Somatostatin regulation by cytokines

BRAIN SOMATOSTATIN Somatostatin

Regulation

by Cytokines

David E. Scarborough Inflammatory states are associatedwith nervous and neuroendocrineresponses, which appear to be mediated through the actions of cytokines. Since endotoxin treatment in the rat is associatedwith declines in thyrotropin (TSH) secretion and growth hormone (GH) secretion, changes that may be explained by stimulation of hy~alamic somatostatin (SRIF), the effects of cytokines on SRIF were examined. in an in vitro model system consisting of fetal rat diencephalic ceils interleukin-1 (K-1). tumor necrosis factor (TNF) and interleukin-6 (R-6) were found to stimulate the synthesis and release of SRIF. This effect developed sfowly over 24 hours and wes dose- and time-dependent. Acute release of SRIF over periods up to 1 hour was not found. The mechanism of cytokine stimulation of SRIF is not known. Since the depletion of glial cells in the cultures inhibits the effect, mediators that depend on the presence glia may be involved. The ability of cytokines to stimulate brain SRIF is likely to prove relevant to our understanding in many araas. including brain development, brain responses to injury, and neuroendocrine changes in chronic illness. Q 1990 by w.3. &~r&rS cOI?Zp84f.

R MANY YEARS immunoregulatory molecules, now termed cytokines, were known principally by the acronyms derived from their activities in bioassay systems. Interleukin-1 (IL-l), for example, was known as EP (endogenous pyrogen) or LAF (lymph~~e-activating factor). Recombinant DNA techniques have allowed us to consolidate many of these bioactivities into single molecules, and have reduced the number of acronyms with which we must contend. They have also led to the realization that cytokines are not confined to the immune system. Rather, cytokines are produced in and act on nearly all body tissues, including those of the nervous and endocrine systems. More broadly, cytokines such as IL-I and tumor necrosis factor (TNF) are understood to be members of a growing family of multifunctional peptide growth factors, a family that includes the various tissue growth factors and the hematopoietic factors. Cytokines are implicated in nervous and neuroendoerine function in many ways.’ In the setting of infection or sepsis central nervous system (CNS) responses occur including fever, anorexia, and sleepiness. IL- 1 binding and receptors are found in brain tissue,2 and most available evidence supports the view that CNS responses to sepsis reflect the effects of IL-1 and other cytokines acting on the brain3 Some of these CNS responses may involve cytokine regulation of s‘omatostatin (SRIF). Anorexia has been correlated with elevated hypothalamic SRIF levels.4 Lin et al recently reported that SRIF may mediate endotoxin fever.* In their study, depletion of hypothalamic SRIF with cysteamine, or blockade of SRIF action with an antagonist, were able to inhibit endotoxin fever. Characteristic neuroendocrine responses to stress and infection include the suppression of thyroid stimulating horF

From the Section of ~n~~r~nofo~~, Louisiana State University School o~Med~cine, Shreveport~ LA. Supported in part by a grantfrom the Edward P. Stiles Trust and by an American Cancer Society Junior Investigator Award. Address reprint requests to David E. Scarborough, MD. Section of Endocrinology, LSU Medical Center, PO Box 33932, Shreveport, LA 71130. 0 1990 by W.B. Saunders Company. ~24-049~/90/3909-2029~03.00/0 108

