Design

This was a prospective case-control study (3:1) whose inclusion criteria for the stroke group (cases) were: patient of any age, symptoms of non-lacunar brain infarction within the previous 12 hours and a brain CT that ruled out any other cause of the neurological deficit. The brain CT was repeated within the first week to demonstrate the presence of a brain infarction and to measure its volume. The hypodensity volume (mL) was calculated in the second scan according to the formula 0.5XaXbXc (where a and b are the largest perpendicular diameters measured in CT and c is the height). Patients with symptoms or signs of an acute infection on admission as well as those with a transient ischemic attack (TIA), lacunar syndrome (pure motor stroke/hemiparesis, ataxic hemiparesis, dysarthria/clumsy had, pure sensory stroke and mixed sensorimotor stroke) or symptoms of brainstem infarction were excluded. Furthermore, patients treated with intravenous thrombolysis or intra-arterial reperfusion therapies were excluded. The controls were patients with acute non-neurological diseases with symptom onset within the previous 12 hours. Cases and controls were matched by gender and age (±5 years). Exclusion criteria for the control group were: history of acute stroke in the last year or current episode of stroke, focal or global neurological symptoms attributable to any central nervous system lesions (whether demonstrated or not) and more than one acute disease at the time of the study. Common exclusion criteria for both groups were: underlying severe conditions including bronchial and heart disease requiring more than 3 hospital admissions in the last year or domiciliary oxygen therapy; disseminated or terminal stage cancer in any location; connective tissue disorders with current activity at the time of evaluation; previous chronic inflammatory diseases such as chronic bronchitis; treatment with anti-inflammatory drugs or calcium channel blockers at symptoms onset; dementia or other deteriorating conditions that preclude an acceptable basal functional status, as determined by Barthel Index (BI) scores lower than 90 and modified Rankin scale (mRS) scores higher than 1; and a history of illicit drug abuse.

Patient management

Both cases and controls were selected among those who came to the emergency department, according to the inclusion and exclusion criteria. Stroke patients were evaluated by a neurologist and managed in a Stroke Unit following the acute stroke management guidelines [18] with special attention to the surveillance of blood pressure, serum glucose levels and body temperature. The initial assessment of the neurological deficit by means of the National Institute of Health stroke scale (NIHSS) was performed on admission and every 24 hours during hospitalization by certified neurologists. Control patients were evaluated following a standard protocol of management depending on the clinical symptoms. The score in the APACHE II (Acute Physiology and Chronic Health Evaluation II) scale score was registered in cases and controls to establish disease severity [19].

Demographic data, previous vascular risk factors and previous treatments were recorded for both groups and registered in a specific database. Furthermore, functional outcomes at 3-months were evaluated by means of the mRS and BI [20]. The mRS is a scale defined by seven grades (0 indicates no symptoms; 1: no disability, despite symptoms; 2: slight disability; 3: moderate disability; 4: moderately severe disability; 5 severe disability; and 6 death). The BI consists of 10 items that measure a person’s daily functioning, specifically the activities of daily living and mobility. The items include feeding, moving from wheelchair to bed and return, grooming, transferring to and from a toilet, bathing, walking on level surface, going up and down stairs, dressing, continence of bowels and bladder. The total scores ranges from 0 to 100, where 0 is the highest functional dependence and 100 the maximum independence.

The study was approved by the local Ethical Committee of La Paz University Hospital. All patients, or next kin if patient was unable to consent, signed the informed consent before the inclusion in the study. The informed consent followed the standards of the Spanish law (Boletín Oficial del Estado of November 15, 2002).

Blood sample management and biochemistry analysis

Laboratory analyses were performed within the first 12 hours from the onset of symptoms and at 72 hours, both in case and controls. Blood samples were extracted by peripheral venous puncture, collected in glass tubes with EDTA K3, centrifuged at 3000g for 15 minutes and the supernatant frozen at –80°C for storage in the biochemistry department. Plasma levels of Glu and other amino acids were measured for physiological amino acids by High Performance Liquid Chromatography (HPLC) following the Pico-Tag© method (Waters Associates) [21] with minor modifications [22]. The reference normal plasma range for Glu in our laboratory was 25 to 110 μmol/l. The pro-inflammatory cytokines IL-6 and the TNF-α were determined by ELISA (Inmunogenetics, S.A.U.). The biochemistry physician was blind to the patients’ clinical characteristics, including their allocation to case or control groups.

