Brain, Behavior, and Immunity 27 (2013) 80–90

Contents lists available at SciVerse ScienceDirect

Brain, Behavior, and Immunity

journal homepage: www.elsevier .com/locate /ybrbi

Cerebellar fastigial nuclear GABAergic projections to the hypothalamusmodulate immune function

Bei-Bei Cao, Yan Huang, Jian-Hua Lu, Fen-Fen Xu, Yi-Hua Qiu ⇑, Yu-Ping Peng ⇑Department of Physiology, School of Medicine, Nantong University, 19 Qixiu Road, Nantong 226001, China

a r t i c l e i n f o

Article history:Received 11 August 2012Received in revised form 21 September 2012Accepted 29 September 2012Available online 6 October 2012

Keywords:CerebellumFastigial nucleusCerebellar–hypothalamic projectionGABALymphocytesHypothalamusIgM antibodyNK cytotoxicity

0889-1591/$ - see front matter � 2012 Elsevier Inc. Ahttp://dx.doi.org/10.1016/j.bbi.2012.09.014

⇑ Corresponding authors. Tel.: +86 513 8505172(Y.-H. Qiu), tel.: +86 513 85051714; fax: +86 513 850

E-mail addresses: yhqiu@ntu.edu.cn (Y.-H. Qiu), yp

a b s t r a c t

Our previous work has shown that the cerebellar fastigial nucleus (FN) is involved in modulation of lym-phocyte function. Herein, we investigated effect of FN c-aminobutyric acid (GABA)-ergic projections tothe hypothalamus on lymphocytes to understand pathways and mechanisms underlying cerebellarimmunomodulation. By injection of Texas red dextran amine (TRDA), an anterograde tracer, into FN,we found that the TRDA-labeled fibers from the FN traveled through the superior cerebellar peduncle(SCP), crossed in decussation of SCP (XSCP), entered the hypothalamus, and primarily terminated inthe lateral hypothalamic area (LHA). Further, by injecting Fluoro-Ruby (FR), a retrograde tracer, in LHA,we observed that the FR-stained fibers retrogradely passed through XSCP and reached FN. Among theseFR-positive neurons in the FN, there were GABA-immunoreactive cells. We then microinjected vigabatrin,which is an inhibitor of GABA-transaminase (GABA-T) that degrades GABA, bilaterally into FN. The viga-batrin treatment increased both number of GABA-immunoreactive neurons in FN–LHA projections andGABA content in the hypothalamus. Simultaneously, vigabatrin significantly reduced concanavalin A(Con A)-induced lymphocyte proliferation, anti-sheep red blood cell (SRBC) IgM antibody level, and nat-ural killer (NK) cell number and cytotoxicity. In support of these findings, we inhibited GABA synthesis byusing 3-mercaptopropionic acid (3-MP), which antagonizes glutamic acid decarboxylase (GAD). Wefound that the inhibition of GABA synthesis caused changes that were opposite to those when GABAwas increased with vigabatrin. These findings show that the cerebellar FN has a direct GABAergic projec-tion to the hypothalamus and that this projection actively participates in modulation of lymphocytes.

� 2012 Elsevier Inc. All rights reserved.

1. Introduction

The cerebellum, the largest subcortical center for motor control,has been reported to regulate nonsomatic physiological function,such as visceral activities (Holmes et al., 2002; Zhuang et al.,2008), feeding behavior (Zhu and Wang, 2008), cognition and work-ing memory (Alexander et al., 2012; Thürling et al., 2012). The im-mune system, as one of targets of nervous and endocrine systems, isalso regulated by the cerebellum. For example, in ‘‘reeler’’ mice, aneurological mutant strain with an abnormally high concentrationof cerebellar norepinephrine, function of T cells and macrophages issuppressed (Green-Johnson et al., 1995); lesion of the vestibulocer-ebellum of rats causes an immunosuppressive effect (Ghoshal et al.,1998). In the recent years, we have investigated immunomodula-tion of the cerebellum, focusing on the cerebellar nuclei, the fasti-gial nucleus (FN) and the interposed nucleus (IN), in view of theirimportant role in transmitting cerebellar output information.Lesions of neuronal somas in bilateral FN with kainic acid lead to

ll rights reserved.

3; fax: +86 513 8505187651506 (Y.-P. Peng).

peng@ntu.edu.cn (Y.-P. Peng).

enhancement of T, B and natural killer (NK) cells (Peng et al.,2005), while lesions of bilateral IN result in suppression of thesecells (Peng et al., 2006). These findings strongly show an involve-ment of the cerebellum in immunomodulation.

However, between the cerebellum and immune system, there isno direct structural connection, so exploring pathways and mech-anisms underlying the cerebellar immunomodulation is highlyimportant to better understand the phenomenon of immunomod-ulation. It has been well known that the hypothalamus is a crucialimmunoregulatory center, which modulates immune cells via thesympathetic nervous system (Madden, 2003; Nance and Sanders,2007; Trotter et al., 2007) and the hypothalamic–pituitary–adrenalaxis (Haddad et al., 2002; Pittman, 2011). Through releasing neuro-transmitters and endocrine hormones, respectively, the twoperipheral pathways directly contact immune cells and regulatetheir function (Baciu et al., 2003; Dong et al., 2002; Eskandariet al., 2003; Sanders and Kohm, 2002; Tayebati et al., 2000; Wrona,2006). Importantly, direct bidirectional connections between thecerebellum and the hypothalamus that constitute cerebellar–hypothalamic circuits have been found (Dietrichs et al., 1994;Haines et al., 1997). They are termed as cerebellar–hypothalamicprojections and hypothalamic–cerebellar projections. The direct

B.-B. Cao et al. / Brain, Behavior, and Immunity 27 (2013) 80–90 81

cerebellar–hypothalamic projections arise from all the threecerebellar nuclei, FN, IN and the dentate nucleus (DN), and termi-nate in several hypothalamic areas and nuclei, including the lat-eral, posterior and dorsal hypothalamic areas (LHA, PHA andDHA, respectively), as well as the dorsomedial and paraventricularnuclei (Cavdar et al., 2001a,b; Dietrichs and Haines, 1984; Hainesand Dietrichs, 1984; Haines et al., 1985, 1990). Based on thesefindings, we hypothesized that the direct cerebellar–hypothalamicprojections transmit cerebellar immunoregulatory information tothe hypothalamus, through which the regulatory information isthen conveyed to immune cells via the two peripheral pathways,sympathetic and endocrine systems. Here, we explored the roleof the direct cerebellar–hypothalamic projections in transmittingcerebellar immunomodulation to provide evidence for thehypothesis.

