As a multisystem disease, nonalcoholic fatty liver disease (NAFLD) is the most common cause of chronic liver disease. At the same time, there is accumulating evidence that NAFLD is not only related to type 2 diabetes, but is likely to promote the progression and development of cardiovascular diseases such as hypertrophy and heart valve calcification, as well as chronic kidney disease.[1–3]
Although the epidemiologic and demographic characteristics of NAFLD vary by region and race, they are generally highly correlated with obesity. However, it is worth mentioning that lean people can also develop NAFLD, and comprise a larger proportion of patients than might be expected. The global prevalence of NAFLD in 2017 was 24%, with higher rates in South America (31%) and Asia (27%). A study by Eslam et al reported that China has the highest rates of incidence and increase of NAFLD, with the number of cases predicted to increase from 2,463,300 in 2016 to 3,145,800 by 2030. A notable clinical challenge of NAFLD is the inter-patient heterogeneity in disease progression, whereby interactions between environmental factors and susceptible polygenes determine the phenotype and development of NAFLD. Genes that contribute to NAFLD susceptibility include PNPLA3, TM6SF2, LYPLAL1, GCKR, and others The PCSK1 single-nucleotide polymorphism (SNP), rs6232, is not a major contributor to obesity in a UK population of European descent, but it may have an age-dependent effect.
PCSK1 encodes the prohormone convertase (PC)1/3. PC1/3 is only expressed in neural and endocrine tissues where it is responsible for cleaving prohormones and proneuropeptides. According to previous reports, various SNPs of PCSK1 are correlated with early-onset obesity, proinsulin disorders, and the promotion of obesity. Stijnen et al proved that PCSK1 rs6232 has a stronger correlation with childhood obesity than adult obesity. Using a systematic review of human genome epidemiology, they demonstrated that PCSK1 rs6232, rs6234, and rs6235 are related to obesity in Caucasians. However, similar reports of such correlations have only focused on the relationship between PCSK1 and obesity. For example, Chang et al reported that common variants of PCSK1 are correlated with obesity in the Chinese population, indicating that the relationship between PCSK1 and NAFLD is worthy of further investigation. Therefore, this study attempted to explore the correlation between PCSK1 and NAFLD by taking the four candidate SNPs (rs6234, rs155971, rs6232, rs3811951) of PCSK1 as the approach in a Han Chinese population.
Materials and methods
This case-control observational study was approved by the Medical Ethics Committee of Shanghai University of Traditional Chinese Medicine (approved No. 2017LCSY069) and conducted in accordance with the 1964 Declaration of Helsinki, as revised in 2013. All patients provided written informed consent in advance of study participation. Patients with NAFLD (n = 732; age range, 60–96 years) and healthy volunteers (n = 823; age range, 60–92 years) participated in this study. In return for participating in the study, we provided some treatment recommendations for NAFLD patients. The clinical diagnosis of NAFLD in each patient was determined by two experienced radiologists through the evaluation of fatty liver on color ultrasound and according to the 2018 criteria of NAFLD of the Chinese Medical Association.
All subjects should meet the following criteria:
- (1) Over 60 years old;
- (2) No history of excessive alcohol consumption. The alcohol equivalent amount for males should be less than 30 g per day, and that for females should be less than 20 g per day;
- (3) Permanent residents of Zhangjiang area, Pudong New Area, Shanghai;
- (4) No blood relationship between the subjects;
- (5) None of the subjects had drug-induced liver disease or autoimmune liver disease;
- (6) No carriers of hepatitis B and C;
- (7) According to the study by Anderson et al, individuals who failed in SNP genotyping were removed (an individual genotyping failure rate should not be higher than 7%. For the four SNPs in this study, any failure should be excluded).
Age, sex, alcohol history, and other medical history were surveyed through questionnaires. The number of individuals at each stage of the study and the reasons for exclusion at each stage are shown in Figure 1. An electronic measuring instrument (Shengyuan, Zhengzhou, Henan Province, China) was used to check the height and weight information of the subjects. Body mass index (BMI) was calculated based on existing information about height and mass: BMI = mass (kg)/height2 (m2). Systolic and diastolic blood pressures were measured using a sphygmomanometer (Biospace, Cheonan, South Korea).
