Enhancement of the p27Kip1-mediated antiproliferative effect of trastuzumab (Herceptin) on HER2-overexpressing tumor cells

Authors

  • Radu Marches,

    Corresponding author
    1. Cancer Immunobiology Center and Department of Microbiology, University of Texas Southwestern Medical School, Dallas, TX, USA
    • Cancer Immunobiology Center, University of Texas Southwestern Medical School, 6000 Harry Hines Blvd., Dallas, TX 75390-8576, USA
    Search for more papers by this author
    • Fax: +214-648-1890

  • Jonathan W. Uhr

    1. Cancer Immunobiology Center and Department of Microbiology, University of Texas Southwestern Medical School, Dallas, TX, USA
    Search for more papers by this author

Abstract

The oncogenic activity of the overexpressed HER2 tyrosine kinase receptor requires its localization in the plasma membrane. The antitumor effect of anti-HER2 antibodies (Abs) is mainly dependent on receptor downregulation and comprises p27Kip1-mediated G1 cell cycle arrest. However, one major limitation of anti-HER2 therapy is the reversibility of tumor growth inhibition after discontinuation of treatment caused by the mitogenic signaling associated with cell surface receptor re-expression. We found that the level of p27Kip1 upregulation, inhibition of Cdk2 activity and magnitude of G1 arrest induced by the humanized Ab trastuzumab (Herceptin, HCT) on BT474 and SKBr3 HER2-overexpressing breast cancer cells correlates with the level of cell surface receptor. Thus, continuous exposure of cells to HCT for 72 hr results in downregulation of the cell surface receptor and a concurrent increase in the level of p27Kip1 protein. Discontinuation of Ab exposure after the first 8 hr results in failure to upregulate p27Kip1 and arrest of cell cycle progression. We show that the lysosomotropic amine chloroquine (CQ) augments receptor internalization in HER2-overexpressing cells either pretreated or continuously treated with HCT and leads to an increased and sustained inhibitory effect. The enhanced CQ-dependent loss of functional HER2 from the cell surface resulted in sustained inactivation of the serine/threonine kinase Akt, upregulation of p27Kip1 protein and inhibition of cyclin E/Cdk2 activity. Potentiation of the inhibitory effect of HCT by CQ was directly related to loss of HER2 from the plasma membrane since prevention of Ab-mediated receptor endocytosis by engagement of the receptor with immobilized HCT abrogated the effect of CQ. © 2004 Wiley-Liss, Inc.

The neu (c-ErbB-2, HER2) proto-oncogene, encoding a 185 kDa RTK, is a member of the EGF receptor family, which includes ErbB-1 (HER1, EGFR), ErbB-3 (HER3) and ErbB-4 (HER4).1 Activation of RTKs is dependent on ligand-mediated homodimerization and subsequent autophosphorylation and transphosphorylation of the intracellular kinase domains of the paired receptors. However, HER2 may be activated even in the absence of a direct ligand because it forms heterodimeric complexes with its family members HER1, HER3 and HER42, 3, 4, 5, 6, 7 upon ligand binding to the ectodomain of these receptors. Once recruited into these heterodimeric complexes, HER2 can modulate ligand-mediated receptor internalization to the recycling route and amplify signal transduction.7, 8, 9, 10 The transforming ability of HER2 has been associated with enhanced cell survival and mitogenic signaling pathways. Amplification of HER2 is observed in cancers of the breast,11, 12, 13, 14, 15 ovary,12, 16 lung17 and stomach;18 and its overexpression predicts poor prognosis in several human cancers.11, 16, 19, 20, 21, 22, 23 At the molecular level, HER2 overexpression can deregulate the G1–S transition by downregulating p27Kip1 and stimulating the cyclin E–cdk2 complex.24 The crucial role of p27Kip1 in HER2-mediated transformation is consistent with clinical observations that a decreased p27Kip1 protein level in breast cancer is associated with poor prognosis.25

