CRYORIOI.OGY

29, 4X-469

Thawed

(1992)

Human

Hepatocytes

in Primary Culture

M. DOU,* G. DE SOUSA,* B. LACARELLE,* M. PLACIDI,* P. LECHENE DE LA PORTE,7 M. DOMINGO,t H. LAFONT,? AND R. RAHMANI**’ *INSERM

U278 FacultP de Pharmacie, 27, Bd Jean Moulin 13385 Marseille U130 Ave Mozart 13008 Marseille, France

cedex 5, France,

TINSERM

In drug metabolism studies, isolated and cultured human hepatocytes provide a useful model for overcoming the difficulty of extrapolating from animal data. In vitro studies with human hepatocytes are scarce because of the lack of livers and suitable methods of storage. After developing a new method for cryopreservation of human hepatocytes, we evaluated the effects of deep freezing storage on their viability, morphology, and functional and toxicological capabilities in classical culture conditions. Freshly isolated human hepatocytes were cryopreserved in medium containing 10% Me,SO and 20% fetal calf serum, using a Nicool ST20 programmable freezer (- I.QC/min for 18 min and - 30”Umin for 4 min). Cells were stored in liquid nitrogen. Viability of thawed human hepatocytes was 50-65% as assessed by erythrosin exclusion test prior to purification on a Percoll density gradient. Morphological criteria showed that thawed human hepatocytes require an adaptation period to the medium after seeding. Functional assessments showed that human hepatocytes which survive freezing and thawing preserve their protein synthesis capabilities and are able to secrete a specific protein, anionic peptidic fraction, which is involved in the hepatic uptake of bile-destined cholesterol. We then studied Midazolam biotransformation to test metabolic functions, and erythromycin toxicity by Neutral Red test (cell viability) and 3-(4,5-dimethylthiazol-2-yl)-diphenyl tetrazolium bromide test (cell metabolism). All of these experiments indicated that thawed human hepatocytes should be used 38 h after seeding for optimum recovery of their functions: membrane integrity, protein synthesis, and stabilization of drug metabolism enzymes. o 1992 Academic PXSS. hc.

Hepatocytes isolated from various species have been shown to represent a suitable in vitro model for assessment of drug transport and metabolic and toxic events at the hepatic level (1, 32). The major inconvenience of this model is that quantitative and qualitative interspecies variabilities are likely to occur in these various processes. The use of human hepatic cells constitutes an interesting alternative for overcoming the difficulty in extrapolating from animal data (6, 19, 33, 37). We recently developed an efftcient procedure for isolating several billion viable hepatocytes from a whole human liver (12). As a single liver perfusion

Received October 24, 1991; accepted March 15, 1992. i To whom reprint requests should be addressed.

results in a large number of cells, and as human livers are scarce, we established a bank of cryopreserved hepatocytes according to a novel methodology. We first compared the functional capabilities of cultured human hepatocytes after and prior to their cryopreservation in liquid nitrogen. To validate the model with reference to fresh cells, we then investigated (i) the morphological aspects of the cells by optical and transmission electronic microscopy, (ii) their capacity to synthesize and release proteins by secretory processes, e.g., anionic peptidic fraction (APF) a specific protein associated with lipids in bile (9), (iii) their reactivity toward erythromytin, a macrolide used as reference in a French multicenter study on acute hepatotoxicity (13), and, finally (iv) their ability to biotransform Midazolam (MDZ) (14, 17), a

454 001l-2240/92 $5.00 Copyright All rights

0 ,992 by Academic Press, Inc. of reproduction in any form reserved.

THAWED

HUMAN HEPATOCYTES

455

dium and filtered through a 60-pm nylon mesh. Three washes in L15 Leibovitz medium supplemented with 10% new born calf MATERIALS AND METHODS serum (NBCS) removed debris and nonparenchymental and damaged cells (3). The Drugs and Chemicals hepatocytes in the final pellet were immeMidazolam (MDZ), 14CMDZ (76.9 &ii diately cryopreserved or used for primary mg) and its metabolites, l-OH-MDZ, 4-OH- culture. MDZ, and di-OH-MDZ and N-DemethylThe viability of isolated human hepatoMDZ were kindly supplied by Hoffman La cytes, as estimated by the erythrosin B exRoche Laboratories (Basel, Switzerland). clusion test, ranged between 70 and 80%. All products for cell culture were The quantity was 10 to 15 billion cells. purchased from Sigma Chemicals Co. The Cryopreservation. Human hepatocytes other chemicals were obtained from com- were resuspended in 4°C HAM F12 mercial sources and were of analytical or COON’S modification medium containing HPLC grade. 2% polyvinyl pyrrolydone (PVP), 2.5% bovine serum albumin Fraction V (BSA), and Obtaining Human Hepatocytes and Use 20% fetal calf serum (FCS), and 10% diZsolation. Human hepatocytes were ob- methyl sulfoxide (M,SO) was then added tained, as previously described (12), from slowly as cryoprotectant. Aliquots of the two whole livers taken under strict ethical hepatocytes were then drawn in sterile conditions from multiorgan (heart, kidney, polypropylene vials (1.8 ml tubes, NUNC and lung) donors, referred to as HL37 (age: Ltd.) at a cell density of about 20 x lo6 32 years; sex: female; death by suicide) and cells/ml. This time lag never exceeded 30 HL38 (age: 40 years; sex: female, death by min. The frozen cycle was controlled by a traffic road accident), by a two step colla- NICOOL ST20 (Air Liquide, France) and genase perfusion technique. No specific the optimum gradient temperature was drug history was reported for any of these - 1.9’Wmin from 4 to -3o”C, and -3O”CI patients. Briefly, first of all, to improve the min from - 30 to - 150°C. Cryopreserved elimination of blood, the liver was exten- hepatocytes were then stored at - 196°C in liquid nitrogen. sively washed in situ with 4°C Eurocollins Thawing. The vials were removed from medium through the portal vein and the abdominal aorta and then flushed with Belzer the liquid nitrogen storage container and liquid. The organ was then transported in rapidly thawed by immersion in a 37°C waice to the laboratory; this time lag never ter bath. Percoll colloidal silica solution exceeded 30 min. The organ was placed in a was used for purification of viable cell acperfusion apparatus and perfused with a cording to their density. First, the hepatocalcium-free buffer (buffer A: NaHCO, cytes were added to a Percoll II solution 0.187%, NaCl 0.83%, KC1 0.05%, glucose (Percoll 37%, Hanks’ buffer salt solution O.l%, 4-(2-Hydroxyethyl)-piperazine 10X (HBSS) 3%, L15 medium 40%, FCS ethane sulfonic acid (Hepes) 0.24%, pH 20%) (v/v; 25/75) and then layered on a 7.4), then with a 37°C collagenase solution Percoll I solution (Percoll63%, HBSS 10x (0.05% in buffer A containing 0.2% CaCl,), 7%, L15 lo%, FCS 20%). After 7 min of under recirculation and continuous oxygen- centrifugation at 4OOg,viable cells were reation. After 20 min, the surrounding Glis- moved from the interface of the centrifuge son’s capsule was disrupted and hepato- gradient and washed twice at 50g in L15 cytes were resuspended in a washing me- medium containing 20% NBCS. benzodiazepine implicated in phase I and phase II metabolic pathways (11, 12).

