The Loss-of-Function S267F Variant in HBV Receptor NTCP Reduces Human Risk for HBV Infection and Disease Progression

Abstract

Background

Sodium taurocholate cotransporting polypeptide (NTCP, SLC10A1) is a hepatocyte receptor for hepatitis B virus (HBV) infection. The natural NTCP S267F variant causes loss of NTCP HBV receptor function. We assessed the association of S267F with HBV resistance, HBV infection clearance, and HBV-related cirrhosis and hepatocellular carcinoma (HCC).

Methods

We tested the effects of S267F in 1117 Han Chinese patients with various HBV infection outcomes using multivariate logistic regression analysis.

Results

The frequency of S267F (T allele) was higher in HBV-resistant healthy controls (n = 179, 4.0%) compared to HBV-infected patients (n = 648, 1.5%); odds ratio (OR) 0.32 (95% confidence interval [CI] 0.15–0.68; P = .003; dominant model). 267F variant genotypes were also associated with reduced risk for cirrhosis (n = 192, 0.5%) and HCC (n = 258, 1.0%) compared to those with chronic HBV infection (n = 202, 3.0%); OR 0.15 (95% CI, 0.03–0.70) and OR 0.21 (95% CI, 0.062–0.72), respectively. There was no association of the S267F variant with spontaneous HBV clearance.

Conclusion

The S267F variant for the HBV cell-entry receptor NTCP was associated with increased resistance to HBV infection and decreased risk for cirrhosis and liver cancer among those with chronic HBV infection.

The NTCP S267F variant diminishes the HBV receptor function of NTCP. This genetic epidemiological study of a full-spectrum HBV outcomes cohort demonstrated its association with human resistance to HBV infection and progression from chronic infection to liver cirrhosis or cancer.

Hepatitis B virus (HBV) infection is a major public health problem, affecting 257 million people worldwide [1]. Individuals with chronic HBV infection (CHB) have considerably higher risk of developing liver cirrhosis and hepatocellular carcinoma (HCC). Current anti-HBV drugs are not effective in curing HBV infection, and identification of therapeutic targets is hindered by a lack of knowledge of HBV cellular interacting targets. Identifying the host genetic factors determining HBV natural outcomes is one of the important research priorities for the discovery of a cure for CHB, as recommended by experts at a recent HBV workshop [2, 3]. Previous genetic association studies have identified HLA-DPA1 and HLA-DPB1 as the major genes associated with HBV infection and clearance [4–6].

Na⁺-taurocholate cotransporting polypeptide (NTCP), encoded by the solute carrier family 10 member 1 gene (SLC10A1), is a transmembrane protein expressed exclusively in the basolateral membrane of the hepatocyte, and accounts for 80% of hepatic uptake of bile salts [7]. Recent landmark work by Li and colleagues identified NTCP as a hepatocellular entry receptor of HBV [8]. Small interfering RNA-mediated knockdown of NTCP expression in primary human hepatocytes, primary Tupaia hepatocytes, and HepaRG cell lines reduced HBV infection, while overexpression of NTCP conferred HBV susceptibility to nonsusceptible hepatocarcinoma cells [8–11]. These studies established NTCP as a functional receptor for HBV infection.

A nonsynonymous variant Ser267Phe (S267F, c.800 C > T, rs2296651) in NTCP is found only on East Asian chromosomes (minor allele frequency 2%–10%) and is absent in African and European populations [9, 12]. The 267F variant diminishes the HBV receptor function of NTCP, as evidenced by loss of HBV binding, and drastically reduced HBV cell entry and HBV replication [9]. The variant also leads to nearly complete loss of the bile acid transporter activity of NTCP [13].

Despite these in vitro results, epidemiological studies showing that the loss-of-function NTCP variant is associated with HBV-mediated clinical outcomes are inconsistent [14–16]. It has not been resolved if NTCP S267F variant is associated with protection against cirrhosis developing into HCC [16]. It is also uncertain whether the variant is differentially associated with distinct HBV outcomes, including resistance to HBV infection and resolution of HBV infection. In addition, the potential interaction of NTCP with the known disease-modifier HLA-DPA1 affecting HBV clinical outcomes has not been reported. Here, we assessed the impact of NTCP S267F on the full spectrum of HBV outcomes (HBV resistance, HBV spontaneous clearance, CHB, cirrhosis, and HCC) in 1117 Han Chinese.

