Journal of Clinical InvestigationVol. 43, No. 5, 1964

Studies on the Intestinal Absorption and IntramucosalLipolysis of a Medium Chain Triglyceride *

MARCR. PLAYOUSTt AND KURT J. ISSELBACHER(From the Department of Medicine, Harvard Medical School, and the Medical Services

[Gastrointestinal Unit], Massachusetts General Hospital, Boston, Mass.)

After absorption from the intestinal tract, fattyacid molecules with chain lengths of ten or lesscarbon atoms are known to be transported un-esterified in the portal blood stream rather thanin the lymph in the form of triglycerides (1-3).It has been reported that in patients with celiacdisease and cystic fibrosis of the pancreas the ef-ficiency of lipid adsorption appears to be relatedto the chain length of the dietary fatty acids (4,5). In these subjects triglycerides of octanoicand decanoic acids (medium chain triglycerides)seem to be more completely absorbed than thosecontaining fatty acids of higher molecular weight.

Differences in the behavior of fatty acids dur-ing absorption are probably due partly to theirphysical properties and partly to variations intheir intramucosal metabolism. The transport ofabsorbed fatty acids of low molecular weight bythe portal route is likely related closely to theirpoor incorporation into mucosal triglycerides (6).This situation may be caused by the low activityof the esterifying enzymes toward these sub-strates (7), the presence of an active lipase, ormore likely, a combination of both factors.

Rat intestinal mucosa was examined, there-fore, for the presence of an enzyme system effec-tive in the hydrolysis of medium chain triglycerides.For these studies trioctanoin was selected as a sub-strate representative of this group of compounds.Active lipolysis of trioctanoin by mucosal homoge-nates was observed, and the intracellular localiza-tion of this lipolytic activity was studied. It waspossible to demonstrate a number of properties

* Submitted for publication October 9, 1963; acceptedJanuary 2, 1964.

Supported in part by U. S. Public Health Servicegrant AM-03014 from the National Institutes of Health.

tWork done during tenure of a fellowship from thePostgraduate Medical Foundation, University of Sydney.Present address: Department of Medicine, University ofSydney, Australia.

serving to distinguish this enzyme system frompancreatic lipase. In the rat, evidence was ob-tained in vivo, suggesting that unhydrolyzed trioc-tanoin may enter mucosal cells directly and thenbe subject to lipolysis by the enzyme systemdescribed.

Materials and Methods

Tripalmitin-carboxyl-C4 and trioctanoin-carboxyl-C14 1were purified by ascending chromatography on platescoated with thin layers of silicic acid; 2 the developingsolvent system contained by volume 40%o ether, 58%ohexane, and 2%o glacial acetic acid. Sodium octanoate-1-C I was converted to the free acid and also purifiedby thin-layer chromatography.

A mixture of medium chain triglycerides containingmore than 90% trioctanoin was a gift; 3 tripalmitin wasobtained commercially.4 Sodium taurocho'ate was syn-thesized from recrystallized cholic acid 5 by the methodof Norman (8).

Experiments with homogenates of intestinal mucosalcells. Female albino (Sprague-Dawley) rats 6 weighingapproximately 200 g were fasted overnight before usein these experiments. Each animal was killed by a blowon the head, and the small intestine was rinsed in situwith ice-cold 0.84%o NaCl, excised, and chilled to 40 Cin 0.278 M mannitol in 0.01 M sodium phosphate buffer,pH 7.4. The jejunal portion of the intestine was washedwith fresh mannitol solution and the mucosa expressedon a glass plate. Fifteen ml of mannitol solution wasadded for each gram of expressed mucosa cells and thehomogenate prepared as described previously (9). Nu-clei and cellular debris were removed by centrifugationat 1,400 X g for 10 minutes; mitochondria sedimented at5,900 X g for 15 minutes and microsomes at 105,000 X gfor 60 minutes. In certain experiments, indicated below,the microsomes were washed in mannitol solution and

1 New England Nuclear Corp., Boston, Mass.2 Silica gel C (Merck). Brinkman Instruments, Great

Neck, N. Y.3 Of V. Babayan, Drew Chemical Co., Boonton, N. J.4 Hormel Foundation, Austin, Minn.5 Special enzyme grade. Mann Research Laboratories,

New York, N. Y.6 Charles River Laboratories, Boston, Mass.

878

LIPOLYSIS OF A MEDIUMCHAIN TRIGLYCERIDE

sedimented again by centrifugation. Brush border frac-tions were prepared by the method described by Millerand Crane (10). Tissue fractions were resuspended in0.154 M KCl in 0.01 M sodium phosphate buffer, pH 7.4,and their protein contents measured by the biuret reaction.

