Neuroprotective effect of neuregulin-1β on spinal cord ischemia reperfusion injury

Abstract

Objective: This study was designed to see if neuregulin-1β (NRG-1β) plays a protective role in spinal cord ischemia and reperfusion injury (SCII).

Design: Animal research.

Setting: China.

Participants: NA.

Interventions: Forty-eight SD rats were randomly divided into control group (n = 16), SCII model group (n = 16) and NRG-1β-treated group (n = 16). In control group, the abdominal aorta was isolated but not clipped. The rats in NRG-1β-treated group were treated with 10μg/kg NRG-1β during developing SCII model.

Outcome Measures: Neurological scores were evaluated. At 3, 6, 12 and 24 h after the reperfusion, rats were killed. Pathological changes of spinal cord were assessed with HE staining, and immunohistochemical staining of matrix metalloproteinases-9 (MMP-9) and tissue inhibitor of metalloproteinase-1 (TIMP-1). MMP-9 and TIMP-1 mRNA levels were assessed using real-time PCR.

Results: NRG-1β reduced the damage of SCII in the rats. The expression of MMP-9 protein and mRNA in NRG-1β treatment group was significantly lower than the model group (P < 0.05) at 6 h, 12 h and 24 h after the perfusion. The expression of TIMP-1 protein and mRNA in the treatment group was significantly higher than the model group at 12 h and 24 h after the perfusion.

Conclusion: NRG-1β reduced the reperfusion damage in rat model of SCII, in which process MMP-9 and TIMP-1 were probably involved.

Introduction

Spinal cord ischemia and reperfusion injury (SCII) is the secondary damage of primary spinal cord injury, which may result from thoracoabdominal aortic and spine operations. SCII can result in delayed neuronal death, neurological dysfunction and immediate or delayed paraplegia.^1,2 However, the pathophysiological mechanism of SCII is still not completely understood, and no effective neuroprotective treatment is available. Several strategies, such as systemic hypothermia and drainage of cerebrospinal fluid, have been implemented to reduce the risk of spinal cord deficits. Unfortunately, none of those therapies is satisfactory. It is still a critical issue to find an effective medicament to prevent and treat SCII.³

Neuregulins (NRG) are a family of four structurally related proteins that are part of the EGF (epidermal growth factor) family of proteins, including NRG-1, NRG-2, NRG-3 and NRG-4. These proteins have been shown to have diverse functions in the development of the nervous system and play multiple essential roles in Schwann cell and oligodendrocyte differentiation, some aspects of neuronal development, as well as the formation of neuromuscular synapses. NRG also plays a role in anti-apoptosis, angiogenesis, inhibiting inflammation, etc. through activation of receptor tyrosine kinases.⁴ NRG is now acknowledged as having promising clinical uses.^5,6 For example, some recent studies showed that NRG could protect the cardiovascular system in patients with myocardial damage.^7,8 Hedhli et al.⁹ reported that NRG/ErbB pathway was involved in cardiocytes viability and angiogenesis. Besides, it is reported that NRG could decrease neuronal apoptosis and relieve cerebellum infarction through improving the microenvironment of neurons in the patients with ischemia/reperfusion injury of brain.¹⁰ NRG also plays a regulating role in the development of the neurological system and regeneration of medullary sheath.¹¹

There are many researches reporting that NRG played a protective role in ischemia/reperfusion injury of heart, brain and liver. As far as we know, it is still unclear about the effect of NRG-1β on SCII. In order to investigate if NRG-1β plays a role in SCII or could reduce the reperfusion injury, we developed rat models of SCII and evaluated the neurological function and histopathological changes. Matrix metalloproteinase (MMP) is calcium-dependent zinc-containing endopeptidase and belongs to a large family of proteases known as the metzincin superfamily. Tissue inhibitor of metalloproteinase (TIMP) can inhibit MMP in a specific way and it plays a critical role in balancing extracellular matrix. MMP-9 and TIMP-1 expression was also investigated.

Methods

Animals

Adult male Sprague–Dawley rats (SPF grade) weighing 200–250 grams were used for the experiments. All rats were purchased from Laboratory Animal Center, Institute of Hygienics and Environment Medicine, Chinese Academy of Military Medical Sciences [SCXK-(military)2014-0001] (Beijing, China). All animal experimental procedures were approved by the Animal Care Committee of University and conducted in accordance with the National Institutes of Health Guide for the Use of Laboratory Animals. Every effort was done to minimize animal’s suffering. The animals were housed in controlled environment at 22–24°C, relative humidity of 50–60% and 12-hour phase shift of the light/dark cycle, and with free access to standard food and water.

