Main

The human epidermal growth factor (Her-2/neu) is a 185 kDa oncoprotein (p185), which is overexpressed in 25–30% of invasive breast cancers (Schechter et al, 1984; Slamon et al, 1987; Olayioye et al, 2000). Her-2/neu overexpression has consistently been found to confer resistance to cytotoxic and endocrine therapy and to account for an aggressive biological behaviour, thereby resulting in shorter disease-free and overall survival in both, patients with early and advanced breast cancer (Slamon et al, 1987).

The Her-2/neu molecule is composed of an extracellular ligand-binding domain, an amphipathic transmembrane region and an intracellular tyrosine kinase domain, which contains a carboxy tail with five major autophosphorylation sites (Ullrich and Schlessinger, 1990). Ligand binding is thought to initiate the formation of homo- and heterodimeric receptor complexes with other growth factor receptor (GFR) class 1 family members such as EGFR, Her-3 and Her-4 into which Her-2/neu is recruited as a preferential dimerisation partner (Tzahar et al, 1996). This process is followed by intrinsic tyrosine kinase-mediated autophosphorylation and mutual phosphorylation of the respective dimerisation partners and ultimately results in activated receptor complexes (Segatto et al, 1990; Tzahar et al, 1996). Alternatively, in vitro Her-2/neu activation has also been demonstrated to occur as a consequence of spontaneous cleavage of its extracellular domain (ECD) thereby resulting in the production of a truncated membrane-bound fragment (p95) with kinase activity (Christianson et al, 1998; Molina et al, 2001). Since the p95 fragment has also been detected in breast cancer specimens, it has been suggested that shedding of the ECD may represent an alternative activation mechanism of Her-2/neu in vivo (Christianson et al, 1998).

Intracytoplasmatic phosphorylated tyrosine residues of the Her-2/neu molecule function as high-affinity binding sites for SH2 domain containing proteins, which link the receptor to intracellular signal transduction processes such as the ras-raf-mitogen activated protein kinase (MAPK) and the phosphatidylinositol-3 kinase (PI-3 K) pathways. Both are believed to be key elements in the regulation of cell proliferation and survival (reviewed in Slichenmyer and Fry, 2001). In this context, phosphorylation of the 1248 tyrosine residue (Tyr1248), which is supposed to constitute the main autophosphorylation site of Her-2/neu, is a key event for downstream signalling (Akiyama et al, 1991; Dougall et al, 1994; DiGiovanna and Stern, 1995; Marshall, 1995).

Monoclonal antibodies targeting the Her-2/neu ectodomain, such as trastuzumab and its murine precursor 4D5, have been shown to abrogate the Her-2/neu activation processes and to interfere with Her-2/neu-dependent gene expression, to modulate cell cycle progression, induce cellular differentiation and to sensitise Her-2/neu overexpressing cells to apoptotic stimuli. These effects are mediated by several distinct mechanisms including the blockade of ligand binding, disruption of homo- and heterodimer formation, induction of receptor internalisation and degradation as well as prevention of cleavage of the ECD (Baselga et al, 2001; Yip and Ward, 2002), all of which ultimately result in a decrease of receptor phosphorylation in vitro (Kumar et al, 1991; Lane et al, 2000). An intact tyrosine kinase activity appears to be the main precondition for growth inhibition induced by trastuzumab, since the regular functions of all of the mentioned mechanisms depend upon Her-2/neu kinase integrity (Xu et al, 1994).

These observations have been successfully translated into clinical use. Thus, trastuzumab has not only demonstrated activity as a single agent, but had also exerted synergistic activity when administered in conjunction with cytotoxic drugs, thus resulting in prolonged progression-free and overall survival in patients with Her-2/neu overexpressing metastatic breast cancer (Slamon et al, 2001; Vogel et al, 2002). Currently, the decision for treating patients with metastatic breast cancer with trastuzumab is based upon the detection of Her-2/neu overexpression and/or amplification of the c-erbB-2 gene determined by immunohistochemistry (IHC) and fluorescence in situ hybridisation (FISH), respectively. However, aside from the intensity of Her-2/neu overexpression, additional factors that would predict the effectivity of trastuzumab as single agent or in combination with cytotoxic treatment are still lacking. Since tyrosine kinase activation is the downstream mechanism of action for Her-2/neu, it has been postulated that tumours exhibiting active Her-2/neu signalling represented by phosphorylation of tyrosine residues might be those most sensitive to treatment with trastuzumab (DiGiovanna and Stern, 1995; Thor et al, 2000).

