Down-Regulation of the erbB-2 Receptor by Trastuzumab (Herceptin) Enhances Tumor Necrosis Factor-related Apoptosis-inducing Ligand-mediated Apoptosis in Breast and Ovarian Cancer Cell Lines that Overexpress erbB-2

INTRODUCTION

Death receptors of the TNF 3 receptor family (e.g., TNF receptor and Fas receptor) induce apoptosis on binding to their specific ligands (e.g., TNF and Fas ligand, respectively) via activation of caspases (1, 2, 3) . These death receptor pathways offer an attractive method to induce apoptosis in cancer cells and may thereby serve as a treatment of cancer. This approach has been hampered due to lack of efficacy and prohibitive toxicity [e.g., TNF fails to induce apoptosis in most cancer cells (4) , and Fas induces lethal liver damage (5)] . TRAIL binds to the death receptors TRAIL-R1 (also called DR4) and TRAIL-R2 (also called DR5/TRICK-2/KILLER) and induces caspase-mediated apoptosis (6, 7, 8) . TRAIL has been reported to induce apoptosis selectively in a variety of cancer cell lines but not in normal cells (9, 10, 11, 12, 13, 14, 15, 16) . Also, animal studies have shown that TRAIL can induce regression of cancer xenografts without toxicity to normal tissues (11 , 17) . The selective induction of apoptosis in cancer cells but not in normal cells has prompted investigation into the use of TRAIL in cancer therapy.

However, not all cancer cells undergo apoptosis when treated with TRAIL. We have previously shown that a majority of breast cancer cells lines are resistant to TRAIL-mediated apoptosis (18) . Similar results have been reported in other cancer cell lines (12 , 19, 20, 21, 22) . Several mechanisms have been described that may control sensitivity to TRAIL-mediated apoptosis. These include the expression of decoy receptors (TRAIL-R3/DcR1 and TRAIL-R4/DcR2) that bind to TRAIL but do not activate the caspase cascade (6 , 7 , 10 , 15) ; the expression of inhibitory downstream molecules such as FLIPs (23 , 24) and IAPs (25) ; and the activation of antiapoptotic transcription factors such as nuclear factor κB (26, 27, 28, 29, 30) . Studies in resistant cancer cell lines have failed to identify the major determinants of TRAIL sensitivity (18 , 20) . Recently, we reported that the combination of chemotherapy and TRAIL enhances TRAIL-mediated apoptosis in breast cancer cell lines. However, this combination also resulted in increased death of normal breast epithelial cells (18) . Thus, it is important not only to overcome resistance to TRAIL but also to target therapy specifically to the cancer cells.

erbB-2 is a member of the epidermal growth factor receptor family and is overexpressed in breast (30%) and ovarian (15–30%) carcinomas (31, 32, 33, 34) . The overall survival rate and time to relapse for patients whose tumors overexpress erbB-2 are significantly shorter than those in patients whose tumors lack erbB-2 overexpression (35 , 36) . erbB-2 overexpression has been demonstrated to enhance proliferative, prosurvival, and metastatic signals in breast cancer cell lines (37, 38, 39) . For example, Akt kinase activity and MAPK activity are up-regulated in cells overexpressing erbB-2 (39, 40, 41) . The anti-erbB-2 antibody, trastuzumab (Herceptin), has clinical activity alone and in combination with chemotherapy in metastatic breast cancer (42, 43, 44) . The mechanisms that account for the biological effect of trastuzumab alone or in combination with chemotherapy are not known completely (45) . In vitro studies suggest that trastuzumab down-regulates the erbB-2 receptor and consequently inhibits downstream pathways involved in cell survival, cell proliferation, and metastasis (46) . As a consequence of erbB-2 down-regulation in breast cancer cell lines, inhibition of proliferation and enhancement of apoptosis have been described (40 , 42) .

In the present work, we investigate whether the combination of trastuzumab and TRAIL could enhance selective TRAIL-mediated apoptosis in cancer cells overexpressing the erbB-2 receptor.

MATERIALS AND METHODS

Cell Lines.

Three erbB-2-overexpressing cell lines (SKBr-3, MDA-MB-453, and SKOv-3) and three cell lines with low erbB-2 expression (MDA-MB-468, SW-626, and MDA-MB-231) were obtained from American Type Culture Collection (Manassas, VA) and grown in culture according to the instructions provided with them. The MCF-7 cell line (with normal erbB-2 expression) was a gift provided by Dr. Ed Chu (Yale Comprehensive Cancer Center, New Haven, CT) and was cultured in RPMI 1640 supplemented with 10% FCS and 1% penicillin/streptomycin. The low erbB-2 receptor-expressing cell line MDA-MB-231 has been demonstrated to be highly sensitive to TRAIL and was used as a positive control for TRAIL treatment (18) .

