Biotechnol. #og. 1001, 7, 554-559
554
Intracellular Ice Formation during the Freezing of Hepatocytes Cultured in a Double Collagen Gel A. Hubel,?M. Toner,+*$ E. G. Cravalho,t M. L. Yarmush,*J*§ and R. G. Tompkins*** Department of Mechanical Engineering, Massachusetts Institute of Technology, Cambridge, Massachusetts 02139, Department of Surgery, Massachusetts General Hospital, Harvard Medical School, Boston, Massachusetts 02114, and Department of Chemical and Biochemical Engineering, Rutgers University, Piscataway, New Jersey 08855
During freezing, intracellular ice formation (IIF)has been correlated with loss in viability for a wide variety of biological systems. Hence, determination of IIF characteristics is essential in the development of an efficient methodology for cryopreservation. In this study, IIF characteristics of hepatocytes cultured in a collagen matrix were determined using cryomicroscopy. Four factors influenced the IIF behavior of the hepatocytes in the matrix: cooling rate, final cooling temperature, concentration of MeZSO, and time in culture prior to freezing. The maximum cumulative fraction of cells with IIF increased with increasing cooling rate. For cultured cells frozen in Dulbecco's modified Eagle's was 70 medium (DMEM), the cooling rate for which 50% of the cells formed ice (BM) OC/min for cells frozen after 1day in culture and decreased to 15 "C/min for cells frozen after 7 days in culture. When cells were frozen in a 0.5 M MezSO DMEM solution, the value of B50 decreased from 70 to 50 "C/min for cells in culture for 1day and from 15 to 10 OC/min for cells in culture for 7 days. The value of the average temperature for cultured cells was only slightly depressed by the addition of MezSO for IIF (TIIF) when compared to the IIF behavior of other cell types. The results of this study indicate that the presence of the collagen matrix alters significantly the IIF characteristics of hepatocytes. Thus freezing studies using hepatocytes in suspension are not useful in predicting the freezing behavior of hepatocytes cultured in a collagen matrix. Furthermore, the weak effect of MeZSO on IIF characteristics implies that lower concentrations of MezSO (0.5 M) may be just as effective in preservingviability. Finally, the value of Bm measured in this study indicates that cooling rates nearly an order of magnitude faster than those previously investigated could be used for cryopreservation of the hepatocytes in a collagen gel.
+
Introduction The establishment of efficient hepatocyte storage protocols is important for a variety of clinical and biotechnological applications including the development of extracorporeal liver support devices and liver cell transplantation. At present, the task of maintaining a supply of hepatocytes through cryopreservation appears essential because adult hepatocytes do not readily proliferate in culture. The capability of freezing liver cells would thus presumably allow the pooling of donor cells and ensure a stable supply of hepatocytes. During the freezing process, the cell in solution is subjected to temperatures below the equilibrium freezing temperature of the solution. As ice forms in the extracellular solution, solute is rejected from the solid phase, and correspondingly, the chemical potential in the unfrozen portion of liquid increases. This abrupt change in chemical potential perturbs the thermodynamic equilibrium between the intracellular and extracellular solutions. The cell responds by expressing water to reach the new equilibrium state, resulting in a competitionbetween heat transfer and mass transfer of water. At high cooling rates, thermodynamic equilibrium cannot be maintained by water transport because the rate at which the chemical potential in the extracellular solution is being lowered is
* Address correspondence to these authors.
+ MIT.
* Harvard Medical School.
5 Rutgers
University.
8756-7938/9 1/3007-0554$02.50/0
much greater than the rate at which water can diffuse out of the intracellular solution. The end result of this imbalance is intracellular ice formation (IIF), which is lethal to the cell (17). For sufficiently low cooling rates, thermodynamicequilibrium with the extracellular solution is achieved by dehydration. Damage at low cooling rates may be associated with the excessive dehydration (19,27), the attainment of high intracellular concentration (16,23), or mechanical effects (22,18). It is clear that in order to avoid or at least minimize the freezing injury, the cooling rate should be high enough to minimize exposure of the cells to high concentrations but low enough to avoid IIF. Freezing studies in the past on isolated hepatocytes have found that, after thawing from a freeze-thaw protocol, a large fraction of the cryopreserved hepatocytes metabolize fluorescence diacetate and attach to surfaces (2,8,9,12, 15) but metabolic functions such as gluconeogenesis, ureogenesis, and protein secretion are severely compromised (3, 9, 10, 15). In contrast to isolated hepatocytes, preliminary freezing studies with hepatocytes in this matrix configuration have been performed to determine the feasibility of freezing the cultured cells for later use in a bioreactor (13). Albumin secretion rates of up to 26% of the cultured hepatocytes were obtained for cells frozen following 7 days in culture. These preliminary results indicate the potential feasibility of cryopreservation and the need for further quantitative study of the freezing behavior of hepatocytes in a collagen matrix. The focus of this investigation is to determine the IIF characteristics of rat hepatocytes cultured in a collagen
0 199 1 American Chemical Society and American Institute of Chemical Engineers
mk
Biotechnol. Prog., 1991, Vol. 7, No. 6
555
PICTORIAL OVERVIEW OF CRYOMICROSCOPE RESEARCH SYSTEM
DEMONSTRATION
PRESSURE REGULATOR
VIDEO MONITOR (PANASONE BTS13U)N)
1 VIDEOCASSETTE RECORDER IPANASONIC AG-53001
VIDEOCAMERA INTSCl
METERING VALVE
17
-
Nq GASCYLINDER
Figure 1. Schematic representation of cryomicroscope research system from ref 4.
matrix using cryomicroscopy and to compare the IIF behavior in the matrix with that of isolated rat hepatocytes. This is the first study to quantify IIF for cells immobilized in an external matrix configuration.
