Jiang et al.

BMC Neurosci (2020) 21:18

https://doi.org/10.1186/s12868-020-00565-5 BMC Neuroscience

RESEARCH ARTICLE Open Access

COX5A over‑expression protects

cortical neurons from hypoxic ischemic
injury in neonatal rats associated with TPI
up‑regulation
Ya Jiang1†, Xue Bai3†, Ting‑Ting Li3†, Mohammed AL‑Hawwas2, Yuan Jin1, Yu Zou1, Yue Hu1, Lin‑Yi Liu4,
Ying Zhang3, Qing Liu3, Hao Yang1, Jun Ma4, Ting‑Hua Wang1*† , Jia Liu1*† and Liu‑Lin Xiong1,2,3*†

Abstract
Background: Neonatal hypoxic-ischemic encephalopathy (HIE) represents as a major cause of neonatal morbid‑
ity and mortality. However, the underlying molecular mechanisms in brain damage are still not fully elucidated. This
study was conducted to determine the specific potential molecular mechanism in the hypoxic-ischemic induced
cerebral injury.
Methods: Here, hypoxic-ischemic (HI) animal models were established and primary cortical neurons were sub‑
jected to oxygen–glucose deprivation (OGD) to mimic HIE model in vivo and in vitro. The HI-induced neurological
injury was evaluated by Zea-longa scores, Triphenyte-trazoliumchloride (TTC) staining the Terminal Deoxynucleotidyl
Transferased Utp Nick End Labeling (TUNEL) and immunofluorescent staining. Then the expression of Cytochrome c
oxidase subunit 5a (COX5A) was determined by immunohistochemistry, western blotting (WB) and quantitative real
time Polymerase Chain Reaction (qRT-PCR) techniques. Moreover, HSV-mediated COX5A over-expression virus was
transducted into OGD neurons to explore the role of COX5A in vitro, and the underlying mechanism was predicted by
GeneMANIA, then verified by WB and qRT-PCR.
Results: HI induced a severe neurological dysfunction, brain infarction, and cell apoptosis as well as obvious neuron
loss in neonatal rats, in corresponding to the decrease on the expression of COX5A in both sides of the brain. What’s
more, COX5A over-expression significantly promoted the neuronal survival, reduced the apoptosis rate, and mark‑
edly increased the neurites length after OGD. Moreover, Triosephosephate isomerase (TPI) was predicted as physical
interactions with COX5A, and COX5A over-expression largely increased the expressional level of TPI.
Conclusions: The present findings suggest that COX5A plays an important role in promoting neurological recovery
after HI, and this process is related to TPI up-regulation.
Keywords: Neonatal hypoxic-ischemic encephalopathy, COX5A overexpression, Neuronal survival, TPI

Background
*Correspondence: [email protected]; [email protected]; Neonatal hypoxic-ischemic encephalopathy (HIE) could
[email protected] impact neuronal development and result in abnormal
†
Ya Jiang, Xue Bai, Ting-Ting Li, Ting-Hua Wang, Jia Liu and Liu-Lin Xiong function in whole life, which may underlie the cause of
contribute equally to this work
1
Laboratory Zoology Department, Institute of Neuroscience, Kunming seizures and other severe neurological deficits in chil-
Medical University, Kunming, China dren [1]. As a destructive disease that causes the death of
Full list of author information is available at the end of the article infants [2–4], HIE occurs in 1.5/1000 live births in China

© The Author(s) 2020. This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing,
adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and
the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material
in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material
is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds
the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://crea‑
tivecommons.org/licenses/by/4.0/. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdo‑
main/zero/1.0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data.
Jiang et al. BMC Neurosci (2020) 21:18 Page 2 of 15

[5], and 0.2–0.3% of live births in the United States are thus indicated that COX5A may play important roles in
recorded with HIE cases. Severely, approximately 25% the process of central nervous system diseases. However,
of HIE infants die in their first year, while another 25% the role of COX5A and its associated mechanism with
survive but with permanent neurological disabilities neonatal HIE remain to be explored.
[6]. The most significant risk factor for HIE is perinatal Here, in order to find out new targets based on COX5A
asphyxia that may occur either in utero or post-natally. and provide the basic evidence for the treatment of HIE
Utero asphyxia is caused by inadequate blood flow or in future, we utilized the animal model of neonatal HI
oxygen supply resulting in breathing difficulty and even injury together with primary cortical neurons subjected
in breathing arrest altogether. Postnatal asphyxia causes to oxygen–glucose deprivation (OGD) as in vitro model
neonatal pulmonary failure such as severe hyaline mem- to analyze the role of COX5A and determine the related
brane disease, meconium aspiration syndrome, pneumo- molecular mechanisms in HIE insults.
nia as well as congenital heart disease [7–10]. However,
brain damage and neuronal cell death through oxidative Materials and methods
stress and excitotoxicity [11] are amongst the outcomes Animals and grouping
of HIE causing high rate of mortality and morbidity in Forty-seven-day-old postnatal (P7) Sprague–Dawley
infants. (SD) male rats (12–15 g) used for in vivo experiments
Increasing treatment studies targeting neonatal HIE and 21-day-old rats for primary cortical neurons cul-
have been developed over the years, and currently, hypo- tures were purchased from the Department of Zoology of
thermia has been considered as a standard therapy. Neo- Kunming Medical University. Experimental procedures
nates receiving hypothermia go through less death and were reviewed following the standard biosecurity and
incidence in IQ score below 70, less severe disability as institutional safety procedures and approved by the Ethic
well as attention-executive dysfunction [12]. However, its Committees at Kunming Medical University in Yunnan
limited effectiveness remains incompletely certain [13]. province, China (reference number: kmmu2018031). All
Besides the hypothermia, the current therapies includ- animals were raised with their mother in plastic cages
ing erythropoietin (EPO) [14], hyperbaric oxygen (HBO) with soft wood and free access to food and water in a
[14], Xenon [15, 16] and melatonin [17] treatments exert temperature (21–25 °C) and humidity (50–60%)-con-
protective effects on the brain and peripheral organs fol- trolled room. Experimental operations and data analysis
lowing HIE, but less than satisfactory. The main reason can’t be done by the same researchers, and data analysis
for this phenomenon is that the molecular mechanisms must be dealt by at least two researchers blinded to the
underlying cell damage in HIE still remain vague, and experimental design.
there lacks an effective intervention against development
of HIE. Thus novel effective interventions based on the Rat model of neonatal HI injury
molecular mechanisms are urgently needed. In this experiment, we employed the classic Rice-Van-
It has been reported that Cytochrome c oxidase subu- nucci method to establish the HI model [23, 24]. Briefly,
nit 5a (COX5A) combines with COX5b to form the P7 SD rats were anesthetized with isoflurane (4% for
COX5, a subtype of cytochrome c oxidase (COX) related induction, 2% for sustained inhalation anesthesia) after
to mitochondria activity [18]. In addition, COX is an they were weighed and numbered. The hypoxic chamber
endogenous metabolic marker and is closely related to was set at 37 °C with humidity 50–80% before the opera-
energy metabolism [19]. For example, employing a 4-ves- tion. Briefly, the midline of the ventral cervical skin was
sel occlusion model of cerebral ischemia in rats, a pre- cut followed by a blunt dissection of parenchyma, to
vious study revealed the activity of COX was decreased expose the right common carotid. Subsequently, the right
at 1 h after ischemia, which may be associated with the common carotid was ligated by an electric coagulator.
neuronal ischemic injury [20]. Moreover, the investiga- Then, the wound was stitched and animals were returned
tion was expanded in the ischemic-reperfusion model to their mother for 1 h before being kept in a hypoxic
to further explore the effects of mitochondrial protein chamber under 8% ­O2 and 92% ­N2 (flow rate 3 L/min) for
synthesis and COX activity, and the results showed that 2 h. Rats in the sham group underwent the same proce-
the degree of COX activity was also decreased to 90.3%, dures, apart from ligation of the right carotid artery.
80.3%, 81.9% and 83% of the control values at 15 min
after cerebral ischemia and 1, 3, 24 h after reperfusion, Zea‑longa score
respectively [21]. Likewise, another study reported that Zea-longa scores were used to evaluate the neurological
COX activity was significantly decreased at 3 h after trau- function in neonatal rats subjected to HI injury, verify-
matic brain injury, and the COX5A subunit was oxda- ing the successful modeling. All rats underwent the zea-
tively modified in the hippocampus [22]. These studies longa score test to evaluate the neurological dysfunction
Jiang et al. BMC Neurosci (2020) 21:18 Page 3 of 15

