In Vitro, In Vivo and Ex Vivo Models for Toxicity Evaluation of Nanoparticles: Advantages and Disadvantages

Abstract

This chapter focus on existing model systems used to evaluate the toxicity of nanoparticles. We will be discussing monolayer and 3D cell based toxicity models, In vivo models like rodents and zebrafish systems. A focus will also be given on ex vivo models like chick embryos. Each toxicity model system will be discussed with its advantages and limitations. The chapter will provide critical information to students and researchers studying nanotechnology about the potential systems to check the toxicity of the nanoparticles developed in the laboratory. This can be used as a quick guide to use a model system to check toxicity based on the different type of particle with informed decisions based on its advantages and disadvantages.

1. Introduction

Nanoparticles (NPs) range from size 1–100 nm [1]. They are made from various materials like polymers, liposomes, dendrimers, and metals like Zinc, Titanium, Gold, and Aluminum [2]. NPs have been found to induce toxicity through the production of reactive oxidative species (ROS). Their small size, greater surface area to volume ratio, and ability to easily penetrate tissue cells, leading to higher chemical reactivity, cause increased ROS production when introduced into the body [3, 4, 5]. Also, due to their very small size, they can be ingested through inhalation and are able to pass through biological barriers into sensitive parts of the body like the lungs, brain, heart, liver, and spleens [6, 7]. A factor that Increases their toxicity is the solubility of the nanomaterial. NPs like zinc oxide and titanium oxide have been found to elicit more toxicity than ceria and titania, which are less soluble [8]. Also, due to the diversity in the use of these NPs in various industries like food, cosmetics, agriculture, biomedical, optics, and technology, it is easy for them to be absorbed and ingested into the human system, affecting the human gut microbiota. They have been observed to accumulate in the stomach, ileum, colon, and duodenum, which poses serious risks and concerns [9, 10]. Hence, the need to study their toxicity in human systems has become necessary [11].

The two major ways of assessing nanoparticle toxicity are in vitro and in vivo methods; ex vivo, although not commonly used, is another form. As shown in Figure 1, here are the common assays used to measure nanoparticle toxicity. With the use of these methods, there has been observed variability in results and even obstruction from the NPs themselves in the assay. In this chapter, we will be considering the various limitations and advantages of the toxicity assays commonly used in nanoparticle toxicity studies to ensure scientists have adequate knowledge and make the best decision on the assays to use for their study.

Figure 1.

Conventional methods of evaluating the cytotoxicity of the nanoparticles [12].

2. In vitro models

2.1 Proliferation assays

Proliferation assays are assays used to check for cellular metabolism in active metabolic cells. They help to ascertain the viability of cells when treated with NPs to check for the toxic effects of the particles on the cells. One of the commonly used assays is the 3-(4,5-Dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) assay. It is a colorimetric assay where 3-[4,5-dimethylthiazole-2-yl]-2,5 diphenyltetrazolium bromide (MTT) is enzymatically reduced by mitochondrial succinate dehydrogenase to formazan crystals, an insoluble product, in the mitochondria of live cells. The breakdown is spectrophotometrically measured to estimate cell viability. In this assay, there is a linear relationship between the color formed when the breakdown to formazan occurs or the absorbance when measured and the viability of the cells. It is a very sensitive and quantitative assay [13, 14]. The MTT assay proves to be advantageous because it produces results quickly under a maximum of 3 to 4 hours. It is also reproducible and does not require manipulation of the target cell [15].

The Alamar Blue Assay is also a colorimetric indicator assay characterized by the reduction of a blue, non-fluorescent dye to a pink-colored substance known as Resazurin [16]. It is both a qualitative and quantitative assay where colorimetric measurement can be taken or physical observation of color change in assessing cell viability or treatment toxicity. It is taken at a wavelength of 570 and 600 nm or 540 and 630 nm using the spectrophotometer. It is excited and emitted at wavelengths of 530–560 and 590 nm respectively [17]. Alamar Blue assay is highly sensitive, requires low cost, it’s easy and safe to use, non-radioactive, and can be used for a large number of sample. It can also be used for both quantitative and qualitative analysis. Another method used to assess proliferation is the incorporation of [³H] thymidine into the DNA of proliferating cells during the S phase of the cell cycle, with the use of autoradiography [18]. As the cells proliferate, new DNA strands are formed. The tritiated thymidine, a radioactive nucleoside, then enters into the new chromosomal strands as the cells divide [19]. [³H] thymidine incorporation method is also widely used in immunological studies because of its high throughput and direct measurement of cell proliferation. It is also sensitive [19]. Another commonly used assay, clonogenic assay is used to assess cell survival and reproductive ability after treatment. The ability of a single cell to reproduce a colony is checked [20].

As shown in Figure 2, MTT assay has its limitations that affect the interpretation of data. Due to the formation of formazan, an insoluble dye, in MTT assay, other assays like XTT (2,3-bis [2-methoxy-4-nitro-5-sulfophenyl]-2H-tetrazolium-5-carboxanilide) are used. With XTT, in the presence of mitochondrial dehydrogenase enzymes, a soluble water product is produced. However, both assays also faced the limitation of reaction under other conditions like acidic media, additives found in culture media, polyphenols, the presence of NPs and O⁻₂. These gives a false positive result of cell viability [22, 23, 24]. In a study carried out on Chinese hamster ovary cells (CHO-K1) after exposure to nano-TiO₂, which increases the formation of O⁻₂, MTT and XTT assay were found to produce an inaccurate analysis of cell viability, because of the ability of the superoxide produced by the NPs to reduce the two salts [25]. Other factors such as cell number, the concentration of the MTT reagent, the treatment applied to the cells, and the toxic effect of the MTT itself has to be critically considered when analyzing the results from the assay. When prostate cancer cell lines (PC-3) were treated with polyethylene glycol-coated gold NPs (Au-PEG-NPs) at a concentration of 5 nm, there was an observed significant optical interference with the MTT assay. However, this effect could be minimized by washing the cells and removing the supernatant before MTT incubation and using appropriate blanks [21].

