INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Sepsis is the leading cause of death in intensive care units, affecting approximately 400,000 patients in the United States annually, and its incidence continues to increase. Despite a better understanding of its pathophysiology, mortality of sepsis remains severe (1). Numerous experimental data have provided a solid rationale for innovative therapies and clinical trials, which have to date not improved the outcome of the syndrome. Differences that exist between animal studies and human disease probably explain why novel agents are efficient in animals, yet fail in clinical trials of sepsis. Experimental models of sepsis are largely based on animal exposure to endotoxin (LPS). Since rodents are constitutively resistant to LPS, Galanos and coworkers (2) demonstrated nearly 20 years ago that mice treated with D-galactosamine (D-GalN) were dramatically sensitized to LPS, allowing reduction of more than 2,500-fold the lethal dose of endotoxin. Based on this observation, several recent studies have been designed and published in prestigious journals using the LPS/D-GalN model (3). However, we addressed the question of whether these studies were relevant to septic shock. Leist and colleagues (8) indeed suggested that LPS in D-GalN-sensitized mice was due to tumor necrosis factor alpha (TNF-) toxicity under conditions of blocked gene transcription, and that accounted for TNF--mediated liver apoptosis and acute liver failure.

Recently, considerable interest has been focused on the TNF- apoptotic pathway, especially because of the involvement of the cysteine protease of the interleukin (IL)-1 converting enzyme (ICE) family referred to as caspases. ICE, referred to as caspase-1, was first described as the enzyme that cleaves inactive pro-IL-1 to mature IL-1 (9). ICE and ICE-related proteases have been further identified as cysteine proteases involved in several, if not all, apoptotic pathways, including the Fas (CD95) and the TNF pathway (10, 11). Eleven caspases have been identified to date in human and in mouse, and allocated into three subgroups with respect to their divergent substrate specificity. Reminiscent of the coagulation or the complement system, the caspases are arranged in a proteolytic cascade that serves to transmit and amplify numerous death signals. Among the terminator caspases has been clearly underlined the role of caspase-3 (CPP32)-like proteases, especially in the liver (12). Direct inhibition of ICE and ICE- related proteases is now available with the use of synthetic peptides, such as YVAD-aldehyde (YVAD-CHO) or YVAD-chloromethylketone (YVAD-CMK), which bind to and block the catalytic site of ICE proteases (12). Using ICE inhibitor YVAD-CMK, we and others recently demonstrated its protective effects on Fas (CD95)-mediated liver apoptosis in vitro as in vivo (13, 14). Characterization of this ICE protease inhibitor prompted us to hypothesize whether YVAD-CMK would prevent caspases activation and lethality induced by low-dose LPS in D-GalN sensitized mice, thus confirming that this experimental setting accounts for a caspase-dependent fulminant apoptotic hepatitis induced by TNF-, and not for septic shock.

We also addressed the question of whether YVAD-CMK could also protect from endotoxic shock induced by high-dose LPS in mice unsensitized with D-GalN. Besides its determinant role during apoptosis, ICE is also involved in the pathophysiology of sepsis (15). Excessive production of the proinflammatory cytokine IL-1 might constitute an important harmful factor during sepsis. This point has been underlined by: (1) the protective effect of a strategy aimed at inhibiting IL-1 signaling by IL-1Ra (16); and (2) the observation that ICE-deficient mice challenged with lipopolysaccharide do not release any IL-1 and are resistant to lethal endotoxic shock (17). Moreover, YVAD-CHO was recently demonstrated to suppress, but only transiently, the release of IL-1 in vivo in a murine endotoxic shock without improving survival (18).

