Document [original]

Toxicology Reports 7 (2020) 1578–1587

Available online 21 November 2020

3D-bioprinted HepaRG cultures as a model for testing long term aflatoxin

B1 toxicity in vitro

Konrad Schmidt, Johanna Berg, Viola Roehrs, Jens Kurreck, Munir A. Al-Zeer*

Department of Applied Biochemistry, Institute of Biotechnology, 4/3-2, Technische Universit¨

at Berlin, Gustav-Meyer-Allee 25, 13355 Berlin, Germany

ARTICLE INFO

Keywords:

3D-bioprinting

Aflatoxin B1

Alternative in vitro model

HepaRG liver cells

ABSTRACT

In recent years 3D-bioprinting technology has been developed as an alternative to animal testing. It possesses a

great potential for in vitro testing as it aims to mimic human organs and physiology. In the present study, an

alginate-gelatin-Matrigel based hydrogel was used to prepare 3D-bioprinted HepaRG cultures using a pneumatic

extrusion printer. These 3D models were tested for viability and metabolic functions. Using 3D-bioprinted

HepaRG cultures, we tested the toxicity of aflatoxin B1 (10 or 20

M) in vitro and compared the results with

2D HepaRG cultures. There was a dose-dependent toxicity effect on cell viability, reduction of metabolic activity

and albumin production. We found that 3D-bioprinted HepaRG cultures are more resistant to aflatoxin B1

treatment than 2D cultures. Although the metabolic activities were reduced upon treatment with aflatoxin B1,

the 3D models were still viable and survived longer, up to 3 weeks, than the 2D culture, as visualized by fluo-

rescence microscopy. Furthermore, albumin production recovered slightly in 3D models after one and two weeks

of treatment. Taken together, we consider using 3D-bioprinting technology to generate 3D tissue models as an

alternative way to study toxicity in vitro and this could also provide a suitable alternative for chronic hepato-

toxicity studies in vitro.

1. Introduction

In vitro cell cultures and animal experiment models are crucial in-

struments in basic research and preclinical studies [1–3]. Even though

cell cultures and animal models are widely used in the field of toxicology

testing, recent advances in 3D-bioprinting technology are emerging as a

useful tool for testing of complex cell environments, toxicity and for

drug discovery in vitro [4,5,42] . Traditional 2D cell culture systems

using cell lines or primary cells do not recapitulate the natural

three-dimensional microenvironment of the respective human tissue or

organ, which in turn leads to changes in gene and protein expression [1,

6]. On the other hand, beyond the inherent complexity of animal

models, the microenvironments of animal tissues do not efficiently

recapitulate many of their human counterparts [7]. The creation of

microenvironments in vitro that capture some features of human tissues

could be enabled by advances in the understanding of 3D-bioprinting

[8]. For this reason, 3D-bioprinting is considered a promising tool to

study infection, cancer, drug screening, and toxicity testing [9,10]. This

can reduce or replace the need for animal testing that does not precisely

reflect human responses in terms of toxicology and pathophysiology.

Although 3D-bioprinting is a cutting edge technology used to fabri-

cate 3D models in vitro that hold tremendous promise, especially in the

fields of biology and medicine, 3D-bioprinting technology and bio-

materials are still being fine-tuned [11]. Bioprinting offers great preci-

sion that enables the exact spatial and temporal deposition of biological

materials, including cells and extracellular matrix in a 3D hierarchal

organization to generate artificial 3D models of native tissues. The

precise deposition and spatial arrangement of these components can

recapitulate the natural architecture and the microenvironments of the

tissue in vitro [11]. Both natural and synthetic polymers, such as collagen

[12], gelatin [13], and alginate [14] etc., are used extensively to fabri-

cate scaffolds for 3D model printing due to on their biocompatibility,

printability and biodegradability properties [15].

Several 3D mini-organs and tissue constructs have been printed and

fabricated in vitro, including skin, heart, bone, cartilage, lung, neurons,

and pancreas using different scaffolds and utilizing various 3D-bio-

printing approaches [11]. Liver 3D mini-models have also been gener-

ated in vitro using various tissue engineering techniques including

3D-bioprinting [16]. Various cell models have been used to fabricate

3D liver model in vitro including the use of human iPSC-derived

* Corresponding author.

E-mail address: [email protected] (M.A. Al-Zeer).

Contents lists available at ScienceDirect

Toxicology Reports

journal homepage: www.elsevier.com/locate/toxrep

https://doi.org/10.1016/j.toxrep.2020.11.003

Received 14 July 2020; Received in revised form 4 November 2020; Accepted 5 November 2020

Toxicology Reports 7 (2020) 1578–1587

1579

hepatocytes [4], HepG2 cells [17], and HepaRG cell line [18]. The

HepaRG cell line has been shown to be stable in vitro, allowing long-term

culture [18]. Liver models using 3D-bioprinted cells are becoming more

common in clinical and basic research, including infection, drug dis-

covery, and toxicity testing [19,20]. As the liver plays a central role in

metabolism and detoxification of chemicals, 3D-bioprinting possesses

great potential to achieve these goals and model complexity that mimics

the liver functions in vitro [43]. Therefore, 3D-bioprining has the po-

tential to become an important tool for in vitro toxicity testing especially

for studies of chronic hepatotoxicity against toxins such as aflatoxins.

Aflatoxins are secondary fungal metabolites, known as mycotoxins,

which are mainly produced by the genus Aspergillus [21,22]. They are

classified as Group 1 carcinogen by the International Agency for

Research in Cancer [23]. Aflatoxins are hepatotoxic and have been

implicated in increasing the risk of hepatocellular carcinoma. Aflatoxin

B1 has been shown to induce cytotoxicity in HepaRG cells [44]. This

hepatotoxicity is mediated by its toxic epoxide, which is produced by

CYP1A2 and CYP3A4 [24]. Further, aflatoxin B1 induces loss of cell

viability due to the induction of apoptosis in HepaRG cells. This

phenotype has been correlated with DNA damage response mediated by

p53 signaling [24,25].

