Single-nucleotide polymorphism screening and RNA sequencing of key messenger RNAs associated with neonatal hypoxic-ischemia brain damage

Abstract

A single-nucleotide polymorphism (SNP) is an alteration in one nucleotide in a certain position within a genome. SNPs are associated with disease susceptibility. However, the influences of SNPs on the pathogenesis of neonatal hypoxic-ischemic brain damage remain elusive. Seven-day-old rats were used to establish a hypoxic ischemic encephalopathy model. SNPs and expression profiles of mRNAs were analyzed in hypoxic ischemic encephalopathy model rats using RNA sequencing. Genes exhibiting SNPs associated with hypoxic ischemic encephalopathy were identified and studied by gene ontology and pathway analysis to identify their possible involvement in the disease mechanism. We identified 89 up-regulated genes containing SNPs that were mainly located on chromosome 1 and 2. Gene ontology analysis indicated that the up-regulated genes containing SNPs are mainly involved in angiogenesis, wound healing and glutamatergic synapse and biological processing of calcium-activated chloride channels. Signaling pathway analysis indicated that the differentially expressed genes play a role in glutamatergic synapses, long-term depression and oxytocin signaling. Moreover, intersection analysis of high throughput screening following PubMed retrieval and RNA sequencing for SNPs showed that CSRNP1, DUSP5 and LRRC25 were most relevant to hypoxic ischemic encephalopathy. Significant up-regulation of genes was confirmed by quantitative real-time polymerase chain reaction analysis of oxygen-glucose-deprived human fetal cortical neurons. Our results indicate that CSRNP1, DUSP5 and LRRC25, containing SNPs, may be involved in the pathogenesis of hypoxic ischemic encephalopathy. These findings indicate a novel direction for further hypoxic ischemic encephalopathy research. This animal study was approved on February 5, 2017 by the Animal Care and Use Committee of Kunming Medical University, Yunnan Province, China (approval No. kmmu2019038). Cerebral tissue collection from a human fetus was approved on September 30, 2015 by the Ethics Committee of Kunming Medical University, China (approval No. 2015-9).

Chinese Library Classification No. R446; R654.5+5; R741

Introduction

Hypoxic ischemic encephalopathy (HIE) refers to a serious neurological syndrome that occurs in the earliest days of life because of placental insufficiency or umbilical cord occlusion in the perinatal period (Allen and Brandon, 2011). HIE is a major neuro-developmental disability in infants, with a prevalence of approximately 1 to 6 per 1000 births (Vannucci, 2000; Chau et al., 2014). Twenty-five percent of HIE patients suffer permanent neurological deficits (Graham et al., 2008). Additionally, worldwide, approximately one million newborns die from HIE each year (Lv et al., 2017). HIE can result in periventricular leukomalacia or developmental retardation, and can cause dysfunction in remote organs (Zeppellini et al., 2001; Wu et al., 2011; Zhao et al., 2012; Alaro et al., 2014; Khatri et al., 2014; Saeed et al., 2014). Despite numerous clinical trials, many neuro-protective strategies have failed to effectively treat HIE patients. The need to expand our understanding of HIE mechanisms and to develop novel therapies is therefore urgent (Ginsberg, 2008; Davies et al., 2019).

A single-nucleotide polymorphism (SNP) is the substitution of a single nucleotide at a certain position in a genome. Generally, SNPs are present at varying percentages within a population (e.g., > 1%) and most do not cause any disorder. However, SNPs have been linked to disorders; for example SNPs in non-coding regions can increase the risk of cancer (Li et al., 2014) or they can influence messenger RNA (mRNA) structure to increase disease susceptibility (Lu et al., 2015). This is in addition to the well-characterized SNPs related to drug metabolism. Research into SNPs is therefore important to identify an individual’s genetic tendency to develop particular diseases (Goldstein, 2001; Lee, 2004; Yanase et al., 2006; Li et al., 2019). SNPs in vasoactive intestinal polypeptide and N-methyl-D-aspartate receptor subunit 3A play roles in cerebral palsy in two-year-old infants after preterm birth (Costantine et al., 2012). Therefore, to observe the role of SNPs in HIE, we aimed to detect key SNPs associated with HIE pathogenesis in rats.

RNA sequencing (RNA-seq), also known as second-generation sequencing, is a powerful tool to analyze transcriptome changes within cells and tissues. This approach enables the collection of data on gene fusion and spliced transcripts, post-transcriptional modifications, gene expression over time and SNP diagnosis (Maher et al., 2009). It also distinguishes RNA populations, such as ribosomal RNAs, transfer RNAs, microRNAs and small RNAs (Ingolia et al., 2012). Recently, this technology has been widely used to investigate the differential expression of genes. Therefore, in the present study, we used RNA-Seq to identify SNPs and changes in gene expression in rats subjected to HIE. We sought to identify SNPs involved in HIE, with the eventual aim of evaluating the genetic predisposition of an individual to developing HIE.

We used seven-day-old rats to establish a HIE model. SNPs and the expression profiles of mRNAs were analyzed in HIE and control brains using RNA-seq and compared. Genes exhibiting SNPs associated with HIE were identified and studied by gene ontology (GO) and pathway analysis to identify their possible involvement in the disease mechanism.

Materials and Methods

Animal care and grouping

Animal experiments and animal handling procedures were approved on February 5, 2017 by the Animal Care and Use Committee of Kunming Medical University, Kunming, Yunnan Province, China (approval No. kmmu2019038) and were performed in accordance with the guidelines of the Unites States National Institutes of Health. Twenty-four specific pathogen-free one-week-old male Sprague-Dawley rats weighing 100–200 g were procured from the Animal Centre of Kunming Medical University (Yunnan, China) (license No. SCXK k2015-0002). The neonate rats were housed at 21–25°C and 45–50% humidity. The animals were exposed to light for 12 hours during the daytime with free access to food and water. Pups were randomly allocated to the sham-operated group (sham, n = 12) and the HIE group (HIE, n = 12). The HIE group was subjected to permanent hypoxia ischemia. The sham-operated group was reared under standard conditions (Figure 1A & B).

Figure 1.

Change in body weight and Zea-longa score in HIE rats.

(A) Rats underwent hypoxia-ischemic brain damage or a sham procedure and were then returned to their cages for recovery. Their mothers normally fed them during recovery. (B) Wounds were sutured and animals numbered for identification. (C) Body weight difference was measured at 24 hours after HIE in the groups of sham and HIE. Zea-longa scores of rats subjected to sham and HIE procedures were performed at 0, 2, 4, 6, 12, 20 and 24 hours post-surgery. The rats with high Zea-longa scores had serious neurological damage. However, there were mild neurofunctional deficits in rats with low scores (Additional Figure 1 ^{(438.8KB, tif)} ). Data are expressed as the mean ± SD, and were analyzed by Student’s t-test. *P < 0.05, vs. sham group. HIE: Hypoxic ischemic encephalopathy.