Although cerebellar–hypothalamic projections have beenreported, traveling routes and used neurotransmitters for the pro-jections need clarification. Our recent studies show that thecerebellar IN has direct glutamatergic and c-aminobutyric acid(GABA)-ergic projections to the LHA and that these projections reg-ulate lymphocyte function (Lu et al., 2012; Wang et al., 2011). How-ever, concerning the cerebellar FN, neither neurotransmitters usedby its projections to the hypothalamus nor role of the projections inimmunomodulation is so clear. Electrophysiological studies revealthat stimulation of FN evokes a monosynaptic inhibitory responseof LHA neurons (Min et al., 1989; Wang et al., 1997; Zhang et al.,2003), suggesting that a direct FN–LHA GABAergic projection existsbetween the cerebellum and hypothalamus. Nevertheless, morpho-logical observation needs to be established to demonstrate the di-rect FN–LHA GABAergic projections. Thus, in the present study,we firstly identified the FN GABAergic projections to the hypothal-amus and then showed modulation of lymphocytes by the projec-tions to provide more evidence for mechanisms and pathwaysunderlying cerebellar immunomodulation.

2. Materials and methods

2.1. Animals

The study was conducted on 170 adult Sprague–Dawley rats ofeither sex weighing 220–240 g (Center of Experimental Animals,Nantong University, China). Animals were housed on a 12 h light/dark cycle with food and water available ad libitum. Among theserats, 20 were used to trace nerve fibers between the cerebellar FNand the hypothalamus, with 10 for anterograde tracing from the FNto the hypothalamus and 10 for retrograde tracing from the LHA tothe FN combined with immunohistochemistry for GABA. In the restof the animals, 72 were randomized to the three groups, intact,FN–saline treatment and FN–vigabatrin treatment, being 24 foreach group, with 7 for measurement of FN–LHA GABAergictransmission and GABA content in the hypothalamus, six for T cellproliferation, six for anti-sheep red blood cell (SRBC) IgM antibodylevel, and five for NK cell number and cytotoxicity; and 78 werealso randomly divided into the three groups, intact, FN–salinetreatment and FN-3-mercaptopropionic acid (3-MP) treatment,being 26 for each group, with seven for determination of FN–LHAGABAergic transmission and GABA content, seven for T cell prolif-eration, five for anti-SRBC IgM antibody, and seven for NK cellnumber and cytotoxicity. All the animals were originally used inthis study, but only those rats placed in an accurate location withthe tracers or drugs were statistically analyzed.

2.2. Anterograde tracing of FN projections to the hypothalamus

Rats were mounted in a stereotaxic frame (David Kopf 902-A,USA) after anesthetized with pentobarbital (55 mg/kg, i.p.). A

volume of 0.3 ll of 5% Texas red dextran amine (TRDA,MW = 10000, Invitrogen) was delivered into unilateral FN by pres-sure injection using the following stereotaxic coordinates: 11.5 mmposterior to the bregma, 1.1 mm left/right to the midline, and6.3 mm ventral to the bregma (Paxinos and Watson, 1998). Theinjection was performed within 4 min and following the injection,the needle remained in the target location for 5 min to avoid thetracer reflux along the needle tract. After survival for 12 days, therats were anesthetized with an overdose of pentobarbital and per-fused transcardially with saline followed by 4% paraformaldehydein 0.1 M phosphate buffer (PB, pH 7.4). Rat brains were post-fixedin 4% paraformaldehyde for 4 h and then cryoprotected in 20%and 30% sucrose in 0.1 M PB overnight at 4 �C. Continuous coronalsections from the cerebellum to the hypothalamus were prepared(30 lm in thickness) with a freezing microtome (Leica CM 1900-1-1, Germany) and observed under a fluorescent microscope (LeicaDML, Germany).

2.3. Retrograde tracing combined with GABA immunohistochemistry

The 0.4 ll of 10% Fluoro-Ruby (FR, Invitrogen) was injected intoone side of LHA using the following coordinates: 4.2 mm posteriorto bregma, 1.2 mm left/right to the midline, and 8.1 mm ventral tothe bregma, where the TRDA-positive fibers terminated mostly.After 8 days of survival, the animals were perfused with saline fol-lowed by a fixative solution containing 4% paraformaldehyde and2.5% glutaraldehyde in 0.1 M PB (pH 7.4). Then 10 lm-thick sec-tions from the cerebellum to the hypothalamus were cut andmounted on glass slides. The FR-labeled neurons in the FN were ob-served. For GABA immunostaining, the cerebellar sections wererinsed in 0.01 M phosphate-buffered saline (PBS) and treated with1% sodium borohydride for 30 min. Subsequently, the sections wereimmersed in 3% normal serum containing 0.3% Triton X-100 for30 min to block non-specific binding of antigens. Then the sectionswere rinsed again in 0.01 M PBS and incubated in mouse monoclo-nal anti-GABA antibody solution (Chemicon, at 1:500 dilution) for18–24 h at 25 �C. After rinsing in 0.01 M PBS again, the sectionswere incubated in FITC conjugated goat anti-mouse IgG antibody(Sigma, at 1:64 dilution) for 4 h at 25 �C. Finally, the cerebellar sec-tions were washed in 0.01 M PBS and observed under a fluorescentmicroscope.

2.4. Injection of vigabatrin or 3-MP in bilateral cerebellar FN

After anesthesia with pentobarbital, rats were placed in a stereo-taxic frame (David Kopf 902-A, USA). Vigabatrin (12 lg in 0.3 ll sal-ine for each side) or 3-MP (60 lg in 0.3 ll saline for each side) wasmicroinjected into bilateral FN according to the same stereotaxiccoordinates as described above. Rats with intact or saline-infusedFN were used as controls. On the third day after the injection, Tlymphocyte proliferation and anti-SRBC IgM antibody level in theserum were examined. For assessment of NK cells, vigabatrin or3-MP was again injected in bilateral FN on the sixth day after thefirst injection. NK cytotoxicity was measured on the nineth dayafter the first injection.

2.5. High-performance liquid chromatography assay for GABA contentin the hypothalamus

On the third day after vigabatrin or 3-MP injection in bilateralFN, the hypothalamus was dissected on ice. The tissue was homog-enized in 0.4 M perchloric acid with 0.1 mM EDTA (1:10 w/v) at 0 �Cand then centrifuged at 15,000 rpm for 20 min at 4 �C. The superna-tants were stored at �80 �C until analysis. GABA content wasmeasured using precolumn derivatization with 6-aminoquinolyl-N-hydroxy-succinimidyl carbamate and fluorescence detection.

82 B.-B. Cao et al. / Brain, Behavior, and Immunity 27 (2013) 80–90

Briefly, 10 ll of each extract from the hypothalamus was injectedinto high-performance liquid chromatography (HPLC) equipment(Waters, USA), which mainly includes 1525 binary pump and2475 multi k fluorescence detector, with an amino acid analysis col-umn (AccQ TagTM, 3.9 � 150 mm, 4 lm beads, Waters). Separationof the mixture of amino acids was carried out using a gradient sys-tem at a flow rate of 1 ml/min. Mobile phase A consisted of sodiumacetate buffer, and mobile phase B was acetonitrile. Fluorometricdetection was performed at excitation and emission wavelengthsof 250 and 395 nm, respectively.