MATERIALS AND METHODS

Full-Spectrum HBV Outcomes Cohort

A protocol review committee at National Institute of Health approved the study design and questionnaire. An internal review board at the National Institute of Health approved the study (IRB number 02CN323-D). Local internal review board approvals from participating hospitals and written informed consent from participants were obtained.

Following a recruitment plan specifically designed for HBV genetic association studies, we enrolled self-reported Han Chinese participants between 2003 and 2007 in several major hospitals specializing in infectious diseases, in Beijing and other cities in Northern China [17]. All recruited participants were ≥40 years old at enrollment, allowing sufficient time for HBV exposure and disease development. By restricting enrollment to those ≥40 years of age, we also avoided confounding by HBV vaccination, which was initiated in the mid-1980s [18] and, starting in 1992, was required for newborns by the Ministry of Health of China (http://www.chinacdc.cn/). HBV transmission mainly occurs during perinatal and early childhood in China and early childhood infection is strongly associated with persistent HBV infection [19]. Hepatitis specialists conducted interviews and physical examination and filled in structured questionnaires, which included comprehensive family history, environmental risk exposures, and health history. We excluded patients with other hepatic viral infections or coinfection (hepatitis A virus, hepatitis C virus, hepatitis D virus, or hepatitis E virus), autoimmune conditions, alcoholic hepatitis, and drug-induced hepatitis.

HBV Outcome Group Definition

The case-control study comprised the full spectrum of HBV natural history: HBV case groups included natural clearance, chronic asymptomatic and symptomatic HBV infection, cirrhosis, and HCC. HBV hypernormal controls were defined as those at least 40 years of age at enrollment lacking serological evidence of previous or current HBV infection (negative for hepatitis B surface antigen [HBsAg], anti-HBs, anti-HBc, hepatitis B e antigen [HBeAg], and anti-HBe) and had alanine aminotransferase (ALT) within the normal range [4, 17]. We used the term “hypernormal” healthy control to make the distinction from the general “normal” population that has high prevalence of anti-HBc, indicative of past HBV infection.

HBV clearance was defined for subjects who were (1) negative for HBsAg and positive for anti-HBs and anti-HBc or (2) anti-HBs positive with no self-reported or clinic/hospital record of hepatitis B vaccination. CHB infection was defined by 2 positive tests at least 1 year apart for HBsAg and anti-HBc, and with no evidence of cirrhosis by imaging and laboratory tests. Case definitions of HBV infection, clearance, chronic infection, cirrhosis, and HCC are in accordance with the predefined criteria [17], based on the diagnosis protocol issued by the Association of Infectious Diseases and Parasites Diseases of China [20].

Criteria for Liver Cirrhosis and HCC

Liver cirrhosis was diagnosed by sonography, computed tomography (CT), or magnetic resonance imaging (MRI); supporting laboratory evidence included low platelet counts, prolonged prothrombin time, and liver synthetic dysfunction. Decompensated liver cirrhosis was diagnosed by at least 1 of the following: severe gastroesophageal varication (3°), history of upper gastrointestinal bleeding or current bleeding, ascites or edema, hepatic encephalopathy, serum albumin <3.5 g/L, total bilirubin >35 μmol/L, and evidence of liver cirrhosis by sonography, CT, or MRI. The diagnosis of HCC was confirmed by histopathological evidence, or at least 2 out of 3 imaging examinations consistent with HCC plus elevated serum α-fetoprotein level.

Virological Laboratory Tests

Serum HBsAg, antibody to HBsAg (anti-HBs), anti-HBc, and antibody to hepatitis C virus (anti-HCV) titers were determined by the chemiluminescent assay (Ortho Clinical Diagnostics, NY) and HBeAg and anti-HBe by the microparticle enzyme immunoassay kit (AxSYM, Abbott, IL). Serum HBV DNA levels were determined by real-time polymerase chain reaction assay with detection limit of 1000 copies/mL. All participants in this study were negative for antibodies to HAV, HDV, and HEV.

