Journal of Chromatography A, 1111 (2006) 233–237

Cell culture and life support system for microbioreactor and bioassay

Yuki Tanaka a,b, Kiichi Sato a,b,c,1, Masayuki Yamato b,d,Teruo Okano b,d, Takehiko Kitamori a,b,c,∗

a Department of Applied Chemistry, School of Engineering, The University of Tokyo, Tokyo, Japanb Core Research for Evolutional Science and Technology (CREST), Japan Science and Technology Corporation, Saitama, Japan

c Integrated Chemistry Project, Kanagawa Academy of Science and Technology (KAST), Kawasaki, Japand Institute of Advanced Biomedical Engineering and Science, Tokyo Women’s Medical University, Tokyo, Japan

Available online 11 July 2005

Abstract

A microchip-based cell culture system was developed and a primary culture of rat hepatocytes was realized in the system. The microchipwas made of glass plates and had a microchannel and a microculture flask inside. The flask inner surface was coated using collagen solution;then FBS and DMEM were added successively. Rat hepatocytes suspended in a medium was introduced into the microchip and incubated at37 ◦C in a humidified atmosphere with 5% CO2. Because of the shortage of dissolved oxygen, the cultured cells in the microchip resultedich©

K

1

nccttdat

stt

SJ

0d

n a significant decrease in viability. To overcome this, a continuous medium flow oxygen and nutrition supplying system was designed andonstructed. The system realized good cell growth for at least 4 days. Liver-specific functions, such as the synthesis of albumin and urea fromepatocytes were confirmed.

2005 Elsevier B.V. All rights reserved.

eywords: Microchip; Hepatocyte; Cell culture; Albumin synthesis

. Introduction

In recent years, integrated chemical systems, which areamed micro total analysis systems (�-TAS) [1] or labs-on-a-hip [2] have become of major interest especially to analyticalhemists due to their desirable characteristics, such as reduc-ions in reagent consumption, required space, and analysisime. Taking advantage of these characteristics, we haveemonstrated many applications, including flow-injectionnalysis, solvent extraction, immunoassay, laser reaction con-rol and enzyme reaction use [3].

We also demonstrated that microchip techniques haveome advantages for a cellular biochemical analysis sys-em [4], because the scale of the liquid microspace insidehe microchip is fitted to the size of the cells. For instance,

∗ Corresponding author. Fax: +81 3 5841 6039.E-mail address: kitamori@icl.t.u-tokyo.ac.jp (T. Kitamori).

1 Present address: Department of Applied Biological Chemistry, Graduatechool of Agricultural and Life Sciences, The University of Tokyo, Tokyo,apan.

by using a microflask fabricated in a microchip, rapid andsecure exchange of media or reagents is achieved by simpleoperations while making continuous microscopic measure-ments. In conventional experiments using a normal culturedish, added reagents access cells only after a long diffusiontime, and rapid chemical stimulation is difficult. On the otherhand, in the microfluidic system, it is easy to realize rapidchemical stimulation with a pressure-driven flow.

Moreover, microchip systems using cultured cells ashighly functional tools have been receiving attention recently,because of potential applications to biochemical studies,bioassay systems and bioreactors. It seems to be useful todevelop a microchip-based culture flask combined with sev-eral chemical processors, in which all procedures requiredfor biochemical experiments using a cultured cell, i.e., cellculture, cell analysis, biological reaction and measurement,can be performed. In most studies, however, long-time cul-ture has not been accomplished. It is important to develop alife support system inside the microchip for long-time cul-turing. Our microfluidic transport method can be useful tosupply oxygen and nutrients.

021-9673/$ – see front matter © 2005 Elsevier B.V. All rights reserved.oi:10.1016/j.chroma.2005.06.053

234 Y. Tanaka et al. / J. Chromatogr. A 1111 (2006) 233–237

Hepatocyte cultures are well known as an assay tool fordrug metabolism studies, biochemical syntheses, and biore-actors. Many researchers have tried to culture hepatocyteswith retention of metabolic activity. Integration of a hepato-cyte culture system into a microchip seems to be an effectiveway to retain metabolic acrivity, because microfluidic systemmimics the in vivo hepatocytes microenvironment. For thispurpose, it is important to know the optimal culture condi-tions inside the microchip.

In this study, a microchip-based system for long-termhepatocyte culturing with maintenance of the functions wasdeveloped. We have designed a new life support system usinga combination of the microfluidic system and the cell culturemicrochip. This life support system is able to supply oxygenand nutrients by the microfluidic technique. Liver-specificfunctions of hepatocytes cultured in the microchip system,such as the synthesis of albumin and urea, were then studied.

