This procedure was established to be used for developing advanced 3D hepatic cultures in vitro, which can provide a more physiologically relevant assessment of the genotoxic hazards associated with nanomaterial exposures over both an acute or long-term, repeated dose regimes.
Due to the rapid development and implementation of a diverse array of engineered nanomaterials (ENM), exposure to ENM is inevitable and the development of robust, predictive in vitro test systems is essential. Hepatic toxicology is key when considering ENM exposure, as the liver serves a vital role in metabolic homeostasis and detoxification as well as being a major site of ENM accumulation post exposure. Based upon this and the accepted understanding that 2D hepatocyte models do not accurately mimic the complexities of intricate multi-cellular interactions and metabolic activity observed in vivo, there is a greater focus on the development of physiologically relevant 3D liver models tailored for ENM hazard assessment purposes in vitro. In line with the principles of the 3Rs to replace, reduce and refine animal experimentation, a 3D HepG2 cell-line based liver model has been developed, which is a user friendly, cost effective system that can support both extended and repeated ENM exposure regimes (≤14 days). These spheroid models (≥500 µm in diameter) retain their proliferative capacity (i.e., dividing cell models) allowing them to be coupled with the ‘gold standard’ micronucleus assay to effectively assess genotoxicity in vitro. Their ability to report on a range of toxicological endpoints (e.g., liver function, (pro-)inflammatory response, cytotoxicity and genotoxicity) has been characterized using several ENMs across both acute (24 h) and long-term (120 h) exposure regimes. This 3D in vitro hepatic model has the capacity to be utilized for evaluating more realistic ENM exposures, thereby providing a future in vitro approach to better support ENM hazard assessment in a routine and easily accessible manner.
Due to the rapid development and implementation of a diverse array of engineered nanomaterials (ENM) across a plethora of human-based applications (e.g., food, cosmetics, clothing, sporting equipment, electronics, transport and medicine), it is inevitable that humans will be exposed to ENM on a regular basis. With this, there are heightened concerns that the novel, size specific physio-chemical characteristics that deem these materials advantageous in numerous applications could cause adverse effects upon human health and the environment concomitantly. Currently many international activities are in place to actively reflect more physiologically relevant exposures to these ENM and assess the potential toxicity of these materials over acute, long-term, and repeated low-dose exposure scenarios.
Hepatic toxicology is key when considering ENM exposure, as it is widely known that the liver is a major site of ENM accumulation post exposure1,2. Moreover, the liver is the primary organ system for metabolism and detoxification of substances that enter systemic circulation3. Based upon the accepted understanding that 2D hepatocyte models do not accurately mimic the complexities of intricate multicellular interactions or appropriately represent metabolic activity observed in vivo, a greater focus into developing robust and physiologically relevant in vitro 3D liver models for in vivo substitute technologies has been established4,5. Utilizing advanced 3D culture technologies improves the longevity of in vitro hepatic models allowing for long term, repeated exposure regimes to be investigated. Additionally, this advanced culture format promotes the formation of enhanced physiological, organotypic features such as bile canaliculi, active transporter processes and improved CYP450 drug metabolizing capabilities, thus improving the predictivity of the models6. Current 3D in vitro hepatic models consisting of mono-cultures (hepatocytes only) or co-cultures (hepatocytes with nonparenchymal cells) exist in several formats, ranging from microtissues or spheroids in ultralow adhesion plates, hanging drop spheroids, cells embedded in matrices and/or scaffolds and microfluidic cell culture platforms, all of which are deemed effective advanced in vitro models for hepatic toxicity assessment6,7. However, the majority of these model systems are high maintenance, require specialized equipment and are expensive. Furthermore, these models are often static (i.e., nondividing cell models) that prevents their use in the assessment of hazard endpoints, such as genotoxicity testing utilizing methods that quantify fixed DNA damage. Genotoxicity is a core prerequisite in regulatory toxicology, and it is a vital component of the risk assessment of any toxicant8. There is no single assay that can be applied to quantify all forms of DNA damage that may arise following exposure to an exogenous agent. However, a core component of the in vitro genotoxicity testing battery is the micronucleus assay, which is a reliable and multifacetted technique that measures gross chromosomal damage9. It is a gold standard technique described by the OECD Test Guideline 487, for assessing in vitro DNA damage and genotoxicity and is part of the test battery requirement for regulatory hazard assessment10,11.