All of the subjects kept fasting overnight before the test. Afterward, physiological indicators such as fasting plasma glucose, total cholesterol, and triacylglycerol of the subjects were collected by an automatic biochemical analyzer (Hitachi, Tokyo, Japan).
DNA extraction, SNP selection, and genotyping
Peripheral blood (5 mL) of each subject was collected by Shanghai Innovation Center of Traditional Chinese Medicine Health Service, Shanghai University of Traditional Chinese Medicine. Total genomic DNA was extracted from the peripheral blood sample by the TIANamp Genomic DNA kit (Tiangen Biotech (Beijing) Co., Ltd., Beijing, China). The concentrations of all DNA samples were measured by a NanoDrop 2000 (Thermo Fisher Scientific, Waltham, MA, USA). Qualified DNA was diluted to 10ng/μL and frozen in a refrigerator for subsequent genotyping.
Four SNPs were identified by searching for obesity variants reported in previous genome-wide association studies: rs6234, rs155971, rs6232, and rs3811951. The SNP genotyping was analyzed using matrix-assisted laser desorption/ionization time-of-flight mass spectrometry in the Bio-X Institutes of Shanghai Jiao Tong University. Data from all subjects were obtained using the same method and in the same batch. The MassARRAY® Analyzer 4 platform (Agena, San Diego, CA, USA) was used to support DNA analyses. The probes and primers were designed by online Assay Design Suite (https://support.agenabio.com/). Multiplex PCR (polymerase chain reaction) was conducted under standard conditions with the Agena iPLEX Reagent Set. The total volume of each reaction system was 5 μL and contained 10ng of genomic DNA. The call rate of all SNPs was more than 95%.
Sample size calculation was performed using the G*Power program based on Cohen method. When the effect intensity was set to 0.1 (corresponding to the “weak” effect), the current sample size showed a significant correlation detection ability of more than 90% (α < 0.05). All participants came from the community physical examination, and neither the subjects nor the doctors knew about the enrollment. It was a double-blinded study.
Two-sided Student t-tests and Pearson Chi-square tests were performed to compare basic demographics (sex, age) and BMI between patients and healthy volunteers (Table 1). The R programming language and software (https://cran.r-project.org/doc/FAQ/R-FAQ.html#Citing-R) were used for data normalization and statistical analyses. Alleles, genotypes, and pairwise linkage disequilibrium were analyzed on the SHEsis platform (http://shesisplus.bio-x.cn/SHEsis.html; Bio-X Institutes, Shanghai, China). Hardy–Weinberg equilibrium was calculated via HaploView Version 4.2. Primitive haplotype development was performed on HaploView, then evaluated via the SHEsis platform. Odds ratios (ORs) and their 95% confidence intervals (CIs) were also computed. The standard for r2 was set at 0.8 for ascertaining the strong pairwise linkage disequilibrium. Age and gender were used as covariables to verify and exclude stratification. To avoid false-positive results, a Bonferroni correction was performed on P-values to obtain Pcorr (Pcorr = P-value × 1/4). Only a Pcorr value < 0.0125 was considered to be statistically significant Fig. 2.
