Insulin in the central nervous system.

In summary, before hypothesizing synthesis of insulin in nonpancreatic tissues, one must determine with some accuracy the insulin concentrations in tissues such as the brain of various species or in IM-9 lymphocytes, or of non-guinea pig insulin in guinea pig tissues. If the concentrations are no more than a few percentage points of the levels initially reported by the NIH laboratory (Havrankova et al., 1978, 1979; Rosenzweig et al., 1980a,b), then some explanation should be given for the erroneously high concentrations that they earlier reported. If the very much lower concentrations that we have reported (Eng and Yalow, 1979, 1980, 1981, 1982; Bauman et al., 1982) are the true levels, then attempts to demonstrate synthesis in extrapancreatic tissues either by amino acid incorporation or by the methodology that has been described by Giddings et al. (1982) are doomed to failure. Our observations that transfer from the periphery can result in insulin levels in the brains of small-brained but not of large-brained animals comparable to or even, on occasion, higher than plasma levels when plasma levels are falling can account for our earlier observations (Eng and Yalow, 1979, 1980) that in rat but not in dog or rabbit brain insulin concentrations may be comparable to plasma levels. Furthermore, the absence of mechanisms in nonendocrine cells for the complex processing of insulin precursors to the 6000-dalton peptide and the absence of proinsulin in the extracts of the variety of tissues reported from the NIH laboratory suggest that the insulin found in these extracts was ultimately derived from pancreatic insulin.