Biochemical and clinical response to a sulfur-restricted diet in ethylmalonic encephalopathy

Abstract

Introduction

Ethylmalonic encephalopathy (EE) is an often-severe inborn error of metabolism caused by biallelic variants in the ETHE1 gene leading to impaired detoxification of hydrogen sulfide (H₂S). H₂S is produced both exogenously by anerobic intestinal bacteria as well as by the endogenous catabolism of the sulfur-containing amino acids methionine and cysteine. Existing therapies including metronidazole, N-acetylcysteine (NAC), and orthotopic liver transplantation (OLT) have been pursued with the objective of reducing or detoxifying exogenously produced H₂S. However, strategies to reduce endogenously produced H₂S using a methionine and cysteine restricted diet are an understudied therapeutic avenue.

Methods

We performed an open-label, single-arm study to evaluate the effects of dietary intervention with a methionine and cysteine restricted diet (20–30 mg/kg/day) on biochemical parameters and overall clinical trajectory in three patients with molecularly confirmed ethylmalonic encephalopathy (two with attenuated phenotypes, one classically affected). All three patients were receiving a combination of medical therapy with metronidazole and NAC and were status-post OLT at the time of diet initiation. Plasma butyrylcarnitine (C4) levels were measured at diagnosis, serially following initiation of medical therapy and OLT, and at regular follow-up visits in a metabolic clinic after diet initiation. Additionally, we obtained untargeted metabolomics studies and directly evaluated ethylmalonate, butyrylcarnitine, isobutyrylcarnitine, isovalerylcarnitine, 2-methylbutyrylcarnitine, glutarylcarnitine, and methylsuccinate levels in the pre- OLT/medical therapy, post- OLT/medical therapy, and post- sulfur-restricted diet states.

Results

We observed a 20–38 % reduction in plasma C4 levels in all three patients following OLT and combination medical therapy with NAC and metronidazole. An 8–10 % reduction in C4 was observed following the introduction of dietary therapy in the two patients with attenuated phenotypes and an 82 % increase in C4 was seen in the patient with the classical phenotype. The metabolic profile as assessed by untargeted metabolomics analysis was largely unchanged in the pre-OLT/medical therapy, post-OLT/medical therapy, and post-diet states.

Conclusions

The modest biochemical response to a sulfur-restricted diet observed in our cohort likely reflects the relatively minor contribution of endogenous sulfur-containing amino acid catabolism to overall H₂S production. Further work is needed to study the impact of dietary intervention on the natural history of EE including diet only trials in the animal model as well as in the pre-OLT period in human participants.

