The current study was performed to determine the impact of polysomy 17 on the interpretation of HER2 testing of invasive breast carcinomas using fluorescent in situ hybridization methods. Current American Society of Clinical Oncology/College of American Pathologists guidelines define HER2-positive tumors as those with >6 HER2 genes per nucleus or those with HER2/CEP17 (chromosome 17) ratio >2.2. These guidelines are potentially contradictory in tumors with polysomy of chromosome 17.
Seventy-two breast carcinoma cases with reported polysomy of chromosome 17 (≥3 CEP17 signals on average) by fluorescent in situ hybridization were identified, and the corresponding HER2 immunohistochemistry was obtained. The HER2 status of the archived samples was reviewed, and the tumors were recategorized according to the 2007 American Society of Clinical Oncology/College of American Pathologists guidelines.
The average CEP17 copy number for the group was 4.5 (range, 3.0-10.4). Thirty-three (45.8%) cases had >6 copies of HER2 per nucleus. Twenty-one cases (29.2%) qualified as HER2 gene amplified using the HER2/CEP17 ratio (>2.2) guideline. All these cases had >6 HER2 signals, which represented 63.6% of all cases with >6 HER2 signals. HER2 protein expression showed significant positive correlations with both HER2 gene copy number and HER2/CEP17 ratio (P < .01, rs = 0.56 and 0.64, respectively).
HER2 status in breast carcinomas has become a standard prognostic marker and is an essential test for the selection of breast cancer patients eligible for the targeted therapy (eg, trastuzumab and lapatinib).1, 2 Trastuzumab, a humanized monoclonal antibody that targets the HER2 protein, is routinely used in the treatment of patients with breast carcinoma overexpressing this protein. Approximately 20% of breast cancers show HER2 overexpression and are characterized by decreased relapse-free time and overall survival.3 Therefore, trastuzumab therapy is widely used as a first-line cancer treatment in breast cancer patients whose primary or metastatic tumors overexpress the HER2 protein.4 Trastuzumab improves response rates and survival, and decreases time to progression when used alone or when added to chemotherapy in metastatic breast cancer.5 Given the significant clinical benefits of trastuzumab in patients with HER2-positive breast cancer, it is of paramount importance to accurately identify all patients eligible for this therapy.
The primary mechanism of HER2 overexpression is amplification of the HER2 gene on chromosome 17,6 which is typically detected by fluorescence in situ hybridization (FISH), although other in situ hybridization methods are now increasingly used (chromogenic and silver in situ hybridization) as well as array comparative genomic hybridization. For FISH analysis, 3 kits approved by the US Food and Drug Administration are available: 1) PathVysion (Abbott Laboratories, Abbott Park, Ill), 2) INFORM (Ventana Medical Systems, Tucson, Ariz), or 3) PHarmDx (DAKO, Glostrup, Denmark). The INFORM kit evaluates only the HER2 gene copy number, and results are based on the absolute HER2 signal count. The PathVysion and PHarmDx kits use 2 probes, a HER2 probe and a CEP17 (chromosome 17 centromere enumeration probe) hybridization control probe, and results are based on the HER2/CEP17 ratio.
Current American Society of Clinical Oncology/College of American Pathologists guidelines now define HER2-positive tumors as those with an average HER2 gene copy number of >6 gene copies per nucleus (for single probe assays) or as a HER2/CEP17 ratio of >2.2 (for double probe assays).3 Typically, similar conclusions are obtained with either the absolute HER2 signal count or the HER2/CEP17 ratio,7 but discrepancies can occur in cancers with increased CEP17 copy number (>3 copies per tumor cell). For example, tumors with increased CEP17 copy number and slightly increased HER2 copy number may be considered amplified by single probe assays but unamplified by double probe assays. It is estimated that this occurs in 2% to 9% of breast cancers.8, 9 However, the reported frequency of CEP17 copy number alteration in breast cancer varies, depending on the study population, selection criteria, and the definition of chromosome 17 polysomy (CEP17).10-13 In the literature, it is commonly assumed that an increase in CEP17 copy number is because of polysomy 17, and these terms have been used interchangeably. However, it is important to recognize that an increase in CEP17 signals does not necessarily represent a true polysomy (ie, gain of the entire chromosome), but rather may represent a focal pericentromeric gain or a partial polysomy.12, 14
We have investigated the effect of increased CEP17 signal number on the interpretation of FISH results in breast cancers diagnosed at our institution. We confirm that increased CEP17 signals may lead to discordant interpretations between the HER2/CEP17 ratio and absolute HER2 gene copy number in a significant proportion of cases.
