Effect of gangliosides on structure and integrity of polyethylene glycol (PEG)-stabilized liposomes

(cid:1) Similar to PEG-lipids, gangliosides are capable of inducing the formation of bilayer disks. (cid:1) PEG-lipids and gangliosides work in an additive manner. (cid:1) Little if any ganglioside can be added to liposomal preparations containing 5 mol% DSPE-PEG without disk formation. (cid:1) The presence of bilayer disks may affect the immunological response to the liposomal formulations.

system (MPS).In order to reduce clearance by the MPS and increase the blood survival time, systemically administered drug nanocarriers are commonly provided with a coat of polyethylene glycol (PEG) [1,2].In case of liposomes, the PEGylation is usually achieved by incorporation of PEG-conjugated lipids (PEG-lipids) in the liposome membrane [3].The PEG-lipids typically consist of PEG with molecular weight between 1 and 5 kDa covalently bound to a lipid anchor in the form of a phospholipid, ceramide, or cholesterol.
Although PEGylated liposomes in general display long blood survival times after a single intravenous injection, repeated administration can lead to rapid elimination of the liposomes from the bloodstream.This seemingly contradictory behavior, referred to as the accelerated blood clearance (ABC) phenomenon [4], is believed to be caused by an immune response to PEG.More specifically, the first dose of PEGylated liposomes initiates robust production of anti-PEG immunoglobulin M (IgM), which then bind to PEG on subsequently injected liposomes and promote their rapid removal from circulation [5][6][7].The development of anti-PEG antibodies does not only hamper the therapeutic efficacy of the liposomal drug formulations but may also lead to severe adverse reactions against PEGylated liposomes and other PEGylated therapeutic products [8].The situation is further complicated by the fact that anti-PEG IgM can be detected in a large proportion of healthy individuals never treated with PEGylated drugs [9,10].Strategies to avoid or suppress anti-PEG immune response are thus urgently needed.
Previous findings indicate that the anti-PEG IgM response and ABC phenomenon can be mitigated through conscious modulation of inherent properties, such as size, structure, PEG-surface density, and chemical composition of the PEGylated nanocarriers [6].Interestingly, a number of studies have reported results suggesting that the anti-PEG immunity of PEGylated liposomes can be attenuated by supplementing the liposomes with high (above 5 mol%) amounts of PEG-lipid [11], as well as by incorporating gangliosides in their membranes [12,13].These conclusions have been drawn under the assumption that the increased PEG-lipid concentrations, and the co-incorporation of PEG-lipids and gangliosides, do not significantly alter the particle size, shape or homogeneity of the liposomal preparations.Since both PEG-lipids and gangliosides are micelle forming lipids, it is not self-evident, however, that the liposomes remain structurally unaltered when these components are incorporated.It is well documented that PEG-lipids can induce the formation of non-liposomal structures when added in concentration exceeding the bilayer saturation limit [14].Thus, depending on the composition and phase state of the lipids, cylindrical micelles or flat, circular bilayer disks can appear in the preparations [15,16].Numerous studies have addressed issues concerning the role of gangliosides in biological membranes.The results and insights gained from these studies suggest that gangliosides may affect membrane structure, and have important influence on, e.g., domain formation and lipid distribution, through a multitude of complex, and composition dependent, mechanisms [17][18][19].Little data is currently available, however, on how the inclusion of gangliosides affects the type and shape of the self-assembled structures found in liposomal preparations.Given the structural similarities between PEG-lipids and gangliosides (see Fig. 1 for molecular structure of DSPE-PEG(2000), GM1 and GD1a) it may be anticipated, however, that these two classes of micelle forming lipids tend to induce lipid assemblies of the same, or similar type.Further, it is plausible that an additive effect could be observed upon their simultaneous presence in the preparations.
The existence of non-liposomal structures, such as bilayer disks, in liposomal preparations can be difficult to detect with dynamic light scattering (DLS) and other indirect techniques.Cryotransmission electron microscopy (cryo-TEM) has, on the other hand, proven to be a very useful technique for the detection and visualization of coexisting lipid structures in complex samples [20].In the present study we employ cryo-TEM for structural characterization of liposomal preparations supplemented with PEGlipids and/or gangliosides.Particular attention is given to systems with the same lipid composition as used in previous studies focused on effects related to anti-PEG immunity [11][12][13].

