Glycogenolysis in the Liver of the Common Frog, Rana Temporaria

It has been known for many years that the metabolism of the frog is subject to seasonal variations. Liver glycogen and total body glycogen attain a maximum value in autumn, and a minimum in spring (Athanasiu, 1899; Pfluger, 1907; Kato, 1910; Bleibtreu, 1911; Goldfederowa, 1926). A few seasonal estimations of blood sugar were reported (Lesser, 1913; Slome, 1936), but the first comprehensive study of the problem was conducted by Smith (1950), who investigated the seasonal changes in blood-sugar level, fat-body development, liver-glycogen content and gonad condition in Rana temporaria.

In view of these seasonal changes, it was felt that temperature might be an important factor in the regulation of amphibian carbohydrate metabolism. It was therefore decided to investigate the fundamental relationship between temperature and glycogenolysis in the frog’s liver, with a view to establishing the role of changing environmental temperature in the observed seasonal changes in carbohydrate metabolism. For this purpose a tissue-slice technique had many advantages, as the number of variables can be greatly reduced. For example, by incubating slices from the same liver at different temperatures, not only can individual variation be controlled, but simultaneous readings over a wide temperature-range are obtainable. Such a technique also lends itself to the in vitro study of the action of drugs and hormones on glucose production.

Frogs were either collected personally in the Wirral Peninsula, or sent in a fresh condition from collectors in Staffordshire and the Norwich area. All frogs were sexually mature and were kept in a tank-room, which was well ventilated and at approximately atmospheric temperature. The daily maximum, minimum and actual temperatures of the tank-room were noted each time a frog was kiUed. Newly arrived frogs were allowed to settle down in the tank-room for at least a day before being used.

As it was necessary to work below room temperature, an incubation apparatus was constructed in a cold-room maintained at a constant temperature of 3⁰ C. The apparatus consisted of six constant-temperature baths thermostatically controlled at approximately 7, 10, 12·5, 15, 18 and 21° C., respectively, and the temperature of each bath was constant to within ± 0·1° C. during each experiment. Each waterbath was mechanically stirred, and a flask-carrier (with provision for two flasks per bath), oscillated above the baths.

The heaters and stirrers of the apparatus were switched on 2 hr. before an experiment was due to begin. A frog was pithed in the tank-room, with minimal struggling by the animal, immediately removed to the cold-room where the liver was excised and dropped into chilled Ringer (pH 7·6). Each lobe of the liver was blotted dry, sliced between ground-glass plates, and the slices (approximately 0·2 mm. thick) dropped into chilled Ringer. A 25 ml. flask containing exactly 5 ml. of Ringer was placed in the appropriate bath 15·30 min. before starting an experiment, and gassed with oxygen for 2 min. immediately prior to the addition of liver. Accurately weighed liver slices (30·50 mg.) were introduced into each flask in turn, and the flasks stoppered. In the case of controlled experiments the same procedure was carried out for a second set of flasks, each containing 5 ml. of the experimental medium. The time of addition of liver to each flask was noted to within 30 sec. In aU experiments a liver slice was weighed and analysed for glycogen. The temperatures of the baths were checked at the beginning and end of each experiment. Exactly 2 hr. after the addition of liver sEces, each flask was removed in turn from the apparatus, and a 2 ml. aliquot withdrawn into a graduated centrifuge tube.

The 2 ml. aliquots from the incubation flasks were deproteinized by the method of Lewy (1946). After centrifugation 5 ml. of the supernatant fluid were withdrawn from each tube for glucose estimation (Hagedorn & Jensen, 1923). Glucose production was calculated as a percentage of the wet weight of liver used in each flask. A blank, in which 2 ml. Ringer was substituted for the 2 ml. of incubation medium, was treated in the same way, and when drugs or other agents were employed in the experimental Ringer, a sample of this medium was also carried through as the experimental blank.

A weighed liver slice was dropped into approximately 1 ml. of 30 % potassium hydroxide solution in a graduated centrifuge tube, and analysed for glycogen by the method of Good, Kramer & Somogyi (1933). The results were expressed as a percentage of the wet weight of tissue.

The Ringer’s solution used in this work was of the following composition: 0·65% NaCl, 0·014% KC1, 0 012% CaClg, 0·001 % NaH_sPO₄. Immediately before use the pH was adjusted to 7·6 by the addition of the necessary amount of a strong solution of Na₂HPO₄. Glass-distilled water was used throughout.