Induction of hypoxia-ischemia brain damage

HIE was induced using the suture occlusion technique as described previously (Ding et al., 2017). Briefly, animals were anesthetized using isoflurane and immobilized. A 0.5 cm skin incision was made along the midline of the neck, and the right common carotid artery was exposed and occluded with an electrocoagulator (Spring Medical Beauty Equipment Co., Wuhan, Hubei Province, China). Following the surgical procedure, the pups were returned to their dams for recovery and feeding for 1 hour. The pups were then placed into an airtight chamber maintained at 37°C and subjected to hypoxia for 2 hours (8% O₂, 92% N₂). The sham group underwent anesthesia and the common carotid artery was exposed but not ligated. The sham group was not exposed to hypoxia.

Measurement of body weight

Animals were weighed on a high precision balance (Shanghai puchun measure instrument Co., Shanghai, China) before surgery and at 24 hours after HIE.

Zea-longa neurological score

To evaluate the success of HIE model establishment and to identify behavioral changes following the operation, animals were assessed for neurological disorders before surgery and 0, 2, 4, 6, 12, 20 and 24 hours post-surgery. The Zea-longa neurological score was performed as described previously with minor modifications (Wang et al., 2010b). Zero points were given for normal behaviors and symmetric double forelimb stretching. The rats whose contralateral forelimb weakness, torso turning or ipsilateral hindlimb could not fully stretch scored 1 point. Affected posture and circling towards the injured side scored 2 points, while 3 point was given to animals that could not weight-bear on the affected side and 4 point indicated that animals could not weight-bear on the affected side and did not exhibit spontaneous locomotor activity or displayed barrel rolling. The animals in the model group with a score ≥ 1 were selected for further analysis.

High-throughput screening

High-throughput screening technologies were performed to search for the key factors participating in the occurrence and development of HIE. The standard as follow: 1) remove the genes reported in the Scientific Citation Index (SCI) paper on the functional and clinical relevance of the cancer species studied in this project; 2) removal of multiple transmembrane protein genes; 3) remove genes that are not explicitly annotated (such as with open reading frame); 4) remove the number of pubMed articles more than 60; 5) existing experimental data to filter genes combined with the key gene database of Kikekien disease.

RNA extraction

At 24 hours post-operation, pups were deeply anesthetized by intraperitoneal injection of pentobarbitone sodium (200–300 μL; 30 mg/m), before perfusion with PBS (pH 7.4) through the heart. The brain cortex was then quickly dissected and placed on dry ice. Total RNA was extracted from the ipsilateral cortex using an RNeasy Mini Kit (Qiagen, Shanghai, China) in accordance with the manufacturer’s protocol. Briefly, RNA from 12 HIE and 12 sham brain samples was pooled into two HIE and two sham pools, which were used for RNA-Seq analysis (performed by Bi-omarker Technologies Co., Beijing, China) (Rodríguez et al., 2014; Seeliger et al., 2014; Wang et al., 2016). Briefly, RNA integrity was confirmed by 2% agarose gel electrophoresis and by analysis on an Agilent Bioanalyzer 2100 System (Agilent Technologies, CA, USA). RNA concentrations were calculated using a NanoDrop 2000 Spectrophotometer (Thermo Fisher Scientific, Wilmington, DE, USA). Finally, ribosomal RNA was depleted in the samples with a Ribo-Zero Ribosomal RNA Removal Kit (Epicentre, Madison, WI, USA).

Construction of mRNA-seq libraries

Sequencing libraries were constructed using the NEBNext^® Ultra^® Directional RNA Library Prep Kit for Illumina^® (NEB, New England, USA) according to the manufacturer’s instructions by Biomarker Technologies Co. (Beijing, China). Briefly, mRNA was fragmented by heating to 94°C for 15 minutes in 5× NEBNext First Strand Synthesis Reaction Buffer. Random primers, reverse transcriptase and murine RNase inhibitor were then added and first strand complementary DNA (cDNA) synthesized at 42°C for 30 minutes. Then, second-strand cDNA was synthesized using a synthesis enzyme mix for 60 minutes at 16°C then for 30 minutes at 20°C. The resulting dsDNA fragments were purified using Agencourt AMPure XP Beads (Beckman Coulter, Beverly, CA, USA). The overhangs were digested to blunt ends with NEB Next End Prep Enzyme Mix and then adaptors linked to the USER Enzyme were ligated to the cDNA. cDNA was then purified using AMPure XP Beads. Finally, the DNA fragments were amplified using Hot Start HiFi PCR Master Mix, and the products were re-purified using the AMPure XP system and library quality was analyzed using an Agilent Bioanalyzer 2100 and qRT-PCR.

Bioinformatic analysis and differential expression analysis

TruSeq PE Cluster Kitv3-cBot-HS was adopted to construct clusters of index-coded samples on an acBot Cluster Generation System (Illumina Inc., San Diego, CA, USA). The RNA library was then sequenced on an Illumina Hiseq platform and paired-end reads generated. The resultant raw reads were further cleaned by eliminating adapter sequences, low-quality reads and poly-N sequences. Furthermore, sequences below 20 nucleotides or more than 30 nucleotides were removed from the clean data to avoid any disruption to downstream analyses. The GC-content and sequence duplication levels of the clean data were then calculated to confirm the quality of the data.

Cufflinks package (version 2.1.1; https://launchpad.net/ubuntu/+source/cufflinks) was used to calculate expression of genes and lncRNAs depending on fragments per kilobase of exon per million reads (FPKM) values (Trapnell et al., 2010). DESeq R package (version 1.10.1) was used for statistical analysis of differential gene expression between groups. The P values were set to < 0.01 based on Benjamini and Hochberg’s method to reduce the false discovery rate. Genes with a log₂ fold expression variation value > 1 were considered to be differentially expressed. EBseq was adopted for the samples without biological replicates (Storey and Tibshirani, 2003).

Gene functional annotation

The non-redundant protein sequence database of the National Center for Biotechnology Information, USA, was used to predict Gene function. Clusters of Orthologous Groups of proteins (KOG/COG) were used to study GO. GO R Packages for comprehensive study of gene function and variations were used to analyze differentially expressed genes and Pfams (Protein families). KEGG (Kyoto encyclopedia of genes and genomes) and KOBAS (KEGG orthology-based annotation system) were applied to identify the statistically enriched pathways of the differentially expressed genes (Mao et al., 2005). An intersection analysis was performed using Venny 2.1 (http://bioinfogp.cnb.csic.es/tools/venny/).

Primary culture of human fetal cortical neurons

Cerebral tissue collection from a 29-day-old human fetus was approved on September 30, 2015 by the Ethics Committee of Kunming Medical University, China (approval No. 2015-9). Informed consent was obtained from the mother. An aborted 29 day-old fetus was collected from the first affiliated hospital of Kunming Medical University and im-mediately stored on ice. The brain was dissected and placed in 75% alcohol for 2 minutes. The cortical tissues were then harvested into cold Dulbecco’s modified Eagle’s medium (DMEM) and cut into 1 mm³ sized blocks. Trypsin (0.25%) digestion for 30 minutes at 37°C was then performed to isolate cells from the cortical tissue. Cells were then rinsed in DMEM containing 10% fetal bovine serum. The tissue suspension was centrifuged at 1000 × g for 10 minutes and the pelleted cells were then resuspended in complete culture medium (Hyclone, Logan, UT, USA) composed of DMEM/high glucose, 10% fetal calf serum and 1% penicillin-streptomycin solution. Neurons were plated in 6-well plates (Corning, New York, USA) coated with poly-D-lysine and laminin (Sigma-Aldrich, St. Louis, MO, USA) at a density of 5 × 10⁵ cells/mL, and incubated at 37°C in a 5% CO₂ atmosphere. Four hours later, the medium was replaced with neurobasal medium containing 2% B27 (Invitrogen, Carlsbad, CA, USA). The culture medium was changed the next day, then half the medium was changed every three days. The cells were then incubated with oxygen-glucose deprivation (OGD). Control group cells were not exposed to OGD.