Preparation of liposomes
Liposome components were weighed or pipetted from stock solutions in chloroform or methanol.A lipid film was obtained by evaporation of the organic solvent, first with a gentle stream of nitrogen then by removing the residual solvent in a vacuum oven (Lab instruments IL, USA) over night.The dry lipid film was suspended in HEPES buffer (10 mM HEPES, 150 mM NaCl, pH 7.4) or, when indicated, in a CF solution (100 mM CF, 10 mM TES buffer, pH = 7.4), to reach a total lipid concentration of 2 or 5 mM.The lipid dispersion was subjected to 15 freeze-thawing cycles (freezing with liquid Nitrogen and thawing with a water bath at 50 °C) where after unilamellar liposomes were prepared by extruding the sample 31 times through a 100 nm pore size filter from Whatman plc (Kent, UK).Liposomes prepared in CF solution were stored for 24 h in room temperature to reach an equilibrium state before starting the experiment.

Determination of lipid content
In order to determine the amount of HEPC and DSPE-PEG (2000) in the samples the phosphorous content was analysed by the method described by Paraskova et al [21].Briefly, the samples were first calcinated at 550 °C for at least 5 h and the obtained ashes were dissolved in 4 mL of MilliQ water.Thereafter 1 mL of a freshly prepared solution consisting of seven parts reagent 1 and three parts reagent 2 were added.Reagent 1 was prepared by mixing the following reagents in the ratio 1:3:10 K(SbO)C 4 H 4 O 6 Á 0.5 H 2 O (2.75 mgÁmL À1 ):(NH 4 ) 6 Mo 7 O 24 Á4H 2 O (4% w/v):H 2 SO 4 (2.5 M).Reagent 2 was an L(+)-ascorbic acid solution (0.1 M).The absorbance at the wavelength 882 nm was measured after 15 min with an HP 8453 UV-Vis spectrometer from Agilent Technologies (Santa Clara, USA).The concentration of phosphorous was determined by use of a standard curve obtained from consecutive dilution of a phosphorous standard solution (0.65 mM) from Sigma Aldrich (St. Louis, USA).The total lipid content was calculated based on the assumption that the phosphorous-free components, cholesterol and gangliosides, contributed in accordance with the initial mixing ratios.

Dynamic light scattering (DLS)
DLS experiments were performed using an ALV/CGS-3 compact goniometer system and an ALV/LSE-5004 light scattering electronics and multiple tau digital correlator from ALV-GmbH (Langen, Germany).The LASER beam had a wavelength of 532 nm with vertically polarized light.All measurements were performed in triplicate and at room temperature with a run time of two minutes and at least two runs.Data acquisition and fitting was performed with the ALV-correlator software (version 3.0).The regularized fit model with the (g2(t)) DLS-exponential was used.

Assessment of carboxyfluorescein encapsulation
Liposomes prepared in CF solution were separated from unencapsulated dye using gel filtration column (PD-10) from GE-Healthcare (Uppsala, Sweden) equilibrated with 10 mM phosphate buffered saline (PBS, 150 mM NaCl, pH = 7.4).Thereafter the samples were diluted to 12 mM to ensure a direct proportionality between fluorescence readings and CF concentration.Subsequently the samples were transferred into a cuvette (Quartz SUPRASIL Ò , Hellma Analytics, Mühlheim, Germany).The liposomes were solubilised by addition of 50 mL of 200 mM Triton X-100 and the fluorescence signal was thereafter recorded with a Fluorolog Ò -3 from Horiba (Kyo ¯to, Japan) operating in the right-angle mode.The excitation and emission wavelengths were set to 495 nm and 520 nm, respectively.All measurements were carried out at 20 °C.

Cryo-transmission electron microscopy (cryo-TEM)
Specimens for cryo-TEM investigation were prepared at controlled temperature (25 °C) and humidity (>90%) within a custom-built environmental chamber.A small (~1 lL) drop of sample was placed onto a copper grid (300 mesh, Agar scientific) coated with a custom-made holey polymer (cellulose acetate butyrate) film, and excess liquid was blotted away with a filter paper.The sample was then vitrified by plunging the grid into liquid ethane held at a temperature just above its freezing point.Thereafter the grid was mounted in a Gatan CT3500 sample holder and transferred to a Zeiss TEM Libra 120 transmission electron microscope (Carl Zeiss AG, Oberkochen, Germany) for viewing.The sample was kept at a temperature below À160 °C during the transfer and viewing processes.The microscope was operating at 80 kV and the observations were made in zero-loss bright-field mode.Digital images were recorded under low dose conditions with a BioVision Pro-SM slow scan CCD camera from Proscan elektronische Systeme GmbH (Scheurig, Germany).

