il-6 stimulation of insulin-like growth factor binding protein (igfbp)-1 production

BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS 228, 611–615 (1996)ARTICLE NO. 1705

IL-6 Stimulation of Insulin-like Growth Factor BindingProtein (IGFBP)-1 Production

Benjamin Samstein, Matthew L. Hoimes, Jie Fan, Robert A. Frost,Marie C. Gelato, and Charles H. Lang1

Departments Surgery and Medicine/Endocrinology, State University of New Yorkat Stony Brook, Stony Brook, New York 11794-8191

Received September 27, 1996

TNFa and IL-1b have previously been shown to increase the IGFBP-1 concentration in plasma and liverunder in vivo conditions. The present study demonstrates that another inflammatory cytokine, IL-6, alsoelevates a 30- to 32-kDa IGF binding protein in the plasma of mice. Moreover, IL-6 produced dose- andtime-dependent increases in IGFBP-1 production by HepG2 cells. The maximal IL-6-induced increase inIGFBP-1 was comparable to that observed with dexamethasone, and this increase was attenuated bydiltiazem or dantrolene, both of which are known to reduce the cytosolic Ca2/ concentration. Finally,incubation of HepG2 cells with TNFa or IL-1b also increased IGFBP-1 in a dose-dependent manner. Theseresults demonstrate that IGFBP-1 production is mediated directly by proinflammatory cytokines and suggestthat this mechanism may be important for the upregulation of IGFBP-1 seen in catabolic conditionsassociated with overexpression of these cytokines. q 1996 Academic Press, Inc.

IGFBP-1 is one of six structurally homologous proteins which can bind IGF-I and IGF-II.Although the exact functional significance of changes in IGFBP-1 within the plasma andtissues remains unclear, numerous studies have indicated that the predominant effect of amolar excess of this binding protein is a generalized inhibition of IGF-I action (1). Hence,alterations in IGFBP-1 levels in vivo may represent an important mechanism for regulatingthe bioactivity of IGF-I, and thereby may modulate key IGF-sensitive metabolic pathways (2).

One of the hallmarks of various inflammatory and catabolic conditions is the rapid andsustained elevation in blood-borne IGFBP-1 (3-9). Gram-negative infection, peritonitis, andendotoxin administration, have also been reported to markedly elevate the content of IGFBP-1 in the liver (3,5-7), which is believed to be the primary source of circulating IGFBP-1.However, the mechanism by which hepatic IGFBP-1 is increased in these conditions remainsobscure. These conditions, in general, are associated with overexpression of inflammatorycytokines, such as TNFa, IL-1 and IL-6 (10,11). Previous studies by our laboratory havedemonstrated that in vivo administration of either TNFa or IL-1b to rats upregulates hepaticIGFBP-1 production (5,6). However, the results from these in vivo studies do not clarifywhether the increase in IGFBP-1 is mediated directly by cytokines or indirectly via alterationsin other known hormonal regulators, such as glucocorticoids and insulin (12,13). Furthermore,although the pleiotropic cytokine IL-6 is known to be a potent stimulus for ‘‘positive’’ acutephase protein secretion in the liver (14), its ability to regulate IGFBP-1 secretion has not beenexamined.

1 Corresponding author. Fax (516) 444-2011. E-mail: [email protected]: IGF, insulin-like growth factor; IGFBP-1, IGF binding protein-1; TNFa, tumor necrosis factor-

alpha; IL-1b, interleukin-1 beta; hrIL-6, human recombinant interleukin-6; NCI, National Cancer Institute; SDS-PAGE, sodium dodecyl sulfate-polyacrylamide gel electrophoresis; BSA, bovine serum albumin; MEM, minimumessential medium.

0006-291X/96 $18.00Copyright q 1996 by Academic Press, Inc.All rights of reproduction in any form reserved.