To generate stable clones, HA-tagged myristylated Akt kinase cDNA in PSRα (gift provided by Dr. Philip Tsichlis; Fox Chase Cancer Research Center, Philadelphia, PA) and the PSRα vector were introduced into the MDA-MB-453 cell line using LipofectAMINE transfection (Life Technologies, Inc., Gaithersburg, MD). Clones were isolated after selection in 800 μg/ml G418 antibiotic (Life Technologies, Inc.).

GST and GST-TRAIL Fusion Protein Production.

The GST and GST-TRAIL fusion protein have been described previously (18) . To produce GST and active GST-TRAIL, GST and GST-TRAIL cDNA plasmids were transformed into DH5α Escherichia coli, and protein expression was induced with 100 mm isopropylthio-β-d-galactosidase (Amersham Pharmacia Biotech Inc., Piscataway, NJ). Bacteria were harvested and lysed by sonication in 0.1% TONE buffer [20 mm Tris-HCl (pH 7.5), 100 mm NaCl, 1 mm EDTA, and 0.1% n-octyl-β-d-glucopyranoside). GST and GST-TRAIL proteins were purified by precipitation with glutathione-Sepharose beads (Amersham Pharmacia Biotech Inc.) and then eluted from the beads with 50 mm glutathione in 0.5% TONE buffer (pH 8.5). The buffer was exchanged for PBS using a PD10 Sephadex column (Amersham Pharmacia Biotech Inc.). The proteins were analyzed by fractionation on 10% SDS-PAGE and visualized with Chromaphor reagent (Promega, Madison, WI). The protein concentrations were measured using a Bio-Rad colorimetric assay (Bio-Rad, Hercules, CA). GST-TRAIL proteins were stable when stored bound to glutathione-Sepharose beads at 4°C for up to 3 months. All experiments were performed with freshly eluted GST-TRAIL proteins because TRAIL activity decreased rapidly on storage at 4°C after elution.

Trastuzumab and TRAIL-mediated Toxicity.

To assess TRAIL-mediated cytotoxicity on trastuzumab (Genentech, South San Francisco, CA) treatment, cells were plated at 2 × 10⁴ cells/well in 96-well microtiter plates and allowed to adhere to the plates overnight. As an isotype-matched antibody, the humanized mouse IgG anti-CD20 rituximab (Genentech) was used. The adherent cells were incubated in the presence or absence of trastuzumab (or rituximab) as indicated in the figure legends for 96 h. Freshly eluted GST-TRAIL fusion protein was added for the last 16 h at the indicated concentrations. Cell viability was assessed by the MTS dye reduction assay (Cell Titer 96 AQ_ueous One Solution Cell Proliferation Assay; Promega).

All MTS data points were repeated six times, and each experiment was carried out at least three times. Results of representative experiments are given as the mean ± SD, and results of multiple experiments are given as the mean ± SE.

Flow Cytometric Detection of Apoptosis.

To assess apoptosis, cells were plated at 5 × 10⁵ cells/60-mm dish, allowed to adhere overnight, and then treated using the same conditions described in the MTS assay. The cells were trypsinized, washed with PBS, fixed in ice-cold methanol, and stored at −20°C overnight. Fixed cells were washed twice with PBS and incubated with DNase-free RNase (1 unit/ml; Roche Molecular Biochemicals, Indianapolis, IN) for 30 min at 37°C. After incubation, nuclei were stained with propidium iodide at 50 μg/ml (Roche Molecular Biochemicals). Stained cells were stored at 4°C and protected from light until flow cytometric analysis. Cells undergoing apoptosis were determined as a percentage of cells with sub-G₀-G₁ DNA content in the DNA histogram compared with the total number of cells present using the FACSort system (Becton Dickinson, Mansfield, MA) and Cell Quest Software (Becton Dickinson, San Jose, CA).

Inhibitors of Caspase Activation and PI3k Activity.

The tetrapeptide caspase inhibitor ZVAD-fmk (Biomol Research Laboratories Inc., Plymouth, PA) was resuspended in DMSO (Sigma Chemical Co., St. Louis, MO) at a concentration of 1 mm. ZVAD-fmk was added to cells treated with or without trastuzumab (as described) at a final concentration of 50 μm 1 h before TRAIL treatment. Control cells were incubated with DMSO at the same concentration as test cells. Cell viability was analyzed by the MTS assay after 16 h of incubation with TRAIL.

