Serum protein adsorption and excretion pathways of metal nanoparticles

Abstract

While the synthesis of metal nanoparticles (NPs) with fascinating optical and electronic properties have progressed dramatically and their potential biomedical applications were also well demonstrated in the past decade, translation of metal NPs into the clinical practice still remains a challenge due to their severe accumulation in the body. Herein, we give a brief review on size-dependent material properties of metal NPs and their potential biomedical applications, followed by a summary of how structural parameters such as size, shape and charge influence their interactions with serum protein adsorption, cellular uptake and excretion pathways. Finally, the future challenges in minimizing serum protein adsorption and expediting clinical translation of metal NPs were also discussed.

Noble metal NPs have gained a lot of attention from researchers, physicians and pharmaceutical industries because their diverse material properties hold great potential in advancing our current diagnostic and therapeutic technologies [1]. Such properties are fundamentally dependent of their size, shape and compositions; as a result, interactions of metal NPs with serum protein adsorption and excretion routes are also strongly influenced by their structures. According to the US FDA, any contrast agents that are injected into the human body must be cleared completely in a reasonable amount of time to be deemed safe [2]. Comprehensive understanding of how these metal NPs may be cleared from the body is crucial in their future clinical translation. In this review, we will mainly use metal nanoparticles as models and illustrate their size-dependent material properties and potential biomedical applications; their interactions with serum protein adsorption and cellular uptake; two major excretion routes and future outlook.

Size-dependent material properties & biomedical applications of metal NPs

Surface plasmon absorption is one of the most interesting properties of metal NPs and strongly depends on structural parameters such as size, shape and surface roughness [3]. For instance, for gold NPs (AuNPs) with dimensions larger than the wavelength of light (R >λ), the energy and bandwidth of surface plasmons can be explained quantitatively with Mie theory. When the size of NPs reaches the electron mean free path (about 50 nm for gold [Au] and silver [Ag]), metallic properties such as surface plasmons, surface-enhanced Raman scattering (SERS), photothermal conversion and other related properties become significantly size dependent [4]. The decrease in the NP size, for example, results in the blue shift of surface plasmon of metal NPs [5]. As the size of the NPs becomes comparable to the electron Fermi wavelength (about 0.5 nm for Au), the continuous electronic band structure of metal NPs breaks into discrete energy states and luminescence instead of surface plasmons is observed in ultrasmall metal NPs [1,6].

The knowledge regarding these unique characteristics of metal NPs has guided many researchers to develop and launch a new generation of nanomedicine that can improve efficacy of photothermal/dynamic therapy (PTT/PDT), molecular imaging and drug delivery [7–10]. For example, several research groups have taken advantage of the strong localized surface plasmon resonance (LSPR) scattering of metal NPs such as Au and Ag, and passivated them with specific targeting ligands that make molecule-specific imaging and diagnosis of diseases possible [4,11]. The exposure of plasmonic NPs to electromagnetic radiation that resonates with the oscillation of the NPs’ surface plasmon, enables them to strongly absorb light and convert it into thermal energy through a series of physical processes [12]. For instance, El-Sayed et al. demonstrated in vitro that Au nanorods (AuNRs) may be utilized as contrast agents for molecular imaging and photothermal cancer therapy [13]. Once the AuNRs were conjugated with anti-EGFR antibodies, these nanostructures were able to bind specifically to the surface of the malignant epithelial cell lines with high affinity, enabling visualization from normal cells and photothermal therapy. Similarly, various other research groups have integrated metal nanostructures into biological systems by illustrating their potential imaging application [14–17]. Precious metal NPs such as those made of Au or palladium (Pd) have been utilized as agents for photoacoustic (PA) molecular imaging [18–22]. Due to their strong optical absorption in the near-infrared (NIR) region, Nie et al. demonstrated the potential of Pd nanosheets (PdNS) as a new class of PA contrast agents [20]. Their results show that these PdNS are stable and effective in optical labeling and visualization of tumor in vivo, which may help in the early detection, diagnosis and treatment of cancer.

Most biomolecules absorb and emit light in the visible region, which is one of several factors that limit the imaging applications of many NPs and organic fluorophores. However, the emission of NIR light is less affected by biomolecules and their autofluorescence is significantly reduced, making these NIR-emitting NPs more suitable for in vivo fluorescence imaging [23,24]. Small nonplasmonic but luminescent metal NPs also show great potential in molecular imaging. As an example, the NIR fluorescence imaging of glutathione-coated AuNPs (GS-AuNPs) was demonstrated and evaluated, in which the renal clearance of the NPs were monitored over a 12-h period [25]. The study also revealed the passive tumor targeting capability of these NPs through the known enhanced permeability and retention (EPR) effect. In addition, our group has fabricated near infrared (NIR) emitting radioactive AuNPs that are not only renal clearable, but also offer the opportunity for both fluorescence and single-photon emission computed tomography (SPECT) imaging. The ability of metal NPs to target cells, depending on how their surface is passivated, may be utilized in the diagnosis and treatment of diseases, through the delivery of medicine. For instance, noble metal NPs such as those made of Au and Ag, have been shown to have therapeutic effects and are capable of carrying and transporting drugs [7,8,10].