611

AID BBRC 5675 / 6910$$1161 10-25-96 08:18:31 bbrca AP: BBRC

Vol. 228, No. 2, 1996 BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS

The purpose of the present study was to determine whether in vivo administration of IL-6alters circulating levels of IGFBP-1, and whether addition of IL-6 to human hepatoma HepG2cells in vitro could directly stimulate production of this binding protein. Since previous studiesindicated that IL-6 mediates acute phase protein secretion via a calcium-dependent mechanism(15), additional experiments aimed to determined whether the effect of IL-6 on IGFBP-1release could be modulated by agents which effectively reduce cytosolic Ca2/ by differentmechanisms. Finally, since TNFa and IL-1b have already been demonstrated to increaseIGFBP-1 in vivo, studies were conducted to determine whether these cytokines could directlystimulate IGFBP-1 production in cultured HepG2 cells.

MATERIALS AND METHODS

In vivo experiment. Fed, male, 9- to 10-week old, Swiss-Webster mice (Taconic Farms, Germantown, NY) weighingbetween 30-35 g were used. On the morning of the experiment, each mouse was mildly restrained by placing it insidea plastic tube. Animals were allowed to rest quietly for 30 min before hrIL-6 (2.5 mg/10 g; Upstate BiotechnologyInc., Lake Placid, NY) or an equal volume (200 ml) of vehicle (0.1% BSA) was injected via percutaneous punctureof the tail vein. Mice were returned to individual cages and sacrificed by decapitation 4 h later. Blood was collectedin heparinized tubes and plasma stored at 070 7C. This time point was chosen because previous studies have indicatedthat IGFBP-1 is increased at this time in animals injected with endotoxin or IL-1b (3,6).

Cell culture. HepG2 cells were grown in 24-well plates (Falcon; B-D, Lincoln Parks, NJ) containing 500 ml ofMEM (Sigma; St. Louis, MO; 1.5 mM calcium) supplemented with 5% FCS, penicillin (100 U/ml), streptomycin(100 mg/ml) and amphotericin B (25 mg/ml) for 5-7 days after subculture, at which time cells were near confluence.HepG2 cells were used because they are cytokine-responsive (16), and have a constitutive secretion of IGFBP-1 thatis under hormonal regulation (17). On the day of the experiment, the medium was replaced with serum-free MEMcontaining either IL-6 (0-100 ng/ml), mrTNFa (0-100 ng/ml; gift of Genetech, South San Francisco, CA), hrIL-1b(0-100 ng/ml; gift of Biological Response Modifiers Program, Division of Cancer Treatment/NCI), dexamethasone(3 mM), endotoxin (Escherichia coli 026:B6; 1 mg/ml; Difco, Detroit, MI), insulin (0.1-10 nM), dantrolene or diltiazem(25-200 mM). The latter three compounds were obtained from Sigma. Cells were incubated for various periods oftime, and the medium collected and stored at 070 7C.

IGFBP-1 determination. The IGFBPs in plasma were determined by ligand blot analysis as previously described(3). Protein content in the supernatant was assayed to ensure that each sample had an equal concentration. Westernblot analysis was used to determine the relative concentration of IGFBP-1 in medium from HepG2 cells (7,8). Sampleswere separated on a 12.5% SDS-PAGE gel under nonreducing conditions. Proteins were electroblotted overnight ontonitrocellulose, and blocked with TRIS-buffered saline containing 1% nonfat dry milk. Membranes were then incubatedwith a 1:2000 dilution of antiserum against human IGFBP-1 (UBI, Lake Placid, NY). Antigen-antibody complexeswere identified with goat anti-rabbit IgG tagged with horseradish peroxidase (Sigma) and exposed to the enhancedchemiluminescence detection system (Amersham, Arlington Heights, IL). Band intensities were determined using alaser densitometer (Hoeffer, San Francisco, CA).

