Journal of Clinical InvestigationVol. 46, No. 8, 1967

Studies in the Ketosis of Fasting *DANIEL W. FOSTERt

(From the Department of Internal Medicine, The University of Texas SouthwesternMedical School at Dallas, Dallas, Texas)

Summary. A series of experiments was performed during the induction ofstarvation ketosis and in the acute reversal of the ketotic state. In contrastto the predictions of two widely held theories of ketogenesis, control ofacetoacetate production by the liver appeared to be unrelated to changes infatty acid mobilization from the periphery, fatty acid oxidation, fatty acidsynthesis, or the acetyl coenzyme A concentration in the liver.

Ketosis of fasting was shown to be reversible within 5 minutes by the in-jection of glucose or insulin. This effect was due to a prompt cessation ofacetoacetate production by the liver. The possibility is raised that theketosis of fasting is due to a direct activation of acetoacetate-synthesizingenzymes secondary to a starvation-induced depression of insulin secretionby the pancreas.

Introduction

It is generally accepted that the accumulation ofketone bodies in the blood during relative or ab-solute carbohydrate deprivation is due to overpro-duction of acetoacetic and P-hydroxybutyric acidsby the liver (1). The mechanism of ketone over-production has not been well defined, though mosttheories attribute the increased rate of synthesisto enhanced oxidation of long chain fatty acids bythe liver. According to this hypothesis, carbohy-drate deprivation causes mobilization of free fattyacids from peripheral fat depots with subsequentuptake by the liver. In the liver the fatty acidsare activated to long chain coenzyme A derivativeswhich are then oxidized at an increased rate toacetyl coenzyme A, the latter compound accumu-lating in the hepatic cell (2-4). In addition, ele-vation of long chain acyl CoA levels is postulatedto contribute to the expanded acetyl CoA pool by

* Submitted for publication August 1, 1966; acceptedMay 8, 1967.

This study was supported by grant CA 08269, U. S.Public Health Service.

t Research Career Development Awardee (5-K3-AM9968), U. S. Public Health Service.

Address requests for reprints to Dr. Daniel W. Foster,Dept. of Internal Medicine, University of Texas South-western Medical School at Dallas, 5323 Harry HinesBlvd., Dallas, Texas 75235.

blocking utilization of acetyl CoA in lipogenesisand the tricarboxylic acid cycle through inhibitionof the acetyl CoA carboxylase (5) and citrate syn-thase reactions (6, 7). The direct stimulus foraccelerated ketogenesis in this formulation is con-sidered to be the increased concentration of acetylCoA in the liver cell, which then secondarily acti-vates acetoacetate synthesis (8, 9).

A second major theory of ketosis differs fromthe above in emphasizing decreased lipogenesisrather than increased fatty acid oxidation as theprimary factor in initiating ketosis (10, 11).Here, too, the immediate stimulus to increasedketone production is considered to be an increasedconcentration of acetyl CoA in the liver.

It is clear that fully developed ketosis is accom-panied by both increased fatty acid oxidation anddecreased fatty acid synthesis. The question to beconsidered is whether alterations in these lipidpathways initiate the accelerated hepatic ketogene-sis found during starvation.

In the studies to be described below an attempthas been made to determine the relationship be-tween fatty acid oxidation, fatty acid synthesis, andketone body production in starvation at varyingtime intervals after the onset of fasting. Similarstudies were performed during recovery from thefasted state. The role of nonesterified fatty acids

1283

DANIEL W. FOSTER

in the blood and of hepatic acetyl CoA concentra-tions in the control of acetoacetate synthesis by theliver has also been assessed. The results indicatethat under the conditions of these experiments he-patic ketogenesis during starvation can be alteredindependently of changes in fatty acid synthesis,fatty acid oxidation, nonesterified fatty acid levelsin the blood, and acetyl CoA concentrations in thehepatic cell. They thus provide evidence that theketosis of fasting may be causally- unrelated tochanges in lipid metabolism and suggest that cur-rent theories of ketosis may have to be modified.

Methods

Treatment of animals

Male Holtzman rats, weighing approximately 200 gand maintained on a balanced diet containing 60%o carbo-hydrate by weight,1 were used in all experiments. Theanimals were allowed to eat ad libitum until 48 hoursbefore use, at which time tube feeding was initiated to as-sure a uniform caloric intake. Immediately after thelast feeding control rats were killed, and additional groupsof animals were killed at varying intervals thereafter asindicated in the Figures. In the recovery studies a simi-lar procedure was followed except that rats were fastedfor 48 hours and killed at precisely timed intervals afterthe administration of 5 ml of a solution containing 7.5 gof the 60% carbohydrate diet by stomach tube or 0.4 mlof 50% glucose intravenously. In some experiments glu-cagon-free insulin2 was given intravenously alone orcombined with glucose.

In vivo-in vitro experiments

Venous blood was obtained for chemical determinationsat time of death, and the liver was quickly removed andplaced in iced buffer. Slices of 0.5 mmthickness wereprepared with an automatic tissue slicer, and sampleswere taken for measurement of glycogen and fatty acidcontent. Five hundred mg of slices was incubated in25-ml center well flasks containing 3.0 ml of Krebs bi-carbonate buffer, pH 7.4, and the isotopic substrate. Forstudies of fatty acid oxidation, palmitate-1-14C was pre-pared as the potassium salt and dissolved in 1% bovineserum albumin freed of fatty acid by the method of Good-man (12). One-tenth of a ml of this preparation, con-taining 1.0 /Ac of radioactivity and 0.1 Mmole of palmitate,was added to each flask. The incorporation of radioac-tivity from the albumin-bound palmitate-l-14C into C02was utilized as the indicator of fatty acid oxidation. Inthe studies of acetoacetate and fatty acid synthesis acetate-2-14C was used as the radioactive substrate. Five ,moles

General Bio'chemicals, Chagrin Falls, Ohio.2Glucagon-free insulin was obtained from the Lilly Re-

search Laboratories, Indianapolis, Ind., through thecourtesy of Dr. W. R. Kirtley.

of acetate and 1 /Ac of radioactivity were present in eachflask. All determinations were performed in duplicate,and slices from the same rat liver were used with eachisotope.

At the end of the incubation period the reaction wasstopped by the addition of 0.25 ml of 10 N H2SO4, andradioactive C02 was collected into 1 N NaOH in thecenter well after shaking 30 minutes in ice (13). Theincubation mixture was then decanted, and the slices werewashed twice with 2-ml vol of water. The washes andincubation mixture were combined and centrifuged, andthe clear supernatant was utilized for isolation of radio-active acetoacetate. The protein pellet remaining fromthe centrifugation of the latter was dissolved in waterand added back to the washed slices in the center wellflasks for determination of fatty acid radioactivity.

