Development and Characterization of Magnetoliposomes for

Drug Delivery Applications

Faria, M.R., Martins, M.B.F., Gonçalves, M.C.

June 2011

Abstract

The objective of the present work is to synthesize and characterize magnetoliposomes, in which astable aqueous ferrofluid with magnetite nanoparticles covered with a surfactant is incorporated into li-posomes. Two different reagents were tested as surfactants, tetramethylammonium hydroxide and phos-phatidylcholine. The incorporation of these particles and synthesis of what are called magnetoliposomeswas successful, using a procedure based on the lipid film method followed by extrusion, with distribution ofthe magnetite associated with the lipidic membrane of the synthesized oligolamellar structures. The lipidused for the formation of the liposomes was soybean phosphatidylcholine and the concentration of ferrofluidwas varied. Cholesterol was also added to the lipid to evaluate its influence in the retention of ferrofluid. TheDried Rehydrated Vesicles method was also tested as an alternative method for the production of magne-toliposomes.

After synthesis, all the samples were characterized through Transmission Electron Microscopy, DynamicLight Scattering, Fourier Transform Infrared Spectroscopy and SQUID magnetometry. The results were sat-isfactory in all techniques, with TEM and DLS defining the structure of the magnetoliposomes very clearlyand with the distribution of the ferrofluid very discernible in the TEM images. All the samples showed sig-nificant magnetization, a result particularly important for the magnetoliposomes samples. In FTIR, someinteresting conclusions were drawn, namely the increase of membrane ordering and decrease of permeabil-ity with cholesterol and with ferrofluid.

Key Words: Nanoparticles, Magnetoliposomes, TEM, DLS, FTIR, SQUID.

1 Introduction

An ideal drug delivery system is supposed to be bothefficient and discreet, delivering the drug to a specificlocation, without being cleared off by the patient’s im-mune system and without interfering with cells otherthan the targeted ones. A lot of the present researchfocuses on cancer chemotherapy, trying to find a ve-hicle for the strong chemicals administered to can-cer patients that target only the cancer cells, thusdiminishing the drug’s severe and harmful toxic ef-

fect on healthy tissues. Also, it is convenient that thissystem is biocompatible and preferably biodegrad-able. Liposomes seem ideal as they have the po-tential of satisfying all these requisites. They consistof an aqueous core entrapped by one or more bilay-ers composed of natural or synthetic lipids. Naturallipids are biologically inert and weakly immunogenic,having low intrinsic toxicity. Their composition canbe varied and they can encapsulate drugs with dif-ferent lipophilicities, either inside the core or inside

1

the membrane. Also, their versatility in terms of ac-tive targeting, such as coupling with antibodies andcoating with ligands targeting proteins expressed oncancer cell membranes, for example, is a great as-set, as is their susceptibility to biological triggers suchas temperature. All these factors can be manipu-lated and studied, making this drug delivery systema grand focus of current research in this area. Su-perparamagnetic nanoparticles, on the other hand,have also been extensively studied for numerous ap-plications, including drug delivery, with an appropri-ate surface chemistry and coating. The objective ofthese particles is to provide a targetable drug deliv-ery through an application of an external magneticfield or to be used as contrast enhancers in magneticresonance imaging. Another application of these par-ticles could be hyperthermia, local heating of a spe-cific tissue to yield its local destruction without affect-ing other healthy cells. In the present work, one aimsto add the benefits of the two systems, synthesizingmagnetoliposomes, that is to say, doping liposomeswith magnetic nanoparticles.

2 Methods

2.1 Ferrofluid Synthesis

Ferrofluid synthesis with stabilization of the mag-netite nanoparticles with tetramethylammonium hy-droxide (TMAOH) was based on the work by Bergeret al. [1]. TMAOH solution 25% in water was fromSigma-Aldrich, as were the Iron (II) Chloride Tetrahy-drate, the Iron (III) Chloride Hexahydrate and the Hy-drochloric Acid. Ammonia was from Pronalab. Thismethod is based on the mixture of Fe2+ and Fe3+

