Antitumor Activity of Epidermal Growth Factor Receptor–Related Protein Is Mediated by Inactivation of ErbB Receptors and Nuclear Factor-κB in Pancreatic Cancer

Pancreatic cancer is the fourth leading cause of cancer mortality in both men and women. Approximately 32,000 Americans each year will develop and also die from this disease (1). The disease is characterized by early locoregional spread and distant metastasis. Thus, the majority of patients present with advanced disease that is not resectable. For these patients, systemic chemotherapy has been largely ineffective, although gemcitabine, with and without other combinations, has shown a modest clinical benefit and has become a standard chemotherapy for advanced pancreatic cancer. The median survival of patients with advanced disease continues to be <6 months (1), suggesting that evaluation of novel targeted therapeutic agents is urgently needed to improve the outcome of patients diagnosed with this deadly disease.

Among many genes, the gene family of tyrosine kinase plays important roles in carcinogenesis of human cancers, including pancreatic cancer (2). The ErbB family of type 1 receptor tyrosine kinases (RTKs) includes four members, epidermal growth factor receptor (EGFR; ErbB1), HER-2 (ErbB2), HER-3 (ErbB3), and HER-4 (ErbB4), which share a similar primary structure and are widely expressed in human tissues. Except for HER-2, binding of receptor-specific ligands to the ectodomain of EGFR, HER-3, and HER-4 results in the formation of homodimeric and heterodimeric kinase active complexes into which HER-2 is recruited as a preferred partner (2, 3).

Increased expression of EGFR and its ligand has been detected in human pancreatic cancer tissues (4). Therefore, blockade of EGFR activity should interrupt EGFR-mediated signal transduction pathways and should result in suppression of tumor growth (5). The major partner of EGFR is HER-2, whose overexpression has also been observed in numerous human cancers where it is associated with multiple drug resistance, higher metastatic potential, and decreased patient survival (6). Aberrant HER-2 expression in pancreatic ductal adenocarcinoma has also been reported in a number of studies, with a prevalence ranging from 7% to 58% (6). HER-3 has impaired kinase activity due to substitutions in crucial residues in the tyrosine kinase domain, but HER-3 does contain six docking sites for the p85 adaptor subunit of phosphatidylinositol 3-kinase (PI3K). HER-3 becomes phosphorylated and functions as a signaling entity when it is dimerized with another erbB receptor (2). The aberrant signaling of erbB family members, such as EGFR, HER-2, and HER-3, has been causally associated with enhanced pancreatic cancer cell proliferation and shorter survival in patients with pancreatic cancer (5–8). However, the role of HER-4 in pancreatic cancer is confusing (9, 10).

Because of deregulated expression and activation of multiple members of the erbB family of RTK, multiple therapeutic strategies have been designed to manipulate EGFR and HER-2. In addition to small-molecule inhibitors of EGFR, monoclonal antibodies (mAb) to EGFR, such as Erbitux (cetuximab, ImClone Systems/Bristol-Myers Squibb, Somerville, NJ), which competitively inhibits ligand binding to the receptor resulting in attenuation of EGFR signaling, showed inhibition of tumor growth (11). Although no small-molecule inhibitors of HER-2 have been developed, mAbs to HER-2, such as Herceptin (Trastuzumab, Genentech, South San Francisco, CA), have been used for the treatment of breast cancers that express high levels of HER-2 (12). Although Erbitux and Herceptin treatments showed signs of success in a limited number of patients with tumors showing increased expression of EGFR or HER-2, failure in others may partly be due to coexpression of multiple EGFR family members leading to an enhanced transforming potential and poor prognosis (13). Therefore, discovering a specific agent that targets multiple members of the erbB family would be important in pancreatic cancer therapy.

