|Home | About | Journals | Submit | Contact Us | Français|
Gel entrapment culture of primary mammalian cells within collagen gel is one important configuration for construction of bioartificial organ as well as in vitro model for predicting drug situation in vivo. Gel contraction in entrapment culture, resulting from cell-mediated reorganization of the extracellular matrix, was commonly used to estimate cell viability. However, the exact influence of gel contraction on cell activities has rarely been addressed. This paper investigated the gel contraction under varying culture conditions and its effect on the activities of rat hepatocyte entrapped in collagen gel within hollow fibers. The hepatocyte activities were reflected by cell viability together with liver-specific functions on urea secretion and cytochrome P450 2E1. Unexpectedly, no gel contraction occurred during gel entrapment culture of hepatocyte under a high collagen concentration, but hepatocytes still maintained cell viability and liver-specific functions at a similar level to the other cultures with normal gel contraction. It seems that cell activities are unassociated with gel contraction. Alternatively, the mass transfer resistance induced by the combined effect of collagen concentration, gel contraction and cell density could be a side effect to reduce cell activities. The findings with gel entrapment culture of hepatocytes would be also informative for the other cell culture targeting pathological studies and tissue engineering.
Collagen matrix, in particular type I collagen, are extensively focused in mediation of cell growth, differentiation, survival and tissue organization via regulating physiological responses of the entrapped cells (Bhadriraju and Chen 2002; Hansen et al. 2006). Initially, interactions between cells and collagen were mostly investigated using the fibroblast–collagen system in the wound-healing field (Souren et al. 1989; Berendsen et al. 2006). It is widely accepted that cell density and collagen concentration regulate the contraction of collagen gel, while no contraction was observed without cell-population. The degree of gel contraction is thus considered as a measurement for cellular activity in fibroblasts (Berendsen et al. 2006).
Collagen similarly regulates cellular functions of hepatocytes in gel entrapment culture for constructing bio-artificial liver (BAL). Two types of gel entrapment cultures are configured by entrapping hepatocytes in collagen gel at inner or outside of hollow fibers (Nyberg et al. 1992). Different from the two-dimensional culture in hepatocyte monolayer, collagen gel within BAL provides a three-dimensional microenvironment as that in liver, which could keep hepatocytes sustaining high liver-specific functions for around 2 weeks (Nyberg et al. 1993). Alternatively, hepatocytes in collagen gel entrapment culture were recommended as an effective model for interpretation of drug hepatotoxicity as well as drug metabolism in vivo (Coecke et al. 1999; Meng et al. 2007). Like fibroblasts, significant gel contraction was also observed in gel entrapment culture of hepatocytes while no such contraction occurred in cell-free gel (Nyberg et al. 1993) and hereby regarded as a crude estimation of hepatocyte viability (Nyberg et al. 1992).
Although gel contraction is a common feature in gel entrapment culture of mammalian cells, gel contraction and its association with cellular functions have never been examined. Further clarifications will be needed to facilitate the development of bioartificial organ as well as in vitro model for drug screening. In this paper, rat hepatocytes were selected as the biological systems to investigate the gel contraction under widely varying culture conditions and its association with the liver-specific functions.
Williams’ E basal medium was purchased from Gibco (Gaithersburg, USA). Bovine serum albumin, Linoleic acid, l-glutamine and Methyl Thiazolyl Tetrazolium (MTT) were purchased from Amresco Inc. (Solon, Ohio, USA). Transferrin and Type IV collagenase were purchased from Sigma (St. Louis, MO, USA). 4-nitrophenol (4-NP) was purchased from Hangzhou Huadong Medicine Group Company (Hangzhou, China). 4-Nitrocatechol (4-NC, Fluka) was kindly offered by Dr. Qiu Xinhui (Institute of Zoology, Chinese Academy of Sciences, Beijing, China). All other reagents were purchased from local chemical suppliers and were of analytical grade. Polysulfone hollow fiber membrane with average pore size of 100 KDa, outer diameter of 1 mm and inner diameter of 0.72 mm was purchased from Yuandong Pharmaceutical Machinery Corporation (Shanghai, China).
Rat tail collagen (type I) was prepared as described by Vinken et al. (2006). The concentrations of collagen were determined by hydroxyproline assay method modified from that of Reddy and Enwemeka (1996).
Rat hepatocytes were isolated from male Sprague–Dawley rats (weighing 200–250 g) by the two-step collagenase perfusion technique (Shen et al. 2006). Hepatocytes with a viability of at least 85% by trypan blue exclusion were used.
