This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Supplemental material
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrowReprints and Permissions
Right arrow Copyright Information
Right arrow Books from ASM Press
Right arrow MicrobeWorld
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Faschinger, A.
Right arrow Articles by Indik, S.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Faschinger, A.
Right arrow Articles by Indik, S.

 Previous Article  |  Next Article 

Journal of Virology, February 2008, p. 1360-1367, Vol. 82, No. 3
0022-538X/08/$08.00+0     doi:10.1128/JVI.02098-07
Copyright © 2008, American Society for Microbiology. All Rights Reserved.

Mouse Mammary Tumor Virus Integration Site Selection in Human and Mouse Genomes{triangledown} ,{dagger}

Alexander Faschinger,1,2 Francoise Rouault,2,3 Johannes Sollner,4 Arno Lukas,4 Brian Salmons,3 Walter H. Günzburg,1,2 and Stanislav Indik1,2*

Research Institute for Virology and Biomedicine, University of Veterinary Medicine Vienna, Vienna A-1210, Austria,1 Christian-Doppler Laboratory for Gene Therapeutic Vector Development, Vienna A-1210, Austria,2 Austrianova Biotechnology GmbH, Vienna A-1210, Austria,3 Emergentec Biodevelopment GmbH, Vienna A-1010, Austria4

Received 21 September 2007/ Accepted 5 November 2007


arrow
ABSTRACT
 
Based on integration site preferences, retroviruses can be placed into three groups. Viruses that comprise the first group, murine leukemia virus and foamy virus, integrate preferentially near transcription start sites. The second group, notably human immunodeficiency virus and simian immunodeficiency virus, preferentially targets transcription units. Avian sarcoma-leukosis virus (ASLV) and human T-cell leukemia virus (HTLV), forming the third group, show little preference for any genomic feature. We have previously shown that some human cells sustain mouse mammary tumor virus (MMTV) infection; therefore, we infected a susceptible human breast cell line, Hs578T, and, without introducing a species-specific bias, compared the MMTV integration profile to those of other retroviruses. Additionally, we infected a mouse cell line, NMuMG, and thus we could compare MMTV integration site selection in human and mouse cells. In total, we examined 468 unique MMTV integration sites. Irrespective of whether human or mouse cells were infected, no integration bias favoring transcription start sites was detected, a profile that is reminiscent of that of ASLV and HTLV. However, in contrast to ASLV and HTLV, not even a modest tendency in favor of integration within genes was observed. Similarly, repetitive sequences and genes that are frequently tagged by MMTV in mammary tumors were not preferentially targeted in cell culture either in mouse or in human cells; hence, we conclude that MMTV displays the most random dispersion of integration sites among retroviruses determined so far.


arrow
INTRODUCTION
 
Integration of proviral DNA into the host cell genome is an essential step in retroviral replication. Determination of the integration profiles of various retroviruses not only is relevant for a more detailed understanding of the early steps of retroviral replication but also might be beneficial for the development of drugs targeting the integration process or for the creation of safer retroviral vectors for gene therapy.

We have analyzed integration targeting of mouse mammary tumor virus (MMTV) in human and murine genomes. The aim of this study was to compare the distribution of the de novo integration sites generated experimentally with what was found for other retroviruses for which similar studies have already been performed. We also investigated the integration patterns of MMTV in cells from two different species and thus were able to analyze whether there is any species-specific effect that would influence the integration targeting.

At present, most of the studies on retroviral integration sites have been conducted with retroviruses or retroviral vectors targeting the human genome. There are only few reports investigating the distribution of integration sites in more than one species. For example, Barr and coworkers analyzed avian sarcoma-leukosis virus (ASLV) and human immunodeficiency virus (HIV) integration target preferences in the chicken genome and compared it with the distribution of the integration sites observed in the human genome. The distributions of de novo integration sites for both viruses were generally similar in both genomes, suggesting that any cellular factor(s) important for insertional targeting are likely conserved between chickens and humans (2).

So far, a number of retroviruses belonging to different genera have been analyzed with respect to their integration site selection (2, 6, 10, 12, 22, 25, 29, 30, 35, 43, 44). Interestingly, retroviruses belonging to diverse genera display different preferences for target site selection (12, 29). It also appears that the integration profiles of various retroviruses can be predicted from the phylogenetic relationships of retroviral integrases (12). The lentiviruses, HIV as well as simian immunodeficiency virus (SIV), tend to integrate within genes or transcription units (2, 6, 22, 29, 35, 44). By contrast, a gammaretrovirus that has been utilized in gene therapy clinical trials for over a decade, murine leukemia virus (MLV), shows a bias in favor of integration near transcription start sites and CpG islands (44). A similar integration pattern was also demonstrated for a member of the Spumavirus genus, foamy virus (FV) (43). The other retroviruses studied so far, ASLV (Alpharetrovirus) and human T-cell leukemia virus (HTLV) (Deltaretrovirus), show the most random distribution of integration sites seen to date (2, 12, 29, 30, 43).

To our knowledge, no work analyzing de novo insertional preferences of Betaretrovirus, the remaining genus in the Orthoretrovirinae subfamily, has been published to date. In this study, we have analyzed the insertion site preferences of MMTV, a betaretrovirus that is known to be associated with mammary adenocarcinomas and T-cell lymphomas in mice (3, 11, 13, 28). MMTV is believed to cause tumors, in most cases via insertional mutagenesis. Insertion of an MMTV provirus in the vicinity of a proto-oncogene leads to an inappropriate transcriptional activation of adjacent genes that cause oncogenic transformation of the infected cells (other mutagenic events resulting in cell transformation, such as promoter insertion or truncations, are less common) and their clonal expansion (9). MMTV-based insertional mutagenesis is one of the most efficient methods to identify previously unknown genes and potential gene pathways that are involved in mammary differentiation and neoplasia. For this reason, the determination of MMTV integration sites in mouse mammary tumors has attracted the interest of many researchers, and a number of MMTV insertion sites have been mapped (7, 8, 15, 16, 27, 31-34, 36, 38, 40, 41). MMTV proviral sequences were often found near one of the Wnt, Notch, or Fgf gene families. Although these common insertion sites (CIS) were determined by use of clonally expanded cells, in which the MMTV enhancer activated the nearby proto-oncogene promoter, it is still not clear whether MMTV shows a bias in favor of integration, for example, within genes or transcription start sites.

