Mouse Mammary Tumor Virus-like ENV Gene Sequences in Human Breast Tumors and in a Lymphoma of a Breast Cancer Patient¹

A great deal of effort has been invested in searching for a tumor virus associated with human breast cancer. However, the existence of such a virus, which has been postulated for many years, has not yet been proven. The B-type retrovirus MMTV³ is undoubtedly the major etiological agent of breast cancer in laboratory mice (1, 2) and, as such, has been the retrovirus most extensively searched for in human breast tumor development. Although a large body of evidence for B-type retroviral involvement in human breast tumors has been reported (3, 4, 5, 6, 7, 8, 9, 10), such reports have often lacked verification from other laboratories and thus have led to much controversy in this field. In addition, the presence of HERV-Ks,which are homologous to MMTV (11, 12, 13), has made it difficult to distinguish between endogenous and exogenous MMTV-like sequences.

Recently, new evidence has suggested a MMTV-like retroviral involvement in human breast cancer (14). In these studies, MMTV-like ENV gene sequences that were 95–99% homologous to MMTV sequences present in mice were found in 38.5% of human breast tumors analyzed. In addition, 66% of these MMTV ENV+ human breast tumors were found to express these sequences (15). The absence of such sequences in over 60% of the human breast tumors examined and in nearly all of the normal human breast tissue analyzed suggested the presence of an exogenous MMTV-like retroviral involvement in the breast tumors harboring the MMTV-like ENV gene sequences. To overcome the presence of endogenous sequences, these researchers selected MMTV ENV gene sequences with very low homology to the HERV-Ks or to any other human gene and searched for these sequences in a panel of human breast cancers using PCR (14). Intrigued by this most recent suggestion of MMTV involvement in human breast cancer, we have attempted to detect these MMTV-like sequences in human breast tumors. Using similar PCR procedures and stringent hybridization techniques, we have detected the presence of MMTV-like ENV gene sequences in 37% of the breast tumors that we have analyzed. DNA sequencing has shown these sequences to be 99–100% homologous to MMTV ENV gene sequences present in inbred mice. We have not detected these MMTV-like ENV sequences in normal breast tissue. However, we have detected these sequences by PCR and stringent hybridization in a T-cell lymphoma of a breast cancer patient who was simultaneously diagnosed with both diseases.

Discarded surgical material from human breast tumors and normal breast tissue was snap-frozen at −70°C in liquid nitrogen and kept at−80°C until processed, in accordance with the protocol approved by the Institutional Review Board. Additional normal breast tissue, which was obtained frozen in dry ice from the Duke University Tissue Procurement Facility, was also kept at −80°C until processed. The human breast tumor cell lines T47D, MCF7, and ZR75 were obtained from the American Type Culture Collection (Manassas, VA). DNA was extracted from human breast tumors, normal human breast tissue, and human breast tumor tissue culture cell lines using the quanidine isothiocyanate procedure (16) or the QlAamp DNA-Mini kit(Qiagen). The T-cell lymphoma DNA was obtained from a paraffin-embedded tissue section using the conditions described by Wright and Manos (17). Concentration of DNA, except for that obtained from the paraffin-embedded section, was determined by absorbance readings in a Pharmacia Biotech Ultrospec 3000.

PCR was performed in 50 μl reactions containing 50 mmKCl, 10 mm Tris-HCl (pH 8.3), 1.5 mmMgCl₂, 200 mm each of the four deoxynucleotide triphosphates, and 2.5 units of Taq polymerase(Perkin-Elmer Corp.). For amplification in both first and nested reactions as well as nonnested reactions, the following cycle scheme was used: (a) 3 min at 94°C; (b) 35 cycles of 1.5 min at 94°C, 2 min at 55°C, and 3 min at 72°C; and(c) 10 min at 72°C (14). Approximately 200 ng of template DNA were used in each first-stage PCR and nonnested PCR. For nested PCR, 1 μl of each of the first-stage products was transferred to a second PCR reaction and further amplified. To detect possible contamination of the master mix components, a reaction without template DNA was routinely tested. p53 primers were used as control for master mix and polymerase activity. The PCR product was analyzed by electrophoresis in 2% agarose gels. φX174 RF DNA cut with the restriction enzyme HaeIII and a 1-kb DNA ladder(Life Technologies, Inc.) were used to identify the size of the PCR products.

