Transcriptional and Translational Regulation of Transforming Growth Factor-β Production in Response to Apoptotic Cells

Abstract

Interaction between apoptotic cells and phagocytes through phosphatidylserine recognition structures (PSRS) results in the production of transforming growth factor-β (TGF-β) which has been shown to play pivotal roles in the anti-inflammatory and anti-immunogenic responses to apoptotic cell clearance. Using 3T3-TβRII and RAWTβRII cells in which a truncated dominant negative TGF-β eceptor II was stably transfected in order to avoid autofeedback induction of TGFβ, we investigate the mechanisms by which TGF-β as produced through PSRS engagement. We show, in the present study, that TGF-β was regulated at both transcriptional and translational steps. P38 MAPK, ERK and JNK were involved in TGF-β transcription whereas translation required activation of Rho GTPase, PI-3 kinase, Akt and mTOR with subsequent phosphorylation of translation initiation factor eIF4E. Strikingly, these induction pathways for TGF-β production were different from those initiated in the same cells responding to LPS or PMA.

Introduction

Deletion of apoptotic cells in vivo involves recognition of surface changes on the cells leading to their ingestion by unique phagocytic mechanisms – a process that we have termed efferocytosis (1–5). Accompanying this recognition, whether or not the apoptotic cell is actually engulfed by the phagocyte, is the generation of anti-inflammatory mediators (6–9) and the establishment of a generally anti-inflammatory and anti-immunogenic local environment. We have earlier suggested that TGF-β is a major mediator of this response and that a number of secondary anti-inflammatory effects result from the autocrine/paracrine actions of the active TGF-β produced.

The TGF-β-family comprises more than 60 structurally related growth and differentiation factors that play important roles in regulation of numerous physiological processes, including cell proliferation, differentiation, apoptosis, early embryonic development, and extracellular matrix protein synthesis (10–14). TGF-β exerts its effects through a heteromeric receptor complex consisting of type I and II transmembrane serine/threonine kinase receptors (15). The most well defined signaling pathways of TGF-β are through Smad family members and the MAP kinase family (11, 16, 17). In mammals, there are three isoforms of TGF-β (TGF-β1, β2 and β3), which are structurally identical and have similar bioactivities. TGF-β may be released as a result of apoptotic cell interaction with inflammatory cells, such as macrophages, but also by structural cells such as airway epithelial cells, endothelial cells and fibroblasts.

During apoptosis, a number of changes occur on the plasma membrane that contribute to recognition of these dying cells by potential phagocytes. One of these is phosphatidylserine (PS), normally restricted to the inner membrane leaflet, but exposed on the outer leaflet as a consequence of loss of membrane phospholipid asymmetry during apoptosis (18, 19). There is considerable evidence to support a major role for recognition of this PS in the production of TGF-β and the anti-inflammatory effects of apoptotic cells (6, 8, 9, 20–22). Thus, in our previous studies, we have demonstrated that interaction of phagocytes with apoptotic cells through PS results in production of the active TGF-β both in vitro and in vivo (6, 8, 9). Its anti-inflammatory effect and TGF-β-dependency has been shown in the inflamed lung instilled with apoptotic Jurkat cells (8) and anti-immunogenic effects likewise following PS administration during an adaptive immune response (22).

The objective of this study was to determine the signal pathways involved in the generation of TGF-β by apoptotic cell stimulation. Unfortunately, the receptor(s) that recognizes PS (PS recognition structures, PSRS) that is responsible for this effect is unknown. A number of recent papers have described candidate PS receptors (23–27) but only one of these (28) was studied for its potential induction of TGFβ. In addition, a number of bridge molecules that bind apoptotic cell PS and link this to various phagocyte receptors have been described (29–33). At this point it is not clear which of these are involved in TGFβ synthesis and release. Accordingly, the experiments herein used whole apoptotic cells as stimulus. In addition, to avoid the complexity of assay systems that may be confounded by the presence of both apoptotic as well as responder cells we have included studies with an IgM monoclonal antibody (mAb217) created earlier by immunizing with macrophages that were active in PS-recognition (34). This antibody was used originally to identify a candidate PS receptor that was originally termed PSR (34) but is now known not to serve this function (35, 36) but rather to belong to the Jumonji family of proteins (JMJD6) and to act as a histone demethylase (37)). We now interpret this original identification as due to non-specific binding of the antibody in the phage display. Nonetheless, this Mab217 antibody binds to, and activates macrophages and other potential phagocytes, mimicking exactly the effects of PS exposed on apoptotic cells in contributing to uptake, and the generation of anti-inflammatory mediators including TGF-β (7, 22, 38). Accordingly, the studies reported here have used intact apoptotic cells and mAb217 to stimulate macrophages and fibroblasts for TGF-β production. Including the antibody stimulation allows us to bypass any confusion that may result from contribution of signaling molecules from the apoptotic cells themselves as well as to provide a coordinated stimulus where needed to avoid the issue of the time it takes individual apoptotic cells and responding cells to interact. TGFβ production may be regulated at many steps, including transcription, translation, secretion and especially activation in the extracellular environment. It is likely that apoptotic cell stimulation can alter each of these steps. However, in this study we have focused on only the first two. The results suggest regulation of both transcription and translation by pathways that differ substantially from those used by other stimuli of TGF-β synthesis.

