| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Advances in Brief |
Molecular Oncology Laboratory, Institute of Comparative Medicine, University of Glasgow Veterinary School, Bearsden, Glasgow, Scotland G61 1QH [J. B., A. T., J. M., E. G., E. C., J. C. N.], and Laboratory of Genomic Diversity, National Cancer Institute, Frederick, Maryland 21702 [S. C., S. J. O.]
 |
ABSTRACT |
|---|
 Top ABSTRACT IntroductionMaterials and MethodsResults and DiscussionREFERENCES |
|---|
|
Introduction |
|---|
|
TopABSTRACTIntroductionMaterials and MethodsResults and DiscussionREFERENCES |
|---|
, ß, and 
retroviruses. Tumors associated with HIV and related lentiviruses have generally been found to lack viral sequences, suggesting that the tumor predisposition conferred by this subfamily of retroviruses is because of indirect mechanisms. Most HIV-1-associated lymphomas are thought to occur secondary to polyclonal B-cell activation and infection with oncogenic viral cofactors such as EBV or HHV-8 (3)
. Similarly, longitudinal monitoring of cats infected with FIV5
has documented impaired cell-mediated immunity and B-cell hyperactivity as factors that may contribute indirectly to lymphomagenesis (4)
, whereas analyses of a series of FIV-associated lymphomas showed that most of these lacked detectable viral sequences (5)
. Despite these generally negative observations, a number of studies have suggested that there may be exceptions to this rule. HIV sequences have been detected in a number of tumors using PCR methods (6
, 7)
. Moreover, monoclonal integration of viral sequences in the tumor DNA was demonstrated for one case of HIV-associated T-cell lymphoma where viral integration was mapped to the FUR gene and was associated with overexpression of the nearby FPS/FES proto-oncogene (8)
. We reported a potentially similar observation in a case of B-cell lymphoma, which arose in a cat infected with FIV, where the tumor cells harbored a single integrated proviral copy detected by Southern blot analysis (9)
. Additional analysis reported here suggests that in this case the consequence of this insertion event may be loss of function rather than transcriptional activation of a host gene. Furthermore, analysis of the human orthologue reveals a previously uncharacterized gene, which maps to a site of frequent deletions and loss of heterozygosity in a number of common human cancers. |
Materials and Methods |
|---|
|
TopABSTRACTIntroductionMaterials and MethodsResults and DiscussionREFERENCES |
|---|
DASH II arms (Stratagene). After ligation and packaging with Gigapack III Gold extract (Stratagene), 
0.5 x 106 plaques were screened with a radiolabelled probe generated by random priming of a 1.2 kb FIV Gag-pol fragment (1242–2431; Ref. 10
). After plaque purification of positive recombinants, phage DNA was prepared and used for analysis by restriction mapping and Southern blot hybridization. A representative clone was selected for subcloning of the EcoRI insert into plasmid pBR328 for additional analysis, generating the plasmid designated pEAG6.
Sequence Analysis.
DNA sequence was determined using a Thermosequenase primer cycle sequencing kit (Amersham Pharmacia Biotech) with the aid of infrared dye-labeled primers (MWG, Germany) and a LiCor 4000 automated sequencer.
Southern Blot Hybridization.
Twenty-µg samples of DNA were digested with EcoRI, fragments separated on a 0.8% agarose TAE gel and transferred onto Hybond N membrane (Amersham Pharmacia Biotech) in 20x SSC. The filter was hybridized with a 94-bp exon 3 probe (PCR product of primers C 5'-GATGAATTTTCAGGATTGTC-3' and D 5'-ACCGCAGACAACTGAAGAG-3', fE3, Fig. 1a
) in RapidHyb (Amersham Pharmacia Biotech) at 65°C, washed 3 x 20 min 0.5x SSC, 0.5% SDS, 60°C, and exposed to X-ray film.
|
; A: 5'-AGACATAGAAATCTCAAGCAG-3'; B: 5'-ACTTGCTCCTAATTTTACCC-3'), which allowed the amplification of a 250-bp fragment from feline but not hamster or mouse template DNA. One-min steps of denaturation (95°C), annealing (52°C), and DNA synthesis (72°C) were performed using 1 µg of template DNA, 50 pmol of each primer, and Taq polymerase (Promega Corporation).