The aim of the present study was to determine whether cleavage of the Her-2/neu ECD is correlated with overall tyrosine kinase activity in vivo and might therefore constitute a clinically relevant ligand-independent mechanism for the activation of Her-2/neu in breast cancer. In addition, we correlated the clinical course of disease of patients treated with trastuzumab with Her-2/neu phosphorylation status to determine the clinical relevance of active Her-2/neu signalling.

Materials and methods

Patient population–inclusion and exclusion criteria

Patients who had received trastuzumab (Herceptin®, Roche Pharmaceuticals, Vienna, Austria)±chemotherapy at our institution between April 2000 and February 2003 in accordance with previously published treatment protocols (Pegram et al, 1998; Cobleigh et al, 1999; Burstein et al, 2001; Miller et al, 2001; Seidman et al, 2001; Slamon et al, 2001; Esteva et al, 2002; Vogel et al, 2002) were identified retrospectively by using pharmacy protocols. Trastuzumab was administered as 4 mg kg−1 loading dose followed by a weekly 2 mg kg−1 maintainance dose, as described in Cobleigh et al (1999), to patients with metastatic breast cancer with grade 2+ or 3+ Her-2/neu overexpression assessed by immunohistochemistry. Patients included into the present analysis were required to have bidimensionally measurable (with both diameters >1.0 cm and at least one lesion with both diameters >1.5 cm) disease (excluding previously irradiated or bone lesions as the only site of measurable disease) with clearly defined margins and radiologically (CT and/or MRI and/or ultrasound) documented tumour progression before initiation of trastuzumab-based treatment.

In addition, patients’ records were required to contain documented response assessment performed every 6–8 weeks (depending on the therapeutic regimen). Response evaluation was performed by independent review of patients records and radiology reports by two investigators and classified in accordance with the Southwest Oncology Group response criteria and end point definitions (Green and Weiss, 1992). Patients who had discontinued treatment before radiological response assessment or had been lost to follow-up (as defined by the last record obtained >8 weeks before the present analysis) were censored as therapeutic failures or deaths, respectively. Patients who had previously received treatment with monoclonal antibodies, vaccines or biological response modifiers were excluded. Further inclusion criteria consisted of availability of intact paraffin-embedded tissue (excluding mechanically altered tissue and fine needle aspirates) from which the original assessment of Her-2/neu overexpression had been performed and deep-frozen (−80°C) sera that had been obtained immediately before the first infusion of trastuzumab. In accordance with our institutional ethical committee guidelines, signed informed consent was obtained from all patients before 8 ml of blood was drawn from the same venous access, which was afterwards immediately used for infusion of trastuzumab. Data on oestrogen and progesterone receptor status of tissue samples from which the original assessment of Her-2/neu overexpression had been performed were available from pathology records.

Determination of Her-2/neu overexpression

In all cases, reassessment of Her-2/neu overexpression and determination of Her-2/neu phosphorylation were performed independently by two experienced pathologists blinded to the clinical course of patients and the results of other tests performed. All testing was performed in the identical tissue material used for initial patient selection for trastuzumab therapy. Her-2/neu protein expression was evaluated on paraffin-embedded tissue using the HercepTest kit (DAKO A/S, Glostrup, Denmark) for immunoenzymatic staining in accordance with the protocol described in the manufacturer's guide: tissue sections were dewaxed in xylene and then rehydrated through ethanol to distilled water. Subsequently, tissue sections were immersed in epitope retrieval solution (DAKO) at 95°C, and then in waterbath at 95°C for a total of 40 min, followed by a 20 min cool-down period at room temperature. Slides were incubated at room temperature with the primary rabbit polyclonal antibody to the Her-2/neu oncoprotein (supplied by the kit manufacturer) on a DAKO Autostainer for 30 min followed by application of peroxidase-blocking reagent. Antibody was localised by incubating slides with the DAKO Visualization Reagent using horseradish peroxidase-conjugated goat anti-rabbit immunoglobulins for 30 min using the DAKO Autostainer. Sections were finally incubated with diaminobenzidine (DAB) as chromogen and counterstained with haematoxylin. To assess the expression of the Her-2/neu oncoprotein accurately, positive controls consisting of freshly cut breast cancer cases known to overexpress Her-2/neu and a control slide consisting of three pelleted, formalin-fixed, paraffin-embedded human breast cell lines with staining intensity scores of 0, 1+ and 3+ (supplied by the kit manufacturer) were included in each staining run. Negative controls were performed by substitution of the HER-2/neu primary antibody by normal rabbit serum (DAKO Negative Control Reagent). Only membrane staining intensity and pattern were evaluated using the 0 to 3+ scale as illustrated in the HercepTest kit scoring guidelines.

