- Email: [email protected]
Elevation of serum CXC chemokine ligand-12 levels predicts poor outcome after aneurysmal subarachnoid hemorrhage
Journal of the Neurological Sciences 362 (2016) 53–58
Contents lists available at ScienceDirect
Journal of the Neurological Sciences journal homepage: www.elsevier.com/locate/jns
Elevation of serum CXC chemokine ligand-12 levels predicts poor outcome after aneurysmal subarachnoid hemorrhage☆ De-Sheng Pan ⁎, Min Yan, Muhammad Hassan, Ze-Bin Fang, Man-Tao Chen Department of Neurosurgery, The First Affiliated Hospital, School of Medicine, Zhejiang University, 79 Qingchun Road, Hangzhou 310003, China
a r t i c l e
i n f o
Article history: Received 14 December 2015 Received in revised form 10 January 2016 Accepted 14 January 2016 Available online 16 January 2016 Keywords: Aneurysm Biomarker CXC chemokine ligand-12 Function outcome Prognosis Subarachnoid hemorrhage Mortality
a b s t r a c t Objective: CXC chemokine ligand-12 (CXCL12) is involved in the innate immune system. Elevation of its level in the peripheral blood is associated with severity and outcome of ischemic stroke. This study aimed to investigate its relation to severity and prognosis following aneurysmal subarachnoid hemorrhage (aSAH). Methods: Serum CXCL12 levels were determined in a total of 182 controls and 182 aSAH patients. Hemorrhagic severity was assessed using the World Federation of Neurological Surgeons (WFNS) scale and modified Fisher grading scale. Unfavorable outcome was defined as Glasgow outcome scale score of 1–3. Prognostic predictors of 6-month mortality and unfavorable outcome were identified using multivariate analysis. Results: The serum CXCL12 levels were significantly higher in patients as compared to controls (14.5 ± 6.7 ng/mL vs. 1.7 ± 0.6 ng/mL, P b 0.001) and were independently associated with WFNS scores (t = 5.927, P b 0.001) and modified Fisher scores (t = 5.506, P b 0.001). Serum CXCL12 levels predicted 6-month mortality and 6-month unfavorable outcome with the area under curves of 0.815 [95% confidence (CI), 0.751–0.868] and 0.809 (95% CI, 0.745–0.864) respectively and were related independently to 6-month mortality (odds ratio, 4.428; 95% CI, 1.977–12.031; P = 0.004) and 6-month unfavorable outcome (odds ratio, 3.821; 95% CI, 1.097–9.251; P = 0.001). Moreover, the predictive values of CXCL12 levels were in the range of WFNS scores and modified Fisher scores. Conclusions: Elevation of serum CXCL12 levels is associated highly with hemorrhagic severity and poor outcome after aSAH, suggesting CXCL12 might have the potential to be a prognostic predictive biomarker of aSAH. © 2016 Elsevier B.V. All rights reserved.
1. Introduction Aneurysmal subarachnoid hemorrhage (aSAH) is a major clinical and health-care problem affecting millions of people worldwide [1–3]. It is important to conduct a precise evaluation of the patient's clinical status in the management of treatment modalities and determination of prognosis. The World Federation of Neurological Surgeons (WFNS) scale and Fisher grading scale are commonly used to evaluate the neurologic status and severity of hemorrhage based on the computed tomography (CT) appearance in aSAH patients [4,5]. In addition, interest in the measurement of blood biomarkers as prognostic predictors of aSAH has increased in the last few years [6–8].
Abbreviations: aSAH, aneurysmal subarachnoid hemorrhage; AUC, area under curve; CI, confidence interval; CT, computerized tomography; CXCL12, CXC chemokine ligand12; CXCR 4, CXC receptor 4; GOS, Glasgow outcome scale; OR, odds ratio; ROC, receiver operating characteristic; SDF-1, stromal cell-derived factor 1; WFNS, World Federation of Neurological Surgeons. ☆ Institution at which the work was performed: The First Affiliated Hospital, School of Medicine, Zhejiang University. ⁎ Corresponding author at: Department of Neurosurgery, The First Affiliated Hospital, School of Medicine, Zhejiang University, 79 Qingchun Road, Hangzhou 310003, China. E-mail address: [email protected] (D.-S. Pan).
http://dx.doi.org/10.1016/j.jns.2016.01.024 0022-510X/© 2016 Elsevier B.V. All rights reserved.
The immune system is a complex and tightly regulated system of effectors against external pathogens and endogenous tissue damage. It can be broadly divided into two component parts, the innate immune system and the adaptive immune system [9,10]. The growing body of evidence has shown that the innate immune system affects brain damage after subarachnoid hemorrhage. Chemokines are small chemoattractant cytokines that are produced in response to damage to call inflammatory cells into the area, thus, regulating the innate immune responses after aSAH [10–12]. The CXC chemokine ligand-12 (CXCL12) is a member of the CXC chemokine subfamily that is constitutively expressed in the bone marrow and other tissues including the brain endothelium and is responsible for regulating the trafficking and localization of bone marrow progenitor cells under steady state and stress conditions [13–15]. Following ischemic stroke, CXCL12 mediates the inflammatory response by recruitment of neural progenitor cells, and the mobilization of bone marrow-derived progenitor cells for tissue regeneration and neovascularization [16]. Other animal experiments had also found that CXCL12 played a significantly beneficial role in acute stroke [17,18]. CXCL12 combing CXC motif receptor 4 could inhibit the caspase-3 pathway by up-regulating Bcl-2/Bax ratio, which protects neurons from apoptosis in rats with traumatic brain injury [19]. Interestingly, elevated circulating CXCL12 level at admission is strongly
54
D.-S. Pan et al. / Journal of the Neurological Sciences 362 (2016) 53–58
associated with the future stroke [20] and future recurrence of ischemic stroke [21], as well as has close relation to stroke severity and lesion volumes [22]. Still, CXCL12 is identified as an independent diagnostic and prognostic marker in patients with acute ischemic stroke [23]. The present study further investigated the potential prognostic role of CXCL12 in Chinese aSAH patients.
2. Methods
2.5. Sampling and laboratory analysis Blood samples were collected from the patients at admission and from controls at study entry. After centrifugation, aliquots of the samples were immediately stored −80 °C until assay. Serum CXCL12 levels were determined in duplicate samples with a quantitative sandwich enzyme-linked immunosorbent assay kit (Quantikine; R&D Systems, Minneapolis, MN, USA) in accordance with the manufactures' instructions. The investigator was blinded to the clinical outcome and neuroimaging findings.
2.1. Ethics statement 2.6. Statistical analysis This work has been carried out in accordance with The Code of Ethics of the World Medical Association (Declaration of Helsinki) and has been also approved by the ethical committee at the First Affiliated Hospital, School of Medicine, Zhejiang University. Still, all participants or their representative were informed of the study protocol and their written informed consent was obtained.
2.2. Participants This is a prospective study from a cohort of patients with confirmed aSAH admitted to the Department of Neurosurgery, First Affiliated Hospital, School of Medicine, Zhejiang University between January 2011 and May 2014. Patients were enrolled in the study with the first-ever non-traumatic SAH, the clinical history of SAH within the last 24 h before admission, the single intracranial aneurysms confirmed by CT angiography with or without digital subtraction angiography and the treatment through clipping or coiling within the 48 h after admission. Exclusion criteria included surgery, trauma or infectious diseases during the preceding 1 month, autoimmune diseases with or without immunosuppressive therapy, prior neurological diseases such as intracerebral hemorrhage and ischemic stroke, previous use of antiplatelet or anticoagulant medication, rebleeding after admission, suspected pseudoaneurysm, and other prior systemic diseases such as uremia, liver cirrhosis, malignancy, chronic heart disease, chronic lung disease, diabetes mellitus and hypertension. The control group consisted of the healthy age- and sex-matched volunteers.
All the results were statistically analyzed using SPSS 19.0 (SPSS Inc., Chicago, IL, USA) and MedCalc 9.6.4.0 (MedCalc Software, Mariakerke, Belgium). Statistical significance was defined as a probability value of less than 0.05. Categorical variables were presented as counts (percentage). Continuous variable were reported as mean ± standard deviation or median (the upper and lower quartiles) as appropriate. The statistical significance of the intergroup observed difference was assessed with the chi-square test, Fisher exact test, Student t test or Mann–Whitney U-test as appropriate. All parameters found to be significant in the univariate analysis were further entered in a binary Logistic regression model to identify prognostic predictors. Linear relationships between CXCL12 levels and other variables were investigated using Spearman's correlation coefficient or Pearson's correlation coefficient as appropriate. We used Bonferroni correction to control for the multiple testing. A multivariate linear regression was carried out further to verify their relationships. Receiver operating curves (ROCs) were generated to determine cutoff values for optimal prognostic predictive sensitivities and specificities. The area under curves (AUCs) were estimated in accordance with the ROC curves and the additive benefits of CXCL12 levels to the WFNS scores and modified Fisher scores were evaluated in a combined logistic-regression model. 3. Results 3.1. Participants characteristics
2.3. Assessment The neurological status was graded according to the WFNS scale [4] and the radiological severity of SAH was classified based on the modified Fisher grade [5]. The WFNS assessments were performed by at least two clinicians. Radiologists helped to determine the modified Fisher grade. When two raters obtained different scores, a judgment of a third senior clinician was considered. Symptomatic cerebral vasospasm was defined as the development of new focal neurological signs, deterioration in level of consciousness, or the appearance of new infarction on CT when the cause was felt to be ischemia attributable to vasospasm after other possible causes of worsening (e.g. hydrocephalus, seizures, metabolic derangement, infection, or oversedation) had been excluded [24,25].
