We have previously shown that lisofylline (LSF) inhibits murine Th1-mediated disease in vivo by blocking IL-12-induced differentiation of Th1 cells. The cellular and molecular mechanisms underlying this inhibition were further explored by testing LSF in several IL-12-responsive model systems in vitro. IL-12-dependent Th1 differentiation was abrogated by LSF and yielded effector T cells that were deficient in proinflammatory cytokine secretion, including IFN-γ, IL-2, and TNF-α. The diminished Th1 phenotype resulted from both a lower frequency of IL-12-derived Th1 clones and a reduced capacity of individual clones to secrete IFN-γ due to lower levels of IFN-γ mRNA. The arrest in Th1 development resulted from a blockade of IL-12 signaling that preceded the Th0 to Th1 transition. Thus, LSF blocked IL-12-enhanced IFN-γ production in anti-CD3-stimulated T cells and prevented IL-12-mediated repression of the transcription factor GATA-3. Lisofylline also inhibited IL-12-induced increases in STAT4 tyrosine phosphorylation, but did not block TCR signaling or inhibit acquisition of IL-12 responsiveness. These findings were extended to show that LSF also inhibits IL-12-dependent responses in human T cells. LSF, which has one asymmetric chiral center, was selectively inhibitory for IL-12 signaling compared with its S-enantiomer (1501-S) and the oxidized side chain analog, pentoxifylline. The results suggest that LSF may be useful as a modulator of Th1-mediated disease in humans.
Thelper cells can be broadly divided into two subpopulations as defined by the types of cytokines they secrete upon secondary antigenic stimulation (1), namely Th1 and Th2 cells. The differentiation of naive T cells into Th1 (inflammatory) or Th2 (anti-inflammatory) effectors is influenced by several factors, including the specific immunogen and Ag dose, the type of APC, and the presence of either IL-12 or IL-4, respectively, upon initial encounter with Ag (2). Evidence for Th polarization in vivo is most evident in instances of chronic antigenic stimulation, and imbalances in the control of Th1 vs Th2 cytokine production contribute to the pathobiology of autoimmune, inflammatory, and allergic diseases.
The pleiotropic cytokine IL-12 is required for the induction of Th1 differentiation, and in addition directly enhances IFN-γ production from NK and T cells (without any additional stimulation), augments T cell proliferation, up-regulates NK and CTL cytolytic activities, and enhances anti-CD3-induced IFN-γ production (3). IL-12 signaling activates a specific set of JAK family kinases (JAK2 and Tyk2) and STAT elements, namely STAT3 and STAT4 (4). During the Th0 to Th1 transition, IL-12R signal transduction also represses the expression of the transcription factor GATA-3, which is nonpermissive for Th1 development and is necessary for Th2 differentiation in response to IL-4 (5).
The efficacy of inhibiting IL-12-mediated signaling and Th1 differentiation has been proven in several models of inflammatory, Th1-dependent disease (6). The anti-inflammatory compound lisofylline (LSF),3 1-(5-R-hydroxhexyl)-3,7-dimethylxanthine, is being used to reduce regimen-related toxicities in patients undergoing bone marrow transplantation (BMT) and appeared to reduce the incidence of severe acute graft-vs-host disease (GVHD) (7), an acknowledged Th1-mediated disease. It was previously shown that LSF prevents onset of experimental autoimmune encephalomyelitis (EAE), the etiology of which requires autoreactive Th1 cells, by inhibiting IL-12 signaling to T cells and by blocking Th1 differentiation in vivo (8). In the present study we have investigated the cellular and molecular bases for inhibition of IL-12 signaling in murine T cells and confirmed the relevance of these to the immunobiology of IL-12 signaling in human T lymphocytes.
