CRISPR correction of the Finnish ornithine delta-aminotransferase mutation restores metabolic homeostasis in iPSC from patients with gyrate atrophy

Hyperornithinemia with gyrate atrophy of the choroid and retina (HOGA) is a severe recessive inherited disease, causing muscular degeneration and retinochoroidal atrophy that progresses to blindness. HOGA arises from mutations in the ornithine aminotransferase (OAT) gene, and nearly one-third of the known patients worldwide are homozygous for the Finnish founder mutation OAT c.1205 T > C p.(Leu402Pro). We have corrected this loss-of-function OAT mutation in patient-derived induced pluripotent stem cells (iPSCs) using CRISPR/Cas9. The correction restored OAT expression in stem cells and normalized the elevated ornithine levels in cell lysates and cell media. These results show an efficient recovery of OAT function in iPSC, encouraging the possibility of autologous cell therapy for the HOGA disease.


Introduction
Hyperornithinemia with gyrate atrophy of choroid and retina (HOGA) is a severe autosomal recessively inherited metabolic disease, causing progressive visual degeneration [1,2] without curative treatment to date. Clinical manifestations include retinochoroidal atrophy, early cataract formation and type-II muscle fiber degeneration. These symptoms start in childhood, progressing to blindness and loss of type-II muscle fibers in adulthood [3,4].
HOGA arises from deleterious mutations in the ornithine aminotransferase (OAT) gene that lead to loss of enzymatic activity and ensuing high ornithine concentrations [5][6][7]. A strict Arg-restrictive diet reduces plasma ornithine concentration and slows down disease progression [8], suggesting ornithine toxicity in certain cell types.
The prevalence of HOGA is particularly high in Finland, where the overall carrier frequency nears 1:151 (versus the global 1;833) [9], with more than 90% of the patients presenting with the pathogenic Finnish mutation OAT c.1205 T > C p.(Leu402Pro) in homozygosity (hereafter HOGA Fin or c.1205 T > C, [9,10]).
A mouse model with OAT deficiency has been constructed [11].
When untreated, these animals recapitulate the progression of the human retinal disease. Moreover, the dietary treatment since weaning prevents retinal degeneration in the murine eye [12]. For human patients, however, this strict diet is difficult to accomplish, and the current treatments have been shown to delay, but not to prevent, the degeneration of the retina and choroid [13][14][15]. Hence, the need for an effective long-term treatment remains unmet. Gene-editing tools offer novel therapeutic opportunities for many genetic diseases [16]. For ophthalmological pathologies, CRISPR/Cas9 systems have proven to efficiently correct specific disease mutations, either by homology-directed repair (HDR) or by single base editing [17,18]. These successful precedents have placed these technologies under the spotlight for the development of advanced therapies, some already in clinical trials [19].
Here, we have generated iPSC lines from two HOGA Fin patients, and subsequently corrected the pathogenic mutation using Cas9-induced HDR. These corrected isogenic iPSC lines served to establish and characterize the first patient-derived iPSC model of HOGA Fin . We observed a restoration of OAT enzyme production and function upon genetic correction of the mutation. The recovered aminotransferase activity lowered the elevated ornithine levels in cell cultures. This further normalized the concentrations of other metabolites in related pathways [20][21][22]. Additionally, the corrected iPSC lines were geno-and phenotypically normal, and we did not detect any off-target genetic modifications, suggesting an acceptable safety profile. As the pathomechanisms of the disease remain unclear and the current treatment does not halt disease progression, we hope that our model will facilitate more detailed studies on pathophysiology and help to develop new therapeutic alternatives.

Patients, material and methods
Additional materials and methods can be found in Supplementary data.

Patients
We collected skin biopsies from two non-related voluntary female patients (aged 18 and 30). The patients were diagnosed in their childhood and have been on arginine-restricted diet ever since. Creatine and lysine supplementation started later in life. Both are homozygous for the most common Finnish OAT mutation c.1205 T > C p.(Leu402Pro).

Ethical consent
The Coordinating Ethics Committee of the Helsinki and Uusimaa Hospital District approved the generation of the induced pluripotent stem cell lines upon informed consent of the donors (Nro #HUS/2754/ 2019).

sgRNA and DNA donor template design
For genome editing of the disease mutation, we designed a suitable guide RNA using online tools (CRISPOR, [25]). We generated the sgRNA by incubating 5 min at 95 • C our gRNA (customized Alt-R CRISPR-Cas9 gRNA, iDT) with Alt-R® CRISPR-Cas9 tracrRNA, ATTO™ 550 (iDT). We designed a dsDNA donor template to include the desired correction along with silent mutations for screening purposes (novel restriction site for Bsh1236i, CG^CG). We generated the dsDNA donor template by overlapping PCR of two primers (Sigma, Table 1).

