Prediction of in vivo kinetic parameters from in vitro systems is a widely accepted practice to estimate pharmacokinetic properties before human drug administration. The key parameter estimated from such studies is intrinsic clearance (CL_int²), and the ability to estimate this parameter for new chemical entities, as part of the drug discovery program and in early stages of the drug development process, is a valuable asset for the pharmaceutical industry. To date, freshly isolated hepatocytes have been considered the in vitro system of choice for these prediction studies because they represent a more realistic physiological situation than other preparations, e.g., hepatic microsomes (Houston and Carlile, 1997). In addition to possessing the full spectrum of drug-metabolizing enzymes, cofactors, and cell membrane receptors, they have a practical advantage due to the rapid dispersion of compound throughout the incubation that facilitates sampling and aids rapid distribution of the drug to metabolizing enzymes. Despite these advantages, there remain some unresolved issues with the use of freshly isolated hepatocyte suspensions. These include identifying appropriate storage conditions, prolonging viability (Skett, 1994) beyond the normal incubation time for drug metabolism studies (approximately 4 h), and maintaining activity of the drug-metabolizing enzymes, particularly the P450 enzymes.

Cryopreserved hepatocytes have the advantage of allowing longterm storage, comparable with that available for hepatic microsomes, and their routine use would help optimize limited human tissue availability and reduce the number of animals required during drug discovery. Hepatocytes from different human donors can be pooled to decrease some of the variability in preparation and to incorporate some of the interindividual variability in enzyme expression between donors. The literature on the optimization of cryopreservation protocols for hepatocytes from both animals and humans shows that functional activity can be maintained (Li et al., 1999; Steinberg et al., 1999) for most drug-metabolizing enzymes. However, the inherent problem of loss of activity within a short incubation period (<4 h) remains.

Despite the large amount of methodological literature available and the existence of several commercial sources, there has been comparatively little effort to explore the use of cryopreserved hepatocytes in predicting in vivo pharmacokinetic parameters. Varied success in estimating in vivo clearance for a range of drug-metabolizing enzymes in various species has been reported (Lau et al., 2002; Shibata et al., 2002; Soars et al., 2002; Naritomi et al., 2003). It is perhaps significant that four of these five studies assessed in vitro metabolism by disappearance of drug substrate. The most comprehensive study involving four animal species and over 20 drugs in each species (Lau et al., 2002) used a high-throughput approach calculating clearance based on drug disappearance from a single 2-h time point. This study chose not to incorporate plasma protein binding on the grounds that their study was evaluating a screening approach to eliminate high-clearance drugs; this resulted in human predictions for low-clearance drugs that were more than 50-fold higher than the observed values. It is important to evaluate comprehensively the ability of cryopreserved hepatocytes to predict in vivo CL_int and to identify any limitations of the model in an animal species where direct comparisons can be made to fresh hepatocytes. Then, the routine use of valuable human tissue can be justified. Such an evaluation can only be achieved by a detailed examination of the kinetics of metabolite formation of a range of prototypic drug substrates to provide comprehensive coverage of the full range of enzymes and pathways of metabolism.

The aim of the current study is therefore to extend the previous investigations into the ability of cryopreserved rat hepatocytes to predict in vitro CL_int in comparison with freshly isolated hepatocytes. Standard methods (Houston, 1994) were used to determine the in vitro CL_int monitored via the metabolite formation for 14 metabolic pathways. The compounds were selected for this evaluation study to provide a wide range of in vivo clearance values (>100-fold) in both rat and human and of physicochemical properties (Table 1). They are tolbutamide (TOL), S-warfarin (WAR), phenytoin (PHE), 7-ethoxycoumarin (7-EC), nordiazepam (NDZ), diazepam (DZ), dextromethorphan (DEX), and propranolol (PROP).

Materials and Methods

Chemicals. OXP and TZ were generously supplied by Wyeth-Ayerst (Princeton, NJ). DZ, 4′-OH DZ, and 4′-OH NDZ were gifts from Roche Ltd. (Welwyn Garden City, Hertfordshire, UK). WAR and HTOL were gifts from Celltech (Cambridge, UK) and Aventis (Strasbourg, France), respectively. DEX, MEM, and DOR were supplied by Hoffman-La Roche (Nutley, NJ). WAR metabolites were purchased from Ultrafine (Manchester, UK), collagenase H from Roche Molecular Biochemicals (Roche Diagnostics, Basel, Switzerland), and FCS from Invitrogen (Carlsbad, CA). Cryopreserved rat hepatocytes and thawing media were purchased from XenoTech (Kansas City, KS). All other substrates and reagents were purchased from Sigma Chemical (Poole, Dorset, UK) or BDH Laboratory Supplies (Lutterworth, Leicester, UK) and were of the highest grade available.

