Regulation of DNA synthesis in fat cells andstromal elements from rat adipose tissue

C. H. Hollenberg, A. Vost

J Clin Invest. 1968;47(11):2485-2498. https://doi.org/10.1172/JCI105930.

The incorporation of tritiated thymidine into the deoxyribonucleic acid (DNA) of adipose fatand stromal cells was followed under a variety of conditions. After in vitro incubation ofadipose slices or up to 2 days after in vivo injection of the isotope, all DNA radioactivity wasin the stromal cell fraction. From 2 to 15 days after thymidine injection total tissue DNAradioactivity was constant, while between 2 and 5 days after injection label in fat cell DNAincreased markedly. Thus new labeled fat cells, initially collected in the stromal pool,required 2-5 days after completion of DNA synthesis to accumulate sufficient lipid to beharvested in the fat cell fraction. Fasting before thymidine injection practically abolishedDNA synthesis in primordial fat cells and reduced less drastically formation of stromalelements. However fasting sufficient to deplete lipid stores by 50% neither destroyed maturefat cells nor impaired their capacity to reaccumulate fat with refeeding. Other studiesevaluated the role of new fat cell formation in the process of lipid accretion accompanyingrefeeding. These experiments indicated that at least during the early phase of rapid weightgain, accumulation of fat was due to deposition of triglyceride in existing cells rather than toaccelerated formation of new fat cells. Studies with hypophysectomized rats demonstratedthat pituitary ablation variably affected stromal DNA synthesis […]

Research Article

Find the latest version:

http://jci.me/105930-pdf

Regulation of DNASynthesis in Fat Cells

and Stromal Elements from Rat Adipose Tissue

C. H. HOLLENBERGand A. VOST

From McGill University Medical Clinic, Montreal General Hospital,Montreal, Quebec, Canada

A B S T R A C T The incorporation of tritiated thy-midine into the deoxyribonucleic acid (DNA) ofadipose fat and stromal cells was followed undera variety of conditions. After in vitro incubationof adipose slices or up to 2 days after in vivoinjection of the isotope, all DNAradioactivity wasin the stromal cell fraction. From 2 to 15 daysafter thymidine injection total tissue DNAradio-activity was constant, while between 2 and 5 daysafter injection label in fat cell DNA increasedmarkedly. Thus new labeled fat cells, initiallycollected in the stromal pool, required 2-5 daysafter completion of DNAsynthesis to accumulatesufficient lipid to be harvested in the fat cellfraction. Fasting before thymidine injection prac-tically abolished DNAsynthesis in primordial fatcells and reduced less drastically formation ofstromal elements. However fasting sufficient todeplete lipid stores by 50% neither destroyedmature fat cells nor impaired their capacity toreaccumulate fat with refeeding. Other studiesevaluated the role of new fat cell formation in theprocess of lipid accretion accompanying refeeding.These experiments indicated that at least duringthe early phase of rapid weight gain, accumulationof fat was due to deposition of triglyceride inexisting cells rather than to accelerated formationof new fat cells. Studies with hypophysectomizedrats demonstrated that pituitary ablation variablyaffected stromal DNA synthesis and nearly abol-ished the formation and (or) maturation of pri-mordial fat cells. In these animals growth hormonemarkedly enhanced thymidine incorporation intostromal DNAbut had no effect on fat cell pre-

Received for publication 10 April 1968 and in revisedform 7 June 1968.

cursors. In intact animals the predominant effect ofgrowth hormone was also on the stromal fraction,although an action of the hormone of lesser mag-nitude on fat cell precursors was also evident.

INTRODUCTION

While information concerning the accretion andmobilization of adipose lipid has accumulatedrapidly over the past decade, there is still sur-prisingly little known about the regulation of fatcell formation, another mechanism that can influ-ence adipose mass. Hindering work in this area isthe difficulty in the histological recognition ofprimitive fat cells and the fact that only a part oftotal adipose tissue deoxyribonucleic acid (DNA)is contained in fat cells (1). The information thatis available has been derived from the use ofradioautography, chemical DNA analysis, andcounting of adipose cells. These data indicate thatthe new fat cells are continually formed in thegrowing animal (2), that expansion of adiposemass by fat feeding is associated with an increasein total adipose tissue DNA (3), and that inhuman and in some experimental forms of obesityan increase in adipose cell number accompanies anincrease in adipose cell size (4, 5). Althoughthese data suggest an influence of diet on fat cellsynthesis, the precise nutritional and hormonalfactors that control the formation and maturationof these elements remain to be defined.

In order to approach this problem, tritiated thy-midine was injected into rats of varying nutritionalstates and adipose tissue removed at later times.The cellular constituents of the tissue were sep-arated into two populations, mature fat cells andstromal vascular elements, by collagenase diges-

The Journal of Clinical Investigation Volume 47 1968 2485

tion (6), and the specific activity of the DNA ofthe two cell pools followed in time. This approachpermitted observations of the maturation time offat cells and of the effects of nutritional variationsand of growth hormone on DNAsynthesis in bothcell populations. Other studies were made of theeffect of fasting on the cellular integrity of maturefat cells.

METHODSMale Wistar rats, initial weight 150-175 g and maintainedon Purina chow, were used in all studies except thosewith hypophysectomized animals. In the latter instance,male Sprague-Dawley animals, initial weight 140-160 g,obtained from Charles River Laboratories (Boston,Mass.) were used no sooner than 10 days after hy-pophysectomy. Both in vitro and in vivo experimentswere performed.

In the in vitro studies, epididymal and lumbar fatpieces from three animals were pooled and sliced, andapproximately 2.5 g were added to 12 ml of Eagle's Lmedium and incubated for varying periods in a Dubnoffshaker at 370C; the gas phase was 95%o 02 and 5%o Co2.To this medium was added 5%o bovine albumin, 200 mg/100 ml of glucose, 0.02 M NaHCO3, 100 U/ml of peni-cillin, 20 /Ag/ml of streptomycin, and varying concentra-tions of unlabeled and tritiated thymidine (thymidine-methyl-8H, [Schwarz Bio Research, Inc., Orangeburg,N. Y.] 6c/mmole). After incubation, the fat pieces weredigested with collagenase in the presence of penicillin,streptomycin, and 1.85 mmunlabeled thymidine, and thespecific activity of DNA in fat cells and stromal cellswere determined as later described.

In the in vivo studies, two schedules were used forintraperitoneal administration of tritiated thymidine.In some studies, a single injection of 22 /Ac/rat was fol-lowed in 1 hr by administration of 80 tnmoles of un-labeled thymidine; animals were sacrificed 20 min afterthe second injection. In other experiments, each rat re-ceived a total of 30 /Ac of labeled thymidine in two in-jections given at 10 and 11 p.m., and animals were sac-rificed at varying times, the earliest being 12 hr afterinjection of the isotope. In experiments in which animalswere followed over days, 25 /Amoles of unlabeled thy-midine was administered daily beginning 12 hr afterisotope injection.

