1 Background

Nanotechnology offers great opportunities for the transformation of materials from bulk to particulate state, in which most of the material components (i.e. nanoparticles) have sizes at the nanometric scale (below 100 nm). Nanosized materials distinctly exhibit higher surface to volume ratios and impressive physicochemical and electromechanical properties useful in different fields of science and technology [68]. Efforts in nanoengineering have thrived to produce valuable nanoparticles (NPs) for various applications across different sectors, such as energy, agriculture and biomedicine. In biomedicine, the nanometric regime of NPs is an asset because most of the biological structures (e.g. proteins and DNA) operate at the same size scale. This underlines the growing applications of NPs in medicine, particularly in the field of bioimaging for diagnosis [69], drug delivery [14], immunotherapy [124] and vaccine development [124]. The use of NPs in medicine (i.e. nanomedicine) has made enormous strides in addressing health problems through effective drug development from the bench to the bedside. As recently discussed in the reviews by Anselmo and Mitragotri [8] and Zhong et al. [159], there are NPs-based products in clinical trials, and up to 30 nanomedicines in the pharmaceutical market. With the outbreak of the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), these statistics have drastically changed due to the urgent COVID-19 vaccine development. There are up to 20 NPs-based vaccines already approved in the clinic, and more than 100 candidates are in clinical trials [28].

Notably, the ability of nanoengineered materials to perform at the nanometric scale is a double-edged sword: while they are desired to provide new solutions to disease shortcomings, they can be detrimental to normal physiological functions or anatomical structures, if not optimized accordingly. For instance, Kolosnjaj-Tabi et al. [72] demonstrated that the nanosized components of inhaled particulate matters (i.e. carbon nanotubes) from air pollution are responsible for lung impairment. Another study reported the crucial contribution of nanometric size as a key element of cytotoxicity: while 20–80 nm silver (Ag) NPs were predicted to be cytotoxic due to high Ag ions release loads, 10 nm AgNPs were found to be way more cytotoxic because of their higher intracellular bioavailability [63]. There are other parameters (such as shape, surface charge and coating) that raise concern in nanotechnology as they increase the accessibility and reactivity of NPs towards the biological structures. With the growing use of biomaterials in nanotechnology, the intrinsic composition of biomedical NPs remains at the heart of development; efforts are often made so that the by-products from NPs degradation appear in biogenic forms [146]. This is a critical element of biosafety since the similarity with endogenous compounds may set the stage for good biocompatibility, but one big concern is the difficulty to distinguish metabolites from NPs degradation versus endogenous metabolites [120]. Nonetheless, formulation scientists are committed to continuously improving the biosafety profiles of NPs and boosting their clinical translation.

The fact that nanomedicine has delivered clinically relevant nanoplatform technologies into the pharmaceutical market is a considerable stride, likely arising from dedicated efforts to improve the quality attributes of individual products, but it is not a reflection of successful implementation of general design rules for the development of ideal nanoproducts. The gap between NPs product output and research input clearly shows the existence of high technical barriers hindering clinical translation, which is seen in the low number of NPs formulations entering the market versus increasing loads of research publications [159]. The lack of unified protocols for NPs evaluation is among the bottlenecks that hold nanomedicine from unfolding fully [154, 156]. The limitations in NPs testing have been recently reviewed [84].

This review briefly discusses the toxicological considerations pertaining to some of the nanoplatforms currently used in the clinic. Based on their inherent biodegradability, the platforms selected for this review include liposomes, poly(lactide-co-glycolide) (PLGA) and iron oxide NPs (IONPs), as representative illustrations of lipid-based, polymeric and metallic systems, respectively. The structural organization, biological fates, nanoparticulate properties and key data related to nanotoxicity of these platforms have been discussed.

2 Main text

2.1 Structural organization of nanoplatforms

The platform nanotechnologies explored in this review have distinct chemical compositions, being made up of phospholipids (liposomes), polymer (PLGA) and metal oxide (IONPs) (Fig. 1). The following paragraphs briefly highlight the structural organization of each platform to establish the fundamental differences that set them apart and may likely be one of the reasons for the differences in their nanotoxicological profiles.