Materials and methods

Antibodies and Reagents

TGF-β1 was from R&D Systems. Lipopolysaccharide (LPS, Escherichia coli 0111:B4) was from List Biological Lab. SB 203580, PD 98059, JNK inhibitor II, wortmannin, LY 294002, rapamycin, and Protease Inhibitor Cocktail Set I were from Calbiochem-Novabiochem. Actinomycin D and cycloheximide were from Sigma Chemical Co. Gene Specific Relative RT-PCR kit was from Ambion Inc. Advantage RT-for-PCR kit was from BD Biosciences. Lipofectamine Plus reagent was from GibcoBRL. Rho Activation Assay kit and recombinant C3 transferase were from Cytoskeleton. Phospho-ERK (E-4), ERK-2 (K 23), p38 (c-20), TGF-β (V)-G, phospho-eIF4E and total eIF4E antibodies were from Santa Cruz Biotech. Anti-p38 MAPK phospho-specific antibody, Phospho-SAPK/JNK (Thr183/Tyr185), phospho-Akt, β-actin, phospho-mTOR and mTOR antibodies were from Cell Signaling. Generation of monoclonal IgM antibody 217 and its IgM control was described by Fadok et al (34). Induction of apoptotic Jurkat cells and characterization of apoptotic and control cells were described previously (7).

Stable transfection of truncated TGF-β receptor II

Stable cell lines of 3T3TβRII, 3T3V, RAWTβRII and RAWV cells were made by transfecting truncated TGF-β receptor II or empty vector. Briefly, pcDNA3.1 plasmids with or without MYC-tagged truncated TGF-β receptor II sequence were transfected into 3T3-L1 and RAW 264 cells using Lipofectamine Plus reagent according to the manufacturer's instructions. Seventy-two hours after transfection, the cells were incubated in the fresh medium containing 500 µg/ml G418 for 4 weeks. Cell colonies resistant to G418 were isolated and screened by limited dilution.

Cell Culture and ELISA Assay

The cells (5.0 × 10⁵ cells/well) were plated in each well of a 24-well tissue culture plate and incubated overnight in Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 10% fetal bovine serum, L-glutamine (2 mM), penicillin (100 U/ml), and streptomycin (100 µg/ml) under a humidified 10% (3T3 cells) or 5% (RAW cells) CO₂ atmosphere at 37°C before stimulation for 18 h in the serum free DMEM. Total TGF-β in cell culture supernatant was quantitated by enzyme-linked immunosorbent assay (ELISA) according to the instructions of the manufacturer (ELISA Tech, Denver, Colorado).

Transient Cell Transfection and Reporter Gene Assays

TGF-β-responsive luciferase reporter pSBE-Luc (SBE, Smad Binding Elements) was transfected into 3T3TβRII and 3T3V cells using Lipofectamine Plus reagent according to the manufacturer's instructions. pSV-β-galactosidase vector (Promega) was co-transfected as internal control to measure differences in transfection efficiency. Luciferase and β-galactosidase activities were measured 18 h after TGF-β stimulation using Luciferase Assay System (Promega) and Galacto-Light (Tropix), respectively. Dominant negative RhoA N19 and constitutively active RhoA V14 (kindly provided by Dr. Jiancheng Hu) and empty vector were transiently transfected into 3T3TβRII cells using Lipofectamine Plus reagent. After 48 h, the cells were stimulated with mAb 217 for 30 min, and total cell lysates were analyzed using immunoblotting for phospho-eIF4E and phospho-Akt.

Relative Quantitative Reverse Transcription Polymerase Chain Reaction of TGF-β Messenger RNA

Total RNA was isolated from cultured cells using TRIzol reagent (GIBCO BRL). The concentration and purity of RNA were evaluated by spectrophotometry at 260 and 280 nm. Reverse transcription (RT) was carried out for 60 min at 42°C with 1 µg total RNA using Advantage RT-for-PCR kit (BD Biosciences). TGF-β mRNA level was determined using Relative Quantitative RT-PCR kit (Ambion Inc.). The cDNA was denatured for 5 min at 94°C, and the amplification was achieved in a temperature cycler (Perkin Elmer GeneAmp PCR System 2400) by 24 cycles of temperature (94°C for 30 s, 57°C for 30 s, and 72°C for 30 s) followed by a 7-min final extension at 72°C. A total of 5 µl of each PCR sample was loaded on a 1.5% agarose gel stained with ethidium bromide. The relative fluorescence of TGF-β vs. 18 S rRNA was analyzed by densitometry.

