Lipopolysaccharide Directly Stimulates Cortisol Secretion by Human Adrenal Cells by a Cyclooxygenase-Dependent Mechanism
http://www.100md.com
内分泌学杂志 2005年第3期
Department of Endocrinology, William Harvey Research Institute, Barts and the London, Queen Mary’s School of Medicine and Dentistry, Queen Mary, University of London, London EC1A 7BE, United Kingdom
Address all correspondence and requests for reprints to: Dr. J. P. Hinson, Department of Endocrinology, Barts and the London, Queen Mary School of Medicine and Dentistry, Suite 12, Dominion House, Bartholomew Close, London EC1A 7BE, United Kingdom. E-mail: j.p.hinson@qmul.ac.uk.
Abstract
Activation of the hypothalamo-pituitary-adrenal axis by bacterial lipopolysaccharide (LPS; endotoxin) is well documented, although there has been uncertainty about whether LPS exerts a direct effect at the level of the adrenal. The present study found that LPS caused a dose-dependent stimulation of basal cortisol secretion by the human adrenocortical cell line, NCI-H295R, without affecting aldosterone. The expression of both Toll-like receptor 2 (TLR2) and TLR4 was demonstrated in these cells, and the specific ligands for TLR4 (purified LPS and lipid A) and TLR2 (Pam3Cys) were found to stimulate cortisol release, suggesting that these receptors may mediate the effects of LPS in adrenal cells, as has been shown in other cell types. LPS was also found to stimulate prostaglandin E2 release by these cells. The effects of LPS on cortisol were attenuated in the presence of both indomethacin and a specific COX-2 inhibitor, but not a COX-1 inhibitor, suggesting an obligatory role for COX-2 activation and prostaglandin synthesis in the adrenal response to LPS.
Introduction
ACTIVATION OF THE hypothalamo-pituitary-adrenal (HPA) axis is an important part of the stress response. It is known that there are multiple examples of cross-talk between the immune and endocrine systems, and in acute infection this is vital to maintain homeostasis. It is well established that bacterial endotoxin, a lipopolysaccharide (LPS) component of the bacterial cell wall, is a potent activator of the HPA axis (1). Although previous studies have shown that LPS exerts this effect principally by stimulating CRH secretion (1), there is also evidence for CRH-independent effects (2, 3). It appears that LPS may act on each tissue comprising the HPA axis, because there is also evidence for a pituitary-independent effect of LPS on the adrenal gland, with LPS causing increased corticosterone secretion in both intact and hypophysectomized rats (4). More recently, it has been shown that LPS inhibits ACTH-stimulated corticosterone secretion by cultured rat zona glomerulosa cells, although the effect on basal secretion was not determined (5). There are conflicting results in rodents, however, because it has been shown that the presence of the hypothalamus is necessary for a corticosterone response to LPS in rats (6). An earlier study also reported that LPS did not have a direct adrenal effect in mice in vivo (7). Preliminary studies from our laboratory have shown that LPS acts directly on human adrenal cells to stimulate cortisol secretion (8).
LPS from Gram-negative bacteria (Escherichia coli LPS) is known to act though Toll-like receptors (TLRs), specifically through TLR4 (9). The TLRs are a receptor family, related to the IL-1 receptor, with multiple ligands and a range of signal transduction pathways (for review, see Refs. 9, 10, 11). Binding of LPS to TLRs is not simple. It appears that a range of other molecules is required for TLRs to recognize the LPS signal. These include CD14, MD-2, and LPS-binding protein, which are all important in the presentation of LPS and its recognition by the TLR (12). Recently, binding of LPS to adrenal cells has been reported (13), and the expression of TLRs has been described in the human adrenal (14).
It is clear that the TLRs signal through a range of different pathways to stimulate the release of cytokines, particularly members of the IL family (10). LPS action has also been linked to prostaglandin (PG) production in some tissues (15), and there is evidence that in intact rats, LPS administration causes up-regulation of COX-2 in the adrenal gland (16). It is not clear whether this is a direct action of LPS on the adrenal, however.
A previous study from our laboratory indicated that LPS exerted a direct action on human adrenocortical cells (8). The present study was designed to extend these findings, determine the dose-response relationship, and elucidate the time course and mechanism of this effect. These studies used the human adrenocortical cell line, NCI-H295R, which produces the normal range of human adrenal steroids (17). It has the advantage of being a pure adrenocortical cell preparation with no contaminating endothelial, medullary, or mast cells that could potentially influence the response to LPS.
Materials and Methods
The NCI-H295R cell line was a gift from Ian Mason (University of Edinburgh, Edinburgh, UK). All tissue culture medium and supplements were obtained from Invitrogen Life Technologies, Inc. (Paisley, UK), with the exception of +1 ITS medium supplement (Universal Biologicals, Gloucester, UK) and Ultroser SF (Biosepra, Cergy-Saint-Christophe, France). All other reagents were obtained from Sigma-Aldrich Corp. (Poole, UK) and were of molecular grade unless otherwise stated.
Crude LPS (E. coli 0127:B8) was obtained from Sigma-Aldrich Corp. A highly purified preparation of LPS (E. coli 0127:B8) was also purchased from Sigma-Aldrich Corp. (prepared by phenolic extraction and gel filtration chromatography; protein content, <1%; RNA, <1%). This is referred to as pure LPS. A stock solution of LPS dissolved in water was aliquoted and stored at –20 C. Lipid A (from E. coli) was dissolved in dimethylsulfoxide (1 mg/ml). Pam3Cys (EMC Microcollections, Tubingen, Germany) was dissolved in water (1 mg/ml). Indomethacin was dissolved in ethanol (10 mg/ml), and forskolin, SC560, and NS398 were each dissolved in dimethylsulfoxide (5 mg/ml). Stock solutions were diluted with cell culture medium to give the final required concentration. Control experiments were carried out with each of the solvents alone at the final concentration in the incubation. None had any effect on cortisol release.