Animals in groups of three were anesthetized withether and exsanguinated by aortic puncture. Epididymaland lumbar fat from the three animals was rinsed in bi-carbonate buffer, pooled minced, and then the fat cellsand stromal elements were separated by the collagenasemethod of Rodbell (6). Despite repeated washes of the fatcells, a pinkish zone, undoubtedly representing stromalfractions, was consistently evident at the fat cell bufferinterface after centrifugation. This zone was found tocontain most of the DNA in the fat cell fraction. Whileit is unlikely that this contamination would influence ex-periments concerned with carbohydrate and lipid metabo-

lism of free fat cells, it would completely vitiate studiesof fat cell DNA synthesis. Several additional procedureswere therefore introduced to reduce stromal contamina-tion. After collagenase digestion, fat cells and stroma wereseparated by 40-sec centrifugation in a clinical centri-fuge; the stroma was aspirated and saved. The fat cellswere redispersed in warm buffer-albumin and filteredthrough a single layer of fine silk hose; the hose re-tained many fine strands of tissue and virtually no lipid,and the residue was transferred to the stromal pool.The filtered fat cells were then washed twice more inbuffer-albumin followed by a third wash in buffer alone;after each wash stromal sediment was added to thestromal pool. While the filtering procedure removed mostof the visible stromal elements, when the fat cells weretransferred with buffer to a 77 X 10 mm plastic tubeand spun for 1 min, a faint thin grayish-pink line wasusually evident at the fat cell buffer interface. This in-terfacial zone was removed from the fat cell pool byfreezing the entire fat cell column with a Cryokwikspray (International Equipment Co., Needham Heights,Mass.) and by slicing the frozen cell column 1-2 mmabove and just below the interface. The upper 30-40 mmof fat cells were processed separately from the inter-facial and stromal elements. An indication of fat cellbreakage and hence of potential nuclear loss from fatcells was obtained by measuring the amount of oil pro-duced during the entire procedure. This amount rarelyexceeded 5% of the triglyceride recovered from the fatcell fraction.

The fat cell and stromal fractions were separatelyhomogenized in 15 ml of cold acetone using a motor-driven Teflon pestle. After they were filtered, the precipi-tates were defatted with 100 ml of cold acetone followedby 50 ml of ethyl ether and then digested overnightin 6 ml of 0.5 N KOH at 370C. DNA was precipi-tated in the cold from the digests by addition of per-chloric acid to 0.5 N, collected by centrifugation, andwashed twice more with 5 ml of 0.5 N perchloric acid.The DNA in the sediment was then hydrolyzed bytreating twice with 1-2 ml of 0.5 N perchloric acid at80'C and aliquots of the pooled extracts taken for radio-active counting and for DNA determination by thediphenylamine reaction, with calf thymus DNA asstandard (7). The acetone-ether washes were saved,evaporated, brought to volume with acetone, and ali-quots taken for radioactive counting and triglyceridedetermination (8).

A number of studies were performed to determine theadequacy of the lipid extraction, washing, and digestionprocedures. The final cold perchloric acid washes weredevoid of radioactivity and, when tritiated thymidine wasadded to unlabeled tissue at the time of homogenization,all radioactivity was recovered in the washes and nonein the hot perchloric acid extracts. When radioactivetriolein was added at the time of homogenization, all ra-

dioactivity was in the acetone-ether washes and none inthe hot perchloric digests. The efficiency of the two hotperchloric acid extractions was established by finding thata third extraction increased recovery of radioactivity

2486 C. H. Hollenberg and A. Vost

by only 6%. Radioactivity in the residual pellet, dis-solved by boiling in 2 N KOH, was inconsequential. Itwas also found that the specific activity of DNA wasthe same in both the first and second hot perchloric acidextracts.

The procedure described produced a distribution ofDNAbetween fat and stromal cells quite different thanthat reported by Rodbell (1). In the present study fatcells contained 15%, interfacial cells 1%, and stromalcells 84% of the recovered DNA. As it has been previ-ously shown that with the collagenase method total re-covery of whole tissue DNA is not achieved (1), dataconcerning absolute content of chemical DNA in thevarious cell fractions must be interpreted with caution.Nonetheless the procedure led to very reproducible re-coveries of fat cell and stromal DNA; in four studiesin which the adipose tissue pool was divided into twoaliquots of equal weight before collagenase digestion,the difference between the two aliquots in fat cell DNAcontent ranged from 3-9%, in stromal DNAfrom 2-10%.-Recoveries of radioactive DNAin the two cellular poolswere equally reproducible.

In some experiments, the specific activity of liver DNAwas determined 80 min after injection of tritiated thy-midine. The liver pieces were homogenized in 0.25 M su-crose containing 0.002 M CaCl2 and spun at 800 g for 15min. The nuclear pellet was resuspended in sucrose-CaCl2 and -the centrifugation repeated. The washed pelletwas dissolved in KOH and radioactive and chemicalDNAdetermined as previously noted. In one experiment,the specific activity of liver DNA was obtained as de-scribed and was compared to that noted when liver washomogenized in acetone and carried through the sameprocedures used for fat and stromal cells; the specificactivities were identical.

In a number of experiments, plasma obtained at thetime of exsanguination was used for determination ofplasma water radioactivity. Plasma water was preparedfrom whole plasma by microsublimation (9). In otherstudies the nature of the radioactivity present in theacetone-ether washes of fat cells was characterized bythin-layer chromatography and by saponification (8).

The radiochemical purity of each batch of thymidine-'H was established as > 99%, by the supplier (SchwarzBio Research, Inc.) using radioautography of paperchromatograms run in two different solvent systems;also, microsublimation of aqueous thymidine-'H resultedin negligible recovery of radioactivity in the volatilefraction. The bovine growth hormone used in thesestudies was obtained from the National Institutes ofHealth, Bethesda, Md. (NIHGH B 13). Radioactivitywas counted in a Nuclear-Chicago liquid scintillationspectrometer by methods previously described (8) andwas converted to disintegrations per minute by using anexternal standard and by correcting for quenching bychannel ratio techniques.

RESULTS

In Vitro incorporation of tritiated thymidineinto adipose tissue DNA. When adipose slices

were incubated with tritiated thymidine and theslices subsequently digested with collagenase, al-most all of the DNA radioactivity was in thestromal cells (Table I). Fat cell DNA usuallycontained 1-2%o of the radioactivity recovered inDNA, an amount that was probably due to re-sidual contamination of this fraction with stroma.Incorporation of thymidine was nearly linear dur-ing 3 hr of incubation, and extent of incorporationwas independent of whether the medium wasEagle's L containing 5%o albumin and glucose, orKrebs-Ringer bicarbonate buffer with albuminonly.

When isolated fat cells were incubated withtritiated thymidine, no radioactivity was recoveredin DNA. Isolated stromal elements did incorporatethymidine, but extent of incorporation was irregu-lar possibly because of damage to the stromal cellsproduced by the collagenase or the washing pro-cedures.