1. Introduction

Ethylmalonic encephalopathy (EE) (MIM#: 602473) is a typically severe, autosomal recessive inborn error of energy metabolism which affects the brain, gastrointestinal system, and vasculature. The classical phenotype was first described in the 1990s by Burlina et al. in a cohort of Italian children with progressive neurological decline and spastic diplegia beginning in infancy as well as orthostatic acrocyanosis and extremity swelling, chronic mucoid diarrhea, and diffuse relapsing petechia [1,2]. Since the original description of EE, there have been limited reports of an attenuated phenotype with variable neuromotor decline and spastic diplegia with minimal or absent vascular and gastrointestinal symptoms [3]. Biochemically, EE is characterized by the urinary excretion of ethylmalonic acid, elevated C4 and C5 acylcarnitine species in plasma, and cytochrome c oxidase (COX) deficiency in muscle and nerve tissue with associated lactic acidemia. Linkage studies by Tiranti et al. mapped the condition to the ETHE1 gene and elucidated its function as a mitochondrial sulfur dioxygenase needed for the catabolism of hydrogen sulfide (H₂S) to thiosulfate and sulfate [[4], [5], [6]]. While there is not a clearly defined genotype-phenotype correlation, the homozygous exon 4 deletion—commonly found in individuals of Mediterranean descent—appears to be associated with the classic phenotype [3,7].Physiologically, H₂S is a vasoactive gasotransmitter produced in large amounts by intestinal bacteria as well as, to a lesser extent, the endogenous catabolism of sulfur-containing amino acids [[8], [9], [10], [11]]. At supraphysiologic concentrations, however, H₂S is toxic to colonic mucosa, vascular endothelial cells, and is a potent inhibitor of COX and short chain acyl-CoA dehydrogenase (SCAD) leading to the classic clinical and biochemical findings [5,12,13]. Briefly, H₂S catabolism begins at the inner mitochondrial membrane with the oxidation of H₂S by sulfide:quinone oxidoreductase which shuttles electrons into the electron transport chain for ATP production and transfers a sulfur atom to glutathione (GSH) to generate glutathione persulfide (GSSH). ETHE1 then oxidizes the outer sulfane sulfur of GSSH, using molecular oxygen, to produce GSH and sulfite. Subsequently, sulfite is further metabolized into either thiosulfate by rhodanese or sulfate by sulfite oxidase —both thiosulfate and sulfate are excreted in urine [6]. Paradoxically, thiosulfate is over-excreted in the urine of EE patients which reflects the metabolism of H₂S by alternative pathways. Advances in the understanding of the biochemical basis of EE has enabled the development of novel treatment strategies including metronidazole to eradicate H₂S producing anaerobic intestinal bacteria as well as NAC¹⁴. The latter acts as a precursor to mitochondrial GSH, providing additional substrate for the production of non-toxic GSSH by sulfide:quinone oxidoreductase [14]. Improved survival was observed in Ethe1^−/− constitutive knock-out mice treated with combination therapy with these two compounds, and clinical, biochemical, and neuroradiographic improvements were seen in a case series of five treated individuals [14]. As the hepatic portal circulation drains H₂S rich venous blood from the intestine, hepatic correction of ETHE1 deficiency should theoretically permit the detoxification of H₂S before reaching the systemic circulation. This effect was first demonstrated by the hepatic correction of Ethe1^−/− in a constitutive knock-out mouse model with an adenoviral vector [15]. Based on the promising findings in the mouse model, liver transplantation was first reported by Dionisi Vici et al. in 2016 in a 9-month-old who demonstrated striking neurological and biochemical improvement [16]. Four additional patients have been subsequently reported in two separate case series by our center all of whom had clinical improvement following transplant, although biochemical abnormalities were not uniformly improved [17,18]. In addition to the H₂S produced by intestinal anaerobes, H₂S is also produced endogenously through the catabolism of the sulfur-containing amino acids methionine and cysteine. Dietary intervention in EE was first reported by McGowan et al. (2004) prior to a complete understanding of EE biochemistry [19]. In that study, urinary excretion of ethylmalonic acid was increased following an oral loading dose of methionine and was reduced following methionine restriction from the diet. Parents reported clinical improvement with increased alertness and vocalization while their child was receiving dietary restriction. In 2018, Boyer et al. reported further clinical and biochemical improvement in an 8-month-old previously on therapy with metronidazole and NAC following the introduction of a methionine and cysteine restricted diet [20]. In this case series, we report the biochemical and clinical findings of three EE patients treated with a methionine and cysteine restricted diet following orthotopic liver transplantation (OLT) and on combination medical therapy.

2. Methods

2.1. Ethics/Institutional review board

This study of dietary intervention for ethylmalonic encephalopathy was conducted in accordance with a protocol that was approved by the Baylor College of Medicine Institutional Review Board (IRB) in accordance with the Helsinki Declaration of 1975, as revised in 2024. Written informed consent was provided by the legal guardians of all participants prior to their participation.

2.2. Sulfur-restricted diet

A methionine and cysteine restricted diet was developed based on a prior report by Boyer et al. (2018) to meet a target methionine and cysteine intake of 20–30 mg/kg/day extrapolated from protocols for sulfite oxidase deficiency and molybdenum co-factor deficiency [20]. The recommended daily allowance for total methionine and cysteine was age and sex matched to each patient to ensure accurate dosing [21]. XMet XCys Maxamaid (Nutricia) as well as SOD Anamix Early Years (Nutricia) were used as the sources of methionine and cysteine free synthetic protein. With respect to intact protein intake, individuals able to take food by mouth were instructed to follow a vegetarian diet since oral intake of food was very limited. Pediasure 1.0 (Abbott) was added to enteral feeds as a source of intact protein for those exclusively fed by gastrostomy tube. Lastly, Duocal (Nutricia) and Pro-Phree (Abbott) were used as a source of protein free nutrition to meet daily estimated energy requirements. Upon review of anthropometrics changes over time and review of plasma amino acid quantifications, formula recipes were titrated at regular visits with a metabolic dietician.

