Application of Thin-layer Chromatography to theQuantitation of Plasma Neutral Lipids

and Free Fatty Acids

Robert T. Louis-Ferdinand, Donald G. Therriault,William F. Blatt, and Milton Mager

With the technical assistance of Edward J. Mat heson

Neutral lipids and free fatty acids were extracted from plasma and separated on thin-layer chromatoplatescoated with silica gel C. The plates were charred by heatingafter spraying with dichromate-sulfuric acid, and then evaluated densitometrically.Plasma analyses obtained by this procedure were compared with parallel determina-tions performed by established chemical technics.

THIN-LAYER CHROMATOGRAPHY (TLC) has been widely applied to the

qualitative analysis of complex biologic lipid mixtures (1, 2), as wellas to the quantitation of selected biochemical entities (3). Quantitationby TLC has been limited,for the most part,to the recovery of separated

components from the chromatoplate, followed by an independentanalysis using established analytic methods. Current technics forquantitative TLC of lipids were reviewed by Privett et a!. (4) in 1962.Recently, Blank et a!. (5) described a method employing densitometryof charred spots for direct quantitative analysis of lipid componentsseparated from a mixture by TLC. This method is more rapid andgenerally more sensitive than the recovery methods, and has the addedadvantage of eliminating errors associated with recovery of the samplefrom chromatoplates.

The purpose of this report is to present a comprehensive study ofthe application of TLC charring technic to the quantitative analysisof plasma neutral lipids and free fatty acids. The method is also com-

pared with conventional analytic procedures. Limitations encounteredusing the TLC charring technic are described in order that they maybe taken into account in adapting the method to the screening and

From the Biochemistry-Pharmacology Division, IT. S. Armny Research Institute of Environ.mental Medicine, Natick, Mass. 01760.

Received for publication Jan. 20, 1967; accepted for publication Mar. 18, 1967.

773

774 LOUIS-FERDINAND ET AL. Clinical Chemistry

routine deterlnination of lipid class composition of complex lipidsobtained from blood.

Materials and MethodsReagents and Apparatus

Highly purified reference lipids* consisting of cholestei’yloleate,

cholesterol,oleicacid,and trioleinwere used in the formulation of the

standards. The radiochemicals* consistedof randomly tagged sterol-’4C,

triolein-1-14C, cholesteryl oleate-14C, and oleic acid-1-14C. Radioactivitywas determined on a Ti’icarb scintillation spectrometer. Commercialchromatoplates coated with silicagel G in a thickness of 250 p,t wereused for most of tileanalyses. All solvents utilizedwere ACS gradeor better.

Extraction

Extraction and analysis of neutral lipids and free fatty acids from

plasma were accomplished by the procedure outlined in Fig. 1. Freshly

drawn plasma (5 ml.) was placed in a 60-ml. separatory funnel andtreated with 40 ml. of chloroform-methanol (2:1). The lower phase wasdecanted and subsequently pooled with the lower phases obtained fromfour additional 20-mi. extractions with tilesame solvent system. To

facilitatetileseparation of phases in the latter20-mi. extractions,2 ml.of normal salinewere added. in allseparations, slow inversion rather

than silaking was used for the mixing procedure. The pooled lowerphases were evaporated to dryness in a rotary evaporator at 10-15#{176}.

The residue was dissolved in chloroform with five separate 2-mi.

washes of anhydrous chloroform. To effect clarification,the extract

was filtered through glass wool, followed by a wash with chloroform

to remove any adsorbed lipid.The combined wash arid extract were

then reduced to 1 ml. under a stream of nitrogen.

Thin-layer Chromatography

Commercial prescored (20 X 20 cm.) silicagel G chromatoplatestwere separated into four 5 X 20 cm. plates, and developed for 2 hr.in a chloroform-methanol (2:1) solvent system to remove organic

contaminants. The plates were then activated by overnight heating at

110#{176}.Following this, tileplates were edged by removing the adsorbent

from tile periphery of the plate, and then divided into two equal chan-nels by a trough through the center.

Applied Science Laboratories, State College, Pa.fAnaltech, Time., Wilmington, Del.

