MECHANISM OF THE DIABETOGENIC ACTION 409 



Let US summarize a few of the known characteristics of alloxan action. 

 (1) Alloxan produces in most animals at a proper dose a selective necrosis 

 of the pancreatic /?-cells so that after several days insulin secretion no 

 longer occurs. (2) Before the final diabetic state develops, alloxan brings 

 about variations in the blood glucose by mechanims which are partly, or 

 perhaps entirely, pancreatic. (3) These early changes in blood glucose are 

 not in themselves the cause of the diabetic state, since counteracting these 

 changes does not modify the eventual destruction of the /5-cells. (4) There 

 is an early and temporary loss of SH groups in various tissues, including 

 the pancreas. (5) The diabetogenic action of alloxan can be antagonized by 

 the administration of glucose, nicotinamide, and Zn++, as well as by numer- 

 ous substances reacting with alloxan. (6) Alloxan is very unstable and it 

 is believed that the effects on the /?-cells must be exerted within a few 

 minutes of an intravenous injection. (7) The action on the /5-cells is probably 

 exerted by alloxan itself rather than any of its breakdown products or 

 metabolites; at least no substance formed from alloxan is diabetogenic 

 with the exception of those readily converted back to alloxan. It may 

 also be convenient to summarize here the evidence for the important 

 conclusion that alloxan exerts its effect within several minutes. First, 

 alloxan is very unstable in the blood and is almost entirely destroyed 

 within 5 min. Second, antagonists must be given either before or within 

 2-3 min of the alloxan injection to be effective. Third, cytological changes 

 in the /5-cells can be seen within 5 min. Finally, ligation of the blood vessels 

 supplying part of the pancreas for 2 min following an alloxan injection 

 protects that portion of the pancreas (Gomori and Goldner, 1945). 



Before discussing the theories of the diabetogenic action, we must try 

 to determine if the action is indeed directly on the /5-cells. One possibility 

 is a specific spasm of the arterioles supplying the islets, as suggested by 

 Poulsen (1946). It is true that alloxan has some vasoconstrictive activity 

 (Macqueen, 1952). Blanching of the region supplied by an injected ar- 

 tery was observed and a transient rise was noted in the blood pressure, 

 which is not abolished by adrenalectomy, nephrectomy, hypophysectomy, 

 decerebration, hexamethonium, or dibenamine. It appears, however, that 

 this explanation is untenable. (1) The cc-cells are also supplied by the same 

 arterioles as the /3-cells and yet show no necrosis, although this might 

 imply a differential sensitivity to anoxia, which is no harder to imagine 

 than a differential sensitivity to alloxan. (2) There is no evidence that such 

 a very selective vasospasm can occur, and it is certainly more difficult 

 to imagine such a selectivity. (3) Epinephrine prevents the diabetogenic 

 action, presumably by a vasoconstriction so that alloxan cannot reach 

 the /3-cells, and yet no damage to the /5-cells occurs from epinephrine. 

 (4) Ligation of blood vessels supplying the pancreas for 1-6 min does not 

 lead to damage to the /?-cells, and this appears to be conclusive evidence 