Fig. 1
figure 1

Schematic representation of clinically approved synthetic nanoplatforms: A showing liposome formation and cargo encapsulation; B illustrating the formation of laden PLGA NPs decorated with stealth polymers (PEG); C illustrative formation of IONPs made of iron oxide (Fe3O4) core coated with carbohydrates (dextran) and decorated with targeting ligands

Full size image

2.1.1 Liposomes

Liposomes are artificial vesicular particles mainly composed of phospholipids, which are amphiphilic compounds naturally found in the cell membranes. Due to their compartmental structure, with a lipophilic membrane enclosing an aqueous core, liposomes offer the advantage of encapsulating molecular cargoes of hydrophilic, amphiphilic and hydrophobic natures (Fig. 1A). Liposomes are made of one or more phospholipid bilayers and produced in a diverse range of sizes, from 20 nm to several micrometres [98], as shown in Fig. 2A, B. The versatility of liposomes in the encapsulation processes has been demonstrated. They are reputed to be smart vehicles for molecular species of variable structure, from small and simple to large and complex molecules [45, 100, 113, 142]. In addition, the ease of surface modification enables the development of liposomes with multiple functionalities, such as stealth, long-circulating and ligand targeting properties [36]. These features set the basis for the applications of liposomes across various areas of nanomedicine, including site-specific drug delivery [17], non-invasive imaging techniques like magnetic resonance imaging (MRI) soft tissue morphological imaging and positron emission tomography (PET) [76] and immunotherapy [60].

Fig. 2
figure 2

Copyright 2004 American Chemical Society

Micrographs illustrating the nanoscale regime of nanoplatforms discussed herein. Cryogenic transmission electron microscopy (TEM) of liposomes loaded with doxorubicin (A) and cisplatin (B)—Reprinted from [136], with permission from Elsevier (licence number 5667521309110). Scanning electron microscopy (SEM) of PLGA NPs after freeze-drying (C) and spray-drying (D)—Reprinted from [123], with permission from Elsevier (licence number 5667530389172). TEM images of 6 nm (E) versus 12 nm (F) IONPs—Reprinted and adapted with permission from [121].

Full size image

2.1.2 PLGA systems

PLGA is a family of FDA-reputed co-polyesters of lactic acid (LA) with glycolic acid (GA) at various LA/GA ratios. Due to their excellent mechanical and processing characteristics, PLGA systems have been widely used for various biomedical applications, such as advanced drug delivery (i.e. sustained release, targeted delivery and protection of drug molecules), tissue engineering and regeneration medicine [90]. Depending on the cargo’s properties, PLGA-based architectures achieve drug product encapsulation through molecular or particulate (micronized solid) dispersion within the polymeric matrix [44], as illustrated in Fig. 1B. The types of cargoes vary from small drug molecules to macromolecules such as proteins and nucleic acids [134]. One of the most attractive features of PLGA systems is the tuneable release characteristics, which can be adjusted to afford long-acting dosage forms for improved patient compliance. Since the ester bonds of PLGA are subject to in vivo hydrolysis (i.e. esterase-mediated digestion), the release kinetics of drug molecules is mainly driven by polymer biodegradation and erosion [39]. Consequently, critical factors controlling drug release rates encompass the polymer molecular weight, LA/GA ratio and size and shape of the matrix. PLGA nanoparticles can exhibit various size patterns depending on the formulation treatments (Fig. 2C, D). Owing to the contribution of drug diffusion rate, additional parameters such as loading capacity, end terminus of polymer (e.g. acid terminated or ester terminated) and drug-polymer interactions have also been identified as key to release kinetics [66]. Drug release characteristics are part of the clinically relevant factors governing PLGA-based product efficacy and safety profiles [105], since PLGA copolymers are reputed biomaterials with commendable biocompatibility, as demonstrated by the existence of several clinically approved PLGA-based formulations [99].

2.1.3 Iron oxide nanoparticles

Iron oxide nanoparticles (IONPs) are one of the privileged ferromagnetic materials endowed with superparamagnetism, which makes them attractive for various sectors, including medicine and biology [4]. Among the types of iron oxides—namely magnetite (Fe3O4), maghemite (γ-Fe2O3) and wüstite (FeO), magnetite exhibits the optimal ferrimagnetism and highest saturation magnetization [152]. Owing to their excellent superparamagnetic properties and small size characteristics (Fig. 2E, F), IONPs with magnetite core, often referred to as superparamagnetic nanoparticles (SPIONs), have been extensively investigated for several biomedical applications, encompassing diagnostic magnetic resonance imaging (MRI), magnetic photothermal therapy and drug delivery [15, 37, 42, 115]. SPIONs are one of the few FDA-approved nanomedicines for clinical uses. Examples of marketed SPION products include Feridex® [21], Feraheme® [141] and NanoTherm® [119], which are used for liver MRI, iron deficiency and local hyperthermia tumour therapy, respectively. As shown in Fig. 1C, an essential constituent of SPIONs for successful biomedical applications is the surface coat from particle functionalization [95], which is mostly made of hydrophilic biomaterials such as carbohydrates [3, 85]. The ability to biofunctionalize SPIONs laid the foundation for controlling their biological fate, ensuring optimal in vivo performance [27]. For instance, this allows for imparting targeting capabilities or addressing protein corona formation to improve bioavailability [131].

