Measurements of dendritic path lengths provide evidence that ethanol-induced lengthening of terminal dendritic segments may result from dendritic regression.

The present study proposed to determine the significance of previously reported ethanol-induced dendritic lengthening in mature cerebellar Purkinje neurons (PN). An analysis of dendritic path lengths (PL) was used to make this determination. A PL is the curvilinear length from the origin of the dendritic root segment to the free tip of a dendritic terminal. It was hypothesized that treatment-induced dendritic lengthening resulted from dendritic growth, and that growth should produce an upward shift in the value of the median PL. The PL and the segment length (SL) for each dendritic terminal were measured in PN of 24 month old male F344 rats, previously fed a liquid diet containing 35% ethanol-derived calories for 48 weeks, to test that hypothesis. Because there are large numbers of terminal dendrites in PN networks a shift in the value of their median PL is a sensitive measure of change in these networks. It was found that almost 30% of the ethanol-exposed cell sample had median PL that were shorter than expected and that the median SL in these same neurons tended to be longer than expected. At the end of the ethanol treatment, half of these rats were weaned from the ethanol diet for a subsequent 8 weeks to test for reversibility of these changes. Following the period of abstinence from dietary ethanol, the significant ethanol-related differences previously found in dendritic lengths were no longer present. A consideration of two models of dendritic regression suggested that elongation of surviving terminal dendritic segments in the ethanol-exposed neurons occurred through deletion of other terminal segments at their vertices (branch points). It was shown that deletions of entire terminal dendritic segments at their vertices would produce increases in the lengths of associated terminal segments in the absence of real growth. Deletions of dendritic segments would also entail loss of PL unique to the deleted segments, thereby causing a shift in the value of the median PL of affected networks.