Insulin-like growth factor-1 secreted by brain microvascular endothelial cells attenuates neuron injury upon ischemia

Authors


Abstract

Insulin-like growth factor (IGF)-1 is essential for the development of the nervous system, and is present in many cell types. Relatively little is known about IGF-1 expression in brain microvascular endothelial cells (BMECs). For in vivo studies, we examined the expression of IGF-1 and insulin-like growth factor-binding protein (IGFBP)-2 after focal cerebral ischemia for 12 h, 24 h, 3 days and 7 days, utilizing a permanent middle cerebral artery occlusion (MCAO) model in rats. For in vitro studies, we examined the levels of IGF-1 and IGFBP-2 in the culture medium or primary culture of BMECs injured by oxygen–glucose deprivation (OGD). Then, we elucidated the protective effects of IGF-1 on cortical neurons injured by OGD and the possible mechanism. In addition, we investigated the effect of BMEC-conditioned medium on IGF-1 receptor expression in neurons. The results showed that IGF-1 expression increased in serum and brain tissue, whereas IGFBP-2 expression decreased in brain tissue of MCAO-injured rats. In primary culture of BMECs, the expression levels of IGF-1 and IGFBP-2 were significantly higher under OGD conditions in culture. IGF-1 administration improved neuron viability upon normoxia or OGD, and upregulated p-Akt expression. This effect was reversed by LY294002, a specific inhibitor of the phosphoinositide 3-kinase–Akt signaling pathway. Furthermore, conditioned medium from OGD-treated BMECs substantially suppressed neuron viability and the expression of IGF-1 receptor simultaneously. These data demonstrate that therapeutic strategies that prioritize the early recovery of the IGF-1 system in BMECs might be promising in ischemic injury.

Abbreviations
BBB

blood–brain barrier

BMEC

brain microvascular endothelial cell

BMEC-CM

brain microvascular endothelial cell-conditioned medium

ECA

external carotid artery

I-CM

ischemia-conditioned medium

IGF

insulin-like growth factor

IGF-1R

insulin-like growth factor-1 receptor

IGFBP

insulin-like growth factor-binding protein

MCAO

middle cerebral artery occlusion

N-CM

normal conditioned medium

OGD

oxygen–glucose deprivation

PI3K

phosphoinositide 3-kinase

TB

Trypan Blue

Introduction

Ischemic stroke is a major cause of morbidity and mortality in humans, and is caused by occlusion of cerebral blood vessels or hemorrhage. The blood–brain barrier (BBB) is a dynamic interface between blood vessels and brain compartments that can efficiently protect nerves from injury. In this functional unit, brain microvascular endothelial cells (BMECs), astrocytes and extracellular matrix constitute the microenvironment that plays an important role in the survival and function of neurons. Emerging data have shown that the brain injury seriously impacts on the cellular network cascade which decides the fate of injury [1]. Consequently, the concept of the BBB has evolved from one of a discrete barrier to one whereby the BBB is best viewed as a modulatory interface between blood and brain [2]. In this context, regulation of this interface via paracrine interactions occurs between the capillary endothelium and its neighboring cells [3].

Insulin-like growth factor (IGF)-1, a circulating hormone generated not only in the liver, regulated by growth hormone, but also locally in many cell types, such as neurons and glia [4, 5], is essential for the development of the nervous system, hippocampal neurogenesis [6, 7], and neurotransmission [8, 9]. IGF-1 regulates cell growth, differentiation and survival of neuronal cell lines, and is among the most potent antiapoptotic growth factors present in eukaryotic cells [10]. IGF-1 is neuroprotective in animal models of stroke, and high serum IGF-1 levels just after ischemic stroke onset are associated with neurological recovery and a better functional outcome, suggesting that the endogenous IGF-1 level impacts on the evolution of cerebral infarction [11, 12].

We have previously reported that the paracrine signaling of BMECs plays different roles in the survival of neurons under normal and injury conditions [2]. In our microarray study comparing the differential gene expression between normal and ischemic BMECs, the mRNA level of IGF-1 was found to change after exposure to oxygen–glucose deprivation (OGD) (unpublished observation). These results indicated that IGF-1 is expressed on BMECs and may be involved in cerebral ischemic injury.