2.2 NPs fates and nanotoxicity profiles

2.2.1 Overview of NPs fates

The increasing use of NPs in biomedicine has motivated continuous efforts to understand their behaviour and fates in biological systems. Many studies have highlighted the relevance of the characteristics of NPs (e.g. particle size, shape and charge) as intrinsic factors that dictate the fates of NPs. However, there are also numerous extrinsic parameters such as administration routes, pharmacokinetics and biodistribution that play pivotal roles [59, 133]. Many of these NPs-unrelated factors depend on various barriers, which stem from the normal physiology/anatomy, pathology and cell mechanisms. These barriers can either grant NPs access to or block them from certain biological areas or functions (Fig. 3). Such barriers can tip the balance between nanotoxicological and nanotherapeutic effects [91]. Therefore, a deeper understanding of the interactions between NPs and biological barriers is crucial for designing effective NPs with minimized side effects.

Fig. 3
figure 3

Reproduced from Meng et al. [91], with permission from Elsevier (licence number 5373331323566)

Schematic illustrations of naturally occurring and pathological barriers that can affect the fates of NPs in the body.

Full size image

Regardless of the targeting features, NPs distribute across various tissues and organs. Their distribution is influenced by factors such as the route of administration, the intrinsic properties of NPs and physiological and pathological conditions [52]. Consequently, different concentrations of NPs can be detected in various organs [79]. Irrespective of the administration routes, NPs eventually enter the bloodstream [81, 156], making their potential impact on the blood (haematotoxicity) a focus in nanotoxicity studies. Additionally, the formation of protein corona on NPs surfaces is commonly considered as a key determinant of their behaviour in the body. This corona affects biodistribution, accumulation and clearance of NPs [6] and essentially determines their biological identity in vivo. Notably, the composition of protein corona greatly influences the fate of NPs. For instance, when the corona is mainly made of opsonins, such as complement proteins and immunoglobulins, it aids in the NPs’ quick recognition and uptake by phagocytic cells [2, 102].

In addition to the plasma clearance resulting from protein corona formation, off-target tissue accumulation of NPs is a concerning issue that can reduce their efficacy and increase side effects. Factors driving off-target biodistribution include uptake by immune cells and tissue entrapment through endothelial fenestrations [43]. On the one hand, the phagocytic cells like monocytes and macrophages influence the ultimate organ accumulation patterns of NPs due to their localization in the body. While these immune cells are ubiquitous, they are primarily found in the liver, spleen, bone marrow and lymph nodes. These organs collectively constitutes the monocyte-macrophage cell system, also called the reticuloendothelial system (RES). Consequently, intensive phagocytosis of NPs by immune cells leads to their sequestration and accumulation in the RES organs [83].

On the other hand, the endothelium fenestrations in some tissues (e.g. liver, spleen, kidney and tumour tissues) have a direct impact on the organ distribution and accumulation of NPs [43]. While pathological fenestrations (as found in tumour tissues) favour passive targeting through the enhanced permeability and retention effect [53, 145], naturally occurring fenestrated endothelia found in the RES organs cause off-target accumulation of NPs, potentially leading to adverse side effects. The liver is the primary component of the RES and comprises both RES cells (i.e. liver macrophages, called Kupffer cells) and non-RES hepatic cells. This dual composition leads to two potential destinies for NPs within the liver: either clearance through bile when engulfed by non-RES cells or retention within the hepatic tissue following phagocytosis by Kupffer cells [83]. In the spleen, the phagocytic cells play a role in the organ’s filtration function, trapping undesired cells (such as ageing red blood cells) and foreign particles like NPs within the splenic tissue. Given the inevitable sequestration by the RES organs, histopathological and biochemical characterizations, as well as organ-function studies, become crucial for nanotoxicity evaluations. This helps ensure that the NPs designed for medical applications are both biocompatible and effective [26, 75, 154].

2.2.2 Nanotoxicological profiles of selected platforms

In vitro screening frequently serves as the first line of nanotoxicity screening, aiding in determining the minimum toxic dose to predict in vivo nanotoxicity [122]. Techniques for this in vitro nanotoxicity assessment typically encompass cell viability assays—subdivided into categories such as cell proliferation, necrosis and apoptosis tests. Additionally, investigations into cytotoxicity mechanisms, like the exploration of oxidative stress and DNA damage are also pivotal in these assessments [154]. In vitro nanotoxicity data primarily serve for guidance [65]. Yet, extrapolating these findings to an in vivo context can pose challenges due to the extremes of in vitro conditions, such as the use of ultra-high doses and extended exposure durations during dose-related cytotoxicity assessments [156]. Therefore, in vivo assessments are recognized as a pertinent and comprehensive method to evaluate nanotoxicity. However, they are also acknowledged as intricate, necessitating animal sacrifice and being time-intensive [154]. In vivo nanotoxicity evaluations mostly focus on histopathological alterations, in addition to analysing various biological and biochemical markers, including haematology, cellular metabolism and clearance mechanisms [26, 75]. The subsequent subsections provide a concise overview of significant in vitro and in vivo nanotoxicity data associated with liposomes, PLGA systems and IONPs.