Serum protein adsorption of metal NPs

Once the metal NPs enter the body's circulatory system, these NPs will inevitably interact with proteins that are present in biological fluids, and the NP's surface will be immediately covered with various types of proteins forming the ‘protein corona’, also known as the ‘bionano interface’ [26]. The formation of protein corona is a dynamic, competitive and complex process, in which low-affinity high-abundance proteins that bind to the NP surface are replaced by high-affinity lower-abundance proteins, due to an ongoing exchange between the adsorbed proteins on the NPs and free proteins present in the biological environment [27]. It is well recognized that the protein corona plays an integral part in the interactions of colloidal metal NPs with living matter. This corona is the primary contact with the cells, including interactions with the cell receptors [27]. Furthermore, the formation of protein corona is believed to be highly affected by the physicochemical properties of the metal NPs as shown in Figure 1, such as size, surface charge, shape, surface roughness as well as their chemical composition [28], which will be discussed in detail in this review.

Figure 1. . The nanoparticles’ physicochemical properties (size, shape and surface charge) and serum protein adsorption are fundamental factors that affect the fate of nanoparticles in vivo.

Size-dependent serum protein adsorption

The size of the NPs in vitro can have a considerable influence on the adsorption of biomolecules, in which the formation of protein corona can alter the NPs’ overall hydrodynamic diameter (HD) [29]. These differences in NP size can affect the identity of the proteins adsorbed on the surface, the thickness of the protein corona formed as well as their binding constants [27,28,30]. For instance, Deng and coworkers investigated how size affects the binding kinetics of the protein fibrinogen to negatively charged poly(acrylic acid)-coated AuNPs with differing HD size from 7 to 22 nm, in which the protein had a higher affinity for the larger NPs and slower dissociation rate [30]. They concluded that even slight changes in the dimensions of the NPs can evidently influence the binding of protein with the NP's surface and within the protein corona. This phenomenon has been observed in other types of inorganic NPs. Tenzer and coworkers used 1D and 2D gel electrophoresis, liquid chromatography-coupled mass spectrometer (LC-MS) and immunoblotting to show that 8, 20 and 125 nm silica NPs (SiNPs) had different types of proteins adsorbed on their surface, when incubated with an equal amount of human plasma for 1 h [31]. According to their analysis, there is an enhanced binding of the lipoprotein clusterin to the small 8 nm SiNPs, while prothrombin or the actin regulatory protein gelsolin preferentially adsorbed on the 125 nm size SiNPs, which contradicted the perception that small proteins only favor small particles [28]. In addition, the thickness of the protein corona that formed, the binding constant, as well as the degree of cooperativity of protein-NP binding depended heavily on the NP size. For example, De Paoli Lacerda et al. discovered that AuNPs of different sizes ranging from 5 to 100 nm had major interaction with proteins that are commonly present in the human plasma (i.e., albumin, fibrinogen, histone, γ-globulin and insulin), where the binding constant and the degree of cooperativity of protein-NP binding increased with the size of the NPs below 50 nm [29]. In addition, the fluorescence quenching of histone by the AuNPs confirmed their hypothesis that the thickness of the protein layer also increased with increasing NP size. This important finding should always be taken into consideration, especially when designing metal NPs intended for clinical use in the future.

Surface charge effect on serum protein adsorption

The binding of proteins onto the surface of NPs is also influenced by the NP's surface charge, and has a significant impact on the NPs’ overall biodistribution and clearance. Several researchers have investigated the effect of protein adsorption on NPs with negatively-, neutrally- and positively-charged surface [32–34]. For instance, Deng et al. incubated the AuNPs of differing surface charge (positively-charged poly(N-(2–aminoethyl)acrylamide) polymer, negatively charged poly(acrylic)acid polymer and neutrally charged poly(N-(2,3–dihydroxypropyl)acrylamide)) with human plasma and determined that both the positively and negatively charged AuNPs had a wide range of proteins bound to their surface with high affinity (primarily fibrinogen), while the AuNPs with neutral charge had very little bound proteins [32]. According to the binding kinetics study, however, the positively charged AuNPs showed a lower affinity and faster release of the protein fibrinogen than the negatively charged NPs. Deng's group hypothesizes that fibrinogen may have specific binding sites that gave preference to the negatively charged particles. The sign of the surface charge, however, does not always indicate the adsorption of proteins. Huhn et al. demonstrated that the negatively and positively charged AuNPs of similar HD with zeta potential of -40 mV and +10 mV, respectively, had comparable amount of the protein human serum albumin (HSA) adsorbed per particle [35]. However, one of the differences was that the positively charged NPs were taken up faster by several cell lines, which may ultimately initiate cytotoxicity. In addition, others have verified how the density of the surface charge can affect the formation of the NP–protein complex [36,37]. Gessner et al. obtained polystyrene NPs with differing surface charge density and incubated them in human plasma [36]. In their study, they observed and concluded that the total amount of proteins bound to the NPs accumulated with increasing surface charge density. Walkey et al. observed a similar phenomenon using AuNPs, where the increase in density of PEG molecules on the NP surface resulted in the decrease of total adsorbed serum proteins at constant NP size [38]. And finally, the initial surface charge of NPs may change once they enter the blood circulation, mainly due to the establishment of proteins on the NP surface. For example, Ehrenberg et al. found that the zeta potential of polystyrene NPs shifted by 5–10 mV for the negatively charged NPs and ∼60 mV for the positively charged NPs after serum protein adsorption [39].

