Suppression of splenic macrophage interleukin-1 secretion following intracerebroventricular injection of interleukin-1β: Evidence for pituitary-adrenal and sympathetic control

CELLULAR

IMMUNOLOGY

132,84-93

(1991)

Suppression of Splenic Macrophage Interleukin-I Secretion following lntracerebroventricular Injection of Interleukin-1 ,O: Evidence for Pituitary-Adrenal and Sympathetic Control’ RICHARD

BROWN,*‘*Zuo Lr,* CATHERINEY. VRIEND,* RAMINDER

NIRULA,*

LORENJANZ,~JASON FALK,~ DWIGHT M. NANCE,#Y~DENNISG.DYCK,~ ANDARNOLDH.GREENBERG*'~ *Manitoba Institute qf Cell Biolog.y, 100 Olivia Street, Winnipeg, Manitoba, Canada R3E OV9, tDepartment o/‘Psychology, University ofA4anitoba, Winnipeg, Manitoba, Canada; and *Department qfAnatomy, Dalhousie University, Halt&x, Nova Scotia, Canada Received July 5, 1990; acceptedAugust 13, 1990 Intracerebroventricular (ICV) injections of intedeukin-1 beta (IL-IO) produced a dose-dependent increase in plasma corticosterone and adrenocorticotropic hormone (ACTH) within 2 hr of injection and then declined over the next 24 hr. Using a potent steroidogenic dose of IL-16 (5 ng). ICV injection resulted in suppression of splenic macrophage IL-l secretion following stimulation by LPS in vitro. Macrophage TGF-P secretion was not affected, indicating a differential action of ICV IL-ID on macrophage cytokine production. Following adrenalectomy (ADX), the suppressive effect of ICV IL- 10 was reversed and resulted in stimulation of macrophage IL-l secretion, indicating that the suppression was mediated by adrenocorticol activation. However, surgical interruption of the splenic nerve to eliminate autonomic innervation ofthe spleen also prevented the macrophage suppressive signal in rats given ICV IL-lp. Furthermore, the combination of ADX and splenic nerve section resulted in a potent stimulatory effect of ICV IL-ID on splenic macrophage IL-I secretion which was greater than either ADX or splenic nerve section alone. These results support the concept of a negative feedback on macrophage IL- 1 secretion by the central action of IL- 10 and indicate that both the hypothalamic-pituitary-adrenal axis and the sympathetic nervous system mediate this effect. 8 1991 Academic Press, Inc.

INTRODUCTION

Evidence for a functional link between the immune system and the central nervous system is accumulating. For instance, cytokine receptors on neurons as well as neuropeptide receptorson leukocytes have been detected(1). One molecule which interacts with both the immune and central nervous systems is interleukin-1 (IL-l). IL-1 is ’ This work was supported by grants from the USDHS, The RJR Nabisco Corporation, the MRC of Canada. AHG is a Terry Fox Scientist of the National Cancer Institute of Canada. 2 Current Address: Department of Psychology, University of Newcastle, Shortland, 2308 Australia. 3 Current Address: Department of Pathology, 236 Basic Medical Sciences Bldg., University of Manitoba, 730 William Avenue, Winnipeg, Manitoba, Canada, R3E OW3. 4 To whom correspondence should be addressed at Manitoba Institute of Cell Biology, 100 Olivia Street, Winnipeg, Manitoba, Canada, R3E OV9.

84 ooos-8749/91 Copyright All

tights

0

$3.00

1991 by Academic

of reproduction

Press,

in any form

Inc. reserved.

