Journal of Neuroimmunology, 41 (1992) 117-120 © 1992 Elsevier Science Publishers B.V. All rights reserved 0165-5728/92/$05.00
117
JNI 02245
Short Communication
Presynaptic inhibitory effect of TNF-a on the release of noradrenaline in isolated median eminence I.J. Elenkov a, K. Kov~cs a E. Duda b E. Stark a and E.S. Vizi a a Institute of Experimental lt,[edicine, Hungarian Academy of Sciences, Budapest, Hungary, and b Institute of Biochemistry, Biological Research Centre, Szeged, Hungary (Received 15 November 1991) (Revision received 20 February 1992) (Accepted 30 April 1992)
Key words." Tumor necrosis factor-a; Median eminence; Noradrenaline release; Presynaptic modulation; Neuro-immune communication
Summary The effect of tumor necrosis factor-a (TNF-a) on the stimulation-evoked release of noradrenaline (NA) from isolated rat median eminence (ME) was investigated, using a low-volume perfusion system. Median eminence, loaded with [3H]noradrenaline, was superfused with Krebs solution and stimulated electrically (2 Hz, 120 shocks). The effect of TNF-a was studied on the $2/S 1 ratio. It was found that stimulation-evoked release of NA from noradrenergic axon terminals in the isolated rat ME was inhibited by TNF-a and this effect was concentration-dependent. In contrast, TNF-a had no effect on the release of [3H]NA from the spleen. Since NA released in the ME might be involved in the modulation of corticotropin-releasing factor (CRF) production, it is suggested that TNF-a, through presynaptic modulation of NA release from noradrenergic nerve terminals in the ME, might regulate CRF and other neurohormone release in this hypothalamic structure.
Introduction It is becoming increasingly apparent that the neuroendocrine and the immune system are engaged in functionally relevant cross-talk (Besedovsky et al., 1983a). Recently, interleukin-1 (IL-1) (Besedovsky et al., 1986; Rivier et al., 1989) and tumor necrosis factor (TNF)-a (Sharp et al., 1989; Bernardini et al., 1990) polypeptides produced during immune response, have been shown to be implicated in the activation of the hypothalamo-pituitary-adrenal axis (HPA). The median eminence, a hypothalamic structure where neurosecretory projections terminate, offers a unique opportunity to study the presynaptic modulation of NA and neurohormone release and their presynaptic interactions because of their well-defined anatomical arrangement: the cell bodies of the neurohormone and NA-contain-
Correspondence to: E.S. Vizi, Institute of Experimental Medicine, Hungarian Academy of Sciences, H-1450 Budapest, P.O. Box 67, Hungary.
ing axons are not in this region (cf. Kordon, 1985; Vizi et al., 1985). Recently, substantial stimulation-evoked and frequency-dependent release of NA from isolated ME was demonstrated (Vizi et al., 1985). The release was not subject to so-called a-2 adrenoreceptor-mediated negative feedback modulation. Therefore it was suggested that NA released in this region may be able to play a tonic modulatory role in hormone release through stimulation of a-2 adrenoreceptors exclusively located on the axon terminals of the hormone-containing neurons (Vizi et al., 1985, 1991). In addition, it has been shown that TNF-a is a potent ACTH secretagogue (Sharp et al., 1989) and its effect can be prevented by CRF-antiserum (Bernardini et al., 1990), suggesting that the effect of TNF on ACTH levels is mediated via CRF production. Since the in vitro release of CRF induced by acetylcholine (Hillhouse et al., 1975) and that induced by electrical stimulation of hypothalamic synaptosomes (Edwardson and Bennett, 1974) is subject to inhibitory modulation by NA, it seemed interesting to study the effect of TNF-a on the
118
S 2 / S 1 =0.91 +_0.04
2.0 ¸
release of NA from the noradrenergic axon terminals in isolated rat ME.
