Introduction

Top Abstract Introduction Materials & Methods Results Discussion References

Interleukin-5 (IL-5) was initially discovered as a murine B-cell growth and differentiation factor (1). However, the major functions of IL-5 in many species, including humans, are most likely exerted on eosinophils. This cytokine plays an important role in the growth (2), differentiation (3), migration (4), survival (5), and priming for effector activities (6) of this cell type. Pertubations of the immune system by allergens (7) or parasites (8) induce bone marrow production and release of eosinophils and their preferential accumulation and persistence in tissues. The in vivo administration of monoclonal anti-IL-5 antibodies in these disease states and the amelioration of eosinophilic infiltration points to the strong causal relationship of IL-5 production and eosinophilia (7). Although mast cells (13) and eosinophils (9) have been reported to synthesize IL-5, the latter suggesting an autocrine role of this cytokine in the recruitment and activation of eosinophils, T cells are most likely the dominant source of this cytokine in vivo (7). In asthma, increased IL-5 mRNA and eosinophil recruitment have been associated with increases in the number and activation state of CD4⁺ T cells in bronchoalveolar lavage (14). In mouse models of allergic pulmonary inflammation, the preferential increase in eosinophils observed in the lungs has been attributed to T-cell production of IL-5 (15).

Unlike other cytokines, little is known regarding IL-5 gene regulation in vitro, particularly in untransformed, mature T cells. Murine and human antigen-specific T-cell clones can be classified into two major subsets, T_H1 and T_H2, based upon the cytokines produced following T-cell receptor stimulation (16, 17). T_H1 cells are distinguished by their production of IL-2, interferon- (IFN-), and lymphotoxin, while T_H2, but not T_H1, cells produce IL-4 and IL-5 (16, 17). Both T-cell subsets produce IL-3 and granulocyte/macrophage colony-stimulating factor (GM-CSF) (16, 17). Less differentiated T-cell clones that produce various combinations of T_H1 and T_H2 cell cytokines have also been described and are designated T_H0 (17). The cytokine production profiles of T_H1 and T_H2 cells parallel their effector functions during immune responses, in that T_H1 cells act as mediators of cell-mediated immune responses and T_H2 cells, by virtue of their synthesis of IL-4 and IL-5, induce allergic immune responses involving elevated IgE levels and selective differentiation of eosinophils (17).

Thus, a better understanding of the regulation of IL-5 gene expression in T cells could provide important insight into the possible dysregulation of this cytokine in diseases characterized by eosinophilia, such as asthma. The studies presented here investigate the in vitro regulation of IL-5 gene expression in human T cells stimulated with anti-()- CD3 monoclonal antibody (mAb) in the presence or absence of a co-stimulatory signal provided by -CD28 mAb, which mimics the effects of interaction of CD28 cell surface molecules on T cells with the counter-receptors, CD80 and CD86, expressed on antigen-presenting accessory cells (18, 19). Two human T-cell systems were used. First, with a T_H0 T-cell clone, it was possible to directly compare aspects of IL-5 gene regulation to IL-2, IL-3, IL-4, IFN-, and GM-CSF. Results obtained with these cells were confirmed and extended in a second cell type, in vitro differentiated human peripheral blood CD4 T_H2 cells, one likely to represent the IL-5-producing cell type in vivo. The results of these studies indicate several unique features of IL-5 gene regulation. First, IL-5 mRNA is significantly more stable than other cytokine transcripts in the presence or absence of CD28 stimulation. Second, in contrast to the other cytokines, initiation of IL-5 transcription has an absolute requirement for new protein synthesis.

	Materials and Methods

Top Abstract Introduction Materials & Methods Results Discussion References

T_H0 cells

The human T_H0 clone, SP-B21, was obtained from H. Yssel (DNAX, Palo Alto, CA) and has been described previously (20). This T-cell clone was grown by biweekly stimulation with irradiated (4,000 rads) allogeneic peripheral blood mononuclear cells (PBMC), irradiated (5,000 rads) Epstein-Barr virus-transformed lymphoblastoid cell line, JY, and purified phytohemagglutinin (Murex, Norcross, GA) (21), in Yssel's medium (21) supplemented with 1.5% human AB serum (Gemini Bioproducts, Calabasas, CA). Three days after each restimulation, the cultures were expanded in this medium containing 20 U/ml recombinant human IL-2 (Biosource, Camarillo, CA). Cells were fed with fresh medium plus IL-2 every 2 to 3 d until they returned to a resting state, at which time (about Day 14) they were restimulated. SP-B21 cells were used for experiments 8 to 10 d after stimulation with allogeneic PBMC and EBV-LCL cells. Yssel's medium was used in all experiments using SP-B21 cells.

CD4 T_H2 Long-term Cultured Cells (LTC)

Peripheral blood was obtained from healthy, informed volunteers at the Schering-Plough Research Institute. Freshly drawn heparinized human blood was diluted 1:1 with Hanks' balanced salt solution (HBSS) (Fisher, Springfield, NJ) and PBMC isolated by Ficoll Hypaque (Pharmacia, Uppsala, Sweden) density gradient centrifugation. After three washes with HBSS, cells were counted and adjusted to 1 × 10⁷ cells/ml in phosphate-buffered saline (PBS) containing 2% fetal calf serum (FCS) (Gibco BRL, Gaithersburg, MD) for subsequent isolation of CD4⁺ T cells or CD4⁺ CD45RA^hiCD45RO^lo T cells (naive CD4⁺ T cells). Unfractionated CD4⁺ T cells were isolated from the PBMC with Dynabeads^® M450 (Dynal Corp., Lake Success, NY), at a 3:1 bead-to-cell ratio. CD4⁺ cells were then removed from the Dynabeads by addition of a 1:10 dilution of Detachabead^® M450 solution. CD4⁺ T-cell separation and bead detachment were performed according to the manufacturer's specifications. The cells were resuspended to a final concentration of 1 × 10⁶ cells/ml in complete RPMI 1640 medium (cRPMI) containing 10% FCS, 10 mM Hepes, 2 mM L-glutamine, 50 U/ml penicillin, 50 mg/ml streptomycin (Fisher Scientific, Springfield, NJ) and 5 × 10⁵ M 2-mercaptoethanol (Sigma Chemical Co., St. Louis, MO). For preparation of naive CD4⁺ T cells, CD4⁺/CD45RO Subset Columns were used (R&D Systems, Minneapolis, MN). FACS analysis demonstrated that CD4⁺ T cells purified by Dynabeads were routinely > 98% pure and > 98% viable. Cells purified by Subset Columns were > 95% CD4⁺ and > 91% CD4⁺/CD45RO and > 98% viable.

