Non-Alcoholic Fatty Liver Disease in Long-Term Type 2 Diabetes: Role of rs738409 PNPLA3 and rs499765 FGF21 Polymorphisms and Serum Biomarkers

Fibroblast growth factor 21 (FGF21) signaling and genetic factors are involved in non-alcoholic fatty liver disease (NAFLD) pathogenesis. However, these factors have rarely been studied in type 2 diabetes mellitus (T2D) patients from admixed populations such as in those of Brazil. Therefore, we aimed to evaluate rs738409 patanin-like phospholipase domain-containing protein (PNPLA3) and rs499765 FGF21 polymorphisms in T2D, and their association with NAFLD, liver fibrosis, and serum biomarkers (FGF21 and cytokeratin 18 levels). A total of 158 patients were included, and the frequency of NAFLD was 88.6%, which was independently associated with elevated body mass index. Significant liver fibrosis (≥F2) was detected by transient elastography (TE) in 26.8% of NAFLD patients, and was independently associated with obesity, low density lipoprotein, and gamma-glutamyl transferase (GGT). PNPLA3 GG genotype and GGT were independently associated with cirrhosis. PNPLA3 GG genotype patients had higher GGT and AST levels; PNPLA3 GG carriers had higher TE values than CG patients, and FGF21 CG genotype patients showed lower gamma-GT values than CC patients. No differences were found in serum values of FGF21 and CK18 in relation to the presence of NAFLD or liver fibrosis. The proportion of NAFLD patients with liver fibrosis was relevant in the present admixed T2D population, and was associated with PNPLA3 polymorphisms.


Introduction
Non-alcoholic fatty liver disease (NAFLD) is one of the most common liver diseases worldwide, affecting 25% of the general population [1]. In Latin America, the estimated prevalence of NAFLD (31%) is even higher than the worldwide rate [1]. The increasing prevalence of fatty liver is closely related to the global growth of type 2 diabetes mellitus (T2D) and obesity, two major drivers of the disease severity and progression [2,3]. Indeed,

Clinical Design and Patients' Selection
This study was designed as single-center longitudinal study conducted in patients with T2D who were followed between 2018 and 2020 at the Outpatient Units of the Endocrinology and Gastroenterology Division (Gastrocentro) of the University of Campinas (UNICAMP), Campinas, Brazil. Patients were invited to participate in the study and evaluated regarding the presence of NAFLD through documentation of liver steatosis on an abdominal ultrasound (US) and clinical and laboratory workups for the exclusion of other chronic liver diseases, such as alcoholic liver disease, hepatitis B and C, autoimmune hepatitis, hemochromatosis, alpha 1-antitrypsin deficiency, and Wilson's disease.
The inclusion criteria were age 18-75 years old and presence of T2D. The exclusion criteria consisted of any other chronic liver disease, previous or current significant alcohol intake (>20 g/day for women and >30 g/day for men), human immunodeficiency virus (HIV), history of exposure to drugs associated with fatty liver (corticosteroids, amiodarone, methotrexate, and/or tamoxifen), previous liver transplantation, presence of other forms of diabetes, presence of relevant systemic diseases that could interfere with liver evaluation, and/or refusal to participate in the study. Patients who were included underwent blood sampling for biochemical analysis, evaluations of SNPs in PNPLA3 and FGF21 genes, and serum biomarkers (CK18 and FGF21). Additionally, the subjects underwent a liver US exam and were non-invasively evaluated for liver fibrosis based on FibroScan ® .

Variables
Demographic and anthropometric data (age, gender, abdominal circumference, and body mass index (BMI)) were obtained, in addition to that regarding the presence of comorbidities (arterial hypertension, dyslipidemia, obesity, coronary arterial disease (CAD), hypothyroidism, chronic kidney disease (CKD), and tobacco use). The presence of CAD was defined by review of medical charts, which evaluated clinical history, electrocardiographic criteria, and/or an ischemic stress test showing CAD, description of angina, and/or previous acute myocardial infarction. CKD was defined as a reduction in glomerular filtration rate to values <60 mL/min/1.73 m 2 for ≥3 months. Regarding T2D, time since diagnosis (>10 years was defined as long-term), medications used, and related complications (retinopathy, diabetic kidney disease and neuropathy) were also recorded.

Definition of Non-Alcoholic Fatty Liver Disease
NAFLD was defined when the liver US (Toshiba Aplio, Toshiba America Medical Systems Inc., Tustin, CA, USA) indicated hepatic steatosis and when other chronic liver diseases, significant alcohol consumption, and/or steatogenic medications were excluded. Patients who had a pattern of chronic liver disease or cirrhosis on US liver evaluation [17] in the absence of other etiology to explain the liver disease were also classified as having NAFLD.