The human hepatocellular carcinoma cell line, HepG2, is used widely for initial hazard assessment screening as the cells are readily available, relatively inexpensive to source, simple to culture and amenable to high throughput screening12,13. When cultured into 3D spherical structures, they have been shown to recapitulate the liver microenvironment well and offer a hepatic model with sufficient proliferative capabilities to support the micronucleus assay3. Further development of the HepG2 spheroid models was established to improve the longevity and liver-like functionality of the model in order to support genotoxicity hazard assessment over long-term, repeated exposure regimes (≤14 days). Thus, in line with the principles of the 3Rs to replace, reduce and refine animal experimentation, the present protocol has been established to provide an advanced 3D in vitro hepatic model capable of reliably evaluating multiple toxicological endpoints (e.g., liver functionality, (pro-)inflammatory markers, cytotoxicity and genotoxicity) following acute, long-term and repeated chemical and ENM exposures in a routine and easily accessible manner.
Here, we present a method to establish a physiologically relevant 3D hepatocyte cell line based in vitro model system for genotoxicity hazard assessment following acute or long-term, repeated ENM exposures. The protocol can be broken down into 6 key stages: culturing cryopreserved HepG2 cells; HepG2 spheroid preparation; HepG2 spheroid transfer from hanging drop to agarose suspension; HepG2 spheroid harvest; micronucleus assay and scoring; and data analysis.
1.Culturing cryopreserved HepG2 cells
NOTE: HepG2 cells, obtained from American Type Culture Collection (ATCC) were cultured in 1x Dulbecco's Modified Eagle Medium (DMEM) with 4.5g/L D-glucose and L-glutamine supplemented with 10% foetal bovine serum (FBS) and 1% penicillin/streptomycin antibiotic.
2. HepG2 spheroid preparation
Figure 1: Counting cells using a haemocytometer. Diagrammatic representation of a haemocytometer highlighting which quadrant to count cells from. Please click here to view a larger version of this figure.
Figure 2: 3D HepG2 in vitro spheroid model preparation. (A) The HepG2 cells seeded in 20 µL drops onto the lid of a 96-well plate. (B) The HepG2 cells post-seeding in the hanging drop model to allow for spheroid formation. Please click here to view a larger version of this figure.
3. HepG2 spheroid transfer from hanging drop to agarose suspension
NOTE: On Day 3 post seeding into hanging drops, the spheroids are transferred into the wells of the same 96-well plate all of which have been previously coated with a fine layer of 1.5% agarose gel.
4. Nanomaterial/Chemical exposure
NOTE: The HepG2 liver spheroid model can support both ENM and chemical based exposure regimes, but the primary focus of this protocol is ENM exposures. Prior to exposure, the test ENM must be suitably dispersed; this can be performed as directed by the NanoGenoTox Dispersion Protocol (Grant Agreement No. 20092101, 2018)15.
5. HepG2 spheroid harvesting
NOTE: Following either chemical or ENM exposure treatments, both cell culture medium or spheroid tissue can be harvested for multiple endpoint analysis. Depending on the endpoint analysis, spheroids can either be harvested individually (e.g., for image analysis) or pooled together (e.g., for cytokinesis block micronucleus assay).
6. Micronucleus assay and scoring
For the manual method of the micronucleus assay, a cytocentrifuge is required to produce a cytodot (a defined, concentrated region of cells) in the centre of the microscope slide. This process supports more efficient scoring of the slide as it allows the scorer to easily locate the cells of interest, as opposed to evaluating a whole slide where the cells can be widely spread.
Figure 3: Cytospin setup to prepare treated cells on microscope slides. (A) Displays the individual components, (i) cuvette funnel, (ii) filter card, (iii) glass microscope slide and (iv) metal support required to cytospin HepG2 cells onto microscope slides. (B) The final cuvette funnel set up. (C) The correct placement of the cuvette funnel within the cytocentrifuge. Please click here to view a larger version of this figure.