To further examine the involvement of IGF-1 in ischemic injury, we first measured the expression levels of IGF-1 and insulin-like growth factor-binding protein (IGFBP)-2 in cerebral ischemia, utilizing a rat model of middle cerebral artery occlusion (MCAO) in vivo and ELISA and immunohistochemistry. Then, we examined IGF-1 and IGFBP-2 expression and release from primary BMECs in response to OGD in vitro. A second series of experiments were designed to study the relationship between BMECs and neurons through administration of IGF-1 released from normal or OGD-injured BMECs.

Results

IGF-1 and IGFBP-2 levels in ischemic brain tissue and serum

The results concerning the changes in the levels of IGF-1 and IGFBP-2 in ischemic brain tissue and serum are summarized in Fig. 1A–D. In the serum, the levels of IGF-1 and IGFBP-2 were less than 200 ng/mL and 400 pg/mL, respectively, in control rats; these levels were significantly increased, and reached their peak values at 3 days and 24 h, respectively, after MCAO. In brain tissue, the IGF-1 level in the ipsilateral cortex after 7 days was significantly lower than in controls. Obviously, the IGFBP-2 levels on both the ipsilateral and contralateral sides were decreased as compared with controls. The immunochemical staining in Fig. 1E shows that IGF-1 expression was increased after occlusion for 1 day or 3 days, whereas IGFBP-2 expression was attenuated to a different degree in MCAO-injured brain tissue.

Figure 1.

(A–D) Changes in IGF-1 and IGFBP-2 levels in serum and brain tissue of the MCAO rat model at the different time points, detected by ELISA. The levels of IGF-1 and IGFBP-2 were significantly increased, and reached their peaks at 3 days and 24 h, respectively, in the serum. The IGF-1 level in the ipsilateral cortex decreased at 7 days. The IGFBP-2 levels in both the ipsilateral and contralateral sides were decreased as compared with controls. (E) Immunohistochemical analysis of IGF-1 and IGFBP-2 in rat brain. Sections of brain tissue were incubated with antibodies against IGF-1 and IGFBP-2 (brown). The immunochemical staining showed that IGF-1 expression was increased after occlusion for 1 day or 3 days, whereas IGFBP-2 expression was reduced to different degrees in MCAO-injured brain tissue. *P < 0.05 versus the control group; **P < 0.01 versus the control group.

IGF-1 and IGFBP-2 levels in BMEC-conditioned medium (BMEC-CM) and BMECs

We assayed IGF-1 and IGFBP-2 levels in normal and OGD BMEC-CM by ELISA. There was no remarkable change in the release of IGF-1 and IGFBP-2 between normal and ischemic BMEC-CM (Fig. 2A). Immunocytochemistry showed an increase in the immunoreactivity of IGF-1 and IGFBP-2 in BMECs of the OGD group as compared with the control group (Fig. 2B). Western blot showed obvious increases in IGF-1 and IGFBP-2 levels in the ischemic group (P < 0.05; Fig. 2C), which were consistent with the findings of the immunofluorescence assay, suggesting that the expression levels of IGF-1 and IGFBP-2 were significantly increased in the OGD-injured BMECs.

Figure 2.

(A) Levels of IGF-1 and IGFBP-2 in cultured media of normal and OGD-induced BMECs determined with by ELISA. There was no substantial change in the release of IGF-1 and IGFBP-2 between normal and ischemic BMEC-CM. (B) Immunofluorescence was used to detect the expression of IGF-1 (green) and IGFBP-2 (red) of normal and OGD-treated BMECs. There was an increase in the immunoreactivity of IGF-1 and IGFBP-2 in the BMECs of the OGD group as compared with the control group. (C) Western blot of IGF-1 and IGFBP-2 expression. β-Actin was used as a loading control. The bands in the ischemic group showed a clear increase in terms of IGF-1 and IGFBP-2. *P < 0.05 versus the control group. I, OGD BMECs; N, normally cultured BMECs.

