The HER-2/neu Oncogene in Breast Cancer: Prognostic Factor, Predictive Factor, and Target for Therapy Jeffrey S. Ross and Jonathan A.

Transcription

1 The HER-2/neu Oncogene in Breast Cancer: Prognostic Factor, Predictive Factor, and Target for Therapy Jeffrey S. Ross and Jonathan A. Fletcher The Oncologist 1998, 3: The online version of this article, along with updated information and services, is located on the World Wide Web at: Downloaded from by guest on March 3, 2014

2 The HER-2/neu Oncogene in Breast Cancer: Prognostic Factor, Predictive Factor, and Target for Therapy JEFFREY S. ROSS, a JONATHAN A. FLETCHER b a Department of Pathology and Laboratory Medicine, Albany Medical College, Albany, New York; b Department of Pathology, Brigham and Women s Hospital and Harvard Medical School, Boston, Massachusetts, USA Key Words. Breast neoplasms Proto-oncogene c-erbb2 HER-2/neu Prognosis Therapy ABSTRACT The HER-2/neu oncogene encodes a transmembrane tyrosine kinase receptor with extensive homology to the epidermal growth factor receptor. HER-2/neu has been widely studied in breast cancer. In this review, the association of HER-2/neu gene and protein abnormalities studied by Southern and slot blotting, immunohistochemistry, enzyme immunoassays, and fluorescence in situ hybridization with prognosis in breast cancer is studied in depth by review of a series of 47 published studies encompassing more than 15,000 patients. The relative advantages of gene amplification assays and frozen/fresh tissue immunohistochemistry over paraffin section immunohistochemistry are discussed. The significance of HER-2/neu overexpression in ductal carcinoma in situ and the HER-2/neu status in uncommon female breast conditions and male breast cancer are also considered. The potential value of HER-2/neu status for the prediction of response to therapy in breast cancer is presented in the light of a series of recently published studies showing a range of impact on the outcome of INTRODUCTION The proto-oncogene HER-2/neu (C-erbB-2) has been localized to chromosome 17q and encodes a transmembrane tyrosine kinase growth factor receptor. The name for the HER-2 protein is derived from Human Epidermal growth factor Receptor, as it features substantial homology with the epidermal growth factor receptor (EGFR) [1-2]. HER-2/neu gene amplification has been associated with the development of breast cancer in animal models [1]. The HER-2/neu protein is a component of a four-member family of closely related growth factor receptors, including EGFR or HER-1 (erb-b1); HER-2 (erb-b2); HER-3 (erb-b3) and HER-4 (erb-b4) (Table 1) [3]. In addition to its association with disease patients treated with hormonal, cytotoxic, and radiation therapies. The evidence that HER-2/neu gene and protein abnormalities in breast cancer predict resistance to tamoxifen therapy and relative sensitivity to chemotherapy regimens including adriamycin is presented. The review will also evaluate the status of serum-based testing for circulating the HER-2/neu receptor protein and its ability to predict disease outcome and therapy response. In the final section, the review will briefly present preliminary data concerning the use of antibody-based therapies directed against the HER-2/neu protein and their potential to become a new modality for breast cancer treatment. The recently presented phase III clinical trial evidence that systemic administration of anti-her2 antibodies (Herceptin ), alone and in combination with cytotoxic chemotherapy in patients with HER-2/neu overexpressing primary tumors, can increase the time to recurrence and overall response rates in metastatic breast cancer is reviewed. The Oncologist 1998;3: Table 1. The HER (ERB) gene family Gene HER-1 (c-erbb-1) HER-2 (c-erbb-2) HER-3 (c-erbb-3) HER-4 (c-erbb-4) Ligand Epidermal growth factor Transforming growth factor alpha Beta cellulin Heparin-binding growth factor Amphiregulin? (Heregulin) Heregulin neu differentiation factor Heregulin neu differentiation factor Correspondence: Jeffrey S. Ross, M.D., Department of Pathology and Laboratory Medicine, Albany Medical College, Mail Code 81, 47 New Scotland Avenue, Albany, New York 12208, USA. Telephone: ; Fax: ; jeffrey_ross@ccgateway.amc.edu Accepted for publication June 9, AlphaMed Press /98/$5.00/0 The Oncologist 1998;3:

3 Ross, Fletcher 238 outcome in gastrointestinal, pulmonary, genitourinary and other neoplasms, amplification of the HER-2/neu gene or overexpression of the HER-2/neu protein has been identified in from 10% to 34% of breast cancers [4-50]. To date, a wide variety of clinical studies, including more than 15,000 patients (Table 2), have evaluated the relationship between HER-2/neu abnormalities and breast cancer outcome, including more than 47 publications concerning the association of gene and/or protein abnormalities with prognosis [4-50]. The techniques used to evaluate HER-2/neu status in breast cancer have included gene-based assays such as Southern and slot blotting, polymerase chain reaction methods, and more Table 2. Summary of selected studies on the correlation of HER-2/neu gene and protein abnormalities with prognosis in breast cancer Impact on prognosis Number/ 1st Author Specimen Method univariate/ year [reference] type/(# cases) of analysis multivariate Comment 1, 1987 Slamon [4] Frozen (189) Southern blot Yes Yes HER-2 amplification predicted overall survival and time to relapse. 2, 1988 Berger [5] Frozen Southern blot Yes - HER-2 protein by IHC correlated with node status and paraffin (51) IHC tumor grade. 3, 1988 van de Vijver [6] Paraffin (189) IHC No No Correlated with size, comedocarcinoma. 4, 1990 Heintz [7] Frozen (50) Southern No No Mitoses, ER/PR negative. analysis 5, 1990 Tsuda [8] Paraffin (176) Southern analysis Yes No Impact on prognosis absorbed by grade. 6, 1990 Borg [9] Fresh (300) Western Yes No Gene amplification predicted relapse and death in node slot blot negative on univariate only. 7, 1990 Paik [10] Paraffin (292) IHC Yes Yes HER-2 protein overexpression second only to nodal status in predicting outcome. 8, 1991 Battifora [11] Paraffin (245) IHC Yes Yes Significant only for low-grade, low-stage cases. 9, 1991 Kallioniemi [12] Paraffin (319) IHC Yes Yes Independent predictor in node-negative and positive. 10, 1991 Clark [13] Fresh (362) Slot blot No No Amplification marginally predictive in node-positive only. 11, 1991 Lovekin [14] Paraffin (782) IHC Yes Yes HER-2 overexpression predicted outcome in early and advanced cases; grade more significant predictor. 12, 1991 McCann [15] Paraffin (314) IHC Yes Yes Predicted outcome in node-negative and node-positive. 13, 1991 Dykens [16] Paraffin (187) IHC Yes Yes Overexpression predicted shortened survival in node-negative, but not in node-positive patients. 