We recently isolated a novel negative regulator of EGFR, termed EGFR-related protein (ERRP), whose expression seems to attenuate EGFR activation (14). We have found that ERRP is expressed in most benign pancreatic ductal epithelium and islet cells, but in tumors, we found a progressive loss in ERRP expression, which could partly contribute to the aggressive tumor cell growth in pancreatic adenocarcinoma (15). We have also observed that recombinant ERRP inhibits EGFR activity and the growth of pancreatic cancer cell line BxPC-3 in vitro (16). In addition, we have shown that ERRP could inhibit growth of multiple cancer cells in vitro, function as a pan-erbB inhibitor, and inhibit tumor growth of colon cancer cells in vivo (17, 18). However, such in vivo studies have not been shown in pancreatic cancer. In this study, we investigated whether ERRP could serve as a pan-erbB inhibitor by examining the effects of recombinant ERRP on the growth and ligand-induced activation of multiple members of erbB family in three pancreatic cancer cell lines that express varying levels of EGFR and other member(s) of its family, specifically HER-2. Additionally, we compared the inhibitory effects of ERRP with that of Erbitux or Herceptin, mAbs to EGFR and HER-2, respectively. Most importantly, we have shown that ERRP inhibits pancreatic tumor growth in vivo in severe combined immunodeficient (SCID) xenograft model and that the antitumor activity was correlated with tumor cell differentiation and inactivation of nuclear factor-κB (NF-κB) activity.

Cells and experimental reagents. Human pancreatic cancer cell lines BxPC-3, HPAC, and PANC-1 were obtained from the American Type Culture Collection (Rockville, MD). Cells were cultured in low-glucose DMEM, containing fetal bovine serum, and antibiotic solution consisting of 100 units/mL penicillin sodium, 100 μg/mL streptomycin sulfate (Life Technologies Bethesda Research Laboratories, Gaithersburg, MD). Primary antibodies for phosphorylated EGFR (phospho-EGFR, Y1173) and phospho-HER-2 (Y887) were obtained from Upstate Biotechnology (Lake Placid, NY). Primary antibodies for phospho-HER-2 (Y1248) and phospho-p65 were purchased from Cell Signaling (Beverly, MA). Primary antibodies for EGFR, HER-2, HER-3, and HER-4 were obtained from Santa Cruz Biotechnology (Santa Cruz, CA). All secondary antibodies were obtained from Pierce (Rockford, IL). Chemiluminescence detection of proteins was done with the use of a kit from Amersham Biosciences (Amersham Pharmacia Biotech, Piscataway, NJ). Protease inhibitor cocktail, 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT), and all other chemicals were obtained from Sigma (St. Louis, MO).

Antibodies to ERRP. Polyclonal antibodies against ERRP were generated as described previously (17). Briefly, rabbits were immuned using an epitope from the “U” region of ERRP comprising 15 amino acids (AVTRPLHPLAQNRVS) that showed no homology with any known sequence in the data base. Western blot analysis of rat liver and gastric mucosa revealed that ERRP antibodies cross-reacted strongly to a protein with molecular mass of about 55 kDa that corresponds well with the calculated molecular mass of ERRP. No cross-reactivity with any member of the EGFR family of proteins (170-190 kDa) was observed suggesting that the antibodies are specific to ERRP.

Generation of recombinant ERRP. ERRP fusion protein was generated using the Drosophila expression system (Invitrogen, Carlsbad, CA) as described previously, and silver staining of the SDS-PAGE resulted in a predominant protein band of M_r 53 to 55 following purification (16, 17). The protein band of M_r 53 to 55 corresponded well with the calculated molecular mass of ERRP, which is composed of 479 amino acids. In the absence of CuSO₄, no 53- to 55-kDa protein was detected. Immunoaffinity-purified ERRP was used in all experiments.

Cell culture and growth assay. Human BxPC-3, HPAC, and PANC-1 pancreatic cancer cells were grown in DMEM supplemented with 5% fetal bovine serum, 100 units/mL penicillin, and 100 μg/mL streptomycin (complete medium) at 37°C in humidified air with 5% CO₂. In some experiments, cells were cultured in serum-free medium wherever indicated. Transforming growth factor-α (TGF-α, 7 nmol/L; Invitrogen) or HB-EGF (5 nmol/L, Sigma) was added to the medium whenever necessary as indicated in the figure legend. To perform the growth assays, cells were incubated overnight at a density of 5,000 per well in 96-well plates and washed in PBS and subsequently in DMEM medium in the absence or presence of the specified experimental conditions as indicated in the figure legend, and MTT assay was done as described earlier (16). Results were plotted as means ± SD of three separate experiments having six determinations per experiment for each experimental condition. Student's t test was used for statistical analysis.

