Endocytosis of Nanomedicines

4.1 Poly(ethylene glycol)-Polylactide nanoparticles

The biodegradable nanoparticles made of D,L-polylactide (PLA) and poly(ethylene glycol-co-lactide) (PEG-co-PLA) blends are widely explored for drug delivery [41–42]. Due to the presence of the partially hydrolyzed PLA the surface of these nanoparticles is usually negatively charged but can be altered by cationic surfactants such as stearylamine. A recent study, examined the effect of such modifications on the uptake of the nanoparticles (around 100 nm) in fibroblast and epithelial cells [41–42]. In both cell types the cationic particles were more readily accumulated in cells compared to anionic particles. However, the pathways of their entry depended on the cell type. In the polarized MDCK epithelial cells, which are normally devoid of caveolae on the apical surface [43–44], the nanoparticles used CME independently of their charge. In contrast, in non-polarized HeLa cervical cancer cells the anionic particles used multiple pathways (CME and caveolae-mediated), while cationic particles appeared to be restricted to CME as well as to macropinocytosis. Interestingly, the nanoparticles charge also determined their intracellular fate. In MDCK cells the cationic particles was routed for transcytosis, while the anionic particles reached the lysosomes [41–42]. This example clearly shows that the entry and subsequent trafficking of the nanoparticles can depend both on the cell type and the particle charge.

4.2 Poly(lactic-co-glycolic acid) (PLGA) nanoparticles

A close example is nanoparticles made of PLGA, a biodegradable synthetic material approved by the United State Food and Drug Administration (FDA) for therapeutic use [45]. The cellular transport of such nanoparticles was also shown to be both cell type and surface charge dependent [23]. In this case the nanoparticles were relatively large (ca. 300 nm) and heterogeneous (polydispersity index 0.2), which obviously complicated interpretations of their cellular uptake. Nevertheless, the cell dependence of the uptake was shown using basic PLGA nanoparticles, which are negatively charged. In the vascular smooth muscle cells (VSMCs) these particles entered predominantly through the CME, however, in rat corneal epithelial cells they utilized the clathrin- and caveolae-independent pathways [46–47]. Furthermore, in VSMCs 85% of these particles was recycled to the cell surface from the early endosomes [46–47] . The remaining 15% were reported to escape the endosomes and reach the cytosol. To increase the endosomal escape the authors subsequently modified the nanoparticles surface with a cationic polymer, poly(L-lysine) (PLL) [23]. The adhesion of nanoparticles at the cell plasma membrane was quantified by AFM in cancer cells. The PLL modification led to nearly five-fold increase in the interaction force. The modified particles were also more rapidly internalized via CME, which allowed considerably increasing the cellular delivery of a model protein encapsulated in such nanoparticles [23].

4.3 Silica-based nanomaterials (SNTs)

CME was observed for “template synthesized” silica nanotubes (SNTs) [48]. Such SNTs provide unique structural features including inner voids for loading bioactive compounds, end functionalization to control compounds release, and inner and outer surfaces that can be differentially functionalized for targeting and biocompatibility. The cationic SNTs (50 nm in diameter and 200 nm long) functionalized with positively charged aminosilane group at the outer surface were shown to internalize via CME and reach lysosomes in cancer cells [48]. In another study mesoporous silica-based nanoparticles (ca. 110 nm) were shown internalize into mesenchymal stem cells and fibroblasts via a CME but not through caveolae [49].

4.4 Chitosan nanoparticles

The cationic chitosan nanoparticles were shown to utilize CME for entry in respiratory epithelium (A549) and intestinal epithelium (Caco-2) cells [50–51]. The nanoparticles used in this study were large (430 nm) and highly polydisperse (polydispersity index 0.5), which means that large particles co-existed with small ones, capable of entry into small vesicles. The selectivity of this formulation towards CME suggests that in this particular case the chemical composition of the material was more important than its size in defining the CME entry mechanism. However, the self-assembled cationic nanoparticles of hydrophobically modified chitosan (360 nm, zeta potential ca. +22mV) are using multiple pathways for cellular entry including CME, caveolae and macropinocytosis [52]. These nanomaterials possibly interacted with the cell membranes through the 5-β-cholanic acid residues used for hydrophobization of chitosan, which altered their trafficking.