We addressed the question of whether pharmacologic inhibition of ICE might be beneficial during sepsis, not only by blocking IL-1 release but also by inhibiting ICE and ICE- related pro-apoptotic proteolytic activities. In distinguishing between these two effects of YVAD-CMK on these two models of sepsis, we wondered whether we could definitely establish that LPS challenge combined with D-GalN does not account for septic shock, but for acute TNF--related liver apoptotic destruction.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Animals and Reagents

Six- to eight-week-old female C57Bl/6 mice (18-20 g) obtained from IFA-CREDO (Paris, France) were used. All experimental procedures were carried out in compliance with French government regulations (Service Vétérinaires de la Santé et de la Production Animale, Ministère de l'Agriculture). Escherichia coli endotoxin (LPS, serotype 0127:B8) and D-galactosamine hydrochloride (D-GalN) were purchased from Sigma Chemical Co. (St. Louis, MO) and reconstituted in pyrogen-free sterile saline. Highly purified recombinant murine TNF- was purchased from Genzyme (Paris, France) and reconstituted in pyrogen-free sterile phosphate-buffered saline (PBS). Tetrapeptide YVAD-CMK (Ac-Tyr-Val-Ala-Asp-chloromethylketone) was purchased from Bachem Biochimie SARL (Basel, Switzerland) and reconstituted in pyrogen-free sterile PBS (4 µmoles in 200 µl).

Experimental Protocol

Mice were injected intraperitoneally with either LPS alone (40 mg/kg), LPS (10 µg/kg or 50 µg/kg) in combination with D-GalN (800 mg/kg), or murine recombinant TNF- (20 µg/kg) with D-GalN (800 mg/kg) in a solution of 200 µl per dose. YVAD-treated mice received YVAD-CMK in 200 µl sterile PBS either 2 h before, simultaneously, or 2 h after LPS or TNF- administration. Control groups in the LPS, LPS/D-GalN, and TNF- received 200 µl sterile saline intraperitoneally. Mice were exsanguinated by retro-orbital puncture 90 min and 4 h after LPS challenge. EDTA plasma and serum were separated immediately and frozen at 80°C until analyzed. Liver damage was evaluated by measuring serum aminotransferases (ALT, alanine aminotransferase and AST, aspartate aminotransferase) using a standard clinical automatic analyzer (Hitachi, type 7150). Survival and clinical signs were assessed on at least eight animals per group. Twenty mice treated with YVAD-CMK 2 h after the lethal LPS/D-GalN challenge were randomly (n = 5) killed 10, 15, 20, and 24 h after LPS/D-GalN.

In Vivo Circulating TNF- and IL-1 Level

Plasma concentrations of TNF- and IL-1 were measured in duplicate experiments by ELISA kits (Genzyme) in eight animals per group.

DNA Internucleosomal Fragmentation Analysis

For electrophoretic detection of DNA internucleosomal fragmentation, minced liver tissue was digested in lysis buffer with 0.1 mg/ml proteinase K for 18 h at 55° C with shaking. DNA was phenol/chloroform extracted and electrophoresed on 2% agarose in TBE buffer.

Hepatic Caspase-1-like and Caspase-3-like Activity

Liver lysates were prepared by Dounce homogenization in a hypotonic buffer (25 mM Hepes, pH 7.5, 5 mM MgCl2, 1mM EGTA, 1 mM PMSF, 1 mg/ml leupeptin and aprotinin). Homogenates were centrifuged at 15,000 rpm for 15 min and the supernatants were used. Protein concentration of samples was determined by the Bradford method using bovine serum albumin as standard. Fifty µg of the extracted proteins were tested in duplicate experiments with the Promega Caspace Kit (Promega, Madison, WI) with 1 mM of fluorescent substrates for caspase-1 ICE (Ac-Tyr-Val-Ala-Asp-aminomethylcoumarin, YVAD-AMC) and caspase-3 CPP32 (Ac-Asp-Glu-Val-Asp-aminomethylcoumarin, DEVD-AMC) in 0.1 ml ICE standard buffer assay (100 mM Hepes-KOH buffer, pH 7.5, 10% sucrose, 0.1% CHAPS, 10 mM DTT) at 30° C for 1 h. In order to assess the specific contribution of caspase-1-like and caspase-3-like enzyme activities, assays were performed in the presence and absence of selective inhibitors for either ICE or CPP32. Levels of AMC released by enzymatic reaction were measured using a spectrofluorometer using an excitation wavelength of 360 nm and an emission wavelength of 460 nm. The fluorescence intensity was calibrated with increasing standard concentrations of AMC. The difference between the substrate cleavage activity levels in the presence and absence of caspases inhibitor reflected the contribution of either ICE or CPP32-like enzyme activity. Caspase activity was calculated from the slope of the AMC calibration recorder trace, and was expressed in picomoles per minute per milligram of protein. Caspase activity was measured for five animals per group and per time point.