Acetoacetate was isolated, with minor modifications, bythe method of Van Slyke (14) as described by Weichsel-baum and Somogyi (15). The insoluble Deniges salt wascollected by centrifugation and washed twice with 20ml of water. The precipitate was dissolved in 1.0 ml of4 N HC1, and a 0.2-ml aliquot was taken for assay ofradioactivity. Control experiments without tissue showedno activity in the Deniges salt from acetate or palmitatealone.

For determination of fatty acid radioactivity 0.5 ml of90%o KOHwas added to the incubation flask, and thecontents were saponified for 1 hour at 15 pounds pres-sure. Nonsaponifiable lipid was extracted with petroleumether and discarded, and the total fatty acids were isolatedand washed after acidification according to the methodof Siperstein and Fagan (13).

Glycogen was isolated after homogenization of samplesof the tissue slices according to Stetten and Boxer (16)and hydrolyzed by the method of Good, Kramer, andSomogyi (17) for assay by glucose oxidase (18). Totalfatty acids were measured gravimetrically after saponi-fication and extraction of fatty acids as described above(13).

Nonesterified fatty acids in the liver were determinedby a modification of the method utilized by White andEngel (19). Livers were frozen in liquid nitrogen andground to a powder in a mortar and pestle cooled inliquid nitrogen. One-half g of frozen powder was thenhomogenized with 10 ml of the fatty acid extraction mix-ture of Dole (20). Titration of free fatty acids was car-ried out by the Trout, Estes, and Friedberg modificationof the latter method (21). With a ratio of extractionmixture to tissue of 20:1, recovery of palmitic acid addedto the powdered tissue was complete (for example, in atypical experiment 0.5 g of liver contained 1.02 sumoles ofnonesterified fatty acids, whereas 0.5 g of liver plus 3.96/Amoles of palmitate yielded 4.94 Amoles of nonesterifiedfatty acids, a recovery of 99%).

Blood sugar was determined by the glucose oxidasemethod (18). Total blood ketones were measured bythe method of Lyon and Bloom (22).

In separate experiments acetyl CoA concentrations inthe liver were measured during the onset of fasting and15 minutes after the reversal of ketosis by the iv adminis-

1284

KETOSIS OF FASTING

tration of glucose and insulin. Acetyl CoA was assayedspectrophotometrically utilizing citrate-condensing en-zyme 3 as described by Ochoa, Stern, and Schneider (23).Preparation of the samples for assay was carried out asdescribed by Wieland and Weiss (4) except that ratswere anesthetized with pentobarbital rather than ethersince the latter anesthetic causes a prompt and sustainedfall in blood ketone concentrations. As has been pointedout by Pearson (24) and Buckel and Eggerer (25), thecoupled assay for acetyl CoA underestimates the concen-tration of this compound because of a shift in the equi-librium of the malate dehydrogenase reaction during theassay. The actual values obtained in these experimentshave been corrected according to the formula of the latterauthors (25). Ninety-five per cent of standard acetylCoA added to the frozen liver powder was recovered inthe perchloric acid extracts in control experiments. Therecovery could be increased to about 99% by a single ad-ditional perchloric acid wash of the precipitated protein.In view of the very small increased yield this wash wasomitted in the experiments shown. Blood was obtainedat the time the liver was removed for measurement ofacetoacetate by the method of Walker (26) modified ac-cording to Kalnitsky and Tapley (27).

In the experiments concerned with the oxidation ofacetoacetic acid-3-"C various tissues and organs wereremoved from the animals as indicated, and 100 mg ofslices was incubated with 10 Mmoles of sodium acetoace-tate containing 1 uc of radioactivity. The hemidiaphragmwas not sliced but used intact. The rate of oxidation ofketone bodies was assayed by "CO2 production fromacetoacetate-3-"C after a 48-hour fast and 10 minutesafter the iv injection of 200 mg of glucose.

In vivo experiments

Fatty acid synthesis. For studies of in vivo fatty acidsynthesis 10 /Amoles of sodium acetate-2-"C containing 2,*c of radioactivity was injected intraperitoneally in 1 mlof isotonic saline. After 30 minutes the animals werekilled and the livers perfused with 20 ml of ice-coldphosphate buffer, 0.1 M, pH 7.4. The tissue was thenminced into a 5-ml vol of water to which was added 1 mlof 90%o KOH. The livers were saponified and extractedfor fatty acids as described above.

Fatty acid oxidation. Fatty acid oxidation was mea-sured in the whole animal by collection and assay of ex-pired C02 for "C content after the iv injection of pal-mitic acid-1-"C. Vena caval and arterial cannulas wereplaced as described in the next section, and the animalwas positioned in a glass metabolic cage through whichC02-free air was drawn by vacuum. The expired airwas collected in bubbling towers containing sodium hy-droxide (28). Palmitate-1-"C was prepared by dis-solving the sodium salt in isotonic saline containing 2.0%obovine serum albumin such that 1.0 ml contained 5,umoles of palmitate and 50 /c of radioactivity. This solu-tion was diluted 2: 1 with plasma from rats fasted 48

3 Crystalline citrate-condensing enzyme from pig heartwas the generous gift of Dr. Paul A. Srere.

hours (28). Six-tenths ml of the final clear solution wasinjected intravenously and "CO2 collected for six 10-min-ute periods. The initial study in each rat was done aftera 48-hour fast. When expired radioactivity had becomenegligible, 3 to 6 hours later, ketosis was reversed by theiv injection of 0.05 U of insulin and 0.1 ml of 50% glu-cose. The latter was included to avoid the possibility ofhypoglycemia with insulin alone. Ten minutes laterpalmitate-1-"C was again injected, and "CO2 was col-lected. Arterial blood ketone levels were measured atintervals throughout the experiment.

In a parallel set of experiments rats were injected withpalmitic acid-1-"C exactly as above except that "CO2 wasnot collected. Five minutes after the injection of isotope,blood was rapidly drawn from the vena cava for deter-mination of the specific radioactivity of the plasma freefatty acids. Fatty acids were extracted and titrated asdescribed previously (20, 21), and a portion of the ex-tracted sample was counted for radioactivity.

Plasma nonesterified fatty acids in fasting and recovery.Rats were tube fed as described above and fasted for 48hours. At the end of the 48-hour fast glucagon-free in-sulin was given intravenously in the doses indicated in theFigures. All animals not treated with insulin were givena control injection of saline to eliminate any differencesdue to the injection technique alone. The animals wereanesthetized with pentobarbital given intraperitoneally 5minutes before blood was drawn. Nonesterified fattyacids and acetoacetate were measured in the same sample.

The isotopic steady state. Rats were anesthetized withpentobarbital, and no. 10 polyethylene catheters wereplaced in the femoral artery on one side and the inferiorvena cava through the femoral vein on the other. In ad-dition a similar catheter was placed in the tail vein.The animals were then placed in individual restrainingcages and allowed to awaken from the anesthesia. One totwo mg of heparin was given intravenously to each ani-mal before the experiment.

