According to the working hypothesis (see above), the study was designed for detecting metabolic panels which are characteristic for the PDR. The study was focused on the metabolites’ qualification (present versus absent) in stratified patients, in order to create the tool-prototype for the “diagnostic tree” (yes/no).
Here, we report preliminary data on a metabolic signature specific for the DR versus controls and for the PDR with comorbidities. Taken into analysis were only data with at least 3 values absent/present regarding the whole data set.
Metabolic panels specific for the DR
Following metabolites were absent in healthy controls but detected in DR patients:
Metabolite cluster Acylcarnitines: Butenylcarnitine, glutaconylcarnitine, pimelylcarnitine, and octadecadienoylcarnitine
Metabolite cluster Amino acid related: 5-Aminovaleric acid
Metabolite cluster Amino acid: Betaine
Metabolite cluster Bile acids: Glycolithocholic acid sulfate
Metabolite cluster Ceramides: Ceramide (d18:1/18:0(OH))
Metabolite cluster Lysophosphatidyl-cholines: Lysophosphatidyl-choline a C24:0, C26:0, C26:1, and C28:0
Metabolite cluster Nucleobases and related: Xanthine
Metabolite cluster Phosphatidyl-cholines: Phosphatidyl-choline aa C24:0, C26:0, C32:3, C34:3, C34:4, C42:1, C42:2, C42:5, Phosphatidyl-choline ae C30:0, C38:0, C40:4, C40:6, and C44:5
Metabolite cluster Triglycerides: Triacylglyceride (16:1_36:1)
Following metabolites were detected in healthy controls but were absent in all DR patients:
Metabolite cluster Acylcarnitines: Decadienoyl-carnitine
Metabolite cluster Bile acids: Glycoursodeoxy-cholic acid, Taurochenodeoxy-cholic acid, and Taurodeoxycholic acid
Metabolite cluster Lysophosphatidyl-cholines: Lysophosphatidyl-choline a C14:0, and C20:3
Metabolite cluster Triglycerides: Triacylglyceride (18:1_36:1), Triacylglyceride (18:2_36:0), Triacylglyceride (18:2_38:6), Triacylglyceride (20:5_36:2), and Triacylglyceride (22:5_34:2)
Unique metabolic signature of proliferative retinopathy
Preliminary data of a unique metabolic signature for DR patients with first hints for a specific profile for PDR patients is summarised in Table 1.
Table 1
Proposed metabolic signature of PDR in the tear fluid of DR patients: comparative analysis of individual metabolites in DR patients stratified by non-PDR versus PDR and DM-comorbidities (with versus without); quantitative characterisation was perform using the detection limit (DL) of corresponding metabolites.
Metabolite | DL [µM] | Healthy individuals: metabolite detected (no/yes against DL) | DM patients with DR: metabolite detected (no/yes against DL) | DM patients with PDR without comorbidities: metabolite detected (no/yes against DL) | DM patients with PDR and comorbidities: metabolite detected (no/yes against DL) |
Metabolite cluster: Acylcarnitines |
Dodecanoyl-carnitine | 0.018 | yes (> DL) | yes (> DL) | no | no |
Tetradecanoyl-carnitine | 0.017 | yes (> DL) | yes (> DL) | no | no |
Tetradecenoyl-carnitine | 0.008 | yes (> DL) | yes (> DL) | no | no |
Metabolite cluster: Amino acids related |
Asymmetric dimethylarginine | 0.080 | yes (> DL) | yes (> DL) | no | no |
Methionine-Sulfoxide | 0.10 | yes (> DL) | yes (> DL) | no | no |
Metabolite cluster: Cholesterol esters |
Cholesteryl ester 15:1 | 0.15 | yes (> DL) | yes (> DL) | no | no |
Metabolite cluster: Fatty acids |
Eicosadienoic acid | 0.80 | yes (> DL) | yes (> DL) | no | no |
Metabolite cluster: Triglycerides |
Triacylglyceride (14:0_34:0) | 0.079 | yes (> DL) | yes (> DL) | no | no |
Triacylglyceride (16:0_32:0) | 0.26 | yes (> DL) | yes (> DL) | no | no |
Accumulated abundant evidence indicates that small non-protein metabolites are stably altered in individuals with T2DM. Further, their profiles might be characteristic for pre-diabetic stages and progression of T2DM and associated pathologies [18]. Consequently, the body fluid (such as blood plasma and tear fluid) metabolomics is considered of potential clinical utility for the patient stratification, disease prediction, targeted prevention and personalisation of treatment in primary, secondary and tertiary DM care.
