Introduction

Abstract Introduction References

Many studies of allergic conditions have found an association of inflammation with increased numbers of activated T cells and increased expression of cytokines, in particular interleukin (IL)-4 and IL-5. Allergic immune responses involving increased immunoglobulin E levels, increased mast cell activation and mediator release, and selective differentiation of eosinophils are regulated by IL-4 and IL-5 (1, 2). Increased activation of CD4⁺ T cells and increased T-cell expression of IL-4 and IL-5, but not messenger RNA (mRNA) for interferon- (IFN-) have been observed in bronchoalveolar lavage fluid (BALF) from allergen-challenged atopic asthmatic (3, 4) and nonatopic asthmatic individuals (5). Bronchial mucosal biopsies from asthmatic patients show increased IL-5 mRNA expression and increased numbers of mucosal eosinophils and activated T cells (6). A preferential activation of cells expressing a Th2 pattern of cytokines (IL-3, IL-4, IL-5, and granulocyte-macrophage colony stimulating factor [GM-CSF]) but not Th1 cytokines (IL-2 and IFN-) has been seen in allergen-induced late-phase cutaneous reactions in atopic patients (7).

Immunohistochemical and in situ hybridization studies of tissue from patients with several allergic states have shown that cell types other than T cells synthesize IL-4 and IL-5. Both mast cells and eosinophils produce these cytokines (8). The relative role of the different cell sources of Th2 cytokines in allergic disease is difficult to assess experimentally. However, in vivo depletion of T cells in a mouse model of pulmonary inflammation reduced both pulmonary eosinophilia and the increased levels of IL-4 and IL-5 mRNA associated with the inflammation, indicating a strong T-cell dependency of these responses (11). In the same model, it was also observed that all lung IL-5 mRNA induced after airway challenge was T-cell associated (12). Therefore, treatments aimed at reducing T-cell activation and IL-5 secretion should be of benefit in reducing allergic inflammation.

A possible source of agents for novel therapies targeted at T-cell activation in allergic inflammation is secondary metabolites of microorganisms. Several families of microbial products are known to have diverse effects on T cells, although their effects on IL-5 production are less well characterized. The inhibitory effects of rapamycin, an immunosuppressive, lipophilic macrolide antibiotic produced by Streptomyces sp., are largely against T-cell proliferation (13). Rapamycin blocks signal-transduction pathways important to the progression of IL-2-stimulated T cells from the G₁ to the S phase of the cell cycle. One member of the macrotetrolide family of antibiotics, tetranactin, is known to inhibit T-cell proliferation and IL-2 production (14). The effects of this family of antibiotics on the production of Th2-type cytokines is unknown. Cyclosporin A (CSA), a cyclic 11-amino-acid-membered peptide of fungal origin, inhibits both T-cell proliferation and cytokine production (13). The evidence for whether human and mouse T-cell- derived IL-5 is sensitive (15, 16) or resistant (17) to the inhibitory effects of CSA is variable. In addition, many of the inhibitory properties of CSA are modulated by CD28-mediated signal-transduction pathways (21). A thorough investigation of the effects of CSA on T-cell production of IL-5 in the presence and absence of CD28 activation has not been undertaken, and would serve to resolve the differing reports of IL-5 sensitivity to CSA.

In this study, we examined and compared the in vitro effects of CSA and macrotetrolides on proliferation and cytokine production by purified, cultured peripheral blood CD4⁺ T cells from normal donors. To examine which cell-activation pathway was affected by these agents, T cells were stimulated with an immobilized anti-()-CD3 monoclonal antibody (mAb) in the absence and presence of a costimulatory signal provided by soluble -CD28 mAb. The latter stimulus mimics the effects of interaction of CD28 cell-surface molecules on T cells with the counterreceptors CD80 and CD86 expressed on antigen-presenting accessory cells (24), and modulates the signals delivered through the T-cell receptor (TCR) alone (21, 23). Because of the cytokine-inhibitory properties of these microbial products, particularly with respect to IL-5, these agents were also examined in vivo in a murine model of IL-5-mediated pulmonary inflammation.

	Materials and Methods

Production of Macrotetrolides

Fermentation of Streptomyces sp. was done in a 2-liter Erlenmeyer flask containing 350 ml of the fermentation medium. The flasks were incubated at 28°C on a rotary shaker at 300 rpm with agitation and 3.5 lpm aeration for 115 h. For isolation, a 4-liter fermentation broth was extracted twice with two volumes of ethyl acetate, the organic layer was removed and dried over anhydrous sodium sulfate, and the solvent was removed. The extract was dissolved in a minimum amount of methylene chloride and loaded on a silica gel column (24 × 3 in) packed with the same solvent. The column was eluted with CH₂Cl₂, CH₂Cl₂: methanol (99:1), CH₂Cl₂: methanol: acetone (8:1:1), and CH₂Cl₂: methanol: acetone (7:2:1). The fractions were monitored through their effects on IL-5 production by -CD3 mAb-stimulated SP-B21 cells (see the following discussion). The active fractions were combined and the solvent was removed. Following separation on a semipreparative Deltapak C-18 column (3 × 30 cm; Waters Inc., Milford, MA) and elution with a mixture of methanol and water (87:13), the methanol was removed from the individual peak eluates under vacuum and the aqueous solution was freeze-dried. Three compounds were isolated. They were low-melting, white, waxy solids, soluble in methanol, chloroform, and ethyl acetate but insoluble in water. These compounds showed only end absorption on UV spectroscopy at 1,740 cm¹. A fast atom bombardment mass spectrum showed molecular ions at m/z (M + H)⁺ 759, 773, and 787, respectively, indicating that the compounds were homologues. The three compounds were identified by spectral methods (¹H nuclear magnetic resonance [NMR] and ¹³C NMR) as the macrotetrolides monactin, dinactin, and nonactin (26). These and other macrotetrolides have been previously isolated in various combinations and proportions from a variety of Streptomyces species (26). They are macrocyclic tetraesters with four R groups, which differ among the nactins as follows: nonactin, R₁ = R₂ = R₃ = R₄ = H; monactin, R₁ = CH₃, R₂ = R₃ = R₄ = H; dinactin, R₁, R₂ = CH₃, R₃, R₄ = H (27).

