Introduction

Small hepatocytes (SHs) are a subpopulation of hepatocytes that have high growth potential in culture1. Although the cells are less than half the size of mature hepatocytes (MHs), they possess hepatic characteristics1. SHs can clonally proliferate to form a colony and can differentiate into MHs by interacting with hepatic non-parenchymal cells (NPCs)2,3 or as a result of treatment with gel derived from Engelbreth-Holm-Swarm sarcoma4. Thus, we consider that SHs may be 'committed progenitor cells' that can further differentiate into MHs. The standard method to obtain an SH-rich fraction, which is a mixture of SHs and NPCs from a liver cell suspension after collagenase perfusion, has used multistep centrifugation. However, with this method, although SHs account for more than 60% of the cells at day 10, it is very difficult to inhibit the growth of NPCs.

Recently, gene expression analysis revealed that CD44 is specifically expressed in SHs but not in MHs5. The CD44 gene encodes for a family of alternatively spliced multifunctional molecules, and CD44 plays a role in adhesion of cells to an extracellular matrix such as hyaluronic acid (HA), collagen (Col) or fibronectin (FN)6. CD44 standard form (CD44s) is composed of a short cytoplasmic tail, a transmembrane region and two extracellular domains, and ten variant forms (v1–v10) exist. SHs have been shown to express CD44s and v6 (ref. 5). CD44s in SHs appears from 3 d after plating; expression increases with the expansion of the SH colony, but decreases with the maturation of SHs. The appearance of CD44v6 is delayed compared with that of CD44s. The expression also disappears with maturation of SHs. Although CD44 is expressed in cultured SHs, no CD44+ hepatocytes are found in the normal rat liver. When the rat liver is severely injured by hepatotoxins such as galactosamine and 2-acetylaminofluorene, CD44+ hepatocytes transiently appear in the periportal regions of the liver lobules. Using an anti-CD44 Ab, we have isolated CD44+ cells from the galactosamine-treated rat liver5. These CD44+ cells possess the characteristics of SHs. However, as we were not able to isolate CD44+ SHs from either a normal adult liver or from a liver with two-thirds removed, we tested HA as a ligand for separating a population of SHs.

HA, a linear polymer of (1-β-4)-D-glucuronic acid (1-β-3)-N-acetyl-D-glucosamine, is a large glycosaminoglycan that can reach a molecular size of 107 Da. It is found in the tissue matrix and body fluids of all vertebrates and has diverse biological roles. These include acting as a vital structural component of connective tissues and playing roles in the formation of loose hydrated matrices that allow cells to divide and migrate, immune cell adhesion and activation, and intracellular signaling7. Such diversity results from the large number of hyaluronan-binding proteins (termed hyaladherins), which exhibit significant differences in their tissue expression, cellular localization, specificity, affinity and regulation. Three HA synthase (HAS) genes, coding for HAS-1, 2 and 3, are recognized to synthesize HA8. HAS is located at the inner cell membrane, where the newly synthesized HA is extruded into the extracellular space9. Synthesized HA is degraded locally in the tissues where it is produced or by the lymph nodes, and the reminder enters the bloodstream10. More than 90% of circulating HA is degraded by hepatic sinusoidal endothelial cells (SECs) through a receptor recycling pathway. Hyaladherins, LYVE-1 and stabilin-1 and 2, but not CD44, are expressed in SEC11,12. Furthermore, a relationship between serum HA levels and liver diseases has been reported13. HA may be related to the induction of SHs in the liver.

We found that SHs cultured on HA-coated dishes could selectively proliferate to form colonies and that the contamination by NPCs with this method was much less than with our previous method. Although we do not know in detail why a population of SHs can be isolated from a normal liver and selectively proliferate on HA-coated dishes, the HA–CD44 interaction may enhance the growth of SHs. In addition, the combination of DMEM/F12 medium and HA-coated dishes allows us to exclude FBS from the culture. Using this protocol, we have isolated human SHs from a normal adult liver. Taking into consideration the application of hepatic stem/progenitor cells to regenerative medicine, the use of proteins derived from an animal, particularly FBS, should be avoided in the culture of the cells. This protocol for isolating and culturing SHs may help researchers in this field to progress in their own investigations.