456

DOU ET AL.

Primary Culture. Fresh or thawed hepatocytes were resuspended in medium I (HAM F12 COON’S medium/Dubelcco’s modified eagle medium (V/V), pH 7.4), containing 10% FCS and supplemented with nutrients (sodium selenite 0.05 l&ml, human transferrin 10 pg/ml, ethanolamine 1 &ml, insulin 0.1 UI/ml, glucagon 2 &ml, thyroxine 0.02 p,g/ml, penicillin 50 UI/ml, netilmicin 100 pg/ml). For metabolic and biochemical studies, cells were seeded in 35mm diameter petri dishes (0.8 x lo6 cells/l 5 ml of medium I) and for toxicological studies, in 96-well microtiter plates (0.3 x lo5 cells/100 t.~lof medium I/well). Cells were then incubated at 37°C under a humidified 5% CO, atmosphere. For thawed hepatocyte culture, dishes and plates were previously coated with type 1 rat tail collagen (10 kg/cm* in 35-mm dishes, and 6.2 kg/cm* for multiwell plates). Medium I was removed after 10-20 h of adhesion and replaced by the same medium without FCS but containing lop6 M dexamethasone (medium II). The medium was renewed daily. Characterization of Human Hepatocytes Morphology. After washing with phosphate buffer saline solution (PBS), cells were fixed up to 6 h at 4°C in 2% glutaraldehyde in 0.1 M cacodylate buffer or PBS, pH 7.4. They were rinsed in the same buffer as fixative and postfixed in 1% 0~0, in PBS. Fixed cells were dehydrated through ethanol and then embedded in Epon 812. Ultrathin sections collected on copper grids were counterstained with uranyl acetate and lead citrate. Biochemical studies. For protein synthesis studies, 0.1 &i/ml of L-[U-‘~C] leucine (sp act: 312 mCi/mmol) was incubated for 24 h in medium II with cell monolayers. (a) Total radiolabeled protein synthesis was assessed by measuring incorporated 14C Leu in intracellular (cells on dishes scraped with PBS) and extracellular proteins separately. Proteins were precipitated

in 10% perchloric acid (PCA) and washed three times in 2% PCA. The final pellet was dissolved in NaOH 0.3 N, neutralized with 35% HCl, and counted by liquid scintillation. The amount of synthetized proteins was expressed in nanomoles per milligram of total proteins as determined by the method of Bradford (5). (b) The secretion capability of the hepatocytes were assessed by measuring the radiolabeled APF contained in a 200~cl.1sample of supernatant, as described elsewhere (9). MDZ biotransformation. For metabolic studies, experiments were initiated by addition of i4C MDZ (0.45 lKi/ml) to achieve a final concentration of 50 FM, as previously described (41). After a 24-h exposure period of MDZ with cell monolayer, the extracellular medium was analyzed by highperformance liquid chromatography (HPLC) using a Hewlett-Packard 1090 chromatograph equipped with an automatic injector and a cooled sampler (11, 12). The radiolabeled compounds were detected by continuous flow liquid scintillation (FLOW ONE Radiomatic Instrument). Intracellular metabolites were quantified after cell scraping with water/methanol (v/v; l/l) and further centrifugation (1OOOOg) to remove proteins. The supernatant was then analyzed by HPLC. Aliquots of extra- and intracellular media was also counted by liquid scintillation (Beckman) to determine the respective percentage of radioactivity in each compartment. The amounts of unchanged MDZ and metabolites were expressed in nanomoles of radiolabeled compound per milligram of total protein. Toxicologic studies. Cells cultured in microtiter plates were exposed to various concentrations of erythromycin (from 0.1 to 2 mg/ml in medium II, against a blank without drug) for 20 h (n = 5 for each concentration). The cytotoxic effects of erythromycin, evaluated as loss of cell metabolic

THAWED

457

HUMAN HEPATOCYTES

functionality after the treatment period, was determined by Neutral Red (NR) test (4) and 3-(4,5-dimethylthiazol-2-yl)diphenyl tetrazolium bromide (MTT) test (28), as previously described (40). The plates were read on a Dynatech MR700 Microelisa reader with a wavelength of 550 nm. Cytotoxicity data were standardized by expressing the absorbance values in the presence of the toxic substance as a percentage of control cell viability. The mid point toxicity, IC50, defined as the concentration of compound which decreased by 50% the highest value obtained with each endpoint, was then determined by interpolation. Statistical analysis. Experiments were performed on two different freshly isolated and cryopreserved human hepatocyte batches, at 14, 38, and 62 h after seeding in order to evaluate the importance of aging of the cells. In all assays, data were computerized and normal distribution was confirmed for each set of data. Results after thawing were compared with those before cryopreservation by Student’s t test. The significant threshold was fixed at 0.05.