2.2.2.1 Liposomes

The medical use of liposomes takes advantage of their nanometric functionalities, but also the possibility of developing different dosage forms for any routes of administration depending on the expected efficacy and safety, as demonstrated by the marketed products [20]. In fact, being made of naturally occurring biodegradable lipids (phospholipids), liposomes are part of materials commonly reputed to have high biocompatibility profiles. This can be illustrated by the phospholipid mixtures (i.e. Beractant, Survanta®) currently commercialized as artificial lung surfactants for prophylaxis treatment of the respiratory distress syndrome in neonates [110, 118]. However, when used as a delivery system, liposomes exhibit variable safety profiles depending not only on the formulation composition but also on the vesicles behaviour in vivo (i.e. biodistribution patterns).

There are many insightful studies illustrating the liposome’s potential to impact the overall safety profile of a therapeutic nano-formulation (Table 1) positively or negatively. The liposomal particle’s inherent factors (such as size and surface charge) have been tuned to control the in vivo fate of liposomes for improved product biocompatibility. For instance, Charrois and Allen investigated the impact of the diameter of liposomes on doxorubicin’s pharmacokinetics and biodistribution to minimize drug cutaneous toxicity while enhancing tumour homing [24]. The differences in the time courses of liposome accumulation in the tumour, skin and paws tissues were obvious, but no preferential accumulation into the tumour tissues was achieved. The influence of surface charge on liposome toxicity has also been demonstrated. As an example, through generation of reactive oxygen species, polycationic liposomes (Lipofectamine™) exhibited much higher lung toxicity than monocationic liposomes made of 1,2-dioleoyl-3-trimethylammonium-propane (DOTAP) [32]. In the same study, no toxicity was observed with neutral and negative liposomes made of phosphatidylcholine and a mixture of phosphatidylcholine and phosphatidylglycerol, respectively.

Table 1 Summary of representative studies evaluating the biocompatibility and toxicity of liposomes
Full size table

In addition to particle size and charge, the influence of liposome surface chemistry (such as PEGylation and ligand decoration) on the formulation’s biocompatibility has been investigated. Biswas et al. observed that stearyl triphenyl-phosphonium PEGylated liposomes showed much better HeLa cells’ viability (IC50 130 µg/mL) than the non-PEGylated liposomes counterpart (IC50 83 µg/mL) [19]. A good example of the influence of targeting on liposome biocompatibility is the study by Allen et al. [5], which also illustrates the advantage of a slower release profile over a faster release behaviour on doxorubicin toxicity. The authors observed that doxorubicin-liposomes decorated with anti-CD19 antibodies increased the mice life span to much higher levels than non-targeted doxorubicin-liposomes, though the latter were found to be still better than plain doxorubicin. The improvement in the drug’s safety profile through liposomal delivery arises from the fact that liposomes hold a great potential to alter drug biodistribution. Another illustrative example of liposomes improving drug toxicity profile is the work by Kuang et al. [74]: this study demonstrated that anionic long-circulating liposomes significantly reduced the accumulation of cisplatin in murine kidneys, which is an elegant strategy to prevent cisplatin’s acute renal toxicity. This is similar to the observation that liposomes suppress the cardiotoxicity of doxorubicin due to reduced accumulation in myocardia tissues [139], which led to successful implementation of marketed doxorubicin liposomal products (i.e. Doxil® and Myocet®).

Liposomes often alter drug circulation in a beneficial way, but liposome accumulation in the RES tissues (liver, spleen, etc.) can be problematic. Therefore, most in vivo studies consider histopathological evaluations of RES organs to assess tissue degeneration or necrosis due to the administered nanoproduct. As shown in Table 1, many authors observed no histopathological changes in the RES tissues of mice following treatment with therapeutic liposomes [71, 140, 157]. Nevertheless, careful safety profiling requires insights beyond histopathological evaluations, looking into molecular biomarker analysis to detect signs of nonnecrotic biological perturbations. In conjunction with this, the work by Knudsen et al. is an appealing illustration, where the biocompatibility of DOTAP cationic liposomes was evaluated [71]. According to the data, the histological, haematological and chemico-clinical evaluations indicated no significant changes after repeated doses, while genotoxicity assessment revealed elevated expression of cytokine genes and DNA strand breaks in the lung and spleen tissues (Fig. 4A). These findings are arguably part of the scarce data demonstrating the need for a thorough organ-function assessment when characterizing the safety profile of emerging NP formulations.

