Lab on a chip-based hepatic sinusoidal system simulator for optimal primary hepatocyte culture
Yoon Young Choi1, Jae Chul Lee1, Jaehyung Kim1, Sang-Hoon Lee2, and Dong-Sik Kim1*
1 HBP Surgery & Liver Transplantation, Department of Surgery, Korea University College of Medicine, Seoul, Korea
2 Department of Biomedical Engineering, Korea University, Seoul, Republic of Korea
*Corresponding author:
Dong-Sik Kim, M.D., Ph.D.
Professor of Surgery
Department of HBP Surgery and Liver Transplantation
College of Medicine, Korea University
145 Anam-ro, Seongbuk-gu
Seoul, 136-701, Korea
E-mail:
Abstract
Primary hepatocyte cultures have been used in studies onliver disease, physiology, and pharmacology. While they are an important tool for in vitro liver studies, maintaining liver-specific characteristics of hepatocytes in vitro is difficult,as these cells rapidly lose their unique characteristics and functions. Portal flowis an important condition to preserve primary hepatocyte functions and liver regeneration in vivo. We have developed a microfluidic chip that does not require bulky peripheral devices or an external power source to investigate the relationship between hepatocyte functional maintenance and flow rates. In our culture system, two types of microfluidic devices were used as scaffolds: a monolayer- and a concave chamber-based device. Under flow conditions, our chips improved albumin and urea secretion rates after 13 days compared to that ofthe static chips. Reverse transcriptionpolymerase chain reaction demonstrated that hepatocyte-specific gene expression was significantly higher at 13 days under flow conditions than when using static chips. For both two-dimensional and three-dimensional culture on the chips,Moreover,three-dimensional culture on the concave chip underflow resulted inthe best performance of the hepatocyte culture in vitro. We demonstrated that flow improvesthe viability and efficiency of long-term culture of primary hepatocytesand plays a key role in hepatocyte function. These results suggestthat this flow system has the potential for long-term hepatocyte culturesas wellas a technique for three-dimensional culture.
Keywords: hepatocyte, microfluidic, shear stress, spheroid,
1. Introduction
Hepatocytes represent 70–80% of the total liver cell population and express most xenobiotic-metabolizing enzyme activities (Guillouzo et al. 1993; Hewitt et al. 2007; Sivaraman et al. 2005). Primary hepatocyte cultures are recognized as one of the most relevant and practical models for the study of drug metabolism and transporter interactions (Hewitt et al. 2007). Although hepatocytes have the unique characteristics of being highly differentiated and capable of growth in vitrovivo,under culture conditions, theylose their ability to properly function in long-term culture(Cho et al. 2010; Khetani and Bhatia 2008; Kojima et al. 2009; Lee et al. 2010; Tanaka et al. 2006, van Poll et al. 2006). This loss of function may be especially important in toxicological and metabolic studies, where the time scale for responses may exceed the time scale for loss of hepatocyte function. To overcome these current limitations, many culture strategies have been developed to simulate the environment of normal hepatic structures, such as cultures on porous membrane gels (Ostrovidov et al. 2004) stacked cultures(Sudo et al. 2005), co-cultures with non-hepatic cells(Miyazawaet al. 2005; Scott et al. 2005), three-dimensional (3D) cultures of hepatocyte aggregates known as spheroids(Glickliset al. 2004), andsandwich cultures (Farghali et al. 1994; Langsch and Bader 2001)and cultured with Extra cellular matrix overlay (Sindhu et al. 1994).
Clinically, selective occlusion of the portal branch causes atrophy of the ipsilateral hepatic lobe and hypertrophy of the contralateral lobe (Rous and Larimore 1920). Also, some research suggests that hemodynamic stress on hepatocytes is the main cause of small-for-size liver graft injury (Man et al. 2001; Xu et al. 2006). Change in hepatic blood flow induces a rapid change in liver volume andhepatic regeneration by activating mediators, such as growth factors and several hormones,at the cellular level (Yokoyama et al. 2007). In such situations, shear stress is thought to inevitably affect the morphology of hepatocytes.Many studies show that shear stress induces cell-cell interactions, albumin production, and urea secretion in vitro and,in several ways helps to maintain the viability of hepatocytes (Kan et al. 1998; Tanaka et al. 2006). Adequate culture conditions with an appropriate shear stress acting on hepatocytes are necessary for flow-based culture experiments with hepatocytes. Recently, microfabrication technology has enabled the development of a microfluidic chip-based culture system that provides control over environmental conditions and generates stable flow profiles. A potential improvement would be to use microfluidic-based cell culture platforms, which enable the creation of cellular environments that mimic a number of important in vivo attributes(Lii et al. 2008; Kim et al. 2007; Tourovskaia et al. 2006), including cell-cell interactions(Schutte et al. 2011; Zervantonakis et al. 2011), oxygen and nutrient delivery(Domansky et al. 2010; Evenou et al. 2011), metabolite removal, and shear stress (Parket al. 2005).
