Direct regulation of IGF-binding protein 1 promoter by interleukin-1β via an insulin- and FoxO-1-independent mechanism

Abstract

The level of insulin-like growth factor-binding protein 1 (IGFBP1), a liver-produced serum protein that regulates insulin-like growth factor-I bioactivity, glucose homeostasis, and tissue regeneration, increases during inflammation. This manuscript describes a novel pathway for the regulation of hepatic IGFBP1 mRNA and protein levels by interleukin (IL)-1β. Experiments with the luciferase reporter system show that IL-1β stimulates transcriptional activity from the 1-kb promoter region of IGFBP1. Although IL-1β stimulation suppresses the insulin activation of protein kinase B, the major upstream regulator of IGFBP1 mRNA transcription, the induction of IGFBP1 by IL-1β did not require an intact insulin response element. Furthermore, neither overexpression nor silencing of FoxO-1 had any effect on the IL-1β-induced increase in IGFBP1 mRNA levels and promoter activity. However, inhibition of the ERK MAP kinases effectively prevented the IL-1β effects. Inhibition of neutral sphingomyelinase, a key player in the IL-1β signaling cascade that acts upstream of ERK, also suppressed the IL-1β effects, while increasing the ceramide, through the addition of C2-ceramide or via treatment with exogenous sphingomyelinase, was sufficient to induce IGFBP1 promoter-driven luciferase activity. Studies in primary rat hepatocytes where the levels of neutral sphingomyelinase were either elevated or suppressed using adenoviral constructs affirmed the key role of neutral sphingomyelinase and ceramide (exerted likely through ERK activation) in the IL-1β-induced IGFBP1 production. Finally, the IL-1β effects on IGFBP1 mRNA production and protein secretion could be abolished by the addition of insulin, either at very late time points or at very high doses.

insulin-like growth factor-binding proteins (IGFBPs) are a family of six serum proteins that bind to insulin-like growth factor-I (IGF-I) and regulate its turnover, transport, and tissue availability (34). Only free nonbound IGF-I is biologically active; therefore, the availability of IGFBPs modulates many of the physiological functions of IGF-I involving cellular growth, survival, proliferation, differentiation, cell motility, and glucose uptake (12). IGFBP1 is particularly important in regulating IGF-I bioactivity. It is the only IGFBP with a tightly regulated expression (21), and a strong inverse correlation exists between the levels of circulating IGFBP1 and those of free IGF-I (34, 51). High baseline IGFBP1 is found to be a reliable predictor for congestive heart failure (25) and a biomarker for early stage alcohol-induced liver disease (37). High levels of IGFBP1 have also been associated with increased risk for cardiovascular mortality and morbidity (20, 56). IGFBP1 may also have IGF-independent functions and can effectively bind the fibronectin receptor and thus affect cell migration and attachment (34). Intracellular roles of IGFBP1 in mitochondria and the nucleus have also been described, implicating it in the regulation of hepatic apoptotic response (36).

IGFBP1 is highly expressed in the liver and endometrium, with some expression also found in the kidney. Liver is the main source of systemic IGFBP1, although, after implantation, IGFBP1 becomes more highly expressed in the endometrium than in the liver. IGFBP1 has a very distinct system of regulation, with insulin playing a key role. During fasting, when the insulin levels are low, IGFBP1 expression is high, and vice versa, the IGFBP1 levels decrease in a fed state, when systemic insulin levels are high. Direct administration of insulin decreases plasma levels of IGFBP1 in humans and rats (7, 9, 22, 53, 54), whereas insulin deficiency induces hepatic IGFBP1 synthesis (46). Insulin suppresses directly the IGFBP1 promoter through protein kinase B (Akt) and Forkhead box O-1 (FoxO-1) transcription factor. In the absence of insulin, FoxO-1 is bound to the insulin response element (IRE) in the IGFBP1 promoter and interacts with coactivators, like p300/CBP and C/EBPβ, to drive IGFBP1 transcription. Insulin, however, activates Akt, which is then translocated to the nucleus (1) where it phosphorylates FoxO-1 (4, 8, 23, 29, 48, 55), causing its nuclear exclusion and, consequently, the suppression of IGFBP1 expression (19).

Inflammation, aging, sepsis, oxidative stress, and other cachectic conditions all increase IGFBP1 expression (2, 14, 30, 32, 49). The specific mechanisms for these effects, however, are not well understood and seemingly involve different systemic mediators and signaling pathways. In vitro, several direct inducers of IGFBP1 mRNA transcription have been identified. These include glucocorticoids, cAMP, thyroid hormones, and phorbol esters. Tumor necrosis factor-α (TNF-α), interleukin (IL)-1α, IL-1β, and IL-6 also seem to induce IGFBP1 mRNA production in vitro in Hep G2 cells, indicating possibly direct effects on mRNA production. However, with the possible exception of IL-6, which has been reported to use a signal transducer and activator of transcription (STAT)-3-mediated pathway to upregulate IGFBP1 mRNA during hepatic injury through the hepatocyte nuclear factor 1 (HNF-1) transcription factor, the signaling pathways of proinflammatory cytokines leading to the stimulation of IGFBP1 production by the liver have not been explored in detail, nor have their possible interrelations with insulin (35). Proinflammatory cytokines, which seem to mediate the increase in IGFBP1 production in many of the aforementioned pathological conditions, are thought to do so by perhaps altering the regulatory actions of insulin. Indeed both IL-1β and TNF-α have been shown to suppress the insulin-signaling pathway at various steps (34).

IL-1β, in particular, is the primary inducer of IGFBP1 biosynthesis during inflammation. IL-1β administration in vivo results in increased levels of circulating IGFBP1 (15), whereas infusion of IL-1β receptor antagonist attenuates sepsis-induced increase in IGFBP1, decrease in IGF-I, and the loss of muscle mass (30). In Hep G2 cells, IL-1β stimulates IGFBP1 mRNA production in a mitogen-activated protein kinase (MAPK)-dependent manner (16, 31). It is not known, however, whether IL-1β influences IGFBP1 production directly or by modulating insulin regulation, nor had the underlying pathways been explored.

IL-1β regulation of mammalian cell functions is achieved through the activation of a unique signaling pathway involving the formation of a receptor signaling complex consisting of IL-1R-associated kinase-1 (IRAK-1), IRAK-4, and TNF-associated factor-6 and the subsequent activation of IκB kinase and the MAPK pathway (40, 44, 57). The mechanism of MAPK activation by IL-1β also involves activation of neutral sphingomyelinase 2 (nSMase-2) and the generation of ceramide at the plasma membrane (18, 26). The IL-1β signaling cascade ends with the activation of several transcription factors, including nuclear factor (NF)-κB, activator protein (AP)-1, and C/EBP.

