TNF-α increases the frequency of spontaneous miniature synaptic currents in cultured rat hippocampal neurons

BRAIN RESEARCH ELSEVIER

Brain Research 659 (1994) 226-230

Research report

TNF-a increases the frequency of spontaneous miniature synaptic currents in cultured rat hippocampal neurons Francesca Grassi

a,* Anna

Maria Mileo b Lucia Monaco c, Antonello Punturieri ~', Angela Santoni a, Fabrizio Eusebi a,b

a Dipartimento di Medicina Sperimentale, Unit,ersita'di Roma 'La Sapienza', 1-00161 Roma, Italy b Laboratorio di Biofisica, Istituto Regina Elena, via delle Messi d'Oro 156, 1-00158 Roma, Italy c lstituto di Istologia, Uni~,ersita' di Roma 'La Sapienza ', 1-00161 Roma, Italy Accepted 21 June 1994

Abstract

Tumor necrosis factor-a (TNF-a) is a cytokine secreted by activated astrocytes and is known to alter evoked synaptic activity in slices of adult rat hippocampus. In this paper we show that TNF-a increases the frequency of spontaneous miniature synaptic currents in cultured hippocampal neurons, acting at nanomolar concentrations. In addition, we show that the mRNA for the 55 kDa TNF-a receptor (TNF-R1) is detected in embryonic rat hippocampal cultures, as well as in acutely dissected embryonic and adult rat hippocampi. Possible transduction pathways mediating the TNF-a effect are discussed.

Keywords: Tumor necrosis factor-a; TNF receptor; Cultured hippocampal neuron; Miniature synaptic current; Spontaneous synaptic activity

1. Introduction

Tumor necrosis factor-a (TNF-oD is a cytokine secreted during host response to infections by activated macrophages [25] and astrocytes [5,15,25,26], consistent with the idea that the latter are immunocompetent cells in the central nervous system [25] (CNS), Alone or in combination with interleukins (ILs), T N F - a induces astrocyte proliferation [25], as well as synthesis and secretion of nerve growth factor (NGF) from astrocytes and fibroblasts [11,28]. T N F - a also modulates neuronal functions, since it enhances evoked synaptic activity in rat hippocampus [23] and spontaneous transmitter release in rat motoneurons [3]. By contrast, T N F - a inhibits long term potentiation (LTP) of synaptic transmission in hippocampal slices [23]. On a slower time scale (24 h instead of 20-60 min), TNF-a also affects Ca 2÷ currents in cultured sympathetic neurons [21].

* Corresponding author. Laboratorio di Biofisica, Istituto Regina Elena, via delle Messi d'Oro 156, 1-00158 Roma, Italy. Fax: (39) (6) 418 0473. 0006-8993/94/$07.00 © 1994 Elsevier Science B.V. All rights reserved SSD1 0 0 0 6 - 8 9 9 3 ( 9 4 ) 0 0 7 6 7 - 7

Several other cytokines share the neuromodulatory ability of T N F - a [19]. Two high-affinity receptors for TNF-a are known, of 55 kDa (TNF-R1) and 75 kDa (TNF-R2) molecular mass. Recent work has demonstrated that TNF-R1 is coupled to several signal transduction pathways [5]. Evidences for the expression of T N F receptors in brain tissues is that rat astrocytes show specific binding of T N F - a [1], and that cultured rat hippocampal and cortical neurons are immunoreactive with p55 T N F R antibody [4]. In this paper, we attempted a characterization of TNF-a effects on hippocampal spontaneous neurotransmission. We detected, by reverse transcriptasepolymerase chain reaction (RT-PCR), the presence of m R N A for TNF-R1 in the adult and embryonic rat hippocampus, as well as in embryonic hippocampal cultures. Therefore, the effects of T N F - a on miniature post-synaptic activity were studied in cultured hippocampal neurons, a system more accessible than brain slices to drugs, which minimizes possible artifacts caused by incorrect cell perfusion. In addition, optical observations of neuronal cells can be performed. We found that TNF-a increases the frequency of sponta-

F. Grassi et al. / B r a i n Research 659 (1994) 226-230

neous miniature post-synaptic currents (minis), with no effect on cytosolic calcium concentration ([Ca2+]i).

