Potential conflict of interest: Nothing to report.
Caspase-8 has been implicated in signaling for apoptotic cell death and for certain nonapoptotic functions. However, knowledge of actual physiological or pathophysiological processes to which this enzyme contributes is lacking. Using a mouse model and employing the conditional knockout approach to delete the caspase-8 gene specifically in the liver, we found that caspase-8 deficiency in hepatocytes facilitates infection of the liver by Listeria monocytogenes, attenuates the hepatocyte proliferation wave during the first 48 hours after partial hepatectomy and, depending on the genetic background of the mice, prompts a chronic inflammatory response to the hepatectomy, as a result of which the proliferation of hepatocytes, although initially suppressed, might later be persistently enhanced, resulting in significant hepatomegaly. Conclusion: These findings indicate that caspase-8 participates in regulation of the cellular response to infection and injury and that it does so by affecting various cellular functions, including cell death, cell proliferation, and induction of inflammation. (HEPATOLOGY 2007.)
The caspase family of cysteine proteases is mainly known for its pivotal role in the induction of apoptosis in animal cells.1 Some of the caspases, characterized by a “prodomain” region located upstream of the proteolytic moiety, serve an initiating role in apoptosis. They become activated upon binding of their prodomains to death-inducing receptors or to adapter proteins associated with such receptors, and once activated they cleave other members of the caspase family, thereby activating them. Caspase-8 (MACH/FLICE/Mch4) is an initiator caspase activated within signaling complexes of receptors of the TNF/NGF family, to which it is recruited by the binding of its prodomain to an adapter protein called Fas-associated death domain (FADD; also called MORT1).2–4 Activation of caspase-8 constitutes a crucial initiating event in the apoptotic death mechanism induced by these receptors (the extrinsic cell-death pathway).5 Both caspase-8 and FADD also contribute, by mechanisms as yet unknown, to various nonapoptotic cellular processes (e.g., see references 5-16).
Although the in vivo functioning of caspase-8 has been explored using a number of transgenic mouse models,5, 13, 14, 16 we still know very little about the enzyme's physiological or pathophysiological significance. Using mice subjected to conditional knockout of caspase-8 by the cyclization recombination enzyme (Cre)/loxP approach, we recently showed that specific deletion of caspase-8 in hepatocytes endows these cells with resistance to the cytotoxic effect of the receptor Fas.14 In this study, we further explored the physiological significance of the deletion of caspase-8 in hepatocytes on liver functions. We show that such deletion compromises the resistance of mice to infection by Listeria monocytogenes, an intracellular pathogen whose eradication from infected hepatocytes depends largely on the ability of cytotoxic T lymphocytes to kill these cells by activating death receptors, particularly Fas. We also show that regeneration of the liver after partial hepatectomy (PHx), a model system for studying tissue response to injury, occurs abnormally when the hepatocytes are devoid of caspase-8. The burst of hepatocyte proliferation triggered by the hepatectomy is significantly decreased in these mice; and, depending on their genetic background, it might also be accompanied by onset of a chronic inflammatory state. These findings implicate caspase-8 in the regulation of several kinds of infection-induced and injury-induced cellular responses, including cell death, cell growth, and inflammation.
Alb, albumin; Cre, cyclization recombination enzyme; Casp, caspase; FADD, Fas-associated death domain; GdCl, gadolinium chloride; H&E, hematoxylin and eosin; PBS, phosphate-buffered saline; PBST, PBS+0.05% Tween 20; PHx, partial hepatectomy; p-STAT 3, phospho signal transducer and activator of transcription 3; TBS, Tris-buffered saline; TBST, TBS+0.05% Tween 20.
Caspase-8 Deletion in Hepatocytes Compromises Resistance to Listeria Infection.
The liver is a major site of replication of L. monocytogenes, a Gram-positive bacterium that invades the cytoplasm of eukaryotic cells and multiplies in it. Eradication of this pathogen from the liver after experimental infection of mice is therefore widely used as a model system for studying mechanisms of immune defense against intracellular pathogens.17 To examine the contribution of caspase-8 function to the immune response in hepatocytes, we compared the recovery from infection at different time points after intravenous inoculation of a sublethal dose of L. monocytogenes in mice with caspase-8–deficient hepatocytes (Casp8F/−:Alb-Cre) to that in their control littermates (Casp8F/+:Alb-Cre).
