Introduction

Delayed cerebral vasospasm (DCVS), early brain injury (EBI), and delayed cerebral ischemia (DCI) are potential devastating complications after aneurysmal subarachnoid hemorrhage (SAH). Despite recent advances in the treatment of SAH, neurological outcomes remain poor, and new therapeutic approaches for preventing and treating DCVS and DCI are highly warranted [1,2,3]. Randomized controlled trials using clazosentan, an ET-1 receptor antagonist, showed a decrease of DCVS but no improvement in the outcome of SAH [4, 5]. This led the scientific community to reevaluate factors other than DCVS that might cause poor clinical outcome after SAH. Inflammation after SAH has been revealed to be an important contributor leading to DCVS, DCI, and worse outcome. Various inflammatory cytokines and interleukins have been described as initiators of the inflammation cascade after blood products are spilled in the subarachnoid space after aneurysm rupture that play a fundamental role in EBI [6,7,8,9]. Preclinical and clinical studies revealed that interleukin (IL)-6 seems to be one of the interleukins involved in the inflammatory response after SAH. Moreover, a growing body of clinical studies has described a strong correlation between IL-6 and worse outcome, corroborating previous findings described in preclinical animal studies where IL-6 was also showed to correlate with DCVS and neuronal cell death [6, 7, 9,10,11,12]. The aim of this study was to perform a systematic review of preclinical and clinical studies that evaluated systemic and cerebral IL-6 levels after SAH and their relation to DCVS, neuronal cell death, and DCI.

Material and methods

Two separate systematic literature searches in the Medline/PubMed database were conducted in January 2021 according to the Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA) guidelines [13]. In both searches, two investigators (DC and SS) independently screened the titles and abstracts and identified articles based on the predefined eligibility criteria following the PRISMA guidelines [13]. The final articles included were selected based on the full text of the eligible studies. Discrepancies in the study selection were discussed between the authors DC and SS until a consensus was reached. Institutional review board approval was not required for this review of published data.

First, a literature search was performed to identify preclinical studies investigating the role of IL-6 in the pathophysiology of DCVS and DCI after SAH. The keywords used were “interleukin-6” and “subarachnoid hemorrhage” using the Boolean operator (AND). The search was restricted to animals by using the MEDLINE PubMed limit “animals.” Eligible studies were in vivo experimental mice, rat, rabbit, cat, dog, pig, goat, and nonhuman primate studies in which IL-6 was investigated in cerebrospinal fluid (CSF), brain parenchyma, brain vessels, and plasma after SAH. Studies on extracranial vessels, organs other than the brain, or in vitro were excluded. From the eligible preclinical articles, we extracted publications details (i.e., authors, journal, date), animal species, SAH induction techniques, and outcome measures (DCVS, DCI, direct vessel observation, neuronal cell death and degeneration, blood–brain barrier disruption, contractile vessel response).

The second literature search was performed to identify clinical studies investigating IL-6 and occurrence of DCVS and DCI after SAH and assessing clinical outcome. The keywords used were “interleukin-6” and “subarachnoid hemorrhage” using the Boolean operator (AND). The search was restricted to humans using the MEDLINE PubMed limit “humans.” The inclusion criteria were all studies evaluating plasma and CSF levels of IL-6 in relation to DCVS, DCI, and patients’ outcome. Studies investigating possible therapeutics approaches other than direct inhibition of IL-6, and relations to infections and hydrocephalus or systemic complications were excluded. From the eligible clinical articles, we extracted publications details (i.e., authors, journal, date), number of patients, IL-6 measurement methods (CSF, plasma), and primary and secondary outcome parameters (DCVS, DCI, clinical outcome).

Results

Animal studies

After duplicates were removed, a total of 61 articles were found that analyzed the effect of IL-6 in animal models (Supplementary Table 1). The initial electronic search yielded 98 potential studies. Of these, 35 articles were excluded after reviewing the title and abstract. The remaining 63 articles underwent full-text analysis. Of those, 2 studies were excluded per the predefined eligibility criteria (Fig. 1). Of the final 61 articles, 44 studies were performed on rats (72.1%), 9 on mice (14.8%), 5 on rabbits (8.2%), and 3 on dogs (4.9%). In rats, the most common SAH model used was the cisterna magna injection model in 19 studies (43.2%), followed by the endovascular perforation model in 18 studies (40.9%) and the prechiasmatic blood injection model in 7 studies (15.9%). In mice, the endovascular perforation model was used in 7 studies (77.8%) and the prechiasmatic blood injection model in 2 studies (22.2%). In rabbits, 3 studies (60%) used the closed cranium extracranial-intracranial shunt model and 2 studies (40%) the double hemorrhage model. In dogs, the models used were the cisterna magna blood injection model in 2 studies (66.7%) and the endovascular perforation model in one study (33.3%) (Fig. 2).