Determination of Her-2/neu gene amplification

In cases of 2+ immunostaining for Her-2/neu, oncoprotein gene copy number of the HER-2/neu (c-erbB-2) gene was determined by dual-colour FISH referring to the numbering of chromosome 17 using the PathVysion® HER-2 DNA probe-kit (Vysis Inc., Downers Grove, IL, USA) according to the manufacturer's directions. The HER2/neu-SpectrumOrange probe contains a DNA sequence specific for the c-erbB-2 human gene locus and hybridises to region 17q11.2–q12 of human chromosome 17. The CEP 17 (chromosome enumeration probe 17)/SpectrumGreen probe contains alpha-satellite DNA that hybridises to the D17Z1 locus (centromere region of chromosome 17). Fluorescence was evaluated by using a Nikon E600 microscope with a Y-Fl Epi-Fluorescence Attachment (Nikon, Tokyo, Japan) and a black-and-white, charge-coupled device (CCD) camera (COHU 4912; Cohu, San Diego, CA, USA) run by Lucia-Fish software (Laboratory-Imaging, Prague, Czech Republic). Fluorescence in situ hybridisation slides were compared with adjacent haematoxylin/eosin slides. Her-2/neu signals and CEP17 signals in at least 20 cancer nuclei per tumour were scored. A cell was considered amplification positive if the Her2/neu : CEP17 ratio exceeded 2.

Determination of specifity of phospho-specific (P-Tyr1248) Her-2/neu antibody (PN2A)

Her-2/neu (p185) overexpressing SKBR3 human mammary carcinoma cells (American Type Culture Collection, Rockville, MD, USA) were cultured in McCoy's 5A modified medium (Gibco BRL, Paisley, Scotland), supplemented with 15% of heat-inactivated fetal calf serum (FCS), glutamine and 50 IU penicillin and 50 μg ml−1 streptomycin (all Gibco) in a humidified atmosphere containing 5% CO2. Confluent SKBR3 cells were cultivated in six-well plates (Corning Inc., NY, USA) and serum starved in medium containing 0.1% FCS overnight. Since epidermal growth factor (EGF) has been shown to stimulate tyrosine phosphorylation of Her-2/neu dramatically (DiGiovanna and Stern, 1995), cells were then incubated with and without EGF (human rEGF, Strathmann Biotech AG, Hamburg, Germany) at 100 ng ml−1 for 2, 4, 8, 10, 30 and 60 min at room temperature. To maximise differential phosphorylation of p185 isolated from EGF-treated cells, the phosphotyrosine phosphatase inhibitor sodium ortho-vanadate (Alexis Biochemicals, QBIOGENE Inc., Carlsbad, CA, USA) was added at a final concentration of 500 μM during the final hour prior to stimulation of cultures that were to be stimulated with EGF. After incubation, cells were lysed in 0.5 ml 1 × Laemli buffer and boiled at 96°C for 10 min.