Materials

Reagents

  • Male F344 rats (Sankyo Lab Service, Tokyo, Japan) weighing 150–200 g

    Caution

    All animal experiments must comply with national and institutional regulations.

  • Ascorbic acid-2 phosphate (Asc2P; Wako Pure Chemical Industries, Osaka, Japan, cat. no. 013-12061) (see REAGENT SETUP)

  • BSA (30% solution;Serological Proteins, IL, cat. no. 82-046-3)

  • Collagenase (Wako Pure Chemical Industries, cat. no. 034-10533; Yakult Pharmaceutical Industry, cat. no. YK-101; Sigma, St. Louis, MO, cat. no. C5138)

  • Dexamethasone (Wako Pure Chemical Industries, cat. no. 041-18861) (see REAGENT SETUP)

  • DMEM/nutrient mixture Ham F-12 (DMEM/F12) (Sigma, cat. no. D8900)

  • Nicotinamide-supplemented medium1 (to grow MHs)

  • L15 medium15 supplemented with a growth factor (to grow MHs)

  • Epidermal growth factor (EGF; BD Biosciences, Bedford, MA, cat. no. 354001) (see REAGENT SETUP)

  • EGTA (Sigma, cat. no. E-0396)

  • Gentamicin solution (50 mg ml−1 ; Sigma, cat. no. G1397)

  • Culture medium stock (see REAGENT SETUP)

  • HA derived from human umbilical cords (Biozyme Laboratories, cat. no. HA1NaL), bovine vitreous humor (Sigma, cat no. H7630), pig skin (Seikagaku Kogyo, cat. no. 400720), rooster comb (Sigma, cat. no. H5388) and Streptococcus (Sigma-Aldrich, cat. no. 53747) (see REAGENT SETUP)

  • HANKS' balanced salt solution (HANKS; Sigma, cat. no. H9269)

  • 10 × Ca2+, Mg2+-free HANKS (Sigma, cat. no. H4641)

  • Wash solution (see REAGENT SETUP)

  • MH wash solution (see REAGENT SETUP)

  • Phenol red–free HANKS

  • HEPES (Dojindo, Kumamoto, Japan, cat. no. 342-01375)

  • Insulin (Sigma, cat. no. I-5500) (see REAGENT SETUP)

  • Insulin–transferrin–selenium (ITS; GIBCO, cat. no. 0459)

  • NaHCO3 (Kanto Chemical, cat. no. 37116-00)

  • Nembutal, 50 mg ml−1 (Dainippon Pharmaceutical, Tokyo, Japan, cat. no. 132141)

  • Nicotinamide (Sigma, cat. no. N3376) (see REAGENT SETUP)

  • Penicillin–streptomycin solution (Sigma, cat. no. P-4333)

  • Percoll PLUS (GE Healthcare Bio-Sciences, Piscataway, NJ, cat. no. 17-5445-01) (see REAGENT SETUP)

  • L-Proline (Sigma, cat no. P5607)

  • Trypan blue (Chroma Technology, VT, cat. no. 1B187) (see REAGENT SETUP)

  • Pre-perfusion solution (see REAGENT SETUP)

Equipment

  • Dishes, 100, 60 and 35 mm (Corning Glass Works, Corning, NY)

  • Non-charged (NC) 60-mm dish (Kord-Valmark, Ontario, Canada, cat. no. 2901)

  • Autoclaved 250-μm nylon filter net (Nippon Rikagaku Kikai, Tokyo, Japan)

  • Cell strainer, 70- μm filter (BD Falcon, cat. no. REF352350)

  • 0.2- μm filter (Mediakap-2; Spectrum Laboratories, CA, cat. no. MEM2M-02B-12S)

  • Paper filter no. 3 (Advantec, Tokyo, Japan)

  • Sterilized 10-cm Petri dish (glass or plastic)

  • Water bath (Teitec Co., Tokyo, Japan, cat. no. EX-B2015250)