T
0 I
(n=S)
1.5 ¸
"6
Materials and Methods
The release of N A from M E was measured as described by Vizi et al. (1985). M E from male adult Wistar rats was dissected by the technique of NegroVilar (Chiocchio et al., 1976). A low-volume four-channel perfusion system was used (Vizi et al., 1985). In brief, median eminence from four to six brains was pooled in a 100-/xl chamber. After the tissue was loaded with [3H]NA and superfused with Krebs solution (NaC1, 113; KCI, 4.7; CaCI2, 2.5; KH2PO4, 1.2; MgSO4, 1.2; NaHCO3, 25.0; and glucose, 11.5 mM), 3-min fractions were collected. Radioactivity was measured and expressed as fractional release by a computer program. The tissue was stimulated (2 Hz, 120 shocks, supramaximal 3-ms impulse duration) twice, at the beginning of the 3rd and the 13th fractions. The first stimulation (S~) was used as a control and the second (S 2) was performed in presence or absence of drug applied. T N F was added to the perfusion fluid 15 min prior to S 2. The change, decrease or increase of the S2/81 ratio, represents the inhibitory or stimulatory effect of drugs on N A release, respectively. The release of N A from rat spleen strips was measured as previously described (Elenkov and Vizi, 1991). H u m a n recombinant T N F - a produced by Escherichia coli cells was purified to homogeneity, yielding a single electrophoretic band of approximately 17 kDa and with specific activity of 20 X 106 u n i t s / m g , as determined by bioassay based on its cytotoxicity on mouse L929 tumor cells. One unit is defined as the amount of T N F - a required to mediate half-maximal cytotoxicity. Last traces of lipopolysaccharide were removed by affinity chromatography on amphotericin B-agarose. T N F - a was dissolved in phosphate-buffered saline (PBS) containing 0.1% bovine serum albumin (BSA). CH-38083 (7,8-(methylenedioxi)-14a-hydroxyalloberbane • HCI, Chinoin, Budapest, Hungary) was dissolved in Krebs solution. All results show the means + SEM. One-way analysis of variance ( A N O V A ) followed by Dunn's and Dunnett's test was used.
1.0
g 0.5 ¸ o
6" -
"°'o
"0-
o
-0-0-0-0
-S 1 0 . 0 . - -
0
"0-0-0,©_
0
-S 2
,
,
10
,
20
,
,-30
40
50
60
rnin.
Fig. 1. Fractional release of [3H]noradrenaline from the isolated median eminence of the rat. 30 min elapsed between the two stimulations (S1 and $2,2 Hz, 120 shocks and 3-ms impulse duration). Each point with vertical lines represents mean_ S.E.M. in 5 experiments. The $2/S 1 ratio was 0.91+0.04. Average weight of the median eminence was 0.30_+0.02 mg (n = 24). For further details see Materials and Methods.
to field stimulation, 1.31 _+ 0.29% and 1.15 + 0.22% of the radioactivity present in the tissue was released by the first ($1) and the second (S:) stimulation (n = 5). The ratio between the amount of radioactivity released by the two consecutive stimulations ( $ 2 / S ~) was 0.91 + 0.04 (n = 5), (Fig. 1). Since T N F - a was dissolved in PBS with 0.1% BSA, Krebs solution contained the same final concentration of PBS and BSA in control experiments. This solution applied 15 min before S 2 in itself did not significantly change stimulation-evoked release ( S a / S 1 = 0.84 + 001; n = 5) as compared to the the normal Krebs solution. At 50 U / m l concentration, T N F - a provoked a slight but non-significant decrease of stimulation-evoked release ( S e / S 1 = 0.76 + 0.04; n = 4). T N F - a applied at 100 U / m l , 250 U / m i and 1000 U / m l concentrations significantly inhibited the stimulation-evoked release ($2/S l = 0.52 + 0.06, n = 5; 0.59 + 0.05, n = 4; and 0.61 + 0.06, n = 5; P < 0.01, P < 0.01 and P < 0.05 respectively). D a t a expressed as percentage inhibition of [3H]NA are shown in Fig. 2.
50.0
g **
40.0 [
30.0-
"6
20.0-
?,2_= c a~
Results
After 45 min loading with [3H]NA, followed by 30 rain washout, the tissue contained 8.09 + 1.89 × 106 B q / g radioactivity as revealed by measuring [3H] (n = 5). At rest, 0.63 + 0.07% of the [3H] content of the tissue was spontaneously released in 3-min fractions and the rate of outflow remained constant. In response
10.0" 1
0.0-
U/ml 50
1O0
250
1000
Fig. 2. Effect of TNF-a expressed as percentage of inhibition of the stimulation-evoked [3H]noradrenaline release from the median eminence. ** and * indicate significant difference (P < 0.01 and 0.05 respectively). TNF was added to the perfusion fluid 15 min prior to S2. 30 min elapsed between the two stimulations (S1 and Sa; 2 Hz, 120 shocks).
119
The effect of T N F - a on NA release from sympathetic nerve terminals of rat spleen was also studied. T N F - a at a concentration of 1000 U / m l applied 15 min prior to $2 did not significantly change the stimulation-evoked release (2 Hz, 360 shocks) of NA ($2/S 1 = 1.09 + 0.04; n = 3), as compared to the control (S2/S 1 = 1.03 + 0.04; n = 3). Since in this preparation the release of NA is under the tonic control of negative feedback modulation to such an extent that no other agonist is able to further reduce the release unless the a-2 adrenoreceptor-modulated inhibition is suspended (Elenkov and Vizi, 1991), CH-38"083, a highly selective a-2 antagonist, was applied during both stimulations. Even under these conditions, TNF-a (100 U / m l ) had no effect on the stimulation-evoked release ($2/S 1 = 0.88 + 0.02; n = 5 vs. $ 2 / S ~ 0.87 + 0.03; n = 4 in control experiments).
