TEXT

Top Abstract Text References

Feline leukemia virus (FeLV) produces acute leukemia and lymphoma in domestic cats and can be transmitted horizontally from animal to animal as an infectious disease. FeLVs, like murine leukemia viruses (MuLVs), do not contain an oncogene, and their precise molecular mechanism of tumorigenesis is not fully understood. It is well established, however, that the U3 region of the leukemia virus long terminal repeat (LTR), which contains binding sites for various transcription factors, plays a key role in their disease pathogenesis (6, 11, 22, 36). Because of selective expression of different transcription factors in different tissues, these enhancer elements can influence tissue tropism and consequently disease specificity and pathogenic potential (18, 34, 41, 42). Duplication of the U3 enhancer region has been shown to be closely related to leukemogenic potential (24, 30). Other specific sequence motifs downstream of the enhancer sequence, or a tandem triplication of 21 bp in the LTR, have also been implicated in the pathogenesis of FeLV- and MuLV-mediated leukemogenesis (1, 19, 37, 40).

Previous studies from our laboratory have shown that the U3 region of the LTR of Moloney murine leukemia virus (Mo-MuLV) can also activate expression of specific cellular genes, such as the major histocompatibility complex class I and T-cell receptor beta (TCR-) genes, the collagenase IV gene (MMP-9), and MCP-1 (12-16, 21, 39). This activation takes place at the level of transcription and is independent of physical location of the effector (U3-LTR sequence) or the responsive genes and is thus a true trans effect rather than due to positional or insertional activation. Dysregulated expression of each of these cellular genes has been documented in various malignancies. We have recently reported that the U3-LTR region of FeLV subgroup A (FeLV-A) exhibits a similar cellular gene transactivational activity (17). Furthermore, we demonstrated that a specific RNA transcript is generated from the U3-LTR region and that the transactivational activity is closely related to the ability of the LTR to generate such a transcript (7, 8, 17). Although we have not fully elucidated the molecular pathways underlying the transactivation of cellular genes by the LTR-specific transcript or the importance of this activation property in leukemogenesis, our findings suggested that the cellular gene activation property of the LTR could play an important role in FeLV- or Mo-MuLV-mediated tumorigenesis.

All eukaryotic genomes, including the human genome, contain several families of retrovirus-like elements (23, 26, 32, 38). These elements, bordered by two LTRs, contain two or three basic retroviral genes, gag and pol (and sometimes env). These elements may have been introduced by infection of germ line cells by exogenous retroviruses and subsequent vertical transmission. The presence of these endogenous sequences may have detrimental effects on the cell, such as expression of viral transcripts and insertional activation or inactivation of important host genes. Conversely, beneficial effects conferred by these endogenous proviral sequences may include host resistance to exogenous virus infection by receptor blockade. Sequence analysis of these endogenous retroviruses indicates that they are frequently riddled with a variety of mutations (4, 9). Genomic DNA from uninfected cats possesses about 8 to 12 copies of endogenous FeLV-related sequences per haploid genome (2, 5, 27-29). Although they are not inducible as infectious virus particles, they can express subgenomic transcripts in a tissue-specific manner. Molecular cloning of these endogenous viral elements from specific-pathogen-free cats and analysis of their nucleotide sequences demonstrated that they are flanked by LTR sequences (34, 35). In this study, we determined whether these endogenous FeLV LTRs can augment cellular gene expression like their exogenous FeLV counterparts and if these LTRs can generate LTR-specific RNA transcripts.

