Skip to main content

Plasma ceramides are associated with MRI-based liver fat content but not with noninvasive scores of liver fibrosis in patients with type 2 diabetes

Abstract

Background

There is growing evidence that ceramides play a significant role in the onset and progression of non-alcoholic fatty liver disease (NAFLD), a highly prevalent condition in patients with type 2 diabetes associated with hepatic and cardiovascular events. However, the relationship between plasma ceramide levels and NAFLD severity in type 2 diabetes remains unclear. The main purpose of the present study was to investigate whether circulating levels of ceramides in patients with type 2 diabetes are associated with liver steatosis assessed by the highly accurate magnetic resonance imaging proton density fat fraction (MRI-PDFF). The secondary objective was to assess the relationship between plasma ceramides and noninvasive scores of liver fibrosis.

Methods

In this cross-sectional single-center study, plasma concentrations of 7 ceramides were measured by liquid chromatography-mass spectrometry in 255 patients with type 2 diabetes (GEPSAD cohort). Liver fat content was assessed by MRI-PDFF, and noninvasive scores of liver fibrosis (i.e. Fibrosis-4 index, NAFLD Fibrosis Score, FibroTest® and Fibrotic NASH Index) were calculated. A validation cohort of 80 patients with type 2 diabetes was also studied (LIRA-NAFLD cohort).

Results

Liver steatosis, defined as a liver fat content > 5.56%, was found in 62.4 and 82.5% of individuals with type 2 diabetes in the GEPSAD and LIRA-NAFLD cohorts, respectively. In GEPSAD, MRI-PDFF-measured liver fat content was positively associated with plasma levels of total ceramides (r = 0.232, p = 0.0002), and 18:0, 20:0, 22:0 and 24:0 ceramides in univariate analysis (p ≤ 0.0003 for all). In multivariate analysis, liver fat content remained significantly associated with total ceramides (p = 0.001), 18:0 (p = 0.006), 22:0 (p = 0.0009) and 24:0 ceramides (p = 0.0001) in GEPSAD, independently of age, diabetes duration, body mass index and dyslipidemia. Overall, similar relationship between plasma ceramides and liver fat content was observed in the LIRA-NAFLD validation cohort. No significant association was found between plasma ceramides and noninvasive scores of fibrosis after adjustment for age in both cohorts.

Conclusions

Plasma ceramide levels are associated with liver steatosis in patients with type 2 diabetes, independently of traditional risk factors for NAFLD. The independent association between plasma ceramides and liver steatosis adds new insights regarding the relationship between ceramides and NAFLD in type 2 diabetes.

Background

Non-alcoholic fatty liver disease (NAFLD) is a very common condition in patients with type 2 diabetes, which is associated with an increased risk of adverse hepatic and cardiovascular events [1]. By definition, NAFLD is characterized by the accumulation of triglycerides in the liver, i.e. steatosis [2]. Actually, it encompasses a wide range of conditions, from simple steatosis or non-alcoholic fatty liver (NAFL) to non-alcoholic steatohepatitis (NASH) characterized by necroinflammation with or without fibrosis, which can progress to cirrhosis and even hepatocellular carcinoma [3, 4].

Ceramides are bioactive lipids involved in several mechanisms related to cardiometabolic diseases [5, 6]. In particular, growing evidence suggests that ceramides play a significant role in the onset and progression of NAFLD [7, 8]. In cellular and animal models, ceramides has been found to contribute to key mechanisms of NAFL and NASH, such as hepatic insulin resistance, endoplasmic reticulum stress and mitochondrial dysfunction, which ultimately lead to steatosis, inflammation and fibrosis [7, 8]. Interestingly, decreasing ceramide levels in rodents prevents the development of both liver steatosis and fibrosis [9,10,11]. In addition, clinical studies have shown that the hepatic ceramide content was associated with liver steatosis and NASH in obese individuals [12, 13].

The interest in circulating ceramide concentrations is currently increasing. Indeed, recent large-scale clinical studies have proposed circulating ceramide levels as biomarkers of cardiovascular events [14,15,16], including in studies specifically conducted in patients with type 2 diabetes [17]. Clinical data on the relationship between circulating ceramides and NAFLD are still limited and quite heterogeneous, whereas the association between the hepatic ceramide content and NAFLD is better documented [12, 13, 18]. While no association was found between plasma ceramides and liver steatosis or fibrosis in the CERADIAB French cohort of patients with type 2 diabetes [19], other cross-sectional studies have yielded more contrasted results [13, 20, 21]. For instance, some plasma ceramide species were associated with liver steatosis, but others not, in the large Dallas Heart Study, which enrolled nondiabetic individuals [20]. In addition, plasma ceramides were elevated in obese individuals with NASH compared those without NASH, but this association was lost after adjustment for age [21]. Lastly, it was recently reported that plasma ceramide correlates with the extent of liver steatosis in two cohorts containing about 20% of diabetic patients [12, 22].

Drawing a definitive conclusion based on the studies investigating the relationship between plasma ceramides and NAFLD is currently challenging, in particular because liver steatosis has been evaluated using either surrogate biomarkers, or computed tomography- or ultrasound-based methods. In fact, these procedures are consensually recognized to be less accurate than magnetic resonance imaging proton density fat fraction (MRI-PDFF), magnetic resonance spectroscopy (MRS) or liver histology [2, 23]. To the best of our knowledge, only two studies reported data on liver fat content measured by MR-based imaging or histology, but they included no or very few patients with type 2 diabetes, making generalizations to type 2 diabetes uncertain [12, 20].