Fig. 1
figure 1

Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) flow chart for preclinical animal studies selection

Fig. 2
figure 2

Subarachnoid hemorrhage model used in various animal species studies

Nearly three-quarters of the studies (45, 73.7%) addressed the research question of EBI, whereas 24 studies (39.3%) examined DCVS. Twelve studies (19.6%) addressed both EBI and DCVS, and four studies (6.6%) did not focus their research question on DCVS and EBI, instead examining only interleukin concentrations.

Overall, IL-6 was measured in brain parenchyma in 36 studies (59%), in plasma/serum in 16 studies (26.2%), in CSF in 14 studies (22.9%), and in cerebral arteries and endothelial cells in 7 studies (11.4%) (some studies measured IL-6 in multiple locations) (Table 1). Of the 36 studies in which IL-6 was measured in brain parenchyma, 34 (94.4%) found an increased expression of IL-6 in parenchyma after SAH. In all the studies in which IL-6 was measured in CSF, an increase of IL-6 expression after SAH was noted. Of the 16 studies in which IL-6 was measured in plasma/serum, 13 (81.3%) showed an increase of IL-6 after SAH induction. An increase of IL-6 expression in cerebral vasculature after SAH was described in all of the studies where IL-6 was measured in the cerebral vasculature. Overall, 52 studies (85.2%) aimed their research on a possible therapeutic substance after SAH and their effect on IL-6 as well (Supplementary Table 1); however, only two studies (3.2%) analyzed the direct inhibition of IL-6 through a receptor antagonist. The first of these analyzed the effect of a direct IL-6 antagonist in vivo with the IL-6 receptor antagonist tocilizumab, which resulted in decreased DCVS, neuronal cell death, and microclot formation [14]. In the second study, an IL-6 antagonist was applied only vitro, with blockade of the membranous IL-6 receptor with a goat polyclonal IL-6R-neutralizing antibody (IL-6R nAb), and resulted in an inhibition of brain endothelial cell barrier disruption after SAH [15].

Table 1 Animal and human studies measuring interleukin-6 in cerebrospinal fluid, plasma/serum, parenchyma, and the cerebral vasculature

Human studies

After duplicates were removed, a total of 30 articles were found that analyzed the effect of IL-6 in clinical studies (Supplementary Table 2). The initial electronic search yielded 105 potential studies. Of these, 25 articles were excluded after reviewing the title and abstract. The remaining 80 articles underwent full-text analysis. Of those, 50 studies were excluded per the predefined eligibility criteria (Fig. 3).

Fig. 3
figure 3

Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) flow chart for human studies selection

On average, 55.9 (7–179) patients were included per study, and 28 (93.3%) of the studies were prospective cohorts. In 17 studies (56.7%), IL-6 was measured in CSF; in 23 studies (76.7%), it was measured in plasma/serum; and in 10 studies (33.3%), it was measured in both (Table 1).

An increase of IL-6 after SAH was found in all of the studies where IL-6 was measured in CSF or plasma/serum, but the increased concentration of IL-6 was especially marked in CSF. In fact, in all 9 studies where IL-6 was measured in both CSF and plasma/serum, significantly higher values of IL-6 were demonstrated in CSF compared with plasma/serum.

