In vivo assessment of liver cell apoptosis as a novel biomarker of disease severity in nonalcoholic fatty liver disease


  • Potential conflict of interest: Nothing to report.


In patients with nonalcoholic fatty liver disease (NAFLD), a liver biopsy remains the only reliable way to differentiate simple steatosis from nonalcoholic steatohepatitis (NASH). Noninvasive methods are urgently needed. Increasing evidence suggests hepatocyte apoptosis is a key mediator of liver injury in NAFLD. The aim of this study was to quantify hepatocyte apoptosis in plasma from patients with NAFLD and correlate it with histological severity. Plasma was obtained from 44 consecutive patients with suspected NAFLD at the time of liver biopsy. Histology was assessed blindly. Caspase-3–generated cytokeratin-18 fragments were measured in situ via immunohistochemistry and in vivo using a novel enzyme-linked immunosorbent assay. Plasma cytokeratin-18 fragments were markedly increased in patients with NASH compared with patients with simple steatosis or normal biopsies (median [interquartile range]: 765.7 U/L [479.6–991.1], 202.4 U/L [160.4–258.2], 215.5 U/L [150.2–296.2], respectively; P < .001). Cytokeratin-18 fragment levels independently predicted NASH (OR 1.95; 95% CI 1.18–3.22; P = .009 for every 50 U/L increase). A cutoff value of 395 U/L calculated using the receiver operating characteristic curve approach showed a specificity of 99.9%, a sensitivity of 85.7%, and positive and negative predictive values of 99.9% and 85.7%, respectively, for the diagnosis of NASH. In conclusion, these findings strongly suggest that determination of hepatocyte caspase activation in the blood is a strong and independent predictor of NASH in human subjects. These data highlight the potential usefulness of this test as a noninvasive diagnostic means of determining histological disease severity in patients with NAFLD. (HEPATOLOGY 2006;44:27–33.)

Nonalcoholic fatty liver disease (NAFLD) is the most common form of chronic liver disease in both children and adults.1, 2 It encompasses a wide spectrum of conditions associated with overaccumulation of fat in the liver ranging from simple steatosis to nonalcoholic steatohepatitis (NASH) and cirrhosis.1 Simple steatosis is the most common form of NAFLD and typically follows a benign nonprogressive clinical course. In contrast, NASH is a potentially serious condition, since as many as 25% of these patients may progress to cirrhosis and experience complications of portal hypertension, liver failure, and hepatocellular carcinoma.3, 4 Emerging data suggest that hepatocyte apoptosis, a specific form of cell death, may play an important role in liver injury and disease progression in NAFLD. We initially demonstrated that caspase activation and liver cell apoptosis are prominent pathological features of human NAFLD.5 Moreover, the degree of apoptosis correlated with the severity of steatohepatitis and the stage of fibrosis. Several experimental models of NAFLD as well as subsequent human studies support these initial observations.6–8

The apoptotic pathway is composed of two arms: the intrinsic pathway (initiated by cellular stress) and the extrinsic pathway (stimulated through a death receptor–mediated process). Both pathways are suspected to be involved in the pathogenesis of NASH.9 In the final common step of apoptosis, the effector caspases (in particular caspase-3 and caspase-7) are activated. These specific intracellular proteases are known to cleave several cellular substrates, including cytokeratin-18 (CK-18), the major intermediate filament protein in the liver. Antibodies against caspase-generated CK-18 fragments have been shown to specifically label early apoptotic cells.10 A recent study reported that these caspase-generated CK-18 fragments could be detected in the blood of patients with chronic hepatitis C infection using a novel enzyme-linked immunosorbent assay kit.11

Liver biopsy, an invasive procedure associated with various complications, is currently the only reliable method of differentiating simple steatosis from NASH and evaluating histological severity of the disease. In light of the increasing prevalence of NAFLD in our population, there is great need to develop accurate and noninvasive techniques to diagnose NASH and assess histological severity. The objective of the present study was to determine the clinical use of quantifying hepatocyte apoptosis in human NAFLD through a noninvasive blood test by comparing the results with those of traditional histological evaluation.


