Inhibition of nonlysosomal calcium-dependent proteolysis by glycine during anoxic injury of rat hepatocytes

GASTROENTEROLOGY1994;106:168-176

Inhibition of Nonlysosomal Calcium-Dependent Proteolysis by Glycine During Anoxic Injury of Rat Hepatocytes J. CHRISTOPHER NICHOLS,* STEVEN F. BRONK,* RONALD L. MELLGREN, t and GREGORY J. GORES* *Center for Basic Research in Digestive Diseases, Departmentof Internal Medicine, Mayo Medical School, Clinic, and Foundation, Rochester, Minnesota; and *Department of Pharmacology, Medical College of Ohio, Toledo, Ohio

Background~Aims: The mechanism by which glycine protects against hepatocyte death during anoxia remains unclear. Nonlysosomal proteolysis, including calpain proteolys|s, has been implicated as a mechanism of lethal cell injury. However, the effect of glycine on nonlysosomal proteolysis is unknown. The aim of this study was to ascertain if glycine cytoprotection is associated with inhibition of nonlysosomal proteolysis. Methods: Rat hepatocyte suspensions were rendered anoxic using an anaerobic chamber. Cell viability was measured by propidium iodide fluorometry. Nonlysosomal protease activity was quantitated by the release of trichloroacetic acid-soluble free amines or tyrosine. Calpain protease activity was measured using a fluorogenic substrate. Results: Glycine and alanine (but not valine) markedly improved cell viability during anoxia in a concentration-dependent manner. During anoxia, the majority of nonlysosomal proteolysis (60%) was dependent on extracellular Ca 2+. Glycine only inhibited that portion of nonlysosomal proteolysis that was dependent on extracellular Ca 2+. Amino acids inhibited the anoxia-stimulated increase in calpain protease activity with the same specificity and concentration-dependence observed for cytoprotection. Glycine was more potent in directly inhibiting purified m-calpain as compared with li-calpain protease activity. Conclusions: Glycine may exert its cytoprotective activity during lethal anoxic hepatocyte injury, in part by inhibiting Ca2+-dependent degradative, nonlysosomal proteases, including calpains.

noxic hepatocyte injury occurs during liver transplantation, liver surgery, shock syndromes, and vascular occlusions. Information regarding the mechanisms of lethal hepatocyte injury during anoxia may lead to improved therapeutic strategies. Although adenosine triphosphate (ATP) depletion is a salient feature of anoxic cell injury, the cellular mechanisms leading to hepatocyte death following ATP depletion remain controversial. The use of cytoprotective agents provides a unique opportunity to explore the mechanisms culminating in lethal anoxic hepatocyte injury. Recently, glycine has been shown to be cytoprotective during ATP depletion in

rat hepatocytes, rabbit renal proximal tubular cells, and human endothelial cells. *-3 Thus, glycine appears to be a useful compound to help identify the fundamental mechanism(s) causing hepatocyte death during anoxia. In hepatocytes, cytoprotection during ATP depletion with metabolic inhibition by potassium cyanide is also provided by L-alanine (but not by D-alanine) or other small neutral amino acids, suggesting that a highly specific, limited amino acid configuration is responsible for this cytoprotection. * Glycine cytoprotection during ATP depletion is not related to preservation of cellular ATP or glutathione and does not require protein synthesis or mitochondrial metabolism of glycine. *'2"4'5 We recently reported that glycine cytoprotection is associated with a decrease in total cellular proteolysis, suggesting that inhibition of degradative proteolysis may be the mechanism responsible for glycine cytoprotection.~ The concentration-response relationship for glycine cytoprotection and inhibition of total cellular proteolysis were similar, further supporting this hypothesis. However, the cellular compartmentation and class of proteases inhibited by glycine were not characterized in our previous study. Cellular proteolysis can be classified as either lysosomal or nonlysosomal. Because lysosomal proteolysis is suppressed by ATP depletion,6 cellular proteolysis during ATP depletion would be expected to be almost exclusively caused by nonlysosomal proteolysis. Indeed, we recently reported virtually complete inhibition of lysosomal proteolysis, but stimulation of nonlysosomal proteolysis during anoxia in rat hepatocytes.7 In chick skeletal muscle, nonlysosomal proteolysis is also enhanced by ATP depletion; the increase in nonlysosomal proteolysis is caused by an activation of Ca2+-dependent proteases, s An increase in Ca2+-dependent, nonlysosomal proteolysis could lead to degradative proteolysis of key cellular proteins, thereby contributing to lethal cell injury. InhibiAbbreviations used in this paper: AMC, 7-amlno-4-methylcoumarin; EGTA,ethylene glycol tetraacetic acid; LDH, lactate dehydrogenase; KRH, Krebs'-Ringer's-HEPES(buffer). © 1994 by the American GastroenterologicalAssociation 0016-5085/94/$3.00

January1994

GLYCINE CYTOPROTECTION 169

tion ofCa2+-dependent, nonlysosomal proteolysis by protease inhibitors protects hippocampal neurons from ischemia, endothelial cells from oxidative injury, and hepatocytes from injury by cystamine. 9-11 More specifically, calpains (which are well characterized Ca2+-dependent, nonlysosomal, cysteine proteases) are thought to contribute to cell injury during anoxia. 9'~2a3 W e recently reported a stimulation of calpainlike activity in rat hepatocytes during anoxia and a delay in the onset of cell death by a highly specific calpain inhibitor. 7 These multiple independent observations suggest that activation of Ca2+-dependent, nonlysosomal proteases (including calpains) may be an important mechanism contributing to lethal anoxic cell injury. Therefore, the overall objective of our study was to determine ifglycine inhibits nonlysosomal, Ca2+-dependent proteolysis during anoxic hepatocellular injury. The specific aims of our study were to answer the following questions. Is glycine as cytoprotective in anoxia as it is during ATP depletion by potassium cyanide? Does glycine inhibit calcium-dependent, nonlysosomal proteolysis and calpainlike protease activity during anoxic injury of rat hepatocytes? Are the concentration-dependence curves for glycine cytoprotection and inhibition of calpainlike protease activity similar? Is glycine inhibition of calpain activity direct or indirect?