CD4⁺ T cells or CD4⁺ naive T cells, prepared as described above, were grown for a total of 20 or 27 d on alternating 3-d stimulation and rest cycles. The cells were stimulated with immobilized -CD3 monoclonal antibody (-CD3 mAb) (10 µg/ml, clone UCHT1; Pharmingen, San Diego, CA) with soluble -CD28 mAb in the presence of 20 U/ml rhIL-2 (Biosource), 10 ng/ml rhIL-4 (Schering-Plough, Kenilworth, NJ) and 10 µg/ml -IFN- antibody (R&D Systems) in cRPMI medium. During the rest phase, the cells were cultured in cRPMI with rhIL-2 and rhIL-4 as above and with 1 µg/ml -IFN- antibody. During both the stimulation and rest phases, the CD4⁺ cells were added to the uncoated or -CD3 mAb coated plates at 4 × 10⁵ cells/ml, 4 ml/well in six-well tissue culture plates (Fisher Scientific) and incubated at 37°C in 5% CO₂. Cells were counted every 3 d to monitor cell expansion. In general, over the course of the 20 d, the cells expanded in the range of one to seven thousand fold. After the growth period, FACS analysis indicated that the cells were > 99% memory CD4⁺CD45RO^hiCD45RA^lo. Cells were used for experiments after the rest cycle on either Day 20 or 27.

T-Cell Stimulation for Cytokine Production

CD4⁺ T cells grown for 20 or 27 d (CD4 LTC) as described above, or SP-B21 cells, at 8 to 10 d after stimulation with irradiated allogeneic PBMC and JY cells, were harvested, washed, and resuspended at 1 × 10⁶/ml in cRPMI or Yssel's medium, respectively. To induce cytokine secretion, 100 µl of cells was added to washed -CD3- coated microtiter plates (10 or 1 µg/ml for CD4 LTC or SP-B21 cells, respectively). Where indicated, -CD28 mAb was added to a final concentration of 1 µg/ml. For all groups, a total of 200 µl was used per well. All controls and experimental groups were tested in quadruplicate. The cultures were incubated for 24 to 48 h at 37°C, 5% CO₂, at which time supernatants were harvested for enzyme-linked immunosorbent assay (ELISA). Primary and secondary antibodies for IL-5 and IL-4 quantitation by ELISA were obtained from Pharmingen and used according to their protocol; the detection limit for these assays was approximately 50 pg/ml. IFN- was measured using an IFN- Cytoscreen Immunoassay Kit (Biosource); the limit of detection was 4 pg/ml, according to the manufacturer's specifications.

T-Cell Stimulation for Determination of Cytokine mRNA Levels

Six-well Falcon tissue culture plates were coated overnight with -CD3 mAb at 10 or 1 µg/ml in PBS as indicated. After washing the plates three times with cold PBS, washed SP-B21 cells or CD4 LTC cells were added at 1 × 10⁶/ml, 4 ml/well, four to six wells/experimental group. Where indicated, soluble -CD28 mAb was 1 µg/ml final concentration. At the indicated times, specific inhibitors were added to the cultures. Cycloheximide (CHX) (Sigma Chemical Co.) was diluted in PBS and used at a final concentration of 10 µg/ml.

Steady-state RNA Analysis and Cytokine mRNA Half-life Determination

For measurement of mRNA half-life (t_1/2), actinomycin D (ActinoD) (Sigma) or cyclosporin A (CSA) (Sandoz, Whippany, NJ) were dissolved in ethanol and added to the indicated cultures at a final concentration of 2.5 µg/ml or 1 µg/ml, respectively, at the indicated time. In preliminary experiments, these concentrations of ActinoD and CSA (Figure 4, legend) were found to block cytokine transcription when added at the time of T-cell stimulation and to be nontoxic during the time period over which mRNA half-life was measured. Cells were harvested at intervals following transcription termination, and supernatants stored at 20°C prior to ELISA. Total cellular RNA was isolated from cultured cells by using Tri Reagent^TM according to the manufacturer's protocol (Molecular Research Center, Inc., Cincinnati, OH). RNA was quantitated by absorbance at 260/280 nm. Denatured RNA (10 to 15 µg) was electrophoresed in a 1.2% agarose/2.2 M formaldehyde gel. In each experiment, two replicate gels were prepared. After staining with ethidium bromide and photographing using a UV transilluminator, RNA was transferred to Genescreen Plus (NEN, Boston, MA) in 10× saline sodium phosphate EDTA (SSPE) and UV crosslinked. Blots were prehybridized overnight and then hybridized for 18 to 24 h in 5× SSPE, 5× Denhardt's, 1% sodium dodecyl sulfate (SDS), 50% formamide, and 100 µg/ml salmon sperm DNA (Sigma Chemical Co.) at 42°C. cDNA probes were labeled with (-³²P)dCTP (> 3,000 Ci/mmol; NEN) using a random primer kit (Boehringer Mannheim, Indianapolis, IN) to a specific activity of  2 × 10⁸ cpm/µg; 2 × 10⁶ cpm per milliliter of hybridization solution was used. After high stringency washes, blots were exposed to Kodak XAR-5 film with enhancing screens (Sigma). These blots were sequentially hybridized, quantitated, stripped, and rehybridized with the final hybridization to glyceraldehyde phosphate dehydrogenase (GAPDH) or -actin cDNA probes. Bands were quantitated by image analysis using the Betascope 603 Blot Analyzer (Betagen, Waltham, MA) using exposure times estimated to be in the linear range of signal strength appropriate to each gene. Normalized RNA values are expressed as [(cytokine counts  background)  (-actin or GAPDH counts  background)] × 1,000. Half-lives were calculated from linear regression best-fit semi-log plots of the natural log of the concentration of RNA versus the time of culture with ActinoD or CSA. t_1/2 values were derived according to the equation t_1/2 = ln2  k_decay, where k_decay = the slope of linear portion of the semi-log plots for each gene (22).

