Quantification of HER family receptors in breast cancer

IHC is the primary technique used to determine protein expression status in a patient sample. It is a simple, fast, easy to implement and relatively inexpensive method for protein detection. Slides are incubated with an antibody directed against the HER receptor protein, labeled, and finally made visible with a chromogen, resulting in a staining localized in the cellular compartment where the protein target is expressed (membrane, cytosol, nucleus). The more the protein is present, the stronger the staining will be. Traditionally, assessment of protein expression is done by visual estimation of staining intensity and is reported as binary (positive versus negative), four-tiered (0, 1+, 2+, and 3+), or semiquantitative continuous variable as for the H score ((% at 0) × 0 + (% at 1+) × 1 + (% at 2+) × 2 + (% at 3+) × 3; range = 0 to 300) results [28].

For companion diagnostic tests, guidelines are generally issued to guide pathologists in the interpretation and scoring of the staining. The HER2 scoring guidelines recommended by ASCO/CAP classified HER2 expression as 0 (no staining or faint incomplete membrane staining observed in ≤10% of tumor cells), 1+ (faint/barely perceptible incomplete membrane staining in >10% of tumor cells), 2+ (circumferential membrane staining that is incomplete and/or weak/moderate and within >10% of tumor cells or complete and circumferential membrane staining that is intense and within ≤10% of tumor cells) or 3+ (circumferential membrane staining that is complete, intense, and within >10% of tumor cells). Tumors with scores 0 and 1+ were considered negative; 2+ was considered equivocal and required FISH reflex testing; 3+ was considered positive and eligible for trastuzumab [29].

Despite the effort to standardize HER2 status determination, current guidelines do not restrict the type and characteristics of IHC assay to be used for HER2 protein expression. The use of FDA approved tests such as HercepTest® (DAKO, Carpinteria, CA, USA), PATHWAY anti-HER-2/neu (Ventana Medical systems, Roche, Tuscon, AZ, USA), InSite™ Her-2/neu (Biogenex, Freemont, CA, USA) as well as fully automated staining systems such as Ventana Benchmark (Ventana Medical systems, Roche, Tuscon, AZ, USA) and Leica Microsystems Bondmax (Leica, Newcastle, UK) may certainly minimize process variability and improve assay repeatability and reproducibility. Nevertheless, many laboratories developed tests with different antibodies directed against other HER2 epitopes (intracellular versus extracellular), which may show non-overlapping specificities and be differently influenced by preanalytical factors [30-32].

Interpretation of membrane staining can be optimized using quantitative image analysis such as the automated quantitative analysis (AQUA) system [33]. AQUA is a fluorescence IHC-based method that provides objective and continuous protein expression scores for tissues by using automated fluorescence microscopy and advanced image analysis algorithms. It is important to note, however, that there are as yet no clinical data related to AQUA’s predictive ability.

Other software applications include Aperio Scanscope (Aperio Technologies; Vista, CA, USA), Definiens (Carlsbad, CA, USA) and Vysis AutoVysion (Abbott Molecular, Des Plaines, IL, USA) among others. These applications can reduce the subjectivity of a traditional scoring system and provide a more reproducible protein expression score [34,35].

HER2 receptor protein extracellular domain (ECD, p105) is released into the circulation after cleavage by matrix metalloproteinases and its levels can be measured in the serum using an enzyme-linked immunosorbent assay approved by the FDA (Siemens Healthcare, Erlangen, Germany). Elevated levels of serum HER2 ECD have been shown to be both prognostic and predictive of response to trastuzumab in HER2-positive tumors [36,37]. Serum ECD values have been suggested, therefore, as an alternative technique for determining HER2 status, although available results are controversial. First, not all patients with HER2-positive tumors appear to have elevated serum ECD values and patients with HER2-negative tumors can also have elevated ECD values. Second, the reported data come from studies including a limited number of patients, thus making current evidences still insufficient to consider basing treatment decisions on ECD levels in routine clinical practice. A large meta-analysis study [38] has combined the data of four trastuzumab trials in metastatic breast cancer and showed that, from the combined dataset (N = 322 patients), there was no correlation between baseline ECD value and tumor response. ECD values decreased upon initiation of combination therapy irrespective of treatment and tumor response. Furthermore, disease progression was not reliably predicted by an increase in ECD levels. Therefore, the use of ECD values in treatment decision making was not recommended.

