Pharmacokinetics and tumor delivery of nanoparticles

Abstract

Nanoparticles (NPs) have been widely used in different areas, including consumer products and medicine. In terms of biomedical applications, NPs or NP-based drug formulations have been extensively investigated for cancer diagnostics and therapy in preclinical studies, but the clinical translation rate is low. Therefore, a thorough and comprehensive understanding of the pharmacokinetics of NPs, especially in drug delivery efficiency to the target therapeutic tissue tumor, is important to design more effective nanomedicines and for proper assessment of the safety and risk of NPs. This review article focuses on the pharmacokinetics of both organic and inorganic NPs and their tumor delivery efficiencies, as well as the associated mechanisms involved. We discuss the absorption, distribution, metabolism, and excretion (ADME) processes following different routes of exposure and the mechanisms involved. Many physicochemical properties and experimental factors, including particle type, size, surface charge, zeta potential, surface coating, protein binding, dose, exposure route, species, cancer type, and tumor size can affect NP pharmacokinetics and tumor delivery efficiency. NPs can be absorbed with varying degrees following different exposure routes and mainly accumulate in liver and spleen, but also distribute to other tissues such as heart, lung, kidney and tumor tissues; and subsequently get metabolized and/or excreted mainly through hepatobiliary and renal elimination. Passive and active targeting strategies are the two major mechanisms of tumor delivery, while active targeting tends to have less toxicity and higher delivery efficiency through direct interaction between ligands and receptors. We also discuss challenges and perspectives remaining in the field of pharmacokinetics and tumor delivery efficiency of NPs.

Graphical Abstract

1. Introduction

Recent rapid advancements in nanotechnology have led to extensive research and wide applications of nanomaterials (NMs) or nanoparticles (NPs) in a number of areas, including consumer products, materials science, and medicine [1, 2]. NPs can be broadly categorized into organic, inorganic, and hybrid NPs. Organic NPs include polymeric, dendrimers, micelles, liposomes, fullerenes and ferritins to name a few [3], while inorganic NPs include gold, quantum dots, silver, etc. (Figure 1). Among all of the organic and inorganic NPs, gold nanoparticles (AuNPs) can be used to treat tumors [4, 5]; silver nanoparticles (AgNPs) can be used as antibacterial agents [6, 7]; iron oxide (FeO) NPs can be used for bioimaging, biosensing, and photothermal therapy; and dendrimers can be used as delivery or carrier systems for drugs and genes to treat various diseases, including cancers [8]. Hybrid NPs are defined as chemical conjugates of multiple organic and/or inorganic materials [9]. Common hybrid NPs include core-shell type hollow-polymer NPs and biomimetic NPs (Figure 1). Hybrid NPs are designed to achieve more complicated functions and take advantages of multiple materials. However, there are still some challenges related to NPs’ biomedical applications, especially the low delivery efficiency of NPs to targeted tumors [6, 10–12]. These wide applications of NPs have also raised concerns on their safety and potential toxicity to human health. A thorough understanding of the absorption, distribution, metabolism, and excretion (ADME) properties of NPs, the tumor delivery efficiency, especially the mechanisms that regulate these processes, is crucial.

Figure 1.

Examples of different nanoparticles that have been characterized for pharmacokinetic and biodistribution properties. Common inorganic nanoparticles include gold, iron oxide, silica and quantum dot nanoparticles, etc. Organic nanoparticles include polymeric, dendrimer, liposome, hydrogel, and micelle nanoparticles, etc. Hybrid nanoparticles contain various types including core-shell type, hollow-polymer and biomimetic nanoparticle type, etc. The three vertical dots on the right panel indicates et cetera. (This figure was created with BioRender.com)

Pharmacokinetics is the science of studying the interactions of drugs with the body in terms of ADME properties using mathematical and experimental methods [13]. Pharmacokinetic parameters that characterize the ADME properties of a drug or a substance include C_max (the maximum observed concentration of the drug collected in bodily material from the animal or human body), T_max (the time it takes to reach the maximum concentration C_max), AUC (area under the concentration curve that represents the total exposure of the drug experienced by the subject in a study), T_1/2 (half-life, which is the time it takes for the drug concentration to decrease by a half), MRT (mean residence time, which is the average time a drug molecule stays in the body), and clearance (the capacity for drug removal by various organs which is defined as the volume of blood from which all drug is removed per time unit). Common drug clearance pathways include renal, hepatic and biliary pathways [14].

In light of the importance of pharmacokinetics in the biomedical applications, safety and toxicity assessments of NPs, our laboratory previously reviewed the pharmacokinetics of different NPs, especially AuNPs [6, 13]. Recently we also reviewed the toxicokinetics, toxicological effects, dose-response relationship, and in vitro to in vivo extrapolation (IVIVE) involved in NP risk assessment [15, 16]. Over the last several years, many new pharmacokinetic studies of different NPs in both healthy and tumor-bearing animals have been published [10, 17]. A summary of representative recent pharmacokinetic or biodistribution studies of NPs in healthy and tumor-bearing animals is provided in Table 1 and Table 2, respectively. However, one key challenge remaining in this field is the low NP delivery efficiency of approximately only 0.7% injected dose (%ID) to the tumor site [6, 10, 13, 18]. The objectives of this review are: (1) to provide an update on the pharmacokinetic properties and their associated mechanisms of NPs in healthy animals; (2) to give an overview on the current state of NP tumor delivery studies and discuss their underlying mechanisms; and (3) to summarize the progress, existing challenges and share our future perspectives in this area.

Table 1.

Selected pharmacokinetic and biodistribution studies of nanoparticles in healthy animals.

References	Type, Size, Surface Coating	Model	Administration Dose (mg/kg)	Administration Method	Sampling Times	Organs/Tissues Examined	PK
Oral Absorption
[41]	Asiatic acid nanostructured lipid carrier (UP-AA-NLC) NPs, ~150 nm; Liposomes	Male Sprague-Dawley (SD) rats (200 ± 20 g), ICR mice (4–5 weeks)	64	Oral	0.5, 2, 4 h	Heart, Liver, Spleen, Lung, Kidney	No
[39]	Thymoquinone-loaded nanostructured lipid carrier; Liposomes	Male Sprague- Dawley rats (180–200 g)	100, 25	Oral, IV	0.25, 0.5, 1, 2, 4, 6, 8, 12, 24 h	Plasma, Blood, Intestine, Liver, Muscle, Kidney	Yes
[222]	Superparamagnetic iron oxide NPs (SPIONs), 20–100 nm	Rat/male	100	Oral	3, 6, 9, 24 h	Plasma, Brain	No
Pulmonary Absorption
[77]	Transferrin-coated gold NPs (Tf-Au NPs) ~25 nm or polyethylene glycol-coated gold NPs (PEG-Au NPs), ~50 nm	Rat/male	1	Pulmonary	24 h	Tracheobronchial lymph nodes, Lung, Kidney, Spleen, Liver	No
[76]	Doxorubicin-conjugated dendrimer, 56 kDa	Male Sprague-Dawley rats (270–320 g)	0.15 mg doxorubicin, 100 μL	Pulmonary	1, 3, 7 d	Plasma, Lung	Yes
[78]	Polyvinyl pyrrolidine (PVP)-coated Ag-NP (50 nm)	Female Wistar rats (200–250 g)	600 μg Ag50-PVP	Pulmonary	3, 21 d	Lung, Liver, Spleen	No
Others
[223]	Collagen-targeted peptide amphiphile (PA)-based NPs	Rat/male	7	IV	1, 2, 3, 6 d	Plasma, Heart, Lung, Liver, Kidney, Spleen	Yes
[224]	Transdermal microneedles (KNP/MN) NPs ~15 nm	Mouse/male & female	150	Transdermal	24, 48 h	Blood, Brain, Heart, Lung, Kidney, Liver, Spleen, Bladder, Intestine, Skin	No
[225]	Poly(ethylene glycol)-coated 15 nm gold NPs or 150 nm diameter silica core/gold NPs	Mouse/NA	0.1 mg Au/kg	IV	9.5, 14.5 d	Blood, Lung, Liver, Kidney, Spleen, Heart	No
[226]	Polyethyleneimine (PEI) and polyethylene glycol (PEG) gold NPs, 50–100 nm	Mouse/male	5 mg Au/kg	IV	1, 5, 15, 30 d	Blood, Liver, Spleen, Kidney, Heart, Brain	No
[227]	Self-assembling polymer nanoparticle (PNP) miRNA-mimics to PEG-peptide gold NPs	Mouse/male	85 nM/20 g	Subcutaneous	1, 4, 11 d	Liver, Kidney, Spleen, Lung, Skin	No

Note: PK, availability of pharmacokinetic parameters; Yes indicates that there is at least one pharmacokinetic parameter, such as the half-life and clearance mentioned in the manuscript; otherwise it is indicated as No; IV indicates that the administration dose is via intravenous injection; NA indicates the sex of animals is unavailable; the unit of the administration dose is mg/kg unless otherwise noted.

