Increase of fucosylated alpha-fetoprotein fraction at the onset of autoimmune hepatitis and acute liver failure

Authors


Abstract

Aim

Increased serum α-fetoprotein (AFP) has been associated with a good prognosis following acute liver failure (ALF), but the levels of the fucosylated fraction of AFP (Lens culinaris agglutinin-reactive fraction of AFP [AFP-L3]) following acute liver injury remain unknown. The aim of the present study was to investigate the clinical significance of AFP and AFP-L3 in patients with acute liver injury.

Methods

We investigated the serum levels of AFP and highly sensitive AFP-L3% in 27 patients with acute-onset autoimmune hepatitis (AIH), 28 patients with acute hepatitis (AH) and 22 patients with ALF at the onset using a highly sensitive immunoassay (micro-total analysis system).

Results

The serum AFP levels were increased in patients with AIH, AH and ALF, but the levels did not significantly differ among them. However, the mean AFP-L3% level was significantly higher in patients with AIH than in patients with AH (P = 0.0039). Moreover, significantly more patients with AIH demonstrated AFP-L3 positivity (≥10%) when compared with patients with AH (P = 0.014). Although the percentage of AFP-L3 positivity increased with AFP levels, at low serum AFP levels (<10 ng/mL), significantly more patients with AIH demonstrated AFP-L3 positivity than did patients with AH (P = 0.024) or ALF (P = 0.038).

Conclusion

We demonstrated for the first time that highly sensitive AFP-L3% levels were increased at the onset of AIH. The mechanism underlying the increase in AFP-L3 remains to be elucidated, but this finding may reflect an alteration of the glycosylation such as hyperfucosylation, which can influence the modifications of self-antigens in hepatocytes.

Introduction

α-Fetoprotein (AFP) HAS been widely used as a serological marker for diagnosing hepatocellular carcinoma (HCC), but increased AFP levels are found not only in patients with HCC but also in patients with chronic hepatitis and liver cirrhosis.[1] In contrast, fucosylated AFP, referred to as the L3 fraction of AFP (Lens culinaris agglutinin-reactive fraction of AFP [AFP-L3]), is a highly specific tumor marker for HCC.[2-4] However, measurement of AFP-L3 has not always been reliable using serum samples with a low total AFP concentration determined by conventional lectin affinity electrophoresis, or a liquid-phase binding assay system (LiBASys assay).[5] Recently, a more sensitive automated immunoassay using an on-chip electrokinetic reaction and separation by affinity electrophoresis (micro-total assay system [μTAS]) has been developed.[6, 7] This highly sensitive system involves microchip capillary electrophoresis and a liquid-phase binding assay system. The assay can detect concentrations of the fucosylated fraction of AFP that exceed 0.3 ng/mL. We have previously reported that the μTAS AFP-L3 is more sensitive for discriminating HCC from benign liver diseases than the conventional LiBASys AFP-L3, particularly in subgroups with lower AFP concentrations and early stage HCC.[8] Other studies have also reported the usefulness of highly sensitive AFP-L3 for the early detection of HCC and the prediction of HCC recurrence after curative treatment.[9-13]

α-Fetoprotein is a single-strand glycoprotein with a molecular weight of 68 kDa, and it belongs to the albumoid gene family, which includes albumin, vitamin d-binding protein, AFP and α-albumin.[14, 15] AFP is expressed at relatively high levels in the fetal liver, particularly in those with congenital abnormalities.[16, 17] Serum AFP levels decline within several weeks after birth and usually remain low throughout life. Reactivation of AFP production in adults occurs during hepatocarcinogenesis and hepatic regeneration.[15] In addition to HCC, the increase of serum AFP levels has been suggested to reflect hepatic regeneration in response to liver injury, with a better survival rate.[17-19] However, the presence of AFP-L3 during acute liver failure (ALF) remains to be elucidated.