PCR products were hybridized on Southern blots (18) under stringent hybridization conditions to a 23-bp probe specific for DNA sequences present in exogenous MMTV ENV sequences (19, 20) but not present in endogenous HERV-K DNA (13). The 23-mer probe, which extended from position 1554–1577 in the MMTV ENV gene (19, 20), was 5′ end-labeled with[³²P]ATP using the RTS T4 kinase labeling kit (Life Technologies, Inc.). Stringent hybridization conditions, as described previously (21, 22), included prehybridization for 5 h in 50% formamide, 5× SSC (20× SSC = 3 m NaCl, 0.3 m sodium citrate), 10 mm NAPO₄buffer, 5× Denhardt’s solution, 100 μg/ml single-stranded salmon sperm DNA, and 0.1% SDS. Hybridization to the ³²P-end-labeled probe was performed in the same buffer at 42°C for 40 h, after which the radiolabeled probe was removed, and the blots were washed in 2× SSC and 0.1% SDS for two 15-min washes at room temperature and then washed in 0.1× SSC in 0.1%SDS for two 15-min washes at 55°C (22).

To obtain sufficient DNA for both DNA sequencing and restriction enzyme digestion, the ethidium bromide-stained 250-bp PCR product of each individual breast tumor analyzed in this way was isolated from 15 separate reactions. The PCR products of these 15 reactions, which were each run separately on 2% agarose gels, were pooled after being isolated from the gels using the Qiagen QIAquick Gel Extraction kit. The isolated 250-bp DNA PCR product from three separate human breast tumors was sequenced by Bio 101, Inc. (Vista, CA). In addition,thirteen 250-bp PCR DNA products from 10 different breast tumors and the 3 breast tumors from which PCR product was sequenced were analyzed by restriction enzyme digestion with the restriction enzymes HaeIII (Boehringer Mannheim) and Sau3A(Life Technologies, Inc.). The digestions were carried out as recommended by the respective vendors and analyzed on ethidium bromide-stained 2% agarose gels. Exogenous MMTV ENV gene sequences homologous to the 250-bp PCR DNA product have one HaeIII site, which, after digestion, results in two DNA fragments of 84 and 161 bp, and two Sau3A sites, which,after digestion, result in three DNA fragments of 71,75, and 99 bp.

To select MMTV ENV gene sequences with little or no homology to HERV-K10, the prototype of the group of human endogenous retroviruses that have homology to MMTV, Wang et al.(14) aligned the sequences of the ENV gene of MMTV with the sequences of the HERV-K10 ENV gene using the IBI/Pastell Sequence Analysis Program. A region of 660 bp of low homology (16%) was localized between MMTV ENV sequences 976-1640. Primers were synthesized for detecting both the 660-bp MMTV-like DNA fragment and a smaller 250-bp MMTV-like DNA fragment between positions 1388–1640 (14).

In our study, to increase sensitivity and specificity and thus reduce the possibility of false positives, we have analyzed human breast tumors for the presence of the 250-bp DNA fragment after nested PCR reactions. The 250-bp sequence (between positions 1388–1640) was amplified by first synthesizing a DNA fragment of 671 bp between positions 976 and 1661 (primer 1, 976–990; primer 4, 1661–1647) and nesting this reaction with primers 2 (1388–1405) and 3 (1640–1626;Table 1). PCR was performed on DNA extracted from human breast tumors, normal human breast tissue, and breast tumor tissue culture cell lines. In addition, a nonnested PCR was done using primers 2 and 3 on DNA extracted from the paraffin-embedded section of the T-cell lymphoma of a breast cancer patient who had been simultaneously diagnosed with both diseases. In Fig. 1,A, a photograph of the ethidium bromide-stained gel of the nested PCR products reveals the presence of the 250-bp product in seven different human breast tumors (Lanes 3–9) and in the MCF7 human breast tumor tissue culture cell line (Lane 1), but not in the DNA from normal breast tissue (Lane 2). Similarly, Fig. 2,A reveals the presence of the 250-bp product in five different human breast tumors (Lanes 4–8) and in the T47D human breast tumor cell line (Lane 1), but not in the ZR75 human breast tumor cell line (Lane 2) or in the normal human breast tissue (Lane 3). Fig. 3 A reveals the presence of a nonnested 250-bp DNA PCR product from the human T-cell lymphoma DNA.