RESULTS

Viability Viability was determined using erythrosin B exclusion test, both before and after cryopreservation. Results (Table 1) demonstrate the effects of deep freezing storage on the viability of human hepatocytes. Frozen cells (20 x lo6 cells/ml) were stored for 15-150 days prior to experiments. After thawing, the viability of hepatocytes (55%, i.e., 11 x IO6 viable cells/ml) was 25 to 35% lower than the values obtained for freshly isolated hepatocytes (80%, i.e., 16 x lo6 viable cells/ml). This loss in viability did not affect the functionality of viable cells. Percoll colloidal silica solution was found not to be toxic for human hepatocytes. Purification on Percoll density gradient allows the percentage of viable cells to be increased for 55 to 87%. Morphology and Fine Structures Fourteen hours after seeding in a culture medium with FCS, both fresh and thawed hepatocytes attached to the plastic and stable monolayers of polygonal cells with granular cytoplasm were established. The hepatocytes in primary culture could be

TABLE 1 Viability before and after Cryopreservation

Liver

Viability before cryopreservation (%I

HL37

78.5

HL38

83

Viability after thawing (o/o)

Duration of the deep freezing storage (days)

Before purification

After purification

Relative decrease in viability (%I

25 75 136 172 16 29 65

54 53.5 55 51.5 56 57 63

84 86.5 89 88 88 87 85

31.21 31.85 29.94 34.39 32.53 31.33 24.1

Note. Viability was determined by erythrosin B exclusion test; 10% of the dye solution (3.6 mg/ml) was added to the cell suspension, and viable cells were counted on Malassez cell. The relative decrease in viability was estimated by comparing the percentage of viable cells after thawing to the percentage of viable cells after isolating.

458

DOU ET AL.

FIG. I. Electronic microscopy: human hepatocytes 14 h after seeding. (a) Fresh cells : Golgi apparatus (g) and associated vesicles (v) are well developed. Rough endoplasmic reticulum (r) is characteristic of functional hepatocyte. M, mitochondria; N, nucleus. (b) Thawed cells : Golgi apparatus (g) is less developed with few associated vesicles. Large vacuoles (V) are present in cytoplasm among vesiculated smooth and rough endoplasmic reticulum (r). Bar, 1 pm.

maintained for up to 5 days in medium II. Light microscopy showed no morphological differences between frozen and nonfrozen cells. Transmission electron microscopy was used to evidence the major ultrastructural differences between fresh and thawed hu-

man hepatocytes, and the influence of the age of the culture. Fourteen hours after seeding, fresh hepatocytes appeared wellpreserved (Fig. la) but the classical polarity of the cell was completely lost. The endoplasmic reticulum (rough and smooth) and dense bodies were well-developed, illus-

FIG. 2. Electronic microscopy: fresh human hepatocytes 36 h after seeding. (a) Cytoplasm area rich in Golgi apparatus (g), vesicles (v), and lipofuscin (Lf). Mitochondtia (m) are clear and often close to rough endoplasmic reticulum cistema (r). S, smooth endoplasmic reticulum. (b) Area between two reassociated hepatocytes showing structures resembling bile canaliculi (BCL). Cytoplasmic vacuoles (V) are present. Golgi apparatus is associated to vesicles, some of them (arrowhead) being lipoproteine secretory vesicles. (c) Cytoplasm is rich in microfilaments (arrows) and microtubules (large arrowhead). Numerous coated vesicles (small arrowheads) appear in the cytoplasm and associated to the plasma membrane facing the extracellular space in contact with medium (Em). These spaces are often filled with unidentified tibtillar material. Bar, 1 pm.

THAWED

HUMAN

HEPATOCYTES

459

460

DOU ET AL.

FIG. 3. Electronic microscopy: thawed human hepatocytes 36 h after seeding. Cytoplasm is rich in smooth endoplasmic reticulum (s), large vacuoles (V), and lipofuscin (Lf). Golgi apparatus (g) and secretory vesicles (sv) are near the plasma membrane in contact with the medium (Em). r, rough endoplasmic reticulum; m, mitochondria. Bar, 1 pm.

trating an intense activity of the cells in these detoxification processes, perhaps because they had adapted to the medium. The Golgi apparatus and its numerous associated vesicles reflected an intense secretory activity. Thawed hepatocytes were slightly altered (Fig. lb); the endoplasmic reticulum was fragmented into numerous vesicles and the Golgi apparatus was less developed. It

appeared as a few small dictyosomes with only few little associated vesicles at the vicinity of the nuclear envelope. Mitochondria were dilated with a clear matrix and few crystae. Thirty-six hours after seeding, fresh hepatocytes were characterized by the presence of intracellular cavities which resembled bile canaliculi (35, 36). When two cells

FIG. 4. Electronic microscopy: human hepatocytes 62 h after seeding. Bar, 1 km. (a) Fresh cells: bile canaliculi structures are more numerous (BCL) and dilated. Numerous lipofuscin (Lf) accumulated all around these BCL. (b) Thawed cells. Cytoplasm lamina showing a dense net of filaments underlining the plasma membrane in contact with medium (Em). Mitochondria (m) are well preserved. Presence of cytoplasmic vacuoles (V) and of small vesicles (white arrowheads) near the plasma membrane in contact with the Petri dish (Ep). s, smooth endoplasmic reticulum. (c) Thawed cells. Contact area between two cells. No BCL is present, in some parts (arrows) the two plasma membranes are in tight contact. Smooth (s) and rough (r) endoplasmic reticulum are well developed.