To determine whether the integration of MMTV proviruses shows any bias in favor of any genomic feature or whether the virus rather exhibits a more random integration pattern, we have cloned and sequenced 468 unique virus-host junction sequences from acutely infected cells. The integration sites were then analyzed and compared with the integration profiles of the other retroviruses for which such analysis has been already performed.


arrow
MATERIALS AND METHODS
 
Virus preparation and cell culture. MMTV-EGFP (molecular clone of MMTV carrying the EGFP gene in the 3' long terminal repeat [3'LTR]) and MMTV(GR) (a wild-type MMTV) virus particles were prepared as described previously (19). Briefly, the production of the viruses was induced by adding 10–6 M dexamethasone (DEX) to CGRES6 or GR producer cell lines 24 h before virus harvest. Supernatants containing the viral particles were filtered through a 0.45-µm filter and concentrated (MMTV-EGFP) by ultracentrifugation at 120,000 x g for 2 h at 4°C. The viral particles were resuspended in a 1/100 volume of fresh medium and used for infection of a human breast cell line, Hs578T, or a mouse mammary cell line, NMuMG. The Hs578T cells infected by MMTV(GR) were propagated in 10–6 M DEX-containing medium for 4 weeks as described previously (18).

Sequencing and mapping of integration sites. Integration sites were cloned by ligation-mediated PCR essentially as described previously (44). Briefly, DNA was harvested 48 h after infection with MMTV-EGFP. DNA from the MMTV(GR)-infected Hs578T cells was harvested 4 weeks after the initial infection. The genomic DNA was digested with MseI, and linker DNA was ligated to the digested ends. Virus-host junction sequences were PCR amplified using one primer complementary to EGFP or to the viral LTR sequence and the other primer complementary to the linker (44). Nested PCR was performed with another primer pair, and the resulting products were cloned into pCR2.1 vector (Invitrogen, CA) and sequenced. The primers used for this study were as follows: EGFP1, 5'-CCA ACG AGA AGC GCG ATC AC-3'; 1370F, 5'-CGT CTC CGC TCG TCA CTT AT-3'; EGFP2, 5'-CTC GGC ATG GAC GAG CTG TA-3'; linker 1, 5'-GTA ATA CGA CTC ACT ATA GGG C-3'; and linker 2, 5'-AGG GCT CCG CTT AAG GGA C-3'.

The sequences were aligned to the human genome (University of California, Santa Cruz, UCSC; hg18, March 2006, NCBI Build 36.1) by use of the BLAT program (http://genome.ucsc.edu/cgi-bin/hgBlat). A sequence was considered a genuine integration event if it contained both the 3'LTR from the nested primer to the end of 3'LTR (CA) and the linker sequence. Additionally, the sequence must match the human or mouse genome with 95% or greater identity, and the match must start immediately (within 3 bases) after the end of LTR sequence. Using these criteria, we mapped 468 integration sites (298 in the human and 170 in the mouse genome). Comparison of integration targeting with other retroviruses was performed using data sets (298 randomly picked sites for each virus) downloaded from GenBank and mapped to the human genome. The following reports and GenBank accession numbers were used for the site selection: for HTLV, reference 12 and numbers EF-580177 to EF-580913; for ASLV, reference 29 and numbers CL528303 to CL528772; for HIV, reference 44 and numbers AY516881 to AY517469; for SIV, references 10 and 17 and numbers AY679815 to AY680027 and AY728482 to AY728804; for MLV, reference 44 and numbers AY515855 to AY516880; and for FV, reference 43 and numbers DU798511 to DU799518. A set of 10,000 random integration sites in human and mouse genomes was generated and analyzed together with the integration sites of the retroviruses and was used as a reference. Locations of the genomic features were downloaded from the UCSC genome database.

CIS in MMTV-induced tumors. Thirty-three CIS commonly tagged in MMTV-induced mammary tumors determined by Theodorou et al. (41) were used for comparison with the de novo integrations. Similarly, for comparison of the in vitro integrations with the more frequently tagged signaling pathways or protein domains, the following data sets were used: FGF, TSP_1, PKINASE, CATION_ATPase, WNT, EGF, RAS, CYCLIN_C, GLA, IBN_N, TRUD for the frequently tagged Pfam protein domains, and mitogen-activated protein kinase signaling pathway, Wnt signaling pathway, focal adhesion, regulation of actin cytoskeleton, calcium signaling pathway, Gap junction, Notch signaling pathway, and Hedgehog signaling pathway for the Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways (41).

Determination of KEGG pathways was performed using the U.S. National Institutes of Health Database for Annotation, Visualization and Integrated Discovery (DAVID) software for KEGG (http://david.abcc.ncifcrf.gov/). Similarly, the Pfam protein domains represented in our data sets were identified using DAVID. Binomial distribution was used for calculation of the probability that a single integrant hits one of the 33 CIS that including a ±50-kb region occupy 8.5 x 106 bases of the mouse genome (2.9 x 109 bases) as follows: P = n!/x! (n – x)! (p)x (1 – p)(n – x), where n equals the number of trials (170), x equals the number of hits (1), and p is the probability of success on a single trial (8.5 x106/2.9 x 109). Nonuniform distribution of CIS was ignored.

Alignment and phylogenetic analysis. Phylogenetic trees were constructed by the neighbor-joining method using alignments of the amino acid sequences of the integrase proteins of MMTV, HTLV, ASLV, HIV, SIV, MLV, and FV. Sequences were analyzed using the CLUSTAL W 1.7 program (42) and software included in the PHYLIP package (14). Bootstrap analysis was made using 1,000 replications.