Cell culture
H295R cells were routinely maintained in 75-cm2 tissue culture flasks in DMEM/Ham’s F-12 Nutrient Mix supplemented with 2% (wt/vol) Ultroser SF, 1% ITS (6.25 mg insulin, 6.25 mg transferrin, 6.25 mg selenium, and 5.35 mg linoleic acid) and 1% penicillin/streptomycin at 37 C under an atmosphere of 95% air/5% carbon dioxide. For all experiments, preincubations and incubations were carried out in serum-free medium which consisted of DMEM/Ham’s F-12 Nutrient Mix supplemented only with 1% penicillin/streptomycin. Cells were passaged 1:2 using a solution of trypsin (0.5 g/liter)/EDTA (0.2 g/liter) every 4 d.
Cell incubations
H295R cells were plated to a density of 500,000 cells/well in six-well plates and were serum-starved for 24 h before treatment. Cells were incubated in serum-free medium in the presence or absence of LPS for different time periods. The effects of the specific TLR agonists were investigated by incubating cells in the presence or absence of these agents for 24 h. Lipid A was used at 1 μg/ml, and Pam3Cys was used at 5 μg/ml. At these concentrations, lipid A is a specific TLR4 agonist (18), and Pam3Cys is a specific TLR2 agonist (19).
To investigate the effects of the cyclooxygenase (COX) inhibitors, indomethacin, SC 560, and NS398, experiments were carried out as described above, using forskolin (10–5 mol/liter) as a positive control. Cells were incubated for 24 h with either LPS or forskolin in the presence or absence of the inhibitor. Indomethacin was used at a concentration of 10–5 mol/liter, SC560 at 10–7 mol/liter, and NS398 at 5 x 10–5 mol/liter. At these concentrations, SC560 is a selective inhibitor of COX-1 (20), and NS398 is a selective inhibitor of COX-2 (21).
At the end of the incubation period, medium was removed from the cells and stored at –20 C until assayed. Cell viability was determined using a 3-[4,5-dimethlthiazol-2-yl]-2,5-diphenyltetrazolium bromide assay (Promega Corp., Madison, WI). None of the treatments used in this study had any effect on cell viability.
Assays
Cortisol was measured by RIA, using an antibody obtained from Bioclin (Cardiff, UK); labeled cortisol was purchased from Amersham Biosciences (Little Chalfont, UK), and the protocol provided was followed. The sensitivity of the assay was 5 pmol/ml. The intraassay variation was 5.2%, and the interassay variation was 7.5%. PGE2 was measured by RIA using a kit provided by Dr. Robert Abayasekara (Royal Veterinary College, London, UK) (22). The limit of detection of the assay was 2 pg/ml PGE2, and intra- and interassay coefficients of variation were 3.5% and 6.7%, respectively. Aldosterone was measured using an in-house assay (23). IL-6 was measured using a specific enzyme immunoassay kit purchased from IDS (Tyne & Wear, UK).
mRNA analysis
RNA was extracted from H295R cells using a QuickPrep Micro mRNA Purification Kit (Amersham Biosciences). The RNA obtained was treated with deoxyribonuclease to ensure that no genomic DNA was present, then subjected to first strand DNA synthesis using the protocol in the kit. The samples were heat inactivated (90 C for 5 min) to denature any RNA-cDNA duplex that had formed and also to inactivate the reverse transcriptase.
Five microliters (equivalent to 10 ng total RNA) were subjected to PCR in a 50-μl reaction volume containing 37.5 μl diethylpyrocarbonate (DEPC)-water, 5 μl 10x PCR buffer, 1 μl deoxy-NTP, 0.5 μl sense primer, 0.5 μl antisense primer, and 0.5 μl Taq polymerase under the following conditions: one cycle of denaturation at 94 C for 5 min and 35 cycles of denaturation at 94 C for 1 min, 1-min primer annealing at the calculated temperature (see Table 1), and 1-min primer extension at 72 C. Ten microliters of the PCR products were electrophoresed through 2% agarose gels, stained with ethidium bromide, and viewed by UV light. The primers used are listed in Table 1. The use of these primers for Toll-2 and Toll-4 has previously been reported (24). Glyceraldehyde-3-phosphate dehydrogenase was used to test the purity of the cDNA. The positive control was mRNA obtained from THP-1 cells, which express all members of the TLR family of receptors (25), and the negative control was DEPC (0.1%, vol/vol)-water. PCR products were sequenced to confirm their identity.
TABLE 1. Primers used for PCR of TLR2 and TLR4 (22 )
Analysis of data
Data were analyzed using ANOVA, followed by Tukey’s comparison test (PRISM, GraphPad, Inc., San Diego, CA).
Results
Endotoxin caused a significant increase in cortisol secretion by H295R cells (Fig. 1). The minimum effective concentration was 5 ng/ml, and a maximal effect was seen at 10 ng/ml. Endotoxin had no effect on aldosterone secretion at any concentration used (data not shown). Analysis of the mRNA species present showed that the genes encoding both TLR2 and TLR4 were expressed in adrenal cells (Fig. 2). The specific TLR ligands, lipid A, and purified LPS, which are specific for TLR4, and Pam3Cys, which is specific for TLR2, all caused significant stimulation of cortisol release (Fig. 3). However, these agents caused an approximately 4-fold increase in cortisol over the basal level, whereas crude LPS consistently caused a greater increase over the basal level.
FIG. 1. Effects of varying concentrations (1 ng/ml to 1 μg/ml) of crude LPS on cortisol release by adrenal H295R cells over a 24-h incubation period. Data shown are the mean ± SD (n = 4). *, P < 0.05; ***, P < 0.001 (compared with control values, by ANOVA).
FIG. 2. Analysis of mRNA from NCI-H295R adrenal cells, showing the expression of TLR2 (A) and TLR4 (B). The positive control was mRNA obtained from THP-1 cells, and the negative control was DEPC (0.1%, vol/vol)-water.
FIG. 3. Effects of lipid A (1 μg/ml), Pam3Cys (Pam; 5 μg/ml), and purified LPS (1 ng/ml) on cortisol release from H295R cells over a 24-h incubation period. ***, P < 0.001 (compared with control values, by ANOVA).
The response to endotoxin was slower in onset than the response to forskolin (Fig. 4). With forskolin stimulation, a significant increase in cortisol was seen after 4 h, but was seen only after 8 h with endotoxin. Endotoxin also caused an increase in PGE2 release from adrenal cells, although forskolin had no effect (Fig. 5). The increased secretion of PGE2 in response to LPS temporally preceded the increase in cortisol. IL-6 release was below the sensitivity of the assay.
