Circulating autoantibodies recognize and bind dying neurons following injury to the brain

While it is known that autoimmune cells can protect against cell damage following traumatic injury of the brain, the role of autoantibodies in brain injury is less clear. Here we present evidence in adult rats that following a cortical lesion of the brain, circulating IgG autoantibodies bind to dying neurons in the vicinity of the lesion. At intervals that ranged from 4 h to 7 days after making a unilateral lesion of visual cortex, we observed neurons near the lesion that were immunopositive for rat IgG. Many of these IgG-positive neurons were in advanced stages of degeneration. The magnitude of the immunostaining observed was directly proportional to the percent reactivity to rat IgG of the antibodies that were used. Preadsorption of the antibodies with rat serum eliminated the immunostaining. In addition, immunostaining for serum albumin in sections through the cortical lesion was negative, supporting the conclusion that the positive staining for IgG does not result from the passive diffusion of serum proteins into injured cells. Instead, the evidence presented here strongly suggests that naturally occurring IgG autoantibodies bind specifically to dying neurons in the injured brain. We propose that this autoantibody binding may participate in the phagocytosis and removal of injured neurons.     

Immunocytochemical localization of glutamate receptor subunits in the brain stem and cerebellum of the turtle Chrysemys picta

The regional distribution of ionotropic (AMPA and NMDA) and metabotropic (mGluR1alpha) glutamate receptor subunits was examined in the brain stem and cerebellum of the pond turtle, Chrysemys picta, by using immunocytochemistry and light microscopy. Subunit-specific antibodies that recognize NMDAR1, GluR1, GluR4, and mGluR1alpha were used to identify immunoreactive nuclei in the brain stem and cerebellum. Considerable immunoreactivity in the turtle brain stem and cerebellum was observed with regional differences occurring primarily in the intensity of staining with the antibodies. The red nucleus, lateral reticular nucleus and cerebellum labeled intensely for NMDAR1 and moderately for GluR1. The cerebellum also labeled strongly for mGluR1alpha. All of the cranial nerve nuclei labeled intensely for NMDAR1 and to varying degrees for GluR1, GluR4, and mGluR1alpha. Counterstaining revealed the presence of neuronal somata where there were no immunoreactive neurons in individual nuclei. This finding suggests that there are subpopulations of immunoreactive neurons within a given nucleus that bear different glutamate receptor subunit compositions. The results suggest that the glutamate receptor subunit distribution in the brain stem and cerebellum of turtles is similar to that reported for rats. Additionally, there is considerable colocalization of NMDA and AMPA receptors as revealed by light microscopy. These results have implications for the organization of neural circuits that control motor behavior in turtles, and, generally, for the function of brain stem and cerebellar neural circuits in vertebrates.     

Recognition by anti-idiotypic anti-thyrotropin-releasing hormone (TRH) antibody of rat TRH receptors.

We asked whether anti-idiotypic antibodies raised against anti-TRH antibody could bind to TRH receptors in the rat anterior pituitary and brain. Six rabbits were immunized with IgG from a rabbit anti-TRH antiserum. One anti-idiotypic antibody caused strong, dose-dependent inhibition in anti-TRH antibody-binding to [125]I-TRH. This inhibition was not observed after treatment with goat anti-rabbit IgG antibody. The anti-idiotypic anti-TRH antibody significantly immunoprecipitated digitonin-solubilized pituitary TRH receptors. When eluates of digitonin-solubilized membranes which were adsorbed by either an anti-idiotypic anti-TRH IgG-, normal rabbit IgG-linked affinity column or control column were analyzed by sodium dodecyl sulfate polyacrylamide gel electrophoresis and visualized by silver stain, only the former eluate showed two bands under nonreducing conditions; one corresponded to a molecular weight marker of 200K, the other to 100K. Western blotting analysis with an anti-idiotypic anti-TRH antiserum showed a single band of molecular weight 56K under reducing conditions. The present study indicates that one can make anti-idiotypic antibodies that specifically recognize TRH receptors.     

