Citrin-deficient patient-derived induced pluripotent stem cells as a pathological liver model for congenital urea cycle disorders

Citrin deficiency is a congenital secondary urea cycle disorder lacking useful disease models for effective treatment development. In this study, human induced pluripotent stem cells (iPSCs) were generated from two patients with citrin deficiency and differentiated into hepatocyte-like cells (HLCs). Citrin-deficient HLCs produced albumin and liver-specific markers but completely lacked citrin protein and expressed argininosuccinate synthase only weakly. In addition, ammonia concentrations in a medium cultured with citrin-deficient HLCs were higher than with control HLCs. Sodium pyruvate administration significantly reduced ammonia concentrations in the medium of citrin-deficient HLCs and slightly reduced ammonia in HLCs differentiated from control iPSCs, though this change was not significant. Our results suggest that sodium pyruvate may be an efficient treatment for patients with citrin deficiency. Citrin-deficient iPSCs are a pathological liver model for congenital urea cycle disorders to clarify pathogenesis and develop novel therapies.

In this study, we aimed to establish in vitro models of urea cycle disorders using patient-specific human iPSCs to mimic specific pathological conditions.Human iPSCs were generated from two patients with citrin deficiency and differentiated into hepatocyte-like cells (HLCs).In addition, we compared iPSCs derived from patients with citrin deficiency to those from control.HLCs differentiated from citrin-deficient patient-specific iPSCs exhibited greater concentrations of ammonia than control iPSCs.Patient-derived citrin-deficient iPSCs can be used as a pathological liver model for congenital urea cycle disorders.

Human iPSC culture
Blood samples were taken from two patients with citrin deficiency (CD1 and CD2), and iPSCs were generated as described below.This study was authorized by the ethics committee of Kansai Medical University (approval number: 2021067), and written consent was obtained from the donors.All experiments were carried out in compliance with relevant institutional guidelines and the principles of the 2013 revised Declaration of Helsinki.Peripheral blood mononuclear cells (PBMCs) were isolated using Ficoll solution (Merck KGaA, Darmstadt, Germany) and were reprogrammed into iPSCs.Episomal vectors encoding human factors (OCT3/4, SOX2, KLF4, L-MYC, LIN28, mp53DD and EBNA; addgene, MA, USA) were transfected into PBMCs by electroporation.The cells were cultured on SNL feeder cells treated with mitomycin C (Fujifilm, Osaka, Japan).After 16-25 days, iPSC colonies were individually selected.These cells were cultured in DMEM/Ham's F-12 medium (Fujifilm) supplemented with 10 μg/mL fibroblast growth factor 2 (FGF2, Fujifilm), 20% knockout serum replacement (KSR, Thermo Fisher Scientific, MA, USA), 1% non-essential amino acids (Thermo Fisher Scientific), 55 mM 2-mercaptoethanol (Thermo Fisher Scientific), and 50 U/mL penicillin/streptomycin (Thermo Fisher Scientific).
Normal human iPSCs (585A1, cell number HPS0354, male, 30's, Japanese) were provided by the RIKEN BRC cell bank and cultured in a feeder-free system as described above.HepG2 cells were maintained in DMEM/Ham's F-12 medium supplemented with 10% fetal bovine serum (GE Healthcare, IL, USA).

Sequence variants analysis
The sequence variants analysis was performed on two patient samples (CD1 and CD2) obtained from peripheral blood at the time of clinical diagnosis.CD1 had a compound heterozygous sequence variants of NM_014251.The primer pairs used for the analysis of the c.Ivs11+1G>A, p.E601*, and c.854del4 sequence variants are shown in Table 1.The PCR products were sequenced by the Sanger method on an automated DNA sequencer using the BigDye Xterminator purification kit (Thermo Fisher Scientific) according to the manufacturer's instructions.