In vitro OGD

In vitro human cortical neurons were prepared to mimic HIE using an OGD protocol (Joerger-Messerli et al., 2018). Briefly, cells were washed once with 0.01 mM PBS before the medium was changed to glucose-free medium. Cells were then transferred Ito a hypoxia chamber (Thermo Scientific, Waltham, MA, USA) and exposed to a gas mixture of 5%CO₂ and 95%N₂ for 2 hours. Control cells were incubated normally, without exposure to OGD.

Quantitative real-time polymerase chain reaction

Total RNA was isolated from human neurons 24 hours post OGD with Trizol reagent (Takara Bio Inc., Otsu, Japan). RNA was reverse transcribed to cDNA with the Revert AidTM First Strand cDNA Synthesis kit (Thermo Scientific). qRT-PCR was then performed to detect the relative expression of mRNA according to previous protocols (Liu et al., 2014), with the primer sequences shown in Table 1. Reactions were performed in a DNA thermal cycler (Bio-Rad, Bole, USA) according to the following standard protocol: initial denaturation at 95°C for 2 minutes, denaturation at 95°C for 15 seconds, amplification at 53°C for 20 seconds, then at 60°C for 30 seconds for a total of 40 cycles. The threshold cycle (Ct) of each sample was recorded, and relative expression was calculated with normalization to β-actin values using the 2^–ΔΔCt method.

Table 1.

Primers used for quantitative real-time polymerase chain reaction

Gene	Sense	Antisense	Product size (bp)
LRRC25	5’-TAT CGG GGC AGT GGT C-3’	5’-CAT AGT CGG GAG TGG AGG G-3’	233
DUSP5	5’-GAG AAG ATT GAG AGT GAG A-3’	5’-ATC CAT TTG TAG TGT AGG T-3’	239
CSRNP1	5’-ATC CAC ACA CTC ACC CGC C-3’	5’-ATC CAC ACA CTC ACC CGC C-3’	193

Statistical analysis

All data are expressed as the mean ± standard deviation (SD). Results were compared by Student’s t-test using SPSS 17.0 (SPSS, Chicago, IL, USA). Multiple comparisons were adjusted by Benjamini-Hochberg. The P value was set at < 0.05.

Results

Establishment of a hypoxia-ischemia brain damage animal model

Rats were weighed before and after HIE surgery. The bodyweight of the HIE rats was significantly reduced 24 hours post-surgery, whereas that of sham rats was not (P = 0.000, Figure 1C). Compared with the sham group, the Zea-longa scores of HIE rats increased over the 24 hour monitoring period after HIE surgery (P = 0.000, Figure 1C). Rats subjected to HIE scored an average of 2 points, with posturing and circling towards the injured side. These results were an indication of brain injury and deterioration of the rats’ neurological function.

The number of single-nucleotide polymorphisms was increased in hypoxia-ischemia brain damage rats

To confirm the correlation of HIE with particular SNPs, we compared the number of SNPs between HIE and sham groups (Additional Table 1). The SNP statistical analysis was based on comparison of the reads of each sample with reference genomes. There were 130248 and 83060 SNPs in the two HIE samples, with 75646 and 49372 SNPs in gene regions (Figure 2A and Additional Table 1). Meanwhile, the control group samples exhibited 69315 and 72933 SNPs in total, with 48728 and 50516 SNPs, respectively, in gene regions (Figure 2A and Additional Table 1). Evidently, the HIE group exhibited a higher total number of SNPs and a higher number of SNPs in gene regions. This indicated that HIE induced genomic SNPs; however, there were no differences of transition, transversion and heterozygosity od SNPS between the control and HIE groups (Figure 2B, Additional Tables 1 and 2 ^{(23.5MB, pdf)} ).

Additional Table 1.

#Sample_ID		Snp_number	Snp_number_gene	Snp_number_intergenic	Transition (%)	Transversion (%)	Heterozygosity (%)
L01	118043		69315	48728	69.31	30.69	42.81
L02	123449		72933	50516	69.25	30.75	46.77
L03	205894		130248	75646	70.40	29.60	47.05
L04	132432		83060	49372	70.77	29.23	47.10

L01 & L02 samples from control group; L03 & L04 samples from hypoxic ischemic encephalopathy group.

Figure 2.

The change of in SNPs in sham and HIE groups.

(A) Number of SNPs in various fields. L03 and L04 groups had more SNPs and genic SNPs and intergenic SNPs compared with L01 and L02. (B) Percentage of transition, transversion and heterozygous SNPs in L01, L02 and L03, L04 groups showed no significant difference. SNP: Single-nucleotide polymorphism. SNP Number: total number of SNPs; Genic SNP: total number of SNPs in gene regions; intergenic SNP: total number of SNPs in inter-gene regions. Transition: Percentage of SNPs of the conversion type relative to the total number of SNPs; transversion: percentage transversion SNPs relative to the total number of SNPs; heterozygosity: percentage of heterozygous SNPs relative to the total number of SNP. L01 & L02 samples from the sham group; L03 & L04 samples from the HIE group. HIE: Hypoxic ischemic encephalopathy; SNPs: single-nucleotide polymorphisms.

Single-nucleotide polymorphism localization

The chromosomal location of SNPs was investigated in sham and HIE groups. The majority of detected SNPs were observed on the largest chromosomes (N1 and N2) as shown in Additional Table 2 ^{(92.9KB, pdf)} . Generally, the numbers of SNPs on each chromosome were higher in the HIE group compared with the control group.

SNP genotypes in the HIE group

To analyze distinct SNPs in the HIE group that might be involved in the etiology of HIE, we screened the SNPs ac-cording to genotype. A total of ten SNP genotypes were identified, including A, C, G, K, M, R, S, T, W, Y. The most common SNP genotype was G (3544) and the least common was W (507) (Figure 3 and Additional Table 3 ^{(23.5MB, pdf)} ).

Figure 3.

Numbers of SNPs in the ten genotypes of the HIE group.

The abscissa represents the genotypes and the ordinate represents the number of each genotype. HIE: Hypoxic ischemic encephalopathy; SNPs: single-nucleotide polymorphisms.