Vol. 13, No. 9 1967 NEUTRAL LIPIDS AND FATTY ACIDS 775

PLASMA ( muh)

40 ml. C:M, 2:1

I UPPER PHASE

LOWER PHASE

I I 1. Extract with 20 ml. C:M, 2:1 4 X

2. 2 ml. 0.9% NaCI

LOWER PHASE UPPER PHASE

I (discard)

POOLED LOWER PHASES

1. To dryness

2. Extract with 10 ml. chloroform (amihvdm-omms)

3. Filter

RESIDUE FILTRATE

(discard) -_________

1. Coneemitrate to 1.0 ml.

2. Apply to silicagel-G thin-layerplate

1. For cholesterolesters,triglycerides, fatty acid esters-

THIN-LAYER Apply I lambda, develop in hexane :ethyl ether:acetic

CHROMATOGRAPHY acid 180:30:2

2. For cholesterol and fatty acids-Apply 4 lambda,

develop in pet. ether: ethyl ether: acetic acid 30:70:1

Char with dichromate-sulfuric acid; quantitate densitometrically

Fig. 1. Extraction amid thin-layer chmromatographic separation of plasma lipids.

Solvent systems used for developing the chromatoplates were pre-

pared daily. A Shandon (Colabs)* chromatography tank designed toaccommodate 20 plates (5 X 20 cm.) was used in these studies.Wicksconsisting of filterpaper strips were interspersed between the chro-matoplates to accelerateequilibrationof the chamber. The tank design

ensures that allplates are held at the same angle during migration.The following solvent systems were used: (1) hexane:ethyl ether:

acetic acid (180:30:2) for cholesterol esters, triglycerides, fatty aci(1esters; and (2) petroleum ether-ethyl ether-aceticacid (30:70 :1) for

cholesterol and fatty acids.

Colnbs, Chicago, Ill.

776 LOUIS-FERDINAND ET AL. Clinical Chemistry

Sample Application

Samples ranging in volume from 1 to 6 /Ll. were applied approxi-mately 2 cm. from the bottom of the chromatoplates. Calibrated i-pi.1)rummond microcap pipets* were found to be ideally suited for this

purpose, since samples could be precisely and rapidly applied withoutdamaging the gel surface.Lightly touching the gel surface was sufficient

to discharge the pipet previously filled by capillary action. The refer-ence lipids were dissolved in chloroform and then combined in 10-mi.ampules. These were evaporated to dryness, the vials sealed undernitrogen, then stored at -15#{176}until ready for use. Immediately prior

to use,the ampules were opened and 0.5ml. of chloroform added. Lipid

concentration was as follows: cholesteryl oleate 7.1 mg./ml.; triolein4.9 mg./ml.; oleicacid 2.2mg./ml.; and cholesterol4.1mg./ml. At least

three separate chromatograms containing the sample and standard were

prepared for each analysis.

Charring and SpotAnalysis

The spots were localized by spraying each air-dried 5 X 20-cm. platefrom a distance of 18 in. with 2 ml. of a 70% sulfuric acid reagentsaturated with potassium dichromate as described by Privett andBlank (6). The plates were charred by heating on a hot plate main-tained at 180#{176}for 50 mm. The charred areas were evaluated byscanning with a Photovolt densitometert equipped with an integratorand a TLC stage. The stage speed was set at 2.5 cm./min. and the chartspeed at 5.2 cm./min.

The following densitometer settings and modifications were made:

1. Stage light source intensity was set at maximum to providegreatest sensitivity.

2. Stage mask used with slit dimensions, 0.3 in. wide by 0.6 in. long.

3. Photocell mask (constructed)-first slit 0.008 in. wide by 0.6 in.long; second slit 0.035 in. wide by 0.6 in. long.

4. Photocell C was used without a filter at response setting No. 3.

5. Adjustments in base line were made with the full light control.

The calculation of absolute lipid content was made by comparing theintegrated area under the curve corresponding to the test spot withthe area of the accompanying reference standard. The instrumentalsettings used provided 20 sq. mm. of densitometer chart area perintegrator unit.

*Bonus Labs, Reading, Mass.

tPhotovolt Corp., New York, N. Y.

Vol. I3, N0. 9, 1967 NEUTRAL LIPIDS AND PATTY ACIDS 777

Chemical Assays

Parallel plasma analyses were conducted as follows: free and totalcholesterol were determined by the method of Leffler and McDonald(7), except that acetone was excluded from the digitonide precipitation

procedure. The procedure of Carison and Wolfstrom (8) was used fortriglyceride evaluation. To effect complete saponification, two dropsof 2.5% KOH were used and the saponification time increased to 60mm. Free fatty acids were assayed by the method of Dole (9).