Karyotyping and confirmation of differentiation into germ layers
Analysis of chromosomal G-bands of CD1-and CD2-iPSCs was performed by Nihon Gene Research Institute (Miyagi, Japan).For the confirmation of differentiation into germ layers, cells were applied to a V-shaped 96-well plate and the medium was replaced every 2 days with DMEM and B27.After 12 days, cells were collected in ISOGEN II (Nippon Gene, Tokyo, Japan).RT-qPCR was then performed on ectoderm (SOX1), mesoderm (Platelet-derived growth factor receptor α: PDGFRα), and endoderm (AFP, SOX17).The primers are shown in Table 1.

Reverse transcription-polymerase chain reaction (RT-PCR)
RNA was isolated from cells using the ISOGEN II reagent.Reverse transcription was performed using ReverTra Ace qPCR RT Master Mix (Toyobo, Osaka, Japan).PCR was performed using a HotStar Taq Plus Master Mix Kit (Qiagen, Venlo, Netherlands) with 1% agarose (Fujifilm).We used the followings condition for PCR: 30 cycles of 95 • C denaturation for 30 s, 57 • C annealing for 30 s, and 72 • C extension for 30 s. Quantitative RT-PCR (RT-qPCR) was performed using a Rotor-Gene Q (Qiagen) and Thunderbird SYBR qPCR Mix (Toyobo).The specific PCR primers used are listed in Table 1.

Immunocytochemistry
Cells were fixed in paraformaldehyde for 30 min, permeabilized using Triton-X for 10 min, and blocked with Blocking One (Nacalai Tesque, Kyoto, Japan).Cells were then incubated with primary antibody overnight at 4 • C and then secondary antibody was applied for 60 min at room temperature.Nuclei were counterstained using Hoechst 33258, and stained cells were evaluated by fluorescence microscopy (BZ-9000, Keyence).The wavelengths used to image fluorescent signals are 360/ 40 nm for DAPI, 470/40 nm for GFP,620/60 nm for Cy5.
The catalog numbers of antibodies for immunocytochemistry used are listed in Table 2.

Ammonia measurement
HCM and ammonium chloride (NH 4 Cl) were incubated together after 15-20 days of differentiation into HLCs.Supernatants were collected after 3, 6, 24 and 48 h.Ammonia test kits (Fujifilm) were used according to the manufacturer's instructions.

Statistical analyses
Statistical analyses were performed using Graph Pad Prism 5. A oneway and two-way analysis of variances (ANOVA) with a Bonferroni post hoc test was used for assessment.P-values <0.05 were considered significant.

Generation of citrin-deficient patient-specific iPSCs
Citrin-deficient iPSCs (CD-iPSCs) were generated from the peripheral blood of patients CD1 and CD2.CD1 was a male patient diagnosed with NICCD who presented with white stools, poor body weight gain, elevated serum ammonia levels (189 μg/dL) and bile acids, and jaundice.CD1 was a 24-year-old patient carrying the c.Ivs11+1 G>A and p. E601* compound heterozygous sequence variants.CD2 was a 48-yearold female patient treated as schizophrenia for a long time but diagnosed with CTLN2 at age 40, after being suspected by mild liver dysfunction (aspartate aminotransferase, 52 U/L; alanine aminotransferase, 52 U/L), fatty liver and hyperammonemia (234 μg/dL) and hypercitrullinemia (949.5 nmol/mL; normal range, 17.0-43.0nmol/mL) [17].CD2 carries the c.Ivs11+1 G>A and c.851del4 compound heterozygous sequence variants and underwent liver transplantation.The c. Ivs11+1 G>A sequence variant is located in intron 11 and is the most frequently reported sequence variant (36.3%) [18].The second most common sequence variant (32.8%) is c.851del4, which is located in exon 9 [18].c.IVS11+1 G>A and c.851del4 are sequence mutations found in east Asia [18], and we consider that the presence of these sequence mutations in two CD patients enhances the value of the model system we have developed.
Four iPSC-like colonies were generated from CD1 and one from CD2.One colony from each patient was selected and cloned.Both CD1-iPSCs and CD2-iPSCs exhibited normal stem cell morphology (Fig. 1A), and normal karyotypes were confirmed by G-banding analysis (Fig. 1B).The expression of pluripotency-related transcripts, including NANOG, OCT4, and SOX2, was confirmed by PCR (Fig. 1C).585A1-, CD1-, CD2-iPSCs were positive for expression of pluripotent protein also such as OCT4, NANOG, SSEA4, and TRA-1-60 by immunostaining (Fig. 1D).An embryoid body-formation assay was performed, and the differentiation into three germ layers was confirmed: ectoderm (SOX1), mesoderm (PDGFRa), and endoderm (AFP and SOX17) (Fig. 1E).Taken together, these experiments demonstrate the self-renewal ability and pluripotency of CD-iPSCs.In addition, sequencing analysis confirmed that CD1-iPSC carries the c.Ivs11+1 G>A sequence variant and the p.E601* sequence variant with a GAA to TAA substitution.CD2-iPSCs carry the c.Ivs11+1 G>A sequence variant and the c.851del4 sequence variant, which causes a frameshift by the deletion of four nucleotides in exon 9 (Fig. 1F).