Materials and Methods Preparation of Hepatocytes. Hepatocytes were isolated from 2-month-old female Lewis rats (200 g) using a modified procedure of Seglen (25) and described in detail elsewhere (13). Animals were anesthetized with ether. The liver was perfused with 300 mL of calcium-free Krebs Ringer bicarbonate buffer, containing 5.5 mM glucose and 20 mM HEPES, pH 7.4, at 50 mL/min. The liver was subsequently perfused with 0.05 7% collagenase (Type IV, Sigma, St. Louis, MO) solution containing 5 mM Ca2+ for 10 min. The resulting cell suspension is filtered and centrifuged at 47g for 5 min. The cell pellet was resuspended to 50 mL, and 12.5 mL of that cell solution was added to 10.8 mL of Percoll (Pharmacia, Piscataway, NJ) and 1.2 mL of lox concentrated Dulbecco’s modified Eagle’s medium (DMEM). This suspension was centrifuged at 50g for 5 min and resuspended in DMEM. Approximately 200-300 million cells were isolated from an 8-g liver with viability as determined using trypan blue exclusion ranging from 90 to 98%. The final cell suspension contained hepatocytes, with nonparenchymal cells constituting less than 1%of the total population. In order to facilitate manipulation of the cultured cells for cryomicroscopy, the culturing system described by Dunn et al. (5)was modified slightly. Twelve-millimetersquare glass cover slips (Bradford Scientific, Epping, NH) were placed on the bottom of a 60-m tissue culture dish (Beckton Dickson, Lincoln Park, NJ), which was then coated with 0.75 mL of rat tail collagen. After approximately 30 min for gelation of the collagen, 2 million hepatocytes were seeded per 60 mm dish. After one day of incubation in 10 percent C02, a second layer of collagen was spread over the single gel of collagen and the layer of cells. After 30 min to allow for gelation of the second collagen layer, 4 mL of complete medium was placed over
the cells in the matrix. Complete medium consisted of DMEM supplemented with 10% fetal bovine serum (Hazelton),0.5unit/mL insulin (Squibb,Princeton, NJ), 0.007 pg/mL glucagon (Lilly, Indianapolis, IN), 0.02 pg/mL epidermal growth factor (Collaborative Research, Bedford MA), 7.5 pg/mL hydrocortisone (Upjohn, Kalamazoo, MI), 200 units/mL penicillin, and streptomycin (Hazelton). Culture medium was changed daily. For the freezing studies, the cover slips with the hepatocytes in the collagen sandwich were cut out of the dish and placed on the freezing stage. This configuration minimized mechanical stress on the cells and maintained integrity of the collagen matrix prior to freezing. The composite system of cover slip, cells, and collagen was optically transparent and observable using transmission light microscopy. Freezing Experiments. The freezing experiments were performed using a cryomicroscope (cf. Figure 1)which consisted of a Zeiss general research microscope (Carl Zeiss, FRG) fitted witha special freezing stage described in detail elsewhere ( 4 ) . Images from the light microscope were transmitted by a camera (Ikegami ITC62, Japan) to a video cassette recorder (Panasonic Model 6300, Japan) and a monitor (Panasonic S1300) to allow direct visual observation and recording of the freezing experiments. A 16X phase-contrast objective (Carl Zeiss) was used with a 2 X optivar (Carl Zeiss) for magnification of the experiments. For this investigation, the freezing studies were performed on a thin convection stage. The sample to be frozen was placed on a composite window consisting of an optically transparent heating layer, thermocouple for sensing temperature, and a thin layer of glass to isolate the thermocouple from the freezing solution. The sample was cooled by chilled dry nitrogen gas flowing underneath the window. Power into the heating layer was controlled by a temperature control unit (Interface Techniques, Cambridge, MA) which allowed the temperature history of the sample to be precisely controlled and specified. The freezing solution for all experiments was Dulbecco’s modified Eagle’s medium (DMEM). For the studies
556
involving cryoprotectants, various concentrations of dimethyl sulfoxide(Me2SO) were added to the DMEM. Final concentrations of MezSO ranging from 0.25 to 2.0 M were used. MezSO was added to the cells in 0.25 M increments, allowing approximately 5 min for equilibration at each step. During a typical freezing experiment, the cells were cooled at a constant cooling rate from an initial temperature of approximately -1.5 "C to a final temperature of -50 "C. Cells in the higher concentrations of MezSO were cooled to -60 "C to ensure that the ice formation had been completed. Only the cells adjacent to the tip of the thermocouple (approximately 200 Mm) were used in order to ensure accurate determination of the actual temperature. There were potential errors in the temperature measurement at the higher cooling rates due to thermal limitations in the freezing stage. At 400 "C/min, there is a potential temperature error of approximately 1.5 "C. The potential for temperature measurement errors decreases linearly with cooling rate, and for cooling rates less than 25 "C/min, temperature measurement errors would be less than 0.1 "C. Seeding of the extracellular solution with a chilled needle was initiated at approximately -1.5 "C. Cells with internal ice exhibit an increase in opacity during the freezingprocess due to light scattering from small ice crystals, which can be observed and recorded. Due to the difficulty of recovering a sample from a cryomicroscope stage, no functional viability test was performed after a freezingprotocol. IIF was correlated with time and temperature by use of electronic superposition of these data on the video images recorded during the freezing experiment.