Results and Discussion

Conditions for Densitometric Analysis

The application of the method is illustrated in Fig. 2 which shows atypical chromatogram obtained from pure lipid standards after spray-ing with sulfuric acid-potassium dichromnate. The extent of charringis determined by densitometry, using an integrator and recorder. Theintegrated area under the recorder tracing corresponding to theseparate spots on the chromatogram is represented by the slash marksat the lower portion of the fignre.

Blank et at. (5) pointed out that special precautions had to be taken

in the preparation of the silicic acid used as an adsorbent, and thatcommercial products invariably contain contaminant organic matterwhich makes them unsatisfactory. If was our experience as well, that

these impurities travel with the solvent, and result in a gradually risingbackground from the bottom of the chromatogram to the solvent front.Figure 3 illustrates the presence of organic impurities in silica gel,which will contribute to a high background. Silica gelG* was extractedwith chloroform-methanol (2:1) and an aliquot of extract correspondingto that obtained from one S X 20 cm. plate was applied at Channel B(Fig. 3). The standard lipid mixture was applied at Channels A and Cfor comparison. It is obvious that the presence of these contaminantswould make quantitation impractical. However, the special processingrecommended by Blank et a!. (5) would have excluded the advantage ofusing commercially available precoated plates. We found that de-veloping these plates for 2 hr. in a chloroform-methanol (2:1) solventsystem caused all organic impurities to move to the top of the chromato-plate. The washed plates could then be used for the separation anddensitometric analysis of neutral lipid with little or no backgroundcharring.

The relationship between tile charred spots and the densitometerresponse was obtained with purified reference lipids: cholesteryl oleate,

*Mallimickrodt, New York, N. Y.

778 LOUIS-FERDINAND ET AL. Clinical Chemstry

cholesterol, oleic acid, and triolein. Tue integrator reading versus the

amount of lipid applied to the chroniatogram is plotted in Fig. 4. Eachpoint represents au average of six separate determinations. in allcases, the integrator response is proportional to the weight of lipidapplied. in order to obtain this relationship, it was iiecessary to takeinto consideration the variation in degree of charring that occurs from

one chromatogram to the other. These variations were normalized bycomparing tile carbon yield froni a standard amount of lipid appliedalongside the sample to be analyzed on each plate. Analysis of the

Cho.E.

F.A.E.

1g.

F.A.

Cho.Origin

Fig. 2. Photodeii.sitomnetricpattern from eliromatogramn containing cholesteryl oleate (Ciw.

K.), metimyl oleate (F.A.E.), trioleimi (Ty.), oleie acid (J”..L) and cholesterol (C/mo.). J)evelop-

lag solvent, hexane : ethyl ether acetic acid (165:15:1).

180/30/2 v/v

.1A B Cl

I

o.

I#{149}Cho.E.

S.

Vol. 13, No. 9, 1967 NEUTRAL LIPIDS AND FATTY ACIDS 779

charred spots of a large number of these standard amounts of lipidpermitted the determination of the average integrator response for thatamount of lipid. These values were then used to correct for any varia-tion due to differences in carbon yield from one plate to the other.

Although Fig. 4 illustrates the direct proportionality between inte-

grator response and weight of lipid, it should be noted tilat the pro-

Hex./E.E./HAc I

Fig. 3. Tlmin-layer chrommmato-

grmmm after chmmmrriimgwith sul-

furic acid-diehromate solutiomi.A. and C. Standard lipid mmiix-

tore containiimg e Imole ste r ol

(C/mo.), oleic acid (F._4.), trio-

Ieiim( Ty.) amid cholesteryl oleate

(Cho. K.). B. Chloroformn : mnetli-

amiol (2:1) extract of silica gelG corresponding to time amount

on a 5 X 20 em. thin-layerplate, 250 / thick.