To perform Western blotting, cell lysates were separated on a 7.5% SDS–polyacrylamide electrophoretic gel (PAGE) with a constant voltage of 200 V for 30 min and transferred to a nitrocellulose membrane (Novex) by electroblotting (20 V, 1 h). SeeBlue® Pre-stained Standards (250, 98 and 64 kDa) (Novex, Invitrogen, Paislay, UK) were used as molecular weight standards. Membranes were incubated in blocking solution for 1 h at room temperature and treated afterwards with 1 μg ml−1 of mouse monoclonal antibody: (A) mouse monoclonal c-erbB-2/HER-2/neu Ab-20 (L87+2ERB19), (B) c-erbB-2/HER-2/neu (phospho-specific) Ab-18 (clone PN2A, both NeoMarkers, Fremont, CA, USA) at room temperature for 2 h. Membranes were incubated in alkaline phosphatase-coupled anti-mouse antibody for 1 h at room temperature after a 30 min wash. Following another washing, the membranes were incubated in 4-nitro blue tetrazolium chloride/5-bromo-4-chloro-3-indolyl phosphate (NBT/BCIP, Roche Diagnostics GmbH, Vienna, Austria) developer solution for 10 min. Membranes were developed and finally washed with bidistilled water.

Determination of Tyr1248 Her-2/neu phosphorylation of tumour samples

Determination of Her-2/neu phosphorylation was performed using the monoclonal antibody PN2A (NeoMarkers, Westinghouse Drive, Fremont, CA, USA). As demonstrated by Western blotting analysis (see above and Green and Weiss, 1992) and immunohistochemistry using peptide blocking experiments (Green and Weiss, 1992) the PN2A antibody specifically recognises phosphorylation of the major autophosphorylation site Tyr-1248 (P-Tyr1248) without crossreactivity with c-erbB-1 (EGFR), c-erbB-3 or c-erbB-4, or unphosphorylated Her-2/neu. Formalin-fixed paraffin-embedded tissue sections (4 μm) were deparaffinised, rehydrated, and endogenous peroxidase-blocked with 2% hydrogen peroxide. Antigen retrieval was performed by placing sections in 10 mmol l−1 citrate buffer (pH 6.0) and microwave treatment for 15 min. Slides were allowed to cool to room temperature (RT), washed with phosphate-buffered saline (PBS) and distilled water, and blocked with Ultra V Block (Lab Vision, Westinghouse Drive, Fremont, CA, USA). PN2A (6 μg ml−1) was applied and sections were incubated at 4°C overnight. After two additional PBS washes, sections were sequentially incubated at RT for 30 min with biotinylated goat anti-polyvalent (Lab Vision, Westinghouse Drive, Fremont, CA, USA) and streptavidin-HRP (Lab Vision, Westinghouse Drive, Fremont, CA, USA). Subsequently, slides were incubated with 3-amino-9-ethylcarbazole (AEC, a widely used chromogen), counterstained with haematoxylin and cover-slipped.

Expression of phosphorylated Her-2/neu was visually assessed using the same scoring system applied for determination of Her-2/neu overexpression (see above). In contrast to evaluation of receptor overexpression of Her-2/neu (considering grade 2+ and grade 3+tumours positive), tumours exhibiting a clearly discernible positive signal for receptor phosphorylation (⩾grade 1+using Herceptest guidelines) on cellular membranes were considered positive, because even weak staining for phosphorylation of the 1248 tyrosine residue of the Her-2/neu molecule (pHER-2/neu) might represent tyrosine kinase activity, that is, active receptor signalling. In contrast, faint cytoplasmatic staining in the absence of membraneous staining was not considered positive. For each assay, pelleted, formalin-fixed, paraffin-embedded human T47D and rhEGF-stimulated SKBR-3 breast cancer cell lines (American Type Culture Collection, Rockville, MD, USA) were used as positive controls. Photomicrographs of immunohistochemical analysis for Her-2/neu overexpression and Her-2/neu phosphorylation are depicted in Figure 1.

Figure 1
figure 1

Immunohistochemical analysis of two (grade 3+) Her-2/neu overexpressing tumours (upper row) differing with respect to Her-2/neu phosphorylation status at the tyr1248 residue (lower row): strong (left) and absent Her-2/neu phosphorylation (right).

Full size image

Determination of serum Her-2/neu ECD levels

Sera obtained immediately before the first infusion of trastuzumab were analysed using a Sequential Solid Phase Sandwich Human Her-2/neu Quantitative ELISA (Her-2/neu Microtiter ELISA, Oncogene Science, Cambridge, MA, USA) according to the manufacturer's instructions. Microplates were washed using an automated washer (Dias Microplate Washer, Dynex Technologies, Denkendorf, Germany). Absorbance reading was performed using an automated reader (FLUOstar Galaxy, BMG Labtechnologies, Offenburg, Germany) and serum ECD concentrations were calculated from absorbance data using the Fluoscan Galaxy software (vers. 4.20-0, BMG Labtechnologies, Offenburg, Germany). Intra- and interassay precision of the assay were always <5% CV.