Results CoolingRate Dependence of IIF. As shown in Figure 2A, the maximum cumulative fraction of cells with IIF increased with cooling rate for hepatocytes frozen in DMEM after 1day and 7 days in culture. Cells in culture for 1 day and 7 days exhibited distinctive IIF behavior when compared to that of freshly isolated cells. The critical cooling rate ( B ~ odefined ), as the cooling rate at which 50 3' 4 of the cells undergo IIF, for cells in culture 1day as well as cells in culture 7 days was 70 and 15 "C/min, respectively. In contrast, freshly isolated cells were observed to have aB50 of 100 "C/min (11). The postthaw appearance of most of the cells with IIF indicated damage at the light microscope level (Le., cell boundaries and nuclei were difficult to distinguish, and in many cases, cells had lysed and disappeared). For a freezing solution of 0.5 M MezSO + DMEM, B50 was depressed from 70 to 50 "C/ min for cells in culture one day and from 15 to 10 "C/min for cells in culture 7 days (Figure 2B). Freshly isolated cells frozen in 0.5 M MezSO + DMEM did not exhibit a significant depression in B50. Average Temperature of IIF. For those cells which underwent IIF at a given cooling rate, it was possible to determine the average temperature for IIF (TIIF).For cells cultured for 1day and frozen in DMEM, the average value of TIIFfor cooling rates between 20 and 400 "C/min was -3.6 "C. Between 20 and 70 "C/min, TIIFwas independent of cooling rate (approximately -3.3 "C); it decreased to -5.7 "C (Figure 3A) by increasing the cooling rate to 400 "C/min. For cells frozen after 7 days in culture, the average value of TIIFfor cooling rates between 10 and 400 "C/min was -5.8 "C and TIIFwas independent of cooling rate for all coolingrates tested (cf. Figure 3A). TIIF for isolated cells frozen in DMEM was included in Figure 3 for comparison purposes (11).
Biotechnol. Prog., 1991, Vol. 7, No. 6 1.0
A
557
Biotechnol. hog.., 1991, Vol. 7, No. 6
:I %
Table I. Internal Ice Formation Parameters For Cultured Hepatocytes day in cooling rate, no. of culture OC/min CPIFO T m f cr,b°C cells DMEM
0
A
-2
L
DAY -1
20 35 50 70 100 130 400 5 10 20 35 50 80 100 400
1
-8
DAY-7
,,
,,
'/
,oul
a 3,
7-----
ISOLATED
-lot
lo
7
I
I
100
lo00
01
#
---_ I 100
10
I
0
0.100 0.577 0.579 1.00
-4.70 -5.97 f 1.68 -6.37 f 0.508 -5.86 f 0.300 -7.08 f 0.547 -5.09 f 0.419 -5.42 f 1.59
1.00
1.00 1.00
10 20 35 50 75 400 10 20 35 50 400
I
I
-2.97 f 1.16 -3.01 f 0.677 -2.51 f 0.800 -2.92 f 0.770 -4.07 f 1.77 -4.25 f 0.389 -5.74 f 1.59 -C
DMEM + 0.5 M MEzSO
1
ISOLATED
0.068 0.026 0.361 0.389 0.688 0.867 1.00 0.00
0.0
0.158 0.158 0.500 0.864 1.00
0.333 0.600
0.855 1.00 1.00
-4.27 f 0.150 -3.93 f 0.320 -5.50 f 0.480 -6.22 f 0.750 -5.54 f 1.98 -5.44 1.87 -5.27 f 0.890 -5.67 f 0.984 -5.60 f 0.656 -7.81 f 3.18
44 50 72 36 32 15 58 12 1 40 19 15 9 17 19 32 19 38 40 22 47 24 55 26 24 39
*
Cumulative fraction of cells with IIF. Average IIF temperature No internal ice formation was observed.
f standard deviation.
lo00
1.0
COOLING RATE ,"C/min
Figure 3. Dependence of average f standard error IIF temperature on cooling rate for cells cultured 24 h (day 1) (0) and cells cultured for 1 week (day 7) ( 0 ) .The freezing solution is (A) DMEM or (B) 0.5 M MezSO. Behavior of isolated hepatocytes (- - -) is also included from ref 11.
0.8
0.6
in MeZSO concentration was roughly equal to the depression of the homogeneous nucleation temperature of the solution as calculated using the method described in ref 24.
