Overexpression of HRad17 mRNA in Human Breast Cancer: Correlation with Lymph Node Metastasis

In Schizosaccharomyces pombe, the products of six genes, Rad1, Rad3, Rad9, Rad17, Rad26, and Hus1, have been identified as essential components of checkpoint pathways. Several of these genes have structural homologues in the budding yeasts, and additional conservation across eukaryotes has been demonstrated, e.g., human homologues of S. pombe Rad3, ataxia telangiectasia mutated (1), and ataxia telangiectasia- and Rad3-related; 2, 3); a human homologue of S. pombe Rad9, HRad9 (4); and a human homologue of S. pombe Rad1, HRad1 (5). Recently, a human homologue of the S. pombe Rad17 checkpoint gene named HRad17 was identified (6, 7, 8). S. pombe Rad17 is known to play critical roles in maintaining genomic stability and integrity, as a G₂-M checkpoint protein in the cell cycle, and to prevent the development of cancer and hereditary diseases; however, the function of HRad17 is still unclear.

It was reported independently that HRad17 could be localized on human chromosome 4q or 5q by fluorescence in situ hybridization analysis (6, 7, 9). These regions are implicated in the etiology of a variety of human cancers, including breast cancer, hepatocellular carcinoma, small cell lung cancer, non-small cell lung cancer, duodenal adenocarcinoma, and head and neck squamous cell carcinoma (1, 10, 11, 12, 13). IHC² and RT-PCR analysis indicates elevated levels of expression of HRad17 in human testis and colon cancer (6). But there are no previous reports studying the clinical significance of HRad17 expression in human cancer. Thus, we investigated its expression in human breast cancer cell lines and in clinical breast cancers by means of RT-PCR and IHC using a monoclonal antibody to evaluate its clinical significance. We also determined whether expression of HRad17 mRNA correlates with p53 mutation, overexpression of oncoprotein c-erbB-2 (also known as HER2/neu), or Ki67 labeling index as a marker of proliferation.

Sixty-four matched normal breast and breast cancer tissue samples were obtained from radical mastectomy performed at the Oita Prefectural Hospital (Beppu, Japan), the National Beppu Hospital (Beppu, Japan), the Matsuyama Red Cross Hospital (Beppu, Japan), and the Medical Institute of Bioregulation, Kyushu University (Beppu, Japan), between June 1998 and May 1999. Patients were excluded if they had received preoperative neoadjuvant chemotherapy or radiotherapy. Specimens were frozen in liquid nitrogen immediately after surgical resection and kept at −90°C until the extraction of RNA. Human breast cancer cell lines MCF-7, MRK-nu-1, YMB-1-E, and YMB-1 were used.

Total RNA was extracted by the acid guanidium thiocyanate-phenol-chloroform extraction procedure. The cDNA was synthesized from 2.5 μg of total RNA as described previously (14). HRad17-specific PCR was performed with primers 5′-TCCTTAGAACAGATTTATGGTTTA-3′ and 5′-ATACTTTACATGAAGTTCTAAGGA-3′ to amplify a 523-bp fragment. To prevent amplification from eventual contamination genomic DNA, these primers are located in different exons. PCR was performed for 26 cycles (45 s at 94°C; 45 s at 60°C; and 45 s at 72°C). Five-μl aliquots of the PCR products were size-fractionated on a 1.5% agarose gel and visualized after ethidium bromide staining. The PCR products were subcloned into the pCR II vector (Invitrogen) and verified by sequencing (ABI100 version 3.3; ABI PRISM). RNA quality was verified by running RT-PCR reactions for GAPDH on each RNA sample (14). Primer sequences for GAPDH were 5GTCAACGGATTTGGTCTGTATT-3′ and 5′-AGTCTTCTGGGTGGCAGTGAT-3′(product size, 560 bp), and PCR was performed for 22 cycles (60 s at 94°C; 60 s at 56°C; and 60 s at 72°C). Each series of RT-PCR reactions had a sample without RNA as a negative control, and all specimens were analyzed at least twice.

In situ RNA hybridization was performed using a nonradioactive RNA color kit (RPN3300; Amersham Pharmacia). A 523-bp PCR product was subcloned into TA vectors (pCR II; Invitrogen). The plasmid DNA was isolated and purified using the RPM kit (BIO 101; Biotechnologies, Inc.), and fragments corresponding to the COOH terminus and 3′ region of Hrad17 were excised using XhoI (Takara Co., Kyoto, Japan). Riboprobes were generated with T7 and SP6 RNA polymerase.