2.4. Endpoint Participants were followed up until death or completion of 6 months after aSAH. The endpoints were unfavorable outcome and death after 6 months. The functional outcome was defined by Glasgow outcome scale (GOS) score. GOS scores were dichotomized in favorable and unfavorable outcomes (GOS of 4–5 vs. GOS of 1–3) [26–28]. For follow-up, structured telephone interviews were performed by one doctor who was blinded to clinical information.
During the study period, 239 aSAH patients were initially evaluated. 57 patients were excluded because of the reasons listed in Fig. 1. Eventually, 182 aSAH patients were included in this study. Additionally, 182 gender- and age-matched healthy subjects were enrolled as controls. This group of aSAH patients was composed of 77 males and 105 females as well as had a mean age of 50.6 ± 11.4 years (range, 34– 77 years). The median admission WFNS scores and modified Fisher scores were 3 (2–3) (range, 1–5) and 3 (2–4) (range, 2–5) respectively. Aneurysmal location is as follows: 48 (26.4%) aneurysms were located at posterior communication artery; 31 (17.0%) aneurysms, internal carotid artery; 37 (20.3%) aneurysms, anterior communication artery; 33 (18.1%) aneurysms, middle cerebral artery; 22 (12.1%) aneurysms, anterior cerebral artery; 8 (4.4%) aneurysms, posterior cerebral artery; 3 (1.7%) aneurysms, vertebral artery. 146 (80.2%) patients had cystic aneurysm. The mean aneurysm diameter was 7.9 ± 5.4 mm (range, 2– 25 mm). 96 (52.8%) patients underwent clipping and other patients obtained endovascular coiling. 37 (20.3%) patients had acute hydrocephalus; 29 (15.9%) patients, intraventricular hemorrhage; 43 (23.6%) patients, underwent external ventricular drain; 52 (28.6%) patients, symptomatic cerebral vasospasm. 3.2. The change of serum CXCL12 levels in aSAH patients Patients had the significant elevation of serum CXCL12 levels as compared with normal controls (14.5 ± 6.7 ng/mL vs. 1.7 ±
D.-S. Pan et al. / Journal of the Neurological Sciences 362 (2016) 53–58
55
Fig. 1. Flow chart illustrating excluded and included patients with aneurysmal subarachnoid hemorrhage.
0.6 ng/mL, P b 0.001). Additionally, the serum CXCL12 levels were obviously higher in non-survivors than in survivors (21.1 ± 5.4 ng/mL vs. 13.2 ± 6.1 ng/mL, P b 0.001) and in patients with unfavorable outcome than with favorable outcome (19.4 ± 4.9 ng/mL vs.12.0 ± 6.0 ng/mL, P b 0.001).
CXCL12 numerically improved the AUCs of WFNS score and modified Fisher score to 0.904 (95% CI 0.852–0.943, P = 0.208) and 0.894 (95% CI 0.840–0.935, P = 0.149) respectively. In Table 2, a univariate analysis found that serum CXCL12 levels higher than 18.1 ng/mL and other variables were significantly
3.3. Correlation analysis
Table 1 The relationship between serum CXC chemokine ligand-12 levels and other parameters following aneurysmal subarachnoid hemorrhage.
In Table 1, some variables including the WFNS scores, modified Fisher scores, acute hydrocephalus, intraventricular hemorrhage, symptomatic cerebral vasospasm, external ventricular drain, blood glucose levels and plasma C-reactive protein levels were associated highly with serum CXCL12 levels. We further used Bonferroni correction to control for the multiple testing. Because all P values, which exhibited statistically significant, were less than 0.002 (0.05/21), their linear relationships were not changed. In Fig. 2, serum CXCL12 levels still remained closely related to WFNS scores (t = 5.927, P b 0.001) and modified Fisher scores (t = 5.506, P b 0.001) in a multivariate linear regression model. 3.4. Mortality prediction 30 patients (16.5%) died within 6 month after aSAH. Based on the ROC curve, the optimal cutoff value of serum CXCL12 levels as an indicator for prognosis of the mortality was projected to be 18.1 ng/mL, which yielded a sensitivity of 80.0% and a specificity of 72.4%, with the AUC at 0.815 [95% confidence interval (CI) = 0.751–0.868] in Fig. 3A. Its AUC was in the range of WFNS score (AUC 0.878, 95% CI 0.821–0.921, P = 0.248) and modified Fisher score (AUC 0.861, 95% CI 0.802–0.908, P = 0.358).When a combined logistic-regression model was configured,
Characteristics
r value
P value
Gender Age (year) WFNS scores Modified Fisher scores Aneurysmal locations Cystic aneurysm Aneurysm diameter (mm) Treatment (clipping/endovascular coiling) Acute hydrocephalus Intraventricular hemorrhage External ventricular drain Symptomatic cerebral vasospasm Admission time (h) Plasma-sampling time (h) Systolic arterial pressure (mm Hg) Diastolic arterial pressure (mm Hg) Blood glucose level (mmol/L) Plasma C-reactive protein level (mg/L) Blood white blood cell count (×109/L) Blood hemoglobin level (g/L) Blood platelet count (×109/L)
0.052 0.127 0.539 0.525 0.015 0.028 0.084 0.056 0.279 0.235 0.309 0.423 0.044 0.048 0.003 0.136 0.260 0.305 0.035 0.128 0.117
0.488 0.087 b0.001 b0.001 0.842 0.703 0.260 0.449 b0.001 0.001 b0.001 b0.001 0.554 0.523 0.970 0.067 b0.001 b0.001 0.640 0.084 0.116
Bivariate correlations were assessed by Spearman's or Pearson's correlation coefficient. WFNS indicates World Federation of Neurological Surgeons.
56
D.-S. Pan et al. / Journal of the Neurological Sciences 362 (2016) 53–58
Fig. 3. Receiver operating characteristic curve analysis of serum chemokine ligand-12 (CXCL12) levels for 6-month mortality (A) and 6-month unfavorable outcome (B) after aneurysmal subarachnoid hemorrhage. 95% CI indicates confidence interval. Fig. 2. Relationships between serum chemokine ligand-12 (CXCL12) levels and the World Federation of Neurological Surgeons (WFNS) scores as well as between serum CXCL12 levels and modified Fisher scores.
associated with the mortality. A regression analysis showed that WFNS score [Odds ratio (OR), 6.054; 95% CI, 2.314–15.837; P b 0.001], modified Fisher scores (OR, 4.110; 95% CI, 1.660–10.175; P = 0.001) and serum CXCL12 levels higher than 18.1 ng/mL (OR, 4.428; 95% CI, 1.977–12.031; P = 0.004) were the independent predictors of mortality.
3.5. Poor outcome prediction In our study, an unfavorable functional outcome was found in 62 patients (34.1%) within 6 months after aSAH. The ROC curve analysis showed that the optimal cutoff value of serum CXCL12 levels to predict unfavorable outcome was 18.0 ng/mL, with the AUC of 0.809 (95% CI = 0.745–0.864), which had specificity and sensitivity values of 71.0% and 80.0% respectively in Fig. 3B. Its AUC was not statistically significantly different from those of WFNS score (AUC 0.869, 95% CI 0.811–0.914, P = 0.158) and modified Fisher score (AUC 0.841, 95% CI 0.779–0.890, P = 0.462). The combined model found that CXCL12 enhanced the AUCs of WFNS score and modified Fisher score to 0.889 (95% CI 0.834–0.931) and 0.867 (95% CI 0.808–0.912) respectively, but the differences were not statistically significant (P = 0.259 and 0.343). In Table 3, serum CXCL12 levels higher than 18.0 ng/mL and other parameters had close relation to the unfavorable outcome using a univariate analysis. WFNS score (OR, 5.890; 95% CI, 2.710–12.803; P b 0.001), modified Fisher scores (OR, 4.844; 95% CI, 2.257–10.397; P b 0.001) and serum CXCL12 levels higher than 18.0 ng/mL (OR, 3.821; 95% CI, 1.097–9.251; P = 0.001) were identified as the
independent predictors of the unfavorable outcome in a binary Logistic regression model. 4. Discussion Although previous papers had found the prognostic and diagnostic predictive significance of serum CXCL12 in ischemic stroke [20–23], the role of CXCL12 has not been explored in aSAH. This study configured the multiple multivariate models and showed some interesting results. Its main findings were that, (1) serum CXCL12 levels of aSAH patients were obviously elevated compared with controls; (2) serum CXCL12 levels had an independent relation to WFNS scores and modified Fisher scores; (3) serum CXCL12 levels could independently predict 6-month mortality and 6-month functional outcome; (4). This biomarker's AUC was in the range of WFNS scores and modified Fisher scores. These data suggests that CXCL12 might be associated with the severity and clinical outcomes of aSAH. CXCL12, also known as stromal cell-derived factor 1 (SDF-1), is subdivided into 6 isoforms: SDF-1a, SDF-1b, SDF-1g, SDF-1d, SDF-1e, and SDF-14. Among these, SDF-1a is the most abundant and plays a variety of roles in different tissues [29]. In ischemic stroke, SDF-1a, binding to the CXC receptor 4 (CXCR 4), is involved in the focal angiogenesis in animal ischemic tissue [30]. In head trauma rats, SDF-1a, combing CXCR4 receptors, could inhibit the caspase-3 pathway by upregulating Bcl-2/Bax ratio, which protects neurons from apoptosis [19]. In vitro experiment has also demonstrated that CXCL12 could promote neuron survival by protecting them from apoptosis [31]. A recent study showed that SDF-1α could help to accelerate clot organization
D.-S. Pan et al. / Journal of the Neurological Sciences 362 (2016) 53–58 Table 2 The comparisons of clinical characteristics and laboratory data between 6-month non-surviving and surviving patients with aneurysmal subarachnoid hemorrhage.