For gel entrapment culture, freshly harvested hepatocytes were mixed with the collagen solution and loaded into hollow fibers by injection as described before (Shen et al. 2006). Then, the hollow fibers containing gel entrapped hepatocytes were put into culture dishes with pre-warmed medium and incubated at 37 °C in a humidified incubator containing 5% CO2. The culture medium was composed of serum-free basal Williams’ E medium supplemented with albumin 500 ug/mL, linoleic acid 4.5 ng/mL, epidermal growth factor 5 ng/mL, dexamethasone 392 ng/mL, glucagon 4 ng/mL, transferrin 6 ng/mL, liver growth factor 2 ng/mL, insulin 0.2 U/mL, penicillin 100 U/mL, streptomycin 100 g/mL, l-glutamine 2 mmol/L, HEPES 15 mmol/L, NaHCO3 2.2 g/L, CuSO4·5H2O 0.1 uM and Na2SeO3 5.1 ng/mL. Culture medium was changed every 48 h.
The diameter of cylindrical gel was measured by a vernier in eyepiece on microscope. The extent of gel contraction was calculated as a percentage of the decrease on diameter over the initial one of 0.72 mm.
As primary hepatocytes in vitro were unable to proliferate, the MTT method was used to evaluate the relative viability of gel entrapped hepatocytes cultured in hollow fibers as previously described (Wang et al. 1996). Briefly, at the culture time of 0, 24, 48, 72, 96, 120, 144 h, gel entrapped hepatocytes were respectively extruded from the hollow fibers with a 5 mL-syringe and immerged in 0.65 mL of the MTT-phosphate buffer solution (1.15 mg/mL) in 24-well plates followed by incubation at 37 °C for 3 h which was interrupted by a short shaking every 1 h. Then, the MTT-phosphate buffer solution was discarded and 1.5 mL of isopropanol was added to the cells. After agitation for 1 h at room temperature, the MTT value was determined by recording the absorbance of the extracts at 570 nm. The cell viability of gel entrapped hepatocytes at the end of culture was presented as a percentage of MTT value normalized to the MTT value of the initially entrapped hepatocytes immediately after loading.
The integral of viable cell number was done as previously described (Renard et al. 1988). Briefly, according to MTT assays at various time points, the time course of viable cell curve could be obtained by the previously established standard curve of viable cell number versus MTT value using freshly isolated hepatocytes. Then the integral of viable cells was calculated by OriginPro 7.5 software through integration over the time course of the viable cell curve and presented by number of cells × h (hours)/mL, which mean the number of viable cells contributing the expression of cell activity over the integrated time for the duration of the integrated time.
Samples were taken every 24 h from culture medium after shaking the culture plates for about 2 min for well mixing. Urea secretion was assayed by Urea Nitrogen Kit from Nanjing Jiancheng Bioengineering Institute (Nanjing, China).
As CYP 2E1 decreased drastically during in vitro culture, the extent of its decrease was used as a sensitive marker to evaluate the relative liver-specific functions among different gel entrapment cultures. The CYP 2E1 activity was detected by the formation of 4-NC by HPLC as described previously (Shen et al. 2006). Briefly, the gel entrapped hepatocytes were extruded from the hollow fibers with a 5 mL-syringe and immerged in 0.5 mL of 4-NP-phosphate buffer solution (1 mM) in 24-well plates followed by incubation at 37 °C for 3 h which was interrupted by about 5 min’s shaking manually every 1 h. At the end of the incubation, the cell mixture was pipetted several times before taking samples for analysis of 4-NC formation by HPLC.
The data were presented as the mean of ±S.D. of three experiments (n = 3) while each experiment was performed in triplicates.
Gel contraction rate was first investigated by culturing rat hepatocytes under various culture conditions with the seeding cell density range of 0.5–2 × 106 cells/mL and collagen concentration range of 1.4–4.5 mg/mL. As indicated in Fig. 1a, when collagen concentration was 1.4 mg/mL, gel diameter rapidly decreased within 48 h of culture and later continuously decreased at a much lower speed. At the endpoint of 144 h, the diameter was only 45, 30 and 20% of the initial one for cell populations of 0.5, 1.0 and 2.0 × 106 cells/mL, respectively. Though in a less severe extent of gel contraction than that shown in Fig. 1a, the gel diameter at a collagen concentration of 2.75 mg/mL similarly decreased with the culture time and seeding cell densities as presented in Fig. 1b. However, at a collagen concentration of 4.5 mg/mL which was far beyond the common collagen usage of 2–3 mg/mL in gel entrapment culture, gel diameter was unexpectedly maintained for the gel entrapment culture at 144 h regardless of the seeding cell density used (Fig. 1c).
