Sleep disruption and other circadian rhythm disturbances are frequently seen in dementia patients. In this study, we examined the suprachiasmatic nucleus (SCN), the putative site of the hypothalamic circadian pacemaker, to determine the nature and degree of pathologic changes caused by severe dementia. Neuropathologic examination indicated that among 30 patients with a clinical history of severe dementia, 22 had Braak and Braak stage V0VI Alzheimer disease, 3 had combined Alzheimer and Parkinson disease, 3 had Pick disease and 2 had severe hippocampal sclerosis. Comparisons were made with a control group composed of 13 age-matched patients with no clinical or pathological evidence of dementia or other CNS disorders. To determine the pathologic involvement within the SCN, human hypothalami were stained with: Nissl, Bielchowsky silver, thioflavin S and specific antibodies directed against vasopressin (VP), neurotensin (NT), neuropeptide Y (NPY), vasoactive intestinal peptide (VIP), beta-amyloid (B/A4) and glial fibrillary acidic protein (GFAP). Pathologic damage was primarily limited to neuronal loss and neurofibrillary tangle formation. Only rare diffuse plaques were noted. The pathologic changes within the SCN were less severe than in the other brain regions. Morphometric analysis was accomplished using a stereological approach to sample the average total number of positively stained neurons and astrocytes in 10 different 0.1mm2 microscopic fields in the dorsal subdivision of the SCN. Patients with Alzheimer disease exhibited a significant decrease in vasopressin (9.75 vs 16.7, p < 0.001) and neurotensin (6.82 vs 9.63, p < 0.002) neurons, as well as a corresponding increase in the GFAP-stained astrocyte/Nissl-stained neuron ratio (0.54 vs 0.10, p < 0.009). These studies provide evidence that both vasopressin and neurotensin neurons are lost in Alzheimer disease, and that the astrocyte/neuron ratio is a reliable indicator of disease-related pathology within the SCN. Taken collectively, our data support the hypothesis that damage to the SCN may be an underlying anatomical substrate for the clinically observed changes in circadian rhythmicity that have been observed in Alzheimer patients.

Author notes

Supported by AG09301, AG 10682, and Harvard Brain Tissue Resource Center MH/NS 31862.