Breast Cancer Research and Treatment 86: 9–18, 2004.© 2004 Kluwer Academic Publishers. Printed in the Netherlands.

Report

Prognostic and predictive impact of the HER-2/neu extracellulardomain (ECD) in the serum of patients treated with chemotherapyfor metastatic breast cancer

Volkmar Müller1,∗, Isabell Witzel1,∗, Hans Joachim Lück4, Günter Köhler4, Gunther vonMinckwitz4, Volker Möbus4, Daniel Sattler4, Waldemar Wilczak2, Thomas Löning2, FritzJänicke1, Klaus Pantel3, and Christoph Thomssen1

1Department of Gynecology, 2Department of Gynecopathology, 3Institute of Tumor Biology, University HospitalHamburg-Eppendorf; 4AGO Study Group, Germany

Key words: chemotherapy, HER-2/neu serum, metastatic breast cancer, paclitaxel, predictive factors

Summary

Background. The extracellular domain of the HER-2/neu-receptor (ECD) is shed from the receptor protein and canbe detected in serum. However, the clinical implication of HER-2/neu ECD measurement must be further evaluated.

Methods. In patients with metastatic breast cancer participating in a trial on first-line chemotherapy, the as-sociation of serum HER-2/neu ECD with progression-free interval, survival, and response was studied. Bloodsamples of patients receiving epirubicin and either cyclophosphamide (EC) or paclitaxel (ET) were collected before(n = 103) and in addition, after three courses of therapy (n = 46).

Results. HER-2/neu ECD levels correlate with HER-2/neu overexpression of corresponding primary tumors de-termined by immunohistochemistry (antibody CB11, p = 0.018) with an optimized cut-off at 15 ng/mL. Elevatedserum levels of HER-2/neu ECD before chemotherapy were correlated with shorter overall survival (p = 0.0097),but not with reduced progression-free survival and response to chemotherapy. In subgroup analyses, patients withelevated pretherapeutic HER-2/neu ECD levels treated with EC showed shorter overall survival (p = 0.0092); nodifference was seen in the ET group. With regard to progression-free survival, patients with elevated HER-2/neuECD levels tended to benefit from ET (p = 0.0341), in patients with low levels no difference was observed betweenEC and ET. A decrease of HER-2/neu ECD levels after three courses of therapy was associated with response totherapy (p = 0.006).

Conclusion. In our group of metastatic breast cancer patients, elevated HER-2/neu ECD levels are associatedwith decreased overall survival. With regard to progression-free survival, particularly patients with high HER-2/neuECD levels seem to benefit from taxane-containing chemotherapy.

Abbreviations: ECD: extracellular domain; ELISA: enzyme-linked immunosorbant assay; kDa: kiloDalton;mL: milliliter; ng: nanogram

Introduction

The HER-2/neu protooncogene encodes a 185 kDatransmembrane receptor with tyrosine kinase activity.Amplification of the HER-2/neu gene and/or overex-pression of the protein correlates with poor prognosis

∗These authors contributed equally to this paper.

of patients with breast cancer in most studies. HER-2/neu overexpression might influence a number ofaspects in breast cancer phenotype. These includereduced endocrine responsiveness, increased meta-static ability and drug resistance. The extracellulardomain of the HER-2/neu-protein (HER-2/neu ECD)

10 V Müller et al.

is cleaved and shed from the receptor and can be de-tected in serum as a protein of approximately 105 kDa[1]. The shedding is actively regulated by a proteolyticprocess [2, 3]. The presence of increased HER-2/neuECD levels in serum is reported to be correlated withhigher tumor burden [4, 5].

The impaired prognosis of patients with tumorsthat overexpress HER-2/neu [6] may be related toresistance to chemotherapy and reduced effectivenessof endocrine treatment. In retrospective analyses ofadjuvant chemotherapy studies with CMF or similarregimen, a reduced disease-free survival was observedfor patients with HER-2/neu overexpressing tumors[7, 8]. Also for endocrine therapy (tamoxifen and othercompounds), experimental data [9] and observationsin patients with primary and metastatic disease sug-gest an adverse outcome in patients with HER-2/neuoverexpression determined by immunohistochemistry[10] or measurement of HER-2/neu ECD levels [11–13]. Regarding the use of anthracyclines in the ad-juvant situation, data suggests an increased responseto anthracycline-containing chemotherapy in patientswith HER-2/neu positive tumors [14–16]. Severalpublications observed an increased efficacy of taxanesin HER-2/neu overexpressing breast cancer [17, 18],a finding that other authors were not able to confirm[19].

The decrease of circulating HER-2/neu ECD dur-ing treatment was found to be associated with responseto therapy in metastatic breast cancer, implicating apotential use of this marker for the monitoring oftherapy [20].

