Cloning of sequences expressed specifically in tumors of rat

Use of serial analysis of gene expression (SAGE) technology

Serial analysis of gene expression, or SAGE, is an experimental technique designed to gain a direct and quantitative measure of gene expression. The SAGE method is based on the isolation of unique sequence tags (9-10 bp in length) from individual mRNAs and concatenation of tags serially into long DNA molecules for a lump-sum sequencing. The SAGE method can be applied to the studies exploring virtually any kinds of biological phenomena in which the changes in cellular transcription are responsible. SAGE is a highly competent technology that can not only give a global gene expression profile of a particular type of cell or tissue, but also help us identify a set of specific genes to the cellular conditions by comparing the profiles constructed for a pair of cells that are kept at different conditions. In this review, we present an outline of the original method, several studies achieved by using the method as a major strategic tool, technological difficulties and intrinsic problems that emerged, and improvements and modifications of the method to cope with these drawbacks. We then present our modified SAGE procedure that generates longer sequence tags (14 bp) rather in detail, and the profile (80K profile) derived from HeLa cells that is composed of 80000 tags obtained from a single library. In addition, a series of smaller profiles (2, 4, 10, 20 and 40K) was made by dividing the 80K profile. When we compared these smaller profiles with respect to tag counts for a number of genes, it became apparent that counts of most gene tags increase stably and constantly as the size of profiles increase, while several genes do not. This may be another problem we have to keep in mind, when the profiles are compared for the identification of 'specific genes'.

Characterization of the cDNA and pattern of expression of a new gene over-expressed in human hepatomas and colonic tumors

A cDNA clone of 6.449 kb ch-TOG (for colonic and hepatic tumor over-expressed gene) initially selected from various human libraries and completed by 5' rapid amplification of cDNA ends (RACE) PCR is described. The original cDNA clone was extracted from an expression library constructed from a human tumoral brain. This library was screened with an antibody raised against the cytochrome P450tu that was shown to be over-expressed in chemically induced mouse hepatic tumors. Using this cDNA as a probe, a full-length cDNA was characterized. Its nucleotide sequence shows no significant similarity with any of the gene sequences collected in the various DNA data bases. The translation of the larger open reading frame leads to a putative protein of 1972 amino acids (molecular mass = 218453 Da). Hybridization analyses on Southern blot and on metaphase chromosomes indicate that this gene is present as a single copy in the genome and is localized on the short arm of chromosome 11. ch-TOG transcripts are present in several human tissues. Over-expression of ch-TOG in neoplastic liver and colon compared with the corresponding normal corresponding tissues is demonstrated. The level of the expression of ch-TOG transcripts was also studied in the various differentiation stages of the human colonic adenocarcinoma cell line Caco-2.

Isolation of cDNA clones corresponding to genes differentially expressed in two colon-carcinoma cell lines differing by their tumorigenicity

In an effort to isolate genes involved in the progression of colonic cells leading to a carcinoma, we used as a model 2 rat colon-carcinoma cell lines selected from the same tumor, differing by their tumorigenicity. When soluble, Triton-X-100 extracted, or cytoskeletal proteins from the progressive PROb cells and the regressive REGb cells were analyzed by SDS-PAGE, minor differences were seen. Furthermore, mRNA-cDNA hybridization analyses showed extensive homology between the 2 mRNA populations. Thus, the homology between the 2 clones is high at both the protein and the mRNA levels. A PROb cDNA library was hybridized with 32P-cDNA synthesized from PROb or REGb mRNA. The clones giving a stronger signal when hybridized with the homologous PROb probe were isolated. The specificity of each clone was confirmed by RNA blotting. Most of the positive clones showed a 2- to 3-fold higher expression in PROb cells when compared with REGb cells. One clone (J 13) corresponded to an mRNA 7- to 10-fold more abundant in PROb cells, and was further studied. No gene amplification was detected by Southern blot analysis, indicating that the difference in mRNA content was most likely due to an increased transcription of this gene in PROb cells. Sequencing of the cDNA showed high homology with the rat ferritin light sub-unit. Over-expression of ferritin in PROb cells as compared with REGb cells was confirmed at the protein level using specific antibodies.