Statistical analysis. Results are presented as means { SE with the number of observations listed in the figurelegends. Statistical comparisons were made by ANOVA followed by Student-Newman-Keuls to determine treatmenteffect. Statistical significance was set at P õ 0.05.

RESULTS AND DISCUSSION

In vivo effect of IL-6. Figure 1 is a ligand blot of plasma obtained from mice 4 h after theinjection of IL-6 or vehicle. In all IL-6 treated mice, the band at 30-32 kDa, which representsboth IGFBP-1 and -2, was markedly increased Ç6-fold above control values. It is not possibleto conclude definitively whether this increase resulted from an elevation in IGFBP-1 and/or -2 because of the inherently poor resolution of these binding proteins in plasma from rodentsby ligand blotting, and the lack of a specific mouse IGFBP-1 antibody to perform Westernblot analysis. However, in previous studies using rats, other inflammatory conditions havebeen reported to increase predominantly IGFBP-1 (5-7). Therefore, it seems highly likely thatthe increased density of the band at 30-32 kDa represents an increase in IGFBP-1. IL-6 alsoproduced a modest increase in the concentration of IGFBP-3 and-4 in the plasma.

Response of HepG2 cells to IL-6. A significant Ç50% increase in the concentration ofIGFBP-1 in the conditioned medium was detected at an IL-6 concentration as low as 1 ng/ml

612



FIG. 1. Ligand blot of IGFBPs in plasma from mice injected with vehicle (lanes 1-4) and IL-6 (lanes 1-3) via atail vein. Molecular weight standards were lysozyme (14,300), carbonic anhydrase (30,000), ovalbumin (46,000), andBSA (69,000; not shown). Densitometry indicated that IL-6 increased plasma levels of IGFBP-1 and/or -2 (i.e., bandat 30-32 kDa) by more than 6-fold.

(Fig. 2). The levels of IGFBP-1 increased in a dose-dependent manner until a plateau wasreach at IL-6 concentrations of between 30-100 ng/ml. The maximal response produced by IL-6 (Ç 5-fold) was essentially identical to that obtained in cells treated with a known maximally-stimulating dose of dexamethasone (18). IL-6 also produced a time-dependent increase inIGFBP-1 by HepG2 cells which was maximal between 8-18 h and appeared to plateau thereafter(Fig. 2). This dose- and time-dependent stimulation of IGFBP-1 is comparable to that observedfor the IL-6 induced stimulation of several acute phase proteins (16).

Cells incubated with either a pharmacological dose of endotoxin or heat-inactivated IL-6failed to stimulate IGFBP-1 production (Fig. 3, left). Under basal conditions, insulin (10-100nM) was able to decrease the IGFBP-1 concentration (Fig. 3, middle), as previously described(18). However, insulin was unable to significantly attenuate the IL-6 induced increase inIGFBP-1 when cells were incubated simultaneously with both peptides (Fig. 3, right). Thislack of response differs from the ability of insulin to attenuate the increased IGFBP-1 producedby dexamethasone (19).

Sepsis and endotoxemia, conditions associated with elevated circulating levels of IL-6, areknown to increase the cytosolic Ca2/ concentration in hepatocytes (20). Moreover, calciumchannel blockers have been demonstrated to reverse or attenuate several of the hepatic manifes-tations of infection (21). Therefore, additional studies were performed to determine whether

FIG. 2. Dose- and time-dependent increases in IGFBP-1 in medium from HepG2 cells incubated with IL-6. Dataare expressed in arbitrary densitometric units (AU). (Left) All doses of IL-6 (1-100 ng/ml) significantly (P õ 0.05)increased the IGFBP-1 concentration in the medium above that seen in the cells without cytokine (‘‘0’’). (Right) Alltime points between 12-36 h showed a significant (P õ 0.05) elevation in IGFBP-1 above time 0. Each bar representsthe mean { SEM of 6-8 cell wells.