Materials

and Methods

Hepatocyte Isolation The use and care of the animals for these studies was reviewed.and approved by the Institutional Animal Care and Use Committee of the Mayo Clinic. Isolated rat hepatocyte suspensions were obtained by collagenase perfusion of livers from fed, male Sprague-Dawley rats (250-300 g) as we have previously described in detail) a

Solutions The basic incubation medium was Krebs'-Ringer'sHEPES (KRH) buffer containing (in mmol/L):NaCl, 115; KH:PO4, 1; CaCl2, 2; KCI, 5; MgSO4, 1.2; and Na-HEPES buffer, 25 (pH 7.4). The incubation medium for the permeabilization of hepatocytes with staphylococcal s-toxin was a sucrose-potassium-based buffer containing (in mmol/L):sucrose, 125; KCI, 75; KH2PO4, 2; MgSO4, 1; and Na-HEPES buffer, 5 (pH 7.5). Measurements of purified calpain protease activity were performed in a buffer containing (in mmol/L):Tris, 50; dithiothreitol, 1; and CaCl2, 2 (pH 7.4).

Anoxia Anoxia (02 < 1.5 torr) was established and maintained within a large anaerobic chamber (model 1025; Forma Scientific Inc., Marietta, OH). v Methylene blue [3,7-bis(dimethylamino)phenothiazine-5-ium chloride] was used in the chamber as a chemical 02 indicator) 5 The chamber atmosphere was

bubbled continuously through a methylene blue solution in an indicator flask to monitor for any significant 02 leak. Methylene blue changes from colorless to an intense blue color at a redox potential of - 2 3 0 mV, which corresponds to an 02 concentration of 1.5 torr. Introduction of O2 into the chamber either by a leak or an error in technique turned the indicator blue within a few seconds. If the indicator solution turned blue, the experiments were stopped and the results discarded. Cell suspensions in sealed tubes were placed in the chamber via an air lock. Buffers were rendered anoxic by constant stirring for 12 hours in the chamber before the experiments. Hepatocytes were maintained at 37°C using a temperaturecontrolled aluminum heater block (Fisher catalog no. 11-718, Pittsburg, PA). The filter fluorometer and all other equipment required for these experiments were housed in the anaerobic chamber.

Determination of Cell Viability Cell viability was determined from the total fluorescence of propidium iodide as previously described. ~4 Fluorescence was monitored using a Sequoia-Turner model 450 filter fluorometer (Sequoia-Turner, Mountain View, CA) with 520 nm (8-nm band pass) excitation and 605-nm (long pass) emission filters.

Determination of Trichloroacetic AcidSoluble Free Amines With Fluorescamine as a Measurement of Proteolysis Hydrolysis of proteins results in the exposure of a primary amino group for every peptide bond broken (except bonds to proline). Measurement of trichloroacetic acid-soluble free primary amines (free amino acids and small peptides) provides a sensitive method to measure proteolysis. We used fluorescamine, a reagent that reacts with primary amines to form a fluorescent product, to quantitate proteolysis in cell suspensions (5 × 105 cells/mL). 7'16 Experiments were performed in the presence of 10 gtmol/L monensin to inhibit lysosomal proteolysis. 7 Protein synthesis was inhibited with 500 ~tmol/L cycloheximide to avoid reincorporation of the amino acids liberated by proteolysis into proteins, s Standard curves were generated with glycine (0-50 ~tmol/L).

Determination of Tyrosine as a Measurement of Proteolysis Hydrolysis of proteins can also be measured by quantirating the release of individual amino acids from proteins. Tyrosine is frequently used as an index amino acid for the measurement of proteolysis because the extracellular and intracellular pools equilibrate rapidly, and tyrosine reacts with 1nitroso-2-napthol to yield a product that can be measured fluorometrically. .7 Therefore, we quantitated proteolysis in hepatocyte suspensions in the presence and absence of extracellular Ca2+, the nonfluorescent calcium ionophore Br-A23187 (10 gmol/L), and glycine by the release of trichloroacetic acidsoluble tyrosine. Hepatocyte suspensions (106 cells/mL) were incubated in 2 mL of buffer at 37°C in the presence of 10 gmol/L monensin and'500 ~tmol/L cycloheximide. At desired

170 NICHOLSET AL.

intervals, the assay was quenched by the addition of 200 gL of 100% trichloroacetic acid. The protein precipitate was removed by centrifugation (2000g for 10 minutes). A 400 ILL aliquot of the supernatant was added to 1800 HI. of 1-nitroso2-napthol reagent (2 vol 0.1% 1-nitroso-2-napthol, 3 vol 3N nitric acid, and 3 vol 0.1N sodium nitrite) and incubated for 30 minutes at 37°C. The supernatant was extracted twice with 2 mL of ethylene chloride to remove the unreacted 1-nitroso2-napthol reagent. Fluorescence was measured in the aqueous phase using a Sequoia-Turner model 450 filter fluorometer with 460-nm (8-nm band pass) excitation and 570-nm (long pass) emission filters. Standard curves were generated with tyrosine (0-20 gM).

Measurement of Calpain Protease Activity We used Suc-Leu-Leu-Val-Tyr-7-amino-4-methylcoumarin (Suc-Led-Leu-Val-Tyr-AMC), a membrane-permeant, calpain substrate, to measure calpain protease activity in assays using hepatocyte suspensions, permeabilized cells, and purified calpains, v Fluorescence was monitored using a Sequoia-Turner model 450 filter fluorometer with 360-nm (8-nm band pass) excitation and 430-nm (long pass) emission filters. Standard curves were generated with AMC (0-300 nmol/L).

