fax: (33) 5-62-74-45-32.
Liver Biology and Pathobiology
Opioid receptor blockade reduces Fas-induced hepatitis in mice
Article first published online: 22 SEP 2004
Copyright © 2004 American Association for the Study of Liver Diseases
Volume 40, Issue 5, pages 1136–1143, November 2004
How to Cite
Jaume, M., Jacquet, S., Cavaillès, P., Macé, G., Stephan, L., Blanpied, C., Demur, C., Brousset, P. and Dietrich, G. (2004), Opioid receptor blockade reduces Fas-induced hepatitis in mice. Hepatology, 40: 1136–1143. doi: 10.1002/hep.20428
- Issue published online: 14 OCT 2004
- Article first published online: 22 SEP 2004
- Manuscript Accepted: 5 JUL 2004
- Manuscript Received: 4 MAR 2004
- Institut National de la Santé et de la Recherche Médicale (INSERM)
- la Ligue contre le Cancer Tarn et Garonne
- Université Paul Sabatier, Toulouse III
Fas (CD95)-induced hepatocyte apoptosis and cytotoxic activity of neutrophils infiltrating the injured liver are two major events leading to hepatitis. Because it has been reported that opioids, via a direct interaction, sensitize splenocytes to Fas-mediated apoptosis by upregulating Fas messenger RNA (mRNA) and modulated neutrophil activity, we assumed that opioids may participate in the pathophysiology of hepatitis. Using the hepatitis model induced by agonistic anti-Fas antibody in mice, we showed that opioid receptor blockade reduced liver damage and consequently increased the survival rate of animals when the antagonist naltrexone was injected simultaneously or prior to antibody administration. Treatment of mice with morphine enhanced mortality. Naloxone methiodide—a selective peripheral opioid antagonist—had a protective effect, but the absence of opioid receptors in the liver, together with lack of morphine effect in Fas-induced apoptosis of primary cultured hepatocytes, ruled out a direct effect of opioids on hepatocytes. In addition, the neutralization of opioid activity by naltrexone did not modify Fas mRNA expression in the liver as assessed with real-time quantitative polymerase chain reaction. Injured livers were infiltrated by neutrophils, but granulocyte-depleted mice were not protected against the enhancing apoptotic effect of morphine. In conclusion, opioid receptor blockade improves the resistance of mice to Fas-induced hepatitis via a peripheral mechanism that does not involve a down-modulation of Fas mRNA in hepatocytes nor a decrease in proinflammatory activity of neutrophils. (HEPATOLOGY 2004.)
Apoptosis is a physiological process involved in the control of cell growth and the elimination of abnormal cells. Of the key molecules from which apoptotic signaling is generated, the Fas (APO-1/CD95) receptor plays a major role. The large tissue distribution of Fas receptor stresses on its crucial role in tissue homeostasis. Under physiological conditions, the Fas receptor is constitutively expressed on hepatocytes. In some hepatic diseases, such as chronic viral hepatitis and acute hepatic failure, liver injury has been associated with an exacerbated Fas-mediated apoptosis.1–3 Liver sensitivity to Fas-mediated apoptosis and consequently hepatic damage correlates with the expression level of Fas receptors on hepatocytes.4–6 Histological analysis of liver injury often reveals the presence of CD11b+ polymorphonuclear leukocytes. Neutrophil invasion is considered a secondary event initiated by primary liver impairment; nevertheless, the active contribution of neutrophils in the pathophysiological mechanisms leading to liver injury has been shown in experimental models.7–10
Many physiological and behavioral functions are under the influence of the opioid system. Opioid peptides are divided into three families—endorphins, enkephalins, and dynorphins—and act through three classes of receptors: the μ opioid receptor (MOR), δ opioid receptor (DOR), and κ opioid receptor (KOR). All three opioid receptor types are antagonized by naltrexone. Morphine—as well as endogenous opioids released under stressful situations—favors Fas-mediated apoptosis in the spleen. This proapoptotic effect elicited by MOR activation was associated with an up-regulation of Fas in splenocytes.11–13In vitro and in vivo experiments have also shown that opioids could alter the functional activity of neutrophils.14–16 The influence of opioids in Fas-mediated apoptosis and in neutrophil activity led us to investigate the involvement of opioids in the sensitivity of liver to apoptosis.
Materials and Methods
Treatment of Mice and Antibody Administration.