Southern blot hybridization (Ref. 18; Figs. 1,B,2,B, and 3,B) was done under stringent hybridization conditions (21, 22) to avoid interference with any endogenous sequences that might interact with the probe. Figs. 1,B, 2,B, and 3,B, in which Southern blots of the gels in Figs. 1,A, 2,A, and 3,A, respectively, were hybridized with the radiolabeled 23-mer oligonucleotide probe (Table 1) that contained MMTV ENV gene sequences from positions 1554–1577, indicate that this MMTV-specific sequence was present in the amplified 250-bp fragment and that the bands in the gel were not artifactual. Our survey of human breast tumors and normal human breast tissue included 73 human breast tumors and 35 normal human breast tissues. Amplification of the 250-bp MMTV-like ENV gene sequence was observed in 27 of 73 human breast tumors. We did not detect the presence of the 250-bp DNA fragment in any of the normal human breast tissue analyzed. The ages of normal controls, a description of the normal breast tissue derived from these controls, the ages of patients whose tumors were either ENV gene positive or negative, the histopathological diagnosis of these breast tumors, and the ethnic background of the patient population are shown in Table 2. The MCF7 and T47D human breast tumor cell lines were positive for the MMTV-like ENV gene sequences, whereas the ZR75 human breast tumor cells were negative. The T-cell lymphoma of one 63-year-old Hispanic female breast cancer patient who had been simultaneously diagnosed with both diseases also contained a 250-bp PCR product that hybridized under stringent hybridization conditions to a probe specific for MMTV ENV gene sequences.

To determine the extent of homology to the MMTV ENV gene throughout the 250-bp stretch, the nested PCR product of three different breast tumors was isolated from ethidium bromide-stained gels and purified using the Qiagen QIAquick Gel Extraction kit, and the DNA was sequenced by Bio 101, Inc. In Fig. 4,the sequences of each of the three breast tumor DNA PCR products are shown (HBT1, HBT2, and HBT3) and aligned with the sequences of the ENV gene of exogenous MMTV found in the GR,C3H, and BR6 strains of mice (19, 20). This domain of the MMTV ENV gene in the GR mouse strain is 100%homologous to that found in the C3H mouse strain and 98% homologous to that found in the BR6 mouse strain (19, 20). The sequences of the three 250-bp DNA PCR products (HBT1–3) from three different human breast tumors were identical. A comparison of the sequences of the 250-bp PCR products of the three different breast tumors with that of MMTV sequences showed the human sequences to be 99% homologous to the MMTV ENV gene of the BR6 mouse strain and 100%homologous to the MMTV ENV gene of the GR and C3H mouse strains. When the 250-bp DNA sequence was compared against all known viral and human genes using the NIH National Center for Biotechnology Information GenBank Blast Program (23), the highest homology recorded was 8%. Restriction enzyme digestion of the 250-bp PCR DNA products from 13 breast tumors (3 of which had been sequenced)with the enzymes HaeIII and Sau3A showed that the products had an identical digestion pattern to each other and to that generated from the homologous exogenous MMTV DNA sequences.

Our results indicate that a 250-bp DNA sequence that is 99%homologous to MMTV ENV gene sequences in the BR6 mouse and 100% homologous to MMTV ENV gene sequences in the GR and C3H mouse strains is present in 37% of the human breast tumors analyzed and is not present in normal human tissue. Both the ethnic background and age range of normal patients and breast tumor patients are quite similar. However, the median age of patients with normal breast tissue and patients with ENV-positive and -negative tumors differs (15 and 32 years, respectively), with over half of our normal controls being under the age of 40 years. However, for five of our breast tumor samples (four ENV-positive samples and one ENV-negative sample), we also had matching normal breast tissue that was ENV negative, and these patients who contributed normal control breast tissue in addition to tumor tissue had a median age of 64 years. To increase sensitivity and specificity and thus reduce the possibility of false positives in our PCR reactions, we analyzed for the presence of this 250-bp DNA fragment in nested reactions, in which we first amplified from human breast tumor DNA a 660-bp sequence that is >95% homologous to MMTV ENVgene sequences. Hybridization to a probe containing sequences specific to the MMTV ENV gene and not to endogenous HERV-K sequences was done under stringent conditions, allowing only exact matching between bases. At no time during these experiments was there any MMTV or MMTV DNA or RNA present in any part of the laboratory.