Statistical analysis

Frequencies of patients’ characteristics were compared by Fisher's exact test, and differences in Her-2/neu ECD levels, age and recurrence-free interval (time from initial diagnosis to appearance of metastatic disease) with Mann–Whitney's U-test. After performance of a normalising transformation on serum Her-2/neu ECD, a multiple analysis of variance (ANOVA) was assessed using serum Her-2/neu ECD as dependent variable. Relationships between serum Her-2/neu ECD and the presence of phosphorylation were adjusted by the grade of Her-2/neu overexpression, the presence or absence of visceral metastases and the number of sites with metastatic disease. The results of the multiple analysis of variance are represented by medians and 95% confidence intervals (95% CI). For all subsequent analyses, grade of Her-2/neu overexpression (2+ vs 3+), presence or absence of Tyr1248 phosphorylation, patient age, recurrence-free interval, anthracycline pretreatment (for early breast cancer or metastatic disease), number of prior chemotherapeutic regimens for metastatic disease, oestrogen receptor (ER) and progesterone receptor (PgR) status, Karnofsky's performance index, presence or absence of visceral metastases (liver and/or lung), number of organs involved by metastatic disease, type of treatment (single-agent trastuzumab vs combination with chemotherapy) and baseline serum Her-2/neu ECD levels were entered as variates. Multiple logistic regression analyses were used to determine whether any of these variates could predict response or clinical benefit (response or disease stabilisation) from trastuzumab-based treatment. In analogy, multiple Cox regression models were utilised to identify the properties of the predictors mentioned above on progression-free and overall survival. Confounders without significant influences were removed by the backward selection method based on the Wald statistic. The related risk (RR), the odds ratio (OR), and 95% CI were calculated with the proportional hazard method in respect of Cox regression and logistic regression. Survival curves (progression-free and overall survival) were compared with the log-rank test. For all analyses, a P-value <5% was considered statistically significant. SPSS statistical software system (SPSS Inc., Chicago, IL, USA, version 10.0) was used for all calculations.

Results

Study population

A total of 69 patients with Her-2/neu overexpressing metastatic breast cancer, who received trastuzumab-based treatment at our institution during the indicated period, were identified. Seven patients were excluded for the present analysis due to the following reasons: nonavailability of tissue samples from which determination of Her-2/neu overexpression had been performed (one patient), mechanically altered tissue samples not amenable for subsequent testing (one patient), nonavailability of sera obtained immediately before the first infusion of trastuzumab (five patients). Thus, the final study population comprised 62 patients (median age 52.6, range 27.6–80.9 years). Patients’ characteristics are depicted in accordance to Her-2/neu phoshorylation status (pHER-2/neu) of tumour samples in Table 1. This analysis covers a median observation period of 22.4 (range 6.1–37.8) months, during which 19 (31%) objective responses (including eight complete responses) were observed, 19 (31%) patients experienced durable disease stabilisation (⩾4 months) and 24 (39%) of patients had primarily progressive disease. As of February 2003, 55 (89%) patients had experienced disease progression and 28 (45%) deaths had occurred, all of which were attributed to disease progression. No patient was lost to follow-up. Median (95% CI) progression-free and overall survival calculated from survival function were 5.7 (95% CI 2.5–8.9) and 23.8 (95% CI 16.7–30.8) months, respectively.

Table 1 Patients and treatment characteristics according to phosphorylation status of Her-2/neu (pHer-2/neu)
Full size table

Specifity of phospho-specific (P-Tyr1248) Her-2/neu antibody (PN2A)

Figure 2 shows the results of Western blotting performed on EGF-stimulated SKBR3 cells by using the PN2A (Figure 2B) and the control Her-2/neu antibody (Figure 2A). Whereas the Her-2/neu antibody detects Her-2/neu independent of its activation status, the PN2A antibody specifically recognises the activated, tyrosine phosphorylated (P-Tyr1248) form of Her-2/neu. This is supported by enhanced tyrosine kinase activities (larger PN2A bands from left to right until point of saturation) corresponding to different treatment periods of SKBR3 cells with EGF (compare Figure 2B, lanes 8, 6-3).