Northern Blot Analysis.
A human multitissue Northern blot (Clontech) was hybridized in RapidHyb with human EST clone (IMAGE clone 1047426) insert, obtained from the United Kingdom MRC HGMP Resource Centre (accession number AA625226). Hybridization was carried out at 65°C, and washing was performed at high stringency (3 x 20 min 0.1x SSC, 0.5% SDS, 60°C).
RT-PCR.
Five µg of total RNA samples were reverse transcribed using a First Strand cDNA synthesis kit (Amersham Pharmacia Biotech) and NotI-dT primer. Amplification was carried out on aliquots of one-fifth of each sample with 30 pmol of each primer pair. Primers used included F-FLJ12973 exon 3 and 4 (exon 3 5'-AAGGAAGAGACTGAAGAACA-3' and exon 4 5'-TGGATTCAGGAGGCTGTCTC-3'), FIV Gag (966F 5'-GGGATTAGACACTAGGCCATCTA-3', 1837R 5'-GACCAGGTTTTCCACATTTATTA-3'), FIV R/U5 (5'-ATTGAACCCTGTCAAGTATCT-3'), or HPRT. Amplification was performed in 1.1x Reddy Mix (ABgene) at 95°C 30 s, 55°C 30 s, and 72°C 30 s. To control for RNA integrity, HPRT amplification was found to require 29 (C), 27 (T), and 25 (0, 1, 2, and 3) cycles (C, T, 0, 1, 2, 3; see Figure
legends). Therefore, other primer pairs were normalized to 32 (C), 30 (T), and 28 cycles (0, 1, 2, and 3). Five-µl aliquots of each reaction were separated on 4% polyacrylamide gels, and detected either by staining with ethidium bromide (Fig. 1b
; Fig. 3b
), or transferred onto Hybond N membrane in 20x SSC and hybridized in 6x SSC at 65°C with either fE3 3 or HPRT probes (Fig. 3d)
. Washes were 3 x 20 min 0.1x SSC, 0.5% SDS, 60°C (fE3) or 0.5x SSC, 0.5% SDS (HPRT).
|
|
Results and Discussion |
|---|
|
TopABSTRACTIntroductionMaterials and MethodsResults and DiscussionREFERENCES |
|---|
To characterize the inserted FIV provirus additionally, a bacteriophage library was prepared from high molecular weight Q254 tumor DNA, and this was screened with a FIV-specific gag-pol probe. Two positive clones were obtained containing apparently identical EcoRI inserts of 11.5 kb, and one of these was analyzed additionally after subcloning into plasmid pBR328 to generate the plasmid clone designated pEAG6. Restriction enzyme mapping and hybridization analysis indicated the presence of an apparently full-length FIV provirus and 2 kb of host flanking sequence. Sequence analysis confirmed this structure, revealing a full-length FIV genome along with 1770 and 69 bp of cellular flanking sequences from the 5' and 3' ends, respectively (Fig. 1a)
. The proviral sequence from lymphoma Q254 was found to be very closely related to the original infecting strain (FIV-G8), with an env gene match of >98% at the amino acid level (data not shown).
Attempts to recover infectious virus from pEAG6 by a sensitive transfection and coculture method (12)
were unsuccessful, suggesting that the tumor-derived viral genome carried a subtle defect that was not evident from the primary sequence. However, analysis of feline fibroblast cells (CrFK) transfected with pEAG6 showed that FIV structural proteins were expressed at similar levels to a control plasmid carrying an intact infectious molecular clone (data not shown). These observations are interesting in the light of our previous studies, which showed that viral sequences were not detectably expressed in the primary tumor (9)
. Attempts to resolve these observations by analysis of viral transcription in the primary tumor were severely constrained by lack of primary tumor material, but sufficient RNA was extracted for limited Northern blot and RT-PCR analysis. Whereas Northern blots did not reveal any detectable FIV transcripts (data not shown), a normalized RT-PCR analysis of tumor RNA using FIV-specific primers revealed their presence at low levels (Fig. 1b)
, suggesting that the provirus may have been expressed at low levels. However, the possibility that this signal arose from low levels of tumor-infiltrating lymphoid or myeloid cells could not be excluded.