Herewith we correlate our findings in the tear fluid with an accumulated knowledge about metabolite clusters, yet qualified and quantified generally in blood.
Metabolite cluster – Acylcarnitines
Carnitine metabolite patterns are known as being altered in blood of DR patients. Moreover, these metabolic profiles demonstrate qualitative and quantitative specificity for T2DM patients with DR, versus without retinopathy as well as patients with PDR versus non-proliferative DR [20]. The carnitine shuttle is an essential attribute of the mitochondrial metabolism. Contextually altered carnitines profile is an independent indicator of mitochondrial stress and compromised mitochondrial health. Carnitines profile measurements are instrumental for identifying the mitochondrial dysfunction associated pathologies such as diabetes, CVD and cancers, amongst others [20]. Due to disease-specific carnitine metabolite set-ups, elevated plasma levels of some acylcarnitine metabolites were associated with increased cardio-vascular risks in T2DM patients.
In our analytical sets, stage/disease specific acylcarnitine metabolites were distinguishable in the tear fluid of the DR against healthy controls as well as in the PDR associated with DM complications.
Metabolite clusters: Amino acids and amino acids related
Plasma and serum amino acid profiles considered reliable biomarker panels for population screening and predictive diagnostics of DM and related complications, since glucose and amino acid metabolisms are closely connected [21]. Meta-analytical studies associate an increased T2DM risks with upregulated blood plasma and serum concentrations of branched-chain, aromatic, alanine, glutamate, lysine, and methionine which were co-detected with enhanced concentrations of carbohydrates, energy-related metabolites as well as some acylcarnitines, triacylglycerols, (lyso)phosphatidylethanolamines and ceramides [22]. These are promising patterns for the T2DM prediction. Further, innovative concepts of a dietary amino acids (AAs) composition have been proposed to improve the overall quality of dietary AAs compositions in targeted T2DM prevention [23].
Our analysis let assume 5-Aminovaleric acid and betaine distinguishable specifically in the tear fluid of the DR patients.
Metabolite cluster: Lysophosphatidyl-cholines
Lysophosphatidylcholines (LPC) are potent inflammatory lipids: significantly increased LPC blood contents correlate well with neurodegeneration and DR progression from the pre-proliferative stage into severe disease [24]. The underlying PDR pathomechanisms involve LPC into blood-retinal barrier damage by promoting oxidative stress injury via TLR4/NF-κB signalling. Further, LPC was proposed to promote pro-angiogenic reactions via VEGFR2 impacting retinal vascular endothelium and inducing vasopermeability [25]. Inhibition of LPC might prevent DM-mediated blood-retinal barrier dysfunction and uncontrolled angiogenesis-related diseases [26].
In our analytical sets, lysophosphatidylcholine profiles detected in the tear fluid were specific for the DR and differ significantly between healthy controls.
Metabolite cluster: Nucleobases and related
In our analytical sets, Xanthine has been found specifically in DR patients while being absent (below the detection limit) in healthy controls. Xanthine is a well-known intermediate product of the purine nucleotide catabolism with urate as end product and thus found in multiple mammalian body fluids [27]. Though, it can bind in vitro to all four nucleotide bases, naturally, it does not incorporate to DNA or RNA due to its instability amongst other reasons [27]. However, oxidative deamination of DNA can lead to Deoxyxanthosine from Deoxyguanosine produced by reactive nitrogen species as a consequence of oxidative and nitrosative stress [27]. Nguyen et al. reported 40-fold increased levels of Xanthine residues in DNA of human lymphoblastoid TK6 cells after exposure to NO [28]. There is evidence that insulin resistance in T2DM might induce the synthesis and deposition of advanced glycation end products, reactive oxygen species, and reactive nitrogen species, leading to cellular stress, protein and mitochondrial dysfunction, and metabolic syndrome, inflammation, and apoptosis [29].
Metabolite clusters: Fatty acids and triglycerides
Lipids and lipoproteins significantly contribute to the pathogenesis of DR. Intervention studies with omega-3 fatty acids demonstrated protective effects against the development and progression of DR. However, the role and impact of individual lipid classes remain underexplored [30]. To this end, association between the hypertriglyceridemia in blood of DM patients and the lower-extremity amputation risk have been demonstrated [31], which is a complication down-stream towards the DR in the overall cascade of DM-complications [17]. Further, blood triglycerides have been demonstrated as a more reliable pre/diabetes biomarker and predictor compared to the total cholesterol, LDL, HDL, LDL/HDL ratio, cholesterol/HDL ratio, and LDL/HDL ratio [32].
