Preconditioning-induced cytoprotection in hepatocytes requires Ca2+ ...

16 downloads 0 Views 450KB Size Report
In different tissues a brief period of ischemia increases the resistance to ... obvious that clinical management of transplantable organs would benefit from a better ...
Research Article

1065

Preconditioning-induced cytoprotection in hepatocytes requires Ca2+-dependent exocytosis of lysosomes Rita Carini1, Roberta Castino2, Maria Grazia De Cesaris1, Roberta Splendore1, Marina Démoz2, Emanuele Albano1 and Ciro Isidoro2,* Laboratories of Pathology1 and of Molecular Pathology2, Dipartimento di Scienze Mediche, Università del Piemonte Orientale ‘A. Avogadro’, via Solaroli 17, 28100 Novara, Italy *Author for correspondence (e-mail: [email protected])

Accepted 6 October 2003 Journal of Cell Science 117, 1065-1077 Published by The Company of Biologists 2004 doi:10.1242/jcs.00923

Summary A short period of hypoxia reduces the cytotoxicity produced by a subsequent prolonged hypoxia in isolated hepatocytes. This phenomenon, termed hypoxic preconditioning, is mediated by the activation of adenosine A2A-receptor and is associated with the attenuation of cellular acidosis and Na+ overload normally occurring during hypoxia. Bafilomycin, an inhibitor of the vacuolar H+/ATPase, reverts the latter effects and abrogates the preconditioning-induced cytoprotection. Here we provide evidence that the acquisition of preconditioning-induced cytoprotection requires the fusion with plasma membrane and exocytosis of endosomal-lysosomal organelles. Poisons of the vesicular traffic, such as wortmannin and 3methyladenine, which inhibit phosphatydilinositol 3kinase, or cytochalasin D, which disassembles the actin cytoskeleton, prevented lysosome exocytosis and also

abolished the preconditioning-associated protection from acidosis and necrosis provoked by hypoxia. Preconditioning was associated with the phosphatydilinositol 3-kinase-dependent increase of cytosolic [Ca2+]. Chelation of free cytosolic Ca2+ in preconditioned cells prevented lysosome exocytosis and the acquisition of cytoprotection. We conclude that lysosomeplasma membrane fusion is the mechanism through which hypoxic preconditioning allows hepatocytes to preserve the intracellular pH and survive hypoxic stress. This process is under the control of phosphatydilinositol 3-kinase and requires the integrity of the cytoskeleton and the rise of intracellular free calcium ions.

Introduction In different tissues a brief period of ischemia increases the resistance to necrosis or apoptosis produced by a subsequent ischemic insult (Murry et al., 1986; Peralta et al., 1997). This phenomenon known as ‘ischemic preconditioning’ has attracted the interest of scientists and clinicians because of the potential implications in organ transplantation. For instance, preconditioning has been shown to protect hepatocytes against injury from ischemia or reperfusion and to improve the success rate of liver transplants ‘taking root’ in rats (Yoshizumi et al., 1998; Yin et al., 1998; Yamagami et al., 1998). More recently, ischemic preconditioning has proved its clinical efficacy in patients undergoing liver resection (Clavien et al., 2000). It is obvious that clinical management of transplantable organs would benefit from a better knowledge of the chemical triggers, the signal pathways and the effector mechanisms responsible for the cytoprotective effect of ischemic preconditioning. Apparently both necrosis and apoptosis are prevented in preconditioned hepatocytes exposed to hypoxia. Studies performed on isolated and perfused liver have shown that resistance to hypoxia induced by preconditioning is associated with the release of nitric oxide and adenosine (Peralta et al., 1997; Peralta et al., 1999), downregulation of caspase 3 activity (Yadav et al., 1999) and decreased production of TNFα by

