed | ISI | ChemPort |
19. Moscatello DK, Holgado-Madruga M, Godwin AK, Ramirez G, Gunn G, Zoltick PW, Biegel JA, Hayes RL and Wong AJ. (1995). Cancer Res., 55, 5536–5539. | PubMed | ISI | ChemPort |
20. Moscatello DK, Montgomery RB, Sundareshan P, McDanel H, Wong MY and Wong AJ. (1996). Oncogene, 13, 85–96. | PubMed | ISI | ChemPort |
21. Nagane M, Coufal F, Lin H, Bogler O, Cavenee WK and Huang HJ. (1996). Cancer Res., 56, 5079–5086. | PubMed | ISI | ChemPort |
22. Nishikawa R, Ji XD, Harmon RC, Lazar CS, Gill GN, Cavenee WK and Huang HJ. (1994). Proc. Natl. Acad. Sci. USA, 91, 7727–7731. | Article | PubMed | ChemPort |
23. O'Rourke DM, Kao GD, Singh N, Park BW, Muschel RJ, Wu CJ and Greene MI. (1998a). Proc. Natl. Acad. Sci. USA, 95, 10842–10847. | Article | PubMed | ChemPort |
24. O'Rourke DM, Nute EJ, Davis JG, Wu C, Lee A, Murali R, Zhang HT, Qian X, Kao CC and Greene MI. (1998b). Oncogene, 16, 1197–1207. | Article | PubMed | ChemPort |
25. O'Rourke DM, Qian X, Zhang HT, Davis JG, Nute E, Meinkoth J and Greene MI. (1997). Proc. Natl. Acad. Sci. USA, 94, 3250–3255. | PubMed |
26. Reardon DB, Contessa JN, Mikkelsen RB, Valerie K, Amir C, Dent P and Schmidt-Ullrich RK. (1999). Oncogene, 18, 4756–4766. | Article | PubMed |
27. Schmidt-Ullrich RK, Contessa JN, Dent P, Mikkelsen RB, Valerie K, Reardon DB, Bowers G and Lin PS. (1999). Radiat. Oncol. Invest., 7, 321–330. | Article |
28. Schmidt-Ullrich RK, Dent P, Grant S, Mikkelsen RB and Valerie K. (2000). Radiat. Res., 153, 245–257. | Article | PubMed | ISI | ChemPort |
29. Schmidt-Ullrich RK, Mikkelsen RB, Dent P, Todd DG, Valerie K, Kavanagh BD, Contessa JN, Rorrer WK and Chen PB. (1997). Oncogene, 15, 1191–1197. | Article | PubMed | ChemPort |
30. Schmidt-Ullrich RK, Valerie K, Fogleman PB and Walters J. (1996). Radiat. Res., 145, 81–85. | Article | PubMed | ChemPort |
31. Wong AJ, Ruppert JM, Bigner SH, Grzeschik CH, Humphrey PA, Bigner DS and Vogelstein B. (1992). Proc. Natl. Acad. Sci. USA, 89, 2965–2969. | Article | PubMed | ChemPort |

Acknowledgements

This research was supported by PHS Grants P01 CA72955 and R01CA65896 (to R Schmidt-Ullrich) from the National Cancer Institute, National Institutes of Health, Department of Health and Human Services; by the Florence and Hyman Meyers Head and Neck Cancer Research Fund; by the Deutsche Krebshilfe, Dr Mildred Scheel Stiftung fellowship support (to G Lammering); and by Grant DAMD17-99-1-9426 (to P Dent) from the US Army. We acknowledge the F-test executed by Kathy Dawson, PhD, VCU, Department of Biostatistics.