The inhibitor of PI3k activity, LY294002 (Alexis, San Diego, CA), was stored at a concentration of 25 mg/ml in DMSO. LY294002 was added to a final concentration of 10 μm for 6 h before TRAIL treatment. Control cells were incubated with DMSO at the same concentration as test cells. Cell viability was analyzed by the MTS assay after 16 h of incubation with TRAIL.

ODNs and TRAIL-mediated Toxicity.

Antisense (5′-CTCCATGGTGCTCAC-3′) and sense (5′-GTGAGCACCATGGAG-3′) phosphorothioate ODNs targeting the 5′ region of the erbB-2 mRNA molecule were obtained from Sigma Chemical Co./Genosys (The Woodlands, TX). The lyophilized ODNs were reconstituted in sterile distilled water to 1 mm, filter-sterilized, and stored in aliquots at −20°C as stock solutions. For subsequent experiments, the stock solutions of ODNs were diluted to give a final concentration of 1 μm. To assess TRAIL-mediated toxicity on ODN treatment, cells were plated at 2 × 10⁴ in 96-well plates, allowed to adhere overnight, and transfected with the ODNs. Diluted ODNs were mixed with 2 μg/ml LipofectAMINE (Life Technologies, Inc.), cells were exposed to the mixture for 4 h, the transfection medium was replaced with fresh culture medium, and the cells were incubated for an additional 48 h. Freshly eluted TRAIL at the concentrations indicated in the figure legends was added to the cells, which were incubated for an additional 16 h. Cell viability was analyzed by the MTS assay.

Isolation and Analysis of Protein Lysates.

Protein was extracted from cells by detergent lysis [1% Triton X-100, 10 mm Tris-HCl (pH 7.5), 150 mm NaCl, 5 mm EDTA, 10% glycerol, 2 mm sodium vanadate, and protease inhibitors (Complete tabs; Roche Molecular Biochemicals)]. Protein lysates were cleared of debris by centrifugation at 15,000 × g for 10 min at 4°C, and the concentration was assessed by the Bio-Rad colorimetric assay (Bio-Rad). Protein samples were boiled in an equal volume of sample buffer [20% glycerol, 4% SDS, 0.2% bromphenol blue, 125 mm Tris-HCl (pH 7.5), and 640 mm β-mercaptoethanol], fractionated by 10–12% SDS-PAGE, and transferred to polyvinylidene fluoride membranes (Millipore, Bedford, MA). Polyclonal rabbit anti-erbB-2 antibody (RB-103-P; Neomarkers, Fremont, CA; 2 μg/ml), mouse monoclonal horseradish peroxidase-conjugated anti-phosphotyrosine antibody (clone 4G10; Upstate Biotechnology, Lake Placid, NY; 0.5 μg/ml), rabbit polyclonal anti-TRAIL-R1 antibody (Calbiochem, San Diego, CA; 2 μg/ml), rabbit polyclonal anti-TRAIL-R2 antibody (Imgenex, San Diego, CA; 1 μg/ml), rabbit polyclonal anti-TRAIL-R3 antibody (Affinity Bioreagents, Golden, CO; 1 μg/ml), anti-TRAIL-R4 antibody (Oncogene Research Products, Cambridge, MA; 1 μg/ml), rabbit polyclonal anti-IAP-1 antibody (R&D Systems Inc., Minneapolis, MN; 1 μg/ml), rabbit polyclonal anti-IAP-2 antibody (R&D Systems Inc.; 1.5 μg/ml), rabbit polyclonal anti-FLIP antibody (Calbiochem; 2 μg/ml), rabbit polyclonal anti-Akt antibody (9272; New England Biolabs, Beverly, MA; 2 μg/ml), rabbit polyclonal anti-phospho-Akt antibody (9271; New England Biolabs; 2 μg/ml), rabbit polyclonal anti-erk-2 antibody (C-14; Santa Cruz Biotechnology, Santa Cruz Biotechnology, CA; 1 μg/ml), mouse monoclonal anti-phospho-erk (Thr 202/Tyr 204) antibody (9106; New Englands Biolabs; 0.5 μg/ml), mouse monoclonal anti-phospho-gsk-3 α/β (9331; New England Biolabs; 1 μg/ml), mouse monoclonal anti-HA (Roche Molecular Biochemicals; 2 μg/ml), rabbit polyclonal anti-HA (Y-11, sc-805; Santa Cruz Biotechnology), and mouse monoclonal anti-α-tubulin antibody (T9026; Sigma Chemical Co.) were used for immunoblotting. Goat antirabbit and antimouse antibodies conjugated with horseradish peroxidase (Amersham Pharmacia Biotech Inc.) was used to visualize immunoreactive proteins at a 1:5000 dilution using SuperSignal (Pierce, Rockford, IL) detection reagent.

