CRISPR/Cas9 engineering of albino cystinuria type A mice

Abstract

Cystinuria type A is a relatively common genetic kidney disease occurring in 1 in 7,000 people worldwide that results from mutation of the cystine transporter rBAT encoded by Slc3a1. We used CRISPR/Cas9 technology to engineer cystinuria type A mice via genome editing of the C57BL/6NHsd background. These mice are an improvement on currently available models as they are on a coisogenic genetic background and have a single defined mutation. In order to use albinism to track Cas9 activity, we co-injected gRNAs targeting Slc3a1 and tyrosinase (Tyr) with Cas9 expressing plasmid DNA into mouse embryos. Two different Slc3a1 mutational alleles were derived, with homozygous mice of both demonstrating elevated urinary cystine levels, cystine crystals, and bladder stones. We used whole genome sequencing to evaluate for potential off-target editing. No off-target indels were observed for the top ten predicted off-targets for Slc3a1 or Tyr. Therefore, we used CRISPR/Cas9 to generate coisogenic albino cystinuria type A mice that could be used for in vivo imaging, further study, or developing new treatments of cystinuria.

1. INTRODUCTION

Cystinuria occurs in 1 in 7,000 globally (Edvardsson et al., 2013). Cystinuria type A results from mutation of SLC3A1 (rBAT protein) and type B results from mutation of SLC7A9 (b^0,+AT protein) (Eggermann, Venghaus, & Zerres, 2012; Sahota, Tischfield, Goldfarb, Ward, & Hu, 2019). Cystinuria is autosomal recessive. High levels of cystine are found in the urine due to defective reabsorption of cystine in the proximal tubule of the kidney. Cystinuria is associated with a high number of surgical procedures, flank pain, chronic urinary tract infections, and chronic kidney disease. Available therapies for cystinuria are lacking in efficacy so animal models in which to test new therapies are needed.

Cystinuria manifests a prominent pediatric phenotype, contributing up to 10% of urolithiasis in children (Claes & Jackson, 2012; Knoll, Zollner, Wendt-Nordahl, Michel, & Alken, 2005). Onset can occur as early as between infancy and two years of age, though median onset is in the early teen years for both male and female patients (Rogers, Kalakish, Desai, & Assimos, 2007). Males are more likely to exhibit the cystinuria phenotype with symptoms before the age of three than females (Rogers et al., 2007), and males tend to do overall worse than females (Claes & Jackson, 2012).

Three different mouse models of cystinuria type A exist (Sahota et al., 2019). One was discovered after ENU mutagenesis resulting in D140G in C3HeB/FeJ mice (Peters et al., 2003), one exists as a spontaneous mutation of E383K in 129S2/SvPasCrl mice (Livrozet et al., 2014), and another was derived from homologous recombination and deletion of exon 1 in C57BL/6J-129/SvJ mice (Ercolani et al., 2010). All of these cystinuria type A strains are on mixed genetic backgrounds. Neither of the point mutation models are available to investigators in the United States. The Ercolani model which is available to U.S. investigators is not useful for testing correction by homologous recombination with Cas9 and related RNA-programmable nuclease systems.

With the goal of using cystinuria type A as a model system for eventual kidney-directed gene therapy strategies, we desired to generate a coisogenic cystinuria type A mouse strain. We created albino mice through mutation of the tyrosinase gene because they would be optimal for in vivo imaging of transgene expression in the whole animal, as darker coat color inhibits detection of luciferase and fluorescent proteins (Bell et al., 2007; Belur, Podetz-Pedersen, Frandsen, & Mcivor, 2007). Lastly, we desired cystinuria type A mice with mutations that could be potentially repaired via genome editing in order that we might test such strategies in an in vivo model. Such strategies could not be recapitulated in a mouse model created by the previously traditional model of knock-in of an antibiotic resistance cassette. For all these reasons, we undertook these studies in order to create a coisogenic albino cystinuria type A mouse line.