In conclusion, the precise biological role of HER-2/neu overexpression and the clinical implication ofHER-2/neu ECD measurements in patients with breastcancer are still not well defined. Therefore, we ex-amined whether HER-2/neu overexpression determ-ined by HER-2/neu ECD has a prognostic impactand a predictive role for response to chemotherapyin a defined group of patients with metastatic breastcancer. Taxane-containing chemotherapies are cost-intensive and therapy provides toxicity. Therefore, itis also of interest to identify patients who will bene-fit from such therapy. For this purpose, we studiedthe association of HER-2/neu ECD concentration withdisease response and survival in a retrospective anal-ysis of sera from patients treated within a randomizedclinical trial for first line chemotherapy in metastaticbreast cancer comparing a taxane and non-taxane regi-men [21]. In addition, HER-2/neu ECD levels during

treatment were examined to evaluate the potential useof HER-2/neu ECD as parameter to monitor therapyresponse.

Material and methods

Patients

Pretreatment serum was obtained from 103 of 597 pa-tients who participated in a randomized clinical multi-center trial on first-line chemotherapy for metastaticbreast cancer [21]. These patients were randomlyassigned to receive either epirubicin and paclitaxel,60/175 mg/m2 (ET) or epirubicin and cyclophosph-amide, 60/600 mg/m2 (EC) every 21 days (maximumof 10 courses). The median age of the patients was56 years (range 33–73 years). Eighty percent hadvisceral disease. With regard to the primary tumorcharacteristics, 45% of the tumors were undifferen-tiated (grade 3), 35% of the patients were estrogenreceptor negative, 38% of the patients had adjuvantchemotherapy, 4.9% with antracyclines. No signifi-cant differences in patient characteristics between oursubgroup and the whole study population of the trialwere observed (data not shown). Median follow-up ofthe study population was 8.9 months (0.5–36 months).

Serum was collected before therapy (n = 103)and after three courses of therapy in a subset of 46patients. Healthy female blood donors (n = 30, me-dian age 52.4 range 28.6–71.2 years) were used ascontrols to verify the normal values of HER-2/neuECD. Response was evaluated every three courses andclassified according to the UICC criteria.

Quantitative analysis of serumHER-2/neu levels

Serum HER-2/neu was quantified by an commer-cially available ELISA (Oncogene Science, part ofBayer Diagnostics, Cambridge, USA). In brief, serumsamples were diluted 1:50 with sample diluting buffercontaining bovine serum albumin and sodium azide.Diluted samples were dispensed into the 96-well platecoated with the monoclonal capture antibody (NB3)and incubated for 3 h at 37◦C. Wells were washedand the detection antibody (TA1) was added. Theplates were incubated for 60 min at 37◦C, washedand then further incubated with antirabbit immun-globulin/horseradish peroxidase conjugate for 30 min

HER-2/neu extracellular domain in metastatic breast cancer 11

at room temperature. After washing, substrate (o-phenylenediamine) was added for 60 min at room tem-perature. The reaction was stopped with 2.5 N H2SO4and absorbance was read at 490 nm by an automatedplate-reader (Titerscan, Labsystems, Finland). TheHER-2/neu ECD concentration was estimated fromthe standard curve. Each sample, standard and con-trol was analyzed in duplicate. Inter- and intra-assaycoefficient of variation were less than 10%.

HER-2/neu immunohistochemistrywith the monoclonal antibody CB11

For HER-2/neu immunohistochemistry from paraffin-embedded tissue sections from primary tumors, theantibody CB 11 (Novocastra Laboratories, Newcastleupon Tyne, UK) was used at 1:40 dilution. TheVentana Enhanced DAB Detection Kit was used forchromogenic vizualisation of the primary antibodywith a Nexes R staining System (Ventana Medical Sys-tems Tucson, USA). Staining was evaluated accordingto the immunoreactive score proposed by Remmeleand Stegner with values between 0 and 12 [22]. Withthis scoring system, that considers both staining in-tensity and proportion of stained cells, only score 9and 12 were considered to strongly overexpress HER-2/neu, corresponding to a 3+ score in the system nowwidely used as ‘DAKO-score’ [23].

Statistical methods

Progression-free interval and overall survival was es-timated with the method of Kaplan–Meier. The differ-ences in progression-free interval and overall survivalin relation to therapy or HER-2/neu ECD levels werecalculated using the logrank-test and Breslow testrespectively. Correlation of HER-2/neu ECD or thetumor marker CA 15-3 with therapy response wascalculated with the Wilcoxon-test. For the correlationof changes in HER-2/neu ECD levels with therapyresponse, an alteration of serum levels more than5% was regarded as increase or decrease. Correla-tion of HER-2/neu levels to prognostic parameters andmetastatic sites was calculated by using the Mann–Whitney U-Test. Correlation between immunohisto-chemistry and HER-2/neu ECD was calculated usingthe Spearman-correlation coefficient. Unless other-wise stated, an error of less than 5% probability (p ≤0.05) was considered as statistically significant. Allp-values correspond to two-sided tests. For statisticalanalyses, SPSS software (version 11.0) was used.