Clarifying the relationship between plasma ceramides and NAFLD severity in patients with type 2 diabetes would add to our understanding of the interplay between ceramides and NAFLD in this population, especially in the era of antidiabetic drugs such as the GLP-1 receptor agonists that can influence plasma ceramide levels [8]. Therefore, the primary objective of the present study was to investigate whether plasma ceramide levels are related to the liver fat content determined by MRI-PDFF in two independent cohorts composed exclusively of patients with type 2 diabetes. The secondary purpose was to evaluate whether there is any relationship between plasma ceramides and noninvasive scores of liver fibrosis that are Fibrosis-4 index (FIB-4), NAFLD Fibrosis Score (NFS), FibroTest®, and Fibrotic NASH Index (FNI).

Methods

Study population and design

The present ancillary study included participants enrolled in the single-center GEnetic Polymorphisms, Steatosis and Diabetes (GEPSAD) study. The study complied with the Declaration of Helsinki, and was approved by our local research ethics committee. All patients gave written informed consent before study inclusion. Participants were recruited from October 2007 to September 2010 upon admission to the Department of Diabetology at the Dijon University Hospital (France). The inclusion criteria were: type 2 diabetes, duration of diabetes ≥ 2 years, age ≥ 18 years, and no concurrent acute or chronic disease. Patients were not eligible if they had causes of liver steatosis other than diabetes and overweight [alcohol consumption ≥ 20 g/day for women and ≥ 30 g/day for men, hepatitis B or C virus infection, or use of drugs known to precipitate steatosis (thiazolidinediones, corticosteroids, and immunosuppressants)] or contraindications to MRI (pacemaker, metallic implants, claustrophobia, or body weight > 150 kg).

As a validation cohort, we enrolled participants from the prospective single-center LIRA-NAFLD trial (ClinicalTrials.gov identifier: NCT02721888) [24]. The inclusion criteria were: poorly controlled type 2 diabetes [glycated hemoglobin A1c (HbA1c) > 7.0%), treatment with metformin and/or sulfonylurea and/or insulin. Exclusion criteria were estimated glomerular filtration rate (eGFR) < 30 mL/min/1.73m2, pregnancy, alcohol abuse, severe liver impairment defined as aspartate aminotransferase (AST) or alanine aminotransferase (ALT) levels > 3 times the upper limit of normal, treatment with dipeptidyl peptidase-4 inhibitors during the three previous months, or previous treatment with thiazolidinediones or glucagon-like peptide-1 receptor agonist.

For the calculation of sample size, it was assumed that the correlation between plasma ceramide levels and liver fat content is about 0.20 (or an r2 of 0.04). Under these assumptions, at least 194 patients were needed to demonstrate a significant effect of plasma ceramide on the liver fat content with an alpha risk of type I error of 5% (bilateral) and a power 1-β of 80%.

During the inclusion visit, patients had a detailed interview, physical examination, liver fat content assessment by MRI-PDFF and fasting blood samples in tubes containing in particular EDTA as preservative for lipidomic analyses. The plasma/serum was immediately separated by centrifugation and analyzed for routine parameters or frozen < -70 °C awaiting lipidomic analyses. Hypertension was defined as systolic blood pressure ≥ 140 mmHg or diastolic blood pressure ≥ 90 mmHg according to American Diabetes Association statement [25], or the use of antihypertensive therapy. Dyslipidemia was defined as triglycerides ≥ 1.70 mmol/L, or high-density lipoprotein (HDL) cholesterol ≤ 1.30 and 1.03 mmol/L for females and males, respectively, or the use of lipid-lowering medications.

Liver fat content assessment by MRI-PDFF

Liver fat content was assessed using triple-echo MRI-PDFF with a 3.0-Tesla Magnetom Trio whole-body system with total imaging matrix technology (Siemens, Erlangen, Germany), as previously described [26]. Hepatic steatosis was defined as a liver fat content > 5.56% of liver tissue weight (i.e. > 55.6 mg triglycerides/g of liver tissue) [27].

Noninvasive scores of liver fibrosis

Liver fibrosis was assessed using the noninvasive scores recommended by the European Association for the Study of the Liver (EASL) to rule-out advanced fibrosis in patients with NAFLD/NASH, namely FIB-4, NFS and FibroTest®. We also considered FNI as it recently was proved to be accurate [28]. FIB-4, NFS and FNI were calculated according to the original formulas: FIB-4 = (age [years] x AST [IU/L]) / ((platelet count [x109/L]) x (ALT [IU/L])1/2); NFS = -1.675 + (0.037 x age [years]) + (0.094 x BMI [kg/m2]) + 1.13 + (0.99 x AST/ALT ratio) – (0.013 x platelet count [x109/L]) – (0.66 x albumin [g/dL]); FNI = 𝑒(− 10.33 + 2.54 x ln AST [IU/L] + 3.86 x ln HbA1c [%] − 1.66 x ln HDL-cholesterol [mg/dL]) / (1 + 𝑒(− 10.33 + 2.54 x ln AST [IU/L] + 3.86 x ln HbA1c [%] − 1.66 x ln HDL-cholesterol [mg/dL])) [28,29,30]. The patented tests (FibroTest® and NashTest®) were performed only in the GEPSAD cohort, according to the manufacturer’s instructions (BioPredictive, Paris, France).

