Diabetes was chemically induced in 10-week-old male Sprague-Dawley rats (n=30; weight, 230–280 g; Chengdu DaShuo Biotech Co., Ltd. Chengdu, China) with STZ (Sigma-Aldrich, St. Louis, MO, USA). The rats received a single intraperitoneal injection of a freshly prepared solution of STZ in citrate buffer (0.01 M; pH 4.5; Junrui Biological Technology Co., Ltd., Shanghai, China) at a dose of 60 mg/kg bodyweight. A corresponding number of weight and age-matched animals (n=30) were maintained as controls, and these rats received a single intraperitoneal injection of citrate buffer only, at a dose of 60 mg/kg bodyweight. The diabetic status of the animals was confirmed by measuring the blood glucose levels of the rats using a glucometer (ACCU-CHEK^® Performa; Roche Diagnostics GmbH, Mannheim, Germany), with fasting blood glucose levels >16.7 mmol/l considered to indicate diabetic conditions. Insulin was not administered to the animals. The rats were housed in a controlled environment at 20–25°C with a 12 h light/dark cycle, and were provided with ad libitum access to food and water. All the animals were maintained and handled in accordance with the guidelines of the Association for Research in Vision and Ophthalmology statement for the Use of Animals in Ophthalmic and Vision Research (7). The present study was approved by the Animal Protection and Ethics Committee of Sichuan University (Chengdu, China).

Following dark adaptation for 60 min in a box, six rats from either the experimental group (n=30) or control group (n=30) received intraperitoneal injections of chloral hydrate solution (Junrui Biological Technology Co., Ltd.) at a concentration of 0.3 ml/100 g bodyweight, and had their pupils fully dilated with compound tropicamide eye-drops prior to assessment. Lidocaine hydrochloride (2%; Shanghai Fosun Zhaohui Pharmaceutical Co., Ltd., Shanghai, China) was applied to the animals for retinal surface anesthesia. The cornea-touch electrode, reference electrode and grounding electrode were placed at the corneal margin of the eye, forehead and end of the tail under the skin, respectively.

The stimulator used was a Ganzfeld full-field (SG-2002; LKC Technologies, Inc. Gaithersburg, MD, USA) dome stimulator, as recommended by the standard F-ERG guidelines provided by the International Society for Clinical Electrophysiology of Vision (ISCEV) in 2004 (8). The stimulus intensity of the standard flash was 2.448 cdxs/m². F-ERG retinal function was assessed using a visual electrophysiological system (MEB9200; Nihon Kohden, Tokyo, Japan). The amplitude and implicit duration of each wave form were analyzed.

After 16 weeks, the rats were anesthetized with 10% chloral hydrate at a dose of 0.3 ml/100 g bodyweight by intraperitoneal injection. The rats were then sacrificed by overdose with 10% chloral hydrate following the removal of the eyes. Each retina was immediately dissected from the eye under a dissecting microscope (model SX-4; Guangzhou Ming-Mei Technology Co., Ltd., Guangzhou, China). These samples were then frozen in liquid nitrogen (Sichuan Qiaoyuan Gas Co., Ltd., Sichuan, China) and stored at −80°C. Microsurgical scissors were used to section the eye along the corneoscleral limbus. The retinas were then immediately dissected from the eye using the SX-4 dissecting microscope. The retinas of the rats were placed into a nuclease-free frozen storage tube prior to being frozen in liquid nitrogen overnight and stored at −80°C. The rat retinas were then placed into a nuclease-free microcentrifuge tube with plastic grinding rods. Retinal RNA was extracted using 1 ml TRIzol reagent (Invitrogen; Thermo Fisher Scientific, Waltham, MA, USA) and quantified at an absorbance of 260 nm using an ultraviolet-visible (UV-Vis) spectrophotometer (NanoDrop 8000; Thermo Fisher Scientific, Inc.). Its integrity was determined using an Agilent 2100 Bioanalyzer (G2939AA; Agilent Technologies, Santa Clara, CA, USA).