In this study, we aimed at optimizing a 3D liver cell model using

bioprinting technology with an alginate/gelatin/Matrigel-based scaf-

fold, which can be used for extrusion-based bioprinting of HepaRG cell

models for toxicity testing. The 3D liver constructs were generated using

a 3D micro-extrusion bioprinter. Viability, cytotoxicity and metabolic

activity of the printed constructs were evaluated. Upon treatment with

AFB1, cells grown in 2D did not survive the damage induced by the

toxin, unlike the 3D-bioprinted constructs. Although AFB1 reduced

metabolic activity in the 3D-bioprinted constructs, cells survived the

AFB1 toxicity and were still viable, as visualized by fluorescence mi-

croscopy. Therefore, a 3D-bioprinted model may pave the way to study

the long-term effect of AFB1 and carcinogenesis in vitro.

2. Materials and methods

2.1. Cell culture

HepaRG cells were obtained from Biopredic International (Saint

Gregoir´

e, France). The cells were cultured in William’s E medium

(Gibco, Dreieich, Germany) supplemented with 2 mM L-Glutamine

(Biowest, Nuaill´

e, France), 10 % fetal bovine serum (FBS; c.c.pro,

Oberdorla, Germany), 50

M hydrocortisone hemisuccinate (Sigma,

Steinheim, Germany), 5

g/mL recombinant human insulin (PAN

Biotech, Aidenbach, Germany), and 1 % penicillin/streptomycin (P/S;

Biowest). The cells were cultured at 37 ◦C and 5 % CO

in a humidified

incubator for 14 days before differentiation and the medium was

renewed every three days. After 14 days, hepatic differentiation was

induced by adding 1.7 % DMSO (Sigma) to the culture medium for

additional 14 days as previously described [20].

2.2. Preparation of cell-scaffold hydrogels

Sodium alginate (4.5 % w/v) and gelatin (6.5 % w/v) powders

(Sigma) were dissolved in William’s E medium on a magnetic stirrer at

1250 min

−1

(overnight at 37 ◦C). The hybrid alginate -gelatin hydrogel

(450

l) was mixed with liquid Matrigel (200

l) (Corning, Tewksbury,

MA, USA), differentiated HepaRG cells, 1.22 M CaSO

(25

l) (Roth,

Karlsruhe, Germany) and William’s E medium containing supplements

(325

l) to obtain the final cell-scaffold mixture bioink composed of

alginate (2 % w/v), gelatin (3 % w/v), Matrigel (20 % w/v), 30 mM

CaSO

and 7 * 10

HepaRG cells/ml, as previously described [26].

Following the initial cross-linking of alginate using CaSO

(8 min after

mixing), the cell-scaffold hydrogel was loaded into the printing cartridge

(Cellink, Gothenburg, Sweden).

2.3. 3D bioprinting

The Cellink INKREDIBLE+3D-printer was utilized for the bio-

printing process which was carried out at room temperature. A

rectangular-shaped construct (1 mm height x10 mm width x10 mm

length) was designed with regularly spaced pores in a grid-like pattern

using Slic3r software. The shape was selected to allow the diffusion of

nutrients, oxygen, and metabolites through the constructs. The hydrogel

was extruded through a 22 G needle at 10–40 kPa to generate 3D-liver

constructs designed by the computer-aided design (CAD) software Rhi-

noceros5 (Robert McNeel & Associates, Seattle, WA, USA). The printed

constructs were further cross-linked using 100 mM CaCl

(Roth) to in-

crease the gelation of alginate. Then, constructs were cultured with

William’s E medium supplemented with 1.7 % DMSO, as well as 20 mM

CaCl

, and incubated at 37 ◦C and 5 % CO

in a humidified incubator.

2.4. Staining with fluorescent DNA dye and immunofluorescence

Printed 3D constructs were fixed for 30 min using 4 % formaldehyde.

The constructs were then permeabilized with 1 % Triton-X-100 (Roth)

for 1 h at room temperature and the nuclei were stained with Hoechst

dye (1

g/ml) (H33342, AppliChem, Darmstadt, Germany). Cellular

distribution was analyzed with the Zeiss Observer. Z1 microscope (Zeiss,

Jena, Germany).

2.5. The live/dead viability assay

To determine cell viability, we used a commercial Live/Dead assay

(Viability/Cytotoxicity kit; ThermoFisher Scientific, Waltham, MA,

USA). Briefly, the 3D liver constructs were incubated with 2

M calcein-

AM and 4 mM ethidium homodimer-1 diluted in 1x HBSS (ThermoFisher

Scientific) for 15 min (37 ◦C, 5 % CO

). Subsequently, samples were

analyzed by fluorescence microscopy (Zeiss Observer. Z1 microscope;

Zeiss, Germany).

2.6. XTT assay

Metabolic activity of HepaRG cells printed constructs or 2D cultures

was determined using the tetrazolium hydroxide salt (XTT) assay ac-

cording to the manufacturer’s instructions (AppliChem, Germany) at

indicated time points. The absorbance at 450 nm, with a reference of

620 nm, was carried out using TriStar Multimode Reader LB942 (Bert-

hold Technologies, Bad Wildbad, Germany). Cell-laden constructs

incubated in 10 % Triton-X-100 (Roth), which was diluted culture me-

dium were used as lysis control. Values were normalized to the lysis

controls.

2.7. Lactate dehydrogenase release

The supernatants from different constructs were collected at each

time point, as indicated, and lactate dehydrogenase colorimetric assays

(Roche, Switzerland) were carried out according to the manufacturer’s

instructions. The absorbance at 490 nm, with a reference of 620 nm, was

measured with the Sunrise microplate reader (Tecan, M¨

annedorf,

Switzerland).

2.8. Albumin measurement

The supernatants from 3D constructs or 2D monolayers were

analyzed and quantified for albumin secretion using the human albumin

enzyme-linked immunosorbent assay (ELISA) kit (Bethyl Laboratories,

Montgomery, TX, USA), according to the manufacturer’s instructions.

The absorbance at 450 nm, with a reference of 620 nm, was carried out

using TriStar Multimode Reader LB942.