14, 1991 Rilke [17] Paraffin (1,210) IHC Yes No Protein overexpression predicted outcome in node-positive only. 15, 1991 Winstanley [18] Paraffin (465) IHC Yes Yes HER-2 protein staining independently predicted survival. 16, 1991 O Reilly [19] Paraffin (172) IHC Yes Yes HER-2 protein overexpression predicted outcome in node-positive, but not in node-negative disease. 17, 1991 Patterson [20] Paraffin (115) Slot blot Yes Yes Gene amplification predicted disease-free interval in node-negative patients. 18, 1992 Toikkanen [21] Paraffin (209) IHC Yes Yes Protein overexpression predicted shortened survival in node-positive patients. 19, 1992 Molina [22] Paraffin (301) IHC Western Yes Yes Both methods of protein levels predicted worse overall survival. 20, 1992 Noguchi [23] Paraffin (234) IHC Yes No Only nodal status correlated independently. 21, 1992 Allred [24] Paraffin (613) IHC Yes - Significant for node-negative low risk only. 22, 1992 Babiak [25] Paraffin (78) Slot blot Yes - Only when combined with aneuploid status. 23, 1992 Tiwari [26] Frozen (61) Southern Yes - HER-2 amplification associated with nodal metastasis. 24, 1992 Gusterson [27] Paraffin (1,506) IHC Yes Yes Predicted outcome in node-positive only. 25, 1993 Bianchi [28] Paraffin (230) IHC Yes No Only diffuse intense staining correlated. 26, 1993 Press [29] Paraffin (210) IHC Yes Yes Predicts relapse in node-negative. Continued on next page.

4 239 The HER-2/neu Oncogene in Breast Cancer Table 2. Summary of selected studies on the correlation of HER-2/neu gene and protein abnormalities with prognosis in breast cancer (continued) Impact on prognosis Number/ 1st Author Specimen Method univariate/ year [reference] type/(# cases) of analysis multivariate Comment 27, 1993 Seshadri [30] Fresh (1,056) Slot blot Yes Yes Independent for both node-negative and positive. 28, 1993 Descotes [31] Frozen (298) Slot blot Yes Yes With PR-negative, node status. Southern 29, 1994 Giai [32] Paraffin (159) Western blot Yes Yes Independent predictor in node-negative. 30, 1994 Muss [33] Paraffin (442) IHC Yes - Predicts response to chemotherapy. 31, 1994 Tetu [34] Paraffin (888) IHC Yes Yes Survival in node-positive, only membranous pattern correlates. 32, 1994 Hartmann [35] Paraffin (340) IHC Yes No Predicts node status. 33, 1994 Jacquemeier [36] Paraffin (81) IHC No No No correlation with therapy response. 34, 1994 Marks [37] Paraffin (230) IHC Yes Yes p53 and node status also independent. 35, 1995 Rosen [38] Paraffin (440) IHC No No Medullary carcinoma is negative. 36, 1995 Quenel [39] Paraffin (942) IHC Yes Yes HER-2 predicted relapse-free and metastasis-free survival. 37, 1996 Sundblad [40] Paraffin (271) IHC Yes Yes Node-positive, HER-2/neu-negative, CEA-positive and bcl-2 positive had the best prognosis. 38, 1996 O Malley [41] Paraffin (107) IHC Yes Yes Both p53 and HER-2/neu staining independently predicted outcome. 39, 1996 Hieken [42] Paraffin; IHC No No HER-2 protein by ELISA, not IHC predictive. cytosol (100) ELISA Yes Yes 40, 1996 Xing [43] Paraffin (37) FISH Yes - HER-2 amplification more predictive than nodal status in young patients. 41, 1997 Dittadi [44] Cytosol (115) ELISA Yes Yes HER-2 protein levels predicted disease-free survival. 42, 1997 Fernandez-Acenero Paraffin (112) IHC Yes No Only TNM stage-independent predictor. [45] 43, 1997 Eissa [46] Cytosol (100) ELISA Yes Yes HER-2 strongest predictor of recurrence in node-negative. 44, 1997 Charpin [47] Frozen (148) IHC Yes Yes Overall and disease-free survival independent of nodal status. 45, 1997 Press [48] Paraffin (324) FISH Yes Yes HER-2/neu gene amplification predicts recurrence and death. 46, 1998 Ross [49] Paraffin (224) FISH Yes Yes 3-tiered amplification system. 47, 1998 Depowski [50] Paraffin (145) FISH Yes Yes HER-2 amplification predicted death independent of nodal status. IHC = immunohistochemistry; ELISA = enzyme-linked immunosorbent assay; FISH = fluorescence in situ hybridization; ER = estrogen receptor; PR = progesterone receptor; CEA = carcinoembryonic antigen. recently, in situ hybridization featuring both fluorescent and nonfluorescent techniques [4-50]. Qualitative and quantitative HER-2/neu protein measurements have been performed by immunohistochemistry on frozen and archival tissues, Western blotting, and enzyme immunoassays (ELISA) [4-50]. The results of the breast cancer outcome studies have not been uniform, and a significant discordance between HER- 2/neu abnormality detection methods has been described. In particular, when immunohistochemistry has been employed on formalin-fixed paraffin-embedded specimens, there has been a significant lack of agreement with gene amplification detection results on the same tumors and inconsistent success at predicting disease prognosis. In this review, the prognostic significance of HER-2/neu abnormalities in breast cancer is considered and the various methods used to detect HER-2/neu abnormalities discussed in detail. The significance of HER- 2/neu abnormalities in ductal carcinoma in situ is also considered, and the potential use of HER-2/neu status in the prediction of response to therapy is presented. A separate section considers the status of serum-based testing for the HER- 2/neu protein for the detection and monitoring of breast cancer. The final section of the review presents the results of recent

5 Ross, Fletcher 240 studies of novel breast cancer therapies directed against the HER-2/neu gene and protein, including the use of anti-her2 antibodies (Herceptin ) in the treatment of metastatic disease. HER-2/neu ABNORMALITIES AND PROGNOSIS IN INVASIVE BREAST CANCER A wide variety of morphology-based and molecularbased prognostic factors and tumor markers have been studied as to their potential to predict disease outcome in breast cancer [51-53]. A partial list of prognostic factors commonly used and under consideration for use in the care of breast cancer patients is included in Table 3. Following the original study by Slamon and coworkers [4] in 1987, many investigators have considered the prognostic potential of the HER-2/neu gene and protein in breast cancer. A synopsis of 47 published studies evaluating more than 15,000 patients with breast cancer for abnormalities of the HER-2/neu gene and protein is presented in Table 2. The original study published by Slamon et al. in 1987 [4] found that HER-2/neu gene amplification determined by Southern analysis independently predicted time to disease relapse and overall survival in 189 patients with breast cancer. Subsequently, Berger et al. reported that HER-2/neu protein overexpression is determined by immunohistochemistry correlated with lymph node status and breast cancer tumor grade [5]. However, an immunohistochemical study Table 3. Prognostic factors in breast cancer Morphology-based Tumor type Tumor size Tumor grade (includes mitotic figure count) Vascular invasion Resection margin status Extent of in situ component Tumor vessel density (requires immunostain) Lymph node status Non-morphology-based