Results

Correlation of serum HER-2/neu ECDwith HER-2/neu overexpression in primarytumors determined by immunohistochemistry

In a subset of 29 patients, paraffin blocks of theprimary tumors were available for determination ofHER-2/neu overexpression by immunohistochemistrywith the monoclonal antibody CB 11. A significantcorrelation of HER-2/neu protein expression in theprimary tumor with the HER-2/neu ECD concentra-tion in serum at the time of metastatic disease wasobserved (R = 0.39, p = 0.018; Spearman logrank).Nine patients with strong immunohistochemical over-expression (Remmele–Stegner score 9 or 12, corres-ponding to DAKO 3+) had a median HER-2/neu ECDconcentration of 45.2 ng/mL (range 9.6–542.6), eightof these patients had HER-2/neu ECD levels higherthan 15 ng/mL, one below 15 ng/mL. The remaining20 patients with weak or no immunohistochemicalstaining had a median HER-2/neu concentration of9.4 ng/mL (range 5.9–41.5 ng/mL), 18 of these pa-tients had ECD levels below 15 ng/mL, two higherthan 15 ng/mL (Figure 1, Table 1). Using χ2-statistics,the optimal cut-off in serum HER-2/neu ECD levelsfor discrimination between HER-2/neu overexpressors

Figure 1. Correlation of immunohistochemical staining from theprimary tumors with the monoclonal antibody CB 11 and serumHER-2/neu ECD level at the time of metastatic disease (n = 29).Immunohistochemical staining intensity was evaluated according tothe scoring system of Remmele and Stegner as described in Materialand methods. The vertical line (score > 8) indicates strongly posi-tive staining corresponding to a DAKO-score of 3+, the horizontalline indicates the cut-off value for increased HER-2/neu ECD of15 ng/mL.

12 V Müller et al.

Table 1. Correlation of pretherapeutic HER-2/neu ECD levels with clinical characteristics and HER-2/neuexpression in the primary tumors

HER-2/neu ECD <15 ng/mL HER-2/neu ECD ≥15 ng/mL

All patients 64 (63.4%) 37 (36.6%)

Treatment regimen

EC 35 19

ET 29 18

n.s.a,b

Age (years)

<50 19 7

≥50 45 30

n.s.a,b

Estrogen receptor status

Positive 33 20

Negative 24 10

n.s.a,b

Site of metastatic disease

Lung 7 5

Liver 29 24

Loco-regional 13 3

p = 0.05b

CA 15-3 levels

Positive (≥28 kU/L) 18 17

Negative (<28 kU/L) 11 3

p = 0.03b

HER-2/neu immunohistochemistry score from the primary tumor (n = 29)

≥9 8 1

<9 2 18

p < 0.001c

a No statistical significant difference. Information for clinical characteristics was not available for all patients.b Calculated with the U-test Mann–Whitney.c Calculated with the Spearman logrank test.

and non-overexpressors concerning the HER-2/neuexpression in the primary tumors determined by im-munohistochemistry was found at 15 ng/mL (χ2 =17.1, p < 0.0001). This HER-2/neu ECD level wasconsidered as cut-off and is comparable to results ofother publications [12, 20].

Serum HER-2/neu ECD levels in healthyindividuals and in breast cancer patients

For verification of the normal range of HER-2/neuECD concentration, serum levels were determined in30 healthy age-matched female blood donors. HER-2/neu ECD concentrations in this control group werebelow the cut-off of 15 ng/mL in all subjects (range5.5–11.1 ng/mL). In 37.9% of patients in our studygroup, elevated concentrations of the HER-2/neu

ECD (>15 ng/mL) were detected (6–800 ng/mL, me-dian 12.73 ng/mL). A significant difference betweenhealthy controls and metastatic breast cancer patients

Figure 2. HER-2/neu ECD levels in patients with metastatic diseasebefore start of chemotherapy and in healthy controls. The horizontalline indicates the cut-off value of 15 ng/mL.

HER-2/neu extracellular domain in metastatic breast cancer 13

Figure 3. Correlation of HER-2/neu ECD levels before therapy with overall survival. The p-values indicate significance calculated with theBreslow test.

Table 2. Correlation of response to chemotherapy for metastatic disease and serum levels of HER-2/neu ECD

Treatment HER-2/neu ECD Complete or partial response No change of disease status Progression of disease

All patients <15 ng/mL 25 (43.9%)a 21 (36.8%) 11 (19.3%)

≥15 ng/mL 14 (40.0%) 12 (34.3%) 9 (25.7%)

Both groups 39 (42.4%) 33 (35.9%) 20 (21.7%)

ET <15 ng/mL 12 (46.2%) 10 (38.5%) 4 (15.4%)

≥15 ng/mL 9 (50.0%) 6 (33.3%) 3 (16.7%)

Both groups 21 (47.7%) 16 (36.4%) 7 (15.9%)

EC <15 ng/mL 13 (41.9%) 11 (35.5%) 7 (22.6%)

≥15 ng/mL 5 (29.4%) 6 (35.3%) 6 (35.3%)

Both groups 18 (37.5%) 17 (35.4%) 13 (27.1%)

a Percent of patients with HER-2/neu ECD levels lower than 15 ng/mL.

was observed (p < 0.0001; U-test Mann–Whitney)(Figure 2).