Autoimmune hepatitis (AIH) is characterized by chronic inflammation of the liver, interface hepatitis, hypergammaglobulinemia and production of autoantibodies.[20, 21] Although AIH typically presents as a chronic necroinflammatory liver disease, an acute presentation occurs in up to 25%, a small minority of whom progress to autoimmune ALF.[22-24] Although patients with AIH usually respond well to corticosteroid treatment, patients progressing to ALF become resistant to corticosteroid treatment, and their prognosis is poor.[23] As mentioned above, AFP has been reported to be a prognostic marker of ALF, but the AFP and AFP-L3 levels in patients with AIH including ALF have not been reported. In the present study, we investigated the AFP and AFP-L3 levels in patients with ALF and acute-onset AIH using a highly sensitive method and evaluated the significance of these markers on the diagnosis of acute-onset AIH.

Methods

Study cohort

This work was conducted in accordance with the Declaration of Helsinki. The study protocol was approved by the Institutional Ethics Committee of Niigata University School of Medicine. Written informed consent for the use of stored serum samples for the study was obtained from the enrolled patients.

A total of 27 patients with acute-onset AIH, 28 patients with acute hepatitis (AH) and 22 patients with ALF who were diagnosed and treated in our hospital between 2010 and 2012 were enrolled in the present study. Three patients diagnosed as autoimmune ALF were excluded from the present study. A diagnosis of AIH was made on the basis of the presence of antinuclear antibody (ANA) and/or anti-smooth muscle antibody (ASMA), according to the classical and simplified criteria defined by the International Autoimmune Hepatitis Group,[25, 26] and liver biopsy findings compatible with AIH. Eligibility criteria for clinically acute-onset AIH were as follows: (i) acute-onset liver injury; (ii) no history of liver injury; and (iii) no signs of chronicity on the basis of physical examination, laboratory data and abdominal ultrasound or computed tomography findings.

A diagnosis of ALF was made according to the diagnostic criteria proposed by the Intractable Hepato-Biliary Diseases Study Group of Japan.[27] Briefly, patients showing prothrombin time (PT) values of 40% or less of the standardized values or international normalized ratios (INR) of 1.5 or more caused by severe liver damage within 8 weeks of symptom onset were diagnosed as having ALF. Patients showing PT values of more than 41% of the standardized values or INR of less than 1.5 caused by acute liver damage were diagnosed as AH. Clinical profiles (age, sex, etiology and the presence of hepatic encephalopathy) and laboratory data at the time of diagnosis were analyzed and compared.

Assays of highly sensitive AFP-L3 and AFP

For this retrospective study, the levels of highly sensitive AFP-L3 and AFP were determined using the μTAS assay (Wako Pure Chemical Industries, Osaka, Japan). The total AFP concentration (ng/mL) in the serum sample was determined by the summation of AFP-L3 and AFP-L1, and then the AFP-L3 level percentage (AFP-L3%) was calculated. The serum level of des-γ-carboxyprothrombin (DCP) was also measured using the μTAS assay. The lower limits of quantitation for AFP and DCP were 0.3 ng/mL and 5 mAU/mL, respectively. The cut-off values for AFP, AFP-L3% and DCP positivity were set at 20 ng/mL, 10% and 40 mAU/mL, respectively.

Statistical analysis

The significance of differences was statistically analyzed by Fisher's exact test, the compared Student's t-test with Welch's correction, or the Mann–Whitney U–test, using SPSS software (version 18; SPSS, Chicago, IL, USA). The level of significance was set at P < 0.05.

Results

Clinical features of the patients

Of 27 patients with AIH, five were male and 22 were female. The mean age at diagnosis was 55.0 ± 16.2 years, which was significantly higher than those of patients with AH (44.7 ± 15.0 years, P = 0.018) and ALF (44.1 ± 16.6 years, P = 0.026). The etiologies of AH were acute hepatitis B virus (HBV) infection in 11 patients (39.3%), drug-induced liver injury (DILI) in eight (28.6%), indeterminate in three (10.7%) and others in six. The etiologies of ALF were HBV infection in 10 patients (45.5%), DILI in six (27.3%), indeterminate in three (13.6%) and others in six. The mean alanine aminotransferase (ALT) level in patients with AIH was 448.4 ± 372.7 IU/L and was significantly lower than those in patients with AH (1809.8 ± 1982.3 IU/L, P = 0.0016) and ALF (2391.7 ± 2410.0 IU/L, P = 0.0012). The mean PT activity in patients with AIH was 76.1 ± 24.7%, which was significantly higher than that in patients with ALF (26.3 ± 10.2%, P < 0.001). Information regarding patient profiles and biochemical features are summarized in Table 1.