Table 2.

Selected pharmacokinetic and biodistribution studies of nanoparticles in tumor-bearing mice.

References	Type, Size, Surface Coating	Model	Administration Dose (mg/kg)	Tumor Type (Cell Line)	Sampling Time	Organs/Tissues Examined	PK
Passive targeting
[188]	Inorganic, Gadolinium metallofullerene-based activatable MRI contrast agent (denoted as RNP) encapsulating doxorubicin(DOX-RNP), ~150 nm	Mouse/NA	5	Cervical cancer (HeLa cells)	1, 4, 24, 48 h.	Blood, Tumor, Heart, Liver, Spleen, Lung, Kidney	No
[189]	Inorganic, Ti3C2 based plate shaped NP ~300 nm	Mouse/female	20	Breast cancer (MDA-MB-231 cells)	4, 24, 48 h	Heart, Liver, Spleen, Lung, Kidney, Tumor	No
[190]	Inorganic, polyethylene glycol capped silver nanoparticles doped with I-131 radionuclide (131I-doped Ag-PEG NPs) ~20 nm	Mouse/male	0.1	Sarcoma	0.25, 0.5, 1, 2 h	Blood, Bone, Muscle, Thyroid, Heart, Lung, Stomach, Intestine, Spleen, Liver, Kidney, Tumor	No
[191]	Organic, Liposome based Curcumin (PG-LipCUR) NP ~120 nm	Mouse/female	5	Breast cancer (4T1 cells)	1, 3, 6, 24 h	Plasma, Heart, Liver, Lung, Kidney, Tumor	Yes
[193]	Rod shaped chelators desferrioxamine B and 1,4,7-triazacyclononane-1,4,7-triacetic acid(NOTA) incorporated with far-red fluorescent dye Cy5 (NOTA-CNC-Cy5) NP ~100 nm	Mouse/female	1.5	Breast cancer (4T1 cells)	1, 4, 12, 24 h	Blood, Heart, Liver, Spleen, Lung, Kidney, Tumor	Yes
[194]	Organic, Non-PEGylated liposomal formulations containing Interferon-gamma (IFN-γ) NP ~120 nm	Mouse/female	0.1 uci/g	Colon cancer (C26 cells)	6, 12, 24, 48 h	Blood, Heart, Liver, Spleen, Lung, Kidney, Tail, Muscle, Bone, Thyroid, Stomach, Intestine, Tumor	No
Active targeting
[192]	Inorganic, Co-encapsulation of plasmonic gold nanorods (GNRs) NP, ~150 nm, targeting ligand: PLGA	Mouse/female	100 mci/100 μL	Ehrlich ascites carcinoma (Ehrlich cells)	2, 4, 24 h	Heart, Lung, Liver, Stomach, intestine, Kidney, Spleen, Thyroid, Muscle, Tumor	No
[195]	Inorganic, Copper based pyridinium (Py)-modified hollow mesoporous silica nanoparticles-based drug reservoir (HMSN-Py) NP, ~250 nm, targeting ligand: folic acid-conjugated polyethylene glycol	Mouse/female	5	Cervical cancer (HeLa cells)	1, 4, 24 h	Heart, Liver, Spleen, Lung, Kidney, Tumor	No
[196]	Organic, DOX-loaded Camellia oleifera protein NP, ~65 nm, targeting ligand: Camellia oleifera protein	Mouse/male	4	Liver cancer (H22 cells)	1, 4, 12, 24, 48 h	Blood, Heart, Liver, Spleen, Lung, Kidney, Tumor	No
[197]	Organic, DOX-loaded HA-photosensitizer conjugate containing reactive oxygen species-sensitive thioketal linkers (TKHCENPDOX), ~90 nm, targeting ligand: hyaluronic acid	Mouse/female	10	Breast cancer (MDA-MB-231 cells)	4, 12, 24, 48 h	Blood, Heart, Liver, Spleen, Lung, Kidney, Tumor	Yes

Note: The administration route of all listed studies is via intravenous injection. NA, information on the sex of the animals is not available; PK, availability of pharmacokinetic parameters; Yes, at least one pharmacokinetic parameter provided in the manuscript. No, no pharmacokinetic parameters are available in the manuscript.

In line with the objectives of this review article, the manuscript is categorized according to different sections: (1) absorption, (2) distribution, (3) metabolism, (4) excretion, (5) tumor delivery, (6) limitations and challenges, and (7) conclusion and future perspectives. Within each section, we review the state-of-the-art and provide some examples on each topic. In the section of Absorption, we introduce absorption processes following three main routes of administration, including gastrointestinal, dermal, and pulmonary routes. We also discuss some potential strategies to efficiently enhance NP uptake via surface modification. In the last two sections, we discuss the limitations and challenges in this field, summarize existing knowledge about ADME and tumor delivery of NPs, and share our future perspectives in the field.

2. Absorption

2.1. Gastrointestinal absorption

Oral drug delivery is the most common route for drug administration in preclinical and clinical use due to its non-invasive nature and convenience compared to other administration routes [19]. The challenges to oral delivery are mainly the hostile gastric and intestinal environments for bioactive drugs like peptides and labile compounds. NPs can be designed to persist in the intestinal lumen long enough to attach to the cell apical surface and then be transported across the intestinal cell membrane via P-glycoprotein transporters [20].

As depicted in Figure 2, the process of NP transit following oral exposure starts from the intake of NPs and continue as they travel through the gastrointestinal tract where NPs may enter into the systemic circulation via the hepatic portal system [21]. Translocation through the epithelium of the villi is a multistep process, including diffusion through the muscus layer, interaction with enterocytes and/or M cells, and uptake via cellular entry or paracellular transport [22]. The most common mechanism for NP uptake into intestinal epithelial cells is endocytosis, which could be caveolae-mediated, clathrin-mediated, caveolae and clathrin-independent, and micropinocytosis [23].

Figure 2.

Schematic diagram of the relationship between an oral dose of nanoparticles and their ultimate fate. After intake of nanoparticles orally, they move through stomach, intestine and intestinal lymphatic system for further distribution. Epithelium of villi is mainly constituted of enterocytes, goblet cells and M cells. Goblet cells can secrete the mucus gel layer. M cells play an important part in lymphoid follicle system, and they are mainly located within follicle associated epithelium in particular transportation of nanoparticles. (This figure was created with BioRender.com)

The absorption efficiency of NPs via the oral route is affected by both biological features of gastrointestinal tract, such as pH, and physicochemical properties of NPs, such as particle size, morphology, surface chemistry, adhesive property of the NP, permeation enhancements, etc. [20, 24, 25]. For example, the pH variance in the gastrointestinal tract can highly affect NP aggregation and alter NP surface charge. The pH environments in the gastrointestinal tract can range from 1 in the stomach to 8 in parts of the intestine. Many pH-responsive NPs were designed to take advantage of this effect [26–28]. In terms of particle size, small NPs (typically < 50 nm) may access regular epithelial cells [24], whereas large NPs (~50–200 nm) are taken up via microfold cells (M cells) of organized lymphoid follicles [25]. For core materials of NPs, it is suggested that some specific materials can enhance oral bioavailability, which include, but is not limited to chitosan, poly(acrylic acid) (PAA), poly(ethylene glycol) (PEG), poly(lactide-co-glycolide) (PLGA), lectins, and dendrimers [29, 30]. Regarding the NP morphology, dendrimers are suggested to be able to enhance the oral absorption. Besides, surface charge can play a crucial role on diffusion through the intestinal mucus layer. Compared to negative-charged NPs, positive and neutral NPs can easily penetrate the mucus layer because of the prevention of mucoadhesion [23]. Accordingly, strategies of increasing oral bioavailability are different through three different segments of the gastrointestinal tract, including stomach, small intestine, and the intestinal lymphatic system. For example, NP formulations can be designed to contain materials that are mucoadhesive to increase gastric retention, thereby increasing the oral bioavailability through the stomach [19]. Polymer NPs with enteric coatings can be used to resist the release of the drug from NPs and stay intact in the stomach at a low pH, but the enteric polymer NPs will dissolve and release the drug from the NPs in the small intestine when the pH increases to 5.0. In this way, enteric polymer coated NPs could target drug to small intestine absorption sites [31–36]. The intestinal lymphatic system can serve as an alternative pathway to increase oral delivery of NPs through two major targets: the chylomicrons in enterocytes and the microfold cells (M cells) in Peyer’s patches [19, 37, 38]. For example, Ansar et al. designed an oral NP (thymoquinone-loaded nanostructured lipid carrier; TQ-NLC) of which the movement through the intestine lymphatic system can escape from the first metabolism and sequentially increase the relative bioavailability (C_max and AUC) in rats [39] (Table 1). In non-lymphoid areas, studies found that the NPs conjugated with nutrients or nutrient-like compounds can be absorbed without or with less lymphoid detection. For example, Florence et al. designed a tomato lectin-conjugated polystyrene beads of which the intestine uptake was significantly increased than plain polystyrene (12.8% vs. 4.2%) [40].