2. Experimental

2.1. Chemical and biochemical reagents

Commercial materials used in the present study wereobtained as follows. Dulbecco’s modified Eagle’s medium(DMEM) was purchased from Invitrogen Corp. (Carlsbad,CsLac

2

w[

F(h

(30 mm × 70 mm), i.e., the cover, middle, and bottom plateswith thicknesses of 170, 100 and 500 �m, respectively. Twosmall access holes for an inlet and an outlet of reagents and alarge hole for cell introduction were mechanically bored intothe cover glass plate. Microchannels and the microcultureflask were made in the middle plate using a highly focusedand intensified CO2 laser beam. The three plates were stackedand joined together by using an optical contact; that is, theplates were polished to an optical smoothness and then lam-inated together without any adhesive in an oven at 1150 ◦C.The channels were 250 �m in width and 100 �m in depth,and the dimensions of the microflask were 100 �m, 1 mm,and 1 cm in depth, width and length, respectively, i.e., 1 �Lin volume.

2.3. Cell isolation and culture

Primary parenchymal hepatocytes from 5-week-old maleWistar rats weighing about 150 g were isolated by in situcollagenase perfusion according to the method of Seglen[6]. Briefly, the liver was first washed in situ with Ca2+-free Hanks-10 mM HEPES solution and then perfusedwith 0.05 wt.% collagenase in Hanks-HEPES solution. Thedigested liver tissues were filtered through cotton gauze, then60 and 150 mesh filters. The cell suspension was centrifugedat 50 × g for 1 min at room temperature and then the super-nefiMa9

p01ndi

2

wsmIbPbSiC

oP

A), and epidermal growth factor (EGF) and fetal bovineerum (FBS) were purchased from Sigma–Aldrich Co. (St.ouis, MO). Collagenase (type 1), nicotinamide, Asc-2Pnd other reagents were obtained from Wako Pure Chemi-al Industries (Osaka, Japan).

.2. Microchip fabrication

A cell culture microchip is illustrated in Fig. 1. The chipas fabricated by the same method as previously reported

5]. The chip was composed of three quartz glass plates

ig. 1. Microchip for cell culture. (A) Illustration of the chip (not to scale).a) Reagent and cell inlet hole, (b) microculture flask, and (c) reagent outletoles. (B) Photograph of the chip.

atant was discarded. Three cycles of washing with minimumssential medium were carried out and the resulting cells werenally suspended in the culture medium described below.ore than 98% of the isolated cells were parenchymal cells

s determined by phase-contrast microscopy and more than0% were viable as measured by trypan blue dye exclusion.

The primary cultures were plated in DMEM sup-lemented with 10% FBS, 100 nmol/L dexamethasone,.5 mg/L insulin, 30 mg/L L-proline, 100 IU/mL penicillin,00 �g/mL streptomycin, 10 ng/mL EGF, 10 mmol/L nicoti-amide, 0.2 mmol/L l-ascorbic acid 2-phosphate, and 1%imethylsulfoxide [7], and incubated at 37 ◦C in a humid-fied atmosphere with 5% CO2.

.4. Assay for synthesized albumin and urea

To evaluate albumin and urea synthesis, conditioned mediaere collected daily to syringes. These supernatants were

tored at −20 ◦C until it was time for assay. Produced albu-in was determined by ELISA (sandwich method). Goat

gG fraction to rat albumin and peroxidase-conjugated rab-it IgG fraction to rat albumin were purchased from ICNharmaceuticals (Aurora, OH) for capture and labeling anti-odies, respectively. Standard rat albumin was obtained fromigma–Aldrich and dissolved in the culture medium for cal-

bration. TMB + Substrate Chromogen (DAKO, Carpinteria,A) was used as a substrate.

A precoated microtiter plate was prepared by the additionf capture antibody solution followed by three washes withBS containing surfactant, and then blocked using Block Ace

Y. Tanaka et al. / J. Chromatogr. A 1111 (2006) 233–237 235

Fig. 2. Rat hepatocytes cultured in the microchip after 4 days. Illustration is not to scale. (A) Without and (B) with medium flow at 0.5 �L/min in the withdrawnmode.

(Dainippon Pharmaceutical, Osaka, Japan). Subsequently,the plates were again washed three times. Sample solutionscontaining rat albumin were incubated with the plates for3.5 h at room temperature followed by washing, and then theplates were filled with the enzyme-labeled antibody solution.The plates were further incubated overnight at 4 ◦C. Afterwashing five times, substrates were reacted for 5 min at roomtemperature, and the resulting products were measured at630 nm with a microplate reader Multiskan Jx (Thermo Lab-systems, Vantaa, Finland).

The produced urea was determined by an enzymatic col-orimetric method. Urea N B kit (Wako Pure Chemical Indus-tries), a quantitative determination kit for urea nitrogen inserum by the urease-indophenol method, was used. A 96-

well microtiter plate was filled with 100 �L of sample perwell and urease and dye solution were added. After incuba-tion at 37 ◦C for 20 min, the plates were read using the platereader at 470 nm.