RhoA Activation Assay

3T3TβRII (5 × 10⁶ cells) cells were plated in 10-cm tissue culture dishes. Clarified cell lysates from 12 h serum-starved cells were used for Rho activity assay according to the manufacturer’s instructions (Cytoskeleton). To confirm equal loading of proteins from each sample, 40 µl of total cell lysate from each sample was blotted with β-actin antibody before binding to GST tagged Rhotekin RBD protein.

Immunoblotting Analysis

Immunoblotting analysis was carried out as described previously with some modification (39). Briefly, cells (3.0 × 10⁵ cells/well) were plated in each well of a 12-well tissue culture plate and incubated in serum free medium overnight before use. Following stimulation, the cells were lysed in lysis buffer (20 mM HEPES (pH7.4), 150 mM NaCl, 1 mM DTT, 0.5% Triton X-100 and 1 X Protease Inhibitor Cocktail Set I), resolved on 10% SDS-PAGE, and blotted to nitrocellulose membranes. The membranes were probed with primary antibodies at 4°C overnight and incubated with horseradish peroxidase-conjugated secondary antibodies for 1 h at room temperature. Proteins were visualized by enhanced chemiluminescence (Amersham Pharmacia Biotech) according to the manufacturer's instructions. Equal loading of proteins in each lane was confirmed by Ponceau S staining or re-probed with corresponding antibodies against the native proteins or β-actin antibody (17).

Statistical analysis

All data are presented as means ± SEM from three or more separate experiments. The means were analyzed using ANOVA for multiple comparisons. When ANOVA indicated significance, the Tukey-Kramer honestly significant difference test for all pairs was used to compare groups. All data were analyzed using JMP (version 5; SAS Institute, Inc., Cary, NC) Statistical Software for the Macintosh.

Results

Effect of apoptotic cells or mAb 217 on TGF-β production

To pursue the mechanisms by which apoptotic cells or mAb 217 induced TGF-β production in a manner that avoided auto-stimulation by TGF-β itself, we made stable cell lines that were unresponsive to TGF-β by transfecting truncated TGF-β receptor II (or empty vector as control) constructs into 3T3-L1 cells and RAW 264 cells (9) to block the paracrine and autocrine effects from TGF-β (40, 41). Blockade of TGF-β signaling was verified using TGF-β-responsive luciferase reporter SBE-luc (SBE, Smad Binding Elements) assay. As shown in Figure 1A, 3T3TβRII cells with truncated TGF-β receptor II did not respond to TGF-β stimulation. Moreover, incubation of these cells with apoptotic Jurkat cells or mAb 217 resulted in reduced amounts of TGF-β in the conditioned medium (Fig. 1B) relative to similarly-treated 3T3V cells, supporting the supposition that in the normal circumstance, the TGF-β that was induced by apoptotic cells could indeed provide positive autostimulation feedback with enhanced overall production of the mediator.

FIGURE 1. Apoptotic cells or mAb 217 induces TGF-β production in 3T3-L1 cells.

A, 3T3-L1 cells stably transfected with truncated TGF-β receptor II (3T3TβRII) or empty vector (3T3V) were transiently co-transfected with SBE-luc and pSV-β galactosidase constructs. After 48 h, the cells were incubated in the presence of TGF-β (10 ng/ml) for 18 h. The luciferase assays, which were normalized to β-galactosidase, are expressed as fold enhancement. *, significantly different from 3T3V cells stimulated with TGF-β. B, 3T3V and 3T3TβRII cells were cultured in the presence of mAb 217 (100 µg), isotype IgM control (100 µg), apoptotic Jurkat cells (ApoJ) or viable Jurkat cells (ViableJ) for 18 h. Total TGF-β in the conditioned medium was analyzed by ELISA. *, significantly different from 3T3V cells.

Effect of apoptotic cells or mAb 217 on TGF-β mRNA expression

As previously described and as shown in Figure 1, interaction of non-professional, or professional phagocytes with apoptotic cells (or stimulation with mAb 217) induces the production of TGF-β (6, 9, 34). In the present study, apoptotic cells or mAb 217 were each shown to induce TGF-β mRNA expression; detectable at 1 h and reaching a plateau from 2 h in 3T3TβRII cells (Fig. 2A). As expected, TGF-β mRNA expression was significantly inhibited by actinomycin D; however, not by the protein synthesis inhibitor, cycloheximide suggesting that new protein synthesis was not required for the induction of TGF-β transcription (Fig. 2B). To rule out the possibility that the increase in TGF-β mRNA was caused by enhancement of TGF-β message stability, 3T3TβRII cells were first treated with PMA (100 nM) overnight to increase the steady state TGF-β mRNA level, and then washed and treated with actinomycin D (10 µg/ml) in the absence or presence of mAb 217. The remaining TGF-β mRNA level after actinomycin D treatment was measured using Relative Quantitative RT-PCR. As shown in Figure 2C, mAb 217 did not affect TGF-β mRNA stability. These findings suggest that the upregulation is at the level of transcription.

