Introduction

Liver plays many essential roles in maintaining normal physiology of body, and one of them is complex metabolism of chemicals, toxins, and drugs. Therefore, it remains one of primary organs assessed for safety and efficacy evaluation of new drugs and chemicals (Okey et al. 1986). A number of drugs are synthesized and evaluated every year as well as chemicals are assessed for their potential toxic effects on biological systems. Toxicological and pharmacological preliminary studies designed to assess their toxicities on human liver rely heavily on extrapolation from animal models. A large number of animals are used for this purpose, which is an expensive and unethical practice.

During the last few years, efforts have been made to develop alternative in vitro models which can be used for evaluating human liver responses against chemicals, drugs, etc. (Clemedson and Ekwall 1999). This can be achieved by culturing either primary hepatocytes or established cell lines. There is a general consent that primary cells are more suitable for such model systems as they are freshly isolated and exhibit very high liver specific functions than the established cell lines. But every time, at least one model animal has to be sacrificed or operated for taking liver explant to initiate the primary culture. Using primary hepatocytes from animals has another major disadvantage that it results in bringing about unexpected differences in experimental data when humans are exposed as species difference leads to differences in the metabolism of drugs and chemicals. On the other hand using established cell lines have several advantages like their ease of maintenance, cryopreservation and revival, economical with regard to cost, time and technical expertise, and similarity of cells being used in different experiments. Although the cell lines become metabolically distanced from their tissue of origin and begin losing their tissue-specific properties after several passages, these can be used for mechanism studies for chemicals which are known to affect basal cell functions like antibiotics, chemotherapeutic agents, organic solvents, heavy metals etc. (Barile 1994). There are several reports which indicate that most cases of organ-specific toxicity is reflected as basal cytotoxicity distributed to the organs (Ekwall 1980, Acosta et al. 1985, Balls and Fentem 1992, Barile et al. 1993, and Hopkinson et al. 1993). Cell line methods measure concentration which represents basal cytotoxicity. When extrapolated to the in vivo situation, these values correspond to human blood concentration and may be interpreted as doses, which pose a risk to all human organs (Barile 1994).

HepG2 is one of the most extensively used cell lines which has been used to evaluate toxicity of chemicals and drugs (Knasmüller et al. 2004). The HepG2 cells from hepatocellular carcinoma retain many metabolic functions of human liver (Knoweles et al. 1980). The cells have been used for toxicological evaluations in two dimensional cultures (Aden et al. 1979). In human liver, hepatocytes are organized in a three-dimensional lobular structure, thus, monolayer two-dimensional culture, although is a good in vitro alternative but fails to accurately resemble the native microenvironment, which can cause difference in normal metabolic processes. Three-dimensional human liver cell culture addresses this shortcoming and can be used to mimic the native liver tissue architecture. This type of three-dimensional cellular growth not only exhibits more or less phenotypic properties comparable to its in vivo counterpart, but also, its functions match with the native tissue.

Three-dimensional growth of cells on polymeric scaffolds utilizing the principles of tissue engineering provides an alternative for monolayer culture. Cells when grown on suitable polymer under appropriate culture conditions can aggregate to form multicellular three-dimensional structures called spheroids. Cells in spheroids undergo structural reorganization and have been observed to exhibit enhanced specific activities and prolonged differentiated states compared to monolayer culture. Many researchers have reported the similarity of spheroid models to intact tissues in terms of structure and specific functions that have been reviewed (Mueller-Klieser 1997). Cells from adult pancreatic islands are capable of self-assembling to form aggregates that secrete insulin (Halban et al. 1987). Three-dimensional organoids were formed from normal human adult breast epithelial stem cells, which were almost identical to those found in vivo (Chang et al. 2001). Epithelial cells of mammary glands are able to reorganize into hollow alveolar-like structures and retain the functions of secreting milk proteins (Barcellos-Hoff et al. 1989). There is a substantial amount of evidence that cells growing in three-dimensional culture are more effective in terms of evaluating drug toxicity than cells in monolayer or dispersed culture (Hoffman 1991 and Dhiman et al. 2005). Many studies have demonstrated the elevated level of drug effect of spheroids compared with cell monolayer (Tokiwa et al. 1997 and Nakazawa et al. 1999) and their use in chemical toxicity (Dilworth et al. 2000).