We recently identified FoxO-1 as a novel and direct downstream target of IL-1β that is activated in response to IL-1β treatment of hepatocytes (11). Having in mind the key role of FoxO-1 in regulating IGFBP1 mRNA transcription, this finding prompted us to elucidate the pathways linking IL-1β with IGFBP1 transcription, with special emphasis on possible interactions with the insulin signaling pathway. Our experiments show that the IL-1β induction of IGFBP1 mRNA levels is due to direct activation of IGFBP1 promoter and mRNA transcription. This process is FoxO-1 independent and involves neutral sphingomyelinase-2 and the MAPK ERK. Furthermore, we also show that high doses of insulin added at the end of IL-1β stimulation can negate the IL-1β effects.

MATERIALS AND METHODS

In vivo studies.

Young (3-mo-old) male C57BL/6 mice (National Institute of Aging, Bethesda, MD) were injected with lipopolysaccharide (LPS, 5.8 mg/kg body wt ip) or phosphate-buffered saline. Liver tissue was collected and immediately frozen in liquid nitrogen. All animals in the study were housed in an American Association for Accreditation of Laboratory Animal Care-approved facility and received humane care according to the Guide for the Care and Use of Laboratory Animals.

Cell cultures and treatments.

Hep G2 and H4-II-E cells (ATCC, Manassas, VA) were maintained in minimal essential medium (MEM) (Invitrogen, Waltham, MA), supplemented with 10% FBS (Atlanta-Biological, Atlanta, GA). HEK 293 cells overexpressing the IL-1β receptor type I (gift from Dr. X. Li, Cleveland Clinic, Cleveland, OH) were maintained in DMEM supplemented with 10% FBS. Typically, cells were seeded at a density of 250,000 cells/well in a 12-well plate. When confluency reached 70–80%, cells were transfected as indicated with 1 μg of purified DNA using Lipofectamine 3000 reagent (Life Technologies, Carlsbad, CA) following the manufacturer's protocol. The following constructs were used for transfection: a green fluorescent protein (GFP)-tagged human FoxO-1 [1,098 pcDNA GFP-FKHR, Addgene plasmid no. 9022, gift from William Sellers (42)], wild-type IGFBP1 promoter [Igfbp1 promoter/pGL3, Addgene plasmid no. 12146, a gift from Domenico Accili (41)], pRL-TK plasmid expressing the Renilla reporter (gift from Dr. Karyn Esser), siRNAi against FoxO-1 (siRNA ID no. s136655; Ambion, Life Technologies), or ΔIRE-IGFBP1 promoter. To create the ΔIRE-IGFBP1 promoter, the IRE (⁻¹²¹CAAAACAAACTTATTT⁻¹⁰⁵) in the plasmid expressing the wild-type promoter was removed by using inverse PCR mutagenesis with primers flanking the desired sites of deletion (primer sequence forward: 5′-AAA AAC CGC GGT GAA CAC GGG GAT CCT A-3′ and reverse: 5′-AAA AAC CGC GGC TTG TGA GCT CCG CAC-3′) and effectively replaced with CCGCGG (a SacII restriction site).

Primary hepatocytes were isolated from Fisher 344 rats (National Institute of Aging, Bethesda, MD) by in situ collagenase (Sigma, St. Louis, MO) perfusion, and the cells were plated (1.5 × 10⁶/dish; viability >80%) in 35-mm tissue culture dishes coated with 0.15 ml of 6.3 mg/ml Matrigel (BD Bioscience, Bedford, MA) in Waymouth's medium. For routine culturing, the medium was supplemented with insulin (0.15 mM) as the only hormone. Before treatment, insulin was withdrawn as indicated in the legends of the Figs. 1–9. Cultures were maintained at 37°C in 5% CO₂ atmosphere, with replacement of the medium every 48 h, commencing 3 h after plating. After treatments, the medium was aspirated, and the Matrigel was reliquidified by incubating for 30 min on ice with PBS containing 5 mM EDTA. The Matrigel was removed by centrifugation at 500 g for 4 min, and the cells were washed once again with PBS and used to prepare various cell extracts. Conditioned media was also collected, concentrated, and used for analyses.

Fig. 1.

Stimulation of hepatic insulin-like growth factor-binding protein 1 (IGFBP1) mRNA levels and protein secretion in response to lipopolysaccharide (LPS) and interleukin (IL)-1β. A: IGFBP1 mRNA expression as determined by real-time PCR in livers of C57B6 mice stimulated with LPS (5.6 mg/g body wt ip) or PBS and killed at the indicated times (n = 4 animals/group). B: IGFBP1 mRNA levels in primary rat hepatocytes cultured under standard conditions (Waymouth's medium containing insulin as the only growth factor) and stimulated with the indicated concentrations of IL-1β for 24 h, based on real-time PCR analysis (n = 3). The abundance of glyceraldehyde-3-phosphate dehydrogenase (GAPDH) was used for normalization. The data shown represent %change with the value of the IGFBP1-to-GAPDH ratio at time 0 being 100%. Inset shows the corresponding IGFBP1 protein abundance in the medium, based on Western blotting. C: IGFBP1 mRNA levels in Hep G2 cells stimulated with IL-1β (10 ng/ml) for 24 h, based on classical RT-PCR analysis (n = 3). The cells were cultured under standard conditions, in the presence of serum. The abundance of β-actin was used for normalization. The data shown represent the IGFBP1-to-β-actin ratio. Insets show representative images of the PCR products on agarose gel (bottom) and corresponding IGFBP1 protein abundance in the medium, based on Western blotting (top). D: IGFBP1 mRNA levels in H4-II-E cells cultured under standard conditions (i.e., in the presence of serum) and stimulated with IL-1β (10 ng/ml) for the indicated times, based on real-time PCR analysis (n = 3). The abundance of β-actin was used for normalization. The data shown represent %change with the value of the IGFBP1-to-β-actin ratio at time 0 being 100%. Inset shows the corresponding IGFBP1 protein abundance in the medium, based on Western blotting. E: IGFBP1 protein abundance in the medium collected from H4-II-E cells cultured under standard conditions and stimulated with IL-1β as indicated for 24 h, based on Western blotting analysis (n = 3). Statistical significance is indicated (*P < 0.05 and **P < 0.01).

Fig. 9.