Shape effect on serum protein adsorption

According to several independent studies, the adsorption of serum proteins onto the surface of nanostructures, as well as the proteins’ conformational change, is also affected by the particles’ geometrical shape [40–42]. Gagner et al. investigated the particle shape effect of 11 nm spherical AuNPs and AuNRs with dimensions ∼10 nm × 36 nm, on protein adsorption and conformation [40]. As a model, the proteins lysozyme (Lyz) and α-chymotrypsin (ChT) were selected due to their well-characterized structures, resemblance in size and earlier characterization in nanospheres. Based on their results, the spherical AuNPs had both Lyz or ChT molecules adsorbed on their surface, but an order of magnitude lower than those adsorbed onto AuNRs. They speculated that this significant difference may be due to the surface areas of the nanostructures, where the spherical AuNPs have approximately 530 nm² compared with 1550 nm² for the AuNRs. In addition, the surface curvature of the spherical AuNPs may have limited the amount of proteins adsorbed, due to steric hindrance between proteins. The surface energies associated with specific atomic orientations are thought to influence the protein adsorption itself, as well as the protein structure conformation. Aside from pure metal nanostructures, the shape effect on protein adsorption has also been observed for other NPs [41,42]. Deng et al. investigated the effect of shape on plasma protein binding to titanium oxide (TiO₂) NPs. They utilized TiO₂ nanorods with average lengths of ∼75 nm and widths of ∼27 nm, and TiO₂ nanotubes with tube diameter of ∼9 nm. Using a 2D gel electrophoresis, they determined that the nanostructures had different protein binding patterns, where IgM and IgG were the major proteins bound to the nanorods while fibrinogen primarily bound to the nanotubes. Since there is still a scarcity in the information regarding how NP shape influences serum protein adsorption, more studies on this topic would certainly help in the fabrication of NPs suitable for in vivo applications.

Composition & surface roughness on serum protein adsorption

Aside from the NP size, surface charge and shape, the NP's composition and surface roughness has also been shown to influence the adsorption of serum proteins, which can ultimately affect their clearance pathways [27]. The effects of chemical composition have been observed with iron oxide nanoparticles (IONPs) in comparison with oil-in-water nanoemulsions [43,44]. Both types of NPs were separately incubated in human plasma and the protein corona was analyzed via 2D gel electrophoresis. According to the results, the IONPs surface was determined to contain significantly more immunoglobulins, fibrinogen and apolipoproteins than the oil-in-water nanoemulsions. The surface roughness or topography of NPs can also affect the types of proteins adsorbed and their interaction at the nanobio interface. For instance, when two types of core-shell superparamagnetic iron oxide NPs (SPIONs), one with smooth Au shell and the other with rough Au shell, were incubated in human plasma for 1 h, it was found that both types of SPIONs had significantly different types of proteins adsorb on their surface [45]. Based on 1D gel electrophoresis and LC-MS/MS analyses, the SPIONs with smooth Au surface had proteins with 310–120 and 70–30 kDa adsorbed on its surface, while the SPIONs with rough Au surface were covered with 120–70 and 30–10 kDa proteins. The researchers proposed that the edges of the jagged surfaces had more scattered surface charge, which favored the adsorption of proteins by means of the van der Waals, electrostatic and hydrogen-bonding interactions.

