Published online before print February 5, 2008
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
expression in human monocyte-derived macrophages
* Division of Paediatric Surgery, Department of Surgery, Li Ka Shing Faculty of Medicine, The University of Hong Kong, China; and
Easter Bush Veterinary Centre, Royal (Dick) School of Veterinary Studies, University of Edinburgh, Roslin, Midlothian, United Kingdom
1Correspondence: Division of Paediatric Surgery, Department of Surgery, The University of Hong Kong, K15, Queen Mary Hospital, Pokfulam, Hong Kong SAR, P.R. China. E-mail: paultam{at}hkucc.hku.hk
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
, a key proinflammatory cytokine, by LPS-stimulated human monocyte-derived macrophages. ALK5 protein was only detectable after LPS stimulation, and the failure of treatment with SB431542 to alter TNF- mRNA expression indicates that regulation is post-transcriptional. The additive effect of blocking TGF-β and p38 MAPK signaling on reducing TNF- but not IL-6 production suggests that there is selectivity in pathway signaling. SB431542 had similar inhibitory effects on TNF- production by human monocytes and endothelial cells as well as macrophages. Furthermore, treatment with SB431542 reduced plasma TNF- levels and tissue damage and thereby, prevented the lethal effects of LPS in a mouse model of septic shock. Our data demonstrate a direct effect of TGF-β signaling via ALK5 on the regulation of TNF- synthesis.
Key Words: ALK5 • inflammation • septic shock
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
, IL-6, and IL-1. Of these cytokines, TNF- is considered to be the most important factor, as it is expressed at high levels during the early phase of septic shock [1
, 2
], and injecting TNF- can reproduce septic shock syndrome in humans [3
]. The up-regulation of TNF- further promotes synthesis of other cytokines and chemokines and facilitates the migration and the accumulation of inflammatory cells into tissues. This increases the release of the cytotoxic reactive oxygen species causing tissue damage. Monocytes, macrophages, neutrophils, and endothelial cells are the main cell types producing TNF- [4
5
6
]. The biosyntheses of TNF- has been well documented. In brief, LPS binds to LPS-binding protein and forms a complex with TLR4, and then, signal transduction via MyD88 and other intracellular components activates NF-
B, leading to the further induction of TNF- mRNA. Post-transcriptional regulation is also a key step in the control of TNF- expression, whereby LPS stimulation activates p38 MAPK and by phosphorylation, controls TNF- mRNA stability and protein production. In vitro experiments have demonstrated that blockade of p38 MAPK with the inhibitor SB203580 reduces TNF- production. However, in vivo, in an experimental model of septic shock, treatment with this compound failed to prevent TNF- release and reduce mortality [7
, 8
]. The failure of the SB203580 treatment may relate to the complexity of the processes underlying septic shock as well as the importance of p38 MAPK in vivo. This observation prompted us to look for additional downstream signaling cascades that might contribute to the regulation of TNF- in the search for more effective therapeutic targets through which to modulate inflammation.
TGF-β is a pleiotropic cytokine that has inflammatory and anti-inflammatory activities depending on the cellular environment. The TGF-β signal transduction is initiated by the binding of the ligand to TGF-β receptor II (TβRII), which forms a complex with the TβRI, initiating phosphorylation of SMAD2 and formation of the receptor-regulated SMAD complex [9
]. This complex then enters the nucleus and regulates transcription. Each of the TβR subunits has several isoforms, and for TβRI expression in endothelial cells, activin receptor-like kinase 1 (ALK1) and ALK5 have been studied in detail [10
]. It has been suggested that the expression of ALK1 is dependent on ALK5 [11
]. The immunosuppressive effects of TGF-β are well documented [12
, 13
]. However, regarding the effects of TGF-β signaling on proinflammatory responses mediated by cells of the innate immune system, information is limited, and most studies have focused on the indirect effects in the context of the recruitment of inflammatory cells to the site of injury (reviewed in ref. [14
]). However, Wiseman and colleagues reported that TNF- mRNA expression was elevated in TGF-β-treated monocytes, but whether this action was a result of the direct effect of TGF-β signaling was not investigated.