FIGURE 2. Effect of apoptotic Jurkat cells or mAb 217 on TGF-β mRNA levels.

A, 3T3TβRII cells were cultured in the presence of viable Jurkat cells, apoptotic cells, isotype IgM (100 µg/ml) or mAb 217 (100 µg/ml) for the time indicated; B, 3T3TβRII cells were pre-treated with the indicated concentrations of actinomycin D (ActD) or cycloheximide (CHX) for 1 h, before stimulation for 4 h; C, 3T3TβRII cells were cultured in the presence of PMA (100 nM) overnight to increase the steady state TGF-β mRNA level, and then the cells were incubated with actinomycin D (10 µg/ml) in the presence or absence of mAb 217 (100 µg/ml) for the time indicated. TGF-β mRNA levels were analyzed by Relative Quantitative RT-PCR.

Requirement for MAP kinases in apoptotic cell-induced transcriptional upregulation of TGF-β production

When 3T3TβRII cells were stimulated with apoptotic Jurkat cells (data not shown) or 217 mAb, phosphorylation of p38 MAPK, ERK and JNK were shown to be increased with time, reaching a maximum at 15 min for p38 MAPK and ERK, and 30 min for JNK (Fig. 3A). To examine involvement of these MAP kinases in the TGF-β production, 3T3TβRII cells were pretreated with SB 203580, a specific p38 MAPK inhibitor, PD 98059, a specific MEK-1 inhibitor, and JNK inhibitor II for one hour, and then incubated with apoptotic Jurkat cells or mAb 217 for 18 hours. As shown in Figure 3B and 3C both TGF-β protein and enhanced mRNA production was reduced by all three inhibitors at concentrations shown to be effective in previous work and without toxicity (17, 42). These findings suggest that MAPKs might play important roles in apoptotic cell-induced TGF-β transcription.

FIGURE 3. Roles of MAP kinases in apoptotic cell or mAb 217-induced TGF-β production.

A, 3T3TβRII cells were stimulated with mAb 217 (100 µg/ml) for the time indicated, total cell lysates were immunoblotted for phospho-p38 MAPK, phospho-ERK and phospho-JNK. B, 3T3TβRII cells were pre-treated with SB 203580 (10 µM), PD 98059 (30 βM) or JNK inhibitor II (30 βM) for 1 h and then stimulated with apoptotic Jurkat cells (upper panel) or mAb 217 (lower panel) for 18 h. Total TGF-β in the conditioned medium was measured by ELISA. *, significantly different from apoJ or mAb 217 alone. C, 3T3TβRII cells were pre-treated with SB 203580 (10 µM), PD 98059 (30 µM) or JNK inhibitor II (30 µM) for 1 h and then stimulated with apoptotic cells (upper panel) or mAb 217 (lower panel) for 4 h. TGF-β mRNA level was analyzed by Relative Quantitative RT-PCR. Ratios of TGF-β and 18 SrRNA were measured using densitometric analysis. *, significantly different from apoJ alone.

Involvement of RhoA activation in apoptotic cell-induced TGF-β translation

Small GTP-binding proteins of the Rho family have been found to play an important role in efferocytosis of apoptotic cells (4, 43–46). Shown in Fig 4A and B, is the activation of RhoA in 3T3TβRII cells after exposure to apoptotic Jurkat cells or mAb 217. There was no change in the total levels of Rho in the cells over the time course of the experiments (data not shown). The Rho activation was completely inhibited by C3 transferase (Fig. 4B). Consistent with a role for Rho in synthesis of TGF-β, C3 transferase suppressed its production (Fig. 4C). However, blockade of Rho activation did not affect TGF-β mRNA expression (Fig. 4D), suggesting that Rho acts at a post-transcriptional step.

FIGURE 4. Requirement for RhoA in apoptotic cell or mAb 217-induced TGF-β production.