Plasma ceramide measurements by liquid chromatography—tandem mass spectrometry

Plasma levels of the seven main ceramide species with a d18:1 sphingoid backbone were determined as previously described [31]. Briefly, 100 µL plasma were mixed with d18:1/17:0 ceramide (Avanti Polar Lipids, Birmingham, AL, USA), used as internal standard, then extracted with 750 µL of 1:2 chloroform/methanol for 10 min. Chloroform (250 µL) was then added and extraction continued for 10 min more. Water (250 µL) was then added and extraction continued for 10 min more. After centrifugation (9,400 g, 5 min), the organic phase was collected. The aqueous phase was acidified with 8 µL of 3 mol/L of hydrochloric acid, and further extracted with 600 µL of chloroform for 10 min. After centrifugation (9,400 g, 5 min), the two organic phases were combined and washed with 800 µL of the upper phase from a chloroform/methanol/water (96.7:93.3:90) mixture. After centrifugation (9,400 g, 5 min), the organic phase was evaporated under vacuum. Extracts were finally dissolved with 100 µL of 60:30:4.5 chloroform/methanol/water. Three microliters of extract were injected on a 6460 MS/MS system (Agilent Technologies, Les Ulis, France). Separation was achieved on a Poroshell C8 column 2.1 × 100 mm, 2.7-µm (Agilent Technologies) using the following gradient conditions (0.3 mL/min): 1 min at 70% mobile phase B, 70–100% B over 4 min, and 5 min at 100% B. Mobile phases A and B consisted of water and methanol, respectively, both containing 1 mmol/L ammonium formate and 0.2% (v/v) formic acid. Individual ceramide species were identified using multiple-reaction monitoring (Supplementary Table 1). Calibration curves were obtained using external standards (Avanti Polar Lipids) for each ceramide molecular species. The total ceramide concentration was calculated by summing the individual ceramide species. A representative chromatogram is shown in Supplementary Fig.  1.

Routine laboratory measurements

Routine biochemical parameters [glucose, creatinine, total cholesterol, HDL cholesterol, triglycerides, AST, ALT, gamma-glutamyl transferase (GGT)] were determined on a Dimension Vista platform using dedicated reagents (Siemens, Saint-Denis, France). Low-density lipoprotein (LDL) cholesterol was estimated by the Friedewald formula when serum triglycerides were ≤ 3.8 mmol/L, or was measured using a direct method when serum triglycerides were > 3.8 mmol/L. eGFR was calculated according to the Modification of Diet in Renal Disease equation [32]. The triglyceride-glucose (TyG) index was calculated as defined in Fedchuck et al. [33]. HbA1c was measured by high performance liquid chromatography (Tosoh G8; Tosoh Bioscience, Tokyo, Japan).

Statistical analysis

Statistical analyses were performed using GraphPad Prism (version 9.1.1). Skewness and kurtosis were calculated for each continuous variable in order to evaluate whether the data distribution matched the Gaussian distribution. For instance, plasma ceramide levels and most of steatosis and fibrosis markers in the present study did not follow a Gaussian distribution, and this was not due to outliers (checked with the robust regression and outlier removal method). Data are reported as medians (interquartile range [IQR]) for non-normally distributed variables or means ± standard deviations (SD) for normally-distributed parameters. A log10 transformation was used for non-normally distributed variables in order to create a Gaussian distribution when necessary for further statistical analyses.

The comparison between two groups of participants was performed using the parametric unpaired Student t test (for normally-distributed variables) and the non-parametric unpaired Mann-Whitney U test (for non-normally distributed variables or when sample size < 30 individuals in a subgroup). Proportions between two groups were compared using the Pearson’s Chi-squared test with Yates’s continuity correction.

For association analysis, univariate Pearson correlation coefficients (r) were determined for continuous variables (using raw data for normally-distributed variables and log10-transformed data for non-normally distributed variables). Point biserial correlation was used to assess the correlations between continuous variables and categorical variables. A Bonferroni correction procedure was applied for controlling false positives in multiple testing, by dividing the desired alpha-level (i.e. 0.05) by the number of comparisons. Multivariate analyses were performed by multivariate linear or logistic regression for continuous and categorical dependent variables, respectively.

A two-tailed probability level of 0.05 was considered statistically significant for all statistical analyses.

Results

Participant characteristics

On the 298 participants with type 2 diabetes enrolled in the GEPSAD cohort, 29 and 14 individuals were excluded from the present study due to the lack of MRI-PDFF and ceramide measurements, respectively. This resulted in a total of 255 participants, whose characteristics are presented in Table 1. Participants exhibited typical features of type 2 diabetes, such as increased body mass index (> 25 and > 30 kg/m2 in 95.3 and 70.6% of patients, respectively), elevated triglyceride levels (≥ 1.70 mmol/L in 50.2% of patients), low plasma HDL-cholesterol concentrations (≤ 1.30 mmol/L in 72.9% of females and ≤ 1.03 mmol/L in 62.4% of males), and elevated TyG index (> 8.38 in 94.5% of participants). AST and ALT ≥ 30 IU/L were found in 64 (25.1%) and 143 (56.1%) participants, respectively.

We also studied another cohort of patients with type 2 diabetes as a validation cohort. The 80 participants were initially enrolled in the LIRA-NAFLD trial, and their main characteristics are presented in Supplementary Table  2.

Table 1 Patient characteristics in GEPSAD

Liver fat content assessment by MRI-PDFF

Liver steatosis, defined as a liver fat content > 5.56% using MRI-PDFF, was found in 159 (62.4%) and 66 (82.5%) participants with type 2 diabetes in GEPSAD and LIRA-NAFLD, respectively. As shown in Table 1 and Supplementary Tables  2, individuals with liver steatosis had higher body mass index (and plasma ALT than individuals without liver steatosis in both cohorts, while levels of HbA1c were similar (p = 0.37 and 0.85 in GEPSAD and LIRA-NAFLD, respectively).

Noninvasive scores of liver fibrosis

Noninvasive scores of liver fibrosis are shown in Table 2 and Supplementary Table  4 for GEPSAD and LIRA-NAFLD, respectively. About three quarters of participants in both cohorts were considered at a low risk of having advanced fibrosis based on the FIB-4 threshold of 1.30 recommended in the European and US guidelines [34, 35]. This increased to 90% in GEPSAD according to the FibroTest® score (< 0.48).

