Background: HER-2 is the target for antibody-based treatment of breast cancer (trastuzumab), which is highly successful in both advanced disease and the adjuvant setting. HER-2 can be analysed by fluorescence in situ hybridisation (FISH) for gene amplification or immunohistochemistry (IHC) for protein overexpression.
Aim: As both methods are known to be influenced by previous tissue processing, to analyse the applicability of both FISH and IHC to decalcified bone metastases of breast cancer.
Methods: A tissue microarray (TMA) was constructed from 149 breast cancer bone metastases. Consecutive TMA sections were analysed by FISH (PathVysion) and IHC (HercepTest).
Results: FISH analysis was interpretable in 113 (85.0%) cases. Amplification was seen in 14 (12.4%) interpretable metastases. HER-2 positivity on IHC analysis was 3+ in 9.8% of cases and 2+ in 11.3%. A comparison of the two techniques revealed high concordance. Of the 14 cases of amplification, 10 (71%) showed 3+ IHC staining, two (14%) showed 2+, one (7%) showed 1+, and one (7%) showed 0+. Three of the four amplified cases that did not show 3+ IHC staining had an equivocal FISH result, with a HER-2/centromere 17 ratio of 1.8–2.2. Of the 13 cases that showed IHC 3+ staining, amplification was present in 10 (77%).
Conclusions: HER-2 FISH analysis has an excellent success rate in highly standardised EDTA-decalcified bone metastases, suggesting that this method is easily applicable to decalcified tissues. The high concordance between IHC and FISH suggests that HER-2 IHC may be equally applicable to EDTA-treated tissues as to the usual formalin-fixed tissues.
Statistics from Altmetric.com
HER-2 gene amplification and protein overexpression occurs in ∼20% of breast cancers and is linked to poor prognosis for patients with these tumours.1–4 HER-2 is also the target of an antibody-based treatment (trastuzumab) which is routinely used in metastatic HER-2-positive breast cancer.5 More recently, adjuvant trastuzumab application has also been shown to be dramatically effective in HER-2-positive breast cancer.6 Because of this, HER-2 status is now routinely assessed in breast cancer biopsy specimens, either by immunohistochemistry (IHC) or fluorescence in situ hybridisation (FISH). Many studies have compared IHC and FISH analysis of HER-2.7–11 Overall, the data suggest that FISH has superior reproducibility,12–14 but IHC has also been successfully applied, especially in laboratories with very high throughput.15 16 In general, FISH or IHC analysis is performed on tissues from primary tumours. However, metastases are also occasionally analysed. Reasons for analysing metastases include the absence of retained primary tumour material at the time of clinically evident metastases and attempts to analyse the current tumour cell population in cases of a recurrence after extensive previous treatment or recurrences many years after treatment of the primary tumour.
Approximately 70% of patients with advanced breast cancer develop bone metastases.17 These tissues typically contain mineralised osseous material and need to be decalcified before histological analysis. As molecular tissue analyses are substantially influenced by previous tissue processing, it is possible that the results of HER-2 analysis may be affected by decalcification steps. To determine the effect of tissue decalcification on HER-2 analysis by FISH and IHC, we studied a consecutive series of 149 breast cancer bone metastases with Food and Drug Administration (FDA)-approved reagents for both techniques.
MATERIALS AND METHODS
A consecutive set of 149 breast cancer bone metastases, which had been paraffin-embedded and from which the retained blocks were large enough to be placed into a tissue microarray (TMA), were included in the study. All blocks were collected at the Institute of Pathology, University Medical Center Hamburg-Eppendorf between January 1996 and February 2007. Patient age ranged from 39 to 97 years (mean 61.8 years). All tissues had been EDTA-decalcified according to a highly standardised procedure. A 10% EDTA solution in distilled water, pH 7.5, was used in an ultrasonic bath (Omnilab 300, Omnilab, Bremen, Germany; cooling system Lauda ecoline RE 206, 15°C). To avoid overdecalcification, all tissues were radiographically (Cabinet X-ray System; Faxitron X-ray Corporation, Lincolnshire, Illinois, USA) controlled. Plastic capsules with tissue samples were radiographed (fig 1) every 24 h and considered appropriately decalcified when radio-opacities consistent with mineralised bone tissue were absent on contact radiography. Decalcifying of fragmented cancellous bone specimens (5–15 mm) took 4–7 days depending on its mineral content.
H&E-stained sections were used to define representative tumour regions. Tissue cylinders (0.6 mm in diameter) were then punched from tumour-containing regions of the donor block using a home-made semi-automated tissue arrayer. Control samples were from the following histologically normal tissues: breast (n = 10), testis (n = 2), endometrium (n = 2), skin (n = 2), skeletal muscle (n = 2), tonsil (n = 2), heart muscle (n = 2), colon mucosa (n = 2), lung (n = 2), lymph node (n = 2), prostate (n = 2) and kidney (n = 2). Figure 2 shows a representative H&E-stained section containing 181 stained tissue samples (fig 2A) and a single spot of an H&E-stained metastasis (fig 2B). Consecutive 4 μm sections were taken using the Paraffin Sectioning Aid System (Instrumentics, Hackensack, New Jersey, USA) and used for FISH and IHC.