Proposed mechanism for IL-1β-induced upregulation of IGFBP1. The mechanisms of IL-1β-induced upregulation of IGFBP1 mRNA transcription in the liver are complex and involve several distinct pathways. In a healthy state, the insulin signaling is the main mechanism for IGFBP1 regulation that facilitates the coordinated regulation of its production in accordance to the feeding state of the organisms. The pathway involves activation of Akt and nuclear export of FoxO-1 transcription factor. IL-1β, however, can negate insulin signaling and can potentially interfere with insulin-dependent regulation of IGFBP1 by downregulation of Akt. IL-1β can also increase IGFBP1 transcription by elevating nuclear FoxO-1 trough JNK activation in an insulin-independent manner. The results from this study delineate a novel pathway allowing for increasing IGFBP1 mRNA and serum IGFBP1 levels in response to IL-1β through direct stimulation of IGFBP1 promoter activity in a manner that does not require FoxO-1 or the insulin response element. This novel pathway seemingly involves activation of nSMase-2 and ERK and can be surpassed by an active insulin pathway.

Cells were treated with rat or human recombinant IL-1β (Life Technologies, Grand Island, NY) and recombinant insulin (Sigma). The inhibitors of JNK (SP-600125; Sigma), and ERK (PD-98059; Cell Signaling Technology, Danvers, MA), were added from stock solutions with DMSO. The inhibitor of neutral sphingomyelinase (GW-4869; Cayman Chemical, Ann Arbor, MI) was delivered according to Luberto et al. (39).

Infection and hepatocytes and adenoviral constructs.

For the purpose of overexpressing nSMase-2 in hepatocytes, recombinant adenovirus expressing the mouse nSMase-2 under a doxycyclin-inducible promoter (Ad-nSMase-2) was used (26). Routinely, hepatocytes were infected 48 h after isolation, and the expression of the transgene was induced by doxycycline (Clontech, Palo Alto, CA) on the day of infection. The expression of the FLAG-tagged nSMase-2 was judged by Western blotting using anti-FLAG antibody and antibody raised against the mouse and rat recombinant truncated nSMase-2. For silencing of nSMase-2 in hepatocytes, the adenovirus-based RNAi silencing approach was used. Candidate double-stranded DNA oligos encoding a sense-loop-antisense sequence to the nSMase-2 mRNA (sh) were designed using BLOCK-iT RNAi Designer (Invitrogen, Carlsbad, CA) and generated by a short annealing procedure. DNA oligos encoding scrambled sense-loop-antisense RNA sequence (scr) and shRNA sequence against β-galactosidase were prepared in parallel. All sequences were then subclonned into pENTR/U6 entry vector (Invitrogen) creating an RNAi cassette that expresses the shRNA against nSMase-2, the scrambled control, and the positive control. After that, the RNAi cassette was excised and subcloned into pAdTrack vector that also encodes GFP (gift from Dr. George Smith, University of Kentucky). Homologous recombination between the pAdTrack shuttle vector containing the RNAi cassette and the adenoviral backbone plasmid (pAdEasy-1) was done in Escherichia coli strain BJ5183 to produce adenovirus expressing shRNA against nSMase-2 (Ad-sh) and adenovirus expressing corresponding scrambled sequence (Ad-scr).

Luciferase/Renilla reporter system for IGFBP1 promoter and ΔIRE-IGFBP1 mutant promoter.

Hep G2 cells (2.5 × 10⁵) were plated in 12-well plates in 1 ml of MEM and cotransfected with 300 ng of luciferase and 30 ng of Renilla reporter plasmids using Lipofectamine 3000 reagent (Life Technologies). pGl3 expressing the luciferase reporter gene driven by either the 1-kB IGFBP1 promoter or the 1-kB promoter with mutated IRE consensus-binding site (ΔIRE-IGFBP1 promoter) was used to assay the IGFBP1 promoter activity, whereas the pRL-TK Renilla reporter plasmid was used as an internal control. Posttransfection (24 h), cells were serum starved for 6 h and then treated with human recombinant IL-1β and/or insulin for a specified time. Posttreatment, the cells were washed with PBS and lysed with 100 μl of passive lysis buffer (Promega, Madison, WI). The luciferase activity was assayed using the Dual Luciferase Assay Reporter System (Promega) following the manufacturer's protocol and normalized for the activity of the Renilla reporter.

Western blotting.

The cell pellet from each dish was resuspended in 50 μl of lysis buffer consisting of 1 mM EDTA, 1% Triton X-100, 1 mM Na₂VO₄, 1 mM NaF, 10 μg/ml aprotinin, 10 μg/ml leupeptin, and 10 mM Tris·HCl, pH 7.4. The cells were then incubated on ice for 20 min and centrifuged at 16,000 g for 10 min at 4°C. The supernatant was used for SDS-PAGE/Western blot analyses. IGFBP1 protein levels were assessed in conditioned medium by loading 20 μl of medium per lane in a nonreducing loading buffer. Proteins were resolved by SDS-PAGE and transferred to Immobilon-P PVDF membranes. Membranes were blocked in PBS containing 0.1% Tween 20 and 5% nonfat milk powder. For detection of specified antigens, anti-IGFBP1, anti-Akt (Cell Signaling Technology), anti-FoxO-1 (Santa Cruz Biotechnology, Dallas, TX), and anti-phospho (p)-ERK1/2 (Cell Signaling Technology) antibodies were used at 1:1,000 dilution. The anti-FLAG M2 monoclonal (Sigma) and anti-β-actin polyclonal antibodies were used to detect the overexpressed nSMase-2 or for loading control at dilutions of 1:2,000 and 1:5,000, respectively. Incubation with the primary antibodies was followed by incubation with either goat anti-rabbit IgG-alkaline phosphatase-conjugated antibody (Sigma) or rabbit anti-mouse IgG alkaline phosphatase-conjugated secondary antibody (Sigma) at dilutions of 1:10,000. Protein-antibody interactions were visualized using the ECF kit (Amersham, Piscataway, NJ) and the Typhoon fluorescent scanning instrument (GE Healthcare Life Sciences, Pittsburgh, PA) and quantified.

nSMase activity assay.

nSMase activity was determined as described previously (43) using 6-[N-(7-nitro-2,1,3-benzoxadiazol-4-yl)amino] (NBD)-sphingomyelin (SM) as a substrate. The generation of NBD-ceramide product was analyzed by high-performance liquid chromatography on a reverse-phase column (Nova Pak, C18; Bio-Rad, Hercules, CA) using methanol-water-orthophosphoric acid (850:150:0.150 by vol) as a mobile phase at the flow rate of 2 ml/min. Product formation was proportional to the amount of added protein for up to 0.01 mg/assay.