Discussion In the present study direct neurochemical evidence was obtained that T N F - a is able to inhibit the stimulation-evoked release of NA from noradrenergic axon terminals in the isolated ME. Since this preparation contains only the axon terminals of noradrenergic neurons and the cell bodies are in the locus coeruleus and the lateral tegmental system (Olson and Fuxe, 1972; Jones and Moore, 1977), the effect of T N F - a on the release of NA evoked by field, i.e. axonal stimulation is located on the varicose axon terminals. In contrast to the effect of T N F - a on ME, it has no effect on NA release from the spleen. A central noradrenergic neuroinhibitory pathway to CRF release was proposed (Ganong, 1977). Therefore our findings that TNF-a was able to presynaptically control the release of NA in ME and that in vitro the inhibitory action of NA is mediated through a-adrenoreceptors (Hillhouse et al., 1975), may provide further explanation of the mode and site of action of TNF-a. It has been shown that TNF-a, similarly to IL-1, activates H P A (Sharp et al., 1989). While there is an increasing consensus that the main site of action of IL-1 is located in the hypothalamus, through the production of CRF (Berkenbosch et al., 1987; Sapolsky et al., 1987), it is currently under debate whether IL-1 may act directly at the pituitary level (Bernton et al., 1.987) or not (Berkenbosch et al., 1987; Sapolsky et al., 1987). Recently Bernardini et al. (1990) provided direct evidence that CRF is involved in the A C T H response induced by TNF-a. The possible link between the action of T N F - a (and probably other cytokines) on CRF (ACTH) and NA release is substantiated by the findings that decreased NA turnover was observed in the hypothalamus of rats during the peak immune
response (Besedovsky et al., 1983b). Recently in vivo data have been obtained that A C T H response to IL-lfl instilled into ME depends on the availability of catecholamines to be released (Matta et al., 1990). Taking into account our recent results that lesions of hypothalamic paraventricular nucleus did not block the effect of lipopolysaccharide (LPS) on plasma A C T H levels in the late phase of the response (Elenkov et al., 1992), and the observation of Moberg (1971) that lesions of ME totally abolished the effect of LPS, it is conceivable that ME might be a target for the effect of LPS a n d / o r related cytokines. By contrast, Sharp et al. (1989) when they intraparenchymally injected TNF-a above the ME, failed to obtain increased plasma A C T H levels, in spite the apparent response after a single i.v. dose of TNF-a. If TNF-a acts through CRF to increase the A C T H level and if NA exerts tonic inhibitory control on CRF-containing neurons through stimulation of a-2 adrenoreceptors, the disinhibition of this control by TNF-a might result in an increase of CRF release and subsequently an increase of A C T H production. Noradrenaline in the ME might be also involved in the vasoconstriction of the portal Vessels, regulating the supply of CRF to the anterior pituitary (Ganong, 1977). Therefore, the inhibition of NA release resulting in a vasodilatation in the ME by TNF-oe indirectly contributes to increase of supply of CRF. Such an effect of rTNF on vessels was shown by Vicaut et al. (1991) to induce a concentration-dependent vasodilatation in the microcirculation of the rat cremaster muscle. Recently, it was suggested (Blalock, 1989) that the immune system may serve as a sensory organ for stimuli that are not recognized by the central and peripheral nervous system. The recognition of these non-cognitive stimuli (e.g. bacteria, tumors, viruses, antigens, etc.) by immunocytes is then converted into information by means of leukocyte-derived hormones. This type of information processing may then be relayed to the neuroendocrine system by different cytokines via affecting the modulatory noradrenergic system in the ME. ME is a hypothalamic structure not protected by the blood-brain barrier, and since noradrenergic input in the ME is not equipped with a-2 adrenoreceptors, i.e. these noradrenergic axon terminals are without negative feedback control (Vizi et al., 1985), they might be ideally designed to receive these modulatory signals from the periphery and to play an interface role in the bidirectional communication between the immune and the neuroendocrine system. In this respect further studies are required to find out whether other cytokines are also able to affect the release of NA, there are receptors sensitive to them and what is the exact mechanism and the link between the effect of TNF-a and A C T H release.