A, 3T3TβRII cells were stimulated with mAb 217 (100 µg/ml) for the time indicated and the levels of activated Rho determined (see methods); B, 3T3TβRII cells were pre-incubated with C3 transferase (0.3 µg/ml) for 48 h and then stimulated with apoptotic Jurkat cells or mAb 217 for 30 min. C, 3T3TβRII cells were pre-incubated with C3 transferase (0.3 µg/ml) for 48 h and then stimulated with mAb 217 for 18 h, TGF-β in the conditioned medium was measured by ELISA. *, significantly different from mAb 217 alone. D, 3T3TβRII cells were pre-incubated with the indicated concentrations of C3 transferase for 48 h and then stimulated with mAb 217 for 4 h, TGF-β mRNA levels were analyzed by Relative Quantitative RT-PCR.

Effect of apoptotic cells on TGF-β translation through RhoA/PI-3K/Akt/mTOR/eIF4E

To examine the mechanisms by which translational regulation of TGF-β occurred, 3T3TβRII cells were stimulated with mAb 217, and phosphorylation of translation initiator factor eIF4E was determined. As shown in Figure 5A, phosphorylated eIF4E became detectable at 5 min and reached maximum at 30 min after stimulation. Importantly, the phosphorylation was inhibited by C3 transferase, and this was further confirmed by overexpression of the dominant negative RhoAN19. Moreover, overexpression of the constitutively active RhoAV14 increased phosphorylation of eIF4E, supporting a requirement of RhoA for TGF-β protein translation (Fig. 5B).

FIGURE 5. Requirement for Rho activation in mAb 217-induced Akt and eIF4E phosphorylation.

A, 3T3TβRII cells were stimulated with mAb 217 (100 µg/ml) for the time indicated and total cell lysates immunoblotted for phospho-eIF4E. B, 3T3TβRII cells were pre-incubated with the indicated concentrations of C3 transferase for 48 h and then stimulated with mAb 217 for 30 min (upper panel). 3T3TβRII cells were transiently transfected with empty vector, dominant negative RhoA N19 or constitutively active RhoA V14. After 48 h, the cells were incubated in the presence or absence of mAb 217 for 30 min (lower panel). Total cell lysates were immunoblotted for phospho-Akt or phospho-eIF4E.

It has been reported that PI-3K, Akt and mTOR can act upstream of eIF4E (reviewed in Ref. (47)). In addition to activation of MAPKs, apoptotic cell or mAb 217 each stimulate phosphorylation of Akt (time course not shown) and, as depicted in Fig 5B, this was inhibited by C3 transferase. Accordingly, when 3T3TβRII cells were treated with the PI-3 Kinase inhibitors wortmannin or LY 294002 for one hour before stimulation, phosphorylation of Akt and eIF4E were both inhibited (Fig. 6A and B). Moreover, constitutively active RhoAV14 increased phosphorylation of Akt and eIF4E (Fig. 5B lower panel).

FIGURE 6. PI3 kinase and Akt activation are downstream of RhoA.

A, 3T3TβRII cells were pre-treated with the indicated concentrations of wortmannin (upper panel) or LY 294002 (lower panel) for 1 h and then stimulated with mAb 217 (100 µg/ml) for 30 min (immunoblotting) or 4 h (TGF-β mRNA). Phospho-Akt and phospho-eIF4E were examined by immunoblotting, and TGF-β mRNA levels were analyzed using Relative Quantitative RT-PCR. B, 3T3TβRII cells were pre-treated with wortmannin (Wort, 100 nM) or LY 294002 (LY, 100 µM) for 1 h and then stimulated with viable or apoptotic Jurkat cells for 30 min. Total cell lysates were immunoblotted for phospho-eIF4E. Ratios of p-eIF4E and β-actin were measured densitometrically. *, significantly different from apoJ alone.

It has been shown that mTOR is an important mediator downstream of PI-3K/Akt for eIF4E phosphorylation (47, 48). Consistent with these findings, rapamycin inhibited mAb 217-induced mTOR phosphorylation as expected but also blocked phosphorylation of eIF4E (Fig. 7A). By contrast, mTOR phosphorylation was not altered by the MAP kinase inhibitors SB 203580, PD 98059 or JNK inhibitor II (Fig. 7A). A role for the PI-3K/Akt/mTOR pathway in TGFβ translation was supported by finding that the PI-3K and mTOR inhibitors were able to block the production of TGF-β protein but had no effect on levels of mRNA (Figs 7B and C). Collectively, these findings suggest that apoptotic cells regulate TGF-β translation through activation of RhoA/PI-3K/Akt/mTOR/eIF4E.

FIGURE 7. RhoA/PI-3K/mTOR pathways are involved in TGF-β protein production.

