Article Text
Abstract
Background A high level of succinylacetone (SA) in blood is a sensitive, specific newborn screening marker for hepatorenal tyrosinemia type 1 (HT1, MIM 276700) caused by deficiency of fumarylacetoacetate hydrolase (FAH). Newborns with HT1 are usually clinically asymptomatic but show liver dysfunction with coagulation abnormalities (prolonged prothrombin time and/or high international normalised ratio). Early treatment with nitisinone (NTBC) plus dietary restriction of tyrosine and phenylalanine prevents the complications of severe liver disease and neurological crises.
Methods and results Six newborns referred for hypersuccinylacetonaemia but who had normal coagulation testing on initial evaluation had sequence variants in the GSTZ1 gene, encoding maleylacetoacetate isomerase (MAAI), the enzyme preceding FAH in tyrosine degradation. Initial plasma SA levels ranged from 233 to 1282 nmol/L, greater than normal (<24 nmol/L) but less than the initial values of patients with HT1 (16 944–74 377 nmol/L, n=15). Four individuals were homozygous for c.449C>T (p.Ala150Val). One was compound heterozygous for c.259C>T (p.Arg87Ter) and an intronic sequence variant. In one, a single heterozygous GSTZ1 sequence variant was identified, c.295G>A (p.Val99Met). Bacterial expression of p.Ala150Val and p.Val99Met revealed low MAAI activity. The six individuals with mild hypersuccinylacetonaemia (MHSA) were not treated with diet or nitisinone. Their clinical course has been normal for up to 13 years.
Conclusions MHSA can be caused by sequence variants in GSTZ1. Such individuals have thus far remained asymptomatic despite receiving no specific treatment.
- Tyrosinemia
- hypersuccinylacetonemia
- MHSA
- GSTZ1/MAAI
- nitisinone
Statistics from Altmetric.com
Introduction
The finding of elevated levels of succinylacetone (SA) has been considered to be pathognomonic for hepatorenal tyrosinemia type 1 (HT1, MIM 276700),1 ,2 a severe autosomal recessive disease that can cause hepatic failure, cirrhosis, liver cancer, renal tubulopathy, rickets and severe porphyria-like neurological crises. These complications are preventable by diagnosis and treatment with nitisinone and dietary restriction of phenylalanine and tyrosine if started during the first month of life.3 HT1 is caused by the deficiency of fumarylacetoacetate hydrolase (FAH), the last enzyme of the degradation of phenylalanine (Phe) and tyrosine (Tyr, figure 1).
Tandem mass spectrometry-based detection of SA is increasingly used for population newborn screening for HT1.4 In the USA, HT1 is included in the Recommended Uniform Screening Panel5 and SA is recognised as the best primary screening marker.6 The longest running neonatal screening programme in the world for HT1 is in Québec, Canada, where HT1 is prevalent because of a genetic founder effect.7 Since screening for HT1 began in 1970, over 3.5 million newborns have been screened. Blood SA has been the first-tier screening marker since 1996.
In Québec, newborns referred for hypersuccinylacetonaemia are tested for liver dysfunction using prolonged prothrombin time (PT) or increased international normalised ratio (INR). In HT1, these measurements are sensitive markers of liver function1 and are available immediately in all referral centres in the province. Individuals referred for hypersuccinylacetonaemia who have high values of PT and/or INR receive the tentative diagnosis of HT1 and are immediately offered treatment. Here we report a subgroup of individuals with mild hypersuccinylacetonaemia (MHSA) who had normal liver function on presentation and who have remained clinically well without treatment.
Materials and methods
Identification and evaluation of individuals with MHSA
In Québec, all hypersuccinylacetonaemic neonates identified by newborn screening are referred to a member of the Québec NTBC Study Group, a group of physicians affiliated with one of five referral centres (CHU de Québec-Université Laval, CHU de Sherbrooke (CHUS), CHU Sainte-Justine, Centre universitaire de santé McGill and Centre Hospitalier de la Sagamie). The initial evaluation includes physical examination and determination of INR and/or PT, and serum alpha fetoprotein (AFP), plasma amino acid chromatography, plasma and urine SA assays and abdominal ultrasound examination.