2. Materials and methods 2.1. Cell culture

Primary cultures of hippocampal neurons, prepared from Sprague-Dawley rat embryos at day 18 (El8) [9], were grown in modified Dulbecco's minimum essential medium, supplemented (5% v/v) with a fraction of horse serum, which supports neuronal survival but inhibits glial proliferation [13]. Astrocytes accounted for 5 - 1 0 % of total cellular content, as assayed by staining with an antibody to glial fibrillary acidic protein [1]. 2.2. RT-PCR

Poly(A) + RNA was prepared from hippocampal cultures (day 10-12 in culture) and from rat hippocampi acutely dissected from 2 pregnant females (killed by cervical dislocation) and their embryos (El8), using the FastTrack mRNA isolation kit (Invitrogen, USA), according to the manufacturer's instructions. First-strand cDNA was generated from poly(A) + RNA by the First-strand cDNA synthesis kit (Pharmacia LKB Biotechnology, Sweden) according to the supplied protocol. The completed first strand cDNA reaction product was directly amplified by PCR in the presence of 2.5 U of Taq polymerase (Perkin Elmer, USA) and 500 nM of each of the following amplification primers: (A) 5'-ATGTTCCAGGTGGAGATT-3', corresponding to bases 562-579 on TNF-RI-cDNA [12]; (B) 5'-GTAGACCCTGGGCCTCCA-3', complementary to bases 952-969 on TNF-RI-cDNA [12]; (C) 5'-TGGGACGATCCAATCCTA-3', corresponding to bases 816833 on cDNA for type 1 IL-I receptorl (IL1-R t) (Gene Bank accession number M95578); (D) 5'-GCCAAGTGGTAAGTGTGT-3', complementary to bases 1713-1730 on IL1-RI-cDNA (Gene Bank accession number M95578). Reactions were placed into a DNA Thermal Cycler 9600 (Perkin Elmer). Cycling conditions were as follows: 15 s at 94°C, 20 s at 52°C and 30 s at 72°C for 25 cycles. An aliquot of 10 /xl of each PCR reaction was electrophoresed on 2% agarose gel in 1 x TAE buffer. The gel was stained in 0.5 mg/ml ethidium bromide and photographed. No products were detected when amplifying non reversetranscribed RNA, excluding that results are due to potentially contaminating genomic DNA. 2.3. Electrophysh~logy and Ca 2+ imaging

Experiments (room temperature, 25°C) were performed, at day 10-15 in culture, on preparations with a well-developed connection network. Neurons were bathed in a normal external solution (NES-4) containing (in mM): 140 NaCI, 2.8 KCI, 4 CaCI 2, 2 MgCI 2, 10 glucose, 10 HEPES-NaOH, pH 7.3. In some experiments, Ca 2+ concentration was lowered to 2 mM (NES-2). When indicated, NES was supplemented with (-)-bicuculline methiodide (Sigma, USA), a blocker of GABA A receptor, or with 6-cyano-7-nitroquinoxaline2,3-dione (CNQX; Tocris Neuramin, UK), a blocker of glutamate receptors (non-N-methyl-o-aspartate types). All reagents were of analytical grade. Neurons were continuously perfused by a gravitydriven system, with a constant fluid level in the bath. The cytokinecontaining NES was simultaneously applied through the perfusion system and added to the bath. Control experiments with bovine serum albumin (BSA, 0.1% w/w, 2 cells; Boehringer-Mannheim, Germany) showed that this procedure and the amount of BSA

