414 4. ALLOXAN 



dehydrogenase. This theory is not so popular today because it is difficult 

 to understand why other SH reagents, some of which reduce pancreatic 

 glutathione (Hultquist, 1958), do not share with alloxan a selective ef- 

 fect on the /5-cells. Also it is not clear how inactivation of glutathione could 

 produce such rapid damage. 



(D) Inactivation of coenzyme A. More recently Cooperstein and Lazarow 

 (1954, 1958) found a rather potent inhibition of a liver acetylating system, 

 part of this being due to inactivation of coenzyme A. If coenzyme A is 

 inactivated in the /5-cells, it would seriously interfere with a-keto acid 

 oxidations and the operation of the cycle, but how this could be selective 

 or responsible for rapid cytological changes (since the /?-cells are resistant 

 to short periods of anoxia) is not understood. 



(E) Inhibition of metalloenzymes. Since alloxan chelates with certain 

 metal ions, and Zn++ is in some manner involved in islet function, it was 

 natural to postulate that alloxan might act on the y5-cells by either removing 

 Zn++ or inactivating a Zn++-dependent enzyme. Furthermore, the only 

 other substances known to produce a similar selective destruction of the 

 /?-cells, accompanied by a typical three-phase blood glucose variation, 

 are 8-hydroxyquinoline (oxine) and diphenylthiocarbazone (dithizon), both 

 eifective chelators of Zn++ and other heavy metal ions (Kadota, 1950; 

 Kadota and Midorikawa, 1951). If Zn++ (0.1 millimole/kg) is injected im- 

 mediately before a diabetogenic dose (0.28 millimole/kg) of alloxan in rats, 

 the incidence of diabetes is appreciably reduced (Lazarow and Patterson, 

 1951). Fe++ is as effective as Zn++, and Co++ is some 5 times more effective. 

 It seems unlikely that the metal ions simply react directly with the alloxan 

 because there is insufficient metal ion (in the case of Co++ the alloxan is 

 injected in 15-fold excess), unless the metal ions act catalytically to inac- 

 tivate the alloxan. Kurita (1955) could detect no protection by Zn++ 

 against the diabetogenic action in rabbits, but this may be due to technical 

 differences or the fact that a higher alloxan dose is required in rabbits. 

 The injection of alloxan in dogs, rabbits, and rats leads to a marked rise 

 in serum Zn++ (more than doubled) and urinary Zn++ (15-fold increase) 

 at around 3 hr (Maske et al., 1952). This might on the surface be interpreted 

 as good evidence for the ability of alloxan to interfere with Zn++ balance, 

 but it appears that the Zn++ actually arises from hemolyzed erythrocytes, 

 since a-tocopherol, which protects against hemolysis but not against islet 

 damage, reduces or abolishes the rise in serum and urinary Zn++. Histo- 

 chemical determination of Zn++ in the islets showed no change in Zn++ 

 content 15 min after the injection of alloxan, although later there is a 

 definite loss of Zn++ from the islets, this perhaps being associated with 

 the necrosis. A complication making it difficult to study the effects of 

 alloxan on Zn++ function and distribution is the fact that alloxanate prob- 



