Major histocompatibility complex class I–related chains A and B (MIC A/B): A novel role in nonalcoholic steatohepatitis^†

Patients.

All enrolled patients were physically examined and a complete set of laboratory parameters was obtained. Individuals aged <18 years and >65 years with different liver pathologies (infection with hepatitis B virus, hepatitis C virus, or human immunodeficiency virus), history of organ transplantation, history of malignancy within the previous 5 years, alcohol or drug abuse within the previous year, autoimmunity, genetic disorders, and therapy with immunosuppressive or cytotoxic agents were excluded. Indication for bariatric surgery was made according to National Institutes of Health guidelines (body mass index ≥40 kg/m², plus comorbidities). Ultrasonographic examination of the liver was performed and biopsies were harvested from all 40 morbidly obese patients undergoing bariatric surgery. The surgeon's choice (for example, adjustable gastric band, Roux-Y, or gastric bypass surgery) was based on the current guidelines adjusted to the patient's clinical status and comorbidities as well as on the clinical expertise. All patients were informed about the additional risks of a wedge liver biopsy during the bariatric procedure. Blood and liver samples were obtained from consented liver transplantation donators used as healthy controls. In addition, to further explore the role of MIC A/B, we also investigated a cohort of 10 patients with NAFL (that is, as a benign form of NAFLD with only simple steatosis). Specimens were split, with one piece stored in 4% formalin solution (Roth, Karlsruhe, Germany) for subsequent histological examination and the other piece stored in RNA-preserving agent (RNAlater; Ambion Applied Biosystems, Darmstadt, Germany) to determine the expression of selected genes.

The study protocol conformed to the ethical guidelines of the 1975 Declaration of Helsinki and was approved by the Institutional Review Board of the University Hospital of Essen. All patients provided written informed consent before enrollment. The patients' baseline characteristics are given in Table 1.

Histological Analysis.

Hematoxylin-eosin staining was performed according to standard techniques. Samples were investigated and quantified according to NAFLD activity score (NAS).17 Steatosis (0–3), hepatocellular ballooning (0–2), and lobular inflammation (0–2) were quantified, respectively. NAS of ≥5 or ≥4 with at least one score for ballooning was defined as NASH. Extent of liver fibrosis was assessed using the modified METAVIR criteria.18

Quantitative Real-Time Polymerase Chain Reaction.

Liver tissue was homogenized with a blade homogenizer (IKA, Staufen, Germany) according to standard laboratory procedures. Total RNA was isolated with the RNeasy mini kit (Qiagen, Hilden, Germany) following the protocol using spin technology. Having spectrophotometrically assured the samples' purities and adjusted their concentrations, 2 μg of each RNA sample was filled up to a total volume of 100 μL with RNAse-free water. Reverse transcription was performed with the Quanti Tect RT kit (Qiagen) according to the manufacturer's instructions. Quantitative real-time polymerase chain reaction (qrt-PCR) of complementary DNA was performed using the iCycler iQ thermal cycler (Bio-Rad, Hercules, CA) with real-time detection system software 3.0a and Genex software (Bio-Rad) in 30 μL reactions containing 15 μL Quanti Tect Sybr Green master mix (Qiagen), 5 μL complementary DNA, 1 μL forward primer, 1 μL reverse primer (at 10 pmol/μL each), and 8 μL aqua dest. Amplification was performed for 15 minutes at 95°C, followed by 40 cycles of 30 seconds at 95°C, 30 seconds at 55°C, and 30 seconds at 72°C. Melting curve data were collected from 95°C to 55°C, at −0.5°C steps for 10 seconds each. Relative gene expressions were calculated from the threshold cycles in relation to housekeeping gene, to untreated controls or healthy donors, respectively.

