Sphingosine 1-Phosphate Cross-activates the Smad Signaling Cascade and Mimics Transforming Growth Factor-β-induced Cell Responses*

Exposure of renal mesangial cells to sphingosine 1-phosphate (S1P) leads to a rapid and transient activation of the mitogen- and stress-activated protein kinases but also the protein kinase B. Here, we show that S1P also induces phosphorylation of Smad proteins, which are members of the transforming growth factor-β (TGF-β) signaling device. However, Smad phosphorylation occurred more slowly with a maximal effect after 20–30 min of S1P stimulation when compared with the rapid activation of the MAPKs. Interestingly, Smad phosphorylation is increased by pertussis toxin, which is in contrast to the complete inhibition of S1P-induced MAPK phosphorylation by pertussis toxin. TGF-β is a potent anti-inflammatory cytokine, which in mesangial cells attenuates the expression of (i) inducible nitricoxide synthase (iNOS) caused by interleukin (IL)-1β, (ii) secreted phospholipase A2 (sPLA2), and (iii) matrix metalloproteinase-9 (MMP-9). These gene products are also down-regulated by S1P in a concentration-dependent manner. Furthermore, the expression of connective tissue growth factor is enhanced by both TGF-β2 and S1P. These effects of S1P are not mediated by the MAPK cascade as neither pertussis toxin nor the MAPK cascade inhibitor U0126 are able to reverse this inhibition. Overexpression of the inhibitory Smad-7 or down-regulation of co-Smad-4 lead to a reversal of the blocking effect of S1P on IL-1β-induced NO release. Moreover, down-regulating the TGF-β receptor type II by the siRNA technique or antagonizing the S1P3 receptor subtype with suramin abrogates S1P-stimulated Smad phosphorylation. In summary, our data show that S1P trans-activates the TGF-β receptor and triggers activation of Smads followed by activation of connective tissue growth factor gene transcription and inhibition of IL-1β-induced expression of iNOS, sPLA2, and MMP-9.

Renal mesangial cells not only play an important physiological role in the regulation of glomerular filtration rate but also importantly contribute to most pathological processes of the renal glomerulus (1)(2)(3). Together with increased extracellular matrix production and inflammatory mediator secretion, me-sangial cell proliferation is a hallmark of many forms of glomerulonephritis ultimately leading to glomerulosclerosis (1,2). The detailed mechanisms regulating cell proliferation, however, are still not completely understood, although the involvement of the mitogen-activated protein kinase (MAPK) 1 cascade is generally assumed (4 -6). During the course of mesangioproliferative glomerulonephritis, the formation of transforming growth factor-␤ (TGF-␤) is dramatically increased (7,8), and antibodies against TGF-␤ attenuate the histological manifestations of the disease (8).
The TGF-␤ family comprises small homo-and heterodimeric polypeptides originally purified from platelets (9), placenta (10), and kidney (11) that regulate quite diverse cellular activities including differentiation, extracellular matrix production, and cell proliferation (12,13). They initiate these cellular responses by binding to the TGF-␤ receptor, which is a 70-kDa transmembrane protein with a cytoplasmic serine/threonine kinase domain (14,15). In the kidney, high affinity TGF-␤ receptors are expressed on glomerular epithelial cells (podocytes) and mesangial cells (16).
