Progress and challenges in development of new therapies for urea cycle disorders

Abstract

Urea cycle disorders (UCD) are inborn errors of metabolism caused by deficiency of enzymes required to transfer nitrogen from ammonia into urea. Current paradigms of treatment focus on dietary manipulations, ammonia scavenger drugs, and orthotopic liver transplantation. In the last years, there has been intense preclinical research aiming at developing more effective treatments for UCD, and as a result, several novel approaches based on new knowledge of the disease pathogenesis, cell and gene therapies are currently under clinical investigation. We provide an overview of the latest advances for the development of novel therapies for UCD.

Introduction

Urea cycle disorders (UCD) are inborn errors of liver metabolism caused by deficiency of enzymes required to transfer nitrogen from ammonia into urea. They have high mortality in both neonatal and late onset cases. Cornerstones of UCD treatment are restriction of dietary protein and oral or intravenous medications activating alternative pathways for ammonia clearance. These drugs include sodium benzoate and sodium phenylacetate, which undergo hepatic conjugation with glycine and glutamine respectively to produce hippuric acid and phenylacetylglutamine that are excreted in the urine (1). Orally administered phenylacetate is poorly palatable, and thus, preferred alternatives are its prodrugs, sodium phenylbutyrate and glycerol phenylbutyrate. In healthy volunteers, phenylbutyrate monotherapy was more effective than benzoate at disposing nitrogen (2). Arginine becomes a conditionally essential amino acid in UCD, and thus, supplementation with arginine, or its precursor citrulline, is needed. Deficiency of N-acetylglutamate synthetase is the only UCD in which ureagenesis may be improved by N-carbamoyl-l-glutamate (3). Dialysis is one of the most rapid methods to reduce acute elevations of plasma ammonia. Hemodialysis is more efficient than peritoneal dialysis in ammonia clearance, and it is currently recommended as treatment of choice (4). Compared to hemodialysis, peritoneal dialysis is less used, although it offers flexibility and autonomy of patients and is attractive in emergency acute hyperammonemia when specialized hemodialysis facilities are not always available. Given these considerations, a liposome-supported peritoneal dialysis to entrap and remove ammonia has been developed (5). However, further studies are needed to confirm the safety and efficacy of this approach (6).

Figure 1

Summary of therapeutic strategies under investigation in urea cycle disorders.

Table 1

Developmental stage of main therapies reviewed herein

Therapeutic approach	Status
Liver and hepatocyte transplantation:
OLT	Clinical trial observational (NCT02740153: active, not recruiting)
LCT	Clinical stage Phase II (NCT00718627: completed)
Mesenchymal stem cells	Clinical stage Phase I/II (NCT01765283: completed)
NO-based and small molecule drugs:
NO-supplementation	Clinical trial (NCT02252770 and NCT03064048: both recruiting)
FXR agonists	Preclinical
Autophagy enhancers	Preclinical
Protein and mRNA replacement therapy:
Delivery of human modified ARG1 protein	Clinical trial Phase II (NCT03378531: recruiting)
Administration of codon-optimized human OTC mRNA complexed with lipid-based nanoparticles	Clinical trial Phase I/II (NCT03767270: not yet recruiting)
Gene replacement therapy:
Administration recombinant AAV8 encoding human OTC	Clinical trial Phase I/II (NCT02991144: recruiting)
Genome editing:
In vivo correction of mutated murine Otc gene	Preclinical
Microbiome engineering:
Ammonia consuming bioengineered bacteria (SYNB1020) Administration of SYNB1020 in hepatic insufficiency and cirrhosis patients	Clinical trial Phase I (NCT03179878: completed) Clinical trial Phase I/II (NCT03447730: recruiting)

In addition to liposomal dialysis, several novel therapeutics with different mode of action have been investigated at both preclinical and clinical levels (Fig. 1 and Table 1). Here, we provide an overview of the latest advances in the development of novel therapies for UCDs.

