Characterizing key nucleotide polymorphisms of hepatitis C virus-disease associations via mass-spectrometric genotyping

  • Authors:
    • Yuta Horiuchi
    • Jason Lin
    • Yui Shinojima
    • Kyoko Fujiwara
    • Mitsuhiko Moriyama
    • Hiroki Nagase
  • View Affiliations

  • Published online on: November 22, 2017
  • Pages: 441-452
Metrics: Total Views: 0 (Spandidos Publications: | PMC Statistics: )
Total PDF Downloads: 0 (Spandidos Publications: | PMC Statistics: )


As more than 80% of hepatocellular carcinoma patients in Japan also suffer from hepatitis C virus infections some time in their medical history, identifying genetic aberrations associated to hepatitis C virulence in these individuals remains a high priority in the diagnosis and treatment of hepatocellular carcinoma. From the BioBank Japan Project, we acquired 480 subjects of hepatocellular carcinoma, chronic hepatitis and liver cirrhosis, and genotyped 131 clinically relevant host single nucleotide polymorphisms to survey the potential association between certain risk alleles and genes to a patient's predisposition to hepatitis C and liver cancer. Among those polymorphisms, we found 12 candidates with statistical significance to support association with hepatitis C virus susceptibility and genetic predisposition to hepatocellular carcinoma. SNPs in genes such as XPC, FANCA, KDR and BRCA2 also suggested likely connections between hepatitis C virus susceptibility and the contraction of liver diseases. Single nucleotide polymorphisms reported here provided suggestions for genes as biomarkers and elucidated insights briefing the linkage of hepatitis C virulence to the alteration of healthy liver genomic landscape as well as liver disease progression.


The seriousness of hepatitis C virus (HCV) infections lies in their elevation of risks pertaining to various life-threatening liver conditions such as liver cirrhosis (LC), chronic hepatitis (CH) as well as hepatocellular carcinoma (HCC). For instance, the estimated risk of hepatocellular carcinoma has been reported to be often 15 to 20 times higher in subjects infected with HCV (1). Statistics also suggest that 75–85% of people infected will become chronic carriers, and 10–15% of HCV infection cases will advance to cirrhosis within the first 20 years (2), along with increased risks of developing HCC. In Japan, ~80% of patients suffering from HCC is caused by prior or concurrent HCV (3), suggesting potential links between possible genetic predisposition and infection incidence. While genetic aberrations can direct and drive carcinogenesis, exact details, especially with the ethnic Japanese population, linking HCV virulence to HCC susceptibility (and to a lesser extent LC and CH) remain opaque. Preventative treatment options for HCV are also lacking in a similar regard, although a meta-analysis by Colombo and Iavarone had suggested the possibility of reducing HCC risks in a small fraction of cases with interferon administration (4). To-date, while various studies have sought to characterize the effect of genotypic alterations in liver diseases, the role of SNPs in liver diseases still remain poorly understood (511). Additionally, while genes such as IFNγ and IL-28B have been reported to harbor potential associations to HCV infections leading to conditions such as fibrosis and jaundice (1214), no clinical screening and characterization studies have been conducted between HCV and severe liver conditions such as CH, LC and HCC.

With an estimated 143 million people infected with hepatitis C worldwide (15), early genetic screening can have a beneficial role in greatly improving the quality of life for those afflicted with liver diseases. With the establishment of the BioBank Japan Project that stores and maintains a number of annotated liver disease cases, we decide to explore the possibility, if any, that certain genetic risk factors are associated with the prevalence of HCV infections to hepatocellular carcinoma in the Japanese population by MassARRAY genotyping, a MALDI-TOF mass spectrometry-based approach well demonstrated in medium- to large-scale cohort studies (16) to analyze a set of specific polymorphisms. We herein present one of the earliest studies attempting to assess several clinically relevant marker gene and variant candidates linking HCV to liver cancer, as well as CH and LC for the purpose of comparison, in the Japanese population using DNA extracted from blood specimens. For each candidate, we then assessed its significance by comparing genotypic frequencies against records from the Japan Single Nucleotide Polymorphisms databank (JSNP), NCBI dbSNP and the Tohoku Medical Megabank (TMM) datasets (1720) to elucidate their potential roles as risk factors in HCV-induced liver diseases and their pertinence in other ethnic groups.

Material and methods

SNP selection

Among published SNPs and polymorphisms in genes relevant to cancer in recent literature, we generated a preliminary list of SNP candidates for screening. To broaden the scope, this list was later expanded to include a small number of SNPs that had few reports of clinical significance to-date, but were well-conserved, disease-associated and believed to impact protein structure by SNPs3D searches (21). In all, 131 SNPs over 4 primer sets were found to meet the MassARRAY prerequisites and selected for analysis (tabulated data of SNPs and corresponding primers available upon request).

Specimen collection and subject demographics

Blood DNA samples from a collection of CH (200), LC (80) and HCC (200) subjects were obtained from the BioBank Japan Project (22). Hepatitis subjects had a gender makeup of 103/97 male/female, with 102 over the age of 50 (mean 54±15). Samples were confirmed to be HCV-negative and cancer-free. The Cirrhosis test group consisted of 45 males and 35 females, with 71 people over the age of 50 (mean 65±11) and similarly confirmed to be HCV-negative and cancer-free. Among HCC patients of 103 men and 97 women, 194 were over 50 (mean 69±8) and all cases were confirmed to be HCV-positive. No other medical history (including history of hepatitis virus B infection) or personal information was obtained. For the purpose of statistical assessment, a simulated control ('healthy') set based on JSNP (last updated May, 2014) and NCBI data (build 142, October, 2014) was used. In cases where multiple genotypes exist from both JSNP and NCBI, prevailing JSNP results were used for the control set. The TMM 2KJPN database (release June, 2016) was also consulted for comparative analysis of variant allele frequencies for the Japanese population.