For signaling, the type II receptor requires both its kinase activity triggering autophosphorylation (15) and association with the 53-kDa type I TGF-␤ receptor, which also interacts with TGF-␤ (14,17). Signaling events that are initiated by TGF-␤ are comprised of the phosphorylation and homotrimerization of Smad proteins, which then assemble with a distinct Smad homotrimer to a heterohexamer and translocate to the nucleus where gene transcription is either activated or suppressed (18). In addition to the Smad pathway, mitogen-activated protein kinase cascades are also activated by TGF-␤ (19,20). Sphingosine 1-phosphate (S1P) is a sphingolipid signaling molecule generated from ceramide by the action of neutral ceramidase followed by sphingosine kinase (for review, see Refs. [21][22][23]. It has gained increasing attention, as S1P has mitogenic properties in many cell types (24,25). Cell surface receptors for S1P belong to the endothelial differentiation gene (Edg) family (26), which also includes the lysophosphatidic acid receptors. Recently, the Edg receptors have been renamed S1P * This work was supported by grants from the Deutsche Forschungsgemeinschaft (HU 842/2-3 and PF361/2-1) and the Dr. H. Schleussner-Foundation for Immune Pharmacology. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
In this study, we show that S1P is able to cross-activate the TGF-␤ signaling pathway in renal mesangial cells and thereby can mimic TGF-␤-induced cell responses including inhibition of proinflammatory gene induction such as the inducible NO synthase (iNOS), secretory phospholipase A 2 (sPLA 2 ), and matrix metalloproteinase-9 (MMP-9).
Cell Culture-Rat renal mesangial cells were cultivated and characterized as described previously (29). Passages 7-21 were used for the experiments in this study.
Cell Stimulation and Western Blot Analysis-Confluent mesangial cells in 60-mm diameter dishes were stimulated with the indicated substances in Dulbecco's modified Eagle's medium containing 0.1 mg/ml of fatty acid-free bovine serum albumin. Thereafter, the medium was withdrawn. The cells washed once with ice-cold phosphate-buffered saline solution and scraped into ice-cold lysis buffer (50 mM Tris/HCl, pH 7.4, 150 mM NaCl, 10% glycerol, 1% Triton X-100, 2 mM EDTA, 2 mM EGTA, 40 mM ␤-glycerophosphate, 50 mM sodium fluoride, 10 g/ml leupeptin, 10 g/ml aprotinin, 1 M pepstatin A, 1 mM phenylmethylsulfonyl fluoride) and homogenized by 10 passes through a 26-gauge needle fitted to a 1-ml syringe. Samples were centrifuged for 10 min at 14,000 ϫ g, and the supernatant was taken for protein determination. Cell extracts containing 50 g of protein were prepared in SDS-sample buffer and subjected to SDS-PAGE. Proteins were transferred on to nitrocellulose paper, and immunostaining was performed as described previously in detail (5). Antibodies were diluted in blocking buffer as indicated in the legends of the figures. Bands were detected by the enhanced chemiluminescence (ECL) method as recommended by the manufacturer.
Cross-linking Experiments-Lysates of unstimulated and stimulated cells in buffer containing 20 mM Hepes, 1% Triton X-100, 2 mM EDTA, 2 mM EGTA were incubated for 30 min at room temperature with 2 mM cross-linker, which was dissolved as 50 mM stock solution in dimethyl sulfoxide. Samples were quenched with 2 mM Tris, pH 7.4, for 15 min. Thereafter, SDS buffer was added, and samples were separated on SDS-PAGE and further analyzed by Western blot analysis.
Determination of NO Formation-Quiescent mesangial cells in 24well plates were stimulated as indicated, and the supernatants were taken for nitrite determination using Griess reagent as described (30).
Determination of Secreted sPLA 2 Activity-Supernatants of stimulated mesangial cells were subjected to sPLA 2 activity assay exactly as described previously (31).
Cell Transfections-Mesangial cells were plated in 30-mm diameter dishes and grown to 50% confluence and then transfected using Lipo-fectAMINE according to the manufacturer's instructions. For transfections 1 g of either pcDNA 3.1 vector alone or pcDNA 3.1 vector containing full-length cDNA of mouse Smad-7 were used. 48 h posttransfection, cells were taken for stimulation. Gene silencing was performed using sequence-specific siRNA reagents of: rat Smad-4 (AA-UACACCGACAAGCAAUGACdTdT and GUCAUUGCUUGUCGGUG-UAUUdTdT) and rat TGF␤R II (AAAGUCGGUUAACAGCGAUCU-dTdT and AGAUCGCUGUUAACCGACUUUdTdT). Mesangial cells were transfected with a 200 nM concentration of the 21-nucleotide duplexes using OligofectAMINE as recommended by the manufacturer (Dharmacon). After 48 h cells were stimulated as indicated in the figure legends to Figs. 9 and 11. The silencing efficiency was detected by Western blot analyses using specific antibodies.