Materials and Methods
Animals and reagents
BALB/c mice were purchased from Charles River Laboratories (B & K Universal, Fremont, CA) and housed under specific pathogen-free conditions. Recombinant murine and human IL-2, IL-12, and IL-4 were purchased from Genzyme (Cambridge, MA). Specific Ab for mouse CD3 (145-2C11), CD24 (J11D), CD25 (3C7), CD28 (37.51), CD54 (3E2), CD69 (H1.2F3), and I-Ad/I-Ed (2G9) and for human CD3 (HIT3a), CD8 (HIT8a), CD14 (M5E2), CD19 (B43), CD45RA (HI100), CD45R0 (UCHL1), and HLA-DR,DP,DQ (TU39) were purchased from PharMingen (San Diego, CA). Specific Ab for CD45R B220 (RA-3) was obtained from Cedarlane (Westbury, NY). GATA-3, STAT4, and STAT5-specific antisera were obtained from Santa Cruz Biotechnology (Santa Cruz, CA). Anti-GSK-3β (clone 7) mAb was obtained from Transduction Laboratories (Lexington, KY), and biotinylated anti-phosphotyrosine 4G10 was obtained from Upstate Biotechnology (Lake Placid, NY). Con A was obtained from Sigma (St. Louis, MO). Anti-rat Ig-coated magnetic beads and magnet, and FITC-conjugated anti-rat Ig were obtained from BioSource International (Camarillo, CA). Protein G-Sepharose was supplied by Amersham Pharmacia (Piscataway, NJ). Anti-Ig- and streptavidin-conjugated HRP second-step reagents were purchased from Pierce (Rockford, IL).
T cell isolation, proliferation and immunofluorescence
Murine splenic T cells were enriched from RBC-depleted splenocytes by negative selection using Ab and magnetic beads (BioSource International). Splenocytes were incubated on ice for 15 min with rat anti-mouse I-Ad/I-Ed, CD24, and CD45R; washed, and resuspended at 107/ml before addition of anti-rat Ig-coated magnetic beads at a ratio of 5:1. Suspensions of magnetic bead and cells were rotated for 30 min at 4 C before application to magnets. T cells remaining in suspension were collected, washed twice, and counted before use. The resulting population was routinely >95% T cells as assessed by immunofluorescent staining. T cell proliferation was measured by culturing cells at 1 × 105/well with insoluble anti-CD3 and/or anti-CD28 for 48, and cultures were pulsed during the last 18 h with [3H]thymidine (0.5 μCi/well). [3H]thymidine incorporation was measured on a Betaplate liquid scintillation counter (Wallac, Turku, Finland). In some assays cell-free supernatants were collected and quantitated for IL-2 secretion.
Immunofluorescent staining for cell surface Ags on murine T cells was conducted by resuspending cells at 107/ml in staining buffer (PBS/0.1%sodium azide/0.25% BSA) with rat Abs specific for CD25, CD54, or CD69 (2 μg/ml). After 30 min, cells were washed three times with staining buffer before addition of secondary, FITC-conjugated goat anti-rat Ab for another 15 min. Cells were washed twice more in staining buffer and fixed in 1% paraformaldehyde before analysis on a Coulter EPICS XL (Hialeah, FL).
Human T cells were purified from peripheral blood of healthy donors. Mononuclear cells were obtained by centrifugation through Ficoll-Paque Plus (Pharmacia Upjohn Biotech, Uppsala, Sweden), and T cells were enriched by rosetting with SRBC. After a second centrifugation through Ficoll to obtain rosettes, resting CD4+ T cells were further purified by negative selection with magnetic beads (Dynal, Oslo, Norway) using Abs specific for CD8, CD14, CD19, and HLA-DR,DP,DQ. T cell purity was assessed functionally by proliferation in response to soluble anti-CD3 and lectins as well as by immunofluorescent staining. Further isolation of memory and naive subsets was derived from these highly purified (>99% T cells) preparations by negative selection with CD45RA- or CD45R0-specific Abs.
Induction of Th1 differentiation
Anti-CD3 (4 μg/ml in PBS) was immobilized to tissue culture plates by incubation for 1 h at 37°C. After washing the plates twice in PBS, murine or human T cells (0.5 × 106 cells/ml) were stimulated in the presence or the absence of IL-12 (1 ng/ml), with or without varying concentrations of LSF. Alternatively, murine T cells were initially stimulated with Con A (2.5 μg/ml). After 7 days, equal numbers of viable cells were restimulated overnight at 5 × 105 cells/ml with insoluble anti-CD3 alone, and supernatants were collected and assayed for IFN-γ by ELISA. For the direct induction of IFN-γ by IL-12 in human T cell subsets, CD45R0- or CD45RA-negative T cells were plated at 2 × 106/ml with titrations of IL-12 with or without 40 μM LSF, and cell-free supernatants were collected 72 h later.