INTRACEREBROVENTRICULAR

IL-I

AND

lMMUNOSUPPRESSION

85

produced by cells of both the immune system (e.g. monocytes, macrophages) and the nervous system (neurons, astrocytes, microglia) (2, 3). IL-l has broad stimulatory actions on the immune system including lymphocyte activation and initiation of the acute phase response (2, 3) as well as potent actions in the brain such as induction of fever, slow-wave sleep, and hunger satiety (3, 4). Recent findings suggest that IL- 1 is a potent inducer of the hypothalamic-pituitary-adrenal axis (HPA) (5-9) and Besedovsky, Del Rey, and colleagues (5,6) have proposed a negative feedback loop involving the release of IL-l from activated leukocytes which stimulates the HPA, resulting in corticosterone induced suppression of IL- 1 production. Another potential immunoregulatory pathway derives from the sympathetic nervous system. Felten and co-workers ( lo- 12) have demonstrated sympathetic innervation of thymus, spleen, marrow, lymph nodes, and gut-associated lymphatic tissue. In the spleen, noradrenergic and peptidergic fibers have been identified terminating among fields of T-lymphocytes and macrophages. To study the possible contribution of the HPA and sympathetic nervous system to immunoregulation, we administered intracerebroventricular (ICV) doses of IL- 1fi to rats and examined splenic macrophage IL- 1 secretion. Our results support the proposal of Besedovsky and colleagues (5, 6) that IL-1 invokes a negative feedback loop via the adrenals. However, the participation of the sympathetic nervous system was observed, suggesting that a functional interaction between the immune system and the brain was mediated via this pathway. MATERIALS

AND METHODS

Animals Adult (300 g) male Sprague-Dawley rats (Charles River, Quebec) were used in all experiments. Animals were housed individually in polythene cages with food and water ad lib. Lighting was maintained on a 12 hr on and 12 hr off regimen.

Reagents Recombinant human interleukin- lfl was obtained from Drs. D. Urdal and C. Henney (Immunex Corporation). Highly purified porcine transforming growth factor (TGFp,) and anti-TGF-P antibody were purchased from R&D Systems (Minneapolis, MN).

All animals were anaesthetized with Nembutal and stainless steel cannulae were implanted into the left lateral ventricle using the stereotoxic coordinates; A-P = 0.8 mm, L = + 1.3, D-V = -3.0 mm ( 13). Adrenalectomized (ADX) rats were obtained from Charles River and cannulated 10 days following ADX. Splenic nerve section was performed at the same time as cannulation. The spleen was exposed by an abdominal incision lateral to the midline. Using a microscope, the splenic nerve was located before its bifurcation with the blood supply into the hilar region of the spleen. The nerve was then dissected away from the vasculature and at least 1 cm of nerve was removed from the animal. Verification of splenic nerve section was achieved by HPLC determination of splenic norepinephrine (NE) content. In all cases, there was an 80-95s reduction in NE content of the spleen at 4 days postsurgery.

86

BROWN

ET AL.

Four days after cannulation, rats were given an ICV injection of either recombinant human IL- lp at doses of 5, 10, or 100 ng in 5 ~1 saline vehicle. Two hours following injection rats were decapitated and trunk blood and spleens were collected and processed. Half of each spleen was taken for the IL-l assay and the other half rapidly frozen on dry ice and then stored at -80°C until NE analysis. IL-l