SNP detection by single-base extension

DNA specimens (5 µg) were diluted to the stock concentration of 10 ng/µl prior to PCR using the iPLEX Gold reagent kit (Sequenom, San Diego, CA, USA). Each specimen (1 µl) was mixed with the primer mix (500 nM) prior to SNP isolation, and 6/96 samples were then randomly selected to confirm the extent of PCR by gel electrophoresis. Four primer mixes, each containing 33 SNPs, were prepared to the same final composition per Sequenom's instructions. PCR products underwent shrimp alkaline phosphatase (0.5 unit) treatment to dephosphorylate unincorporated deoxynucleotides at 37°C for 40 min followed by 85°C for 5 min. Extension reactions were performed following the iPLEX-Extend protocol by preparing each sample in a solution of 0.62 µl water, 0.2 µl 10X iPLEX buffer plus, 0.2 µl iPLEX termination mix, 0.94 µl extend primer mix, 0.04 µl iPLEX enzyme for reaction as follows: 94°C (step 1, 30 sec), 94°C (step 2, 5 sec), 52°C (step 3, 5 sec), 80°C (step 4, 5 sec), in which steps III and IV were repeated for 5 cycles followed by 40 cycles of steps II–IV before the final step at 72°C for 3 min. Following instructions from the SpectroChip Chip and Resin kit (Sequenom), samples were characterized by mass spectrometric analysis on a MassARRAY compact MALDI-TOF mass spectrometric analyzer. Oligonucleotides were binned with replicates per manufacturer's recommendations to minimize calling failures. Data were acquired and processed using EpiTYPER 1.0 in the Sequenom MassARRAY Workstation suite. For illustrative purposes, some data and cluster plot/spectrum snapshots were exported to XML and tab-delimited text files, respectively, for replotting and annotation with R 3.2 (available

Statistical analysis

Statistical analyses were performed using SPSS 15.0 (IBM Corporation, Armonk, NY, USA) and R. Allele and genotypic differences between healthy (see Specimen collection) and diseased subjects as well as ethnicity were evaluated by the χ2 test of significance. Ancestral/reference and alternative allele were reordered for presentation according to dbSNP conventions; in certain cases where the ancestral allele was ambiguous, e.g. PMS2 (rs1805321), the statistically dominant allele in dbSNP was selected as the ancestral allele (e.g. C for rs1805321). Comparisons to the TMM dataset were determined in R, but not used to determine significance as only allele frequencies were available. A predefined α-level of 0.05 was used for statistical assessment. Per the exploratory and targeted nature of this study, P-values were not corrected for multiple comparisons; instead, the authors advised readers to use corrected significance levels, for instance the Dunn-Šidák corrected α′=0.00039 (m=131) when cross-examining results here with other non-targeted genotyping association studies to account for potential consequences of multiple hypothesis testing.

Changes to local protein structure

As specimens obtained for this study came from a Japanese population, we consulted annotated records from TMM to determine potential changes to local protein structures. For non-synonymous SNPs meeting the significance level, such changes were predicted by evaluating the impact of amino acid substitutions by a composite probability from HumVar PolyPhen-2 (2.2.2r398) and Grantham scores (23,24). PolyPhen2 FPR thresholds were appraised at 10/20%. For the purpose of calculation, the Grantham value of all non-identical amino acid replacements were transformed to a kernel density function in R and re-evaluated as cumulative probabilities. Scores were averaged by the geometric mean (square root of the product of two probabilities).

Interaction and gene enrichment analysis

Potential interactions among the list of significant candidate genes were analyzed by GeneMANIA to identify the top 25 related genes. Interactions were selected to allow at most 25 genes with 25 attributes, and filtered by the presence of possible associations via co-expression, co-localization, genetic and physical interactions, pathway information as well as predicted interactions. Networks were analyzed and visualized in Cytoscape. Statistically enriched Reactome pathways (version 58) were identified by the Panther overrepresentation test with a predefined α-level of 0.05 to assess individual pathway significance.

Linkage disequilibrium analysis

Potential linkage disequilibria were examined using LDproxy in the LDlink suite (25) to possible linkages among the significant SNPs, using the Japanese population genotype data from the Phase 3 of the 1000 Genomes Project as input. Proxy locus pairs with correlation coefficients (R2) >0.80, and RegulomeDB scores of 1-5 (including subgroups of 1–3) were retained for further analysis. Variant pair associations, as a function of the correlation coefficient between loci pairs, were visually examined in Cytoscape as network diagrams to identify common functionally active proxy variants.

Cohort allele frequency dilution

VCF files containing individual variant data information from the Phase 3 1000 Genomes Project (26) were acquired via the Data Slicer (GRCh37 release 89, last accessed August, 2017) for each of the 12 candidate polymorphisms. To test the effect on changes in candidate allele frequencies if samples from this population were mixed with the cohort. n=0–480 randomly selected genotypes from G1000 were mixed with the cohort, also randomly selected to make up a final sample size of 480, in order to examine the resultant reference (A) and alternative (B) allele counts as well as the mean allele frequencies (%B) as a function of n. In a separate validation, both the cohort and G1000 populations were randomly sampled (100 individuals/set) to evaluate the likelihood that both populations contained different mean allele frequencies. Hundred random trials were performed for both validations in R, with statistical significance by two-sample t-tests.


We found 12 out of the 131 SNPs investigated (Table I) to be potentially associated; among those, 10 SNPs, 1, [α-methylacyl-CoA racemase (AMACR) rs34677], 2 (ARHGAP8, rs2071762), 3 (KDR, rs2305948), 4 (SELE, rs5361), 5 (XPC, rs2228001), 6 (FANCA, rs2239359), 8 (BRCA2, rs766173), 9 (PCDHGB7, rs17208397), 11 (PMS2, rs1805321) and 12 (ADAMTS16, rs16875054) were non-synonymous; SNP 7 (LRRN3, rs17439799) was intronic and 10 (ERCC4, rs2020955) was synonymous. A full table containing genotypes, alleles and call rates for the 131 selected SNPs are available upon request and at Discrepancies did exist in allele distribution between database annotations, e.g. WRN (rs1346044) contained T/C alleles in NCBI but A/G in JSNP.

Table I

Genotypes, alleles and impact on local protein structure of statistically significant SNPs and corresponding JSNP and dbSNP entries.

Table I

Genotypes, alleles and impact on local protein structure of statistically significant SNPs and corresponding JSNP and dbSNP entries.

