In human fibroblasts, N-phosphoacetyl-L-aspartate (PALA) and γ-radiation induce reversible and irreversible p53-mediated G1 cell cycle arrest, respectively. By coupling the premature chromosome condensation technique to fluorescence in situ hybridization, we found no evidence of DNA damage after PALA treatment. We used representational difference analysis (cDNA-RDA) to study changes in gene expression after PALA treatment and γ-radiation in normal human fibroblasts. The mammary-derived growth inhibitor (MDGI) gene was expressed in PALA-treated cells. Ectopic MDGI expression arrested PALA-treated but not irradiated RKO cells. Expression of an antisense RNA against MDGI resulted in partial G1 escape of PALA-treated human fibroblasts. The tumor necrosis factor stimulated gene 6, TSG-6, seems to be under the control of p53 and is only and specifically induced upon PALA treatment. In irradiated cells we have identified `novel' genes that are differentially expressed, along with known genes not previously linked to cell cycle control. Some of these `novel' genes correspond to clones in the expressed sequence tag (EST) database; one of them shows identity with ESTs mapping to a region on chromosome 7, where gene(s) involved in replicative senescence and frequently deleted in tumors are located. Thus, PALA treatment and γ-irradiation elicit a pattern of differential gene expression that could contribute to a quiescence or senescence-like phenotype.

Introduction

The tumor suppressor p53 is an important factor for the maintenance of genetic stability. Ionizing radiation and chemotherapeutic drugs used in cancer treatments directly damage DNA and increase p53 levels, leading to arrest in the G1 phase of the cell cycle or apoptosis (16). G1 arrest is in part dependent on a pathway that involves p53-dependent transcriptional activation of p21WAF1/CIP1/SDI1 (7), an inhibitor of the G1 cyclin–Cdk complexes, such as cyclin D–Cdk4/Cdk6 and cyclin E–Cdk2, which in turn inhibits phosphorylation of pRb (8,9). There is evidence that p53 can also be activated following hypoxia in the absence of DNA damage (10) and may respond to metabolic signals. While many p53-responsive anti-cancer agents cause DNA damage directly, disruption of nucleotide pools by antimetabolites such as the CAD inhibitor N-phosphonacetyl-L-aspartate (PALA) (11) could cause DNA damage during progression through S phase. However, G0/G1 synchronized cells with a functional p53 pathway arrest in G1 in the presence of PALA, while G0/G1 synchronized cells lacking a functional p53 pathway are able to progress into S phase (12). This suggests that p53 helps to maintain genetic stability by preventing DNA replication during metabolic depletion, indicating the existence of a different p53-dependent arrest function.

Previously, we found that, in response to γ radiation-induced DNA damage, normal human diploid fibroblasts undergo prolonged, probably permanent, arrest similar to senescence (6). In contrast, the G1 arrest caused by PALA treatment is reversible (12,13). Chromosome DNA damage, which could be a secondary effect, probably requires transit through S phase, and is not readily observed by classic cytogenetic analysis of such cells. These results suggest that important differences exist in G1-arrested cells following irradiation or antimetabolite treatment.

Different genes could be expressed depending on the nature of the signal (DNA damage or metabolite depletion) being transduced to the p53-mediated arrest pathway. In PALA-treated cells we found no evidence of DNA damage by fluorescence in situ hybridization (FISH) coupled to premature chromosome condensation (PCC). Both treatments induce changes in gene expression as judged by cDNA representational difference analysis (cDNA-RDA) (14). The expression of some of these genes in p53-defective treated cells suggests their independence from p53 trans-activation. Lack of expression of these genes in a panel of tumor cell lines suggests a role in tumorigenesis. One of the newly identified genes shows identity with expressed sequence tags (ESTs) mapping to a chromosomal region where gene(s) involved in replicative senescence are located and frequently deleted in tumors. Thus, PALA treatment and γ-irradiation elicit a pattern of differential gene expression that could contribute to a quiescence or senescence-like phenotype.

Materials and methods

Cell culture and treatments

WS1 (normal human embryonic skin fibroblasts) cells expressing a neomycin resistance gene (WS1neo) and the HPV16 E6 protein (WS1E6), or an antisense oriented mammary-derived growth inhibitor (MDGI) cDNA fragment were prepared by retroviral gene transduction and used within 12 passages after infection. RKO and H1299 cells expressing the MDGI full-length cDNA were generated by retroviral gene transduction. Cells expressing E6 have a non-functional p53 pathway. HeLa tumor cells have no functional p53; HL60 human leukemia cells do not express p53; H1299 is a non-small cell lung carcinoma cell line lacking p53 expression due to a 5′ intragenic deletion; U2OS osteosarcoma tumor cells and RKO colorectal carcinoma cells are considered p53 wild-type. Cells were maintained in Dulbecco's modified Eagle's medium (DMEM) supplemented with 1× non-essential amino acids (Gibco-BRL, Life Technologies) and 10% fetal bovine serum (FBS; Euroclone Ltd, UK) at 37°C in a humidified atmosphere containing 5% CO2. All cells were split 1:4, 1:6 approximately every 3 days for maintenance. Human fibroblasts synchronized in G0 by serum deprivation were split and released in medium plus 10% dialyzed FBS and 100 μM PALA (NSC-224131) provided by the Drug Biosynthesis and Chemistry Branch of the National Cancer Institute (Bethesda, MD) for additional 72 h, or the cells were irradiated (4–6 Gy) at room temperature with a 137Cs γ-irradiator.

Metaphase aberrations

Asynchronously growing human fibroblasts were treated with PALA (100 μM) only or with PALA and caffeine (1 mM) and 2 × 106 cells were seeded in 75 cm2 flasks. Cultures were then incubated in 0.2 μg/ml colcemid (Gibco-BRL, Life Technologies) for 24 h and mitotic cells were harvested. Cell swelling was induced by treatment for 10 min with 75 mM KCl at 37°C; cells were then fixed with three changes of methanol:acetic acid (3:1) and dropped on to clean ice-cold glass microscope slides. Slides were stained in 2.5% Giemsa for 10 min and metaphase spreads were scored for chromosomal aberrations (13).