2.3. Standard biochemical evaluation

Plasma butyrylcarnitine was measured by LC-MS/MS as part of a quantitative plasma acylcarnitine profile (ARUP Laboratories, Test Code: 0040033). Creatine (Cr) corrected urine thiosulfate levels was measured by LC-MS/MS (Mayo Clinic Laboratories, Test ID: FSULU).

2.4. Untargeted metabolomics analysis

Untargeted metabolomics analysis was performed as previously described [22,23] by Baylor Genetics (BG) on samples of plasma prior to OLT/medical therapy, following OLT/medical therapy, and following diet initiation. Briefly, small molecules are extracted from blood plasma in an 80 % methanol solution and aliquoted for three separate mass spectrometry assays: (1) gas chromatography coupled mass spectrometry, (2) LC-MS/MS, and (3) LC-MS/MS in ion negative mode. Each sample is run alongside invariant anchor specimens to enable semiquantitative analysis. Integrated intensity values are calculated for each analyte's chromatographic peak and are log transformed and median scaled. Semi-quantitative Z-scores for each analyte are calculated using an in-house reference population.

3. Results

3.1. Clinical case histories

3.1.1. Patient 1(P1)

Patient 1 is a now 10-year-old African-American female, previously described by Lim et al. 2021 [18], who first presented to the general genetics clinic at 4 years of age in the setting of global developmental delay, lower extremity spasticity, and basal ganglia injury on MRI brain. She is the fifth child of a non-consanguineous couple. From the developmental perspective, she met early milestones. She walked at 12 months of age and was able to walk up and down stairs while holding a railing at the time of her initial evaluation. With respect to her fine motor skills, she was able to feed herself with her hands and remove her shoes. With respect to her language, she was able to follow simple one step commands such as “no” and “come here” and had a 20–30-word vocabulary. Whole exome sequencing (WES) revealed likely pathogenic compound heterozygous variants in ETHE1 (NM_014297.3: c.407C > T (p.Thr136Ile) and NM_014297.3: c.263C > T (p.Ser88Leu)) highly suggestive of ethylmalonic encephalopathy. On initial presentation she did not have a history of petechial rashes, loose or frequent stools, or acrocyanosis. Review of newborn screening (NBS) records showed normal first and second screens. Urine organic acid analysis revealed ethylmalonic and methylsuccinic aciduria and acylcarnitine profile revealed markedly elevated butyrylcarnitine of 1174 nmol/L (reference range, < 600 nmol/L). Untargeted metabolomic studies of plasma similarly revealed elevated ethylmalonate (Z-score = +5.48), butyrylcarnitine (Z-score = +3.02), glutarylcarnitine (Z-score = +3.42), and methylsuccinate (Z-score = +1.35). Metronidazole 30 mg/kg/day, N-acetylcysteine 105 mg/kg/day, and levocarnitine 1000 mg daily were started at the time of diagnosis. At 5 years of age, she presented with a significant metabolic decompensation in the setting of a febrile illness and asthma exacerbation with profound lactic acidosis and multi-focal metabolic strokes involving the deep gray nuclei, corticospinal tracts and the supratentorial and infratentorial cortices. This decompensation was associated with significant developmental regression including loss of independent ambulation and ability to feed herself. Given prior report of successful OLT [16] she underwent uncomplicated OLT at 5 years of age. Post-OLT, she was able to regain some developmental milestones including the ability to pull to stand and very limited food intake by mouth. At 8 years of age, a sulfur-restricted diet consisting of metabolic formula and intact protein with a vegetarian diet was started. Following initiation of her diet, she demonstrated continued improvement in her developmental skills including taking some steps with assistance although she has not returned to her prior baseline. Precise quantification of her developmental skills, however, is limited by the lack of formal developmental assessment.