120
References Berkenbosch, F.B., van Oers, J., Del Rey, A., Tilders, F. and Besedovsky, H.O. (1987) Corticotropin-releasing factor-producing neurons in the rat activated by interleukin-l. Science 238, 524526. Bernardini, R., Kamilaris, T.C., Calogero, A.E., Johnson, E.O., Gomez, M.T., Gold, P.W. and Chrousos, G.P. (1990) Interactions between tumor necrosis factor-a, hypothalamic corticotropin-releasing hormone, and adrenocorticotropin secretion in the rat. Endocrinology 126, 2876-2881. Bernton, E.W., Beach, J.E., Holaday, J.W., Smallridge, R.C. and Fein, H.G. (1987) Release of multiple hormones by a direct action of interleukin-1 on pituitary cells. Science 238, 519-521. Besedovsky, H.O., del Rey, A.E. and Sorkin, E. (1983a) What do the immune system and the brain know about each other? Immunol. Today 4, 342-346. Besedovsky, H.O., del Rey, A.E., Sorkin, E., Da Prada, M., Burri, R. and Honegger, C. (1983b) The immune response evokes changes in brain noradrenergic neurons. Science 221,564-566. Besedovsky, H.O., Del Rey, A., Sorkin, E. and Dinarello, C.A. (1986) Immunoregulatory feedback between interleukin-I and glucocorticoid hormones. Science 233, 652-654. Blalock, J.E. (1989) A molecular basis for bidirectional communication between the immune and neuroendocrine systems. Physiol. Rev. 69, 1-32. Chiocchio, S.R., Negro-Vilar, A. and Tramezzani, J.H. (1976) Acute changes in norepinephrine content in the median eminence induced by orchidectomy or testosterone replacement. Endocrinology 99, 629-635. Edwardson, J.A. and Bennett, G.W. (1974) Modulation of corticotrophin-releasing factor release from hypothalamic synaptosomes. Nature 251,425-427. Elenkov, l.J. and Vizi, E.S. (1991) Presynaptic modulation of release of noradrenaline from the sympathetic nerve terminals in the rat spleen. Neuropharmacology 30, 1319-1324. Elenkov, l.J., Kovfics, K., Kiss, K., Bert6k, L. and Vizi, E.S. (1992) Lipopolysaccharide is able to bypass corticotrophin-releasing factor in affecting plasma ACTH and corticosterone levels: evidence from rats with lesions of the paraventricular nucleus. J. Endocrinol. 133, 231-236.
Ganong, W.F. (1977) Neurotransmitters involved in ACTH secretion: catecholamines. Ann. N.Y. Acad. Sci. 297, 509-517. Hillhouse, E.W., Burden, J. and Jones, M.T. (1975) The effect of various putative neurotransmitters on the release of corticotrophin-releasing hormone from the hypothalamus of the rat in vitro. I. The effect of acetylcholine and noradrenaline. Neuroendocrinology 17, 1-11. Jones, B.E. and Moore, R.Y. (1977) Ascending projections of the locus coeruleus in the rat. II. Autoradiographic study. Brain. Res. 127, 23-53. Kordon, C. (1985) Neural mechanisms involved in pituitary control. Neurochem. Int. 7, 917-925. Matta, S.G., Singh, J., Newton, R. and Sharp, B.M. (1990) The adrenocorticotropin response to interleukin-1/3 instilled into rat median eminence depends on the local release of catecholamines. Endocrinology 127, 2175-2182. Moberg, G.P. (1971) Site of action of endotoxins on hypothalamicpituitary-adrenal axis. Am. J. Physiol. 220, 397-400. Olson, L. and Fuxe, K. (1972) Further mapping out of central noradrenaline neuron systems: Projections of the 'subcoeruleus' area. Brain Res. 43, 289-295. Rivier, C., Vale, W. and Brown, M. (1989) In the rat, interleukin-la stimulates adrenocorticotropin and catecholamine release. Endocrinology 125, 3096-3102. Sapolsky, R., Rivier, C., Yamamoto, G., Plotsky, P. and Vale, W. (1987) Interleukin-1 stimulates the secretion of hypothalamic corticotropin-releasing factor. Science 238, 522-524. Sharp, B.M., Matta, S.G., Peterson, P.K., Newton, R., Chao, C. and Mcallen, K. (1989) Tumor necrosis factor-~ is a potent ACTH secretagogue: comparison to interleukin-1/3. Endocrinology 124, 3131-3133. Vicaut, E., Hou, X., Payen, D., Bousseau, A. and Tedgui, A. (1991) Acute effects of tumor necrosis factor on the microcirculation in the rat cremaster muscle. J. Clin. Invest. 87, 1537-1540. Vizi, E.S., Harsing, L.G. Jr., Zimanyi, I. and Gaal, G. (1985) Release and turnover of noradrenaline in isolated median eminence: lack of negative feedback modulation. Neuroscience 16, 907-916. Vizi, E.S., Kiss, J. and Elenkov, I.J. (1991) Presynaptic modulation of cholinergic and noradrenergic neurotransmission: interaction between them. NIPS 6, 119-123.