View larger version (19K): [in this window] [in a new window]   View larger version (76K): [in this window] [in a new window]   Figure 4.   Measurement of cytokine mRNA half-lives following CSA addition 3 h after stimulation of CD4 LTC. CD4 LTC isolated by Dynabeads were grown for 20 d as described in MATERIALS AND METHODS. On Day 20, the ratio of IL-5 to IFN- secreted was 7.5 following -CD3 mAb stimulation. CD4 LTC were stimulated with immobilized -CD3 mAb (10 µg/ml) with soluble -CD28 mAb (1.0 µg/ml) for 3 h, at which time CSA was added (t = 0 h). Cells were harvested at the indicated times for isolation of total RNA. Results are shown as percent remaining mRNA versus time (A) and were derived from image analysis of blots (B), subsequently exposed to film as shown. Percent remaining cytokine mRNA was calculated as described in Figure 2. t1/2 values (Table 3, CSA, -CD28) were derived as described in MATERIALS AND METHODS. The results shown are representative of two similarly conducted experiments with different donor CD4 LTC. CSA added to CD4 LTC cells 10 min before stimulation for 3 h (at which time mRNA was harvested) completely inhibited IL-3 (data not shown), IL-4, IL-5, and GM-CSF but not GAPDH mRNA levels (3 h, lane 2) compared with cells similarly treated in the absence of CSA (0 h, lane 1).

View larger version (21K): [in this window] [in a new window]   Figure 2.   Measurement of cytokine mRNA half-lives following ActinoD addition 6 h after stimulation of SP-B21 cells. SP-B21 cells were stimulated with immobilized -CD3 (1.0 µg/ml) for 6 h, at which time ActinoD was added (t = 0 h). Cells were harvested at the indicated times for isolation of total RNA. Results are shown as percent remaining mRNA versus time and were derived from image analysis of blots, subsequently exposed to film. Percent remaining cytokine mRNA for a given time was calculated as follows: [(cytokine counts  background counts at t = x h)  (cytokine counts  background counts at t = 0 h)]. t1/2 values (Table 2, ActinoD, 6 h) were derived as described in MATERIALS AND METHODS. The results shown are representative of two similarly conducted experiments and one experiment in which 10 µg/ml -CD3 was used.

View larger version (24K): [in this window] [in a new window]   Figure 3.   Measurement of cytokine mRNA half-lives following CSA addition 3 h after stimulation of SP-B21 cells. SP-B21 cells were stimulated with immobilized -CD3 (1.0 µg/ml) for 3 h, at which time CSA was added (t = 0 h). Cells were harvested at the indicated times for isolation of total RNA. Results are shown as percent remaining mRNA versus time and were derived from image analysis of blots, subsequently exposed to film. Percent remaining cytokine mRNA was calculated as described in Figure 2. t1/2 values (Table 2, CSA, 3 h) were derived as described in MATERIALS AND METHODS. The results shown are representative of two similarly conducted experiments. CSA added to SP-B21 cells 10 min before stimulation with immobilized -CD3 for 3 h (at which time mRNA was harvested) inhibited IL-2, IL-3, IL-4, IL-5, IFN-, GM-CSF, and GAPDH mRNA levels 96, 97, 95, 87, 78, 95, and 8%, respectively, compared with cells similarly treated in the absence of CSA.

RNA Analysis: Nuclear Run-on Assay

Unless otherwise noted, SP-B21 cells were stimulated for 3 h with -CD3 in the presence and absence of 10 µg/ml CHX under the culture conditions described above. Two sets of each experimental group were established, one for the isolation of nuclei and the other for simultaneous isolation of total RNA for Northern analysis. The transcription in freshly isolated nuclei was carried out essentially as described elsewhere (23). Nuclei from 1 × 10⁸ SP-B21 cells were used per experimental group. Cells were scraped from the culture plate, washed once in cold PBS, and lysed in 10 mM Tris-HCl (pH 7.4), 10 mM NaCl, 3 mM MgCl₂, and 0.5% NP-40 (Sigma Chemical Co.), harvested by centrifugation, washed, and resuspended in 200 µl of 50 mM Tris-HCl (pH 8.3), 40% glycerol, 5 mM MgCl₂, and 0.1 mM EDTA. To this was added an equal volume of 2× reaction buffer (10 mM Tris-HCl [pH 8.0], 5 mM MgCl₂, 0.3 M KCl, 5 mM dithiothreitol, and 1 mM each ATP, CTP, and GTP), 2 µl RNAsin (Promega, Madison, WI), and 100 µCi -[³²P]UTP (NEN). The reaction was incubated at 30°C for 30 min. Transcripts were isolated using Trisolv (Biotecx, Houston, TX), followed by extraction with phenol/chloroform/isoamyl alcohol and precipitation at 20°C with an equal volume of isopropanol. RNA pellets were resuspended in 10 mM Tris-HCl (pH 7.4), 10 mM EDTA, 0.2% SDS, 0.6 M NaCl, and 5× Denhardt's (Sigma Chemical Co.). For each treatment, equal counts per minute (approximately 0.5 to 1.5 × 10⁶ cpm) were hybridized in 500 µl to the indicated purified cDNA inserts (250 ng/slot blot), which had been UV crosslinked to nylon membranes. Slot blots were prehybridized overnight and hybridized for 72 h at 65°C and washed, as described above for Northern analysis.

cDNA Probes

The following human cDNA purified inserts were used as probes in Northern analysis and nuclear run-on experiments: -actin, #78554; GM-CSF, #57595; IL-4, #57593; IL-3, #59399; and GAPDH, #57091 (ATCC, Rockville, MD). The following human cDNA inserts of the designated size were excised from the pCD-SR vector with BamHI and were originally obtained from K. Moore (DNAX, Palo Alto, CA): IL-2 (1,021 bp), IL-5 (1,025 bp), IFN- (1,402 bp).