Eight- to 10-week-old female BALB/c mice (19–20 g of body weight) were purchased from Janvier (Le Genest Saint Isle, France). Animals were intraperitoneally injected with 350 μL PBS with or without 2 mg of the opioid receptor antagonist naltrexone hydrochloride (Sigma Chemical, St. Louis, MO). Other groups of mice were intraperitoneally injected with 1 mg of morphine (Francopia, Paris, France) in the presence or absence of naltrexone. Twenty-four hours later, all the mice were intraperitoneally injected with 6 μg of either hamster anti-mouse Fas Jo2 monoclonal antibody (mAb) (PharMingen, San Diego, CA) or hamster anti–keyhole limpet hemocyanin (KLH) Ha4/8 mAb (PharMingen). In other series of experiments, drugs including naloxone methiodide (Sigma Chemical)—an antagonist of all three opioid receptor subclasses that is unable to cross the blood–brain barrier—were injected following anti–Fas Jo2 antibody administration. All treatments were performed between 9:00 and 11:00 A.M. Starting from antibody administration, the survival rate of animals was monitored each hour until the tenth hour and daily for 1 week. Alanine aminotransferase activity in serum was quantified using a standard colorimetric method. All experiments involving animals were performed in compliance with the relevant laws and institutional guidelines (INSERM) and were approved by the local ethics committee (Midi-Pyrénées, France).
Livers removed from mice were immediately transferred into Duboscq fixative or 10% formol to be embedded in paraffin. Four-micrometer liver sections were then stained with hematoxylin-eosin. Cells with an apoptotic morphology characterized by chromatin and cytoplasm condensation and cell shrinkage were enumerated. For each sample, fifteen fields focusing on the portal tract were subjected to double-blind examination by two observers (M. J. and P. B.). Liver apoptosis severity expressed as a percentage was scored according to the ratio between the average of apoptotic hepatocytes and the average of hepatocytes in a normal control mouse.
Hepatocytes undergoing apoptosis were stained using rabbit anti–active caspase-3 mAb (clone C92-605, PharMingen). Formol-fixed liver sections were placed in 10 mmol/L Na-citrate buffer (pH 6) and heated twice to a boil for 10 minutes. Endogenous biotin was blocked before adding rabbit anti–active caspase-3 mAb for 30 minutes at room temperature. Bound anti–caspase-3 antibodies were revealed using biotin-labeled, swine anti–rabbit immunoglobulin (Ig) antibodies (DakoCytomation, Carpinteria, CA) and alkaline phosphatase–labeled streptavidin (StreptABComplexes/AP kit, DakoCytomation). Slides were then counterstained with hematoxylin.
DNA Fragmentation Analysis.
Genomic DNA was prepared from liver using the High Pure PCR Template Preparation Kit from Roche Molecular Biochemicals (Mannheim, Germany) according to the manufacturer's procedure. DNA samples were run in 0.7% agarose gel and revealed via ethidium bromide staining.
In Vivo Neutrophil Depletion.
Mice were intraperitoneally injected either with 100 μg of depleting rat anti–mouse Ly-6G and Ly-6C (Gr-1) (clone RB6-8C5; PharMingen) mAb or with 100 μg of control rat anti-DNP mAb (clone LO-DNP-11; LO/IMEX, University of Louvain, Brussels, Belgium). A second antibody injection was performed 16 hours later. Neutrophil depletion was determined via examination of blood smears stained with Giemsa.
Polymerase Chain Reaction Analysis.
Total RNA was isolated from previously crushed liver or brain by guanidine isothiocyanate-phenol-chloroform extraction using ready-to-use TRIzol reagent (GIBCO-BRL, Paisley, Great Britain). RNA was reverse-transcribed with moloney murine leukemia virus reverse transcriptase (GIBCO-BRL) using random hexamer oligonucleotides for priming. In negative controls the reverse transcriptase was replaced by H2O. Complementary DNA (cDNA) encoding MOR, DOR, and KOR were amplified using polymerase chain reaction (PCR) using the 5′-GTGTGTAGTGGGCCTCTTTGG-3′ forward and 5′-TCTATGGACCCCTGCCTGTATT-3′ reverse primers, the 5′-CCGGTACACCAAATTGAAGACC-3′ forward and 5′-AACTGGAGCATGCATACCACTG-3′ reverse primers, and the 5′-TGTTTGTCATCATCCGATACACG-3′ forward and 5′-AAACTGCAAGGAGCATTCAATG-3′ reverse primers, respectively. Forward and reverse primers for β-actin were 5′-TGGAATCCTGTGGCATCCATGAAAC-3′ and 5′-TAAAACGCAGCTCAGTAACAGTCCG-3′. The amplification procedure consisted of an initial denaturation step at 94°C for 2 minutes followed by 40 cycles of three steps of 45 seconds each, including denaturation at 94°C, primer annealing at 59°C, and primer extending at 72°C. β-Actin cDNA was amplified using a 25-cycle PCR protocol. All protocols included a final extending step at 72°C for 7 minutes. The fragment sizes for MOR, DOR, KOR, and β-actin were 414 bp, 381 bp, 401 bp, and 349 bp, respectively.