NAFLD, nonalcoholic fatty liver disease; NASH, nonalcoholic steatohepatitis; CK-18, cytokeratin-18; AST, aspartate aminotransferase; ALT, alanine aminotransferase.

Patients and Methods

Patient Characteristics.

The study was approved by the Cleveland Clinic Institutional Review Board, and all patients gave written informed consent prior to participation. Our cohort consisted of 44 consecutive patients undergoing liver biopsy for clinical suspicion of NAFLD by their treating hepatologist. All of the patients had persistently elevated liver enzymes for more than 3 months and evidence of clinical components of the metabolic syndrome (any of the following: obesity, diabetes, hypertension, hyperlipidemia) in the absence of other causes of elevated aminotransferases. Patients were excluded if alcohol consumption was >30 g/d for males and >20 g/d for females and if other liver diseases were detected via serological testing and imaging studies. Blood was obtained from each patient at the time of liver biopsy. Demographic, clinical, and laboratory data were collected. Two patients were excluded because of an alternative diagnosis on liver histology (necrosis and hepatocyte dropout of unknown etiology and primary biliary cirrhosis); one patient was excluded because of a hemolysed blood sample. Patients were subsequently divided into four groups according to their histological findings (see Liver Histology): normal biopsy, “not NASH” (simple fatty liver), “borderline NASH,” and “definitive NASH.” Subjects with “borderline NASH” (n = 2) were not included in the analysis, because according to Kleiner et al.,12 these patients cannot be classified into a clear-cut category. A total of 39 patients with clinically suspected NAFLD were used to perform the final statistical analysis. In addition, 35 healthy age-matched controls from blood bank donors without clinical signs or symptoms of liver disease, and no history of chronic illnesses, were analyzed.

Liver Histology.

The histological diagnosis was established using hematoxylin-eosin and Masson trichrome stains of formalin-fixed paraffin-embedded liver and graded in a blinded fashion by an experienced hepatopathologist (L. Y.) according to the NAFLD scoring system recently proposed by the National Institute of Diabetes and Digestive and Kidney Diseases NASH Clinical Research Network.12 According to this scoring system, the degree of steatosis and inflammatory activity is measured using a 9-point scale (steatosis = 0–3; lobular inflammation = 0–3; ballooning = 0–2). The NAFLD activity score is the unweighted sum of steatosis, lobular inflammation, and hepatocellular ballooning scores. In accordance with the report of the NASH Clinical Research Network, a NAFLD activity score of ≥5 corresponded to a diagnosis of “definitive NASH,” a score of 3–4 corresponded to “borderline NASH,” and a score of <3 corresponded to “not NASH.” The stage of fibrosis was similarly measured using a 6-point scale (1a, b = mild (1a)/moderate (1b) zone 3 perisinusoidal fibrosis; 1c = portal fibrosis only; 2 = zone 3 and portal/periportal fibrosis; 3 = bridging fibrosis; 4 = cirrhosis).


Paraffin-embedded liver tissue was cut, deparaffinized, and hydrated as previously described.5 Immunohistochemistry was performed with a mouse monoclonal antihuman M30 antibody (Roche Diagnostics) for detection of the caspase cleavage product of CK-18. The samples were incubated with the primary antibody for 1 hour at room temperature diluted 1:50 in the blocking solution. After washing with phosphate-buffered saline, the sections were incubated with the ready-to-use secondary antibody (DAKO Corporation, Carpinteria, CA) for 30 minutes at room temperature. After washing in phosphate-buffered saline, the samples were incubated with a ready-to-use streptavidin-peroxidase conjugate in phosphate-buffered saline containing carrier protein and antimicrobial agents (DAKO Corporation, Carpinteria, CA) for 30 minutes at room temperature. After washing with phosphate-buffered saline, the samples were stained with 3,3′-diaminobenzidine (Vector, Burlingame, CA) for 2 to 5 minutes, washed in phosphate-buffered saline, counterstained with hematoxylin for 2 to 3 minutes, and dehydrated by transferring them through increasing ethanol solutions (30%, 50%, 70%, 80%, 95%, 100% ethanol). Following dehydration, the slices were soaked twice in a xylene bath at room temperature for 5 minutes, mounted, and examined.