3.1.2. Patient 2 (P2)

Patient 2 is the now 12-year-old older brother of patient 1, previously described by Lim et al. 2021 [18]. He first presented to our clinic at 5 years of age with a similar history of global developmental delay, lower extremity spasticity, and basal ganglia injury. Unlike his sister, however, he did not meet early developmental milestones as he did not crawl until 1 year of age and did not walk until 2 years of age. Similar to his sister, he did not have a history of petechial rashes, loose or frequent stools, or acrocyanosis. Review of NBS records likewise revealed normal first and second screens. With respect to his biochemical studies, urine organic acid analysis revealed ethylmalonic and methylsuccinic aciduria and acylcarnitine profile revealed a markedly elevated butyrylcarnitine of 1532 nmol/L (reference range, < 600 nmol/L). Untargeted plasma metabolomics studies similarly revealed elevated ethylmalonate (Z-score = +5.26), butyrylcarnitine (Z-score = +3.63), glutarylcarnitine (Z-score = +2.20), and methylsuccinate (Z-score = +2.35). Known familial variant testing revealed the presence of the same compound heterozygous variants found in his sister. Metronidazole 30 mg/kg/day, N-acetylcysteine 105 mg/kg/day, and levocarnitine 1000 mg daily were started at the time of diagnosis. At 6 years of age, he presented with an acute metabolic stroke involving the bilateral basal ganglia, in the setting of a streptococcal pharyngitis infection, which was associated with motor regression. He underwent uncomplicated OLT at 6 years of age. Following OLT, he demonstrated improvement in his developmental milestones including cruising, ability to “high-five”, some limited expressive speech including “hi” and “bye”, as well as generally increased interaction with his environment. At 9 years of age a sulfur-restricted metabolic formula and intact protein with a vegetarian diet were started. Following initiation of his diet, he had continued developmental improvement including ability to ambulate with a gait trainer and undress himself independently. Precise quantification of his developmental skills, however, is limited by lack of formal developmental assessment. He has undergone serial neuroimaging studies which have shown stable lesions.

3.1.3. Patient 3 (P3)

Patient 3 is a now 9-year-old female, previously described by Tam et al. 2019 [17], who first presented to care in the context of a borderline elevated C4 of 1.41 μmol/L (borderline >1.3; abnormal >1.8) and an elevated C4/C2 ratio of 0.1168 μmol/L (elevated >0.05) on first NBS. She is the second child of consanguineous Jordanian parents. Follow-up in the newborn period was not pursued due to her family's move outside of the country. At four months of age, she developed profuse non-bloody diarrhea, petechia, and failed to achieve further developmental milestones. MRI brain performed at 9 months of age at an outside facility revealed signal abnormalities in the bilateral basal ganglia and corona radiata. Her family returned to the United States when she was 13 months old and re-established care with our clinic. On initial developmental history at 13 months old she was unable to sit without support, lift her head in a prone position, or roll over. She did smile socially and regarded others. With respect to her biochemical studies, urine organic acid analysis revealed ethylmalonic and methylsuccinic aciduria and acylcarnitine profile revealed markedly elevated butyrylcarnitine of 2313 nmol/L (reference range, < 600 nmol/L). Untargeted metabolomics studies of plasma similarly revealed elevated ethylmalonate (Z-score = +4.93), butyrylcarnitine (Z-score = +2.95), and methylsuccinate (Z-score = +2.84), glutarylcarnitine was not detected. Molecular testing ultimately revealed a homozygous deletion in exon 4 of ETHE1 prompting initiation of treatment with metronidazole 30 mg/kg/day and N-acetylcysteine 105 mg/kg/day which led to a modest improvement in her petechia and loose stools. She underwent OLT at 18 months of age which was complicated by intraoperative hepatic artery thrombosis and ischemia of the left lobe of the donor liver. Despite revision of her hepatic artery anastomosis and use of thrombolytics, persistent thrombosis of the anastomosis led to chronic left hepatic lobe infarction and smoldering biliopathy. Despite these complications, she experienced developmental gains post-transplant including imitation of speech and ambulation with a gait trainer. At 5 years of age, she started the sulfur-restricted diet, she was fed exclusively per naso-jejunostomy and naso-gastrostomy tube and received measured intact protein with Pediasure 1.0. Clinically, she demonstrated improvement in her chronic emesis and petechia, although she did develop worsening diarrhea. She ultimately underwent repeat OLT at 6-years-of age which was complicated by intraoperative portal vein thrombosis and post-operative hepatic artery thrombosis leading to smoldering ischemic biliopathy, recurrent cholangitis, and ultimately the development of cirrhosis with portal hypertension and varices, without portosystemic shunting.