Figure 2.

Factors affecting the final optical density (OD) measurements in the MTT assay. These include the concentration of MTT reagent and the proportion that actually enters the cell, cellular metabolic activity (which is highly dependent on a multitude of variables including treatments to the cells, biological effect of culture media, cell density, and impedance of cell metabolism due to toxic effects of MTT), cell number, timing of formazan crystals extrusion (which could impede further MTT uptake), chemical interference such as abiotic reduction of MTT by culture media, the tested treatment, or released cellular content, optical interference by all the background components, time of incubating cells with MTT reagent and/or tested treatment, and ultimately the optical measurement. Chemical structure of MTT and formazan are illustrated inside the cell: MTT consists of a tetrazole ring core containing four nitrogen atoms (1) surrounded by three aromatic rings including two phenyl moieties (2) and one thiazolyl ring (3). Reduction of MTT results in disruption of the core tetrazole ring and the formation of formazan. Red arrows and the “-” sign indicate disruption of MTT reduction on the normal metabolic activity of the cells and the impeding effect of the formazan crystals (when presenting on the cell surface) on further uptake of MTT reagent by cells [21].

A study on the in vitro assessment of the toxicity of carbon based NPs HiPco single-walled carbon nanotubes (SWCNT), arc discharge SWCNT and Printex 90 carbon black NPs was done to initiate a more reliable method with the use of clonogenic assay [26]. This study was performed to avoid the interaction of carbon based NPs with the indicator dyes used in colorimetric assays, leading to inaccurate interpretations on toxicity studies [27, 28]. Graphene oxide and TiO2 NPs have also been observed to interact with the dyes [29, 30]. This assay rules out the possibility of the nanoparticle reacting with the assay itself. It is recommended as a very useful tool for testing cytotoxicity since colony number and size are taken into account. This makes it an effective differentiating tool between cell viability and cell proliferation [26]. It is also a sensitive method since colony size depicts the division rate and proliferation after the cells are treated [31].

For Alamar Blue assay, because of its photosensitivity, it is affected by light exposure and has to be done in the dark. Cell density is also a factor that can affect the assay reading, hence cells must be cultured in high population to prevent slow growth and allow adequate dye reduction. The assay is also limited to a pH range of 7.0 and 7.4, and optimum temperature of 37°C. Longer incubation times could also lead to the reduced dye being bleached to a pink color. This quenches the fluorescence and could cause misinterpretation of the fluorescent signals. There is also the need for a positive and negative control to rule out non-specific interaction with the chemistry [17]. Specifically, with the use of nanomaterials, it is necessary to check toxicity using a combination of other assays because of the diversity and physiochemical properties of the nanomaterials, which could affect the cell and also interfere with the chemistry of the assay [32, 33].

The [³H] thymidine proliferation method is toxic and also requires radioactive facilities, which can be very expensive. Also, there is the need for proper waste management of the radioactive materials used with this method [34, 35]. It has also been found to interfere with the target cells by inducing cell cycle arrest, apoptosis and fragmenting the DNA [34, 36, 37].

One of the major cons of Clonogenic assay is that it is time-consuming as it takes about 10–14 days to perform. This is quite long compared to the other assays described above which take only a few hours to a day [38].

2.2 Apoptosis assays

Apoptosis Assay is another assay carried out to evaluate the cytotoxicity of nanomaterials. It measures the extent of DNA damage and cell death when cells are treated. The release of free radicals due to oxidative stress is a pointer to DNA damage and nanomaterials have been found to elicit such reactions [39, 40]. One of the methods used to study apoptosis is the Annexin-V assay. This assay works on the principle of annexin-V binding to phosphatidylserine (PS) which is externalized on the plasma membrane due to activation of caspase-dependent pathway. In combination with Propidium Iodide (PI) which stains the nucleus indicating the last stage of cell death, both dead and apoptotic cells can be identified [41, 42]. Annexin-V and PI assays have been used to detect apoptosis in HeLa and human HepG2 hepatoma cells treated with gold and silica NPs respectively [43, 44]. Annexin-V does not penetrate the cell but only binds to phosphatidylserine on the extracellular membrane of the cell; hence it does not cause cell damage or affect the intracellular components of the cell. It is highly sensitive as it binds to only PS amidst other molecules on the cell surface. It produces bright fluorescent signals and it’s easy to perform [41]. There are different options available for its labeling and also it can be performed both in vitro and in vivo [45].

Comet assay is used to determine DNA damage, single- and double-stranded DNA breaks and the mutagenicity of treatment both in vivo and in vitro [46, 47, 48, 49, 50]. It is one of the commonly used assay to assess the genotoxic effects of engineered nanomaterials like TiO₂, SiO₂, Zinc oxide NPs [51, 52, 53, 54, 55]. Damaged bases are detected when nucleoids are incubated with endonuclease III and formamidopyrimidine DNA glycosylase (FPG), specific to oxidized pyrimidine and purines, respectively [56]. Comet assay is a sensitive technique that is quick, cheap and easy to do [57]. It has been used both in vitro and in vivo and hence can be used to check the toxic effects of treatment in a particular organ or tissue and on any animal model [49, 58]. It can also be done on the first or specific site of contact.