Th0 Cells

The human Th0 clone SP-B21 (obtained from H. Yssel, DNAX, Palo Alto, CA) has been described previously (28, 29). This T-cell clone was grown by biweekly stimulation with irradiated (4,000 rads) allogeneic peripheral blood mononuclear cells (PBMC), the irradiated (5,000 rads) Epstein-Barr virus-transformed lymphoblastoid cell line JY, and purified phytohemagglutinin (Murex, Norcross, GA) in Yssel's medium (28) supplemented with 1.5% human AB serum (Gemini Bioproducts, Calabasas, CA). Three days after each restimulation, the cultures were expanded in this medium, containing 2 ng/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 JY cells. Yssel's medium was used in all experiments done with SP-B21 cells.

Purification of CD4⁺ Cells

Peripheral blood was obtained from normal, healthy individuals at the Schering-Plough Research Institute. PBMC were isolated with Ficoll-Hypaque (Pharmacia, Uppsala, Sweden) density-gradient centrifugation. 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% fetal calf serum, 10 mM N-2-hydroxyethylpiperazine- N'-ethanesulfonic acid, 2 mM L-glutamine, 50 U/ml penicillin, 50 µg/ml streptomycin (Fisher Scientific, Springfield, NJ), and 5 × 10⁵ M 2-mercaptoethanol (Sigma, St. Louis, MO). Fluorescence-activated cell sorting (FACS) analysis demonstrated that CD4⁺ T cells purified with Dynabeads were routinely > 98% pure. FACS analysis of CD4⁺ T cells isolated as described previously showed approximately 45% naive (CD4⁺CD45RA^hiCD45RO^lo) and 40% memory (CD4⁺CD45RA^loCD45RO^hi) cells (30).

CD4⁺ Cell Culture

Six-well Falcon tissue culture plates (Becton Dickinson) were coated with 10 µg/ml -CD3 mAb (clone UCHT1; Pharmingen, San Diego, CA) in phosphate-buffered saline (PBS) and incubated overnight at 4°C. Purified CD4⁺ cells were added to the -CD3 mAb-coated plates at 4 × 10⁵ cells/ml, 4 ml per well, and incubated for 3 d at 37°C, under 5% CO₂. On Day 3 the cells were recultured at 4 × 10⁵ cells/ ml in cRPMI containing recombinant human IL-2 (2 ng/ ml; Biosource, Camarillo, CA) in uncoated six-well tissue-culture plates for 3 d, at which time the cells were stimulated as indicated. Over the 6-d culture period, the CD4⁺ T cells expanded approximately 5- to 10-fold and remained highly pure (> 98%), indicating that no contaminating cell population grew during the culture period.

Cell Stimulation for Proliferation and Cytokine Production

CD4⁺ T cells grown for 6 d as described previously, or SP-B21 cells, were stimulated (1 × 10⁵ cells/well) with the indicated amount of -CD3 mAb in the absence or presence of -CD28 mAb (1 µg/ml; Pharmingen) in a total volume of 200 µl per well in 96-well plates. All controls and experimental groups of CD4⁺ T cells were tested in quadruplicate. SP-B21 cells were used in duplicate. Stock solutions of CSA (Sandoz, Whippany, NJ) were made at 10 mg/ml in ethanol, stored at 70°C, and diluted appropriately in cRPMI on the day of use. Solvent control cultures contained the same concentration of ethanol as contained in the highest concentration of compound, which was no greater than 0.1%. The cultures were incubated for 24 h at 37°C under 5% CO₂, at which time supernatants were harvested for determination of cytokine levels by enzyme-linked immunosorbent assay (ELISA). At this time cells were also assayed for viability, designated as 3-(4,5-dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)-2-(4-sulfophenyl)-2H-tetrazolium (MTS) activity, using the CellTiter 96 AQueous Kit (Promega, Madison, WI). When indicated, replicate sets of cultures were incubated for the indicated time, pulsed with 1 µCi/well [³H]thymidine ([³H]TdR; NEN, Boston, MA), and harvested at the indicated time onto glass fiber filters for scintillation counting.

Cytokine ELISAs

Primary and secondary antibodies for quantitation of IL-5 and IL-4 by ELISA were obtained from Pharmingen and used according to the manufacturer's protocol; the detection limit for these assays was approximately 20 pg/ml. IFN- was measured with an IFN- Cytoscreen Immunoassay Kit (Biosource); the limit of detection was 4 pg/ml according to the manufacturer's specifications. IL-2 was measured with the Quantakine Kit (R&D, Minneapolis, MN); the limit of detection was 7 pg/ml according to the manufacturer's specifications. Percent inhibition of cytokine production, cell viability, and [³H]TdR incorporation were calculated for each concentration of compound as follows: [(solvent control  compound group)  solvent control] × 100. The concentration of compound that caused 50% inhibition of the response in the solvent control cultures was designated as the IC₅₀ concentration. There was not detectable cytokine synthesis in the absence of stimulation.

