engleski

Relationship between hippocampal neurofibrillary degeneration and neuronal loss in aging and Alzheimer's disease

Using the optical disector we counted number of hippocampal neurons and NFTs in Nissl and modified Bielschowsky stained sections in 12 control and 12 brains with the diagnosis of definite AD as determined by CERAD criteria. The slope of age-related neuronal loss was steeper in AD patients than in controls for all the fields (in thousands of neurons per year: -215 vs -120 for CA1, -68 vs -34 for subiculum, -12 vs -5 for CA2/3, -36 vs +3 for CA4, and -17 vs +26 for granule cell layer-GCL). However, significant AD-related neuronal loss was found only in GCL (22.8% p=0.01) and CA2/3 (30.3% p=0.01). Although neuronal loss was greater in older than younger AD patients in all the fields, it was close to significance only for CA1 (linear regression of neuron number vs age r=-0.48 2p=0.115). The number of NFTs negatively correlated with age of AD subjects in all the fields, except CA2/3 (r=0.16), but was significant only in CA1 (r=-0.71 2p=0.01). Significant positive correlation between neuron number and NFTs was obtained only in CA4 (r=0.57 2p=0.05), and was much weaker in subiculum (r=0.47 2p=0.12) and CA1 (r=0.36 2p=0.25). The proportion of number of NFTs/number of neurons in AD group was in relatively narrow band of values (mean&plusmn ; ; ; ; SD: CA1 0.52&plusmn ; ; ; ; 0.21, subiculum 0.32&plusmn ; ; ; ; 0.21, CA2/3 0.18&plusmn ; ; ; ; 0.15, CA4 0.12&plusmn ; ; ; ; 0.14, GCL<<0.01), and did not correlate with duration of the disease. It was also not much influenced by the age of subjects since the slope of proportion vs age was negligible: -0.6% per year for CA1, +0.4% for subiculum, +0.6% for CA2/3, -0.1% for CA4). When effects of aging and NFTs were added in AD subjects, neuron loss still could not be explained for minimally 22, 8% or 2325000 of GC, 18, 5% or 320000 of CA2/3, 23.0% or 274000 of CA4 and 3.9% or 98000 of subicular neurons (we assumed for this purpose that all counted NFTs were intracellular and insoluble, and calculated only minimal mean differences). Excluding GCL, the difference between neuron loss and number of NFTs was highest in CA4 (4.3 neurons per tangle were lost), followed by CA2/3 (2.6) and subiculum (1.2). Only a part of neuron loss in AD was attributable to aging. Since in all fields, except CA1, there were greater losses of neurons in AD than the number of NFTs + aging would suggest, we supposed that other pathological changes leading to neuron death exist. Strong negative correlation between age at death and CA1 NFTs may mean that &#8216; heathier&#8217; CA1 neurons need a higher level of exposure to pathogen to overcome a &#8216; threshold&#8217; for functional damage during AD, with consequent higher degree of characteristic pathology. Although not significantly, the number of NFTs rised with age of AD subjects only in CA2/3 field, which may be interesting since Mizutani and Shimada reported that NFTs frequency in the CA2 region in their series was a useful indicator for the neuropathological diagnosis of AD. Irrespectably to duration of AD, the proportion of NFTs to neuron number was limited, reaching some plateau values. We therefore postulated that NFTs numbers are related to the absolute number of topographically and cytochemically susceptible neurons and not that with time eventually all the neurons will develop NF changes. AD-related neuronal loss was most profound in GCL and CA2/3, presumably due to the earliest involvement of these fields in pathological process by anterograde changes originating in entorhinal cortex. Since CA1 and subicular neurons were most proned to NF degeneration and GC the least, we concluded that the pattern of neuron loss in hippocampus do not match the pattern of NFTs, but that it precedes in time and exceeds in number, due to mechanisms other than neurofibrillary degeneration. The authors would like to thank Dr. Paul Coleman for critical comments.

Alzheimer's disease; hippocampus; neurofibrillary tangle; neuron loss

nije evidentirano