Functional Evaluation of 3D Liver Models Labeled with Polysaccharide Functionalized Magnetic Nanoparticles

Establishing a rapid in vitro evaluation system for drug screening is essential for the development of new drugs. To reproduce tissues/organs with functions closer to living organisms, in vitro three-dimensional (3D) culture evaluation using microfabrication technology has been reported in recent years. Culture on patterned substrates with controlled hydrophilic and hydrophobic regions (Cell-ableTM) can create 3D liver models (miniature livers) with liver-specific Disse luminal structures and functions. MRI contrast agents are widely used as safe and minimally invasive diagnostic methods. We focused on anionic polysaccharide magnetic iron oxide nanoparticles (Resovist®) and synthesized the four types of nanoparticle derivatives with different properties. Cationic nanoparticles (TMADM) can be used to label target cells in a short time and have been successfully visualized in vivo. In this study, we examined the morphology of various nanoparticles. The morphology of various nanoparticles showed relatively smooth-edged spherical shapes. As 3D liver models, we prepared primary hepatocyte–endothelial cell heterospheroids. The toxicity, CYP3A, and albumin secretory capacity were evaluated in the heterospheroids labeled with various nanoparticles. As the culture period progressed, the heterospheroids labeled with anionic and cationic nanoparticles showed lower liver function than non-labeled heterospheroids. In the future, there is a need to improve the method of creation of artificial 3D liver or to design a low-invasive MRI contrast agent to label the artificial 3D liver.


Introduction
Establishing a rapid in vitro evaluation system for drug screening is essential for the development of new drugs [1][2][3]. The liver is an essential organ in vivo, and many attempts have been made to artificially mimic the liver in vitro [4][5][6][7][8]. These cell culture techniques are used in many fields including drug discovery, medicine, and cosmetics. For example, two-dimensional (2D) culture via a substrate or extracellular matrix (ECM) is simple and effective for studying cell interactions with drugs [9,10] and is used for screening in drug development. However, they exhibit behavior that differs from that of cell proliferation and morphology. To reproduce tissues and organs with functions that are similar to those of living organisms, in vitro evaluation using three-dimensional (3D) culture has been conducted in recent years [4][5][6][7][8]11,12]. Specifically, it has been reported that 3D culture is superior to 2D culture in terms of tissue/organ structure, cell differentiation ability, proliferation, drug toxicity, and metabolism, with results closer to those obtained in vivo [5,[13][14][15].
As a safe and minimally invasive diagnostic method, MRI contrast agents are widely used to contrast the condition of the blood vessels, various organs, and lesion areas. In particular, iron-and gadolinium-based contrast agents are expected to produce a high contrast effect through interactions with the surrounding protons. We focused on an MRI contrast agent already approved as a drug, polysaccharide magnetic nanoparticle composite (Resovist ® ; ATDM), and synthesized several types of magnetic nanoparticles with different properties (cationic and anionic groups, particle size, zeta potential, etc.) [27][28][29][30][31]. In addition, we investigated the stability of the synthesized nanoparticles in the culture medium and their cellular uptake. The results showed that cationic polysaccharide magnetic particles (TMADM) effectively take up various cells [27][28][29][30][31]. In particular, TMADM can be labeled on pancreatic islets in a short time and visualized after implantation in mice [29]. Furthermore, transmission electron microscopy of the internal structure of human hepatocellular carcinoma cells (HepG2) monospheroids labeled with TMADM showed that TMADM accumulated in the central and inner cell lysosomes of spheroids [30].
We then examined the morphology of the four synthesized magnetic nanoparticles. For the 3D liver models, we prepared primary hepatocyte-endothelial cell heterospheroids, that is, miniature livers with liver-specific functions. The toxicity, CYP3A, and albumin secretory capacity were evaluated in the heterospheroids labeled with several polysaccharide magnetic iron oxide nanoparticles.