Western blot. Cells were lysed in lysis buffer [50 mmol/L Tris (pH 7.5), 100 mmol/L NaCl, 1 mmol/L EDTA, 0.5% NP40, 0.5% Triton X-100, 2.5 mmol/L sodium orthovanadate, 10 μl/mL protease inhibitor cocktail, and 1 mmol/L phenylmethylsulfonyl fluoride] by incubating for 20 minutes at 4°C. The protein concentration was determined using the Bio-Rad assay system (Hercules, CA). Total proteins were fractionated using SDS-PAGE and transferred onto nitrocellulose membrane for Western blotting as described earlier (16). For reprobing, membranes were incubated for 30 minutes at 50°C in buffer containing 2% SDS, 62.5 mmol/L Tris (pH 6.7), and 100 mmol/L 2-mercaptoethanol; washed; and incubated with desired primary antibody.

BxPC-3 xenografts. Four-week-old female ICR-SCID mice were obtained from Taconic Laboratory (Germantown, NY). The mice were adapted to animal housing and BxPC-3 xenografts were developed as described earlier (19). Briefly, each mouse received 10⁷ BxPC-3 cells (in serum-free RPMI 1640) s.c. in each flank area. When s.c. tumors developed to ∼1,500 mg, the tumors were excised, and serial propagation was accomplished by trimming extraneous material, cutting the tumors into fragments of 20 to 30 mg, which were then transplanted s.c. using a 12-gauge trocar into the flanks of a new group of mice for maintenance of tumors as well as for experimental purpose.

For the subsequent drug efficacy trials, small fragments of the BxPC-3 xenograft were implanted s.c. and bilaterally into naive, similarly adapted mice. Mice were checked three times per week for tumor development. Once transplanted, BxPC-3 fragments developed into palpable tumors (60-100 mg); groups of five animals were removed randomly and assigned to different treatment groups. Using this model, the efficacy of ERRP was studied. ERRP at 1 mg/kg was given as s.c. injection for seven injections every other day. Mice in the control and ERRP-treated group were followed for measurement of s.c. tumors, changes in body weight, and side effects of the drugs. Tumors were measured two times per week. Tumor weight (mg) was calculated using the formula (A × B²) / 2, where A and B are the tumor length and width (in mm). To avoid discomfort in the control group, animals were euthanized when their total tumor burden reached 2,000 mg. Tumor tissues were harvested for histologic, immunohistochemical, and NF-κB activity analysis. All studies involving mice were done under Animal Investigation Committee–approved protocols. Tumor weights in SCID mice were plotted against time.

Alcian blue and nuclear fast red staining. ERRP-treated or control mice were sacrificed by cervical dislocation after general anesthesia. Immediately after sacrifice the tumors were removed and cut in two pieces. One half was frozen in liquid nitrogen and stored at −70°C for further nuclear protein extraction and measurement of DNA-binding activity of NF-κB by electrophoretic mobility shift assay (EMSA), and the other half was fixed in 4% buffered formalin and paraffin embedded using standard protocol. Four-micrometer-thick consecutive sections were cut from formalin-fixed, paraffin-embedded tissue blocks stained with Alcian blue and Nuclear Fast Red, respectively. Stained sections were dehydrated, mounted in xylene, and analyzed using standard light microscopy.

Immunohistochemical determination of Ki67, phospho-EGFR, and p65. The expression of Ki67 (proliferative marker), phospho-EGFR, and p65 was detected in histologic sections of tumor xenografts. Sections were cut from formalin-fixed, paraffin-embedded tissue blocks; collected on 3-ethoxy-aminoethyl-silane-treated slides; and allowed to dry overnight at 37°C. Sections were dewaxed in xylene, rehydrated through graded concentrations of ethanol to distilled water, immersed in 10 mmol/L citrate buffer (pH 6.0), and processed in a thermostatic water bath for 40 minutes at 98°C for antigen retrieval. Anti-Ki67 (clone MIB1; BioGenex Laboratories, San Ramon, CA; dilution 1:100), anti-phospho-EGFR, and anti-p65 antibodies were applied on three slides for each case, and incubations were done overnight at room temperature in a humidified atmosphere followed by a 30-minute incubation of secondary antibody. Slides were then incubated with streptavidin peroxidase and visualized using the 3,3′-diaminobenzidine chromogen (Lab Vision Corp., Fremont, CA).