	Results

Top Abstract Introduction Materials & Methods Results Discussion References

Cytokine production characteristics of the SP-B21 clone and the long-term cultured peripheral blood CD4⁺ cells (CD4 LTC) are shown in Table 1. Cytokine production was stimulated through the T-cell receptor with immobilized -CD3 mAb in the presence or absence of soluble -CD28 mAb. When stimulated with -CD3 mAb, SP-B21 cells produced large amounts of IL-5 (range: 5.8 to 36 ng/ ml) and lesser amounts of IFN- (0.26 to 1.3 ng/ml), such that the mean ratio of IL-5:IFN- was 33 in seven independent experiments. In three experiments in which cytokine levels in cultures co-stimulated by -CD28 mAb were compared with levels in cultures stimulated with -CD3 mAb alone, only a small increase (1.3-fold) was observed in the amount of IL-5 or IFN- produced in the co-stimulated cultures. SP-B21 cells also produced IL-2 in response to -CD3 mAb stimulation, and this was increased about threefold by optimal co-stimulation (data not shown). Although this clone produces high levels of IL-5, its production of IL-2 and IFN- classifies it as of the T_H0 phenotype (16, 24).

CD4 LTC produced significant amounts of IL-5 (range: 0.36 to 5.3 ng/ml) and lesser amounts of IFN- (0.08 to 0.46 ng/ml) when stimulated with -CD3 mAb alone, such that the mean ratio of IL-5:IFN- was 19 in seven independent experiments. In CD4 LTC cultures co-stimulated with -CD28 mAb, the ratio of IL-5 to IFN- was 34, reflecting a less than twofold increase versus -CD3-stimulated cultures. No detectable IL-2 was found in the supernatants of CD4 LTC stimulated with a broad concentration range of both -CD3 and -CD28 mAb. The ratio of IL-5: IFN- produced, as well as the lack of IL-2 production, is indicative of the differentiation of these cells to a T_H2 phenotype, since freshly isolated naive CD4⁺ T cells do not produce IL-5 (25). While co-stimulation had a larger effect on IL-5 and IFN- production by CD4 LTC (four-fold and 1.8-fold, respectively) than SP-B21 cells (1.3-fold), the increase in the ratio of IL-5 to IFN- in both co-stimulated CD4 LTC and SP-B21 cells versus -CD3 mAb- stimulated cultures was less than twofold. Thus, SP-B21 cells and CD4 LTC cells are similar in that both IL-5 and IFN- are only weakly increased by co-stimulation.

Kinetics of Cytokine mRNA Expression in SP-B21 Cells

Initial experiments were conducted to determine the time course of IL-5 gene expression in SP-B21 cells, in response to stimulation by plate bound -CD3 mAb. This kinetic information was necessary to design and interpret subsequent experiments aimed at estimating the transcription and degradation rates of IL-5 and their modulation by different modes of T-cell stimulation or drug treatment. Two experiments are shown in which the kinetics of cytokine mRNA expression was measured by Northern analysis in SP-B21 cells stimulated through the T-cell receptor with -CD3 mAb for various times from 0.5 to 5.0 h (Figures 1C and 1D) or from 3 to 24 h (Figures 1A and 1B). Maximal mRNA expression for IFN- (Figures 1A and 1C) and GM-CSF (Figures 1B and 1C) was seen 3 h after stimulation, while that for IL-3, IL-4, and IL-5 was 4 to 5 h, 3 to 5 h, or 4 to 5 h, respectively (Figures 1A and 1C). Maximum IL-2 mRNA levels were detected 2 h after stimulation (Figure 1D). A rapid loss of steady-state mRNA levels of all cytokines was also observed within 3 h of the time of maximal gene expression (Figures 1A, 1B and 1D). IL-5 mRNA levels declined approximately 50% within 2 h from a peak at 5 h after stimulation. Measurable levels of IL-4, IL-5, IFN- and GM-CSF mRNA were seen at 24 h, with IL-5 having decreased the least relative to its respective time point of maximal expression (Figure 1A). This time course of expression indicated that IL-5 mRNA, like other cytokines, was rapidly induced and degraded following T-cell receptor stimulation.

View larger version (21K): [in this window] [in a new window]   Figure 1.   Kinetics of cytokine mRNA expression in SP-B21 cells. (A) Cells were stimulated with immobilized -CD3 for 3, 5, 6, 7, 8, 10, 13, and 24 h, at which time the cells were harvested for isolation of total RNA. Northern analysis was done as described in MATERIALS AND METHODS. Normalized RNA (y-axis) was calculated as follows: [(cytokine counts  background counts)  (GAPDH counts  background counts)] × 1,000. Background counts were determined by measuring counts at three different locations on the blot using a Betascope. (B) Because of the different degree of expression, GM-CSF is depicted separately from the other genes. (C ) Cells were stimulated for 0.5, 1, 1.5, 2, 3, 4, and 5 h and analysis was done as described in panel A. (D) Because of the different degree of expression, IL-2 is depicted separately from the other genes. The results shown are representative of other time course experiments (data not shown) and contained in experiments shown in Figures 2-5 (samples to which no inhibitors were added).

View larger version (46K): [in this window] [in a new window]   Figure 5.   Protein synthesis inhibition completely blocks IL-5 but not IL-2, IL-3, IL-4, IFN-, or GM-CSF gene expression. (A) Cycloheximide was added to cells 10 min before addition of SP-B21 cells to -CD3-coated plates. Three hours after stimulation, cells were harvested for the isolation of total RNA. (B) CHX was added to SP-B21 cells 10 min before addition to -CD3-coated plates (t = 0 h) or 1 h after stimulation. Cells were harvested at 2 h after stimulation. Northern analysis was done as described in MATERIALS AND METHODS. Addition of CHX to unstimulated cells did not induce gene expression of any cytokine gene.