RNA Analysis: Steady-State mRNA

Six-well Falcon tissue culture plates were coated overnight with -CD3 mAb (1 µg/ml) in PBS. SP-B21 cells were added at 1 × 10⁶/ml, 4 ml/well, to 4 to 6 wells per experimental group. Compounds were added to the cultures at the indicated concentrations at the time of cell stimulation. At the indicated time, total cellular RNA was isolated from cultured cells with Tri Reagent according to the manufacturer's protocol (Molecular Research Center, Inc., Cincinnati, OH). RNA was quantitated by measurement of absorbance at 260/280 nm. Denatured RNA was electrophoresed in a 1.2% agarose/2.2 M formaldehyde gel. After staining with ethidium bromide and photographing with a UV transilluminator, RNA was transferred to Genescreen Plus (NEN) in 10× saline sodium phosphate-ethylenediamine-tetraacetic acid (SSPE) and UV cross-linked. Blots were prehybridized overnight and then hybridized for 18 to 24 h in 5× SSPE, 5× Denhardt's solution, 1% sodium dodecyl sulfate (SDS), 50% formamide, and 100 mg/ml salmon sperm DNA (Sigma) at 42°C. Complementary DNA (cDNA) probes were labeled with (-³²P)deoxycytidine triphosphate ([-³²P]dCTP) (NEN; > 3,000 Ci/mmol), using a random primer kit (Stratagene, La Jolla, CA), to a specific activity of  2 × 10⁹ cpm/µg; 2 × 10⁸ cpm/ml hybridization solution was used. After high-stringency washes, blots were exposed to Kodak XAR-5 film (Kodak, Rochester, NY) with enhancing screens (Sigma). These blots were sequentially hybridized, quantitated, stripped, and rehybridized, with the final hybridization being made to glyceraldehyde-3-phosphate dehydrogenase (GAPDH) or -actin cDNA probes.

RNA Analysis: Nuclear Run-On Assay

SP-B21 cells were stimulated for 3 h with immobilized -CD3 mAb (1 µg/ml). 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 blot analysis. The transcription in freshly isolated nuclei was done essentially as described elsewhere (31), with modifications (32). Nuclei from 1 × 10⁸ SP-B21 cells were used per experimental group, as described previously. Nuclei were labeled in reaction buffer (10 mM Tris-HCl, pH 8.0; 5 mM MgCl₂; 0.3 M KCl; 1 mM each adenosine triphosphate [ATP], CTP, and guanosine triphosphate [GTP]; 5 mM dithiothreitol; ribonuclease inhibitor [Promega]; and 100 µCi -[³²P]uridine triphosphate [UTP; NEN]), and were incubated at 30°C for 30 min. Labeled transcripts were isolated by 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 ethylenediaminetetraacetic acid (EDTA), 0.2% SDS, 0.6 M NaCl, and 5× Denhardt's solution (Sigma). For each treatment, labeled RNA (approximately 0.5-1.5 × 10⁶ cpm) was hybridized in 500 µl to the indicated purified cDNA inserts (250 ng/slot blot), which had been UV cross-linked to nylon membranes. Slot blots were prehybridized overnight, hybridized for 72 h at 65°C, and washed, as described previously for Northern blot analysis.

cDNA Probes

The following purified human cDNA inserts (American Type Culture Collection, Rockville, MD) were used as probes in Northern blot analysis and nuclear run-on experiments: -actin, No. 78554; GM-CSF, No. 57595; IL-4, No. 57593; IL-3, No. 59399; and GAPDH, No. 57091. 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 of DNAX: IL-2 (1,021 bp), IL-5 (1,025 bp), and IFN- (1,402 bp).

In Vivo Allergic Pulmonary Inflammation

Detailed procedures for inducing allergic pulmonary inflammation in mice have been described previously (33). In brief, B6D2/F1 mice (Jackson Laboratory, Bar Harbor, ME), between 6 and 10 wk of age, were sensitized by intraperitoneal injection of 7.5 µg ovalbumin (OVA) adsorbed to 2 mg alum gel suspended in 0.5 ml normal saline on Days 0 and 5. Nonsensitized control animals received alum only. Seven days after the second sensitization, the animals were challenged by two exposures to OVA (0.5%) aerosolized with an ultrasonic nebulizer (Ultra-Neb 99; DeVilbiss, Somerset, PA). The two exposures were of 1 h duration and were separated by 4 h. One day after challenge, lungs from groups of mice were perfused through the pulmonary artery via the right ventricle with saline to remove peripheral blood, and were lavaged with 0.5 ml saline containing 0.1% EDTA. Total cell and eosinophil numbers were enumerated according to standard hematologic procedures. Cytospins of BALF were prepared and stained with Leukostat stain (Fisher, Pittsburgh, PA) for differential counts.

