Molecular Biological Determination of HER2 Status Using Both DNA and RNA Approaches. A Concordance Study with IHC Assessment

1. Introduction

The human gene for epidermal growth factor 2 (HER2) [also known as ERBB2 or epidermal growth factor receptor (EGFR)2] encodes a 185 kDa transmembrane glycoprotein with tyrosine kinase activity. It is overexpressed and/or amplified in 15-20% of all breast cancers. Amplification of HER2 has been shown to be related to tumor size, lymph node metastasis, high S-phase fraction, aneuploidy, and low levels of steroid hormone receptors, factors that may increase the rate of tumor cell proliferation. Recent studies show a str

ong correlation between HER2 gene amplification and tamoxifen resistance. Anti-HER2 therapy consists of the administration of the monoclonal antibody trastuzumab (Herceptin), which is effective in the case of metastasis and/or in combination with chemotherapy [1].

According to the guidelines of the American Society of Clinical Oncology/College of American Pathologists (ASCO/CAP), there are now established techniques for determining HER2 status in primary tumours based on immunohistochemistry (IHC) or fluorescence/chromogenic in situ hybridization analysis (FISH/CISH) [2]. Nevertheless, there is increasing evidence of equivocal samples with a negative FISH result, which are therefore not eligible for antiHER2 therapies. Initial estimates put the number of such cases at around 8 [3]

The aim of the study is to introduce the determination of HER2 status using molecular biology methods, both at the level of the number of copies of the HER2 gene and at the level of its gene expression, so that we can evaluate with the advantage of two levels (both DNA and cDNA), which will help us to read particularly ambiguous and doubtful cases accurately.

2. Materials and methods

We examined samples of formalin-fixed, paraffin-embedded (FFPE) primary breast cancer tumors obtained from the Pathological Anatomy Department of Bulovka College Hospital (FNB). The samples were routine surgical specimens that were fixed in formalin, processed and stored according to standard histologic protocols and the rules of the local medical ethics committee of the FNB. The selection of tumor material was based on a previous clinical examination for HER2 performed by immunohistochemistry (IHC).

2.1. DNA/RNA preparation

On the original slides of each of the samples, a representative area containing tumor cells was marked by a pathologist (Dr. Z. Špurkova). Thereafter, 3-5 microtome slices (diameter 5 µM) were obtained from the corresponding area in each paraffin block, and RNA/DNA was isolated using CE-IVD certified diagnostics: FFPE RNA Purification Kit Dx (Norgen Biotek, Thorold, Ontario, CA) or AmoyDx FFPE DNA/RNA kit (Xiamen, China). The clear advantage of the AmoyDx kit was the simultaneous isolation of both RNA and DNA without affecting RNA yields compared to the Norgen Biotek extraction. The concentration of the DNA/RNA was determined using a Qubit 4 Fluorometer (Thermo Fisher Scientific, Waltham, MA, USA). Fluorometric analysis was performed by Quant-iT^TM dsDNA HS Assay Kit Quant -iT^TM RNA Assay Kit (Thermo Fisher Scientific, Waltham, MA, USA) for, respective DNA and RNA.

2.2. Targets

Real-time PCR assays for the target gene HER2 and the reference gene APP (encoding amyloid precursor protein, were run on both levels, as DNA so cDNA. As expected, HER2 overexpression was present in all tumors with HER2 gene amplification but was uncommon in breast cancers without gene amplification [4]. The APP gene has been selected as a „housekeeping“/“backbone“ marker. Design of primers for gene expression study was performed in the ExonSurfer software allowing to avoid amplification of genomic template (Table n. 1) [5].

2.3. Reverse transription and real-time PCR of cDNA (gene expression level)

cDNA was synthesized using ProtoScript II First Strand cDNA Synthesis Kit (New England BioLab., Ipswich, MA, USA) according to the manufacturer’s instruction. The Random Primer Mix yields shorter cDNAs on average and can be used for the detection of multiple short RT-PCR products, appropriately designed for fragmented nucleic acids.

The 20-µl reaction mixture contained 2x Luna® Universal qPCR Master Mix (New England BioLab., Ipswich, MA, USA), 0,05 µM forward primer, 0,05 µM reverse primer (all from Generi-Biotech s.r.o., H. Kralove, CZ) and 2,5 ul DNA or cDNA. See the New England Biolabs for the exact protocol [6]. In general, the volume of cDNA product should not exceed 1/10 of the PCR reaction volume. The PCR program started with one cycle at 95°C for 2 mins. The amplification was run for 40 cycles with short denaturation at 95°C for 15sec, and annealing with extension at 60°C for 45 sec. All reactions were performed in a Connect CFX real-time PCR system (Bio-Rad, Hercules, CA, USA).

2.4. IHC assessment

A total of 10 breast cancer tissue samples were analysed immunohistochemically using the c-erbB-2 oncoprotein antibody (DAKO, Glostrup, Denmark) with internal (on slide) positive and negative controls. Staining was performed in a Ventana immunohistochemistry staining system.

Immunostaining was evaluated by a pathologist and was performed according to ASCO guidelines [7]: negative for 0 (no or week incomplete staining <10% tumour cells) and 1+ (faint or barely perceptible incomplete membrane staining > 10% tumour cells); equivocal for 2+ (weak to moderate complete membrane staining >10% tumour cells) and positive for 3+ (>10% tumour cells with circumferential complete, intense membrane staining). The samples were then sent to the reference laboratory of the Institute of Pathology VFN Prague (Czech Republic), where the determination was performed using the certified Ventana HER/Neu kit, clone 4B5. Some samples with a negative and indeterminate result (positivity 0+,1+,2+) were subsequently examined by FISH, whereby no amplification of the HER2 gene could be detected (Table 2). FISH was not performed on definitely positive tumors (3+) (samples no. 3, 7).