Kupffer cells (Peralta et al., 2001). We have shown that the preconditioning-induced cytoprotection can be reproduced in isolated rat hepatocytes by a short cycle of hypoxiareoxygenation or by direct stimulation of the adenosine A2Areceptor with the agonist CGS21680 (Carini et al., 2001a). The signaling pathway was shown to involve a trimeric G-inhibitor protein, the phospholipase C, PKC isoenzymes and the p38MAPK (Carini et al., 2000; Carini et al., 2001a). In preconditioned hepatocytes the hypoxia-induced acidosis and the consequent Na+ overload, critical alterations for the appearance of hypoxic damage (Carini et al., 1997a; Carini et al., 1999), are prevented (Carini et al., 2000; Carini et al., 1995). Both these effects do not occur in preconditioned hepatocytes treated with bafilomycin A, an inhibitor of the vacuolar H+/ATPase pump (Carini et al., 2001b). The latter observation indicates that the H+/ATPase pump, normally located on endosomal-lysosomal membrane, might be responsible for the attenuation of the hypoxic acidosis in preconditioned hepatocytes. Several reports by our and other laboratories show that some of the above mentioned signal transducers act in fact as regulators of the endocytic membrane traffic (Ogier-Denis et al., 1995; Chiarpotto et al., 1999; Baldassarre et al., 2000; Petiot et al., 2000), implying that movement of vacuolar acidic organelles could be linked to

Key words: Cell death, Cathepsin D, Ischemia, Exocytosis, Signal transduction

1066

Journal of Cell Science 117 (7)

some biochemical features of preconditioning. The concept that vesicle trafficking controls the cell surface expression of proteins has recently received confirmation in various cell models. Thus, in hepatocytes the cell surface exposition of death receptors (Fas and TNFR1) was shown to rely on endocytic vesicle traffic (Feng et al., 2000). Also, in lymphocytes the cell surface expression of the Fas ligand was shown to depend upon exocytic insertion of lysosomal-like cytotoxic granules (Bossi et al., 1999). Similarly, degranulation in activated neutrophils was shown to result in the insertion of the vacuolar H+/ATPase on the plasma membrane (Nanda et al., 1996). Based on these observations, we have hypothesized that endosome and lysosome translocation to the cell periphery and fusion with plasma membrane is the (principal) mechanism through which the cytoprotective effect of preconditioning is established. Indeed, a rapid movement of endosomallysosomal organelles would be compatible with the fact that preconditioning establishes within minutes (5 to 10 minutes of hypoxia is sufficient). Here we show that in preconditioned hepatocytes such acidic organelles move in fact from the perinuclear region toward the plasma membrane and fuse with it. This was demonstrated by showing the peripheral localization of cathepsin D-positive organelles, the surface exposition of Lamp-1 and the extracellular release of soluble lysosomal enzymes in hypoxic-preconditioned hepatocytes. The same effects were produced by stimulating the adenosine A2A-receptor with CGS21680, a condition that also confers cytoprotection. We also provide evidence that inhibition of lysosome exocytosis by disrupting the actin cytoskeleton or blocking the activity of (phosphoinositide 3-kinase) PI3K precludes the acquisition of preconditioning-induced cytoprotection from hypoxia as well as the associated attenuation of acidosis and of Na+ overload. Finally, we demonstrate that elevation of cytosolic free calcium ions levels in preconditioned hepatocytes is mandatory for both exocytosis of endosomal-lysosomal organelles and acquisition of cytoprotection. To our knowledge this is the first report showing the occurrence of PI3K-mediated and calciumdependent exocytosis of lysosomes upon stimulation of the adenosine A2A-receptor and the link between this cellular event and the protection from hypoxic cell death. Materials and Methods Materials Collagenase (Type I), N-(2-hydroxyethyl)-piperazine-N′-(2ethanesulfonic acid) (HEPES), phenylmethylsulphonyl fluoride, propidium iodide, wortmannin (WM), 3-methyladenine (3MA), cytochalasin D (Cyt D), CGS21680 were purchased from Sigma Chemical Co (St Louis, MO). EGTA-AM [ethylene glycol bis-(βaminoethyl ether) N,N,N′,N′-tetraacetic acid acetoxymethyl ester] was from Calbiochem (St Diego, CA). All the other chemicals were of analytical grade and were purchased from Merck (Darmstad, Germany) if not otherwise specified. Hepatocyte isolation, treatments and estimation of cell viability Rat hepatocytes were freshly isolated by collagenase liver perfusion of male Wistar rats (180-250 g) (Harlan Italy, S. Pietro al Natisone, Italy), as previously reported (Carini et al., 2000; Carini et al., 2001a). The use and the care of the animals were approved by the Italian Ministry of Health and by the University Commission for Animal Care following the criteria of the Italian National Research Council.