2.6. Carboxyl fluorescein succinimidyl ester/CD3 double-labelingmeasurement of T lymphocyte proliferation

On the third day after vigabatrin or 3-MP injection into bilateralcerebellar FN, the mesenteric lymph nodes were harvested from theanesthetized rats by celiotomy. Lymphocytes were obtained bygently squeezing the lymph nodes and then washed three timeswith RPMI 1640 culture medium (Gibco). After erythrocytes werelysed by sterilized distilled water, the lymphocytes were thenresuspended in RPMI 1640 culture medium at a concentration of1 � 107 cells/ml. Carboxyl fluorescein succinimidyl ester (CFSE,Invitrogen) was added to the cell suspension at a final concentra-tion of 2 lM, which was incubated for 15 min at 37 �C in a waterbath in the dark. The labeling was terminated by adding the samevolume of 100% fetal calf serum to quench the free CFSE at roomtemperature for 5 min. The labeled cells were then washed threetimes with sterile RPMI 1640 culture medium, counted andresuspended in complete culture medium at a concentration of1 � 106 cells/ml. The complete culture medium consisted of RPMI1640 medium, 10% heat-inactivated calf serum, 2.5 � 10�2 MHEPES (Sigma), 1 � 10�3 M sodium pyruvate, 5 � 10�5 Mb-mercaptoethanol and antibiotics (100 U/ml penicillin, 100 U/mlstreptomycin). These cells were incubated with 5 lg/ml concanav-alin A (Con A, Sigma) in an incubator (ESPEC BNA-311, Japan) at37 �C in 5% CO2 for 72 h. After incubation, the cultured cells werethen harvested, washed twice with RPMI 1640 culture mediumand resuspended in 100 ll of 0.01 M PBS per sample. Then0.25 lg of PE conjugated anti-CD3 antibody (eBioscience) wasadded to each sample to identify T lymphocytes. Each sample wasincubated for 30 min and analyzed using a FACSCalibur flow cytom-eter (BD Biosciences) by acquiring 5000 cells. FACS data were ana-lyzed using CellQuest software (BD Biosciences).

2.7. Enzyme-linked immunosorbent assay for anti-SRBC IgM antibodylevel in the serum

Rats were immunized via intraperitoneal injection of 2 ml sterilesaline containing 1.25 � 108 SRBC per ml. On the third day afterimmunization, the rats were treated with vigabatrin or 3-MP. Bloodwas taken from the right ventricle on the fifth day after immuniza-tion, when IgM response was peak, and serum was collected by cen-trifugation at 1000 rpm for 10 min and detected for anti-SRBC IgMantibody level by enzyme-linked immunosorbent assay (ELISA).Hemoglobin-free SRBC membranes were prepared by lysing defi-brinated SRBC with Tris–EDTA repeatedly. A 96-well microtiterplate (Costar, Corning Incorporated, USA) was coated with 100 llSRBC membranes per well at 4 �C for at least 16 h. After washingwith 0.05% Tween 20 (Sigma), the plate was blocked with blockingbuffer for 1 h to avoid non-specific binding. Diluted serum samples(1:500 in 0.01 M PBS) were added to the wells (125 ll/well) andincubated at 37 �C for 1 h. Hundred and twenty five microlitre ofgoat anti-rat IgG-HRP (Serotec, Oxford, UK, 1:3000 dilution) wasadded to each well, which was incubated for 1 h at 37 �C. Afterwashing with PBS/Tween 20, 3,305,50-tetramethylbenzidine withhydrogen peroxide was added to each well, which was incubated

for 10 min. The reaction was stopped with 0.5 M H2SO4. Absorbancewas determined with a multi-mode microplate reader (Bio Tek,USA) at 450 nm.

2.8. Flow cytometric assay for NK cell number and cytotoxicity

On day three following the second injection of vigabatrin or3-MP in bilateral FN, the spleens were collected and splenic mono-nuclear cells were isolated by density gradient centrifugation on aFicoll–Hypaque gradient (specific gravity: 1.085). After washingtwice with RPMI 1640 culture medium, the splenic mononuclearcells were suspended in complete culture medium in a culture flaskand incubated at 37 �C in 5% CO2 for 2 h. Non-adherent cells, aseffector cells, were collected and resuspended in the complete cul-ture medium at a concentration of 4 � 106 cells/ml. For counting NKcell percentage in splenic mononuclear cells, the non-adherent cellswere centrifugated and incubated with an anti-rat NKR-P1A-APCantibody (0.25 lg in 100 ll PBS, Invitrogen) for 30 min at roomtemperature protecting from light. These cells were then deter-mined by using a FACSCalibur flow cytometer (BD Biosciences)equipped with CellQuest software. For analysis of NK cell cytotox-icity, the target cells, YAC-1 cell line (Shanghai Institute of Biochem-istry and Cell Biology, Chinese Academy of Science) that is aMoloney leukemia virus-induced mouse lymphoma with notedsensitivity to NK cells, were labeled with 100 nM of calcein acetoxy-methyl (CAM, Fluka) for 15 min at 37 �C in 5% CO2. The labeled YAC-1 cells were washed twice, counted and adjusted to 4 � 105 cells/ml. The effector cells (mononuclear cells) were co-incubated withthe target cells (YAC-1 cells) at the ratio of 10:1 for 2 h at 37 �C ina humidified 5% CO2 incubator. After the incubation period,0.1 mM ethidium homodimer-1 (EH-1, Fluka) of 2 ll was combinedwith these cells (each ml) at a final concentration of 200 nM for15 min at room temperature to stain dead cells. The FACSCaliburflow cytometer (BD Biosciences) equipped with CellQuest softwaremeasured these samples. To exclude the possibility of spontane-ously dead YAC-1 cells, we measured background death of YAC-1cells in the absence of mononuclear cells in each experiment. Lessthan 5% spontaneous lysis of target cells was observed in theseexperiments and it was subtracted from percentage of totally deadYAC-1 cells in each sample. NK cell cytotoxicity was expressed aspercentage of specifically dead YAC-1 cells relative to total YAC-1cells.

2.9. Statistical analysis

Results were expressed as mean ± standard error. Statisticalanalysis was performed with Statistics Package for Social Science(SPSS, 16.0). The data were submitted to the one-way analysis ofvariance (ANOVA), followed by Student–Newman–Keul’s test tocompare the data of all groups between each other. Differenceswere considered statistically significant at p < 0.05.

