Two commonly reported incidental variants in OTC are associated with late-onset disease

Summary

Asymptomatic individuals with pathogenic variants in OTC, the gene encoding ornithine transcarbamylase are increasingly being identified through cascade testing, carrier screening, or as secondary findings from genome-wide sequencing tests. However, guidance for counseling and management of such individuals is currently lacking. We selected two common OTC variants for phenotypic and functional characterization: NM_000531.6:c.118C>T p.(Arg40Cys) and NM_000531.6:c.1061T>G p.(Phe354Cys). The former is the most frequently reported pathogenic/likely pathogenic missense variant present in gnomAD, and the latter has been frequently encountered in our clinical practice. We performed a retrospective chart review at our center, queried the database of the Urea Cycle Disorders Consortium, and performed a literature review to create cohorts of individuals with these variants. Functional studies were pursued using a validated yeast-based assay. We identified 14 individuals (6 females, 8 males) with the p.(Arg40Cys) variant and 14 individuals (5 females, 9 males) with the p.(Phe354Cys) variant. There were no reported episodes of neonatal hyperammonemia in males and no hyperammonemic events reported in females with either variant. In our functional assay, both variants reduced yeast growth to the hypomorphic range. Our findings support the classification of both p.(Arg40Cys) and p.(Phe354Cys) variants in OTC as hypomorphic variants that are typically associated with late-onset OTCD in males.

Incidental pathogenic or likely pathogenic variants in OTC, the gene encoding ornithine transcarbamylase are increasingly reported in the context of widespread genetic testing. These findings provide opportunities for preventing life-threatening hyperammonemia but also raise challenges for the management of asymptomatic individuals. In this study, we report clinical and functional studies of two OTC common variants.

Introduction

Urea cycle disorders (UCDs) are inborn errors of metabolism characterized by an inability to efficiently dispose of waste nitrogen generated from normal protein turnover, catabolism of surplus dietary protein, and catabolism of endogenous proteins for energy production in stressed states.¹ The urea cycle, in its complete form, exists exclusively within the hepatocytes and consists of five “classical” catalytic enzymes, a regulatory co-factor synthesizing enzyme, and two transporters linking the mitochondrial and cytoplasmic compartments. Blockades in this cycle lead to hyperammonemia and its associated neurotoxicity.

Ornithine transcarbamylase (OTC) catalyzes the second step of the urea cycle through the addition of the carbamoyl group of carbamoyl phosphate (CP) to ornithine (ORN) to form citrulline and phosphate within the mitochondrial matrix. OTC deficiency (OTCD) (MIM: 311250), an X-linked disorder, is the most common UCD and accounts for over half of all UCDs.² Phenotypic manifestations of OTCD may occur in hemizygous males and heterozygous females. There is wide phenotypic variability, ranging from severe neonatal-onset disease to a non-neonatal onset (i.e., late-onset) phenotype, which is associated with hyperammonemia predisposition in males and females and with manifestations later in infancy through adulthood.³ The phenotype in females is variable and ranges from apparently asymptomatic to subtle chronic symptoms, including protein aversion, recurrent gastrointestinal complaints, developmental delay, executive functioning deficits, and psychiatric disturbances, and to infantile-onset hyperammonemia with acute liver failure.^4,5

OTC is included in many carrier screening panels due to the risk for life-threatening hyperammonemia and availability of therapies to prevent this complication. It is included in the American College of Medical Genetics and Genomics (ACMG) recommended list of conditions that should be included in carrier screening and is on the ACMG list of reportable secondary findings from exome and genome sequencing.^6,7,8Thus, incidentally, pathogenic and likely pathogenic variants in OTC are increasingly identified in apparently asymptomatic individuals.^9,10