Correlation of HER-2/neu ECD levelswith patient characteristics

Patients with visceral metastasis had significantlyhigher levels of HER-2/neu ECD compared with thosepatients suffering from locoregional metastasis (p =0.044; U-test Mann–Whitney). No correlation wasseen between HER-2/neu ECD levels and age orsteroid hormone receptor status (Table 1).

Impact of HER-2/neu ECD levels on survival

Increased levels of pretherapeutic HER-2/neu ECD inserum of the patients with metastatic breast cancerfrom our study had a significant negative impact onoverall survival time (p = 0.0097, Breslow). Thisimpact was only observed in the first year of follow-up and the median survival in the patient group withelevated serum HER-2/neu ECD levels was not dif-ferent from those with non-elevated ECD levels (15.4months vs. 14.7 months) (Figure 3(a)). In a subgroup

analysis, the negative impact of elevated HER-2/neuECD levels on overall survival was significant par-ticularly in patients treated with EC (p = 0.0114logrank, p = 0.0092 Breslow). Best survival wasseen in patients with low ECD-levels and first-line ECtreatment (Figure 3(b)). In patients treated with ET, nosignificant difference for overall survival was observedbetween patients with low and high HER-2/neu ECDlevels (Figure 3(c)).

Predictive value of HER-2/neu ECD levels forresponse to therapy and progression-free interval

No significant correlation between the response tochemotherapy and HER-2/neu ECD levels was ob-served for the entire patient population. With regardto the EC-treated patients with elevated HER-2/neuECD levels, a decreased overall response rate (com-plete and partial responses) of 29.4% was observedcompared to 41.9% in patients with non-elevated ECDlevels (p = 0.059, χ2 test). No differences betweenthe response rates between elevated and non-elevatedlevels were observed in the ET-treated patient group(50.0 vs. 46.2%; Table 2).

14 V Müller et al.

Figure 4. Correlation of HER-2/neu ECD levels before therapy with progression free interval. The p-values indicate significance calculatedwith the Breslow test.

Figure 5. HER-2/neu ECD levels in serum samples before initiation of therapy and after 3 months of therapy. The lines and numbers indicatethe median HER-2/neu ECD values.

We found no influence of pretherapeutic HER-2/neu ECD levels on progression-free-interval in theentire patient group (Figure 4(a)) and the group withlow HER-2/neu ECD levels (Figure 4(b)). However,we found a significantly longer progression-free inter-val if patients with increased HER-2/neu ECD levelswere treated with ET (30 weeks) compared to EC (14weeks; p = 0.0341, Breslow) (Figure 4(c)).

Course of HER-2/neu ECD levels during therapy

HER-2/neu ECD levels in the study group wereelevated in 37.9% of patients before therapy. Inpretherapeutic sera, HER-2/neu ECD levels werecorrelated with the CA 15-3 values (p = 0.03, datanot shown).

In 46 patients we were able to study HER-2/neuECD levels during the course of treatment. Of thesepatients, 42 were evaluable of objective response. Inpatients with complete (CR) or partial remission (PR),the median serum HER-2/neu ECD level droppedfrom 17.1 ng/mL (pretherapeutic) to 11.4 ng/mL afterthree courses of therapy. A decrease of the medianHER-2/neu ECD levels was associated with responseto therapy (p = 0.006, Wilcoxon-test). No signifi-

cant changes of median ECD levels were observed inpatients with NC or PD (Figure 5).

With regard to the patients who had initially ele-vated HER-2/neu ECD levels (ECD > 15 ng/mL) andresponded to therapy (CR or PR; n = 9), the levelsdecreased from a median value of 60.6 to 20.6 ng/mL(p = 0.023, Wilcoxon). In none of these patientswith initially elevated serum HER-2/neu ECD valuesand CR or PR, an increase of HER-2/neu ECD levelsduring therapy was observed.

Discussion

We studied the HER-2/neu ECD concentrations inserum samples from a defined population of patientswith metastatic breast cancer. These patients were re-cruited retrospectively from a randomized clinical trialthat compared two different chemotherapy regimensfirst-line treatment of metastatic breast cancer. The in-clusion criteria of that trial led to a selection of highrisk patients with a high proportion of visceral metas-tasis. Thus, the median survival of these patients wasonly 15.2 months.