The in vitro three-dimensional growth of cells depends on both the animal cells and matrix material. An ideal matrix should provide support to the cells for the adhesion, growth, and proliferation. It should have spatial and compositional properties that could attract the cells and guide them for their activity (Liu et al. 1999). The matrix should functionally imitate the extracellular matrix of the animal tissue and direct the cells to grow in three-dimensional shapes. A number of natural and synthetic polymers are currently being employed as tissue engineering scaffolds. Chitosan [β (1→4)-2-amino-2-deoxy-d-glucose] is one such material, which possesses appropriate properties to act as scaffold material. It is a unique polysaccharide derived from chitin, the second most abundant natural polymer next to cellulose. It is the major constituent of shells of shrimp and crab (Sandford and Steinnes 1991). The significant properties which make chitosan a suitable biomaterial for tissue engineering matrices include its (a) biocompatibility (Klokkevold et al. 1992), (b) biodegradability (Hirano 1989) without the toxicity of degraded products, (c) availability of reactive groups, (d) nontoxicity, (e) antimicrobial properties, (f) ease of chemical modification (Hirano et al. 1989), (g) high affinity to proteins, (h) easy fabrication into films, fibers, porous scaffolds and microspheres (Muzzarelli et al. 1989), and above all, (i) structural similarities to glycosaminoglycans [Muzzarelli et al. (1988)] , major components of extra cellular matrix of liver tissues. A number of studies have shown that the chitosan is a promising material to be used in liver tissue engineering applications (Yagi et al. 1997, Elçin et al. 1998, Khanna et al. 2000, Yang et al. 2001, Taek et al. 2002, Li et al. 2003, Wang et al. 2003, and Ding et al. 2004).

In the present study, the human hepatocyte cell line, Hep G2 has been grown in three dimensions using unmodified chitosan as matrix. Our approach involves the use of chitosan for the improved orientation and functions of Hep G2 cells. Herein, we report the three-dimensional “liver-like tissue” model with improved functionality taking Hep G2 cells and using chitosan as synthetic extracellular matrix for in vitro tissue engineering applications like the preliminary evaluation of the toxicity of drugs and other chemicals.

Materials and Methods

Materials.

Chitosan with 85% degree of deacetylation was purchased from Sigma, St. Louis, MO. Tissue culture media (MEM), sodium bicarbonate, and sodium carbonate were also purchased from Sigma, St. Louis, MO. Fetal bovine serum was bought from Gibco BRL, MO. Urea and Albumin estimation kits were purchased from Bayer, Germany. Other chemicals were of analytical grade and purchased locally.

Preparation of chitosan films.

Two percent (w/v) chitosan solution was prepared by dissolving chitosan in 1 M acetic acid. The chitosan films were prepared by dropping 1 ml of chitosan solution per well of 24 wells tissue culture plates (Nunc, Roskilde Denmark). The plates were kept in laminar hood at room temperature for drying the chitosan solutions into films. Excess of acetic acid from the films was neutralized by pouring sterile 0.1 M NaOH solution for 30 min, followed by repeated washing of films with autoclaved phosphate-buffered saline (PBS; pH 7.4). The neutralization process was monitored by measuring the pH of PBS taken out from the wells after each washing. The process was continued till the change in pH of PBS ceased. After complete neutralization, the plates were dried and sterilized by keeping them in UV light for 24 h. Control plates without chitosan were also treated in the same manner. The prepared plates were kept in sterile vacuum desiccators till the use.

Maintenance of HepG2 cell line.

The HepG2 cells are human hepatocellular carcinoma. The cells were procured from National Centre for Cell Science (NCCS), Pune (Maharashtra), India. These were routinely grown in MEM medium (Sigma) supplemented with 10% fetal bovine serum (Gibco), 1 mM sodium pyruvate (Sigma), 1.5 g/l, and sodium bicarbonate (Sigma) in tissue culture flasks (Nunc) in 37° C at 5% CO2 incubator for maintenance. The media of the cells was changed every third day, and the subculture was done on a weekly basis with 1:4 splitting ratio. The viability of the cells was determined by taking the cell suspension and staining with 0.4% Trypan blue at 1:5 ratios. Viable cells exclude the dye, while the dead cells retain it. The cells were counted using hemocytometer. In each experiment, the cell suspension with >95% viability was used.