  • Vascular clamp, bulldog type (Fine Science Tools, Foster City, CA, cat. no. 18050-35)

  • Peristatic pump (Tokyo Rika Instruments, Tokyo, Japan, RP-1000)

  • Silicon tube (φ4.76 × 7.94 mm2)

  • Butterfly needle, 18 gauge (Terumo, Osaka, Japan, cat. no. SV-18CLK)

  • O2 gas (95% O2/5% CO2)

  • Neubauer improved hemocytometer (Sigma, cat. no. Z359629)

Reagent setup

  • HA solution Measure flakes or powder of HA; UV irradiate HA on a plastic tray for 1 h. After irradiation, place HA into a 50-ml tube and adjust the concentration of HA stock solution to 10 mg ml−1 by adding sterilized PBS. For UV irradiation, use a standard clean bench or safety cabinet equipped with a UV lamp.

    Caution

    UV is harmful to skin and eyes.

  • 1,000 × insulin stock solution (500 μg ml−1) Add 100 mg insulin to 100 ml ddH2O and then add 1.2 ml 1 N HCl. Adjust to 200 ml with ddH2O and filter with a 0.2-μm filter.

    Critical

    Insulin dissolves in acidic solution.

  • Pre-perfusion solution Add approximately 850 ml ddH2O into a 1,000-ml graduated cylinder; add 100 ml 10 × Ca2+, Mg2+-free HANKS, 190 mg EGTA, 1 ml 1,000 × insulin stock solution and stir and adjust pH to 7.5 with 7 ml 1 M NaHCO3. Adjust to 1,000 ml and filter with a 0.2-μm filter. Distribute into each bottle (150 ml). Store at 4 °C until use.

    Critical

    EGTA solution should be warmed to 37 °C before use.

  • Perfusion solution To 200 ml HANKS, add 1 ml 1,000 × insulin stock solution and collagenase (100 U ml−1). Shake gently.

    Critical

    Prepare HANKS with insulin before the experiment. Add collagenase to pre-warmed perfusion solution just before use and immediately dissolve by gentle shaking.

  • Wash solution To 500 ml HANKS, add 0.5 ml 1,000 × insulin stock solution, 2 ml penicillin–streptomycin solution and 0.5 ml gentamicin solution.

  • MH wash solution To 500 ml HANKS, add 0.5 ml 1,000 × insulin stock solution, 3.3 ml BSA, 2 ml penicillin–streptomycin solution and 0.5 ml gentamicin solution.

  • Percoll solution14 Add 2.4 ml 10 × HANKS and 21.6 ml Percoll into a 50-ml conical tube. Mix gently upside down several times. Store at 4 °C until use.

  • Trypan blue stock solution (0.1%) Add 100 mg trypan blue to 100 ml phenol red–free HANKS. Filter with a paper filter.

  • Dexamethasone stock solution Add 39.2 mg dexamethasone to an autoclaved brown bottle. Add 10 ml ethanol (10−2 M stock solution); dilute to 100 × (10−4 M) with autoclaved ddH2O. Store at 4 °C until use.

    Critical

    10−4 M stock solution must be used within 3 months.

  • EGF stock solution (10 μg ml−1) Add 10 ml autoclaved ddH2O to an EGF vial. Store 1-ml aliquots in microcentrifuge tubes at −20 °C until use.

  • Nicotinamide stock solution (1M) Add 12.21 g nicotinamide to 100 ml PBS. Filter with a 0.2-μm filter and store at 4 °C until use.

  • Asc2P stock solution (100 mM) Add 2.90 g Asc2P to 100 ml PBS, filter with a 0.22-μm filter and store at 4 °C in a 100-ml brown bottle until use.

  • Culture medium stock Add the following reagents to a 1,000-ml beaker: DMEM/F12 (15.56 g), HEPES (1.20 g), L-Proline (30 mg), penicillin–streptomycin (8 ml), ddH2O up to 1,000 ml. Mix using a magnetic stir bar, add 2.20 g NaHCO3, adjust pH to 7.6 with 1 N NaOH and filter with a 0.2-μm filter. Store at 4 °C until use.