Gender (male/female) Age (year) WFNS scores Modified Fisher scores Aneurysmal location Posterior communication artery Internal carotid artery Anterior communication artery Middle cerebral artery Anterior cerebral artery Posterior cerebral artery Vertebral artery Cystic aneurysm Aneurysm diameter (mm) Treatment (clipping/endovascular coiling) Acute hydrocephalus Intraventricular hemorrhage External ventricular drain Symptomatic cerebral vasospasm Admission time (h) Plasma-sampling time (h) Systolic arterial pressure (mm Hg) Diastolic arterial pressure (mm Hg) Blood glucose level (mmol/L) Plasma C-reactive protein level (mg/L) Blood white blood cell count (×109/L) Blood hemoglobin level (g/L) Blood platelet count (×109/L) Serum CXCL12 levels higher than 18.1 ng/mL
Non-survivors n = 30
Survivors n = 152
P value
15/15 55.1 ± 11.9 4 (3–4) 4 (3–4)
62/90 49.7 ± 11.1 2 (2–3) 3 (2–3)
0.351 0.017 b0.001 b0.001 0.260
8 (26.7%) 5 (16.6%) 2 (6.7%) 9 (30.0%) 3 (10.0%) 2 (6.7%) 1 (3.3%) 25 (83.3%) 7.5 ± 4.6 18/12
40 (26.3%) 26 (17.2%) 35 (23.0%) 24 (15.8%) 19 (12.5%) 6 (3.9%) 2 (1.3%) 121 (79.6%) 8.0 ± 5.5 78/74
15 (50.0%) 11 (36.7%) 18(60.0%) 16 (53.3%) 10.6 ± 3.7 12.6 ± 4.6 129.6 ± 21.6 78.0 ± 13.7 12.7 ± 3.8 17.7 ± 4.9 7.6 ± 3.0 119.4 ± 21.7 190.2 ± 36.0 24 (80.0%)
22 (14.5%) 18 (11.8%) 25 (16.5%) 36 (23.7%) 9.7 ± 3.7 14.0 ± 5.1 131.8 ± 23.5 78.1 ± 11.6 10.5 ± 4.1 15.3 ± 3.9 8.4 ± 3.0 118.4 ± 23.8 181.9 ± 47.0 42 (27.6%)
0.639 0.656 0.384 b0.001 0.002 b0.001 0.001 0.234 0.170 0.636 0.969 0.008 0.004 0.223 0.825 0.365 b0.001
The variables are presented as percentages, mean ± standard deviation or median (the upper and lower quartiles). The comparisons were assessed using the chi-square test, Fisher exact test, Student t test or Mann–Whitney U test as appropriate. WFNS indicates World Federation of Neurological Surgeon; CXCL12, CXC chemokine ligand-12.
and endothelial cell proliferation after implantation of coils for embolization of aneurysm [32]. Thus, CXCL12 might be involved in the pathogenesis of many neurodegenerative diseases including aSAH and could exert a protective effect on acute brain injury. Chemokines are expressed in neuron, astrocytes and microglia in central nervous system [33]. SDF-1 expression is associated with reactive astrocytes [34]. It has been reported that microglia are activated and chemoattracted by SDF-1 [35]. Furthermore, microglia activation leads to tumor necrosis factor-alpha production in astrocytes [36]. Still, SDF-1α up-regulates interleukin-6 through CXCR4, phosphatidylinositol-3-kinase/serine/threonine kinase, extracellular signal-regulated kinase, and nuclear factor-kappaB-dependent pathway in microglia [37]. These results provide the support for the notion that CXCL12 plays a regulatory role in neuroinflammation. Given the fact that neuroinflammation contributes to the pathogenesis of diffuse brain injury after aSAH [38], the potential role of CXCL12 is established for the regulation of diffuse brain injury after aSAH. However, a mechanism linking CXCL12 and neuroinflammation after aSAH remains to be studied. The accumulating evidence has demonstrated that CXCL12 levels have significant prognostic predictive values for some neurologic diseases [22,23,39,40]. Elevated serum CXCL12 levels have been found after acute ischemic stroke. Moreover, serum CXCL12 was positively correlated with infarct volume, stroke severity, mortality and unfavorable functional outcome [22,23]. In our study, elevation of serum CXCL12 levels was found; CXCL12 was verified to be associated with aSAH severity reflected by WFNS score and modified Fisher score; CXCL12 was identified as an independent predictor for poor outcomes
57
Table 3 The factors correlated with 6-month functional outcome after aneurysmal subarachnoid hemorrhage.
Gender (male/female) Age (year) WFNS scores Modified Fisher scores Aneurysmal location Posterior communication artery Internal carotid artery Anterior communication artery Middle cerebral artery Anterior cerebral artery Posterior cerebral artery Vertebral artery Cystic aneurysm Aneurysm diameter (mm) Treatment (clipping/endovascular coiling) Acute hydrocephalus Intraventricular hemorrhage External ventricular drain Symptomatic cerebral vasospasm Admission time (h) Plasma-sampling time (h) Systolic arterial pressure (mm Hg) Diastolic arterial pressure (mm Hg) Blood glucose level (mmol/L) Plasma C-reactive protein level (mg/L) Blood white blood cell count (×109/L) Blood hemoglobin level (g/L) Blood platelet count (×109/L) Serum CXCL12 levels higher than 18.1 ng/mL
Unfavorable outcome n = 62
Favorable outcome n = 120
P value
28/34 53.3 ± 12.3 3 (3–4) 4 (3–4)
49/71 49.2 ± 10.7 2 (1–3) 3 (2–3)
0.575 0.019 b0.001 b0.001 0.076
20 (32.3%) 11 (17.7%) 6 (9.7%) 10 (16.1%) 8 (12.9%) 5 (8.1%) 2 (3.2%) 49 (79.0%) 7.5 ± 4.8 36/26
28 (23.3%) 20 (16.7%) 31 (25.8%) 23 (19.2%) 14 (11.7%) 3 (2.5%) 1 (0.8%) 97 (80.8%) 8.1 ± 5.6 60/60
24 (38.7%) 20 (32.3%) 29 (46.8%) 31 (50.0%) 9.3 ± 3.1 13.9 ± 5.0 135.5 ± 22.5 76.8 ± 11.9 12.0 ± 4.4 17.0 ± 4.4
13 (10.8%) 9 (7.5%) 14 (11.7%) 21 (17.5%) 10.1 ± 4.0 13.8 ± 5.1 129.4 ± 23.2 78.7 ± 12.0 10.2 ± 3.9 15.0 ± 4.0
8.6 ± 3.7
8.1 ± 2.6
118.7 ± 23.7 184.0 ± 45.7 44 (71.0%)
118.5 ± 23.4 182.9 ± 45.5 24 (20.0%)
0.773 0.520 0.302 b0.001 b0.001 b0.001 0.001 0.201 0.886 0.096 0.312 0.008 0.003 0.245 0.954 0.870 b0.001
The variables are presented as percentages, mean ± standard deviation or median (the upper and lower quartiles). Comparisons were assessed using the chi-square test, Fisher exact test, Student t test or Mann–Whitney U test as appropriate. WFNS indicates World Federation of Neurological Surgeon; CXCL12, CXC chemokine ligand-12.
including mortality and unfavorable outcome; interestingly, its predictive value was in the range of WFNS score and modified Fisher score. Overall, CXCL12 in serum was correlated with hemorrhagic severity and clinical outcomes after aSAH. Our study has two limitations. Firstly, blood samples were collected only once after aSAH. Serial follow-up studies of CXCL12 are needed because its levels can change over time, hence additional measurements in the days thereafter would have been of interest. Secondly, we measured CXCL12 in serum, not in cerebral spinal fluid. It is still uncertain whether peripheral CXCL12 levels reflect similar changes in the central nervous system. This aspect of study warrants to be explored in future. 5. Conclusions The results of this study demonstrated that elevated serum CXCL12 levels had close relation to WFNS scores and modified Fisher scores, in addition to 6-month mortality and 6-month unfavorable outcome, suggesting CXCL12 might have the potential to be a prognostic predictive biomarker after aSAH. Competing interests The authors declare that they have no competing interests. Conflict of interest The authors report no conflicts of interest concerning the materials or methods used in this study or the findings specified in this article.