The clinical approach to hypersuccinylacetonaemia and normal liver function has evolved as clinical experience has grown. In all cases, however, for each neonate with MHSA, two options for clinical management have been explored with the parents, that is, treatment with nitisinone and restricted diet or non-treatment. The following points are discussed. For patients with HT1 detected in Québec, treatment with diet and nitisinone started before 1 month of age effectively prevents liver disease, at least for the two decades of follow-up14 and presumably longer in patients with good adherence. Notions from basic research are discussed, including evidence that SA is a toxin.15 In situations of dietary non-adherence, potential complications of treatment include nutritional deficiencies and hypertyrosinemia with corneal lesions.16 Learning problems have been reported in some patients with HT1 treated with nitisinone and diet.17 Treatment with nitisinone and diet imposes some restrictions on lifestyle and is expensive.18 The consequences of maleylacetoacetate isomerase (MAAI) deficiency in humans are not known. Gene-targeted MAAI-deficient mice, a potential model for humans with MAAI deficiency, are generally asymptomatic but can develop liver failure with metabolic stress like tyrosine loading.8 ,19 For families with MHSA, it is explained that treatment with nitisinone plus a special diet would be expected to normalise SA concentration. Families are offered the choice between treatment with diet and nitisinone or no treatment, with clinical, laboratory and imaging follow-up.
In practice, non-treated individuals with presumed MHSA and normal liver function have been followed, with clinical examination and PT and/or INR. Following the initial encounter, blood and urine SA, AFP and liver ultrasound results become available for individuals at any of the referral centres. Today, individuals with normal clinical examination, normal coagulation testing and SA values in the zone defined in this article are considered to have MHSA. For these patients, examination and measurement of PT and/or INR, plasma AFP, alanine aminotransferase, amino acids and SA and urine SA have been recommended every 6 months, with abdominal ultrasound examination yearly and abdominal MRI every 2 years.
SA determination and AFP levels in classical patients with HT1 and individuals with MHSA
In newborns referred for elevated SA in blood spots, diagnostic SA measurements were performed in plasma and urine by a sensitive method using gas chromatography–mass spectrometry (GC-MS) with stable isotope dilution, essentially as described.20
We assembled the first diagnostic plasma and urine SA measurements and serum AFP measurements obtained from these individuals with normal coagulation tests and from patients with HT1, taken prior to beginning NTBC treatment. Diagnostic SA measurements were all performed in the Biochemical Genetics Laboratory of CHUS, using the above method. AFP values were determined in the clinical biochemistry laboratories of the different participating centres.
Variant detection and in silico predictions
Individuals were tested for the French–Canadian FAH founder mutation, c.1062+5G>A, using an allele-specific PCR method. In individuals not homozygous for this sequence variant, FAH gene sequencing was performed, plus deletion/duplication analysis if both mutant alleles were not identified. If this analysis was negative, GSTZ1 gene sequencing was performed. When available, parental DNA samples were tested to document segregation. Variants in the coding region were examined using SIFT, PolyPhen2 and Mutation Taster algorithms for in silico prediction of their biological effects. To assess the potential effects of the intronic sequence variant on splicing, Human Splice Finder 3.0 (http://www.umd.be/HSF3/HSF.html) was used.
Preparation of bacterial expression plasmids for GSTZ1/MAAI, homogentisate 1,2-dioxygenase and FAH
Normal and mutant human GSTZ1/MAAI cDNAs were cloned for bacterial expression. Also, cDNAs for normal human HGD (homogentisate 1,2-dioxygenase) and FAH were cloned to serve as reagents for the coupled enzyme assay as described below.
