Factors to consider in the assessment of viability of cryopreserved islets of Langerhans

20, 151-160 (1983)

CRYOBIOLOGY

Factors

to Consider

in the Assessment of Viability Islets of Langerhans

of Cryopreserved

DENNIS B. MCKAY’ AND ARMAND M. KAROW, JR. Department of Pharmacology, University of Cambridge, Cambridge, United Kittgdorn, Department of Pharmacology. Medical College of Georgia, Augusta, Georgia With the development of techniques for the isolation and transplantation of the pancreatic islets of Langerhans, research has been directed toward low-temperature storage of isolated islets as a means of preservation. Low-temperature preservation (cryopreservation) of pancreatic islets would allow storage of sufficient quantities of this tissue for transplantation and would permit histocompatibility matching. While the literature contains reports of “successful” islet cryopreservation (1,2,4, 7-9) their evidence engenders an unwarranted enthusiasm; islet cryopreservation is difficult and the techniques are far from perfected. For successful cryopreservation several questions must be considered. What cryoprotectant should be used? Does the cryoprotectant exert any toxic effects? What cooling rates produce optimum survival? Can islet cells recover from the insult of low-temperature preservation? If lowtemperature preservation is to be an answer for long-term storage of isolated islets of Langerhans, then much work must be done to assess the feasibility of such a solution. Few studies have examined the problem and the results are controversial. The studies presented here were designed to determine the cryoprotectant and toxic effects of dimethyl sulfoxide (Me,SO) on islet Received November 12, 1982; accepted November 15, 1982. ’ Present address: Room 904-Howard, Memorial Sloan-Kettering Cancer Center, 1275 York Ave., New York, N.Y. 10021.

and

function, to minimize any toxic manifestations of Me&SO, and to characterize the effects of low-temperature preservation on islet function. The results of these studies address only a few of the many questions that need to be answered before clinical application of cryopreserved islet transplantation occurs. MATERIALS

AND

METHODS

Islet isolation. The islets of Langerhans were isolated from 250-350 g male Sprague-Dawley rats by the method of Lacy and Kostianovsky (5). Briefly, the rats were anesthetized with sodium pentobarbital intraperitoneally. After exposing the pancreas through a midline abdominal incision, the bile duct was occluded distally at the duodenum and then cannulated proximally at the liver. Hanks’ balanced salt solution (20 ml) was injected via the cannula to distend the pancreas and facilitate its removal. Two pancreases were pooled, minced, and digested with collagenase (Type IV, Worthington, Freehold, N.J.) for 10 min at 37°C with vigorous agitation. The amount of collagenase (3.5-4.5 mg/ml) was adjusted according to the specific activity of the enzyme which varied from 17.5-225 U/mg. After washing, the islets were separated from the pancreatic debris by handpicking under a dissecting microscope. Static incubations. For the experiments investigating the effects of Me,SO on insulin release, Me&SO was removed prior to the islets being stimulated in static incuba151 OOII-2240183 $3.00 Copyright 0 1983 by Academic Press, Inc. All rights of reproduction in any form reserved.

152

MC KAY AND

tions. In these experiments after Me&SO treatment five islets were placed into small conical incubation vials (autoanalyzer cups, 2 ml) with 400 ~1 of a modified Kreb’s solution (7). The solution had been equilibrated with 5% CO, in oxygen, and the pH was adjusted to 7.4 prior to each experiment. Each of the vials was then capped and placed into a 37°C Dubnoff metabolic shaker bath. After 60 min aliquots of the incubation media were removed from each vial and frozen for subsequent assay of insulin content. Islet perifusion. The perifusion apparatus used in the experiments has been previously described by us (7). The perifusion apparatus consists of a modified 13-min Swinney filter chamber, connected to a peristaltic pump by 60 cm of tubing (I.D. 76 mm). The chamber is submerged directly into a reservoir containing a modified Kreb’s solution. The pump draws the medium into the chamber past the islets at a flow rate of approximately 1 ml/min. Samples are collected continuously over specified time intervals and frozen for subsequent determination of insulin content. MezSO treatment. The procedure for Me,SO treatment for the experiments designed to investigate the effects of the cryoprotectant on insulin release has been previously described (7). Briefly, islets were treated with one of several concentrations of Me&SO for 60 min at either 37 or 10°C. After removal of Me&SO the islets were then placed into static incubation or perifused. Me,SO administration was either by a one-step (rapid) procedure, whereby islets were placed directly into the Me&SOcontaining media, or by a gradual (stepwise) procedure, whereby the Me,SO concentration was increased at one-fourth increments (i.e., 0.25 x final concentration) at 5-min intervals by medium replacement. Culturing conditions. After isolation, the islets were placed into culture for 24 hr prior to being used experimentally. Culturing was done in a 37°C incubator in which