We selected two commonly encountered OTC variants for phenotypic and functional characterization. We identified NM_000531.6:c.118C>T p.(Arg40Cys) as the most frequently reported pathogenic/likely pathogenic missense variant in OTC in the gnomAD version 4.1.0 database, with an estimated population frequency of 1 in 64,800 for hemizygous males and 1 in 36,900 for heterozygous females.¹¹ A similar frequency of approximately 1 in 57,300 was reported in a recent study using two large US biobanks.¹⁰ The NM_000531.6:c.1061T>G p.(Phe354Cys) variant was selected as it has been frequently encountered in our clinical practice and has an estimated population frequency of 1 in 371,000 for hemizygous males and 1 in 101,000 for heterozygous females in gnomAD.¹¹ These two variants have been reported in males with acute hyperammonemia and in apparently asymptomatic males who were ascertained through carrier screening, family testing, or as secondary findings. Both variants are classified as pathogenic by most laboratory submissions in ClinVar.^{10,12,13,14,15,16,17,18,19,20} Moreover, in a yeast-based, high-throughput screening assay of 1,570 OTC variants, both variants reduced yeast growth to a hypomorphic range.²¹ To understand the biochemical and phenotypic consequences of these variants and to inform counseling, we conducted a clinical study and replicated functional studies of OTC enzyme activity in this yeast model system.²¹

Subjects, material, and methods

Chart, literature, and database review

We performed a retrospective chart review of individuals who were heterozygous or hemizygous for the NM_000531.6:c.118C>T p.(Arg40Cys) or NM_000531.6:c.1061T>G p.(Phe354Cys) variants and who were evaluated in the Texas Children’s Hospital (TCH) Metabolic Genetics Clinic from 1995 to 2024. This chart review was conducted in accordance with a protocol approved by the Baylor College of Medicine’s institutional review board (IRB).

To identify additional individuals with these variants, we searched the Longitudinal Study of Urea Cycle Disorders database. This multicenter natural history study performed by the Urea Cycle Disorders Consortium (UCDC) of the NIH Rare Diseases Clinical Research Network includes data on 928 participants with various UCDs. The search included data collected between March 2006 and October 2024. From 2007 to 2015, study procedures were approved by the IRBs of all participating clinical sites. From 2015 onward, the study was conducted under a protocol that was approved by the UCDC central IRB at Children’s National Medical Center. All participants or their legal guardians provided informed consent prior to participation in the study. To supplement these two searches, we performed a literature review for publications or abstracts containing these two variants using PubMed, Google Scholar, Mastermind,²² and LitVar2.²³

From these sources, the following data (if available) were collected from all subjects: sex, age at diagnosis, reason for testing, presenting symptoms at time of first metabolic decompensation, provisional diagnosis at time of first metabolic decompensation, triggering factors preceding initial decompensation, plasma ammonia and peak plasma ammonia level at first hyperammonemic episode, interventions for treatment of first hyperammonemic episode (e.g., use of nitrogen scavengers, hemodialysis), plasma amino acid profile and urine orotic acid level at diagnosis, presence or absence of symptoms of chronic hyperammonemia, long-term therapies (e.g., protein-restricted diet, citrulline/arginine supplementation, oral nitrogen scavengers), age at last evaluation, number of lifetime crises, lifetime peak glutamine and citrulline levels, presence of neurological sequelae, and survival. OTCD was categorized as neonatal onset if metabolic decompensation occurred within 4 weeks after birth and as late onset if clinical manifestations occurred after 4 weeks of age.

Functional yeast-based assay and structural modeling

In a validated assay for human OTC protein function,²¹ yeast cells lacking the yeast OTC ortholog (arg3Δ) do not grow in medium lacking arginine, whereas arg3Δ mutant cells harboring a codon-harmonized version of human OTC (yOTC) grow robustly. In this assay, the yeast OTC ortholog ARG3 open reading frame was knocked out and replaced with a selectable kanamycin-resistance gene to generate the arg3Δ deletion strain (null). yOTC, which encodes the human protein but more closely matches the codon utilization frequency of ARG3, was generated. yOTC was integrated into the yeast genome under the control of the native ARG3 promoter to generate a wild-type strain. Yeast expression constructs harboring the p.(Arg40Cys) and p.(Phe354Cys) variants were similarly transformed into the arg3Δ deletion strain. Six individual isolates of each variant were tested. Strains were grown to saturation in 96-well plates in rich liquid media (containing arginine) overnight at 30°C. Then, cells from the source plate were robotically pinned in triplicate to solid agar plates containing minimal media (lacking arginine) and grown at 30°C for 72 h. Plates were photographed and a pseudo-patch “volume” for each pinned isolate was generated from the images, consisting of the product of the patch area and the mean patch pixel intensity (grayscale). Lastly, a linear model was used to obtain variant growth estimates, as previously described,²¹ producing estimates in the range null (arg3Δ) = 0 to wild-type (yOTC) = 1.