227

included in cytokine buffers (0.0{)6%) do not affect resting neuronal parameters. Whole cell data, obtained using an Axopatch 200 amplifier (Axon Instruments, USA), were digitized at 2 or 4 kHz (pClamp ver. 5.5, Axon Instruments) and digitally filtered (8-pole Bessel) at 1 or 2 kHz. A semi-automatic function for AutesP (Garching Instrumente, Munich, Germany, kindly provided by H. Zucker) counted the number of peaks within each detected event by a threshold crossing criterion. Mini frequency was calculated by counting the number of peaks in each 1 s epoch. Values were averaged over control periods (10-15 min), and at the plateau of TNF-a treatment. Event peak amplitude and decay constant (r, time to 1/e of peak amplitude) were measured only in isolated minis. Electrophysiological recordings were performed using borosilicate glass electrodes filled with (in raM): 100 CsCI, 18 tetraethylammonium chloride, 0.9 CaCI2, 1.8 MgCI 2, 10 EGTA, 9 HEPES-CsOH, 1.8 Na-ATP; pH 7.3 (3-5 MI2 resistance). Ca 2+ imaging experiments were performed on neurons loaded (90 rain, 37°C) with the cell permeant fluorescent probe Fura 2acetoxymethyl ester (3.5 ~zM; Calbiochem, USA), as previously described [10].

3. Results

3.1. Cytokine receptor expression We investigated the presence of a receptor for TNFa in acutely excised adult and embryonic rat hippocampi, as well as in cultured embryonic hippocampal neurons, focusing on the transducing TNF-R1. Given the small quantities of cellular material available for embryonic preparations, the most sensitive assay, RTPCR, was used. As shown in Fig. 1 (upper panel), a 400 bp fragment was detected when m R N A from adult and embryonic hippocampi, cultured hippocampal neurons and rat heart was amplified with oligonucleotides A and B, corresponding to TNF-R1. No amplification product was detected in the m R N A prepared from the human bladder cancer cell line RT112, which expresses TNF-R1 (A.P. and A.S., in preparation), thus confirming that our oligonucleotides specifically identify rat TNF-R1. As a control, we tested our preparations for the presence of a m R N A for IL1-R t, the expression of which in the hippocampus has been demonstrated [6]. Using oligonucleotides specific for IL1-RI (C and D) a product of 800 bp was amplified from hippocampal neurons and heart tissue, but not from resting T cells (Fig. 1, lower panel). The size of the amplified products was identical to that predicted by the corresponding cDNA. The specificity of these fragments was assessed by southern blot analysis with internal probes (data not shown).

3.2. Effect of TNF-a on spontaneous post-synaptic currents

Minis were recorded from cultured hippocampal neurons, in the presence of tetrodotoxin (TTX, 1 /zM,

228

k (;rassi et al. / Brain Research 659 (1994) 226-230

1

2

3

4

5

..... 408 bp

1

2

3

4

5

..,

847 bp

Fig. 1. RT-PCR analysis of mRNAs for TNF-RI and ILI-R I. Upper panel: expression of TNF-R1 in the following preparations: lane 1, adult rat hippocampus; lane 2, embryonic hippocampus; lane 3, cultured hippocampal neurons; lane 4, rat heart; lane 5 (negative control), human bladder cancer cell line RT112. Lower panel: expression of IL1-R I. Lanes 1-4, as in the upper panel; lane 5 (negative control), Percoll fractionated resting rat T cells (IV fraction). DNA fragments of pBr328 were used as size markers (leftmost lanes). The amplified fragments with the corresponding size are indicated by arrows.

