CRISPR/Cas9-mediated GJA8 knockout in rabbits recapitulates human congenital cataracts

Cataracts are the leading cause of vision loss in the world, although surgical treatment can restore vision in cataract patients. Until now, there have been no adequate animal models for in vivo studies of artificial lens safety and drug interactions. Genetic studies have demonstrated that GJA8 is involved in maintaining lens opacity and proper lens development. In this study, a cataract model with GJA8 gene knockout was developed via co-injection of Cas9/sgRNA mRNA into rabbit zygotes. Our results showed that gene mutation efficiency in the GJA8 locus reached 98.7% in embryos and 100% in pups, demonstrating that the Cas9/sgRNA system is a highly efficient tool for gene editing in rabbits. In agreement with other studies, our genetic and histology results showed that impaired GJA8 function caused microphthalmia, small lens size and cataracts. In summary, our novel rabbit model of cataracts will be an important drug-screening tool for cataract prevention and treatment.


Results
CRISPR/Cas9-mediated gene targeting of GJA8 in zygotes. In order to disrupt GJA8 in rabbits, two sgRNAs targeting the CDS of GJA8 were designed (Fig. 1A). To clone the sgRNA sequence into the pUC57-T7-gRNA vector, a Bbs I enzyme cut site was added beside the complementary DNA oligonucleotides (Table S1).
To determine the efficiency of the CRISPR/Cas9 system in modifying the GJA8 gene, in vitro transcribed mRNA from Cas9 and sgRNAs was microinjected into the zygote and cultured until blastocyst stage. As shown in Table 1, 85.6% of injected embryos developed to blastocyst stage, indicating that there were no significant differences in cleavage and development rate between non-injected and microinjected embryos. To determine the mutation efficiency, genomic DNA was extracted from a single blastocyst, and the PCR products were sequenced. As shown in Fig. 1B and Table 1, mutation efficiency was as high as 98.7% in the injected blastocyst. This result was confirmed by the T7E1 assay and PCR Sanger sequencing data. Together, this indicated that dural sgRNA directed CRISPR/Cas9 system was efficiently mutated rabbit GJA8 in our study (Fig. 1C).
Generation of GJA8 knockout rabbits via zygote injection. A total of 52 and 58 injected zygotes (pronuclear stage) were transferred into the oviducts of two surrogate rabbits ( Table 2). After 30 days gestation, two recipient mothers gave birth to 11 rabbit pups. Genomic DNA from ears was isolated and used to detect  mutations by PCR and sequencing. As expected, all 11 pups had mutated GJA8, and the indels in the founders ranged from 12-81 bp deletions ( Fig. 2A and Table S2). This result was also confirmed by T7E1 assay, which showed that all F0 rabbits (100%) had the GJA8 mutation. In addition, genotype data indicated that 45.5% of F0 rabbits had the GJA8 mosaicism (F0-1, F0-5, F0-7, F0-10 and F0-11), revealing that CRISPR/Cas9 generated mosaicism was commonly detected in F0 animals (Fig. 2B). Since the GJA8 is mainly expressed in eye lenses, we therefore performed a chimera analysis on lenses from the F0-8 rabbit. As shown in Fig. S1A, the same mutated sequences and cleavage bands were found in between the lenses and ears, suggesting that no chimera mutations were presented in the eyes of GJA8 mutated rabbit. We then examined whether the gene mutations reduced of protein levels or changed the cataract phenotype. As shown in Fig. 2C, protein levels were reduced in the lenses of adult GJA8 (+ /−) rabbits when compared to WT counterparts. Photographs showed eight of 11 F0 GJA8 mutated rabbits appeared to develop cataracts with lens opacities (Fig. 2D, Fig. S1B), when compared to WT littermates. Additionally, the histological H&E staining showed that the lens inner fiber cells of GJA8 mutants lens were severely distorted (Fig. 2E)   well-aligned inner fiber cells of cataract-free WT rabbits. These observations confirmed that GJA8 mutations affected lens fiber cells during embryonic lens development and provide a novel animal model that recapitulates human congenital cataracts.