Neuronal cell viability and p-Akt expression

As shown in Fig. 3A, IGF-1 increased the cell viability not only of normal cultured BMECs, but also of OGD-injured neurons (P < 0.01), indicating that the actions of IGF-1 are positive. The basal cell viability of neurons was not affected significantly by LY294002 or PD98059 in comparison with the control group, whereas the increased cell viability of OGD-injured neuron caused by IGF-1 was clearly inhibited by LY294002, but was not influenced by PD98059 (Fig. 3A). Figure 3B shows that IGF-1 could activate p-Akt in both normally cultured and OGD-injured neurons. LY294002, but not PD98059, significantly suppressed p-Akt expression in OGD-injured neurons. Total Akt expression in all groups showed no evident differences. These results suggested that the instructive effect of IGF-1 on the viability of OGD-injured neurons was conducted through the Akt pathway but not through the extracellular signal-related kinase pathway.

Figure 3.

(A) Cell viability of primary cortical neurons under normal conditions or OGD with different treatments, determined with the 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl-tetrazolium bromide assay. LY294002 is an Akt inhibitor, and PD98059 is an extracellular signal-related kinase inhibitor. IGF-1 increased the cell viability not only of normally cultured BMECs, but also of OGD-injured neurons. The basal cell viability of neurons was not affected significantly by LY294002 or PD98059. The increased cell viability of OGD-injured neurons caused by IGF-1 was significantly inhibited in the presence of LY294002, but was not influenced by PD98059. (B) Expression of p-Akt on neurons under different conditions, determined with western blotting. Akt served as a loading control. The experiment was repeated three times, with similar results. IGF-1 activated p-Akt in both normally cultured and OGD-injured neurons. LY294002, but not PD98059, significantly suppressed p-Akt expression in OGD-injured neurons. Total Akt expression in all groups showed no evident difference. *P < 0.05 versus the control group; **P < 0.01 versus the control group; ##P < 0.01 versus the OGD group.

Apoptosis of neurons and IGF-1 receptor (IGF-1R) expression

For normally cultured neurons, OGD treatment increased the apoptosis rate, whereas normal conditioned medium (N-CM) significantly decreased the apoptosis rate of OGD-damaged neurons (P < 0.05; Fig. 4A). Ischemia-conditioned medium (I-CM) clearly increased the apoptosis rates of both normal and OGD-injured neurons, indicating that I-CM aggravated the injured neuron, as characterized by a further increase in apoptosis rate. In addition, I-CM suppressed IGF-1R expression in both normal and OGD-injured neurons markedly as compared with the effect of N-CM. These results demonstrated that N-CM was therapeutic, whereas I-CM aggravated OGD-induced neuronal injury, which is related to IGF-1R expression on the neurons (Fig. 4B).

Figure 4.

(A) Changes in the neuronal apoptosis rate after different treatments, determined with flow cytometry. OGD treatment increased the apoptosis rate, whereas N-CM significantly reduced the apoptosis rate of OGD-damaged neurons. I-CM clearly increased the apoptosis rate of normal neurons and damaged ischemic neurons substantially. (B) Expression of IGF-1R in normal or OGD-injured neurons incubated with N-CM or I-CM was measured by western blotting. β-Actin served as a loading control. I-CM suppressed IGF-1R expression in both normal and OGD-injured neurons markedly as compared with the effect of N-CM. **P < 0.01 versus basal cultured conditioned media (B-CM) in the control group; #P < 0.05 versus B-CM in the OGD group; ▲P < 0.05 versus N-CM in the OGD group; ▲▲P < 0.01 versus N-CM in the OGD group; $$P < 0.01 versus N-CM in the normal group.

Discussion

IGF-1 is an endocrine and autocrine/paracrine growth factor that is the primary mediator of the effect of growth hormone [13]. Low levels of IGF-1 are associated with poor outcome in elderly patients with stroke, suggesting that the endogenous IGF-1 level impacts on the evolution of cerebral infarction [12]. IGF-1 has been reported to protect neurons from both oxidative stress and apoptosis induced by various stimuli [14-26]. Administration of IGF-1 reduces neuronal loss and infarct volume, and increases glial proliferation in the brain, following experimental cerebral ischemia [19-27]. In addition to having neuroprotective activities, IGF-1 is also a potent angiogenic factor required for cerebral angiogenesis during development and adulthood [20]. IGFBP-2 is a member of the IGF family, and binds and negatively modulates the activity of IGF-1 and IGF-2 [21]. IGF-1 has been shown to protect neural cells against acute ischemic injury, and IGF-1 has been discussed as a vascular protective factor [22]; a role for IGF-1 in BMECs after ischemic injury has not been fully demonstrated.