After collection of samples of arterial blood at 0 time,a constant infusion of sodium acetoacetate-3-"C wasstarted through the vena caval catheter utilizing a micro-perfusion pump. The specific activity of the acetoacetatewas 2.8 ,tc per ,umole, administered so that approximately25 ,umoles was given per hour. An isotopic steady statewas rapidly obtained, usually by 10 minutes after thestart of the infusion. At this point glucose or insulinwas given intravenously, and changes in the specific ac-tivity of the blood acetoacetate were followed.

One-tenth-ml samples of the arterial blood were col-lected directly into calibrated micropipettes and trans-ferred immediately into 0.4 ml of 6.25%o trichloroaceticacid. After centrifugation the supernatant solution wasanalyzed for acetoacetate content by the method ofWalker (26). One-tenth ml of the supernatant was neu-tralized by the addition of 0.1 ml of 1.4 Msodium acetate,and 1 ml of the diazo reagent was added. The reactionwas allowed to proceed for 30 minutes at room tempera-ture and was then stopped by the addition of 0.25 ml of4 N HCl. After standing for 8 minutes to destroy anyoxalacetate present (27), the solution was extracted with

1285

DANIEL W. FOSTER

TABLE I

Hepatic glycogen and fatty acid concentrations and blood glucose levels with the onset offasting and during recovery*

Hours after onset of fasting Hours after refeeding

Determinations 0 12 24 48 6 12

Blood glucose 151 i 18 107 i 11 90 ± 6 87 ± 8 160 ± 9 188 ±4Liver glycogen 42.1 i 7.5 6.8 i 1.9 3.7 i 0.6 5.9 ±t 0.4 13.7 + 1.8 19.5 ± 0.9Liver fatty acids 34.2 ± 2.2 31.6 i 3.0 36.0 i 1.6 40.6 i 2.2 26.6 i 3.4 25.2 i 1.6

(total)

* Experimental details are given in the text. Results for glycogen and fatty acid are given as milligrams per 1,000mgwet weight of liver and are tabulated as means and standard errors of the means for nine animals at each time interval.Blood glucose is given as milligrams per 100 ml whole blood.

01 12

FIG. 1.SUIE2FFArCD YTESS.ATYAI

A T

1*1

til tof th xeietlpoedr r ulndi the

Ittt

fsgo t rw

~40- 4

20- 2

12 24 36 -48 612 24

FIG. 1. STUDIES OF FATTY ACID SYNTHESIS, FATTY ACID

OXIDATION, AND ACETOACETATESYNTHESIS IN LIVER SLICES

WITH THE ONSET OF FASTING AND DURING RECOVERY. De-

tails of the experimental procedure are outlined in the

text. Groups of rats were killed at the indicated times

after the last tube feeding. At the end of 48 hours of

fasting the remaining rats were refed the high carbohy-

drate diet by stomach tube. All experiments were per-formed in duplicate, and each point represents the meanand standard error of the mean of the results in nine rats.The results in the liver are tabulated as millimicromolesof substrate converted to product by 500 mg of slices ina 1-hour period. Points for the fatty acid oxidation ex-periments have been multiplied by 10 for easier visualiza-tion. Total blood ketones are tabulated as acetoneequivalents.

1 or 2 ml of ethyl acetate, and absorbancy was measuredat 450 m/A in a Beckman DU spectrophotometer. Theremaining portion of the sample was used for isolationand assay of the radioactive acetoacetate as describedabove. In some experiments acetoacetate and f3-hydroxy-butyrate concentrations were measured enzymatically bythe method of Williamson, Mellanby, and Krebs (29).Acetoacetate levels were comparable when measured bythe two methods, and specific activities of acetoacetate and,8-hydroxybutyrate were identical in the steady state.

Assay of radioactivity

Samples to be assayed for radioactivity were preparedas follows: "4C02. The NaOH contained in the centerwell of the flasks was diluted 1:10 with water, and 1.0ml was added to 15 ml of the solution described by Bray(30). Fatty acid-"C. The washed pentane extract con-taining the long chain fatty acids was evaporated to dry-ness under nitrogen and dissolved in 10 ml of toluenecontaining 400 mg of 2,5-diphenyloxazole and 5 mg of1,4-di [2- (5-phenyloxazolyl) ] benzene per 100 ml. Aceto-acetate-"4C. Two-tenthsml of the acid solution containingthe solubilized Deniges salt was added to 3.8 ml of ethanoltogether with 16 ml of the toluene phosphor solution.

All samples were counted in a liquid scintillation spec-trometer with and without internal standards added toeach vial to correct for differential quenching in theseparate solvent systems. The results were adjusted tothe fatty acid solvent system where counting efficiencywas approximately 50%.

Materials

All materials were obtained from commercial sourcesand were of the highest available grade.4 Purity of theethyl acetoacetate-3-14C used for preparation of the sodiumsalt was determined by gas chromatography. A singlesymmetrical peak with the precise retention time of au-thentic ethyl acetoacetate was obtained. This peak con-tained all of the injected radioactivity. The ethyl ester

4Radioactive substrates were obtained from the NewEngland Nuclear Corp., Boston, Mass. The D(-) -f3-hydroxybutyric acid dehydrogenase was obtained fromBoehringer and Sons through Calbiochem, Los Angeles,Calif.

1 286

KETOSIS OF FASTING

TABLE II

CO2 and acetoacetate production from acetate-2-14C and palmitate-1-'4C in fasting*

Acetate-2-14C Palmitate-1-14C

Time fasted C02 Acetoacetate C02 Acetoacetate

hours mpmolest mjmoles X 10t0 290 -+-- 32 15i:1.0 24 1.8 0.96 i 0.07

12 326 ± 30 81 ± 6.8 23 1.0 9.2 ± 3.324 285 ± 20 130 ± 20.0 20 ± 2.1 10.5 ± 3.848 330 ± 25 108 ± 4.8 28 ±4 1.5 7.2 ± 1.1

* The conversion of acetate-2-14C and palmitate-1-14C to CO2 and acetoacetate is listed for the indicated period offasting in nine animals. Results are given as millimicromoles of substrate converted to product per 500 mg of slices perhour. Slices from the same animals were used with each isotope under the experimental conditions listed in Figure 1.

t Direct quantitative comparison of the rates of oxidation of acetate and palmitate cannot be made from these databecause of differences in the amount of substrate added and differences in tissue pool sizes. Approximately 6% of theadded acetate-2-14C was recovered as -4CO2; the figure for palmitate-1-'4C was 2 to 3%.

was hydrolyzed in 1 N NaOH, evaporated to drynessunder vacuum, and neutralized before use. Sodium ace-toacetate solutions were assayed enzymatically (29).

Results

Changes in the liver with the onset of fasting.Changes in blood sugar, liver glycogen, and totalhepatic fatty acids in fasting and recovery are re-corded in Table I. The blood sugar and glycogenconcentrations decreased in the expected fashionwith the onset of starvation. As has been previ-ously noted, total fatty acids of the liver were un-changed by fasting (31, 32). This is in contrastto the situation in experimental diabetes wherefat content of the liver does increase and appearsto be correlated with the degree of ketosis (3).