FISH and PCC analysis

PCC was used to visualize chromosome damage in G1-arrested PALA (100 μM) treated cells (15). Chinese hamster ovary (CHO) cells were used as the mitotic inducer population. These cells were cultured at 37°C in a 5% CO2 humidified incubator in DMEM supplemented with 1× non-essential amino acids, 1% penicillin/streptomycin and 10% FBS. Mitotic CHO cells were obtained by gentle shake-off following overnight incubation in medium containing 0.2 μg/ml colcemid.

Human fibroblasts were fused with mitotic CHO cells to induce chromosomal condensation. Interphase human fibroblasts were collected by centrifugation and mixed with mitotic CHO cells (ratio 1:1). The mixture was washed twice in phosphate-buffered saline (PBS) and the pellet was resuspended in 0.1 ml/106 of pre-warmed polyethylene glycol (PEG) 1500 added dropwise to allow cell fusion. Following PEG dilution over a period of 10 min in medium without serum, cells were centrifuged, resuspended in complete medium containing 0.2 μg/ml colcemid and then incubated for an additional 90 min at 37°C. At the end of the incubation period, cells were processed as above and FISH was done according to the published procedures with minor modifications (16) and using total human DNA biotinylated as a probe. Slides were mounted in antifade solution (Cytifluor, London) containing 2 μg/ml of propidium iodide (Sigma, St Louis< MO) and examined with a Zeiss Axioskop microscope equipped for epifluorescence. Nuclei and metaphases were photographed; images were scanned and digitally processed with Adobe Photoshop.

RNA isolation and northern blot analysis

Total cellular RNA was isolated using the RNAeasy kit (Qiagen GmbH). Fifteen micrograms of each sample was resolved on a denaturing MOPS–formaldehyde agarose gel. Electrophoresis, transfer to nylon membranes (Hybond-N; Amersham–Pharmacia), hybridization and autoradiography were done according to the manufacturer's instructions. Membranes were hybridized using as probes the DNA fragments derived from PCR amplification labeled with [α-32P]dCTP by the random primed method (Amersham–Pharmacia).

cDNA-RDA and DNA sequencing

cDNA-RDA was performed according to the published procedure (14) on total RNA isolated from human fibroblasts after 72 h of PALA treatment or γ-irradiation. cDNA was prepared by reverse transcription of total RNA using an oligo(dT) primer and Superscript II reverse transcriptase (Life Technologies). The sequences of the synthetic oligonucleotides used in cDNA-RDA were as follows: R-Bam-24, 5′-AGCACTCTCCAGCCTCTCACCGAG-3′; R-Bam-12, 5′-GATCCTCGGTGA-3′; J-Bgl-24, 5′-ACCGACGTCGACTATCCATGAACA-3′; J-Bgl-12, 5′-GATCTGTTCATG-3′; N-Bam-24, 5′-AGGCAACTGTGCTATCCGAGGGAG-3′; N-Bam-12, 5′-GATCCTCCCTCG-3′. The double-stranded cDNA was digested with the restriction enzyme DpnII, ligated to specific adapters (R-Bam 12-to-24-mer) and amplified by PCR using specific primers (R-Bam24) to generate the `representations'. After the generation of these `representations' the adapters were removed with DpnII and the digested DNA was phenol extracted and ethanol precipitated to generate the `driver'. A fraction (20 μg) of this digested `representation' was gel purified and the products (free of the adapters) were ligated to the J-Bgl 12–24 adapter to generate the `tester'. For the first subtractive hybridization, 0.4 μg J-Bgl-ligated tester was mixed with 40 μg of the driver (molar ratio 1:100). For each subtraction, PCR reactions were set up with diluted hybridization mix using the J-Bgl24 oligonucleotide as a primer. Products were combined giving the first difference product (DP1). J-adapters on DP1 were replaced with N-Bam24 adapters and the process of subtractive hybridization and selective amplifications (using alternately N-Bam and J-Bgl adapters) was reiterated to generate DP2, DP3 and DP4. The molar ratio between tester and driver in DP2, DP3 and DP4 was 1:400, 1:80 000 and 1:400 000, respectively. The DNA sequence of fragments cloned in the BamH1 site of the plasmid pUC19 was determined by a cycle sequencing procedure using the Thermo Sequenase kit (Amersham-Pharmacia) and [α-33P]ddNTPs.

Cloning of the full-length MDGI cDNA by RT–PCR

Total RNA was isolated from human placenta using the RNAeasy kit (Quiagen GmbH) and 2 μg were DNase treated and reverse transcribed into cDNA using Superscript II reverse transcriptase and oligo(dT) as a primer. cDNA was used as a template in PCR reactions with the primers 5′-ATGGATCCACTATGGCGGACGCCTT-3′ and 5′-GGCCTTGGATCCGCTTTATTGACC-3′ in order to obtain the full-length MDGI cDNA. The PCR product, after BamHI digestion, was cloned into the plasmid pGEM-t-easy (Promega), sequenced and subcloned into the BamHI site of the retroviral vector pSLX to allow stable expression in infected cells.

Flow cytometric cell cycle analysis

For experiments on asynchronous cultures, cells were seeded 24 hr prior to treatment at densities to prevent contact inhibition. Cultures were pulse-labeled with 10 μM BrdU (a thymidine analog) for 4 h 72–144 h after treatment, harvested and subjected to flow cytometric cell cycle analysis as previously described (6). G0 synchrony of WS1-asMDGI cells was achieved by serum deprivation. For experiments on serum-deprived cultures, the FBS concentration in the medium was reduced from 10% to 0.1% for 72 h and cells were released in medium containing 10% FBS. BrdU (50 μM) was added at the time of release for continuous labeling of cells undergoing DNA synthesis in the presence of PALA. Cultures were harvested and subjected to flow cytometric cell cycle analysis 72 h after release as previously described (6). Data analysis was done using Cell Quest. Cellular debris and fixation artifacts were gated out, and the G0/G1, S and G2/M fractions were quantified. Experiments were repeated at least twice; 10 000 events were analyzed for each sample and representative experiments are shown.