Assessment of Cell Surface Death Receptor Expression.

The total TRAIL surface binding was determined by flow cytometry by measuring the binding of GST-TRAIL or GST alone. After 72 h of incubation in the presence or absence of trastuzumab, the SKBr-3 cells were washed once with cold PBS and harvested using EDTA (2.5 μm) in cold PBS. The cells were pelleted and resuspended in cold PBS containing 1% FCS, 0.02 mm sodium azide, and 0.5 mm EDTA. The cells were then incubated with freshly eluted GST or GST-TRAIL at 20 μg/ml for 1 h at 4°C. After this, the cells were washed once with cold PBS, resuspended in the solution described above, and then incubated with a mouse FITC-conjugated anti-GST antibody (Santa Cruz Biotechnology) at 10 μg/ml for 45 min at 4°C. Finally, the cells were washed once in cold PBS, and surface staining was determined using the FACSort system (Becton Dickinson) and Cell Quest Software (Becton Dickinson). The specific surface expression of TRAIL-R1 and TRAIL-R2 was determined by measuring the binding of a mouse anti-TRAIL-R1 antibody (Santa Cruz Biotechnology) or a mouse anti-TRAIL-R2 antibody (Imgenex), respectively. A purified mouse IgG1 immunoglobulin (PharMingen, San Diego, CA) was used as an isotype-matched antibody. Cells were treated and incubated as described above. After the incubation with mouse anti-TRAIL-R1 (20 μg/ml), mouse anti-TRAIL-R2 (20 μg/ml), or mouse isotype immunoglobulin (20 μg/ml), the cells were incubated with a goat FITC-conjugated antimouse IgG (Sigma Chemical Co.; 10 μg/ml). The surface staining was determined as described above.

Akt Kinase Assays.

Stable clones of MDA-MB-453 cells expressing the HA-tagged myristylated Akt kinase were plated at 2 × 10⁶ in 100-mm dishes, allowed to adhere overnight, and treated with trastuzumab as described above. HA-tagged Akt kinase was immunoprecipitated by using mouse monoclonal anti-HA antibody, and Akt kinase activity was assayed by detecting phosphorylation of gsk-3 α/β protein under conditions recommended by the manufacturer (Akt kinase assay kit; New England Biolabs). The assay was performed in triplicate.

Statistical Analysis.

Statistical comparison of mean values was performed using Student’s t test (paired and unpaired). All Ps are two-tailed. Interactions between TRAIL and trastuzumab were classified by the fractional inhibition method as follows: when expressed as the fractional inhibition of cell viability, additive inhibition produced by both inhibitors (i) occurs when i_1,2 = i₁ + i₂; synergism occurs when i_1,2 > i₁ + i₂; and antagonism occurs when i_1,2 < i₁ + i₂ (47) . The synergism was confirmed by dose-effect analysis using Syncalc software (Biososft, Cambridge, United Kingdom; Ref. 48 ).

RESULTS

Trastuzumab Enhances TRAIL-mediated Apoptosis in erbB-2-overexpressing Cancer Cell Lines.

Breast and ovarian cancer cell lines expressing different levels of the erbB-2 receptor were assessed for sensitivity to TRAIL-mediated apoptosis in the presence or absence of trastuzumab pretreatment (Fig. 1A) ⇓ . Treatment of erbB-2-overexpressing cell lines (SKBr-3, MDA-MB-453, and SKOv-3) with trastuzumab significantly enhances TRAIL-mediated growth inhibition. These cells show low or modest sensitivity to either reagent alone, but the combination results in 30–40% growth inhibition. The combined effect of trastuzumab and TRAIL in the erbB-2-overexpressing cell lines was similar to the effect of TRAIL alone in the TRAIL-sensitive cell line MDA-MB-231 (18) . In contrast, in cells with low erbB-2 receptor expression, there was no enhancement of TRAIL-mediated toxicity by pretreatment with trastuzumab in either the TRAIL-resistant (MCF-7, MDA-MB-468, and SW-626) or TRAIL-sensitive (MDA-MB-231) cell lines. The level of the erbB-2 receptor for each cell line is shown in the bottom panel of Fig. 1A ⇓ . Longer exposure demonstrated the presence of low levels of erbB-2 receptor in MDA-MB-468, SW-626, and MDA-MB-231 cells (data not shown). The enhancement of TRAIL-mediated toxicity by trastuzumab antibody required pretreatment for at least 48 h. Cotreatment or shorter pretreatments did not result in enhanced TRAIL-mediated toxicity. The enhancement of TRAIL-mediated toxicity by trastuzumab does not reflect a shift in the kinetics of TRAIL-mediated apoptosis. Longer incubation with TRAIL alone (i.e., 48 or 72 h) resulted in only slight increases in TRAIL-mediated toxicity. Pretreatment with trastuzumab resulted in a significant increase of TRAIL-mediated toxicity when TRAIL was added for 24, 48, or 72 h. Similarly, no shift in kinetics was seen when the length of trastuzumab preincubation was varied (data not shown). The lack of inhibitory effect by trastuzumab on cells expressing low levels of the erbB-2 receptor is consistent with previous reports that demonstrated that only cells overexpressing erbB-2 are inhibited by trastuzumab (49) . Interestingly, in cells expressing low levels of the erbB-2 receptor (i.e., MCF-7, SW-626, or MDA-MB-231), trastuzumab induced slight growth enhancement. Trastuzumab, like other antibodies against erbB-2, is a weak agonist for the erbB-2 receptor (42 , 50 , 51) , and this may account for the growth in cell lines expressing low amounts of erbB-2.

