Comparative Polymerase Chain Reaction Analysis of c-myc Amplificationon Archival Breast Fine-Needle Aspiration Materials¹

FNA⁴is a simple, fast, and cost-effective method for evaluating breast lesions (1). However, morphological analysis alone on FNA materials may not be accurate for defining premalignant lesions. Detecting these lesions is important for defining an individual’s risk for developing breast cancer (2). FNA materials potentially provide several advantages over tissue samples for the evaluation of tumor markers. These advantages include the minimally invasive nature of the procedure and the fact that quantitative analysis of biomarker expression can be carried out because the cells obtained by FNA are fresh and enhancing controlled fixation and biomarker quantitation.

The overall aim of this project was to develop a schema whereby multiple biomarkers can be analyzed using the minimum amount of cytological materials. Multiple biomarker analysis is essential for individual risk assessment of breast cancer development because breast cancer, like any other epithelial malignancy, develops through multiple pathways. Two separate studies were carried out on the same FNA smears. The first one was to measure multiple biomarkers (DNA ploidy,G-actin, and p53) simultaneously on single-cell basis using the state-of-the-art QFIA technique (3). The purpose of the present study was to determine whether genomic DNA can be extracted from the cells on the FNA smears so that biomarkers such as c-myc amplification can be analyzed using a novel,semiquantitative, cPCR technique. The preliminary results of the present study will also determine whether c-mycamplification can be used as an intermediate end-point marker for breast cancer.

The oncogene c-myc is a key regulator of cell cycle progression (from G₁ into S phase), and its overexpression can induce either cell proliferation or apoptosis (4). Using various techniques previously, c-mycoverexpression has been detected in 20–60% of primary breast cancers and in 60% of benign biopsies from patients who subsequently developed breast CA (5, 6). These findings indicate that c-myc amplification may be an important biomarker for breast cancer. However, c-myc amplification has not been evaluated on breast FNA samples, particularly archived materials.

In this study, archived FNA smears collected from 1995 were used. These smears had previously analyzed by QFIA analysis prior to genomic DNA extraction for PCR analysis. The results of both of the studies demonstrate that it is feasible to study multiple biomarkers with different methods on archival FNA smears. Our approach of combining in situ fluorescence labeling and image analysis with subsequent DNA analysis using PCR technique provides a powerful approach for analyzing multiple biomarkers for individual breast cancer risk assessment.

A detailed description of FNA materials used for this study and the results of a two-year follow-up study of these cases have been published elsewhere (7). In brief, there were 265 incidents of breast FNA at University of California-Los Angeles Medical Center in 1995. These FNA materials were reviewed systematically, and the cytological interpretation of the cases was confirmed by two cytopathologists (S. K. A. and S. L. H.). Our initial design was to carry out two sets of QFIA analysis; each set required two alcohol-fixed slides (one served as negative control). Therefore, only those cases who had at least four alcohol-fixed slides were carried out by QFIA analysis for DNA ploidy, p53, and G-actin (3). This yielded a total of 72 cases (20 CA, 18 PBDA, and 34 benign) from which 56 cases had an adequate number of cells on the smear to carry out the QFIA analysis. Of 16 cases that had an inadequate number of cells, 3 (15%) were in the CA group, 4 (22%) in the PBDA group, and 9 (26%) in the benign group. The difference of proportion of inadequate cases among each group did not reach statistical significance (P > 0.05 byχ ² test).

After QFIA analysis, the smears were used for the analysis of c-myc amplification using cPCR technique. The QFIA analysis was strictly limited on the epithelial cells. To minimize the contamination problem caused by inflammatory cells or other cells types, areas of slide that contained epithelial cells were marked with a doting pen, and they were specifically scraped into Eppendorf tubes for subsequent PCR analysis. From the 56 QFIA-analyzed cases, 46 yielded a sufficient amount of genomic DNA for PCR analysis. Among the 10 cases that did not have sufficient material, 4 (13.3%) were benign,3 (16.6%) were PBDA, and 3 (12.5%) were CA; therefore, there was no disproportion of benign versus malignant cases among those patients who did not have sufficient material for PCR analysis. An additional seven cases that did not undergo QFIA analysis (because only dried smears were available) were also included in the c-mycanalysis. This study was approved by the Human Subject Protection Committee of University of California at Los Angeles. To ensure patients’ confidentiality, patients’ name and other identification information were removed before the analysis.