Subcellular, regional and immunohistochemical localization of adenosine deaminase in various species

Immunohistochemical and subcellular fractionation techniques were employed to compare the cellular and subcellular localization of adenosine deaminase (ADA) in various brain regions of several mammalian species. A relatively restricted distribution of ADA-immunoreactive neurons in rat brain was previously reported. Mouse brain exhibited a pattern similar in many respects to rat and, in addition, contained intensely immunostained neurons in lateral habenula and hippocampus. Glial immunostaining was absent or very light in rat but evident in mouse. Prominent immunoreactive fibers and neurons were observed in hamster spinal cord and anterior hypothalamus, respectively. ADA-immunostaining in guinea-pig was localized to presumptive fibers in the superficial layers of spinal cord dorsal horn and to glial cells throughout the brain. Demonstration of specific immunostaining in rabbit was not possible. ADA activity was far more heterogeneously distributed in rat and most brain areas in guinea-pig and rabbit contained up to 5-fold and 10-fold higher levels of activity, respectively, compared with rat. Crude synaptosomal (P2) fractions of rat cortex contained a greater proportion of ADA activity than those of rabbit cortex. Within rat, relatively high activity was found in P2 fractions of whole hypothalamus, cerebellum, and hippocampus. ADA activity was greater in P2 fractions of rat anterior compared with whole hypothalamus and the greatest proportion of the enzyme in this fraction was localized to purified synaptosomes. The large variations in the activity and cellular location of ADA in the animals examined suggest species differences in mechanisms governing adenosine metabolism in brain and possible differences in the relationships between cellular metabolism, ADA and the neuroregulatory role of adenosine in the CNS.     

Estrogen receptor-immunostaining of neuronal cytoplasmic processes as well as cell nuclei in guinea pig brain

We have recently developed an immunocytochemical technique for staining estrogen receptors in cell nuclei of some cells in the guinea pig brain. With optimization of this technique, we have now been successful in staining estrogen receptor-immunoreactivity in processes of many neurons in the guinea pig brain that also contain estrogen receptor-immunoreactive cell nuclei. While reaction product is visible in cell nuclei under a wide variety of conditions, neuronal processes are darkly immunostained only with modifications of the procedure optimized for maximum sensitivity, for example, the multiple-bridge immunocytochemical procedure. Omission of Triton X-100 and/or dimethylsulfoxide, usually used to increase penetration of the antibodies, had no effect on immunostaining in processes or cell nuclei, and immunostaining was apparent in both vibratome-cut and freezing microtome-cut sections. Injection of estradiol, but not progesterone, caused the total loss of cytoplasmic estrogen receptor-immunostaining, consistent with the idea that the cytoplasmic immunostaining, like the cell nuclear immunostaining is due to the presence of estrogen receptors. In all neuroanatomical regions containing estrogen receptor-immunoreactive cell nuclei, associated processes of some, but not all cells were also immunostained. However, in certain areas, such as the midbrain central gray and the preoptic area, cytoplasmic staining was particularly dark. The cellular characteristics that result in immunostaining in some estrogen receptor-immunoreactive cells, and not in others is under investigation.     