Fas messenger RNA (mRNA) expression level was quantified in liver from 7 PBS-treated and 7 naltrexone-treated BALB/c mice using real-time quantitative PCR. Primers for amplification of Fas cDNA, 18S rRNA, and hypoxanthine phosphoribosyl transferase (HPRT) cDNA were designed by using Primer Express Software (Applied Biosystems, Foster City, CA). Forward and reverse primers were 5′-AACATGGAACCCTTGAGCCA-3′ and 5′-AGGCGATTTCTGGGACTTTGT-3′ for Fas, 5′-AGTCCCTGCCCTTTGTACACA-3′ and 5′-CCGATTTCCTCACTAAACC-3′ for 18S rRNA, and 5′-CTGGTGAAAAGGACCTCTCG-3′ and 5′-TGAAGTACTCATTATAGTCAAGGGCA-3′ for HPRT. Amplification was performed with an ABI Prism 7000 Sequence Detector (Applied Biosystems) using the PCR SYBR Green sequence detection system (qPCR Mastermix plus; Eurogentec, Seraing, Belgium). Results were expressed as the number of Fas (CD95) mRNA copies normalized to 18S rRNA and HPRT mRNA molecules.
Primary Mouse Hepatocyte Culture.
Livers from normal BALB/c mice digested in situ with 0.03% collagenase I (56 U/mL) (Worthington Biochemical Corporation, Lakewood, NJ) were excised and minced in William's medium (GIBCO-BRL) supplemented with 10% fetal calf serum. Cell suspensions were filtered through a 100-μm filter, and hepatocytes were isolated via Percoll density centrifugation. Collagen-coated 24-well cell cultures (BD Bioscience, San Diego, CA) were seeded with hepatocytes (7 × 104/well) in William's medium 10% fetal calf serum. Twenty-four hours later, cells were washed and incubated in William's medium without fetal calf serum with or without drugs and/or anti-Fas antibody. Cellular viability was examined by quantifying metabolically active cells (CellTiter 96R AQueous One Solution Cell Proliferation Assay, Promega Corporation, Madison, WI). Viability of the cells cultured in medium alone was determined to be 100%.
Results and Discussion
The influence of the opioid system on liver sensitivity to Fas (APO-1/CD95)-mediated apoptosis was examined in normal (i.e., unstressed) BALB/c mice housed in conventional conditions. Mice weighting 19–20 g were intraperitoneally injected with either PBS or naltrexone 24 hours prior to agonistic anti–Fas Jo2 antibody administration (0.3–0.31 mg/kg of body weight). Nineteen (50%) of 38 PBS-treated mice died within 10 hours following injection of anti-Fas mAb, and 8 (21%) survived after 24 hours (Fig. 1). By contrast, 33 (85%) of 39 mice pretreated with naltrexone were still alive 10 hours after Jo2 administration, and 27 (69%) survived beyond 48 hours (P < .001, χ2 test). The beneficial effect of naltrexone was not observed when the amount of anti-Fas antibody administrated was higher than 0.35 mg/kg of body weight—a dose leading to the death of all mice within 10 hours (data not shown). No mortality was observed in mice injected with an irrelevant hamster anti-KLH mAb. Liver apoptosis was examined on liver section stained with hematoxylin-eosin or with a specific anti–active caspase-3 antibody. Livers were excised from mice 6 hours after Jo2 antibody administration (0.31 mg/kg). Hepatocytes exhibiting morphological characteristics of apoptosis on tissue section dyed with hematoxylin-eosin (Fig. 2A) or specifically labeled with an anti–active caspase-3 antibody (Fig. 2B) were enumerated. Percentages of apoptotic hepatocytes estimated via hematoxylin-eosin staining were similar to those determined via anti–active caspase-3 antibody immunostaining (Pearson's correlation > 0.96, P < .01; data not shown). For 8 (57%) of 14 PBS-treated mice, anti-Fas antibody administration resulted in apoptosis of less than 25% of hepatocytes. In liver from the 6 other animals (43%), the relative number in apoptotic hepatocytes was higher than 45%. By contrast, livers from all 12 naltrexone-treated mice exhibited less than 25% of apoptotic hepatocytes (Fig. 2C), (P = .02, ANOVA). Livers from naltrexone-treated mice surviving anti-Fas antibody treatment completely regenerated within 48 hours (data not shown). The contrast in liver apoptosis magnitude between the two groups was confirmed via serum alanine aminotransferase activity. Alanine aminotransferase levels in serum from naltrexone-treated mice were significantly reduced compared with those in serum from PBS-treated mice (P = .05, Mann-Whitney test) (Fig. 2D). Typical apoptotic DNA laddering was only observed in PBS-treated mice (Fig. 2E). No apoptosis was observed in mice injected with control hamster IgG. Thus, the neutralization of endogenous opioid activity in normal mice reduced the severity of Fas-induced hepatitis.