Histology and TUNEL Assay

Livers were excised and immediately transferred to formalin-acetic acid-alcohol fixatives. Samples cut at 3 µm were stained with hematoxylin eosin staining (HES). For YVAD-CMK-treated animals challenged with LPS/D-Gal, histologic examination of the liver harvested at various times of sacrifice 1 to 7 d after experimental procedure was realized. Liver sections from different lobes were assessed for apoptotic process severity (marked condensation of chromatin and cytoplasm, cell shrinkage, and apoptotic bodies) by an observer unaware of the treatment group assignment. Apoptosis severity (%) was evaluated by a ratio of the mean number of apoptotic hepatocytes upon total hepatocytes of 10 fields. Apoptotic cells were determined using an in situ Cell Death Peroxydase Detection Kit (Boehringer, Meylan, France; terminal deoxynucleotidyl transferase-mediated dUTP nick end labeling, TUNEL Kit), according to the manufacturer's protocol. Color revelation was realized with diaminobenzidine (DAB). DNase I treatment of additional sections were performed and served as positive controls. Specificity of the TUNEL assay was tested with samples for which the terminal transferase (TdT) was not added on the slides. Heart, spleen, and lungs were also harvested, transferred to formalin-acetic acid-alcohol fixatives, embedded in paraffin, and 3-µm sections were stained with HE for analysis.

Statistical Analysis

All data are given as mean ± SE. Chi-square, Student's t test and two-way ANOVA analysis followed by the Bonferroni t test were performed. p Value < 0.05 was considered significant.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Effect of YVAD-CMK after LPS Challenge in D-galactosamine-sensitized Mice

One hundred percent lethality was induced in D-GalN sensitized mice within 10 h with LPS doses of 10 µg/kg, a 4,000-fold lower than the lethal dose of LPS required when injected without D-GalN (Table 1). Histopathologic analysis performed on liver sections taken from control animals 6 h after D-GalN/LPS challenge showed characteristic features of apoptosis (> 95%), widespread destruction of liver architecture, and erythrocyte agglutination (Figure 1a). Apoptosis in the liver was clearly confirmed at the same time, since DNA internucleosomal fragmentation was detected on agarose gel electrophoresis (Figure 2a, lane 2), and numerous apoptotic cells were evidenced by the TUNEL assay, as shown in Figure 2b. Careful histologic examination of other organs (heart, spleen, lung) revealed no injury, underlining the high and unique sensitivity of the liver in this TNF-dependent model.

View larger version (149K): [in this window] [in a new window]   Figure 1.   Histology of liver sections from control and YVAD-CMK-treated mice after D-GalN/LPS administration (original magnification: ×500). (a) Control mouse liver 6 h after a lethal challenge shows massive pan-lobular and multi-lobular apoptotic necrosis greater than 95%, peliosis, and erythrocyte agglutination. (b) Liver section from D-GalN-sensitized mice co-injected with LPS and YVAD-CMK displays few scattered apoptotic bodies. No necrosis nor neutrophilic and lymphoid infiltrates are observed. Liver sections from two different YVAD-CMK post-treated mice 15 h after D-GalN/LPS challenge show focal rare apoptosis (c) or massive apoptosis greater than 50% (d ).

View larger version (122K): [in this window] [in a new window]   Figure 2.   (a) Liver DNA agarose gel electrophoresis 6 h after LPS/D-GalN challenge, demonstrating internucleosomal fragmentation for control mice (lane 2), and absence of DNA laddering in YVAD-CMK treated mice (lane 3). One Kb DNA size marker is shown in lane 1. (b) TUNEL assay 6 h after LPS/D-GalN challenge in control animals, demonstrating numerous apoptotic hepatocytes with DAB dark brown positive staining nuclei, erythrocytes, and complete destruction of liver architecture. (c) TUNEL assay 6 h after LPS/D-GalN challenge in YVAD-CMK-treated animals, showing only a few scattered apoptotic hepatocytes, with complete preservation of liver architecture.

