Early Brain Injury After Subarachnoid Hemorrhage: Incidence and Mechanisms

Abstract

Aneurysmal subarachnoid hemorrhage is a devastating condition causing significant morbidity and mortality. While outcomes from subarachnoid hemorrhage have improved in recent years, there continues to be significant interest in identifying therapeutic targets for this disease. In particular, there has been a shift in emphasis towards secondary brain injury that develops in the first 72 hours after subarachnoid hemorrhage. This time period of interest is referred to as the early brain injury period, and comprises processes including microcirculatory dysfunction, blood-brain-barrier breakdown, neuroinflammation, cerebral edema, oxidative cascades, and neuronal death. Advances in our understanding of the mechanisms defining the early brain injury period have been accompanied by improved imaging and non-imaging biomarkers for identifying early brain injury, leading to the recognition of an elevated clinical incidence of early brain injury compared to prior estimates. With the frequency, impact, and mechanisms of early brain injury better defined, there is a need to review the literature in this area to guide preclinical and clinical study.

Introduction

Aneurysmal subarachnoid hemorrhage (SAH) accounts for 5% of strokes annually, but up to 27% of stroke-related years of potential life lost^1-4. In recent decades, survival from SAH has improved due to modern neuroimaging technologies, novel aneurysm treatment modalities, increased access to organized neurocritical care, streamlined workflows allowing more rapid hospitalization after SAH, and widespread use of nimodipine therapy following trials showing reduced neurologic morbidity with nimodipine administration^5-9. In tandem with these improvements, translational research in SAH has continued to grow with the aim of identifying targetable molecular pathways that may reduce the population impact of SAH¹⁰.

Historically, much attention was directed towards preventing large artery vasospasm causing delayed cerebral ischemia (DCI) after SAH^{11, 12}. However, therapies that reduced angiographic vasospasm have not translated to improved patient outcomes¹³. Driven by results from preclinical and clinical studies including the Clazosentan to overcome neurological ischemia and infarction occurring after subarachnoid hemorrhage (CONSCIOUS) randomized clinical trials^{13, 14}, a growing body of evidence implicates processes besides large artery vasospasm as important contributors to the poor patient outcomes associated with SAH¹⁵. In particular, substantial attention has been focused on the pathophysiological processes that occur immediately after SAH^{11, 16}. During this critical early period, many patients develop secondary brain injury in the form of microcirculatory dysfunction, Blood-Brain-Barrier (BBB) breakdown, neuroinflammation, oxidative cascades, cerebral edema, and neuronal cell death¹¹. Secondary brain injury in this acute time is commonly referred to as Early Brain Injury (EBI), with many investigators elucidating its independent contribution to poor patient outcome, its relationship to subsequent development of DCI, and its molecular underpinnings¹⁷

Alongside these advances in understanding the mechanisms of EBI, our ability to measure this form of secondary brain injury has increased dramatically in recent years^{18, 19}. Early techniques to identify EBI relied entirely upon assessment of the presence or absence of global cerebral edema (GCE) on imaging around the time of patient admission and was reported as a dichotomous indicator¹⁸. With this approach for identification of EBI, only 8% of SAH patients were found to have EBI on admission, while another 12% developed EBI soon after admission²⁰. Since these early reports, measurement strategies for identification of EBI have evolved and include more quantitative scales including the semi-quantitative subarachnoid hemorrhage early brain edema score (SEBES)²¹ and the quantitative measurement of selective sulcal volume (SSV)¹⁸. Recently, Yuan et al. described an advancement to the quantitative technique for identifying EBI when they developed an automated approach for measuring SSV¹⁹. With these more quantitative approaches for measuring EBI, identification of this important secondary brain injury process has increased. Specifically, EBI as identified by qualitative GCE assessment was observed in 37% of patients, while semi-quantitative SEBES measurement led to the diagnosis of EBI in 49% of patients and quantitative measurement of SSV led to the diagnosis of EBI in 61% of patients¹⁹. Therefore, modern imaging techniques have revealed that EBI is more prevalent than previously believed.

As the frequency, impact, and underlying mechanisms of EBI have come to the fore of the field of SAH, it is critical to pause and review the literature to guide future preclinical and clinical studies. In the present work, we review state-of-the art strategies to identify EBI in SAH patients, demonstrate its impact on patient outcomes, and review the underlying mechanisms that are linked to EBI pathophysiology. To synthesize the current state of the literature on this crucial topic, the indexing databases PubMed, Scopus, and Embase were searched with the query “(Early Brain Injury) AND (Subarachnoid Hemorrhage)” from January 1, 2000 to December 31, 2021. Biomarkers of EBI are detailed in Supplement 1 and animal models of EBI are detailed in Supplement 2. To synthesize the current understanding of EBI, we first discuss primary disturbances immediately following SAH and subsequently present secondary disturbances that occur downstream of these initial events. We then discuss molecular mechanisms that cause secondary disturbances, placing emphasis on those with greater translational potential.