Size and Zeta Potential Measurement
The average hydrodynamic diameter and the polydispersity index of nanoparticles were determined by dynamic light scattering (DLS) using an ELSZneo (Otsuka Electronics Co., Ltd., Osaka, Japan) fitted with a 638 nm laser beam at a fixed angle of 165°. Measurements were taken at 25 °C, with 0.89 × 10 −2 Pa.s viscosity and a refractive index of 1.33. The nanoparticles were concentrated at 0.025 mg/mL in 2.5 mM phosphate buffer solution (pH 7.4), to enable measurements (performed in triplicate), in order to assure a convenient scatter intensity on the detectors. Three measurements were carried out for each sample and mean values with standard deviations were calculated. The zeta potential measurements were carried out using a ZetaSizer nano ZS (Malvern Instruments, Malvern, UK). Each sample was properly diluted with 0.025 M phosphate buffer and maintain the temperature at 25 °C.

SEM and TEM Micrographs of Polysaccharide Magnetic Iron Oxide Nanoparticles
The morphology of polysaccharide magnetic iron oxide nanoparticles was observed by a field emission scanning electron microscope (FE-SEM) (Hitachi High-Technologies Co., SU9000) (Tokyo, Japan). Ultrasonically dispersed nanoparticles onto a collodion

Size and Zeta Potential Measurement
The average hydrodynamic diameter and the polydispersity index of nanoparticles were determined by dynamic light scattering (DLS) using an ELSZneo (Otsuka Electronics Co., Ltd., Osaka, Japan) fitted with a 638 nm laser beam at a fixed angle of 165 • . Measurements were taken at 25 • C, with 0.89 × 10 −2 Pa.s viscosity and a refractive index of 1.33. The nanoparticles were concentrated at 0.025 mg/mL in 2.5 mM phosphate buffer solution (pH 7.4), to enable measurements (performed in triplicate), in order to assure a convenient scatter intensity on the detectors. Three measurements were carried out for each sample and mean values with standard deviations were calculated. The zeta potential measurements were carried out using a ZetaSizer nano ZS (Malvern Instruments, Malvern, UK). Each sample was properly diluted with 0.025 M phosphate buffer and maintain the temperature at 25 • C.

SEM and TEM Micrographs of Polysaccharide Magnetic Iron Oxide Nanoparticles
The morphology of polysaccharide magnetic iron oxide nanoparticles was observed by a field emission scanning electron microscope (FE-SEM) (Hitachi High-Technologies Co., SU9000) (Tokyo, Japan). Ultrasonically dispersed nanoparticles onto a collodion membrane were observed at an accelerating voltage of 15 kV. The structure and morphology of polysaccharide magnetic iron oxide nanoparticles were observed using a transmission electron microscopy (TEM) (H-7650 type zero A, Hitachi) (Tokyo, Japan). Ultrasonically dispersed nanoparticles onto a collodion membrane were observed at an accelerating voltage of 120 kV.

Cells and Animals
Bovine carotid artery normal endothelial cells (HH) (JCRB0099) were obtained from the Japanese Collection of Research Bioresources Cell Bank (Osaka, Japan). Male Sprague-Dawley rats (6-7 weeks old, specific pathogen-free) weighing approximately 200-250 g were purchased from SLC Japan. The rats were housed under specific pathogen-free conditions with a 12-h light/dark cycle and had free access to food and water. Rat studies were approved by the review committee of Nagoya University Graduate School of Medicine. Rat hepatic parenchymal cells (primary hepatocytes) were prepared from male Sprague-Dawley rats by the collagenase perfusion method [32] Cell viability was >85%, as determined by the trypan blue dye exclusion test using trypan blue stain (Gibco) at a final concentration of 0.2%.

3D Spheroids Culture of Hepatocytes and Endothelial Cells
The 12-well cell culture plates for 3D spheroids culture (Cell-able TM ) were supplied by Sumitomo Bakelite (Tokyo, Japan). HH cells (3 × 10 5 ) were inoculated onto the cell array and cultured in 1 mL of DMEM containing 10% FBS and antibiotics (100 U/mL penicillin and 100 U/mL streptomycin). Rat hepatic parenchymal cells (primary hepatocyte) (4 × 10 5 /well) were inoculated into the HH-precultured cell array to form heterospheroids. The heterospheroids were cultured in Williams Medium E containing 10% FBS, 1 µmol/L insulin, 1 µmol/L dexamethasone, and antibiotics. The heterospheroids were incubated for two days at 37 • C in an incubator with a humidified 5% CO 2 atmosphere. The heterospheroids were then incubated with polysaccharide magnetic iron oxide nanoparticles (0, 0.1, 0.5, 1, 5, 15, 30, and 60 µg Fe/mL) for 24 h at 37 • C. Hepatic function and in vitro cytotoxicity started 3 d after inoculation of the heterospheroids.