58
D.-S. Pan et al. / Journal of the Neurological Sciences 362 (2016) 53–58
Acknowledgment The authors thank all staffs in the Department of Neurosurgery, The First Affiliated Hospital, School of Medicine, Zhejiang University for their technical support. References [1] Y. Wang, Y. Tian, D. Wang, H. Wei, Z. Zhao, R. Jiang, et al., High angiopoietin-1 levels predict a good functional outcome within 72 h of an aneurysmal subarachnoid hemorrhage: a prospective study from a single center, J. Neurol. Sci. 356 (1–2) (2015) 72–76. [2] C. Lindgren, S. Söderberg, L.O. Koskinen, M. Hultin, L. Edvardsson, S. Naredi, Longterm subarachnoid haemorrhage survivors still die due to cerebrovascular causes, Acta Neurol. Scand. 132 (2015) 410–416. [3] A. Rodríguez-Rodríguez, J.J. Egea-Guerrero, Z. Ruiz de Azúa-López, F. MurilloCabezas, Biomarkers of vasospasm development and outcome in aneurysmal subarachnoid hemorrhage, J. Neurol. Sci. 341 (1–2) (2014) 119–127. [4] C. Drake, Report of World Federation of Neurological Surgeons Committee on a universal subarachnoid hemorrhage grading scale, J. Neurosurg. 68 (6) (1988) 985–986. [5] C.M. Fisher, J.P. Kistler, J.M. Davis, Relation of cerebral vasospasm to subarachnoid hemorrhage visualized by computerized tomographic scanning, Neurosurgery 6 (1) (1980) 1–9. [6] Y.F. Shen, W.H. Wang, W.H. Yu, X.Q. Dong, Q. Du, D.B. Yang, et al., The prognostic value of plasma thrombospondin-1 concentrations after aneurysmal subarachnoid hemorrhage, Clin. Chim. Acta 448 (2015) 155–160. [7] X.D. Chen, J. Sun, C. Lu, H.J. Ba, M.H. Chen, J.H. Lin, et al., The prognostic value of plasma soluble CD40 ligand levels following aneurysmal subarachnoid hemorrhage, Thromb. Res. 136 (1) (2015) 24–29. [8] J.Y. Cai, X.D. Chen, H.J. Ba, J.H. Lin, C. Lu, M.H. Chen, et al., Identification of plasma adrenomedullin as a possible prognostic biomarker for aneurysmal subarachnoid hemorrhage, Peptides 59 (2014) 9–13. [9] S.T. Shibutani, T. Saitoh, H. Nowag, C. Münz, T. Yoshimori, Autophagy and autophagy-related proteins in the immune system, Nat. Immunol. 16 (10) (2015) 1014–1024. [10] S.P. Gadani, J.T. Walsh, J.R. Lukens, J. Kipnis, Dealing with danger in the CNS: the response of the immune system to injury, Neuron 87 (1) (2015) 47–62. [11] C.Z. Chang, S.C. Wu, C.L. Lin, A.L. Kwan, Valproic acid attenuates intercellular adhesion molecule-1 and E-selectin through a chemokine ligand 5 dependent mechanism and subarachnoid hemorrhage induced vasospasm in a rat model, J. Inflamm. (Lond.) 12 (2015) 27. [12] G.H. Kim, C.P. Kellner, D.K. Hahn, B.M. Desantis, M. Musabbir, R.M. Starke, et al., Monocyte chemoattractant protein-1 predicts outcome and vasospasm following aneurysmal subarachnoid hemorrhage, J. Neurosurg. 109 (1) (2008) 38–43. [13] Q. Guo, B.L. Gao, X.J. Zhang, G.C. Liu, F. Xu, Q.Y. Fan, et al., CXCL12-CXCR4 axis promotes proliferation, migration, invasion, and metastasis of ovarian cancer, Oncol. Res. 22 (5) (2015) 247–258. [14] T. Nagasawa, CXCL12/SDF-1 and CXCR4, Front. Immunol. 6 (2015) 301. [15] F. Arenzana-Seisdedos, SDF-1/CXCL12: a chemokine in the life cycle of HIV, Front. Immunol. 6 (2015) 256. [16] M. Li, R.M. Ransohoff, Multiple roles of chemokine CXCL12 in the central nervous system: a migration from immunology to neurobiology, Prog. Neurobiol. 84 (2) (2008) 116–131. [17] A.M. Robin, Z.G. Zhang, L. Wang, R.L. Zhang, M. Katakowski, L. Zhang, et al., Stromal cell-derived factor 1alpha mediates neural progenitor cell motility after focal cerebral ischemia, J. Cereb. Blood Flow Metab. 26 (1) (2006) 125–134. [18] X. Yu, D. Chen, Y. Zhang, X. Wu, Z. Huang, H. Zhou, et al., Overexpression of CXCR4 in mesenchymal stem cells promotes migration, neuroprotection and angiogenesis in a rat model of stroke, J. Neurol. Sci. 316 (1–2) (2012) 141–149.
[19] W. Mao, X. Yi, J. Qin, M. Tian, G. Jin, CXCL12 inhibits cortical neuron apoptosis by increasing the ratio of Bcl-2/Bax after traumatic brain injury, Int. J. Neurosci. 124 (4) (2014) 281–290. [20] R.C. Schutt, M.D. Burdick, R.M. Strieter, B. Mehrad, E.C. Keeley, Plasma CXCL12 levels as a predictor of future stroke, Stroke 43 (12) (2012) 3382–3386. [21] X.L. Gu, L. Liu, X.D. Lu, Z.R. Liu, Serum CXCL12 levels as a novel predictor of future stroke recurrence in patients with acute ischemic stroke, Mol. Neurobiol. (2015)http://dx.doi.org/10.1007/s12035-015-9151-0. [22] P. Liu, J.W. Xiang, S.X. Jin, Serum CXCL12 levels are associated with stroke severity and lesion volumes in stroke patients, Neurol. Res. 37 (10) (2015) 853–858. [23] X.X. Duan, G.P. Zhang, X.B. Wang, H. Yu, J.L. Wu, K.Z. Liu, et al., The diagnostic and prognostic value of serum CXCL12 levels in patients with ischemic stroke, Neurol. Sci. 36 (12) (2015) 2227–2234. [24] J. Claassen, G.L. Bernardini, K. Kreiter, J. Bates, Y.E. Du, D. Copeland, et al., Effect of cisternal and ventricular blood on risk of delayed cerebral ischemia after subarachnoid hemorrhage: the Fisher scale revisited, Stroke 32 (9) (2001) 2012–2020. [25] J.A. Frontera, A. Fernandez, J.M. Schmidt, J. Claassen, K.E. Wartenberg, N. Badjatia, et al., Defining vasospasm after subarachnoid hemorrhage: what is the most clinically relevant definition? Stroke 40 (6) (2009) 1963–1968. [26] X.D. Zhu, J.S. Chen, F. Zhou, Q.C. Liu, G. Chen, J.M. Zhang, Relationship between plasma high mobility group box-1 protein levels and clinical outcomes of aneurysmal subarachnoid hemorrhage, J. Neuroinflammation 9 (2012) 194. [27] X.D. Zhu, J.S. Chen, F. Zhou, Q.C. Liu, G. Chen, J.M. Zhang, Detection of copeptin in peripheral blood of patients with aneurysmal subarachnoid hemorrhage, Crit. Care 15 (6) (2011) R288. [28] J.W. Pan, L.N. He, F. Xiao, J. Shen, R.Y. Zhan, Plasma gelsolin levels and outcomes after aneurysmal subarachnoid hemorrhage, Crit. Care 17 (4) (2013) R149. [29] B. Nash, O. Meucci, Functions of the chemokine receptor CXCR4 in the central nervous system and its regulation by μ-opioid receptors, Int. Rev. Neurobiol. 118 (2014) 105–128. [30] Y. Wang, J. Huang, Y. Li, G.Y. Yang, Roles of chemokine CXCL12 and its receptors in ischemic stroke, Curr. Drug Targets 13 (2) (2012) 166–172. [31] M.Z. Khan, R. Brandimarti, S. Shimizu, J. Nicolai, E. Crowe, O. Meucci, The chemokine CXCL12 promotes survival of postmitotic neurons by regulating Rb protein, Cell Death Differ. 15 (10) (2008) 1663–1672. [32] Y. Gao, Q. Wang, X. Cui, Y. Liu, T. Zheng, C. Chen, et al., Controlled release of stromal cell-derived factor-1α from silk fibroin-coated coils accelerates intra-aneurysmal organization and occlusion of neck remnant by recruiting endothelial progenitor cells, Int. J. Clin. Exp. Pathol. 7 (12) (2014) 8366–8380. [33] A. Bajetto, R. Bonavia, S. Barbero, T. Florio, A. Costa, G. Schettini, Expression of chemokine receptors in the rat brain, Ann. N. Y. Acad. Sci. 876 (1999) 201–209. [34] W.D. Hill, D.C. Hess, A. Martin-Studdard, J.J. Carothers, J. Zheng, D. Hale, et al., SDF-1 (CXCL12) is upregulated in the ischemic penumbra following stroke: association with bone marrow cell homing to injury, J. Neuropathol. Exp. Neurol. 63 (1) (2004) 84–96. [35] S. Tanabe, M. Heesen, I. Yoshizawa, M.A. Berman, Y. Luo, C.C. Bleul, et al., Functional expression of the CXC-chemokine receptor-4/fusin on mouse microglial cells and astrocytes, J. Immunol. 159 (2) (1997) 905–911. [36] P. Bezzi, M. Domercq, L. Brambilla, R. Galli, D. Schols, E. De Clercq, et al., CXCR4activated astrocyte glutamate release via TNFalpha: amplification by microglia triggers neurotoxicity, Nat. Neurosci. 4 (7) (2001) 702–710. [37] D.Y. Lu, C.H. Tang, W.L. Yeh, K.L. Wong, C.P. Lin, Y.H. Chen, et al., SDF-1alpha upregulates interleukin-6 through CXCR4, PI3K/Akt, ERK, and NF-kappaB-dependent pathway in microglia, Eur. J. Pharmacol. 613 (1–3) (2009) 146–154. [38] K.A. Hanafy, The role of microglia and the TLR4 pathway in neuronal apoptosis and vasospasm after subarachnoid hemorrhage, J. Neuroinflammation 10 (2013) 83. [39] C. Tingjun, L. Zhaohui, J. Zhaocai, L. Zihao, X. Quangang, H. Dehui, et al., Changes of CXCL12, CXCL14 and PDGF levels in the brain of patients with idiopathic demyelinating optic neuritis and neuromyelitis optica, J. Neuroimmunol. 279 (2015) 1–6. [40] W. Tang, X. Wang, Y. Chen, J. Zhang, Y. Chen, Z. Lin, CXCL12 and CXCR4 as predictive biomarkers of glioma recurrence pattern after total resection, Pathol. Biol. (Paris) 63 (4–5) (2015) 190–198.