Ursodeoxycholic acid, a main component in bile acids which can be absorbed from the intestine, was used as surface modification strategy to increase the bioavailability of asiatic acid in cancer fibrosis treatment [41] (Table 1). In addition, for dendrimers, two potential mechanisms by which oral absorption could be enhanced include: (1) alteration of the barrier function of intestinal epithelium by dendrimers and thus increasing the permeability for the co-administered drug with the dendrimer acting as a “recipient”; and (2) the dendrimer complex itself can enhance the permeability and cross the intestinal epithelium [30, 42]. Herein, polyamidoamine (PAMAM) dendrimers were most extensively studied in oral delivery. An in vitro study using Caco-2 cells suggested that both positive- and negative-charged dendrimers can be internalized as short as 20 min [43].

Taking advantages of the characteristics that increase oral absorption mentioned above, some oral-delivered NPs were developed for various cancer treatments. Lee et al. developed a pH-responsive organic-inorganic hybrid nanocomposite (EGAC-BSA: bovine serum albumin with 3-aminopropyl functionalized magnesium phyllosilicate [AC-BSA] coated with the pH-sensitive polymer Eudragit^®L100–55) as an effective oral delivery system for protein drugs [44]. Clathrin-mediated endocytosis and paracellular transport were the major pathways in oral absorption via M cells. Compared to rapid drug release of 80% for AC-BSA, the coated NPs of EGAC-BSA showed a much slower drug release of 15% at a low pH of 1.2, which indicated that the NPs succeeded in protecting the protein drug against the harsh gastrointestinal environment compared to the free unbound protein drug. Up to a 2.1–3.8 fold higher cellular uptake in human colon adenocarcinoma cells in vitro and significantly enhanced intestinal permeation of drug absorption compared to free drug indicated that the organic-inorganic hybrid nanoparticle EGAC-BSA is a promising and effective oral delivery system [44]. In another instance, Tariq et al. developed biodegradable epirubicin loaded poly-lactic-co-glycolic acid nanoparticles (EPI-NPs) (235 nm, −27.5 mV), which showed enhanced endocytic update across enterocytes and M cells by 3.9-fold improvement in oral bioavailability compared to EPI drug itself [45]. Overall, absorption via M cells to the lymphatic system following oral administration is one of the potential pathways for oral delivery of NPs.

2.2. Dermal absorption

Skin is the largest organ of the body and plays an essential role in protection as a barrier against outer environment insults, and in homeostasis of internal cells, metabolism and deposition. Absorption through the skin is a common route of exposure to NPs but not an efficient site for NP delivery due to very low absorption [46–48]. The primary barrier for dermal absorption of chemicals or NPs is the stratum corneum [49]. Intercellular, transcellular, and transappendageal pathways through hair follicles and sweat glands are four basic pathways that NPs can potentially penetrate through the skin [50]. Figure 3 shows a schematic of dermal absorption of NPs through different pathways. In brief, NPs with diameters less than 10 nm are physically able to penetrate stratum corneum through diffusion, intercellular or transcellular route to reach dermis [51]. While physically penetrable, NPs <10 nm will not be dermally absorbed and enter into blood circulation in part phagocyte engulfment will hardly occur [46, 47]. For example, Ryman-Rasmussen et al. compared the pig skin penetration for the quantum dots (core/shell diameters ~10 nm for major axis, hydrodynamic diameter <50 nm) with different core/shell sizes, shapes, and surface coatings, and found that spherical 4.6 nm core/shell diameter quantum dots 565 and ellipsoid 12 × 6 nm core/shell quantum dots 655 with either neutral (polyethylene glycol) or cationic (polyethylene glycol-amine) coatings can penetrate the stratum corneum and be localized within the epidermal and dermal layers by 8 h after dermal exposure, but it required 24 h for anionic carboxylic acid-coated QD 655 to penetrate the stratum corneum [52]. Their study suggests that the stratum corneum layer of the skin is permeable to smaller sizes (<10 nm) of NPs with different physicochemical properties. This size limitation of skin stratum corneum penetration sometimes can be extended to 20 nm, such as iron-based particles [53].

Figure 3.

Overview of nanoparticle dermal absorption depending on size and other physicochemical properties. Diameters less than 10 nm are able to penetrate stratum corneum and reach to dermis by intercellular or transcellular route. Diameters in 20–40 nm are able to to reach to dermis via transappendageal route. Diameters in 40–100 nm are not able to penetrate stratum corneum and only few can reach to dermis. Two transappendageal pathways include a. hair follicles and b. sweat glands. Other two pathways that NPs penetrate through the skin include c. intercellular and d. transcellular. (This figure was created with BioRender.com)

When NPs’ diameters exceed 20 nm, transappendageal routes such as hair follicles and aqueous pores comes into play. It was shown that 40 nm NPs, but not 750 or 1500 nm NPs can penetrate deeply into vellus hair openings and through the follicular epithelium [54]. The rigid hair shaft can act as a geared pump to move NPs deeper into the hair follicle. It was reported that the hair follicle infundibulum and canal has a 10 times longer storage of NPs within the hair follicles than the storage reservoir capacity of the stratum corneum [55]. Even though the hair follicle is a good storage site for NPs, it is not a good site for absorption of NPs into the systemic circulation. When NPs’ diameters ae larger than 40 nm, they do not easily penetrate stratum corneum. For example, titanium dioxide NPs (primary size 10 × 50 nm with mean agglomerates of 200 nm) and zinc oxide NPs (mean size of 140 nm) in different sunscreen formulations can only penetrate the stratum corneum to reach the epidermal layer of the skin, but no titanium or zinc was detected in the skin perfusate samples, indicating minimal transdermal absorption [56]. On the other hand, larger NPs may reach the dermis layer through transappendageal routes [52, 54, 57, 58]. It was shown that NPs in the size range of 300–400 nm had an efficient storage behavior of up to 10 days in the hair follicles after dermal exposure [55, 59].

The influence of the dermal absorption of NPs can be multifactorial [51]. Besides the size discussed above, the solvents, the location and skin condition at the application site can all influence dermal absorption. The skin condition is influenced by the skin integrity, dimensions of orifices, aqueous pores, lipidic fluid paths, density of appendages, and regional variations [49, 60, 61]. A number of in vitro and in vivo studies have been carried to investigate the penetration of NPs through normal or damaged skin [56, 62–65]. For example, in the pig skin penetration study of titanium dioxide and zinc oxide NPs mentioned above, the authors further showed that UVB-damaged skin had slightly enhanced penetration of titanium dioxide and zinc oxide NPs in sunscreen formulations, but no transdermal absorption was detected [56]. However, not all skin barrier perturbation will increase penetration. Zhang and Monteiro-Riviere [65] compared the penetration of QD565 (14 nm) and QD655 (18 nm) among healthy, flexed, tape-stripped, and abraded rat skin and found that QD565 and QD655 only penetrate the stratum corneum in healthy and tape-stripped skin, but QD655 can penetrate into the viable dermal layers of abraded rat skin at 8 and 24 h after dermal exposure. This study demonstrated that barrier perturbation by tape stripping does not cause penetration, but abrasion enable quantum dots to penetrate deeper into the dermal layers.

In addition, NPs’ own physicochemical properties including surface coating, zeta potential, shape, binding affinity with macromolecules, metabolism, and diffusion coefficient are also influential in dermal absorption [52, 66–68]. Furthermore, physicochemical properties of the NPs dispersing a penetrating molecule also play a role in influencing dermal absorption. Synergisms and interactions exist between dispersed NPs and skin components [58]. When a formulation of drug containing NPs are applied to skin, all of the ingredients will trigger a direct contact to the skin area. Volatility, skin sweat sebum components and NP physicochemical properties on different combinations will cause synergism and interactions via different circumstances [58]. The diffusion coefficient of the molecules in the vehicle can be reduced by the viscous formulations, thus blocking or retarding the skin partitioning, and as a result, influencing dermal absorption. For example, extreme lipophilic formulations will compete with stratum corneum lipophilicity, reducing the partitioning of the NPs from formulation into stratum corneum lipids, thereby reducing absorption rate. In contrast, occlusive hydrophilic formulations might increase the absorption rate via increasing partitioning and altering or damaging stratum corneum as shown in both in vitro and in vivo studies [58, 69, 70]. Due to a high proportion of lipid content, the stratum corneum is naturally negatively charged. As a result, typically positive-charged NPs have a higher penetration capacity through the stratum corneum [50].