Assay

Spleens were disaggregated through stainless steel wire mesh and washed in (UMEM. Red cells were removed by NH&l lysis and remaining cells were washed, resuspended in aMEM, 0.1% BSA, 20 mM Hepes, and plated into 24-well plates (Nunclon) at a concentration of 2 X lo6 splenocytes/ml, 1 ml per well. Following 2 hr incubation at 37°C in 5% CO2 wells were washed to remove nonadherent cells and 1 ml of medium was added containing either 0, 1, 10, or 100 pg LPS (Salmonella typhimurium, Sigma) and the remaining adherent cells (>95% nonspecific esterase positive) were incubated for 48 hr at 37°C in 5% COZ. Supernatants were removed from these cultures, centrifuged, and frozen at -80°C pending bioassay. For the IL- 1 bioassay, the LBRM-33, 1AS-CTLL-2 assay system was employed ( 14). Briefly, LBRM and CTLL2 cells were maintained in RPM1 1640 (GIBCO), 2.3 mM Hepes, 1% nonessential amino acids, 1% sodium pyruvate, 1Op5M 2-mercaptoethanol, and 5% fetal calf serum (GIBCO) were adjusted to lo5 cells per milliliter and 100 ~1 of this dilution was added to wells (96-well flat-bottomed, Nunclon). Standards were prepared at concentrations of 3, 10, 30, 100, and 1000 pg/ml of rh IL- I@. A 20-~1 aliquot of these standards or supernatant from macrophage cultures was added to the wells along with phytohemagglutinin (GIBCO) to a final concentration of l/500 and plates were incubated for 24 hr at 37°C in 5% COZ. After 24 hr plates were centrifuged and 100 ~1 of the supernatant was added to wells containing 5 X IO3 CTLL-2 cells in 100 ~1 medium and incubated a further 24 hr. CTLL-2 cells were routinely maintained in the same medium as LBRM cells with the addition of IL-2. About 4 hr prior to harvest, wells were pulsed with 0.5 &i [3H]thymidine and cells, harvested onto glass fiber paper, and placed into scintillant, and incorporation was determined by counting discs in a Beckman p scintillation counter. IL- I concentration of the supernatant was determined by comparing the [3H]TdR incorporation of the macrophage supernatant samples with the IL-l standards and then converting to pg/ml. Neutralization of IL-l activity was accomplished by preincubating 20 ~1 of macrophage supernatant and 8 ~1 of rabbit anti-human IL- 1 antibody (Genzyme, Boston, MA) or normal rabbit IgG (Teknika, Westchester, PA) overnight at 4°C before adding to the assay. Rabbit anti-IL- 1 antibody was preabsorbed with LBRM cells at a 4: 1 vol: vol ratio at 4°C for 1 hr. This was repeated four times until no LBRM toxicity was detected by trypan blue exclusion, similar to the normal IgG control. TGF-p Assay The method of assay for TGF-P was the CCL-64 mink lung growth inhibition assay as described previously (15). CCL-64 mink lung epithelial cells maintained in DME (GIBCO) with 10% fetal bovine serum were trypsinized, washed with aMEM and resuspended in DME, 0.2% fetal bovine plasma, 10 mM Hepes, penicillin (25 U/ml)

INTRACEREBROVENTRICULAR

IL-I

AND

IMMUNOSUPPRESSION

87

and streptomycin, and plated at a concentration of 5 X lo5 cells per 0.5 ml in 24-well Costar dishes (Flow). Acid-activated supernatant was neutralized and added with or without anti-TGF-P antibody (R & D Systems) 3 hr after cells were initially plated. Following 22 hr incubation, cells were pulsed with 0.25 &i [‘251]IUdR for 2-3 hr at 37°C. Cells were fixed in 1 ml methanol-acetic acid (3) (vol/vol). After 1 hr at room temperature, wells were washed twice with 2 ml of 80% methanol. Cells were lysed in 1 ml of 1 N NaOH for 30 min at room temperature and [‘251]IUdR was counted on an LKB gamma counter. A standard curve of porcine TGF-PI was included and data were expressed as pg TGF-P/ 1Ohsplenocytes.

ACTH and C‘orticosteroneRadioimmunoassa~~ Blood was centrifuged in EDTA and plasma collected and stored at -20°C pending assay. An RIA kit (Nichols Institute Diagnostics) was used for plasma ACTH determinations according to the manufacturer’s instructions. Corticosterone assay followed the method described previously ( 16).

Norepinephrine Assa?, Frozen spleen was weighed and homogenized in 0. I A4 percloric acid containing 0.1 mM EDTA to yield a final tissue concentration of 20 mg per milliliter. Homogenates were centrifuged at 1500 rpm for 10 min and 1.O ml of the supernatant was taken for alumina extraction using the ESA Plasma Catacholamine methodology with an internal standard. Samples were separated by HPLC using an ESA Model 5700 solvent delivery model with a CSC-S ODS2 5-pm column. Analysis and quantitation was accomplished with a Coulochem 5 100 A electrochemical detector and Shimadzu CR60 1 Chromatopat Integrator. Data were expressed as nanogram NE per gram wet tissue weight. RESULTS

Plasma ACTH and Corticosterone Releasefi)llowing ICV Interleukin-l/3 Figures IA and 1B show the effect of 0.5, 5, 10, or 100 ng IL- 10 ICV on plasma ACTH and corticosterone 2 hr postinjection. These doses produced a marked activation of the hypothalamic-pituitary-adrenal axis in agreement with previous studies (6. 7). Plasma ACTH levels returned to control levels by 8 hr while corticosterone levels were still elevated at this time point. Levels of both ACTH and corticosterone were normal by 24 hr (Figs. 2A and 2B).