FIG. 4. Time course of cortisol release by adrenal H295R cells. A, Control incubation; B, effects of 10 μmol/liter forskolin; C, effects of 10 ng/ml crude LPS. Data shown are the mean ± SD (n = 4). *, P < 0.05; **, P < 0.01; ***, P < 0.001 (compared with control values for that time point, by ANOVA).
FIG. 5. Time course of PGE2 release by adrenal H295R cells. A, Control incubation; B, effects of 10 μmol/liter forskolin; C, effects of 10 ng/ml crude LPS. Data shown are the mean ± SD (n = 4). ***, P < 0.001 (compared with control values for that time point, by ANOVA).
The cortisol response to endotoxin was inhibited by indomethacin, whereas both basal cortisol secretion and the response to forskolin were unaffected (Fig. 6A). The use of more specific COX-1 and COX-2 inhibitors showed that COX-2 inhibitor, NS398, significantly attenuated the cortisol response to LPS, without affecting either basal or forskolin-stimulated secretion (Fig. 6B). The COX-1 inhibitor, SC560, had no significant effect on basal or stimulated cortisol release (Fig. 6C).
FIG. 6. Effects of different COX inhibitors on the cortisol responses of adrenal H295R cells to crude LPS and forskolin (10 μmol/liter). A, Effects of indomethacin (10 μmol/liter); B, effects of NS398 (5 μmol/liter), a selective COX-2 inhibitor; C, effects of SC560 (0.1 μmol/liter), a selective COX-1 inhibitor. Data shown are the mean ± SD (n = 4). *, P < 0.05; **, P < 0.01; ***, P < 0.001 (compared with control values). ++, P < 0.01; +++, P < 0.001 (compared with LPS alone).
Discussion
The results of this study confirm the previous finding of the expression of TLRs in human adrenocortical cells (14). Binding of LPS to rat adrenal cells has previously been reported, but the receptor was not further characterized (13). A consensus has emerged that TLR-4 is the major receptor for bacterial LPS (9, 10), and its expression in adrenal cells suggests that the adrenal cortex may be a novel target for the actions of endotoxin. The data presented confirm our earlier finding that bacterial endotoxin exerts a direct stimulatory effect on cortisol secretion, but not on aldosterone secretion, by human adrenocortical cells (8). Although some previous studies have demonstrated an effect of endotoxin on corticosterone secretion by the rat adrenal in hypophysectomized rats in vivo (4), others have failed to show an effect in rodents (6, 7). It is not clear from in vivo studies whether the effect seen was a direct action of LPS on the adrenocortical cells. The rat adrenal gland contains numerous mast cells, and the degranulation of these mast cells results in significant stimulation of adrenal steroidogenesis (26). When an adrenal cell line is used, no such confounding factors are present.
It is known that endotoxin can act through a variety of signaling molecules, including the TLRs. This study has demonstrated the expression of genes encoding TLR2 and TLR4, and the use of specific ligands for these receptors provides indirect evidence to suggest that both receptor types may be functional in these cells. However, the discrepancy in magnitude of response between the effects of the specific ligands and the effects of crude LPS suggest that this response may be mediated by either a different receptor or an additive or cooperative action of the two receptor subtypes. This is the subject of an ongoing study in our laboratory.
The dose-dependency of the response to LPS was very similar to that previously reported in other cell types (15), with a threshold of 5 ng/ml. The maximally effective concentration in the present study was found to be 10 ng/ml, although the effects of higher concentrations were not investigated in the study by Wang and co-workers (15). The dose-response characteristics of the specific TLR ligands were not investigated in the present study, although it was established that Pam3Cys was toxic to these cells at concentrations higher than that used in this study (5 μg/ml; data not shown).
The time course of the response to LPS was noteworthy. Forskolin was used to stimulate the cAMP-dependent pathway, because these cells are relatively unresponsive to ACTH (17). The cortisol response to forskolin was evident at 4 h, but there was no increase in cortisol in response to LPS until 8 h. Previous studies using human subjects in vivo have shown a two-stage cortisol response to LPS administration (27). The first part of the response was seen about 3–4 h after LPS administration and coincided with the peak ACTH response. Both cortisol and ACTH returned to basal levels by 8 h, and cortisol, but not ACTH, then increased again, with a peak about 11–12 h after LPS administration. In light of the present results, it seems likely that this second phase of the cortisol response may be due to a direct action of LPS on the adrenal cortex.
Endotoxin, but not forskolin, significantly stimulated PG synthesis, and a comparison of the time course of response revealed that the increase in PG release preceded the increase in cortisol secretion. The use of inhibitors of PG biosynthesis suggested that formation of PGs is an obligatory step in the adrenal response to LPS stimulation. Previous studies have implicated adrenal COX-2 in the hypothalamo-pituitary-adrenal response to endotoxin, because treatment of intact rats with endotoxin caused an increase in adrenal COX-2 expression (16). However, the finding that this effect was blocked by dexamethasone suggests that the change in adrenal COX-2 may have been a result of activation of the whole axis, rather than a direct effect of endotoxin on the adrenal gland. The results of the present study suggest that LPS may act directly on the human adrenal cortex to stimulate COX-2.
In other tissues, LPS acts to stimulate cytokine production. In the adrenal gland, production of PGE2 appears to be simply an intermediary in the signal transduction pathway leading to increased cortisol biosynthesis. It has previously been suggested that IL-6 may have a role in the rodent adrenocortical response to LPS, because IL-6 knockout mice had an attenuated glucocorticoid response (28). However, IL-6 release in this study was below the assay detection limit; thus, it was not possible to determine whether LPS altered adrenal IL-6 secretion. It is possible that the adrenal effect of LPS is a late part of the physiological response to infection, and in this regard it is probably significant that the adrenal response to LPS stimulation is much slower than the response to forskolin.
In conclusion, these studies demonstrate a direct stimulatory effect of bacterial endotoxin on cortisol secretion by human adrenal cells. This provides additional evidence for immune-adrenal interactions and may be a significant part of the physiological response to infection.