Wang et al. reported that the imidazole-based, ionic liquid microemulsions formulated as organic polymeric NPs (~ 40 nm) can alter stratum corneum fluidity to facilitate penetration of the active drug ingredient released from the nanoformulation [71]. In this study, the tight arrangements of keratinocytes were disrupted by the NPs as the stratum corneum properties differed after interacting with NPs when delivering the drug dencichine (Den). In vitro experiments showed that the drug was absorbed approximately 10 times more in the NP formulation than in traditional aqueous solution. No significant toxicity and skin irritation were observed in this study. The authors concluded that ionic liquid microemulsion, as one of the heterogenous spherical shapes of non-aggregated droplets with sizes less than 100 nm organic NPs, can be a promising and desirable treatment in dermal absorption through increasing the stratum corneum fluidity to facilitate absorption of the active drug ingredient release from the NPs [71].

2.3. Pulmonary absorption

Lungs are a vital and attractive target for the pulmonary administration of NPs in animals and humans [72, 73]. This route, compared to conventional oral administration, offers many advantages including high surface area and rapid absorption with high vascularization and recirculation, better permeability, and limited proteolytic activity [74, 75]. In addition, lungs are also responsible for oxygen exchange to all of the cells. Multiple aerosolized NPs were designed in the pre-clinical or clinical stages for not only tumors but also many respiratory diseases, and some of them have been approved for marketing (Table 1) [75–79]. In humans, the lung consists of five lobes with the right lung consisting of three and the left two. Bronchi and smaller air passages, alveoli, and lymph tissues compose the remainder of the lungs. The blood-gas barrier thickness in the mouse (0.32 μm) is similar to that of the rat (0.38 μm) but smaller than humans (0.62 μm) [80], suggesting increased permeability in these rodents. The deposition of NPs can be described by four different mechanisms by which NPs deposit in respiratory tract based on the particle size, interception, impaction, sedimentation and Brownian diffusion [81, 82]. In addition, biomolecular coronas are of great importance in pulmonary deposition. Some of the lipid NPs will directly interact with biological fluids from lung surfactants when inhaled. Proteins and lipids, the major biomolecules in these fluids, can adsorb onto the lipid NP surface, forming a layer termed the “biomolecular corona” that defines the new biological identity of the NPs and alters the subsequent biodistribution [83, 84].

When aerosols are being inhaled, particles deposit through most of the airways. This process is controlled and dominated by inertia impaction. Airflow reaches the maximum at the trachea and decreases when air passes deeper into lungs. When the lungs expand their elasticity, upper airways’ geometry would be changed accordingly. The carinii bifurcation represents the main impact site for the aerosol being transported into other organs after being absorbed. NPs that are not deposited in the upper airways will continue to be transported through the airway depending on residual gravitational settling force. Extremely small particles rely on Brownian motion and random-walk diffusional process on the walls in the periphery of lung [82]. The pulmonary absorption bioavailability is dictated not only by NP size but also by structural and physicochemical properties including molecular weight, charge, ionization, and presence or absence of hydrophobic regions. The details of physicochemical effects on NP aerosol deposition were summarized in a previous review [75]. NP aerosols may be modulated by molecular dispersion depending on the forms of solution or solid states [82, 85]. Afterwards, mucociliary clearance and transport takes place with the movement of mucus removing particles from the airways. The mucociliary clearance takes place under the assistance of coordinated ciliary beat and epithelial cells from the airway at the end of esophagus. For those particles or droplets that are inhaled deeply into lungs, they will be taken up by macrophages and transport to lymph nodes. The full process is depicted in Figure 4.

Figure 4.

A. Lung structure in human beings. The lung consists of five lobes in total with the right lung consisting of three and the left two. B. Lung structure in mice (similar with rats). The lung consists of five lobes in total with the right lung consisting of four and the left one. C. Different mechanisms in pulmonary absorption. Major absorption mechanisms are (a) inertial impaction, (b) interception, (c) sedimentation, and (d) diffusion. Then nanoparticles go through mucociliary and cell mediated transport and then get absorbed into blood and organ systems. (This figure was created with BioRender.com)

Examples of NP charges and properties that may affect absorption are illustrated in the following case. Rosiere et al. developed a new folate-grafted copolymer of polyethylene glycol (PEG) and chitosan, F-PEG-HTCC in solid lipid nanoparticles (SLN) of approximately 200–300 nm. The dose related to 1 mg/kg chemotherapy drug paclitaxel was administered to female BALB/c mice by administering through the noninvasive endotracheal route using an aerosolizing system (Microsprayer model IA-1C, Penn-Century, Philadelphia, USA) in treating lung tumors. They found that pulmonary exposure to paclitaxel (from the coated SLN) was prolonged up to 6 h and greatly limited systemic distribution. Compared to inhalation of Taxol itself, the F-PEG-HTCC coated SLN led to 7-fold and 32-fold increased concentrations at 1 and 6 h after exposure, respectively. The positively charged liposome NPs greatly enhanced lung mucosa absorption. In addition, F-PEG-HTCC-coated SLN showed increased antitumor activity and decreased pulmonary adverse effects compared to conventional formulations following inhalational exposure [86].

3. Distribution

Upon absorption following different routes of exposure, NPs will circulate in the bloodstream and then be distributed to different organs and tissues. One of the major challenges of tissue distribution of NPs is a short blood circulation time, thus low distribution to the target organ, including tumors [10, 11]. For example, liposome NPs consisting of an aqueous cavity surrounded by one or several phospholipid bilayer membranes, are versatile since they once encapsulated may contain both hydrophilic and hydrophobic compounds [87]. However, the biocompatibility, biodegradability and structure of encapsulating compounds with different characteristics may still result in short blood circulation times with rapid clearance, even in a few minutes, which is a major drawback seen with liposome delivery vehicles [88, 89]. To overcome this drawback, polyethylene glycol (PEG) was added as surface coatings to extend blood-circulation half-life up to a few hours. Allen et al. successfully extended blood circulation half-life of liposome PEG derivative from 2 h up to 20 h, showing a great potential in altering half-lives via PEG modifications [90, 91]. Opsonization exists in liposomes [92, 93], while PEGylated liposomes with a mean diameter of 100–150 nm have decreased opsonization and phagocytosis by the reticuloendothelial system (RES). Reduction of the particle size and creating surface modification with a layer of amphiphilic polymer chains such as PEG can greatly reduce opsonization and clearance by the RES system [88, 91, 94–96].

Tissue distribution is the process of NPs’ transport from blood to interstitial fluids and cells in different organs, tumors, other tissues. As described elsewhere [6, 97–99], for both inorganic and organic NPs, size, surface modifications, surface charge, opsonizations, blood flow, vascular permeability, plasma protein corona content, mononuclear phagocytic system, complement system and interaction of NPs with tissue cells could modulate the extent of biodistribution. When NPs are absorbed, the majority of NPs would be diffused or through endocytosis to enter the bloodstream and then distributed to organs and tissues. Distribution of NPs to the target organ is important in maximizing the therapeutic functionality and minimizing the side effects in both in clinical diagnosis and therapeutics [100].

When NPs are in the bloodstream, the process of opsonization, one of the most important biological barriers in controlling drug delivery, will happen. This process is called biomolecular corona formation. Biomolecular corona formation is an important factor that can alter the NP physicochemical properties, trigger conformational changes in both NPs and the adsorbed macromolecules especially proteins, thereby changing the pharmacokinetic and biodistribution profiles of NPs in the body [101–106]. Among different types of biomolecular coronas, protein coronas are the most extensively studied. The composition and structural organization of the adsorbed proteins are important in determining the fate of the NP-protein complexes. Opsonin proteins include immunoglobulins and components of the complement system such as C3, C4, and C5, other blood serum proteins such as laminin, fibronectin and type I collagen [15, 107–109]. They will bind to the NPs, which allows macrophages of the RES to easily recognize and remove these “foreign” bodies. Taking polymeric NPs as an example, whenever non-biodegradable, these NPs will accumulate in liver and spleen, thereby potentially eliciting toxicity and other side effects [110–112]. The bound opsonin proteins may undergo a conformational transition from inactive to active, and when presented in the blood stream may interact with specialized receptors from phagocytes alerting the presence of foreign material. After opsonization has occurred, the next step of their clearance from the bloodstream for further processing is the attachment of the phagocyte to the NPs via the surface bound opsonins. In addition, non-specific adherence of phagocytes to the cell surface could occur. This process is typically due to the attachment of the opsonin proteins with more hydrophobic NP surface [113–116].