Nucleotide sequence accession numbers. All integration site sequences have been deposited into GenBank under accession numbers EU018613 to EU019080.


arrow
RESULTS
 
Mapping of MMTV integration sites in the mouse and human genomes. Many genetic approaches that are routinely used with other retroviruses are complicated when they are applied to studies conducted with MMTV, mainly due to the poor replicative capacity of the virus in vitro. The titers obtained in cell culture are considerably lower than titers observed for other retroviruses, which makes analysis of MMTV integration targeting more difficult. Nevertheless, using linker-mediated PCR (LM-PCR) we could determine, on average, up to 10 integration sites in a single-round infection experiment. In order to develop a large data set suitable for statistical analyses and comparisons with the integration site specificities of other retroviruses, we thus performed a series of single-round infection experiments.

The second technical difficulty in working with MMTV is the presence of multiple endogenous MMTV sequences in the available susceptible mouse cell lines. MMTV-specific primers used for the LM-PCR could thus hybridize with the endogenous sequences, which would then lead to the amplification of the endogenous segments instead of the de novo virus-host junctions. As the unique de novo junctions are underrepresented in the infected cells (one unique de novo integration compared to 2 to 10 endogenous proviruses per cell in most inbred mouse strains [9]), one might expect that the endogenous sequences amplified in the LM-PCR would predominate over the de novo integrations. We circumvented this problem by using an infectious molecular clone of MMTV carrying the EGFP gene in the 3'LTR (MMTV-EGFP) (19). The presence of the EGFP gene on the integrated proviral sequence allowed us to use EGFP-specific primers, thereby specifically amplifying only those virus-host junction sequences that resulted from the de novo infection events.

We used the recombinant MMTV-EGFP virus for the infection of the murine mammary cell line NMuMG as well as human breast cells, Hs578T. It had long been believed that human cells are not susceptible to MMTV; however, our previous studies revealed that among others, the human cell line Hs578T sustains MMTV infection and also supports replication of the virus (18, 19). This cell line was infected with the MMTV-EGFP virus in single-round infection experiments, in which the DNA was extracted from the infected cells 48 h after infection. Additionally, as this human cell line does not contain endogenous MMTV sequences, we also infected these cells with a wild-type MMTV [MMTV(GR)]. In this case, the infected cells were cultured in the presence of 10–6 M DEX to support the replication of the virus for 4 weeks. At the time of genomic DNA extraction, all of the cells in culture were infected, as could be demonstrated by an immunofluorescence staining of Gag proteins (18). Genomic DNA extracted from either the MMTV-EGFP- or the MMTV(GR)-infected human cells was then submitted to LM-PCR as described previously (44). Sequences were trimmed and mapped to the human genome (UCSC; hg18, March 2006, NCBI Build 36.1). A total of 298 unique integration sites were mapped from the infected Hs578T cells. Of these, 82 and 216 integrations were cloned from the MMTV-EGFP- and MMTV(GR)-infected cells, respectively. Local features at each integration site were examined, and the data sets obtained for both viruses, MMTV-EGFP and MMTV(GR), were compared. No statistically significant differences in any of the analyzed genomic features were found, and so these two data sets were pooled and used for the further analyses reported here.

NMuMG murine cells infected with MMTV-EGFP were analyzed in a similar manner. The genomic DNA was extracted 48 h after the infection and amplified by LM-PCR. The PCR products were cloned and sequenced and the positions of the integration sites were determined using the mouse genome database (UCSC; mm8, Feb. 2006, NCBI Build 36). A total of 170 unique integration sites were determined.

Integration in transcription units. First, we sought to assess whether MMTV displays a tendency to integrate within transcription units by measuring the proportion of MMTV integrants within RNA polymerase II genes by use of human or mouse reference sequence (RefSeq) genes (Table 1). This analysis revealed that of the 298 unique integrations mapped in the human genome, 105 (35.35%) occurred in the defined RefSeq gene set. Similarly, analysis of the integrations in the mouse genome showed that 55 of 170 (32.72%) integration events mapped within RefSeq genes (Table 1). When the integration preferences of MMTV determined in human and mouse genomes, respectively, were compared to the computer-generated random sites found in RefSeq, no significant difference was observed (35.92% and 34.99%, respectively) (Table 1).


View this table:
[in this window]
[in a new window]

  		TABLE 1. Integration frequency near genomic features

Next, we compared the frequency of MMTV integration in transcription units observed for human cells with those of six other retroviruses (HIV, MLV, SIV, FV, ASLV, and HTLV) by using the same number (298 per virus) of randomly picked integrations, as well as to the in silico-generated random sites. As previously reported, the highest frequency of proviral integrations in transcription units was seen for HIV (70.0%; P < 0.0001) and SIV (75.1%; P < 0.0001). MLV, ASLV, and HTLV showed a modest, but significant, preference for integration within RefSeq genes (45.9%, 46%, 44.6%; P < 0.0001). FV, in contrast, displayed no significant preference for RefSeq genes (28.0%) compared to random sites (35.0%). Similarly to FV, MMTV did not show any evidence for a tendency to integrate in RefSeq genes either in human or in mouse cells (Table 1).

The distribution of integration sites within the transcription units was also analyzed. All transcription units were divided into eight bins, starting from the transcription start site. In agreement with previous analyses, MLV integrations were preferentially located near the transcription start of genes, with 15.1% of integrations found in the first bin (P < 0.0001). HIV and SIV, the two viruses integrating preferentially into genes, tended to integrate in the middle of genes. A modestly higher frequency of integrations in the first four bins was found for HTLV. MMTV, like ASLV, showed roughly even distribution throughout all eight bins, irrespective of whether the integrations in the mouse or the human genome were analyzed (Fig. 1A).