Analysis of a series of feline-rodent hybrid cell lines was undertaken to assign the FIV integration locus (FIVi-1) to a feline chromosome. Primers A and B (Fig. 1, a and c)
were used to amplify a fragment of 250 bp from a panel of 38 cat-rodent hybrids. The lowest level of discordancy was noted with markers on feline chromosome B3, providing an unambiguous assignment to this chromosome (Fig. 1d)
.
Searches for matches to known genes by Basic Local Alignment Search Tool analysis using the flanking sequences revealed a highly significant match to a human gene (FLJ12973/Q9H967) that maps to chromosome 15q15. This location is consistent with the presence of an orthologue on feline chromosome B3, which displays syntenic alignment with human chromosomes 14 and 15, and a short region of chromosome 3 (13)
. The FLJ12973 gene is highly conserved, as shown by its close homology to its murine orthologue on chromosome 2 (Q9CTV4 at band E5) and by matches to ESTs from other mammalian species (data not shown). As shown in Fig. 1c
, feline-human sequence match is evident over the entire exon 3 of FLJ12973, which is coextensive with the feline exon apart from a single codon deletion. Homology extends 55 bp upstream and 9 bp downstream of the canonical splice acceptor and donor signals, which mark the exon boundaries (Fig. 1c)
. These observations unequivocally identify the feline orthologue of FLJ12973 as the gene interrupted by proviral insertion at FIVi-1 in lymphoma Q254.
The structure of the human FLJ12973 gene in Fig. 2a
, which is based on the annotated draft sequence, shows a gene comprising 13 exons and covering 40 kb of the human genome on chromosome 15q15. The first exon occurs within a CpG island, which also marks the boundary with the next 5' gene (MFAP1), which is in the opposite orientation. The relative position of FIV insertion in the orthologous feline gene is also indicated, which is between exons 3 and 4, in the same transcriptional orientation as the cellular gene. The exon structure of cDNAs derived from the human gene are indicated below, including the full-length transcript, which was derived from the NT2 teratocarcinoma cell line (AK023035). An interesting variation emerged from our analysis of an EST clone (IMAGE clone 1047426), which we obtained from the United Kingdom MRC HGMP Resource Centre (AA625226). Complete sequence analysis of this clone revealed a 1126-bp insert, which differed from the full-length transcript by an alternative splice in exon 2 and an internal deletion spanning from the middle of exon 9 to the extreme 3' end of exon 13. The deletion site is marked by a 5-bp homology (CCACC) between these exons (Fig. 2a)
, suggesting that this is not a normal splicing variant. Intriguingly, the same motif is repeated at the site of natural deletion variants of the MECP2 gene, which is responsible for Rett syndrome, a progressive neurological disorder (14)
. As the EST clone came from the Soares library, which was generated from heterogenous RNA sources, the origins of this deletion are unclear, but the possible existence of natural variants of this gene clearly merits additional study.
|
. The longest possible product is a 626 amino acid protein (MSRSG), whereas the alternatively spliced AA625226 cDNA clones suggests an alternative initiation (MCPKS). However, alignment with an EST from the orthologous murine gene (AK020004) shows that neither of these is conserved in the mouse. Use of the first conserved initiator codon would generate a 529 amino acid protein with the NH2-terminal sequence MKNTS. This sequence displays the closest match to the Kozak consensus (15)
. Clearly, additional direct analysis will be required to established the true NH2 terminus of FLJ12973 and its orthologues. The truncated cDNA AA625226 also encodes a frame-shift truncation with respect to the full-length gene product (Fig. 2b)
. Notably, the insertion site of FIV between exons 3 and 4 is 3' with respect to all of these potential translational initiation sites.
Analysis of FLJ12973 for known protein family domains revealed a series of five tandem WD40 repeats in a 250 amino acid COOH-terminal domain (Fig. 2, b and c)
. This protein-protein interaction motif is found in an expanding range of protein families of which the roles include transcriptional regulation, apoptosis induction, and signal transduction (16)
. Alignment with database sequences using the Smith and Waterman algorithm (MPSRCH)6
revealed that the closest human gene homologue is DDB2, the small subunit of a DNA damage-binding protein complex, which is defective in xeroderma pigmentosum patients of complementation group E (17)
. As shown in the diagram in Fig. 2c
, the DDB2 protein is remarkably similar to FLJ12973 in overall structure. It shares five tandem WD40 repeats and can be aligned with the COOH-terminal portion of FLJ12973 over its entire length with very few gaps.