Use of calcium-binding proteins to map inputs in vestibular nuclei of the gerbil

We wished to determine whether calbindin and/or calretinin are appropriate markers for vestibular afferents, a population of neurons in the vestibular nuclear complex, or cerebellar Purkinje inputs. To accomplish this goal, immunocytochemical staining was observed in gerbils after lesions of the vestibular nerve central to the ganglion, the cerebellum, or both. Eleven to fourteen days after recovery, the brain was processed for immunocytochemical identification of calretinin and calbindin. After lesion of the vestibular nerve, no calretinin staining was seen in any of the vestibular nuclei except for a population of intrinsic neurons, which showed no obvious change in number or staining pattern. Calbindin staining was reduced in all nuclei except the dorsal part of the lateral vestibular nuclei. The density of staining of each marker, measured in the magnocellular medial vestibular nucleus, was significantly reduced. After the cerebellar lesion, no differences in calretinin staining were noted. However, calbindin staining was greatly reduced in all nuclei. The density of staining, measured in the caudal medial vestibular nucleus, was significantly lower. After a combined lesion of the cerebellum and vestibular nerve, the distribution and density of calretinin staining resembled that after vestibular nerve section alone, whereas calbindin staining was no longer seen. This study demonstrates that calretinin and calbindin are effective markers for the identification of vestibular afferents. J. Comp. Neurol. 386:317–327, 1997. © 1997 Wiley-Liss, Inc.     

IgG Leakage May Contribute to Neuronal Dysfunction in Drug-Refractory Epilepsies With Blood-Brain Barrier Disruption

ABSTRACT: Focal epilepsies are often associated with blood-brain barrier disruption. In 4 entorhinal cortex tissue samples and 13 hippocampal samples from patients with pharmacoresistent temporal lobe epilepsy, we observed immunoglobulin G (IgG) leakage in the parenchyma and IgG-positive neurons that had evidence of neurodegeneration, such as shrinkage and eosinophilia. These findings were not present in samples from 12 nonepileptic control subjects. To complement these findings, we used a rat in vivo model that mimics the development of limbic epilepsy with blood-brain barrier disruption. During epileptogenesis, IgG leakage and neuronal IgG uptake increased concomitantly with the occurrence of seizures. Immunoglobulin G accumulation in neurons was selective, particularly for interneurons and pyramidal neurons. Immunohistochemistry and electron microscopy showed that IgG uptake in the rat neurons was associated with eosinophilia, shrinkage, and ultrastructural degenerative changes. Moreover, IgG-positive neurons lost their NeuN immunohistochemical staining. Together, these observations suggest that IgG leakage is related to neuronal impairment and may be a pathogenic mechanism in epileptogenesis and chronic epilepsy.     

Distribution of substance P-like Immunoreactivity in the chameleon brain

The distribution of substance P-like immunoreactivity in the chameleon brain and spinal cord was studied with immunohistochemical methods using polyclonal antibodies against substance P. In the telencephalon, immunoreactive cell bodies and fibers were located primarily in the striatum and in the globus pallidus. In addition, few substance P-like fibers were observed in the cortical areas, in the septum, and in the amygdala. In the diencephalon, a high density of immunostained neurons and fibers were seen in the periventricular and ventrolateral hypothalamus. Another group of cell bodies was located in the optic tectum and particularly in the stratum griseum central. A large number of immunoreactive fibers were also detected in the thalamic nuclei and in the median eminence. In the mesencephalon, few immunoreactive neurons were observed in the ventral tegmental area, in the substantia nigra, and in the nucleus reticularis isthmi. These latter nuclei, the periventricular area, the posterior commissure, the nucleus lentiformis mesencephali, the oculomotor nucleus, and the raphe nuclei contained a dense plexus of substance P immunoreactive fibers. No immunoreactive cell bodies were observed in raphe nuclei. In the spinal cord, no substance P-like immunoreactive neurons were observed, but a large number of substance P immunostained fibers were seen in the dorsal and lateral part of the dorsal horn and surrounding the dorsal parts of the central canal. The results of the present study are discussed with respect to those obtained in other species of reptiles, the main differences concerning the lateral septum, the habenula, the area of the paraventricular organ, and the raphe nuclei.     