Using standard plasmid cloning techniques, we created cDNA constructs. Each was cloned into the pET30a(+) expression vector (Novagen). GST1/MAAI was flanked by BamHI and HindIII sites immediately 5′ and 3′, respectively, of the initiation ATG and stop codons. Similarly, the HGD cDNA (corresponding to sequence NM_000187.3) was flanked by EcoRI and HindIII sites and FAH (corresponding to sequence NM_000137.2) by BamHI and XhoI sites. (Details of cloning are available on request).
Several GSTZ1/MAAI haplotypes are prevalent. The reference cDNA sequence (U86529.1) is on the A haplotype.21 Both missense sequence variants tested in GSTZ1/MAAI were found to be on the D haplotype. The vector sequence was changed to the D haplotype by introducing the following three single nucleotide substitutions in the GSTZ1/MAAI cDNA: G at cDNA position 94 (encoding Glu at codon 32), 124G (Gly42) and 245T (Met82). The sequences of all recombinant plasmids used for expression were confirmed by complete sequencing of the cDNA.
Bacterial expression and purification of recombinant HGD, GSTZ1/MAAI and FAH proteins
Recombinant plasmids were selected in Escherichia coli DH5α, purified and transformed into E. coli BL21. To induce protein expression, bacteria were inoculated into 3 mL of Lysogeny broth (LB) medium containing 50 μg/mL kanamycin, grown overnight at 37°C, diluted into 150 mL LB medium containing 50 μg/mL kanamycin and incubated at 37°C until optical density at 600 nm (OD600) reached 0.6. Then isopropyl-β-d-thiogalactoside was added to a final concentration of 0.2 mmol/L. Incubation was continued at 20°C overnight. Cells were harvested by centrifugation, suspended in 50 mM sodium phosphate, 300 mM NaCl pH 7.5 and then lysed by sonication. Lysates were centrifuged at 12 000 g for 15 min at 4°C. The crude supernatant containing HGD was stored at −80°C. Supernatants containing expressed GSTZ1 (MAAI) and FAH were purified using His Trap HP columns (GE Healthcare, Sweden) as described.22 Purified MAAI and FAH and crude protein extract of HGD were stored at –80° C.
GSTZ1/MAAI activity assays
MAAI activity was measured in two ways. First, the dichloroacetate (DCA) dehalogenation activity of GSTZ1/MAAI was tested, with DCA as substrate and glyoxylate production measured colorimetrically at 535 nm.23 Second, in a coupled assay including HGD and FAH,8 MAA production from homogentisic acid was recorded as absorption at 330 nm, then MAAI was added and MAAI activity was measured as the rate of decrease in absorption.
Results
We describe six individuals referred following newborn screening for tyrosinemia, who had MHSA and normal liver function on initial evaluation (table 1).
Molecular analysis of individuals with MHSA
In all six cases, GSTZ1 sequence variants were detected (table 1). Individual 1 had a premature termination mutation in exon 5, c.259C>T (p.Arg87Ter), predicted to be deleterious by American College of Medical Genetics (ACMG) criteria.24 This mutation was in trans with a variant in intron 2, c.68-12G>A, that is predicted to create a new splice acceptor site at position –10 with respect to exon 3. The resulting transcript would encode three new amino acids in the MAAI peptide and create a frameshift thereafter. Sequencing of GSTZ1 in individuals 4 and 5 revealed that they were apparently homozygous for c.449C>T (p.Ala150Val) in exon 7 of GSTZ1. Individuals 2 and 3 were from the same ethnic/geographic group as individuals 4 and 5; in individuals 2 and 3, the presence of c.449C>T was investigated as a first-line test, both were c.449C>T homozygotes. Individual 6 was heterozygous for the c.295G>A point sequence variant (p.Val99Met) in exon 5; no other variant was detected by sequencing or deletion/duplication analysis of GSTZ1.