Characteristics of EBI

Following aneurysm rupture, primary disturbances activate downstream pathways that produce EBI, a form of secondary brain injury that strongly predicts poor long-term outcome after aneurysmal SAH (for detailed discussion, see Supplement 1)²². These primary disturbances include extravasation of hemorrhagic blood, a rapid rise in intracranial pressure (ICP), acute vasospasm, diminished cerebral blood flow (CBF), disrupted cerebral autoregulation, and brain swelling⁵. Mass effect, toxicity of blood products, and an inability for the cerebral vasculature to maintain a consistent CPP activates downstream cascades, causing the poor neurological state patients demonstrate acutely following SAH^{23, 24}. The downstream secondary pathological cascades classically attributed to SAH-induced EBI include microvascular dysfunction, BBB disruption, cerebral edema, neuroinflammation, oxidative cascades, and neuronal death²⁵ (Figure 1.).

Figure 1.

Characteristic secondary disturbances mediated by pathways that occur during the EBI period and lead to loss of neuronal integrity. Major mediators of each process are outlined. Several mediators function in producing the phenotypes of EBI through multiple mechanisms and are listed on separate pathways. NFkB = Nuclear factor kappa B, TLR-4 = Toll-like receptor 4, iNOS = Inducible nitric oxide synthase, NLRP3 = NLR family pyrin domain containing 3, Bcl2/Bax = B-cell lymphoma 2/Bcl-2-associated X protein, PI3k/Akt = Phosphoinositide 3-kinase/Akt, AIM2 = Absent in melanoma 2, eNOS = Endothelial nitric oxide synthase, L-type VDCC = L-type voltage dependent calcium channel, CaSR = Calcium-sensing receptor, H2S = Hydrogen sulfide, AQP4 = Aquaporin-4, HIF-1a = Hypoxia inducible factor 1 subunit alpha, VEGF = Vascular endothelial growth factor, MMP-9 = Matrix-metalloproteinase-9, ZO-1 = Zonula occludens-1, TREM-1 = Triggering receptor expressed on myeloid cells-1, Figure created using Biorender.com.

Primary Disturbances

The development of brain injury begins immediately following SAH with a rapid rise in ICP from hemorrhagic blood extravasation into the subarachnoid spaces, ventricles, and parenchyma²⁶. This rise in ICP, paired with acute vasoconstriction and microthrombosis, disrupts cerebral perfusion pressure (CPP) and CBF, generating global ischemia^{2, 5}. In severe SAH, global ischemia can cause unconsciousness and complete perfusion arrest²². These events also contribute to the “thunderclap headache” classically described by patients experiencing SAH. SAH also causes pulmonary, cardiac, and systemic inflammatory dysfunction, which may exacerbate processes contributing to secondary brain injury in the EBI period²⁷.

Alongside brain ischemia, extravasation of blood and hemoglobin breakdown products into the brain stimulate pathological cascades that later induce secondary brain injury^{2, 28, 29}. Breakdown of heme-containing blood generates toxic substrates, which catalyze reactions that produce reactive oxygen species (ROS) and activate neuroinflammation³⁰. Pathways activated by toxic blood breakdown products also disrupt the microcirculation and BBB, and generate cerebral edema^{11, 30, 31}. Neuronal loss from cell damage caused by these pathologic cascades contributes to the neurologic dysfunction seen in the acute period following SAH^{32, 33}.

Secondary Disturbances

Microcirculatory Dysfunction

Downstream of primary disturbances from SAH, deleterious pathways disrupt cerebral homeostasis. These changes may be first encountered with changes in the microvasculature. While the behavior of large vessels following SAH has been well-studied in DCI, the role of small vessels including pial arterioles in early microcirculatory dysfunction has garnered attention^{34, 35}. Recent data suggests that both endothelial cells and smooth muscle cells are implicated in acute microvasospasm³⁵. While the precise etiology of early microcirculatory dysfunction remains unclear, potential mediators of acute microvasospasm and microthrombosis include nitric oxide (NO) and ionized calcium (Ca²⁺)^{36, 37}. In the state of secondary brain injury following SAH, there is a characteristic decrease in NO and increase in Ca²⁺ concentration^{38, 39}. Therefore, biological systems dictating the behavior of these molecular mediators such as nitric oxide synthase (NOS) and L-type voltage dependent calcium channels (VDCCs) are important contributors to microcirculatory dysfunction, and more generally the overall phenotype of EBI.

Blood-Brain-Barrier Disruption

Several damaging pathways in the post-hemorrhagic brain act on the BBB. In the quiescent state, the BBB maintains homeostatic and autoregulatory functions in the brain, regulating molecular transfer and preventing toxic insults⁴⁰. The BBB also conserves the immune privilege of the central nervous system (CNS), inhibiting immune cell entry in periods of stasis and regulating leukocyte infiltration during inflammation⁴¹. Structurally, the BBB is formed by endothelial cells, which are supported by astrocytes and pericytes that perform maintenance and repair functions^{42, 43}. Between endothelial cells, tight junctions and adherens junctions prevent paracellular molecular transfer. Disruption of the BBB during SAH rapidly compromises its function, causing leukocytic infiltration, influx of toxins, and stimulation of inflammatory and oxidative cascades^{31, 44}. Indeed, complete BBB disruption can be observed as early as 30 minutes after SAH^{42, 44}. Initially, ischemia causes rapid apoptosis of endothelial cells and perivascular astrocytes, causing structural deterioration of the BBB⁴⁵. Blood products further disrupt the BBB by initiating the synthesis of inflammatory cytokines and enzymes that degrade the BBB. Ultimately, these pathways lead to loss of tight junctions, degradation of the extracellular matrix (ECM), and disruption of the basal lamina⁴⁶. Rendered incapable of its baseline neuroprotective functions, the BBB worsens EBI by allowing unregulated cell, toxin, and ion movement into the brain.