Serum Biomarkers
For analysis of M65 CK18 and FGF21 levels, blood samples were collected in dry tubes, and serum was stored at −80 • C until analysis. Commercial enzyme-linked immunosorbent assay (ELISA) kits from Finetest EH2820 and EH0130 (Wuhan Fine Biotech Co., Ltd., Wuhan, China) were used for the evaluation of CK18 (measured in ng/mL) and FGF21 (measured in pg/mL), respectively, as per the manufacturers' recommendations.

DNA Extraction and Genotyping
Genomic DNA was extracted using the salting-out technique from blood samples collected in EDTA tubes. DNA quantification was determined by spectrophotometry (GeneQuant DNA/RNA Calculator, Pharmacia, LKC Biotechnology, Uppsala, Sweden). After DNA extraction from leucocytes, the samples were stored at −20 • C until evaluation of the polymorphisms. Genotyping of the SNPs rs738409 in PNPLA3 and rs499765 in FGF21 was performed by real-time polymerase chain reaction (Applied Biosystems, Ther-moFisher Brand, Foster City, CA, USA) using the TaqMan TM SNP Genotyping Assay, human/PNPLA3 (C_7241_10) and TaqMan TM SNP Genotyping Assay, human/FGF21 (C_987279_10) (ThermoFisher, Foster City, CA, USA), respectively, as per the manufacturer's recommendations. The equipment used for reading was Step One Plus from Applied Biosystems.

Ethical Considerations
This study was approved by the Ethics Committee of UNICAMP (Approval No. 2,766,627). The protocol was conducted in accord with the ethical guidelines of the 2013 World Medical Association Declaration of Helsinki [19]. Informed consent was obtained from participants prior to starting the study.

Statistical Analysis
To analyze the demographic, clinical, and biochemical variables, we obtained the frequencies (absolute and percentage) of the categorical variables and descriptive measures (mean, standard deviation, median, and minimum and maximum values) for the quantitative variables. To assess the association between categorical variables, the chisquared test was used, and when necessary, Fisher's exact test. For numerical variables, the Mann-Whitney and Kruskal-Wallis tests were used. Univariate and multivariate logistic regression were performed to assess factors related to the presence of NAFLD, NAFLD with significant fibrosis (≥F2), and cirrhosis (F4), according to TE measurement. A stepwise method of variable selection was used in the multivariate logistic regression analysis. Odds ratio (OR) measures with 95% confidence intervals (95% CI) were obtained. A probability value of <0.05 was considered significant. Regarding the genetic polymorphisms, the Hardy-Weinberg equilibrium was evaluated. Heterozygosity and the polymorphic information content (PIC) were also calculated. The Statistical Analysis System (SAS) for Windows software package, version 9.4 (SAS Institute Inc., 2002-2008, Cary, NC, USA) was used for the statistical analyses conducted by a biomedical statistician from the Statistics Service at the School of Medical Sciences of UNICAMP.

Results
A total of 306 patients were evaluated for study inclusion. Of these, 148 patients did not meet the inclusion criteria or became ineligible during the study evaluation, as shown in Figure 1. The study population consisted of 158 patients, among whom 62% (98) were women. The mean age was 61.2 years, and the mean duration of diabetes was 18.2 years. Regarding comorbidities, 90.5% (143) had hypertension, 83.5% (132) had dyslipidemia, 24.7% (39) had CAD, and 20.9% (33) had CKD. Regarding T2D complications, 60.8% (96) had retinopathy, 38.6% (61) had neuropathy, and 52.5% (83) had diabetic kidney disease. Most patients (80%, 127) were taking insulin, and their median HbA1C was 8.10%. The median waist circumference was 99 cm, and median BMI was 30.25 kg/m 2 , with 84.8% (134) of the patients presenting as overweight or obese; 81.6% (129) of the patients were taking statins. On laboratory evaluation, a minority of the patients had elevated aminotransferases levels (12.7% (20) for ALT >35 U/L and 5.7% (9) for AST >35 U/L). The main characteristics of the study patients are detailed in Table 1.

Factors Associated with NAFLD
Based on an age-adjusted univariate logistic regression analysis, the factors associated with the presence of NAFLD were obesity (OR: 4.6; 95% CI: 1.  Table 2. No differences were found in serum values of FGF21 and CK18 biomarkers in regard to the presence of NAFLD. In patients with or without NAFLD, the median FGF21 serum values were 387.2 versus 358.3 pg/mL (p = 0.1104); and for CK18, 0.7 versus 0.8 ng/mL (p = 0.7492).