A ribbon diagram structure of human OTC was generated in PyMol (PyMOL Molecular Graphics System, version 3.0, Schrödinger) using the crystal structure of human OTC complexed with the bi-substrate CP-ORN analog N-phosphonacetyl-l-ORN (PDB: 1OTH).²⁴

Results

Phenotype associated with OTC p.(Arg40Cys) variant

We identified 14 individuals (6 females, 8 males) from 9 families; 10 were from the literature search,^{10,12,13,14,15,16,17} 3 from TCH, and 1 from the UCDC database (Table 1). These individuals were identified due to acute hyperammonemia (n = 6/14), cascade testing due to a family member’s diagnosis (n = 4/14), secondary findings from genome-wide sequencing tests (n = 2/14), carrier screening (n = 1/14), or prenatally in the context of positive maternal carrier screening (n = 1/14). The clinical and biochemical phenotypes are detailed in Table S1.

Table 1.

Clinical features of individuals with the p.(Arg40Cys) and p.(Phe354Cys) variants in the literature, the UCDC database, and at our center

	p.(Arg40Cys)	p.(Phe354Cys)
Demographics
Total n	14	14
Males n	8	9
Females n	6	5
Manner of presentation
Neonatal hyperammonemia n, males	0	0
Neonatal hyperammonemia n, females	0	0
Late-onset hyperammonemia n, males	6	2
Median age of first decompensation, years (range)	25 (9–66)	14 (13–68)
Late-onset hyperammonemia n, females	0	0
Reason for ascertainment of variant cascade testing n, males	1	4a
Cascade testing n, females	3	1
Carrier screening n, males	0	0
Carrier screening n, females	1	2
Prenatal diagnosis n, males	1	2
Prenatal diagnosis n, females	0	1
Incidental finding n, males	0	1
Incidental finding n, females	2	1
Follow-up
Death at initial decompensation n	3	1
Age range of unaffected males at last follow-up, years	0.25–40	3–55
Pre-/perinatal complications n, femalesb	0	0
Review of public databases
gnomAD (version 4.1.0) n, males	6	1
gnomAD (version 4.1.0) n, females	11	4

^aIndividual P28 was identified at 55 years old after his grandson (P15) was diagnosed with OTCD and died. He ultimately presented with his first episode of acute hyperammonemia at 68 years old.

^bExplicitly known for P14 (R40C) and P24 (F354C). The prenatal course for P14 was notable for severe morning sickness and numerous episodes of prolonged fasting and, ultimately, the uneventful delivery of an unaffected child with peripartum and postpartum metabolic precautions. For P24, the delivery of P22 was complicated by uterine atony and postpartum hemorrhage. Serial ammonia levels in the peripartum period were normal.

All six individuals (P1, P2, P5, P7, P8, P9) identified due to symptomatic hyperammonemia were males (Table S1). The median age of first hyperammonemic episode was 25 years (range 9–66 years). Identifiable triggers for hyperammonemia were present in all but one individual. Three males died from their first hyperammonemic crisis; two had been discharged after symptomatic improvement with intravenous fluids and/or antibiotics without a definitive diagnosis only to develop fatal encephalopathy in the days following discharge (P5 and P8). OTC enzyme activity levels from liver tissue were available for P1 and P2 and were between 0% and 3%.¹²

There were no instances of neonatal hyperammonemia in males in our cohort with the p.(Arg40Cys) variant. The two males identified prenatally or by cascade testing are without known hyperammonemia at 3 months and 40 years, respectively, at last follow-up. None of the six females in our cohort had documented hyperammonemia, including during parturition (P14).

Phenotype associated with OTC p.(Phe354Cys) variant

We identified 14 individuals (5 females, 9 males) from 8 families (Table 1). Six individuals were identified from the literature,^18,19,20 5 from TCH and 3 from the UCDC longitudinal study. These individuals were identified due to acute hyperammonemia (n = 2/14), cascade testing due to a family member’s diagnosis (n = 5/14), secondary findings from genome-wide sequencing tests (n = 2/14), carrier screening (n = 2/14), or prenatally in the context of positive maternal carrier screening (n = 3/14). The clinical and biochemical phenotypes are detailed in Table S2.

