- Email: [email protected]
A rare common integration site of proviruses of the mouse mammary tumor virus in P-type mammary tumors of mouse strain GR
VIROLOGY
156,
229-237
(1987)
A Rare Common
Integration Site of Proviruses of the Mouse Mammary in P-Type Mammary Tumors of Mouse Strain GR MARCUS SCHUERMANN’
Division
of Molecular
Biology, Plesmanlaan Received
AND
Tumor Virus
ROB MICHALIDES2
The Netherlands Cancer Institute, Antoni van Leeuwenhoekhuis, 12 1, 1066 CX Amsterdam, The Netherlands June
17, 1986;
accepted
September
28,
1986
The mouse mammary tumor virus (MMTV) can induce mammary tumors in mice by proviral activation of the cellular oncogenes int-1 or int-2. Activation of these genes, however, is observed in only a few hormoneand pregnancydependent mammary tumors of the mouse strain GR. To study the possible involvement of other oncogenes we cloned three MMTV proviral-host fragments (MT 40,42, and 53) from different mammary tumors of GR with a single acquired MMTV provirus. From a genomic library of normal mouse DNA we isolated phages with insert DNAs that covered 2030 kb of the uninterrupted regions. Suitable probes devoid of repetitive DNA sequences were isolated in order to screen other mammary tumors for MMTV proviral integrations in these regions. Only two mammary tumors, MT 40 and 42, showed integration of extra MMTV proviruses within the same region. The integrations occurred only 60 bp apart. The other mammary tumors, however, did not contain MMTV proviral integrations in this region, nor in the MT 53 region. Using mouse-hamster somatic cell hybrid DNA, the MT 40742 integration region was assigned to mouse chromosome 7, and the second region, MT 53, to chromosome 16. The two regions bear no homology to known cellular oncogenes. We did not observe any mRNA being expressed from these cloned segments either in tumors or in normal mammary glands. These findings indicate that plaque(P)-type mammary tumors in mouse strain GR do not originate from MMTV provirus insertions in a particularly favored integration region, but that there may be a variety of if’I&?grStiOf’I Sites in these tUfIIOrS. 0 1997 Academic press, IK.
INTRODUCTION
Mtv-2 locus in mouse strain GR is responsible for the high incidence of pregnancy-dependent mammary tumors in this strain (Van Nie et al., 1977). Most of the virally induced mammary tumors are of clonal origin, as can be deduced from the pattern of integrated extra MMTV proviruses (Cohen et al., 1979; Cohen and Varmus, 1980). This clonality of the tumor has made it possible to use integrated extra proviral DNA as a marker to identify essential cellular genes activated due to proviral insertion. Two common integration regions for MMTV proviruses have been identified in the mouse genome, int-1, which is frequently activated in C3H mammary tumors (Nusse and Varmus, 1982) and int-2 in mammary tumors of mouse strain BR6 (Peters et al., 1983). In plaque (P)-type mammary tumors of GR, however, a few MMTV integrations have been found in the int-1 or int-2 regions. Of a series of 25 GR mammary tumors classified as P-type tumors, only two tumors contained MMTV integrations in the int-1 and one tumor showed an integration in the int-2 locus (Michaiides, 1984). We therefore searched for another common MMTV integration region in GR Ptype mammary tumors, using the same transposon tagging technique to detect common integration areas as has been applied previously (Nusse and Varmus, 1982; Peters et a/., 1983).
Slowly transforming retroviruses can cause a variety of tumors in different species by insertion of a proviral copy next to cellular protooncogenes, which subsequently become activated under the influence of strong viral enhancers or promoters (for reviews, see Varmus, 1982; Bishop, 1983; Varmus, 1984; Nusse et al., 1984a). The mouse mammary tumor virus (MMTV) belongs to this group of retroviruses and causes a high incidence of mammary tumors in females of susceptible inbred strains. MMTV was originally found as an exogenous milk-transmitted particle present in mouse strains with a high mammary tumor incidence, such as Rlll and C3H. In addition, inbred strains carry endogenous proviruses in their germ line (Cohen and Varmus, 1980). Two of these endogenous MMTV proviruses are associated with the mammary tumor induction loci Mtv-1 and Mtv-2 (Michalides et a/., 1981 a). The Mtv-1 locus is present in C3Hf and DBAf mice and controls the development of late and infrequently appearing mammary tumors (Van Nie and Verstraeten, 1975). The
’ Present address: European Molecular Biology Laboratory, Postfach 10 22 09, D-6900 Heidelberg, Federal Republic of Germany. ’ To whom correspondence should be addressed. 229
0042-6822187
$3.00
Copyright 8 1997 by Academic Press, Inc. All rights of reproduction in any form reserved.
230
SCHUERMANN
MATERIALS
AND
METHODS
Mouse strains Mice from strains GR, 020/Mtv-2+, 020 x GR, 020 X STS/A were from the breeding colony of the Netherlands Cancer Institute, Amsterdam. DNA-RNA
preparation
DNA and RNA from liver or tumor material were extracted from the same tissue using 5% (w/v) citric acid during homogenization. The nuclei were pelleted and used for DNA extraction as described by Nusse and Varmus (1982). The RNA-containing supernatant was further processed according to Schibler et a/. (1980). Southern
analysis
Restriction enzyme digestions, gel electrophoresis, DNA transfer, and hybridization experiments were carried out in general according to Maniatis et al. (1982). Restriction enzymes were purchased from either New England Biolabs or Boehringer Mannheim. In some hybridization experiments 9% dextran sulfate was added to the hybridization mixture. Probes labeled with 32P by nick-translation generally reached a specific activity of approximately 2 X 1 O* cpm/pg DNA. Northern
analysis
Either poly(A)-RNA (3-5 rg) or total RNA (15 rg) from samples was used per lane. Poly(A)-containing RNA was isolated by passage of total RNA through columns of oligodeoxythymidilate cellulose. The RNA was then electrophoresed on a 1% agarose slab gel, transferred to a nitrocellulose filter, and hybridized with 32P-labeled probes as described previously by Thomas (1980). Molecular
cloning
of proviral integration
sites
The majority of techniques employed in the isolation of molecular clones followed standard protocols (Maniatis et a/., 1982). Briefly, high molecular weight DNA was digested to completion with EcoRl and separated on a 0.8% agarose gel. DNA in the region containing the respective extra band was cut out, electroeluted, and ligated into the separated EcoRl arms of the bacteriophage vector XgtWES XB. After in vitro packaging, the recombinant phages were plated on Escherichia co/i 2600 and screened using an MMTV-envelopespecific probe. Resulting positive clones were picked and propagated in small culture, and the DNA inserts after EcoRI digestion were checked on gel electrophoresis for the correct size. The appropriate host-viral junction fragments were then subcloned into pBR 322 vector for large-scale DNA preparation and fragments containing mainly cellular DNA were transferred into
AND MICHALIDES
pUC 18 or pUC 8 vectors. These inserts were then mapped with frequently cutting restriction enzymes, and specific subfragments of unique-sequence cellular DNA were detected, using a method described by Steinmetz et al. (1980) and were used as probes. Construction and analysis of a GR mouse liver DNA library in EMBL3 vector In general all procedures followed protocols published by Frischauf et a/. (1983). Fresh livers from young GR mice were used for the preparation of high molecular weight DNA. To achieve the proper insert length, the DNA was partially digested with Sau3A, followed by treatment with calf intestine phosphatase (Boehringer Mannheim). The partially digested DNA was then ligated into the BarnHI site of the EMBL3 arms, packaged using a Boehringer Mannheim packaging kit, and plated with an efficiency of 2 X lo5 recombinants per microgram of DNA. The screening and analysis of positive clones were done in the same way as in the first cloning procedure, now using probes specific for the respective integration regions. For subcloning of these fragments, pUC 18 plasmid DNA was chosen. DNA sequencing DNA sequence analysis was carried out by the method of Maxam and Gilbert (1980) on 5 and 20% polyacrylamide gels in 0.5 X TBE buffer containing 7 M urea. Chromosomal assignment and 53 derived probes
of mammary
tumor 40
To map these integration regions we used DNAs from mouse hamster somatic cell hybrids, kindly provided by Dr. John Hilkens (Department of Tumor Biology, The Netherlands Cancer Institute). Characterization of these cell hybrid clones for chromosomal markers has been described elsewhere (Hilkens eta/., 1986). The hybrids generally retained all hamster chromosomes but segregated mouse chromosomes at random. The selected panel consisted of DNAs of 21 hybrids for localization of the MT40142 region and of 16 hybrids for localization of the MT53 area. The DNAs were digested with EcoRl and hybridized to nick-translated probes. The filters were washed under stringent conditions (0.1 %I SSC, 0.1% SDS, 65°C). RESULTS Characterization of extra MMTV integrations mammary tumors of GR
in
Eleven mammary tumors of mouse strain GR were examined by Southern analysis for the acquisition of new proviral MMTV DNA. We used restriction enzymes
INTEGRATION REGION OF MMTV PROVIRUS IN GR MAMMARY
which cut the provirus internally (Cohen et a/., 1979) such as EcoRl and BarnHI, thereby allowing the estimation of the number of extra MMTV proviral copies by the use of appropriate probes. In this way the MMTV-LTR probe was used to identify the extra proviral sequences in mammary tumors, and the MMTVenvelope probe (MMTV-env) to distinguish between right (env+) and left parts of the MMTV provirus. Most tumors showed many acquired extra copies with weak intensities, indicating multiple integration events at later stages of tumor development (Fig. 1). Only a few tumors harbored one extra proviral copy (Fig. 1, lanes 2, 4, and 10). We first studied whether the int-1 or int-2 regions were also involved in these P-type mammary tumors. We hybridized the tumor DNAs to int-1 and int2 probes using different restriction enzymes to cover an area as large as possible of these integration regions according to published restriction maps (Nusse and Varmus, 1982; Peters eta/., 1983; Dickson eta/., 1984). In this way an area was scanned where nearly 90% of the MMTV integrations near the int-1 gene in C3H mammary tumors had been located (Nusse and Varmus, 1982) and 80% of the integration events at the int-2 locus in BR6 mammary tumors (Peters et al., N
40
41
42
46
49
5051
52
53
54
kb 23
TUMORS
231
N 42
N 53
kb 23-
9.4-
2.31
2
3
4
5
6
FIG. 2. Cloning
of the extra MMTV proviral DNA fragments from DNAs of mammary tumors 42 and 53. The EcoRl inserts of the cloned phages were compared with the extra MMTV proviral EcoRl fragments of the original tumor DNAs by hybridization to an MMTWenv probe. Lanes 1 and 4, GR liver DNA; lanes 3 and 6, phage clones derived from the EcoRI-digested tumor DNA, enriched for the 5-to 6-kb area (see Materials and Methods), next to their original tumor DNAs MT 42, lane 2; MT 53, lane 5.
