Stroke is an important cause of disability and death globally, resulting in more than 6 million deaths per year27,28. TIA condition offers a great opportunity to avoid the development of cerebrovascular disease3,5 as up to one out of five ischemic strokes are heralded by transient ischemic symptoms29. TIA provides a chance to early establish the best management to prevent a devasting ischemic stroke. In our cohort of consecutive TIA patients we observed that in association to previous described predictors of stroke recurrence like motor weakness30,31, LAA4,5,7,15,32 and DWI abnormality4,6,8. Interestingly, lipidomics identified metabolites related to bionerergetics pathways and oxidative stress response differentially expressed in SR group.
The use of blood biomarkers that support in stroke diagnosis and early identification of subjects with high-risk of recurrence is currently a challenge 33 due to the interference of a broad diversity of factors such as age, concomitant pathologies, type of stroke, severity, and cerebral location, among others. However, their interest or their potential utility in clinical practice is high. As cerebrovascular disease is prevalent worldwide28 and TIA patients are heterogeneous in terms of SR5,30, the use of biomarkers related to SR could help the assessment of the individual risk of SR and management decisions22, especially in places without direct access to vascular imaging.
Recent studies analyzing different biofluids (serum and urine) from a metabolomic approach have demonstrated, comparing stroke patients with healthy controls, the presence of specific metabolic profiles ascribed to changes in fatty acids, amino acids, choline metabolism, phospholipids, sphingolipids, and folate one-carbon cycle34–38. These few works collectively reveal the complexity of analyzing and discern metabolic events associated with stroke and the identification of unambiguous biomarkers. Brain ischemia is originated by the transient or permanent occlusion of blood flow to the brain tissue that, in turn, reduces energy supply, alters membrane ionic balance, depolarizes neuronal membrane, increases intracellular Ca2 + concentrations and activates calcium-dependent proteases which, ultimately, leads to the neuronal death38,39. Additional cell damaging mechanisms include alterations of the blood brain barrier and subsequent increase in cerebral oxidative damage and neuroinflammatory response40, as well as metabolic alterations affecting lipid metabolism. Effectively, hypoxic stress (and other cerebral pathological states) induces a cerebral increased content of lipid droplets (LDs), the lipid storage organelles composed of a core of TG and sterol esters surrounded by a phospholipid monolayer and different associated proteins41, predominantly in glial cells and in lesser degree in neurons42. This accumulation of LDs is suggested as a support for energy supply, as well as a neuroprotective mechanism against the stress-induced lipotoxicity42. Remarkably, diverse studies using animal models of ischemia-reperfusion demonstrated that the limited regenerative ability of the injured brain is associated with the formation of inhibitory lipids in the damaged region43. In any case, the result is a severe brain damage and long-term functional impairments whose recovery is challenging and very often inadequate39.
In this context, our aim was to perform a plasma lipidomic analysis among consecutive TIA patients to find new biomarkers of early SR, and further to find predictors of SR. We obtained the plasma lipidomic profiling of TIA patients by non-targeted lipidomics and discovered that, although the global difference is minor, the SR patients presented specific lipidomic changes compared to TIA non-SR patients. The lipidomic profile of patients with SR consisted of a very restrictive set of lipids made up of 5 TG, 1 DG, and 1 plasmalogen. The observed changes in these lipid classes require special attention because the metabolic pathways and cell mechanisms behind them can be crucial in the physiopathology of SR.
There are two functional categories associated with the different lipid classes identified: bioenergetics, and antioxidant protection. Thus, TG are bioenergetic compounds that compose the lipid droplets, and they are also present in neural cells39. DG are components of cell membranes and lipid mediators, but also precursors for biosynthesis of TG39. Finally, plasmalogens are structural components of cell membranes44 and phospholipid monolayer of LDs43, and they also have antioxidant properties45 that help to maintain lipid layer integrity.
Our results indicate a significantly low abundance of these particular lipid species in SR patients compared to non-SR subjects. The observed low abundance of particular DG and TG lipid species in plasma from SR patients points to a low accumulation/formation of cerebral LDs indicating a patient-specific response to stress conditions and suggesting a defective ischemia-associated stress response of SR patients. In another hand, the detected differential plasmalogen also requires a special attention. In human brain, phosphatidylethanolamines (PE) are quantitatively the major phospholipid46,47 and the predominant form is the alkenyl-PE (plasmalogens). Interestingly, their total fatty acid composition indicates a large content of polyunsaturates, and the positional distribution of fatty acids shows that position-1 is occupied primarily by the palmitic acid (16:0), stearic acid (18:0), and oleic acid (18:1) groups both in the white and gray matter, whereas position-2 consists of the highly polyunsaturates and these are more numerous in the gray matter than in the white matter. The lipid species detected in low abundance in plasma from SR patients is PE(P-18:0/18:2) which is not a usual plasmalogen species, maybe because is pending of further remodeling, or because is a lipid species more typically constitutive of the phosphoslipid monolayer present in LDs41. Assuming this fact, the low presence of this plasmalogen reinforces the idea of the defective accumulation of LDs composition in SR patients, as well as the deficient response of these patients based on their antioxidant properties. Importantly, these findings are in line with previous observations in animal models of ischemia-reperfusion43,48 and in ischemic stroke patients38, suggesting that this lipid set express a condition of impaired stress in SR patients compared to TIA non-SR patients.
Although we think that our results could be generalizability as we identified previous described predictors of SR a bigger cohort could increase the statistical power of the lipidomic analysis. In addition, we admit that lipidomic analysis could be influenced by many uncontrolled conditions. Therefore, our results should be confirmed in other independent cohorts.
In conclusion, the lipidomic profiles of TIA subjects with non-SR and SR were different, with minor but significant changes. The observed changes in lipid patterns, especially PE(P-18:0/18:2), suggest pathophysiological mechanisms associated with the different formation of LDs that is translated to plasma level as consequence of a more intensive or high-risk ischemic condition related to early SR. The determination of these differential metabolites which are related to bionerergetics pathways and oxidative stress could improve the assessment of individual risk of SR and management decisions. In addition, our findings encourage the investigation of new potential pharmacological interventions.