Incubation Conditions for Substrates. All substrates were prepared in DMF to give final concentrations in incubations as detailed in Table 2. Linearity studies were performed with respect to cell density, incubation time, and optimal conditions for metabolite formation selected, as also detailed in Table 2.

Preparation and Incubation of Fresh Hepatocyte Suspensions. Hepatocytes were prepared by the collagenase perfusion method (Berry and Friend, 1969). Briefly, male Sprague-Dawley rats (weight 200–250 g; Charles River, Margate, Kent, UK) were sacrificed by cervical dislocation, and the liver was removed. The two largest lobes were cannulated and perfused at 6 to 8 ml/min with Earle' balanced salt solution (EBSS) containing 0.25 mM EGTA for 4 min. The EGTA was flushed out with EBSS for 4 min before perfusion with EBSS containing 1 mM calcium chloride, 0.68 mg/ml collagenase H, and 0.07 mg/ml trypsin inhibitor for 8 to 12 min until the cell matrix was seen to break. The cells were gently teased out of the liver capsule into Williams medium E (WME), pH 7.4, containing 1% (w/v) BSA, and the cell suspension was filtered through nylon mesh and centrifuged for 3 min at 100g. The supernatant was discarded, the cell pellet was resuspended in WME containing 1% BSA (w/v), and the centrifugation step was repeated twice. The final cell pellet was resuspended in WME containing 0.1% BSA, and the viability was obtained using the trypan blue exclusion method. Only preparations with a viability greater than 85% were used. The cell suspension was diluted to the relevant cell density in WME based on the number of viable cells. All incubations were performed in Eppendorf tubes in a Thermomixer (Eppendorf AG, Hamburg, Germany) set at 37°C and 900 rpm. The incubation volume was 1 ml for all compounds. Incubations were initiated by the addition of 5 μl of substrate in DMF (final solvent concentration of 0.5% v/v) after 5-min preincubation of the cell suspension. Termination of incubations was by snap-freezing in liquid nitrogen. All incubations were performed in duplicate, and samples were analyzed as described below.

Preparation and Incubation of Cryopreserved Hepatocyte Suspensions. Cryopreserved rat hepatocytes (from male Sprague-Dawley rats) were thawed according to the vendor's instructions. Briefly, hepatocytes were thawed at 37°C for 1.5 min and centrifuged through supplemented Dulbecco's modified Eagle's medium containing Percoll for 5 min at 90g. The supernatant was removed and hepatocytes were washed by centrifugation in Dulbecco's modified Eagle's medium for 3 min at 60g. The pellet was resuspended in Waymouth MB 752/1 medium (Sigma Chemical) containing 5% FCS, and viability was assessed by trypan blue exclusion. The average viability of cryopreserved hepatocytes was 84.9 ± 3.8%. The cell suspension was diluted to the relevant cell density in Waymouth's medium containing 5% FCS based on the number of viable cells and transferred (0.125 ml) into each well of a 24-well plate (Corning Glassworks, Corning, NY). All incubations were performed in an incubator (Sanyo Gallenkamp PLC, Leicester, UK) set at 37°C with 5% CO₂ and >95% humidity. After 5-min preincubation, reactions were initiated by addition of 0.125 ml of prewarmed Waymouth's medium containing 5% FCS and 1.25 μl of substrate in DMF followed by mixing for 10 s at 400 rpm. The final incubation volume was 0.25 ml and solvent concentration was 0.5% (v/v) in all incubations. Reactions were terminated by snap-freezing in liquid nitrogen, and samples were analyzed as described below.