Table II indicates the effect of medium concen-tration of thymidine and of prior fasting of theanimals on incorporation of thymidine. In thesestudies adipose slices were incubated with tritiatedthymidine and radioactivity in the DNAof sub-sequently isolated stromal cells determined. Abovea medium thymidine concentration of 5 Mmoles/liter, incorporation was nearly independent of the

TABLE IIn Vitro Incubation of Adipose Slices with Tritiated

Thymidine. DNARadioactivity in Fat Cellsand Stromal Cells*

Thymidineincorporationt %Total

DNAradio-Duration of Strom'l Fat activity in

Expt. incubation cells cells stromal cells

Mi,.

1 60 513 40 99180 1348 204 92

2 180 1270 80 983 180 862 93 974 180 2090 95 99

DNA, deoxyribonucleic acid.* All samples were incubated in Eagle's L medium. Inexperiments 1-3, the specific activity of thymidine in themedium was 29.4.uc/Mmole at a medium concentration of5 pumoles/liter. In experiment 4, the specific activity ofthymidine was 5.9 pc/,smole and the medium concentra-tion, 250;&moles/liter.t /u4moles of thymidine incorporated/mg of DNA.

DNA Synthesis in Adipose Tissue 2487

TABLE I IEffect of Medium Concentration of Thimidine and of

Fasting on In Vitro Incorporation of Thymidineinto Adipose Tissue Stromal DNA*

Thymidine incorporation:Medium concentra-

Expt. tion of thymidine Fed 48-hr fasted

umoles/liter1 5 1485

250 2080

2 5 1270 413 250 2090 95

DNA, deoxyribonucleic acid.* In all experiments adipose slices were incubated for 3 hr.In experiment 1, the medium was Krebs-Ringer bicarbo-nate, in experiments 2 and 3, Eagle's L; all media weresupplemented with albumin and glucose.Jpu~moles of thymidine incorporated/mg of stromal DNA.

quantity of thymidine added, for a 50-fold increasein medium concentration led to less than a 2-foldincrement in incorporation. Tissue obtained from48-hr fasted animals incorporated less thymidinethan did tissue from fed animals, and this differ-ence was as marked at a medium thymidine con-centration of 250 /Amoles/liter as at 5 /umoles/liter.If it is assumed that the high medium thymidineconcentration increased the size of the intracellularthymidine pool, the latter finding suggests thatthe reduction in incorporation noted with tissuefrom fasted rats was not due to label dilution inan expanded nucleotide pool. If this had been thecase incubation with a high concentration ofthymidine would have reduced differences betweenfed and fasted groups in nucleotide pool size andthus would have reduced differences in extent ofincorporation.

Although incorporation of tritiated thymidineinto the DNAof fat cells was small irrespective ofwhether adipose slices or free fat cells were used,consistently, with both preparations, radioactivitywas recovered in the acetone-ether extracts ofthese cells. It was also noted that when adiposeslices were incubated in Krebs-Ringer bicarbon-ate, the presence of glucose increased the radio-activity of the acetone-ether washes 40-fold, while,as prevoiusly noted, glucose had no effect onrecovery of radioactivity in stromal DNA. Thenature of the radioactivity in acetone-ether wasdetermined by thin-layer chromatography; 75%of the label migrated with either triglyceride

2488 C. H. Hollenberg and A. Vost

(577%o), free fatty acid (4%o), or diglyceride(13%), while only 21% remained at the origin,the site where thymidine was recovered in thesame system. When the triglyceride was elutedand saponified, the saponified fraction contained90%o of the radioactivity all of which ran withfree acid on thin layer. Hence the bulk of theradioactivity in acetone-ether appeared to be inglyceride fatty acid. The influence of mediumglucose on extent of incorporation of label intothis fraction suggested that tritium derived fromthymidine had been incorporated into fatty acidssynthesized during incubation. Tritium could havebeen made available for incorporation into fattyacids by formation of tritiated water or by pro-duction of other thymidine metabolites such asfl-aminoisobutyric acid (10) which could con-ceivably serve as fatty acid precursors. Tritiatedwater was in fact produced during incubation butonly when adipose slices were present; after 3 hrof incubation 1.5%o of the added radioactivity waspresent in medium water.

Specifi activity of adipose tissue and liver DNA80 min after injection of tritiated thymidine. Inthe first series of in vivo experiments, labeled thy-midine was administered to fed, fasted, and refedrats and adipose samples and, in some instances,liver removed 80 min later. The data derivedfrom these studies, shown in Fig. 1, are expressedper unit of total tissue DNA. As was found in thein vitro experiments, 48 hr of fasting markedlydepressed thymidine incorporation into adiposetissue DNA. When 48-hr fasted rats were refedfor 3 days, values similar to those observed in fedanimals were obtained, while the specific activityof adipose DNAderived from animals refed for 5days was usually higher than in continuously fedanimals. In fed animals, no correlation was notedbetween the specific activities of adipose and liverDNA; however, as-with fat tissue, fasting reducedincorporation of thymidine into liver DNA. Inconformity with the results of the incubationstudies, when fat tissue was removed from fedanimals 60 min after injection of labeled thymi-dine and digested with collagenase, almost all ofthe DNAradioactivity was present in the stromalfraction (Table III).

Radioactivity was also present in the acetone-ether washes of the adipose samples removed afterthymidine injection (Fig. 2). The radioactivity

rM x at/ADWOuDN

2

301

20-

10

ADIPOSE TISSUE

6

S

4

3

2

FED I1 2-3 S 1-11REFED

LIVER

SWMx WS3/MeUVIr DNA

±

FIGURE 1 Specific activity of adipose and liver DNA80 min after thymidine injection. Heights of bars andbrackets represent means and standard errors of themean, respectively. Data are expressed per unit of totaltissue DNA, and the number of experiments in eachseries is shown at the foot of the columns. In the fastedseries, all groups were fasted for 48 hr except one inwhich this interval was 18 hr; in the adipose tissue ex-periments the fasted data were derived from five groups.In the refed experiments all animals were starved for48 hr before the start of refeeding. The duration of therefeeding interval in days is shown below the variouscolumns.

present in these washes fell with fasting, increasedmarkedly when fasted animals were refed for 2-3days, and then returned toward levels seen in fedanimals with longer periods of refeeding. Thispattern is of course very reminiscent of the changesin lipogenesis induced in adipose tissue by fastingand refeeding. Thin-layer chromatography of anacetone-ether extract of tissue obtained from a5 day refed group revealed that 75% of the radio-activity travelled with triglyceride, 9% with di-glyceride, 10%o with free fatty acid, and 6% re-mained at the origin. When the triglyceride wassaponified, 80% of the radioactivity was in thesaponified fraction. Thus, as in the in vitro studies,most of the radioactivity recovered in acetone-ether had the characteristics of glyceride fatty acid.