The key role of HER2 in signaling cell proliferation and transformation made it an attractive therapeutic target in HER2-overexpressing human tumors. MAbs directed against the HER2 ectodomain could specifically inhibit the growth of tumor cells overexpressing HER2.26, 27, 28, 29, 30 The major mechanism of anti-HER2-mediated inhibition of cell transformation is suppression of HER2 expression on the cell surface28, 31, 32, 33, 34 due to Ab-simulated endocytosis of the receptor via coated pits.35, 36 Indeed, retention of nascent HER2 in the endoplasmic reticulum reversed the HER2-mediated transformation associated with cell surface localization.37 One such potent anti-HER2 Ab, 4D5,38 was humanized;39 and the resulting Ab, HCT, is clinically used for the therapy of HER2-overexpressing metastatic breast cancers.40, 41 HCT, like its parental Ab 4D5, markedly downregulates HER2 expression by stimulating receptor endocytosis42 and thereby arrests cells in the G1 phase of the cycle via p27Kip1.24, 43 However, monotherapy with HCT has limited antitumor effects44 and requires sustained exposure to the Ab. Thus, when treatment in vitro or in vivo with either 4D5 or HCT is discontinued, overexpression of HER2, proliferation and tumor growth resume.28, 45

Our objective was to improve the therapeutic effect of HCT. One possible method to increase HCT potency is to prevent receptor recycling by modulating intracellular routing and/or retention of the endocytosed Ab–receptor complex. Here, we show that alkalinization of the endosomal compartment with CQ potentiates the p27Kip1-mediated inhibitory effect of HCT on HER2-overexpressing tumor cells by increasing the extent and duration of the Ab-mediated internalization of the receptor.

Abbreviations

Ab, antibody; CQ, chloroquine; ECL, enhanced chemiluminescence; EGF, epidermal growth factor; EGFR, EGF receptor; HCT, trastuzumab (Herceptin); MAb, monoclonal antibody; MAPK, mitogen-activated protein kinase; MFI, mean fluorescence intensity; NDF, Neu differentiation factor; PI, propidium iodide; PI-3K, phosphatidylinositol-3-kinase; RTX, Rituxan; TBS-T, 20 mM TRIS-HCl (pH 7.5), 150 mM NaCl and 0.05% Tween-20.

MATERIAL AND METHODS

Cells

The HER2-overexpressing human breast cancer cell lines BT474 and SKBr3 and the human breast MDA-MB-231 cell line, which expresses low levels of HER2, were purchased from the ATCC (Rockville, MD). Both cell lines were grown in DMEM (GIBCO, Grand Island, NY) supplemented with 10% FCS and 2 mM L-glutamine.

Antibodies

Her81 is a mouse IgG1 MAb generated against an extracellular epitope of HER2 that binds the receptor with a KD of 1.9 × 10−9 M. Her81 and the isotype-matched control G28-5 (anti-CD40) were used for immunofluorescent staining. The clinically marketed recombinant humanized anti-HER2 Ab HCT (Genentech, South San Francisco, CA) and the isotype-matched control RTX (humanized MAb to human CD20,Genentech) were used for in vitro treatment of BT474 cells. For immunoblot analysis and immunoprecipitation, the following Abs were used: rabbit anti-ErbB2/neu (#2242), -Akt (#9272) and -phospho(Ser473)-Akt (#9271; all from New England Biolabs, Beverly, MA); mouse anti-phosphotyrosine PY20 (#610000; BD Transduction Laboratories, San Diego, CA); mouse anti-p21WAF1 (#15091A; Pharmingen, San Diego, CA); and mouse anti-actin (#8432) and rabbit anti-p27Kip1 (#528), -cyclin D1 (#718), -Cdk2 (#163) and -cyclin E (#198; all from Santa Cruz Biotechnology, Santa Cruz, CA).

[3H]-Thymidine incorporation assay

Six replicate cell samples were plated into 96-well plates at 2 × 104 cells/well in culture medium. Plated cells were left to adhere for 8 hr and then incubated for 72 hr at 37°C in the presence or absence of 20 μg/ml HCT or RTX and 5 μM CQ. For the last 8 hr, 1 μCi [3H]-thymidine was added to each well. Cells were harvested on fiberglass filters in 2 steps, before and after trypsinization, and the filter-bound radioactivity was counted with a scintillation counter. In some experiments, HCT or control Ab was coated onto the surface of the microtiter plates. To this end, Abs diluted in PBS were added to wells and the plates wrapped in plastic film and incubated overnight at 37°C. Cells were plated after the unbound Ab was removed by washing the wells 3 times with PBS and finally with medium.