2 4 6 8 10 0 2 4 6 8 10

780 LOUIS.FERDINAND ET AL. Clinical Chemistry

portionality (slope of tile line) varies somewhat from one class oflipid to the other. This is not unexpected in light of the relationshipbetween the degree of charring and the distance the substance travelson the chromatogram. Figure 5 illustrates the integrator response as a

80

60

40

a)‘I,

C0a.sf10a)

100

080

I-

Dl60C

40

20

0

Wt.in micrograms

20

Fig. 4. Integrator respomise

veloped on chromatoplates.

front photodensitometer vs. weight of individual lipids de-

function of R for cholesterol. The Rr values were obtained by usingvarious ratios of diethyl ether, petroleum ether, and acetic acid in thedeveloping solvent system. The same relationship held true for oleicacid, triolein, and cholesterol oleate. Figure 6 illustrates the absolutedistance traveled on the chromatogram for various amounts of oleic

acid. These results are in excellent agreement with those reported byRouser et at. (10) who studied the charring phenomenon of phospho-lipids migrating to different positions on the chromatogram. Theyreported that the extent of charring is directly proportional to thedistance of migration of the particular lipid class. They attribute thisdirect relationship between the distance traveled and the increase incharring to the fact that spots spread during chromatography and thatthe spread increases with increase ill distance traveled.

It would seem at first sight that our results and those of Rouser et at.(10) are in direct contradiction to tile findings of Blank et at. (5) who

12 p Cholesterol

I . I . I m

Vol. 13, No. 9, 1967 NEUTRAL LIPIDS AND FATTY ACIDS 781

6 8 10Distance Migrated in cm.

reported constant peak areas for tripalmitin from lt values 0.4 to 0.8.They pointed out (5) that the increase in spot size resulting from in-creased distance traveled is compensated by a decrease in density ofthe spot. If one examines Fig. 6 cai’efully, he will note that the increased

charring with increased distance traveled is less, tile smaller tile amount

of lipid applied. This suggests that under certain circumstances, tile

decrease in density of the spot may well compensate for the larger spotresulting from increased distance traveled. It is obvious that under theconditions which Blank et at. used (5), the peak area was a function ofspot density as well as spot size. linder our conditions, however, themaui contributing factor seems to be spot size. This is further supported

by the results presented in Fig. 7. Carbon paper circular discs ofvarying diameter were placed between two glass plates and densitom-eter readings were obtained just as with the lipid chromatograms. Since

60U,

0 50a.U,

a,Fig. 5. Integrator response 40

from photodensitometer V5. Em

value of cholesterol obtained byvarying ratio of diethyl ether, ,

petroleum ether, and acetic acid . 20

in developing solvent.10

00! 02 03 04 05 06

Rf

80 A

a606. Integrator response V

tance traveled on chromatoplateof different weights of oleic o 40 #{163}

acid. Migration distance con-trolled by varying developing ‘- o

time.

20 _-‘

__- ‘2.2pg.

782 LOUIS-FERDINAND ET AL. Clinical Chemistry

carbon )apeI tends to block out the light, tile densitometer response

is a function of spot size only. Comparison of the integrator responseobtained from the carbon paper spots with that obtained from lipidspots of equivalent size showed reasonably close agreement. The

apparent nolulinearity of the relationship between spot area and

integrator response is probably the result of some light diffusing

through the carbon paper.

since the photodensitonueter respoiise is affected by tile distance thespot travels on tile chromatoplate, it was essential to select a solvent

system that would allow the respective lipids in plasma to move suf-

ficiently far to obtain maximal response. A solvent system consisting

of hexaiie:ethyl ether:acetic acid (180:30:2) was satisfactory for the(letermnination of cholesterol esters and triglyceride. However, fatty

acids and cholesterol did not travel sufficiently far in this system, anda second solvent system consisting of petroleum ether :ethyl ether :aceticacid (30 :70 :1) was used. Figure 8 illustrates the relative migration

distance of the respective lipids in each of tile solvent systems.

In the actual analysis of plasma lipid extracts, know’n amounts of

suitable standards were applied to tile same plate for comparison.

rfllese standards were dissolved ill the same solvent and applied to

the plate in the same volume as the unknown sample. These precautions

were taken to provide compensation for the variation in carbon yieldfrom one plate to the other.

Lipid Extraction

That the extraction p1ocedIlre described above was adequate for the

quantitative removal of all neutral lipid and free fatty acid from

120

60 - Fig. 7. Integrator respomise

0 from photodensitonteter vs. car-

:: . bon paper spot area.