Specific activity of fat cell and stromal DNAvarying times after injection of tritiated thymidine.It was evident from the preceding studies thatduring brief in vitro or in vivo exposure of adiposetissue to tritiated thymidine, radioactivity was in-corporated only into stromal DNA. These resultswere anticipated since there is adequate histologi-cal evidence that mature fat-filled adipose cellsdo not display mitotic activity (11). It was con-

I

TABLE IIIDNARadioactivity in Fat Cells and Stromal Cells

60 min after Injection of Tritiated Thymidine

DNAspecificDNAradioactivity* radioactivity

Stromal Fat Stromal FatExpt. cells cells cells cells

1 36.60 0.28 69.0 2.52 49.25 0.44 89.5 5.1

eI Animals in groups of three were injected with 30 juc oflabeled thymidine/rat and fat removed 1 hr later.

1* a 4 * dpm X 103goFEAS dmp X 10-3/mg of DNA.

sidered likely however that among the labeledstromal elements were new, primordial fat cellswhich would only be collected in the fat cell poolafter differentiation and accumulation of lipid.Thus experiments were designed to follow fat celland stromal cell DNAradioactivity over a longerperiod of time. In these studies, simultaneous in-jection of many animals with tritiated thymidinewas followed by sacrifice of groups of three ratsat various times from 12 hr to 15 days later. Thesedata are shown in Fig. 3; 0 time values representtissues excised 12 hr after injection.

During the first 2 days of the study, radioactivityin stromal DNA fell markedly. Thereafter agradual decline in stromal specific activity was

REFED

FIGuRE 2 Radioactivity in acetone-ether washes of adi-pose tissue removed 80 min after thymidine injection.Data are expressed per unit of total tissue DNA. Means,standard errors of the mean, numbers of experiments,and duration of refeeding are as shown in Fig. 1.