Sample Hydrolysis, Extraction, and Analysis. Samples for 7-EC, PHE, NDZ, DZ, DEX, and WAR were hydrolyzed with an equal volume of β-glucuronidase with sulfatase activity (1000 U/ml for 7-EC and 200 U/ml for all other substrates) in 60 mM sodium acetate (pH 4.5) for 1 h at 37°C. With the exception of 7-EC, all substrates and their metabolites were extracted from the cell suspension by liquid-liquid extraction after addition of internal standards by rotary mixing and centrifuging in ethyl acetate (NDZ, DZ, and PROP), tert-butylmethylether (PHE and DEX), or chloroform (TOL and WAR). Samples were reconstituted in mobile phase and analyzed by high-performance liquid chromatography-UV. 7-EC samples were analyzed by fluorimetry after addition of glycine/NaOH buffer. Metabolite levels in samples were quantified against standard curves for the metabolites. Table 3 shows the analytical conditions for each compound.

Data Analysis. All data were analyzed using GRAFIT 4.0 software (Erithacus Software, Horley, Surrey, UK). The kinetic model selected for fitting of the data was based on visual inspection of the data on Eadie-Hofstee or Hanes plots. For TOL, DZP, and WAR metabolites and DOR formation in cryopreserved suspensions, the Michaelis-Menten equation was used and CL_int estimated using the ratio of V_max to K_M (Houston, 1994). For 7-EC and PHE metabolites, analysis was performed assuming two-site kinetics and total CL_int determined by addition of the individual CL_int values for the low- and high-affinity components. For MEM and DOR formation from DEX in fresh suspensions, the Hill equation was used and CL_max was determined (Houston and Kenworthy, 1999). Where more than one metabolite was measured, the individual CL_int values were added to give total CL_int. Analysis of PROP was performed assuming mono-exponential depletion where CL_int was calculated from the dose to the area under the curve ratio.

Binding Experiments Using Equilibrium Dialysis. Compounds, at various concentrations, were mixed with WME containing 0.03% BSA or Waymouth's medium containing 5% FCS to mimic the conditions in fresh and cryopreserved suspensions, respectively. The mixtures were subject to equilibrium dialysis against WME or Waymouth's medium without protein supplement at 37°C using a Dianorm apparatus (Diachema AG, Zurich, Switzerland). Diachema membranes were used (molecular weight cutoff, 50 kDa; diameter, 63 mm) and cells were rotated at 5 rpm for 4 h. Dialysis on each compound was performed in triplicate. On completion of the dialysis period, both the protein and media fractions were removed from the cell and samples were analyzed as above. Fraction unbound values were determined using the concentration of free compound in the dialysate and the total concentration of compound added to the dialysis chamber.

Results

In vitro CL_int was determined for eight commonly used probe substrates in freshly isolated and cryopreserved rat hepatocyte suspensions. For all compounds with the exception of PROP, the kinetic parameters, V_max and K_M (or S₅₀), were determined for the major metabolites of each compound, and the ratio of these parameters was used to estimate in vitro CL_int. Due to the large number of PROP metabolites, the depletion approach was used where the disappearance of substrate was monitored over time (10 time points), and in vitro CL_int was estimated from the ratio of the initial substrate concentration and the area under the depletion curve.

The metabolites studied displayed a wide variety of kinetic profiles, including typical Michaelis-Menten kinetics (HTOL, WAR metabolites 4-, 6-, 7-, 8-, and 10-OH, and DZ metabolites 4′-OH DZ, TZ, NDZ), two-site kinetics (7-HC and 4-OH PHE), sigmoidal kinetics [OXP and DEX metabolites (DOR, MEM)], and substrate inhibition (4′-OH NDZ). In the majority of cases (13 of 14), the same type of kinetics was observed in both hepatocyte models. The only exception to this was the loss of sigmoidicity for DOR formation in cryopreserved suspensions. Some changes in the contribution of each metabolite to total metabolism were also observed for four substrates. DOR formation was the dominant pathway for DEX metabolism in cryopreserved suspensions in comparison with fresh, where DOR and MEM have equal contributions. The second site of 4-OH PHE formation had a less important role in cryopreserved suspensions compared with fresh cells for PHE. For WAR, there was a change in the major metabolite from 7-OH to 8-hydroxywarfarin in cryopreserved suspensions accompanied by a loss of 10-hydroxywarfarin. The contribution of OXP and 4′-OH NDZ to NDZ metabolism was equal in cryopreserved suspensions compared with fresh where 4′-OH NDZ was the major metabolite.

