Progression of nonalcoholic fatty liver disease (NAFLD) and more specifically one of its serious forms, nonalcoholic steatohepatitis (NASH), is accompanied by elevated concentrations of proinflammatory cytokines, such as tumor necrosis factor alpha (TNFα), in the blood plasma and liver [1, 2]. Abnormally high TNFα promotes hepatic inflammation, lipid deposition and peroxidation, stimulates activation of Kupffer cells and hepatocyte apoptosis, and leads to insulin resistance [3]. As plasma TNFα levels go back to normal, the liver function recovers [4, 5].

Proteins belonging to the TNF family exert their biological effects by interacting with TNFR superfamily receptors [6]. TNFα-binding receptors (mbTNFR) are represented by two types of transmembrane proteins: mbTNFRI and mbTNFRII. The intracellular region of mbTNFRI carries a death domain absent in mbTNFRII. Once activated, the death domain triggers either apoptosis or necroptosis [7]. Another type of TNFα receptors are soluble sTNFRs, a product of mbTNFR ectodomain shedding mediated by ADAM metalloproteinases [8]. sTNFRs bind to TNFα and act as mbTNFR antagonists preventing activation of TNFα-signaling pathways. Low concentrations of soluble TNFα receptors can be found in the blood serum and urine of healthy individuals. Patients with chronic viral hepatitis [9], cirrhosis [10], or NAFLD [11, 12] have elevated levels of circulating TNFR, which indicates inflammation and activation of T-cell immunity, in particular CD8+ T-cells that express metalloproteinase ADAM- 17 [8]. It is hypothesized that both levels and ratio of soluble to membrane-bound TNFα receptors play a significant role not only in inducing hepatocyte death and damage to the liver, but also in the regeneration and homeostasis of this organ [13, 14, 15, 16]. It appears that the ratio of soluble to membrane-bound TNFRI and TNFRII largely determines the intensity of immune response and inflammatory reactions. It has been shown that mutations in the TNFRSF1A and TNFRSF1B genes affect sTNFRI and sTNFRII concentrations in the blood plasma and the number of mbTNFRI and mbTNFRII proteins on the surface of innate immunity cells [17]. Therefore, we can hypothesize that polymorphisms of genes coding for TNFα receptors may substantially contribute to the etiology and pathogenesis of liver diseases, including NAFLD. At present, attempts are made to establish an association between polymorphic variants of TNFR-encoding genes and NAFLD. The data is still scarce, describing mostly a link between TNFRSF1A or TNFRSF1B polymorphisms and biliary cirrhosis, alcoholic liver disease and hepatocellular carcinoma [18, 19, 20]. Associations between polymorphisms of TNFα receptor genes and NAFLD are hardly reported in the literature. That said, we decided to investigate how TNFRSF1A and TNFRSF1B polymorphisms contribute to the development of NAFDL in Karelian residents.

METHODS

Venous blood sample collection was aided by the Department of Propedeutics of Internal Diseases and Hygiene (Institute of Medicine, Petrozavodsk State University) and the Laboratory for Clinical Diagnostics of the Clinical Hospital at Petrozavodsk Station (Russian Railways JSC ). The study recruited 110 male and 132 female patients with NASH (242 patients in total) and 151 healthy individuals (64 males and 87 females). The healthy donors also underwent a medical checkup by the doctors of the Clinical Hospital at Petrozavodsk Station (Russian Railways JSC). All participants were divided into 2 groups: healthy controls with no clinical symptoms of NAFLD (mean age of 48.04 ± 2.26 years) and patients with NASH (mean age of 50.14 ± 2.46 years). The age did not differ significantly between the groups (U = 132.5; р = 0.637). The study included individuals of both sexes who gave informed consent to participate. Among other general criteria for inclusion were: Karelian residency, negative HBsAg and hepatitis C antibody tests (no chronic viral hepatitis), the absence of alcoholic, drug-induced or autoimmune liver diseases confirmed by medical history and clinical or laboratory tests. The main group included patients with a first-time diagnosis of mild to moderate NASH (prior to treatment). Exclusion criteria for both groups were: infectious or inflammatory diseases within a month before the study, pregnancy or lactation, smoking, diabetes mellitus, body mass index ≥ 30 kg/m2, drug therapy, intake of hepatotropic drugs. The diagnosis was established based on standard clinical, laboratory, instrumental and histological tests. The following blood parameters were evaluated: ALT, AST, and ALP (measured on the RandomAccessF-15 analyzer by BioSystems, Spain). Ultrasound scans revealed enlarged liver and increased parenchymal echogenicity in all patients with NASH. In some cases, the diagnosis was confirmed by liver biopsy.