Liver and hepatocyte transplantation

Orthotopic liver transplantation (OLT) is currently the only method for long-term restoration of hepatic urea cycle activity. Ammonia levels normalize in liver transplant recipients, but the biochemical defect is not corrected in other tissues. Therefore, in argininosuccinate lyase (ASL) deficiency in which the defective enzyme is more widely expressed, the systemic phenotype persists post-OLT. A recent retrospective review of data from the United Network for Organ Sharing Database including 278 pediatric and adult UCD patients who had received OLT (7) showed that 5-year patient survival was 89%, whereas allograft survival ranged from 77% to 92%, depending on the age at transplant. Limited outcome data suggest that OLT may stabilize or improve neurocognitive outcomes, particularly when performed early in life (8,9). More data are expected by a current study investigating differences in mortality, neurocognitive outcomes, and quality of life between UCD patients treated medically or receiving OLT (NCT no. 02740153).

Human heterologous liver cell transplant (LCT) has been proposed as an alternative or “bridge” therapy until OLT, particularly in infants with highest risks and complications. Compared to OLT, surgical complications are significantly less in LCT (10). Cells may be prepared and cryopreserved from human donor organs, including those not suitable for whole-organ transplantation, and potentially multiple recipients may benefit from a single donor (11). Cells are infused via portal vein and must engraft into the native liver. However, either transplanted cells must have a selective growth advantage over recipient cells or endogenous hepatocytes need to be removed from the liver to make space for proliferation and expansion of transplanted cells and to achieve liver repopulation with corrected hepatocytes. In tyrosinemia type 1, hepatocytes corrected for the fumarylacetoacetate hydrolase deficiency have a strong selective growth advantage and space is generated by the enzyme defect affecting viability of hepatocytes (12). In contrast, transplanted cells in UCD have limited ability to proliferate and do not have a selective growth advantage over diseased cells. Therefore, they are frequently lost over time due to rejection or cellular senescence (13). For these reasons, LCT has shown only a modest increase in ureagenesis that does not appear to be clinically relevant (10,14). Therefore, LCT may have attenuated but did not prevent hyperammonemia crises (10). In addition to immunosuppressive therapies, LCT recipients required continuation of ammonia-scavenging therapies and dietary protein restriction, although liberalization of the diet was well tolerated in some patients (13).

Compared to mature hepatocytes, transplantation of mesenchymal stem cells that can differentiate into hepatocytes in vivo offers several potential advantages, including a greater capacity to repopulate after engraftment and immune-tolerogenic properties (15). Additionally, stem cells can be produced in large-scale culture, and they are not highly dependent on donor organ availability. A phase 1/2 clinical trial showed partial restoration of ureagenesis in some stem cell recipients that appeared to peak at 6 months following initial infusion (16).

An increasing number of preclinical studies have shown that autologous hepatocytes derived from induced pluripotent stem cells (iPSCs) are effective in repair of damaged livers (17). Patient's specific hepatocytes derived from iPSCs can be corrected by genome editing tools and infuse into the patient by the procedure used for hepatocyte transplantation (18). However, this approach is hampered by some of the same limitations of LCT, i.e. the lack of selective growth advantage, which is not a feature for UCD, and accordingly, transplantation of genome-edited hepatocytes resulted in only modest improvement of survival of mice with argininemia (19).

Nitric oxide-based and small molecule drugs

Besides impaired ammonia detoxification, deficiency of ASL was found to be associated with nitric oxide (NO)-deficiency resulting in systemic manifestations (20). Hence, NO donors were successfully used to rescue several defects in ASL-deficient mice such as the reduced body weight, reduced survival, and increased blood pressure (21,22). Efficacy of NO synthase-independent NO source (e.g. sodium nitrite) was further shown in an ASL-deficient subject with severe hypertension refractory to antihypertensive medications (21). Based on these studies, there are ongoing clinical trials evaluating NO supplementation in ASL-deficient patients (NCT nos. 02252770 and 03064048) (23). However, NO synthesis does not appear to be affected in proximal defects of the urea cycle, including ornithine transcarbamylase (OTC) deficiency (24). Therefore, at this time NO sources can be considered as a treatment only for ASL and possibly argininosuccinate synthetase 1 (ASS1) deficiencies. Nevertheless, this treatment does not affect hyperammonemia but only those disease features due to NO depletion, such as increased blood pressure and possibly others (e.g. neuropsychiatric symptoms).