  • Preparation of culture medium Mix DMEM/F12 stock medium (500 ml), BSA (1.67 ml), nicotinamide stock solution (5.50 ml), Asc2P stock solution (5 ml), ITS (5 ml), EGF stock solution (0.5 ml), dexamethasone stock solution (0.5 ml) and gentamicin (0.5 ml).

Equipment setup

  • Perfusion system See Figure 1a. The system is composed of a water bath, vascular clamp, peristatic pump, silicon tube (φ4.76 × 7.94 mm2) and 18-gauge butterfly needle.

    Figure 1: Experimental setup.
    figure 1

    (a) Setting up the perfusion apparatus. Pre-perfusion solution is warmed to approximately 39 °C and bubbled with O2. (b) Position of the inserted needle in the portal vein of a rat.

    Full size image

Procedure

Preparation of HA-coated dishes

Timing At least 1 d before the experiment

  1. 1

    Dilute the stock HA solution to 100 μg cm−2 with PBS, fill Petri dishes and incubate at 37 °C overnight. The next day, discard the solution and wash once with PBS. Aspirate PBS and leave to dry on a clean bench with UV irradiation for 30 min.

    Caution

    UV is harmful to skin and eyes.

    Troubleshooting

Isolation of liver cells

Timing 1–1.5 h

  1. 2

    Settle the perfusion apparatus in the warmed water bath (Fig. 1a). Pour the pre-perfusion solution into the apparatus before the experiment and bubble it with 95% O2/5% CO2 gas at a flow rate of 0.5 l min−1.

    Critical Step

    The equipment should not be left for more than 30 min before proceeding to the next step.

    Troubleshooting

  2. 3

    After light anesthesia by ether, anesthetize a rat with an i.p. injection of nembutal (5 mg per 0.1 ml per 100 g body weight).

    Troubleshooting

  3. 4

    Cut the abdominal wall using surgical scissors and open the abdominal cavity to look at the portal vein.

  4. 5

    Ligate a common bile duct and splenic vein together using a surgical thread at the portion nearest the portal vein (Fig. 1b).

    Critical Step

    This is to avoid loss of the solution.

    Troubleshooting

  5. 6

    Insert a butterfly needle filled with pre-perfusion solution into the portal vein 1.5–2.0 cm from the bifurcation of the portal vein, stop the tip of the needle at a position close to the bifurcation of the portal vein and clamp the needle with a surgical clip (Fig. 1b).

    Troubleshooting

  6. 7

    Start the perfusion at a flow rate of 30 ml min−1.

  7. 8

    Cut the inferior vena cava and heart as soon as the flow starts.

    Critical Step

    Cut the inferior vena cava at the portion beneath the right kidney and the thoracic cavity, and then cut the heart to flow the perfusate out of the cadaver. Washing out the blood completely from the liver and preventing an increase of intra-hepatic pressure is important to the success of the preparation.

    Troubleshooting

  8. 9

    When the amount of the pre-perfusion solution becomes small, add collagenase to the perfusion solution and then pour it into the perfusion apparatus.

    Critical Step

    To avoid a decrease in collagenase activity, dissolve it after the pre-perfusion solution flows. Do not mix vigorously.

    Troubleshooting

  9. 10

    Flow the solution at a flow rate 15–20 ml min−1.

    Critical Step

    The flow rate should be decided by rat body weight.

    Troubleshooting

  10. 11

    Stop the flow before air bubbles move into the liver when the solution flows out from the reservoir.

    Troubleshooting

  11. 12

    Cut the liver from the abdominal cavity and transfer it to a sterilized Petri dish.

  12. 13

    Prepare a 100-ml beaker with 70–80 ml wash solution and add a small amount of the wash solution to the Petri dish.

    Critical Step

    From this step onward, all procedures should be done in sterilized conditions.

  13. 14

    Peel the hepatic capsule as carefully as possible and, to drop the digested cells, shake the liver into the beaker.