Figure 1
View larger version (41K):
[in this window]
[in a new window]

  		FIG. 1. Frequency of retroviral integration (MMTV in human [MMTV h] and mouse [MMTV m] genomes, HTLV, ASLV, HIV, SIV, MLV, and FV) within and around transcription units, transcription start sites, and CpG islands. (A) Integration sites of the retroviruses were mapped relative to RefSeq transcription units divided into eight bins and plotted as the percentage of all integrations within each bin. (B) Integration frequencies near transcription start sites of RefSeq genes. Percentages of all integrations per kb for each interval near the transcription start sites of RefSeq genes were plotted for each virus. (C) Integration near CpG islands. Regions upstream and downstream of CpG islands were scored for the percentages of integrations per kb for each virus. The scattered line represents a mean random value calculated using the computer-generated random sites. The random value for the mouse genome is omitted.

Taking these data together, we conclude that independently of species, MMTV does not preferentially integrate within genes or any segment within transcription units.

Integration near transcription start sites and CpG islands. Previously, it was reported that some retroviruses, such as MLV and to a lesser extent FV, favor integration in the vicinity of transcription start sites (43, 44). Thus, it was not surprising that when the data sets obtained for MLV and FV were analyzed, the proportions of MLV (16.1%) and FV (6.0%) integrations located within ±2 kb of transcription start sites were significantly greater than expected for random integrations for both viruses (P values of <0.0001 and <0.01, respectively). A weak bias in favor of transcription start sites was also seen for HTLV (5%; P < 0.07) (Table 1; Fig. 1B). For MMTV, with 1.7% integrations found in human and 2.4% in mouse cells, no significant preference for transcription start sites compared to computer-generated random sites (3.2% and 2.6%, respectively) was found. Similarly, no preference for ±2 kb of transcription start sites was observed for ASLV, HIV, and SIV (3.0%, 2.0%, and 0.6%, respectively) (Table 1; Fig. 1B).

CpG islands are thought to be associated with transcription initiation in vertebrates. These regions remain unmethylated in all cell types and often surround promoters of housekeeping genes (1). Studies of integration tendencies in CpG islands thus not only may reveal preferences of various retroviruses for integration near transcription start sites but also can possibly uncover an influence of DNA methylation on integration. Using data sets for the six previously analyzed retroviruses, we were able to generate results which confirmed those found by others (10, 12, 29, 30, 43, 44). Again, the strongest preference for integration near CpG islands was found for MLV, with 20.1% integrations within ±2 kb of CpG islands, followed by FV, with 9.4% of integrants found in the same region. A weak preference for CpG islands was seen also for HTLV (7.7% in the ±2-kb window; P < 0.01). All the other viruses (ASLV, HIV, SIV) including MMTV showed no preference for CpG islands compared to random data sets for both human and mouse cells (Table 1; Fig. 1C).

Integration in repetitive elements. We also determined the frequency of integrations in repetitive elements, which account for nearly half of the human and mouse genome sequences. Of these, long interspersed nuclear elements (LINEs), short interspersed nuclear elements (SINEs), and LTR retrotransposons (LTRRs) are the most abundant. No tendency toward integration into these regions was determined for MMTV in either mouse or human cells (Table 1). Among the analyzed retroviruses, HTLV is noteworthy for its tendency to avoid integrating in the most frequent repeats. Low numbers of integrations (7.0% [P < 0.01], 13.7% [P < 0.01], and 5% [P < 0.05]) compared what was seen for to the in silico-generated integrants (12.6%, 19.6%, and 8.1%) were observed for SINEs, LINEs, and LTRRs, respectively. Additionally, significantly greater numbers of the HTLV integrants were found in nonrepetitive compared to random sequences (68% and 54.8%, respectively; P < 0.001). A similar trend was also seen for HIV and MLV (63% [P < 0.01] and 68.6% [P < 0.001], respectively). A significantly decreased probability for integration within LINEs was also detected for MLV and FV (10.5% [P < 0.001] and 15.1% [P < 0.05], respectively). ASLV, on the other hand, showed a modest trend in favor of integration into LINEs (24.3% [P < 0.05]) (Table 1).

Other repetitive elements, such as DNA repeat elements, simple repeats (microsatellites), low-complexity repeats, satellite repeats, RNA repeats (including RNA, tRNA, rRNA, snRNA, and scRNA), and other repeats, represent a minor portion of the overall genome, and no significant number of MMTV integrations within any of these genomic segments could be observed (Table 1).

Comparison of de novo integrations with commonly tagged sites detected in tumors. Recently, high-throughput screening of MMTV proviral insertion sites (664 independent virus-host flanking sequences) in 160 mammary tumors, arising after infection of BALB/c mice with C3H-MMTV, identified 33 CIS frequently targeted by MMTV in tumors (41). The availability of the database of the identified integration sites prompted us to investigate whether MMTV tends to integrate in the CIS identified in the above-mentioned study. For tumor cells, the most frequent genomic loci hosting MMTV proviral sequences belonged either to Wnt or to Fgf genes (about 60% and 40% of tumors harbor the MMTV-tagged Wnt and Fgf genes, respectively) (9, 41). When the de novo MMTV integrations determined in either mouse or human genomes were screened for the presence of the integrants within ±50 kb of the Wnt or Fgf genes, no such integration event was found (we included the ±50-kb gene-flanking regions because enhancer-mediated activation is known to act over long distances [9]). Likewise, the other CIS identified in the study (41) were not tagged by MMTV after infection of human cells. One out of 170 proviral insertions identified in the mouse genome had integrated into the Odz1 gene, one of the CIS (Table 2). If we take into account the probability that one of the 170 integrations occurs in the vicinity of a CIS (including ±50-kb flanking regions occupying 8.4 x 106 nucleotides in the mouse genome) is 0.32, then such a result is not surprising and probably reflects random integration. One and two integrations in the human orthologues of mouse CIS were also identified for ASLV and HIV, respectively.