Cerebral Edema

Cerebral edema develops rapidly following SAH, and is closely related to BBB disruption⁴⁷. Specifically, vasogenic edema occurs due to endothelial cell dysfunction, tight junction destruction, and basal lamina degradation⁴⁸. Cytotoxic edema, or ionic edema, is the second component of cerebral edema in EBI and occurs in the setting of cellular homeostasis failure⁴⁸. After SAH, cytotoxic edema is generated by osmotic shifts caused by dysfunctional cell membrane pumps downstream of ischemic energy failure¹¹. Early studies of cerebral edema after SAH characterized its development as vasogenic due to apoptosis and death of endothelial cells and perivascular astrocytes¹¹. Data derived from diffusion-weighted MRI of patients with SAH clarified the important role of cytotoxic edema in EBI, revealing that both forms of edema mutually contribute to cerebral edema in EBI⁴⁹. Alleviating cerebral edema is a promising strategy for improving outcomes in SAH, as Cao et al. demonstrated that hydrogen sulfide (H₂S) can alleviate cerebral edema during EBI and improve neurological function in rats⁵⁰.

Neuroinflammation

Destruction of the BBB and development of cerebral edema are accompanied by robust immunologic and oxidative cascades. The immunologic cascades in EBI are primarily mediated by resident microglial cells and infiltrating leukocytes⁵¹. After SAH, leukocytic infiltration from the peripheral blood is one of the earliest events, where infiltrating neutrophils stimulate resident microglia and activate innate responses⁵². Once neuroinflammatory signaling has begun, central chemokine expression increases recruitment of neutrophils, monocytes, and other immune cells from the periphery. Accumulation of neutrophils in the cerebral microvasculature and parenchyma was observed as early as 10 minutes after experimental SAH in a study by Ye et al.⁵³. Activated neutrophils release inflammatory cytokines following their arrival in the brain, further stimulating resident cells⁵⁴. Alongside leukocytic influx, hemorrhagic blood products stimulate toll-like receptors (TLRs), primarily TLR-4, which then initiate de-novo synthesis of inflammatory cytokines and intrinsic CNS inflammation pathways⁵⁵. These pathways also impact the behavior of resident immune cells, inducing microglial activation and polarization. Neuroinflammatory responses after SAH have a detrimental clinical impact, both acutely and in the long-term. Indeed, elevated levels of inflammatory cytokines are associated with poorer clinical outcomes after SAH^{2, 56}.

Oxidative Cascades

Along with vulnerability to inflammation, the brain is highly susceptible to oxidative damage following SAH, in part due to its significant oxygen requirement⁵⁷. The metabolic demand of cerebral function leads to the generation of free radicals as by-products of normal cellular respiration⁵⁸. To mollify the toxic effects of these by-products, the brain utilizes endogenous glutathione-dependent processes to maintain oxidative balance. In the CNS, the function of glutathione relies on the enzymes superoxide dismutase, catalase, and glutathione peroxidase, as well as the cystine/glutamate antiporter system Xc− ^{29, 59}. At rest, these enzymes prevent excess production of free radicals.

Following SAH, several processes cause detrimental oxidative reactions¹¹. Elevated metabolic demands and impaired cellular respiration lead to excess free radical generation, overpowering intrinsic antioxidant systems⁵⁹. Blood products that enter the brain in SAH cause lipid peroxidation, worsening free radical generation²⁹. Oxidative damage can contribute significantly to secondary brain injury. Indeed, the degree of lipid peroxidation after SAH correlates with clinical prognosis⁶⁰.

The toxic effects of blood and hemoglobin breakdown products as oxidative and neurotoxic agents have direct effects on neuronal health, which is demonstrated by studies targeting these products. Specifically, therapeutic us of haptoglobin, a physiologic hemoglobin scavenger, has the ability to mitigate neuronal damage and delayed vasospasm after SAH in a CD 163-mediated system^{61, 62, 63}. A mouse model using exogenous haptoglobin administration conducted by Garland et al. demonstrated attenuation of behavioral deficit, glial cell disruption, and synaptic changes compared to control mice, confirming the effect of blood product toxicity³¹.

Apoptosis and Cell Death

BBB disruption, cerebral edema, neuroinflammation, and oxidative cascades initiate events that result in neuronal death^{33, 64}. Several mechanisms of neuronal death are present in EBI, some well-studied, and others that are only recently described. These mechanisms include apoptosis, pyroptosis, ferroptosis, and necroptosis³³. In addition, autophagy is involved in the downstream outcomes of EBI, where it is both protective and deleterious⁶⁵.