Fig. 4
figure 4

A Level of DNA damage expressed as strand breaks (SB), 24 h (plain bars) and 48 h (dashed bars) following IV injection of nanoparticles (0–100 mg/kg body weight) to Wistar rats (n = 8)—illustrating increasing DNA damage due to tissue exposure to liposomes (*p < 0.05 vs control). Reprinted from [71], an article published under the terms of the Creative Commons Attribution-NonCommercial-ShareAlike Licence (CC BY-NC-SA 3.0). B Change in Chinchilla rabbits body weight (n = 10) following daily IV injection of doxorubicin (Dox)-loaded PLGA NPs with and without poloxamer 188 coat (P188)—illustrating the toxicological benefit of Dox encapsulation in NPs. Reprinted from [111], with permission from Elsevier (licence number 5474400895276). IONPs concentration-dependent MTT cytotoxicity (C1) (*p < 0.05; **p < 0.01 vs control) and % induction of ROS generation (C2) (*p < 0.01; **p < 0.001 vs control) in human lung alveolar epithelial cells following incubation over 24 h (n = 3)—highlighting oxidative stress as mechanism for IONPs cytotoxicity. Reprinted from [34], with permission from Elsevier (licence number 5474441325417)

Full size image
2.2.2.2 PLGA systems

The inherent biosafety of PLGA arises from the fact that its biodegradation releases lactic and glycolic acids as by-products, which both enter the Kreb’s cycle for safe metabolization into carbon dioxide and water [128]. However, careful biocompatibility profiling entails thorough toxicological consideration including not only polymer intrinsic properties but also relevant biological parameters such as tissue-polymer interactions or tissue exposure to the laden polymeric matrix, which can lead to intolerable inflammatory or immunological responses [117]. In this context, types of PLGA-based constructs (i.e. implants, microparticles or NPs) are of paramount importance due to inconsistent in vivo responses observed. For instance, implants can cause prolonged local reactions throughout polymer degradation, while NPs may lead to intracellular disturbances due to high cell penetration and broad biodistribution [86]. The biocompatibility of all the PLGA systems has been recently reviewed [35]. As a biocompatible vehicle, PLGA NPs have demonstrated some potential to control drug product biosafety. For instance, the haemolytic effect of the antibiotic amphotericin B was significantly reduced when used in the form of amphotericin B-loaded PLGA NPs, which was solely attributed to the slow release behaviour of NPs [93]. In another study, PLGA NPs exhibited dose-dependent cytotoxicity, but the overall safety profile was much better than that of dendrimers, since the former showed acceptable cytocompatibility up to 180 μg/mL (despite their larger sizes, 140 nm), while the latter were highly cytotoxic at 2.8 10–4 M (though being of smaller sizes, 10 nm) [31].

The impact of surface chemistry on cytocompatibility of PLGA NPs was elucidated by Gossmann et al. [46]. These authors observed that non-decorated PLGA NPs loaded with didodecyldimethylammonium bromide exhibited much higher cytotoxicity on Caco-2 cell line than their PEGylated formulation counterparts (EC50 54.8 vs 996.5 µg/mL, respectively); this was further explained by enhanced NPs dispersion and absence of aggregation due to PEG chains [47]. A similar observation was reported by Cai et al., who noted that PEG-PLGA NPs containing poly-L-orithine/fucoidan were less cytotoxic than non-PEGylated NPs of the same composition [22]. Another variable parameter investigated in PLGA NPs biosafety profiling encompasses the routes of administration. In a comparative toxicological study, amphotericin B laden PLGA NPs administered orally and intraperitoneally demonstrated no nephrotoxicity nor hepatotoxicity in Wistar rats irrespective of the routes, while marketed deoxycholate amphotericin B formulation caused hepatic cellular alteration following administration through the same routes [93]. As illustrated in Table 2, many studies reported various biocompatibility profiles of PLGA NPs depending on the chemical composition and functionalities [51, 92, 125, 132]. Among the important observations from PLGA nanotoxicity studies is the meticulous analyses of biosafety that explored broader toxicological parameters. An illustrative example is the study by He et al. [55, 56]. The authors established the biosafety of monomethoxy-PEG-PLGA-poly(L-lysine) NPs by assessing (1) protein synthesis, cell membrane integrity and chromatin agglutination in Huh7, L02 and RAW 264.7 cells; (2) the release of interleukin-1β, tumour necrosis factor-α and transforming growth factor-β1 from THP-1 cell-derived macrophages; and (3) the potential impact on embryonic development using zebrafish embryos. Moreover, nanotoxicity assessment is required to involve chronic toxicity studies in addition to acute toxicity studies to better inform on the biosafety profiles of nanoproducts. For example, acute toxicity studies recently showed no difference between plain doxorubicin versus doxorubicin-loaded PLGA NPs [111]; however, chronic toxicity data demonstrated that the encapsulation of doxorubicin in PLGA NPs significantly improved drug safety profile (Fig. 4B). Rigorous exploration of nanotoxicity being one the utmost requirements for translational development, such extensive safety profiling appears to be highly inspirational for nanotechnologists.

