478 VOGT : ASSAYS OF ADRENOCORTICAL HORMONES [Vol. 76 Assays of Adrenocortical Hormones on Small Laboratory Animals BY MARTHE VOGT (Presented at the meeting of the Biological Methods Group on Tuesday, October 24h, 1950) All biological assays of cortical hormojnes are camed out on adrenalec- Mice and young rats are the animals of choice because Practical tomised animals. of the ease and speed with which their adrenals can be removed. details are given for the performance of three such assays. ALL attempts at assaying cortical hormones either on normal animals or on isolated organs have so far failed; only methods using adrenalectomised animals have met with some measure of success. In most tests the potency of the assayed material is measured by partial restoration of a functional deficiency resulting from the removal of the adrenals.One test, however, which uses as a measure of activity the fall in circulating eosinophils produced by cortical hormone, is carried out on adrenalectomised animals for the sole purpose of excluding effects due to stimulation of the animal’s own adrenals. Since, then, adrenalectomy has to precede any further assay procedure, it is obvious that those species that are easiest to adrenal- ectomise are also most suitable for these hormone assays. Since the rat is the animal on which mass-adrenalectomy is most conveniently performed, all the tests I shall discuss were originally tried on rats. Soon, however, it was found that, in all but one of these tests, mice show a greater sensitivity to the hormone, not only because of their smaller body weight, but also because of an inherent greater susceptibility.In spite, therefore, of their being somewhat less pleasant to handle and to anaesthetise than rats, all except one of the tests are nowadays done on mice.August, 19611 ON SMALL LABORATORY ANIMALS 479 There are three biological assays of cortical hormone on rodents of which enough is (1) The survival test in low environmental temperature, originated in 1938 by Selye and Schenker-l (“Cold test.”) (2) The test using deposition of liver glycogen in fasting mice given glucose parenterally, described by Venning, Kazmin and Bell2 (3) The test using the fall of circulating eosinophils in the mouse, devised by Speirs and Meyer.3 A method very similar to that of Venning, Kazmin and Bell2 was published simul- taneously by Eggleston, Johnston and D~briner.~ Fewer people seem to have used this second method, so that I shall omit its description.There is no reason, however, to suppose that it is less good. Before describing the different methods, it will be useful to compare their merits or disadvantages- (a) Of the three tests, the “cold test” is easiest to perform, since it requires observation of survival times for a period of an average of 12 to 14 hours, but no further chemical or biological measurements on the adrenalectomised animals. (b) The amount of material required is a little less for the “cold test” than for the liver glycogen test and much less for the eosinophil test. (c) The “cold test” is the only test applicable both t o desoxycorticosterone-like compounds and to compounds oxygenated at Cll.The former compounds, however, have to be administered in a different way owing to their slow action and do not give a graded response at different dose levels. The other two methods, as far as we know, assay exclusively compounds carrying an oxygen atom in the 11-position. (d) The precision is highest for the glycogen deposition test, lowest for the eosinophil test and appears to be intermediate for the “cold test.” The desirable number of animals on each dose or compound is not less than 10 for all tests. (e) Standard and unknown have to be assayed simultaneously in every experiment that uses the “cold test,” whereas a standard curve can be constructed for the glycogen deposition test.This means that experiments can be done with unknown solutions alone, and the results read off on a previously prepared standard curve. It is hoped, though it is not certain, that the same will be true for the eosinophil test. (f) The “cold test” gives significant differences between different doses only if the group of rats is made up of litter mates of very nearly the same weight. No such restriction applies to the other two tests. It may be asked whether there is no method of testing for the influence of cortical substances on electrolyte metabolism. Many such tests have been tried, but none has been found satisfactory. Recently, however, Spencer5 has developed a method that, though laborious, appears to be very promising. It permits the assay of small quantities of deoxy- corticosterone acetate (DOCA) by measuring the rate of sodium excretion after a given load of sodium chloride injected subcutaneously.There has not been enough time since the publication of the method for experience to be gained with it in other laboratories. The method is lengthy and requires precision in all manipulations; it appears to require 0-5 to 4 pg of DOCA per mouse and 12 “selected” adrenalectomised mice per test. The 12 animals are selected from 18 in a preliminary sodium load test carried out without injecting any steroids. On the next day, test, standard and control solutions are given intraperitoneally to 4 mice each and the sodium excretions measured; similar tests are performed on the sub- sequent two days, changing over the mice used for a particular solution until all mice have in turn received test, standard and control.A brief description of each test follows, though there is little to add to the published accounts except for the third test. known to justify their use as routine or screening tests. (1) “COLD TEST”- The principle of the test is to accelerate the death of adrenalectomised rats by keeping them at a low environmental temperature. Whereas death at room temperature would occur in 4 to 20 days, the cold environment (in the absence of food) precipitates it, so that it occurs within as many hours. It can be delayed by administration of cortical hormone, and the length of this delay is a measure of the potency of the hormone preparation. The480 VOGT : ASSAYS OF ADRENOCORTICAL HORMONES wol.76 exposure to low temperature not only shortens the duration of the assay, but also increases its sensitivity and, to a lesser degree, its precision. Age and size of the rat and the litter from which the animal is taken considerably influence its survival time and make careful matching of the groups imperative for obtaining significant differences in survival times with varying doses of hormone (Vogt6). The optimal age and size of the animals depends on the particular strain, successful assays requiring vigorous, fast-growing animals. Mice can be used instead of rats, but offer no advantage. Suitable ages and sizes for a particular colony of Wistar rats are, for instance, 21 to 25 days and 37 