Triglyceride substrates were dissolved in a smallamount of ether and homogenized in 10%o crystalline bo-vine albumin; 7 the ether was then evaporated at 35 to40° C. Crude pancreatic lipase8 was dissolved in 0.154M KCl containing 0.01 M sodium phosphate, pH 7.4.

Incubations were carried out in stoppered test tubesat 370 C for 10 minutes. The usual incubation mixtureconsisted of 1 ml of lipid-albumin suspension, 0.3 ml offivefold concentrated, low calcium, Krebs-Ringer phos-phate buffer, pH 7.4 (11), 20 ,umoles of sodium tauro-cholate, and 0.5 ml of tissue suspension or lipase solu-tion. The final volume was 2 ml. Incubations wereterminated by placing the test tubes in an alcohol-icebath at - 15° C for 1 minute. After this, cold chloro-form-methanol (2: 1) was added. The lipids were thenextracted by the method of Folch, Lees, and SloaneStanley (12).

In experiments with tripalmitin as the radioactive sub-strate, the solvents containing the extracted lipids wereevaporated at 400 C under a stream of nitrogen and theresidues redissolved in chloroform. Samples were as-sayed for total radioactivity in a liquid scintillationspectrometer and the lipids fractionated by thin-layerchromatography as previously described (13).

In the experiments with trioctanoin as the substrate,the chloroform-methanol added at the end of the incuba-tion contained 6 ,tmoles of carrier trioctanoin and 42,gmoles of carrier octanoic acid, and the pH during theFolch extraction was adjusted to be below 3 (the pK ofoctanoic acid is 4.9). The volatility of octanoic acidmade necesary a modification of the usual fractionationprocedure. The fatty acids present in the lower (chloro-form) phase were extracted into dilute aqueous NaOH.To prevent the formation of emulsions the NaOH wassaturated with NaCl. The aqueous fraction was acidified,and the lipids were re-extracted into hexane. A suitablefraction of the hexane fraction was added directly intoa counting bottle containing scintillation fluid (13) andits radioactivity determined. The neutral lipid fractionwas analyzed further by thin-layer chromatography. Re-covery of C14-octanoic acid after its addition to incubationmixtures containing boiled microsomes was 92 to 96%.Neither evaporation of the hexane fractions derived fromnormal incubations nor subsequent thin-layer chromatog-raphy of the residues demonstrated significant contami-nation by monoglycerides or other higher glycerides.

Experiments on lipid absorption in vivo. These stud-ies employed female rats weighing about 200 g. Bileand pancreatic juice were excluded from the intestinallumen by double ligation of the common bile duct at thesite of its entrance into the duodenum. Twenty-fourhours later, under ether anesthesia, a 20- to 30-cm seg-

7Nutritional Biochemical Corp., Cleveland, Ohio.8 Sigma Chemical Company, St. Louis, Mo.

ment of lower duodenum and proximal jejunum was iso-lated between silk sutures, with care taken not to ob-struct blood flow. The segment was washed throughwith 50 ml of 0.84% NaCl, and the radioactive lipid,suspended in 10% albumin, was introduced in a volumeof about 1 ml. After 10 minutes, a specimen of bloodwas withdrawn from the portal vein, the isolated seg-ment of gut removed intact, and its contents washed outwith 50 ml of NaCl solution. After rapid chilling ofthe empty sac, a mucosal homogenate was prepared asdescribed above using 0.154 M KCl. The homogenatewas then centrifuged at 900 X g for 10 minutes to sedi-ment remaining intact cells, cell debris, and nuclei.Portions of the sediment, supernatant fraction, portalvein plasma, and contents of sac lumen were extractedwith chloroform-methanol and the lipids fractionated inthe manner described above. The amount of radioactivityabsorbed was calculated by subtracting the amount re-covered in the intestinal lumen from the administereddose. Radioactive lipids in the 900 X g sediment of themucosal homogenate were considered to represent pri-marily adsorption on cell walls, whereas those recoveredfrom the supernatant (cytoplasmic) fraction were re-garded as intracellular (14).

Results

Hydrolysis of trioctanoin by subcellular frac-tions of intestinal mucosa. Table I shows thatC14-trioctanoin is readily hydrolyzed upon incuba-tion with washed microsomes from intestinal epi-thelial cells. In contrast, the incubation of equi-molar amounts of C14-tripalmitin with intestinalmicrosomes under the same conditions resulted invery little hydrolysis.