Drugs used. Adrenaline-, adrenaline B.P.; noradrenaline: Bayer ‘Levophed’; thyroxine: sodium thyroxine (B.D.H.); insulin: insulin (B.D.H. 40); ergotamine: Sandoz ‘Femergin’; dibenzyline: dibenzyline (Smith, Kline and French) (made up as follows: dibenzyline, 5·0 g.; absolute alcohol, 48·5 ml.; propylene glycol, 48·4101.; hydrochloric acid, 0·2%).

The glucose produced by 100 mg. of liver in 2 hr. at each of the six temperatures was calculated. These six glucose values were plotted against the incubation temperatures to give a temperature-glycogenolysis curve, and it was found that the curves could be divided into two main groups:

1. Those curves in which the increase in glycogenolysis with rise in temperature was linear over the range 7-21° C.

2. Those curves where the rates of glycogenolysis lay on two straight lines intersecting at any point between 10 and 18⁰ C., but most frequently between 12 and 14⁰ C. The rate of glycogenolysis increased progressively from 7° C. to about 13⁰ C., but further rise in temperature caused a disproportionate increase in glucose production. Smith (1951, 1952a) obtained a similar type of curve for the rate of beat of the isolated perfused frog’s heart, and the temperature-glycogenolysis curves showing this sharp increase in rate above a critical temperature have been termed type B to conform with Smith’s terminology.

The temperature coefficient (Q₁₀) for the overall glucose production from 7 to 21⁰ C. was calculated for all the temperature-glycogenolysis curves. In addition, the Q₁₀ of the line joining the 7 and 10° C. values of the type B curves, extrapolated to 21⁰ C., was calculated. A frequency histogram was constructed from the Q₁₀ values of all linear curves. This showed two peaks at 1·65 and 1·95, with a minimum at 1-75, and the linear curves were divided into two groups on this basis. Those with a Q_lo greater than 1-75 were termed high linear, and those with a Q₁₀ less than i·75 were termed low linear.

The average high linear, low linear and type B curves were obtained by reading off the glucose production at 7, 10, 12·5, 15, 18 and 21⁰ C. from each particular curve in that group, and calculating the mean. The data for each group, together with the number of livers, and the temperature coefficients, are given in Table 1, on which Fig. 1 is based. A few curves were obtained which could not be assigned to any of the three types described above. Some showed a sudden small rise in glucose production at a particular temperature, in an otherwise linear or type B curve, and a few curves appeared which had an extremely low glucose production at all temperatures.

From Table 1 it will be seen that there is no significant difference in the average glucose production at 7⁰ C. for all three types of curve, and that the 21⁰ C. glucose values of the type B and high linear curves are very similar. The mean high linear, low linear and type B curves had Q₁₀ values of 1·99, 1·48 and 194 respectively, and the Q₁₀ of the lower portion of the type B curve extrapolated to 21° C. (1·45) was almost identical with that for the mean low linear curve (1·48), so that when these two latter lines are drawn with the same co-ordinates, they are almost coincident. There is no significant difference between the Q₁₀ values of these two lines, or between the Q₁₀ values of the high linear and type B curves. Howliver, the difference between the Q₁₀ values of the low linear and high linear, and the low linear and type B curves is highly significant (P < 0 · ooi in each case).

During 18 months of experimental work, there were 68 type B, 13 low linear and 8 high linear curves. Type B curves were obtained in livery month of the year, the low linears appeared mainly during the period February-April, and sporadically at other times, while the high linears occurred irregularly throughout the year. The few curves which showed a very low glucose production at all temperatures occurred, with the exception of two cases, in March and April, during the spawning and post-spawning period. Liver-glycogen content was low at these times (1-3 %), and it was observed that when the liver-glycogen content was less than 1 %, glucose production was reduced at all temperatures. With higher fiverglycogen values of 2 % upwards, glucose production was not restricted.

The pattern of the type B curve, which indicated a disproportionate increase in glucose production above a critical temperature, suggested a potentiation of some glycogenolytic agent. In view of this, it was decided to investigate the action of various drugs and endocrine preparations on the liver, both in vivo and in vitro. Except where otherwise stated, the control and experimental slices were obtained from the same liver, to eliminate individual variation between animals.

One set of liver slices was incubated for 2 hr. with freshly prepared adrenaline-Ringer (1 in 10⁷), and the control slices with unmodified Ringer. A total of six experiments were carried out during January and June. The glucose values at 7, 10, i2 · 5, 15, 18 and 21⁰ C. were read off from each curve and the mean curve calculated. These data, together with Q₁₀ values and standard errors, are given in Fig. 2. The 7⁰ C. value of the control curve in Fig. 2 is considerably higher than that for the mean curves shown in Fig. 1, and is very close to the value at 7⁰ C. obtained after adrenaline. It was thought that the relatively high level of glucose production shown by the control slices at this temperature might possibly be due to the presence of circulating adrenergic material in the animal at the time of sacrifice. If any such active material was carried over in the liver slices, then perhaps its destruction would be so retarded at the lowest temperatures that it would continue to stimulate glycogenolysis and lead to an enhanced glucose value at this end of the temperature-range.

