New mutant and congenic mouse stocks expressing the murine leukemia virus-associated thymocyte surface antigen GIX.

For several reasons the G(IX) antigen (1) has a prominent place in current work on murine leukemia virus (MuLV): In the prototype G(IX+) mouse strain 129, the G(IX) trait is mendelian, and is expressed selectively (though not exclusively) on thymocytes. Thus, expression of this cell surface component is under the control of cellular genes and is subject to the controls governing the differentiation of T lymphocytes (2). Although the 129 mouse produces no demonstrable leukemia virus such as that found in the AKR strain, it was soon realized that G(IX) antigen must in some way be related to MuLV, because productive infection with MuLV is frequently associated with appearance of G(IX) antigen on cells that are genotypically G(IX-), most notably on MuLV-infected rat cells, or cells that belong to other differentiation pathways (1). The basis of this connection between G(IX) and MuLV has recently become clear from the demonstration that G(IX) is one of MuLV envelope. Therefore, our working hypothesis is that the presence of G(IX) is one of the antigens present on gp69/71 (3,4), the major glycoprotein component of the MuLV envelope. Therefore, our working hypothesis is that the presence of G(IX) antigen always denotes the presence of gp69/71 (though not all variants of gp69/71 need necessarily carry G(IX)). Study of the circumstances under which G(IX) is expressed on the cell surface is thus potentially a powerful approach to understanding how the expression of C-type viral genomes is controlled. Such studies are greatly facilitated by the availability of mutant and congenic strains of inbred mice which differ from the nonmutant or partner strains only with respect to one or another manifestation of the viral genome. It is for this reason that we record here (Table I) some details of two G(IX) mutant and two G(IX) congenic stocks derived in our colonies at Memorial Sloan-Kettering Cancer Center (MSKCC). In addition, to these four strains, Table I includes data for the three relevant partner strains, and for strain AKR, for comparison. These eight strains all differ from one another with respect to one or more MuLV-related traits.

Antigens and antibodies cross-reactive to the murine mammary tumor virus in human breast cyst fluids.

Human breast cyst fluids were shown to contain low concentrations of IgA (15-78 micrograms/ml) and IgG (33-145 micrograms/ml). The IgA:IgG ratios in individual breast cyst fluids ranged from 1:0.6 to 1:4. These levels are considerably higher than their ratio in serum (1:7). IgA from 33% of the 40 fluids examined, and IgG from 10% of the fluids, reacted with the murine mammary tumor virus (MuMTV). The reactivity was detected by an enzyme-linked immunosorbent assay that measures antibody binding to both the envelope glycoprotein and core protein of the virus. In a second series of experiments. IgA from 28% of 40 breast cyst fluids reacted only with MuMTV while IgA from 30% of the fluids was reactive with both MuMTV and the Rauscher murine leukemia virus. Antigen reactive with antiserum to the 28,000-dalton MuMTV core protein (p28), was also identified in a 165,000-g pellet fraction from breast cyst fluids. In individual fluids, the extent of IgA binding to MuMTV was positively correlated (P less than or equal to 0.01) with the binding of anti-p28 antibody to the pellet of the breast cyst fluid. Fractions with the buoyant density of retroviruses (1.16-1.18 g/ml) or their cores (1.21-1.25 g/ml) were isolated from breast cyst fluids. These fractions contained a DNA polymerase capable of utilizing the reverse transcriptase-specific template, dG12-18 x poly rCm. In addition, they reacted with antiserum to MuMTV p 28 but not with antiserum to the 30,000-dalton Rauscher murine leukemia virus core protein.

Leukocyte migration inhibition among breast cancer patients in response to various oncogenic viruses.