KAROW

the air was humidified and enriched with 5% CO,. The culture medium was CMRL1066 supplemented with 10% fetal calf serum. In those experiments where the culturing time was increased to 20-24 hr, gentamicin sulfate (10 mg/liter) was added and the glucose concentration was increased to 180 mg/dl. Insulin assay. Insulin content was determined by radioimmunoassay (12) using porcine insulin (Eli Lilly, Indianapolis, Ind.) as the standard and porcine [lz51]insulin (New England Nuclear, Boston, Mass.) as the tracer.

FIG. 1. Apparatus for slowly cooling the islets of Langerhans. Five beakers (250,400,600,800, and 1000 ml) were stacked one inside the other and suspended 2-3 cm over liquid nitrogen contained in an evacuated dewar. The islets in 2-ml cryotubes were placed around the periphery of a 150 ml beaker containing approximately 20 ml of 95% ethanol. This beaker containing the islets was placed inside the smallest of the five beakers. Cooling was achieved by conduction through the beaker walls. A copper-constantan thermocouple inserted into a “dummy” cryovial was used to monitor temperature.

FACTORS IN THE ASSESSMENT OF CRYOPRESERVED ISLETS

0 14

1.4

07

153

2.8

Me2SO(M)

FIG. 2. Insulin release from islets after pretreatment with Me,SO using the rapid protocol. Insulin release from islets after pretreatment with Me,SO using the rapid protocol at either 10 (solid bar) or 37°C (striped bar) or the stepwise protocol at 37°C (open bar). Results are expressed as a percentage of insulin released relative to nontreated controls. The absolute values for each of the control and experimental groups are found in Table 1. Values represent means k SEM (n = 15-38).

Slow cooling system. The slow cooling apparatus (Fig. l), originally described by Fahy (3), consists of a stack set of Pyrex beakers (150- lOOO-mlcapacity) suspended in a liquid nitrogen dewar. The liquid nitrogen level was maintained 2- 3 cm below the stacked beakers. Cooling rates were monitored by a copper-constantan thermocouple which was inserted into a “dummy” cryotube (2-ml capacity, Nunc, Denmark). The freezing protocol has been described

by us (7). The Me,SO-treated islets, 50 islets/O.2 ml/tube, were cooled slowly (0.5-0.6Wmin) to -6°C seeded, and then cooled at a rate of 0.25-0.35”CYmin to -75°C. The islets were stored at that temperature for 20-24 hr and then thawed (2-3Wmin) by suspending in an evacuated dewar (22°C). The addition and dilution of Me,SO was accomplished by the stepwise protocol. Insulin synthesis. The protein synthesiz-

TABLE 1 Quantitative Data on the Effects of Me$O, Treatment Protocol, and Temperature on Glucose-Induced Insulin Release” Insulin release (FIJ insulin/islet/60 min)

[Me801

37°C Me&SO pretreatment (rapid)

10°C Mea0 pretreatment (rapid)

37°C Mea0 pretreatment (serial)

0.14 M Control 0.7 M Control 1.4M Control 2.8 M Control 5.6 M Control

75.1 k 4.3 106.2 k 4.2 60.8 + 3.1 99.2 k 5.0 62.1 k 2.2 102.9 2 5.0 38.8 k 2.8 108.5 + 4.8 12.4 i 1.0 112.8 r 4.9

91.1 +- 7.7 97.5 ” 8.8 134.5 + 3.2 132.5 ” 6.3 93.3 ” 5.8 96.0 2 6.4 68.3 t 4.0 113.7 k 4.9 6.7 ?I 0.5 105.0 k 4.5

102.7 k 4.1 110.5 + 3.6 97.1 + 4.8 110.5 k 3.6 94.9 I?-4.7 110.5 2 3.6 30.1 + 1.9 110.5 2 3.6

n Prior to glucose stimulation in static incubations, islets were pretreated with one of several concentrations of Me$O for 60 min by the rapid or serial protocol at either 37 or 10°C. Results are expressed as means + SEM (n = 15-38 groups of islets).