Measurement of Cytokine mRNA Half-life

The process of mRNA degradation in SP-B21 cells was directly assessed by measuring the t_1/2 of IL-5 mRNA following the addition of the commonly used transcriptional inhibitor, ActinoD. ActinoD, a DNA intercalating agent, is a global inhibitor of new transcription mediated by RNA polymerase I, II, and III (22). Transcription was terminated by the addition of ActinoD either 3 h (Table 2), 5 h (Table 2) or 6 h (Figure 2, Table 2) after -CD3 mAb stimulation of SP-B21 cells. RNA was harvested at multiple time points after termination of transcription. In this manner, several estimates of mRNA half-life were determined, which varied because of the differing times of transcription termination in relationship to the time course of gene expression (Figure 1). Because of the early peak and decline of IL-2 mRNA levels (Figure 1D), IL-2 mRNA was not always detected in these studies. Half-life values are shown in Table 2 and, for some time points of transcription termination, graphic display (Figures 2, 3, and 4A) of image analysis quantitation or autoradiographs (Figure 4B) are also shown.

When transcription was terminated by ActinoD addition 3 h after -CD3 mAb stimulation of SP-B21 cells (Table 2), t_1/2 values for IL-3, IL-4, IFN- and GM-CSF were estimated to be 1.2, 2.4, 4.0, and 1.0 h, respectively. In contrast, the t_1/2 of IL-5 mRNA was > 6 h. Assessment of time points later than 6 h to establish the t_1/2 of IL-5 was not possible because of ActinoD-induced toxic effects. In order to obtain a t_1/2 value for IL-5, studies were done in which ActinoD was added at later times after stimulation (5 and 6 h after stimulation), when, based on the kinetics of IL-5 mRNA expression (Figure 1A), it was likely that degradation was greater than synthesis. For several cytokine genes, the later the addition of ActinoD after stimulation, the shorter the t_1/2 value, indicative of changing rates of mRNA degradation over the time course of gene expression. IL-4, IFN-, and GM-CSF transcript stability (Table 2) was progressively decreased as transcription was stopped at progressively later times (3, 5, or 6 h). The exception to this progressive change in the rate of degradation was IL-3, whose degradation rate did not differ over time. In contrast, the stability of IL-5 transcripts was significantly shorter when transcription was stopped at 5 versus 3 h, respectively, and then increased between 5 and 6 h after transcription was stopped (Table 2). Importantly, IL-5 consistently displayed the longest t_1/2 values. When transcription was terminated by the addition of ActinoD 5 or 6 h after stimulation with -CD3 mAb (Table 2), the t_1/2 value for IL-5 mRNA was at least 1.7 h longer than the other cytokine gene transcripts.

Co-stimulation of T cells through ligation of CD28 on T cells has been found to increase the stability of cytokine mRNA (19). Although co-stimulation of SP-B21 cells with -CD28 mAb, only weakly increased secreted cytokine levels (1.3-fold, Table 1) as compared with -CD3 mAb stimulation alone, it was of interest to determine if the highly stable IL-5 transcripts observed following -CD3 mAb stimulation could be further stabilized by CD28 co-stimulation and whether IL-5 mRNA remained the most stable transcript. SP-B21 cells were stimulated with immobilized -CD3 mAb in the presence of -CD28 mAb and transcription was terminated 5 h after stimulation with ActinoD, in parallel with SP-B21 cells stimulated with -CD3 mAb alone (Table 2, 5 h). All cytokine mRNA transcripts tested were stabilized to a small degree (increases were 44, 44, 0, 5, and 22%, respectively, for IL-3, IL-4, IL-5, IFN-, and GM-CSF). IL-5 was more stable (by 1.5 h) than the other cytokine transcripts examined (data not shown).

For all studies in which ActinoD was used to stop transcription (Table 2), the decay curves for IL-3, IL-4, IFN-, and GM-CSF displayed biphasic decay with an initial faster decay that occurred within the first 1 to 2 h followed by a slower secondary decay phase (Figure 2). This is indicative of a slowing of the rate of decay possibly due to the termination of transcription of other gene products required for mRNA degradation. A slowing of IL-5 mRNA degradation did not occur until 4 h after ActinoD addition (Figure 2). It was possible the long t_1/2 of IL-5 mRNA observed in these experiments, as well as the biphasic decay observed with the other cytokine genes (Figure 2), was due to the inhibition of transcription by ActinoD of other genes, which normally function to regulate the degradation of cytokine transcripts.

To address this possibility with a more specific inhibitor, additional mRNA t_1/2 studies were done using CSA, as has been done by others (28). CSA indirectly blocks the transcription of cytokine genes induced by T-cell receptor activation, such as by -CD3 mAb stimulation. By inhibiting the Ca⁺-calmodulin-dependent phosphatase, calcineurin (29), CSA prevents the cytoplasmic to nuclear translocation of NF-AT, a transcription factor involved in cytokine gene activation. Thus, the nonspecific, indirect effects of ActinoD are avoided. In addition, cytotoxicity, judged by decreased integrity of 18S and 28S rRNA, seen in cells incubated with ActinoD for longer than 6 h, was not observed with CSA (data not shown). In -CD3 mAb- stimulated SP-B21 cells, CSA (1 µg/ml) added before stimulation was found to be a potent inhibitor of cytokine gene expression (Figure 3, legend) and thus a useful tool for mRNA stability studies. In addition, nuclear run-on assays demonstrated that transcription initiation of IL-2, IL-3, IL-4, IL-5, GM-CSF, and IFN- in CSA (1.0 µg/ml)-pretreated, stimulated SP-B21 cells was equivalent to that in unstimulated cells and completely inhibited as compared with -CD3-stimulated SP-B21 cells (data not shown). Others have demonstrated the blockade of IL-5 gene transcription by CSA in human PBMC (30).

To terminate transcription for assessment of cytokine mRNA stability, CSA was added to SP-B21 cells 3 h after stimulation with -CD3 (Figure 3, Table 2). For each of these cytokine genes, the t_1/2 value determined by transcription termination with CSA added at 3 h after stimulation (Table 2) was considerably shorter than that determined with ActinoD added at 3 h after stimulation (Table 2). Faster degradation of IL-5 mRNA in the presence of CSA (t_1/2  6.5 h) as compared with ActinoD (t_1/2 > 6 h) was less obvious than for the other cytokine genes. Importantly, the t_1/2 value for IL-5 was at least 5 times greater than that for the other cytokine genes.