Functional Examination of Hepatocyte-Endothelial Cell Heterospheroids Labeled with Polysaccharide Magnetic Iron Oxide Nanoparticles
Hepatocyte-endothelial cell heterospheroids labeled with polysaccharide magnetic iron oxide nanoparticles were also assayed calorimetrically for metabolic activity using formazan formation (Cell Counting Kit-8 (WST-8); Dojindo Molecular Technologies Inc., Kumamoto, Japan). Rat primary hepatocytes on the cell array were assayed for CYP3A activity by testosterone 6-β-hydroxylation, as described previously [33]. Cytochrome P450 subfamily CYP3A activity was estimated by the rate of testosterone 6-β-hydroxylation. The heterospheroids were incubated with 150 µmol/L testosterone (T1500; Sigma, MO, USA) for 4 h, and the formation of 6-β-hydroxytestosterone was determined by high performance liquid chromatography with a reversed-phase analytical column (TSKgel Super-ODS18197, TOSOH, Tokyo, Japan). Albumin production was assessed by measuring its accumulation in the culture medium for 24 h using a rat albumin ELISA (Enzyme-Linked Immunosorbent Assay) quantification kit according to the manufacturer's instructions (E111-125; Bethyl Laboratories, Inc., Montgomery, TX, USA).

Statistical Data Analysis
Data are presented as mean ± standard error. Each experiment was repeated at least in triplicates (n ≥ 3). Statistical significance was determined using Welch's t-test for comparisons between two groups; Statistical significance was set at p-values (p < 0.05, and p < 0.01).

Morphology, Size, and Zeta Potential of Polysaccharide Magnetic Iron Oxide Nanoparticles
We synthesized and evaluated various magnetic polysaccharide iron oxide nanoparticles (Scheme 1), as described previously [28,31]. The morphology of various nanoparticles was observed by FE-SEM and TEM ( Figure 1). The morphology of various nanoparticles showed relatively smooth-edged spherical shapes and the aggregate formation by FE-SEM micrograph ( Figure 1A). The morphology of various nanoparticles showed nearly spherical shapes and the aggregate formation by TEM micrograph ( Figure 1B). The nanoparticle size and zeta potential were measured from dynamic light scattering and Zetasizer, respectively (Table 1). TMADM and EADM were efficiently transduced into cells because of their positive charges [28]. In contrast, the negatively charged nanoparticles ATDM and CMDM were transduced to a lesser extent into the cells [28].

Morphology, Size, and Zeta Potential of Polysaccharide Magnetic Iron Oxide Nanoparticles
We synthesized and evaluated various magnetic polysaccharide iron oxide nanoparticles (Scheme 1), as described previously [28,31]. The morphology of various nanoparticles was observed by FE-SEM and TEM ( Figure 1). The morphology of various nanoparticles showed relatively smooth-edged spherical shapes and the aggregate formation by FE-SEM micrograph ( Figure 1A). The morphology of various nanoparticles showed nearly spherical shapes and the aggregate formation by TEM micrograph ( Figure 1B). The nanoparticle size and zeta potential were measured from dynamic light scattering and Zetasizer, respectively (Table 1). TMADM and EADM were efficiently transduced into cells because of their positive charges [28]. In contrast, the negatively charged nanoparticles ATDM and CMDM were transduced to a lesser extent into the cells [28].   Figure 2 shows a phase-contrast photomicrograph of hepatocyte-endothelial cell heterospheroids labeled with various nanoparticles. After creating the endothelial cell monospheroids (Figure 2a,b), the heterospheroids were prepared by seeding primary hepatocytes on the monospheroids. We have previously reported that the prepared hepatocyteendothelial cell heterospheroids have hepatocyte-specific structures and functions [5,17]. The heterospheroids were incubated with 60 μg-Fe/mL various nanoparticles (TMADM, EADM, ATDM, and CMDM) in a culture medium for 24 h at 37 °C. The morphology color of the heterospheroids labeled with various nanoparticles (Figure 2e-l) is darker compared to the control group (non-labeled heterospheroids) (Figure 2c,d). These results are consistent with those reported for HepG2 cell monospheroids [30,31].