1AMACR (34677)T/GHCC42214917530320GT41.72530.0000689319188.18110.0167C/A (3717/379)0.17930.6720Q>H0.320240.0470.123
2ARHGAP8 (2071762)T/CHCC387378189149229N/S69271108C/T (2456/1580)P>L0.018980.5210.097
3KDR (2305948)T/CHCC02515818325341N/S6.23510.0443C/T (3662/436)5.26320.0218V>I0.999290.0790.281
4SELE (5361)A/CHCC180201823622AL8.02620.0046T/G (3953/145)9.37380.0022S>R0.9441100.6370.775
5XPC (2228001)A/CHCC908015185260110GT13.11230.0014G/T (1694/2398)19.18130.0000Q>K0.000530.1790.000
6FANCA (2239359)T/CHCC10549816225965GT7.07760.0290C/T (652/3396)3.42260.0643G>S0.000560.2000.000
7LRRN3 (17439799)T/CHCC120551018529575GT6.69970.0351T/C (3113/981)2.56060.1096Int
8BRCA2 (766173)T/GHCC14047118832749GT10.57510.005166862596
A/C (3569/521)0.02660.8703N>H0.034680.2790.097
9PCDHGB7 (17208397)G/CHCC152601583106AL5.08170.024224521861
G/C (3955/129)1.56560.2109V>L0.261320.0890.153
10ERCC4 (1799801)T/CHCC120611119230183GT7.20550.027252072800
T/C (3060/1032)2.43960.1183Syn
11PMS2 (1805321)T/CHCC13507413776198GT56.20440.00005259684N/AP>S0.001740.2950.017
12ADAMTS16 (16875054)A/GHCC91110512529221GT16.34290.000348414945
G/A (3559/537)0.55310.4570A>T0.590580.2110.352

[i] Maximum χ2 statistics against either curated JSNP or dbSNP genotype or allele counts and corresponding P-values. ID, SNP index to be used throughout this article. ST, significance type (statistical significance against control genotype 'GT' or allele 'AL' frequencies, respectively); A/B, A and B alleles; DP, disease phenotype; Mut, mutation type (Int, intron; Syn, synonymous); PD, PolyPhen2 HumDiv score; PV, PolyPhen2 HumVar scores; GV, Grantham value; PG, Grantham probability of impact from amino acid substitution; GMI, geometric mean impact on protein structure.

We also evaluated the likelihood that these variants differed from corresponding records in the TMM databank by their maximum χ2 statistics between the JSNP and dbSNP datasets as a potential screen to confirm the ethnic specificity of those variants to the Japanese population, but did not assign significance based on these statistics as the TMM databank contained considerably less publicly available information than JSNP or dbSNP at the time of analysis. Most of the variants appeared to have allele frequencies that varied across different phenotypes, some with only one subgroup that strongly deviated from the other two, e.g. 3 in only HCC, and others in multiple subgroups, e.g. 4 and 8. Generally, the HCC subgroup frequently exhibited allele frequencies that deviated from the other two phenotypes, for example candidates 5 and 6. For SNPs with significant differences in genotype and allele frequency, all but 1 (AMACR) expressed differential compositions across different ethnic groups (Table II), suggesting that they may be specific to the Japanese population.

Table II

Stratified comparisons of significant SNPs across different ethnic groups.

Table II

Stratified comparisons of significant SNPs across different ethnic groups.

CytobandRs# SymbolNCBI SSIDPopulationGroupsP-value
5p13.2rs3467712675253JPT, HCB, CEU, YRIG10.0960.089
22q13.31rs207176212527661JPT, HCB, CEU, YRIG1****
4q12rs230594848430022JPT, HCB, CEU, YRIG1****
1q24.2rs536148428825JPT, HCB, CEU, YRIG10.001**
3p25.1rs222800144411054JPT, HCB, CEU, YRIG10.0090.004
XPC66856826JPT, HSP, CEU, AAM,G20.0010.002
68853585JPT, HCB, CEU, YRIG10.0090.007
16q24.3rs223935916702260JPT, HCB, CEU, YRIG1****
FANCA43872290AoD Japanese, AfricanG3N/A**
American, Caucasian, Chinese
66860952JPT, HSP, CEU, AAM,G2****
69355410JPT, HCB, CEU, YRIG1****
7q31.1rs1743979924455640JPT, HCB, CEU, YRIG1****
13q13.1rs7661735586314JPT, HCB, CEU, YRIG1****
BRCA266862596JPT, HSP, CEU, AAM,G2****
69130493JPT, HCB, CEU, YRIG1****
5q31.3rs1720839724621861JPT, HCB, CEU, YRIG1****
16p31.12rs179980152072800JPT, HCB, CEU, YRIG1****
ERCC466861998JPT, HSP, CEU, AAM,G2****
5p15.32rs1687505448414945JPT, HCB, CEU, YRIG1****

{ label (or @symbol) needed for fn[@id='tfn2-ijo-52-02-0441'] } Reference datasets from NCBI dbSNP with specified SSIDs. G1, Asian/European/Sub-Saharan African; G2, Asian/European/African American/Sub-Saharan African; G3, Asian/Caucasian/African American. HapMap groups: JPT, Japanese (Tokyo, Japan); HCB, Han Chinese (Beijing, China); CEU, Utah residents with ancestry from northern and western Europe; YRI, Yoruba (Ibadan, Nigeria)

** P<0.001; N/A, data not available. GT, genotype; AL, allele.

We also explored whether potential linkage disequilibria existed among candidates in Table I to ascertain the possibility of non-random association with other alleles as means to confirm their true phenotypic implications on the liver diseases. While some of those SNPs did have proxy pairs that would suggest some extent of linkage disequilibria (available upon request), we observed no common proxy variants among them (Fig. 1B and C). Annotations of those proxy variants also suggested either synonymous mutations or the lack of clear biological function. As not all variants in Table I were on the same chromosome, and most linked proxy pairs were outside the immediate proximity (only 12/170 were within 1,000 bp; data available upon request), potential associations on allele frequencies of SNPs in Table I by those proxy variants most likely had no effect on disease outcomes.


Roughly 10% of the SNPs genotyped in this study displayed critical roles in liver functions and various cancer phenotypes. For instance, AMACR is an ubiquitously expressed enzyme responsible for bile branched-chain fatty acid metabolism (27) that undergoes increases in mRNA and protein expressions in colorectal and prostate cancers as well as pre-cancerous hyperplastic polyps in the large intestine (28,29). Co-increases of COX-2 expression and AMACR could also potentially lead to immune suppression, neoplastic changes and tumor invasion (30). As polymorphisms in AMACR have been implicated in prostate and colorectal tumorigenesis (28), its possible contribution in HCV infection-induced disease phenotypes shall not be overlooked. Similarly, despite the lack of prior evidence suggesting direct association of ARHGAP8 to HCV-induced HCC, the loss of Rho-suppressing ARHGAP7 has been implicated in multiple cancers including HCC (31). Regulatory features of RhoGAP via Rho has been said to attribute to reductions in hepatic fibrosis, suggesting that a loss of function variant may induce liver cirrhosis in a manner independent from HCV infection. Adding to the ability for statins to attenuate liver fibrosis in chronic HCV infections (32), inter-regulatory feedback of Rho acylation by members of the ARHGAP family, for instance ARHGAP8, may well attribute to the aberrant dysregulation, via HCV infections, during HCC development (33).