The total RNA samples were first treated with DNase I (New England Biolabs, Inc., Ipswich, MA, USA) to degrade any possible DNA contamination, and the digestion products were then purified with magnetic beads (Dynabeads^® mRNA Purification kit; Invitrogen; Thermo Fisher Scientific, Inc.). Subsequently, the mRNA was enriched using oligo (dT) magnetic beads (for eukaryotes; Dynabeads^® mRNA Purification kit). The mRNA was then mixed with fragmentation buffer (Ambion^® RNA Fragmentation Reagents; Ambion; Thermo Fisher Scientific, Inc.) and fragmented into short fragments (~200 bp). The first strand of cDNA was then synthesized using random hexamer-primed reverse transcription (Super Script^® II Reverse Transcriptase; Invitrogen; Thermo Fisher Scientific, Inc.) according to the manufacturer's protocol. Second strand buffer (Invitrogen; Thermo Fisher Scientific, Inc.), deoxynucleotide triphosphates (New England Biolabs, Inc.), RNase H (Second Strand Master Mix; Invitrogen; Thermo Fisher Scientific, Inc.) and DNA polymerase I (Second Strand Master Mix; Invitrogen; Thermo Fisher Scientific, Inc.) were added to synthesize the second strand. The double strand cDNA was then purified with magnetic beads. End reparation was then performed, and the adaptors were then ligated to the ends of the fragments using a ClaSeek Library Preparation kit (Thermo Fisher Scientific, Inc.) according to the manufacturer's protocol. The ligation products were selected by size and purified on a Tris-acetate-EDTA-agarose gel (Sigma-Aldrich). Finally, the fragments were enriched by polymerase chain reaction (PCR) amplification using Platinum^® Pfx DNA Polymerase (Invitrogen; Thermo Fisher Scientific, Inc.) and GeneAmp^® system 9700 (Applied Biosystems; Thermo Fisher Scientific, Inc.) according to the manufacturer's protocol. PCR products were purified with magnetic beads and dissolved in 50 µl Epstein-Barr solution (Agencourt^® AMPure^® XP Beads-PCR Purification; Beckman Coulter, Inc., Brea, CA, USA). During the quality control step, an Agilent 2100 Bioanalyzer was used to qualify and quantify the sample library.

The library products were prepared for sequencing with an Ion Proton platform (Ion Torrent™; Thermo Fisher Scientific, Inc.). Data filtering was performed to obtain high-quality reads, and the clean reads were saved as '.bam' files and used as the original sequencing results. All the sequence reads were mapped to the reference genome sequences. The maximum number of mismatches allowed for the mapping was set at two.

The expression level for each gene was determined by the number of reads uniquely mapped to the specific gene, and by the total number of uniquely mapped reads in the sample. To determine the expression levels of various genes and compare them between samples, the variable read per kilobase of exon per million mapped reads (RPKM) method was used (9). The statistical significance of the differential expression of each gene was determined, according to the P-value. Fold-change (FC) differences between the control and diabetic groups of rats were calculated. The P-values were adjusted for multiplicity to control the false discovery rate (FDR). Differentially expressed genes (DEGs) were defined as those with an |FC| >2 and FDR <5%.

GO enrichment analysis provides all the GO terms, which are significantly enriched among DEGs, compared with the background genome, and filters the DEGs that correspond to biological functions (10). This method first maps all the DEGs to GO terms in the GO database (http://www.geneontology.org/) and calculates the numbers of genes for every term. It then uses a hypergeometric test to identify significantly enriched GO terms among the DEGs in relation to the background genome. Using non-redundant (NR) annotation, the Blast2GO version 3.0 program (Biobam, Valencia, Spain) was used to obtain GO annotations for the DEGs. Subsequently, WEGO software (http://wego.genomics.org.cn/cgi-bin/wego/index.pl) (11) was used to generate GO functional classifications for the DEGs and determine the distribution of gene functions of a species from the macro level.

Genes usually interact during certain biological functions, and a pathway-based analysis can assist in establishing the the biological functions of a gene (12). KEGG is a major public pathway-associated database (13). Pathway enrichment analysis identifies significantly enriched metabolic pathways and signal transduction pathways in DEGs, relative to the whole background genome.

To validate the RNA-seq findings in the present study, 10 genes, identified as differentially expressed, were randomly selected for RT-qPCR validation. The RT-qPCR was performed using eight samples of purified RNA, including four from the diabetes group and four from the control group. Total RNA was first reverse-transcribed into cDNA using a RevertAid First Strand cDNA Synthesis kit (Thermo Fisher Scientific, Inc.) according to the manufacturer's instructions. The primers were designed using Primer Express 3.0 (v. 0.4.0; http://bioinfo.ut.ee/primer3-0.4.0/) and the sequences are listed in Table I. GAPDH was used as a reference control. Subsequent qPCR was performed with a Mastercycler^® ep realplex (Eppendorf, Germany) using SYBR Premix Ex Taq™ II (Takara Bio, Inc., Otsu, Japan) in accordance with the manufacturer's protocol. The qPCR was conducted with the following thermal cycling conditions: 95°C for 30 sec, followed by 40 cycles of 95°C for 5 sec and 60°C for 30 sec, and three replicates were performed for each amplification. Relative quantification analyses were performed using the comparative Cq method, and relative gene expression levels were calculated using the 2^−∆∆Cq method (14).