Of the 13 studies where an association between DCVS and IL-6 concentrations in either CSF or plasma/serum was analyzed, 9 (69.2%) described a statistically significant association between IL-6 and DCVS, with an additional study showing a positive trend. On the other hand, of 15 studies (50%) where a possible correlation between IL-6 and the occurrence of delayed ischemic neurologic deficit (DIND) and DCI was analyzed, 10 studies (66.6%) described a statistically significant association between IL-6 concentrations (either in CSF or plasma/serum) and DCI/DIND, with additional 2 studies (13.3%) showing a positive trend. Of 16 studies in which an association between IL-6 concentrations and outcome was analyzed, 12 studies (75%) showed that higher concentrations of IL-6 either in CSF or plasma/serum corresponded with worse patient outcome, with an additional study showing a trend of higher CSF IL-6 concentrations in patients with worse outcome. No studies analyzed any kind of direct inhibitor of IL-6 as therapeutic agent.

Discussion

In this systematic review, we identified 61 animal studies that evaluated the effects of IL-6 in SAH models. Nearly all of the studies in which IL-6 was measured in CSF and parenchyma—100% and 94.4%, respectively—showed increased expression of IL-6 after SAH induction. Moreover, an increased expression of IL-6 in plasma/serum was described in 81.3% of the studies. These results were mirrored in human studies. All clinical studies in which IL-6 was measured in CSF or plasma/serum demonstrated an increase of IL-6 after SAH. Both in animal models and in humans, IL-6 concentrations after SAH have been described to be significantly higher in CSF compared with plasma/serum, further illustrating that the main inflammatory reaction after SAH is mostly compartmentalized intracranially, as a response to the presence of blood in the subarachnoid space [16].

IL-6 in relation to DCVS and DCI

Osuka et al. [12] described an increase of IL-6 in CSF after SAH, which was considerably higher in patients experiencing DCVS. In the same study, they described that a single application of IL-6 in the CSF of dogs after SAH caused DCVS. Similar results were reported by our research group as well: we observed that the intrathecal application of IL-6 in rabbits led to an increased expression of the vasoconstrictor endothelin (ET)-1 and to DCVS [6].

The effect of IL-6 on vasoconstriction on arteries was demonstrated by Bowman et al. [17] in a rat femoral artery vasospasm model, where a significant increase of IL-6 was found in blood-exposed spastic arteries and the administration of a polyclonal antibody against IL-6 reduced vasospasm. Apoptotic damage to the endothelium and smooth muscle cells of cerebral arteries is known to be a factor in obstructing the physiologic vasoregulation and breaking down the blood–brain barrier after SAH [18,19,20,21]. In this context, IL-6 has been shown to be an important mediator in apoptosis, and barrier disruptive effects and inhibition of IL-6 attenuated vascular inflammation and endothelial barrier disruption of retinal endothelial cells [22]. Aihara et al. described higher mRNa expression of IL-6 in the endothelial wall of spastic basilar artery in dogs after SAH, a finding that was confirmed by other studies that showed a higher expression of IL-6 in endothelium of spastic cerebral arteries after SAH [23,24,25,26]. The role of IL-6 in endothelial damage and subsequently disruption of the blood–brain barrier has been described by Blecharz Lang et al. [15], who noted a significant overexpression of IL-6 and the IL-6 receptor (IL-6R) in the endothelium after SAH. Moreover, they showed that the in vitro inhibition of IL-6 with a polyclonal IL-6R neutralizing antibody led to an inhibition of brain endothelial cell barrier disruption through direct inference with the IL-6 signaling pathway. Although it seems unlikely that IL-6 induces DCVS directly, it is more likely that it works indirectly as a mediator by stimulating activation of CSF monocytes that in turn secrete ET-1 or as a result of its role in impairing cerebral autoregulation [7, 8, 27, 28].

The preclinical results were mirrored on clinical studies. In the early 1990s, studies described an upregulation of different inflammatory cytokines after SAH and correlated the upregulation with the occurrence of DCVS and worse clinical outcomes [7, 29,30,31]. In 1993, Mathiesen et al. [29] first described a marked increase of IL-6 in CSF (up to 300-fold higher than systemically) in a prospective cohort of SAH patients. Moreover, they noticed that mean CSF IL-6 concentrations were increased in patients with DCI. After this study, the role of inflammation and especially of IL-6 gained interest in the research field of EBI, DCVS, and DCI after SAH. During the next decades, more studies identified IL-6 as one of the most upregulated interleukins after SAH and named it a crucial player in the inflammation cascade after SAH (Fig. 4).