Cell attachment studies.

The cell attachment studies were done on chitosan films and control plates without chitosan. Briefly, the sterilized films were incubated with 2 ml MEM complete medium per well. After 3 h, the medium was replaced with 1 ml cell suspension having cell density of 1 × 106 cells per ml. The films and control plates were allowed to incubate at 37° C in a 5% CO2 incubator. At different time intervals, the medium from each well from both plates was carefully collected in centrifuge tubes. The tubes were spinned at 800 rpm for 10 min, supernatant from each tube was discarded, and cell pellet was dispersed in 1 ml PBS. The unattached cells were counted using hemocytometer.

Determination of seeding density for spheroid formations.

The appropriate seeding density that would yield HepG2 cell spheroids was determined by taking the cells at three different concentrations of cells; 1 × 104, 3 × 104, and 5 × 104/ml and seeding them onto chitosan and control plates. The plates were kept at 37° C in a 5% CO2 incubator. The cells were observed under optical microscope for spheroid formation after every 24 h.

Morphological studies.

The morphology of the three-dimensional Hep G2 cell spheroids was observed through an inverted optical microscope (Leica, Germany) at ×20 magnification, and photographs were taken at days 1, 3, 6, and 9 of both control cells as well as spheroids. The viability of the cells was determined (at the same time of proliferation studies) by incubating them with (3-(4, 5-Dimethylthiazol-2-yl)-2, 5-diphenyltetrazolium bromide) (MTT) and then observing the formazon production by the live cells.

Cell proliferation studies.

The proliferation of cells was determined by MTT assay by plotting formazon absorbance against time. For cell culture experiments, chitosan and control plates were UV sterilized and washed with sterilized double distilled water. Each chitosan plate had some wells left blank (without chitosan) for the purpose of making them control for MTT assay. Seeded on the 24-well plates containing only chitosan films (not in blank wells) were 5 × 104 cells/well and control plates (leaving few wells blank like in chitosan). In blank wells of both chitosan and control plates, equal volume of complete media (with out cells) was poured. After inoculation, the plates were kept at 37° C in a 5% CO2 incubator. After every 24 h, in two wells of chitosan plate (one with chitosan film and other with only media) and two wells of control plate (one with cells and other with only media), 100 μl MTT (5 mg/ml in PBS of pH 7.4) was added in each well. The plates were again kept at 37° C in a 5% CO2 incubator till purple crystals of formazon appeared (Visual observation was made at this time.). The media from the well was carefully removed without disturbing the cells. Acidic isopropanol (0.1 N HCl in isopropanol), 1 ml, was added to each well to dissolve the crystals by keeping them in incubator at 37° C for 1 h. The solution from each well was collected in sterile centrifuge tubes spinned at 13,000 rpm for 2 min to remove any cell debris. The absorbance of the solutions was recorded at 570 and 690 nm, and background subtraction was done.

Albumin secretion analysis.

For albumin secretion studies, again, 5 × 104 cells/ well were seeded on the 24-well chitosan plates and control plates and keeping them at 37° C in a 5% CO2 incubator. The media samples were collected after every 24-h interval. The assay for albumin secretion was performed by using a commercial kit (Bayer, Leverkusen, Germany) as per given protocol. Briefly, the samples were brought at room temperature. Standard pure albumin (5 mg/ml) and samples were added with Bromo-cresol green reagent and incubated for 1 min. The color change was measured spectrophotometrically at 628 nm. The concentration of albumin in the samples was determined by extrapolating the obtained values with the standard albumin curve.

Urea synthesis assay.