Tumor tissue nuclear protein extraction and EMSA. Tumor specimens excised from ERRP-treated or control mice were processed for the isolation of nuclear proteins as described by Banerjee et al. (20). EMSA was done by incubating 10 μg of nuclear extract with IRDye-700-labeled NF-κB oligonucleotide. The incubation mixture included 2 μg of poly(deoxyinosinic-deoxycytidylic acid) in a binding buffer. The DNA/protein complex formed was separated from free oligonucleotide on 8.0% native polyacralyamide gel using buffer containing 50 mmol/L Tris, 200 mmol/L glycine (pH 8.5), and 1 mmol/L EDTA and then visualized by Odyssey Infrared Imaging System using Odyssey Software Release 1.1. Antiretinoblastoma immunoblotting with nuclear protein was done as loading control.

Constitutive expression of erbB family of RTK in pancreatic cell lines. Initial studies were done to examine the relative levels of EGFR and its family members in three pancreatic cancer cell lines, such as BxPC-3, HPAC, and PANC-1 by Western blot analysis. All three cell lines expressed high levels of EGFR and HER-2, respectively (Fig. 1A), but HER-4 could not be detected in any of these cell lines. The latter could be due to the lack of sensitivity of the commercial antibody used. BxPC-3 and HPAC expressed low and modest levels of HER-3, respectively, and PANC-1 did not show any detectable level of HER-3 (Fig. 1A).

We also examined the ligand-induced activation of EGFR, HER-2, and HER-3 in these cells. Treatment of cells with TGF-α (7 nmol/L) for 20 minutes showed dramatic increase in phosphorylation of EGFR (Y1173) and phosphorylation of HER-2 (Y1248) in all three cell lines tested (Fig. 1B). TGF-α also induced phosphorylation of HER-2 on Tyr⁸⁸⁷ residue in BxPC-3 cells but not in HPAC and PANC-1 cells (Fig. 1B), indicating cell type–dependent phosphorylation on specific sites in response to TGF-α. This initial characterization of pancreatic cancer cells was important for subsequent experiments as indicated below, particularly with respect to complexities of cellular signaling by erbB family of RTK.

ERRP-, Erbitux-, or Herceptin-induced cell growth inhibition in BxPC-3, HPAC, and PANC-1 cells. We examined the growth-inhibitory effects of ERRP, Erbitux, and Herceptin using the MTT assay in three human pancreatic cancer cell lines BxPC-3, HPAC, and PANC-1, which express varying levels of EGFR, HER-2, and HER-3. MTT provides a measure of mitochondrial dehydrogenase activity within the cell and thereby offers an indication of cellular proliferation status. ERRP showed strong growth inhibition in all three cell lines tested. ERRP inhibited the growth of BxPC-3 cells in a dose-dependent manner, revealing ∼50% growth inhibition with a dose of 5 μg/mL, and 80% to 90% inhibition with a dose of 10 μg/mL after 72 hours of treatment (Fig. 2). These results are consistent with our published report (16).