Hepatocytes were suspended at a final cell density of 106/ml in KrebsHenseleit-HEPES (KHH) buffer containing 118 mM NaCl, 4.7 mM KCl, 1.2 mM KH2PO4, 1.3 mM CaCl2, 25 mM NaHCO3– and 20 nM HEPES at pH 7.4. Hepatocytes were preconditioned either by exposure to CGS21680 or by a hypoxic-reoxygenation cycle as previously described (Carini et al., 2000; Carini et al., 2001a). The inhibitors WM, 3MA and CytD were added 15 minutes before preconditioning and were present throughout the following incubation. Hepatocytes were then incubated for 60 minutes at 37°C in sealed bottles under 95% O2/5% CO2 (control condition) or 95% N2/5% CO2 (hypoxia). Substances were used at the following concentrations: CGS21680, 1 µM; WM, 250 nM; 3MA, 10 mM; CytD, 20 µM. Hepatocyte viability was determined by standard LDH assay, the Trypan Blue exclusion test and by measuring the fluorescence of hepatocytes stained with propidium iodide according to the method of Gores et al. (Gores et al., 1989). For the latter, 106 hepatocytes were loaded with 10 µg/ml propidium iodide in 1 ml KHH buffer and the fluorescence was determined in a spectrofluorometer at 520 nm and 605 nm excitation and emission wavelengths, respectively. Parallel aliquots of hepatocytes were permeabilized with digitonin (375 µM) prior to loading with propidium iodide in order to obtain the maximal staining. Extent of cell death was deduced from the ratio of fluorescence intensity measured in non-permeabilized vs digitonin-permeabilized samples. At the beginning of the experiments hepatocyte viability ranged between 80-85%. Immunofluorescence At the end of the treatment hepatocytes were seeded on polylisinecoated glass coverslips, allowed to adhere for 5 minutes and then fixed in absolute methanol. This method allowed rapid cell attachment and proved valid for morphological studies since the integrity of subcellular structures in living cells was well maintained, despite cell polarity being lost. Endosomal-lysosomal organelles were traced by immunodetection of cathepsin D (CD), a soluble lysosomal enzyme, and of Lamp-1, a lysosomal membraneassociated glycoprotein. Cell morphology could be better appreciated by immunostaining of actin filaments. CD immunolocalization was performed by using a specific rabbit antiserum (Dragonetti et al., 2000), Lamp-1 and actin were revealed by using specific mouse monoclonal antibodies, respectively purchased from BD Transduction Laboratories (Lexington, KY) and Sigma. Specific secondary antibodies, either conjugated with Texas Red or FITC, were purchased from Sigma. As a negative control, cells were incubated with the secondary antibody alone or with pre-immune antiserum. The experiment was repeated three times and for each experimental condition three coverslips were prepared. At least four fields with about 10-20 cells per field have been analyzed in each coverslip with a laser confocal immunofluorescent microscope (Leica DMIREZ, Leica Microsystems, Heidelberg, Germany). Representative images have been selected by two independent investigators. The surface expression of Lamp-1 was evaluated in non-permeabilized hepatocytes by cytofluorometric analysis. For this purpose isolated hepatocytes were subjected to preconditioning treatment, stained in suspension for Lamp-1 and then analyzed with a fluorescent activated cytofluorometer (FACSCAN, Beckton Dickinson, Mountain View, CA). Similarly, fluorescence associated with intracellular CD was evaluated in permeabilized hepatocytes (by using the FIX & PERM kit from CALTAG Laboratories, Burlingame, CA) stained with anti-CD antibodies as above. Optimal permeabilization and intracellular fluorescent staining was set using actin as the reference antigen. At least 100,000 events were analyzed. The experiments were repeated twice. Based on the setting with cells labeled only with the secondary antibody, values lower than 101 arbitrary units of fluorescence intensity (abscissa axis) were considered negative. Cell positivity corresponds to the area below the