4.5 Surface modified nanoparticles that target CME

A special case is nanomaterials, which are modified to display surface ligands that utilize CME. Such ligands can include, for example, mannose-6-phosphate, Tf, nicotinic acid, etc [53]. Notably, the cellular trafficking of such modified materials may differ from the trafficking of the ligand alone [54]. For example, quantum dots (QDs) (ca. 50 nm) modified with Tf utilized CME as the initial stage of entry, however, in contrast to Tf alone they were not routed to lysosomes and not recycled to the cell surface. Instead these modified QDs resided in the perinuclear endosomes. In another example, QDs (30 nm) modified with Shiga Toxin were entering cells, but unlike the toxin did not accumulate in the Golgi and were routed to the endosomes. Such effect was likely to be due to the ability of the QDs to interfere with the cellular trafficking since they were also shown to alter the transport of unconjugated ligands (Shiga toxin, Ricin B) [54]. Therefore, in attempts to target nanomaterials though specific ligands and cellular pathways one should take into account a possibility that such nanomaterials can alter the normal pathways employed by these ligands. Further examples of those effects are discussed below for unimers and micelles of Pluronic^® block copolymers [55].

5.1 Polymeric micelles with cross-linked anionic core

Core-cross linked polymeric micelles (cl-micelles) of poly(ethylene oxide)-b-poly( methacrylic acid) (PEO-b-PMA) copolymer are unique hydrogel-like nanomaterials, which have swollen cores of a cross-linked PMA network surrounded by a PEO shell [63]. At extracellular pH 7.4 they have strong negative charge (zeta potential −18 mV), which decreases at pH 5.0 (−7 mV), while the cl-micelles contract from ca. 150–160 nm to ca. 110 nm. These cl-micelles display selective entry in cancer cells but do not enter normal epithelial cells due to their ability to target differences in endocytosis mechanisms between these cells (Fig. 6) [26]. The internalization of cl-micelles in the cancer cells proceeds predominately through caveolae-mediated endocytosis. In confluent normal epithelial cells this endocytosis route is absent at the apical side [44] and the cl-micelles sequester in tight junction (TJ) regions of the cell membrane without entering the cells. Furthermore, contrary to conventional wisdom in cancer cells the cl-micelles bypass the early endosomes and reach the lysosomes within 30 min. where they release drug in a pH dependent fashion. Hence, micelles loaded with a cytotoxic drug are toxic to cancer cells but not to normal epithelial cells, where they do not enter [26].

5.2 DOXIL^®

DOXIL^® is a slightly negatively charged PEGylated liposomes (86 nm), which encapsulate doxorubicin hydrochloride and are used to treat patients with metastatic ovarian cancer. Surprisingly, little was known until recently about trafficking of DOXIL^® nanoparticles into cells. Our study presents evidence that these nanoparticles utilize caveolae-mediated endocytosis to internalize into epithelial cancer cells [26]. Similarly to the cl-micelles, after the entry through caveolae the DOXIL^® nanoparticles are accumulated in lysosomes, where the drug is apparently released [26].

5.3 Polysiloxane nanoparticles

Amphiphilic self-assembled nanoparticles (ca. 100 nm) of poly(3-aminopropyl)siolxane (PAPS) modified with stearic acid residues and galactose were shown to selectively target caveolae in human aortic endothelial cells [64]. Even more interestingly, these nanoparticles can activate eNOS cell signaling and promote NO formation. This strongly suggests that the nanoparticles disrupt eNOS binding to the caveolin-1 scaffolding domains and provides evidence that synthetic nanomaterials can activate signal transduction leading to pharmacological effects due to interactions with specific micro-domains at the cell membrane.