Results

Chromosomal aberrations are not detected after treatment with PALA by PCC–FISH analysis

The exquisite sensitivity of the p53-mediated arrest pathway suggests that PALA treatment could keep cells arrested in G0/G1 by generating a low amount of DNA damage. Previously, we reported the absence of chromosomal aberrations in G0-synchronized WS1neo cells and then released and in asynchronous NHF-3 cells treated with PALA and analyzed for the presence of chromosomal aberrations (12). In contrast, nearly 30% of NHF-3 cells show aberrant chromosomes after irradiation (6). In order to reveal any DNA breakage caused by PALA, asynchronously growing human fibroblasts were treated with PALA and exposed to caffeine to determine if inhibition of repair increased the number of unrepaired chromosome breaks resulting in chromosomal aberrations. As shown in Table I, there was a modest increase in the number of cells bearing chromosomal metaphase aberrations after treatment with PALA plus caffeine for 24 h (aberrations in 1/200 metaphases) or 72 h (1/100) compared with untreated cells (0/400) and cells treated with PALA for 24 h (1/500) or 72 hours (0/100). The same cytogenetic analysis of p53-deficient HT1080 and WS1-E6 cells showed the presence of an increased number of DNA breaks after PALA treatment (Table I). These data indicate that PALA treatment does not readily induce DNA damage detectable by conventional metaphase chromosome analysis. To investigate whether chromosomal damage such as nicks or gaps could be generated in G1 by PALA and therefore be responsible for the cell cycle arrest observed, we used the PCC technique (15) coupled to FISH analysis. Following fusion between interphase and mitotic cells, the mitotic cells induce a prophase-like reaction in interphase cells, which results in the condensation of chromatin into discrete structures called prematurely condensed chromosomes. A biotinylated human genomic DNA was used as a probe to allow detection of DNA damage caused by the presence of nicks and gaps. We induced chromosome condensation of PALA-treated G1-arrested human fibroblasts by fusing them with mitotic hamster cells. FISH of the resulted PCC metaphases showed absence of DNA fragmentation revealed as additional chromosomes in these G1-arrested cells (Figure 1A). As a control, PCCs were induced in fibroblasts asynchronously grown for 72 h in the presence of PALA. As shown in Figure 1B, the PCC technique causes chromosomal fragmentation, revealing the presence of DNA damage that is otherwise masked. These results confirm that PALA is able to keep cells in G1 without inducing DNA damage; the presence of additional PCCs in asynchronously grown PALA-treated cells only reflects the fact that PALA can induce DNA damage during the S-phase transition and that some unrepaired DNA breaks are unmasked by the mechanical stress caused by the PCC technique. We concluded that DNA repair did not occur in G1-synchronized PALA-treated cells as revealed by BrdU incorporation, because of unscheduled DNA synthesis, detected by a FITC-conjugated mouse monoclonal anti-BrdU antibody (data not shown).

Differentially expressed genes detected by cDNA-RDA in irradiated or PALA-treated human fibroblasts

To investigate whether differential gene expression could be detected depending on the signal transduced to the p53-mediated arrest pathway, we used the cDNA-RDA technique. This method allows the isolation of differentially expressed genes in isogenic cells by coupling subtractive hybridization to PCR. cDNA-RDA was performed on total RNA extracted from human fibroblasts after 72 h of PALA treatment or γ-irradiation. It is known that irradiation or PALA treatment in asynchronous human fibroblasts arrest cells both in G1 and G2/M. To study differential expression of genes presumably responsible for the G1 arrest, cells were previously synchronized in G0/G1 by serum starvation. Synthesized double-stranded cDNAs were digested with the restriction enzyme DpnII, ligated to specific adapters (12-to-24-mer) and amplified by PCR. These PCR products, called `representations'(Figure 2A) were used in subsequent steps of subtractive hybridization followed by PCR amplifications in order to generate the `difference products'. Following each PCR reaction, the old adapter was removed and a new one was added using a different tester:driver ratio. Using this approach, four difference products (DP1–4) were obtained (Figure 2A). The final difference products, DP3 and DP4, were digested and cloned into the BamHI site of the pUC19 plasmid vector. The cloned fragments were used as probes in Southern blots of the `representations' (Figure 2B). DNA fragments that gave a differential hybridization signal in the `representations' were used in northern hybridization experiments to confirm that they really corresponded to mRNA differentially expressed in the treated cells (Figure 2B).

Thirty-five fragments, from a total of 60 cloned into the plasmid vector pUC19, gave a different hybridization signal by Southern hybridization of the `representations' and 17 of these fragments displayed a different expression pattern when hybridized against total RNA isolated from PALA and γ-irradiated cells. Twelve out of the 17 positive clones were sequenced by cycle sequencing and the resulting sequences were compared (Table II) with sequences in the GenBank database.

Some of the PCR fragments, derived from the cDNA-RDA experiments done using the `representation' derived from the PALA-treated cells as a tester, were used as probes in northern hybridization experiments, in order to assess any dependence on p53 status and cell cycle position. Northern blots were performed using total RNA isolated from WS1neo and WS1-E6 cells. Northern blots were also performed on total RNA isolated from tumor cells in order to assess the gene expression pattern. These experiments are shown in Figure 3A. Here we probed northern blots with the MDGI gene and observed a basal level of expression in asynchronous cells, which was increased in starved cells. Northern blot analysis showed that the MDGI gene was not expressed in irradiated G1-arrested cells, whereas it was expressed in G1-arrested PALA-treated cells. A similar profile of expression was seen in p53-deficient (WS1-E6) cells. RT–PCR experiments on total RNA extracted from PALA-treated RKO and H1299 cells showed lack of MDGI expression in these cells (G.Seidita, unpublished results). These findings suggest that MDGI expression is not influenced directly by p53 status. When the human tumor necrosis factor-inducible (TSG-6) gene was used as a probe in northern hybridization experiments, expression was only observed in PALA-arrested cells. Interestingly, lack of expression was observed both in asynchronous and synchronized cells, indicating a specific induction of this gene upon PALA treatment. The fact that there was no expression in WS1-E6 cells suggests that this gene could be under the control of p53. The human phosphotyrosine independent ligand p62 for the p56lck Src homology domain 2 (SH2) gene was expressed differently in PALA-treated cells than in irradiated cells. However, expression of this gene does not seem to be under p53 control because there is a basal level of expression in WS1-E6 cells. Expression of MDGI and TSG-6 genes was not detected in most tumor cells tested, with the exception that HeLa cells expressed MDGI. Expression of the p62 gene was detected in all tumor cells. Northern analysis was also carried out on total RNA isolated from the same cells indicated above, using as a probe DNA fragments obtained from cDNA-RDA of the representation derived from γ-irradiated cells as a tester. The plasminogen activator inhibitor (PAI-1) showed increased expression after γ-irradiation in WS1neo cells, and this was more pronounced for the band corresponding to the 3.0 kb transcript (Figure 3B). PAI-1 was not expressed after PALA treatment. A similar pattern was observed in WS1-E6 cells, suggesting that PAI-1 expression does not depend on p53 status. This is in contrast with a previous report suggesting that the PAI-1 gene can be trans-activated by p53 (17). We detected a constitutive level of PAI expression also in the p53-containing RKO cell line but not in the p53-containing U2OS cell line. This result suggests that, in this case, the presence of a wild-type p53 alone is not sufficient for PAI expression.