Correlation of cellular uptake with serum protein adsorption

In the physiological environment, the cells encounter NPs in their nonpristine state or form with the proteins adsorbed on their surface. The absence and presence of preadsorbed proteins on the same NPs have been shown to affect how NPs are localized inside the cells [46]. For example, 50 nm silica NPs (SiNPs) were incubated with cells: one in the absence of serum proteins (‘bare-SiNPs’) and the other in the presence of serum proteins. As expected, the cells interacted with the particles differently. The bare-SiNPs were endocytosed and engulfed in vesicles at a higher concentration, and some were adhered more strongly to the cell membrane than those SiNPs with preadsorbed proteins. The study also revealed that the types of proteins adsorbed are determined by the location where the NPs were situated. After the cell studies, the NPs were recovered and the proteins adsorbed on their surface were investigated. The majority of the proteins adsorbed on the SiNPs (with preadsorbed serum) were immunoglobulins, apolipoproteins and complement proteins, whereas the bare-SiNPs had mostly cytosolic proteins and membrane-associated proteins. According to Aggarwal et al., the most commonly adsorbed proteins detected on nearly all NPs are those that are highly abundant in the blood, which were albumin, immunoglobulin (IgG), fibrinogen and apolipoproteins [47]. Aside from the NPs’ size and surface charge, the geometrical shape also influences their endocytosis. Many researchers have synthesized different types of NPs with varying shapes, which can affect the NPs’ cellular uptake and ultimately their tissue biodistribution [48–50]. Generally, most NPs can be endocytosed by cells through a receptor-mediated process, while others can be endocytosed nonspecifically [51,52]. Once the NPs begin to be endocytosed, their shape can influence the kinetics of cellular uptake. Gratton et al. determined that nanorods had the highest uptake by HeLa cells, followed by spheres, cylinders and cubes with size larger than 100 nm [53]. For NPs below this size, however, spherical NPs have been shown to be internalized at a higher rate than nanorods and changing the aspect ratio of nanorods can influence cellular uptake, where higher aspect ratio reduces cell internalization [54].

Excretion routes

When noble metal NPs used for imaging or drug delivery are intravenously injected, they directly enter the circulatory system, interact with various proteins and agents of the immune system, are transported throughout different organs and tissues, and are finally excreted out of the body as illustrated in Figure 2. Excretion is an important biological process that prevents damage and toxicity by removing unwanted materials from the body. There are two primary excretion routes – the renal (urinary) pathway and the hepatobiliary (hepatic) pathway. The renal pathway is a complex process that involves glomerular filtration, which is highly dependent of molecular size and surface charge. The capillary walls of the glomerulus typically have a filtration-size threshold (FST) of <5.5 nm, which means that molecules with an HD size within this value can be filtered by the kidneys while the larger ones (HD >8 nm) remain in the blood circulation [55]. Because clearance pathways are strongly size-dependent, the physicochemical properties of metal NPs such as size, surface charge and shape are the major fundamental factors that influence their clearance.

Figure 2. . Following their entrance in the vascular system, nanoparticles with distinct sizes, surface charges and shapes encounter proteins that may eventually alter their physicochemical properties.

These changes could result in their opsonization by cells of the reticuloendothelial system such as those in the liver and spleen, while others may undergo renal excretion. In the kidneys, not all nanoparticles that are smaller than the filtration-size threshold can be excreted. As shown, positively charged nanoparticles are drawn toward the negatively charged capillary endothelium walls, while others can easily pass through into the Bowman's space.

d: Diameter; HD: Hydrodynamic diameter.

Hepatobiliary pathway

Foreign matter that cannot be filtered via the renal route and remain in the bloodstream is excreted via the hepatobiliary pathway in the liver. Kupffer cells (also known as reticuloendothelial cells) are resident macrophages in the liver capable of phagocytizing foreign materials present in the blood of the hepatic sinus [56]. These cells, which are part of the reticuloendothelial system (RES), play a significant role in liver clearance through the selective endocytosis of opsonized antigens [57]. Generally, large molecules and NPs (>10 nm) that are administered into the body are taken up by the cells of the immune system and accumulate in the organs of the RES, where they are metabolized and excreted via the hepatic route [48]. Semmler-Behnke et al. studied the biodistribution of 18 nm core-size AuNPs that were intravenously injected into rats [58]. According to their results, the large AuNPs were immediately taken up by the liver with ∼94% of the injected dose (%ID), and negligible amount of Au was detected in the urine (<0.1%ID). Several other research groups have fabricated NPs ∼10 nm, which were very close to the FST, but were still rapidly cleared by the liver. For instance, Cho et al. studied the effect of size on the kinetics of PEG-coated AuNPs [59,60]. They determined that the 4 and 13 nm core-size NPs had higher concentration in the blood at 24-h postinjection (pi.) compared with the 100 nm NPs, which were completely cleared from the blood by 24 h pi. Their results correlated with the biodistribution studies, where the 100 nm NPs were quickly taken up by the liver and spleen within 30 min of pi. However, a limited number of large particles have been shown to still be excreted in the urine. Balogh and coworkers have demonstrated that ∼13% ID of large 22 nm AuNPs were able to be cleared from the body renally, while more than 30%ID/g accumulated in the liver alone [61]. Other metal NPs, such as Pd, Ag and bimetallic NPs, have also been intensively studied in the past few years [20,62,63]. Nie et al., investigated biodistribution of Pd nanosheets with HD ∼18 nm and found that ∼25%ID/g of Pd accumulated in the liver and ∼8%ID/g in the spleen at the 24 h pi. Nevertheless, these results strongly suggest that the size of metal NPs is still the major and dominant factor in directing their excretion via the hepatobiliary pathway [20].