To assess the effect of trastuzumab further and to demonstrate the specificity of trastuzumab, the SKOv-3 cells were incubated without antibody, with trastuzumab, or with the isotype-matched antibody rituximab and then treated with varying concentrations of TRAIL (Fig. 1B) ⇓ . A significant enhancement of TRAIL-mediated toxicity by trastuzumab was seen, even at concentrations of TRAIL that alone had little or no toxic effect (i.e., 0.025 or 0.1 μg/ml). In contrast, rituximab had no toxic effect alone or in combination with TRAIL. Rituximab also had no toxic effect alone or in combination with TRAIL in other erbB-2-expressing or non-erbB-2-expressing cells (data not shown).

When cells were incubated with varying concentrations of trastuzumab and then treated with TRAIL (1 μg/ml), enhancement of TRAIL-mediated toxicity could be seen at 10-fold lower doses of trastuzumab (data not shown). When the sum of each treatment was compared with the combination, the combined treatment was statistically greater than the sum of the individual treatments (P < 0.04 and P < 0.01 by unpaired and paired two-tailed t tests, respectively). Thus the enhancement of TRAIL-mediated toxicity by trastuzumab in the erbB-2-overexpressing cell lines was more than additive. By both fractional inhibition analysis (47) and dose-effect analysis (48) , the effects of the combination were synergistic compared with either reagent alone.

TRAIL has been demonstrated to induce apoptosis by a caspase-dependent mechanism (3 , 7) . To determine whether the increased inhibition by the combination of trastuzumab plus TRAIL seen in the MTS assays was due to the induction of apoptosis, the effect of caspase inhibition was measured (Fig. 2A) ⇓ . SKBr-3 cells were incubated with or without trastuzumab, and 1 h before the addition of TRAIL, the caspase inhibitor ZVAD-fmk was added to the cultures. ZVAD-fmk completely inhibited the toxicity of TRAIL either alone or in combination with trastuzumab. Caspase inhibition only slightly blocked the toxicity of trastuzumab alone. The partial inhibition of trastuzumab-induced toxicity results because there was no inhibitor present for most of the time that trastuzumab was present. When ZVAD-fmk is added at the same time as the trastuzumab, an almost complete inhibition of trastuzumab-induced toxicity is observed (data not shown). This is consistent with trastuzumab inducing apoptosis via a caspase-dependent mechanism. In previous work (18) , we have shown that chemotherapy could induce TRAIL-mediated apoptosis by augmenting activation of caspase-3. By itself, TRAIL weakly activates caspase-3, but trastuzumab did not increase this activation (data not shown). This is consistent with the demonstrated existence of both caspase-3-dependent and -independent mechanisms of apoptosis (52) .

To quantitate the induction of apoptosis in the cultures, the fraction of cells with sub-G₀-G₁ DNA content was measured by flow cytometry (Fig. 2B) ⇓ . Whereas trastuzumab or TRAIL alone each significantly increased the fraction of apoptotic cells compared with control cells, the combination of trastuzumab plus TRAIL results in a synergistic increase in the fraction of apoptotic cells. In these experiments, TRAIL induced greater apoptosis than trastuzumab, whereas Fig. 1A ⇓ shows that the trastuzumab treatment resulted in greater growth inhibition than TRAIL. The measurement of apoptosis by flow cytometry (Fig. 2B) ⇓ shows only those cells undergoing apoptosis at the time of the measurement, whereas the MTS assay shows the cumulative effect of ongoing apoptosis over the entire period. Trastuzumab was present in the medium for 72 h before TRAIL treatment. Thus, the greater growth inhibition induced by trastuzumab as compared with TRAIL observed in Fig. 1A ⇓ likely reflects ongoing apoptosis over the entire 96-h incubation. Together, the results of caspase inhibition and flow cytometry demonstrate that the enhancement of TRAIL-mediated toxicity by trastuzumab is due to the induction of apoptosis.