Mucrophuge IL-I Production

qfirr

ICV Interleukin-l/3

Within 2 hr of ICV injection of 5 ng IL-l@ a marked suppression of macrophage IL-l production in response to LPS stimulation was observed, as well as a reduction in basal levels released by nonstimulated macrophages (Fig. 3A). IL-l values, when expressed as pg/ml/pg macrophage DNA or protein, were identical to those shown in Fig. 3. No variation was detected in the total macrophage recovery from spleens following IL-l ICV compared to the saline control group. A similar suppressive effect on macrophage IL- 1 secretion was seen with a 100 ng dose of IL- 1p ICV (not shown). To assesswhether this was a generalized suppression of macrophage activity, the same

88

BROWN

ET AL.

1400 A

1400

1200

12CUl loo0 =; E i3 LEO0 I 6 4800

g1LMCl E c t”o lz tG 8600 e 0 O400

400

200

200

0

0 0.5

0

5

10

loo

Icv rll-1 I3 (ng)

FIG. 1. Effect of ICV injection of 0.5, 5, IO, 100 ng IL-I/3 or saline vehicle on (A) serum corticosterone and (B) serum ACTH levels 2 hr postinjection. ** = P < 0.00 1, * = P < 0.01 using Student’s t test compared to vehicle injection. n = 5 rats/group.

supernatants were also evaluated for TGF-P activity. Secretion of TGF-@ was not effected (Fig. 3B). The reduction in IL-l activity in the LBRM-CTLL2 assay following IL-l ICV was not due to a direct suppression of IL-2 responsiveness of the CTLL-2 cells. Supernatants from macrophages treated with IL-l or saline ICV had no effect on IL-2-induced CTLL-2 proliferation (data not shown). Since the IL-l had a profound effect on ACTH and corticosterone release, and corticosterone is able to suppress macrophage IL- 1 secretion (17), we next examined

A Vehicle n rlL-18 5 ng

OL 2

8

Time (HIS) Poet Injection

24

I

I

I

2

8

24

Time (HIS) Post Injection

FIG. 2. Time course of plasma ACTH and corticosterone on 2, 8, and 24 hr postinjection of 5 ng IL-lb ICV. ** = P < 0.001, * = P < 0.05 using Student’s t test compared to vehicle injection. n = 5 rats/group.

INTRACEREBROVENTRICULAR

IL-l

AND

IMMUNOSUPPRESSION

89

140 A Vehicle n rll-18 5 ng

rlL-18 5 I

0s (pS/ml)

IControl 10 p&ml

FIG. 3. (A) Suppression of splenic macrophage IL-I production by ICV injection of 5 ng IL-16 (9 I .32) = 9.5h. P i 0.0 I). Spleens were collected 2 hr postinjection. n = 5 rats/group. This experiment was repeated with similar results. (B) TGF-@ production by macrophages from ICV IL-I stimulated rats stimulated with IO ng LPS.