References
Beishuizen A, Thijs LG 2003 Endotoxin and the hypothalamo-pituitary-adrenal (HPA) axis. J Endotoxin Res 9:3–24
Elenkov IJ, Kovacs K, Kiss J, Bertok L, Vizi ES 1992 Lipopolysaccharide is able to bypass corticotropin-releasing factor in affecting plasma ACTH and corticosterone levels: evidence from rats with lesions of the paraventicular nucleus. J Endocrinol 133:231–236
Schotanus K, Makara GB, Tilders FJ, Berkenbosch F 1994 ACTH response to a low dose but not a high dose of bacterial endotoxin in rats is completely mediated by corticotropin releasing hormone. Neuroimmunomodulation 1:300–307
Suzuki S, Oh C, Nakano K 1986 Pituitary-dependent and -independent secretion of CS caused by bacterial endotoxin in rats. Am J Physiol 250:E470–E474
Enrique de Salamanca A, Garcia R 2003 Rat glomerulosa cells in primary culture and E. coli lipopolysaccharide action. J Steroid Biochem Mol Biol 85:81–88
Yasuda N, Greer MA 1978 Evidence that the hypothalamus mediates endotoxin stimulation of adrenocorticotrophic hormone secretion. Endocrinology 102:947–953
Spinedi E, Suescun MO, Hadid R, Daneva T, Gaillard RC 1992 Effects of gonadectomy and sex hormone therapy on the endotoxin-stimulated hypothalamo-pituitary adrenal axis: evidence for a neuroendocrine-immunological sexual dimorphism. Endocrinology 131:2430–2436
Vakharia K, Renshaw D, Hinson JP 2002 Bacterial lipopolysaccharide directly stimulates cortisol secretion in human adrenal cells. Endocr Res 28:357–361
Janssens S, Beyaert R 2003 Role of Toll-like receptors in pathogen recognition. Clin Microbiol Rev 16:637–646
Beutler B, Hoebe K, Du X, Ulevitch RJ 2003 How we detect microbes and respond to them: the Toll-like receptors and their transducers. J Leukocyte Biol 74:479–485
Murphy TJ, Paterson HM, Mannick JA, Lederer JA 2004 Injury, sepsis and the regulation of Toll-like receptor responses. J Leukocyte Biol 75:400–407
Palsson-McDermott EM, O’Neill LAJ 2004 Signal transduction by the lipopolysaccharide receptor, Toll-like receptor 4. Immunology 113:153–162
Enrique de Salamanca A, Portoles MT, Garcia R 2000 Binding of Eschericia coli lipopolysaccharide to fasciculata-reticularis and glomerulosa cells evaluated by flow cytometry. J Cell Biochem 7:386–394
Bornstein SR, Schumann RR, Rettori V, McCann SM, Zacharowski K 2004 Toll-like receptor 2 and Toll-like receptor 4 expression in human adrenals. Horm Metab Res 36:470–475
Wang T, Qin L, Liu B, Liu Y, Wilson B, Eling TE, Langenbach R, Taniura S, Hong J-S 2004 Role of reactive oxygen species in LPS-induced production of prostaglandin E2 in microglia. J Neurochem 88:939–947
Cover PO, Slater D, Buckingham JC 2001 Expression of cyclooxygenase enzymes in rat hypothalamo-pituitary-adrenal axis: effects of endotoxin and glucocorticoids. Endocrine 16:123–131
Rainey WE, Bird IM, Mason JI 1994 The NCI-H295R cell line: a pluripotent model for human adrenocortical function. Mol Cell Endocrinol 100:45–50
Paik Y-N, Schwabe RF, Bataller R, Russo MP, Jobin C, Brenner DA 2003 Toll-like receptor 4 mediates inflammatory signalling by bacterial lipopolysaccharide in human hepatic stellate cells. Hepatology 37:1043–1055
Redecke V, Hacker H, Datta SK, Fermin A, Pitha PM, Broide DH, Raz E 2004 Cutting edge: activation of Toll-like receptor 2 induces a Th2 immune response and promotes experimental asthma. J Immunol 172:2739–2743
Smith CJ, Zhang Y, Koboldt CM, Muhammad J, Zweifel BS, Shaffer A, Talley JJ, Masfarrer JL, Seibert K, Isakson PC 1998 Pharmacological analysis of cyclo-oxygenase-1 in inflammation. Proc Natl Acad Sci USA 27:13313–13318
Vane JR, Bakhle YS, Botting RM 1998 Cyclo-oxygenases 1 and 2. Annu Rev Pharmacol Toxicol 38:97–120
Cheng Z, Elmes M, Kirkup SE, Abayasekara DR, Wathes DC 2004 Alteration of prostaglandin production and agonist responsiveness by n-6 polyunsaturated fatty acids in endometrial cells from late-gestation ewes. J Endocrinol 182:249–256
Kapas S, Orford CD, Barker S, Vinson GP and Hinson JP 1992 Studies on the intracellular mechanism of action of -melanocyte-stimulating hormone (-MSH) on rat adrenal zona glomerulosa. J Mol Endocrinol 9:47–54
Naik S, Kelly EJ, Meijer L, Pettersson S, Sanderson IR 2001 Absence of Toll-like receptor 4 explains endotoxin hyporesponsiveness in human intestinal epithelium. J Paediatr Gastroenterol Nutr 32:449–453
Zarember KA, Godowsky PJ 2002 Tissue expression of human Toll-like receptors and differential regulation of Toll-like receptor mRNAs in leukocytes in response to microbes, their products, and cytokines. J Immunol 168:554–561
Hinson JP, Vinson GP, Pudney J, Whitehouse BJ 1989 Adrenal mast cells modulate vascular and secretory responses in the intact adrenal gland of the rat. J Endocrinol 121:253–260
Vedder H, Schreiber W, Yassouridis A, Gudewill S, Galanos C and Pollmacher T 1999 Dose-dependence of bacterial lipopolysaccharide (LPS) effects on peak response and time course of the immune-endocrine host response in humans. Inflamm Res 48:67–74
Van Enckevort FH, Sweep CG, Span PN, Demacker PN, Hermsen CC, Hermus AR 2001 Reduced adrenal response to bacterial lipopolysaccharide in interleukin-6 deficient mice. J Endocrinol Invest 24:786–795(K. Vakharia and J. P. Hin)
Address all correspondence and requests for reprints to: Dr. J. P. Hinson, Department of Endocrinology, Barts and the London, Queen Mary School of Medicine and Dentistry, Suite 12, Dominion House, Bartholomew Close, London EC1A 7BE, United Kingdom. E-mail: j.p.hinson@qmul.ac.uk.