The studies presented above were done in a single T_H0 clone, in which the relevance of features of IL-5 gene regulation to IL-5-producing T_H2 cells in vivo is unknown. Consequently, additional stability studies, using CSA as the transcription inhibitor, were undertaken in peripheral blood CD4 LTC grown under conditions which drive the differentiation of a T_H2 cytokine phenotype (Table 1). In these cells, the genes for two cytokines, IL-4 and IL-5, diagnostic for T_H2 cells were measured as well as IL-3 and GM-CSF, cytokines produced by T_H1, T_H2, and T_H0 cells (16, 24). The kinetics of cytokine gene expression in -CD3 mAb-stimulated CD4 LTC measured by Northern analysis with hourly time points was similar but not identical to that in SP-B21 cells. IL-3, IL-4, and GM-CSF mRNA levels were maximal between 3 and 4 h, whereas IL-5 mRNA levels were highest between 9 and 18 h (data not shown).

Although t_1/2 values identical to those derived from SP-B21 cells were not found, similar conclusions can be made. As for the SP-B21 cells, CSA (1 µg/ml) completely inhibited de novo transcription of IL-3, IL-4, IL-5, and GM-CSF when added prior to stimulation of CD4 LTC with -CD3 mAb (data not shown). The stability of IL-5 mRNA (t_1/2 > 6 h) in -CD3-stimulated CD4 LTC, to which CSA was added 3 h after stimulation, was longer by at least 5 h than that of IL-3, IL-4, or GM-CSF (Table 3). At 6 h after the addition of CSA to CD4 LTC, the last time point measured, the IL-5 mRNA level was 100% of the level at the time of transcription termination. Thus, in CD4 LTC, IL-5 transcripts (t_1/2 > 6 h) were more stable than IL-3, IL-4, and GM-CSF transcripts, by at least 5 h (Table 3).

The effect of -CD28 mAb co-stimulation on mRNA stability in CD4 LTC was also determined (Figure 4A, Table 3). CSA was used to terminate transcription and was found to completely block de novo transcription of IL-3 (data not shown), IL-4, IL-5, and GM-CSF (Figure 4B, lane 3) in -CD28 mAb-co-stimulated CD4 LTC. As compared to stimulation with -CD3 mAb alone, -CD28 mAb co-stimulation had little effect (less than 40% change) on the stability of cytokine mRNA (Table 3). IL-5 transcripts were significantly more stable than IL-3, IL-4, and GM-CSF transcripts. As in SP-B21 cells, for all cytokines tested in CD4 LTC, t_1/2 values, derived by the termination of transcription with CSA (Table 3), were less than those determined with ActinoD (data not shown) and the later slower phase of mRNA decay seen with ActinoD was not observed except where residual transcript levels were less than 20% (Figure 4A). Importantly, IL-5 gene expression in SP-B21 cells and CD4 LTC differs from that of the other cytokines tested in that it displays a long t_1/2, suggesting differences in the mechanism of degradation. Further examination of the nature of these mechanisms was done primarily in SP-B21 cells.

In several experiments reflected in Figures 3 and 4A, following an initial decrease in IL-5 mRNA levels at 1 h after the addition of CSA, an increase in IL-5 mRNA levels occurred within the next 1 to 2 h. This may reflect a secondary wave of transcription, independent of -CD3 mAb stimulation, which was not observed with other cytokine mRNAs. This may be mediated by IL-2, produced early after cell stimulation (Figure 1D), which is known to induce IL-5 in a CSA-resistant manner (30, 31).

Requirement of Protein Synthesis for Induction of IL-5 Gene Expression

Decay of most eukaryotic mRNAs including cytokine mRNAs is dependent on translation initiation (32). In view of this, the effect of the translation inhibitor, CHX, on IL-5 gene expression was evaluated at different times in relationship to -CD3 mAb stimulation of SP-B21 cells. As shown in Figure 5A, CHX added to SP-B21 cells prior to -CD3 mAb stimulation differentiated three groups of cytokines that were totally, partially, or not at all inhibited by this treatment. The accumulation of IL-5 steady-state mRNA measured 3 h after stimulation was totally blocked by the addition of CHX 10 min prior to stimulation. Quantitation of radioactive counts from 10 experiments consistently demonstrated an absence of IL-5 mRNA. In contrast, IL-3 and IL-4 mRNA levels were significantly but not totally inhibited by this treatment (Figure 5A). The mean % inhibition ± SD of seven experiments for IL-3 and IL-4 was 32 ± 29 and 66 ± 25, respectively. Lastly, both GM-CSF and IFN- mRNA levels were not consistently or greatly inhibited by the addition of CHX prior to stimulation (Figure 5A). In the majority of experiments (six of seven), GM-CSF mRNA levels were minimally inhibited (< 17% inhibition) or were elevated (81% increase). Similar results were observed for IFN-. A complete lack of IL-5 steady-state mRNA was also observed in CHX pretreated -CD3 mAb-stimulated CD4 LTC, while IL-4 and GM-CSF mRNA were only partially inhibited (data not shown).

Because of the kinetics of IL-2 mRNA expression (Figure 1D), the effects of CHX on IL-2 gene expression could not be determined with cells harvested at 3 h (Figure 5A). Therefore, studies were done in which cells were harvested 2 h after stimulation (Figure 5B). A significant but not complete reduction (60%) in IL-2 mRNA levels was seen when CHX was added 10 min before stimulation (0 h). IL-5 steady-state mRNA was totally blocked by CHX added at 0 h (Figure 5B). GAPDH mRNA levels demonstrated that this was not a toxic concentration of CHX (Figures 5A and 5B). Addition of CHX to unstimulated SP-B21 cells did not induce cytokine gene expression, and measurement by ELISA of cytokines secreted by -CD3 mAb-stimulated cells showed that CHX was effective in blocking protein synthesis (data not shown). Other protein synthesis inhibitors (anisomycin and puromycin) gave similar results (data not shown). Thus, IL-5 mRNA accumulation does not occur in the absence of protein synthesis, in contrast to that of IL-2, IL-3, IL-4, GM-CSF, and IFN-.