Molecular Mechanisms of EBI

With the broad characteristics of EBI well-defined, recent work has identified pathways contributing to secondary brain injury with therapeutic potential that warrant explicit discussion and may be the subject of future preclinical and clinical studies.

Microcirculatory Dysfunction

Nitric Oxide Pathways

Studies examining delayed neurological deterioration after SAH previously identified NO as a targetable mediator of DCI with clinical potential⁶⁶. The influence of NO on EBI has also been the subject of more contemporary investigations^{13, 14}. In the brain, NO is produced from L-Arginine by NOS, which has 3 forms – endothelial NOS (eNOS), inducible NOS (iNOS), and neuronal NOS (nNOS)⁶⁷. Baseline microvascular stability and tone is maintained by eNOS, which becomes disrupted within 1 hour of experimental SAH⁶⁸. This is followed by a rapid decrease in NO levels³⁸. A potential role of NO in early microcirculatory dysfunction was recently described by Lenz et al., who demonstrated that eNOS −/− mice had less microvessel perfusion and experienced more acute microvasospasms in an experimental SAH model than their wild-type counterparts³⁶. Additional investigations demonstrated that exogenous delivery of NO following SAH attenuates cerebral microvasospasm, improves neurologic outcome, and provides neuroprotection in preclinical models³⁶. Furthermore, a possible mechanistic connection between NO depletion, carbon dioxide sensitivity, and early microvasospasm and microthrombosis was posited by Friedrich et al. in mice⁶⁹. Due to the baseline function of NO in platelet aggregation inhibition, NO depletion after SAH may exacerbate microthrombosis in the EBI period³⁵. Therefore, targeting NO-dependent pathways may have a role in future studies of EBI.

Intracellular Calcium Pathways

Changes in intracellular Ca²⁺ levels occur as early as 15 minutes following SAH, and increased intracellular Ca²⁺ in endothelial and smooth muscle cells has been noted to increase cerebral vasospasm¹⁵. Mechanistically, Ca²⁺ activates myosin light chain, which then induces contraction and spasm of small cerebral vessels⁷⁰. Recent animal studies have implicated the role of L-type VDCC’s in mediating Ca²⁺ influx and enhancement of pressure-dependent vasoconstriction³⁴. This abnormal arteriolar constriction contributes to neurological deficits in the EBI period³⁷. To measure the effects of targeting L-type VDCC’s, Nystoriak et al. selectively targeted these receptors with nimodipine, which prevented SAH-induced increases in Ca²⁺ and vascular tone in rats³⁴. These authors further demonstrated that enhanced arterial tone after SAH was driven by membrane depolarization³⁴.

Additional Mechanisms of Microcirculatory Dysfunction

Though NO and Ca²⁺ are major mediators of microvessel dysfunction, these molecules act via additional pathways to exacerbate secondary brain injury. For instance, NO conversion to s-nitrosoglutathione compounds after SAH has been implicated in the pathogenesis of EBI-related cellular stress and apoptosis⁷¹. NO may also exacerbate post-hemorrhagic neuronal apoptosis through modulation of caspase-12⁷². Calcium sensing receptor (CaSR) has recently received attention in EBI, as Wang et al. found that CaSR expression increased in neurons, astrocytes, and microglia after experimental SAH in mice⁷³. Further study demonstrated that CaSR agonism promoted GCE and deterioration of neurological function, while CaSR antagonism was neuroprotective⁷³. Finally, the hydrogen sulfide system is involved in anti-apoptotic signaling and may serve as a target in EBI due to its demonstrated ability to modulate L-type VDCC’s and improve neurologic function after SAH in a rat model⁷⁴. This study further demonstrated that administration of a calcium channel agonist weakened the therapeutic effect of hydrogen sulfide, providing insights to a potentially targetable pathway⁷⁴.

Blood-Brain-Barrier Disruption

Acute Blood-Brain-Barrier Disturbances

Molecular pathways leading to the phenotype of BBB disruption are initiated immediately following the entry of blood into the subarachnoid space, ventricles, and parenchyma. Blood products activate TLR-4, leading to transcription of tumor necrosis factor α (TNF- α), IL-1β, IL-6, IL-8, and IL-12 in a nuclear factor kappa-B (NF-κB)-dependent manner⁴². NF-κB and inflammatory cytokines up-regulate matrix metalloproteinase-9 (MMP-9), an enzyme that degrades zonula-occludens 1 (ZO-1) and other structural proteins⁷⁵. This compromises the structural integrity of tight junctions, accelerating the development of BBB disruption (Figure 2). In addition, MMP-9 degrades components of the extracellular matrix (ECM) including collagen IV, fibronectin, and laminin⁴². Finally, some evidence suggests that BBB disruption occurs in a biphasic manner, with early BBB disruption reflecting the primary disturbances of SAH and late BBB disruption driven by secondary apoptosis¹⁷.

Figure 2.

MMP-9 stimulation in EBI leads to the degradation of ZO-1 and other components of the ECM, causing disruption of the BBB and allowing unregulated paracellular movement. Figure created using Biorender.com.