in aqueous media, coprecipitation of the correspond-ing hydroxides and upon the addition of a strong al-kali, the fast aging of those oxides under vigorousstirring to form magnetite. 1.0 mL of FeCl2 solution(2.0 M in 2.0 M HCl) was combined with 4.0 mL ofFeCl3 solution (also 2.0 M in 2.0 M HCl) and the so-lution was magnetically stirred vigorously. 50 mL ofan aqueous ammonia solution (0.7 M) were added

dropwise to this mixture and the magnetite startedto precipitate. With the help of a magnet, the su-pernatant was discarded and the rest centrifuged for1 minute at 1000 rpm. After another decantation,the precipitate was washed with bidistilled water andcentrifuged and decanted again. 8 ml of TMAOHwere then added to the solution with vigorous stirringand the excess liquid was poured off. This ferrofluid(TMAOH:Fe3O4 (8 ml/0.3 g)) was relatively stable inaqueous solution.

A different attempt to a stable ferrofluid was per-formed, by stabilizing magnetite nanoparticles withphosphatidylcholine (PC), from egg (EPC) or soy-bean (SPC). It was based on the protocol by Giri etal. [2]. PC is added during the synthesis of mag-netite. It started with a preparation of a 0.2 M so-lution of FeCl2 and FeCl3 salts with a molar ratio2:1. To 5 mL of this solution, one added a solutionof PC (0.046 g) in methanol (1 mL), all the while un-der magnetic stirring. Afterwards, 50 mL of a 0.93 Mammonia solution were added dropwise to the pre-vious solution, also under stirring. Two different pro-cedures were attempted from this point, keeping thesolution under stirring for 15 minutes at 60 ◦C or fol-lowing the former procedure and decant immediatelythe solution. The precipitate was washed (and de-canted) four times with water, one with acetone andone with methanol. From the four synthesis only themagnetite covered with SPC and without heat treat-ment (SPC:Fe3O4) was incorporated into liposomes,due to its better apparent stability in solution.

2.2 Magnetoliposome Synthesis

Generally speaking, to form liposomes, the lipid orlipids are first mixed and dissolved in an organic sol-vent (chloroform in the present work) and this solventis removed in a rotary evaporator (rotavapor r-144 ofBÜCHI) to yield a lipid film. Afterwards, the film hy-dration takes place, by adding an aqueous mediumto the container with agitation above the lipid phasetransition temperature, so that hydration takes placein the fluid phase of the lipid. In the case of thepresent work, based on [3], 0.076 g of SPC were

2

dissolved in chloroform (4 mL) and dried in a rotaryevaporator under reduced pressure at 40 ◦C. Afterthe complete removal of the chloroform, the lipid filmwas hydrated for 24 h with an aqueous solution offerrofluid (5 mL) in order to obtain a lipid concentra-tion of 20 mmol/L and afterwards 10 minutes of bathsonication were applied prior to extrusion. Extrusionwas made progressively with membranes of 800 nm,600 nm, 400 nm and 200 nm, twice for each mem-brane, except for the 200 nm, which was used fourtimes, with membrane exchange between the sec-ond and the third. After this process of sample siz-ing, other characterization methods were applied tothe samples. Some problems were found when ex-truding the sample as most ferrofluid got stuck on thepolycarbonate membranes, particularly when pass-ing from 400 nm to 200 nm. This is probably dueto the not complete efficiency of the surfactant men-tioned previously and it was partially solved by syn-thesizing the ferrofluid on the day before incorpora-tion and leaving it to settle overnight in solution, dis-carding the agglomerates that deposited on the flaskand using only the supernatant for hydration of thevesicles.

Another method for preparing liposomes isthrough Dried Rehydrated Vesicles (DRV). The lipo-somes prepared through this method have the capac-ity to entrap larger quantities of hidrophilic solutes,when compared to other types of liposomes [4]. Inthe case of the present work, the advantage that wasaimed for was the high entrapment efficiency of theferrofluid. This efficiency is useful, not only in econ-omy of reagents but also as less lipid is requiredto achieve a good local concentration of magnetite.Generally speaking this technique involves freeze-drying the empty vesicles with their consequent dis-ruption after the regular procedure of film formationand hydration. Afterwards, the vesicles suffer two hy-drations, the first one with around 20 % of the finalhydration volume with a concentrated solution of thesubstance to entrap and the second one, performedafter a certain amount of time, variable with the in-tended application, with the remaining hydration vol-ume of distilled water. This results in theory in a

higher encapsulation efficiency precisely due to thecontrolled rehydration step.