Analysis of In Vivo Cytokine mRNA Levels by Semiquantitative Polymerase Chain Reaction

Perfused lungs were immediately frozen on dry ice and stored at 70°C. Lung tissue was solubilized in guanidinium isothiocyanate, and RNA from the tissue was purified by cesium chloride ultracentrifugation (34). A semiquantitative polymerase chain reaction (PCR) was used to measure steady-state levels of mRNAs for IFN-, IL-4, and IL-5 as described (12, 35). Briefly, RNA (5 µg) was reverse transcribed with avian myeloblastosis virus reverse transcriptase (Boehringer Mannheim, Indianapolis, IN) in duplicate, and the product was serially diluted. Aliquots of each dilution were amplified with AmpliTaq DNA polymerase (Perkin Elmer, Norwalk, CT), using a GeneAmp PCR System 9600 Thermal Cycler (Perkin Elmer) under conditions determined in earlier experiments to be optimal for exponential amplification. A standard curve for each analyzed gene was generated by using a sample of RNA containing all cytokines examined, and unknown sample dilutions were quantitated relative to this curve. The RNA used for generating the standard curve was from BALB/c splenocytes stimulated in vitro with concanavalin A. Derived values were normalized by reference to the constitutively expressed mRNA that encodes hypoxanthine phosphoribosyltransferase.

Animal Care and Use

The study was conducted in accordance with the National Institutes of Health Guide for the Care and Use of Laboratory Animals and the Animal Welfare Act, in a program approved by the American Association for the Accreditation of Laboratory Animal Care.

	Results

During the process of screening microbial products for inhibitors of IL-5 biosynthesis, an extract was identified that inhibited IL-5 production in cultures of -CD3 mAb-stimulated SP-B21 cells, a human Th0 T-cell clone. The cytokine production pattern of this clone has been studied previously (36). Isolation and purification of the individual components of the extract resulted in the identification of a previously known family of macrotetrolide antibiotics (27). Purified nonactin, monactin, and dinactin were tested for effects on -CD3 mAb-stimulated SP-B21 cells. Monactin (Figure 1A) and dinactin (Figure 1B) had similar inhibitory effects on IL-5 production (IC₅₀ = 20 ng/ml), and displayed weak inhibitory effects on cell viability (30% inhibition at 1,000 ng/ml) in the MTS assay. In contrast, nonactin (Figure 1C) was a less potent inhibitor of IL-5 production (IC₅₀ = 200 ng/ml). CSA was a potent inhibitor of IL-5 production by -CD3 mAb-stimulated SP-B21 cells (IC₅₀ = 50 ng/ml); complete inhibition was observed at 100 ng/ml, without evidence of toxicity (Figure 1D). For further characterization of the effects of macrotetrolides on T-cell function, dinactin was selected because of its 10-fold greater potency than nonactin on IL-5 production, its similar dose-response profile to that of monactin, and its larger yield over monactin in the purification process.

View larger version (25K): [in this window] [in a new window]   Figure 1.   Effects of the macrotetrolides monactin, dinactin, and nonactin on IL-5 production by -CD3 mAb-stimulated SP-B21 cells. SP-B21 cells were activated with immobilized -CD3 mAb (0.3 µg/ml) in the presence of the indicated concentrations of monactin (A), dinactin (B), nonactin (C), and CSA (D). Values represent the means of duplicate cultures that differed by less than 20%; the results are representative of two experiments. Supernatants were collected at 24 h, and IL-5 was measured with an ELISA. Cell viability was assessed by using the MTS assay. IL-5 levels in control cultures were 24.9 ± 1.06 ng/ml.

The effects of CSA and dinactin on human T-cell proliferation and cytokine production were tested with CD4⁺ T cells purified from peripheral blood of healthy individuals and cultured for 6 d as described previously (30). CD4⁺ T cells were stimulated with immobilized -CD3 mAb in the presence or absence of soluble -CD28 mAb. Dinactin inhibited CD4⁺ T-cell proliferation with an IC₅₀ of 10 to 20 ng/ml, and complete inhibition of [³H]TdR incorporation was observed at higher concentrations (Figures 2A and 2C). The dose response to dinactin was identical when T cells were stimulated through the TCR alone (-CD3) or costimulated with -CD28 mAb (Figures 2A and 2C). Moreover, the dose response to dinactin was unaltered by the inclusion of exogenous IL-2 in the mAb-stimulated cultures (Figure 2C). Stimulation in the presence or absence of -CD28 mAb resulted in distinctly different responses to CSA (Figures 2B and 2D). Almost complete inhibition of proliferation was seen with -CD3 mAb-stimulated T cells in the presence of > 100 ng/ml CSA (IC₅₀ = 10 ng/ml). In contrast, when T cells were stimulated with -CD3 mAb in the presence of -CD28 mAb, there was at least a 10-fold decrease in potency of CSA compared with that of T cells stimulated with -CD3 mAb alone (Figure 2D). In other experiments, only weak inhibition of proliferation ( 25%) was seen with all concentrations of CSA in -CD3 mAb plus -CD28 mAb-stimulated cultures (Figure 2B). The inclusion of IL-2 in the cultures resulted in a shift in the dose-response curve such that little inhibition of T-cell proliferation was seen in the presence of CSA regardless of the mode of stimulation (Figure 2D).

