A common mouse mammary tumor virus integration site in chemically induced precancerous mammary hyperplasias

Mammary carcinomas can be induced by chemical and hormonal as well as viral carcinogens. Irrespective of the class of inducer, these tumors develop in discrete stages, of which alveolar hyperplasia is one of the earliest identifiable. Since carcinogenesis by the mammary tumor virus is now thought to involve proviral activation of adjacent cell genes at specific loci, we sought to determine if a similar mechanism also played a role in chemical and hormonal carcinogenesis and if its role was stage specific. Three high-tumor-incidence BALB/c hyperplastic alveolar nodule outgrowths of two different etiologies were found to have exogenous mouse mammary tumor virus proviruses integrated at the same site in the genome. This common site of integration is not within the bounds of the int-1 and int-2 loci into which proviruses detected at these loci are clustered in MMTV-induced mammary tumors. All three HANs are commonly impaired in end-point differentiation. We propose that mouse mammary tumor virus integration at this site is responsible for a specific abnormality in differentiation associated with the preneoplastic phenotype.     

Mammary tumors from GR mice contain more than one population of mouse mammary tumor virus-infected cells

Spontaneous mammary tumors in the GR mouse strain contain several acquired copies of mouse mammary tumor virus (MMTV) DNA that are not present in normal organ DNA and that are detectable by restriction endonuclease analysis and the Southern blotting procedure. Hormone-responsive and -independent cell populations were selected from spontaneous GR mammary tumors by grafting the tumors in castrated male GR mice in the presence and absence of female sex hormones. Analysis of the acquired MMTV DNA copies revealed differences between hormone-responsive and -independent cells derived from the same tumor; however, specific MMTV DNA fragments could not necessarily be correlated with the hormone responsiveness of the tumor. In some cases more than one proviral pattern could be detected for both hormone-responsive and hormone-independent cells. These results suggest that spontaneous GR mammary tumors are made up of more than one population of hormone-responsive and -independent cells which can be distinguished by their MMTV-specific proviral restriction fragments.     

Restriction endonuclease map of endogenous mouse mammary tumor virus loci in GR, DBA, and NFS mice

Mouse mammary tumor virus (MMTV) is integrated in the genome of most mice as an endogenous provirus. Two of these MMTV proviral loci (Mtv-1 and Mtv-2) are associated with virus expression and tumorigenicity. We prepared restriction endonuclease maps of the endogenous MMTV proviruses in two strains, DBA and GR, which contain the Mtv-1 and Mtv-2 loci, plus a third strain, NFS, which has a low mammary tumor incidence. We find that all these mouse strains have certain MMTV loci in common even though their origins are widely divergent. We also find that some integrated MMTV proviruses appear to have undergone alterations or deletions when compared with MMTV exogenous proviral DNA. We have thus made a comprehensive characterization of MMTV loci in these mouse strains which could serve as a basis for the study of their differences in expression.     

Surface structure of mouse mammary tumor virus

The structure of the envelope of mouse mammary tumor virus (MuMTV) was studied by using negative stain, thin sections and freeze-drying-shadowing. The surface of the viral membrane was found to contain a reticular structure composed mainly of hexagons and occasional pentagons. The spacing between the nearest corners was about 74 Å. The viral projections were attached to this reticulum and each projection was surrounded by five or six immediate neighbors. The average interprojection distance was 73.7 ± 10.3 Å. The projections have two distinguishable components, knobs at the distal ends and thin stalks that connect the knobs to the viral member-reticulum. The knobs measure 54.4 ± 10.3 Å in diameter and the average length of the projections was 94.9 ± 8.4 Å. The projection knobs were found to be composed of at least three subunits 15–25 Å in diameter. To explain the pattern of distribution of the surface projections in relation to the reticular structure, we propose that MuMTV contains two types of projections.Freeze-drying and shadowing revealed a highly regular hexagonal array of pits on the viral surface. Pits in fivefold symmetry were also observed. They were spaced 218 ± 43 Å apart and were about 80–100 Å in diameter.Considering the symmetry of the surface projections, reticular structure and pits, we conclude that the shape of the envelope of mouse mammary tumor virus is quasi-icosahedral or like a geodesic dome.     

Site-specific rearrangements in the long terminal repeat of extra mouse mammary tumor proviruses in murine T-cell leukemias

Extra MMTV proviruses in T-cell leukemias of GR and C57/BL10 mice contain alterations in their long terminal repeat (LTR) sequence. The four different leukemias studied contain different deletions, but common hallmarks were observed around the recombination sites. At the 5' end of the deletions we observed a common nonamer sequence, AGACAGGTG, in two leukemias and an almost identical sequence, AGAGCAGGTG, in two other leukemias. At the 3' end of the deletions we invariably found a common stretch of five nucleotides, TTAAA. Three of the four leukemias showed nonconserved crossover sites. The deletions in two leukemias were replaced with neighboring sequences, generating direct repeats. The MMTV LTR characteristic open reading frame and glucocorticoid response element were altered in all four rearranged MMTV LTRs. These results demonstrate site specific rearrangements in the LTR of extra MMTV proviruses in T-cell leukemias and suggest that these rearrangements might permit expression of MMTV in a new target cell.     