Table 1 -
Basic characteristics of NAFLD patients and healthy control participants
||71.630 ±0 6.34
||71.760 ±0 6.31
||26.920 ±0 3.31
||24.320 ±0 3.45
||145.160 ±0 21.45
||138.630 ±0 24.09
||83.540 ±0 11.55
||81.610 ±0 12.82
||6.500 ±0 1.69
||6.240 ±0 1.73
||2.020 ±0 1.20
||1.740 ±0 1.12
||5.140 ±0 0.94
||5.110 ±0 0.99
||1.150 ±0 0.24
||1.210 ±0 0.28
||3.210 ±0 0.85
||3.160 ±0 0.86
||348.630 ±0 85.04
||334.780 ±0 89.32
||22.440 ±0 13.50
||20.310 ±0 13.86
||24.050 ±0 9.14
||23.680 ±0 9.38
||33.700 ±0 34.88
||30.750 ±0 27.30
||43.810 ±0 2.128
||43.660 ±0 2.49
||14.440 ±0 5.01
||14.760 ±0 5.86
||3.240 ±0 1.63
||3.270 ±0 1.90
||75.620 ±0 20.66
||75.280 ±0 26.86
The data are expressed as the mean ± SD with the exception of male [n (%)]. ALB = albumin, ALP = alkaline phosphatase, ALT = alanine aminotransferase, AST = aspartate aminotransferase, BMI = body mass index, Control = healthy volunteer group, DBIL = direct Bilirubin, DBP = diastolic blood pressure, FPG = fasting plasma glucose, GGT = gamma-glutamyl transferase, HDL-C = high density lipoprotein cholesterol, LDL-C = low density lipoprotein cholesterol, NAFLD = nonalcoholic fatty liver disease patient group, SBP = systolic blood pressure, TBIL = total bilirubin, TC = total cholesterol, TG = triacylglycerol, UA = uric acid.
Characteristics of subjects
There were no significant differences in age or sex between the 732 NAFLD patients and 823 healthy volunteers (Table 1). However, the BMI of the NAFLD group was significantly higher than that of the healthy control group (P < 0.05).
Genetic association between PCSK1 SNPs and NAFLD
No deviation from the Hardy–Weinberg equilibrium was observed for any of the four PCSK1 SNPs in our population. In the comparison of NAFLD patients with healthy volunteers, we found that none of the distributions of the four SNPs, whether allelic or genotypic, had any significant correlation with the occurrence of NAFLD (Table 2).
Table 2 -
Allelic and genotypic distributions [n
(%)] of four SNPs of PCSK1
in patients with NAFLD and healthy volunteers
||OR (95% CI)
ORs (95% CI) were assessed using Pearson Chi-square test. CI = confidence interval, Control = healthy volunteer group, HWE = Hardy–Weinberg equilibrium, NAFLD = nonalcoholic fatty liver disease patient group, OR = odds ratio, P-value = P-value of Pearson chi-squared test for rs6234, rs155971, rs6232, and rs3811951 genotypes, PCSK1=proprotein convertase subtilisin/kexin type 1, SNP = single-nucleotide polymorphism.
*P-value of Pearson chi-squared test for rs6234, rs155971, rs6232, and rs3811951 alleles.
In addition to SHEsis analysis, five different genetic models (codominant, dominant, recessive, overdominant, and log-additive) were used to identify the correlation between genotype frequency and NAFLD. As shown in Table 3, the recessive model for rs3811951 appeared to show a relationship (OR = 1.077; 95% CI = 0.924–1.256; P = 0.04); however, the significance did not withstand Bonferroni correction (Pcorr > 0.05).
Table 3 -
Genotype frequency analysis of four SNPs of PCSK1
in five genetic models
Pcorr (recessive) = P-value of recessive model after Bonferroni correction, SNPs = single-nucleotide polymorphisms. The genotype frequencies of rs6234, rs155971, and rs3811951 were statistically different in the recessive model after Bonferroni correction. The genotype frequencies of rs6234, rs155971, and rs3811951 were statistically different in the codominant model, dominant model, recessive model, overdominant model, and log-additive model. The genotype frequency of rs6232 was statistically different in the codominant model. PCSK1=proprotein convertase subtilisin/kexin type 1. "—" means that it is not applicable to Dominant and other models since rs6232 has only TT and CT genotypes, but no CC genotypes.
Association between PCSK1 SNPs and NAFLD in haplotype
The squared correlation (r2) was assessed for all random SNP combinations. A haplotype block composed of rs6234 and rs3811951 was created via SHEsis. Using haplotype analysis, we identified rs6234 and rs3811951 as a strong linkage disequilibrium region in PCSK1 (D′ = 0.95; r2 = 0.81) (Table 4 and Fig. 2). None of the three haplotypes was significantly associated with the incidence of NAFLD (Table 5).
