Potential conflict of interest: Nothing to report.
Only half of patients with chronic hepatitis C virus (HCV) infection and genotype-1 show a sustained antiviral response to the current antiviral therapy. The reason this treatment fails is unclear, and no reliable marker exists that predicts the treatment outcome. In the present study, we investigated the apoptotic activation of caspases in HCV patients undergoing antiviral therapy with regard to the treatment outcome. We determined caspase activation in sera from patients who were either responding or nonresponding to antiviral therapy by using two novel caspase assays, an immunological and a luminometric enzyme test. We found that compared with nonresponding individuals, responding patients showed significantly (P < .05) increased caspase activity, which was closely correlated with virus elimination (r = 0.81). The cutoff value of serum caspase activity was determined, which correctly predicted the treatment outcome with a sensitivity of 70% and a specificity of 82% (area under the curve 0.845; 95% CI). In conclusion, hepatic caspase activity might play a role in HCV clearance and could also predict the efficacy of antiviral therapy. (HEPATOLOGY 2006;43: 1311–1316.)
Chronic hepatitis C virus (HCV) infection is characterized by inflammatory liver damage and a long viral persistence associated with an increased risk of developing cirrhosis and hepatocellular carcinoma. The treatment options for patients infected with HCV, especially for those with genotype 1, are not satisfactory. Although antiviral therapy has been improved with the introduction of pegylated interferon (IFN) and ribavirin, about 50% of patients do not respond to treatment.1, 2 In addition, treatment is expensive and associated with significant side effects. So far, however, no predictive response markers exist, and it remains unclear which patients could benefit from antiviral therapy.
There is increasing evidence that apoptosis plays an important role in HCV-associated liver damage.3–6 Apoptosis is essentially controlled by caspases, a family of intracellular cysteine proteases that cleave a variety of protein substrates, thereby inducing the demise of the cell. Among the different substrates are proteins involved in DNA replication, cell-cycle control, signal transduction, and structural proteins including members of the cytokeratin type I family.7, 8 There is also increasing evidence that caspases have different nonapoptotic functions including regulation of the immune response.9
Recent data show that activation of caspases is considerably up-regulated in HCV-infected liver and correlates with the grade of inflammatory disease activity and early stages of fibrosis.10, 11 By employing a novel ELISA that selectively recognizes a caspase-generated neoepitope of cytokeratin-18 (CK-18), which is abundantly expressed in hepatocytes, we recently demonstrated that caspase activation can also be detected in serum from patients with HCV infection, but not in those of healthy control individuals. Interestingly, serological detection of caspase activation was associated with more progressive fibrotic liver alterations, even in patients with normal aminotransferase levels.12
Caspases can be activated by two principal pathways, namely, the extrinsic and the intrinsic death pathway.13 The intrinsic pathway is initiated at the mitochondrion by the release of cytochrome c, whereas the extrinsic pathway involves death receptors such as CD95 or tumor necrosis factor–related apoptosis-inducing ligand (TRAIL) receptors. Although hepatocyte apoptosis may occur by a variety of mechanisms, death-receptor-mediated apoptosis is a particularly prominent process in the liver.14 Both CD95 and its ligand CD95L have been shown to be up-regulated in HCV infection.15, 16 Moreover, IFNs modulate apoptotic pathways in the liver by several mechanisms. IFN-γ but also IFN-α/β induce expression of death receptors, death ligands, and certain caspases and thereby sensitize cells to apoptosis.17, 18 These gene-inducing effects of IFNs are largely controlled through the STAT pathway. Additional important targets of the IFN response are the 2′,5′-oligoadenylate system and double-stranded RNA-dependent protein kinase PKR. PKR is not only involved in the translational control of viral proteins but also in apoptosis signaling by up-regulating several death-promoting molecules such as CD95, CD95L, or caspase-8.19 Thus, the IFN response is closely associated with caspase activation and apoptosis, giving rise to several amplification loops.
A current key question is whether caspase activation and apoptosis are involved in the resolution or persistence of HCV infection. Since IFNs play a major role in the treatment of HCV infection, we asked whether the responsiveness to antiviral therapy, i.e. the resolution of HCV infection, is associated with induction of caspase activation. We therefore compared the intrahepatic caspase activity in sera from patients either showing a sustained antiviral response, or a relapse or no viral clearance during IFN therapy. Our data demonstrate that patients who show a sustained response reveal significantly elevated serum levels of caspase activity compared to patients that failed to clear the HCV infection.