Antisense downregulation of a mouse mammary tumor virus activated protooncogene in mouse mammary tumor cells reverses the malignant phenotype

Activation of the protooncogene Wnt-1 by insertion of the mouse mammary tumor virus (MMTV) is known to cause mammary tumors in mice. Wnt-1 expression in mammary glands has been postulated to confer direct local growth stimulation of mammary epithelial cells leading to their acquisition of a preneoplastic state. Wnt-1 expression also induces morphological alterations in cultured normal mammary cells. However, it has not been determined whether or not transformed mammary cells require continuous Wnt-1 expression for their ability to form tumors in vivo. To address this question, we constructed antisense and sense Wnt-1 expression vectors containing a synthetic promoter composed of five high-affinity glucocorticoid response elements (GRE5). This promoter is at least 50-fold more inducible by dexamethasone than the promoter contained in the long terminal repeats of MMTV. The vectors were introduced into a mouse mammary tumor cell line (R/Sa-MT) that expresses high levels of endogenous Wnt-1 mRNA and forms rapidly growing tumors when transplanted into syngeneic hosts. Of the 12 stably transfected cell lines established (9 with antisense and 3 with sense constructs), 2 antisense cell lines (R/Sa-MT/antisense) and 1 sense cell line (R/Sa-MT/sense) were examined for inducibility by dexamethasone of antisense and sense Wnt-1 RNAs, changes in endogenous Wnt-1 RNA expression, and changes in cell morphology. The growth patterns of the cells in vitro and in vivo were also examined. Our results show that (1) the levels of the expression of endogenous Wnt-1 mRNA and protein were reduced significantly (>80%) in those cells (R/Sa-MT/antisense) that expressed antisense Wnt-1 RNA at high levels following exposure to dexamethasone, compared to the R/Sa-MT/sense and R/Sa-MT control cells and (2) transplantation of the R/Sa-MT/antisense cells produced smaller tumors ( approximately 0.2 cm in 16 weeks) compared to the tumors ( approximately 2.0 cm in 8 weeks) that were produced by the R/Sa-MT/sense and R/Sa-MT cells. We therefore suggest that Wnt-1 expression is required not only for the transformation of normal mammary cells into tumor cells, but also for the maintenance of their tumorigenicity.     

Gene order of the mouse mammary tumor virus glycoproteins

Immunochemical and tryptic peptide mapping techniques were used to show that the mouse mammary tumor virus (MMTV) envelope glycoproteins gp52 and gp36 are distinct components derived from a common glycosylated precursor polypeptide of 75,000 daltons (gPr75). Because both gp52 and gp36 are derived from a common precursor polypeptide and therefore have a common initiation site, we have been able to determine their gene order within the viral genome. The gene order was deduced from three different types of experiments. The first approach measured the differential inhibition by NaCl hypertonic shock on initiation of gp52 and gp36 synthesis. The second approach measured the kinetics of appearance of various MMTV proteins following the synchronized reinitiation of polypeptide synthesis resulting from NaCl hypertonic shock. The third approach analyzed polypeptides released from polyribosomes after a series of variable-length short pulses with [3H]amino acids. Our results indicate that the gene order for gPr75 is H2N-gp52-gp36-COOH and 5'-gp52-gp36-3' within the MMTV genome.     

The polypeptide composition of mouse mammary tumor virus

The polypeptide composition of mouse mammary tumor virus isolated from several strains of mice, has been examined by polyacrylamide gel electrophoresis. Gel Patterns, as revealed by staining and ¹⁴C amino acid labelling, showed 10 polypeptide species reproducibly associated with the virus. Three of these contain carbohydrate moieties as determined by incorporation of labelled glucosamine.     

Sequence homology of deoxyribonucleic acid to mouse mammary tumor virus genome in human breast tumors.