In this study, we built two types of microfluidic chips – monolayer and concave –in which primary hepatocytes were exposed to a continuous and stable flow of media. The monolayer chamber was fabricated using a soft-lithograph procedure, which enabled the adhesion of hepatocytes (Whitesideset al. 2001). The concave chamber was fabricated with a meniscus-based method for culturing hepatocyte spheroids (Lee et al. 2013). The chip that we used enables the generationofa flow rate that mimicsthat of the artificial sinusoidal system over hepatocytes. This flow rate maintainshepatocyte viability and function and allows for removal of waste products and a continuous supply of oxygen and nutrients; it creates an environment as similar to in vivo as possible without bulky peripheral devices and an external power source. In these systems, the flow rates were controlled by the polyethylene glycol (PEG) concentration. The purpose of this study was to evaluate albumin and urea synthesis and the morphology of hepatocytes under various flow rate conditions. Our results suggest that this flow-based culture system maintains the biological functions of hepatocytes in vitro as well as providing a cost-effective and self-powered screening platform for pharmaceutical development.
2. Materials and methods
2.1. Fabrication of flat and concave chips
Microfluidic devices were composed entirely of two substrates. The flat chip was fabricated through a replica-molding process by using soft lithography, and the concave chip was fabricated using the surface tension of polydimethylsiloxane (PDMS) pre-polymer, as we have previously described (Lee et al. 2013; Whitesides et al. 2001).
2.2. Preparation of flat and concave chips
Both chips were sterilized by autoclaving (120˚C for 30 min) and were dried in an oven. The flat chamber was coated with 5g/mL poly-L-lysine overnight and 5g/mL collagen type Ⅰfor 1 h to improve cell adhesion. After coating, flat chips were rinsed three times before seeding cells in phosphate-buffered saline (Invitrogen, Carlsbad, CA, USA). Unlike the flat chips, the concave chips were coated with 3% bovine serum albumin (BSA) overnight to prevent cell adhesion to chips and the formation of spheroids.
2.3. Isolation of primary hepatocytes
Hepatocytes were isolated from adult male Sprague-Dawley rats (DBL, Seoul, South Korea) weighing 250 g using a two-step collagenase perfusion procedure (Seglen 1976). The viability of harvested hepatocytes was greater than 90%. Primary hepatocytes were cultured in high-glucose Dulbecco’s Modified Eagle’s Medium supplemented with 20 mM HEPES (Sigma-Aldrich, St. Louis, MO, USA), 25 mM NaHCO3, 30 mg/L proline, 10% fetal bovine serum (FBS), 25 U/mL penicillin, 25 mg/mL streptomycin, 10 mg/mL gentamicin, 10 ng/mL epidermal growth factor (EGF), 50 ng/mL insulin, 0.1 mM dexamethasone, 10 mM nicotinamide, and 100 mM L-ascorbic acid.
2.4. Osmosis-driven flow
PDMS cubic chambers with one cellulose membrane window were fabricated using conventional protocols (Park et al. 2007); this is hereafter referred to as the osmotic pump. Bonding of a cellulose membrane with a PDMS chamber was performed using the PDMS solution as adhesive glue. As a preliminary study, we conducted osmosis experiments to evaluate the pumping capability of the osmotic pump.Deionized water was used as a buffer solution, and polyethylene glycol (PEG; Sigma-Aldrich; molecular weight = 2000) solution was used as a driving agent.
2.5. Hepatocyte monolayer cultures on flat chip and spheroid cultures on concave chip
For monolayer culture, 100l of hepatocytes suspension (106 hepatocytes/ml) in culture mediaon flat chips, hepatocyte suspensions (adjusted to 1 × 105 cells/100L)were directly seeded on collagen-coated flat chipsin culture medium. The collagen that we used in flat chip was used only enhancing the adhesion of hepatocytes. Also, collagen overlay method was not used in this experiment in order to compare the flow effect.When hepatocytes had attached to the substrate in 12hours after cell seeding,non-adherent cells were washed out, and a flow of culture medium was applied. Medium was replaced every 48 hours for the duration of the experiment. For culture on concave chips, 100l of hepatocytes suspension (106hepatocytes/ml)a primary hepatocyte suspension (1 × 105 cells/100L)was loaded into the microchannel using a micropipette, allowing the cells to be trapped within the concave microwells. Most cells were evenly dockedaggregatedwithin concave microwells. Cells were left in the incubator for 30 min without flow for stabilization of cells within the microwells, after which a flow of culture medium was gently applied to remove cells that did not dock within the microwells. Identical amounts of hepatocytes remained within the concave microwells. We inserted 1000l micropipette tip as inlet reservoirs which continuously provided the culture medium. And then flexible polyurethane tube was inserted into the outlet of microchipsOnce the chips were enclosed with a flexible polyurethane tubeand connected to the osmotic pumpto initiate flow, the microfluidic device was transferred to a humidified CO2 incubator at 37˚C.