MATERIALS AND METHODS
Retrospective review of pathology reports from Creighton Medical Laboratories (Creighton University, Omaha, Neb) identified 72 patients whose breast cancers had ≥3 CEP17 signals per nucleus on average by FISH. The cohort represented approximately 12% of all tested cases in the period 2003 to 2007. The study was approved by the Creighton University Institutional Review Board.
Fluorescent In Situ Hybridization
Formalin-fixed paraffin-embedded tissue sections were used in all cases.
Two-color FISH was performed on 3.5 μm-thick sections from the paraffin blocks. Before hybridization, sections were deparaffinized, dehydrated in 100% ethanol, and air dried. Commercially available Locus Specific Identifier HER2 probe (190 Kb Spectrum Orange directly labeled fluorescent DNA probe) and a CEP17 probe (5.4 Kb Spectrum Green directly labeled fluorescent DNA) were used according to the manufacturer's recommendations (PathVysion, Abbott Molecular, Des Plaines, Ill). Thirty nuclei were scored per sample, and the number of HER2 (orange) and CEP17 (green) signals were recorded. A ratio of HER2 to CEP17 >2.2 was defined as gene amplification; polysomy 17 was defined as ≥3 CEP17 signals per nucleus (average for 30 cells).
Immunohistochemical assay for HER-2/neu (IgG1, Clone CB11, Ventana Medical Systems) expression was performed using a commercially available detection kit and automated staining procedures (Benchmark, Ventana Medical Systems).
Automated Cell Imaging System (ACIS, ChromaVision Medical Systems, San Juan Capistrano, Calif) was used for measuring the percentage of cells with membranous staining of HER2 protein. This system combines color-based imaging technology with automated microscopy to provide quantitative information on the intensity and the percentage of cells with positive staining; pathologists reviewed the images on screen and selected tumor-rich areas for analysis. The Automated Cell Imaging System scoring system for HER2 protein was as follows: score <1.0 (negative, equivalent to 1+), score 1.0 to 2.4 (borderline, 2+), and score ≥2.4 (positive, 3+).
Single Nucleotide Polymorphism Array
Single nucleotide polymorphism (SNP) array karyotyping was performed on 2 selected cases with borderline score (2+) of HER-2/neu protein on immunohistochemistry (IHC) and polysomy 17 (≥3 signals of CEP17 per nucleus) on FISH. After tumor enrichment via manual microdissection, DNA was obtained from 10 μm paraffin sections as described previously,15 and 250K Nsp Assay Kits (Affymetrix, Santa Clara, Calif) were used according to the manufacturer's protocol, except for increased starting genomic DNA. One microgram of genomic DNA was digested with Nsp restriction enzyme, ligated to adaptors, and amplified by polymerase chain reaction (PCR) using a universal primer. After purification of PCR products with SNP Clean magnetic beads (Agencourt Biosciences, Beverly, Mass), amplicons were quantified, fragmented, labeled, and hybridized to 250K Nsp arrays. After washing and staining, the arrays were scanned to generate CEL files for downstream analysis.
Data acquired from the Affymetrix Gene-Chip Operating System v4.0 were analyzed using Affymetrix Gene-Chip Genotyping Analysis Software v4.1. Copy number analysis was performed with Copy Number Analyzer for Affymetrix GeneChip arrays v3.0, as previously described.16
A nonparametric chi-square test was used for testing associations between variables. For correlation purposes, the nonparametric Spearman correlation rank was used. All statistical tests were 2-sided, and P values <.05 were considered statistically significant. Statistical analysis was performed using the Statistical Package for Social Sciences software (v17.0; SPPS Inc, Chicago, Ill).
Clinicopathologic Characteristics of the Cohort
All but 2 patients were women. The patient's age ranged between 34 and 99 years (mean, 58.3 years). The study included 68 primary and 4 metastatic breast carcinomas. The majority of the cases were invasive ductal carcinomas of no special type (65 cases, 90.3%). The remaining cases included 3 mucinous carcinomas (4.2%), 1 case of invasive lobular carcinoma (1.4%), 1 case of mammary Paget disease in association with invasive ductal carcinoma (1.4%), and 2 cases of ductal carcinoma in situ (2.8%).