We used mouse monoclonal antibodies 31E9 for HRad17 protein (6, 15), DO-7 (Dako Co.) for p53, and NCL-Ki67-MM1 (NovoCastra Laboratories, Ltd.) for Ki67. Rabbit polyclonal antibody (Nichirei Co., Tokyo, Japan) was used for the detection of overexpression of c-erbB-2. Five-μm-thick sections of formalin-fixed and paraffin-embedded tissue samples were deparaffinized by incubation in xylene and rinsed in graded ethanol-water solutions. Antigen retrieval for p53 and Ki67 (16) staining was performed by heating the samples in a microwave oven for 20 min at 98°C, in citric buffer (pH 6.0). Samples were blocked with normal horse serum, and endogenous peroxidases were quenched with 0.3% H₂O₂ in methanol. Sections were incubated with anti-HRad17 (1:200), anti-p53 (1:100), anti-Ki67 (1:200) or anti-c-erbB-2 (1:100) antibodies for 30 min at room temperature, rinsed, and then incubated further with the peroxidase-conjugated secondary antibody for 30 min at room temperature. Detection substrate was 3,3-diaminobenzidine, and all sections were counterstained with Meier’s Hematoxylin before mounting. Negative controls were run simultaneously with preimmune immunoglobulin. The percentage of breast cancer cells showing a positive immunohistochemical reaction in a representative section of each tumor was determined for p53 and Ki67 by counting the number of positively stained cells in 1000 cancer cells. For p53, tumors were considered positive if >10% of all cancer cells were stained. For Ki67, tumors were classified as high or low labeling index by comparing their value with the median of all samples. Overexpression of c-erbB-2 was judged as positive when >10% of cancer cells were stained with the antibody. Two independent pathologists observed all specimens.

The clinical variables considered in this study, including the stage of disease and various pathological factors, are shown in Table 1. The relationship between HRad17 mRNA expression and the clinicopathological factors were analyzed using the χ² test and Student’s t test. A stepwise logistic regression model was used for the multivariate analysis of independent predictors of axillary lymph nodes metastasis using the StatView software package (Abacus Concepts, Inc.). All tests were considered significant when the P was <0.05.

HRad17 mRNA was expressed in all of four human breast cancer cell lines, MCF-7, MRK-nu-1, YMB-1-E, and YMB-1, by RT-PCR (Fig. 1).

A total of 64 matched normal and primary breast tumors were examined for expression of HRad17 mRNA. HRad17 mRNA was expressed in 35 of 64 tumor specimens (54.7%), whereas no amplified product was recognized in the matched normal tissues (Fig. 2). As shown in Table 1, there were no significant differences between the HRad17 mRNA expression status and each clinical factor of age, menopausal state, histological subtype, lymphatic permeation, vascular invasion, or overexpression of c-erbB-2 and estrogen receptor or progesterone receptor status. However, the expression of HRad17 mRNA showed a tendency to correlate with tumor size and expression of p53 protein, although it did not reach a statistically significant difference (P = 0.06 and P = 0.10, respectively). Interestingly, the expression of HRad17 mRNA was significantly associated with lymph node metastasis and Ki67 index (P = 0.001 and P = 0.006, respectively). Twenty-four (68.6%) of 35 cases with HRad17 mRNA expression showed lymph node metastasis, whereas only 8 (27.6%) of 29 cases without HRad17 mRNA expression showed lymph node metastasis. Twenty-three (65.8%) of 35 cases with HRad17 mRNA expression showed a high Ki67 labeling index, whereas only 9 (31.1%) of 29 cases without HRad17 mRNA expression showed a high Ki67 labeling index. The Ki67 labeling index ranged from 0.3% to 66.5% (median, 8.4%; mean, 12.5%). The average Ki67 labeling index was 15.4 ± 14.1% with HRad17 mRNA expression group, whereas it was higher than that of 9.0 ± 13.9% without HRad17 mRNA expression group (P = 0.07).

The independent predictors of lymph node metastasis were determined by stepwise logistic regression analysis. The following variables were subjected to the multivariate analyses: (a) tumor size (≤3.0 cm versus ≥3.1 cm); (b) nodal palpability in the axilla (nonpalpable versus palpable); (c) clinical stage (stage I versus stages II–IV); (d) lymphatic permeation (negative versus positive); (e) vascular invasion (negative versus positive); and (f) expression of HRad17 mRNA (negative versus positive). As shown in Table 2, all of these parameters correlate with lymph node metastasis by univariate analysis. Tumor size (≤2.0 cm versus ≥2.1 cm), overexpression of c-erbB-2, and Ki67 labeling index were excluded because there were no significant differences with respect to lymph node metastasis. Of these variables, the lymphatic permeation and the expression of HRad17 mRNA were the only ones proven to be the independent predictors of the lymph node metastasis (Ps < 0.0001).

In situ hybridization was also performed to confirm that the expression of the gene represented by the PCR fragment is much higher in breast carcinoma tissues than in normal tissues. High expression of the gene represented by the PCR fragment was observed in breast carcinoma tissues, whereas very little expression was found in normal breast (Fig. 3). These results confirmed that the PCR fragment-containing gene is indeed overexpressed by breast carcinoma tissues.