154

MC KAY AND KAROW 100 mg/dl

glucQ*

100 mgldl

glucose

400 mg/dl

glucose

J 200 0 %

Z

150

s 2 3 a

loo

50

250

4

200

& 0 z 3

150

2 z 3 a

400

mgtdl

glucose

;

no

glucose

loo

50

400 mg dl glucose

100 mg dl glucose

.

10

I

20

I

30

I

40

I

50

6C

1

70

L SO

I

.

5

I

90

100

110

120

TIME MINI FIG. 3. Approximately 125 islets/chamber per experiment were initially perifused for 40 min with a 100 mg/dl glucose medium to establish the basal release rate. At f = 40 the islets were stimulated with 400 mg/dl glucose. During the final 40-min period, the islets were perifused in a glucose-free medium to establish the glucose dependency of the stimulated response. Upper panel. The dynamic insulin response from untreated control islets (n = 4). Middle panel. The dynamic insulin response from islets pretreated (rapid addition) with 1.4 M Me,SO at 37°C. After Me,SO washout (rapid dilution), approximately 125 islets/chamber per experiment were perifused at 37°C (n = 4). Lower panel. The dynamic insulin response from islets pretreated (rapid addition) with 1.4 M Me,SO at 10°C. After Me,SO washout (rapid dilution), approximately 125islets/chamber per experiment were perifused at 37°C (n = 7). From Ref. (6).

155

FACTORS IN THE ASSESSMENT OF CRYOPRESERVED ISLETS TABLE 2 Quantitative Data on the Dynamic Release of Insulin from Isolated Perifused Islets of Langerhans after Experimental Treatmenta

Dynamic response parameters* Mean basal release rate (pU insuliniminlgroup) Peak response, 1st phase (FLUinsulin/group) Peak response, 2nd phase (pU insulin/group) Total insulin, 1st phase (t = 40-45) (FLJ insulin/group) Total insulin, 2nd phase (t = 46-80) (PU insulin/group)

Control 37°C (12= 4)

MeSO treated (1.4 M) (n = 4)

10°C (n = 7)

7.9 k 3.0

28.3 e 2.6’

11.5 2 4.4

174.6 t 11.6

91.5 IfI 9.4’

130.9 k 12.8

245.8 5 7.7

192.9 + 12.3’

228.9 k 8.9

611.0 + 9.4

289.3 t 25.8<

350.4 k 48.2

6992.0 k 363.6

5245.5 k 364.7’

5786.4 + 423.5

37°C

n Prior to perifusion, islets were incubated in the presence of 1.4 M MeSO at either 37 or 10°C. After 60 min, the Me30 was removed through rapid dilution. Nontreated control islets were incubated simultaneously at 37°C and handled in an identical manner. For each experiment after MeSO washout approximately 125 islets were perifused at 37°C during which time the data was collected. Results are expressed as mean + SEM. * The mean basal release rate refers to the rate of insulin release during the first 40-min period of perifusion with 100mg/dl glucose (nonstimulatory). The peak responses refer to the maximum rate of insulin release during either the tirst phase (t = 40-45) or second phase (t = 46-80) of glucose-induced insulin release. c Significant differences from controls; P < 0.05. (From Ref (6).)

ing capacity of the islets was determined by the method of Schatz et al. (10) as modified by us (7). Briefly, islets were incubated for 180 min in an amino acid-containing medium supplemented with 300 mg/dl glucose. During the 180-min incubation, aliquots were removed at 60-min intervals for insulin assay. After the 180-min incubation, the islet cells were disrupted and their protein content precipitated with 10% TCA. Separation of the islet protein was done on a Sephadex G-50 fine column. RESULTS

Islet Function After MezSO Pretreatment To determine the effects of Me,SO on islet function in the absence of freezing, islets were treated with one of several concentrations of Me,SO for 60 min at 37 and