Table 1. Patient characteristics at admission
FactorsAIHAHALF
  1. *P < 0.01 (compared to AIH), **P < 0.05 (compared to AIH), †P < 0.01 (compared to AH).
  2. AH, acute hepatitis; AIH, autoimmune hepatitis; ALF, acute liver failure; ALT, alanine aminotransferase; AST, aspartate aminotransferase; EBV, Epstein–Barr virus; HAV, hepatitis A virus; HBV, hepatitis B virus; HEV, hepatitis E virus; IgG, immunoglobulin G; INR, international normalized ratio; PT, prothrombin time; SD, standard deviation.
No. of patients272822
Age (mean ± SD, years)55.0 ± 16.244.7 ± 15.0**44.1 ± 16.6**
Sex (female/male)22/514/149/13
Etiology   
HAV/HBV/HEV/EBV2/11/1/21/10/0/0
Drug/alcohol/indeterminate8/1/36/2/3
Hepatic encephalopathy (0/I–II/III–IV)25/2/028/0/08/6/8
Laboratory examinations (Mean ± SD) (range)
Albumin (g/dL) (range)3.6 ± 0.7 (2.1–4.8)3.8 ± 0.6 (2.5–4.7)3.2 ± 0.6 (2.4–4.5)
IgG (mg/dL) (range)2389.8 ± 1176.1 (856–6578)1594.4 ± 504.1* (774–3270)1420.2 ± 337.7* (771–1867)
AST (IU/L) (range)363.5 ± 326.9 (36–1478)1579.5 ± 2261.9* (145–9314)3380.8 ± 3826.5* (22–11720)
ALT (IU/L) (range)448.4 ± 372.7 (39–1214)1809.8 ± 1982.3* (28–7459)2391.7 ± 2410.0* (54–7530)
Total bilirubin (mg/dL) (range)7.3 ± 8.3 (0.5–25.9)3.8 ± 4.0 (0.7–17.9)10.8 ± 7.8 (2.6–27.2)
Platelet counts (×104/mm3) (range)17.1 ± 7.0 (3.5–32.7)18.0 ± 6.5 (6.3–39.1)10.1 ± 6.6*, (4.1–35.8)
PT (%) (range)76.1 ± 24.7 (36.3–126)74.7 ± 19.0 (47–131)26.3 ± 10.2*, (6–39)
PT (INR) (range)1.21 ± 0.36 (0.86–2.33)1.13 ± 0.14 (0.89–1.45)2.85 ± 1.41*, (1.70–7.02)
Death/liver transplantation0/00/06/1

The mean immunoglobulin G level in patients with AIH was 2389.8 ± 1176.1 mg/dL, which was significantly higher than those in patients with AH (1594.4 ± 504.1 mg/dL, P < 0.001) and ALF (1420.2 ± 337.7 mg/dL, P < 0.001). ANA was positive in 21 patients with AIH (77.8%), in eight patients with AH (35.7%) and in two patients with ALF (9.1%). ASMA was positive only in seven patients with AIH (23.3%). Three patients with AIH (10.0%) fulfilled the criteria of ALF and died of liver failure.