Since the extent of gel contraction is a cumulative process exerted by viable cells, the integral of viable cell number under various culture conditions was calculated based on the data shown in Fig. 1 for illustration of its effect on gel contraction. Unexpectedly, the gel contraction increased with viable cell number lacking a linearship relationship and was largely affected by collagen gel concentration (data not shown here).
For a convenient comparison, the final gel contraction extents of all gel entrapment cultures at 144 h are presented in Fig. 2a. It could be more obviously seen that, except for the gel entrapment culture at high concentration of collagen (4.5 mg/mL), gel contraction extent significantly increased with the seeding cell density and decreased with an increase in the collagen concentration at a collagen concentration of 1.40 mg/mL as well as 2.75 mg/mL. Hence, it seems that gel contraction directly increased with seeding cell density but inversely with collagen concentration.
Also cell viability was assayed under comparable gel entrapment cultures with various collagen concentrations and cell seeding densities (Fig. 2b). Cell viability decreased significantly with the increasing cell density under identical collagen concentration. This inverse association between cell viability and seeding cell density was in opposite to the correlation between gel contraction and seeding cell density. Nevertheless, at comparable gel entrapment cultures with identical cell densities but varying collagen concentration, no significant differences of cell viabilities were observed among the three cultures.
Hence, by comparing of Fig. 2a and b, it seems that cell viability was unrelated to gel contraction.
To compare the cellular performance of entrapped cells, the cumulative cell function on urea synthesis is presented in Fig. 3a with identical collagen concentration of 2.75 mg/mL. The cumulative urea production increased with culture time more drastically for cells cultured at cell density of 2.0 × 106 cells/mL than that at lower seeding density (0.5 and 1.0 × 106 cells/mL). The urea secretion of hepatocytes cultured at the other two collagen concentrations (data not shown) showed a similar trend as in Fig. 3a. Figure 3b demonstrates the corresponding time course of integral of viable cell number at a collagen concentration of 2.75 mg/mL. At each time point, the integral of viable cell number increased with seeding cell density. Hepatocytes entrapped in the two other collagen concentrations showed a similar trend as in Fig. 3b (data not shown). Then the accumulative urea secretion and its corresponding integral of viable cell at each time (0, 24, 48, 72, 96, 120, 144 h) were drawn in Fig. 4 with trend lines to investigate their relationship. The slopes of those trend lines in Fig. 4 are summarized in Table 1, from which it can be seen that urea secretion per viable cell decreased with the increasing seeding cell density. The difference on urea secretions among various collagen concentrations was insignificant when initial seeding cell density was fixed. Considering that gel contraction was strongly impacted by collagen concentration, the non-correlation of urea secretion with gel contraction could be concluded.
Another liver-specific function, CYP 2E1 activity on basis of viable cells, is presented in Fig. 5. It can be seen that the CYP 2E1 activity shared a similar trend for the three collagen concentrations. This metabolic activity generally decreased with increasing seeding cell density, excepting that a higher function of CYP 2E1 activity was presented for a cell density of 1.0 × 106 cells/mL, compared with the two other cell densities at identical collagen concentration of 2.75 mg/mL at 144 h. It seems that CYP 2E1 activity is almost unrelated to collagen concentrations or gel contraction.
Overall, cell viability as well as liver-specific functions decreased with seeding cell density and were unassociated with gel contraction.
It is well known that the cytoskeleton of the cell provides the tensional force to counteract the mechanical load of extracellular matrix and such contractile forces generate gel contraction (Kolodney and Wysolmerski 1992). In this way, the cell can reorganize the extracellular matrix and rebuild micro-environment in vivo. But the relationship between gel contraction and cellular performance has never been addressed in previous studies.
Our results show that gel contraction mediated by hepatocytes increased with seeding cell density and inversely with collagen concentration, corresponding well with fibroblast or human bronchial epithelial cells (HBEC) mediated gel contraction (Liu et al. 1998). It is reasonable that a dense cell population with a highly viable cell number could provide more integrin to elicit shrinkage force of collagen fiber and thus facilitate gel contraction. Normally, gel contraction was observed for a collagen concentration range of 0.5–2.0 mg/mL (Liu et al. 1998; Wu et al. 1996). Thus, it appeared unexpected that no significant contraction occurred at a high collagen concentration of 4.50 mg/mL. We premised that increasing collagen concentration could reduce the moisture among collagen fibers and cause more compactness and polymerization of collagen gel. In addition, the blockage of cell mobility caused by stronger polymerization of dense collagen could counteract the contractile forces of hepatocytes and thus cause no gel contraction.