The first synthesis was made with SPC and abreak of 24 h between the first and second hydra-tions. Other synthesis were performed with a break of2 h between hydrations. Two more synthesis with thesame differences in the hydration period were madewith SPC and Cholesterol (20%). The final lipid con-centration was 20 mmol/L in all cases. All the mag-netoliposomes samples were passed through a chro-matography column using bi-distilled water as eluentin order to exclude the ferrofluid that wasn’t incorpo-rated.

2.3 Characterization Methods

Four methods were used to characterize thenanoparticles obtained, Dynamic Light Scattering(DLS), Transmission Electron Microscopy (TEM),Fourier Transform Infrared Spectroscopy (FTIR) andSQUID magnetometry. TEM was the one exploredmore thoroughly, as it gives a more realistic percep-tion of what our samples looks like, and allows oneto measure its size and address roughly the statusof one’s experiment, thus providing a standing pointfrom which to evolve. In this work TEM imaging wasobtained using the electron microscope H-8100 fromHitachi and the sample preparation consisted in plac-ing a drop of the sample on top of a carbon grid,letting it dry completely, placing one drop of chemi-cal contrast Phosphotungstic Acid (PTA), waiting 45sthen wiping off the excess contrast with some fil-ter paper. There were some problems with the in-teraction of the contrast with the ferrofluid, as it re-acted well with the vesicles and allowed us to visual-ize and discern the liposomes membranes, but whenprogressively higher concentrations of ferrofluid wereused, the film became too dark and the images be-came difficult to interpret and discern. Some of thesamples were analyzed without contrast precisely forthis reason, to clarify with certainty the position of themagnetite nanoparticles. DLS was used to measurethe diameter of the nanoparticles through a Zetasizernano S series equipment by Malvern. FTIR analysis

3

was performed to check whether there was some de-tectable difference between empty vesicles and vesi-cles loaded with ferrofluid, using a FTIR spectropho-tometer Nicolet 5700. The samples were preparedwith the pellet method, in which the powder sam-ples were ground together with 200 mg of KBr andthe powder mixture is then compressed to yield pel-lets that were analyzed in the equipment. A back-ground spectrum was obtained with a pellet contain-ing only KBr. Principal Component Analysis (PCA)was performed on the spectra, inspired by what isdone by several authors to differentiate and auto-matically classify yeast strains [5]. They are ableto take spectra with a high degree of similarity and,through PCA, differentiate several bacterial isolatesbased upon the small differences between the spec-tra that this method enhances. PCA is commonlyemployed in the interpretation of the infrared spec-tral data variance and was thus applied on the sam-ples analyzed by FTIR on the present work to checkwhether it allowed differentiation of the several mag-netoliposomes among them or of the liposomes withand without ferrofluid. If successful, it would allow fora fast classification of these vesicles, with a methodthat, as far as the author’s knowledge, wasn’t yet ap-plied to this area of research. Finally, the magne-tization curves of the samples were obtained witha SQUID magnetometer. The model used in thepresent work was the Magnetic Property Measure-ment System by Quantum Design, with a sensitivityup to 10−8 emu. The sample was placed in a non-magnetic gelatin capsule and compressed with paperto prevent movement and breakage during vacuum.The magnetic field was ramped from 0 to 60000 Oeat 300 K.

3 Results

3.1 Ferrofluid

The characterization of the magnetite nanoparticlescovered with TMAOH started with a TEM analy-sis, with an image of the sample TMAOH:Fe3O4

(8 ml/0.3 g) presented in figure 1. The magnetite par-ticles were relatively monodisperse and presented asize around 10-15 nm without the surfactant, and asize around 50-60 nm when covered with TMAOH.Their size is not very easy to control through syn-thesis by coprecipitation, but an advantage of thismethod is its simplicity and the large quantity ofnanoparticles that can be synthesized.

Figure 1: TEM imaging of the synthesized ferrofluidTMAOH:Fe3O4 (8 ml/0.3 g).