Sigma). Mini frequency, dependent on the extracellular Ca 2+ concentration, remained almost stable for up to 40 min (3 control cells, Fig. 2a). In few neurons, spontaneous events were completely abolished by either CNQX (10 /xM) or bicuculline (10 /xM), but in most cells co-application of the two drugs was required to fully block spontaneous activity (not shown). This suggests that these neurons do not release glycine, in spite of their responsiveness to exogenous glycine [9]. Bicuculline strongly reduced baseline noise (see Fig. 2b) and holding current, in agreement with the described tonic secretion of G A B A [24]. The decay constant of excitatory (CNQX-sensitive) and inhibitory (bicuculline-sensitive) minis followed different gaussian distributions (Fig. 2d) with %x = 5.03 _+ 0.42 ms (n = 14; mean + S.E.M.) and ~'in = 20.2 _+ 0.9 ms (n ----15). At - 5 0 mV holding potential, the amplitude of excitatory and inhibitory minis was Iex = 22.6 + 1.2 pA (n = 14) and Iin = 31.5 _+ 1.8 pA (n = 15), respectively. The effect of human recombinant TNF-a (Boehringer-Mannheim) on mini frequency was initially investigated in neurons bathed in NES-4, discarding cells with basal mini frequency below 0.3 Hz. In each dish, control activity was recorded from one neuron for 10 to 15 min, then T N F - a (10 nM) was applied. Within 2-5

rain, mini frequency rose to a plateau 3.2 ± 1.1 times higher than basal (mean_+ S.E.M. calculated from 6 responsive cells out of 8 cells tested, 6/8; Fig. 2c), with comparable effects on inhibitory and excitatory events. In these 6 cells, the peak current amplitude and the decay rate of the two mini populations became 1~,,,:: 2 1 . 1 ± 2 . 5 pA, r e × = 6 . 4 + 1.0 ms and li,1 = 3 3 . 4 + 4 . 3 pA, ~%=24.1 + 1.5 ms (Fig. 2b,d), not significantly different from control values. Mini frequency did not decrease after a 5-10 min wash (Fig. 2c). Beyond this time, the total recording period became too long to obtain reliable results. A similar effect was elicited when TNF-a concentration was reduced to 3 nM (2.65 _+ 0.57 fold increase, 4 / 7 ) (e.g. Fig. 2b) or to 1 nM (3.5 fold increase, 1/2). In neurons equilibrated in NES-2, mini frequency irreversibly rose 1.70 + 0.17 ( 4 / 6 ) times above control, 1-4 min after treatment with TNF-a 3 nM (not shown). In the absence o f T Y X (n = 2), TNF-a (3 nM) induced a 3-fold increment in the frequency of both the total post-synaptic activity (i.e. including minis) and of the large post-synaptic currents induced by action potentials (not shown). A large inter-culture variability was observed, as TNF-a (at any concentration) raised mini frequency by at least 1.2 fold over basal in m o r e than 80% of the neurons tested in 4 cultures, and never or occasionally in 4 other preparations. For the results reported above, all the neurons tested have been counted, but only responsive cells have been considered for averages. T N F - a (3 or 10 nM) did not significantly change [Ca2+]i in both Fura-2 loaded neurons and glial ceils during the first 15-20 min after cytokine application (not shown). In each neuron (magnification x 1000), the Ca 2+ signal was examined in the soma and in those neuronal processes sufficiently loaded to detect dye fluorescence. In neurons bathed in NES-4 with TTX, basal [Ca2+]i was 39 _+ 2 nM (n = 192). Although TNFcomparably facilitates spontaneous transmitter release at GABAergic and glutamatergic synapses, the effect does not appear to be due to an increased [Ca2+]i.

4. Discussion In this paper we have identified a mRNA for TNFRI in adult and embryonic rat hippocampus, as well as in hippocampal cultures. We have also shown that, in cultured hippocampal neurons, T N F - a increases the frequency of miniature post-synaptic currents, with no effect on current amplitude or decay rate, indicating that T N F - a acts mainly on presynaptic terminals. This effect of TNF-a is not significantly dependent on the extracellular Ca 2+ concentration (2 or 4 mM), and is not mediated by an increase of [Ca2+] i, as shown by fluorescence Ca 2+ imaging.