Table 2 Liver fat content and noninvasive scores of liver fibrosis in GEPSAD

Plasma ceramides and liver steatosis

The main purpose of the study was to investigate the relationship between plasma ceramide levels and liver fat content determined by MRI-PDFF in patients with type 2 diabetes. In GEPSAD, participants with liver steatosis according to MRI-PDFF had higher circulating concentrations of total ceramides (median [IQR], 5.53 [4.31–7.52] vs. 5.07 [3.86–6.46] µmol/L, p = 0.02), 18:0 ceramide (0.126 [0.088–0.164] vs. 0.071 [0.056–0.098] µmol/L, p = 0.009), 20:0 ceramide (0.084 [0.058–0.107] vs. 0.071 [0.056–0.098] µmol/L, p = 0.05), 22:0 ceramide (0.595 [0.446–0.809] vs. 0.544 [0.361–0.706] µmol/L, p = 0.01) and 24:0 ceramide (3.12 [2.25–4.27] vs. 2.78 [1.95–3.64] µmol/L, p = 0.01) compared to the patients without liver steatosis.

Figure 1 shows the univariate correlations between plasma ceramides and liver fat content assessed by MRI-PDFF. After Bonferroni adjustment, the liver fat content remained positively associated with plasma total ceramides (r = 0.232, p = 0.0002), 18:0 (r = 0.241, p = 0.0001), 20:0 (r = 0.183, p = 0.0003), 22:0 (r = 0.255, p < 0.0001) and 24:0 ceramides (r = 0.256, p < 0.0001) in GEPSAD. We were interested in grouping together ceramide species with very long-chain saturated fatty acids (VLSFA), i.e. with at least 20 carbon atoms, because they are linked to health outcomes in epidemiological studies in a different manner than the ceramides with long-chain fatty acids (LCFA, i.e. the 16:0 and 18:0 ceramides) [36]. In GEPSAD, the liver fat content correlated with the plasma concentration of ceramides containing VLSFA (r = 0.264, p < 0.0001), but not with ceramides containing LCFA (r = 0.096, p = 0.127).

Not surprisingly, liver fat content correlated with age, diabetes duration, body mass index and dyslipidemia in univariate analysis (data not shown). Using a multivariate linear regression model including all these variables (Table 3), the liver fat content measured by MRI-PDFF remained independently associated with plasma total ceramides in GEPSAD [standardized β (95% confidence intervals (CI)), 0.201 (0.082–0.320), p = 0.001]. Using a multivariate logistic regression approach, plasma levels of total ceramides were independently linked to the presence of liver steatosis, i.e. with liver fat content > 5.56% [odds-ratio (OR) (95% CI), 8.2 (1.6–46), p = 0.014)] (Supplementary Table  3).

Fig. 1
figure 1

Univariate correlations between plasma ceramide concentrations and hepatic markers of NAFLD in GEPSAD and LIRA-NAFLD (shown as a heatmap). Only correlations remaining statistically significant after Bonferroni correction were reported in the heatmap (i.e. when p < 0.0063). Liver fat content corresponds to the value determined by MRI-PDFF. FibroTest® and NashTest® were not performed in LIRA-NAFLD (grey cells). Abbreviations: FIB-4, fibrosis-4 index; FNI, fibrotic NASH index; LCFA, long-chain fatty acid; NFS, NAFLD fibrosis score; VLSFA, very long-chain saturated fatty acid

Table 3 Multivariate linear regression analysis for MRI-PDFF-based liver fat content (log)

Regarding ceramide molecular species in multivariate linear regression analysis (Table 3), the liver fat content measured by MRI-PDFF was independently associated with the plasma concentrations of 18:0 ceramide [standardized β (95% CI), 0.170 (0.050–0.290), p = 0.006], 20:0 ceramide [0.138 (0.019–0.257), p = 0.02], 22:0 ceramide [0.203 (0.084–0.322), p = 0.0009] and 24:0 ceramide [0.235 (0.117–0.353), p = 0.0001] in GEPSAD. The plasma concentration of ceramides with VLSFA was independently linked to the liver fat content, while the plasma concentration of ceramides with LCFA were not.

We aimed to confirm these results using the LIRA-NAFLD validation cohort, composed of 80 patients with type 2 diabetes. Similarly to GEPSAD, the liver fat content measured by MRI-PDFF in LIRA-NAFLD was also positively associated with the plasma concentration of total ceramides (r = 0.244, p = 0.029), 18:0 ceramide (r = 0.293, p = 0.008), 22:0 ceramide (r = 0.320, p = 0.003), 24:0 ceramide (r = 0.253, p = 0.016) and ceramides containing VLSFA (r = 0.280, p = 0.012) in univariate analysis (Fig. 1). In multivariate analysis, liver fat content remained independently associated with plasma total ceramides, 18:0 ceramide, 22:0 ceramide, 24:0 ceramide, and with ceramides containing VLSFA (Table 3).

Plasma ceramides and noninvasive scores of liver fibrosis

The secondary objective of the study was to investigate the relationship between plasma ceramides and noninvasive scores of liver fibrosis. The participants were divided into two groups according to the thresholds of noninvasive scores usually recommended to rule-out significant fibrosis (27,31). In GEPSAD, the circulating concentrations of total ceramides were higher in participants with FIB-4 < 1.30 than in those with FIB-4 ≥ 1.30 (median [IQR], 5.53 [4.36–7.42] vs. 4.95 [3.77–6.41] µmol/L, p = 0.04), and were similar between participants with NFS < or ≥ -1.455 (6.04 [4.58–7.47] vs. 5.36 [4.17–7.10] µmol/L, p = 0.28), FibroTest® score ≤ or > 0.48 (5.40 [4.18–7.24] vs. 4.91 [4.14–7.08] µmol/L, p = 0.38) and FNI < or ≥ 0.10 (5.55 [4.40–6.59] vs. 5.33 [4.17–7.28] µmol/L, p = 0.83).