Permeabilization of Hepatocytes With ~Toxin The effect of glycine on heparocyte calpainlike activity was also assessed in permeabilized cells in which extracellular and intracellular concentrations of glycine are equal. StaphyloCoccal 0t-toxin has been used to permeabilize many cell types, inclhding rat hepatocytes) s or-toxin interacts with the plasma membrane to form stable transmembrane pores 2.5 nm in diameter. The selective permeabilization of the plasma membrane permits the passage of small molecules --<1000 mol wt (e.g., small fluorogenic substrates and amino acids); however, large cytosolic proteins (e.g., calpains mol wt of 110 kilodaltons) are retained within the cell) 9 Hepatocytes (106 cells/mL) in a sucrose-potassium buffer were incubated for 30 minutes with 200 hemolytic units/mL CZ-toxin.The efficiency and selectivity of or-toxin permeabilization at 30 minutes was assessed by counting the percentage of cells stained by ttypan blue and by measuring lactate dehydrogenase (LDH) release into the supernatant. Aliquots of cell-free buffer for the LDH assay were obtained by separating the cells from the buffer by passing the suspension gently through a 0.2 gtm syringe filter (Nalgene Co., Rochester, NY). LDH activity was assayed by monitoring nicotinamide adenine dinucleotide consumption during the conversion of pyruvate to lactate, a° Nicotinamide adenine dinucleotide absorbance was monitored at 340 nm and 23°C using a spectrophotometer (Beckman DU 7400; Beckman Instruments, Palo Alto, CA).

Isolation of m-Calpain.and g-Calpain m-Calpain was purified from bovine heart tissue and g-calpain from human erythrocytes as previously described in detail. 2~'22 The isolated enzyme was stored in 50 mmol/L morpholinepropane sulfonic acid, 0.2 mmol/L ethylenedi-

GASTROENTEROLOGYVol. 106, No. 1

aminetetraacetic acid, 1 mmol/L dithiothreitol, and 50% glycerol (pH 7.0) at -20°C.

Statistical Analysis All data represent a minimum of three separate experiments (each performed in triplicate) and are expressed as means _ SE unless otherwise indicated. Differences between groups were analyzed using analysis of variance for repeated measures. A post hoc analysis using the Bonferroni test was used to account for multiple comparisons. Half-maximal concentrations providing cytoprotection were determined by plotting the data on semilog paper. All statistical analysis was performed with the statistical software package InStat from GraphPAD (San Diego, CA). Enzyme kinetic data were calculated using the software package Enzfitter from Elsevier Biosoft (Cambridge, England).

Materials Fluorescamine, Br-A23187, propidium iodide, trichloroacetic acid, monensin, glycine, valine, alanine, staphylococcal 0~-toxin, glycerol, dithiothreitol, Tris buffer, and fraction V of bovine serum albumin were obtained from Sigma Chemical Co. (St. Louis, MO); HEPES buffer was obtained from United States Biochemical Corporation (Cleveland, OH); collagenase type D was obtained from Boehringer Mannheim Biochemicals (Indianapolis, IN); Suc-Leu-Leu-Val-Tyr-AMC and Suc-AlaAIa-Phe-AMC were obtained from Bachem Bioscience, Inc. (Philadelphia, PA); AMC was obtained from Enzyme Systems Products (Livermore, CA); and digitonin was obtained from Calbiochem (La Jolla, CA). All other reagents were of analytical grade from the usual commercial sources.

Results Cytoprotection by Glycine During Anoxic Hepatocellular Injury In our previous study, glycine cytoprotection was observed in hepatocytes depleted of ATP by metabolic inhibition with potassium cyanide. * Because ATP depletion by potassium cyanide may not be a valid model of anoxia, 23 we first confirmed that glycine was also cytoprotective during anoxia. Cell viability during anoxia was significantly improved in the presence of 2 mmol/L glycine (Figure 1A). After 4 hours of anoxia, cell viability was only. 13.2% + 2% in the absence ofglycine, whereas cell viability in the presence of 2 mmol/L glycine was 73.8% + 0.7% (P --< 0.01). Indeed, after 4 hours, cell viability of anoxic hepatocytes incubated in the presence of 2 mmol/L glycine was virtually identical to aerobic controls (82.6% + 1.0% vs. 73.8% -+ 0.7%). To determine the specificity of amino acid cytoprotection, cell viability was measured in the presence of other amino acids. L-Alanine was also cytoprotective; the addition of 5 mmol/L I.-alanine improved cell viability to 62.1% _-Z4.5% compared with 15.5% --+ 5.4% in the absence of L-alanine (Figure 1B). In contrast, L-valine (in concentra-

January 1994

GLYCINE CYTOPROTECTION 171

100

viously reported in hepatocytes during ATP depletion by potassium cyanide. ~ Because maximal cytoprotection was observed with 2 mmol/L glycine and 5 mmol/L alanine, we used these concentrations ofglycine and alanine, respectively, for the remainder of our studies. Glycine Inhibits C a 2 + - D e p e n d e n t N o n l y s o s o m a l Proteolysis in Hepatocyte Suspensions

° m

J~ ¢o

50

--

o m

¢D 0

I

I

I

I

I

0

60

120

180

240

100

0 Minutes Rgure 1. Glycine and L-alanine protect against cell killing during anoxia. Hepatocyte suspensions (10 s cells/mL) were incubated in KRH buffer containing 0.2% bovine serum albumin and 1.0 pmol/L propidium iodide at 37°C. (A) Cells were incubated under aerobic or anaerobic conditions, in the absence of amino acids, or in the presence of 2.0 mmol/L glycine or 2 mmol/L L-valine. (/3) Cells were incubated under aerobic or anaerobic conditions, in the absence of amino acids, or the presence of 5 mmol/L L-alanine. O, aerobic + glycine; A, aerobic, absence of amino acids; Q, anoxic + glycine; & , anoxic, absence of amino acids; 0 , anoxic + valine; I , anoxic + alanine; I-I, aerobic + alanine.