DNA Synthesis in Adipose Tissue 2489

~~~~~R 0 r 2k10I

T

m la

220 DPMx 10'3/MGDNA

200

180 --x STROMALCELLS1 o-no FAT CELLS

160A140 -

120 -

100 \80 I

60.

40 - ~ ~ -E.20

0 2 4 6 8 10 114 1%DAYS

FIGURE 3 Specific activity of fat cell and stromal DNAvarying times after thymidine injection. Data are ex-pressed per unit of fat cell or stromal DNA. The ar-row indicates the time of thymidine injection which was12 hr before 0 time. Six groups of animals were used ateach period except at days 12 and 15 when five groupswere sacrificed. Brackets indicate standard errors of themean.

observed which was due in part to an increase inunlabeled stromal DNAthat occurred with weightgain. While the reason for the abrupt decline instromal radioactivity is not known, it may havebeen due to rapid turnover of a fraction of thecellular elements of this pool or to migration fromthe tissue of cells such as leucocytes and macro-phages labeled in situ. At 0 time only 1% of thetotal DNAradioactivity was in fat cells, a propor-tion which was identical with that noted in the invitro and short-term in vivo experiments andwhich probably represents the extent of stromalcontamination of the fat cell pool.

The appearance of radioactivity in fat cell DNAfollowed a pattern quite distinct from that of thestromal cells. The specific activity of adipocyteDNA rose slightly from 0 to 2 days, markedlyfrom 2 to 5 days, and slowly thereafter. It was

clear that this change in specific activity was dueto redistribution of radioactivity between fat celland stromal DNA and not to net accretion oflabel in DNAover the experimental interval. Asshown in Table IV, after the initial decline intotal DNA radioactivity that occurred between 0and 2 days, radioactivity in the total cellular poolremained constant up to 15 days. Thus the in-crease in fat cell radioactivity that was evident at-5 days was due to a diversion of DNAlabel fromthe stromal into the fat cell pool, a process thatwas essentially complete 7 days after injection ofthe isotope. It is of interest that of the radio-activity present in stromal DNA12 hr after thy-midine administration, only about 10% was ulti-mately recovered in the fat cell pool, and thatthere was no correlation between the stromal spe-cific activity at 0 time and fat cell specific activity7-15 days later.

These data are entirely consistent with the con-cept that within the stromal fraction was a pool ofprimordial fat cells of variable size, and that thesecells required 2-5 days after completion of DNAsynthesis to accumulate enough lipid to be har-vested in the fat cell fraction. It was also evidentthat estimates of DNA synthesis in primitive fatcells could be made on the basis of the specificactivity of the fat cells collected at least 7 daysafter thymidine administration, a period sufficientfor maturation of these elements.

When tissue derived from 48-hr fasted rats wasused it was found that the minimum amount ofcellular lipid necessary to allow collection of cellsin the fat pool was 0.010-0.015 ug/cell, whereasthe average lipid content of the fat cells used in theexperiments shown in Fig. 3 was 0.080 jug/cell.1

1 These calculations were based on the assumptionthat the DNAcontent per cell was 7 X 10' Ag.

TABLE IVRadioactivity in Adipose DNA Varying Times after Thymidine Administration*

Days after injection ... 0 2 5 7 12 15

Total DNAradioactivityl 35.91 1 4.41 18.15 i 1.86 20.04 i 2.26 19.57 :1 3.87 19.01 ± 1.93 17.25 :4 2.42%Total DNAradioactivity 1.0 i 0.3 3.4 i 1.3 11.2 ± 2.5 14.6 i 3.0 14.4 i 2.6 22.7 i 8.1

in fat cells

2490 C. H. Hollenberg and A. Vost

DNA, deoxyribonucleic acid.* Total DNAradioactivity represents sum of radioactivity in fat cell and stromal fractions. Six groups of rats were usedat each interval except at days 12 and 15 when five groups were sacrificed. Standard errors are indicated.t dpm X 10-3.

It is of interest that Hirsch, using rats of similarsize but entirely different methods of fat cell col-lection and sizing, also found the mean lipidcontent per cell to be 0.080 ug (4), while Gold-rick who employed the collagenase system followedby repeated fat cell washes was also able to collectfat cells containing as little as 0.01 pg of lipidper cell (12).

Effect of fasting and short periods of refeedingon thymidine incorporation into fat cell andstromal DNA. While the 80-min in vivo studieshad demonstrated that fasting markedly diminishedincorporation of thymidine into adipose DNAandthat a short period (2-3 days) of refeeding didnot increase total DNAsynthesis above that seenin fed animals, these experiments did not providedata as to the possible differential behavior of fatcells and stromal elements under these conditions.In particular the possibility had not been excludedthat hyperplasia and more rapid maturation ofprimordial fat cells had accompanied the rapidweight gain seen. during the first several days ofrefeeding. To explore this problem experimentsanalogous to those shown in Fig. 3 were per-formed with, in this instance, fasted and fasted-refed rats.

In the fasted series, animals were starved for36 hr, labeled thymidine administered, and therats kept without food until the initial group wassacrificed 12 hr after thymidine injection. The restof the animals were then allowed to feed ad lib.and were sacrificed at varying times. To study theeffects of refeeding, rats were fasted for 48 hr andthen refed chow. 36 hr after the start of refeeding,the time at which the per cent of increase in ratweight was most rapid, all animals in this serieswere injected with labeled thymidine, and 12 hrafter the injection the first group of refed rats wassacrificed. Adipose tissue was obtained from othergroups at later times. The data from these studiescompared to the results obtained in fed animalsthat had been matched for weight and injectedsimultaneously with the refed animals are shownin Fig. 4. 0 time values were derived from tissuesremoved 12 hr after thymidine administration.

In the experiments with fasted animals, radio-activity in fat cell DNA was barely detectableeven when tissue was removed 2 wk after thymi-dine injection, hence 2 wk after completion of fast-ing. This interval was sufficient to permit complete

FAT CELLS STROMALCnLS

0-.O PMS*_-- Ua

- M

DAYS

FIGURE 4 Effect of fasting and of refeeding on incor-poration of thymidine into fat cell and stromal DNA.Data are expressed per unit of fat cell or stromal DNA.The number of groups in the fed series has been indicatedin Fig. 3. In the refed series five groups of animals wereused at each time, while in the fasted series one groupwas sacrificed at 0 time, one at 6 days, and five at 14days. The arrow indicates the time of thymidine injection,the brackets the standard errors.

restoration of lipid stores as estimated by lipidcontent of the tissues removed and lipid content perfat cell and hence was presumably sufficient to en-sure complete harvesting of all labeled cells thathad differentiated into adipocytes. Incorporationof thymidine into stromal DNAwas also markedlyreduced by fasting, but this reduction was not ascomplete as noted with fat cells; consistently, thespecific activity of stromal DNAexceeded that offat cell DNA.

Although 36 hr of refeeding restored to a con-siderable extent incorporation of thymidine intofat cell DNAand had a similar but less markedeffect on the stromal fraction, the rate of DNAsynthesis in fat cells and stroma of refed animalswas always less than in fed rats. Furthermore,maturation of fat cells occurred no more rapidlyin the refed than in the fed series; in both in-stances radioactivity in fat cell DNA increasedmarkedly between day 2 and day 5 of the study.

From the data obtained in the refed experimentsit was possible to assess the contribution of newfat cell formation to the process of lipid repletionthat occurred during the refeeding interval. It wasobvious that, relative to continuously fed animals,refeeding for 36 hr had not induced hyperplasiaof primordial fat cells nor more rapid maturationof these elements. Furthermore it was evident thatduring the first 2 days of refeeding, an intervalduring which weight gain and lipid accumulation

DNA Synthesis in Adipose Tissue 2491