at 0 h, 2 h, 4 h, 8 h, and 16 h after HI. The specific scor- sections were washed for three times with 0.01 M PBS
ing criteria were as follows: 0 points—no signs of nerve buffer (PH = 7.45), 5 min each. Next, 3% hydrogen per-
injury; 1 point—the contralateral forepaw lost ability to oxide was added onto the sections followed by incubat-
become fully stretched; 2 points—animals turns to one ing at 37 °C for 15 min and washed as described above.
side while walking; 3 points—walking is unstable, falling In order to block the non-specific binding in the tissue,
to one side; 4 points—loss of consciousness [25]. a drop of 5% goat serum (Solarbio, S9070) was added
on the section and incubated at 37 °C for 30 min. Sub-
Tissue acquisition sequently, the primary antibodies (COX5A, Zhongshan-
For morphological analysis, rats in the sham (n = 5 for jinqiao, rabbit, 1:50) were added and 2% goat serum was
TTC staining, n = 3 for immunohistochemical and immu- used as the negative control, lefting overnight in a refrig-
nofluorescent staining) and HI (n = 5 for TTC staining, erator at 4 °C. The next day, sections were washed with
n = 3 for immunohistochemical and immunofluorescent 0.01 M PBS buffer for three times, followed by addition
staining) groups were euthanatized at 16 h after HI under of enhancement solution, incubation at 37 °C for 20 min
deep anesthesia with 4% isoflurane (sustained inhalation and washing with 0.01 M PBS three times. Subsequently,
anesthesia) for 2 min. Then, after the perfusion of 0.9% the secondary antibody (Zhongshanjinqiao, goat anti-
normal saline followed by 4% paraformaldehyde, the rabbit IgG) was added and the sections were incubated
brain was harvested and put into 4% paraformaldehyde at 37 °C for 1 h, and then washed with PBS. Next, DAB
for more than 72 h. With paraffin embedded, the brain staining was performed for 5 min following the manu-
Sects. (5 μm) were prepared for immunohistochemical facturer’s protocol. To perform nuclei staining, the sec-
staining and immunofluorescent staining, and the core tions were treated with hematoxylin solution for 5 min,
infarction regions were the key observations. and soaked in 1% HCl-ethanol solution for 10 s followed
For molecular biology analysis, rats in the sham (n = 6 by 5% ammonia treatment for 1 min. Afterwards, they
for qRT-PCR and n = 6 for WB) and HI (n = 6 for qRT- were dried with serial dilutions of ethanol (75%, 80%,
PCR and n = 6 for WB) groups were sacrificed at 16 h 85%, 95%, and 100%) and treated with xylene solution for
after the surgery under deep anesthesia by continuously 2 min, finally sealed with neutral resins. The immune-
inhaling 4% isoflurane for 2 min followed by perfusion positive pictures were captured using a light microscope
with 0.9% normal saline. Thereafter, the white infarcted (Leica DMI 6000 B inverted microscope, Germany). For
area and non-infracted area of the contralat-eral hemi- quantification, three sections of each brain with 50 μm
sphere were removed and stored at -80 °C for further WB apart were selected, and five fields of each slice were ran-
and qRT-PCR. domly collected at 100×. The positive cells (dark brown
labeled) in each field were quantified via Image-Pro Plus
Triphenyte‑trazoliumchloride (TTC) staining 6.0 software (MediaCybernetics, Silver Spring, MD,
To evaluate the brain damages after HI, TTC staining USA).
was performed to observe the infarction of brain tis-
sues in neonatal HI rats. The whole brain from the sham Primary cortical neuron cultures
and HI rats was quickly removed at 16 h after rats being and COX5A‑over‑expression HSV vector transduction
deeply anaesthetized with 4% isoflurane, brain tissues COX5A overexpression transduction was conducted
was taken out (operating on ice) and frozen in a refrig- for the further functional verification of COX5A’s role
erator at − 20 °C for 10 min before being cut into five in the process of HI. The primary cortical neurons were
pieces (2 mm each). Afterwards, the sections were placed isolated and cultured from freshly dissected brains of rat
in 1% TTC solution at 37 °C and incubated for 30 min. pups at postnatal day one. In brief, the cortical tissues
The sections were then captured using the digital cam- were dissociated to disperse the cells in a neuro-basal
era. The non-ischemic necrotic area was pale red, and the medium (Gibco, California-USA) supplemented with
ischemic necrotic tissue was white. The ratio of infarct 2% B27 and 0.5 mM of glutamine. The cells were plated
was analyzed using Image J software [26]. on poly-l-lysine and laminin coated vessels at the den-
sity as described above and were incubated at 37 °C with
Immunohistochemistry in vivo 5% ­CO2. The media were then replaced on alternating
The immunohistochemistry was employed to detect days as described before [27]. Thereafter, a low toxicity
COX5A expression. The collected brain was paraffin- HSV-COX5A virus was purchased from Sky Bio (Beijing-
embedded and cut into 5 µm sections, then fixed on China) to explore the function of COX5A according to
the glass slides. Before immunohistochemical stain- the manufacturer’s protocol. Briefly, HSV-COX5A virus
ing, the sections were deparaffinized and a circle was was transducted into the neurons after 5 days of incu-
drawn around the tissue with a PAP pen. Thereafter, bation at a MOI = 10. GFP, emitting green fluorescence,
Jiang et al. BMC Neurosci (2020) 21:18 Page 4 of 15