117
JNI 02245
Short Communication
Presynaptic inhibitory effect of TNF-a on the release of noradrenaline in isolated median eminence I.J. Elenkov a, K. Kov~cs a E. Duda b E. Stark a and E.S. Vizi a a Institute of Experimental lt,[edicine, Hungarian Academy of Sciences, Budapest, Hungary, and b Institute of Biochemistry, Biological Research Centre, Szeged, Hungary (Received 15 November 1991) (Revision received 20 February 1992) (Accepted 30 April 1992)
Key words." Tumor necrosis factor-a; Median eminence; Noradrenaline release; Presynaptic modulation; Neuro-immune communication
Summary The effect of tumor necrosis factor-a (TNF-a) on the stimulation-evoked release of noradrenaline (NA) from isolated rat median eminence (ME) was investigated, using a low-volume perfusion system. Median eminence, loaded with [3H]noradrenaline, was superfused with Krebs solution and stimulated electrically (2 Hz, 120 shocks). The effect of TNF-a was studied on the $2/S 1 ratio. It was found that stimulation-evoked release of NA from noradrenergic axon terminals in the isolated rat ME was inhibited by TNF-a and this effect was concentration-dependent. In contrast, TNF-a had no effect on the release of [3H]NA from the spleen. Since NA released in the ME might be involved in the modulation of corticotropin-releasing factor (CRF) production, it is suggested that TNF-a, through presynaptic modulation of NA release from noradrenergic nerve terminals in the ME, might regulate CRF and other neurohormone release in this hypothalamic structure.
Introduction It is becoming increasingly apparent that the neuroendocrine and the immune system are engaged in functionally relevant cross-talk (Besedovsky et al., 1983a). Recently, interleukin-1 (IL-1) (Besedovsky et al., 1986; Rivier et al., 1989) and tumor necrosis factor (TNF)-a (Sharp et al., 1989; Bernardini et al., 1990) polypeptides produced during immune response, have been shown to be implicated in the activation of the hypothalamo-pituitary-adrenal axis (HPA). The median eminence, a hypothalamic structure where neurosecretory projections terminate, offers a unique opportunity to study the presynaptic modulation of NA and neurohormone release and their presynaptic interactions because of their well-defined anatomical arrangement: the cell bodies of the neurohormone and NA-contain-
Correspondence to: E.S. Vizi, Institute of Experimental Medicine, Hungarian Academy of Sciences, H-1450 Budapest, P.O. Box 67, Hungary.
ing axons are not in this region (cf. Kordon, 1985; Vizi et al., 1985). Recently, substantial stimulation-evoked and frequency-dependent release of NA from isolated ME was demonstrated (Vizi et al., 1985). The release was not subject to so-called a-2 adrenoreceptor-mediated negative feedback modulation. Therefore it was suggested that NA released in this region may be able to play a tonic modulatory role in hormone release through stimulation of a-2 adrenoreceptors exclusively located on the axon terminals of the hormone-containing neurons (Vizi et al., 1985, 1991). In addition, it has been shown that TNF-a is a potent ACTH secretagogue (Sharp et al., 1989) and its effect can be prevented by CRF-antiserum (Bernardini et al., 1990), suggesting that the effect of TNF on ACTH levels is mediated via CRF production. Since the in vitro release of CRF induced by acetylcholine (Hillhouse et al., 1975) and that induced by electrical stimulation of hypothalamic synaptosomes (Edwardson and Bennett, 1974) is subject to inhibitory modulation by NA, it seemed interesting to study the effect of TNF-a on the
118
S 2 / S 1 =0.91 +_0.04
2.0 ¸
release of NA from the noradrenergic axon terminals in isolated rat ME.