- m I

20 40 60 80 100 I10 140

Spot Area mm2

P.E./E.E./HAc

j 30/70/1 v/v

Hex./E.E./HA1

180/30/2 v/vJIs..;Cho.E.

Vol. I3, No. 9, I67 NEUTRAL LIPIDS AND FATTY ACIDS 783

ChaE,Tg,FAE

LCho.

Origin

Fig. 8. Photogra p1msof timimi-layer chromatogramsa fter ehmarrimig with sulfuric acid-dichiromatesolution. A. and D. l)og plasma mieutral lipid extract. B. aud F. Standard lipid mmmixtimrecon-

taimiing ill order of mmseemidimigR values: cholesterol (C/mo.), oleic acid (F.A. ), trioleimm( Ty.),

metlmyl oleate (F.A.E.), cholesteryl oleate (C/mo. E.). C. and E. Stamidard mixture eommtaimtimmgimi

order of asceimdimig R values mono-, di-, amid trioleiim.

plasma was demonstrated liv the recovery of ad(ied ‘4(_’la1)(’ie(l lipids.Known amounts of ‘4C-labeled lil)i(ls were added to fresil plasmasamples. Following extraction, the ra(iiOaCtivitv in the final ChllOFO-

form concentrate was measured. A 14C-toiuene internal standard was

Esterified cholesterol*

TriglycerideFree fatty acids*

Cholesterol

Vol. 13, No. 9, 1967 NEUTRAL LIPIDS AND FATTY ACIDS 78

total obtained by TLC and that obtained gravimetrically after silicic

acid column chromatography. The rather large variability among thesamples measured gravimetrically suggested that there may have beensome contaminating phospholipid eluted along with the neutral lipids.Phosphorus determination of the chloroform eluates showed that thiswas the case. When corrections are made for the phospholipid presentin the eluate, based 011 the phosphorus determinations, agreement be-

tween the two methods is reasonably good.

Comparative Studies

Normal dog and human plasma extracts were used to compare TLC

with cited chemical methods. Assay results shown in Table 2 fall within

the ranges for normal values given by Ikels (ii) and Henry (12). Thetriglyceride and cholesterol values obtained by TLC agree well withthose obtained by tile chemical assays. A comparison of the methodson the basis of individual determinations is presented in the last columnwhere the mean differences and the standard deviation from tile mean

of the individual differences ai’e presented. In all cases, except foresterified cholesterol, the TLC values tend to be slightly higher thanthe chemical assay. However, with the exception of free fatty acid,these differences are not statistically significant. In light of the poorreproducibility of the TLC method for the determination of free fattyacids (Table 1), it is not surprising that a large discrepancy existsbehveen the methods. Free fatty acids are present in plasma in suchsmall quantities that the spot on the chromatogram does Ilot give a

Table 2. TLC ANALYSES VS. CHEMiCAL ANALYSES OF Doo ANt) HUMAN PLASMA

NEUTRAL LIPIDS AND FREE FATTY Acmns

No.

TLC

(31g./100 ml.± S.D.)

Chem. aoay(Mg./IOO ml. ± S.D.)

.1! eon differe,Iee

(Chem. aooay -

TLC ± S.D.)

DOG

Esterified cholesterol*

TriglycerideFree fatty acids*

Cholesterol

S

SS

8

80 ± 1Sf68 ± 14

9 ± 538 ± 9.0

105 ± 3250 ± 9

0.6 ± 0.6f37 ± 11

23 ± 20- 18 ± I1

-S ± S-1 ± 15

HUMAN

20 96 ± 22 139 ± 21 43 ± 2317 109 ± 52 lOS ± 28 -1 ± 36

20 15±6 7 ±4t -8±620 48 ± 8 42 ± 19 -6 ± 17

* By paired t test, values were significamitly differemit at p < 0.01.

t Obtained by using 250 average molecular weight per equivalent.Calculated as cholesterol which is esterifled omi the basis of cholesterol uleate.

786 LOUIS.FERDINAND ET AL. Clinical Chemistry

charring response adequate tor reliable qualititation. Nor is it practicalto illcrease the sample size to obtain a greater reading, for tile higherconcelltratioll of cholesterol in the sample would present difficulties ill

its determination. One solution which we have hot investigated is to

(letermnine free fatty acids separately using a larger saflhl)le.