mone (TSH) secretion, the stimulation or suppression of growth hormone (GH) secretion (depending on species), variable effects on prolactin secretion, and the stimulation of ACTH secretion. Increasing evidence implicates cytokines in these responses. IL-1 stimulates the hypothalamic~pituitary-adrenal axis, perhaps at multiple levels, including the stimulation of hypothalamic corticotropin releasing factor (CRF) release.6 IL- 1 and TNF both suppress thyroid function, an effect also mediated at more than one level of the hypoth~amic-pitui~-~yroid axis, and one which likely involves the suppression of hy~t~lamic th~otropin-releasing hormone (TRH) secretion. 7.8 Intracerebroventricular administration of IL-l9 or y-interIeron’0 results in suppression of TSH secretion. This effect could be mediated by a number of mechanisms, including suppression of TRH or stimulation of SRIF release. It is noteworthy that SRIF neurons recently have been found to make special ~nnections with hypothalamic TRH neurons,’ ’ suggesting that some cytokine effects on TRH might be mediated indirectly through SRIF. The usual response of GH secretion to stress or endotoxin in rats is suppression,‘2 an effect that could be mediated by various pathways, including increased SRIF release. In favor of this latter ~ssibility is the observation that the declines in TSH and GH levels that follow endotoxin administration are attenuated by pretreatment with SRIF antiserum.‘* On the other hand, SRIF release from the median eminence (sampled by push-pull cannula) is not increased following endotoxin administration.13 Localized infusion of IL-I in the vicinity of growth hormone releasing factor (GRF) neurons in the arcuate nucleus suppresses GH secretion.‘4 SRIF has been reported to inhibit hypothalamic GRF,” and thus might mediate even this effect. Cancer patients being treated with TNF have elevated levels of prolactin, cortisol. and ACTH, but GH levels are not elevated,r6 suggesting a blunting of the usual stress-related rise in GH. This latter observation could also be explained by TNF stimulation of SRIF release. Al1 the effects described up to this point can be attributed to cytokines circulating in the blood and produced either by leukocytes or other tissues extrinsic to the CNS. Whether circulating cytokines can gain access to the CNS in sufficient amounts to cause such responses is an open question. The similar fevers produced by peripheral venous and local brain Mefabolism, Vol39, No 9, SuppI

(September),

1990: pp 108-l 11

SOWlATOSTATlN

REGULATION

109

BY CYTOKINES

injection of pyrogenic cytokines suggest, but do not prove, that this occurs. Moreover, cytokines have now been found to be intrinsic to the brain. IL-16 has been visualized, by immunocytochemistry, in neurons of the postmortem human h~~~arn~” and in the rat brain and pituitary.‘* A similar network of TNF-~sitive neurons is seen in the mouse brain.” Further, IL-1 appears to be a product of neuroglia. IL 1 alpha and beta mRNA are detected in cultures of microglial cells stimulated with endotoxin,2’ and bioassayable IL- 1 TNF activities have been linked to both microglia and astroglia.2’*22 Increased glia-associated cytokine production occurs during late gestation in the rat bminz3 and is also associated with brain injury.24 Many cytokines and growth factors have been detected in the brain or have been found to be trophic or mitogenic for brain cells. These include acidic and basic fibroblast growth factors (FGFs), IL-l, TNF, platelet-derived growth factor (PDGF), IL-3, and colony-stimulating factors. Given the potential involvement of SRIF in the nervous and neuroendocrine responses to cytokine stimulation, and in view of the extensive evidence that cytokines are intrinsic to the brain and hypothalamus, the study of cytokine effects on brain SRIF takes on additional importance. In vitro study of SRIF regulation using primary cultures of fetal rat brain is a well~stablish~ technique.2s Using this model, we have shown that IL-l@ markedly stimulates the synthesis and release of immunoreactive SRIF peptide from rat diencephalic cells in vitro.26 This effect is obtained with concentrations of IL-1 as low as lo-” mol/L and is dosedependent up to a maximally effective dose of approximately 10m9mol/L (Fig 1). This effect is detectable at as early as 24 hours and is accompanied by increases in levels of SRIF mRNA, as measured by Northern blot analysis of cell extracts. Although increased levels of SRIF peptide are found in both culture media and cell extracts, most of the increase in the culture content of the peptide derives from increases in the

Control

12

11 -

10

9

8

7

Log [ hrIL-1 ]

Fig 1, Fetal rat diencephalic cells were exposed to human recombinant IL-18 for 5 days. Cells were fad on the sixth day in vitro with media supplemented with R-1, again on day 8. and ware harvested for SRIF RIA on day 11. In this experiment the threshold dose of 0.01 nmol/L is evident. Also, the maximal afkctive dose is seen to be 1 to 10 nmol/L (no statisticallysignificant diirence was observed over the range of 1 to 100 nmol/L). Pversus control wells: # = ,017; * i ,001. There were three to five wells for each condllon: bars = SEM.