Metalloproteinases in Blood-Brain-Barrier Disruption

Pathways modulating MMP-9 are the subject of recent studies, which have clarified its relationship with other mechanisms of EBI. Triggering receptor expressed on myeloid cells 1 (TREM-1) is a neuroinflammatory amplifier upregulated following SAH that influences MMP-9 activity⁷⁶. Sun et al. found that TREM-1 promoted secondary brain injury by both stimulating neuroinflammation and inducing MMP-9-mediated degradation of ZO-1 in rats⁷⁵. Stability of the BBB is also impacted by oxidative cascades. For example, NF-E2-related factor 2 (Nrf2) is an antioxidant transcriptional regulator essential for protection against oxidative stresses, and was identified by Guo et al. as an important upstream target for minimizing BBB disruption in a rat model^{31, 77}. In another animal study, ursolic acid, a modulator of oxidative stress, was found to reduce BBB disruption in a dose-dependent manner⁷⁸.

Targeting Blood-Brain-Barrier Pathways

Because of its function, targeting MMP-9 can reduce BBB disruption⁷⁹. Xu et al. recently demonstrated that direct MMP-9 inhibition with celastrol promoted maintenance of BBB integrity, minimized brain edema, reduced peripheral albumin leakage into the CSF, and preserved neurological status in a rat study of SAH⁴⁴. Antagonism of TLR-4 has also shown promise for reducing disruption of the BBB in EBI by reducing downstream MMP-9 transcription and activation⁸⁰. Additional targets have been identified to reduce BBB disruption. Anti-vascular endothelial growth factor (VEGF) antibodies can attenuate BBB disruption, either via direct targeting of VEGF or the VEGF receptor⁸¹, perhaps due to VEGF-A serving as an inducer of BBB disruption⁴².

Cerebral Edema

Cerebral edema in EBI can be broadly divided into two categories: cytotoxic edema and vasogenic edema. Both forms of cerebral edema contribute to the phenotype of global edema commonly observed in the EBI period.

Cytotoxic Edema

Cytotoxic edema is characterized by cellular swelling due to excess water, and in EBI, results from water and ion channel dysfunction involving aquaporin 4 (AQP4) in astrocytes, pericytes, and other cells⁵¹. At baseline, AQP4 assists in the maintenance of osmotic homeostasis⁸². However, altered cell permeability during EBI causes dysfunction of these transporters to and worsens cerebral edema, as the aquaporins cannot function appropriately and eliminate water from the brain⁸³. The precise role of AQP4 in worsening EBI-associated edema remains unclear, as animal models have demonstrated that complete AQP4 deficiency worsens cerebral edema while partial antibody blockade of AQP4 improves cerebral edema⁸³. The processes contributing to cerebral edema are amplified by downstream effects of neuroinflammation and oxidative cascades⁵¹. In particular, microglial activation has been identified as a major cause of AQP4 disorganization and subsequent edema in EBI⁸⁴. Therapeutic targeting of AQP4-related pathways has demonstrated potential in preclinical models. Recently, Hou et al. examined the effect of nimodipine on the glymphatic system in mice and found that nimodipine conferred neuroprotective effects and reduced brain edema by mitigating glymphatic dysfunction in a cAMP/PKA-mediated pathway that modulated AQP4 expression, position, and polarization in pericytes⁸⁵.

Vasogenic Edema

Another aspect of cerebral edema that has garnered attention is the activity of VEGF⁴⁸. In the state of cerebral ischemia that follows SAH, baseline degradation of hypoxia inducible factor-1 (HIF-1) is reduced, leading to upregulation of VEGF⁸⁶. VEGF then induces the formation of capillary fenestrations, which allows for further leakage of fluid, worsening vasogenic edema⁸⁷. Liu et al. demonstrated that inhibition of VEGF reduced neurological impairments and brain edema following SAH using a mouse model⁸⁸. In a study targeting upstream mediators of VEGF, Xu et al. observed reduced VEGF expression and reduced brain edema when HIF-1 blockade was performed in rats⁸⁹.

Neuroinflammation

Following SAH, the brain carries a markedly pro-inflammatory phenotype, which progressively develops after ictus⁹⁰. First, acute signaling is initiated with the induction of inflammatory cytokines and the NLRP3 inflammasome. Acute inflammatory signaling is modulated by both resident cells and infiltrating cells. Second, sustained responses occur through microglial polarization, which enables these intrinsic CNS immune cells to mediate chronic inflammation or reparation after SAH.

Early Signaling in Neuroinflammation

Characterizing the early inflammatory phenotype after SAH are the cytokines IL-6, IL-10, TNF-α, and IL-1β, which are produced both by resident cells and infiltrating cells²⁸. Synthesis of these cytokines is contingent upon signaling through TLR-4, which is highly expressed in the CNS by microglia, astrocytes, and endothelial cells at baseline, and by neurons in pathologic states⁹¹. After SAH, TLR-4 expressed by these cells is first activated by blood products including heme and oxyhemoglobin (Figure 3). The majority of inflammatory pathways contributing to EBI act through the transcription factors NF-κB, mitogen-activated protein kinase (MAPK), and interferon regulatory factor (IRF) in a MyD88/TRIF-dependent manner²⁸. It is believed that TLR-4 signals through the signaling adapter MyD88 in early stages of neuroinflammation and through the signaling adapter TRIF in later stages of neuroinflammation^{55, 91}. Human studies have demonstrated that elevated TLR-4 levels in peripheral macrophages are associated with poor functional recovery after SAH⁹².