7. Data analysis
Figure 4: Micronucleus assay scoring decision tree. Schematic decision tree to highlight the necessity for different scoring and cytotoxicity assessment procedures when using the micronucleus assay with 3D models following acute or long-term exposure regimes. Acute (≤24 h) exposures allow use of the cytokinesis blocked micronucleus assay, while long-term (≥24 h) exposures require the mononuclear version of the assay; both of which are described in OECD Test Guideline 487. Please click here to view a larger version of this figure.
Suitability of this cell-line based 3D liver spheroid model for long-term culture and genotoxic hazard assessment was evaluated by conducting baseline characterization to determine the viability and liver-like functionality of the model over the duration of 14 days in culture as well as its applicability for the micronucleus assay.
Baseline Characterization of the 3D HepG2 Liver Spheroid Model
Prior to any in vitro toxicological assessment, it is important to check that the 3D HepG2 spheroids have formed properly before performing the agarose transfer or chemical/ENM treatment. HepG2 spheroids produced using the hanging drop method usually take 2 – 3 days post seeding (4000 cells/spheroid) to form compact, spherical shaped spheroids with an average diameter of 495.52 µm W x 482.69 µm H as shown in Figure 5A-5C. HepG2 spheroids that have formed correctly and are acceptable to be used for in vitro toxicological assessment must have a compact, spherical shaped structure with a smooth surface and no visual projections. Figure 5 provides examples of good quality (Figure 5D-F) and a poor quality (Figure 5G-I) spheroids. The latter of which should be discarded. Typically, 90-95% of spheroids formed per plate will form correctly and be viable for further experimentation.
Figure 5: Light microscopy images displaying the natural morphology of the HepG2 spheroids formed via the hanging drop method. (A-C) show Day 2 and (D-I) Day 4 HepG2 liver spheroids post seeding. (D-F) are examples of good quality HepG2 spheroids whilst (G-I) shows poorly formed spheroids. All images were taken on a X20 objective using a microscope. The scale bar represents 20 µm. Please click here to view a larger version of this figure.
To further confirm HepG2 spheroid viability, a basic colorimetric Bromocresol Green Albumin (BCG) Assay or Urea Assay can be performed to assess their liver-like functionality. Liver-like functionality was assessed in line with viability using the Trypan Blue Exclusion Assay over a 14 day culture period to determine the longevity of the liver spheroid model and establish if it could support long-term or repeated ENM/chemical based hazard assessment (Figure 6). Albumin concentration remained consistent over the duration of the culture period. Urea production displays an increase in the concentration of urea produced per spheroid over a week in culture before reaching a plateau by day 7. It is important to note that the levels of albumin and urea produced in the 3D HepG2 spheroids are substantially higher than that observed in the same cell line cultured in a 2D format. Indeed, 2D cultures of HepG2 cells, peak albumin and urea levels were 0.001 mg/mL and 0.010 ng/µL respectively. Furthermore, in previous work published by Shah et al. using an almost identical HepG2 spheroid system, the authors highlight a notable improvement in metabolic activity (CYP1A1 and CYP1A2) in the 3D HepG2 in vitro model systems when compared to the 2D cultured HepG2 cells5.
Figure 6: 14-Day baseline characterisation data for HepG2 liver spheroids. Following transfer from hanging drop, (A) highlights the viability of the HepG2 spheroid model over a 14 day period whilst (B) and (C) highlight the liver-like albumin and urea functionality respectively. Mean data ± SEM presented, n = 4. Please click here to view a larger version of this figure.
With the inevitable development of a necrotic core, a known limitation of 3D liver spheroid cultures, the viability of this HepG2 based model had to be established to demonstrate it was able to sustain long-term (5-10 day) exposure regimes whilst maintaining the proliferative capability required to support the micronucleus assay5. Indeed, this 3D liver spheroid model has been shown to retain >70% viability over 10 days in culture. Based on this and in conjunction with the sustained liver-like functionality observed over the ≥14 day culture period, this 3D liver spheroid model can thus support long-term, repeated ENM exposure regimes up to 10 days long (i.e., before viability of the spheroids drop below 70%). For reference, it is advised that albumin levels for HepG2 spheroids seeded at 4000 cells/spheroid should be ≥20.0 ng/µL whilst urea production should be ≥0.25 ng/µL before conducting an in vitro toxicological assessment with this model.