was rapid, any new fat cells formed would not havematured and thus would not have contributed sig-nificantly to the lipid stores of the tissue. Henceduring this interval lipid repletion must have beendue entirely to accretion of fat by existing butdepleted fat cells. Between the 2nd and 4th dayof refeeding (day 0-day 2 of the study), the lipidcontent of the excised tissues and the lipid contentper fat cell increased from 30 to 60%o of fed values.While it is not certain that the lipid deposited dur-ing this interval was also in existing cells pre-dominantly, it is likely that this was so. Becauseof the time required for maturation of new fatcells, only those cells formed during the first 24 hrof refeeding could conceivably have been in thefat cell pool by the end of the 4th refeeding day(day 2 of the study). As new fat cell formationwas negligible at the end of the fasting period andstill less than that seen in fed animals 36 hr afterrefeeding, it is unlikely that synthesis of new fatcells had contributed in a major fashion to therapid accretion of lipid that occurred not only dur-ing the first 2 days of refeeding but also during theensuing 48 hr interval. The possibility that hyper-plasia of primordial fat cells was a significantfeature later in refeeding remains to be explored.

Radioactivity in acetone-ether washes after thy-midine injection. As in the in vitro and short-term in vivo experiments, radioactivity was con-sistently present in the acetone-ether washes offat cell fractions derived from tissues excised 12hr or longer after thymidine administration (TableV). Once again, most of this radioactivity had thecharacteristics of glyceride fatty acid, and incorpo-ration of label into this fraction was greater inrefed than in fed animals.

In a series of continuously fed animals, radio-activity in fat cell lipid was followed up to 25 daysafter thymidine injection, and these data, ex-pressed per milligram of fat cell DNA, are shownin Fig. 5. Radioactivity in this fraction increasedup to 10 days after isotope administration andthereafter gradually declined; the decline was dueboth to a reduction in recovered radioactivity andto the increase in DNAcontent of the fat cell poolthat occurred with weight gain. As intravenouslyinjected thymidine disappears very rapidly fromthe circulation (10), it was evident that the con-tinued labeling of adipose lipid must have beendue to persistence in the circulation of a tritiated

TABLE VRadioactivity in Acetone-Ether Washes of Fat Cells

12 hr after Thymidine Administration

Fed 36-hr refed

dpm X 10-3/mg of fat cell DNA 513 1030

%Distribution of radioactivity*CE 2 tr.TG 93 99FFA 2 tr.DGand Chol 2 tr.PL 1 tr.

%of TGradioactivity in 71 97saponified fraction

* DNA, deoxyribonucleic acid.* CE, cholesterol ester; TG, triglyceride; FFA, free fattyacid; DG, diglyceride; Chol., cholesterol; PL, phospho-lipid.

derivative of thymidine that could be incorporatedinto adipose tissue fatty acid. Irrespective of theroute by which tritium entered fat cell lipid, netaccumulation of label was essentially complete 10days after injection. Thus abrupt reductions inadipose lipid radioactivity induced over a briefperiod after this time must have been due tochanges in the rate of fat mobilization. In laterstudies use was made of this fact to estimateextent of lipid depletion induced by fasting.

In searching for a tritiated thymidine metabolitethat could be incorporated into adipose lipid, at-tention was focused on production and turnover oftritiated water which is known to accumulate

DPMX 103/MGFAT CELL DNA

400

300

200

100

0 2 4 6 8 10 12 14 16 18 20 22 24 26DAYS

FIGURE 5 Radioactivity in acetone-ether washes of fatcells varying times after thymidine injection. The datawere derived from fed animals, and the number of groupsused at the varying times were: 0 time, nine; 2 days, nine;5 days, seven; 7 days, eight; 9 days, five; 12 days,seven; 15 days, eight; 25 days, five. Brackets representstandard errors.

2492 C. H. HoUenberg and A. Vost

shortly after injection of tritiated thymidine (10).Fig. 6 indicates the change in plasma water radio-activity from 12 hr (0 time) to 15 days afteradministration of labeled thymidine. If it is as-sumed that newly formed tritiated water hadequilibrated with total body water by 0 time, itwas evident that over the 12 hr subsequent tothymidine injection, 40-50%o of the administeredradioactivity had been incorporated into bodywater. This figure is very similar to that obtainedin man (10). Radioactivity in plasma water de-clined exponentially at the same rate in fed andrefed animals and 10 days after injection, the timeat which net accretion of radioactivity in adiposelipid had ceased, plasma water contained only 10%oof the radioactivity present 12 hr after isotopeinjection.

In a few observations made in animals fasted

500

400

300

2001

oa90so7060so40

30

20~

DPMX 10-3/ml PLASMAH20

(x FED*REFED

I0

0 2 4 0 5DAYS

10 12 14

FIGURE 6 Radioactivity in plasma water varying timesafter thymidine injection. At each of days 0, 2, and 5 thefed series was comprised of four groups, at days 7, 12,and 15, of two groups. The animals in the refed serieshad been fasted for 48 hr and fed for 36 hr before thy-midine injection (given 12 hr before 0 time), and thisseries was comprised of three groups at days 0, 2, and5, two groups at days 7 and 12, and one at day 15.Brackets indicate standard errors of the fed series.

before thymidine administration and then allowedto feed, a similar disappearance rate of tritiatedwater was observed. Although very little radio-activity (or lipid) was recovered in the acetone-ether washes of fat cells obtained from these ani-mals before feeding was started, when tissueswere removed varying times during the feedingperiod, radioactivity in adipose lipid was not onlydemonstrable but usually greater than in samplestaken from fed animals at the same time. Thusthe rebound in lipogenesis produced by refeed-ing resulted in more extensive incorporation oftritium, possibly derived from tritiated water, intonewly formed lipid. However, as shown in Fig. 4,despite the presence of tritiated water throughoutthe period of observation of the fasted group anddespite enhanced incorporation of label into adi-pose lipid, very little radioactivity was recoveredin fat cell DNA. Furthermore no increase instromal DNA radioactivity was evident over theperiod of observation. These data provided con-vincing evidence that tritium derived from tritiatedwater was not incorporated into the DNAof eithercellular pool.

Effect of fasting on mature fat cells. From thedata obtained in refed animals it was concludedthat repletion of adipose lipid during at least theearly stages of refeeding had been accomplished byrefilling of existing cells. This conclusion impliedthat those cells depleted of fat by starvation hadsurvived and had remained capable of lipid accu-mulation with refeeding. This assumption wasexamined in the next study.

Fed rats were injected with tritiated thymidineand divided into two groups, one of which wasfed continuously for 25 days before sacrifice. Theother group was fed for 10 days, a period longenough to ensure maturation of fat cells, fastedfor 2 days, and then fed again for 13 days. Thefinal feeding interval was sufficient to restorelipid content of excised tissues and lipid contentper fat cell to values seen in continuously fed ani-mals and hence should have been long enough toallow complete harvesting of previously depletedfat cells. If fasting had destroyed depleted fat cellswith loss of DNAor had significantly reduced thecapacity of these cells for subsequent lipid accre-tion, recovery of radioactivity in fat cell DNAwould have been substantially less in this group.Alternatively if these cells had retained their in-

DNA Synthesis in Adipose Tissue 2493

tegrity and their ability to reestablish activities ofthose enzymes involved in lipid synthesis andassimilation, fat cell DNA radioactivity wouldhave been comparable in the two series.

As shown in Table VI, the two groups of ani-mals displayed very similar values not only forlipid content but also for radioactivity and spe-cific activity of fat cell and stromal cell DNA.However, radioactivity in fat cell lipid was 50%oless in the fasted than in fed animals, a valuewhich should approximate the extent of lipid de-pletion produced during the fasting period. Whilevariation between groups precluded detection ofsmall changes, it is evident that fasting sufficientto reduce lipid stores by one-half produced neitherextensive destruction of fat cells or stromal ele-ments nor significant impairment in the capacityof depleted fat cells to accumulate lipid withrefeeding.

Effect of growth hormone on thymidine incor-poration in fat cell and stromal DNA. Growthhormone is known to exert a variety of effects onadipose tissue metabolism; these include enhance-ment of fat mobilization (13), ultimate inhibitionof fatty acid synthesis (14), and increased invitro incorporation of thymidine into the DNAofadipose pieces when the hormone is given before

TABLE VIEffect of Fasting on Mature Fat and Stromal Cells*

Fed Fasted

Radioactivity in fat cell 7.0 :1: 1.3 8.5 i 0.9DNAt

Radioactivity in stromal 25.2 i 6.9 27.8 i 6.8DNAt

Specific activity in fat 30.2 :1: 5.6 38.2 3= 4.7cell DNA§

