626 



NA TURE 



[October 27. 1923 



animals when the blood sugar is alwut 0*045 per cent., 

 thev may full to be olwcrved until a much lower level 

 has been reached. This is particularly the case when 

 a lar^e dose of insulin is given some time after food. 

 The liability of symptoms to occur at the above 

 percentage of blood sugar has served as a useful basis 

 for measurement of the dosage of insulin, one unit being 

 defined as the amount which can lower the blood sugar 

 to the convulsive level of 0045 per cent, within four 

 hours in rabbits weighing alwut 2 kilogrammes. Since 

 this " physiological " unit, as it is called, is said to be 

 stronger than is necessary for certain clinical purposes, 

 it has been decided for the present to use, as the 

 " clinical " unit, one that is one-third the above strength. 

 This question of the physiological assay of insulin is 

 receiving much attention at present. 



These observations clearly pointed the way to the 

 next problem, namely, the cause for the lowering of 

 blood sugar. At first it seemed as if this should soon 

 be solved, for, as already mentioned, it has been found 

 that insulin not only causes glycogen to be deposited 

 in the liver of diabetic animals, when sugar is given, 

 li 1! lUo causes the respiratorx (juotient to become 

 liiiscd in a manner to indicate that there is increased 

 combustion of carbohydrate. It seemed likely that 

 stimulation -of the same processes in the normal Jinimal 

 under insulin must be responsible for the disappearance 

 of glucose from the blood. But the experimental 

 evidence goes to show that the mechanism of action 

 is much more complex. It has recently been found 

 that glycogen is not formed when insulin is given 

 along with sugar to normal animals (McCormick, 

 O'Brien, and E. C. Noble) ; indeed, when given to those 

 that are well fed it is decreased in amount (Dudley and 

 Marriam), and no certain evidence can be obtained 

 from respiratory experiments that more active com- 

 bustion of glucose is a necessary accompaniment of 

 the lowering of blood sugar. The consumption of 

 oxygen greatly increases in dogs preceding the onset 

 of symptoms, and the respiratory quotient usually 

 becomes somewhat raised (Dixon, Eadie, and Pember), 

 but in mice changes of an opposite character occur 

 (Dudley, Laidlaw, Trevan, and I3oock) ; in rabbits the 

 symptoms are at first like those in dogs, though less 

 marked, and subsequently like those in mice. 



The glucose which disappears is apparently neither 

 oxidised to carbon dioxide and water nor polymerised 

 into glycogen. For the present the problem is un- 

 solved, but we must not lose sight of the possibility 

 that insulin changes both glucose and glycogen into 

 some intermediary product which we fail to identify 

 with our present methods of analysis, either because 

 it does not give the reducing reactions upon which the 

 detection and estimation of sugars depend, or because 

 it is not precipitated by alcohol after treatment with 

 strong alkali, which is the characteristic property of 

 glycogen. That such substances may exist in the 

 tissues is indicated by the fact that it is impossible 

 after injecting large amounts of sugar into animals to 

 recover much more than one-half by chemical analysis 

 of the entire body. 



Whatever may be the nature of the mechanism by 

 which the blood sugar becomes lowered, there is no 

 doubt that it acts in the tissues and not in the blood 

 itself (Eadie). Thus, the addition of insulin to blood 



NO. 2817, VOL. 112] 



incubated outside the body does not alter the rate at 



which the sugar disappears from it, 

 isolated heart is perfused with a suitabli 

 containing sugar the addition of insulin tu t 

 causes this to lose its sugar more rapidly (H< 

 Latchford). The hyp(jglyca-mic effect of ii 

 for a period which varies in different an; 

 rabbits the blood sugar often iK'gins to rist agam in 

 about an hour after the injection, but this dejKnds very 

 largely on the amount of glycogen which is stored in 

 the liver. This becomes converted into gluco.se to 

 replace that lost from the blood, so that well-fed 

 animals show much quicker recovery and can with- 

 stand much larger doses of insulin, without the develof)- 

 ment of symptoms, than those previously sta 

 This mobilisation of the carlx)hydrate reserves 

 appear to depend on the transmi.ssion to the livtr 

 through its nerve supply of mes.sages set up by the 

 hypoglycemic condition, for Bum has shown that if 

 the nerve pathway (splanchnic) be locked by the 

 drug ergotonine, insulin causes, in well-fed animals, 

 a more profound degree of hypoglycaemia than 

 otherwise. 



Another important effect of insulin is on the hyper- 

 glycemia due to other causes than removal of the 

 pancreas. So far as it has been investigated, insulin 

 is capable of preventing the development of h\ptT- 

 glycemia in all these conditions. Most attention has 

 been paid to its effect on the hyperglycemia due to 

 epinephrin, partly because of the possibility that 

 insulin may be assayed by determining the amount 

 necessary to antidote a known amount of epinephrin, 

 which itself can be accurately assayed, and partly 

 because an investigation of the physiological antagonism 

 between these two hormones may throw some light 

 on the mechanism of the action of insulin. One 

 example may be given to illustrate this. We have seen 

 that insulin causes the glycogen stored in the liver to 

 become less in amount, acting in this regard like 

 epinephrin, although probably much less quickly. 

 When insulin is given along with epinephrin, however, 

 glycogen disappears from the liver much more slowly 

 than with epinephrin alone, indicating that under 

 certain conditions the pancreatic hormone arrests 

 rather than stimulates the breakdown of glycogen. 

 When these two hormones are in excess in the body, 

 one prevents the other from causing glycogen to 

 disappear. Another curious result of a somewhat 

 similar nature has been obtained by Bum with 

 pituitrin, a hormone derived from the pituitary gland. 

 When it is given along with epinephrin it also prevents 

 hyperglycemia, and when it is given with insulin it 

 prevents hypoglycemia. These results indicate the 

 ver\' puzzling nature of the problem of the action 

 of insulin, and they show that this action may 

 possibly be linked with that of other hormones in the 

 animal. 



Chemically, insulin usually gives the biuret test,*and 

 it behaves in its general properties not unlike a proteose. 

 It may, however, be a much simpler substance, since 

 active preparations have been obtained from the 

 pancreas of the skate, in which no biuret test could be 

 obtained. Its chemical identity being unknown, it is 

 of course out of the question that it could at present be 

 prepared sjTithetically. 