expressed from a co-infection with a GFP-expressing rinsed three times with PBS before being incubated with
virus in neurons after transduction. Then the infection fluorescence-labeled secondary antibodies of Dylight 594
rate was calculated (GFP -positive cells/total neurons). (anti-mouse, abbkine, 1:200) at 37 °C for 1 h. Then DAPI
Detection of the green fluorescence and qRT-PCR were was used to counterstain the nuclei. Five fields in each
employed to demonstrate the successful transduction well were randomly collected at 400× with a fluorescence
and over-expression of COX5A. microscope (Leica, CM1860, Germany). The COX5A
­ uj+ cells, aver-
positive cells and average cell number of T
Oxygen–glucose deprivation (OGD) age length of the neurite (about calculating 15 neurons)
After culturing for 7 days, OGD was established to were quantified by Image-Pro Plus 6.0 software.
mimic the HI condition in vitro. In detail, following three
washes with PBS, the neurons were incubated using glu- Terminal deoxynucleotidyl transferased Utp nick end
cose-free medium (Gibco, USA) in an anaerobic chamber labeling (Tunel) in vivo and in vitro
containing 5% ­CO2 and 95% ­N2 at 37 °C for 90 min as Tunel staining was performed to determine the apoptotic
described earlier [27]. Subsequently, cells were returned occurrence in the cortical infarction region of neonatal
to original medium, then placed in a normoxic chamber HI rats and cultured primary cortical neurons subjected
with 95% air and 5% ­CO2 for 16 h before further testing. to OGD. Briefly, the brain slices (n = 3 brains, and three
sections of each brain with 50 μm apart) and cell sec-
Immunofluorescent staining in vivo and in vitro tions (n = 6 in each group) were rinsed three times with
In order to examine the neuronal viability after HI injury 0.01 M PBS buffer for 5 min, and 50 μl of 0.01% sodium
in vivo and in vitro, immunofluorescence staining was citrate and 0.1% TritonX-100 was added onto each slice
carried out. In vivo, the brain slices (n = 3 brains, and and incubated for 30 min at 37 °C. Then, the slices were
three sections of each brain with 50 μm apart) were sub- incubated with the tunel mixture reagent (tunel label
sequently deparaffinized and washed three times with solution: Tune enzyme solution = 9:1) for 1 h at 37 °C.
0.01 M PBS and pre-incubated for 30 min with 0.3% Tri- For the negative control, only the tunel label solution
tonX-100 in 5% normal goat serum. Then they were incu- was added dropwise. After that, counterstaining of the
bated overnight at 4 °C with primary antibodies against nuclei was performed using DAPI staining. Finally, tunel
NeuN (mouse, Bioss, 1:100). The next day, the primary positive cells were labeled by red fluorescence and can
antibodies were removed and sections were rinsed with be co-stained with DAPI staining, which showed blue
0.01 M PBS buffer for three times before being incu- fluorescence. For quantification, five fields on each slice
bated with the fluorescence-labeled secondary antibod- around the core infarction regions were randomly cho-
ies of Alexa 488 (anti-mouse, Invitrogen, 1:100) for 1 h sen for imaging with a fluorescence microscope (Leica,
at 37 °C. Counterstaining of the nuclei was performed CM1860, Germany) at 200×. Then, about 200 cells in
using DAPI staining. As NeuN was located on the cyto- each field were quantified by Image-Pro Plus 6.0 soft-
plasm and nucleus of neurons, so the positive staining ware. Then, apoptosis rate was calculated as (tunel posi-
was the nuclei of neurons labeled with green fluorescence tive cells/DAPI cells) %.
and co-stained with DAPI, which showed blue fluores-
cence. Finally, the slices were observed under a fluores- Quantitative real‑time polymerase chain reaction
cent microscope (Leica, CM1860, Germany) at 200×. In (qRT‑PCR)
detail, five fields of each slice were randomly selected for QRT-PCR was performed to detect the mRNA expres-
the estimation of relative number of NeuN positive cells sion of COX5A and other candidate genes after COX5A
(green fluorescence labeled) using Image-Pro Plus 6.0 over-expression. The primers were designed using Pre-
software. mier 5.0 software, then verified via the BLAST software
In vitro, to study the expression of COX5A in OGD and synthesised by Takara Bio Inc (Takara, Japan). Total
condition and the effect of COX5A on Neuronal Class III RNAs from the fresh cortex and cultured neurons (n = 6
β-Tubulin ­(Tuj1+) in neurons, the primary neuronal cells per group) were extracted using the RNAiso plus kit
from the normal, OGD, OGD + NC, OGD + COX5A (TaKaRa, JAPAN) and cDNA Synthesis was performed
groups were cultured on glass cover slips. In brief, the using the ReviertAid Kit (Thermo Fisher Scientific Inc.).
cell sections (n = 6 in each group) were fixed with 4% The levels of Gstp1, Sod2, Rho-GDIa, TPI, COX5A and
formalin and blocked with 0.3% TritonX-100 in 5% nor- β-actin mRNA expression in the samples were estimated
mal goat serum for 30 min at 37 °C. The slips were then by qRT-PCR (CFX-96, Bio-Rad, USA) according to the
incubated with anti-COX5A (Mouse, Santa Cruz, 1:100) manufacturer’s instructions using the primers described
and anti-beta III Tubulin antibody (Mouse, abcam, 1:200) in Table 1. Briefly, each reaction was performed in a vol-
respectively at 37 °C for 16–18 h. Then, the cells were ume of 20 μl mixture consisting of 10 μl of SYBR Green
Jiang et al. BMC Neurosci (2020) 21:18 Page 5 of 15

Table 1 The premier sequences

Factor Forward Reverse Annealing
temperature
(°C)

β-actin 5′GAA​GAT​CAA​GAT​CAT​TGC​TCC3′ 5′TAC​TCC​TGC​T TG​C TG​ATC​CA3′ 52

COX5A 5′GTT​GGA​CCA​ATC​ATA​GGC​GCT3′ 5′CAA​TGT​CGA​TCA​CAT​GCA​CCA3′ 56
Rho-GDIa 5′ TAC​AAG​CCT​CCA​GCC​CAG​AA 3′ 5′ AAG​GTC​AAG​CGG​GTC​ACA​AT3′ 53
Gstp1 5′ TCA​AGG​C TC​GCT​CAA​GTC​CA 3′ 5′ TTG​CAT​CGA​AGG​TCC​TCC​AC 3′ 52
Sod2 5′ CCC​TGA​CCT​GCC​T TA​CGA​C T 3′ 5′ AGC​GAC​C TT​GCT​CCT​TAT​TG 3′ 53
TPI 5′ GGG​GCA​ACT​GGA​AGA​TGA​A 3′ 5′ CCT​GGC​GAA​GTC​GAT​GTA​G 3′ 52