to 50 g. The variation of weight within a litter should not exceed a few grams, the precision of the results being highest when this variation is smallest.If X samples are tested on X groups of rats in a particular experiment, it is desirable to have 10 litters of X rats each, and to allot one rat of each litter to every group. The effect of variations in weight within a litter is minimised by making the mean weight of all groups equal. It is also advisable to make the sex ratio the same in all groups. The rats are adrenalectomised on one morning, well fed overnight and used for the assay on the following day. The material to be tested is injected subcutaneously in four doses spaced at intervals of one hour and a half. After the first injection, all animals are simul- taneously placed in a cold-room or large refrigerator and only removed for brief periods for further injections or rapid inspection in order to determine the times of death. If doubt is felt about whether a rat is dead or not, it is counted as dead when its corneal reflex has disappeared.The mean survival time of each group measures the potency of each sample. There is a straight line relationship between mean survival time and log-dose of hormone, provided the samples are free from toxic substances. A control group injected with the solvent and two groups treated with different doses of a standard have to be used in every assay in which a quantitative result is desired. If the samples differ in potency by less than 250 per cent. they will rarely produce significant differences in mean survival time.A suitable total dose of extract is 0.1 ml per rat. This may produce an increase of 18 to 4 hours in mean survival time over the controls, depending among other things on the temperature in the refrigerator. The temperature should be somewhere between +2” and +7” C, according to the vitality of the rats, which will vary with their size and with the strain, (2) THE GLYCOGEN-DEPOSITION TEST- The principle of the test is the observation that fasted, adrenalectomised mice only deposit injected glucose as liver glycogen when supplied with cortical steroids carrying an oxygen atom at C1,. Mice weighing from 20 to 25 g are fed on a specified diet, adrenalectomised and kept a t a constant temperature. Food is withdrawn on the third post-operative day, and the test is begun on the following morning.Seven hypodermic injections of a mixture of glucose, alcohol and the test substance are given at inte:rvals of 45 to 60 minutes. The mouse is anaesthetised one hour after the last injection. The liver is excised and hydrolysed and its glycogen content is determined. A glycogen deposition of about 30 mg per 100 g of liver is significant. Since there are seasonal variations in the response, frequent checking with a standard preparation is desirable, but is not required on every occasion as it is in the “cold test.” The most accurate results are obtained when between 10 and 40pg of cortisone are given to each mouse. The test involves liver glycogen estimations on every animal. Not less than 6 mice are required for each sample. (3) THE MOUSE EOSINOPHIL TEST- The test is based on the fact that injection of‘ cortical steroids oxygenated at CI1 causes a fall in the circulating eosinophils in the blood of any mammal.The assay is still in the experimental stage, but its sensitivity is such that this fact alone warrants its discussion in this survey. The counting of eosinophils in the blood of the mouse requires some practice, but by using the solution recommended by Speirs and Meier,3 which dissolves all blood cells except the eosinophils, counting is greatly facilitated. In the description of the details of the test, I shall quote the procedure developed by Dr. Bibile at the Pharmacological Laboratory, Edinburgh (unpublished), but I wish to emphasise that it is still in the process of development.From 6 to 10 large mice are adrenalectomised, special care being taken to remove all fat surrounding the adrenal along with the gland, in order to avoid regeneration of glandular tissue.August, 19611 ON SMALL LABORATORY ANIMALS 481 The animals are kept at a steady temperature and supplied with 0.9 per cent. sodium chloride solution instead of drinking water. On the fifth post-operative day, the mouse is warmed until its tail vessels are well dilated, and a drop of blood is obtained from a tail vein for a first eosinophil count. At the same time an injection of 5 pg of adrenaline is given subcutaneously, and a second eosinophil count is made 3 hours later. Those mice responding with a fall in the count are suspect of glandular remnants and are discarded.Immediately after obtaining the second sample of blood, the test substance is injected and its effect on the eosinophil count is observed 4 hours later. A fall indicates adrenocortical activity in a sensitive strain. The test is applicable to doses of 2 pg of cortisone, and possibly to less. The dose-response curve, however, appears to be even flatter than that for the “cold test.” I t is possible to improve the practicability of the test by using adrenalectomised mice maintained with implants of desoxycort icost erone, If we assay a mixture like a cortical extract, it may or may not be justified to use a simple steroid, say cortisone, as a standard. In the “cold test” it is not justified because the dose - response curves are not parallel. In the glycogen deposition test it has been the practice to use Kendall’s compound E as standard, and this probably introduces little, if any, error, since in that test the potency of an extract is likely to be determined by its content of compounds E and the very similar F.The same may hold for the eosinophil test, but there is as yet no information on this point. It is the ardent wish of every worker in the field of biological assays to see the biological test disappear and the assay replaced by a chemical estimation. Until recently, this hope was faint indeed for adrenocortical hormones. New chemical micro-methods, particularly chromatography, have also opened new avenues in steroid chemistry. I hope that the task of biological assays will, in the not too distant future, be relegated to its proper field, namely, to help the chemist in the isolation and identification of compounds manufactured by the adrenal gland, and to test synthetic substitutes. The question of standards has not yet been raised. To-day, the picture is changing. REFERENCES 1. 2. 3. 4. 5. 6. Selye, H., and Schenker, V., Proc. Soc. Exp. Biol., 1938, 39, 618. Venning, E. H., Kazmin, V. E., and Bell, J. C., Endocrinology, 1946, 38, 79. Speirs, R. S., and Meyer, R. K., Ibid., 1949, 45, 403. Eggleston, N. M., Johnston, B. J., and Dobriner, K., Ibid., 1946, 38, 197. Spencer, A. G., Nature, 1950, 166, 32. Vogt, M., J . Physiol., 1943, 102, 341. THE DEPARTMENT OF PHARMACOLOGY UNIVERSITY OF EDINBURGH