Therapeutic mAbs targeting the extracellular domains of EGFR and HER-2 have been developed, which showed efficacy in the treatment of patients with cancer. Anti-EGFR and anti-HER-2 antibodies are widely used in the management of patients with breast cancer and some other solid cancers. Compared with ERRP, we tested the effects of Erbitux and Herceptin on pancreatic cancer cell growth. We observed that the degree of cell growth in all three cell lines by Erbitux and Herceptin was significantly lower compared with ERRP. Although Erbitux inhibited cell growth in BxPC-3 and PANC-1 cells in a dose-dependent manner, 50% inhibition was achieved with a dose of 20 μg/mL after 72 hours of treatment (Fig. 2). Moreover, compared with the other two cell lines, HPAC cells were much less sensitive to Erbitux, as treatment of this cell line with a dose of 20 μg/mL induced only 15% inhibition in cell growth (Fig. 2). Herceptin resulted in cell growth inhibition comparatively at higher concentrations (10 μg/mL for PANC-1 and 20 μg/mL for BxPC-3 and HPAC). Herceptin at a dose of 20 μg/mL only inhibited cell growth by about 15% in all three cell lines (P < 0.01; Fig. 2). These results showed a greater responsiveness of these cell lines to ERRP compared with either Erbitux or Herceptin, suggesting that ERRP could be a better therapeutic agent for pancreatic cancer. We subsequently tested the inhibitory effects of Erbitux, Herceptin, and ERRP on growth factor–induced activation of EGFR.

Effects of ERRP, Erbitux, or Herceptin on TGF-α- and HB-EGF-induced EGFR activation in pancreatic cancer cell lines. EGFR activity is believed to be required for growth of most pancreatic cancer cells in vivo and in vitro. To determine whether and to what extent ERRP affects EGFR function, BxPC-3, HPAC, and PANC-1 cells were serum starved for 48 hours and subsequently incubated with 5 μg/mL ERRP, Erbitux, or Herceptin for 90 minutes followed by exposure to 7 nmol/L TGF-α or 5 nmol/L HB-EGF for 15 minutes. Figure 2D shows a marked activation of EGFR (pY1173-EGFR) in TGF-α or HB-EGF-treated cells, whereas pretreatment of cells with ERRP or Erbitux resulted in near complete inhibition of EGFR phosphorylation in all three cell lines tested. Herceptin treatment modestly reduced TGF-α- or HB-EGF-induced EGFR phosphorylation in BxPC-3 and PANC-1 cells but had no effect in HPAC cells. Total EGFR level was not affected by ERRP treatment (Fig. 2D), suggesting specific effects of ERRP on the activation of EGFR. To further test the effect of ERRP on HER-2 phosphorylation, we conducted additional experiments as presented below.

Effects of ERRP on TGF-α- and HB-EGF-induced HER-2 activation in pancreatic cancer cell lines: Y1248 and Y887 phosphorylation. We compared the effects of ERRP with Ertitux or Herceptin on ligand-induced activation (tyrosine phosphorylation) of HER-2 in three pancreatic cancer cell lines. Exposure of cells to TGF-α-induced (7 nmol/L) or HB-EGF-induced (5 nmol/L) marked phosphorylation of HER-2 on Tyr¹²⁴⁸ (Y1248) in all three cell lines tested (Fig. 3A). In all the cell lines, the ligand-induced activation of HER-2 was inhibited by recombinant ERRP, Erbitux, or Herceptin. Although ERRP or Erbitux treatment showed near-complete inhibition of phosphorylation of HER-2 (Y1248), Herceptin pretreatment for 90 minutes reduced TGF-α- or HB-EGF-induced HER-2 phosphorylation by 60% to 80% in all three cell lines (Fig. 3A).

We also examined TGF-α- or HB-EGF-induced HER-2 phosphorylation on Tyr⁸⁸⁷ in these three cell lines. We observed that significant HER-2 phosphorylation on Tyr⁸⁸⁷ in BxPC-3 cells induced by either TGF-α or HB-EGF (Fig. 3B) but not in HPAC or PANC-1 cells (data not shown). These results suggest that HER-2 phosphorylation on Tyr⁸⁸⁷ is cell line dependent. In BxPC-3 cells, pretreatment with ERRP (5 μg/mL) or Erbitux (5 μg/mL) abolished TGF-α- or HB-EGF-induced HER-2 phosphorylation (Y887) almost completely. Pretreatment of cells with Herceptin for 90 minutes inhibited HER-2 phosphorylation by 80% (Fig. 3B). Subsequently, we tested these phenomenon on HER-3 activation as presented below.