Genotoxicity Assessment of Engineered Nanomaterials
For genotoxicity assessment, the micronucleus assay was used to determine the presence of micronuclei following both acute (24 h) and long-term (120 h) ENM exposures. Aflatoxin B1 is a known liver carcinogen16,17 and is a recommended positive control for the micronucleus assay. Optimization experiments have shown that 0.1 µM of Alfatoxin B1 induces a significant positive (≥2.0 fold increase) genotoxic response in 3D HepG2 liver spheroids and thus is used in every micronucleus assay conducted with this model. To ensure the validity of the micronucleus assay results using the HepG2 spheroid model, the background micronucleus frequency for HepG2 cells used in this 3D in vitro model should lie within a range of 0.6% – 1.2%. As a result, Alfatoxin B1 should induce a genotoxic response of at least two-fold higher than that seen with the negative control; thus, 0.1 µM of Alfatoxin B1 should induce a micronuclei frequency between 1.5% – 3.0%. Using these control parameters, ENM associated genotoxicity in vitro can then be reliably assessed. Based on OECD Test Guideline 487, it is important to note that when testing an ENM or chemical, the concentrations selected should not induce more than 55% ± 5% cytotoxicity (indicated by a reduction in CPBI or RVCC values in relation to the negative control)11. Figure 7 illustrates the data generated when Aflatoxin B1 and two ENMs (titanium dioxide (TiO2) and sliver (Ag)) were evaluated following both acute and long-term exposures in the HepG2 spheroids, and subsequent genotoxic potential was analyzed using the micronucleus assay. Both ENMs assessed were tested at a noncytotoxic, low dose of 5.00 µg/mL over an acute (24 h) exposure and long-term (120 h) exposure regime. A similar trend for genotoxicity across both TiO2 and Ag ENMs can be observed, whereby the elevated genotoxicity response that resulted following 24 h exposure was not evident after a long-term 5 day exposure. This was despite sustained genotoxicity induced by the Aflatoxin B1 positive control at both time-points.
Figure 7: Genotoxicity assessment following TiO2 and Ag ENM exposure on HepG2 liver spheroids. Genotoxicity (micronucleus frequency) assessment using the micronucleus assay post (A) acute (24 hour) and (B) long-term (120 hour) exposure to 5.00 µg/mL of TiO2 and Ag ENM. Negative control is a media only, whilst the positive control is 0.1 µM of Aflatoxin B1. Mean data (n=2) presented ± SD. Significance indicated in relation to the negative control: * = p≤ 0.05. Please click here to view a larger version of this figure.
Applications for 3D hepatic models vary considerably depending on the particular biochemical endpoint or adverse outcome pathway being targeted. Each model has its benefits and limitations, from interdonor variation in primary human hepatocyte (PHH) models to reduced cytochrome p450 activity in cell-line based models, but all are valuable in their own right6,12,18,19. When assessing genotoxicity there are limitations in the models compatibility with regulatory approved endpoints such as the in vitro micronucleus assay, as active proliferation is required. This is necessary, as genotoxicity assessment requires the quantification of fixed DNA damage to be assessed post cell division when there is opportunity for DNA repair to correct transient lesions. Unfortunately, highly differentiated hepatocyte (i.e., HepaRG) based spheroids or PHH microtissues, which are deemed to exhibit the most physiologically relevant liver-like characteristics form static (non-proliferative) models12,19,20. As a result, the 3D HepG2 spheroid model presented here provides a suitable, alternative model able to support genotoxicity testing. HepG2 cell-line based spheroids have sufficient actively dividing cells on the outer surface of the spheroids whilst maintaining basic liver-like characteristics, such as albumin and urea production and some CYP450 activity5,12,19. Principally this in vitro liver model has been developed to complement the micronucleus assay, as this is one of the two in vitro assays recommended in the battery for genotoxicity testing8,10,11,21. However, the model can be readily applied to DNA sequencing analysis and gene expression (RNA) technologies, while it has the potential to be further adapted and utilized for other DNA damage endpoints, such as the comet assay. Nonetheless, it is important to consider the role that ENM interference plays in some endpoint analyses. For example, flow cytometry-based analyses may not be suitable for ENM genotoxicity assessment specifically due to particle interference22.