Table 4 -
Linkage disequilibrium between four SNPs of PCSK1
D′ values are shown above, and squared correlation (r2) values below the diagonal. The linkage disequilibrium analysis D′ is calculated by SHEsis. PCSK1=proprotein convertase subtilisin/kexin type 1, SNPs = single-nucleotide polymorphisms.
Table 5 -
Estimated haplotype frequency analysis between the NAFLD and healthy control groups
||OR (95% CI)
*Pearson P-value, Fisher P-value for haplotypes GCTA, GTTA, and CCTG. Case = nonalcoholic fatty liver disease (NAFLD) patient group, CI = confidence interval, Control = healthy volunteer group, OR = odds ratio.
The unit of LD is the D-value that measures the deviation between the observed haplotype frequency and the expected frequency at equilibrium. The value of D ranges from −0.25 to 0.25. Since D-value is strictly dependent on allele frequency, it is not available for expressing actual LD intensity. The most commonly used measurements of LD are D′ and r2. D′ reflects the history of population recombination, which is suitable for studying the degree of population linkage imbalance and r2 reflects the degree of allele correlation. D′ = D/Dmax. The value of D′ ranges from −1 to 1.
As the coding gene for PC1/3, PCSK1 was one of the first genes to be associated with unifactorial early-onset obesity. The enzymatic activity of PC1/3 is indispensable for the activating cleavage of many peptide hormone precursors, including proinsulin, proglucagon, and proghrelin. The close relationship between PC1/3 and physiological processes including glucose homeostasis has been reported in a variety of studies. Rohden et al showed that the bioactivation of pro-GIP (glucose-dependent insulinotropic peptide) and proglucagon were affected by low levels of PC1/3 mRNA in jejunal cells of patients with obesity and type 2 diabetes mellitus. Zhu et al found that PC1/3-null mice were not obese but did not grow normally, with about 60% of the size of normal mice at 10 weeks. O’Rahilly et al reported a case of compound heterozygosity for PC1/3 in a female patient with severe obesity, persistent amenorrhea, and postprandial hypoglycemia. Heni et al showed that rs6232 and rs6235 of PCSK1 could affect the transformation of proinsulin stimulated by glucose in a German population, which inspired our study. Furthermore, Heni et al found that rs6232 influenced glucose homeostasis in a manner independent of BMI and proinsulin concentration.
There are several reports indicating that mutations in PCSK1, particularly the SNPs, are directly related to obesity. Muhsin et al constructed a mouse model with a point mutation in PCSK1 that exhibits phenotypic traits of obesity, adephagia, and hyperinsulinemia that are consistent with those in humans. Using a systematic review of the literature, Nead et al collected phenotypic and genetic data of 331,175 individuals from different races, proving through calculation that the PCSK1 SNPs rs6232 and rs6234/rs6235 are associated with BMI variation and susceptibility to obesity. Löffler et al studied genetic data from 52 children with elevated proinsulin and/or glucose tolerance defects, identifying PCSK1 rs6232 and rs6234 as polygenic risk variants for childhood obesity. Philippe et al found a new heterozygous nonsense mutation of PCSK1 (P. arg80*) that led to truncation of the coding pro-peptide to <2 exons (14 exons in the wild-type) and co-segregated with obesity in three generations of a French family.
None of our results showed a significant correlation between PCSK1 SNPs and the occurrence of NAFLD in the Han Chinese. We believe that PCSK1 may not be an important cause of NAFLD or even obesity in this population. In a study involving 3210 Han Chinese of both sexes, Qi et al found no significant associations between PCSK1 rs6234 and obesity, overweight, BMI, waist circumference, or body fat percentage.