Both c.295G>A (p.Ala150Val) and c.449C>T (p.Val99Met) change the codons of conserved amino acids. Both are predicted to be deleterious by the PolyPhen2, SIFT and Mutation Taster programmes (see online supplementary table S1). By ACMG criteria, both are classified as variants of unknown significance. Both c.449C>T and c.295G>A occur on the D haplotype of GSTZ1 because the individuals are homozygous at each of the three coding sequence nucleotides that define this haplotype.
supplementary data
In vitro assay of MAAI activity
Enzymatic activity of GSTZ1/MAAI was assayed in two fashions (figure 2). In an assay using DCA as substrate (figure 2A), c.449C>T (p.Ala150Val) produced 64% of wild-type activity and c.295G>A (p.Val99Met), 86%. MAAI activity was also measured with a coupled assay in which MAA is first produced enzymatically by HGD, then its disappearance is measured after addition of normal or mutant MAAI;13 using this assay, c.449C>T (p.Ala150Val) showed 56% of control activity and c.295G>A (p.Val99Met), 27% (figure 2B).
Clinical and biochemical description of individuals with MHSA
Initial evaluation
In the six individuals with MHSA, plasma SA level at first assay following screening ranged from 233 to 1282 nmol/L (median, 358 nmol/L) versus an assay reference range of <24 nmol/L. By comparison, in the 15 most recently diagnosed patients with mutation-proven HT1 in Québec, all of whom were tested by the same laboratory, plasma SA levels ranged from 16 944 to 74 377 nmol/L (median, 39 454 nmol/L figure 3A). In urine, initial SA concentrations in individuals with MHSA ranged from 73 to 4103 µmol/mol creatine (median, 522) versus <34 μmol/mol creatine for the assay reference range and from 59 921 to 1 195 273 μmol/mol creatine (median, 290 391) in patients with classic HT1 (figure 3B). In both plasma and urine, there was a complete separation between the HT1 and MHSA groups, and a 10-fold or greater difference between the highest MHSA value and the lowest value in patients with HT1. INR and/or PT were normal in all individuals with MHSA. The initial AFP levels of individuals with MHSA were normal. In comparison, the levels of AFP of patients with HT1 before treatment with NTBC were all increased. Initial plasma amino acid profiles of all individuals with MHSA were unremarkable, whereas most patients with HT1 showed elevation of plasma tyrosine.
Follow-up
Each of the six MHSA families chose non-treatment. None received nitisinone or dietary restriction. All agreed to long-term clinical surveillance. None has presented clinically detectable symptoms attributable to liver or kidney dysfunction when followed for up to 13 years. Assessments of liver and kidney function by clinical, biochemical and imaging techniques have been normal. No episodes consistent with tyrosinemic neurological crises25 have been noted.
AFP levels have remained within age-appropriate reference ranges (figure 4A). Figure 4B, C shows the course of plasma and urine levels of SA in individuals with MHSA. SA levels tended to diminish with time although nearly all measurements remained above the reference range. Plasma amino acid profiles and alanine aminotransferase results remained essentially normal in all individuals with MHSA.
Interrogation of the ExAC database
All four GSTZ1 sequence variants described (see online supplementary table S1 and figure S1) were reported in the ExAC database (http://exac.broadinstitute.org/) but were rare. One p.Val99Met homozygote was reported: a person of northern European origin sampled at 41 years of age, for whom additional information is unavailable.
Discussion
We describe a new biochemical condition of importance for clinicians and biochemists involved with newborn screening or with follow-up of ‘screen-positive’ infants. Although experience is limited to a small number of cases, consideration of MHSA is important for accurate diagnosis and for informed treatment decisions.
At the first encounter with a screened neonate referred for hypersuccinylacetonaemia, the physician is confronted with the pressing question of whether treatment with NTBC and diet is required. We have found the presence of liver dysfunction using PT and/or INR as a marker to be useful in this context because it is immediately available in all reference centres. So far, all patients with HT1 have had perturbed coagulation tests. In contrast, no individual eventually designated as MHSA had this. Many medical conditions and technical factors can influence coagulation test results. Hence, prolonged coagulation results in a screened neonate may be due to conditions other than HT1. It is important to rapidly perform specific testing by measurement of SA in plasma and/or urine. SA is specific for MHSA and HT1. Therefore, samples for plasma and urine SA determination are obtained at the initial encounter and results are available within days, even for samples collected in remote centres.