Statistical analysis
Data were analysed by unpaired t-test using Origin 2018 from the OriginLab Corporation (Northampton, USA) in order to test for significant differences (p < 0.05).

Cryo-TEM investigations to assess structural effects induced by inclusion of PEG-lipids and porcine gangliosides in the liposome preparations
We began our investigations by studying samples prepared in the absence of gangliosides but with different concentrations of PEG-lipid.As expected, micrographs obtained from samples containing 5 mol% DSPE-PEG(2000) display mainly liposomes (Fig. 2a).Noteworthy, a small number of flat, circular bilayer disks were observed to coexist with the liposomes.The amount of these discoidal structures, henceforth referred to as lipodisks, increases with increasing PEG-lipid concentration (Fig. 2b-e).As the population of lipodisks grow, the liposomes reduce in number.Thus, samples containing 15 mol% PEG-lipid are clearly dominated by lipodisks, and liposomes are only rarely observed in samples containing 20 mol% PEG-lipid.From the micrographs shown in Fig. 2 it is moreover evident that the average size of the lipodisks decreases with increasing PEG-lipid concentration, and that samples containing 40 mol% PEG-lipid include structures having a size similar to that observed for pure PEG-lipid micelles (Fig. 2f).The results displayed in Fig. 2 originate from cryo-TEM analyses of samples with total lipid concentration corresponding to 5 mM.Noteworthy, samples with 2 mM total lipid concentration displayed a very similar structural behaviour.
The micrographs presented in Fig. 3 show that the formation of lipodisks is promoted by inclusion of gangliosides in the lipid mixtures.Hence, a considerable amount of rather large lipodisks can be seen in samples containing 5 mol% PEG-lipid and 2.5 mol% gan-glioside (Fig. 3a).Upon increasing the ganglioside concentration to 5 and 10 mol% the population of disks grows at the expense of the liposomes (Fig. 3b, c).Although the disks are rather polydisperse in size it is apparent that their average diameter decreases with increasing ganglioside concentration.From a comparison of the micrographs shown in Fig. 2 and Fig. 3 it can be noted that liposomes formed in the presence of ganglioside appear on average larger than liposomes supplemented with PEG-lipid alone.
Cryo-TEM results shown in Fig. 4 disclose that gangliosides induce formation of discoidal structures also in the absence of PEG-lipids.As evident from Fig. 4a, samples composed of HEPC, cholesterol and 15 mol% ganglioside contain, apart from liposomes, a considerable fraction of lipodisks.Samples containing 40 mol% ganglioside are dominated by small disks and liposomes are only rarely observed (Fig. 4b).Sample composed of ganglioside alone display small micelles (Fig. 4c).

Particle size and size distributions as determined by DLS measurements
The ability of gangliosides to promote the formation of nonliposomal structures was further investigated by DLS measurements.Data collected for samples containing PEGylated liposomes (HEPC:Chol:DSPE-PEG(2000) 62:33:5) prepared in the absence of ganglioside suggest a relatively monodisperse population of particles with hydrodynamic radius 51 ± 6 nm.Samples supplemented with ganglioside display a considerably broader size distribution (Fig. 5a) and, as judged by number weighted data (Fig. 5b), the preparations contain a substantial fraction of particles with a hydrodynamic radius that is smaller than what is expected for PEGylated liposomes.Noteworthy, the DLS measurements indicate that these particles progressively decrease in size and become more numerous as the ganglioside concentration is increased from 2.5 to 10 mol%.
Fig. 5c and d show DLS data collected for samples devoid of ganglioside but supplemented with varying amounts of DSPE-PEG (2000).For samples containing more than 5 mol% PEG-lipid, the number weighted intensity distributions suggest a significant population of structures with a size smaller than that of PEGylated liposomes.The non-liposomal structures decrease in size and increase in number with increasing PEG-lipid concentration.Samples supplemented with 20 mol% PEG-lipid contain, as judged by the particle distribution, very few structures with a size compatible with liposomes.

Encapsulation efficiency of CF
Fluorescence measurements involving the hydrophilic probe CF were used to assess how changes in PEG-lipid and ganglioside concentrations affect the entrapped aqueous volume of the preparations (see Section 2.4 for details).As seen in Fig. 6, gangliosidefree preparations containing 5 mol% PEG-lipid display higher fluorescence intensity, indicative of higher entrapped volume, than ganglioside-free preparations including 15 mol% PEG-lipid.A decrease in fluorescence intensity is noted also upon increasing the ganglioside content from 5 to 10 mol% in preparations containing 5 mol% PEG-lipid.

