The Animal Care and Use Committee of the Sichuan Provincial People’s Hospital authorized all animal experiments, which followed the ARVO guidelines for the use of animals in ophthalmology and vision research.

Asrgl1 KO mice were generated by using the CRISPR/Cas9 system on C57BL/6J genetic background. The Cas9 mRNA and two single guide RNAs targeting a region upstream of exon 3 and downstream of exon 4 of Asrgl1 exon 3 were microinjected into mouse zygotes (Figure 1). gRNA1 sequence (matching reverse strand of gene) was as follows: TCT​GTC​GCT​AGC​AAC​CAA​CAA​GG. gRNA2 sequence (matching forward strand of gene) was as follows: ATA​TTG​CTA​GAC​AGG​CTT​GGA​GG. As confirmed by Sanger sequencing and PCR genotyping analysis, the conventional translation start site was successfully eliminated.

Mouse ERG detection was performed on an Espion Visual Electrophysiology System (Diagnosys, Lowell, MA, United States). Briefly, female Asrgl1 KO mice and female controls were dark-adapted overnight 1 day before detection. In the next morning, before ERG, animals were anesthetized and the eyes were dilated with a drop of tropicamide and phenylephrine, as well as tetracaine (0.5%). Throughout the experiment, a heating platform was used to keep the body temperature at 37°C. Gold wire loops were used to record dark-adapted ERGs in response to flashes with intensities ranging from 0.003 to 20 cds/m². After 20 min of complete light adaptation, cone-mediated ERGs were recorded with white flashes.

The eyes with the whole retina and optic nerve were chosen for imaging by hematoxylin and eosin (H&E) staining. Female wild-type (WT) and female KO mice’s eyes were enucleated, marked for orientation on the nasal side, fixed overnight in 1.22% glutaraldehyde and 0.8% paraformaldehyde in 0.08 M phosphate buffer, embedded in paraffin, and cut into 5-μm slices. H&E staining was used on sections taken from five places around the optic nerve. The rows of photoreceptors in the outer nuclear layer were counted using H&E-stained slices. Every 200 μm from the optic nerve, three measurements of the outer nuclear layer were recorded and averaged. The optic nerve was given the number 0 mm.

Enucleated eyeballs were removed from female WT and female Asrgl1 KO mice for immunohistochemistry, marked on the nasal side for orientation, fixed for 1 h in 4% paraformaldehyde in 100 mM phosphate buffer (pH 7.4), and cryoprotected in 30% sucrose. For sectioning, tissues were immersed in an optimal cutting temperature (OCT) solution and frozen on dry ice. For 30 min, sections were blocked and permeabilized in phosphate buffer with 10% normal goat serum and 0.2% Triton X-100, then labeled with different primary antibodies overnight at 4°C. Then, the slides were washed and incubated with secondary antibodies for 2 h. Finally, the slides were washed and covered with coverslips. The primary antibodies used in our study were as follows: rhodopsin (1D4, cat. #MA1-722, Thermo Fisher, Waltham, MA, United States), Na-K pump (cat. #MA3928, Thermo Fisher, Waltham, MA, United States), opsin (red/green, cat. #AB5405, Millipore, Burlington, MA, United States), and 594-conjugated peanut agglutinin (PNA) (cat. #RL1072, Vector laboratories, Burlingame, CA, United States). Secondary antibodies (Alexa Fluor 488 and Alexa Fluor 594) and DAPI (cat. #D8417, Sigma, St Louis, MO, United States, 1:2,000 dilution) were also used in the experiments. Confocal microscopy was used to image eye sections (LSM800, Carl Zeiss, Jena, Germany).

TRIzol (Invitrogen, Austin, TX, United States) was used to extract total RNA from the retinas of female Asrgl1 KO mice and female normal controls, according to the manufacturer’s protocol. cDNA was made as previously described. An Eppendorf Mastercycler personal PCR equipment (Eppendorf, Germany) was used to amplify Asrgl1 genotyping products, which were then analyzed on a 1.5% agarose gel electrophoresis. Using the SYBR PCR Master Mix kit (Applied Biosystems, Foster City, CA) and the 7500 Fast Real-Time PCR detection machine, equal amounts of cDNA from the retinas of Asrgl1 KO mice and normal controls were submitted to PCR. Specific primers used in our study were listed in Supplementary Data S1.