Off-target analysis in the F0 of GJA8 gene knockout rabbits. One of the major concerns when using the CRISPR/Cas9 system is off-target mutagenesis, which has been widely reported in human cell lines 22 , mice 20 and zebrafish 23 . To test whether off-target mutagenesis occurred in the GJA8 knockout rabbits, we performed Sanger sequencing and the T7E1 cleavage assay on the PCR products from 7 POTS. As shown in Fig. 3A-D, none of the sequencing reads had mutations, suggesting that off-target mutagenesis was eliminated by the co-injection of Cas9 and sgRNAs into rabbit zygotes. The information about POTS was listed in Table S3 Genotype and phenotype of the F1 GJA8 knockout rabbits. To study whether the deletions or indels were heritable, genotypes of the F1 pups (F0-7 × F0-4, F0-5 × F0-8 and F0-7 × F0-9) were determined by PCR and T-cloning Sanger sequencing. As shown in Fig. 4A-C and Table S4, all of the F1 rabbits had the mutation. The F1-1 rabbit was a bi-allelic mutants (− /− ) while the others were mono-allelic mutant (+ /− ). The T7E1 cleavage assay confirmed this result (Fig. 5A). To determine whether the GJA8 protein had been disrupted, equal amounts of protein from a WT and F1-23 rabbit were used for western blots. As shown in Fig. 5B, GJA8 protein was detected in the WT but sharply decreased in the F1-23 rabbit. In addition, the predicted 3D models showed GJA8 protein structure was obviously disrupted in the F1-4 and F1-1 rabbits, compared with WT rabbit (Fig. 5C). Taken together, these data demonstrating the deletions or indels of GJA8 were inheritable in our cataract rabbit model. The phenotype of our cataract rabbit model was also examined in the F1 generation. In contrast to WT, GJA8 (+ /− ) rabbits showed opacity, smaller lens and obvious microphthalmia (Fig. 5D,E and G and Fig. S1C). Disorganized fiber cell layers were also observed in the founder of our cataract rabbit model (Fig. 5F). Disrupted gap junctions in lens fibers of GJA8 knockout rabbits. In order to investigate whether gap junctions were affected by GJA8 mutation, the gap junctions were examined by thin-section immunolabeling and electron microscopy in GJA8 (+ /− ) and WT rabbits. According to immunolabeling results, weak fluorescent signals of GJA8 protein were observed in the outermost fiber cells of F1-23 (WT/-52) rabbit, but stronger punctate signals were seen in WT rabbits (Fig. 6A). Furthermore, the transmission electron microscopy results revealed a much smaller gap junction in the cortical fibers of the GJA8 mutant rabbit, compared to longer and regular gap junction in the WT rabbit (Fig. 6B).

Discussion
In this study, we found that the embryo survival and development were not affected by cytoplasmic injections. Also, the efficiency of gene modification was as high as 100% in newborn by co-injection of sgRAN and Cas9 mRNAs into rabbit's zygotes, which suggesting the cytoplasm injection using CRISPR/Cas9 system is a feasible way to perform gene editing in animals. In addition, Sanger sequencing showed that the indels in the founder rabbits ranged from 12-81 bp. These results suggested that the dual sgRNA directed CRISPR/Cas9 system improved the knockout efficiency, providing a strategy to facilitate gene knockout and large deletions of lncRNA genes 24 . Moreover, CRISPR/Cas9 based cytoplasmic injection has several advantages over traditional methods, like somatic cell nuclear transfers (SCNT), in which abnormal reprogramming in somatic cell clones and low cloning efficiency have been reported in several mammals 25 . Therefore, CRISPR technology shows great promise as a genome editing technique in various vertebrate model systems.