5.4 QDs

The effects of surface modifications on cellular entry were recently examined using CdSe/ZnS core-shell ellipsoid QDs (diameters 6 nm, 12 nm), which were modified with PEG, PEG-amine or poly(acrylic acid) [65]. The resulting QDs-based nanomaterials differed both in surface charge and size, being neutral (45 nm), positive (20 nm) and negative (18 nm), respectively. Of the three materials the negatively charged particles displayed the greatest uptake in skin cells. Moreover, the negatively charged particles entered the cells primarily through caveolae, while the neutral and positive particles did not appear to use this pathway [65].

5.5 Abraxane^®

Aside from charge the specific interactions with the receptor located in caveolae may define the cellular entry of nanomaterials. This is a case of Abraxane^®, a nanoparticle albumin-bound (nab) form of paclitaxel (130 nm) approved by the FDA for metastatic or relapsed breast cancer [66]. Abraxane^® nanoparticles take advantage of caveolae-mediated transcytosis for efficient delivery of the drug to the tumor sites [67]. They bind to gp60, the albumin receptor present in caveolae of endothelial cells, and transport across the vascular walls to the tumor interstitial spaces [68]. After entering the interstitial spaces the Abraxane^® nanoparticles are captured by SPARC (secreted protein, acidic and rich in cysteine) that is selectively secreted by the tumors [69–70]. The SPARC-nanoparticles complexes are taken up in tumor cells resulting in selective tumor cytotoxicity.

5.6 Surface-modified nanoparticles that target caveolae

As the significance of caveolae for the delivery of nanomaterials has emerged the attempts have been made to modify the surface of these nanomaterials to target proteins displayed in caveolae. One such target was identified by Oh et al. by in vivo quantitative proteomics of caveolae isolated from tumor lung endothelium [22]. They determined that these caveolae are enriched with aminopeptidase P (APP). Furthermore, the monoclonal antibody to APP was shown to be transported in vivo within seconds across endothelium to the lung tissue. Hence, the colloidal gold nanoparticles modified with antibody to APP were found to concentrate in caveolae of lung microvasculature [22].

Another well known targeting molecule is cyclic RGD peptide, which binds with α_vβ₃ integrin receptor located in caveolae. Recently, Oba et al. used thiolated c(RGDfK)-PEG-b-PLL copolymer for cellular delivery of the pDNA [71]. This copolymer binds with the DNA molecules through its cationic PLL chains resulting in core-shell polyion complex nanoparticles having core of electrostatically coupled PPL/DNA and a shell of PEG decorated with the cyclic RGD peptide. The core of such nanoparticles was additionally stabilized by cross-linking PLL aminogroups. They were shown to enter cells through caveolae and localize in the perinuclear regions in HeLa cells. The increased entry of the RGD-modified complexes compared to the non-modified analogs was accompanied by a superior DNA transfection efficiency in these cells.

8.1 PRINT micro-and nanoparticles

These particles are designed through a top-down lithographic fabrication method called PRINT (Particle Replication In Non-wetting Templates) [80]. PRINT technology results in the formation of nearly monodisperse nanoparticles that have precisely controlled shape, size and charge. In an elegant study Gratton et al., evaluated internalization pathways of three different series of micro- and nano-particles made from the cross-linked PEG-based hydrogels [81]. All materials were obtained by UV-copolymerization of the same mixture of monomers in different PRINT molds and therefore are believed to have the same chemical composition, but different sizes and shapes. The materials were positively charged and had relatively close zeta potentials (ranging from +21mV to +41mV). They included cubic microparticles (2 µm, 3 µm, and 5 µm), cylindrical microparticles (0.5 × 1 µm, 1 × 1 µm) and cylindrical nanoparticles (200 × 200 nm; 100 × 300 nm; 150 × 450 nm). Surprisingly, all these different particles including the largest ones were shown to internalize in HeLa cells (Fig. 7). However, the nanoparticles seemed to enter cells more rapidly than the microparticles. Furthermore, the shape of the nanoparticles also mattered as the extended cylinders (150 × 450 nm) entered faster than the equal cylinders of nearly same volume (200 × 200 nm), or cylinders of lower volume (100 × 300 nm). All nanoparticles appeared to employ multiple pathways of entry including CME, caveolae-mediated endocytosis and to a lesser extent macropinocytosis. It also appeared that CME and caveolae were more prominent for the entry of nanoparticles compared to the microparticles. Interestingly, when the charge of these materials was inverted to negative (−34mV) by acylating their amino groups their entry became negligible.