With the progressive delay in the time of addition of CHX to -CD3-stimulated SP-B21 cells, increases in IL-5 steady-state mRNA were observed (Figure 6). Addition of CHX 1 h after -CD3 stimulation resulted in minimal but measurable IL-5 mRNA when harvested 2 h (Figure 6, Gp 6; Figure 5B) or 4 h (Figure 6, Gp 7) after stimulation. When addition of CHX occurred 2 h (Gp 8) or 3 h (Gp 9) after stimulation and cells were harvested 4 h after -CD3 stimulation, no decrease in IL-5 mRNA levels compared with untreated stimulated cells was found (Gp 3). Importantly, unlike what has been described for other cytokines, no evidence of superinduction of IL-5 mRNA was seen.

View larger version (32K): [in this window] [in a new window]   Figure 6.   Early requirement for protein synthesis for IL-5 mRNA accumulation. CHX (10 µg/ml) was added to SP-B21 cells 10 min before addition to -CD3-coated plates (0 h) or 1, 2, or 3 h after stimulation of SP-B21 cells. Cells were harvested for the isolation of total RNA at 2, 4, or 6 h after stimulation. Northern analysis was done as described in MATERIALS AND METHODS and quantitated as in the legend to Figure 1. ns = not significantly different than background counts. Results shown are representative of three similar experiments.

These results indicate an early requirement for protein synthesis in IL-5 gene expression. This requirement may exist in at least two pathways: either a pathway important to initiation of transcription or, alternatively, a pathway critical to IL-5 mRNA stabilization. To distinguish between these possibilities, nuclear run-on assays, which measure the level of transcription independent of mRNA degradation, transcript processing, or nuclear-to-cytoplasm translocation, were conducted. The rate of transcription was measured in unstimulated SP-B21 cells or in cells stimulated in the presence or absence of CHX. The time point chosen for nuclear RNA isolation was 3 h after stimulation, presumably representing the phase of active IL-5 transcription, based upon the kinetics of steady-state mRNA levels (Figures 1A and 1C).

Two representative nuclear run-on assays using SP-B21 cells are shown in Figure 7. Transcription of the constitutively expressed genes, -actin and GAPDH, was detected as expected in all treatment groups. A small upregulation of -actin was observed in cells stimulated in the presence of CHX, as has been described by others (33). In contrast, with the exception of GM-CSF, no cytokine transcription above background signals was detected in unstimulated SP-B21 cells. This may represent true constitutive transcription of GM-CSF or, alternatively, may be due to cross-hybridizing sequences in the cDNA probe used (34). Stimulation of SP-B21 cells with -CD3 mAb induced a measurable increase in the transcription rate of IL-2, IL-3, IL-4, IL-5, GM-CSF, and IFN-. IL-2 was not always detected in these experiments (Figure 7, left panel), most likely due to the time of assay being past the peak of maximal transcription (Figure 1). Treatment of cells with CHX 10 min before -CD3 stimulation revealed differences in the cytokine genes. IFN- transcription was unaffected by inhibition of protein synthesis, while a small decrease in GM-CSF transcription was noted. Importantly, CHX addition prior to stimulation greatly blocked the transcription of the IL-2, IL-3, IL-4, and IL-5 genes. However, for only IL-5 did the absence of transcription (Figure 7) correspond to an absence of steady-state mRNA (Figure 5). Nuclear run-on assays were also done in CD4 LTC; however, the rate of transcription was weaker than in SP-B21 cells. While IL-3, GM-CSF, and actin signals were measurable, transcription of IL-4 and IL-5 was barely detectable. Therefore, our conclusions are based on nuclear run-on experiments with SP-B21 cells only.

View larger version (40K): [in this window] [in a new window]   Figure 7.   Nuclear run-on analysis of cytokine transcription in the presence and absence of CHX. SP-B21 cells, in the presence or absence of CHX (10 µg/ml) added 10 min earlier, were added to uncoated or -CD3-coated tissue culture flasks. After 3 h of incubation, cells were harvested, nuclei isolated, and run-on assays conducted as described in MATERIALS AND METHODS. Nylon blots containing UV-crosslinked cytokine, -actin, or GAPDH cDNAs were hybridized to the labeled nuclear transcripts for 72 h, washed, and exposed on film for approximately 10 d. Left and right panels represent two identically conducted experiments.

	Discussion

Top Abstract Introduction Materials & Methods Results Discussion References

To study the regulation of IL-5 gene expression in human T cells, two cell systems were used. One cell type used was a human T_H0 clone, SP-B21, which produces a broad array of cytokines following stimulation, thereby allowing direct comparison of aspects of IL-5 gene regulation with those of other cytokine genes. In addition to the T_H0 clone SP-B21, polyclonal CD4⁺ T cells with a T_H2 phenotype were used. These populations were generated from long-term growth (20 to 27 d) of peripheral blood CD4⁺ T cells under conditions that favored the development of the T_H2 cytokine profile. The differentiation of a T_H2 phenotype from CD4⁺ T cells required the inclusion of a co-stimulatory signal, provided by -CD28 mAb here, as well as IL-4 and anti-IFN- mAb, as has been described by others (26, 27, 35). Similar to the SP-B21 cells, these cells secreted high levels of IL-5 when appropriately stimulated and generated large numbers of cells (> 10⁹) required for this type of study. The study of IL-5 regulation in these T_H2 T-cell populations derived from multiple donors is likely to be representative of the dominant IL-5 producing cells in vivo.

Although stability studies of IL-2 (28, 36, 37), IL-3 (38), IL-4 (38), GM-CSF, and IFN- (19) have been done in various human cell systems, there has been no direct comparison of the t_1/2 values of these transcripts in a single cell system using T-cell receptor stimulation. In addition, while murine IL-5 mRNA in a mouse T-cell lymphoma has been reported to be very stable (39), specific t_1/2 values have not been reported. By using two different agents to arrest transcription at different times after cytokine mRNA induction by -CD3 mAb stimulation in both a T_H0 clone and polyclonal T_H2 cells, it has been clearly demonstrated here that the t_1/2 of IL-5 mRNA is considerably longer (by at least 2 h) than that of other cytokine transcripts produced by the same cells. Thus, this is likely to be a feature intrinsic to the IL-5 gene itself, since it is observed in different species, transformed and nontransformed T cells, and under different modes of T-cell stimulation. The long half-life of IL-5 mRNA is not insignificant; small changes in mRNA levels can lead to large changes in gene product because on the order of 10⁴ moles of protein are synthesized per mole of mRNA (40).