Flow cytometry

Cells were harvested by trypsinization and counted prior to staining. For cell cycle analysis, 106 cells were washed with PBS, fixed in 70% ice-cold ethanol and stored at −20°C overnight. Fixed cells were washed twice with PBS and stained overnight at 4°C in the dark with 40 μg/ml PI containing 100 μg/ml RNase A. Cells were filtered through a 105 μm-pore-size nylon mesh (Small Parts, Miami Lakes, FL), and a total of 10,000 events were collected on a FACScan (Becton Dickinson, Mansfield, MA). DNA histograms were analyzed using Paint-A-Gate software (Becton Dickinson). For the detection of cell surface levels of HER2, 5 × 105 cells were incubated for 30 min on ice with 1 μg FITC-conjugated Her81 or isotype-matched control. After incubation, cells were washed twice with PBS, resuspended in 1 ml 1% paraformaldehyde in PBS and analyzed on a FACScan.

Radioiodination

Her81 (70 μg) was iodinated with 0.5 mCi of carrier-free Na125I (Amersham, Piscataway, NJ) by incubation for 20 min on ice in a 10 × 75 mm borosilicate tube coated with 10 μg of Iodo-Gen (Pierce, Rockford, IL). Ab-bound iodine was separated from free 125I by spinning the solution on a MicroSpin G-25 column (Amersham) for 2 min at 735g. Collected and radiolabeled Ab was brought to 1 ml with PBS containing 1 mg/ml BSA. Specific activity of 125I-labeled Her81, measured in a gamma counter, ranged between 4 and 5 × 103 cpm/ng Ab.

Competitive binding assay

Competition reactions (total volume 0.2 ml HBSS with 5% FBS) were conducted by incubating 1 × 106 BT474 cells with 125I-Her81 (at 75% of the saturating concentration) in the presence of increasing amounts of unlabeled Abs for 1 hr on ice. Nonspecific binding was determined in the presence of a 50-fold molar excess of unlabeled Her81. Cell-bound Ab was separated from unbound Ab by layering triplicate aliquots (60 μl) of the binding reaction on top of an oil mixture (340 μl dioctylphthalate/dibutylphthalate at 1.1/1 ratio; Sigma, St. Louis, MO) in a 0.4 ml microcentrifuge tube (Fisher, Pittsburgh, PA) and pelleting the cells through the oil by centrifugation at 600g for 20 min. After 2 hr storage at −80°C, the tip of each tube was cut from the body and applied to a gamma counter to measure bound radioactivity (the tip). Specific binding was calculated by subtracting nonspecific binding from total binding.

Microscopy

To examine the effect of combination treatment on cell morphology, BT474 cells were cultured in a 96-well plate for 72 hr with experimental concentrations of Abs in the presence or absence of CQ and photographed using an inverted microscope (Zeiss, Oberkochen, Germany; Axiovert S100TV). For immunofluorescent staining, cells were grown and treated on 8-well Lab-Tek chamber slides (Nalge Nunc, Rochester, NY). At the end of incubation, plastic chambers were removed and slides were rinsed with PBS, fixed with 3% paraformaldehyde for 15 min at room temperature, washed twice with PBS and permeabilized for 10 min with 0.1% Triton X-100 in PBS. After blocking with 1% albumin for 30 min, cells were incubated with FITC–Her81 for 1 hr at room temperature. Slides were washed twice with PBS, rinsed with distilled water, air-dried and mounted with Vectashield mounting medium (Vector, Burlingame, CA). Slides were examined with a fluorescence microscope (Axiophot, Zeiss).

Western blot and kinase assay

Cell monolayers grown in 100 mm dishes were washed twice with PBS, scraped into ice-cold lysis buffer [50 mM TRIS-HCl (pH 8.0), 120 mM NaCl, 0.5% NP40, 100 μM Na3VO4, 10 μg/ml aprotinin, 10 μg/ml leupeptin, 1 mM PMSF] and incubated for 20 min on ice. Lysates were cleared by 10 min centrifugation at 15,000g at 4°C. Protein concentration in cell lysates was measured using the bicinchoninic acid protein detection kit (Pierce). For immunoblot analysis, equal amounts of total protein were resolved by SDS-PAGE and transferred to PVDF membranes. Membranes were blocked with TBS-T containing 5% BSA and incubated overnight at 4°C with primary Ab in TBS-T. Blots were washed in TBS-T and incubated with antimouse or antirabbit IgG coupled to horseradish peroxidase (Amersham), and immunoreactive bands were visualized using an ECL system (Amersham).