In addition to size, shape also plays a key factor in hepatic clearance of metal NPs. Black et al. determined that 35%ID/g of the metal nanospheres were detected in the liver at 24 h pi. while the nanodisks, nanorods and nanocages were observed to be 55, 52 and 63%ID/g, respectively [64]. On the other hand, only 5%ID/g of the nanospheres were found in the spleen while the others were >40%ID/g at 24 h pi. This shape effect on clearance and tissue biodistribution is also observed in other inorganic NPs [50,65]. For example, Decuzzi et al. observed that spherical silicon-based NPs had less accumulation in the RES organs (<13%ID) compared with the cylindrical NPs (∼20%ID) [51]. It is noteworthy that all of the tissue distribution data provided here is obtained via elemental analysis, but other quantification methods can be used as an alternative approach [66].

Since many nonbiodegradable metal NPs currently being developed are larger than the FST, a large majority of the administered NPs inevitably accumulate in the organs of the RES. There have been numerous reports regarding metal NPs toxicity in vitro and in vivo over the past decade due to accumulation, as summarized in several reviews [67–69]. For example, Cho et al. studied the in vivo toxicity of 13-nm PEGylated AuNPs by intravenously injecting them to mice [60]. They determined that these NPs accumulated in the liver and spleen up to 7 days pi. Furthermore, they observed that the NPs induced acute inflammation and apoptosis in the liver. Chen et al. have also assessed the toxicity of citrate-capped AuNPs with different sizes that were administered intraperitoneally [70]. They determined that the AuNPs (8–37 nm) induced severe sickness in mice and concluded that the systemic toxicity was due to the injury of the liver, spleen and lungs. Although the uptake of large metal NPs in the organs of the RES has been frequently observed, low accumulation had been shown for some NPs [71–73]. For instance, Zhang et al. determined that >50 nm mesoporous-silica-coated upconversion fluorescent nanoparticles (UCNPs) that were intratumorally injected into melanoma tumor-bearing mice had negligible uptake in the organs such as the kidneys, liver and spleen, but with high uptake in the tumor [72]. In addition, their in vivo studies showed tumor growth inhibition in the PDT-treated mice, which may serve as a platform for noninvasive cancer therapy in the future. However, it is important to note that the route of administration of metal NPs have an effect on how they are distributed in the body, which is another factor that should be considered [8]. Furthermore, whether these metal NPs can undergo enterohepatic circulation or not, is an issue that still needs to be addressed. As it has been observed in drug molecules, enterohepatic circulation occurs by biliary excretion and intestinal reabsorption of a solute or drug, which could still result in toxicity [74]. Nevertheless, extensive studies regarding the accumulation-induced toxicity of nonbiodegradable metal NPs are still needed.

Urinary pathway

Size plays a dominant role in the fate of foreign matter in the human body. For small molecules such as proteins and carbohydrates, the FST of the kidneys has been intensively researched and studied [55]. For example, the carbohydrate inulin has an average HD of about 3 nm, and it has been shown to pass through the renal pathway with 100% efficiency, while only about 9% of an antibody Fab’ fragment (HD ∼6 nm) can be detected in the urine [75,76]. On the other hand, metal NPs have a broader size distribution, which makes it more challenging to pinpoint their actual kidney FST compared with biomolecules. There are several types of NPs that can be cleared from the body through the renal route [2,48,77,78], but currently, only a few metal NPs are capable of passing through the kidney filtration barrier as listed in Table 1, namely AuNPs, PdNS, Au nanoclusters (AuNCs), copper NPs (CuNPs), luminescent porous silicon NPs (LPSiNPs) and copper⁶⁴ alloyed AuNCs [20,77,79–82]. For renal clearable AuNPs, there are several different ways to fabricate them, one of which is the use of thiolated ligands [77,83]. Luminescent AuNPs (core: 1.7 nm/HD: 2.3 nm) with 600 nm emission have been synthesized at room temperature by the reduction of HAuCl₄ with glutathione (GSH), a tri-amino acid peptide [77]. Tu et al. and other research groups have been able to construct AuNPs (core: 2.5 nm/HD: 3.3 nm) with near infrared (NIR) emission, via the thermal reduction of Au³⁺ with GSH at 90–95°C [25,83]. Several other researchers have utilized the versatility of PEG in synthesizing luminescent AuNPs with renal clearance and long blood circulation times [84]. For instance, our group fabricated PEGylated AuNPs (core: 2.3 nm/HD: 5.5 nm) via a facile one-pot synthesis method by the thermal reduction of HAuCl₄ with thiolated PEG (PEG-SH) ligands. PEG-SH has also been used to directly synthesize Cu⁶⁴-AuNCs (core: 2.8 nm/HD: 4.3 nm), in which nonradioactive Cu(NO₃)₂ and HAuCl₄ were mixed together, followed by the addition of PEG methyl ether thiol (mPEG-SH) [79]. More recently, ultrasmall PdNS-GSH have also been shown to be capable of entering the renal pathway. Tang et al. synthesized ultrasmall PdNS (core: 4.4 nm/HD: --) by reacting palladium(II) acetylacetonate in the presence of poly(vinylpyrrolidone) (PVP) and sodium bromide (NaBr) in a solvent comprising N,N-dimethylpropionamide (DMP)-water mixture [85]. The nanosheets were dispersed in a solution containing GSH to obtain the renal clearable PdNS-GSH.

