Growth arrest in human T-cells is controlled by the non-coding RNA growth-arrest-specific transcript 5 (GAS5)

The control of T-cell proliferation and cell death is a key area in the cellular physiology of the immune system, as T-cells with required specificities must be stimulated to undergo many rounds of multiplication to produce an effective immune response to combat challenge by a pathogen. The events at the end of the multiplication phase, growth arrest and subsequent cell death, are equally important because the controlled termination of the immune response is vital for the efficiency of the immune system and for the prevention of immune system pathologies, i.e. leukemia, lymphoma and autoimmune disease.

Growth-arrest-specific transcript 5 (GAS5) was originally isolated from a subtraction cDNA library designed to identify genes that are upregulated in cells undergoing growth arrest (Schneider et al., 1988). Subsequent studies have shown that Gas5 transcripts have several patterns of alternate splicing but that the putative open reading frame is small. The open reading frame has been conserved only very poorly during relatively short periods of evolution, as demonstrated by a number of disruptions caused by frameshift mutations in several mouse strains (Muller et al., 1998), and by an interruption by a stop codon after the first 13 amino acids in rat Gas5 (Raho et al., 2000). It has therefore been deduced that any important biological activity of Gas5 must be mediated through introns that encode multiple small nucleolar RNAs (snoRNAs) (Smith and Steitz, 1998).

Gas5 has been independently identified by a differential-display strategy, which indicated that it is relatively overexpressed in a mouse strain that has increased susceptibility to hyperthermia-induced neural tube defects (Vacha et al., 1997). Other differential display studies have shown that Gas5 is overexpressed after benzo(a)pyrene treatment of rat vascular smooth muscle cells (Lu et al., 2002), and after amino acid deprivation of murine embryonal carcinoma cells (Fontanier-Razzaq, 2002). Stimulation of Gas5 expression on amino-acid deprivation of embryonal carcinoma has also been demonstrated by northern blotting (Fleming et al., 1998).

Recently, a fragment of Gas5 cDNA has been isolated from a retroviral cDNA expression library by using an unbiased functional screen for genes that control apoptosis in lymphocytes (Williams et al., 2006). Since the significance of changes in the level of GAS5 expression has not previously been investigated, we were prompted to analyse the functional effects of experimental manipulation of GAS5 transcript levels in both normal and leukemic human T-cells. These analyses clearly demonstrate a crucial role for GAS5 in the control of growth arrest, apoptosis and the cell cycle that has not previously been reported for any non-coding RNA.

Since GAS5 transcripts are subject to complex processing, we examined the effects of overexpression of several GAS5 cDNAs in the cloned CEM-C7 leukemic human T-cell line (Williams et al., 1998; Norman and Thompson, 1977). Transfection with GAS5 in the expression plasmid pCMVSPORT6 significantly inhibited the growth of the CEM-C7 cell population (Fig. 1A). This involved both increased apoptosis (Fig. 1B,C and the sub-G0 population in Fig. 1D) and slowing of the cell cycle with an increase in the proportion of cells in G₁ and a decrease in the proportion of cells in S phase (Fig. 1D). Transfection with the GAS5 expression construct also substantially reduced the colony-forming ability of the population (Fig. 1E). Significantly, these effects were also induced by GAS5 cDNAs which did not contain the putative open reading frame (GAS5-1B and GAS5-2B; Fig. 2), and were also observed in parallel studies using the human leukemic T-cell line Jurkat (Fig. 3).

Since overexpression studies do not necessarily indicate physiological importance, we used RNA interference (RNAi) to downregulate endogenous GAS5 expression in the CEM-C7 T-cell line (downregulation of 75-90% for all three small interfering RNAs (siRNAs) targeting GAS5 was confirmed by quantitative RT-PCR; Fig. 4A). The CEM-C7 T-cells treated with control siRNAs entered stationary phase and showed a subsequent reduction in viable cell density as normal (Fig. 4B). By sharp contrast, whereas densities of viable cells in cultures treated with the GAS5 siRNAs were barely distinguishable from the control populations during log phase, they diverged markedly from control populations undergoing growth arrest at 48-96 hours and continued to increase in cell density for several more days (Fig. 4B). All three GAS5 siRNAs thus delayed the onset of the stationary phase of growth. Analysis of the effects produced by the GAS5 siRNAs showed that all the GAS5 siRNAs increased both the total number of cells accumulated (Fig. 4B) and the proportion of cells in S phase normally observed at 96 hours (Fig. 4D, Table 1). This was accompanied by a clear reduction in the rate of apoptosis (Fig. 4C, and sub-G0 peak at 96 hours, Fig. 4D, Table 1).