We evaluated whether the enhancement of TRAIL-mediated apoptosis by trastuzumab preincubation resulted from changes in the expression of proteins known to modulate TRAIL sensitivity (Fig. 3A) ⇓ . There was no change in protein expression of TRAIL-Rs, IAP-1, IAP-2, or FLIP. In addition, on trastuzumab treatment, there was no change in the surface expression of the death-inducing receptors (TRAIL-R1 and TRAIL-R2; Fig. 3B ⇓ ). There was also no change in total TRAIL binding, indicating that there is also no change in the surface levels of the decoy receptors (TRAIL-R3 and TRAIL-R4). Thus, alterations in expression of different components of the TRAIL pathway do not account for the increase in TRAIL-mediated apoptosis.

Trastuzumab Down-Regulates the erbB-2 Receptor.

Incubation of cells overexpressing the erbB-2 receptor with trastuzumab for 96 h resulted in down-regulation of the erbB-2 protein and a decrease in the amount of phosphorylated erbB-2 receptor (Fig. 4) ⇓ . Short incubations (5 min to 8 h) of cells with trastuzumab resulted in increased activation of the erbB-2 protein (as measured by phosphorylation of the receptor) and no decrease in protein levels (data not shown). Longer incubations (>24 h) resulted in a progressive decrease and inactivation of the erbB-2 protein with maximum down-regulation at 72–96 h. This time course of erbB-2 down-regulation is consistent with the long preincubation necessary for trastuzumab to enhance TRAIL toxicity.

To assess whether down-regulation of the erbB-2 receptor is sufficient to enhance TRAIL-mediated apoptosis, cells were incubated with antisense ODNs. Antisense ODNs have previously been shown to down-regulate the erbB-2 receptor in breast cell lines (53 , 54) . The SKBr-3 cell line was transfected with antisense or sense ODNs or mock-transfected, and then cells were treated with TRAIL (Fig. 5A) ⇓ . The antisense ODNs induced significantly more toxicity (15–20%) than mock-transfection (P < 0.02) or sense ODNs (P < 0.02). A slight increase in toxicity by sense ODNs alone (5–10%) was observed compared with the mock transfection, but this increase was not statistically significant (P = 0.15). There was a significant increase in TRAIL-mediated apoptosis in cells transfected with antisense ODNs (30–40%; P < 0.001) but not in cells transfected with sense ODNs (P = 0.25) compared with mock-transfected cells (Fig. 5A) ⇓ . By fractional inhibition analysis, the interaction between antisense ODNs and TRAIL was synergistic. Lysates obtained from cells transfected under the same conditions showed a significant down-regulation of the erbB-2 receptor in cells transfected with antisense ODNs, but not in mock-transfected cells or in cells transfected with sense ODNs (Fig. 5B) ⇓ . Thus, down-regulation of the erbB-2 receptor by two distinct mechanisms enhanced TRAIL-mediated apoptosis.

Down-Regulation of the erbB-2 Receptor Inhibits Akt Kinase Activity but not MAPK Activity.

erbB-2 overexpression results in activation of different downstream pathways such as the Akt kinase and MAPK pathways, which leads to cell proliferation and cell survival (39 , 41) . To assess whether down-regulation of the erbB-2 receptor affects these downstream pathways, cells overexpressing the erbB-2 receptor (SKBr-3 and MDA-MB-453) or cells with low erbB-2 expression (MDA-MB-468) were treated with trastuzumab, and then the activities of the Akt kinase and MAPK pathways were measured. Both the Akt kinase and MAPK are activated in response to growth factor receptor activation by phosphorylation of specific serine and threonine residues (55 , 56) . The activation of Akt kinase and MAPK can be measured using phospho-specific antibodies (55 , 56) . Trastuzumab treatment resulted in a significant reduction of Akt kinase activity as measured by a phospho-Akt kinase-specific antibody, whereas there was no decrease in total Akt protein levels (Fig. 6A) ⇓ . Therefore, the change in phosphorylation represents a decrease in activity and not down-regulation of the Akt protein. Trastuzumab induced a similar decrease in Akt kinase activity in the SKOv-3 cell line (data not shown). In contrast, there was no change in the activity of MAPK as measured by a phospho-MAPK-specific antibody (Fig. 6A) ⇓ . In cells expressing low levels of the erbB-2 receptor (MDA-MB-468 and MDA-MB-231), trastuzumab did not induce changes in either Akt kinase or MAPK activity. Interestingly, the MDA-MB-231 cell line, which is very sensitive to TRAIL-mediated apoptosis, has very low levels of phospho-Akt. These pathways were also assessed in cells treated with antisense or sense ODNs (Fig. 6B) ⇓ . As observed with trastuzumab, there was a significant reduction in Akt kinase activation and no change in MAPK activation in cells transfected with antisense ODNs. There was no change in the activity of either of these kinases in mock-transfected or sense ODN-transfected cells.