The recent recognition of interactions of the urea cycle with bile acid metabolism (25) and autophagy (26) has expanded the spectrum of potential interventions to improve ammonia detoxification in UCD. Activation of farnesoid X receptor (FXR), a bile acid-dependent regulator of metabolism, was found to stimulate amino acid catabolism and ammonia detoxification through ureagenesis and glutamine synthesis (25). Moreover, the FXR agonist obeticholic acid was found to induce enzymes of the urea cycle, namely ASS1, ASL and arginase 1 (ARG1), and to reduce hyperammonemia (25). However, whether this molecule that is given to patients with primary biliary cholangitis (27) is effective in improving hyperammonemia in UCD murine models remains to be investigated. Furthermore, enhancement of liver autophagy was found to promote ammonia clearance by increasing ureagenesis as consequence of improved recycling of intermediates and energy (26). These studies also suggest that enhancement of hepatic autophagy is an attractive target for therapy of UCD.

Protein and mRNA replacement therapy

Conventional protein-based enzyme replacement therapy can be used in lysosomal storage diseases because lysosomal enzymes are naturally taken up by cells through the mannose-6-phosphate receptor. On the other hand, cytosolic or mitochondrial enzymes are generally not taken up by cells, and therefore, urea cycle enzyme replacement therapy to restore hepatic ammonia detoxification presents a greater challenge. However, argininemia might be a candidate for enzyme replacement therapy. Hyperammonemia is uncommon in argininemia, and most neurologic manifestations are secondary to increased blood levels of arginine. Infused pegylated human recombinant ARG1 used to deplete arginine in cancer patients was effective at reducing plasma arginine in murine models of argininemia (28). Although mouse survival did not improve likely because the pegylated enzyme did not enter hepatocytes and did not improve hyperammonemia that accounts for lethality, administrations of arginase have clinical potential because elevated plasma arginine rather than severe hyperammonemia is responsible for the neurological phenotype of arginase deficiency (28). The therapeutic potential of recombinant ARG1 in patients with argininemia is currently under evaluation in a phase 2 clinical trial (NCT no. 03378531).

In contrast to protein replacement, delivery of mRNA to the liver is more readily available for UCD and has shown promising results in inborn errors of liver metabolism. Delivery of mRNA is based on systemic administration of lipid nanoparticles containing synthetic mRNA engineered to be immunologically inactive and stable (29). Intravenously injected mRNA molecules are largely taken up by the liver, and once they have entered into hepatocytes, they use the cell translational and posttranslational machineries to produce the therapeutic protein, overcoming the hurdles of recombinant protein therapy. Moreover, in contrast to viral vector-mediated gene delivery, mRNA therapy has no risks of insertional mutagenesis (30). A drawback of this approach is the requirement of multiple mRNA infusions because of short-term duration of delivered mRNA inside the cells. Repeated administrations of human OTC mRNA normalized blood ammonia and improved survival in OTC-deficient mice (31). Safety and tolerability of an optimized mRNA therapeutic agent are under clinical investigation in patients with OTC deficiency (NCT no. 03767270). A similar approach with ARG1 mRNA has also been investigated in cell lines but not yet in an argininemia mouse model (32).

Extracellular vesicles were found to restore both activity and urea production through the transfer of ASS1 enzyme and mRNA in ASS1-deficient cells (33). These vesicles are membrane bound particles secreted by cells in the extracellular milieu and have been implicated in intercellular exchange of proteins, lipids, and nucleic acids. This observation suggest that extracellular vesicles might be used as cargoes for a combined delivery of protein and mRNA. However, several issues related to production and characterization of such vesicles for clinical use and risk of immune rejection need to be carefully evaluated in vivo to establish the clinical relevance of this approach.

