Biochemical Society Transactions

Cellular Delivery of Therapeutic Macromolecules

Gene delivery by cationic lipids: in and out of an endosome

D. Hoekstra, J. Rejman, L. Wasungu, F. Shi, I. Zuhorn


Cationic lipids are exploited as vectors (‘lipoplexes’) for delivering nucleic acids, including genes, into cells for both therapeutic and cell biological purposes. However, to meet therapeutic requirements, their efficacy needs major improvement, and better defining the mechanism of entry in relation to eventual transfection efficiency could be part of such a strategy. Endocytosis is the major pathway of entry, but the relative contribution of distinct endocytic pathways, including clathrin- and caveolae-mediated endocytosis and/or macropinocytosis is as yet poorly defined. Escape of DNA/RNA from endosomal compartments is thought to represent a major obstacle. Evidence is accumulating that non-lamellar phase changes of the lipoplexes, facilitated by intracellular lipids, which allow DNA to dissociate from the vector and destabilize endosomal membranes, are instrumental in plasmid translocation into the cytosol, a prerequisite for nuclear delivery. To further clarify molecular mechanisms and to appreciate and overcome intracellular hurdles in lipoplex-mediated gene delivery, quantification of distinct steps in overall transfection and proper model systems are required.

  • cationic lipid
  • endocytosis
  • gene delivery
  • hexagonal phase
  • lipoplex
  • transfection


The ability of introducing and expressing genes into cells (‘transfection’) has boosted developments in both therapeutic and cell biological exploitation of this technology [15]. Therapeutic applications focus particularly on curing diseases of genetic origin by introducing the defective gene(s) in malfunctioning target cells, whereas cell biology benefits from the possibility of investigating biological properties of expressed protein(s). Since DNA molecules are internalized poorly by cells, a crucial aspect of this technology entails the efficiency of DNA delivery into cells, and a variety of carrier systems are currently being exploited, including virus-derived and non-viral vehicles, such as synthetically prepared cationic lipids (Figure 1) and polymers.

Figure 1 Examples of cationic lipids for transfection

DOTAP, N-[1-(2,3-dioleyl)propyl]-N,N,N-trimethylammonium chloride; SAINT-2, N-methyl-4-(dioleyl)methylpyridinium.

Although non-viral-system-mediated transfection is still inferior compared with viral systems, safety hazards encountered with the latter, including immunological and oncogenic side effects, justify research to improve non-viral transfection efficiency. Part of the approach relies on the synthesis of novel cationic lipids and polymers. However, given current insight into how cationic lipid–DNA (‘lipoplexes’) or polymer–DNA complexes (‘polyplexes’) transfect cells, it seems equally useful to further explore mechanisms of gene delivery to better appreciate and overcome potential hurdles in overall transfection, which range from complex assembly to intracellular trafficking and transcription efficiency. For example, such studies may prompt the rational development of lipoplexes, containing properly complexed and transcription-competent DNA, targeted into those intracellular pathways that lead to the highest transfection efficiency. Also, dissociation of DNA from the carrier, as a vital step in nuclear delivery efficiency, is essential, and recent work suggests that a precise quantification of DNA delivery, including dissociation from the carrier and eventual transcription efficiency, is necessary for the proper appreciation of barriers in gene delivery [6]. Therefore in vitro experiments using sophisticated technology and model systems are essential, but are not necessarily sufficient to warrant in vivo success [1,7]. Thus additional barriers exist in vivo that are insufficiently matched by in vitro simulation, particularly the role of serum, and, to a lesser degree, the need for device modulation to avoid preliminary clearance of DNA vectors by macrophages and to achieve targeting to desired tissue and/or cells. To avoid non-specific binding in vivo of the usually net-positively charged lipoplexes, PEG [poly(ethylene glycol)]-derivatized lipids are often included [810], although the so-called stealth properties conveyed in this manner may require a proper adjustment (i.e. a decrease) of the net positive charge [9]. Here, we will briefly discuss some recent developments in advancing our understanding of the mechanisms of lipoplex-mediated transfection, with an emphasis on lipoplex internalization and trafficking, leading to expulsion of DNA into the cytosol as a prerequisite for nuclear gene delivery. For more detailed recent overviews, see [3,5,1013].