basal and LPS-stimulated macrophage activity in adrenalectomized rats following either IL-1 or saline ICV (Fig. 4A). Adrenalectomy reversed the suppression of the macrophage resulting in an increase in IL-1 secretion. Furthermore. IL-1 production in ADX rats was significantly enhanced with and without stimulation of macrophages by LPS. All ADX mice had extremely low levels of circulating corticosterone (Fig. 4A). To assesswhether the suppressive signal could also be mediated by a sympathetic innervation of the spleen, the splenic nerve was surgically sectioned prior to ICV injection of IL-lp. The success of the splenic sympathectomy was confirmed by observing a reduction in norepinephrine content of sympathectomized spleens to 5- 15% of control levels. Sympathectomy of the spleen prevented IL- 1B ICV from suppressing macrophage IL- I secretion compared to saline ICV injected controls, despite significant increases in corticosterone levels (Fig. 4B). Macrophage IL-I secretion was not significantly different either with or without LPS stimulation. We next examined the effect of the combination of ADX and splenic nerve section. This resulted in a profound stimulatory effect on macrophage IL-1 secretion in response to LPS to levels considerably higher than that seen with either ADX or splenic nerve section alone (Fig. 4C). We confirmed that the increase in macrophage supernatant seen in adrenalectomized and splenic nerve sectioned rats was due to IL- 1. Supernatants were preincubated in either rabbit anti-IL- 1 antiserum or normal rabbit lg prior to assaying. All proliferative activity was neutralized by the antiserum (Fig. 5). DISCUSSION The present results show that ICV IL-10 administration suppressed IL- 1 release from LPS-stimulated splenic macrophages. The suppressive effects of corticosteroids

90

BROWN ET AL. L-

4oaI-

A. AOX

..

600 600

600

T

0 ; 600 cf a 9 400

/-

I 200

100

1

400’ I iit 200 D

NS

ii Ortb0sterOns

Ltk

NE

l .

f NS LJIL ;I

L3.

600

G m 2 5

Sympathectomy

T

600

z

604

_, $

B

400

5 400 $

200

200 $i

J

g

300

l .

NS

lhntrd

LPS

:. ADX b Sympathectomy

z w 2

600 f P.

.

4 000 2.

p” A

I L 200 al

400

NE

FIG. 4. Effect of 5 ng IL-ID or saline ICV on splenic macrophage IL-1 production following with or without LPS stimulation (A) ADX (adrenalectomy), (B) splenic nerve section and (C) ADX and splenic nerve section rats. Macrophage cultures were stimulated with 10 fig of LPS as described under Methods. Rats receiving 5 ng IL- 1j? ICV are marked in shaded boxes (m), while those receiving saline vehicle ICV are open boxes (Cl). Spleens were assayedfor norepinephtine content and serum assayedfor corticosterone on each animal to confirm the successof the adrenalectomy and nerve section. ** = P < 0.01, * = P < 0.05 using Student’s t test compared to vehicle injection. NS, not significant: NE, norepinephrine; n = 5-8 rats/ group.

on macrophage IL- 1 production have been well documented ( 17, 18). Given both the high levels of corticosterone elicited by ICV IL- 1p 5 ng and also the elimination of the suppressive effect of IL-l@ on macrophage IL-1 secretion following ADX, the corticosterone-induced suppression of IL- 1 secretion seemed an appropriate hypothesis. Several observations from the reported experiments, however, suggested that the regulation of macrophage function following ICV IL- 1 was more complex. Firstly, ADX did not simply eliminate an inhibitory effect but allowed stimulation of macrophage IL-l secretion following ICV IL- lp injection. This macrophage IL-I

INTRACEREBROVENTRICULAR

IL-l

0 Normal

1400

AND

IMMUNOSUPPRESSION

91

Rabbit Ig

n Rabbii Anti - IL-l

1200 -6 B a 1000 ‘; s 7 800 YZ E 600 400 200

FIG. 5. Immunoreactive IL-I in the supernatants of macrophages from adrenalectomized and nervesectioned rats treated ICV with IL- I Stimulation of IL-2 production by the LBRM-33 cell line was neutralized by rabbit anti-IL-l antibody. but not control rabbit immunoglobulin. ** = P < 0.01. * = P < 0.05 using Student’s t test compared to control rabbit immunoglobin: n = 5-8 rats per group.