In this study population, we found a strong cor-relation between elevated HER-2/neu serum levels

HER-2/neu extracellular domain in metastatic breast cancer 15

at the time of metastatic recurrence and overexpres-sion of HER-2/neu in the primary tumor determinedby immunohistochemistry. Three out of 29 cases inour population showed discordant results. It has beensuggested that the HER-2/neu status determined byimmunohistochemistry or gene amplification mightchange from primary tumors to later occurring meta-static sites in some patients [23, 24]. In addition, theconcentration of serum HER-2/neu ECD is probablynot only dependent on the extent of overexpression inthe single tumor cell but also on the tumor load and theproportion of HER-2/neu receptor molecules that areshed into the circulation. With regard to the correla-tion between overexpression in the primary tumor andelevated HER-2/neu ECD levels in metastatic patients,we calculated the best cut-off at a level of 15 ng/mL.This value corresponds well to the published values[20] and has also been proposed by the manufacturerof the assay [12].

The measurement of HER-2/neu ECD at the timeof metastatic disease could offer an advantage forthe determination of HER-2/neu status since it mightbetter reflect the current situation of HER-2/neu ex-pression in the metastatic tumors. In addition, serumsamples can be easily obtained and repeated determin-ations can be performed. Currently, the most importantindication for HER-2/neu testing is the evaluation fortreatment with the humanized anti-HER-2/neu anti-body trastuzumab. This therapeutic approach has beenshown to be effective for patients with metastaticbreast cancer [25] and is now studied as adjuvanttreatment [26]. First reports indicate that HER-2/neuECD measurement might help to select patients forthis therapy and also to monitor therapy response [20].

Patients from our study with elevated HER-2/neuECD levels at the time of detection of metastasis hada significantly worse prognosis with regard to overallsurvival than those with normal values. This obser-vation confirms results from other authors who havealso seen a worse prognosis in patients with elevatedHER-2/neu ECD levels in metastatic disease [27–30].Kandl et al. first reported that detection of HER-2 inthe serum of 79 breast cancer patients was associ-ated with an impaired prognosis [31]. SubsequentlyWillsher and colleagues reported on 119 breast cancerpatients that HER-2/neu ECD levels are of prognosticsignificance, but no predictive value for response tochemotherapy was observed [5]. In retrospective anal-yses, Hayes et al. [30] and Lipton et al. [12] morerecently also found an unfavorable course of disease inmetastatic breast cancer patients with elevated HER-

2/neu ECD levels. However, most of these studiesexamined patients treated with endocrine therapy.

The prognostic impact of elevated HER-2/neuECD levels in our study is only significant within thefirst year of observation. This might be explained bythe selection of high-risk metastatic patients and bythe fact that patients received different second-lineregimen following the first-line chemotherapy. Par-ticularly in EC treated patients, elevated HER-2/neuECD levels predicted significantly shorter overall sur-vival compared to low ECD levels. Since we did notobserve this effect in ET-treated patients, we conclude,that the taxane treatment might outweigh the potentialprognostic disadvantage of HER-2/neu overexpres-sion. This hypothesis is supported by the observation,that elevated HER-2/neu ECD levels are predictive forlower response rates only in EC-treated patients butnot in those treated with the taxane combination (ET).

The primary endpoint of the original clinical trialin which our patients were treated was the compar-ison between the treatment arms for progression freeinterval and response rates. Patients in this study weretreated with highly efficient anthracycline-containingregimens with or without taxanes. Significant dif-ferences with regard to response, progression freeinterval and overall survival were only found in sub-groups of the entire trial population [21]. Analysesshowed no significant difference between the patientscharacteristics, response and survival of our subgroupand the whole study population (data not shown). Ourstudy for measurement of HER-2/neu ECD levels wasinitiated in order to examine if it is possible to identifya subgroup of patients that shows an increased benefitfrom the taxane therapy.

In our entire patient population, elevated HER-2/neu ECD levels had no predictive impact onresponse rates and progression-free interval. Insubgroup analyses, patients treated with EC showeda decreased response rate if they had elevated HER-2/neu ECD levels compared with low levels while inpatients treated with ET, no differences were observed(Table 2). Patients with increased HER-2/neu ECDlevels had significantly prolonged progression-free in-terval if treated with ET compared to EC (Figure 4(c)).These results suggest the potential use of HER-2/neuECD determination to predict therapy response andthereby to select patients for taxane containing ther-apy. Konecny et al. examined another patient subsetfrom the same chemotherapy trial using FISH analysisfor HER-2/neu gene amplification in the primary tu-mor. In line with our results, they found an increased