Both individuals identified because of symptomatic hyperammonemia were males, ages 13 and 14 years (P15 and P16). P15 died because of a hyperammonemic crisis presumably triggered by a protein-rich meal. He had spontaneous improvement but developed fatal encephalopathy 1 day after discharge that was refractory to nitrogen scavengers and hemodialysis. No information for P16 was available apart from genotyping and age at presentation.¹⁹ In P15, 1.8% residual OTC activity was detected in liver tissue at autopsy.¹⁸

There were no instances of neonatal hyperammonemia in males in our cohort with the p.(Phe354Cys) variant. Five of six males identified prenatally or by cascade testing are without known hyperammonemia at between 3 and 55 years of age at last evaluation. None of the five females in our cohort had hyperammonemia, including during parturition (P24).

Functional yeast-based assay and structural modeling

A functional yeast-based assay was used to measure the impact of the p.(Arg40Cys) and p.(Phe354Cys) human OTC variants.²¹ The assay exploits the requirement of OTC activity for yeast growth in the absence of Arg3, the yeast OTC ortholog. In arginine-lacking minimal media, the arg3Δ deletion strain is unable to grow unless complemented by the activity of human OTC. Thus, the level of yeast growth can be used as a proxy for enzymatic activity of a particular variant, relative to wild-type. Multiple isolates of yeast, each containing one of the variants integrated into the genome in single copy, along with the arg3Δ deletion strain (relative growth = 0.0) and OTC wild-type strain (relative growth = 1.0), were assessed for growth on minimal medium lacking arginine. Normalized growth estimates for the variants in this assay were 0.276 (SE = 0.0047) for p.(Arg40Cys) and 0.361 (SE = 0.0050) for p.(Phe354Cys) (Figure 1). These findings are consistent with prior estimates²¹ and place both variants in the hypomorphic range (5%–90% of wild-type growth), which is associated with late-onset disease.²¹ Neither variant occurs at a region of substrate binding or catalysis, including the CP binding pocket, ORN binding pocket, or SMG loop (Figure 2). The latter contains an invariant Ser-Met-Gly motif and is involved in the capping of the active site pocket following binding of ORN and CP.²⁴

Figure 1.

Both variants of interest grow in the hypomorphic range in a validated yeast-based functional assay

arg3Δ yeast cells are arginine auxotrophic and thus cannot grow in arginine-free media. arg3Δ yeast cells harboring the codon-harmonized version of human OTC, yOTC, grow robustly. Cells harboring the p.(Arg40Cys) and p.(Phe354Cys) variants have growth estimates of 0.276 (SE = 0.0047) and 0.361 (SE = 0.0050), respectively, when normalized to the null and wild-type growth.

Figure 2.

Both variants of interest are located outside of key catalytic and substrate binding domains

Crystal structure of human OTC complexed with the bi-substrate CP-ORN analog N-phosphonacetyl-l-ornithine (PALO) as elucidated by Shi et al.²⁴ The C-terminal extension (residues 345–354) is shown in green. The SMG loop (residues 264–276) is shown in orange. The positions of residues of interest, Phe354 and Arg40, are depicted with stick representations in gold and pink, respectively. Image generated with PyMol (PDB: 1OTH).

Discussion

The latest revision of the Society for the Study of Inborn Errors of Metabolism suggested guidelines for diagnosis and management of UCDs poses an important question: “what interventions are appropriate in which situations?” However, specific guidance for managing asymptomatic individuals with pathogenic or likely pathogenic OTC variants identified incidentally or prenatally is lacking.²⁵

This study attempts to inform clinical decision-making for the most commonly reported pathogenic/likely pathogenic variant in OTC p.(Arg40Cys) in gnomAD,¹¹ a reference population database in which the majority of participants are not expected to have a severe Mendelian phenotype, and a second variant frequently encountered in our clinical practice, p.(Phe354Cys). We leveraged clinical data from the literature, a retrospective chart review at a tertiary care center for UCDs, and a longitudinal cohort of individuals with OTCD coupled with functional data from a validated yeast assay. These findings support the classification of both variants as hypomorphic variants causative of late-onset disease in males.

While late-onset OTCD has, on average, lower morbidity and mortality,²⁶ a recent Japanese retrospective cohort study comprising 82 individuals with late-onset OTCD reported that nearly 25% of individuals required hemodialysis at the time of first hyperammonemic episode and that 26% of these individuals suffered permanent neurologic sequalae or death.²⁷ Diagnostic delay is likely a contributor to these poor outcomes. In our case series, two individuals presented with their first metabolic decompensation with a priori knowledge for risk of hyperammonemia, which enabled prompt initiation of appropriate treatment and survival of their first decompensation without neurological sequelae.