Using protein corona fingerprinting approach, Walkey et al. [117] characterized the identity and quantity of each adsorbed protein on the surface of a library of 105 surface-modified AuNPs. The authors also applied bioinformatic methods to develop a multivariate linear model that uses the protein corona fingerprint to predict cell association 50% more accurately than a traditional model that uses physicochemical parameters, such as size, aggregation state, and surface charge. More recently, another research group integrated machine learning and meta-analysis approaches to build a quantitative model that can predict the functional composition of the protein corona and the cell recognition of the NPs (with a determination coefficient of R² > 0.80) [118].

An important factor regarding protein corona formation is time-dependence. The surface of NPs and the composition of protein coronas will change with time during circulation in the vascular system, which will in turn impact the composition of protein coronas and the subsequent fate of the NPs. Sahneh et al. [119] developed a pharmacokinetic model to simulate the impact of biocorona formation kinetics on interspecies extrapolations of NP biodistribution, and found that different blood circulation time scales between different species contribute to differences in NP-project corona formation kinetics, which in turn causes differences in pharmacokinetics and tissue distribution of NPs between rodents and humans, in part because of a longer blood circulation time in humans allows for further biocorona evolution. Lazarovits et al. [120] developed a supervised deep learning model that can predict NP biological fate (i.e., percentage of distribution to different organs) by considering the time-dependence using protein composition data extracted from NP-protein coronas collected at different time points following intravenous injection. Their study provides possibilities to both predict NPs’ biodistribution and also to engineer surface chemistries that are specifically designed to enhanced targeted delivery.

Another important factor that substantially alters NP-protein corona composition is healthy versus diseased state of an individual. Multiple studies have shown that NP-protein corona compositions in serum from diseased animals or humans were different from those in normal serum, resulting in different biodistribution profiles of NPs between healthy versus diseased subjects [121, 122]. For example, it was report that albumin and alpha-2-macroglobulin concentrations were much lower in non-small cell lung cancer patient plasma than healthy people, and the result was similar in the mouse [121].

After the binding and phagocytosis of the foreign NPs, endocytosis and exocytosis of NPs will take place. NPs are able to transfer in between cells via endocytosis and exocytosis. The secretion of the enzymes and other oxidative-reactive chemical factors including nitric oxide, hydrogen peroxide and superoxide will degrade the phagocytosed material during endocytosis and exocytosis [123]. The whole process and a schematic view are in Figure 5.

Figure 5.

The opsonization, phagocytosis, endocytosis and exocytosis of nanoparticles in the distribution process. The bound of opsonin proteins will undergo conformational changes and be activated. Then phagocytosis will take place. Endocytosis and exocytosis will be involved into transporting the nanoparticles to other organs. Nanoparticle characteristics to activate complement normally include particle size, zeta potential and surface coating, etc. (This figure was created with BioRender.com)

In terms of tissue distribution of NPs, many studies have consistently reported that liver and spleen are the two major organs where NPs are distributed to and accumulated in following systemic exposure. Since a large portion of NPs accumulate in the liver and spleen, this greatly affects its drug delivery efficacy and safety to the target tissue, such as tumor and other organs, especially when NPs are not loaded with drugs. The nonspecific accumulation of empty nanocarriers in the body not only decreases the pharmacological effects but also raises safety issues. A study showed that a high concentration of empty nanocarriers inhibited the endo-lysosomal transport process and reduced escape of the protein delivered by the NPs, thus reducing the protein transport efficiency of mesoporous silica NPs [124]. Therefore, the potential accumulation and elimination of the NPs after release of its cargo should attract greater attention. Studies have demonstrated that NPs accumulate to a very low degree in mouse kidneys and lungs [125, 126]. Liang et al. investigated biodistribution of long circulating quantum dot NPs and found that at 24-h after injection, the accumulation was relatively low in spleen, lung and kidney (0.64%, 0.46% and 2.03%), while liver had the highest amount of accumulation (22.01 % and 46.96% at day 1 and day 7) among all studied organs [125].

Besides distribution to major RES organs, NPs can also be distributed to and release from the brain. Kumar et al. summarized the temporal organ distribution of NPs from 116 studies and found that NP concentration achieved the peak within 72 hours in the most organs, such as muscle, stomach, brain, kidney, and heart [98]. Unlike other organs, the NP drainage from brain is limited by blood-brain barrier as only a small proportion of NPs can be eliminated via the blood-brain barrier [127]. Liu et al. reported an elimination pathway through glymphatic system in which NPs (AuNPs as an example) can be transported from brain parenchyma to the cerebrospinal fluid, and subsequently eliminated to RES [128].

For NPs that carry a drug, modification of the surface properties of the NPs can alter the biodistribution and tumor delivery of the NPs. Ekdawi et al. performed a spatial-temporal quantification of tumor macro- and micro-distribution of liposomes in an orthotopic xenograft mouse model [129]. For liposome-based drug formulations, combining with the heterogeneous nature of the tumor microenvironment diminishes in vivo efficacy as a result of transport-limiting properties. Ekdawi et al. performed biodistribution studies in SCID mice by orthotopic implantation. They combined the liposome-based drug NGR-PEG3400-DSPE, near-infrared (NIR) fluorescent dye, Genhance^™ 680 (GH680) and incorporated them into the liposome NPs (1.22 mg lipid/g mouse; 0.7 mg I per g mouse; 2 μg GH680 per g mouse). The biodistribution results were collected up to 120 h post injection. CT scan showed that in normal tissues, spleen and liver were two largest organs of distribution (25.2 ± 3.14 and 23.4 ± 3.21% ID/cm³, respectively, at 72 h), with kidneys being the next organ with a high distribution (7–8% ID/cm³ at 72 h, n = 4). As for tumor accumulation of the liposomes, an increased uptake was shown until up to 72 h followed by a gradual decrease as time goes on (4.0 ± 0.59% ID/cm³ at 72 h, n = 4). Overall, the modifications of liposomal encapsulations combined with increased vascular permeability or density could lead to increase in NP tumor delivery efficiency. Additionally, another useful strategy to increase tumor delivery of NPs is surface coating with zwitterionic polymers, such as PLGA, poly(2-methacryloyloxyethyl phosphorylcholine) (PMPC), and polydopamine (PDA), which were shown to extend the blood circulation time of NPs and enhance NP distribution to tumors [130–132].

Another unique property that significantly affects biodistribution of NPs is agglomeration and aggregation [15]. Agglomeration refers to the phenomenon that NPs are associated into a loose cluster that can be broken down relatively easily through mechanical forces, and thus is a reversible process. When NPs are associated with each other tightly forming a strong, dense, and stable particle collective that is irreversible, this process is termed aggregation. Agglomeration and aggregation can occur before entering into the systemic circulation (i.e., in the NPs suspension or in the gastrointestinal tract) or within the systemic circulation. When agglomeration or aggregation takes place, the size of the NPs increases to various degrees depending on the extent of association of NPs. In general, following extravascular administration, increased size of NPs due to either agglomeration or aggregation will result in less absorption due to reduced permeability across cell membrane. When NPs are in the blood circulation, increased size of NPs may enhance uptake of NPs by phagocytic cells to different tissues, thereby changing the half-life of NPs in the plasma [133].

4. Metabolism

Metabolism is a process of NPs being processed or altered in the body, resulting in a change in the molecular structure, composition, and/or physicochemical NP properties through degradation, dissolution, drug release, aggregation, or opsonization [15]. The pathways and the extent of metabolism of NPs depend on the type of NPs (inorganic vs. organic) and the physicochemical properties of the NPs. Most of the inorganic NPs such as gold, quantum dots, and silica are relatively stable in the body with minimum or negligible metabolism [125, 134]. However, for silver NPs, they can be metabolized to become silver ion (Ag⁺) especially in acidic environments (e.g., in gastric fluid), which precipitates with Cl⁻, leading to particle growth and facilitating particle aggregation by decreasing their electrostatic repulsion in the solution, and this process is pH-dependent [135]. On the other hand, organic NPs can often be degraded or metabolized in the body. These organic NPs like liposomes, solid lipid NPs, micelles, nanogels, and dendrimers can be biodegraded into hydrophilic small molecule metabolites for further excretion through urine or bile [91, 97]. Also, single-walled carbon nanotubes can be biodegraded by human neutrophil enzyme myeloperoxidase in neutrophils and to a lesser degree in macrophages [136]. Of note, it was shown that the biodegraded nanotubes did not cause an inflammatory response when aspirated into the lungs of mice, suggesting that the extent of metabolism or biodegradation of carbon nanotubes may be a major determinant of the scale and severity of the associated inflammatory toxicity in humans [136].