T
0 I
(n=S)
1.5 ¸
"6
Materials and Methods
The release of N A from M E was measured as described by Vizi et al. (1985). M E from male adult Wistar rats was dissected by the technique of NegroVilar (Chiocchio et al., 1976). A low-volume four-channel perfusion system was used (Vizi et al., 1985). In brief, median eminence from four to six brains was pooled in a 100-/xl chamber. After the tissue was loaded with [3H]NA and superfused with Krebs solution (NaC1, 113; KCI, 4.7; CaCI2, 2.5; KH2PO4, 1.2; MgSO4, 1.2; NaHCO3, 25.0; and glucose, 11.5 mM), 3-min fractions were collected. Radioactivity was measured and expressed as fractional release by a computer program. The tissue was stimulated (2 Hz, 120 shocks, supramaximal 3-ms impulse duration) twice, at the beginning of the 3rd and the 13th fractions. The first stimulation (S~) was used as a control and the second (S 2) was performed in presence or absence of drug applied. T N F was added to the perfusion fluid 15 min prior to S 2. The change, decrease or increase of the S2/81 ratio, represents the inhibitory or stimulatory effect of drugs on N A release, respectively. The release of N A from rat spleen strips was measured as previously described (Elenkov and Vizi, 1991). H u m a n recombinant T N F - a produced by Escherichia coli cells was purified to homogeneity, yielding a single electrophoretic band of approximately 17 kDa and with specific activity of 20 X 106 u n i t s / m g , as determined by bioassay based on its cytotoxicity on mouse L929 tumor cells. One unit is defined as the amount of T N F - a required to mediate half-maximal cytotoxicity. Last traces of lipopolysaccharide were removed by affinity chromatography on amphotericin B-agarose. T N F - a was dissolved in phosphate-buffered saline (PBS) containing 0.1% bovine serum albumin (BSA). CH-38083 (7,8-(methylenedioxi)-14a-hydroxyalloberbane • HCI, Chinoin, Budapest, Hungary) was dissolved in Krebs solution. All results show the means + SEM. One-way analysis of variance ( A N O V A ) followed by Dunn's and Dunnett's test was used.
1.0
g 0.5 ¸ o
6" -
"°'o
"0-
o
-0-0-0-0
-S 1 0 . 0 . - -
0
"0-0-0,©_
0
-S 2
,
,
10
,
20
,
,-30
40
50
60
rnin.
Fig. 1. Fractional release of [3H]noradrenaline from the isolated median eminence of the rat. 30 min elapsed between the two stimulations (S1 and $2,2 Hz, 120 shocks and 3-ms impulse duration). Each point with vertical lines represents mean_ S.E.M. in 5 experiments. The $2/S 1 ratio was 0.91+0.04. Average weight of the median eminence was 0.30_+0.02 mg (n = 24). For further details see Materials and Methods.
to field stimulation, 1.31 _+ 0.29% and 1.15 + 0.22% of the radioactivity present in the tissue was released by the first ($1) and the second (S:) stimulation (n = 5). The ratio between the amount of radioactivity released by the two consecutive stimulations ( $ 2 / S ~) was 0.91 + 0.04 (n = 5), (Fig. 1). Since T N F - a was dissolved in PBS with 0.1% BSA, Krebs solution contained the same final concentration of PBS and BSA in control experiments. This solution applied 15 min before S 2 in itself did not significantly change stimulation-evoked release ( S a / S 1 = 0.84 + 001; n = 5) as compared to the the normal Krebs solution. At 50 U / m l concentration, T N F - a provoked a slight but non-significant decrease of stimulation-evoked release ( S e / S 1 = 0.76 + 0.04; n = 4). T N F - a applied at 100 U / m l , 250 U / m i and 1000 U / m l concentrations significantly inhibited the stimulation-evoked release ($2/S l = 0.52 + 0.06, n = 5; 0.59 + 0.05, n = 4; and 0.61 + 0.06, n = 5; P < 0.01, P < 0.01 and P < 0.05 respectively). D a t a expressed as percentage inhibition of [3H]NA are shown in Fig. 2.
50.0
g **
40.0 [
30.0-
"6
20.0-
?,2_= c a~
Results
After 45 min loading with [3H]NA, followed by 30 rain washout, the tissue contained 8.09 + 1.89 × 106 B q / g radioactivity as revealed by measuring [3H] (n = 5). At rest, 0.63 + 0.07% of the [3H] content of the tissue was spontaneously released in 3-min fractions and the rate of outflow remained constant. In response
10.0" 1
0.0-
U/ml 50
1O0
250
1000
Fig. 2. Effect of TNF-a expressed as percentage of inhibition of the stimulation-evoked [3H]noradrenaline release from the median eminence. ** and * indicate significant difference (P < 0.01 and 0.05 respectively). TNF was added to the perfusion fluid 15 min prior to S2. 30 min elapsed between the two stimulations (S1 and Sa; 2 Hz, 120 shocks).
119
The effect of T N F - a on NA release from sympathetic nerve terminals of rat spleen was also studied. T N F - a at a concentration of 1000 U / m l applied 15 min prior to $2 did not significantly change the stimulation-evoked release (2 Hz, 360 shocks) of NA ($2/S 1 = 1.09 + 0.04; n = 3), as compared to the control (S2/S 1 = 1.03 + 0.04; n = 3). Since in this preparation the release of NA is under the tonic control of negative feedback modulation to such an extent that no other agonist is able to further reduce the release unless the a-2 adrenoreceptor-modulated inhibition is suspended (Elenkov and Vizi, 1991), CH-38"083, a highly selective a-2 antagonist, was applied during both stimulations. Even under these conditions, TNF-a (100 U / m l ) had no effect on the stimulation-evoked release ($2/S 1 = 0.88 + 0.02; n = 5 vs. $ 2 / S ~ 0.87 + 0.03; n = 4 in control experiments).