With the onset of fasting after a high carbohy-drate diet, profound changes occurred in the pat-terns of fatty acid and acetoacetate synthesis by ratliver slices. These changes are shown graphicallyin Figure 1. By 12 hours after the final tube feed-ing fatty acid synthesis had fallen almost tenfoldfrom 111 m~umoles per hour to 16 mnumoles perhour. At the same time a fivefold rise of aceto-acetate synthesis occurred, from 15 to 81 mjumolesper hour. Simultaneously blood ketones increasedfrom 1.0 to 5.0 mg per 100 ml, continuing to riseto 9.0 mg per 100 ml at 48 hours. At 24 hoursfatty acid synthesis had fallen almost to zero andremained at this level at 48 hours. In contrast tothe striking changes in fatty acid and acetoacetatesynthesis, fatty acid oxidation was unchanged dur-ing the first 24 hours, rising slightly but with nostatistical significance at 48 hours.

It should be emphasized that the indicator offatty acid oxidation in these experiments was "4CO2

production from palmitate-1-"4C. Two factorscould conceivably interfere with the adequacy ofthis procedure. First, since palmitate oxidation toCO2 requires breakdown of the long chain fattyacyl CoA to acetyl CoA and subsequent oxidationof acetyl CoA to CO2, a depression of the latterreaction might mask increased oxidation of palmi-tate to acetyl CoA. Second, the radioactive palmi-tate added to the slices might be diluted by an ex-panded hepatic fatty acid pool with the sameeffect. The data of Table II indicate that 14CO2production from acetate-2-14C was not impaired bya 48-hour fast though acetoacetate production in-creased sevenfold. An essentially identical patternwas found in slices from the same liver when pal-mitate- 1-14C was the substrate as would be ex-pected from the fact that acetyl CoA is an obliga-tory intermediate in CO2 and acetoacetate produc-tion from long chain fatty acids. Since 14CO2production from acetate was unimpaired, it is

TABLE III

Nonesterified fatty acids in the liver duringfasting and recovery*

Hepatic nonesterifiedTime fasted fatty acids

hours m.moles/g0 524 ± 13

24 1,324 ± 11548 717 i 13948 + 15 minutes treated 642 ± 87

* Nonesterified fatty acids in the liver were measuredas described in the text. Results represent the means andstandard errors of the means from duplicate determinationsin four animals at each time interval and are given asmillimicromoles of nonesterified fatty acid per gram ofliver. Treated animals were given 0.05 U of insulin and0.4 ml of 50% glucose intravenously 15 minutes beforebeing killed.

1287

DANIEL W. FOSTER

/80k

160tk

4 /40

0)120

.1 /00

't 80

o 600

E 40

20

0 2 3t Hogurs

Gzucose

FIG. 2. STUDIES OF FATTY ACID SYNTHESIS, FATTY ACIDOXIDATION, AND ACETOACETATESYNTHESIS IN LIVER SLICESDURING RECOVERYFROMFASTING. At the end of a 48-hourfast animals were refed by stomach tube (time Q). Eachpoint represents the mean and standard error of the meanin duplicate experiments from six animals.

clear that an increased oxidation of palmitate-1-14Cto acetyl CoA would be reflected by an increasedrecovery of radioactivity in CO2 as well as aceto-acetate provided that isotope dilution in an ex-panded fatty acid pool had not occurred. Whenhepatic free fatty acid levels were measured, asshown in Table III, a definite increase was foundat 24 hours that decreased to near fed levels at48 hours. Correction of the 14CO0 recoveriesfrom palmitate listed in Table II by these data in-dicates a maximal increase of palmitate oxidationat 24 hours of about twofold and at 48 hours of1.6-fold. The increase in fatty acid oxidation wasthus considerably smaller than the tenfold changesoccurring in acetoacetate and fatty acid synthesisin the same time interval.

On the basis of these experiments, it appearedthat accelerated ketogenesis by the liver and a risein blood ketones occurred without equivalentchange in fatty acid oxidation by hepatic tissue.Since fatty acid synthesis fell concomitantly with

the rise in acetoacetate production and to equiva-lent degree, the results were compatible with theviewpoint that decreased lipogenesis might initiateoverproduction of ketones by the liver. Studiesperformed at earlier intervals after the onset offasting were unsuccessful in differentiating thesetwo effects.

Changes in the liver zuith recovery from fasting.In view of the fact that fatty acid synthesis andacetoacetate production varied simultaneously andin reciprocal fashion with the onset of fasting,even at the earliest time intervals, these pathwayswere next studied in the recovery from fasting inan attempt to see whether an increased synthesisof fatty acids could be demonstrated before a fallin acetoacetate synthesis. Such a sequence wouldbe expected if decreased fatty acid synthesis did,in fact, initiate acetoacetate overproduction in theliver.

In the initial experiments, shown in Figure 1,fasting was terminated after 48 hours by adminis-tration of 5 ml of the high carbohydrate diet bystomach tube. Fatty acid synthesis, fatty acid oxi-dation, acetoacetate synthesis, and blood ketoneswere measured at 6 and 12 hours. At the 6-hourperiod blood ketones and acetoacetate synthesis hadboth returned to normal, and fatty acid synthesishad increased from barely detectable levels to 80miumoles per hour. By 12 hours lipogenesis hadincreased to above normal levels as is typical forrefeeding after starvation (33). Fatty acid oxida-tion was not significantly changed. Thus reversalof ketosis was accompanied by reversal of thechanges in fatty acid and acetoacetate synthesis inthe liver, but again the two pathways appeared tochange simultaneously.

To study this relationship further, experimentswere next performed at earlier intervals during therecovery from fasting. As before, rats were fastedfor 48 hours and then tube fed the high carbohy-drate diet. Studies were carried out at 1, 2, and3 hours after refeeding with the results shown inFigure 2. At 1 hour blood ketones fell from 9.5to 2.7 mg per 100 ml, and acetoacetate synthesisdecreased from 145 to 65 m~tmoles per hour. Sur-prisingly, however, no increase in fatty acid syn-thesis occurred, even up to 3 hours. These find-ings indicate that changes in acetoacetate synthesisin the liver were not dependent upon changes infatty acid synthesis.