Specific activity in 19.2 i 6.1 19.9 1 6.9stromal DNA§

Radioactivity in fat cell 64.1 :h 8.8 33.8 :1 4.011lipidt

Radioactivity in fat cell 287.5 ± 44.8 156.6 4 17.411lipid/mg of fat cell DNA

DNA, deoxyribonucleic acid.* Means and standard errors of six experiments are shown.Fasted series represents those animals starved for 2 daysbeginning 10 days after thymidine injection and refed for13 days before sacrifice.t dpm X 10-3.§ dpm X 10-3/mg of DNA.1t Fed vs. fasted P < 0.01.

removal of the tissue (15). While the effects ofthis agent on free fatty acid release and lipo-genesis are undoubtedly due to alterations in fatcell function, it is not known which of the cellularelements of the tissue responds to the hormonewith an increased rate of DNAsynthesis. Indeedit would appear paradoxical that a hormone thatacts to reduce adipose lipid would also have apronounced effect on the rate of multiplication ofthose cells destined to store fat. An effect ofgrowth hormone on increasing growth of support-ing and vascular tissue would be a more under-standable phenomenon. A series of experimentswere therefore performed to determine whetherthis hormone stimulates DNA synthesis in boththe fat cell and stromal fractions of adipose tissueor whether it has a more discriminatory effect.

In these studies growth hormone was adminis-tered over 36 hr to fed hypophysectomized rats,fed intact animals, and intact animals fasted dur-ing the period of hormone administration. Thymi-dine was injected into control and growth hor-mone-treated groups 6 hr after the last dose ofhormone, and animals were killed no sooner than8 days and more commonly 12-16 days after iso-tope administration. The results are shown inFig. 7. In the control hypophysectomized group,incorporation of thymidine into stromal DNAwas'variable, while consistently little if any radioactiv-ity was recovered in fat cell DNA. When growthhormone was given to hypophysectomized animals,DNAsynthesis in the stromal fraction was mark-edly increased, but the hormone had no apparenteffect on the incorporation of thymidine into theDNA of fat cell precursors. A differential effectof the hormone on stromal and fat cell DNAcon-tent was also evident. Despite the fact that growthhormone had been given 1-2 wk before tissuesampling, the quantity of DNAextracted from thestromal pool of the treated animals was signifi-cantly greater than that obtained from control rats.No such difference was evident in fat cell DNAcontent.

It was thought possible that the reduction inDNAradioactivity noted in fat cells of hypophy-sectomized animals was due to the smaller amountof food consumed by these animals as compared tonormal rats. However several facts were not inkeeping with this interpretation. The lipid contentper fat cell in both the control and hormone-

2494 C. H. Hollenberg and A. Vost

HYPOXFED INTACT FED INTACT FASTED

240220200

180160140120100

8060

40

20n

DPMX WYVMODNA

r--" -e

FAT STROMA FAT STOM~A FAT sTRxACELLS CELLS CELLS

FiGuRE 7 Effect of growth hormone on incorporation of thymidine intofat cell and stromal DNA. Data are expressed per unit of fat cell orstromal DNA; the height of the entire bar represents the results de-rived from the growth hormone experiments, the hatched area the con-

trol studies. Brackets indicate standard errors. All growth hormone-treated rats received 3.0 mg of bovine growth hormone I-P in divideddoses over 36 hr and were given tritiated thymidine 6 hr after the lastgrowth hormone injection. Control animals were either not injected or

given saline.

In the hypophysectomized series (hypox. fed) there were eight controland five growth hormone-treated groups. Adipose tissue was removed8-16 days after thymidine injection. In the intact fed series there were30 control and 9 growth hormone-treated groups, and in this seriestissue was removed 12-16 days after isotope administration. In the intactfasted series 13 control and 9 growth hormone-treated groups were used;the hormone was administered during the 2 day fast, and feeding was

begun 12 hr after thymidine injection. Tissue was obtained 12-16 daysafter the beginning of the refeeding period. The differences between thecontrol and growth hormone groups were significant (P < 0.01) in thefollowing instances: hypophysectomized stroma, fed intact stroma, fastedintact fat cells and stroma.

treated hypophysectomized animals was 0.06-0.07,ag, a value similar to that obtained in intact ratsof the same weight and much greater than seen

in fasted intact rats. Moreover in control hypophy-sectomized animals, the specific activity of stromalDNAfar exceeded that of fasted intact animals.Hence inadequate nutrition was probably not themajor reason for the absence of radioactivityin the DNAof fat cells derived from hypophysec-tomized rats, and it would appear more reasonableto attribute this finding to an effect of hypophysec-tomy on fat cell formation and (or) maturation.

While a specific action of growth hormone on

stromal proliferation was the most obvious expla-nation of the different'effects of the hormone on

fat cell and stromal DNAradioactivity, it was evi-dent that pituitary ablation may have abolished a

fat cell response to growth hormone that wouldhave been demonstrable in the presence of normalpituitary function. To explore this possibility a

second series of experiments was performed in fedintact rats. In this instance once again growthhormone clearly stimulated thymidine'incorpora-tion into stromal but not fat cell DNA. Even thisexperiment was not considered as providing de-finitive evidence that the hormone only affectsstromal DNA synthesis, for it' was possible thatDNAformation in primordial fat cells was pro-

gressing maximally in untreated animals. As ithas been shown that prior administration of

DNA Synthesis in Adipose Tissue 2495

GROlhTHHORMOIE

CONTRO

CFt1 T

Ir

r---- _j

growth hormone to fasted rats enhances the subse-quent in vitro incorporation of thymidine intoDNA of adipose slices (15), a third series ofintact animals was given the hormone during a48 hr fasting period. In this instance growth hor-mone produced a small but significant increase inDNA radioactivity extracted from fat cells re-moved 12-16 days after isotope injection; how-ever, this increment was only one-third that notedin the stromal fraction. Thus it is evident thatgrowth hormone stimulated DNA synthesis inboth stromal cells and fat cell precursors, althoughthe major effect of the hormone was undoubtedlyon the stromal pool. It was also clear that removalof the pituitary resulted in an inhibition of fat cellformation and (or) maturation that was not re-versed by a brief period of growth hormone treat-ment.

DISCUSSION

The results of these experiments demonstrate thefeasibility of studying DNAsynthesis in both thefat cell and stromal elements of adipose tissueunder in vivo conditions. To approach this pro-blem it was found necessary to purify the fat cellfraction by methods somewhat more elaboratethan the multiple washes usually employed. Inexperiments done before the introduction of theseadditional procedures, when adipose tissue wasremoved within 12 hr of thymidine injection, aconsiderable fraction of the radioactive DNA ofthe tissue was recovered in the fat cell fraction,a finding also reported by others (16). The knownlack of mitotic activity displayed by lipid-contain-ing adipose cells casts serious doubt on the validityof these observations and suggested the possibilityof residual stromal contamination despite repeatedfat cell washes. This apparent paradox was easilyresolved when the fat cell pool was purified morethoroughly.

In this study thymidine incorporation intoDNAhas been used as an index of synthesis ofnew DNA; it is recognized, however, that it couldalso have been a function, at least in part, of DNArepair. However, if the latter process had signifi-cantly influenced the results, it is likely that allcells would have been labeled initially, not justthe stromal fraction. It was also assumed that theextent of thymidine incorporation was a measureof the rate of new cell formation, and while no