Prior to drug therapy, 10 ml of venous blood were collected into EDTA-containing vacuum test tubes, of which 250 μL were used for DNA extraction. Some venous blood was used to obtain 200 μL plasma samples for measuring TNFα concentrations. The remaining blood volume was used for biochemistry tests.

The study was approved by the Committee on Medical Ethics of Petrozavodsk State University and Ministry of Health and Social Development of the Republic of Karelia (Protocol 39 dated November 15, 2017).

TNFα concentrations were measured in randomly selected blood plasma samples by ELISA using the Human TNFα Platinum ELISA kit (eBioscience, Austria). It total, 30 plasma samples of healthy donors (mean age of 49.11 ± 1.81 years) and 60 samples of patients with NASH (mean age of 49.95 ± 2.74 years) were tested; male and female samples were equally represented. The age did not differ significantly between the groups (U = 181.5; р = 0.535). Optical density of the solution was measured on the microplate reader Sunrise (Tecan, Austria) at 450 nm wavelength and 620 nm reference wavelength.

DNA was extracted from the peripheral blood on microcolumns using the K-Sorb kit (Syntol, Russia). Quality and quantity of the obtained DNA were evaluated on the SmartSpec spectrophotometer (Bio-Rad, USA).

To amplify the region of the TNFRSF1A gene harboring position 339 (rs767455), the following primers were used: forward 5’agtggctgaggttaggac3’ and reverse 5’ctatgcccgagt ctcaac3’ described in [21]. To amplify the region of the TNFRSF1B gene harboring position 587 (rs1061622), the following primers were used: forward 5’gcacacatcgtcactctc3’ and reverse 5’aaggagtgaatgaatgagac3’described in [21]. Polymerase chain reaction (PCR) was carried out in the iCycler iQ5 (Bio-Rad, USA) using a reaction mix by Evrogen, Russia. PCR products containing rs767455 were incubated with 1 unit Bsе1 I restriction endonuclease (SibEnzyme, Russia) for 3 hours at 65 °С. PCR products containing rs1061622 were incubated with 1 unit Fat I restriction endonuclease (SibEnzyme, Russia) for 1 hour at 55 °С. Then DNA fragments were separated in 1.5% agarose gel using the tris-acetate buffer.

The obtained data were processed in Statgraphics 2.1. Differences in allele and genotype frequencies between the two groups were assessed using the χ2 test; differences in biochemical parameters were assessed using the nonparametric Mann–Whitney–Wilcoxon U test. The latter was employed because distribution in the groups was not normal. To assess how different genotypes affected blood biochemistry, the Kruskal–Wallis test was used. To estimate the risk of developing NASH, we calculated the odds ratio (OR) and the 95% confidence interval (CI) [22]. Differences were considered significant at p < 0.05.

RESULTS

fig. 1 and fig. 2 show electrophoresis of rs767455-and rs1061622- containing PCR products after restriction digest.

TNFRSF1AT>C (rs767455) allele and genotype frequencies have been analyzed in patients with NASH and healthy controls.

The datasets were tested for deviations from the Hardy- Weinberg equilibrium. Both healthy controls and patients with NASH demonstrated deviations for allele and genotype frequencies (χ2 = 8.25 (df = 2, p < 0.05), χ2 = 21.64 (df = 2, p < 0.05), respectively).

tab. 1 shows that frequencies of T>C (rs767455) alleles and genotypes did not differ between patients with NASH and healthy donors.

We have also analyzed the frequencies of TNFRSF1B 587T>G alleles and genotypes in patients with NASH and healthy controls.

The two studied groups did deviate from the Hardy- Weinberg equilibrium (χ2 = 0.30 (df = 2, p > 0.05), χ2 = 4.16 (df = 2, p > 0.05) for healthy donors and patients with NASH, respectively).

tab. 2 shows that TNFRSF1B 587T>G allele and genotype frequencies differed between the healthy donors and patients with NASH. The G allele was far more frequent in patients with NASH than in healthy individuals. Carriers of the G allele are at a higher risk of developing NASH (OR = 4.83; 95% CI: 2.72–8.57).