Concerning the stabilization of magnetitenanoparticles with PC, TEM imaging showed that theparticles also appeared to be fairly monodisperse,with size distribution from 5 nm to 15 nm and a roundshape. TEM imaging of sample SPC:Fe3O4 withoutheat treatment is shown in figure 2.

Figure 2: TEM imaging of the synthesized ferrofluidSPC:Fe3O4.

The other ferrofluids synthesized with PC as sur-factant were also analyzed through TEM imaging andyielded similar results. All the samples seemed ratherheterogeneous concerning PC distribution, and thephospholipid seems to lead to agglomeration of thenanoparticles, instead of stabilization and dispersion.This was common to the four procedures, and dif-

4

ferences amongst them weren’t found, at least mi-croscopically. Macroscopically, when suspended inchloroform or even water, the nanoparticles stabi-lized with soybean phosphatidylcholine without heat-ing yielded the apparently most stable suspension,with the other three forming clearly visible agglomer-ates. For this reason only, these were the particleschosen to attempt to incorporate into liposomes.

FTIR measurements of magnetite and ferrofluidTMAOH:Fe3O4 (8 ml/0.3 g) were also obtained andare depicted in figure 3.

Figure 3: FTIR spectra of magnetite and ferrofluidTMAOH:Fe3O4 (8 ml/0.3 g).

Analyzing first the spectrum of magnetite, we seea peak around 2364 cm−1 that can be attributed to thepresence of some carbon dioxide [6] that wasn’t sub-tracted on the background due to possible changes inthe atmosphere inside the equipment when the sam-ples are switched. The clear bands around 570 cm−1

and 400 cm−1 (in this spectra we only see the be-ginning of this peak, cause the limit of our analy-sis was precisely the 400 cm−1), are attributed tothe Fe-O stretching mode [7] and are therefore themost characteristic region of the magnetite spectra.Concerning the ferrofluid spectrum, one sees a verybroad band towards the end, with the highest pointon 530 cm−1 that is attributed to the magnetite pres-ence. Another very prominent peak is the one around3450 cm−1, that is attributed to the O-H stretch ofwater or the hydroxide groups of the TMAOH [5].The last well defined peak around 1650 cm−1 couldbe attributed to the stretching vibration of the C-N

groups [5, 6]. The presence of some maghemiteon the samples is also possible, as their character-istic bands for the Fe-O groups are on 700 cm−1,630 cm−1 - 660 cm−1 and 620 cm−1 [7]. On our spec-tra, a hint of this bands is present, particularly in theferrofluid, so the presence of some maghemite is astrong possibility.

The obtained magnetization measurements formagnetite are depicted in figure 4.

Figure 4: Magnetization curve obtained for the syn-thesized magnetite nanoparticles. A saturation mag-netization of 76 emu/g was obtained.

The absence of hysteresis corroborates the hy-pothesis of superparamagnetism of the synthesizednanoparticles, making them ideal for biomedical ap-plications, where a remanent magnetization isn’t de-sirable, and a rapid relaxation of the magnetic mo-ments to random directions when the field is removedis important. However, more complete measure-ments with temperature are needed in order to con-clude with certainty that the particles are in fact inthe single domain state, namely the detection of theblocking temperature. A saturation magnetization of76 emu/g was obtained. However, as the quantityof sample used in this case and with ferrofluid wassmaller than 1 mg to prevent saturation of the SQUID,which corresponds to a very small volume, one hasto consider an error on the magnetization values ofthese samples up to 25 %. Nevertheless, taking intoaccount that the bulk magnetite saturation magneti-zation is 82 emu/g [8], 76 emu/g is still about 92 %of the bulk saturation magnetization. And even con-sidering the maximum error of 25 % to the lowervalue, the saturation magnetization of the synthe-

5

sized nanoparticles would still comprise about 70 %of the bulk saturation magnetization. This reductionis due to the size of the nanoparticles, as an increaseon the surface area leads to an increase on the num-ber of surface atoms with respect to bulk atoms. Themagnetic moments of these surface atoms are lessaligned than those of the atoms in the bulk and thisleads to a reduction in magnetization. The reason forthis lack of alignment is still a topic of discussion [8].Also, the increase of the contribution from impuritiesand oxides at the surface layer helps decrease thissaturation magnetization.