If the assumed 7⁰ C. control value is used (shown by dotted line in Fig. 2), then the Q₁₀ values of the curves with and without adrenaline are almost identical, as adrenaline increased the rate of glucose production of liver slices at all temperatures. Fig. 2 also shows that the critical temperature at which rapid increase of glycogenolysis occurs is lower in the presence of adrenaline (12 · 5° C. compared with 15° C.). Other experiments were carried out with adrenaline in January, February, July and October in connexion with other work, without corresponding unmodified Ringer controls. As a matter of interest, the mean curve was calculated for a total of fifteen experiments made in the presence of adrenaline (including those described above), and the curve thus obtained is shown in Fig. 2. This latter curve was nearer the high linear form than that based on data from only six experiments, and again, this might be regarded as evidence that the extra glucose production commenced at a lower temperature (in this case in the neighbourhood of 7⁰ C.) in the presence of adrenaline. This mean curve had a Q₁₀ value of 2·08, which is in good agreement with the value of the mean high linear curve in Fig. 1 (1·99).

Eight experiments were carried out during February, March and September. Noradrenaline bitartrate (1 in 10⁶) raised the glucose production of the liver slices in all except one case. The results obtained were averaged, and the data, shown graphically in Fig. 3, show that noradrenaline increased the glucose production at all temperatures from 7 to 21⁰ C. The 12·5,18 and 21⁰ C. values of the curve obtained in the presence of noradrenaline are joined directly, because the higher mean value at 15⁰ C. was due to a single experiment (such irregular curves have been mentioned in the previous section). The <2₁₀ of the experimental curve (1’89) is fairly close to that of the control (1·96), indicating that noradrenaline had proportionally the same effect at all temperatures. In Fig. 3, the break in the mean control curve occurs at 16·5° C., while the break in the mean curve obtained in the presence of noradrenaline occurs at 10° C. This indicates that, as in the case of adrenaline, noradrenaline leads to the earlier appearance of the extra glucose production associated with a type B curve.

The work of Smith (1950, 1954) suggested that thyroxine might be an important factor in the carbohydrate metabolism of the frog. In the present work, liver slices were incubated with Ringer plus thyroxine (1 in 1o⁶) and unmodified Ringer. Thyroxine only affected glucose production in one experiment, where it converted a type B curve to the high linear form. This result could not be repeated, and it was concluded that thyroxine had no action on the glucose production of the frog’s liver in vitro.

The effect of thyroxine in vivo was then investigated. Injections were made into the dorsal lymph sac of 1 ml. of Ringer plus thyroxine (1 in 10⁶), 48 and 24 hr. before sacrifice, and control animals were similarly injected with unmodified Ringer. It was found that thyroxine administered in vivo had no effect on glucose production over the temperature range 7-21° C.

The thyroids of all animals used in this work were examined histologically. The state of activity of the thyroid in individual frogs was not correlated either with the type of temperature-glycogenolysis curve obtained, or with the rate of glucose production. It was concluded that thyroxine does not directly influence the glucose production of frog-liver slices. This fact, howliver, does not mean that thyroxine may not indirectly influence glycogenolysis in the living animal, and in fact Smith (1954) has shown that thyroxine may play a significant role in governing the blood-sugar level of frogs subjected to excitement.

Frogs were injected with four units of insulin in the dorsal lymph sac 19 hr. prior to an experiment, and control frogs were injected with the same volume of Ringer. Liver slices were taken from one animal in each group, and incubated for 2 hr. in unmodified Ringer. Three experiments were carried out, and a typical example is given in Fig. 4. Insulin administered in vivo lowered the glucose, production of liver slices at all temperatures to well below the range of the seasonal curves in Fig. 1. An experiment carried out 2 days after the injection of four units of insulin showed that glucose production was still very low.

The effect of insulin in vitro was also investigated. Liver slices obtained from a normal animal were incubated with Ringer plus insulin (10 units/litre), and with unmodified Ringer. As no decrease in glucose production was obtained, the concentration was increased to 10 units/100 ml. in subsequent experiments, but the insulin was still without effect on glucose production.