Hypoxanthine-phosphoribosyltransferase 1 primers were used to normalize the samples. PCR primers (all obtained from Eurofins MWG Operon, Ebersberg, Germany) were as follows: hypoxanthine-phosphoribosyltransferase 1, 5′-GAC-CAG-TCA-ACA-GGG-GAC-AT-3′ (forward) and 5′-CTT-GCG-ACC-TTG-ACC-ATC-TT-3′ (reverse); MIC A, 5′-GTA-TTG-GGA-CCG-GAA-CAC-AC-3′ (forward) and 5′-ATG-CTC-TGG-AGG-GTG-TGA-GA-3′ (reverse); MIC B, 5′-TGC-CAT-GAA-GAC-CAA-GAC-AC-3′ (forward) and 5′-GGG-GCA-CTG-TTC-TCC-TGA-T-3′ (reverse); NKG2D, 5′-TTC-AGA-TAT-CCC-CAA-GGC-TG-3′ (forward) and 5′-TGA-TCT-GCT-GGC-CTT-CTC-TT-3′ (reverse); tumor necrosis factor–related apoptosis-inducing ligand (TRAIL)–death receptor 5 (DR5), 5′-CAC-TGG-AAT-GAC-CTC-CTT-TTC-3′ (forward) and 5′-CTT-CCG-GCA-CAT-CTC-AGG-3′ (reverse); CD95/Fas, 5′-CAA-AGC-CCA-TTT-TTC-TTC-CA-3′ (forward) and 5′-TTT-GGT-TTA-CAT-CTG-CAC-TTG-G-3′ (reverse); collagen 1α (I), 5′-AAC-AGC-CGC-TTC-ACC-TAC-AG-3′ (forward) and 5′-GGA-GGT-CTT-GGT-GGT-TTG-GT-3′ (reverse); and α-small muscle actin, 5′-TTC-GTT-ACT-ACT-GCT-GAG-CGT-GAG-A-3′ (forward) and 5′-AAG-GAT-GGC-TGG-AAC-AGG-GTC-3′ (reverse).

Enzyme-Linked Immunosorbent Assays.

CD95/Fas and Fas ligand concentrations were determined by sandwich enzyme-linked immunosorbent assay (ELISA) methods. Plates precoated with human Fas/TNF RSF6 and human Fas ligand/TNF SF6 monoclonal antibodies (Quantikine ELISA kit, R&D Systems, Wiesbaden, Germany) were blocked by adding 15% bovine serum albumin, washed, and incubated with the standard or patients' sera. Patients' sera were diluted 1:3 in 7.5% bovine serum albumin. Absorbance was measured at 450 nm.

Apoptosis.

Cell death markers M30 (for apoptosis) and M65 (for overall cell death) were assessed both in the sera of patients and healthy controls using the M30 (Apoptosense) and M65 ELISA kit (both from Peviva, Bromma, Sweden) following the manufacturer's instructions. Whereas M30 is a cytokeratin-18 neo-epitope only exposed upon apoptotic cleavage by activated caspase-3,19 M65 reflects total cleaved and uncleaved cytokeratin-18.

MIC A/B are induced upon cellular distress conditions such as DNA damage, malignant transformation, or intracellular infection.20-23 Therefore, sections were counterstained with 4′,6-diamidino-2-phenylindole–containing ProLong antifade reagent (Invitrogen, Karlsruhe, Germany), and apoptotic hepatocytes were quantitated by way of terminal deoxynucleotidyl transferase-mediated dUTP nick-end labeling (TUNEL) assay, which enzymatically labels free 3′ OH ends of damaged DNA with a fluorescently labeled nucleotide as described.24 Cells displaying TUNEL-labeled fluorescent nuclei were quantified by counting the number of positive cells per high-power field. A total of 10 high-power fields were analyzed for each patient with excitation and emission wavelengths of 380 and 430 nm, respectively, using an inverted laser scanning confocal microscope (LSM 510, Carl Zeiss Micro-Imaging, Göttingen, Germany) equipped with a ×40 NA 1.4 lens and LSM 510 imaging software. Data are expressed as the number of TUNEL-positive cells per 10 high-power fields.

Immunohistochemistry for Hepatic NK Cells and MIC A/B.

Tissue sections were preincubated with Block-ace (Dako Cytomotion Inc., Carpinteria, CA) for 10 minutes at 37°C to block nonspecific binding of the primary antibody. Endogenous peroxidase activity was blocked with 0.3% H₂O₂and 0.1% NaN₃in distilled water for 10 minutes at room temperature. Sections were then incubated overnight with 1:500 diluted biotin-conjugated anti-CD57 primary antibody (Invitrogen, Karlsruhe, Germany) for hepatic NK cells and with MIC A/B–specific monoclonal antibodies (AbD Serotec, Münster, Germany), rinsed three times with phosphate-buffered saline, and incubated with avidin-biotin peroxidase complexes (Vector Laboratories, Burlingame, CA). Histochemical development was achieved by employing the 3,3′-diaminobenzidine substrate kit (Vector Laboratories). Finally, sections were counterstained for 3 minutes with hematoxylin and coverslipped with mounting medium for light microscopy.