View this table:
[in this window]
[in a new window]

  		TABLE 2. De novo MMTV integrants found in CIS determined for MMTV-induced tumorsa

The recent study of Theodorou and colleagues also revealed that genes often targeted in mammary tumors belong to the same gene family or act in the same oncogenic pathway (41). Therefore, we sought to examine if some of the commonly tagged pathways or protein families were overrepresented in our data set. First, we identified all genes, including ±50-kb flanking regions, associated with MMTV integrations and then we determined any signaling cascades in which these MMTV-tagged genes might be involved by using the U.S. National Institutes of Health DAVID software for KEGG pathways (see the supplemental material). In a similar way, we identified any Pfam protein domains (DAVID) present in our data set (see the supplemental material). The protein domains and pathways were then compared to those that are significantly overrepresented in the MMTV-tagged mammary tumors (41). Data from the other six retroviruses as well as the random data sets were also analyzed. We could not detect any bias favoring the signaling pathways overrepresented in the MMTV-tagged tumors (Table 3). A modestly greater number of genes encoding proteins containing the epithelium growth factor (EGF) domain was detected among those integration sites targeted by MMTV in the mouse genome compared to what was seen for random integrations (4.6% of the tagged genes encode protein with the EGF domain [P < 0.05]) (Table 4). One of these genes, Odz1, belongs to the CIS. The other genes (3110045G13Rik, Celsr2, Ltbp1, Nrxn3, Selp, and Vldlr) have not previously been shown to host MMTV proviruses. Not even a weak preference for these groups (or any other group of genes) was found in the MMTV-infected human cells.


View this table:
[in this window]
[in a new window]

  		TABLE 3. Frequency of de novo integrations in genes that belong to commonly tagged KEGG pathways in MMTV-induced mammary tumors


View this table:
[in this window]
[in a new window]

  		TABLE 4. Frequency of de novo integrations in genes encoding proteins with Pfam protein domains commonly tagged in MMTV-induced mammary tumors


arrow
DISCUSSION
 
MMTV belongs to the Betaretrovirus genus, the only genus of the Retroviridae for which a large-scale mapping study of de novo retroviral integration sites has not been performed so far. In this study, we investigated the integration site selection after infection of human and mouse mammary cells with MMTV. A total of 468 integration sites were obtained. Of these, 298 unique integration sites were determined from independent infection events in human cells, which further substantiate our previously published data that human cells are infectible with MMTV (19) and also support the replication of the virus (18).

Since MMTV is a relatively "low-titer virus" and since the mouse genome contains endogenous MMTV loci, we circumvented some of the difficulties in working with MMTV by performing a series of infection experiments using a recombinant MMTV-EGFP virus. In parallel, since the human genome does not contain endogenous MMTV sequences, we infected the human breast cells with wild-type MMTV(GR). In the presence of DEX, a glucocorticoid stimulating the major MMTV promoter, the virus eventually infected all cells in culture (determined by immunofluorescence Gag imaging), thereby making the determination of the integration sites easier (18).

We sought to minimize any possible selection bias of integration sites by avoiding antibiotic selection of cell clones or extended growth of infected cells in cell culture. Comparisons of the integration sites targeted by MMTV-EGFP (harvested 48 h after infection) and MMTV(GR), where further infections and superinfections are possible, did not reveal any obvious differences. Prior reports from others have also shown that short-term drug selection does not significantly influence the populations of recovered integration sites (22). We did not observe any prominent LM-PCR product that would reflect early integration events, lending credence to the idea that 4 weeks of cultivation time is not long enough for the significant clonal expansion of early infectants.

The integration profile determined after MMTV infection of human cells resembled that seen for mouse cells. In both cases, we did not observe preferential targeting of genes, in contrast to what has previously been shown for HIV and SIV. Likewise, we have not seen integration favoring transcription start sites as has been reported for MLV and FV. We did find, on the other hand, that MMTV, irrespective whether the virus infects mouse or human cells, appears to integrate randomly, as has been shown for ASLV and HTLV. The random dispersion of the integration events in both species suggests either that there is no cellular factor(s) that interacts with the MMTV preintegration complex (PIC) and helps to tether this complex to the chromosomal DNA or, more likely, that potential cellular factor(s) are conserved between mice and humans. Another possible reason for the random dispersion of the MMTV integration sites might also be the lack of expression of the cellular factor(s) in the cell lines used in the study. Although it cannot be ruled out, it does seem rather unlikely that the PIC-interacting partner(s) would be absent in these two cell lines specifically. We have selected mouse and human mammary cell lines because mammary epithelial cells are major targets during in vivo infection.

Not much is known about cellular factors interacting with PICs, especially with the PIC of MMTV. Recently, it was shown that LEDGF/p75 binds to HIV integrase and plays a role in HIV integration site selection (6, 23, 24). Another cellular protein that interacts with the PIC is BAF. This protein, which, in contrast to LEDGF/p75, binds to MLV as well as to HIV PICs, is known to be a factor that blocks the self-destructive integration of proviral DNA into its own genome in vitro (5, 21, 37). Recently, it was also shown that BAF, together with its own binding partner, the nuclear-envelope-associated protein emerin, is required for the appropriate localization of HIV cDNA before chromatin engagement. In contrast to HIV, emerin is not required for the appropriate localization of MLV cDNA before chromatin tethering (20). Rather, the PIC of MLV interacts with other members of the LEM family of inner nuclear membrane and nucleoplasmic proteins such as LAP2{alpha} (39). Further work is needed to uncover whether such proteins also interact with PICs of viruses displaying random distributions of integration sites.

MMTV, analogously to ASLV and HTLV, generates 6-bp-long duplications of host sequences flanking the integrated proviruses (26). Additionally, Derse and coworkers suggested that the integration profiles of various retroviruses could be predicted from the phylogenetic analysis of retroviral integrases (12).