Northern Blot Analysis-Total RNA was isolated using guanidinium isothiocyanate solution. 30 g of RNA was separated by electrophoresis on 1% formaldehyde-agarose gels. RNA was transferred to a nylon membrane and cross-linked by UV light. Blots were hybridized with a 470-bp reverse transcriptase-PCR product of rat iNOS (forward primer, CCT TTG CTA CTG AGA CAGG; reverse primer, GGG ATC TGA ATG CAA TGTT), with a 416-bp reverse transcriptase-PCR product of rat IIA-sPLA 2 (forward primer, GGT CCT CCT GTT GCT AGC AG; reverse primer, CTT TGC AAA ACT TGT TGG GG), with a 700-bp probe of rat MMP-9 (32), and with a 357-bp reverse transcriptase-PCR product of rat CTGF (forward primer, GCC CTG TGA AGC TGA CCT AG; reverse primer, GAG TTC GTG TCC CTT ACT CCC). All probes were labeled with [␣-32 P]dCTP using the Multiprime DNA labeling system (Amersham Biosciences). Hybridization was carried out at 42°C for 20 h, and the membranes were exposed on an imaging system. To correct for variations in RNA amounts, blots were finally rehybridized with a 32 P-labeled GAPDH cDNA probe. Statistical Analysis-Statistical analysis was performed using oneway analysis of variance followed by a Bonferroni's post hoc test for multiple comparisons (GraphPad InStat Version 3.00 for Windows NT, GraphPad Software, San Diego, CA).

Sphingosine 1-Phosphate Activates the S1P Receptors and
Leads to Increased MAPK Activities-Stimulation of mesangial cells with S1P leads to a rapid activation of all three major MAPK cascades. Activation of these cascades can be detected by measuring the phosphorylation of the different members of these cascades. As seen in Fig. 1, a substantial increase in the phosphorylation of the classical p42/p44-MAPKs occurs already after 2 min of S1P stimulation (Fig. 1A, upper panel), and this increase in the degree of phosphorylation is maintained for 10 min but declined rapidly thereafter. In contrast, the total p42/p44 protein levels did not change upon stimulation (Fig.  1A, lower panel). A similar rapid and transient increase in phosphorylation is seen for the stress-activated p38-MAPK ( Fig. 1B), the stress-activated protein kinases SAPK/JNKs (Fig. 1C), and also for the protein kinase B (Fig. 1D).
Sphingosine 1-Phosphate Cross-activates the TGF-␤ 2 Receptor and Stimulates the Smad Signaling Cascade-In addition to the rapid activation of the three MAPK cascades, S1P also causes an enhanced phosphorylation of Smad proteins. As seen in Fig. 2A, upon S1P stimulation phosphorylation of Smad-2 at Ser 465 and Ser 467 is increased with a more delayed time course when compared with the MAPKs. Maximal phosphorylation occurs after 15-20 min of stimulation. As a positive control TGF-␤ 2 was employed, which is a well known activator of Smads leading to increased phosphorylation of these transcription factors (13,33). TGF-␤ 2 causes a comparably delayed phosphorylation of Smad-2 as found for S1P (Fig. 2B). Smad-1 was also investigated, as Smad-1 is known to transduce signals from another member of the TGF-␤ receptor superfamily, the bone morphogenetic protein receptor (34). Smad-1 causes an increase in phosphorylation at Ser 463/465 upon S1P and TGF-␤ 2 stimulation (Fig. 2, A and B). Moreover, Smad-3 is another important member of the Smad family that is activated by the TGF-␤ 2 receptor (35). To elucidate whether S1P also phosphorylates Smad-3, mesangial cells were labeled with [ 32 P]orthophosphate and then subjected to stimulation with S1P and TGF-␤ 2 followed by cell lysis and immunoprecipitation of Smad-3. Fig. 2C shows that both S1P and TGF-␤ 2 induce phosphorylation of Smad-3 at 50 kDa. Surprisingly, two phosphorylated bands appear at ϳ50 and 55 kDa. This strongly suggests that other Smad members like the 55-kDa Smad-1 or -2 are also recognized by the anti-Smad-3 antibody, although the antibody (FL-425) was raised against recombinant full-length Smad-3. The stimulating effect of S1P on Smad phosphorylation occurs in a concentration-dependent manner as seen in Fig. 3. However, only concentrations of 100 nM and above (for Smad-2) and 1 M and above (for Smad-1) significantly increase Smad phosphorylation.