Abstract
Activation of the hypothalamo-pituitary-adrenal axis by bacterial lipopolysaccharide (LPS; endotoxin) is well documented, although there has been uncertainty about whether LPS exerts a direct effect at the level of the adrenal. The present study found that LPS caused a dose-dependent stimulation of basal cortisol secretion by the human adrenocortical cell line, NCI-H295R, without affecting aldosterone. The expression of both Toll-like receptor 2 (TLR2) and TLR4 was demonstrated in these cells, and the specific ligands for TLR4 (purified LPS and lipid A) and TLR2 (Pam3Cys) were found to stimulate cortisol release, suggesting that these receptors may mediate the effects of LPS in adrenal cells, as has been shown in other cell types. LPS was also found to stimulate prostaglandin E2 release by these cells. The effects of LPS on cortisol were attenuated in the presence of both indomethacin and a specific COX-2 inhibitor, but not a COX-1 inhibitor, suggesting an obligatory role for COX-2 activation and prostaglandin synthesis in the adrenal response to LPS.
Introduction
ACTIVATION OF THE hypothalamo-pituitary-adrenal (HPA) axis is an important part of the stress response. It is known that there are multiple examples of cross-talk between the immune and endocrine systems, and in acute infection this is vital to maintain homeostasis. It is well established that bacterial endotoxin, a lipopolysaccharide (LPS) component of the bacterial cell wall, is a potent activator of the HPA axis (1). Although previous studies have shown that LPS exerts this effect principally by stimulating CRH secretion (1), there is also evidence for CRH-independent effects (2, 3). It appears that LPS may act on each tissue comprising the HPA axis, because there is also evidence for a pituitary-independent effect of LPS on the adrenal gland, with LPS causing increased corticosterone secretion in both intact and hypophysectomized rats (4). More recently, it has been shown that LPS inhibits ACTH-stimulated corticosterone secretion by cultured rat zona glomerulosa cells, although the effect on basal secretion was not determined (5). There are conflicting results in rodents, however, because it has been shown that the presence of the hypothalamus is necessary for a corticosterone response to LPS in rats (6). An earlier study also reported that LPS did not have a direct adrenal effect in mice in vivo (7). Preliminary studies from our laboratory have shown that LPS acts directly on human adrenal cells to stimulate cortisol secretion (8).
LPS from Gram-negative bacteria (Escherichia coli LPS) is known to act though Toll-like receptors (TLRs), specifically through TLR4 (9). The TLRs are a receptor family, related to the IL-1 receptor, with multiple ligands and a range of signal transduction pathways (for review, see Refs. 9, 10, 11). Binding of LPS to TLRs is not simple. It appears that a range of other molecules is required for TLRs to recognize the LPS signal. These include CD14, MD-2, and LPS-binding protein, which are all important in the presentation of LPS and its recognition by the TLR (12). Recently, binding of LPS to adrenal cells has been reported (13), and the expression of TLRs has been described in the human adrenal (14).
It is clear that the TLRs signal through a range of different pathways to stimulate the release of cytokines, particularly members of the IL family (10). LPS action has also been linked to prostaglandin (PG) production in some tissues (15), and there is evidence that in intact rats, LPS administration causes up-regulation of COX-2 in the adrenal gland (16). It is not clear whether this is a direct action of LPS on the adrenal, however.
A previous study from our laboratory indicated that LPS exerted a direct action on human adrenocortical cells (8). The present study was designed to extend these findings, determine the dose-response relationship, and elucidate the time course and mechanism of this effect. These studies used the human adrenocortical cell line, NCI-H295R, which produces the normal range of human adrenal steroids (17). It has the advantage of being a pure adrenocortical cell preparation with no contaminating endothelial, medullary, or mast cells that could potentially influence the response to LPS.
Materials and Methods
The NCI-H295R cell line was a gift from Ian Mason (University of Edinburgh, Edinburgh, UK). All tissue culture medium and supplements were obtained from Invitrogen Life Technologies, Inc. (Paisley, UK), with the exception of +1 ITS medium supplement (Universal Biologicals, Gloucester, UK) and Ultroser SF (Biosepra, Cergy-Saint-Christophe, France). All other reagents were obtained from Sigma-Aldrich Corp. (Poole, UK) and were of molecular grade unless otherwise stated.
Crude LPS (E. coli 0127:B8) was obtained from Sigma-Aldrich Corp. A highly purified preparation of LPS (E. coli 0127:B8) was also purchased from Sigma-Aldrich Corp. (prepared by phenolic extraction and gel filtration chromatography; protein content, <1%; RNA, <1%). This is referred to as pure LPS. A stock solution of LPS dissolved in water was aliquoted and stored at –20 C. Lipid A (from E. coli) was dissolved in dimethylsulfoxide (1 mg/ml). Pam3Cys (EMC Microcollections, Tubingen, Germany) was dissolved in water (1 mg/ml). Indomethacin was dissolved in ethanol (10 mg/ml), and forskolin, SC560, and NS398 were each dissolved in dimethylsulfoxide (5 mg/ml). Stock solutions were diluted with cell culture medium to give the final required concentration. Control experiments were carried out with each of the solvents alone at the final concentration in the incubation. None had any effect on cortisol release.
Cell culture
H295R cells were routinely maintained in 75-cm2 tissue culture flasks in DMEM/Ham’s F-12 Nutrient Mix supplemented with 2% (wt/vol) Ultroser SF, 1% ITS (6.25 mg insulin, 6.25 mg transferrin, 6.25 mg selenium, and 5.35 mg linoleic acid) and 1% penicillin/streptomycin at 37 C under an atmosphere of 95% air/5% carbon dioxide. For all experiments, preincubations and incubations were carried out in serum-free medium which consisted of DMEM/Ham’s F-12 Nutrient Mix supplemented only with 1% penicillin/streptomycin. Cells were passaged 1:2 using a solution of trypsin (0.5 g/liter)/EDTA (0.2 g/liter) every 4 d.