229

F. Grassiet al. /Brain Research 659 (19941 226-230

a

NES-2

b

NES-4 NES-2

KJJ,'///JA

r..'l/////.q

12 I~

09

~

06

~

O3

+BIC

~

O0

C 12

~

9

-" '-3X p~ S

g" 6 E

~

3

~

o 0

600

1200

1800

0

time (s)

10

20

30

40

50

0

lO

2O

3O

40

50

(ms)

1; (ms)

Fig. 2. Effect of TNF-a on whole-cell miniature post-synaptic currents in cultured hippocampal neurons. Each panel represents data from a different neuron, all clamped at - 5 0 mV. a: mini frequency, oscillating around 0.30 Hz (continuous line) in NES-2 increased to (I.50 Hz when bathing and perfusion solution was changed to NES-4, and returned close to control value (0.26 Hz) upon wash with NES-2. The time course was constructed by averaging the number of minis within 10 s-epochs. Bars describe perfusion protocol, b: bicuculline selectively blocked slowly-decaying inhibitory minis, whereas CNQX abolished fast excitatory events, before (left) and after (right) treatment with TNF-a (3 nM in NES-4). Note the marked reduction of baseline noise in the presence of bicuculline (all traces filtered at 1 kHz). A decreased holding current is masked by the alignment of the traces, c: mini frequency rose within 5 minutes from bath and perfusion application of TNF-a (10 nM). The time course was constructed as in a. This was the second largest increase in mini frequency observed. Note that wash with NES did not decrease mini frequency, d: histograms of mini decay constants (r), before (left) and after TNF-a application (right). Mean values were: 7e× = 4.8 + 0.2 ms (n = 60) and ~'in = 22.3 + 0.6 ms (n = 124) for control; %x = 5.2 _+ 0.2 ms (n = 1101 and "Fin 27.4 + 0.5 ms (n = 350) during the plateau of TNF-a effect. Note that in this as in few other neurons, tin increased by about 20% after TNF-a application. Best fitting curves are superimposed. =

The reported increase of miniature synaptic current frequency is unlikely to be related to the observed inhibition of LTP in adult rat hippocampus [23], as the induction of LTP has been shown to increase mini amplitude rather than frequency [14]. TNF-R1 is coupled to several transduction pathways [27], all capable of inducing cellular responses on a time scale of few minutes. Stimulation of TNF-R1 induces the production of diacylglycerol, which activates protein kinases C (PKC) [22,26,27], but fails to raise [Ca2+]i [22]. Activation of TNF-R1 also stimulates phospholipase A 2 and sphingomyelinase. In addition, ceramide [26,27] and nitric oxide (NO) [25] are produced. At least two of these pathways might cause the observed TNF-a-induced increase of synaptic activity. Stimulation of PKC raises the frequency of miniature endplate currents at the neuromuscular junction [7], and of inhibitory minis in cultured rat hippocampal slices [2]. NO increases the frequency of excitatory minis in newborn rat hippocampal cultures [18]. Alternatively, the TNF-a-induced enhancement of mini frequency might be mediated by events caused by the stimulated secretion of NGF, such as the activation of opiod receptors [20]. However, NGF secretion is reported to be slower than the increase of mini frequency [11,28] (6-12 h vs. 5-10 rain).

Available experimental data do not allow us to discriminate whether TNF-a acts on receptors expressed in neurons, glia or both. In principle, the cytokine could indirectly affect neuronal function through receptors expressed in astrocytes, as astrocytes play an important neuroregulative role [16] and may directly signal to cultured neurons [17]. Alternatively, TNF-R1 expressed in neurons might respond with the stimulation of PKC and the consequent increase of mini frequency to T N F - a secreted by astrocytes, as hypothesized for IL-1 [6]. We observed that TNF-a action on mini frequency is independent of cytokine concentration in the range 1 to 10 nM. This agrees with the reported [22] saturation of T N F - a effects at concentrations beyond 0.5 nM. In conclusion, in the present paper we demonstrate that nanomolar concentrations of TNF-a may facilitate spontaneous transmitter release in cultured hippocampal neurons. We also demonstrate the presence of m R N A for TNF-R1 in all the hippocampal preparations tested. Interestingly, the encoding of TNF-R1 does not appear to be induced by our culture conditions, as it is also observed in mRNAs from ex vivo preparations. The sensitivity of the putative TNF-a receptor appears to be compatible with the binding affinity reported for other brain cells (0.7 nM) [1], since