Initially, we measured nonlysosomal proteolysis in hepatocyte suspensions during anoxia in the presence (Ca2+-dependenr, nonlysosomal proteolysis) and absence of extracellular Ca 2+ (Ca2+-independent, • nonlysosomal proteolysis) using the fluorescamine-based assay (Figure 3). After 2 hours ofanoxia, nonlysosomal proteolysis was twofold greater in the presence of extracellular Ca 2+ (219.9 -+ 4.3 vs. 90.7 + 31.9 nmol free amines" 106 cells -t. 120 min-~; P -----0.01). Because glycine contains a free amine, and therefore cannot be used with the fluorescamine-based assay, nonlysosomal" proteolysis was also quantitated using the tyrosine-based assay (Table 1). Using this assay, nonlysosomal proteolysis was also twofold greater in the presence of extracellular Ca 2+ during anoxia (45.9 + 4.3 vs. 21.1 + 6.2 nmol tyrosine-106 cells -z- 120 min-l; P -- 0.01). Under aerobic conditions, nonlysosomal proteolysis was also significantly greater in the presence of extracellular Ca 2+. However, total nonlysosomal proteolysis was 32% greater during anoxia than under aerobic conditions. In the absence of extracellular Ca 2+, there was no significant difference in nonlysosomal proteolysis between aerobic and anoxic cells. Thus, the majority of nonlysosomal proteolysis during anoxic and aerobic conditions is Ca2+-dependent, and the Ca2+-de-

100 Glycins b

o ~

~• 50 "N tions of 0.5, 1, 2, 5, and 10 mmol/L) did not protect against loss of cell viability during anoxia. Glycine cytoprotection was concentration-dependent with half-maximal protection at 0.4 mmol/L glycine and maximal protection at 2 mmol/L glycine (Figure 2). L-Alanine cytoprotection was also concentration-dependent and equally efficacious, but less potent than glycine (Figure 2). Half-maximal protection for I.-alanine was at 1.3 mmol/L and maximal protection at 5 mmol/L L-alanine. Thus, amino acid cytoprotection during anoxia showed the same specificity and concentration-dependence pre-

0

•

I

I

0

I

I

I

2 3 4 Amino acid, mM

I

I

5

Rgure 2. Amino acid cytoprotection during anoxic hepatocellular injury is concentration dependent. Cells were incubated in an anaerobic chamber in the presence of various concentrations of glycine (0) or L-alanine ( l l ) for 4 hours. All other conditions are as described in Rgure 1. Cell viability was assessed by propidium iodide~fluommetry.

172

NICHOLS ETAL.

I

GASTROENTEROLOGY VoI. IO6, No. 1

' I

i

I

' J

_=8 0"--

150

,<

RE 100 .C ~

e-

m

~

,

0

0

" 30

I 60

i

I 90

i

I 120

Minutes Rgure 3, The increase in nonlysosomal proteolysis during anoxia is predominantly calcium dependent, Hepatocyte suspensions (5 x 10 s cells/mL) were incubated in an anaerobic chamber in 1.3 mL of calcium-free KRH buffer containing 0.5 mmol/L ethylene glycol tetraacetic acid (EGTA) (e) or KRH buffer containing 2 mmol/L Ca2+ (HI) at 37°C for 2 hours. Proteolysis was determined by measuring trichloroacetic acid-soluble free amines via a fluorescent assay using fluorescamine.

pendent portion of nonlysosomal proteolysis is stimulated in afioxia. To determine ifglycine inhibits Ca2+-dependent, nonlysosomal proteolysis in hepatocytes, we initially measured nonlysosomal proteolysis in cells exposed to 10 ~tmol/L Br-A23187, a nonfluorescent Ca2+-ionophore, in the presence and absence of extracellular Ca2+ (Table 1). The calcium ionophore increased proteolysis fourfold in Ca2+-containing buffer compared with Ca2+-free buffer. Glycine (2 mmol/L) inhibited nonlysosomal proteolysis by 43% in hepatocytes treated with Br-A23187 in the presence ofextracellular Ca2+ but had no effect on nonlysosomal proteolysis in cells treated with Br-A23187 in the absence of extracellular Ca-'+. Likewise, glycine also inhibited nonlysosomal proteolysis during both aerobic and anoxic conditions in the presence of extracellular Ca2+. Glycine (2 mmol/L) inhibited nonlysosomal proteolysis by 27% and 35% after 2 hours of incubation under aerobic and anoxic conditions in Ca2+-containing media, respectively. In contrast, glycine had no effect on nonlysosomal proteolysis in cell suspensions incubated in Ca 2+free media. Thus, glycine appears to inhibit only that component of nonlysosomal proteolysis that is dependent o n extracellular Ca 2+.

Inhibition of Calpainlike Protease Activity by Glycine During Anoxic Hepatocellular Injury The fluorogenic, membrane-permeant, calpain substrate Suc-Leu-Leu-Val-Tyr-AMC was used to moni-