Immunoprecipitation and Western blotting
Proteins from detergent-soluble lysates or immunoprecipitates were separated by SDS-PAGE followed by Western blotting to examine their expression. T cells stimulated under various conditions were lysed in Nonidet P-40 lysis buffer (10 mM Tris (pH 7.4), 150 mM NaCl, 1% Nonidet P-40, 1 mM EDTA, and 0.25% sodium deoxycholate) containing 0.5% SDS and boiled at 95 C for 10 min. Lysates were spun at 14,000 × g for 30 min, and detergent-soluble fractions were carefully removed from the insoluble pellet. The protein concentration for each sample was determined by bicinchoninic acid protein assay (Pierce). Immediately before electrophoresis, equal amounts of lysate were boiled for 5 min in 1× reducing sample buffer (6.5 mM Tris, (pH 6.8), 1.25% SDS, and 12.5% glycerol), cooled, loaded onto gel, and separated by SDS-PAGE. Proteins were transferred onto Optitran (Schleicher & Schuell, Keene, NH) overnight. Membranes were blocked with 5% BSA in 1× TBS-T (Tris-buffered saline containing 0.1% Tween-20) for >2 h before incubation with primary Ab. Blots were washed three times in TBS-T followed by incubation with second-step reagents. After three additional washes in TBS-T, proteins were visualized by enhanced chemiluminescence (ECL, Amersham, Aylesbury, U.K.).
For analysis of STAT4 and STAT5, splenic T cells (106/ml) were stimulated with Con A for 3 days, washed, and then rested overnight (2 × 106/ml) before harvesting and three more washes. Blasts were subsequently incubated at 5 × 106/ml in the absence or the presence of increasing concentrations of LSF. After 4 h, T cells were stimulated with either IL-2 or IL-12 for 10 min before cell harvesting and generation of detergent-soluble lysates as described above. One microgram of STAT4- or STAT5-specific antisera was added per 0.25 mg of lysate, rotated at 4 C for 2 h, and followed by addition of protein G-Sepharose. Immunoprecipitates were pelleted by centrifugation and washed four times with lysis buffer before SDS-PAGE, transfer to membrane, and Western blotting with anti-phosphotyrosine or STAT4- or STAT 5-specific antisera.
ELISAs for IFN-γ, TNF-α, and IL-2, and ELISPOT analysis
ELISAs specific for murine IFN-γ, TNF-α, and IL-2 and for human IFN-γ were purchased from Genzyme. The IFN-γ-specific ELISPOT was developed by coating Millipore IP plates (Millipore, Burlington, MA) with 100 μl of anti-IFN-γ Ab R4-6A2 (PharMingen) at 10 μg/ml in PBS for 1 h at 37°C. Plates were washed four times with PBS and blocked with 200 μl of 10% FBS for 30 min at 37°C. After aspiration of wells, equal numbers of T cells from experimental groups were added and restimulated with Con A at 4 μg/ml. After incubation for 2 or 3 days, the plates were washed five times with PBS followed by five washes with PBS and 0.05% (v/v) Tween-20 (Sigma). Biotinylated anti-murine IFN-γ Ab XMG1.2 (PharMingen) at 1 μg/ml in PBS with 4% BSA was added to each well and incubated at room temperature for 1 h. After five more washes with PBS/Tween-20, streptavidin-conjugated HRP (Pierce) was added at 1 μg/ml, and incubation proceeded at room temperature for 30 min, followed by five additional washes. To develop the ELISPOT, one tablet of 3-amino-9-ethyl-carbazole (Sigma) was dissolved in 2.5 ml of dimethylformamide (Sigma), diluted to 50 ml with 0.2 N acetate/acetic acid buffer (pH 5.0), and 35 μl of hydrogen peroxide was added just before use. Plates were incubated for 5 min, washed with water, and allowed to dry.
To measure the relative production of IFN-γ during Th0 to Th1 maturation, Nunc Maxi-sorp (VWR, Brisbane CA) 96-well plates were coated with anti-IFN-γ and washed as described above, except that cells were stimulated with Con A or PHA (Sigma) at 4 μg/ml plus IL-12 at 1 ng/ml, with or without serial dilutions of LSF, 1501-S, or PTX. After incubation for 2 or 3 days and washing exactly as described above, the plates were incubated with 100 μl of 3,3′,5,5′-tetramethyl-benzidine (Sigma) for 5 min, the reaction was stopped by the addition of 100 μl of 0.1 N H2SO4, and absorbance was read at 450 nm. For human T cells, the assay was exactly the same, except that Abs specific for human IFN-γ capture (NIB42) and detection (biotinylated 4S.B3) were used in the sandwich technique (PharMingen).