Effects of ERRP on TGF-α- and HB-EGF-induced erbB3 activation in pancreatic cancer cells. BxPC-3 and HPAC cells expressed low or modest levels of HER-3 protein, respectively, whereas no detectable levels of HER-3 were noted in PANC-1 (Fig. 1A). TGF-α or HB-EGF treatment for 20 minutes showed a marked activation (tyrosine phosphorylation) of HER-3 in HPAC cells but not in BxPC-3 or PANC-1 cells. Again, pretreatment of cells with ERRP for 90 minutes abrogated TGF-α- or HB-EGF-induced HER3 phosphorylation by 80% (Fig. 3C). Erbitux treatment had no apparent effect on TGF-α- or HB-EGF-induced HER-3 phosphorylation. These results suggest that ERRP but not Erbitux is a strong inhibitor for both TGF-α- and HB-EGF-induced HER-3 activation. To our surprise, Herceptin almost completely inhibited HER-3 phosphorylation induced by TGF-α but had no effect on HB-EGF-induced activation of HER-3 (Fig. 3C). These results clearly suggest that ERRP could be considered as a pan-erbB inhibitor and thus may have broader use for the inhibition of tumor growth. Therefore, for the first time, we have conducted in vivo experiments in an animal model of pancreatic cancer to test the antitumor activity of ERRP and have examined whether biological targets of ERRP could be affected in animal tumors treated with ERRP.

Effect of ERRP on pancreatic tumor growth in vivo. To determine whether systemic therapy with purified ERRP protein could stunt tumor growth in animals, we established BxPC-3 human pancreatic cancer xenografts in SCID mice as described earlier (19). Mice were checked two times per week for tumor development. Once transplanted BxPC-3 fragments developed into palpable tumors (60-100 mg), ERRP was given at 1 mg/kg as s.c. injection for a total of seven injections every other day. Mice in the control and ERRP-treated groups were followed for measurement of s.c. tumors, changes in body weight, and side effects of ERRP treatment. ERRP treatment significantly inhibited tumor growth (P = 0.0007 versus vehicle) compared with untreated control (Fig. 4). There was no difference in mortality or body weight between the two groups (data not shown), indicating that ERRP treatment was not toxic to the animals. We subsequently asked the most important question of whether the antitumor activity of ERRP could be correlated with changes in the biological markers that are known to be altered, as shown earlier in our in vitro studies (16). The answer to this question is presented below.

Effect of ERRP on cancer cell proliferation and phospho-EGFR expression in vivo. To determine the biological effects of the treatment of animal tumors with ERRP, tumors were harvested from the mice for immunohistochemical analysis. Figure 5A and B shows H&E staining of sections from control and ERRP-treated mice, respectively. These sections showed a poorly differentiated adenocarcinoma characterized by sheets and clusters of epithelial cells with atypical, moderately pleomorphic and hyperchromatic nuclei. The nucleocytoplasmic ratio was high, but the cells had moderate amount of cytoplasm. Tissue sections were also stained with Alcian blue and Nuclear Fast Red for measuring the nature of epithelial cells and mucin production, key features in pancreatic cancer. Figure 5C shows Alcian blue staining in control tumor, and Fig. 5D shows staining in ERRP-treated animal tumor. In many foci, the nests of cells were punctuated by intracytoplasmic vacuoles, many of which contain basophilic mucin. Some also contained necrotic debris and apoptotic cells, which are common findings in pancreatic adenocarcinomas. The amount of mucin production seemed significantly more abundant in those tumor tissues of animals that had received ERRP. This was verified by Alcian blue stain (done at pH 2.5), which highlighted the mucin in the cells. The Ki67 nuclear labeling index, which was determined by immunohistochemical staining with mib antibody, was 20% in the ERRP-treated tumors compared with 38% in the control tumors (Fig. 5F and E, respectively). These results are consistent with our in vitro data showing that ERRP is a powerful agent for the inhibition of pancreatic cancer cell growth.