0.4
Discussion The cryomicroscopic characterization of the IIF behavior presented in this investigation is an important tool for describing the freezing behavior of hepatocytes in the collagen matrix. The results of this study show that there is a critical cooling rate above which IIF is observed in frozen cultured hepatocytes. Furthermore, our results indicate that the IIF characteristics (Bw, TIIF,and the depression in BMand TIIFwith the addition of MezSO) of hepatocytes cultured in a collagen double gel matrix are distinct from that observed with isolated hepatocytes and vary with time in culture. Therefore, freezing studies performed on isolated hepatocytes would not necessarily be predictive for cultured hepatocytes. Furthermore, freezing protocols determined for given culture conditions cannot be applied to another set of culture conditions even for the same cell type. It appears that all of these relevant parameters (cooling rate, final cooling temperature, concentration of MeZSO, and day in culture before freezing) must be precisely controlled and specified in each particular case to produce reliable and reproducible results. For cells frozen at less than 130 "C/min in DMEM after 1day in culture, IIF was observed a t relatively high temperatures (TIIF -3.7 "C). This freezing behavior
-
0.2
0
-5
-10
-15
- 20
TEMPERATURE, 'c Figure 4. Cumulative fraction of cells with ice as a function of temperature for cells cultured 1 week frozen at 400 OC/min in various concentrations of MeZSO. Each data point is a single cell.
contrasts with the measured values of TIIFfor other cells and embryos (Table 11) as well as isolated hepatocytes (TIIF -7.7 "c). Permeating cryoprotective agents (CPA's) such as MezSO have been shown to affect the IIF behavior of cells (7, 14, 20, 21, 26). For hepatocytes in the collagen matrix frozen in 0.5 M MezSO + DMEM after 1 day and 7 days in culture, B50 decreased from 70 to 50 "C/min and from 15 to 10 "C/min, respectively. In contrast, isolated cells frozen in 0.5 M MeZSO had the same B50 as cells frozen without MezSO present. The depression in BW with the addition of MezSO for the cells in the matrix may indicate
-
Blotechnol. Prog., 1991, Vol. 7, No. 6
558
0
-50
r
'
DAY-1
0.0
I
I
I
0.5
1 .o
1.5
2 .o
Me2S0 CONCENTRATION, M
*
Figure 5. IIF temperature standard error for cultured hepatocytes frozen after 1 day and 7 days in culture and freshly isolated cells (11) as a function of MezSO concentration. The solid line is the equilibrium freezingtemperature for the solution. The dashed line is the homogeneous nucleation temperature. Table 11. Internal Ice Formation Parameters for Other Biological Systems tem range T~mfor MezSO AT: biological system for !IF, O C IIF, O C concn O C ref human lymphocytes -15 to -25 -12 0M 7 hamster ova
-10to-18 -42to-76 mouse fibroblasts -4t0-18 -23to-37 mouse ova -8to-18 -30to-60 Drosophila melanogaster -3 to -28 embryos -5to-45 isolated rat hepatocytes -5 to -8 -14to-22
-50 -13 -57.1 -11.3 -29.2
10% v01
38
OM
26
1M OM
44
2M
18
20
-11
OM
-45 -13.3 -25 -7.7 -16.8
1M 0M
34
1M
12
28
OM 2M
14 21
I1
9
Depression in TIIFwith the addition of MezSO in the specified concentration.
a fundamental difference between the freezing behavior of the hepatocytes in the matrix and freshly isolated cells. There is the potential for an interaction between the CPA and the extracellular matrix or a cell-specific interaction or a combination of these effects which may account for the observed differences in freezing behavior. The addition of CPA has been observed to depress TIIF, which has been attributed to two different effects: (1)a colligative effect resulting from the depression in the homogeneous nucleation temperature due to the addition of solute and (2) a cell-specific interaction with the CPA, which has been observed experimentally in several different cell types (7,14,26). For both isolated and cultured hepatocytes, TIIFwas depressed approximately 8 "C with the addition of 2 M MezSO, which corresponded roughly to the depression in the homogeneous nucleation temperature (cf. Figure 5). In contrast, various cells and embryos listed in Table I1 exhibited 18-44 OC depression in TIIFwhen frozen in solutions with 2 M MezSO. Thus, increasing the concentration of CPA for hepatocytes does not provide a significant benefit in terms of depression in Tm.Significantdepression in TIIFwould allow the cultures to be frozen a t a higher cooling rate to a lower temperature, therefore reducing the exposure time to the potentially damaging extracellular concentrations which develop during freezing.