Figure 2
figure 2

Western blot analysis demonstrating the phosphorylation state specifity of the PN2A antibody. Specifity of (A) anti-Her-2/neu (p185) antibody and (B) anti-phospho (tyr1248) Her-2/neu antibody (PN2A) is shown by Western blot analysis of whole-cell lysates. SKBR3 cells were treated without (lanes A1 and B8) and with 100 ng ml−1 EGF for 2 (lanes A3 and B6), 4 (lanes A4 and B5), 8 (lanes A5 and B4), 10 (lanes A6 and B3), 30 (lanes A7 and B2) and 60 (lanes A8 and B1) min (kDa, molecular weight in kilodalton, molecular weight markers at 250, 98 and 64 kDa).

Full size image

Frequencies of Her-2/neu gene amplification, protein overexpression and receptor activation

Immunohistochemical analysis for Her-2/neu expression demonstrated grade 2+ overexpression in eight (13%) and grade 3+ staining in 54 (87%) of tumour samples. FISH analysis of grade 2+ Her-2/neu overexpressing samples did not reveal gene amplification in any of the specimens examined. Positive immunohistochemical staining of the cell membranes for phosphorylation of the Her-2/neu 1248 tyrosine residue (tyr1248) was observed in 15 (24%) tissue samples with intense staining detected in four, moderate staining in three and weak, but clearly discernible staining for receptor phosphorylation in eight tissue samples, respectively. Moderate or intense staining for tyr1248 (pHer-2/neu) was associated with grade 3+ Her-2/neu overexpression in all cases, whereas weak staining for tyr1248 was observed in five grade 3+ and three grade 2+ Her-2/neu overexpressing tumours.

Correlation of serum Her-2/neu ECD levels with pHER-2/neu status

Patients with Her-2/neu phosphorylation presented with significantly higher serum Her-2/neu ECD levels before initiation of trastuzumab-based treatment (median 148.2, range 6.1–510.0 ng ml−1) as compared to patients without Her-2/neu phosphorylation (median 28.5, range 5.2–6076.2 ng ml−1; Mann–Whitney test: P=0.010, Figure 3).

Figure 3
figure 3

Serum Her-2/neu ECD levels are depicted in accordance to Her-2/neu overexpression (grade 2+ vs 3+) and Her-2/neu phosphorylation status of tumours (extreme outliers not depicted).

Full size image

Serum Her-2/neu ECD levels have also been reported to correlate with the intensity of immunohistological Her-2/neu overexpression–with lower values observed in patients with grade 2+ Her-2/neu-positive tumors–and to correlate with tumour mass (Leitzel et al, 1992; Krainer et al, 1997; Kostler et al, manuscript submitted). In our patient collective, serum ECD levels (median, range) also tended to be higher in patients with grade 3+ Her-2/neu overexpressing tumours (45.2, 5.2–6076.2 ng ml−1) than in patients with grade 2+ Her-2/neu overexpressing tumours (15.4, 11.2–245.4 ng ml−1), but without statistical significance (P=0.734). Therefore, we also performed a correlation between serum Her-2/neu ECD levels and pHer-2/neu expression in the subset of grade 3+ Her-2/neu overexpressing tumours. In this subset, pHer-2/neu expressing cancers were again found to have higher median (range) serum ECD values of 108.4 (6.1–510.0) ng ml−1, as compared to patients without pHer-2/neu expression (35.9, 5.2–6076.2 ng ml−1). However, this difference was not significant (P=0.061), presumably because of the small sample size. Figure 3 depicts serum Her-2/neu ECD levels according to Her-2/neu overexpression and phosphorylation status.