2.6. Cell viability
Cell viability was assessed by incubating spheroids with 50 mM calceinAM and 25 mg/mL ethidium homodimer-1 in culture medium for 40 min at 37˚C, and then imaging cells with an inverted fluorescence microscope.
2.7. Functional assessment
Albumin and urea secretion were analyzed by measuring their concentrations in medium conditioned by either monolayers cultured on flat chips or spheroids cultured on concave chips. After culturing for 1, 3, 5, 7, 9, 11, or 13 days, the medium that remained in the tube connected to the osmotic pump was collected for albumin and urea measurements. The tube was re-filled with fresh autoclaved water to re-initiate osmotic pump.and replaced with fresh medium, and albumin and urea concentrations were measured.
2.8. Reverse transcription-polymerase chain reaction (RT-PCR) assay
Primary hepatocyte cells were cultured on flat and concave chips for 3, 5, 7, 9, 11, or 13 days in vitro. In flat chip, for detached hepatocytes from the flat surface, the cells were treated with Tryp-LE TM express (Invitrogen, Carlsbad, CA) and then collected the cell suspension into tube.Also, in concave chip, spheroids were gently retrieved using the micropipetteand subsequently resuspended in hepatocyte culture medium. Then both cell suspensions were centrifuged in 1200rpm for 2min. Cell pellets digestion in TRI-Reagent was performed, followed by chloroform extraction and precipitation with isopropyl alcohol.The cells were gently retrieved and subsequently resuspended in hepatocyte culture medium. Digestion in TRI-Reagent was performed, followed by chloroform extraction and precipitation with isopropyl alcohol.cDNA was synthesized from two microgram of total RNA using PrimeScriptTM 1st strand cDNA synthesis kit (TAKARA, Japan) according to the manufacturer’s instructions. One microliter of the cDNA reaction mixture was subjected to PCR amplification using gene-specific primers (Table. 1) the AccuPower® PCR PreMix (BIONEER, Korea).The PCR condition were 95˚C for 5 minutes followed by 25 cycles at 95˚C 30 seconds, 56˚C for 30 seconds, and 72˚C for 30 seconds. PCR products were analyzed on a 2% agarose gel.
2.9. Acquisition of gel images and quantitative analysis
Images of the RT-PCR ethidium bromide-stained agarose gels were acquired with an Olympus High Performance CCD camera (Olympus Corporation,Japan), and quantification of the bands was performed with ImageJ. The ratio between each sample and GAPDH was calculated to normalize for initial variations in sample concentration and as a control for reaction efficiency. Mean and standard deviation of all experiments performed were calculated after normalization withGAPDH.
2.10. Statistical analysis
Results are expressed as means ± standard deviation. Experiments for albumin and urea at specific wall shear stresses were conducted in five independent experiments. Hepatocyte-specific gene expression assessments under flow were conducted in triplicate. Statistical analysis of data was performed using the Kruskal-Wallis and Mann-Whitney tests, with P < 0.05 considered significant. SPSS statistics23.0 (IBMCorporation, Armonk, NY, USA) for Windows was used for all statistical analyses.
3. Results
3.1Evaluation of osmotic pump
The flow volume generated by the osmotic pump depends on the molar concentration of the PEG solution. To determine the relationship between the osmosis-driven flow rate and PEG concentration, osmotic pumps were prepared. We then estimated the flow rate of0.1M and 0.25M PEG solutions in the chip using a computational fluid dynamic simulation (COMSOL, USA), as previously described (Lee et al. 2013). In our system, medium flow was stably maintained over 14 days. The average velocity was 8.75 × 10-6 m·s-1 for the 0.1M PEG solution and 2.06 × 10-5 m·s-1 for the 0.25M PEG solution. The volumetric flow rates of the medium were set at 0.251L/min for the 0.1M PEG solution and 0.628L/min for the 0.25M PEG solution.Microchip height was maintained at 100m, resulting in wall shear stresses of 1.01 × 10-6 Pa·s for the 0.1M PEG solution and 2.51 × 10-6 Pa·s for the 0.25M PEG solution. The average transferred volume of culture medium was 180L /12h in the 0.1M PEG solution and 450L /12h in the 0.25M PEG solution (Fig.1).