3. Results

3.1. Traveling courses and ending locations of neuronal projectionsfrom the cerebellar FN to the hypothalamus

TRDA, an anterograde tracer with fluorescence, was microin-jected into unilateral FN and 12 days later, continuous brain sec-tions from the cerebellum to the hypothalamus were cut toidentify neuronal projections from the FN to the hypothalamus.Of the ten rats that received TRDA injections, eight were accuratein FN location and confined to FN areas. An example for TRDA-labeled neurons in FN area was illustrated in Fig. 1A. The TRDA-labeled neurons sent fibers that traveled primarily in the superior

B.-B. Cao et al. / Brain, Behavior, and Immunity 27 (2013) 80–90 83

cerebellar peduncle (SCP) with the shapes of dots or cords (Fig. 1B).These fibers crossed in decussation of SCP (XSCP) (Fig. 1C),penetrated in contralateral SCP (Fig. 1D), and entered the hypothal-amus. In the hypothalamus, the TRDA-positive fibers terminatedmostly in LHA and partly in PHA with the characteristics of varicos-ities or diffuse plexus (Fig. 1E).

3.2. The cerebellar FN has GABAergic output to LHA

To clarify that GABA is a neurotransmitter for FN–LHA neuronalprojections, we employed a combined examination with retro-grade tracing and GABA immunostaining. FR, a retrograde tracerwith fluorescence, was microinjected into unilateral LHA and onthe eighth day after the injection, cerebellar sections were treatedwith GABA immunofluorescence staining. We observed that the FRlabels were confined to LHA (Fig. 2A1). The FR-labeled fibers retro-gradely traveled through XSCP (Fig. 2A2) and reached FN and othertwo nuclei (IN and DN) of the cerebellum (Fig. 2A3). Therefore, inthe FN, these FR-stained neurons were retrogradely traced fromthe LHA (Fig. 2B1). Simultaneously, many GABA-immunoreactivecells were seen in the FN (Fig. 2B2). Merging these red FR-labeledcells and green GABA-positive neurons in the FN, we observedsome small yellow cells (Fig. 2B3). These yellow cells representedFN GABAergic neurons that sent axons to the LHA. These resultsshowed that there was a direct GABAergic projection from the cer-ebellar FN to LHA.

3.3. Changes of FN–LHA GABAergic transmission following vigabatrinor 3-MP injection into bilateral FN

After vigabatrin, an inhibitor of GABA-transaminase (GABA-T)that degrades GABA, or 3-MP, an antagonist of glutamic acid decar-boxylase (GAD) that synthesizes GABA, was microinjected intobilateral cerebellar FN, changes of FN–LHA GABAergic transmissionwere assessed by FR retrograde tracing combined with GABAfluorescence immunohistochemistry. As described above, in FNthe FR-labeled neurons, which were retrogradely traced fromLHA, and the GABA-immunoreactive neurons were clearly seen(Fig. 3). The GABA-immunoreactive neurons in FN were increasedby vigabatrin treatment but decreased by 3-MP. Importantly, the

Fig. 1. Histological observation of neuronal projections from the cerebellar FN to thehypothalamus with 30 lm in thickness on the 12th day after the cerebellar FN was microFN area (A). The TRDA-positive fibers travel in ip silateral SCP (B), cross in XSCP (C), go inpartly in PHA (E1). The characteristics of nerve endings with the finely beaded-like varicothe shapes of dots or cords are designated with arrowheads and magnified in relative u

percentage of FR/GABA double-positive neurons in FR single-labeled cells was significantly elevated by the vigabatrin injectionbut notably reduced by the 3-MP treatment (Fig. 3). There were nostatistical differences in the percentage of FR/GABA neuronsbetween intact and saline-treated FN (Fig. 3). These results indi-cated that vigabatrin injection in bilateral FN enhanced FN–LHAGABAergic transmission, whereas 3-MP impaired the transmission.

3.4. Effects of injection of vigabatrin or 3-MP in the FN on GABAcontent in the hypothalamus

On the 3rd day after vigabatrin or 3-MP injection in bilateral FN,the content of GABA in the hypothalamus was examined. The vig-abatrin treatment in the FN notably raised GABA content in thehypothalamus, while 3-MP remarkably diminished GABA contentin the hypothalamus, in comparison with control rats with intactor saline-infused FN (Fig. 4). There were no significant differencesin GABA content of the hypothalamus between the two groups ofcontrol rats. These data further showed that vigabatrin or 3-MPinjection in bilateral cerebellar FN enhanced or attenuatedFN–hypothalamic GABAergic transmission, respectively.

3.5. Vigabatrin attenuates but 3-MP augments T lymphocyteproliferation after injection into bilateral cerebellar FN

On the 3rd day after vigabatrin or 3-MP injection into bilateralcerebellar FN, Con A-induced lymphocyte proliferation was deter-mined by CFSE combined with anti-CD3 antibody labeling. Theproportion of proliferated T cells by Con A stimulation was signif-icantly reduced by vigabatrin treatment but remarkably increasedby 3-MP (Fig. 5). To show the changes of T cell proliferation werenot caused by a possible alteration of T cell number, we simulta-neously measured T cell number after the drugs acted. The vigabat-rin or 3-MP treatment in the cerebellar FN did not significantlyalter the number of T lymphocytes (Fig. 5), demonstrating thatthe drug-induced changes in the proliferative response of T cellswere not associated with T lymphocyte number. In addition, nostatistical differences were found in the proportion of proliferatedT cells between intact and saline-treated control rats.

hypothalamus. Brain sections were continuously cut from the cerebellum to theinjected with the anterograde tracer, TRDA. The injection site of TRDA is confined toto contralateral SCP (D), enter the hypothalamus and mostly terminate in LHA and

sities in LHA can be seen via the magnification (E2). The typical traveling fibers withp-right boxes. 3 V: third ventricle; scale bars = 100 lm.

Fig. 2. Retrograde tracing and GABA immunohistochemistry for FN–LHA neuronal projections. The retrograde tracer FR was microinjected into LHA and 8 days later,immunohistochemistry for GABA was performed on the cerebellar sections. The FR-stained area shows that the FR injection is confined to LHA (A1). These FR-positive fibersretrogradely travel through XSCP (A2), and reach the cerebellar nuclei, FN, IN and DN (A3). In FN, the FR-labeled neurons, which retrogradely come from LHA, can be clearlyseen (B1). By using GABA immunohistochemistry on the same FN section, we observe many GABA-immunoreactive neurons (B2). Merging the FR retrograde tracing signals(red) and the GABA immunoreactivity (green) obtains the small yellow neurons (indicated by the arrowheads, B3). These yellow cells represent GABAergic neurons in FNprojections to LHA. Aq: aqueduct; scale bars = 100 lm. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of thisarticle.)

84 B.-B. Cao et al. / Brain, Behavior, and Immunity 27 (2013) 80–90

3.6. Vigabatrin treatment of the cerebellar FN reduces but 3-MPtreatment elevates levels of anti-SRBC IgM antibody in the serum

On the third day following injection of vigabatrin or 3-MP intobilateral FN, levels of anti-SRBC IgM antibody in the serum weremeasured by ELISA. The anti-SRBC IgM antibody levels in the ser-um were markedly lower in vigabatrin-treated rats but were high-er in 3-MP-treated rats than in control animals with intact orsaline-treated FN (Fig. 6). Between the two groups of control rats,no statistical differences were observed in the anti-SRBC IgM anti-body levels.