For proteolytic slide pretreatment, a commercial kit was used (paraffin pretreatment reagent kit; Vysis, Downers Grove, Illinois, USA). A spectrum orange-labelled HER-2 probe was used with a spectrum green-labelled centromere 17 reference probe (PathVysion; Vysis-Abbott, Abbott Molecular, Abbott Laboratories, Abbott Park, Illinois, USA). Before hybridisation, sections were deparaffinised, air-dried, dehydrated and then denatured for 5 min at 74°C in 70% formamide/2 × standard saline citrate (SSC) solution. After overnight hybridisation at 37°C in a humidity chamber, slides were washed and counterstained with 0.2 μM 4′,6-diamidino-2-phenylindole in an anti-fade solution. In contrast with the manufacturer’s directions, we did not count HER-2 and centromere 17 signals in tumours with unequivocal amplification status. Instead, we estimated the mean number of HER-2 and centromere 17 signals for tumour samples such as those described previously.16 Only in the case of borderline results did we score 20 tumour cells for HER-2 and centromere 17 counts. Tumours were reported as positive (HER-2/centromere 17 ratio >2.2), equivocal (ratio 1.8–2.2) or negative (ratio <1.8).18
The HercepTest (Dako, Glostrup, Denmark) was used exactly as suggested by the manufacturer. Antigen retrieval of the deparaffinised tissue sections was performed in a water bath at 95–99°C for 50 min followed by peroxidase blocking and incubation with the prediluted primary antibody. Cell line test slides provided by the manufacturer were used as positive and negative controls. One immunostained slide was scored by one pathologist (GS) on a scale of 0 to 3+ according to the FDA-approved guidelines.19 Discontinuous membrane staining was scored as 1+, unequivocal continuous membrane staining with weak to moderate intensity was scored as 2+, and strong and complete membrane staining of >30% of invasive tumour cells was scored as 3+.
Contingency table analysis and Fisher’s exact tests were used to study the relationship between HER-2 IHC and FISH.
A total of 133 of 149 arrayed cancer samples (89.3%) contained tumour tissue on the consecutive slides used for IHC and FISH. The remaining 16 samples were not informative because of absence of tissue on the TMA (n = 1) or a lack of unequivocal tumour cells in the arrayed samples (n = 15).
The TMA was only hybridised once. A total of 113 of 133 (85.0%) tumour-containing tissue spots were interpretable by FISH. The amplification status was determined by estimate in 105 cases and by manual scanning of 20 cells in eight cases. HER-2 amplification was present in 14 cases (12.4%). Of these, 12 cancers showed positive FISH results including HER-2 gene clusters, and two cancers had equivocal amplification without distinct clusters and a HER-2/centromere 17 ratio of 1.8–2.2. Figure 3 shows examples of HER-2 amplified and non-amplified metastases.
Of the 14 cases of HER-2 gene amplification, 12 showed two copies of chromosome 17, and two showed three copies of centromere 17. In this study, all 11 (9.7%) cases of polysomy were identified among the metastases with FISH-negative results.
All non-informative spots on TMA were interpreted in FISH analysis. These had an estimated score of two centromere 17 and two FISH HER-2/neu copies, but were classified as non-informative because of an insufficient number of tumour cells in the punched tissue and/or significant numbers of signals that were weak and/or occurred over the cytoplasm.
Table 1 shows the IHC results in comparison with the FISH data. Of 133 interpretable cases, there were 13 (9.8%) with 3+ staining, 15 (11.3%) with 2+ staining, 38 (28.6%) with 1+ staining, and 67 (50.4%) with 0+ staining. This positivity rate was slightly higher than for FISH. Only 10 (76.9%) of the 13 cases with 3+ staining were amplified. On the other hand, concordance with IHC results was good for the 12 tumours with classical amplification containing gene clusters (p<0.001). IHC staining was 2+ for two of the tumours and 3+ for the remaining 10. Furthermore, neither of the two cancers with equivocal (ratio 1.8–2.2) amplification showed 3+ overexpression; one showed 1+ positivity, and the other showed no immunostaining (0+). Figure 2C shows an example of a non-amplified but IHC 3+ positive cancer.
This study shows that both IHC and FISH can be applied to EDTA-decalcified bone metastases. The success and concordance rates of the two methods are comparable to their application to non-osseous non-decalcified formalin-fixed tissues. To avoid overdecalcification, decalcification at our institution is performed with daily radiographic monitoring of the degree of progressive demineralisation of the bone specimens. Because of the use of such specific and highly standardised processing, we were unable to further define the effects of other methods of bone tissue decalcification on the reproducibility of the IHC and FISH results.