Another factor that may have affected our results is age. A study of 20,249 adults of European descent conducted by Kilpeläinen et al found that PCSK1 rs6232 was significantly associated with BMI (P = 0.021) and obesity (P = 0.022) in the younger group (under the median age of 59), but not in the older group. These results suggest that the PCSK1 rs6232 is not a major factor in common obesity in the UK general population, and that the effect of rs6232 may be related to age.
Although NAFLD and obesity often co-exist, there may be associations between NAFLD and SNPs of various genes that are also affected by age, sex, and moreover, race. As mentioned above, even in obesity-related studies, uncertainty remains whether PCSK1 is involved in the occurrence and development of obesity in people of Han Chinese descent. Therefore, we suspect that the PCSK1 SNPs examined in this study are also unlikely to be major contributors to NAFLD in this population.
Our research has some limitations. First, although the sample size satisfied power greater than 90%, there was still the possibility of false negative. Secondly, the current diagnostic criteria for NAFLD are based on the subjective judgment of ultrasound doctors, lacking objective scoring criteria and not completely accurate. This may lead to missed or mis-checked situations, which will affect our final results. Finally, only four SNPs from PCSK1 were selected for the correlation study with NAFLD, which could not prove the correlation between PCSK1 gene and NAFLD in the Han Chinese population. Therefore, in the case of increasing the number of SNP loci, selecting another set of samples supplemented by updated NAFLD diagnostic criteria will help obtain more stable results in subsequent studies.
Multiple statistical methods failed to show any significant correlations between four obesity-related PCSK1 SNPs (rs6234, rs155971, rs6232, and rs3811951) and NAFLD in a Han Chinese population. Further research on the genetic susceptibility component of NAFLD in this population is needed and will require larger genetic datasets.
XL and YS participated in the design and review of this manuscript. BL, NW, DR, YB, FYang, SY, FJ, RW, and GH participated in the collection of samples. XY and FYuan participated in the data analysis. XY, LL, LJ, KH, MF, KS, XW, and QL participated in literature research. XY was mainly responsible for the manuscript writing. All authors approved the final version of the manuscript.
This work was supported by Innovation Funding in Shanghai (Nos. 20JC1418600 and 18JC1413100), the National Nature Science Foundation of China (Nos. 82071262 and 81671326), Natural Science Foundation of Shanghai (Nos. 20ZR1427200 and 20511101900), Shanghai Municipal Science and Technology Major Project (No. 2017SHZDZX01), and the Shanghai Leading Academic Discipline Project (No. B205). The content is solely the responsibility of the authors and does not represent the official views of the funding agencies.
Institutional review board statement
This study was approved by the Medical Ethics Committee of Shanghai University of Traditional Chinese Medicine (approved No. 2017LCSY069).
Declaration of participant consent
The authors certify that they have obtained the participant consent forms. In the forms, participants have given their consent for their images and other clinical information to be reported in the journal. The participants understand that their names and initials will not be published and due efforts will be made to conceal their identity.
Conflicts of interest
There are no conflicts of interest.
Editor note: GH is an Editorial Board member of Journal of Bio-X Research. She was blinded from reviewing or making decisions on the manuscript. The article was subject to the journal’s standard procedures, with peer review handled independently of this Editorial Board member and their research groups.
. Byrne CD, Targher G. NAFLD: a multisystem disease. J Hepatol. 2015;62 (1 Suppl):S47–S64.
. Musso G, Gambino R, Tabibian JH, et al. Association of non-alcoholic fatty liver disease with chronic kidney disease: a systematic review and meta-analysis. PLoS Med. 2014;11:e1001680.
. Kuchay MS, Krishan S, Mishra SK, et al. Effect of empagliflozin on liver fat in patients with type 2 diabetes and nonalcoholic fatty liver disease: a randomized controlled trial (E-LIFT Trial). Diabetes Care. 2018;41:1801–1808.
. Eslam M, Valenti L, Romeo S. Genetics and epigenetics of NAFLD and NASH: clinical impact. J Hepatol. 2018;68:268–279.