Decrease in Akt Kinase Activity Enhances TRAIL Toxicity.

The data above suggest that the enhancement of TRAIL-mediated apoptosis by down-regulation of the erbB-2 receptor (by either trastuzumab or antisense ODNs) results from decreased Akt kinase activity. The activation of Akt kinase by growth factor receptors is mediated by PI3k activity (56) . To assess whether the decrease of Akt kinase activity is responsible for the observed increase in TRAIL-mediated apoptosis described above, cells overexpressing erbB-2 receptor (SKBr-3) or expressing a low amount of erbB-2 receptor (MDA-MB-468) were treated with a PI3k inhibitor (LY294002) before the addition of TRAIL. In both cell lines, LY294002 was toxic by itself. There was a significant enhancement in TRAIL-mediated toxicity in both cell lines by pretreatment with LY294002 (Fig. 7A) ⇓ . Immunoblotting with phospho-Akt demonstrated that LY294002 pretreatment resulted in decreased Akt activity (Fig. 7B) ⇓ .

To confirm that the Akt kinase is the critical target affected by erbB-2 down-regulation, a constitutively active (membrane-targeted) Akt kinase was introduced into the MDA-MB-453 cell line, which overexpresses the erbB-2 receptor. Vector control or the activated Akt kinase clones were treated with trastuzumab alone, TRAIL alone, or the combination, and toxicity was assayed as described in the Fig. 1 ⇓ legend. Treatment of the vector clone with the combination of trastuzumab + TRAIL resulted in 35–40% inhibition compared with the control cells. This was similar to the effect in the parental MDA-MB-453 cells, as seen in Fig. 1A ⇓ . Expression of the constitutively active Akt kinase in two independent clones resulted in a significant reduction of toxicity with each reagent alone or the combination (Fig. 8A) ⇓ . There was almost complete abrogation of the synergism between trastuzumab and TRAIL in the clones expressing the active Akt kinase. Lysates from these clones were immunoblotted using erbB-2 and phospho-Akt antibodies (Fig. 8B) ⇓ . The vector and the active Akt kinase clones all showed down-regulation of the erbB-2 receptor. Reduced levels of phospho-Akt after trastuzumab incubation were observed. However, because the transfected constitutively active Akt kinase is a small fraction of the total Akt kinase and is myristylated, the phosphorylation status does not reflect its activity accurately. The exogenous constitutively active Akt kinase was immunoprecipitated via its HA epitope, and its activity was measured by an in vitro kinase assay using a specific phospho-gsk-3 antibody (Fig. 8C) ⇓ . No suppression of the exogenous Akt kinase activity by trastuzumab treatment was seen.

DISCUSSION

In the work presented here, we demonstrate that down-regulation of the erbB-2 receptor by either trastuzumab or antisense ODNs enhances TRAIL-mediated apoptosis in breast and ovarian cancer cells that overexpress erbB-2 (Figs. 1 ⇓ 2 ⇓ 3 ⇓ 4 ⇓ 5 ⇓ ). Cross-linking of the erbB-2 receptor by trastuzumab activates the receptor (57) . Recent reports have demonstrated that down-regulation of erbB-2 by anti-erbB-2 antibodies results in recruitment of the cbl protein, ubiquitination, and proteasome-dependent degradation of the activated receptors (46) . Antisense ODNs, on the other hand, block synthesis of the erbB-2 protein (53 , 58) . Thus, the enhancement of TRAIL-mediated apoptosis by erbB-2 down-regulation is independent of the mechanism through which down-regulation is achieved.

Trastuzumab specifically enhances TRAIL-mediated apoptosis in erbB-2-overexpressing cells. There was no enhancement of TRAIL-mediated apoptosis in all of the cell lines tested that express low levels of the erbB-2 receptor (Fig. 1A) ⇓ . The selectivity of this effect results in targeting TRAIL only to those cells that overexpress erbB-2 receptor, thus potentially avoiding toxicity to normal cells that express low levels of erbB-2. Furthermore, other ways to down-regulate the receptor, such as antisense ODNs, might also be expected to selectively sensitize only the cancer cells that overexpress erbB-2.