Assembly and phase behaviour of lipoplexes

Liposomes prepared from cationic amphiphiles interact with polyanionic DNA (or RNA), thereby spontaneously forming lipoplexes. During this event, the DNA structure undergoes significant changes [14,15], probably in a cationic lipid- and helper lipid-dependent manner [10,1618], and double-stranded DNA can become partially ‘melted’ [15]. This feature should be taken into account particularly in the context of improvement of lipoplex-mediated transfection given the (very) poor transcription efficiency of genes delivered by lipoplexes, compared with viral systems [6]. For cellular internalization, lipoplex size could be an important parameter, and the relative colloidal stability of lipoplexes appears to be in part dictated by its polymorphic properties, complexes preferring a lamellar phase being less susceptible to growth in size than those displaying an inverted hexagonal HII phase [10,19]. Since the charge ratio (+/−) of the assembled complexes usually exceeds a value of 1, the net positive charge promotes highly efficient binding to cells via electrostatic interactions with negative charges at the cell surface. Often the transfection efficiency is enhanced if the lipoplexes also contain a so-called helper lipid, such as zwitterionic DOPE (1,2-dioleoyl-sn-glycero-3-phosphoethanolamine) [18,20,21]. The presence of the helper lipid affects the association of the DNA with the lipoplexes [10], but particularly modifies aggregate morphology of the complex by promoting the hexagonal HII morphology under physiologically relevant conditions. Indeed, several studies in search of a correlation between the structural properties of lipoplexes and their transfection efficiency, employing different cationic lipids, have revealed that lipoplexes that adopt the HII phase strongly facilitate intracellular release of DNA and display the highest transfection efficiency [2224]. The geometry of the cationic lipid co-determines the ability of the lipoplex to adopt such a non-lamellar inverted hexagonal-phase structure, dictated by a packing parameter which predicts that, when the area of the hydrocarbon tails is much larger than that of the headgroup, the cationic lipid tends to adopt a bilayer-destabilizing hexagonal HII phase. More recently, other non-lamellar phases have been described [19,25,26], which, by virtue of bilayer-destabilizing features, may similarly operate in the cytosolic release of DNA from intracellular compartments, following lipoplex internalization.

Internalization and intracellular trafficking

There is convincing evidence that endocytosis represents the major pathway of lipoplex entry into cells that leads to productive gene expression, rather than that genes are translocated into the cytosol at the level of the plasma membrane (Figure 2; [2,4,13]). However, various endocytic pathways operate in eukaryotic cells, i.e. clathrin-dependent and -independent pathways, the latter including phagocytosis, macropinocytosis and caveolae-mediated internalization. The relative contribution of each pathway in lipoplex internalization has been poorly defined to date, although involvement of the clathrin-mediated pathway has been firmly established, while evidence is emerging that entry may also occur via macropinocytosis [27].

Figure 2 Schematic representation of intracellular trafficking pathways for non-viral gene delivery vehicles in transfection

Transfection-productive entry of lipoplexes and polyplexes into cells can be accomplished via clathrin-mediated endocytosis (a), caveolae-mediated endocytosis (b) and macropinocytosis (c). None of these pathways is mutually exclusive. Polyplexes have been shown to enter cells via all three pathways, although the mechanism involved in plasmid release and nuclear targeting (N) from caveosomes (Cav) and macropinosomes (MP) remains to be determined. Also recycling of polyplexes via recycling endosomes (RE) has been reported [47]. For lipoplexes, neither caveolae-mediated endocytosis nor recycling has been demonstrated to date. A pore-like mechanism possibly mediates release of DNA from early endosomes (EE) in the case of lipoplexes, although release from late endosomes (LE) cannot be excluded. Because of plasmid degradation, significant entry via lysosomes (LY) is highly unlikely.