The in vitro CL_int values estimated from both hepatocyte models covered a wide range from 0.1 to 98 μl/min/10⁶cells and are shown in Table 4 along with the corresponding V_max and K_M values. Figure 1 compares the in vitro CL_int values determined for the individual metabolites of the compounds in fresh and cryopreserved rat hepatocyte suspensions. In Fig. 1, the metabolites are coded according to the degree of accuracy of the in vitro CL_int estimation in cryopreserved suspensions in comparison with fresh suspensions. The degree of accuracy was assessed in three categories: good estimations, overestimations, and underestimations, based on whether the difference in the mean values was within 2-fold. For HTOL, PROP, and the three DZ metabolites, there was a good estimation of in vitro CL_int from cryopreserved suspensions compared with that in fresh suspensions. For WAR, there was an overestimation of total in vitro CL_int by approximately 125% that was accompanied by an increase in 4-, 6-, and 8-OH, but good agreement for 7-hydroxywarfarin metabolism. For the other six metabolites (formed from PHE, NDZ, 7-EC, DEX), there was an underestimation of in vitro CL_int observed in cryopreserved suspensions compared with fresh. This was particularly pronounced (4–16%) for pathways showing atypical Michaelis-Menten kinetic profiles (DOR, MEM, OXP, and 4′-OH NDZ), but less so (25–45%) for pathways showing biphasic Michaelis-Menten kinetics (7-HC and 4-OH PHE).

• Download figure
• Open in new tab
• Download powerpoint

Fig. 1.

In vitro CL_int estimates for individual metabolites in fresh and cryopreserved rat hepatocyte suspensions.

The metabolites represented are 7-HC (1), 4-OH PHE (2), HTOL (3), 4′-OH NDZ (4a), OXP (4b), 4-hydroxywarfarin (5a), 6-hydroxywarfarin (5b), 7-hydroxywarfarin (5c), 8-hydroxywarfarin (5d), DOR (6a), MEM (6b), 4′-OH DZ (7a), NDZ (7b), TZ (7c), and PROP (8). In the case of OXP and MEM in both models and DOR in fresh suspensions, the values are CL_max. Compounds giving good agreement between in vitro CL_int in cryopreserved suspensions and fresh suspensions are shown by open circles; those showing an underprediction are shown relative to fresh suspensions by closed squares; and those displaying an overprediction relative to fresh suspensions are shown by closed triangles. The solid line is a line of unity. Data are represented as mean ± S.D. of at least three determinations.

The changes in in vitro CL_int were reflected in changes in the individual kinetic parameters, V_max and K_M (or S₅₀), for each metabolite. Figures 2 and 3 compare the V_max and K_M (or S₅₀) values estimated in fresh and cryopreserved rat hepatocyte suspensions where the metabolites are again coded according to the degree of accuracy in the in vitro CL_int estimation. Figure 2 shows that generally where a good or overestimation of in vitro CL_int is obtained in cryopreserved suspensions relative to fresh cells for the metabolites, no change in V_max was observed. Significant decreases were observed, however, for certain metabolites exhibiting an underestimation of in vitro CL_int in cryopreserved suspensions relative to fresh cells (notably OXP, PHE, and 4′-OH NDZ). From Fig. 3, it can be seen that an overestimation of in vitro CL_int for 4-, 6-, and 8-hydroxywarfarin was accompanied by a decrease in the K_M of the individual metabolites in cryopreserved suspensions. An increase in K_M or a loss of sigmoidicity (increased S₅₀) was related to a decrease in in vitro CL_int for some metabolites (7-HC, MEM, and DOR, respectively). A general trend emerged where a markedly lower CL_int estimation from cryopreserved suspensions compared with fresh cells was associated with either a decrease in V_max or an increase in K_M (or S₅₀), whereas an increase in in vitro CL_int estimation was associated with a decrease in K_M.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 2.

V_maxestimates for individual metabolites in fresh and cryopreserved suspensions.

The metabolites represented are 7-HC (1), 4-OH PHE (2), HTOL (3), 4′-OH NDZ (4a), OXP (4b), 4-hydroxywarfarin (5a), 6-hydroxywarfarin (5b), 7-hydroxywarfarin (5c), 8-hydroxywarfarin (5d), DOR (6a), MEM (6b), 4′-OH DZ (7a), NDZ (7b), and TZ (7c). Symbols are identical to the ones used in Fig. 1. The solid line is a line of unity. Data are represented as mean ± S.D. of at least three determinations.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 3.