Once NPs are in the blood circulation, most NPs less than the glomerular filtration size limit (~5.5 nm) will undergo renal excretion by the kidney and leave the body via the urine [137]. Biodegradable NPs with size more than 5.5 nm in diameter can interact with the negative charged glomerular capillary wall (with NPs’ surface charge), a portion of these NPs would be able to enter the Bowman’s cavity, pass through the glomerular capillaries and get reabsorbed by renal tubule and return to the blood [138–141]. Non-biodegradable and yet large size NPs (NP size larger than 5.5 nm in diameter), however, may be taken up by phagocytes of the RES system and retain in the body for a relatively longer time (i.e., up to several months) [142]. The long retention time is due to the complexity of metabolism and excretion processes: NPs will undergo hepatobiliary elimination via transcytosis through hepatocytes in the liver, resulting in transport via the bile canaliculi into the biliary system, then into the gastrointestinal tract with eventual elimination in feces. Sometimes the whole process will even extend from several months to years [14, 143–145]. In recent years, studies have shown that NPs larger than the kidney filtration threshold of ~5.5 nm can be found intact in the urine following injection in mice [146]. The exact mechanism remains to be elucidated, but it has been suggested that larger sizes of NPs may bypass the glomerular filtration barrier through the proximal convoluted tubules to be excreted out [146]. It is worth noting that, if the NP size is too large (more than 300–500 nm), they will still undergo the disassembling and breakdown process (normally happen in blood circulation) to be smaller than the liver sinusoidal endothelia cell fenestrae filtration size limit to gain access via fenestrae into the spaces of Disse [14]. A flowchart of NPs’ fate including metabolism and excretion is illustrated in Figure 6.

Figure 6.

A Flowchart of nanoparticle fate of metabolism and excretion. A. Nanoparticles with the size larger than 5.5 nm will go through hepatobiliary process. Nanoparticles enter the liver via the portal vein, then traverse the hepatic sinusoid or be sequestered in liver resident Kupffer cells. After this process, nanoparticles will then be filtered out into the space of Disse and endocytosed by hepatocytes. Nanoparticles will then transcytose through the hepatocytes and enter through bile ducts. After entering through the bile ducts, nanoparticles will pass through gallbladder or enter into the common bile duct, then be excreted into the duodenum of the small intestines. Eventually, nanoparticles will pass through the gastrointestinal tract and be eliminated into feces. B. Nanoparticles with the size smaller than 5.5 nm will go through renal process and be eliminated by kidney into urine. Nanoparticles will enter the kidney via the portal vein, then be filtered out into glomerulus and travel from limb of loop of Henle to collecting tube, then be eliminated into urine. (This figure was created with BioRender.com)

For non-biodegradable NPs, a classic example would be from Kolosnjaj-Tabi et al. in determination of one-year fate on iron oxide-coated AuNPs in mice [147]. The 5-nm diameter gold core was coated with iron oxide and formed into a core-shell morphology, with the hybrid NP complex reaching a diameter of 16 nm after coating. After intravenous injection of 50 μmol iron/kg of NPs into the pathogen-free female 8-week-old C57BL/6 mice, the authors observed the iron oxide shell of PEG-coated hybrid NPs was degraded and removed from both liver and spleen, while the amphiphilic polymer-coated NPs retained up to 10% of the initial iron oxide dose in these organs. Also, NPs were absorbed by intracellular vesicles of liver Kupffer cells and still detected after one year of intravenous administration. The study demonstrates that inorganic NPs may remain for quite a long period of time in part because of the non-biodegradable property.

Some of the inorganic NPs like degradable silver NPs could react with different biomolecules, resulting in oxidation, reduction, hydrolysis and/or binding reactions, thereby being dissolved and releasing silver ion or be transformed into silver sulfide [6, 148, 149]. The complete metabolism process of silver NPs is presented in our earlier review article [6].

For biodegradable NPs (usually organic NPs), chemical compositions and physicochemical properties are major factors to consider. Neutral and hydrophilic materials are effective in evading macrophage uptake and can rapidly degrade into metabolites, while some other conventional long chain moieties would not degrade easily and thus remain in the body [150]. Ma et al. orally administered the lipid-based nano-vehicle micelles using male Sprague Dawley rats to study the micelle NP fate. Fluorescence labeled mixed micelle (MM) NPs (0.8 mL) were administered and the MMs survived the gastrointestinal environment with a relatively high integrity for about 4 h. Thereafter, a relatively high amount of MM NPs was accumulated in the intestinal villi surface, and a relatively low amount of these NPs stayed in the basolateral tissue. Meanwhile, beginning at 2 h after administration and lasting for about 12 h, MM NPs were detected in the liver with fluorescence signals which peaked at 8 h, while other major organs including lung, spleen and kidney had no fluorescence signals detected, indicating the major organ for MMs deposition was liver. In addition, in vitro evaluation showed that during the transport of MMs across the Caco-2 and Caco-2/HT29-MTX cell monolayers, micelle NPs could be easily broken down within epithelial cells and release the drug, which makes up a majority of bioavailable drug concentration [151].

5. Excretion

There are differences between biodegradable and nonbiodegradable NPs in terms of excretion. Biodegradable NPs could be broken down into their constituent materials and undergo excretion as small molecules [14]. There are normally three major routes for excretion for small molecules, including urine, feces, and exhaled air. Other minor excretory routes for small molecules generally include milk, hair, tears, saliva, and nails. As for NPs, renal and hepatobiliary excretion are two main pathways for intravenously administered NPs [145]. Some NPs can go through hepatobiliary elimination via transcytosis through hepatocytes in the liver, which mean the route of elimination would be bile canaliculi transportation into the biliary system, gastrointestinal tract and eventually feces [14, 143, 144]. The hepatobiliary excretion process is usually long and time consuming and can take up to one to two years for NPs, which caused possible inconsistencies and obstacle in performing quantitative or semiquantitative methods via evaluating organ biodistribution using fluorescence imaging or radiolabeled methods [14, 152]. NP size and composition would be two major factors in determining the interactions of liver cells with the NPs.

Poon et al. evaluated hepatobiliary excretion of AuNPs with surfaces modified with PEG and functionalized with Alexa Fluor 750 (AF750) using NHS-ester conjugation chemistry at approximately 50–200 nm after intravenous administration in BALB/c mice [14]. The authors used indocyanine green (ICG) as a positive control and confirmed that AuNPs existed along the hepatobiliary pathway in liver, bile (from the gallbladder), intestines, and feces, indicating that AuNPs can go through the canonical hepatobiliary pathway for excretion. In addition, they also performed intravenous injection of synthesized 100 nm dual labeled liposomes to determine the differences between biodegradable and nonbiodegradable NPs. At 30 min, an increased concentration of biodegradable NPs in major organs such as liver, kidney, spleen and gastrointestinal tract was observed, while at 4 h a decrease of plasma concentration was detected. Kupffer cells most likely contributed to the degradation of liposomes during the period of decrease in plasma concentrations. The study suggests that biodegradable NPs or NPs that are larger than 5.5 nm can be disassembled, broken down, or metabolized and may return to the blood circulation, followed by excretion through renal and hepatobiliary pathways. For nonbiodegradable NPs larger than 5.5 nm, most of them may retain long-term in Kupffer cells. However, it is important to note that if Kupffer cells are saturated or incapacitated by other mechanisms or if Kupffer cell can be avoided, then the NPs may undergo hepatobiliary excretion [14].

6. Tumor delivery

NPs can be specifically designed to preferentially target tumor tissues. A high delivery efficiency of NPs to the tumor site is important to protect normal cells from toxicity and also ensure the efficacy of the nanomedicine. There are two primary mechanisms of cancer targeting for NPs, including passive and active targeting.