Genotyping of the NTCP S267F Variant

The S267F, rs2296651 C > T polymorphism was genotyped using a commercial TaqMan allele discrimination assay (Applied Biosystems, assay ID: C__16184554_10) on the ABI 7900HT sequencer detector system according to the manufacturer’s protocol. For quality control, water controls were included on each plate and 10% of samples were duplicated. No contaminated water control or genotype mismatches between duplicates was observed.

Statistical Analysis

All analyses were performed using SAS version 9.2 (SAS Institute, Cary, NC). Logistic regression models were applied to case-control comparisons for the different phenotypic outcomes using a dominant genetic model, with or without adjusting for sex, age groups (40–44, 45–54, and ≥55 years old). We did not test the recessive model because there were only 2 homozygous 267F/F carriers. To account for potential population substructures among patients, we adjusted for the first 2 principal components generated by Eigensoft [21] using genotypic data from the Affymetrix single nucleotide polymorphism array 6.0 . Odds ratios (OR) and P values were determined using a conditional logistic regression test. Student t test was used to analyze quantitative differences between groups. All tests were 2-sided and a P value <.05 was considered statistically significant.

RESULTS

The demographic characteristics and NTCP S267F genotype distributions in multiple HBV outcome groups are presented in Table 1. The NTCP S267F genotype distributions in hypernormal controls and all samples combined did not deviate from Hardy-Weinberg equilibrium expectations (P > .05).

Table 1.

Distribution of NTCP 267F in Multiple HBV Outcome Groups

HBV Infection Outcome	n	Female Sex, %	Age, Year Mean (± SD)	ALT ≥ 45 IU/L, %	HBV DNA Level, log10 Copies/mL (± SD)	267 T Allelea Frequency, %
HBV resistant	179	39.33	49.03 (7.33)	1.12	…	3.9
Clearance	286	28.67	49.92 (9.74)	8.50	…	1.7
Chronic hepatitis	202	30.22	47.03 (6.19)	35.64	4.73 (1.97)	3
Cirrhosis	192	36.98	50.63 (8.46)	60.94	5.23 (1.64)	0.52
Hepatocellular carcinoma	258	35.66	51.18 (10.28)	58.53	4.38 (1.70)	1

Abbreviations: ALT, alanine aminotransferase; HBV, hepatitis B virus; NTCP, sodium taurocholate cotransporting polypeptide.

^a267 T allele specifies 267F.

NTCP S267F and Resistance to HBV Infection

To determine if the NTCP S267F variant affects risk for CHB, we compared genotype frequencies between the hypernormal control individuals (n = 179) and those with CHB, including HBV-cirrhosis and HBV-HCC (n = 648) (Table 2). The 267F (allele T) variant genotypes were significantly elevated in HBV-resistant hypernormal controls (4.0%) compared to those with CHB (1.5%), with reduced risk for HBV infection (dominant model, OR = 0.37; 95% confidence interval [CI], 0.18–0.76; P = .005). Adjusting for sex, age, and the first 2 principal components in this cohort did not substantially affect association (OR_adj = 0.32; 95% CI, 0.15–0.68; P = .003) (Table 2 and Figure 1).

Table 2.

Distribution of NTCP S267F in HBV-Resistant and Infected Groups

S267F Genotypea	HBV Resistant, n (%) (Total = 179)	HBV Infected, n (%) (Total = 648)	Genetic Model	OR (95% CI)	P
CC	166 (92.74)	630 (97.22)	Dominant	1
CT/TT	13 (7.26)	18 (2.78)	Crude	0.37 (0.18–0.76)	.007
			Adjustedb	0.35 (0.17–0.73)	.005
			Adjustedc	0.32 (0.15–0.68)	.003
T allele	14 (3.91%)	19 (1.47%)	Additivec	0.15 (0.04–0.61)	.008

Abbreviations: CI, confidence interval; HBV, hepatitis B virus; NTCP, sodium taurocholate cotransporting polypeptide; OR, odds ratio.