K_Mestimates for individual metabolites in fresh and cryopreserved suspensions.

The metabolites represented are 7-HC (1), 4-OH PHE (2), HTOL (3), 4′-OH NDZ (4a), OXP (4b), 4-hydroxywarfarin (5a), 6-hydroxywarfarin (5b), 7-hydroxywarfarin (5c), 8-hydroxywarfarin (5d), DOR (6a), MEM (6b), 4′-OH DZ (7a), NDZ (7b), and TZ (7c). In the case of OXP and MEM in both models and DOR in fresh suspensions the values are S₅₀. Symbols are identical to the ones used in Fig. 1. The solid line is a line of unity. Data are represented as mean ± S.D. of at least three determinations.

Equilibrium dialysis was performed to estimate the free fraction of compound available for metabolism in both fresh and cryopreserved hepatocyte models. The extent of binding to BSA or FCS in fresh and cryopreserved hepatocyte suspensions was similar and therefore did not account for the difference in values of in vitro CL_int obtained for the two models.

Discussion

To date, the most successful in vitro system for the prediction of in vivo CL_int has been freshly isolated hepatocyte suspensions (Houston and Carlile, 1997). However, due to the limitations of this model, newer approaches such as cryopreservation need to be validated. Most investigators determining in vivo clearance with cryopreserved hepatocytes (Lau et al., 2002; Shibata et al., 2002; Soars et al., 2002; Naritomi et al., 2003) have used the substrate depletion approach to estimate in vitro CL_int. Although some success has been reported, a detailed kinetic examination covering a range of enzymes, pathways of metabolism, and CL_int values is required for a full evaluation of the utility of cryopreserved hepatocytes Therefore, in the present study, the in vitro CL_int of eight commonly used probe substrates (TOL, WAR, PHE, 7-EC, NDZ, DZ, DEX, and PROP) was determined in fresh and cryopreserved rat hepatocyte suspensions. The compounds were selected because they displayed over a 100-fold range of in vivo clearance values in both rat and human and a wide variety of physicochemical properties such as Log P and pK_a, and they are metabolized by a variety of P450 enzymes (as shown in Table 1). The metabolite formation method was used for all compounds with the exception of PROP, for which the substrate depletion method was used.

The in vitro CL_int values determined for the compounds in freshly isolated rat hepatocytes in this study are generally in agreement with those in the literature (Tables 1 and 4). The exceptions were the in vitro CL_int values determined for PHE and TOL, which in the present study are lower than those of Ashforth et al. (1995). This was due to increased sensitivity of the analytical methods allowing lower substrate concentrations to be studied and/or inclusion of a larger number of substrate concentrations, which improves the accuracy of the prediction upon modeling of the data. Fresh hepatocyte suspensions have already been shown to be good predictors of in vivo CL_int (Houston and Carlile, 1997), and it can therefore be assumed that since the in vitro CL_int values determined in the present study are similar to those determined previously, they will give good estimates of in vivo CL_int and hepatic clearance using a liver model.

The results show that cryopreserved hepatocytes give similar in vitro CL_int values to freshly isolated rat hepatocytes for certain compounds, namely TOL, DZ (all three metabolites), and PROP. HTOL and the three DZ metabolites exhibit typical Michaelis-Menten kinetics in their metabolite formation, and the in vitro CL_int of PROP was determined by substrate depletion. Cryopreserved hepatocytes gave an overestimation of in vitro CL_int compared with fresh cells for WAR as a result of decreased K_M for three of its four metabolites. An under-estimation of in vitro CL_int was observed for 7-EC, PHE, NDZ, and DEX in cryopreserved hepatocytes as a result of decreased V_max or increased K_M (or S₅₀) values. Interestingly, these compounds do not exhibit single-site Michaelis-Menten enzyme kinetics in their metabolite formation; two-site kinetics were observed for 7-EC and PHE, and sigmoidicity and substrate inhibition for DEX metabolites and 4′OH NDZ, respectively. The latter two types of atypical Michaelis-Menten kinetic profiles resulted in particularly pronounced (4–16%) differences in CL_int values, whereas the differences were much less (25–45%) in the cases of biphasic Michaelis-Menten kinetics.