Lysosome exocytosis in hepatocyte preconditioning curve starting from values of fluorescence intensity higher than 101 arbitrary units and is given as a percentage of the total area. β-Hexosaminidase assay and CD western blotting Hepatocytes (106/ml KHH buffer) were incubated for 60 minutes at 37°C under control conditions after being preconditioned or not in the absence or the presence of inhibitors. The activity of the lysosomal β-hexosaminidase was assayed in hepatocyte homogenates (106 cells sonicated in 0.36 ml phosphate buffer containing 0.25% sodium desossicholate) and in incubation media. For the assay, 18 µl of cell homogenates and 50 µl of incubation media (corresponding to 50×103 hepatocytes and the respective secretion) were incubated for 60 minutes at 37°C in sodium-citrate buffer at pH 4.5 with the substrate p-nitrophenyl-N-acetyl β-D glucosaminide. Fluorescence was measured at 405 nm in a spectrofluorometer (Beckman DU530). This assay reveals only the mature β-hexosaminidase resident within endosomal-lysosomal organelles and therefore it is useful to monitor the exocytosis from these organelles. Enzyme activity was expressed as mU/mg of cell protein. Secreted activity is expressed as percent of total (intracellular plus extracellular) β-hexosaminidase. Enzyme assays were run in duplicate and repeated at least three times for each sample. Secreted CD molecular forms were revealed by standard western blotting techniques using specific rabbit anti-rat CD immune serum (Dragonetti et al., 2000). Proteins secreted by the hepatocytes were TCA-precipitated from aliquots of incubation media normalized per number of cells, separated by SDSpolyacrylamide (12.5%) gel electrophoresis and electroblotted onto nitrocellulose filter. CD-related bands were revealed by incubation with the anti-CD antiserum followed by a peroxidaseconjugated goat-anti-rabbit antibody and subsequent peroxidaseinduced chemiluminescence reaction as recommended by the manufacturer (Amersham). Intensity of the bands was estimated by densitometry. Determination of intracellular pH Cytosolic pH was measured as previously reported in detail (Carini et al., 1999; Carini et al., 2001b) using the fluorescent indicator dye 2′,7′-bis(carboxyethyl)-5,6-carboxyfluorescein-acetoxymethyl ester (BCECF-AM) (Molecular Probes, Eugene OR). For pH probe loading the hepatocytes were incubated for 30 minutes at 25°C in KHH buffer containing 5 µg/ml BCECF-AM. Calibration values were obtained for each experiment by incubating hepatocytes in media at different pH containing 10 µM K+/H+ ionophore nigericine and 120 mM K+. Fluorescence was measured at 450/530 nm wavelength pair using a Hitachi 4500 spectrofluorometer. Measurement of intracellular Na+ concentration Intracellular Na+ levels were measured as detailed previously (Carini et al., 1995; Carini et al., 2001b) using the fluorescent sodium-binding benzofuran isophthalate acetoxymethyl ester (SBFI-AM) (Molecular Probes, Eugene OR) as Na+ probe. Briefly, the hepatocytes were incubated for 60 minutes at 25°C in KHH buffer containing 10 µM SBFI-AM, washed and re-suspendend in fresh KHH medium for further treatments. At each time-point aliquots of hepatocytes were centrifuged and re-suspended in fresh medium for measurements. Changes in SBFI-AM fluorescence were monitored using the Hitachi 4500 spectrofluorometer set at 345 and 385 nm excitation and at 510 nm emission wavelengths. The ratio of fluorescence values at 345 nm and 385 nm excitation was calculated after correction for spontaneous SBFI-AM fluorescence. Calibration of SBFI-AM fluorescence was carried out with hepatocytes incubated in solutions of known Na+ concentrations in the presence of the Na+ ionophore gramicidin D (2 µM).