In this study, we have investigated the effect of inhibiting ALK5-mediated signaling on the regulation of TNF- expression. Blockade was achieved using the inhibitor SB431542, which has proved successful in blocking TGF-β-induced cancer cell proliferation in anti-tumor therapy [14
]. In addition to in vitro studies, we demonstrate in an experimental model of septic shock in mice that SB431542 in vivo can lower plasma levels of TNF- and prevent the lethal effects of LPS. Collectively, our findings suggest that the expression and function of ALK5 are essential in the regulation of TNF- production and therefore, may potentially be applied to the management of TNF--mediated diseases.
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
LPS injection and mouse model of septic shock
The mice (25–30 g) were injected i.p. with saline or LPS (3.5 mg/kg), dissolved in saline. The survival and mortality of mice were monitored for 8 days. SB431542 (10 mg/kg) or vehicle (DMSO) was i.v.-injected 2 h before LPS administration. Tissue collected from the lung, spleen, liver, and kidney of mice killed 18 h after LPS treatment was fixed in 10% formalin and embedded in paraffin. The plasma was collected 1.5 h after LPS injection.
Histology and immunohistochemistry analysis
Sections were stained with H&E for histology analysis. For immunohistochemical analysis, sections were quenched endogenous peroxidase activity with 3% H2O2. Antigen retrieval was performed with citrate buffer. For F4/80 staining, the sections were digested with Proteinase K. Sections were incubated with primary antibodies [rabbit polyclonal antipolymorphonuclear neutrophils (anti-PMN), 1:3000; rabbit anti-iNOS, 1:100; rat anti-F4/80, 1:50; goat anti-MCP-1, 1:100; and goat anti-RANTES, 1:300] for 16 h at 4°C, and the Dako EnVisionTM system (or StreptABC in the case of goat antibodies) with peroxidase (3'-diaminobenzidene tetrahydrochloride; Dako Corp., Denmark) was used for a reveal-positive signal.
Human PBMC isolation and macrophage cultures
PBMCs were isolated from healthy donors by Ficoll-paqueTM Plus (Amersham Bioscience, Piscataway, NJ, USA) gradient centrifugation. To obtain a PBMC-derived macrophage, PBMCs were cultured in RPMI 1640, supplemented with 10% human AB serum (Sigma Chemical Co.), 1% penicillin/streptomycin (Gibco, Grand Island, NY, USA), and 4 ng/ml GM-CSF (Peprotech, Rocky Hill, NJ, USA) for 7 days until they maturated into macrophages and were used for experiments between Days 7 and 10. For the experiments in PBMCs, the cells were cultured in RPMI 1640, supplemented with 10% human AB serum and 1% penicillin/streptomycin for 2 days.
RT-PCR and quantitative PCR (qPCR)
Total RNA was purified using RNeasy kits with DNase I digestion (Qiagen GmBH, Germany), according to the manufacturer’s instructions. RNA (1 µg) was reversed-transcribed, and PCR was performed using Taq polymerase (Invitrogen Corp., Carlsbad, CA, USA) with selected primers (Table 1
). qRT-PCR was performed using the kit by Applied Biosystems (Foster City, CA, USA) on an ABI Prism 7700 sequence detector, and the fold changes were calculated according to the manufacturer’s instructions.
|
View this table: [in a new window] |
Table 1. Primers Used in Experiments
|
and IL-6 content in triplicate using ELISA kit (R&D Systems, Minneapolis, MN, USA).
Western blot analysis
Cells were lysed in radioimmunoprecipitation assay buffer with proteinase inhibitors (Roche Applied Science, Nutley, NJ, USA). Protein (20 µg) was separated by electrophoresis and transferred onto a polyvinylidene difluoride membrane (Amersham Pharmacia Biotech Inc., Little Chalfont, UK). The membranes were incubated at 4°C with anti-pSMAD2, ALK5, and actin antibodies (1:1000) and then incubated with HRP-conjugated antibody (1:1000). The signals were revealed by an ECL kit (Amersham Biosciences).