Genetic correction of c.1205 T > C in mutant iPSC
For each electroporation experiment, we dissociated 2 × 10^6 patient-derived iPSC to single cells with StemPro Accutase Table 1 gRNA and oligos used for off-target screening and dsDNA donor template production.

Off-target analysis
We selected and designed PCR primers ( Table 1, Sigma) for the top-6 off-target sites based on the off-target score for the given gRNA, as found in the online tools https://benchling.com and CRISPOR. Sequences were confirmed by Sanger sequencing of the PCR products (see Table 1 for primers).

Ornithine measurement in cell media
We measured Ornithine concentration in iPSC media (n = 4/cell line) using Ornithine Assay Kit (abcam, BioVision). Results are expressed in μM. We based the statistical analysis on Tukey test for multiple comparison of group means. For targeted analysis of ornithine-related metabolites, FIMM metabolomics unit sonicated and analyzed the soluble fraction with Thermo Vanquish UHPLC+ system, coupled to a Q Exactive Orbitrap quadrupole mass spectrometer (ThermoFisher Scientific). Gradient on a SeQuant ZIC-pHILIC (2.1 × 100 mm, 5-μm particle) column (Merck) was used with Mobile phase A (20 mM ammonium carbonate, pH 9.4), and B (acetonitrile). Instrument control operated with the Xcalibur 4.1.31.9 software (ThermoFisher Scientific).

Targeted metabolomic analysis
Metabolite peaks were integrated with TraceFinder™ 5.1 software using retention times confirmed with standards (inhouse library and library kit MSMLS-1EA, Merck). Data quality was monitored with pooled human serum as Quality Control (QC). We ran post-measurement analysis with data normalized to total protein (Pierce BCA Protein Assay Kit, ThermoFisher) within each prepared sample. Volcano plots generated in Rstudio with the 'EnhancedVolcano' package. Results shown as a matrix of fold change (cut at 0.5) versus p.value (cut at 0.05).

Genetic correction of disease mutation in patient iPSC
We established HOGA Fin iPSC lines and subsequently corrected the HOGA Fin mutation with our selected CRISPR/Cas9 design (Fig. 1A). We obtained 55 monoclonal colonies. Our editing approach, comprising the homologous recombination of a double-strand DNA (dsDNA) donor template, showed a recombination efficiency close to 49% (Fig. 1B). It generated a total of 27 edited clones, 18 with one and 9 with both alleles corrected (heterozygotes and homozygotes, respectively).
We detected and excluded from further analysis 2 heterozygotes and 19 non-corrected clones with unintentional alterations (indels) in the targeted region.
Nonetheless, the analysis of the top six predicted off-target regions did not reveal off-target modifications in any of the edited iPSC lines (Table 1).

Characterization of iPSC lines
To ensure the quality and function of all our iPSCs, we tested their genetic integrity and pluripotency ability (see Supplementary data). A PCR for EBNA1 and OriP in 8 original patient cell lines confirmed that our iPSC lines did not retain the vectors employed for reprogramming (data not shown). All iPSC lines, before and after correction, presented distinctive hiPSC morphology and positive immunostainings of pluripotency markers (Nanog, Lin28, Tra-1-60, Supplementary Fig. 1). Furthermore, quantitative measurement by qPCR of pluripotency markers OCT4, Nanog and Sox2 in corrected cell lines revealed expected expression patterns, indicating that the editing process did not affect these characteristic pluripotency genes (data not shown). During embryoid body differentiation, all cell lines successfully differentiated into the three germ layers (Supplementary Fig. 1). Importantly, neither the reprogramming nor the editing process affected the karyotypes of edited or non-edited cell lines, as shown by G-banding of iPSCs from both patients (Supplementary Fig. 2). Along with the genetic results, these analyses suggest that the editing process did not impair neither the function nor characteristics of our iPSCs.

Analysis of protein expression
The predominant Finnish HOGA-causing mutation c.1205 T > C p. (Leu402Pro) likely destabilizes the protein structure, ultimately leading to its degradation, as seen in patient fibroblasts [7]. Accordingly, we observed comparable mRNA expression in Hel24.3 (stable iPS cell line used as a control, c.1205 T/T), mutant iPSC, and iPSC edited in homozygosity by qPCR. We did not detect, however, OAT protein in our iPSC derived from homozygous HOGA Fin patients. After the genetic correction, the iPSC lines edited in hetero-and homozygosity showed a strong production of OAT in the Western Blot analysis (Fig. 2A).