^aC specifies S267 and T specifies 267F.

^bAdjusted by age, stratified by gender.

^c Adjusted by age, stratified by gender, plus first 2 principal components.

Figure 1.

Allele frequency distribution of sodium taurocholate cotransporting polypeptide (NTCP) 267F in various hepatitis B virus (HBV) outcome groups. P values were from multivariate logistic regression. Abbreviation: HCC, hepatocellular carcinoma.

Interaction Between NTCP and HLA-DPA1 on HBV Resistance

Carriage of the NTCP 267F variant genotypes (OR = 0.37; 95% CI, 0.17–0.75), HLA-DPA1 rs3077 T allele (0.50; 95% CI, 0.35–0.71), and increasing age (1.29; 95% CI, 1.02–1.63) were independently associated with HBV resistance. However, we observed no interaction between these 3 terms (P for interaction = .43).

NTCP S267F and HBV Clearance

We then compared participants with spontaneous HBV clearance (n = 286) with participants with CHB (n = 648). The NTCP 267F variant genotypes were not significantly associated with HBV clearance in the crude or adjusted models (P ≥.53).

Interaction of NTCP and HLA-DPA1 Variants on HBV Clearance

We next tested the joint and added effects of NTCP with HLA-DPA1 in viral clearance, because the latter is a major predictor of HBV spontaneous clearance [4, 6]. We found no interaction between HLA-DPA1 rs3077 and NTCP S267F for HBV clearance (dominant model, interaction term, P = .79). These results suggest that NTCP, serving as an entry receptor for infection of hepatocytes, does not play a role in the host immune response that mediates HBV clearance.

NTCP S267F and Risk for Cirrhosis

To determine if the S267F variant affects risk of developing cirrhosis, we compared the S267F variant genotypes in the cirrhosis group with the chronic hepatitis alone group. We included the decompensated cirrhosis patients but not those with compensated cirrhosis to avoid the potential overlaps between those with compensated cirrhosis and those with early cirrhosis in the CHB group not detected by imaging diagnosis. We found that the carriage frequency of 267F was lower in the cirrhosis group (0.52%) than in the CHB group (2.97%), conferring an 85% reduction in risk (dominant, OR_adj = 0.15; 95% CI, 0.03–0.74; P = .02) for developing cirrhosis (Table 3). Further adjusting with HLA-DPA1 rs3077 did not affect the cirrhosis outcome (P = .02) nor did rs3077 itself, suggesting that NTCP S267F but not HLA-DPA1 rs3077 reduces cirrhosis development.

Table 3.

Distribution of NTCP Variant Genotypes in Chronic Hepatitis and Cirrhosis Groups

S267F Genotypea	Cirrhosis (Total = 192)	Chronic Hepatitis (Total = 202)	Genetic Model	OR (95% CI)	P
CC	190 (92.74)	191 (97.22)	Dominant	1
CT/TT	2 (7.26)	11 (2.78)	Crude	0.18 (0.04–0.84)	.014
			Adjustedb	0.16 (0.03–0.75)	.02
			Adjustedc	0.15 (0.03–0.70)	.017
T allele	2 (0.52%)	12 (2.97%)	Additivec	0.15 (0.03–0.74)	.019

Abbreviations: CI, confidence interval; NTCP, sodium taurocholate cotransporting polypeptide; OR, odds ratio.

^aC specifies S267 and T specifies 267F.

^bAdjusted by age and gender.

^cAdjusted by age and gender plus first 2 principal components.

NTCP S267F and Development of HCC

We also assessed the association between NTCP S267F and HCC by comparing HCC participants (n = 258) with participants with CHB or cirrhosis (n = 394). S267F was not significantly associated with the risk for HCC development (OR = 0.53; 95% CI, 0.19–1.53; P = .24). S267F was also not significantly associated with risk for HCC when compared to cirrhosis alone (n = 192; OR = 1.55; 95% CI, 0.28–8.58; P = .62).