AFP, AFP-L3% and DCP levels

Table 2 summarizes the AFP, AFP-L3% and DCP levels at the onset. The mean AFP levels of patients with AIH, AH and ALF were 32.9 ± 48.0, 56.2 ± 164.8 and 29.0 ± 42.4 ng/mL, respectively. Although there was no significant difference among them, 22 patients (78.6%) with AH showed AFP levels less than 10 ng/mL (Fig. 1). The mean AFP-L3% level of patients with AIH was 12.4 ± 9.1% and was significantly higher than that in patients with AH (5.6 ± 7.5%) (P = 0.0039). The number of AFP-L3 positive patients with AIH (16 patients, 59.3%) was significantly higher when compared with patients with AH (seven patients, 25.0%) (P = 0.014). The levels of AFP and AFP-L3% in patients with AIH were similar to those in patients with ALF, and no significant difference was shown.

Table 2. α-Fetoprotein and its fucosylated fraction
FactorsAIHAHALF
  1. *P = 0.0039 (compared to AIH), **P = 0.014 (compared to AIH).
  2. AFP, alpha-fetoprotein; AFP-L3, Lens culinaris agglutinin-reactive fraction of AFP; AH, acute hepatitis; AIH, autoimmune hepatitis; ALF, acute liver failure; DCP, des-γ-carboxyprothrombin.
No. of patients272822
AFP (ng/mL) (range)32.9 ± 48.0 (1.7–200)56.2 ± 164.8 (1.0–765.1)29.0 ± 42.4 (0.6–119.5)
AFP-L3% (%) (range)12.4 ± 9.1 (0.5–35.6)5.6 ± 7.5* (0.5–26.8)13.1 ± 17.8 (0.5–58.4)
Percentage of AFP-L3 positivity (≥10) (%)59.3 (16/27)25.0** (7/28)40.9 (9/22)
DCP (mAU/mL) (range)63.2 ± 151.7 (0–794)43.4 ± 67.2 (8–376)44.8 ± 53.7 (6–252)
Figure 1.

Percentages of serum α-fetoprotein (AFP) levels in patients with autoimmune hepatitis (AIH), acute hepatitis (AH) and acute liver failure (ALF). Percentages of serum AFP levels (<10 ng/mL, 10–100 and >100 ng/mL) in patients with AIH (n = 27), AH (n = 28) and ALF (n = 22) are shown. image, <10 ng/mL; image, 10–100 ng/mL; image, >100 ng/mL.

Relationships between AFP and AFP-L3%

We then analyzed the relationships between AFP/AFP-L3% levels and the ALT level, but no significant difference was observed between the AFP/AFP-L3% levels and the ALT level (Fig. 2). Although we further analyzed the relationships between AFP/AFP-L3% levels and other laboratory parameters, including cholinesterase, PT value and the ratio of direct bilirubin to total bilirubin (D/T ratio), no significant differences were observed between the AFP/AFP-L3% levels and these parameters (data not shown). However, increased AFP-L3% levels were correlated with increased AFP, and the AFP-L3% levels were proportional to the AFP levels in patients with AIH (R2 = 0.194, P = 0.021), AH (R2 = 0.274, P < 0.001) and ALF (R2 = 0.415, P = 0.023) (Fig. 3a,b). Interestingly, among patients showing AFP levels less than 10 ng/mL, significantly more patients with AIH (6/15 patients, 40.0%) demonstrated AFP-L3 positivity than patients with AH (2/22 patients, 9.1%) (P = 0.024) and ALF (1/14 patients, 7.1%) (P = 0.038) (Fig. 3c).

Figure 2.

Associations between α-fetoprotein (AFP) and alanine aminotransferase (ALT) and between Lens culinaris agglutinin-reactive fraction of AFP percentage (AFP-L3%) and ALT in patients with autoimmune hepatitis (AIH), acute hepatitis (AH) and acute liver failure (ALF). Associations between serum AFP (ng/mL) and ALT (IU/L), and associations between serum AFP-L3% (%) and ALT (IU/L) in patients with AIH (a), AH (b) and ALF (b) are shown. Each plot shows the serum AFP or AFP-L3 and ALT in each case. (a) image, AIH; (b) image, AH; image, ALF.

Figure 3.