1067

Evaluation and chelation of free cytosolic calcium ions Cytosolic free calcium ions levels were determined by using the fluorescent cell permeable dye Fura2-AM (Sigma) as previously detailed (Carini et al., 1997b). Hepatocytes were loaded with this dye by a 15 minute incubation in KHH medium containing 4 µM Fura2AM. Cells were then washed to remove the excess and further incubated to allow complete de-esterification of Fura2-AM. Ca2+dependent Fura-2 fluorescence was measured with a computerassisted fluorometer (Perkin Elmer LS-5B) positioning the excitation wavelength alternately at 340 nm or 380 nm and the emission wavelength at 509 nm. Calibration was done by measuring the fluorescence in cells permeabilized with 10 µg/ml digitonin. Cytosolic free Ca2+ concentration was calculated assuming a Fura-2 Kd of 225 nM. To inhibit intracellular calcium signaling the membrane-permeable calcium chelant EGTA-AM was employed. For this purpose isolated hepatocytes were loaded with EGTA-AM (15 minutes at room temperature in KHH containing 25 µM EGTA-AM) prior to the preconditioning treatment. Hepatocytes were then processed for viability assay and fluorescence analysis in suspension (Lamp-1 surface expression) or on glass coverslip (CD subcellular localization) as described above. Statistics All experiments on cell viability, [Na+]i, pHi and [Ca2+]i concentrations were done in triplicate and repeated at least three times. Data were expressed as mean ± s.d. Statistical analysis was performed with the Instat-3 statistical software (GraphPad Software, San Diego, CA) using a one-way ANOVA test with Bonferroni’s correction for multiple comparisons when more than two groups were analyzed. Normality of data distribution of all groups was verified by the Kolmogorov and Smirnov test. Significance was taken at a P value of less than 0.005.

Results Preconditioning is associated with endosome and lysosome translocation to the cell periphery and fusion with plasma membrane We first sought to determine the effect of preconditioning on the subcellular localization of acidic vacuolar compartment. To this end, hepatocytes were either maintained for 20 minutes under oxygen fluxing (control) or preconditioned with a cycle of 10 minutes of hypoxia followed by 10 minutes of reoxygenation. Cells were then processed for confocal immunofluorescence analysis using the soluble protease cathepsin D (CD) and the membrane glycoprotein Lamp-1 as tracers of endosomal-lysosomal organelles (Démoz et al., 1999; Sarafian et al., 1998). The image in Fig. 1A (left panels) shows that, in control hepatocytes, endosomes and lysosomes are distributed throughout the cytoplasm with a preferential accumulation in the perinuclear region, in accord with their usual location in normal cells (Matteoni et al., 1987; Démoz et al., 1999). By contrast, in non-preconditioned hepatocytes killed by prolonged exposure to hypoxia CD fluorescence was greatly reduced and diffused in the cytoplasm, suggestive of lysosome rupture (not shown). The population of hypoxicpreconditioned hepatocytes appeared heterogeneous for endosome and lysosome localization. In almost 40 to 50% of the hepatocytes these organelles appear intact and many of them accumulate at the periphery of the cell, close to the plasma membrane (Fig. 1A, right panels). About 20 to 30% of the hepatocytes showed a distribution of CD-positive organelles