Strikingly, mice treated with YVAD-CMK before or simultaneously to D-GalN/LPS injection were completely resistant to the lethal effect of endotoxin (Table 1). Moreover, even when given an LPS dose more than five times greater than the lethal dose in the control group, YVAD-CMK still prevented death (Table 1). By marked contrast, liver apoptosis in YVAD-CMK-treated mice was strikingly inhibited, evidenced in isolated foci, and evaluated to be less than 5%. Liver architecture was completely preserved (Figure 1b), whereas no neutrophil or lymphoid infiltrates were observed. The apoptotic process was dramatically inhibited by YVAD-CMK treatment, since no DNA fragmentation was evidenced on agarose gel electrophoresis (Figure 2a, lane 3), a technique that lacks sensitivity. However, some isolated apoptotic hepatocytes were observed by a more sensitive method, TUNEL assay, and are shown in Figure 2c. No signs of hepatic injury were further evidenced by histologic examination of the liver harvested from 1 to 7 d after the LPS/D-GalN challenge, confirming that YVAD-CMK conferred almost full protection. The histologic appearance of the other organs analyzed was also normal, suggesting that YVAD-CMK, at the dosage used, had no toxic effect, as already observed in our previous study (13). Finally, hepatic enzymes, which are known to be highly correlated to the extent of liver injury in this model (8), showed dramatic increases within 8 h in control mice, whereas only a slight and significantly less important increase was observed in YVAD-CMK- treated mice (Figure 3).

View larger version (10K): [in this window] [in a new window]   Figure 3.   AST and ALT serum levels (mean ± SD) 6 h after LPS/ D-GalN in control and YVAD-CMK pretreated mice. Experiments were performed in eight animals for each group. *p < 0.05.

YVAD-CMK still conferred a significant protection (70%) when administered 2 h after lethal challenge (Table 1). Moreover, death that occurs within 8-10 h in control mice was significantly delayed up to 24 h for the nonsurviving YVAD-CMK-treated mice. To investigate whether a curative effect of YVAD-CMK might be missed because of its short half-life, we treated mice 24 h before D-GalN/LPS lethal challenge with YVAD-CMK. We still observed the same protective effect in this model (data not shown). These data led us to hypothesize that, rather than a limited half-life time of this anti-caspase, its partial loss of effect in post-treatment might be related to its inability to completely block the already switched-on cascade of activated caspases. We therefore killed YVAD-CMK post-treated mice randomly between 10 and 24 h after LPS challenge and examined their liver. In 70% of mice, liver appeared almost completely protected, as apoptotic landmarks were evaluated to less than 5% (Figure 1c). For the remaining mice, apoptotic process (evaluated to 40 to 60% of the liver at 15 h) seemed to have only been delayed, which could explain why these animals ultimately die from liver failure (Figure 1d).

Inhibition of Caspase-3-like Dependent Activity by YVAD-CMK in LPS/D-GalN Challenge

Administration of LPS/D-GalN in the control group induced a time-dependent increase in caspase-3-like activity, as evidenced by the proteolytic cleavage of the caspase-3 substrate Ac-DEVD-AMC presented in Figure 4. A 16-fold increase in caspase-3-like activity was measured 6 h after lethal challenge, fully consistent with the detection of DNA laddering and the presence of apoptotic injury in the liver at the same time (Figures 1 and 2). The decrease in hepatic caspase-3-like activity observed at H7 was related to liver destruction, hemorrhage, and leakage of caspase as well as transaminases in the plasma (A.M., manuscript in preparation). By contrast, no caspase-1-like activity was observed, since no release of the caspase-1 substrate Ac-YVAD-AMC was detected at any time point in the control group (Figure 4). Note that neither D-GalN alone nor LPS (low dose or high dose) caused an increase in caspase-1-like and caspase-3-like activity, or DNA internucleosomal fragmentation in the liver (data not shown). By contrast, caspase-3-like protease activity was completely inhibited by YVAD-CMK, while remaining without effect on the baseline caspase-1 activity (Figure 4). This anti-caspase effect of YVAD-CMK was fully consistent with the absence of significant apoptotic liver injury detected by plasma transaminases measurements, liver histologic analysis, DNA agarose gel, and TUNEL assay.

