Temporal patterns of inflammatory biomarkers measured in the cerebrospinal fluid of patients with aneurysmal subarachnoid hemorrhage using multiplex Proximity Extension Assay technology. CURRENT STATUS: POSTED

Background Neuroinflammation has been extensively studied in the context of subarachnoid hemorrhage (SAH) in recent years. A common approach is correlation of levels of single or few inflammatory biomarkers with clinical parameters such as cerebral vasospasm and outcome. However, the complexity of the inflammatory response post SAH may require a more comprehensive overall approach by analyzing multiple biomarkers simultaneously. Methods Twenty-nine patients with SAH requiring insertion of external ventricular drainage were enrolled and ventricular cerebrospinal fluid was collected at days 1, 4 and 10 after hemorrhage. The levels of 92 inflammatory biomarkers were simultaneously measured in the samples using Proseek Multiplex Inflammation 1® assay (Olink Proteomics, Uppsala, Sweden) based on Proximity Extension Assay (PEA) technology. Thirty-eight proteins were excluded from further analysis due to lack of >10% of measurable values. Temporal patterns for each of the remaining 54 proteins were illustrated on graphs with medians and quartiles. Wilcoxon matched pairs test was used for comparison of the medians between time points. Results Four different patterns were visually observed with an early peak and gradually decreasing trend (9 proteins), middle peak (14 proteins), late peak after a gradually increasing trend (27 proteins) and no specific pattern (4 proteins). Several biomarkers showed statistically significant increasing trends (defined as day 1 < day 4 < day 10 values) and late peaks (day 4 < day 10), such as chemokines CXCL6 and CCL23, or significant early peaks (that is, day 1 > day 4) such as leukemia inhibitor factor (LIF) and macrophage inflammatory factor-1β (MIP-1β). No statistically significant decreasing trends (defined as day 1 > day 4 > day 10) were observed. Two proteins showed statistically significant middle peaks (chemokine CCL28 and Delta and Notch epidermal growth factor-related receptor-DNER). Conclusion The comprehensive data set provided in this study may act as an illustration of an inflammatory profile of the acute phase of SAH showing groups of biomarkers with similar temporal patterns of activation. Further studies can be designed based on these data in order to explore potential implications on early and


Background:
Spontaneous subarachnoid hemorrhage (SAH) comprises approximately 5% of all strokes with an incidence of around 9 cases per 100.000 per year and aneurysm rupture being the cause in 85% of cases. [1] Brain injury in SAH occurs both at the time of the bleeding itself, termed primary injury, but also during the following days to weeks, a phenomenon known as secondary injury. Despite advances in the critical care of SAH patients and aneurysm treatment methods, mortality and morbidity rates among the survivals from the initial bleeding remain high, reflecting a lack of effective treatments targeting the pathophysiological mechanisms that underlie the secondary injury.
Delayed cerebral ischemia (DCI) develops in approximately 30-40% of SAH patients and is considered to be a major cause of unfavorable outcome. [2] For decades the condition was attributed to cerebral vasospasm (CV), i.e. the narrowing of basal cerebral arteries seen early on angiography and persisting for up to two weeks post SAH leading to decreased cerebral blood flow (CBF) and infarctions in the affected territories. [3] However, about 20% of SAH patients develop DCI without evidence of CV and only 30% of patients with CV actually suffer from DCI. [4] Moreover, the randomized multicenter CONCSIOUS-1 trial with the endothelin receptor antagonist Clazosentan failed to show any effect on functional outcome or incidence of cerebral ischemia despite significant decrease of CV. [5,6] Thus, an uncoupling of angiographic vasospasm and DCI became apparent and new concepts emerged as potential underlying mechanisms of delayed brain injury. [7,8] Early brain injury (EBI) refers to the events occurring within the first 72 h from ictus and includes the primary injury and its direct consequences. Elevation of intracranial pressure (ICP), global ischemia, impairment of CBF autoregulation, cortical spreading depolarization (CSD), disruption of blood-brain barrier (BBB), cell death, oxidative stress and inflammatory processes are among the mechanisms that are activated shortly after aneurysm rupture and evolve during the following days. [9] The processes triggered by the EBI are now believed to play an important role in the development of DCI. [9,10] Neuroinflammation may be a mechanistic link between these two conditions and has been studied extensively in the past years. [11][12][13] Many clinical and experimental studies report levels of single or few inflammatory biomarkers (such as cytokines) in the peripheral blood, cerebrospinal fluid (CSF) and cerebral extracellular fluid through microdialysis (MD) in SAH patients or animal models and further correlate these biomarkers with clinical parameters and outcome. [14][15][16][17][18][19][20][21][22] Although this strategy provides useful insight in the interplay between inflammation and the disease course it fails to account for the complexity of the involved mechanisms that cannot be adequately described by a single (or few) parameter(s) alone.
The aim of the present study was to provide a comprehensive inflammatory profile of the acute phase of SAH by simultaneously measuring the levels of 92 inflammatory biomarkers in the CSF at days 1, 4 and 10 after admission using multiplex Proximity Extension Assay (PEA) technology. The biomarkers were then categorized into different groups according to their temporal expression.