Total RNA from T cells was isolated by Trizol extraction (Life Technologies, Baltimore, MD). The RNase protection analysis of IFN-γ and Th1 cytokine expression was measured by Riboquant kit (PharMingen) and conducted according to the manufacturer’s instructions. Expression levels of IFN-γ mRNA were quantitated by normalizing to levels of control L32 genes.
LSF inhibits murine Th1 differentiation and production of proinflammatory cytokines
We have previously shown that LSF blocks IL-12-induced Th1 differentiation both in vivo and in vitro, as assessed by inhibition of IFN-γ secretion from T cells restimulated with Ag and anti-CD3, respectively (8). To determine whether the production of other proinflammatory cytokines was also abrogated in vitro, the levels of IL-2 and TNF-α were measured in activated T cell supernatants derived from cells that had undergone differentiation in the presence of IL-12 and LSF. As shown previously for IFN-γ production, IL-12 augmented Th1 differentiation compared with T cells that were stimulated with anti-CD3 alone, and this was inhibited by LSF in a dose-dependent fashion with an estimated concentration causing half-maximal inhibition (IC50) of approximately 16 μM (Fig. 1,A). When the same culture supernatants were examined for IL-2, another prototypical Th1 cytokine, a similar pattern emerged. IL-2 production was enhanced by costimulation of T cells with IL-12 and was inhibited if cocultured with LSF (Fig. 1,B). Although TNF-α is not produced exclusively by Th1 cells (9), TNF-α production was assessed because of its known proinflammatory properties, which have been implicated in a number of inflammatory diseases (10); TNF-α production was also inhibited by LSF in a dose-dependent fashion, albeit less effectively than were the prototypical Th1 cytokines IFN-γ and IL-2 (Fig. 1,C). The inhibition of proinflammatory cytokine production by LSF was specific for Th1 cells differentiated in the presence of IL-12 and was not due to nonspecific inhibition of T cell cytokine secretion. For example, even the relatively low levels of IFN-γ and TNF-α produced by Th2 cells that underwent differentiation in the presence of IL-4 (data not shown) were unaffected when cells were treated with IL-4 in the presence of LSF (Fig. 1, A and C).
LSF decreases the frequency and potency of IL-12-induced, murine Th1 lymphocytes
The ability of LSF to block IL-12-induced Th1 maturation may result from a decrease in the number of effector Th1 cells, a relative decrease in IFN-γ production at the clonal level, or both. To distinguish among these possibilities, ELISPOT analysis was performed on TCR-restimulated T cells that had been primed to undergo Th1 differentiation in the presence of IL-12 and LSF. T cell differentiation was induced as before, except that the T cells were initially activated with Con A with or without IL-12 and/or LSF, and restimulated with Con A, instead of anti-CD3. It was previously determined that this caused no differences in IL-12-mediated Th1 differentiation or the ability of LSF to inhibit in this assay when IFN-γ was measured by conventional ELISA (data not shown). The frequency and magnitude of IFN-γ secretion by Th1 cells were qualitatively assessed by colorimetric development of the ELISPOT assay.
Compared with T cell cultures primed with Con A alone, those costimulated in the presence of IL-12 produced greater numbers of individual IFN-γ secreting clones, as assessed by the number of spots in wells (Fig. 2,A). As expected, costimulation with IL-4 suppressed the appearance of IFN-γ-secreting T cells. In addition, the relative intensity of individual spots with IL-12 was greater than that without cytokine, indicating the enhanced ability of individual clones to produce larger amounts of IFN-γ. If Th1 differentiation with IL-12 was induced in the presence of LSF, a concentration-dependent decrease in the number of IFN-γ-producing clones was observed, indicating that the maturation of clonal Th1 effectors by IL-12 was blocked by LSF (Fig. 2,A). The decrease in IFN-γ secretion at the clonal level was coincident with a generalized decrease in the relative intensity of individual spots, showing that LSF not only blocked Th1 differentiation at the cellular level, but also inhibited the ability of individual clones to become high level IFN-γ producers (Fig. 2,A). The decreased frequency and intensity of clonal spots correlated with a decrease in IFN-γ as measured in the supernatants by standard ELISA (data not shown). When T cells that were differentiated in the presence of IL-12 and LSF were assessed for their levels of IFN-γ mRNA after restimulation, a dose-dependent decrease in IFN-γ mRNA was observed in cultures primed in the presence of LSF (Fig. 2 B). Thus, LSF decreases the frequency and potency of Th1 effectors generated in the presence of IL-12, which correlates with decreased levels of IFN-γ mRNA.