Cytokine mRNA stability was also examined in both cell systems in response to stimulation by -CD3 stimulation in the presence of CD28 costimulation. Co-stimulation by -CD28 mAb mimics the effects of interaction of CD28 cell surface molecules on T cells with the counter-receptors CD80 and CD86, expressed on antigen-presenting accessory cells. An important consequence of CD28 co-stimulation in T cells in vitro is increased production of cytokines, which is largely the result of an increased stability of cytokine mRNAs (19). In general, we observed less than a 40% increase in cytokine mRNA stability induced by CD28 co-stimulation in either SP-B21 or CD4 LTC and IL-5 mRNA was consistently more stable than the other cytokine genes. This contrasts with the large increases (500%) in IL-2, IFN-, and GM-CSF mRNA stability observed by others (19) in freshly isolated primary, peripheral blood T cells. This lack of sensitivity of mRNA half-lives to the effects of CD28 co-stimulation in SP-B21 and CD4 LTC is also reflected in the degree to which CD28 mAb treatment enhanced the cytokine secretion by those cells. As shown in Table 1, a less than 1.3-fold enhancement in IL-5 or IFN- levels was observed in SP-B21 cells by CD28 co-stimulation and only a two- to fourfold increase in IL-5 or IFN- levels was seen in co-stimulated CD4 LTC. It is known that long-term cultured T-cell clones, such as SP-B21 cells, unlike naive cells are not absolutely dependent on co-stimulation for cytokine production, although with suboptimal T-cell receptor stimulation, co-stimulation can achieve maximal cytokine production (41). In addition, the long-term growth conditions used for the differentiation of CD4⁺CD45RO T cells to CD4⁺CD45RO⁺ T_H2 phenoytpe also eliminated an absolute requirement for CD28 co-stimulation, because -CD3 mAb stimulation alone induced significant cytokine secretion (Table 1). The results with CD4 LTC may predict that CD28 activation-induced increases in cytokine mRNA stability in memory or effector T cells in vivo may be minimal as compared with CD28 activation-induced increases in naive T-cell production of IL-2. The latter is sufficiently large to account, importantly, for an autocrine to paracrine shift in IL-2 production (42).

It is noteworthy that, in both cell types tested, of all the cytokine mRNAs examined except IL-5, the t_1/2 values determined with CSA were considerably shorter than those determined with ActinoD. A similar, large decrease in the IL-5 t_1/2 was not seen, although the highly stable nature of IL-5 mRNA and the difficulty in determining exact t_1/2 values under all conditions prevents drawing the same conclusion for this cytokine. The shorter t_1/2 values determined with CSA suggests that ActinoD may block the transcription of other genes whose products regulate cytokine mRNA decay or, alternatively, that inhibition by CSA of a restricted set of Ca⁺-induced signals accelerates cytokine mRNA decay. Given that ActinoD can alter total cellular RNA metabolism (22), the former is the more likely. In addition, of the cytokine transcripts examined in SP-B21 cells, only IL-5 mRNA had a t_1/2 that was longer than that which would be expected from the kinetics of steady-state mRNA decline (Figure 1A). This indicates that the degradation pathway that targets IL-5 is distinct from that which targets IL-2, IL-3, IL-4, IFN-, or GM-CSF in -CD3 mAb-stimulated SP-B21 cells and involves proteins whose transcription is blocked by ActinoD and CSA.

There is evidence that cytokine mRNA degradation is mediated by rapidly turned-over proteins, in the absence of which mRNA is relatively stable (32, 43). One potential target for these labile proteins involved in mRNA degradation are the AU-rich elements found in the 3' untranslated region (UTR) of many cytokine genes and proto-oncogenes (19). An inducible cytoplasmic factor, requiring both de novo transcription and translation, has been described in T cells that binds specifically to cytokine AU-rich regions (44, 45) and may play a role in targeting GM-CSF mRNA for degradation. It is possible that different 3' AU-rich cytokine transcripts may be selectively degraded by sequence-specific factors. IL-5 mRNA possesses six copies of the AUUUA motif dispersed over a 1-kb region of the UTR, while GM-CSF contains five closely linked copies, and IL-4, two single copies (46). A detailed examination of the AU-rich elements in the 3'-UTR of the IL-5 transcript and the proteins that interact with these sequences has not been undertaken yet is necessary to explain the unusually long t_1/2 of this mRNA molecule.

IL-5 was unique among the cytokines examined in having an absolute requirement for protein synthesis for de novo transcription. The requirement for protein synthesis for IL-5 transcription was absolute and was reflected in both an absence of transcription inititation as well as undetectable IL-5 steady-state mRNA levels when cells were stimulated following exposure to CHX. In contrast, while transcription of the IL-2, IL-3, and IL-4 genes was highly dependent on new protein synthesis as shown in the nuclear run-on assay, this was not as stringent as for IL-5, since only a partial reduction in steady-state mRNA levels for these cytokines was observed. This can be explained by the occurrence of a low degree of IL-2, IL-3, and IL-4 transcription (below the limits of detection in the run-on experiments) followed by stabilization of those transcripts (33, 36, 38, 43). Thus, IL-5 gene activation has an absolute requirement for protein synthesis. In addition, IL-5 transcript degradation is controlled by newly synthesized proteins in a manner that is distinct from the other cytokine genes, since no evidence of superinduction, as seen with other cytokines (33, 36, 38, 43), was found in the absence of protein synthesis.

The in vivo administration of anti-IL-5 antibodies in several animal models of eosinophilia (7, 8), induced by a variety of agents, has clearly demonstrated that eosinophilia is controlled in a dominant fashion by IL-5. Eosinophilia, whether induced by parasite infection (8) or regular antigen exposure, such as in asthma (9), is often persistent in nature and is associated with activated T cells with a T_H2 cytokine profile (7, 14). The long half-life of IL-5 mRNA in activated T cells indicates the expression of this gene is not as tightly regulated temporally as other T-cell- derived cytokine genes and may provide a mechanism by which, once the gene is activated, IL-5 protein synthesis can occur over long periods of time. Further definition of the dependency of IL-5 gene expression on de novo protein synthesis or of the sequence elements or proteins responsible for the stability of the IL-5 transcript is required.