Construction of Artificial Hepatic Lobule-Like Spheroids on a Three-Dimensional Culture Device

Abstract

One major purpose of cell culture is the reconstruction of physiological structures. Using bovine aortic epithelium cell line HH (JCRB0099) as feeder cells and rat primary hepatocytes, we constructed hepatic lobule-like spheroids on a cell array plate designed for three-dimensional (3D) culture. Microfabricated patterning of the cell array with poly(ethyleneglycol) brushes promotes the formation of spheroids at 100-μm diameter at 100-μm intervals. Our standard protocol is to seed with feeder HH cells and then seed with primary hepatic parenchymal cells. The composite cell spheroids thus obtained are called heterospheroids. Feeder cells that were attached to the plate migrated and encompassed the spheroidal hepatocyte mass. Electron microscopy revealed Disse space-like structures characterized by hepatocyte-rooted microvilli rooted between hepatocyte and feeder epithelial HH cells. Differentiated hepatic functions such as albumin synthesis and cytochrome P450 subfamily CYP3A activities were maintained for 28 days in the heterospheroid versus monospheroid and monolayer cultures. In addition, glucuronide conjugation activity was maintained at a high level in heterospheroids. These results indicate that structurally similar hepatic lobules were formed in a microfabricated cell array coculture system and that the culture conditions are beneficial for maintaining differentiated hepatic functions.

Introduction

The liver is a seemingly monotonous organ but has a naturally well-organized microstructure. The minimum unit is a hepatic lobule wherein several types of cells cooperatively perform functions such as synthesis, detoxification, homeostasis, etc. (2). Artificial liver reconstruction is a popular topic in the field of regenerative medicine, cell and organ biology, and drug development. Morphological reconstitution may yield great advantages in long-term maintenance of hepatic functions. Usually, isolated primary hepatocytes are cultured on flat surfaces in monolayers. However, conventional monolayer culture does not provide satisfactory maintenance of liver functions. To improve culture conditions and output, a three-dimensional (3D) culture technology is being pursued.

Hepatic spheroids formed in 3D culture systems maintain liver functions for a longer period than do monolayer hepatocytes (6,9). A unique 3D culture system is the cell array that features a combination of micropatterning with poly(ethyleneglycol) brushes and bovine carotid epithelial HH cells as feeder cells. Micropatterning standardizes the size and the density of spheroids at a 100-μm diameter at 100-μm intervals. This standardization prohibits inside necrosis and provides a uniform cell density in the culture well. Furthermore, the feeder cells support the attachment and cultivation of hepatocytes (1,6). While micropatterning was performed initially only on glass plates, we developed a method to print the micropattern on conventional cell culture plates (12-, 24-, and 96-well plates, Cell-able) to create a more convenient tool for research. In this communication, we report the results obtained from a 12-well cell array.

Materials and Methods

Cells and Culture Plate

Bovine aortic cell line HH (JCRB0099) was obtained from the Health Science Research Resource Bank (Osaka, Japan). Rat hepatocytes were prepared from Wistar rats by the collagenase perfusion method (7). Collagenase (for cell isolation, 032-10534) was purchased from Wako Pure Chemicals (Osaka, Japan). Twelve-well cell culture plates for 3D culture (Cell-able) were supplied by Transparent (Chiba, Japan). For a morphological study, spheroids were formed on a glass-based cell array (21-mm diameter round, Transparent). A collagen type I-coated 12-well plate (CellTight, MS-0012K; Sumitomo Bakelite, Tokyo, Japan) served as a control. Culture medium was RM100 medium (Transparent), which is based on Dulbecco's modified Eagle's medium, containing 10% fetal bovine serum.

Hepatocyte Culture

Prior to inoculation with hepatocytes, 8 × 10⁴ bovine aortic epithelium HH feeder cells were seeded into each well. Wistar rat hepatic parenchymal cells (4 × 10⁵/well) were inoculated onto the HH-precultured cell array to form heterospheroids. In control groups, only hepatocytes (4 × 10⁵/well) were cultured on the cell array (monospheroids) or on a conventional culture plate (monolayer). Cells were cultured for 28 days, and differentiated hepatic activities were determined as described below.

Light and Electron Microscopy

Hepatocyte and feeder cell heterospheroids were cultured on cell arrays for 48 h, fixed with paraformaldehyde, and embedded in epoxy resin. For light microscopy, thin sections were stained with toluidine blue. Ultra-thin sections were processed for transmission electron microscopy.