3.7. Vigabatrin injection in bilateral FN impairs but 3-MP injectionenhances NK cells

Vigabatrin or 3-MP was microinjected into bilateral cerebellarFN on the 1st and sixth days, respectively, and cytotoxicity andnumber of NK cells were assessed by flow cytometric assay onthe nineth day. Quantity of dead YAC-1 cells killed by NK cellswas obviously less in vigabatrin-treated rats than in intact orsaline-treated rats (Fig. 7A). Simultaneously, this treatmentdecreased percentage of NK cells in splenic mononuclear cells(Fig. 7A). However, the amount of dead YAC-1 cells killed by NKcells was significantly more in 3-MP-treated animals than in con-trol animals with intact or saline-injected FN (Fig. 7B). Similarly,the number of NK cells in mononuclear cells was increased by thetreatment with 3-MP (Fig. 7B). No significant differences werefound between intact and saline-treated control rats in the cytotox-icity or number of NK cells (Fig. 7).

4. Discussion

By using TRDA anterograde tracing, we found that the TRDA-labeled fibers arising from the cerebellar FN traveled through SCP,crossed in XSCP, and primarily terminated in LHA. Since the antero-grade tracer, TRDA, is not transmitted to another neuron by thesynapse (Köbbert et al., 2000), the observation of traveling pathwaysand ending locations of the TRDA-positive fibers demonstrates a

direct neuronal projection from the cerebellar FN to LHA. This resultsupports these reports showing direct cerebellar–hypothalamicprojections (Dietrichs and Haines, 1984; Haines et al., 1985, 1990).Further, by injection of the retrograde tracer, FR, in LHA, we observedthat the FR-labeled nerve fibers traveled through XSCP and reachedthe cerebellar FN, IN and DN. Likewise, FR is not conveyed by the syn-apse to another neuron (Köbbert et al., 2000), so these FR-positiveneurons in the cerebellar nuclei directly came from LHA. These con-sistent results obtained from the anterograde and retrograde trac-ings confirm a direct neuronal projection from the cerebellarnuclei to the hypothalamus. It has been reported that routes of allnerve fibers to and from the cerebellum pass through the three cer-ebellar peduncles, the superior, middle, and inferior cerebellarpeduncles, among which output fibers arising from the cerebellarnuclei pass through SCP (Ito, 2002). Our present results that the cer-ebellar FN sent axons to the hypothalamus through SCP support thisreport.

In the cerebellar FN, there are GABAergic neurons (Chung et al.,2009; Sultan et al., 2002; Uusisaari and Knöpfel, 2008).Electrophysiological studies reveal that stimulation of FN evokesa monosynaptic inhibitory response of LHA neurons, suggesting adirect GABAergic projection from FN to LHA (Min et al., 1989;Wang et al., 1997; Zhang et al., 2003). In the current study, weobserved by combined use of FR retrograde tracing and GABAimmunofluorescent histochemistry that in the FR-labeled neurons,which came retrogradely from LHA to FN, there were GABA immu-noreactive cells. These FR/GABA double-stained cells in the cere-bellar FN represent GABAergic neurons that project fibers to LHA.Thus, the present morphological observation provides direct andfurther evidence for the FN–LHA GABAergic projections.

In the brain, GABA, as a major inhibitory neurotransmitter, isfinely regulated for its concentration by two main enzymes: onegoverning its synthesis (GAD) and the other one responsible forits catabolism (GABA-T) (Mausset et al., 2001). 3-MP is a GADantagonist competitive with glutamate to inhibit GABA synthesis(Crick et al., 2007). Since 3-MP can easily enter cells for its lipo-philic properties and is transported to axonal terminals by axoplas-mic transportation, it decreases GABA concentration in thelocations of local injection and remote endings (Engel et al.,

Fig. 3. Changes in number of GABA-immunoreactive neurons in FN projections to LHA after vigabatrin or 3-MP injection into bilateral FN. The retrograde tracer FR wasinjected in LHA and five days later, vigabatrin or 3-MP was microinfused into bilateral FN. On day three following the vigabatrin or 3-MP treatment, the cerebellum was cutinto 10 lm-thick sections and the sections were stained with GABA fluorescence immunohistochemistry. The left panels in (A) and (B) reflect FR-stained neurons in FN withvarious treatments. These neurons were retrogradely traced from LHA. The middle panels in (A) and (B) denote GABA-immunoreactive neurons in FN with these treatments.The right panels in (A) and (B) show FR/GABA double-labeled neurons in FN with the different treatments. These double-labeled yellow cells represent GABAergic neurons inFN–LHA projections. They were significantly increased by the vigabatrin treatment but remarkably decreased by the 3-MP injection (C). These experiments were repeated forfive times. ⁄⁄p < 0.01, compared with intact FN; ++p < 0.01, compared with saline-treated FN. Scale bars = 100 lm. (For interpretation of the references to color in this figurelegend, the reader is referred to the web version of this article.)

B.-B. Cao et al. / Brain, Behavior, and Immunity 27 (2013) 80–90 85

Fig. 4. Effects of vigabatrin or 3-MP injection into bilateral FN on GABA content in the hypothalamus. HPLC assay was used to test GABA content in the hypothalamus afterinjection of vigabatrin (A) or 3-MP (B) into bilateral FN. GABA was eluted at the 10th peak according to the standard amino acid samples (A and B). The data are mean andstandard error of seven respective experiments (C). ⁄p < 0.05, compared with intact rats; +p < 0.05, compared with saline-treated rats.

86 B.-B. Cao et al. / Brain, Behavior, and Immunity 27 (2013) 80–90

2001; Netopilová et al., 1995). On the other hand, after GABA actson the postsynaptic cell, it is removed from the synaptic cleft intolocalized nerve terminals and glial cells by specific membrane-bound transport molecules. Following removal from the synapse,GABA is either recycled to the readily releasable neurotransmitterpool (GABAergic nerve terminals only) or metabolized (neuronsand glial cells) to an inactive molecule by action of the mitochon-drial enzyme GABA-T (Kwan et al., 2001). Compounds capable ofinhibiting the enzymatic activity of GABA-T are expected tofacilitate GABAergic transmission by increasing the presynapticavailability and subsequent release of GABA (Jackson et al.,2000). Vigabatrin, an irreversible inhibitor of GABA-T, can induce

a dramatic increase in GABA concentration in the locations of localinjection and remote endings (Mausset et al., 2001). In this study,injection of 3-MP in bilateral cerebellar FN led to a decrease bothin number of FN GABAergic neurons projecting to LHA and in GABAcontent in the hypothalamus. On the contrary, injection of vigabat-rin in bilateral cerebellar FN resulted in an increase both in numberof FN GABAergic neurons projecting to LHA and in GABA content inthe hypothalamus. These data demonstrate that the 3-MP treat-ment in FN impairs but the vigabatrin treatment enhances FN GAB-Aergic transmission to the hypothalamus.