Figure 3.

Hemorrhagic blood products stimulate neuroinflammation through TLR-4 signaling, which occurs in a MyD88-dependent manner early in EBI and a TRIF-dependent manner later in EBI. Figure created using Biorender.com.

NLRP3 Inflammasome-mediated Neuroinflammation

Another major driver of neuroinflammation is the NLRP3 inflammasome, which is composed of a sensor (NLRP3), an adaptor (ASC), and an effector (caspase 1)⁹³. The NLRP3 inflammasome is activated in response to cellular damage signals and regulates the cleavage of pro-IL-1β and pro-IL-18 by caspase 1, which generate active forms of each cytokine, subsequently amplifying NF-κB activation, gene transcription, and inflammation⁹³. Several studies have reported that regulating NLRP3 activity may be a promising approach to reduce EBI by altering microglial reactivity and reducing inflammation-associated vascular dysfunction. For example, Wei et al. reported that brain tissue damage was mitigated by inhibiting NLRP3 activation with RVD1 in a rat endovascular perforation model, while Dodd et al. used the selective NLRP3 inhibitor MCC950 to demonstrate reduced inflammation, cerebrovascular dysfunction, and tight junction disruption in a murine anterior circulation autologous blood injection model^{93, 94}. Further study of the NLRP3 inflammasome’s role in EBI should be pursued given its apparent role at the intersect of neuroinflammation and other secondary brain injury pathways⁹³.

Microglial Polarization in EBI

The role of microglial polarization in neuroinflammation and EBI has received increased attention in recent years⁷⁵. At rest, microglia exhibit the M0 phenotype, but are primed to polarize towards one of two terminal phenotypes. Once polarized to a given state, microglia perform specific functions in response to sustained inflammatory or anti-inflammatory signaling. Polarization from the M0 to the M1 phenotype is promoted by NF-κB-related pathways, while polarization from the M0 to the M2 phenotype is promoted by IL-4⁹⁵. M1 microglia are pro-inflammatory and secrete the cytokines IL-6, TNF-α, IL-1β, and IL-16⁹⁵. Survival of M1 microglia is promoted by inducible nitric oxide synthase (iNOS), which blocks certain forms of cell death⁹⁶. Targeting of iNOS therefore provides an opportunity to attenuate neuroinflammation⁹⁶. M2 microglia are anti-inflammatory and secrete the cytokines TGF-β, IL10, and CD 206⁹⁵. The M2 phenotype is neuroprotective, serving key functions of promoting neurogenesis, axonal reparation, angiogenesis, and re-myelination (Figure 4). Because of the potential role of M2 microglia in attenuating EBI, Wei et al. suggested that modulating microglial polarization towards the M2 phenotype may provide a promising strategy to reduce neuroinflammation⁹⁷.

Figure 4.

Polarization of M0 to the M1 phenotype is mediated by numerous signaling molecules associated with neuroinflammation, while IL-4 primarily mediates polarization to the M2 phenotype. Figure created using Biorender.com.

Oxidative Cascades

Organelle Dysfunction

Organelle dysfunction following SAH is a major source of free radicals⁹⁸. In the state of relative ischemia after SAH, mitochondria are impaired and produce excess free radicals that overpower intrinsic antioxidant systems and damage mitochondrial proteins, DNA, and lipids⁹⁹. The resulting increase in mitochondrial membrane permeability promotes apoptosis of neurons via leakage of mitochondrial proteins including cytochrome C, which initiates apoptotic cascades mediated by caspases and Bax¹⁰⁰. Mitochondrial damage in EBI is amplified by calcium ion accumulation, which disrupts the mitochondrial membrane potential. Mitochondria attempt to restore oxidative balance via substrate oxidation and oxygen consumption, but are overwhelmed by excess free radicals¹⁰¹. To counter the deleterious downstream effects of mitochondrial dysfunction, cells employ mitophagy, selective autophagy of defective mitochondria, to maintain neuronal homeostasis¹⁰². However, excess mitochondrial fission worsens the production of free radicals⁹⁸.

Mitochondrial Reactive Oxygen Species

Reactive oxygen species (ROS) of mitochondrial origin are known to be potent activators of the NLRP3 inflammasome and other mediators of EBI, initiating many deleterious pathways¹⁰³. These free radicals also cause DNA damage that activates extrinsic cell death pathways¹⁰⁴. Recently, Li et al. demonstrated that minocycline protected against NLRP3 inflammasome-induced inflammation as well as P53-associated apoptosis in an endovascular perforation model of rats¹⁰³. Further study of NLRP3 activated by oxidative damage has shown that preventing the downstream effects of NLRP3 is a promising therapeutic strategy, with two recent studies utilizing the compounds pterostilbene and resveratrol as targeted inhibitors in rodent models^{52, 105}.