tor calpainlike protease activity in cell suspensions. We have previously validated the use of this membrane-permeant substrate for measuring calpainlike protease activity in hepatocytes. 7 Hydrolysis of the calpain substrate was measured at 30 minutes of anoxia because at this time point, ATP depletion is > 9 0 % but cell lysis has not occurredT; thus, measurements of calpain activity represent a response to injury as opposed to a postnecrotic phenomenon. Calpainlike protease activity increased 6.7fold after 30 minutes of anoxia (0.0370 + 0.001 vs. 0.248 _ 0.035 nmol AMC-106 cells-'.min-~; P -0.01) (Figure 4). Glycine inhibited calpainlike protease activity in a concentration-dependent manner with halfmaximal inhibition at 0.2 mmol/I, glycine and maximal inhibition at 2 mmol/L glycine. Likewise, L-alanine also inhibited calpainlike protease activity in a concentrationdependent manner with half-maximal protection at 0.5 mmol/L L-alanine and maximal protection at 5 mmol/L L-alanine. In contrast, L-valine (in concentrations up to 5 mmol/L only) decreased calpainlike protease activity by 34% and did not return calpain protease activity to basal aerobic values. Thus, calpainlike protease activity during anoxia is inhibited by amino acids with the same specificity and similar concentration-dependence observed for cytoprotection. The specificity of glycine inhibition of hydrolysis of the fluorogenic calpain substrate Suc-Leu-Leu-Val-TyrAMC by cell suspensions was assessed using Suc-AlaAIa-Phe-AMC, commonly used as a metalloprotease substrate. "~4"25 Proteolytic hydrolysis of Suc-Ala-AlaPhe-AMC did not significantly increase in hepatocyte suspensions after 30 minutes of anoxia (0.20 _ 0.03 vs. 0.24 + 0.03 nmol AMC-106 cells-'.min-'; P = not significant). Furthermore, glycine had no effect on the rate of hydrolysis of Suc-Ala-Ala-Phe-AMC (Figure 5). The inhibition of protease activity by glycine during anoxia using a peptidyl-AMC substrate appears to be specific for calpain activity because inhibition by glycine was not observed using a peptidyl-AMC substrate with differing protease specificity.

Glycine Inhibits Calpain Protease Activity in Permeabilized Hepatocytes and Using Purified Calpains To eliminate transport variables, calpain protease activity was measured in (x-toxin-permeabilized hepatocyte suspensions. In all suspensions, the permeabilization of hepatocytes was > 9 5 % as assessed by trypan blue staining, and LDH leakage was < 20%, showing selective permeabilization of hepatocytes by or-toxin as previously reported./s Inhibition of calpainlike protease activity in permeabilized cells was concentration-dependent with half-maximal inhibition at 2.2 mmoL/L and maximal in-

January 1 9 9 4

GLYCINE CYTOPROTECTION

173

Table 1. Inhibition of Nonlysosomal Proteolysis by Glycine in the Presence of Extracellular Calcium

Nonlysosomal proteolysis (nmo/ tyrosine 1 0 s cells/time) a Treatment Anoxic +Ca 2. - C a 2+ Difference b (%)

P

No glycine

Glycine (2 mmol/L)

Difference ~ (%)

P

45.9 __+4.3 21.1 _+ 6.2 - 54 "=0.05

29.6 _+ 6.2 21.1 -~ 7.3

-35 --

<0.05 NS

34.8 _+ 1.1 18.0 _+ 1.5 -48 <0.01

25.5 + 2.3 17.5 _+ 1.3 -44 <0.01

-27 --

<0.05 NS

21.0 __ 2.1 5.3 _ 2.9

11.9 ~ 1.5 8.9 ~ 1.5

-43 --

<0.05 NS

NS

Aerobic

+Ca 2+ - C a 2+ Difference b (%)

P Br-A23187 +Ca 2+ - C a 2+ Difference b (%)

P

- 75

<:0.01

NS

NOTE. Nonlysosomal proteolysis was .measured by quantitating the release Of trichloroacetic acid-soluble tyrosine in hepatocyte suspensions as described in Materials and Methods. Cells were incubated in KRH buffer containing 0.5 mmol/L EGTA and no added Ca 2. (-Ca 2+) or KRH buffer containing 2 mp~ol/L Ca 2+ (+Ca2+). Br-A23187 was used at a concentration of 10 llmol/L. Data represent the mean _+ SE for three separate experiments (each performed in triplicate) and were tested by unpaired Student's t test. aTime for measurement of proteolysis was 30 minutes with Br-A23187 and 120 minutes in the anoxic and aerobic experiments. bPercent differences are only shown for significant values.

hibition at 10 mmol/L. The Michaelis constant for the hydrolysis of Suc-Leu-Leu-Val-Tyr-AMC was significantly increased in permeabilized cells by 2 mmol/L glycine (14.5 + 0.81 vs. 96.0 - 16.4 ~mol/L; P -- 0.01), and the Vm~ Was decreased (1.05 + 0.01 vs. 0.16 _+ 0.01 nmol AMC' 106 cells -* "min-*; P -- 0.01). As assessed by Lineweaver-Burk analysis, these results indicate thfit glycine inhibited calpain protease activity in a direct, uncompetitive manner consistent with an allosteric mechanism of inhibition. Calpains exist in two isozymic forms, m-calpain and ~-calpain, which require millimolar and micromolar Ca2+ for in vitro activity, respectively. 26 To ascertain if glycine directly inhibits m-calpain and g-calpain protease activity, we measured the in vitro activity of highly purified calpains in the presence and absence of glycine. Glycine directly inhibited m-calpain protease activity with half-maximal inhibition at 3.5 mmol/L and > 9 0 % inhibition at 10 mmol/L (Figure 6). In contrast, glycine only inhibited g-calpain protease activity by 10% at 3.5 mmol/L and 75% at 10 mmol/L. Thus, glycine more potently inhibits m-calpain than ~-calp~lin protease activity.