Comparison of structural effects induced by ganglioside components GD1a and GM1
Data reported in the previous sections were retrieved from samples supplemented with total ganglioside extract from porcine brain, in which the components GM1, GD1a, GD1b and GT1b represent the vast majority of gangliosides [22].To complement these investigations, cryo-TEM analyses were carried out of samples containing 5 mol% PEG-lipid and 10 mol% of isolated ganglioside GM1 or GD1a (see Fig. 1 for structures).The micrographs shown in Fig. 7 confirm for both samples a coexistence between liposomes and bilayer disks.It is evident, however, that the ratio of disks to liposomes is considerably higher in samples complemented with GD1a than with GM1.

Cholesterol-free preparations
To assess the influence of the lipid phase-state on the propensity for disk formation, samples containing ganglioside, or PEGlipid, but lacking cholesterol were analysed by cryo-TEM.As shown in Fig. 8a, micrographs obtained from samples composed of HEPC and 20 mol% ganglioside (total extract) expose large, often bent or twisted, bilayer disks.A few liposomes, displaying the angular shape typical of liposomes in the gel-phase state [23], could also be detected in the micrographs.HEPC samples devoid of ganglioside but supplemented with 20 mol% PEG-lipid displayed a rather homogeneous population of small bilayer disks (Fig. 8b).
The micrographs revealed no liposomes in the latter samples.