Human protein cytokine arrays
The cultured supernatants were applied to a human cytokine antibody array 3.1 (H0109809, RayBiotech, Norcross, GA, USA). The array membranes were processed according to the manufacturer’s instructions. The results with immunoreactivity were assessed and quantified by using Scion image software (Scion Corp., Frederick, MD, USA).
Statistical analysis
Survival data were performed by the Kaplan-Meier method and the log-rank test using SPSS software (SPSS Inc., Chicago, IL, USA). Immunopositive cell counting was analyzed and expressed as mean ± SEM. The Student’s t-test or the one-way ANOVA followed by the appropriate post-hoc test was used when appropriate. Statistical significance was set at P< 0.05.
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
expression in human PBMC-derived macrophages. To establish whether TGF-β signaling is directly involved in TNF- synthesis, first, ALK5 expression was investigated in macrophages activated with LPS and IFN-
. Under resting conditions, ALK5 could not be detected. However, ALK5 is expressed in response to activation with LPS and IFN- stimulated. Similarly, these proinflammatory stimuli increased levels of the TGF-β intracellular signaling molecule pSMAD2 (Fig. 1A
and 1B
).
 View larger version (51K): [in a new window] |
Figure 1. ALK5 expression in activated macrophages. The expression of TNF- (A) in matured human PBMC-derived macrophages was stimulated with LPS (1 µg/ml) and IFN- (50 ng/ml) for 18 h. The cell lysates were used for protein detection; n=7 (B). UT, untreated.
|
was reduced by ALK5 blockade expression. The addition of SB431542 (10 µM) reduces TNF- production by 61.17 ± 23% and 58.36 ± 38%, respectively, for LPS and LPS + IFN--stimulated macrophages (P<0.05; n=3 individual blood donors; a representative was shown in Fig. 2 A
). The blockade of TGF-β signaling is confirmed by reduction in pSMAD2 (Fig. 2B)
. The SB431542 treatment inhibited TNF- expression in a dose-dependent manner (Fig. 2D)
. IL-6 was also induced by LPS-stimulated macrophages, but its expression was not affected by SB431542 treatment (Fig. 2C)
.
 View larger version (24K): [in a new window] |
Figure 2. Effects of SB431542 on TNF-, pSMAD2, and IL-6 expression in a dose-dependent manner. Cells were pretreated for 2 h, with or without SB431542, and then incubated with LPS and LPS + IFN- for a further 18 h. The cell lysates and supernatant were collected and stored at –20°C. In cell culture, SB431542 was dissolved in DMSO and used at 10 µM or as indicated; the DMSO concentration was never higher than 0.05%. The doses of LPS and IFN- were used as in A. Supernatants were collected for TNF- (A; dotted bars, without SB431542; and hatched bars, with SB431542). (B) Measurement and ALK5/pSMAD2 detection in cell lysates (n=3). (C) IL-6 (dotted bars, without SB431542; and hatched bars, with SB431542). (D) Effects of different doses are presented.
|
expression mRNA expression and stability by TGF-β, mRNA levels were evaluated after LPS/IFN- stimulation, with and without addition of the ALK5 inhibitor by qRT-PCR. We observed that TNF- mRNA was markedly increased after LPS/IFN- stimulation. However, there was no alternation in transcripts for TNF- in response to LPS/IFN- treatment, including post-treatment with SB431542 (at 10 µM or 25 µM), which implies that the control of TNF- expression is at the post-transcriptional level (Fig. 3A
and 3B
). Furthermore, mRNA levels for ALK5 and TGF-β were also unaltered in the treatment groups as compared with naïve macrophages (Fig. 3A)
.