Identification of a novel transcription unit in the human insulin-like growth factor-II gene

The human insulin-like growth factor-II (hIGF-II) gene has until now been thought to be composed of eight exons, including three independent leader exons. In the present study two additional exons, one leader exon and one alternatively used ordinate exon, have been newly identified. They were abundantly expressed in human histiocytoma tissue, generating mRNA species of about 5.0 kb in length. The new leader exon shows significant sequence similarity with the rE1 exon, previously reported to be transcribed only in the rat, and is mapped at nearly the same genomic location as in the rat. On the other hand, sequence similarity with another exon in the corresponding region of the rat genome was also found. It was, however, obvious that the rat sequence would not work as an active exon, since both splice acceptor and donor sites were deviated considerably from the consensus sequences. It has thus become apparent that the complex transcription unit of a single-copy hIGF-II gene comprises at least 10 exons, including four leader exons, one alternative exon and three common protein-coding exons.

Expression of insulin-like growth factor II, alpha-fetoprotein and hepatitis B virus transcripts in human primary liver cancer

Insulin-like growth factor II is a fetal growth factor structurally and functionally related to insulin and insulin-like growth factor I. Its mRNA expression is developmentally regulated in human liver, the reexpression of insulin-like growth factor II fetal transcripts being often observed in primary liver cancer. Insulin-like growth factor II and alpha-fetoprotein mRNAs were studied in 16 human primary liver cancers, most of which were highly differentiated. Hepatitis B virus transcripts were also analyzed in the tumors from hepatitis B virus chronic carriers. alpha-Fetoprotein mRNA was detected in only four tumors and in one nontumorous cirrhotic tissue; all these samples also displayed insulin-like growth factor II fetal transcripts. Furthermore, fetal insulin-like growth factor II mRNAs were observed in five tumors and six nontumorous cirrhotic areas not expressing alpha-fetoprotein mRNA. The presence of hepatitis B virus RNA was only observed in tissues not expressing alpha-fetoprotein or fetal insulin-like growth factor II mRNA. In conclusion, fetal insulin-like growth factor II transcripts are more frequently observed than alpha-fetoprotein mRNA in highly differentiated liver cancers and in surrounding cirrhotic areas. The reexpression of fetal insulin-like growth factor II transcripts might then be a marker of early steps of liver cell transformation.

Expression of a mouse replacement histone H3.3 gene with a highly conserved 3' noncoding region during SV40- and polyoma-induced Go to S-phase transition

We have isolated and sequenced a mouse replacement variant histone H3.3 cDNA. It corresponds to the most abundant mRNA expressed from a unique gene by the use of one out of three polyadenylation sites. The 3' non coding region of H3.3 is very long (approximately 1100 nt) and highly conserved throughout evolution since it is about 95% homologous to the 3' non coding region of the chicken H3.3B gene. We studied the expression of the H3.3 gene during SV40- and polyoma-induced mitotic host reaction in confluent, Go-arrested primary mouse kidney cell cultures. H3.3 replacement variant mRNA steady state levels increased during the Go to S-phase transition, apparently as the result of two mechanisms: one related to cell growth, whereas the other was linked to cellular DNA synthesis. The latter mechanism was however far less pronounced than with replication histone variant mRNAs. The biological implications of these results are discussed.

Major urinary protein as a negative tumor marker in mouse hepatocarcinogenesis

A cDNA clone isolated by differential screening of a mouse liver cDNA library using RNA probes from normal adult liver and from hepatocellular tumors showed, on northern blot analysis, a single transcript of 1 kb present at high levels in normal adult mouse liver but at 15- to 100-fold reduced levels in spontaneous and p-dimethylaminoazobenzene- or urethan-induced liver tumors. The transcript, which was absent in two-wk-old mouse liver, is coded by a developmentally regulated gene. Nucleotide sequence analysis demonstrated that the isolated clone encoded a major urinary protein (MUP). Immunocytochemistry analysis, using a goat antibody to MUP on histologic liver sections containing diethylnitrosamine-induced liver nodules, showed that the decrease in MUP levels also occurred in microscopic nodules. Therefore, MUP decrease represents an early event in the development of mouse liver tumors, and it may be used as a tumor marker in mouse hepatocarcinogenesis.