Gene replacement therapy

UCD have long been considered good targets for gene therapy because of their severity and the need to deliver the therapeutic gene only to the liver to achieve normalization of ammonia. Multiple vector systems have been investigated, but overall, adeno-associated viral (AAV) vectors have emerged as the most attractive for liver-directed gene therapy based on their safety and efficacy. Correction of the urea cycle via AAV-mediated liver-directed gene therapy has been reported in mouse models of OTC deficiency (34), ASS1 deficiency (35), ASL deficiency (23,36), and ARG1 deficiency (37). Nevertheless, its translation into the clinic has been a major challenge due to: (1) the cell-autonomous enzyme deficiency that requires a high percentage of hepatocyte gene transfer to achieve phenotype correction; (2) the achievement of adequate levels of transgene expression with sufficient rapidity in the context of a severe neonatal disease; and (3) the loss of non-integrating vector genomes as a consequence of liver growth. Long-term correction of urea cycle activity might thus require vector readministration, which in turn is hampered by induction of immune responses against the reinfused vector (38). Furthermore, early timing of gene delivery along to very high doses appear to be critical factors for genotoxicity and risks of hepatocellular carcinoma after AAV gene delivery, at least in rodents. Notably, in mice that developed cancer, the AAV was integrated in the microRNA-341 within the Rian locus that has no orthologs in the human genome (39). Moreover, several long-term studies including larger animal models such as dogs (40) and non-human primates (41,42) have not raised concerns about AAV liver genotoxicity thus far. Nevertheless, the number of AAV-treated subjects remains small, and careful follow-up and surveillance are warranted.

Preclinical studies in OTC deficient mice (34,43) have led to a currently ongoing phase 1/2, open-label dose-finding safety clinical trial study investigating a serotype 8 AAV (AAV8) vector encoding OTC in adults with late-onset OTC deficiency (NCT no. 02991144). For other UCD, there have been preclinical studies, but they have not yet resulted in clinical trials. A single intravenous injection of an AAV8 expressing ASS1 under the control of a liver specific promoter resulted in reduced plasma ammonia and citrulline concentrations and significant phenotypic improvement of survival, growth, and skin phenotype (35). Among UCD, citrullinemia is attractive because blood citrulline levels is a readily available disease biomarker.

Compared to other UCD, ASL-deficient patients have a systemic phenotype, which involves brain, liver, kidney, gut and peripheral arteries. This observation is consistent with the high rate of neurocognitive impairment, epilepsy, and ataxia despite a lower rate of hyperammonemia episodes in this disorder compared to other UCD. These features make ASL deficiency a difficult target for gene therapy, and indeed, liver-directed gene therapy resulted in correction of the ureagenesis defect but not the remaining phenotypic defects, particularly the brain disorder (21,44).

Neonatal intravenous injections of AAVrh10 delivering ARG1 to liver cells were effective in rescuing biochemical and neurological abnormalities of argininemia mice (37), and these results are consistent with the hypothesis that cerebral involvement occurs as a consequence of the hyperargininemia, as previously discussed.

Carbamoyl phosphate synthetase 1 (CPS1) deficiency is particularly challenging for gene therapy because of its severity often resulting in early neonatal death and large size of the gene encoding CPS1 that cannot be packaged within AAV vectors (45). AAV vectors can indeed accommodate sequences up to 4.5–5 kb in size, whereas the inclusion of sequences greater than 5 kb reduces significantly the in vivo potency (46). Strategies to overcome this obstacle include smaller version of the gene encoding a functionally active CPS1 enzyme that still need to be developed or AAV dual vector strategy in which a large transgene is split into halves that are packaged separately and provided as a dual AAV vector mix. By this strategy, a target cell needs to receive a copy of both halves of transgene that once combined lead to generation of an mRNA transcript encoding the complete large gene (47).