A, 3T3TβRII cells were pre-treated with indicated concentrations of rapamycin overnight, SB 203580 (SB, 10 µM), PD 98059 (PD, 50 µM), JNK inhibitor II (JNKII, 50 µM), wortmannin (Wort, 100 nM) or LY 294002 (LY, 100 µM) for 1 h and then stimulated with mAb 217 (100 µg/ml) for 30 min. Total cell lysates were immunoblotted for phospho-mTOR and phospho-eIF4E. B, 3T3TβRII cells were pre-treated with rapamycin (100 nM) overnight, wortmannin (100 nM) or LY 294002 (50 µM) for 1 h and then stimulated with apoptotic Jurkat cells for 18 h. Total TGF-β in the conditioned medium was analyzed by ELISA. *, significantly different from apoJ alone. C, 3T3TβRII cells were pre-treated with rapamycin (Rapa, 100 nM) overnight, wortmannin (100 nM) or LY 294002 (100 µM) for 1 h and then stimulated with apoptotic Jurkat cells for 4 h. TGF-β mRNA levels were analyzed using Relative Quantitative RT-PCR. Ratios of TGF-β and 18 SrRNA were measured densitometrically.

Apoptotic cells or mAb 217 regulate TGF-β production via unique signal pathways

PMA and LPS are both able to stimulate TGF-β production (7, 49, 50). As shown in Fig 8A, PMA was found to induce increased levels of TGF-β protein in our system in a fashion that was inhibited by blockade of the MAP kinases but not by wortmannin or rapamycin. As expected, the MAP kinase inhibitors also prevented the upregulation of TGF-β mRNA in response to PMA (Fig 8B) but in this case, no effect of wortmannin or rapamycin was seen. PMA has been reported to regulate protein translation through p38 MAPK-mediated eIF4E phosphorylation (51–53) and as shown in Fig 8C, both SB 203580 and JNK inhibitor II did block PMA-induced eIF4E phosphorylation. It appears, therefore, that p38 MAPK, ERK and JNK are involved in both TGF-β transcription and possibly translation induced by PMA but that the PI-3K and mTOR pathways are not required for this stimulus.

FIGURE 8. Effect of inhibitors on PMA-induced TGF-β production.

A, 3T3TβRII cells were pre-treated with rapamycin (Rapa, 100 nM) overnight, SB 203580 (SB, 10 µM), PD 98059 (PD, 30 µM), JNK inhibitor II (JNKII, 30 µM) or wortmannin (Wort, 100 nM) for 1 h and then stimulated with PMA (100 nM) for 18 h. Total TGF-β in the conditioned medium was analyzed by ELISA. *, significantly different from PMA alone. B, 3T3TβRII cells were pre-treated with rapamycin (Rapa, 100 nM) overnight, SB 203580 (SB, 10 µM), PD 98059 (PD, 50 µM), JNK inhibitor II (JNKII, 50 µM) or wortmannin (Wort, 100 nM) for 1 h and then stimulated with PMA (100 nM) for 4 h. TGF-β mRNA levels were analyzed using Relative Quantitative RT-PCR. C, 3T3TβRII cells were pre-treated with rapamycin (Rapa, 100 nM) overnight, SB 203580 (SB, 10 µM), PD 98059 (PD, 50 µM), JNK inhibitor II (JNKII, 50 µM) or wortmannin (Wort, 100 nM) for 1 h and then stimulated with PMA (100 nM) for 30 min. Total cell lysates were immunoblotted for phospho-eIF4E.

Similar to PMA stimulation, LPS-induced TGF-β protein production was inhibited by SB 203580, PD 98059, JNK inhibitor II but not by wortmannin or rapamycin (Fig. 9A). However, LPS-induced TGF-β mRNA expression was only substantially blocked by the p38 MAPK inhibitor, SB 203580 (Fig. 9B). Surprisingly, LPS-induced eIF4E phosphorylation was inhibited by PD 98059 (MEK inhibitor) and JNK inhibitor II but not by SB 203580, wortmannin or rapamycin (Fig 9C). These findings suggest a) TGF-β production in response to different stimuli is differently regulated (Table 1); b) apoptotic cells or mAb 217 induce TGF-β production through unique signaling pathways, by which p38 MAPK, ERK and JNK are involved in transcription, and RhoA/PI-3K/Akt/mTOR/eIF4E are involved in translation (Fig. 10).

FIGURE 9. Effect of inhibitors on LPS-induced TGF-β production.

A, RAWTβRII cells were pre-treated with rapamycin (Rapa, 100 nM) overnight, SB 203580 (SB, 10 µM), PD 98059 (PD, 30 µM), JNK inhibitor II (JNKII, 30 µM) or wortmannin (Wort, 100 nM) for 1 h and then stimulated with LPS (100 ng/ml) for 18 h. Total TGF-β in the conditioned medium was analyzed by ELISA. *, significantly different from LPS alone. B, RAWTβRII cells were pre-treated with rapamycin (Rapa, 100 nM) overnight, SB 203580 (SB, 10 µM), PD 98059 (PD, 50 µM), JNK inhibitor II (JNKII, 50 µM) or wortmannin (Wort, 100 nM) for 1 h and then stimulated with LPS (100 ng/ml) for 4 h. TGF-β mRNA levels were analyzed using Relative Quantitative RT-PCR. C, RAWTβRII cells were pre-treated with rapamycin (Rapa, 100 nM) overnight, SB 203580 (SB, 10 µM), PD 98059 (PD, 50 µM), JNK inhibitor II (JNKII, 50 µM) or wortmannin (Wort, 100 nM) for 1 h and then stimulated with LPS (100 ng/ml) for 30 min. Total cell lysates were immunoblotted for phospho-eIF4E.