Sequences coding for part of oncogene-induced transin are highly conserved in a related rat gene

The transin gene is induced by oncogenes and epidermal growth factor (EGF). We report here the isolation of a related gene (transin-2 gene). The structures of these genes are very similar. Indeed, a stretch of 428 nucleotides of the transin gene containing both exon and intron sequences is 98% conserved in the transin-2 gene. However, the putative promoter regions of the genes show little sequence homology, apart from a short element related to a sequence involved in control of transcription by cyclic AMP or a tumour promoter. Expression of the transin-2 gene, unlike that of the transin gene, is not induced by EGF, dibutyryl cyclic AMP or cytochalasin D. Nevertheless, transin-2 RNA is expressed in several transformed rat embryo fibroblast cell lines, and can be induced by a tumour promoter. The proteins transin and transin-2 are approximately 71% homologous in sequence. Both proteins show significant sequence homology with two connective tissue degrading metalloproteases. These homologies raise the possibility that expression of transin and transin-2 in transformed cells might play a role in tumour invasion.

Isolation of the oncogene and epidermal growth factor-induced transin gene: complex control in rat fibroblasts

Various oncogenes or epidermal growth factor (EGF) induce transcription of a 1.9-kilobase RNA (transin RNA) in rat fibroblasts. The induction by EGF can be blocked by cycloheximide. Thus the response of the transin gene to EGF appears to require de novo protein synthesis. Transin RNA induction is specific to EGF, as neither insulin, platelet-derived growth factor, fibroblast growth factor, nor transforming growth factor beta could elicit the same response. However, transforming growth factor beta could block the EGF induction of transin RNA. Whereas the calcium ionophore A23187 and the tumor promoter TPA, either alone or administered together, did not increase transin RNA levels, TPA could synergise with a serum factor to effect such an increase. Dibutyryl cyclic AMP also induced transin RNA. Treatment of cells with the microfilament-disrupting agent cytochalasin B, but not the microtubule-disrupting agent colcemid, resulted in an increase in transin RNA levels, suggesting a role for the cytoskeleton in control of transin gene expression. The transin RNA does not contain repeated sequences and appears to be encoded by a single-copy gene. The protein sequence encoded by the last four exons of the transin gene shows some homology to two regions of the heme-binding protein hemopexin.

Cellular events during hepatocarcinogenesis in rats and the question of premalignancy

The cellular, biochemical, and genetic changes that occur in the liver of rats exposed to chemical hepatocarcinogens are reviewed. Multiple new cell types appear in the liver of carcinogen-treated rats including foci, nodules, ducts, oval cells, and atypical hyperplastic areas. The application of phenotypic markers for these cell types suggests that hepatocellular carcinomas may arise from more than one cell type, including a putative liver stem cell that proliferates following carcinogen exposure. Study of DNA, RNA, and proteins produced by hepatocellular carcinomas and putative premalignant cells has so far failed to identify a gene or gene product clearly associated with the malignant or premalignant phenotype. Understanding the cellular lineage from normal cell through putative premalignant cell to cancer is critical to understanding the process of carcinogenesis. Application of new immunological (monoclonal antibody, transplantation) and molecular biological (gene cloning, oncogene identification) approaches to this problem holds promise that the process of hepatocarcinogenesis will be better known in the near future.