3.2. Biochemical response to sulfur-restricted diet

Plasma butyrylcarnitine levels were measured at diagnosis, throughout the post-OLT/medical therapy period, and at regular metabolic clinic follow-up visits while on diet (Fig. 1). Butyrylcarnitine levels at diagnosis were 1174 nmol/L for P1, 1532 nmol/L for P2, and 2313 nmol/L for P3 (reference range, < 600 nmol/L). Mean butyrylcarnitine levels across multiple measurements in the post-OLT/medical therapy period were 944 nmol/L (SD = 85, n = 2) for P1 (20 % reduction), 949 nmol (n = 1) for P2 (38 % reduction), and 1449 nmol/L (SD = 551, n = 6) for P3 (37 % reduction). Mean butyrylcarnitine levels across multiple measurements in the post-diet state were 851 nmol/L (SD = 214, n = 4) for P1 (10 % reduction), 873 nmol/L (SD = 122, n = 6) for P2 (8 % reduction), and 2640 (n = 1) for P3 (82 % increase). Cr corrected urinary thiosulfate levels were measured serially in the post-OLT/medical therapy and post-diet states (Fig. 1). In the post-OLT/medical therapy state, mean Cr corrected urinary thiosulfate levels were 77 mg/g Cr (SD = 60, n = 2) for P1 (reference range, < 7.8 mg/g Cr), 64.5 mg/g Cr (SD = 48, n = 4) for P2, and 138.5 mg/g Cr (SD = 86, n = 2) for P3. In the post-diet state, mean Cr corrected urinary thiosulfate levels were 54.9 mg/g Cr (SD = 48, n = 3) for P2 (15 % reduction) and 39.6 (SD = 17.7, n = 5) for P3 (71 % reduction). Post-diet measurements were not available for P1.

Fig. 1.

Plasma and urine biomarkers measured at diagnosis and over the course of treatment. (A) Plasma C4 levels (mean, ± standard deviation). (B) Creatine corrected urinary thiosulfate levels (mean, ± standard deviation). Upper limit of normal for each assay is represented by the dashed horizontal line. Abbreviations: OLT, orthotopic liver transplant; MT, medical therapy.