1068

Journal of Cell Science 117 (7)

Fig. 1. Preconditioning induces the translocation of endosomal-lysosomal organelles to the extreme periphery of the cell. Hepatocytes were incubated under control conditions prior to (control) or after being preconditioned by a cycle of hypoxia-reoxygenation (A) or by a 15 minute incubation with 1 µM CGS21680 (B). Cells were allowed to attach to glass coverslips, fixed and permeabilized, and processed for immunofluorescence confocal microscopy. Endosomes and lysosomes were identified by immunofluorescent detection of CD or Lamp-1. Representative images are shown. (A) Hypoxic preconditioning (PC) caused the dislocation of endosomes and lysosomes from the perinuclear region (see controls, Co) toward the cellular periphery. (B) CGS21680-preconditioned hepatocytes were stained for both CD (red fluorescence, panel b) and for actin (green fluorescence, panel c). Staining of cortical actin marked the cell border. Cell morphology can be appreciated in the phase contrast image (panel a). The translocation of endosomes and lysosomes to the extreme periphery of the cells can be appreciated in panel d, showing the overlap of actin and CD staining. The arrow in panel d points to the plasma membrane region in which cortical actin appears disassembled, as expected in the exocytosis process (Miyake et al., 2001), while lysosomal CD appears to be extruded from the cell (see also panels b and c). (C) Typical cytofluorograms of cell surface expression of Lamp-1 are shown. The positivity for this lysosomal membrane protein is increased in hepatocytes preconditioned by transient hypoxia (PC) or by CGS21680 treatment (CGS).

similar to that observed in controls, whereas the remaining hepatocytes (about 30%) appeared partially or totally devoid of lysosomal organelles (not shown). Similar pictures were obtained by immunostaining endosomes and lysosomes with antibodies specific for an integral membrane glycoprotein (Fig. 1A). The appearance of Lamp-1 on the cell surface proves that at least some of these organelles fused with the plasma membrane in preconditioned hepatocytes. We extended our investigation to a chemical-induced preconditioning system based on the stimulation of the adenosine A2A-receptor with the agonist CGS21680 (Carini et al., 2001a). In this case we also observed that endosomes and lysosomes localized to the periphery of the cell in a large fraction of hepatocytes (Fig. 1B). Hepatocytes were also stained for actin, which localizes close to the plasma membrane, to mark the cell border (Fig. 1Bc). The image in Fig. 1B (panel d), showing the overlap of cortical actin (green) and CD (red) staining, demonstrates that, in these hepatocytes, endosomes and lysosomes move to the extreme periphery of the cell upon preconditioning. To quantify the extent of lysosome-plasma membrane fusion we analyzed by cytofluorometry the expression of Lamp-1 on the cell surface in non-permeabilized control and preconditioned hepatocytes and obtained the following values (percent of positivity, average of two experiments in duplicate): control, 1%; hypoxicpreconditioned, 35%; CGS221680-preconditioned, 33% (Fig. 1C).

To corroborate these data we also analyzed by cytofluorometry the intracellular content of CD in permeabilized hepatocytes. Compared with controls, in preconditioned hepatocytes about 29% cells (average of two experiments in duplicate) were judged negative for intracellular CD (arbitrary units of fluorescence). Taken together, these data are consistent with the exocytosis of a large fraction of lysosomes that leads to the insertion of lysosomal membrane proteins in the plasma membrane and extracellular release of the lumenal content in preconditioned hepatocytes. This event occurred in about one-third of the hepatocytes subjected to preconditioning treatments. The morphological features described above occurred in living cells, as the preconditioning treatments do not affect cell viability (see below). Preconditioning-induced exocytosis of endosomallysosomal organelles was further demonstrated biochemically, based on the assumption that if fusion of these organelles with plasma membrane takes place then soluble enzymes normally confined within them should be found at higher levels in the extracellular milieu. In fact, the proportion of βhexosaminidase activity measured in the extracellular medium, compared with that found in the cell, was increased in hepatocytes preconditioned either by a brief hypoxicreoxygenation cycle or by exposition to the adenosine A2Areceptor agonist CGS21680 (Fig. 2A). Conversely, the activity measured within the cells was 31.27±4.3 mU/mg and

