VOL. 12 (1953) EFFECTS OF INSULIN ON MELANOMA 335 



10.88 (range 7.28 to 13.80); and at 0.8 unit per ml 11. 15 (range 7.29 to 14.44). 

 In a similar experiment in which crystalline zinc-insulin was compared with low- 

 zinc insulin, the control (^co was 9.4; with 4 and 0.8 units of crystalline zinc- 

 insuhn per ml, 13.80 and 14.44, respectively; and with the same levels of low-zinc 

 insulin 13.47 and 13.54, respectively. The results of these, and other experiments, 

 show conclusively that 0.8 unit of insulin per ml is approximately as effective as 4 

 units per ml in stimulating the (^cq., o^ melanoma slices. However, crystalline zinc- 

 insulin (unlike amorphous low-zinc insulin) is sparingly soluble in water. While the 

 routine use of 4 to 0.8 units of insulin per ml in the present experiments assured a 

 saturating concentration of insulin per vessel, it was desirable to determine the actual 

 limiting concentrations capable of elevating Qqq^. For this purpose crystalline zinc- 

 insulin was dissolved at alkaline pH (as described under methods) to give a clear 

 solution which was immediately neutralized. Dilutions of such solution were added 

 directly to the Warburg vessels in the routine way. In two such experiments it was 

 found that as little as 0.08 unit per ml (= 3 y per ml) was as effective as ten times this 

 amount (42 % and 39 % average stimulation respectively) . Other experiments confirmed 

 this finding. Obviously the limiting concentration is below 0.08 unit per ml. Further 

 experiments have indicated that the limiting concentration of insulin varies somewhere 

 between 0.04 and 0.004 unit per ml (50 mg wet weight of tissue in a total fluid volume 

 of 2.5 ml). 



Solutions of zinc-insulin (0.15 units per ml) in 0.05 A^ NaOH were completely in- 

 activated by heating for one hour at 100° C. While the extremely small amounts of 

 insulin (3 y or less per ml) that were effective in raising the (2c6 would indicate a high 

 degree of specificity, avidin, a protein unlikely to contain insulin, was run in comparison 

 with the hormone. In two experiments, where the two proteins were compared at weights 

 equivalent to 0.8 to 4.0 units of crystalhne-zinc insulin, only the insulin significantly 

 increased the <2c6 (avidin 8 and 9% ; insuhn 37 and 47% respectively). Further work is 

 needed to determine the conditions necessary to produce optimal sensitivity to the 

 hormone. As will be described below, exposure of the melanoma-bearing mice to elevated 

 temperatures (38° to 35° C) increases the subsequent in vitro reponse to insulin. Since in 

 the present investigation it was desirable to maintain a saturating concentration of 

 insuhn; and to make additions in distilled water {e.g. zinc, phenol, magnesium, etc.), 

 1 to 4 units of insulin per ml were used in most experiments. 



Effect of temperature on in vitro response to insuli?i* 



When the Q^^ values of a number of S-91 tumors from mice maintained at ambient 

 room temperatures (24° C ± 4°) were compared, considerable inter-tumor variation in 

 response to insulin was encountered. Thus, for example, in 20 experiments, involving 

 20 different tumors in Hank-Simms medium, the average (Jco^ without added insulin 

 was 10.2, and with added insuhn (4 units per ml) was 12.6 (average stimulation-24%). 

 The ranges were 6.3 to 15.2 minus insuhn and 7.6 to 19.6 plus insulin (o to 50% stimu- 

 lation). On the other hand when the tumor-bearing mice were held at controlled tempera- 

 tures before making in vitro determination of tumor metabolism, the responses to insulin 

 were much more uniform. For example, when the tumors were from mice exposed to 



* The authors wish to acknowledge the assistance of Mrs. Elinor Brent in the conduct of 

 these experiments. 



References p. 346. 