Down-regulation of the erbB-2 receptor did not alter total or surface expression of TRAIL-Rs or other proteins that have been shown to modulate TRAIL-mediated apoptosis (Fig. 3A) ⇓ . The mechanism underlying the enhancement of TRAIL-mediated apoptosis is inhibition of the prosurvival Akt kinase pathway. Akt/protein kinase B, the cellular homologue of the viral oncoprotein v-Akt, is a serine/threonine kinase that has been identified as an important component of prosurvival signaling pathways (59) . Once activated, Akt exerts antiapoptotic effects through phosphorylation of substrates that directly regulate the apoptotic machinery such as Bad and caspase-9 (60) . In addition, Akt also phosphorylates substrates that indirectly inhibit apoptosis such as the human telomerase reverse transcriptase subunit, the forkhead transcription family members, or the IκB kinases (61, 62, 63) . All cell lines overexpressing the erbB-2 receptor, which are resistant to TRAIL-mediated apoptosis, show high levels of Akt kinase activity. Similarly, two cell lines expressing low levels of erbB-2 receptor (MDA-MB-468 and SW-626) also show high levels of Akt activity and resistance to TRAIL-mediated apoptosis. In contrast, the MDA-MB-231 cell line, which expresses low levels of erbB-2 receptor, has low Akt activity and is extremely sensitive to TRAIL-mediated apoptosis.

In all of the erbB-2-overexpressing cell lines tested, there was a decrease in activation of the Akt kinase when the erbB-2 receptor was down-regulated by trastuzumab or antisense ODNs (Fig. 6) ⇓ . Although erbB-2 down-regulation was seen in cells with low erbB-2 expression, there was no change in Akt kinase activation or biological effect by either agent alone or the combination. erbB-2 down-regulation did not decrease MAPK activation. Both Akt kinase and MAPK are activated in response to growth factor receptor activation and up-regulated when erbB-2 receptor is overexpressed (39 , 40 , 42) . The reason for the selective decrease in Akt kinase activity is unknown.

The Akt kinase is activated via the PI3k signaling pathway (59) . Similar enhancement in TRAIL-mediated apoptosis was observed when the phosphorylation of Akt kinase was blocked by using an inhibitor of PI3k activity. However, LY294002 treatment resulted in a significant enhancement of TRAIL-mediated apoptosis in erbB-2-overexpressing cells as well as in cells with low erbB-2 expression (Fig. 7) ⇓ . These data suggest that other pathways keep the Akt kinase active in the cell lines with low erbB-2 expression. The expression of a constitutively active form of Akt kinase in erbB-2-overexpressing cell lines almost abrogated the effect of trastuzumab alone, TRAIL alone, or the combination, indicating that Akt kinase is a major determinant in cell survival in erbB-2-overexpressing cell lines (Fig. 8) ⇓ . Therefore, down-regulation of Akt kinase activity will result in an enhancement of TRAIL-mediated apoptosis independent of the underlying mechanisms. Thus, targeting the Akt kinase pathway may enhance TRAIL-mediated apoptosis in many cell types, even those that do not overexpress the erbB-2 receptor.

It is not surprising that inhibition of the antiapoptotic Akt kinase pathway would sensitize cells to death receptor-induced apoptosis. Consistent with this, previously published work has shown that erbB-2 overexpression inhibits TNF-induced apoptosis by activation of Akt kinase and that an anti-erbB-2 antibody enhances TNF-induced apoptosis (64 , 65) . It is also likely that the enhanced efficacy of chemotherapy by trastuzumab is mediated through this Akt kinase-dependent mechanism.

In summary, down-regulation of erbB-2 by trastuzumab enhances TRAIL-mediated apoptosis specifically in erbB-2-overexpressing breast and ovarian cancer cell lines. The mechanism behind the interaction can be explained at least in part by down-regulation of the prosurvival Akt kinase pathway. Trastuzumab has been described to have collateral toxicity in normal tissues (66) . More recent data suggest that TRAIL may be toxic to normal human liver cells (67) . Therefore, combinations of therapy that maintain the efficacy of TRAIL and trastuzumab at lower doses may be important to their clinical usefulness. Our results indicate that the interaction between trastuzumab and TRAIL is still present after reducing the dose of each reagent. The trastuzumab concentrations that sensitized cells to TRAIL-mediated apoptosis were within a range clinically achieved in patients (68) . Therefore, the identification of this synergistic interaction between these two drugs with effects targeted specifically to cancer cells is attractive as a new treatment modality. These data are in vitro experiments with cancer cell lines, and hence the efficacy and specificity of this combination remain to be tested in animal models.