Interleukin-6: Important Mediator of Vasospasm Following Subarachnoid Hemorrhage

Abstract

The correlation of neuroinflammation with the development of cerebral vasospasm following subarachnoid hemorrhage has been well documented in the literature; both clinical and pre-clinical. The exact mechanisms by which this process occurs, however, are poorly elucidated. Recent evidence indicates that interleukin-6 is not only an important prognostic biomarker for subarachnoid hemorrhage and subsequent vasospasm development but also an integral component in the progression of injury following initial insult. In this review, we briefly highlight other pathways under investigation and focus heavily on what has been discovered regarding the role of interleukin 6 and cerebral vasospasm following subarachnoid hemorrhage. A proposed mechanistic pathway is highlighted in written and graphical format. A discussion regarding the human correlative findings and initial pre-clinical mechanistic studies is addressed. Finally, in the future investigation section, innovative developments and a clear description of areas warranting further scientific inquiry are emphasized. This review will catalyze continued discovery in this area of emerging significance and aid in the quest for effective vasospasm treatment where limited clinical therapeutics currently exist.

1. INTRODUCTION

Vasospasm and delayed cerebral ischemia (DCI) are well-known complications of subarachnoid hemorrhage (SAH). 70% of patients with SAH that present to the hospital develop cerebral vasospasm on imaging over the course of treatment, and 30% have symptomatic vasospasm [1]. Little is known about why cerebral vasospasm occurs in a delayed fashion following SAH or the best methods to prevent it. Several proposed pathways have been postulated, which we briefly highlight in the alternative pathways section. Recent evidence points to an increased inflammatory cascade mediated by interleukin 6 (IL-6) as a key component of vasospasm development and subsequent DCI, which is the focus of this review [2]. Human data indicates CSF IL-6 surges at 3-5 days after SAH, often preceding vasospasm in a predictable manner [3]. In particular, this inflammatory surge has been associated temporally with blood-brain barrier (BBB) breakdown mediated through the disruption of tight junction proteins and by activation of endothelin-1 [4]. Furthermore, after the BBB breakdown occurs, toll-like receptor 4 signaling contributes to the initiation of apoptosis in microfoci adjacent to SAH in the surrounding parenchyma [5]. Blocking the apoptotic cascade within these regions in pre-clinical models has therefore been found to be beneficial in decreasing morbidity following SAH [6]. Additionally, targeting IL-6 directly prior to the initiation of the inflammatory cascade contributes to reduced development of vasospasm and the subsequent toll-like receptor response [7]. Further, intracisternal injection of IL-6, but not IL-8, is sufficient to drive vasospasm in dogs [8]. In this review, we highlight what has already been discovered regarding IL-6 and vasospasm development and provide insights regarding areas warranting further investigation in the quest to develop meaningful therapeutics.

2. ALTERNATIVE PATHWAYS

Cerebral vasospasm following SAH is a topic of growing clinical significance. Most work to date has looked at endothelial damage and physiologic changes in vascular responsiveness [9]. As such, measurements of vasoconstrictor and vasodilator substances such as nitric oxide, endothelin, arachidonic acid metabolites, and prostaglandins have been heavily investigated [10]. Furthermore, changes in vascular smooth muscle contractility have been examined by looking at G protein-coupled receptor changes and myofilament Ca²⁺ responsiveness [11]. Only recently has investigation expanded towards glial-centric mechanisms [12]. The glia/endothelial interface is emerging as an important component in the initiation of vasospasm.

Regarding this interface, a pathway of significance is endogenous nitric oxide synthetase (eNOS). Immediately following SAH, eNOS is downregulated predisposing to vasospasm [13]. The response is most pronounced in patients with CC genotype making them most susceptible to vasospasm [14]. Statins have been found to upregulate this pathway early, providing reduced vasospasm in pre-clinical models [15, 16]. eNOS is upregulated naturally in a delayed fashion at the astrocyte footplate/endothelial junction following SAH and helps contribute to nitric-oxide production, ultimately alleviating vasospasm [17]. A topic of growing importance is downstream mechanisms that are induced by eNOS and nitric-oxide that cause subsequent vasodilation.

A separate topic that has emerged in the literature is the role of oxidative stress in vasospasm. Yang and colleagues have proposed that sources of oxidative stress following SAH include upregulated enzymatic activity, disruption in mitochondrial respiration, and degradation of extracellular hemoglobin [18]. Wu and colleagues have shown that the release of H2O2 by the mechanisms above directly induces cerebral vasospasm and, when blocked, prevents sustained vasoconstriction [19]. A proposed mechanism is a direct increase of lipid peroxide by H₂O₂ [20]. In rats, oxidative stress contributes to endothelial cell injury by peroxidation, which causes sustained vasoconstriction [21]. Likewise, human data support that patients with hyperoxemia and increased oxidative free radicals are more likely to develop vasospasm during hospital admission [22].