As shown in Fig. 1, few correlations between plasma ceramides and fibrosis scores remained significant after Bonferroni correction in univariate regression in both cohorts. Age was strongly associated with all scores of fibrosis in both cohorts (p < 0.0001 for all). Using multivariate linear regression analysis, the plasma concentration of total ceramides was not significantly associated with FIB-4 [standardized β (95% CI), -0.109 (-0.225–0.006), p = 0.06], NFS [-0.077 (-0.203–0.049), p = 0.23], FibroTest® [0.063 (-0.051–0.176), p = 0.28] and FNI [0.004 (-0.122–0.131), p = 0.95] in GEPSAD after adjustment for age. Similar results were obtained using a multivariate logistic regression approach by considering thresholds usually recommended to rule-out significant fibrosis (Table 4) [28, 34]. Since preliminary evidence suggests that FIB-4 outperforms NFS and FibroTest® to rule in advanced fibrosis in individuals with type 2 diabetes [34, 37], we also assessed the relationship between plasma ceramides and FIB-4 at the rule-in threshold for advanced fibrosis using multivariate analysis. Thus, no plasma ceramide level was associated with a FIB-4 ≥ 2.67 in GEPSAD after adjustment for age (e.g., estimate [95% CI], OR = 0.024 [3 × 10− 4–1.70], p = 0.09 for total ceramides).

Table 4 Multivariate logistic regression analysis for noninvasive scores of fibrosis

In the LIRA-NAFLD validation cohort, the plasma concentration of total ceramides was also not associated with the FIB-4 [standardized β (95% CI), -0.092 (-0.276–0.093), p = 0.32], NFS [1.5 × 10− 5 (-3.4 × 10− 4–3.7 × 10− 4), p = 0.93], and FNI [1.0 × 10− 5 (-1.3 × 10− 5–3.3 × 10− 5), p = 0.41] scores after adjustment for age. Overall, the multivariate logistic regression approach yielded similar findings than in GEPSAD using the rule-out thresholds for advanced fibrosis (Table 4).

Discussion

Here, our key finding was that plasma ceramide levels correlate positively with the degree of liver steatosis, namely with the liver fat content assessed by MRI-PDFF expressed as percentage, in two independent cohorts including overall 335 individuals with type 2 diabetes. The association between plasma ceramides and liver fat content was independent of traditional risk factors for NAFL. Interestingly, an Australian study has also shown that plasma ceramide levels are linked to the extent of liver steatosis using histology, which is another reference method [12]. But, this study was conducted in obese patients of which only 22% had type 2 diabetes, making generalization to all patients with type 2 diabetes uncertain. Our results suggest that this link also exists in patients with type 2 diabetes. Another clinical study, conducted in 149 patients with type 2 diabetes, found a positive association between plasma total ceramides and liver steatosis [19], but using surrogate biomarkers of steatosis (i.e. SteatoTest® and fatty liver index) that are extensively recognized as being less accurate than MR-based imaging methods [34]. In contrast to our results, Apostolopoulou et al. found that the circulating levels of total ceramides were not significantly higher in obese individuals with biopsy-proven liver steatosis than in those without liver steatosis, but this study included only 7 individuals in each group [13].

Regarding ceramide molecular species, we found an independent association between liver fat content and the plasma levels of the 18:0, 20:0, 22:0 and 24:0 ceramide species. Similar results were also observed in the Australian study mentioned above, which was conducted in obese participants including 22% with type 2 diabetes [12]. A US study found that plasma 18:0 ceramide was independently associated with the extent of steatosis, but the 20:0, 22:0 and 24:0 species were not [22]. We hypothesized that the difference with our results regarding these ceramides with VLSFA could be due to the fact that liver steatosis was assessed by non-contrast computed tomography, which is a less accurate method than MRI-PDFF, and perhaps also because of a particular study population of Asian individuals living in USA. In contrast to our results, the Dallas Heart Study found that the plasma level of 18:0 ceramide was not associated with MRS-measured liver fat content in a large population without diabetes [20]. However, the link between plasma ceramides and liver steatosis can be quite different in patients with type 2 diabetes since the plasma concentration of 18:0 ceramide is largely increased in this population [38,39,40]. Overall, we found an independent association in our two cohorts between liver fat content and plasma ceramides with VLSFA. Interestingly, decreasing specifically these ceramide species in mice reduces liver steatosis [41].

Our secondary aim was to investigate the link between plasma ceramides and noninvasive scores of liver fibrosis, since cellular and animal studies suggest that ceramides could play a role in hepatic fibrogenesis [9]. Overall, we found no association between plasma ceramide levels and noninvasive scores of liver fibrosis in both cohorts. In particular, this was confirmed using the FIB-4, FibroTest® and NFS scores, which are all recognized as useful for ruling-out advanced fibrosis in guidelines [2, 34]. Other clinical studies on ceramides and liver fibrosis are sparse, but a lack of relationship between plasma ceramides and FibroTest® has been already observed in two French cohorts of patients with type 2 diabetes with fewer participants than in our study [19]. In a context other than type 2 diabetes, it has been demonstrated that plasma ceramide concentrations are not linked to severe fibrosis stage assessed by liver biopsy in individuals with chronic hepatitis C [42]. However, it has been shown that plasma ceramides decreased with the severity of liver cirrhosis, but in a cirrhotic population including only 2% patients with NAFLD [43]. Overall, our results suggest that plasma ceramides are not useful as clinical biomarkers of liver fibrosis in patients with type 2 diabetes.

The main strength of our study is that it reports on the relationship between plasma ceramides and the extent of liver steatosis using the accurate MRI-PDFF procedure, and also that it focuses specifically on a population with type 2 diabetes. MRI-PDFF is recognized by the EASL to be the most accurate non-invasive method for detecting and quantifying steatosis [34]. However, our study has limitations that are worth discussing. Firstly, our two cohorts are relatively small, which can diminish the power of our correlations. Secondly, medications for management of diabetes and dyslipidemia may change ceramide metabolism and subsequently the relationship between plasma ceramides and liver outcomes. Although the effects of lipid-lowering drugs on ceramides have not been studied extensively, it has been reported that statins reduce plasma ceramide levels, unlike ezetimibe [44, 45]. However, introducing the use of statins in our multivariate analysis did not significantly change the association between plasma ceramide levels and liver fat content (data not shown), suggesting that statin use was not a confounding factor in our study. Thirdly, the lack of liver biopsy or magnetic resonance elastography to diagnose fibrosis should be mentioned as a limitation for accurately investigating the relationship between fibrosis and plasma ceramides. However, the noninvasive scores of liver fibrosis used in the present study are well-validated, and the guidelines consider that they are sufficiently accurate to rule-out advanced fibrosis in individuals with NAFLD [2, 34]. Lastly, the GEPSAD study has a cross-sectional design, and as such, demonstrates correlation between plasma ceramides and liver fat content, but not causation.