RT-PCR.

Cellular or tissue RNA was isolated using Trizol reagent. cDNA synthesis was performed using 1.5 μg RNA and 4 units of Superscript II reverse transcriptase (Invitrogen) per sample according to the manufacturer's recommendations. Classical RT-PCR analysis was performed with Taq DNA polymerase using the following primers for the mouse IGFBP1: 5′-ctacccatggagtgggaaga-3′ (forward) and 5′-tgccctttcaaagcagaact-3′ (reverse). The PCR products were visualized by ethidium bromide staining after 1.8% agarose gel electrophoresis. For the quantitative real-time PCR, the Taqman rat IGFBP1 gene expression (Rn00565713_m1) and rodent glyceraldehyde-3-phosphate dehydrogenase (GAPDH) or β-actin control assays (Applied Biosystems, Foster City, CA) were used. Reactions were performed in triplicate in 96-well plates on an ABI Perkin Elmer Prism 7700 Sequence Detection System and analyzed with SDS 1.9.1 software (Applied Biosystems). The levels of IGFBP1 mRNA were normalized to GAPDH or β-actin mRNA.

Statistical analysis.

Student's t-test and two-way ANOVA were used. Percent change of A over B was calculated as (A − B)/B × 100 and, when discussed in the text, rounded to the nearest 0 or 5.

RESULTS

IL-1β stimulates IGFBP1 mRNA and protein expression in liver and in liver cell lines.

The effects of inflammation on hepatic IGFBP1 were studied in mice and in several hepatic cell lines and primary hepatocytes (Fig. 1). As anticipated, the intraperitoneal administration of LPS (5.6 mg/g body wt) led to a mild inflammatory response as evidenced by the robust increase in plasma proinflammatory cytokine levels, including those of IL-1β, and by the transient elevation in body temperature (data not shown). These changes were accompanied by a marked increase in IGFBP1 mRNA levels in the livers of the LPS-injected animals (Fig. 1A). Notably, direct addition of IL-1β to primary rat hepatocytes (Fig. 1B), to human Hep G2 hepatocarcinoma cells (Fig. 1C), or to rat H4-II-E hepatic cells (Fig. 1, D and E) resulted in similar effects on IGFBP1. IGFBP1 mRNA levels increased upon IL-1β stimulation in a dose- and time-dependent manner. In primary hepatocytes (Fig. 1B), the magnitude of this increase ranged from 60% (at 1.25 ng/ml, not significant) to 300% (at 10 ng/ml). In Hep G2 cells (Fig. 1C), the change reached 340%, whereas in HII2E rat hepatoma cells (Fig. 1D) the increase was between 50 and 90%, depending on the time of IL-1β treatment. The levels of IGFBP1 protein secreted in the medium were also significantly higher in the IL-1β-treated cells, and, overall, they paralleled the changes in mRNA transcription (Fig. 1, B–E, insets). These data show that inflammation in general, and IL-1β in particular, causes significant increase in IGFBP1 mRNA transcription, protein production, and secretion.

IL-1β effects on IGFBP1 mRNA and protein are seen in the absence of insulin.

Insulin is the main determinant of IGFBP1 expression and secretion. It has been presumed that proinflammatory cytokines increase IGFBP1 mRNA levels by counteracting insulin-dependent suppression. Consistent with other observations that proinflammatory cytokines can counteract insulin signaling (50), IL-1β had a significant inhibitory effect on the insulin-induced Akt phosphorylation in our system within 15 min of stimulation (Fig. 2A). In turn, insulin had no effect on IL-1β-induced JNK activation (Fig. 2A). These observations support the notion that IL-1β suppresses the insulin signaling cascade, and, therefore, the increases in the IGFBP1 mRNA and protein levels induced by IL-1β could potentially be caused by counteraction of the insulin effects. Because insulin and other growth factors are consistently present in the tissue culture media, we then investigated whether the IL-1β effects on IGFBP1 required the presence of insulin. For that purpose, we cultured cells in serum-free medium for 16–18 h and then treated them with IL-1β in the presence or absence of insulin. Importantly, treatment of H4-II-E cells for 24 h and primary rat hepatocytes for 18 h with different doses of IL-1β led to significant increases in IGFBP1 mRNA transcription (Fig. 2B) and protein secretion (Fig. 2, C–E), in the presence and in the absence of insulin. Insulin treatment alone led to an ∼60% decrease in IGFBP1 mRNA level (Fig. 2B) and between 75 and 85% decline in IGFBP1 protein levels (Fig. 2D). The IL-1β-induced increase in IGFBP1 mRNA in the absence and presence of insulin was 50 and 75%, respectively (Fig. 2B). The levels of secreted IGFBP1 increased by 210% in the absence of insulin, and, respectively, by 420, 230, and 330% in the presence of 0.1, 0.5, and 1.0 μg/ml of insulin. These data indicate that IL-1β effect can be seen in the absence of insulin, and, therefore, the underlying pathway does not involve solely a suppression of insulin regulation (Fig. 2D). It should be noted that the magnitude of the IL-1β response was not affected by low doses of insulin, but it was somewhat suppressed by high doses (0.5 and 1 μg/ml) (Fig. 2D).

Fig. 2.

IL-1β stimulation of IGFBP1 is independent of insulin. A: IL-1β inhibition of insulin-induced protein kinase B (Akt) phosphorylation. H4-II-E rat hepatoma cells were serum-deprived overnight and treated with insulin (0.5 μg/ml) and IL-1β (25 ng/ml) for 15 min as indicated. Akt and JNK phosphorylation was determined by Western blotting using antibody against the phosphorylated (active) forms of the proteins with β-actin as loading control. B: IGFBP1 mRNA levels in H4-II-E rat hepatoma cells that were serum starved overnight and then treated with insulin (0.1 μg/ml) and IL-1β (25 ng/ml) for 24 h as indicated. The abundance of GAPDH was used for normalization. The data shown represent %change with the value of the IGFBP1-to-GAPDH ratio in the control being 100%. C and D: IGFBP1 protein abundance in the medium. H4-II-E rat hepatoma cells were serum deprived (0.1% FBS) overnight and treated with insulin and IL-1β for 24 h as indicated. IGFBP1 abundance was determined by Western blotting. A representative blot (C) and quantification (D) are shown. The data shown represent %change with the abundance of the IGFBP1 protein in the medium of untreated cells being 100%. E: Western blot showing IGFBP1 protein abundance in the medium of primary rat hepatocytes cultured in the absence of serum/insulin and stimulated with IL-1β and insulin as indicated for 18 h. Data are averages ± SD (n = 3). Statistical significance is indicated (*P < 0.05 and **P < 0.01 for IL-1β effects and ##P < 0.01 for insulin effects; ns, nonsignificant).