Sirius Red Staining for Liver Fibrosis.

We proceeded according to Arteel et al.25 Direct red 80 and Fast green FCF (color index 42053) were obtained from Sigma-Aldrich Diagnostics (Taufkirchen, Germany). Red-stained collagen fibers were quantified in liver sections by digital image analysis as described.26

Statistical Analysis.

All data represent the measurements of three separate experiments and are expressed as the mean ± standard error unless otherwise indicated. Statistical analyses were performed using SAS version 8.2 software (SAS Institute, Cary, NC). Qualitative characteristics were analyzed using the χ² test or Fisher's exact test as appropriate. Correlation coefficients reported are rank (Spearman) correlations. Statistical significance was assumed at P < 0.05. This explorative analysis performed no adjustment for multiple testing. Descriptive statistics were computed for all variables and included means and standard deviations or medians. Differences between concentrations were evaluated statistically by one-way analysis of variance, repeated-measures analysis of variance, or paired Student t test.

Hepatic NK Cells and NKG2D, MIC A/B Transcripts Are Increased in NASH.

We initially sought to investigate the role of hepatic NK cells in patients with NASH following bariatric surgery through immunohistochemical analysis. As depicted in Fig. 1A, an increase of this cell type was observed in patients with NASH, whereas NK cell numbers were significantly lower in patients with NAFL (13.7 ± 0.9 versus 5.5 ± 0.8 NK cells/10 hpf, p = 0.001). In contrast, almost no NK cells were observed in controls (Fig. 1A). Likewise, qrt-PCR also demonstrated a significant 13.1-fold increase in NKG2D messenger RNA (mRNA) expression as compared with healthy controls (P < 0.01, Fig. 1B). Interestingly, NAFL patients revealed lower NKG2D mRNA transcription levels than patients with NASH (9.0 ± 5.1 for NAFL vs. 13.1 ± 1.8 for NASH, p = 0.27, n.s. = not significant). In addition, MIC A/B mRNAs were increased in the livers of patients with NASH versus control livers by 3.6- and 15.8-fold (P < 0.001, Fig. 1C). We observed a significant reduction of MIC B levels in individuals with NAFL as compared to patients with NASH whereas transcripts for MIC A were almost identical in both groups. Next, the tissue distribution of MIC A/B was examined by immunohistochemistry. The diffuse cellular staining pattern of MIC A/B in patients with NASH in Fig. 1C is no artifact but rather typical of this kind of stain.27 In contrast, control hepatocytes were negative while individuals with NAFL weakly expressed MIC A/B.

TRAIL–DR5 and CD95/Fas Transcripts Are Increased in NASH.

Hepatic NK cells play a critical role in TRAIL- and Fas-mediated liver injury. The liver harbors many NK cells,28 and approximately 30% to 40% of these constitutively express TRAIL.29 We thus questioned whether the expression of TRAIL–DR5 and CD95/Fas mRNAs is increased in NASH livers. Indeed, a 2.7-fold increase in TRAIL–DR5 and a 3.6-fold increase in CD95/Fas mRNA were observed in patients with NASH compared with controls (Fig. 2A). TRAIL–DR5 mRNA was significantly less increased in individuals with NAFL compared with patients with NASH (1.1 ± 0.1 versus 2.7 ± 0.2, P = 0.01). In contrast, CD95/Fas transcripts were also reduced in the NAFL group, but this difference was not significant when compared with patients with NASH (2.5 ± 0.5 versus 3.6 ± 0.9; P = 0.16). Up-regulation of CD95/Fas in patients with NASH was further confirmed through ELISA and demonstrated a significant increase as compared with both healthy controls and individuals with NAFL (P < 0.05). However, up-regulation of CD95/Fas was not accompanied by enhancement of the Fas ligand (Fig. 2B). Subsequently, histopathological examination was performed along with determination of serum alanine aminotransferase and aspartate aminotransferase values. The NAS is a well-established scoring system for patients with NASH as a progressive form of NAFLD.17 As expected, histopathological examinations from patients with NASH displayed an increase in NAS as compared with controls. In addition, a marked elevation in serum alanine aminotransferase and aspartate aminotransferase values was also observed in patients with NASH versus controls (Fig. 2C). Liver enzymes from individuals with NAFL were within the reference ranges.