View larger version (21K): [in this window] [in a new window]   Figure 2.   Differential inhibitory effects of dinactin and CSA on proliferation of CD4+ T cells. CD4+ T cells grown as described in MATERIALS AND METHODS were stimulated with -CD3 mAb (10 µg/ml) in the presence or absence of soluble -CD28 mAb (1.0 mg/ml) in the presence of the indicated concentrations of dinactin (A, C) or CSA (B, D). The cultures in (A) and (C) or (B) and (D) were derived from a single donor. In some cultures, exogenous IL-2 (2 ng/ml) was added (C, D). The cultures in (A) and (B) were incubated for 24 h, pulsed with [3H]TdR at 1 µCi/well for 24 h, and harvested at 48 h. The cultures in (C) and (D) were incubated for 20 h, pulsed with [3H]TdR at 1 µCi/well for 4 h, and harvested at 24 h. Values represent the mean percent inhibition of quadruplicate cultures. SEM were less than 15% of the mean. [3H]TdR incorporation in control cultures for (A) and (B) was -CD3 = 51,578 ± 2,563 cpm; -CD3 plus -CD28 mAb = 48,165 ± 2,826 cpm; and medium alone = 485 ± 109 cpm. [3H]TdR incorporation in control cultures for (C) and (D) was -CD3 = 3,778 ± 526 cpm; -CD3 plus -CD28 mAb = 10,890 ± 711 cpm; -CD3 plus IL-2 = 8,710 ± 397 cpm; -CD3 plus -CD28 mAb and IL-2 = 11,240 ± 697 cpm; medium alone = 118 ± 49 cpm; IL-2 alone = 1,717 ± 152 cpm. The experiments shown are representative of six different donors, each tested with dinactin and CSA.

T-cell proliferation is known to be controlled by IL-2 (37). To assess directly the ability of the antibiotics described here to affect T-cell growth differentially, CD4⁺ T cells were starved of IL-2 for 18 h and then recultured with fresh IL-2 alone for 24 h in the presence of each of the microbial products. As has been reported elsewhere (21), CSA had no effect on T-cell proliferation (data not shown) under conditions in which exogenous IL-2 was supplied. In contrast, dinactin (data not shown) inhibited IL-2-induced T-cell proliferation with the same dose-response curve as that seen with -CD3 mAb-stimulated T cells in either the presence or absence of soluble -CD28 mAb (Figures 2A and 2C).

The CD4⁺ T-cell cultures were used to test the capacity of the microbial products to inhibit cytokine production. In eight independent experiments, both IL-5 and IFN- were detected in the culture supernatants of CD4⁺ T cells from multiple donors after restimulation with -CD3 mAb following the 6-d culture period (Table 1). Thus, these cells, at the bulk culture level, had a Th0 cytokine profile, in that both IL-5 and IFN- were produced. The inclusion of soluble -CD28 mAb in the cultures induced an approximate 5-fold increase in both IL-5 and IFN- levels, and thus did not alter the twofold ratio of IL-5 to IFN- produced without costimulation (Table 1). To detect IL-2 and IL-4, costimulation with -CD28 mAb was required, since both cytokines were generally undetectable in T-cell cultures activated with -CD3 mAb only (data not shown).

Dinactin and CSA were tested for their effects on cytokine production through the use of -CD3 mAb and -CD28 mAb stimulation of CD4⁺ T cells to measure multiple cytokines. Dinactin inhibited cytokine production with IC₅₀ values ranging from 10 ng/ml for IL-4 and IL-5 to 30 or 60 ng/ml for IFN- or IL-2, respectively (Figure 3A). Complete inhibition of IL-4 and IL-5 but not of IFN- or IL-2 production was seen. The inhibitory effect of dinactin on cytokine production was not due to cell death, because less than 30% inhibition of cell viability was noted at all concentrations tested in the MTS assay (Figure 3A). CSA also inhibited cytokine production with an IC₅₀ for IL-2, IL-4, and IFN- of 40 to 60 ng/ml (Figure 3B). Interestingly, CSA-mediated inhibition of IL-5 production by CD4⁺ T cells costimulated with -CD3 mAb in the presence of -CD28 mAb was biphasic and was significantly weaker than that observed for the other cytokines. As shown for two donors, at CSA concentrations of less than 125 ng/ml, IL-5 levels were enhanced above that found in the control cultures (Figures 3B and 3C). Concentrations of CSA above 250 ng/ml were weakly inhibitory (Figures 3B and 3C). The dose-dependent enhancement of IL-5 production was reproducible, as shown in Table 2. In 14 independent experiments with 14 different donors, low levels of CSA enhanced IL-5 synthesis. The maximum increase in IL-5 synthesis (81.8 ± 11.2%) was obtained with 0.06 ± 0.01 µg/ml CSA. A biphasic dose-response curve and concentration-dependent CSA-induced increases in the production of other cytokines were not observed. In contrast to these findings in costimulated cultures, IL-5 production induced by -CD3 mAb alone was highly sensitive to the effects of CSA (IC₅₀ = 10 ng/ml); complete inhibition was observed at 30 ng/ml (Figure 3C).

View larger version (20K): [in this window] [in a new window]   Figure 3.   Differential inhibitory effects of dinactin and CSA on CD4+ T-cell cytokine production. (A, B) CD4+ T cells, grown as described in MATERIALS AND METHODS, were stimulated with -CD3 mAb (10 µg/ml) and soluble -CD28 mAb (1.0 µg/ml) in the presence of the indicated concentrations of dinactin (A) and CSA (B). After 24 h of culture, cell viability was measured with the MTS assay and culture supernatants were collected for determination of cytokine levels with ELISA. Values represent the mean percent inhibition of quadruplicate cultures. SD were less than 15% of the mean. Mean cytokine levels (ng/ml) in control cultures were: IL-2, 5.4 ± 0.48; IL-4, 6.6 ± 0.16; IL-5, 3.4 ± 0.20; and IFN-, 1.6 ± 0.13. (C ) CD4+ T cells, grown as described previously, were stimulated with -CD3 mAb (10 µg/ml) in the presence or absence of soluble -CD28 mAb (1.0 µg/ml) in the presence of the indicated concentrations of CSA. After 24 h of culture, culture supernatants were collected for measurement of IL-5 levels with ELISA. Values represent the mean percent inhibition of IL-5 in quadruplicate cultures. SD were less than 15% of the mean. IL-5 levels (ng/ml) in control cultures (no CSA) were 5.70 ± 0.57 (-CD3) and 2.45 ± 0.143 (-CD3 plus -CD28 mAb). Results with CSA are representative of 14 donors (Table 2). Results with dinactin are representative of six different donors.