Because naltrexone improved Fas-induced hepatitis, it could be expected in turn that morphine aggravated it. The percentage of mice surviving anti-Fas antibody injection was lower in the morphine-treated animal group compared with that of the PBS-treated group. The difference in mortality of mice was observed as early as 10 hours following anti-Fas antibody administration (Fig. 3). Mice injected with naltrexone alone or together with morphine were similarly protected from hepatitis, suggesting that morphine and endogenous opioid effects were mediated via classical naltrexone-sensitive opioid receptors.
Numerous consensual studies have reported that opioids induce a reduction in the number of splenocytes in mice.11–13, 17 Opioid-induced splenocyte apoptosis was prevented by either opioid receptor antagonists or recombinant Fas-Ig or anti-FasL antibodies.11, 12 A similar opioid effect in adrenalectomized and sham-operated mice ruled out a mechanism dependent on the hypothalamic–pituitary–adrenal axis.12, 13 The major role of the MOR in the sensitization of splenocytes to Fas-mediated apoptosis was supported by the abrogation of the opioid-induced spleen atrophy in MOR knockout mice.13, 17In vitro, experiments performed on immune cells—including lymphocytes and macrophages—showed that a direct interaction with morphine resulted in up-regulation of Fas mRNA.11, 18 Together, these data assume that spleen cell sensitization to Fas-mediated apoptosis induced by opioids was a paracrine/autocrine MOR-mediated effect associated with an increase in Fas mRNA. By comparison, we investigated whether the neutralization of opioid receptor activity was associated with a lower expression of Fas mRNA in the liver. Fas mRNA was quantified via real-time PCR in the livers of PBS-treated and naltrexone-treated mice. For each mouse, the number of Fas mRNA copies was normalized to the HPRT housekeeping gene mRNA and to 18S ribosomal RNA. As shown in Fig. 4, the relative amounts of Fas (CD95) mRNA in the liver did not significantly differ between the two groups of mice (P > .05, Mann-Whitney test). Accordingly, a modulation of Fas mRNA synthesis was unlikely to account for the protective effect of naltrexone simultaneously injected with anti-Fas antibodies in mice (Fig. 5A). The inefficiency of naltrexone to prevent death in mice injected with anti-Fas antibodies 2 hours earlier (Fig. 5A) highlights the primordial contribution of endogenous opioids at the beginning of Fas-induced hepatitis. Taken together, our data indicate the protective effect of opioid receptor blockade in Fas-induced hepatitis model is not due to a down-modulation of Fas mRNA expression in the liver.
Activation of opioid receptors within the central nervous system may stimulate the hypothalamic–pituitary–adrenal axis and the sympathetic nervous system, resulting in glucocorticoid and catecholamine release, respectively.13, 19, 20 However, the demonstration that glucocorticoids as well as β-adrenergic agonists protected hepatocytes from apoptosis21–24 was incompatible with an indirect mechanism involving neutralization by naltrexone of central opioid receptors. This presumption was attested by investigating the potency of naloxone methiodide, an antagonist of all the three opioid receptor subclasses unable to cross the blood–brain barrier, to protect mice against Fas-induced hepatitis. Most of the mice injected with PBS plus anti-Fas antibody and all of the mice injected with both morphine and anti-Fas antibody died, while those administered with anti-Fas antibody and morphine plus naloxone methiodide survived (Fig. 5B). However, although the proapoptotic effect of opioids was dependent on peripheral opioid receptors, none of the three receptor subclasses antagonized by naloxone methiodide was detected in livers from normal mice as assessed using reverse-transcriptase PCR (Fig. 6A). This absence of opioid receptors in normal livers indicates that the opioid effect antagonized by naltrexone did not involve a direct opioid receptor–mediated interaction with hepatocytes. Considering that opioids are metabolized and could be taken up by the liver through transporters such as organic anion-transporting polypeptide,25–28 the enhancing effect of morphine on Fas-induced apoptosis was further investigated on primary-cultured mouse hepatocytes. As has been reported previously,29, 30 anti-Fas antibody alone did not modify the viability of mouse hepatocytes in vitro (Fig. 6B). Cytolytic activity of anti-Fas antibody against primary-cultured hepatocytes was triggered by actinomycin D (Fig. 6C). The direct killing activity of anti-Fas antibody in the absence as well as in the presence of actinomycin D was not altered by morphine, suggesting that the sensitization of liver to Fas-induced apoptosis in vivo does not involve a direct action of opioids on hepatocytes. Together, these data exclude a mechanism of action of opioids similar to the prevalent one involved in opioid-induced spleen cell apoptosis.