The VeraTag™ proximity-based assay (HERmark® Breast Cancer Assay; Monogram Biosciences, Inc., South San Francisco, CA, USA) enables precise quantitative measurements of total HER-2 expression and HER2 homodimers in formalin-fixed, paraffin-embedded (FFPE) tissue specimens [39,40]. The HERmark assay was developed based on a proprietary proximity-based technology platform that enables accurate quantification of proteins and protein-protein complexes through the release of a fluorescent tag (VeraTag reporter, Monogram Biosciences) conjugated to a pair of monoclonal antibodies directed to unique epitopes on the HER2 receptor in molecular proximity [40]. The continuous total HER-2 expression results are grouped as HERmark negative, HERmark equivocal, and HERmark positive. The threshold for a positive HERmark test is based on the comparison with HER2 tests performed in 1,090 breast tumor reference samples (central IHC and central in situ hybridization) from three different study cohorts. The HERmark assay can detect HER2 at amounts of 2,500 up to more than 1 million receptors per cell, and is thus said to be 7 to 10 times more sensitive than IHC. The assay has been validated according to the specifications prescribed by the Clinical Laboratory Improvement Amendments and is performed only in the CAP-certified clinical reference laboratory at Monogram Biosciences (US). VeraTag™ proximity-based assays have been developed also to measure total EGFR, EGFR-EGFR homodimers and EGFR-HER2 heterodimers [40-42], p95 [43], total HER3, HER3-HER3 homodimers and HER3-phosphoinositide 3-kinase (PI3K) complex heterodimers [44] and the phosphorylated forms of EGFR, HER2, and HER3.

Clinical application of protein-protein interactions has uncovered many potential targets for novel drug development or drug resistance mechanisms [45], with the MDM2-p53 interaction [46,47] and B-Raf inhibition being examples of recent successes [48]. More recently, incorporation of protein interaction data was shown to also improve the predictive performance of prognostic gene expression signatures [49,50]. Despite the importance of adjunct information supplied by the protein interactome configuration to improve the existing prognostic signatures for predicting patient outcome [50], this protein interaction information has rarely been incorporated in diagnostic/prognostic assays.

Fluorescence lifetime imaging microscopy (FLIM) is based on quantifying the non-radiative transfer of energy between the donor and acceptor fluorophores and can only occur when the two molecules are no further apart than 10 nm, consistent with being in molecular contact [51-53]. Various automated imaging platforms, including ours, measure Förster resonance energy transfer (FRET) - the decrease in donor lifetime, the gold standard for FRET measurements (reviewed recently in [54]) - to directly monitor validated protein-protein interactions [55-61] and post-translational modifications, including conformational changes, in cultured cells [58,62-66]. A two antibody FRET/FLIM approach was originally applied, by ourselves and others, to human cancer tissues to detect the nano-proximity between a donor fluorophore-conjugated anti-protein kinase C or anti-EGFR antibody, and an acceptor fluorophore-labeled phospho-specific antibody, providing highly specific quantification of phosphorylation [67,68]. Detailed methodology for sample preparation and instrumentation can be found elsewhere [69,70]. We have now extended this method to measure endogenous protein-protein interactions in archived pathological material [71]. The presence of autofluorescence in stromal and epithelial components may cause difficulties in accurately determining the fluorescence lifetime of fluorophores in FFPE tissue samples [72]. By circumventing the autofluorescence issue using a new analysis algorithm [73], we have recently described the first clinical utilization of this refined FLIM assay (using Alexa546 and Cy5 as donor and acceptor fluorophores, respectively) to quantify the level of HER1-HER3 dimer formation in FFPE tissues from basal-like breast cancer patients who were treated with a neoadjuvant anti-EGFR treatment (cetuximab or panitumumab) [74]. Moreover, we have demonstrated the existence of EGFR-HER4 dimers in breast cancer cells and how these dimers are important for cell motility [75].

Liquid chromatography-tandem mass spectrometry-based proteomics has emerged as the most effective method to study complex proteomes. In this approach, the proteins representing a proteome are analyzed after enzymatic digestion by liquid chromatography coupled to mass spectrometry (MS). Although this approach is a powerful tool to identify proteins in complex biological samples [76,77], it is not optimal for systematic quantification of these proteins because of the stochastic nature and the limited sensitivity of the approach. During the past few years, targeted proteomics has been shown to be complementary to the more widely used discovery proteomic methods. In targeted proteomics, only pre-determined peptide ions are selected for detection and quantification in a sample. The main MS approach supporting targeted proteomics is selected reaction monitoring (SRM), where specific MS assays are generated a priori and used to selectively detect and quantify proteins of interest in a sample. This approach can provide objective quantification and multiplex capabilities with high sensitivity and in an antibody-free setting [78-80]. SRM methods have long been used to quantify low-abundance protein targets in plasma [81] but application of these techniques to FFPE tissue samples has, until recently, been hindered by incomplete solubilization of samples [82,83]. The Liquid Tissue-(SRM) diagnostic technology platform is a newly developed proteomic method that overcomes this limitation, allowing for precise protein quantification in FFPE tissues. Microdissected FFPE tumor tissues are subjected to Liquid Tissue processing to reverse formalin crosslinks. This is followed by trypsinization to completely solubilize all of the protein in the sample. This tryptic peptide mixture is then subjected to SRM analysis using stable isotope-labeled control peptides for accurate quantification [83-85]. Multiple reports have demonstrated that comparable results may be obtained between formalin fixed and matching frozen tissue [84,86]. The reliability of this approach for analysis of proteins in any biological sample including FFPE patient tumor tissue has been demonstrated [87-91], thus widening the application of MS to patient-derived tissue with a consequent profound impact on patient stratification and targeted cancer therapeutics.