It will he note(l that the esterified cholesterol values determined byare consistently lower than tilose obtained by chemical assay.

This (hoes hot lleCessarily ihidliCate that tile TLC values are in error.

It should be pointed out that the cholesterol method of Leffler andMcI)onald (7) involves tile determmatioii of total cholesterol without

prior saponification. The esterified cholesterol is then obtailled by

(lifferences l)etween free and total cholesterol. If tile total cholesterol

value by this method were high, it would he reflected in high esterifled

cholesterol values. it has been pointed out by Brown (13, 14) tilat tile

direct metho(1 of cholesterol determination overestimated the level of

cholesterol iii serum by approximately 36 ± 22 mg./l00 1111. it is in-terestuig that the TLC method differs from tile chemical assay by

approximately tile same amount.

It would seeni then that there is good agreenieiit between the ru L(’

(letermihlatiohls of esterified cholesterol amid the chemical assay pro-cedures involving a prior saponification step. From regression curves,the following equations were obtained to predict the chemical assay

values of esterified cholesterol from values obtained by TLCIIuniaii plasma

Chemical assay = 0.401 TLC + 100.8.])og plasma:

(‘henlical assay = 1.464 rri( - 12.2.

References

1. Mamigold, H. K., amid MaIms, 1). C., Fmaetionatioim 01’ fats, oils, amid waxes on tlmiim layersof silicic acid. J. A at. Oil Clmenti,sts’ Soc. 37, 383 (1960) -

2. Malimms, 1). C., aimd Mammgold, H. K., Ammalysis of complex lipid mmmixtures by tlmimi-layerehmromnimtogt-aphy amid comimpleniemitary mnethods. .1. A tim. Oil C/tern isis’ Soc. 37, 576 (1960).

3. Morris, C. J. 0. R., Thin-layer chromnatography of proteins ott Sephadcx G-lOO and 0.20(1.J. Cimroniafoq. 16, 167 (1964).

4. Privett, 0. S., Blank, M. L., Codding, I). \V ., ammilNickell, E. C., Lipid analysis by

tjuantitative tlmiim-layer chromatography. .1. A at. Oil Chemists’ Soc. 42, 381 (1965).5. Blank, M. L., Sehmit, .J .A., and Privett, 0. S., Quantitative ammalysis of lipids by thimm-

layer chmomimatographiy. J. Aiim. Oil Chemists’ Soc. 41, 371 (1964).6. Privett, 0. S., and Blamik, M. L., (‘liarrimig eommditiomms for the qimantitative analysis of

niommo-, di-, amid triglyeeri(les by thin-layer chmmonumtograplmy. J. A am. Oil Clmeinists’ Soc.39, 520 (1962).

7. Leffler, H. H., amid MeI)onal,1, A., Estimation of cholesterol in serumu. Ant.,!. Cliii. Pat/tot.39, 311 (1963).

8. (‘1 rlson, L. A., a ad Volfst roni, L. B., I )etermmmimmatiou of gly,.eri,les immIilood serum. Cli,,.C/tim. Ada 4, 197 (1959).

vol. 13, No. 9, I96 NEUTRAL LIPIDS ANb FATTY ACIDS 787

9. I)ole, V. P., A relation between non-esterified fatty acids in plasma amid lIme metabolismimof glucose. J. Clin. Invest. 35, 150 (1956).

10. Rouser, G., Kritchevsky, G., Gahli, C., amid Heller, D., I)etermnimmation of polar lipids:Quantitative column amid thin-layer chromatography. J. Ant. Oil Chemists’ Soc. 42, 215(1965).

II. Ikels, K. G., The Separatiomi of Serum Lipids into Four Major Coinponemits by SilicieAcid

Chromatography. School of Aerospace Medicine, Report 61-108, September 1961.12. Henry, R. J., Clinical Chemistry: Principles and Technics. Harper, New York, 1964, p. 862.13. Brown, W. P., The reliability of methods for cholesterol determinatiomi in studies on

coronary disease. Med. J. Asstr&ia 2, 467 (1961).14. Brown, W. D., Errors in time determination of serum cholesterol. Australia,, J. Exp. Biol.

Med. Sci. 39, 209 (1961).