A Facstly AcidSynrMhheais

* Fczaty AcidOxidaiton,

A Acetoacecataz's

o Bzood Ketone2S

1288

'O

KETOSIS OF FASTING

TABLE IV

Acetyl coenzyme A levels in the liver during fasting and recovery from fasting*

15 minutesDetermination 0 12 hours 24 hours 48 hours treated

Acetyl CoA, 34.4 i 2.8 65.9 i 4.0 70.5 ± 4.9 50.5 ± 3.2 47.1 i 3.1mumoles/g liver (6) (6) (6) (8) (8)

p

DANIEL W. FOSTER

TABLE V

Fatty acid synthesis in vivo in the reversal of ketosis*

Condition Fatty acid synthesis

m/imoles substrate convertedFed 735 4 2748-hour fast 55 i 6.41 hour retreated 43 ± 10.23 hour retreated 97 ± 32

* Experimental conditions are described in the text.The results are given as the means and standard errors ofthe means in six animals at each time interval.

therefore undertaken to determine how rapidlyreversal of ketosis might occur. The results ofsuch experiments are shown in Figure 3. Ratswere fasted 48 hours and then rapidly injected with0.025 U of glucagon-free insulin intravenously.Samples of blood were drawn in the fasted stateand at 5, 10, 20, and 30 minutes after treatment.Blood ketones began to fall by 5 minutes and at 10minutes had decreased from the fasting mean of10.0 to 3.0 mg per 100 ml. At the 20-minute pe-riod no further fall was noted, and at 30 minutesa definite increase in ketosis occurred. The re-bound in acetoacetate concentration appeared tofollow the development of hypoglycemia of about30 mg per 100 ml. Reversal of ketosis also beganwithin 5 minutes when insulin was replaced with0.3 ml of 50% glucose.

The fact that insulin alone reversed the ketosisof fasting suggests that the response to glucose inthe previous studies was mediated totally or inpart through physiologic stimulation of insulin se-cretion by the pancreas.

Fatty acid synthesis and oxidation in vivo dur-ing recovery from fasting. The data of Figure 2indicated that fatty acid synthesis and oxidationwere unchanged for up to 3 hours after the reversalof fasting ketosis when measured in the liverslice in vitro. Although close parallel existed be-tween in vitro acetoacetate synthesis and blood ke-tone levels and although the changes in fatty acidsynthesis were those expected with fasting, it waspossible that the time course of these changesmight be different in vivo. A comparison of fattyacid synthesis in the intact animal in the fed andfasting state is shown in Table V. Forty-eighthours after initiation of the fast, fatty acid synthesisfell from 735 to 55 mpumoles per half hour, a de-crease of 93% from the rate in the fed animal. No

change in the rate of synthesis was seen at 1 and 3hours after refeeding.

The rate of oxidation of palmitic acid-1-14C to14CO0 was measured in five animals before andafter reversal of ketosis. The results were similarin all. A typical experiment is shown in Figure4. In the fasted state 24%o of the injected radio-activity was recovered in CO2 in 60 minutes.Three hours later the animal was treated with in-sulin and glucose. Acetoacetate concentration inthe blood immediately before reversal of ketosiswas 11.6%o. Eight minutes after the administra-tion of insulin and glucose, acetoacetate decreasedto 4.6 mg per 100 ml in the expected fashion. At10 minutes palmitate-1-14C was injected. Thecumulative recovery curve of isotope in CO, wasalmost superimposable on the pretreatment curve,and again 24%c of the injected radioactivity wascollected in the 60-minute period. Acetoacetateconcentration at the end of the experiment was1.6 mg per 100 ml. Measured specific activity ofunesterified fatty acids in the blood 5 minutes afterthe injection of the radioactive palmitate was 3,120+ 64 cpm per umole in four fasted animals and2,520 ± 99 cpm per ,umole in four animals treatedwith glucose and insulin.

aLL

LIJ >-,

5z

Wz E

a.w

00lq-cuo'

300-0 48 hour fasted*-0- 10 minute treated

25_

20

15 ,

10 -

5 l

I' 1

-0 10 20 30 40 50MINUTES

60

FIG. 4. PALMITATE-1-14C OXIDATION TO 14CO2 IN VIVOIN THE REVERSAL OF KETOSIS. Experimental details areoutlined in the text. Fasted and treated curves wereobtained in the same animal. Acetoacetate concentrationin the blood immediately before the injection of insulinand glucose was 11.6 mg per 100 ml. Eight minutes af-ter insulin acetoacetate was 4.6 mg per 100 ml, and atthe end of the experiment the level was 1.6 mg per 100ml. Palmitate-1-'4C in the treated curve was adminis-tered 10 minutes after insulin and glucose.

1290

KETOSIS OF FASTING

These experiments support the conclusionsdrawn from in vitro data and indicate that undercircumstances where ketosis has been reversed invivo no change occurs in fatty acid synthesis oroxidation.

Plasma nonesterified fatty acids in fasting andrecovery. The relationship of ketosis to plasmanonesterified fatty acids was next studied. In theinitial experiments, acetoacetate and nonesterifiedfatty acids were measured in inferior vena cavalblood during the onset of fasting and 10 minutesafter the iv injection of 0.05 U of insulin. Theresults are shown in Figure 5. By 12 hours freefatty acids had increased from 113 uEq per L to421 /uEq per L. From 12 to 48 hours a moregradual increase occurred, the peak level being530 ptEq per L. Acetoacetate concentrations roseonly slightly during the first 12 hours but by 48hours had increased about fivefold from a begin-ning level of 1.2 mg to a final concentration of 6.6mgper 100 ml. Ten minutes after the injection ofinsulin both the ketoacid and nonesterified fattyacids had clearly decreased.

In an effort to dissociate the two phenomena, ad-ditional experiments were performed. As shownpreviously in Figure 3, it had been observed that

6.0 F

- 5.00'E

4.0

I-,,,3.0

0w 2.00

1.0 I-

0'0 12 24 36 48 10min.

HOURS- I

FIG. 5. NONESTERIFIED FATTY ACIDS ANDACIN THE BLOOD IN FASTING AND RECOVERY. Nfatty acids and acetoacetate were measuredvena caval blood. After a 48-hour fast 0.05cagon-free insulin was given intravenously.injected with insulin received a control injecline. Each point represents the mean and stalof the mean of the results obtained in six anin

0C'Ew-

_-

0I-

w0

4L

0jw

wz

o 48hrs. 5min. 10min.TIME

FIG. 6. NONESTERIFIED FATTY ACIDS AND ACETOACETATEIN THE BLOODWITH INSULIN TREATMENT(0.5 U). Non-esterified fatty acids and acetoacetate were measured inblood obtained from the abdominal aorta. After a 48-hour fast 0.5 U of insulin was given intravenously. Eachpoint represents the mean and standard error of the meanof the results obtained in eleven animals except for thefed controls, where six animals were used.

the administration of insulin caused a rapid drop700 in the concentration of acetoacetate in the blood

until hypoglycemia supervened, at which time a600 rebound in ketosis occurred. To exaggerate this

response a group of fasted animals were given 0.5500 -J U of insulin intravenously, a dose ten times greater

W than that used in the experiment of Figure 5.400 The results are shown in Figure 6. Five minutes300 ,, after the injection of insulin the mean acetoace-

z tate level in the blood decreased from 6.0 to 3.6200 mgper 100 ml (p = < 0.005). Plasma nonesteri-

fied fatty acids also fell from 216 to 103 ,uEq per L100 (p = < 0.01). The mean blood sugar at this time

was 64 mg per 100 ml. At 10 minutes acetoace-0 tate rose sharply to 6.1 mg per 100 ml (p = <

0.05) as the blood sugar fell to 41 mg per 100 ml.ETOACETATE In contrast, plasma nonesterified fatty acid con-Jonesterified tinued to decrease, reaching a low of 53 uEq per

in inferior L.U of glu- Additional evidence that the ketosis of fasting

tion osf nsat may be unrelated to an elevation of free fatty acidsndard error in the blood is provided by the data of Figure 7.nals. In this experiment animals fasted for 48 hours

o NEFA* Acetoocetote

F

IF

tInsulin

1291

DA. IEL W. FOSTER

6.01-

5.0A0cEw-

I-

0wC.