Kinase assays were performed with lysates prepared as described above. Equal amounts (400 μg) of lysate were incubated overnight at 4°C with 20 μl of protein A-agarose precoated with 2 μg of either anti-Cdk2 or anti-cyclin E. Immunoprecipitates were washed 3 times with ice-cold lysis buffer and twice with ice-cold kinase buffer [25 mM TRIS-HCl (pH 7.5), 5 mM β-glycerophosphate, 2 mM dithiothreitol, 0.1 mM Na3VO4, 10 mM MgCl2] before the kinase assay was performed. The reaction was started by incubating each immunoprecipitate in 30 μl kinase buffer containing 2 μg histone H1 (Roche, Indianapolis, IN), 2 μCi [γ-32P] ATP and 50 μM cold ATP for 30 min at 30°C. Kinase reactions were stopped by addition of 30 μl of 2× Laemmli sample buffer. Samples were then boiled for 5 min and the reaction products separated by 10% SDS-PAGE and autoradiographed.

RESULTS

Effect of treatment with a combination of CQ and HCT on the loss of HER2 from the cell surface

To monitor the level of membrane HER2 in BT474 and SKBr3 cells after treatment with HCT, we used FITC-labeled Her81, an Ab non-cross-reactive with HCT and specific for the extracellular domain of HER2. The distinct specificities of HCT and Her81 were determined by competition experiments using radiolabeled Her81 and increasing concentrations of unlabeled HCT, RTX (isotype-matched control for HCT), G28-5 (isotype-matched control for Her81) or Her81 (self-inhibition). As indicated in Figure 1, only unlabeled Her81 was able to compete with radiolabeled Ab for binding to HER2, whereas HCT or the isotype-matched control showed no inhibition.

Figure 1.

Her81 and HCT recognize different epitopes on the extracellular domain of HER2. Competition experiments were conducted by incubating 1 × 106 BT474 cells with 125I-labeled HER81 at 75% of the saturating concentration (detected from the saturation curve, see inset) in the presence of increasing concentrations of unlabeled Abs in 200 μl of medium for 1 hr on ice. Bound radioactivity was determined as described in Material and Methods and expressed as a percentage in the absence of unlabeled competitor.

Following treatment with HCT, the level of cell surface HER2, as detected by flow cytometry, was reduced. Thus, after 72 hr of incubation in the presence of HCT, levels of surface HER2 on BT474 and SKBr3 cells were reduced to 62.3 ± 5.8% and 77.4 ± 8.2%, respectively, compared to untreated control, as judged by MFI (Fig. 2). The combination of HCT with a nontoxic concentration of CQ resulted in additional diminishment of cell surface receptors to 49.3 ± 7.9 % (BT474) and 56.5 ± 8.2 % (SKBr3) (Fig. 2). This decrease corresponds to reductions of more than 90,000 (BT474) and 150,000 (SKBr3) receptors/cell since both cell lines express on their surface approximately 7 × 105 HER2 receptors/cell.46 When cells were released in HCT-free, CQ-free medium after the first 8 hr of incubation, when HER2 levels were reduced to 72.9% (BT474) and 79.2% (SKBr3), they re-expressed plasma membrane HER2 by 72 hr at almost the same level as untreated controls. However, when such cells were grown in medium containing CQ, cell surface re-expression of HER2 in both cell lines was limited to <80%, i.e., more than 125,000 receptors/cell fewer than in cells grown in CQ-free medium (Fig. 2). We next tested whether the CQ-mediated decrease of cell surface expression of HER2 resulted from increased intracellular accumulation. Immunofluorescence analysis indicated that in control cells HER2 was concentrated at the plasma membrane, regardless of the presence of CQ (Fig. 3). In contrast, treatment with HCT resulted in loss of HER2 from the plasma membrane and accumulation within cytoplasmic vesicles, which was increased by cotreatment with CQ (Fig. 3). The difference was even more evident in cells pretreated with HCT. Thus, in the absence of CQ, the receptor recycles back to the plasma membrane, whereas in the presence of CQ the receptor remains primarily intracellular (Fig. 3).

Figure 2.

Effect of combination treatment on surface expression of HER2. Cells were incubated in 6-well plates for 72 hr in the absence or presence of 20 μg/ml HCT or control Ab with or without 5 μM CQ. Cells were pulsed with HCT continuously (72 hr) or for the first 8 hr (8 hr); then, medium containing HCT was aspirated, and monolayers were washed once with culture medium and replenished with the same volume of HCT-free culture medium supplemented or not with 5 μM CQ. After incubation, cells were harvested by trypsinization and 5 × 105 cells of each treatment were stained with 1 μg FITC-conjugated Her81 and analyzed by flow cytometry. Bars represent percentage of MFI relative to untreated controls ± SD for 3 independent experiments. MFI for each sample was corrected by subtracting background fluorescence given by binding of FITC-labeled isotype-matched control, specific for CD40 (G28-5).