While mutations in DNA mismatch or strand-break repairs have always been intimately associated with tumor development, the presence of risk alleles in XPC, a gene critical in nucleotide excision repair (34), further highlights the importance of exogenous genomic editing by HCV in HCC. XPC has been said to elevate the rise of skin cancer 1,000-fold and escalating organ neoplasms as much as 10-fold (35) through its role in xeroderma pigmentosum, a pre-cancerous condition. From this, it is likely that defects in XPC may also increase one's susceptibility to HCC. Additionally, aberrations in FANCA, a known tumor suppressor (36) as well as BRCA2, a DNA repair gene, can have serious repercussions leading to oncogenesis, most notably breast cancer, both in Japanese and other ethnic cohorts (3740). Among these reports, candidate 8 also belongs to the group of 25 breast cancer potential risk alleles, sufficiently justifying the need for further investigation into its mechanistic role in HCV-induced HCC. Likewise, mutations in PMS2 may be similarly implicated in response to HCV virulence as the gene was most remarkably known for its role in DNA mismatch repair and links to microsatellite instability as in hereditary nonpolyposis colorectal cancer (4143).

Our genotyping results also revealed some curiosities on the role of functional surrogates in HCV-induced oncogenesis, particularly through potential risk alleles in ADAMTS16. ADAMTS16 belongs to a family of secreted metalloproteases that tends to be fairly tissue limited, and the discovery of its association to HCV was unexpected as ADAMTS16 was primarily expressed only in the lung and brain; with that said, recent literature did suggest the possibility of high-quantity shuttling of ADAMTS13 and ADAMTS19 to unexpected destinations, such as tumors and nearby tissues in cases of osteosarcoma, melanoma and colon cancer (44). Surfacing of ADAMTS16 in HCC samples could also infer the occurrence of a similar phenomenon by mutations or other mechanisms. Interestingly, while some in the ADAMTS family were implicated in diseases such as angio-inhibition (45) and Ehlers-Danlos syndrome (46), most remain relatively unknown and inconclusive (44,47,48). As such, risk alleles in ADAMTS16 as well as the observation that these metalloproteases were typically downregulated in cancer as a consequence of promoter hypermethylation (49), would suggest that polymorphisms in ADAMTS16 may potentially push the gene to displace and redistribute itself and other members in the family, possibly via extracellular matrix remodeling (50), and become oncogenic. Additionally, not all conditions were genetically predisposed to the same polymorphisms (Table III) or reacted identically to the virulence of HCV. For example, a mutation to 3 would more likely drive onset in HCC over LC or CH, while subjects with a mutation to 6 would perhaps present HCC phenotypes differently. Odds ratios (OR >1, P<0.05) also inferred differential susceptibility for subjects with mutations in XPC and BRCA2, as XPC showed elevated odds for CH and BRCA for all subgroups. Just as the BRCA genes helped physicians to assess breast cancer susceptibility, the combination of BRCA and perhaps a combination of several aforementioned SNPs could also help elucidate risks to HCC by the process of elimination.

Table III

Stratified genotypic comparisons of significant SNPs among disease phenotypes.

Table III

Stratified genotypic comparisons of significant SNPs among disease phenotypes.

ID Symbol dbSNP rsIDAllelesGrouping (or JSNP/NCBI set ID)Genotype
All P-valueHCC P-valueLC P-valueCH P-value
NCBI 126752530.0320.1060.1470.002
NCBI 68931918********
JSNP IMS-JST038996**0.0170.0020.003
NCBI 125276610.1110.9880.8180.053
NCBI 692711080.0650.8620.6590.014
NCBI 716479510.1110.9880.8180.053
JSNP IMS-JST0074390.0120.0560.1020.029
NCBI 484300220.6470.3740.4380.603
NCBI 688991720.6510.4010.4610.599
JSNP IMS-JST0634100.1580.0440.3020.389
NCBI 484288250.1880.0910.4130.091
NCBI 687843050.2010.0950.4240.095
NCBI 444110540.0110.5230.0920.656
NCBI 668568260.010.3650.1080.847
NCBI 688535850.0120.3260.1850.555
JSNP IMS-JST0867940.0030.0010.3710.165
NCBI 167022600.1690.3660.1910.581
NCBI 43872290
NCBI 668609520.1690.3660.1910.581
NCBI 693554100.1690.3660.1910.581
JSNP IMS-JST0104820.0870.0290.5040.488
NCBI 24455640**0.845**0.926
NCBI 55863140.0190.008**0.01
NCBI 668625960.0130.005**0.006
NCBI 691304930.0130.005 <0.0010.003
NCBI 246218610.0040.0260.040.002
NCBI 689515340.0040.0260.040.002
NCBI 520728000.1560.0270.2370.07
NCBI 668619980.1560.0270.2370.07
NCBI 5259684********
NCBI 48414945****0.0740.007
NCBI 66315014****0.0740.007
NCBI 68925406****0.0740.007

{ label (or @symbol) needed for fn[@id='tfn3-ijo-52-02-0441'] } The full table including genotypic frequencies, allele frequencies, dominant (AA+AB/BB) and recessive allele models (AA/AB+BB), is available as Table S4 upon request. P, P-values of significance of a particular subgroup compared against all three disease phenotypes (All) or individual subgroups. Bold values indicate genes with call rate >80%.

** , P<0.001; –, not available.

In some situations nearby variants may act in concert and contribute jointly to particular phenotypes; to examine the possibility of these nonrandom associations, we next explored possible interactions and the presence of common functionally active proxy variants associated with polymorphisms identified in this study. While the presence of a small interaction revealed the presence of common interacting partners with the 10 candidate genes, for instance SMARCA4, PPP2CA and XPO1 (Fig. 1A), linkage disequilibrium analysis painted a slightly different picture, alternatively suggesting that these candidates were genotypically pairwise independent, and contributed cumulatively to the outcome of each disease. Additionally, multivariate analysis across genders and ages revealed no strong covariances among the candidates (data not shown). Interestingly, enrichment analysis highlighted strong connections to DNA repair and post-translational modifications, two biologically orthogonal functions, for the candidate genes and their interacting partners (table available upon request). Enrichment in the two nonparallel pathways would thus posit an overall additive effect that reflected the candidates' genotypic independence. While defects in DNA repair were commonplace in cancer as a consequence of unregulated division, no mechanistically direct explanation could answer the functional enrichment in protein modification as a consequence; as such, the hypothesis that both pathways, via functional alteration by the candidates genotyped, acted in concert to achieve this additive effect reflected in the association of HCV to HCC, appeared to be a plausible one and hopefully could be further pondered.