The direct leukocyte migration inhibition (LMI) test was done with leukocytes from healthy women and from patients with primary operable breast cancer, benign breast disease, or head and neck cancer. Purified preparations of murine mammary tumor virus (MuMTV), Mason-Pfizer monkey virus (MPMV), and murine leukemia virus (MuLV) were used. For each virus, the lowest 10th percentile of the LMI responses of controls was used to discriminate positive from negative responses. Leukocytes from 46 of 94 (49%) breast cancer patients responded to MuMTV, which was significantly different from each of the other test groups: Positive reactions were observed in only 9 of 67 (13%) healthy persons, 2 of 32 (6%) patients with benign breast tumors, and 2 of 20 (10%) patients with head and neck cancer. Although leukocytes from 29% of the breast cancer patients responded to MPMV, this response did not significantly differ from that observed in healthy women (14%), in patients with benign disease (20%), or in patients with head and neck cancer (20%). The leukocytes from the breast cancer patients were not reactive to MuLV. LMI tests to both MuMTV and extracts of MCF-7 cultured breast cancer cells were done in 36 breast cancer patients and 40 healthy women simultaneously. Of the breast cancer patients, 75% responded to MuMTV and/or MCF-7 antigen as compared to 18% of the controls (P less than 0.005). These data suggest that leukocytes from breast cancer patients are presensitized to MuMTV and antigen(s) present in MCF-7 breast cancer cell line, but not to MPMV or MuLV.

Immunization of mice against murine mammary tumor virus infection and mammary tumor development.

Formalin-inactivated whole murine mammary tumor virus (MuMTV), VuMTV membranes, the acid-soluble component of MuMTV, and purified MuMTV glycoprotein with a molecular weight of 55,000 (gp55; also designated as gp52) were used as vaccines in an attempt to identify the MuMTV antigen(s) that can protect mice from exogenous MuMTV infection and subsequent tumor development. Formalin-inactivated whole MuMTV, MuMTV membranes, and purified MuMTV gp55 were effective immunogens, whereas the acid-soluble component of MuMTV (which consists mainly of MuMTV gp55) failed to protect mice from challenge with live virus. These results suggest that (a) MuMTV gp55 is the major immunizing antigen and (b) its native conformation must be maintained for it to be an effective vaccine.

The morphology of murine oncornaviruses following different methods of preparation for electron microscopy.

The effect of different preparative procedures for electron microscopy on the size and shape of murine oncornaviruses has been studied. With conventional negative staining procedures using neutral sodium phosphotungstate, both murine mammary tumor virus and murine leukemia virus appeared in head-and-tail forms, with a peak head diameter of 122 and 130 nm, respectively. Negative staining with uranyl accetate gave round virions with peak diameters of 148 and 130 nm. Prefixed virus was round with peak diameters of 141 and 130 nm, respectively, in phosphotungstate, and 148 and 117 nm, respectively, in uranyl acetate. With thin sections, the peak diameters were 143 and 123 nm. The preservation of the spherical shape of the virus was obtained by glutaraldehyde fixation dehydration in alcholic solutions of uranyl acetate, and critical point drying. Under these conditions the viruses had peak diameters of 99 and 82 nm, respectively. The size of murine mammary tumor virus has always been found to be larger than murine leukemia virus in all preparations except for negative staining with neutral sodium phosphotungstate. Shadowing of the virion preparations revealed considerable flattening of the particles in all cases except for critical point drying. Negatively stained preparations did not cast any shadow, and thus thethickness of the particles could not be evaluated. Virus can be reversibly converted from spherical to head-and-tail forms by altering osmotic strength. Under most of the conditions used, murine mammary tumor virus gave a bimodal size distribution with significant numbers of particles that were smaller than the major virus size.

Identification of the virions in the in vitro L1210(V) leukemia cell lines by morphological, virological, and immunological techniques.

In vitro L1210 (V) cell lines contained abundant intracytoplasmic A-particles, numerous C-type particles, a small number of B-type particles, and occasional intracisternal A-particles. The intracytoplasmic A-particles were incorporated into both spiked (B-type) and smooth-surfaced (C-type) particles formed at the budding site. Both B-and C-type particles also developed by gradual accululation of neucleooid material. The particles, particularly the C-type, exhibited a wide range of densities. The cells showed strong surface immunofluoresence for both murine mammary tumor virus and Gross murine leukemia virus antigens and variable degrees of cytoplasmic immunoflurescence for the protein antigens (p1 to p6) of Rauscher leukemia virus. The cells, the culture supernatant, and the purified virus each gave positive reactions with murine mammary tumor virus and murine leukemia virus antisera by immunodiffusion. The viral particles failed to infect C57BL, C57BL/6 X DBA/2F1 (hereafter called BD2F1), BALB/c, Af,and RIIIf mice. Howver, the cells were highly tumorigenic in BD1F-1 mice, moderately tumorigenic in BALB/c mice, but not tumorigenic in C57BL, Af, and RIIIf mice.