K. Schmidt et al.

Toxicology Reports 7 (2020) 1578–1587

1580

2.9. Treatment with toxins

Differentiated HepaRG cells were treated with 10, or 20

M aflatoxin

B1 (AFB1) (Sigma-Aldrich). The AFB1 was dissolved in dimethylsulf-

oxide (DMSO). Control cells were treated with the same volume of

DMSO for 1, 2, 5 and 7 days of incubation. For long-term culture cells

were treated with single dose 10, or 20

M AFB1 for one week then cells

were cultured for additional two weeks without the toxin. A single dose

of doxorubicin (Dox) (Sigma-Aldrich) was used to treat the constructs

for 1, 2, and 3 days. Dox was also dissolved in DMSO to a final con-

centration of 10, or 20

2.10. Statistical analysis

Data were analyzed using Student’s t-test (GraphPad Prism 6,

GraphPad Software, Inc., La Jolla, CA, USA). Data are represented as

mean ±SD, p-values are considered significant by *p ≤0.05; **p ≤0.01;

***p ≤0.001; **** p ≤0.0001.

3. Results

3.1. Preparation and characterization of HepaRG cell-scaffold bioink

Bioink optimization is a critical step in bioprinting. To assess the

suitability of the hydrogel composed of alginate-gelatin-Matrigel (2 %, 3

%, and 20 % respectively) for 3D-bioprinting, differentiated HepaRG

cells were used to generate 3D constructs for toxicity testing studies in

vitro. The preparation of the printing process was performed as previ-

ously described [26]. Briefly, the hydrogel was pre-heated to 37 ◦C and

loaded into a three ml-syringe, cells, Matrigel and CaSO

were loaded

Fig. 1. Characterization of bioprinted 3D constructs. (A) Qualitative viability staining of the 3D printed mature HepaRG cells in the constructs at the indicating time

points using the Live/Dead staining; calcein-AM (live cells in green) and ethidium homodimer-1 (dead cells in red). (B) Spatial distribution of mature HepaRG cells in

3D constructs using alginate/gelatin/Matrigel as a bioink at the following time points: 1, 7, 14, and 21 days post-printing visualized by nuclear Hoechst staining

(blue). Graphs are representative images of three independent experiments. (C) Metabolic activity of mature HepaRG cells in the 3D constructs was determined by

tetrazolium hydroxide salt (XTT) assays at the indicated time points post-printing. Absorbance was carried out at 450 nm. (D) Albumin secretion of mature HepaRG

3D constructs quantified using enzyme-linked immunosorbent assay (ELISA) analysis at the indicated time points. Results are shown as mean ±SD of three inde-

pendent experiments. * p ≤0.05; ** p ≤0.01; *** p ≤0.001; **** p ≤0.0001.

K. Schmidt et al.

Toxicology Reports 7 (2020) 1578–1587

1581

into another syringe. Using a Luer-lock adapter, both syringes were

connected, and the hydrogels mixed thoroughly. They were then loaded

onto a printing cartridge and printed using the INKREDIBLE +bio-

printer (Cellink). The optimized hydrogel consisting of

alginate-gelatin-Matrigel enabled the printing procedure, and protected

the integrity of the 3D constructs, cell viability, and metabolic activity of

the cells.

The viability of the printed cells was confirmed by live/dead staining

(Fig. 1A). The 3D constructs were qualitatively analyzed using calcein-

AM (green) to stain living cells and ethidium homodimer-1 (red) to

stain dead cells after 1, 7, 14, and 21 days of culture, followed by

fluorescence microscopy. No obvious increase in the number of dead,

ethidium homodimer-1-positive cells, was detected at any of the time

points. Furthermore, the number of living, calcein-AM-positive cells,

was relatively stable during the entire time course (Fig. 1A). Further, we

analyzed the spatial distribution of the differentiated HepaRG cells

within the 3D constructs using fluorescence microscopy. First, the nuclei

were stained with Hoechst DNA dye (blue) and the distribution of the

cells was visualized at the indicated time points (Fig. 1B). Cells were

homogeneously distributed in the printed 3D constructs without sig-

nificant differences during the time course of the experiment (Data not

shown). Taken together, we conclude that the 3D-printed liver con-

structs are viable, and the cells are distributed evenly throughout the

hydrogel.

Using XTT assays, we quantified the metabolic activity of the Hep-

aRG cells in the 3D printed constructs (Fig. 1C). Consistent with the cell

viability obtained using the live/dead assay, the metabolic activity in

the 3D constructs increased significantly at day 7 and 21 compared to

day 1 (Fig. 1C). Furthermore, we analyzed albumin secretion using

ELISAs to determine the metabolic activity of the hepatic cells in 3D

constructs (Fig. 1D). Albumin secretion is one of the main characteristics

of hepatocytes and its production reflects a specific-function of liver

cells. The amount of albumin secreted from 3D printed cells increased

over time, and this increase was statistically significant at days 7, 14 and

21 (Fig. 1D). Taken together, our results showed substantial increases in

cell viability and metabolic activity over time in the 3D cell culture

system.

Fig. 2. Assessment of aflatoxin B1 toxicity on liver cells grown as 2D monolayers. (A) Monolayers of mature HepaRG cells, treated with DMSO, 10

M AFB1, and 20

M AFB1 at the indicated time points were tested for cell survival using live-dead staining. Live cells fluoresce green, whereas dead cells fluoresce red. Graphs are

representative images of three independent experiments. (B) Metabolic activity of mature HepaRG cells grown in 2D upon AFB1 treatment at the indicated time

points. Activities were determined using tetrazolium hydroxide salt (XTT) assays. Absorbance was carried out at 450 nm. (C) Albumin levels of mature HepaRG cells

in 2D monolayers upon AFB1 exposure were quantified at the indicated time points using ELISA analysis. Results are shown as mean ±SD of three independent

experiments. * p ≤0.05; ** p ≤0.01; *** p ≤0.001; **** p ≤0.0001.