As is shown in Figure 7, passive targeting is designed to utilize the difference between cancer cells and normal tissues. In tumors, neovascularization and large pores in the vascular wall can be triggered by high proliferation of cancer cells, thus changing permeation characteristics compared to normal vessels. Macromolecules including NPs are enabled by the defective angiogenesis to leak from blood vessels that supply the tumor and accumulate within the tumor tissue. Moreover, the expressions of some specific inflammatory factors are elevated, such as prostaglandins, bradykinin, nitric oxide, interleukins, interferon gamma, VEGF, and HIF-1α, which can increase the inter-endothelial cell gaps and subsequently vascular permeability [153]. These characteristics of solid tumors lead to the enhanced permeability and retention (EPR) effect, one of the driving forces of passive targeting [154]. In this regard, the NP size will greatly influence the permeability and penetrability of NPs into the tumor tissue besides other physicochemical characteristics of NPs [155]. Smaller NPs will be more permeable in tumor vasculature while still will not permeable into normal vessels [156, 157]. Large NPs, on the other hand, will not benefit from the EPR effect but rather be phagocytized and cleared by the immune system [158]. In an investigation of internalization in HeLa cells, polyethylene glycol hydrogels (100 nm – 5 μm) showed a clear size-dependent trend that larger particles had slower internalization kinetics than smaller ones [159]. After entering the tumor region, the NPs with relatively small sizes remain sufficiently penetrated in the dense matrix, while they are also easily pumped back into the bloodstream by the high interstitial fluid pression of the tumor. In contrast, large-size NPs can retain in the tumor well, whereas they are difficult to penetrate deeply [160]. For example, the comparison of size-controlled drug-silica nanoconjugates (i.e., different sizes of 20, 50, and 200 nm, but similar other physiochemical properties, including drug loadings, release kinetics, surface PEG densities, zeta potentials, and pharmacokinetics profiles) demonstrated that small NPs (20 & 50 nm) can more efficiently target tumor tissue passively than larger NPs (200 nm), and the NPs in 20 nm showed both more rapid accumulation and clearance (0.82 vs. 1.4 %ID/g at 48 h post injection) of radiation signal in the tumor than those in 50 nm, which implies fast tumor penetration but non-persistent accumulation of smaller sizes of NPs [161]. Accordingly, multiple aggregation strategies were used to correspond the effects of particle size on tumor delivery efficiency of which details will be discussed below. Other than EPR effect, tumor microenvironment would also influence passive targeting. Many cancer cells exhibit increased glycolysis, which is a series of metabolic processes by which one molecule of glucose is catabolized to two molecules of pyruvates [162]. During this process, glycolysis produces an acidic environment where pH-sensitive NPs would be triggered and release the drug within the vicinity of tumor cells [163].

Figure 7.

Passive and active targeting of nanoparticles to normal tissue cells and cancer cells. Enhanced permeability and retention (EPR) effect and tumor microenvironment are main determinants for passive targeting. Nanoparticles can also be specifically designed to target through direct interaction between ligands and receptors on the cell membrane (i.e., active targeting). Transferrin receptors, folate receptors, glycoprotein (such as lectin), and epidermal growth factor receptor (EGFR) are four major types of receptors on cancer cell membrane. (This figure was created with BioRender.com)

Passive targeting is usually enhanced by size changes. To achieve better penetration and accumulation in tumor in the meantime, the particle size can be changed by shrinkage strategy or aggregation strategy. Shrinkage strategy is usually achieved by biodegradation to release loading drug or active agent [155]. The stimulation may include pH, overexpressed enzyme, and reactive oxygen species [160]. The responsive chemical groups to pH include amino polymers [164] and 2,3-Dimethylmaleicanhydride [165], while those to enzymes include matrix metalloproteinase [166], hyaluronidase [167], etc. For example, a tumor microenvironment double responsive NP of LDC (the complex of laponite, doxorubicin, and chito-oligosaccharides) can be degraded by tumor-surrounding lysozymes efficiently and consequently shrink from the initial 100 nm into 30 nm to facilitate drug penetration deeply into tumors [168]. In contrast, the basic idea of aggregation strategy is that the initial small-sized NPs are first delivered to tumor region for deep penetration, and later form large agglomerates for better tumor accumulation after specific stimuli, such as enzyme, pH, light, temperature, and redox [160].

Although passive targeting benefits the NPs to be delivered to tumor, lack of specificity and differences in permeability of blood vessels can cause a low delivery efficacy compared to active targeting [169]. Active targeting, however, specifically targets NPs to cancer cells through direct interactions between ligands of the NPs and receptors on the cancer cell membrane. The ligands on the surface of NPs are selected to target the molecules (i.e., the receptors) that are overexpressed on the surface of cancer cells, which allows them to distinguish targeted cancer cells from healthy cells [169]. The types of targeting moieties include monoclonal antibodies, peptides, amino acids, vitamins, and carbohydrates [170]. These targeted moieties, also called ligands, bind to receptors that are widely overexpressed in cancer cells but have lower expression in normal cells. Common receptors include transferrin receptor, P-selectin, folate receptor, glycoproteins, and the epidermal growth factor receptor (EGFR). Transferrin normally delivers iron into cells (Figure 7) [171]. Given the fact that transferrin receptors are overexpressed in cancer cells and expressed at a relatively low level in healthy cells, transferrin-conjugated NPs are commonly used as targeting methods to deliver NPs to cancer cells [171–173]. Higher cellular uptake efficiency has been shown in transferrin-modified NPs compared to other non-modified drug NPs by up to 1.5 fold [174]. P-selectin is known to overexpress in metastatic cancer cells and vasculature surrounding tumor cells, but has a low expression in adjacent normal tissues. Accordingly, NPs containing a surface coating of fucoidan that can target P-selectin of tumor cells can facilitate passing of NPs through the vascular barrier of tumors [175]. In addition, the vitamin folic acid is internalized by a folate receptor that is expressed on a few normal cell types. The fact that the alpha isoform of folate receptor (FR-α) is overexpressed in approximately 40% of human cancers makes it popular to design a folate receptor-targeting strategy by folate-conjugated NPs [176, 177]. Thirdly, different types of glycoproteins including lectin are also common since lectins are non-immunological proteins that recognize and specifically bind to certain carbohydrates. Lastly, epidermal growth factor receptor (EGFR) is also overexpressed in different cancer cells. Conjugating EGFR using modified ligands in targeting EGFR-overexpressed cancer cells is a promising method for NP drug delivery [178–181].

Regardless of targeting strategy, several physicochemical characteristics were suggested to affect NP tumor delivery efficiency. In a recent meta-analysis, particle size, core materials, and surface charge can significantly influence the tumor delivery efficiency in tumor-bearing mice [10]. The impacts of particle size on passive targeting were discussed above. For core materials, a recent meta-analysis showed high tumor delivery efficiency for dendrimers and gold NPs in organic and inorganic NPs, respectively. Researchers are continuing to develop NPs with new materials with great potentials of tumor delivery. For example, multiple biomimetic NPs, are a new type of hybrid NPs emerged as novel drug delivery carriers in recent years (Figure 1). Compared to conventional NPs, biomimetic NPs can integrate native cell components (e.g., proteins, cell membrane) as the coating of the NPs, which can take advantages of biological features and functions of native cells to improve biocompatibility, target specificity, retention time, as well as reduce adverse immune responses [182]. In a design of macrophage-derived exosome-coated PLGA NPs, the in vivo pharmacokinetics showed remarkably extended blood retention time than PLGA NPs without coating (t_1/2: 1.3 vs. 7.9 h), as well as 2.22-fold higher fluorescence intensity in tumor [183]. Human serum albumin-based NPs coated with biomimetic 4T1 cancer cell membrane was shown to have improved colloidal stability and targeting ability to breast cancer cells [184]. In terms of surface charge, positive-charged NPs are mechanistically internalized by cells with high efficiency through electrostatic interactions since cell membrane are negatively charged [159]. Compared to negative-charge NPs, positive-charged NPs were suggested to be internalized into HeLa cells with larger amount (100% vs. 60%) [185]. Additionally, studies have shown that protein corona formation could modify passive and active targeting effects of NPs to tumor cells [186, 187]. Fleischer and Payne found that amine-terminated NPs denatured albumin that was adsorbed on the surface of NPs and changed the secondary structure of the adsorbed protein, and this can subsequently lead to a loss of specificity toward the albumin receptor on the cell surface in favor for scavenger receptors. The phenomenon indicates that the formation of biomolecular coronas can result in an alteration of the cell surface receptors targeted by NPs, which could decrease or even block the targeting efficiency [186]. Accordingly, many studies were conducted to increase the tumor delivery efficiency of NPs by changing NP’s physicochemical properties [188–197] (Table 2).

Besides the NP’s physicochemical properties, the tumor type is also a significant predictor of tumor delivery efficiency [10]. Therefore, some tumor-specific NPs were designed. Sobol et al. discovered that specific pancreatic tumor targeted antibody-labeled AuNPs increased tumor uptake into both subcutaneous and orthotopic pancreatic xenografts (24.0 ± 11.6 %ID/g vs. 4.0 ± 1.2 %ID/g), which accounts for a 4–7 times increase in deposition into tumors that expressed the target antigen [198]. 5B1 humanized antibody was utilized as the specifically targeted antibody and IgG label was used as the negative control. This study found that the accumulation of IgG (control) labeled AuNPs had significantly less accumulation compared with the 5B1 targeted labeled AuNPs, suggesting that with active targeting moiety on the surface of the 5B1-targeted AuNPs, the accumulation of AuNPs at the site of the tumor via EPR is further enhanced and retained by active tumor targeting [198].