Table 2 Summary of representative studies evaluating the biocompatibility and toxicity of some PLGA-based NPs
Full size table
2.2.2.3 Iron oxide nanoparticles

SPIONs are associated with nanometric sizes and high surface to volume ratios that promote the biological reactivity of particles (as the case for all NPs in general). However, due to enhanced reactivity and propensity to diffuse through biological barriers, the nanoscale regime of SPIONs is the motor for their interference with the structure and functions of tissues or organs [129]. Therefore, commendable efforts have been dedicated to SPIONs’ nanotoxicity assessment to improve the safety profile of each individual product under development. Many reviews have recently discussed the toxicity and biocompatibility of SPIONs [10, 147, 152, 158]; hence, the few data presented in Table 3 solely serve as illustrative references for emerging nanotechnologies.

Table 3 Summary of representative studies evaluating the biocompatibility and toxicity of SPIONs
Full size table

Apart from the general NP toxicological parameters (related to particle characteristics and pharmacological factors), the intrinsic toxicity of SPIONs are attributed to both the iron core and surface coatings [10]. The toxicity mechanisms inherent in the iron core include the generation of reactive oxygen species (ROS) causing oxidative stress [34], and the iron overload that alters iron homeostasis leading ultimately to cell death (Fig. 4C1, 2). Following endocytosis, iron particles are metabolized in iron ions that further diffuse into the nucleus and mitochondria, where they react with oxygen and hydrogen peroxide through oxidative reduction to generate ROS [9, 155]. Regarding the surface chemistry side, studies have demonstrated the chief influence of coatings on SPIONs’ biocompatibility profiles, due to their impact on particle size, surface charge and oxidative site accessibility. For instance, PEGylation has shown great potential to reduce the toxicity of SPIONs both in vitro [106] and in vivo [114]. The positive impact of surface chemistry on SPION toxicity was also observed with non-polymeric coatings, such as curcumin and silica layers that were used by Malvindi et al. [88] and Bhandari et al. [16], respectively. Nevertheless, most coating–toxicity relationships have been established in vitro, hence there is a pending need for further confirmation under in vivo conditions. In fact, an instructive study showed that PEGylated SPIONs that were reported to be nontoxic in vitro were found to cause liver and kidney injury [127], likely because of in vivo xenobiotic metabolization (based on the intracellular toxicity mechanisms explained early). Another important parameter is the SPION surface charge, which is consistent with some of the general nanotoxicity rules, postulating that positively charged particles are toxic while neutral and negatively charged particles are expected to be safe. Cationic coatings like polyethylenimine (PEI) have shown the potential to yield toxic SPIONs, while neutral or poorly charged PEGylated particles were found to be safer under the same conditions [38, 58].

While many studies have demonstrated the possibility of tuning the characteristics of SPIONs for enhancing biocompatibility by means of cell viability or proliferation assays, some authors have investigated subcellular and molecular perturbations for a diligent biosafety profiling. As an example, Pongrac et al. evaluated the potential toxicity of SPIONs with different coatings (D-mannose or poly-L-lysine) using murine neural stem cells as a regenerative medicine model [112]. Data showed that both uncoated and coated SPIONs (regardless of the coating) exhibited adverse effects at subcellular levels (e.g. loss of mitochondrial homeostasis, DNA damage, etc.), while cell viability remained unchanged after 24 h of incubation with up to 200 μg/mL. Such studies evidence the need for careful characterization of nanotoxicity to engineer NPs with a good safety profile. Despite their clinical relevance, SPIONs under development exhibit inherent toxicity that requires meticulous toxicological testing. The fact that such intrinsically toxic materials have made it to the clinic lends hope for futuristic developments of safer nanoproducts.