8.2 Pluronic^® block copolymers

Another interesting example is the cellular entry of amphiphilic triblock copolymer of poly(ethylene oxide) (PEO) and poly(propylene oxide) (PPO), Pluronic^® P85 (P85) [82]. Below the critical micelle concentration (CMC) this copolymer exists as single molecule chains (unimers) and above CMC forms 14.6 nm aggregated micelles with a hydrophobic PPO core and hydrophilic PEO shell. Both the copolymer and micelles are electrostatically neutral. This copolymer was shown to utilize multiple pathways to gain entry in mammalian cells depending on its aggregation state (Fig. 8A) [55]. In particular, P85 unimers internalize predominantly through caveolae-mediated endocytosis, however, in cells devoid of caveolae the copolymer exploits caveolae- and clathrin-independent endocytosis. Interestingly the unimers were shown to bypass early endosomes/lysosomes, transport to endoplasmic reticulum, and eventually reach mitochondria [21]. Furthermore, the copolymer succeeds in entering nearly impenetrable cells, such as primary brain endothelial cells and primary neurons where the copolymer moves anterograde from cell body to axons/dendrites (Fig. 8B). On the contrary P85 micelles internalize exclusively through CME [55]. Furthermore, at concentrations above CMC P85 inhibits caveolae-mediated endocytosis while having little or no effect on CME [55].

8.3 PS nanoparticles

Contrary to Pluronic^® micelles PS nanoparticles do not disintegrate upon dilution. However, their cellular trafficking can also depend on their size. For example, Lai et al. have recently shown that negatively charged 43 nm PS nanoparticles can enter HeLa cells through CME and reside in the endo-lysosomal compartment [83]. The reduction of the nanoparticle size to 24 nm resulted in a shift of the entry mechanism to clathrin-and caveolae-independent pathway. Furthermore, it also altered the processing of the nanoparticles, which in this case bypassed lysosomes and localized near the perinuclear space. Recent work also evaluated the entry of 200 nm positively charged (amine-modified) or negatively charged (carboxylate-modified) PS nanoparticles [26]. Independent of their charge such nanoparticles were shown to enter caveolae-deficient fibroblasts as well as confluent MDCK cells that lack apical caveolae.

8.4 Poly(amidoamine) (PAMAM) dendrimers

Dendrimers are repeatedly branched, monodisperse and usually highly symmetric compounds, which have been widely researched for delivery of therapeutic and diagnostic agents [84]. They have large number of surface groups that can immobilize drugs, enzymes, targeting moieties or other imaging agents. The amine- modified cationic dendrimers or carboxylate-modified anionic PAMAM dendrimers were shown to utilize multiple routes of entry inside Caco-2 and B16f10 melanoma cells.[85–86]. The entry mechanisms appeared to depend on the dendrimer generation. For example, a lower generation amine-modified dendrimer G2 was entering cells primarily through CME, while the higher generation dendrimer G4 utilized multiple routes [85–86]. The positively charged dendrimers have been shown to enter early endosomes and then route towards the lysosomes within 20 min. [85], [87]. In contrast, the negatively charged dendrimers appeared to accumulate in lysosomes at later points. Finally, in experiments studying the transport of the dendrimers across the polarized cell monolayers both cationic as well as anionic dendrimers were shown to open up the TJ and cross the monolayers by a paracellular route as well as transcellular route [88].