Fifty-two human breast tumors were screened for the presence of DNA homology to mouse mammary tumor virus (MMTV) using molecularly cloned MMTV proviral genomic DNA probes and dot-blot hybridization. Seven patients were found to contain an entire provirus (gag, pol, env, and LTR positive at high stringency). Fifty percent (5/10) of patients having a first degree relative with breast carcinoma were found to have DNA homology to the gag-pol portion of the MMTV genome when hybridization and washing was performed at moderate (56C) stringency. Thirty-nine percent (7/18) of patients with any positive family history and 23 percent (8/34) of patients with a negative family history demonstrated homology under these parameters. Of the patients positive for gag-pol at moderate stringency, fewer had taken exogenous hormones than the sample group (20 percent vs 52 percent), more were parous (93 percent vs 68 percent), estrogen receptor positive (69 percent vs 48 percent), and male (13 percent vs 4 percent). At higher stringency (62C) no correlation to family history, hormone use or sex was detected, but positivity was noted among estrogen and progesterone receptor positive patients (67 percent vs 48 percent). Under lower stringency wash conditions, mismatched MMTV-related sequences are identified suggesting the existence of an endogenous gene with partial homology to MMTV. High stringency hybridization may identify a related retrovirus with significant homology to MMTV.     

Characterization and chromosomal distribution of endogenous mouse mammary tumor viruses of European mouse strains STS/A and GR/A

The endogenous mouse mammary tumor virus (MMTV) proviral copies in two genetically dissimilar mouse strains, STS/A, a European mouse strain, and BALB/c, were characterized. STS/A carries the same four MMTV proviral copies as GR.Mtv-2-; these strains share also most of the isoenzyme markers and are therefore highly related. Cellular DNA of GR.Mtv-2- contains a partial MMTV provirus that is not present in STS/A. GR.Mtv-2- is derived from GR; they differ in the locus Mtv-2 that contains one MMTV provirus. Expression of this Mtv-2 endogenous MMTV provirus is directly linked to mammary tumorigenesis in GR. MMTV proviral loci were studied using restriction enzyme analysis and the Southern transfer procedure using liver DNAs from recombinant inbred strains between BALB/c and STS/A. All segregating MMTV-specific EcoRI fragments were identified to MMTV proviral loci and most of these were localized by studying the cosegregation of the Mtv units and known chromosomal markers. Since STS/A, GR.Mtv-2-, and GR are highly related, the five complete endogenous MMTV proviruses of GR were located on the following chromosomes: Mtv-2 on chromosome 18, Mtv-3 on 11, Mtv-19 on 1, Mtv-20 on 4, whereas Mtv-8 has tentatively been located on chromosome 18 by Callahan et al. (R. Callahan, D. Gallahan, and Ch. Kozak (1984), J. Virol. 49, 1005-1008). GR and GR.Mtv-2 furthermore contain two incomplete MMTV proviral elements, one of which is also present in STS/A.     

Identification of human homologues of the mouse mammary tumor virus receptor

Summary.  Human breast tumors contain DNA sequences with high homology to the Mouse Mammary Tumor Virus (MMTV) env gene. env encodes the membrane glycoprotein that binds to the MMTV receptor (Mtvr) in mouse tissues and is required for infection. If humans become infected with MMTV, identification of a human Mtvr might shed light on the mechanism of infection. Here, we identified two human genes, Mtvr1 and Mtvr2, encoding proteins highly related to the mouse receptor. Mtvr-related sequences were also detected in other mammalian species. Thus, zoonotic transmission of MMTV is a real possibility given the existence of highly homologous MMTV receptors. Received July 23, 2001 Accepted October 20, 2001     

DNA of mouse mammary tumor virus‐like virus is present in human tumors influenced by hormones

Mouse mammary tumor virus (MMTV) is a hormonally regulated, oncogenic virus of mice. MMTV‐like virus DNA has previously been detected in human breast cancers, liver disease, and liver cancers. It is hypothesized that local hormonal effects might be of primary importance in determining MMTV‐like virus detection in human tumors. MMTV‐like virus envelope (env) DNA was determined using nested PCR in 89 ovarian, 147 prostate, 50 endometrial, 141 skin, and 51 lung cancers. Viral‐positive sequences were compared with published MMTV‐like viral sequences from human breast cancer, liver cancer and MMTV. Immunohistochemistry for estrogen receptor (ER‐α) and progesterone receptor (PgR) was performed on a subset of tumors. MMTV‐like virus env DNA was detected in ovarian cancers (14/89; 16%), prostate cancers (53/147; 36%), endometrial cancers (5/50; 10%), skin cancers (13/141; 9%) but not in lung cancers (0/51). Phylogenetic analysis of the viral‐positive sequences showed no clustering of the isolates according to tissue type. A significant association was observed between the presence of hormone receptors and detection of MMTV‐like virus in the human cancers screened (P = 0.01). A significant association between MMTV‐like virus and PgR was noted in skin cancers (P = 0.003). Therefore, unlike the mouse model, the detection of MMTV‐like env sequences in human cancers in addition to breast indicates that MMTV‐like viral expression is not breast cancer‐specific and may relate to hormone‐dependent viral expression. J. Med. Virol. 82:1044–1050, 2010. © 2010 Wiley‐Liss, Inc.