A final area of scientific and clinical interest is the role of apoptosis in vasospasm and, more importantly, DCI. Apoptosis begins at similar time points to vasospasm in mouse models of SAH [23]. Cahill and colleagues support that p53-induced apoptosis may be a key contributor to the onset of vasospasm [24]. This has been further verified in a canine model where p53 induced the caspase cascade, which particularly selects endothelial cells and initiates vasospasm [25]. Not surprisingly, caspase inhibitors block endothelial apoptosis and have been shown to decrease vasospasm [26]. In addition to early endothelial injury, Hanafy found that microglia induce activation of toll-like receptors and contributes to delayed apoptosis and vasospasm, which may account for the wide window of clinically reported vasospasm [27]. Recent data indicates that patients with impaired nuclear factor-like 2 (Nrf2) regulation are likely to have reduced antioxidant genes and thus are most susceptible to delayed injury from apoptosis [28]. The mechanisms linking vasospasm to DCI are still being elucidated, but the permanent changes induced by the apoptotic cascade are thought to be an important component [29]. The remainder of this review will focus on the interleukin-6 cascade.

3. INTERLEUKIN-6 CASCADE

Previous groups have postulated that sphingomyelinase cleavage of sphingomyelin initiates the IL-6 surge as it does in peripheral organs, but this was proven not to be the key mechanism within the brain following SAH [2]. Passive diffusion of peripheral IL-6 into the brain does not occur, and it is thought that the IL-6 surge is intrinsic to the brain itself [30]. A potential cause is that IL-6 levels vary significantly within the CSF verse plasma following SAH, likely due to the relatively protected BBB interface [31]. H₂S-producing enzymes released from the milieu surrounding subarachnoid blood can damage endothelial barriers, thereby upregulating endothelin-1 activity [32]. Further discussion on how IL-6 is activated within the brain once endothelin-1 is upregulated is addressed in detail below in the preclinical section. Briefly, endothelin-1 upregulation acutely disrupts tight junction proteins causing a leaky BBB interface. This sudden BBB disruption causes an acute inflammatory surge mediated by reactive astrocytes in an otherwise immune-privileged brain. Migroglia also expresses IL-6 after exposure to blood products, as occurs after SAH [33]. This brain-intrinsic inflammatory surge causes significant IL-6 release and triggers the downstream cascades linked to vasospasm. Matrix metalloproteinase 9 activity has uniquely been shown to be upregulated at similar times to IL-6 within the brain following SAH [34]. This upregulation may disrupt the local cytoarchitecture and allow the recruitment and migration of microglia to regions adjacent to subarachnoid blood [35]. The STAT3 pathway is upregulated by IL-6 sequentially and is known to be important in the microglial phenotype switch that occurs after SAH [36]. Not surprisingly, the increases in IL-6 shortly precedes the surge in monocyte chemoattractant-1 (MCP-1) and vasospasm [37]. Postulated mechanisms for the link to MCP-1 include MAPK and STAT3 signaling initiated through the membrane-bound IL-6 receptor [38]. The pathway has been shown to cause peripheral macrophage recruitment, which is likely influential in triggering vasospasm following SAH due to cross-talk between microglia and peripheral macrophages [39]. Further work is needed to understand this important interaction and the dynamic response of cerebral vasculature. Once vasospasm occurs, it triggers an upregulation in toll-like receptor 4, which also has mechanistic links to the STAT3 pathway. Toll-like receptor 4 signaling induces the apoptotic cascade that results from delayed cerebral ischemia associated with vasospasm [27].

4. HUMAN STUDIES

Not only has IL-6 been shown to increase in SAH patients, but CSF levels are highest in patients that subsequently develop symptomatic vasospasm [40, 41]. The key predictive marker is a peak in IL-6 starting on day 3 following SAH but extending to day 5 [42, 43]. The increased CSF levels of IL-6 within CSF have been directly correlated to DCI [44, 45]. Furthermore, patients with the highest CSF levels of IL-6 were more likely to require shunt placement due to hydrocephalus [46]. Interestingly, ICP spikes >20 were associated with a sudden CSF spike of IL-6 as well [47]. Multivariate analysis has shown that a CSF IL-6 surge is the single most predictive marker for overall poor outcome following SAH [48]. Serum IL-6 can also remain elevated for weeks in patients with DCI [49]. In light of these findings, several groups have proposed IL-6 as a prognostic biomarker and a potential therapeutic target [50]. A percent change or threshold cut-off has yet to be determined, however, and warrants further investigation. One of the complicating features for clinical trials is varying levels of baseline IL-6 due to patient comorbidities and/or genetic factors [51]. Because of these high pre-injury variations, serum IL-6 has been unreliable in predicting vasospasm or delayed injury, lending credence for the need of CSF IL-6 as a more tangible biomarker [52]. This was further verified by Sarrafzadeh and colleagues comparing serum and CSF IL-6 levels in 38 SAH patients showing the importance of CSF levels over plasma levels as a predictive tool [53]. Even though systemic IL-6 levels have proven to be a poor biomarker for acute events, elevated serum IL-6 does indicate poor overall outcome and increased likelihood for systemic issues [49]. Pilot studies have shown that anti-inflammatory medications that decrease IL-6 are associated with better outcomes following SAH, but underlying mechanistic studies have yet to be conducted [54].