Accurate and sensitive measurement of plasma SA in the nanomolar range has been the most discriminant marker for the diagnosis of MHSA. There is a nearly 1000-fold separation between the plasma SA levels of patients with HT1 before nitisinone treatment (16 944–74 377 nmol/L) and of normal controls (upper reference limit, 24 nmol/L, figure 3A). Individuals with MHSA fell in an intermediate range, with initial plasma SA values of 233–1282 nmol/L (median, 358 nmol/L), about 10-fold above the reference range but about 100-fold less than levels of patients with HT1. Complete separation between these three groups was also observed for the urine SA results (figure 3B). Of note, this description applies to samples obtained following newborn screening, in the first weeks of life. In individuals with MHSA, SA levels tended to be highest in the neonatal period, then to decrease over time, although generally remaining above the normal range.
It is possible that, in jurisdictions outside Quebec, a fraction of individuals with MHSA might be misdiagnosed as having HT1. Inspection of newborn bloodspot SA data available from the Region 4 MS/MS Collaborative Project (http:www.clir-r4s.org/) suggests that some individuals with MHSA described here in our study would likewise have screened positive in a number of other neonatal screening programmes for SA. Online supplementary figure S2A shows cumulative data compiled from 54 participating programmes. Online supplementary figure S2B provides a breakdown of the normal population SA data provided by individual programmes (anonymised) as well as a depiction of the distribution of SA cut-off values used by each programme. More detailed information and discussion, including description of various methodologies for bloodspot SA assay and their possible influence on ranges of observed values, can be found in ref. 6.
However, during follow-up testing of ‘screen-positive’ newborns, the mild elevations of plasma and urine SA of individuals with MHSA would not be detected by all diagnostic laboratories. Most clinical laboratories do not use a dedicated, sensitive assay for SA, relying instead on analysis of urine organic acids by GC-MS without stable isotope dilution. In this setting, typical limits of detection for SA are about 1–2 mmol/mol creatine (1000–2000 μmol/mol creatine), with limits of quantitation about 5–10 mmol/mol creatine, which is sufficient for the diagnosis of classical HT1. Few clinical laboratories measure SA in plasma. Some individuals with MHSA with ‘screen-positive’ results might subsequently have been dismissed as false positives if the confirmatory laboratory testing followed the algorithm recommended in 2009 by the ACMG (http://www.acmg.net). This algorithm is based on urine organic acids, plasma AFP and plasma amino acids, all of which would probably give normal or unremarkable results in most individuals with MHSA, considering our observations with the Quebec's individuals with MHSA described here. However, any individuals with MHSA receiving a presumptive diagnosis of HT1 and treated thereafter with NTBC would become biochemically indistinguishable from patients with true HT1 under treatment, with normal liver function and normal SA. Some MAAI-deficient patients may, therefore, have received or be receiving treatment with NTBC and dietary restriction.
Do the sequence variants in GSTZ1 cause the MHSA observed in our cohort? GSTZ1 encodes MAAI, the only enzyme other than FAH known to be related to SA production (figure 1). All four of the GSTZ1 sequence variants found are predicted to reduce MAAI function and one, c.259C>T (p.Arg87Ter), meets the stringent ACMG criteria for pathogenicity by sequence analysis alone.19 In individual 1, the c.259C>T sequence variant occurs opposite a sequence variant predicted to interfere with splicing and to cause premature termination near the N-terminal of MAAI. Each of these two sequence variants is predicted to severely reduce MAAI capacity. Therefore, individual 1 is predicted to have absent or very low MAAI activity. Interestingly, he also has the highest levels of SA among the individuals with MHSA. In individuals with MHSA, the in vivo level of MAAI activity cannot be deduced with certainty from the current results. Direct measurement of MAAI activity and protein in liver would give a more precise indication of this but, because of the benign clinical course to date, tissue sampling has not been requested and no such samples are available.