16 V Müller et al.

response to taxane containing therapy compared withnon-taxane regimen in patients with HER-2/neu geneamplification [32]. With respect to HER-2/neu over-expression in the primary tumor, for which more datais available, several clinical and experimental reportssuggest that tumors with HER-2/neu overexpressionmay have an improved response to increased levels ofdoxorubicin [15, 16, 33]. The role of taxanes in HER-2/neu positive breast cancer is less clear. Althougha relationship between improved response to taxanetherapy and HER-2/neu overexpression was observedby some investigators [17, 18, 34], contrary resultshave been described [19]. One study examining levelsof circulating HER-2/neu ECD has not demonstrateda relationship between paclitaxel sensitivity and re-sponse: In the context of the ECOG 1193 trial formetastatic breast cancer, Stender et al. assayed forHER-2/neu ECD by a different enzyme immunoassayas used for our study, with 23% of paclitaxel-treatedpatients testing ‘positive’. No association betweenHER-2/neu ECD positivity and clinical response topaclitaxel was noted, but a worse survival was ob-served for the patients with increased levels of HER-2/neu ECD [35]. Also Hayes et al. were not able tofind a predictive value of HER-2/neu ECD measure-ments in patients with metastatic breast cancer [30]. Incontrast, Colomer et al. found also a lower likelihoodof response and impaired overall survival in patientswith increased HER-2/neu ECD levels in a study in-volving 58 patients using the doublet of paclitaxeland doxorubicin [36]. This group also reported sim-ilar findings in patients treated with gemcitabine andpaclitaxel [37]. The different cut-off levels for HER-2/neu ECD positivity and the different assay systemsmight play a role explaining different results of thesestudies.

A decrease of HER-2/neu ECD in serum of pa-tients with metastatic breast cancer correlates withresponse to chemotherapy as also indicated by a reportpublished by Schwartz et al. [38]. This might help tomonitor therapy with HER-2/neu ECD measurements,especially in those patients with elevated HER-2/neuECD levels.

In conclusion, we observed a prognostic impactof HER-2/neu ECD levels. Patients with high HER-2/neu ECD levels showed a shorter overall survival. Inaddition, we found a predictive impact of HER-2/neuECD levels. Patients with elevated HER-2/neu ECDlevels had a lower response rate when treated withEC chemotherapy. Taxane-containing chemotherapy

was able to overcome the negative prognostic impactof elevated ECD levels. Measurement of HER-2/neuECD serum levels can be used for therapy monitoringin HER-2/neu overexpressing disease. Therefore, ourresults add important evidence for the clinical use ofthis serum test.

Acknowledgements

The authors wish to thank Mrs A. Andreas,Mrs K. Baack and Mrs S. Santjer for excellent tech-nical assistance, Professor J. Berger, Institute for Sta-tistics, University Hospital Hamburg-Eppendorf, forhelp with the statistical analysis, and Prof P. Kühnl(Abt. für Transfusionsmedizin University HospitalHamburg) for contributing serum samples.

References

1. Wu JT, Zhang P, Astill ME, Lyons BW, Wu LH: Identifica-tion and characterization of c-erbB-2 proteins in serum, breasttumor tissue, and SK-BR-3 cell line. J Clin Lab Anal 9:141–150, 1995

2. Codony-Servat J, Albanell J, Lopez-Talavera JC, Arribas J,Baselga J: Cleavage of the HER2 ectodomain is a pervanadate-activable process that is inhibited by the tissue inhibitor ofmetalloproteases-1 in breast cancer cells. Cancer Res 59:1196–1201, 1999

3. Molina MA, Codony-Servat J, Albanell J, Rojo F, Arribas J,Baselga J: Trastuzumab (herceptin), a humanized anti-HER2receptor monoclonal antibody, inhibits basal and activatedHER2 ectodomain cleavage in breast cancer cells. Cancer Res61: 4744–4749, 2001

4. Leitzel K, Teramoto Y, Konrad K, Chinchilli VM, VolasG, Grossberg H, Harvey H, Demers L, Lipton A: Elevatedserum c-erbB-2 antigen levels and decreased response to hor-mone therapy of breast cancer. J Clin Oncol 13: 1129–1135,1995

5. Willsher PC, Beaver J, Pinder S, Bell JA, Ellis IO, BlameyRW, Robertson JF: Prognostic significance of serum c-erbB-2protein in breast cancer patients. Breast Cancer Res Treat 40:251–255, 1996

6. Slamon DJ, Clark GM, Wong SG, Levin WJ, Ullrich A,McGuire WL: Human breast cancer: correlation of relapseand survival with amplification of the HER-2/neu oncogene.Science 235: 177–182, 1987

7. Allred DC, Clark GM, Tandon AK, Molina R, Tormey DC,Osborne CK, Gilchrist KW, Mansour EG, Abeloff M, EudeyL: HER-2/neu in node-negative breast cancer: prognostic sig-nificance of overexpression influenced by the presence of insitu carcinoma. J Clin Oncol 10: 599–605, 1992

8. Stal O, Sullivan S, Wingren S, Skoog L, Rutqvist LE,Carstensen JM, Nordenskjold B: c-erbB-2 expression and ben-efit from adjuvant chemotherapy and radiotherapy of breastcancer. Eur J Cancer 31A: 2185–2190, 1995