Table 1. Somatostatin Stimulation by hrTNFa and hrfL-lg Culture No.

Control

1

1,105

2

795

3

1,511

TNF

IL-l

TNF t IL-1

1,579

1,924

2,174

1,192

2,045

2,618

1.819

4,125

5,788

NOTE. THe results of exposing primarY cultures of fetal rat diencephalon to TNFa and IL-l&

alone and in combination, are shown. Total SRIF

content is expressed as picograms per well. Cytokine exposures lasted 4 to 7 days. TNF was tested at 10” mol/L and IL- 1 at 10-k mol/L (culture 1) or IO-’ mol/L (cultures 2 and 3). The relative magnitude of changes varied with the timing and duration of cytokine treatment and was most pronounced when several days elapsed batween the last media change and harvest (culture 3). For each culture the differences were significant (each value equals the mean of six wells; P for comparisons between each treatment condition and any other condition were all 1.02).

extracellular fraction. We have also evaluated the effect of acute exposures to human recombinant (hr)I~l~ on the release of SRIF. No significant increase in release was detected during incubations up to 4 hours in length using IL-l/3 at concentrations between lo-‘* mol/L and lo-* mol/L. Recently we have extended these studies to include TNF and IL-la. TNF also stimulates the accumulation of SRIF in primary cultures of fetal rat brain2’ The threshold dose of hrTNF varies with the particular preparation, but overall is similar to that of IL- 1. IL- 1B appears somewhat more potent on a molar basis than TNF and the combination of the two is markedly synergistic (Table 1). IL- 1cyhas limited sequence homology to IL-l@, but appears to act at the same receptor in most systems. In our model hrIL- 1(Yshows the same pattern of SRIF stimulation as that previously found for ILlP.*’ In contrast to TNF, the combination of IL- 18 and IL1a: at high doses is neither additive nor synergistic, suggesting that they may act at the same receptor in this system as well. IL-6 is a cytokine that is induced by both IL 1 and TNF and also shares many of their actions. Preliminary data from our laboratory suggests that IL-6 is also trophic for SRIF in vitro. Basic fibroblast growth factor, a cytokine that shares bioactivities and sequence homology with IL- 1, has been shown to stimulate SRIF release and to stabilize SRIF mRNA in vitro in a tumor cell line.29 Taken together these results suggest that SRIF stimulation is a property both of IL-l and of a number of biolo~cally related cytokines. The mechamism of cytokine stimulation of SRIF is not known. IL-l stimulates cyclic AMP (CAMP) formation in some systems, and forskolin appears to release SRIF in vitro via a CAMP-dependent mechanism. IL- 1 stimulates prostaglandin (PG) E2 release in cultured astrocytes30 but PGE2 does not simulate h~th~amic SRIF in vitro.” Other phospholipid mechanisms may be involved since IL-1 appears able to stimulate diacyglycerol production through novel pathways. 32 IL-l activates central noradrenergic neurons,33 and norepinephrine stimulates SRIF release in vitro in some studies,25 suggesting another possible mechanism. CRF stimulates SRIF release in vitro,” and thus IL-1 stimulation of CRF could lead secondarily to an increase in SRIF. As noted earlier, a number of the cytokines found to stimulate SRIF are also mitogens for neuroglial cells. The hypothesis that SRIF stimulation is mediated through effects