-
1983). Using this approach, we did not observe any MMTV integrations in the int-1 or int-2 regions in the 11 GR P-type mammary tumors examined. Three tumors, MT 40, 42, and 53, with only a single acquired provirus, as determined by EcoRI, BarnHI, and Bg/ll digestions (Fig. 1, lanes 2, 4, and lo), were selected for further cloning experiments.
9.4. 6.6-
43-
Cloning of the viral-cellular junction fragments of the extra MMTV provirus in GR MT 40, 42, and 53 232.0-
1234567891t)ll FIG. 1. Analysis of MMTV proviruses in normal tissue and mammary tumors. Cellular DNA (20 pg) was digested to completion with EcoRl and, after Southern blotting, analyzed for the presence of MMTVDNA-containing fragments using an MMTV-LTR probe. Lane 1, GR liver DNA; lanes 2 through 11, DNAs of GR mammary tumors. HindIlldigested X DNA was used as a molecular weight marker. Right-half fragments of extra bands most suitable for cloning are marked by an arrowhead.
Viral-cellularjunction fragments with the right EcoRl part of the extra MMTV provirus in mammary tumors 40, 42, and 53 (5.5, 5.6, and 5.8 kb in size; see arrowheads in Fig. 1) were chosen to clone. Cellular DNA fragments enriched for the appropriate size by gel separation after EcoRl digestion were cloned in XgtWES XB arms. Approximately 2 X 1 O5 recombinant phages per microgram of DNA were obtained in each case. Those phages containing viral sequences as inserts were identified by screening with an MMTV-env probe. Electrophoresis of EcoRl digestions of the cloned phage DNAs next to digestions of original tumor DNAs confirmed the identity of the clones (Fig. 2). To obtain a useful hybridization probe, we subcloned the phage insert DNAs of each tumor, called XMT40/1, XMT42/
232
SCHUERMANN
1, and XMT53/1, into plasmid vectors, and isolated fragments free of both viral DNA and repeated cellular DNA sequences present in the mouse genome. The resulting cellular probes were first hybridized to tumor DNAs to confirm the alterations of the corresponding regions. The MT 40 cellular DNA probe did indeed recognize the alteration in tumor 40 (Fig. 3) but also in tumor 42 (see below). The MT 53 cellular DNA probe recognized the alteration in tumor 53 DNA (data not shown). Extra MMTV proviruses of GR mammary tumors 40 and 42 are integrated in the same region The extra MMTV integration regions cloned from the three tumors were characterized by restriction enzyme mapping. The resulting maps showed that integration regions of tumors 40 and 42 contained identical restriction fragments toward the right EcoRl site, but differed only slightly in the left Hinfl fragment, also containing part of the MMTV-LTR (Fig. 4). We did exclude that LTR sequences had undergone rearrangements. The difference between these clones therefore should be located directly in the juxtaposed cellular se-
N 40
42
2
3
N
40
42
2
3
kb
2.3-
Al
Bl
FIG. 3. Cloning of the MMTV integration region of mammary tumors 40 and 42. Identification of the cloned integration region by MMTVLTR probe and a cellular MT 40 DNA probe. (A) Hybridization to MMTV-LTR probe to identify the MMTV proviral copies; (6) the same blot rehybridized with probe 40/l, which was derived from the cellular region flanking the integration of the extra MMTV provirus in MT 40. Arrowheads indicate the position of the extra bands in tumors 40 and 42, which were recognized by both probes.
AND
MICHALIDES
quences. Clone MT53/1, however, showed a completely different pattern. For a more precise localization of the difference in the integration regions in these two tumors, we sequenced the viral-host DNA transitions of MT40 and MT42 (Fig. 5). MT42 contained an additional 60 bp at the precise 3’ end of the MMTV-LTR with respect to MT 40. To demonstrate that this 60 bp difference is not the result of cloning procedures, we used a probe covering the viral-host transition, an Il/lspl-Pvull fragment from hMT42/1 (shown in Fig. 4) and hybridized the MT 40 and 42 tumor DNAs digested with Hinfl to this probe. We still observed the 60-bp difference in these two tumor DNAs (Fig. 6, arrowheads). This clearly demonstrates that the difference was already present in the original tumor DNA. This is particularly relevant because tumors 40 and 42 occurred within the same mouse, but in different mammary glands.
Screening of other Mtv-2-associated mammary tumors for integration of extra MMTV proviruses integration regions MT40/42 and 53
in
We constructed a restriction enzyme map of the uninterrupted loci of MT 40 and 53 to recruit suitable restriction fragments for analysis of other mammary tumors for integration of MMlV extra proviruses in these regions. We also constructed a library of GR mouse liver DNA in EMBL3 vector to extend these two regions. For the MT53 region, we isolated a 13-kb fragment from a positive recombinant phage, and for the MT40/42 region two phage clones containing overlapping inserts covering together approximately 28 kb. Subsequent restriction enzyme analysis provided complete maps of the environments around the respective MMTV provirus integrations (now termed the MT40 and MT53 regions), shown in Figs. 7 and 8. This enabled us to isolate probes from several upstream and downstream regions and to use these in a more extensive screening assay (see Figs. 7 and 8). The P-type mammary tumors used were negative for integrations of MMTV proviruses in the int-1 or int-2 regions. The DNAs of these tumors were assayed for integration events in the MT40 or MT53 region using the various probes indicated in Figs. 7 and 8 and the restriction enzymes EcoRl or BarnHI for digestions of the tumor DNAs. In this way, we covered approximately 30 kb of the MT40 region and 20 kb of the MT53 region. A total of 54 P-type mammary tumors of GR were examined for integration of extra MMTV proviruses in the MT40 region and 17 tumors for integration in the MT53 region. None of these tumors, however, showed any MMn/ integrations into either the MT40 or the MT53 integration region.
INTEGRATION
S Hf MT
REGION
OF
MMTV
PROVIRUS
Hf
Av
HHfAv
IN
GR MAMMARY
Pv
E
40
S Hf
AvM
Hf
4
H HfAv
40/l
X
Pv
probe
Hf
E
42 i
t
S Hf MT
233
Hf
X
I
MT
TUMORS
Av
53
Hf Pv
Av
Hf
\I
I
I
Bg
42/l
Hf
Pv
I
I
I
0
proba
I
S
S E
Bg
I 5311 probe
0.2
0.4
0.0
0.6
1.0
1.2
1.4
1.6
I
I
I
I
I
I
I
1
FIG. 4. Restriction maps of the three MMTV integration regions from tumors 40, 42, and 53. Filled boxes, right MMTV sequences; line, cellular DNA region flanking the respective integration sites. Probes derived from the cellular part Restriction enzyme sites: Av, Avall; Bg, Bglll; E, EcoRI; H, HindIll; Hf, Hinfl; M, Mspl; Pv, Pvull; S, Sack X. Xbal.