Enrichment analysis for genes containing SNPs in HIE rats

Differentially expressed genes in both the sham and HIE groups were studied, as was enrichment analysis of GO data derived from sequence analysis. We detected 89 up-regulated genes, including Csrnp1, Dusp5, Lrrc25, Cxcl1, Clca4l, and Serpine1 exhibiting SNPs in the HIE group compared with the sham group (Figure 4 and Additional Table 4). Compared with the sham group, HIE induced more genes to contain SNPs. These genes were involved in cell adhesion, peptidyl-serine phosphorylation, protein autophosphorylation, cytoplasm, cytosol, cell junction, axon, dendrite, dendritic spine, synapse, postsynaptic membrane, phospholipid binding, protein kinase binding, PDZ domain binding and protein complex binding. These functions were in addition to processes common with the sham group, including postsynaptic density, neuronal cell body and ATP binding with neuronal cell body (Figure 4A & B). In addition, pathway analysis revealed that only genes in the sham group containing SNPs were mainly involved in mitogen-activated protein kinase, sphingolipid, adrenergic signaling in cardiomyocytes, circadian entrainment, retrograde endocannabinoid and dopaminergic synapse pathways (Figure 4A). Meanwhile, up-regulated genes containing SNPs only in the HIE group, mainly participated in mitogen-activated protein kinase, circadian entrainment, retrograde endocannabinoid signaling, glutamatergic synapse, long-term depression and oxytocin signaling pathways (Figure 4B). These were quite different from those of the sham group.

Figure 4.

Enrichment analysis of GO terms for differentially expressed mRNAs.

The three most significant items of each GO term are listed in the bar chart. The horizontal axis shows the GO annotations and pathway, and the vertical axis shows the P value. (A) The GO analysis and pathway analysis of the differentially expressed genes in the sham group. Blue represents decreasing expression. Red represents increasing expression. The right chart shows the top nine differentially expressed genes. (B) The GO and pathway analysis of the differentially expressed genes in HIE group. Blue represents decreasing expression. Red represents increasing expression. The right chart shows the top 24 differentially expressed genes. BP: Biological processes; CC: cellular components; GO: gene ontology; MF: molecular functions.

Additional Table 4.

Gene ID	False Discovery rate	Log2FC (fold change)	Regulated Gene name
gene10928	0.007321917	1.234849228 Up	Cecr2
gene11610	0.000494348	1.440158186 Up	Chd7
gene11832	0.000510142	1.401960952 Up	B4galt1
gene12406	0.036291268	1.405219558 Up	Ror1
gene12937	0.00059812	2.539193652 Up	Fgr
gene13033	0.017403119	1.11615355 Up	LOC684122
gene13264	8.00E-07	1.998225126 Up	Casz1
gene1410	0.006976111	1.195557824 Up	Proser3
gene14458	1.01E-08	1.517016795 Up	Elmsan1
gene15930	0.003616995	1.456447854 Up	Irak3
gene16622	1.82E-14	2.088731276 Up	Ccdc134
gene16625	0.001539985	1.275741357 Up	Shisa8
gene16799	1.04E-05	1.01371392 Up	Ano6
gene17912	0.014602678	1.127546269 Up	Arid3b
gene17977	0.033618102	1.652847862 Up	Spesp1
gene17997	0.001029297	1.770125454 Up	Smad6
gene18421	1.46E-05	2.000879142 Up	Slco2a1
gene18711	1.69E-05	1.280236316 Up	Itga9
gene18726	0.004030785	1.288195305 Up	X ylb
gene18736	1.23E-11	3.638238928 Up	Csrnp1
gene18827	0.025594256	3.10825301 Up	Ccr1
gene19018	2.67E-16	Inf Up	LOC100912849
gene19083	0.000139538	2.188379763 Up	Runx2
gene19085	0.030428257	1.484082283 Up	Clic5
gene20150	0.002857127	1.217668196 Up	V asn
gene20655	0.019537489	1.38160466 Up	Tcf7
gene20938	0.000337032	2.83969408 Up	Hs3st3b1
gene21624	0.028824658	1.562356011 Up	Pctp
gene21784	0.000108425	2.220223383 Up	Arl5c
gene21921	2.95E-18	1.480593354 Up	Stat3
gene21922	6.95E-17	1.57894954 Up	Ptrf
gene22053	0.000267204	1.567816761 Up	Itgb3
gene22098	7.88E-09	1.368885314 Up	Pecam1
gene22145	0.007564055	1.691362364 Up	Fam20a
gene22728	5.87E-13	1.237620155 Up	Ets2
gene23174	0.000252217	2.317403614 Up	Leprel1
gene23185	0.001227681	1.037587703 Up	Lpp
gene23858	1.32E-46	5.16528331 Up	Serpine1
gene24177	0.011318539	1.216444191 Up	V sig10
gene24687	3.47E-07	3.017963325 Up	Ptprc
gene25476	0.009700318	1.158533434 Up	Tgfbr3
gene25582	0.000392403	2.521693129 Up	Fgf5
gene25612	5.20E-08	1.768312007 Up	Shroom3
gene25643	1.44E-22	6.759783682 Up	Cxcl1
gene26603	1.77E-05	1.022215948 Up	Flnb
gene27014	0.004656377	1.004953379 Up	Lats2
gene27729	0.014227908	1.010355395 Up	Dcp1a
gene2782	0.001952483	1.028077146 Up	Zfp143
gene27887	0.035139802	3.219479213 Up	Lrrc25
gene28211	0.014214731	1.011538396 Up	Pragmin
gene28758	9.32E-05	1.426349769 Up	Ror2
gene28759	2.07E-15	1.516883601 Up	Nfil3
gene28852	1.67E-07	1.478320873 Up	Rnf144b
gene28867	7.98E-05	1.029774301 Up	Rbm24
gene28923	0.012709411	1.454983568 Up	Mak
gene28946	3.28E-06	1.664387255 Up	Bmp6
gene29280	1.94E-05	2.469452509 Up	Sfrp4
gene2954	0.01129745	1.184598231 Up	Arhgap17
gene29939	4.79E-05	1.17526293 Up	Nrg2
gene30479	1.26E-06	1.923075394 Up	Zfp516
gene3108	9.78E-09	2.654171021 Up	Bag3
gene31215	2.09E-06	1.651489706 Up	Irf8
gene31237	0.017930871	1.346726574 Up	Zfp469
gene31693	0.004789907	2.684390722 Up	RT1-A1
gene31715	0.002824524	1.266958747 Up	Itpr3
gene31863	6.56E-07	2.527943172 Up	Itgb2
gene31869	0.003597344	1.206319056 Up	Col18a1
gene32290	0.013296976	1.495250541 Up	Aim1
gene33491	0.000675436	1.245128971 Up	Lonrf3
gene4319	7.61E-14	3.35208246 Up	Dusp5
gene4981	0.031396721	1.007595524 Up	Fam105a
gene5150	4.11E-06	1.815866502 Up	Cp
gene5422	1.93E-07	1.32403003 Up	W wtr1
gene5687	7.30E-35	Inf Up	Sh2d2a
gene6379	5.43E-07	2.585755151 Up	Synpo2
gene6410	0.003597344	1.301344945 Up	Zgrf1
gene6465	0.005514585	1.771361524 Up	Dkk2
gene6493	0.031604667	1.320650436 Up	Tacr3
gene6563	0.02891601	1.643902108 Up	Gbp5
gene6564	0.000229551	2.077167857 Up	LOC685067
gene6587	7.44E-05	6.178383959 Up	Clca4l
gene7067	0.029239889	1.105943242 Up	Ptgs1
gene7238	4.90E-11	1.174738611 Up	gene7238
gene8116	2.92E-52	3.079275962 Up	Cd44
gene9608	1.17E-08	1.836606238 Up	Cdk6
gene9626	0.020857722	3.432909322 Up	Tfpi2
gene9721	1.10E-06	1.640092399 Up	Mdfic
gene9738	0.00386661	2.268566993 Up	Cftr
gene9891	7.54E-13	1.752861704 Up	Cald1

Further gene ontology analysis of the up-regulated genes containing SNPs in the HIE rats

The majority of mRNA targets in the HIE group were associated with 14 biological processes, including angiogenesis involved in wound healing, bone morphogenetic protein signaling pathway, cell adhesion, immune response, integrin-mediated signaling pathway, regulation of innate immune system, negative regulation of cytokine-mediated im-munity, neutrophil chemotaxis, palate development and positive regulation of protein phosphorylation (Figure 5). The GO results concur with the molecular functions analysis because they indicated that genes primarily participating in growth factor activity were ligands for protein kinases, the intracellular calcium activated chloride channel and epigenetic regulator binding regions within DNA (Figure 5). Similarly, a cellular components analysis showed that these differentially expressed genes encode proteins localized on the external side of the plasma membrane, and with adhesion cellular components, and the integrin alpha9-beta1 complex (Figure 5).