We also analyzed changes in protein local secondary and quaternary structures in order to attribute the impact of non-synonymous mutations to the connection between HCV and HCC to ascertain the underlying routes to disease progression. Although DNA repair mechanisms may be biochemically similar across different types of cancer, focused studies on BRCA2 and FANCA still remain relatively unexplored in HCV-induced HCC. Especially since hepatic tissues are constantly under oxidative stress, genome repair during acute and chronic HCV infection and the oncogenesis of HCC are critical in understanding disease progression. We observed that ADAMTS16, SELE and KDR all carried polymorphisms with potentially damaging mutations; as a matter of fact, both ADAMTS16 and SELE were associated with HCC and CH, suggesting a possible path of disease progression upon HCV infection from CH to HCC via these mutations (Table I). Nonetheless, the fact that KDR also manifested a damaging mutation and was associated with HCC would also suggest a possible role for the gene, as KDR-mediated mechanisms could support HCV infection (51), subsequently directing cells to the carcinogenesis of HCC.

A difficulty in establishing genetic connections between diseases and HCV infections is that the virus can remain dormant in the liver, leaving the patient with normal liver functions and no symptoms. As such, disease progression in early stages becomes retarded. These associated SNPs can, in theory, display no phenotypical effect during early onset, which can delay diagnosis and treatment altogether. As both acute and chronic HCV infections could transform healthy livers, the full and exact extent of the infection would be difficult to gauge. HCV carriers may present normal ALT levels from normal liver panel screenings and the amount of fat deposits in the tissue may also mask the presence of HCVs, liver biopsies are likely to be inconclusive in these cases for determining the stage of infection. As such, extended genotyping studies beyond the one described herein, alongside functional characterizations of BRCA2 and FANCA, may be a practical option clinically for diagnosis.

Among the different genotyping technologies, MassARRAY proved to be a capable choice, demonstrating high sensitivities in various applications such as human papillomavirus genotyping for cervical cancer (52) or classification of ectopic crypts in colorectal cancers (53) via KRAS and BRAF mutations. Although we were unable to call every SNP on the initial test, we nonetheless obtained call rates surpassing 90% in duplicate typing of most SNPs, suggesting reasonably minimal genotyping errors (Fig. 2A and B). Mass spectrometry-based genotyping offers excellent sensitivity in analyte detection and is thus an ideal approach for characterizing specimens from tissue depositories, where amounts of DNA may be very limited. Extension (1-bp) coupled with mass spectrometric-based characterization by MassARRAY was a suitable choice in assessing candidate SNPs for studies of similar scale. We do, however, recognize that such a method is limited to the number of SNP candidates determined a priori, so the discovery of novel SNPs can be more challenging. Nonetheless, for cancers and precancerous conditions dependent on a small of oncogenic drivers, this method is a sensible and attractive option at the clinical setting.

The MassARRAY method allowed us to maintain a reasonable subject size of 480, although the relatively lower number of LC cases might have marginally reduced the statistical power in this study. While increasing the number of test subjects could certainly be beneficial in improving the statistical power, such an approach was not always possible in practice. In consideration for the potential loss in statistical power, we performed two validation simulations to see whether these differences in allele frequencies were still preserved if subsets of our cohort were 'diluted' with a different population sample. By taking genotypes from randomly selected G1000 individuals and mixing them with (also randomly selected) subjects in the cohort, we assembled subsets of 480 individuals and checked the frequencies of the candidate risk allele. Plotting such changes as a function of different numbers of G1000 individuals (Fig. 2C) revealed moderate frequency shifts for all of the candidate polymorphisms. Frequencies of these risk alleles thus appeared to be condition-specific, as they were certainly susceptible to dilution with subjects from a different and considerably more heterogeneous population of various ethnicities and health states. Additionally, bootstrapping our cohorts to smaller subsets of 100 individuals for direct comparison to sets of equal sizes in the G1000 population also confirmed differences in allele frequencies (Fig. 2D). While such tests could not truly offset benefits of increasing cohort sizes, results nonetheless demonstrated a reasonable level of statistical robustness considering the size of our cohort.

Genotyping more candidates, e.g. Tolloid-like 1 protein (54), from overlooked biochemical pathways (for instance, stress responses, inflammation and cell cycles) could allow us to further understand the manner in which HCV infections evolve phenotypically. Characterizing changes in known and nearby polymorphisms in genes implicated in insulin biosynthesis, inflammation and adipose tissue remodeling, etc., would be highly useful in deciphering risk factors of obesity to HCV-induced HCC; in a similar vein, alcohol metabolism and various other pathways related to opioid metabolism would also be highly useful in risk assessment of substance abuse to HCC. Additionally, this type of candidate expansion would be particularly more beneficial in analyzing ethnically more heterogeneous populations. As most candidate SNPs, with the exception of AMACR (rs34677), exhibited pairwise independence across non-ethnically Japanese datasets, our results hinted that HCV-induced liver disease progressed differently from other infection-driven oncogenesis such as the case with Helicobacter pylori and stomach adenocarcinoma. Based on the differences across different datasets, we would recommend increasing the number of candidates for genotyping to validate this hypothesis. Nonetheless, our approach here could still serve as a useful tool for early diagnosis based on liver disease examination.

An ongoing discussion in liver cancer is the possibility of oncogenic addiction, in which cancer progression is often controlled by a handful of driver genes subsequently leading to uncontrollable growth and eventual transformation to tumor. Horizontally integrating SNP-based oncogenesis information from one cancer type to another can be a useful approach in deciphering the puzzle. For liver diseases that may ultimately lead to cancer, identifying function-altering SNPs is a useful tool for facilitating earlier diagnosis and expediting treatment. Furthermore, for cancers which the theory of oncogenic addiction well applies, the ability to identify driver mutations quickly also has significant therapeutic implications. Based on our previous knowledge of SNPs in liver diseases and oncogenes, we utilized a targeted SNP-screening approach using a MassARRAY-based protocol from a set of 131 SNPs, and were able to identify 12 positions with implications ranging from disease onset, survival rates and in other metrics. Additionally, we were also able to highlight the significance of BRCA2 and FANCA germ-line mutations and associate them to liver cancers. Most of these HCV-dependent mutation candidates were non-synonymous, and their association with other risk factors further suggested their potential roles as biomarkers for the liver conditions described in this study.