. Kilpeläinen TO, Bingham SA, Khaw KT, et al. Association of variants in the PCSK1
gene with obesity in the EPIC-Norfolk study. Hum Mol Genet. 2009;18:3496–3501.
. Stijnen P, Ramos-Molina B, O’Rahilly S, et al. PCSK1
mutations and human endocrinopathies: from obesity to gastrointestinal disorders. Endocr Rev. 2016;37:347–371.
. Stijnen P, Tuand K, Varga TV, et al. The association of common variants in PCSK1
with obesity: a HuGE review and meta-analysis. Am J Epidemiol. 2014;180:1051–1065.
. Chang YC, Chiu YF, Shih KC, et al. Common PCSK1
haplotypes are associated with obesity in the Chinese population. Obesity (Silver Spring). 2010;18:1404–1409.
. Fan JG, Wei L, Zhuang H, et al. Guidelines of prevention and treatment of nonalcoholic fatty liver disease (2018, China). J Dig Dis. 2019;20:163–173.
. Anderson CA, Pettersson FH, Clarke GM, et al. Data quality control in genetic case-control association studies. Nat Protoc. 2010;5:1564–1573.
. Lachenbruch PA. Statistical power analysis for the behavioral sciences (2nd ed.). J Am Stat Assoc. 1989;84:1096–1097.
. Shi YY, He L. SHEsis, a powerful software platform for analyses of linkage disequilibrium, haplotype construction, and genetic association at polymorphism loci. Cell Res. 2005;15:97–98.
. Barrett JC, Fry B, Maller J, et al. Haploview: analysis and visualization of LD and haplotype maps. Bioinformatics. 2005;21:263–265.
. Li Z, Zhang Z, He Z, et al. A partition-ligation-combination-subdivision EM algorithm for haplotype inference with multiallelic markers: update of the SHEsis http://analysis.bio-x.cn
). Cell Res. 2009;19:519–523.
. Ramos-Molina B, Martin MG, Lindberg I. PCSK1
variants and human obesity. Prog Mol Biol Transl Sci. 2016;140:47–74.
. Rohden F, Costa CS, Hammes TO, et al. Obesity associated with type 2 diabetes mellitus is linked to decreased PC1/3 mRNA expression in the Jejunum. Obes Surg. 2014;24:2075–2081.
. Zhu X, Zhou A, Dey A, et al. Disruption of PC1/3 expression in mice causes dwarfism and multiple neuroendocrine peptide processing defects. Proc Natl Acad Sci U S A. 2002;99:10293–10298.
. O’Rahilly S, Gray H, Humphreys PJ, et al. Brief report: impaired processing of prohormones associated with abnormalities of glucose homeostasis and adrenal function. N Engl J Med. 1995;333:1386–1390.
. Heni M, Haupt A, Schäfer SA, et al. Association of obesity risk SNPs in PCSK1
with insulin sensitivity and proinsulin conversion. BMC Med Genet. 2010;11:86.
. Muhsin NIA, Bentley L, Bai Y, et al. A novel mutation in the mouse Pcsk1
gene showing obesity and diabetes. Mamm Genome. 2020;31:17–29.
. Nead KT, Li A, Wehner MR, et al. Contribution of common non-synonymous variants in PCSK1
to body mass index variation and risk of obesity: a systematic review and meta-analysis with evidence from up to 331 175 individuals. Hum Mol Genet. 2015;24:3582–3594.
. Löffler D, Behrendt S, Creemers JWM, et al. Functional and clinical relevance of novel and known PCSK1
variants for childhood obesity and glucose metabolism. Mol Metab. 2016;6:295–305.
. Philippe J, Stijnen P, Meyre D, et al. A nonsense loss-of-function mutation in PCSK1
contributes to dominantly inherited human obesity. Int J Obes (Lond). 2015;39:295–302.
. Qi Q, Li H, Loos RJ, et al. Association of PCSK1
rs6234 with obesity and related traits in a Chinese Han population. PLoS One. 2010;5:e10590.