One day after infection, bacterial loads in the organs of the Casp8F/−:Alb-Cre mice were similar to those in their control littermates. On the 6th day, however, bacterial titers in both the livers and spleens of the Casp8F/−:Alb-Cre mice were 10-fold to 100-fold higher than in the controls (Fig. 1A,B). By the 14th day, the pathogen had been totally cleared from both livers and spleens of the control mice, whereas titers in the livers of the Casp8F/−:Alb-Cre mice remained high (Fig. 1C). The prolonged infection resulted in inflammation (Fig. 1D,E) and development of necrotic lesions in the liver (Fig. 1F,G), as well as increased proliferation of hepatocytes (Fig. 1H,I). By the 6th day after infection, about 15% of the Casp8F/−:Alb-Cre mice, but none of the control mice, died.
Effects of Caspase-8 Deficiency in Hepatocytes on Recovery from PHx: Attenuation of the Early Growth Response.
To evaluate the contribution of caspase-8 to tissue recovery from injury, we assessed the effect of caspase-8 deletion from hepatocytes on liver regeneration after PHx. Consistent with prior reports, we found that PHx prompts a burst of hepatocyte proliferation (reviewed in Fausto et al.18). In the livers of Casp8F/−:Alb-Cre mice, however, proliferation, and also the induction of several molecular changes associated with the G1/S transition (increased expression of cyclin A, D, and E, and phosphorylation of the retinoblastoma protein) occurred to a significantly lesser extent than in their control littermates (Fig. 2). This decrease was observed after 1/3 PHx (Fig. 2A,C–E, and left part of F), as well as after 2/3 PHx (Fig. 2B and right part of F), which leads to more robust and better synchronized DNA synthesis and more effective progression through the cell cycle.19 Because mortality during the first few hours after resection was significantly higher after 2/3 PHx, all subsequent analyses of the effect of caspase-8 deficiency were restricted to recovery of mice from 1/3 PHx.
Effects of Caspase-8 Deficiency in Hepatocytes on Recovery from PHx: Improved Recovery of the Ischemic Lesion Site.
The strict control of cell growth in vivo is impressively manifested by the ability of the liver to maintain its normal size and the accurate recovery of its original size following resection.20, 21 In various pathological conditions, however, this control fails, resulting in abnormal enlargement (hepatomegaly).22–24 To further assess the effect of caspase-8 deficiency on recovery of the liver after PHx we monitored changes in liver volume by the use of MRI. Prior to hepatectomy, the average liver volume in the Casp8F/−:Alb-Cre mice was identical to that in the Casp8F/+:Alb-Cre mice (data not shown). After PHx, however, the 2 groups differed significantly in their kinetics of liver growth (Fig. 3C). On T1-weighted spin echo MRI, 2 regions were distinguishable in the hepatectomized livers: the ischemic lesion site that was generated as a consequence of the resection, and which was found on histological analysis to contain necrotic tissue as well as apoptotic cells (Supplementary Fig. 1), whose area gradually decreased during regeneration; and the rest of the liver, which increased in size to compensate for the loss of the dissected tissue. The decrease in size of the ischemic lesion site in the Casp8F/−:Alb-Cre mice was significantly more rapid than that in the control mice (Fig. 3A-C), suggesting that absence of caspase-8 in the hepatocytes promotes rapid healing or adsorption of the tissue at that site.
Effects of Caspase-8 Deficiency in Hepatocytes on Recovery from PHx: Persistent Late Hepatocyte Proliferation.
Surprisingly, although the initial proliferative response of the caspase-8–deficient hepatocytes was milder than that of the normal hepatocytes, the increase in size in the rest of the liver in the Casp8F/−:Alb-Cre mice was not slower than that in their control littermates, but significantly faster. Moreover, whereas the liver in control mice stopped growing once it reached its original size, the eventual size of the liver in the Casp8F/−:Alb-Cre mice was significantly larger than normal (Fig. 3D), reaching 120% of the pre-PHx size.
On assessing hepatocyte proliferation at a later stage after hepatectomy, we found that whereas in the control mice cell proliferation in the liver subsided after the initial burst, proliferation of the hepatocytes of the Casp8F/−:Alb-Cre mice persisted for several weeks after resection; thus eventually, despite its initial suppression, it significantly exceeded that of the normal mice (Fig. 2A and Fig. 2C, lower panels). A similar late persistent post-PHx increase was observed in hepatocyte levels of cyclin A and E (Fig. 2F).
Effects of Caspase-8 Deficiency in Hepatocytes on Recovery from PHx: A Chronic Inflammatory Response.