Concerning the ferrofluid, the magnetizationcurve obtained for the sample TMAOH:Fe3O4

(8 ml/0.3 g) is depicted in figure 5.

Figure 5: Magnetization curve obtained for the syn-thesized ferrofluid TMAOH:Fe3O4 (8 ml/0.3 g). A sat-uration magnetization of 100 emu/g was obtained.

The shape of the ferrofluid magnetization curve is,as expected, similar to the magnetite curve, shownpreviously. The difference is mainly the saturationmagnetization that reaches the value of 100 emu/g.As mentioned previously for the magnetite samplethis sample may present an error in the saturationmagnetization value up to 25 %. Nevertheless, onecan compare the two analysis and discern clearlythat the ferrofluid saturation magnetization is in prin-ciple higher than that of the magnetite. Guardia etal. [9] covered magnetite nanoparticles with a dif-ferent surfactant (oleic acid) and obtained, similarlyto the present work, a higher saturation magnetiza-tion of the ferrofluid with respect to the magnetitenanoparticles with no surfactant. They claim thatthe oleic acid stabilizes the surface of the nanoparti-

cles, through the new O2− surface ligands of the sur-factant, and helps reduce the surface spin disorder,thus increasing the magnetization and approximatingit from that of the bulk magnetite. The same explana-tion could be applied in the present work. Also, theincrease in size of the ferrofluid particles, as agglom-erates of five and six particles are formed and stabi-lized together, could yield to a reduction of the sur-face effects mentioned previously and thus increasespin alignment at the surface of the nanoparticles.

3.2 Magnetoliposomes

Focusing first the vesicles prepared by theDRV method, TEM image for the sampleSPC:Chol(24 h)‖TMAOH:Fe3O4 is presented in fig-ure 6.

Figure 6: TEM image of magnetolipo-somesSPC:Chol(24 h)‖TMAOH:Fe3O4 obtainedthrough the DRV method without chemical contrast.

Through the process of extrusion, the color of thesamples was almost lost, so that this method didn’tappear very successful. TEM images were collectedwithout chemical contrast, to assess whether somemagnetite was still present in the samples and toevaluate this ferrofluid distribution on the sample. Be-tween the samples analyzed with different hydrationperiods and PC or PC:Chol formulations, the differ-ences in color, TEM or DLS analysis weren’t verydiscernible. The presence of magnetite is very clearin these samples, although not in a very high con-centration. The magnetite is present in agglomer-

6

ates and doesn’t present any clear interaction withthe lipidic membrane. One can assume that this fer-rofluid is in fact inside the liposomes, as the sampleswere passed through the chromatography column .With contrast, the images obtained weren’t very dis-cernible, the film became very dark possibly due tothe ferrofluid present, making it almost impossible todraw conclusions from these images, although lipidicmembranes were discerned in some of them.

Concerning the classical film method, the re-sults were quite different. The concentration of fer-rofluid was varied from 1.25 g of magnetite permole of lipid (SPC‖TMAOH:Fe3O4 (1.25 g/mol)),to 25 g/mol (SPC‖TMAOH:Fe3O4 (25 g/mol)) and75 g/mol (SPC‖TMAOH:Fe3O4 (75 g/mol)). Also,the effect of the addition of cholesterol was studied(SPC:Chol‖TMAOH:Fe3O4 (25 g/mol)). The vesiclesobtained are oligolamellar structures (have aroundfive bilayers surrounding them) and the ferrofluidis distributed in association with these layers, asshown in the TEM image for the SPC‖TMAOH:Fe3O4

(75 g/mol) sample,figure 7.

Figure 7: TEM image of magnetoliposomesSPC‖TMAOH:Fe3O4 (75 g/mol).