Conclusions

Plasma levels of ceramides are associated with the degree of liver steatosis in patients with type 2 diabetes, but not with noninvasive scores of liver fibrosis. Our results help to clarify the relationship between plasma ceramides and NAFLD severity, and therefore add to the knowledge of the interplay between ceramides and NAFLD in type 2 diabetes. However, longitudinal studies are needed to determine whether plasma ceramide levels can predict the progression to advanced stages of NAFLD in patients with type 2 diabetes.

Data availability

The datasets supporting the conclusions of this article are available from the corresponding author on reasonable request.

Abbreviations

ALT:

alanine aminotransferase

AST:

aspartate aminotransferase

CI:

confidence interval

EASL:

European Association for the Study of the Liver

EDTA:

ethylenediaminetetraacetic acid

eGFR:

estimated glomerular filtration rate

FIB-4:

fibrosis-4 index

FNI:

fibrotic NASH index

GEPSAD:

genetic polymorphisms, steatosis and diabetes

GGT:

gamma-glutamyl transferase

GLP-1:

glucagon-like peptide-1

HDL:

high-density lipoprotein

IQR:

interquartile range

LCFA:

long-chain fatty acid

LFC:

liver fat content

LDL:

low-density lipoprotein

MR:

magnetic resonance

MRI-PDFF:

magnetic resonance imaging proton density fat fraction

MRS:

magnetic resonance spectroscopy

NAFL:

non-alcoholic fatty liver

NAFLD:

non-alcoholic fatty liver disease

NASH:

non-alcoholic steatohepatitis

NFS:

NAFLD fibrosis score

OR:

odds-ratio

SD:

standard deviation

TyG:

triglyceride-glucose

VLSFA:

very long-chain saturated fatty acid

References

  1. Lomonaco R, Godinez Leiva E, Bril F, Shrestha S, Mansour L, Budd J, et al. Advanced liver fibrosis is common in patients with type 2 Diabetes followed in the outpatient setting: the need for systematic screening. Diabetes Care. 2021;44:399–406.

    Article  PubMed  PubMed Central  Google Scholar 

  2. European Association for the Study of the Liver (EASL), European Association for the Study of Diabetes (EASD). European Association for the Study of Obesity (EASO). EASL–EASD–EASO clinical practice guidelines for the management of non-alcoholic fatty Liver Disease. Diabetologia. 2016;59:1121–40.

    Article  Google Scholar 

  3. Machado MV, Diehl AM. Pathogenesis of nonalcoholic steatohepatitis. Gastroenterology. 2016;150:1769–77.

    Article  CAS  PubMed  Google Scholar 

  4. Diehl AM, Day C. Cause, pathogenesis, and treatment of nonalcoholic steatohepatitis. N Engl J Med. 2017;377:2063–72.

    Article  CAS  PubMed  Google Scholar 

  5. Choi RH, Tatum SM, Symons JD, Summers SA, Holland WL. Ceramides and other sphingolipids as drivers of Cardiovascular Disease. Nat Rev Cardiol. 2021;18:701–11.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  6. Green CD, Maceyka M, Cowart LA, Spiegel S. Sphingolipids in metabolic Disease: the good, the bad, and the unknown. Cell Metab. 2021;33:1293–306.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  7. Hajduch E, Lachkar F, Ferré P, Foufelle F. Roles of ceramides in non-alcoholic fatty Liver Disease. J Clin Med. 2021;10:792.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  8. Zhu C, Huai Q, Zhang X, Dai H, Li X, Wang H. Insights into the roles and pathomechanisms of ceramide and sphigosine-1-phosphate in nonalcoholic fatty Liver Disease. Int J Biol Sci. 2023;19:311–30.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  9. Jiang M, Li C, Liu Q, Wang A, Lei M. Inhibiting ceramide synthesis attenuates hepatic steatosis and fibrosis in rats with non-alcoholic fatty Liver Disease. Front Endocrinol. 2019;10:1–8.

    Article  Google Scholar 

  10. Chaurasia B, Tippetts TS, Monibas RM, Liu J, Li Y, Wang L, et al. Targeting a ceramide double bond improves insulin resistance and hepatic steatosis. Science. 2019;365:386–92.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  11. Xia JY, Holland WL, Kusminski CM, Sun K, Sharma AX, Pearson MJ, et al. Targeted induction of ceramide degradation leads to improved systemic metabolism and reduced hepatic steatosis. Cell Metab. 2015;22:266–78.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  12. Ooi GJ, Meikle PJ, Huynh K, Earnest A, Roberts SK, Kemp W, et al. Hepatic lipidomic remodeling in severe obesity manifests with steatosis and does not evolve with non-alcoholic steatohepatitis. J Hepatol. 2021;75:524–35.

    Article  CAS  PubMed  Google Scholar 

  13. Apostolopoulou M, Gordillo R, Koliaki C, Gancheva S, Jelenik T, De Filippo E, et al. Specific hepatic sphingolipids relate to insulin resistance, oxidative stress, and inflammation in nonalcoholic steatohepatitis. Diabetes Care. 2018;41:1235–43.