Table 1 . Current renal clearable metal nanoparticles (via intravenous injection).

NP	Size (core/HD)	Surface coating	BioD.	Renal clearance	Ref.
64Cu-Au alloy NP	2.5 nm/4.3 nm	PEG350	Li ∼15%ID/g Sp <15%ID/g Ki ∼15%ID/g	Yes (24 h pi.)	[79]
AuNPs	2.5 nm/3.3 nm	GSH	Li ∼3%ID/g Sp ∼2.5%ID/g Ki <20%ID/g	Yes (12 h pi.)	[25]
AuNPs	1.7 nm/2.2 nm	GSH	Li 3.7%ID/g Sp 0.3%ID/g Ki 8.8%ID/g	Yes (24 h pi.)	[77]
198[Au]AuNPs	2.6 nm/3.0 nm	GSH	Li <5%ID/g Sp <3%ID/g Ki <10%ID/g	Yes (24 h pi.)	[86]
PdNS	4.4 nm/--	GSH	Li <10%ID/g Sp <5%ID/g Ki <5%ID/g	Yes (15 days pi.)	[85]
AuNPs	2.4 nm/6.6 nm	DTDTPA	Li <6%ID/g Sp <5%ID/g Ki <12%ID/g	Yes (24 h pi.)	[87]
LPSiNPs	126 nm/151 nm	Dextran	Li <25%ID/g Sp ∼60%ID/g Ki <5%ID/g		[82]
	–-nm/<6nm	Dextran	Li <5%ID/g Sp <10%ID/g Ki <5%ID/g	Yes (4 weeks pi.)
AuNPs	2.3 nm/5.5 nm	PEG1000	Li <5%ID/g Sp <5%ID/g Ki <10%ID/g	Yes (24 h pi.)	[84]
CuNPs	2.0 nm/2.7 nm	GSH	Li ∼10%ID Sp <0.5%ID Ki <1.0%ID	Yes (24 h pi.)	[88]

BioD: Biodistribution; DTDTPA: Dithiolated polyaminocarboxylate; HD: Hydrodynamic diameter; ID: Injected dose; Ki: Kidney; Li: Liver; NP: Nanoparticle; pi.: Postinjection; Sp: Spleen.

Renal clearance of metal NPs is dependent not only on size, but also their surface charge. Many small metal NPs with HD ≤6 nm have been shown to still accumulate in vital organs of the RES but at a lower concentration, and this observed phenomena must be due to other factors such as the NP's surface charge. The charge on the NP's surface may have potential electrostatic interactions between each other, as well as with proteins that are present in the biological environment, which may result in the increase of HD. The high ionic strength that is inherent in the native physiological environment is known to significantly decrease the stability of metal nanoparticles. For instance, several types of NPs such as iron oxide (IO), gold (Au) and silver (Ag) NPs have suffered from low stability and formed large aggregates in buffered solutions or solutions containing high salt concentrations [77,89–91]. Chen et al. determined that ∼2–3.5 nm cysteine-coated AuNPs were fairly stable in aqueous solution but immediately formed aggregates (HD >100 nm) in phosphate-buffered saline (PBS) solution [77]. As a result, even though their core size is smaller than the FST, severe accumulation of these gold nanoparticles was found in the liver (∼55%ID) after intravenouos injection due to aggregation. Interestingly, our group has demonstrated that the negatively charged GS-AuNPs (∼2 nm) have low serum protein adsorption and can serve as ‘surface ligands’ to stabilize other NPs [90,92]. For instance, the incorporation of GS-AuNPs on the surface of superparamagnetic iron oxide NPs (SPIONs) to form the hybrid nanostructures (HBNPs) resulted in the reduction of serum protein adsorption compared with the pure SPIONs [90]. The incubation of serum proteins with the SPIONs led to the increase in the HD of the particles, with tenfold the amount of proteins adsorbed compared with the HBNPs. Furthermore, the stability of the SPIONs and HBNPs in different salt solutions (i.e., NaCl and MgCl₂) were evaluated and compared, where the HBNPs evidently showed higher stability. The surface charge effect on NP clearance has been observed before in other types of NPs [34,93]. NPs with a neutral surface charge most likely result in longer circulation time in the blood and lower uptake by the organs of the RES due to reduced opsonization [47,94]. For instance, Papisov and coworkers observed that the positively charged (poly-lysine)-dextran-coated IONPs had much faster blood clearance than the negatively charged succinate-(poly-lysine)-dextran-coated IONPs, which showed a comparable biodistribution and pharmacokinetics with the neutrally charged dextran-coated IONPs [95]. This phenomenon was also observed in metal NPs. For example, Balogh et al. showed that the positively charged AuNPs accumulated in the kidneys more, whereas the negatively and neutrally charged NPs had higher percentage in the liver [61]. In another independent study, Hirn and coworkers also investigated the effect of NP surface charge, where positively charged cysteamine-coated AuNPs or negatively charged thioglycolic acid coated AuNPs of equal size (2.8 nm) were intravenously injected into adult female rats. After 24 h, the biodistribution studies revealed that the NPs with positive charge had significantly lower accumulation in the liver, but a much higher concentration in the blood. They also determined that the positively charged NPs had slightly higher accumulation in the kidneys. This result is expected because the glomerular capillary walls in the kidneys have fixed negative charges which makes the kidney filtration of molecules charge selective [55].