molecular markers Estrogen and progesterone receptors Cell proliferation index (immunostains) S phase by flow or image cytometry DNA ploidy by flow or image cytometry Growth factors Oncogenes (HER-2/neu, myc, ras) Tumor suppressor genes (p53) Proteases (cathepsin D) Cell cycle regulators (cyclins, cyclin-dependent kinases) Plasminogen system (PAI-1, upai) by van de Vijver et al. in 1988 [6] failed to show a significant association of HER-2/neu protein immunostaining with disease outcome. In 1990, Heintz et al. reported that the HER- 2/neu gene copy number by Southern analysis correlated with mitotic activity and negative estrogen and progesterone receptor status but could not link gene expression with disease outcome [7]. During the next seven years, however, most large studies found substantial correlation of HER- 2/neu status with disease outcome. This was especially true for studies using either gene-based analysis or immunohistochemistry with anti-her-2/neu antibodies having sensitivity in processed tissues and restricting positive cases to those with a characteristic membranous staining pattern (Fig. 1), [10-12, 14-35, 37, 39-50]. In a 1994 study on 442 patients, Muss et al. [33] reported a significant predictive value for HER-2/neu protein overexpression on the response to cytotoxic chemotherapy and overall disease outcome. In 1996, a study of lymph node negative breast cancer by O Malley and coworkers using a semiquantitative immunohistochemical assessment of HER-2/neu protein levels found an independent prediction with both overall and disease-free survival on multivariate analysis [41]. In another recent study of HER-2/neu protein levels, Hieken et al. found that HER-2/neu protein overexpression measured by an ELISA technique on fresh tumor cytosols independently predicted disease outcome in breast cancer [42]. Another two recent studies reporting adverse prognosis in patients with elevated HER-2/neu protein in tumor cytosols by the ELISA method include the 1997 publications on 115 patients by Dittadi and coworkers [44] and on 100 patients by Eissa et al. [46]. The ELISA method has also been used to measure serum HER-2/neu protein levels (see below). Several immunohistochemical studies found significant correlation between HER-2/neu protein immunoreactivity and disease outcome on univariate analysis (see section on Figure 1. Overexpression of the HER-2/neu protein by immunohistochemistry. Note the positive membranous staining pattern creating a weaved basket appearance (peroxidase-antiperoxidase).

6 241 The HER-2/neu Oncogene in Breast Cancer comparison of HER-2/neu gene and protein detection techniques below) but were unable to demonstrate independent predictive status for immunoreactivity on multivariate analysis [23, 35, 45]. Gene-based studies of HER-2/neu gene amplification in breast cancer include a negative study by Heintz et al. [7] and a 1990 study by Tsuda et al. [8] in which a prognostic impact of the HER-2/neu gene copy number on univariate analysis was absorbed by tumor grade on multivariate analysis. In 1992, however, a study by Babiack et al. [25] reported that the HER-2/neu gene copy number and aneuploidy were predictive of outcome on univariate analysis, and Tiwari et al. reported that gene amplification was associated with nodal metastases [26]. In 1993, Seshadri et al. [30] showed univariate and multivariate significance for HER-2/neu gene amplification as determined by slot-blotting in the prediction of disease recurrence among 1,056 node-negative and node-positive patients. Descotes et al. [31] (using both Southern and slot-blotting methodologies) also reported independent adverse prognostic significance for HER-2/neu gene amplification cases. Given that Southern and slot-blotting procedures are expensive and time-consuming and require fresh or frozen tissue, the fluorescence in situ hybridization (FISH) technique was implemented to measure HER-2/neu gene copy number on formalin-fixed archival specimens (Fig. 2). In two previous studies, the FISH method was found to be more sensitive than Southern blotting for the detection of HER-2/neu gene amplification [48, 54]. The FISH technique has out-performed a number of solid matrix blotting techniques designed to detect HER-2/neu DNA and RNA as well as immunohistochemistry designed to detect HER-2/neu protein in formalin-fixed paraffin-embedded tissues [48]. The FISH technique has been described as a rapid, reproducible, and extremely reliable method of detecting HER-2/neu gene amplification [48]. In addition, FISH can readily be performed on archived paraffin blocks stored for long periods and has been successfully applied to fine-needle aspiration biopsies [55]. In a study using the FISH method, Xing et al. reported that HER-2/neu gene amplification was more predictive than lymph node status and was the strongest independent predictor of outcome in breast cancer [43]. The recently published study by Press et al. [48] successfully utilized FISH on archived specimens to predict short- and long-term outcome in node-negative breast cancer. This study employed a twocategory system of HER-2/neu amplification status: four or fewer signals/nucleus = unamplified, and greater than four signals/nucleus = amplified. A recent study derived from a subset of the cases reported by Press et al. [48] features a three-tiered amplification scoring system with an equivocal borderline amplified group featuring greater than three but fewer than 10 signals/nucleus [49]. The results of the threetiered system are similar to that of the two-tiered system and confirm the significant association of HER-2/neu gene amplification with early recurrence, recurrence at any time, and disease-related death in node-negative breast cancer initially treated by surgery alone (Fig. 3). Both the two-tier and the three-tier amplification scoring systems similarly found the adverse impact of HER-2/neu gene amplification to be independent of tumor size, grade, and estrogen receptor status [48-49]. In another study using a two-tier system HER-2/neu gene amplification again predicted disease-related death independent of the original nodal status in a combined series of node-negative and node-positive patients [50]. Figure 2. Amplification of the HER-2/neu gene detected by fluorescence in situ hybridization. Note the numerous bright intranuclear signals seen individually and in clusters (Biotinylated HER-2/neu probe [Oncor, Inc.; Gaithersburg, MD];avidin anti-avidin detection, DAPI counterstain). Figure 3. Kaplan-Meier survival curve for the prediction of disease-related death in lymph node-negative breast cancer by amplification of the HER- 2/neu oncogene detected by fluorescence in situ hybridization. Patients with amplified tumors had a significantly increased risk of dying from their disease compared with patients in whom primary tumors were not amplified (p < ).

