Apolipoprotein E (ApoE, encoded by APOE) is a polymorphic glycoprotein serving the function of cholesterol transport. APOE consists of three common alleles, ε2, ε3, and ε4, and there are six genotypes (ε2/2, ε2/3, ε2/4, ε3/3, ε3/4, and ε4/4). APOE may act as a strong genetic cause of ICH, with ε2 and ε4 showing positive, and ε3 showing negative, association with ICH (Talha et al., 2020). Marini et al. (2019) reported that APOE ε2 and ε4 alleles appear to affect LICH risk, and Biffi et al. (2010) confirmed the findings and found a novel association between APOE ε4 and DICH. Moreover, O’Donnell et al. (2000) found that patients with alleles ε2 and ε4 had a 28% 2-year recurrence rate of ICH, as compared with only 10% in carriers of ε3/ε3 genotype. As for racial differences, genetic studies for ICH occurrence have produced conflicting results. Zhang et al. (2014) reported a significantly higher frequency of APOE ε4 allele presented in Asian, European and American patients compared with ICH-free individuals. However, Nie et al. (2019) revealed that both APOE ε2 and ε4 alleles acted as risk factors for ICH only in European and American populations, and neither of them had an effect in Asian populations. The pathophysiology underlying the association between APOE polymorphisms and ICH is generally accepted to be mediated by CAA. It is supposed that APOE ε4 enhances vascular amyloid deposition, while APOE ε2 promotes severe CAA responsible for vessel rupture. Moreover, the APOE ε4 allele may increase the risk of developing hypertension.

Heritability estimates have identified that APOE genetic variations play substantial roles not only in occurrence of ICH, but also in hematoma volume and ICH outcome (Devan et al., 2013). Brouwers et al. (2012a,b) demonstrated that APOE ε2 was associated with hematoma expansion and CTA spot sign in patients with LICH. Biffi et al. (2011) found that APOEε2 increased the risk of poor functional outcome (mRS > 2) and mortality at 3 months in LICH. Moreover, functional dependency (Math et al., 2019) as well as poor survival (McCarron et al., 2003) following ICH was observed in patients who possessed the APOE ε4. APOE ε2 seems to promote the severity and rupture of diseased blood vessels, leading to a larger range of bleeding. APOE ε4 appears to amplify neuroinflammatory responses in the setting of ICH, causing increased cerebral edema.

Glutathione peroxidase 1 (GPX1, encoded by GPX1) is an intracellular antioxidant enzyme involved in vascular defense through detoxifying hydrogen. Forgione et al. found that genotypes containing the T allele of C593T polymorphism of GPX1 doubled the risk of LICH compared with CC genotype (Pera et al., 2008). The mechanism is possibly related to amyloid-associated pathology. Deficiency of GPX1 leads to endothelial dysfunction, which could increase susceptibility to oxidative stress-related injury in amyloid-induced vascular damage. Besides, a lack of enzyme GPX1 increases susceptibility to amyloid toxicity in cultured cortical neurons.

Complement C3b/C4b receptor 1 (CR1, encoded by CR1), a member of complement reactivation family, mediates immune responses through binding with the complement component (3b/4b). A genetics of Cerebral Hemorrhage on Anticoagulation (GOCHA) study identified that the CR1 rs6656401 polymorphism influenced occurrence as well as recurrence of CAA-ICH (Biffi et al., 2012). The CR1 genetic variability leading to higher risk of LICH is thought to be involved in the clearance of amyloid plaques. Rogers et al. (2006) demonstrated that the CR1 protein was bound to Aβ42 peptide at its C3b ligation adhesion site, while Crehan et al. (2013) found the ability of microglia to phagocytose Aβ was impeded through blocking CR1, resulting in the clearance of Aβ.

Integrin subunit alpha V (ITGAV, encoded by ITGAV) is a glycoprotein belonging to a family of type I transmembrane glycoprotein receptors and mediates a number of cellular functions binding its extracellular ligand to transmit signal. Dardiotis et al. (2017) found that the rs7565633 polymorphism of ITGAV was independently associated with LICH risk. One consideration is that ITGAV may affect the integrity of the BBB, as it is implicated in the steadiness of endothelial cell junctions and the tight adhesion of endothelium to the astrocytic perivascular endfeet. Moreover, ITGAV can mediate Aβ-induced toxicity to neurons.

Protein kinase C eta (PKC eta, encoded by PRKCH) is a serine-threonine kinase essential for cellular signaling transduction and functional regulation of cells, including proliferation, differentiation, and apoptosis. The 1425G/A polymorphism in PRKCH enhances the kinase activity. Chen et al. (2012) demonstrated a borderline association between of 1425A and LICH. PKC is involved in the development of atherosclerosis in humans, and the pathophysiology has been proposed to be mediated by inflammation and immunity.