All patients were treated between March 2000 and August 2002 in a controlled study according to the guidelines of the Ethics Committee at the Hanover Medical School.20 We selectively recruited 23 HCV patients with no other cause of liver disease, 11 of whom had a sustained antiviral response and 12 of whom had a nonsustained antiviral response (Table 1). All patients had genotype 1 except for one patient in the responder and one patient in the relapser group, who had genotype 3. HCV disease activity was assessed according to the index of ISHAK.21 There were no significant differences in the stage of fibrosis (ISHAK F), histological grade of inflammation (ISHAK A-D), or virus load between the patient groups (Table 1). The presence of anti-HCV antibodies and HCV RNA was quantified by an ELISA (Abbott Diagnostics, Germany) and the Amplicor HCV Monitor test (Roche Diagnostics, Germany), respectively. Genotyping of HCV was performed by a reverse hybridization assay (Inno LiPA HCV II; Innogenetics, Belgium). Patients received daily injections of 9 μg of consensus interferon (Yamanouchi Pharma GmbH, Heidelberg, Germany) and a standard dose (1000/1200 mg per day) of ribavirin.20 Four patients from both the responder and the nonsustained responder groups, in addition, had received a higher initial dose of IFN (18 μg/d for the first 8 weeks), which did not influence the sustained antiviral response compared to patients treated with an initial lower IFN dose.20 The treatment was performed for 48 weeks (responder patients) and stopped in patients with detectable viremia at month 6 (nonresponder patients). Sustained antiviral response was defined as HCV-RNA negativity in serum 24 months after treatment. Relapse summarized patients with an initial treatment response (HCV-RNA negativity) but HCV recurrence after 48 weeks. Serum samples from patients with either a sustained viral response (n = 11, mean age 44 ± 3), a viral relapse (n = 5, mean age 53 ± 3.1) or no viral response (n = 7, mean age 51 ± 2.8) were taken before therapy, on days 1, 3, and 14, as well as after 24 weeks and at the end of treatment after 48 weeks. The sera were stored at −20°C.
Table 1. Clinical, Virological, and Histological Characteristics of Randomly Selected Patients
Number of patients
Age (mean ± SEM
48 ± 1.9
44 ± 3
51 ± 2.8
53 ± 3.1
Sex (% male)
Stage of fibrosis ISHAK F (mean ± SEM)
3.0 ± 0.4
2.8 ± 0.6
3.9 ± 0.8
2.3 ± 0.9
Necroinflammatory score ISHAK A-D (mean ± SEM)
5.0 ± 0.5
4.6 ± 0.9
5.0 ± 0.3
6.0 ± 0.95
Mean viral load before therapy (× 106/mL ± SEM)
2.7 ± 0.6
2.1 ± 0.5
3.6 ± 1.6
2.7 ± 1.2
Serological Detection of Caspase Activity.
Caspase activity in serum samples was determined by two independent methods. For the quantitative measurement of the apoptosis-associated neoepitope in the C-terminal domain of cytokeratin-18 (amino acids 387–396), we used the M30-Apoptosense ELISA kit (Peviva, Sweden)12, 22 according to the instructions of the manufacturer. In addition, we established a novel luminescent substrate assay for active caspase-3 and caspase-7 (Caspase-Glo assay, Promega, Germany).22 The assay provides a luminescent substrate with the caspase cleavage sequence DEVD in a reagent optimized for caspase activity and luciferase activity. Following cleavage of the substrate at the DEVD peptide by caspase-3 and caspase-7, aminoluciferin is released, resulting in light production in a luciferase reaction, which can be measured in relative light units (RLU) by a luminometer. First, patients' sera were diluted 1:1 in buffer containing 50 mmol/L Tris-HCl (pH 7.4), 10 mmol/L KCl, and 5% glycerol. Then, 10 μL of the diluted serum was incubated with 10 μL of the caspase substrate for 3 hours at room temperature. Finally, the luminescence of the samples was measured in a luminometer. A statistical analysis comparing the concentration of the M30 antigen (U/L) or caspase activity (RLU) in the different patient groups at various times was performed using the two-tailed t test for equality of means. The predictive value of the assay was determined by receiver operating characteristics (ROC) analysis. The statistical analysis was confirmed by a professional statistician. A P value of less than .05 was considered significant. All assays were performed in duplicate. Intraassay variation was less than 4%, and interassay variation was less than 10%.