That the substrate specificity of the microsomallipase differed from that of pancreatic lipase isapparent from the results of incubations of tri-octanoin and tripalmitin in the presence of pan-creatic lipase. Whereas the microsomal enzymewas predominantly active in the hydrolysis oftrioctanoin, pancreatic enzyme attacked both sub-strates to an approximately equal extent (TableI).

Relatively little mono- or diglyceride was iso-lated after the microsomal hydrolysis of trioc-tanoin. This was probably due to the action ofmonoglyceride lipase, which has previously beendemonstrated in intestinal cell microsomes (13).Monoglyceride lipase has a maximal activity onmonoglycerides containing fatty acids with eightto twelve carbon atoms (13).

Results of studies on the intracellular localiza-tion of lipolytic activity for trioctanoin are shown

879

MARCR. PLAYOUSTAND KURT J. ISSELBACHER

TABLE I

Hydrolysis of trioctanoin and of tripalmitin by washed intestinal microsomes and by pancreatic lipase*

Distribution of radioactivelabel after incubation

Source of Fatty acid Triglyc- Diglyc- Monoglyc- Fattyenzyme Substrate liberated eride eride eride acid

mjsmoles/J10 %, % %O

mznulesTrioctanoin 155 52 3 6 39

Intestinalcellmicrosomes

Tripalmitin 20 93 1 1 5

Trioctanoin 194 41 6 7 46Pancreatic

lipaseTripalmitin 176 39 3 14 44

* Each incubation mixture contained 140 mjsmoles of either C'4-trioctanoin or C14-tripalmitin, 100 mg crystallinebovine albumin, 180 Mmoles NaCl, 84 Emoles KCl, 30 Mmoles sodium phosphate, 2 Emoles CaC12, 2 Mmoles MgSO4, and 20ismoles sodium taurocholate in a total volume of 2 ml and at pH 7.4. The reaction was started by the addition either ofwashed intestinal microsomes (1.6 mg protein) or of pancreatic lipase (0.3 mg protein). The incubations were carriedout at 37° C for 10 minutes and were terminated by placing the tubes in an alcohol-ice bath at -15° C for 1 minute, fol-lowed by the addition of 2:1 chloroform-methanol as described in Methods. The data shown represent the mean valuesof three sets of experiments carried out in duplicate.

in Table II. The highest specific activity waspresent in the microsomal fraction, although thetotal activity in the soluble fraction was appre-ciable. At least part of the activity in the solublefraction might have been due to contaminationby pancreatic lipase, a suggestion supported bythe significant hydrolysis that occurred when C14-tripalmitin was incubated with either the wholehomogenate or the soluble fraction. On the other

TABLE II

Hydrolysis of trioctanoin by homogenates of rat intestinalmucosa: intracellular localization of enzyme activity*

Protein TotalSpecific in activity in

Fraction activity fractionj fractiont

mEmolesfatty acidliberated/min/mgprotein

Whole homogenate 3.1 100 100Nuclei (unwashed) 1.3 38 13Mitochondria 3.5 8 9Microsomes 7.0 17 38Soluble fraction 4.1 32 43Sumof fractions 95 103Isolated brush borders 0.9 5.6 2.1

(prepared separately)

* Each incubation was carried out for 10 minutes at 370 C with 156mpmoles C'4-trioctanoin and 0.8 to 1.4 mg tissue protein. See Table Ifor the other components of the incubation mixtures and the methodof terminating the reactions.

t Figures refer to relative amounts of protein or total enzyme activityin fraction as compared to whole homogenate. All analyses werecarried out in duplicate and are representative of three sets of experi-mnents,

hand, as indicated in Table I, hydrolysis of tri-palmitin by the washed microsomes was minimal.

Lipolysis of trioctanoin by isolated brush bor-ders was slight (Table II). Furthermore, thetotal activity of the homogenate was not appre-ciably reduced by the removal of the brush bor-der fraction.

Some properties of the microsomal trioctanoinlipase. Although extensive purification and iso-lation of the trioctanoin lipase from cell fractionswas not carried out, characterization of some of itsproperties was attempted by studies with washedmicrosomes. It was of particular interest to com-pare these properties with those of pancreaticlipase.

A time study showed that the rate of hydroly-sis of trioctanoin by intestinal microsomes waslinear for 10 minutes. Additional experimentsshowed that the quantity of fatty acid liberatedwas roughly proportional to the amount of micro-somal protein present in the incubation mixture.