Table 2 The antibody information of WB

Protein Manufacturer Lot number Dilution ratio

COX5A (primary antibody) Santa cruz sc-376907 1:500

β-actin (primary antibody) Bioss bsm-33139 M 1:2000
TPI (primary antibody) Bioss bs-4042R 1:300
HRP, Goat Anti-Mouse IgG (secondary antibody) Abbkine A21020 1:500
HRP, Goat Anti-Mouse IgG (secondary antibody) Abbkine A21010 1:500

master mix, 1 μl of cDNA, 7.8 µl of water, and 0.6 µl were visualized with the ECL (ECL Western blotting
each of forward and reverse primers. Finally, the data kit) luminescence solution. The quantitative analy-
were analyzed using a comparative critical threshold sis was carried out by Image J software. The data were
(Ct) method where relative expression was calculated as expressed as a ratio of the protein of interest band to
­2−∆∆Ct method. β-actin band optical density values.

Western blotting (WB)

WB was employed with the intent to detect the pro- Bioinformatics prediction
tein expression of COX5A in the cortex of neona- GeneMANIA provides many large, publicly available
tal HI rats and of TPI in OGD neurons after COX5A biological datasets to find related genes. These include
overexpression. The right and left cortex from rats protein–protein, protein-DNA and genetic interac-
of the sham and HI group and primary neuronal cells tions, pathways, reactions, gene and protein expres-
from the normal, OGD, OGD + NC, OGD + COX5A sion data, protein domains and phenotypic screening
groups (n = 6 wells in each group) were collected and profiles. Data is regularly updated. The relationship
lysed with RIPA buffer containing a proteinase inhibi- between Gstp1, Sod2, Rho-GDIa, TPI and COX5A
tor cocktail (Roche) on ice for 30 min, then the lysates was predicted using the GeneMANIA website (http://
were centrifuged at 12,000 rpm for 10 min at 4 °C. genem​ania.org/).
Protein concentrations were determined by BCA
protein quantification kit. The protein samples were
boiled and denatured with a loading buffer and 40 μg Statistical analysis
of protein was added to each well and run at 120 V for Data were expressed as mean ± standard deviation
90 min. After that, the proteins in the gel were trans- (SD). The comparison between two groups was per-
ferred to a PVDF membrane at a current of 200 mA. formed using a Student’s t test. For comparison of
The PVDF membrane was then blocked with 5% skim multiple groups in vitro experiments, ANOVA with
milk powder for 90 min at room temperature, and fur- least-significant difference (LSD) or Dunnett’s T3 post
ther incubated with related primary antibodies (seen in hoc test was applied, if equal variances were found, LSD
Table 2) at 4 °C overnight. β-actin was used as an inter- was performed; otherwise, Dunnett’s T3 was used [28].
nal control. Thereafter, membranes were incubated All statistical analyses were performed with SPSS18.0
with the secondary antibodies for 2 h (seen in Table 2). software (IBM Corporation, NY, USA). p < 0.05 was
Finally, the membranes were developed and the bands considered statistically significant.
Jiang et al. BMC Neurosci (2020) 21:18 Page 6 of 15

Results HI caused neuron loss

HI induced cerebral infarction and cell apoptosis Immunofluorescent staining showed that the nuclei
TTC staining and zea-longa scoring were used to of neurons were labeled with NeuN. The fluorescence
evaluate the reliability of the HI model establishment intensity weakened and positive cells of NEUN specific
in this study. The zea-longa test showed an obvious for neurons decreased in the HI group, in comparison to
increase in the score of rats in the HI group compared sham group (HI vs. sham, p < 0.05, Fig. 2a, b).
with the sham group (HI vs. sham, p < 0.05, Fig. 1a).
TTC staining showed that the infarcted brain tissue COX5A expression was decreased after HI injury in vivo
was white without detectable staining, while the unaf- and in vitro
fected brain tissue was stained red. Compared with To detect the expression changes of COX5A in the cer-
the sham group, a quantitative analysis showed that ebral cortex after HI, we performed immunohistochem-
the infarct volume in HI group was obviously higher istry, qRT-PCR and WB. The immunohistochemistry
than that in the sham group (HI vs. sham, p < 0.01, results showed that COX5A was localized in the nucleus
Fig. 1b, c), which further confirmed the HI model was of the cortex at 16 h (Fig. 3a, b). Compared with the sham
established successfully. In addition, the TUNEL stain- group, the positive cells of COX5A were significantly
ing revealed that the apoptosis rate in the infraction reduced in the left and right cortex in the HI group (HI
region of the right cortex in the HI group was signifi- vs. sham, p < 0.01, Fig. 3c, d). COX5A protein and mRNA
cantly increased in comparison with the sham group expression were significantly diminished in the left and
(HI vs. sham, p < 0.01, Fig. 1d, e). right cortex as compared with the sham group revealed
by WB (HI vs. sham, p < 0.01, Fig. 3e, f ) and qRT-PCR
detection (HI vs. sham, p < 0.05, Fig. 3g, h).

Fig. 1 Successful establishing of HIE model in vivo. a The zea-longa scores in sham and HI groups. N = 20/group. b The comparison of TTC
staining between sham and HI groups, and the white staining represented the ischemic area. N = 5/group. c The percentage of infarct area in
sham and HI groups. d Tunel assay showed the apoptotic cells in right cortex. Red fluorescence represents the positive tunel staining, and blue
is the nucleus staining with DAPI. N = 3/group. e Apoptosis rate in the right cortex, revealed by the tunel/DAPI. HI hypoxic-ischemic, h hour, TTC​
Triphenyte-trazoliumchloride, tunel terminal deoxynucleotidyl transferasedUtp nick end labeling staining. Data are presented as the mean ± SD.
*p < 0.05, **p < 0.01. Scale bar = 100 μm
Jiang et al. BMC Neurosci (2020) 21:18 Page 7 of 15

Fig. 2 Evaluation of neuronal loss after HI. a Representative images of immunofluorescent staining of NeuN (a neuron marker) in the right cortex.
Green staining represents the NeuN positive cells, and the nucleus is stained by blue. b Bar chart of relative number of NeuN positive cells per field.
N = 3/group. HI: hypoxic-ischemic. Data are presented as the mean ± SD. *p < 0.05. Scale bar = 100 μm