An extracellular matrix represents a distinctly different extracellular environment from that experienced by cells in suspension. A difference in the extracellular environment has implications not only in terms of characterizing the external environment during the freezing process but also for the properties of the cultured cell. The doublegel culturing system has been shown to affect cellular level functions such as protein synthesis and secretion, detoxification, and bile salt production and secretion (5, 6). Changes in cell function from a suspended state to an attached state may reflect differences in the membrane structure which may be manifested in changes in the hydraulic permeability. In addition to structural differences in the membrane, hepatocytes in the matrix are attached to the extracellular matrix and therefore volumetric changes are limited. The manner in which the attachment of the plasma membrane to the collagen gel affects the water transport from the cells is unknown. Finally, the diffusionof water out of the cell may be affected by the collagen gel, although it is rather unlikely that the resistance of the gel (
554
Intracellular Ice Formation during the Freezing of Hepatocytes Cultured in a Double Collagen Gel A. Hubel,?M. Toner,+*$ E. G. Cravalho,t M. L. Yarmush,*J*§ and R. G. Tompkins*** Department of Mechanical Engineering, Massachusetts Institute of Technology, Cambridge, Massachusetts 02139, Department of Surgery, Massachusetts General Hospital, Harvard Medical School, Boston, Massachusetts 02114, and Department of Chemical and Biochemical Engineering, Rutgers University, Piscataway, New Jersey 08855
During freezing, intracellular ice formation (IIF)has been correlated with loss in viability for a wide variety of biological systems. Hence, determination of IIF characteristics is essential in the development of an efficient methodology for cryopreservation. In this study, IIF characteristics of hepatocytes cultured in a collagen matrix were determined using cryomicroscopy. Four factors influenced the IIF behavior of the hepatocytes in the matrix: cooling rate, final cooling temperature, concentration of MeZSO, and time in culture prior to freezing. The maximum cumulative fraction of cells with IIF increased with increasing cooling rate. For cultured cells frozen in Dulbecco's modified Eagle's was 70 medium (DMEM), the cooling rate for which 50% of the cells formed ice (BM) OC/min for cells frozen after 1day in culture and decreased to 15 "C/min for cells frozen after 7 days in culture. When cells were frozen in a 0.5 M MezSO DMEM solution, the value of B50 decreased from 70 to 50 "C/min for cells in culture for 1day and from 15 to 10 OC/min for cells in culture for 7 days. The value of the average temperature for cultured cells was only slightly depressed by the addition of MezSO for IIF (TIIF) when compared to the IIF behavior of other cell types. The results of this study indicate that the presence of the collagen matrix alters significantly the IIF characteristics of hepatocytes. Thus freezing studies using hepatocytes in suspension are not useful in predicting the freezing behavior of hepatocytes cultured in a collagen matrix. Furthermore, the weak effect of MeZSO on IIF characteristics implies that lower concentrations of MezSO (0.5 M) may be just as effective in preservingviability. Finally, the value of Bm measured in this study indicates that cooling rates nearly an order of magnitude faster than those previously investigated could be used for cryopreservation of the hepatocytes in a collagen gel.
+
Introduction The establishment of efficient hepatocyte storage protocols is important for a variety of clinical and biotechnological applications including the development of extracorporeal liver support devices and liver cell transplantation. At present, the task of maintaining a supply of hepatocytes through cryopreservation appears essential because adult hepatocytes do not readily proliferate in culture. The capability of freezing liver cells would thus presumably allow the pooling of donor cells and ensure a stable supply of hepatocytes. During the freezing process, the cell in solution is subjected to temperatures below the equilibrium freezing temperature of the solution. As ice forms in the extracellular solution, solute is rejected from the solid phase, and correspondingly, the chemical potential in the unfrozen portion of liquid increases. This abrupt change in chemical potential perturbs the thermodynamic equilibrium between the intracellular and extracellular solutions. The cell responds by expressing water to reach the new equilibrium state, resulting in a competitionbetween heat transfer and mass transfer of water. At high cooling rates, thermodynamic equilibrium cannot be maintained by water transport because the rate at which the chemical potential in the extracellular solution is being lowered is
* Address correspondence to these authors.
+ MIT.
* Harvard Medical School.
5 Rutgers
University.
8756-7938/9 1/3007-0554$02.50/0
much greater than the rate at which water can diffuse out of the intracellular solution. The end result of this imbalance is intracellular ice formation (IIF), which is lethal to the cell (17). For sufficiently low cooling rates, thermodynamicequilibrium with the extracellular solution is achieved by dehydration. Damage at low cooling rates may be associated with the excessive dehydration (19,27), the attainment of high intracellular concentration (16,23), or mechanical effects (22,18). It is clear that in order to avoid or at least minimize the freezing injury, the cooling rate should be high enough to minimize exposure of the cells to high concentrations but low enough to avoid IIF. Freezing studies in the past on isolated hepatocytes have found that, after thawing from a freeze-thaw protocol, a large fraction of the cryopreserved hepatocytes metabolize fluorescence diacetate and attach to surfaces (2,8,9,12, 15) but metabolic functions such as gluconeogenesis, ureogenesis, and protein secretion are severely compromised (3, 9, 10, 15). In contrast to isolated hepatocytes, preliminary freezing studies with hepatocytes in this matrix configuration have been performed to determine the feasibility of freezing the cultured cells for later use in a bioreactor (13). Albumin secretion rates of up to 26% of the cultured hepatocytes were obtained for cells frozen following 7 days in culture. These preliminary results indicate the potential feasibility of cryopreservation and the need for further quantitative study of the freezing behavior of hepatocytes in a collagen matrix. The focus of this investigation is to determine the IIF characteristics of rat hepatocytes cultured in a collagen
0 199 1 American Chemical Society and American Institute of Chemical Engineers
mk
Biotechnol. Prog., 1991, Vol. 7, No. 6
555
PICTORIAL OVERVIEW OF CRYOMICROSCOPE RESEARCH SYSTEM
DEMONSTRATION
PRESSURE REGULATOR
VIDEO MONITOR (PANASONE BTS13U)N)
1 VIDEOCASSETTE RECORDER IPANASONIC AG-53001
VIDEOCAMERA INTSCl
METERING VALVE
17
-
Nq GASCYLINDER
Figure 1. Schematic representation of cryomicroscope research system from ref 4.
matrix using cryomicroscopy and to compare the IIF behavior in the matrix with that of isolated rat hepatocytes. This is the first study to quantify IIF for cells immobilized in an external matrix configuration.