TUNEL (terminal deoxynucleotidyl transferase biotin-dUTP nick end labeling) assay is a commonly, wisely used tool for the detection of DNA damage caused by both apoptosis, DNA fragmentation and necrosis, caused by exposure to toxic materials [59, 60, 61]. When DNA fragmentation occurs, the 3^′ hydroxyl termina becomes free. Then the enzyme terminal deoxynucleotidyl transferase incorporates labeled dUTP into the free end, which causes the staining observed in the assay [62]. It can be detected by light microscopy, fluorescence microscopy or flow cytometry [63]. TUNEL assay is fast as it can be completed within 3 hours. It has high sensitivity and versatile with the use of various techniques, as single cells can be detected using fluorescence microscopy or few cells using flow cytometry [63, 64].

Flow Cytometry is a very common and versatile technique used in the toxicity analysis of nanomaterials when applied to cells. It works on the principle of detecting fluorescently labeled cells when light beams are passed through the cell suspension. The technique analyses both the size and granularity of the cell populations, which gives a characteristic of the type of cells being studied [65]. With the use of flow cytometry, large populations can be analyzed with valid statistical results and this is done within a short period of time. This method can be used to evaluate the uptake of NPs by the cells, cell death, and expression of certain proteins [66].

However, these assays have their own set of limitations. For example, Annexin-V assay cannot be effectively used to obtain high throughput due to strong background signals when there are unbound labeled annexin-V. It can be difficult to optically image tissues when stained with annexin-v because of its short lifetime and slow diffusion in the body [67]. It can produce false positive results when it binds to negatively charged aldehyde adducts [68]. Also, Engineered NPs have been found to interfere and hinder the detection of DNA damaged from oxidation with the comet assay when ions are released when the particles are dissolved [69]. Cell toxicity, even in the absence of DNA damage, can lead to false-positive results in Comet Assay. There are also variability of protocols from the assay which can make it difficult to compare results with other laboratories, as there is no standardization [57]. For TUNEL assay, there is the limitation of non-specific staining because it labels all the free 3′ hydroxyl termini end irrespective of the cause of it. Hence, it can label non-apoptotic cells, which gives a false positive result. This means it can stain cells undergoing DNA repair, cells secretes factors causing proliferation in neighboring cells, or cells damaged through other means [70]. It is also costly. For flow cytometry, freshly prepared samples are required. If samples are kept for a long period of time before analysis, it could affect the cell properties and the analysis. Also, the use of the flow machine requires competent and skillful hands to accurately analyze and interpret the data [71]. NPs smaller than 100 nm experience low light scattering which could reduce the sensitivity of the detection [72].

2.3 Necrosis assays

Necrosis assays are generally used to check for membrane integrity of the cells after treatment. As shown in Figure 3, NPs can adhere to membranes causing changes in their structure and function, hence it’s important to carry out these assays. Neutral Red (2-amino-3 methyl-7-dimethylaminophenazoniumchloride) uptake assay is a widely used assay used to measure cytotoxicity. The Neutral Red dye is a weak, cationic, cell permeable dye taken up by viable cells and localized in the lysosome. It permeates into the cell by nonionic passive dilution and binds to the phosphate groups found on the lysosome matrix. The dye is extracted from the viable cells and spectrophotometrically measured [74, 75]. It is a sensitive dye that measures the integrity of the cell membrane and measures the viability of the cells [76]. The Neutral Red uptake assay is very sensitive and requires less equipment. It does not face much interference and does not have unstable reagents like other assays like MTT, XTS, etc. [77]. It is simple and can be used to detect only viable cells. It can also be done together with estimation of total protein content [78]. However, the use of Neutral Red, a fluorescent-based assay, is prone to interference from the NPs, which could lead to quenching of the dye. In a study to check the toxicity effect of silver NPs, Neutral Red, amongst other dyes, were observed to face interference from the particles [79]. Also, when applied on certain compounds that are volatile and insoluble in water, it faces problems in analysis as it mainly works when soluble in water [76]. Some chemicals could also cause the transformation of the dye into insoluble crystals, which could increase the toxicity estimation of the assay, giving a false negative result. When total protein estimation is done after the neutral red uptake assay, it could lead to reduction of the amount of protein estimated [78].

Figure 3.

The toxicity mechanisms induced by nanoparticles [73].

Trypan Blue Exclusion Assay is used to measure cytotoxicity of nanoparticle treatment on cells. The Trypan Blue is a dye that is absorbed by dead cells, leaving out viable cells and giving the actual number of viable cells after treatment.

Trypan Blue Exclusion Assay is very simple and quick to do, and it does not require any technical know-how. However, Trypan Blue assay is less sensitive and less reliable compared to other assays. It can be tedious and time-consuming when done on large number of samples [80]. Because of the use of hemocytometer, the possibility of making counting errors is present. Also, if the cells are not properly diluted or poorly dispersed in the counting chamber, it can contribute to the error [81]. The assay is insensitive as it cannot distinguish between live cells and cells that are gradually losing cell function. Another disadvantage of this assay is it can cause toxicity to mammalian cells [82].