View larger version (17K): [in this window] [in a new window]   Figure 4.   Caspase-1-like and caspase-3-like activity in the liver. Hepatic caspase activity was determined by measuring the proteolytic cleavage of the caspase-1-like substrate Ac-YVAD-AMC and the caspase-3 substrate Ac-DEVD-AMC after LPS/D-GalN challenge in control mice (black bars) or YVAD-CMK-treated mice (white bars). Experiments were performed in five animals for each group and each time point. *p < 0.05.

Effect of YVAD on TNF--mediated Fulminant Liver Failure

D-galactosamine-sensitized mice challenged with a lethal dose (20 µg/kg) of murine recombinant TNF- (rTNF-) were treated with either saline or YVAD-CMK. Time course and histopathologic features of liver injury and death induced by LPS in D-GalN-sensitized mice were completely reproduced with rTNF- treatment in D-GalN-sensitized mice, with a slight delay (about 2 h) accounting for the LPS-induced release of TNF- by Küpffer cells (Table 1). Death occurred in control rTNF--treated mice within 8 h, preceded by liver enzyme release, caspase-3 increased activity in the liver, and massive liver apoptosis (data not shown). By contrast most mice treated with YVAD-CMK before or simultaneously with rTNF- challenge 100% and 75%, respectively, were protected from liver destruction and death (Table 1). Moreover, we still observed a 50% protection of mice treated with YVAD-CMK 2 h after the lethal administration of rTNF- (Table 1). The decreasing ability of YVAD-CMK to rescue from TNF--induced hepatic injury when injected 2 h after rTNF- challenge is coherent with the hypothesis that the ongoing activation of proteases cascade finally reaches a point of no return (10).

Effect of YVAD-CMK on Endotoxic Shock

We further addressed the question of whether YVAD-CMK would protect from an endotoxic shock induced by high-dose LPS in unsensitized mice. Mice pretreated 2 h before LPS challenge with either saline or YVAD-CMK were injected intraperitoneally with a lethal dose of endotoxin (40 mg/kg). Mice from both groups experienced lethargy, febrile shaking, and diarrhea, all features of a systemic inflammatory response. Multiple organ failure and death were observed similarly in control and YVAD-treated mice within 24-48 h (Table 1). Careful histologic examination did not reveal any liver apoptotic injury contrasting with diffuse alveolar inflammation and hemorrhage. Note again the important differences that exist between this high-dose LPS model, with mortality observed within 48 h, and the LPS/D-GalN model, where mortality occurs within 10 h. Moreover, multiple organ injury is observed in the LPS challenge, whereas exclusive liver apoptotic injury is only evidenced in the LPS/D-GalN model.

Effect of YVAD-CMK on Cytokine Release after LPS Challenge

Curiously, YVAD-CMK did not affect systemic IL-1 release nor, as expected, TNF in the high-dose LPS challenge (Figure 5). A typical time course for the production of these cytokines in response to LPS was observed in both control and YVAD-CMK-treated mice, with TNF- and IL-1 reaching the peak level at 90 min and 4 h, respectively. The absence of YVAD-CMK effects on IL-1 release was also observed in the low-dose LPS challenge in D-GalN sensitized mice (Figure 5).