K. Schmidt et al.

Toxicology Reports 7 (2020) 1578–1587

1582

3.2. Assessment of aflatoxin B1 toxicity using HepaRG 2D culture

To study the sensitivity of cells cultured in 2D to toxicity stemming

from AFB1, HepaRG cells were seeded onto a12-well plate at a density of

5*10

cells per well. Cells were then treated with a single dose of AFB1

(10 and 20

M) or DMSO, which served as a control. Following treat-

ment with AFB1, cells were stained and analyzed for cell viability using

live/dead stain and fluorescence microscopy. AFB1 treatment (10

resulted in a reduction of viable cells (calcein-AM positive-cells shown in

green) after 24 and 48 h (Fig. 2A). Similar results were obtained from the

Fig. 3. Effect of aflatoxin B1 on liver cells in 3D printed constructs. (A) Cell viability of mature HepaRG in 3D constructs upon treatment with AFB1 (10

M or 20

for the indicated time periods was visualized using the live-dead staining; calcein-AM (living cells in green) and ethidium homodimer-1 (dead cells in red). (B)

Number of green-stained cells quantified with ImageJ in constructs treated with AFB1 for the indicated time points. Bars indicate the means ±SD, n ≥3 images. (C)

Metabolic activity of the mature HepaRG cells treated with DMSO or AFB1 (10

M or 20

M) inside the 3D printed alginate/gelatin/Matrigel constructs was

determined by the XTT assay at the indicated time points. Absorbance was measured at 450 nm. (D) Levels of albumin secreted from mature HepaRG cells in 3D

constructs upon DMSO or AFB1 (10

M or 20

M) treatment were quantified at the indicated time points using ELISA analysis. Results are shown as mean ±SD of

three independent experiments. * p ≤0.05; ** p ≤0.01; *** p ≤0.001; **** p ≤0.0001.

K. Schmidt et al.

Toxicology Reports 7 (2020) 1578–1587

1583

cells treated with 20

M AFB1; however, the toxic effect of AFB1 was

more pronounced after 48 h post-treatment (Fig. 2A). After five and

seven days of treatment, cell cultures showed signs of severe cytotox-

icity, as the number of viable cells detected among the AFB1-treated

cells was very low or undetectable, in contrast to the control DMSO-

treated cells that reached confluency within two days after initial

seeding (Fig. 2A).

Next, we quantified the reduction of the tetrazolium salt XTT by

dehydrogenase enzymes to determine the metabolic activity of the 2D

HepaRG cell cultures upon AFB1 treatment (Fig. 2B). Consistent with the

cell viability obtained using the Live/Dead stain, AFB1 treatment

resulted in a significant reduction of the metabolic activity of the 2D

cultures after 24 and 48 h (Fig. 2B). Further, the metabolic activity in the

2D cell cultures treated with AFB1 decreased over time and was statis-

tically no longer measurable on days five and seven of culture.

To assess the impact of AFB1 on the hepatic metabolism of HepaRG

cell culture, albumin secretion was analyzed using ELISAs. In the DMSO-

treated differentiated HepaRG cells, the amount of secreted albumin

increased over time; however, the increase was only statistically sig-

nificant at day five (Fig. 2C). In contrast, AFB1 treatment inhibited al-

bumin production strongly, without obvious differences in albumin

secretion between the different concentrations tested (Fig. 2C). Taken

together, AFB1 toxicity showed substantial increases in toxicity over

time in the 2D cell culture system. This toxicity was characterized by

reduced metabolic activity and reduction of albumin secretion over the

time course of the experiment.

3.3. Assessment of aflatoxin B1 toxicity using 3D printed HepaRG

constructs

In 3D printed HepaRG constructs, a slight increase in sensitivity to

AFB1 toxicity was apparent during acute and chronic exposure to the

toxin using Live/Dead stain (Fig. 3A). Unlike the 2D cell culture model,

AFB1 was not strongly toxic to the cells on days one and two (Fig. 3A).

Further, AFB1 treatment did not show the substantial increases in

toxicity over time that leads to the complete loss or death of the printed

cells, which is evident from the strong green signal from the living cells

and only small fraction of red, dead cells after five and seven days

(Fig. 3A). Quantitative analysis of the number of green-stained cells

showed that AFB1 treatment did not significantly decrease the number

of living cells in all conditions except for constructs treated with 20

M-

AFB1 for 5 days (Fig. 3B). Our data suggest that printed HepaRG cells

maintained in the 3D constructs and exposed to AFB1 had significantly

higher viability after seven days, confirming their resistance against the

hepatic cytotoxicity compared to 2D cell cultures.

We then sought to assess the metabolic activity of untreated and

AFB1-treated cells in these models by quantifying the reduction of the

tetrazolium salt XTT by dehydrogenase enzymes in the 3D constructs

(Fig. 3C). For DMSO-treated 3D constructs, the XTT level increased to a

minor extent over the one-week culture period. Further, AFB1 treatment

slightly, but significantly, decreased the metabolic activity of the 3D

constructs one day post-treatment (Fig. 3C). The level of metabolic ac-

tivity two days post-treatment was similar to the metabolic level on day

one post-treatment (Fig. 3C). Similarly, longer treatment (five- and

seven-days post-treatment) showed an insignificant decrease in the

metabolic activity of the 3D constructs in comparison to days one and

day two.

To assess whether the metabolic capacity of the 3D printed HepaRG

constructs was maintained upon AFB1 treatment, we quantified the al-

bumin secretion to represent metabolic activity in untreated and treated

3D constructs using ELISAs. When comparing untreated and treated

constructs, it was noticeable that AFB1 significantly inhibited the

secretion of albumin at the indicated time points (Fig. 3D). However, a

slight but significant increase of albumin secretion was observed in cells

treated with 10

M AFB1 at days five and seven in comparison to days

one and two (Fig. 3D). In fact, when comparing the ratios of albumin

concentration and metabolic activity in 2D and 3D microenvironments,

the 3D constructs were more resistant than their 2D counterparts. The

metabolic activity was found to be reduced thirteen-fold when

comparing the control DMSO-treated cells with 20

M AFB1 in 2D cul-

tures after seven days. This reduction was only five-fold in the 3D con-

structs. Similar tendencies were found in albumin secretion; sixteen-fold

reduction in 2D cultures, and only eight-fold reduction in 3D constructs

when comparing DMSO and 20

M AFB1 conditions. These results

suggest an important role for AFB1 in the induction of cytotoxic effect in

3D bioprinted HepaRG cells, which perhaps explains the reduction of

metabolic activity and albumin production of the liver cells upon

treatment with the toxin.

Next, we investigated whether prolonged survival of the 3D bio-

printed HepaRG constructs was connected to the properties of the bioink

used in the present study. We used doxorubicin (Dox) to check for

cytotoxicity in HepaRG constructs. Indeed, Dox treatment induced

strong cell death in the printed HepaRG cells three days post-treatment

as visualized by microscopy using the live/dead stain (Fig. 4A), unlike

AFB1 treatment. Additionally, metabolic activities of the printed Hep-

aRG constructs were significantly reduced compared to AFB1 treated

constructs (Fig. 4B). Thus, our data suggest that the bioink did not play a

critical role in the cytotoxic effect of the toxins used in this study.