Figure 5.

Gene ontology analysis of the genes containing SNPs.

A map of the 14 biological processes (BP), 4 cellular components (CC) and 4 molecular functions (MF) for the up-regulated genes containing SNPs after HIE. The vertical axis shows the P value. BMP: Bone morphogenetic protein.

Validation of candidate genes of interest

To further screen the gene sequencing data for genes that are truly involved in HIE, a high-throughput screening via PubMed searching was carried out. This revealed 27 factors that have not been studied in HIE (Additional Table 5). Intersection analysis of these 27 genes with the 89 up-regulated genes exhibiting SNPs using Venny 2.2 software (http://bioinfogp.cnb.csic.es/tools/venny/index.html), identified three genes, CSRNP1, DUSP5 and LRRC25, that possessed SNPs (Figure 6A). Therefore, they were assessed in the following experiment. Human fetal cortical neurons were cultured, and OGD was employed to monitor HIE in vitro. Two hours after OGD, the neurons exhibited obvious damage, as indicated by broken cell bodies and axons (Figure 6B). Meanwhile, qRT-PCR showed that the expressional levels of CSRNP1, DUSP5 and LRRC25 were significantly up-regulated 24 hours post OGD, compared with the normal group (CSRNP1: P = 0.000; DUSP5: P = 0.000; LRRC25: P = 0.000; Figure 6C–E).

Additional Table 5.