Hepatocyte Apoptosis Is Up-regulated in NASH.

In order to examine potential effects of MIC A/B on hepatocyte death rates in NASH, we quantified the intrahepatic rates of apoptotic (by M30 and confirmatory TUNEL assays) and overall cell death (by M65 assay). As expected, M30 expression was significantly elevated in patients with NASH (440.5 ± 71.1 U/L) as compared with controls (94.4 ± 30.9 U/L; P < 0.05) (Fig. 3A), which went along with a likewise significantly increased M65 expression in patients with NASH versus controls (784.9 ± 129.9 U/L versus 230.3 ± 43.5 U/L; P < 0.05). In contrast, M30 levels were significantly reduced in individuals with NAFL (139.5 ± 20.9 U/L), which was again reflected in the significantly lower M65 expression in NAFL (298.2 ± 35.4 U/L). To further investigate whether an increase in MIC proteins is associated with enhanced liver injury and fibrosis, we determined correlation coefficients both for patients with NAFL and NASH. As shown in Fig. 3B, increased MIC A mRNA expression in patients with NAFL revealed a relevant correlation with NAS (r² = 0.87), whereas the correlation coefficient for the stage of fibrosis was less evident (r² = 0.68) (Fig. 3B). Higher correlation coefficients were found for MIC B transcripts in relation to NAS (r² = 0.93) and fibrosis stage (r² = 0.79). In patients with NASH, significant positive correlations were found for MIC A mRNA and NAS (r² = 0.89) as well as the stage of fibrosis (r² = 0.83) (Fig. 3C). Moreover, the correlation coefficients of MIC B transcripts to the severity of disease in NASH were also significant for both NAS (r² = 0.72) and fibrosis stage (r² = 0.85). In the final common step of the apoptotic machinery, activated effector caspases—in particular caspase-3 and caspase-7—cleave cytokeratin-18 as the major hepatic intermediate filament protein. Cytokeratin-18 fragment levels independently predict NASH. A cutoff value of 395 U/L holds 99.9% specifity and 85.7% sensitivity for the diagnosis of NASH.30 Regarding the role of MIC A/B in NASH severity, we further stratified the levels of these stress-induced proteins versus apoptosis-indicating M30 levels. Figure 3D shows a significant correlation within the NASH cohort. Confirmatory TUNEL assay demonstrated numerous clusters of apoptotic cells on an altered hepatic micro-architecture in obese patients with NASH (Fig. 3E). In contrast, only few apoptotic cells were observed in control livers or livers affected by fatty infiltration in patients with NAFL. Consistent with the previous results, the number of TUNEL-positive hepatocytes was also significantly higher in NASH patients than in controls and in individuals with NAFL. Quantitation of TUNEL-positive hepatocytes demonstrated a four-fold increase in patients with NASH compared with healthy controls (8.8 ± 0.8 versus 2.2 ± 0.2 TUNEL-positive cells per 10 high-power fields; P = 0.005) (Fig. 3E). Because hepatocyte apoptosis can induce liver inflammation and lead to fibrosis,31 we next investigated markers of liver fibrosis during MIC A/B–mediated liver injury.

Markers of Hepatic Fibrogenesis Are Enhanced in NASH.

Hepatic stellate cells are the principal cell type responsible for collagen deposition in the liver.32 We thus quantified transcripts indicating hepatic stellate cell activation by qrt-PCR. As expected, mRNA for α-smooth muscle actin, a cardinal marker for hepatic stellate cell activation, was increased 2.6-fold in patients with NASH as compared with controls (Fig. 4A). Conversely, α-small muscle actin transcripts were not up-regulated in individuals with NAFL. To ascertain whether HCS activation was associated with enhanced hepatic fibrogenesis, mRNA for hepatic collagen 1α (I) was quantified next. Collagen 1α (I) mRNA expression was increased 4.4-fold in patients with NASH versus controls (Fig. 4B). However, collagen 1α (I) mRNA was marginally up-regulated in NAFL. Hepatic collagen protein deposition was identified in liver specimens by Sirius red staining (Fig. 4C) and subjected to quantitative morphometry.26 Collagen staining by Sirius red was increased 2.7-fold in patients with NASH versus controls (Fig. 4C). Moreover, collagen deposition in livers of individuals with NAFL was significantly lower (5.7 ± 1.1%; P < 0.05) (Fig. 4C).