Contents lists available at ScienceDirect
Journal of the Neurological Sciences journal homepage: www.elsevier.com/locate/jns
Elevation of serum CXC chemokine ligand-12 levels predicts poor outcome after aneurysmal subarachnoid hemorrhage☆ De-Sheng Pan ⁎, Min Yan, Muhammad Hassan, Ze-Bin Fang, Man-Tao Chen Department of Neurosurgery, The First Affiliated Hospital, School of Medicine, Zhejiang University, 79 Qingchun Road, Hangzhou 310003, China
a r t i c l e
i n f o
Article history: Received 14 December 2015 Received in revised form 10 January 2016 Accepted 14 January 2016 Available online 16 January 2016 Keywords: Aneurysm Biomarker CXC chemokine ligand-12 Function outcome Prognosis Subarachnoid hemorrhage Mortality
a b s t r a c t Objective: CXC chemokine ligand-12 (CXCL12) is involved in the innate immune system. Elevation of its level in the peripheral blood is associated with severity and outcome of ischemic stroke. This study aimed to investigate its relation to severity and prognosis following aneurysmal subarachnoid hemorrhage (aSAH). Methods: Serum CXCL12 levels were determined in a total of 182 controls and 182 aSAH patients. Hemorrhagic severity was assessed using the World Federation of Neurological Surgeons (WFNS) scale and modified Fisher grading scale. Unfavorable outcome was defined as Glasgow outcome scale score of 1–3. Prognostic predictors of 6-month mortality and unfavorable outcome were identified using multivariate analysis. Results: The serum CXCL12 levels were significantly higher in patients as compared to controls (14.5 ± 6.7 ng/mL vs. 1.7 ± 0.6 ng/mL, P b 0.001) and were independently associated with WFNS scores (t = 5.927, P b 0.001) and modified Fisher scores (t = 5.506, P b 0.001). Serum CXCL12 levels predicted 6-month mortality and 6-month unfavorable outcome with the area under curves of 0.815 [95% confidence (CI), 0.751–0.868] and 0.809 (95% CI, 0.745–0.864) respectively and were related independently to 6-month mortality (odds ratio, 4.428; 95% CI, 1.977–12.031; P = 0.004) and 6-month unfavorable outcome (odds ratio, 3.821; 95% CI, 1.097–9.251; P = 0.001). Moreover, the predictive values of CXCL12 levels were in the range of WFNS scores and modified Fisher scores. Conclusions: Elevation of serum CXCL12 levels is associated highly with hemorrhagic severity and poor outcome after aSAH, suggesting CXCL12 might have the potential to be a prognostic predictive biomarker of aSAH. © 2016 Elsevier B.V. All rights reserved.
1. Introduction Aneurysmal subarachnoid hemorrhage (aSAH) is a major clinical and health-care problem affecting millions of people worldwide [1–3]. It is important to conduct a precise evaluation of the patient's clinical status in the management of treatment modalities and determination of prognosis. The World Federation of Neurological Surgeons (WFNS) scale and Fisher grading scale are commonly used to evaluate the neurologic status and severity of hemorrhage based on the computed tomography (CT) appearance in aSAH patients [4,5]. In addition, interest in the measurement of blood biomarkers as prognostic predictors of aSAH has increased in the last few years [6–8].
Abbreviations: aSAH, aneurysmal subarachnoid hemorrhage; AUC, area under curve; CI, confidence interval; CT, computerized tomography; CXCL12, CXC chemokine ligand12; CXCR 4, CXC receptor 4; GOS, Glasgow outcome scale; OR, odds ratio; ROC, receiver operating characteristic; SDF-1, stromal cell-derived factor 1; WFNS, World Federation of Neurological Surgeons. ☆ Institution at which the work was performed: The First Affiliated Hospital, School of Medicine, Zhejiang University. ⁎ Corresponding author at: Department of Neurosurgery, The First Affiliated Hospital, School of Medicine, Zhejiang University, 79 Qingchun Road, Hangzhou 310003, China. E-mail address: [email protected] (D.-S. Pan).
http://dx.doi.org/10.1016/j.jns.2016.01.024 0022-510X/© 2016 Elsevier B.V. All rights reserved.
The immune system is a complex and tightly regulated system of effectors against external pathogens and endogenous tissue damage. It can be broadly divided into two component parts, the innate immune system and the adaptive immune system [9,10]. The growing body of evidence has shown that the innate immune system affects brain damage after subarachnoid hemorrhage. Chemokines are small chemoattractant cytokines that are produced in response to damage to call inflammatory cells into the area, thus, regulating the innate immune responses after aSAH [10–12]. The CXC chemokine ligand-12 (CXCL12) is a member of the CXC chemokine subfamily that is constitutively expressed in the bone marrow and other tissues including the brain endothelium and is responsible for regulating the trafficking and localization of bone marrow progenitor cells under steady state and stress conditions [13–15]. Following ischemic stroke, CXCL12 mediates the inflammatory response by recruitment of neural progenitor cells, and the mobilization of bone marrow-derived progenitor cells for tissue regeneration and neovascularization [16]. Other animal experiments had also found that CXCL12 played a significantly beneficial role in acute stroke [17,18]. CXCL12 combing CXC motif receptor 4 could inhibit the caspase-3 pathway by up-regulating Bcl-2/Bax ratio, which protects neurons from apoptosis in rats with traumatic brain injury [19]. Interestingly, elevated circulating CXCL12 level at admission is strongly
54
D.-S. Pan et al. / Journal of the Neurological Sciences 362 (2016) 53–58
associated with the future stroke [20] and future recurrence of ischemic stroke [21], as well as has close relation to stroke severity and lesion volumes [22]. Still, CXCL12 is identified as an independent diagnostic and prognostic marker in patients with acute ischemic stroke [23]. The present study further investigated the potential prognostic role of CXCL12 in Chinese aSAH patients.
2. Methods
2.5. Sampling and laboratory analysis Blood samples were collected from the patients at admission and from controls at study entry. After centrifugation, aliquots of the samples were immediately stored −80 °C until assay. Serum CXCL12 levels were determined in duplicate samples with a quantitative sandwich enzyme-linked immunosorbent assay kit (Quantikine; R&D Systems, Minneapolis, MN, USA) in accordance with the manufactures' instructions. The investigator was blinded to the clinical outcome and neuroimaging findings.
2.1. Ethics statement 2.6. Statistical analysis This work has been carried out in accordance with The Code of Ethics of the World Medical Association (Declaration of Helsinki) and has been also approved by the ethical committee at the First Affiliated Hospital, School of Medicine, Zhejiang University. Still, all participants or their representative were informed of the study protocol and their written informed consent was obtained.
2.2. Participants This is a prospective study from a cohort of patients with confirmed aSAH admitted to the Department of Neurosurgery, First Affiliated Hospital, School of Medicine, Zhejiang University between January 2011 and May 2014. Patients were enrolled in the study with the first-ever non-traumatic SAH, the clinical history of SAH within the last 24 h before admission, the single intracranial aneurysms confirmed by CT angiography with or without digital subtraction angiography and the treatment through clipping or coiling within the 48 h after admission. Exclusion criteria included surgery, trauma or infectious diseases during the preceding 1 month, autoimmune diseases with or without immunosuppressive therapy, prior neurological diseases such as intracerebral hemorrhage and ischemic stroke, previous use of antiplatelet or anticoagulant medication, rebleeding after admission, suspected pseudoaneurysm, and other prior systemic diseases such as uremia, liver cirrhosis, malignancy, chronic heart disease, chronic lung disease, diabetes mellitus and hypertension. The control group consisted of the healthy age- and sex-matched volunteers.
All the results were statistically analyzed using SPSS 19.0 (SPSS Inc., Chicago, IL, USA) and MedCalc 9.6.4.0 (MedCalc Software, Mariakerke, Belgium). Statistical significance was defined as a probability value of less than 0.05. Categorical variables were presented as counts (percentage). Continuous variable were reported as mean ± standard deviation or median (the upper and lower quartiles) as appropriate. The statistical significance of the intergroup observed difference was assessed with the chi-square test, Fisher exact test, Student t test or Mann–Whitney U-test as appropriate. All parameters found to be significant in the univariate analysis were further entered in a binary Logistic regression model to identify prognostic predictors. Linear relationships between CXCL12 levels and other variables were investigated using Spearman's correlation coefficient or Pearson's correlation coefficient as appropriate. We used Bonferroni correction to control for the multiple testing. A multivariate linear regression was carried out further to verify their relationships. Receiver operating curves (ROCs) were generated to determine cutoff values for optimal prognostic predictive sensitivities and specificities. The area under curves (AUCs) were estimated in accordance with the ROC curves and the additive benefits of CXCL12 levels to the WFNS scores and modified Fisher scores were evaluated in a combined logistic-regression model. 3. Results 3.1. Participants characteristics
2.3. Assessment The neurological status was graded according to the WFNS scale [4] and the radiological severity of SAH was classified based on the modified Fisher grade [5]. The WFNS assessments were performed by at least two clinicians. Radiologists helped to determine the modified Fisher grade. When two raters obtained different scores, a judgment of a third senior clinician was considered. Symptomatic cerebral vasospasm was defined as the development of new focal neurological signs, deterioration in level of consciousness, or the appearance of new infarction on CT when the cause was felt to be ischemia attributable to vasospasm after other possible causes of worsening (e.g. hydrocephalus, seizures, metabolic derangement, infection, or oversedation) had been excluded [24,25].