Interestingly, the S1P-induced Smad-2 phosphorylation is not abolished by pertussis toxin an inhibitor of G i/o proteins (36) but rather increased (Fig. 4, upper panel), while the activation of the p42/p44-MAPK cascade by S1P is dose-dependently blocked by pertussis toxin (Fig. 4, middle panel). Again, the total amount of MAPK does not change throughout the stimulation period (Fig. 4, lower panel). Pertussis toxin alone has no effect on Smad nor p42/44-MAPK phosphorylation. These data propose that the p42/p44-MAPK may have a negative effect on Smad activation, which has also been suggested by several other groups in different cell types (37)(38)(39).
Sphingosine 1-Phosphate Stimulation Mimics TGF-␤ 2 -induced Cell Responses-We then investigated whether S1P is able to mimic TGF-␤-mediated cell responses. In mesangial cells, TGF-␤ 2 inhibits the formation of NO (40), prostaglandins (41), and extracellular matrix (42) caused by pro-inflammatory cytokines. As seen in Fig. 5A, IL-1␤-induced nitrite formation, which reflects the inducible NO synthase activity, is blocked by TGF-␤ 2 . A similar reduction is seen with S1P, showing first significant reductions at 100 nM S1P and reaching approximately 90% inhibition at 10 M S1P. The reduction of NO release is paralleled by a transcriptional down-regulation of IL-1␤-induced iNOS mRNA as demonstrated by a Northern blot analysis (Fig. 5B). Moreover, S1P also reduces sPLA 2 activity, which is considered to be the rate-limiting step in prostaglandin formation (Fig. 6A). Yet, the inhibitory effects of Thereafter, cells were harvested, and Western blot analyses were performed using specific anti-phospho-Smad-1 and -2 antibodies and total Smad-2 antibodies at a dilution of 1:1000 each. Bands were detected by the ECL method. Data are representative of two independent experiments giving similar results.
TGF-␤ 2 and S1P were less pronounced as compared with NO release. Again, the reduced sPLA 2 activity is due to reduced mRNA transcription of the group IIA sPLA 2 (Fig. 6C) and also to reduced protein synthesis and secretion (Fig. 6B). Furthermore, MMP-9 mRNA expression is also blocked by S1P in a concentration-dependent manner as seen in a Northern blot analysis (Fig. 7A). Similar to the NO release, the reduction of MMP-9 mRNA expression by S1P is only detectable by concentrations of 100 nM and above (Fig. 7A). This reduction of IL-1␤-triggered MMP-9 mRNA expression by S1P and TGF-␤ 2 is also reflected on the amount of MMP-9 protein secreted into the supernatant (Fig. 7B). Stimulation of cells with S1P or TGF-␤ 2 alone has no effect per se on either iNOS, sPLA 2 , or MMP-9 expression (data not shown).