Hepatic Differentiated Functions

Albumin production was assessed by measuring its accumulation in the culture medium in 24 h, using a rat albumin ELISA kit (E111-125; Bethyl, TX). Cytochrome P450 subfamily CYP3A activity was estimated by the rate of testosterone 6-β-hydroxylation. Cultured hepatocytes (heterospheroid, monospheroid, and monolayer) were incubated with 100 μmol/L testosterone (T1500; Sigma, MO) for 4 h, and the formation of 6-β-hydroxytestosterone was determined by HPLC with a C18 reverse-phase column (CD C18; Imtakt, Kyoto, Japan). Other metabolites (2-α-, 16-α-, and 6-β-hydroxytestosterone and testosterone glucuronide) were also determined by reverse-phase column ultraperformance liquid chromatography (UPLC; Waters Corp., MA).

Results

Microscopic Observation of Cell Array Surface

Circles of 100-μm diameter are aligned regularly at 100-μm intervals on the bottom of the culture plate (Fig. 1A). There are approximately 8,000 circles in each well of a 12-well plate and 800 in each well of a 96-well plate (Fig. 1A). The magnified photograph (Fig. 1B) shows the microfabrication more clearly. The cell culture area is inside the circle coated with collagen type I; the outer moiré-like area is the surface covered with superhydrophilic poly(ethyleneglycol) molecules that inhibit the attachment of protein and cells (Fig. 1B). The culture protocol is illustrated schematically in Fig. 1C. After seeding with bovine aortic-origin HH feeder cells (light pink), primary hepatic parenchymal cells (yellow) are inoculated. The red line marked P denotes poly(ethyleneglycol) layer where cells cannot spread.

Figure 1.

Microscopic appearance of the cell culture surface of a cell array (A, B) and schematic illustration of 3D coculture spheroid (C). Original magnifications: 25× (A) and 50× (B). Scale bars: 100 μm (A, B). (A) An arrow indicates the edge of the well. (B) The magnification of the square in (A). The cell culture area is inside the clear circle coated with collagen type I, and the outer moiré-like pattern is the surface covered with superhydrophilic poly(ethyleneglycol) molecules that inhibit the attachment of proteins and cells. (C) Yellow and light pink cells represent primary hepatic parenchymal cells and bovine aortic HH cells (feeder). P, poly(ethyleneglycol) layer.

The effective culture area is 20% of the total bottom surface. Simply stated, an adequate number of cells in a cell array is 20% of a conventional nonfabricated plate, as is the case for HH cells. On the other hand, cells such as hepatocytes that tend to aggregate can be cultured at 50–100% of the rate in conventional plates.

Morphology of Hepatocyte-Feeder Cell Heterospheroids on the Cell Array

Figure 2A shows a phase-contrast downward view of hepatocyte–feeder cell heterospheroids cultured for 48 h. With careful investigation, plain color HH cells and yellowish hepatocytes are distinguishable in each spheroid. The horizontal section stained with toluidine blue is shown in the same horizontal scale (Fig. 2B). Uniform spheroids were cultured regularly on the cell array. Each spheroid consists of 40–50 hepatocytes. A typical spheroid is shown in detail (Fig. 2C). HH cells are recognized by their spindle shape and hepatocytes by their box shape and rich dense granules of mitochondria. Interestingly, HH cells migrate from the original basement position to enwrap hepatocyte spheroids (arrowheads in Fig. 2C).

Figure 2.

Hepatocyte–feeder cell heterospheroids formed on the cell array. Rat hepatocytes were cultured for 48 h on the feeder cell–precultured cell array. Phase contrast micrograph of a downward view (A) and horizontal section stained with toluidine blue (B, C). Uniform spheroids were cultured regularly on the cell array. A typical spheroid is shown in (C). Arrowheads indicate HH cells that migrated from the culture plate and enwrapped spheroidal hepatocyte masses. Scale bars: 100 μm (A, B) and 10 μm (C).

An electron micrograph reveals the formation of a Disse space-like structure between a hepatocyte and epithelial feeder cell (Fig. 3). The most notable feature is the presence of hepatocyte-rooted microvilli. No similar structure was seen among the HH cells or hepatocytes (data not shown).

Figure 3.

Electron micrograph of hepatocyte (Hp) and feeder (HH) cell contact area. Disse space-like structure with hepatocyte-rooted microvilli was observed (A). m, mitochondria. (B) Magnification of square in (A). Arrows in (B) indicate the microvilli.