The eukaryotic initiation factor 4AII gene (eIF4AII) was expressed in irradiated WS1neo cells but not in PALA-treated WS1neo cells. The pattern of expression of eIF4AII in WS1-E6 cells suggests that its expression does not depend on p53 status, because there is no functional p53 protein in these cells. For fragments 57 and 55, which corresponded to cloned ESTs, expression was observed only after irradiation in WS1neo cells, indicating a specific induction of these genes. PAI-1 gene expression was not detected in most tumor cells with the exception of RKO, whereas eIF4AII gene expression was detected in all tumor cells. Fragments 57 and 55 were not expressed in any of the tumor cells analyzed except RKO cells.

MDGI participates in PALA-induced G1 arrest in human cells

To investigate whether the differential expression of MDGI observed after PALA treatment plays any role in the p53-mediated arrest response in human cells, human MDGI cDNA containing the full-length coding region of MDGI was isolated by RT–PCR, verified by sequencing and subcloned into the retroviral vector pSLX to allow stable expression in infected cells. To evaluate whether p53 contributes to the response associated with MDGI expression, after PALA treatment, both p53-positive (RKO) and p53-negative (H1299) cells were infected with recombinant retroviruses carrying the full-length MDGI cDNA. As shown in Figure 4A, stably transfected RKO cells (RKO-MDGI) showed a marked reduction in the proportion of S-phase cells when treated with PALA, with the majority of the cells arrested in G1 and G2/M both at 72 h and 144 h. In contrast, untransfected control cells (RKO) and cells transfected with empty vector alone (RKO-pSLX) did not arrest in G1 and cell progression in presence of PALA resulted in increased mortality of these cells as revealed by an increasing number of cells with a sub-G1 content of DNA. Such apoptotic cells made up >80% of the population for RKO cells and ∼60% for RKO-pSLX cells at 144 h. These results suggest that ectopic MDGI expression in arresting cells, in response to PALA, could protect them from undergoing apoptosis. Interestingly, the RKO-MDGI cells recovered from the arrest when PALA was removed, suggesting that MDGI, when ectopically expressed in these cells, participates in transient arrest. Western experiments to assess the p53 protein level following MDGI over-expression showed no p53 increase in untreated RKO-MDGI cells. Also, RKO and RKO-MDGI cells showed, when treated with PALA, a similar p53 increase (G.Costanzo, unpublished results). Both stably transfected H1299 (H1299-MDGI) and control H1299 cells were resistant (Figure 4B) to PALA and did not arrest at the two time points analyzed. To assess whether MDGI over-expression in RKO cells could be responsible for a generalized arrest response following stressful signals, we analyzed by FACScan the cell cycle progression after γ-irradiation. As shown in Figure 4C, it seems that MDGI participates specifically in the PALA-induced arrest pathway, because the cell cycle distribution of RKO-MDGI cells after treatment with 6 Gy of γ-rays was similar to that observed for untransfected RKO cells treated in the same way. These results raise the possibility that the lack of MDGI expression could, despite the presence of normal p53, render cells refractory to PALA-induced G1 arrest.

To explore this possibility further, we cloned the MDGI fragment in the antisense orientation into the retroviral vector pSLX. The recombinant retroviral vector obtained, pSLX-asMDGI, was used to infect normal human fibroblasts (WSI). The infected cells (WS1-asMDGI) were then used to investigate whether the presence of antisense MDGI is sufficient to override G1 arrest after PALA treatment. Figure 4D shows the cell cycle distribution, determined by FACScan, of WS1-asMDGI cells untreated and after PALA treatment. WS1-asMDGI cells behaved normally when starved by serum deprivation and then released, indicating the absence of any deleterious effect of expression of the antisense MDGI. However, synchronized WS1-asMDGI cells escaped, at least in part, the G1 arrest after PALA treatment. Continuous BrdU labeling indicated that a higher percentage of WS1-asMDGI cells (27%) than WS1 control cells (5.5%) can progress into S-phase.

Discussion

Signal transduction pathways that lead to G0/G1 arrest in the presence of DNA damage appear to be similar to those or in the presence of depleted ribonucleotide pools (6,12), but there are subtle differences, suggesting that some unique components are involved in the two types of pathway. Here we investigate whether DNA damage and metabolite depletion trigger different expression of genes involved in the p53-mediated arrest pathway. To this end, we studied first the presence of DNA damage in G0/G1-arrested PALA-treated cells by using a combination of PCC and FISH. We did not detect DNA damage in G0/G1-arrested PALA-treated cells. These results are consistent with what we and others have observed previously by FACScan analysis (6,12). Some of the asynchronous PALA-treated cells arrested in G2 could have some unrepaired DNA breaks (single strand breaks) that are unmasked by unprogrammed chromosomal condensation as revealed by PCC analysis. Indeed, PALA does not seem to induce persistent chromosomal damage that could account for the cell cycle arrest as revealed by the use of the repair inhibitor caffeine in asynchronous NHF3 cells.

We then investigated, using the cDNA-RDA method (14), whether metabolite depletion or DNA damage can trigger differential expression of genes involved in the p53-mediated arrest in human fibroblasts. Using this approach we were able to identify differentially expressed genes following different stress signals. Some of these genes, which are expressed in arrested cells by serum deprivation, seem to be involved in the maintenance of cells in the G1 phase of the cell cycle following PALA treatment.