IL-1β effects on IGFBP1 mRNA and protein are also FoxO-1-independent.

FoxO-1 is a key transcription factor that drives IGFBP1 mRNA transcription by binding to the IRE in its promoter. Our earlier studies found that IL-1β could stimulate marked increases in FoxO-1 nuclear levels and activity through a mechanism that involves JNK but that is Akt independent. Therefore, it was possible that the IL-1β effect on IGFBP1, albeit independent of the suppressive effects of insulin on IGFBP1 expression, can still be FoxO-1 mediated. We explored this possibility using siRNAi against FoxO-1. Endogenous FoxO-1 was rendered practically undetectable in H4-II-E cells treated with the siRNAi for 48 h (Fig. 3A). The level of IGFBP1 was very low in cells treated with FoxO-1 siRNAi, reaching only about 50% of the level in control-treated cells (Fig. 3B). Such decline was consistent with Foxo-1 being required for the constitutive IGFBP1 mRNA transcription. Interestingly, however, the magnitude of IL-1β stimulation of IGFBP1 mRNA was similar in cells treated with scrambled control RNAi (when IGFBP1 expression increased by 70%) and those treated with FoxO-1 siRNAi (when an increase of 120% was observed). Together, these results show that FoxO-1 transcription factor is not required for the IL-1β induction of IGFBP1.

Fig. 3.

FoxO-1 is not required for IL-1β stimulation of IGFBP1 mRNA. A: protein abundance of IGFBP1 and FoxO-1 in H4-II-E rat hepatoma cells treated with small-interfering RNA (siRNA) against FoxO-1 or with scrambled (Scr) control for 48 h. Cells were stimulated with IL-1β for 24 h under serum-free conditions. IGFBP1 levels were determined in the medium, whereas FoxO-1 and β-actin were assayed in cell homogenates. B: quantification of the IGFBP1 protein shown in A. The data shown represent %change with the abundance of the IGFBP1 protein in the medium of untreated cells being 100%. Values are averages + SE (n = 2). Statistical significance is indicated (**P < 0.01 and #P < 0.05).

IL-1β activates a 1-kb proximal IGFBP1 promoter region.

To determine whether IL-1β simulates the transcription of IGFBP1 mRNA, the dual luciferase reporter system was used. IL-1β treatment of Hep G2 cells transfected with a luciferase construct driven by a 1-kB IGFBP1 promoter region led to a gradual increase in the luciferase activity in a time-dependent manner (Fig. 4A). The luciferase activity increased as early as 6 h posttreatment by 1.2-fold and rose further by 1.7- and 2.1-fold following 16 and 24 h of stimulation, respectively. These effects of IL-1β were dose dependent (Fig. 4B). The magnitude of the change was 1.6-fold at 2.5 ng/ml and increased to 1.7-fold at 10 ng/ml and to 2-fold at 50 ng/ml. As expected, insulin treatment alone had the opposite effect and suppressed the luciferase activity by 65% within 3 h and by 40% within 6 h of treatment (Fig. 4C). This insulin effect was transient and disappeared at 16 and 24 h as opposed to the IL-1β effect that was maximal at these time points.

Fig. 4.

IL-1β stimulation of IGFBP1 promoter activity in the presence and absence of insulin. Luciferase activity of Hep G2 cells transfected with the dual luciferase reporter system. A construct containing the 1-kB IGFBP1 promoter was used. A–C: cells were serum starved for 6 h and then treated with IL-1β (25 ng/ml, A) or insulin (10 μg/ml, C) for the indicted times or with IL-1β at the indicated concentrations for 24 h (B). All treatments were done in serum-free medium. D: cells treated with IL-1β (25 ng/ml) in the presence and absence of insulin (10 μg/ml) for 24 h. E: cells treated with IL-1β (25 ng/ml, 24 h) and insulin (10 μg/ml, added 3 h before IL-1β treatment). F: cells treated with IL-1β (25 ng/ml, 24 h) and insulin (10 μg/ml, present only during the final 3 h of IL-1β treatment). Data shown represent the activity of the luciferase normalized for that of the Renilla. Statistical significance is indicated (*P < 0.05, **P < 0.01, and ***P < 0.005 for IL-1β effects and #P < 0.05 and ##P < 0.01 for insulin effects).

We further tested the combined effects of insulin and IL-1β on IGFBP1 promoter. Consistent with the observed differences in the time course on insulin and IL-1β effects, insulin had no effect on the IL-1β-induced increase in IGFBP1 promoter activity when added together with the cytokine for 24 h. A small but statistically significant inhibitory effect was seen at 6 h (Fig. 4D). To eliminate the possibility that the simultaneous addition of IL-1β and insulin interfered with proper binding of insulin to the insulin receptor, additional experiments were done. In these experiments, insulin was added either 3 h before the IL-1β addition (Fig. 4E) or during the last 3 h of the IL-1β treatment period (Fig. 4F). Adding insulin as a pretreatment led to an outcome that was identical to the changes seen when the two molecules were added simultaneously (compare Fig. 4, E with D). However, as shown in Fig. 4F, the late addition of insulin diminished the IL-1β-induced increases in promoter activity. Together, these results indicate that: 1) IL-1β and insulin regulate IGFBP1 mRNA transcription via direct and independent pathways that also follow different time courses, and 2) when both IL-1β and insulin signaling pathways are active, the net effect on IGFBP1 is determined by the magnitude of the IL-1β stimulation and the insulin-induced suppression.

IL-1β effects on IGFBP1 promoter activity are FoxO-1 independent.