To determine whether dinactin inhibited cytokine production pre- or posttranscriptionally, its effects on transcriptional activation and steady-state mRNA levels were examined (Figure 4). Transcriptional activation was assessed with the nuclear run-on assay, which measures the level of transcription independent of mRNA degradation, transcript processing, or nuclear-to-cytoplasmic translocation (38). For these studies, the Th0 T-cell clone SP-B21, rather than primary CD4⁺ T cells, was used because of stronger signals. In addition, it was not possible to obtain sufficient CD4⁺ T cells from individual donors for such studies. Other studies of cytokine gene expression have utilized the SP-B21 clone and shown that it has similarities to primary CD4⁺ T cells (32).

View larger version (40K): [in this window] [in a new window]   View larger version (28K): [in this window] [in a new window]   Figure 4.   Effects of dinactin, CSA, and rapamycin on cytokine transcriptional activation and steady-state mRNA levels. (A) Untreated SP-B21 cells and SP-B21 cells in the presence of CSA (1,000 ng/ml), dinactin (100 ng/ml), or rapamycin (100 ng/ml; Calbiochem) added 10 min earlier were added to uncoated or -CD3 mAb-coated tissue-culture flasks. After 3 h of incubation, cells were harvested, the nuclei isolated, and run-on assays conducted as described in MATERIALS AND METHODS. Nylon blots containing UV cross-linked cytokine and -actin cDNAs were hybridized to the labeled nuclear transcripts for 72 h, washed, and exposed on film for approximately 10 d. Lanes 1-4 and 5-7 represent two identically conducted experiments. (B) Total cellular RNA was harvested at 3 h from replicate cultures used in Exp. 1 (lanes 1- 4) for nuclear run-on analysis. Northern blot analysis was done as described in MATERIALS AND METHODS, using 15 µg RNA per sample. Exposure times for autoradiograms were 3 d for IL-4 and IL-5, 4 d for IFN-, and 5 h for GAPDH.

In unstimulated SP-B21 T-cell cultures (Figure 4A, lane 5), levels of IL-2, IL-3, and IFN- transcription were undetectable in the nuclear run-on assay, whereas a low level of transcription of IL-4, IL-5, and GM-CSF was observed. The low level of transcription seen for some cytokines may reflect residual, selective stimulation resulting from the conditions under which the cells were grown. However, upon -CD3 mAb stimulation of SP-B21 T cells, increased transcription was observed for all of the cytokines tested (Figure 4A, lanes 1 and 6) except IL-2. IL-2 was not detected in these experiments, most likely because the time of assay was past the peak of its maximal transcription (data not shown). Addition of CSA (1,000 ng/ml) prior to stimulation decreased the level of transcription of all the cytokines tested as compared with the levels produced by stimulated cells in the absence of inhibitor (Figure 4A, lane 2 versus lane 1 and lane 7 versus lane 6). CSA reduced the level of cytokine transcription to that observed in unstimulated cultures (Figure 4A, lane 7 versus lane 5), indicating blockade of transcriptional activation. This effect was also reflected in decreased steady-state mRNA levels for IL-4, IL-5, and IFN- (Figure 4B, lane 2 versus lane 1). Rapamycin did not alter the degree of transcriptional activation of any genes tested (Figure 4A, lane 4 versus lane 1), nor did it alter steady-state mRNA levels for IL-4, IL-5, or IFN- (Figure 4B, lane 4 versus lane 1). Similar to rapamycin, dinactin had no effect on cytokine transcriptional activation (Figure 4A, lane 3 versus lane 1) or on steady-state mRNA levels for IL-4, IL-5, or IFN- (Figure 4B, lane 3 versus lane 1). These results indicate that dinactin acts post-transcriptionally to inhibit cytokine production. Concentrations of monactin and dinactin as high as 2,000 and of nonactin as high as 1,000 ng/ml did not inhibit cytokine steady-state mRNA levels (data not shown). None of the three inhibitors tested altered the steady-state mRNA levels of the housekeeping gene GAPDH (Figure 4B, lanes 2-4 versus lane 1), which was consistent with the minimal effect of these inhibitors on actin transcription in the nuclear run-on experiments (Figure 4A, lanes 2-4 versus lane 1, and lane 7 versus lane 5).

Because of their effects on cytokine production, CSA and dinactin were tested for their ability to block pulmonary eosinophilia in a murine model of allergic pulmonary inflammation. In addition, because of the activity of CSA at the level of cytokine transcription (Figure 4), IL-4, IL-5, and IFN- mRNA levels in the lung were examined in CSA-treated mice.

Sensitized mice were treated intraperitoneally with dinactin (0.1, 1.0, and 10.0 mg/kg body weight [mpk]) 1 h before and 6 h after challenge with nebulized ovalbumin. Eosinophils in the BALF were measured 24 h after challenge (Figure 5A). As described previously (33, 39), in sensitized, challenged mice there was a significant increase in the number of eosinophils in the BALF as compared with that in unsensitized, challenged mice (Figure 5A) or in sensitized, unchallenged mice (Figure 5B). In dinactin-treated mice (0.1 and 1.0 mpk), eosinophils were reduced by 60% as compared with the count in untreated mice (Figure 5A). Interestingly, a higher dose of dinactin (10 mpk) was ineffective in reducing the number of BALF eosinophils.