Microsomal suspensions were kept for severaldays at 0 to 4° C without significant loss ofactivity. Freezing and thawing caused only asmall loss of activity, but there was marked in-hibition of activity when the microsomal prepara-tion was immersed for 5 minutes in a water bathat 50° C (Table III). This contrasts with pan-creatic lipase, which was stable when subjected

880

LIPOLYSIS OF A MEDIUMCHAIN TRIGLYCERIDE

TABLE III

Effect of heating on lipolytic activity of intestinal microsomesversus pancreatic lipase*

Hydrolysis of tri-octanoin by

Durationof heating Intestinal Pancreaticto 500 C microsomes lipase

minutes %of activity remainingafter heating

0 100 1002 57 1005 25 98

10 22 84

* Incubations were carried out for 10 minutes at 370 Cwith 176 m~smoles C14-trioctanoin and either intestinal cellmicrosomes (0.6 mg protein) or pancreatic lipase (0.2 mgprotein). The microsomes and the pancreatic lipase hadbeen immersed in a 500 Cwater bath for the time indicated.See Table I for other constituents of the incubation mix-tures and the method of terminating the reaction. Thedata are representative of three separate experiments car-ried out in duplicate.

to this temperature for 5 minutes. Also, the in-testinal cell lipase was inhibited by fluoride ionsand by p-chloromercuribenzoate (PCMB) to agreater extent than pancreatic lipase (Table IV).

A pH activity curve demonstrated that lipolysisof trioctanoin by intestinal microsomes was ac-tive over a wide pH range, the maximal hydroly-sis occurring between pH 5.8 and 6.3 (Figure1). With the same substrate and the same ex-perimental conditions, the pH activity curve forpancreatic lipase showed a very broad pH opti-mumbetween 6.3 and 8.2 (Figure 1).

Absorption of C"4-trioctanoin by the rat, inztivo. When C14-trioctanoin was introduced into

TABLE IV

Effect of potassium fluoride and p-chloromercuribenzoate(PCMB) on trioctanoin hydrolysis*

Hydrolysis HydrolysisConcen- by intes- by pan-

tration of tinal creaticAddition inhibitor microsomes lipase

moles %of %ofcontrol control

None 0 100 100KF 0.02 12 87KF 0.2 2 35PCMB 10-4 29 100

* Incubations were carried out for 10 minutes at 370 Cwith 170 msmoles C'4-trioctanoin and either intestinalmicrosomes (0.6 mg protein) or pancreatic lipase (0.2 mgprotein). KF or PCMBwas present in the concentrationsindicated. See Table I for the other constituents of theincubation mixtures and the method of terminating thereaction. The data shown are the mean values of threeseparate sets of experiments.

250 - PANCREATIC LIPASE

i, 2500ok ,

k.~ ~~~~~~A

150-

k50 -MICROSOMAL

4 5 6 7 8 9 10

p H

FIG. 1. EFFECT OF PH ON HYDROLYSIS OF TRIOCTANOINBY MICROSOMALVERSUS PANCREATIC LIPASE. Each in-cubation mixture contained 160 m/umoles C"4-trioctanoin,1.1 mg microsomal protein or 0.12 mg pancreatic lipase,10 mg albumin, and one of the following buffers: 40Amoles succinate (pH 4.1 to 5.6), 25 /Amoles phosphate(pH 5.8 to 7.4), 60 /Amoles Tris (pH 7.7 to 8.8), and 40/Amoles carbonate-bicarbonate (pH 8.9 to 10.1). Othercomponents of the incubation mixtures and the conditionsof incubation were the same as described in Table I.

an isolated loop of small intestine of an anesthe-tized rat, 58% of the administered lipid wasabsorbed during a 10-minute period. A typicalresult of three separate experiments is shown inTable V. Analysis of the lipid in the lumen atthe end of the experiment showed that the radio-active label was still in triglyceride (> 99%o).This confirmed the satisfactory exclusion of lipaseactivity from the lumen and suggested that thenature of the lipid presented to the mucosal cellsfor absorption was predominantly triglyceride.

A homogenate was prepared from the mucosaof the isolated sac of small intestine, and aftersedimentation of cell walls, debris, and nuclei,the intracellular (cytoplasmic) lipids were ex-tracted and analyzed for radioactivity. In con-trast to the composition of the lipid in the lumen,10%o of the label in the intracellular lipids wasin free fatty acids and 90%o in the neutral lipid.The distribution of radioactivity was even morestriking in the lipids of the portal vein plasma,where 80%o of the label was in free fatty acids.