Importantly, the 3-MP injection in bilateral cerebellar FN causedan augment of T cell proliferation, anti-SRBC antibody level, and NK

Fig. 5. Effects of vigabatrin or 3-MP injection in bilateral FN on T lymphocyte proliferation. On the third day after vigabatrin or 3-MP was infused into bilateral FN,proliferative response of lymphocytes from the mesenteric lymph nodes to Con A was assessed by flow cytometric analysis based on CFSE and anti-CD3 antibody staining. In(A) and (B), proportion of T lymphocytes in lymph node cells is shown in (P1), and proportion of proliferated T lymphocytes in T cells appears in (P2). The data are from six orseven separate experiments (C). ⁄⁄p < 0.01, compared with intact rats; ++p < 0.01, compared with saline-treated rats.

**++

*+

0.0

0.5

1.0

1.5

2.0

2.5

3.0

Vigabatrin 3-MP

IgM

OD

val

ue

IntactSalineVigabatrin/3-MP

Fig. 6. Influences of vigabatrin or 3-MP treatment in bilateral FN on anti-SRBC IgMantibody levels in the serum. On the third day following vigabatrin or 3-MPinjection into bilateral FN of rats that had been intraperitoneally injected with SRBC2 days earlier, peripheral sera of the rats were collected and measured for anti-SRBCIgM antibody levels by ELISA. Each bar in the statistical diagram indicates the meanand standard error of six or five replications for vigabatrin or 3-MP, respectively.⁄p < 0.05, ⁄⁄p < 0.01, compared with intact rats; +p < 0.05, ++p < 0.01, compared withsaline-treated rats.

B.-B. Cao et al. / Brain, Behavior, and Immunity 27 (2013) 80–90 87

cell number and cytotoxicity, while the vigabatrin injection in bilat-eral cerebellar FN brought an attenuation of T cell proliferation,anti-SRBC antibody level, and NK cell number and cytotoxicity.The opposite actions of the 3-MP and vigabatrin treatment in FNon T, B and NK cells strongly show that FN GABAergic neurons areactively involved in modulation of peripheral lymphocyte activity.We have previously indicated that lesions of bilateral cerebellarFN with kainic acid led to an enhancement of T and NK cells (Penget al., 2005), suggesting that the cerebellar FN participates in mod-ulation of the immune system. Here, we provide further evidenceand present that FN GABAergic neurons are a mechanism underly-ing FN immunomodulation. Notably, the immune response changescaused by the 3-MP or vigabatrin treatment in cerebellar FN wereaccompanied by the alteration in FN GABAergic projections to thehypothalamus. Therefore, it is suggested that FN GABAergic projec-tions to the hypothalamus are a route by which the cerebellummodulates peripheral immune response. This view is supportedby our recent reports indicating that cerebellar IN has direct gluta-

Fig. 7. Effects of vigabatrin or 3-MP injection into bilateral FN on NK cell number and cytotoxicity. On day nine after vigabatrin or 3-MP was first infused into bilateral FN,splenic mononuclear cells were separated and detected for NK cell number and cytotoxicity against YAC-1 cells by use of flow cytometry with NKR-P1A or CAM/EH-1 labeling.In (A1) and (B1), each fluorescent dot image has four plots, among which the upper left (UL) represents number of dead NK cells labeled by EH-1, the upper right (UR) isproportion of dead YAC-1 cells stained by both CAM and EH-1, the low left (LL) denotes percentage of intact NK cells labeled by neither CAM nor EH-1, and the low right (LR)indicates number of living YAC-1 cells stained by CAM. In (A2) and (B2), the values are percentage of NK cells in splenic mononuclear cells. The statistical data for vigabatrin(A3) and 3-MP (B3) are from five or seven independent experiments. NK cell cytotoxicity = [(UR/(UR + LR)killed � (UR/UR + LR)spontaneous)] � 100%. ⁄p < 0.05, ⁄⁄p < 0.01,compared with intact rats; +p < 0.05, ++p < 0.01, compared with saline-treated rats.

88 B.-B. Cao et al. / Brain, Behavior, and Immunity 27 (2013) 80–90

matergic and GABAergic projections to LHA and these projectionsregulate lymphocytes (Lu et al., 2012; Wang et al., 2011).

In the classical knowledge, the cerebellum serves as a subcorti-cal motor control center by receiving inputs from the cerebral cor-tex and sensory systems and by sending outputs via the deepcerebellar nuclei, FN, IN, and DN, to motor neurons in the cortex(via a relay in the thalamus) and in the brainstem (Hall, 2004).However, over recent decades, the cerebellum has been expandedabout its structural connections and functional roles. In particular,the findings of direct bidirectional connections between the cere-bellum and the hypothalamus that constitute cerebellar–hypotha-lamic circuits including cerebellar–hypothalamic projections and

hypothalamic–cerebellar projections (Dietrichs et al., 1994; Haineset al., 1997) provide a substantial basis for extending the under-standing of cerebellar function. For example, via the cerebellar–hypothalamic projections, the cerebellum regulates nonsomaticphysiological function, such as visceral activities (Onat and Cavdar,2003; Zhu et al., 2006) and feeding behavior (Zhu and Wang, 2008).Immune function, which could be considered as a nonsomaticphysiological function, has been also reported to be modulated bythe cerebellum. Green-Johnson et al. (1995) reveal that in ‘‘reeler’’mice, a neurological mutant strain with an abnormally high con-centration of cerebellar norepinephrine, function of T cells andmacrophages is suppressed. Subsequently, Ghoshal et al. (1998)

B.-B. Cao et al. / Brain, Behavior, and Immunity 27 (2013) 80–90 89

report that lesion of the vestibulocerebellum of rats causes animmunosuppressive effect. These findings suggest the cerebellumis involved in mediating immunomodulation. We previouslyobserved that lesions of the two cerebellar nuclei, FN and IN, respec-tively, produced opposite effects on immune cells (Peng et al., 2005,2006). Since cerebellar information is exported ultimately viathe cerebellar nuclei, the immune changes caused by FN or INlesions represent an immunomodulation by the cerebellum.

However, between the cerebellum and immune system, there isno direct structural connection. The current results showing thatthe FN–hypothalamic GABAergic projections influenced lympho-cyte responses suggest that the hypothalamus is a relay in convey-ing the immunoregulatory information from the cerebellum. It hasbeen well known that the hypothalamus, a crucial center regulatingvisceral activities, has definite immunomodulatory effect by itsdirect control of sympathetic nerves and endocrine glands that con-tact immune cells via releasing neurotransmitters and hormones,respectively (Baciu et al., 2003; Sanders and Kohm, 2002; Wrona,2006). Consequently, by the connection of the cerebellum withthe hypothalamus via the bidirectional cerebellar–hypothalamiccircuits, cerebellar function can be more extensive and intensive.On the one hand, via the hypothalamic–cerebellar projections andthe classical inputs from the cerebral cortex and sensory systems,the cerebellum integrates the diverse information it receives aboutmovement status, visceral activities, and immune response. On theother hand, by the cerebellar–hypothalamic projections and theclassical outputs to motor neurons in the cortex and in the brain-stem, the cerebellum provides a coordinated set of modulation ofsomatic movement, visceral activities, and immune response.