The frequency of HER-2 amplification observed in this series of breast cancer bone metastases (12.4%) is in the range reported for primary breast tumours or nodal metastases in studies using TMAs for analysing large breast cancer cohorts.20–22 This may argue against a particular role for HER-2 amplification in the process of bone metastasis in breast cancer as previously suggested by reports of HER-2 amplification in isolated tumour cells in bone marrow derived from HER-2-negative breast cancers.23
The success rate of fluorescence in situ hybridisation (FISH) for analysing HER-2 in breast cancer bone metastasis samples that have been carefully decalcified is very high.
Similar results can be obtained for immunohistochemical (IHC) analysis of HER-2 on carefully decalcified tissue.
However, because of the variability in tissue processing for IHC and the difficulty in fully standardising decalcification, the use of FISH may be preferable for routine laboratory HER-2 analysis of bone metastases.
Although only one TMA hybridisation experiment was carried out, HER-2 FISH analysis was successful in 85.0% of cases. This is higher than in other TMA FISH studies performed in our laboratories. Using the same HER-2 FISH probe, we previously had a success rate of 77%.20 We can assume that some tissue samples that could not be interpreted in our TMA analyses would become “analysable” if large sections were used, because these would probably contain tumour areas that had suffered less damage during tissue handling than the randomly selected small regions placed in our TMAs. It is also certain that additional hybridisations using modified protease pretreatments would have further increased the success rate in this study. It is noteworthy that decalcification of our material was highly standardised by the use of daily radiography of tissue samples in order to control bone density status. Under such conditions, our data show that FISH is as applicable to decalcified bone tissue as to commonly processed tissue from primary tumours.
The concordance rate of FISH in cases with 3+ IHC was 76.9%. This is what is typically achieved in formalin-fixed primary tumours. For example, in the BCIR6 005/006 trials, in which more than 2000 breast cancer samples submitted to our reference laboratory from all five continents were re-evaluated, we confirmed amplification in 78.2% of 3+ IHC cases.24 A rate of false-positive HER-2 IHC results of ∼20% is concerning. It is known, however, that IHC results are substantially influenced by preanalytical tissue processing. The difficulties of standardising IHC performed in paraffin-embedded tissue specimens have been reported in the literature on HER-2 in breast cancer, with frequencies of positivity ranging from 2% to more than 50%.12 25–27 It was hoped that the use of a standardised test with standardised control cell lines would solve this problem. It soon became clear that this kit could not prevent errors.27 28 Frequencies of HER-2 positivity in the range 30–50% have also been reported in studies using the FDA-approved HercepTest.7 12 29 It is now clear that inappropriate fixation of breast cancer specimens can dramatically alter IHC testing and lead to false-positive HER-2 results.30–32 To compensate for false positivity due to fixation problems, a modified scoring system has been recommended that takes into account the level of non-neoplastic epithelium, which was found to provide greater specificity.33 However, this is obviously not possible in HER-2 analyses of metastases. Our data suggest that, at least in cases in which decalcification is performed very carefully, preservation and accessibility of the HER-2 protein may be comparable to those in formalin-fixed primary tumours.
False-negative HER-2 IHC results are also well documented and vary in frequency from antibody to antibody.27 In a study using tissue microarrays, we found an IHC HER-2 positivity (2+/3+) rate of 15.6% in samples that were interpretable by FISH, but only of 7.6% in cases that were not interpretable by FISH.34 The most likely explanation is that “uncharacterised tissue damage” led to both non-interpretable FISH assays and a decrease in IHC immunostaining in a large fraction of the tissues. The use of slides that are not freshly cut can also lead to substantial discrepancies in HER-2 IHC results.35 Not all cases with negative IHC and positive FISH results must be considered as false-negative IHC cases, however. Both IHC-negative FISH-positive metastases found in this study had an equivocal FISH ratio of 1.8–2.2 and HER-2 copy number of 3.10 and 5.00. Cancers with such a low level of amplification are unlikely to express HER-2 at the levels observed for classical amplified cancers with 20–50 HER-2 gene copies. The clinical significance of these cases is not known, as no studies on the response to trastuzumab for low-level amplifications have been undertaken.
In summary, the success rate of FISH is very high for breast cancer bone metastasis samples that have been carefully decalcified. The observed concordance of HER-2 FISH and IHC data suggests that the reliability of HER-2 IHC is good if the technique is applied to EDTA-decalcified tissues, with daily radiographic monitoring to avoid overdecalcification and tissue damage. However, because of the variability in tissue processing for HER-2 IHC and the difficulty in fully standardising decalcification in routine laboratory procedures, the use of primary FISH may be preferable for bone metastases.
Competing interests: None.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.