Fig. 4
figure 4

Number of animal and human studies analyzing the role of interleukin-6 after subarachnoid hemorrhage published per year from 1993 to 2021

Several studies described the significant association between IL-6 and DCVS [7, 16, 32,33,34,35,36,37,38]. Similar to a preclinical study published by our group [7, 8], Fassbender et al. [7, 8] demonstrated a correlation between IL-6 and ET-1 in patients with SAH.

DCVS is not the only factor leading to poor outcome after SAH. In this context, DCI plays a relevant role as well. DCI is characterized by neuronal cell death and, even here, IL-6 has been shown to plays a role. Clinical studies have shown equivalent results to those from 10 animal studies that described a statistically significant association between IL-6 concentrations (either in CSF or plasma/serum) and DCI/DIND, with additional 2 studies showing a positive trend [7, 10, 12, 35, 38,39,40,41,42]. From previous rabbit studies, we know that a single intrathecal application of IL-6 in rabbits can lead to neuronal cell death [6]. After SAH, IL-6 seems to act as an intermediate that induces the aggregation of inflammatory cells in a lesion area, increasing the release of oxygen-free radicals from neutrophils and collaborating with TNF-α to promote neuronal cell apoptosis through calcium overload in the cells [43, 44].

IL-6 in relation to microclot formation

Differences in the time course and location of DCVS and DCI have questioned the cause-and-effect relationship: that is, clinically not all patients who develop DCVS have DCI, and not all patients with DCI have DCVS [45,46,47,48,49]. Microclot formation has been considered as a possible contributing factor for DCI, because of microvessel occlusion that leads to neuronal ischemia, degeneration, and apoptosis [48, 50, 51]. After aneurysmal rupture, blood products are spilled out within the subarachnoid space. The degradation and breakdown of red blood cells results in the deposition of methemoglobin, heme, and hemin, which lead to the activation of TLR4 and the subsequent initiation of the inflammatory cascade [52, 53]. Immunomodulatory cells such as microglia are activated and, together with endothelial cells, upregulate the secretion of IL-6, which is released into both the serum and cerebrospinal fluid after SAH [29, 54]. The expression of IL-6 has been noted to be increased in the cerebral arterial wall [55]. Different studies have described prothrombogenic actions of IL-6 [56,57,58]. Moreover, IL-6 together with TNF-α and IL-1 is known to induce tissue factor upregulation, which in turn promotes a procoagulant state among endothelial cells [59]. IL-6 inhibits the cleavage of ultralarge von Willebrand factor, resulting in platelet aggregation and adhesion of the vascular endothelium, thus possibly causing thrombosis in the microvessels [60]. These endothelial and hemostatic dysfunctions with development of microclots are highly involved in the pathophysiology of EBI and the development of DCI [14, 61, 62]. Different preclinical SAH models in mice and rabbits have confirmed microclot formations after SAH [14, 61, 63]. Human autopsy studies on SAH patients also demonstrated the presence of microclots and a correlation with location and severity of ischemia [48, 50, 51]. We previously demonstrated in a rabbit SAH model how the IL-6 receptor antagonist tocilizumab significantly reduced microclot formation, neuronal cell death, and DCVS [14].

Considering the poor outcome of SAH patients as a result of DCVS and DCI, new therapeutic approaches in treating and preventing DCVS and DCI are highly warranted. As we have seen, IL-6 is marker of neuroinflammation after SAH and seems to be relevant in the pathophysiology of DCVS and DCI. Therefore, it seems reasonable to develop new therapies against neuroinflammation to prevent the aggravation of EBI and improve outcome of patients with SAH. Given that animal models have shown that use of an IL-6 antagonist led to a decrease of DCVS, DCI, and microclot formation [14, 15, 17] and considering the availability of medications such as the IL-6 receptor antagonist tocilizumab, which is widely used in the treatment of rheumatoid arthritis [64,65,66] and was recently used in clinical studies for treatment of patients affected by COVID-19 [67,68,69], new doors have opened for possible future clinical trials using an IL-6 receptor antagonist in an effort to decrease DCVS and DCI after SAH to improve patient outcomes.

Conclusions

IL-6 is a marker of intracranial inflammation and seems to play an important role in the pathophysiology of DCVS and DCI after SAH in preclinical and clinical studies. Its inhibition might be a possible therapeutic approach to improve the outcome of SAH patients.