3.2. Morphological observations of hepatocytes cultured on flat and concave flow chips
After the flat and concave microfluidic chips were uniformly filled with hepatocytes in suspension, the cells were cultured overnight in an incubator without flow perfusion to allow the attachment of cells to their substrate or the aggregation of cells in concave flow chips. Hepatocytes aggregated uniformly into spheroids in each concave microwell, owing to the funneling effect and lack of necessary cell attachment protein on the surface of the concave microwellmorphological effect of the concave microwell structure. The viability of the hepatocyte monolayersunder osmotic flowexposed to the 0.1M and 0.25M PEG solutions was approximately 95%, even after 7 days in culture. The viability of spheroids exposed to this medium was higher than that ofthe hepatocyte monolayer (Fig. 2a). In addition, hepatocyte viability was lower under static conditions than flow conditions. These results demonstrate that medium flow and 3D culturing positively affect the viability of hepatocytes. In these systems, convective flow rates provide a continuously fresh nutrient supply and long-term stability of flow rate over several days. This osmotic pump-based flow chip enabled the mimicking of artificial sinusoidal flow, controlled by the concentration of PEG (0.05 M, 0.1 M, or 0.25 M). We investigated the effect of various flow rates on the stability of both monolayer- and concave-cultured hepatocytes on flow chips. On monolayer chips, flow rate did not affect morphological features; the original monolayer of hepatocytes exposed to various flow rates were firmly attached to the flow chips and maintained their morphology past day 14. However, on concave chips, cultured spheroids exposed to high flow rates lost their stability, and cells were dispersed outside the concave microwells.Those exposed to low flow rates, however, were stable without cell loss until day 13 (Fig. 2b).
3.3. Functional validation of hepatocytes cultured on flat and concave flow chips
To determine the relationship between hepatocytes and flow, we established three different experimental groups: (1) a monolayer culture without flow, (2) a monolayer culture with various flow rates, and (3) a spheroid culture with flow. In the monolayer culture without flow, hepatocytes were seeded in a flat chamber, and no osmotic pump was used. In the monolayer culture with flow, an osmotic pump was used, and the effects of flow on hepatocytes were observed. We compared the performance of hepatocytes in static culture and flow culture models by measuring secreted albumin and urea concentrations, which showed significant flow-dependent differences. Of the three models, static cultures lost their hepatocyte-specific functions most rapidly. In contrast, monolayer and spheroidscultures under flow conditions maintained their function for a longer period than those in static culture conditions. In additionMoreover, the 3D, spheroid culture with flow group exhibited the highest albumin and urea production after 13 days (Fig. 3). In additionparticular, urea secretion of the spheroid culture with flow was significantly higherthat ofthanthe monolayer culture with flow rates of 0.1M (p=0.016) and 0.25M (p=0.016) PEG solution.other experimental group (P =0.016)We also compared a static 3D culture and the 3D culture with flow and found asignificant difference in the amounts of albumin and urea produced. The 3D flow culture condition showed increased albumin secretion and improved urea synthesis when compared with the 3D static cultures. These results demonstrate that flow systems affect the preservation maintenanceof hepatocyte function in two- and three-dimensional culture conditions.
3.4. Hepatocyte-specific gene expression of hepatocytes cultured on flat and concave flow chips
We analyzed the hepatocyte-specific functions of cells cultured under flow conditions by evaluating the expression of gene markers from days 3 to 13 (Fig. 4). We found that UGT1A5was highly expressed in hepatocytes grown under flow conditions, contrastingwith a lower expression in hepatocytes grown under static culture; this indicates that phase 2 metabolic enzymes are affected by flow. An analysis of the time-dependent expression of CYP1A2, a marker of xenobiotic metabolism, showed that CYP1A2 expression was higher underflow conditions than under static conditions. Moreover, CYP1A2 expression in spheroids remained relatively high until day 13. These results suggest that spheroids hepatocytescultured underflow conditions can preserve their liver-specific functionality to a greater extent than monolayers hepatocytescultured under either flow orstatic conditions. In addition, expression of Mrp1, a multi-drug resistance marker, was high in the concave flow culture and remained at a high level throughout the experiment. This result is supported by a quantitative analysis of liver-specific gene expression, which also shows greater expression of liver function-related genes in the 2D and3Dflow model than in the monolayer flowstatic culturemodel. These results demonstrate that hepatocytes cultured in flow systems maintained their functionality and viability better than those in static culture systems. Collectively, the gene expression data indicate that hepatocyte functions were higher under media flow.was affected by the flow of the medium.and the 3D state of the cells.