Search

Type 2 diabetes is a complex, multifactorial disease with varying presentation and underlying pathophysiology. Recent studies using data-driven cluster analysis have led to a stratification of type 2 diabetes into novel subgroups based on six clinical measurements. Whether these subgroups truly correspond to the underlying phenotypic differences is nevertheless unclear. Here, we apply an unsupervised, data-driven clustering method (Similarity Network Fusion) to characterize type 2 diabetes in two independent cohorts involving 1,134 subjects in total based on integrated plasma lipidomics and peptidomics data without pre-selection. Logistic regression was then used to explore clustering based on ≥ 180 circulating lipids and 1,195 protein biomarkers, alongside clinical signatures. Two subgroups were identified, one of which associated with elevated C-peptide levels, diabetic complications and more severe insulin resistance compared to the other. GWAS analysis against 403 type 2 diabetes risk variants revealed associations of several SNPs with clusters and altered molecular profiles. We thus demonstrate that heterogeneity in type 2 diabetes can be captured by circulating omics alone using an unsupervised bottom-up approach. Such multiomics signatures could reflect pathological mechanisms underlying type 2 diabetes and thus may help inform on precision medicine approaches to disease management.

We have deployed a multi-omics approach in large cohorts of patients with existing type 2 diabetes to identify biomarkers for disease progression across three molecular classes, metabolites, lipids and proteins. A Cox regression analysis for association with time to insulin requirement in 2,973 patients in the DCS, ANDIS and GoDARTS cohorts identified homocitrulline, isoleucine and 2-aminoadipic acid, as well as the bile acids glycocholic and taurocholic acids, as predictive of more rapid deterioration. Increased levels of eight triacylglycerol species, and lowered levels of the sphingomyelin SM 42:2;2 were also predictive of disease progression. Of ∼1,300 proteins examined in two cohorts, levels of GDF-15/MIC1, IL-18RA, CRELD1, NogoR, FAS, and ENPP7 were associated with faster progression, whilst SMAC/DIABLO, COTL1, SPOCK1 and HEMK2 predicted lower progression rates. Strikingly, identified proteins and lipids were also associated with diabetes incidence and prevalence in external replication cohorts. Implicating roles in disease compensation, NogoR/RTN4R improved glucose tolerance in high fat-fed mice and tended to improved insulin signalling in liver cells whilst IL-18R antagonised inflammatory IL-18 signalling towards nuclear factor kappa-B in vitro. Conversely, high NogoR levels led to islet cell apoptosis. This comprehensive, multi-disciplinary approach thus identifies novel biomarkers with potential prognostic utility, provides evidence for new disease mechanisms, and identifies potential therapeutic avenues to slow diabetes progression.

Type 2 diabetes is a multifactorial disease with multiple underlying aetiologies. To address this heterogeneity, investigators of a previous study clustered people with diabetes according to five diabetes subtypes. The aim of the current study is to investigate the etiology of these clusters by comparing their molecular signatures. In three independent cohorts, in total 15,940 individuals were clustered based on five clinical characteristics. In a subset, genetic (N = 12,828), metabolomic (N = 2,945), lipidomic (N = 2,593), and proteomic (N = 1,170) data were obtained in plasma. For each data type, each cluster was compared with the other four clusters as the reference. The insulin-resistant cluster showed the most distinct molecular signature, with higher branched-chain amino acid, diacylglycerol, and triacylglycerol levels and aberrant protein levels in plasma were enriched for proteins in the intracellular PI3K/Akt pathway. The obese cluster showed higher levels of cytokines. The mild diabetes cluster with high HDL showed the most beneficial molecular profile with effects opposite of those seen in the insulin-resistant cluster. This study shows that clustering people with type 2 diabetes can identify underlying molecular mechanisms related to pancreatic islets, liver, and adipose tissue metabolism. This provides novel biological insights into the diverse aetiological processes that would not be evident when type 2 diabetes is viewed as a homogeneous disease.

Aims/hypothesisFive clusters based on clinical characteristics have been suggested as diabetes subtypes: one autoimmune and four subtypes of type 2 diabetes (T2D). In the current study we replicate and cross-validate these T2D clusters in three large cohorts using readily measured variables in the clinic.MethodsIn this cross-sectional study, 15,940 individuals were clustered based on age, BMI, HbA1c, random or fasting C-peptide and HDL in three independent cohorts. Clusters were cross-validated against the original clusters based on HOMA measures. In addition, between cohorts, clusters were cross-validated by re-assigning people based on each cohort’s cluster centres.ResultsFive distinct T2D clusters were identified and mapped back to the original four ANDIS clusters. Using C-peptide and HDL instead of HOMA-B and HOMA-S three of the clusters mapped with high sensitivity (80.6 – 90.7%) to the previously identified Severe Insulin Deficient (SIDD), Severe insulin resistant (SIRD) and Obese (MOD) clusters. The previously described ANDIS MARD cluster could be mapped to the two milder groups in our study – one characterised by high HDL, and the other having not any extreme characteristic (MDH cluster). When these two milder groups were combined they mapped well to the previously labelled MARD cluster (sensitivity 79.4%). In the cross-validation between cohorts, particularly the SIDD and MDH cluster cross-validated well with sensitivities ranging from 73.3% to 97.1%. SIRD and MD showed a lower sensitivity ranging from 36.1% to 92.3% where individuals shifted from SIRD to MD and vice versa.Conclusions/interpretationClusters based on C-peptide instead of HOMA measures result in clusters that resemble those based on HOMA measures, especially for SIDD, SIRD and MOD. By adding HDL, the MARD cluster based upon HOMA measures resulted in the current clustering in two clusters with one cluster having high HDL levels. Cross-validation between cohorts showed generally a good resemblance between cohorts. Together, our results show that the clustering based on clinical variables readily measured in the clinic (age, HbA1c, HDL, BMI and C-peptide) results in informative clusters that are representative of the original ANDIS clusters and stable across cohorts.

SummarydsSwissKnife is an R package that enables several powerful analyses to be performed on federated datasets. The package works alongside DataSHIELD and extends its functionality. We have developed and implemented dsSwissKnife in a large IMI project on type 2 diabetes, RHAPSODY, where data from 10 observational cohorts have been harmonised and federated in CDISC SDTM format and made available for biomarker discovery.Availability and implementationdsSwissKnife is freely available online at https://github.com/sib-swiss/dsSwissKnife. The package is distributed under the GNU General Public License version 3.Contactmark.ibberson@sib.swiss