View larger version (15K): [in this window] [in a new window]   Figure 5.   Dinactin and CSA inhibit eosinophil influx into BALF of OVA-challenged allergic mice. (A) Mice were sensitized or not, as described in MATERIALS AND METHODS, and all mice were challenged. Sensitized mice were treated with dinactin 1 h before and 6 h after challenge with 0.1, 1.0, or 10 mpk dinactin given intraperitoneally. Dinactin was administered in methycellulose as a vehicle. At 24 h after challenge, BALF was collected and eosinophils were counted. Values represent means ± SEM of six mice per group. (B) All groups of mice were sensitized and challenged; a control group of sensitized, unchallenged mice was included. All groups contained nine mice. Sensitized mice were treated 24 h and 1 h before challenge with 25, 50, or 100 mpk CSA in olive oil given orally. For determination of BALF cellular content, five mice per treatment group were used at 24 h after challenge. The other four mice per treatment group were used at 6 h after challenge for quantitating lung cytokine mRNA levels (Figure 6). Results are the average of two experiments, each done on five mice per group. Statistical significance was determined by analysis of variance with Fisher's protected least significant difference test (StatView; Abacus Concepts Inc., Berkeley, CA). *Significantly different value (P < 0.05) than for untreated sensitized, challenged animals.

View larger version (19K): [in this window] [in a new window]   Figure 6.   CSA inhibits cytokine mRNA levels in lungs of OVA-challenged allergic mice. Animal treatment protocol is as described in legend to Figure 5. Six hours after challenge, four mice per treatment group were euthanized and RNA was isolated from pooled lung tissue. Steady-state mRNA levels for IFN- (A), IL-4 (B), and IL-5 (C) were measured with a semiquantitative PCR. Results are the average of two experiments, each done on five mice per group.

CSA dose-dependently inhibited pulmonary eosinophilia in allergic mice when given orally at 24 h and 1 h before challenge (Figure 5B). As observed in other animal models (40), high doses were required for efficacy. CSA (50 or 100 mpk) reduced the number of eosinophils in the BALF by 94% or 98% (Figure 5B), respectively, to levels comparable to those seen in unchallenged animals. The lowest tested dose of CSA reduced the number of eosinophils in the BALF by 62%. A similar pattern of reduction of T cells in the BALF was also observed (data not shown). Animals treated with CSA also showed reduced levels of cytokine mRNA in the lung at 6 h after challenge (Figure 6). The lowest dose of CSA administered (25 mpk) effectively inhibited all challenge-induced increases in IL-4 and IFN- mRNA (Figures 6A and 6B). Interestingly, IL-5 mRNA was less susceptible to CSA-mediated inhibition (Figure 6C). Maximum inhibition was obtained with 50 mpk CSA; 25 mpk CSA reduced IL-5 mRNA levels insignificantly.

	Discussion

In this study we examined and compared the effects of CSA and a macrotetrolide antibiotic, dinactin, on human T-cell proliferation and cytokine production under conditions of stimulation via the TCR alone (-CD3 mAb) or in combination with costimulatory signals (-CD3 mAb and -CD28 mAb). Because of the ability of CSA and of dinactin to inhibit IL-5, these agents were also examined in an in vivo murine model of IL-5-mediated pulmonary inflammation. Several findings are notable and have implications for the treatment of IL-5-mediated inflammation. Both dinactin and CSA reduced pulmonary eosinophilia when given to sensitized mice near the time of airway antigen challenge, although their in vitro T-cell-inhibitory profiles differed. Dinactin was a potent in vitro inhibitor of T-cell proliferation and cytokine production (IC₅₀ = 20 ng/ ml), and in contrast to CSA showed no evidence of selectivity for any T-cell-stimulation pathway. In addition, unlike CSA, dinactin inhibited cytokine production by a post-transcriptional mechanism. Importantly, the induction of IL-5 in CSA-treated T cells displayed a biphasic pattern in the presence of a costimulatory signal. In contrast to other cytokines, IL-5 induced by -CD3 mAb and -CD28 mAb stimulation was only minimally inhibited by CSA, and at some concentrations of CSA was significantly increased above control levels. Additionally, the resistance of IL-5 mRNA to CSA observed in vitro was corroborated in vivo, and suggests a role for CD28 in antigen-induced IL-5 production in the lung.

Both dinactin and CSA were inhibitory to T-cell function, although with clearly different mechanisms of action. All tested stimuli that induced T-cell proliferation were inhibited similarly by dinactin (and monactin), for which minimal to near maximal inhibitory effects were obtained within a 2- to 3-fold range of concentrations. In no case was exogenous IL-2 able to reverse this inhibition, as was seen with CSA. This indicates that dinactin may block a step common to multiple pathways of induction of T-cell growth. A similar macrotetrolide, tetranactin, has been reported to have ionophoric properties in that it depleted intracellular Ca²⁺ concentrations while promoting Na⁺ influx in unstimulated and mitogen-stimulated rat T cells (14), and this is therefore likely to be a macrotetrolide class effect.