4.0 F

3.0 F

2.01-

1.0

OL0 48hrs. 5min. 10min.

TIMEFIG. 7. NONESTERIFIED FATTY ACIDS AND

IN THE BLOODWITH INSULIN TREATMENT(Iperimental details are as listed for Figure0.01 U of insulin was injected at the endfast. Each point represents the results froi

were treated with 0.01 U of insulin irUnder these circumstances no changeacetoacetate concentrations in the bloofact that nonesterified fatty acidpromptly at 5 minutes and at 10 nnear 0.

These experiments demonstrate tltate and nonesterified fatty acid conc4the blood can be varied independentlythe conclusion that ketogenesis needally related to changes in long chain fabilization and oxidation.

Recovery from ketosis in the isostate. The very rapid fall in blood Ithe administration of insulin or glucoquestion of the mechanism of this reiit possible that cessation of hepatic ovof ketones could occur within 5 ntreatment, or did the reversal of therepresent an increased peripheral oxiitone bodies? To answer this questicfasted for 48 hours and then infusedacetate-3-14C until a constant specifiblood acetoacetic acid had been obcause of the rapid turnover of acetoa

intact rat, this could be accomplished without aprimer dose of acetoacetate-3-14C. When the iso-

350 topic steady state was reached, the animal wastreated with glucose or insulin intravenously

300 through the tail vein catheter. Since the specificactivity of the blood acetoacetate under these cir-

250 cumstances is the equilibrium product of nonradio-_^ active ketones synthesized by the liver and the

200 Cr radioactive material being infused, it is clear that3 cessation of hepatic synthesis would result in a

150 E rise in specific activity of the blood, whereas an in-W creased peripheral utilization would cause no

100 change. A typical example of such an experimentis shown in Figure 8 where it can be seen that theinjection of insulin was followed by a prompt fallin arterial blood acetoacetate concentration andthat concomitant with this fall a sharp rise oc-

0 curred in its specific activity. Identical resultswere obtained when glucose was given alone.Thus recovery from fasting ketosis is accompanied

1ACETOACETATE by a marked diminution of acetoacetate production0.01 U). Ex-6

6 except that by the liver.of the 48-hour Acetoacetate oxidation by various tissues beforem six animals. and after treatment zeith glucose. Although the in

vivo experiments reported above suggested thatitravenously. decreased hepatic synthesis was the major mecha-

occurred ind despite thes decreasedminutes were

iat acetoace-entrations inand support

not be caus-atty acid mo-

topic steadyketones afterse raised theversal. Waserproductionainutes afterketotio state

dation of ke-)n, rats wereI with aceto-c activity oftained. Be-Lcetate in the

S:

(I)

0 15 30 45 60 75MINUTES

th-

0\

I-

(3

FIG. 8. THE REVERSAL OF KETOSIS IN THE ISOTOPICSTEADY STATE. The rat was fasted 48 hours and acetoace-tate-3-16C administered as described in the text. At theindicated time 0.025 U of insulin was given intravenously.Specific activity is recorded as counts per minute permilligram of acetoacetic acid. (For conversion of aceto-acetate concentrations to acetone equivalents, multiply by0.58.)

6 An analysis of the kinetics of acetoacetate metabolismwill be presented in a separate report.

[oNEFA| Acetoacetate|

1292

KETOSIS OF FASTING

TABLE VI

Oxidation of acetoacetate-3-'4C by various tissues of the fasted rat before and after glucose administration*

Tissue Untreated Treated p value

Diaphragm 4.0 -1- 0.70 4.1 i 0.79 NSSkeletal muscle 1.1 4 0.14 1.4 0.19 NSKidney 2.5 i 0.08 2.4 i 0.01 NSAdipose tissue 0.28 + 0.05 0.28 4 0.07 NSHeart 5.6 41 0.44 5.8 :1 0.85 NSBlood ketones, mg/100 ml -8.9 i 1.1 3.6 :1 1.0

DANIEL W. FOSTER

acid oxidation or fatty acid synthesis, though al-terations in these pathways, particularly the latter,did occur with fasting. Strong confirmation ofthis conclusion was obtained by studies in the in-tact animal. The oxidation of albumin-bound pal-mitate-1-14C to 14CO2 was unchanged 10 minutesafter the injection of glucose and insulin at a timewhen blood acetoacetate concentration had droppedto less than 50%o of the original fasting level. Al-though some caution must be exercised in the in-terpretation of these results because of lack ofknowledge of the specific activity of the tissuefatty acid pools, the fact that the measured spe-cific activity of the unesterified fatty acids in theblood of fasted and treated animals was essentiallythe same 7 and the observation that the liver freefatty acid pool was unchanged with treatment (Ta-ble III) suggest that the results may accuratelyreflect absolute rates of fatty acid oxidation.

Fatty acid synthesis in vivo likewise was un-changed on reversal of ketosis. Significant in-creases in lipogenesis did not occur until at least3 hours after refeeding. This is in contrast tostudies by Fain, Scow, Urgoiti, and Chernick (41)in the pancreatectomized diabetic rat in vivo wherefatty acid synthesis increased in 30 minutes afterthe injection of 12 U of insulin. The reason forthe difference in time required for restoration oflipogenesis in fasting and diabetes is not known.

The results of both in vitro and in vivo studiesthus indicated that ketosis could be interruptedwithout changes in either fatty acid oxidation orsynthesis. If, as seems likely, recovery from keto-sis represents a reversal of the processes initiatingketosis, then the overproduction of acetoacetate bythe liver cannot be considered to be the secondaryconsequence of alterations in these lipid pathways.