To address this problem, we examined the immunohistochemical expression in the ischemic brain and the serum concentrations of IGF-1 and IGFBP-2 following MCAO in rats. The IGF-1 level in the serum peaked at 3 days. We then examined IGF-1 levels in both the ipsilateral and contralateral sides of the brain, and found that the peak level of IGF-1 in the brain appeared at 24 h after MCAO, which was consistent with the results of immunohistochemistry, characterized by large amounts of dark brown staining. These results indicated that the levels of IGF-1 in the brain are not always correlated with serum levels [23], although studies have indicated that IGF-1 can cross the BBB [24-36]. Emerging data show that increased expression of IGF-1, together with that of IGFBP-2 and IGFBP-3, occurs in response to unilateral hypoxic ischemic injury in the infant rat [27, 28] or to focal ischemia in adult rats [29]. We found that the level of IGFBP-2 in the serum peaked at 24 h after MCAO. In brain tissue, the IGFBP-2 levels on both the ipsilateral and contralateral sides decreased remarkably during different durations of hypoxia. In agreement with the ELISA results, we also found, by immunohistochemistry, that IGFBP-2 expression was suppressed, with reduced positive staining. A reason that may have contributed to our detection of no consistent correlation between the levels of IGFBP-2 in serum and brain tissue is as follows: although IGFBP-2 secretion increases in the early phase of hypoxia, more IGFBP-2 is transported out of brain tissue to bind IGF-1, owing to the incremental requirement for IGF-1 of OGD-injured brain tissue.

IGF-1 is known to be expressed in various cell types in the central nervous system, such as microglia [30], neurons [31], and glia [32], whereas it is not clear whether IGF-1 can be detected in BMECs. To this end, we investigated IGF-1 and IGFBP-2 expression in BMECs. Both immunofluorescence and western blotting implicated BMECs as a major source of IGF-1 and IGFBP-2 in the brain. IGF-1 and IGFBP-2 expression levels were significantly increased after cerebral ischemia injury. The result for IGFBP-2 obtained in vivo is not in agreement with that obtained in vitro, which might indicate a stress-induced increase after OGD treatment for 6 h. If we prolonged the treatment time, the effect of OGD might be consistent with the in vivo result. ELISA showed no significant difference between normal and OGD-injured BMECs, which was not consistent with the immunological results. These data indicate that secretion of IGF-1 and IGFBP-2 is not synchronous with their expression. The increased IGFBP-2 level under OGD conditions might bind IGF-1, so as to cause the decline in IGF-1 secretion, resulting in no significant increase in the level in culture medium.

As IGF-1 has been shown to have definite neuroprotective effects, studies on the mechanism have been carried out in recent years. The phosphoinositide 3-kinase (PI3K)–Akt pathway is thought to be the most important way by which IGF-1 protects cells from apoptosis [33, 34]. The IGF-1-mediated antiapoptotic function is closely related to several signaling pathways, including mitogen-activated protein kinase and glycogen synthase kinase-3β [35-37]. Our findings verified that, under OGD, IGF-1 had a definite protective effect on primary cortical neurons. The viability of neurons in the presence of IGF-1 was obviously improved as compared with the OGD group, whereas the PI3K blocker significantly inhibited the neuroprotection afforded by IGF-1, indicating that the PI3K pathway played an important role in mediating this neuroprotection. The effect of IGF-1 in promoting neuron viability was not inhibited by a blocker of the extracellular signal-related kinase pathway, PD98059. The above results showed that the PI3K–Akt pathway was important for IGF-1 to protect nerves.

