Molecular markers are studied for their potential to act as prognostic or predictive factors. A prognostic factor influences the clinical outcome independently of treatment, whereas a predictive factor correlates with prognosis because it is linked to the response to a particular therapy. HER2 may function as both a prognostic and a predictive factor; in addition, HER2 can be the target of therapy.
The HER2/neu oncogene, also referred to as c-erbB-2/neu, encodes a protein with a molecular weight of 185,000 daltons (p185). The gene product is a transmembrane tyrosine kinase receptor belonging to a family of epidermal growth factor receptors (EGFRs) that are structurally related to the human EGFR.[1,2] The other members of this family are HER1 (also known as EGFR), HER3, and HER4. This receptor family is known as the type 1-receptor tyrosine kinases.
Ligands for this family of receptors include the epidermal growth factor (EGF) and neuregulins, also known as neu differentiation factors or heregulins. At least six different ligands, the EGF-like ligands, activate the EGF receptor and cause formation of homodimersan event believed to activate the intrinsic tyrosine kinase, resulting in transautophosphorylation of tyrosine residues. EGF-like ligands can also induce heterodimerization between other members of the HER family, forming heterodimers such as HER1/HER2, HER1/HER3, and HER1/HER4 (Figures 1 and 2). The HER2 receptor is partially homologous to the EGFR. However, to date, unlike for EGFR, HER3, and HER4, no ligand for HER2 has been identified.
The second class of ligands for the HER receptors, collectively termed neuregulins, bind directly to HER3 and HER4, but not to HER2 or the EGFR. It is hypothesized that the main role of HER2 may be to dimerize with the other members of the HER family of receptors. Interestingly, HER2 is the preferred heterodimerization partner within the family and is frequently transactivated by EGF-like ligands or neuregulins resulting from the formation of heterodimerization with other members of the HER family. This heterodimerization allows the participation of HER2 in signal transduction, even in the absence of a cognate ligand.
The HER2 gene was first identified as a transforming oncogene in the DNA of chemically induced neuroblastomas in the rat. HER2 is overexpressed or amplified in approximately 20% to 25% of human breast cancers. The HER2 gene product is composed of a cytoplasmic domain with tyrosine kinase activity, a transmembrane domain, and an extracellular domain that may be shed from the surface of breast cancer cells.
HER2 may be involved in the pathogenesis and clinical aggressiveness of HER2-overexpressing tumors. Indeed, evidence supports a direct role for HER2 overexpression in the pathogenesis and poor clinical outcome of human tumors. When the mutated gene is transfected into mouse fibroblast cells (NIH-3Y3), it causes transformation, and the resulting cells are tumorigenic in nude mice. Transgenic mice that overexpress the neu gene (the rodent homolog of the human HER2 gene) develop breast cancer. Specific antibodies to the extracellular domain of the human HER2/neu gene product inhibit the growth of experimental tumors that overexpress the gene.
HER2 has been shown to be overexpressed in several human carcinomas including, but not limited to, breast, ovarian, gastric, colon, and non-small-cell lung cancer. Cellular proliferation is regulated by extracellular factors that trigger signal transduction cascades from surface receptors through cytoplasmic effectors and ultimately control progression through the cell cycle. To date, the exact mechanisms by which oncogenic HER2 affects cell proliferation and the cell-cycle regulatory components involved have not been identified.
HER2 Overexpression as a Predictive Factor
Amplification of the HER2 proto-oncogene and overexpression of its protein product in patients with breast cancer have been linked to a poor prognosis; ie, a more aggressive clinical course and shortened survival.[12,13] Although inconclusive, data have also suggested that HER2 overexpression may be useful as a predictive factor, raising the possibility of pretreatment selection of patients who might benefit from particular therapeutic strategies. Recent data indicate that patients with lymph node-positive breast cancer whose tumors overexpress HER2 may obtain additional benefit from adjuvant anthracycline-containing chemotherapy (as opposed to non-anthracycline-containing regimens).[14,15]
The role of HER2 status in predicting response to adjuvant systemic therapy and the method of defining HER2 status require further exploration before definitive recommendations can be made on how best to utilize this biomarker. Further investigations are also necessary to define the responsiveness of HER2-positive tumors to hormonal therapy with selective estrogen-receptor modulators (SERMs) and aromatase inhibitors.
Preclinical studies have suggested that estrogen-dependent cultured human breast cancer cell lines are rendered hormone-independent after transfection with multiple copies of the HER2/neu gene. Some studies have suggested that patients whose tumors overexpress HER2/neu are less likely to respond to tamoxifen(Drug information on tamoxifen) and may have a worse outcome, compared to patients with normal HER2/neu expression.[18,19] However, other studies have reported no worse outcome with the use of adjuvant tamoxifen in patients with HER2-overexpressing tumors.[20,21]
The American Society of Clinical Oncology (ASCO) recently updated its recommendations for the use of tumor markers. In this document, ASCO recommends the evaluation of HER2/neu status in all primary breast cancers, at either the time of diagnosis or the time of recurrence. However, the data are currently insufficient to recommend the routine use of HER2 overexpression to identify patients with a higher risk of relapse. Complicating the evaluation of the published data on HER2 as a prognostic factor is the lack of uniform use of assays in assessing HER2 by immunohistochemistry or fluorescent in situ hybridization (FISH).
The need for accurate detection of HER2 status is becoming more apparent, as therapeutic decisions are influenced by this information in both the adjuvant and advanced-stage setting.