613



FIG. 3. (Left) Representative Western blot of IGFBP-1 in conditioned medium from HepG2 cells incubated withvehicle (‘‘control’’, lane 1), endotoxin (ET, lane 2), heat-inactivated IL-6 (HI, lane 3), and IL-6 (30 ng/ml; lane 4).(Middle) Western blot of insulin-induced inhibition of constitutive basal IGFBP-1 production. Cells were incubatedfor 20 h with increasing doses of insulin. (Right) Representative Western blot of IGFBP-1 in medium from cellsincubated { insulin and { IL-6. Lane 1, no insulin or IL-6; lanes 2 and 3, cells incubated with IL-6 (3 ng/ml) {insulin (10 nM); lanes 4 and 5, cells incubated with IL-6 (30 ng/ml) { insulin (10 nM).

the IL-6 induced change in IGFBP-1 secretion by HepG2 cells was calcium dependent. In thisstudy, the calcium-channel blocker diltiazem was used to inhibit receptor-mediated influx ofCa2/ across the plasma membrane, whereas dantrolene was used to prevent Ca2/ release fromintracellular stores. Incubation of cells with diltiazem greatly inhibited the IL-6 induced increasein IGFBP-1 (Fig. 4, left). Similarly, addition of dantrolene to cells also prevented stimulationof IGFBP-1 secretion by IL-6 (Fig. 5, right). Moreover, ionomycin, which elevates cytosolicCa2/, increased IGFBP-1 production by HepG2 cells in the absence of IL-6 (Fig. 4). In contrast,neither diltiazem nor dantrolene significantly reduced constitutive production of IGFBP-1 underbasal, non-stimulated conditions (data not shown). These results strongly suggest that increasesin the intracellular calcium concentration are capable of stimulating IGFBP-1 production, andthat the IL-6 induced increase in IGFBP-1 is mediated, at least in part, by calcium-dependentprocesses.

Influence of TNFa and IL-1b. Although previous studies indicate that TNFa and IL-1bincrease IGFBP-1 levels in plasma and liver when injected in vivo (5,6), it remained to beelucidated whether these cytokines produce their effects directly or act indirectly via increasesin other hormonal regulators. Figure 5 illustrates that both TNFa and IL-1b produce a dose-dependent increase in IGFBP-1 concentration in HepG2 cells, indicating a direct hepatic effect.This response was absent in cells treated with heat-inactivated cytokines (data not shown). Ittherefore appears likely that in vivo these cytokine signals originate from activated Kupffercells which, because of their juxtaposition, modulate IGFBP-1 production by hepatocytes.

FIG. 4. Representative Western blot of IGFBP-1 in conditioned medium from HepG2 cells incubated with IL-6alone or in the presence of diltiazem or dantrolene. (Left) Cells were incubated with vehicle (C) or IL-6 { differentdoses (50-200 mM) of the calcium channel antagonist diltiazem, which was dissolved in MEM prior to addition tocells. (Right) Cells incubated with vehicle (DMSO; final concentration 0.2%) and no IL-6 (lane 1) or IL-6 { differentdoses of dantrolene (0-100 mM) dissolved in DMSO (£ 0.2%). Lane 6 is from cells stimulated with ionomycin (1mM). Triplicate determinations from 3 separate experiments were performed for all studies.

614



FIG. 5. Dose-dependent increase in IGFBP-1 concentration in conditioned medium from HepG2 cells incubatedwith increasing concentrations of TNFa (left) or IL-1b (right). Cytokines were dissolved in MEM prior to additionto cells. Triplicate determinations from 3 separate experiments were performed.

SUMMARY

We have shown that the proinflammatory cytokines IL-6, IL-1b and TNFa can directlystimulate IGFBP-1 production by HepG2 cells, and that the ability of all the cytokines examinedto enhance IGFBP-1 release occurs at concentrations which have been detected in the bloodduring various inflammatory conditions. Furthermore, the IL-6 induced increase cannot beovercome by the inhibitory effects of insulin and appears mediated via a calcium-dependentmechanism. These data, in conjunction with in vivo findings, support a physiological role forthese cytokine signals as regulators of IGFBP-1 synthesis during inflammation and, hence, asimportant regulators of IGF-I bioactivity and bioavailability to tissues.