The three GAS5 siRNAs substantially protected colony-forming ability after treatment with the glucocorticoid analogue dexamethasone (three- to sevenfold; Fig. 5) and after transient serum withdrawal (two- to ninefold; Fig. 5). Observations made with the siRNAs, taken together with the overexpression experiments, indicate that GAS5 expression is both necessary and sufficient for the normal growth arrest of the T-cell line.

Since the behaviour of cell lines often differs from that of the corresponding normal cells, we manipulated the endogenous level of GAS5 expression in normal lymphocytes that were stimulated with the mitogen phytohaemagglutinin (PHA) (Greaves and Bauminger, 1972). PHA stimulation produces several days of proliferation which is normally followed by growth arrest and increased apoptosis in the stationary phase. These effects were seen in the lymphocytes treated with control siRNAs (Fig. 6A). However, the PHA-stimulated cultures treated with GAS5 siRNAs showed reduced levels of endogenous GAS5 transcripts (Fig. 6D) which was accompanied by substantially increased viable cell numbers in the stationary phase, both in the presence (Fig. 6A) and in the absence of serum (Fig. 6B). Apoptosis was also reduced by the GAS5 siRNAs (Fig. 6C), as for the leukemic T-cell populations, showing that physiological GAS5 expression is also required for normal growth arrest in peripheral blood lymphocytes. Overexpression of GAS5 produced by transfection with GAS5-expression constructs, however, was sufficient by itself to induce growth arrest in normal peripheral blood lymphocytes (Fig. 6E), as found for the T-cell lines.

Since murine Gas5 has originally been isolated on the basis of its preferential expression in the growth-arrest phase of cell growth (Schneider et al., 1988), we examined the expression pattern of GAS5 splice variants in primary lymphocytes entering growth arrest. Three reverse transcriptase (RT)-PCR products of 250 bp, 450 bp and 600 bp were amplified, corresponding to GAS5-01, GAS5-3A and GAS5-1B, respectively (Fig. 2A). Fig. 6F shows that the expression of these alternatively spliced variants varies in different ways with time in culture. The abundance of GAS5-1B and GAS5-3A splice variants significantly reduced as culture density increased and reached saturation, whereas the proportion of GAS5 expressed as GAS5-01 was increased. Real-time RT-PCR was used to quantify the level of GAS5 transcripts in these samples relative to the level of GAS5 in the cells before the start of the experiment. In agreement with previous reports, our analysis shows that the level of GAS5 increased overall as the culture progressed, as the medium became exhausted and the cells reached saturation (Fig. 6G).

Since GAS5 cDNAs that do not contain the poorly conserved putative open-reading frame are functionally active in inducing growth arrest (Fig. 2), GAS5 appears to act as a non-coding RNA (Mattick, 2005). snoRNAs, such as those contained in the introns of GAS5 RNA, are involved in the processing of ribosomal RNA (Kiss, 2002) and may have other functions that remain to be identified (Matera et al., 2007); e.g. snoRNA HBII-52 has recently been shown to regulate alternative splicing of serotonin receptor 2C (Kishore and Stamm, 2006). Other snoRNAs have been implicated in the control of RNA editing (Vitali et al., 2005). Nevertheless, the profound and specific effects on the growth and survival of normal and leukemic T-cells demonstrated here could not have been anticipated.

Although this is the first report of any direct investigation of the functional effects of changes in GAS5 expression, previous studies have highlighted its potential importance. Expression of Gas5 has recently been reported to be subject to significant change in Ba/F3 cells transfected with oncogenic kinases that are associated with myeloproliferative disorders (Lelievre et al., 2006), and some radiation-induced thymomas carry rearrangements between Gas5 sequences and Notch 1 gene sequences (Tsuji et al., 2003). In addition, a snoRNA encoded at a different locus but, again, associated with an ORF without coding potential has been mapped to the chromosome breakpoint t(3;6)(q27;q15) in a human B-cell lymphoma (Tanaka et al., 2000). Most recently, genetic analysis of a mouse model of SLE (systemic lupus erythematosus) has indicated that Gas5 may well be involved its pathology (Haywood et al., 2006). This observation is of particular interest because it is widely acknowledged that genetic defects that affect the control of apoptosis may play a crucial role in the pathology of autoimmune diseases in general, and in the pathology of SLE in particular (Lorenz et al., 2000). In addition, the human chromosomal locus 1q25 at which the GAS5 gene is encoded has been associated with SLE in genetic studies in humans (Johanneson et al., 2002; Tsao, 2003; Tsao, 2004). The dramatic effects of changes in GAS5 expression on the control of T-cell apoptosis and proliferation we report here may well help to explain these observations, and should stimulate investigation, first, of the role of GAS5 in the physiological control of growth arrest and, second, of the potential involvement of GAS5 dysregulation in autoimmune disease, leukemias and lymphomas, and other cancers.