Our data are thus in close agreement with recently published data in which nonnested PCR detected 660-bp MMTV-like ENV gene sequences in 38.5% of the breast tumors from fresh tissue and detected 250-bp MMTV ENV DNA sequences in 39.5% of the breast tumor DNA extracted from paraffin-embedded sections that were 95–99%homologous to MMTV (14). Our results are also similar to those reported by Imai (24) concerning an 82-year-old Japanese woman whose breast tumor contained sequences that hybridized to a probe from the feral mouse JYG-MMTV genome on a Southern blot. Cloning and sequence analysis were identical to the MMTV ENVsequence of MMTV strain GR.

The possibility of an exogenous retrovirus as the potential etiological agent of some breast cancers is intriguing. An exogenous retrovirus,HTLV-1, has been described as the etiological agent of adult T-cell leukemia (25). HTLV-1 has been reproducibly isolated from adult T-cell leukemia patients, and the leukemic cells invariably contain an HTLV-1 provirus, whereas other cells from these patients do not. Little, if any, HTLV-1 replication or gene expression is detectable in vivo by analyzing primary leukemic cells from humans. HTLV-1 keeps the virion burden low, but it promotes gradually accelerating amplification of cells harboring transcriptionally quiescent proviral DNA over the host organism’s lifetime. We did not detect MMTV-like ENV DNA sequences in 35 of 35 normal breast tissues or in 46 of 73 human breast tumors analyzed, and these results strongly suggest an exogenous viral origin and a possible viral role in a percentage of human breast tumors. Clearly, ENV sequences alone do not make up an entire virus;therefore, we and others are currently examining breast tumors for the presence of additional exogenous MMTV-like sequences to the rest of the viral genome as well as for a viral entity. In the mouse system,exogenous MMTV is passed through the milk of the mother to the offspring (1, 2). One means of HTLV-1 infection has been found to be through nursing (25). However, available epidemiological studies on human breast cancer have not revealed any link between breast feeding and breast cancer development (26).

In addition to causing the development of mammary tumors in inbred strains of mice, MMTV is also involved in the induction of T-cell lymphomas in these mice (27, 28, 29). Thus, our finding that one of the breast cancer patients who had been diagnosed simultaneously with non-Hodgkin’s T-cell lymphoma appears by PCR and stringent hybridization results to also have the 250-bp MMTV ENV DNA sequence in the cells of the T-cell lymphoma is a very interesting result. Although synchronicity of malignancy is rare, studies from our department and other studies have reported the occurrence of breast cancer and lymphoma in the same patient and have suggested that non-Hodgkin’s lymphoma may occur more frequently than expected in women with breast cancer (30, 31, 32, 33, 34, 35). Over 20 years ago, an increased incidence of lymphoma was seen in mice inoculated with human mammary cancer extracts (36). More recently, the frequent development of murine T-cell lymphomas after implantation of human inflammatory breast cancer cells in nude mice was demonstrated (37). We are currently collecting tumor tissue from individuals who have been diagnosed with both breast cancer and lymphoma and from individuals who have been diagnosed with lymphoma alone to analyze for the presence of MMTV-like DNA sequences.

Since submission of the manuscript, we have detected the presence of MMTV-like ENV gene sequences in both a breast tumor and a non-Hodgkin’s B-cell lymphoma of a 62-year-old white female who was simultaneously diagnosed with both diseases. In addition, we have found MMTV-like ENV gene sequences in the non-Hodgkin’s lymphomas of 3 of 19 patients who were diagnosed with lymphoma alone.

We thank Dr. Susan Fineberg and the Department of Pathology at the Montefiore Medical Center for help in procuring breast tumor and normal breast tissue samples. We greatly appreciate the help of Giovanna Cavasotto in the clerical preparation of the manuscript.