110

DAVID E. SCARBOROUGH

on the glia therefore deserves serious consideration. Forskolin and CAMP-stimulated SRIF release is attenuated in primary cultures of rat cortex depleted of glia by treatment with cytosine arabinoside (AK+C).‘~ Observations from our laboratory also implicate glial cells in that the IL-1 dose-dependence of glial mitogenesis parallels that for SRIF stimulation.35 In addition, we have observed that Am-C treatment temporarily abrogates the IL-l response (unpublished observations, Jan 1989). The recent report that SRIF is synthesized in cerebellar astrocytes in vitro36 suggests that cytokine-stimulated glia may even secrete SRIF directly. Although most evidence suggests that neurons in primary brain cell cultures from late gestation are largely postmitotic, the possibility remains that cytokines might stimulate the proliferation of SRIF neurons or their precursors. Another pathway for glial stimulation of SRIF secretion would be the cytokine-stimulated release from astrocytes of an excitatory amino acid such as glutamate. Glutamate has been found to stimulate SRIF release in vitro via N-methylD-aSpartate receptors.37 Although some or all of the above pathways may be active, there is as yet no conclusive evidence in favor of any one of them. Definition of the mechanism of IL- 1 SRIF stimulation will require additional studies. Further, despite the apparent similarities among IL-l, TNF, and the other cytokines, their effects on SRIF may prove to be mediated quite differently. The finding of a connection between cytokines and brain SRIF metabolism has relevance in many areas. Regional brain SRIF levels are increased in Huntington’s disease and decreased in Alzheimer’s disease. Rats that have undergone the electrophysiologic treatment known as amygdaloid kindling have increased brain SRIF. In the previously cited study of SRIF and endotoxin fever, elevated hypothalamic SRIF levels were found following endotoxin treatment.’ Do cytokines and SRIF participate in the pathogenesis of febrile seizures? Does the increase in IL-l observed in fetal rat brain during late gestation relate to the embryogenesis of SRIF or other peptidergic neurons? Cytokines are under intense investigation as antitumor agents. On the other hand, given the mitogenicity of cytokines for glial cells and the evidence for glial production of cytokines, cytokines could act as autocrine or paracrine growth

factors for glial tumors. SRIF receptors have been detected in such tumors. Since SRIF is inhibitory to the release of many peptides, it may inhibit the release of autocrine growth factors. Consistent with this hypothesis, SRIF has been reported to inhibit the growth of meningioma cells in vitro. Clearly, the elucidation of interactions between cytokines and brain SRIF should contribute to our understanding of the biology of glial tumors and of the potential CNS side effects of cytokine chemotherapies. As discussed, cytokine stimulation of SRIF may play a significant role in the neuroendocrine response to acute or chronic infection. The “euthyroid sick syndrome” and the stunting of growth in chronically ill juveniles are but two of the clinical circumstances in which this pathway may be involved. If cytokines do modulate brain SRIF in vivo, we may ask what value such a mechanism might have for the organism. The major effects of stimulation of hypothalamic SRIF would be to suppress GH and TSH secretion. Growth hormone has been shown to have permissive and enhancing effects on various immune parameters. On the other hand, clinically significant immune deficiency has not been noted in GH-deficient humans. Even so, more subtle, but useful, changes in the immune response can be imagined owing to modulation of GH secretion. However, any such speculations will have to address the fact that the direction of change is opposite in rats and humans. The utility of inhibition of thyroid function in infectious illness is similarly a problematic issue. Although thyroid hormones doubtless can affect immune parameters, the survival value of thyroid function inhibition has not been proven. It may be argued, as it is for other types of illness, that lowering of thyroid hormone helps conserve metabolic resources. In summary, recent work in the field of immunoregulation has revealed the presence of a new class of neural and neuroendocrine regulatory molecules, the cytokines. Although investigation in this area is just beginning, it is likely that a number of neuropeptides and hormones will prove subject to various forms of cytokine regulation. Somatostatin regulation by cytokines, as reviewed here, is an example of this sort of phenomenon, and we can look forward to many more such findings as work in this field continues.