Chromosomal location of the MT40 and MT53 integration regions Chromosomal location of the MT40 and MT53 integration regions might provide indications for a possible insertion of extra MMTV proviruses in these tumors near other known cellular oncogenes. We therefore hybridized DNAs from a panel of well-characterized mouse-hamster somatic cell hybrids to probes of the MT40 and MT53 integration regions. The results with probe 40/l showed the highest concordancy with markers on chromosome 7 for 20 out of 21 independent hybrid clones. In the case of probe 53/l, taking DNAs from 16 of these clones, a 100% correlation was found for chromosome 16. Since chromosome 7 contains the cellular oncogenes Ha-ras-1 , c-fes, and int-2 (Kozak et a/., 1983; Peters et a/., 1984a), we hybridized the recombinant phage insert DNAs to 3zP-labeled probes of these three
MT 42
MT 40
I
long terminal are indicated
kb
repeat below.
oncogenes. However, we did not observe any positive hybridization, nor did the previously published restriction maps of Ha-ras-1 , c-fes, or int-2 show any similarity to those of MT40. These data indicate that the cloned 40/l region is different from these genes and is separated from them by at least 20-25 kb. Mammary tumors 40 and 42 also did not show any expression of int-2 on Northern blots (data not shown). It is therefore highly unlikely that MMTV integrations in these two tumors resulted in int-2 gene activation. Search for functional and MT53 regions
genes in the MT40
Proviral insertions around int-1 and int-2 genes at the 5’site commonly occur in opposite transcriptional orientation; insertions on the 3’ site instead occur in the same transcriptional orientation as that of the int genes (Nusse et al., 1984b; Dickson et a/., 1984). The
IGTAAAGTATT CCCCTGATAT GTATGACCCC TTGGGATCAA TTCCCAGTAG TTTAAGAAAG,
. ..TGACCCTCAC
GTCGGCCGAC TGCGGCiCTA ATGAAGCTAG GGAGAATTGA
GATGACGAAG ATGGAGATGG GGAGGATGGA GATGAGGAAG ATATTGGTGG GAGGATGGTC ATAGAAACCA GAACGATAGG... FIG. 5. DNA sequence of the viral-host DNA transition of MT40 and MT42. The sequenced using XMT40/1 and XMT42/1. The extra 60-bp cellular DNA that is present The viral sequence ends just at the 3’ end of the viral LTR (Kennedy et a/., 1982).
viral-host in MT42
DNA transitions of MT40 and MT42 were in comparison with MT40 is indicated on top.
234
SCHUERMANN N
40
42
AND
MICHALIDES
regions as a result of MMTV proviral insertion the limited areas covered by the probes used.
within
DISCUSSION
bp
1630
510 390 344 296 220
FIG. 6. Comparison of the MMTV integration sites in tumors 40 and 42. DNA of tumors 40 and 42 (lanes 2 and 3) and GR liver DNA (lane 1) were digested with Hinfl, blotted, and hybridized with the 42/l probe (see Fig. 7). Hinfl-digested pBR322 DNA was used as a size marker. The uninterrupted cellular Hinfl fragment, about 1.7 kb in size, is marked by an asterisk. Arrowheads show the different positions of the interrupted fragment due to insertion of MMTV proviral DNA (about 660 and 720 bp in size).
int-1 and int-2 genes are therefore, in general, located upstream of the integrated MMTV provirus with respect of its transcription orientation. In order to detect expression of genes that might become activated as a result of MMTV proviral integrations in the MT40 or MT53 regions, we used different probes of these regions covering approximately 8 kb upstream as well as 1 kb downstream of the integration site, and hybridized them to the respective tumor RNAs on Northern blots. We observed no hybridization to RNA from tumors 40 and 42 or from tumor 53. There was also no detectable hybridization to control RNA, such as liver and normal mammary gland RNA. When rehybridized to a P-actin probe, all RNA samples showed a consistent signal in the expected range. Thus, although the results are still preliminary, it appears that no transcriptional unit becomes activated in the MT40 and MT53 integration
We have cloned and characterized two integration regions of extra MMTV proviruses from three different GR mammary tumors. These tumors contained only one extra MMTV provirus and appeared clonal with respect to this acquired MMTV copy. In two of the mammary tumors, MT 40 and 42, these cloned regions represent a common area in which the MMTV proviruses integrated only 60 bp apart. In none of 54 P-type mammary tumors of GR did we observe MMTV proviral integrations in this region. Therefore this region is only rarely implicated in the early development of at least some of the GR mammary tumors. Comparison of the sequences around the integrations of the extra MMTV proviruses in tumors 40 and 42 did not reveal any sequences in common and therefore does not imply a homologous recombination as an explanation for this region specific integration. Although mammary tumors 40 and 42 were present in the same mouse, we assume that they represent two different tumors, since (i) they were found in two different mammary glands, respectively the third and fifth right mammary gland, and were equally big at the time of sacrificing the mouse, and (ii) it is highly unlikely that the rearrangement involved only a cellular DNA fragment leaving the viral LTR exactly intact (Fig. 5). Even if this highly improbable event did occur, it should have taken place in a tumor cell of tumor 42 that would have given rise to a clonal outgrowth of tumor 40 in another mammary gland. This we believe to be highly unlikely. Mammary tumors 40 and 42 are therefore two different tumors with an MMTV proviral copy integrated into the same region, being clonally amplified at subsequent cell divisions. The nearby integration of extra MMTV proviruses in tumors 40 and 42 may represent fortuitous integration events in a particular favored genomic area. A similar rare common integration region has also been described by Garcia et al. in BALB/cfC3H in mammary and kidney adenocarcinomas (Garcia et al., 1986). These regions may indeed represent preferred targets for retroviral integration and bear common cellular recognition sequences as proposed by Cohen and Murphey-Corb for baboon endogenous virus (Cohen and Murphey-Corb, 1983). However, direct sequence comparisons of virus-host cell junction fragments in other systems and tumors did not yield any sequence homologies or repeated DNA sequences with respect to host DNA (Van Ooyen and Nusse, 1984; Selten et a/., 1985), which also is not evident from our sequence data. Nevertheless, the implication of these observed
INTEGRATION
probes
REGION
40/3 ,..’-.._, B ,/ ,,/’
screened alr*a*
OF
‘...
..,,
..,
MMTV
PROVIRUS
42/l I 4wm.... B ....” .,..’
n
E
IN GR
“‘..... .__.,,(,,
MAMMARY
(,,,,,.,..........
BE , , ,(,........
“....,,,
TUMORS
235
..“”
. . 40/2 .. .‘. “.,,E 1
.“’
E
)rMT40/42clones x40/1
-
-
HN
phage
ONA
FIG. 7. Topography of the MT40 region. The integrated MMTV provirus here corresponds to position 0. Open boxes, long terminal repeats of MMTV; filled boxes, cellular probes used for the screening of the respective areas indicated. The genomic areas of the clones obtained in the first and second cloning steps are indicated. The restriction enzyme sites for BarnHI (S), EcoRl (E), Kpnl (K), and Xbal (X) are given.
rare common integration mains to be studied.
regions in tumorigenesis
re-
53 region, except for tumors 40 and 42. It is striking that of the 11 tumors examined only tumors 40 and 42, with tumor 53, show a more intense extra MMlV proviral band, indicating a clonal outgrowth of the tumor. These clonal outgrowths of P-type tumors occur less frequently; most of the P-type mammary tumors represent semiclonal outgrowths with respect to their
Do P-type mammary tumors of GR share a common MMTV integration region? In none of the 54 P-type mammary tumors of GR did we observe any MMTV integrations in the MT40/42 or
E
K
B
Eli.
I
X
X
X
i
probes screened sreas
E ........... , _._...... . B
a E
XMT53
E
-
-
FIG. 8. Topography
of the MT53
region.
For legends,
see Fig. 7.
phege
DNA
236
SCHUERMANN
extra MMTV proviruses (Maclnnes et al., 1981; Michalides et al., 1981 b). Proviral insertion often results in an elevated expression of the juxtaposed genes mediated by enhancers of the integrated provirus. A search for constitutive mRNA expression from functional neighboring genes in MT 40, 42, and 53, however, has also failed so far, although there still remains the possibility that the genes of interest were not covered by the probes used, were further away from the integration site, or were expressed in amounts too low to be detected by the respective probes. Our findings are therefore too inconclusive to exclude that genes of the integration regions MT 40, 42, and 53 do play a direct role in the development of tumors 40, 42, and 53. They do, however, exclude these regions as frequently observed integration sites in P-type mammary tumors of GR. The present data do not allow us to formulate a general mechanism of mammary tumor development in mouse strain GR. Nevertheless, it is also worthwhile to consider the low incidence of int-1 and int-2 positive mammary tumors found in GR mice. Together with our findings, these results strongly suggest that MMTV proviral insertion in a particularly favored integration region does not act in general as an initial event in GR mammary tumorigenesis, but that there may be a great diversity of insertion sites in these tumors. The early expression of endogenous MMTV associated with the MTV-2 locus increases the likelihood that GR mammary gland cells become infected with MMTV at very early stages. Histologically, early tumor stages in GR mice show a much greater variety in pattern than other mouse strains (Van Nie, 1980; Percy et a/., 1980), indicating that the affected cells retain a capacity to undergo differentiation into more than one cell type (Hynes et a/., 1984). It is quite conceivable that different oncogenes contribute to the transformation of mammary gland cells at distinct stages of differentiation. This could also explain the low incidence of int-1 or int-2 positive tumors, when int-1 or int-2 would be implicated in the transformation of mammary gland cells in later stages of differentiation. A second point concerns the clonality of GR mammary tumors. It has been demonstrated in serially transplanted GR mammary tumors that the pattern of acquired MMlV proviruses is altered when hormonesensitive tumors become hormone independent (Michalides et a/., 1982; Maclnnes et al., 1981). From these studies it was concluded that the original tumor already harbored different cell populations. GR tumors at later stages may have undergone more selection and as a consequence may have become more homogeneous. This assumption is in contrast to the observed clonal selection of proviral insertions around int2 found in pregnancy-dependent BR6 mammary tumors
AND
MICHALIDES
(Peters eta/,, 198413). The semiclonality of extra MM-IV proviruses in most of the P-type mammary tumors suggests that de nova MMTV integration events may not be directly involved in the earliest stages of mammary gland transformation which lead to P-type mammary tumors and that they merely mark the more progressively growing subclones of these tumors. MMTV insertion, also in the known int-1 and int-2 regions, would in that case activate cellular genes that may be implicated in one of the multiple steps of progression of GR mammary tumors (Michalides, 1984). Our findings in this study would be in line with this view, since a common integration region for extra MMTV proviruses was only found in GR mammary tumors 40 and 42, which show a clonal outgrowth of the tumor, but not in 54 other P-type mammary tumors. Other mechanisms that may account for the induction of mammary tumors in mouse strain GR which are not solely dependent on insertional mutagenesis by MMTV provirus, but which are nevertheless likely to contain additional clonal extra MMTV proviruses, are (i) activation of an oncogene juxtaposed to the Mtv-2associated provirus itself, which may occur in a stageand tissue-specific manner, and (ii) growth stimulation of normal mammary cells via paracrine effects exerted by the large number of multiple preneoplastic lesions in early GR mammary glands due to massive MMTV replication (Van Nie, 1980; Michalides, 1984). Some cells of these lesions may become more easily transformed and stimulated to subsequent outgrowth. Which among these possibilities provides the mechanism for tumorigenesis in GR mammary tumors remains to be determined.