Materials and Methods Preparation of Hepatocytes. Hepatocytes were isolated from 2-month-old female Lewis rats (200 g) using a modified procedure of Seglen (25) and described in detail elsewhere (13). Animals were anesthetized with ether. The liver was perfused with 300 mL of calcium-free Krebs Ringer bicarbonate buffer, containing 5.5 mM glucose and 20 mM HEPES, pH 7.4, at 50 mL/min. The liver was subsequently perfused with 0.05 7% collagenase (Type IV, Sigma, St. Louis, MO) solution containing 5 mM Ca2+ for 10 min. The resulting cell suspension is filtered and centrifuged at 47g for 5 min. The cell pellet was resuspended to 50 mL, and 12.5 mL of that cell solution was added to 10.8 mL of Percoll (Pharmacia, Piscataway, NJ) and 1.2 mL of lox concentrated Dulbecco’s modified Eagle’s medium (DMEM). This suspension was centrifuged at 50g for 5 min and resuspended in DMEM. Approximately 200-300 million cells were isolated from an 8-g liver with viability as determined using trypan blue exclusion ranging from 90 to 98%. The final cell suspension contained hepatocytes, with nonparenchymal cells constituting less than 1%of the total population. In order to facilitate manipulation of the cultured cells for cryomicroscopy, the culturing system described by Dunn et al. (5)was modified slightly. Twelve-millimetersquare glass cover slips (Bradford Scientific, Epping, NH) were placed on the bottom of a 60-m tissue culture dish (Beckton Dickson, Lincoln Park, NJ), which was then coated with 0.75 mL of rat tail collagen. After approximately 30 min for gelation of the collagen, 2 million hepatocytes were seeded per 60 mm dish. After one day of incubation in 10 percent C02, a second layer of collagen was spread over the single gel of collagen and the layer of cells. After 30 min to allow for gelation of the second collagen layer, 4 mL of complete medium was placed over
the cells in the matrix. Complete medium consisted of DMEM supplemented with 10% fetal bovine serum (Hazelton),0.5unit/mL insulin (Squibb,Princeton, NJ), 0.007 pg/mL glucagon (Lilly, Indianapolis, IN), 0.02 pg/mL epidermal growth factor (Collaborative Research, Bedford MA), 7.5 pg/mL hydrocortisone (Upjohn, Kalamazoo, MI), 200 units/mL penicillin, and streptomycin (Hazelton). Culture medium was changed daily. For the freezing studies, the cover slips with the hepatocytes in the collagen sandwich were cut out of the dish and placed on the freezing stage. This configuration minimized mechanical stress on the cells and maintained integrity of the collagen matrix prior to freezing. The composite system of cover slip, cells, and collagen was optically transparent and observable using transmission light microscopy. Freezing Experiments. The freezing experiments were performed using a cryomicroscope (cf. Figure 1)which consisted of a Zeiss general research microscope (Carl Zeiss, FRG) fitted witha special freezing stage described in detail elsewhere ( 4 ) . Images from the light microscope were transmitted by a camera (Ikegami ITC62, Japan) to a video cassette recorder (Panasonic Model 6300, Japan) and a monitor (Panasonic S1300) to allow direct visual observation and recording of the freezing experiments. A 16X phase-contrast objective (Carl Zeiss) was used with a 2 X optivar (Carl Zeiss) for magnification of the experiments. For this investigation, the freezing studies were performed on a thin convection stage. The sample to be frozen was placed on a composite window consisting of an optically transparent heating layer, thermocouple for sensing temperature, and a thin layer of glass to isolate the thermocouple from the freezing solution. The sample was cooled by chilled dry nitrogen gas flowing underneath the window. Power into the heating layer was controlled by a temperature control unit (Interface Techniques, Cambridge, MA) which allowed the temperature history of the sample to be precisely controlled and specified. The freezing solution for all experiments was Dulbecco’s modified Eagle’s medium (DMEM). For the studies
556
involving cryoprotectants, various concentrations of dimethyl sulfoxide(Me2SO) were added to the DMEM. Final concentrations of MezSO ranging from 0.25 to 2.0 M were used. MezSO was added to the cells in 0.25 M increments, allowing approximately 5 min for equilibration at each step. During a typical freezing experiment, the cells were cooled at a constant cooling rate from an initial temperature of approximately -1.5 "C to a final temperature of -50 "C. Cells in the higher concentrations of MezSO were cooled to -60 "C to ensure that the ice formation had been completed. Only the cells adjacent to the tip of the thermocouple (approximately 200 Mm) were used in order to ensure accurate determination of the actual temperature. There were potential errors in the temperature measurement at the higher cooling rates due to thermal limitations in the freezing stage. At 400 "C/min, there is a potential temperature error of approximately 1.5 "C. The potential for temperature measurement errors decreases linearly with cooling rate, and for cooling rates less than 25 "C/min, temperature measurement errors would be less than 0.1 "C. Seeding of the extracellular solution with a chilled needle was initiated at approximately -1.5 "C. Cells with internal ice exhibit an increase in opacity during the freezingprocess due to light scattering from small ice crystals, which can be observed and recorded. Due to the difficulty of recovering a sample from a cryomicroscope stage, no functional viability test was performed after a freezingprotocol. IIF was correlated with time and temperature by use of electronic superposition of these data on the video images recorded during the freezing experiment.