7. Limitations and Challenges

While there is substantial progress as reviewed above, there are still some challenges in the field of nanomedicine. Safety and efficacy of nanomedicines are always the two primary concerns in this field. In the past several years, the majority of nanomedicine research reported unsatisfactory low tumor delivery efficiencies of most NPs, and some clinically tested NP drugs have been withdrawn due to safety concerns [10, 11, 199]. This phenomenon is, in part, due to a lack of thorough understanding of the in vivo fate of NPs, difficulties in the extrapolation of pharmacokinetics of NPs from animals to humans due to significant species differences in anatomy and disposition processes, and a lack of computational models to simulate or predict the in vivo pharmacokinetics and tissue distribution of different NPs between species [16, 200, 201]. Upon entering into the systemic circulation, a large portion of NPs accumulate in the liver and spleen. These nonspecific accumulation of NPs not only causes loss of pharmacological activity, but also increase potential safety hazards [145]. A phenomenon of greater concern is that the accumulation of empty nano-carriers in non-target healthy organs could inhibit the endosomal transport process thereby reducing NP delivery and transport efficiency to the target site [202]. Therefore, potential toxicity of NPs in non-specific RES organs such as liver and spleen should be given appropriate attention in NP risk assessment [15, 203]. Recently, intelligent nanosystems, such as pH-sensitive CaCO₃ NPs that are stable under physiological conditions and can release drug in the weakly acidic environment of the tumor, can result in a gradual drug accumulation when distributed to tumor site, thereby decreasing distribution to non-target organs [204]. Xu et al. designed a 4-carboxyphenylboronic acid-decorated redox nanomicelle which can specifically target to the tumor site during distribution in the body, thereby reducing the toxic side effects of non-selective anticancer drugs on non-target organs [205]. However, more detailed mechanistic studies need to be conducted to develop more intelligent NP delivery systems that reduce toxicity while increasing specific tumor delivery efficiency.

As mentioned in previous sections, metabolic processes may yield toxic metabolites of NPs. Some polymeric NPs (e.g., polystyrene) when degraded could release monomers that cause toxicologic effects to various organisms including carcinogenicity [206]. Premature metabolism of NPs could lead to early drug leakage during the biodistribution process, which can easily result in nondirected tissue toxicity, and this untargeted release lowers delivery efficiency to the target tissue [207]. Excessively slow metabolism will lead to long-term accumulation and potential toxicity [150]. Overall, understanding the role of metabolism of NPs in altering pharmacodynamic activity by adjusting their physicochemical properties to improve sustained release and enhanced effects remains to be investigated. The roles of various factors, including metabolic enzymes, the RES system, transporters and other biodistribution processes, as well as NP inherent properties such as size, zeta potential, pH, molecular weight on metabolism of NPs should be studied in combination. Quality evaluation standards based on existing regulatory guidance should be established with more details emphasizing the connections between ADME characteristics and physicochemical properties of NPs, in order to further decipher the toxicity and delivery efficacy of NPs [97, 208].

8. Conclusion and Future Perspectives

In conclusion, the review gives an introduction about factors that influence the ADME processes of different types of NPs. Pharmacokinetics of NPs is dependent upon the size, type, surface charge, and other physicochemical properties discussed above. Regarding absorption, scientists have worked for decades to improve absorption via various routes including oral, dermal, and pulmonary with which different approaches have been developed to bypass the relevant obstacles. Some NPs were modified and developed to be low pH resistant in order to overcome the harsh environment in gastrointestinal environments, while some other NPs were modified to change physicochemical properties for a better solubility to increase absorption [51, 71].

As for biodistribution of NPs, opsonization (e.g., protein corona formation), endocytosis (including phagocytosis pinocytosis and receptor-mediated endocytosis) and exocytosis are the major mechanistic steps involved in distributing both organic and inorganic NPs. In the body, metallic and inorganic NPs are mostly distributed to the liver and spleen, regardless of the NP type. Organic NPs share the same properties with accumulation primarily in liver and spleen followed by lower amounts in the kidney [6, 129]. Multiple factors including vascular permeability, plasma protein content, biomolecular corona formation, as well as RES system and complement system interaction with NPs can all influence their distribution process. Normally, NPs in the blood will distribute relatively rapidly to the liver, spleen and bone marrow due to the fact that the sinusoidal endothelial capillaries in liver and bone marrow are in direct contact with the vascular space [125, 209]. Compared to liver and bone marrow, the accumulation in the kidney would be relatively low due to lower maximal uptake and release rate constants of the kidney [125, 210]. NPs are less distributed to muscle and brain since larger sizes of NPs rarely penetrate muscle tissue due to normal or tight endothelium and the existence of blood-brain barrier [211, 212].

Metabolism of NP tend to differ between organic NPs and inorganic NPs. Inorganic NP tend to not get transformed except for the soluble ones such as AgNPs. AuNPs for example will stay in body for months and not be metabolized. In contrast, for some inorganic NPs like AgNPs, they may be dissoluted and release Ag+. Interestingly, AgNPs may then precipitate and even reform AgNPs depending upon the pH of the environment where ions diffuse to [213]. As for organic NPs, they tend to be biodegradable. Specifically, organic NPs can be biodegraded into hydrophilic small molecule metabolites for further excretion through urine or bile. For biodegradable NPs, specific chemical compositions and physicochemical properties will determine their susceptibility to metabolic processes. Neutral and hydrophilic NPs are more prone to be degraded into metabolites due to their ability to escape macrophage uptake. Some long chain NPs including polyacrylamide hydrogel are easier to be degraded into small molecules [150]. Metabolism plays an important role in the toxicity of NPs as seen in the example of the metabolism of single-walled carbon nanotube mentioned above [136].

The hepatobiliary and renal elimination pathways are major excretory pathways for both organic and inorganic NPs. Other excretory organs such as skin, lung or lymphatic system would also contribute to the elimination of NPs depending on how they were administered. Biodegradable and non-biodegradable NPs share different excretory and elimination process since biodegradable NPs could break down into their constituent materials and undergo elimination as small molecules. The rate of excretion of NPs directly impacts their circulation time, efficacy, and toxicity. Different NP size will lead to different excretion mechanisms and fates as described above [214]. Biocorona proteins would greatly modify surface NPs’ physiochemical properties and thereby influencing biodistribution, excretion, and toxicity of NPs [60, 101, 102, 104, 215].

Passive and active targeting strategies are the most important characteristics in the design of nanomedicines. EPR effect is an important driving force in passive targeting. Permeability differences between normal tissues and tumor cells is purported to be the reason for efficacy of the EPR effect. The EPR effect modulates delivery to solid tumors, tumor blood vessels and inflamed tissue since they exhibit a pore size around 200–780 nm for the diffusion of NPs [209, 216]. The enlarged pore size creates a “leaky vasculature system” in the solid tumor and should facilitate non-specific NP delivery, yet only low efficiency could be achieved [10, 209]. This could be due to the greater phagocytosis of larger sizes of NPs. In order to achieve increased specificity and delivery efficiency, active targeting by modifying NPs properties comes into play, that is incorporating ligands, monoclonal antibodies, peptides, amino acids, vitamins and carbohydrates to serve as modifying functional groups in NPs [217, 218].

Thanks to extensive recent research, many different NPs are being designed to overcome different ADME obstacles and mostly importantly, more mechanisms are being discovered that modulate NP disposition. As these mechanisms are being deciphered for both organic and inorganic NPs, study designs which optimize pharmacokinetic properties should gradually increase the percentage of injected dose reaching the intended target, making such targeting strategies more efficient.

At this juncture, in order to continue to improve our understanding of ADME and tumor delivery of NPs, we recommend the following considerations for future studies: (1) for every pharmacokinetic and biodistribution study, all information regarding the physicochemical properties of NPs and the study design (e.g., tumor model, targeting strategy, and animal model) should be reported; (2) when possible, dose- and time-dependence should be considered; (3) the original raw time-concentration data of NPs in plasma and tissues should be reported; (4) pharmacokinetic and biodistribution data of different NPs can be curated to build databases that may be further analyzed using more advanced pharmacostatistical methods, such as physiologically based pharmacokinetic (PBPK) models, machine learning, and artificial intelligence models. Recent studies have shown that machine learning and AI approaches can be integrated with pharmacokinetic modeling approaches to build AI-based PBPK models to support discovery and development of small molecular drugs [219]. We believe this AI-based PBPK modeling approaches are applicable to NPs and useful to support nanomedicine development [120, 220, 221].