ACKNOWLEDGMENTS We thank Roe1 Nusse for helpful discussions and for carefully reading the manuscript. Mouse-hamster somatic cell hybrid DNAs were kindly provided by John Hilkens; int-1 probe was obtained from Roel Nusse; int-2 probe was a gift from Clive Dickson of the Imperial Cancer Research Fund Laboratories. M. Schuermann was supported by the Deutsche Forschungsgemeinschaft.
REFERENCES BISHOP, J. M. (1983). Cellular oncogenes and retroviruses. Annu. Rev. Biochem. 52, 301-354. COHEN, J. C., SHANK, P. R., MORRIS, V. L.. CARDIFF, R., and VARMUS, H. E. (1979). Integration of the DNA of mouse mammary tumor virus in virus-infected normal and neoplastic tissue of the mouse. Cell 16, 333-345. COHEN, J. C., and VARMUS, H. E. (1980). Proviruses of mouse mammary tumor virus in normal and neoplastic tissues from GR and C3Hf mouse strains. J. Virol. 35, 298-305. COHEN, J. C., and MURPHEY-CORB, M. (1983). Targeted integration of baboon endogenous virus in the BEVI locus on human chromosome 6. Nature (London) 301, 129-l 32. DICKSON, C., SMITH. R., BROOKES. S., and PETERS, G. (1984). Tumorigenesis by mouse mammary tumor virus: Proviral activation of a
INTEGRATION
REGION
OF
MMTV
cellular gene in the common integration region int-2. Ce//37,529536. FRISCHAUF, A. M., LEHRACH, H., POUSTRA, A., and MURRAY, N. (1983). Lambda replacement vectors carrying polylinker sequences. J. Mol. Biol. 170, 827-842. GARCIA, M., WELLINGER, R., VESSAZ, A., and DIGGELMANN, H. (1986). A new site of integration for mouse mammary tumor virus proviral DNA common to BALBlcf(C3H) mammary and kidney adenocarcinemas. EMBO J. 5, 127-l 34. HILKENS, J., CUYPERS, H. T., SELTEN. G., KROEZEN, V., HILGERS, J., and BERNS, A. (1986). Genetic mapping of the Pim-1 oncogene to mouse chromosome 17. Somatic Cell Mol. Genet. 12, 8 l-88. HYNES, N. E., GRONER, B., and MICHALIDES, R. (1984). Mouse mammary tumor virus: Transcriptional control and involvement in tumorigenesis. Adv. Cancer Res. 41, 155-l 79. KENNEDY, N., KNEDLITSCHEK, G., GRONER, B., HYNES, N. E., HERRLICH, P., MICHALIDES, R., and VAN OOYEN, A. J. J. (1982). Long terminal repeats of endogenous mouse mammary tumor virus contain a long open reading frame which extends into adjacent sequences. Nature (London) 295, 622-624. KOZAK, C. A., SEARS, J. F., and HOGGAN, M. D. (1983). Genetic mapping of the mouse oncogenes c-Ha-ras-1 and c-fes to chromosome 7. J. Viral. 47, 217-220. MACINNES, I. I., CHAN. E. C. M. L., PERCY, D. H.. and MORRIS, V. L. (1981). Mammary tumors from GR mice contain more than one population of mouse mammary tumorvirus infected cells. Virology 113,119-129. MANIATIS, T., FRITSCH, E. F., and SAMBROOK, J. (1982). “Molecular Cloning: A Laboratory Manual.” Cold Springs Habor Laboratory, Cold Springs Harbor, NY. MAXAM, A. M., and GILBERT, W. (1980). Sequencing of end-labeled DNA with base-specific chemical cleavages. /n “Methods in Enzymology” (L. Grossman and K. Moldave, Eds.), Vol. 65, pp. 499560. Academic Press, New York. MICHALIDES, R., VAN NIE, R., NUSSE, R., HYNES, N. E., and GRONER, B. (1981 a). Mammary tumor induction loci in GR and DBAf mice contain one provirus of the mouse mammary tumor virus. Cell 23, 165-173. MICHALIDES, R., WAGENAAR, E., GRONER, B., and HYNES, N. E. (1981 b). Mammary tumor virus proviral DNA in normal murine tissue and non-virally induced mammary tumors. J. Viral. 39, 367-376. MICHALIDES, R., WAGENAAR, E., and SLUYSER, M. (1982). Mammary tumor virus DNA as a marker for genotypic variance within hormone-responsive GR mouse mammary tumors. Cancer Res. 42, 1154-1158. MICHALIDES, R. (1984). Amplification of mouse mammary tumor virus DNA in mammary tumors of GRIA mice, a model for hormone sensitive mammary tumors. In: “Hormones and Cell Regulation” (J. E. Dumont and 1. Nunez Eds.), Vol. 8, pp. 255-268. Elsevier, Amsterdam. NUSSE. R., and VARMUS, H. E. (1982). Many tumors induced by the mouse mammary tumor virus contain a provirus integrated in the same region of the host genome. Ce// 31, 99-109.
PROVIRUS
IN GR
MAMMARY
TUMORS
237
NUSSE, R., VAN OOYEN, A., WESTAWAY, D., FUNG, Y. K., and VARMUS, H. E. (1984a). Oncogene activation by proviral insertion. In: “Cancer Cells 2/0ncogenes and Viral Genes” (G. F. van de Woude, A. J. Levine, W. C. Topp, and J. D. Watson, Eds.), pp. 205-209. Cold Springs Harbor Laboratory, Cold Springs Harbor, NY. NUSSE, R., VAN OOYEN, A., Cox, D., FUNG, Y. K. T., and VARMUS, H. (1984b). Mode of proviral activation of a putative mammary oncogene (int-1) on mouse chromosome 15. Nature (London) 307, 131-136. PERCY, D. H., MORRIS, V. L., and MACINNES, J. I, (1980). Comparison of biologic behavior and histology of transplanted mammary tumors in GR mice. J. Nat/. Cancer Inst. 69, 933-938. PETERS, genesis region 377. PETERS, tumor mouse
G., BROOKES. S., SMITH, R., and DICKSON, C. (1983). Tumoriby mouse mammary tumor virus: Evidence for a common for provirus integration in mammary tumors. Cell 33, 369G. C., KOZAK, C., and DICKSON, C. (1984a). Mouse mammary virus integration regions int-1 and int-2 map on different chromosomes. Mol. Cell. Biol. 4, 375-378.
PETERS, G., LEE, A. E., and DICKSON, C. (1984b). Activation of a cellular gene by mouse mammary tumor virus may occur early in mammary tumor development. Nature (London) 309, 273-275. SCHIBLER, U., TOSI, M., PITTET, A.-C., FABIANI, L.. and WELLAUER, P. K. (1980). Tissue-specific expression of mouse a-amylase genes. J. Mol. Biol. 142, 93-l 16. SELTEN, G., CUYPERS, H. T.. and BERNS, A. (1985). of the putative oncogene Pim-1 in MuLV induced EMBO J. 4, 1793-1798.
Proviral activation T-cell lymphomas.
STEINMETZ, M., HOECHTL, J., SCHNELL, H., GEBHARD, W., and ZACHAU, H. G. (1980). Cloning of V region fragments from mouse liver DNA and localization of repetitive DNA sequences in the vicinity of immunoglobulin gene segments. Nucleic Acids Res. 8, 1721-l 729. THOMAS, P. S. (1980). Hybridization of denatured fragments transferred to nitrocellulose. Proc. 77,5201-5205.
RNA and small DNA Nad. Acad. Sci. USA
VAN NIE, R., and VERSTRAETEN, A. A. (1975). Studies mission of mammary tumor virus by CBHf mice. 922-932.
of genetic transhf. J. Cancer 16,
VAN NIE, R., VERSTRAETEN, A. A., and DE MOES. J. D. (1977). Genetic transmission of mammary tumor virus by GR mice. Inf. J. Cancer 19,383-390. VAN NIE, R. (1980). Mammary tumorigenesis in the GR mouse strain. In “Mammary Tumors in the Mouse” (1. Hilgers and M. Sluyser, Eds.). pp. 202-267. Elsevier, Amsterdam. VAN OOYEN, A., and NUSSE, R. (1984). Structure and nucleotide sequence of the putative mammary oncogene int-1 ; Proviral insertions leave the protein encoding domain intact. Cell 39, 233-240. VARMUS, H. E. (1982). Form and function of retroviral proviruses. Science 216, 812-820. VARMUS, H. E. (1984). Annu. Rev. Gener.