Measurement of Caspase-Generated CK-18 Fragments in the Blood.

A blood sample was obtained from each patient at the time of liver biopsy, processed to plasma and stored frozen at −80°C. The plasma was subsequently used for quantitative measurement of the apoptosis-associated neo-epitope in the C-terminal domain of CK-18 using the M30-Apoptosense ELISA kit (PEVIVA; Alexis, Grünwald, Germany). All assays were performed in triplicate, and the absorbance was determined using a microplate reader (Molecular Devices M2, Sunnyvale, CA).

Statistical Analysis.

Descriptive statistics were computed for all variables. These include means and standard deviations or medians, as well as 25th and 75th percentiles for continuous factors. For categorical variables, frequencies and percentages were estimated. Kruskal-Wallis and Dunn's tests were used to assess whether there were any significant differences in terms of continuous clinical or serological characteristics between any of the three subject groups. Chi-square or Fisher's exact tests were used for categorical factors. Spearman's correlation coefficient was used to estimate the association of plasma CK-18 levels with age, body mass index, aspartate aminotransferase (AST), alanine aminotransferase (ALT), AST/ALT ratio, and stage of fibrosis. A logistic regression analysis was used to assess the association between plasma levels of CK-18 fragments and the likelihood of having definitive NASH. Variables that were found to be associated with CK-18 levels in the univariable analysis or those known to be associated with NASH severity (AST/ALT ratio, diabetes, hyperlipidemia, body mass index) were studied. To predict the presence of NASH with optimal sensitivity and specificity, receiver operating characteristic curve analysis was used to estimate potential cutoff values of plasma CK-18 fragments. A P value of .05 was considered statistically significant. SAS version 9.1 software (SAS Institute, Cary, NC) and R 2.0.1 software (The R Foundation for Statistical Computing) were used to perform all analyses.


Patient Characteristics.

The main clinical and serological features of the patients are described in Table 1. Patient age (50.8 ± 11.1 years) and sex (53.9% females) did not differ significantly between the groups (P = .76), whereas body mass index was significantly higher in patients with “definitive NASH” compared with patients with simple steatosis and normal liver biopsies (P < .001). AST/ALT ratio was not significantly different between the groups (P > .10). Thirty-one percent of the patients had clinical diabetes, 46.2% had hypertension, and 46.2% had hyperlipidemia (P > .15). Thirty-eight percent of patients with a diagnosis of NAFLD had positive autoantibodies (antinuclear, anti–smooth muscle, and/or antimitochondrial). Table 2 lists the histological characteristics of the patients: 8 patients (20.5%) had a NAFLD activity score compatible with “not NASH” (simple steatosis), and 21 patients (53.8%) had “definitive NASH.” Ten patients (25.6%) with a clinical suspicion of NAFLD had a normal liver biopsy. None of the patients in the normal biopsy or “not NASH” groups had fibrosis. Of the 39 subjects, only 2 (5.1%) had cirrhosis, both of whom also had “definitive NASH.”

Table 1. Clinical and Serological Characteristics of the Patients
FactorAll Patients (N = 39)Normal Biopsy (n = 10)Not NASH (Simple Steatosis) (n = 8)NASH (n = 21)P Value*
  • Abbreviations: BMI, body mass index; Q25, 25th percentile; Q75, 75th percentile; ANA, antinuclear antibody; SMA, smooth muscle antibody; AMA, antimitochondrial antibody.