Morphology of Hepatocyte-Endothelial Cell Heterospheroids Labeled with Polysaccharide
Magnetic Iron Oxide Nanoparticles Figure 2 shows a phase-contrast photomicrograph of hepatocyte-endothelial cell heterospheroids labeled with various nanoparticles. After creating the endothelial cell monospheroids (Figure 2a,b), the heterospheroids were prepared by seeding primary hepatocytes on the monospheroids. We have previously reported that the prepared hepatocyteendothelial cell heterospheroids have hepatocyte-specific structures and functions [5,17]. The heterospheroids were incubated with 60 µg-Fe/mL various nanoparticles (TMADM, EADM, ATDM, and CMDM) in a culture medium for 24 h at 37 • C. The morphology color of the heterospheroids labeled with various nanoparticles (Figure 2e-l) is darker compared to the control group (non-labeled heterospheroids) (Figure 2c,d). These results are consistent with those reported for HepG2 cell monospheroids [30,31]. Materials 2022, 15, x FOR PEER REVIEW 6 of 16

Metabolic Activity of Hepatocyte-Endothelial Cell Heterospheroids
The metabolic activities of hepatocyte-endothelial cell heterospheroids are shown in Figure 3. The heterospheroids assessed in vitro cytotoxicity (the WST-1 assay) of various nanoparticles ( Figure 3). As the concentration of nanoparticles increased, the heterospheroids showed a decrease in formazan, a metabolite formed by mitochondrial dehydrogenase. The heterospheroids labeled with a high amount of nanoparticles inhibited cell proliferation. The heterospheroids labeled with anionic nanoparticles (ATDM and CMDM) showed higher activity than the cultured heterospheroids labeled with cationic nanoparticles (TMADM and EADM). Relative WST-1 activity was significantly different in the control group (non-labeled heterospheroids) and the labeled groups (control group vs. various nanoparticles (concentration higher than 5 μg-Fe/mL), p < 0.05; control group vs. TMADM and EADM (concentration higher than 30 μg-Fe/mL), p < 0.01).

Metabolic Activity of Hepatocyte-Endothelial Cell Heterospheroids
The metabolic activities of hepatocyte-endothelial cell heterospheroids are shown in Figure 3. The heterospheroids assessed in vitro cytotoxicity (the WST-1 assay) of various nanoparticles ( Figure 3). As the concentration of nanoparticles increased, the heterospheroids showed a decrease in formazan, a metabolite formed by mitochondrial dehydrogenase. The heterospheroids labeled with a high amount of nanoparticles inhibited cell proliferation. The heterospheroids labeled with anionic nanoparticles (ATDM and CMDM) showed higher activity than the cultured heterospheroids labeled with cationic nanoparticles (TMADM and EADM). Relative WST-1 activity was significantly different in the control group (non-labeled heterospheroids) and the labeled groups (control group vs. various nanoparticles (concentration higher than 5 µg-Fe/mL), p < 0.05; control group vs. TMADM and EADM (concentration higher than 30 µg-Fe/mL), p < 0.01).