    Article  CAS  PubMed  Google Scholar 

  14. Wang DD, Toledo E, Hruby A, Rosner BA, Willett WC, Sun Q, et al. Plasma ceramides, mediterranean diet, and incident Cardiovascular Disease in the PREDIMED trial. Circulation. 2017;135:2028–40.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  15. Laaksonen R, Ekroos K, Sysi-Aho M, Hilvo M, Vihervaara T, Kauhanen D, et al. Plasma ceramides predict cardiovascular death in patients with stable coronary artery Disease and acute coronary syndromes beyond LDL-cholesterol. Eur Heart J. 2016;37:1967–76.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  16. Katajamäki TT, Koivula M-K, Hilvo M, Lääperi MTA, Salminen MJ, Viljanen AM, et al. Ceramides and phosphatidylcholines associate with Cardiovascular Diseases in the elderly. Clin Chem. 2022;68:1502–8.

    Article  PubMed  Google Scholar 

  17. Alshehry ZH, Mundra PA, Barlow CK, Mellett NA, Wong G, McConville MJ, et al. Plasma lipidomic profiles improve on traditional risk factors for the prediction of cardiovascular events in type 2 Diabetes Mellitus. Circulation. 2016;134:1637–50.

    Article  CAS  PubMed  Google Scholar 

  18. Béland-Bonenfant S, Rouland A, Petit J-M, Vergès B. Concise review of lipidomics in nonalcoholic fatty Liver Disease. Diabetes Metab. 2023;49:101432.

    Article  PubMed  Google Scholar 

  19. Carlier A, Phan F, Szpigel A, Hajduch E, Salem JE, Gautheron J, et al. Dihydroceramides in triglyceride-enriched VLDL are associated with nonalcoholic fatty Liver Disease severity in type 2 Diabetes. Cell Rep Med. 2020;1:100154.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  20. Neeland IJ, Singh S, McGuire DK, Vega GL, Roddy T, Reilly DF, et al. Relation of plasma ceramides to visceral adiposity, insulin resistance and the development of type 2 Diabetes Mellitus: the Dallas Heart Study. Diabetologia. 2018;61:2570–9.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  21. Anjani K, Lhomme M, Sokolovska N, Poitou C, Aron-Wisnewsky J, Bouillot J-L, et al. Circulating phospholipid profiling identifies portal contribution to NASH signature in obesity. J Hepatol. 2015;62:905–12.

    Article  CAS  PubMed  Google Scholar 

  22. Gadgil MD, Sarkar M, Sands C, Lewis MR, Herrington DM, Kanaya AM. Associations of NAFLD with circulating ceramides and impaired glycemia. Diabetes Res Clin Pract. 2022;186:109829.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  23. Lee SS, Park SH, Kim HJ, Kim SY, Kim M-Y, Kim DY, et al. Non-invasive assessment of hepatic steatosis: prospective comparison of the accuracy of imaging examinations. J Hepatol. 2010;52:579–85.

    Article  CAS  PubMed  Google Scholar 

  24. Petit JM, Cercueil JP, Loffroy R, Denimal D, Bouillet B, Fourmont C, et al. Effect of liraglutide therapy on liver fat content in patients with inadequately controlled type 2 Diabetes: the Lira-NAFLD study. J Clin Endocrinol Metab. 2017;102:407–15.

    PubMed  Google Scholar 

  25. American Diabetes Association Professional Practice Committee. 10. Cardiovascular disease and risk management: standards of medical care in diabetes-2022. Diabetes Care. 2022;45:S144–74.

  26. Guiu B, Petit J-M, Loffroy R, Ben Salem D, Aho S, Masson D, et al. Quantification of liver fat content: comparison of triple-echo chemical shift gradient-echo imaging and in vivo proton MR spectroscopy. Radiology. 2009;250:95–102.

    Article  PubMed  Google Scholar 

  27. Szczepaniak LS, Nurenberg P, Leonard D, Browning JD, Reingold JS, Grundy S, et al. Magnetic resonance spectroscopy to measure hepatic triglyceride content: prevalence of hepatic steatosis in the general population. Am J Physiol Endocrinol Metab. 2005;288:E462–468.

    Article  CAS  PubMed  Google Scholar 

  28. Tavaglione F, Jamialahmadi O, De Vincentis A, Qadri S, Mowlaei ME, Mancina RM, et al. Development and validation of a score for fibrotic nonalcoholic steatohepatitis. Clin Gastroenterol Hepatol. 2023;21:1523–1532e1.

    Article  CAS  PubMed  Google Scholar 

  29. Angulo P, Hui JM, Marchesini G, Bugianesi E, George J, Farrell GC, et al. The NAFLD fibrosis score: a noninvasive system that identifies liver fibrosis in patients with NAFLD. Hepatology. 2007;45:846–54.

    Article  CAS  PubMed  Google Scholar 

  30. Sterling RK, Lissen E, Clumeck N, Sola R, Correa MC, Montaner J, et al. Development of a simple noninvasive index to predict significant fibrosis in patients with HIV/HCV coinfection. Hepatology. 2006;43:1317–25.

    Article  CAS  PubMed  Google Scholar 

  31. Tran TTT, Postal BG, Demignot S, Ribeiro A, Osinski C, Pais de Barros J-P, et al. Short-term palmitate supply impairs intestinal insulin signaling via ceramide production. J Biol Chem. 2016;291:16328–38.

    Article  PubMed  PubMed Central  Google Scholar 

  32. Levey AS, Bosch JP, Lewis JB, Greene T, Rogers N, Roth D. A more accurate method to estimate glomerular filtration rate from serum creatinine: a new prediction equation. Modification of Diet in Renal Disease Study Group. Ann Intern Med. 1999;130:461–70.

    Article  CAS  PubMed  Google Scholar 

  33. Fedchuk L, Nascimbeni F, Pais R, Charlotte F, Housset C, Ratziu V, et al. Performance and limitations of steatosis biomarkers in patients with nonalcoholic fatty Liver Disease. Aliment Pharmacol Ther. 2014;40:1209–22.