Cell incubations
H295R cells were plated to a density of 500,000 cells/well in six-well plates and were serum-starved for 24 h before treatment. Cells were incubated in serum-free medium in the presence or absence of LPS for different time periods. The effects of the specific TLR agonists were investigated by incubating cells in the presence or absence of these agents for 24 h. Lipid A was used at 1 μg/ml, and Pam3Cys was used at 5 μg/ml. At these concentrations, lipid A is a specific TLR4 agonist (18), and Pam3Cys is a specific TLR2 agonist (19).
To investigate the effects of the cyclooxygenase (COX) inhibitors, indomethacin, SC 560, and NS398, experiments were carried out as described above, using forskolin (10–5 mol/liter) as a positive control. Cells were incubated for 24 h with either LPS or forskolin in the presence or absence of the inhibitor. Indomethacin was used at a concentration of 10–5 mol/liter, SC560 at 10–7 mol/liter, and NS398 at 5 x 10–5 mol/liter. At these concentrations, SC560 is a selective inhibitor of COX-1 (20), and NS398 is a selective inhibitor of COX-2 (21).
At the end of the incubation period, medium was removed from the cells and stored at –20 C until assayed. Cell viability was determined using a 3-[4,5-dimethlthiazol-2-yl]-2,5-diphenyltetrazolium bromide assay (Promega Corp., Madison, WI). None of the treatments used in this study had any effect on cell viability.
Assays
Cortisol was measured by RIA, using an antibody obtained from Bioclin (Cardiff, UK); labeled cortisol was purchased from Amersham Biosciences (Little Chalfont, UK), and the protocol provided was followed. The sensitivity of the assay was 5 pmol/ml. The intraassay variation was 5.2%, and the interassay variation was 7.5%. PGE2 was measured by RIA using a kit provided by Dr. Robert Abayasekara (Royal Veterinary College, London, UK) (22). The limit of detection of the assay was 2 pg/ml PGE2, and intra- and interassay coefficients of variation were 3.5% and 6.7%, respectively. Aldosterone was measured using an in-house assay (23). IL-6 was measured using a specific enzyme immunoassay kit purchased from IDS (Tyne & Wear, UK).
mRNA analysis
RNA was extracted from H295R cells using a QuickPrep Micro mRNA Purification Kit (Amersham Biosciences). The RNA obtained was treated with deoxyribonuclease to ensure that no genomic DNA was present, then subjected to first strand DNA synthesis using the protocol in the kit. The samples were heat inactivated (90 C for 5 min) to denature any RNA-cDNA duplex that had formed and also to inactivate the reverse transcriptase.
Five microliters (equivalent to 10 ng total RNA) were subjected to PCR in a 50-μl reaction volume containing 37.5 μl diethylpyrocarbonate (DEPC)-water, 5 μl 10x PCR buffer, 1 μl deoxy-NTP, 0.5 μl sense primer, 0.5 μl antisense primer, and 0.5 μl Taq polymerase under the following conditions: one cycle of denaturation at 94 C for 5 min and 35 cycles of denaturation at 94 C for 1 min, 1-min primer annealing at the calculated temperature (see Table 1), and 1-min primer extension at 72 C. Ten microliters of the PCR products were electrophoresed through 2% agarose gels, stained with ethidium bromide, and viewed by UV light. The primers used are listed in Table 1. The use of these primers for Toll-2 and Toll-4 has previously been reported (24). Glyceraldehyde-3-phosphate dehydrogenase was used to test the purity of the cDNA. The positive control was mRNA obtained from THP-1 cells, which express all members of the TLR family of receptors (25), and the negative control was DEPC (0.1%, vol/vol)-water. PCR products were sequenced to confirm their identity.
TABLE 1. Primers used for PCR of TLR2 and TLR4 (22 )
Analysis of data
Data were analyzed using ANOVA, followed by Tukey’s comparison test (PRISM, GraphPad, Inc., San Diego, CA).
Results
Endotoxin caused a significant increase in cortisol secretion by H295R cells (Fig. 1). The minimum effective concentration was 5 ng/ml, and a maximal effect was seen at 10 ng/ml. Endotoxin had no effect on aldosterone secretion at any concentration used (data not shown). Analysis of the mRNA species present showed that the genes encoding both TLR2 and TLR4 were expressed in adrenal cells (Fig. 2). The specific TLR ligands, lipid A, and purified LPS, which are specific for TLR4, and Pam3Cys, which is specific for TLR2, all caused significant stimulation of cortisol release (Fig. 3). However, these agents caused an approximately 4-fold increase in cortisol over the basal level, whereas crude LPS consistently caused a greater increase over the basal level.
FIG. 1. Effects of varying concentrations (1 ng/ml to 1 μg/ml) of crude LPS on cortisol release by adrenal H295R cells over a 24-h incubation period. Data shown are the mean ± SD (n = 4). *, P < 0.05; ***, P < 0.001 (compared with control values, by ANOVA).
FIG. 2. Analysis of mRNA from NCI-H295R adrenal cells, showing the expression of TLR2 (A) and TLR4 (B). The positive control was mRNA obtained from THP-1 cells, and the negative control was DEPC (0.1%, vol/vol)-water.
FIG. 3. Effects of lipid A (1 μg/ml), Pam3Cys (Pam; 5 μg/ml), and purified LPS (1 ng/ml) on cortisol release from H295R cells over a 24-h incubation period. ***, P < 0.001 (compared with control values, by ANOVA).
The response to endotoxin was slower in onset than the response to forskolin (Fig. 4). With forskolin stimulation, a significant increase in cortisol was seen after 4 h, but was seen only after 8 h with endotoxin. Endotoxin also caused an increase in PGE2 release from adrenal cells, although forskolin had no effect (Fig. 5). The increased secretion of PGE2 in response to LPS temporally preceded the increase in cortisol. IL-6 release was below the sensitivity of the assay.