Number	Gene name	Synonyms	Description	GeneSummary	Transcription quantity	Sequence ID	Cellular localization	Literature quantity from Pubmed	Novoseek disease relationships for the gene	MalaCards disease relationships for the gene	Symbol	Number	False discovery rate	Fold change	Regulated	Function
3437	IFIT3	CIG-49|GARG-49|IFI60|IFIT4|IRG2|ISG60|P60|RIG-G	Interferon-induced protein with tetratricopeptide repeats 3		4	NM_001031683(1473), NM_001289759(1317), NM_001549(1473), NM_001289758(1317)	Cytoplasm. Mitochondrion	42	0	2	Ifit3	gene4056	1.83E-10	5.8044	Up	Negative regulation of cell proliferation
59271	EVA1C	B18|B19|C21orf63|C21orf64|FAM176C|PRED34|SUE21	Eva-1 homolog C (C. elegans)		2	NM_001286556(1317), NM_058187(1326)	Membrane; Single-pass type I membrane protein (Potential)	10	0	0	Eva1c	gene22665	5.12E-11	4.2549	Up	Plasma membrane,integral component of membrane, carbohydrate binding
51804	SIX4	AREC3	SIX homeobox 4	This gene encodes a member of the homeobox family, subfamily SIX. The drosophila homolog is a nuclear homeoprotein required for eye development. Studies in mouse show that this gene product functions as a transcription factor, and may have a role in the differentiation or maturation of neuronal cells. (provided by RefSeq, May 2010)	1	NM_017420(2346)	Nucleus (By similarity)	18	0	0	Six4	gene14292	7.99E-09	4.2156	Up	Regulation of synaptic growth at neuromuscular junction, regulation of protein localization
10361	NPM2	-	Nucleophosmin/nucleoplasmin 2		3	NM_001286681(411), NM_182795(645), NM_001286680(645)	Nucleus (By similarity). Note=Found in the oocyte nucleus before nuclear membrane breakdown, after which it is redistributed to the cytoplasm (By similarity)	13	0	0	Npm2	gene27231	0.0000146	4.1234	Up
115265	DDIT4L	REDD2|Rtp801L	DNA-damage-inducible transcript 4-like		1	NM_145244(582)	Cytoplasm (By similarity)	14	0	1	Ddit4l	gene6516	5.36E-10	3.8915	Up
64651	CSRNP1	AXUD1|CSRNP-1|FAM130B|TAIP-3|URAX1	Cysteine-serine-rich nuclear protein 1	This gene encodes a protein that localizes to the nucleus and expression of this gene is induced in response to elevated levels of axin. The Wnt signalling pathway, which is negatively regulated by axin, is important in axis formation in early development and impaired regulation of this signalling pathway is often involved in tumors. A decreased level of expression of this gene in tumors compared to the level of expression in their corresponding normal tissues suggests that this gene product has a tumor suppressor function. (provided by RefSeq, Jul 2008)	1	NM_033027(1770)	Nucleus (By similarity)	15	0	0	Csrnp1	gene18736	1.23E-11	3.6382	Up
10981	RAB32	-	RAB32, member RAS oncogene family	Small GTP-binding proteins of the RAB family, such as RAB32, play essential roles in vesicle and granule targeting (Bao et al., 2002 (PubMed 11784320)).(supplied by OMIM, Aug 2009)	1	NM_006834(678)	Mitochondrion. Cytoplasmic vesicle, phagosome. Cytoplasmic vesicle, phagosome membrane; Lipid-anchor; Cytoplasmic side (By similarity). Note=Recruited to phagosomes containing S.aureus or M.tuberculosisExplore the universe of human proteins at neXtProt for RAB32: NX_Q13637Post-translational modifications:	19	0	0	Rab32	gene46	4.88E-28	3.4791	Up
2634	GBP2	-	Guanylate binding protein 2, interferon-inducible	Interferons are cytokines that have antiviral effects and inhibit tumor cell proliferation. They induce a large number of genes in their target cells, including those coding for the guanylate-binding proteins (GBPs). GBPs are characterized by their ability to specifically bind guanine nucleotides (GMP, GDP, and GTP). The protein encoded by this gene is a GTPase that converts GTP to GDP and GMP. (provided by RefSeq, Jul 2008)	1	NM_004120(1776)	Cytoplasm. Nucleus. Golgi apparatus membrane; Lipid-anchor	31	0	0	Gbp2	gene6566	0.0000554	3.4162	Up
1847	DUSP5	DUSP|HVH3	Dual specificity phosphatase 5	The protein encoded by this gene is a member of the dual specificity protein phosphatase subfamily. These phosphatases inactivate their target kinases by dephosphorylating both the phosphoserine/threonine and phosphotyrosine residues. They negatively regulate members of the mitogen-activated protein (MAP) kinase superfamily (MAPK/ERK, SAPK/JNK, p38), which are associated with cellular proliferation and differentiation. Different members of the family of dual specificity phosphatases show distinct substrate specificities for various MAP kinases, different tissue distribution and subcellular localization, and different modes of inducibility of their expression by extracellular stimuli. This gene product inactivates ERK1, is expressed in a variety of tissues with the highest levels in pancreas and brain, and is localized in the nucleus. (provided by RefSeq, Jul 2008)	1	NM_004419(1155)	Nucleus (Potential)	58	4	1	Dusp5	gene4319	7.61E-14	3.3521	Up
126364	LRRC25	MAPA	Leucine rich repeat containing 25		1	NM_145256(918)	Membrane; Single-pass type I membrane protein (Potential)	8	0	0	Lrrc25	gene27887	0.0351398	3.2195	Up
8638	OASL	OASLd|TRIP-14|TRIP14|p59 OASL|p59-OASL|p59OASL	2’-5’-Oligoadenylate synthetase-like		3	NM_001261825(1155), NM_198213(768), NM_003733(1545)	Isoform p56: Nucleus, nucleolus. Cytoplasm|Isoform p30: Cytoplasm	39	0	0	Oasl	gene24225	0.0000296	3.2178	Up
57820	CCNB1IP1	C14orf18|HEI10	Cyclin B1 interacting protein 1, E3 ubiquitin protein ligase	HEI10 is a member of the E3 ubiquitin ligase family and functions in progression of the cell cycle through G(2)/M.(supplied by OMIM, Apr 2004)	3	NM_182852(834), NM_021178(834), NM_182849(834)	Nucleus. Chromosome. Note=May associate with segregating chromosomes during metaphase and anaphaseExplore the universe of human proteins at neXtProt for CCNB1IP1: NX_Q9NPC3Post-translational modifications:	21	0	0	Ccnb1ip1	gene26728	0.0000691	3.1892	Up
56833	SLAMF8	BLAME|CD353|SBBI42	SLAM family member 8	This gene encodes a member of the CD2 family of cell surface proteins involved in lymphocyte activation. These proteins are characterized by Ig domains. This protein is expressed in lymphoid tissues, and studies of a similar protein in mouse suggest that it may function during B cell lineage commitment. The gene is found in a region of chromosome 1 containing many CD2 genes. (provided by RefSeq, Jul 2008)	1	NM_020125(858)	Membrane; Single-pass type I membrane protein	10	0	0	Slamf8	gene25120	7.95E-09	3.0308	Up
79094	CHAC1	-	ChaC glutathione-specific gamma-glutamylcyclotransferase 1		2	NM_024111(795), NM_001142776(660)	Golgi apparatus, trans-Golgi network (By similarity). Cytoplasm, cytosol	14	0	0	Chac1	gene8349	7.27E-12	2.9496	Up
64092	SAMSN1	HACS1|NASH1|SASH2|SH3D6B|SLy2	SAM domain, SH3 domain and nuclear localization signals 1	SAMSN1 is a member of a novel gene family of putative adaptors and scaffold proteins containing SH3 and SAM (sterile alpha motif) domains (Claudio et al., 2001 (PubMed 11536050)).(supplied by OMIM, Mar 2008)	3	NM_001256370(1326), NM_022136(1122), NM_001286523(915)	Nucleus. Cytoplasm (By similarity). Cell projection, ruffle (By similarity). Note=Shuttles between cytoplasm and nucleus. Colocalizes with the actin cytoskeleton and actin-rich membrane ruffles (By similarity)	19	0	0	Samsn1	gene22523	0.0006012	2.8449	Up
2615	LRRC32	D11S833E|GARP	Leucine rich repeat containing 32	This gene encodes a type I membrane protein which contains 20 leucine-rich repeats. Alterations in the chromosomal region 11q13-11q14 are involved in several pathologies. (provided by RefSeq, Jul 2008)	2	NM_001128922(1989), NM_005512(1989)	Membrane; Single-pass type I membrane protein	26	0	0	Lrrc32	gene2350	0.000000131	2.7125	Up
55647	RAB20	-	RAB20, member RAS oncogene family		1	NM_017817(705)	Golgi apparatus. Cytoplasmic vesicle, phagosome. Cytoplasmic vesicle, phagosome membrane; Lipid-anchor; Cytoplasmic side (By similarity). Note=Highly enriched on apical endocytic structures in polarized epithelial cells of kidney proximal tubules (By similarity). Recruited to phagosomes containing S.aureus or M.tuberculosis	11	0	0	Rab20	gene28479	0.000000736	2.7008	Up