  • *

    P values correspond to the comparison of the three subject groups. Kruskal-Wallis tests for continuous factors or Pearson's chi-square for categorical variables were used.

  • Positive imaging of fatty liver included ultrasound, CT, or MRI.

Age (yr), mean (SD)50.8 (11.1)48.9 (7.9)46.8 (17.8)53.1 (9.0)NS
BMI (kg/m2), mean (SD)31.5 (4.0)28.6 (2.7)31.1 (4.8)33.1 (3.6).008
AST (U/L), median (Q25, Q75)58.0 (46.0, 76.0)59.0 (43.0, 66.0)50.0 (35.0, 79.5)61.0 (46.0, 76.0)NS
ALT (U/L), median (Q25, Q75)73.0 (54.0, 104.0)78.5 (59.0, 118.0)71.0 (47.0, 154.0)82.0 (44.0, 101.0)NS
AST/ALT ratio, mean (SD)0.8 (0.4)0.7 (0.2)0.8 (0.4)0.9 (0.4)NS
Ferritin, median (Q25, Q75)176.0 (75.6, 397.0)75.0 (53.0, 96.0)397.0 (119.0, 434.4)249.2 (130.4, 485.0).01
Sex, % female53.9505057.1NS
Race, % Caucasian82.19062.585.7NS
Diabetes, %30.82037.533.3NS
Hypertension, %46.2305052.4NS
Hyperlipidemia, %46.2602547.6NS
Positive ANA, %21.633.337.510NS
Positive SMA, %26.511.142.927.8NS
Positive AMA, %7.414.314.30NS
Positive imaging for fatty liver, %62.22071.480NS
Table 2. Histological Characteristics of Patients
FactorAll Patients (N = 39)Normal Biopsy (n = 10)Not NASH (Simple Steatosis) (n = 8)NASH (n = 21)
  1. NOTE. All values are expressed as percentages. Score, grade, and activity score were determined according to Kleiner et al.12

Lobular inflammation    
Activity grade    

Hepatocyte Apoptosis Is Increased in the Liver of Patients With “Definitive NASH.”

Caspase-3 activation and hepatocyte apoptosis have been shown to be prominent features of different experimental models of NAFLD—as well as human NAFLD—and to correlate with disease severity.5, 13 To determine whether caspase-3–generated CK-18 fragments are also increased in the livers of NAFLD patients, immunohistochemistry using the M30 monoclonal antibody was performed (Fig. 1). The immunoreactivity product was readily identified in liver tissue from patients with “definitive NASH” but was rarely detected in patients with simple steatosis or normal liver biopsies (Fig. 1). These data are consistent with previous studies and confirm the presence of hepatocyte apoptosis in patients with NASH using a complementary and highly specific technique.

Figure 1.

Caspase-3–generated CK-18 fragments are readily detected in the livers of patients with “definitive NASH.” Immunohistochemistry for caspase-generated CK-18 fragments was performed in the liver tissue using a mouse monoclonal antihuman antibody (M30) that selectively recognizes the caspase cleavage–generated neo-epitope of CK-18. (Original magnification ×200.)

Caspase-3–Generated CK-18 Fragments Are Markedly Increased in the Blood of Patients With “Definitive NASH.”

Plasma levels of CK-18 fragments ranged from 105.5 to 2,306.4 U/L (median 516.7 U/L, interquartile range 246.4–804.1 U/L). These levels were significantly higher than those observed in the 35 healthy controls (median 234, 25th percentile 197, 75th percentile 289; P < .001)”. More importantly, CK-18 fragment levels were strikingly higher in patients with “definitive NASH” (NAFLD activity score ≥5) compared with those with “not NASH” (simple steatosis) or normal biopsies (median [25th, 75th percentile]: 766 U/L [480, 991], 202 U/L [160, 258], 215 U/L [150, 296], respectively; P < .001) (Fig. 2). Only 3 patients in the “definitive NASH” group had a CK-18 value that overlapped with the range of the values observed in the other two groups. CK-18 fragment levels showed a weak correlation with body mass index (r = 0.36; P = .024; 95% CI 0.05–0.67) and stage of fibrosis (r = 0.55; P < .001; 95% CI 0.28–0.82), but not with age (P = .85), serum AST/ALT ratio (P = .17), serum AST (P = .057), or serum ALT (P = .55). They did not differ significantly according to the presence or absence of history of diabetes, dyslipidemia, or hypertension (P > .30).