Template-specific requirements for DNA synthesis by the Mason-Pfizer monkey virus DNA polymerase: unique aspects.

The biochemical properties of DNA polymerase purified from Mason-Pfizer monkey virus were studied, with respect to synthetic and natural template-primer utilization. Thes studies revealed the following new information about the Mason-Pfizer monkey virus enzyme: (a) Mason-Pfizer monkey virus polymerase was found to prefer template: primer molar nucleotide ratios of 2.5-5: 1 for optimal rates of synthesis with poly(C) .(dG)12-18 as template-primer. (b) Poly(A)-directed synthesis was stimulated by the addition of low concentrations of inorganic phosphate to the reaction mixture. (c) Poly(2' -O-methyl-cytidylate), poly(rCm), was the only template studied for which Mn2+ proved the preferred divalent cation. Combinations of divalent cations stimulated rather than inhibited poly(rCm)-directed poly(dG) synthesis by the Mason-Pfizer monkey virus enzyme. (d) Heteropolymeric regions of rabbit globin mRNA and avian myeloblastosis virus 70 S RNA could be copied by the Mason-Pfizer monkey virus polymerase with oligo(dT), oligo(U) or in the case of avian myeloblastosis virus RNA, endogenous primers. In all such studies, Mg2+ was the preferred divalent cation and a distinct preference for the DNA primer in the reverse transcription of natural RNAs was observed. These new findings necessitated comparative studies with the DNA polymerases from Rauscher murine leukemia virus and murine mammary tumor virus, as representative type C and type B retroviruses. Although the Mason-Pfizer monkey virus enzyme was found to share some properties in common with both type C and type B mammalian viral enzymes, certain of the above properties rendered it unique among the polymerases examined.

Evidence for a prokaryotic promoter in the murine mammary tumor virus long terminal repeat.

The long terminal repeat (LTR) of C3H murine mammary tumor virus (MuMTV) is approx. 1.3 kb long. HaeIII digestion of a cloned PstI fragment containing the left-end LTR generated four fragments of sizes 0.56, 0.41, 0.34 and 0.14 kb, one of which (0.41 kb) had a promoter activity in Escherichia coli. This was demonstrated by replacing the bacterial promoter for the neomycin-resistance (NmR) gene in the plasmid pKC56 with the HaeIII fragments. Only the 0.41-kb fragment that contains sequences from the U3 region of the LTR was found to contain a promoter, as shown by the expression of the drug-resistance phenotype in the recombinant plasmid. The strength of this promoter was comparable to or greater than that found with the parental NmR gene promoter. S1 nuclease mapping of the NmR gene transcript indicated that the initiation of this transcript occurred within the 0.41-kb LTR fragment from a site approx. 10 bp upstream from the 3' end. A comparison of the known DNA sequences in the MuMTV LTR with those found in bacterial promoters revealed that a 'Pribnow box', the initiation signal for the prokaryotic promoters, is present in the 0.41-kb LTR fragment upstream from the initiation site. Furthermore, in a recombinant plasmid that contained the complete LTR the same promoter sequences appeared to be involved in the initiation of RNA transcription. The 0.34-kb LTR fragment, which contains sequences derived from the U3 and U5 regions of the LTR, did not possess promoter activity in E. coli. However, it was found to induce deletions of adjacent plasmid DNA sequences. The deletions were specifically initiated from the downstream end of the LTR-fragment insert. The presence of a prokaryotic promoter in the MuMTV LTR, together with the observation that certain LTR sequences can induce deletions, analogous to those caused by transposable elements, in recombinant plasmids suggest that the MuMTV LTR may have evolved from such elements.

Characterization and chromosomal localization of the human homologue of a rat AMP-activated protein kinase-encoding gene: a major regulator of lipid metabolism in mammals.