Figure 3.

Localization of HER2 after incubation with HCT in the absence or presence of CQ. Cell monolayers were incubated on chamber slides with HCT or control Ab in the absence or presence of CQ. After removal of plastic chambers, Lab-Tek slides were rinsed with PBS, fixed, permeabilized and stained with FITC–Her81. Slides were examined under a fluorescent microscope. Magnification ×100.

Intensity and duration of the inhibitory effect of HCT on DNA synthesis and cell cycle progression are increased by CQ

We next analyzed the antiproliferative effect of the enhanced receptor downregulation. Inhibition of the proliferation of BT474 and SKBr3 cells on monolayers by treatment with HCT alone required its continuous presence (Fig. 4). The membrane HER2 re-expression after HCT removal resulted in resumption of cell growth (Fig. 4). Conversely, the CQ-mediated delay of receptor re-expression after HCT removal resulted in persistence of the inhibitory effect (Fig. 4). Mock treatment of cells with isotype-matched RTX in the absence or presence of CQ had no effect. Growth of MDA-MB-231 cells, showing little HER2 expression, was not inhibited by HCT alone or in combination with CQ (Fig. 4). Taken together, these results indicate that the effect of CQ is strictly related to the Ab-mediated endocytosis of HER2. The extent and nature of the inhibitory effect were determined by analysis of cell cycle status (Fig. 5). Continuous treatment with HCT resulted in accumulation of cells in the G1 phase and a concurrent reduction of the S phase. Re-entry of cells into S phase at 64 hr after HCT removal corresponded to the increase in [3H]-thymidine incorporation. The reversibility of HCT-mediated inhibition with a very low level of apoptosis indicates that the inhibitory effect is mainly cytostatic rather than cytotoxic. This conclusion is supported by the low level of DNA fragmentation, as judged by the proportion of sub-G1 phase cells (Fig. 5). The accumulation of cells in G1 after HCT removal was higher in the presence of CQ in both BT474 and SKBr3 cells compared to cells released in CQ-free medium. Cotreatment with HCT and CQ also partially changed the nature of the inhibitory effect. Thus, addition of CQ increased the percentage of cells with DNA fragmentation cultured for 72 hr and pulsed continuously with HCT (Fig. 5).

Figure 4.

Effect of combination treatment on DNA synthesis. BT474, SKBr3 or MDA-MB-231 (negative control) cells were incubated in 96-well flat-bottomed plates for 72 hr as indicated above. [3H]-Thymidine (1 μCi/well) was added 8 hr prior to harvest. Average percentages of residual [3H]-thymidine incorporation were calculated from the 6 replicate samples of 3 independent experiments. Data are expressed as average percentage of incorporation in control untreated cultures ± SD.

Figure 5.

Cell cycle status. Cells were grown in 100 mm dishes and incubated for 72 hr with medium, 20 μg/ml RTX or HCT in the absence or presence of 5 μM CQ. The last 2 samples were exposed for the first 8 hr to 20 μg/ml HCT; then, the Ab was removed, and cell monolayers were washed and incubated for the remaining 64 hr in HCT-free culture medium in the absence or presence of 5 μM CQ. Cells (106), harvested by trypsinization, were stained with PI and analyzed by flow cytometry. Histograms are representative of 3 experiments.

Biochemical changes associated with enhancement of receptor downregulation

Several lines of evidence have indicated that p27Kip1 is one of the key mediators of G1 arrest in HER2-overexpressing tumor cells treated with HCT,43 4D5 (the parental Ab of HCT)24, 47 or tyrosine kinase inhibitors.48 Indeed, continuous incubation with HCT alone led to upregulation of p27Kip1 in both cell lines. The upregulation of p27Kip1 protein induced by continuous exposure to HCT alone was enhanced by an additional 30% (BT474) and 36% (SKBr3) when CQ was combined with HCT (Fig. 6). Furthermore, cells that remained in G1 phase when released from HCT treatment in CQ-containing medium showed markedly increased levels of p27Kip1 compared to cells released in CQ-free medium (123.9% vs. 78.7% in BT474 and 155.4% vs. 96.4% in SKBr3) (Fig. 6). CQ, alone or in the presence of control Ab, had no effect. Regardless of treatment, no change in cyclin D1 level was found (Fig. 6).

Figure 6.