After a frog is injected with insulin, there is an interval of 4-5 hr. between time of injection and fall in blood-sugar level (Smith, 1953). In view of this, liver slices were left in contact with Ringer plus insulin (10 units/100 ml.) for hr. at 3⁰ C. immediately prior to incubation, and the control slices were treated in the same way with unmodified Ringer. Even after this preliminary treatment, the liver slices still failed to respond to insulin in vitro. The slices were left in contact with their respective media at a low temperature, because it was felt that at a higher temperature they might lose a considerable proportion of their glycogen content before incubation commenced. For the same reason, this preliminary period was not extended beyond hr.

The present work suggested the presence of a sympathomimetic agent in the liver of the frog. In view of this, the effect of adrenergic blocking agents was investigated on liver slices incubated with otherwise unmodified Ringer. In each experiment, the liver slices were obtained from the same animal.

Slices were incubated with Ringer plus ergotamine tartrate (1 in 5 × 10⁴) and unmodified Ringer. No decrease in glucose production was observed, and in most cases ergotamine slightly raised the glucose production at all temperatures. These experiments were repeated with added adrenaline (1 in io⁷) in both control and experimental Ringer, but no blockade was observed.

Concentrations of either 50 or 100 mg. dibenzyline/litre were used, and the liver slices were in contact with either the dibenzyline Ringer or the control Ringer for 30 min. at 3⁰ C. immediately prior to incubation (Harvey, Wang & Nickerson, 1952). At the end of the 30 min. period, the slices were washed in a small amount of their respective media. The experimental slices were then incubated with freshly prepared dibenzyline Ringer, and the control slices with control Ringer, for 1 hr. only, on account of the instability of dibenzyline in aqueous solution. In each experiment, the control and experimental slices were taken from the same liver.

Eleven experiments were carried out, seven with 50 mg. dibenzyline/litre, and four with too mg. dibenzyline/litre. In two cases there was no suppression of glycogenolysis, but in the remaining nine experiments glucose production was significantly lowered in the presence of dibenzyline. An experiment showing successful blockade is given in Fig. 5. The control curve was a type B, with a marked increase in glucose production about 12·5° C. The slices in the experimental medium produced less glucose at all temperatures compared with the controls. There is a good deal of scatter among the glucose values for the dibenzyline-Ringer, but they do approximate to the low linear type of curve, as can be seen from Fig. 5 where a straight line (Q₁₀ 1·51) has been drawn between the points.

The mean curve for all eleven experiments was obtained by reading off the glucose values from each particular curve at 7, 1o, 12·5, 15, 18 and 21° C. These data, together with Q₁₀ values and standard errors, are given graphically in Fig. 6. The effect of temperature on the rate of glucose production was decreased in the presence of dibenzyline, and at 21° C. the experimental and control glucose values are significantly different (P less than 0-01).

Kepinov (1946) found that the environmental temperature affected the rate of liver-glycogen loss in captive starved frogs. In the present work, experiments were carried out during the summer, to see if there was any relationship between the environmental temperature at which frogs were kept previous to sacrifice, and the rate of glycogenolysis in liver slices incubated over the temperature range 7-21⁰ C.

In July, frogs were kept in the cold-room at 3⁰ C. for periods of up to a fortnight. Control frogs were kept in the tank-room, at approximately atmospheric temperature, for the same length of time. All frogs were unfed. For each experiment, one frog from each group was killed, and liver slices taken from each animal were incubated with unmodified Ringer. The glucose production of livers from the cold-treated animals varied, and no correlation could be found with temperature. It was concluded that the rate of glucose production in vitro depended only on the actual temperature at which the liver slices were incubated, and not on the temperature to which the animal had previously been exposed.

The liver-glycogen content was estimated in all frogs used in this work, except in a few cases where the very small livers yielded insufficient slices for both incubation and glycogen estimation. The average liver-glycogen content of the monthly frog samples, and the pattern of seasonal change, was essentially similar to that described by Smith (1950).

The present work suggests that the low linear curve and that part of the type B curve below the temperature of the intersection both represent the uncomplicated action of temperature on glycogenolysis. Smith (1951) postulated that the low linear (type E) heart curve was the simplest relationship between pulse-rate and temperature, and it is believed that this also applies to the effect of temperature on glycogenolysis.