THAWED

HUMAN

HEPATOCYTES

Em

461

462

DOU ET AL.

were associated, the intercellular spaces with microvilli and bordered by junctional complexes were always present (Fig. 2b). The cytoplasm of these cells (Fig. 2a) was rich in Golgi elements and lipofuscin. The smooth endoplasmic reticulum was developed, on the contrary the rough endoplasmic reticulum was restricted to citerna close to more or less swollen mitochondria. Highly characteristic was the presence of filaments and microtubule-like structures in these cells (Fig. 2c) and the abundance of coated vesicles in the vicinity of the plasma membrane in contact with the medium (Fig. 2c, small arrowheads). These coated vesicles are indicative of an intense vesicular circulation in this part of the cell. Thawed hepatocytes (Fig. 3) seemed to have recovered a functional appearance: well-developed Golgi apparatus with secretory vesicles appeared near the plasma membrane; the smooth endoplasmic reticulum was abundant. Cytoplasmic vacuoles and lipofuscin were the sign of an intense turnover in which the lysosomal system may be involved. But in any case, the presence of intracellular bile canaliculi-like (BCL) structures was observed and microfilaments or microtubules were not drastically enhanced as is the case for fresh cells. Sixty-two hours after seeding, the bile canaliculus-like structures in fresh cells were more dilated (Fig. 4a) and filaments and microtubule-like structures were always abundant. Except for a marked accumulation of lipofuscin containing dense bodies all around the BCL, the cytoplasm appeared normal. Thawed cells (Fig. 4b) were strikingly flattened: the cytoplasm lay on the support as a thin lamelle no thicker than the diameter of the nucleus. The membrane in contact with the medium was smooth with scarce protrusions and was underlined by a regular layer of filaments (arrows). When two cells were associated (Fig. 4c) we never observed BCL structures. Adhesion between the cells involved desmosomes and gap junctions (Fig. 4c,

short arrows). The endoplasmic reticulum (rough and smooth) was abundant and the cells appeared well developed. Eighty-five hours after seeding, fresh cells (Fig. 5a) were still alive but with a less intense activity. Endoplasmic reticulum and Golgi apparatus were less developed. Microtilaments and microtubules were still overlying the cytoplasm. Lastly, thawed human hepatocytes appeared deeply altered (Fig. 5b). Biochemical

Functionality

of Hepatocytes

Total protein synthesis. Parallel to a wellpreserved morphology, high viability, and plating efficiency, thawed human hepatocytes were shown to maintain their ability to synthesize proteins (Table 2). However, differences between experimental values at 14 and 86 h were statistically significant (P d 0.05) at 14 and 86 h after seeding. In contrast and although secretion rate (extracellular/total) was lower than that of fresh cells, the synthesis capability of thawed hepatocytes was considered similar to those before cryopreservation at 38 and 62 h. Whatever the age of the cells, protein synthesis and secretion level in fresh cultures were constant whereas they seemed increased in thawed hepatocytes. Synthesis and secretion of a specific protein APF. Protein secretion is a good crite-

rion to assesscell functionality. Analysis of extracellular medium allowed estimation of both synthesized and secreted protein. APF was synthetized and secreted in bile as has been demonstrated in isolated rat hepatocytes studies (9). The results presented here clearly show that the APF was synthetized and secreted by human hepatocytes. After cryopreservation, it appeared that two- to fourfold less APF was secreted by thawed human hepatocytes, as compared with fresh cells (see Table 3). These results are in agreement with the observations established for total proteins. The maximum protein recovery in the extracellular medium occurred to 38 h after seeding. The

THAWED

HUMAN HEPATOCYTES

463

FIG. 5. Electronic microscopy: human hepatocytes 85 h after seeding. (a) Fresh cells. Microtilaments (arrows) and microtubules (arrowheads) are still abundant in cytoplasmic area rich in little vesicles. g, golgi apparatus; m, mitochondria; r, rough endoplasmic reticulum; s, smooth endoplasmic reticulum. (b) Thawed cells. Endoplasmic membranes are vesiculated and mitochondria (m) are swollen, with their membranes disrupted. Li, lipid droplet; N, nucleus. Bar 1 pm.

time course differed between fresh and thawed cells. Fresh hepatocytes conserved for a longer time their capability of synthesis and secretion of APF and of all the proteins. Midazolam

Metabolism

Intracellular unchanged MDZ concentrations did not exceed 8.1 k 0.15 and 5.7 k 0.02 nmol/mg of proteins, before and after cryopreservation, respectively, whereas extracellular unchanged MDZ concentrations were 80.7 + 7.7 and 120.3 + 24.6 nmol/mg of proteins (i.e., 67.6 + 5.4% and 77.9 2 5.9% of total radioactivity, respectively). Figure 6 presents HPLC radiochromatograms of MDZ and its metabolites in the

extracellular compartment after biotransformation by fresh or thawed hepatocytes. Both HPLC profiles showed phase I (mono- and dihydroxylated derivatives) and phase II (polar derivatives) metabolites of MDZ. Results in Table 4 demonstrated that in both cases the major metabolites were polar derivatives referred to as GluMDZ and X. Hydroxylated derivatives (DiOH-, 4-OH-, and mainly I-OH-MDZ) represented only a minor part. Sixty-two hours after seeding, MDZ metabolism was drastically different in fresh and thawed cells. To identify X derivative, we separated phase II metabolites from extracellular media by diethyl ether extraction (pH 9) in order to remove unchanged MDZ and phase I metabolites. Aqueous phase was then hydrolyzed by l3 glucuronidases, leading to