In the bright-filed images, we observed that com- HSV‑COX5A vector was successfully transducted
pared with the normal group, differentiation and in the primary neuronal cells
development of neurons was slower in the OGD group The successful transduction of HSV-COX5A vector into
90 min after OGD operation (Fig. 4a). In addition, the primary cortical neurons was demonstrated by the
COX5A mRNA expression was also decreased in the green fluorescence emitted by GFP in Fig. 5a, and the
OGD cells (OGD vs. normal, p < 0.05, Fig. 4b). Moreo- infection rate reached about 92%. What’s more, qRT-
ver, immunofluorescent staining further revealed that PCR showed that the mRNA expression of COX5A was
the number of COX5A positive cells was decreased significantly increased in the COX5A over-expression
significantly in OGD group as compared with the Nor- group in both normal and OGD conditions as compared
mal group (OGD vs. normal, p < 0.01, Fig. 4c, d). with the NC (COX5A-ORF vs. NC, p < 0.01, Fig. 5b, c).
Jiang et al. BMC Neurosci (2020) 21:18 Page 8 of 15

Fig. 3 The expression of COX5A after HI. a, b Representative images of immunohistochemical staining of COX5A in left- and right-cortex, Red
arrows represent the COX5A positive cells, which were stained by the dark brown colour. N = 3/group. c, d Bar charts of the COX5A positive cells in
left- and right- cortex. N = 3/group. e, f WB results of the COX5A protein expression in left- and right-cortex. N = 6/group. g, h qRT-PCR results of the
COX5A mRNA expression in left- and right-cortex. N = 6/group. HI hypoxic-ischemic, COX5A cytochrome c oxidase subunit 5a, WB western blotting,
qRT-PCR quantitative real-time polymerase chain reaction. Data are presented as the mean ± SD. *p < 0.05, **p < 0.01. Scale bar = 200 μm

It demonstrated that COX5A was successfully over- (Fig. 6a). Quantitative evaluation revealed that the OGD
expressed in the cortical neurons. treated cells had overall shorter length of neurites and
less cell number relative to normal group (OGD vs. nor-
Over‑expression of COX5A promoted the neuronal mal, p < 0.01, Fig. 6a, d, e). In addition, OGD also led to
outgrowth and decreased the cell apoptosis increased cell apoptosis indicated by the elevated apop-
To study the influence of COX5A overexpression on neu- tosis rate (OGD vs. normal, p < 0.01, Fig. 6b, f ). However,
rites outgrowth and neuronal apoptosis, immunofluores- COX5A over-expression in OGD neurons showed a sig-
cent staining of Tuj1 and TUNEL assay were employed nificant elevation in the cell number, and increase in the
Jiang et al. BMC Neurosci (2020) 21:18 Page 9 of 15

Fig. 4 The expression of COX5A in OGD neurons. a The bright field of neurons in normal and OGD groups. Scale bar = 100 μm. b Quantitative
analysis of relative mRNA expression of COX5A in neuron. N = 6/group. c Immunofluorescent staining of COX5A in normal and OGD groups. Red
staining represents the COX5A positive cells, and the nucleus is stained by blue, n = 3/group. Scale bar = 200 μm. d The COX5A positive cells per
field were calculated. N = 6/group. OGD oxygen–glucose–deprivation, COX5A cytochrome c oxidase subunit 5a, qRT-PCR quantitative real-time
polymerase chain reaction. Data are presented as the mean ± SD. *p < 0.05, **p < 0.01

neurite lengths as well as decrease in apoptosis rate in the mRNA expression of Rho-GDIa and Sod2 was sig-
comparison with the OGD + NC group (OGD + COX5A nificantly reduced (OGD + COX5A vs. OGD + NC,
vs. OGD + NC, p < 0.05, p < 0.01, Fig. 6d–f ). These p < 0.05, Fig. 7b, e), while the Gstp1 and TPI expres-
revealed COX5A over-expression could maintain the sion was significantly increased (OGD + COX5A vs.
neurites and inhibit neuronal apoptosis so as to improve OGD + NC, p < 0.01, Fig. 7c, d). Altogether, the findings
neurological function. revealed that only TPI expression was positively corre-
lated with COX5A. WB analysis further confirmed the
COX5A over‑expression markedly up‑regulated TPI protein expression of TPI was increased after COX5A
expression overexpression in OGD neurons (OGD + COX5A vs.
GeneMANIA was used to explore the underlying molec- OGD + NC, p < 0.05, Fig. 7f ).
ular mechanism of COX5A over-expression in regulating
neurites outgrowth and cell apoptosis in the cortical neu-
rons. Interestingly, four genes (Gstp1, TPI, Rho GDIa and Discussion
Sod2) were screened out. What’s more, it was found that In this study, employing a rat model with neonatal
Gstp1 and TPI are co-localized with COX5A, and the TPI hypoxic-ischemic in vivo, and OGD neuronal cell injury
was physically interacted with COX5A (Fig. 7a). Moreo- model in vitro, we found that the expression of COX5A
ver, the qRT-PCR verification showed that as compared was significantly decreased after HI with more neuronal
to the normal group, the mRNA levels of Rho-GDIa and damages and apoptosis in the right cortical injury area.
Gstp1 were markedly increased while the expression of In addition, over-expression of COX5A effectively pro-
TPI was reduced in OGD group, and Sod2 levels didn’t moted the outgrowth of neuronal neurite and reduced
respond to the OGD (OGD vs. normal, p < 0.01, Fig. 7b– apoptosis in neurons subjected to OGD, and the poten-
e). Conversely, in the OGD cells over-expressing COX5A, tial molecular mechanisms are closely related to the
Jiang et al. BMC Neurosci (2020) 21:18 Page 10 of 15

Fig. 5 Transduction and verification of HSV contained COX5A over-expression. a The effective transduction of the HSV virus into neurons at
MOI = 10. Green fluorecence corresponded to a green fluorescent protein which was a marker protein in HSV vector. Gray corresponded to neuron
in the bright field with PH channel. N = 6/group. b, c Verification of COX5A expression using qRT-PCR in normal and OGD conditions, respectively.
N = 6/group. OGD oxygen–glucose–deprivation, NC negative control, GFP green fluorescent protein, COX5A-ORF cytochrome c oxidase subunit 5a
over-expression, Data are presented as the mean ± SD. **p < 0.01. Scale bar = 100 μm

