2024 Volume 71 Issue 2 Pages 193-197
The post-hoc study, derived from our previous prospective observational study, investigated the association between fasting serum proinsulin levels and hepatic steatosis in people with type 2 diabetes. The severity of hepatic steatosis was assessed using the fatty liver index. A total of 268 participants were divided into three groups: low (n = 110), moderate (n = 75), and high fatty liver index (n = 83). In both the crude and age/sex-adjusted analysis, logarithm-transformed proinsulin was significantly higher in the high fatty liver index group than in the low or moderate groups (all p < 0.01). The moderate fatty liver index group showed higher logarithm-transformed proinsulin than the low group (both p < 0.01). Positive associations between proinsulin and fatty liver index shown in this study would support an involvement of hepato-pancreatic crosstalk in the pathophysiology of type 2 diabetes.
PROINSULIN is an immature, 86-amino acid peptide that is synthesized in the endoplasmic reticulum of pancreatic beta cells and is cleaved into equimolar concentrations of insulin and C-peptide [1]. Metabolic stress, including oxidative stress and endoplasmic reticulum stress, cause beta cell dysfunction, leading to increased secretion of incompletely processed proinsulin [2]. Therefore, increased proinsulin levels are not only a biomarker for beta cell dysfunction, but also for increased risk of type 2 diabetes and impaired glucose tolerance [3-6].
Prevalence of hepatic steatosis is increasing worldwide, especially as a comorbidity of type 2 diabetes [7, 8]. Additionally, the presence of hepatic steatosis is associated with the risk of developing type 2 diabetes [9]. Previously, we showed that fasting proinsulin levels were positively associated with the severity of hepatic steatosis in a Japanese population-based study [10]. However, it is unclear whether a similar association exists in people with type 2 diabetes. Therefore, in the present study, we investigated the association between proinsulin levels and hepatic steatosis in individuals with type 2 diabetes.
We previously performed a prospective observational study [11, 12]. In brief, Japanese people with type 2 diabetes aged 20 years or older from four institutions were recruited, and clinical information and fasting blood samples were collected. The exclusion criteria for participants are described elsewhere [11]. Data obtained as part of the above study were analyzed in this post-hoc analysis. Specifically, clinical information including age, sex, anthropometric measurements, and duration of diabetes and biochemical parameters including plasma glucose, HbA1c, serum C-peptide, aspartate aminotransferase (AST), alanine aminotransferase (ALT), γ-glutamyl transpeptidase (γ-GTP), triglycerides (TG), high density lipoprotein (HDL) cholesterol, estimated glomerular filtration rate (eGFR), and proinsulin from fasting blood samples were included. Proinsulin levels were measured using a proinsulin enzyme-linked immunosorbent assay kit (Mercodia, Uppsala, Sweden); other biochemical parameters were measured using standard techniques. The severity of hepatic steatosis was assessed using the fatty liver index, which is calculated using the following formula: ((exp (0.953 * ln (TG) + 0.139 * BMI + 0.718 * ln (γ-GTP) + 0.053 * waist circumference – 15.745)/1 + exp (0.953 * ln (TG) + 0.139 * BMI + 0.718 * ln (γ-GTP) + 0.053 * waist circumference – 15.745)) * 100 [13]. This study was approved by the institutional review board of Hokkaido University Hospital (017-0147), carried out in accordance with the Declaration of Helsinki, and registered with the University Hospital Medical Information Network Center (UMIN000029993). All participants provided written informed consent.
Steatosis was classified in accordance with previously determined fatty liver index values corresponding to no steatosis (<30), borderline steatosis (≥30 and <60) and definite steatosis (≥60) [14]. In this analysis, groups were named low, moderate, and high fatty liver index, respectively. Clinical information including anthropometric and biochemical data were compared between the three groups. A Kruskal-Wallis or chi-square test was used to examine significant differences between groups. Fasting proinsulin levels exhibited skewed distributions; therefore, these values were natural logarithm (ln)-transformed. Analysis of covariance followed by Tukey’s honestly significant difference test for multiple post-hoc comparisons were used to compare these values between the three groups, with age and sex incorporated as covariates. Additionally, a similar comparison was performed, stratified by BMI (< or ≥25 kg/m2) and HbA1c (< or ≥7.0%) to verify the effect of obesity and glucose tolerance. Values are expressed as median (interquartile range) or the number (%) of subjects. p < 0.05 was considered statistically significance. JMP 10 (SAS Institute Inc., Cary, NC, USA) was used to carry out statistical analysis.