Since TGF-␤ 2 is also known to positively affect transcription of certain genes like the connective tissue growth factor (CTGF) (43), we investigated whether this positive effect is also mimicked by S1P. As seen in Fig. 8A, 4 h of stimulation with S1P S1P (in M) or TGF-␤ 2 (in ng/ml). Thereafter, supernatants were collected and taken for sPLA 2 activity assay (A) or used to precipitate proteins and subjected to Western blot analysis using a monoclonal anti-sPLA 2 antibody at a dilution of 1:60 (B). Cells were extracted for RNA and sPLA 2 , and GAPDH mRNA expression was measured by Northern blot analysis (C) as described under "Experimental Procedures." Results in A are expressed as percent of maximal IL-1␤ stimulation and are means Ϯ S.D. (n ϭ 4). *, p Ͻ 0.05; **, p Ͻ 0.01; ***, p Ͻ 0.001, considered statistically significant compared with the IL-1␤-stimulated value. Data in B and C are representative of two to four independent experiment giving similar results.
up-regulates CTGF protein expression in a dose-dependent manner, but for significant changes at least 100 nM of S1P is required. In parallel, CTGF mRNA expression is also enhanced by S1P (Fig. 8B).
To elucidate whether the inhibitory effect of S1P on IL-1␤induced NO release is mediated by the MAPK cascade or by the Smad signaling cascade, we tested pertussis toxin. However, neither the inhibitory effect of S1P nor that of TGF-␤ 2 on IL-1␤-induced NO release is reversed by pre-treatment with pertussis toxin (Fig. 9A), thus indicating that the MAPK cascade is not involved in these inhibitory effects exerted by S1P. This conclusion is further substantiated by the finding that the classical MAPK cascade inhibitor U0126 does not reverse the inhibitory effect of either S1P or TGF-␤ 2 (Fig. 9A). In contrast, transient overexpression of mesangial cells with a construct encoding the full-length inhibitory Smad-7 (44) causes a reversal of the blocking effect of TGF-␤ 2 and also S1P on IL-1␤induced NO release (Fig. 9B), suggesting that the inhibitory effect of S1P on NO release is indeed mediated by the Smad signaling cascade. However, the reverting effect of Smad-7 overexpression on the S1P-mediated response seems only moderate and cannot be increased under any condition. This is due to the low transfection efficiency of mesangial cells, which is a well known problem of these cells (45). Furthermore, we transfected mesangial cells with siRNA of Smad-4, which is a co-Smad required for translocation of Smad-1, -2, and -3 to the nucleus. As seen in Fig. 9C, also depletion of Smad-4 is able to significantly reverse the effect of S1P on NO release, resembling the partial effect of Smad-7 overexpression.
TGF-␤R II Moves to a Higher Molecular Complex upon S1P and TGF-␤2 Stimulation-To elucidate whether S1P stimulation induces a heterologous receptor interaction, we used two chemical cross-linkers, the non-cleavable amino-reactive bifunctional DSS and the photoreactive, unselective BASED. Ly-sates from cells stimulated with either S1P or TGF-␤ 2 were incubated with or without a 2 mM concentration of the two cross-linkers. As seen in Fig. 10, the TGF-␤ R II stains as a 70-kDa protein in control lysates. In the presence of DSS, a shift toward three slower migrating complexes (at about 100, 140, and Ͼ200 kDa) is detected. Upon S1P or TGF-␤ 2 stimulation, the intensity of the 100-and 140-kDa complexes increased slightly. The same complexes are also observed in the presence of BASED although the intensity is less pronounced. Moreover, since BASED is a cleavable cross-linker sensitive to dithiothreitol, we added dithiothreitol to the BASED-cross-linked lysates. Under this condition the higher molecular complexes disappear again (Fig. 10).