REFERENCES 1. Scarborough DE, Reichlin S: Cytokines, the brain and aging. Prog NeuroEndocrinlmmunol I : IO-15, 1988 2. Katsuura G, Gottschall PE, Arimura A: Identification of a highaffinity receptor for interleukin-I beta in rat brain. Biochem Biophys Res Commun 156:61-67, 1988 3. Plata-Salamln CR, Oomura Y, Kai Y: Tumor necrosis factor and interleukin-I@ Suppression of food intake by direct action in the central nervous system. Brain Res 448: 106- 114,I988 4. Ho LT, Chem YF, Lin MT: The hypothalamic somatostatinergic pathways mediate feeding behavior in the rat. Experientia 45: 161-162, 1989 5. Lin MT, Uang WN, Ho LT: Hypothalamic somatostatin may mediate endotoxin-induced fever in the rat. Naunyn Schmiedebergs Arch Pharmacol 339:608-6 12, 1989 6. Bateman A, Singh A, Kral T, et al: The immune hypothalamicpituitary-adrenal axis. Endocr Rev l&92- I 12, 1989

7. Dubuis JM, Dayer JM, Siegrist-Kaiser CA, et al: Human recombinant interleukin-16 decreases plasma thyroid hormone and thyroid stimulating hormone levels in rats. Endocrinology 123:2175 2181, 1988 8. Pang X, Hershman JM, Mire11CJ, et al: Impairment of hypothalamic-pituitary-thyroid function in rats treated with human recombinant tumor necrosis factor-alpha (cachectin). Endocrinology 12576-84, 1989 9. Rettori V, Jurcovicova J, McCann SM: Central action on interieukin-1 in altering the release of TSH, growth hormone, and prolactin in the male rat. J Neurosci Res 18:179-183. 1987 10. Gonzalez MC, Riedel M: Effects of human recombinant gamma-interferon on the release of growth hormone, thyroid stimulating hormone and prolactin in the male rat. Program of the 7 1st Annual Meeting of the Endocrine Society, Seattle, WA, 1989, p 63 (abstr no. 164)

SOMATOSTATIN REGULATION BY CYTOKINES

11. Toni R, Jackson IMD, Lechan RM: Somatostatinergic innervation of TRH neurons in the rat hypothalamic pamventricular nucleus: Evidence for unique specialization of contacts with thyroid hormone-responsive cells. Program of 7 1st Annual Meeting of the Endocrine Society, Seattle, WA, 1989, p 338 (abstr no. 1262) 12. Kasting NW, Martin JB: Altered release of growth ho~one and thyrotropin induced by endotoxin in the rat. Am J Physiol243: E332-E337, 1982 13. Fukata J, Kasting NW, Martin JB: Somatostatin release from the median eminence of unanesthetized rats: Lack of correlation with pharmacologically suppressed growth hormone secretion. Neuroendocrinology 40: 193-200, 1985 14. Lumpkin MD, Hartmann DP: R~mbin~t human interleukin-I beta acts within the hypothalamic arcuate nucleus to inhibit pulsatile growth hormone secretion. Program of 71st Annual Meeting of the Endocrine Society, Seattle, WA, 1989, p 220 (abstr no. 789) 15. Tannenbaum GS, McCarthy GF, Reaudet A: Inhibitory role of somatostatin on growth hormone-releasing factor within the hypothalamus. Program of 7 1st Annual Meeting of the Endocrine Society, Seattle, WA, 1989, p 31 (abstr no. 36) 16. Nolten WE, Goldstein D, EhrIich EN, et al: Endocrine changes associated with tumor necrosis factor administration in cancer patients. Program of 7 1st Annual Meeting of the Endocrine Society, Seattle, WA, 1989, p 145 (abstr no. 491) 17. Breder CD, Dinarello CA, Saper CB: Interleukin-I immunoreactive innervation of the human h~othaiamus. Science 240: 321-324, 1988 18. Lechan RM, Toni R, Clark BD, et ah Interleukin-I beta localization in rat brain and pituitary. Program of 7 1st Annual Meeting of the Endocrine Society, Seattle, WA, 1989, p 263 (abstr no. 96 I) 19. Breder CD, Saper CB: Tumor necrosis factor immunoreactive innervation in the mouse brain. Sot Neurosci 141280, 1988 (abstr no. 512) 20. Gebicke-Haerter PJ, Batter J, Schobert A, et al: Lipopolysaccharide-free conditions in primary astrocyte cultures allow growth and isolation of microglial cells. J Neurosci 9: 183-l 94, 1989 21. Hetier E, Ayala J, Denefle P, et al: Brain macrophages synthesize interleukin-I and interleukin-1 mRNAs in vitro. J Neurosci Res 21:391-397, I988 22. Sawada M, Kondo N, Spurns A, et ah Production of tumor necrosis factor-alpha by microglia and astrocytes in culture. 3rain Res 491:394-397, 1989 23. Giulian D, Young DG, Woodward J, et al: Interleukin-I is an astroglial growth factor in the developing brain. J Neurosci 8:709714, 1988