Discussion Activation of Ca2+-dependent, nonlysosomal proteases has been proposed as a mechanism of lethal cell injury. 1°-13 The importance of Ca2+-dependent, nonlysosomal proteolysis as a mechanism of cell necrosis is suggested by the observations that specific protease inhibi-

tors protect various cell types from lethal injury. 9-'1 The degradarive effects of nonlysosomal proteolysis would be further accentuated during anoxic cell injury because protein synthesis is markedly suppressed, preventing replacement of degraded proteins. "~7 Although the role of Ca 2+ in mediating lethal hepatocyte injury remains controversial, recent data have shown an increase of cytosolic-free calcium (Cai2+) during anoxic injury of rat hepatocytes. 28 Removal of extracellular Ca 2+ prevents the increase of cytosolic-free calcium and delays cell death. 28 In the current study, Ca2+-dependent, nonlysosomal proteolysis was suppressed by glycine under anoxic conditions and in the presence of a calcium ionophore. Inhibition of Ca2+-dependent, nonlysosomal proteolysis is consistent with previous observations showing that glycine protects against loss of cell viability despite sustained increases of cytosolic-free calcium to micromolar concentrations. 29 Our data indicate that glycine cytoprotection against anoxic hepatocellular injury is associated with a direct inhibition ofCa"+-dependent, nonlysosomal proteases, including calpains. Thus, degradative, nonlysosomal proteolysis appears to be an important mechanism contributing to lethal injury of hepatocytes durin~ anoxia. Calpain protease activity was measured to further characterize and identify the Ca2+-dependent, nonlysosomal proteases inhibited by glycine during anoxia. The rationale for studying the effect ofglycine on calpain proteolysis was based on several published observations. First, calpains are Ca2+-dependent, nonlvsosomal, cvsteine pro-

174

NICHOLS ET AL.

GASTROENTEROLOGYVol. 106, No. 1

0.3

'!


o.1

0 0.0

1

2

3

4

5

A m i n o acid, m M Rgure 4. Glycine and L-alanine inhibit calpainlike protease activity during anoxia. Hepatocyte suspensions (106 cells/mL) in KRH buffer containing 0.2% bovine serum albumin were incubated in an anaerobic cham~)er at 370C. Cells were incubated in the presence of various concentrations of glycine (0), L-alanine (11), or L-valine ( • ). After 30 minutes of anoxic incubation, Suc-Leu-Leu-VaI-Tyr-AMC (20 pmol/L) was added to the cell suspensions, and the initial rate of peptidylAMC hydrolysis was quantitated for a period of 10 minutes by measuring the generation of free AMC fluorometrically. The dotted line represents basal activity of calpainlike protease activity under aerobic conditions in the absence of amino acids.

teases. 26 Second, activation ofcalpain proteolysis has been associated with anoxic myocardial, neuronal, and hepatocellulhr injury.9'12a3 Finally, calpain protease inhibitors retard bleb formation and delay the onset of cell injury) ° We observed that the amino acid specificity and concentration-dependence of amino acid cytoprotection during anoxia was closely associated with inhibition of calpainlike.protease activity. These data suggest glycine is cytoprotective, in part, by inhibiting calpain protease activity. The fluorometric assay in cell suspensions does not distinguish between the two different isozymes of calpains, m-calpain and gt-calpain. Therefore, we used purified m-calpain and li-calpain to determine which calpain isozyme was inhibited by glycine. Although glycine directly inhibited both isoenzymes, glycine more potently inhibits m-calpain than g-calpain protease activity. In hepatocytes, the activity of m-calpain is fivefold greater than the activity of l~-calpain, explaining the stronger correlation between the concentration-dependence for glycine inhibition of total hepatocellular calpain activity and isolated m-calpain. 31 The fivefold difference in calpain inhibition by glycine between cell suspensions and the purified protease could be caused by greater intracellular than extracellular concentrations of glycine. In fact, under physiological conditions, intraceUular concentrations of glycine are 4 - 6 fold greater than extracellular concentrations in renal

proximal tubular cells. 32 To eliminate potential concentration gradients of glycine across the plasma membrane, we measured calpain protease activity in (1-toxin permeabilized hepatocytes. In permeabilized cells, the concentration of glycine required for maximal inhibition of calpain protease was identical to that required with the purified protease. These results suggest that intracellular concentrations of glycine are greater than extracellular concentrations during anoxic injury of rat hepatocytes. Glycine transport into the cells by Na+-coupled transport processes (i.e., amino acid transport systems Gly, ACS, and Ala) would be expected to decrease as Na + gradients collapse because of the ATP depletion of anoxia) 3 However, glycine is also transported by several Na+-independent processes (i.e., amino acid transport systems ACS and Leu)) 3 These Na+-independent transporters may be active during anoxia, leading to greater intracellular than extracellular concentrations of glycine. The discrepancy between the concentrations ofglycine required to inhibit calpain protease activity in intact cells compared with the purified enzymes may also be caused by mechanisms leading to indirect inhibition of calpains by glycine within the cell. For example, calpains contain a glycinerich domain that is important in calpain binding to acidic intracellular phospholipids) <35 The interaction of calpains with acidic phospholipids results in activation of the calpains, even at physiological calcium concentrations) 5 Perhaps high concentrations of glycine also bind to acidic phospholipids, thereby preventing binding and activation of calpains.

~p..~uc-Ala-Ala-Phe-AM

100

%

E E

° ~

E o

#

s0

_

o

aI-Tyr-AMC

o

>

o

I

I

0

1

~ 2

!

I

±

3

4

5

Glycine, mM Rgure 6. Glycine inhibition of calpainlike protease activity is specific. Hepatocyte suspensions (10 s cells/mL) in KRH buffer containing 0.2% bovine serum albumin and various concentrations of glycine were incubated in an anaerobic chamber at 37°C. After 30 minutes of anoxic incubation, the calpain substrate SucoLeu-Leu-VaI-Tyr-AMC (0) or the metalloprotease substrate Suc-Ata-Ala-Phe-AMC (I-1) was added at a concentration of 20 p.mol/L. The initial rate of peptidylAMC hydrolysis was quantitated for a period of 10 minutes by measuring the generation of free AMC fiuorometrically.

January 1994

GLYCINE cY'rOPROTECTION 175

2.0

~';

1.5

<

,.o

r-

0.5

0.0 0

5

10

Glycine, mM Rgure 6. Glycine directly inhibits m-calpain protease activity. Two micrograms of purified m-calpain (0) or ll-calpain ( I ) was incubated in 1.5 mL of Tris-dithiothreitol buffer containing 2 mmol/L Ca2÷ and various concentrations of glycine at 37°C. Calpain protease activity was assessed by the rate of hydrolysis of the fluorogenic calpain substrate Suc-Leu-Leu-VaI-Tyr-AMC (40 Hmol/L). The initial rate of peptidyl-AMC hydrolysis was quantitated for a period of 10 minutes by measuring the generation of free AMC fluorometrically.