3. Results

3.1. Assessment of HER2 status

We developed a protocol to determine the copy number of the HER2 gene by quantitative RT-PCR/qPCR using SYBR green dye I, with which we examined 10 blind DNA samples from tumour tissue of breast cancer patients. The content of the target DNA in tumor samples were quantified by using standard curves according to the Control Genomic Human DNA (ref.no. 4312660 Themo Fisher Scientific, Waltham, MA, USA) and the same PCR protocol as for gene expression profiling (written in the section 2.3.). In order to correctly determine the ratios in FFPE samples, we were forced to replace the 168bp amplicon of the TBP reference gene (TATA-binding protein gene) with a shorter one (APP, 124bp, see Table 1). HER2 copy number in each sample, was normalized on the basis of its content of the „backbone“ reference gene that is located at 21q21, which has not been found to exhibit alterations in breast cancer patients [1]. and amplification efficiency was found to be similar to HER2 gene [8]. Standard curves for both the targets; HER2 and the APP reference gene for each run were constructed using threefold serial dilution ranging from 1 ng/µl to 50 pg/µl of the Control DNA (cat.no. 4312660, Thermo Fisher Scientific, Waltham, MA, USA) and CFX Manager software (Bio-Rad, Hercules, CA, USA). Supplementary file S2 contains such runs with constructed calibration curves, then HER2 copy number was calculated as a ratio HER2 (ng/μl) to APP (ng/μl). To test the reproducibility, we analyzed ten samples in duplicates. Amplification of HER2 has been already defined as 5 copies of this gene above average ploidy of the tumor sample [9]. However, as we found small differences between APP and HER2 amplification efficiency, the threshold for a positive result was appropriately shifted (Supplementary Figure S1a).

The samples 3 and 7 were strongly positive (in agreement with IHC assessment 3+) and the sample no. 4 appeared slightly above the cut-off value (Figure 1a). Dual concept for evalutation revealed concordant results also on gene expression level. (Figure 1b).

Relative quantification of HER2 expression was performed using the ΔCt method. For HER2-positive samples, the increase in the difference between the Ct values of HER2 and APP transcripts in tumour tissue essentially corresponded to the overexpression of the HER2 gene (Figure 1b). The real-time PCR gene expression data for another positive sample (no. 7) are shown in Supplementary Figure S3.

Different studies demonstrate that quantitative real-time PCR approaches are valuable tools for the assessment of HER2 gene overexpression and can provide a reliable alternative to FISH and IHC determinations [10,11,12]. Generally, PCR-based assays can detect changes in HER2 gene number as well as gene expression. Q-PCR assays based on DNA/cDNA analyses are describing as sensitive enough, faster, easy to perform, specific and cost-effective to measure HER2 alterations and able to analyse multiple samples simultaneously [13]. The high sensitivity of real-time qPCR means that even minute amounts of DNA or RNA can be detected in FFPE tissue [4], opening up the possibility of performing retrospective clinical and molecular studies on the large sample archives stored in pathology institutes [14]. The design of short amplicon primers and the use of random primers in reverse transcription appear to be crucial both for long-term archived FFPE blocks and for the precise calculation of ratios between the focused biomarkers. In addition, linkage to reference genes enables the HER2-status to be monitored in time course and therefore it can reflect on therapy efficiency. Determined gene expression ratios are also useful in HER2-Low breast cancer patients [15].

Strongly positive samples from IHC decising (samples no. 3, 7) were also positive using molecular method determination, consistently at both DNA and RNA levels. Regarding to the requirements on content of tumor tissue, usually it is about 20% [16] it seems in agreement with another studies [3] that molecular approaches are rather sensitive and less demanding on percentage of tumour cells, with fractions of amplified cells as small as 5%. Our system offers a decisive advantage: the possibility to perform a double evaluation, which facilitates the interpretation of the results, enhances result interpretation flexibility and quality control (Figure 2).

Three samples that were classified as positive by real-time PCR were also positive in the IHC detection. This indicates that real-time PCR does not give false-positive results, which is consistent with a previously published study [17]. We hope that the LDT -assay developed by our laboratory shows better analytical performance than FISH (also detectable in the current study) and would be useful also in patients with HER2-low breast cancer (sample no. 4), so that more patients could benefit from anti-HER2 treatment. The consideration here is that the combination of the recommended IHC and FISH/CISH methods still leads to results that fall into the equivocal range and could exclude true positive samples, and that qPCR methods could be useful as a complement.

5. Conclusions

While IHC requires knowledge for reading to achieve reproducibility, long-term skilled techniques and use of expensive antibodies and reagents that threaten to pro-expire; on the contrary, molecular-biological approaches would consider where large numbers of patients are not expected. We anticipate their use especially in small laboratories without the necessary equipment (a fluorescent microscope), everywhere the risk of overexposure to chemistry increases and the huge manual workload may compromise the quality of results.

The increasing number of publications and new diagnostics (e.g. MammaTyper®, OncotypeDX® test, Xpert® Breast Cancer STRAT4) on the successful implementation of HER2 determination by molecular biology methods speaks more and more for the idea that not the inherent limitations of the method but its good standardization is important for its complementary use as a third alternative.

6. Patents