Since the normal feline genome contains endogenous FeLV-related sequences, we asked whether these sequences also possess similar transactivational activity toward specific cellular genes. We tested two endogenous FeLV proviral clones, CFE-6 and CFE-16, obtained from pathogen-free feline placental DNA (35), for this purpose. One of these two clones, CFE-6, possesses all of the viral structural genes (gag, pol, and env) and LTRs with sizes very similar to those of complete exogenous viruses. Clone CFE-16, in contrast, has an approximately 4.0-kb deletion which truncates both pol and env genes. Neither of these two clones, however, generates any infectious virus particles upon transfection (35). We tested these two clones for the ability to transactivate the collagenase IV gene. BALB/3T3 cells were cotransfected with these clones along with a reporter plasmid wherein the chloramphenicol acetyltransferase (CAT) gene was placed under the control of the collagenase IV gene promoter sequences from 517 to +62 (517/+62 Coll-CAT). We have shown previously that the U3-LTR region from both FeLV-A and Mo-MuLV can strongly activate this collagenase gene promoter (13, 17). To avoid interference from endogenous FeLV sequences present in the genome of the transfected cells, we did not use cells of feline origin in this assay. As shown in Fig. 1, the CAT activity induced by the endogenous FeLV clones CFE-6 and CFE-16 (1.5 and 1.8 times control, respectively) was not significantly different from that generated by the control backbone vector plasmid pTZ19U alone. In contrast, the exogenous full-length FeLV-A clone p61E and its LTR clone p61E-LTR activated reporter expression significantly, up to 4.5- and 12-fold, respectively, as reported previously (17). These data thus demonstrate that endogenous FeLV sequences do not possess transactivational activity toward the 517/+62 collagenase gene promoter in BALB/3T3 cells. In separate experiments, we found that these two endogenous FeLVs, in contrast to their exogenous counterparts (17), do not transactivate the MCP-1 or major histocompatibility complex class I gene promoters either (data not shown).

View larger version (50K): [in this window] [in a new window]   FIG. 1.   Transcriptional activation of a collagenase IV gene promoter reporter by endogenous FeLV proviral clones. One microgram of the 517/+62 Coll-CAT reporter plasmid was cotransfected with 10 µg of clone CFE-6 or CFE-16 into BALB/3T3 cells by the DEAE-dextran method. Exogenous full-length FeLV-A clone 61E (10 µg) and its LTR subclone 61E-LTR (7.5 µg) were used as positive controls. Cotransfection with 7.5 µg of backbone vector plasmid pTZ19U was used to determine the constitutive basal expression of the collagenase IV gene promoter reporter vector. Transfections with clones 61E, CFE-6, and CFE-16 were done in duplicate. Efficiency of transfection was monitored by cotransfection of 1 µg of an expression plasmid for green fluorescent protein for each plate. Forty-eight hours after transfection, cells were washed with phosphate-buffered saline and assessed microscopically for green fluorescence under UV light to normalize transfection efficiency. CAT assay was performed on the cell lysates, and products were separated by thin-layer chromatography. These experiments were repeated three times. The thin-layer chromatogram of one representative experiment is shown. Autoradiographs were photographed by AlphaImager 3.4, and quantitative analysis of the percent conversion (fold activation) for each sample was done by densitometric analysis of the image using the AlphaEase program (Alpha Innotech). Ac-Cam, acetylated chloramphenicol; Cam, chloramphenicol.