Urea cycle dysfunction in CD-HLC
Citrin deficiency is a secondary urea cycle disorder caused by a shortage of aspartate supply from mitochondria to the cytoplasm in the urea cycle [2].To investigate whether ammonia is consumed via the urea cycle in CD-HLCs, NH 4 Cl (1 mM) was added to the medium, and the ammonia concentration in the medium, not cell lysate, was measured.After the addition of NH 4 Cl, the ammonia concentration decreased in 585A1-HLCs, indicating NH 4 Cl consumption in HLCs.However, in CD-HLCs, the ammonia concentration did not decrease, and NH 4 Cl consumption was not observed at every time point (Fig. 4A).After 48 h of NH 4 Cl treatment, ammonia concentration decreased significantly in 585A1-HLCs compared with CD-HLCs (Fig. 4B).This demonstrates a dysfunction in the urea cycle in CD-HLCs.
Sodium pyruvate is an alternative treatment for urea cycle disorders, including citrin deficiency [19].To examine whether the addition of sodium pyruvate improves urea cycle dysfunction, HCM was supplemented with various concentrations of sodium pyruvate from day 11 to day 15 of culture, and then NH 4 Cl was added to the HCM on day 16 for 24 h of co-culture.Sodium pyruvate addition significantly reduced ammonia concentrations in a dose-dependent manner in CD-HLCs.Slight decreases were also observed in 585A1-HLCs, but these changes were not significant (Fig. 4C).Seven technical replicates were performed for Fig. 4A,B and three technical replicates were performed for Fig. 4C.CD1-, CD2-, and 585A1-HLCs were positive for expression of hepatic protein such as ALB, HNF4α, ASGR1, and AFP.HepG2 cells were used as a positive control for HLCs.Scale bar, 100 μm.(D) mRNA expression of hepatocyte-associated genes in HLCs.PCR showed that CD1-and CD2-, and 585A1-HLCs express ALB, ASGR1, and AFP.(E) Immunoblot analysis for citrin and ASS protein in CD1-, CD2-, and 585A1-HLCs.Citrin was not expressed in CD1-and CD2-HLCs and ASS was expressed in CD1-and CD2-HLCs more weakly than in 585A1-HLCs.GAPDH was used as a control.(F) Immunoblot analysis for CPS1, OTC, ASL, and ARG1 protein in 585A1-, CD1-, CD2-HLCs.PHH and HepG2 were used as positive controls.Enzymes of the urea cycle were expressed in 585A1-, CD1-, and CD2-HLCs.GAPDH was used as a control.

