Monounsaturated fatty acid oligomerization is responsible for the agglutination activity of heated virgin olive oil

Monounsaturated fatty acid oligomerization is responsible for theagglutination activity of heated virgin olive oil

Ioannis S. Patrikios*, Philippos C. Patsalis

The Cyprus Institute of Neurology and Genetics, Department of Cytogenetics, PO Box 23462, Nicosia, Cyprus

Received 9 December 2002; accepted 23 June 2003

Abstract

The present study focuses on the isolation, purification and characterization of a molecule formed when virgin olive oil is heated

at 130 �C for 24 h in air, that is found to be a strong agglutinin. The hemagglutinating activity of the novel molecule isolated fromthe heated olive oil was evaluated against human red blood cells (RBCs). Thin layer chromatography (TLC) of the oil mixtureshowed appearance of high molecular weight molecules, dimers and polymers. 13C nuclear magnetic resonance (NMR) and mass

chromatography–mass spectroscopy (GC–MS) were used for structure elucidation. A linear oligomerization of monounsaturatedfatty acids is involved. Light microscopy and electron microscopy were used to characterize and visualize the agglutination process.Agglutination without lysis or fusion was observed. The unheated olive oil and the isolated compound were also tested in-vitro

against normal and malignant colon and breast cells. The results showed the highest reduction of tumor cells with the isolated novelcompound. We conclude that virgin olive oil when heated in air produces oligomerization/polymerization of free unsaturated fattyacid possibly oleic acid (OA) that is a strong hemagglutinin against human RBCs with possible anti-cancer properties but with

unknown nutritional effects on human health.# 2003 Elsevier Ltd. All rights reserved.

Keywords: Agglutinins; Monounsaturated fatty acids; Virgin olive oil

1. Introduction

Processes in which polyunsaturated fatty acids arepolymerized by heating to a high temperature are dis-closed in the patent literature (Goebel, 1949, 1953).There are a variety of methods to dimerize unsaturatedfatty acids (mostly linoleic, but sometimes oleic) includ-ing thermal, clay-catalyze and peroxide-catalyzed meth-ods. Oligomers (C36, C44 and C54) are usedcommercially in many different applications, includingsolid and liquid polyamide resins, urethane resins, cor-rosion inhibitors, varnishes, soaps, polymer modifiers,oil additives, and epoxy curing agents. These oligomersare liquids probably because of the presence of so manystructural isomers (Leonard, 1979).One of the reactions (thermal or clay-catalyzed) com-

mon to all unsaturated, straight-chain aliphatic acidsand their alkyl esters is self-condensation to form high-molecular-weight dibasic and polybasic acids. One

molecule of an unsaturated fatty acid will react withanother to form a dicarboxylic acid with about doublethe original molecular weight. The resulting dimer is nota single entity, but is a collection of isomers groupedtogether under the designation of ‘‘dimer acids’’. Thesecommercial dimer acids are mostly mixtures of 36-car-bon dibasic acids, smaller amounts of 54-carbon tribasicacids, still higher-molecular-weight polybasic acids, andtrace levels of monomer (Leonard, 1979). The methylester of OA was thermally polymerized (in the absenceof clay) and the principal component of the dimer mix-ture was shown to be the straight chain (linear) dimericacid (McMahon & Crowell, 1974). The thermal poly-merization of oleic acid mostly yields the acyclic form(by a free-radical mechanism), but when linoleic acid isthe precursor the monocyclic, monocyclic structurespredominate (via a Diels–Alder reaction). One of themechanisms for thermal dimerization proposed byMeyers, Goebel, and Barrett (1960) involves a free-radical mechanism. A free-radical mechanism for ther-mal dimerization is very plausible when OA is thestarting material. Transfer of a hydrogen atom fromone oleate molecule to another can lead to the forma-tion of two organic radicals that can combine and form

0963-9969/$ - see front matter # 2003 Elsevier Ltd. All rights reserved.

doi:10.1016/S0963-9969(03)00119-4

Food Research International 36 (2003) 985–990

www.elsevier.com/locate/foodres

* Corresponding author. Tel.: +357-2239-2694; fax: +357-2239-

2793.