The MDGI gene is a putative tumor suppressor, involved in the differentiation of breast cells, that is able to slow cell growth when overexpressed in cells derived from breast cancer (18). The MDGI gene is a candidate for response to a metabolic signal in that it suppresses the mitogenic effects of epidermal growth factor (19). Our findings that MDGI over-expression in RKO cells results in G1 arrest, after PALA treatment, suggest a role for the MDGI gene in p53-mediated G1 arrest. This arrest is likely to be p53 dependent because it was not observed in p53-deficient H1299 cells infected with the same recombinant retroviruses harboring the full-length MDGI cDNA. The observed G1 arrest in RKO-MDGI cells appeared to be transient, disappearing upon removal of PALA, suggesting activation of the p53-dependent checkpoint responsible for the arrest induced by ribonucleotide depletion (12).

The involvement of the MDGI gene in the PALA-induced arrest pathway is stressed by our findings that some (27%) of the normal human (WS1) fibroblasts infected with the recombinant retrovirus harboring the MDGI fragment in the antisense orientation escaped arrest after PALA treatment. The fact that we observed a similar increase in the p53 protein level in RKO and RKO-MDGI PALA-treated cells, together with the observation that only RKO-MDGI cells were able to arrest in G1, strongly suggests that MDGI could work upstream of p53 or in cooperation with p53 in the PALA-induced G1 arrest.

The p62 gene was also differentially expressed. This gene encodes a phosphotyrosine-independent ligand of the SH2 domain of the tyrosine kinase p56lck, a member of the c-Src family of cytoplasmic tyrosine kinases (20). Binding of p62 to the SH2 domain of p56lck could prevent the catalytic activity of this tyrosine kinase affecting downstream transducers and, in turn, cellular progression. In addition, p62 has been suggested to belong to a class of ubiquitin-binding proteins, thereby affecting a signal transduction pathway through ubiquitination and protein degradation (21). As we have shown (Figure 3), MDGI and p62 expression appear to be independent of p53. On the other hand, TSG-6 seems to be under the direct control of p53, given its lack of expression in p53-deficient (WS1-E6) cells. In addition, TSG-6 mRNA was not detected in untreated cells but was readily induced by PALA in normal human fibroblasts. TSG-6 was originally identified as a tumor necrosis factor/interleukin-1-inducible gene (22) and has a potent anti-inflammatory activity along as well as inhibiting protease action (23). We speculate that the product of TSG-6 could, like PAI-1, inhibit the plasmin/plasminogen pathway.

PAI-1 was differentially expressed in γ-irradiated cells although its expression is probably not under the direct control of p53 (Figure 3). Previous in vitro studies (17) suggested that p53 binds to the PAI-1 promoter; overexpression of a temperature-sensitive mutant human p53, at the permissive temperature of 32.5°C, resulted in an increase in both the 3.0 and 2.2 kb PAI-1 transcripts when compared with control cells. However, this increase, which was higher for the 3.0 kb transcript, could reflect a senescent-like status because of p53 overexpression in these cells, rather than being linked to p53 directly. Our results showing that PAI-1 is expressed in p53-deficient (WS1-E6) cells following irradiation do not support the idea that accumulation of cellular p53 could directly affect PAI-1 expression. Interestingly, PAI-1 transcription was elevated in late passages (senescent) of human fibroblasts (24) principally for the 3.0 kb mRNA form, and in young fibroblasts in which two cyclin-dependent inhibitors, p16Ink4a and p21WAF1/CIP1/SDI1, were ectopically expressed (25). Senescent cells, similarly to irradiated fibroblasts, appear irreversibly arrested in the G1 phase, therefore PAI-1 increased transcription could be considered a marker of the senescent-like phenotype suggested for irradiated fibroblasts (6). PAI-1 is known to be induced by a transforming growth factor β (TGFβ) signaling pathway probably mediated by Smad proteins. Phosphorylation of Smad3 by a TGFβ receptor is believed to be an essential step in signal transduction by TGFβ for the activation of the PAI-1 promoter. Overexpression of Smad protein combinations can mimic the transcriptional effect of TGFβ on the PAI-1 promoter (26). In addition, dominant-negative Smad3 reduces the stimulation of PAI-1 (27). Eukaryotic initiation factor 4AII is a 407 aa factor that is required for mRNA binding to ribosomes. It has been characterized as an RNA-dependent ATPase and RNA helicase belonging to a family of RNA helicases termed DEAD box proteins. The murine eIF4AII gene is preferentially expressed in organs with low proliferative capacity. eIF4AII mRNA levels increased ∼3-fold in cells synchronized by nutrient starvation, whereas eIF4AI mRNA levels, which are 10-fold more abundant than eIF4AII in growing cells, remained unchanged (28). This suggests that eIF4AII mRNA synthesis is associated preferentially with the growth-arrested state of the cell.

Overcoming the G1 growth arrest due to senescence could be a critical and essential step in the development of human cancer. Senescence could be then considered a major tumor-suppressing mechanism that the cell activates to avoid tumorigenesis. In addition to already characterized genes, we have identified two novel genes (corresponding to fragments 55 and 57) that are characterized by a different pattern of expression in irradiated cells. Fragment 57 shows high homology with one EST cluster that maps to a region on chromosome 7 that is deleted in some tumors and whose expression restores a cellular senescence-like behavior in immortalized human fibroblasts (29,30). Its functional characterization may shed light on its involvement in the regulation of senescence-like cellular mechanisms involved in tumor development.

Our results suggest that expression of genes other than those directly trans-activated by p53 are necessary to maintain cells arrested in G1. In particular, the MDGI gene is a new candidate for involvement in the transient cell cycle arrest observed in PALA-treated cells. The PAI-1 and eIF4AII genes, as well as the novel gene corresponding to fragment 57 participating in cellular senescence, may be responsible for the permanent cell cycle arrest observed following γ-irradiation.