Tumors from all treatment groups were also analyzed for the expression of phospho-EGFR as well as phospho-p65 (active subunit of NF-κB). The expression levels of phospho-EGFR were significantly lower in tumors from the ERRP-treated mice (Fig. 6A) than those from vehicle-treated control mice (Fig. 6B), indicating that ERRP could down-regulate phospho-EGFR in vivo, similar to those observed in vitro. In our earlier report (16), we showed that ERRP could down-regulate the DNA-binding activity of NF-κB in a cell culture model. However, such studies have not been done in any in vivo setting. For this reason, we measured DNA-binding activity of NF-κB in tumor tissues of ERRP-treated animals and compared our results with untreated animal tumors as presented below.

Inhibition of NF-κB activity by ERRP in vivo. NF-κB plays an important role in EGFR-related cell signaling and ultimate cell growth and survival. To determine whether ERRP could affect the function of NF-κB in vivo, we examined the changes in NF-κB activity in tumor tissues using EMSA. Tumors excised from vehicle or ERRP-treated mice were homogenized, and nuclear proteins were isolated. EMSA with nuclear extracts from ERRP-treated or vehicle-treated xenograft tumor tissues revealed that treatment with ERRP strongly inhibited the DNA-binding activity of NF-κB compared with untreated controls (Fig. 4). The specificity of NF-κB DNA-binding activity was confirmed by supershift assay using anti-p65 antibody, and the equal protein loading was confirmed by retinoblastoma Western blot analysis, which showed no changes in its expression.

An immunohistochemical analysis for activated p65 (phospho-p65) was also carried out within tumor sections. Tumor sections from animals treated with ERRP (Fig. 6D) showed significantly reduced staining intensity compared with untreated control (Fig. 6C). These results showed that treatment of animals with ERRP reduced DNA-binding activity of NF-κB, which clearly suggests that ERRP was able to target specific downstream signaling molecules of the EGFR signaling (16), further implicating the role of ERRP as a pan-erbB inhibitor that targets multiple members of the erbB family of tyrosine kinase (14, 21).

Preponderance of EGFR and HER-2 in a wide variety of solid tumors has prompted extensive drug development efforts to design pharmacologic inhibitors of EGFR and HER-2. Indeed, a number of small-molecule inhibitors of EGFR, as well as mAbs to EGFR (Erbitux), have been developed (22). Although no small-molecule inhibitors of HER-2 have been developed, mAbs to HER-2 (Herceptin) have been used for the treatment of breast cancers that express high levels of HER-2. Although Erbitux and Heceptin treatments showed signs of success in a limited number of patients with tumors having increased expression of EGFR or HER-2, failure in others may partly be due to coexpression of multiple EGFR family members, leading to enhanced transforming potential and poor prognosis (13). Therefore, identification of inhibitor(s) targeting multiple members of the EGFR family is likely to provide a therapeutic benefit to a broad range of the patient population. ERRP, a recently isolated negative regulator of EGFR, possesses substantial homology to the extracellular ligand-binding domain of EGFR and its family members (14). Previously, we reported that ERRP inhibited cell growth of BxPC-3 pancreatic cancer cells in vitro (16). We also reported that recombinant ERRP is a pan-erbB inhibitor in breast and colon cancer cell lines (18). ERRP as a potential therapeutic agent for epithelial cancers also came from the observation that recombinant ERRP inhibits the growth of xenografts of colon cancer HCT-116 cells in SCID mice (23). Our current data show that ERRP is also a pan-erbB inhibitor of pancreatic cancer cell lines, similar to those observed in colon cancer cells (18).

The EGFR family comprises four distinct receptors: EGFR/ErbB1, HER-2/ErbB2, HER-3/ErbB3, and HER-4/ErbB4. The pancreatic cancer cells frequently overexpress EGFR, HER-2, HER-3, and, less frequently, HER-4, as well as six ligands that bind directly to EGFR (6). Overexpression of EGFR, HER-2, and HER-3 has also been implicated in the development and progression of pancreatic cancer (6, 8). It has been reported that HER-4 was exclusively found in nonmetastatic but not in metastatic tumors (9, 10). We also found that ERRP abrogates ligand-induced activation of EGFR, HER-2, and HER-3 in pancreatic cancer cells. In contrast, Erbitux abrogated the ligand-induced activation of EGFR and HER-2 but not HER-3. Compared with ERRP and Erbitux, Herceptin was found to be a much weaker inhibitor for TGF-α- or HB-EGF-induced EGFR and HER-2 activation. This is supported by the observation that Herceptin only partially reduced the ligand-induced EGFR and HER-2 activation. Moreover, Herceptin abrogated TGF-α-induced but not HB-EGF-induced HER-3 activation. Taken together, these results suggest that ERRP is an effective inhibitor for all three erbB family members, EGFR, HER-2, and HER-3, in pancreatic cancer.