LSF abrogates the differentiation of naive human T cells into Th1 and Tc1 subsets
To extend the previous results from murine to human T cells, the ability of LSF to inhibit IL-12-induced differentiation of naive human T cells into Th1 and Tc1 effectors was evaluated in highly purified CD4+ and CD8+ T cells isolated from peripheral blood. CD40-mediated IL-12 production by APC is triggered through CD154 (CD40 ligand) expression on activated T cells (3). To limit endogenous IL-12 secretion in the differentiation assay, cell preparations were rigorously depleted of contaminating APC as assessed by FACS analysis for B cell (CD19) and macrophage (CD11b, CD14) cell surface Ags; the lack of APC was also functionally confirmed by the inability of T cells to proliferate in response to soluble anti-CD3 or to lectins such as PHA (data not shown). To obtain naive T cells, this population was further depleted of the CD45R0+ memory cells. To recover sufficient numbers of T cells after 1 wk of stimulation in the absence of APC, all T cell cultures were initially stimulated with a combination of insoluble anti-CD3 and insoluble anti-CD28 with exogenous IL-2.
CD4+ and CD8+ human T cells that were primed with IL-12 displayed enhanced Th1 and Tc1 differentiation compared with duplicate cultures that were stimulated in the absence of IL-12 (Fig. 3). Both Th1 and Tc1 cells secreted enhanced levels of IFN-γ, with Tc1 cells producing up to 30-fold more IFN-γ than Th1 on a per cell basis. The magnitude of IFN-γ secretion was proportional in a relatively linear fashion to the number of restimulated T cells. When T cells were induced to undergo differentiation with IL-12 in the presence of 40 μM LSF, the pattern of inhibition previously seen with murine T cells was observed. Thus, LSF treatment led to greatly reduced levels of effector IFN-γ secretion regardless of the number of Th1 or Tc1 cells that were induced to secrete IFN-γ during restimulation.
The effects of LSF on naive and memory human T cells
In addition to blocking naive T cell differentiation into either the Th1 or Tc1 inflammatory phenotype in response to IL-12, LSF was assessed for its ability to inhibit IL-12 signaling in human memory T cells as well. Since memory T cells, by definition, have previously seen Ag and can be polarized toward particular Th subsets, the effects of LSF were not evaluated in an assay that measures helper subset differentiation. Instead, LSF was tested for its capacity to block IFN-γ secretion in response to IL-12 alone, exploiting the capacity of IL-12 to induce IFN-γ secretion directly from T cells in the absence of any other stimulus (3).
Highly purified, CD4+ naive (CD45R0 negative) and memory (CD45RA negative) T cells were induced to secrete IFN-γ in response to IL-12 in a concentration-dependent fashion, and as expected, the memory pool produced much larger amounts of IFN-γ (Fig. 4). This was not due to the presence of preactivated cells within the memory pool, as the T cell preparations were intentionally depleted of MHC class II-positive cells. If T cells were cocultured with LSF along with IL-12, significant inhibition of IL-12 induced IFN-γ secretion was observed regardless of the concentration of IL-12 employed. In general, memory cells producing greater amounts of IFN-γ in response to higher IL-12 concentrations were more resistant to inhibition by LSF. Nonetheless, these data show that LSF abrogates IL-12 signaling in naive and memory T cells in the absence of any additional stimulus, demonstrating its ability to block an IL-12-dependent response.
The inhibitory effects of LSF are demonstrable early during Th0 to Th1 development and do not affect T cell activation
Upon initial encounter with Ag, naive T cells are primed to become either Th1 or Th2 depending on the immunogen, the type of APC, and the presence of IL-12 or IL-4, respectively (2). Since there are no absolute cell surface Ag that define Th1 or Th2 cells, the Th phenotype can be established only after secondary stimulation through the TCR in the absence of the polarizing cytokines, IL-12 and IL-4 (9). IL-12 not only enhances Th1 differentiation, but also augments the acute production of IFN-γ from anti-CD3-activated T cells. It was therefore hypothesized that if LSF inhibited Th1 differentiation by blocking IL-12 signaling, a block in IL-12-enhanced IFN-γ production from anti-CD3-stimulated T cells should be demonstrable during priming, when the Th0 to Th1 transition occurs.