AMP-activated protein kinase (AMPK) phosphorylates and inactivates acetyl-CoA carboxylase and beta-hydroxy beta-methylglutaryl-coenzyme A (HMG-CoA) reductase which are the major enzymes involved in fatty acid and lipid biosyntheses. The AMPK gene from rat (rAMPK) has recently been cloned [Carling et al., J. Biol. Chem. 269 (1994) 11442-11448]. In order to study the structure and function of the human AMPK gene (hAMPK), we have cloned the gene, and report in this communication its nucleotide (nt) sequence, tissue distribution and chromosomal location. Our results show that the ORF of hAMPK encodes 552 amino acids (aa) (62.250 kDa) and is highly conserved with rAMPK with identities of 97.3 and 90% at the aa and nt levels, respectively. The hAMPK gene bears homology to a yeast protein kinase-encoding gene (snf1) that regulates carbohydrate metabolism, and also with three other genes encoding SNF1-like kinases from different plant species, namely Arabidopsis thaliana, Hordeum vulgare and Secale cereale. As determined by fluorescent in situ hybridization of a human metaphase chromosome spread, hAMPK maps to chromosome 1p31. The size of the hAMPK transcript is 8.5 kb and the transcription start point (tsp) is located approx. 46 bp upstream from the ATG codon. While 10-15% of AMPK is alternatively spliced in most tissues of the rat, our RT-PCR analyses of the hAMPK mRNA did not reveal the presence of any alternatively spliced form of the gene in human tissues. An interesting aspect of AMPK is that its expression, unlike in rat liver, could not be detected in human liver, and thus the purported role of the gene in controlling fatty-acid synthesis in the human liver remains to be determined.

Colonization characteristics of a murine mammary tumor cell line that metastasizes frequently to the heart.

Metastasis by mouse mammary tumor cells is usually confined to lung. This paper describes the metastatic behavior of an established mouse mammary tumor cell line, 4526, that in addition to lung and liver metastasis, shows a high rate of heart metastases. The tumor cells were inoculated into the fourth mammary fat pad of syngeneic mice and their pattern of distant colonization was analysed qualitatively as well as quantitatively. We found that the cell line produced 100, 70 and 40% metastases to the lung, liver and heart, respectively. While the lung metastases appeared primarily as nodular masses, the liver metastases occurred both as nodular and diffuse masses. In addition, we observed that the metastatic load of each of the different lung lobes of individual mice was proportional directly to its relative size, and there seemed to be an inverse relationship between the occurrence of lung and liver metastases in individual mice. As compared to lung and liver metastases, heart metastases were found to be localized internally, usually in the cavity and wall of the ventricle. Furthermore, hearts with metastases revealed destruction of cardiac tissue and blockage of the cavity space. Our results show that 4526 cells are phenotypically stable, since the metastatic behavior of several clonal derivatives of the cell line obtained from lung, liver and heart colonies were found to be identical to that of the parental cell line. Thus this cell line, because of its unparalleled metastatic characteristics, offers a model for investigations into the biology of mammary tumor cell metastasis, especially heart metastasis.

Mixed inocula of mouse mammary tumour cell subpopulations result in changes of organ-specific metastasis.

Tumour and metastatic phenotypes, the pattern of mouse mammary tumour virus (MMTV) integration and expression, and the expression of a metastasis associated gene, nm23, were examined in three mammary tumour cell subpopulations, 66, 168 and 4526. Tumour growth, host survival, metastatic aggressiveness, and the distribution of different cell types in metastasis resulting from mixed cell inocula were also analysed. The results of these studies indicated that the cell lines were distinguishable from each other both phenotypically and genotypically. However, a rearrangement of the mammary tumour specific protooncogene, int-1, caused by MMTV was found to be a unique characteristic of the cell line 4526. Therefore, int-1 was used as a stable marker to examine the genotype of the metastatic colonies that developed in mice bearing tumours of mixed cell inocula. Highly metastatic 4526 cells influenced the metastatic range of poorly metastatic 66 cells. Line 66 cells that normally colonize only to lungs were also found to colonize liver when inoculated together with the liver-metastasizing 4526 cells. This acquired metastatic phenotype of 66 cells was transient. On the contrary, mixed cell inocula of 4526 and non-metastatic 168 cells did not produce any colony of 168 cells. The metastatic aggressiveness of 4526 cells was inhibited by both 66 and 168 cells. Furthermore, the metastatic behaviour of mixed inocula differed depending on the relative abundance of the component populations in the mixtures. These findings suggest that interaction between cells of different metastatic phenotypes may result in changes of their metastatic behaviour.