The association between serum Her-2/neu ECD levels and Her-2/neu phosphorylation status was adjusted by the grade of Her-2/neu overexpression and the presence/absence of visceral metastases in a multiple analysis of variance (ANOVA) with normalised Her-2/neu ECD as criterion variable. As shown in Table 2, patients with Her-2/neu phosphorylation had an adjusted median serum Her-2/neu ECD concentration of 41.2 (95% CI 17.5–97.3) ng ml−1, which was significantly higher as compared to patients without Her-2/neu phosphorylation having an adjusted median serum ECD level of 18.5 (95% CI 9.3–36.7) ng ml−1 (P=0.046). The adjusted median serum Her-2/neu ECD concentration of patients with visceral metastases amounted to 59.4 (95% CI 32.4–109.0) ng ml−1 in comparison to patients without visceral metastases having 12.8 (95% CI 5.0–33.1) ng ml−1 (P=0.001). The grade of Her-2/neu overexpression and the number of sites with metastatic disease did not show a significant correlation with serum Her-2/neu ECD levels.

Table 2 Multiple analysis of variance (ANOVA) applied to Her-2/neu ECD as normalised criterion variable
Full size table

Correlation of pHer-2/neu status, histopathological and patients’ characteristics with response to trastuzumab-based treatment

Both objective response rates–seven out of 15 (47%) vs 12 out of 35 (34%, P=0.197)–and rates of clinical benefit–12 out of 15 (80%) vs 26 out of 47 (55%, P=0.129)–to trastuzumab-based treatment tended to be higher in patients with tumours exhibiting tyr1248-phosphorylated Her-2/neu as compared to those without pHer-2/neu expression.

Multiple logistic regression analyses were performed to evaluate the predictive role of Her-2/neu tyr1248 phosphorylation status upon response and benefit from trastuzumab-based treatment. In all analyses, grade of Her-2/neu overexpression, patient age, recurrence-free interval, anthracycline pretreatment, number of prior chemotherapeutic regimens for metastatic disease, ER and PgR status, Karnofsky's performance index, presence or absence of visceral metastases, number of organs involved by metastatic disease, type of treatment (single-agent trastuzumab vs combination with chemotherapy) and baseline serum Her-2/neu ECD levels were entered as variates.

In univariate analysis, none of the variates mentioned above significantly predicted response with significance levels below 10%. Her-2/neu phosphorylation status was the only variate showing a trend to predict benefit (for descriptive values see Table 3).

Table 3 Clinical and histopathological predictors of response (complete or partial), clinical benefit (response or disease stabilisation), progression-free survival (PFS) and overall survival (OAS) with significance levels below 10% (P<0.1) in univariate and multivariate analyses RR=related risk; 95% CI=95% confidence interval
Full size table

In multivariate analysis, no significant predictors of response to trastuzumab-based treatment were identified; again, Her-2/neu phosphorylation status was the only covariate that revealed a trend to significant prediction of response to trastuzumab-based treatment, whereas no predictors of clinical benefit were identified in multivariate analysis (Table 3).

Correlation of pHer-2/neu status, histopathological and patients’ characteristics with progression-free and overall survival from trastuzumab-based treatment

In patients with tyr1248 Her-2/neu phosphorylation, the median (95% CI) progression-free survival calculated from survival curves was 11.7 (95% CI 5.2–18.3) months and thus significantly longer as compared to the median (95% CI) progression-free survival of 4.5 (3.4–5.6) months observed in patients without pHer-2/neu staining tumours (log-rank test: P<0.0095). Figure 4 depicts progression-free survival curves according to Her-2/neu phosphorylation status. Median progression-free survival (4.4, 95% CI 2.8–6.0 months) of patients with Her-2/neu ECD levels below 15 ng ml−1 did not differ significantly from progression-free survival of patients with Her-2/neu ECD above 15 ng ml−1 (6.7, 95% CI 2.3–11.2 months, log-rank test: P=0.129).

Figure 4
figure 4

Progression-free survival in patients receiving trastuzumab-based treatment for Her-2/neu overexpressing metastatic breast cancer in accordance to activation status of Her-2/neu. Full line: unphosphorylated Her-2/neu, dotted line: phosphorylated Her-2/neu (log-rank test P<0.0095).

Full size image

Univariate Cox regression analysis revealed that grade 3+ Her-2/neu overexpression, presence of Her-2/neu phosphorylation and low number of organs involved by metastatic disease were significant predictors of longer progression-free survival, whereas a good performance status showed a trend to a reduced risk of progression (descriptive values are depicted in Table 3). Multivariate Cox regression analysis showed that grade 3+ Her-2/neu overexpression, presence of visceral metastases and Her-2/neu tyr1248 phosphorylation were independent predictors of longer progression-free survival, whereas a higher number of organs affected by metastatic disease represented a significant adverse predictor of progression-free survival (Table 3).