HER-2/neu extracellular domain in metastatic breast cancer 17

9. Benz CC, Scott GK, Sarup JC, Johnson RM, Tripathy D,Coronado E, Shepard HM, Osborne CK: Estrogen-dependent,tamoxifen-resistant tumorigenic growth of MCF-7 cells trans-fected with HER2/neu. Breast Cancer Res Treat 24: 85–95,1993

10. De Placido S, De Laurentiis M, Carlomagno C, Gallo C,Perrone F, Pepe S, Ruggiero A, Marinelli A, Pagliarulo C,Panico L, Pettinato G, Petrella G, Bianco AR: Twenty-yearresults of the Naples GUN randomized trial: predictive factorsof adjuvant tamoxifen efficacy in early breast cancer. ClinCancer Res 9: 1039–1046, 2003

11. Yamauchi H, O’Neill A, Gelman R, Carney W, Tenney DY,Hosch S, Hayes DF: Prediction of response to antiestrogentherapy in advanced breast cancer patients by pretreatmentcirculating levels of extracellular domain of the HER-2/c-neuprotein. J Clin Oncol 15: 2518–2525, 1997

12. Lipton A, Ali SM, Leitzel K, Demers L, Chinchilli V, EngleL, Harvey HA, Brady C, Nalin CM, Dugan M, Carney W,Allard J: Elevated serum HER-2/neu level predicts decreasedresponse to hormone therapy in metastatic breast cancer. J ClinOncol 20: 1467–1472, 2002

13. Molina R, Jo J, Filella X, Zanon G, Farrus B, Munoz M, LatreML, Pahisa J, Velasco M, Fernandez P, Estape J, Ballesta AM:C-erbB-2, CEA and CA 15.3 serum levels in the early diag-nosis of recurrence of breast cancer patients. Anticancer Res19: 2551–2555, 1999

14. Muss HB, Thor AD, Berry DA, Kute T, Liu ET, Koerner F,Cirrincione CT, Budman DR, Wood WC, Barcos M: c-erbB-2expression and response to adjuvant therapy in women withnode-positive early breast cancer. N Engl J Med 330: 1260–1266, 1994

15. Paik S, Bryant J, Park C, Fisher B, Tan-Chiu E, Hyams D,Fisher ER, Lippman ME, Wickerham DL, Wolmark N: erbB-2 and response to doxorubicin in patients with axillary lymphnode-positive, hormone receptor-negative breast cancer. J NatlCancer Inst 90: 1361–1370, 1998

16. Thor AD, Berry DA, Budman DR, Muss HB, Kute T,Henderson IC, Barcos M, Cirrincione C, Edgerton S, AllredC, Norton L, Liu ET: erbB-2, p53, and efficacy of adjuvanttherapy in lymph node-positive breast cancer. J Natl CancerInst 90: 1346–1360, 1998

17. Baselga J, Seidman AD, Rosen PP, Norton L: HER2 overex-pression and paclitaxel sensitivity in breast cancer: therapeuticimplications. Oncology (Huntingt) 11: 43–48, 1997

18. Gianni L, Capri G, Mezzelani A, Valagussa P, Greco M,Bertuzzi A, Grasselli G, Tarenzi G, A. Moliterni, PilottiS, Bonadonna G: HER-2/neu (HER2) amplification and re-sponse to doxorubicin/paclitaxel (AT) in women with meta-static breast cancer. Proceedings of ASCO Abstract No. 491,1997

19. Sjostrom J, Collan J, von Boguslawski K, Franssila K,Bengtsson NO, Mjaaland I, Malmstrom P, Ostenstad B,Wist E, Valvere V, Bergh J, Skiold-Petterson D, Saksela E,Blomqvist C: C-erbB-2 expression does not predict responseto docetaxel or sequential methotrexate and 5-fluorouracilin advanced breast cancer. Eur J Cancer 38: 535–542,2002

20. Esteva FJ, Valero V, Booser D, Guerra LT, Murray JL,Pusztai L, Cristofanilli M, Arun B, Esmaeli B, Fritsche HA,Sneige N, Smith TL, Hortobagyi GN: Phase II study ofweekly docetaxel and trastuzumab for patients with HER-2-overexpressing metastatic breast cancer. J Clin Oncol 20:1800–1808, 2002

21. Lück HJ, Thomssen C, Untch M, Kuhn W, Eidtmann H,DuBois A, Olbricht S, Möbus V, Steinfeld D, Bauknecht T,Schroeder W, Jackisch C: Multicentric phase III study in firstline treatment of advanced metastatic breast cancer (ABC).Epirubicin/paclitaxel (ET) vs epirubicin/cyclophosphamide(EC). A Study of the AGO Breast Cancer Group. Proceedingsof ASCO Abstract No. 280, 2000

22. Remmele W, Stegner HE: Recommendation for uniform defi-nition of an immunoreactive score (IRS) for immunohis-tochemical estrogen receptor detection (ER-ICA) in breastcancer tissue. Pathologe 8: 138–140, 1987