RNA was isolated from the retinas of female Asrgl1 KO mice and female normal controls, and the integrity of the RNA was determined using the Bioanalyzer 2100 system’s RNA Nano 6000 Assay Kit (Agilent Technologies, CA, United States). The library was then produced and tested for transcriptome sequencing using the Agilent Bioanalyzer 2100 instrument. The index-coded samples were clustered using the TruSeq PE Cluster Kit v3-cBot-HS (Illumina) on a cBot Cluster Generation System according to the manufacturer’s instructions. The library preparations were sequenced on an Illumina Novaseq device after cluster creation, yielding 150-bp paired-end reads. Raw data (raw readings) in fastq format were processed using in-house perl scripts, resulting in clean, high-quality data. The level of gene expression was measured after reads were mapped to the reference genome. All of the downstream analyses were based on clean, high-quality data. Prior to differential gene expression analysis, the read counts for each sequenced library were modified by one scaling normalization factor using the edgeR tool package. The edgeR R package was used to perform differential expression analysis of two situations. The P values were adjusted using the Benjamini and Hochberg method. The criterion for significant differential expression was set at 0.05 corrected P-value and log2 fold change. The raw data have been uploaded to the Genome Sequence Archive (https://ngdc.cncb.ac.cn/gsa/browse/CRA005193), and the assigned accession number was CRA005193.

Top biological functions as well as canonical pathways associated with DEGs generated from the RNA-seq between the Asrgl1 KO retina and normal controls were identified by Ingenuity pathway analysis (IPA) (Qiagen). Fisher’s exact test was used to determine the likelihood of any biological function or canonical pathway being explained only by chance. Algorithmically, molecular interaction networks are created based on the connectivity of the molecules. Fisher’s exact test was used to determine network scores, which corresponded to −log 10 (p value).

Statistical significance was established using Prism 7.0 software and the Student’s t-test or one-way ANOVA (GraphPad Software, La Jolla, CA, United States). The data are presented as the mean SEM unless otherwise noted in the text and figure legends. The ERG datasets in Figure 2 and the ONL thickness measurement dataset in Figure 3 were subjected to ANOVA tests. The number of animals used in each experiment is indicated by n. In the figure legends, the total number of animals used in each experiment is also listed. The following p values are indicated by asterisks in the figures: *≤0.05, **≤0.01, and ***≤0.001.

ASRGL1 is widely expressed in all human tissues studied, with the highest levels of expression in the brain, female tissues including the uterine cervix and fallopian tube, and male tissues like the testis (Edqvist et al., 2015). To investigate the functions of AsrgL1 in vivo, Asrgl1 KO were generated using CRISPR/Cas9 technology. The target sites were selected as exon 3 and exon 4 (Figure 1A). Genotyping analysis using PCR primers flanking the deletion site revealed the successful generation of Asrgl1 knockout mice (Supplementary Figure S1). To validate the genomic Asrgl1 ablation, mouse DNA samples were subjected to Sanger sequencing analysis, and results showed genomic deletion of 1,388 bp in Asrgl1 KO mice (Figure 1B). Both Sanger sequencing and electrophoresis analysis of cDNA product from Asrgl1 KO retinas revealed a 301-bp deletion of exon 3 and exon 4 in Asrgl1 KO (Figure 1C–E). All the above data showed the successful generation of Asrgl1 KO by CRISPR/Cas9 technology.

ERG examination was first performed to test the in vivo function of Asrgl1 in the retina. In three 6-month-old Asrgl1 KO mice and littermate controls, no abnormal ERG response was observed (data not shown). However, all Asrgl1 KO mice exhibited reduced scotopic and photopic responses at 8 months of age. The mean a- and b-wave amplitudes of Asrgl1 KO were reduced by approximately 15% and 30%, respectively, in scotopic response tests (Figure 2A1–A4), which indicated altered rod cell function. When compared to WT mice, Asrgl1 KO had a somewhat flat wave under light-adapted conditions, showing that cone cell function was also impaired in Asrgl1 KO (Figure 2B1–B3). Although, by H&E staining, the retinas from Asrgl1 KO mice revealed no difference in the retinal ONL thickness at 6 months of age (Supplementary Figure S2), the thickness of the ONL started to decrease around 9 months of age (Figure 3A1, A2) and gradually intensified at 12 months (Figure 3B1, B2) and 15 months of age (Figure 3C1, C2). The photoreceptor layer including the inner segment and outer segment decreased around 9 months (Figure 3A1, A4) and intensified at 12 months (Figure 3B1, B4) and 15 months of age (Figure 3C1, C4). There was no obvious change in the inner nuclei layer (INL) of Asrgl1 KO at 9 months of age (Figure 3A1, A3), and the thickness of the INL in Asrgl1 KO was slightly decreased at 15 months of age (Figure 3C1, C3).