7 Ross, Fletcher 242 Table 4. Summary of studies of HER-2/neu gene and protein in the prediction of prognosis in breast cancer Correlation status Number of studies (%) Number of cases (%) All cases None 6 (13%) 1,222 (8%) Univariate 13 (28%) 3,884 (25%) significance only Independent on 28 (60%) 10,142 (67%) multivariate analysis Totals 47 (100%) 15,248 (100%) Cases with no correlation Immunohistochemistry-based 4 (67%) 810 (66%) Gene-based 2 (33%) 412 (34%) Totals 6 (100%) 1,222 (100%) Cases with univariate but not multivariate significance Immunohistochemistry-based 5 (83%) 2,126 (92%) Gene-based 1 (16%) 176 (8%) Totals 6 (100%) 2,302 (100%) COMPARISON OF METHODS OF DETECTION OF HER-2/neu ABNORMALITIES Of the 47 studies considering 15,248 patients listed in Table 2, 30 (64%) employed immunohistochemistry (IHC) on fixed paraffin-embedded tissues to measure HER-2/neu protein overexpression. As seen in Table 4, 6 (13%) of the studies encompassing 1,222 (8%) of the patients showed no correlation between HER-2/neu status and outcome; 13 (28%) of the studies encompassing 3,884 (25%) of the patients reported prognostic significance on univariate analysis only; and 28 (60%) of the studies encompassing 10,142 (67%) of the patients reported independent prognostic significance on HER-2/neu gene or protein multivariate analysis. Of the six published studies involving 1,222 patients which did not link HER-2/neu abnormalities with prognosis, four (67%) studies of 810 (66%) patients employed IHC on paraffin-embedded tissues as the HER-2/protein detection technique and two (33%) studies involving 412 (34%) patients used gene-based assays. In addition, of the six studies of 2,302 patients in which HER-2/neu abnormalities predicted outcome only on univariate but not on multivariate statistical analysis, five (83%) were IHC-based and one (16%) was based on gene amplification measurements. The prognostic and predictive potential of HER-2/neu protein immunostaining in breast cancer has recently been reviewed by Allred and coworkers [56]. Examples of the IHC-based noncorrelating studies (Table 2) include the 1994 publication on 81 patients by Jacquemeier et al. [36], showing no association of HER-2/protein staining with response to breast cancer therapy and the 1994 report by Rosen et al. [38] on 440 patients which did not relate immunoreactivity with disease outcome. Examples of gene-based noncorrelating studies include the 1990 Southern analysis of 50 cases by Heintz et al. [7], in which amplification correlated with proliferation and estrogen receptor negative status only, and the 1991 study on 362 patients by Clark et al. [13] in which HER- 2/neu gene amplification detected by slot blotting was marginally predictive in node positive patients only. Although only four studies reported to date used the FISH-based gene amplification assay [43, 48-50], four of four (100%) correlated HER-2/neu with outcome on univariate analysis and three of three (100%) showed independent significance when multivariate analysis was performed. The various methods of determining HER-2/neu gene and/or protein abnormalities in breast cancer specimens are compared in Table 5. Immunohistochemistry has been the Table 5. Comparison of methods of detection of abnormalities of the HER-2/neu gene and protein IHC- IHC- Method Frozen tissues Paraffin tissues ELISA Southern blot FISH Abnormality detected Protein Protein Protein Gene Gene overexpression overexpression overexpression amplification amplification On slide method Yes Yes No No Yes Affected by dilution of target No No Yes Yes No Affected by storage No Yes No No No Can be performed on needle biopsies and FNAs No Yes No No Yes Can be performed on tumors <1 cm Difficult Yes No No Yes Standardized technique No No No No Yes Sensitivity High Good High Good High Specificity High Good High High High FDA-approved No No No No Yes

8 243 The HER-2/neu Oncogene in Breast Cancer predominant method utilized. However, there are significant issues that can impact on IHC, especially when performed on archival, fixed, paraffin-embedded tissues. Many laboratories perform the staining on referred specimens and cannot control the time and nature of tissue fixation, the method of tissue processing, or the temperature of the paraffin embedding procedure, all of which can influence HER-2/neu protein antigen loss. Prolonged storage can also be a problem, and significant loss of tumor-marker immunostaining intensity has been identified, particularly when specimens are stored as unstained slides [57]. The impact of the fixative has been considered and shown to have a significant impact on HER-2/neu immunostaining [58]. Using cell line controls, different antibodies have differing staining patterns depending on how the cells were fixed [58]. Studies of the various commercially available antibodies have also demonstrated a wide variety of sensitivity and specificity for fixed, paraffin-embedded tissues. In a study by Busmanis et al. [59], a panel of six antibodies showed a wide variation in staining patterns including occasional cytoplasmic immunoreactivity (a pattern considered to be nonspecific by most investigators). In a study of a large panel of antibodies, Press et al. similarly reported a wide range of detection rates using a large tissue block containing multiple breast tumors [60]. The use of nonstandardized antigen retrieval (amplification) techniques further compounds the problem and introduces the potential for false-positive staining. The lack of an agreed-upon scoring system for interpreting HER-2/neu protein IHC is another significant issue. Recent attempts to reach consensus on immunohistochemical staining interpretation show promise for dealing with this issue [61]. It should be mentioned that IHC on frozen sections has shown substantial correlation with HER-2/neu gene-based assays [48] and would be an ideal method of detection if not obviously limited by the general lack of availability of fresh or frozen material for this approach. The ELISA technique, when