 View larger version (25K): [in a new window] |
Figure 3. Expression of TNF-, TGF-β1, and ALK5 genes. Human PBMC-derived macrophages pretreated with 10 µM or 25 µM SB431542 for 4 h and the mRNA levels were measured by using semi-qPCR for TNF-, TGF-β, ALK5, and β-actin (A) and qPCR for TNF- (B). The p38 MAPK inhibitor SB203580 and SB431542 (25 µM each) were added to cell cultures alone or together 2 h before activation with LPS and IFN-. Secreted TNF- (C) and IL-6 (D) were measured (n=3).
|
expression is well documented [16
]. As ALK5 inhibitor SB431542 has only a minimal effect on p38 MAPK activity [15
], we have used the combination of this compound and SB203580, a p38 MAPK inhibitor, to examine the functional relationship between p38 MAPK and ALK5 signaling in the regulation of TNF- expression. The addition of SB203580 inhibited LPS/IFN--induced production of TNF- by macrophages (Fig. 3C)
. When the p38 MAPK and ALK5 pathways were inhibited by SB203580 and SB431542, the expression of TNF- was further reduced, suggesting an additive effect. However, regarding IL-6 production, when the inhibitors were added individually, only SB203580 reduced TNF- expression, and this was not enhanced by the addition of SB431542 (Fig. 3D)
.
ALK5 regulates TNF- expression in human monocytes and endothelial cells
Similar to macrophages after exposure to LPS, monocytes also produce TNF-, which is markedly but not completely inhibited by the addition of SB431542 at 10 µM, and the addition of a higher concentration (25 µM) of the compound cannot totally inhibit pSMAD2 activity (Fig. 4 A
). This difference may, in part, be a result of the constitutive expression of ALK5 in monocytes (Fig. 4B)
. As ALK5 is constitutively expressed on monocytes but not macrophages, we investigated the kinetics of expression of TNF- in response to LPS, which was more rapid in monocytes (Fig. 4C)
. The production of TNF- reached a plateau at 4 h in monocytes, whereas in macrophages, this was not achieved until 8 h after activation. We noted that TNF- production in endothelial cells, which have constitutive expression of ALK5, had similar kinetics to monocytes and that LPS from different bacteria behave in the same way (data not shown), suggesting the TGF-β signaling plays a general role in the LPS-induced TNF- expression.
 View larger version (20K): [in a new window] |
Figure 4. Human monocyte expression ALK5. Human monocytes isolated from PBMC pretreated with SB431542 were cultured with LPS + IFN-. Supernatants were collected for measurement of TNF- production (A) and the cell lysates for TGF-β and pSMAD2 detection (n=3; B). Production of TNF- was compared between monocytes and macrophages after LPS and IFN- stimulation (n=2; C).
|
, a multicytokine array was used. TNF- production after SB431542 treatment was reduced by 67% and 51% in macrophages and monocytes, respectively, which had been activated with LPS/IFN- (Fig. 5
). Of the cytokines investigated, TNF- expression was inhibited the most. MCP-1 and RANTES, which have been reported to be induced by TNF- [17
, 18
], were also reduced in macrophages but less so in monocytes. However, their reduced levels could be a result of lowered TNF- production or a direct effect of SB431542 on these genes. Regarding increased cytokine production, we noted that in activated macrophages, SB431542 MDC and IL-7 were up-regulated the most and for monocytes, IL-10 and oncostatin M. Furthermore, oncostatin M can inhibit LPS-induced TNF-, and so, its induction by SB431542 would also function to suppress proinflammatory responses.
 View larger version (50K): [in a new window] |
Figure 5. Multicytokine array analysis. Supernatants of activated macrophages or monocytes were used to investigate the effect of SB431542 on other inflammatory cytokines using human cytokine antibody array with 42 cytokines. The visual spots were quantified, and the table represents the relative changes compared with control (n=1 each; representative of macrophage was shown). GRO, Growth-related oncogene; MDC, macrophage-derived chemokine; ENA-78, epithelial-derived, neutrophil-activating factor 78.
|
level in mice septic shock model and IL-6. However, the mechanisms underlying these effects were not defined. Extending our studies about the in vitro effects of SB431542 on TNF- synthesis, we investigated if inhibition of TGF-β signaling was protective against septic shock. SB431542 (10 mg/kg), administered 2 h before LPS injection, resulted in a significant reduction in mortality in a mouse model of septic shock (Fig. 6 A
). The reduction of mortality corresponded with a decline of plasma level of TNF-, 30% compared with the vehicle control (Fig. 6B)
.