Associations between α-fetoprotein (AFP) and Lens culinaris agglutinin-reactive fraction of AFP percentage (AFP-L3%) in patients with autoimmune hepatitis (AIH), acute hepatitis (AH) and acute liver failure (ALF). (a,b) Associations between serum AFP (ng/mL) and AFP-L3% (%) in patients with (a) AIH, (b) AH and (b) ALF are shown. Each plot shows the serum AFP and AFP-L3 in each case. (c) Percentages of AFP-L3 positive patients among different AFP levels (<10, 10–100 and >100 ng/mL) are shown. (a) image, AIH R2 = 0.194, P = 0.021; (b) image, AH R2 = 0.274, P < 0.001; image, ALF R2 = 0.415, P < 0.023; (c) image, AIH; image, AH; image, ALF.

Discussion

In the present study, we revealed that highly sensitive AFP-L3% was increased in patients with acute liver injury, including acute-onset AIH and ALF, using the μTAS system, which can detect AFP-L3 concentrations that exceed 0.3 ng/mL. Although elevated AFP levels have been associated with ALT elevation, hepatocyte regeneration and hepatic fibrosis in patients with benign liver diseases such as chronic viral hepatitis,[1, 17, 28] AFP and AFP-L3% levels were not associated with increased ALT levels in the studied patients. However, increased AFP-L3% levels were correlated with increased AFP levels in patients with ALF. The AFP and AFP-L3% levels in patients with AIH were similar to those in patients with more severe liver injuries diagnosed as ALF; however, among patients showing low AFP levels, the AFP-L3% level was significantly higher in patients with AIH than in patients with AH or ALF.

Following acute liver injury with extensive necrosis, an increase in AFP is interpreted as a sign of dedifferentiated hepatic regeneration.[17, 29] AFP has been studied predominantly in patients with non-acetaminophen-induced liver failure, and the majority of studies have found that an elevation of AFP is associated with a favorable prognosis.[19, 29, 30] A study limited to patients with acetaminophen-induced liver injury also suggested that AFP levels more than 20 ng/mL are associated with a favorable outcome.[29] On the other hand, other studies showed that higher absolute values of AFP on admission did not predict a favorable outcome in patients with ALF, although a rising level of AFP over the first 3 hospital days frequently indicated survival.[17] Yang et al. also reported the remarkable usefulness of serial AFP monitoring for a more accurate predictor of prognosis in ALF with chronic HBV infection.[19] In the present study, the median values of AFP and AFP-L3% were 2.5 ng/mL (range, 0.6–119.5) and 0.5% (range, 0.5–58.4%) in 15 surviving patients with ALF and 23 ng/mL (range, 0.9–93.2) and 15.6% (range, 0.5–41.7%) in six deceased and one liver transplanted patient, respectively. Although these values appeared to be higher in the deceased and liver transplanted patients, we could not find any significant relationship between AFP/AFP-L3% and prognosis of patients with ALF. We also found that the number of patients showing an AFP level greater than 10 ng/mL was higher in patients with AIH than in patients with AH. Moreover, although the degree of liver injury in patients with AIH was not as severe as that in patients with ALF, the AFP levels in patients with AIH were similar to those in patients with ALF. Consistent with our results, Schiødt et al. reported that patients with AIH demonstrated higher AFP levels than those with acute liver injuries such as ischemia and acetaminophen-induced liver injury.[17] This may suggest that absolute AFP levels are related more closely to the characteristic pattern of injury for each etiology than to the acuity of illness or the likelihood of survival.[17] Although we could not evaluate the serial AFP levels and the association between the changes in AFP level and survival, our findings may provide new insight into the pathophysiology of AIH, particularly liver regeneration in AIH.