Immunohistochemical study of microtubule-associated protein 2 and ubiquitin in chronically aluminum-intoxicated rabbit brain

Summary Experimental neurofibrillary change was produced in rabbit brains by daily subcutaneous aluminum tartrate injection for 40 days. The production of experimental neurofibrillary changes was confirmed by immunostaining with antibodies against neurofilament triplet proteins and the brain tissue was studied immunohistochemically with antibodies against microtubule-associated protein (MAP) 2 and ubiquitin. The hippocampal neurons of the chronically aluminum-intoxicated rabbit brain showed diminished staining of dendrites by anti-MAP2 antibody. The length of anti-MAP2-positive dendrites in hippocampus was significantly shorter than that of the control brain. In the cortex somata of a subset of pyramidal neurons were intensively stained by anti-MAP2 antibody, while the MAP2 immunoreactivity of distal dendrites was diminished. The immunostaining by anti-ubiquitin antibody revealed the positive staining of the neurons bearing experimental neurofibrillary changes in the lower brain stem nuclei. It is speculated that MAP2 dislocation and ubiquitination are accompanying phenomena of the production of experimental neurofibrillary changes in chronically aluminum-intoxicated rabbit brains.Supported in part by grants from Ministry of Education of Japan and the Sandoz Gerontological Research Foundation     

Purification and characterization of rabbit tissue factor

Anti-human tissue factor IgG column was used to partially purify rabbit tissue factor. The rabbit tissue factor was then further purified on SDS-polyacrylamide gel electrophoresis to prepare an essentially homogeneous rabbit tissue factor. This was used as an antigen to raise a high titer, monospecific, polyclonal antiserum for rabbit tissue factor. Immunostaining of a crude Triton extract of rabbit brain acetone powder and purified tissue factor with the anti-rabbit tissue factor IgG yielded a single major band with an apparent molecular weight of 45 kD, which corresponds to the mobility on SDS-PAGE of purified rabbit tissue factor apoprotein. The anti-rabbit tissue factor IgG was coupled to Affi-Gel-15 and used as an immunoadsorbent column to purify large amounts of rabbit brain tissue factor in a rapid, one-step technique.     

Immunohistochemical Localization of Kininogen in Rat Spinal Cord and Brain

Kininogen localization has been determined by immunocytochemistry in rat spinal cord and brain using a kinin-directed kininogen monoclonal antibody. In the spinal cord, there were immunostained neurons and fibers in laminae I, II, VII, and IX, intensely stained fibers in the superficial layers of the dorsal horn, and immunoreactive glial and endothelial cells. Small neurons, satellite cells, and Schwann cells immunostained distinctly in the dorsal root ganglion. In the brain stem, there were immunoreactive neurons and fibers in the tractus solitarius and nucleus, trigeminal spinal tract and nuclei, periaqueductal gray matter, vestibular nuclei, cochlear nuclei, trapezoid body, medial geniculate nucleus, and red nucleus. Immunostained neurons and fibers were also found in cerebellum (dentate nucleus), cerebral cortex (layers III and V), hippocampus (pyramidal cell layer), and corpus callosum. Glia and endothelial cells stained in all brain regions. The widespread location of kininogen in neurons and their processes, as well as in glial and endothelial cells, indicates more than one functional role, including those proposed as a mediator, a calpain inhibitor, and a kinin precursor, in a variety of neural activities and responses.     

Identification of PSF, the polypyrimidine tract-binding protein-associated splicing factor, as a developmentally regulated neuronal protein.

The polypyrimidine tract-binding protein-associated splicing factor (PSF), which plays an essential role in mammalian spliceosomes, has been found to be expressed by differentiating neurons in developing mouse brain. The sequence of a fragment of mouse PSF was found to be remarkably similar to that of human PSF. Both the expression of PSF mRNA in cortex and cerebellum and PSF immunoreactivity in all brain areas were high during embryonic and early postnatal life and almost disappeared in adult tissue, except in the hippocampus and olfactory bulb where various neuronal populations remained PSF-immunopositive. Double-labeling experiments with anti-PSF antibody and anti-neurofilaments or anti-glial fibrillary acidic protein antibodies on sections of cortex, hippocampus, and cerebellum indicate that PSF is expressed by differentiating neurons but not by astrocytic cells. In vitro, mouse PSF was found to be expressed by differentiating cortical and cerebellar neurons. Radial glia or astrocyte nuclei were not immunopositive; however, oligodendrocytes differentiating in vitro were found to express PSF. The restricted expression of PSF suggests that this splicing factor could be involved in the control of neuronal-specific splicing events occurring at particular stages of neuronal differentiation and maturation.     