Dinactin also inhibited cytokine production by stimulated -CD3 mAb plus -CD28 mAb in normal peripheral blood CD4⁺ T cells cultured in vitro. Although there was a slightly stronger and more complete inhibitory effect on the Th2 cytokines IL-4 and IL-5 (IC₅₀ = 10 ng/ml) than on the Th1 cytokines IFN- and IL-2 (IC₅₀ = 30-60 ng/ml), dinactin, unlike CSA, produced no evidence of resistance to its effects under conditions of costimulation. In addition, provision of exogenous IL-2 did not modulate the effects of dinactin on cytokine production, indicating that the effects of dinactin were not due to low cell viability and did not occur subsequently to blockade of IL-2. Moreover, a major distinction in the mechanism of action of macrotetrolides and CSA was revealed by gene expression studies. Whereas CSA blocks cytokine gene transcription by inhibiting the Ca²⁺-calmodulin-dependent phosphatase calcineurin (41), dinactin had no effect on transcriptional activation nor on steady-state mRNA levels of the cytokine genes examined. This clearly indicates that the effects of dinactin are post-transcriptional. When administered to sensitized mice 1 h before and 6 h after aerosol antigen challengetiming that may optimize the effects of a post-transcriptional cytokine inhibitordinactin significantly reduced pulmonary eosinophilia at low doses. Dinactin was less effective in vivo than would be predicted on the sole basis of its potent inhibition of cytokine production in vitro. Degradation of dinactin to less active metabolites may contribute to this. Alternatively, dinactin may not block other mechanisms contributing to pulmonary eosinophilia. However, because of the ionophore properties of the macrotetrolides (14) and their broad, nonselective effects on T-cell function (42), it is unlikely that these agents can be pursued further as therapeutic agents.

Several studies, using a variety of different T cells and stimuli, and generally using single concentrations of CSA, variably indicate that IL-5 gene expression or secreted protein is either reduced (15, 16) or not reduced by CSA (17). In a detailed comparison of cytokine production in human T cells stimulated through the TCR alone or in conjunction with a costimulatory -CD28 mAb in the presence of a broad concentration range of CSA, we have shown that IL-5 is highly resistant to CSA only under conditions of costimulation. In contrast to the case with other cytokines, only weak inhibition of IL-5 was seen at CSA concentrations above 250 ng/ml, and at concentrations below 125 ng/ml, IL-5 levels were greater than in non-CSA-treated cultures. Although it is known that CD28 costimulation decreases the sensitivity of several cytokines to CSA (20, 21, 23), the biphasic dose-response curve and concentration-dependent CSA-induced increases that we observed for IL-5 were not observed with IL-2, IL-4, or IFN-. The mechanism responsible for this exceptional resistance of IL-5 to CSA under costimulatory conditions is unknown, but may relate to IL-2 levels. We and others have observed that IL-2 production is negligible when T cells are stimulated through the TCR, that IL-2 increases under conditions of costimulation (23, 25), and that IL-2 is less inhibited by CSA under the latter than under the former conditions (21, 23). In addition, it has been reported that IL-2 alone can induce IL-5 gene expression in T cells (15), and transcriptional activation of IL-5 by IL-2 is CSA resistant (16). Thus, under costimulatory conditions and low CSA concentrations, and when some IL-2 is available, IL-2 itself may induce IL-5 as seen in our study, such that a biphasic dose-response relationship between IL-5 and CSA results.

The results described here serve to explain the divergent reports in the literature with regard to CSA and IL-5. The wide variety of stimuli used across many studies is likely to have differential dependency on the CD28 pathway in activating T cells. Those stimuli with a high degree of dependency would reveal resistance of IL-5 to CSA, whereas those not involving CD28 would demonstrate inhibition of IL-5 by CSA.

The effectiveness of CSA in reducing IL-5-mediated pulmonary eosinophilia is likely to be influenced by the degree to which this biologic response involves the CD28 pathway in T-cell activation. The use of antibodies to CD80/CD86 (43), the counterreceptors for CD28, or of agents that block CTLA4 (43), a molecule homologous to CD28, in several mouse models of pulmonary inflammation have clearly shown a role for the CD28 pathway in the eosinophil influx that is elicited by airway antigen challenge. In this context, the in vivo results that we obtained with CSA are interesting. We observed a complete, dose-dependent inhibition of eosinophils in the BALF of challenged animals, with only a partial reduction in IL-5 mRNA levels in the lung. For each concentration of CSA tested there was a greater reduction in eosinophils than of IL-5. The use of anti-IL-5 mAb in this model (46) has demonsrated that although IL-5 is a major contributor to pulmonary eosinophilia, other mechanisms also contribute to it. Thus, CSA is likely to inhibit eosinophils by other mechanisms in addition to IL-5 inhibition.

Importantly, the resistance of IL-5 to CSA seen in our in vitro studies was also observed in lung IL-5 mRNA levels in vivo. Although virtually all challenge-induced increases in IL-4 or IFN- mRNA were eliminated by the lowest dose of CSA tested (25 mpk), IL-5 mRNA levels were decreased by only 68% at the highest dose of CSA (100 mpk) tested. These results corroborate the resistance of IL-5 to CSA observed in vitro, and suggest a role for CD28 in the in vivo induction of IL-5.

Several studies have found that CSA reduces allergen-induced late-phase responses in animal models (47) and in patients with atopic mild asthma (48), and that it reduces disease exacerbations in steroid-dependent patients with severe asthma (49). These studies provide further direct evidence for a role of T cells and their products in the eosinophilic inflammation characteristic of asthma. Because of the immunosuppressive properties of CSA and the poor risk-to-benefit ratio of its use in all but the most severely asthmatic individuals, additional efforts to identify other IL-5-inhibitory microbial products is warranted. On the basis of the findings presented here, agents that target the CD28 pathway of T-cell activation may be effective inhibitors of IL-5.