Further, in the hypothalamus, the LHA has been reported to reg-ulate immune response by exporting information to endocrineglands and sympathetic nerves that innervate immune cells (Wro-na, 2006; Wrona and Trojniar, 2003; Wrona et al., 2003). Accord-ingly, the present findings demonstrating that the FN GABAergictransmission to LHA affected peripheral lymphocytes suggest apossibility that FN GABAergic projections alter LHA neuronal activ-ity via a synaptic connection and subsequently the LHA neuronsmodulate lymphocytes via the sympathetic nerves or endocrinehormones. We previously found that at the time when immunechanges were caused by lesions of bilateral cerebellar FN, contentof norepinephrine, which is a major neurotransmitter of sympa-thetic nerves, in the spleen was reduced (Peng et al., 2005). This re-sult supports our present hypothesis, suggesting that in theperiphery, the sympathetic nerves transmit immunomodulatinginformation from the cerebellum via a relay in the hypothalamus.

In conclusion, a direct GABAergic projection from cerebellar FNneurons to the hypothalamus is demonstrated. Changing the FN–hypothalamic GABAergic transmission by 3-MP or vigabatrin treat-ment in bilateral FN affects T, B, and NK cells. These findings revealthat FN GABAergic projections to the hypothalamus regulate theimmune system and also imply that FN–hypothalamic GABAergicprojections are a route by which the cerebellum modulates im-mune response.

Conflict of interest statement

All authors declare that there are no conflicts of interest.

Acknowledgments

This work was supported by grants 30870819 and 81271323from the National Natural Science Foundation of China,BK2010278 and BK2011386 from the Natural Science Foundationof Jiangsu Province of China, CP12012003, BK2012014, BK2012015and K2009047 from the Nantong Applied Research Program ofChina,

and a project funded by the Priority Academic Program Develop-ment of Jiangsu Higher Education Institutions.

References

Alexander, M.P., Gillingham, S., Schweizer, T., Stuss, D.T., 2012. Cognitiveimpairments due to focal cerebellar injuries in adults. Cortex 48, 980–990.

Baciu, I., Hriscu, M., Saulea, G., 2003. Hypothalamic mechanisms of immunity. Int. J.Neurosci. 113, 259–277.

Cavdar, S., Onat, F., Aker, R., Sehirli, U., San, T., Yananli, H.R., 2001a. The afferentconnections of the posterior hypothalamic nucleus in the rat using horseradishperoxidase. J. Anat. 198, 463–472.

Cavdar, S., San, T., Aker, R., Sehirli, U., Onat, F., 2001b. Cerebellar connections to thedorsomedial and posterior nuclei of the hypothalamus in the rat. J. Anat. 198,37–45.

Chung, S.H., Marzban, H., Hawkes, R., 2009. Compartmentation of the cerebellarnuclei of the mouse. Neuroscience 161, 123–138.

Crick, E.W., Osorio, I., Bhavaraju, N.C., Linz, T.H., Lunte, C.E., 2007. An investigationinto the pharmacokinetics of 3-mercaptopropionic acid and development of asteady-state chemical seizure model using in vivo microdialysis andelectrophysiological monitoring. Epilepsy Res. 74, 116–125.

Dietrichs, E., Haines, D.E., 1984. Demonstration of hypothalamo-cerebellar andcerebello-hypothalamic fibers in a prosimian primate (Galago crassicaudatus).Anat. Embryol. 170, 313–318.

Dietrichs, E., Haines, D.E., Roste, G.K., Roste, L.S., 1994. Hypothalamocerebellar andcerebellohypothalamic projections: circuits for regulating nonsomaticcerebellar activity? Histol. Histopathol. 9, 603–614.

Dong, J., Mrabet, O., Moze, E., Li, K., Neveu, P.J., 2002. Lateralization andcatecholaminergic neuroimmunomodulation: prazosin, an alpha1/alpha2-adrenergic receptor antagonist, suppresses interleukin-1 and increasesinterleukin-10 production induced by lipopolysaccharides.Neuroimmunomodulation 10, 163–168.

Engel, D., Pahner, I., Schulze, K., Frahm, C., Jarry, H., Ahnert-Hilger, G., Draguhn, A.,2001. Plasticity of rat central inhibitory synapses through GABA metabolism. J.Physiol. 535, 473–482.

Eskandari, F., Webster, J.I., Sternberg, E.M., 2003. Neural immune pathways andtheir connection to inflammatory diseases. Arthritis Res. Ther. 5, 251–265.

Ghoshal, D., Sinha, S., Sinha, A., Bhattacharyya, P., 1998. Immunosuppressive effectof vestibulo–cerebellar lesion in rats. Neurosci. Lett. 257, 89–92.

Green-Johnson, J.M., Zalcman, S., Vriend, C.Y., Nance, D.M., Greenberg, A.H., 1995.Suppressed T cell and macrophage function in the ‘‘reeler’’(rl/rl) mutant, amurine strain with elevated cerebellar norepinephrine concentration. BrainBehav. Immun. 9, 47–60.

Haddad, J.J., Saade, N.E., Safieh-Garabedian, B., 2002. Cytokines and neuro-immune-endocrine interactions: a role for the hypothalamic–pituitary–adrenal revolvingaxis. J. Neuroimmunol. 133, 1–19.

Haines, D.E., Dietrichs, E., 1984. An HRP study of hypothalamo-cerebellar andcerebello-hypothalamic connections in squirrel monkey (Saimiri sciureus). J.Comp. Neurol. 229, 559–575.

Haines, D.E., Sowa, T.E., Dietrichs, E., 1985. Connections between the cerebellumand hypothalamus in the tree shrew (Tupaia glis). Brain Res. 328, 367–373.

Haines, D.E., May, P.J., Dietrichs, E., 1990. Neuronal connections between thecerebellar nuclei and hypothalamus in Macaca fascicularis: cerebello–visceralcircuits. J. Comp. Neurol. 299, 106–122.

Haines, D.E., Dietrichs, E., Mihailoff, G.A., McDonald, E.F., 1997. The cerebllar–hypothalamic axis: basic circuits and clinical observations. Int. Rev. Neurobiol.41, 83–107.

Hall, W.C., 2004. Movement and its central control. In: Purves, D., Augustine, G.J.,Fitzpatrick, D., Hall, W.C., LaMantia, A.S., McNamara, J.O., Williams, S.M. (Eds.),Neuroscience. Sinauer Associates, Inc., Massachusetts, pp. 435–452.