2.4 Oxidative stress assays

Due to the high surface area to volume ratio of NPs, they elicit the production of reactive oxidative species (ROS) from the cells as shown in Figure 4, which could lead to further cellular damage. The ROS produced is measured using some of the oxidative stress assays.

Figure 4.

Induction of oxidative stress by nanomaterials [83].

2′,7′-dichlorofluorescein diacetate (DCFH-DA) is a nonionic, nonpolar fluorophore that penetrates the membrane and is sensitive to ROS. When DCFH-DA is internalized by the cell, it is enzymatically hydrolyzed by cytosolic esterases into dichlorofluorescein (DCFH), which remains in the cytosol. The nonfluorescent DCFH is then oxidized by hydroxyl radicals and cellular ROS into highly fluorescent dichlorofluorescein (DCF), which is then analyzed by flow cytometry or fluorescence microscopy [84].

Tritiated borohydride and 2,4-dinitrophenylhydrazine (DNPH) are used to measure protein carbonyl levels, which are indicative of oxidative damage. Protein carbonyls are reduced to alcohols by borohydride. This reduction is measured spectrophotometrically at wavelength absorbance of 340 nm. Also, when protein carbonyls react with DNPH a stable 2,4-dinitrophenyl (DNP) hydrazone product is generated and its absorbance is read between 360 and 390 nm [85]. DNPH is highly sensitive and specific and result from the assay can further be improved with the use of high-performance liquid chromatography (HPLC) or Western blotting [86].

Another means of assessing the production of ROS is through analysis of lipid peroxidation products like malondialdehyde (MDA) and 4-hydroxyl-2-nonenal (4-HNE). MDA is formed when large polyunsaturated fatty acids (PUFAs) are broken down and during the metabolism of arachidonic acid (AA) [87]. The by-product reacts with thiobarbituric acid (TBA) at pH 3.5 to produce MDA-TBA which is measured by a spectrophotometer at wavelengths of 532 nm and fluorescently at 553 nm [88]. 4-HNE is a by-product of AA and peroxidation of PUFAs. It reacts with primary amines to form Schiff’s bases and thiol. It also reacts with amino compounds to form Michael’s adduct which is detected by HNE-protein adduct ELISA assays [89, 90].

DCFH-DA, However, is prone to inaccurate interpretation due to nonspecific enzymatic oxidation and photooxidation [84, 91]. With the use of DNPH, larger samples are required for the analysis [92]. The lipid peroxidation is nonspecific since TBA reacts with other molecules other than MDA, which could lead to false positive results.

3. Ex vivo models

Using in vitro models for studying nanoparticle mediated toxicity comes with the set of challenges as discussed in the previous section. Animal models like rodents or zebrafish embryos are ideal alternatives for studying toxicity induced by NPs, however, there are various limitations including the cost of maintenance, biosafety etc. One also needs to follow appropriate ethical guidelines and have a moral compass to ensure whether inducing pain or sacrificing the animal can be avoided by alternate studies.

Ex vivo studies where tissue slices are cultured outside the host organism, experiments on fertilized eggs, organs on chip studies are more reliant than in vitro studies which are often based on monolayer culture of cells. The following section will briefly describe the prominent ex vivo models.

3.1 Precision cut tissue slices

In precision cut tissue slices, tissues are collected from a model organism or human biopsy samples for testing the effects of nanoparticle treatment. This approach provides a more comparable biological scenario rather than using immortalized or primary cells cultured in monolayer. As the organ tissues provide a better biological replica of the effect of nanoparticle in the biological system as a whole. Predominantly, this method has been used to study aerosol effects or nanocarriers developed for treating lung disorders [93, 94, 95, 96]. Using liver cut slices to study hepatotoxicity is also common [97, 98, 99, 100].

One of the major challenges of this method involves the penetration of nanocarriers into the organ slices. In a biological system, like our body, circulatory system ensures the uptake/delivery of nanocarriers. Without this in the organ slices which also has several physical barriers makes it difficult for the passive entry of NPs. As seen in some studies, the NPs are seen around the surface of the tissue slices [101]. Besides this, sometimes slicing a tissue induces inflammation around the cut region. This is unfavorable while studying toxicity effects. An inflamed tissue could introduce bias into the experimental outcomes. Inflammation could activate cells and induce necrosis or apoptosis which will interfere with toxicity results for nanoparticle testing.

3.2 Organ-on-a-chip

A 3D culture with multiple kinds of cell population sufficiently supported with extra cellular matrix and growth factors is better than monolayer culture in terms of nanotoxicology analysis. However, these kinds of system are considered to be static and does not mimic vascular perfusion or the sheer stress in biological systems. Recently, microfluidics have been integrated to 3D culture of different population to generate micro physiological systems called organ-on-a-chip. A very simplistic model of organ-on-a-chip will have a single type of cell lined over the microfluidic channel with a continuous flow of media. In advanced models like lung-on-a chip, blood brain barrier-on-a chip or blood retinal barrier-on-a-chip model, multiple types of cells are used with microfluidic channel mimicking biological processes like breathing, blinking [102, 103, 104, 105, 106, 107, 108, 109, 110, 111, 112, 113, 114, 115] etc.

These devices employ different perfusion rates and with the kind of material used, introduce mechanical strains which mimics biological circulation events like breathing, heartbeat, peristalsis, blinking, twitching etc. By using transparent materials for channel construction, the device also allows live visualization. As compared to conventional in vitro techniques, this provides a better physiological parallel with intricate designs to replicate organ architecture. Recently, there have been studies to develop multiple organ on a chip interconnected devices which could essentially be called body-on-chip to develop a model of whole organism rather than studying the effects on individual organs [116]. With intricate designs to carefully recreate biological systems, this could to a larger extent eliminate animal based testing.