Subjects

Adult male Long-Evans rats (weight 250 to 300g) were used. Rats were housed with free access to food and water and at a room temperature of 21 ± 2°C, relative humidity of 45 ± 15% and a 12h light/ dark cycle (7:00-19:00).

Experimental groups

We conducted a randomized, blind study. Rats were assigned to five groups: 1) ES (n = 10), which underwent a right internal carotid artery (ICA) embolization with autologous clot; 2) sham-operated group (n = 6) without embolization; 3) pMCAO (n = 6) with permanent MCAO together with transient bilateral common carotid artery ligation for 60 min; 4) sham-operated group (n = 6) without pMCAO; and 5) Controls which underwent tissue stress by leg compression (LC) for 180 min (n = 6). These animals were subjected to limb muscle compression in back paw without carotid or cerebral damage.

After 72h, animals were re-anesthetized for euthanasia by transcardial perfusion with a saline solution followed by fixing solution (4% paraformaldehyde and 0.1% glutaraldehyde in 10% buffered formalin phosphate).

ES model: surgical procedures

Anesthesia was induced by a solution of ketamine (25 mg/mL), diazepam (2 mg/mL), and atropine (0.1 mg/mL) at a dose of 2.5 ml/kg by intraperitoneal injection. Analgesia was provided by meloxicam 2 mg/kg by a subcutaneous route. The femoral vein and artery were cannulated for continuous monitoring of physiological parameters (glycemia, blood gases, blood pressure and heart rate) (Monitor Schiller AG CH 6340BAAR), and for extraction of samples. Body temperature was also monitored and maintained at 36.5 ± 0.5°C. The external carotid artery (ECA) was also cannulated to introduce embolus. This consisted of a 3 mm long by 0.4 mm wide thrombus obtained from arterial blood coagulated in a polyethylene tube (Centracath Vygon 19 G, inside diameter 0.5 mm) at 37.5°C for 40 min. As previously described [15],[23], the thrombus was introduced into the ICA through a catheter located in the ECA, enabling the blood flow to impel it up to the bifurcation of the intracranial ICA where it impacts due to its diameter, causing interruption of blood flow to the middle cerebral artery (MCA). The location of the clot and proper occlusion of MCA were verified by an angiography. Angiography (Stenoscop General Electric. Exposure: 40 Kv, 0.5 mA) with 0.3 ml of non-ionic contrast (Iohexol Omnitrast 300 Schering) was performed at 20 min after embolization to verify the arterial occlusion and again at 120 min to check whether recanalization had occurred. Animals that did not present arterial occlusion after the first angiography or those that spontaneously recanalized (animals that show MCA patency on the second angiogram) were rejected. Using these criteria, we ensured that all animals included in the study had an MCA occlusion and did not present early reperfusion.

The sham-operated animals underwent the entire surgical procedure except for embolization.

pMCAO model: surgical procedures

The anesthesia induction and the monitoring of physiological parameters were similar to those previously described for ES.

The surgical procedure to induce permanent focal cerebral ischemia was a variant of that described by Chen et al [16] and Liu et al [17]. A small craniectomy was made above the rhinal fissure over the right MCA branch, which was permanently ligated just before its bifurcation between the frontal and parietal branches with a 9-0 suture. Complete blood flow interruption was confirmed using an operating microscope. Both common carotid arteries were then transitorily occluded for 180 min. A thermistor probe was placed under the temporal muscle and over the cerebral artery region to measure brain temperature. Transient occlusion of common carotid arteries helps to reduce the variability of infarct volume in this model [24].

The sham-operated animals underwent the entire surgical procedure except for MCA ligation.