Discussion In the present study direct neurochemical evidence was obtained that T N F - a is able to inhibit the stimulation-evoked release of NA from noradrenergic axon terminals in the isolated ME. Since this preparation contains only the axon terminals of noradrenergic neurons and the cell bodies are in the locus coeruleus and the lateral tegmental system (Olson and Fuxe, 1972; Jones and Moore, 1977), the effect of T N F - a on the release of NA evoked by field, i.e. axonal stimulation is located on the varicose axon terminals. In contrast to the effect of T N F - a on ME, it has no effect on NA release from the spleen. A central noradrenergic neuroinhibitory pathway to CRF release was proposed (Ganong, 1977). Therefore our findings that TNF-a was able to presynaptically control the release of NA in ME and that in vitro the inhibitory action of NA is mediated through a-adrenoreceptors (Hillhouse et al., 1975), may provide further explanation of the mode and site of action of TNF-a. It has been shown that TNF-a, similarly to IL-1, activates H P A (Sharp et al., 1989). While there is an increasing consensus that the main site of action of IL-1 is located in the hypothalamus, through the production of CRF (Berkenbosch et al., 1987; Sapolsky et al., 1987), it is currently under debate whether IL-1 may act directly at the pituitary level (Bernton et al., 1.987) or not (Berkenbosch et al., 1987; Sapolsky et al., 1987). Recently Bernardini et al. (1990) provided direct evidence that CRF is involved in the A C T H response induced by TNF-a. The possible link between the action of T N F - a (and probably other cytokines) on CRF (ACTH) and NA release is substantiated by the findings that decreased NA turnover was observed in the hypothalamus of rats during the peak immune
response (Besedovsky et al., 1983b). Recently in vivo data have been obtained that A C T H response to IL-lfl instilled into ME depends on the availability of catecholamines to be released (Matta et al., 1990). Taking into account our recent results that lesions of hypothalamic paraventricular nucleus did not block the effect of lipopolysaccharide (LPS) on plasma A C T H levels in the late phase of the response (Elenkov et al., 1992), and the observation of Moberg (1971) that lesions of ME totally abolished the effect of LPS, it is conceivable that ME might be a target for the effect of LPS a n d / o r related cytokines. By contrast, Sharp et al. (1989) when they intraparenchymally injected TNF-a above the ME, failed to obtain increased plasma A C T H levels, in spite the apparent response after a single i.v. dose of TNF-a. If TNF-a acts through CRF to increase the A C T H level and if NA exerts tonic inhibitory control on CRF-containing neurons through stimulation of a-2 adrenoreceptors, the disinhibition of this control by TNF-a might result in an increase of CRF release and subsequently an increase of A C T H production. Noradrenaline in the ME might be also involved in the vasoconstriction of the portal Vessels, regulating the supply of CRF to the anterior pituitary (Ganong, 1977). Therefore, the inhibition of NA release resulting in a vasodilatation in the ME by TNF-oe indirectly contributes to increase of supply of CRF. Such an effect of rTNF on vessels was shown by Vicaut et al. (1991) to induce a concentration-dependent vasodilatation in the microcirculation of the rat cremaster muscle. Recently, it was suggested (Blalock, 1989) that the immune system may serve as a sensory organ for stimuli that are not recognized by the central and peripheral nervous system. The recognition of these non-cognitive stimuli (e.g. bacteria, tumors, viruses, antigens, etc.) by immunocytes is then converted into information by means of leukocyte-derived hormones. This type of information processing may then be relayed to the neuroendocrine system by different cytokines via affecting the modulatory noradrenergic system in the ME. ME is a hypothalamic structure not protected by the blood-brain barrier, and since noradrenergic input in the ME is not equipped with a-2 adrenoreceptors, i.e. these noradrenergic axon terminals are without negative feedback control (Vizi et al., 1985), they might be ideally designed to receive these modulatory signals from the periphery and to play an interface role in the bidirectional communication between the immune and the neuroendocrine system. In this respect further studies are required to find out whether other cytokines are also able to affect the release of NA, there are receptors sensitive to them and what is the exact mechanism and the link between the effect of TNF-a and A C T H release.