5. PRE-CLINICAL STUDIES

In a rat model, Song and colleagues found that dexmedetomidine drastically reduced IL-6 and subsequently the severity of vasospasm [55]. Inhibiting IL-6 in a rabbit model of SAH also reduced vasospasm and DCI [56]. A proposed mechanism for the IL-6 response is through activation of endothelin-1 [57]. Endothelin-1 has been linked to disruption of tight junction proteins contributing to microfoci disruptions of the BBB [58]. Blocking the initial BBB disruption has reduced the IL-6 response [59]. This beneficial effect is thought to be mediated by limiting exposure of IL-6 receptors at the brain endothelial barrier to surrounding inflammation from reactive astrocytes [4]. If BBB disruption does occur, then endothelin-1 indirectly activates membrane-bound IL-6 receptors adjacent to areas where subarachnoid blood has accumulated and allows for interaction with the inflammatory milieu from the irritated brain parenchyma. The particular cell type involved with IL-6 release is reactive astrocytes [60]. Matsumoto and colleagues found using a rat model that IL-6 receptor activation is directly linked to histidine-rich glycoprotein, which is found in plasma and only interacts with the IL-6 receptor at sites of BBB disruption mediated by endothelin-1 [61]. The release of IL-6 is thought to be from reactive astrocytes that occur near sites of BBB disruption [62]. The binding of IL-6 to the activated IL-6 receptor has been shown to trigger the STAT3/SOCs3 cascade in a rat model of SAH [63]. The cascade in addition to recruiting microglia and influencing their phenotype switch as outlined above activates the toll-like receptor 4 system [64]. In areas of vasospasm, the toll-like receptor system begins the process of apoptotic progression and leads to neuronal cell death in the susceptible injured brain parenchyma [65]. This has been directly linked to the manifestation of DCI [66]. A proposed mechanistic cascade linking IL-6 to vasospasm is outlined in Fig. (1).

Fig. (1).

Mechanistic pathway showing how interleukin-6 is activated and the role it plays in development of cerebral vasospasm.

6. FUTURE INVESTIGATION

Simon and Grote aptly point out that limited work has been done to understand the mechanistic link between elevated IL-6 and symptomatic vasospasm [67]. Recent advancements in aggregation assays have allowed better quantitative measurement of IL-6 with a fast turn-around time [68]. Dengler and colleagues have shown that utilizing the lateral flow immunoassay chip-test, they can have data within 20 minutes of sample collection [69]. Further ELISA characterization can determine the functionality of the released IL-6 within the CSF [70]. Another promising option is in vivo real-time measurements of IL-6 changes via cerebral microdialysis monitors [71]. An area of emerging importance is how IL-6 levels change after clipping versus coiling [50]. It will be important to have viable time-sensitive assays in order to advance towards clinical trials. As alluded to before, an area of vital importance in transitioning towards clinical trials is establishing a percent change threshold for IL-6 that could be used as an early warning sign for cerebral vasospasm. This will likely require multi-institutional collaboration in order to garner sufficient patient data and to confirm inter-region validity.

Another area requiring investigation is understanding the time course of the IL-6 response [72]. Several groups have looked at individual components of the pathway as outlined in the pre-clinical section, but an adequate mechanistic time course linking IL-6 to vasospasm has yet to be determined. Additionally, the exact cell types involved have not been clearly delineated. A murine model seems the most feasible to tease out the intricacies of the mechanism as both inhibitory pharmacology as well as genetic knockout can be employed [73]. Parra and colleagues have shown that cerebral vasospasm occurs in a comparative time window in mice to that observed in humans following SAH [74]. This is appealing in that mechanistic links as well as proposed therapeutic interventions can be readily correlated. Another advantageous consideration is the incorporation of multidisciplinary collaboration where both clinical and pre-clinical staff bring individual expertise to improve therapeutic relevance [75].

CONCLUSION

Vasospasm continues to be a significant clinical concern for patients suffering from SAH. Current options for the treatment of symptomatic vasospasm are limited. Correlative studies have shown a strong association between the inflammatory surge following SAH and the subsequent development of cerebral vasospasm. IL-6 has emerged as an important cytokine not only to be used as a biomarker but also integral in the mechanistic progression towards vasospasm. In this review, we evaluated both the clinical and pre-clinical data for the role of IL-6 in the development of vasospasm. By compiling and synthesizing the available studies, a working mechanistic pathway has been proposed. An improved understanding of the mechanism and timing of events will aid in designing the initial clinical trials where IL-6 percent change thresholds can be flushed out. Furthermore, good work is warranted regarding the potential therapeutic benefit of targeting IL-6 to prevent cerebral vasospasm.