The present observation that Casp8F/−:Alb-Cre hepatocytes continued to proliferate long after PHx would appear to be consistent with the continuously increasing size of the liver in these mice at a late stage after hepatectomy. However, the fact that this more rapid volume increase was already discernible a few days after hepatectomy (Fig. 3D), when the proliferation rate of the caspase-8–deficient hepatocytes was still lower than normal, suggested that additional factors contribute to this difference as well.
A functional MRI protocol combined with hypercapnia and hyperoxia provides a sensitive measure of perfusion and hemodynamic alterations resulting from a variety of pathological changes.25 In this study, in control mice PHx was followed by a decrease in both ΔSco2 and ΔSo2, reflecting a decrease in liver vascularity and blood content. In contrast, in the livers of Casp8F/−:Alb-Cre mice, 4 days after PHx both parameters increased (Fig. 4), as a result of increased blood volume and flow. Such an increase was found to occur in association with an inflammatory state (H. Barash, unpublished data). Histological analysis indeed revealed an increased content of leukocytes in the livers of the hepatectomized Casp8F/−:Alb-Cre mice, indicative of inflammation. Staining with the anti-F4/80 antibody indicated that the accumulating leukocytes (8.5 ± 4.3% of the area of histological sections of the liver at 14 days post-PHx, compared to 2.3 ± 0.7% in control, Casp8F/+:Alb-Cre mice), were macrophages (Fig. 5A,B). Western blot and immunohistochemical analyses revealed a significant increase in phosphorylated STAT-3 both in these macrophages and in the hepatocytes of the Casp8F/−:Alb-Cre mice (Fig. 5B,C).
As mentioned above, inflammation and enhanced hepatocyte proliferation in the Casp8F/−:Alb-Cre mice were also observed after their infection with L. monocytogenes (Fig. 1D,E,H,I). Our mice were kept in a specific pathogen-free facility; nevertheless, in view of this effect of Listeria infection, we attempted to further exclude the possibility that the inflammation and enhanced liver growth observed after PHx reflect an effect of some pathogen that escaped our notice. Accordingly, we repeated the experiments with mice that were rederived by cesarean section, placed with foster gnotobiotic mothers, and then maintained in germ-free isolators. PHx of these rederived mice initiated the same chronic inflammatory state as that observed before rederiving (data not shown).
Effects of Caspase-8 Deficiency in Hepatocytes on Recovery from PHx: Persistent Late Hepatocyte Proliferation Occurs as a Consequence of the Chronic Inflammatory Response.
Comparative histological analyses of different sections from the livers of hepatectomized Casp8F/−:Alb-Cre mice during the time of persistent hepatocyte proliferation disclosed that variation between sections in the extent of cell proliferation correlated with the numbers of macrophages that accumulated in the corresponding regions (Fig. 5D), suggesting that hepatocyte proliferation and the inflammatory state of the liver are causally related. To further examine this relationship, we injected mice with GdCl, an agent that induces transient depletion of Kupffer cells,26 on days 10 and 12 after PHx, the time at which hepatocyte proliferation and the Kupffer-cell content were at their highest. In addition to substantially decreasing the inflammatory cells that had accumulated in the Casp8F/−:Alb-Cre liver, this treatment also practically wiped out the increase in hepatocyte proliferation in these mice (Fig. 5D, bottom panel; Fig. 6). Moreover, the excessive increase in the size of their livers (probably due in part to an inflammation-related increase in blood volume and flow and in Kupffer cell numbers, and in part to an increase in hepatocyte number that mainly occurs at that time) was curtailed. These findings indicate that the constitutive proliferation of hepatocytes in Casp8F/−:Alb-Cre mice at a late stage after hepatectomy and the excessive increase of liver size are consequences of the persistent inflammation that occurs in their livers.
Effects of Caspase-8 Deficiency in Hepatocytes on Recovery from Partial Hepatectomy: Occurrence of the Chronic Inflammatory Response Varies with Genetic Background.
Our experiments were first performed with mice of mixed genetic background. Once we had established conditional deletion in mice of pure C57Bl/6 background (see Supplementary Materials and Methods), we examined the consequences of caspase-8 deficiency in hepatocytes for the recovery from PHx in these mice as well. As in the mice of mixed background, the early proliferative response to PHx in C57Bl/6 mice with caspase-8–deficient hepatocytes was blunted (Fig. 2). The latter mice, however, did not exhibit an inflammatory response to the PHx, nor did they display a protracted enhancement of cell proliferation (data not shown).
The concordant variation observed in the occurrence of inflammation and protracted growth, depending on genetic background, further indicated that these two changes are causally related. It also indicated that the impact of caspase-8 deficiency on the initial proliferative response to PHx is independent of its pro-inflammatory consequence.