We have shown in the case of exogenous FeLVs that the minimum LTR sequences necessary for transactivational activity do not encode a protein product. Instead, this LTR sequence generates a specific RNA transcript, and as in the case of Mo-MuLV, this transcript appears to be related to the transactivational activity of the LTR (8, 17). We therefore wished to determine if endogenous FeLVs can generate an LTR-specific RNA transcript. We used a reverse transcriptase (RT)-PCR-based technique to detect any such transcripts. This technique is based on the fact that all regular retroviral transcripts are terminated at a polyadenylation site within the R region of the 3' LTR. As such, reverse transcription with primers complementary to sequences downstream of this polyadenylation site and subsequent PCR amplification with a primer binding in the U3 region will detect only LTR-specific transcripts, not regular viral transcripts (8, 17). To design specific primers for this purpose, we first compared the FeLV-A (clone 61E) LTR sequences with the two endogenous FeLV LTR sequences used in this study (Fig. 2). As previously noted (3), sequences in the R and U5 regions of these LTRs are highly homologous between exogenous and endogenous FeLVs but are quite divergent in the U3 regions. We used a primer (P4) in the RT reaction with a sequence complementary to the U5 region. When used together with a 5'-PCR primer (P2) in these RT-PCRs, P4 will detect only LTR-specific transcripts from exogenous FeLV-A. In other RT reactions, we also used primer P3, which is complementary to the R region. This primer can reverse transcribe both LTR-specific and regular viral transcripts. Because identical P3 and P4 primer binding sequences are present on LTRs from both endogenous and exogenous FeLVs, these primers were also used in the RT-PCRs to detect LTR-specific RNA transcripts generated by endogenous FeLVs. As we had not yet determined the precise origin of any putative LTR transcript from endogenous LTRs, we chose two different 5'-PCR primers, P5 and P6. These primers were designed from the sequence of an endogenous FeLV (CFE-6) LTR (positions 207 to 187 and 151 to 131, respectively) (Fig. 2). We isolated total cellular RNA from uninfected feline embryo fibroblast cells (line AH927) and performed RT-PCR to determine if endogenous-LTR-specific RNA transcripts are generated. Total cellular RNA from these cells was isolated by lysing actively growing AH927 cells in the presence of a denaturing reagent (4 M guanidine thiocyanate) followed by phenol-chloroform extraction and isopropanol precipitation (17). RNA samples were treated with RQ1 RNase-free DNase (Promega Corporation) at a concentration of 0.1 U/µl at 37°C for 30 min. To demonstrate that AH927 cells indeed contain these endogenous sequences, we used the same primer pairs to PCR amplify the LTRs from the genomic DNA of these cells. At the same time, we also used molecular clones of exogenous FeLV-A (p61E) and endogenous FeLV (CFE-6) to test the specificity of these primers. As shown in Fig. 3A, primer pairs P5-P4 and P6-P4 amplified 331- and 274-bp fragments, respectively, both from AH927 genomic DNA and from a plasmid containing a molecular clone of CFE-6 (lanes 7 to 10). This demonstrates that AH927 cells contain endogenous FeLV LTR sequences similar to CFE-6 in size and sequence. Neither the P5-P4 nor P6-P4 pair, however, amplified any RNA transcripts in RT-PCR using total RNA from uninfected AH927 cells (lanes 3 through 6). RNA from exogenous FeLV-A 61E-infected AH927 cells was reverse transcribed with primer P4 and PCR amplified with primer P2 (lanes 1 and 2). As shown previously, a 350-bp amplified product was detected in this sample (lane 2), but no product was generated when the RT enzyme was omitted from the reaction mixture (lane 1).

View larger version (30K): [in this window] [in a new window]   FIG. 2.   Nucleotide sequence alignment of exogenous FeLV-A LTR (61E) with endogenous FeLV LTRs CFE-6 and CFE-16. Sequence information was obtained from GenBank (accession numbers are M18247, M21479, and M21480, respectively) and from reference 3. Sequence alignment was performed by the MegAlign program available in the sequence analysis program package LASERGENE from DNASTAR, Inc., Madison, Wis. Positions of the RT and PCR primers used in the study are shown in boxes with the primer name above (for 61E) or below (for CFE-6). Complete information on these primers is available in Table 1.