Table 1.

Comparison of kinases and signal pathways involved in TGF-β production.

	p38 MAPK	ERK	JNK	Rho/PI3K/mTOR/eIF4E
ApoJ or mAb 217	Transcription	Transcription	Transcription	Translation
PMA	Transcription Translation	Transcription	Transcription Translation	Not involved
LPS	Transcription	Translation	Translation	Not involved

FIGURE 10. Summary of the mechanisms by which apoptotic cells or mAb217 induce TGF-β production.

Apoptotic cells or mAb217 up-regulate TGF-β mRNA expression through p38 MAPK, ERK and JNK, and enhance protein translation through and enhance protein translation through RhoA/PI3K/Akt/mTOR/eIF4E signaling pathway. PSRS – Phosphatidylserine recognition structures.

Discussion

The ability of apoptotic cells to signal for their quiet, non-inflammatory and non-immunogenic removal in vivo is critical for normal tissue homeostasis and for resolution of inflammation. Just as there is considerable redundancy in the recognition and removal receptors and pathways for apoptotic cells, it seems likely that the anti-inflammatory effects are also mediated by various separate mechanisms. One of these is the selective stimulation by apoptotic cells of the anti-inflammatory and anti-immunogenic mediator, TGFβ. Thus, in previous studies (8, 22) we have shown that the ability of PS-exposing apoptotic cells or PS liposomes to block ongoing pulmonary inflammation or adaptive immunity was significantly dependant on the production of active TGFβ.

To explore the signaling mechanisms by which apoptotic cells induce TGF-β production it was first necessary to rule out the autocrine and paracrine stimulating effects of TGF-β itself. Accordingly, we set up a cell culture system using cells stably transfected with the dominant negative, truncated TGF-β receptor II. These were shown not to respond to added TGF-β and, as expected, stimulation with apoptotic cells showed lower overall levels of TGF-β produced. This observation supports the presence of autocrine/paracrine effects of TGF-β in enhancing the amounts of the molecule normally generated by apoptotic cells. The monoclonal antibody 217, which was raised against PS recognizing macrophages, binds to, and activates cells in a fashion that mimics exactly the effects of apoptotic cells in terms of efferocytosis and the generation of anti-inflammatory mediators (7, 22, 38). Unfortunately, however, attempts by us and a number of other laboratories have so far been unsuccessful in clearly identifying the antigen(s) with which it interacts. It was used herein as an adjunct to stimulation with apoptotic cells to initiate TGF-β production without the complexities of adding whole cells into the system. In all cases the stimulating effects of mAb217 and intact apoptotic cells were qualitatively identical, although as expected, not always in quantity or exact time course.

We demonstrated that apoptotic cells or mAb 217 increased total cellular levels of TGF-β mRNA without affecting its stability. Furthermore, the upregulated TGF-β transcription was inhibited by SB 203580, PD 98059 and JNK inhibitor II suggesting the involvement of all three MAP kinases (p38 MAPK, ERK and JNK). Notably, each individual inhibitor almost completely suppressed protein or message expression. Presumably, TGF-β mRNA level is not proportionally related to TGF-β protein, and vice versa (54). TGF-β contains predicted AP-1, Egr-1, and SP-1 binding sites in its promoter regions (41, 55–58). The AP-1-type proteins mediating this induction may act at one or more AP-1 sites in the two promoters of TGF-β gene and three putative, overlapping AP-1 binding sites about 200 base pairs downstream of TGF-β (52, 59, 60). Consistent with these, we have observed that blockade of mAb 217-induced ERK and JNK phosphorylation inhibited phospho-c-jun, although, surprisingly, SB 203580 increased mAb 217-induced phospho-c-jun (data not shown). The role of p38 MAPK in regulation of gene expression was not well understood, although transcriptional factors such as activating transcription factor-2, SRF accessory protein-1 and CCAAT-enhancer binding protein-β are known as substrates of p38 MAPK (61). It is important to note that Otsuka et al (62, 63) have earlier reported that PS-liposomes induce TGFβ production with a requirement for ERK and PI3Kinase. However, they did not address possible selective effects on transcription and translation.