Akt is an important serine/threonine protein kinase of the PI3K pathway, and activated Akt can cause cell apoptosis by phosphorylating proteins such as BAD and FKHRL [38]. Our western blot findings demonstrated that p-Akt expression in OGD-injured neurons was upregulated in the presence of IGF-1, and finally improved cell viability. On the other hand, the increased p-Akt expression was clearly inhibited by LY294002, which was in agreement with the cell viability findings. However, whether PD980159 was added or not added to the OGD-injured neurons treated with IGF-1 had no influence on p-Akt expression, which was also in agreement with the cell viability findings.

Changes in the apoptosis rate showed that normal BMEC-CM improved the survival of both normally cultured and OGD-injured neurons, whereas ischemic BMEC-CM was harmful for neuron survival. IGF-1R is one of four transmembrane receptors [IGF-1R, IGF-2 receptor (IGF-IIR), IR, and hybrid receptors of IGF and IR] that constitute the IGF-1R signaling system in addition to the three circulating ligands (IGF-I, IGF-II, and insulin) and multiple regulatory IGF-binding proteins (IGFBP-1 to IGFBP-6) [39-41]. IGF-1 exerts its biological action through binding to its receptor, IGF-1R. Owing to the suppression of IGF-1R in both normally cultured and OGD-injured neurons by I-CM, IGF-1 does not have neuroprotective and antiapoptotic functions. In addition to the IGF-1 system, there are common growth factors and receptors between nerves and blood vessels, including vascular endothelial growth factor, angiopoietin, glial-derived neurotrophic factor, and basic fibroblast growth factor [42]. These factors are secreted from BMECs, and the corresponding receptors are located on the surfaces of nerve cells, inducing neurogenesis, migration, and axon growth. These might partly explain the possible reason why BMEC exhibits protective effects on neurons. Further study is necessary to determine whether other factors are involved and what changes occur during the injury.

In conclusion, this study provides clear evidence of IGF-1 and IGFBP-2 changes after MCAO injury in vivo, and demonstrates the expression of IGF-1 and IGFBP-2 in OGD-injured BMECs in vitro. We also show that PI3K signaling pathway may play a key role in the neuroprotective and antiapoptotic effects of IGF-1.Therefore, agents that block IGF–IGFBP interaction, and enhance access to IGF-IR, are promising targets for therapeutic intervention against brain ischemic injury.

Experimental procedures

Animals

All experiments were carried out with the approval of the Animal Care and Use Committee of Beijing University of Chinese Medicine. All randomly assigned male Sprague-Dawley rats, purchased from Beijing Vital River Laboratory Animals (Beijing, China), were 8–10 weeks old. Rats were housed in a pathogen-free facility with 12-h light/dark cycles and continuous access to food and water. Rats were acclimatized for at least 4 days before any experiment, and were fasted overnight before any procedure.

MCAO

The MCAO model was established exactly as described previously [43]. Briefly, rats were anesthetized with 10% chloral hydrate (350 mg/kg body weight, intraperitoneal). The left common carotid artery was exposed through a midline incision, and was carefully dissected free from surrounding nerves and fascia. The occipital artery branches of the external carotid artery (ECA) were then isolated, and the occipital artery and superior thyroid artery branches of the ECA were coagulated. The ECA was dissected further distally. The internal carotid artery was isolated and carefully separated from the adjacent vagus nerve. A 0.25-mm-diameter fishing suture was coated with poly(l-lysine), and its tip was rounded by heating near a flame. The filament was inserted through the ECA into the internal carotid artery and then into the circle of Willis to occlude the origin of the left middle cerebral artery. Rats in the sham group underwent similar surgery as that in the model group, but without suture insertion. The core body temperature was maintained between 36.5 °C and 37.5 °C during and after the procedures. Ten to 12 rats were included in each group.

Measurement of IGF-1 and IGFBP-2 in brain tissue and serum

Twelve hours, 1 day, 3 days and 7 days after occlusion, the rats were anesthetized with chloral hydrate, and the blood was collected through the abdominal aorta. The samples were centrifuged at 3500 g for 15 min, and the serum was collected for the measurement of IGF-1 and IGFBP-2. The striata and cortexes of the left hemisphere were removed, and the contralateral and ipsilateral tissues were homogenized with 1 × RIPA buffer plus proteinase inhibitors on ice, and centrifuged at 14 000 g for 15 min at 4 °C. The supernatant was taken for the measurement of IGF-1, and protein quantification was conducted with the bicinchoninic acid method. ELISA steps were performed according to the protocol provided with the ELISA kit (R&D Systems). The IGF-1 concentration of each group was calculated by use of a standard curve. Five to six rats were used in each group, and the experiment was repeated three times.