It is however important to note that several factors such as age have not been well explored in this study, as illuminated by the relatively smaller sample sizes of subjects under 50 years old in all three disease groups; additionally, at 80 samples, the statistical power for SNPs associated to cirrhosis may also be lower than the other two groups. Other factors such as hepatitis B infection, substance use or obesity could also influence disease progression. It is now common knowledge that sharing syringe needles is a common way that HCV is transmitted between persons, and that alcohol abuse has been said to worsen chronic HCV progression (55). Obesity, mostly through nonalcoholic fatty liver disease and type 2 diabetes, can also have an effect on the outcome of liver cancer (56). In the Japanese population, hepatitis B infection attributed to ~16% of HCC cases (2), a fraction ~1/5 of HCV infections; while relatively minor compared to HCV, hepatitis B infections still present a great public health concern and future genotyping studies should also consider the inclusion of genes potentially implicated in these infections. Along this note, it is also important to note that the status of HCV infections such as virus titers, genotypes and history of prior treatments may also affect the progression of liver cancer and thus critical to be taken into consideration in the cohort selection. Nonetheless, the incorporation of in silico screening and whole-exome sequencing data to identify and refine the list of SNP candidates for future MassARRAY studies should provide definitive improvements for more reliable characterization of disease associations and medium-size clinical cohort studies.


This study was supported by the Academic Frontier Project for Private Universities of 2006, Grant-in-Aid for Scientific Research (B) JP26290060 and Research on Innovative Areas (JP26112517, JP25134718 and JP16H01579) whilst being conducted as part of the BioBank Japan Project sponsored by the Ministry of Education, Culture, Sports, Science and Technology as well as the Japan Agency for Medical Research and Development.



hepatocellular carcinoma


single nucleotide polymorphisms


hepatitis C virus


chronic hepatitis


liver cirrhosis


polymerase chain reaction


deoxyribose nucleic acid


Japanese Single Nucleotide Polymorphisms databank


Tohoku Medical Megabank


α-methylacyl-CoA racemase


Rho GTPase-activating proteins


a disintegrin and metalloproteinase with thrombospondin motifs


the 1000 Genomes Project



El-Serag HB: Hepatocellular carcinoma. N Engl J Med. 365:1118–1127. 2011. View Article : Google Scholar : PubMed/NCBI


Chen SL and Morgan TR: The natural history of hepatitis C virus (HCV) infection. Int J Med Sci. 3:47–52. 2006. View Article : Google Scholar : PubMed/NCBI


Yoshizawa H: Hepatocellular carcinoma associated with hepatitis C virus infection in Japan: Projection to other countries in the foreseeable future. Oncology. 62(Suppl 1): 8–17. 2002. View Article : Google Scholar : PubMed/NCBI


Colombo M and Iavarone M: Role of antiviral treatment for HCC prevention. Best Pract Res Clin Gastroenterol. 28:771–781. 2014. View Article : Google Scholar : PubMed/NCBI


Alavian SM and Haghbin H: Relative importance of hepatitis B and C viruses in hepatocellular carcinoma in EMRO countries and the Middle East: A systematic review. Hepat Mon. 16:e351062016. View Article : Google Scholar : PubMed/NCBI


Tang L, Marcell L and Kottilil S: Systemic manifestations of hepatitis C infection. Infect Agent Cancer. 11:292016. View Article : Google Scholar : PubMed/NCBI


Kato N, Ji G, Wang Y, Baba M, Hoshida Y, Otsuka M, Taniguchi H, Moriyama M, Dharel N, Goto T, et al: Large-scale search of single nucleotide polymorphisms for hepatocellular carcinoma susceptibility genes in patients with hepatitis C. Hepatology. 42:846–853. 2005. View Article : Google Scholar : PubMed/NCBI


Gu X, Qi P, Zhou F, Ji Q, Wang H, Dou T, Zhao Y and Gao C: An intronic polymorphism in the corticotropin-releasing hormone receptor 2 gene increases susceptibility to HBV-related hepatocellular carcinoma in Chinese population. Hum Genet. 127:75–81. 2010. View Article : Google Scholar


Segat L, Milanese M, Pirulli D, Trevisiol C, Lupo F, Salizzoni M, Amoroso A and Crovella S: Secreted protein acidic and rich in cysteine (SPARC) gene polymorphism association with hepatocellular carcinoma in Italian patients. J Gastroenterol Hepatol. 24:1840–1846. 2009. View Article : Google Scholar : PubMed/NCBI


Tang KS, Lee CM, Teng HC, Huang MJ and Huang CS: UDP-glucuronosyltransferase 1A7 polymorphisms are associated with liver cirrhosis. Biochem Biophys Res Commun. 366:643–648. 2008. View Article : Google Scholar


Kim YS, Cheong JY, Cho SW, Lee KM, Hwang JC, Oh B, Kim K, Lee JA, Park BL, Cheong HS, et al: A functional SNP of the Interleukin-18 gene is associated with the presence of hepatocellular carcinoma in hepatitis B virus-infected patients. Dig Dis Sci. 54:2722–2728. 2009. View Article : Google Scholar : PubMed/NCBI


Tanaka Y, Nishida N, Sugiyama M, Kurosaki M, Matsuura K, Sakamoto N, Nakagawa M, Korenaga M, Hino K, Hige S, et al: Genome-wide association of IL28B with response to pegylated interferon-alpha and ribavirin therapy for chronic hepatitis C. Nat Genet. 41:1105–1109. 2009. View Article : Google Scholar : PubMed/NCBI


Nalpas B, Lavialle-Meziani R, Plancoulaine S, Jouanguy E, Nalpas A, Munteanu M, Charlotte F, Ranque B, Patin E, Heath S, et al: Interferon gamma receptor 2 gene variants are associated with liver fibrosis in patients with chronic hepatitis C infection. Gut. 59:1120–1126. 2010. View Article : Google Scholar : PubMed/NCBI


Tillmann HL, Thompson AJ, Patel K, Wiese M, Tenckhoff H, Nischalke HD, Lokhnygina Y, Kullig U, Göbel U, Capka E, et al: German Anti-D Study Group: A polymorphism near IL28B is associated with spontaneous clearance of acute hepatitis C virus and jaundice. Gastroenterology. 139:1586–1592. 1592.e12010. View Article : Google Scholar


GBD 2015 Disease and Injury Incidence and Prevalence Collaborators: Global, regional, and national incidence, prevalence and years lived with disability for 310 disease and injuries, 1990–2015: a systematic analysis for the Global Burden of Disease Study 2015. Lancet. 388:1545–1602. 2016. View Article : Google Scholar


Clendenen TV, Rendleman J, Ge W, Koenig KL, Wirgin I, Currie D, Shore RE, Kirchhoff T and Zeleniuch-Jacquotte A: Genotyping of single nucleotide polymorphisms in DNA isolated from serum using sequenom MassARRAY technology. PLoS One. 10:e01359432015. View Article : Google Scholar : PubMed/NCBI