It should be noted that the increased fatty acidoxidation theory of ketosis requires two phases:increased mobilization of free fatty acids fromperipheral fat stores and increased oxidation ofthese fatty acids in the liver. To study the mobili-zation phase, nonesterified fatty acids and aceto-acetate concentrations in the blood were compared

7 As noted in Figures 5 to 7 free fatty acids in the blooddecreased promptly on administration of insulin. Sincethe injected palmitate exchanges rapidly with a pool offatty acid about 100 times larger than that of the blood(28, 40), the fall in concentration of nonesterified fattyacids in the blood would affect the specific activity of thetissue pool to a negligible extent.

during the onset of fasting and in recovery fromthe fasted state. These experiments indicated thatboth components increased with starvation, non-esterified fatty acids showing the greater percent-age rise at 12 hours. On retreatment with 0.05 Uof insulin, both acetoacetate and free fatty acidlevels decreased at 10 minutes in near parallelfashion. When the insulin dose was increased to0.5 U, a different pattern was observed. At 5minutes acetoacetate fell sharply as expected andwas accompanied by a fall in nonesterified fattyacids. At 10 minutes, however, a definite dissoci-ation occurred. Acetoacetate concentration in-creased in the characteristic rebound phenomenonof hypoglycemia, whereas plasma nonesterifiedfatty acids continued to decline. When the dose ofinsulin was decreased to 0.01 U, blood acetoacetatelevels did not fall despite a marked decrease infree fatty acids over the 10-minute period. Ifketosis were dependent on mobilization of fattyacids from the periphery, a depression of free fattyacid concentrations in the blood would have to beaccompanied by a decrease in acetoacetate levelsduring this period in view of the repeated observa-tion that ketones fall by 5 minutes with adequateinsulin. It thus appears, in confirmation of thefatty acid oxidation data, that no necessary rela-tionship exists between nonesterified fatty acidsin the blood and starvation ketosis.

As discussed previously, current theories that in-voke either increased fatty acid oxidation or de-creased fatty acid synthesis as initiating steps inthe development of ketosis consider the increasedacetoacetate synthesis of starvation to be the re-sult of an expanded hepatic acetyl CoA pool. Se-quential measurements of acetyl CoA indicatedthat concentrations of this compound increased at24 hours to levels about double those of the fedstate, a result compatible with a controlling role foracetyl CoA in acetoacetate synthesis. On the otherhand, at 48 hours the acetyl CoA concentration de-creased, whereas acetoacetate concentrations in theblood increased still further. The critical point,however, is that the administration of glucose andinsulin to animals fasted for 48 hours caused arapid fall of acetoacetate in the blood without de-crease in the concentration of acetyl CoA in theliver. This course makes unlikely the thesis thataccelerated ketogenesis in the liver is primarily re-lated to an expanded acetyl CoA pool.

1294

KETOSIS OF FASTING

It is of interest that the concentrations of bothfree fatty acids and acetyl CoA increased in theliver at 24 hours and decreased at 48 hours. Sincenonesterified fatty acids in the blood continued torise throughout the 48-hour fast, it seems likelythat the major source of fatty acid contributing tothe acetyl CoA pool in the liver is hepatic ratherthan peripheral triglyceride.8

In regard to the metabolism of acetyl CoA, thepresent studies showed no impairment of acetate-2-14C oxidation to 14CO2 during 48 hours of star-vation. This observation in the intact cell, to-gether with the recent demonstration that citratesynthase activity measured directly in liver ex-tracts is unchanged or slightly increased in fasting(43), lends no support to the concept (6, 7, 9)that palmityl CoA-induced inhibition of the latterenzyme plays a significant role in starvationketosis.

The rapidity with which ketosis was reversibleby the administration of insulin or glucose wassurprising. Amatruda and Engel (44) had previ-ously shown that insulin reversed the ketosis offasting in 1 hour, and Fain and associates (41)reported that blood ketones in the diabetic ratdecreased 30 minutes after the injection of insulin.In the studies reported here, however, acetoacetatelevels began to decline within 5 minutes aftertreatment with insulin or glucose. Moreover, theresults of experiments in the isotopic steady stateclearly showed this decrease to be due to an abruptcessation of ketone body production by the liver.

The fact that changes in acetoacetate synthesiscan be dissociated from changes in fatty acid mo-bilization from the periphery, fatty acid oxidation,fatty acid synthesis, and acetyl CoA concentrationin the liver by kinetic studies raises the possibilitythat the increased acetoacetate production accom-panying starvation is a primary rather than a sec-ondary event. That is, it seems possible that star-vation produces an activation of acetoacetate-syn-

8 Repeated attempts were also made to measure longchain fatty acyl CoA levels in the liver as described byTubbs and Garland (34) and Bortz and Lynen (35). Al-though the former authors state that recovery of addedpalmityl CoA was complete, recoveries of palmityl CoAin these experiments were only 10 to 40% utilizing theirconditions for isolation, hydrolysis, and assay of the longchain CoA esters. Attempts to measure the protein-bound fatty acyl CoA with pork brain palmityl CoAdeacylase of high specific activity prepared according toSrere, Seubert, and Lynen (42) were also unsuccessful.

thesizing enzymes independent of changes in lipidmetabolism leading to alteration in acetyl CoAproduction or utilization. This would not, ofcourse, imply that increased fatty acid oxidationand decreased fatty acid synthesis do not play arole in assuring substrate for acetoacetate synthesisin ketosis, only that such changes do not initiatethe ketotic state.

The mechanism by which a primary stimulationof acetoacetate synthesis might be accomplished isunknown. The observation that the injection ofinsulin alone reverses the ketosis of fasting withinminutes by causing cessation of ketone synthesisin the liver suggests the possibility that the con-trol of acetoacetate synthesis may be mediated byinsulin itself. The corollary of this possibilitywould be that the ketosis of fasting may representa primary activation of acetoacetate-synthesizingenzymes as a result of starvation-induced depres-sion of insulin secretion by the pancreas.

AcknowledgmentsThe author wishes to express his appreciation to Dr.

M. D. Siperstein for many helpful discussions in thecourse of this work. The expert technical assistance ofMiss M. Joanne Guest is gratefully acknowledged.

References1. Bressler, R. The biochemistry of ketosis. Ann. N. Y.

Acad. Sci. 1963, 104, 735.2. Fritz, I. B. Factors influencing the rates of long-

chain fatty acid oxidation and synthesis in mam-malian systems. Physiol. Rev. 1961, 41, 52.

3. Scow, R. O., and S. S. Chernick. Hormonal controlof protein and fat metabolism in the pancreatecto-mized rat. Recent Progr. Hormone Res. 1960, 16,497.

4. Wieland, O., and L. Weiss. Increase in liver acetyl-coenzyme A during ketosis. Biochem. biophys.Res. Commun. 1963, 10, 333.

5. Bortz, W. M., and F. Lynen. The inhibition of acetylCoA carboxylase by long chain acyl CoA deriva-tives. Biochem. Z. 1963, 337, 505.

6. Tubbs, P. K. Inhibition of citrate formation by long-chain acyl thioesters of coenzyme A as a pos-sible control mechanism in fatty acid biosynthesis.Biochim. biophys. Acta (Amst.) 1963, 70, 608.

7. Wieland, O., and L. Weiss. Inhibition of citrate-synthase by palmityl-coenzyme A. Biochem. bio-phy. Res. Commun. 1963, 13, 26.