2. RESULTS AND DISCUSSION

We previously characterized the metabolic consequences of cystinuria type A on a mixed genetic background (Woodard et al., 2019). We sought to develop a cystinuria type A model on a coisogenic background that would be albino for in vivo imaging. Given speed, precision, and lower cost, we used a CRISPR/Cas9 genome editing approach to simultaneously create Slc3a1^−/−/Tyr^−/− mice on the C57BL/6NHsd background. The strategy involved simultaneous targeting of exon 2 of Slc3a1 (Figure 1a) and exon 1 of Tyr. We reasoned that Tyr knockout would allow us to easily identify mice wherein genome editing had occurred, allowing us to use coat color as a readout of genome editing and a surrogate marker of simultaneous Slc3a1 editing. We first evaluated the ability of our designed gRNA to mediate editing of Slc3a1 in mouse embryonic fibroblast NIH3T3 cells. Surveyor assay analysis revealed editing of Slc3a1 when the gRNA was present in the pX330-U6-Chimeric-BB-CBh-hSpCas9 (pX330-Slc3a1) plasmid (Figure 1b).

FIGURE 1.

Design and confirmation of genome editing of Slc3a1 and Tyr. (a), gRNA targeting strategy of exon 2 (blue box) of Slc3a1 and exon 1 (blue box) of Tyr. (b), confirmation of genome editing of Scl3a1 via surveyor assay after attempted editing in NIH3T3 cells.

Pronuclear microinjection of pX330-Slc3a1 and pX330-Tyr ultimately produced 14 pups of which 6 were black, 2 were mosaic, and 6 were albino. White coat color of the pups indicated gene editing activity via albinism conferred by editing Tyr. Analysis of black pups revealed no editing of Slc3a1. White pups were analyzed via TIDE analysis and two Slc3a1 mutations were identified being c.527_531delTTGAG;insA (p.Val176GluX13) and c.526insTT (p.Glu177LeuX14) (Figure 2) (protein sequence of WT and mutations in Supporting Information). Both indels resulted in frameshift mutation in exon 2 expected to disrupt the Scl3a1 transcript. The two different mutations were subsequently bred to homozygosity for further analysis (Figure 2b and c).

FIGURE 2.

Generation of Slc3a1 knockout alleles. (a), TIDE analysis of WT mouse. (b) TIDE analysis of heterozygous and homozygous c.527_531delTTGAG;insA allele. (c), TIDE analysis of heterozygous and homozygous c.526insTT allele. Sequencing analysis demonstrates the complementary strand.

Cystinuria type A mice exhibit on average a 51-fold elevation in urinary cystine concentration (Figure 3a). We confirmed cystinuria in the c.527_531delTTGAG;insA and c.526insTT male and female mice using HPLC as described previously (Woodard et al., 2019). Both mutant Slc3a1−/− lines exhibited pathognomonic hexagonal cystine crystals in their urine (Figure 3b). Elevated urinary cystine and cystine crystalluria confirmed functional knockout of Slc3a1. Cystinuria type A mice have previously been demonstrated to form bladder stones which can be visualized by X-ray analysis (Ercolani et al., 2010; Woodard et al., 2019). Therefore, we evaluated the male and female mice of the two different mutations overtime in their ability to form bladder stones (Figure 4a and b). Both c.527_531delTTGAG;insA and c.526insTT male Slc3a1−/− mice exhibited a higher rate of stone formation compared to their female counterparts (Figure 4b). We did not observe stone formation in c.527_531delTTGAG;insA female mice. The c.526insTT female mice exhibited a statistically lower rate of stone formation compared to male knockout mice (p=0.037), though the rate was not statistically different from c.527_531delTTGAG;insA female mice or WT animals (p=0.25) (Figure 4b). Without doing a side-by-side comparison, the age of onset of stone formation appeared similar to that reported for other cystinuria type A strains (Zee et al., 2017).

FIGURE 3.

Slc3a1 genome editing results in cystinuria. (a), urinary cystine concentration at 8–12 weeks of age in male or female WT, c.526insTT, or c.527_531delTTGAG;insA mice (N=3–10±SEM). *, p<0.05 via T test. (b), pathognomonic hexagonal urinary cystine crystals found in the urine of c.526insTT and c.527_531delTTGAG;insA allele animals at 8 weeks of age.

FIGURE 4.

Slc3a1 editing results in bladder stone formation. (a), X-ray analysis of a typical c.526insTT or c.527_531delTTGAG;insA male mouse. Bladder stones are outlined by the blue circles. A WT (C57BL/6NHsd) mouse is shown as a control. (b), analysis of stone formation over time (N=5–17). Slc3a1−/− male mice exhibited a higher rate of stone formation overtime compared to female Slc3a1−/− mice (log-rank Mantel-Cox test).