Although the majority of the cloned PCR products derived from cDNA-RDA experiments correspond to known genes, the involvement of MDGI, p62 and TSG-6 in the arrest pathway has not been shown previously. The PCR products from the cDNA-RDA experiments that used γ-irradiated cells showed homology with few genes already present in the GenBank database, suggesting that we have identified novel genes that are specifically expressed following γ-irradiation. Currently we are in the process of obtaining full-length cDNAs corresponding to the isolated cDNA fragments and will then start their functional characterization as it relates to their involvement in the cell cycle arrest response as well as their expression in senescent cells.

Table I.

PALA-induced chromosomal aberrations in p53 and p53+ cells

Cell strain/line  Treatment  Abnormal metaphases/total metaphases  Chromatid breaks  Chromatid exchanges(a) 
a Chromatid exchanges (including non-sister-chromatid interchanges such as triradials and quadriradials). 
NHF3  untreated  0/400  –  – 
  PALA 24 h  1/500  – 
  PALA 24 h + caffeine  1/200  12  – 
  PALA 72 h  0/100  –  – 
  PALA 72 h + caffeine  1/100  – 
HT1080  untreated  1/200  – 
  PALA 24 h  6/200 
  PALA 72 h  5/100 
WS1neo  untreated  0/100  –  – 
  PALA 24 h  0/100  –  – 
WS1-E6  untreated  0/100  –  – 
  PALA 24 h  5/100 
Cell strain/line  Treatment  Abnormal metaphases/total metaphases  Chromatid breaks  Chromatid exchanges(a) 
a Chromatid exchanges (including non-sister-chromatid interchanges such as triradials and quadriradials). 
NHF3  untreated  0/400  –  – 
  PALA 24 h  1/500  – 
  PALA 24 h + caffeine  1/200  12  – 
  PALA 72 h  0/100  –  – 
  PALA 72 h + caffeine  1/100  – 
HT1080  untreated  1/200  – 
  PALA 24 h  6/200 
  PALA 72 h  5/100 
WS1neo  untreated  0/100  –  – 
  PALA 24 h  0/100  –  – 
WS1-E6  untreated  0/100  –  – 
  PALA 24 h  5/100 
Table II.

GenBank matches of the fragments obtained from cDNA-RDA of WS1 cells

Fragment no.  GenBank match 
Tester derived from PALA treated cells 
88  TSG-6 mRNA 
23  C3 component of complement 
26  human phosphotyrosine independent ligand p62 for the SH2 domain of tyrosine kinase p56lck 
30  cathepsin X 
34  MDGI (heart fatty acid binding protein) 
41  F-actin 
Tester derived from γ-irradiated cells 
no match 
no match 
human mRNA for eIF4AII 
13  endothelial plasminogen activator inhibitor PAI-1 
55  EST H91867 (chromosome 6) 
57  EST AA191741 (chromosome 7) 
Fragment no.  GenBank match 
Tester derived from PALA treated cells 
88  TSG-6 mRNA 
23  C3 component of complement 
26  human phosphotyrosine independent ligand p62 for the SH2 domain of tyrosine kinase p56lck 
30  cathepsin X 
34  MDGI (heart fatty acid binding protein) 
41  F-actin 
Tester derived from γ-irradiated cells 
no match 
no match 
human mRNA for eIF4AII 
13  endothelial plasminogen activator inhibitor PAI-1 
55  EST H91867 (chromosome 6) 
57  EST AA191741 (chromosome 7) 
Fig. 1.

Detection of DNA damage by premature chromosome condensation coupled to FISH analysis. Labelled total human DNA was used as a probe. The occurrence of DNA damage in G1-arrested PALA-treated cells was detected using PCC after fusion with hamster mitotic cells and FISH. Examples of PCC metaphases: (A) G0/G1 synchronized WS1neo cells treated with PALA for 72 h; (B) asynchronous WS1neo cells treated with PALA for 72 h. PCC metaphases shown in (B) have >46 human chromosomes (brighter).

Fig. 1.

Detection of DNA damage by premature chromosome condensation coupled to FISH analysis. Labelled total human DNA was used as a probe. The occurrence of DNA damage in G1-arrested PALA-treated cells was detected using PCC after fusion with hamster mitotic cells and FISH. Examples of PCC metaphases: (A) G0/G1 synchronized WS1neo cells treated with PALA for 72 h; (B) asynchronous WS1neo cells treated with PALA for 72 h. PCC metaphases shown in (B) have >46 human chromosomes (brighter).

Fig. 2.

Difference products obtained by cDNA-RDA on WS1neo cells treated with PALA or irradiated showed a stepwise reduction in complexity and enrichment in specific bands. (A) cDNA `representations' (M, 123 bp DNA ladder; 1, PALA-treated WS1neo; 2, irradiated WS1neo) and difference products (M, lambda phage DNA HindIII digested; lane 1, tester derived from PALA-treated cells, driver from irradiated cells; 2, tester derived from irradiated cells, driver from PALA-treated cells. Southern and northern hybridizations were performed to confirm the specificity of the fragments obtained from cDNA-RDA. (B) Southern blots of the cDNA `representations' were hybridized using as probes the DNA fragments (5, 34, 8) derived from the cloned difference products. Northern hybridization of WS1neo total RNA (15 μg/lane) extracted from treated cells after 72 h were hybridized as for the Southern blots. The human GAPDH probe was used to ensure equal RNA loading. p, PALA; γ, γ-irradiation.

Fig. 2.

Difference products obtained by cDNA-RDA on WS1neo cells treated with PALA or irradiated showed a stepwise reduction in complexity and enrichment in specific bands. (A) cDNA `representations' (M, 123 bp DNA ladder; 1, PALA-treated WS1neo; 2, irradiated WS1neo) and difference products (M, lambda phage DNA HindIII digested; lane 1, tester derived from PALA-treated cells, driver from irradiated cells; 2, tester derived from irradiated cells, driver from PALA-treated cells. Southern and northern hybridizations were performed to confirm the specificity of the fragments obtained from cDNA-RDA. (B) Southern blots of the cDNA `representations' were hybridized using as probes the DNA fragments (5, 34, 8) derived from the cloned difference products. Northern hybridization of WS1neo total RNA (15 μg/lane) extracted from treated cells after 72 h were hybridized as for the Southern blots. The human GAPDH probe was used to ensure equal RNA loading. p, PALA; γ, γ-irradiation.