The reason for the ferrofluid distribution in asso-ciation with the membranes is still unclear, and couldpossibly be attributed to a specific interaction of thelipid headgroups and magnetite, already mentionedin several works [10, 11], where the chemisorption ofthe polar headgroups to the iron oxide surface wasdescribed. Also, on [12], a similar interaction with

the lipidic membrane was found with iron ions, andthey also document the affinity between the phos-phate groups of PC and the iron ions, with attractiveCoulombic interactions playing the main role. The notcomplete stability of our ferrofluid and the possiblyincomplete coverage of the magnetite nanoparticles,might have also lead to this curious distribution, as itallowed magnetite to contact directly with the phos-pholipids and thus promote the referred interaction.

TEM images were taken to the sampleSPC:Chol‖TMAOH:Fe3O4(25 g/mol) without chem-ical contrast to check the magnetite distributionaround the vesicles, as shown in figure 8.

Figure 8: TEM image of magnetoliposomesSPC:Chol‖TMAOH:Fe3O4(25 g/mol).

The ferrofluid is associated clearly with the lipo-some membranes, allowing the visualization of struc-tures with a diameter around 200 nm.

Incorporation of the ferrofluid synthesized with PCinstead of TMAOH was also tried in PC:Chol lipo-somes but the TEM images taken (not shown) re-vealed little success as it seemed that the mag-netite nanoparticles remained aggregated and notdisposed around the liposomes.

FTIR measurements were taken for empty lipo-somes and magnetoliposomes SPC‖TMAOH:Fe3O4

(1.25 g/mol), SPC‖TMAOH:Fe3O4 (25 g/mol),SPC‖TMAOH:Fe3O4 (75 g/mol) (two samples) andSPC:Chol‖TMAOH:Fe3O4 (25 g/mol). The result ispresented in figure 9.

The band around 3450 cm−1, common to all sam-

7

ples could be attributed to some remaining water, de-spite the lyophilization. The bands around 2925 cm−1

and 2853 cm−1 are attributed to antisymmetric andsymmetric methylene vibrations as mentioned previ-ously [2, 5]. The peak around 2361 cm−1 is assumedto correspond to some remaining carbon dioxide in-side the equipment with the change of atmospheresince the background was captured, as mentionedpreviously with the magnetite.

Figure 9: FTIR spectra of empty liposomes and mag-netoliposomes with different concentrations of fer-rofluid.

At 1736 cm−1 we find the peak correspondingto the C=O bond of ester functional groups of fattyacids [5]. The band around 1465 cm−1 is due to bend-ing vibrations of the CH2 group and bands around1233 cm−1, 1090 cm−1, 970 cm−1 and 820 cm−1 areattributed to the presence of PO3−

4 group vibrationmode [2, 5]. Having stated and labeled all peaks,there is also one at 580 cm−1 that could be attributedto the presence of magnetite on the magnetolipo-somes, as it is relatively prominent on the samplesthat present magnetite and almost inexistent on theempty liposomes. However, the most amazing differ-ences became apparent when analyzing more partic-ularly the methylene vibrations bands, shown in fig-ure 10.

Several studies [13–15] have related the infor-mation of these bands, particularly the symmetricstretching band around, 2850 cm−1, to the ordering ofmembrane lipids, stating that the shift of this band to-wards higher wavenumbers and its broadening points

to a higher membrane fluidity and consequent disor-der of the membrane lipids due to addition of Gaucheconformers. With the addition of cholesterol, there isa small deviation to the right end of the spectra to-wards the lower wavenumbers, which, based on whatwas mentioned would indicate a decrease in chaindisorder of the lipidic membrane and a higher mem-brane rigidity, which is the role that cholesterol shouldbe playing.

Figure 10: Amplification of FTIR spectra of figure 9,showing the methylene symmetric stretching band.

Concerning the empty liposomes when comparedto the magnetoliposomes, there doesn’t appear to bea discernible shift, at least not one that we can pin-point with certainty but there is a broadening of theempty liposomes when compared to the magnetoli-posomes, either with PC or with PC:Chol. Accordingto [15], this translates to an increase of the mobilityand disorder of the acyl chains in empty liposomeswhen compared to magnetoliposomes. Other studiesneed to be made in order to confirm this hypothesis,namely studies with temperature variation to checkwhether and how this broadening and these shifts areaffected by temperature, to take into account the be-havior of these membranes and the exact role of theferrofluid in terms of its stability or lack thereof.