As 3D HepaRG constructs showed a more robust hepatic phenotype

after one-week of AFB1 exposure, unlike the 2D culture, we sought to

determine whether 3D printed HepaRG cells would survive a longer

exposure to AFB1 (Fig. 5). Interestingly, cells in the printed 3D con-

structs were still viable two- and three-weeks after the initial treatment,

as visualized by fluorescence microscopy using the Live/Dead stain

(Fig. 5A). However, quantification of cells fluoresces in green revealed a

significant decrease in number of living cells after two and three weeks

of culture in vitro (Fig. 5B). In agreement with the Live/Dead stain data,

the metabolic activity of cells in 3D constructs exhibited a slight

reduction two- and three-weeks post-treatment, measured using XTT

assays (Fig. 5C). Albumin levels two weeks post-treatment were similar

to those measured one week post-treatment suggesting that cells are still

viable and metabolically active (Fig. 5D). Taken together, functional

characterization revealed that 3D-bioprinted HepaRG constructs

exhibited a more robust hepatic phenotype than 2D cultures during

prolonged AFB1 treatment and survived the cytotoxicity exerted by

AFB1. This feature will enable longer-term and repeated exposure to

allow the study of the hepatotoxicity of toxins, chemicals, and

carcinogens.

4. Discussion

In vitro hepatic cell culture models based on 3D microenvironment

have been shown to better reflect the in vivo behavior of liver cells than

2D cultures [27]. Models using 3D cell cultures have been used in

clinical and basic research because they have proper polarization,

cell-cell and cell-matrix contacts, and hepatic functions [28,29]. These

advanced systems are currently being evaluated for infectious disease,

toxicity testing, and cancer research [4,5]. Despite its great potential,

bioprinting is still very challenging to use for toxicity testing in a 3D

microenvironment post-printing. Thus, in this study we employed 3D

bioprinting technology to study the feasibility and reproducibility of

using 3D HepaRG constructs for toxicity testing in vitro. Bioprinted

constructs were generated using the INKREDIBLE+3D-printer. The

generated 3D constructs were highly viable, exhibited prolonged sur-

vival, reduced cell death, and prolonged secretion of albumin over the

time course of culture in vitro compared to 2D cell cultures. Upon AFB1

treatment, 2D constructs reduced metabolic activity and increased cell

death, which indicates a loss of hepatocyte functionality. However, 3D

constructs were still viable and metabolically active but AFB1 treatment

showed substantial increase in toxicity over time indicating that partial

loss of hepatocytic activity. In contrast, AFB1 treatment resulted in

complete destruction of the 2D monolayer within five days

K. Schmidt et al.

Toxicology Reports 7 (2020) 1578–1587

1584

Fig. 4. Effect of doxorubicin on liver cells in 3D

printed constructs. (A) Cell viability of mature

HepaRG in 3D constructs upon treatment with AFB1

or Dox (10

M or 20

M) for the indicated time

periods was visualized using the live-dead staining;

calcein-AM (living cells in green) and ethidium

homodimer-1 (dead cells in red). (B) Metabolic ac-

tivity of the mature HepaRG cells treated with

DMSO, AFB1, or Dox inside the 3D printed con-

structs was determined by the XTT assay at the

indicated time points. Absorbance was measured at

450 nm. Results are shown as mean ±SD of three

independent experiments. * p ≤0.05; ** p ≤0.01;

*** p ≤0.001; **** p ≤0.0001.

K. Schmidt et al.

Toxicology Reports 7 (2020) 1578–1587

1585

post-treatment. These 3D bioprinted liver constructs may thus serve for

prolonged toxicity testing in vitro.

Several models including the traditional 2D cultures, 3D organoids,

and spheroids have been employed using primary human hepatocytes,

and the HepG2 and HepaRG cell lines for cytotoxicity testing in vitro [16,

29,30]. It has been shown that 3D models including organoids and

spheroids exhibit increased viability and functional stability compared

to 2D monolayers, highlighting their utility for chronic toxicity studies in

vitro [31]. Human primary hepatocytes are considered the gold standard

for testing toxicity in vitro [32]. Yet primary spheroids or 3D models

have been not considered for toxicity testing, in part, due to the diffi-

culties in handling. In the present study, we used human HepaRG he-

patocytes due to their metabolic competence to evaluate the

appropriateness of 3D constructs for toxicity testing. Our 3D constructs

exhibited long-term stability and cell viability, which is suitable for

chronic toxicity testing in vitro similar to the 3D organoids and spheroids

[33]. It has been shown that AFB1 treatment induces cell toxicity and

loss of metabolic activity of HepaRG spheroid model ([34], [35]).

Further, it has been reported that HepaRG spheroids were more sensitive

to AFB1 exposure than the 2D culture [45]. This increased in toxicity

was suggested to be correlated with different gene expression profile for

cells cultured in 3D microenvironment. In agreement with our results, it

has been also demonstrated that the metabolic activity of HepaRG 2D

cell culture is highly reduced upon treatment with AFB1 in vitro [36].

Not only, AFB1 treatment diminished the metabolic activity of liver cells

but also reduced albumin production in our 3D bioprinted cells as well

as signs of stress and enhanced toxicity due to toxin treatment. However,

no previous reports have investigated the direct of effect of AFB1 on

albumin production in vitro using HepaRG cells and this needs further

investigation. In conclusion, 3D models of liver cells result in an overall

improvement of cell viability and are functionality suitable for chronic

toxicity testing.

Animal models have been evaluated to acute and chronic doses of

AFB1 toxicity in vivo [37,38]. Although, mice display low sensitivity to

liver toxicity of AFB1 [39], a recent study has shown that a single dose of

AFB1 can induce hepatic damage and inflammation in mice model.

Moreover, the hepatic lipid droplets have been suggested to play crucial

roles in the trapping, and detoxifying of AFB1 [37]. In contrast to mice

models, rats are among the animals that are sensitive to AFB1that can

induce liver damage [40]. Recently published data demonstrated that

AFB1 induced hepatic damage through the alteration of gene expression

of lipids in rats [38]. Interspecies differences in the susceptibility to

AFB1-induced liver toxicity restrict the use of animal models for human

related studies, therefore, 2D and 3D hepatic models are considered

more reliable models for studies of AFB1 toxicity in vitro.