Physiological monitoring: mortality

In all animals, the femoral artery was cannulated during surgery and ischemia for continuous monitoring of physiological parameters (glycemia, blood gases and blood pressure) (Monitor Omicron ALTEA RGB medical devices). Temperature was maintained at 36.5 ± 0.5°C. A deviation of less than 20% from normal mean values of physiological parameters was accepted; animals with values outside these normal limits were rejected.

In the ES, 4 rats died before the 72^th hour due to the severity of the brain infarction, and they were excluded from the analysis. There was no mortality before the sacrifice in the pMCAO and LC rats.

Functional evaluation

Before the procedure and at 24 and 48 hours after surgery and leg compression, each animal was given a score on the neurological scale described by Rogers [25],[26]: 0 = No deficit; 1 = failure to extend left forelimb; 2 = decreased grip of the left forelimb while tail pulled; 3 = spontaneous movement in all directions, contralateral circling if pulled by the tail; 4 = circling or walking to the left; 5 = movement only when stimulated; 6 = unresponsive to stimulation; 7 = death.

All animals were weighed preoperatively and immediately before the animals were euthanized at 72 hours. Post-operative weight loss as a percentage of preoperative weight was calculated as: 100 × (preoperative weight – pre-euthanization weight)/preoperative weight.

Evaluation of infarct volume by hematoxylin-eosin (H&E)

After euthanization, the brains were removed and fixed in 10% buffered formalin for 24h at 4°C. Brains were sectioned at the optic chiasma and at the infundibular stalk. The resultant blocks of brain between these two cuts were then embedded in paraffin and sectioned into 5 μm-thick coronal slices. Every twentieth slice (for a total of four slices [numbers 1, 21, 41 and 61], which were separated by 100 μm from each other) was stained with H&E. H&E staining allows for the identification of ischemic lesions as well-defined pale areas. The infarct volume was thus measured for these sections as previously described [24],[27]. Lesion volumes were calculated as a percentage of the volume of the contralateral hemisphere using the following formula: % lesion volume = (volume of the contralateral hemisphere – ipsilateral intact volume)/volume of contralateral hemisphere × 100.

Cell death

Cell death was assessed by marking nuclear DNA fragments in situ by immunohistochemistry using the TUNEL method (biotin-dUTP nick end-labeling mediated by terminal deoxynucleotidyl transferase; TdT-FragEL DNA fragmentation detection kit, Oncogene Research Products), following the manufacturer’s instructions, and counterstaining with methyl green. TUNEL-positive cells were counted in all animals in the same predetermined slice of brain located in the central infarct area (slice number 46) using a 40X objective on an optic microscope (Olympus) with analysis software (Image-Pro Plus).

In the ES, the TUNEL-positive cells were counted in the frontal, lateral and piriform areas of the cortex, and in the medial and lateral striatum. The mean of the five counts in each area was calculated. Counts were made both in the embolized and contralateral hemispheres. The results were presented as the total for the embolized hemisphere and then separately for the cortex and striatum [23].

In the pMCAO, we identified cells death in the cortex of both hemispheres based on their nuclear morphology and the dark color [26].

Quantification of biochemical markers

Blood samples were obtained at 3 and 72 hours after ischemia, collected in plastic tubes containing EDTA and then immediately centrifuged. Plasma was frozen at -80°C until analysis.

Glu was determined by HPLC as previously described for humans. The pro-inflammatory cytokines, IL-6 and TNF-α, were determined by ELISA (Inmunogenetics, S.A.U.).

Data analysis

Statistical analyses were performed with the SPSS package 15.0 for Windows (SPSS Inc., Chicago, Illinois, USA). A univariate analysis was performed with the X ² test for dichotomous variables. Continuous variables were tested using the t-test, the Mann-Whitney or Wilcoxon test when appropriate. The Mann-Whitney, and Kruskal-Wallis tests were used to compare the values of physiological parameters, functional evaluation scores, lesion volumes, number of TUNEL positive cells and plasma levels for Glu and inflammatory cytokines between the study groups, and Wilcoxon test for comparison within study groups. The Spearman correlation coefficient was used to analyze the relationship between Glu and cytokines, infarct size, number of TUNEL positive cells and functional evaluation. In patients, the correlations were adjusted by the brain infarct size, dividing the sample into two groups: large infarcts (third tertile) and medium sized infarcts (first and second tertile). P-values less than 0.05 were considered significant.