Endoplasmic Reticulum Stress

Intrinsic cellular systems are also involved in attenuating oxidative cascades. Overproduction of free radicals and excess calcium levels both lead to endoplasmic reticulum (ER) stress via disruption of protein folding and activation of the unfolded protein response (UPR)¹⁰⁶. The UPR terminates with activation of downstream cascades using pathways modulated by three major sensors: PKR-like ER kinase, inositol-requiring enzyme 1, and activating transcription factor 6¹⁰⁶. The behavior of the ER stress-UPR pathway is contingent upon the degree of ER stress, with low levels of ER stress initiating pro-survival signaling and severe stress causing apoptotic signaling^106,107. During the EBI period, ER stress was demonstrated as neuroprotective by Yan et al. in a rat model, likely via inhibition of apoptosis and activation of autophagy¹⁰⁷.

Hemoglobin-derived Oxidative Cascades

Extrinsic processes further amplify oxidative cascades initiated by intrinsic processes. Hemoglobin leads to free radical generation due to the oxidative effects of oxy-hemoglobin and freed iron, which produce superoxide and hydrogen peroxide¹⁰⁸. Freed iron is particularly damaging due to its ability to deposit in brain tissue and contribute to brain edema and neuronal death in the EBI period¹⁰⁹. Several preclinical studies have examined the use of the iron chelator deferoxamine after SAH, which significantly reduced brain edema and neurologic deficit in multiple studies, thus reflecting the key role iron overload plays in EBI and revealing a promising therapeutic target^{110, 111}.

Free Radical Targets

Recent studies have identified additional sources of free radicals that may serve as viable therapeutic targets. For example, Won et al. observed that the production of arachidonic acid derivatives by the enzymes cyclooxygenase and lipoxygenase generated superoxide anions as byproducts following SAH¹¹². Free radical accumulation is also worsened by excess neuroinflammation since infiltrating neutrophils produce ROS during their oxidative bursts¹¹³. Indeed, Ye et al. demonstrated that inhibition of leukotriene B4 reduced oxidative damage in EBI by minimizing neutrophil activity in rats⁵³(Supplement 3).

Cell Death

Apoptosis

Mechanisms of cell death in EBI are heterogeneous and begin with apoptosis. Friedrich et al. observed neuronal apoptosis as early as 10 minutes following SAH in rats, making it the earliest form of cell death in EBI¹¹⁴. Apoptosis occurs via two major pathways: the intrinsic pathway and the extrinsic pathway⁶⁴. The intrinsic pathway is caspase-dependent and occurs due to cytochrome c leakage from the mitochondria, while the extrinsic pathway is caspase-independent and is initiated by the binding of either TNF or Fas ligand to death receptors¹¹⁵. Activation of caspase-3 into cleaved caspase-3 is stimulated by pathways including the PI3K/Akt pathway and the ENT1/NRLP3 inflammasome pathways¹¹⁵. Stimulation of the anti-apoptotic protein Bcl2 and inhibition of the pro-apoptotic protein Bax have both been demonstrated to minimize apoptosis and improve neurological outcomes in animal models of EBI¹¹⁶. Mounting evidence has identified caspase inhibition as a promising strategy to ameliorate EBI following SAH¹¹⁷.

Autophagy

Autophagy is an energy-dependent pathway that recycles damaged organelles and cellular debris via formation of double-membraned autophagosomes that fuse with lysosomes¹⁰². In EBI, autophagy occurs due to disordered organelle function induced by ischemia, and has traditionally been viewed as a vital process promoting neuronal survival¹¹⁸. Studies have demonstrated that promoting autophagy attenuates brain injury during EBI, and that inhibition of autophagy exacerbates brain injury^{65, 119}. Despite the apparent benefits of stimulating autophagy, excess autophagy causes autophagic cell death¹²⁰. Wang et al. demonstrated that fibroblast growth factor 2 (FGF2) can inhibit excess autophagy via activation of the PI3K/Akt pathway following SAH in rats¹²⁰.

Pyroptosis

Pyroptosis is a form of inflammatory cell death that is highly contingent on the activity of inflammasomes and the cytokines IL1β and IL18 to induce pyroptotic cell death¹²¹. In EBI, the NLRP3 and AIM2 inflammasomes are the best-studied mediators of pyroptosis¹²¹. Recently, Yuan et al. demonstrated that knockdown of caspase-1 using a lentiviral vector reduced pyroptosis after EBI in mice¹²¹. In another study, Chen et al. found that atorvastatin inhibited activity of both the NLRP3 and AIM2 inflammasomes in a mouse model of EBI, attenuating pyroptotic neuronal death¹²².

Necroptosis

Another form of cell death in EBI is necroptosis, which has features of necrosis and apoptosis¹²³. Necroptosis is mediated by TNFα and its receptor TNFR1, and is inhibited by necrostatin-1¹²⁴. Yang et al. demonstrated that inhibition of necroptosis using necrostatin-1 attenuated cerebral edema and improved neurobehavioral scores in an autologous blood injection model of rats¹²⁴. This effect may be partially explained by the relationship between necroptosis and disruption of the BBB¹²⁵.