A new family of LTR-like sequences abundantly expressed in rat tumors

We report the identification of two genome DNA fragments containing middle repetitive sequences abundantly expressed in various rat tumors but rarely in normal tissues. These fragments included homologous regions which belonged to a new family of long terminal repeat (LTR) like sequences, designated RAL elements; one displayed the solitary type and the other a provirus structure. The element was transcribed in a strand specific fashion and started from the presumptive cap site within the RAL element. The presumptive polyA addition site within the element was also utilized as evidence of the analysis of a cDNA clone containing the RAL element. This evidence suggested that transcriptional control signals within the element were functioning. Run on assay revealed that expression of the element was regulated at the transcriptional level.

The mRNA coding for the secreted protease transin is expressed more abundantly in malignant than in benign tumors

Transin RNA is a 1.9-kilobase RNA transcript induced by oncogenes in rat embryo fibroblast cell lines. We show that RNA species complementary to a cloned transin cDNA are present in mouse skin squamous cell carcinomas induced by a classical initiation-promotion protocol but not in premalignant, benign papillomas or in normal epidermis. A single application of a tumor promoting phorbol ester to normal epidermis elicits a transient increase in these RNA levels. Transin RNA encodes a secreted protease, an activity consistent with a functional role for enhanced expression of transin RNA in the progression of benign, encapsulated tumors to malignant, invasive carcinomas.

Isolation of cellular genes differentially expressed in mouse NIH 3T3 cells and a simian virus 40-transformed derivative: growth-specific expression of VL30 genes

We constructed and screened a cDNA library made from simian virus 40 (SV40)-transformed NIH 3T3 cells, and we isolated cDNAs representing genes that are differentially expressed between the parental cell and its SV40-transformed derivative. We found only a small number of cDNAs representing such genes. Two isolated cDNA clones represented RNAs expressed at elevated levels in the transformed cell line in a manner relatively independent of growth conditions. The expression of two other cDNAs was growth specific because transformed cells and nonconfluent parental cells contained higher levels of the homologous RNAs than did confluent, contact-inhibited parental cells. Another cDNA was well expressed in confluent parental and confluent transformed cells, but not in nonconfluent cells. The expression of some of these cDNAs varied strikingly in different mouse cell lines. Thus the genotype or histories of different cell lines can also affect the expression of certain genes. Interestingly, the only cDNA isolated that was expressed exclusively in the transformed cell was from an SV40 message. We focused on a growth-specific cDNA which we show is derived from a mouse endogenous retrovirus-like family called VL30. We sequenced the 3' long terminal repeat (LTR) of this transcriptionally active VL30 gene. This LTR has good homology with other VL30 LTR sequences, but differences occur, particularly upstream of the VL30 promoter. We found that VL30 gene expression varied in different mouse cell lines such that C3H cell lines had very low levels of VL30 transcripts relative to NIH 3T3 cell lines. However, Southern analysis showed that both cell lines had about the same number of VL30 genes homologous to our probe and that the position of the majority of these genes was conserved. We discuss possible explanations for this difference in VL30 expression.

Isolation of the oncogene and epidermal growth factor-induced transin gene: complex control in rat fibroblasts

Various oncogenes or epidermal growth factor (EGF) induce transcription of a 1.9-kilobase RNA (transin RNA) in rat fibroblasts. The induction by EGF can be blocked by cycloheximide. Thus the response of the transin gene to EGF appears to require de novo protein synthesis. Transin RNA induction is specific to EGF, as neither insulin, platelet-derived growth factor, fibroblast growth factor, nor transforming growth factor beta could elicit the same response. However, transforming growth factor beta could block the EGF induction of transin RNA. Whereas the calcium ionophore A23187 and the tumor promoter TPA, either alone or administered together, did not increase transin RNA levels, TPA could synergise with a serum factor to effect such an increase. Dibutyryl cyclic AMP also induced transin RNA. Treatment of cells with the microfilament-disrupting agent cytochalasin B, but not the microtubule-disrupting agent colcemid, resulted in an increase in transin RNA levels, suggesting a role for the cytoskeleton in control of transin gene expression. The transin RNA does not contain repeated sequences and appears to be encoded by a single-copy gene. The protein sequence encoded by the last four exons of the transin gene shows some homology to two regions of the heme-binding protein hemopexin.