A GWAS by Woo et al. (2014) found that several variants on chr12q21.1, an intergenic region near the Thyrotropin-releasing hormone-degrading ectoenzyme (TRHDE), were associated with LICH in Europeans, with peak association detected at rs11179580. However, replication in Americans failed to corroborate the positive association (Woo et al., 2014). It is unknown whether the effect estimated in the discovery phase varies across races and how genetic variants on chr12q21.1 contribute to susceptibility to ICH (Tables 1, 2 and Figures 1A,B).

Angiotensin I converting enzyme (ACE, encoded by ACE) is a key enzyme of the rennin-angiotensin system, which converts angiotensin I into angiotensin II and inactivates bradykinin. Chen et al. (2008) examined the associations of two polymorphisms (ACE A-240T and ACE I/D) with ICH and found that haplotype ACE T-D polymorphism played minor a role in female DICH risk independent of hypertension, which is consistent with a later study found that ACE DD genotype was independently associated with deep hematoma (Kalita et al., 2011). In addition to the effect on ICH occurrence, the ACE I/D was reported to influence HICH recurrence as well, with 87.1% of patients deep seated (Misra et al., 2012). Furthermore, Qin et al. (2013) revealed that DD genotype had a greater risk in HICH than that in general ICH, and a lower risk in the subgroup with controls including hypertension than that excluding hypertension, indicating that ACE I/D polymorphism and hypertension may have a synergistic effect for ICH. Analyses by race suggested that ACE I/D polymorphism was associated with individual susceptibility to ICH in Asians, but not Europeans (Li et al., 2020). The reigning hypothesis of the mechanism is that ACE may promote hypertension by enhancing serum ACE levels. Converted by ACE, angiotensin II is involved in the hypertensive development processes, including arteriolar proliferation, smooth muscle cell death and collagen deposition. Moreover, angiotensin II can induce endothelial dysfunction through inflammatory mechanisms. Additionally, chronically high levels of plasma ACE have been shown to increase thickness and stiffness of vascular walls.

The chr1q22 locus contains two genes, polyamine-modulated factor 1 (PMF1) gene PMF1 and solute carrier family 25-member 44 (SLC25A44) gene SLC25A44. PMF1 is a nuclear protein required for mitosis and polyamine metabolism. SLC25A44 belongs to the SLC25 family of mitochondrial carrier proteins. Woo et al. (2014) identified several variants related with DICH susceptibility located on chr1q22. For the most robust rs2984613, each additional C allele increased the occurrence rate of DICH by 24% (Woo et al., 2014). The variants within 1q22 may alter expression of PMF1 and disrupt polyamine metabolism, which links with cerebrovascular disease through excitatory neurotransmitter receptor regulation and BBB breakdown.

Collagen type IV alpha 1, 2 chains (COL4A1 and COL4A2, encoded by COL4A1 and COL4A2, respectively), contribute to type IV collagen, a principal component of the vascular basement membrane. Lin et al. (2018) found that the rs544012 AC and rs679505 AA genotypes of COL4A1 were independently associated with the risk of HICH. Meanwhile, they also demonstrated that rs532625 AA of COL4A1 was associated with higher susceptibility to disability at 3 and 6 months and susceptibility to death/disability at 6 months in HICH patients (Rannikmäe et al., 2015). Rannikmäe et al. (2015) identified three intronic single nucleotide polymorphisms (SNPs) in COL4A2, including rs9521732, rs9521733, and rs9515199, contributing to sporadic cases of DICH. Later, in a larger population, nine SNPs in COL4A2 associated with DICH were identified (Rannikmäe et al., 2017). The accumulation of mutant collagen in vascular endothelial cells and pericytes contributes to basement membrane defects, which weaken the blood vessels and affect vascular function. It has been revealed that COL4A1 and COL4A2 missense mutations could change evolutionarily conserved amino acids and attribute to increased intracellular retention of COL4A1 and COL4A2.

Matrix metalloproteinases (MMPs) are a family of zinc/calcium containing endopeptidases and play a crucial role in degrading the extracellular matrix (ECM). Among MMPs, gelatinases, MMP2 (encoded by MMP2) and MMP9 (encoded by MMP9), are uniquely significant in digesting elastin and collagen of the vascular basement membrane. Chen et al. (2015) and Ho et al. (2015) found that the variants in MMP2 (rs2285053) and MMP9 (rs3787268 and rs2250889) promoters could affect DICH risk in view of age and gender, respectively. It is supposed that MMPs contribute to ICH through degrading ECM of vessels. With abnormal presence of MMPs, the deposition of vascular collagen is insufficient, leading to weakness of the vessel walls and the formation of aneurysm. Furthermore, MMPs pathway also plays roles in tissue remodeling and inflammation reactions in ICH.