performed on tumor cytosols made from fresh tissue samples, avoids the potential antigen damage associated with fixation, embedding, and uncontrolled storage. In the three published studies including 315 patients listed in Table 2, ELISA-based measurements of HER-2/neu protein in tumor cytosols have uniformly correlated with disease outcome [39, 41, 43]. However, the small size of breast cancers diagnosed in an era of enhanced screening generally precludes tumor tissue ELISA methods because insufficient tumor tissue is available to produce a cytosol. Western blotting can also detect HER-2/neu protein overexpression in both tumor cytosols and archival tissues but is generally cumbersome and impractical for routine specimens. Southern and slot blotting were the first gene-based HER- 2/neu detection methods used in breast cancer specimens. These methods can be significantly hampered when tumor cell DNA extracted from the primary carcinoma sample is diluted by DNA from benign breast tissue and inflammatory cells. The FISH technique allows simultaneous morphologic assessment, such that evaluation of gene amplification can be restricted to invasive carcinoma cells. This approach has proven to be more sensitive than Southern analysis for the detection of HER-2/neu abnormalities in breast cancer [48]. Of the 15 gene-based studies listed in Table 2, the two (13%) noncorrelating studies used Southern (50 patients) and slotblotting (362 patients) methods. As mentioned above, the FISH-based assays of HER-2/neu gene amplification have uniformly predicted an adverse disease outcome. In summary, both HER-2/neu gene amplification and protein overexpression have been associated with an adverse outcome in breast cancer, with gene-based methods and protein detection on fresh or frozen samples obtaining the most consistent results. Many of the published studies have used a relatively short clinical follow-up period for a disease prone to late recurrences. Thus, abnormalities of HER-2/neu may actually identify patients at greater risk for early disease relapse, and evidence confirming the ability of the marker to predict overall relapse rates must await continuing long-term studies of tumor outcome. HER-2/neu ABNORMALITIES IN DUCTAL CARCINOMA IN SITU (DCIS) HER-2/neu protein overexpression was first reported in in situ breast cancer by van de Vijver et al. in 1988 and was associated with the comedocarcinoma variant [6]. As seen in Table 6, numerous subsequent studies confirmed this relationship between HER-2/neu protein immunoreactivity and the high-grade and comedo subtypes of DCIS as well as with other unfavorable prognostic factors [62-73]. HER-2/neu protein expression has been associated with DCIS extent [68, 73], negative staining for hormone receptor proteins [65, 67, 71], high cell proliferation rates [64, 67], DNA aneuploidy [65], p53 protein overexpression [64, 67, 71], and association with invasive disease [63]. Given the increased incidence of the diagnosis of DCIS on breast biopsies generated by the detection of small lesions on mammography, there has been significant recent interest in prognostic markers that could further subtype the disease and guide follow-up therapy. As seen in Table 6, HER-2/neu abnormalities appear to identify a particularly virulent form of DCIS featuring a tendency for a high nuclear grade and comedo-type necrosis. Further evaluation of the HER-2/neu status in these lesions appears warranted to confirm whether this marker can be clinically useful in stratifying patients into low-risk groups which may be followed conservatively and high-risk groups that may require extensive post-biopsy surgical procedures to prevent recurrence and to rule out invasive disease with an aggressive phenotype.

9 Ross, Fletcher 244 Table 6. HER-2/neu abnormalities in DCIS HER-2 1st Author Detection Year [reference] Cases method Observations 1990 Lodato [62] 33 IHC HER-2 protein staining positive in comedo type and negative in cribiform and micropapillary types Allred [63] 59 IHC DCIS with HER-2 protein staining is more likely to be associated with invasive component Zafrani [64] 127 IHC Comedo and high-grade DCIS associated with high proliferation index, p53 positive and HER-2 positive status Brower [65] 35 IHC Comedo DCIS associated with high S phase, aneuploidy, ER/PR negative and HER-2 positive Leal [66] 40 IHC Aneuploidy and positive HER-2 protein staining correlated with high tumor grade Bobrow [67] 105 IHC Comedo DCIS associated with high-grade, high S phase, p53 positive, PR negative, and HER-2 positive status DePotter [68] 76 IHC HER-2 protein staining correlated with large cell subtype and extent of disease Inaji [69] 59 IHC HER-2 protein staining correlated with comedo DCIS and inversely correlated with ps2 staining Albonico [70] 50 IHC HER-2 protein staining correlated with comedo DCIS subtype Bose [71] 43 IHC Comedo DCIS associated with positive p53 and HER-2 and negative ER and PR Moreno [72] 94 IHC HER-2 protein staining correlated with high-grade DCIS Mack [73] 70 IHC HER-2 protein staining correlated with size of lesion and histologic grade. IHC = immunohistochemistry; ER = estrogen receptor; PR = progesterone receptor; DCIS = ductal carcinoma in situ. HER-2/neu ABNORMALITIES IN OTHER TYPES OF BREAST MALIGNANCIES AND PROLIFERATIVE DISEASE HER-2/neu gene expression has generally not been specifically implicated in the progression or prognosis assessment of lobular breast cancer [74]. HER-2/neu protein overexpression has been a consistent feature of both mammary and extramammary Paget s disease [75]. Studies of ductal carcinoma in situ and Paget s disease extension patterns have led some researchers to suggest that the HER-2/neu protein may function both as a growth factor receptor and as a cell motility factor. HER-2/neu protein overexpression has been associated with an absence of neuroendocrine breast cancer differentiation [76]. Immunohistochemical staining for HER- 2/neu protein has varied in reports considering benign breast disease, such