10°C. The effects of Me,SO pretreatment on subsequent glucose-induced insulin release are illustrated in Fig. 2 and Table 1. Using the one-step or rapid treatment protocol at 37”C, there is significant depression (P < 0.05) of subsequent stimulated insulin release at all concentrations tested. Lowering the temperature of exposure significantly improved (P < 0.05) the insulin secretory response up to a Me,SO concentration of 2.8 M (Fig. 2). The effects of Me,SO on the dynamic pattern of insulin release is illustrated in Fig. 3 and Table 2. The biphasic pattern of insulin release is still intact after 1.4 M Me,SO pretreatment at both 37 and 10°C. However, there are some quantitative effects on insulin release (Table 2). With pretreatment at 37°C there is a significant increase (P < 0.05) in the mean basal release rate. We also find sig-

156

MC

00 mg/dl

Glucose

KAY

AND

KAROW

mq/dl

1 400

Glucose

!

mq/dl

100

Glucose

-

MelSO-free control

-

1.4

M Me2S0

pretreated

10

20

30

40

50

60

70

80

90

100

110

120

TIME(min)

FIG. 4. Comparison of the effects of 1.4 M Me,SO pretreatment (serial) at 4°C to Me,SO-free pretreated controls. Islets from one rat were divided into two groups. Islets in one group were pretreated with 1.4M Me,SO (serial protocol) for 60 min at 4°C. Islets in the second group were pretreated in identical manner except for the omission of Me,SO from the incubation medium. After washout, both groups were perifused simultaneously at 37°C (125 islets/chamber).

niticant depression (P < 0.05) of insulin release in both phases of the biphasic response. With pretreatment at 10°C the only significant effect is a decrease in the total amount of insulin released during the first phase of stimulation. Gradual addition and dilution of Me,SO

reduces its adverse effects when this was done at 37°C. A significantly improved (P < 0.05) insulin response resulted when compared to that of the rapid protocol up to 1.4 M Me,SO (Fig. 2). Using the information from these studies, the condition that would result in a

TABLE 3 Insulin Release from Islets Cooled at a Rate of 0.3WMin Nonfrozen control Culture durationa Mean basal release rate PU insulin/islet/min Mean stimulated release rate FU insulinlisletimin A stimulated release rate PU insulin/islet/min

20-24b

Frozen-thawed

2-4 hr* (n = 7)

(n = 3)

2-4< (n = 6)

20-24 hrr (n = 5)

0.06 + 0.02”

0.07 -t 0.03

0.16 k 0.04

0.12 2 0.03

1.31 5 0.09

0.92 t 0.17

0.38 A 0.07

0.49 k 0.05

1.24 r 0.08

0.86 5 0.15

0.23 2 0.05

0.42 4 0.05

n The mean basal release rate refers to the rate of insulin release 15 min prior to stimulation while the islets were perifused with 100 mg/dl glucose (nonstimulatory). The mean stimulated release rate refers to the rate of insulin released during the period of perifusion with 400 mg/dl glucose (stimulatory). A values that refer to the amount of insulin secreted in excess of nonstimulated (basal) release. b Postisolation culture duration. c Post-thaw culture duration. rl Results are expressed as means k SEM.

FACTORS IN THE ASSESSMENT OF CRYOPRESERVED ISLETS

minimum amount of inhibition of insulin release would be hypothermic exposure to the cryoprotectant using the stepwise addition and dilution protocol. When this was done with 1.4 M Me,SO the dynamic response pattern closely paralleled the Me,SO-free controls (Fig. 4).

Islet Function after Freezing Our next consideration was at what cooling rate should islets be frozen. In our system the cooling rate which resulted in optimum survival was 0.3”Umin. With the faster cooling rates several manifestations of freeze-induced damage became apparent (see Ref. (7)). The slow rate consistantly resulted in islets that were glucose responsive, i.e., under appropriate moment-tomoment control by glucose. The insulin secretory responses for these optimally cooled islets are found in Table 3 and Fig. 5. The biphasic nature of the dynamic response is intact, but slightly atypical; the second phase is depressed. Although increasing the duration of post-thaw culture results in a doubling of the net insulin output upon stimulation (Table 3), there is no improvement in the dynamics of the insulin response (Fig. 5). Two additional aspects of islet function were investigated simultaneously: The ability of optimally cooled islets to synthesize insulin and their ability to release insulin during a more prolonged stimulatory period. As Table 4 illustrates, both groups of frozen-thawed islets synthesize insulin in the same relative amounts as nonfrozen control groups. When we look at insulin release during the 180-min incubation (Fig. 6), those islets with a 2- to 4-hr post-thaw culture period release less insulin up to 120 min, but after 180 min no difference from the nonfrozen group can be demonstrated. For those islets with a 20- to 24-hr postthaw culture period, insulin release closely parallels that of the nonfrozen group.