230

F. Grassiet al. / Brain Research 659 (1994) 226-230

the cytokine acts on neurons at nanomolar concentrations. Such T N F - a concentration is in the range detected in the serum of rats after systemic administration of IL-2 [8], and all this might confer a physiopathological significance to the TNF-a-induced increase of synaptic activity here described.

[12]

[13]

Acknowledgments The work was supported by CNR A C R O grants to F.E. and A.S., and by an AIRC grant to F.E. We thank Drs. Stefano Vicini and Stefano Alema' for critical reading of the manuscript and helpful suggestions, and Dr. Sergio Fucile for performing [Ca2+]i measurements.

References

[14]

[15]

[16]

[17] [18]

[1] Benveniste, E.N., Sparacio, S.M. and Bethea, J.R., Tumor necrosis factor-a enhances interferon-y-mediated class II antigen expression on astrocytes, J. Neuroimmunol., 25 11989) 209219. [2] Capogna, M., Scanziani, M., G~ihwiler, B.H. and Thompson, S.M., Stimulation of protein kinase C increases inhibitory synaptic transmission in the rat hippocampus, Soc. Neurosci. Abstr., 19 (part 2) (1993) 902. [3] Caratsch, C.G., Santoni, A. and Eusebi, F., Interferon-a,/3 and tumor necrosis factor enhance the frequency of miniature and plate potentials at rat neuromuscular junction, Neurosci. Lett., 166 (1994) 97-100. [4] Cheng, B., Christakos, S. and Mattson, M.P., Tumor necrosis factors protect neurons against metabolic-excitotoxic insults and promote maintenance of calcium homeostasis, Neuron, 12 (1994)139-153. [5] Chung, I.Y. and Benveniste, E.N., Tumor necrosis factor-a production by astrocytes, J. Immunol., 144 (1990) 2999-3007. [6] Cunningham, E.T. Jr and De Souza, E.B., Interleukin 1 receptors in the brain and endocrine tissues, Irnmunol. Today, 14 (1993) 171-176. [7] D'Angelo, E., Rossi, P., Tanzi, F. and Taglietti, V. Protein kinase C facilitation of acetylcholine release at the rat neuromuscular junction, Eur. J. Neurosci., 4 (1992) 823-831. [8] Ellison, M.D. and Merchant, R.E., Appearance of cytokine-associated central nervous system myelin damage coincides temporally with serum tumor necrosis factor induction after recombinant inteleukin-2 infusion in rats, J. Neuroimmunol., 33 (1991) 245-251. [9] Grassi, F., CI--mediated interaction between GABA and glycine currents in cultured rat hippocampal neurons, Brain Res., 594 (1992)115-123. [10] Grassi, F., Giovannelli, A., Fucile, S. and Eusebi, F., Activation of the nicotinic acetylcholine receptor mobilizes calcium from caffeine-insensitive stores in C2C12 mouse myotubes, Pfl/ig. Arch., 422 (1993) 591-598. [11] Hattori, A., Tanaka, E., Murase, K., Nobuhiro, 1., Chatani, Y., Tsujimoto, M., I-Iayashi, K. and Kohno, M., Tumor necrosis

[19]

[20]

[21]

[22]

[23]

[24]

[25] [26] [27]