While the prevention of hyperammonemic encephalopathy and its associated sequalae is the focus of the metabolic management of OTCD, a “pauci-symptomatic” OTCD phenotype is increasingly recognized in apparently asymptomatic individuals with normal biochemical studies. Neurocognitive testing, including in highly educated individuals with OTCD, reveals a neurobehavioral profile consistent with a nonverbal learning disability with deficits in visuospatial reasoning, motor dexterity, and executive functioning, which may be unmasked when challenged.⁵ Functional near-infrared spectroscopy studies, which use a wearable optical device to measure cortical hemodynamics, have demonstrated deficient activation of the left prefrontal cortex during executive functioning tasks.²⁸ Neurobiochemical abnormalities, including depletion of myoinositol in posterior and frontal white matter on magnetic resonance spectroscopy suggestive of deep white matter injury, have also been reported in asymptomatic individuals.²⁹

Certainly, all apparently asymptomatic individuals should be made aware of signs and symptoms of hyperammonemia and potential triggering factors. These include fasting, perioperative period, rapid weight loss, high-protein intake, intercurrent illnesses, medications like valproic acid and systemic corticosteroids, and pregnancy and parturition. Along with this counseling, an emergency letter should be provided outlining appropriate interventions in the setting of metabolic crisis. Emergency bracelets should be considered.

Clinical decision-making may be guided further by comprehensive neuropsychiatric testing and multimodal neuroimaging including magnetic resonance spectroscopy, although guidelines for their use are not currently available. Eventually, such biomarkers may be used to further risk stratify individuals and help establish the basis for a linear or logistic regression model-based risk score calculator.³⁰

The key limitations of this work include the retrospective design and small sample sizes. Additionally, phenotypic information in published case reports was limited. Our own cohort lacked the deep phenotyping, which could have enabled the detection of more subtle effects of urea cycle dysfunction. Lastly, underrecognition of UCDs in adults may have led to underreporting of affected individuals in the literature.

In conclusion, clinical and functional data support the p.(Arg40Cys) and p.(Phe354Cys) variants in OTC as causative of late-onset hyperammonemia in males. The recommended interventions for long-term management of apparently asymptomatic individuals identified through carrier screening, cascade testing, or secondary findings, however, remain without a satisfactory answer. However, anticipatory guidance regarding potential risk factors for hyperammonemia and emergency letters indicating the need to check ammonia levels in the setting of intercurrent illness or symptomatic presentations may be lifesaving. As the cohort of such individuals continues to expand through widespread genetic testing, so does the opportunity to identify biomarkers and other predictive tools for risk stratification.

Data and code availability

All relevant phenotypic information for patients in the TCH or UCDC cohorts are available in the supplemental tables provided. The ClinVar accession numbers for the p.(Arg40Cys) and p.(Phe354Cys) variants are VCV000011013.30 and VCV000097104.23, respectively.

Acknowledgments

This was work funded in part by the UCDC (grant no. U54HD061221), which is part of the National Institutes of Health (NIH) Rare Diseases Clinical Research Network (RDCRN), supported through a collaboration between the National Center for Advancing Translational Sciences, the Eunice Kennedy Shriver National Institute of Child Health and Human Development (NICHD), and the National Institute of Diabetes and Digestive and Kidney Diseases. The UCDC has also been supported by the O’Malley Foundation, the Rotenberg Family Fund, the Dietmar Hopp Foundation, the Kettering Fund, and the National Urea Cycle Disorders Foundation. Research reported in this publication was also supported in part by the Clinical Translational Core of the Baylor College of Medicine Intellectual and Developmental Disabilities Research Center (grant no. P50HD103555) that is funded by the NICHD. R.S.L., G.A.C., and A.M.D. were funded by a grant (no. R01HD114863) from the NIH/NICHD. The content is solely the responsibility of the authors and does not necessarily represent the official views of the NIH. L.C.B. holds a Career Award for Medical Scientists from the Burroughs Wellcome Fund. M.T.S. was supported by the T32 GM07526 Medical Genetics Research Fellowship Program.

Declaration of interests

The authors declare no competing interests.