These data suggested that trioctanoin moleculesentered the mucosal cells without prior hydrolysisin the lumen. Furthermore, to explain the pres-ence of radioactive label in the fatty acid fractions

881

MARCR. PLAYOUSTAND KURT J. ISSELBACHtI

TABLE V

Absorption of trioctanoin, octanoic acid, and tripalmitin by rat intestine, in vivo*

Intracellular C14-lipidrecovered from mucosa C14-lipid in portal

Composition of intestinal sac§ vein plasmaluminal lipid

Substrate C'4-lipid at end of Fatty Neutral Fatty Neutraladministeredt absorbed in experiment$ acid lipid acid lipid

%of ad- cpm cpm/ml plasmaministered

doseC14-Trioctanoin, 58 >99% 500 4,000 600 100

100 jImoles, triglyceride5 X 106 cpm

C14-octanoic acid, 71 >99% 31,400 6,500 7,900 2002.5 jsmoles, fatty acid5 X 106 cpm

C14-tripalmitin, 6 >99% <50 390 0 015 jmoles, triglyceride3.5 X 106 cpm

* Bile and pancreatic enzymes were excluded from the intestine by bile duct ligation. Twenty-four hours later theladioactive lipid was introduced into a 20- to 30-cm segment of small intestine. After 10 minutes portal vein blood wassampled, the gut sac excised and emptied, and a mucosal cell homogenate prepared. For further details see Methods.

t Lipid substrate was suspended in 1 ml 10% crystalline bovine albumin.t C'4-lipid absorbed was calculated by subtracting the amount of radioactivity recovered in the sac lumen from the

administered dose.§ Intracellular C'4-lipid is the supernatant fraction remaining after sedimentation of intact cells, cell debris, and

nuclei at 900 X g for 10 minutes.

of the mucosal homogenate and the portal vein,hydrolysis of the trioctanoin evidently must haveoccurred within the mucosal cells.

It could be argued that these findings mightbe explained by the liberation of free fatty acidin the intestinal lumen, followed by its very rapidabsorption and subsequent re-esterification withinthe mucosal cells or removal as unesterified fattyacid in the portal blood stream. Were this thecorrect mechanism one would expect that, iflabeled octanoic acid were introduced into an iso-lated loop of small intestine, the distribution ofradioactive lipids in the mucosal homogenateshould be similar to that obtained with trioc-tanoin. A study with C14-octanoic acid ratherthan C14-trioctanoin was therefore performed. Atracer dose of C14-octanoic acid was used to ex-clude the possibility of overloading the mucosalesterification mechanism. As shown in Table V,the analysis of the lipid in the intestinal mucosarevealed 80% of the label still in the fatty acidfraction; this is in contrast to the results obtainedduring the absorption of C'4-trioctanoin, whenonly 10% of the radioactivity was found in thefatty acid and 90%o in the neutral lipid.

The absorption of trioctanoin was comparedwith the behavior of a long chain triglycerideunder the same in vivo conditions. C14-tripal-mitin was introduced into a segment of rat smallintestine. Only 6%o was absorbed in 10 minutes,whereas 58%o of a considerably larger quantity oftrioctanoin had been absorbed in the same timeinterval. Furthermore, in the tripalmitin ex-periment very few counts were found in theintramucosal lipids and none in the lipids of theportal vein plasma.

DiscussionIt seems well established that the chain length

of a fatty acid influences in some manner theroute of its absorption from the intestinal tract(i.e., portal versus lymphatic system) and alsowhether it leaves the mucosal cell in the form oftriglyceride or fatty acid. Bloom, Chaikoff, andReinhardt showed that when C14-palmitic acidwas absorbed from the intestine of rats, 92%could be recovered in the lymph, whereas the re-coveries in lymph of labeled myristic and decanoicacids were only 15 to 55% and 13%, respectively(1). In experiments in two healthy dogs (15)

882

LIPOLYSIS OF A MEDIUMCHAIN TRIGLYCERIDE

we have found that less than 3% of orally ad-ministered C14-trioctanoin appeared in the drain-age from surgical lymph fistulas; the small amountof label that was present in the lymph was inthe form of triglyceride. The chemical natureof the absorbed lipid transported in the portalvein has been studied by Borgstrom (3). He fedC14-decanoic acid to rats and demonstrated thatthe radioactive label in the portal plasma wasalmost completely in the unesterified fatty acidfraction. Studies in a patient with chyluria (16)and in a child with chylothorax (17) indicate thatthe route of absorption in man appears to be thesame as in animals.