FISH Results and Interpretation
The HER gene and CEP17 copy number, HER2/CEP17 ratio, and HER2 protein expression results are summarized in Table 1 and Figure 1 (Euler diagram). Average CEP17 copy number for the cohort was 4.5 and ranged from 3.0 to 10.4 (Fig. 2). HER2 gene amplification as defined as the ratio of HER2/CEP17 >2.2, was identified in 21 cases (29.2%). All these cases had >6 HER2 signals per nucleus. More than 6 HER2 copies per nucleus were observed in an additional 12 cases without an increased HER2/CEP17 ratio, for a total of 33 cases (45.8% of all cases); using the criterion of >6 HER2 signals per nucleus as positive for amplification, these 33 cases would be categorized as HER2 amplified. These findings, therefore, demonstrate that discrepant interpretation of gene amplification status was detected in 12 (36.4%) cases when the number of CEP17 copies was taken into account. Of these 12 cases, HER-2/neu protein IHC was available for 10 cases: 3 cases had HER2 overexpression (score 3+), and 6 cases had borderline score (2+), whereas 1 case was negative (score 1+).
Table 1. Results of the Immunohistochemistry and Fluorescent In Situ Hybridization in a Cohort of 72 Cases with CEP17 Polysomy
HER2 Gene Copy No.
HER2 indicates human epidermal growth factor receptor 2.
>6 copies, 33 (45.8%)
>2.2, 21 (29.2%)
≤6 copies, 39 (54.2%)
<2.2, 51 (70.8%)
CEP17 Copy No.
HER-2/neu Protein Expression
>6 copies, 9 (12.5%)
≥2.4 (score 3+), 22 (36.1%)
≤6 copies, 63 (87.5%)
<2.4 (score 0 to 2+), 39 (63.9%)
Interestingly, HER2/CEP17 ratios <1.0 were observed in 9 (12.5%) cases, of which 3 (4%) cases had a HER2/CEP17 ratio ≤0.7. One of these cases had a ratio of <0.5 (1.71 HER2 signals and 4.06 CEP17 signals, ratio = 0.42), indicating that amplification of the centromeric region may not be accompanied by amplification of the HER2 gene region in some cases.
SNP Array Karyotyping and Interpretation
Cytogenomic arrays provide high-resolution, genome-wide copy number information. Affymetrix 250K Nsp mapping arrays contain 250,000 probes genome-wide, including 4854 probes on chromosome 17, and were used in 2 cases in our series to further investigate the relationship between CEP17, HER2 gene, and other loci on chromosome 17. SNP array analysis of the first case, with a FISH HER2/CEP17 ratio of 0.7 (2.3 of 3.3), revealed that copy number variability occurring along chromosome 17 may be undetected when using 1 or 2 FISH probes to determine chromosome 17 copy number. In this case, the copy number at the CEP17 locus is 3, whereas that at the HER2 locus is 2, generating a ratio of 0.67 (Fig. 3). Another case with similar FISH result (HER2/CEP17 ratio 0.7 [5.7 of 7.7]) showed discrete amplifications of both the centromeric region and the HER2 locus along with complex cytogenetic changes that included a relative loss of 17p and a relative gain of most of the 17q chromosome.
Correlation Between Protein Expression and HER2 Gene Status
HER2 protein expression results were available for 61 cases. Nine (14.8%) cases were negative (scores 0-1+), 30 (49.2%) cases were borderline (score 2+), and 22 (36%) cases were positive (score 3+).
HER2 protein expression positively correlated with both HER2 gene copy number and HER2/CEP17 ratio (P < .01, rs = 0.56 and 0.64, respectively). A trend toward positive correlation was found between HER2 protein expression and CEP17 copy number, but did not reach statistical significance (P = .067).
Notably, 6 (27.3%) of 22 cases with HER2 protein scores of 3+ had no HER2 gene amplification (ratio, <2.2). However, 4 of these 6 cases harbored >6 copies of the HER2 gene, fulfilling the absolute copy number criterion for HER2 gene amplification.3 In the borderline protein expression category (score 2+), only 4 of 30 cases had HER2 gene amplification (ratio, <2.2; 13.3%). None of the cases with score 0 to 1+ showed HER2 gene amplification by FISH.
HER2 status is routinely assessed in all patients with a new diagnosis of invasive breast carcinoma. However, the most accurate method of assessing HER2 status is yet to be determined, and interpretation guidelines3 as given by American Society of Clinical Oncology/College of American Pathologists may give conflicting results, depending on whether the laboratory uses a single probe kit or a double probe kit. The most recent American Society of Clinical Oncology/College of American Pathologists guidelines for HER2 testing define HER2 amplification by FISH as >6 HER2 gene copies per nucleus or a ratio (HER2 gene signals to chromosome 17 signals) of >2.2.3 Although this appears rather straightforward, abnormalities of chromosome 17 in breast cancer are frequent and may include whole chromosome copy number gains (polysomy 17) or losses (monosomy 17), focal copy number gains and losses, and other structural rearrangements.10 These abnormalities of chromosome 17 can lead to discrepant interpretations of FISH data, depending on which criterion is used.