Off-target effects have been described in Cas9-mediated knockout mice 20 and zebrafish 26 . However, the potential off-target effects were not found in the present study, which could be due to the low concentration (50 ng/μ l) of sgRNA and Cas9 mRNAs used. We hypothesized that low concentrations would reduce the off-target effects by causing the sgRNA and Cas9 mRNAs to transiently act on targeting sites and degrade immediately after targeting the gene. In addition, it is particularly important to avoid the mismatches of seed sequences (8-12 bases close to PAM) when design the sgRNA. Furthermore, using cytoplasm microinjection instead of somatic cell nuclear transfer also reduces off-target effects. In fact, to reduce the off-target mutation, the inactivated structural domain of Cas9 (Cas9D10A) or using truncated sgRNAs have been reported to be used in other groups [27][28][29][30] . Results of the our study revealed that knocking out of the GJA8 gene in rabbits, which is important for the development and function of vertebrate lens 5 , was sufficient to recapitulate the human cataract phenotype. Previous studies have been reported that the GJA8 (+ /− ) mice appeared to have normal eyes and lenses, while only GJA8 (− /− ) mice developed microphthalmia with smaller lenses compared to WT mice 10 . We found, however, that GJA8 (+ /− ) rabbits developed cataracts had microphthalmia and smaller lenses. These data were consisted with the clinical data of human GJA8 cataract phenotypes, due to the dominant gene mutations [5][6][7] . Our results demonstrated that GJA8 (+ /− ) rabbits have dominant cataracts, which was more similar to human cataracts than the mouse in inheritance pattern. Yet, not all these rabbit pups in the F0 or F1 generations exhibited the cataract phenotype. The different lengths of deletion fragments could possibly cause these different lens phenotypes. According to the genotype analysis, cataracts were not observed in rabbits with deletion fragment lengths that were multiples of 3, such as F0-4 (WT/-57bp), F0-6 (WT/-51bp) and F0-10 (WT/-51bp) (− 44, + 3) rabbits in the F0 generation. These results suggested that GJA8 transcription was not terminated but produced a mutated protein. The 3D structure indicating that these mutated protein structures of GJA8 were unstable in the rabbits (Fig. 5C). In summary, our study showed that knocking out GJA8 in rabbit via CRISPR/Cas9 system causes human-like cataracts. This novel rabbit cataract model would provide a convenient way to screen new drugs for cataract prevention and treatment.

Materials and Methods
Ethical statement. New Zealand rabbits were housed in the Laboratory Animal Center of Jilin University.
All animal protocols were approved by the Animal Care Center and Use Committee of Jilin University. All experiments were performed according to the guidelines approved by Jilin University.
Microinjection and embryo transfer. The female New Zealand White rabbits (6-8 months old) were superovulated with FSH (60 IU) at intervals of 12 h for 6 times. After the last injection, female rabbits were mated with the male rabbits. Females then received a 100 IU human chorinonic gonadotrophin (HCG) injection. At 18 h post HCG injection, female rabbits were euthanized, and oviducts were flushed with 5 ml DPBS-BSA for zygotes collection. Rabbit embryos at the pronuclear stage (around 18-20 h post-mated) were collected and transferred into oocyte manipulation medium, which contained 9.5 g TCM-199, 0.05 g NaHCO 3 (Sigma, S4019), 0.750 g Hepes (Sigma, H3784), 0.05 g penicillin, 0.06 g streptomycin, 1.755 g NaCl, 3.0 g BSA, and 1 L Milli Q H2O. A mixture of Cas9 and sgRNA mRNA (200 ng/μ l and 50 ng/μ l, respectively) was microinjected into the embryo cytoplasm to knock out the GJA8 gene. The injected embryos were transferred into EBSS medium for short-term culture at 38.5 °C, 5% carbon dioxide and humidity conditions. 30-50 injected embryos were transferred into the oviduct of a recipient rabbit.