Albumin Synthesis of Hepatocyte-Endothelial Cell Heterospheroids
The amount of albumin secreted by hepatocyte-endothelial cell heterospheroids is shown in Figure 5 and Figure 6, respectively. Figure 5 shows the albumin secretion by the heterospheroids on day 3. As the concentration of various nanoparticles increased, albumin secretion of the heterospheroids decreased for various nanoparticles. Albumin secretion was significantly different in the control group (non-labeled heterospheroids) and the

Albumin Synthesis of Hepatocyte-Endothelial Cell Heterospheroids
The amount of albumin secreted by hepatocyte-endothelial cell heterospheroids is shown in Figures 5 and 6, respectively. Figure 5 shows the albumin secretion by the heterospheroids on day 3. As the concentration of various nanoparticles increased, albumin secretion of the heterospheroids decreased for various nanoparticles. Albumin secretion was significantly different in the control group (non-labeled heterospheroids) and the labeled groups (control group vs. TMADM and EADM (concentration higher than 15 or 30 µg-Fe/mL), p < 0.05; control group vs. ATDM and CMDM (60 µg-Fe/mL), p < 0.05; control group vs. TMADM and EADM (60 µg-Fe/mL), p < 0.01). The heterospheroids labeled with anionic nanoparticles (ATDM and CMDM) showed higher albumin secretory capacity than the cultured heterospheroids labeled with cationic nanoparticles (TMADM and EADM).

Discussion
It is critical to evaluate the effects of contrast agents and drugs, such as nanoparticles, on cells for purposes of medical and clinical research. We expect to modify the clinical MRI contrast agents (Resovist ® ) and use them as a visualization tool for cell transplantation. The most common compounds utilized in MRI for contrast enhancement are iron oxide and gadolinium [34]. Feridex ® is a superparamagnetic iron oxide (SPIO) colloid with low molecular weight dextran with a particle size of 120-180 nm. Resovist ® is carboxydextrane-coated SPIO with a particle size of 45-60 nm. Primovist ® is gadolinium ethoxybenzyl dimeglumine (Gd-EOB-DTPA) with a molecular weight of 725.71. This study focused on ATDM (Resovist ® ) and evaluated the other iron-based polysaccharide nanoparticles (TMADM, EADM, and CMDM) for labeling target cells. We synthesized and evaluated characteristic features of various nanoparticles (ATDM, TMADM, EADM, and CMDM), as described previously [28,31]. The morphology of magnetic polysaccharide nanoparticles by FE-SEM and TEM showed relatively smooth-edged spherical shapes and nearly  Figure 6 shows the albumin secretion by the heterospheroids on days 3, 4, and 5. Statistically, significant difference was observed in the control group and the labeled groups with 60 µg-Fe/mL various nanoparticles (TMADM and EADM: days 3, 4, and 5 p < 0.01; ATDM and CMDM: days 3, 4, and 5 p < 0.05). As the culture period progressed, the albumin secretion of the labeled groups decreased significantly (control group vs. TMADM and EADM: day 3 (15 or 30 µg-Fe/mL, p < 0.05) vs. days 4 and 5 (30 µg-Fe/mL, p < 0.01); control group vs. ATDM and CMDM: day 3 (60 µg-Fe/mL, p < 0.05) vs. days 4 and 5 (15 or 30 µg-Fe/mL, p < 0.05). As the culture period progressed, the heterospheroids labeled with anionic nanoparticles (ATDM and CMDM) showed higher albumin secretory capacity than the cultured heterospheroids labeled with cationic nanoparticles (TMADM and EADM).