The apoptosis-sensitive cloned human T-leukemic cell line CEM-C7 CKM1 (Williams et al., 1998; Norman and Thompson, 1977) and Jurkat JKM1 cell line (Williams et al., 1998) were maintained in RPMI-1640 medium (Sigma) supplemented with 10% heat-inactivated foetal calf serum (FCS; Hyclone), 2 mM L-glutamine and 200 μg/ml gentamycin (Sigma), at 37°C in a 5% CO₂ humidified incubator.

Whole blood was collected into heparinized tubes from healthy volunteers. The blood was mixed with 1/5 volume of 6% (w/v) dextran solution and allowed to stand for 45 minutes at room temperature to allow sedimentation of the bulk of the erythrocytes (Hansel et al., 1990). The leukocyte-rich supernatant was decanted and layered onto 10-ml cushions of Lymphoprep (Axis-Shield; catalogue no. 1114547) and centrifuged at 800 g for 25 minutes at 18-20°C. The mononuclear cells, which accumulated at the interface between the plasma and the Lymphoprep, were then resuspended in cold PBS and recovered following centrifugation (1500 g, 30 minutes). Cells were resuspended in complete RPMI supplemented with 2.5 μg/ml PHA (phytohaemagglutinin; Sigma, catalogue no. L1668) and incubated at 37°C in a 5% CO₂ humidified incubator. After 24 hours, the non-adherent cells (depleted of macrophages) were harvested, resuspended at a density of 10⁶ cells per ml in complete RPMI medium containing 2.5 μg/ml PHA and incubated for 3-4 days at 37°C before being used for the experiments.

Cell viability was determined by nigrosin-exclusion analysis. The CaspaTag TM fluorescent caspase activity kit (Intergen; catalogue no. S7300-025) was used to detect active caspases in the cells as a marker for apoptosis, according to the manufacturer's instructions. This assay uses a fluorochrome inhibitor of caspases (FLICA reagent), fam-VAD-fmk (e.g. Bedner et al., 2000; Smolewski et al., 2001). Detection was performed using a Nikon Eclipse E400 fluorescence microscope. The ratio of FLICA-positive cells to total cells (labelled with a DNA dye) is given as a measure of the proportion of cells undergoing caspase-dependent apoptosis.

The different GAS5 ESTs in the expression vector pCMVSPORT6 or the vector alone, were introduced into CEM-C7 and Jurkat cells by electoporation [20 μg DNA at 248 V (CEM-C7), 293 V (Jurkat), 1050 μF in 0.4-cm cuvettes (Bio-Rad) at room temperature]. Efficiency of transfection for Jurkat and CEM-C7 cells was 60-70%. Expression of GAS5 was determined after 24 hours by real-time RT-PCR. Peripheral blood lymphocytes were cultured in complete RPMI medium supplemented with 2.5 μg/ml PHA for 5 days. On the day of the transfection, 5×10⁶ cells were centrifuged and resuspended in 100 μl of human T-cell nucleofector solution (Amaxa; catalogue no. VPA-1002) in the presence of 2 μg of the appropriate DNA (pmaxGFP, pCMVSPORT and different GAS5 ESTs). Cells were transfected using the Nucleofector (Amaxa; program T-23). Samples were removed from the cuvette immediately, transferred to 3 ml of Iscove's medium and incubated at 37°C. After 4 hours, the medium was changed to one that contains 2.5 μg/ml of PHA. Functional studies were carried out after 24 hours.

Preparation of cells (Hansel et al., 1990) and nuclear propidium iodide (PI)-staining procedure for cell cycle analysis was performed according to standard procedures (Kastan et al., 1991; White et al., 1990). Cells (1×10⁶) were suspended in 200 μl PBS, and fixed in 2 ml ice-cold 70% ethanol in 30% PBS. After a 30-minute incubation on ice, the cells were centrifuged for 5 minutes at 2000 rpm in a tabletop centrifuge and the supernatant aspirated. Cells were resuspended in 970 μl PBS, 3 μl RNase (DNAse-free; Sigma; catalogue no. R4642) and 40 μl of PI (1 mg/ml) and incubated for 30 minutes at 37°C before analysis using the MoFlo flow cytometer (Dako Cytomation).