One limiting factor of spheroid models that actively undergo cell division is their size. Optimization of seeding density is critical as there needs to be enough cells that allow the model to continue to proliferate; but not too high a cell number, which results in the spheroid becoming overly compact, leading to an increased necrotic core. The cause of this necrosis is believed to be restricted oxygen and nutrient diffusion, as the limit of this diffusion is thought to be approximately 100 – 150 µm of tissue23,24. However, this does depend of the cell type, cell number, scaffold interactions and culture conditions25. Since, it has been shown that approximately 700 µm diameter is the limit for avoiding premature onset of necrosis in the center of C3A spheroids, seeding 4000 HepG2 cells per spheroid ensures the diameter of the model at the time of exposure is ≤500 µm26. Furthermore, Shah et al. established that HepG2 cells seeded above 5000 cells per spheroid exhibited a 25% reduction in viability following 7 days in culture, which could pertain to the average diameter of 680 µm and limited availability of nutrients in a 20 µL hanging drop5. To overcome this, the model devised in the present protocol undergoes a critical step where the hanging drop is transferred to agarose coated wells following initial formation of the spheroid. This ensures a greater volume of culture medium is present to sustain the ever-growing number of cells within the spheroids. As a result, the HepG2 spheroid model remains over 70% viable following 10 days in culture and can be utilized for long-term hazard assessment in vitro.
Whilst the HepG2 spheroid model can support both acute and long-term exposure regimes, refreshing cell culture medium during extended culture periods is restricted for this model as complete replacement of the medium is not advised due to the potential loss of the spheroids. It is presumed that with ENM exposures, the tendency for homogenous ENM dispersions to agglomerate and sediment is high. However, it is notable that the rate at which an ENM sediments can vary depending on the particle parameters (e.g., size, shape and density) and can be determined theoretically using the in vitro sedimentation, diffusion and dosimetry (ISDD) model, or its recent derivatives, often referred to when regarding ENM (suspension) exposure approaches27,28. With this is mind, it is assumed that if only 50% of the cell culture medium is carefully removed from the surface of the cell culture, the disruption and subsequent removal of the ENM dose should in theory be minimal. However, with Brownian motion at play, this may not strictly be the case and further work into the deposition and sedimentation of each particular ENM to be tested should be undertaken to ensure the correct dosimetry is retained throughout the long-term exposure regimes27. Principally this is a potential limitation to consider when performing repeated dosing regimes as this could be critical to the final, accumulated concentration. Chemical based exposures on the other hand, whilst not without their own limitations to consider, offer a more simplistic approach in that chemical substances tend to remain in solution and thus a direct replacement of the original chemical concentration in addition to the newly added concentration ensures that any chemical lost during media refreshment is replaced accordingly29. Future applications would include evaluating the suitability of the model for repeated exposure regimes over long-term culture periods as repeated dosing strategies are crucially important for assessing the ability of a particular organ system to ameliorate or overcome the adverse effects, if any, induced by bioaccumulation of a xenobiotic substance.
In conclusion, this 3D in vitro hepatic model has the capacity to be utilized for evaluating a range of realistic exposure scenarios, thereby providing a future in vitro approach to better support both ENM and chemical hazard assessment in a routine and easily accessible manner.
The authors have nothing to disclose.