E-mail address: [email protected] (I.S. Patrikios).

a dimer (Leonard, 1979). The dimer was postulated tobe either acyclic or a six-membered cyclic dimer.There are a myriad of possible isomers, positional and

geometrical isomers of the double bond, as well as struc-tural isomers resulting from head-to-head or head-to-tailalignment of the starting material as well as for the cyclicdimers (McMahon&Crowell, 1974). Head-to-headmeansthat both carboxyl groups are at the same side and head-to-tail means that the carboxyl groups are at the oppositeside. When the oligomer is in the linear form (I), thehydrocarbon chains are free to rotate about the connectingC–C bond and the carboxyl groups can be at any position;but when the oligomer is in the ring form the carboxylgroups are frozen in one relative position (II–III).

Methyl oleate can be dimerized by di-tert-butyl peroxideto yield a mixture of isomers of an acyclic dimer. The tert-butoxy radical could remove a hydrogen atom from the 11or 8 carbon atoms of oleate. These allyl radicals wouldhave two resonance structures each, and all could coupleto give dimers. The following mechanism was proposed byPaschke, Peterson, Harrison, and Wheeler (1964).

The free radical species A, B, C and D couple ran-domly to form dimer, resulting in 10 species of dimer,AA, AB, AC, AD, BB, BC, BD, CC, CD, and DD.These dimers represent joining to the same extent at posi-tions 8, 9, 10 and 11 of the oleate segments of the dimer.At least six simple hydroperoxide products are

formed in the autoxidation of methyl oleates 1–6:

The most accepted mechanism for oleate autoxidationinvolves hydrogen abstraction at carbon-8 and carbon-11 since the hydrogen atoms at positions 8 and 11 areboth allylic and, thus, both would be expected to bereactive. The interactions between the unpaired electronon carbon-8 and carbon-11 and the p-electrons of theadjacent double bond produce two resonance-stabilizedallylic radicals (similar to di-tert-butyl peroxide oxida-tion mechanism) with delocalized electrons distributedover three carbon atoms, as shown below (7–8).The radicals have partial double bond character

between the carbons in the allylic system, and theseradicals can exist as trans–trans, trans–cis, cis–trans andcis–cis isomers.

.In our hands, pure oleic acid does not agglutinate

red cells, but heated samples agglutinate several speciesof red blood cells (Patrikios, Britton, Bing, & Russell,1994).

2. Materials and methods

2.1. Materials

HPLC grade solvents and fatty acids were obtainedfrom Sigma (St. Louis, MO). The lipid standards were

986 I.S. Patrikios, P.C. Patsalis / Food Research International 36 (2003) 985–990

obtained from Larodan Fine Chemicals (Malmo,Sweden).

2.2. Hemagglutination titer assays

Human RBCs were washed twice with 1 mM phos-phate-buffered saline (0.85%)–0.01% sodium azide(PBS-N) (pH 7.4). The volume of packed RBCs wasnoted and diluted to 5% RBC (v/v) with PBS-N. PBS-N(100 ml) was placed in the control well No. 1 of amicrotiter plate and 50 ml into the wells numbered 2–12.Lipid preparations (50 ml, 0.3–0.5 mg/ml) were seriallydiluted in the wells numbered 2–12. These dilutionswere done in duplicate. PBS-N (50 ml) were added to theserially diluted wells followed by 25 m1 of 5% RBCs toall wells. The plates were shaken on a Tektator Shaker(Stuart Scientific, UK) for 3 min and incubated at roomtemperature. The plates were read at half-hour intervalsby tilting them and observing the bottoms of the wells.The first well of each row served as the control.Hemagglutination was recognized as the button-like

settling of cells in contrast to flowing in the controlwells. Titer is expressed as the highest dilution of testsamples which still gives agglutination. Specific titer isdefined as titer per mg lipid per ml.