PNPLA3 and FGF21 Polymorphisms
The distribution of the genotypes for both evaluated SNPs was consistent with the Hardy-Weinberg equilibrium. The heterozygosity coefficient and PIC values for the PN-PLA3 SNP were 0.5878 and 0.5013, respectively. For the FGF21 SNP, these values were 0.6264 and 0.5494, respectively. The frequencies of the genotypes for both SNPs did not differ between patients with and without NAFLD, as shown in Table 5. Patients with the G/G genotype of the SNP rs738409 in PNPLA3 had higher GGT values compared to C/C and C/G genotypes (71.1 ± 57.2 versus 42.8 ± 42.3 and 38.0 ± 30.5 U/L, respectively; p = 0.0336), in addition to higher AST values (32.8 ± 24.7 versus 19.2 ± 6.9 and 19.3 ± 4.9 U/L; p = 0.0039). AST levels were elevated more frequently in individuals with the G/G genotype (25%) compared to other genotypes (p = 0.0079). In addition, GG carriers had higher TE values than CG carriers (p = 0.0154), as shown in Table 6. Table 6. Demographic, clinical, and biochemical characteristics of the genotypes of the single nucleotide polymorphism in PNPLA3 in patients with T2D (n = 148).  Regarding SNP rs499765 in FGF21, patients with the CG genotype showed lower GGT values than those with the CC genotype (33 ± 28.4 versus 53.7 ± 41.2 U/L; p = 0.0013). No other associations with biochemical and clinical variables and non-invasive assessment of liver fibrosis were detected. There was no association of serum FGF21 and CK18 values with the genotypes of SNPs rs738409 in PNPLA3 (p = 0.3416 and 0.4177, respectively) and rs499765 in FGF21 (p = 0.5633 and 0.2403, respectively).