We have also assessed the effect of the TNFRSF1B polymorphism (rs1061622) on liver function tests and plasma TNFαlevels (tab. 3). No significant differences were observed regarding the studied parameters between carriers of two different genotypes in the compared groups. The genotype did not have any effect on blood biochemistry both in patients with NASH and healthy controls (р > 0.05).

DISCUSSION

We have attempted to establish an association between two polymorphisms rs767455 and rs1061622 of genes TNFRSF1A and TNFRSF1B, respectively, and susceptibility to NASH. According to the literature, these polymorphisms are associated with a few inflammatory diseases and abnormal levels of TNFα in the blood plasma [23]. The rs767455 polymorphism of gene TNFRSF1A is a synonymous mutation at position 36 of exon 1. Synonymous mutations are known to disrupt mRNA splicing, alter mRNA structure and affect protein folding [24]. It has been shown that adenine to guanine substitution at position 36 of TNFRSF1A leads to a CCA to CCG codon change, disrupting translation [25]. In combination with other TNFRSF1A mutations (haplotype T-A-T at rs4149570-rs767455-rs1800692), it leads to the reduced abundance of exon 2-skipping products [26]. We have not established an association between the rs76 7455polymorphismandsusceptibilityto NASH in the study participants. However, we have discovered an association between the G allele carriership (rs1061622, TNFRSF1B) and the risk of this disease.

Thers 1061622 polymorphism of the TNFRSF1B gene is a thymine to guanine substitution at position 587 of exon 6 that leads to a methionine to arginine amino acid substitution at position 196 of the protein’s transmembrane domain, near the site of proteolytic cleavage by ADAM metalloproteases. This mutation affects ectodomain shedding (cleavage of the intracellular fragment of the transmembrane protein and its release into the extracellular matrix). Some researchers have shown that TT (Met196) genotype carriers have lower levels of sTNFRII than those with the Arg196 receptor variant [27]. Other authors report that carriers of TT+TG genotypes at this locus have higher levels of sTNFRII in the blood plasma than donors with the GG genotype [28].

Thus, the rs1061622 polymorphism can alter the ratio of membrane-bound to soluble TNFRII both in health and inflammation. Patients with liver diseases have elevated levels of sTNFRI and sTNFRII in the blood plasma and liver that positively correlate with disease severity [10, 12, 29, 30]. However, the role of increased ectodomain shedding of TNFα receptors in inflammation is not absolutely clear. Elevated concentrations of sTNFRII accompanied by reduced number of mbTNFRII on cell surface can trigger mbTNFRI-mediated signaling pathways leading to apoptosis [7]. Besides, soluble TNFR can act as physiological attenuators of TNFα activity, competing for the ligand with membrane-bound receptors. However, it appears that soluble receptors are capable of stabilizing and preserving circulating TNFα and thus act as its agonists [31].

The Met196 andArg196 variants of TNFRII differ in their ability to mediate TNF signaling and trigger apoptosis or necroptosis. Epithelial HeLaS3 cells transfected with the pcDNA3.1 plasmid containing the Arg196 allele of TNFRII demonstrated reduced activity of the nuclear factor kB and poor recruitment of TRAF2 upon stimulation with recombinant TNFα [32]. Subsequent activation of TNFRI signaling pathway in these cells induced apoptosis while in the cells transfected with the plasmid containing the wild type Met196, survival rates were better. Importantly, NASH is accompanied by the death of hepatocytes [33]. We hypothesize that the rs1061622 polymorphism of the TNFRSF1B gene contributes to the development and progression of NASH through activation of signaling pathways that induce hepatic cell death.

The rs1061622 polymorphism affects the levels of proinflammatory cytokines, which provides another explanation of its involvement into the etiology and pathogenesis of NASH [34]. We have studied plasma concentrations of TNFα in healthy and diseased carriers of different alleles and genotypes to reveal no significant differences between the groups. However, we cannot claim the absence of any effect of the TNFRSF1B polymorphism rs1061622 on TNFα levels because of a small sample size (especially true for the controls) and some other factors that may affect this parameter. Therefore, the role of this polymorphism in the development of NASH needs to be further investigated.

CONCLUSIONS

No association has been found between the rs767455 T>C TNFRSF1A polymorphism and the development of NASH in Karelian residents. We have however discovered an association between the rs1061622 T>G TNFRSF1B polymorphism and the disease. This polymorphic marker can be implicated in the genetic predisposition to NASH among the residents of Karelia.