Discussion
The urea cycle is an important protein metabolism pathway in the liver that produces urea from ammonia.Urea cycle disorders are genetic diseases characterized by abnormalities in enzymes involved in this cycle, leading to conditions such as hyperammonemia.Enzymes including CPS1, OTC, ASS, ASL, and ARG1 mediate the urea cycle and comprises four amino acids: ornithine, citrulline, argininosuccinic acid, and arginine.[20].Citrin is expressed in the liver, kidney, and heart and is involved in the transport of aspartic acid [1,21].Citrin deficiency causes an inability to supply aspartic acid to the urea cycle, resulting in secondary urea cycle disorders [1].In this study, iPSCs were generated from two citrin-deficient patients with different compound heterozygous sequence variants.The disease mechanism of CD is thought to be loss of citrin function.Therefore, it is highly unlikely that the variant seen in the patient is in cis.Considering this disease mechanism, plus the fact that all three sequence variants are loss-of-function, we consider that the patients in these cases are likely to be compound heterozygotes.On the other hand, CD-HLCs exhibited normal liver function regarding the production of albumin and liver-specific proteins, but also showed loss of citrin protein and decreased expression of ASS.In addition, we found that CD-HLCs exhibit decreased ammonia consumption compared to HLCs derived from control iPSCs.These abnormalities mimic the pathophysiology of citrin deficiency.Although an effective treatment for citrin deficiency has yet to be established (apart from liver transplantation), treatment with sodium pyruvate and arginine has been demonstrated clinically [19].For instance, the administration of sodium pyruvate to CTLN2 patients reduces the frequency of encephalopathy, increases energy intake and BMI, and reduces ammonia levels [22][23][24].Our results clarify that pyruvate administration ameliorates ammonia concentrations in CD-HLCs, suggesting that it might be an adequate treatment for patients with citrin deficiency.Based on these results, CD-HLCs have great potential as a preclinical research tool for analyzing therapeutic efficacy in vitro.
The pathophysiology of each urea cycle disorder differs depending on the site of the lesion, and the optimal treatment must be selected for each disease.For example, patients with urea cycle disorders restrict protein intake to minimize ammonia accumulation, while hyperammonemia in citrin deficiency is treated with a high-protein, lowcarbohydrate diet [25,26].Furthermore, the administration of pyruvate and arginine remains controversial for some diseases.However, there are several problems with urea cycle disorder research; patient sample collection is challenging, especially for organs such as the liver, and differences between human and mouse NADH shuttles prevent the reproduction of full pathophysiological phenotypes in citrin-deficient animal models.The absence of citrin leads to quantitative decrease of ASS and hyperammonemia, but Citrin KO mice do not show hyperammonemia and decrease of hepatic ASS activity [27].It is possible that the high activity of glycerol-3-phosphate shuttle in mice may compensate for any decrease in malate-aspartate shuttle activity by transporting NADH-reducing equivalents back into mitochondria [31].Doubleknockout mice with disrupted mitochondrial glycerol-3-phosphate dehydrogenase (mGPD) are currently being used and manifest hyperammonemia.However, it is not clear how inhibiting the transport of reducing equivalents leads to hyperammonemia.Knocking out mGPD also may have created other metabolic dysfunctions, which limit the application of the double KO model to humans with CD [28].
In this study, HLCs were differentiated from patient iPSCs, thereby enabling the development of novel therapies and research using disease models of specific pathological conditions.Genomic repair of urea cycle sequence variants may enable cell-based clinical treatments without immunosuppression.However, this study has limitations.The iPSCderived HLCs have been supposed to be relatively immature and resemble fetal hepatocytes rather than mature, adult hepatocytes [29].Therefore, the possibility cannot be ruled out that an adequate urea cycle has not been formed.But Fig. 3-F shows that the expression of urea cycle enzymes is similar in both control HLCs and CD-HLCs.Another limitation is that the differentiation induction method does not reproduce liver ammonia metabolism because living hepatocytes have different metabolic functions that vary depending on their localization along the sinusoids within hepatic lobule [30].Mature hepatocytes must be evaluated with a normal urea cycle and improved induction methods.Furthermore, this study is lacking isogenic controls which are considered the best control in such studies and all comparisons are made to a single control cell line.
Previous reports have clarified the effects of L-arginine using HLCs differentiated from citrin-deficient patient-specific iPSCs, but mild improvements in urea levels were observed [31].On the other hand, urea production was measured using control and CD-HLCs by the method of Kim et al.; however, with less than detectable levels in our study.Therefore, we assessed the urea cycle based on ammonia concentration.The possibility of measuring urea, citrulline, and glutamine would be helpful in elucidating the disease mechanism.Moreover, sodium pyruvate significantly reduced ammonia levels in CD-HLCs.The precise mechanism by which sodium pyruvate improves the metabolic abnormalities in CD remains incompletely understood.Recent studies suggest that exogenously supplied pyruvate enter the cytoplasm, and convert to lactic acid.At this conversion, NADH is transformed into NAD + and lead to a reduction in the cytosolic NADH/NAD + ratio.In cases of CD, the ingestion of carbohydrates results in the metabolism of glucose through the glycolytic pathway, leading to the production of NADH and the progression of metabolic dysfunction.However, the introduction of pyruvate helps to counteract this process, alleviate glycolysis inhibition, and provide oxaloacetate and aspartate to activate cytosolic urea production [19,24,32].Additionally, the secondary effects of pyruvate increase cellular energy metabolism [33,34], and a previous study using a liver-perfusion system in citrin knockout mice has revealed that pyruvate ameliorates ureogenesis from ammonia by reoxidizing NADH and supplying energy [35].Although detailed mechanisms must be elucidated, our study is the first in vitro study to demonstrate the efficacy of pyruvate in patients with CD.