View larger version (24K): [in this window] [in a new window]   Figure 5.   TNF and IL-1 plasma concentrations 1.5 h and 4 h in control untreated and YVAD-CMK-treated mice after LPS and LPS/ D-GalN (mean ± SD). Experiments were performed in eight animals for each group.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Our study underlines the striking difference existing between two widely recognized models of lethal endotoxemia. First, as previously reported (2), treatment of mice with D-galactosamine dramatically increased their sensitivity to the lethal effects of lipopolysaccharide several thousand-fold. The onset of lethality after treatment with D-galactosamine and lipopolysaccharide occurred faster than with lipopolysaccharide alone; all animals died within 8 h, by contrast with the delayed mortality observed in the LPS group. LPS-challenged animals without D-GalN were febrile, weak, prostrate, and lethargic, suffering from diarrhea. LPS on its own induced a systemic proinflammatory process but no specific liver injury, while other organ injuries, such as diffuse alveolitis and hemorrhage, resulted in multiple organ failure and death within 24 to 48 h.

By contrast, LPS toxicity in D-GalN-treated mice actually resulted exclusively from severe apoptotic liver injury and complete destruction, and not from the systemic inflammatory response, as it was previously reported (8). There is now full agreement in the literature that death in the D-GalN/LPS model is due to TNF- toxicity and accounts for TNF--mediated liver apoptosis and acute liver failure (8). LPS as well as recombinant TNF- alone, at doses as low as 10 µg/kg, and 20 µg/kg, respectively, do not induce any cytotoxicity. Hepatotoxicity of low-dose LPS and TNF- is indeed associated with a lethal apoptotic process initiated only under the condition of blocked transcription (8). It is known that TNF- induces apoptosis in hepatocytes as in other cell types, provided that the gene transcription is blocked with transcriptional inhibitors such as actinomycin D. This transcriptional inhibition appears to specifically block numerous nuclear factor-kappa B (NF-B)-dependent genes induced by TNF, which are known to counteract the pro-apoptotic effects of TNF-R1 activation (19). Several NF-B-dependent genes candidates exist, among them the Mn SOD and NOS2, but as yet are not fully identified. It is known that D-galactosamine induces a selective transcriptional block in hepatocytes. This amino sugar is exclusively metabolized in hepatocytes, leading to a selective depletion of uridine nucleotides with severe transcriptional arrest, beginning 0.5 h after the injection, persisting approximately 3 h, and recovering after 5 h (8). This information was previously suggested by Galanos and colleagues because sensitization by D-GalN disappeared when lipopolysaccharide was administered 4 h after D-galactosamine (2).

Neutralization of TNF- by either monoclonal antibodies, inhibitor of TNF- processing, or anti-TNF- pharmacologic agents, such as chlorpromazine, tunicamycin, quinine, or linomide (5, 6, 22), completely prevents mortality in this model. LPS challenge in D-galactosamine-sensitized mice knock-out for the TNF-55 kD type I receptor and TNF- itself is completely non-lethal (4, 25, 26).

Our present data on protection conferred by YVAD-CMK from TNF--induced caspase-3-dependent liver apoptosis, liver failure, and death provide a new evidence of the involvement of TNF- cytotoxicity in this model. Indeed, YVAD-CMK beneficial effects, at the high dose used in this study, resulted from its broad spectrum anti-caspase activity in the liver, as we and others previously observed in Fas- and transforming growth factor beta (TGF-)-mediated liver apoptosis (13, 14, 27). Besides structural and functional homologies, Fas, TNF-, and TGF- appear to recruit similar or at least partially overlapping pathways, among which are the activation of cysteine proteases of the IL-1 converting enzyme (ICE) family, referred to as caspases, and which represent the downstream event that leads to the irreversible stage of apoptosis (11, 27, 28).

The beneficial properties of YVAD-CMK in the D-GalN/ LPS model did not result from any effect on systemic cytokine release. The absence of effects on TNF release, combined with the inhibition of hepatic caspase-3-like activity, clearly indicated that YVAD-CMK beneficial action in D-GalN/LPS accounted for its ability to block the apoptotic cytotoxicity of already released proinflammatory cytokines. YVAD-CMK did not inhibit the LPS-induced release of IL-1 and TNF- in LPS/D-GalN-treated mice, in contrast to other known inhibitors of LPS/D-GalN-mediated lethal shock that diminish cytokine production, namely TNF- (5, 6). This is an important point because complete inhibition of TNF- release may be deleterious (29), and inhibition of proteases after induction of sepsis can block or overcome the cytotoxicity of already released proinflammatory cytokines.