In such a phylogenetic analysis, the integrase of MMTV is clustered in a branch comprising integrases of ASLV and HTLV (Fig. 2), two retroviruses that display little integration preference for any genomic feature, providing a basis for the prediction that the integration profile of MMTV will resemble that of ASLV and HTLV. In accordance with this prediction, we observed that the integration site preference of MMTV most closely resembles that of ASLV and HTLV. However, we did not observe any preference for integration within genes. Whereas 44.8% and 46% of the integrated HTLV and ASLV proviruses, respectively, were found within genes, only 32.7% and 35.4% of the MMTV integrants in mouse and human cells, respectively, landed within RefSeq regions, a frequency that is not significantly different from that seen for the computer-generated random sites. Additionally, unlike for ASLV and HTLV, we did not see a significant difference in frequencies of MMTV integration in repetitive elements compared with random sites. HTLV also does not preferentially integrate into LINEs, SINEs, and LTRRs, whereas ASLV shows a slight bias in favor of LINEs. Taken together, it appears that MMTV displays the most random integration site distribution among retroviruses for which this has been determined to date.


Figure 2
View larger version (13K):
[in this window]
[in a new window]

  		FIG. 2. Phylogenetic tree constructed by the neighbor-joining method based on an alignment of amino acid sequences of the integrase proteins of MMTV, HTLV, ASLV, HIV, SIV, MLV, and FV. Bootstrap values adjacent to each node represent percentages of 1,000 trees supporting the clustering. The length of duplication generated by each virus during integration and the preferential targeting of each virus are shown. The total branch lengths between the two species are proportional to the distance for each pair of species. The scale bar shows the distance.

The random distribution of the integration sites might raise the question of whether the integration was mediated by the MMTV integrase or whether it was a nonspecific, integrase-independent process. The fact that we, using both MMTV-EGFP and MMTV(GR) viruses for infection, have determined the 6-bp-long target site duplications (data not shown and reference 18, respectively) that are specific for MMTV (26), as well as the fact that all the sequences included in the analysis were terminated be the canonical CA dinucleotide, lead us to believe that the insertions of proviral sequences into the host genome were legitimate integrations mediated by the MMTV integrase.

MLV-derived vectors are currently the most widely used vectors in clinical gene transfer protocols. Clinical trials with MLV-based vectors revealed that, under certain circumstances, insertional mutagenesis of the LMO2 gene led to the development of leukemias in some of the treated X-linked severe combined immune deficiency patients (reviewed in reference 4). Taking into account the random integration targeting exhibited by MMTV, it can be speculated that the construction of a hybrid vector carrying the MMTV integrase would improve safety. Integrase was previously demonstrated to be a major determinant of retroviral insertion targeting, and a hybrid HIV carrying the integrase of MLV targeted sites with a specificity close to that of MLV (22). Obviously, this concept would require further studies to elucidate whether the MMTV integrase alone would be sufficient for random site selection or whether other factors that are known to participate in the integration reaction, such as Gag and terminal LTR sequences, are needed. Concerns regarding the safety of such vectors could also be raised due to the fact that MMTV is known as an insertional mutagen. However, we have not observed an increased integration targeting of the genes previously described, as CIS are often tagged by MMTV in mammary tumors. Likewise, we have not seen a significant bias in favor of genes encoding proteins containing domains or involved in signaling pathways commonly found in MMTV-induced tumors. Thus, it may be possible to improve the safety of gene therapy vectors by using an integrase which would not direct proviral DNA to the vicinity of or within genes. The integrase of MMTV, the virus showing the most random distribution of integration sites yet, might be a good option for such an improvement.


arrow
ACKNOWLEDGMENTS
 
This work was supported by the Christian Doppler Forschungsgesellschaft.


arrow
FOOTNOTES
 
* Corresponding author. Mailing address: Research Institute for Virology and Biomedicine, University of Veterinary Medicine Vienna, Vienna A-1210, Austria. Phone: 43(0)1250772333. Fax: 43(0)1250772690. E-mail: stanislav.indik{at}vu-wien.ac.at Back

{triangledown} Published ahead of print on 21 November 2007. Back

{dagger} Supplemental material for this article may be found at http://jvi.asm.org/. Back