Discussion
The present study evaluated a long-term T2D population at a tertiary university center and found an 88.6% prevalence of NAFLD, among which more than a quarter of participants had significant liver fibrosis based on TE assessment. Elevated BMI was independently associated with fatty liver, and also with significant liver fibrosis, along with LDL and GGT levels. In contrast, cirrhosis was associated with GGT levels and with the GG genotype of the SNP rs738409 in PNPLA in the multivariate logistic regression analysis. The evaluated PNPLA3 and FGF21 SNPs were not associated with NAFLD among T2D patients, but the PNPLA3 GG genotype carriers had higher GGT and AST values compared to the other genotypes, and higher TE LSM values than the CG carriers. Participants with the FGF21 CG genotype showed lower GGT values than the CC genotype carriers.
T2D and NAFLD are strongly related. The pooled prevalence of NAFLD diagnosed by US in T2D patients in a recent systematic review and meta-analysis was 59.21%, and this increased to 91.62% when it was diagnosed by either liver biopsy, non-invasive markers, or radiologic methods [4]. Among 561 T2D American patients, 30% of them followed at endocrinology outpatient clinics, Lomonaco et al. (2021) reported a 70% prevalence of steatosis detected by controlled attenuation parameter elastography [5]. The higher NAFLD detection rate in the present study could be attributed to the cohort profile, composed exclusively from tertiary center outpatients with long-term T2D (mean of 18 years) and poor glycemic control (mean HbA1C of 8.5%). In addition, almost 85% of the patients were overweight or obese, a factor that was independently associated with NAFLD in the present study. Indeed, evidence from a meta-analysis of 21 cohort studies showed that obesity independently led to a 3.5-fold increase in the risk of developing NAFLD [20]. The predominant environmental metabolic conditions, such as obesity and long-term T2D with poor glycemic control, may have hindered an association between rs738409 in PNPLA and NAFLD in this population.
Liver fibrosis is a strong predictor of all-cause and liver-related mortality in NAFLD patients, even at moderate stages (F2) [21,22]. In the study population, significant liver fibrosis (≥F2), as evaluated by TE, was found in 26.7% of NAFLD patients, a value higher than the 15% reported by Lomonaco et al. (2021) [5]. In addition to the aforementioned population characteristics, this difference in the prevalence of significant liver fibrosis could be due to the lower LSM cut-off definition adopted in the present study (LSM ≥ 7.0 kPa). Advanced liver fibrosis assessed by liver biopsy is described in 17% of T2D patients with NAFLD [4]. Due to a higher proportion of patients in the intermediate category of serum non-invasive markers of fibrosis, such as NAFLD Fibrosis Score and FIB4 [23], and the possible suboptimal performance of these in T2D patients [24], some authors have suggested to first use TE in this context to rule out advanced fibrosis [25].
Along with obesity, higher GGT levels were also independently associated with significant liver fibrosis in the present cohort, as previously described [26,27]. GGT may be involved in the modulation of the redox equilibria within cells and neighborhood, and low antioxidant defenses are correlated with elevated GGT levels [28,29]. The reduced ability of the liver to synthesize LDL with fibrosis progression could justify the association of LDL levels with significant liver fibrosis in this population [30]. Moreover, measurement of serum lipids and metabolites may assist to unravel the metabolic changes during NAFLD liver fibrosis progression [31]. Elevated aminotransferases in the present cohort were uncommon, since only 12.7% had ALT and 5.7% had AST above the reference values; findings that were quite similar to those described by Lomonaco et al. (2021) of 10% and 6%, respectively [5]. On the other hand, in patients with NAFLD, the mean ALT values were higher in those with significant fibrosis, but still within normal range according to the laboratory reference. This finding may be associated with the concepts of "normal" and "healthy" serum ALT levels, which motivated a proposal to reduce the ALT normal laboratory cut-off points to <19 U/L in women and <30 U/L in men with chronic hepatitis C or NAFLD [32].
The GG genotype of the PNPLA3 SNP rs738409 was independently associated with cirrhosis in NAFLD patients with T2D. This SNP is responsible for the largest proportion of the genetic predisposition related to NAFLD, and is related to the whole spectrum of the disease [11,33]. In admixed populations, such as in Brazil, this SNP was also related to liver fibrosis [14,34]. However, data exclusively in T2D patients are scarce in Brazil. Machado et al. (2019) evaluated T2D Brazilian patients and reported significant liver fibrosis (TE LSM ≥ 7.9 kPa) in 25.2% of the cases, and a GG genotype frequency in rs738409 of 12% [15]; these results were similar to the present study values of 26.8% and 10.8%, respectively. In addition, significant liver fibrosis was independently associated with the G allele of rs738409 [15]. As recently reported by Gabriel-Medina et al. (2022), if a genetic risk factor, such as PNPLA3 rs738409, coexists with T2D, a major metabolic driver of NAFLD progression, advanced liver fibrosis may result even more frequently [35]. This SNP could be useful for personalized risk prediction of liver fibrosis in T2D subjects in the future. Accordingly, a recent large prospective UK Biobank study consisting of 22,812 T2D participants demonstrated that PNPLA3 rs738409 was independently associated with an increased risk of severe liver disease during follow-up [36].
The clinical significance of the CG genotype in FGF21 rs499765, which was associated with lower GGT values than the CC genotype, needs to be further elucidated since it is the first description of this SNP in Brazilian patients. The same finding is true for FGF21 serum levels in admixed populations. As only 12 patients with cirrhosis were included, the role of the serum biomarkers in this group remains unclear. The relationship between T2D and NAFLD is bidirectional, and prospective data have shown that NAFLD patients have a 2.19-fold increased risk of incident diabetes over time [37]. Interestingly, a recent Korean longitudinal study reported that PNPLA3 SNP rs738409 may modulate the risk of incident T2D in subjects both with and without NAFLD [38].
The present study brings some limitations. Evaluation of the use of antidiabetics that could potentially have an effect on NAFLD, such as pioglitazone and sodium glucose co-transporter-2 inhibitors, was not performed in this population. The lack of liver histological analysis hampered the steatohepatitis evaluation and the assessment of its impact on the results. In addition, the absence of histological data impaired the performance and usefulness of serum M65 CK18 and FGF21, biomarkers used for discrimination of steatohepatitis in NAFLD [39]. On the other hand, the lack of liver histological analysis could be acceptable, as non-invasive liver fibrosis evaluation is current clinical practice and broadens the applicability of the study findings beyond specialized liver centers. It should be noted that there is a proposal to change the term NAFLD to metabolic dysfunction-associated fatty liver disease (MAFLD), which does not currently discriminate steatosis from steatohepatitis [40,41]. The sample size was moderate, but the study provides a portrait of T2D patients from a mixed-race population with long-term disease, with findings that mirror those from larger cohorts [5,27], and it also adds a genetic evaluation.

Conclusions
In conclusion, in this evaluation of an admixed T2D population, we found that NAFLD with liver fibrosis by TE was relevant and associated with SNP rs738409 in PNPLA3. More studies are expected to clarify the involvement of the FGF21 rs499765 SNP and of serum FGF21 levels in T2D and NAFLD patients. Informed Consent Statement: Informed consent was obtained from all subjects involved in the study.

Data Availability Statement:
The raw data required to reproduce these results are available from the corresponding author upon reasonable request.