View larger version (44K): [in this window] [in a new window]   FIG. 3.   RT-PCR analysis of the cellular RNA transcripts from uninfected and FeLV-A-infected AH927 cells. Total cellular RNA was isolated from actively growing AH927 cells by guanidine thiocyanate extraction followed by phenol-chloroform extraction as described elsewhere (17). Genomic DNA from AH927 cells was isolated by the sodium dodecyl sulfate-proteinase K digestion method (31). PCR products were separated on 2% agarose gels. PstI-cut lambda DNA was used as molecular weight markers (M). The migration positions and sizes of the amplified products are indicated. To demonstrate specificity of the primers used, simple PCR analysis of genomic or plasmid DNA was also carried out and analyzed in the same gel. (A) Analysis with oligonucleotide P4 as the RT- and 3'-PCR primer. Lanes 1, 3, and 5, no RT; lanes 1 and 2, RNA from FeLV-A 61E-infected AH927 cells; lanes 3 to 6, RNA from uninfected AH927 cells; lane 11, control (Cont) PCR with P4 and P2 primers with no template added. (B) Analysis with oligonucleotide P3 as RT- and 3'-PCR primer. Lanes 1, and 3, no RT; lanes 1 to 4, RNA from uninfected AH927 cells. (C) Analysis of total RNA from FeLV-A 61E-infected AH927 cells. Lanes 1 and 3, no RT; lanes 1 to 6, RNA from FeLV-A 61E-infected AH927 cells. Exogenous FeLV-A plasmid clones 61E-LTR (lanes 7 and 10) and endogenous FeLV plasmid clone CFE6-LTR (lanes 8 and 11) were used as PCR controls. Lane 9 is a control PCR with P3 and P6 primers with no template added. (D) Schematic diagram of the FeLV LTR and locations of primers used for RT-PCR. T, TATA box; A, polyadenylation site.

Similar results were obtained when the other RT primer, P3 (which should detect both viral and LTR-specific transcripts), was used (Fig. 3B). Primer pairs P5-P3 and P6-P3 amplified 228- and 171-bp fragments, respectively, both from AH927 genomic DNA and from a plasmid containing the molecular clone CFE-6 (lanes 5 to 8). The failure to generate a PCR product in reactions using primer P5 or P6 and exogenous FeLV-A molecular clone p61E as the template (lanes 9 and 10) demonstrated the specificity of these two primers for endogenous FeLV LTR sequences. Occasionally we detected a faint band with the P4-P5 (Fig. 3A, lane 4) or P3-P5 (Fig. 3B, lane 2) primer pair, but bands were never detected using a more internal primer (P6) in any RT-PCR (Fig. 3A, lanes 6; Fig. 3B, lane 4). It is therefore highly unlikely that these occasional faint bands were generated from an endogenous LTR transcript. These data thus suggest that LTR transcripts are not made by endogenous FeLV sequences in AH927 cells and that there is negligible production of normal viral transcripts from them.

Although expression from endogenous FeLV elements has been reported in placenta, fetal lymphoid tissues, and some FeLV-negative lymphomas, their level of expression is low (5, 25, 28). In contrast, substantial levels of expression from endogenous FeLV sequences have been detected from certain FeLV-positive tumor cell lines, such as F422 and FL-74 (25). We therefore wished to determine whether infection of AH927 cells with exogenous FeLV could induce expression of the endogenous FeLV. We analyzed total RNA extracted from FeLV-A 61E-infected AH927 cells for the presence of endogenous FeLV LTR-specific RNA transcripts. The FeLV-infected cell line was generated by transfecting AH927 cells with exogenous FeLV-A 61E. After three passages, FeLV core protein antigen p27 production was detected with a Viracheck enzyme immunoassay kit (Synbiotics, Inc., San Diego, Calif.), confirming the infection. As shown in Fig. 3C and previously (17), primer pairs P3-P2 and P4-P2 amplified 247- and 350-bp exogenous FeLV-specific fragments, respectively (lanes 2 and 5), from this FeLV-A 61E-infected AH927 RNA. However, no endogenous FeLV LTR-specific fragment was amplified when the P3-P6 or P4-P6 primer pair was used (lanes 4 and 6). In control PCRs, primer pair P3-P2 amplified a 247-bp fragment from the exogenous FeLV-A plasmid clone 61E-LTR (lane 7), as well as a 269-bp fragment from the endogenous FeLV plasmid clone CFE6-LTR (lane 8). The P3-P6 primer pair amplified a 178-bp product from pCFE6-LTR (lane 11) but generated no p61E-LTR (lane 10). These results thus show that endogenous FeLV-specific LTR transcripts are not generated in AH927 cells, even when they are infected with exogenous FeLVs.