A correlation between clathrin-mediated endocytosis and lipoplex-mediated transfection is supported convincingly by several pieces of evidence, including the use of inhibitors of endocytosis, electron microscopy, co-localization with pathway-specific markers and, particularly, mutants, defective in clathrin-mediated endocytosis, that fail to internalize lipoplexes and concomitantly show a proportional decrease in transfection efficiency [2,2830]. A combined approach is essential, since the use of inhibitors alone may give rise to an ambiguous outcome because of toxic side effects. However, complete inhibition is often not observed, implying that different pathways may be operating simultaneously or may take over if one pathway is blocked. Insight into the exact contribution of a pathway and parameters governing the entry of lipoplexes into such a pathway is relevant for establishing optimal transfection protocols. Accordingly, entry along a distinct pathway should be linked directly to gene expression capacity.

Size appears to be an important parameter that (co-)determines the mechanism of particle entry into cells. Since the size of lipoplexes (from 100 nm to several micrometers) is difficult to control, these studies were carried out with latex beads of defined size showing that particles with a diameter of 200 nm or less entered cells almost exclusively via the clathrin-coated pathway [31]. Unexpectedly, a rapidly increasing contribution of caveolae-mediated endocytosis was seen for particles with a larger diameter, and 500 nm beads almost exclusively entered via this pathway. Rather than by size-dependent differences governing entry in a distinct pathway, a correlation between lipoplex size and transfection efficiency has been discussed, so far largely in terms of kinetics of (intra)cellular processing [10,32,33]. For lipoplexes with a size of approx. 200 nm, clathrin-mediated endocytosis appears to be the major route of entry [29,30]. Certain polyplexes are internalized via both clathrin-mediated endocytosis and caveolae [30,3436]. Intriguingly, in one of these studies, it was claimed that only caveolae-mediated internalization caused productive transfection [30]. Consistent with a particle-size-dependent entry mechanism, larger polyplexes may have preferred a cellular entry route via caveolae [37]. Whether caveolae-mediated internalization of large lipoplexes (i.e. beyond 300 nm) may similarly occur is unknown. However, this appears to be feasible, since several bacteria enter cells via caveolae [38]. Most importantly, caveolae-mediated trafficking seems to proceed along a non-acidic and non-digestive pathway, thereby becoming particularly attractive for lipoplex targeting, provided that effective escape of plasmid from caveosomes occurs. In this manner, lysosomal digestion would be avoided, as this is thought to contribute negatively to transfection (Figure 2).

Do lipoplexes ‘exploit’, analogous to viruses, distinct cellular receptors for entry into cells, following initial electrostatic interaction? Recently, we observed that the binding affinity of lipoplexes to the apical and basolateral cell-surface domains in (polarized) epithelial cells differs, and that distinct receptors, i.e. the β1 integrin receptor present exclusively at the basolateral surface, mediate clathrin-mediated lipoplex internalization in these and several other cell types (I. Zuhorn, G. Robillard and D. Hoekstra, unpublished work). The parameters driving such an interaction between (ligand-devoid) lipoplexes and these cell-surface receptors remain to be determined. Surface-bound lipoplexes seem to laterally redistribute and/or cluster before internalization [9]. Whether interaction with endocytic receptors capable of internalizing lipoplexes is secondary to an initial random electrostatic interaction is unknown, but since antibodies against the β1 integrin receptor block transfection, but not lipoplex binding, such a scenario appears to be possible. Accordingly, these data suggest that multiple binding entities may exist on the cell surface, but that a pool of (specific) receptors, such as β1 integrins, can play a role in the (specific) internalization of lipoplexes, leading to productive transfection.