In a previous study, we found that a highly selective, membrane-permeant calpain inhibitor only partially protected against loss of cell viability during anoxia, despite complete inhibition of calpain protease activity/In the current study, cell survival was improved by glycine to a viability virtually identical to aerobic controls. The greater magnitude of protection afforded by glycine as comparecl with a selective calpain inhibitor may be explained by inhibition of nonlysosomal proteases in addition to calpains by glycine. Indeed, glycine inhibition of nonlysosomal proteolysis is twofold greater than that observed using a specific calpain inhibitor (35% vs. 16% inhibition, respectively). 7 Thus, glycine appears to not only inhibit calpains but also other Ca- -dependent, nonlysosomal proteases. Multiple Ca2+-dependent, nonlysosomal proteases may contribute to lethal cell injury during anoxia. For example, Ca--dependent, serinelike proteases have been recently described and found to contribute to pulmonary endothelial cell injury. 36'37 Glycine may also directly inhibit other calcium-dependent proteases by a direct allosteric effect and indirect mechanisms as discussed above for calpains. An alternative explanation for enhanced cytoprotection by glycine would include inhibition of other critical processes leading to cell death in addition to proteolysis. For example, glycine may inhibit the activity of other degradative hydrolases such as phospholipases or endonucleases, which may also contribute to irreversible cell injury, j" The results of this study may have implications for decreasing anoxic liver injury during organ preservation

for transplantation surgery. Addition of glycine or other inhibitors of nonlysosomal, Ca2+-dependent proteolysis to organ-storage solutions may improve organ viability following cold ischemic storage. For example, addition of glycine to Collin's solution results in functional improvement of dog kidneys after 48 hours of cold ischemia. ~9 Glycine in the rewarming media improves hepatocyte survival following 96 hours of cold ischemia in University of Wisconsin preservation solution. 4° The cytoprotective effect of glutathione in the University of Wisconsin solution may be related to its metabolic breakdown and release ofglycine. 2 Moreover, addition ofprotease inhibitors to the preservation solution enhances graft and animal survival in a rodent model of liver transplantation. 4. Further investigations, including studies using in vivo models, will ultimately be required to elucidate the physiological and therapeutic role ofglycine in anoxic and ischemic liver injury.

References 1. Dickson RC, Bronk SF, Gores GJ. Glycine cytoprotection during lethal hepatocellular injury from adenosine triphosphate depletion. Gastroenterology 1992;102:2098-2107. 2. Weinberg JM, Davis JA, Abarzua M, Rajan T. Cytoprotective effects of glycine and glutathione against hypoxic injury to renal tubules. J Clin Invest 1987;80:1446-1454. 3. Weinberg JM, Varani J, Johnson KJ, Roeser NF, Dame MK, Davis JA, Venkatchalam MA. Protection of human umbilical vein endothelial cells by glycine and structurally similar amino acids against calcium and hydrogen peroxide-induced lethal cell injury. Am J Pathol 1992;140:457-471. 4. Weinberg JM, Davis JA, Abar-zua M, Rajah T. Relationship between cell adenosine triphosphate and glutathione content and protection by glycine against hypoxic proximal tubule cell injury. J Lab Clin Meal 1989;113:612-622. 5. Weinberg JM, Venkatachalam MA, Garzo-Quintero R, Roeser NF, Davis JA. Structural requirements for protection by small amino acids against hypoxic injury in kidney proximal tubules. FASEB J 1990;4:3347-3354. 6. Plomp PJ, Gordon PB, Meijer AJ, Hoyvik H, Seglen PO. Energy dependence of different steps in the autophagic-lysosomal pathway. J Biol Chem 1989;264:6699-6704. 7. Bronk SF, Gores GJ. pH dependent, non-lysosomal proteolysis contributes to lethal anoxic injury of rat hepatocytes. Am J Physiol 1993; 264:G744-G751. 8. Fagan JM, Wajnberg EF, Culbert L, Waxman L. ATP depletion stimulates calcium-dependent protein breakdown in chick skeletal muscle. Am J Physiol 1992;262:E637-E643. 9. Lee KS, Frank S, Vanderklish P, Arai A, Lynch G. Inhibition of proteolysis protects hippocampal neurons from ischemia. Proc Natl Acad Sci USA 1991;88:7233-7237. 10. Geeraerts MD, Ronveaux-Dupal MF, Lemasters JJ, Herman B. Cytosolic free Ca2. and proteolysis in lethal oxidative injury in endothelial cells. Am J Physiol 1991;30:C889-C896. 11. Nicotera P, Hartzell P, Baldi C, Svensson S-A, Bellomo G, Orrenius S. Cystamine induces toxicity in hepatocytes through the elevation of cytosolic Ca2÷ and the stimulation of a nonlysosomal proteolytic system. J Biol Chem 1986;261:14628-14635. 12. Tolnadi S, Korecky B. Calcium-dependent proteolysis and its inhibition in the ischemic rat myocardium. Can J Cardiol 1986; 2:442-447. ~, 13. lizuka K, Kawaguchi H, Yasuda H. Calpain is activated during

176

NICHOLS ET AL.