Figure 2.

CK-18 fragments are significantly increased in the blood of patients with “definitive NASH” compared with patients with “not NASH” (simple steatosis) as well as patients with clinical suspicion of NAFLD but normal biopsy (median [25th, 75th percentile]: 766 U/L [480, 991], 202 U/L [160, 258], 215 U/L [150, 296], respectively; P < .001). A scatter plot of the data is presented, with each diamond representing one subject and the solid lines representing the median values for each group. The dotted line shows the area of overlap between the three groups.

Caspase-3–Generated CK-18 Fragments as an Independent Predictor of NASH in Patients With Suspected NAFLD.

The risk of definitive NASH on liver biopsy increased with increasing plasma levels of caspase-3–generated CK-18 fragments (P = .019). For every 50 U/L increase in CK-18 levels, the likelihood of having “definitive NASH” increased 86% (OR 1.86; 95% CI 1.23–2.82). To ascertain whether plasma CK-18 fragment levels independently predicted the presence of “definitive NASH,” we used a multivariable logistic regression analysis. Variables associated with CK-18 fragment levels and those that are known to be associated with NASH severity (age, body mass index, AST/ALT ratio, diabetes, and hyperlipidemia) were studied as possible confounders of the association between NASH and plasma levels of CK-18 fragments. The adjusted OR of 1.95 and 95% CI (1.18–3.22) were similar to the unadjusted OR, confirming that elevated levels of CK-18 fragments served as an independent predictor of “definitive NASH” (P < .05). Because none of the subjects in the “not NASH” group had fibrosis, we were unable to adjust for this factor in the multivariable analysis. Using the area under the receiver operating characteristic curve approach, we next calculated potential cutoff values to separate patients with “definitive NASH” from those with “not NASH” (Fig. 3). The area under the curve was estimated to be 0.93 (95% CI 0.83–1.00) and was found to be significantly higher than 0.5 (chance assignment). Two different cutoff values were calculated (Table 3). The first value was selected to minimize the rate of false positive results. Using this approach, a cutoff value of 395 U/L accurately predicted “definitive NASH” with a specificity of 99.9%, a sensitivity of 85.7%, and positive and negative predictive values of 99.9% and 85.7%, respectively. The second value was calculated to minimize the false negative rate. In this case, a value of 380.2 U/L gave a specificity of 94.4%, a sensitivity of 90.5%, and positive and negative predictive values of 95% and 89.5%, respectively.

Figure 3.

Caspase-3–generated CK-18 fragments accurately predict “definitive NASH” in patients with suspected NAFLD. Approximate area under the curve = 0.93. A cutoff value of 395.0 U/L accurately predicted “definitive NASH” versus “not NASH” with a specificity of 99.9%, a sensitivity of 85.7%, and positive and negative predictive values of 99.9% and 85.7%, respectively. A value of 380.2 U/L gives a specificity of 94.4%, a sensitivity of 90.5%, and a positive and negative predictive value of 95.0% and 89.5%, respectively.

Table 3. Diagnostic Value of Plasma CK-18 Fragment Determination for Predicting NASH
Validity MeasuresCK-18 Fragment Level Cut-off Value
395 U/L380.2 U/L
  1. NOTE. All values are expressed as percentages.

