Subfractionation of human peripheral blood lymphocytes by partitioning in two-polymer aqueous phase systems. Characteristics of a small subpopulation of cells with high partition coefficient

Mixtures of aqueous solutions of dextran and poly(ethylene glycol) form immiscible 2-phase systems that are suitable for the separation of viable, functionally-competent cells on the basis of their surface properties. With the appropriate choice of ionic composition, these systems will separate cells on the basis of charge-related surface characteristics. We have previously shown that in charge-sensitive phase systems, human B lymphocytes have a low partition coefficient (K) and T cells have an intermediate K. The cell subpopulation with the highest K values contains large null lymphocytes that have most of the antibody-dependent and natural killer activities. The present study further characterizes this latter cell subpopulation. 8-33% of peripheral blood lymphocytes in normal persons had high K values. About 50% of these cells had typical Giemsa-staining granules characteristic of killer cells and one third had Fc receptors, but lacked complement receptors. Although 15-30% of the cells with high K values formed rosettes with sheep erythrocytes, there was virtually no response to T-cell mitogens such as PHA and Con A. Mixing cells having high K values with mitogen-responsive T cells from intermediate cell fractions revealed no evidence of suppressor-cell activity to explain the poor mitogen response. The studies indicate that a heterogeneous population of cells share charge-related surface properties resulting in a high partition coefficient. These cells comprise a significant proportion of human peripheral blood lymphocytes and include virtually all of the large granular lymphocytes as well as a large population without this morphology that shares the feature of unresponsiveness to T-cell mitogens.     

Hallucinations and associated factors in Alzheimer's disease.

In a consecutive sample of 230 community-dwelling patients with probable Alzheimer's disease, a structured interview yielded evidence of current hallucinations in 29.1% and misperceptions in another 11.3%. Visual and auditory modalities were similarly represented in apparent hallucinations. Hallucinations prior to the current monitoring period were rare among patients with misperceptions or with no perceptual abnormality. The probability of hallucinations was associated with the severity of cognitive dysfunction, the degree of other behavioral disturbances, and the presence of extrapyramidal signs. A logistic regression model predicting hallucinations based on these diverse clinical features accurately classified 87.0% of the sample.