Fluorescently labeled PEGylated gold nanoparticles were tested on tumor on a chip model which provided better insights into the circulation, elimination and uptake of nanoparticles in tumor microenvironment [117, 118]. Similarly Organ-on-a-Chip model are used in understanding the effects of shape of NP on its toxicity [119]. Understanding the effect of flow rates, especially while NP mediated targeted therapy for crossing BBB [120], the underlying changes in toxicity profile introduced by surface modifications of the particle [121] etc. are few of the recent advances in NP mediated research which has recently been revolutionized by Organ-on-a-Chip models. Specifically focusing on toxicity studies, lung-on-a-chip models have provided better insights into pulmonary toxicity of nanoparticles, specifically, TiO₂, Silica NPs [108, 122] etc. Also, the effect of NPs on crossing placental membrane has been evaluated using Organ-on-a-Chip model. In one such model, placental membrane integrity and maternal immune cell response were negatively challenged by TiO₂ NPs [123]. These models also enable the researchers to evaluate toxicity of NPs in both static and mobile conditions as opposed to static conditions in in vitro studies [124].

One major disadvantage of the devices is the use of precursors which adsorbs drugs while testing. This reduces the amount of drugs available for interacting with the cells of interest. The widely used material polydimethylsiloxane (PDMS) [125] is known to adsorb certain drugs which is a disadvantage to the application. Polysulphone based materials are considered to be an alternative for PDMS. However, they lack transparency and are difficult to tune the mechanical properties. If careful considerations are introduced while designing the architecture and appropriate precursor materials are chosen, organ-on-chip can be a lot desired application for nanoparticle toxicity analysis (Figure 5).

Figure 5.

The lung/liver-on-a-chip platform. (a) A photograph of the chip system comprising the pump main unit with four pump heads, the PEEK chip, and the reservoir plate. (b) a schematic view of the chip comprising four circuits is shown; each circuit includes two compartments to house the lung and liver tissues, respectively. The cross-section schemas of the plates show the path of the tubes and channels and the relative depth of each well. (c) a close-up view of the two compartments showing the groove pattern on the bottom of the wells. (d) Effect of chip materials on absorption of nicotine. A solution of 10 mM nicotine in phosphate-buffered saline (PBS) was kept in the wells of the chips made from PEEK or PDMS for 8 hours before it was collected. The concentrations of nicotine were then measured using liquid chromatography coupled to high-resolution accurate mass spectrometry. Nicotine concentrations remaining in solution are expressed as % relative to the stock solution. Data are presented as mean ± SEM. N = 3. PEEK: Polyetheretherketone; PDMS: Polydimethylsiloxane [126].

3.3 Chick chorioallantoic membrane (CAM) assay

Angiogenesis and neovascularization is a characteristic trait of multiple disorders including cancer and retinopathies. Preventing the formation of new leaky vasculature is thus critical in treating these disorders. NPs developed as potential therapeutic options thus need to be evaluated for their anti-angiogenic properties. Evaluating this on developing chick embryos is both cost and time effective other than the simplicity of the model [127, 128, 129, 130]. Briefly, the assay is performed by making a hole in the eggshell and inserting a membrane which delivers the NPs of choice. At the end of study, the eggs are opened, and the blood vessels are quantified using bright field imaging. One of the major advantages of this method is its reproducibility and simplicity to conduct in small scale laboratories.

3.3.1 Ex-vivo tumor models on CAM

Cancer cells of human origin are transplanted to CAM that covers chicken embryo. After 3 days of transplantation, a tumor with host species features will be developed. The tumor features multiple cell type of human origin, with rich vasculature and extracellular matrix. In terms of tumor microenvironment and cell types, this model is a closer approximation to in vivo models compared to 3D organoid studies. Also, the tumor formation process is rapid and eliminates longer waiting period as compared to rodents. Also, immune compromised animals are costlier in contrast to chick embryos which poses naturally underdeveloped immune system in early developmental stages. Besides these advantages, the rich nutrients present in chick embryo encourages effective angiogenesis in the tumor model. This facilitates drug and nanoparticle testing and monitoring of reduction of micro blood vessels and other tumor characteristics (Figure 6).

Figure 6.

Patient tumor sample transplanted on the CAM membrane. (A) Tumor formed by transplanting minced sample of ovarian cancer patient tumor. (B) Tumor is eliminated after intravenous injection of PMO-1 containing doxorubicin. (C) Chick embryo major organs look normal 3 days after injection [130].

3.3.2 Anti-angiogenesis studies for retinopathies

These are studies where implants are prepared and investigated on its ability to sustainably deliver therapeutic agents in ocular regions. Neovascularization and resultant increased ocular pressure lead to a plethora of ocular disorders including diabetic retinopathy, macular degeneration etc. Nanoparticle mediated implants or delivery systems attempt to target pro-angiogenic markers in the eye. However, this needs to be carefully implemented with increased attention on not to induce any inflammation and reduction in angiogenesis.

Briefly fertilized eggs will be candled to spot the air sac and blood vessels. A small hole of 1×1cm will be made into the egg shell at air sac region. Similarly, a cut is made near the vascular region without disturbing the CAM. Sucking out air from the air sac will distant the CAM from the egg membrane at the vascular region. Once this is achieved, a small incision is made at the CAM of vascular region where implants are introduced through sterilized filters. These filters containing the particle are then incubated over desired time frame. At the end of experiment duration, CAM is fixed and imaged to count the no of blood vessels compared to untreated controls (Figure 7 represents multiple ways of performing CAM assay for angiogenesis studies).