Reverse phase protein array (RPPA) and collaborative enzyme enhanced reactive-immunoassay (CEER) are nano-scaled dot blot platforms allowing the detection of multiple proteins (both total and phosphorylated) in many samples simultaneously. They do not require large amounts of sample but are not suitable for FFPE tissue. For RPPA protein lysates are immobilized onto microarrays and then probed with the primary antibodies of choice. Detection is performed by quantification of the labels (fluorescent, colorimetric or other kinds) bound to either the primary or, more often, the secondary antibody added to amplify the signal. RPPA allows testing hundreds of samples at the same time and multiplexing is performed by analyzing multiple arrays spotted with the same protein lysates with different antibodies [92].

CEER takes advantage of the immunocomplexes formed between antibodies printed on a nitrocellulose microarray surface with the target molecules in cell lysates. Once the complexes are formed, two detector antibodies (one conjugated to glucose oxidase and another conjugated to horse radish peroxidase (HRP)) are added. Target detection (expressed as computational units (CU)) requires the presence of both detector antibodies, and the enzyme channeling event between glucose oxidase and HRP will not occur unless both antibodies are in close proximity [93]. The main difference with RPPA is that, instead of protein lysate, antibodies are immobilized in cellulose arrays. This means that, contrary to RPPA, CEER is capable of measuring the expression of dozens of targets simultaneously in the same sample.

Further studies are needed to prove the clinical relevance of the above described methods.

FISH is considered the gold standard method for gene amplification status determination. FISH uses fluorescently labeled probes (usually red) that are complementary to a part of the target gene. After hybridization to the complementary DNA on the slide, the probes can be visualized with a fluorescence microscope. A second probe labeled with a different fluorochrome (usually green) directed against the centromeric region of the chromosome containing the target gene is generally used as control for polysomy. The number of copies of the target gene and centromere probe can be estimated and the ratio determined.

Similar to CISH, silver in situ hybridization (SISH) technology uses a non-fluorescent method where the HRP bound to the probe catalyses the reduction of silver acetate to produce a black signal. Several studies showed a good correlation between FISH, CISH, SISH, and IHC for HER2 status determination [94-100].

Recently released ASCO/CAP guidelines recommended that HER2 must be considered in situ hybridization (ISH)-positive based on a single-probe average HER2 copy number ≥6.0 signals/cell or dual-probe HER2/CEP17 ratio ≥2.0 or dual-probe HER2/CEP17 ratio <2.0 with an average HER2 copy number ≥6.0 signals/cell [27]. Whether the centromere control probes for polysomy 17 are really necessary is a matter of debate given that it has been proven by several studies that true polysomy 17 is very rare in breast carcinomas [101]. Concurrent evaluation of several chromosome 17 genes using multiple-probe FISH or multiplex ligation-dependent probe amplification showed that focal amplifications encompassing the centromere - and not true polysomy - are the most common explanation for increases in CEP17 signals [102,103]. These results suggest that CEP17 copy number assessment by standard ISH is not a useful surrogate for polysomy 17. Compared with IHC, ISH assays, in which the target gene copy number is counted, are considered to be more quantitative analytically. However, ISH is not a direct measurement of the protein and just because a change in gene copy number is observed does not necessarily mean that it is expressed. In addition, the procedure is time consuming and new 'fast' FISH assays are under development to reduce the turnaround time [104].

PCR-based techniques such as multiplex ligation-dependent probe amplification [105] have several advantages over ISH-based assays. First, they are more quantitative and results are easier to interpret. Second, they require only small amounts of DNA and are not affected by DNA degradation, thus performing well with FFPE samples. Third, they can be multiplexed, allowing simultaneous interrogation of multiple genes or different parts of genes, representing an ideal and low cost prescreening tool. Head-to-head comparisons between IHC, FISH, and CISH have shown good correlation among technologies [106-109]. The main weaknesses of PCR-based assays are that they do not preserve tissue morphology, may require sample macro- or microdissection to enrich for tumor content, heterogeneity can be missed and contamination with normal or ductal carcinoma in situ may lead to both false-negative and false-positive results.