PEGylated and zwitterionic AuNPs, which are known to have neutral or very low surface charge, have also been shown to have longer blood circulation compared with positively charged and negatively charged NPs [84]. PEG molecules, used as surface ligands, have been shown to decrease the electrostatic charge on the NPs’ surface [84,96]. As a result, the NPs physiological stability can be dramatically improved due to the reduction or prevention of the protein corona formation, which slows down the uptake of the PEGylated NPs by phagocytic cells of the RES and prolongs their blood circulation. Arvizo et al. conducted a study regarding how the NPs’ surface chemistry affect their pharmacokinetics, biodistribution and tumor uptake [97]. Four groups of ∼2 nm core size AuNPs with different surface chemistry (neutral, zwitterionic, negative and positive charge) were independently injected in mice, and it was observed that both the PEGylated (neutral) and zwitterionic AuNPs had longer blood circulation, while the negatively and positively charged NPs had relatively short half-lives. Furthermore, the zwitterionic and negatively charged AuNPs had higher accumulation in the liver compared with the PEGylated and positively charged NPs. They concluded that charge is an important key in determining how NPs interact with biological systems.

Although most renal clearable NPs observed to date are spherical, there are a few NPs with other geometrical shapes that are also capable of renal clearance. For instance, nonmetal based NPs such as single walled carbon nanotubes (SWCNTs) that resemble the shape of nanorods have been detected in the urine with a clearance efficiency of ∼65%ID [98]. In addition, PdNS below 10 nm have also been detected in the urine but at a lower efficiency of ∼31% at 15 days pi. [85]. Hence, the geometrical shape of nanostructures intended for in vivo applications is equally important and must be considered in fabricating NPs with decreased accumulation in the body.

Most of the metal NPs that are currently renal clearable (shown in Table 1) do not have material properties that are apparent in their larger counterparts. As mentioned previously, metal NPs less than 5 nm have no LSPR properties, which limit their application in PDT or PTT. However, Qian et al. have observed and demonstrated that the 2.2 nm AuNPs they fabricated exhibited a surface plasmon resonance at approximately 520 nm, which indicated that the particles in this size regime can also have metallic properties [99]. With proper surface functionalization, these NPs can be designed to enable their renal clearance in the future. In addition, although there is insufficient information regarding the drug delivery application of renal clearable metal NPs, their use as drug carriers may be hampered by their limited capacity or payload due to their ultrasmall size. For example, ∼100 nm mesoporous SiNPs were loaded with ∼2.4 mg doxorubicin (DOX) per gram NP, while sub-10 nm particles such as mesoporous SiNPs and UCNPs were loaded with only ∼3 Cy5.5 dyes per NP and 72 µg DOX/mg NP, respectively [100–102]. The drug loading capacity of inorganic NPs, however, can be tuned depending on the surface coating [103,104]. Sherlock et al. was able to design ∼5 nm FeCo/graphitic carbon shell nanocrystals (NCs) efficiently loaded with ∼200 to ∼2000 DOX per NC through the strong π-stacking on the graphitic shell [105]. Similar to many inorganic NPs, organic NPs such as those made of dendrimers or polymers have been utilized in drug delivery, as well as in multimodality imaging [106]. By contrast, these organic NPs like liposomes are biodegradable, which is a big advantage that led to their advancement in the field [107]. Therefore, reducing the accumulation-induced toxicity of nonbiodegradable metal NPs is crucial in nanomedicine.