With respect to the metabolomics analysis, we directly evaluated the key EE associated metabolites: ethylmalonate, butyrylcarnitine, isobutyrylcarnitine, isovalerylcarnitine, 2-methylbutyrylcarnitine, glutarylcarnitine, and methylsuccinate in the pre-OLT/medical therapy, post-OLT/medical therapy, and post‑sulfur restricted diet states (Fig. 2). Ethylmalonate levels were highly elevated in the pre-OLT/medical therapy (mean Z-Score = +5.22), post-OLT/medical therapy (mean Z-score = +4.06), and post-diet states (mean Z-score = +4.53). Butyrylcarnitine levels were elevated pre-OLT/medical therapy (mean Z-Score = +3.20), post-OLT/medical therapy (mean Z-score = +3.45), and post-diet states (mean Z-score = +3.76). Isobutyrylcarnitine levels were elevated in the pre-OLT/medical therapy (mean Z-Score = +2.88), post-OLT/medical therapy (mean Z-score = +2.08), and post-diet states (mean Z-score = +2.51). Glutarylcarnitine levels were elevated in the pre-OLT/medical therapy (mean Z-Score = +2.81), decreased in the post-OLT/medical therapy (mean Z-score = +1.64), and increased post-diet states (mean Z-score = +3.84). Isovalerylcarnitine levels were elevated in the pre-OLT/medical therapy (mean Z-Score = +1.82), post-OLT/medical therapy (mean Z-score = +1.84), and post-diet states (mean Z-score = +2.27). 2-methylbutyrylcarnitine was decreased in the pre- OLT/medical therapy state (mean Z-Score = +1.50) and the post- OLT/medical therapy (mean Z-Score = +0.76) and increased in the post-diet state (mean Z-Score = +1.97). Methylsuccinate levels were elevated in the pre-OLT/medical therapy (mean Z-Score = +2.18), decreased post-OLT/medical therapy (mean Z-score = +0.33), and increased in the post-diet states (mean Z-score = +1.74). Samples for post-OLT/medical therapy metabolomics analysis were obtained 20 days post-OLT for P1, 7 days post-OLT for P2, and 33 days post-OLT for P3. Samples for post-diet metabolomics analysis were obtained approximately one-year on diet for P1 and P2 and approximately one and half years on diet for P3.

Fig. 2.

The profile of EE diagnostic metabolites measured by semi-quantitative untargeted metabolomics in the pre-OLT/medical therapy, post-OLT/medical therapy, and post- sulfur-restricted diet states. Heat map visualization of EE diagnostic metabolites, Z-score values are displayed within cells; missing data are shown as an empty cell marked with an “X”. Abbreviations: OLT, orthotopic liver transplant; MT, medical therapy.

In addition to directly evaluating diagnostic metabolites, we also took an agnostic approach to metabolomics profiling. One thousand one hundred and thirty-two unique metabolites were detected in at least one sample across all runs. We extracted metabolites with an average Z-score across samples of greater than +1.75 or less than −1.75 in the pre-transplant state. Among the 20 metabolites with an average Z-score > +1.75, all seven EE diagnostic metabolites were detected (Fig. 3).

Fig. 3.

The metabolome of EE patients in the pre-OLT/medical therapy, post-OLT/medical therapy, and post- sulfur-restricted diet states. 1132 unique metabolites were detected in at least one sample across all runs. Shown are metabolites with an average Z-score across samples of greater than +1.75 or less than −1.75 in the pre-transplant state. Among the 20 metabolites with an average Z-score > +1.75, all seven EE diagnostic metabolites were detected (red star). Z-score values are displayed within cells; missing data are shown as an empty cell marked with an “X”. Abbreviations: OLT, orthotopic liver transplant; MT, medical therapy.

4. Discussion

This open-label, single-arm study aimed to assess the effect of dietary intervention with a sulfur-restricted diet on the clinical manifestations and biochemical profile of three participants with ethylmalonic encephalopathy. With respect to the biochemical phenotype, plasma C4 levels and urinary thiosulfate levels at diagnosis were greater in P3, the classically affected patient detected by NBS, compared to the siblings P1 and P2 who presented in early childhood to general genetics clinic with an attenuated, predominately neurodevelopmental phenotype. This is consistent with a recent large literature review of 70 patients reporting a statistically significant difference in C4 levels between classical and attenuated phenotypes, supporting the use of C4 as a clinically useful biomarker [3]. A reduction in plasma C4 levels ranging from 20 to 38 % was observed in all three patients following OLT and combination medical therapy with NAC and metronidazole consistent with prior reports of biochemical improvement with these interventions [[16], [17], [18],20]. A more modest reduction in C4 ranging from 8 to 10 % was observed in the two patients (P1 and P2) with an attenuated phenotype following the introduction of dietary therapy. However, an 82 % increase in C4 was seen in the patient with the classical phenotype (P3) following diet initiation which exceeded the initial value at the time of diagnosis. This is most likely explained by the substantial transplant complications faced by this patient, including hepatic artery and portal vein thrombosis, which ultimately progressed to liver cirrhosis with portal hypertension by the time a post-diet acylcarnitine profile was obtained.