Independent sample t-tests were performed to assess for any significant differences in the measured variables between the control and diabetic groups. All values are reported as the mean ± standard error of the mean, unless otherwise stated. Data were analyzed using Statistical Package for the Social Sciences software, version 18.0 (SPSS, Inc., Chicago, IL, USA). P<0.05 was considered to indicate a statistically significant difference.

All the STZ-treated rats exhibited characteristics of diabetes, and their blood glucose levels were significantly higher, compared with those in the age-matched control rats. However, changes in weight occurred more slowly in the diabetic rats, compared with the age-matched control rats, following STZ treatment (Table II) and showed symptoms of polyuria. The age-matched control rats had normal glucose levels, showed no signs of polyuria and gained weight consistently until the animals were sacrificed.

The ERG responses were defined according to the current ISCEV standards (15). The a-wave is a fast negative ERG component, obtained primarily from the maximal combined response. The leading edge reflects the membrane current in the photoreceptors. The a-wave amplitude is measured from the baseline to a-wave trough. The b-wave is a large positive ERG potential. Under scotopic conditions, the rising phase (up to the peak) is directly generated by bipolar cells and Müller cells. The implicit duration of the b-wave is measured from flash onset to the peak of the b-wave (16–18). The oscillatory potentials (OPs) are three to five low-amplitude, high-frequency wavelets, superimposed on the ascending limb of the ERG b-wave. The OPs are considered to result from feedback between the amacrine cells and bipolar cells, and/or feedback from ganglion cells to amacrine cells. The peak to trough amplitudes of the waves are frequently combined to provide an overall measure of the OP amplitude. Alternatively, the amplitudes of individual wavelets may be recorded. The latter may be preferable, as OPs are complex in origin, and individual wavelets may be generated at different sites within the retina (19,20).

The amplitudes of the a-wave, b-wave, OP1, OP2 and ∑OP were all significantly reduced (Table III) in the diabetic group, compared with the control group (P<0.05). Furthermore, the implicit b-wave durations in the diabetic group were significantly longer (Table IV) than those in the control group (P<0.001). No significant differences were found in the implicit durations of the a-wave, OP1, OP2, OP3 or ∑OP between the control and diabetic groups.

To obtain triplicate results, three samples were obtained from the control and the diabetic groups of animals, with each sample obtained from a pair of rat retinas. In total, six RNA-Seq libraries were constructed, and over 14,000,000 clean reads were generated in each library. Subsequently, ~87.9% of the total reads were mapped to the reference genome. The detailed mapping statistics are listed in Table V.

The triplicate samples from the control and diabetic groups were assayed for DEGs (Fig. 1), and a total of 868 genes were found to be differentially expressed, with 565 upregulated genes and 303 downregulated genes. The 10 most markedly upregulated and downregulated genes are listed in Table VI.

The GO classification comprises cellular component, molecular function and biological process domains. Based on sequence homology, the 868 DEGs identified in the present study were categorized into 49 functional groups. In the cellular component, molecular function and biological process GO classification categories, 26, 12 and 11 functional groups were identified, respectively (Fig. 2). Among these groups, the cellular process, cell and cell parts, and binding were the predominant in each of the three categories. A high percentage of the altered genes in the biological processes in the DEGs were involved in response to stimulus (349/637 genes; 54.8%), regulation of biological processes (331/637 genes; 52.0%) and metabolic processes (331/637 genes; 52.0%). Furthermore, 94 apoptotic and apoptotic regulatory genes, and 19 inflammatory genes, were also detected as having changed.

The KEGG pathway significant enrichment analysis revealed that the DEGs were involved in 217 pathways, including CAMs, complement and coagulation cascades, and antigen processing and presentation.

To validate the RNA-seq findings, the present study prepared rat retinas from separate groups of rats in each group qPCR analysis. In total, 10 genes were selected, comprising Gryaa, Gryab, Htra2, Pitx3, Fgf2, Casp3, Stat3, Fabp7, Jak3, Xiap and Gapdh. Changes in the expression levels of these genes were determined using RT-qPCR. These results were similar to those obtained following the RNA-seq.