Effect of combination treatment on protein levels of HER2 and some of its major downstream signaling targets. Protein from lysates (100 μg) was resolved on 7.5% (HER2), 10% (Akt, P-Akt, actin) or 12% (cyclin D1, p27, p21) SDS-PAGE and analyzed by Western blotting.

The PI-3K–Akt pathway, but not the MAPK pathway, controls the level of p27Kip1 in HER2-overexpressing cells.49 PI-3K-mediated activation of the serine/threonine kinase Akt is also implicated in the resistance to apoptosis of HER2-overexpressing cells. Akt kinase is activated via PI-3K by phosphorylation of Thr308 and Ser.47350 Immunoblotting using anti-phospho Ser473–Akt indicates that both cell lines contain Akt in an active form (Fig. 6), as previously reported.51 Akt inactivation correlated with the extent of HCT-mediated loss of HER2 from the plasma membrane. Thus, receptor downregulation induced by addition of CQ resulted in a decrease of residual Akt phosphorylation from 70% to 40.2% in BT474 cells and from 34.3% to 17.7% in SKBr3 cells continuously incubated with HCT. Addition of CQ also prevented the increase in Akt activity in cells pretreated with HCT. This was reflected by the marked suppression of Akt phosphorylation from 125% to 65.8% in BT474 cells and from 125% to 70.2% in SKBr3 cells. The decline in Akt activity was not due to a change in the total Akt protein level (Fig. 6). Notably, levels of Akt inactivation mirrored levels of p27Kip1 upregulation in the different treatments, thus confirming its key role in cell proliferation. The inactivation of Akt might also be partly responsible for the sought increase in apoptosis at the expense of G1 arrest during the combination treatment, although the cells remained considerably resistant to apoptosis. The constant level of p21Waf1 throughout the treatment (Fig. 6) might explain the preservation of viability, as previously suggested.52 Continuous incubation of BT474 cells for 72 hr with HCT in the presence of CQ yielded a major fragment, with an electrophoretic mobility of approximately 160 kDa in addition to the 185 kDa HER2 band (Fig. 6). Incubation with HCT alone, CQ or combinations of RTX and CQ did not generate the band with the lower electrophoretic mobility. The 160 kDa comigrating band was not phosphorylated and was not the result of partial proteosome-mediated degradation of mature, plasma membrane–localized HER2 since coincubation with 1 μM MG132, the irreversible proteosome inhibitor, failed to inhibit or attenuate its level (data not shown). Therefore, we assume that the lower m.w. and the lack of phosphotyrosine content could represent an incompletely glycosylated nascent HER2 protein.

p27Kip1 is a binding partner of the cyclin E–Cdk2 complex that is capable of inhibiting its catalytic activity.53 Therefore, we next investigated the kinase activity of Cdk2 and its cyclin partner, cyclin E. Cdk2 and cyclin E were immunoprecipitated from cell lysates and analyzed for catalytic activity in vitro using histone H1 as a substrate. While high levels of Cdk2-associated kinase activity were present in control as well as in untreated cells, this activity was additionally decreased in immunoprecipitates from combination treatment (Fig. 7). The inactivity of Cdk2–cyclin E complexes is consistent with the observed increase in p27Kip1 levels. Collectively, these data indicate that the potentiation of G1 arrest induced by HCT in HER2-overexpressing cells is mediated by the sustained upregulation of p27Kip1 and inactivation of Cdk2-associated kinase activity.

Figure 7.

In vitro kinase assays of Cdk2 and cyclin E. Cdk2 and cyclin E were immunoprecipitated from whole-cell lysates (400 μg protein/sample). One-half of the immunoprecipitate was used to determine the relative kinase activity of Cdk2 or cyclin E using histone H1 as a substrate in the presence of [γ-32P] ATP (top). The other half of the immunoprecipitate was used to detect Cdk2 or cyclin E protein by Western blotting (WB) (bottom). Graphs represent kinase activity of each sample relative to that of the untreated sample, normalized to the amount of Cdk2 or cyclin E protein in the immunoprecipitate. Levels of radioactively phosphorylated histone H1 and of the immunoprecipitated protein were quantified by densitometry.