To confirm the results in Fig. 3 indicating that IL-1β effects on IGFBP1 mRNA and protein levels are FoxO-1 independent, the IGFBP1 promoter-driven luciferase reporter system was used in cells overexpressing FoxO-1. FoxO-1 overexpression alone increased IGFBP1 promoter activity between three- and fivefold (Fig. 5, A and B). This is consistent with the well-established transcriptional functions of FoxO-1 (19). To confirm that the overexpressed protein is sensitive to regulation, insulin, which causes nuclear export of FoxO-1 and suppression of FoxO-1-dependent promoter activities, was used as a control. As anticipated, the insulin treatment inhibited the IGFBP1 promoter activity; notably, this inhibition was three- to fourfold greater in FoxO-1-overexpressing cells (i.e., from 35 to ∼15) than in control cells (i.e., from 8 to ∼3) based on absolute values of the normalized luciferase activity (Fig. 5A). In contrast, the magnitude of the IL-1β-induced stimulation of IGFBP1 promoter activity was similar for the control (i.e., from 8 to 18) and FoxO-1-overexpressing (from 28 to 42) cells in terms of absolute change (Fig. 5B). These results provide initial evidence that the effects of IL-1β do not involve FoxO-1. To directly show that, we deleted the IRE in the 1-kB IGFBP1 promoter construct where FoxO-1 binds. Hep G2 cells were then transfected with either wild-type (WT) or mutant (ΔIRE) 1-kB IGFBP1 promoter construct in the presence and absence of FoxO-1 overexpression. As anticipated, the ΔIRE was irresponsive to FoxO-1 overexpression (Fig. 5C); importantly, IL-1β activated the ΔIRE and the WT construct to a similar degree (Fig. 5D), confirming that IRE was not involved. Adding insulin either alone or as a cotreatment during the last 3 h of IL-1β incubation resulted in suppression of the WT but not the ΔIRE promoter, thus verifying that the mutant was also irresponsive to insulin (Fig. 5E). Together, these results conclusively show that IL-1β stimulation of the IGFBP1 promoter activity is independent of FoxO-1 and does not require IRE.

Fig. 5.

IL-1β stimulation of IGFBP1 promoter activity is independent of FoxO-1. Luciferase activity of Hep G2 cells. A and B: cells transfected with the dual luciferase reporter system using the 1-kB IGFBP1 promoter, cotransfected with FoxO-1 or control vector, serum starved for 6 h, and treated with insulin (10 μg/ml, 6 h, A) or IL-1β (50 ng/ml, 20 h, B). C: cells transfected with luciferase construct driven by the wild-type 1-kB IGFBP1 promoter (WT) or with a construct carrying a deletion of the FoxO-1-binding site (ΔIRE) and cotransfected with FoxO-1 or control (empty) vector. D: cells transfected with WT or ΔIRE luciferase construct, serum starved for 6 h, and treated with IL-1β (25 ng/ml, 24 h). E: cells transfected as described for D and treated with insulin (10 μg/ml, 6 h), IL-1β (25 ng/ml, 24 h), or with IL-1β with or without insulin during the last 3 h before harvest. The activity of Renilla reporter was used for normalization. The data shown in A–D represent normalized luciferase activity. The data shown in E are %change in the normalized luciferase activity over nontreated controls. The broken line represents 100%. Statistical significance is indicated (*P < 0.05 and **P < 0.01; the lines under the asterisks link the groups being compared).

IL-1β effects on IGFBP1 promoter activity involve the MAPK ERK1/2 and ceramide.

Our earlier studies have shown that, in hepatocytes, the IL-1β signaling cascade involves the activation of neutral sphingomyelinase and generation of ceramide at the plasma membrane (10, 18, 26). Furthermore, IL-1β-induced ceramide production is important for the proper activation of JNK and ERK in response to IL-1β stimulation (18). To test the participation of these two MAPKs in IL-1β stimulation of IGFBP1 promoter activity, we used two inhibitors [SP-600125 (a JNK inhibitor) and PD-98059 (an ERK inhibitor)]. Inhibition of JNK had no effect on the magnitude of IGFBP1 promoter activation by IL-1β. In contrast, the inhibition of ERK had profound effects (Fig. 6A). The basal activity of the IGFBP1 promoter was significantly suppressed, and the effects of IL-1β were completely negated (Fig. 6A). These results likely indicate that, in our cell culture system, transcription factors that maintain the basal IGFBP1 promoter activity are ERK substrates and require the kinase to be functionally active. One such substrate is C/EBP that, in a complex with P300 and FoxO-1, binds to and activates the IGFBP1 promoter. Our results, however, also indicate that IL-1β-induced ERK activation is necessary for the induction of IGFBP1 promoter activity by IL-1β.

Fig. 6.

Effect of inhibitors and sphingolipids on the IGFBP1 promoter activity. Normalized luciferase activity in Hep G2 cells transfected with the dual luciferase reporter system using the 1-kB IGFBP1 promoter or the ΔIRE construct. A: cells treated with IL-1β (50 ng/ml, 20 h) in the presence of inhibitor of JNK (SP-600125), ERK (PD-98059), or DMSO as a vehicle control. B: cells treated with IL-1β in the presence or absence of inhibitor of neutral sphingomyelinase (GW-4869) or appropriate vehicles. C: cells treated with ceramide (60 μM, 20 h) or bacterial sphingomyelinase (0.1 U/ml, 20 h). D: Hep G2 cells cotransfected with FoxO-1 or vector control and treated with ceramide (30 μM, 20 h). Data shown represent the activity of the luciferase normalized for that of the Renilla. Statistical significance is indicated (*P < 0.05 and **P < 0.01).

Because IL-1β-induced ceramide generation is required for the IL-1β effects on ERK, we next used a highly selective inhibitor of neutral sphingomyelinase (GW-4869) to block IL-1β-induced ceramide accumulation. GW-4869 successfully prevented IL-1β-induced activation of the IGFBP1 promoter activity (Fig. 6B). In contrast, the addition of C2- or C6-ceramide or treatment with bacterial sphingomyelinase, all of which serve to elevate the intracellular ceramide content and are sufficient to activate ERK (data not shown), resulted in small but statistically significant increases in the IGFBP1 reporter activity (Fig. 6C). Ceramide effects did not require intact IRE, and the ΔIRE-promoter was stimulated by ceramide to the same extent as the WT promoter (Fig. 6D, left). Like IL-1β, ceramide had an additive effect with FoxO-1 (Fig. 6D, right). Together, these results indicate that the pathway responsible for induction of IGFBP1 promoter activity in response to IL-1β most likely involves neutral sphingomyelinase, ceramide, and ERK and does not involve JNK or FoxO-1.

Physiological role of neutral sphingomyelinase-2 in regulation of IGFBP1 mRNA in liver.

We next studied the significance of this novel pathway for IGFBP1 regulation in a physiologically relevant system by disturbing (either overstimulation or silencing) the neutral sphingomyelinase. Primary rat hepatocytes were infected with adenovirus overexpressing the IL-1β-inducible form of the sphingomyelinase, nSMase-2 (Fig. 7), or shRNA against the rat nSMase-2 (Fig. 8) at 48 h before stimulation with IL-1β. As shown in previous studies, IL-1β stimulation of control noninfected hepatocytes leads to transient activation of nSMase activity (Fig. 7A) and the respective accumulation of ceramide (data not shown) that precedes the activation of ERK and the induction of IGFBP1 mRNA. Hepatocytes overexpressing nSMase-2 have a higher ceramide content and higher nSMase-2 activity [data not shown (26)], but also exhibit a more pronounced activation of ERK (Fig. 7, B and C). This confirms that nSMase-2 is upstream of ERK in the IL-1β signaling cascade. More importantly, the overexpression of nSMase-2 and the respective overstimulation of ERK were associated with increased basal and IL-1β-induced secretion of IGFBP1 from the hepatocytes (Fig. 7D). This confirms the significance of nSMase-2 and ERK in regulation of IGFBP1 expression.