2.3 Discussion about NPs characteristics affecting product toxicity

2.3.1 Impact of liposomes characteristics on product toxicity

The characteristics of liposomes, such as size, surface charge and composition, are crucial in determining their toxicity and safety. Smaller liposomes can target tissues more effectively due to enhanced permeability and retention, but may also present higher toxicity levels. For example, smaller liposomes have shown faster drug release and greater toxicity compared to larger counterparts like Doxil liposomes [13]. Strategies such as charge neutralization and PEGylation have been employed to mitigate toxicity and enhance safety. The lipid composition must be carefully balanced to maintain efficacy while minimizing adverse effects. In vitro drug leakage tests are vital for assessing the stability of the lipid bilayer and the encapsulated drug, ensuring controlled drug release under physiological conditions. This is essential for consistent drug delivery to the diseased tissues, which could influence toxicity. As far as drug retention and release are concerned, the internal environment of the liposome, including its volume and ionic concentrations, is critical for drug loading and stability, impacting the toxicity profile [13].

The pharmacokinetics and biodistribution of liposomes are influenced by their size, affecting therapeutic efficacy and toxicity. Smaller liposomes are designed to encapsulate drugs effectively and have extended circulation times, allowing for efficient targeting of specific sites, such as tumour tissues. This targeted delivery is crucial for minimizing exposure to healthy tissues and reducing systemic toxicity. Liposomal drug delivery systems, particularly those modified to reduce cardiotoxicity, optimize the delivery of anticancer drugs, reducing their toxicity [137]. PEGylated liposomes, or ‘stealth’ liposomes, are designed to evade the immune system and prolong circulation time. PEGylation helps these liposomes avoid clearance by the mononuclear phagocyte system, enhancing drug delivery to tumour sites and reducing toxicity. The pharmacokinetic profile of doxorubicin-loaded PEGylated liposomes differs significantly from doxorubicin alone, due to their size and stealth properties, which allow for targeted and sustained drug release, potentially reducing toxicity to non-target tissues [137].

The surface charge of liposomes affects their tissue distribution, cellular uptake and clearance. Positively charged lipids are linked with higher toxicity, which can be reduced by incorporating negatively charged phospholipids. The ratio of cationic to anionic lipids is pivotal for achieving a balance between safety and efficacy. For instance, it was found that a molar ratio of cationic lipid (DOTAP) to anionic lipid (POPS) of 7:3 achieved the best safety-efficacy balance [144]. The review by Inglut et al. deeply discussed the surface characteristics affecting liposomes toxicity to healthy tissues and unfavourable immune responses [62]. The formulation composition is also critical because it dictates the structural organization of liposomes (such as the morphology and lamellarity), ultimately influencing product safety. The morphology and lamellarity of liposomes affect drug loading and release, which can influence toxicity, as it was previously reported that changes in the particle shape of Doxil liposomes led to complement activation [136]. In summary, the use of liposomes in drug delivery systems requires a thorough understanding of their characteristics to ensure safety and efficacy. Analytical methods and careful experimentation are necessary to elucidate the mechanisms of liposome toxicity, facilitating their safe application in medical treatments [1].

2.3.2 Impact of PLGA nanoparticles characteristics on product toxicity

PLGA is widely researched for drug delivery applications due to its biodegradable nature. Its degradation products are metabolized via the Krebs cycle and are safely cleared by the body. Numerous studies have confirmed PLGA's biocompatibility across various biological tissues, underscoring its suitability for medical use. However, the safety profile of PLGA formulations can be unpredictable due to discrepancies in formulation technologies. These discrepancies arise from the diversity in particulate formulation characteristics and composition, highlighting the necessity for comprehensive nanotoxicity studies to balance risk and benefit trade-offs. The safety and toxicity of PLGA nanoparticles are influenced by several factors, including size, surface charge, biodegradability, concentration and stability. Smaller nanoparticles, with their larger surface area-to-volume ratio, exhibit higher reactivity, which may lead to increased toxicity. This is supported by studies such as [80], which demonstrate that smaller PLGA nanoparticles can significantly trigger the release of TNF-α. However, these nanoparticles did not show notable cytotoxicity at concentrations up to 300 μg/mL [153]. Furthermore, the propensity of nanoparticles to adsorb proteins is a critical consideration, because it influences their in vitro cytotoxicity. Smaller particles, for instance, tend to adsorb more biomolecules, potentially contributing to their cytotoxic effects [153].