Genome editing

Genome editing is the correction of mutant sequences in the genomic DNA and is achieved by delivering chimeric proteins made of a DNA-sequence-specific binding domain and an endonuclease inducing DNA site-specific double-strand breaks, and a template encompassing the wild-type sequence to be used as a substrate for repair by homology-directed repair. Genome editing tools include zinc-finger nucleases, transcription activator-like effector nucleases, and the clustered regularly interspaced short palindromic repeats (CRISPR)/CRISPR-associated protein 9 (Cas9). In vivo genome editing is based on delivery to target cells of multiple vectors or lipid nanoparticles carrying components of the editing machinery (48,49). However, sustained expression of delivered editing machinery can lead to DNA damage and improper genome editing that ultimately might predispose to cancer. Moreover, off-target cleavage is a limitation of genome editing tools (50). Improvements in guide-design and new generation of CRISPR/Cas9 with higher specificity are critical for efficient on-target cleavage and low levels of off-target cleavage (51,52).

In OTC-deficient mice, a genome editing strategy entailing injection in newborn mice of two AAV vectors, one expressing Cas9 and the other a guide RNA and the donor DNA, resulted in correction of the mutation in about 10% of hepatocytes. However, in adult OTC-deficient mice, the percentage of gene correction was lower and was associated with worsening of the phenotype and larger deletions that affected also the Otc gene (53). These studies illustrate that this approach is not yet ready for clinical translation and requires careful further optimization in preclinical studies.

Synthetically engineered bacteria for ammonia detoxification

The field of synthetic biology and microbiome engineering holds great promise for treatment of several human diseases through development of smart microbes that can detect or treat diseases (54). Ammonia is produced by bacterial metabolism in the gut. Antibiotics are used to inhibit growth of urease-producing bacteria in the gut, thus decreasing ammonia production and absorption through the gastrointestinal tract. For example, rifaximin, a minimally absorbed antibiotic, has shown efficacy in hyperammonemia, but its long-term use raises concerns for the development of antimicrobial resistance. Inoculation of predefined gut microbiota with minimal urease expression in place of preexisting microbiota was found to be effective at reducing fecal ammonia and improving neurobehavioral deficits and mortality in a liver failure (thioacetamide-induced) mouse model of hyperammonemia (55). However, the efficacy of this approach has not been investigated in UCD. In contrast, a simpler approach based on administration of a genetically engineered ammonia hyperconsuming strain of Lactobacillus plantarum was found to reduce ammonia levels and mortality in rodents with hyperammonemia, including OTC-deficient mice (56). Moreover, an orally delivered probiotic Escherichia coli Nissle 1917 engineered to convert ammonia into arginine was effective in reducing systemic ammonia levels and improving survival in mouse models of hyperammonemia, including OTC-deficient mice. A phase 1 dose-escalation study in adult healthy volunteers showed that this bioengineered bacteria was well tolerated and showed in vivo activity (57) and is currently being evaluated in human patients with cirrhosis (NCT no. 03447730) and could be next considered in UCD.

Conclusions

Despite current therapies, cumulative morbidity is still high in UCD, and thus, several approaches have been investigated to improve clinical outcomes. Recent mechanistic insights have shown the involvement of ASL in systemic NO production, and autophagy and FXR in ureagenesis and glutamine synthesis. Interestingly, this new knowledge has potential for the development of novel therapies.

In UCD, cell therapies have generally shown limited efficacy and clearly require further development to overcome the limited engraftment and short-term correction of transplanted cells. Although gene therapy has suffered clinical setbacks, new vector developments and results from clinical trials of liver-directed gene therapy in hemophilias (58,59) have sparked enthusiasm for the development of gene therapy in UCD, particularly OTC deficiency. Gene therapy for other UCD such as CPS1 and ASL deficiencies is more challenging because of either large size of the gene or need of gene transfer to multiple tissues besides the liver to achieve phenotypic correction. Clinical translation of gene therapy is hampered by the high percentage of hepatocyte required for correction, loss of correction due to liver growth, genotoxicity and immune responses against viral vectors. Nevertheless, strategies to overcome these obstacles can be foreseen through optimization of available gene transfer and genome editing tools. In the meantime, repeated dosing of mRNA might be more readily available and can be potentially be used as a bridge to gene replacement therapy once liver growth is completed.

Acknowledgements

N.B.-P. is supported by the European Research Council (IEMTx) and Fondazione Telethon, Italy (TCBP37TELC and TCBMT3TELD). N.A.M's work is generously supported by the Rashid Family Fund.

Conflict of Interest statement. None declared.