[28]

factor stimulates the synthesis and seclc~ion ~1 biotogicall~ active nerve growth factor in non-neuronal cells. J. Biol. ('hem.. 268 11993) 2577-2582. Himmler, A., Maurer-Fogy. 1.. Kroenke, M., Scheurich. P.. Pfizenmaier, K., Lantz, M., Hauptmann, R., Stratowa. C. and Adolf, G.R., Molecular cloning and expression of human and rat tumor necrosis factor receptor chain (p60) and its soluble derivative, tumor necrosis factor-binding protein, DNA ('ell BioZ, 9 11990) 705-715. Kaufman, L.M. and Barrett, J.M., Serum factor supporting long-term survival of rat central neurones in culture, Science, 220 (1986) 1394-1396. Manabe. T., Renner, P. and Nicoll, R.A., Post-synaptic contribution to long-term potentiation revealed by the analysis of miniature synaptic currents, Nature, 355 (1992)50-55. Morganti-Kossmann, M.C., Kossmann, T. and Wahl, S.M., Cytokines and neuropathology, Trends Pharmacol. Sci., 13 11992) 286-291. Murphy, S., Simmons, M.L., Agullo, L., Garcia, A., Feinstein. D.G., Galea, E., Reis, D.J., Minc-Golomb, D. and Schwartz, J.P., Synthesis of nitric oxide in CNS glial cells, Trends Neurosci., 16 (1993) 323-328. Nedergard, M., Direct signaling from astrocytes to neurons in culture of mammalian brain cells, Science, 263 11994) 1768-1771. O'DelI, T.J., Hawkins, R.D., Kandel, E.R. and Arancio O., Tests of the roles of two diffusible substances in long-term potentiation: evidence for nitric oxide as a possible early retrograde messenger, Proc. Natl. Acad. Sci. USA, 88 (1991) 1128511289. Patterson, P.H. and Nawa, H., Neuronal differentiation factors/cytokines and synaptic plasticity, Cell 72/Neuron. 10 (Suppl.) 11993) 123-137. Shen, K.F. and Crain, S.M., Nerve growth factor rapidly prolongs the action potential of mature sensory ganglion neurons in culture, and this effect requires activation of Gs-coupled excitatory K opioid receptors on these cells, J. Neurosei., (in press). Soliven, B. and Albert, J., Tumor necrosis factor modulates Ca 2+ currents in cultured sympathetic neurons, J. Neurosci., 12 (1992) 2665-2671. Schiitze, S., Berkovic, D., Tomsing, O., Unger, C. and Kr6nke, M., Tumor necrosis factor induces rapid production of l'2'diacylglycerol by a phosphatidylcholine-specific phospholipase C, J. Exp. Med., 174 (1991) 975-988. Tancredi, V., D'Arcangelo, G., Grassi, F., Tarroni, P., Palmieri, G., Santoni, A. and Eusebi, F., Tumor necrosis factor alters synaptic transmission in rat hippocampal slices, Neurosci. Lett., 146 (1992) 176-178. Valeyev, A.Y., Cruciani, R.A., Lange, G.D., Smallwoood, S. and Barker, J.L., CI- channels are randomly activated by continuous GABA secretion in cultured embryonic rat hippocampal neurons, Neurosci. Lett., 155 (1993) 199-203. Vassalli, P., The pathophysiology of tumor necrosis factors, Annu. Rev. Immunol., 10 (1992) 411-452. Vilcek, J., and Lee, T.H., Tumor necrosis factor, Z Biol. Chem., 266 11991) 7313-7316. Wiegmann, K., Schiitze, S., Kampen. E., Himmler, A., Machleidt, T. and Kr6nke, M., Human 55-kDa receptor for tumor necrosis factor coupled to signal transduction cascades, J. Biol. Chem., 267 (1992) 17997-18001. Yoshida, K., Kakihana, M., Chen, L.S., Ong, M., Baird, A. and Gage, F.H., Cytokine regulation of nerve growth factor-mediated cholinergic neurotrophic activity synthesized by astrocytes and fibroblasts, J. Neurochem., 59 (1992) 919-931.