The result of the PCA mentioned in the methodssection is depicted in figure 11. This technique wasapplied to the second derivative of the spectra, ob-tained through the finite differences method, basedon what was done in [5], where they state that thesecond derivative enhances by itself the differences

8

between the various spectra. 97.2 % of the varianceis accounted for by the first two principal components,so one can consider the scores presented in figure 11to be very representative of our data.

Figure 11: Principal Component Analysis of the FTIRspectra belonging to the empty liposomes and mag-netoliposomes samples from figure 9.

First of all, this result is to be regarded carefully,as the few samples analyzed aren’t enough to make itstatistically significant, so it should only be regardedas a pointer for future attempts. Looking at this pre-liminary results, the outlook seems positive, as thereis a clear distinction between magnetoliposomes andthe empty liposomes. Even the very small concen-tration of 1.25 g of magnetite per mole of PC wasplaced in the same cluster as the other magnetolipo-somes, which is a very good indication of the possiblesuccess of this method in identifying the presence orabsence of any amount of ferrofluid on synthesizedsamples. Secondly, a rather curious observation con-cerning the ferrofluid concentration is observed, asthe concentrations seem to be placed on the scoregraph in ascendent (or descendent depending onperspective) order. This could point towards a suc-cessful classification of the vesicles through ferrofluidconcentration. And finally PC:Chol, although in ap-proximately the same concentration as the sampleof PC with 25 g of magnetite per mole of lipid, re-mained on a lower position when compared to theother lipids, which could point towards its differentia-tion with respect to lipid formulation. As mentioned,

this result however satisfactory and curious, isn’t aguaranty of the success of the PCA method to theclassification of liposomes, mainly due to the smallnumber of samples analyzed, but it is a first step inthe, hopefully, right direction.

The magnetization measurements were onlymade for two magnetoliposomes samples, for lackof further availability of the SQUID magnetometer.Sample SPC‖TMAOH:Fe3O4 (75 g/mol) result isshown in figure 12 (a) and the result for sampleSPC:Chol‖TMAOH:Fe3O4 (25 g/mol) is shown in fig-ure 12 (b).

(a)

(b)

Figure 12: Magnetization curve obtained fora) the magnetoliposomes SPC‖TMAOH:Fe3O4

(75 g/mol) and b) the magnetoliposomesSPC:Chol‖TMAOH:Fe3O4 (25 g/mol).

Taking into account that the amount of sampleanalyzed for both cases also revolved around a fewmg and the fact that in those samples, a very smallamount of ferrofluid was present as some losses arebound to occur during the incorporation process, thefact that a magnetization signal was detected is a re-markable achievement. As expected, in such a smallquantity, no saturation is achieved, and larger field

9

intensities are needed to increase small steps on themagnetization.

4 Conclusion

The synthesis of a ferrofluid with TMAOH and itsincorporation in magnetoliposomes was successful.However, its colloidal stability, when in aqueous so-lution, wasn’t ideal, and future work on this areashould be started by improving this ferrofluid and as-suring its complete stability in solution. This wouldprovide a more successful incorporation and a moreaccurate measure of the ferrofluid that actually re-mains inside the liposomes. Some curious observa-tions were made in the magnetoliposomes synthe-sis, which could be useful in future developmentsof this technology, namely distribution of the mag-netite nanoparticles inside the oligolamellar struc-tures membranes, possibly in the hydrophilic zone,as our ferrofluid is an aqueous solution. In FTIR, weconcluded that cholesterol, as expected, induced ahigher membrane rigidity with a more ordering of thelipidic chains. Ferrofluid, interestingly, also increasedthe stability of the lipidic membrane, thus possibly re-ducing spontaneous leakage of liposome content, inthe possibility of a real application for this technol-ogy. Principal Component Analysis was also appliedto the FTIR spectra of these vesicles and at leasttwo clear clusters were found of magnetoliposomesand empty liposomes. A possibility of grouping thesamples according to concentration of ferrofluid wasalso verified, so that some kind of automatic classi-fication could perhaps be established with more ofthese samples. This should be measured for far moresamples to have some solid statistical meaning andthat would be an interesting next step in this workin order to validate these results and possibly givethem some real utility. Finally, magnetization mea-surements confirmed unequivocally the presence ofiron in sufficient amount to account for a measurablemagnetization of these vesicles, which was a remark-able result. As far as the author’s knowledge these

are the first magnetization curves obtained for mag-netoliposomes. However, magnetization measure-ments more precise with, for instance, temperaturedependence need to be made in order to assure thesuperparamagnetism of these nanoparticles.