S267F, however, was significantly associated with a reduced risk for developing HCC (OR = 0.21; 95% CI, 0.062–0.72; P = 0.013; Figure 1), when its carriage was compared between HCC and CHB alone (without cirrhosis).

NTCP S267F and Overall Disease Progression

Subsequently, we asked whether S267F affects overall disease progression rather than HCC specifically. When we combined the HCC and cirrhosis groups together and then compared it with the CHB group, 267F was significantly associated with reduced disease progression to HCC or cirrhosis (OR = 0.19; 95% CI, 0.06–0.55; P = .0023; full adjustment) (Figure 1). Moreover, 267F was associated with reduced risk for overall disease progression in a trend test (HCC vs cirrhosis, vs CHB; OR = 0.15; 95% CI, 0.03–0.74; P = .019; with full adjustment OR = 0.15, P = .017; Figure 1). These results suggest that S267F may affect overall HBV disease progression to cirrhosis or HCC. From cirrhosis to HCC there is, however, no additional protection afforded by carriage of S267F.

NTCP S267F and HBV DNA Levels

Finally, we tested whether S267F had an impact on HBV DNA levels. We found that among all HBV infected participants with HBV-DNA data available, HBV DNA levels were slightly lower in 267F carriers (n = 14; mean 4.31 ± 1.84 SD copies/mL) than noncarriers (n = 455; 4.83 ± 1.82 SD copies/mL), but the difference was not significant (P = .29).

DISCUSSION

Recent in vitro studies found that HBV utilizes NTCP, a hepatocyte membrane bile acid uptake protein, as a receptor for hepatocyte entry [8, 10, 11]. An NTCP HBV receptor function-disabling variant, S267F [8], occurs specifically in Asian populations where HBV infection is highly prevalent [18, 19]. In this study, we evaluated the genetic impact of S267F on the full spectrum of clinical outcomes in persons with HBV infection. We found that carriage of 267F confers host resistance to HBV infection and progression to cirrhosis or HCC, with 65% lower risk for HBV infection and 85% lower risk of having cirrhosis. It was not, however, associated with spontaneous HBV clearance, nor with altered risk for HCC from cirrhosis. We observed that the frequency of the protective 267F variant genotypes gradually decreased following unfavorable progression of disease: HBV resistance → chronic hepatitis → cirrhosis and HCC. These findings strongly support a beneficial role of this NTCP-disabling variant and in vivo involvement of NTCP in HBV pathogenesis.

We demonstrated that the NTCP S267F variant was associated with decreased risk for HBV infection and decreased risk for development of cirrhosis. These might be directly caused by the NTCP S267F variant change that leads to loss of NTCP binding to HBV pre-S1 and diminished receptor efficiency for hepatocyte entry by establish HBV infection [9], perhaps by a mechanism of reduced transmission among hepatocytes and decreased inflammation that leads to cirrhosis, or by alteration of other functions of NTCP, such as loss of ability to transport bile salt. Our results are largely consistent with previous studies [15, 16, 22]. Peng et al found that southern Chinese S267F carriers were resistant to HBV infection by comparing CHB with healthy controls [15]. Hu et al suggested that S267F was associated with reduced risk for cirrhosis by comparing patients with cirrhosis or cirrhotic-HCC to those that did not progress to cirrhosis or cirrhotic-HCC in a longitudinal cohort [16]. Our study expands on previous studies in that we included the full spectrum of HBV-associated outcomes with well-defined nonoverlapping phenotypes.

Although we found that the S267F variant was generally protective, it did not prevent progression from cirrhosis to HCC. Whether the NTCP variant directly influences HCC risk is an important question. Hu et al found that S267F was associated with reduced risk for HCC when comparing HCC with CHB [16]. We found a similar significant protective effect when HCC was compared to CHB alone. However, when comparing HCC with the cirrhosis group, the association was nonsignificant, which suggests that patients with cirrhosis have a similar risk for developing HCC regardless of their S267F genotype status. We consider that the latter comparison is more clinically important and practical as most HCCs develop from cirrhosis [22]. A study in a Korean population also did not find a protective role in HCC development [23]. It is biologically plausible that a disabled HBV receptor reduces HBV replication and HBV-induced chronic inflammation, processes that lead to destruction and regeneration of hepatocytes with resultant scarring, increasing the probability of cell division errors that initiate the hepatocarcinogenesis process [24]; whether NTCP is directly involved in oncogenesis remains to be determined.HBV NTCP entry inhibitors are currently under development [25, 26]; for these to be effective, administration in the early stages of HBV infection will be essential.