Synergism between HCT and CQ is related to modulation of cell surface expression of HER2

We next asked if the effect of CQ is related to modulation of cell surface HER2 expression rather than representing a synergism between 2 downstream signaling pathways, one triggered by CQ and the other by HCT. To this end, we cultured BT474 cells in the presence of equal amounts of soluble or plastic-immobilized HCT in medium with or without CQ and analyzed the effect on cell proliferation. In contrast to soluble HCT, immobilized HCT, which binds the receptor but prevents its endocytosis, not only abrogates the inhibitory effect of HCT regardless of the presence of CQ but stimulates cell growth, as judged by both the higher level of [3H]-thymidine incorporation (Fig. 8a) and the extent of confluence of the cell monolayers (Fig. 8b). This indicates that HER2 removal from the plasma membrane and intracellular “trapping” of the receptor are the basis of the enhanced and prolonged antitumor effect of the HCT–CQ combination.

Figure 8.

Effect of CQ on BT474 cells exposed to soluble vs. plastic-bound HCT or RTX. Cells (2 × 105/ml) were cultured in 96-well plates for 72 hr in (a) control medium, (b) 5 μM CQ, (c) 20 μg/ml HCT, (d) plastic-bound HCT, (e) 20 μg/ml HCT + 5 μM CQ, (f) plastic-bound HCT + 5 μM CQ, (g) 20 μg/ml RTX + 5 μM CQ and (h) plastic-bound RTX + 5 μM CQ and pulsed with 1 μCi/well [3H]-thymidine for the last 8 hr. (a) Average percentages of residual [3H]-thymidine incorporation from the 6 replicate samples of 3 independent experiments. Data are expressed as the average percent of incorporation in control untreated cultures ± SD. (b) Growth pattern (morphology) of cells before harvest. Cells were photographed on an inverted microscope at ×10.

DISCUSSION

The oncogenic activity of the constitutively activated HER2 receptor under conditions of overexpression is associated with its localization in the plasma membrane54 and related mainly to HER2/neu gene amplification but also to a relatively slow rate of endocytosis.55 Suppression of cell surface expression of HER2 represents a modality to overcome its mitogenic signaling that has been successfully exploited by the use of specific Abs against HER2 on HER2-overexpressing cells.31, 34 In this respect, the stronger inhibitory activity achieved by targeting distinct epitopes of HER2 with cocktails of Abs reflects a more efficient receptor downregulation.30, 56, 57 HCT, which is used for the treatment of metastatic breast cancer overexpressing HER2, has the capacity to stimulate receptor downregulation42 and induce a cytostatic effect defined by a block in cell cycle progression mediated by p27Kip1.43 However, when treatment is discontinued, overexpression of HER2, proliferation and tumor growth resume.45 Therefore, the extent and duration of the Ab-induced downregulation of the surface receptors are critical for their inactivation and the outcome of treatment. In the present study, we used the lysosomotropic amine CQ, which increases the pH of endosomal compartments, to prevent uncoupling of the receptors from Ab in the sorting endosomes and their recycling to the cell surface. Indeed, alkalinization of the endosomal sorting compartments with monensin has been successfully used to increase TGF-β- and NDF-mediated downregulation of ErbB-1 and ErbB-3, respectively.9, 58 Our data demonstrate that in vitro treatment with HCT in the presence of CQ induces more efficient and persistent removal of the receptor from the cell surface due to intracellular trapping and results in potentiation of the inhibitory effect of HCT in both intensity and duration. The increased overall inhibitory effect persisted for almost 3 days after HCT removal and consisted mainly in G1 arrest. The magnitude of the inhibitory effect was closely related to the enhancement of Akt inactivation and upregulation of its downstream target, p27Kip1. Active Akt (as a result of PI3K-mediated phosphorylation at Thr308 and Ser473) promotes survival, tumorigenesis59 and cell proliferation.49 We found that combination treatment led to decreased Ser473 phosphorylation of Akt, with no change in protein level. This might account, at least in part, for the increased apoptosis at the expense of G1 arrest. Interestingly, quinazoline tyrosine kinase inhibitors were also reported to suppress cell growth and prevent Akt phosphorylation in EGFR- or HER2-overexpressing tumor cells, leading to p27Kip1-mediated G1 arrest.48, 49, 60 The p27Kip1 protein level of cells pretreated or continuously treated with HCT was considerably higher in the presence of CQ than in its absence, which correlated with cyclin E–Cdk2 inactivation and G1 arrest. However, there was considerable resistance to apoptosis even after treatment with the HCT–CQ combination. The constitutively elevated levels of p21Waf1 cyclin-dependent kinase inhibitor in HER2-overexpressing cells could explain this resistance, as previously suggested.52 This is not unique for HER2-overexpressing cells as a considerable number of independent observations point to the antiapoptotic role of p21Waf1 in different cell types.61, 62, 63, 64, 65, 66 Although the magnitude of HCT potentiation by CQ is comparable in both cell lines, their sensitivities to HCT are different. This variability between cellular responses to anti-HER2 treatment has been previously reported24, 67 and may reflect cooperation with other growth factor receptors, such as EGFR, in the activation of additional intracellular pathways.68 Indeed, although both cell lines express similar levels of HER2, SKBr3 cells express 3-fold more EGFR than BT474 cells (2.2 × 105vs. 7 × 104 receptors/cell).46 The high levels of EGFR could determine the formation of heterodimers with HER269 with deficient endocytosis70 able to stimulate alternative signaling pathways that bypass anti-HER2 inhibition, as suggested.71 This might explain why not all patients with metastatic breast carcinomas overexpressing HER2 respond to treatment with HCT.40, 41