up-regulation of TPI expression. This may provide a new hypoxia for 2 h. As a result, the cerebral injuries were
idea for future clinical treatment with HI injury. mainly concentrated in the right side of the brain [24],
thus, we focused on the right cerebral hemisphere in the
The HI model was successfully established in P7 rats later observation of brain damage.
In this study, the neonatal HI model was successfully
established based on the classic Rice-Vannucci method of Decreased expression of COX5A induced the neuronal
neonatal HI [23, 24]. Previously, an MRI study compared injury
the HI model by Rice-Vannucci and the neonatal stroke In the present study, the expression of COX5A was
filament occlusion, which revealed that the neonatal decreased after HI injury. Multiple studies [34–37]
stroke injury is restricted in the middle cerebral artery, reported that the expression of COX5A decreased in
while it spreads collaterally in the Rice-Vannucci HI a variety of central nervous system diseases, which in
model [29] Therefore, the Rice-Vannucci model of neo- turn caused an imbalance in neuronal energy regula-
natal HI has been used the most in the basic study. The tion. Moreover, Wei HL reported that down-regulation
zea-longa scores were used to evaluate the neurological of COX5A seriously impaired the sensory function in a
function in ischemic model, and also applied to verify the neuroplastic model of SD rat after dorsal root ganglion
hypoxic-ischemic model establishment in neonatal rats resection [38]. What’s more, the down-regulation of
[30, 31], Moreover, literature proved that the brain injury COX5A led to mitochondrial damage and dysfunction,
of P7 rats equals that of full-term or near-term human further accelerated disease progression in the course of
fetuses [32]. Additionally, P7 rats represent the peak HIE disease [39, 40]. COX5A reduction has also been
brain growth, which occurs at term humans and is equiv- demonstrated close correlation with acute myocardial
alent to 34 weeks’ gestation [33]. Therefore, HI model in infarction [41] and diabetes [42]. However, little infor-
the present study was established in P7 neonatal rats by mation has been reported about the role of COX5A in
the right common carotid artery ligation and subsequent neuronal growth after SCI injury. Exactly in this study,
Jiang et al. BMC Neurosci (2020) 21:18 Page 11 of 15

Fig. 6 The role of COX5A over-expression on neurites outgrowth and cell apoptosis in vitro. a Representative images of Tuj1 immunofluorecent
staining in each group, DAPI labeled nucleus (blue), Tuj1 labeled neurons (red). Scale bar = 50 μm. b TUNEL assay in neurons after COX5A
over-expression under OGD condition. DAPI labeled nucleus (blue), red color represents positive apoptotic cells. Scale bar = 100 μm. c Quantitative
histogram of average cell number per field. d Bar chart of average length of neuites. e The apoptosis rate of the neurons, which was indicated
by tunel/DAPI. OGD oxygen–glucose–deprivation, NC negative control, COX5A cytochrome c oxidase subunit 5a over-expression, tunel terminal
deoxynucleotidyltransferasedUtp nick end labeling. N = 6/group. Data are presented as the mean ± SD. *p < 0.05, **p < 0.001
Jiang et al. BMC Neurosci (2020) 21:18 Page 12 of 15

Fig. 7 The molecular mechanism of COX5A over-expression on neurite outgrowth after OGD injury. a The relationship between Gstp1, Sod2,
Rho-GDIa, TPI and COX5A predicted by GeneMANIA. b The mRNA expression of Rho GDP-dissociation inhibitor1. c The mRNA expression of Gstp1.
d The mRNA expression of TPI. e The mRNA expression of Sod2. f The protein expression of TPI in each group. OGD oxygen glucose deprivation,
NC negative control, COX5A cytochrome c oxidase subunit 5a over-expression, Gstp1 glutathione S-transferasp1, Sod2 Superoxide dismutase 2,
Rho-GDIα guanine dissociation inhibitor α, TPI triosephosephate isomerase. N = 6/group. Data are presented as the mean ± SD. *p < 0.05, **p < 0.001

we demonstrated the cell number reached about 60 per OGD neurons. Nevertheless, the decreased trend was sig-
field and the each neurite length exceeded 100 μm under nificantly reversed in OGD neurons when overexpressing
normal condition, which were conversely reduced a lot in COX5A. These indicated the decrease of COX5A may
Jiang et al. BMC Neurosci (2020) 21:18 Page 13 of 15

induce the neuronal damage and cell apoptosis, while up-regulation in this study could be useful to improve
overexpression of COX5A could promote the neurites neurological function in the HIE insults, and its role is
length and depressed neuronal apoptosis, strengthening positively correlated with COX5A.
the crucial role of COX5A in the process of HIE.
Conclusion
Over‑expressing COX5A for neuronal improvement Collectively, we demonstrated the decrease of COX5A
is associated with TPI up‑regulation in the cortex after HI and OGD injury, whereas, over-
Our study found that the over-expression of COX5A expressing COX5A in OGD condition could largely
promoted the outgrowth of neurites and reduced apop- improve cell function. The possible mechanism is asso-
tosis in neurons subjected to OGD, therefore uncovering ciated with TPI up-regulation. These findings may con-
its crucial role in the recovery of HIE. As for the role of tribute to the mechanism explanation of HIE and provide
COX5A, a previous study found that the COX5A expres- new idea for the treatment of HIE in future clinic trial.
sion was up-regulated in the injured area after acupunc-
ture treatment with rats after spinal cord injury. Besides,
Abbreviations
XiYang YB constructed COX5A over-expression trans- OGD: Oxygen–glucose–deprivation; NC: Negative control; COX5A:
genic mice and found that hippocampal neurons have Cytochrome c oxidase subunit 5a over-expression; Gstp1: Glutathione
enlarged cell bodies and increased dendritic complex- S-transferasp1; Sod2: Superoxide dismutase 2, Rho-GDIα: Guanine dissociation
inhibitor α; TPI: Triosephosephate isomerase; TUNEL: Terminal Deoxynucle‑
ity, and hippocampal-dependent spatial learning and otidyl TransferasedUtp Nick End Labeling; TTC​: Triphenyte-trazoliumchloride;
memory ability got improved [43]. These indicated that HI: Hypoxic-ischemic; WB: Western blotting; qRT-PCR: Quantitative real-time
increasing COX5A expression may exert neuroprotec- polymerase chain reaction; GFP: Green Fluorescent Protein.