    Troubleshooting

  14. 15

    Filter the cell suspension through a 250-μm nylon filter net into a new 100-ml beaker.

  15. 16

    Filter the cell suspension through a 70-μm filter into four 50-ml conical tubes using a 25-ml pipette and then adjust each tube to an equal volume with the wash solution (approximately 40 ml).

  16. 17

    Centrifuge the tubes at 50g for 1 min at 4 °C.

    Troubleshooting

  17. 18

    Collect supernatants and transfer to new conical tubes. Repeat Steps 15–17 a total of 3 times.

    Critical Step

    Many SHs are included in the supernatant after low-speed centrifugation. This procedure is carried out to remove the majority of MHs from the cell suspension.

    Pause point

    Cells in tubes may be kept on ice for 1–3 h. Avoid long-term preservation.

  18. 19

    If you want to isolate MHs, proceed directly to Step 19B. To isolate SHs, proceed with Step 19A.

    1. A

      Isolation of SHs

      1. i

        Centrifuge the supernatant at 50g for 5 min at 4 °C.

        Critical Step

        This procedure is used to remove hematopoietic cells.

        Troubleshooting

      2. ii

        Discard the supernatant and add 40 ml wash solution to the tubes.

        Troubleshooting

      3. iii

        Centrifuge at 50g for 5 min at 4 °C after gentle pipetting to dissociate the cell pellet.

      4. iv

        Add 40 ml wash solution and then centrifuge at 150g for 5 min at 4 °C.

        Critical Step

        This procedure is carried out to damage some MHs contained in the suspension.

        Troubleshooting

      5. v

        Discard the pellet, add 40 ml wash solution and then centrifuge at 150g for 5 min at 4 °C.

      6. vi

        Discard the supernatant, pour 20 ml culture medium into each tube and gather the suspension into two 50-ml conical tubes.

      7. vii

        Centrifuge the suspension at 50g for 5 min at 4 °C.

        Troubleshooting

      8. viii

        Add 10 ml culture medium to each tube and gather the suspension into one tube.

      9. ix

        Add 0.5 ml cell suspension to 1.5 ml trypan blue solution and pipette gently.

        Critical Step

        Tubes should be on ice.

      10. x

        Count the viable cells as soon as trypan blue is added. For counting, use a Neubauer improved hemocytometer. Count the number of cells with trypan blue–negative nuclei which exist in nine masses. Use this formula to obtain the cell number: the number of viable cells (cells ml−1) = [(number of cells inside nine masses)/9] × 4 × 104.

        Critical Step

        Cells with trypan blue–positive nuclei are dead. Count the number of cells that are smaller than typical MHs and larger than NPCs (15–20 μm in diameter). The overall viability of cells will not be good, but most small cells, including SHs, are viable. As SH colony formation depends on the cell density, the number of viable cells, which have the potential to attach to the dishes, is important in this step.

        Pause point

        Cells in tubes may be kept on ice for 1–3 h. Avoid long-term preservation.

      11. xi

        Adjust the concentration of the cells to 1 × 105 cells ml−1 and then seed them on HA-coated dishes.

        Troubleshooting

      12. xii

        Place dishes in a 5% CO2/95% air incubator at 37 °C.

      13. xiii

        Replace the medium with fresh medium after 3 h.

        Critical Step

        This procedure excludes unattached cells.

      14. xiv

        Renew medium every other day.

        Critical Step

        Two times a week is enough when the number of attached cells is small.

      Timing 1–1.5 h

    2. B

      Isolation of MHs

      1. i

        Add 20 ml MH wash solution to a pellet in each tube (from Step 17) and gather the suspension into two 50-ml conical tubes.

      2. ii

        Centrifuge the tubes at 50g for 1 min at 4 °C after gentle pipetting to dissociate the cell pellet.

        Critical Step

        As MHs are more labile to a shock-like pipetting or centrifugation than SHs, gentle pipetting is important.

        Troubleshooting

      3. iii

        Discard the supernatant and add 40 ml MH wash solution to each tube.

      4. iv

        Centrifuge at 50g for 1 min at 4 °C.