Out of an endosome

To avoid degradation in lysosomes, the plasmid has to escape into the cytosol before reaching this organelle (Figure 2). Thus prompt release from an endosomal compartment presumably constitutes one of the critical steps in determining the efficiency of transfection. Indeed, disruption of endosomal integrity by osmotic shock strongly promotes nuclear delivery of plasmid and subsequent gene expression [21,33]. However, little insight is available into how endosomal membrane destabilization and the concomitant dissociation and release of plasmids proceed in situ. Whether DNA dissociates before or concomitant with endosomal membrane perturbation is unclear, as well as whether DNA transfers across a perturbed membrane or requires complete lysis of the endosome. The latter has been proposed for polyplex-mediated delivery (‘proton-sponge’ mechanism [39]), which involves osmotic lysis of the compartment rather than destabilization upon tight interactions between polyplex and endosomal membrane. In contrast, the fact that lipoplexes that preferentially adopt non-lamellar phases, including the hexagonal HII [2224] and HI [19,25] or a cubic phase [26], all strongly promote transfection is consistent with a mechanism involving lipoplex-induced endosomal membrane destabilization. Indeed, inclusion of PEGylated lipid derivatives or phospholipids such as phosphatidylcholine in lipoplex formulations, both stabilizing membrane bilayer structure, inhibits nucleic acid dissociation, in spite of their efficient cellular internalization [10,40,41]. Moreover, interaction of lipoplexes with liposomes that contain acidic phospholipids such as PS [1,2-acyl-sn-glycero-3(phospho-L-serine)] facilitates the lamellar to non-lamellar hexagonal-phase transition of lipoplexes [12,16,24,42], concomitantly causing the dissociation of nucleic acid. These observations are in line with the proposal that PS, localized within the endosomal membrane, but possibly translocated into lipoplexes, attached to the inner leaflet of the endosomal membrane (Figure 2), is instrumental in the transfection mechanism [43]. On the other hand, cationic lipids may translocate via the interacting non-lamellar intermediates (perhaps relying in part on fusion) into the endosomal membrane, and a local increase may give rise to membrane destabilization and pore formation of transient stability, determined by the rate of lateral diffusion of surrounding membrane lipids [12,44]. In this manner, a plasmid may acquire access into the cytosol. To characterize further the molecular mechanisms underlying lipoplex-induced endosmal membrane destabilization, a model system, simulating such an event by studying the interaction of lipoplexes with giant unilamellar liposomes, allowing direct microscopic visualization of the interaction and biophysical characterization by means of laser-scanning imaging, may prove to be highly valuable [45]. In this study, intraliposomal delivery of polynucleotides was observed without lysis, suggesting that DNA translocation across a (endosomal) membrane via a cationic lipid-induced pore-like mechanism may be possible [12,44]. Consistent with this notion, using doubly labelled lipoplexes, fluorescently tagged nucleic acids accumulate in the nucleus, whereas similarly probed lipid markers of the carrier remain associated with endo/lyso-somal compartments [46].

More recent work, in which a comparison was made between adenovirus compared with lipoplex-mediated transfection, calls for bottlenecks in gene delivery, other than the efficiency of complex internalization, endosomal release and nuclear delivery [6]. Sophisticated quantification of gene delivery by these vectors revealed that, even though a comparable level of escape from endosomes and transfer of genes into the nucleus may occur, the lipoplex-delivered reporter gene required three orders of magnitude more intracellular gene copies to reach a similar level of expression than observed for the virus-delivered gene. Not only do these experiments emphasize the need for quantification in order to properly appreciate the relative significance of (intra)cellular barriers in gene delivery, but they also emphasize the need for lipoplex formulations that improve gene transcription efficiency itself.

Taken together, although considerable progress has been made in recent years to define barriers in lipoplex-mediated transfection, numerous details are still incomplete. Endocytosis represents a major pathway of entry of lipoplexes, but more insight is needed in the relative contribution of a distinct pathway for lipoplex processing, in terms of effectiveness of internalization, efficiency of gene escape from endosomes, macropinosomes and/or caveosomes and eventual transfection efficiency. Technology for properly quantifying these distinct pathways as well as sophisticated simulation of these events is indispensable, and, as described above, options are within reach. Appreciating parameters that govern transfection by means of polyplexes and viral systems will be beneficial to further improve lipoplex-mediated transfection. Moreover, as such, this kind of work will be equally insightful into the cell biology of intracellular trafficking pathways.


  • Cellular Delivery of Therapeutic Macromolecules: Biochemical Society Focused Meeting in association with the Royal Society of Chemistry, the Royal Pharmaceutical Society of Great Britain and the Academy of Pharmaceutical Sciences held at Cardiff University, U.K., 29–31 August 2006. Organized and Edited by S. Akhtar (Cardiff, U.K.), M. Gait (MRC - LMB, U.K.), M. Gumbleton (Cardiff, U.K.) and A. Jones (Cardiff, U.K.).

Abbreviations: DOPE, 1,2-dioleoyl-sn-glycero-3-phosphoethanolamine; PEG, poly(ethylene glycol); PS, 1,2-acyl-sn-glycero-3(phospho-L-serine)


View Abstract