 View larger version (23K): [in a new window] |
Figure 6. Effect of SB431542 in a mouse model of septic shock. SB431542 or vehicle was used to pretreat animals for 2 h before LPS challenge. The animal survival was recorded (n=6 in each group; P<0.001; A). TNF- levels in the plasma were measured after 90 min and 18 h of LPS treatment by ELISA (thick, hatched lines, vehicle; thin, hatched lines, SB431542-pretreated; ***, P<0.001, compared between two groups; B).
|
 View larger version (90K): [in a new window] |
Figure 7. Analysis of the cellular infiltration after the administration of SB431542 to LPS-treated mice, which were treated with vehicle control (A, C, and E) or with SB431542 (B, D, and F) and were killed; the tissues were analyzed (n=3 in each group) after 18 h exposure to LPS. H&E (A and B) and immunohistochemical staining for PMNs (C and D) and macrophages (F4/80; E and F) were performed. The immunopositive cells were counted and compared in lung, liver, spleen, and kidney (cross-hatched bars, normal; thick, hatched bars, vehicle; thin, hatched bars, SB431542-treated; *, P<0.05; **, P<0.01; and ***, P<0.001, when comparing SB431542 pretreated with vehicle control).
|
expression, especially in the lung [22
, 23
], and the high expression of these molecules would contribute to the recruitment of inflammatory cells to the lungs. MCP-1-positive cells were localized mainly in terminal bronchioles, whereas RANTES-positive cells were detected in mononuclear cell infiltrate. Treatment with SB431542 markedly reduced the number of MCP-1- and RANTES-expressing cells (Fig. 8B
, 8D
, and 8F
).
 View larger version (151K): [in a new window] |
Figure 8. Immunohistochemistry analysis of iNOS and chemokine expression after treatment with SB431542. The expression of iNOS, MCP-1, and RANTES was determined in lung tissue sections in vehicle control (A, C, and E) and in the SB431542 treatment group (B, D, and F) exposed to LPS in vivo.
|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
is a key inflammatory mediator in septic shock, and it has been demonstrated that the initial high levels of TNF- induced contribute directly to the severity of septic shock [24
]. In agreement is the observation that in a mouse model of septic shock, the administration of a high dose of TNF- leads to the development of disease [25
]. Inhibition of p38 MAPK decreases TNF- expression in vitro and reduces mortality in animal models of septic shock. Increased survival is accompanied by diminished cellular infiltration and lower serum TNF- levels [26
, 27
], although this is controversial [8
]. In humans, reducing TNF- had also been achieved [28
] by the administration of a p38 MAPK inhibitor. However, the clinical application of the inhibitor in the treatment of the septic shock has not been reported.
In this study, the expression of ALK5 in macrophages induced by LPS stimulation suggested that TGF-β signaling was involved in the innate immune system in response to pattern receptor recognition, although whether the effect is a benefit or deleterious in providing protection against microbes is not known. The use of an ALK5 inhibitor (SB431542) provided evidence that activation of the TGF-β signaling pathway is proinflammatory through its capacity to increase TNF- expression. We observed that secreted TGF-β1, detectable in culture supernatants of unstimulated macrophages (
200 pg/ml), was decreased after LPS stimulation (data not shown). This finding, together with the observation that there is constitutive expression of TβRII (data not shown) in naive macrophages, implies that the induction of ALK5 with LPS/IFN- stimulation and the subsequent formation of the TβR complex after the binding of TGF-β1 is a unique step that results in the initiation of signal transduction leading to the post-transcriptional regulation of TNF- mRNA in macrophages. Furthermore, we observed using Actinomycin D that treatment with SB431542 had no direct effect on the degradation of mRNA for TNF- (data not shown).