We used whole genome sequencing to evaluate for off-target editing. We analyzed the top 10 potential off-target sites as scored by the CRISPR Finder tool (http://www.sanger.ac.uk/htg/wge/) for Cas9 cleavage activity for each gRNA (Slc3a1 and Tyr) revealed no off-target editing (Supporting Information). We used Venny 2.1.0 (https://bioinfogp.cnb.csic.es/venny) to evaluate for overlap of indels between c.526insTT and c.527_531delTTGAG;insA strains (Supporting Information). Importantly, indels for Scl3a1 and Tyr were present in both strains (Table). Given that the other overlapping indels did not correlate with predicted off-targets, our conclusion is that these indels are likely resultant from spontaneous mutation and subsequent breeding (Montoliu & Whitelaw, 2018).

Table.

Indels common to c.526insTT and c.527_531delTTGAG;insA strains.

Ensembl Gene	c.526insTT	c.527_531delTTGAG;insA	Gene Symbol
ENSMUSG00000044783	nonframeshift deletion	nonframeshift deletion	Hjurp
ENSMUSG00000063681	frameshift deletion	frameshift deletion	Crbl
ENSMUSG00000006412	frameshift deletion	frameshift deletion	Pfdn2
ENSMUSG00000052563	frameshift insertion	frameshift insertion	D930048N14Rik
ENSMUSG00000006056	frameshift insertion	frameshift insertion	Calcoco2
ENSMUSG00000072812	frameshift deletion	frameshift deletion	Ahnak2
ENSMUSG00000089901	frameshift deletion	frameshift deletion	Gm8113
ENSMUSG00000022255	frameshift deletion	frameshift deletion	Mtdh
ENSMUSG00000094296	frameshift deletion	frameshift deletion	Gm21798
ENSMUSG00000045350	nonframeshift deletion	nonframeshift deletion	Faml86a
ENSMUSG00000001870	frameshift deletion	frameshift deletion	Ltbpl
ENSMUSG00000024131	frameshift insertion	frameshift deletion	Slc3al
ENSMUSG00000045150	nonframeshift deletion	nonframeshift deletion	Olfi-1161
ENSMUSG00000028357	nonframeshift insertion	nonframeshift insertion	Kifl2
ENSMUSG00000043962	frameshift deletion	frameshift deletion	Thrap3
ENSMUSG00000062518	frameshift insertion	nonframeshift insertion	Zfp534
ENSMUSG00000035203	frameshift insertion	frameshift insertion	Epnl
ENSMUSG00000054676	frameshift deletion	frameshift deletion	1600014C10Rik
ENSMUSG00000004651	frameshift deletion	frameshift deletion	Tyr
ENSMUSG00000095891	frameshift insertion	frameshift insertion	GmlO717
ENSMUSG00000062649	stopgain	stopgain	Olfr44
ENSMUSG00000042254	frameshift deletion	frameshift deletion	Cilp
ENSMUSG00000032803	frameshift deletion	frameshift deletion	Cdv3
ENSMUSG00000079600	frameshift deletion	frameshift deletion	Gm17604

WGS results revealed common indels between the two cystinuria type A mouse lines generated. Importantly, Slc3a1 and Tyr were mutated in both lines.

In conclusion, we have successfully generated new cystinuria type A albino mouse models using CRISPR/Cas9 genome editing. These coisogenic strains can be used for new therapeutic approaches directed at cystinuria (Sahota et al., 2019; Zee et al., 2017), and for in vivo imaging of drugs or delivery vectors targeting the underlying genetic disease.