Establishment of rat SCII model

Forty-eight rats were randomized into three groups with 16 rats in each group, including control group, SCII group (SCII model group) and NRG-1β-treated group. The rat model of SCII was established as previously described.^11–13 Briefly, after 8h-fasting, the rats underwent the following procedures. The rats were weighed and intraperitoneally injected with 5% chloral hydrate (0.6 ml/100 g), and then were fastened on the operation table. The hair in rat’s abdominal area was removed and the skin was sterilized with anerdian. An incision of 4–5 centimeters was made along the abdomen midline to expose the abdominal cavity. The intestine was wrapped with wet gauze in order to fully expose the abdominal aorta and inferior vena cava. The abdominal aorta and inferior vena cava were gently separated by forceps. In the control group, the abdominal incision was closed, while in the other two groups the abdominal aorta was clipped with an artery clap in the distal part of the left renal artery. The intestines were put back in the abdominal cavity. The abdominal incision was covered with wet gauze. After 30 min ischemia, the artery clamp was removed for blood perfusion. The incision was closed and blood perfusion time was recorded. The rats had free access to food and water. Every four rats were randomly selected and killed for the following experiments at 3, 6, 12 and 24 h after perfusion (four rats for each time point). In NRG-1β group, NRG-1β (10 μg/kg; 396-HB-050, R&D Systems Inc, Minneapolis, USA) was injected into the rat’s caudal vein according to the previous reports^14–16 when artery clamp was removed.

Assessing the rats’ neurological scores

For assessing spinal function after the injury, neurological scores were evaluated¹⁷ according to modified Tarlov score¹⁸ at 3 h after the reperfusion and 3 min before the rats were killed. Specifically, it is rated as score 0 if the rat was in complete paralysis and had no reaction to acupuncture in the hind limbs. It’s rated as score 1 if the rat was in complete paralysis and could reacted to acupuncture in the hind limbs but couldn’t move. It’s rated as score 2 when the rat could move its hind limbs but wasn’t able to stand. It’s rated as score 3 when the rat could stand but wasn’t able to walk. It’s rated as score 4 if the rat could walk unsteadily for several steps. It’s rated as score 5 if the rat was able to walk slowly. It’s rated as score 6 it the rat could walk normally.

HE staining

Every four rats were randomly selected and killed for the following experiments at each time point. The rats in the SCII model group and in NRG-1β-treated group were anesthetized by intraperitoneal injection of 5% chloral hydrate (0.6 ml/100 g) at 3, 6, 12 and 24 h after the reperfusion. The rats were killed by air embolism through injection of 3 ml air into the caudal vein. The rats in control group were anesthetized at 3, 6, 12 and 24 h after the peritoneal cavity was closed and were killed as described above. After the rat heart was exposed, a tube was inserted into the ascending aorta through the ventriculus sinister. The auricula dextra was cut open and administered with 250 ml normal saline (4℃). When the drainage was clear, the spine L1–L6 was removed. The lumbosacral spinal cord was fixed in 4% paraformaldehyde for HE staining and immunohistochemical staining; the other sections were preserved at 80℃ for the experiment of Real-time PCR.

All samples were embedded in paraffin. Serial transversal sections of 4μm were cut and processed for HE staining. Specifically, the sections were dewaxed and rehydrated, following with hematoxylin staining for 3 min, washing with water for 5 min, dipping in hydrochloric acid alcohol for 3 times, washing with water for 5 min, eosin staining for 1 min and washing with water for 5 min. Finally, the sections were dehydrated and sealed with gum. Five fields were randomly selected for each sample, the staining intensity and distribution, gray matter and white matter of spinal cord, neurons appearance, etc. were observed in light microscope (OLYMPUS CX71, Japan) and recorded. Three independent experiments were done in each group. The average value was used for analysis.