Caspase-8 was shown in this study to affect cellular responses to infection and injury by contributing to a number of different cellular functions. Absence of caspase-8 in hepatocytes attenuated the resistance of mice to Listeria infection. It also affected, in several ways, regeneration of the liver after PHx. Healing of the lesion site during the first few days after hepatectomy occurred more rapidly in the absence of caspase-8. In contrast, throughout the rest of the liver significantly less cell proliferation was observed at that time than in the control littermates not lacking caspase-8. In addition, in mice with mixed genetic background, we found the effect of caspase-8 deficiency was reversed after this period of initial growth stimulation in the liver. Rather than ceasing to proliferate once the liver reached its original size, the caspase-8–deficient hepatocytes kept on proliferating, resulting in an abnormally enlarged liver. This sustained growth was associated with substantial accumulation of macrophages and activation of STAT-3, both characteristic of inflammation. Injection of GdCl, which interferes with the function of the Kupffer cells and induces their elimination,27 resulted in arrest of the delayed hepatocyte proliferation, indicating it occurs as a consequence of the inflammation. This arrest is particularly notable in view of the fact that the hepatocyte proliferation facilitated early after PHx does not decrease, but rather is enhanced by GdCl treatment, apparently as a consequence of the induction of TNF by GdCl.28, 29
The contribution of caspase-8 to immune defense against pathogens is consistent with our notion of the physiological role of the extrinsic cell-death pathway to which it contributes. The liver is the main site of clearance of Listeria from the circulation and is also a major site of persisting Listeria infection. Arrest of the infection is therefore largely dependent on the ability of cells of the immune system to kill infected hepatocytes. Whereas early resistance to the pathogen depends mostly on a local innate immune response involving Kupffer cells and neutrophils, final clearance of the bacteria is mediated by T lymphocytes. These cells too employ different armaments at different stages of the operation of immune defense. At an early stage of infection, CD8+ T cells kill Listeria-infected hepatocytes via the effects of perforin and granzymes.17 Because of their ability to directly activate mitochondrial apoptotic mechanisms and effector caspases, granzymes cause death in a way that does not depend on the initiating role of caspase-8. Consistently, absence of caspase-8 in the hepatocytes seemed to have no effect on the yield of Listeria in the early stages after infection. Later, however, CD8+ T cells employ the Fas ligand for cell killing.30 Fas triggers death by activating caspase-8; therefore, hepatocytes that are devoid of this enzyme are completely resistant to Fas-induced death.14 The larger amounts of Listeria observed here on the 6th day of infection of mice with caspase-8–deficient hepatocytes, and the persistence of infection in the liver even on the 14th day, probably reflect the failure of T lymphocytes to eliminate infected hepatocytes through Fas-induced triggering of the apoptotic function of caspase-8. Continuous release of Listeria from the infected hepatocytes, and the consequent recurrent infection of cells in the spleen, apparently accounted for the increased titer of the pathogen observed on the 6th day of infection not only in the caspase-8–deficient liver but also in the spleen, despite normal caspase-8 levels in that organ.
A contribution of caspase-8 to other mechanisms of defense against Listeria cannot however be excluded. As shown in a recent analysis of the effect of caspase-8 deficiency on B lymphocyte function, caspase-8 also contributes to the signaling activity of Toll receptors,15, 16 which play an important role in the immune response to infecting bacteria.
The apoptotic function of caspase-8 might well also account for the more rapid rate of recovery of the lesion site in the Casp8F/−:Alb-Cre mice after hepatectomy. Our preliminary histological analysis disclosed that the lesion contained large numbers of dying cells. We also detected in this region activation of caspase-3, as well as the accumulation of T lymphocytes whose cytotoxic activity might contribute to caspase-8–dependent cell death (Supplementary Fig. 1). Further analysis of the impact of caspase-8 deficiency on this region should clarify whether the more rapid decrease in size of the lesion in mice with caspase-8–deficient hepatocytes reflects decreased cell death, enhanced adsorption, or some other functional effect of caspase-8 deficiency.