464

DOU ET AL. TABLE 2 Protein Synthesis in Fresh and Thawed Hepatocytes IC proteins

T14 h T38 h T62 h T86 h

EC proteins

Total synthesis

Fresh

Thawed

Fresh

Thawed

Fresh

113.62 2 2.89 150.73 f 0.35 152.16 t 38.08 170.1 k 1.07

81.29 k 27.21 149.73 * 57.06 211.46 * 101.35 292.85* -+ 15.2

63 2 8.6 80.33 t 1.59 85.15 k 1.36 92.9 2 1.07

16.58* t 5.84 25.53 ” 7.69 46.81* k 17.93 83.93 2 18.16

176.62 -+ 11.5 231.06 * 1.25 237.31 2 36.72 262.99 k 18.78

Thawed 97.88* k 26.29 175.26 + 57.41 258.26 2 116.38 376.78% k 29.6

Percentage of secretion Fresh

Thawed

35.56 e 2.52 34.77 k 0.5 36.38 2 6.23 35.43 k 2.93

16.53 + 13.07 16.03 2 8.23 19.12 f 6.23 22.26 t 4.64

Note. Neosynthesis of both intracellular (IC) and secreted (EC) proteins is expressed in nmol of Cl4 Leu incorporated in PCA precipitable proteins/mg of total proteins and represents the mean 5 SD (n = 6) from two experiments, each performed in triplicate after cryopreservation. * P < 0.05 compared to values in fresh hepatocytes

complete Glu-MDZ conversion to a product coeluting with the I-OH-MDZ standard, whereas metabolite X was found to be insensitive. After incubation with Helix Pomatia juice of the same aqueous phase, Glu-MDZ and X were both hydrolyzed into two compounds which comigrated with I-OH-MDZ and original MDZ (results not shown). Hepatotoxicity

tests

Figure 7 shows the erythromycin toxicity to human hepatocytes in primary culture before and after cryopreservation, using MTT assay. Percentages of cell viability at 14 and 62 h revealed a lack of correlation between fresh and thawed cells. Most of TABLE 3 Synthesis and Secretion of APF by Isolated Human Hepatocytes Incubated without Bile Salts Time

Fresh

Thawed

14 h 38 h 62 h 86 h

10.8 20.1 20.4 38.4

3.6 7.7 5.4 4.9

Note. Samples were tested as described under Materials and Methods. Each plot represents the determination of APF newly synthetized in the supematant expressed in pg APF/mg of protein in the medium.

the data were significantly different (P < 0.05), especially on the first day of culture. Conversely, comparable dose-response curves were obtained at 38 h. Table 5 gives the IC50 values (pg/ml) for both NR and MTT tests. These tests led to similar protiles of drug toxicity according to the age of the cells and before or after cryopreservation. Thirty-eight hours after seeding, the erythromycin concentration required to reduce 50% of cell viability was 0.5 mg/ml. It was identical after thawing, as compared to prior cryopreservation. DISCUSSION

Cryopreservation of isolated human or animal hepatocytes has previously been reported with various degrees of success depending on the criteria used to assess cell viability and functionality (7, 15, 16, 18, 2427,29,38). Since human hepatocytes in primary culture are increasingly used to overcome the interspecies variability in drug metabolism and toxicity (10, 17, 21, 37), and since human livers are scarce, we developed a protocol for human hepatocyte cryopreservation which allows the maintenance of most of their functionalities after long-term storage. To constitute an available human hepatocytes bank, we evalu-

THAWED

HUMAN HEPATOCYTES

465

thawing, viable hepatocytes represented about 50% of the initial population (determined by exclusion of the erythrosin dye), even after 1 year storage. Several authors have used isosmotic Percoll gradient for purification of viable cells (7,24, 31, 39). We recommend the use of two Percoll solution densities, thus allowing rapid separation of viable hepatocytes from dead cells and debris, without damage. With the methodology, thawed human hepatocytes after seeding are able to 30 attach on a coated support and can survive several days (more than a week) as a stable monolayer of polygonal cells. Most studies (for a review see (34)) have shown the importance of culture medium composition on the protein synthesis which is characteristic of parenchymental hepatic cells (20) and on the maintenance of differentiation states of the hepatocytes. Biochemical and morphological studies on thawed human hepatocytes demonstrated a net recovery of the synthesis apparatus afI ti ter 24 h adaptation. As early as 38 h, the 5 10 15 i +--30 RETENTION TIME (min) rough endoplasmic reticulum and the Golgi FIG. 6. HPLC chromatogram of MDZ and its meapparatus were well-developed and vesitabolites. Metabolites were identified according to cles resembling secretory vesicles were obtheir retention times relative to those of the standards. served. Like rat hepatocytes (7, IS), thawed human hepatocytes were able to ated (i) several morphological parameters synthesize nearly as many proteins as fresh and (ii) biochemical, (iii) metabolic, and (iv) cells. In contrast, the secretory process toxicologic functions of cells in primary seemed affected. De Loecker et al. (8) reculture before and after cryopreservation. ported that preservation of protein syntheMost studies have stressed the impor- sis and membrane transport may not run tance of the composition of the freezing me- parallel. We therefore evaluated the secredium, as well as of the cooling rate for the tion of a specific apoprotein synthesized by optimization of hepatocyte cryopreserva- isolated and cultured hepatocytes, APF, tion (7, 15, 18, 25, 26, 29). In our study, which was found to be associated to lipids M,SO was used in combination with PVP in bile, and which is involved in the hepatic (2) BSA and FCS to improve the recovery uptake of bile-destined cholesterol (9). This and the functionality of cryopreserved he- protein may be a marker of the active sepatocytes (7, 25, 26, 29). Unlike Gomez et cretion of bile by hepatocytes, while total al. (18), but in agreement with other au- protein secretion is an index of plasma sethors (7, 26, 29), we used a slow cooling cretion. Thawed human hepatocytes segradient ( - 1.9YYmin) in the first 18 min of creted less APF than did fresh cells, perfreezing, followed by a rapid temperature haps because the secretory apparatus was decrease of -3O”CYmin for 4 min. After not reorganized after freezing and thawing,