Immunohistochemical staining of MMP-9 and TIMP-1

For immunohistochemical staining of MMP-9 and TIMP-1, SABC kit was used. The experiment was performed as the following. The sections were fixed at 70℃ for 2 h, and then were dewaxed and rehydrated, followed with treatment by 3%H₂O₂ and incubation in dark for 30 min. After washing with PBS (5 min × 3 times), the sections were boiled in sodium citrate solution for 15 min. The sections were washed with PBS after cooling to room temperature. The goat serum was added onto the sections prior to incubation at 37℃ for 30 min. The serum was removed and the sections were stained with 1:100 rabbit anti-rat MMP-9 monoclonal antibody or 1:100 rabbit anti-rat TIMP-1 monoclonal antibody (BOSTER Biological Technology co.ltd, Wuhan, China) and were incubated at 37℃ for 2 h. The sections were washed with PBST (5 min × 4 times). As for control, PBS was used to replace monoclonal antibody. The secondary antibodies, Biotinylated goat anti-rabbit IgG (BOSTER Biological Technology co.ltd, Wuhan, China) were used and were incubated at 37℃ for 0.5 h. The sections were washed with PBST (5 min × 4 times). After DAB staining, the sections were washed with water for 5 min. Diaminobenzidine (DAB) Liquid System (ZLI-9017, Golden Bridge international Inc., Beijing) was used to perform the detection according to the manufacturer’s instruction. For the negative control, 1×PBS was replaced for primary antibody. All slides were counterstained with haematoxylin. The images were captured by Zeiss LSM 510 laser scanning confocal microscope (Carl Zeiss Incorporated, Thornwood, NY, USA). For image analysis, the staining-positive cells were counted (200×) in four consecutive sections of each sample were counted and the average values were used for analysis. Three independent experiments were done in each group.

Detection of MMP-9 mRNA and TIMP-1 mRNA level using real-time PCR

Total cellular RNA was isolated by TRIzol LS reagent following the manufacturer’s protocol. The concentration of RNA was determined by spectrophotometric analysis. The primers were designed by using Premier 5.0 software and synthesized by Beijing Sunbiotech co., Ltd (Beijing, China). For internal standards, β-actin was used. For the synthesis of MMP9, amplification fragment of 162 bp was used. The forward primer was 5′-TGGTTCTTCCTTCAAGAATCG-3′, and the reverse primer was 5′-GAGGGATCATCTCGGCTACC-3. For the synthesis of TIMP1, amplification fragment of 249 bp was used. The forward primer was 5′-CCAACCCACCCACAGACAG-3′, and the reverse primer was 5’-GCCCGCGATGAGAAACTC-3. The cDNA synthesis was done according to PrimeScriptTMII 1st Strand cDNA Synthesis Kit (TaKaRa, Japan). The genes were amplified according to the instruction of TaKaRa SYBR^® Premix Ex Taq™II. Specifically, synthesis of first-strand DNA was performed in 20 μl reaction mix (1 μl Oligo Dt Primer (50 μM), 1 μl dNTP (10 mM), 10 μl H₂O and 8 μl RNA) which was incubated at 65℃ for 5 min and then at 0℃ for 2 min. Then the mixture was supplemented with 4ul reverse transcriptase buffer, 0.5 μl RNase Inhibitor and 1 μl PrimeScript reverse transcriptase, and incubated at 42℃ for 1 h. Heat-inactivation (at 70℃) of reverse transcriptase for 15 min was done. Finally, it was preserved at −80℃. Amplification was performed in a total volume of 20ul containing 10ul SYBR Premix Ex Tap II, 1 μl primers, 1 μl cDNA, 0.4 μl ROX Reference Dye and 6.6 μl deionized water. After a common initial denaturing step of 0.5 min at 95°C, the amplification reaction was carried out. Ct value was obtaine. Method 2^–△△Ct was used for data analysis using ABI 7300 Real Time PCR System. Three independent experiments were done in each group.

Statistical analysis

Data are presented as Mean ± SD. Comparisons among groups were performed using one-way analysis of variance. Pairwise comparison was done using SNK-q test. P < 0.05 was used as significant. All analyses were conducted using SPSS software version 19.0 (SPSS, Inc., Chicago, IL, USA).

Results

Neurological score of rats’ hind limbs

In the control group, the rats had normal walking ability. In the SCII model group, the rats began to show dyskinesia at 3 h after the perfusion, and neurological score decreased with time. In NRG-1β-treated group, the neurological score decreased with time, but not to the same extent as the model animals; the neurological score remained significantly higher than that of the model group (P < 0.05; Figure 1).

Figure 1.

Neurological scores of the rats hind limbs in the three groups. *P < 0.05, compared with model group.