response did not require LPS stimulation as macrophages from rats receiving ICV IL1p without LPS activation released significantly higher levels of IL- 1 to the supernatant. The question of the origin of the observed stimulatory signal observed in ADX rats injected with IL-l/3 is not immediately apparent but one candidate source was the sympathetic innervation of the spleen. Felten and co-workers (lo- 12) have demonstrated extensive sympathetic innervation of the spleen with tyrosine hydroxylase positive and peptidergic nerve terminals distributed among fields of T lymphocytes and, most notably for our experiments, macrophages. These authors have also described synapse-like connections between nerves and lymphocytes in the spleen (10-12). Our results show that sectioning of the splenic nerve, like adrenalectomy, blocked the suppressive effects of ICV IL- 1P on macrophage IL- 1 secretion. This indicated that sympathetics were not delivering a stimulatory signal but rather that a second negative regulatory pathway existed through the sympathetic innervation of the spleen. Furthermore, this occurred in the presence of significant increases in corticosterone, which reached levels comparable to rats with an intact splenic nerve. This would suggest that these levels of corticosterone were either insufficient to induce macrophage suppression, or that sectioning of the splenic nerve reduced the effectiveness of adrenocortical mediated suppression. Confirming that both pathways contributed to the suppressive signals, the combination of ADX and splenic nerve section resulted in the highest levels of IL-l production by splenic macrophages following ICV IL-l/3 injection. Taken together, the results of the present experiments suggest that the relationship between the HPA and macrophage IL-l production is not as simple as the model proposed by Besedovsky and colleagues (5, 6). They hypothesized that macrophage IL-l is secreted in response to antigen or other stimuli, enters the blood supply, and reaches the CNS thereby activating the HPA, possibly at the level of CRF release (18). ACTH is released from the pituitary which in turn stimulates corticosterone release

92

BROWN

ET AL.

from the adrenal cortex resulting in suppression of immune function, which in our experiment was measured as macrophage IL-I secretion, It is clear from the above experiments that, in addition to the adrenal pathway, signals mediated through the splenic nerve can also interfere with the IL-l-mediated CNS effects. Furthermore, the most pronounced macrophage IL-1 secretion in response to ICV IL-ID was seen in the ADX-nerve section group. Besedovsky et al. (19) have reported similar results in evaluating immune responses to sheep erythrocytes following surgical denervation of the spleen combined with ADX. This combination resulted in what these researchers described as “removal of restraint” on the ability of splenocytes to produce plaqueforming cells in response to immunization with sheep erythrocytes. The nature of this macrophage stimulatory signal is not known. It is possible that the ability of macrophages to autoactivate through the secretion of monokines such as TNF-a (20, 2 1) is normally down-regulated by adrenal and sympathetic control. That is, these regulatory pathways may mediate a negative tonic control on the macrophage which, in their absence, could continuously autoactivate in situ when minimally stimulated. Immunoregulatory neuropeptides and norepinephrine which are present within the splenic sympathetic autonomies may modulate macrophage function on their release following stimulation of this pathway. For example, norepinephrine can block LPS-induced interleukin-1 production in peritoneal macrophages (22). Substance P, on the other hand, enhances phagocytic activity (23) and augments macrophage CSF-l-induced myelopoeisis (24). A recent report by Sundar et al. (25) noted that substeroidogenic doses of IL-I@ ICV into rats resulted in depressed cell-mediated responses including IL-2 production, mitogen-induced proliferation, and NK activity. Because of the absence of steroidogenesis and the rapidity of the observed effects by 15 min postinfusion, the authors suggested an autonomic pathway as the mediator of these effects. Our results confirm the existence of such a pathway. In conclusion, we report the participation of the HPA and the sympathetic nervous system in the regulation of macrophage IL-l production. The relationship between the HPA and the sympathetic nervous system in the regulation of splenic macrophages is apparently complex. While surgical disruption of the function of either or both pathways results in dramatic alterations in IL-I production, the interactions of these systems in the intact animal are yet to be clarified. ACKNOWLEDGMENTS We thank Drs. David Urdal and Chris Henney of lmmunex for their gift of rIL-lfi. The fine technical assistance of Angela Kemp and Michael Talgoy was greatly appreciated. We thank Agnes Warkentin for typing this manuscript.

REFERENCES 1. 2. 3. 4. 5. 6. 7.