It has been reported previously that levels of expression from some endogenous FeLVs may be determined by cis-acting elements present in the upstream cellular sequences (3). The two endogenous FeLV clones that we used in our cellular gene transactivational activity studies (CFE-6 and CFE-16) contain cell-derived sequences adjacent to the LTRs. It was formally possible that the inability of these clones to cause gene transactivation was due to the presence of those cell-derived flanking sequences. We therefore isolated the LTRs from the endogenous FeLVs and assessed their gene transactivational activity potential. We first identified the minimum region of the exogenous FeLV-A LTR that was sufficient to confer transactivational activity by generating eight 5'- and/or 3'-end deletion constructs from the exogenous FeLV LTR subclone p61E-LTR. Various regions of the LTR were amplified by PCR using specific primer pairs and subcloned into the pGEM3Z vector (Promega). The nucleotide sequences of these clones were all verified before they were used in any experiment. The transactivational activities of these smaller LTR constructs were then analyzed in cotransfection experiments with the 517/+62 Coll-CAT reporter plasmid in BALB/3T3 cells. One of these clones, 61E-H, which includes FeLV-A LTR sequence from 248 to 39, was found to be the smallest region sufficient for wild-type-level collagenase gene promoter transactivation (Fig. 4A and B) (detailed analysis of these clones will be reported elsewhere). Primers M5 and M7 (Table 1) were used to generate clone 61E-H.

View larger version (32K): [in this window] [in a new window]   FIG. 4.   Analysis of transactivational activity and LTR-specific RNA transcript production by the endogenous LTR clones. (A) Schematic diagram of the exogenous (Exo) and endogenous (Endo) FeLV LTR clones. Specific LTR fragments were PCR amplified using primer pairs indicated in the text and cloned into the pGEM3 vector. (B) Transactivational activity of the LTR clones. Each LTR-containing plasmid (7.5 µg) was cotransfected with 1 µg of 517/+62 Coll-CAT reporter plasmid into BALB/3T3 cells. CAT activities in these cells were analyzed 48 h later as described for Fig. 1. Cotransfection of pGEM3 and 517/+62 Coll-CAT was used to determine basal expression of the reporter. This assay was performed three times with similar results. One representative chromatogram is shown. (C) RT-PCR analysis of LTR-specific RNA transcript production by the LTR clones. Individual clones were transfected in BALB/3T3 cells as described above; 48 h later, total RNA was extracted from the transfected cells. RNA samples were DNase treated prior to RT-PCR analysis as described for Fig. 2. RT- and 3'-PCR primers were P3 for 61E-LTR and CFE6-LTR, P19 for 61E-H, P20 for CFE6-3, and P4 for 61E. The 5'-PCR primer for LTR clones of exogenous origin (61E, 61E-LTR, and 61E-H) was P2; the 5'-PCR primer for LTR clones of endogenous origin (CFE6-LTR and CFE6-3) was P5. PstI-digested lambda DNA was used as molecular weight markers (M). Sizes of the amplified products are indicated.