Histopathological changes in the groups

In control group, there was a clear boundary between the gray matter and the white matter of the spinal cord, with neurons in polygon appearance, and nucleus and nucleolus with clear outline. In the model group, at 3 h after the perfusion, there was no significant difference in the pathological changes of spinal marrow between the model group and the control group. At 6 h after the perfusion, some neurons shrank with slight vacuolus clearance; nucleolus was vague in appearance; the boundary between cytoplasm and nucleus was blurred; slightly loosened tissue organization and infiltration of neutrophile granulocyte was observed. At 12 h after the perfusion, neurons obviously shrank with vacuolus clearance; nucleus was with deep staining and vague in appearance; obvious loosened tissue organization and more infiltration of neutrophile granulocyte was observed. At 24 h after the perfusion, the histopathological changes worsen, and a lot of neurons showed degenerative necrosis with loose and disordered tissue organization. In NRG-1β-treated group, the pathological changes began showing difference from the control group at 6 h after the perfusion, and worsen with time but still better than the model group (Figure 2).

Figure 2.

HE staining of spinal cord tissue (×200). Three independent experiments were done in each group, and five fields were randomly selected for each sample. A, control group; B–D, model group at 6, 12 and 24 h after the reperfusion. a–d, NRG-1β-treated group at 3, 6, 12 and 24 h after the reperfusion. The short arrows indicate normal neurons. The thick arrows indicate neurons death or apoptosis. The thin arrows indicate loosened tissue organization.

Immunohistochemical changes of MMP-9 and TIMP-1 in the three groups

There were no positive MMP-9-staining or positive TIMP-1-staining cells in control group. In model group, at 3 h after the perfusion, MMP-9-staining positive cells with light brownish-orange cytoplasm were observed in the spinal anterior horn and the surroundings of central canal. The amount of staining-positive cells increased with time. In NRG-1β-treated group, there was no statistical difference in the amount of MMP-9-staining positive cells between NRG-1β-treated group and the model group at 3 h after the perfusion. At 6, 12 and 24 h after the perfusion, MMP-9-staining positive cells were significantly less than those of the model group. As for TIMP-1-staining, at 3 and 6 h after the perfusion, the positive staining cells were neuron cells and vascular endothelial cells; at 12 h after the perfusion, the positive staining cells were neuron cells and neuroglial cells. In NRG-1β-treated group, TIMP-1-staining positive cells showed a similar trend to that of the model group, and were significantly more than those of the model group at 12 and 24 h after perfusion (P < 0.05; Figures 3–4, Table 1).

Figure 3.

Immunohistochemical staining of MMP-9 in model group and NRG-1β-treated group (×400). Three independent experiments were done in each group, and five fields were randomly selected for each sample. A1–C1, model group at 6, 12 and 24 h after the reperfusion. a1–c1, NRG-1β-treated group at 6, 12 and 24 h after the reperfusion. The staining-positive neurons are indicated by arrows.

Figure 4.

Immunohistochemical staining of TIMP-1 in model group and NRG-1β-treated group (×400). Three independent experiments were done in each group, and five fields were randomly selected for each sample. A1–C1, model group at 6, 12 and 24 h after the reperfusion. a1–c1, NRG-1β-treated group at 6, 12 and 24 h after the reperfusion. The staining-positive neurons are indicated by arrows.

Table 1. Immunohistochemical changes of MMP-9 and TIMP-1 in the three groups (n = 4, $\overline{x}\pm s$).

Group	Indicator	Time post-reperfusion
		3 h	6 h	12 h	24 h
Control group	MMP-9	0.0	0.0	0.0	0.0
	TIMP-1	0.0	0.0	0.0	0.0
SCII model group	MMP-9	9.0 ± 1.6Δ	23.8 ± 1.7Δ	28.3 ± 1.5Δ	34.8 ± 2.6Δ
	TIMP-1	11.8 ± 0.9Δ	12.3 ± 1.5Δ	7.8 ± 0.9Δ	7.8 ± 1.5Δ
NRG-1β-treated group	MMP-9	8.5 ± 0.6	17.8 ± 0.9*	20.8 ± 3.5*	30.0 ± 2.2*
	TIMP-1	12.3 ± 0.9	13.8 ± 0.9	10.5 ± 1.7*	10.3 ± 0.9*