Median (95% CI) overall survival in patients without pHer-2/neu expression was 23.8 (18.1–29.5) months and did not differ significantly from overall survival observed in patients with pHer-2/neu expressing tumours, which had not been reached at the time of these analyses (log-rank test: P=0.6842). Likewise, median overall survival of patients with Her-2/neu ECD levels below 15 ng ml−1 was not reached during the observation period. Patients with Her-2/neu ECD levels above 15 ng ml−1 had a median overall survival of 19.6 (95% CI 13.2–26.6 months, log-rank test: P=0.143).

In univariate Cox regression analysis for overall survival grade 3+ Her-2/neu overexpression, a low number of organs affected by metastatic disease and good performance status were significant predictors for longer overall survival, whereas serum Her-2/neu ECD levels observed at baseline did not cause significant effects (descriptive values are depicted in Table 3). In multivariate analyses, a low number of organs affected by metastatic disease and grade 3+ Her-2/neu overexpression yielded a reduced risk of death within the observation period (Table 3).

Discussion

The present investigation was performed to evaluate a potential correlation of tyrosine phosphorylated Her-2/neu with cleavage of its ECD in breast cancer patients, and to determine the clinical consequences of Her-2/neu activation with respect to the efficacy of trastuzumab-based treatment. We here report that, in Her-2/neu overexpressing tumours, increased serum Her-2/neu ECD levels are indeed associated with active Her-2/neu tyrosine kinase signalling, as represented by Tyr1248-phosphorylated Her-2/neu. Our findings strongly support the clinical relevance of previous in vitro observations, which have demonstrated that spontaneous proteolytic cleavage of the ECD represents a ligand-independent activation mechanism of Her-2/neu (Molina et al, 2001).

In this context, it is interesting to note that the detection of increased serum levels of the Her-2/neu ECD has been associated with increased resistance to endocrine therapy in Her-2/neu overexpressing metastatic breast cancer (Lipton et al, 2002), thereby suggesting that resistance to antihormonal agents was associated with active Her-2/neu-dependent signalling. On the other hand, serum Her-2/neu ECD levels have been associated with increased response rates and prolonged progression-free survival to subsequent trastuzumab-based treatment (Esteva et al, 2002; Kostler et al, manuscript submitted). It can, therefore, be hypothesised that not only the mere Her-2/neu overexpression, but also its activation status are crucial for susceptibility to the biological effects of trastuzumab in vivo.

Based on these assumptions, we have determined the impact of the Her-2/neu phosphorylation on the clinical outcome of patients with Her-2/neu overexpressing metastatic breast cancer receiving trastuzumab-based treatment. In the present analysis, we have found a trend towards higher response rates and higher rates of clinical benefit in patients with activated Her-2/neu. Notably, progression-free survival to trastuzumab-based treatment was more than doubled in patients with tumours exhibiting activated Her-2/neu as compared to patients lacking tyrosine-phosphorylated Her-2/neu. Whether this observation can be attributed to an increased and prolonged sensitivity to trastuzumab-based treatment or represents an intrinsic biologic characteristic of tumours exhibiting tyr1248 phosphorylation of Her-2/neu remains to be determined in future prospective trials evaluating the predictive value of pHer-2/neu for various forms of treatment.

In addition, the absence of Tyr1248 phosphorylation does not necessarily preclude Her-2/neu-dependent signalling: although Her-2/neu activation is thought to be a hierarchical event in which Tyr1248 is the ultimate C-terminal site to be phosphorylated before downstream signalling occurs (Akiyama et al, 1991; Harder et al, 1994), it is conceivable that interactions of other phosphorylated (non)tyrosine residues with heterodimerisation partners resulting in heterodimerisation partner-dependent secondary signalling (Graus-Porta et al, 1997; Olayioye et al, 19982000; Ouyang et al, 2001) could also be targeted by trastuzumab. The characterisation these activation states, of stimulatory and inhibitory interactions and their correlation with the clinical effects of trastuzumab will clearly have a profound effect on our biologic understanding of receptor signalling and ultimately refine targeted therapies.