Patients Responding to Antiviral Therapy Showed Increased Caspase Activity in the Serum Compared to Patients With No Response or With a Viral Relapse.
To investigate hepatic caspase activation with respect to the outcome of antiviral therapy, we recruited HCV patients who were responding (n = 11) or nonresponding (n = 7) to IFN therapy or patients with a viral relapse (n = 5). Caspase activation was measured in serum samples using an ELISA that specifically recognizes a caspase-generated neoepitope of cytokeratin-18 (CK-18), which is abundantly expressed in hepatocytes but not in nonepithelial cells and lymphocytes.12 Patients not responding to antiviral therapy showed moderate caspase activity (171 ± 24 U/L) before therapy, which remained rather constant during the course of treatment and increased slightly after 24 weeks (Fig. 1A). Patients with a viral relapse had a caspase activity of 195 ± 65 U/L before therapy. Strikingly, however, responder patients showed strongly and significantly (P < .05) elevated levels of the caspase-generated CK-18 fragment, which was most evident before therapy (471 ± 95 U/L) and within the first 3 days of treatment. Thereafter, caspase activity declined, and no significant differences were obtained when the groups with a sustained and non-sustained response were compared after 14 and 24 weeks (P = .6 and P = .2, respectively).
Determination of the Cutoff Value of Serological Caspase Activity to Predict the HCV Treatment Outcome.
To determine the predictive discriminating value of caspase activity for a therapeutic response, we performed a ROC analysis comparing patients with a sustained and non-sustained response (Fig. 1B). Caspase activity above or below 254 U/L before treatment correctly predicted the clinical outcome with a sensitivity of 70% and a specificity of 82% (area under the curve 0.845; 95% CI). Thus, these data suggest that measurement of caspase activity might be a sensitive, noninvasive method of predicting treatment outcome.
Serological Detection of Caspase Activity Correlated With Virus Load in HCV Patients Responding to Antiviral Therapy.
In all responders, caspase activity correlated with the virus load during the course of antiviral therapy. A representative example of a responder patient is shown in Fig. 2A, which demonstrates that the extent of caspase activity almost coincided with virus elimination following IFN treatment. Caspase activity in this patient correlated (r = 0.936) with viremia (Fig. 2B). The correlation between caspase activity and virus load reached a mean r of 0.81 ± 0.037 in all responding patients tested. In contrast, nonresponders generally showed lower levels of caspase activity before treatment (Fig. 2D).
To verify the data obtained with the ELISA, we in addition established a novel luminometric enzyme assay for the measurement of caspase activity.22 To this end, serum samples were incubated with the caspase substrate DEVD-aminoluciferin, which, when cleaved by caspase-3 and caspase-7, generates a luminescent signal that can be measured in a luciferase reaction. As shown for a representative example of a responder patient (Fig. 2C), both the immunological and the enzymatic assay gave very similar results, underlining the reliability of the caspase measurements.
The present study has demonstrated for the first time that HCV patients who do not clear viral infection during IFN/ribavirin therapy or show viral relapse after antiviral treatment have significantly lower levels of caspase activity than do patients who resolve their HCV infection. Very intriguing is the finding that caspase activity in the responder patients was already elevated before the antiviral treatment. To assess caspase activation in serum samples, two independent methods were employed. In the first assay, we measured the concentration of a caspase-generated neoepitope of CK-18. We have recently shown that this assay is largely liver specific in HCV-infected patients, because CK-18 expression is restricted to epithelial cells and is abundant in hepatocytes, whereas nonepithelial cells do not express the antigen.12, 22 The second assay directly assessed the proteolytic activity of caspases by a luminometric technique.12, 22 That both assays showed a similar extent of serum caspase activation underlines the reliability of the measurements. Our data therefore indicate that caspase activity might be a powerful marker for predicting responses to HCV antiviral therapy.