Long-term survival of cells transfected with GAS5-expressing constructs or vector alone was assessed by the ability of the cells to form colonies in soft agar. An equal proportion of culture from each experimental condition was diluted in 5 ml Iscove's medium (Sigma) containing 20% heat inactivated FCS, 10% CEM-C7-conditioned medium and 0.3% noble agar (Difco), and plated in 60-mm dishes overlaid with 2.5 ml Iscove's complete medium containing 10% cell-conditioned medium. Colonies were counted following an incubation for 2-3-weeks at 37°C in 5% CO₂ and 95% air.

Three different GAS5 siRNAs (small interfering RNAs) were designed by Ambion [siRNAs ID: 290458 (GAS5siRNA2); 290460 (GAS5siRNA1) and 290459 (GAS5siRNA3); Reference sequence AF141346]. Negative-control siRNA (siRNA catalogue no. 4605) and GAPDH-targeting positive-control siRNA (Am 4605) were purchased from Ambion. All siRNAs were HPLC purified, annealed before use. To analyse siRNA-transfection efficiency, siRNA duplexes were labelled with Cy3 using the Silencer™ siRNA labelling kit (Ambion; catalogue no. 1632), following the manufacturer's instructions, and transfection efficiencies (fluorescently labelled cells after 48 hours) were 70-80%. On the day before transfection, cells were sub-cultured in RPMI supplemented with 10% FCS. On the day of transfection, 10⁶ Cells (CEM-C7, Jurkat or primary lymphocytes) were centrifuged and washed once in Optimem 1 (Invitrogen; catalogue no. 51985-026) before resuspension in 400 μl Optimem. Cells were then incubated with 20 nM or 100 nM siRNA duplex for 10 minutes at room temperature in a 0.4-cm electroporation cuvette. Cells were electroporated for 25 mseconds at 248 V (CEM-C7) or 293 V (Jurkat and primary lymphocytes) and 1050 μF using a Biorad Gene Pulser. Following electroporation, cells were incubated at room temperature for 20 minutes prior to transfer to six-well plates containing Iscove's medium (Sigma) supplemented with 2 mM glutamine and 20% heat-inactivated FCS. The analysis of specific silencing of GAS5 expression was carried out after 48 hours, using real-time RT-PCR.

Total RNA from 10⁷ cells was isolated using Trizol (GIBCO BRL; catalogue no. 15596-026) according to the manufacturer's instructions. 5 μg of RNA was reverse transcribed using SuperscriptTM II RNase H reverse transcriptase and random primers (Promega) according to the manufacturer's instructions (Invitrogen; catalogue no. 18064). One tenth of the reverse transcribed RNA was used in the PCR reaction. The GAS5 transcripts were detected using GAS5-fwd primer located in exon 9 (5′-GAAATGCAGGCAGACCTGTTATCC-3′) and GAS5-reverse primer located in exon 12 (5′-GACTACCTCAGAGTACCGTGTTCT-3′). PCR products were sub-cloned using the TOPO PCR cloning Kit (Invitrogen) and their identities verified by sequence analysis (MWG Biotech).

Real-time RT-PCR was performed using 2 μl of the cDNA prepared as described for RT-PCR, above (equivalent to 500 ng of the total RNA) and TaqMan MGB probes and primers specific to human GAS5 (exon 12; designed by Applied Biosystems, forward primer 5′-CTTCTGGGCTCAAGTGATCCT-3′; reverse primer 5′-TTGTGCCATGAGACTCCATCAG-3′; reporter 5′-CCTCCCAGTGGTCTTT-3′) with eukaryotic 18S rRNA as an endogenous control (Applied Biosystems; assay ID, Hs99999901_s1), according to the manufacturer's instructions. Quantification of GAS5 trancripts in cells transfected with GAS5-expressing constructs compared with pCMVSPORT6-transfected cells was determined using the comparative C_T method, by using pCMVSPORT6-transfected cells as calibrators. The ABI Prism 7000 sequence detection system was used to measure real-time fluorescence, and data analysis was performed using ABI Prism 7000 SDS software.

Data are presented as the mean ± standard error of the mean (± s.e.m.). Statistical significance was determined by analysis of variance using Origin 6.1; P<0.01 was considered statistically significant.

We thank the Wellcome Trust, BBSRC and the Breast Cancer Campaign for financial support and Alan P. Johnstone and Stanislav Glazewski for advice and constructive comments on the manuscript.