Since in normal circumstances most dietarytriglyceride is digested to readily absorbable fattyacid and monoglyceride, the nature of the lipidmixture presented to the mucosal cell is probablythe same in all instances, regardless of the chainlengths of the fatty acids present. It follows thatthe subsequent partition between the medium andlong chain fatty acids must be due to differencesin their metabolism or behavior within the intes-tinal wall. Re-esterification of the long chainfatty acids appears to be a very active process (7),and the concentration of free fatty acids in themucosal cells is always low. The transport ofthe triglycerides out of the cells is presumably de-pendent on their close association with cell mem-brane structures (18) leading to the formation ofchylomicrons. Their appearance in the lymphand their complete absence from portal blood maybe due to the inability of triglycerides to diffuseacross the endothelial cells of the intestinal capil-laries.

The fate of the fatty acids of shorter chainlength, such as octanoic acid, might be expectedto be determined by the balance between esterifi-cation and lipolysis. If a significant concentrationof the octanoic acid is maintained, its physicalproperties (19) are such that it can readily diffuseout of the mucosal cells and through the endo-thelium into the blood stream. Since the rate ofportal flow is so much greater than lymph flow,even a limited release of unesterified fatty acidsfrom mucosal cells would favor the blood as themajor route of absorption. Possible factors tend-ing to keep the medium chain fatty acids in thefree, unesterified form are 1) a low activity ofthe esterifying mechanisms for these substrates

(7) and 2) the presence of intracellular lipolyticsystems. Evidence of the presence of intestinallipases has been presented by a number of in-vestigators. A monoglyceride lipase can be dem-onstrated in intestinal epithelial subcellular frag-ments (13, 20, 21), particularly in the micro-somes; this activity is not altered by washing theparticles to remove traces of pancreatic lipase(13). Since the absorption of intact monoglycer-ide molecules has been demonstrated in experi-ments with doubly-labeled lipid (22, 23), themucosal cells normally contain substrates uponwhich the monoglyceride lipase can act.

On the other hand, there is no good evidencethat the usual dietary triglycerides (containinglong chain fatty acids) can be absorbed unhydro-lyzed. The physiological role, therefore, of theintracellular lipases demonstrated by some in-vestigators to act on long chain triglycerides isunclear (24). In these previous studies contami-nation of tissue preparations by the highly activepancreatic lipase that bathes the mucosa was notalways excluded. Also, some lipase estimationswere performed at an alkaline pH using turbidi-metric methods and synthetic detergents. In thepresent lipase experiments we attempted to avoidthese possible sources of error. Lipolytic activitywas estimated directly by measuring the amountof fatty acid liberated from a radioactive triglycer-ide substrate. Incubations were performed closeto the physiological pH range and with tauro-cholate and albumin rather than with syntheticdetergents.

The hydrolysis of trioctanoin observed in thepresent experiments appears not to have been dueto contamination by pancreatic lipase for thefollowing reasons: 1) the highest specific activityof the lipase acting upon trioctanoin was in themicrosomal rather than the soluble fraction andwas not influenced by repeated washing; 2) thesubstrate specificity of this lipase was differentfrom that of pancreatic lipase; and 3) propertiesof the two enzyme systems, such as heat inactiva-tion and inhibition by fluoride ions and sulfhydrylagents, readily distinguished between them (TableIV).

Esterification of octanoic acid by cell-free in-testinal homogenates has previously been re-ported from this laboratory to be much less activethan esterification of palmitic acid (7). Appar-

883

MARCR. PLAYOUSTAND KURT J. ISSELBACHER

ently the small quantity of trioctanoin that maybe synthesized within the cell can be readily hy-drolyzed by the enzyme system we have demon-strated. The highest specific activity of the lipasethat hydrolyzes trioctanoin was in the microsomalfraction, which is the same cellular fraction in-volved in the majority of the reactions leading tointestinal triglyceride synthesis. Significantly,when C'4-octanoic acid is incubated with hamsterintestinal slices (25) and with rat everted gutsacs (26), there is very little incorporation ofthe label into glyceride esters. These findingsprobably reflect both the low rate of esterificationof the free acid and rapid lipolysis of any tri-octanoin synthesized. During the absorption oftrioctanoin in the intact animal, these same twomechanisms would favor the presence of unes-terified octanoic acid in the mucosal cells and itstransport in the portal blood as discussed above.