tive effect on the central nervous system diseases. Simi- Acknowledgements

larly, our study found that the expression of COX5A was None.
reduced in HIE rats and OGD neurons, and transduction
Authors’ contributions
of HSV-COX5A over-expressing virus into cortical neu- All authors are responsible for the accuracy of data in this study, and approved
rons in vitro reduced apoptosis and increased neuronal the final version of the manuscript. YJ, JL and LLX designed and supervised
neurite lengths to alleviate the OGD injury, therefore the whole study; XB and TTL performed all the experiments; JM, LYL, and MA
contributed the newly added experiments in the revised stage; YJ, YZ, YH, YZ,
supporting the positive role of COX5A in neural repair QL, and HY participated in the part of experiments and data analysis. XB, JL,
after HI. For the mechanism of COX5A, it was well- LLX and THW revised the final paper. All authors read and approved the final
established that COX5A over-expression could increase manuscript.
CGRP densities in lamina of the spinal cord [44] to fur- Funding
ther promote wound healing. Whereas, the present study This study was supported by the Key Grant of biological Medicine from Yun‑
concealed that the over-expression of COX5A could nan Province (No. 2018ZF007-4), and the Program Innovative Research Team
in Science and Technology in Yunnan Province (No. 2017HC007), and Yunnan
markedly up-regulate the TPI level in OGD neurons, Applied Basic Research Project-Union foundation (No. 2017FE467). They paid
which suggests TPI may be the downstream targets of for the experimental materials and for the expenses incurred in the process of
COX5A in the HI process. As a glycolytic enzyme [45], data collection and analysis.
mutations of TPI in coding gene revealed its association Availability of data and materials
with neurological phenotypes [46]. Also, TPI deficiency The datasets analyzed during the current study available from the correspond‑
is associated with progressive neurological dysfunction ing author on reasonable request.
and commonly leads to the death of the carrier in early Ethics approval and consent to participate
years of childhood [47]. Recently, through development Experimental procedures were reviewed following the standard biosecurity
of genetically engineered variants of TPI, it was hypoth- and institutional safety procedures and approved by the Ethic Committees
at Kunming Medical University in Yunnan province, China (reference number:
esized that this enzyme has non-catalytic function that kmmu2018031).
is important for the survival of neurons, which explains
the neurological complications associated with TPI defi- Consent for publication
Not applicable.
ciency [48]. Nevertheless, TPI is identified as a target for
arginine methyltransferase 5 and cyclin dependent kinase Competing interests
2 [49, 50]. These findings demonstrate the strict regula- The authors confirm that there is no conflict of interest in this article.

tion of TPI by cell cycle pathways, and explain their neu- Author details
rological importance. In fact, new evidence showed the 1
Laboratory Zoology Department, Institute of Neuroscience, Kunming Medi‑
firm relationship between cell cycle imbalance and neu- cal University, Kunming, China. 2 School of Pharmacy and Medical Sciences,
Division of Health Sciences, University of South Australia, Adelaide 5000,
rological disorder [51], in which, Alzheimer’s disease is South Australia. 3 National Traditional Chinese Medicine Clinical Research Base
considered one example where failure in cell balancing and Western Medicine Translational Medicine Research Center, Department
results in massive neurodegeneration [52]. Together, TPI of Cardiac and Cerebral Diseases, Affiliated Traditional Chinese Medicine
Jiang et al. BMC Neurosci (2020) 21:18 Page 14 of 15