YVAD-CMK did not inhibit IL-1 release in our experiments in both LPS/D-GalN and high-dose LPS challenged mice (Figure 5). This result was unexpected, since the YVAD-CMK aldehyde counterpart, YVAD-CHO, was previously shown to suppress the release of IL-1 in vivo in a murine model of endotoxic shock (18). However, this reversible inhibitor only led to a transient decrease of circulating IL-1 levels and did not affect LPS-induced lethality (18). Note also that no increase in hepatic caspase-1 activity was observed in LPS/ D-GalN challenged control mice (Figure 4), as well as in high-dose LPS-treated mice (data not shown), whereas IL-1 secretion in both groups was detected in the plasma (Figure 5). This point raises the question of whether Küppfer cells should actually be the most important cells accounting for IL-1 production in response to LPS. Recently, Schumann and coworkers (30) reported on the LPS-induced caspase-1 activation and IL-1 release by cultured monocytic THP-1 cells. Caspase-1 activity was blocked in vitro by YVAD-CMK, resulting in a moderate inhibition of IL-1 release after LPS stimulation. We therefore concluded from this study that YVAD-CMK, since having effect in vitro, could probably target monocytes/ macrophages in vivo. However, YVAD-CMK, at the dose used in our in vivo study, did not inhibit IL-1 plasmatic release in response to LPS. This suggests that either YVAD-CMK was unable to block monocytes/macrophages caspase-1 activity in vivo, or that IL-1 can be released in vivo in a caspase-1-independent manner.

During the course of our studies, very similar results were reported by Jaeschke and associates (31), demonstrating that caspase-3 was activated in the liver in response to Salmonella endotoxin in D-galactosamine-challenged CH3HeB/LPS-sensitive mice, accounting for TNF--dependent hepatic apoptosis, neutrophil recruitment, and subsequent necrosis (31). Furthermore, no increase in hepatic caspase-1 activity was detected, definitely ruling out a critical role for caspase-1 in TNF--induced liver apoptosis by contrast with Fas pathway (14). By using a broad spectrum anti-caspase inhibitor, zVAD-FMK, at very important doses, these authors could inhibit caspase-3 activation and apoptosis, but they did not study the protective effect of this anti-caspase inhibitor in a lethal experiment. However, the results of this study, using different genetic background, endotoxin, and anti-caspase inhibitor, underline clearly the important role of TNF-induced apoptosis in the liver in response to LPS/D-galactosamine, which indeed accounts for caspase-3-dependent fulminant liver failure, and not for septic shock.

TNF- plays a pivotal role in the pathogenesis of several inflammatory diseases, including sepsis. By giving evidence of the in vivo inhibition by YVAD-CMK of TNF--mediated apoptosis, our results, combined with others (31), provide a significant insight in the therapeutic approach of TNF-- related diseases. Future prospects for caspases inhibitors appear clearly devoted to the treatment of acute liver diseases. The role of TNF- and Fas (CD95) in several liver diseases is continuously underlined in either septic (32), T-cell-dependent inflammatory and viral liver injury (8, 33), toxic liver disease (34), ischemia/reperfusion injury (35), or liver graft rejection (36). Very recently, an immunomodulator, linomide, was also demonstrated to provide protection from both TNF-- and Fas-mediated liver apoptosis (5, 37). Note that the inhibition of TNF- release or activity, which occurs during linomide treatment (5), is also known to be deleterious in some experimental models (29). Moreover, the precise mechanisms of action of linomide and its ability to provide a curative effect during TNF-- and Fas-mediated liver failure remain unknown. By contrast, YVAD-CMK, which does not affect cytokine production, can block or overcome the cytotoxicity of already released proinflammatory cytokines. Our results with YVAD-CMK inhibition of both TNF-- and Fas-mediated hepatocyte apoptosis provide strong evidence that a simple galenic formulation of caspases-inhibiting drugs constitutes a new promising therapeutic strategy for acute liver disease involving uncontrolled apoptosis.