Gene mutation detection in embryos and pups by PCR. The injected embryos were collected at the blastocyst stage. Genomic DNA was extracted with an embryo lysis buffer (1% NP40) at 50 °C for 20 minutes and 90 °C for 5 minutes in a BIO-RAD PCR Amplifier. Genomic DNA from wild type (WT) and GJA8 knockout (KO) rabbit pups was isolated using the TIANamp Genomic DNA Kit (TIANGEN, Beijing, China). The DNA was amplified with 2× Taq Plus MasterMix (TIANGEN) and PCR primers used to detect mutation were as follows: F-5′ CGAGAATGTCTGCTACGATGAG-3′ , and R-5′ CCGGAAACCATACAGGAAGTAG-3′ (Fragment size of GJA8-WT allele: 367 bp). The PCR products were purified and cloned into the pGM-T vector (TIANGEN, Beijing, China), and then analyzed by Sanger sequencing. The colonies were picked and sequenced to confirm exact gene mutations.
Off-target assay. Seven potential off-target sites (POTS) for each sgRNA were predicted to analyze site-specific cleavage by the CRISPR/Cas9 system according to an online design tool (http://crispr.mit.edu/). The PCR products of the POTS were sequenced and confirmed by T7E1 enzyme digestion, as previously described 31 . Primers for POTS determination were listed in Supplementary Tables S2.
Hematoxylin and eosin (H&E) staining. Eyes tissues from WT and GJA8 KO rabbits were fixed in 4% paraformaldehyde for 48 h, then embedded in paraffin wax and sectioned for slides. Slides were stained with hematoxylin and eosin (H&E) and viewed under a Nikon ts100 microscope.
Western blotting. Imediately after rabbits were euthanized, the eyes were ground under liquid nitrogen. The powder of eye tissues was lysed in a protein lysis buffer on ice for 30 min. Protein concentrations were measured using the BCA Protein Assay Kit (Beyotime). Anti-GJA8 polyclonal antibody (1:2,000; Abcam 199102) was used to measure protein levels. Anti-Beta actin monoclonal antibody (1:2,000; Proteintech 60008-1) was used as an internal control.
3 D structure models of the WT and GJA8 mutant proteins were built from their amino acid sequences according to the web site: http://swissmodel.expasy.org/ 32 . Immuno-fluorescence assay. Eye lenses from WT and GJA8 KO rabbits were fixed in 4% paraformaldehyde for 48 h, embedded in paraffin wax and sectioned for slides. Thin sections were de-waxed with dimethylbenzene and dehydrated using an ethanol gradient (100%, 95% and 80%). The slides were placed in 0.1% Triton solution at 4 °C for 1 h. Slides were blocked with 5% goat serum for half an hour at 37 °C and rinsed with PBS 3 times. Slides were incubated with primary antibody (1:200, Abcam 199102) at 4 °C for overnight. Slides were then incubated with and anti-rabbit secondary antibody (1:500, Invitrogen A-11012) in 5% goat serum for 1 h at room temperature in the dark. Finally, cell nuclei were counter stained with hochest 33342 for 7 min in the dark. Confocal laser microscopy was used to examine the GJA8 fluorescence.
Electron microscopy analysis. For morphological analysis by electron microscopy, WT and GJA8 KO eye lenses were cut into small pieces, and fixed in 0.1 M cacodylate buffer (2% glutaraldehyde, 2.5% formaldehyde, pH 7.2) for 2-4 hours at room temperature. As previous study described 33 , the fixed lenses were rinsed and treated with OsO 4 , tannic acid, and uranyl acetate. Then, they were embedded in Epon. Ultrathin sections were cut with a diamond knife, mounted on copper grids and viewed under an electron microscope (Hitachi H-7650) at 80 Kv.
Statistical analysis. Percentage for in vitro embryo development in the 2-cell, morular and blastocyte groups was compared by chi-square test. A value of p < 0.05 was considered statistically significant.