arrow
REFERENCES
 
    1
  1. Alberts, A., A. Johnson, J. Lewis, M. Raff, K. Roberts, and P. Walter. 2002. Molecular biology of the cell, 4th ed. Garland Science, Oxford, United Kingdom.
    2. 2
  2. Barr, S. D., J. Leipzig, P. Shinn, J. R. Ecker, and F. D. Bushman. 2005. Integration targeting by avian sarcoma-leukosis virus and human immunodeficiency virus in the chicken genome. J. Virol. 79:12035-12044.[Abstract/Free Full Text]
    3. 3
  3. Bittner, J. J. 1936. Some possible effects of nursing on the mammary gland tumor incidence in mice. Science 84:162.[Free Full Text]
    4. 4
  4. Bushman, F. D. 2007. Retroviral integration and human gene therapy. J. Clin. Investig. 117:2083-2086.[CrossRef][Medline]
    5. 5
  5. Chen, H., and A. Engelman. 1998. The barrier-to-autointegration protein is a host factor for HIV type 1 integration. Proc. Natl. Acad. Sci. USA 95:15270-15274.[Abstract/Free Full Text]
    6. 6
  6. Ciuffi, A., M. Llano, E. Poeschla, C. Hoffmann, J. Leipzig, P. Shinn, J. R. Ecker, and F. Bushman. 2005. A role for LEDGF/p75 in targeting HIV DNA integration. Nat. Med. 11:1287-1289.[CrossRef][Medline]
    7. 7
  7. Clausse, N., D. Baines, R. Moore, S. Brookes, C. Dickson, and G. Peters. 1993. Activation of both Wnt-1 and Fgf-3 by insertion of mouse mammary tumor virus downstream in the reverse orientation: a reappraisal of the enhancer insertion model. Virology 194:157-165.[CrossRef][Medline]
    8. 8
  8. Clausse, N., R. Smith, C. M. Calberg-Bacq, G. Peters, and C. Dickson. 1993. Mouse mammary-tumor virus activates Fgf-3/Int-2 less frequently in tumors from virgin than from parous mice. Int. J. Cancer 55:157-163.[Medline]
    9. 9
  9. Coffin, J. M., S. H. Hughes, and H. E. Varmus. 1997. Retroviruses. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY.
    10. 10
  10. Crise, B., Y. Li, C. Yuan, D. R. Morcock, D. Whitby, D. J. Munroe, L. O. Arthur, and X. Wu. 2005. Simian immunodeficiency virus integration preference is similar to that of human immunodeficiency virus type 1. J. Virol. 79:12199-12204.[Abstract/Free Full Text]
    11. 11
  11. Dekaban, G. A., and J. K. Ball. 1984. Integration of type B retroviral DNA in virus-induced primary murine thymic lymphomas. J. Virol. 52:784-792.[Abstract/Free Full Text]
    12. 12
  12. Derse, D., B. Crise, Y. Li, G. Princler, N. Lum, C. Stewart, C. F. McGrath, S. H. Hughes, D. J. Munroe, and X. Wu. 2007. HTLV-1 integration target sites in the human genome: comparison with other retroviruses. J. Virol. 81:6731-6741.[Abstract/Free Full Text]
    13. 13
  13. Dudley, J., and R. Risser. 1984. Amplification and novel locations of endogenous mouse mammary tumor virus genomes in mouse T-cell lymphomas. J. Virol. 49:92-101.[Abstract/Free Full Text]
    14. 14
  14. Felsenstein, J. 1989. PHYLIP - phylogeny inference package (version 3.2). Cladistics 5:164-166.
    15. 15
  15. Gallahan, D., and R. Callahan. 1987. Mammary tumorigenesis in feral mice: identification of a new int locus in mouse mammary tumor virus (Czech II)-induced mammary tumors. J. Virol. 61:66-74.[Abstract/Free Full Text]
    16. 16
  16. Gray, D. A., C. M. McGrath, R. F. Jones, and V. L. Morris. 1986. A common mouse mammary tumor virus integration site in chemically induced precancerous mammary hyperplasias. Virology 148:360-368.[CrossRef][Medline]
    17. 17
  17. Hematti, P., B. K. Hong, C. Ferguson, R. Adler, H. Hanawa, S. Sellers, I. E. Holt, C. E. Eckfeldt, Y. Sharma, M. Schmidt, C. von Kalle, D. A. Persons, E. M. Billings, C. M. Verfaillie, A. W. Nienhuis, T. G. Wolfsberg, C. E. Dunbar, and B. Calmels. 2004. Distinct genomic integration of MLV and SIV vectors in primate hematopoietic stem and progenitor cells. PLoS Biol. 2:e423.[CrossRef][Medline]
    18. 18
  18. Indik, S., W. H. Gunzburg, P. Kulich, B. Salmons, and F. Rouault. 2007. Rapid spread of mouse mammary tumor virus in cultured human breast cells. Retrovirology 4:73.[CrossRef][Medline]
    19. 19
  19. Indik, S., W. H. Gunzburg, B. Salmons, and F. Rouault. 2005. Mouse mammary tumor virus infects human cells. Cancer Res. 65:6651-6659.[Abstract/Free Full Text]
    20. 20
  20. Jacque, J. M., and M. Stevenson. 2006. The inner-nuclear-envelope protein emerin regulates HIV-1 infectivity. Nature 441:641-645.[CrossRef][Medline]
    21. 21
  21. Lee, M. S., and R. Craigie. 1994. Protection of retroviral DNA from autointegration: involvement of a cellular factor. Proc. Natl. Acad. Sci. USA 91:9823-9827.[Abstract/Free Full Text]
    22. 22
  22. Lewinski, M. K., M. Yamashita, M. Emerman, A. Ciuffi, H. Marshall, G. Crawford, F. Collins, P. Shinn, J. Leipzig, S. Hannenhalli, C. C. Berry, J. R. Ecker, and F. D. Bushman. 2006. Retroviral DNA integration: viral and cellular determinants of target-site selection. PLoS Pathogens 2:e60.[CrossRef][Medline]
    23. 23
  23. Llano, M., D. T. Saenz, A. Meehan, P. Wongthida, M. Peretz, W. H. Walker, W. Teo, and E. M. Poeschla. 2006. An essential role for LEDGF/p75 in HIV integration. Science 314:461-464.[Abstract/Free Full Text]
    24. 24
  24. Llano, M., M. Vanegas, O. Fregoso, D. Saenz, S. Chung, M. Peretz, and E. M. Poeschla. 2004. LEDGF/p75 determines cellular trafficking of diverse lentiviral but not murine oncoretroviral integrase proteins and is a component of functional lentiviral preintegration complexes. J. Virol. 78:9524-9537.[Abstract/Free Full Text]
    25. 25
  25. MacNeil, A., J. L. Sankale, S. T. Meloni, A. D. Sarr, S. Mboup, and P. Kanki. 2006. Genomic sites of human immunodeficiency virus type 2 (HIV-2) integration: similarities to HIV-1 in vitro and possible differences in vivo. J. Virol. 80:7316-7321.[Abstract/Free Full Text]
    26. 26
  26. Majors, J. E., and H. E. Varmus. 1981. Nucleotide sequences at host-proviral junctions for mouse mammary tumour virus. Nature 289:253-258.[CrossRef][Medline]
    27. 27
  27. Marchetti, A., F. Buttitta, S. Miyazaki, D. Gallahan, G. H. Smith, and R. Callahan. 1995. Int-6, a highly conserved, widely expressed gene, is mutated by mouse mammary tumor virus in mammary preneoplasia. J. Virol. 69:1932-1938.[Abstract]
    28. 28
  28. Michalides, R., E. Wagenaar, J. Hilkens, J. Hilgers, B. Groner, and N. E. Hynes. 1982. Acquisition of proviral DNA of mouse mammary tumor virus in thymic leukemia cells from GR mice. J. Virol. 43:819-829.[Abstract/Free Full Text]
    29. 29
  29. Mitchell, R. S., B. F. Beitzel, A. R. Schroder, P. Shinn, H. Chen, C. C. Berry, J. R. Ecker, and F. D. Bushman. 2004. Retroviral DNA integration: ASLV, HIV, and MLV show distinct target site preferences. PLoS Biol. 2:E234.[CrossRef][Medline]
    30. 30
  30. Narezkina, A., K. D. Taganov, S. Litwin, R. Stoyanova, J. Hayashi, C. Seeger, A. M. Skalka, and R. A. Katz. 2004. Genome-wide analyses of avian sarcoma virus integration sites. J. Virol. 78:11656-11663.[Abstract/Free Full Text]
    31. 31
  31. Nusse, R., and H. E. Varmus. 1982. Many tumors induced by the mouse mammary tumor virus contain a provirus integrated in the same region of the host genome. Cell 31:99-109.[CrossRef][Medline]
    32. 32
  32. Peters, G., S. Brookes, R. Smith, and C. Dickson. 1983. Tumorigenesis by mouse mammary tumor virus: evidence for a common region for provirus integration in mammary tumors. Cell 33:369-377.[CrossRef][Medline]
    33. 33
  33. Peters, G., C. Kozak, and C. Dickson. 1984. Mouse mammary tumor virus integration regions int-1 and int-2 map on different mouse chromosomes. Mol. Cell. Biol. 4:375-378.[Abstract/Free Full Text]
    34. 34
  34. Roelink, H., E. Wagenaar, and R. Nusse. 1992. Amplification and proviral activation of several Wnt genes during progression and clonal variation of mouse mammary tumors. Oncogene 7:487-492.[Medline]
    35. 35
  35. Schroder, A. R., P. Shinn, H. Chen, C. Berry, J. R. Ecker, and F. Bushman. 2002. HIV-1 integration in the human genome favors active genes and local hotspots. Cell 110:521-529.[CrossRef][Medline]
    36. 36
  36. Schuermann, M., and R. Michalides. 1987. 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.[CrossRef][Medline]
    37. 37
  37. Segura-Totten, M., and K. L. Wilson. 2004. BAF: roles in chromatin, nuclear structure and retrovirus integration. Trends Cell Biol. 14:261-266.[CrossRef][Medline]
    38. 38
  38. Shackleford, G. M., C. A. MacArthur, H. C. Kwan, and H. E. Varmus. 1993. Mouse mammary tumor virus infection accelerates mammary carcinogenesis in Wnt-1 transgenic mice by insertional activation of int-2/Fgf-3 and hst/Fgf-4. Proc. Natl. Acad. Sci. USA 90:740-744.[Abstract/Free Full Text]
    39. 39
  39. Suzuki, Y., H. Yang, and R. Craigie. 2004. LAP2alpha and BAF collaborate to organize the Moloney murine leukemia virus preintegration complex. EMBO J. 23:4670-4678.[CrossRef][Medline]
    40. 40
  40. Theodorou, V., M. Boer, B. Weigelt, J. Jonkers, M. van der Valk, and J. Hilkens. 2004. Fgf10 is an oncogene activated by MMTV insertional mutagenesis in mouse mammary tumors and overexpressed in a subset of human breast carcinomas. Oncogene 23:6047-6055.[CrossRef][Medline]
    41. 41
  41. Theodorou, V., M. A. Kimm, M. Boer, L. Wessels, W. Theelen, J. Jonkers, and J. Hilkens. 2007. MMTV insertional mutagenesis identifies genes, gene families and pathways involved in mammary cancer. Nat. Genet. 39:759-769.[CrossRef][Medline]
    42. 42
  42. Thompson, J. D., D. G. Higgins, and T. J. Gibson. 1994. CLUSTAL W: improving the sensitivity of progressive multiple sequence alignment through sequence weighting, position-specific gap penalties and weight matrix choice. Nucleic Acids Res. 22:4673-4680.[Abstract/Free Full Text]
    43. 43
  43. Trobridge, G. D., D. G. Miller, M. A. Jacobs, J. M. Allen, H. P. Kiem, R. Kaul, and D. W. Russell. 2006. Foamy virus vector integration sites in normal human cells. Proc. Natl. Acad. Sci. USA 103:1498-1503.[Abstract/Free Full Text]
    44. 44
  44. Wu, X., Y. Li, B. Crise, and S. M. Burgess. 2003. Transcription start regions in the human genome are favored targets for MLV integration. Science 300:1749-1751.[Abstract/Free Full Text]