The molecular 18, 553-612.
genetics
of cellular
oncogenes.
156,
229-237
(1987)
A Rare Common
Integration Site of Proviruses of the Mouse Mammary in P-Type Mammary Tumors of Mouse Strain GR MARCUS SCHUERMANN’
Division
of Molecular
Biology, Plesmanlaan Received
AND
Tumor Virus
ROB MICHALIDES2
The Netherlands Cancer Institute, Antoni van Leeuwenhoekhuis, 12 1, 1066 CX Amsterdam, The Netherlands June
17, 1986;
accepted
September
28,
1986
The mouse mammary tumor virus (MMTV) can induce mammary tumors in mice by proviral activation of the cellular oncogenes int-1 or int-2. Activation of these genes, however, is observed in only a few hormoneand pregnancydependent mammary tumors of the mouse strain GR. To study the possible involvement of other oncogenes we cloned three MMTV proviral-host fragments (MT 40,42, and 53) from different mammary tumors of GR with a single acquired MMTV provirus. From a genomic library of normal mouse DNA we isolated phages with insert DNAs that covered 2030 kb of the uninterrupted regions. Suitable probes devoid of repetitive DNA sequences were isolated in order to screen other mammary tumors for MMTV proviral integrations in these regions. Only two mammary tumors, MT 40 and 42, showed integration of extra MMTV proviruses within the same region. The integrations occurred only 60 bp apart. The other mammary tumors, however, did not contain MMTV proviral integrations in this region, nor in the MT 53 region. Using mouse-hamster somatic cell hybrid DNA, the MT 40742 integration region was assigned to mouse chromosome 7, and the second region, MT 53, to chromosome 16. The two regions bear no homology to known cellular oncogenes. We did not observe any mRNA being expressed from these cloned segments either in tumors or in normal mammary glands. These findings indicate that plaque(P)-type mammary tumors in mouse strain GR do not originate from MMTV provirus insertions in a particularly favored integration region, but that there may be a variety of if’I&?grStiOf’I Sites in these tUfIIOrS. 0 1997 Academic press, IK.
INTRODUCTION
Mtv-2 locus in mouse strain GR is responsible for the high incidence of pregnancy-dependent mammary tumors in this strain (Van Nie et al., 1977). Most of the virally induced mammary tumors are of clonal origin, as can be deduced from the pattern of integrated extra MMTV proviruses (Cohen et al., 1979; Cohen and Varmus, 1980). This clonality of the tumor has made it possible to use integrated extra proviral DNA as a marker to identify essential cellular genes activated due to proviral insertion. Two common integration regions for MMTV proviruses have been identified in the mouse genome, int-1, which is frequently activated in C3H mammary tumors (Nusse and Varmus, 1982) and int-2 in mammary tumors of mouse strain BR6 (Peters et al., 1983). In plaque (P)-type mammary tumors of GR, however, a few MMTV integrations have been found in the int-1 or int-2 regions. Of a series of 25 GR mammary tumors classified as P-type tumors, only two tumors contained MMTV integrations in the int-1 and one tumor showed an integration in the int-2 locus (Michaiides, 1984). We therefore searched for another common MMTV integration region in GR Ptype mammary tumors, using the same transposon tagging technique to detect common integration areas as has been applied previously (Nusse and Varmus, 1982; Peters et a/., 1983).
Slowly transforming retroviruses can cause a variety of tumors in different species by insertion of a proviral copy next to cellular protooncogenes, which subsequently become activated under the influence of strong viral enhancers or promoters (for reviews, see Varmus, 1982; Bishop, 1983; Varmus, 1984; Nusse et al., 1984a). The mouse mammary tumor virus (MMTV) belongs to this group of retroviruses and causes a high incidence of mammary tumors in females of susceptible inbred strains. MMTV was originally found as an exogenous milk-transmitted particle present in mouse strains with a high mammary tumor incidence, such as Rlll and C3H. In addition, inbred strains carry endogenous proviruses in their germ line (Cohen and Varmus, 1980). Two of these endogenous MMTV proviruses are associated with the mammary tumor induction loci Mtv-1 and Mtv-2 (Michalides et a/., 1981 a). The Mtv-1 locus is present in C3Hf and DBAf mice and controls the development of late and infrequently appearing mammary tumors (Van Nie and Verstraeten, 1975). The
’ Present address: European Molecular Biology Laboratory, Postfach 10 22 09, D-6900 Heidelberg, Federal Republic of Germany. ’ To whom correspondence should be addressed. 229
0042-6822187
$3.00
Copyright 8 1997 by Academic Press, Inc. All rights of reproduction in any form reserved.
230
SCHUERMANN
MATERIALS
AND
METHODS
Mouse strains Mice from strains GR, 020/Mtv-2+, 020 x GR, 020 X STS/A were from the breeding colony of the Netherlands Cancer Institute, Amsterdam. DNA-RNA
preparation
DNA and RNA from liver or tumor material were extracted from the same tissue using 5% (w/v) citric acid during homogenization. The nuclei were pelleted and used for DNA extraction as described by Nusse and Varmus (1982). The RNA-containing supernatant was further processed according to Schibler et a/. (1980). Southern
analysis
Restriction enzyme digestions, gel electrophoresis, DNA transfer, and hybridization experiments were carried out in general according to Maniatis et al. (1982). Restriction enzymes were purchased from either New England Biolabs or Boehringer Mannheim. In some hybridization experiments 9% dextran sulfate was added to the hybridization mixture. Probes labeled with 32P by nick-translation generally reached a specific activity of approximately 2 X 1 O* cpm/pg DNA. Northern
analysis
Either poly(A)-RNA (3-5 rg) or total RNA (15 rg) from samples was used per lane. Poly(A)-containing RNA was isolated by passage of total RNA through columns of oligodeoxythymidilate cellulose. The RNA was then electrophoresed on a 1% agarose slab gel, transferred to a nitrocellulose filter, and hybridized with 32P-labeled probes as described previously by Thomas (1980). Molecular
cloning
of proviral integration
sites
The majority of techniques employed in the isolation of molecular clones followed standard protocols (Maniatis et a/., 1982). Briefly, high molecular weight DNA was digested to completion with EcoRl and separated on a 0.8% agarose gel. DNA in the region containing the respective extra band was cut out, electroeluted, and ligated into the separated EcoRl arms of the bacteriophage vector XgtWES XB. After in vitro packaging, the recombinant phages were plated on Escherichia co/i 2600 and screened using an MMTV-envelopespecific probe. Resulting positive clones were picked and propagated in small culture, and the DNA inserts after EcoRI digestion were checked on gel electrophoresis for the correct size. The appropriate host-viral junction fragments were then subcloned into pBR 322 vector for large-scale DNA preparation and fragments containing mainly cellular DNA were transferred into
AND MICHALIDES
pUC 18 or pUC 8 vectors. These inserts were then mapped with frequently cutting restriction enzymes, and specific subfragments of unique-sequence cellular DNA were detected, using a method described by Steinmetz et al. (1980) and were used as probes. Construction and analysis of a GR mouse liver DNA library in EMBL3 vector In general all procedures followed protocols published by Frischauf et a/. (1983). Fresh livers from young GR mice were used for the preparation of high molecular weight DNA. To achieve the proper insert length, the DNA was partially digested with Sau3A, followed by treatment with calf intestine phosphatase (Boehringer Mannheim). The partially digested DNA was then ligated into the BarnHI site of the EMBL3 arms, packaged using a Boehringer Mannheim packaging kit, and plated with an efficiency of 2 X lo5 recombinants per microgram of DNA. The screening and analysis of positive clones were done in the same way as in the first cloning procedure, now using probes specific for the respective integration regions. For subcloning of these fragments, pUC 18 plasmid DNA was chosen. DNA sequencing DNA sequence analysis was carried out by the method of Maxam and Gilbert (1980) on 5 and 20% polyacrylamide gels in 0.5 X TBE buffer containing 7 M urea. Chromosomal assignment and 53 derived probes
of mammary
tumor 40
To map these integration regions we used DNAs from mouse hamster somatic cell hybrids, kindly provided by Dr. John Hilkens (Department of Tumor Biology, The Netherlands Cancer Institute). Characterization of these cell hybrid clones for chromosomal markers has been described elsewhere (Hilkens eta/., 1986). The hybrids generally retained all hamster chromosomes but segregated mouse chromosomes at random. The selected panel consisted of DNAs of 21 hybrids for localization of the MT40142 region and of 16 hybrids for localization of the MT53 area. The DNAs were digested with EcoRl and hybridized to nick-translated probes. The filters were washed under stringent conditions (0.1 %I SSC, 0.1% SDS, 65°C). RESULTS Characterization of extra MMTV integrations mammary tumors of GR
in
Eleven mammary tumors of mouse strain GR were examined by Southern analysis for the acquisition of new proviral MMTV DNA. We used restriction enzymes
INTEGRATION REGION OF MMTV PROVIRUS IN GR MAMMARY
which cut the provirus internally (Cohen et a/., 1979) such as EcoRl and BarnHI, thereby allowing the estimation of the number of extra MMTV proviral copies by the use of appropriate probes. In this way the MMTV-LTR probe was used to identify the extra proviral sequences in mammary tumors, and the MMTVenvelope probe (MMTV-env) to distinguish between right (env+) and left parts of the MMTV provirus. Most tumors showed many acquired extra copies with weak intensities, indicating multiple integration events at later stages of tumor development (Fig. 1). Only a few tumors harbored one extra proviral copy (Fig. 1, lanes 2, 4, and 10). We first studied whether the int-1 or int-2 regions were also involved in these P-type mammary tumors. We hybridized the tumor DNAs to int-1 and int2 probes using different restriction enzymes to cover an area as large as possible of these integration regions according to published restriction maps (Nusse and Varmus, 1982; Peters eta/., 1983; Dickson eta/., 1984). In this way an area was scanned where nearly 90% of the MMTV integrations near the int-1 gene in C3H mammary tumors had been located (Nusse and Varmus, 1982) and 80% of the integration events at the int-2 locus in BR6 mammary tumors (Peters et al., N
40
41
42
46
49
5051
52
53
54
kb 23
TUMORS
231
N 42
N 53
kb 23-
9.4-
2.31
2
3
4
5
6
FIG. 2. Cloning
of the extra MMTV proviral DNA fragments from DNAs of mammary tumors 42 and 53. The EcoRl inserts of the cloned phages were compared with the extra MMTV proviral EcoRl fragments of the original tumor DNAs by hybridization to an MMTWenv probe. Lanes 1 and 4, GR liver DNA; lanes 3 and 6, phage clones derived from the EcoRI-digested tumor DNA, enriched for the 5-to 6-kb area (see Materials and Methods), next to their original tumor DNAs MT 42, lane 2; MT 53, lane 5.