2.3. Thin-layer chromatography

Silica gel G plates (20�20 cm) (Macherey-Nagel,GmbH & Co. Germany) were prewashed in the devel-oping solvent system: isooctane/isopropyl alcohol/ace-tic acid (95:5:1,v/v/v), air dried for 30 mm andactivated by heating for 1 h at 120 �C under vacuum (15mm). From a stock solution of 0.5 mg/ml in chloro-form/methanol (1:1 v/v), 20–30 mg of sample was spot-ted onto silica gel plates using 2 ml Microcap pipettes(Blaubrand, Germany). Before development, the plateswere dried using a hand-dryer on a cool setting for 5min. The chromatography chamber, 26 cm�7 cm�24cm (d�w�h), was saturated with vapor from the sol-vent system for 30 min before development of plates.The plates were allowed to develop until the solventfront was about 2 cm from the top, removed and air-dried for 30 min. The plates were visualized in iodinevapor.For extraction from the plates the bands were visua-

lized by a strip exposed to iodine and each band wasscraped off the remaining undeveloped plate andeluted with chloroform/methanol (1:1 v/v) as in Kates(1986).

2.4. Heated olive oil preparation

Olive oil was heated at 130 �C for 24 h in a metalheating block (USA/Scientific, Olala, FL) in glass tubesopen to air.

2.5. Preparation of olive oil dispersions

Samples under investigation dissolved well in ethanol,but high concentrations yielded separation of phaseswhen mixed with PBS-N.Dispersions were prepared by diluting the samples to

0.5 mg/ml in ethanol. Dispersions of lipid extracts werefiltered through LC PVDF Acrodisc membranes (Gel-man) with size cut-off 0.2 mm, to remove particulatematter.

2.6. Synthesis of fatty aid methyl esters

Diazomethane was generated by using the Mini-Dia-zald apparatus (Technical Bulletin AL-180) (Aldrich,Milwaukee, WI). The yellow ether distillate contain-ing the diazomethane was added slowly to the fattyacid to be esterified and the solution was allowed tostand at room temperature in a hood until the etherevaporated.

2.7. Light microscopy

Photographs were taken using a Cytovision Analyzer(Medicell Co. Nicosia, Cyprus), attached to a ZeissAxioskop microscope. Agglutinated cells were takenfrom microtiter plates containing samples which werebeing titered and allowed to develop for approximately30 min. The following preparations were used: (1)untreated human RBCs, (2) most active lipid extractfrom heated virgin olive oil (1 mg/ml), (3) unheatedvirgin olive oil preparations (1 mg/ml). All samples weretotally dissolved in pure ethanol. Fresh human RBCswere washed once with citrate because the cells retaintheir shape much better in citrate than in PBS, andtherefore allow better photos to be taken. However, itshould be noted that cells washed in citrate will onlymaintain their integrity for approximately 2 h, and forthat reason, only that portion of drawn blood neededfor a single titer plate is washed.

2.8. Nuclear magnetic resonance spectroscopy

The 13C NMR spectra was recorded on an FT NMRspectrometer NR/400 (Bruker). Samples were picked upin 1 ml of CDCl3 (99.8%)+1% TMS (Cambridge Iso-tope, Woburn, MA) and transferred to NMR tubes(507-PP) (Wilmad, Buena, NJ) and filled up to 3 cm.The spectral width was 17 241 Hz for 13C NMR and thedelayed time between pulses was 4.0 s. The number ofscans collected for 13C NMR was 600.

2.9. Mass spectrometry

Mass spectra were obtained with Finnigan MAT SSQ70 spectrometer using chemical ionization (CI) with

I.S. Patrikios, P.C. Patsalis / Food Research International 36 (2003) 985–990 987

ammonia (134 EV) for molecular ions and electronimpact (El) (70 EV) for fragmentation patterns. Thetemperature program ran from room temperature to600 �C at 60� per min.

3. Studies of normal and malignant cell membrane

changes

Normal and malignant cells (donated by the cytoge-netic department of the Cyprus Institute of Neurologyand Genetics and from the Nicosia General Hospital)and cell proliferation and adhesion had been studiedafter incubation of the cells with the isolated lipidmolecules (defined as lipid agglutinins) produced as aresult of olive oil heat processing. The ideal conditionsfor the experiments of the cells had been established.The studies included concentration optimization of thesamples to be tested that affect membrane and or celladhesion. Briefly, the cells released from stock culturesby trypsin treatment, seeded in 15 cm diameter glasspetri dishes at 3�106 cells per dish, and grown in Dul-becco’s modified Eagle’s medium (DMEM) supple-mented with 10% fetal calf serum (FCS) for 72 h untilthey will be 70–80% confluent. Forty-eight hours beforeharvesting, medium supplemented with the isolatedcompounds in PBS (0.5 mg/ml) with continuous shak-ing for different time intervals up to 12 h. The mediumremoved and the cells rinsed twice with PBS. Cells har-vested by mild trypsinization (0.25% trypsin solution inPBS on the fourth day, at which time cell culturesreached saturation densities).