3. MATERIALS AND METHODS

3.1. Generation of Slc3a1^−/−/Tyr^−/− mice using CRISPR/Cas9 gene editing

The pX330-U6-Chimeric-BB-CBh-hSpCas9 plasmid, which contains a human codon-optimized SpCas9 and chimeric gRNA, was used for gene editing (Addgene, #42230) (Cong et al., 2013). Exon 2 of Slc3a1 was retrieved from the Mouse Dec. 2011 GRCm38/mm10 assembly on the UCSC genome browser (https://genome.ucsc.edu/cgi-bin/das/mm9/dna?segment=chr17:85431783,85431963). The gRNA sequence AGACTTCAGATACGCTGTTG, was subcloned into pX330 to create pX330-Slc3a1 with the primers SLCex2g1top CACCGAGACTTCAGATACGCTGTTG and SLCex2g1bottom AAACCAACAGCGTATCTGAAGTCTC. Gene editing of mouse Slc3a1 was confirmed by transfecting pX330-Slc3a1 into NIH3T3 cells using Lipofectamine 2000 (Fisher, Waltham, MA) and an IDT Surveyor Mutation detection kit (Integrated DNA Technologies, Coralville, IA), similar to how we have described previously (Luo et al., 2017). The gRNA sequence targeting Tyr, CCATGGATGGGTGATGGGAGTCC, was subcloned into pX330 to created pX330-Tyr. All plasmids were confirmed using DNA sequencing. Microinjections were completed using 5 ng/μl pX330-Slc3a1 + 5 ng/μl pX330-Tyr. Three pronuclear microinjection days were completed using C57BL/6NHsd mice (Envigo, Indianapolis, IN). 341 embryos were injected, 270 embryos survived (79% survival), 12 embryo transfer surgeries were completed, 16 pups born (6%), 3 pups died, and 7 of the 13 living pups were albino.

3.2. Genotyping for Slc3a1 mutations

Genomic DNA was obtained by DNeasy (Qiagen) from clipped tails from mice less than 21 days old. Primers SLC_Surv_F2 GGCTCTCCCTTCAGGTATCC and SLC_Surv_R1 TCCGTGTCCTACTCGATTGG were used for Taq PCR (Advantage 2, Clonetech) on 2 ul of each tail DNA per 20 ul PCR rxn. Thermocycling was performed with the following PCR program: Initial denature 94C for 5m, followed by 40 cycles of denature 94C for 30s, anneal 61C for 30s, and polymerize 72C for 45s, and a final polymerization of 72C for 10m to produce a 408 bp product which was gel-extracted from a 1% agarose gel. TIDE analysis (https://tide.deskgen.com/) was used to determine original genome editing events and to confirm genotypes of homozygous mice (Brinkman, Chen, Amendola, & van Steensel, 2014). Subsequently, mice were genotyped by Transnetyx (Cordova, TN).

3.3. Phenotypic analysis of mice

Urine amino acid concentrations were determined by reverse phase HPLC using a modified version of the methods of Bidlingmeyer et al. as previously described (Bidlingmeyer, Cohen, & Tarvin, 1984; Woodard et al., 2019). Urine creatinine was measured using the University of Alabama-Birmingham O’Brien Center Core C Biomarkers Laboratory by LC-MS/MS (Woodard et al., 2019). To monitor stone formation over time, mice were taken to the Vanderbilt University Institute of Imaging Science, anesthetized with isoflourane and oxygen until immobile, then placed in a Faxitron 2000 X-ray machine at setting 35 for an exposure time of 4 seconds as described previously (Woodard et al., 2019). Urinary cystine crystals were visualized using light microscopy.

3.4. Analysis of off-target Cas9 activity

Whole genome sequencing (WGS) was performed on one F3 mouse of each of the homozygous Slc3a1−/− alleles (c.526insTT, c.527_531delTTGAG;insA) and a WT (C57BL/6NHsd) mouse purchased from Envigo. Genomic DNA was prepared using a Dneasy kit from Qiagen (Germantown, MD) according to the manufacturers instruction. WGS and analysis was performed by Genewiz (South Plainfield, NJ). The top 10 potential off-target sites for each gRNA (Slc3a1 and Tyr) were identified using the Welcome Trust Sanger Institute Genome Editing website (http://www.sanger.ac.uk/htg/wge/).

3.5. Animal care

Slc3a1^−/− and wild type mice were bred and maintained in an AAALAC accredited vivarium at Vanderbilt University Medical Center in temperature regulated rooms (22 ± 2°C) with a 12-hour lights on-off cycle and fed as described previously (Woodard et al., 2019). All care and procedures were carried out according to the Institutional Animal Care and Use Committee of the Nashville Tennessee Valley Healthcare System VA and Vanderbilt University Medical Center. All studies were in compliance with the US National Research Council’s Guide for the Care and Use of Laboratory Animals, the US Public Health Service’s Policy on Humane Care and Use of Laboratory Animals, and Guide for the Care and Use of Laboratory Animals.

3.6. Statistical analysis

The T test was used for comparison urine cystine levels between groups. The log-rank Mantel-Cox test was used for survival curve (% without stones) comparison between groups.