Conclusions
In conclusion, we investigated the pathogenesis of the urea cycle disorder using iPSCs derived from patients with citrin deficiency by inducing differentiation into HLCs.Cells derived from patients with urea cycle disorders showed inadequate ammonia metabolism compared with normal HLCs.Sodium pyruvate reduced ammonia in cells from patients with citrin deficiency, but sodium pyruvate did not alter ammonia significantly in cells from healthy subjects.These findings suggest that pyruvate may be an efficient treatment for citrin-deficient patients.These hepatic urea cycle models from human iPSCs are useful tools for studying pathogenesis and developing novel therapies for patients with urea cycle disorders.

Fig.
Fig. Differentiation of iPSCs into HLCs.(A) Timeline and factors involved in the differentiation of iPSCs into HLCs.The differentiation protocol was performed by modifying a previous method.(B) Phase contrast images of differentiated CD2-iPSCs.Magnification: 100× and 200×.Scale bars, 200 μm and 100 μm.(C) Expression of hepatic markers measured by RT-qPCR.β-actin was used as a reference gene.Results are presented as mean ± SD of 4 samples.Statistical analysis was performed using one-way ANOVA with Bonferroni's test.**, P < 0.01 vs undifferentiated iPSCs.

Fig. 4 .
Fig. 4. Evaluation of ammonia consumption in CD-HLCs.Each group was treated with NH 4 Cl for 48 h and ammonia concentrations were measured by absorbance spectrophotometry.(A) Ammonia concentration decreased significantly in 585A1-HLCs compared with CD-HLCs.Closed triangle, 585A1-HLCs; open circle, CD1-HLCs; closed circle, CD2-HLCs.Results are presented as means ± SD of 7 samples.Statistical analysis was performed using two-way ANOVA with Bonferroni's test.*, P < 0.01, **, P < 0.001585A1-HLCs vs CD-HLCs.(B) Ammonia concentration decreased significantly in 585A1-HLCs compared with CD-HLCs at 48 h.Results are presented as means ± SD of 7 samples.Statistical analysis was performed using one-way ANOVA with Bonferroni's test.**, P < 0.01 vs before NH 4 Cl treatment.(C) Effect of sodium pyruvate on ammonia consumption in HLCs.White bar without sodium pyruvate, grey bar with 10 mM sodium pyruvate, black bar with 100 mM sodium pyruvate.Results are presented as means ± SD of 3 samples.Statistical analysis was performed using one-way ANOVA with Bonferroni's test.**, P < 0.01 vs without sodium pyruvate.

Table 1
Primers used in PCR analysis

Table 2
Catalog numbers of antibodies for immunocytochemistry.