An enzyme-linked immunoassay for the detection of antibodies to the mouse mammary tumor virus: application to human breast cancer.

An enzyme-linked immunoassay (ELISA) was developed, using the mouse mammary tumor virus (MMTV) fixed to wells of a microtiter plate, for the determination of antibodies to MMTV. The intensity of the final color change was dependent upon virus or viral antibody concentration. MMTV antibody was readily detectable in sera diluted as much as 1 : 2800. Fixed MMTV bound antibodies to an internal viral protein (p 28) as well as to viral envelope components (gp 52, gp 34), demonstrating that the virus was rendered permeable by our procedure. Applying this assay to human sera, significant differences (P less than 0.005) in IgG binding to MMTV were detected between sera of breast cancer patients, benign breast disease patients and healthy individuals. 26% of breast cancer-derived sera contained MMTV binding antibody; 10% of benign sera or 8% of normal sera were also positive. The reactivity of human IgG with MMTV was blocked by prior incubation of the virus with antisera to gp 34 or, to a lesser extent, with gp 52. The results demonstrate that MMTV antibodies can be quantitated by this simple, rapid and inexpensive procedure.

Cloning in a plasmid of an MMTV from a wild Chinese mouse: sequencing of the viral LTR.

Plasmid subcloning by conventional techniques of full length exogenous mouse mammary viruses (MMTV) has not been realized because of the involvement of host-mediated structural changes in the viral gag gene. To circumvent this problem, an alternative subcloning method, excision of phagemid (pBluescript SK) from lambda ZAP II, was successfully used to subclone a novel exogenous MMTV (JYG-MMTV) provirus fragment containing an intact gag gene. Sequence analysis revealed that the LTR of this virus is significantly different from the LTR of C3H-MMTV in the U3 region.

Dietary regulation of mammary tumorigenesis in RIII/Sa mice: investigation of a possible mechanism.

The effects of caloric restriction on the incidence of mammary tumor development, the levels of the expression of mouse mammary tumor virus (MMTV)- and prolactin-RNA, as well as the levels of serum prolactin, were investigated in virgin RIII/Sa mice, a strain known to display a high incidence of spontaneous mammary tumor development. Of the 54 mice fed a low-calorie (LC; 10 kcal/day) diet containing low fat (LF; 5% corn oil) for a period of 72 weeks, only seven mice were found to develop mammary tumors, an incidence of 13%. By contrast, the cumulative tumor incidence in a similar sized group of mice, fed a high-calorie (HC; 16 kcal/day) low fat-containing diet was 73%. Estimation of the relative levels of MMTV-RNA, as determined by Northern and slot blot hybridizations, in the mammary glands of mice fed LCLF and HCLF diets for 8, 10, 16, 28, and 36 weeks revealed that the LCLF diet-fed mice expressed 4-15-fold less RNA than the HCLF diet-fed mice. Interestingly, however, the LCLF diet did not appear to exert any effect on the expression of prolactin RNA even though it reduced the levels of serum prolactin. We suggest that in RII/Sa mice the modulation of MMTV-induced mammary tumors by dietary calorie is linked to the secretion of serum prolactin which, in turn, affects the replication of MMTV required for mammary cell transformation.

Biotransformation and protein binding of N-(4-hydroxyphenyl)retinamide in murine mammary epithelial cells.

The biotransformation of N-(4-hydroxyphenyl)retinamide (HPR), and interactions of parent compound and/or metabolites with the cellular retinoid binding proteins (CRBPs) and cellular retinoic acid binding proteins (CRABPs) were examined in murine mammary tumor virus (MuMTV)-induced murine mammary tumor cells (GR-3A) grown in monolayer cell culture. Soluble fractions (cytosols) obtained from the extracts of GR-3A cells after high speed centrifugation were found to contain proteins of approx. 15,000 daltons which bound retinol and retinoic acid, but did not bind HPR or HPR metabolites. Moreover, HPLC analysis of GR-3A cell extracts demonstrated that [3H]retinol and [3H]retinoic acid were not detected in cells that had been exposed to [3H]HPR for 48 h. These findings, that under in vitro conditions there is no appreciable enzymatic hydrolysis of HPR to retinoic acid or conversion to retinol, suggests that the metabolism and cytological effects of HPR may be distinct from those of retinol or retinoic acid within murine mammary epithelial cells.