Hospital, Southwest Medical University, Luzhou 646000, China. 4 Department 22. Opii WO, et al. Proteomic identification of oxidized mitochondrial
of Neurosurgery, The First People’s Hospital of Zhaotong, Zhaotong 657000, proteins following experimental traumatic brain injury. J Neurotrauma.
China. 2007;24(5):772–89.
23. Rice JE, Vannucci RC, Brierley JB. The influence of immaturity on hypoxic-
Received: 6 December 2019 Accepted: 17 April 2020 ischemic brain damage in the rat. Ann Neurol. 1981;9(2):131–41.
24. Chen W, et al. Matrix metalloproteinases inhibition provides neuroprotec‑
tion against hypoxia-ischemia in the developing brain. J Neurochem.
2009;111(3):726–36.
25. Carter RJ, Morton J, Dunnett SB. Motor coordination and balance in
rodents. Curr Protoc Neurosci. 2001;15(1):8–12.
26. Rousselet E, Kriz J, Seidah NG. Mouse model of intraluminal MCAO: cer‑
References
ebral infarct evaluation by cresyl violet staining. J Vis Exp. 2012;69:e4038.
1. Dixon BJ, et al. Neuroprotective strategies after neonatal hypoxic
27. Liu XH, et al. Hyperbaric oxygenation reduces long-term brain injury
ischemic encephalopathy. Int J Mol Sci. 2015;16(9):22368–401.
and ameliorates behavioral function by suppression of apoptosis in a rat
2. Doycheva D, et al. Granulocyte-colony stimulating factor in combination
model of neonatal hypoxia-ischemia. Neurochem Int. 2013;62(7):922–30.
with stem cell factor confers greater neuroprotection after hypoxic-
28. Kim HY. Statistical notes for clinical researchers: Sample size calculation 3.
ischemic brain damage in the neonatal rats than a solitary treatment.
Comparison of several means using one-way ANOVA. Restor Dent Endod.
Transl Stroke Res. 2013;4(2):171–8.
2016;41(3):231–4.
3. Shetty J. Neonatal seizures in hypoxic-ischaemic encephalopathy–
29. Ashwal S, et al. Comparison of two neonatal ischemic injury models
risks and benefits of anticonvulsant therapy. Dev Med Child Neurol.
using magnetic resonance imaging. Pediatr Res. 2007;61(1):9–14.
2015;57(Suppl 3):40–3.
30. Zhang ZB, et al. MiRNA-7a-2-3p inhibits neuronal apoptosis in oxygen-
4. Volpe JJ. Perinatal brain injury: from pathogenesis to neuroprotection.
glucose deprivation (OGD) model. Front Neurosci. 2019;23(13):16.
Ment Retard Dev Disabil Res Rev. 2001;7(1):56–64.
31. Xiong LL, et al. Suppression of PDGF induces neuronal apoptosis after
5. Kurinczuk JJ, White-Koning M, Badawi N. Epidemiology of neonatal
neonatal cerebral hypoxia and ischemia by inhibiting P-PI3K and P-AKT
encephalopathy and hypoxic-ischaemic encephalopathy. Early Hum Dev.
signaling pathways. Brain Res. 2019;15(1719):77–88.
2010;86(6):329–38.
32. Clancy B, Darlington RB, Finlay BL. Translating developmental time across
6. Logitharajah P, Rutherford MA, Cowan FM. Hypoxic-ischemic encephalop‑
mammalian species. Neuroscience. 2001;105(1):7–17.
athy in preterm infants: antecedent factors, brain imaging, and outcome.
33. Bennet L, et al. Cell therapy for neonatal hypoxia-ischemia and cerebral
Pediatr Res. 2009;66(2):222–9.
palsy. Ann Neurol. 2012;71(5):589–600.
7. Barkovich AJ, et al. Proton spectroscopy and diffusion imaging on the first
34. Antequera D, et al. Cytoplasmic gelsolin increases mitochondrial activity
day of life after perinatal asphyxia: preliminary report. Am J Neuroradiol.
and reduces Abeta burden in a mouse model of Alzheimer’s disease.
2001;22(9):1786–94.
Neurobiol Dis. 2009;36(1):42–50.
8. Bruce DA, Schut L. Spinal lipomas in infancy and childhood. Childs Brain.
35. Callahan LA, Supinski GS. Downregulation of diaphragm electron
1979;5(3):192–203.
transport chain and glycolytic enzyme gene expression in sepsis. J Appl
9. Ferreira GC, McKenna MC. l-Carnitine and acetyl-l-carnitine
Physiol. 2005;99(3):1120–6.
roles and neuroprotection in developing brain. Neurochem Res.
36. Liang HL, Ongwijitwat S, Wong-Riley MT. Bigenomic functional regulation
2017;42(6):1661–75.
of all 13 cytochrome c oxidase subunit transcripts in rat neurons in vitro
10. Zarifi MK, et al. Prediction of adverse outcome with cerebral lactate level
and in vivo. Neuroscience. 2006;140(1):177–90.
and apparent diffusion coefficient in infants with perinatal asphyxia.
37. Ongwijitwat S, Wong-Riley MT. Is nuclear respiratory factor 2 a master
Radiology. 2002;225(3):859–70.
transcriptional coordinator for all ten nuclear-encoded cytochrome c
11. Northington FJ, Chavez-Valdez R, Martin LJ. Neuronal cell death in neona‑
oxidase subunits in neurons? Gene. 2005;360(1):65–77.
tal hypoxia-ischemia. Ann Neurol. 2011;69(5):743–58.
38. Wei H. Role of COX5A in neuroplasticity following partial dorsal root rhi‑
12. Shankaran S. Therapeutic hypothermia for neonatal encephalopathy. Curr
zotomy and electro-acupuncture. Kunming: Kunming Medical University;
Treat Options Neurol. 2012;14(6):608–19.
2012.
13. Laptook AR, et al. Effect of therapeutic hypothermia initiated after 6 hours
39. Zhu C, et al. Involvement of apoptosis-inducing factor in neuronal
of age on death or disability among newborns with hypoxic-ischemic
death after hypoxia-ischemia in the neonatal rat brain. J Neurochem.
encephalopathy: a randomized clinical trial. JAMA. 2017;318(16):1550–60.
2003;86(2):306–17.
14. Wei L, et al. Hyperbaric oxygenation promotes neural stem cell prolifera‑
40. Modjtahedi N, et al. Apoptosis-inducing factor: vital and lethal. Trends
tion and protects the learning and memory ability in neonatal hypoxic-
Cell Biol. 2006;16(5):264–72.
ischemic brain damage. Int J Clin Exp Pathol. 2015;8(2):1752–9.
41. Hui J, Chun Y, Da Z. Effects of Chinese herbs capable of replenishing qi,
15. Schmidt M, et al. Xenon attenuates cerebral damage after ischemia in
nourishing yin and activating blood circulation and their compatibility
pigs. Anesthesiology. 2005;102(5):929–36.
on differentially expressed genes of ischemic myocardium. Chin Sci Bull.
16. Dingley J, et al. Xenon provides short-term neuroprotection in neonatal
2009;54(18):3278–82.
rats when administered after hypoxia-ischemia. Stroke. 2006;37(2):501–6.
42. Lu Y, et al. Predict the neurological recovery under hypothermia after
17. Yildiz EP, Ekici B, Tatli B. Neonatal hypoxic ischemic encephalopathy: an
cardiac arrest using C0 complexity measure of EEG signals. Conf Proc IEEE
update on disease pathogenesis and treatment. Expert Rev Neurother.
Eng Med Biol Soc. 2008;2008:2133–6.
2017;17(5):449–59.
43. XiYang YB, et al. Sodium channel voltage-gated beta 2 plays a vital role in
18. Allen LA, et al. Isoforms of yeast cytochrome c oxidase subunit V affect
brain aging associated with synaptic plasticity and expression of COX5A
the binuclear reaction center and alter the kinetics of interaction with the
and FGF-2. Mol Neurobiol. 2016;53(2):955–67.
isoforms of yeast cytochrome c. J Biol Chem. 1995;270(1):110–8.
44. Fukuda R, et al. HIF-1 regulates cytochrome oxidase subunits to optimize
19. Perez-Gracia E, Torrejon-Escribano B, Ferrer I. Dystrophic neurites of senile
efficiency of respiration in hypoxic cells. Cell. 2007;129(1):111–22.
plaques in Alzheimer’s disease are deficient in cytochrome c oxidase.
45. Albery WJ, Knowles JR. Free-energy profile of the reaction catalyzed by
Acta Neuropathol. 2008;116(3):261–8.
triosephosphate isomerase. Biochemistry. 1976;15(25):5627–31.
20. Racay P, et al. Mitochondrial calcium transport and mitochondrial dys‑
46. Tanaka KR, Zerez CR. Red cell enzymopathies of the glycolytic pathway.
function after global brain ischemia in rat hippocampus. Neurochem Res.
Semin Hematol. 1990;27(2):165–85.
2009;34(8):1469–78.
47. Orosz F, Olah J, Ovadi J. Triosephosphate isomerase deficiency: facts and
21. Racay P, et al. Ischemia-reperfusion induces inhibition of mitochondrial
doubts. IUBMB Life. 2006;58(12):703–15.
protein synthesis and cytochrome c oxidase activity in rat hippocampus.
48. Roland BP, et al. Evidence of a triosephosphate isomerase non-catalytic
Physiol Res. 2009;58(1):127–38.
function crucial to behavior and longevity. J Cell Sci. 2013;126(Pt
14):3151–8.
Jiang et al. BMC Neurosci (2020) 21:18 Page 15 of 15

49. Kim C, et al. Regulation of post-translational protein arginine methylation 52. Yang Y, Geldmacher DS, Herrup K. DNA replication precedes neuronal cell
during HeLa cell cycle. Biochim Biophys Acta. 2010;1800(9):977–85. death in Alzheimer’s disease. J Neurosci. 2001;21(8):2661–8.
50. Lee WH, et al. Functional inactivation of triosephosphate isomerase
through phosphorylation during etoposide-induced apoptosis in HeLa
cells: potential role of Cdk2. Toxicology. 2010;278(2):224–8. Publisher’s Note
51. Khurana V, Feany MB. Connecting cell-cycle activation to neurodegenera‑ Springer Nature remains neutral with regard to jurisdictional claims in pub‑
tion in Drosophila. Biochim Biophys Acta. 2007;1772(4):446–56. lished maps and institutional affiliations.

Ready to submit your research ? Choose BMC and benefit from:

• fast, convenient online submission

• thorough peer review by experienced researchers in your field
• rapid publication on acceptance
• support for research data, including large and complex data types
• gold Open Access which fosters wider collaboration and increased citations
• maximum visibility for your research: over 100M website views per year

At BMC, research is always in progress.

Learn more biomedcentral.com/submissions