Genomic DNA was extracted from FNA smears, many of which had undergone QFIA studies as described above, using a slightly modified method by de Melo et al. (8). Cells were first scraped into a sterile 1.5-ml Eppendorf tube using a razor blade. Typically, from one to two cellular smears, about 10 μg of genomic DNA was produced. The resulting pellet was resuspended in 400 μl of 6 m guanidium hydrochloride, 30 μl of 20% sodium sarcosyl, 30 μl of 7.5 m ammonium acetate, and 10 μl of proteinase K (10 mg/ml). Then, the entire mixture was heated at 40°C for 1 h. An additional 10 μl of proteinase K was added to the mixture, and the new mixture was heated for another hour at 60°C if the powder was not completely dissolved.

cPCR analysis of c-myc oncogene amplification was carried out according to the protocol of Ribot et al.(9), with minor modifications. This method determines the amplification status of a given gene by comparing the signal intensity of PCR product in a specimen DNA to the control DNA. The rationale of this technique is that the signal intensity of the PCR products reflects the copy number of the gene, and, by comparing the signals in sample versus control DNA, the presence of gene amplification in a specific specimen can be detected. However, instead of using peripheral blood DNA from a normal individual as a control source as proposed by Ribot, DNA from two cell lines were used. One cell line (HL-60) is a high c-myc expressor, and the other(K562) is a known low expressor for c-myc. These cells served as positive and negative controls, respectively, for each batch of experiments. To be noted, these cells were treated in a way simulating the actual FNA smears (i.e., cells were cytospined on slides, alcohol-fixed, and stained with regular Papanicolaou before QFIA analysis).

The cPCR analysis required a calibration step using a control gene that was not amplified in tumor DNA. In our study, GAPDHwas used as the control gene, and the up and the down primer sequences for this gene were GCC TGC TTC ACC ACC TTC TTG and GTC CAC TGG CGT CTT CAC CAC, respectively. For c-myc, the up and the down primer sequences were ATG CCC CTC AAC GTT AGC TT and GTG GGC AGC TCG AAT TT,respectively. Furthermore, the PCR amplification was performed on three different dilutions (1:10, 1:5, and 1:2.5) of the DNA master solution(5ng/μl). The purpose of using multiple-dilution strategy is two-fold: (a) it served as an internal control for the reproducibility; and (b) it kept the PCR amplification within the exponential phase. For the calibration step, 10 μl from each of the three dilutions of the template DNA master solution was used to undergo PCR reaction for the control gene GAPDH. The intensity of the bands was measured by a densitometer, and the difference of the intensity between sample and control (HL-60 and K562)DNA was used to calibrate the amount of diluted DNA added to the reaction. PCR reaction using the calibrated amount of diluted DNA was repeated for the control gene to confirm the accuracy of the calibration. For the subsequent analysis of c-mycamplification, the same calibrated amount of diluted template DNA was used. The PCR cycles for both the calibration and the analysis steps were 94°C for 40 s, 55°C for 1 min, and 72°C for 3 min for 35 cycles with 1.5 mmMgCl_2.

c-myc amplification was defined as positive if the mean DNA band intensity in a sample was higher than the DNA band intensity in negative control cells (K562). The degree of c-mycamplification was further graded into +, ++, and +++ based on densitometer measurements. χ² analysis was carried out to determine the significance of the difference of c-myc amplification among three study groups.

Table 1 lists the study cases, their FNA and tissue diagnosis, and their corresponding biomarker results. Among the 53 cases analyzed, 24 were benign (fibroadenoma or others), 12 were PBDA, and 17 were considered malignant or suspicious for malignancy. Twelve of the 17 cases with a malignant diagnosis were confirmed by the subsequent tissue diagnosis. Among the 17 cases of malignancy, there were 10 cases of infiltrating ductal adenocarcinoma, 2 cases of intraductal papilloma with carcinoma-in situ components, and the remaining cases were clinically diagnosed as late-stage metastatic breast cancer.

Fig. 1 shows representative PCR results for c-myc amplification in positive control cells (HL-60), negative control cells(K562), a benign FNA sample (1-BN), and a malignant FNA sample (2-CA). The top panel shows the PCR products of the control gene GAPDH before calibration. The band intensity increased as the dilution of the initial DNA added to the reaction buffer decreased. The lower the template DNA was diluted,the denser the band for GAPDH was after PCR amplification. The intensity of the band densities was measured by a densitometer, and the intrasample variation of the band intensity for each dilution was used to calibrate the amount of template DNA added to the PCR reaction. The middle panel shows the GADPH gene after calibration, and the lower panel shows the results of c-myc amplification after calibration. In this representative example, the malignant sample(2-CA) clearly showed higher band intensity than the negative control (K562) and the benign sample(1-BN) and was defined as positive (+) for c-mycamplification.

