Demographic and clinical characteristics of the study participants:
Table (2) provides demographic and clinical characteristics of the study participants. Of the 75 stroke patients; 29 were diabetic, 44 were hypertensive, 26 were smokers and 14 suffered from ischemic heart disease. All stroke groups exhibited significantly higher serum levels of cholesterol, LDL, VLDL and risk ratio, without any significant differences in triglycerides, HDL-cholesterol and prothrombin time compared to healthy controls.
Serum expression levels of mRNA HIF1-α and LncRNAs in healthy controls and stroke groups:
Compared to healthy controls, the expression levels of mRNA HIF1-α and lncRNA HIF1A-AS2 were markedly upregulated, while those of lncRNA LINK-A were substantially downregulated in all stroke groups at p =0.000. Notably, mRNA HIF1-α and lncRNA HIF1A-AS2 reached their highest level in hemorrhagic group, whereas thrombotic group displayed the least increase as compared to other stroke groups. On the other hand, lncRNA LINK-A was significantly low in hemorrhagic group compared to thrombotic and embolic groups, without any significance difference in its expression level between both ischemic groups (Fig. 1).
Serum levels of MDA and TAC in healthy controls and stroke groups:
In all stroke patients, the serum levels of oxidative stress biomarkers were significantly high in comparison with healthy controls. MDA levels were markedly high in hemorrhagic group compared with thrombotic and embolic groups, without any significant difference between ischemic groups. On the other side, increased serum levels of TAC did not show significant difference among the three studied groups of stroke patients (Fig. 2).
Serum levels of VEGF, ANG1 and BDNF in healthy controls and stroke groups:
In all stroke groups, serum levels of VEGF and ANG-1 were significantly high, whereas BDNF levels were significantly low as compared to healthy controls (Fig. 2). Obviously, serum VEGF was markedly increased in hemorrhagic stroke compared to thrombotic and embolic groups, without any significant difference between the last two groups. Similarly, the hemorrhagic group showed the highest ANG-1 serum levels while the embolic group showed the lowest level. However, there was no significant difference in BDNF serum level between stroke groups.
Serum protein expression levels of PI3K, P-Akt, VEGFR2 and TIE2 in healthy controls and stroke groups:
Results in figure (3) illustrated that the serum protein levels of PI3K, P-Akt, VEGFR2 and TIE2 were significantly higher in all groups of stroke patients versus healthy controls, without any significant difference between the stroke groups versus each other.
Correlation analysis of different serum molecular markers and clinical data in all stroke groups:
To evaluate the usefulness of circulating mRNA HIF1-α and lncRNAs; HIF1A-AS2 and LINK-A as stroke biomarkers, we tested whether their levels were associated with stroke risk factors and biochemical markers.
Pearson’s correlation analysis revealed significant positive correlation of HIF1-α and HIF1A-AS2 with each other and with diabetes mellitus, hypertension, TAGs, serum-cholesterol, LDL, VLDL, risk ratio, prothrombin time, MDA, TAC, VEGF, ANG1, and significant negative correlation with lncRNA LINK-A and BDNF. On the other side lncRNA LINK-A was correlated in opposite manner (Table 3).
ROC curve and positivity rate:
ROC curve analysis was applied to test the possible use of mRNA HIF1-α, lncRNAs; HIF1A-AS2 and LINK-A in predicting hemorrhagic stroke patients from ischemic ones as presented in figure (5) and tables (4,5). All biomarkers were efficient in predicting hemorrhagic from ischemic patients. For HIF1-α, the optimal cut-off value was 7.17-fold change giving 80% sensitivity and 84% specificity and area under the curve (AUC) equivalent to 0.83, whereas for HIF1A-AS2 and LINK-A the optimal cut-off values were 7.83- and 0.28-fold change giving 80% and 92% sensitivity and 82% and 94% specificity, with AUC equivalent to 0.867 and 0.914, respectively. Regarding the positivity rates for HIF1-α, lncRNA HIF1A-AS2 and lncRNA LINK-A, they were 80, 80 and 92 %, respectively in the serum of the hemorrhagic stroke patients compared to 16, 18 and 6 %, respectively in ischemic stroke patients (Fig. 4).
Univariate logistic regression analysis for predicting hemorrhagic stroke:
To predict the risk of hemorrhagic stroke, univariate logistic regression analysis was conducted. Data indicated that HIF1-α, HIF1A-AS2, LINK-A, VEGF and ANG1 were significant predictors for hemorrhagic stroke with odds ratios; 1.984, 2.201, 2.537, 1.016 and 1.243, respectively. Remarkably, LINK-A showed the highest odds ratio (2.537) (Table 5).