Finally, the involvement of the TGF-␤R II in S1P signaling was investigated by depleting cells of the receptor using siRNA. Fig. 11A shows that down-regulation of the TGF-␤R II abrogates not only TGF-␤ 2 -stimulated (Fig. 11A, upper right panel) but also S1P-induced Smad-1 phosphorylation (Fig. 11A, upper  left panel). To detect the silencing efficiency of the siRNA treatment a Western blot analysis of the same cell lysates using an antibody against TGF-␤R II is shown in Fig. 11A, lower panels. The subtype of S1P receptor involved in Smad activation at least includes the S1P 3 because suramin, which has been reported to be a selective S1P 3 receptor antagonist compared with the other S1P receptor subtypes in vitro (46), antagonizes S1P-stimulated Smad-1 phosphorylation (Fig.  11B), whereas it has no effect on MAPK phosphorylation (data not shown). DISCUSSION Exogenous S1P triggers many important cellular responses including proliferation (24,25), differentiation (47), cytoprotection (48), and cell migration (49). It has been extensively documented that S1P binds to specific cell surface receptors the endothelial differentiation gene receptors (Edg) (26), which have now been renamed to S1P 1-5 receptors (27). Despite the existence of these cell membrane receptors, S1P generated inside the cell can also trigger intracellular events independent of the S1P receptors. However, the identity of these putative intracellular targets of S1P still remains to be unvealed.
In this study, we present the first evidence that exogenous S1P, by binding to its receptor, cross-activates the Smad signaling cascade in renal mesangial cells, which is classically triggered by the cytokine TGF-␤. Thus, S1P can mimic TGF-␤mediated cell responses, which adds a new facet of action to this highly interesting molecule, and this may well be of clinical advantage.
TGF-␤ exerts a number of apparently opposing biological effects. Under chronic renal disease, it exerts profibrotic activities leading to glomerulosclerosis and interstitial fibrosis (50). In acute disorders, TGF-␤ acts as a potent anti-inflammatory agent to reduce pro-inflammatory parameters including NO and prostaglandins in resident tissue cells and macrophages (40,51,52). In addition, TGF-␤ suppresses the respiratory burst of macrophages and stimulates the synthesis of IL-1 receptor antagonist. Furthermore, TGF-␤ is a potent immunosuppressive agent, which by far exceeds the potency of drugs like cyclosporin A (50,53), and it has been proposed to promote the migration and homing of T cells (50,54).  The finding that MAPK inhibition does not reverse S1Pinhibited NO release but rather potentiates the inhibiting effect (Fig. 9A) together with the data that pertussis toxin even increased Smad activation (Fig. 4), strongly suggests a negative effect of the MAPKs on Smad activation. Such a negative regulatory effect on Smad-2 and -3 has already been reported for epithelial cells (38). When transfecting these cells with oncogenic Ras, a loss of TGF-␤-induced cell response was obtained, which was due to inhibited Smad nuclear translocation. This effect was reversed by mutations of the MAPK phosphorylation site within Smad-2 or -3 (38). Moreover, Sowa et al. (39) reported that in mouse osteoblasts not only the classical MAPKs but also the SAPK/JNKs are able to negatively regulate Smad-3-induced transcriptional activity.
Additional strong evidence for the involvement of the Smad signaling cascade in S1P-mediated cell responses is given by the findings that: (i) overexpression of Smad-7, which is an endogenous inhibitor of the Smad signaling cascade (45), is able to reverse, at least partially, the negative effect of S1P on NO release (Fig. 9B), and (ii) cellular depletion of Smad-4 by the siRNA technique exerts the same effect as Smad-7 overexpression in terms of nitrite release (Fig. 9C).
Presently, it is not yet resolved whether S1P-triggered activation of the MAPK cascades and activation of the Smad signaling cascade are mediated by the same S1P receptor, but it is tempting to speculate that different receptor subtypes including pertussis toxin-insensitive ones may be involved. In this context, it was reported that S1P 1 also couples to G 13 (55). Moreover, S1P 2 and S1P 3 can also signal through G 12/13 and G q (55,56), and Ohmori et al. (57) showed that S1P 3 receptor activation leads to a pertussis toxin-insensitive focal adhesion kinase-dependent cell motility. Our data with suramin ( Fig.  11B) clearly suggest that at least the S1P 3 receptor is involved in the Smad cross-activation by S1P. Although being a rather unselective antagonist at various purinoceptors and growth factor receptors, suramin has been reported to be a selective S1P 3 receptor antagonist compared with the other S1P receptor subtypes in vitro (46). Whether additional S1P receptor subtypes are also involved in the Smad activation remains to be investigated.