We used alginate/gelatin/Matrigel based-bioink which produced a

Fig. 5. Long-term culture of mature HepaRG cells in 3D printed following aflatoxin B1 treatment. (A) 3D constructs treated with AFB1 (10

M or 20

M) for one

week then constructs were cultured for additional 2 weeks with the toxin showed reduction in cell viability as visualized using the live-dead staining; calcein-AM

(living cells in green) and ethidium homodimer-1 (dead cells in red). (B) Number of green-stained cells quantified with ImageJ revealed a significant decrease in

number of living cells after two and three weeks of culture. Bars indicate the means ±SD, n ≥3 images. (C) Metabolic activity of the mature HepaRG cells treated

with DMSO or AFB1 (10

M or 20

M) inside the 3D printed alginate/gelatin/Matrigel constructs was determined by the XTT assay at the indicated time points.

Absorbance was measured at 450 nm. (D) Albumin secretion from mature HepaRG cells cultured in the 3D constructs upon treatment with DMSO or AFB1 (10

M or

M) was quantified at the indicated time points using ELISA analysis. Results are shown as mean ±SD of three independent experiments. * p ≤0.05; ** p ≤0.01;

*** p ≤0.001; **** p ≤0.0001.

K. Schmidt et al.

Toxicology Reports 7 (2020) 1578–1587

1586

stable matrix with a high porosity that sustained cell viability and

metabolic activity of bioprinted cells, as previously described [26].

Blending the three components increased the printability, viability, and

retained the shapes of the bioprinted constructs in the cultured medium

post-printing. In our previous study, we demonstrated that the algina-

te/gelatin/Matrigel bioink has increased porosity using scanning elec-

tron microscopy. Thus, we chose to use the same blend composed of 2 %

alginate, 3 % gelatin, and 20 % Matrigel for our bioink and printed the

same cell number in a grid-shape constructs. Remarkably, the observed

longer viability of 3D HepaRG cultures upon AFB1 treatment was not

linked to the bioink composition, as doxorubicin induced strong toxicity

against the 3D cultures, suggesting the suitability of this bioink in testing

for toxicity in vitro.

In conclusion, our data presented here indicate that 3D bioprinting

using HepaRG cells extends cellular function in vitro, relative to 2D cell

cultures. Our data corroborates previous findings, which demonstrated

that liver cells can survive for longer periods in a 3D microenvironment

post-printing [20], in which cytotoxicity in the 3D constructs was

continuously monitored over a long period of time; up to 3 weeks.

Although it is still not clear how HepaRG cells in 3D bioprinted con-

structs can sustain cell viability upon AFB1 treatment for so long, we

believe that our work highlights the crucial role of cell-cell and

cell-matrix contact for epithelial cell function and viability, which sus-

tain the epithelial phenotype and provide a 3D microenvironment that

enhances cell resistance to toxicity exerted by toxins in vitro such as

AFB1 [41]. Therefore, for cytotoxicity testing, responses may be more

adequately simulated in liver 3D bioprinted models.

Author statement

K.S. performed the experiments. V.R. helped with the conduction of

the albumin ELISA. J.B. contributed to the analysis of the data. M.A.A.Z.

conceived the project, performed the experiments, analyzed the data,

and wrote the manuscript. J.K. supervised the project and wrote and

edited the manuscript.

Declaration of Competing Interest

The authors report no declarations of interest.

Acknowledgements

We are thankful to Bernd Krostitz for the great technical support and

to Erik Wade for careful proofreading and helpful comments. Financial

support by the Bundesinstitut für Risikobewertung, Germany grant

number 1328-568, is gratefully acknowledged. We acknowledge sup-

port by the Open Access Publication Fund of Technische Universit¨

Berlin (TU Berlin), Germany.

References

[1] N. Charbe, P.A. McCarron, M.M. Tambuwala, Three-dimensional bio-printing: a

new frontier in oncology research, World J. Clin. Oncol. (2017), https://doi.org/

10.5306/wjco.v8.i1.21.

[2] D.-J. Cheon, S. Orsulic, Mouse models of cancer, Annu. Rev. Pathol. Mech. Dis.

(2011), https://doi.org/10.1146/annurev.pathol.3.121806.154244.

[3] D.G. Hackam, D.A. Redelmeier, Translation of research evidence from animals to

humans [9], J. Am. Med. Assoc. (2006), https://doi.org/10.1001/

jama.296.14.1731.

[4] A. Faulkner-Jones, C. Fyfe, D.J. Cornelissen, J. Gardner, J. King, A. Courtney,

W. Shu, Bioprinting of human pluripotent stem cells and their directed

differentiation into hepatocyte-like cells for the generation of mini-livers in 3D,

Biofabrication (2015), https://doi.org/10.1088/1758-5090/7/4/044102.

[5] A. Roth, T. Singer, The application of 3D cell models to support drug safety

assessment: opportunities & challenges, Adv. Drug Deliv. Rev. (2014), https://doi.

org/10.1016/j.addr.2013.12.005.

[6] J.J.M. Landry, P.T. Pyl, T. Rausch, T. Zichner, M.M. Tekkedil, A.M. Stütz, A. Jauch,

R.S. Aiyar, G. Pau, N. Delhomme, J. Gagneur, J.O. Korbel, W. Huber, L.

M. Steinmetz, The genomic and transcriptomic landscape of a hela cell line, G3

Genes| Genomes| Genet. (2013), https://doi.org/10.1534/g3.113.005777.

[7] M. Mak, R.D. Kamm, M.H. Zaman, Impact of dimensionality and network

disruption on microrheology of cancer cells in 3D environments, PLoS Comput.

Biol. (2014), https://doi.org/10.1371/journal.pcbi.1003959.

[8] M. Weinhart, A.C. Hocke, S. Hippenstiel, J. Kurreck, S. Hedtrich, 3D organ models -

revolution in pharmacological research? Pharmacol. Res. 139 (2019) 446–451.

[9] E.L. LeCluyse, R.P. Witek, M.E. Andersen, M.J. Powers, Organotypic liver culture

models: meeting current challenges in toxicity testing, Crit. Rev. Toxicol. (2012),

https://doi.org/10.3109/10408444.2012.682115.

[10] T. Liu, C. Delavaux, Y.S. Zhang, 3D bioprinting for oncology applications, J. 3d

Print. Med. (2019), https://doi.org/10.2217/3dp-2019-0004.