-
1983). Using this approach, we did not observe any MMTV integrations in the int-1 or int-2 regions in the 11 GR P-type mammary tumors examined. Three tumors, MT 40, 42, and 53, with only a single acquired provirus, as determined by EcoRI, BarnHI, and Bg/ll digestions (Fig. 1, lanes 2, 4, and lo), were selected for further cloning experiments.
9.4. 6.6-
43-
Cloning of the viral-cellular junction fragments of the extra MMTV provirus in GR MT 40, 42, and 53 232.0-
1234567891t)ll FIG. 1. Analysis of MMTV proviruses in normal tissue and mammary tumors. Cellular DNA (20 pg) was digested to completion with EcoRl and, after Southern blotting, analyzed for the presence of MMTVDNA-containing fragments using an MMTV-LTR probe. Lane 1, GR liver DNA; lanes 2 through 11, DNAs of GR mammary tumors. HindIlldigested X DNA was used as a molecular weight marker. Right-half fragments of extra bands most suitable for cloning are marked by an arrowhead.
Viral-cellularjunction fragments with the right EcoRl part of the extra MMTV provirus in mammary tumors 40, 42, and 53 (5.5, 5.6, and 5.8 kb in size; see arrowheads in Fig. 1) were chosen to clone. Cellular DNA fragments enriched for the appropriate size by gel separation after EcoRl digestion were cloned in XgtWES XB arms. Approximately 2 X 1 O5 recombinant phages per microgram of DNA were obtained in each case. Those phages containing viral sequences as inserts were identified by screening with an MMTV-env probe. Electrophoresis of EcoRl digestions of the cloned phage DNAs next to digestions of original tumor DNAs confirmed the identity of the clones (Fig. 2). To obtain a useful hybridization probe, we subcloned the phage insert DNAs of each tumor, called XMT40/1, XMT42/
232
SCHUERMANN
1, and XMT53/1, into plasmid vectors, and isolated fragments free of both viral DNA and repeated cellular DNA sequences present in the mouse genome. The resulting cellular probes were first hybridized to tumor DNAs to confirm the alterations of the corresponding regions. The MT 40 cellular DNA probe did indeed recognize the alteration in tumor 40 (Fig. 3) but also in tumor 42 (see below). The MT 53 cellular DNA probe recognized the alteration in tumor 53 DNA (data not shown). Extra MMTV proviruses of GR mammary tumors 40 and 42 are integrated in the same region The extra MMTV integration regions cloned from the three tumors were characterized by restriction enzyme mapping. The resulting maps showed that integration regions of tumors 40 and 42 contained identical restriction fragments toward the right EcoRl site, but differed only slightly in the left Hinfl fragment, also containing part of the MMTV-LTR (Fig. 4). We did exclude that LTR sequences had undergone rearrangements. The difference between these clones therefore should be located directly in the juxtaposed cellular se-
N 40
42
2
3
N
40
42
2
3
kb
2.3-
Al
Bl
FIG. 3. Cloning of the MMTV integration region of mammary tumors 40 and 42. Identification of the cloned integration region by MMTVLTR probe and a cellular MT 40 DNA probe. (A) Hybridization to MMTV-LTR probe to identify the MMTV proviral copies; (6) the same blot rehybridized with probe 40/l, which was derived from the cellular region flanking the integration of the extra MMTV provirus in MT 40. Arrowheads indicate the position of the extra bands in tumors 40 and 42, which were recognized by both probes.
AND
MICHALIDES
quences. Clone MT53/1, however, showed a completely different pattern. For a more precise localization of the difference in the integration regions in these two tumors, we sequenced the viral-host DNA transitions of MT40 and MT42 (Fig. 5). MT42 contained an additional 60 bp at the precise 3’ end of the MMTV-LTR with respect to MT 40. To demonstrate that this 60 bp difference is not the result of cloning procedures, we used a probe covering the viral-host transition, an Il/lspl-Pvull fragment from hMT42/1 (shown in Fig. 4) and hybridized the MT 40 and 42 tumor DNAs digested with Hinfl to this probe. We still observed the 60-bp difference in these two tumor DNAs (Fig. 6, arrowheads). This clearly demonstrates that the difference was already present in the original tumor DNA. This is particularly relevant because tumors 40 and 42 occurred within the same mouse, but in different mammary glands.
Screening of other Mtv-2-associated mammary tumors for integration of extra MMTV proviruses integration regions MT40/42 and 53
in
We constructed a restriction enzyme map of the uninterrupted loci of MT 40 and 53 to recruit suitable restriction fragments for analysis of other mammary tumors for integration of MMlV extra proviruses in these regions. We also constructed a library of GR mouse liver DNA in EMBL3 vector to extend these two regions. For the MT53 region, we isolated a 13-kb fragment from a positive recombinant phage, and for the MT40/42 region two phage clones containing overlapping inserts covering together approximately 28 kb. Subsequent restriction enzyme analysis provided complete maps of the environments around the respective MMTV provirus integrations (now termed the MT40 and MT53 regions), shown in Figs. 7 and 8. This enabled us to isolate probes from several upstream and downstream regions and to use these in a more extensive screening assay (see Figs. 7 and 8). The P-type mammary tumors used were negative for integrations of MMTV proviruses in the int-1 or int-2 regions. The DNAs of these tumors were assayed for integration events in the MT40 or MT53 region using the various probes indicated in Figs. 7 and 8 and the restriction enzymes EcoRl or BarnHI for digestions of the tumor DNAs. In this way, we covered approximately 30 kb of the MT40 region and 20 kb of the MT53 region. A total of 54 P-type mammary tumors of GR were examined for integration of extra MMTV proviruses in the MT40 region and 17 tumors for integration in the MT53 region. None of these tumors, however, showed any MMn/ integrations into either the MT40 or the MT53 integration region.
INTEGRATION
S Hf MT
REGION
OF
MMTV
PROVIRUS
Hf
Av
HHfAv
IN
GR MAMMARY
Pv
E
40
S Hf
AvM
Hf
4
H HfAv
40/l
X
Pv
probe
Hf
E
42 i
t
S Hf MT
233
Hf
X
I
MT
TUMORS
Av
53
Hf Pv
Av
Hf
\I
I
I
Bg
42/l
Hf
Pv
I
I
I
0
proba
I
S
S E
Bg
I 5311 probe
0.2
0.4
0.0
0.6
1.0
1.2
1.4
1.6
I
I
I
I
I
I
I
1
FIG. 4. Restriction maps of the three MMTV integration regions from tumors 40, 42, and 53. Filled boxes, right MMTV sequences; line, cellular DNA region flanking the respective integration sites. Probes derived from the cellular part Restriction enzyme sites: Av, Avall; Bg, Bglll; E, EcoRI; H, HindIll; Hf, Hinfl; M, Mspl; Pv, Pvull; S, Sack X. Xbal.
Chromosomal location of the MT40 and MT53 integration regions Chromosomal location of the MT40 and MT53 integration regions might provide indications for a possible insertion of extra MMTV proviruses in these tumors near other known cellular oncogenes. We therefore hybridized DNAs from a panel of well-characterized mouse-hamster somatic cell hybrids to probes of the MT40 and MT53 integration regions. The results with probe 40/l showed the highest concordancy with markers on chromosome 7 for 20 out of 21 independent hybrid clones. In the case of probe 53/l, taking DNAs from 16 of these clones, a 100% correlation was found for chromosome 16. Since chromosome 7 contains the cellular oncogenes Ha-ras-1 , c-fes, and int-2 (Kozak et a/., 1983; Peters et a/., 1984a), we hybridized the recombinant phage insert DNAs to 3zP-labeled probes of these three
MT 42
MT 40
I
long terminal are indicated
kb
repeat below.
oncogenes. However, we did not observe any positive hybridization, nor did the previously published restriction maps of Ha-ras-1 , c-fes, or int-2 show any similarity to those of MT40. These data indicate that the cloned 40/l region is different from these genes and is separated from them by at least 20-25 kb. Mammary tumors 40 and 42 also did not show any expression of int-2 on Northern blots (data not shown). It is therefore highly unlikely that MMTV integrations in these two tumors resulted in int-2 gene activation. Search for functional and MT53 regions
genes in the MT40
Proviral insertions around int-1 and int-2 genes at the 5’site commonly occur in opposite transcriptional orientation; insertions on the 3’ site instead occur in the same transcriptional orientation as that of the int genes (Nusse et al., 1984b; Dickson et a/., 1984). The
IGTAAAGTATT CCCCTGATAT GTATGACCCC TTGGGATCAA TTCCCAGTAG TTTAAGAAAG,
. ..TGACCCTCAC
GTCGGCCGAC TGCGGCiCTA ATGAAGCTAG GGAGAATTGA
GATGACGAAG ATGGAGATGG GGAGGATGGA GATGAGGAAG ATATTGGTGG GAGGATGGTC ATAGAAACCA GAACGATAGG... FIG. 5. DNA sequence of the viral-host DNA transition of MT40 and MT42. The sequenced using XMT40/1 and XMT42/1. The extra 60-bp cellular DNA that is present The viral sequence ends just at the 3’ end of the viral LTR (Kennedy et a/., 1982).
viral-host in MT42
DNA transitions of MT40 and MT42 were in comparison with MT40 is indicated on top.