The ability of TGF-β to mediate proinflammatory and anti-inflammatory activity has been described previously [29
]. In studies using mice in which TGF-β1 is overexpressed [30
] or deficient [31
, 32
], as well as SMAD3 knockout mice [33
], immune responses to LPS were augmented, as determined by cytokine expression or cellular infiltration. These outcomes may reflect different stages of inflammation. In the initial stages, TGF-β may recruit monocytes, and their differentiation into macrophages and inhibiting TGF-β signaling at this stage may reduce or delay the inflammatory response. In contrast, TGF-β, by reducing IFN- expression [34
] and enhancing IL-10 production [35
], would limit the development of inflammation, and consequently, inhibition or deletion of TGF-β would enhance immune responses. In these circumstances, total depletion of TGF-β, which occurs in the TGF-β and Smad3 knockout mice, would enhance inflammation. Here, in LPS-induced sepsis, we observed that enhanced survival increase was accompanied by a reduction in inflammation following SB203580 treatment during the early stages of septic shock, although the inhibitor may not completely inhibit the SMAD2 activation. However, at a later stage of inflammation, TNF- levels are comparable between control animals and those receiving the inhibitor.
Post-transcriptional events are central in controlling TNF- expression by regulating translational initiation, mRNA stability, and polyadenylation. Adenosine-uridine (AU)-rich elements (ARE) present in the 3'-untranslated region of TNF- regulate the expression of TNF- by binding transacting factors such as the zinc-finger protein tristetraproin and RNA-binding proteins Hel-N1, HuR, AU factor 1, T cell-restricted intracellular antigen 1 (TIA-1), and TIA-1-related protein (TIAR) [36
]. Regarding the control of TNF- synthesis, the contribution of p38 MAPK has been studied in the most detail. It regulates mRNA production and stability and is itself controlled through the activation of the downstream gene MAPK-activated protein kinase 2. Many ARE-binding proteins are regulated by p38 MAPK and are involved in mRNA stability, with the possible exception of TIA-1 and TIAR. Whether these molecules relate to TGF-β signaling is under investigation.
It has been suggested that monocytes are the most important cell type in septic shock [37
]; their constitutive expression of ALK5 and responsiveness of LPS highlights the importance of antimicrobial responses in septic shock. However, the production of large amounts of TNF- induced by LPS in monocytes may cause tissue damage through enhanced expression of targets of TNF-, such as RANTES [22
] and MCP-1 [23
], and by neutrophil recruitment. Here, we report that SB431542 reduces the number of RANTES and MCP-1-secreting cells induced by LPS administration in vivo. Our results confirm the potential importance of monocytes in septic shock, and that by understanding TNF- regulation, more effective options for treating septic shock may be revealed.
The inhibition of TNF- expression and reduction in mortality following septic shock in animals by the ALK5 inhibitor SB431542 suggest that TGF-β signaling could have a beneficial role in the development of septic shock. The protective effects could be mediated through several mechanisms, including reduced synthesis of TNF-, directly or through the regulation of other mediators such as RANTES and MCP-1, which would prevent neutrophil and macrophage infiltration; and the down-regulation of TGF-β-related genes, for example, plasminogen activator inhibitor-1, which is suggested to be downstream of ALK5 in endothelial cells [10
] and is involved in diffuse intravascular coagulation that occurs in septic shock [38
]. The association of ALK5 with the production of TNF- is not only applicable to inflammatory diseases such as septic shock but also may help to explain the limited effect of SB431542 in cancer since TNF-, which has anti-cancer activity.
Received August 7, 2007; revised November 7, 2007; accepted December 4, 2007.
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
in cultured human endothelial cells stimulated with lipopolysaccharide or interleukin-1 Arterioscler. Thromb. Vasc. Biol. 20,410-415 release in patients with severe peritonitis: role of tumor necrosis factor- converting enzyme cleavage Crit. Care Med. 33,1359-1364[CrossRef][Medline]-supported proliferation of leukemia and lymphoma cell lines J. Biol. Chem. 275,21086-21093-induced expression of monocyte-chemoattractant protein-1 in endothelial cells Blood 93,857-865 in endotoxemic TGF-β1 transgenic mice J. Leukoc. Biol. 63,31-39[Abstract] production Ann. Rheum. Dis. 59(Suppl. 1),i3-i5
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