1!O --iiT--

110

466

DOU ET AL. TABLE 4 MDZ Biotransformation by Fresh and Thawed Human Hepatocytes 14 h

GLU X diOH 40H 10H

38 h

62 h

Fresh

Thawed

Fresh

Thawed

Fresh

Thawed

10.76 f 2.14 17.16 2 0.59 0.39 e 0.27 0.91 k 0.64 3.96 2 2.8

9.47 2 4.26 10.76* 2 1.27 1.37 k 0.68 0.74 k 0.64 4.9 k 2.56

12 * 4.1 22.05 k 2.34 0.3 f 0.21 0.3 4 0.21 2.16 ? 1.52

8.93 ” 4.56 10.28* L 1.46 1.28 k 0.68 0.37 k 0.26 1.77 + 1.25

12.62 * 4.09 31.57 2 5.01 0

2.17*

0 1.52 2 1.07

10.79* 0.67 0 2.42

Note. Results are expressed in nmol/mg of total protein, and represent the mean ? SD (n = 6) from two experiments each performed in triplicate after thawing (except at 62 h). * P < 0.05 compared to values of fresh cells.

while synthesis and detoxication systems were restored. Indeed, 38 h after seeding, fresh hepatocytes, in contrast to thawed ones, showed a proliferation of cytoplasmic filaments from attachment area on dishes, and neocanaliculi (35, 36)) expressing a new polarization of the cells. These results are in complete accordance with that of Oyamada and Mori (30) who have shown, by immunocytolocalization of actin and tubulin in hepatocytes, that these structures correspond to microfilaments and tubules. The partial recovery of the secretion capa-

bilities of thawed human hepatocytes may be due to the modification of the cytoarchitecture or the composition of the plasmic membrane, or both. Further studies would be needed to elucidate this point. To test the metabolic capabilities of thawed human hepatocytes, we examined the biotransformation of MDZ, a watersoluble imidazo-benzodiazepine, whose metabolic pathways have been elucidated both in viva and in vitro (11). This compound was particularly interesting since (i) its biotransformation process involved

FIG. 7. In vitro toxicity of erythromycin was determined after 20 h exposure by MTT reduction at 14(A), 38(B), and 62(C) h after seeding, before (0) and after (A) cryopreservation. Results represent the mean of two separate experiments performed in triplicate on thawed cultures. Bars denote SD. Statistics were performed using Student’s t test.

THAWED

HUMAN

TABLE 5 Comparison of the ICSO Values @g/ml) Obtained with Neutral Red and MTT Assays on Fresh and Thawed Hepatocytes MTT

Neutral red

14 h 38 h 62 h

Fresh

Thawed

Fresh

Thawed

255 451 603

657 485 441

353 554 672

740 510 480

phase I and phase II reactions (14, 22), (ii) most of the metabolites are already known and available, and (iii) in vitro human metabolism has been investigated using hepatocytes in suspension (12). MDZ was extensively metabolized by fresh human hepatocytes in primary culture and metabolites were mainly phase II derivatives: Glu-MDZ, resulting from conjugation of 4-OH- and mainly l-OH-MDZ to glucuronic acid; and X, probably resulting from a sulfo conjugation. The latter metabolite has never been evidenced in suspension of human hepatocytes or in microsomes (12), or with pig or rat hepatocytes in primary culture (results not shown). MDZ biotransformation was qualitatively but not quantitatively identical when using human hepatocytes after cryopreservation: there was 1.6-, 2.15, and 2.9fold less metabolite X in extracellular medium of thawed culture than if fresh cells at 14, 38, and 62 h after seeding, respectively. However, in metabolic studies on rat and dog cryopreserved hepatocytes (31), Powis et al. reported that the decrease in conjugate formation was not due to a loss of sulfo and glucuronyl transferase enzyme activities, but probably to a depletion of highenergy cofactors. These results led us to consider thawed human hepatocytes, as well as fresh ones, as representing a suitable model for investigations of drug metabolism and toxicity (23). We therefore compared hepatocyte toxic response toward erythromycin before and after cryopreservation, using NR test (cell