Pre-OLT/medical therapy urinary thiosulfate levels were not available for all three participants. However, a decrease in urinary thiosulfate excretion was seen following diet initiation with a 15 % reduction seen in P2 and 71 % reduction seen in P3, although there was large variation seen in these samples. Interestingly, while global untargeted metabolomics analysis provided an agnostic profile of our patients' biochemical states which correlated with predicted changes in EE, it was not changed in the post-OLT/medical therapy state. This is most likely explained by the limitations of this semi-quantitively methodology and emphasizes the importance of quantitative, validated biochemical studies when assessing response to interventions in inborn errors of metabolism. While our untargeted metabolomics analysis identified classic EE associated metabolites, we did not identify alterations in redox metabolites including reductions in NAD⁺/NADH and glutathione previously reported in untargeted metabolomics studies of EE patient-derived fibroblasts [24]. Similarly, we did not observe markers of impaired bioenergetics including alterations in Kreb's cycle intermediates [24]. We suspect that this discrepancy likely reflects differences inherent to cell versus biofluid based metabolomics studies. Notably, the fibroblast extract based metabolomics study by Sahebekhtiari et. (2016) did not report elevations in the canonical EE metabolites.

While a sulfur-restricted diet has not been previously studied in the Ethe1^−/− knock-out mouse, extraintestinal conditional knock-out studies have provided insight into the contribution of endogenous sulfur-containing amino acid catabolism to H₂S production compared to the exogenous contribution of the enteric flora. Interestingly, brain, muscle, and liver specific conditional knock-out mice do not show the clinical or biochemical hallmarks of the constitutive knock-out animal suggesting that the overall contribution of endogenous sulfur containing amino acid catabolism to H₂S production in EE is relatively minor [13]. The findings of this study are supportive of this observation from the animal model.

A major limitation and confounding factor in this study is that all participants were status-post OLT and on combination medical therapy with NAC, carnitine, and metronidazole at the time of diet initiation. This may have masked small effects on the biochemical profile and confounds the evaluation of the clinical trajectory. Notably, the vascular complications and progressive cirrhosis of the transplanted liver observed in P3 offers a natural experiment to gain insight into the contribution of the sulfur-restricted diet in the setting of impaired hepatic correction. At the time that the post-diet acylcarnitine profile was obtained, the patient had developed significant portal hypertension with the return of C4 levels to the pre-transplant range, which would suggest that the diet provided minimal biochemical benefit. Importantly, the underlying H₂S mediated vasculopathy associated with P3's classical EE phenotype may have contributed to the vascular complications of her two transplants including portal vein and hepatic artery thrombosis and it is currently unknown how dietary therapy might reduce or prevent this. Therefore, future work implementing the diet in the mouse model as well as in the pre-transplant period in human patients together with formalized neurodevelopmental assessment may uncover subtle biochemical and clinical benefits.

CRediT authorship contribution statement

Steven H. Lang: Writing – review & editing, Writing – original draft, Investigation, Formal analysis, Data curation. Andres Caceres Salgado: Writing – review & editing, Project administration, Investigation, Formal analysis, Data curation. Matthew T. Snyder: Writing – review & editing, Methodology, Data curation, Conceptualization. Brandy Rawls-Castillo: Writing – review & editing, Project administration, Methodology. Aaron Williams: Writing – review & editing, Project administration. Charul Gijavanekar: Writing – review & editing, Formal analysis, Data curation. Sarah H. Elsea: Writing – review & editing, Project administration, Data curation. Xia Wang: Formal analysis, Data curation. Mary Elizabeth M. Tessier: Writing – review & editing, Methodology. Claudia Soler-Alfonso: Writing – review & editing, Writing – original draft, Supervision, Project administration, Methodology, Investigation, Funding acquisition, Conceptualization. Fernando Scaglia: Writing – review & editing, Writing – original draft, Resources, Project administration, Methodology, Investigation, Formal analysis, Data curation, Conceptualization.

Funding

This work has received no external funding.

Declaration of competing interest

The authors have no financial or conflicts of interest to disclose.

Data availability

Data will be made available on request.