Our results also indicate that NTCP 267F/F does not provide full protection against HBV infection or progression to clinical outcomes. Among the 2 homozygous 267F/F carriers in our study group, one was HBV infected and the other had no antigen or antibody markers for HBV infection. The 267F/F patient was a 45-year-old male with asymptomatic CHB (positive for HBsAg, anti-HBc, and anti-HBe), normal ALT, and undetectable HBV-DNA. This incomplete protection for the S267F homozygote has been observed in other studies, in which 12%–16% of 267F/F homozygous individuals were infected with HBV [15, 16]. These data indicate the likely presence of other minor receptors for HBV entry or, alternatively, that S267F dramatically reduces but does not completely disable the HBV entry function of NTCP [9]. It is also possible that there are some HBV strains that have evolved mutations that facilitate binding of variant NTCP. This is similar to the situation for the heterozygous null Δ32 deletion in CCR5, the major cell-entry receptor for HIV. Individuals homozygous for Δ32 enjoy near-complete resistance to infection, while heterozygotes for Δ32 progress to AIDS more slowly than noncarriers [27].

We did not find significant association of NTCP S267F, either independently or in interaction with HLA-DPA1, with HBV spontaneous clearance. HLA-DPA1 rs3077 is a major predictor of HBV spontaneous clearance, as previously shown in this cohort (OR = 2.7; P = 3.95 × 10^–11) [4] and in other studies as summarized in a recent metaanalysis [6]. NTCP S267F was also not associated with HBV clearance from studies carried out in Taiwan [16] and Southern Chinese [28]. Lack of interaction between those 2 host genetic factors is most likely due to their distinct roles in HBV infection. NTCP only serves as an entry receptor for infection of hepatocytes, but does not play a role in host immune response against HBV, which is primarily afforded by HLA class II proteins [4, 6]. Similarly, HCV clearance is mainly attributed to products of the IL28B/IFNL4 and HLA genes [29].

We reported a comprehensive evaluation of genetic associations of NTCP S267F variant with clinical outcomes after HBV exposure. A strength of our study is that it was designed specifically to screen for genetic factors that affect each stage of HBV infection, which should minimize the selection bias often confounding convenience samples. However, we were unable to test the recessive model due to the low frequency of the 267F allele in the general population. Nevertheless, the consistency of our results for association of HLA-DPA1 and NTCP further corroborates the influence of these genetic factors on HBV infection [4–6, 15, 16, 30]. Additional variants in NTCP have also been reported in HBV infection with conflicting results [14, 31, 32], and need to be studied in our sample set.

In summary, by assessing all the major outcomes of HBV infection, we found that the NTCP 267F variant was associated with decreased the risk of HBV infection and progression to cirrhosis or liver cancer, but not with clearance; on the other hand, HLA DPA1 is strongly associated with resistance to HBV infection or resolution of HBV infection, but has no association with progression to clinical endpoints. We demonstrated for the first time that the NTCP S267F variant does not interact with HLA-DPA1 rs3077 variant to affect HBV infection and chronic liver disease progression. Our population genetic data support a role for NTCP as the primary entry receptor of HBV in vivo.

Disclaimer. The content of this publication does not necessarily reflect the views or policies of the Department of Health and Human Services, nor does mention of trade names, commercial products, or organizations imply endorsement by the US Government. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

Financial support. This work was supported by the National Cancer Institute, National Institutes of Health (NIH) (contract number HHSN26120080001E); and the Intramural Research Program of the NIH, National Cancer Institute, Center for Cancer Research.

Potential conflicts of interest. All authors: No reported conflicts of interest. All authors have submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Conflicts that the editors consider relevant to the content of the manuscript have been disclosed.