3. Results and discussion

3.1. Pre-treatment of a microchip

The microchip for the cell culture was washed with 0.1%NaOH and rinsed with pure water, followed by autoclavingat 120 ◦C for 15 min. The inside walls of the microflask werecoated with 0.1% collagen type I solution for 15 min, and then

236 Y. Tanaka et al. / J. Chromatogr. A 1111 (2006) 233–237

Fig. 3. Concept of the medium fluidic system.

rinsed with PBS. Next, medium supplemented with 10% FBSwas introduced into the treated microchip and washed awayafter 10 min incubation; this gave a fibronectin coating onthe microflask and channels. After that, cells suspended witha small amount of the fresh medium, were dropped throughthe cell introduction hole of the chip. Cells moved into themicroflask by capillary action. The chip was put into a 10 cmdish which was filled with the medium, and incubated in aCO2 incubator at 37 ◦C with 5% CO2.

3.2. Hepatocyte culture in a microchip

Primary culturing was done for 4 days in the microchip.Because metabolic rate and oxygen consumption of the hep-atocytes are very high, a shortage of nutrients and/or oxygenmay occur in the center of the microculture flask. Hepato-cytes grew well only near the inlet holes. The cultured cellsin the center of the microflask, i.e., far from the inlet holes,resulted in a significant decrease in viability (Fig. 2A).

To overcome this, it is necessary to supply fresh mediumcontinuously. By using microfluidic techniques, a continuousmedium flow system was designed and constructed (Fig. 3).This system consisted of a microchip holder, a Teflon capil-lary and a syringe. At first, flow rate was investigated from0.1 �L/min to 1 �L/min with injection or withdraw mode.FhoosisItcTBo

3

cd

Fig. 4. Liver-function test. Solid circles, medium collected from themicrochip with 1000 hepatocytes; open circles, medium collected from themicrochip without hepatocytes. (A) Urea and (B) albumin synthesis.

function. To evaluate the performance of the hepatocytes,albumin synthesis was monitored for 4 days. Fig. 4A showsthat the ability of hepatocytes to synthesize albumin increasedduring the first 3 days of culture. This result meant that thehepatocytes in the microchip with the medium fluidic sys-tem were active in albumin biosynthesis. Albumin secretedby hepatocytes was quantified and found to be 5 ng per 1000cells per 1 day. This value was comparable to that obtainedfrom hepatocytes in a primary culture in a cell culture dishand to a reported value.

Urea secretions from the hepatocytes were also exam-ined. Throughout the 4-day culture period, high levels ofurea biosynthesis were detected at relatively constant rates(Fig. 4B). The hepatocytes in the microchip with the mediumfluidic system exhibited urea secretion and maintenance offunction. The value of urea secretion depends on the value ofnitrogen source in the culture medium. However, the presentvalue was comparable to that obtained from hepatocytesin a primary culture in a cell-culture dish with the samemedium.

4. Conclusions

We developed a culture system for adult rat hepatocytesthat allowed the cells to live a long time without losing nor-mcimt

or high flow rates, cells were washed away. On the otherand, with low flow rates, cells lost their viability becausef a shortage of nutrients and oxygen. We concluded theptimum flow rate was concluded to 0.5 �L/min (data nothown). For some experiments like chemical stimulation, thenjection mode is suitable. If conditioned medium is neces-ary for further experiments, the withdrawn mode is suitable.n the injection mode, however, air bubble interference inhe microflask was found in some cases. Because air bubblesause serious damage to the cells, they should be avoided.herefore, in most cases, the withdrawn mode is preferable.y using the system, good cell growth was realized through-ut the whole microflask for at least 4 days (Fig. 2B).

.3. Assay for synthesized albumin and urea

One of the primary goals of the cell culture is to sustain theell-specific functions of cultured cells. Thus, albumin pro-uctivity was examined, as an indicator of the liver specific

al functions while inside a microchip. We designed andonstructed a medium fluidic system utilizing the microflu-dic technique. We showed that cells could live inside the

icrochip and they retained their liver-specific functions. Wehought the cell culture microchip with the medium flow sys-

Y. Tanaka et al. / J. Chromatogr. A 1111 (2006) 233–237 237

tem would also be useful for high throughput bioassays andpharmacokinetic studies.

Acknowledgements

This work was supported in part by a Grant-in-Aid forScientific Research from the Ministry of Education, Science,Sports and Culture of Japan and the Nanotechnology Projectof the Ministry of Agriculture, Forestry and Fisheries ofJapan. We gratefully acknowledge financial support from theNew Energy and Industrial Technology Development Orga-nization (NEDO) of the Ministry of Economy, Trade andIndustry, Japan.

References

[1] D.J. Harrison, A. von den Berg, Micro Total Analysis Systems ’98,Kluwer Academic Publishers, Dordrecht, 1998.

[2] D.J. Graves, Trends Biotechnol. 17 (1999) 127.[3] T. Kitamori, M. Tokeshi, A. Hibara, K. Sato, Anal. Chem. 76 (2004)

52A.[4] E. Tamaki, K. Sato, M. Tokeshi, K. Sato, M. Aihara, T. Kitamori,

Anal. Chem. 74 (2002) 1560.[5] H.M. Sorouraddin, A. Hibara, T. Kitamori, Fresenius J Anal Chem.

371 (2001) 91.[6] P.O. Seglen, Methods Cell Biol. 13 (1976) 29.[7] C. Tateno, K. Yoshizato, Am. J. Pathol. 148 (1996) 383.