In our analytical sets, triglycerides profiles detected in the tear fluid were evidently specific for the PDR and differ significantly between healthy controls and DR patients; further, the absence of eicosadienoic acid was specific for the PDR.
Metabolite cluster: Bile acids
Bile acids play an important role in regulating glucose, energy and lipid associated metabolic pathways. Contextually, metabolic diseases such as obesity, dyslipidemia and T2DM have been associated with dysregulated bile acid homeostasis [33]. Moreover, bile acids are involved in a control over inflammation regulation via signalling pathway based on the nuclear farnesoid X receptor and Takeda G protein-coupled receptor 5. Bile acid signalling and the gut microbiome have been proposed as the targets for treating diabetes. To this end, bile acids have been proposed as a protective factor against DR with great therapeutic potential in the retinal disorder management [34, 35].
In our analytical set, specifically the glyolithocholic acid sulfate belonging to the metabolite cluster of bile acids was detected in DR patients but absent in healthy controls, while glycoursodeoxy-cholic acid, taurochenodeoxy-cholic acid and taurodeoxycholic acid were determined in healthy controls but absent in DR patients.
Metabolite cluster: Ceramides
Via their lipotoxicity, ceramides contribute to the diabetogenic effects implicated in inflammation, β-cell apoptosis and insulin resistance, and are involved in the initiation and progression of microvascular and macrovascular complications of DM relevant for cardiovascular disease, blindness, nephropathy, peripheral neuropathy, and lower-limb amputation [36–38].
In cultured cells and isolated tissues, ceramides promote mitochondrial dysfunction, disrupted vasodilatation and apoptosis [39]. Accumulated research data indicate that the targeted reduction of the ceramide biosynthesis and their decreased circulating levels is beneficial for preventing diabetes and DM-related complications [37]. To this end, serum ceramides are considered reliable biomarkers of adverse cardiovascular disease outcomes. Pre-clinical studies demonstrated protective affects against DM and CVD in mice and rats treated with ceramides inhibitors.
In our analytical sets, ceramide fractions were evidently specific for the DR patients.
Metabolite cluster: Phosphatidyl-cholines
Phosophatidyl-cholines are essential in controlling levels of circulating lipoproteins like very low-density (VLDLs) and high density (HDL) lipoproteins [39]. They are phospholipid components of the class of plasma lipoproteins and needed for lipoprotein assembly and secretion. Multiple studies demonstrated a positive association of increased levels of phosphatidyl-cholines with obesity, metabolic syndrome, insulin resistance, and gestational diabetes [39, 40].
In the eye, choline and phosphatidyl-choline are factors included in tears and the meibum. In the retina, a choline deficiency is associated with a distorted differentiation of retinal neuronal cells and cytoarchitectural defects; moreover, increased phosphatidyl-choline levels are linked to lowered risks for diabetes and cardiovascular diseases [41].
However, our data set demonstrated the presence of several phosphatidyl-choline forms specifically for DR patients, while being absent in healthy controls.
Metabolite cluster: Cholesterol esters
Ample data is available about cholesterol and its implication in cardiovascular disease, and cholesterol ester-enriched foam cells are a main symptom of atherosclerotic plaques [42]. It is also well established that T1DM and T2DM patients have an increased probability of developing cardiovascular disease as a severe diabetes complication caused by augmented cholesterol levels [43]. However, even after lowering cholesterol levels, a residual risk remains, which has been associated to remnant lipoprotein particles (RLP) [43]. Chait et al. defined RLPs as postlipolytic partially triglyceride depleted particles, and the longer RLPs remain in circulation, the greater their enrichment with cholesteryl esters [43]. Cholesterol esters are generated from cholesterol and crucial factors of plasma lipoproteins and cell membranes involved in multiple metabolic processes in the body [44]. In the retina, mainly nonesterified cholesterol is found in a well-controlled homeostasis and reduced cholesterol levels lead to retinal degeneration [44]. Its role in DR is still not fully understood but multiple studies support a positive correlation between dysregulated lipid metabolism and DR [44].
The presented MS analysis demonstrated an absence of cholesterol esters the PDR (with and without comorbidities), while being present in DR patients and healthy controls.