        Troubleshooting

      5. v

        Discard the supernatant and add 25 ml MH wash solution to each tube.

      6. vi

        Pipette gently to dissociate the cell pellet.

      7. vii

        Pour 25 ml cell suspension into Percoll solution and gently mix (upside down several times).

      8. viii

        Centrifuge at 50g for 15 min at 4 °C.

        Critical Step

        This step is performed to remove dead MHs. After centrifugation, dead MHs and NPCs float on the solution.

        Troubleshooting

      9. ix

        Discard the supernatant and add 40 ml MH wash solution to each tube.

      10. x

        Centrifuge the tubes at 50g for 1 min at 4 °C after gentle pipetting to dissociate the cell pellet.

        Troubleshooting

      11. xi

        Discard the supernatant and add 40 ml culture medium to each tube.

      12. xii

        Centrifuge the tubes at 50g for 1 min at 4 °C after gentle pipetting to dissociate the cell pellet.

        Troubleshooting

      13. xiii

        Add 20 ml culture medium to each pellet and gather the suspension into one tube.

      14. xiv

        Add 0.5 ml cell suspension to 1.5 ml trypan blue solution and suspend.

      15. xv

        Count the number of viable and dead MHs as described in Step 19A(x).

        Critical Step

        Keep cells in tubes on ice.

      16. xvi

        Adjust cell density to 2–6 × 105 viable cells ml−1 and then seed the cells on dishes.

        Troubleshooting

      17. xvii

        Place the dishes in a 5% CO2/95% air-incubator at 37 °C.

      18. xviii

        1–2 h later, replace the medium with fresh medium.

        Critical Step

        If researchers need to grow MHs, we suggest using nicotinamide-supplemented medium1 or L15 medium15 supplemented with a growth factor.

      Timing 1–1.5 h

Troubleshooting

Troubleshooting advice can be found in Table 1.

Table 1 Troubleshooting table.
Full size table

Timing

Steps 2–18, isolation of liver cells: 1–1.5 h

Steps 19A(i–xiv), isolation of SHs: 1–1.5 h

Steps 19B(i–xviii), isolation of MHs: 1–1.5 h

Anticipated results

A critical component of this method is to use HA-coated dishes and serum-free medium for SH culture. As CD44 is not expressed in MHs but is expressed in SHs, we first examined whether SHs could selectively proliferate on HA-coated dishes under our standard culture conditions (DMEM + 10% FBS). As shown in Figure 2a, compared with the NC-, non-, rat tail Col- and FN-coated dishes, SHs could selectively proliferate on HA-coated dishes. However, there was much less growth of NPCs in HA-coated dishes than in other dishes. The average size of SH colonies on HA-coated dishes was larger than that of SH colonies on NC-, Col- and FN-coated dishes (Fig. 2b). Although the SH growth on non-coated dishes was as good as that on HA-coated dishes, NPCs also grew (Table 2).

Figure 2: Growth of liver cells on dishes coated with extracellular matrices.
figure 2

(a) Phase-contrast micrograph of small hepatocytes (SHs) and non-parenchymal cells at days 7, 14 and 21 after plating. Cells are cultured in DMEM supplemented with 10% FBS. Arrows show SH colonies and asterisks show NPCs. The expansion of SH colonies is clear and there is little contamination by NPCs on the hyaluronic acid (HA)–coated dish compared with other dishes. At day 21 on HA-coated dishes, proliferated cells, which are marked by arrows, are all SHs. The magnification of the photos is the same; scale bar = 140 μm. (b) The growth of SH colonies on various dishes. Columns and bars show the average and s.d. of three dishes, respectively.