The authors would like to acknowledge that this research has received funding from the European Union’s Horizon 2020 research and innovation programme for the PATROLS project, under grant agreement No.760813
Aflotoxin B1 | Sigma Aldrich, UK | A6636-5MG | |
Agarose | Sigma Aldrich, UK | A9539-50G | |
Autoclave Tape | |||
BCG Albumin Assay | Sigma Aldrich, UK | MAK124 | |
Bovine Serum Albumin Powder | Sigma Aldrich, UK | A9418 | |
Cell Freezing Aid | Thermo Fisher Scientific, UK | 5100-0001 – Mr Frosty | |
Centrifuge | Eppendorf | 5810 R | |
Cytochalasin B | Merck, UK | 250233 | |
Cytology Metal Clips | |||
Cytospin 4 Centrifuge | ThermoFisher Scientific, UK | CM00730202 | |
DMEM with 4.5g/L D-Glucose, L-Glutamine | GIBCO, Paisley, UK | 41965-039 | |
DMEM, phenol-red free with 4.5g/L D-Glucose, L-Glutamine with Hepes | GIBCO, Paisley, UK | 21063-029 | |
DPX Mounting Medium | FisherScientific, UK | D/5330/05 | |
Ethanol | FisherScientific, UK | 10048291 | |
FBS | GIBCO, Paisley, UK | 10270-106 | |
Filter Cards for Shandon Cytospin | FisherScientific, UK | 15995742 | |
Frosted Glass Slides | ThermoFisher Scientific, UK | ||
Giemsa's Stain Improved R66 Solution, Gurr | VWR Chemicals, UK | MFCD00081642 | |
Glass Coverslips (24 x 60) | Deckglaser, VWR | ECN631-1575 | |
Haemocytometer and Coverslip | |||
Immersion Oil for Microscope | Zeiss, UK | 518F, ISO8034 | |
Laminar Class II Tissue Culture Hood | Scanlaf Mars | ||
Light Microscope | Zeiss, UK | Axiovert 40C | |
Liquid Nitrogen | |||
Methanol | FisherScientific, UK | 10284580 | |
Microwave | |||
Non-Filtered, Sterile 200µl and 1000µl Pipette tips | Greiner-Bio-One, UK | ||
NuncMicroWell 96-Well Microplates | ThermoFisher Scientific, Denmark | 167008 | |
P1000 and P200 micropipettes | |||
P300 and P50 multi-channel pipettes | |||
PBS pH 7.4 1X, MgCl2 and CaCl2 Free | GIBCO, Paisley, UK | 14190-094 | |
Pen/Strep | GIBCO, Paisley, UK | 15140-122, Penicillin/Strepmyocin 100X or 10,000U/ml | |
Phosphatase Buffer Tablets | GIBCO, Paisley, UK | 10582-013 | |
Pipette Boy | |||
Simport Scientific CytoSep Funnels for Shandon Cytospin 4 Centrifuges | FisherScientific, UK | 11690581 | |
Sonifier SFX 550 240V CE 1/2" – Probe | Branson, USA | 101-063-971R | |
T-25 and T-75 Tissue Culture Flask | Greiner-Bio-One, UK | T-25 (690175) and T-75 (660175) | |
Trypan Blue Solution | Sigma Aldrich, UK | T8154-100mL | |
Urea Assay Kit | Sigma Aldrich, UK | MAK006 | |
Virkon Disinfectant | DuPont, UK | Rely+On Virkon | |
Water Bath (37˚C) | Grant JBNova 18 | ||
Weighing Balance | |||
Xylene | FisherScientific, UK | 10588070 | |
0.05% Trypsin-EDTA | GIBCO, Paisley, UK | 5300-054 | |
0.2mL and 1.0mL Eppendorf Tubes | Greiner-Bio-One, UK | ||
0.45µm Filter Unit | Millex HA, MF-Millipore, UK | SLHA033SS | |
1.0mL Syringe | BD Plastipak, FisherScientific, UK | 300185 | |
20mL LS Scintillation Glass Vials, 22-400 Foil Lined PP Caps | DWK Life Sciences GmbH, Germany | WHEA986581 | |
37˚C and 5% CO2 ISO Class 5 Hepa Filter Incubator | NUAIRE DHD Autoflow | ||
3mL Pasteur Pipette | Greiner-Bio-One, UK | ||
50mL Conical Falcon Tubes | Greiner-Bio-One, UK | ||
50mL or 100mL Glass Bottles | |||
50mL Skirted Falcon Tubes | Greiner-Bio-One, UK | ||
5mL, 10mL and 25mL Pipettes | Greiner-Bio-One, UK | ||
9.4cm Square, Petri Dish | Greiner-Bio-One, UK | 688161 |