SCII: spinal cord ischemia reperfusion injury; NRG-1β: neuregulin-1β; MMP-9: matrix metalloproteinase-9; TIMP-1: tissue inhibitor of metalloproteinase-1. ^Δ indicates P < 0.05 compared with control group; * indicates P < 0.05 compared with model group.

mRNA levels of MMP-9 and TIMP-1 in the three groups

In the control group, there’s little expression of MMP-9 and TIMP-1. In model group, the expression of MMP-9 increased with time and was significantly higher than that of control group at 6, 12 and 24 h after the perfusion. As for TIMP-1 mRNA, it slightly increased and then decreased at 12 and 24 h after the perfusion. In NRG-1β-treated group, MMP-9 mRNA level was lower than that of the model group; the expression level of TIMP-1 showed no statistical difference from that of the model group (Figure 5).

Figure 5.

Levels of MMP-9 mRNA and TIMP-1 mRNA. *P < 0.05, compared with model group. Three independent experiments were done in each group.

Discussion

As the secondary damage of primary spinal cord injury, SCII contributes to delayed neurological dysfunction and even paralysis. It is a critical issue to find effective therapies for SCII. In this study, we mainly found that NRG-1β improved neurological scores and histopathological results in the rat models of SCII. Generally, NRG-1β decreased MMP-9 expression and increased TIMP-1 expression level in the SCII rats.

In the present study, 10 μg/kg NRG-1β was administered into the rat models of SCII in order to investigate the role of NRG-1β in SCII. After the model was built, Tarlov score¹⁸ was used to assess the rats’ neurological function. The rats’ neurological score of the model group was lower than control group, and decreased with time. The neurological score in NRG-1β-treated group was significantly higher than that of the model group. Histopathological examination also showed that NRG-1β reduced pathological changes. It indicated that NRG-1β could reduce spinal cord ischemia reperfusion injury in rat model of SCII. Those results were in accordance with the previous study on rat models of brain ischemia and reperfusion.^13,19 As reported, NRG-1 on axons of central nerve system neurons could interact with its receptor, ErbB4, and promote the myelination of that axon.²⁰ NRG-1 promoted myelination and was decreased in schizophrenic patient.

In the present study, MMP-9 and TIMP-1 expression levels were assessed by immunohistochemical staining and q-PCR to see if NRG-1β played a protective role in SCII. The expression of MMP-9 protein and mRNA in NRG-1β treatment group was significantly lower compared with the model group (P < 0.05) at 6, 12 and 24 h after the perfusion. This was in accordance to the experiment results of de Castro et al.²¹ As the specific inhibitor of MMP-9, TIMP-1 slightly increased 3 and 6 h after the perfusion and then decreased at 12 and 24 h after the perfusion. We thought that during the first 6 h, the negative feedback of inflammation reaction caused the recruitment of TIMP-1 in the injured spinal cord; but the increasing SCII might affect the secretory cells, which resulted in TIMP-1 decrease. After NRG-1β treatment, MMP-9 decreased while TIMP-1-staining positive cells significantly increased at 12 and 24 h after perfusion (P < 0.05).

Although there are many reports concerning the role of MMP-9 in nerve injury, the specific mechanism is unclear. For example, Wu et al.²² proved that MMP was involved in many pathological processes in post-central nervous injury but the molecular mechanism was not elucidated. Duchossoy et al.²³ investigated whether scar formation could be associated with the activation of MMP-2 and MMP-9 which were found to be transiently upregulated in the spinal cord wound. In the condition of acute spinal cord injury, ischemia and anoxia could act as a stimulator to inflammatory cell, neurons, endothelial cells and glial cell and further increased MMP expression. Fan et al.²⁴ studied the role of MMP-9 in apoptosis in rat model of SCII, and found that MMP-9 expression and cell apoptosis increased in the rat models. Besides, there was a positive relationship between MMP-9 level and apoptosis (r = 0.936, P < 0.05). It suggested that MMP-9 was involved in secondary damage of spinal cord injury and induced neuronal apoptosis. We should do further work to investigate the specific role of MMP-9 and TIMP-1 in the SCII, and to elucidate the molecular mechanism of how NRG-1β regulates MMP-9 and TIMP-1 expression.

Disclaimer statements

Contributors None

Funding This research was funded by the Science & Technology Fund of Tianjin Municipal Commission of Health and Family Planning (KZ025).

Conflict of interest The authors declared that they have no conflict of interests.

Acknowledgements None