Ferroptosis

Lastly, recent studies have identified novel mechanisms of cell death implicated in EBI¹⁰⁰. First described in 2012, ferroptosis is an iron-dependent form of programmed cell death that is characterized by excessive lipid peroxidation¹²⁶. This is accompanied by an impaired glutathione peroxidase pathway and diminished antioxidant activity in cells, leading to oxidative cell death. In EBI, inhibitors of ferroptosis including ferrostatin-1 have been demonstrated to reduce brain edema and neurological deterioration in in-vivo models¹²⁷. Qu et al. showed that inhibition of ferroptosis reduces neuroinflammation, BBB impairments, and oxidative stress in a mouse model of EBI¹²⁸. More generally, iron-mediated toxicity following acute SAH has received considerable attention, with recent human studies correlating iron deposits in cortical gray matter to cognitive outcome¹²⁹.

Conclusion and Future Directions

EBI is a multifaceted form of secondary brain injury with a heterogeneous pathophysiology. Recent studies have confirmed the influence of several mechanisms and phenotypes that contribute to the poor neurological state following SAH. Future investigation of biomarkers should apply automated quantification of SSV to study the genetic basis of post-SAH acute brain injury in larger human cohorts. Promising molecular targets such as NO, Ca²⁺, MMP-9, AQP4, VEGFs, TLR and related cytokines, microglia, Nrf2, the NLRP3 inflammasome, ER stress pathwasy, and cell death pathways may demonstrate therapeutic benefit in future clinical studies and warrant further investigation (Table 1). Additional mechanisms contributing to EBI such as acute white matter injury should be explored in future studies¹³⁰.

Table 1.

Potential therapeutic targets and agents in EBI

Injury mechanism	Molecular mediators	Potential therapeutic agents
Microcirculatory disruption
	Endothelial nitric oxide synthase69	Fasudil, Atorvastatin, Simvastatin
	L-Type voltage dependent calcium channel34	Nimodipine
	Calcium sensing receptor74	Ronacaleret, encalaret, NPS-2143 (Calcilytics)
	Hydrogen sulfide75	Bay K8644, Ambroxol
	5-HT1B receptor11	Isamoltane hemifumarate
	Extracellular Ca2+ influx11	Magnesium
	NO-mediated pathways72	S-nitrosoglutathione
BBB disruption
	Matrix metalloproteinase-944	Celastrol, Osteopontin
	Nuclear factor erythroid 2-related factor 231, 78	Mitoquinone
	Toll-like receptor 479, 81	Ursolic acid
	Vascular endothelial growth factor82	Bevacizumab
	Zonula occludens-111	Argatroban, Hydroxyfasudil
Cerebral edema
	Aquaporin-485	Nimodipine
	Hypoxia-inducible facor-190	Perfluorooctyl-bromide
	Vascular endothelial growth factor11	PP1 (Src family kinase inhibitor)
	C-Jun phosphorylation, Aquaporin-411	SP600125
	Vascular endothelial growth factor, Aquaporin-411	Melatonin
Neuroinflammation
	Nuclear factor kappa B11	N-acetylcysteine, estradiol
	MyD8828	Progesterone, Resveratrol, ST2825
	TRIF28	Pepinh-TRIF, Resatorvid
	Toll-like receptor 480, 94	Eritoran, Naloxone, TAK-242
	Inducible nitric oxide synthase95	L-NIL, Aminoguanidine
	NLRP3 inflammasome53,97,104	RVD1, MCC950, pterostilbene, resveratrol, minocycline
	Peroxisome proliferator-activated receptor gamma96	Erythropoietin
	Apolipoprotein E pathway11	ApoE-mimetic protein
Oxidative cascades
	Heme and blood products62	Haptoglobin
	Free iron109	Deferoxamine
	Cyclooxygenase113	Aspirin, naproxen, ketorolac
	Lipoxygenase113	Zileuton, meclofenamate
	Oxygen free radicals11	Tetramethypyrazine
Cell death pathways
	Bcl2/Bax signaling117	Melatonin
	Caspases122	Z-IETD-FMK, Ac-DEVD-CHO
	Autophagy11	Rapamycin, simvastatin
	PI3K/Akt121	FGF2, PNU-282987
	NLRP3/AIM2 inflammasomes123	Atorvastatin
	Necroptosis125	Necrostatin 1
	Ferroptosis128	Ferrostatin 1

With the occurrence of EBI in patients increasingly measurable, and its underlying mechanisms well-defined, human studies and clinical trials are necessary to identify suitable therapeutics that target this form of secondary brain injury^{11, 19}. Active trials in SAH are still primarily focused on reducing DCI, and neglect therapeutic targets relating to EBI^{23, 131}. Prospective study targeting EBI will require substantial collaboration between translational scientists, experienced clinical investigators, and funding agencies. However, this is the clear next step to reduce the healthcare burden of SAH and improve patient outcomes.

Non-standard Abbreviations and Acronyms:

DCI

delayed cerebral ischemia

BBB

blood-brain-barrier

EBI

early brain injury

SEBES

subarachnoid hemorrhage early brain edema score

SSV

selective sulcal volume

GCE

global cerebral edema

CBFI

cortical blood flow insufficiency

VBM

voxel-based morphometry

WFNS

world federation of neurosurgeons

ROS

reactive oxygen species

VCDD

voltage dependent calcium channel

NOS

nitric oxide synthase

CaSR

calcium sensing receptor