as fibrocystic mastopathy and fibroadenoma. In general, when staining interpretation is limited to a membranous staining pattern, the staining of benign or proliferative breast tissue both in cancerfree patients and patients with adjacent infiltrating and in situ breast carcinomas has been absent to weak in intensity [77-79]. When Southern blot analysis was performed on both benign female breast tissue and male breast tissue showing gynecomastia, no significant amplification was identified [79]. HER-2/neu STATUS IN MALE BREAST CARCINOMA Male breast cancer comprises approximately 0.05% of new breast cancer diagnoses per year in the United States (1,000 new cases). A variety of prognostic markers have been studied in male breast cancer, including routine histologic parameters of size, grade, and lymph node status, as well as newer molecular markers of disease progression. In a study by Gattuso and coworkers [80], HER-2/neu protein overexpression measured by immunohistochemistry was present in 35% of male breast carcinoma patients and emerged as a statistically significant predictor of disease outcome on univariate analysis. Similarly, a study by Joshi and coworkers in 1996 found HER-2/neu protein immunostaining to predict shortened survival in male breast cancer [81]. However, three additional IHC studies found a similar range of overexpression as seen in female breast cancer but could not correlate immunostaining with disease outcome [82-84]. In another immunohistochemical study of HER-2/neu expression in male breast carcinoma, the expression rate was found to be half (17%) of that in women (33%) [85]. In the latter study, it was commented that male breast carcinomas are often characterized by large tumor cells similar to the type associated with HER- 2/neu gene amplification in women. Although male breast carcinoma is a relatively rare form of malignancy, it appears that further studies of HER-2/neu gene and protein status in male breast carcinoma are warranted to determine whether this

10 245 The HER-2/neu Oncogene in Breast Cancer oncogene has clinical utility for the management of patients with this disease. HER-2/neu STATUS AND THE PREDICTION OF RESPONSE OF BREAST CANCER TO THERAPY In addition to the considerable interest in the HER-2/neu gene and protein as prognostic factors, investigators have also evaluated potential applications of HER-2/neu testing for predicting response of breast cancer to therapy. However, compared with the prognostic studies, the reports of therapy response prediction are fewer, involve smaller patient groups, and have been generally less well controlled [32, 33, ]. Table 7 lists 20 published studies involving more than 4,500 patients, tumor cell lines, and xenografts that used a variety of detection techniques to test the relationship between HER- 2/neu abnormalities and response to therapy. The most common association of HER-2/neu protein overexpression with therapy response, reported in nine of the studies, is the apparent resistance of these tumors to hormone therapy alone [86-88, 90, 92, 97, 99]. Several studies found that HER- 2/neu-overexpressing tumors were specifically resistant to tamoxifen therapy [92, 97]. However, in a recent IHC-based Table 7. Studies of the prediction of HER-2/neu abnormalities on the response of breast cancer to therapy HER-2/neu 1st Author Specimen detection Year [reference] type (# cases) method Findings 1992 Wright [86] Paraffin (72) IHC HER-2/neu protein overexpression predicts lack of response to hormone therapy Gusterson [27] Paraffin (1,506) IHC HER-2/neu protein overexpression predicts lack of response to CMF chemotherapy Muss [33] Paraffin (442) IHC Overexpression associated with increased chemosensitivity to adriamycin Giai [32] Paraffin (159) Western Overexpression predicted reduced response to tamoxifen alone and CMF blot chemotherapy Leitzel [87] Serum (300) EIA High serum levels of HER-2/neu protein predicted lack of response to hormone therapy Berns [88] Cytosols (259) Southern HER-2/neu gene amplification predicted lack of response to hormonal therapy. analysis 1995 Stahl [89] Frozen tissues Flow HER-2/neu protein overexpression predicted poor response to CMF chemotherapy (152) cytometry versus radiation Archer [90] Paraffin (92) IHC Neither HER-2, ras, p21 or p53 protein levels predicted response rates to hormonal therapy Yu [91] Cell lines Transfection Overexpression of HER-2/neu protein conferred resistance to taxol Carlomagno [92] Paraffin (145) IHC HER-2/neu protein overexpression correlated significantly with resistance to tamoxifen Haffty [93] Paraffin (40) IHC HER-2/neu protein overexpression predicted local recurrence in patients receiving surgery and radiation only Revillion [94] Serum (33) ELISA Serum HER-2/neu protein levels at diagnosis did not predict response rates to chemotherapy Baselga [95] Paraffin IHC HER-2 overexpression associated with three times greater sensitivity to paclitaxel Pegram [96] Cell lines; Transfection No significant impact in vitro or in vivo (athymic mice) on drug sensitivity. xenografts 1997 Newby [97] Paraffin (155) IHC HER-2 protein staining predicted failure of response to tamoxifen therapy Yamauchi [99] Serum (94) ELISA High HER-2 protein serum levels predict lack of response to hormonal therapy Burke [98] Not stated Not stated Her-2/neu significantly predicted response to adjuvant chemotherapy and (226) radiotherapy using an artificial neural network analysis system Elledge [100] Paraffin (205 ) IHC HER-2/neu protein overexpression did not predict tamoxifen resistance when ER status was included Bianco [101] Paraffin (433) IHC HER-2/neu protein overexpression predicted resistance and adverse outcome in patients treated with tamoxifen Ravdin [102] Paraffin (595) IHC CAFT chemotherapy superior to tamoxifen in HER-2/neu positive cases (not statistically certain). IHC = immunohistochemistry; ELISA = enzyme-linked immunosorbent assay; EIA = enzyme immunoassay; CMF = cytoxan, methotrexate, 5-fluorouracil; ER = estrogen receptor.