Murine T cells that were primed with Con A alone displayed enhanced IFN-γ production in the presence of IL-12 (Fig. 5), which was clearly demonstrable by 48 h after activation and which in our model system is a time that coincides with cell commitment to the Th1 phenotype (data not shown). When T cells were costimulated with Con A and IL-12 along with LSF, a dose-dependent decrease in IFN-γ production was observed, such that IFN-γ secretion returned to the levels observed in T cells that were stimulated with Con A alone in the absence of IL-12 (Fig. 5,A). We have previously shown that the block in IL-12-mediated Th1 differentiation was not achieved by the enantiomer of LSF, 1501-S, or by its oxidized, achiral analogue PTX (8). In agreement with the inability of these structurally related compounds to abrogate IL-12-induced Th1 differentiation, both were shown to be ineffective as inhibitors of IL-12-enhanced IFN-γ secretion from Con A-activated T cells (Fig. 5, B and C). LSF was also evaluated for a similar effect on IL-12-enhanced IFN-γ production in partially purified human T cells that were stimulated with PHA and IL-12. As observed above for murine T cells, the augmented IFN-γ production in response to IL-12 in PHA-activated human T cells was also abrogated by LSF (Fig. 5 D).
IL-12-induced Th1 differentiation requires both the presence of IL-12 and stimulation through the TCR, which allows for optimal acquisition of IL-12 responsiveness. This is due in part to up-regulation of the IL-12R β2-chain to form a high affinity receptor (11). The ability to inhibit Th1 differentiation in response to exogenous IL-12, therefore, may result from a block in IL-12 signaling or suboptimal TCR activation. However, it was previously shown that LSF does not block several TCR-generated signals, as measured by 1) IFN-γ secretion in response to anti-CD3 alone, 2) acquisition of IL-12 responsiveness with respect to IL-12-induced proliferation of T cell blasts, and 3) the recovery of Th1 cells after 1 wk of priming with anti-CD3 alone or anti-CD3 and IL-12 (8).
Additional parameters of T cell activation were examined in the presence of LSF to rule out any other possible inhibitory effects on T cell activation. The stimulation of murine splenic T cell proliferation with insoluble anti-CD3 was not inhibited by LSF at concentrations up to160 μM, i.e., 10-fold the IC50 observed for inhibition of IL-12-induced Th1 differentiation (Fig. 6,A). Similarly, T cell proliferation stimulated with a combination of insoluble anti-CD3 and insoluble anti-CD28 was not abrogated by LSF (Fig. 6,B), whereas stimulation by either anti-CD3 or anti-CD3 and anti-CD28 was blocked by CsA (Fig. 6, E and F). Note that at concentrations equal to or in excess of 250 μM, LSF did inhibit T cell proliferation in response to anti-CD3 alone or with costimulation by anti-CD28 (Fig. 6, C and D). At these higher concentrations of LSF the predicted nonspecific methylxanthine effects are observed, e.g., inhibition of the type IV cAMP phosphodiesterases (PDE) (12, 13, 14), which, as shown for other methylxanthines, block lymphocyte proliferation by a cAMP-dependent mechanism (15, 16) (data not shown).
Murine T cells were also assessed for the up-regulation of early (CD69) and late (CD54) activation Ags in response to Con A plus IL-12 in the presence of LSF. CD69 and CD54 were both up-regulated 24 h after Con A stimulation, which increased after 48 h, but which was not enhanced by IL-12 costimulation (Fig. 6, G and H). Up-regulation of neither activation Ag was affected by coincubation of the cells with 40 μM LSF (Fig. 6, G and H), nor was there any decrease in the mean fluorescence of positive cells (data not shown). Additional studies examined up-regulation of CD25 expression in response to either Con A or Con A with IL-12 as well as IL-2 secretion in response to anti-CD3, and none of these responses was inhibited by LSF (data not shown). In all cases examined to date, there is no evidence for LSF inhibition of T cell activation at concentrations that block IL-12 signaling and IL-12-mediated Th1 differentiation.