The potential for such misinterpretations is significant, given that polysomy 17 is relatively common in breast carcinomas, although the reported frequency of this finding varies in the literature.10-13 In a recently published series by Vanden Bempt et al, >40% of breast carcinomas were found to harbor increased CEP17 copy number.11 Our recent study on apocrine carcinoma of the breast also revealed increased CEP17 copy number in 33% of the apocrine carcinomas of the breast.17 The present study also revealed a smaller overall proportion (12% of all tested cases) of increased CEP17 copy number cases, composed predominantly of invasive breast carcinomas of no special type. In addition, increased CEP17 copy number is frequently found in tumors showing HER2 overexpression, including those with a borderline (2+) score, as confirmed in our study.18-20 Our series included unselected (no prior IHC determination of HER-2/neu protein) and selected (equivocal IHC staining results) cases, reflecting our referral laboratory's mixture of cases received from different institutions.
HER2 status determination by FISH depends on the criteria used.21 Our study indicates that determination of HER2 amplification status may show discordant results, depending on whether CEP17 copy number was taken into account. Indeed, more than one-third of the studied cases harboring >6 copies of the HER2 gene did not show HER2 gene amplification (ratio, >2.2). Importantly, a majority of these cases had a borderline score (2+) on IHC, and therefore were not amenable for the targeted therapy. Similarly, increased CEP17 copy number appears to contribute to the discordant results between protein expression and gene amplification (IHC 3+/FISH negative), because 6 (27%) of 22 cases with a 3+ result on IHC did not show a HER2/CEP17 ratio >2.2. Notably, 4 of 6 cases demonstrated >6 copies of the HER2 gene. Therefore, these cases did not fit within the HER2-amplified breast carcinoma category.21,22 Taken together, a subgroup of borderline (2+) breast carcinomas represented a majority of the cases with increased CEP17 copy number in our study, and only 13% of the cases showed HER2 gene amplification. This finding is in line with previous studies that confirmed that breast cancers with an equivocal IHC score (2+) harbored CEP17 polysomy instead of HER2 gene amplification.18, 23,24
It is common in the literature to use the term polysomy17 when there is an increase in CEP17 signal by FISH. However, FISH analysis is a targeted assay and cannot assess the copy number of an entire chromosome. An increased number of CEP17 signals may represent a focal gain in the centromeric region of chromosome 17 rather than a true polysomy 17. Affymetrix 250K Nsp mapping arrays contain 4854 probes on chromosome 17. By generating a SNP array karyotype, one can discern between true polysomy 17 and focal gain of CEP17. Our reanalysis of published data (GEO dataset record GSE10099) revealed that true polysomy of chromosome 17 is a rare event, present in 1% of all analyzed cases.25 Most increases in CEP17 copy number by FISH are because of focal gains rather than true polysomy 17. This finding has been reported by others.14, 26
Interestingly, CEP17 copy number may have a predictive therapeutic value; increased CEP17 copy number appears to be a predictive marker for anthracycline-based chemotherapy in breast cancer.27, 28
Similar to gains in CEP17 copy number as seen on FISH testing, deletions of the CEP17 copy number do not necessarily correlate with deletion of the entire chromosome. In a previous study by Tubbs et al, HER2 monoallelic deletion (HER2/CEP17 ≤0.7) was demonstrated in 2% (12 of 742) of breast carcinomas.29 We likewise found deletion of the HER2 gene in a subset of cases with polysomy 17 (3 of 72, 4%). This finding is also supported by the SNP array karyotype of the 2 cases included in the present study.
In summary, we found that increased CEP17 copy number (≥3 copies of CEP17) is seen in 12% of breast tumors undergoing routine assessment of HER2 gene status. Half of these cases exhibited an equivocal (2+) score on IHC. Furthermore, a significant proportion of cases showing increased CEP17 copy number led to discrepant interpretations based on which criterion was used (HER2/CEP17 ratio vs absolute HER2 gene copy). However, positive gene dosage (>6 HER2 genes or HER2/CEP17 ratio >2.2), regardless of the evaluation method used, is positively correlated with HER2 protein expression. As a result of our findings, we propose that the average number of HER2 genes per nucleus be reported alongside the average HER2/CEP17 ratio, to accurately identify all patients eligible for trastuzumab treatment.
We thank Mirza Bašić, BSc, MSc for his excellent technical support and Shera F. Kash, PhD for analyzing breast cancer data.
CONFLICT OF INTEREST DISCLOSURES
Dr. Jill Hagenkord is a cofounder and chief medical officer in iKaryos Diagnostics, which uses Affymetrix and single nucleotide polymorphism arrays for commercial testing.