Full size image
Table 2 Purity of small hepatocytes cultured in serum-free medium and on hyaluronic acid–coated dishes.
Full size table

Next, we examined whether SHs could proliferate in serum-free medium. Although we reported that SH colonies could expand in serum-free medium after re-plating SH colonies4,16, it was very difficult for SHs in serum-free medium to grow to form colonies on conventional dishes. When we used serum-free DMEM/F12 and HA-coated dishes, many colonies developed (Fig. 3a), whereas few colonies grew in serum-free DMEM (Fig. 3b). We usually use HA derived from human umbilical cords. However, the effect of HA is not different among commercially available forms of HA derived from pig skin, bovine vitreous humor, rooster comb and Streptococcus (data not shown). To observe the appearance of SHs in serum-free culture, both nicotinamide and growth factors (either a single factor or a combination of EGF, hepatocyte growth factor and transforming growth factor-α) must be included in the medium1. These results suggest that ingredients contained in Ham F12 medium are important for SHs to grow in the serum-free culture, because the use of DMEM alone without FBS cannot support the growth of SHs. However, when the specific agents included in Ham F12 medium were added to DMEM and SHs were cultured in the serum-free medium, the growth of SHs was neither enhanced nor inhibited (data not shown). Therefore, we now hypothesize that the balance of ingredients in Ham F12 medium may be important for SHs to grow in serum-free culture. On the other hand, adding transferrin and selenium is different from the original recipe for culturing rat SHs2. Although SHs appear and grow without those agents, both are necessary to maintain their growth in serum-free culture. Without transferrin and selenium, the growth of SHs becomes worse after 1–2 weeks of culture (current data not shown; refer to ref. 1). Figure 3a shows a typical SH colony in serum-free culture. Although it was difficult to avoid MH survival in DMEM with FBS, the use of HA-coated dishes and serum-free medium suppressed MH survival. Most attached MHs died within 1 week (Fig. 3c). In this protocol, the purity of SHs was 84.5% ± 3.2% at day 10 (Table 2). Some SECs (SE1+), stellate cells (desmin+), liver epithelial cells (vimentin+) and Kupffer cells (ED1/2+) were observed.

Figure 3: The growth of small hepatocyte (SH) colonies on hyaluronic acid (HA)–coated dishes cultured in serum-free medium.
figure 3

(a) Phase-contrast micrograph of a typical SH colony at day 7. Scale bar = 100 μm. (b) The number of SH colonies on HA-coated dishes cultured in serum-free DMEM or DMEM/F12 medium at day 10. Many colonies grew in the serum-free DMEM/F12 medium, whereas there were few in the serum-free DMEM. Columns and bars show the average of the total number of colonies per 60-mm dish and s.d. of three dishes, respectively. (c) Few mature hepatocytes (MHs) can survive in the serum-free medium on HA-coated dishes. The growth of MHs and SHs is shown as the ratio day 7/day 1.

Full size image

Characterization of SHs on HA was carried out using immunocytochemistry and RT-PCR. All SHs were immunocytochemically stained with anti-CD44 (Fig. 4a) and anti-CK8 Abs (Fig. 4b). As shown in Figure 4c, RT-PCR shows that SHs express mRNAs of hepatic marker genes such as albumin, transferrin, tyrosine aminotransferase, hepatocyte nuclear factor-4α and CK8 as plentifully as MHs at day 10. In addition, CD44 is not expressed in MHs but is expressed in SHs. These results reveal that cell colonies grown in serum-free DMEM/F12 on HA have quite similar characteristics to SHs observed using our standard culture method (DMEM/FBS). Until now, we have not observed any difference between the cells.

Figure 4: Characteristics of small hepatocyte (SH) colonies on hyaluronic acid (HA)–coated dishes.
figure 4

Immunocytochemistry for (a) CD44 and (b) CK8 at day 10. Scale bars = 100 μm. (c) Expression of mRNAs of hepatic marker genes. SHs were cultured in serum-free medium on HA-coated dishes at day 7. RNA of mature hepatocytes (MHs) is derived from isolated MHs. SHs express many hepatic marker genes. CD44 is expressed in SHs but not MHs. These are representative data.

Full size image

The SH is a hepatocyte progenitor cell that is committed to differentiate into an MH. Manipulation of the growth and maturation of SHs is easy, and they can be cryopreserved17. Cryopreserved SHs can maintain the abilities of growth and maturation18. Maturated SHs in serum-free culture can be used for studies of hepatic drug metabolism, liver diseases and regenerative medicine.