definite proof of this assumption is available, itwould appear to be a reasonable inference sincepolyploidy is not a feature of the cellular constitu-ents of adipose tissue (3). A further assumptioninherent in the study was that radioactivity recov-ered in DNA had been introduced solely duringthe process of thymidine incorporation. Mentionhas already been made of one finding that supportsthis assumption; in fed and fasted animals fol-lowed for 2 wk after injection of the isotope, nonet accretion of DNA label was evident despitepersistent circulation throughout the period of atleast one radioactive thymidine metabolite, tri-tiated water. Finally it must be appreciated thatall studies of DNAsynthesis which involve incor-poration of a labeled precursor into DNA arebased on the assumption that changes in radio-activity in product reflect alterations in rates ofsynthesis and not in specific activity of intracellu-lar precursors. In the present study the failure ofchanges in medium thymidine concentration to,abolish differences in DNA radioactivity betweenfed and fasted rats and the increase produced bygrowth hormone in the chemical DNAcontent ofstroma suggest that, at least in these instances, theisotopic data did in fact reflect changes in ratesof DNA formation. A more direct approach tothis problem could not be made in this study forthe first requirement, separation of primordial fatcells from the rest of the stromal fraction, couldnot be accomplished. Furthermore it was evidentthat since collagenase digestion altered thymidineflux across subsequently isolated stromal cells,accurate measurement of nucleotide pools in thesecells was impossible.

The picture of the kinetics of the cellular ele-ments of adipose tissue that emerges from theseexperiments contains several features of interestand a number that are unexplained. It is evidentthat of the DNA synthesized in adipose tissueduring the first 12 hr after thymidine injection,one-half was lost from the tissue during the next48 hr by unknown mechanisms and that only 10%was ultimately recovered in cells containing fat.Maturation of fat cells occurred rapidly, the inter-val between completion of DNA synthesis andsignificant lipid accumulation being between 2 and5 days, and it is clear that the processes of differ-entiation and maturation of primordial fat cellswere not hastened by refeeding after a period of

2496 C. H. HoUenberg and A. Vost

fasting. As in other tissues (17, 18) fasting de-pressed DNA synthesis in fat cells and stromalelements, and it is of some interest that this effectwas more profound in the fat cell fraction.

The demonstration that mature fat cells werenot destroyed by an acute fast of duration suffi-cient to reduce lipid stores by 50%o is consonantwith the observation that in obese man, fastingdoes not reduce the number of existing fat cellsbut merely empties them (4). It is also clear thatfat cells depleted of lipid by fasting retained thecapacity for reaccumulation of triglyceride whenrefeeding was begun, and that the rate of refillingof existing cells was more rapid than the rate ofmaturation of new fat cells. Because of this fact,lipid repletion during at least the early stages ofrefeeding was due entirely to refilling of existingfat cells and was not significantly influenced bynew fat cell formation. Thus while it is very likelythat protracted under- and over-nutrition will in-fluence adipose stores by affecting both the pro-duction of fat cells and the lipid content per cell, itis probable that abrupt short-term changes in adi-pose mass predominantly reflect changes in cellu-lar lipid content. Hausberger, using principallyhistological techniques, has previously reached asimilar conclusion (5).

The data derived from the experiments withhypophysectomized and growth hormone-treatedanimals can be considered as providing only intro-ductory information on pituitary control of fat cellformation and maturation. It is evident that pitui-tary ablation almost completely abolished forma-tion and (or) maturation of fat cells, and that thiseffect was not influenced by brief treatment withgrowth hormone. The specific consequences ofhypophysectomy that produce this phenomenonremain to be defined, and it is not clear as towhether removal of the pituitary resulted in areduction in the rate of DNAsynthesis in primor-dial fat cells or in arrest of cellular differentiation.

There is less ambiguity in the interpretation ofthe growth hormone studies; in all systems used,administration of this hormone produced a markedincrease in DNAsynthesis in the stromal fractionand in intact animals caused a lesser augmentationof DNA formation in cells destined to becomeadipocytes. The abolishment by growth hormoneof the inhibitory effect of fasting on DNA syn-thesis, a phenomenon first described by Mura-

kawa and Raben (15), is particularly intriguingand remains completely unexplained. It would beof obvious interest to determine whether proteinand RNAsynthesis are involved in this hormoneeffect.

The differences between fat cells and stromalelements in responsiveness to growth hormonewas one of several observations that suggested thatDNA synthesis in the two cellular pools wascontrolled independently, at least to some extent.In fed animals, the ultimate recovery of radio-active DNA in fat cells was unrelated to radio-activity present in stromal DNA12 hr after injec-tion of isotope. Furthermore nutritional variationsappeared to influence more significantly incorpo-ration of thymidine into primordial fat cell ratherthan stromal DNA. In relation to the latter phe-nomenon it is pertinent to note that in those formsof human and experimental obesity associatedwith hyperinsulinemia, increases in fat cell num-ber have been described (4, 5). Obviously consid-eration must be given to the possibility that altera-tions in nutrition produced changes in fat cellDNA synthesis due to fluctuations in insulinaction on primitive fat cells. Indeed it would notbe surprising if such a control system was foundto exist, for it is well established that insulin invitro can enhance DNA synthesis in mammarygland explants (19).

ACKNOWLEDGMENTSThe authors would like to thank Dr. R. Patten for hishelp at several stages of this study. Miss M. Faluhelyiand Mrs. R. Osiek rendered valuable technical assistances.

This work was supported by grants from the MedicalResearch Council of Canada and from the Quebec HeartFoundation.

REFERENCES

1. Rodbell, M. 1964. Localization of lipoprotein lipasein fat cells of rat adipose tissue. J. Biol. Chem. 239:753.

2. Hellman, B., and C. Hellerstrbm. 1961. Cell renewalin the white and brown fat tissue of the rat. ActaPathol. Microbiol. Scand. 51: 347.

3. Peckham, S. C., C. Entenman, and H. W. Carroll.1962. The influence of a hypercaloric diet on grossbody and adipose tissue composition in the rat. J.Nutr. 77: 187.

4. Hirsch, J., J. L. Knittle, and L. B. Salans. 1966. Celllipid content and cell number in obese and nonobesehuman adipose tissue. J. Clin. Invest. 45: 1023.(Abstr.)

DNA Synthesis in Adipose Tissue 2497

5. Hausberger, F. X. 1965. Effect of dietary and endo-crine factors on adipose tissue growth. In Handbookof Physiology. A. E. Renold and G. F. Cahill, Jr.,editors. The Williams & Wilkins Company, Balti-more. 1st edition. 519.

6. Rodbell, M. 1964. Metabolism of isolated fat cells.I. Effects of hormones on glucose metabolism andlipolysis. J. Biol. Chem. 239: 375.

7. Burton, K. 1956. A study of the conditions andmechanism of the diphenylamine reaction for thecolorimetric estimation of deoxyribonucleic acid.Biochem. J. 62: 315.

8. Hollenberg, C. H. 1966. The origin and glyceridedistribution of fatty acids in rat adipose tissue. J.Clin. Invest. 45: 205.

9. Vaughan, B. E., and E. A. Boling. 1961. Rapid assayprocedures for tritium-labeled water in body fluids.J. Lab. Clin. Med. 57: 159.

10. Rubini, J. R., E. P. Cronkite, V. P. Bond, and T. M.Fliedner. 1960. The metabolism and fate of tritiatedthymidine in man. J. Clin. Invest. 39: 909.

11. Simon, G. 1965. Histogenesis. In Handbook of Physi-ology. A. E. Renold and G. F. Cahill, Jr., editors.The Williams & Wilkins Company, Baltimore. 1stedition. 101.

12. Goldrick, R. B. 1967. Morphological changes in the

adipocyte during fat deposition and mobilization. Am.J. Physiol. 212: 777.

13. Raben, M. S., and C. H. Hollenberg. 1959. Effect ofgrowth hormone on plasma fatty acids. J. Clin. Invest.38: 484.

14. Goodman, H. M. 1965. Early and late effects ofgrowth hormone on the metabolism of glucose inadipose tissue. Endocrinology. 76: 1134.

15. Murakawa, S., and M. S. Raben. 1968. Effect ofgrowth hormone and placental lactogen on DNAsynthesis in rat costal cartilage and adipose tissue.Endocrinology. In press.

16. Kazdovai, L., T. Braun, and P. Fabry. 1967. In-creased DNAsynthesis in epididymal adipose tissue ofrats ref ed after a single fast. Metab. Clin. Exptl. 16:1174.

17. Leduc, E. H. 1949. Mitotic activity in the liver ofthe mouse during inanition followed by ref eeding withdifferent levels of protein. Am. J. Anat. 84: 397.

18. Lahtiharju, A. 1966. Influence of glucocorticoid, min-eralocorticoid and starvation on DNA synthesis ofepidermal and gastric cells in the mouse. Growth.30: 449.

19. Turkington, R. W. 1968. Hormone-induced synthesisof DNAby mammary gland in vitro. Endocrinology.82: 540.

2498 C. H. Hollenberg and A. Vost