Immunohistochemistry

The rats were killed and the brains were removed for immunohistochemistry. Briefly, paraffin-embedded sections were deparaffinized in Histoclear (National Diagnostics, Atlanta, GA, USA), rehydrated, and washed with 0.01 m NaCl/Pi. Endogenous peroxidase was inhibited with 3% hydrogen peroxide in methanol at room temperature for 30 min. After being blocked in 10% goat serum for 30 min, sections were incubated with antibodies against IGF-1 (1 : 500; Santa Cruz), IGF-1R (1 : 500; Santa Cruz) and IGFBP-2 (1 : 500; Santa Cruz) at 4 °C overnight. Following washing in 0.01 m NaCl/Pi, the sections were incubated with horseradish peroxidase-conjugated secondary antibody at 37 °C for 2 h. Peroxidase was revealed by dipping the sections in a mixture containing 0.05% 3,3′-diaminobenzidine for 5 min. The sections were then counterstained with hematoxylin, coverslipped, and observed. Alternative sections of the same brain were processed in the same way, but without primary antibody, and used as controls.

Primary BMECs and neuronal cultures

Primary rat BMEC isolation and culture were performed according to a method previously described [44]. Briefly, the isolated cerebral gray matter was homogenized and then passed through 200-μm and 74-μm mesh. The material left on the 74-μm mesh screen was digested with type II collagenase (0.2%). The microvessel fragments were collected, resuspended in complete culture medium, and seeded in gelatin-coated Petri flasks. Cells from the third passage were used for the subsequent experiments. The BMECs were identified by immunofluorescence labeling with antibody against factor VIII-related antigen. The rate of BMEC factor VIII-related antigen positivity was > 99%.

Cortical neurons were isolated from 17-day embryos of pregnant mice. The cortex was cut into pieces, and tissue fragments were treated with 0.125% trypsin at 37 °C for 20 min. The digestion was stopped with DMEM supplemented with 10% fetal bovine serum, and this was followed by pipetting and passaging through 74-μm mesh. The cells were resuspended at a density of 1 × 106 cells/mL in serum-free, high-glucose Neurobasal medium (Invitrogen) supplemented with B27 and Hepes (Sigma-Aldrich), and then cultured on poly(d-lysine)-coated 24-well dishes. Cells were maintained in growth medium at 37 °C in a 95% air/5% CO2 humidified atmosphere. Half of the initial culture medium was removed at day 4 and replaced with fresh medium.

OGD establishment and preparation of different BMEC-CMs

The BMECs were cultured in OGD conditions for 6 h to mimic cerebral ischemia in vitro, as previously described [45]. Briefly, the BMEC complete culture medium was changed to DMEM (glucose-free), and the culture flask was placed in a sealed tank with a continuous low-flow 93% N2 and 7% CO2 mixture for 6 h. When the third passage of BMECs grew to 90% confluence, serum-free culture medium was used to collect the BMEC-CM. The BMEC complete medium was replaced with serum-free medium, and harvested as conditioned medium from the normal cultured BMECs. The N-CM and the I-CM were collected and centrifuged at 1000 g for 10 min to remove the cell debris.

Measurement of IGF-1 and IGFBP-2 in BMEC-CM and BMECs

A 1-mL BMEC suspension was seeded into 24-well plates at a concentration of 1 × 105 cells/mL. IGF-1 and IGFBP-2 in N-CM and I-CM were detected by ELISA. ELISA steps were performed according to the protocol provided with the ELISA kit. The contents of IGF-1 and IGFBP-2 in each group were calculated by use of a standard curve. The experiment was repeated in triplicate, and six wells were used for each group.