The major effect of CQ is related to the buffering capacity of intracellular acidic compartments (endosomes, lysosomes and Golgi complexes). Other suggested effects include intercalation with the DNA by noncovalent binding, especially to the CG sequence. The double-, but not the mono-, protonated form of CQ (which prevails at acidic pH) binds noncovalently to DNA via the cationic charge on the aromatic ring.72 This can potentially affect the structure of the DNA, preventing DNA polymerase and other DNA-binding proteins from functioning properly. In our preliminary studies, we found that at concentrations higher than 30 μM CQ becomes increasingly inhibitory, as detected by 3H-thymidine incorporation. However, the weak, noncovalent binding of CQ to DNA limits its toxic effect compared to other drugs, such as cisplatin, which bind covalently to DNA and produce irreversible intra- or interstrand crosslinks, resulting in inhibition of DNA synthesis and transcription.73, 74 Moreover, CQ up to 100 μM has no effect on cyclin or Cdk expression levels.75 We did not find any significant cytotoxic effects or changes in levels of Cdk2 or cyclin E induced by CQ at the concentration used. The major effect of CQ is due to the increase in Ab-mediated internalization of HER2 and does not represent a synergism between 2 signaling pathways induced by CQ and HCT, respectively. Thus, engagement of HER2 on BT474 cells by equivalent amounts of plastic-immobilized HCT, which prevents endocytosis of Ab–receptor complexes, totally abrogates the effect of CQ. Moreover, in contrast to the inhibitory effect of soluble HCT, immobilized HCT stimulates BT474 cells, as judged by the higher rate at which monolayers reach confluence and the DNA replication rates. This indicates that crosslinking of the receptor per se generates a stimulatory, rather than an inhibitory, signal when it does not induce its internalization. Not surprisingly, the tumor-stimulatory effect of some anti-HER2 Abs is associated with prolonged receptor crosslinking on the cell surface, whereas the inhibitory Abs induce rapid endocytosis.34 The failure of EGF and TGF-β ligands to stimulate endocytosis of another member of the family, ErbB-1, due to either an endocytosis-defective receptor76 or low-affinity binding77 also results in stronger mitogenic signals. This shows the essential role of HER2 localization in the plasma membrane for signaling and supports the key role of CQ-stimulated intracellular receptor trapping as the basis for the enhanced and prolonged antitumor effect of the HCT–CQ combination.

The extent to which proteasomal processing is involved in downregulation of receptors is unclear. Previous reports have suggested that EGF- or Ab-mediated lysosomal degradation of EGFR depends on proteosomal activity.78 Inhibitory anti-HER2 Abs can enhance the ubiquitination and degradation of HER2 mediated by the adaptor protein c-Cbl.79 However, we did not find any involvement of proteosomal degradation during the combination treatment. Regardless of the underlying mechanism(s), our results demonstrate that CQ potentiates the antitumor effect of HCT on HER2-overxpressing cells and is characterized by (i) increased extent and duration of receptor downregulation; (ii) increased upregulation of p27Kip1 and inactivation of its upstream regulator, Akt; (iii) increased apoptosis and retarded G1 arrest after HCT removal. The potential clinical relevance for the treatment of HER2-overexpressing tumors is clear; and in vivo experiments are, therefore, under way. One major prospect is that administration of CQ with HCT could shorten the duration of treatment and achieve a maximal therapeutic effect with lower doses of HCT. In addition, the low cell surface HER2 may also influence tumor responsiveness to radio- and chemotherapy.

Acknowledgements

We thank Dr. E.S. Vitetta for providing Her81 antibody and Dr. X. Wang for critical reading of the manuscript. We also thank Dr. S. Meng for help with fluorescence microscopy.

Ancillary