Extravasated neutrophils were observed in damaged livers from mice injected with anti-Fas antibody. Quantification of extravasated neutrophils in liver from PBS-treated mice and naltrexone-treated mice revealed a striking relationship (Pearson's correlation > 0.96; P < .01) between the number of infiltrating neutrophils and the number of apoptotic hepatocytes (Fig. 7A). The correlation found in PBS-treated mice and in naltrexone-treated mice was similar. The absence of neutrophils in livers from control mice injected with the irrelevant anti-KLH antibody (Fig. 7B) and in nonapoptotic livers from PBS- and naltrexone-treated mice injected with anti-Fas antibody (Fig. 7A) suggests that neutrophil infiltration is initiated by hepatocyte apoptosis. Our observations are in agreement with previous findings showing that hepatocyte apoptosis induced by injecting normal mice with moderate doses of agonistic anti-Fas antibody results in CXC chemokine release responsible for the recruitment of neutrophils.31 At these antibody doses, both anti-Fas antibodies and infiltrating neutrophils already described as promoting epithelial cell apoptosis8, 32, 33 might contribute to hepatocyte death. We then investigated whether neutrophils expressed opioid receptors. CD11b/Gr-1– bearing granulocytes were isolated from the blood of 10 normal mice via cytofluorometric cell-sorting. Reverse-transcriptase PCR experiments failed to amplify cDNA-encoding opioid receptors in a preparation containing more than 98% CD11b/Gr-1–bearing cells, mainly neutrophils (data not shown). This latter result—excluding circulating polymorphonuclear leukocytes as target cells for opioids—was in agreement with our observation that livers from PBS-treated mice and naltrexone-treated mice with a comparable rate of apoptotic hepatocytes displayed a similar neutrophilic infiltration (Fig. 7A, C, D). The reduction in the number of extravasated neutrophils in livers from naltrexone-treated mice was associated with the decrease in hepatocyte apoptosis (Fig. 7A), but not with an alteration of the initial neutrophilic recruitment (Fig. 7C, D). Because it could not be excluded that opioid receptors were found in activated neutrophils that penetrated into damaged livers, an inhibition by naltrexone of their activity, such as NO release,34–36 remained plausible. To assess such a paracrine opioid effect, we compared normal mice and neutrophil-depleted mice to determine the ability of morphine to magnify Fas-induced liver apoptosis. Mice were injected twice with either granulocyte-depleting anti–Gr-1 mAb or irrelevant anti-DNP mAb prior to morphine administration. Anti–Gr-1 antibody treatment resulted in the depletion of 84 + 6% peripheral blood neutrophils as assessed by Giemsa staining. Twenty-four hours later, mice were injected with anti-Fas antibodies and liver apoptosis was estimated. Morphine-mediated amplification of hepatocyte apoptosis was not abrogated in the absence of neutrophils (Fig. 8).
Taken together, our results show that endogenous opioids as well as morphine sensitize hepatocytes to Fas-induced apoptosis through a mechanism that does not involve opioid receptors within the central nervous system. In contrast with the mechanism involved in spleen cell apoptosis, the peripheral proapoptotic effect of opioids in the liver is neither mediated via opioid receptors expressed on parenchymal liver cells nor associated with upregulation of Fas mRNA synthesis. Neutrophilic infiltration following the initial Fas-induced apoptosis of hepatocytes is not altered by naltrexone treatment. In addition, the persistence of the morphine effect in the absence of neutrophils suggests a mechanism of action that is independent of neutrophils. It has been reported that hepatocyte apoptosis could be generated by initial Fas triggering in nonparenchymal liver cells.37, 38 Considering that nonparenchymal cells—including endothelial cells and Kupffer cells—express opioid receptors,39, 40 it could be hypothesized that the opioid effect is associated with a modulation of their sensitivity to Fas-induced apoptosis.
In summary, our study suggests that a preventive peripheral neutralization of opioid activity could be instrumental in attenuating Fas-induced hepatitis and, conversely, that morphine treatment is detrimental to patients with hepatic impairment.
The authors wish to thank Michel March for technical assistance and Drs Isabelle Maridonneau-Parini and Claire Racaud-Sultan for helpful discussions.