HRad17 protein was detected by IHC. Strong staining was frequently seen in the nucleus of the cancer cells, but faint staining was seen in the cytoplasm of the surrounding normal breast epithelial cells (Fig. 4). There was a difference in staining intensity and localization of HRad17 protein between cancer cells and the surrounding normal epithelial cells. In most cases, strong staining of the nucleus was observed frequently in an advancing margin rather than in the central area of the tumors. A representative case of immunohistochemical staining for HRad17, p53, and Ki67 is shown in Fig. 4.

In human malignancy, identifying new genes that are associated with tumor growth, metastasis, and prognosis is very important in advancing the understanding of cancer biology. The size of the tumor and the number of metastatic lymph nodes are the most powerful prognostic markers for breast cancer patients (17). Consequently, most breast cancer patients are offered postoperative chemoendocrine therapy according to their metastatic nodal status, tumor size, and hormone receptor status (17). Recently, several molecular markers, such as Ki67 (18, 19), p53 (20), c-erbB-2 (21), vascular endothelial growth factor (22), and several others, were reported as predictors of lymph nodes metastasis or prognosis for breast cancer patients. Most of these markers were investigated by IHC or enzyme immunoassay using resected specimens postoperatively. With a view to clinical usefulness, early detection techniques that clearly predict lymph node metastasis before an operation may improve the planning of disease treatment. However, as mentioned above, information on important predictive factors for lymph node metastasis are obtained by studying resected specimens after operation. We thus focused on RT-PCR expression of HRad17 mRNA in the resected specimens as a possible useful molecular predictor of lymph node metastasis. In this study, only resected samples were available. If it proves to be a useful marker in the resected samples, the methodology can be applied to preoperative biopsy samples. We would then be able to get relevant prognostic information before surgery.

We demonstrated that the expression of HRad17 mRNA was correlated significantly with both histological lymph node metastases and high Ki67 index; moreover, it tended to correlate with tumor size and p53 status. This suggested that the tumors with overexpression of HRad17 mRNA might have more potential to recur after operation. The multivariate analysis demonstrated that HRad17 expression status and lymphatic vessel permeation were independent factors for lymph node metastasis. The status of the lymphatic vessel permeation is diagnosed by pathological examination of the resected primary tumors. Thus, it is impossible to obtain the information before operation. On the other hand, the expression status of HRad17 mRNA can be performed by RT-PCR using preoperative needle biopsy samples (23). Thus, for preoperative evaluation of lymph node metastasis, the study of Hrad17 mRNA could become a useful tool. We are now starting a clinical investigation to assess the concordance rate of RT-PCR expression of HRad17 mRNA between preoperative needle biopsy samples and subsequent resected specimens. First, this is needed to evaluate the expression status of HRad17 in various specimens from breast cancer tissues, because the display of breast cancer is often complex, involving intraductal, microinvasive, multifocal neoplastic and preneoplastic lesions in many samples. Then, the use of a marker on a small biopsy sample would seem to be more informative than that in a complete surgical resection specimen. Furthermore, if adequate information about axillary nodal status was obtained before surgery, we could avoid axillary node dissection and its uncomfortable complications. A small breast tumor without overexpression of HRad17 could be treated by local resection of the primary site only and without axillary node dissection or, at most, a sentinel node biopsy only, in place of a complete axillary dissection (24, 25). According to our findings, many early-staged small breast cancer patients may be curable with much less invasive surgery.

Although little was known about the function of HRad17, there are some experimental reports suggesting that HRad17 and p53 may have a role in G₂-M checkpoint control of the cell cycle in human cancer (7, 26). Interestingly, our results showed a tendency for a correlation between the expression of p53 mutant protein and HRad17 mRNA (P = 0.100) and a significant correlation between the Ki67 labeling index and HRad17 mRNA (P < 0.01). Ki67 is a proliferation protein expressed in late G₁, S, G₂, and M phases of the cell cycle; and Ki67 has shown prognostic utility in a number of studies (18, 19). To clarify the molecular interaction between lymph node metastasis and cell cycle checkpoint proteins, additional studies are needed. Biomolecular analysis of the properties of interactions between HRad17 and other G₂-M checkpoint proteins or DNA replication proteins are now required to fully establish the role that this novel gene plays role in cell cycle checkpoint control.

Note Added in Proof

Since the original submission of this paper, overexpression of HRad17 by RT-PCR and IHC has been shown to be correlated with lymph node metastasis in lung cancer (27).

We thank Drs. Yoshiaki Rai, Hideya Tashiro, Chiaki Shirasaka, and Yasuaki Emi for their surgical samples; Drs. Kouichi Tsuji, Yasuji Yoshikawa, Shouichi Era, and Ayako Gamachi for their pathological advice; Drs. Tohru Utsunomiya, Keishi Yamashita, Keishi Yoshinaga, and Takaaki Masuda for their helpful discussions, and Noriko Aoki, Kazue Ogata, and Daisuke Mori for their excellent technical assistance.