Holmes, M.J., Cotter, L.A., Arendt, H.E., Cass, S.P., Yates, B.J., 2002. Effects of lesions ofthe caudal cerebellar vermis on cardiovascular regulation in awake cats. BrainRes. 938, 62–72.

Ito, M., 2002. Historical review of the significance of the cerebellum and the role ofPurkinje cells in motor learning. Ann. N. Y. Acad. Sci. 978, 273–288.

Jackson, M.F., Esplin, B., Capek, R., 2000. Reversal of the activity-dependentsuppression of GABA-mediated inhibition in hippocampal slices from c-vinylGABA (vigabatrin)-pretreated rats. Neuropharmacology 39, 65–74.

Köbbert, C., Apps, R., Bechmann, I., Lanciego, J.L., Mey, J., Thanos, S., 2000. Currentconcepts in neuroanatomical tracing. Prog. Neurobiol. 62, 327–351.

Kwan, P., Sills, G.J., Brodie, M.J., 2001. The mechanisms of action of commonly usedantiepileptic drugs. Pharmacol. Ther. 90, 21–34.

Lu, J.H., Mao, H.N., Cao, B.B., Qiu, Y.H., Peng, Y.P., 2012. Effect ofcerebellohypothalamic glutamatergic projections on immune function.Cerebellum Online: http://dx.doi.org/10.1007/s12311-012-0356-8.

Madden, K.S., 2003. Catecholamines, sympathetic innervation, and immunity. BrainBehav. Immun. 17 (suppl 1), S5–S10.

Mausset, A.L., de Seze, R., Montpeyroux, F., Privat, A., 2001. Effects of radiofrequencyexposure on the GABAergic system in the rat cerebellum: clues from semi-quantitative immunohistochemistry. Brain Res. 912, 33–46.

Min, B.I., Oomura, Y., Katafuchi, T., 1989. Responses of rat lateral hypothalamicneuronal activity to fastigial nucleus stimulation. J. Neurophysiol. 61, 1178–1184.

Nance, D.M., Sanders, V.M., 2007. Autonomic innervation and regulation of theimmune system (1987–2007). Brain Behav. Immun. 21, 736–745.

90 B.-B. Cao et al. / Brain, Behavior, and Immunity 27 (2013) 80–90

Netopilová, M., Drsata, J., Kubová, H., Mares, P., 1995. Differences betweenimmature and adult rats in brain glutamate decarboxylase inhibition by 3-mercaptopropionic acid. Epilepsy Res. 20, 179–184.

Onat, F., Cavdar, S., 2003. Cerebellar connections: hypothalamus. Cerebellum2, 263–269.Paxinos, G., Watson, C., 1998. The Rat Brain in Stereotaxic Coordinates, 4th ed.

Academic Press, San Diego.Peng, Y.P., Qiu, Y.H., Cao, B.B., Wang, J.J., 2005. Effect of lesions of cerebellar fastigial

nuclei on lymphocyte functions of rats. Neurosci. Res. 51, 275–284.Peng, Y.P., Qiu, Y.H., Qiu, J., Wang, J.J., 2006. Cerebellar interposed nucleus lesions

suppress lymphocyte function in rats. Brain Res. Bull. 71, 10–17.Pittman, Q.J., 2011. A neuro-endocrine-immune symphony. J. Neuroendocrinol. 23,

1296–1297.Sanders, V.M., Kohm, A.P., 2002. Sympathetic nervous system interaction with the

immune system. Int. Rev. Neurobiol. 52, 17–41.Sultan, F., Konig, T., Mock, M., Their, P., 2002. Quantitative organization of

neurotransmitters in the deep cerebellar nuclei of the Lurcher mutant. J.Comp. Neurol. 452, 311–323.

Tayebati, S.K., Bronzetti, E., Morra Di Cella, S., Mulatero, P., Ricci, A., Rossodivita, I.,Schena, M., Schiavone, D., Veglio, F., Amenta, F., 2000. In situ hybridization andimmunocytochemistry of alpha 1-adrenoceptors in human peripheral bloodlymphocytes. J. Auton. Pharmacol. 20, 305–312.

Thürling, M., Hautzel, H., Küper, M., Stefanescu, M.R., Maderwald, S., Ladd, M.E.,Timmann, D., 2012. Involvement of the cerebellar cortex and nuclei in verbal andvisuospatial working memory: A 7T fMRI study. Neuroimage 62, 1537–1550.

Trotter, R.N., Stornetta, R.L., Guyenet, P.G., Roberts, M.R., 2007. Transneuronalmapping of the CNS network controlling sympathetic outflow to the rat thymus.Auton. Neurosci. 131, 9–20.

Uusisaari, M., Knöpfel, T., 2008. GABAergic synaptic communication in theGABAergic and non-GABAergic cells in the deep cerebellar nuclei.Neuroscience 156, 537–549.

Wang, J.J., Pu, Y.M., Wang, T., 1997. Influences of cerebellar interpositus nucleus andfastigial nucleus on the neuronal activity of lateral hypothalamic area. Sci.China, Ser. C Life Sci. 40, 176–183.

Wang, F., Cao, B.B., Liu, Y., Huang, Y., Peng, Y.P., Qiu, Y.H., 2011. Role ofcerebellohypothalamic GABAergic projection in mediating cerebellarimmunomodulation. Int. J. Neurosci. 121, 237–245.

Wrona, D., 2006. Neural–immune interactions: an integrative view of thebidirectional relationship between the brain and immune systems. J.Neuroimmunol. 172, 38–58.

Wrona, D., Trojniar, W., 2003. Chronic electrical stimulation of the lateralhypothalamus increases natural killer cell cytotoxicity in rats. J. Neuroimunol.141, 20–29.

Wrona, D., Jurkowski, M., Luszawska, D., Tokarski, J., Trojniar, W., 2003. The effectsof lateral hypothalamic lesions on peripheral blood natural killer cellcytotoxicity in rats hyper- and hyporesponsive to novelty. Brain Behav.Immun. 17, 453–461.

Zhang, Y.P., Ma, C., Wen, Y.Q., Wang, J.J., 2003. Convergence of gastric vagal andcerebellar fastigial nuclear inputs on glycemia-sensitive neurons of lateralhypothalamic area in the rat. Neurosci. Res. 45, 9–16.

Zhu, J.N., Wang, J.J., 2008. The cerebellum in feeding control: possible function andmechanism. Cell Mol. Neurobiol. 28, 469–478.

Zhu, J.N., Yung, W.H., Kwok-Chong Chow, B., Chan, Y.S., Wang, J.J., 2006. Thecerebellar–hypothalamic circuits: potential pathways underlying cerebellarinvolvement in somatic–visceral integration. Brain Res. Rev. 52, 93–106.

Zhuang, J., Xu, F., Frazier, D.T., 2008. Hyperventilation evoked by activation of thevicinity of the caudal inferior olivary nucleus depends on the fastigial nucleus inanesthetized rats. J. Appl. Physiol. 104, 1351–1358.