234
SCHUERMANN N
40
42
AND
MICHALIDES
regions as a result of MMTV proviral insertion the limited areas covered by the probes used.
within
DISCUSSION
bp
1630
510 390 344 296 220
FIG. 6. Comparison of the MMTV integration sites in tumors 40 and 42. DNA of tumors 40 and 42 (lanes 2 and 3) and GR liver DNA (lane 1) were digested with Hinfl, blotted, and hybridized with the 42/l probe (see Fig. 7). Hinfl-digested pBR322 DNA was used as a size marker. The uninterrupted cellular Hinfl fragment, about 1.7 kb in size, is marked by an asterisk. Arrowheads show the different positions of the interrupted fragment due to insertion of MMTV proviral DNA (about 660 and 720 bp in size).
int-1 and int-2 genes are therefore, in general, located upstream of the integrated MMTV provirus with respect of its transcription orientation. In order to detect expression of genes that might become activated as a result of MMTV proviral integrations in the MT40 or MT53 regions, we used different probes of these regions covering approximately 8 kb upstream as well as 1 kb downstream of the integration site, and hybridized them to the respective tumor RNAs on Northern blots. We observed no hybridization to RNA from tumors 40 and 42 or from tumor 53. There was also no detectable hybridization to control RNA, such as liver and normal mammary gland RNA. When rehybridized to a P-actin probe, all RNA samples showed a consistent signal in the expected range. Thus, although the results are still preliminary, it appears that no transcriptional unit becomes activated in the MT40 and MT53 integration
We have cloned and characterized two integration regions of extra MMTV proviruses from three different GR mammary tumors. These tumors contained only one extra MMTV provirus and appeared clonal with respect to this acquired MMTV copy. In two of the mammary tumors, MT 40 and 42, these cloned regions represent a common area in which the MMTV proviruses integrated only 60 bp apart. In none of 54 P-type mammary tumors of GR did we observe MMTV proviral integrations in this region. Therefore this region is only rarely implicated in the early development of at least some of the GR mammary tumors. Comparison of the sequences around the integrations of the extra MMTV proviruses in tumors 40 and 42 did not reveal any sequences in common and therefore does not imply a homologous recombination as an explanation for this region specific integration. Although mammary tumors 40 and 42 were present in the same mouse, we assume that they represent two different tumors, since (i) they were found in two different mammary glands, respectively the third and fifth right mammary gland, and were equally big at the time of sacrificing the mouse, and (ii) it is highly unlikely that the rearrangement involved only a cellular DNA fragment leaving the viral LTR exactly intact (Fig. 5). Even if this highly improbable event did occur, it should have taken place in a tumor cell of tumor 42 that would have given rise to a clonal outgrowth of tumor 40 in another mammary gland. This we believe to be highly unlikely. Mammary tumors 40 and 42 are therefore two different tumors with an MMTV proviral copy integrated into the same region, being clonally amplified at subsequent cell divisions. The nearby integration of extra MMTV proviruses in tumors 40 and 42 may represent fortuitous integration events in a particular favored genomic area. A similar rare common integration region has also been described by Garcia et al. in BALB/cfC3H in mammary and kidney adenocarcinomas (Garcia et al., 1986). These regions may indeed represent preferred targets for retroviral integration and bear common cellular recognition sequences as proposed by Cohen and Murphey-Corb for baboon endogenous virus (Cohen and Murphey-Corb, 1983). However, direct sequence comparisons of virus-host cell junction fragments in other systems and tumors did not yield any sequence homologies or repeated DNA sequences with respect to host DNA (Van Ooyen and Nusse, 1984; Selten et a/., 1985), which also is not evident from our sequence data. Nevertheless, the implication of these observed
INTEGRATION
probes
REGION
40/3 ,..’-.._, B ,/ ,,/’
screened alr*a*
OF
‘...
..,,
..,
MMTV
PROVIRUS
42/l I 4wm.... B ....” .,..’
n
E
IN GR
“‘..... .__.,,(,,
MAMMARY
(,,,,,.,..........
BE , , ,(,........
“....,,,
TUMORS
235
..“”
. . 40/2 .. .‘. “.,,E 1
.“’
E
)rMT40/42clones x40/1
-
-
HN
phage
ONA
FIG. 7. Topography of the MT40 region. The integrated MMTV provirus here corresponds to position 0. Open boxes, long terminal repeats of MMTV; filled boxes, cellular probes used for the screening of the respective areas indicated. The genomic areas of the clones obtained in the first and second cloning steps are indicated. The restriction enzyme sites for BarnHI (S), EcoRl (E), Kpnl (K), and Xbal (X) are given.
rare common integration mains to be studied.
regions in tumorigenesis
re-
53 region, except for tumors 40 and 42. It is striking that of the 11 tumors examined only tumors 40 and 42, with tumor 53, show a more intense extra MMlV proviral band, indicating a clonal outgrowth of the tumor. These clonal outgrowths of P-type tumors occur less frequently; most of the P-type mammary tumors represent semiclonal outgrowths with respect to their
Do P-type mammary tumors of GR share a common MMTV integration region? In none of the 54 P-type mammary tumors of GR did we observe any MMTV integrations in the MT40/42 or
E
K
B
Eli.
I
X
X
X
i
probes screened sreas
E ........... , _._...... . B
a E
XMT53
E
-
-
FIG. 8. Topography
of the MT53
region.
For legends,
see Fig. 7.
phege
DNA
236
SCHUERMANN
extra MMTV proviruses (Maclnnes et al., 1981; Michalides et al., 1981 b). Proviral insertion often results in an elevated expression of the juxtaposed genes mediated by enhancers of the integrated provirus. A search for constitutive mRNA expression from functional neighboring genes in MT 40, 42, and 53, however, has also failed so far, although there still remains the possibility that the genes of interest were not covered by the probes used, were further away from the integration site, or were expressed in amounts too low to be detected by the respective probes. Our findings are therefore too inconclusive to exclude that genes of the integration regions MT 40, 42, and 53 do play a direct role in the development of tumors 40, 42, and 53. They do, however, exclude these regions as frequently observed integration sites in P-type mammary tumors of GR. The present data do not allow us to formulate a general mechanism of mammary tumor development in mouse strain GR. Nevertheless, it is also worthwhile to consider the low incidence of int-1 and int-2 positive mammary tumors found in GR mice. Together with our findings, these results strongly suggest that MMTV proviral insertion in a particularly favored integration region does not act in general as an initial event in GR mammary tumorigenesis, but that there may be a great diversity of insertion sites in these tumors. The early expression of endogenous MMTV associated with the MTV-2 locus increases the likelihood that GR mammary gland cells become infected with MMTV at very early stages. Histologically, early tumor stages in GR mice show a much greater variety in pattern than other mouse strains (Van Nie, 1980; Percy et a/., 1980), indicating that the affected cells retain a capacity to undergo differentiation into more than one cell type (Hynes et a/., 1984). It is quite conceivable that different oncogenes contribute to the transformation of mammary gland cells at distinct stages of differentiation. This could also explain the low incidence of int-1 or int-2 positive tumors, when int-1 or int-2 would be implicated in the transformation of mammary gland cells in later stages of differentiation. A second point concerns the clonality of GR mammary tumors. It has been demonstrated in serially transplanted GR mammary tumors that the pattern of acquired MMlV proviruses is altered when hormonesensitive tumors become hormone independent (Michalides et a/., 1982; Maclnnes et al., 1981). From these studies it was concluded that the original tumor already harbored different cell populations. GR tumors at later stages may have undergone more selection and as a consequence may have become more homogeneous. This assumption is in contrast to the observed clonal selection of proviral insertions around int2 found in pregnancy-dependent BR6 mammary tumors
AND
MICHALIDES
(Peters eta/,, 198413). The semiclonality of extra MM-IV proviruses in most of the P-type mammary tumors suggests that de nova MMTV integration events may not be directly involved in the earliest stages of mammary gland transformation which lead to P-type mammary tumors and that they merely mark the more progressively growing subclones of these tumors. MMTV insertion, also in the known int-1 and int-2 regions, would in that case activate cellular genes that may be implicated in one of the multiple steps of progression of GR mammary tumors (Michalides, 1984). Our findings in this study would be in line with this view, since a common integration region for extra MMTV proviruses was only found in GR mammary tumors 40 and 42, which show a clonal outgrowth of the tumor, but not in 54 other P-type mammary tumors. Other mechanisms that may account for the induction of mammary tumors in mouse strain GR which are not solely dependent on insertional mutagenesis by MMTV provirus, but which are nevertheless likely to contain additional clonal extra MMTV proviruses, are (i) activation of an oncogene juxtaposed to the Mtv-2associated provirus itself, which may occur in a stageand tissue-specific manner, and (ii) growth stimulation of normal mammary cells via paracrine effects exerted by the large number of multiple preneoplastic lesions in early GR mammary glands due to massive MMTV replication (Van Nie, 1980; Michalides, 1984). Some cells of these lesions may become more easily transformed and stimulated to subsequent outgrowth. Which among these possibilities provides the mechanism for tumorigenesis in GR mammary tumors remains to be determined.