11 Ross, Fletcher 246 study including 200 patients, HER-2/neu protein staining failed to predict tamoxifen resistance in estrogen-receptor positive cases [100]. This was also the case in a previously published study using an ELISA technique to determine the HER-2/neu protein status [94]. In a twenty-year update of an adjuvant chemotherapy trial from Italy, not only did HER- 2/neu protein overexpression predict resistance to tamoxifen treatment, HER-2/neu-positive patients treated with tamoxifen actually had an adverse outcome compared with untreated patients [101]. In an initial report of the SWOG Biological Correlative Study, patients with HER-2/neu protein overexpressing tumors appeared more responsive to cytotoxic chemotherapy with a CAFT regimen than when treated with tamoxifen alone, although this relationship could not achieve complete statistical significance [102]. HER-2/neu protein overexpression has also been linked to CMF resistance in clinical specimens (cytoxan, methotrexate, 5-fluorouracil) [32, 87, 88] and taxol resistance in cell lines [91]. However, another study found that HER-2/neu positive tumors were three times more sensitive to taxol [95], and HER-2/neu overexpression has also been associated with enhanced response rates to chemotherapy regimens containing adriamycin [16]. HER-2/neu immunostaining has successfully predicted local recurrence in patients receiving surgery and radiation alone [93]. Recently, an artificial neural network approach utilized HER-2/neu status along with other Table 8. Serum-based studies of HER-2/neu protein in breast cancer prognostic markers to successfully predict response to adjuvant chemotherapy and radiation [98]. It appears that although strong trends have been presented in the published studies, such as the resistance to tamoxifen and sensitivity to adriamycin for HER-2/neu-positive tumors, more studies are needed using appropriate control arms to confirm these important associations. Should this be accomplished, it would seem likely that HER-2/neu testing would be of even greater value in the management of breast cancer patients and possibly achieve status as a standard of practice similar to hormone receptor testing. SERUM HER-2/neu ANTIGEN LEVELS AS A TUMOR MARKER Antibodies similar to those used for HER-2/neu protein detection in breast tumor tissues have been applied to human serum samples to test whether circulating HER-2/neu receptor protein levels can predict the progression of breast cancer [88, 94, 99, ]. Table 8 is a summary of 12 published studies on 1,960 patients featuring HER-2/neu serum protein enzyme-based immunoassays. Seven of the studies (1,330 patients) found a significant correlation of serum HER-2 protein levels with either disease recurrence, metastasis, or shortened survival [103, 106, 98, ]. Two studies (379 patients) reported no significant association of serum levels with prognosis [104, 107]. Of the studies in which serum 1st Author Sample Year [reference] number Method Findings 1994 Isola [103] Serum (227) ELISA Serum HER-2 protein levels predicted development of metastasis and treatment resistance Kandl [104] Serum (79) ELISA Serum HER-2 protein did not correlate with tissue IHC status but did predict shortened survival Wu [105] Serum (124) ELISA Serum HER-2/neu protein levels correlated with levels measured from tumor cytosols; breast cancer had higher levels than prostate or ovarian cancers Andersen Serum (168) ELISA Serum levels correlated with tumor gene amplification and detected metastatic disease [106] but not local recurrence Leitzel [87] Serum (300) EIA High serum levels of HER-2/neu protein predicted lack of response to hormone therapy Volas [107] Serum (300) EIA Serial HER-2/neu protein levels did not predict disease course or response to hormonal therapy Molina [108] Serum (200) EIA Serum HER-2/neu protein levels were an early predictor of metastasis Willsher [109] Serum (119) ELISA Serum HER-2/neu protein levels had prognostic significance at all stages; did not predict the response to hormonal therapy Revillion [94] Serum (33) ELISA Serum HER-2/neu protein levels at diagnosis did not predict response rates to chemotherapy Yamauchi [99] Serum (94) ELISA High HER-2 protein serum levels predict lack of response to hormonal therapy Fehm [110] Serum (211) EIA Elevated in 12% of locally advanced breast cancer patients; independent predictor of disease-free survival Mansour [111] Serum (105) EIA Significant predictor of short-term disease outcome (p = 0.007). EIA = enzyme immunoassay; ELISA = enzyme-linked immunosorbent assay.

12 247 The HER-2/neu Oncogene in Breast Cancer HER-2 protein levels predicted response to therapy, three studies found that elevated serum HER-2/neu protein predicted therapy resistance [98, 103, 106], whereas three studies did not demonstrate this association [94, 107, 109]. Interestingly, elevated serum HER-2/neu levels correlate with decreased survival and absence of clinical response to hormonal therapy, even when estrogen receptor assays are positive [98]. However, in a recent study, serial serum HER-2/neu protein levels did not correlate with the clinical course in a group of patients with metastatic breast carcinoma treated with second-line hormonal therapy [107]. NEW THERAPIES IN BREAST CANCER DIRECTED AT THE HER-2/neu GENE AND PROTEIN Table 9 lists a series of studies in which the HER-2/neu receptor protein served as a therapeutic target. Although most of the listed studies were performed in vitro using cell lines and xenografts [112, 113, 115, ] the preliminary results of clinical trials in humans have been promising [114, 116, 121, 122]. Baselga and coworkers have reported a phase II study of 45 women with HER-2/neu overexpressing metastatic breast cancer resistant to multiple previous therapies, in which 11.6% and 37% of patients developed objective response and disease stabilization, respectively, when treated with monoclonal antibody rhumab HER-2 (Herceptin ) Table 9. Selected studies of breast cancer treatment targeting the HER-2/neu gene and protein directed toward the HER-2/neu protein [116]. Herceptin has a favorable toxicity profile and has recently achieved a significant overall response rate of 15% when given alone in metastatic breast cancer [121]. Most recently, the results of a randomized multinational controlled phase III clinical trial using Herceptin in combination with first-line chemotherapy in metastatic breast cancer have been reported [122]. In 469 patients with HER-2/neu overexpressing primary tumors, Herceptin increased the time to disease progression and response rate when given in combination with either adriamycin-cytoxan or paclitaxel [122]. Although the adverse event incidence in this study was not increased by Herceptin therapy, the incidence of myocardial toxicity classically associated with adriamycin was slightly increased in patients who received the drug [122]. If the preliminary results can be confirmed on large patient groups, then the use of anti-her- 2/neu antibody therapy may have significant potential as a new method of breast cancer treatment whether given alone or in combination with other conventional treatment modalities. At the time of this review, similar clinical trials are in progress at multiple centers. Additional approaches to anti-her-2/neu therapy, under preliminary investigation, include the use of antisense gene therapy probes [117] and anti-her-2/neu protein vaccination [123, 124]. 1st Author Cases Treatment Year [reference] treated method Results 1991 Shepard [112] Nude mice mumab 4D5 Enhances TGF-alpha cytotoxicity and blocks progression of xenografts in nude mice Zhang [113] Cell lines Emodin Tyrosine kinase inhibitor dephosphorylates HER-2 protein and prevents anchoragedependent and independent cell growth Valone [114] 10 patients MDX-210 Phase I trials; 2 (20%) of 10 patients responded; low toxicity. antibody 1996 Jinno [115] Cell lines A0011 antibody Enhanced adriamycin cytotoxicity by antibody targeting. + adriamycin 1996 Baselga [116] 46 patients rhumab HER-2 Low toxicity; 12% overall response rate in patients with distant metastases. antibody 1996 Liu [117] Cell lines Antisense Inhibition of proliferation, HER-2/neu protein production, and anchorageoligonucleotides independent growth Wright [118] Cell lines sfv antibody Selective killing of breast cancer cells that overexpress HER-2/neu protein. delivered by adenovirus 1997 Heijnen Cell lines 520C9 antibody Synergistic effect of both agents on PMN-directed killing of breast [119] + G-CSF cancer cells in culture Witters [120] Cell lines 4D5 antibody Synergistic antiproliferative effect of both agents on breast cancer cells in culture. + tamoxifen 1998 Cobleigh 222 patients Herceptin 15% overall response rate for Herceptin used alone in metastatic disease. [121] 1998 Slamon [122] 469 patients Herceptin Herceptin plus chemotherapy (AC or paclitaxel) increased time to progression and response rates in patients with metastatic disease.