[11] I. Matai, G. Kaur, A. Seyedsalehi, A. McClinton, C.T. Laurencin, Progress in 3D

bioprinting technology for tissue/organ regenerative engineering, Biomaterials

(2020), https://doi.org/10.1016/j.biomaterials.2019.119536.

[12] A. Lee, A.R. Hudson, D.J. Shiwarski, J.W. Tashman, T.J. Hinton, S. Yerneni, J.

M. Bliley, P.G. Campbell, A.W. Feinberg, 3D bioprinting of collagen to rebuild

components of the human heart, Science (2019), https://doi.org/10.1126/science.

aav9051.

[13] X. Wang, Q. Ao, X. Tian, J. Fan, H. Tong, W. Hou, S. Bai, Gelatin-based hydrogels

for organ 3D bioprinting, Polymers (2017), https://doi.org/10.3390/

polym9090401.

[14] C.M. Hwang, S. Sant, M. Masaeli, N.N. Kachouie, B. Zamanian, S.H. Lee,

A. Khademhosseini, Fabrication of three-dimensional porous cell-laden hydrogel

for tissue engineering, Biofabrication (2010), https://doi.org/10.1088/1758-

5082/2/3/035003.

[15] D. Fan, Y. Li, X. Wang, T. Zhu, Q. Wang, H. Cai, W. Li, Y. Tian, Z. Liu, Progressive

3D printing technology and its application in medical materials, Front. Pharmacol.

(2020), https://doi.org/10.3389/fphar.2020.00122.

[16] V.M. Lauschke, D.F.G. Hendriks, C.C. Bell, T.B. Andersson, M. Ingelman-Sundberg,

Novel 3D culture systems for studies of human liver function and assessments of

the hepatotoxicity of drugs and drug candidates, Chem. Res. Toxicol. (2016),

https://doi.org/10.1021/acs.chemrestox.6b00150.

[17] H. Jeon, K. Kang, S.A. Park, W.D. Kim, S.S. Paik, S.H. Lee, J. Jeong, D. Choi,

Generation of multilayered 3D structures of HepG2 cells using a bio-printing

technique, Gut Liver (2017), https://doi.org/10.5009/gnl16010.

[18] S. Klein, D. Mueller, V. Schevchenko, F. Noor, Long-term maintenance of HepaRG

cells in serum-free conditions and application in a repeated dose study, J. Appl.

Toxicol. (2014), https://doi.org/10.1002/jat.2929.

[19] T. Grix, A. Ruppelt, A. Thomas, A.K. Amler, B.P. Noichl, R. Lauster, L. Kloke,

Bioprinting perfusion-enabled liver equivalents for advanced organ-on-a-chip

applications, Genes (2018), https://doi.org/10.3390/genes9040176.

[20] T. Hiller, J. Berg, L. Elomaa, V. R¨

ohrs, I. Ullah, K. Schaar, A.C. Dietrich, M.A. Al-

Zeer, A. Kurtz, A.C. Hocke, S. Hippenstiel, H. Fechner, M. Weinhart, J. Kurreck,

Generation of a 3D liver model comprising human extracellular matrix in an

alginate/gelatin-based bioink by extrusion bioprinting for infection and

transduction studies, Int. J. Mol. Sci. (2018), https://doi.org/10.3390/

ijms19103129.

[21] M.S. Mohsenzadeh, N. Hedayati, B. Riahi-Zanjani, G. Karimi, Immunosuppression

following dietary aflatoxin B1 exposure: a review of the existing evidence, Toxin

Rev. (2016), https://doi.org/10.1080/15569543.2016.1209523.

[22] M.G.Y. Theumer, Khoury Henneb, Tadrist Snini, Puel I. Canlet, Audebert Oswald,

Genotoxicity of aflatoxins and their precursors in human cells, Toxicol. Lett.

(2018), https://doi.org/10.1016/j.toxlet.2018.02.007.

[23] (IARC), I.A. for R. on C, Monographs on the evaluation of carcinogenic risks to

humans. The Monographs, 2002.

[24] E.P. Gallagher, K.L. Kunze, P.L. Stapleton, D.L. Eaton, The kinetics of aflatoxin B1

oxidation by human cDNA-expressed and human liver microsomal cytochromes

P450 1A2 and 3A4, Toxicol. Appl. Pharmacol. (1996), https://doi.org/10.1006/

taap.1996.0326.

[25] D. Bovard, S. Sandoz, K. Luettich, S. Frentzel, A. Iskandar, D. Marescotti,

K. Trivedi, K. Guedj, Q. Dutertre, M. Peitsch, J. Hoeng, A lung/liver-on-a-chip

platform for acute and chronic toxicity studies, Lab Chip (2018), https://doi.org/

10.1039/c8lc01029c.

[26] J. Berg, T. Hiller, M.S. Kissner, T.H. Qazi, G.N. Duda, A.C. Hocke, S. Hippenstiel,

L. Elomaa, M. Weinhart, C. Fahrenson, J. Kurreck, Optimization of cell-laden

bioinks for 3D bioprinting and efficient infection with influenza A virus, Sci. Rep.

(2018), https://doi.org/10.1038/s41598-018-31880-x.

[27] F. Pampaloni, E.G. Reynaud, E.H.K. Stelzer, The third dimension bridges the gap

between cell culture and live tissue, Nat. Rev. Mol. Cell Biol. (2007), https://doi.

org/10.1038/nrm2236.

[28] X. Ma, X. Qu, W. Zhu, Y.S. Li, S. Yuan, H. Zhang, J. Liu, P. Wang, C.S.E. Lai,

F. Zanella, G.S. Feng, F. Sheikh, S. Chien, S. Chen, Deterministically patterned

biomimetic human iPSC-derived hepatic model via rapid 3D bioprinting, Proc.

Natl. Acad. Sci. U. S. A. (2016), https://doi.org/10.1073/pnas.1524510113.

[29] W. Zhu, X. Ma, M. Gou, D. Mei, K. Zhang, S. Chen, 3D printing of functional

biomaterials for tissue engineering, Curr. Opin. Biotechnol. (2016), https://doi.

org/10.1016/j.copbio.2016.03.014.