We also demonstrated that caspase activity is associated with virus clearance, although our data certainly do not determine whether caspase activation is the cause or a consequence of viral clearance. Nevertheless, these results imply a role for caspases in the antiviral immune response and HCV clearance. It is quite established that the development of a vigorous and multispecific T-cell response is required for HCV clearance.23, 24 However, although an HCV-specific T-cell response in peripheral blood has been associated with viral clearance, recent data demonstrate that an intrahepatic T-cell response in particular is required for the efficient resolution of infection, at least in chimpanzee models.25 An important effector mechanism by which cytotoxic T lymphocytes (CTLs) and NK cells could control virus infection is the induction of apoptosis by the receptor- or perforin/granzyme B-mediated pathways. Both these processes are involved in the control of infection by several viruses.26, 27 Interestingly, FADD, a component of the death receptor signaling pathway, has recently been implicated in a novel innate immune mechanism for the control of virus replication,28 suggesting that death receptor systems might be integral components of the antiviral response.
For efficient viral elimination, certainly various other factors, in particular IFNs, are required. IFNs activate PKR- and STAT-dependent pathways, both of which are involved in the inducible expression of several proapoptotic factors.17–19 These include death receptors and death ligands, FADD as well as caspase-8, which sensitize virally infected hepatocytes to CTL-mediated caspase activation and apoptosis. In contrast, HCV can exert several mechanisms to evade the antiviral effects of IFN, which might contribute to the ineffective immune response observed in many HCV patients.29
In addition to impaired IFN signaling and altered cytokine profiles, several viral and host-mediated factors have been implicated in the reduced response to IFN-α therapy.30, 31 These include age, sex, stage of fibrosis, and obesity, which were not significantly different between our patient groups. Whether caspases are directly involved in viral clearance is unknown. A more recent report that is consistent with our data demonstrated that nonresponder HCV patients might have elevated levels of antiapoptotic Bcl-2.32 This together with the present findings could suggest that reduced expression of antiapoptotic molecules and the increased activation of apoptotic caspases would result in not only liver cell apoptosis but also possibly in clearance of HCV. It has been also shown that caspases and other proapoptotic mediators exert their antiviral role by controlling replication of several viruses.33, 34 Moreover, the role of caspases might be not restricted to apoptotic processes.9 Caspases can cleave different cytokine precursors of active cytokines and thus can create an environment that could be essential for HCV clearance. An important cytokine activated by caspases is IL-18 (IFN-γ-inducing factor), which plays a crucial role in the antiviral response and has also been associated with the therapeutic outcome of anti-HCV therapy.35, 36 Evidence is emerging that the intrahepatic production of IFN-γ, which can be induced by caspase-mediated generation of IL-18, is more important for viral resolution than is the response to type I IFNs.25 This idea is supported by the evidence that HCV rapidly induces but is not controlled by IFN-α/β, whereas viral clearance has greater correlation with the production of IFN-γ by intrahepatic HCV-specific T cells. It therefore can be assumed that IFN-γ will also modulate apoptotic pathways in an HCV-infected liver. An interesting candidate for this is TRAIL, which is inducible by viral infection and IFNs.37, 38 Indeed, we observed that serum levels of TRAIL also were more strongly induced in responder than in nonresponder and relapse patients in the course of IFN therapy (data not shown). As the TRAIL receptor/ligand system seems to have evolved as a prime physiological mechanism of elimination of virus-infected cells, TRAIL might be involved in triggering caspase activation during HCV infection.
The elevated caspase activation measured in sera of responding patients presumably reflects an efficient host immune response. Both nonapoptotic and apoptotic functions of caspases might be relevant to this process. Whether this response is HCV specific or caused by general activation of CTLs or NK cells remains to be investigated. Nevertheless, the increased caspase activation, as determined by the CK-18 cleavage, presumably involves an intrahepatic rather than a peripheral immune response. So far, intrahepatic T-cell responses are only poorly understood in HCV infection, as their analysis requires the study of liver biopsies. Thus, although we are aware that prospective studies with larger patient cohorts are necessary, we suggest that the serum measurement of caspase activity could serve as a valuable and noninvasive predictive marker for the response to antiviral therapy. Not only could this prevent enormous costs, but it also could spare nonresponding patients from the significant side effects of IFN therapy.