To date, no attempt has been made to deter-mine if the lipase that hydrolyzes trioctanoin isdistinct from other mucosal lipases or esterases(13, 24, 27). Trioctanoin was the only mediumchain triglyceride used in our experiments. Fur-ther studies are needed to determine if the micro-somal enzyme that splits trioctanoin is also activein the hydrolysis of other medium chain triglycer-ides as well as mono- and diglycerides. Thismight be anticipated because these substrates havesimilar physical and chemical properties (19).

The physical characteristics of the mediumchain triglycerides, particularly their solubilityand diffusibility when compared with triglyceridesof higher molecular weights, suggested the pos-sibility that transfer from intestinal lumen to mu-cosal cells might occur without prior lipolysis. Inthe anesthetized rat, trioctanoin was readily andrapidly absorbed by an isolated loop of small in-testine despite the apparent absence of pancreaticlipase. The experimental evidence favored theentrance of the unhydrolyzed triglyceride into themucosa and its subsequent intracellular hydrolysisand transport as unesterified fatty acid in theportal blood. It could be argued that the trioc-tanoin might have been hydrolyzed during itspassage through the cell membrane by an enzymelocated in the microvilli or brush borders of themucosa. This possibility seems unlikely in viewof the low lipolytic activity demonstrated in brushborder preparations in vitro. In our experiments,

furthermore, the trioctanoin was introduced as asuspension in albumin. Since bile salts had beensurgically excluded from the lumen and no mono-glyceride was added, the question of absorption ofthe trioctanoin from a micellar solution did notarise.

In normal circumstances the presence of pan-creatic lipase ensures the extensive lipolysis oftrioctanoin before its absorption, so that quanti-tatively the transport of the intact triglyceride isof limited significance. Patients with impairedfat absorption due to pancreatic or bile salt de-ficiency have, however, been reported to absorbmedium chain triglycerides more completely thanthe usual dietary fats (28, 29). A similar trendhas been noted in absorptive studies on childrenwith fibrocystic disease of the pancreas (4, 5).It is reasonable to speculate that, although pos-sibly unimportant in the normal subject, the di-rect adsorption of medium chain triglycerides andtheir subsequent intracellular hydrolysis may be-come an important consideration in the thera-peutic approach to patients with impaired intra-luminal digestion of fat.

Summary1. Extensive lipolysis of C14-trioctanoin was

observed when it was incubated with subcellularfractions of rat intestinal mucosa. The specificactivity of the enzyme system that splits trioc-tanoin was greatest in the microsomal fraction.Tripalmitin, a long chain triglyceride, was nothydrolyzed by this microsomal enzyme system.

2. The lipase that hydrolyzes trioctanoin wasfound to be distinct from pancreatic lipase. Thesubstrate specificities of the two enzyme systemsare dissimilar, and there are marked differencesin the effects of heating, fluoride ions, and p-chloromercuribenzoate.

3. In vivo experiments using isolated loops ofrat intestine indicate that C14-trioctanoin canenter the mucosal cells without prior hydrolysis inthe lumen. The demonstration of radioactivefatty acids in the intestinal mucosa and portalvein plasma suggests that the triglyceride washydrolyzed within the mucosal cells. Under thesame experimental conditions C14-tripalmitin wasonly poorly absorbed.

4. We postulate that a mucosal lipase such asthat which acts on trioctanoin may have a role

884

LIPOLYSIS OF A MEDIUMCHAIN TRIGLYCERIDE

in the hydrolysis of medium chain triglyceridesthat are either ingested or synthesized within theintestinal cell. Furthermore, such an enzyme sys-tem by keeping medium chain fatty acids withinthe mucosa in the free, unesterified form mayseem to favor their portal rather than lymphaticroute of transport.

AcknowledgmentsThe authors are indebted to Miss Dorothy Budz and

Mrs. Marilyn Kozacko for valuable technical assistance.

References1. Bloom, B., I. L. Chaikoff, and W. 0. Reinhardt.

Intestinal lymph as pathway for transport of ab-sorbed fatty acids of different chain lengths.Amer. J. Physiol. 1951, 166, 451.

2. Kiyasu, J. Y., B. Bloom, and I. L. Chaikoff. Theportal transport of absorbed fatty acids. J. biol.Chem. 1952, 199, 415.

3. Borgstr6m, B. Transport of '4C-decanoic acid inporta and inferior vena cava blood during absorp-tion in the rat. Acta physiol. scand. 1955, 34, 71.

4. Van de Kamer, J. H., and H. A. Weijers. Malab-sorption syndrome. Fed. Proc. 1961, 20 (suppl. 7),335.

5. Fernandes, J., J. H. van de Kamer, and H. A.Weijers. Differences in absorption of the variousfatty acids studied in children with steatorrhea.J. clin. Invest. 1962, 41, 488.