Methods: Ethics
The study was conducted in accordance with Declaration of Helsinki for human studies and approved by Uppsala University Ethics Committee. All participants or their next of kin gave written consent for participation in the study.

Patients
Demographic data of the 29 patients included in the study are presented in Table 1     Results: Visual inspection of the linear NPX graphs revealed four different temporal patterns (Table 3): an early peak at day 1 followed by a decreasing trend (Group A: 9 proteins), a middle peak at day 4 (Group B: 14 proteins), an increasing trend with a late peak at day 10 (Group C: 27 proteins) and finally no specific pattern (Group D: 4 proteins). Figures 1-4 illustrate typical graphs of median protein expression levels (linear NPX) and upper and lower quartiles over time for each group. Medians were then compared between time points for each protein using Wilcoxon matched pairs test. Table 4 shows the results (p-value considered significant at < 0.05). Table 3 Groups Four different groups were identified based on visual inspection of the graphs, that is an early peak followed by a decreasing trend (Group A), a middle peak (Group B), a late peak after an increasing trend (Group C) and finally no specific pattern (Group D). Table 4 Expression levels of all proteins and time points.  (4,(5)(6)(7)9) 0,31 4,9 (3,3-9,5) 0,92 IL-10RB 3,4 (2,9 − 6,2) < 0,01 7,1 (4,7-9,6) 0,26 7,9 (4,3-12) < 0,01

Discussion:
In this study we present a concise data set of the sequential production of multiple biomarkers of inflammation in the CSF of patients during the first 10 days post SAH. To our knowledge, this is the largest data set to describe the inflammatory profile of SAH ever reported with 92 inflammatory biomarkers included, many of which never studied in SAH patients previously. This is also the first report of Proximity Extension Assay (PEA) technology being used for measuring biomarker expression levels in the context of SAH. Similar studies were recently published in patients with traumatic brain injury (TBI) and trigeminal neuralgia, as well as numerous other non-neurosurgical conditions (such as neuropathic pain, cardiovascular diseases, gastric cancer, etc.). [29][30][31] The analyses in the present study were performed in CSF compartment alone (not plasma or cerebral interstitial fluid) as this seems more suitable to describe the disease pathophysiology, given also that major early and late clinical complications (i.e. CV and chronic hydrocephalus) spatially correlate best with this compartment. Many of the included proteins have previously been associated to SAH inflammation, for example IL-1ra, IL-6, IL-8, TNF-a, LIF, MCP-1, and VEGF-A. [32] On the other hand, there is scarce or non-existing literature on many other proteins, some of which showed interesting temporal patterns and statistically significant peaks and trends, such as LIF, CCL11, CCL28, 4E-BP1, CD40, CXCL6, CXCL9, and IL-18. [20,[33][34][35][36] Group A Nine biomarkers showed higher day 1 values and decreasing trends throughout the observation period (Tables 3 and 4). Four of them (CCL11, LIF, MIP-1β and TRAIL) showed statistically significant early peaks with day 1 > day 4 levels but no statistically significant decreasing trends were observed.
MIP-1β (also known as CCL4) and CCL11 are members of the chemokine family (C-C motif) and are involved in chemotaxis of macrophages and activated T-cells, respectively, as well as other proinflammatory actions. Their early peak can possibly be associated to the recruitment of leukocytes at the site of the bleeding. LIF is a cytokine involved in activation of signaling pathways that regulate cell growth among other actions. A similar early peak in serum has been observed previously. [37] TRAIL, a member of the TNF superfamily (also known as TNFSF10) involved in apoptosis, has not been studied in the SAH literature.