The observed attenuation of the initial growth response to hepatectomy as a result of caspase-8 deficiency is reminiscent of the way caspase-8 contributes to lymphocyte growth. In B lymphocytes, caspase-8 contributes to the growth induced by certain growth-stimulating agents but not by others,16 and in T lymphocytes caspase-8 deficiency compromises growth only in certain differentiation states.13, 14 Likewise, the decreased growth response of the hepatocytes as a consequence of caspase-8 deficiency is only partial and transient, indicating that caspase-8 participates not in the growth mechanisms themselves but rather in one of several alternative mechanisms by which the growth process is regulated. Both in lymphocytes and in hepatocytes, the absence or the arrest of function of FADD (an adapter protein that is recruited together with caspase-8 to receptors of the TNF/NGF family for death induction) results in growth suppression similar to that observed in the absence of caspase-8.6, 7, 8, 31 This suggests that, as in death induction, the growth-promoting effect of caspase-8 in these cells depends on its association with FADD.
Only by in vitro studies of the effect of caspase-8 deletion on the function of isolated hepatocytes can the mechanisms underlying the findings of this study be fully deciphered. However, the present investigation illustrates the unique advantage of in vivo studies of the function of a protein. By applying cell-type-specific deletion of a protein to explore its function in a particular cell type in vivo, it is possible to determine how the deleted protein affects interactions of this cell with cells of other types. In this way, we found that hepatectomy of livers with caspase-8–deficient hepatocytes eventually triggers inflammation. The functional changes underlying this inflammation are likely to involve different types of cells, engaged in a complex dialogue that cannot be discerned by studying their individual functions in culture: the caspase-8–deficient hepatocytes trigger, probably through the generation of chemokines and other inflammatory mediators, accumulation and activation of macrophages, which in turn induce, perhaps via other mediators, persistent growth of the hepatocytes.
In fact, the other functional consequences of caspase-8 deficiency that were noticed in this study might also not be fully understood without taking into account the interactions of cells of different types within the liver. The initial growth-stimulatory response triggered by hepatectomy is inflicted by mediators, at least some of which are generated within the site of the lesion and in response to it.32 The more rapid healing of the lesion observed when the hepatocytes lack caspase-8 might well result in reduced generation of these growth-stimulatory mediators. Thus, in addition to a possible cell-autonomous effect of caspase-8 deficiency in hepatocytes on their growth, this deficiency might also affect their growth indirectly, by altering the generation of mediators that stimulate this growth. Likewise, we cannot exclude the possibility that the enhancement of Listeria growth in livers with caspase-8–deficient hepatocytes reflects not only a cell-autonomous deficiency (resistance of the hepatocytes to T cell–induced death), but also deficiency of other, more complex defense mechanisms, namely activation of immune cells by the infected hepatocytes in a way that depends on the expression of caspase-8 in the latter cells. There is indeed evidence that the activity of apoptotic caspases within cells can potentiate the immune response against them (see, for example, Albert et al.33 and Casares et al.34).
The mechanisms underlying the effect of caspase-8 deficiency on the homeostasis of liver cells are of particular interest in view of the evidence that in several kinds of cancers, including hepatocellular carcinoma, caspase-8 is frequently deficient or mutated.35, 36–41 Our finding that caspase-8 deficiency facilitates a chronic inflammatory state and persistent cell growth in response to injury might provide a clue to the way in which such deficiency promotes tumor development.
Like extracellular regulators, intracellular signaling proteins can also be found to demonstrate evolutionary preservation of a general physiological function that is common to the different functional consequences that they inflict in different organisms.42 Up to now, caspase-8 has been mainly implicated in the control of cell death. The findings of the present study indicate a more general role for caspase-8 as a regulator of defensive responses against infection and injury, both through apoptotic and through nonapoptotic functions. This notion is consistent with the involvement of the Drosophila caspase-8 and FADD orthologs in the induction of defense mediators,43 and with the evolution of mammalian caspases that serve specifically either to catalyze the generation of the inflammatory mediators IL-1, IL-18, and IL-33 (caspases 1, 4, 5, 1144) or to suppress it (caspase-1245). It is also in line with the involvement of the paracaspase and metacaspase families (which are evolutionarily related to the caspases) in various aspects of the cellular response to infection, injury, and stress.46
We thank Dr. Steffen Jung for advice and help in studying the recovery of the mice from infection with Listeria monocytogenes; Dr. Ori Brenner for advice on histological analysis; Drs. Eugene Varfolomeev and Alon Harmelin for helpful advice; Dr. Eli Pikarsky's laboratory (Hadassah University Hospital) for helpful tips on immunostaining; Shoshana Grossfeld for maintenance of the mice; Jadi Natan and Calanit Raanan for assistance in preparation of slides for histology; Inna Kolesnik, Tatiana Shalevich, and Dvir Mintz for genotyping the mice; and Shirley Smith for scientific editing.