Our data further showed that ERRP but not Erbitux or Herceptin inhibited cell growth of all three pancreatic cancer cell lines efficiently. Erbitux inhibited BxPC-3 and PANC-1 cell growth slightly, without inhibiting HPAC cell growth. This may be due to complex regulation of erbB signaling in HPAC cells. Herceptin only slightly inhibited cell growth of all three cell lines tested when high doses were used because it only modestly reduced ligand-induced activation of EGFR and HER-2. Therefore, Erbitux and Herceptin could only be effective in a limited number of patients with pancreatic cancer. Our results also suggest that EGFR or HER-2 expression level alone is not likely to be the sole determinant influencing the antiproliferative activity of Erbitux or Herceptin in pancreatic cancer cells. Erbitux, an EGFR inhibitor, did not show a strong inhibitory effect on cell growth of all three EGFR-positive cell lines, although it inhibited ligand-induced phosphorylation of EGFR. The same was also true for Herceptin; although Herceptin reduced ligand-induced HER-2 phosphorylation (activation), it failed to show strong growth inhibitory effect on these HER-2-positive cell lines.

It has been documented that activation of EGFR leads to the activation of a transcription factor (NF-κB), which has been shown to be activated in a variety of cancers. NF-κB activation due to activation of PI3K/AKT usually results in antiapoptotic signaling (24, 25). Increasing evidence of dysregulated NF-κB-associated pathways has been found in various human pancreatic cancer cell lines and primary tumors, which supports the role of NF-κB in pancreatic cancer. In our previous study, we found that ERRP inhibits EGF-induced NF-κB activation in BxPC-3 cells in vitro (16). In this study, our results show, for the first time, that NF-κB activity is significantly inhibited in the tumors of ERRP-treated animals compared with untreated controls. Thus, the down-regulation of NF-κB signaling could be achieved during ERRP-mediated pancreatic tumor growth inhibition in vivo, which clearly recapitulates our in vitro findings.

Consistent with our in vitro results (16), ERRP treatment significantly inhibited pancreatic cancer cell growth in vivo in the SCID xenograft model, which could partly be attributed to decreased proliferation as evidenced by reduced Ki-67 immunoreactivity in the tumors of ERRP-treated animals. Moreover, the levels of mucin, a typical feature of glandular differentiation (26), seemed to be substantially higher in ERRP-treated mice. These observations are interesting and novel in that the ERRP could be a powerful agent for tumor differentiation, a good prognostic feature for pancreatic cancer.

In summary, our data show that ERRP but not Erbitux or Herceptin is universally effective in inhibiting growth of BxPC-3, HPAC, and PANC-1 pancreatic cancer cells. ERRP also inhibits TGF-α- or HB-EGF-induced activation of EGFR, HER-2, and HER-3. In contrast, Erbitux or Herceptin only modestly or partially reduced ligand-induced activation of these erbB family of tyrosine kinase. Our results suggest that ERRP is an effective pan-erbB inhibitor and thus a potent antitumor agent for pancreatic cancer, resulting from inactivation of erbB and NF-κB signaling. We strongly believe that ERRP could be a novel therapeutic agent for pancreatic cancer expressing different levels and subclasses of erbB family of tyrosine kinase.

Grant support: National Cancer Institute/NIH grant 1R01CA101870-02 (F.H. Sarkar), University of Texas M.D. Anderson Cancer Center Specialized Programs of Research Excellence grant S P20CA 101936-02 on pancreatic cancer (J. Abbruzzese), and Puschelberg Foundation.

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.

We thank Carrie Koerner for editorial assistance.