Table 2 summarizes the results of c-myc amplification measured by cPCR analysis in all of the 53 samples. It was observed that 16(94.1%) of 17 malignant lesions, 5 (41.7%) of 12 PBDA, and 4 (16.7%)of 24 other benign lesions (fibroadenoma or fibrocystic disease) had positive c-myc amplification. The overall difference of c-myc amplification among these groups was highly significant by χ² analysis (P =0.0002).

Table 3 compares c-myc amplification with other biomarkers (G-actin, p53, and DNA ploidy) among three groups of FNA samples. The c-myc was concomitantly negative in 14 of 20 cases with G-actin, 13 of 19 with p53, and 11 of 15 with DNA. Of about 20 cases, in which both c-myc and QFIA markers were successfully analyzed, none of the cases were positive concomitantly for c-myc with either G-actin or p53, except one case with DNA ploidy. In contrast, of 14 cases in the cancer group, c-myc was positive concomitantly in 13 cases with G-actin, 5 with p53, and 11 with DNA ploidy.

The breast cancer model provides a significant opportunity for the development of tumor markers. These tumor markers can be used to define an individual’s risk of developing breast cancer, and they are urgently needed in both clinical and epidemiological settings (10). There are hundreds of potential tumor markers reported in the literature; however, to determine the validity of these markers is rather a complex process (11). It is well recognized that multiple markers should be used for accurate determination of individual risk because multiple pathways are involved in the development of a malignant tumor (12). Therefore,novel methods are needed for evaluating multiple markers simultaneously using limited sample volume such as FNA material.

This laboratory has developed an approach that allows multiple marker analysis to be carried out on both the phenotypic and the genotypic level on single cell basis. The first step is to quantitatively evaluate multiple phenotypic tumor markers (G-actin/p53/DNA)on single-cell and in situ basis simultaneously, using a novel QFIA method. The second step, as we reported in the current study, is to obtain genomic DNA from the cells that had undergone QFIA study for genotypic marker analysis such as c-mycamplification using PCR technique. The results of these studies indicate that by using this approach, multiple marker analysis can be carried out even with limited cytological materials. On the basis of these findings, it is reasonable to assume that other markers, such as HER-2/neu and BRCA1 and 2, can also be evaluated.

Although genomic DNA has commonly been extracted from fresh or archival tissues or recently collected cytological materials, there have been few studies demonstrating the extraction of DNA from archival cytologically stained and additional immunofluorescence-labeled FNA smears. Our findings indicate that with refined technology such as laser microdissection technique, cells with certain morphological features or with a specific phenotype can be selected out to determine specific molecular events underlying such a phenotype in a quantitative fashion. This approach will have tremendous impact not only on improving the cytological diagnosis of cancer but also on understanding the molecular mechanisms of cancer. For example, one can compare the expression of certain oncogenes or tumor suppressor genes between two cells that appear normal morphologically but one of which has abnormal phenotypes (such as increased G-actin) to determine the sequence of the premalignant events.

This study not only shows that one can extract DNA from archival FNA smears but also demonstrates that a semiquantitative analysis of c-myc amplification can be carried out using the extracted DNA material. For c-myc analysis, we used a novel cPCR technique. Although it may seem cumbersome because it requires three PCR reactions with different dilutions for each specimen, it is rather simple and fast in practice. The advantage of this technique is that it does not require an artificially introduced internal control for the reaction. However, a potential pitfall of this method is that even with careful measurements and volume calculations, the resulting band intensity may still vary from the expected value. Although this problem may be partially overcome by using three series dilutions, at best the resulting data are only semiquantitative, and, therefore, c-myc amplification is graded into four scales (−, +, ++,and +++) based on the intensity reading.

Our finding that c-myc is amplified in 16.7% of benign cases and 41.7% of PBDA cases indicates that c-mycmay not be a specific marker for cancer but probably an early risk indicator. Some of these cases have a marked increase of c-myc amplification (for example, case 17, Table 1). Whether such cases will have substantially higher risk to develop breast malignancy is exactly the question that needs to be answered. The fact that c-myc is positive in 4 of 18 G-actin-negative and 4 of 17 p53-negative benign lesions suggests that c-myc amplification occurs earlier than G-actin and p53 and, therefore, is less specific than G-actin and p53 for breast cancer.

In summary, this study demonstrates the approach for multiple biomarker analysis for breast cancer on FNA material. It is possible that a similar approach can be used for other types of cytological specimens such as urine. The information obtained and the methods developed from this study will eventually provide a foundation for large-scale, both retrospective and prospective, studies of biomarkers for breast cancer. The ultimate objective of the biomarker evaluation is to develop intermediate end-point markers for individual risk assessment that are crucial for the success of the chemoprevention study of breast cancer. The results of this study suggest that c-myc amplification may be a biomarker of breast cancer risk; however, additional large,prospective studies are needed to confirm the present observation.