Results CoolingRate Dependence of IIF. As shown in Figure 2A, the maximum cumulative fraction of cells with IIF increased with cooling rate for hepatocytes frozen in DMEM after 1day and 7 days in culture. Cells in culture for 1 day and 7 days exhibited distinctive IIF behavior when compared to that of freshly isolated cells. The critical cooling rate ( B ~ odefined ), as the cooling rate at which 50 3' 4 of the cells undergo IIF, for cells in culture 1day as well as cells in culture 7 days was 70 and 15 "C/min, respectively. In contrast, freshly isolated cells were observed to have aB50 of 100 "C/min (11). The postthaw appearance of most of the cells with IIF indicated damage at the light microscope level (Le., cell boundaries and nuclei were difficult to distinguish, and in many cases, cells had lysed and disappeared). For a freezing solution of 0.5 M MezSO + DMEM, B50 was depressed from 70 to 50 "C/ min for cells in culture one day and from 15 to 10 "C/min for cells in culture 7 days (Figure 2B). Freshly isolated cells frozen in 0.5 M MezSO + DMEM did not exhibit a significant depression in B50. Average Temperature of IIF. For those cells which underwent IIF at a given cooling rate, it was possible to determine the average temperature for IIF (TIIF).For cells cultured for 1day and frozen in DMEM, the average value of TIIFfor cooling rates between 20 and 400 "C/min was -3.6 "C. Between 20 and 70 "C/min, TIIFwas independent of cooling rate (approximately -3.3 "C); it decreased to -5.7 "C (Figure 3A) by increasing the cooling rate to 400 "C/min. For cells frozen after 7 days in culture, the average value of TIIFfor cooling rates between 10 and 400 "C/min was -5.8 "C and TIIFwas independent of cooling rate for all coolingrates tested (cf. Figure 3A). TIIF for isolated cells frozen in DMEM was included in Figure 3 for comparison purposes (11).
Biotechnol. Prog., 1991, Vol. 7, No. 6 1.0
A
557
Biotechnol. hog.., 1991, Vol. 7, No. 6
:I %
Table I. Internal Ice Formation Parameters For Cultured Hepatocytes day in cooling rate, no. of culture OC/min CPIFO T m f cr,b°C cells DMEM
0
A
-2
L
DAY -1
20 35 50 70 100 130 400 5 10 20 35 50 80 100 400
1
-8
DAY-7
,,
,,
'/
,oul
a 3,
7-----
ISOLATED
-lot
lo
7
I
I
100
lo00
01
#
---_ I 100
10
I
0
0.100 0.577 0.579 1.00
-4.70 -5.97 f 1.68 -6.37 f 0.508 -5.86 f 0.300 -7.08 f 0.547 -5.09 f 0.419 -5.42 f 1.59
1.00
1.00 1.00
10 20 35 50 75 400 10 20 35 50 400
I
I
-2.97 f 1.16 -3.01 f 0.677 -2.51 f 0.800 -2.92 f 0.770 -4.07 f 1.77 -4.25 f 0.389 -5.74 f 1.59 -C
DMEM + 0.5 M MEzSO
1
ISOLATED
0.068 0.026 0.361 0.389 0.688 0.867 1.00 0.00
0.0
0.158 0.158 0.500 0.864 1.00
0.333 0.600
0.855 1.00 1.00
-4.27 f 0.150 -3.93 f 0.320 -5.50 f 0.480 -6.22 f 0.750 -5.54 f 1.98 -5.44 1.87 -5.27 f 0.890 -5.67 f 0.984 -5.60 f 0.656 -7.81 f 3.18
44 50 72 36 32 15 58 12 1 40 19 15 9 17 19 32 19 38 40 22 47 24 55 26 24 39
*
Cumulative fraction of cells with IIF. Average IIF temperature No internal ice formation was observed.
f standard deviation.
lo00
1.0
COOLING RATE ,"C/min
Figure 3. Dependence of average f standard error IIF temperature on cooling rate for cells cultured 24 h (day 1) (0) and cells cultured for 1 week (day 7) ( 0 ) .The freezing solution is (A) DMEM or (B) 0.5 M MezSO. Behavior of isolated hepatocytes (- - -) is also included from ref 11.
0.8
0.6
in MeZSO concentration was roughly equal to the depression of the homogeneous nucleation temperature of the solution as calculated using the method described in ref 24.