Once we had identified the smallest region of exogenous FeLV LTR necessary for gene transactivation, we constructed analogous, similarly sized LTR clones from the full-length endogenous FeLV clone CFE-6. A 362-bp endogenous FeLV CFE-6 LTR subclone (CFE6-LTR) was constructed by PCR, using primer pair P7-P8 (Fig. 2 and 4A). This clone contains the same LTR region as the exogenous FeLV clone 61E-LTR. A second, smaller 227-bp endogenous LTR subclone containing the LTR region similar to clone 61E-H (CFE6-3) was also generated by PCR, using primer pair M5c-M7c (Fig. 2 and 4A). Following verification of their nucleotide sequences, clones CFE6-LTR and CFE6-3 were tested for the ability to transactivate the collagenase gene promoter in cotransfection assays. Neither of these clones could transactivate the 517/+62 Coll-CAT reporter in cotransfection assays (Fig. 4B). These results suggest that the inability of the endogenous FeLV LTRs to transactivate the collagenase gene promoter in our experiments is not due the presence of adjacent cellular sequences.

Since we have reported previously that feline and murine LTRs with transactivational activities also generate LTR-specific RNA transcripts and that these transcripts are required for transactivational activity (8, 17), we next determined if endogenous FeLV LTRs generate such a transcript. We transfected the endogenous FeLV LTR clones CFE6-LTR and CFE6-3, as well as the exogenous FeLV-A LTR clones 61E-LTR and 61E-H, into BALB/3T3 cells. Forty-eight hours later, RNA was isolated from these transfected cells by the guanidine isothiocyanate extraction followed by DNase treatment. The presence of LTR-specific RNA transcripts was ascertained by RT-PCR using primers specific for the individual clones tested. The RT primers used for the smaller LTR clones 61E-H and CFE6-3 were P19 and P20, respectively (Table 1 and Fig. 2); for the other two clones (61E-LTR and CFE6-LTR), the P3 RT primer was used. In PCRs, P2 and P5 were used as 5' primers for exogenous and endogenous clones, respectively. As shown in Fig. 4C, LTR-specific transcripts were detected from both of the LTRs of exogenous FeLV origin (247 bp for 61E-LTR and 170 bp for 61E-H). However, no amplified products of the appropriate size were generated by either of the endogenous FeLV LTR clones (228 bp for CFE6-LTR and 141 bp for CFE6-3). RT-PCR analysis of an RNA preparation from FeLV-A 61E-infected AH927 cells with primers P4 and P2 generated a 350-bp amplified product, as shown previously (17). These data thus demonstrate that endogenous FeLV LTRs do not produce RNA transcripts and further suggest a relationship between generation of LTR-specific RNA transcripts and gene transactivational activity by the FeLV LTR.

This study investigated the ability of endogenous FeLV LTRs to activate cellular gene expression. The LTRs of leukemia viruses play a central role in disease pathogenesis. They regulate transcription from the viral genome through the enhancer activity of the U3 region, enhancer multiplication, and the stem-loop structure in the R region (6, 10, 18, 33, 36). Additionally, we have hypothesized that the cellular gene transactivational activity of leukemia virus LTRs could also play an important role in leukemogenesis (13-15, 17). We report here that although exogenous FeLV-A LTRs could transactivate the collagenase IV gene and MCP-1 promoters, two nearly full-length endogenous FeLV clones and their isolated LTRs did not possess similar activity. The U3-LTR sequences from endogenous FeLVs differ substantially from the analogous region in the exogenous FeLV LTRs (Fig. 2). Although endogenous FeLV LTRs can function as enhancers in transient reporter assays (3), their inability to activate cellular gene expression in trans is likely the result of this sequence divergence. In parallel pilot experiments studying murine retroviruses, we have found that the LTRs of nonleukemogenic ecotropic MuLVs (gift from Arifa Khan) (20) also fail to demonstrate transactivational activity of cellular genes, in contrast to the strong transactivational activity of LTRs from exogenous, leukemogenic MuLVs. As observed with the FeLVs, the nucleotide sequences of these endogenous viral LTRs differed in the U3 region from the LTRs of leukemogenic MuLV.