Lens culinaris agglutinin-reactive fraction of AFP is a highly specific tumor marker for HCC.[2-4] Although increases in serum AFP levels are observed in patients with chronic liver diseases, AFP-L3 is scarcely detected in benign liver diseases.[31] Moreover, the advantage of AFP-L3 measured by the conventional method is limited due to its low sensitivity, particularly in patients with a total AFP below 20 ng/mL.[28] Therefore, there was no previous report describing AFP-L3 levels in patients with acute-onset AIH. Nakagawa et al. have shown that α1-6 fucosyltransferase (Fut8) is involved in the fucosylation of AFP,[32] and high expression of Fut8 was observed in non-cancerous liver cirrhotic tissues as well as in HCC tissues.[33] However, the levels of guanosine 5′-diphospate (GDP)-fucose, which is a donor substrate for Fut8, were reported to be significantly higher in HCC tissues than in cirrhotic and normal liver tissues,[34] and the enhanced expression of the human homolog of GDP-4-keto-6-deoxymannose-3,5-epimerase-4-reductase (FX), which contributes to the synthesis of GDP-fucose, was associated with the increase of GDP-fucose in HCC. Therefore, both Fut8 and FX have been shown to regulate the production of AFP-L3.[31] Moreover, Nakagawa et al. reported that fucosylation is a possible signal for the secretion of hepatic glycoproteins into bile and that the fucosylation-based sorting machinery may be disrupted in the liver-bearing HCC.[31, 35] Then, AFP-L3 would increase in the sera of HCC patients. Because low levels of AFP-L3 are detected in the sera of patients with benign liver diseases,[31] severe liver damage such as ALF would cause the disruption of fucosylation-based sorting machinery, resulting in the increase of AFP-L3 in sera. Therefore, the increase of AFP-L3 in patients with AIH may also result from a severe liver damage.

Fucosylation is one of the most important types of glycosylation in a number of pathological conditions, including cancer and inflammation.[36] Glycosylation is a post-translational modification of many secreted proteins, and it has been known that structural changes in the glycan structures of serum proteins are an indication of liver damage.[37] Interestingly, pro-inflammatory cytokines have been shown to influence the expression levels of a number of Golgi glycosyltransferases, which can result in an alteration in glycosylation.[38] Because many self-antigens are glycosylated, such alterations in glycosylation may affect the epitope modification of self-antigens recognized by T cells. In fact, there are a number of possible pathways by which changes in glycosylation of self-antigens can result in the loss of tolerance.[38] Although the significance of AFP-L3 on the pathogenesis of AIH remains unknown, the increase in AFP-L3, particularly in patients with AIH who showed a total AFP below 10 ng/mL, may reflect a hyperfucosylation status of glycoproteins produced in the liver. We hypothesize that an alteration in glycosylation such as hyperfucosylation may involve the post-translational modifications of self-antigens and that the consequence of changes in glycosylation may generate neo-self-epitopes, resulting in the loss of tolerance and the development of autoimmunity against hepatocytes. Therefore, we consider that the status of protein glycosylation should be further investigated by analyzing the mRNA and protein levels of the corresponding glycosyltransferases in patients with AIH. Consistent with our results, Sakurai et al. reported that serum AFP-L3 levels in patients with fulminant hepatic failure at the onset of encephalopathy was significantly increased.[39] Interestingly, these authors also reported that the residual liver volume significantly correlated with the serum AFP-L3 level, but not with AFP. Therefore, the increase of AFP-L3 in patients with AIH may be associated with an increased ability of hepatocyte regeneration.

In conclusion, we have demonstrated for the first time that the highly sensitive AFP-L3% was increased in patients with acute liver injury, including acute-onset AIH and ALF, and that the AFP-L3% level was significantly higher in patients with AIH than in patients with AH and ALF, particularly among patients showing low AFP levels. Although the significance of AFP-L3 on the pathogenesis of AIH remains unknown, we speculate that increases in AFP-L3 may reflect severe liver damage that causes the disruption of fucosylation-based sorting machinery of glycoproteins into bile, increased liver regeneration capability and a hyperfucosylation status that influences the modifications of self-antigens in hepatocytes. Our results, which were obtained using a highly sensitive assay, may provide new insights into the pathogenesis of AIH.

Acknowledgments

This work was supported in part by Grants-in-Aid for Scientific Research (C) (24590963 to S. Y.) from Japan Society for the Promotion of Science (JSPS), and Grants-in Aid from the Ministry of Health, Labor and Welfare of Japan to the Study Group of Intractable Hepatobiliary Diseases.