Fig. 7.

Neutral sphingomyelianse-2 (nSMase-2) and ERK are components of the pathway for IGFBP1 mRNA upregulation in response to IL-1β. Hepatocytes were isolated from rats and cultured for 48 h under standard condition. As indicated, hepatocytes were infected with adenovirus overexpressing nSMase-2 and treated with doxycycline (1 μg/ml) to induce the expression of the transgene. Later (48 h), the medium was exchanged to such without insulin for 6 h, and cells were treated with IL-1β (10 ng/ml for A and D or as indicated in B and C) in medium containing no insulin (A–C) or to medium containing insulin at the indicated concentrations (D). A: activation of nSMase-2 in response to IL-1β stimulation. B: activation of ERK1/2 by IL-1β or nSMase-2 overexpression. ERK-1 activation was assayed using Western blotting and antibodies against the phosphorylated form of ERK. C: quantification of ERK1/2 phosphorylation. Data are represented as %change with the abundance in nontreated cells being 100%. Values are averages of three replicates. D: the levels of IGFBP1 determined by Western blotting in samples of conditioned medium collected 16 h after IL-1β stimulation, using antibody against the rat IGFBP1. The levels of β-actin in the corresponding cellular extracts were used as a loading control. The levels of the overexpressed nSMase-2 were monitored using nSMase-2-specific antibody. Results are representative of two independent experiments. Statistical significance is indicated (**P < 0.01).

Fig. 8.

Adenovirus-mediated silencing of nSMase-2 in primary rat hepatocytes. A: short-hairpin RNA (shRNA) against the rat and mouse nSMase-2. The double-stranded DNA oligo nucleotide sequence was subclonned into pENTR/U6 entry vector, and the RNAi cassette was excised and subcloned into pAdTrack vector that also encodes green fluorescent protein (GFP). Homologous recombination between the pAdTrack shuttle vector containing the RNAi cassette and the adenoviral backbone plasmid (pAdEasy-1) generated the final adenoviral product expressing shRNA against nSMase-2 (Ad-sh). B: efficiency of infection of primary hepatocytes with adenovirus expressing shRNA against nSMase-2 and GFP. Hepatocytes were cultured on Matrigel for 48 h and infected with Ad-shRNAi at multiplicity of infection (MOI) of 5. GFP fluorescence was observed under a fluorescent microscope in live cells. C: efficiency of silencing of nSMase-2 protein synthesis. HEK293 cells were infected with Flag-tagged rat nSMase-2 (Ad-nSMase-2) at an MOI of 1, and expression was stimulated with doxycycline (1 μg/ml). Later (1 h), the cells were also infected either with Ad-sh or the same adenovirus expressing scrambled sense-loop-antisense sequence (Ad-scr). The expression of the FLAG-tagged nSMase-2 was judged by anti-FLAG Western blotting. D: silencing of nSMase-2 prevents the IL-1β-induced increase in nSMase-2 activity. Hepatocytes were cultured for 48 h, infected with Ad-shRNAi at an MOI of 5, and stimulated with IL-1β 48 h later. Total hepatocyte nSMase activity was measured using 6-[N-(7-nitro-2,1,3-benzoxadiazol-4-yl)amino] (NBD)-sphingomyelin (SM) as a substrate. E: silencing of nSMase-2 prevents IL-1β-induced synthesis and secretion of IGFBP1. Hepatocytes from rats were infected with Ad-nSMase-2 at an MOI of 5, and expression was stimulated with doxycycline (1 μg/ml). Later (48 h), the medium was exchanged to such without insulin, and cells were treated with IL-1β (10 ng/ml) for 24 h. IGFBP1 protein was measured in the medium using Western blotting.

Notably, the apparent overstimulation of the pathway resulting from the nSMase-2 overexpression rendered it somewhat resistant to insulin suppression (Fig. 7D). This is clearly seen when comparing the effectiveness of insulin at high dose (2.5 ng/ml) in negating the IL-1β effects in control and nSMase-2-overexpressing hepatocytes (Fig. 7D, bottom): insulin seems to diminish the effectiveness of IL-1β in inducing IGFBP1 mRNA only in control but not in nSMase-2-overexpressing hepatocytes, where the IL-1β pathway is overstimulated. In this latter case, IL-1β treatment effectively stimulates the IGFBP1 production even when insulin is present at high concentration.

We then tested whether nSMase-2 activity was required for the IL-1β induction of IGFBP1 in primary hepatocytes. For that purpose, shRNAi against the rat nSMase-2 (Fig. 8A) was cloned into an adenoviral vector (also expressing a GFP tag), and the resulting construct (Ad-sh-RNAi) or the respective control expressing a scrambled sequence (Ad-scr-RNAi) was used to infect primary hepatocytes. The infection was highly efficient, with almost 100% of the cells being GFP positive (Fig. 8B). Because the level of endogenous nSMase-2 in hepatocytes is very low, and only trace amounts can be seen with nSMase-2 antibody, the efficiency of silencing was tested using a dual expression system where hepatocytes were infected simultaneously with two adenoviruses, one expressing the rat nSMase-2 and another, either the shRNAi or the scrRNAi sequence. There was a substantial decline in the levels of the overexpressed nSMase-2 in cells infected with Ad-sh-RNAi (Fig. 8C). The same adenoviral construct also completely suppressed the IL-1β stimulation of the endogenous nSMase-2 activity (Fig. 8D). This is consistent with the already established role of nSMase-2 as the only bona fide signaling neutral sphingomyelinase that is IL-1β inducible. More importantly, the upregulation of IGFBP1 secretion in response to IL-1β was completely suppressed in cells with silenced nSMase-2 (Fig. 8E). Together, these results clearly show that nSMase-2 is an essential step in the IL-1β signaling pathway that leads to IGFBP1 upregulation in hepatocytes.