3.1. Electron microscopy

The experiment was performed as described in Kates(1986). Briefly, agglutinated cells were taken frommicrotiter plates containing samples which were beingtitered and allowed to develop for approximately 30min. Sample of human RBC’s with the unheated virginolive oil in phosphate buffer saline (1 mg/ml, control)and sample of human RBCs with the oligomer (isolatedfrom the heated virgin olive oil, l mg/ml) in phosphatebuffer saline, were fixed in 2.5% gluteraldehyde. After fix-ation the samples (8–10 ml) were centrifuged for 10 min at2600�g. The sediment was removed and placed in 1.5 mleppendorf tubes with 0.1 M phosphate buffer and cen-trifuged for 15 min at 7700�g. The buffer was removed,sediment was mixed with warmed agar and then placed at�20 �C for 5 min in order to solidify. The blocked speci-mens were washed with 0.1 m phosphate buffer, thenomiscated, dehydrated, and embedded in Araldite. Fromthese resin blocks, ultra thin sections were cut and stainedwith uranylacetate and lead citrate before being examineusing a JEOL transmission electron microscope.

4. Results and discussion

In addition to being concerned about the amount offat in our diet, we should also be careful about heatingoils to high temperatures and about exposing them tooxygen. The way in which oils are manufactured, pro-cessed and used affects their safety and nutritionalvalue.There are many studies in the literature about oxida-

tion of olive oil. Patrikios (2002) reported the hemaglu-tinating properties of oxidation and thermaldegradation products of olive oil after processing andcooking. In this study, we are interested in studying themolecular structure that is formed under the aboveconditions, and which is responsible for the hemagglu-tinating activity.The components of the heated olive oil were resolved

by thin layer chromatography on silica gel in a systemthat discriminates by the number of carboxyl groups(Table 1). Bands were cut out, eluted and titered againstHuman RBCs. The identification of the components ofthe fraction was carried out by comparison with Rfs ofthe components of a heated sample of oleic acid andstandard commercial monomer, linear dimer and trimer(Emery). Some byproducts in the heated olive oil sam-ple showed the same Rf values as the corresponding Rfsof the monomeric, dimeric and polymeric byproducts ofthe heated sample of oleic acid and the commercial oli-gomers. This might be the result of hydrolysis, whichcan take place under these conditions (heating in air), ofthe triacylglycerols backbone, releasing oleic acid or anyother unsaturated fatty acid, resulting in dimerizationand polymerization which have been shown to producestrong hemagglutinins.

Table 1

Rf values of the components of OA and virgin olive oil, unheated and

heated (24 h at 130�) on silica gel TLC in isooctane / isopropyl alco-

hol/acetic acid (95:5: 1, v/v/v)

Item

Monomer

OA

Heated OA

0.30

0.28

0.26

0.18

Virgin olive oil

Heated virgin olive oil

0.31

0.25

0.19

a The identification of the components of the fractions was by

comparison with Rfs of the components of standard commercial

monomer, linear dimer and trimer (Emery).b The concentration was approximately 0.5 mg/ml. Specific titer is

defined as titer per mg lipid per ml.