23. Mass R: The role of HER-2 expression in predicting responseto therapy in breast cancer. Semin Oncol 27 (Suppl. 11):46–52, 2000

24. Tanner M, Jarvinen P, Isola J: Amplification of HER-2/neu andtopoisomerase IIalpha in primary and metastatic breast cancer.Cancer Res 61: 5345–5348, 2001

25. Slamon DJ, Leyland-Jones B, Shak S, Fuchs H, Paton V,Bajamonde A, Fleming T, Eiermann W, Wolter J, PegramM, Baselga J, Norton L: Use of chemotherapy plus a mono-clonal antibody against HER2 for metastatic breast cancerthat overexpresses HER2. N Engl J Med 344: 783–792,2001

26. Leyland-Jones B: Trastuzumab: hopes and realities. LancetOncol 3: 137–144, 2002

27. Fehm T, Maimonis P, Katalinic A, Jager WH: The prognosticsignificance of c-erbB-2 serum protein in metastatic breastcancer. Oncology 55: 33–38, 1998

28. Isola JJ, Holli K, Oksa H, Teramoto Y, KallioniemiOP: Elevated erbB-2 oncoprotein levels in preoperativeand follow-up serum samples define an aggressive diseasecourse in patients with breast cancer. Cancer 73: 652–658,1994

29. Yang H, Wanner IB, Roper SD, Chaudhari N: An optimizedmethod for in situ hybridization with signal amplification thatallows the detection of rare mRNAs. J Histochem Cytochem47: 431–446, 1999

30. Hayes DF, Yamauchi H, Broadwater G, Cirrincione CT,Rodrigue SP, Berry DA, Younger J, Panasci LL, MillardF, Duggan DB, Norton L, Henderson IC: Circulating HER-2/erbB-2/c-neu (HER-2) extracellular domain as a prognosticfactor in patients with metastatic breast cancer: Cancerand Leukemia Group B Study 8662. Clin Cancer Res 7:2703–2711, 2001

31. Kandl H, Seymour L, Bezwoda WR: Soluble c-erbB-2 frag-ment in serum correlates with disease stage and predicts forshortened survival in patients with early-stage and advancedbreast cancer. Br J Cancer 70: 739–742, 1994

32. Konecny G, Thomssen C, Pegram M, Luck HJ, Untch M,Pauletti G, Dandekar U, Ramos L, Kuhn W, Eidtmann H,DuBois A, Olbricht S, Steinfeld D, Moebus V, Jaenicke F,Slamon D: HER-2/neu gene amplification and response to pac-litaxel in patients with metastatic breast cancer. Proceedings ofASCO Abstract No. 88, 2001

33. Harris LN, Yang L, Tang C, Yang D, Lupu R: Induction ofsensitivity to doxorubicin and etoposide by transfection ofMCF-7 breast cancer cells with heregulin beta-2. Clin CancerRes 4: 1005–1012, 1998

34. Baselga J, Norton L, Albanell J, Kim YM, Mendelsohn J:Recombinant humanized anti-HER2 antibody (Herceptin) en-hances the antitumor activity of paclitaxel and doxorubicinagainst HER2/neu overexpressing human breast cancer xeno-grafts. Cancer Res 58: 2825–2831, 1998

18 V Müller et al.

35. Stender M, Neuberg D, Wood W, Sledge G: Correlation ofcirculating c-erbB-2 extracellular domain (HER-2) with clin-ical outcome in patients with metastatic breast cancer (MBC).Proceedings of ASCO Abstract No. 541, 1997

36. Colomer R, Montero S, Lluch A, Ojeda B, Barnadas A,Casado A, Massuti B, Cortes-Funes H, Lloveras B: CirculatingHER2 extracellular domain and resistance to chemotherapy inadvanced breast cancer. Clin Cancer Res 6: 2356–2362, 2000

37. Colomer R, Llombart A, Lluch A, Ojeda B, Barnadas A,Caranana V, Fernandez Y, De Paz L, Guillem V, Montero S,Alonso S: Biweekly paclitaxel and gemcitabine in advancedbreast cancer. Phase II trial and predictive value of HER2

extracellular domain (ECD). Proceeings of ASCO AbstractNo. 373, 2000

38. Schwartz MK, Smith C, Schwartz DC, Dnistrian A, Neiman I:Monitoring therapy by serum HER-2/neu. Int J Biol Markers15: 324–329, 2000

Address for offprints and correspondence: Professor Dr ChristophThomssen, Klinik und Poliklinik für Frauenheilkunde, Universitäts-klinikum Hamburg-Eppendorf, Martinistr. 52, D-20246 Hamburg,Germany; Tel.:+49-40-42803-3510; Fax: +49-40-42803-4355;E-mail: christoph.thomssen@medizin.uni-halle.de