120
References Berkenbosch, F.B., van Oers, J., Del Rey, A., Tilders, F. and Besedovsky, H.O. (1987) Corticotropin-releasing factor-producing neurons in the rat activated by interleukin-l. Science 238, 524526. Bernardini, R., Kamilaris, T.C., Calogero, A.E., Johnson, E.O., Gomez, M.T., Gold, P.W. and Chrousos, G.P. (1990) Interactions between tumor necrosis factor-a, hypothalamic corticotropin-releasing hormone, and adrenocorticotropin secretion in the rat. Endocrinology 126, 2876-2881. Bernton, E.W., Beach, J.E., Holaday, J.W., Smallridge, R.C. and Fein, H.G. (1987) Release of multiple hormones by a direct action of interleukin-1 on pituitary cells. Science 238, 519-521. Besedovsky, H.O., del Rey, A.E. and Sorkin, E. (1983a) What do the immune system and the brain know about each other? Immunol. Today 4, 342-346. Besedovsky, H.O., del Rey, A.E., Sorkin, E., Da Prada, M., Burri, R. and Honegger, C. (1983b) The immune response evokes changes in brain noradrenergic neurons. Science 221,564-566. Besedovsky, H.O., Del Rey, A., Sorkin, E. and Dinarello, C.A. (1986) Immunoregulatory feedback between interleukin-I and glucocorticoid hormones. Science 233, 652-654. Blalock, J.E. (1989) A molecular basis for bidirectional communication between the immune and neuroendocrine systems. Physiol. Rev. 69, 1-32. Chiocchio, S.R., Negro-Vilar, A. and Tramezzani, J.H. (1976) Acute changes in norepinephrine content in the median eminence induced by orchidectomy or testosterone replacement. Endocrinology 99, 629-635. Edwardson, J.A. and Bennett, G.W. (1974) Modulation of corticotrophin-releasing factor release from hypothalamic synaptosomes. Nature 251,425-427. Elenkov, l.J. and Vizi, E.S. (1991) Presynaptic modulation of release of noradrenaline from the sympathetic nerve terminals in the rat spleen. Neuropharmacology 30, 1319-1324. Elenkov, l.J., Kovfics, K., Kiss, K., Bert6k, L. and Vizi, E.S. (1992) Lipopolysaccharide is able to bypass corticotrophin-releasing factor in affecting plasma ACTH and corticosterone levels: evidence from rats with lesions of the paraventricular nucleus. J. Endocrinol. 133, 231-236.
Ganong, W.F. (1977) Neurotransmitters involved in ACTH secretion: catecholamines. Ann. N.Y. Acad. Sci. 297, 509-517. Hillhouse, E.W., Burden, J. and Jones, M.T. (1975) The effect of various putative neurotransmitters on the release of corticotrophin-releasing hormone from the hypothalamus of the rat in vitro. I. The effect of acetylcholine and noradrenaline. Neuroendocrinology 17, 1-11. Jones, B.E. and Moore, R.Y. (1977) Ascending projections of the locus coeruleus in the rat. II. Autoradiographic study. Brain. Res. 127, 23-53. Kordon, C. (1985) Neural mechanisms involved in pituitary control. Neurochem. Int. 7, 917-925. Matta, S.G., Singh, J., Newton, R. and Sharp, B.M. (1990) The adrenocorticotropin response to interleukin-1/3 instilled into rat median eminence depends on the local release of catecholamines. Endocrinology 127, 2175-2182. Moberg, G.P. (1971) Site of action of endotoxins on hypothalamicpituitary-adrenal axis. Am. J. Physiol. 220, 397-400. Olson, L. and Fuxe, K. (1972) Further mapping out of central noradrenaline neuron systems: Projections of the 'subcoeruleus' area. Brain Res. 43, 289-295. Rivier, C., Vale, W. and Brown, M. (1989) In the rat, interleukin-la stimulates adrenocorticotropin and catecholamine release. Endocrinology 125, 3096-3102. Sapolsky, R., Rivier, C., Yamamoto, G., Plotsky, P. and Vale, W. (1987) Interleukin-1 stimulates the secretion of hypothalamic corticotropin-releasing factor. Science 238, 522-524. Sharp, B.M., Matta, S.G., Peterson, P.K., Newton, R., Chao, C. and Mcallen, K. (1989) Tumor necrosis factor-~ is a potent ACTH secretagogue: comparison to interleukin-1/3. Endocrinology 124, 3131-3133. Vicaut, E., Hou, X., Payen, D., Bousseau, A. and Tedgui, A. (1991) Acute effects of tumor necrosis factor on the microcirculation in the rat cremaster muscle. J. Clin. Invest. 87, 1537-1540. Vizi, E.S., Harsing, L.G. Jr., Zimanyi, I. and Gaal, G. (1985) Release and turnover of noradrenaline in isolated median eminence: lack of negative feedback modulation. Neuroscience 16, 907-916. Vizi, E.S., Kiss, J. and Elenkov, I.J. (1991) Presynaptic modulation of cholinergic and noradrenergic neurotransmission: interaction between them. NIPS 6, 119-123.