Biochemical markers

Figure 1 shows the biochemical markers in patients. The cases group tended to present lower IL-6 plasma levels than controls in the first measurement (P = 0.068), and significantly lower levels in the second measurement (P = 0.038). Furthermore, IL-6 plasma levels rose over time in both groups but were significantly only in the cases (P = 0.007) (Figure 1A). TNF-α plasma levels were similar in the cases and control groups in the <12 h and the 72 h measurement. Moreover, TNF-α plasma levels increased over time in the cases and control groups, although did not reached the statistical significance (Figure 1B). Glu plasma levels were similar in the cases and control groups in both the first and the second measurement. Furthermore, the Glu levels tended to decrease in the cases group (P = 0.053) and did not differ over time in the control group (Figure 1C).

Finally, stroke patients were divided according to the timing of the first blood sample: <6 hour (35/51) and 6-12 hours (16/51). When comparing the <6 h and the 6-12 h groups, there were no statistically differences in plasma levels (median [IQR]) of IL-6 (3.3 [11] vs. 4 [9.6]), Glu (118.6 [101] vs. 87 [72]) and TNF-α (5.3 [4] vs. 6 [10.4]) (P NS).

Correlations between biochemical markers, stroke severity, volume of infarct and stroke outcome

In the control group, IL-6 at <12 h tended towards a negative correlation with the 3-month BI score. IL-6 at 72 h also correlated with clinical severity in the control group. TNF-α did not correlate with any of these parameters in the control group. Glu at <12 h tended towards a correlation with clinical severity in all control subjects, while Glu at 72 h correlated negatively with clinical severity in large infarct control group. Moreover, Glu at 72 h correlated positively with clinical outcome measured by the BI (Table 2).

Biochemical markers

The biochemical marker levels in the five groups of rats are shown in Figure 3. The ES had higher IL-6 plasma levels than the pMCAO rats in both the 3 h and 72 h measurements (P = 0.002). Furthermore, the IL-6 plasma levels were higher in the ES than the LC group (P = 0.002). The IL-6 plasma levels were higher in the ES than in its sham at 3 h (P = 0.002) but not at 72 h. There were no differences in any IL-6 plasma level between the pMCAO and its sham. IL-6 levels at 3 h were similar between pMCAO and LC but, at 72 h, were higher in pMCAO (P = 0.009). The IL-6 plasma levels increased significantly over time in the ES (P = 0.046), but not in the pMCAO or LC groups (Figure 3A).

Regarding the TNF-α (Figure 3B), the ES showed higher plasma levels than the pMCAO at 3 h (P = 0.002) and 72 h (P = 0.002). Moreover, TNF-α levels at 3 h and 72 h were higher in ES than in LC (P = 0.026 and P = 0.002, respectively) rats. TNF-α levels were lower in pMCAO than in LC at 3 h (P = 0.002) but not at 72 h. The TNF-α plasma levels were higher for the ES than for its sham at 3 h (P = 0.004) but not at 72 h, and higher for the pMCAO than for its sham both at 3 h (P = 0.015) and at 72 h (P = 0.002). Furthermore, TNF-α increased over time in both ischemic models (P = 0.027 for the ES, and P = 0.028 for the pMCAO) but remained stable in the LC group.

Regarding the Glu (Figure 3C), the ES presented lower plasma levels than the pMCAO at 3 h (P = 0.002) and at 72 h (P = 0.002). There were no differences in Glu levels between the ES or pMCAO and their corresponding shams. Moreover, Glu levels were lower in ES than in LC at 3 h (P = 0.002) and 72 h (P = 0.004), although there were no significant differences between pMCAO and LC. The plasma Glu levels increased over time in ES (P = 0.028), pMCAO and LC (P = 0.028).

Correlations between biochemical markers, functional evaluation, volume of infarct and cell death

On the other hand, the only significant correlation found in the pMCAO was a negative correlation between TNF-α levels at 72h and cell death in the cortex (Table 4).