8.5 Non-viral gene delivery agents and DNA

Nanosized complexes of cationic lipids with DNA (“lipoplexes”) or synthetic polycations with DNA (“polyplexes”) have been used for gene delivery. Clearly, the cellular trafficking of such nanomaterials can strongly depend on their composition, size, shape, and surface characteristics. Since the polyelectrolyte complexes of DNA can undergo multiple structural transitions [89] it is hard to precisely control their characteristics, especially in the cellular environment. As a result there are often conflicting reports involving such systems. For example, one study reported that cellular entry of polyethyleneimine (PEI)-based polyplexes depends both on the cell type as well as PEI structure [90]. More specifically, it was shown that polyplex based on branched PEI in the monkey kidney cells (COS-7) predominantly utilize CME [91] while in HeLa cells both CME and caveolae-mediated endocytosis. In contrast, polyplex based on linear PEI appeared to internalize through CME independently of the cell type. To the contrary, another study suggested that linear PEI/DNA polyplex use both pathways while branched PEI/DNA polyplex as well as polyamidoamine (PAMAM) dendrimer/DNA polyplex use only caveolae [86]. This clearly contradicts to another report that branched PEI-based polyplexes use both CME and caveolae-mediated endocytosis [92]. Our own data suggested that polyplexes based on linear PEI or PEG-g-PEI graft copolymer utilize both pathways [93]). Similarly, polyplexes based on PEG-g-PLL graft copolymer were also shown to utilize CME, caveolae-mediated endocytosis and macropinocytosis [94]. Lipoplexes can enter cells through multiple pathways and their entry depends on the lipoplex composition. For example, lipoplexes based on DOTAP (N-[1-(2,3-dioleoyloxy)propyl]-N,N,N-trimethylammonium methyl sulfate) [92] and Lipofectamine were shown to internalize by CME. In contrast, DMRIE-C (1,2-dimyristyloxypropyl-3-dimethyl-hydroxy ethyl ammonium bromide)-based lipoplex internalizes through caveolae [95]. However, following the initial internalization stages both pathways appear to converge towards the late endosomes or lysosomes, where DNA delivered with Lipofectamine- and DMRIE-based lipoplexes is found. However, it should be noted that blocking of the lysosomal transport did not improve the transfection yield, which suggests that the DNA may utilize a different pathway to reach its nuclear destination.

An ideal gene delivery agent after initial uptake must escape degradation in lysosomes, and enable translocation of DNA to the nucleus. In a classic study Bausinger and colleagues followed the intracellular trajectory of linear PEI/DNA polyplexes by fluorescence live cell imaging [96]. They were able to show that at the initial stage polyplexes attached to the cell membrane are co-localized with actin filaments, while at the later stage they appear to move along the microtubules. A similar observation was made by Suh et al. who suggested that branched PEI/DNA polyplexes entrapped in endosomes are transported by motor-protein driven machinery along the microtubules through the cytoplasm towards the nucleus [97]. They also speculate that after reaching the perinuclear space the DNA can diffuse to the nucleus (Fig. 9). This appears to contradict a widespread belief that DNA can be released directly to cytoplasm from endosomal/lysosomal compartments through destabilization of these compartments by PEI (“proton sponge” hypothesis) [98]. The nuclear import of the DNA is cell-cycle dependent and is one of the most challenging barriers for non-viral gene delivery. Surprisingly, it was recently shown that synthetic polymers can help to overcome this barrier through the activation of specific cellular trafficking machinery [93]. In particular, Pluronic^® block copolymers added to cells enhance uptake and nuclear import of plasmid DNA delivered with polyplexes (Fig. 9). Notably, this effect appears to be due to the ability of the block copolymers to activate NFκB signaling in the cells. NFκB was shown to bind to cytosolic DNA and facilitate its transport to the nucleus through nuclear import machinery [99]. Consequently, the effects of the Pluronic^® copolymer were dependent on the DNA promoter [93]. The copolymer enhanced nuclear import of pDNA containing NFκB binding sites, but had no effect on the import of pDNA without these sites. Interestingly, Pluronics^® have been also shown to increase transcription of the DNA delivered into the nucleus in transiently and stably transfected cells [100]. This phenomenon may also involve activation of the NFκB transcription machinery. Overall, the interest in the effects of the Pluronics^® on gene delivery has been recently propelled by findings that these copolymers can greatly enhance the delivery of the naked DNA in vivo [101]. Very little is known how the DNA alone is transported into the cells, however, it was shown recently that naked DNA in serosal cells in vivo can internalize through macropinocytosis [102].