Group B
Among the 14 proteins included in this group only two (CCL28 and DNER) showed statistically significant middle peaks, that is significantly higher day 4 values than both day 1 and day 10.
Production of the chemokine CCL28 is induced by other proinflammatory cytokines and its chemotactic actions are exerted on B and T cells and eosinophils. Delta and Notch-like Epidermal growth factor-Receptor (DNER) is an activator of NOTCH1 pathway. None of these biomarkers have been studied in a SAH context earlier and their potential involvement in the SAH complications, mainly CV that coincides temporally with the observed middle peaks, should be examined.
It should be noted that the majority of the biomarkers in this group showed a significant increase in their levels between day 1 and day 4 (11/14, Table 4). In almost all cases though, with the exception of the two biomarkers named above, these dynamics seemed to wear off between day 4 and day 10.
Interestingly, IL-6 was recently studied by our group using a quantitative routine monoclonal antibody-based method in CSF samples from 44 patients with severe SAH. The results showed a very similar temporal IL-6 pattern as in the present study, with increasing values between day 1 and day 4 followed by decreasing values towards day 10, although remaining higher than day 1 values. [38] Group Signaling pathway molecule 4E-BP1 has been implicated in the development of vasospasm in a canine SAH-model but no reports on human studied are available. [35] Protein CD40, a member of TNF family, is found on antigen-presenting cells and mediates multiple inflammatory responses. Elevated serum levels of CD40 have been associated with poor outcome and severity of neurological deficits in SAH patients. [36,39] IL-18, a strong proinflammatory cytokine involved in the synthesis of inflammatory mediators, has recently been shown to be a predictor of early brain injury and clinical prognosis in SAH patients as elevated concentrations correlated to cerebral edema and acute hydrocephalus. [20] It should be noted that the observed temporal pattern of IL-18 in that study differed from our study as we demonstrated a late peak of this cytokine. Urokinase (or uPA) is a serum protease that activates plasminogen to plasmin which in turn leads to thrombolysis and tissue degradation. Plasma concentrations of its receptor (soluble uPA-receptor) was not shown to correlate with neurological outcome post SAH. [40] General considerations The great variability of the expression levels and temporal patterns of the measured biomarkers is an indicator of the complexity of the inflammatory response after SAH. Many of the included proteins are well established in the SAH research both in humans and in preclinical animal models while others are novel in the SAH field. Their exact role as well as interplay with each other is not easy to establish, especially considering the fact that many of these substances are described to play both a detrimental and a beneficial role in the disease course depending on the time after bleeding. [41] This study may offer guidance for further research on groups of biomarkers based either on their families or their pattern of activation (e.g. early vs late peak) to identify potential underlying inflammatory mechanisms of SAH complications as well as novel targets of intervention in order to prevent/treat these conditions and improve outcome.

Limitations
A limitation with the study is that the PEA analysis method in its presently available form does not give absolute protein concentrations. However, the protein expression levels illustrate the relative concentrations and how these change over time. They also illustrate the relations between the levels of the different biomarkers. The study may be limited by the relatively small number of patients included. Another drawback may be the fact that the study is limited to the CSF compartment. Similar analyses could be performed in the CSF, cerebral interstitial fluid and plasma, giving the opportunity to compare protein levels in the different fluid compartments providing a more thorough inflammatory profile of the acute phase of the disease. Comparisons with healthy individuals could also serve as an indicator of the intensity of activation for each biomarker. Moreover, correlations of the biomarker expression levels with clinical parameters were not performed in this study, as the main goal was to provide general information of as many proteins as possible in order to look for patterns of expression for further investigation. Finally, measurements in more than three time points could possibly describe the temporal patterns of each protein more accurately.

Conclusion:
The temporal patterns of expression of multiple inflammation-related proteins in the acute phase of SAH are reported in this study providing an inflammatory profile of the disease that facilitates further