Supporting evidence for this hypothesis was obtained from the effect of dibenzyline added to the incubation medium. In the presence of dibenzyline there was a general tendency for type B curves to be changed towards the low linear form, and in one case an almost linear relation was achieved (Fig. 5). The mean curve for eleven experiments with dibenzyline (Fig. 6) shows that the extra glucose production appeared at a higher temperature and was considerably reduced in comparison with the mean control curve. If the initial linear part of the curve after dibenzyline is extrapolated to 21⁰ C. then a line of Q₁₀ 1-49 is obtained, which is almost identical with that of the mean low linear curve in Fig. 1 (1-48). The increase in mean glucose production at 18° C. could be due to adrenergic material breaking through the blockade at higher temperatures. If it may be assumed that the lower portion of the mean curve after dibenzyline (Fig. 6) represents the plain action of temperature on glycogenolysis, then that portion of the mean type B curve (Fig. 1) below the intersection must also represent the simple direct action of temperature, since these two lines, extrapolated to 21⁰ C., have very similar Q₁₀ values (1-49 and 1-45 respectively). The disproportionate increase in glucose production shown by the type B curve can only be explained by the potentiation of some glycogenolytic agent above a critical temperature.

It is known that a sympathomimetic principle does occur in the liver of the frog. Richardet (1926) found that when the medium used for perfusing a frog’s liver was subsequently perfused through the isolated frog’s heart, an increase in frequency and amplitude resulted. He showed that the inotropic and chronotropic effects were not due to an alteration in pH, or to the presence of glucose, and concluded that the liver had delivered a sympathomimetic substance to the perfusate. The necessary enzyme system for such a synthesis does exist in certain organs (Beyer, Blaschko, Bum & Langemann, 1950; Schmiterlow, 1951; Blaschko, 1942). Bacq (1933) and Stehle & Ellsworth (1937) put forward the theory that this liver sym-pathin is noradrenaline or a similar substance. Work on the comparison of the effects of noradrenaline and liver sympathin indicated that this was probably true (Greer, Pinkston, Baxter & Brannon, 1938), and Gaddum & Goodwin (1947) concluded that there is good evidence for the theory that liver sympathin is noradrenaline, or possibly tyramine. Since the present work was carried out, both noradrenaline and adrenaline have been reported in the liver of the frog (Ostlund, 1954). It would be interesting to know whether the amounts of these amines present varied with the temperature of the animal at the time of sacrifice.

The increased glucose production obtained with added adrenaline confirms the results of previous workers (Fluch, Greiner & Loewi, 1935; Kepinov, 1937), as it caused an increased glucose production at all temperatures from 7 to 21⁰ C. Also, in the presence of adrenaline, the disproportionate increase in glucose production observed in the type B curves occurred at a lower temperature. Added noradrenaline also had both these effects on glucose production. This lowering of the temperature at which the increased glucose production occurred suggested that added adrenaline or noradrenaline was protecting endogenous sympathetic material from enzymic destruction, and that in the absence of either of these agents, sympa-thin was being destroyed at the lower temperatures and was thus unable to manifest any effect on glycogenolysis. These facts support the hypothesis that the form of the type B curve is due to the potentiation of liver sympathin above a critical temperature.

The results obtained with insulin both in vivo and in vitro agree with previous reports on the perfused isolated frog’s liver (Issekutz & Szende, 1934). The failure of insulin to act in vitro is probably attributable to the time interval which elapses between administration of insulin in vivo and lowering of blood-sugar level (Issekutz, 1924; Smith, 1953), because when insulin is administered previously to the intact animal, the glucose production of the perfused liver is reduced (Kenese & Kovacs, 1949), as was also found with liver slices in the present work.

The general conclusion to be drawn from the experimental work presented in this paper is, therefore, that the relation between temperature and rate of glycogenolysis is fundamentally linear. Above a certain critical temperature (12-14⁰ C.), howliver, it is generally found that glucose production increases more rapidly with further rise in temperature, probably owing to stimulation by endogenous sympathomimetic material. This in vitro relation between temperature and glycogenolysis may be considered in relation to some observations reported by Smith (19526), showing that the liver glycogen content of male frogs living under natural conditions in the autumn varied inversely with temperature over the range 8-20° C. The present work, if applicable in vivo, suggests that glycogenolysis in the intact frog would increase rapidly with rising temperature between 13 and 21⁰ C. It was over precisely this part of the temperature-range that Smith found the most rapid change in liver-glycogen content. This correspondence between in vitro and in vivo findings suggests that the occurrence of sympathomimetic stimulation of glycogenolysis at the higher environmental temperatures may be an important factor in controlling the glycogen content of the frog’s liver in the summer and autumn months.

I wish to thank Dr C. L. Smith for his valuable advice and criticism during the course of this work. I should also like to thank Dr K. Carter, of Smith, KEne and French Ltd., for the sample of dibenzyline given to Dr Smith.

This paper represents part of a thesis submitted to the University of Liverpool for the Degree of Ph.D. I am grateful to the University for a 2-year post-graduate research grant, and also to D.S.I.R. for a grant for 3 months.