Rabbit IgG distribution in skin, spinal cord and DRG following systemic injection in rat

In order to determine the distribution of antibodies such as anti-NGF following systemic injection in neonates, immunocytochemical techniques were used to examine the localization of rabbit IgG in rat skin, DRG, and spinal cord after treatments with normal rabbit serum or purified rabbit IgG. Daily subcutaneous injections beginning on postnatal day 2 or on day 15 were given for three days. On the fourth day the animals were sacrificed and tissues were processed for rabbit IgG-IR. In the dorsal and ventral spinal cord, staining intensities suggest a substantial increase in the blood–brain barrier during the first two weeks after birth. Staining intensity in the epidermis of the glabrous skin from the hindpaw was substantially lower than in the adjacent dermis. In addition, IgG infrequently accumulated intracellularly in intensely stained patches in the epidermis. IgG was also able to reach relatively high intracellular concentrations in a small number of sensory neurons. The IgG staining pattern in the skin was similar when anti-NGF itself was administered to the animals. The results are discussed in the context of the effects of anti-NGF on the development of nociceptive afferents.     

Dual enhancement of triple immunofluorescence using two antibodies from the same species

Triple immunofluorescence method with two mouse monoclonal antibodies and another rabbit polyclonal antibody was established with catalyzed reporter deposition (CARD) amplification on thick floating sections from the rat cerebellum. One of the monoclonal antibodies (anti-calbindin), diluted maximally, probed with anti-mouse IgG-horseradish peroxidase (HRP) and amplified with Cy5-conjugated tyramide, immunolabeled cerebellar Purkinje cells and their arborization. Subsequently, a rabbit polyclonal IgG (anti-glial fibrillary acidic protein (anti-GFAP)), probed with anti-rabbit IgG-HRP, amplified with biotin-tyramide and visualized with fluorescein-isothiocyanate (FITC)-streptavidin, immunolabeled Bergmann's glia. Another mouse monoclonal IgG (anti-SNAP25), probed with anti-mouse IgG-rhodamine without CARD amplification, selectively visualized synaptic sites, because the maximal dilution of the other monoclonal antibody (anti-calbindin) was below the detection threshold of this anti-mouse IgG-rhodamine. Separation of the two signals (calbindin and SNAP25), each detected through mouse monoclonal antibody, was then based on the difference of sensitivity either with or without CARD amplification. Triple immunofluorescence is possible when just one of the three primary antibodies is from different species. Intensification of two of the three signals provides further advantages to examine immunolocalization of multiple epitopes on histological sections.     

Immunoperoxidase labelling of rat brain sections with sera from patients with paraneoplastic cerebellar degeneration and systemic neoplasia

Sera from patients with systemic cancer found by immunofluorescence staining to have antibodies to human cerebellar cell populations were reacted with vibratome sections of rat cerebellum and examined by peroxidase-antiperoxidase (PAP) methods. Seven patients with clinically or pathologically confirmed paraneoplastic cerebellar degeneration and two neurologically normal patients with high titers of anticerebellar antibodies were studied. Sera from all antibody-positive patients, but not from controls, produced intense staining of brain sections. Sera from patients with ovarian adenocarcinoma reacted predominantly with Purkinje cells and neurons within brainstem nuclei. Sera from patients with oat cell carcinoma and one patient with ductal carcinoma of the breast produced nuclear and cytoplasmic staining of neurons throughout the central nervous system. Serum from a patient with Hodgkin's disease labeled the peripheries of Purkinje cells and Golgi II cells. Serum from a patient with mixed mesodermal sarcoma of the ovary labeled Purkinje cells, basket cells, and scattered astrocytes. Staining of extraneural tissues was not observed. This study confirms the presence of antineural antibodies in patients with systemic neoplasia with and without paraneoplastic cerebellar degeneration and suggests that the antigens recognized by this antibody response may vary with the associated neoplasm.     