0.4
Discussion The cryomicroscopic characterization of the IIF behavior presented in this investigation is an important tool for describing the freezing behavior of hepatocytes in the collagen matrix. The results of this study show that there is a critical cooling rate above which IIF is observed in frozen cultured hepatocytes. Furthermore, our results indicate that the IIF characteristics (Bw, TIIF,and the depression in BMand TIIFwith the addition of MezSO) of hepatocytes cultured in a collagen double gel matrix are distinct from that observed with isolated hepatocytes and vary with time in culture. Therefore, freezing studies performed on isolated hepatocytes would not necessarily be predictive for cultured hepatocytes. Furthermore, freezing protocols determined for given culture conditions cannot be applied to another set of culture conditions even for the same cell type. It appears that all of these relevant parameters (cooling rate, final cooling temperature, concentration of MeZSO, and day in culture before freezing) must be precisely controlled and specified in each particular case to produce reliable and reproducible results. For cells frozen at less than 130 "C/min in DMEM after 1day in culture, IIF was observed a t relatively high temperatures (TIIF -3.7 "C). This freezing behavior
-
0.2
0
-5
-10
-15
- 20
TEMPERATURE, 'c Figure 4. Cumulative fraction of cells with ice as a function of temperature for cells cultured 1 week frozen at 400 OC/min in various concentrations of MeZSO. Each data point is a single cell.
contrasts with the measured values of TIIFfor other cells and embryos (Table 11) as well as isolated hepatocytes (TIIF -7.7 "c). Permeating cryoprotective agents (CPA's) such as MezSO have been shown to affect the IIF behavior of cells (7, 14, 20, 21, 26). For hepatocytes in the collagen matrix frozen in 0.5 M MezSO + DMEM after 1 day and 7 days in culture, B50 decreased from 70 to 50 "C/min and from 15 to 10 "C/min, respectively. In contrast, isolated cells frozen in 0.5 M MeZSO had the same B50 as cells frozen without MezSO present. The depression in BW with the addition of MezSO for the cells in the matrix may indicate
-
Blotechnol. Prog., 1991, Vol. 7, No. 6
558
0
-50
r
'
DAY-1
0.0
I
I
I
0.5
1 .o
1.5
2 .o
Me2S0 CONCENTRATION, M
*
Figure 5. IIF temperature standard error for cultured hepatocytes frozen after 1 day and 7 days in culture and freshly isolated cells (11) as a function of MezSO concentration. The solid line is the equilibrium freezingtemperature for the solution. The dashed line is the homogeneous nucleation temperature. Table 11. Internal Ice Formation Parameters for Other Biological Systems tem range T~mfor MezSO AT: biological system for !IF, O C IIF, O C concn O C ref human lymphocytes -15 to -25 -12 0M 7 hamster ova
-10to-18 -42to-76 mouse fibroblasts -4t0-18 -23to-37 mouse ova -8to-18 -30to-60 Drosophila melanogaster -3 to -28 embryos -5to-45 isolated rat hepatocytes -5 to -8 -14to-22
-50 -13 -57.1 -11.3 -29.2
10% v01
38
OM
26
1M OM
44
2M
18
20
-11
OM
-45 -13.3 -25 -7.7 -16.8
1M 0M
34
1M
12
28
OM 2M
14 21
I1
9
Depression in TIIFwith the addition of MezSO in the specified concentration.
a fundamental difference between the freezing behavior of the hepatocytes in the matrix and freshly isolated cells. There is the potential for an interaction between the CPA and the extracellular matrix or a cell-specific interaction or a combination of these effects which may account for the observed differences in freezing behavior. The addition of CPA has been observed to depress TIIF, which has been attributed to two different effects: (1)a colligative effect resulting from the depression in the homogeneous nucleation temperature due to the addition of solute and (2) a cell-specific interaction with the CPA, which has been observed experimentally in several different cell types (7,14,26). For both isolated and cultured hepatocytes, TIIFwas depressed approximately 8 "C with the addition of 2 M MezSO, which corresponded roughly to the depression in the homogeneous nucleation temperature (cf. Figure 5). In contrast, various cells and embryos listed in Table I1 exhibited 18-44 OC depression in TIIFwhen frozen in solutions with 2 M MezSO. Thus, increasing the concentration of CPA for hepatocytes does not provide a significant benefit in terms of depression in Tm.Significantdepression in TIIFwould allow the cultures to be frozen a t a higher cooling rate to a lower temperature, therefore reducing the exposure time to the potentially damaging extracellular concentrations which develop during freezing.
An extracellular matrix represents a distinctly different extracellular environment from that experienced by cells in suspension. A difference in the extracellular environment has implications not only in terms of characterizing the external environment during the freezing process but also for the properties of the cultured cell. The doublegel culturing system has been shown to affect cellular level functions such as protein synthesis and secretion, detoxification, and bile salt production and secretion (5, 6). Changes in cell function from a suspended state to an attached state may reflect differences in the membrane structure which may be manifested in changes in the hydraulic permeability. In addition to structural differences in the membrane, hepatocytes in the matrix are attached to the extracellular matrix and therefore volumetric changes are limited. The manner in which the attachment of the plasma membrane to the collagen gel affects the water transport from the cells is unknown. Finally, the diffusionof water out of the cell may be affected by the collagen gel, although it is rather unlikely that the resistance of the gel (