467

HEPATOCYTES

viability in function of NR uptake: incorporation and accumulation in lysosomes) and MTT test (reduction of soluble yellow MTT by mitochondrial succinic deshydrogenases). Results were in agreement with the other assessments used to determine the time-dependence criteria in evaluating cell viability after thawing. In conclusion, the criteria used to assess the functionality of cryopreserved human hepatocytes showed that cell damage induced by freezing was at least partly reversible since a longer period of culture appears to reduce functional differences between frozen and nonfrozen cells. Indeed, thawed human hepatocytes have to be used 38 h after seeding for a better recovery of their functions: membrane integrity, membrane transport, protein synthesis, and stabilization of drug metabolism enzyme. The next step of this study will be to evaluate, with a large number of known chemicals, the validity of using cryopreserved adult human hepatocytes in an in vitro model for metabolic and toxicologic studies. ACKNOWLEDGMENTS The authors thank the staff for the organ transplantation units of Assistance Publique a Marseille (Professors Bricot, Di Marino, and Rampal). REFERENCES 1. Guillouzo, A., Ed. “Liver Cells and Drugs .” Vol. 164. John Libbey Eurotext Ltd. INSERM 1988. 2. Barnard, T. Ultrastructural effects of the high molecular weight cryoprotectants Dextran and Polyvinyl Pyrrolidone on liver and brown adipose tissue in vitro. J. Microsc. 120 (l), 93-104. 3. Bojar, H., Basler, M., Fuchs, F., Dreyfurst, R., Staib, W., and Broelsch, C. H. Preparation of the parenchymal and non parenchymal cells from adult human liver : Morphological and biochemical characteristics. J. C/in. Biochem. 14, 527-532 (1976). 4. Borenfreund, E., Puemer, J. A. Toxicity determined in vitro by morphological alterations and Neutral Red absorption. Toxicol. Left. 24, 119124 (1985). 5. Bradford, M. A rapid and sensitive method for the quantification of microgram quantities of pro-

468

DOU ET AL.

tein utilizing the principle of protein dye binding. Anal. Biochem. 12, 248-254 (1976). 6. Cano, J. P., Rahmani, R., Fabre, G., Richard, B., Lacarelle, B., Bore, P., Bertault-Peres, P., de Sousa, G., Fabre, I., Placidi, M., Coulange, C., Ducros, M., and Rampal, M. “Human Hepatocytes as an Alternative Model to the Use of Animal in Experiments.” Proceedings of International Symposium on liver cells and drugs. John Libbey Eurotext Ltd. INSERM, 1987. 7. Chesne, C., and Guillouzo, A. Cryopreservation of isolated rat hepatocytes : A critical evaluation of freezing and thawing condition. Cryobiology 25, 323-330 (1988). 8. De Loecker, R., Fuller, B. J., Gruwez, J., and De Loecker, W. The effects of cryopreservation on membrane integrity, membrane transport and protein synthesis in rat hepatocytes. Ctyobiology 27, 143-152 (1990). 9. Domingo, N., Botta, D., Martigne-Cros, M., Lechene, P., Pak-Leung, P., Hauton, J. C., and Lafont, H. Evidence for synthesis and secretion of APF-Bile lipid associated protein by isolated rat hepatocytes. B.B.A. 1044, 243-248 (1990). 10. Dou, M., de Sousa, G., Lacarelle, B., Placidi, M., and Rahmani, R. Functional characterization (morphological, metabolic and toxicological) of human hepatocytes after cryopreservation. “In vitro” Methods in cellular pharmacotoxicology. Deuxitme congrts de la Societt de Pharmace-Toxicologic Cellulaire S.P.T.C., Montpellier, March l-2, 1990. 11. Fabre, G., Crevat-Pisano, P., Dragma, S., Cow, J., Barra, Y., and Cano, J. P. Involvement of the macrolide antibiotic inductible cytochrome P450 LM,, in the metabolism of Midazolam by microsomal fractions prepared from rabbit liver. Biochem. Pharmacol. 37 (lo), 1947-1953 (1988). 12. Fabre, G., Rahmani, R., Placidi, M., Combalbert, J., Covo, J., Cano, J. P., Coulange, C., Ducros, M., and Rampal, M. Characterization of Midazolam metabolism using human hepatic fractions and hepatocytes in suspension obtained by perfusing whole human livers. Biochem. Pharmacol. 37 (22), 4389-4397 (1988). 13. Fautrel, A., Chesne, C., Guillouzo, A., de Sousa, G., Placidi, M., Rahmani, R., Braut, F., Pichon, J., Hoellinger, H., Vintezou, P., Diarte, I., Melcion, C., Cordier, A., Lorenzon, G., Benicourt, M., Vannier, B., Fournex, R., Peloux, A. F., Bichet, N., Gouy, D., Cano, J. P., and Lounes, R. A multicenter study of acute in vitro cytotoxicity in rat liver cells. Toxicol. In Vitro, 5 (1991). 14. Fragen, R. J., Gahl, F., and Caldwell, N. A water-