Figure 7.

Representative images of chorioallantoic membrane (CAM) variants. (A) in-ovo setup by windowing method on day of incubation; (B) ex-ovo setup in a petri plate; (C) ex-ovo setup in a glass-vertical view; (D) ex-ovo setup in a glass-horizontal view; (E) ex-ovo setup on plastic cups, image taken by a camera and; (F) ex-ovo setup in plastic cups, image taken by a Chemidoc (charge-coupled device (CCD) camera) [131].

For a simplistic and cost-effective model like CAM assay, most often the only disadvantages lies in the frequent contamination of samples. This can be eliminated with limited exposure of the opened CAM to outside air and by following stringent sterilization practices [132].

4. In vivo models

Nanocarrier formulations are tested in in vivo models as pre-clinical studies to evaluate its feasibility to escalate to clinical trials and into commercial market at the end of pipeline. The maximum dosage of NPs which could be safely tolerated, the pharmacokinetics and elimination window of NPs from the tested organisms, the accumulation and effect of long-term exposure is analyzed during these experiments. Vertebrates and invertebrate groups of animals are used for these studies to scale up from simplistic biological networks to understanding the effects on complex organisms genetically closer to human beings. Good lab practices are stringently adhered to while conducting these studies with appropriate ethical standards.

4.1 Invertebrates

Invertebrate models often have a shorter lifespan which aids researchers in testing nanoparticle toxicities. Due to their shorter life cycle, it’s feasible to understand and compare the effects of exposure of NPs in their developmental stages. It also benefits to conduct multiple rounds of testing within a shorter duration of time. The most established invertebrate model systems for nanoparticle toxicity include Caenorhabditis elegans and D. melanogaster.

4.1.1 C. elegans

C. elegans are nematodes which can grow up to 1 mm in size in its fully developed adult stage. They are often used for understanding nanotoxicity through oral uptake which is also the major form of nanotoxicity in human beings [133, 134, 135, 136, 137, 138]. They pose around 70–80% of gene homology with humans and have around 70% of major signal transduction pathways conserved as compared to human beings. They are also transparent in nature allowing to visualize and track the accumulation of fluorescent labeled NPs. (Figure 8).

Figure 8.

Induction of C. elegans major stress or host defense responses by SiNP treatment. Fluorescent images of worms treated with H₂O (a), SiNPs (b), and tunicamycin (c) with a GFP reporter for ER stress. Fluorescent images of worms treated with H₂O (d), SiNPs (e), and ethidium bromide (f) with a GFP reporter for mitochondrial stress. Fluorescent images of transgenic worms carrying the GFP reporter for oxidative stress subjected to H₂O (g), SiNPs (h), and H₂O₂ (i) treatment. Fluorescent images of transgenic worms carrying the GFP reporter for innate defense subjected to H₂O (j), SiNP treatment (k), and physical injury (l). The same magnification was used in all of the images (m). Quantitative analysis of fluorescent intensity fold change of worms treated with H₂O and SiNPs and a positive control group corresponding to (a–l). N ≥ 20. Error bars represent mean ± SEM; ** p < 0.01 [139].

Major studies conducted on testing nanoparticle formulation and its effect on C. elegans has found that there have been significant changes in oxidative response, reproduction and lysosomal signaling after oral uptake of particles. For example, treatment with TiO₂ NPs have reported to alter the expression levels of glutathione-s-transferase gene. Exposure to NPs in C. elegans could also be through their vulval slit or opening. The nematodes being hermaphrodite, interaction of NPs at vulval site and spermathecae, where sperms are stored and oocyte fertilization happens, could provide preliminary results on how NPs affect reproduction in organisms. Interestingly, a recent study has reported that exposure of TiO₂ NPs in C. elegans leads to decreased expression levels of pod-2 a gene known to have role in reproduction in the nematodes [133, 136]. Another study has reported the exposure of silver NPs leading to altered expression levels of proteases involved in lysosomal pathway related genes.

Generation of mutant strains of C. elegans has helped in understanding the effects of common particles like graphene oxide (GO), Silver, cadmium quantum dots and TiO₂. Tracing of these NPs is often reliable and easy in C. elegans due to the transparency of its body. However, C. elegans lacks organs like a well-developed lung, kidney, heart etc. which makes it difficult to draw comparisons with higher order species. Also, it only has a 70% homology with human genome with some critical signaling pathways completely absent.

4.1.2 D. melanogaster

D. melanogaster or fruit fly is another model well used for oral toxicity of NPs. Like C. elegans, fruit fly poses different life development stages which makes it suitable to study the effect of NPs on different life stages. Effects of NPs in the gut cells, eye and wing development are often studied for testing the toxicity of NPs (Figure 9). Along with this, behavioral studies are also conducted where crawling speed and path is monitored [140, 141]. Immune pathway in fruit flies is well studied and they also display a similarity of autophagy related genes with human beings. This makes D. melanogaster an ideal model for studying immune response and rate of autophagy with respect to NPs like GO [142]. The effect of nanoparticle exposure to the organisms is also analyzed by studying the effects on reproduction. Offspring number, morphology, development life stages are analyzed to understand the same [140, 143, 144, 145].

Figure 9.