References[1] Berger, P., Adelman,N.B., Beckman, K.J., Campbell, D.J.,

Ellis, A.B., Lisensky, G.C., Preparation and Properties of anAqueous Ferrofluid, 1999, J. Chem. Ed. 76, 943-948.

[2] Giri, J., Thakurta,S.G., Bellare, J., Nigam, A.K., Bahadur,D., Preparation and Characterization of Phospholipid Sta-bilized Uniform Sized Magnetite Nanoparticles, Journal ofMagnetism and Magnetic Materials, 2005, 293, 62-68.

[3] Sabaté, Raimon, Barnadas-Rodríguez, Ramon, Callejas-Fernández, José, Hidalgo-Álvarez, Roque, Estelrich,Joan,Preparation and Characterization of Extruded Magne-toliposomes, International Journal of Pharmaceutics, 2008,347, 156-162.

[4] Antimisiaris, S.G., Weissig, V. (Ed.), Preparation of DRV Li-posomes, Liposomes: Methods in Molecular Biology, vol.605, Humana Press, 2010.

[5] Lin, M., Al-Holy, M., Chang, S., Huang, Y., Cavinato, A.G.,Kang, D., Rasco, B.A., Rapid discrimination of Alicyclobacil-lus strains in apple juice by Fourier transform infrared spec-troscopy, International Journal of Food Microbiology, 2005,105, 369-376.

[6] Leibniz Institute for Age Research / Fritz Lipmann Institute,[www.imb-jena.de](27th April 2011).

[7] Namduri, H., Nasrazadani, S., Quantitative Analysis of IronOxides Using Fourier Transform Infrared Spectrophotome-try, Corrosion Science, 2008, 50, 2493-2497.

[8] Gubin, Sergey P., Magnetic Nanoparticles, Wiley-VCH Ver-lag GmbH & Co., Germany, 2009.

[9] Guardia, P., Batlle-Brugal, B., Roca, A.G., Iglesias, O.,Morales, M.P., Serna, C.J., Labarta, A., Batlle, X., SurfactantEffects in Monodisperse Magnetite Nanoparticles of Con-trolled Size, Submitted to Journal of Magnetism and Mag-netic Materials.

[10] De Cuyper, M., Soenen, S.J.H., Weissig, V. (Ed.), CationicMagnetoliposomes, Liposomes: Methods in Molecular Biol-ogy, vol. 605, Humana Press, 2010.

[11] De Cuyper, M., Joniau, M., Mechanistic Aspects of theAdsorpiton of Phospholipids Onto Lauric Acid StabilizedFe3O4 Nanocolloids, Langmuir, 1991, 7, 647-652.

[12] Faure, C., Meyre, M., Trépout, S., Lambert, O., Lebraud, E.,Magnetic Multilamellar Liposomes Produced by In Situ Syn-thesis of Iron Oxide Nanoparticles: Magnetonions, J. Phys.Chem. B, 2009,113, 8552-8559.

[13] Severcan, F., Baykal, U., Suzer, S., FTIR studies of vitaminE-cholesterol-DPPC membrane interactions in CH2 region,Fresenius J Anal Chem, 1996, 355, 415-417.

[14] Rebolj, K., Bakrac, B., Garvas, M., Ota, K., Sentjurc, M.,Potrich, C., Coraiola, M., Tomazzolli, R., Serra, M.D., Macek,P., Sepcic, K, EPR and FTIR studies reveal the importanceof highly ordered sterol-enriched membrane domains for os-treolysin activity, Biochimica et Biophysica Acta 1798, 2010,891-902.

[15] Zhao, L., Feng, S., Effects of cholesterol component onmolecular interactions between paclitaxel and phospholipidwithin the lipid monolayer at the air-water interface, Journalof Colloid and Interface Science, 2006, 300, 314-326.

10