Sphingolipid profiling as a biomarker of type 2 diabetes risk: evidence from the MIDUS and PREDIMED studies

Abstract

Background: Type 2 diabetes (T2D) has become a worldwide pandemic. While ceramides may serve as intermediary between obesity-related lipotoxicity and T2D, the relationship with simple glycosphingolipids remains uncertain. The aim of this study was to characterize the associations between blood glycosphingolipid and ceramide species with T2D and to identify a circulating sphingolipid profile that could serve as novel biomarker for T2D risk.

Methods: Cross-sectional relationship between sphingolipid levels, insulin resistance, and T2D prevalence were evaluated in 2,072 American adults from MIDUS cohort. Prospectively, the association between sphingolipid species and the incidence of T2D was analyzed using a case-cohort design nested within the PREDIMED trial (250 cases and a random sample of 692 participants, with 3.8 years of median follow-up). Circulating levels of sphingolipid species in both populations were measured using LC/MS. Hazard ratios were estimated with weighted Cox regression models using Barlow weights.

Results: In American adults, only CER18:0 and CER22:0 were linked to insulin resistance and a higher prevalence of T2D. Conversely, three lactosylceramides (LCER 14:0, 16:0, and 24:1) showed a strong inverse relationship with both insulin resistance and T2D. These findings led to development of two sphingolipid scores. In the prospective analysis, these scores consistently predicted a reduced risk of T2D incidence in PREDIMED (HR: 0.64, 95% CI 0.44 to 0.94 and 0.58, 0.40 to 0.85 respectively) between extreme quartiles, with 5-year absolute risk differences of 9.6% (95% CI: 0.3-20.5%) and 11.4% (1.0-21.6%). They were validated in the same trial with samples obtained after 1 year of follow-up.

Conclusions: Our findings support the potential usefulness of circulating sphingolipid profiles as novel biomarkers for T2D risk. Moreover, this study opens the door for future research on the predictive value and possible protective roles of lactosylceramides in T2D.

Keywords: Ceramides; Insulin resistance; Lactosylceramides; Sphingolipids; Type 2 diabetes.

Fig. 1

Association of blood sphingolipid levels, insulin resistance and T2D prevalence in MIDUS. A Linear regression model for the association between sphingolipid levels and HOMA-IR. Model 1: bivariate, Model 2: adjusted for sample of origin, race, sex, and age, Model 3: Model 2 plus BMI; and a Model 4: Model 3 plus statin use. Color label represents the standardized regression coefficient (positive in red, and negative in blue) of those associations. Asterisks denote significant associations with a Bonferroni-adjusted p < 0.05. B Odds ratios for T2D (with 95%CI) per SD in blood levels of sphingolipid species, adjusted for sample of origin, race, sex, age, BMI, and statin use. C Odds ratios for T2D (with 95%CI) per increase in serum levels of sphingolipid classes as well as a set of selected species and LCER/CER22 ratio, adjusting for sample, race, sex, age, BMI, and statin use

Fig. 2

Estimates of the cumulative incidence of T2D based on quartiles of selected sphingolipid species in PREDIMED. Quartiles of A baseline levels (T0) of the sum of selected LacCer species (LCER14:0, LCER16:0, and LCER24:1), B baseline levels (T0) of the ratio of the selected lactosylceramides to CER22:0, C levels of the selected LacCer after one year of intervention (T1), and D levels of the ratio of these LacCer to Cer22:0 after one year of intervention (T1). Hazard Ratio (95% CI) for quartile 4 (highest concentration of LacCer) compared to quartile 1 (reference) and the p-value are shown. In both cases, Cox regression was performed, adjusting for age, sex, BMI level, intervention group, and recruitment center