FIG. 4. Time course of cortisol release by adrenal H295R cells. A, Control incubation; B, effects of 10 μmol/liter forskolin; C, effects of 10 ng/ml crude LPS. Data shown are the mean ± SD (n = 4). *, P < 0.05; **, P < 0.01; ***, P < 0.001 (compared with control values for that time point, by ANOVA).
FIG. 5. Time course of PGE2 release by adrenal H295R cells. A, Control incubation; B, effects of 10 μmol/liter forskolin; C, effects of 10 ng/ml crude LPS. Data shown are the mean ± SD (n = 4). ***, P < 0.001 (compared with control values for that time point, by ANOVA).
The cortisol response to endotoxin was inhibited by indomethacin, whereas both basal cortisol secretion and the response to forskolin were unaffected (Fig. 6A). The use of more specific COX-1 and COX-2 inhibitors showed that COX-2 inhibitor, NS398, significantly attenuated the cortisol response to LPS, without affecting either basal or forskolin-stimulated secretion (Fig. 6B). The COX-1 inhibitor, SC560, had no significant effect on basal or stimulated cortisol release (Fig. 6C).
FIG. 6. Effects of different COX inhibitors on the cortisol responses of adrenal H295R cells to crude LPS and forskolin (10 μmol/liter). A, Effects of indomethacin (10 μmol/liter); B, effects of NS398 (5 μmol/liter), a selective COX-2 inhibitor; C, effects of SC560 (0.1 μmol/liter), a selective COX-1 inhibitor. Data shown are the mean ± SD (n = 4). *, P < 0.05; **, P < 0.01; ***, P < 0.001 (compared with control values). ++, P < 0.01; +++, P < 0.001 (compared with LPS alone).
Discussion
The results of this study confirm the previous finding of the expression of TLRs in human adrenocortical cells (14). Binding of LPS to rat adrenal cells has previously been reported, but the receptor was not further characterized (13). A consensus has emerged that TLR-4 is the major receptor for bacterial LPS (9, 10), and its expression in adrenal cells suggests that the adrenal cortex may be a novel target for the actions of endotoxin. The data presented confirm our earlier finding that bacterial endotoxin exerts a direct stimulatory effect on cortisol secretion, but not on aldosterone secretion, by human adrenocortical cells (8). Although some previous studies have demonstrated an effect of endotoxin on corticosterone secretion by the rat adrenal in hypophysectomized rats in vivo (4), others have failed to show an effect in rodents (6, 7). It is not clear from in vivo studies whether the effect seen was a direct action of LPS on the adrenocortical cells. The rat adrenal gland contains numerous mast cells, and the degranulation of these mast cells results in significant stimulation of adrenal steroidogenesis (26). When an adrenal cell line is used, no such confounding factors are present.
It is known that endotoxin can act through a variety of signaling molecules, including the TLRs. This study has demonstrated the expression of genes encoding TLR2 and TLR4, and the use of specific ligands for these receptors provides indirect evidence to suggest that both receptor types may be functional in these cells. However, the discrepancy in magnitude of response between the effects of the specific ligands and the effects of crude LPS suggest that this response may be mediated by either a different receptor or an additive or cooperative action of the two receptor subtypes. This is the subject of an ongoing study in our laboratory.
The dose-dependency of the response to LPS was very similar to that previously reported in other cell types (15), with a threshold of 5 ng/ml. The maximally effective concentration in the present study was found to be 10 ng/ml, although the effects of higher concentrations were not investigated in the study by Wang and co-workers (15). The dose-response characteristics of the specific TLR ligands were not investigated in the present study, although it was established that Pam3Cys was toxic to these cells at concentrations higher than that used in this study (5 μg/ml; data not shown).
The time course of the response to LPS was noteworthy. Forskolin was used to stimulate the cAMP-dependent pathway, because these cells are relatively unresponsive to ACTH (17). The cortisol response to forskolin was evident at 4 h, but there was no increase in cortisol in response to LPS until 8 h. Previous studies using human subjects in vivo have shown a two-stage cortisol response to LPS administration (27). The first part of the response was seen about 3–4 h after LPS administration and coincided with the peak ACTH response. Both cortisol and ACTH returned to basal levels by 8 h, and cortisol, but not ACTH, then increased again, with a peak about 11–12 h after LPS administration. In light of the present results, it seems likely that this second phase of the cortisol response may be due to a direct action of LPS on the adrenal cortex.
Endotoxin, but not forskolin, significantly stimulated PG synthesis, and a comparison of the time course of response revealed that the increase in PG release preceded the increase in cortisol secretion. The use of inhibitors of PG biosynthesis suggested that formation of PGs is an obligatory step in the adrenal response to LPS stimulation. Previous studies have implicated adrenal COX-2 in the hypothalamo-pituitary-adrenal response to endotoxin, because treatment of intact rats with endotoxin caused an increase in adrenal COX-2 expression (16). However, the finding that this effect was blocked by dexamethasone suggests that the change in adrenal COX-2 may have been a result of activation of the whole axis, rather than a direct effect of endotoxin on the adrenal gland. The results of the present study suggest that LPS may act directly on the human adrenal cortex to stimulate COX-2.
In other tissues, LPS acts to stimulate cytokine production. In the adrenal gland, production of PGE2 appears to be simply an intermediary in the signal transduction pathway leading to increased cortisol biosynthesis. It has previously been suggested that IL-6 may have a role in the rodent adrenocortical response to LPS, because IL-6 knockout mice had an attenuated glucocorticoid response (28). However, IL-6 release in this study was below the assay detection limit; thus, it was not possible to determine whether LPS altered adrenal IL-6 secretion. It is possible that the adrenal effect of LPS is a late part of the physiological response to infection, and in this regard it is probably significant that the adrenal response to LPS stimulation is much slower than the response to forskolin.
In conclusion, these studies demonstrate a direct stimulatory effect of bacterial endotoxin on cortisol secretion by human adrenal cells. This provides additional evidence for immune-adrenal interactions and may be a significant part of the physiological response to infection.