Lysosome exocytosis in hepatocyte preconditioning 23.12±3.8 mU/mg in control and preconditioned hepatocytes, respectively. Movement of endocytic vesicles requires the integrity of the cytoskeleton and involves various signaling enzymes. Cytochalasin D (Cyt D), which affects the actin cytoskeleton, and wortmannin (WM) and 3-methyladenine (3MA), which inhibit the lipid kinase PI3K, have been shown to interfere with the normal trafficking of endosomal-lysosomal organelles (Cordonnier et al., 2001; Brown et al., 1995; Punnonen et al., 1994). We therefore checked the efficacy of these drugs to inhibit endosome and lysosome translocation to the periphery and fusion with plasma membrane associated with preconditioning. As predicted, when hepatocytes were treated in the presence of 3MA (see below for details) the preconditioning-induced increase in β-hexosaminidase in the medium was completely prevented. In fact, the level of βhexosaminidase was much lower than that observed under basal conditions in control hepatocytes (Fig. 2A). This outcome is consistent with the morphological data showing the clustering of endosomes and lysosomes at one pole of the nucleus in preconditioned hepatocytes treated with 3MA (not shown). We attempted to better define whether fusion of CD-

1069

positive organelles with plasma membrane involved mainly endosomes or lysosomes. For this purpose we took advantage of the fact that the molecular forms of CD accumulate in different proportions in these organelles and can therefore be exploited as markers to discriminate between endosomes and lysosomes (Chiarpotto et al., 1999; Dragonetti et al., 2000). In rat cells CD is present as a 52 kDa precursor (proCD) within the endoplasmic reticulum and Golgi complex, as a 43 kDa mature single-chain in endosomal compartments, and as a 31 + 13 kDa mature double-chain in lysosomes (Démoz et al., 1999; Dragonetti et al., 2000). In preconditioned hepatocytes the extracellular release of the three CD molecular forms was nearly doubled, an effect completely reversed by WM, 3MA or Cyt D (Fig. 2B). Under basal conditions (control) hepatocytes released the three forms of CD (only the 31 kDa large chain of the double-chain is visible in the gel) in the medium, but in different proportions (Fig. 2C). In preconditioned hepatocytes the secretion of proCD (from preendosomal organelles) and the mature double-chain CD (from lysosomes) was stimulated by a factor of three, whereas secretion of the mature single-chain form of CD (typically resident within endosomes) was stimulated by nearly 1.5-fold

Fig. 2. Preconditioning induces the secretion of lysosomal hydrolases: inhibition by WM, 3MA and CytD. Isolated hepatocytes were subjected to hypoxic- or CGS21680-preconditioning (PC or CGS samples, respectively) or not (Co, controls) and further incubated for 60 minutes under control conditions. In some samples preconditioning and subsequent incubation were performed in the presence of 10 mM 3MA, 250 nM WM or 20 µM Cyt D as indicated (see the Materials and Methods section for details). The extracellular release of β-hexosaminidase activity (A) and CD protein (B) were determined in the incubation media by enzyme assay and western blotting, respectively. (A) There is an excess of secreted β-hexosaminidase activity in media from preconditioned hepatocytes compared with that from controls. This secretion is largely inhibited by 3MA. (B) The CD-related bands identified by western blotting were quantified by densitometry. Total CD found in media from preconditioned hepatocytes (PC and CGS lanes) was approximately double that found in media from control (Co) or preconditioned hepatocytes treated with 3MA, WM or Cyt D. Data in A and B are means±s.d. of three separate experiments. In A, the difference of PC vs Co and of CGS vs Co data was statistically significant (P