Alteration in the LTR enhancer sequence and/or the presence of negative cis-acting control elements in adjacent cellular sequences have been implicated in the repression of transcription from endogenous FeLV sequences (3). In our transient transfection assays, we used a nearly full-length FeLV clone (CFE-6), as well as an endogenous FeLV clone containing a large deletion (CFE-16). Previous studies using transient transfection reported that transcription from CFE-16, but not from CFE-6, could be demonstrated (3). In the case of CFE-6, the lack of expression was shown to be due the presence of inhibitory cellular sequences upstream of the LTR. In our assays, however, neither of these endogenous viruses could transactivate the collagenase gene promoter (Fig. 1). To eliminate the possibility that adjacent cellular sequences included in the proviral clones might have influenced their activity, we studied isolated regions of the LTR alone. The smaller endogenous FeLV LTR constructs, CFE6-LTR and CFE6-3, used in our study represent equivalent regions to a region of the exogenous FeLV LTR that is known to have transactivational activity. Since CFE6-LTR and CFE6-3 clones had no cellular sequences associated with them and still failed to demonstrate transactivational activity, the possibility of a cis-acting negative regulatory effect by adjacent cellular sequences was ruled out.

Although we have demonstrated that both Mo-MuLV and FeLV LTRs can transactivate cellular gene expression, the mechanism underlying this activity is not completely understood. We have previously demonstrated that these viruses produce LTR-specific RNA transcripts and that the generation of these transcripts is necessary for transactivational activity (7, 8, 17). In the present study, we were unable to detect any endogenous LTR-specific transcripts in the feline embryo fibroblast cell line AH927. Genomic DNA from these cells, however, did contain FeLV sequences similar to the endogenous feline proviral clone CFE-6 (Fig. 3). No regular viral transcripts were detected in AH927 cells either, suggesting that transcription from these naturally occurring endogenous FeLV sequences does not normally take place in AH927 cells. This finding is consistent with a recent report that endogenous FeLV-specific expression was not detectable in this cell line (25). In contrast, the same investigators found that certain FeLV-positive tumor cell lines, such as F422 and FL-74, did express their endogenous FeLV elements. Infection of AH927 cells with FeLV-A, however, did not induce expression from the endogenous FeLV sequences in our studies. We do not yet know whether the lack of inducible expression of the endogenous FeLV elements in AH927 cells is specific for this cell line. It may be informative to examine endogenous virus LTR-mediated gene transactivation in other feline cell lines, such as A3201 and FL74, where endogenous FeLV expression can reportedly be induced (25).

Earlier studies have shown that transient transfection of certain endogenous FeLV clones into murine fibroblasts such as NIH 3T3 cells could result in transcription of viral genes (3, 35). We therefore attempted to demonstrate LTR-specific RNA transcript production by endogenous FeLVs in murine fibroblasts (BALB/3T3), to address the potential concern that failure to detect such transcripts in AH927 cells was due to cell line specificity. Production of LTR-specific RNA transcripts by endogenous viral LTRs did not occur in BALB/3T3 cells, whereas LTRs from exogenous FeLVs did generate such transcripts in this cell line. Consistent with our hypothesis that the LTR transcript mediates transactivation of cellular genes, LTR clones from endogenous FeLVs exhibited no transactivational activity when they were transfected in BALB/3T3 cells, in contrast to the significant transactivational activity exhibited by LTRs from exogenous FeLV clones. Identical results were obtained when these experiments were repeated in NIH 3T3 cells (data not shown). Together, our findings suggest that the failure of the endogenous FeLV LTRs to induce gene transactivation in our study is not cell line specific.

Although we have observed a consistent direct relationship between LTR-specific RNA transcript production and cellular gene transactivation in our experiments, we have not yet demonstrated a causal relationship between these activities for the FeLVs. Construction of chimeric exogenous-endogenous FeLV LTRs is under way to more precisely define LTR sequence requirements for the transactivation function and transcript production. The data from such experiments will be critical in designing transactivation-deficient mutant proviruses for evaluation of the biological relevance of LTR-mediated host cell gene activation in tumorigenesis.