Fig. 3.

Northern blot analysis of differentially expressed genes following (A) PALA treatment and (B) γ-irradiation treatment in p53-competent, p53-deficient and tumor cells. Northern blot analysis was performed on total RNA (15 μg) isolated from p53-competent (WS1neo), p53-deficient (WS1-E6), treated or untreated, and tumor cells with different p53 status (HeLa, HL60, RKO, U2OS and H1299). The DNA fragments used as probes, obtained from the cDNA-RDA method, are indicated to the right of each blot. Bottom panels show the ethidium bromide stained gels before transfer. p, PALA; γ, γ-irradiation; S, synchronized; A, asynchronous.

Fig. 3.

Northern blot analysis of differentially expressed genes following (A) PALA treatment and (B) γ-irradiation treatment in p53-competent, p53-deficient and tumor cells. Northern blot analysis was performed on total RNA (15 μg) isolated from p53-competent (WS1neo), p53-deficient (WS1-E6), treated or untreated, and tumor cells with different p53 status (HeLa, HL60, RKO, U2OS and H1299). The DNA fragments used as probes, obtained from the cDNA-RDA method, are indicated to the right of each blot. Bottom panels show the ethidium bromide stained gels before transfer. p, PALA; γ, γ-irradiation; S, synchronized; A, asynchronous.

Fig. 4.

Cell cycle effects of the full-length MDGI cDNA and of the antisense-oriented MDGI fragment in human cells. Cells were fixed at the indicated times, stained with anti-BrdU–FITC and propidium iodide and analyzed by flow cytometry. On the left are flow cytometry dot plots (x-axis, propidium iodide; y-axis, log FITC) showing the distribution of cells in the different cell cycle phases (G1, S and G2/M). The histograms on the right represent the mean number of cells distributed in the different cell cycle phases and apoptotic cells (A) respectively. (A and B) Representative dot plots of untreated and PALA-treated cells [(A) RKO; (B) H1299] and their MDGI-expressing counterparts. (C) Representative dot plots of untreated and irradiated RKO cells and their MDGI-expressing counterparts. (D) Representative dot plots of normal WS1 human fibroblasts infected with recombinant retrovirus harboring the MDGI antisense. In the lower profiles, in which WS1-asMDGI and WSI control cells are compared, BrdU was present for 72 h.

Fig. 4.

Cell cycle effects of the full-length MDGI cDNA and of the antisense-oriented MDGI fragment in human cells. Cells were fixed at the indicated times, stained with anti-BrdU–FITC and propidium iodide and analyzed by flow cytometry. On the left are flow cytometry dot plots (x-axis, propidium iodide; y-axis, log FITC) showing the distribution of cells in the different cell cycle phases (G1, S and G2/M). The histograms on the right represent the mean number of cells distributed in the different cell cycle phases and apoptotic cells (A) respectively. (A and B) Representative dot plots of untreated and PALA-treated cells [(A) RKO; (B) H1299] and their MDGI-expressing counterparts. (C) Representative dot plots of untreated and irradiated RKO cells and their MDGI-expressing counterparts. (D) Representative dot plots of normal WS1 human fibroblasts infected with recombinant retrovirus harboring the MDGI antisense. In the lower profiles, in which WS1-asMDGI and WSI control cells are compared, BrdU was present for 72 h.

1
To whom correspondence should be addressed Email: adileon@unipa.it

We thank Dr Rosario Billetta (School of Medicine, University of Chile) for thoughtful readings of the manuscript and Angelo Polisano (University of Palermo) for technical assistance. This work was supported by grants from Associazione Italiana per la Ricerca sul Cancro (AIRC) and Ministero dell'Università della Ricerca Scientifica (MURST).