Albumin Synthesis and CYP3A Activity Under Various Culture Conditions

Rat primary hepatocytes cultured on a cell array with feeder cells (heterospheroid) maintained albumin synthesis and cytochrome P450-mediated monooxygenation much better than did hepatocytes cultured without feeder cells (monospheroid) or those cultured in a monolayer on a conventional plate (Fig. 4).

Figure 4.

Time course changes in albumin synthesis (A) and drug-metabolizing activities (B) of rat primary hepatocytes cultured under various conditions. Hepatocytes were cultured on a cell array with feeder cells (heterospheroid), without feeder cells (monospheroid), and on a conventional culture plate (monolayer). Albumin synthesis was measured by 24-h accumulation. Testosterone 6-β-hydroxylation was determined by 4-h incubation with 100 μmol/L testosterone. Each point indicates the mean ± SD (n = 3).

Albumin synthesis on day 3 was almost the same and increased on day 6 in all three groups (Fig. 4A). At its maximum, heterospheroid activity was elevated approximately twofold above that of monospheroids and monolayer. Then, heterospheroid activity decreased gradually, maintaining 62–25% of peak activity, still higher than the activity on day 3. The activities of monospheroids and monolayers fell rapidly after day 6.

Cytochrome P450 activity (CYP3A estimated as testosterone 6-β-hydroxylation) also increased in heterospheroids and was maintained until day 28. In contrast, hepatocytes of monospheroids and monolayers showed the highest activity on day 3 and deceased thereafter.

Table 1 shows a single-point study of other CYPs and glucuronidation activities on day 7. Again, hepatocyte heterospheroid was higher than in monolayer hepatocytes, as indicated by all four metabolites.

Table 1.

Comparison of Cytochrome P450 Activities and Glucuronide Conjugation of Rat Primary Hepatocytes in Heterospheroid and Monolayer Culture

	Heterospheroid*	Monolayer*
2-α Hydroxylation (2C)†	1.18	<0.25‡
16-α Hydroxylation (2B, 2C)	3.30	1.18
6-β Hydroxylation (3A)	2.35	<0.25
Glucuronide	90.5	28.8

^*Activities (nmol/h/well) on day 7.

^†Codes in parentheses indicate CYP subfamilies responsible for each metabolite.

^‡<0.25: under the detection limit.

Discussion

In vitro reconstitution of the physiological properties of live organs is an attractive but still unattained scientific goal. Sandwich culture (3), gel entrapment culture (8,10), organoid culture (5), coculture with nonhepatic parenchymal cells (9), and other methods have been reported to improve hepatocyte culture conditions. While the cell array has a simple structure in comparison to other devices, it promotes the natural activity and behavior of hepatocytes. Moreover, the use of feeder cells provides optimized conditions for hepatocytes. Our previous study showed that cryopreserved human hepatocytes with poor adhesion characteristics can be used for long-term culture experiments (1). In other words, feeder cells are the ultimate culture matrix.

Since the culture area is reduced to 20% of the total surface of the well bottom, high-density spheroidal culture is feasible even with smaller numbers of hepatocytes. In fact, our preliminary study indicated that the CYP-specific activity per cell number increased when smaller numbers of hepatocytes were cultured. In the case of cryopreserved human hepatocytes, this characteristic has a cost advantage.

Biologically, restriction of the culture area is likely to suppress the growth of the feeder cells that overgrow and are quickly released into suspension in conventional culture plates. This growth suppression may be the result of cell–cell contact inhibition. Induction of feeder cell growth arrest provides a wider range of time for inoculation of the hepatocytes. In addition, our recent work showed that cell arrays preseeded with feeder HH cells can be cryopreserved and stored as ready-to-use plates (4).

The most striking finding in this work is the existence of Disse space-like structures in the hepatocyte–feeder cell heterospheroids. This may be attributed to the origin of feeder HH cells that were established from the blood vessel epithelium. Unfortunately, intrahepatic microbile ducts were not observed in the heterospheroids. The culture time (48 h) may have been insufficient for spheroid maturation.

In comparison with the other culture conditions, heterospheroids maintained better hepatic functions. Not only CYP activities but also testosterone conjugation, a successive step in drug metabolism, was better maintained by heterospheroid culture.

In conclusion, our 3D cell array plate promotes the formation of hepatic lobule-like spheroids that have a structure similar to a Disse space between the endotheliumorigin feeder cells and hepatocytes. In addition to morphological reconstitution, the spheroid culture maintained hepatic differentiated functions.