8.6 Surface modified nanoparticles that target multiple pathways

To increase cellular delivery of various nanomaterials such as liposomes, iron nanoparticles, siRNA- and DNA-containing polyplexes, etc., their surfaces were modified with cell penetrating peptides (CPPs), also known as Protein Transduction Domains (PTDs) [103–104]. Such ligands usually employ multiple pathways for intracellular trafficking. They represent small cationic polypeptides of from 10 to 30 amino acids. Some of the best characterized CPPs are TAT peptide, penetratin, transportan, poly-arginine, rabies virus glycoprotein (RVG) peptide, etc. For example, human immunodeficiency virus TAT fusion protein (>30 kDa) and TAT peptides (1 to 5 kDa) enter cells by cholesterol-dependent macropinocytosis as the primary entry mechanism [105–106]. Their transport does not seem to involve CME or caveolae. Yet, since the TAT peptide was shown to enter cells even when macropinocytosis was inhibited, alternative mechanisms were also postulated, which involve direct translocation of the peptide across the plasma membrane [107]. In a recent example, Torchillin et al. modified PEGylated liposomes and PEG-polyphosphatidylethanolamine micelles with 1) the targeting antibodies exposed at the liposome surface above the PEG layer and 2) the TAT peptide hidden within this layer [108]. Such double functionality is believed to ensure delivery of the liposomes to the target cells, where the antibody layer is cleaved and the TAT peptide exposed, which enables subsequent particle delivery into cells (Fig. 10). In another example the RVG peptide was shown to cross the brain endothelial cells and subsequently enter the neuronal cells using nicotinic acetylcholine receptor (nAchR) [109]. This peptide was further fused with nonamer arginine residues to form complexes with siRNA, which could deliver siRNA to the brain and enable protection against fatal viral encephalitis in mice. In a subsequent study, this peptide was conjugated to PAMAM dendrimers through a PEG spacer and then complexed with DNA, yielding polyplexes with RVG functionalities exposed at the surface [110]. These polyplexes were shown to internalize in the brain endothelial cells through CME and caveolae-mediated endocytosis, and deliver the DNA to the brain in vivo. Notably, their transport mechanism appeared to be more dependent on the GABA_B receptor than on nAchR.

In addition to CPPs, various other ligands were used to enhance cellular delivery of both nanoparticles and water-soluble polymers. For example, the HPMA copolymer was recently modified with an antibody against prostate specific membrane antigen (PSMA) [111]. The cellular internalization of such antibody-modified HPMA in the prostate cancer cells overexpressing PSMA was greatly enhanced compared to the copolymer conjugated to a non-specific IgG. The entry of the antibody-HPMA conjugates involved CME, macropinocytosis, as well as clathrin- and caveolae-independent endocytosis. At the later stages of trafficking the conjugate appeared to be routed to the late endosomes. Muro et al., have also showed that in polystyrene particles of micron and submicron sizes of different shapes (spherical or elliptical) can be internalized into endothelial cells by targeting intercellular adhesion molecule 1 (ICAM-1) [112]. The authors suggest such particles enter through caveolae-, clathrin-, and macropinocytosis-independent pathway, which, however depends on actin. Furthermore, the ICAM-1-targeted micron sized particles remained in the prelysosomal compartments whereas the submicron sized particles reached the lysosomes more rapidly. Moreover, it appeared that particle shape influenced the rate of endocytosis, with spheres internalized more rapidly than elliptical disks.