    Article  CAS  PubMed  Google Scholar 

  34. European Association for the Study of the Liver. Clinical Practice Guideline Panel, Chair:, EASL Governing Board representative:, panel members: EASL Clinical Practice guidelines on non-invasive tests for evaluation of Liver Disease severity and prognosis – 2021 update. J Hepatol. 2021;75:659–89.

    Article  Google Scholar 

  35. ElSayed NA, Aleppo G, Aroda VR, Bannuru RR, Brown FM, Bruemmer D, et al. 4. Comprehensive medical evaluation and assessment of comorbidities: standards of care in diabetes-2023. Diabetes Care. 2023;46:49–67.

    Article  Google Scholar 

  36. Lemaitre RN, King IB. Very long-chain saturated fatty acids and Diabetes and Cardiovascular Disease. Curr Opin Lipidol. 2022;33:76–82.

    Article  CAS  PubMed  Google Scholar 

  37. Boursier J, Canivet CM, Costentin C, Lannes A, Delamarre A, Sturm N, et al. Impact of type 2 Diabetes on the accuracy of noninvasive tests of liver fibrosis with resulting clinical implications. Clin Gastroenterol Hepatol. 2023;21:1243–1251e12.

    Article  CAS  PubMed  Google Scholar 

  38. Haus JM, Kashyap SR, Kasumov T, Zhang R, Kelly KR, Defronzo RA, et al. Plasma ceramides are elevated in obese subjects with type 2 Diabetes and correlate with the severity of insulin resistance. Diabetes. 2009;58:337–43.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  39. Zhang L, Hu Y, An Y, Wang Q, Liu J, Wang G. The changes of lipidomic profiles reveal therapeutic effects of exenatide in patients with type 2 Diabetes. Front Endocrinol. 2022;13:677202.

    Article  Google Scholar 

  40. Denimal D, Bergas V, Pais-de-Barros J-P, Simoneau I, Demizieux L, Passilly-Degrace P, et al. Liraglutide reduces plasma dihydroceramide levels in patients with type 2 Diabetes. Cardiovasc Diabetol. 2023;22:104.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  41. Park W-J, Park J-W, Merrill AH, Storch J, Pewzner-Jung Y, Futerman AH. Hepatic fatty acid uptake is regulated by the sphingolipid acyl chain length. Biochim Biophys Acta. 2014;1841:1754–66.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  42. Li J-F, Qu F, Zheng S-J, Ren F, Wu H-L, Liu M, et al. Plasma sphingolipids: potential biomarkers for severe hepatic fibrosis in chronic Hepatitis C. Mol Med Rep. 2015;12:323–30.

    Article  CAS  PubMed  Google Scholar 

  43. Grammatikos G, Ferreiròs N, Waidmann O, Bon D, Schroeter S, Koch A, et al. Serum sphingolipid variations associate with hepatic decompensation and survival in patients with Cirrhosis. PLoS ONE. 2015;10:e0138130.

    Article  PubMed  PubMed Central  Google Scholar 

  44. Tarasov K, Ekroos K, Suoniemi M, Kauhanen D, Sylvänne T, Hurme R, et al. Molecular lipids identify cardiovascular risk and are efficiently lowered by simvastatin and PCSK9 deficiency. J Clin Endocrinol Metab. 2014;99:E45–52.

    Article  PubMed  Google Scholar 

  45. Ng TWK, Ooi EMM, Watts GF, Chan DC, Weir JM, Meikle PJ, et al. Dose-dependent effects of rosuvastatin on the plasma sphingolipidome and phospholipidome in the metabolic syndrome. J Clin Endocrinol Metab. 2014;99:E2335–40.

    Article  CAS  PubMed  Google Scholar 

Download references

Acknowledgements

This study was supported by the Department of Clinical Research of the University Hospital of Dijon Bourgogne. The authors thank Suzanne Rankin (Department of Clinical Research, University Hospital of Dijon Bourgogne) for the language revision of the manuscript, and are indebted to Isabelle Simoneau, Philippe Loisel and Véronique Grivet for their help during the recruitment of participants (all from the Clinical Research Unit of the department of Endocrinology and Diabetology, University Hospital of Dijon Bourgogne).

Funding

This work was supported by a grant from the University Hospital of Dijon Bourgogne (AOI 2010).

Author information

Authors and Affiliations

Authors

Contributions

D.D., B.V. and J-M.P. were involved in the conception, design, and conduct of the study. A.R., B.B., B.V. and J-M.P. participated in the recruitment of patients. D.D, J-P.d.B, L.D. and J-M.P. participated in the acquisition of data and analysis. D.D, S.B.B. and J-M.P. interpreted data. D.D. wrote the first draft of the manuscript, and all authors edited, reviewed, and approved the final version of the manuscript. D.D. is the guarantor of this work and, as such, had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Corresponding author

Correspondence to Damien Denimal.

Ethics declarations

Ethics approval and consent to participate

The study complies with the Declaration of Helsinki and was approved by our regional ethics committee.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.

Additional information

Publisher’s Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Electronic supplementary material

Below is the link to the electronic supplementary material.

Supplementary Table 1

. Quantified ceramide species and conditions for mass spectrometry acquisition.

Supplementary Table 2

. Patient characteristics in LIRA-NAFLD.

Supplementary Table 3

. Multivariate logistic regression analysis for liver fat content > 5.56%.

Supplementary Table 4

. Liver fat content and noninvasive scores of liver fibrosis in LIRA-NAFLD.

Supplementary Fig. 1

. A representative chromatogram of external and internal standards.

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data.

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Denimal, D., Béland-Bonenfant, S., Pais-de-Barros, JP. et al. Plasma ceramides are associated with MRI-based liver fat content but not with noninvasive scores of liver fibrosis in patients with type 2 diabetes. Cardiovasc Diabetol 22, 310 (2023). https://doi.org/10.1186/s12933-023-02049-2

Download citation

  • Received:

  • Accepted:

  • Published:

  • DOI: https://doi.org/10.1186/s12933-023-02049-2

Keywords