Previously, the active interaction between collagen and the embedded cells was assumed to enhance cellular functions (Wu et al. 1996; Yamada et al. 2001) and, furthermore, high cell density culture was highly suggested to enhance cell viability and hepatocellular functions such as albumin and urea secretion and detoxification (Dvir-Ginzberg et al. 2003). Nevertheless, in this paper, the non-correlation between gel contraction and cell activity was unexpectedly concluded as a result of the fact that cell viability as well as liver-specific functions per viable cell were lower in the gel entrapment cultures at higher cell densities regardless of the more severe gel contraction observed. Moreover, the fact that the liver-specific functions expressed normally at a high collagen concentration of 4.50 mg/mL without gel contraction provided a solid support for the non-association of cell activities with gel contraction. More caution is necessary concerning the previous claim that gel contraction is a marker for cell viability and functions (Berendsen et al. 2006; Nyberg et al. 1992).
To our opinion, the non-association of cell functions with gel contraction might result from the intricate effect of diffusion of nutritions. The providing of nutritions was largely determined by diffusion across the 3D gel-cell matrix (Millis et al. 2002; Legallais et al. 2001). Thus, the diffusion coefficient across gel-cell matrix has been extensively explored and we summarized these results in Table 2. From this table, it can be seen that high collagen concentrations alone significantly reduced diffusion (Ramanujan et al. 2002; Kanamori et al. 2000). In addition, the entrapment of bacterial or mammalian cells in gel obviously inhibits mass transfer by decreasing the diffusivity as a result of enhanced tortuosity (Mateus et al. 1999; Nicholson and Phillips 1981; Nicholson and Sykova 1998). As the embedded viable cells mentioned above possibly exacerbate the reduced diffusivity, our previous entrapment of dead cells without causing contraction could better reflect the contribution of cells alone to the reduced diffusivity (Wu et al. 2005). Thus in this paper the contribution of cells (by taking into account seeding cell density as well as cell activity) on mass transfer resistance was studied. Further, the increased reduction of the diffusivity in loading with viable cells than with dead cells illustrated the negative contribution of contraction alone (Wu et al. 2005). As a result, when cell densities were fixed in gel entrapment culture, increasing collagen concentrations could trigger two counteractive impacts on mass transfer: reducing mass transfer across collagen; enhancing mass transfer due to less severe gel contraction. This could result in the comparable cellular activities among these cultures with varying collagen concentrations. Differently, the increasing cell density in gel entrapment cultures with identical collagen concentration caused two synergistic side effects on lowering mass transfer via enhancing tortuosity as well as gel contraction, finally lowering cell viability. Taken together, gel entrapment at high cell density could largely inhibit mass transfer and thus cell performance.
Nevertheless, high cell density as well as high cellular functions were required for clinical use of BAL and in vitro model of drug screening platform. Dvir-Ginzberg et al. (2003) adopted alginate scaffolds with high porosity and large pore sizes to improve mass transfer for sustaining high cell viability under high cell loading. In consideration of the adverse effect of gel contraction on cellular functions, we strongly recommend that gel contraction should be avoided due to its negative effect on mass transfer. One strategy is the addition of antibodies for binding with integrin of hepatocytes and thus blocking gel contraction (Gullberg et al. 1990, 1992). The other alternative could be the addition of matrix metalloproteinase inhibitors to prevent cell migration and contraction (Wong et al. 2004).
In summary, this paper indicates that hepatocytes at a high collagen concentration failed to elicit gel contraction but sustained cell viability and liver-specific functions to some extent. The non-association of gel contraction with cell activities is in contrast to the previous claim on the positive association between gel contraction and cell activities. Due to the adverse effect of mass transfer resistance on reduced cell activities induced by gel contraction, reducing gel contraction would be desirable to improve gel entrapment culture of hepatocytes as well as other mammalian cells. Our findings will benefit tissue engineering and pathological studies as well.
This research was supported by Grants (No. 30772614) from National Natural Science Foundation of China (NSFC) and partly by Hi-Tech Research and Development (863) Program of China (2006AA02A140).