soluble benzodiazepine, R021-3981, for induction of anesthesia. Anesthesia 49, 41 (1978). 15. Fuller, B., Grout, B., and Woods, R. Biochemical and ultrastructural examination of cryopreserved hepatocytes in rat. Cryobiology 19,493502 (1982). 16. Fuller, B., Woods, R., Howe, S., Nutt, L., Attenburrow, V., and Hobbs, K. E. F. Metabolic effects of cryopreservation on isolated hepatocytes at rat. Cryobiology 17, 618-620 (1980). 17. Gerecke, M. Chemical structure and properties of Midazolam compared with other Benzodiazepine. Br. J. Clin. Pharmacol. 16, IlS-15s (1983). 18. Gomez, M. J., Lopez, P., and Castell, J. V. Biochemical functionality and recovery of hepatocytes after deep freezing storage. In Vitro 20, 826-832 (1984). 19. Guguen-Guillouzo, C., Champion, J. P., Brissot, P., Glaie, E. D., Launois, B., Bourel, M., and Guilouzo, A. High yield preparation of isolated human adult hepatocytes by enzymatic perfusion of the liver. Cell Bid. Znt. Rep. 6, 625-628 (1982). 20. Guillouzo, A. Plasma protein production by cultured adult hepatocytes. In “Isolated and Cultured Hepatocytes” (A. Guillouzo and C. Guguen-Guillouzo Eds.), pp. 155-170. Le Cditions INSERM, Paris, John Libbey, London eurotext, 1986. 21. Guillouzo, A., Begue, J. M., Champion, J. P., Gascoin, M. N., and Guguen-Guillouzo, C. Human hepatocytes culture : A model of pharmace-toxicological studies. Xenobiotica 15, 635-641 (1985). 22. Heizmann, P., Eckert, M., and Zeigler, W. H. Pharmacokinetics and bioavailability of midazolam in man. Br. J. C/in. Pharmacol. 16, 43s (1983). 23. Inaba, T., Makowka, L., Rotstein, L., Mahon, W. A., Falk, R. E., Falk, J. E., and Phillips, M. J. Aminopyrin metabolism in cryopreserved isolated rat hepatocytes. Can. J. Physiol. Pharmacol. 59, 408-411 (1981). 24. Innes, G. K., Fuller, B. J., and Hobbs, K. E. F. Functional testing of hepatocytes following their recovery from cryopreservation. Cryobiology 25, 23-30 (1988). 25. Le Cam, A., Guillouzo, A., and Freychet, P. Ultrastructural and biochemical studies of isolated adult rat hepatocytes prepared under hypoxic conditions. Exp. Cell. Res. 98, 382-395 (1976). 26. Loretz, L. J., Li, A. P., Flye, M. W., and Wilson, A. G. E. Optimisation of cryopreservation procedures for rat and human hepatocytes. Xenobiorica 19 (5), 489-498 (1989). 27. Maganto, P., Cienfuegos, J. A., Santamaria, L.,

THAWED

28.

29.

30.

31.

32.

33.

34.

35.

HUMAN HEPATOCYTES

Eroles, G., Andres, S., Castillo-Olivares, J. L., and Municio, A. M. Cryopreservation and transplantation of hepatocytes: An approach for culture and clinical application. Cryobiology 25, 311-322 (1988). Mosmann, T. Rapid calorimetric assay for cellular growth and survival. Application to proliferation and cytotoxicity assays. J. Immunol. Methods, 65, 55-63 (1983). Novicki, D. L., Irons, G. P., Strom, S. C., Jirtle, R., and Michalopoulos, G. Cryopreservation of isolated rat hepatocytes. In vitro 18, 393-399 (1982). Oyamada, M., and Mori, M. Immunohistochemical demonstration of tubulin and actin in rat hepatocytes in situ using a perfusion extractionfixation procedure. J. Histochem. Cytochem. 33 (12) 1197-1204 (1985). Powis, G., Santone, K. S., Melder, D. C., Thomas, L., Moore, D. J., and Wilke, T. J. Cryopreservation of rat and dog hepatocytes for studies of xenobiotic metabolism and activation. Drug Metab. Disp. 15 (6), 826832 (1987). Rahmani, R., Richard, B., Fabre, G., and Cano, J. P. Extrapolation of preclinical pharmacokinetic data to therapeutic drug use. Xenobiotica 18, 71-78 (1988). Reese, J. A., and Bryard, J. L. Isolation and culture of adult hepatocytes from liver biopsies. In Vitro 17, 935-941 (1982). Reid, L. M., Narita, M., Fujita, J., Murray, Z., Liverpool, C., and Rosenberg, L. Matrix and hormonal regulation of differentiation in liver cultures. In “Isolated and Cultured Hepatocytes” (A. Guillouzo and C. Guguen-Guillouzo, Eds.), pp. 225-258. John Libbey, London/Paris (1986). Remy, L., Chalvet, C., Ripert, J. P., and Gero-

36.

37.

38.

39.

40.

41.

469

lami, A. Intracellular lumina and bile canaliculi in rat hepatocytes in vitro: A cytochemical study. Acta. Hisrochem. 85, 87-92 (1989). Remy, L., Reynier, M. O., Hashieh, I. A., and Gerolami, A. La lumitre intracellulaire dans les hepatocytes de rat en culture. Etude d’une infrastructure cytoplasmique originale. T.C.R. Acad. Sci. Paris, t. 307, serie III, 63-68 (1988). Richard, B., Fabre, G., de Sousa, G., Fabre, I., Rahmani, R., and Cano, J. P. Interspecies variability in mithoxanthrone metabolism using primary culture of hepatocytes isolated from rat, rabbit and human. Biochem. Phurmacol. 41(2), 255-262 (1991). Rijntjes, P. J. M., Moshage, H. J., Van Gemert, P. J. L., De Waals, R., and Yap, S. H. Cryopreservation of adult human hepatocytes: the influence of deep freezing storage on viability, cell seeding, survival, tine structures, and albumin synthesis in primary culture. J. Hepatol. 3, 7-18 (1986). Santone, K. S., Medler, D. C., and Powis, G. Studies of chemical toxicity to fresh and cryopreserved rat hepatocytes. Toxicol. Appl. Pharmucol. 97, 370-376 (1989). de Sousa, G., Dou, M., Barbe, D., Lacarelle, B., Placidi, M., and Rahmani, R. Freshly isolated or cryopreserved human hepatocytes in primary culture : Influence of drug metabolism on hepatotoxicity. Toxicol. In Vitro 5 (1991). de Sousa, G., Dou, M., Lacarelle, B., Placidi, M., and Rahmani, R. Evaluation of metabolic activities (Phase I and Phase II) of cryopreserved human hepatocytes. European journal of drug metabolism and pharmacokinetics, abstract 138. Fourth European Congress of Biopharmaceutics and Pharmacokinetics. Geneve, April, 17-19, 1990.