hypoxic myocardial cell injury. Biochem Med Metab Biol 1991; 46:427-431. 14. Gores GJ, Nieminen A, Reishman KE, Dawson TL, Herman B, Lemasters JJ. Extracellular acidosis delays onset of cell death in ATP-depleted hepatocytes. Am J Physio11988; 255:C315-C322. 15. Murphy JG, Smith TW, Marsh JD. Mechanisms of reoxygenationinduced calcium overload in cultured chick embryo heart cells. Am J Physiol 1988;254:Hl133-H1141. 16. Groskreutz JL, Bronk SF, Gores GJ. Ruthenium red delays the onset of cell death during oxidative stress of rat hepatocytes. Gastroenterology 1992; 102:1030-1038. 17. Waalkes TP, Udenfriend S. A fluorometric method for the estimation of tyrosine in plasma and tissues. J Lab Clin Med 1957; 50:733-736. 18. McEwen BF, Arion WJ. Permeabilization of rat hepatocytes with staphylococcal aureus ~z-toxin. J Cell Biol 1985;100:19221929. 19. Ahnert-Hitger G, Mach W, Fo;auhr K.I, Gratzl M. Poration by atoxin and str~ptolysin O: an approach to analyze intracellular processes. Methods Cell Biol 1989;31:63-90. 20. Bergmeyer HU. Methods of enzymatic analysis. New York: Academic, 1984:574-578. 21. Mellgren RL, Mericle MT, Lane RD. Proteolysis of the calcium"dependent protease inhibitor by myocardial calcium-dependent protease. Arch Biochem Biophys 1986;246:233-239. 22. Mellgren RL. Interaction of human erythrocyte multicatalytic proteinase with polycations. Biochim Biophys Acta 1990; 1040:2834. 23. Aw TE, Jones DP. Cyanide toxicity in hepatocytes under aerobic and anaerobic conditions. Am J Physiol 1989; 257:C435-C441. 24. Mundy DI, Strittmatter WJ. Requirement for metalloprotease in exocyto~is: evidence in mast cells and enterochromaffin cells. Cell 1985;40:645-656. 25. Roe JL, Farech YA, Strittmatter WJ, Lennary WJ. Evidence for involvement of metalloproteases in a step in sea urchin gamete fusion. J Cell Biol 1988;107:539-544. 26. Mellgren RL. Calcium-dependent proteases: an enzyme system active at cellular membranes? FASEB J 1987; 1:100-115. 27. Buc-Calderon P, Lefebvre V, van Steenbrugge M. Inhibition of protein synthesis in isolated hepatocytes as an immediate response to oxygen limitation. In: Hochachka PW, ed. Surviving hypoxia, mechanisms of control and adaptation. Boca Raton, FL: CRC Press, 1993:271-280. 28. "Gasbarrini A, Bode AB, Farghali H, Bender C, Francavilla A, Van Thiel D. Effect of anoxia on intracellular ATP, Na=+, Ca,2+, M~ 2. , and cytotoxicity in rat hepatocytes. J Biol Chem 1992; 267:6654-6663. 29. Weinberg J, Davis JA, Roeser NF, Venkatachalam MA. Role of increased cytosolic free calcium in the pathogenesis of rabbit proximal tubule cell injury and protection by glycine or acidosis. J Clin Invest 1991;87:581-590.

GASTROENTEROLOGY Vol. 106, No. 1

30. Nicotera P, Hartzell P, Davis G, Orrenius S. The formation of plasma membrane blebs in hepatocytes exposed to agents that increase cytosolic Ca2+ is mediated by the activation of a nonlysosomal proteolytic system. FEBS Lett 1986;209:139-144. 31. Murachi T. Calpain and calpastatin. Trends Biochem Sci 1983; 8:167-169. 32. Weinberg JM, Nissim I, Roeser NF, Davis JA, Schultz S, Nissim I. Relationships between intracellular amino acid levels and protection against injury to isolated proximal tubules. Am J Physiol 1991; 260:F410-F419. 33. Kilberg M. Amino acid transport in isolated rat hepatocytes. J Membr Biol 1982;69:1-12. 34. Murachi T. Intracellular regulatory system involving calpain and calpastatin. Biochem Int 1989;18:263-294. 35. Suzuki K, Imajoh S, Emori Y, Kawasaki H, Minami Y, Ohno S. Regulation of activity of calcium activated neutral protease. Adv Enzyme Regul 1988;27:153-167. 36. Liao D, Gurtner GH. Calcium dependence of the serine proteases involved in oxidant activation of phospholipases A2 (abstr). FASEB J 1993;7:A346. 37. Steiner DF, Smeekens SP, Ohagi S, Chan SJ. The new enzymoiogy of precursor processing endoproteases. J Biol Chem 1992; 267:23435-23438. 38. Trump BF, Berezeky IK. The role of cytosolic Ca2÷ in cell injury, necrosis, and apoptosis. Curt Opin Cell Biol 1992;4:227-232. 39. Mangino MJ, Murphy MK, Grabau GG, Anderson CB. Protective effects of glycine during hypothermic renal ischemia-reperfusion injury. Am J Physiol 1991;261:F841-F848. 40. Marsh DC, Vreugdenhil PK, Mack VE, Belzer FO, Southard JH. Glycine protects hepatocytes from injury caused by anoxia, cold ischemia, and mitochondrial inhibitors, but not injury caused by calcium ionophores or oxidative stress. Hepatology 1993; 17:91-98. 41. Takei Y, Marzi I, Kauffman FC, Currin RT, Lemasters JJ, Thurman RG. Increase in survival time of liver transplants by protease inhibitors and a calcium channel blocker, nisoldipine. Transplantation 1990;50:14-20.

Received May 14, 1993. Accepted August 31, 1993. Address requests for reprints to: Gregory J. Gores, M.D., Center for Basic Research in Digestive Diseases, Mayo Medical School, Clinic, and Foundation, Rochester, Minnesota 55905. Supported by grants DK45331 and HL36573 from the National Institutes of Health and by the Mayo Foundation. Preliminary portions of this work were presented at the 95th meeting of the American Gastroenterological Association and published in abstract form (Gastroenterology 1993; 104:A963). The authors thank Sara Erickson for secretarial assistance and Drs. R. K. Pearson and William Karnes, Jr., for reviewing the manuscript and providing helpful and constructive comments.