The expression pattern of the human FLJ12973 gene was explored by Northern blot analysis using a multitissue blot (Clontech; Fig. 2d
). As can be seen from this analysis, the 1.2 kb EST probe (IMAGE clone 1047426) detected two transcripts-a major species at around 2.8 kb and a less abundant RNA of 4.5 kb. The smaller transcript is similar in size to that expected from the full-length cDNA sequence (AK023035), and a notable feature is the expression of FLJ12973 in all of the tissues examined. Expression levels varied very little across tissues, although some differences in relative abundance of the two transcripts were noted. Similarly, analyses of murine and feline tissue blots showed ubiquitous and relatively uniform patterns of expression, although in these cases only a single mRNA species was observed, with sizes of 4.5 kb and 3.5 kb, respectively (data not shown).
In the light of the foregoing observations, it was important to determine the effects of proviral insertion on the structure and expression of the feline FLJ12973 gene. The use of an exon 3 (fE3) probe for Southern blot analysis of a range of feline DNA samples (Fig. 3a)
confirmed that the locus was clonally rearranged in lymphoma Q254, which displayed an EcoRI fragment of 11.5 kb (R) corresponding precisely to the size predicted for the cloned provirus and its flanking sequences. A coincident fragment was detected with an FIV-specific probe (Fig. 3a
, right panel). The 2-kb EcoRI fragment corresponding to the unoccupied FIVi-1 site (U) was proportionately reduced in intensity in Q254 tumor DNA (T). Analysis of control tissue (C) from the same cat excluded the possibility of an inherited polymorphism at this locus. These results confirm that the FIV provirus was acquired somatically in lymphoma Q254 and indicate that insertion of this element at FIVi-1 preceded clonal expansion of the lymphoma cells.
From the position of insertion within the gene (Fig. 2a)
, we considered it possible that a truncated or fused mRNA species might be generated. As discussed above, limited availability of primary tumor material constrained expression analysis. A single Northern blot was generated and analyzed with a probespanning exons 3/4 of the feline gene, revealing no detectable signal in Q254 tumor RNA (data not shown). However, whereas this analysis ruled out gross overexpression, the low control hybridization signal meant that no conclusion could be drawn with regard to possible transcriptional down-regulation of the gene. Additional analysis by RT-PCR was undertaken to test for the presence of fusion transcripts initiating in viral sequences. As shown in Fig. 3b
, primers spanning FIV LTR and exon 4 of F-FLJ12973 detected a transcript specifically in RNA derived from Q254 tumor cells, showing that read-through transcripts were indeed present, although presumably at low levels. Sequence analysis confirmed that these represented unspliced transcripts containing 3'LTR and host F-FLJ12973 intron 3 sequences (Fig. 3c)
. To test for the effect of insertion on the normal transcription pattern of F-FLJ12973, RT-PCR was also carried out using primers from exons 3 and 4. As can be seen in Fig. 3d
, this assay readily detected transcripts in a range of feline cell lines (and other tissues; data not shown), with the exception of tumor Q254. The presence of transcripts in control tissue (salivary gland) from the same animal demonstrated that this evident loss of expression was not a constitutive defect.