Journal of Virology, February 2008, p. 1360-1367, Vol. 82, No. 3
0022-538X/08/$08.00+0     doi:10.1128/JVI.02098-07
Copyright © 2008, American Society for Microbiology. All Rights Reserved.



This article has been cited by other articles:

  • Brady, T., Lee, Y. N., Ronen, K., Malani, N., Berry, C. C., Bieniasz, P. D., Bushman, F. D. (2009). Integration target site selection by a resurrected human endogenous retrovirus. Genes Dev. 23: 633-642 [Abstract] [Full Text]  
  • Shun, M.-C., Botbol, Y., Li, X., Di Nunzio, F., Daigle, J. E., Yan, N., Lieberman, J., Lavigne, M., Engelman, A. (2008). Identification and Characterization of PWWP Domain Residues Critical for LEDGF/p75 Chromatin Binding and Human Immunodeficiency Virus Type 1 Infectivity. J. Virol. 82: 11555-11567 [Abstract] [Full Text]  
  • Kim, S., Kim, N., Dong, B., Boren, D., Lee, S. A., Das Gupta, J., Gaughan, C., Klein, E. A., Lee, C., Silverman, R. H., Chow, S. A. (2008). Integration Site Preference of Xenotropic Murine Leukemia Virus-Related Virus, a New Human Retrovirus Associated with Prostate Cancer. J. Virol. 82: 9964-9977 [Abstract] [Full Text]  

This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Supplemental material
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrowReprints and Permissions
Right arrow Copyright Information
Right arrow Books from ASM Press
Right arrow MicrobeWorld
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Faschinger, A.
Right arrow Articles by Indik, S.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Faschinger, A.
Right arrow Articles by Indik, S.

Copyright © 2009 by the American Society for Microbiology.   For an alternate route to Journals.ASM.org, visit: http://intl-journals.asm.org | More Info»