Conclusion

In summary, metal NPs with size-dependent material properties have great potential in biomedicine, but their structural parameters such as size, shape, and charge must be taken into account in order to minimize serum protein adsorption, cellular uptake, and toxicity. Several non-biodegradable metal NPs have been developed with HD below the FST and proper surface chemistry that reduces the formation of protein corona, which give them the capability to be cleared renally. Although metal NPs can also be excreted via the hepatic pathway, this route can pose potential health risks. Hence, the renal route is preferable for NPs intended for clinical use. As summarized in this review, even the renal clearable metal NPs still have slight accumulation in vital organs such as the liver and kidneys. As a result, better design and development of metal NPs in the future can certainly help reach the requirement for their complete elimination from the body.

Future perspective

The complete elimination of administered NPs in the body is a critical requirement for their translation into clinical use. Therefore, investigation of the toxicity, biodistribution and clearance of nonbiodegradable metal NPs in vivo is of utmost importance, because it can provide us with a better understanding on how to improve their design. As mentioned previously, in order for metal NPs to be cleared from the body they must meet the minimum size requirement with a final HD <6 nm.

After satisfying the minimum size requirement, the NPs’ surface charge and shape must be taken into account, as well as understanding how protein corona may form on the NPs. We have reviewed earlier that even some inorganic NPs that meet the size requirement, still accumulate in the liver and other vital organs, and are unable to be cleared in the urine. Generally, positively charged NPs can be effectively phagocytosed or internalized by cells through electrostatic interactions, since the plasma membrane of cells have a net negative charge [108]. As mentioned previously, the glomerular capillary walls in the kidneys have fixed negative charges, which influences the renal clearance of NPs, especially those with positive charge. PEGylating the metal NPs can render their surface charge to be neutral, which can decrease their accumulation in the kidneys and also prolong their circulation in the bloodstream [84,109]. On the other hand, zwitterionization of the NPs can also reduce the accumulation in the organs of the RES, as has been shown by our group, except these zwitterionic NPs have faster renal clearance than the PEGylated NPs. The use of metal nanostructures that form large protein coronas, has limited application in biomedicine, and physicochemical properties such as size, shape and surface charge can influence protein adsorption. For instance, larger NPs tend to attract more proteins than smaller NPs in such a way that the thickness of the protein layer adsorbed increased with particle size [29]. The process of PEGylation and zwitterionization is one strategy to reduce or eliminate the formation of the protein corona [84,96,109]. In addition, functionalizing or precoating the NPs’ surface with proteins or peptides has also shown to reduce adsorption of other serum proteins present in the blood. In terms of shape, spherical NPs have shown to be less targeted by proteins compared with the other NPs with different geometrical shapes, due to difference in surface curvature and surface energies. Therefore, the successful manipulation of NPs’ interaction with proteins and cells in vivo can definitely ensure the safety and effectiveness of nanomaterials in biomedicine.

Before these NPs can benefit nanomedicine, characterizing the effect of NPs on biological systems is important. Current results have highlighted the effects of the physicochemical properties of NPs, in which size, shape and surface charge can not only influence the formation of protein corona and vice versa, but also the NPs cellular uptake and biological clearance. The fabrication of metal NPs with size less than the FST have shown to be more suitable for in vivo applications, due to less accumulation in vital organs and lower cytotoxicity compared with their larger counterparts. Similarly, the surface charge and geometrical shape also affects how NPs interact with proteins and cells. Although it has been established that these basic properties play an integral part in clearance, the formation of protein corona surrounding the NPs still remains difficult to control and not ideal for targeting applications due to the potential RES uptake. Therefore, a better and more complete understanding of the NP–protein interactions is still necessary in order to utilize NPs in their full potential in vivo.

Executive summary.

Serum protein adsorption of metal nanoparticles

• The physicochemical properties of noble metal nanoparticles (NPs) such as size, surface charge and shape are main factors that influence the formation of protein corona and the NPs’ excretion.

• Serum protein adsorption in metal NPs typically increases with increasing NP size, decreases with increasing surface PEG density and increases with increasing NP aspect ratio (nanorods vs nanospheres).

• PEGylation and zwitterionization of metal NPs can reduce the formation of protein corona.

Excretion routes

• There are two primary excretion routes: renal (urinary) and hepatic (hepatobiliary).

• Generally, metal NPs with hydrodynamic diameter (HD) less than the kidney filtration size threshold (HD ≤6 nm) can be cleared through the renal pathway, while those above this value proceed through the hepatic route. Some metal NPs with HD ≤6 nm can still accumulate in the organs of the reticuloendothelial system, due to factors such as surface charge and shape.

• Metal NPs can be excreted via hepatic route, which can pose potential health risks. Therefore, renal route is the preferred pathway for NPs intended for clinical use.

Future perspective

• Understanding the NP–protein complex fundamentally can certainly benefit nanomedicine in the future.

• The investigation of the toxicity, biodistribution and clearance of nonbiodegradable metal NPs in vivo is of utmost importance, because it can provide us with a better understanding on how to improve their design.