Positive predictive value99.995.0
Negative predictive value85.789.5


NAFLD has been increasingly recognized as one of the most common forms of chronic liver disease, affecting up to one third of the United States adult population.14 Natural history data suggest that patients with simple fatty liver have a relatively benign clinical course,15 whereas those with NASH are at particular risk for disease progression and may develop cirrhosis and its complications, including hepatocellular carcinoma.3, 4, 16, 17 At present, the available noninvasive tests to assess disease severity of NAFLD include clinical signs and symptoms, routine laboratory and radiological imaging tests, and combinations of clinical and blood test results. Unfortunately, these tests are of limited use, and liver biopsy remains the only reliable means of diagnosing NASH and grading the severity of liver damage.18 There is, therefore, an urgent need to develop and validate noninvasive tests that accurately distinguish NASH from simple steatosis and determine the stage and grade of disease. Such tests would not only aid clinicians in the selection of patients for liver biopsy but would also allow for noninvasive assessment of disease progression and therapeutic response.

Emerging data suggest that hepatocyte apoptosis may be a key component of the “second hit” involved in the progression of simple steatosis to NASH. Terminal deoxynucleotidyl transferase-mediated dUTP nick end labeling assay and immunohistochemical detection of active caspase-3 in a well-characterized NAFLD patient population have shown that hepatocyte apoptosis is a prominent pathological feature of human NAFLD.5 Moreover, a positive correlation exists between hepatocyte apoptosis and degree of inflammation and stage of fibrosis. Human and experimental studies in various animal models of NAFLD and in vitro models of hepatocyte steatosis are now providing further evidence that fat accumulation in liver cells may lead to an increase in apoptotic cell death.19 Hepatocyte apoptosis has also been linked to liver fibrogenesis.20 For instance, in animal models of cholestasis, attenuation of hepatocyte apoptosis also reduces fibrogenesis.21 Engulfment of apoptotic bodies by hepatic stellate cells stimulates the fibrogenic activity of these cells and may be one mechanism through which hepatocyte apoptosis promotes fibrosis.22

In a recent study, Bantel et al.11 reported the first assessment of a blood biomarker for hepatocyte apoptosis. Many forms of apoptosis involve the activation of caspases, which are intracellular proteases that cleave aspartate residues. Indeed, the signature cleavage of proteins after aspartate moieties by caspases is unique to this family of caspases. Furthermore, the cleaved protein generates new epitopes for which antibodies can be developed. Bantel et al. took advantage of this information to measure caspase-specific cleavage products of CK-18, a protein that is abundant in hepatocytes, thereby providing specificity for the source of this cleaved protein in blood via ELISA. This group demonstrated that hepatitis C virus patients had elevated levels of blood caspase–generated CK-18 cleavage fragments compared with healthy controls. Caspase activity levels mirrored the degree of steatosis in these patients23 and correlated with fibrosis in a specific subset of subjects.11 Using this approach in the present study, we were able to demonstrate that in vivo quantification of hepatocyte apoptosis accurately predicts NASH. Plasma caspase–generated CK-18 cleavage fragments were strikingly increased in patients with “definitive NASH” compared with those with “not NASH” and patients with suspected NAFLD but normal liver biopsy. Moreover, caspase activity levels independently predicted the presence of NASH. Because of its high sensitivity, specificity, and positive and negative predictive value, this test has the potential to become an important instrument in clinical practice. Larger studies are needed to confirm these observations.

In summary, our findings suggest that noninvasive monitoring of hepatocyte apoptosis in the blood of patients with NAFLD is a reliable tool to differentiate NASH from “not NASH” in patients with suspected NAFLD. To allow this blood test to become an important instrument in daily clinical practice, we are in the process of planning a large multicenter prospective validation study.


The authors would like to thank Dr. Stanley Hazen for providing the blood samples from healthy controls and Teresa A. Markle and Michael P. Berk, General Clinical Research Center technologists, for their excellent work and dedication.