I

10

I

I

1

,

20

30

40

50

,

60

I

I

I

70

60

90

I

loo

I

110

Time (min)

FIG. 5. The dynamic insulin response from untreated control islets (upper panel) (n = 4). The dynamic pattern of insulin release from frozen-thawed islets after either 2-4 hr (middle panel, n = 6) or 20-24 hr (lower panel, n = 5) of post-thaw culture. Quantitative values (means ? SEM) for insulin release rates are found in Table 3. (Modified from McKay and Karow, Ref. (7)).

DISCUSSION

Dimethyl sulfoxide has many potentially harmful physica and biochemical prop-

158

MC KAY AND KAROW TABLE 4 Protein Synthesis from Islets Cooled at a Rate of 0.3”CIMin Percentage of total cpm incorporated into islet protein

Nonfrozen Control 2- to 4-hr postisolation Frozen-thawed 2- to 4-hr post-thaw Frozen-thawed 20- to 24-hr post-thaw

Peak 1

Peak II

Peak III

70.0 -t- 12.9d

12.1 + 5.6

17.9 k 7.9

67.1 2 5.7

10.7 -t 5.0

21.4 ” 7.6

79.9 2 4.9

7.5 i 3.6

11.1 t_ 3.4

a Protein peak I eluted in the void volume and represents islet protein with a molecular weight >20,000. b Protein peak II was undetermined by our laboratory, but Schatz et al. (10) demonstrated this peak to be proinsulin. (’ Protein peak III represents the insulin containing fraction as determined by insulin radioimmunoassay and calibration with [iZ51]insulin. d The results are expressed as means t SEM.

erties. It also exerts a broad spectrum of Me&30 might produce on insulin release in pharmacological effects (see Ref. (11)). the absence of freezing. Then, by defining Several investigators (1, 2, 8, 9) are using the effects of the cryoprotectant on the Me&30 as a cryoprotectant to increase the nonfrozen tissue, one may now be able to post-thaw viability of frozen islets in con- distinguish damage due to the freezecentrations as high as 2 M, yet little is thaw process from that of merely treatment known of the effects of this agent on islet with the cryoprotectant alone. Addition function in the absence of freezing. The ini- and removal of the cryoprotectant does tial studies presented here were designed to result in inhibition of subsequent glucosedetermine the extent of adverse effects that induced insulin release. However, lowering

c ;wz IiiF u-c ZC i,’ 2 .’

1500,

1000.

-

nonfrozen

-

frozen-thawed,

control 2-4 hr culture

A--.A

frozen-thawed.

20-24

‘3 500-

hr culture

O0

60

120 TIME

180

(min)

FIG. 6. Insulin secretion during the 3 hr of incubation with [3H]leucine. Aliquots (20 ~1) were removed from the amino acid incubation medium of each group of SOislets at 60-min intervals and assayed for insulin content. Values were normalized by subtraction of the amount of insulin found at t = 0 from each of the time periods for each experiment. Results are expressed as means ? SEM (n = 4-7).

FACTORS

IN THE ASSESSMENT

the exposure temperature to Me&SO and stepwise treatment does minimize the amount of inhibition. Using this type of cryoprotectant treatment protocol, the cooling rate which results in optimum survival is 0.3”C/min. These islets synthesize insulin and release insulin biphasically upon glucose stimulation, although the second phase of insulin release appears depressed. Of importance, though, is that insulin release from these islets is under the moment-to-moment control by glucose. This was not always the case with islets frozen under suboptimal conditions. Insulin release from these islets either was delayed or failed to return to nonstimulatory values when glucose levels were reduced (7). These results also emphasize the importance of post-thaw culturing on islet function. Glucose-induced insulin release was increased twofold by increasing the duration of post-thaw culture. This is in contrast to our nonfrozen control group where insulin release was reduced by 30% over the same period of culture. For these studies we have used both the total quantity of insulin released upon stimulation, as well as the dynamic pattern of release, as indices of viability. These studies indicate the time interval during which glucose-induced insulin release is measured may also be important. Figure 6 indicates that with shorter stimulatory periods, larger discrepancies between nonfrozen and frozen-thawed groups exist. If short stimulatory periods are used to compare nonfrozen and cryopreserved islets, then viability would appear low. However, with longer stimulatory periods, less difference between groups is observed. Why this discrepancy during the early stages of insulin release exists probably relates to some form of freeze-induced damage. SUMMARY

