Search

Senile dementia of the Alzheimer's type can be diagnosed with certainty only by examining neurofibrillary tangles and neuritic plaques under the microscope. Recently, it has been suggested that the condition is linked to specific neurotransmitter systems, with a decline of cortical acetylcholine, choline acetyltransferase, cholinergic neurones projecting to the cortex, cortical noradrenaline content, locus coeruleus neurones and cortical somatostatic content. Using immunocytochemical methods, we here report that somatostatin-immunoreactive processes are present in neuritic plaques in human Alzheimer's specimens. These data, as well as other reports of non-cholinergic changes, strongly imply that Alzheimer's disease cannot be linked exclusively to cortical cholinergic elements, as proposed previously. Rather, our data on plaque and somatostatin co-localization and distribution patterns suggest that Alzheimer's neuropathology may involve primarily the loss of selective cortical neurones that are targets of the implicated transmitter systems and that plaque formation may result from the degeneration of presynaptic and postsynaptic neurites of large projection neurones in layers III and V. Given the neurochemically heterogeneous input to these cells, it is not surprising that several neurotransmitter systems, one of which is somatostatin, are implicated in the pathology of Alzheimer's disease.

Following the discovery of the growth hormone release-inhibiting factor somatostatin from extracts of ovine hypothalamus1, an N-terminally extended somatostatin of 28 amino acids has been identified in mammalian tissue2–6. The original peptide, somatostatin-14 (SS14), corresponds to the C-terminus of somatostatin-28 (SS28). Both SS28 and SS14 have biological activity3,7–9, occur in several rat brain regions10–15, are present in cell bodies and nerve terminals11–14,16 and can be released in vitro upon depolarization in a calcium-dependent manner16–19. Further, high-affinity binding sites were described for SS14, which also bind SS28 (refs 20–23). Recently, a dodecapeptide which corresponds to the N-terminus of somatostatin-28, somatostatin-28(1–12), has been characterized in rat hypothalamus24. Radioimmunological25 and immunohistochemical13 studies have indicated the presence of SS28(1–12)-like immunoreactivity in several cortical and subcortical regions of the rat brain. This peptide was found to be unevenly distributed with the highest concentration in the hypothalamus, and preferentially localized to dendritic and axonal processes and terminals. These observations suggest that SS28(1–12) may be a neurotransmitter. In this study, we describe a calcium-dependent release of a SS28(1–12)-like peptide from hypothalamic slices in vitro. This finding supports a neurotransmitter function for this peptide.

The concept of a hypothalamic neurohumoral control for anterior pituitary secretion postulates the existence of a growth hormone-releasing factor (GRF) of neuronal origin that stimulates the pituitary gland to release growth hormone (GH). Such a compound has not yet been isolated and characterized from the brain, although there is extensive physiological and biochemical evidence for its existence (reviewed in ref. 2). However, a 44-amino-acid amidated peptide having the physiological properties of GRF as well as chemical similarities was recently isolated from a human pancreatic tumour that had caused acromegaly. Two shorter biologically active fragments of 40 and 37 residues were also isolated. The synthetic replicates of these human pancreas GRF (hpGRF) peptides specifically stimulate GH release in vitro and in vivo. Assuming similarity or identity between the putative hypothalamic GRF and the tumour-derived hpGRF, we have used immunohistochemistry to search for hpGRF-like immunoreactivity in the brain. We report here that antisera against the hpGRF1-40 peptide specifically stain neuronal cell bodies in the arcuate nucleus of the primate hypothalamus, with fibres projecting to the median eminence and ending in contact with portal vessels. This topography is characteristic of a neuronal system elaborating a releasing factor. These results provide evidence that hypothalamic GRF is very similar, if not identical, to hpGRF.

A potent and specific growth hormone-releasing factor (GRF) was recently isolated and characterized from a human islet cell tumour of the pancreas that caused acromegaly1–3. Antibodies raised against the synthetic replicate of this peptide have allowed the immunohistochemical identification of GRF-producing neurones within the primate central nervous system4. Such neurones are found mainly in the arcuate nucleus4–6 in human and monkey hypothalamus, suggesting that this nucleus is a primary source of GRF. We have further investigated this hypothesis by studying the anatomical organization of GRF neurones in rat hypothalamus, using an antibody raised against the recently characterized rat hypothalamic GRF7 in normal animals and in animals neonatally treated with monosodium glutamate (MSG), a treatment which results in the selective destruction of arcuate nucleus neurones8–10. We present here the results which show that GRF-producing neurones are located mainly in the arcuate nucleus of rats. MSG treatment results in the complete loss of GRF- immunoreactive cell bodies within this nucleus and provokes a selective disappearance of GRF-immunoreactive fibres in the median eminence. These results show that the arcuate nucleus is the origin of the GRF-containing fibres that project to the median eminence and establish the MSG-treated rat as an in vivo model for studying growth hormone secretion in the absence of neurohumoral GRF.