DISCUSSION

Increases in the circulating levels of IGFBP1 contribute to the onset of catabolic state in inflammation and are associated with increased frailty and weakness, as well as with impaired tissue regeneration and glucose homeostasis (3, 28, 33). The mechanisms behind these increases in IGFBP1 are complex and involve changes in hormonal homeostasis, as well as elevated blood levels of proinflammatory cytokines like IL-1β and TNF-α. Increased plasma IGFBP1 has been well documented in humans during septic shock (47) and in animal models of experimental endotoxemia (6, 13, 38). At least in rats, the administration of IL-1β receptor antagonist effectively prevents the IGFBP1 increases in blood and in the liver during sepsis (30), suggesting that IL-1β is a key systemic mediator regulating IGFBP1 mRNA expression. The intracellular signaling mechanisms involved in these effects, however, remain poorly understood. The goal of this manuscript is to decipher the key steps of the pathway(s) of IL-1β-induced stimulation of IGFBP1 in the liver with specific emphasis on the possible interactions with the insulin-dependent regulation.

In a healthy state, insulin is the main determinant of IGFBP1 mRNA transcription by virtue of regulating the nuclear localization and activity of FoxO-1 transcription factor. In the absence of insulin, FoxO-1 (that has a nuclear localization signal and is constitutively targeted to the nucleus) is bound to the IRE in the IGFBP1 promoter and drives its transcription. Increases in systemic insulin level, however, lead to the phosphorylation and activation of Akt that, after relocating to the nucleus, phosphorylates FoxO-1 inducing FoxO-1 nuclear export and, consequently, a suppression of IGFBP1 mRNA transcription. This mechanism ensures strict regulation of the hepatic IGFBP1 production in a manner that reflects the fasted/fed state of the organisms.

Negation of insulin-dependent suppression of IGFBP1 mRNA transcription has been thought to explain the increases in IGFBP1 production during inflammation because all main inflammatory cytokines, including TNF-α, IL-1β, and IL-6, inhibit the insulin signaling pathway upstream of FoxO-1 (50). IL-1β in particular has been shown to downregulate the expression of insulin receptor substrate-1 and to inhibit Akt phosphorylation (5, 17, 24, 27, 45, 52). This present study also found a substantial decline in the magnitude of insulin-induced Akt activation in the presence of IL-1β. At the same time, however, our experiments show that suppression of insulin-dependent regulation is not the major mechanism of IL-1β stimulation of IGFBP1. For one thing, IL-1β effects are seen both in the presence and in the absence of insulin. For another, IL-1β stimulates directly a luciferase reporter driven by the 1-kB IGFBP1 promoter, and deletion of the IRE does not diminish this effect. Last but not least, the IL-1β and insulin effects follow very different time course: while the insulin-dependent suppression of the IGFBP1 promoter activity is a relatively rapid event and occurs within the first several hours after insulin addition, it takes at least 6 h to observe the activation of the IGFBP1 promoter following IL-β stimulation. Together, these observations delineate a novel and direct pathway for the regulation of IGFBP1 promoter in response to IL-1β.

Our earlier studies showed that IL-1β could increase FoxO-1 nuclear content via a mechanism that involves JNK activation but that is Akt and insulin independent (11). Therefore, we specifically tested the involvement of FoxO-1 and JNK and found that both molecules were dispensable for the stimulation of IGFBP1 promoter by IL-1β. Instead, our studies found that ERK was involved (Fig. 9). The plasma membrane-localized neutral sphingomyelinase-2 and ceramide were also essential components of the IL-1β pathway leading to activation of IGFBP1 promoter. Activation of nSMase-2 and transient accumulation of ceramide are early signaling events in the IL-1β cascade and are required for optimal activation of several key steps in the pathway, including IRAK-1, JNK, ERK, FoxO-1, and C/EBP (10, 11, 18, 26). Suppressing the activity of both nSMase-2 and ERK either by pharmacological inhibitors or via genetic means alleviated the increases in IGFBP1 mRNA evoked by IL-1β. In contrast, nSMase-2 overexpression potentiated the IL-1β effects on both ERK and IGFBP1, indicating that nSMase-2 and its product ceramide acted upstream of ERK.

Other proinflammatory cytokines, including TNF-α, IL-1α, and IL-6, have also been reported to stimulate the expression of IGFBP1 mRNA in vitro (3). These three cytokines engage distinct signaling pathways, none of which resembles the signaling cascade initiated by IL-1β. Therefore, the underlying mechanisms leading to increases in IGFBP1 mRNA levels in each case are clearly different. In the case of TNF-α and IL-1α, for example, the effects seen on IGFBP1 mRNA levels were not recapitulated by corresponding changes in IGFBP1 promoter activity. In fact, 18 h following stimulation, both IL-1α and TNF-α suppressed the luciferase activity driven by the same 1-kB IGFBP1 promoter examined in this present study, indicating that TNF-α and IL-1α might stimulate transcription through cis-acting DNA sequences located outside of the 1-kB promoter, or that these cytokines induce elevation of IGFBP1 mRNA by acting at the level of mRNA stability (3). In turn, the IL-6-induced stimulation of IGFBP1 mRNA was shown to involve an activation of the 1-kB proximal promoter, similarly to IL-1β. These effects, however, were mediated by STAT-3 and activator protein-1 (c-Fos/c-Jun), factors acting on the IGFBP1 promoter via the HNF-1 site (35).

Another important observation of this study is that IL-1β effect on IGFBP1 is apparently sensitive to cotreatment with insulin. Although it is unclear whether a similar trend is present in vivo, at least in vitro, insulin cotreatment can effectively diminish the IL-1β-induced stimulation of IGFBP1. This is another confirmation for the existence of two distinct pathways by which insulin and IL-1β regulate IGFBP1 expression through distinct cis-acting DNA sequences in the IGFBP1 promoter. In conclusion, this manuscript delineates a novel pathway for the regulation of IGFBP1 mRNA that involves ERK, nSMase-2, and ceramide and does not require insulin, Akt, or the IRE. These results indicate the existence of alternative approaches of suppressing inflammation-induced elevation in IGFBP1 without interfering with systemic levels of IL-1β.

GRANTS

This work was supported by National Institute on Aging Grants AG-019223 and AG-026711.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

L.S., D.B., A.D., S.S., A.K., and N.G. performed experiments; L.S., D.B., and M.N.N.-K. analyzed data; L.S., D.B., and M.N.N.-K. interpreted results of experiments; L.S., D.B., A.D., S.S., A.K., N.G., and M.N.N.-K. approved final version of manuscript; D.B. and M.N.N.-K. prepared figures; A.K. and M.N.N.-K. edited and revised manuscript; M.N.N.-K. conception and design of research; M.N.N.-K. drafted manuscript.