387496	RASL11A	-	RAS-like, family 11, member A	RASL11A is a member of the small GTPase protein family with a high degree of similarity to RAS (see HRAS, MIM 190020) proteins.(supplied by OMIM, Nov 2008)	1	NM_206827(729)	Nucleus, nucleolus (By similarity). Note=Associates with rDNA transcription unit throughout the cell cycle (By similarity)	10	0	0	Rasl11a	gene23588	7.73E-10	2.6636	Up
55303	GIMAP4	IAN-1|IAN1|IMAP4|MSTP062	GTPase, IMAP family member 4	This gene encodes a protein belonging to the GTP-binding superfamily and to the immuno-associated nucleotide (IAN) subfamily of nucleotide-binding proteins. The encoded protein of this gene may be negatively regulated by T-cell acute lymphocytic leukemia 1 (TAL1). In humans, the IAN subfamily genes are located in a cluster at 7q36.1. (provided by RefSeq, Jul 2008)	1	NM_018326(990)	Cytoplasm, cytosol (By similarity)	17	0	1	Gimap4	gene10131	0.0055691	2.6318	Up
10457	GPNMB	HGFIN|NMB	Glycoprotein (transmembrane) nmb	The protein encoded by this gene is a type I transmembrane glycoprotein which shows homology to the pMEL17 precursor, a melanocyte-specific protein. GPNMB shows expression in the lowly metastatic human melanoma cell lines and xenografts but does not show expression in the highly metastatic cell lines. GPNMB may be involved in growth delay and reduction of metastatic potential. Two transcript variants encoding different isoforms have been found for this gene. (provided by RefSeq, Jul 2008)	2	NM_002510(1683), NM_001005340(1719)	Membrane; Single-pass type I membrane protein (Potential). Melanosome. Note=Identified by mass spectrometry in melanosome fractions from stage I to stage IV	48	3	2	Gpnmb	gene10145	2.82E-20	2.594	Up
121549	ASCL4	HASH4|bHLHa44	Achaete-scute family bHLH transcription factor 4	Basic helix-loop-helix transcription factors, such as ASCL4, are essential for the determination of cell fate and the development and differentiation of numerous tissues (Jonsson et al., 2004 (PubMed 15475265)).(supplied by OMIM, Mar 2008)	1	NM_203436(522)	Nucleus (By similarity)	8	0	0	Ascl4	gene15593	0.0009807	2.5786	Up
55008	HERC6	-	HECT and RLD domain containing E3 ubiquitin protein ligase family member 6	HERC6 belongs to the HERC family of ubiquitin ligases, all of which contain a HECT domain and at least 1 RCC1 (MIM 179710)-like domain (RLD). The 350-amino acid HECT domain is predicted to catalyze the formation of a thioester with ubiquitin before transferring it to a substrate, and the RLD is predicted to act as a guanine nucleotide exchange factor for small G proteins (Hochrainer et al., 2005 (PubMed 15676274)).(supplied by OMIM, Mar 2008)	2	NM_017912(3069), NM_001165136(2961)	Cytoplasm, cytosol	13	0	0	Herc6	gene10306	0.0044702	2.5675	Up
131450	CD200R1	CD200R|HCRTR2|MOX2R|OX2R	CD200 receptor 1	This gene encodes a receptor for the OX-2 membrane glycoprotein. Both the receptor and substrate are cell surface glycoproteins containing two immunoglobulin-like domains. This receptor is restricted to the surfaces of myeloid lineage cells and the receptor-substrate interaction may function as a myeloid downregulatory signal. Mouse studies of a related gene suggest that this interaction may control myeloid function in a tissue-specific manner. Alternative splicing of this gene results in multiple transcript variants. (provided by RefSeq, Jul 2008)	4	NM_138939(567), NM_170780(978), NM_138940(498), NM_138806(1047)	Isoform 1: Cell membrane; Single-pass type I membrane protein|Isoform 4: Cell membrane; Single-pass type I membrane protein|Isoform 2: Secreted|Isoform 3: Secreted	46	0	0	Cd200r1	gene22955	0.0018284	2.4729	Up
51166	AADAT	KAT2|KATII	Aminoadipate aminotransferase	This gene encodes a protein that is highly similar to mouse and rat kynurenine aminotransferase II. The rat protein is a homodimer with two transaminase activities. One activity is the transamination of alpha-aminoadipic acid, a final step in the saccaropine pathway which is the major pathway for L-lysine catabolism. The other activity involves the transamination of kynurenine to produce kynurenine acid, the precursor of kynurenic acid which has neuroprotective properties. Two alternative transcripts encoding the same isoform have been identified, however, additional alternative transcripts and isoforms may exist. (provided by RefSeq, Jul 2008)	4	NM_016228(1278), NM_182662(1278), NM_001286683(1278), NM_001286682(1290)	Mitochondrion (Potential)	32	0	1	Aadat	gene28019	0.0042495	2.4221	Up
23704	KCNE4	MIRP3	Potassium channel, voltage gated subfamily E regulatory beta subunit 4	Voltage-gated potassium (Kv) channels represent the most complex class of voltage-gated ion channels from both functional and structural standpoints. Their diverse functions include regulating neurotransmitter release, heart rate, insulin secretion, neuronal excitability, epithelial electrolyte transport, smooth muscle contraction, and cell volume. This gene encodes a member of the potassium channel, voltage-gated, isk-related subfamily. This member is a type I membrane protein, and a beta subunit that assembles with a potassium channel alpha-subunit to modulate the gating kinetics and enhance stability of the multimeric complex. This gene is prominently expressed in the embryo and in adult uterus. (provided by RefSeq, Jul 2008)	1	NM_080671(666)	Membrane; Single-pass type I membrane protein	26	0	0	Kcne4	gene19690	0.0239597	2.4019	Up
54626	HES2	bHLHb40	Hes family bHLH transcription factor 2		1	NM_019089(522)	Nucleus (By similarity)	8	0	0	Hes2	gene13331	0.0025872	2.3469	Up
84617	TUBB6	HsT1601|TUBB-5	Tubulin, beta 6 class V	turn composed of alpha- and beta-tubulin polymers. Each microtubule is polarized, at one end alpha-subunitsare exposed (-) and at the other beta-subunits are exposed (+). Microtubules act as a scaffold to determinecell shape, and provide a backbone for cell organelles and vesicles to move on, a process that requiresmotor proteins. The major microtubule motor proteins are kinesin, which generally moves towards the (+) endof the microtubule, and dynein, which generally moves towards the (-) end. Microtubules also form thespindle fibers for separating chromosomes during mitosis.	8	NM_001303526(1230), NM_032525(1341), NM_001303530(903), NM_001303527(1125), NM_001303528(1146), NM_001303529(903), NM_001303524(1341), NM_001303525(315)	Cytoplasm, cytoskeleton (By similarity)	45	0	1	Tubb6	gene30325	2.86E-30	2.3358	Up
152007	GLIPR2	C9orf19|GAPR-1|GAPR1	GLI pathogenesis-related 2		6	NM_022343(465), NM_001287014(219), NM_001287010(387), NM_001287013(510), NM_001287012(189), NM_001287011(315)	Golgi apparatus membrane; Lipid-anchor. Note=Binds lipid-enriched microdomains of Golgi membranes not only by ionic interactions but also through the myristate	22	0	0	Glipr2	gene11917	4.52E-08	2.2159	Up
91543	RSAD2	2510004L01Rik|cig33|cig5|vig1	Radical S-adenosyl methionine domain containing 2		1	NM_080657(1086)	Endoplasmic reticulum membrane; Peripheral membrane protein; Cytoplasmic side. Golgi apparatus. Endoplasmic reticulum. Lipid droplet (By similarity). Mitochondrion. Mitochondrion inner membrane. Mitochondrion outer membrane. Note=Infection with human cytomegalovirus (HCMV) causes relocation to the Golgi apparatus and to cytoplasmic vacuoles which also contain HCMV proteins glycoprotein B and pp28. Interaction with human cytomegalovirus/HHV-5 protein vMIA/UL37 results in its relocalization from the endoplasmic reticulum to the mitochondria	45	0	0	Rsad2	gene13889	0.000000123	2.1947	Up
8676	STX11	FHL4|HLH4|HPLH4	Syntaxin 11	This gene encodes a member of the syntaxin family. Syntaxins have been implicated in the targeting and fusion of intracellular transport vesicles. This family member may regulate protein transport among late endosomes and the trans-Golgi network. Mutations in this gene have been associated with familial hemophagocytic lymphohistiocytosis. (provided by RefSeq, Jul 2008)	1	NM_003764(864)	Membrane; Peripheral membrane protein (Potential). Golgi apparatus, trans-Golgi network membrane; Peripheral membrane protein (By similarity)	44	0	5	Stx11	gene63	0.0000149	2.153	Up