References

1
Kastan,M.B., Onyekwere,O., Sidransky,D., Vogelstein,B. and Craig,R.W. (
1991
) Participation of p53 protein in the cellular response to DNA damage.
Cancer Res.
 ,
51
,
6304
–6311.
2
Kuerbitz,S.J., Plunkett,B.S., Walsh,W.V. and Kastan,M.B. (
1992
) Wild-type p53 is a cell cycle checkpoint determinant following irradiation.
Proc. Natl Acad. Sci. USA
 ,
89
,
7491
–7495.
3
Clarke,A.R., Purdie,C.A., Harrison,D.J., Morris,R.G., Bird,C.C., Hooper,M.L. and Wyllie,A.H. (
1993
) Thymocyte apoptosis induced by p53-dependent and independent pathways.
Nature
 ,
362
,
849
–852.
4
Lowe,S.W., Ruley,H.E., Jacks,T. and Housman,D.E. (
1993
) p53-dependent apoptosis modulates the cytotoxicity of anticancer agents.
Cell
 ,
74
,
957
–967.
5
Lowe,S.W., Schmitt,E.M., Smith,S.W., Osborne,B.A. and Jacks,T. (
1993
) p53 is required for radiation-induced apoptosis in mouse thymocytes.
Nature
 ,
362
,
847
–849.
6
Di Leonardo,A., Linke,S.P., Clarkin,K. and Wahl,G.M. (
1994
) DNA damage triggers a prolonged p53-dependent G1 arrest and long-term induction of Cip1 in normal human fibroblasts.
Genes Dev.
 ,
8
,
2540
–2551.
7
El-Deiry,W.S., Harper,J.W., O'Connor,P.M. et al. (
1994
) WAF1/CIP1 is induced in p53-mediated G1 arrest and apoptosis.
Cancer Res.
 ,
54
,
1169
–1174.
8
Harper,J.W., Adami,G.R., Wei,N., Keyomarsi,K. and Elledge,S.J. (
1993
) The p21 Cdk-interacting protein Cip1 is a potent inhibitor of G1 cyclin-dependent kinases.
Cell
 ,
75
,
805
–816.
9
Dulic,V., Kaufmann,W.K., Wilson,S.J., Tlsty,T.D., Lees,E., Harper,J.W., Elledge,S.J. and Reed,S.I. (
1994
) p53-dependent inhibition of cyclin-dependent kinase activities in human fibroblasts during radiation-induced G1 arrest.
Cell
 ,
76
,
1013
–1023.
10
Graeber,T.G., Peterson,J.F., Tsai,M., Monica,K., Fornace,A.J. and Giaccia,A.J. (
1994
) Hypoxia induces accumulation of p53 protein, but activation of a G1-phase checkpoint by low-oxygen conditions is independent of p53 status.
Mol. Cell. Biol.
 ,
14
,
6264
–6277.
11
Collins,K.D. and Stark,G.R. (
1971
) Aspartate transcarbamylase. interaction with the transition state analogue N-(phosphonacetyl)-l-aspartate.
J. Biol. Chem.
 ,
246
,
6599
–6605.
12
Linke,S.P., Clarkin,K.C., Di Leonardo,A., Tsou,A. and Wahl,G.M. (
1996
) A reversible, p53-dependent G0/G1 cell cycle arrest induced by ribonucleotide depletion in the absence of detectable DNA damage.
Genes Dev.
 ,
10
,
934
–947.
13
Di Leonardo,A., Linke,S.P., Yin,Y. and Wahl,G.M. (
1993
) Cell cycle regulation of gene amplification.
Cold Spring Harbor Symp. Quant. Biol.
 ,
58
,
655
–667.
14
Hubank,M. and Schatz,D.G. (
1994
) Identifying differences in mRNA expression by representational difference analysis of cDNA.
Nucleic Acids Res.
 ,
22
,
5640
–5648.
15
Pandita,T.K. and Hittelman,W.N. (
1992
) Initial chromosome damage but not DNA damage is greater in ataxia telangiectasia cells.
Radiat Res.
 ,
130
,
94
–103.
16
Pinkel,D., Landegent,J., Collins,C., Fuscoe,J., Segraves,R., Lucas,J. and Gray,J. (
1988
) Fluorescence in situ hybridization with human chromosome-specific libraries: detection of trisomy 21 and translocations of chromosome 4.
Proc. Natl Acad. Sci. USA
 ,
85
,
9138
–9142.
17
Kunz,C., Pebler,S., Otte,J. and von der Ahe,D. (
1995
) Differential regulation of plasminogen activator and inhibitor gene transcription by the tumor suppressor p53.
Nucleic Acids Res.
 ,
23
,
3710
–3717.
18
Huynh,H.T., Larsson,C., Narod,S. and Pollak,M. (
1995
) Tumor suppressor activity of the gene encoding mammary-derived growth inhibitor.
Cancer Res.
 ,
55
,
2225
–2231.
19
Yang,Y., Spitzer,E., Kenney,N. et al. (
1994
) Members of the fatty acid binding protein family are differentiation factors for the mammary gland.
J. Cell Biol.
 ,
127
,
1097
–1109.
20
Park,I., Chung,J., Walsh,C.T., Yun,Y., Strominger,J.L. and Shin,J. (
1995
) Phosphotyrosine-independent binding of a 62-kDa protein to the src homology 2 (SH2) domain of p56lck and its regulation by phosphorylation of Ser-59 in the lck unique N-terminal region.
Proc. Natl Acad. Sci. USA
 ,
92
,
12338
–12342.
21
Vadlamudi,R.K., Joung,I., Strominger,J.L. and Shin,J. (
1996
) p62, a phosphotyrosine-independent ligand of the SH2 domain of p56lck, belongs to a new class of ubiquitin-binding proteins.
J. Biol. Chem.
 ,
271
,
20235
–20237.
22
Lee,T.H., Wisniewski,H.G. and Vilcek,J. (
1992
) A novel secretory tumor necrosis factor-inducible protein (TSG-6) is a member of the family of hyaluronate binding proteins, closely related to the adhesion receptor CD44.
J. Cell Biol.
 ,
116
,
545
–557.
23
Wisniewski,H.G., Hua,J.C., Poppers,D.M., Naime,D., Vilcek,J. and Cronstein,B.N. (
1996
) TNF/IL-1-inducible protein TSG-6 potentiates plasmin inhibition by inter-alpha-inhibitor and exerts a strong anti-inflammatory effect in vivo.
J. Immunol.
 ,
156
,
1609
–1615.
24
Mu,X.C., Staiano-Coico,L. and Higgins,P.J. (
1998
) Increased transcription and modified growth state-dependent expression of the plasminogen activator inhibitor type-1 gene characterize the senescent phenotype in human diploid fibroblasts.
J. Cell. Physiol.
 ,
174
,
90
–98.
25
McConnell,B.B., Starborg,M., Brookes,S. and Peters,G. (
1998
) Inhibitors of cyclin-dependent kinases induce features of replicative senescence in early passage human diploid fibroblasts.
Curr. Biol.
 ,
8
,
351
–354.
26
Yingling,J.M., Datto,M.B., Wong,C., Frederick,J.P., Liberati,N.T. and Wang,X.F. (
1997
) Tumor suppressor Smad4 is a transforming growth factor β-inducible DNA binding protein.
Mol. Cell. Biol.
 ,
17
,
7019
–7028.
27
Mucsi,I. and Goldberg,H.J. (
1997
) Dominant-negative SMAD-3 interferes with transcriptional activation by multiple agonists.
Biochem. Biophys. Res. Commun.
 ,
232
,
517
–521.
28
Williams-Hill,D.M., Duncan,R.F., Nielsen,P.J. and Tahara,S.M. (
1997
) Differential expression of the murine eukaryotic translation initiation factor isogenes eIF4A(I) and eIF4A(II) is dependent upon cellular growth status.
Arch. Biochem. Biophys.
 ,
338
,
111
–120.
29
Liang,H., Fairman,J., Claxton,D.F., Nowell,P.C., Green,E.D. and Nagarajan,L. (
1998
) Molecular anatomy of chromosome 7q deletions in myeloid neoplasms: evidence for multiple critical loci.
Proc. Natl Acad. Sci. USA
 ,
95
,
3781
–3785.
30
Nakabayashi,K., Ogata,T., Fujii,M., Tahara,H., Ide,T., Wadhwa,R., Kaul,S.C., Mitsui,Y. and Ayusawa,D. (
1997
) Decrease in amplified telomeric sequences and induction of senescence markers by introduction of human chromosome 7 or its segments in SUSM-1.
Exp. Cell Res.
 ,
235
,
345
–353.