Immunohistological study on brains of Alzheimer's disease using antibodies to fetal antigens,C-series gangliosides and microtubule-associated protein 5

Summary An immunohistological study of Alzheimer's brains was performed using antibodies to C-series gangliosides and microtubule-associated protein 5 (MAP5), and their staining patterns were compared with those of antibodies to tau and -amyloid precursor protein. Antibodies to C-series gangliosides and MAP5, both of which are known to preferentially expressed in the fetal brains, immunostained dystrophic neurites of senile plaques, neurofibrillary tangles and neuropil threads abundant in 3rd and 5th layers in the cerebral cortex, all of which are considered to be pathological hallmarks of Alzheimer's disease. The immunostaining patterns of these structures by antibodies to C-series gangliosides and MAP5 were similar to those by the antibody to tau. These three antibodies also immunostained some neurons in Alzheimer's brain, although their staining patterns were slightly different from one another; i.e., both diffuse and granular patterns were seen by the antibody to tau, but only granular pattern by the antibodies to C-series gangliosides and MAP5. These neurons immunostained by these three types of antibodies appeared to be the precursors of the classical neurofibrillary tangles, as positively stained neurons were not seen in the brains of non-demented cases. The presence of fetal antigens such as the C-series gangliosides and MAP5 in Alzheimer's brain may suggest that regeneration or sprouting of neurons is ongoing in association with the re-induction of gene expression characteristic for the brain in the early stage of development.     

Immunohistochemical localization of intracellular plasma proteins in the human central nervous system

Summary The regional distribution of plasma protein immunoreactivity was studied in the postmortem central nervous system (CNS) of normal subjects 18 to 78 years old. Samples taken from various areas of brain and spinal cord were processed for peroxidase-antiperoxidase immunocytochemistry using polyclonal antibodies against plasma albumin, prealbumin, ₁-acid glycoprotein, ₁-macroglobulin, IgG, transferrin, haptoglobin, hemopexin, fibrinogen, as well against the glial fibrillary acidic and S-100 proteins. Many neurons of the spinal cord, cranial nerve nuclei, pontine nuclei, cerebellar dentate nucleus, red nucleus, thalamus and hypothalamus showed strong immunostaining for albumin and moderate to strong staining for ₁-acid, IgG, transferrin, haptoglobin, as well as relatively weak immunoreactivity against other plasma proteins. Less intense staining was seen in the nucleus basalis, putamen and Purkinje cells. In contrast, most cerebral cortical neurons were negative except for a few positively stained pyramidal neurons in the hippocampus and in layers III and V of the association neocortex, although more positive pyramidal neurons were observed in the motor and sensory neocortices. Reaction products were also seen in axons of motor and sensory long tracts. These findings suggest that plasma proteins may be transported to spinal cord and brain stem neurons by peripherally projecting nerves and that a series of anterograde and retrograde transneuronal transfers are responsible for the accumulation of plasma proteins in relay nuclei and in other CNS neurons.     