9.1 Effect of the particle charge

First, the cellular entry definitely depends on the nanomaterial charge. At present the majority of reports suggest that positively charged nanomaterials predominantly internalize through CME with some fraction utilizing macropinocytosis. Examples include cationic nanoparticles of very different origin - stearylamine-coated PEG-co-PLA, PLGA modified with PLL, amino group-modified SNTs, chitosan, etc [41–42] . However, there are some exceptions, most notably, PEI-based polyplexes, which are strongly cationic and yet may utilize multiple pathways including caveolae-mediated endocytosis [92]. It should be noted, however, that such complexes are usually formulated with excess of polycation and in cellular environment may additionally bind serum proteins, which can drastically affect their net charge and composition and alter trafficking. Furthermore, the large excess of polycations used in cell transfection experiments may also disturb the normal cell trafficking mechanism. Another exception involves positively charged PRINT particles which can utilize multiple pathways to gain cellular entry [81].

On the other hand, negatively charged nanoparticles, such as DOXIL^®, cl-micelles, and QDs, are more likely to utilize caveolae-mediated endocytosis [26],[65]. Exceptions include some carboxylate-modified PS nanoparticles and negatively charged PLGA nanoparticles, which can enter cells through caveolae-independent pathways [46– 47]. Since cell membranes are generally negatively charged, it is widely believed that negatively charged nanomaterials should internalize slower compared to their positively charged counterparts. This appears to be a case of negatively charged PRINT nanoparticles. Yet a striking exception is negatively charged QD, which were shown to internalize much faster compared to neutral or positively charged QD [65]. It is unclear from literature whether neutral nanomaterials show any preference for specific cellular entry routes. In case of neutral Pluronic^® micelles the CME pathway appears to be the main route of entry. However, this may be due to the ability of this block copolymer to inhibit caveolae-mediated endocytosis at micellar concentrations [55].

9.2 Effect of the particle size

It has been long believed that the size of nanomaterials may play a paramount role in their inclusion within different endocytic vesicles that greatly vary in sizes. Specifically, the need to keep particle small, between 10 to 100 nm, to enter the endocytic vesicles was postulated by many and became a foundation for current definition of nanomedicine by various agencies worldwide. However, while the small size may beneficial for a rapid entry into cells there appears to be no size cut off limit up to at least 5 µm to gain cellular entry of some materials through pinocytosis. The largest particles are, perhaps, more likely to enter cells through macropinocytosis as was suggested for PRINT microparticles [81]. Other notable example of size effects is negatively charged PS nanoparticles, which showed CME based entry for 43 nm particles and caveolae-mediated entry for 25 nm particles [83] . A major problem in dissecting the effects of the size is high polydispersity of many nanomaterials. However, as discussed above in some cases, such as highly polydisperse chitosan nanoparticles [51], the size of the particles may play much lesser role in defining the entry pathway than the chemical composition of the nanomaterial.

9.3 Effect of the particle shape

The effects of the particle shape were previously described for phagocytosis [113]. However, to best of our knowledge for pinocytosis the PRINT nanoparticles remain the only clear example of relationship between the particle geometry and cellular entry [81, 114]. Clearly, to define such effects the nearly monodisperse nanoparticles need to be manufactured. Other than shape, however, the compressibility of the material may matter for soft nanoparticles, such as nanogels and the like [114].

9.4 Effect of the cell type

Perhaps, the most exciting, challenging and poorly explored area is the relationship between the nanomaterial cellular trafficking and cell type. Few examples discussed above underscore that the cell type may be critical in defining the nanoparticle entry and final destination in the cells. Notably, it even appears that the differential endocytic pathways in normal and tumor cells: may be a gateway for selective targeting of novel nanomaterials into tumors [26]. Most studies of the nanomaterial trafficking performed so far did not emphasize the link between cell origin and availability of various endocytic pathways present in these cells. At the same time, the known cellular pathways may be differentially presented or even totally absent in selected cell types depending on the cell phenotype, and, in some cases, even growing conditions, such as cell density, presence of growth factors, etc. Obviously, in-depth understanding of the cell biology and it’s relation to nanomaterials science is most critical for advancement of this area of nanomedicine and drug delivery.