2.4. Endpoint Participants were followed up until death or completion of 6 months after aSAH. The endpoints were unfavorable outcome and death after 6 months. The functional outcome was defined by Glasgow outcome scale (GOS) score. GOS scores were dichotomized in favorable and unfavorable outcomes (GOS of 4–5 vs. GOS of 1–3) [26–28]. For follow-up, structured telephone interviews were performed by one doctor who was blinded to clinical information.
During the study period, 239 aSAH patients were initially evaluated. 57 patients were excluded because of the reasons listed in Fig. 1. Eventually, 182 aSAH patients were included in this study. Additionally, 182 gender- and age-matched healthy subjects were enrolled as controls. This group of aSAH patients was composed of 77 males and 105 females as well as had a mean age of 50.6 ± 11.4 years (range, 34– 77 years). The median admission WFNS scores and modified Fisher scores were 3 (2–3) (range, 1–5) and 3 (2–4) (range, 2–5) respectively. Aneurysmal location is as follows: 48 (26.4%) aneurysms were located at posterior communication artery; 31 (17.0%) aneurysms, internal carotid artery; 37 (20.3%) aneurysms, anterior communication artery; 33 (18.1%) aneurysms, middle cerebral artery; 22 (12.1%) aneurysms, anterior cerebral artery; 8 (4.4%) aneurysms, posterior cerebral artery; 3 (1.7%) aneurysms, vertebral artery. 146 (80.2%) patients had cystic aneurysm. The mean aneurysm diameter was 7.9 ± 5.4 mm (range, 2– 25 mm). 96 (52.8%) patients underwent clipping and other patients obtained endovascular coiling. 37 (20.3%) patients had acute hydrocephalus; 29 (15.9%) patients, intraventricular hemorrhage; 43 (23.6%) patients, underwent external ventricular drain; 52 (28.6%) patients, symptomatic cerebral vasospasm. 3.2. The change of serum CXCL12 levels in aSAH patients Patients had the significant elevation of serum CXCL12 levels as compared with normal controls (14.5 ± 6.7 ng/mL vs. 1.7 ±
D.-S. Pan et al. / Journal of the Neurological Sciences 362 (2016) 53–58
55
Fig. 1. Flow chart illustrating excluded and included patients with aneurysmal subarachnoid hemorrhage.
0.6 ng/mL, P b 0.001). Additionally, the serum CXCL12 levels were obviously higher in non-survivors than in survivors (21.1 ± 5.4 ng/mL vs. 13.2 ± 6.1 ng/mL, P b 0.001) and in patients with unfavorable outcome than with favorable outcome (19.4 ± 4.9 ng/mL vs.12.0 ± 6.0 ng/mL, P b 0.001).
CXCL12 numerically improved the AUCs of WFNS score and modified Fisher score to 0.904 (95% CI 0.852–0.943, P = 0.208) and 0.894 (95% CI 0.840–0.935, P = 0.149) respectively. In Table 2, a univariate analysis found that serum CXCL12 levels higher than 18.1 ng/mL and other variables were significantly
3.3. Correlation analysis
Table 1 The relationship between serum CXC chemokine ligand-12 levels and other parameters following aneurysmal subarachnoid hemorrhage.
In Table 1, some variables including the WFNS scores, modified Fisher scores, acute hydrocephalus, intraventricular hemorrhage, symptomatic cerebral vasospasm, external ventricular drain, blood glucose levels and plasma C-reactive protein levels were associated highly with serum CXCL12 levels. We further used Bonferroni correction to control for the multiple testing. Because all P values, which exhibited statistically significant, were less than 0.002 (0.05/21), their linear relationships were not changed. In Fig. 2, serum CXCL12 levels still remained closely related to WFNS scores (t = 5.927, P b 0.001) and modified Fisher scores (t = 5.506, P b 0.001) in a multivariate linear regression model. 3.4. Mortality prediction 30 patients (16.5%) died within 6 month after aSAH. Based on the ROC curve, the optimal cutoff value of serum CXCL12 levels as an indicator for prognosis of the mortality was projected to be 18.1 ng/mL, which yielded a sensitivity of 80.0% and a specificity of 72.4%, with the AUC at 0.815 [95% confidence interval (CI) = 0.751–0.868] in Fig. 3A. Its AUC was in the range of WFNS score (AUC 0.878, 95% CI 0.821–0.921, P = 0.248) and modified Fisher score (AUC 0.861, 95% CI 0.802–0.908, P = 0.358).When a combined logistic-regression model was configured,
Characteristics
r value
P value
Gender Age (year) WFNS scores Modified Fisher scores Aneurysmal locations Cystic aneurysm Aneurysm diameter (mm) Treatment (clipping/endovascular coiling) Acute hydrocephalus Intraventricular hemorrhage External ventricular drain Symptomatic cerebral vasospasm Admission time (h) Plasma-sampling time (h) Systolic arterial pressure (mm Hg) Diastolic arterial pressure (mm Hg) Blood glucose level (mmol/L) Plasma C-reactive protein level (mg/L) Blood white blood cell count (×109/L) Blood hemoglobin level (g/L) Blood platelet count (×109/L)
0.052 0.127 0.539 0.525 0.015 0.028 0.084 0.056 0.279 0.235 0.309 0.423 0.044 0.048 0.003 0.136 0.260 0.305 0.035 0.128 0.117
0.488 0.087 b0.001 b0.001 0.842 0.703 0.260 0.449 b0.001 0.001 b0.001 b0.001 0.554 0.523 0.970 0.067 b0.001 b0.001 0.640 0.084 0.116
Bivariate correlations were assessed by Spearman's or Pearson's correlation coefficient. WFNS indicates World Federation of Neurological Surgeons.
56
D.-S. Pan et al. / Journal of the Neurological Sciences 362 (2016) 53–58
Fig. 3. Receiver operating characteristic curve analysis of serum chemokine ligand-12 (CXCL12) levels for 6-month mortality (A) and 6-month unfavorable outcome (B) after aneurysmal subarachnoid hemorrhage. 95% CI indicates confidence interval. Fig. 2. Relationships between serum chemokine ligand-12 (CXCL12) levels and the World Federation of Neurological Surgeons (WFNS) scores as well as between serum CXCL12 levels and modified Fisher scores.
associated with the mortality. A regression analysis showed that WFNS score [Odds ratio (OR), 6.054; 95% CI, 2.314–15.837; P b 0.001], modified Fisher scores (OR, 4.110; 95% CI, 1.660–10.175; P = 0.001) and serum CXCL12 levels higher than 18.1 ng/mL (OR, 4.428; 95% CI, 1.977–12.031; P = 0.004) were the independent predictors of mortality.
3.5. Poor outcome prediction In our study, an unfavorable functional outcome was found in 62 patients (34.1%) within 6 months after aSAH. The ROC curve analysis showed that the optimal cutoff value of serum CXCL12 levels to predict unfavorable outcome was 18.0 ng/mL, with the AUC of 0.809 (95% CI = 0.745–0.864), which had specificity and sensitivity values of 71.0% and 80.0% respectively in Fig. 3B. Its AUC was not statistically significantly different from those of WFNS score (AUC 0.869, 95% CI 0.811–0.914, P = 0.158) and modified Fisher score (AUC 0.841, 95% CI 0.779–0.890, P = 0.462). The combined model found that CXCL12 enhanced the AUCs of WFNS score and modified Fisher score to 0.889 (95% CI 0.834–0.931) and 0.867 (95% CI 0.808–0.912) respectively, but the differences were not statistically significant (P = 0.259 and 0.343). In Table 3, serum CXCL12 levels higher than 18.0 ng/mL and other parameters had close relation to the unfavorable outcome using a univariate analysis. WFNS score (OR, 5.890; 95% CI, 2.710–12.803; P b 0.001), modified Fisher scores (OR, 4.844; 95% CI, 2.257–10.397; P b 0.001) and serum CXCL12 levels higher than 18.0 ng/mL (OR, 3.821; 95% CI, 1.097–9.251; P = 0.001) were identified as the
independent predictors of the unfavorable outcome in a binary Logistic regression model. 4. Discussion Although previous papers had found the prognostic and diagnostic predictive significance of serum CXCL12 in ischemic stroke [20–23], the role of CXCL12 has not been explored in aSAH. This study configured the multiple multivariate models and showed some interesting results. Its main findings were that, (1) serum CXCL12 levels of aSAH patients were obviously elevated compared with controls; (2) serum CXCL12 levels had an independent relation to WFNS scores and modified Fisher scores; (3) serum CXCL12 levels could independently predict 6-month mortality and 6-month functional outcome; (4). This biomarker's AUC was in the range of WFNS scores and modified Fisher scores. These data suggests that CXCL12 might be associated with the severity and clinical outcomes of aSAH. CXCL12, also known as stromal cell-derived factor 1 (SDF-1), is subdivided into 6 isoforms: SDF-1a, SDF-1b, SDF-1g, SDF-1d, SDF-1e, and SDF-14. Among these, SDF-1a is the most abundant and plays a variety of roles in different tissues [29]. In ischemic stroke, SDF-1a, binding to the CXC receptor 4 (CXCR 4), is involved in the focal angiogenesis in animal ischemic tissue [30]. In head trauma rats, SDF-1a, combing CXCR4 receptors, could inhibit the caspase-3 pathway by upregulating Bcl-2/Bax ratio, which protects neurons from apoptosis [19]. In vitro experiment has also demonstrated that CXCL12 could promote neuron survival by protecting them from apoptosis [31]. A recent study showed that SDF-1α could help to accelerate clot organization
D.-S. Pan et al. / Journal of the Neurological Sciences 362 (2016) 53–58 Table 2 The comparisons of clinical characteristics and laboratory data between 6-month non-surviving and surviving patients with aneurysmal subarachnoid hemorrhage.