Restoration of metabolic pathways
Given that patients present with elevated ornithine in circulation and body fluids, we measured ornithine concentration in cell media to assess OAT function in our cell model. Media from all corrected cell lines presented ornithine levels comparable to those in media from control iPSC lines (Hel24.3), significantly lower than those of the homozygousmutant HOGA iPSCs (Fig. 2B). Our results indicate that the correction of the OAT mutation leads to the recovery of the enzymatic activity.
OAT dysfunction triggers an intracellular imbalance of ornithine related metabolites (Fig. 2C). To evaluate the effects of protein function recovery, we explored these metabolic pathways in edited versus nonedited patient lines, using liquid chromatograph-mass spectrometry (LC-MS targeted metabolomics). We discovered that the levels of intracellular ornithine significantly decrease in edited cell lines when compared to the mutants (Fig. 3A-B). In addition, we found a significant increase in the intracellular levels of creatine, creatine phosphate, alanine, arginine, lysine, glutamine, and proline ( Fig. 3A-B).

Discussion
Our results show the restoration of OAT enzymatic function after genetic correction in a novel iPSC model for HOGA Fin . Patients with HOGA present with complex metabolic imbalance, related to the hyperornithinemia caused by OAT dysfunction. Here, we investigated these intracellular metabolic changes by comparing the OAT-mutant and homozygously corrected cells. We observed restoration of metabolic homeostasis upon recovery of OAT activity, including increased levels of lysine, glutamate, glutamine, creatine and creatine phosphate reported depleted in patients [3,4,21].
While the mechanisms behind chorio-retinal degeneration remain elusive, type-II muscle fiber atrophy is thought to arise from the failure to generate enough creatine and creatine phosphate. The ornithinemediated inhibition of L-arginine:glycine amidinotransferase (AGAT) plays a key role in the pathology [21]. AGAT produces guanidinoacetate, the precursor of creatine, from arginine. In line with these previous reports, we could show substantially reduced creatine and creatine phosphate concentration in the OAT-mutant cell lines. This emphasizes the importance of creatine supplementation usually included in the dietary treatment [26] . Additionally, HOGA patients present with low plasma concentrations of lysine [27], and oral lysine treatment has lowered plasma ornithine levels [28] by the competition of lysine and ornithine in the tubular reabsorption of amino acids. Interestingly, we observed lower intracellular concentration of lysine in the OAT-mutant iPSC lines.
As part of the molecular characterization, we analyzed the quantitative OAT expression levels and OAT production. Our results show that mutant iPSC express OAT at the mRNA level but fail to produce the protein. This suggests that the Finnish-founder mutation under study, OAT c.1205 T > C p.(Leu402Pro), might destabilize the structure of the protein, ultimately leading to its degradation.
The current treatment strategies usually improve the muscular prognosis but only decelerate the progression of the retinal symptoms. Although long-term therapeutic alternatives have been proposed [29,30], neither enzyme nor gene replacement therapies exist to date.
Here, we sought to explore the genetic correction of the mutation directly in patient-derived iPSC. This approach allows the production of functional, autologous cells with the OAT expression under its physiological genetic control. Our iPSC feature lowered intracellular ornithine levels compared to their non-edited counterparts. The cell culture media of all our edited cell lines showed decreased ornithine concentration, albeit less pronounced in heterozygotes. This accords with previous observations of reduced OAT activity in carriers, although they show no clinical symptoms [31,32].
Importantly, the edition did not produce any detectable off-target modifications, neither did it affect the pluripotent quality of the iPSCs nor their chromosomal integrity. However, the occurrence of undesired, on-target alterations makes it currently necessary to obtain monoclonal lines, as the effects of these changes on the phenotype remain uncertain. These events challenge the immediate utilization of the presented genome editing approach as an in vivo therapeutic alternative. Nevertheless, we believe that our results offer new opportunities for the modelling of this disease, mechanistic studies, and the evaluation of novel therapies for gyrate atrophy. We predict that the constant advancement of CRISPR/Cas9 editing, and delivery systems will bring promising optimization strategies to make new therapeutic alternatives a reality.

Funding
The work was funded by Academy of Finland (Grant 308481), Foundation for Paediatric Research, Sigrid Juselius Foundation, Paulo Foundation, Mary & Georg C. Ehrnrooth Foundation and Finnish Cultural Foundation. FIMM metabolomics unit was supported by HiLIFE and Biocenter Finland.

Declaration of Competing Interest
The authors declare no conflict of interest.