13 Ross, Fletcher 248 OTHER USES OF HER-2/neu IN BREAST CANCER. Antibodies to the HER-2/neu protein have been used for radionucleotide imaging in breast cancer. In a 1994 study, Allan and coworkers used Tc99m-labeled monoclonal antibodies to HER-2 protein and reported encouraging results [125]. In addition to a potential use as a detector of occult metastasis, this approach may also allow for antibody-targeted chemo or radiation therapy in patients with recurrent or metastatic disease. CONCLUSIONS The preponderance of evidence indicates that HER- 2/neu gene amplification and protein overexpression are associated with an adverse outcome in breast cancer. Although controversy exists over the best detection method for measuring abnormalities of HER-2/neu, the majority of data favor the conclusion that assays of gene amplification such as the FISH technique and immunohistochemistry on REFERENCES 1 Slamon DJ, Clark GM. Amplification of C-ERB-B2 and aggressive breast tumors? Science 1988;240: De Potter CR. The neu oncogene: more than a prognostic indicator? Hum Pathol 1994;25: Lupu R, Cardillo M, Harris L et al. Interaction between ERBreceptors and heregulin in breast cancer tumor progression and drug resistance. Sem Cancer Biol 1995;6: Slamon DJ, Clark GM, Wong SG et al. Human breast cancer: correlation of relapse and survival with amplification of the Her-2/neu oncogene. Science 1987;235: Berger MS, Locher GW, Saurer S et al. Correlation of c-erb B2 gene amplification and protein expression in human breast carcinoma with nodal status and nuclear grading. Cancer Res 1988;48: van de Vivjer MJ, Peterse JL, Mooi WJ et al. Neu-protein overexpression in breast cancer. N Engl J Med 1988;319: Heintz NH, Leslie KO, Rogers LA et al. Amplification of the c-erb B-2 oncogene in prognosis of breast adenocarcinoma. Arch Pathol Lab Med 1990;114: Tsuda H, Hirohashi S, Shimosato Y et al. Correlation between histologic grade of malignancy and copy number of c-erbb-2 gene in breast carcinoma. A retrospective analysis of 176 cases. Cancer 1990; 65: Borg A, Tandon AK, Sigurdsson H et al. HER-2/neu amplification predicts poor survival in node-positive breast cancer. Cancer Res 1990;50: frozen sections or enzyme immunoassays on fresh tumor cytosolic protein produce more consistent results than immunohistochemistry on formalin-fixed, paraffin-embedded tissue sections. Although requiring further statistical confirmation, evidence also indicates that breast cancers with HER-2/neu gene or protein abnormalities are resistant to treatment with tamoxifen alone and relatively more sensitive to cytotoxic therapy regimens containing adriamycin. Based on the promising new information concerning the response of metastatic breast cancer in patients with HER- 2/neu overexpressing primary tumors to treatment with anti- HER-2 antibodies (Herceptin ), it appears that further basic and clinical research studies of this gene and protein will likely continue well into the 21st Century. ACKNOWLEDGMENT Dr. Ross and Dr. Fletcher are consultants for Oncor, Inc., Gaithersburg, Maryland. 10 Paik S, Hazan R, Fisher ER et al. Pathologic findings from the nations surgical adjuvant breast and bowel project: prognostic significance of erb B2 protein overexpression in primary breast cancer. J Clin Oncol 1990;8: Battifora H, Gaffey M, Esteban J et al. Immunohistochemical assay of neu/c-erb B-2 oncogene product in paraffin-embedded tissues in early breast cancer: Retrospective follow-up study of 245 stage I and II cases. Modern Pathol 1991;4: Kallioniemi OP, Holli K, Visakorpi T et al. Association of C- erb B2 protein over-expression with high rate of cell proliferation, increased risk of visceral metastasis and poor long-term survival in breast cancer. Int J Cancer 1991;49: Clark GM, McGuire WL. Follow-up study of HER-2/neu amplification in primary breast cancer. Cancer Res 1991;51: Lovekin C, Ellis IO, Locker A et al. C-erb B2 oncoprotein expression in primary and advanced breast cancer. Br J Cancer 1991;63: McCann AH, DeDervan TA, O Regan M et al. Prognostic significance of C-erb B2 and estrogen receptor status in human breast cancer. Cancer Res 1991;51: Dykens R, Corbett IP, Henry J et al. Long term survival in breast cancer related to overexpression of the C-erb B2 oncoprotein: an immunohistochemical study using monoclonal antibody NCL-CB11. J Pathol 1991;163: Rilke F, Colnaghi MI, Cascinelli N et al. Prognostic significance of HER-2/neu expression in breast cancer and its relationship to other prognostic factors. Int J Cancer 1991;49:44-49.