Possible mechanism of nanoparticle-induced mortality in adult Drosophila. (A) Location of spiracles in drosophila: sp1, mesothoracic spiracle; sp2, metathoracic spiracle; sp3 to sp9, abdominal spiracles, image from Lehnmann et al. (B) SEM image shows mesothoracic and metathoracic spiracle of an adult drosophila (blue square) Center row (C − E): SEM images of spiracles in unexposed drosophila; sp1 (C), sp2 (D), both 20–50 μm, and an abdominal spiracle (E) at 5 μm. Bottom row (F − H): Spiracles are covered/decorated with nanomaterials (see arrows) after dry exposure of adults to CB (F); MWNTs (G); CB (H) differential toxicity of carbon nanomaterials in drosophila: Larval dietary uptake is benign, but adult exposure causes locomotor impairment and mortality [140].

The studies of nanotoxicity using fruit flies are limited to survival, developmental stages, eclosion rate, fertility and geotaxis performance analysis. Though it provides a significant addition in terms of understanding the toxicity as compared to in vitro studies, parallels cannot be drawn between human beings. One of the major challenges in using D. melanogaster as a model organism is the absence of an adaptive immunity in the organism. Other limitations include insufficient evidence of the cognitive capabilities of the organism especially affecting behavioral studies. Also, vertebrate specific genetic disorder models cannot be developed in D. melanogaster. Other than this, unlike other models, amount of ingestion of NPs per flies cannot be accurately standardized as they are not gavage fed. Even after the above-mentioned limitations they remain one of the simplest models along with C. elegans to develop transgenic lines by breeding.

4.2 Vertebrate models

Vertebrate models for studying nanotoxicity include zebrafish, rabbit rodent models like mice, rat, hamsters etc. They share higher similarities with human beings in terms of the respiratory, circulatory, and nervous system. However, studying long term exposure of NPs in these organisms, especially rodents, becomes challenging due to their longer lifespan and gestation periods. The following section will briefly discuss the recent reports of using these vertebrate models for nanotoxicology studies and their limitations.

4.2.1 D. rerio

Zebrafish or D. rerio is one of the few vertebrate models with shorter breeding and offspring rearing period. This allows it to be a vertebrate model with ease to study developmental stages and the effect of NPs on life cycle. Also, oral and circulatory introduction of NPs is possible through zebrafish. The oral toxicity studies are conducted by introducing NPs in embryonic medium or fish water. The NPs are introduced in the blood stream directly using micro injections or intravitreal, intraperitoneal, intraventricular injections. Zebrafish models have been used to develop blood brain barrier models, tumor models and studied to see the effect and penetration properties of these nano formulations [146, 147, 148, 149, 150].

Toxicity of nanoparticle formulation is often assessed by survival analysis, cardiac rhythm studies, morphological changes in eye, spine and fin development, edema in cardiac sac etc. Behavioral changes like swimming pattern, response to tapping and light are also investigated. Like fruit flies, conservation of autophagy related genes to human beings allows using zebrafish models for studying the autophagy related gene expression with respect to nanoparticle treatment. Recently, treatment of ZnO nanoparticle in zebrafish model has displayed an increase in inflammation related gene over expression [151]. There are also reports of siler NPs affecting the gut microbiota of zebrafish [152]. Embryonic zebrafish studies have also reported the effect of silica based NPs in reducing the blood pressure in zebrafish and vascular endothelial cells (Figure 10) [153].

Figure 10.

Inflammatory response and vascular endothelial cell dysfunction induced by SiNPs. (a,B) SiNPs increased the recruitment and chemotaxis of neutrophils in caudal vein of Tg(mpo:GFP) zebrafish. (C,D) SiNPs inhibited the expression of vascular endothelial cells in Tg(fli-1:EGFP) zebrafish. n = 30, data are expressed as mean ± standard deviation from three independent experiments (*p < 0.05). Scale bar: 100 μm [153].

One of the major limitations of using zebrafish is the inability to use them for studying respiratory effects of aerosol based nano formulations. Other than this limitation they are also not ideal for breast cancer and prostrate cancer models as they lack the appropriate tissue of origin in their body architecture.

4.2.2 Rabbit and rodent models

Rabbits were classified as rodents till the early 20th century. They along with models with mice, rat and limitedly hamsters have been used for studying pharmacokinetics and pharmacodynamics of nanoparticle formulations. Their organs are also harvested and used for tissue distribution studies as they have striking similarity in tissue characteristics with human beings. Ocular and dermal toxicity of NPs are mostly studied in rabbits. Recently, silver NPs were tested on shaved skin regions on albino rabbits and the toxicity was analyzed using prefixed criteria. Dry skin, scaling in doses lower than 100 ppm and erythema in higher doses up to 4000 ppm was observed as part of this study [154]. Nano-hydroxyapatite was intravenously introduced to New Zealand white rabbits and it was noted that they does not affect liver function, and renal function in the animals [155]. In another study conducted to understand the toxicity of aflatoxin B1, treatment with curcumin and ZnO NPs prevented lipid and protein degradation via oxidation and showed better liver health as compared to aflatoxin B1 treated groups [156]. In corneal fibrosis (haze) model in rabbit using excimer laser performing -9D photorefractive keratectomy (PRK), nanoparticle formulation containing BMP7 topical application 5 minutes after PRK reduced the corneal haze by 50 percent with no toxicity [157]. Nephroprotective effects of NPs have also been studied using rabbit models. Studies using CaO NPs however have reported significant toxicity in liver and kidney after exposure [158].