Farrar, W. L., Kilian, P. L., Ruff; M. R., Hill, J. M,, and Pert, C. B., J. Immunol. 139, 459, 1987. Fontana, A., Kristensen, F., Dubs, R., Gemsa, D., and Weber, E., J. Immunol. 129, 2413, 1982. Dinarello, C. A., FASEB J. 2, 108, 1988. Oomura, Y., Ph.wio/. Behnv. 44, 555, 1988. Del Rey, A., Besedovsky, H., Sorkin, E., and Dinarello, C. A.. Ann. N. Y. Acad. Sci. 496, 85, 1987. Besedovsky, H., Del Rey, A., Sorkin, E., and Dinarello, C. A., Science 233, 652. 1986. Sapolsky, R., Rivier, C., Yamamoto. G.. Plotsky, P.. and Vale, W., Science 238, 522. 1987.

INTRACEREBROVENTRICULAR

IL-l

AND

IMMUNOSUPPRESSION

93

8. Uehara. A., Gottschall, P. E.. Dahl, R. E., and Arimura, A., Endocrinology 121, 1580. 1987. 9. Beach. J. E., Smallridge, R. C.. Kinzer. C. A., Bernton. E. W.. Moladay, J. W., and Fein, H. G., L[fc sci. 44, I. 1989. IO. Felten, D.. Felten, S. Y., Carlson. S. L.. Olschowka, J. A., and Livnat, S.. J. Immune/. 135, 755. 1985. 1 I. Felten. D. L., Livnat. S., Felten, S. Y.. Carlson, S. L.. Bellinger. D. L.. and Yeh, P., Bruin Ra Bulletin. 13, 693, 1984. 12. Felten. D. L.. Felten, S. Y., Bellinger. D. L.. Carlson. S. L.. Ackerman, K. D., Hadden. K. S., Olschowki, J. A., and Livnat. S.. Immunol. Rev. 100, 225. 1987. 13. Paxinos. G., and Watson. C.. In “The Rat Brain in Stereotaxic Coordinates.“Academic Press: Sydney. 1982. 14. Conlon. P. J., J. fmmunol. 131, 1280, 1983. 15. Khalil, N.. Bereznay. O., Sporn, M.. and Greenberg, A. H., J. E.x/j, M&. 170, 727, 1989. 16. Greenberg. A. H., Dyck. D. G., Sandler. L. S., Pohajdak. B., Dreisel. K. M., and Grant, D.. J. ~a[/. Cuncrr

111.u. 72, 653. 1984.

17. Snyder. D. S.. and Unanue, E. R., J. fmmunol. 129, 1803. 1982. 18. Berkenbosch, F.. Van Oers. J., Del Rey. A., Tildens. F., and Besedovsky. H.. .Qience 240, 321, 1987. 19. Besedovsky. H. 0.. Del Rey. A., Sorkin. E.. Da Prada, M.. and Keller. H. H., Cd Immunol. 48, 346, 1979.

20. Philip, R.. and Epstein, L. B.. hiu/l/re (London) 303, 86, 1986. ‘2I. Hori, K.. Ehrke. M. J.. Mace. K., Maccubin, D., Doyle. M. J.. C&.&t, 47,2793.

y.. and Mihich, E.. C:(uzcer Rey.

1987.

22. Koff. W. C.. Fann. A. V.. Dunegan, M. A.. and Lachman, L. B.. Lymph&& Rev. 5, 239. 1986, 23. Bar-Shavit. Z., Goldman. R.. Stabinsky. Y., Gottleib. P., Fridkin, M.. Teichberg. V. I.. and Blumberg. S., Uioc/1cv~7. B~oph,~~s RPS. C’ommun 94, 1445. 1980. 24. Moore, R. N.. Osmond. A. P.. Dunn, J. A.. Joshi. J. G.. and Rouse, B. T., J. Immwd. 141, 2699. 1988.

25. Sundar. S. K.. Becker. K. J.. Curpial, M. A.. Carpenter, M. D.. Rankin, L. A., Fleener. S. L.. Ritchie. J. C.. Simson. P. E., and Weiss, J. M.. Proc, h’url. Acud. &I. c:S,4 86, 6398. 1989.