The Journal of Neuroscience, February 1988, 8(2): 543654 Corticotropin-Releasing Factor in Cerebellar Afferent Systems: A Combined lmmunohistochemistry and Retrograde Transport Study Sharon Cummings,’ Burt Sharp,* and Robert Eldel Departments of ‘Cell Biology and Neuroanatomy and *Medicine, The University of Minnesota, Minneapolis, Minnesota 55455 The flocculus and paraflocculus of cat and sheep cerebellum to other hypophysiotropic hormones such as leuteinizing hor- were studied with immunohistochemical methods, using mone-releasinghormone (Witkin et al., 1982), thyrotropin-re- antisera to corticotropin-releasing factor (CRF). CRF im- leasing hormone (Johanssonand Hokfelt, 1980), growth hor- munoreactivity was present within 3 populations of varicose mone-releasingfactor (Sawchenkoet al., 1985),and somatostatin nerve fibers. One population of CRF-immunoreactive (CRF- (Johanssonet al., 1984), was present not only within hypophys- IR) fibers appeared to appose Purkinje cell somata and to iotropic neurons, but also occurred in widely distributed neu- follow their dendrites into the molecular layer. This arrange- ronal systems(Fischman and Moldow, 1982; Olschowka et al., ment suggested they were climbing fibers. A second group 1982; Cote et al., 1983; Cummingset al., 1983; Schipper et al., of CRF-IR profiles reminiscent of mossy fibers was widely 1983; Swansonet al., 1983; Merchenthaler, 1984). distributed throughout the granule cell layer. A third popu- In previous studies we reported the presenceof CRF-im- lation of CRF-IR fibers was present as a beaded plexus lying munoreactive (CRF-IR) fibers in rat cerebellar cortex (Cum- parallel to the pial surface, above and subadjacent to the mings et al., 1983) and in the flocculus and paraflocculusof cat Purkinje cell layer. The fibers of this plexus extended into and monkey cerebellum (Cummings et al., 1985). However, the Purkinje cell layer and surrounded these somata. reports conflict on the existenceof CRF immunoreactivity with- The source of some of the CRF-IR fibers within the floc- in the cerebellum. Other immunohistochemical studies have culus and paraflocculus was determined by a retrograde reported CRF fibers in rat cerebellum (Olschowka et al., 1982; axonal transport study utilizing the fluorescent tracer Fast Merchenthaler, 1984), and in the anterior vermis of human blue (FB) in combination with the immunohistochemical lo- cerebellum (Powers et al., 1986), but somehave been unable to calization of CRF. It was determined that CRF-IR perikarya verify this (Swanson et al., 1983). Varying levels of CRF im- within the inferior olivary nucleus gave rise to a population munoreactivity in the cerebellum have been reported in ra- of climbing fibers within those lobules. Furthermore, all di- dioimmunoassay studies (Fischman and Moldow, 1982; Cote visions of the inferior olive were found to contain CRF-IR et al., 1983), but other investigations have been unable to detect somata. This latter finding suggests the potential for CRF- the CRF immunoreactivity in that region by radioimmunoassay IR climbing fiber projections from the inferior olive to other (Palkovits et al., 1985). A radioimmunoassaystudy utilizing 4 regions of the cerebellar cortex. The existence of CRF-IR antisera directed against either the rat or ovine sequencesde- mossy fibers and fibers within the ganglionic plexus sug- tected CRF-IR in cerebellarextracts only with the ovine antisera gests the possibility of CRF-IR afferent projections from oth- (Skofitsch and Jacobowitz, 1985). Recently, however, specific er regions of the brain stem to the flocculus and parafloc- binding sites for lZ51-Tyr-oCRF (Wynn et al., 1984) and for culus. Nle21,‘251-Tyr32-oCRF(De Souza et al., 1986), were identified within the cerebellar cortex. In the latter study, the specific The 4 1 amino acid peptide, corticotropin-releasingfactor (CRF), binding of the radiolabeled CRF receptor ligand was reported sequencedand synthesizedby Vale and colleaguesin 1981 (Spiess to be higher in the cerebellum than in several other areasof the et al., 1981; Vale et al., 1981) was localized immunohisto- rat CNS. The existence of CRF binding sitesin the cerebellum chemically in the hypothalamus soon after its discovery (Bloom strongly suggestsa functional role for a CRF-like peptide within et al., 1982; Paul1et al., 1982; Cummings et al., 1983; Swanson that region. et al., 1983). However, it becameapparent that CRF, similar The present study was undertaken to clarify discrepancieson the existence of CRF within the cerebellum. We had observed consistent populations of CRF-IR fibers in the flocculus and Received Jan. 20, 1987; revised June 29, 1987; accepted July 21, 1987. paraflocculus, and therefore restricted this study to those re- This work