Of the 284 participants from our previous analysis [11], one was excluded due to a missing waist circumference. Fifteen participants were excluded because their proinsulin levels were not available (n = 12) or their values were undetectable (n = 3). Therefore, a total of 268 participants were included in this analysis. Clinical information and biochemical characteristics of these participants stratified by severity of hepatic steatosis are shown in Table 1. With increasing fatty liver index, BMI, waist circumference, FPG, HbA1c, C-peptide, AST, ALT, γ-GTP, and TG were all significantly higher (all p < 0.01) and age and HDL cholesterol were significantly lower (both p < 0.01). In both the crude and age/sex-adjusted analysis, ln (proinsulin + 1) was significantly higher in the high fatty liver index group than in the low or moderate groups (all p < 0.01). The moderate fatty liver index group showed higher ln (proinsulin + 1) than the low group (both p < 0.01; Table 2). When these participants were divided into two groups based on BMI (< or ≥25 kg/m2) or HbA1c (< or ≥7.0%), similar associations were observed; however, there were no statistically significant differences between the moderate and high groups in the BMI <25 population and the low and moderate groups in the BMI ≥25 population (Supplementary Tables 1 and 2).
Anthropometric and biochemical characteristics of 268 participants
| Fatty liver index | p value | |||
|---|---|---|---|---|
| Low group | Moderate group | High group | ||
| n | 110 | 75 | 83 | |
| Age (years) | 71 (63, 77) | 70 (66, 78) | 62 (54, 69) | <0.01 |
| Number of females, n (%) | 52 (47) | 32 (43) | 31 (37) | 0.39 |
| BMI (kg/m2) | 22.1 (20.6, 23.6) | 26.1 (24.3, 26.9) | 29.1 (26.9, 31.4) | <0.01 |
| Waist circumference (cm) | 83 (79, 88) | 92 (89, 95) | 100 (94, 107) | <0.01 |
| Duration of diabetes (years) | 15 (8, 23) | 15 (10, 24) | 11 (6, 20) | 0.05 |
| FPG (mg/dL) | 128 (109, 147) | 142 (127, 161) | 145 (127, 166) | <0.01 |
| HbA1c (%) | 7.0 (6.5, 7.6) | 7.2 (6.8, 7.7) | 7.3 (6.9, 7.9) | <0.01 |
| C-peptide (ng/mL) | 1.3 (0.8, 1.7) | 1.7 (1.2, 2.2) | 2.8 (2.1, 4.0) | <0.01 |
| AST (U/L) | 20 (17, 23) | 22 (19, 29) | 25 (20, 34) | <0.01 |
| ALT (U/L) | 17 (12, 25) | 22 (16, 31) | 30 (19, 44) | <0.01 |
| γ-GTP (U/L) | 18 (15, 24) | 27 (21, 38) | 51 (30, 68) | <0.01 |
| TG (mg/dL) | 82 (64, 109) | 126 (98, 153) | 177 (130, 258) | <0.01 |
| HDL cholesterol (mg/dL) | 59 (46, 69) | 51 (42, 62) | 45 (40, 56) | <0.01 |
| LDL cholesterol (mg/dL) | 92 (73, 118) | 93 (78, 111) | 92 (78, 109) | 0.90 |
| eGFR (mL/min/1.73 m2) | 64 (52, 74) | 64 (54, 77) | 71 (56, 85) | 0.13 |
| Fatty liver index | 14 (9, 20) | 42 (37, 50) | 79 (69, 89) | <0.01 |
Values are expressed as median (interquartile range) or the number (%) of participants. A Kruskal-Wallis test, or chi-square test was used to compare each parameter between groups. ALT: alanine aminotransferase, AST: aspartate aminotransferase, BMI: body mass index, eGFR: estimated glomerular filtration rate, FPG: fasting plasma glucose, γ-GTP: γ-glutamyl transpeptidase, HDL: high density lipoprotein, LDL: low density lipoprotein, TG: triglycerides
Fasting proinsulin levels in participants grouped by severity of hepatic steatosis
| Fatty liver index | |||
|---|---|---|---|
| Model 1 | Low group (n = 110) | Moderate group (n = 75) | High group (n = 83) |
| Ln (Proinsulin + 1) | 1.93 (1.79, 2.07) | 2.39 (2.22, 2.56)** | 3.09 (2.93, 3.25)** ## |
| Model 2 | |||
| Ln (Proinsulin + 1) | 1.94 (1.80, 2.08) | 2.40 (2.23, 2.57)** | 3.05 (2.89, 3.22)** ## |
Values are normalized by natural logarithmic transformation and expressed as least squares means (95% confidence interval). Analysis of covariance and Tukey’s honestly significant difference tests were used to compare proinsulin levels between groups. Model 1: crude; Model 2: adjusted for age and sex.