9.5 Understanding subcellular trafficking of nanomaterials

Delivery of polymers and nanomaterials to specific intracellular compartments, e.g. different cellular organelles, cytoplasm etc., has become increasingly popular due advent of drugs that target specific regions within the cells. For example, recently, a hexadentate-PLGA copolymer chemically conjugated to a selective MAPK inhibitor was shown to internalize and release drug in the cytosol of tumor cells [115]. Pluronic^® block copolymers can reach mitochondria where they exert unique pharmacological activities like ATP depletion in multidrug resistant cancer cells [116]. Some materials like cl-micelles can be routed to lysosomes and employ lysosomal pH as a trigger for release of a cytotoxic drug precisely within the cancer cells [26]. On the contrary, non-viral gene delivery systems, should bypass lysosomes to prevent degradation of DNA, siRNA or antisense oligonucleotides.

Therefore, the studies of cell trafficking should extend beyond the initial entry point of the nanomaterials towards precise definition of the entire nanomaterial cellular itinerary. In such studies one must be careful not to misinterpret the initial stages of entry as a guarantee for the nanomaterial’s final destination. Thus, caveolae-mediated endocytosis is still broadly believed to bypass lysosomes but in certain cases such as cl-micelles the entry through this pathway may destine to the lysosomes [26]. In another case Pluronic^® unimers enter through caveolae and then pass through ER to mitochondria – the pathway, which still is quite peculiar for biological molecules [21]. Needless to say the impetus should be given in the future to gain fundamental understanding of sorting of nanomaterials. One might even speculate that certain motif(s) (e.g., hydrophobic or chargeable groups), within the synthetic polymer might govern its sorting to a specific compartment and on removing such motif the trafficking may change. Well known analogs of such motifs are present in proteins (e.g. KDEL amino acid sequence targeting the ER, or GPI anchor on the plasma membrane) [117–118]. However, we also need to realize that synthetic water-soluble polymers and nanomaterials can exploit pathways that are unique for biomacromolecules and are not fully explored.

9.6 Understanding active targeting of nanomaterials

The concept of active targeting of nanomaterials using antibodies, polypeptides, aptamers and other targeting groups, has generated tremendous interest and on a surface appears to be simple [3, 84]. It may be more complicated than it is seems because both water-soluble polymers and nanoparticles are large enough to engage in considerable non-specific surface interactions with cells and organelles that may be comparable with the specific interactions of the ligands attached to polymers and nanomaterials. One obvious remedy could be, firstly, to modify the surface of the nanoparticle with “inert” polymer that would have minimal interactions with the cells, and, secondly, to attach one or multiple targeting ligand(s) at the inert layer to allow for specific binding of the targeting ligands with their receptor [4]. The problem, however, is in identifying such an “inert” polymer to mask the surface of the carrier. The current “gold standard” is PEG but there are clear indications that even this polymer attached to nanomaterials can activate complement, and otherwise engage the body immunity [119].

With this caution, there are several exciting examples of using targeting ligands for delivery of nanomaterials into cells. Some of these examples, e.g. APP-modified gold nanoparticles [22], RVG-peptide/siRNA complexes [109–110], TAT-modified liposomes [104] and others are discussed above. Such biospecific, surface-modified nanomaterials are rapidly internalized in cells. Conventional wisdom assumes that they might follow a similar intracellular itinerary as the ligands used for their modification. However, this might not be the case, as illustrated above by an example of QD, which are modified with Tf or CTB [54]. Furthermore, Pluronic^® block copolymers were shown to inhibit caveolae-mediated endocytosis at higher concentrations (0.01% and above) [55]. Even more interestingly they have shown an ability to enhance nuclear import of polyplex-delivered DNA [93]. By interacting with specific membrane microdomains, such as caveolae, polymers and nanomaterials may affect multiple signal transduction pathways. Another example of this sort is polysiloxane nanoparticles that appear to activate of eNOS signaling [64]. These examples suggest that polymers and nanomaterials can also influence intracellular transport. It is expected therefore that introduction of surface modifications in nanomaterial may result in a complex trafficking behavior, as nanomaterials may have an intrinsic ability to affect select cellular functions.