The surface charge of PLGA nanoparticles also plays a crucial role in determining their safety profile. For example, negatively charged PLGA nanoparticles were found to induce a higher inflammatory response in cells [94]. However, a study by Dailey et al. [30] found that PLGA nanoparticles, despite their small size and large surface area, elicited the lowest inflammatory responses compared to non-biodegradable polystyrene nanoparticles. This finding supports the Generally Recognized As Safe (GRAS) status of PLGA's polymeric body. However, surface modifications, such as coatings, can introduce toxicity to PLGA nanoparticles. For instance, Grabowski et al. [48] observed that cytotoxicity and inflammatory responses in A549 human lung epithelial cells varied with surface modification. Solutions containing chitosan, for example, induced similar toxicity levels as corresponding PLGA/chitosan nanoparticle suspensions. Such modifications can alter the surface charge of PLGA nanoparticles, affect their stability in various environments and lead to size variations, ultimately impacting their safety profile [94]. Overall, the safety and efficacy of PLGA nanoparticles are multifaceted, contingent on their size, surface charge, biodegradability and concentration. These factors collectively influence the nanoparticles' interaction with biological systems, their biodistribution and their potential to cause inflammatory responses or other toxic effects.

2.3.3 Impact of SPIONs characteristics on product toxicity

Surface modification techniques, notably PEGylation, are pivotal in mitigating the toxicity of nanoparticles. This strategic alteration aids in circumventing the reticuloendothelial system, thereby significantly reducing toxicity. Such modifications render nanoparticles, particularly SPIONs, more apt for diagnostic and therapeutic applications. Notably, PEGylated SPIONs demonstrate enhanced stability and a marked reduction in undesirable interactions with plasma proteins, a critical factor for ensuring safety in medical applications [78, 114]. In a study by Park et al., different types and concentrations of coatings on SPIONs were found to have varying morphological effects on cells: citric acid-coated SPIONs at high concentrations led to the formation of cytoplasmic vesicles and eventual cell death, while PEGylated SPIONs showed less cytotoxicity [106].

The particle size of SPIONs is a significant determinant in their interaction with biological systems. Smaller particles, due to their distinct biodistribution and toxicokinetics, behave differently compared to their larger counterparts. Particles under 10 nm are typically excreted through renal clearance, while those exceeding 200 nm are predominantly sequestered by the spleen. The ideal size range for medical administration is established between 10 and 100 nm. This size-dependent behaviour, particularly the enhanced cellular penetration of smaller nanoparticles, potentially leads to varied biological effects, as detailed in [87]. Furthermore, the internalization and degradation rates of SPIONs within cells are crucial to their safety. For instance, SiO2-coated SPIONs have been observed in MRI scans within stem cells for durations of 8–12 weeks, likely attributable to the stability of SiO2 in cellular environments [78].

SPIONs are known to induce oxidative stress, DNA damage and caspase activation, all of which are vital considerations in evaluating their safety. The generation of reactive oxygen species, a key player in SPION-induced genotoxicity, can lead to extensive cellular damage and apoptosis, underscoring the critical role of oxidative stress in nanoparticle-induced toxicity [114].

Biocompatibility emerges as a paramount factor in determining the safety profile of nanoparticles. SPIONs that exhibit biocompatibility show reduced toxicity in vivo, although comprehensive toxicity profiling is essential for their clinical application. Additionally, the impact of nanoparticles on immune cells, such as T-lymphocytes, warrants careful consideration, as discussed in Prabhu et al.'s study [114]. This summarized overview underscores the intricate relationship between the characteristics of SPIONs and their biological interactions. It highlights the necessity for thorough research and meticulous design to ensure the safety and effectiveness of SPIONs in biomedical applications.

3 Conclusions

Successful implementation of nanomedicines in the clinic has motivated continuous investigations to develop high value nanoparticles (NPs) for medical applications. The ability of NPs to operate on the same size scale as biological structures (i.e. nanoscale) is one of the most intriguing features underlying the effectiveness of NPs in medicine. However, sufficient evidence demonstrated that the nanoparticulate nature also sets the ground for potential toxicity, which is likely one of the key challenges affecting their translational development from bench to bedside. In this review, toxicological data from selected clinically approved nanoplatforms (liposomes, PLGA and IONPs) were briefly outlined. This may serve as a practical illustration of the lack of design rules for safe and quality design in nanoengineering since, even for such clinically established nano-systems, there are still limited insights into the direct correlation between the properties of particles and nanotoxicological profiles. The effectiveness and safety of nano-formulations are intricately influenced by factors such as particle size, surface charge, chemical composition, surface coating, solubility and aggregation state. These characteristics play a pivotal role in shaping the interactions of NPs with biological systems, thereby determining the overall efficacy and safety of the nano-formulations. Although most studies reported no obvious histopathological, apoptotic or necrotic changes caused by NPs, the analysis of the summarized data indicates the lack of organ-function studies as well as the difficulty to compare nanotoxicity data from different protocols, given the discrepancies in doses, models, etc., that make comparison beyond standards. Therefore, unified nanotoxicity assays are highly desired to encourage systematic elucidation of particulate/molecular mechanisms of nanotoxicity and allow the clinical translation of nanomedicines to fully unfold.