A wide range of assay methods have been used to assess the HER2 status of fresh and archival surgical specimens from breast carcinoma patients. Methods to assess protein overexpression include immunohistochemistry, enzyme immunoassay, and Western blot analysis; methods to evaluate gene amplification include FISH, Southern blot analysis, and polymerase chain reaction (PCR); and methods to assess messenger (m)RNA overexpression include Northern blot analysis and in situ hybridization (Table 1).[22,23] In addition, circulating levels of the shed extracellular domain of the HER2 receptor protein can be detected by serum enzyme-linked immunosorbent assay (ELISA).
All of these assays have been used in research laboratories, but some have not been routinely used in hospitals because they require specialized equipment and/or the use of radioisotopes. On the other hand, a wide range of factors affect both HER2 testing and determination of HER2 status, including standardization of methods, standardization of antibodies (monoclonal/polyclonal) used for immunohistochemistry, agreement on the adoption of common grading criteria, scoring systems, and correlation of various assays.
Immunohistochemistry and FISH are the techniques routinely recommended for determining HER2 status, and both have been approved by the US Food and Drug Administration (FDA) for use in the selection of patients for therapy with the monoclonal antibody trastuzumab(Drug information on trastuzumab) (Herceptin). The other available techniques should be used for research purposes only. However, many factors may compromise successful immunohistochemical testing of HER2 status.
Form of FixationThe fixative should preserve antigenic integrity and limit extraction, diffusion, or displacement of antigen during subsequent processing. Formalin-fixed, paraffin(Drug information on paraffin) wax-embedded tumor tissue samples are appropriate for immunohistochemical assessment of HER2, whereas other methods of tissue fixation can adversely affect reactivity. The rate of false-positive cases rises to 50% with the use of an unsuitable fixative. Evidence suggests that HER2 protein reactivity may deteriorate in fixed paraffin wax sections after prolonged storage; the immunohistochemistry test should, therefore, be performed on block sections that have been stored for no more than a few months.
Antigen RetrievalExcessive antigen retrieval can artifactually increase immunoreactivity of cancer cells and bring about spurious membrane reactivity in normal breast epithelial cells, as well as carcinoma cells.
AntibodySeveral antibodies are commercially available, including CB11 (mouse antihuman monoclonal antibody, Ventana, Tucson), TAB 250 (mouse antihuman monoclonal antibody, Zymed, San Francisco), and HercepTest (rabbit antihuman HER2/neu polyclonal antibody, DAKO, Carpinteria, Calif). No single antibody has been consistently demonstrated to be superior in terms of sensitivity and specificity (Figure 3). The polyclonal antiserum used in the FDA-approved HercepTest, however, has been subjected to comprehensive standardization of immunohistochemical procedures.
Scoring SystemHercepTest is the currently recommended scoring method. A semiquantitative system based on the intensity of cell-membrane immunostaining and the percentage of positive cells, it has a score range from 0 to 3+. Samples scoring 3+ are regarded as unequivocally positive and 0/1+ as negative. Borderline 1+/2+ and 2+ require confirmation with an alternative method, preferably FISH.
When dealing with the analysis of an image, both brightness and spatial resolution play crucial roles. Basically, the finer the resolution, the closer we approach the original appearance of the image. The quality of resolution, in turn, is significantly impaired by a so-called contouring phenomenon, which mainly affects the perimeter evaluation of images, such as HER2 immunostain on the cell membrane. Taken together, these optical phenomena are likely to influence interpretation of immunohistochemistry. Interobserver variation in the assessment of staining is considerable and can lead to misclassification of HER2 status. A simplified immunohistochemistry scoring method is advisable.
The accuracy, sensitivity, and reproducibility of HER2 immunohistochemical assay can be substantially improved with the use of an image analyzer system to quantify the immunohistochemical staining. The Automated Cellular Imaging System (ACIS, ChromaVision, San Juan Capistrano, Calif) automatically scans the immunohistochemically stained slide, and clearly distinguishes and quantifies cell membrane staining from cytoplasmic staining using so-called color-space transformation proprietary technology.
With the ACIS system, HercepTest is scored as a number between 0 and 4, with a score < 2 indicating a negative result and > 2, a positive result. FISH analysis is requested for scores ranging from 0.5 to 1.9. When compared with the FISH standard, the ACIS immunohistochemical assay notably reduces intraobserver and interobserver disagreement, thus improving the concordance rate and sensitivity of the manual immunohistochemical assay.
False-Positive ResultsAnother pitfall of immunohistochemistry is the possibility of false-positive results. In over 90% of cases, HER2 protein overexpression is related to HER2/neu gene amplification, which results in increased mRNA transcription and, ultimately, increased synthesis of the glycoprotein receptor. When technical artifacts (eg, excessive antigen retrieval) can be ruled out, HER2 overexpression in the absence of genomic amplification could be the effect of a transcriptionally up-regulated overexpression.[28,30] HER2 is a growth factor, and enhanced transcription in the absence of gene amplification is a well-recognized mechanism of cellular function through enhanced production of mRNA by phosphorylation of tyrosine kinase acting on growth factors and other regulators of cell growth and proliferation.
Transiently enhanced transcription unlinked with the corresponding oncogene amplification has been immunohistochemically seen in colonic regenerating epithelium and is a well-recognized mechanism in cellular homeostasis. This does not seem to be the case for HER2 in breast cancer, however, and most HercepTest-positive cases with a score of 2+ and a normal gene copy should be regarded as true false-positives, unresponsive to trastuzumab. In contrast, gene amplification without protein overexpression is an artifact and can only be due to a failure of the immunohistochemistry assay; ie, posttranslational modifications of the mRNA transcript may go undetected by inappropriate or inadequate antibody detection methods.