The samples collected for estimating albumin were used for the quantitative determination of urea. The urea estimation was done using a urease-based commercial kit (Bayer) as per given protocol. Briefly, 5-ml glass tubes were taken for setting up the reactions, and each was added with 100 μl of urease. For blank and standard, 10 μl complete media without cells and standard pure urea was used. Ten microliters of each sample was added in a separate tube. The reaction was allowed to stand for 10 min at 37° C. Then, 1.5 ml of reagents A (phenol) and B (hypochlorite) were added in each tube and incubated for 15 min at 37° C. The color formed was measured spectrophotometrically at 546 nm. The concentration of urea in the sample was determined by extrapolating the obtained values with the standard urea curve.

Statistical analysis.

The data obtained by studying cell proliferation, albumin secretion, and urea synthesis were analyzed by t test using Sigmastat 2.0 software.

Results

Cell attachment kinetics.

Throughout the experiment, Chitosan matrix exhibited more cell attachment, i.e., less number of unattached cells than the control polypropylene plate (Fig. 1). During the first hour, around 20% more cells could settle down and adhere on the chitosan matrix than the control plate. By the fifth hour, there were more than approximately 80% cells adhering on chitosan, while on the control plate, approximately 65% cells adhered.

Figure 1.
figure 1

Attachment kinetics of Hep G2 cells on chitosan films and control tissue culture plate. Each point represents the mean of the results from three different experiments.

Full size image

Determination of seeding density for spheroid formation.

The optimal seeding density was found to be 5 × 104/ml to achieve spheroid formation of Hep G2 cells for specified period.

Morphological studies.

After 24 h of the seeding of the cells (day 1), the cells on both control pate and chitosan showed complete attachment to the surface (Fig. 2 a,b). The spheroids began to form on chitosan within 72 h (day 3), and it was difficult to identify individual cells in spheroids, but cells on the control plate continued to grow as flat monolayer (Fig. 3 a,b). The viability of the spheroids was confirmed by adding MTT and formation of purple formazon by them (Fig. 3 c). After 144 h (day 6), the spheroids were bigger in size and tightly organized (Fig. 4 c) producing formazon (Fig. 4 d), while the cells on the control plate increased in number but still were maintaining monolayer form (Fig. 4 a,b). The spheroids on day 9 were observed to be quite large (Fig. 5 c) and were found to be live as indicated by the formazon production by their cells (Fig. 5 d), but the cells on the control plate, as observed under microscope, were confluent and appeared to produce less formazon crystals (Fig. 5 a,b).

Figure 2.
figure 2

Optical photographs showing the growth of Hep G2 cells on first day on (a) control plate, (b) chitosan film, and (c) chitosan film with formazon producing cells. All photographs are at ×20 magnification.

Full size image
Figure 3.
figure 3

Optical photographs showing the growth of Hep G2 cells on the third day on (a) control plate, (b) chitosan film, and (c) chitosan film with formazon producing cells. All photographs are at ×20 magnification.

Full size image
Figure 4.
figure 4

Optical photographs showing the growth of Hep G2 cells on the sixth day on (a) control plate, (b) control plate with formazon producing cells, (c) chitosan film, and (d) chitosan film with formazon-producing cells. All photographs are at ×20 magnification.

Full size image
Figure 5.
figure 5

Optical photographs showing the growth of Hep G2 cells on the ninth day on (a) control plate, (b) control plate with formazon producing cells, (c) chitosan film, and (d) chitosan film with formazon producing cells. All photographs are at ×20 magnification.

Full size image

Cell proliferation studies.

For proliferation assay, 5 × 104/ml per well was taken, as it was found to be the most appropriate density to form spheroids for the experiments. Proliferation of cells on chitosan was initially almost similar to the control plate (Fig. 6) until day 4, but as the number of cells increased, the growth kinetics was found to be slightly lowered than the control plate. The maximum growth was observed on day 8 in both chitosan matrix and control plate. On the ninth day, the cells on both surfaces began dying.

Figure 6.
figure 6

Growth kinetics of Hep G2 cells on chitosan films and control plate. Each point represents the mean of the results from three different experiments.

Full size image

Albumin secretion studies.

The results indicate that albumin secretion by spheroids kept increasing up to day 3, then gradually decreased until the ninth day, while in monolayer, a sharp decrease in albumin secretion was observed until the ninth day after a slight increase on the second day (Fig. 7). The HepG2 cell spheroids on chitosan plates exhibited significantly more albumin secretion throughout the experiment than the cells growing on the control plate.