Effect of malignant transformation upon the cellular retinoid binding proteins in cultured murine mammary cells.

Cellular retinol (CRBP) and retinoic acid binding proteins (CRABP) were measured in normal (C57Bl) and Murine Mammary Tumor Virus (MuMTV)-induced murine mammary tumor cells (C57BlfRIII) grown in monolayer culture. High speed supernatant fractions (cytosols) from transformed cells contained elevated levels of CRBP and CRABP (0.276 and 1.410 pmol/mg protein, respectively) compared to cytosol from non-transformed murine mammary cells which contained only low levels of CRBP (0.099 pmol/mg protein). Our findings suggest that malignant transformation of mammary epithelial cells, induced by MuMTV infection, results in changes in the expression of cellular retinoid binding proteins and possibly in the sensitivity of these cells to various retinoids.

Polyclonal and monoclonal antibodies monospecific to MMTV LTR orf protein produced in E. coli.

Monoclonal and polyclonal antibodies specific to an open reading frame of the mouse mammary tumor virus long terminal repeat were generated using an open reading frame-beta-galactosidase fusion protein produced in E. coli. Both antibodies reacted with the open reading frame-beta-galactosidase fusion protein but not with beta-galactosidase alone using an immunoblotting technique. It is concluded that these antibodies were specific for the protein encoded by the open reading frame of the mouse mammary tumor virus long terminal repeat.

Overexpression of the p53 gene product in canine mammary tumors.

p53, a tumor suppressor gene, is a target of genetic alternations in many human and animal cancers. Compared to normal tissues, cancer tissues overexpress mutant p53 protein thus allowing their detection by a number of immunochemical procedures. To what extent the expression of mutant p53 correlates with dog mammary tumorigenesis has not been fully studied. In the present study, 20 spontaneously arising canine mammary tumors were examined for overexpression of mutant p53. Two different monoclonal antibodies, BP53-12 and PAb122, which recognize different epitopes of the p53 product, were used. The canine tumors in the present study exhibited five different histological types: i) osteosarcoma (n=7); ii) carcinosarcoma (n=4); iii) solid carcinoma (n=5); iv) complex carcinoma (n=3); and v) tubulopapillar carcinoma (n=1). The positive ratios against BP53-12 and PAb122 antibodies were 50% (10/20) and 60% (12/20) respectively. Among these positive samples, 35% (7/20) reacted to both antibodies. Finally, 15 out of 20 tumors showed positivity against one of the monoclonal antibodies. Mostly, as in human mammary tumor cells, BP53-12 staining was observed in the nuclei of tumor cells. PAb122 staining, however, was confined to cytoplasm of osteosarcoma or carcinosarcoma cells. To confirm the location of the staining, immunoelectron microscopy was done. The results showed that the cytoplasm of cartilage cells in the sarcomas had positive staining. These results indicate that anti-p53 antibodies BP53-12 and PAb122, generated against human p53 are cross reacting with the same molecule in canine cells and that the role of p53 in tumorigenesis is not only confined to tumors in human. Our finding suggests that a combination of p53 monoclonal antibodies should be used to screen, not only canine mammary tumors but also human mammary tumors, to obtain a better tumor prognosis.

Characterization of rare mammary tumours appearing on the neck of RIII/Sa mice infected with mouse mammary tumour virus.

RIII/Sa and C3H mice harbour milk-borne mouse mammary tumour virus (MMTV) and develop mammary tumours at a high incidence. These mammary tumours usually arise ventrally and/or on the sides of the animals. In the present study, some mice of both strains were observed to have tumours in the dorsal neck area. Histological analysis of the tumours indicated their similarity to mammary tumours induced by MMTV oncogenesis. The neck tumours were found by thin-section electron microscopy to contain both type A and type B particles that are hallmarks of MMTV infection. In addition, the neck tumour DNA possessed insertion mutations of Wnt-1 and Fgf-3 proto-oncogenes, the activation of which play important roles in the development of mouse mammary tumours. These neck tumours appear to be mammary tumours that arise in the context of in-situ mammary tissue, similar to rare 'ectopic' human breast cancers that arise in the axillary region and other sites remote from the breast.