Given the clear differences evident in the in vitro kinetics between fresh and cryopreserved hepatocyte suspensions, the overall rank order of CL_int values for the 15 pathways can only be described as roughly similar (Fig. 1). However, the type of kinetics observed was the same for most metabolites.

Drug binding to BSA or FCS in the incubation media for fresh and cryopreserved suspensions was found to be similar and therefore does not explain the CL_int differences seen between models. Although the same strain of rat was used for fresh and cryopreserved hepatocytes, the sources were different: the fresh hepatocytes were isolated from rats raised in the United Kingdom, and the cryopreserved hepatocytes were from the United States. This may have some contribution to the difference in clearance values between the two models; however, probably a more important contributing factor to the differences involves the incubation conditions required for each of the two models. Freshly isolated hepatocytes were incubated in Eppendorf tubes with agitation, as is common practice for clearance studies. In contrast, cryopreserved hepatocytes were incubated, as recommended by the vendor, in a 24-well plate in an incubator with 5% CO₂ and without agitation. Incubating hepatocytes without agitation may cause an uneven distribution of substrate throughout the incubation media, possibly resulting in concentration gradient effects and impaired entry into hepatocytes. For higher CL_int compounds (e.g., PHE, 7-EC, DEX, and NDZ), entry into the hepatocyte may be the rate-limiting step for clearance. In contrast, the lack of agitation is likely to have much less impact on the low CL_int compounds.

Interestingly, other studies comparing these two models have used the same incubation conditions for both systems. Lau et al. (2002) incubated both fresh and cryopreserved hepatocytes in 24-well plates without agitation, and although there was a good correlation between the two systems, there was a degree of underestimation of the in vivo clearance (see Introduction). Naritomi et al. (2003) used the same incubation conditions for both fresh and cryopreserved hepatocytes, although the type of incubation vessel and whether samples were agitated during incubation were not documented. Once again, there was a good correlation between the two systems, but a combination of scaling factors was required to give good estimations of the in vivo data.

However, cryopreserved hepatocytes gave similar high clearances for PROP and DZ to fresh hepatocytes, indicating that the incubation conditions are not the sole cause of the decrease in activity. There was no relationship between the differences seen in in vitro CL_int and the particular P450 enzyme(s) involved in the metabolism of these compounds. Although the poorly correlated DEX and NDZ are CYP3A and CYP2D substrates, so are the well correlated PROP and DZ. We conclude that there is no general loss of metabolizing capacity in hepatocytes subjected to cryopreservation, confirming some observations in the literature (Li et al., 1999; Steinberg et al., 1999). The possibility that the cryopreservation process may cause degradation or decrease the activity of particular P450 enzymes leading to a decrease in turnover as indicated by the decreases in V_max accompanying the decrease in in vitro CL_int requires further investigation with more selective P450 probes.

In contrast to the high CL_int compounds, cryopreserved hepatocytes gave a higher estimate of in vitro CL_int than fresh hepatocytes for WAR. Since the incubation times for both models were similar, these overpredictions cannot be attributed to improved stability of cryopreserved suspensions. However, they may be linked to the differences in incubation conditions and the effect this may have on the particular P450 enzyme (CYP2C family for the metabolism of WAR and TOL). It can be speculated from these results that cryopreserved suspensions may give better routine predictions than fresh suspensions of in vitro CL_int for low turnover compounds, although further work is required to confirm these observations. This may be an important advantage for the pharmaceutical industry, where identification of compounds with low CL_int is a common goal.

The results of these studies with cryopreserved hepatocyte suspensions indicate that this system can give accurate predictions for compounds exhibiting typical Michaelis-Menten kinetics and/or low turnover compounds. However, they are less successful in providing accurate predictions for compounds showing more complex kinetics. In addition to the differences between some mean in vitro CL_int values, there is a higher degree of variation between batches of cryopreserved hepatocytes compared with fresh suspensions as indicated by larger standard deviations. In conclusion, our studies do not endorse the observations of previous investigators (Lau et al., 2002; Naritomi et al., 2003), who have concluded that cryopreserved and fresh hepatocytes are equally valuable for determination of CL_int values. Further work is required to determine the reasons behind the significant differences in turnover compared with freshly isolated hepatocytes before the use of cryopreserved hepatocytes as a prediction tool becomes routine.