Rodent models of nanotoxicity mainly includes rats and mice. Immune compromised, genetically altered rodents are very commonly used in nanotoxicology studies and have been reviewed better in various book chapters and reviews elaborately [159, 160, 161, 162]. Briefly they are used for central nervous system disorder models, cancer models, hepatotoxicity models, aerosol treatment models. They provide relatively more homology with human beings; however, the laboratory maintenance of rodents and rabbits are often challenging. Also, historical trait analysis with respect to nanoparticle formulation is often limited by the number of offspring produced and longer gestational periods. Besides this, there is no accepted duration of days or guidelines recommending the number of days nanoparticle exposure in rodents to be monitored. This largely leads to inaccuracies and inconstancies in reporting toxicity of the same or similar nano formulation by different research groups.

4.2.3 Primate models

Primate models for nanotoxicity analysis rose to prominence after a study conducted on Rhesus macaque by intravenously injecting 25 mg/kg of phospholipid-micelle-encapsulated CdSe-Cds-ZnS QDs [163]. The study term was 90 days and the authors observed significant changes in behavioral patterns including loss of sleep, appetite, body weight, physical activity etc. By the end of study term, most of these effects were reversed to original state, however the particle displayed accumulation on liver and kidneys. Further studies on this were not carried out to understand the long term implications of hepatic accumulation. Quantum dots are generally considered to be safe to administer after conducting studies on in vitro and other in vivo models. However, this study indicated the need of behaviorally closer primate models to draw significant conclusion on nanoparticle toxicity before human trials.

Following this study, there were multiple reports of nanoparticle toxicity analysis using primates. Polylyisne conjugated DNA NPs for targeting retinal pigment epithelium was studied using baboons showing no inflammatory response in the eye [164]. Also in cynomolgus monkeys, the safety of gadolinium based NPs for imaging purpose was evaluated [165]. In another studies, PEG-bl-PPS polymerosomes were found to be nontoxic in non-human primates [166]. Also, cargo of siRNA in cyclodextrin with transferrin ligands as targeting moieties were also found to be relatively safe in cynomolgus monkeys [167]. However, there were elevated levels of creatinine and nitrogen along with certain inflammatory cytokines at higher dosage. In another study, mice and Rhesus macaque were compared on its effect of CdSe/CdS/ZnS semiconductor NPs in placental crossing and miscarriage rates. Interestingly, in the rodent models there was not toxicity recorded even though the particles were shown to cross the placenta. There was no miscarriage in rodents and fetus displayed no abnormalities. However, in primate models there was a 60 percentage rate of miscarriage establishing the fact that primate models are far superior in compared to rodent models for toxicity analysis [168].

Even though they are ideally the closest to understanding human body’s response to nano formulations, testing these formulations on them require more human resources and expertise along with stricter ethical guidelines in handling them. Also, these magnificent creatures are often sacrificed at the end of the study to harvest organs and understand tissue damage. Most of them being social organisms, this would have larger implications on their family group and could even lead closer ones to depression. Infant carrying is one such response displayed by mothers losing infants in primates. Secondary responses of curiosity and stress to death of infants or members is often displayed by these primates [169, 170, 171, 172]. A closer evaluation of the morality and scientific rationale should be evaluated before conducting such studies.

5. Conclusions

The past few decades witnessed the advent of nanoparticles and their potential use in multiple fields of biomedical sciences. From drug delivery to semiconductor devices, nanoparticles find applications around us. Informed use of nanomaterials, especially on its toxicity is highly relevant as more and more studies report the hazardous effects of these particles. The current chapter discussed in vitro, in vivo, ex vivo models for evaluating nanoparticle toxicity. As we analyze the plethora of assays conducted to study, in some cases, same particles in multiple model systems, we understand the varying toxicity reports. Such studies challenge the dangerous assumption of deeming a NP to be nontoxic by simply analyzing in vitro and in some cases rodent models. The need of non-human primate models closer to the genetic and physiological profile of human beings vs. the morale of sacrificing animal life for our benefit need to be carefully questioned. Alternate strategies like organ-on -chip models require further refinement and balance in incorporating parameters to better mimic study conditions.

Acknowledgments

Authors acknowledge the financial support provided by Blazer foundation and Medical Biotechnology Program at Department of Biomedical Sciences, UIC College of Medicine Rockford.

Conflict of interest

The authors express no conflict of interest.

Appendices and nomenclature

NP	Nanoparticles
ROS	Reactive Oxygen Species
MTT	3-[4,5-dimethylthiazole-2-yl]-2,5 diphenyltetrazolium bromide
XTT	2,3-bis[2-methoxy-4-nitro-5-sulfophenyl]-2H-tetrazolium-5-carboxanilide
CHO-K1	Chinese hamster ovary cells
Au-PEG-NPs	PEG coated gold nanoparticles
SWCNTs	Single Walled Carbon Nanotubes
GO	Graphene Oxide
FPG	formamidopyrimidine DNA glycosylase
TUNEL	terminal deoxynucleotidyl transferase biotin-dUTP nick end labeling
DCFH-DA	2′,7′-dichlorofluorescein diacetate
DNPH	2,4-dinitrophenylhydrazine
PUFA	Poly unsaturated fatty acids
TBA	Thio barbituric acid
CAM	Chick Chorioallantoic Membrane
PEG-bl-PPS	Polyethylene glycol-bI-polypropylene sulfide
AuNPs	Gold nanoparticles
PEEK	polyetheretherketone
PDMS	polydimethylsiloxane