To investigate the pathological changes underlying the Asrgl1 knockout mice’s retinal degeneration phenotype, immunostaining was performed with retinal cryosections in Asrgl1 KO and WT mice. There were no changes in rhodopsin and M-opsin expression in Asrgl1 KO at 8 months of age (Supplementary Figure S3A, B). However, decreased rhodopsin protein was observed in the photoreceptor of 12- and 15-month-old Asrgl1 KO mice, respectively (Figure 4A, C), compared to WT mice, which indicated a shortened outer segment of 12- and 15-month-old Asrgl1 KO mice. Rhodopsin is a key outer segment protein that is required for adequate vision, and its manufacture and transit are tightly regulated (Palczewski, 2006). Decreased rhodopsin is cytotoxic and leads to rod death. Decreased Na-K ATP staining in 15-month-old Asrgl1 KO mice indicated shortened inner segment. M-opsin, a cone cell marker, and PNA (which binds to the sheaths of the cone matrix in all three types of cones) were also used to stain retinal sections. The number of cone cells was reduced in 12-month-old Asrgl1 KO mice and drastically decreased in 15-month-old Asrgl1 KO mice (Figure 4B,D). A TUNEL assay was utilized to determine whether the retinal degeneration was caused by apoptosis. There were no changes in TUNEL-positive cells in 8-month-old Asrgl1 KO mice (Supplementary Figure S3C). TUNEL-positive cells were substantially more abundant at 15 months of age in Asrgl1 KO retinas than in WT retinas (Figure 4E,F), and TUNEL-labeled nuclei were mostly found in the ONL, indicating continuous photoreceptor cell death in Asrgl1 KO retinas.

To investigate the transcriptional effect of Asrgl1 ablation, mouse retinal tissues from 8-month-old Asrgl1 KO (n = 3) and WT mice (n = 4) were subjected to RNA-seq analysis. There were 149 DEGs identified in Asrgl1 mice (Supplementary Table S2), among which 89 were upregulated (p ≤ 0.01 and log2 ratio ≥1) and 60 (p ≤ 0.01 and log2 ratio ≤−1) were downregulated. Volcano plots of differentially expressed genes are shown in Figure 5A. IPA was used to identify numerous highly enriched biological processes related with DEGs in order to investigate the biological functions enriched in DEGs (Table 1). Gastrointestinal disease (p value: 4.79E−02 to 2.61E−05, 13 genes assigned), organismal injury and abnormalities (p value: 4.94E−02 to 2.61E−05, 31 genes assigned), and cancer (p value: 4.79E−02 to 3.23E−04, 18 genes assigned) were discovered as the top three highly enriched terms in the perspective of diseases and disorders. Cellular development (p value: 4.16E−02 to 3.14E−05, seven genes assigned), cell morphology (p value: 4.79E−02 to 1.07E−03, six genes assigned), and cell death and survival (p value: 4.94E−02 to 1.22E−03, seven genes assigned) were found as the top three highly enriched terms in the perspective of molecular and cellular functions. Digestive system development and function (p value: 3.38E−02 to 2.61E−05, seven genes assigned), organ morphology (p value: 4.48E−02 to 2.61E−05, 13 genes assigned), and hematological system development and function (p value: 4.60E−02 to 3.14E−05, seven genes assigned) were shown as the top three enriched terms in the perspective of physiological system development and functions.

Canonical pathway analysis could inform the key metabolism and signaling pathways in which the DEGs may be involved. The main metabolism and signaling pathways in which DEGs may be engaged could be informed via canonical pathway analysis. In this study, canonical molecular pathways enriched by DEGs were explored by bioinformation analysis of IPA software. Figure 5B shows the top five canonical pathways in the retina of Asrgl1 KO mice that are significantly associated with DEGs. The glucocorticoid receptor signaling (p value = 9.36E−06), intrinsic prothrombin activation pathway (p value = 2.78E−04), and xenobiotic metabolism AHR signaling pathway (p value = 2.04E−03) were the top three major pathways altered in retinal tissues in Asrgl1 KO mouse. The molecules that were changed in the glucocorticoid receptor signaling were IL1β, KRT5, KRT7, KRT13, KRT14, KRT15, KRT19, KRT80, and KRT6B. Our results firstly suggested an association of Asrgl1 with GR signaling in the pathological process of retinitis pigmentosa.

The molecular interaction networks were then created and ranked based on the connectivity of the detected DEGs. The network of “Cell Morphology, Embryonic Development” (score = 49) (Figure 6) was the most enriched network alerted in retinal tissues of Asrgl1 KO mice. Notably, several genes including NFE2L2, IL-4, Foxp3, and Fos were the core “nodes” with the most connections in the network. To validate the expression levels of the main altered genes in glucocorticoid receptor signaling pathway and network of “Cell Morphology, Embryonic Development,” real-time PCR was performed. As Figure 7 shows, the expression of KRT5, KRT7, KRT13, KRT14, KRT15, KRT19, and KRT80 in glucocorticoid receptor signaling pathway all increased in Asrgl1 knockout mice compared to the normal controls. Interestingly, only the expression of Foxp3 was decreased among the selected “central node” genes, namely, NFE2L2, IL-4, Foxp3, and Fos. As Foxp3 is a transcriptional regulator that is required for regulatory T-cell development and inhibitory activity (Treg) (Li et al., 2015), its function in Arsgl1-related RP needs to be explored further in follow-up studies.