[30] R. Joss´

e, C. Aninat, D. Glaise, J. Dumont, V. Fessard, F. Morel, J.M. Poul,

C. Guguen-Guillouzo, A. Guillouzo, Long-term functional stability of human

HepaRG hepatocytes and use for chronic toxicity and genotoxicity studies, Drug

Metab. Dispos. (2008), https://doi.org/10.1124/dmd.107.019901.

[31] C.C. Bell, A.C.A. Dankers, V.M. Lauschke, R. Sison-Young, R. Jenkins, C. Rowe, C.

E. Goldring, K. Park, S.L. Regan, T. Walker, C. Schofield, A. Baze, A.J. Foster, D.

P. Williams, A.W.M. van de Ven, F. Jacobs, J. van Houdt, T. L¨

ahteenm¨

aki,

J. Snoeys, S. Juhila, L. Richert, M. Ingelman-Sundberg, Comparison of hepatic 2D

sandwich cultures and 3d spheroids for long-term toxicity applications: a

multicenter study, Toxicol. Sci. (2018), https://doi.org/10.1093/toxsci/kfx289.

K. Schmidt et al.

Toxicology Reports 7 (2020) 1578–1587

1587

[32] M.J. G´

omez-Lech´

on, L. Tolosa, I. Conde, M.T. Donato, Competency of different cell

models to predict human hepatotoxic drugs, Expert Opin. Drug Metab. Toxicol.

(2014), https://doi.org/10.1517/17425255.2014.967680.

[33] J. Li, R.S. Settivari, M.L. LeBaron, M.S. Marty, Functional comparison of HepaRG

cells and primary human hepatocytes in sandwich and spheroid culture as

repeated-exposure models for hepatotoxicity, Appl. In Vitro Toxicol. (2019),

https://doi.org/10.1089/aivt.2019.0008.

[34] S.C. Ramaiahgari, S. Waidyanatha, D. Dixon, M.J. DeVito, R.S. Paules, S. Ferguson,

Three-dimensional (3D) HepaRG spheroid model with physiologically relevant

xenobiotic metabolism competence and hepatocyte functionality for liver toxicity

screening, Toxicol. Sci. (2017), https://doi.org/10.1093/toxsci/kfx122.

[35] C. Bell, V. Lauschke, S. Vorrink, H. Palmgren, R. Duffin, T. Andersson,

M. Ingelman-Sundberg, Transcriptional, functional, and mechanistic comparisons

of stem cell–derived hepatocytes, HepaRG cells, and three-dimensional human

hepatocyte spheroids as predictive in vitro systems for drug-induced liver injury,

Drug Metab. Dispos. (2017), https://doi.org/10.1124/dmd.116.074369.

[36] C. Aninat, A. Piton, D. Glaise, D. Le Charpentier, S. Langou¨

et, F. Morel, C. Guguen-

Guillouzo, A. Guillouzo, Expression of cytochromes P450, conjugating enzymes

and nuclear receptors in human hepatoma HepaRG cells, Drug Metab. Dispos.

(2006), https://doi.org/10.1124/dmd.105.006759.

[37] N. Hammoudeh, C. Soukkarieh, D.J. Murphy, A. Hanano, Involvement of hepatic

lipid droplets and their associated proteins in the detoxification of aflatoxin B1 in

aflatoxin-resistance BALB/C mouse, Toxicol. Rep. 7 (2020) 795–804, https://doi.

org/10.1016/j.toxrep.2020.06.005.

[38] O.A. Rotimi, S.O. Rotimi, C.D. Duru, O.J. Ebebeinwe, A.O. Abiodun, B.O. Oyeniyi,

A.F. Faduyile, Acute aflatoxin B1 – induced hepatotoxicity alters gene expression

and disrupts lipid and lipoprotein metabolism in rats, Toxicol. Rep. 4 (2017)

408–414, https://doi.org/10.1016/j.toxrep.2017.07.006.

[39] Z. Ilic, D. Crawford, D. Vakharia, P.A. Egner, S. Sell, Glutathione-S-transferase A3

knockout mice are sensitive to acute cytotoxic and genotoxic effects of aflatoxin

B1, Toxicol. Appl. Pharmacol. 242 (3) (2010) 241–246, pmid:19850059.

[40] J.M. Cullen, B.H. Ruebner, L.S. Hsieh, D.M. Hyde, D.P. Hsieh, Carcinogenicity of

dietary aflatoxin M1 in male Fischer rats compared to aflatoxin B1, Cancer Res. 47

(7) (1987) 1913–1917, pmid:3102052.

[41] L.L. Li, S.L. Qiao, W.J. Liu, Y. Ma, D. Wan, J. Pan, H. Wang, Intracellular

construction of topology-controlled polypeptide nanostructures with diverse

biological functions, Nat. Commun. (2017), https://doi.org/10.1038/s41467-017-

01296-8.

[42] Wei Sun, Binil Starly, C. Andrew Daly, A. Jason Burdick, Jürgen Groll, Gregor

Skeldon, Wenmiao Shu, Yasuyuki Sakai, Marie Shinohara, Masaki Nishikawa, The

bioprinting roadmap, Biofabrication. . 12 (2) (2020), https://doi.org/10.1088/

1758-5090/ab5158.

[43] V.Y. Soldatow, E.L. Lecluyse, L.G. Griffith, I. Rusyn, In vitro models for liver

toxicity testing, Toxicology Research 2 (1) (2012) 23–39, https://doi.org/10.1039/

C2TX20051A.

[44] Y. Yokoyama, Y. Sasaki, N. Terasaki, T. Kawataki, K. Takekawa, Y. Iwase,

T. Shimizu, S. Sanoh, S. Ohta, Comparison of Drug Metabolism and Its Related

Hepatotoxic Effects in HepaRG, Cryopreserved Human Hepatocytes, and HepG2

Cell Cultures, Biological and Pharmaceutical Bulletin 41 (5) (2018) 722–732,

https://doi.org/10.1248/bpb.b17-00913.

[45] D. Mueller, L. Kr¨

amer, E. Hoffmann, S. Klein, F. Noor, Toxicology in Vitro : an

International Journal Published in Association with BIBRA, 3 D organotypic

HepaRG cultures as in vitro model for acute and repeated dose toxicity studies. 28

(1) (2013) 104–112, https://doi.org/10.1016/j.tiv.2013.06.024.

K. Schmidt et al.