Tissue inhibitors of metalloproteinases (TIMPs) are a family of glycosylated proteins that negatively regulate the proteolytic activity of MMPs. TIMP1 (encoded by TIMP1) and TIMP2 (encoded by TIMP2) act as the main endogenous inhibitors of MMP9 and MMP2, respectively. Two variants (rs7503607 and rs7503726) of TIMP2 and one variant (rs4898) of TIMP1 have been identified as DICH risk factors among subjects of different age and gender (Chen et al., 2015; Ho et al., 2015). TIMPs have been postulated to affect vascular ECM remodeling and breakdown of the BBB by modulating the balance between matrix synthesis and degradation through MMPs inhibition. Moreover, it has been demonstrated that the expression of TIMPs had effects on the development of hypertension, atherosclerosis and intracranial aneurysms.

Tumor necrosis factor (TNF, encoded by TNF) is a proinflammatory and immunomodulatory cytokine implicated in the inflammatory process. The TNF polymorphisms in its promoter region can lead to an increase of TNF concentration. Chen et al. (2010) reported that the minor alleles −1031C and −308A were positively associated with DICH risk in males, whereas the −863A was inversely associated with DICH risk in females after adjusting for confounding factors such as hypertension, diabetes, etc. TNF can result in the weakening and rupture of the vessel wall through inducing the production of MMPs. Moreover, TNF can reduce the release of tissue plasminogen activator (t-PA) and stimulate the release of plasminogen activator inhibitor type 1 to take the antifibrinolytic effects, which may be involved in early activation of the hemostatic mechanism during a DICH event (Tables 1, 2 and Figures 1A,B).

Methylenetetrahydrofolate reductase (MTHFR, encoded by MTHFR) is a ubiquitous cytosolic enzyme playing a key role in homocysteine metabolism. The C677T variant in MTHFR can alter the amino acid sequence and enzymatic activity of MTHFR. Several large-scale meta-analyses have shown that MTHFR C677T polymorphism was associated with ICH risk (Kang et al., 2013; Zhao and Jiang, 2013; Hu et al., 2016). A prospective study demonstrated that MTHFR TT genotype was a predictor of ICH independent of hypertension (Hultdin et al., 2011). The reduction of enzyme activity caused by MTHFR variant leads to high levels of plasma homocysteine, which may cause endothelial dysfunction responsible for atherosclerosis. Other proposed pathogenetic mechanisms are related to alterations of the normal procoagulant-anticoagulant balance and vascular smooth muscle cell proliferation.

Cholesteryl ester transfer protein (CETP, encoded by CETP) is an important molecule of the lipid metabolism system, mediating exchange of cholesteryl esters and triglycerides between lipoproteins of different density. Anderson et al. (2016) demonstrated an association between twelve CETP variants and ICH risk in nearly 6,000 participants, with the strongest association detected at the rs173539 locus. Several variants in CETP that result in lower plasma CETP levels and activity have been identified as increasing high-density lipoprotein cholesterol (HDL-C) and total cholesterol (TC) concentrations but lowering low-density lipoprotein cholesterol (LDL-C) (Xu et al., 2016). Dysregulation of lipids influences endothelial function and initiates the formation of early atherosclerotic lesions in arterial vasculature, which can lead to decreased vascular compliance and increased vascular pressure.

Rho associated coiled-coil containing protein kinases (ROCKs), referred to as ROCK1 (encoded by ROCK1) and ROCK2 (encoded by ROCK2), belong to the AGC family of serine/threonine kinases. Serving as effectors of the RhoA small GTPase, ROCKs are involved in diverse cellular processes, including cell contraction, migration, proliferation and apoptosis. Yang et al. (2018) found that three polymorphisms (rs288980, rs7237677, and rs978906) in ROCK1 and ROCK2 were significantly associated with ICH independent of hypertension, which was observed having positive relationship with ROCK1 polymorphism when following 4,128 subjects. The underlying mechanism for the association is likely a consequence of cerebrovascular atherosclerosis due to dysfunction of endothelial cell and vascular smooth muscle caused by abnormal activation of RhoA/ROCK pathway.