Gender (male/female) Age (year) WFNS scores Modified Fisher scores Aneurysmal location Posterior communication artery Internal carotid artery Anterior communication artery Middle cerebral artery Anterior cerebral artery Posterior cerebral artery Vertebral artery Cystic aneurysm Aneurysm diameter (mm) Treatment (clipping/endovascular coiling) Acute hydrocephalus Intraventricular hemorrhage External ventricular drain Symptomatic cerebral vasospasm Admission time (h) Plasma-sampling time (h) Systolic arterial pressure (mm Hg) Diastolic arterial pressure (mm Hg) Blood glucose level (mmol/L) Plasma C-reactive protein level (mg/L) Blood white blood cell count (×109/L) Blood hemoglobin level (g/L) Blood platelet count (×109/L) Serum CXCL12 levels higher than 18.1 ng/mL
Non-survivors n = 30
Survivors n = 152
P value
15/15 55.1 ± 11.9 4 (3–4) 4 (3–4)
62/90 49.7 ± 11.1 2 (2–3) 3 (2–3)
0.351 0.017 b0.001 b0.001 0.260
8 (26.7%) 5 (16.6%) 2 (6.7%) 9 (30.0%) 3 (10.0%) 2 (6.7%) 1 (3.3%) 25 (83.3%) 7.5 ± 4.6 18/12
40 (26.3%) 26 (17.2%) 35 (23.0%) 24 (15.8%) 19 (12.5%) 6 (3.9%) 2 (1.3%) 121 (79.6%) 8.0 ± 5.5 78/74
15 (50.0%) 11 (36.7%) 18(60.0%) 16 (53.3%) 10.6 ± 3.7 12.6 ± 4.6 129.6 ± 21.6 78.0 ± 13.7 12.7 ± 3.8 17.7 ± 4.9 7.6 ± 3.0 119.4 ± 21.7 190.2 ± 36.0 24 (80.0%)
22 (14.5%) 18 (11.8%) 25 (16.5%) 36 (23.7%) 9.7 ± 3.7 14.0 ± 5.1 131.8 ± 23.5 78.1 ± 11.6 10.5 ± 4.1 15.3 ± 3.9 8.4 ± 3.0 118.4 ± 23.8 181.9 ± 47.0 42 (27.6%)
0.639 0.656 0.384 b0.001 0.002 b0.001 0.001 0.234 0.170 0.636 0.969 0.008 0.004 0.223 0.825 0.365 b0.001
The variables are presented as percentages, mean ± standard deviation or median (the upper and lower quartiles). The comparisons were assessed using the chi-square test, Fisher exact test, Student t test or Mann–Whitney U test as appropriate. WFNS indicates World Federation of Neurological Surgeon; CXCL12, CXC chemokine ligand-12.
and endothelial cell proliferation after implantation of coils for embolization of aneurysm [32]. Thus, CXCL12 might be involved in the pathogenesis of many neurodegenerative diseases including aSAH and could exert a protective effect on acute brain injury. Chemokines are expressed in neuron, astrocytes and microglia in central nervous system [33]. SDF-1 expression is associated with reactive astrocytes [34]. It has been reported that microglia are activated and chemoattracted by SDF-1 [35]. Furthermore, microglia activation leads to tumor necrosis factor-alpha production in astrocytes [36]. Still, SDF-1α up-regulates interleukin-6 through CXCR4, phosphatidylinositol-3-kinase/serine/threonine kinase, extracellular signal-regulated kinase, and nuclear factor-kappaB-dependent pathway in microglia [37]. These results provide the support for the notion that CXCL12 plays a regulatory role in neuroinflammation. Given the fact that neuroinflammation contributes to the pathogenesis of diffuse brain injury after aSAH [38], the potential role of CXCL12 is established for the regulation of diffuse brain injury after aSAH. However, a mechanism linking CXCL12 and neuroinflammation after aSAH remains to be studied. The accumulating evidence has demonstrated that CXCL12 levels have significant prognostic predictive values for some neurologic diseases [22,23,39,40]. Elevated serum CXCL12 levels have been found after acute ischemic stroke. Moreover, serum CXCL12 was positively correlated with infarct volume, stroke severity, mortality and unfavorable functional outcome [22,23]. In our study, elevation of serum CXCL12 levels was found; CXCL12 was verified to be associated with aSAH severity reflected by WFNS score and modified Fisher score; CXCL12 was identified as an independent predictor for poor outcomes
57
Table 3 The factors correlated with 6-month functional outcome after aneurysmal subarachnoid hemorrhage.
Gender (male/female) Age (year) WFNS scores Modified Fisher scores Aneurysmal location Posterior communication artery Internal carotid artery Anterior communication artery Middle cerebral artery Anterior cerebral artery Posterior cerebral artery Vertebral artery Cystic aneurysm Aneurysm diameter (mm) Treatment (clipping/endovascular coiling) Acute hydrocephalus Intraventricular hemorrhage External ventricular drain Symptomatic cerebral vasospasm Admission time (h) Plasma-sampling time (h) Systolic arterial pressure (mm Hg) Diastolic arterial pressure (mm Hg) Blood glucose level (mmol/L) Plasma C-reactive protein level (mg/L) Blood white blood cell count (×109/L) Blood hemoglobin level (g/L) Blood platelet count (×109/L) Serum CXCL12 levels higher than 18.1 ng/mL
Unfavorable outcome n = 62
Favorable outcome n = 120
P value
28/34 53.3 ± 12.3 3 (3–4) 4 (3–4)
49/71 49.2 ± 10.7 2 (1–3) 3 (2–3)
0.575 0.019 b0.001 b0.001 0.076
20 (32.3%) 11 (17.7%) 6 (9.7%) 10 (16.1%) 8 (12.9%) 5 (8.1%) 2 (3.2%) 49 (79.0%) 7.5 ± 4.8 36/26
28 (23.3%) 20 (16.7%) 31 (25.8%) 23 (19.2%) 14 (11.7%) 3 (2.5%) 1 (0.8%) 97 (80.8%) 8.1 ± 5.6 60/60
24 (38.7%) 20 (32.3%) 29 (46.8%) 31 (50.0%) 9.3 ± 3.1 13.9 ± 5.0 135.5 ± 22.5 76.8 ± 11.9 12.0 ± 4.4 17.0 ± 4.4
13 (10.8%) 9 (7.5%) 14 (11.7%) 21 (17.5%) 10.1 ± 4.0 13.8 ± 5.1 129.4 ± 23.2 78.7 ± 12.0 10.2 ± 3.9 15.0 ± 4.0
8.6 ± 3.7
8.1 ± 2.6
118.7 ± 23.7 184.0 ± 45.7 44 (71.0%)
118.5 ± 23.4 182.9 ± 45.5 24 (20.0%)
0.773 0.520 0.302 b0.001 b0.001 b0.001 0.001 0.201 0.886 0.096 0.312 0.008 0.003 0.245 0.954 0.870 b0.001
The variables are presented as percentages, mean ± standard deviation or median (the upper and lower quartiles). Comparisons were assessed using the chi-square test, Fisher exact test, Student t test or Mann–Whitney U test as appropriate. WFNS indicates World Federation of Neurological Surgeon; CXCL12, CXC chemokine ligand-12.
including mortality and unfavorable outcome; interestingly, its predictive value was in the range of WFNS score and modified Fisher score. Overall, CXCL12 in serum was correlated with hemorrhagic severity and clinical outcomes after aSAH. Our study has two limitations. Firstly, blood samples were collected only once after aSAH. Serial follow-up studies of CXCL12 are needed because its levels can change over time, hence additional measurements in the days thereafter would have been of interest. Secondly, we measured CXCL12 in serum, not in cerebral spinal fluid. It is still uncertain whether peripheral CXCL12 levels reflect similar changes in the central nervous system. This aspect of study warrants to be explored in future. 5. Conclusions The results of this study demonstrated that elevated serum CXCL12 levels had close relation to WFNS scores and modified Fisher scores, in addition to 6-month mortality and 6-month unfavorable outcome, suggesting CXCL12 might have the potential to be a prognostic predictive biomarker after aSAH. Competing interests The authors declare that they have no competing interests. Conflict of interest The authors report no conflicts of interest concerning the materials or methods used in this study or the findings specified in this article.