988 I.S. Patrikios, P.C. Patsalis / Food Research International 36 (2003) 985–990

Fig. 1 shows Human RBCs that were suspended in3.8% citrate (control) and Human RBCs which weremixed with the isolated oligomer preparation. As seenin Fig. 1, mixtures of chains and rosettes were observedby light microscopy (LM), but no fusion or lysis wasapparent. Agglutination was found to be time-depen-dent and involve cell clumping but not cell fusion. Theagglutination, which we observed, most likely, is due toa partial insertion of the oligomer formed into red cellmembrane in a manner which links red cell rather thanaffecting membrane permeability properties.The effect of the hemagglutinins on Human RBCs

was visualized by electron microscopy (EM) (Fig. 2).Human RBCs were suspended as above and mixed withthe isolated oligomer preparation. The results confirmthe above findings. Since centrifugation of the samplewas used in the procedure of slide preparation, the RBCin the control, where there is no clumping formation,tent to be settled in an order manner.The preliminary experimental results of untreated

olive oil and the oligomer isolated from the olive oilheated preparations against the malignant cells showedthat at low concentrations both have an inhibitory effect

of the tumor proliferation, with the oligomer to be mostactive. More extensive studies are required towardsthese findings.Fig. 3 shows the chemical ionization (CI) mass spec-

troscopy of the oligomer isolated from the heated sam-ple of olive oil after methylation resulting to thecorresponding dimethyl dioleate with a molecularweight of 590. These findings confirmed the structure ofacyclic ‘‘dehydro’’ dimer that we proposed and whichhad been first characterized by Paschke et al. (1964) ascontaining two double bonds (cis and/or trans) andcarbon to carbon linkages at the 8, 9, 10 and 11 car-bons. The allylic free radical species couple randomly toform dimer, resulting in 10 species of dimer.

Fig. 1. Light micrographs of human erythrocytes in citrate buffer (pH

7.4): (a) with no addition, (b) with the isolated virgin olive oil oligomer

preparation (1.0 mg/ml).

Fig. 2. Electron micrographs of human erythrocytes in citrate buffer

(pH 7.4): (a) with no addition, (b) with the isolated virgin olive oil

oligomer preparation (1.0 mg/ml).

I.S. Patrikios, P.C. Patsalis / Food Research International 36 (2003) 985–990 989

13C and 1H NMR also confirm the structures includ-ing the formation of cis and trans double bonds. Thesignals for tertiary carbon at 47.7 ppm (multiplet) indi-cate the acyclic linkages in the dimer. The signals at 27.4

ppm and 32.6 ppm indicate the allylic methylenes, cisand trans, respectively.We conclude that olive oil can undergo nutritional

damage when heated in air with a possible direct effecton human health.

Acknowledgements

The author wishes to thank the Cyprus Institute ofNeurology and Genetics for hosting the project, and inparticular Dr. Kyriakos Kyriakou who provided exper-tise in electron microscopy. This work was supported bygrant from the Research Promotion Foundation ofCyprus RPF 06/99.

References

Goebel, C. G. (1949). Polymerization of unsaturated fatty acids. US

Patent 2,482,761.

Goebel, C. G. (1953). Method for manufacturing polymerized fatty

acids. US Patent 2,664,429.

Leonard, E. C. (1979). Dimer Acids. In Kirk-Othmer encyclopedia of

chemical technology, (3rd ed., pp. 768–782).

McMahon, D. H., & Crowell, E. P. (1974). Characterization of pro-

ducts from clay catalyzed polymerization of tall oil fatty acids.

Journal of the American Oil Society, 51, 522–527.

Kates, M. (1986). Techniques in lipidology. New York: North Holland/

American Elsevier.

Myers, L. D., Goebel, C. G. & Barrett, F. O. (1960). Process for

polymerizing unsaturated fatty acids. US Patent 3,076,003.

Paschke, R. F., Peterson, L. E., Harrison, S. A., & Wheeler, D. H.

(1964). Dimer acid structure. The dehydro-dimer from methyl oleate

and di-t-butyl peroxide. Journal of the American Oil Chemists’

Society, 41, 56–60.

Patrikios, I. S. (2002). Lipid extracts isolated from heat processed food

show a strong agglutinating activity against human red blood cells.

Food Research International, 35, 535–540.

Patrikios, S. I., Britton, O.’N., Bing, D. K., & Russell, C. S. (1994).

Heating unsaturated fatty acids in air produces hemagglutinins.

Biochimica et Biophysica Acta, 1212, 225–234.

Fig. 3. Ion and mass chromatograms of dimethyl dioleate (since CI

was with ammonia, the actual m/e is observed mass—18).

990 I.S. Patrikios, P.C. Patsalis / Food Research International 36 (2003) 985–990