The evaluation of serial measurements of the NK cell activity in man

Studies on the NK cell activity require extensive controls because it is known that a series of variables can influence both the quality of effector cells and of target cells. This known variation of the susceptibility to lysis in one target cell line or even in sublines requires "standards", if separately performed experiments should be compared. In some laboratories, the experiments have been controlled for the variation of the target cell's susceptibility to lysis; others control for the differences of NK cell activities among subjects or for the variation of the NK cell activity in individuals from day to day. Each of these methods concentrates the main point of control to a particular variable. To demonstrate the power of several variables in relation to an induced and expected variation of the NK cell activity, we describe in this paper the results of a factorial analysis of variance testing PB-MNC from healthy donors against two target cells. The in vitro treatment of PB-MNC by IFN was chosen as induced variation of NK cell activity. As variables of the system, we defined subjects, days, target cells, and the effect of IFN. It is shown that the induced variable (IFN-treatment) counts for the most powerful variance component (32% of the total variance). Other large variance components are the differences of NK cell activities among subjects (20%), the day to day variation of the subjects' NK cell activities (13%), and the day to day variation of the susceptibility of target cells to lysis (13%). Since it is almost impossible to control for all variables in an experiment, some variables have to be selected which should be controlled for. The quantification of the variables, which is possible by a factorial analysis of variance, is useful for that purpose. This method of data analysis allows to estimate further the relevance of a particular expected effect if several variables, as in NK cell tests, influence the result of the experiment.