111

24. Giulian D, Lachman LB: Interleukin-I stimulation ofastro8lial proliferation after brain injury. Science 228:497-499, 1985 25. Peterfieund RA, Vale Ww: Somatostatin secretion from the hypothalamus. Adv Exp Med Biol 188:183-200, 1985 26. Scarborough DE, Lee SL, Dinarello CA, et al: fntedeukin-la stimulates ~matos~tin bi~ynth~s in primary cultures of fetal rat brain. Endocrinology 124549-551, 1989 27. Scarborough DE, Dinarello CA: Tumor necrosis factor alpha augments basal and interleukin- 1 beta-stimulated somatostatin synthesis by rat diecephalic cells in vitro. Pro8ram of 7 1st Annual Meeting of the Endocrine Society, Seattle, WA, 1989, p 103 (abstr no. 323) 28. Scarborough DE: Interleukin-1 stimulation of ~matos~tin synthesis in vitro. Pro@am 2nd Intemation~ Pituitary Congress, Palm Springs, CA. 1989, p M24 (abstr) 29. Zeytin FN, Rusk SF, De Lellis R: Growth hormone-releasing factor and fibroblast growth factor regulate somatostatin gene expression. Endocrinology 122: 1 l33- 1136, 1988 30. Hartung H, Schafer B, Heininger K, et ak Recombinant interleukin-I beta stimulates eicosanoid production in rat primary culture astrocytes. Brain Res 489: I l3- 119, 1989 3 I. Ojeda SR, Negro Vilar A, Arimura A, et al: On the hypothalamic mechanism by which prostaglandin E2 stimulates growth hormone release. Neuroendocrinology 3 I: l-7, 1980 32. Kester M, Simonson MS, Mene P, et al: Interleukin-I generates tmnsmembmne signals from phospholipi~ through novel pathways in cultured rat mesanlgial cells. J Clin Invest 83:7 18-723, 1989 33. Dunn AJ: Systemic interleukin-I administration stimulates hypothalamic norepinephtine metabolism parallelling the increased plasma corticosterone. Life Sci 43:429-435, 1988 34. Tapia-Arancibia L, Pares-Herbutt! N, Astier H, et al: Adenylate cyclase activation is not sufficient to stimulate somatostatin reiease from dispersed cerebral cortical and dien~ph~ic cells in #a-free cubures. Brain Res 450:101-I IO, I988 35. Scarborough DE, Lee SL, Reichlin S: Effects of interleukin-1 on somatostatin synthesis in fetal rat brain cell cultures, in Goetzl El (ed): Neuroimmune Networks: Physiology and Diseases. New York, NY, Liss, 1989, pp 83-87 36. Shinoda H, Marini AM, Cosi C, et al: Brain region and gene specificity of neuropeptide gene expression in cultured astrocytes. Science 245415-420, 1989 37. Tapia-Arancibia L, Astier H: Glutamate stimulates somatostatin release from diencephalic neurons in primary culture. Endocrinology 123:2360-2366, 1988