was supported in part by a doctoral dissertation fellowship from the gions. Hypothesizing that CRF-IR fiber populationswithin those Graduate School of the The University of Minnesota (SC.). We wish to thank lobules might arise from extrinsic sources,a retrograde tracing Dr. Kenneth E. Miller for helpful advice, Dr. Virginia Seybold for critical reading of the manuscript, and Dr. Georgia Bishop for assistance in interpretation of data study was undertaken, in combination with immunohistochem- from the cat inferior olive. ical localization of CRF, in order to ascertain possiblesources. Corresoondence should be addressed to Sharon Cummines. Ph.D.. Denartment of Anatomy and Neuroscience Program, The Ohio State U%ersity,‘4072 Graves The focus of this initial double-label study was directed pri- Hall, 333 W. 10th Ave., Columbus, OH 43210. marily to the inferior olive, the only known sourceof cerebellar Copyright 0 1988 Society for Neuroscience 0270-6474/88/020543-12$02.00/O climbing fibers. 544 Cummings et al. - CRF in Cerebellar Afferent Systems ebellum is wholly crossed. Therefore, the fluorescent dye was trans- Materials and Methods ported only to the left side. The inferior olivary nuclei on both sides of Localization of CRF in flocculus and paraflocculus. Cats (body weight the brain stem were observed and compared with bright-field micros- l-2 kg; n = 5) and lambs (n = 2) were included in this study. The copy in order to determine that the pattern of staining for CRF im- animals were obtained from the Research Animal Resources Division munoreactivity was identical and that the appearance or specificity of of the University of Minnesota or from other ongoing investigations. immunostaining on the left side, to which there was FB transport, was All animals were anesthetized with sodium pentobarbital and perfused not altered by the presence of the fluorescent dye. transcardially with saline followed by a fixative of picric acid/phosphate Characterization ofantisera. The lamb tissue was incubated overnight (0.1 M) buffered 2% paraformaldehyde (pH 7.3) (Stephanini et al., 1967) in ovine CRF primary antiserum (l/500) (ImmunoNuclear Corp., Still- at room temperature. The brains were removed, immersed in the same water, MN) characterized and described previously (Cummings et al., fixative for 90 min and stored overnight in 5% sucrose in 0.1 M Sor- 1983). The cat tissue was incubated in a primary antiserum (l/5000) enson’s phosphate buffer, pH 7.3 The flocculus and paraflocculus were directed against rat/human CRF produced in rabbits with a synthetic dissected free from the cerebellum. The tissues were treated by 1 of 2 rat CRF-keyhole limpet hemocyanin conjugate. Controls for the spec- procedures: (1) the individual lobules were sectioned at 50 or 75 pm on ificity of immunohistochemical localizations of this rat CRF antiserum a sliding microtome and processed by the peroxidase-antiperoxidase were performed by incubating sections with antiserum that had been (PAP) method of Stemberger (1979) modified for free-floating sections pretreated overnight at 4°C with an excess of synthetic rat CRF (10 r~pl as previously described (Cummings et al., 1983) or (2) the lobules were ml of diluted CRF antiserum). All of the staining observed was found embedded together within a single block of brain paste in an orientation to represent immunohistochemically specific localizations because no to permit sectioning of the lobules in the transverse or sagittal plane, staining was found in structures on sections that served as absorption and 10 pm sections of the frozen tissue block were cut in a cryostat. controls. Furthermore, the PAP procedure was conducted on filter paper Slide mounted sections were hydrated with PBS and processed for the (Whatman No. 1) upon which various synthetic peptides had been indirect immunofluorescence technique of Coons (1958) as modified immobilized with’vapor-phase paraformalhehyde fixation according to and ureviouslv described for CRF (Onstott and Elde. 1984). Sections the method of Larsson (198 1). The peptides tested included ovine CRF, were counterstained with ethidium bromide, a fluorescent Nissl stain rat CRF, Lys*-vasopressin, Arg8-vasopressin, oxytocin, dynorphin,,-,,,, (Schmued et al., 1982) in order to define Purkinje cell and granule cell Met-enkephalin, and Leu-enkephalin. Each peptide was tested at quan- somata. Sections were mounted in a fade-retarding medium (Platt and tities varying from 10 nmol to 10 pmol/blot. The antiserum to rat CRF Michael, 1983) examined and photographed using a Zeiss standard stained blots containing as little as 10 pmol of rat or ovine CRF. In fluorescence microscope fitted with epi-illumination. Fluorescein-iso- contrast, this antiserum did not stain blots containing 10