ACKNOWLEDGMENTSThis work was supported by NIH Grants GM 38032, AA 11290, and DK 49316.

REFERENCES1. Lee, P. D. K., Conover, C. A., and Powell, D. R. (1993) Proc. Soc. Exp. Biol. Med. 204, 4–29.2. Lewitt, M. S., Denyer, G. S., Cooney, G. J., and Baxter, R. C. (1991) Endocrinology 129, 2254–2256.3. Fan, J., Molina, P. E., Gelato, M. C., and Lang, C. H. (1994) Endocrinology 134, 1685–1692.4. Wojnar, M. M., Fan, J., Frost, R. A., Gelato, M. C., and Lang, C. H. (1995) Am. J. Physiol. 268, R970–R977.5. Fan, J., Bagby, G. J., Gelato, M. C., and Lang, C. H. (1995) Am. J. Physiol. 269, R1204–R1212.6. Fan, J., Wojnar, M. M., Theodorakis, M., and Lang, C. H. (1996) Am. J. Physiol. 270, R621–R629.7. Lang, C. H., Fan, J., Cooney, R., and Vary, T. C. (1996) Am. J. Physiol. 270, E430–E437.8. Lang, C. H., Fan, J., Frost, R. A., Gelato, M. C., Sakurai, Y., Herndon, D. N., and Wolfe, R. R. (1996) J. Clin.

Endocrinol. Metab. 81, 2474–2480.9. Frost, R. A., Fuhrer, J., Mariuz, P., Lang, C. H., and Gelato, M. C. (1996) Clin. Endocrinol. 44, 501–514.

10. Siaud, P., Assenmacher, I., and Barbanel, G. (1994) Am. J. Physiol. 267, R164–R170.11. Ertel, W., Morrison, M. H., Wang, P., Ba, Z. F., Ayala, A., and Chaudry, I. H. (1991) Ann. Surg. 214, 141–148.12. Luo, J., Reid, R. E., and Murphy, L. J. (1990) Endocrinology 127, 1456–1462.13. Gelato, M. C., Alexnder, D., and Marsh, K. (1992) Diabetes 41, 1511–1519.14. Heinrich, P. C., Castell, J. V., and Andus, T. (1990) Biochem. J. 65, 621–636.15. Rose, S., Baumann, H., Jahreis, G. P., and Sayeed, M. M. (1994) Shock 1, 87–93.16. Mackiewicz, A., Speroff, T., Ganapaathi, M., and Kushner, I. (1991) J. Immunol. 146, 3032–3037.17. Powell, D. R., Lee, P. D. K., DePaolis, L. A., Morris, S., and Suwanichkul, A. (1993) Growth Regul. 3, 11–13.18. Arany, E., Strain, A. J., Hube, M. J., Phillips, I. D., and Hill, D. J. (1993) J. Cell. Physiol. 155, 426–435.19. Lewitt, M. S., Saunders, H., and Baxter, R. C. (1996) J. Cell. Physiol. 166, 121–129.20. Rose, S., Thompson, K. D., and Sayeed, M. M. (1992) Am. J. Physiol. 263, R5530–R558.21. Song, S-K., Karl, I. E., Ackerman, J. J. H., and Hotchkiss, R. S. (1993) Proc. Natl. Acad. Sci. USA 90, 3933–

3937.22. Biffl, W. L., Moore, E. E., Moore, F. A., Barnett, C. C., Silliman, C. C., and Peterson, V. M. J. Leukocyte Biol.

59, 569–574.

615


il-6 stimulation of insulin-like growth factor binding protein (igfbp)-1 production

Documents