Discussion
Due to the large polar headgroup, DSPE-PEG(2000) exhibit high positive spontaneous curvature and self-assembles in aqueous solution into globular micelles [24].When added to other lipids, micelle-forming PEG-lipids tend to drive the lipid assemblies into structures with more positive curvature.As a consequence, only a limited amount of PEG-lipid can be incorporated into the essentially flat lipid bilayer of a liposome.Above this bilayer saturation limit, which typically corresponds to less than 10 mol% PEG-lipid, non-liposomal structures begin to appear in the samples [14][15][16].The type and nature of the structures formed depend on several factors, including lipid composition and phase state.Under conditions where the lipids adapt a liquid disordered phase state, thread-like cylindrical micelles tend co-exist with the liposomes [16,25].Structures of a fundamentally different type and geometry appear, however, under other conditions.Thus, several studies have shown that planar, circular bilayer disks (frequently referred to as PEG-stabilized lipodisks) may form as an intermediate structure between liposomes and globular lipid/PEG-lipid mixed micelles [14][15][16]26].The PEG-lipids accumulate in this case primarily at the highly curved rim of the disks [27,28], and with  increasing PEG-lipid concentration the disks gradually shrink in size.This structural behaviour is well documented for systems based on saturated phosphatidylcholines in the gel-phase state [15,26], and has also been reported for saturated and unsaturated phospholipids in mixtures with 40 mol% cholesterol [14,16,29].
Results obtained in the current study show that major structural changes, as expected, occur upon addition of DSPE-PEG (2000) to mixtures composed of HEPC and 33 mol% cholesterol.The cryo-TEM analyses reveal that the PEG-lipid induces the formation of lipodisks, which shrink in size and become more numerous with increasing PEG-lipid concentration (Fig. 2).Since a small population of disks can be detected already in samples supplemented with 5 mol% PEG-lipid, it is likely that this concentration represents, or is close to, the bilayer saturation limit.It should in this context be mentioned that no disks but only liposomes were observed in samples supplemented with 2.5 mol% PEG-lipid (data not shown).
Gangliosides share important properties with PEG-lipids.Thus, they are composed of two long hydrocarbon chains connected to a bulky polar headgroup (Fig. 1), and self-assemble in dilute aqueous solution into micelles that, depending on the ganglioside subclass, adapt an ellipsoidal or close to spherical shape [30,31].Considering the similarities between gangliosides and PEG-lipids, it is plausible to assume that they are capable of bringing about similar structural changes when added to bilayer-forming lipids.In sup-port of this assumption, a very recent study report on the presence of discoidal lipid assemblies, presumably small lipodisks, in sonicated samples composed of 1,2-dioleoyl-sn-glycero-3-phosphocho line (DOPC) and 35 mol% GM1 [31].The cryo-TEM investigations carried out in the present study verify that gangliosides indeed are quite potent in inducing the formation of lipodisks in systems based on HEPC and 33 mol% cholesterol.As obvious from Fig. 7, GM1 is less effective than GD1a in promoting disk formation.This comes as no surprise, since GD1a due to its larger polar headgroup and higher spontaneous curvature [30] is expected to have a lower bilayer saturation limit than GM1.The fact that some liposomes and comparably large disks are still detected in samples containing 40 mol% Gd1a (Fig. 7b) indicates, however, that GD1a is less potent than DSPE-PEG(2000) when it comes to driving lipid assemblies into structures with high average curvature (compare Fig. 2e).Results from our cryo-TEM investigations of cholesterol-free HEPC samples give further support to this view.The small disks present in the PEG-lipid-supplemented sample (Fig. 8b) obviously have a total rim, or edge, area that is much larger than that of the larger disks observed in the ganglioside-containing sample (Fig. 8a).The observation that PEG-lipids have a stronger propensity than gangliosides to induce highly curved structures fits well with published data suggesting that Gd1a forms ellipsoidal micelles composed of about 225 monomers [30], whereas the micelles formed by DSPE-PEG(2000) are spherical in shape and built from around 75 monomers [24].
The formation of PEG-stabilized lipodisks is coupled to a partial component segregation and accumulation of the PEG-lipids at the disk rim [27,28].In the region of coexistence between liposomes and disks it is furthermore clear that there is an uneven distribution of the components also between the different types of lipid assemblies.Thus, as compared to the liposomes, the disks are enriched with PEG-lipids.It is possible, albeit not yet confirmed, that also the cholesterol content may vary depending on the type and size of the self-assembled structures.A similar uneven intraand inter-aggregate distribution of the molecular species can be expected also in the systems supplemented with gangliosides.In case of gangliosides extracted from natural sources, variations in the structure of the polar headgroups and/or the hydrophobic tails give rise to a large number of discrete molecular species and open up for the possible coexistence of a wide range of structures displaying different composition, size and shape.This may partly explain the structural diversity revealed by cryo-TEM and the broad peaks observed in the DLS measurements (Fig. 5).
Given that both gangliosides and PEG-lipids are capable of inducing disk formation, an additive effect can be anticipated upon their simultaneous inclusion in HEPC/cholesterol-based liposome preparations.Hence, it can be hypothesized that very little, if any, ganglioside can be added to samples containing 5 mol% PEG-lipid without inducing the formation of disks.Data from our cryo-TEM investigations verify this assumption and show the coexistence of liposomes and disks in samples supplemented with 2.5 mol% or more ganglioside (Fig. 3).Noteworthy, as indicated by cryo-TEM and verified by DLS (Fig. 5b), the population of disks grows at the expense of the liposomes with increasing ganglioside concentration.
We have at present no explanation for the increased liposome size observed in the presence of gangliosides.Since all preparations were repeatedly extruded through filter membranes with pore size 100 nm, it appears, however, that the lipid membranes either become softer in the presence of gangliosides, thus allowing more than 100 nm liposomes to deform and slip through the pores, or that the large liposomes form subsequent to the extrusion process.In support of the former hypothesis, a clear fluidizing effect of GM1 on POPC membranes has been reported in studies by Fricke et al. [32].Results from molecular dynamics simulations, as well as experimental studies, point however towards GM1 having the opposite effect in DOPC [31] and DPPC membranes [19,[33][34].Formation of large liposomes in already extruded samples may possibly occur through the fusion of unstable/insufficiently stabilized bilayer disks into larger disks, which eventually close on themselves and form liposomes.A similar process has been shown to be responsible for the growth of small sonicated liposomes in response to the addition of conventional micelle forming surfactants [35].
Apart from contributing to the fundamental knowledge concerning phase and structural behaviour in dilute aqueous lipid/ PEG-lipid/ganglioside systems, the results and insights gained from the present study have implications for the use of gangliosidemodified PEGylated liposomes in drug delivery applications.In particular, the possible presence of non-liposomal structures in the liposomal preparations needs to be acknowledged and considered.In case of formulations designed with the intention to encapsulate drugs in the aqueous core of liposomes, it is evident that the presence of lipid disks may have a detrimental impact on the encapsulation efficiency and loading capacity of the formulations.Structurally polydisperse drug formulations are problematic, however, also from a more general perspective, not least since their in vivo use may lead to unpredicted and unwanted effects on biodistribution and pharmacokinetics.
Finally, the findings of the present study raise some important questions concerning the origin of and mechanisms behind the attenuated anti-PEG immune response observed upon inclusion of gangliosides in PEGylated HEPC/cholesterol liposomes [12,13].Mima et al. report data from studies in mice models suggesting an attenuating effect of gangliosides on the occurrence of the ABC phenomenon, as well as on the production of anti-PEG IgM [12].A clear and positive correlation was observed between the strength of the immunosuppressive effect and the ganglioside content in the liposomal preparations.Thus, a weaker anti-PEG IgM response was noted for liposomes modified with 10 than 5 mol% ganglioside.Experiments designed to distinguish between different ganglioside components showed furthermore that liposomal preparations supplemented with GD1a tended to supress the anti-PEG IgM production to a considerably higher degree than those supplemented with GM1.Studies in rat models performed by Zhang et al. suggest, moreover, that the anti-PEG IgM production elicited by GM1-supplemented PEGylated liposomes gradually decreases when the ganglioside content is increased in the range between 2.5 and 15 mol% [13].The attenuated anti-PEG IgM response observed for the ganglioside-modified liposomes has been suggested to be connected to the gangliosides capacity as Siglec (sialic acid-binding immunoglobulin-type lectin) ligands.More specifically, the co-presentation of PEG and gangliosides on the liposome surface has been proposed to lead to immunological tolerance against PEG via a mechanism based on the binding of ganglioside to B cell inhibitory co-receptors [12].Notably, the ganglioside concentrations needed to bring about the mitigating effects on the immune response observed in the abovementioned studies all fall within the range where liposomes according to our investigations are likely to co-exist with lipid disks.Further, alterations that according to our studies lead to lower liposome/disk ratios appear to be connected to weaker anti-PEG IgM responses.In light of these findings it appears justified to question whether the attenuated immune response is mediated by the gangliosides per se, or rather due to the fact that a significant part of the liposomes are replaced by a population of smaller, discoidal and heavily PEGylated structures.It is in this context noteworthy that a previous study by Ishida et al. indicate that the induction of the ABC phenomenon can be significantly suppressed by increasing the PEG-lipid content in HEPC/choles-terol liposomes from 5 to 15 mol% [11].Since, as judged from our investigations, preparations supplemented with 15 mol% PEG-lipid tend to contain comparatively few liposomes and instead be dominated by disks (Fig. 2c), it is tempting to conclude that PEG-stabilized lipodisks are less immunogenic than PEGylated liposomes.Interestingly, data put forward by Kaminskas et al. suggest that soy PC/PEG-lipid mixed micelles provoke less production of anti-PEG IgM than PEGylated soy PC liposomes [30].Results of the same study moreover indicate that the PEGylated micelles are not subject to recognition by anti-PEG IgM, and therefore not susceptible to ABC.The detailed reasons behind the differences in immunological behaviour observed for PEGylated micelles and liposomes are at present not fully understood.It is worth noting, however, that certain properties typical of PEGylated mixed micelles, such as a comparatively small size and high PEGsurface density, also apply to PEGylated lipid disks.The fact that a structural transformation from liposomes to micelles, or disks, leads to a higher number of PEGylated particles in the preparations could potentially also contribute to the observed differences in immunological behaviour.