Figure 7.
figure 7

Albumin secretion by Hep G2 cells in spheroids and control monolayer. Each point represents the mean of the results from three different experiments.

Full size image

Urea synthesis assay.

Like albumin, the spheroids on chitosan showed significant increase in urea synthesis on the second day then a gradual decrease up to the eight day, and on the ninth day, the amount of urea was decreased very sharply [Fig. 8]. In control cells, increase in urea concentration was observed on day 2, but from the third day, it was sharply decreased up to the fifth day and then maintained a very low amount.

Figure 8.
figure 8

Urea synthesis by Hep G2 cells in spheroids and control monolayer. Each point represents the mean of the results from three different experiments.

Full size image

Statistical analyses.

The proliferation data of HepG2 cells over chitosan and over the control plate were analyzed by t test. The data passed the normality test (P = 0.112) and the equal variance test (P = 0.409). The difference in the mean values of the two groups was not great enough to reject the possibility that the difference was due to random sampling variability. There was not a statistically significant difference between the control and chitosan plate (P = 0.587). When the albumin secretion data were compared using t test, the data passed the normality test (P = 0.162) and the equal variance test (P = 0.959). The difference between the mean values between the data was found to be greater than what would be expected if occurring by chance. There was a statistically significant difference (P = <0.001). The power of performed test with alpha was 0.05:1.00. The urea synthesis data, when compared using t test, also passed the normality test (P = 0.343) and the equal variance test (P = 0.283). The difference between the mean values between the data was found to be greater than what would be expected if occurring by chance. There was a statistically significant difference (P = <0.001). The power of performed test with alpha was 0.050:0.996.

Discussion

Herein, we have been able to show that HepG2 cells can be grown on unmodified chitosan matrices into three-dimensional spheroids with liver tissue-specific functions closer to the native tissues. The HepG2 are the human hepatocellular carcinoma cells, which retain many liver specific functions like albumin secretion, urea synthesis, and glucose secretion functions. The cell line has been found useful in studying mechanism of toxicity (Aden et al. 1979). The attachment kinetics of HepG2 cells on the chitosan matrix was studied to examine the adhesive property of the matrix. For tissue engineering applications, the surface must have the adhesive property to attach the cells properly. An adherent cell, when it comes in contact with a surface, requires an adhesive force for anchoring it firmly on the surface. Only after a successful adherence will the cells grow and begin their life cycle. The adhesive force involves many nonspecific interactions like Van der Waal’s interactions, electrostatic interactions, etc. and specific forces like bonds between the cell adhesion molecules and their receptors. The attachment kinetics was studied by counting the unattached cells at a predetermined time interval. This method was chosen, as it was easier to count the unattached cells than to count the attached cells after their removal from the surface. As expected, the HepG2 cells kept attaching at higher number than the control chitosan plate. This is owing to the fact that the structure of chitosan is similar to glycosaminoglycans, one of the components of extracellular matrix of liver, involved in cell–matrix interaction in vivo.

Seeding density of cells is a very important criterion to begin cell proliferation studies in vitro. When the cell culture experiment was initiated with low seeding densities of 1 × 104/ml and 3 × 104/ml, the cells did not divide uniformly due to long intercellular distances and lack of proper cell–cell interactions, although they formed aggregates only when sufficient number of cells became available after a number of cell divisions. On the other hand, when the seeding density was increased to 7 × 104/ml, the cells achieved confluence (state of saturation) early forcing to abandon the experiment much before its scheduled time frame. For our experiments, the seeding density of 5 × 104/ml was found to be optimum where cells could aggregate on the surface of chitosan and grew into spheroids in a shorter culture period.