Discussion
It is critical to evaluate the effects of contrast agents and drugs, such as nanoparticles, on cells for purposes of medical and clinical research. We expect to modify the clin-ical MRI contrast agents (Resovist ® ) and use them as a visualization tool for cell transplantation. The most common compounds utilized in MRI for contrast enhancement are iron oxide and gadolinium [34]. Feridex ® is a superparamagnetic iron oxide (SPIO) colloid with low molecular weight dextran with a particle size of 120-180 nm. Resovist ® is carboxydextrane-coated SPIO with a particle size of 45-60 nm. Primovist ® is gadolinium ethoxybenzyl dimeglumine (Gd-EOB-DTPA) with a molecular weight of 725.71. This study focused on ATDM (Resovist ® ) and evaluated the other iron-based polysaccharide nanoparticles (TMADM, EADM, and CMDM) for labeling target cells. We synthesized and evaluated characteristic features of various nanoparticles (ATDM, TMADM, EADM, and CMDM), as described previously [28,31]. The morphology of magnetic polysaccharide nanoparticles by FE-SEM and TEM showed relatively smooth-edged spherical shapes and nearly spherical shapes, respectively. Similar to previous reports, dextran-based nanoparticles exhibit aggregate formation [35,36]. The particle size of ATDM by TEM or DLS was approximately 23-30 nm and 57.9 ± 0.5 nm, respectively. TEM sizes are smaller than those obtained by DLS. Therefore, TEM determines the actual size of dried nanoparticle formulations. On the other hand, DLS measures the size of ATDM in an aqueous solution. These results are consistent with previous reports by Resovist ® , who measured the hydrodynamic diameter of Resovist ® in an aqueous solution and incorporated the surface-bound water layer in DLS [35,36].
Many researchers have reported numerous comparisons between the 2D and 3D cultures of drugs [1][2][3]5,[13][14][15]37,38]. Most importantly, the selection of 3D cultures allows for longer retention of cellular function and better test sensitivity than 2D cultures. Therefore, 3D culture techniques have been widely used in drug discovery, medicine, and cosmetics. The human body is composed of various cells, among which the cells of the liver play an essential role in metabolism, detoxification, bile production, and secretion [4][5][6][7][8][11][12][13][14][15][39][40][41]. In order to investigate the effect of nanomaterials on the liver, a 3D liver model with liver-specific functions must be created in vitro. Several studies have been reported on the effects of nanomaterials on 3D liver models [30,[42][43][44][45] [45]. Various types of nanoparticles can be evaluated for liver functions in 3D liver models [30,[42][43][44][45], and the obtained results can support in vivo [43]. In our previous reports, the heterospheroids of endothelial cells and primary hepatocytes exhibited liver-like structures, and maintained the albumin secretory capacity, CYP3A activity, and glucuronidation for an extended period [4,5]. In this study, the heterospheroids were used to investigate the effects of various nanoparticles.
The morphology of the heterospheroids labeled with various nanoparticles (Figure 2) did not differ significantly from that of the HepG2 cell monospheroids labeled with nanoparticles [30]. The heterospheroids labeled with ATDM and CMDM exhibited higher WST-1 activity than that labeled with TMADM and EADM (Figure 3). This result shows that the nanoparticles have anionic and cationic characteristics, respectively (Table 1).
In our previous reports, Yukawa H. et al. reported the viability of adipose tissuederived stem cells (ASCs) labeled with ATDM or TMADM [46]. The viability of ASCs was significantly different in non-labeled ASCs and ASCs labeled with TMADM (concentration 5-100 µg-Fe/mL, p < 0.05). On the other hand, the viability of ASCs was not significantly different in non-labeled ASCs, and ASCs labeled with TMADM (concentration 5-100 µg-Fe/mL). The proliferation rate of ASCs labeled with TMADM (concentration 10-50 µg-Fe/mL) was also maintained compared to non-labeled ASCs. Anionic iron-based polysaccharide nanoparticles (ATDM and CMDM) have been reported to exist primarily near the cell surface in 2D culture [27,30,31,46]. The cationic nanoparticles (TMADM and EADM) are reported to be mainly taken up by cells and accumulate in lysosomes [30,31,46]. The quantitative determinations of the ATDM and TMADM internalized into ASCs by measuring the amount of Fe derived from ATDM or TMADM using ICP-AES [46]. The amount of Fe in cells labeled with TMADM was about four-fold higher than that of ATDM. No significant differences in cell function were observed compared to ASCs labeled with ATDM and non-labeled ASCs.