Fibrinogen alpha chain (FGA, encoded by FGA) is one type of three chains comprising fibrinogen molecules, which is crucial for intravascular thrombus formation and clot viscoelastic properties. Jagiełła et al. (2014) reported that AA genotype of FGA Thr312Ala was correlated with a higher prevalence of ICH. The same FGA polymorphism has been reported to cause amino-acid substitution and influence clot structure and properties by facilitating factor XIII cross-linking of fibrin fibers (Standeven et al., 2003). It is assumed that FGA Thr312Ala variant could increase the risk of ICH involving the coagulation pathway, since the disturbance of blood coagulation contributes to the occurrence of hemorrhage.

Serine/threonine/tyrosine kinase 1 (STYK1, encoded by STYK1) is a novel oncogene with kinase domain and is important for regulating intracellular signaling pathways and mediating diverse cellular and developmental processes. An exome-wide association study (EWAS) revealed that rs138533962 polymorphism of STYK1 had significant associations with ICH occurrence and the prevalence of some intermediate phenotypes, such as hypertension, diabetes mellitus, and hypo-HDL-cholesterolemia (Yamada et al., 2017). Given STYK1’s roles in promoting angiogenesis and changing the vascular morphology during tumor growth, it is deduced that STYK1 may have an effect on the remodeling of blood vessels in the brain (Table 3 and Figures 1A,B).

Secreted by either endothelial cells or platelets, von Willebrand Factor (vWF, encoded by VWF) mediates platelet adhesion and aggregation at the site of vascular injury. Appelboom et al. (2013) found that VWF rs216321 polymorphism was significantly associated with relative hematoma growth and approached a borderline association with absolute hematoma growth. It has been reported that the rs216321 variant could decrease the ability of vWF to activate platelets and enhance the ability of vWF to carry factor VIII (Vaidya et al., 2010). Furthermore, combined exposure to hypertension and low levels of vWF largely increased the risk of ICH compared with their separate effects (Johansson et al., 2004), indicating a possible synergistic interaction between them.

17p12 was identified as a genome-wide significant susceptibility locus for hematoma volume of DICH by Marini et al. (2018). Specifically, it was found that rs11655160 polymorphism within the locus was associated with a 6.6-mL decrease in mean ICH volume. Moreover, this same variant was found to be a predictor of lower admission Glasgow coma scale (aGCS) and poor functional outcome (defined as modified Rankin Scale > 2, mRS > 2) at 3 months after ICH (Marini et al., 2018). 22q13 was identified as a locus for ICH volume in LICH, but failed to be replicated (Marini et al., 2018). The biological mechanism underlying the findings is still unknown.

Haptoglobin (Hp, encoded by HP) is an acute phase plasma glycoprotein that promotes the stabilization and clearance of hemoglobin (Hb). In humans, HP has two major HP alleles, HP1 and HP2, represented as three genotypes: HP1–1, HP2–1, and HP2–2. Murthy et al. (2015) showed that compared with carriers of HP 1-1 phenotype, patients with HP 2-1/2-2 had a higher occurrence rate of poor functional outcome (mRS > 2) after ICH. Released after ICH, Hb causes the damage of vascular endothelium mediated by free radicals, and leads to disruption of the BBB, resulting in cerebral edema and death of brain parenchymal cells. Hp accelerates the rapid clearance of Hb via CD163 scavenger receptors forming Hp–Hb–CD163 complex on macrophages, exerting its antioxidant activities.

Interleukin 6 cytokine family signal transducer (IL6ST, encoded by IL6ST) performs signal transduction in the processes of immune response, inflammation, the acute phase response and hemopoiesis. Carriers with the minor allele (G) of IL6ST rs10940495 were associated with an 84% decrease of having a poor outcome (mRS > 2) at 6 months compared with major allele homozygotes (AA) carriers (El Husseini et al., 2018). Neuroinflammation is thought to cause the secondary brain injury and influence outcomes following ICH, and the formation and homodimerization of the complex of IL-6, IL-6R, and IL6ST can trigger the two main signaling pathways involved in the process of neuroinflammation. Besides, Soluble IL6ST has been found to influence blood pressure in patients with stoke.

Complement factor H (CFH, encoded by CFH) is a primary regulatory component of the alternative pathway, and is crucial for restricting complement action to activating surfaces and regulating complement activation. Appelboom et al. showed that CFH rs1061170 polymorphism was independently predictive of mortality at discharge and 6-months, as well as survival duration after ICH onset (Appelboom et al., 2011). In the setting of ICH, thrombin deposits at the hematoma site to activate complement, which then mediates inflammatory reactions such as recruitment of inflammatory cells, release of cytokines and lysis of erythrocytes, contributing to BBB disruption, cerebral edema development and cerebral injury.