** p < 0.01 vs. low group; ## p < 0.01 vs. moderate group
In the present study, we showed that fasting proinsulin levels were positively associated with fatty liver index in people with type 2 diabetes. This association persisted regardless of the presence or absence of obesity or the status of glucose management. Given that proinsulin is a marker of pancreatic beta cell dysfunction, this study confirms the association between beta cell dysfunction and hepatic steatosis in people with type 2 diabetes previously observed in population-based studies [10, 15].
Since it has been reported that the ratio of proinsulin/C-peptide is a sensitive biomarker of beta cell dysfunction [16, 17], we compared natural logarithm-transformed proinsulin/C-peptide ratios between the three groups. Similar to ln (Proinsulin + 1), a positive association was observed between ln (Proinsulin/C-peptide + 1) and fatty liver index (Supplementary Table 3). This result further supports our finding that beta cell dysfunction is associated with hepatic steatosis in people with type 2 diabetes.
The positive relationship between proinsulin and fatty liver we observed suggests the existence of hepato-pancreatic crosstalk. Although causality cannot be proven from the present study design, hepatic steatosis is hypothesized to cause beta cell dysfunction, possibly via humoral factors. Specifically, secretion of fetuin-A and S100A6 increases in hepatic steatosis, and these factors adversely affect pancreatic beta cell function [18, 19]. Conversely, excessive beta cell burden, indicated by elevated proinsulin, could have promoted hepatic steatosis through inappropriate hyperinsulinemia [20]. However, longitudinal studies are needed to clarify any causal relationship.
An association between proinsulin levels and hepatic steatosis was observed not only in the general population [10, 21] but also in people with type 2 diabetes. The finding of this association in people with type 2 diabetes is of mechanistic interest. Even in healthy beta cells, proinsulin secretion may be elevated when beta cells are stressed by the presence of hepatic steatosis, as described above. In people with type 2 diabetes, proinsulin levels are elevated because of an impairment in proinsulin processing [22]. Moreover, it is possible that proinsulin may be additively elevated in these people because of the presence of hepatic steatosis. This is interesting because our results suggest that there is a mechanism for this additive elevation of proinsulin in people with type 2 diabetes.
One limitation of this study is its low sample size, particularly in the high fatty liver index, BMI <25 group and the low fatty liver index, BMI ≥25 group, which may have influenced the presence or absence of statistical significance between some groups. Other limitations were that the fatty liver index was used as an indirect way to evaluate hepatic steatosis, and that only Japanese participants were included. Regarding the former, it has been reported that there is a strong correlation between fatty liver index and hepatocellular lipid content [23, 24]. Moreover, hepatic steatosis was evaluated by the hepatic steatosis index, which reflects the grade of fatty liver present [25]. As shown in Supplementary Table 4, a similar positive association was observed with a parameter other than the fatty liver index. More importantly, because this study is based on clinical practice data, consideration of the effects of antihyperglycemic agents is required. In our study, use of sodium-glucose cotransporter-2 inhibitors and glucagon-like peptide-1 receptor agonists were significantly different between the three groups (Supplementary Table 5). To rule out the effect of these agents, proinsulin levels were compared among participants not using these agents. The results were similar to the overall analysis (Supplementary Table 6).
In conclusion, a positive association between proinsulin levels and fatty liver index was observed in people with type 2 diabetes. This association suggests that hepato-pancreatic crosstalk is involved in the pathophysiology of type 2 diabetes; however, further studies are required to determine the exact nature of this relationship.
We thank Edanz (https://jp.edanz.com/ac) for editing a draft of this manuscript.
The authors declare no conflict of interest.