In previous reports, Kermanizadeh A. et al. evaluated the toxicity and hepatic function of human liver cell heterospheroids (hepatocyte and non-parenchymal cells) by exposure to various nanomaterials (Ag, ZnO, MWCNT, and a positively charged TiO 2 ) [42]. The repeated exposure to nanomaterials for the heterospheroids was more toxic than a single exposure. The cytotoxicity and viability of the heterospheroids were markedly reduced by exposure to nanomaterials. On the other hand, there was no significant difference in albumin secretion of heterospheroids by nanomaterial exposure. Albumin is a protein produced by the liver and is a standard indicator of liver function. These differences in albumin secretion of the heterospheroids may depend on the type of nanomaterials and various cells. Tee J.K. et al. created the 3D liver models (heterospheroids) using LO2 cells and sinusoidal endothelial cells [44]. The structure of these heterospheroids (Tee J.K. et al. [44]) is similar to that of our created heterospheroids (Enosawa S. and Miyamoto Y. et al. [5]). Because the hepatocyte is the core of the heterospheroids, and the endothelial cells are located on the outside heterospheroids. As the concentration of TiO 2 nanoparticles increased, the sinusoidal endothelial cells in the heterospheroids decreased compared to the controls (heterospheroids without TiO 2 nanoparticles) [44]. TiO 2 nanoparticles weakened the adhesion between the sinusoidal endothelial cell layers and the core of the hepatocytes in the heterospheroids and detached the sinusoidal endothelial cells in the heterospheroids [44]. In our previous reports, Enosawa S. and Miyamoto Y. et al. reported that the albumin secretion in the hepatocyte-endothelial cell heterospheroids was approximately 2-fold higher than in hepatocyte monospheroids in short-term cultures [5]. In this study, we also showed that the albumin secretion of hepatocyte-endothelial cell heterospheroids labeled with TMDM (day 5, 60 µg-Fe/mL) was approximately 2-fold lower than that of non-labeled heterospheroids (day 5, 0 µg-Fe/mL) ( Figure 6A). The decrease in albumin secretion may be due to detaching the outer endothelial cells of the heterospheroids by labeling the nanoparticles.
In this study, we focused on anionic polysaccharide magnetic iron oxide nanoparticles (Resovist ® ) and synthesized four types of magnetic iron oxide nanoparticle derivatives with different properties. In our previous reports, the ASCs labeled with TMADM (≥30 mg-Fe/mL) could be detected by MRI imaging and could be traced for at least 14 days after transplantation with the skin in the mice [46]. Therefore, we expect to modify the clinical MRI contrast agents and use them as a visualization tool for artificial 3D liver transplantation. However, the liver function of hepatocyte-endothelial cell heterospheroids labeled with various nanoparticles decreased even at low concentrations (≥5 or 15 mg-Fe/mL). Pos-sible causes are (1) primary hepatocyte damage and (2) detachment of endothelial cells outside the heterospheroids [44] with the use of nanoparticles. To create an artificial 3D liver, the selection of cell sources is essential, including primary cells [47,48], immortalized cells [49], lineage cells, and stem cells [50,51]. Human hepatic parenchymal and nonparenchymal cells are difficult to obtain due to a lack of donors. The hepatocytes differentiated from induced pluripotent stem cells (iPSCs) [52][53][54][55], embryonic stem cells (ESCs) [56], and tissue stem cells [57] are expected to be a new cell source. Recently, in vitro 3D culture techniques have been used to create liver and multi-organ models [6,58]. Koike H. and Takebe T. et al. reported a technique for continuously creating liver, bile, and pancreas (HBP) structures from 3D cultures of human PSCs [58]. Peng W.C. and Nusse R. et al. have successfully cultured hepatocytes for more than six months in 3D culture [59]. These hepatocytes grown in 3D culture were transplanted into mice livers and significantly repopulated [59]. In the future, there is a need to improve the method of creation of artificial 3D liver or to design a low-invasive MRI contrast agent to label the artificial 3D liver.

Conclusions
In this study, we examined the morphology of the four synthesized polysaccharide magnetic iron oxide. The morphology of various nanoparticles showed relatively smoothedged spherical shapes. For the 3D liver models, we prepared primary hepatocyte-endothelial cell heterospheroids. The toxicity, CYP3A, and albumin secretory capacity were evaluated in the heterospheroids labeled with various nanoparticles. As the culture period progressed, the heterospheroids labeled with anionic and cationic nanoparticles showed lower liver function than non-labeled heterospheroids. In the future, there is a need to improve the method of creation of artificial 3D liver or to design a low-invasive MRI contrast agent to label the artificial 3D liver.