With the development of techniques for the isolation and transplantation of pancre-

OF CRYOPRESERVED

ISLETS

159

atic islets of Langerhans, research has been directed toward low-temperature storage of islets as a means of preservation. For successful islet cryopreservation several factors must be considered. In these studies we have investigated the effects of the cryoprotectant dimethyl sulfoxide (Me,SO) on islet function in the absence of freezing. We have found that Me,SO pretreatment can inhibit subsequent glucose-induced insulin release, but this effect can be minimized by hypothermic exposure to the cryoprotectant using a stepwise addition and dilution protocol for treatment. By studying islet function after freezing and thawing, we have found also that a slow cooling rate (0.3”Cmin) results in optimal survival and that islet function can be significantly improved by increasing the duration of postthaw culture. The results of these studies address only a few of the many questions that need to be answered before clinical application of cryopreserved islet transplantation occurs. REFERENCES 1. Bank, H. L., Davis, R. E., and Emerson, D. Cryogenic preservation of isolated rat islets of Langerhans: Effect of cooling and warming rates. Diabetalogia 16, 195- 199, (1979). 2. Bank, H. L., and Reichard, L. Cryogenic preservation of isolated islets of Langerhans: Twostep cooling. Cryobiology 18, 489-496, (1981). 3. Fahy, G. M. Analysis of “solution effects” injury: Cooling rate dependence of the functional and morphological and sequellae of freezing in rabbit renal cortex protected with dimethyl sulfoxide. Crybiology 18, 550-570, (1981). 4. Ferguson, S., Allsopp. R. N., Taylor, R. M. R., and Johnston, I. D. A. Isolation and long term preservation of pancreatic islets from mouse, rat and guinea pig. Diabetologia 12, 1 IS- 121, (1976). 5. Lacy, P. E., and Kostianovsky, M. Method for the isolation of intact islets of Langerhans from the rat pancreas. Diabetes 16, 35-39, (1967). 6. McKay, D. B., and Karow, A. M. Glucose induced insulin release from DMSO-treated rat islets of Langerhans. Res. Commun. Chem. Pathol. Pharmacol. 30, 15-27, (1980). 7. McKay, D. B., and Karow, A. M. A functional analysis of isolated rat islets of Langerhans:

160

MC KAY AND KAROW

Effects of dimethyl sulfoxide and lowtemperature preservation. Cryobiology, 20, 41-50, (1983). 8. Rajotte, R. V., Scharp, D. W., Downing, R., Preston, R., Molnar, G. D., Ballinger, W. F., and Greider, M. H. Pancreatic islet banking: The transplantation of frozen-thawed rat islets transported between centers. Crybiology 18, 357-369, (1981). 9. Rajotte, R. V., Stewart, H. L., Voss, W. A. G., Shnitka, T. K., and Dossetor, J. B. Viability studies of frozen-thawed rat islets of Langerhans. Cryobiology 14, 116- 120, (1977). 10. Schatz, H., Maier, V., Hinz, M., Nierle, C., and

Pfeiffer, E. F. Stimulation of H-3-leucine incorporation into the proinsulin and insulin fraction of isolated pancreatic mouse islets in the presence of glucagon, theophylline and cyclic AMP. Diabetes. 22, 433-441, (1973). 11. Shlafer, M. Pharmacological consideration of cryopreservation. In “Organ Preservation for Transplantation” (A. M. Karow, Jr., and D. E. Pegg, Eds.), 2nd ed. Dekker, New York, 1981. 12. Wright, P. H., Makalu, D. K., Vichick, D., and Sussman, K. E. Insulin immunoassay by back titration; some characteristics of the technique and the insulin precipitant action of alcohol. Dinbetes 20, 30-45, (1971).