Altered expression and activation of transcription factors that regulate Th1 development
The differentiation of Th0 cells into either Th1 or Th2 cells is influenced by several transcription factors that act both positively and negatively with regard to the Th phenotype and that are thought to regulate maturation of effector function (17, 18, 19, 20). GATA-3 is a transcription factor whose expression is necessary for Th2 development and whose expression is actively repressed by IL-12 signaling (5, 21). Furthermore, forced ectopic expression of GATA-3 during primary T cell activation precludes the normal maturation of Th1 cells induced by IL-12 (5). To determine whether the effects of LSF on Th1 differentiation are consistent with the expression of GATA-3 in developing Th1 cells, expression of GATA-3 was measured in activated murine T lymphocytes cultured in the presence of IL-12, with or without LSF. In agreement with the results of Ouyang et al. (5), GATA-3 expression was repressed in cells costimulated with IL-12 relative to that in cells stimulated with Con A alone and was greatly enhanced when naive T cells were costimulated with IL-4 (Fig. 7). Consistent with its inhibitory effects on Th1 differentiation, LSF derepressed GATA-3 protein levels that occurred in the presence of IL-12, with expression returning to levels seen in cells stimulated with Con A alone, but not to the elevated levels seen when T cells were primed with IL-4 (Fig. 7). LSF did not alter GATA-3 expression in T cells stimulated with Con A alone (data not shown) and had little effect on the expression of GSK-3β or β-catenin (Fig. 7), indicating that the changes in GATA-3 protein expression were specific and were not due to global changes in protein expression.
To determine whether LSF blocked more proximal events that are necessary for IL-12R-mediated signaling (22, 23), the ability to inhibit IL-12-induced increases in STAT4 tyrosine phosphorylation was assessed. While IL-12-stimulated T cell blasts displayed rapid increases in STAT4 tyrosine phosphorylation that peaked after 10 min of stimulation (data not shown), T cell blasts that were incubated with LSF before IL-12 addition exhibited a concentration-dependent decrease in STAT4 activation (Fig. 8,A). Western blotting with STAT4-specific reagents showed that this was specific for tyrosine phosphorylation and was not due to generalized decreases in STAT4 expression. To test whether LSF inhibited activation of other STAT isoforms, duplicate samples of T cell blasts were also pretreated with LSF before activation with IL-2. Similar to that observed with IL-12 stimulation, LSF inhibited IL-2-mediated tyrosine phosphorylation of STAT5 in a dose-dependent fashion without decreasing the overall levels of STAT 5 (Fig. 8 B).
IL-12 is a pleiotropic cytokine that regulates the differentiation of naive Th0 cells into proinflammatory Th1 cells (6), the Th subset that mediates inflammatory diseases such as GVHD (24). The data herein show that LSF is an effective inhibitor of IL-12 signaling in both murine and human T cells at concentrations that do not affect T cell activation or proliferation in response to TCR signaling. These findings extend those of a previous study in which LSF showed efficacy in the Th1-mediated autoimmune model EAE and the concurrent reduction of Th1 differentiation in vivo (8).
It was found that LSF was an effective inhibitor of IL-12-driven Th1 differentiation in vitro. The resulting effector population was deficient in the production of several proinflammatory cytokines upon secondary TCR stimulation, including IFN-γ, IL-2, and TNF-α (Fig. 1). The effect on Th1 differentiation by LSF was 2-fold, leading to decreases in the number of differentiated Th1 clones that secrete IFN-γ as well as a reduction in the relative amount of IFN-γ produced by individual clones (Fig. 2,A). Inhibition of IFN-γ production was regulated at the mRNA level, since T cells that were primed in the presence of IL-12 and LSF showed reduced levels of steady state IFN-γ mRNA after TCR restimulation (Fig. 2,B). It is unclear whether LSF inhibits expression or activation of transcription factors in Th1 cells that are necessary to drive high level IFN-γ production, although LSF does derepress GATA-3 expression in response to IL-12 (Fig. 7), to a level that is nonpermissive for Th1 maturation (5). We are currently examining whether LSF alters epigenetic modifications to chromatin, such as DNA methylation, that are operative in differentiating Th1 or Th2 cells (25, 26).
The inhibition of IL-12-mediated Th1 differentiation by LSF correlated with an early block in IL-12-dependent signaling during the Th0 to Th1 transition, shortly after T cells are stimulated through the TCR. Thus, IL-12-enhanced IFN-γ production by TCR-stimulated cells was abrogated by LSF in both murine and human T cells (Fig. 5. A and D), and consistent with our previous observations for IL-12-mediated Th1 maturation (8), the inhibitory effect of LSF on IL-12-enhanced IFN-γ production was stereoselective and was not blocked by 1501-S or by PTX. Importantly, LSF does not appear to abrogate IL-12 signaling by limiting acquisition of IL-12 responsiveness, since LSF is not inhibitory for anti-CD3-induced T cell activation or proliferation (Fig. 6), nor does it block IL-12-dependent proliferation of T cell blasts if present during anti-CD3-induced blastogenesis but removed before addition of IL-12 (8).