8. Krebs, H. Biochemical aspects of ketosis. Proc. roy.Soc. Med. 1960, 53, 71.

9. Wieland, O., L. Weiss, and I. Eger-Neufeldt. Enzy-matic regulation of liver acetyl-CoA metabolism inrelation to ketogenesis. Advanc. Enzyme Regulat.1964, 2, 85.

1295

DANIEL W. FOSTER

10. Siperstein, M.D., and V. M. Fagan. Studies on therelationship between glucose oxidation and inter-mediary metabolism. II. The role of glucose oxi-dation in lipogenesis in diabetic rat livers. J. clin.Invest. 1958, 37, 1196.

11. Siperstein, M. D. Inter-relationships of glucose andlipid metabolism. Amer. J. Med. 1959, 26, 685.

12. Goodman, D. S. Preparation of human serum al-bumin free of long-chain fatty acids. Science 1957,125, 1296.

13. Siperstein, M. D., and V. M. Fagan. Studies on therelationship between glucose oxidation and inter-mediary metabolism. I. The influence of glycolysison the synthesis of cholesterol and fatty acid innormal liver. J. clin. Invest. 1958, 57, 1185.

14. Van Slyke, D. D. Studies of acidosis. VII. The de-termination of p-hydroxybutyric acid, acetoaceticacid, and acetone in urine. J. biol. Chem. 1917,32, 455.

15. Weichselbaum, T. E., and M. Somogyi. A methodfor the determination of small amounts of ketonebodies. J. biol. Chem. 1941, 140, 5.

16. Stetten, D., Jr., and G. E. Boxer. Studies in carbo-hydrate metabolism; I. The rate of turnover ofliver and carcass glycogen, studies with the aid ofdeuterium. J. biol. Chem. 1944, 155, 231.

17. Good, C. A., H. Kramer, and M. Somogyi. The de-termination of glycogen. J. biol. Chem. 1933, 100,485.

18. Saifer, A., and S. Gerstenfeld. The photometric mi-crodetermination of blood glucose with glucoseoxidase. J. Lab. clin. Med. 1958, 51, 448.

19. White, J. E., and F. L. Engel. Lipolytic action ofcorticotropin on rat adipose tissue in vitro. J. clin.Invest. 1958, 37, 1556.

20. Dole, V. P. A relation between non-esterified fattyacids in plasma and the metabolism of glucose. J.clin. Invest. 1956, 35, 150.

21. Trout, D. L., E. H. Estes, Jr., and S. J. Friedberg.Titration of free fatty acids of plasma: a study ofcurrent methods and a new modification. J. Lipid.Res. 1960, 1, 199.

22. Lyon, J. B., Jr., and W. L. Bloom. The use of fur-fural for the determination of acetone bodies inbiological fluids. Canad. J. Biochem. 1958, 36,1047.

23. Ochoa, S., J. R. Stern, and M. C. Schneider. Enzy-matic synthesis of citric acid. II. Crystalline con-densing enzyme. J. biol. Chem. 1951, 193, 691.

24. Pearson, D. J. A source of error in the assay ofacetyl-coenzyme A. Biochem. J. 1965, 95, 23 c.

25. Buckel, W., and H. Eggerer. Zur optischen Bestim-mung von Citrat-Synthase und von Acetyl-CoenzymA. Biochem. Z. 1965, 343, 29.

26. Walker, P. G. A colorimetric method for the esti-mation of acetoacetate. Biochem. J. 1954, 58, 699.

27. Kalnitsky, G., and D. F. Tapley. A sensitive methodfor estimation of oxaloacetate. Biochem. J. 1958,70, 28.

28. McCalla, C., H. S. Gates, Jr., and R. S. Gordon, Jr.C1"02 excretion after the intravenous administra-tion of albumin-bound palmitate-l-C" to intact rats.Arch. Biochem. Biophys. 1957, 71, 346.

29. Williamson, D. H., J. Mellanby, and H. A. Krebs.Enzymic determination of D(-)-p3-hydroxybutyricacid and acetoacetic acid in blood. Biochem. J.1962, 82, 90.

30. Bray, G. A. A simple efficient liquid scintillator forcounting aqueous solutions in a liquid scintillationcounter. Analyt. Biochem. 1960, 1, 279.

31. Feigenbaum, A. S., and H. Fisher. Changes in fattyacid composition in nutritional fatty degenerationof the liver. I. Effect of starvation. Brit. J. Nutr.1963, 17, 31.

32. Scow, R. O., S. S. Chernick, and M. S. Brinley.Hyperlipemia and ketosis in the pregnant rat.Amer. J. Physiol. 1964, 206, 796.

33. Tepperman, J., and H. M. Tepperman. Effects ofantecedent food intake pattern on hepatic lipo-genesis. Amer. J. Physiol. 1958, 193, 55.

34. Tubbs, P. K., and P. B. Garland. Variations in tissuecontents of coenzyme A thio esters and possiblemetabolic implications. Biochem. J. 1964, 93, 550.

35. Bortz, W. M., and F. Lynen. Elevation of long chainacyl CoA derivatives in livers of fasted rats. Bio-chem. Z. 1963, 339, 77.

36. Beatty, C. H., R. D. Peterson, R. M. Bocek, andE. S. West. The effect of insulin and deprivationof food on metabolism of acetoacetic acid by mus-cle. J. biol. Chem. 1964, 239, 2106.

37. Langdon, R. G. Hormonal regulation of fatty acidmetabolism in Lipide Metabolism, K. Bloch Ed.New York, John Wiley, 1960, p. 238.

38. Opie, L. H., J. R. Evans, and J. C. Shipp. Effect offasting on glucose and palmitate metabolism ofperfused rat heart. Amer. J. Physiol. 1963, 205,1203.

39. Fritz, I. B., and E. Kaplan. Effects of glucose onfatty acid oxidation by diaphragms from normaland alloxan-diabetic fed and starved rats. Amer.J. Physiol. 1959, 198, 39.

40. Bragdon, J. H., and R. S. Gordon, Jr. Tissue distri-bution of C" after the intravenous injection of la-beled chylomicrons and unesterified fatty acids inthe rat. J. clin. Invest. 1958, 37, 574.

41. Fain, J. N., R. 0. Scow, E. J. Urgoiti, and S. S.Chernick. Effect of insulin on fatty acid synthe-sis in vivo and in vitro in pancreatectomized rats.Endocrinology 1965, 77, 137.

42. Srere, P. A., W. Seubert, and F. Lynen. Palmitylcoenzyme A deacylase. Biochim. biophys. Acta(Amst.) 1958, 33, 313.

43. Srere, P. A., and D. W. Foster. On the proposedrelation of citrate enzymes to fatty acid synthesisand ketosis in starvation. Biochem. biophys. Res.Commun. 1967, 26, 556.

44. Amatruda, T. T., Jr., and F. L. Engel. The role ofthe endocrine glands in ketosis. I. The ketosis offasting. Yale J. Biol. Med. 1959, 31, 303.

1296