A 2-mL BMEC suspension was seeded into six-well plates at a concentration of 1 × 106 cells/mL. BMECs were divided into two groups: normal cultured BMECs and OGD BMECs. The cells were scraped in lysis buffer. The insoluble material was removed by centrifugation at 12 000 g for 20 min. The protein content was measured with the bicinchoninic acid method, and 50 μg of protein was then separated by SDS/PAGE in a 12.5% polyacrylamide gel and transferred to a 0.45-μm nitrocellulose membrane. Nonspecific binding sites were blocked with NaCl/Tris (40 mm Tris, pH 7.6, 300 mm NaCl) containing 5% nonfat dry milk for 1 h at 37 °C. The membrane was then incubated with primary antibodies against IGF-1 (1 : 1000; Santa Cruz), IGFBP-2 (1 : 1000; Santa Cruz), and β-actin (1 : 3000; Santa Cruz). The membranes were then incubated with the corresponding horseradish peroxidase-conjugated secondary antibodies (1 : 5000). Immunoreactive proteins were detected by enhanced chemiluminescence according to the instructions of the manufacturer (Pierce, Rockford, IL, USA).

Next, immunocytochemistry was used to show the expression of IGF-1 and IGFBP-2. BMECs (1 × 104 cells per well, 1 mL per well) were seeded into gelatin-coated 24-well plates with coverslips. BMECs were washed with NaCl/Pi, and then fixed with 4% formaldehyde for 10 min at room temperature. BMECs were incubated with 0.1% Triton X-100 for 10 min, and then incubated with 3% H2O2 at room temperature. Five per cent BSA in NaCl/Pi was used to block nonspecific binding, and BMECs were incubated overnight at 4 °C with primary antibodies against IGF-1 and IGFBP-2 (1 : 100), and then with the fluorescence-conjugated secondary antibody (1 : 500). The slices were visualized by fluorescence microscopy.

Neuronal viability and p-Akt detection after IGF-1 signaling pathway-blocking treatment

A 100-μL BMEC suspension was seeded into 96-well plates at a concentration of 1 × 104 cells/mL. Then, 10 μm LY294002 (PI3K channel blocker) and 10 μm PD98059 (MEK1 channel blocker) were used in the presence or absence of 100 ng/mL IGF-1 for the control and OGD-treated neurons. Cell viability was then evaluated with the 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl-tetrazolium bromide assay. Trypan Blue (TB) staining was also performed. After each treatment, cells were stained with 0.4% TB (Sigma-Aldrich). Unstained cells were regarded as viable, and stained cells were regarded as dead. The total cell number and the number of TB-positive cells were counted under a light microscope with a magnification of × 200 by an independent, blinded investigator. Survival value was calculated by using the following formula: detected unstained cells/detected whole cells shortly before treatment [46]. Experiments were repeated with at least three separate batches of culture. Western blot analyses were performed to quantify p-Akt expression in neurons according to standard protocols, as in the western blotting process described above. Antibodies against anti-p-Akt (1 : 500; Cell Signaling) and Akt (1 : 1000; Cell Signaling) were used. The experiment was repeated in triplicate.

Apoptosis and IGF-1R expression on neurons exposed to different BMEC-CMs

Serum-free primary neuronal cultures (normal culture or OGD-treated) were subjected to 6 h of incubation with the collected N-CM or I-CM. Our former work showed that 100% BMEC-CM was the appropriate concentration for the following experiment. Three wells were used for each group. After they had been incubated for 6 h, the cells were resuspended in propidium iodide staining solution with RNaseA, and incubated for 30 min at 37 °C. The apoptosis rate was then determined with flow cytometry. Western blotting and immunofluorescence were conducted as described above.

Statistical analysis

All results are summarized as mean ± standard deviation. spss (SPSS) and Excel (Microsoft) software were used for further data and statistical analysis. One-way analysis of variance followed by the Student–Newman–Keuls multiple-range test or post hoc tests was used to determine statistically significant differences among the groups. A P-value of < 0.05 was considered to be statistically significant.

Acknowledgements

This research was supported by grants from the National Natural Science Foundation of China (Grant No. 81102679), the Fundamental Research Funds for the Central public welfare research institutes (ZZ070824), the National Science & Technology Major Project of China (No. 2009ZX09502-014), and the Ministry of Education and Key Laboratory of Chinese Internal Medicine.

Ancillary