References
Beishuizen A, Thijs LG 2003 Endotoxin and the hypothalamo-pituitary-adrenal (HPA) axis. J Endotoxin Res 9:3–24
Elenkov IJ, Kovacs K, Kiss J, Bertok L, Vizi ES 1992 Lipopolysaccharide is able to bypass corticotropin-releasing factor in affecting plasma ACTH and corticosterone levels: evidence from rats with lesions of the paraventicular nucleus. J Endocrinol 133:231–236
Schotanus K, Makara GB, Tilders FJ, Berkenbosch F 1994 ACTH response to a low dose but not a high dose of bacterial endotoxin in rats is completely mediated by corticotropin releasing hormone. Neuroimmunomodulation 1:300–307
Suzuki S, Oh C, Nakano K 1986 Pituitary-dependent and -independent secretion of CS caused by bacterial endotoxin in rats. Am J Physiol 250:E470–E474
Enrique de Salamanca A, Garcia R 2003 Rat glomerulosa cells in primary culture and E. coli lipopolysaccharide action. J Steroid Biochem Mol Biol 85:81–88
Yasuda N, Greer MA 1978 Evidence that the hypothalamus mediates endotoxin stimulation of adrenocorticotrophic hormone secretion. Endocrinology 102:947–953
Spinedi E, Suescun MO, Hadid R, Daneva T, Gaillard RC 1992 Effects of gonadectomy and sex hormone therapy on the endotoxin-stimulated hypothalamo-pituitary adrenal axis: evidence for a neuroendocrine-immunological sexual dimorphism. Endocrinology 131:2430–2436
Vakharia K, Renshaw D, Hinson JP 2002 Bacterial lipopolysaccharide directly stimulates cortisol secretion in human adrenal cells. Endocr Res 28:357–361
Janssens S, Beyaert R 2003 Role of Toll-like receptors in pathogen recognition. Clin Microbiol Rev 16:637–646
Beutler B, Hoebe K, Du X, Ulevitch RJ 2003 How we detect microbes and respond to them: the Toll-like receptors and their transducers. J Leukocyte Biol 74:479–485
Murphy TJ, Paterson HM, Mannick JA, Lederer JA 2004 Injury, sepsis and the regulation of Toll-like receptor responses. J Leukocyte Biol 75:400–407
Palsson-McDermott EM, O’Neill LAJ 2004 Signal transduction by the lipopolysaccharide receptor, Toll-like receptor 4. Immunology 113:153–162
Enrique de Salamanca A, Portoles MT, Garcia R 2000 Binding of Eschericia coli lipopolysaccharide to fasciculata-reticularis and glomerulosa cells evaluated by flow cytometry. J Cell Biochem 7:386–394
Bornstein SR, Schumann RR, Rettori V, McCann SM, Zacharowski K 2004 Toll-like receptor 2 and Toll-like receptor 4 expression in human adrenals. Horm Metab Res 36:470–475
Wang T, Qin L, Liu B, Liu Y, Wilson B, Eling TE, Langenbach R, Taniura S, Hong J-S 2004 Role of reactive oxygen species in LPS-induced production of prostaglandin E2 in microglia. J Neurochem 88:939–947
Cover PO, Slater D, Buckingham JC 2001 Expression of cyclooxygenase enzymes in rat hypothalamo-pituitary-adrenal axis: effects of endotoxin and glucocorticoids. Endocrine 16:123–131
Rainey WE, Bird IM, Mason JI 1994 The NCI-H295R cell line: a pluripotent model for human adrenocortical function. Mol Cell Endocrinol 100:45–50
Paik Y-N, Schwabe RF, Bataller R, Russo MP, Jobin C, Brenner DA 2003 Toll-like receptor 4 mediates inflammatory signalling by bacterial lipopolysaccharide in human hepatic stellate cells. Hepatology 37:1043–1055
Redecke V, Hacker H, Datta SK, Fermin A, Pitha PM, Broide DH, Raz E 2004 Cutting edge: activation of Toll-like receptor 2 induces a Th2 immune response and promotes experimental asthma. J Immunol 172:2739–2743
Smith CJ, Zhang Y, Koboldt CM, Muhammad J, Zweifel BS, Shaffer A, Talley JJ, Masfarrer JL, Seibert K, Isakson PC 1998 Pharmacological analysis of cyclo-oxygenase-1 in inflammation. Proc Natl Acad Sci USA 27:13313–13318
Vane JR, Bakhle YS, Botting RM 1998 Cyclo-oxygenases 1 and 2. Annu Rev Pharmacol Toxicol 38:97–120
Cheng Z, Elmes M, Kirkup SE, Abayasekara DR, Wathes DC 2004 Alteration of prostaglandin production and agonist responsiveness by n-6 polyunsaturated fatty acids in endometrial cells from late-gestation ewes. J Endocrinol 182:249–256
Kapas S, Orford CD, Barker S, Vinson GP and Hinson JP 1992 Studies on the intracellular mechanism of action of -melanocyte-stimulating hormone (-MSH) on rat adrenal zona glomerulosa. J Mol Endocrinol 9:47–54
Naik S, Kelly EJ, Meijer L, Pettersson S, Sanderson IR 2001 Absence of Toll-like receptor 4 explains endotoxin hyporesponsiveness in human intestinal epithelium. J Paediatr Gastroenterol Nutr 32:449–453
Zarember KA, Godowsky PJ 2002 Tissue expression of human Toll-like receptors and differential regulation of Toll-like receptor mRNAs in leukocytes in response to microbes, their products, and cytokines. J Immunol 168:554–561
Hinson JP, Vinson GP, Pudney J, Whitehouse BJ 1989 Adrenal mast cells modulate vascular and secretory responses in the intact adrenal gland of the rat. J Endocrinol 121:253–260
Vedder H, Schreiber W, Yassouridis A, Gudewill S, Galanos C and Pollmacher T 1999 Dose-dependence of bacterial lipopolysaccharide (LPS) effects on peak response and time course of the immune-endocrine host response in humans. Inflamm Res 48:67–74
Van Enckevort FH, Sweep CG, Span PN, Demacker PN, Hermsen CC, Hermus AR 2001 Reduced adrenal response to bacterial lipopolysaccharide in interleukin-6 deficient mice. J Endocrinol Invest 24:786–795(K. Vakharia and J. P. Hin)