Taken together, these results suggest that the consequence of FIV insertion in the feline FLJ12973 gene may be loss of function. The status of the unaffected allele in the Q254 tumor cells is of interest in this respect. Southern blot analysis (Fig. 3a)
shows a normal sized EcoRI fragment corresponding to the unaffected allele, showing that at least part of the normal gene was retained in the tumor cell genome. The possibility that a deletion has occurred elsewhere in the gene cannot be excluded, whereas it is also conceivable that epigenetic silencing has occurred by a mechanism such as promoter hypermethylation, which can affect suppressor genes (18)
. In this respect it is interesting to note the close structural relationship of FLJ12973 to DDB2, one of the components of the xeroderma pigmentosum group E complex. DNA repair defects are commonly observed in lymphomas, and this class of gene has also been shown to be subject to epigenetic regulation (19)
. Such an event could conceivably mediate tumor initiation, particularly in the context of a B-cell population under proliferative stress, as seen in FIV infection (20)
. However, other loss-of-function effects potentially relevant to oncogenesis should not be discounted, as WD40 domain protein families execute many regulatory functions (21)
. The possibility of a wider role for FLJ12973 inactivation also merits additional study. Human chromosome 15q is the site of frequent deletions in ovarian (22)
and sporadic colorectal cancers (23)
, whereas the latter tumor type is also associated with an inherited susceptibility gene, CRAC1, which maps to a similar location (24)
. It will clearly be important for future studies to characterize the FLJ12973 gene in greater detail and to examine the biological consequences of functional inactivation.
This study reveals additional evidence of the direct mutagenic capacity of the immunosuppressive lentiviruses and their potential contribution as direct agents in carcinogenesis. The possibility of such a role has generally been ignored in recent analyses of AIDS-related malignancies. Our findings should rekindle interest in this field of study and raise awareness of the potential risks associated with lentivirus-based vectors, which are being developed currently for clinical use. It will be important to monitor the sites of integration of these agents in vivo and to examine their capacity to silence as well as activate host cell genes.
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
1 J. B. was supported by the Wellcome Trust United Kingdom. Other aspects of this work were supported by awards from Cancer Research United Kingdom and the Leukemia Research Fund. 
2 Present address: Sydney Cat Specialists, 70 Roberts St. Camperdown, NSW 2050, Australia.
3 These authors contributed equally to this work.
4 To whom requests for reprints should be addressed, at Molecular Oncology Laboratory, Institute of Comparative Medicine, University of Glasgow Veterinary School, Bearsden, Glasgow, Scotland G61 1QH. Phone: 44-141-330-5771; Fax: 44-141-330-4874; E-mail: j.c.neil{at}vet.gla.ac.uk
5 The abbreviations used are: FIV, feline immunodeficiency virus; EST, expressed sequence tag; RT-PCR, reverse transcription-PCR; HPRT, hypoxanthine phosphoribosyltransferase; LTR, long terminal repeat.
6 Internet address: http://www.ebi.ac.uk.
7 Internet address: http://genome.ucsc.edu.
Received 5/22/02. Accepted 10/31/02.
|
REFERENCES |
|---|
|
TopABSTRACTIntroductionMaterials and MethodsResults and DiscussionREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  R. Natrajan, J. Louhelainen, S. Williams, J. Laye, and M. A. Knowles High-Resolution Deletion Mapping of 15q13.2-q21.1 in Transitional Cell Carcinoma of the Bladder Cancer Res., November 15, 2003; 63(22): 7657 - 7662. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Cancer Research | Clinical Cancer Research |
| Cancer Epidemiology Biomarkers & Prevention | Molecular Cancer Therapeutics |
| Molecular Cancer Research | Cancer Prevention Research |
| Cancer Prevention Journals Portal | Cancer Reviews Online |
| Annual Meeting Education Book | Meeting Abstracts Online |
indicating the LTRs. b, RT-PCR analysis of FIV expression in lymphoma Q254. Primers from the FIV Gag gene (966F, 1837R) were used to screen for FIV specific products. The amplification products were separated on a polyacrylamide gel and detected by staining with ethidium bromide. Samples include Q254 lymphoma (T), salivary gland (C). Control samples include no added RNA (0), feline fibroblast RNA (1 and 3), feline lymphoma F422 (2) and the FL4 cell line (+), which is persistently infected with FIV. c, sequence of the 5' 600bp of EcoRI clone pEAG6, showing the position and sequence of a conserved exon and its match to the FLJ12973 gene on human chromosome 15. The human DNA sequence is depicted only for the region of match. Primer sequences as defined above are underlined. d, assignment of the FIV integration site FIVi-1 to feline chromosome B3. DNA from a series of 38 rodent x cat somatic cell hybrid clones was screened for the presence of FIVi-1 sequences by PCR using primers A and B (a and c). The histogram shows the percentage of discordance between FIVi-1 and the specified feline chromosomes as defined by 28 established marker genes 