Conclusions
Considerable attention has over the years been focused on the structural effects caused by inclusion of micelle-forming surfactants/amphiphiles in liposomal preparations.A remarkable discovery made in previous investigations is that PEG-lipids possess an unusual ability to induce the formation of comparatively large, long-lived discoidal bilayer structures [14][15][16].Data reported in the present study reveal that gangliosides, similar to PEG-lipids, promote formation of bilayer disks when mixed with lipids in the gel-or liquid ordered phase state.Importantly, results from comprehensive cryo-TEM investigations show that PEG-lipids and gangliosides work in an additive manner.Hence, very little, if any, ganglioside can be added to HSPC:cholesterol liposomes supplemented with 5 mol% DSPE-PEG(2000) without disk formation.The presence of non-liposomal structures can have important impact on the efficacy and safety of liposomal drug formulations, and may also complicate interpretation and evaluation of data on, e.g., immunological effects.In light of the structural behaviour revealed in the current report it is clear that further investigations, involving careful adjustment of PEG-lipid/ganglioside ratios to avoid disks formation, are needed to confirm or refute the proposed ability of gangliosides to attenuate the immunogenicity of PEGylated liposomes [12,13].Taken together with data from previous immunological investigations [11,36], findings of the present study suggest that PEG-stabilized lipodisks may provoke less anti-PEG immune response than PEGylated liposomes.Further studies focused on the immunological response elicited by PEGstabilized lipodisks are warranted, not least since accumulating evidence point towards lipodisks as promising and versatile drug nanocarriers [37][38][39][40].