Growth-related changes in specific mRNAs upon lectin activation of human lymphocytes

A cDNA library in lambda gt10 was constructed from the cytoplasmic poly(A) +RNA of human peripheral blood lymphocytes after 72 hr of phytohemagglutinin stimulation, with the aim of assessing selective gene expression as a result of lymphocyte activation. Thirteen recombinants were isolated by the use of an enriched probe and differential screening. These clones were categorized into two groups with respect to their hybridization to mRNA. In the first group three recombinants were isolated, which hybridized to single discrete mRNAs in the size range 0.7-1.7 kb. The mRNAs corresponding to these clones were present at elevated levels in activated lymphocytes, but the kinetics of increase differed. The 0.7-kb mRNA coded for by clone p1L1 increased two-fold at 6 hr and remained elevated over 72 hr, as did beta-actin mRNA. The 1.7-kb mRNA coded for by clone p9L2 increased two- to three-fold after 6 hr and was maximally expressed after 24 hr exposure to phytohemagglutinin, coincident with the onset of DNA replication, and maintained this level up to 72 hr. The 1.0-kb mRNA coded by p10L2F which was rare in resting cells increased 25- to 30-fold after 6 hr, prior to overall transcriptional increases and reached peak levels after 72 hr when a substantial proportion of the cells were in the S and G2 + M phases of the cell cycle. This clone was undetectable or very rare in the leukemic T-lymphoblast cell line CCRF-CEM. The second group of clones, consisting of the remaining 10 recombinants, did not hybridize to discrete bands, but to a smear on RNA blots.(ABSTRACT TRUNCATED AT 250 WORDS)

Isolation of cellular genes differentially expressed in mouse NIH 3T3 cells and a simian virus 40-transformed derivative: growth-specific expression of VL30 genes

We constructed and screened a cDNA library made from simian virus 40 (SV40)-transformed NIH 3T3 cells, and we isolated cDNAs representing genes that are differentially expressed between the parental cell and its SV40-transformed derivative. We found only a small number of cDNAs representing such genes. Two isolated cDNA clones represented RNAs expressed at elevated levels in the transformed cell line in a manner relatively independent of growth conditions. The expression of two other cDNAs was growth specific because transformed cells and nonconfluent parental cells contained higher levels of the homologous RNAs than did confluent, contact-inhibited parental cells. Another cDNA was well expressed in confluent parental and confluent transformed cells, but not in nonconfluent cells. The expression of some of these cDNAs varied strikingly in different mouse cell lines. Thus the genotype or histories of different cell lines can also affect the expression of certain genes. Interestingly, the only cDNA isolated that was expressed exclusively in the transformed cell was from an SV40 message. We focused on a growth-specific cDNA which we show is derived from a mouse endogenous retrovirus-like family called VL30. We sequenced the 3' long terminal repeat (LTR) of this transcriptionally active VL30 gene. This LTR has good homology with other VL30 LTR sequences, but differences occur, particularly upstream of the VL30 promoter. We found that VL30 gene expression varied in different mouse cell lines such that C3H cell lines had very low levels of VL30 transcripts relative to NIH 3T3 cell lines. However, Southern analysis showed that both cell lines had about the same number of VL30 genes homologous to our probe and that the position of the majority of these genes was conserved. We discuss possible explanations for this difference in VL30 expression.

12-O-Tetradecanoylphorbol-13-acetate (TPA)-induced gene sequences in human primary diploid fibroblasts and their expression in SV40-transformed fibroblasts

We have isolated cDNA sequences from TPA-treated primary human fibroblasts, which indicate RNA species that are coordinately regulated after treatment of these cells with either ultraviolet light, mitomycin C, the UV-induced factor EPIF, or TPA. The levels of RNA are elevated in Bloom syndrome (cells of two out of three patients). After transformation with SV40 one of the sequences is overexpressed while another one is reduced. Both genes maintain their inducibility by the agents mentioned.