ANNA-3 anti-neuronal nuclear antibody: marker of lung cancer-related autoimmunity

Two anti-neuronal nuclear antibodies (ANNA-1 and ANNA-2) are markers of paraneoplastic neurological autoimmunity related to small-cell carcinoma. ANNA-2 is also related to breast carcinoma. Here we define a third IgG specificity (ANNA-3), identified in 11 patients (10 adults) by immunofluorescence screening of sera from approximately 68,000 patients with suspected paraneoplastic neurological syndromes. ANNA-3 binds prominently to nuclei of cerebellar Purkinje neurons, not to cytoplasm, granular neurons, or enteric neurons, but distinctively to renal glomerular podocytes. Western blots revealed an approximately 170 kDa antigen, in cerebellum and small-cell carcinoma. IgG eluted from this protein reproduced Purkinje and podocyte nuclear staining. ANNA-2 in 8 of 32 cases bound to podocyte nuclei but not to the 170 kDa protein. Healthy subjects and control neurological and cancer patients lack ANNA-3. Neurological accompaniments, subacute and usually multifocal, included sensory/sensorimotor neuropathies, cerebellar ataxia, myelopathy, brain stem and limbic encephalopathy. All of 9 adults followed had an intrathoracic neoplasm, seven biopsied within 7 months (five small-cell lung carcinomas and two adenocarcinomas, one lung, one esophagus) and two imaged, one early, the other 3 years later. Thus, immunohistochemical and Western blot criteria can now identify six IgG markers of neurological autoimmunity related to small-cell carcinoma, their frequency being ANNA-1 > collapsin response-mediator protein-5 > amphiphysin > Purkinje cell cytoplasmic antibody-2 = ANNA-2 = ANNA-3.     

Intraneuronal IgG in the central nervous system: uptake by retrograde axonal transport

The uptake of immunoglobins by CNS neurons was studied in rats. Rats were injected IP with solutions containing large amounts of rabbit IgG. Immunocytochemical staining of sections of the neuraxis revealed uptake of rabbit IgG by motor neurons of the CNS with axons projecting outside of the blood-brain barrier, including ventral horn motor neurons and cranial nerve motor nuclei neurons as well as in neurons projecting to the hypothalamus and area postrema. Staining was also noted in certain large neurons of the reticular formation and in Purkinje cells, as well as diffusely in the hypothalamus, area postrema, the pia mater, and associated vasculature and larger penetrating vessels. Uptake of rabbit IgG by lumbar spinal cord motor neurons projecting to the sciatic nerve was prevented by ligation of the sciatic nerve. These experiments support the hypothesis that certain central neurons take up immunoglobins from the periphery by retrograde axonal transport. The function of this process is not known, but it may have significance for the pathogenesis of motor and autonomic neuropathies and neuronopathies.     

Distribution of adrenomedullin-like immunoreactivity in the rat central nervous system by light and electron microscopy

Adrenomedullin is a peptide of marked vasodilator activity first isolated from human pheochromocytoma and subsequently demonstrated in other mammalian tissues. Using a polyclonal antiserum against human adrenomedullin-(22-52) amide and the avidin-biotin peroxidase complex technique, we have demonstrated by light and electron microscopy that adrenomedullin-like immunoreactivity is widely distributed in the rat central nervous system. Western blotting of extracts of different brain regions demonstrated the fully processed peptide as the major form in the cerebellum, whereas a 14-kDa molecular species and a small amount of the 18-kDa propeptide were present in other brain regions. Immunoreactive neurons and processes were found in multipolar neurons and pyramidal cells of layers IV-VI of the cerebral cortex and their apical processes, as well as in a large number of telencephalic, diencephalic, mesencephalic, pontine and medullary nuclei. Cerebellar Purkinje cells and mossy terminal nerve fibers as well as neurons of the cerebellar nuclei were immunostained, as were neurons in area 9 of the anterior horn of the spinal cord. Immunoreactivity was also found in some vascular endothelial cells and surrounding processes that probably originated from perivascular glial cells. Electron microscopy confirmed the light microscopy findings and showed the reaction product in relation to neurofilaments and the external membrane of small mitochondria. Immunoreactive terminal boutons were occasionally seen. The distribution of adrenomedullin-like immunoreactivity in the central nervous system suggests that it has a significant role in neuronal function as well as in the regulation of regional blood flow.