ACKNOWLEDGMENTS We thank Roe1 Nusse for helpful discussions and for carefully reading the manuscript. Mouse-hamster somatic cell hybrid DNAs were kindly provided by John Hilkens; int-1 probe was obtained from Roel Nusse; int-2 probe was a gift from Clive Dickson of the Imperial Cancer Research Fund Laboratories. M. Schuermann was supported by the Deutsche Forschungsgemeinschaft.
REFERENCES BISHOP, J. M. (1983). Cellular oncogenes and retroviruses. Annu. Rev. Biochem. 52, 301-354. COHEN, J. C., SHANK, P. R., MORRIS, V. L.. CARDIFF, R., and VARMUS, H. E. (1979). Integration of the DNA of mouse mammary tumor virus in virus-infected normal and neoplastic tissue of the mouse. Cell 16, 333-345. COHEN, J. C., and VARMUS, H. E. (1980). Proviruses of mouse mammary tumor virus in normal and neoplastic tissues from GR and C3Hf mouse strains. J. Virol. 35, 298-305. COHEN, J. C., and MURPHEY-CORB, M. (1983). Targeted integration of baboon endogenous virus in the BEVI locus on human chromosome 6. Nature (London) 301, 129-l 32. DICKSON, C., SMITH. R., BROOKES. S., and PETERS, G. (1984). Tumorigenesis by mouse mammary tumor virus: Proviral activation of a
INTEGRATION
REGION
OF
MMTV
cellular gene in the common integration region int-2. Ce//37,529536. FRISCHAUF, A. M., LEHRACH, H., POUSTRA, A., and MURRAY, N. (1983). Lambda replacement vectors carrying polylinker sequences. J. Mol. Biol. 170, 827-842. GARCIA, M., WELLINGER, R., VESSAZ, A., and DIGGELMANN, H. (1986). A new site of integration for mouse mammary tumor virus proviral DNA common to BALBlcf(C3H) mammary and kidney adenocarcinemas. EMBO J. 5, 127-l 34. HILKENS, J., CUYPERS, H. T., SELTEN. G., KROEZEN, V., HILGERS, J., and BERNS, A. (1986). Genetic mapping of the Pim-1 oncogene to mouse chromosome 17. Somatic Cell Mol. Genet. 12, 8 l-88. HYNES, N. E., GRONER, B., and MICHALIDES, R. (1984). Mouse mammary tumor virus: Transcriptional control and involvement in tumorigenesis. Adv. Cancer Res. 41, 155-l 79. KENNEDY, N., KNEDLITSCHEK, G., GRONER, B., HYNES, N. E., HERRLICH, P., MICHALIDES, R., and VAN OOYEN, A. J. J. (1982). Long terminal repeats of endogenous mouse mammary tumor virus contain a long open reading frame which extends into adjacent sequences. Nature (London) 295, 622-624. KOZAK, C. A., SEARS, J. F., and HOGGAN, M. D. (1983). Genetic mapping of the mouse oncogenes c-Ha-ras-1 and c-fes to chromosome 7. J. Viral. 47, 217-220. MACINNES, I. I., CHAN. E. C. M. L., PERCY, D. H.. and MORRIS, V. L. (1981). Mammary tumors from GR mice contain more than one population of mouse mammary tumorvirus infected cells. Virology 113,119-129. MANIATIS, T., FRITSCH, E. F., and SAMBROOK, J. (1982). “Molecular Cloning: A Laboratory Manual.” Cold Springs Habor Laboratory, Cold Springs Harbor, NY. MAXAM, A. M., and GILBERT, W. (1980). Sequencing of end-labeled DNA with base-specific chemical cleavages. /n “Methods in Enzymology” (L. Grossman and K. Moldave, Eds.), Vol. 65, pp. 499560. Academic Press, New York. MICHALIDES, R., VAN NIE, R., NUSSE, R., HYNES, N. E., and GRONER, B. (1981 a). Mammary tumor induction loci in GR and DBAf mice contain one provirus of the mouse mammary tumor virus. Cell 23, 165-173. MICHALIDES, R., WAGENAAR, E., GRONER, B., and HYNES, N. E. (1981 b). Mammary tumor virus proviral DNA in normal murine tissue and non-virally induced mammary tumors. J. Viral. 39, 367-376. MICHALIDES, R., WAGENAAR, E., and SLUYSER, M. (1982). Mammary tumor virus DNA as a marker for genotypic variance within hormone-responsive GR mouse mammary tumors. Cancer Res. 42, 1154-1158. MICHALIDES, R. (1984). Amplification of mouse mammary tumor virus DNA in mammary tumors of GRIA mice, a model for hormone sensitive mammary tumors. In: “Hormones and Cell Regulation” (J. E. Dumont and 1. Nunez Eds.), Vol. 8, pp. 255-268. Elsevier, Amsterdam. NUSSE. R., and VARMUS, H. E. (1982). Many tumors induced by the mouse mammary tumor virus contain a provirus integrated in the same region of the host genome. Ce// 31, 99-109.
PROVIRUS
IN GR
MAMMARY
TUMORS
237
NUSSE, R., VAN OOYEN, A., WESTAWAY, D., FUNG, Y. K., and VARMUS, H. E. (1984a). Oncogene activation by proviral insertion. In: “Cancer Cells 2/0ncogenes and Viral Genes” (G. F. van de Woude, A. J. Levine, W. C. Topp, and J. D. Watson, Eds.), pp. 205-209. Cold Springs Harbor Laboratory, Cold Springs Harbor, NY. NUSSE, R., VAN OOYEN, A., Cox, D., FUNG, Y. K. T., and VARMUS, H. (1984b). Mode of proviral activation of a putative mammary oncogene (int-1) on mouse chromosome 15. Nature (London) 307, 131-136. PERCY, D. H., MORRIS, V. L., and MACINNES, J. I, (1980). Comparison of biologic behavior and histology of transplanted mammary tumors in GR mice. J. Nat/. Cancer Inst. 69, 933-938. PETERS, genesis region 377. PETERS, tumor mouse
G., BROOKES. S., SMITH, R., and DICKSON, C. (1983). Tumoriby mouse mammary tumor virus: Evidence for a common for provirus integration in mammary tumors. Cell 33, 369G. C., KOZAK, C., and DICKSON, C. (1984a). Mouse mammary virus integration regions int-1 and int-2 map on different chromosomes. Mol. Cell. Biol. 4, 375-378.
PETERS, G., LEE, A. E., and DICKSON, C. (1984b). Activation of a cellular gene by mouse mammary tumor virus may occur early in mammary tumor development. Nature (London) 309, 273-275. SCHIBLER, U., TOSI, M., PITTET, A.-C., FABIANI, L.. and WELLAUER, P. K. (1980). Tissue-specific expression of mouse a-amylase genes. J. Mol. Biol. 142, 93-l 16. SELTEN, G., CUYPERS, H. T.. and BERNS, A. (1985). of the putative oncogene Pim-1 in MuLV induced EMBO J. 4, 1793-1798.
Proviral activation T-cell lymphomas.
STEINMETZ, M., HOECHTL, J., SCHNELL, H., GEBHARD, W., and ZACHAU, H. G. (1980). Cloning of V region fragments from mouse liver DNA and localization of repetitive DNA sequences in the vicinity of immunoglobulin gene segments. Nucleic Acids Res. 8, 1721-l 729. THOMAS, P. S. (1980). Hybridization of denatured fragments transferred to nitrocellulose. Proc. 77,5201-5205.
RNA and small DNA Nad. Acad. Sci. USA
VAN NIE, R., and VERSTRAETEN, A. A. (1975). Studies mission of mammary tumor virus by CBHf mice. 922-932.
of genetic transhf. J. Cancer 16,
VAN NIE, R., VERSTRAETEN, A. A., and DE MOES. J. D. (1977). Genetic transmission of mammary tumor virus by GR mice. Inf. J. Cancer 19,383-390. VAN NIE, R. (1980). Mammary tumorigenesis in the GR mouse strain. In “Mammary Tumors in the Mouse” (1. Hilgers and M. Sluyser, Eds.). pp. 202-267. Elsevier, Amsterdam. VAN OOYEN, A., and NUSSE, R. (1984). Structure and nucleotide sequence of the putative mammary oncogene int-1 ; Proviral insertions leave the protein encoding domain intact. Cell 39, 233-240. VARMUS, H. E. (1982). Form and function of retroviral proviruses. Science 216, 812-820. VARMUS, H. E. (1984). Annu. Rev. Gener.
The molecular 18, 553-612.
genetics
of cellular
oncogenes.