58
D.-S. Pan et al. / Journal of the Neurological Sciences 362 (2016) 53–58
Acknowledgment The authors thank all staffs in the Department of Neurosurgery, The First Affiliated Hospital, School of Medicine, Zhejiang University for their technical support. References [1] Y. Wang, Y. Tian, D. Wang, H. Wei, Z. Zhao, R. Jiang, et al., High angiopoietin-1 levels predict a good functional outcome within 72 h of an aneurysmal subarachnoid hemorrhage: a prospective study from a single center, J. Neurol. Sci. 356 (1–2) (2015) 72–76. [2] C. Lindgren, S. Söderberg, L.O. Koskinen, M. Hultin, L. Edvardsson, S. Naredi, Longterm subarachnoid haemorrhage survivors still die due to cerebrovascular causes, Acta Neurol. Scand. 132 (2015) 410–416. [3] A. Rodríguez-Rodríguez, J.J. Egea-Guerrero, Z. Ruiz de Azúa-López, F. MurilloCabezas, Biomarkers of vasospasm development and outcome in aneurysmal subarachnoid hemorrhage, J. Neurol. Sci. 341 (1–2) (2014) 119–127. [4] C. Drake, Report of World Federation of Neurological Surgeons Committee on a universal subarachnoid hemorrhage grading scale, J. Neurosurg. 68 (6) (1988) 985–986. [5] C.M. Fisher, J.P. Kistler, J.M. Davis, Relation of cerebral vasospasm to subarachnoid hemorrhage visualized by computerized tomographic scanning, Neurosurgery 6 (1) (1980) 1–9. [6] Y.F. Shen, W.H. Wang, W.H. Yu, X.Q. Dong, Q. Du, D.B. Yang, et al., The prognostic value of plasma thrombospondin-1 concentrations after aneurysmal subarachnoid hemorrhage, Clin. Chim. Acta 448 (2015) 155–160. [7] X.D. Chen, J. Sun, C. Lu, H.J. Ba, M.H. Chen, J.H. Lin, et al., The prognostic value of plasma soluble CD40 ligand levels following aneurysmal subarachnoid hemorrhage, Thromb. Res. 136 (1) (2015) 24–29. [8] J.Y. Cai, X.D. Chen, H.J. Ba, J.H. Lin, C. Lu, M.H. Chen, et al., Identification of plasma adrenomedullin as a possible prognostic biomarker for aneurysmal subarachnoid hemorrhage, Peptides 59 (2014) 9–13. [9] S.T. Shibutani, T. Saitoh, H. Nowag, C. Münz, T. Yoshimori, Autophagy and autophagy-related proteins in the immune system, Nat. Immunol. 16 (10) (2015) 1014–1024. [10] S.P. Gadani, J.T. Walsh, J.R. Lukens, J. Kipnis, Dealing with danger in the CNS: the response of the immune system to injury, Neuron 87 (1) (2015) 47–62. [11] C.Z. Chang, S.C. Wu, C.L. Lin, A.L. Kwan, Valproic acid attenuates intercellular adhesion molecule-1 and E-selectin through a chemokine ligand 5 dependent mechanism and subarachnoid hemorrhage induced vasospasm in a rat model, J. Inflamm. (Lond.) 12 (2015) 27. [12] G.H. Kim, C.P. Kellner, D.K. Hahn, B.M. Desantis, M. Musabbir, R.M. Starke, et al., Monocyte chemoattractant protein-1 predicts outcome and vasospasm following aneurysmal subarachnoid hemorrhage, J. Neurosurg. 109 (1) (2008) 38–43. [13] Q. Guo, B.L. Gao, X.J. Zhang, G.C. Liu, F. Xu, Q.Y. Fan, et al., CXCL12-CXCR4 axis promotes proliferation, migration, invasion, and metastasis of ovarian cancer, Oncol. Res. 22 (5) (2015) 247–258. [14] T. Nagasawa, CXCL12/SDF-1 and CXCR4, Front. Immunol. 6 (2015) 301. [15] F. Arenzana-Seisdedos, SDF-1/CXCL12: a chemokine in the life cycle of HIV, Front. Immunol. 6 (2015) 256. [16] M. Li, R.M. Ransohoff, Multiple roles of chemokine CXCL12 in the central nervous system: a migration from immunology to neurobiology, Prog. Neurobiol. 84 (2) (2008) 116–131. [17] A.M. Robin, Z.G. Zhang, L. Wang, R.L. Zhang, M. Katakowski, L. Zhang, et al., Stromal cell-derived factor 1alpha mediates neural progenitor cell motility after focal cerebral ischemia, J. Cereb. Blood Flow Metab. 26 (1) (2006) 125–134. [18] X. Yu, D. Chen, Y. Zhang, X. Wu, Z. Huang, H. Zhou, et al., Overexpression of CXCR4 in mesenchymal stem cells promotes migration, neuroprotection and angiogenesis in a rat model of stroke, J. Neurol. Sci. 316 (1–2) (2012) 141–149.
[19] W. Mao, X. Yi, J. Qin, M. Tian, G. Jin, CXCL12 inhibits cortical neuron apoptosis by increasing the ratio of Bcl-2/Bax after traumatic brain injury, Int. J. Neurosci. 124 (4) (2014) 281–290. [20] R.C. Schutt, M.D. Burdick, R.M. Strieter, B. Mehrad, E.C. Keeley, Plasma CXCL12 levels as a predictor of future stroke, Stroke 43 (12) (2012) 3382–3386. [21] X.L. Gu, L. Liu, X.D. Lu, Z.R. Liu, Serum CXCL12 levels as a novel predictor of future stroke recurrence in patients with acute ischemic stroke, Mol. Neurobiol. (2015)http://dx.doi.org/10.1007/s12035-015-9151-0. [22] P. Liu, J.W. Xiang, S.X. Jin, Serum CXCL12 levels are associated with stroke severity and lesion volumes in stroke patients, Neurol. Res. 37 (10) (2015) 853–858. [23] X.X. Duan, G.P. Zhang, X.B. Wang, H. Yu, J.L. Wu, K.Z. Liu, et al., The diagnostic and prognostic value of serum CXCL12 levels in patients with ischemic stroke, Neurol. Sci. 36 (12) (2015) 2227–2234. [24] J. Claassen, G.L. Bernardini, K. Kreiter, J. Bates, Y.E. Du, D. Copeland, et al., Effect of cisternal and ventricular blood on risk of delayed cerebral ischemia after subarachnoid hemorrhage: the Fisher scale revisited, Stroke 32 (9) (2001) 2012–2020. [25] J.A. Frontera, A. Fernandez, J.M. Schmidt, J. Claassen, K.E. Wartenberg, N. Badjatia, et al., Defining vasospasm after subarachnoid hemorrhage: what is the most clinically relevant definition? Stroke 40 (6) (2009) 1963–1968. [26] X.D. Zhu, J.S. Chen, F. Zhou, Q.C. Liu, G. Chen, J.M. Zhang, Relationship between plasma high mobility group box-1 protein levels and clinical outcomes of aneurysmal subarachnoid hemorrhage, J. Neuroinflammation 9 (2012) 194. [27] X.D. Zhu, J.S. Chen, F. Zhou, Q.C. Liu, G. Chen, J.M. Zhang, Detection of copeptin in peripheral blood of patients with aneurysmal subarachnoid hemorrhage, Crit. Care 15 (6) (2011) R288. [28] J.W. Pan, L.N. He, F. Xiao, J. Shen, R.Y. Zhan, Plasma gelsolin levels and outcomes after aneurysmal subarachnoid hemorrhage, Crit. Care 17 (4) (2013) R149. [29] B. Nash, O. Meucci, Functions of the chemokine receptor CXCR4 in the central nervous system and its regulation by μ-opioid receptors, Int. Rev. Neurobiol. 118 (2014) 105–128. [30] Y. Wang, J. Huang, Y. Li, G.Y. Yang, Roles of chemokine CXCL12 and its receptors in ischemic stroke, Curr. Drug Targets 13 (2) (2012) 166–172. [31] M.Z. Khan, R. Brandimarti, S. Shimizu, J. Nicolai, E. Crowe, O. Meucci, The chemokine CXCL12 promotes survival of postmitotic neurons by regulating Rb protein, Cell Death Differ. 15 (10) (2008) 1663–1672. [32] Y. Gao, Q. Wang, X. Cui, Y. Liu, T. Zheng, C. Chen, et al., Controlled release of stromal cell-derived factor-1α from silk fibroin-coated coils accelerates intra-aneurysmal organization and occlusion of neck remnant by recruiting endothelial progenitor cells, Int. J. Clin. Exp. Pathol. 7 (12) (2014) 8366–8380. [33] A. Bajetto, R. Bonavia, S. Barbero, T. Florio, A. Costa, G. Schettini, Expression of chemokine receptors in the rat brain, Ann. N. Y. Acad. Sci. 876 (1999) 201–209. [34] W.D. Hill, D.C. Hess, A. Martin-Studdard, J.J. Carothers, J. Zheng, D. Hale, et al., SDF-1 (CXCL12) is upregulated in the ischemic penumbra following stroke: association with bone marrow cell homing to injury, J. Neuropathol. Exp. Neurol. 63 (1) (2004) 84–96. [35] S. Tanabe, M. Heesen, I. Yoshizawa, M.A. Berman, Y. Luo, C.C. Bleul, et al., Functional expression of the CXC-chemokine receptor-4/fusin on mouse microglial cells and astrocytes, J. Immunol. 159 (2) (1997) 905–911. [36] P. Bezzi, M. Domercq, L. Brambilla, R. Galli, D. Schols, E. De Clercq, et al., CXCR4activated astrocyte glutamate release via TNFalpha: amplification by microglia triggers neurotoxicity, Nat. Neurosci. 4 (7) (2001) 702–710. [37] D.Y. Lu, C.H. Tang, W.L. Yeh, K.L. Wong, C.P. Lin, Y.H. Chen, et al., SDF-1alpha upregulates interleukin-6 through CXCR4, PI3K/Akt, ERK, and NF-kappaB-dependent pathway in microglia, Eur. J. Pharmacol. 613 (1–3) (2009) 146–154. [38] K.A. Hanafy, The role of microglia and the TLR4 pathway in neuronal apoptosis and vasospasm after subarachnoid hemorrhage, J. Neuroinflammation 10 (2013) 83. [39] C. Tingjun, L. Zhaohui, J. Zhaocai, L. Zihao, X. Quangang, H. Dehui, et al., Changes of CXCL12, CXCL14 and PDGF levels in the brain of patients with idiopathic demyelinating optic neuritis and neuromyelitis optica, J. Neuroimmunol. 279 (2015) 1–6. [40] W. Tang, X. Wang, Y. Chen, J. Zhang, Y. Chen, Z. Lin, CXCL12 and CXCR4 as predictive biomarkers of glioma recurrence pattern after total resection, Pathol. Biol. (Paris) 63 (4–5) (2015) 190–198.