Figure 6.

The validation of candidate genes of interest.

Screening for candidate genes by PubMed searching and SNP sequencing. (A) Three genes (CSRNP1, DUSP5 and LRRC25) were identi-fied by Venny 2.1 from 27 genes detected by HCS and 89 differentially up-regulated genes containing SNPs. (B) The morphology of fetal cortical neuron growth in normal and oxygen-glucose deprivation (OGD) groups. The red arrows indicate fetal neurons. Scale bar: 50 μm. (C–E) The relative expression (2^–ΔΔCt) of CSRNP1, DUSP5 and LRRC25 in normal and OGD groups at 24 hours post OGD. Data are expressed as the mean ± SD, and were analyzed by Student’s t test. *P < 0.05. HCS: High-throughput screening; SNP: single-nucleotide polymorphism.

Discussion

In this study, we performed a comprehensive analysis of SNPs and mRNA expression in cortical samples from rats subjected to HIE and sham operation. RNA sequencing revealed that HIE induced SNPs in mRNAs, which were mainly from genes on chromosomes 1 and 2. Phenotype analysis indicated the SNPs were commonly associated with a G phenotype and scarcely with a W phenotype. In addition, a total of 89 up-regulated genes containing SNPs were found in the HIE group, and were mainly involved in angiogenesis, wound healing biological process and glutamatergic synapse. This is in addition to long-term depression signaling pathways, which were closely correlated with the pathogenesis of HIE. Moreover, intersection analysis identified CSRNP1, DUSP5 and LRRC25 as the most relevant to HIE (Figure 7). These genes were also up-regulated in human neurons after OGD, which may be related to the pa-thology of HIE; therefore, they may be new targets for HIE therapy.

Figure 7.

Flow chart of experimental strategy.

A hypoxic ischemic encephalopathy model was prepared and the neurological score of animals was determined. The cortex of each animal was extracted 24 hours post-operation. RNA was extracted and mRNA-sequencing (seq) libraries were prepared. After quality control of the mRNA library, mRNA-seq and gene ontology (GO) analysis were performed. Finally, bioinformatic analysis was conducted.

HIE pathogenesis is correlated with SNPs detected by RNA sequencing

It is acknowledged that some SNPs are associated with specific disorders and are the main reason for differences in disease susceptibility. A variety of human diseases, including sickle-cell anemia, β-thalassemia and cystic fibrosis are correlated with SNPs (Ingram, 1956; Chang and Kan, 1979; Hamosh et al., 1992). Equally, the seriousness of the disease and the way our body reacts to therapies are also manifestations of genetic variation. A single base variation in poly (ADPRibose) Polymerase-1, for example, is connected with gastrointestinal tumors (Martín-Guerrero et al., 2017); the rs1495741 genetic variant and smoking are strongly associated with the risk of bladder cancer (Ma et al., 2016); and the aldehyde dehydrogenase 2 Glu504Lys SNP is a candidate risk factor for a wide range of chronic diseases, including cancer, cardiovascular disease, and late-onset Alzheimer’s disease, linked to lifestyle factors such as alcohol consumption and the presence of other genetic variations (Zhao and Wang, 2015). In addition, accumulating evidence shows that lung injury is associated with genetic variations (Wang et al., 2010a; Trittmann et al., 2014, 2016; Cho et al., 2015; Liu et al., 2016). In the present study, a large number of SNPs in HIE rats were detected via RNA-seq, indicating that HIE-induced brain injury was associated with the SNPs. Moreover, further study found 89 genes exhibiting SNPs that were up-regulated after HIE.

Gene ontology analysis of the up-regulated genes containing SNPs

A bioinformatics analysis was performed to predict the potential functions of the differentially expressed mRNAs containing SNPs in HIE rats. GO analysis categorized genes functions according to biological process, cellular com-ponent, and molecular function. Kyoto encyclopedia of genes and genomes analysis clarified the potential signaling pathways that genes might participate in. Compared with the control group, 89 genes exhibiting SNPs were up-regulated in the brain from the HIE group. This indicated that the effect of HIE-induced brain injury was a complex multigene process. The inflammatory response, together with excitotoxic and oxidative responses, are major con-tributors to ischemic injury in both the immature and adult brain (Ruscher et al., 2013; Fernández-Tajes et al., 2014). Angiogenesis is an important process in the recovery of brain ischemia-induced injury (Zheng et al., 2018). In our study, the GO analysis revealed that SNP-containing genes that were up-regulated 24 hours after HIE were mainly enriched for GO terms associated with the angiogenesis, wound healing, negative regulation of cytokine-mediated signaling pathway and negative regulation of innate immune response. Therefore, together with previous studies, our results revealed that SNPs in mRNA may influence biological processes after HIE.

A molecular function analysis revealed that the 89 up-regulated genes containing SNPs were mainly involved in growth factor activity, protein kinase binding, transcription regulatory region DNA binding, and intracellular calcium-activated chloride channel activity. The human genome contains about 560 protein kinase genes, accounting for about 2% of all human genes (Manning et al., 2002). Furthermore, kinases regulate the majority of cellular pathways, especially those involved in mechanistic cellular signaling and signal transduction (Murphy et al., 2014; Eyers and Murphy, 2016). This suggests that the up-regulated genes containing SNPs were also involved in growth factor activity, which is important for recovery from HIE.

HIE typically results in serious long-term sequelae, mainly because of damage caused to neurons or axons in the acute phase (Busl and Greer, 2010; Shankaran et al., 2012). Therefore, long-term neurological deficits from HIE are partially caused by inhibited axon regrowth. In the current study, pathway analysis indicated that the 89 up-regulated genes containing SNPs were mainly enriched in glutamatergic synapse and long-term depression signaling pathways, indi-cating that SNPs in mRNAs also participated in pathways involved in long-term neurological function.

CSRNP1, DUSP5 and LRRC25 are potential targets for HIE therapy

To select the most relevant genes involved in HIE, we performed high-throughput screening by searching PubMed and we used PCR to verify candidates. We identified 27 genes that have not been studied in HIE. In addition, 89 up-regulated genes containing SNPs were found in the HIE group by RNA sequencing. We intersected the 27 factors from high-throughput screening and the 89 up-regulated genes exhibiting SNPs using Venny. Three genes, CSRNP1, DUSP5 and LRRC25 were identified. qRT-PCR showed that the relative expression of CSRNP1, DUSP5 and LRRC25 was up-regulated in OGD-treated human fetal cortical neurons group compared with the control group. These data indicate that the three genes are candidate targets for HIE therapy, and they will inform further research on HIE pathogenesis.

In conclusion, this study’s findings indicate that HIE is accompanied with an increased number of SNPs, which often exhibited in G phenotype and rarely the W phenotype. Additionally, 89 up-regulated genes containing SNPs were involved in angiogenesis, wound healing, negative regulation of cytokine-mediated signaling pathway, negative regulation of innate immune response and palate development, which may contribute to the pathogenesis and biochemical characteristics of HIE in neonatal rats. Finally, CSRNP1, DUSP5 and LRRC25 were verified by high throughput screening as the most relevant genes containing SNPs to HIE. However, we did not investigate the influence of SNP overexpression or knockout on HIE in the current study. In spite of this, our results provide robust evidence for advancing the development of HIE therapy.

Additional files:

Additional Figure 1 ^{(438.8KB, tif)} : The correlation analysis of linear regression among weight, Zea-longa score and the number of SNPs found that there is no significant diffidence between the weight and the number of SNPs as well as the Zea-longa score in HIE group.

Additional Figure 1

The correlation analysis among weight, Zea-longa score and the number of SNPs found that there is no significant diffidence between the weight and the number of SNPs as well as the Zea-longa score in HI group.

(A, B) Correlation linear analysis among the weight, Zea-longa score and the number of SNPs. Two points on the line represent sham group, the two points of dispersion represent HI group.

Additional Table 1: The total number of single-nucleotide polymorphisms (SNPs) in sham and hypoxic ischemic encephalopathy (HIE) groups.

Additional Table 2 ^{(92.9KB, pdf)} : The number of single-nucleotide polymorphisms (SNPs) in all of samples.

Additional Table 3 ^{(23.5MB, pdf)} : The number of single-nucleotide polymorphisms (SNPs) in A, C, G, K, M, R, S, T, W, Y genotypes.

Additional Table 4: The upregulated genes in hypoxic ischemic encephalopathy (HIE) compared with the sham group.

Additional Table 5: Advances in genes research.