Whereever the branching point is located within the S1P cascade, at least one link leads to the TGF-␤R II. This is supported by the findings that S1P stimulation increases the appearance of higher molecular complexes of TGF-␤R II (Fig.  10) and the studies using siRNA of the receptor that abrogates the S1P-stimulated Smad activation (Fig. 11). Further studies will be needed to unravel the exact cross-link between these two receptors as well as the identity of the cross-linked proteins appearing in Fig. 10.
Recently, Sato et al. (58) reported that ceramide exerts a stimulating effect on basal and TGF-␤-induced collagen promoter activity in fibroblasts, whereas exogenous S1P and overexpression of the sphingosine kinase-1 inhibited promoter activity. This contrasts to our data, which show clearly no effect of ceramide or ceramide 1-phosphate on TGF-␤ signaling in mesangial cells (data not shown) but a clear positive effect of S1P. A possible explanation for these contrasting findings could be that fibroblasts express a different pattern of S1P receptors than mesangial cells. It may be speculated that fibroblasts are especially rich in G i -coupled S1P receptors, which strongly activate the MAPK cascade and, in turn, exert a negative effect on TGF-␤ signaling (37)(38)(39).
During the last years it has become evident that different types of receptors including G-protein-coupled receptors not only undergo homodimerization but also heterodimerization or oligomerization with other receptor classes and thereby trigger a cross-talk between different signaling cascades (59). In this context, it is worth mentioning that the S1P receptors also cross-activate other growth factor signaling cascades. Recently, it was shown that vascular endothelial growth factor, epidermal growth factor, and also platelet-derived growth factor receptors can be transactivated by S1P stimulation and this leads to an enhanced mitogenic response in various cell types (60 -63). In the present study we provide evidence for a novel type of transactivation between GPCRs exemplified by the S1P receptors and the TGF-␤ receptor family. Mesangial cells do express S1P 1-5 (28), and it has recently been suggested that especially S1P 2 and S1P 3 are involved in the mitogenic function of S1P in these cells (28). The authors also propose a crossactivation of the PDGF receptor cascade.
A further link between S1P and a typical TGF-␤ response, the immunosuppression, was recently presented independently by two groups (64 -66). It was shown that the immunosuppressive drug FTY720 acts as a S1P mimetic and prevents homing of T lymphocytes to lymph organs (64,65). Obviously, FTY720 serves as a substrate of the sphingosine kinase, and the phosphorylated form of FTY720 exerts the same immunosuppressive effects as the FTY720 itself suggesting that FTY720 is only a prodrug activated by intracellular phosphorylation. Interestingly, the phosphorylated form of FTY720 is able to bind to S1P receptors, thus rendering it a potentially useful agonist of the S1P receptor (64, 65). Moreover, differential binding affinities FIG. 12. Schematic overview of the hypothetical cross-communication between S1P receptors and TGF-␤ signaling in mesangial cells. ERK, extracellular signal-regulated protein kinase; S1P-R, S1P receptor; I and II, TGF-␤ receptor types 1 and 2, respectively. have been shown for the S1P 4 receptor, which exerts a 20-fold higher affinity for the phospho-FTY720 than for the natural ligand S1P (65).
The fact that S1P can exert anti-inflammatory and immunosuppressive effects in vivo makes the regulation of S1P formation by the S1P generating enzymes a key task for future research. These enzymes not only include the sphingosine kinases of which two isoforms have been described, the sphingosine kinase 1 and sphingosine kinase 2, but also the upstream located neutral ceramidase, which exerts an equally important function by supplying the substrate for the sphingosine kinases.
In summary, we have shown that S1P cross-activates the TGF-␤ signaling cascade and leads to activation of at least three Smad proteins, i.e. Smad-1, -2, and -3, with subsequent gene transcription (Fig. 12). In this way, S1P exerts an antiinflammatory potential just similar to TGF-␤ and may represent a novel target for therapeutic strategies to cope with inflammatory and/or immune diseases.