The viability of the cells in spheroids was evaluated by incubating the cells with MTT. MTT is a yellow compound which is converted into purple formazon in the reaction catalyzed by mitochondrial reductase enzyme in living cells (Mosmann 1983). Thus, the purple formazon is the indicator of live cells. During morphological studies, the spheroids began to form on chitosan within 72 h (day 3), and it was difficult to identify individual cells in spheroids, but cells on the control plate continued to grow as flat monolayer (Fig. 3 a,b). Furthermore, to confirm that the spheroids formed on chitosan plate were not a clump of dead cells, MTT was added in one of the wells, and the purple formazon production by the spheroids indicated their viability (Fig. 3 c).

The formation of three-dimensional structures from living cells depends not only upon cell–cell interactions but also upon cell–matrix interactions. In case of the poly propylene tissue culture control plate, the two dimensional cell–cell contacts were present, but no three-dimensional cell–cell interactions and cell–matrix interactions were occurring. On the other hand, chitosan structurally mimics glycosaminoglycans and maintains the integrity of the tissue. In general, the extracellular matrix components signal the cells to orient and organize themselves in a definite pattern to form a three-dimensional architecture. Thus, the cells cultured on chitosan matrix organized themselves into three-dimensional spheroids.

The MTT assay, apart from live–dead cell analysis, was also used for proliferation studies, as commonly used hemocytometer-assisted counting of mechanically separated cells from three-dimensional spheroids was found to be inaccurate. In the formed spheroids, the overall growth kinetics was lowered compared to control cells because cells started reorganizing in the form of three-dimensional spheroids on chitosan plate, and reorganization of cells in three dimensions induced them to differentiate rather than proliferate. Although the cells in spheroids kept proliferating and increasing in number up to the eighth day, on the ninth day, the proliferation stopped, but survival of the cells maintained in spheroids. On the other side, cells on monolayer started dying.

Many researchers have shown that the functions, viability, and proliferation of the cells are related to their morphology and organization of the cells (Roberts and Soames 1993, Hamamoto et al. 1998, Young et al. 2000, Baldwin and Saltzman 2001). To study the functionality of the spheroids, albumin secretion assay was performed, as secretion of albumin is one of the markers of function of hepatocytes. The HepG2 cell spheroids on chitosan plates exhibited significantly more albumin secretion throughout the experiment than the cells growing on the control plate. This happened as the cells forming spheroids were in three-dimensional structures, resembling more closely with the in vivo tissue architecture. This induced them to produce more albumin compared to the cells on the control plate. An adult human being produces around 12 g of albumin per day (Tavill and McCullough 1992). The average number of liver cells is assumed to be 2 × 1011, which is comparable to 60 μg/106 cells per day (Khalil et al. 2001). Our observations match with this value, i.e., ranging from 42–61 μg/106 cells/d.

Urea production is another key marker of hepatocytes, which is the main end-product of protein synthesis. On the control plate, the cells synthesized very less amount of urea compared to spheroids on chitosan. This may again be due to the three-dimensional organization of cells into spheroids on chitosan surface, resulting into more differentiated and functionally active state of the cells comparable to the native tissue. Urea, 24.96 g/l, is synthesized by an adult human being (Rudman et al. 1973), which is equivalent to about 200 μg/106 cells/24 h.

The statistical analysis shows that the proliferation of HepG2 cells on chitosan was not significantly different, but the chitosan was inducing cells to secrete more albumin and synthesizing more urea than the cells over control plate, comparable to liver tissue.

From the above observations, it is evident that the chitosan induced the HepG2 cells to reorganize into three-dimensional spheroids, and this three-dimensional architecture made the cells behave more closely to the native tissues in terms of their functionality due to improved cell–cell and cell–matrix interactions.

Conclusion

HepG2 cells organized into three-dimensional spheroids on chitosan. The spheroids exhibited three-dimensional cellular organizations with proper cell to cell and cell to matrix interactions, which mimics native liver cytoarchitecture. Furthermore, the cells in spheroids showed liver-specific functions in terms of urea and albumin secretions comparable to their in vivo counterparts. It shows that chitosan induced Hep G2 cells to form three-dimensional spheroids up-regulating the liver specific functions. On the other hand, the control monolayer cells lacked proper cell to cell contact and cell to matrix interaction, which is a must for tissue-specific gene expression. These miniature liver tissue-like spheroids can be a superior model than monolayer culture for the preliminary drug and chemical testing applications.