6. Borgstr6m, B. in Lipide Metabolism, K. Bloch, Ed.New York, John Wiley and Sons, 1960, p. 138.

7. Dawson, A. M., and K. J. Isselbacher. The esteri-fication of palmitate-1-C' by homogenates of in-testinal mucosa. J. clin. Invest. 1960, 39, 150.

8. Norman, A. Preparation of conjugated bile salts us-ing mixed carboxylic acid anhydrides. ArkivKemi 1955, 8, 331.

9. Senior, J. R., and K. J. Isselbacher. Direct esterifica-tion of monoglycerides with palmityl coenzyme Aby intestinal epithelial subcellular fractions. J.biol. Chem. 1962, 237, 1454.

10. Miller, D., and R. K. Crane. The digestive functionof the epithelium of the small intestine. II. Lo-calization of disaccharide hydrolysis in the isolatedbrush border portion of intestinal epithelial cells.Biochim. biophys. Acta (Amst.) 1961, 52, 293.

11. Cohen, P. P. Methods for preparation and study oftissues in Manometric Techniques, W. W. Umbreit,R. H. Burris, and J. F. Stauffer, Eds. Minneapolis,Burgess, 1959, p. 149.

12. Folch, J., M. Lees, and G. H. Sloane Stanley. Asimple method for the isolation and purification of

total lipides from animal tissues. J. biol. Chem.1957, 226, 497.

13. Senior, J. R., and K. J. Isselbacher. Demonstrationof an intestinal monoglyceride lipase: an enzymewith a possible role in the intracellular completionof fat digestion. J. clin. Invest. 1963, 42, 187.

14. Dawson, A. M., and K. J. Isselbacher. Studies onlipid metabolism in the small intestine with ob-servations on the role of bile salts. J. clin. In-vest. 1960, 39, 730.

15. Playoust, M. R., and K. J. Isselbacher. Unpublishedobservations.

16. Blomstrand, R., N. A. Thorn, and E. H. Ahrens, Jr.The absorption of fats, studied in a patient withchyluria. I. Clinical investigation. Amer. J. Med.1958, 24, 958.

17. Fernandes, J., J. H. van de Kamer, and H. A.Weijers. The absorption of fats studied in a childwith chylothorax. J. clin. Invest. 1955, 34, 1026.

18. Palay, S. L., and L. J. Karlin. An electron micro-scopic study of intestinal villus. II. The pathwayof fat absorption. J. biophys. biochem. Cytol.1959, 5, 373.

19. Deuel, H. J., Jr. The Lipids. New York, Inter-science Publishers, 1951, vol. 1, p. 57.

20. Tidwell, H. C., and J. M. Johnston. An in vitrostudy of glyceride absorption. Arch. Biochem.1960, 89, 79.

21. Pope, J. L., R. E. Askins, and J. C. McPherson.The specificity of an intestinal lipase for mono-glycerides. Fed. Proc. 1962, 21, 259.

22. Skipski, V. P., M. G. Morehouse, and H. J. Deuel,Jr. The absorption in the rat of a 1,3-dioleyl-2-deuteriostearyl glyceride-C1' and a 1-monodeuteri-ostearyl glyceride-C14. Arch. Biochem. 1959, 81, 93.

23. Johnston, J. M., and J. L. Brown. The intestinalutilization of doubly-labeled a-monopalmitin. Bio-chim. biophys. Acta (Amst.) 1962, 59, 500.

24. DiNella, R. R., H. C. Meng, and C. R. Park.Properties of intestinal lipase. J. biol. Chem.1960, 235, 3076.

25. Gelb, A. M., and J. I. Kessler. Effect of fatty acidstructure on esterification by the small intestinein vitro. Amer. J. Physiol. 1963, 204, 821.

26. Playoust, M. R., and K. J. Isselbacher. Unpublishedobservations.

27. Aldridge, W. N. Some esterases of the rat. Bio-chem. J. 1954, 57, 692.

28. Hashim, S. A., H. B. Roholt, and T. B. Van Itallie.Pancreatogenous steatorrhea treated with mediumchain triglyceride. Clin. Res. 1962, 10, 394.

29. Iber, F. L., E. Hardoon, and M. H. Sprague. Use ofeight and ten carbon fatty acids as neutral fat inthe management of steatorrhea (abstract). Clin.Res. 1963, 11, 185.

885