The Rho family of small GTPases, including Rho, Rac and Cdc42, are involved in regulation of a variety of cellular functions such as cell migration and adhesion, proliferation, vesicle trafficking, bacterial ingestion and inflammation (64–67). Interaction of apoptotic cells initiates a delayed activation of Rho in responding cells and this is known to negatively regulate subsequent efferocytosis ((4, 46, 68, 69) and Gardai et al unpublished). Accordingly we then investigated the possible involvement of Rho in apoptotic cell-induced TGF-β production. Inactivation of RhoA by C3 transferase inhibited apoptotic cell or mAb 217-induced TGF-β protein production, but did not affected TGF-β mRNA expression. These findings suggested that posttranscriptional regulation of TGF-β generation is another important control point in regulating the activity of this anti-inflammatory mediator. Further exploration of the downstream effectors leading to TGF-β translation revealed that Rho inhibition resulted in lower levels of Akt and eIF4E phosphorylation.

The mammalian target of rapamycin (mTOR) is a central regulator of translation and cell proliferation (70–72). Two major substrates for mTOR are the serine/threonine kinase p70^S6K and the 4E-binding protein 4EBP-1. Phosphorylation of 4EBP-1 by mTOR results in release of the cap-binding protein translation initiation factor, eukaryotic initiation factor (eIF4E) (47, 73), which is inactive when bound to hypo-phosphorylated 4EBP-1. Moreover, eIF4E activity is also regulated by phosphorylation and enhances translation rates of cap-containing mRNAs (74), which include TGF-β (54, 72, 75–77). The upstream regulator of mTOR in this circumstance appears to be PI3-kinase/Akt (48, 72). PI3-kinase has been reported to be upstream of Rac and Rho (78, 79). On the other hand, RhoA also has been shown to prevent myoblast death by inducing the PI3-K/Akt pathway (80). In the present study, we suggest that PI3-kinase is a downstream signal mediator from Rho, which then leads to apoptotic cell-induced TGF-β translation through Akt/mTOR/eIF4E. TGF-β itself is known to activate Rho (81–84), however, this was avoided by use of the 3T3TβRII cells or RAWTβRII cells (Fig. 9). Positive or negative involvement of RhoA in TGF-β production has been reported (85, 86). Our findings suggest that the mechanisms leading to TGF-β regulation might be cell type and stimulus dependent.

MAP kinases (p38 MAPK, ERK and JNK) have been shown to regulate cytokine production at both transcription and translation (17, 87–92). Intriguingly, when we examined a potential role for p38 MAPK, ERK or JNK in induction of TGF-β to non-apoptotic cell stimuli PMA or LPS, we did indeed find evidence of their involvement in its translational regulation via eIF4E (51, 53). However, apoptotic cell-induced TGF-β translation appears to be regulated independent of p38 MAPK, ERK and JNK. Thus, in the present study, we compared apoptotic cells or mAb 217 to PMA and LPS for stimulation of TGF-β production. Although all three stimuli activated p38 MAPK, ERK and JNK, the outcome of the same kinase activation was entirely different (Table 1).

These experiments have begun to address the signal pathways involved in enhanced transcription and translation of TGFβ in response to apoptotic cells. Other points of regulation likely include secretion and extracellular activation of the molecule. Of note, an earlier study in vivo (8) suggested that the administration of apoptotic cells to an ongoing inflammatory site induced an immediate release of TGFβ that was not dependent on protein synthesis, suggesting that indeed apoptotic cell recognition might enhance TGFβ liberation from the cell. There are many potential mechanisms by which latent TGFβ may be activated, but at this point, a possible role for the interaction of the apoptotic cell in this process has not been addressed. These two steps in the process of achieving functional TGFβ are areas for important future research. It will also be critical to sort out the contribution of the various PS-recognizing receptors and bridge molecules involved in its production, release and activation, just as it has now become equally important to sort out their various contributions to the uptake and clearance process.

In summary, these studies show that apoptotic cells up-regulate TGF-β mRNA expression through p38 MAPK, ERK and JNK, and enhance protein translation through a newly defined PSRS/RhoA/PI-3K/Akt/mTOR/eiF4E signaling pathway (Fig. 10). The present findings demonstrate that the mechanisms leading to apoptotic cell-induced TGF-β production can be distinguished from other stimuli.

Abbreviations

TGF-β

transforming growth factor-β

PS

phosphatidylserine

PSRS

phosphatidylserine recognition structures

LPS

lipopolysaccharide

MAPK

mitogen-activated protein kinase

ERK

extracellular signal-regulated kinase

JNK

c-jun N-terminal kinase

MEK

MAPK/ERK kinase

mTOR

mammalian target of rapamycin

4EBP-1

4E-binding protein-1

eIF4E

eukaryotic initiation factor 4E