The early blockade of IL-12 signaling by LSF during Th1 differentiation was also manifest as derepression of GATA-3 expression. Since GATA-3 is nonpermissive for IL-12-dependent Th1 differentiation (5), its derepression by LSF is consistent with the drug effect on T cell biology. It is also noteworthy that LSF restored GATA-3 expression to the levels observed in the absence of IL-12 costimulation, but did not raise GATA-3 levels to those seen when T cells were induced to undergo Th2 differentiation in the presence of IL-4 (Fig. 7). This is consistent with our observations that there is no polarization toward a Th2 phenotype in vitro when IL-12-induced Th1 differentiation was inhibited by LSF (data not shown), and that inhibition of Th1-dependent EAE in vivo by LSF did not result in enhanced Th2 cytokine production by Ag-restimulated cells ex vivo (8).
One of the earliest events in IL-12-mediated signaling cascades is the coordinate activation of Jak/STAT family members that help regulate nuclear translocation of STAT 3/4 and drive cytokine-dependent gene expression (4). LSF blocked IL-12-induced tyrosine phosphorylation of STAT4 in a dose-dependent fashion without causing decreases in STAT4 expression (Fig. 8,A). Surprisingly, LSF also inhibited IL-2-mediated activation of STAT5 (Fig. 8,B), even though LSF had no effect on anti-CD3-induced T cell activation or proliferation (Fig. 6) or on IL-2 secretion or IL-2-induced proliferation of T cell blasts (data not shown). The capacity of LSF to block IL-2-dependent STAT5 activation without affecting IL-2-induced proliferation may relate to the independence of IL-2-induced mitogenesis from STAT5 activation (27, 28, 29), whereas IL-12-induced mitogenesis and Th1 differentiation are critically dependent on STAT4 (22, 23). The ability of LSF to inhibit cytokine-dependent responses, including IL-4-mediated Th2 differentiation (8), could result from its ability to preferentially block activation of distinct STAT members and the relative dependence of cytokine receptor-mediated responses on particular STATs. Alternatively, it is also possible that the inhibition of STAT4 and STAT5 activation by LSF may be unrelated to its efficacy at mitigating cytokine responses and Th1 disease in vivo.
The ability of cAMP-specific or nonselective PDE inhibitors to ameliorate EAE and other Th1-mediated disease is largely attributed to their efficacy at blocking APC production of IL-12 and TNF-α as well as T cell activation (30, 31). Elevated levels of intracellular cAMP have also been shown to block Jak/STAT activation (32, 33, 34) and promote IL-10 secretion by APC (35, 36), causing immune deviation and a shift toward Th2 predominance, as seen in studies using PTX (37, 38). The pharmacological activities of certain methylxanthines at levels ranging from several hundred micromolar to millimolar concentrations correlates with their inhibitory effects on classes of cyclic nucleotide PDEs (13, 14). At the relatively low concentrations of LSF here employed to block IL-12 signaling, IL-12 secretion by APC in response to LPS or anti-CD40 is not inhibited, and LSF treatment does not cause Th2 polarization (8) or inhibit T cell activation (Fig. 6). Furthermore, the inability of LSF to block T cell activation or proliferation is inconsistent with PDE inhibition, since PDE activity has been shown to be up-regulated and necessary for proper T cell responsiveness (39). Coupled with the inability of PTX or 1501-S to block IL-12-dependent signaling (Fig. 5), the data suggest that LSF may mitigate IL-12 signaling via a different mechanism that is at least in part cAMP independent.
The ability of LSF to inhibit IL-12-dependent signaling in murine T cells has now been extended to include human T cells as well. Although LSF apparently reduced the incidence of severe GVHD following BMT in humans (7), the underlying mechanism remained unclear. Our experimental models in vitro and our in vivo doses are within the range of concentrations measured in plasma from clinical trial patients that were administered LSF. Thus, the present data support a mechanism by which LSF may inhibit Th1-associated diseases such as GVHD by abrogating IL-12 signaling and the generation of inflammatory Th1 effectors.
We thank Drs. Robert Lewis and Jack Singer for critical reading of the manuscript, and Lara Porter and Kima Nalis for technical assistance.
This work was supported by National Institutes of Health Small Business Technology Transfer Grant 2R42NS35762-02.
Abbreviations used in this paper: LSF, lisofylline; BMT, bone marrow transplantation; GVHD, graft-vs-host disease; EAE, experimental autoimmune encephalomyelitis; ELISPOT, enzyme-linked immunospot; 1501-S, S-enantiomer of LSF; PTX, pentoxifylline; PDE, phosphodiesterase.