Hirakawa M, Tanaka T, Hashimoto Y, Kuroda M, Takagi T and Nakamura Y: JSNP: A database of common gene variations in the Japanese population. Nucleic Acids Res. 30:158–162. 2002. View Article : Google Scholar :


Haga H, Yamada R, Ohnishi Y, Nakamura Y and Tanaka T: Gene-based SNP discovery as part of the Japanese Millennium Genome Project: Identification of 190,562 genetic variations in the human genome Single-nucleotide polymorphism. J Hum Genet. 47:605–610. 2002. View Article : Google Scholar


Sherry ST, Ward MH, Kholodov M, Baker J, Phan L, Smigielski EM and Sirotkin K: dbSNP: The NCBI database of genetic variation. Nucleic Acids Res. 29:308–311. 2001. View Article : Google Scholar :


Kuriyama S, Yaegashi N, Nagami F, Arai T, Kawaguchi Y, Osumi N, Sakaida M, Suzuki Y, Nakayama K, Hashizume H, et al: The Tohoku Medical Megabank Project: Design and Mission. J Epidemiol. 26:493–511. 2016. View Article : Google Scholar : PubMed/NCBI


Yue P, Melamud E and Moult J: SNPs3D: Candidate gene and SNP selection for association studies. BMC Bioinformatics. 7:1662006. View Article : Google Scholar : PubMed/NCBI


Nakamura Y: The BioBank Japan Project. Clin Adv Hematol Oncol. 5:696–697. 2007.PubMed/NCBI


Adzhubei IA, Schmidt S, Peshkin L, Ramensky VE, Gerasimova A, Bork P, Kondrashov AS and Sunyaev SR: A method and server for predicting damaging missense mutations. Nat Methods. 7:248–249. 2010. View Article : Google Scholar : PubMed/NCBI


Grantham R: Amino acid difference formula to help explain protein evolution. Science. 185:862–864. 1974. View Article : Google Scholar : PubMed/NCBI


Machiela MJ and Chanock SJ: LDlink: A web-based application for exploring population-specific haplotype structure and linking correlated alleles of possible functional variants. Bioinformatics. 31:3555–3557. 2015. View Article : Google Scholar : PubMed/NCBI


Auton A, Brooks LD, Durbin RM, Garrison EP, Kang HM, Korbel JO, Marchini JL, McCarthy S, McVean GA and Abecasis GR; 1000 Genomes Project Consortium: A global reference for human genetic variation. Nature. 526:68–74. 2015. View Article : Google Scholar : PubMed/NCBI


Levin AM, Zuhlke KA, Ray AM, Cooney KA and Douglas JA: Sequence variation in alpha-methylacyl-CoA racemase and risk of early-onset and familial prostate cancer. Prostate. 67:1507–1513. 2007. View Article : Google Scholar : PubMed/NCBI


Daugherty SE, Platz EA, Shugart YY, Fallin MD, Isaacs WB, Chatterjee N, Welch R, Huang WY and Hayes RB: Variants in the alpha-Methylacyl-CoA racemase gene and the association with advanced distal colorectal adenoma. Cancer Epidemiol Biomarkers Prev. 16:1536–1542. 2007. View Article : Google Scholar : PubMed/NCBI


FitzGerald LM, Thomson R, Polanowski A, Patterson B, McKay JD, Stankovich J and Dickinson JL: Sequence variants of alpha-methylacyl-CoA racemase are associated with prostate cancer risk: A replication study in an ethnically homogeneous population. Prostate. 68:1373–1379. 2008. View Article : Google Scholar : PubMed/NCBI


Pruthi RS, Derksen E and Gaston K: Cyclooxygenase-2 as a potential target in the prevention and treatment of genitourinary tumors: A review. J Urol. 169:2352–2359. 2003. View Article : Google Scholar : PubMed/NCBI


Ng IO, Liang ZD, Cao L and Lee TK: DLC-1 is deleted in primary hepatocellular carcinoma and exerts inhibitory effects on the proliferation of hepatoma cell lines with deleted DLC-1. Cancer Res. 60:6581–6584. 2000.PubMed/NCBI


Trebicka J and Schierwagen R: Statins, Rho GTPases and KLF2: New mechanistic insight into liver fibrosis and portal hypertension. Gut. 64:1349–1350. 2015. View Article : Google Scholar : PubMed/NCBI


Khan FS, Ali I, Afridi UK, Ishtiaq M and Mehmood R: Epigenetic mechanisms regulating the development of hepatocellular carcinoma and their promise for therapeutics. Hepatol Int. 11:45–53. 2017. View Article : Google Scholar


Haiman CA, Hsu C, de Bakker PI, Frasco M, Sheng X, Van Den Berg D, Casagrande JT, Kolonel LN, Le Marchand L, Hankinson SE, et al: Comprehensive association testing of common genetic variation in DNA repair pathway genes in relationship with breast cancer risk in multiple populations. Hum Mol Genet. 17:825–834. 2008. View Article : Google Scholar


de Boer J and Hoeijmakers JH: Nucleotide excision repair and human syndromes. Carcinogenesis. 21:453–460. 2000. View Article : Google Scholar : PubMed/NCBI


Krupa R, Sliwinski T, Morawiec Z, Pawlowska E, Zadrozny M and Blasiak J: Association between polymorphisms of the BRCA2 gene and clinical parameters in breast cancer. Exp Oncol. 31:250–251. 2009.PubMed/NCBI


Carreira A, Hilario J, Amitani I, Baskin RJ, Shivji MK, Venkitaraman AR and Kowalczykowski SC: The BRC repeats of BRCA2 modulate the DNA-binding selectivity of RAD51. Cell. 136:1032–1043. 2009. View Article : Google Scholar : PubMed/NCBI


Johnson N, Fletcher O, Palles C, Rudd M, Webb E, Sellick G, dos Santos Silva I, McCormack V, Gibson L, Fraser A, et al: Counting potentially functional variants in BRCA1, BRCA2 and ATM predicts breast cancer susceptibility. Hum Mol Genet. 16:1051–1057. 2007. View Article : Google Scholar : PubMed/NCBI


Ishitobi M, Miyoshi Y, Ando A, Hasegawa S, Egawa C, Tamaki Y, Monden M and Noguchi S: Association of BRCA2 polymorphism at codon 784 (Met/Val) with breast cancer risk and prognosis. Clin Cancer Res. 9:1376–1380. 2003.PubMed/NCBI


Sliwinski T, Krupa R, Majsterek I, Rykala J, Kolacinska A, Morawiec Z, Drzewoski J, Zadrozny M and Blasiak J: Polymorphisms of the BRCA2 and RAD51 genes in breast cancer. Breast Cancer Res Treat. 94:105–109. 2005. View Article : Google Scholar : PubMed/NCBI


Peltomäki P: Deficient DNA mismatch repair: A common etiologic factor for colon cancer. Hum Mol Genet. 10:735–740. 2001. View Article : Google Scholar : PubMed/NCBI


Yuan ZQ, Gottlieb B, Beitel LK, Wong N, Gordon PH, Wang Q, Puisieux A, Foulkes WD and Trifiro M: Polymorphisms and HNPCC: PMS2-MLH1 protein interactions diminished by single nucleotide polymorphisms. Hum Mutat. 19:108–113. 2002. View Article : Google Scholar : PubMed/NCBI


Umar A, Boland CR, Terdiman JP, Syngal S, de la Chapelle A, Rüschoff J, Fishel R, Lindor NM, Burgart LJ, Hamelin R, et al: Revised Bethesda Guidelines for hereditary nonpolyposis colorectal cancer (Lynch syndrome) and microsatellite instability. J Natl Cancer Inst. 96:261–268. 2004. View Article : Google Scholar : PubMed/NCBI


Cal S, Obaya AJ, Llamazares M, Garabaya C, Quesada V and López-Otín C: Cloning, expression analysis, and structural characterization of seven novel human ADAMTSs, a family of metalloproteinases with disintegrin and thrombospondin-1 domains. Gene. 283:49–62. 2002. View Article : Google Scholar : PubMed/NCBI


Vázquez F, Hastings G, Ortega MA, Lane TF, Oikemus S, Lombardo M and Iruela-Arispe ML: METH-1, a human ortholog of ADAMTS-1, and METH-2 are members of a new family of proteins with angio-inhibitory activity. J Biol Chem. 274:23349–23357. 1999. View Article : Google Scholar : PubMed/NCBI


Colige A, Sieron AL, Li SW, Schwarze U, Petty E, Wertelecki W, Wilcox W, Krakow D, Cohn DH, Reardon W, et al: Human Ehlers-Danlos syndrome type VII C and bovine dermatosparaxis are caused by mutations in the procollagen I N-proteinase gene. Am J Hum Genet. 65:308–317. 1999. View Article : Google Scholar : PubMed/NCBI


Rodriguez-Lopez J, Pombo-Suarez M, Loughlin J, Tsezou A, Blanco FJ, Meulenbelt I, Slagboom PE, Valdes AM, Spector TD, Gomez-Reino JJ, et al: Association of an sSNP in ADAMTS14 to some osteoarthritis phenotypes. Osteoarthritis Cartilage. 17:321–327. 2009. View Article : Google Scholar


Hu X, Chen H, Jin M, Wang X, Lee J, Xu W, Zhang R, Li S and Niu J: Molecular cytogenetic characterization of undifferentiated embryonal sarcoma of the liver: A case report and literature review. Mol Cytogenet. 5:262012. View Article : Google Scholar : PubMed/NCBI


Kumar S, Rao N and Ge R: Emerging Roles of ADAMTSs in Angiogenesis and Cancer. Cancers (Basel). 4:1252–1299. 2012. View Article : Google Scholar


Bonnans C, Chou J and Werb Z: Remodelling the extracellular matrix in development and disease. Nat Rev Mol Cell Biol. 15:786–801. 2014. View Article : Google Scholar : PubMed/NCBI


Goldman O, Han S, Sourisseau M, Dziedzic N, Hamou W, Corneo B, D'Souza S, Sato T, Kotton DN, Bissig KD, et al: KDR identifies a conserved human and murine hepatic progenitor and instructs early liver development. Cell Stem Cell. 12:748–760. 2013. View Article : Google Scholar : PubMed/NCBI


Basu P, Chandna P, Bamezai RNK, Siddiqi M, Saranath D, Lear A and Ratnam S: MassARRAY spectrometry is more sensitive than PreTect HPV-Proofer and consensus PCR for type-specific detection of high-risk oncogenic human papillomavirus genotypes in cervical cancer. J Clin Microbiol. 49:3537–3544. 2011. View Article : Google Scholar : PubMed/NCBI


Kim MJ, Lee EJ, Chun SM, Jang SJ, Kim DS, Lee DH and Youk EG: The significance of ectopic crypt formation in the differential diagnosis of colorectal polyps. Diagn Pathol. 9:2122014. View Article : Google Scholar : PubMed/NCBI


Matsuura K, Sawai H, Ikeo K, Ogawa S, Iio E, Isogawa M, Shimada N, Komori A, Toyoda H, Kumada T, et al Japanese Genome-Wide Association Study Group for Viral Hepatitis: Genome-wide association study identifies TLL1 variant associated with development of hepatocellular carcinoma after eradication of hepatitis C virus infection. Gastroenterology. 152:1383–1394. 2017. View Article : Google Scholar : PubMed/NCBI


Peters MG and Terrault NA: Alcohol use and hepatitis C. Hepatology. 36(Suppl 1): S220–S225. 2002.PubMed/NCBI


Aleksandrova K, Stelmach-Mardas M and Schlesinger S: Obesity and liver cancer. Recent Results Cancer Res. 208:177–198. 2016. View Article : Google Scholar : PubMed/NCBI

Related Articles

Journal Cover

February 2018
Volume 52 Issue 2

Print ISSN: 1019-6439
Online ISSN:1791-2423

Sign up for eToc alerts

Recommend to Library

Copy and paste a formatted citation
Horiuchi, Y., Lin, J., Shinojima, Y., Fujiwara, K., Moriyama, M., & Nagase, H. (2018). Characterizing key nucleotide polymorphisms of hepatitis C virus-disease associations via mass-spectrometric genotyping. International Journal of Oncology, 52, 441-452.
Horiuchi, Y., Lin, J., Shinojima, Y., Fujiwara, K., Moriyama, M., Nagase, H."Characterizing key nucleotide polymorphisms of hepatitis C virus-disease associations via mass-spectrometric genotyping". International Journal of Oncology 52.2 (2018): 441-452.
Horiuchi, Y., Lin, J., Shinojima, Y., Fujiwara, K., Moriyama, M., Nagase, H."Characterizing key nucleotide polymorphisms of hepatitis C virus-disease associations via mass-spectrometric genotyping". International Journal of Oncology 52, no. 2 (2018): 441-452.