The two missense sequence variants in GSTZ1 detected in the other individuals with MHSA were expressed in vitro and tested by enzyme assay. In each case, the level of GSTZ1/MAAI activity in recombinant mutant proteins was lower than normal, although more residual activity was seen than is typical for the causal mutations of many inborn errors.26 Three considerations suggest that these GSTZ1/MAAI sequence variants suffice to produce hypersuccinylacetonaemia. First, the functional effect in cells may be greater than that expected from in vitro measurements. Many missense sequence variants cause increased protein degradation in vivo, which may be underestimated by in vitro expression.26 Second, sequence variants may have effects unrelated to the translated peptide, affecting splicing or transcription, or additional undetected deleterious sequence variants may be present on the mutant allele. Third, and most important, the occurrence of MHSA in partial deficiency of MAAI is predicted by metabolic control theory: metabolite levels are influenced to some extent by any change in the activity of each step of linear or branched metabolic pathways.27 ,28 Mild elevation of SA concentration is, therefore, expected in partial deficiency of MAAI. Conversely, the branched nature of the pathway for MAAI metabolism, in which MAAI has alternative routes of degradation in addition to isomerisation by GSTZ1/MAAI, may contribute to the very modest elevation of SA seen in patients with MHSA with GSTZ1/MAAI sequence variants.
Individual 6 has the lowest levels of SA among the MHSA group and also has a documented deleterious variant on only one GSTZ1 allele. The other five individuals with MHSA have deleterious variants on both GSTZ1 alleles, consistent with autosomal recessive transmission. Perhaps the second GSTZ1 allele of individual 6 has an unidentified deleterious variant. Alternatively, perhaps he has a single abnormal GSTZ1 allele. Clearly, most MAAI-deficient heterozygotes do not have substantial hypersuccinylacetonaemia and they are not identified by newborn screening: even making a very conservative estimate for the incidence of homozygous genetic MAAI deficiency (eg, one per million, less than the observed number of homozygous individuals), nearly 100 heterozygotes would be born annually in Québec. Conversely, even though most MAAI-deficient heterozygotes are not identified by screening, some such individuals might possibly have elevated SA. We speculate that rarely individuals with a single deleterious GSTZ1 allele may have MHSA, and that this phenomenon might be restricted to a small number of specific sequence variants in GSTZ1 or to individuals who also have a low activity of the pathways of MAA metabolism other than MAAI (figure 1). As the other pathways become better known, this hypothesis could be tested directly in apparently heterozygous individuals with mildly elevated SA.
Each MHSA family has so far opted for non-treatment. The lack of detectable liver disease in these individuals after up to 13 years of follow-up is reassuring, but does not guarantee long-term normal hepatic function. Therefore, we offer ongoing surveillance. To date, no evidence of liver dysfunction has been detected in individuals with MHSA, but none of them has had a serious illness that severely stressed the liver. We currently permit an unrestricted natural diet but discourage artificially high protein intake such as consumption of protein supplements, which would increase the substrate load on MAAI and FAH. Maintenance of intracellular levels of glutathione (GSH) is a consideration because GSH is the cofactor of GSTZ1/MAAI, and also because GSH can itself mediate the non-enzymatic degradation of MAAI (figure 1). This pathway may be particularly important for tyrosine degradation if MAAI is deficient. During infections, when protein catabolism is accelerated, if pharmacological treatment of fever is judged to be necessary, we favour the use of non-steroidal anti-inflammatory antipyretics such as ibuprofen rather than acetaminophen, because under some conditions, acetaminophen can deplete GSH.29 Also, GSTZ1/MAAI is essential for and is inactivated by the metabolism of DCA, a drug and an environmental toxin.18 DCA has been used to treat some patients with mitochondrial disease or cancer.30 ,31 Individuals with MHSA are predicted to be slow metabolisers of DCA. One individual heterozygous for c.295G>A (p.Val99Met), described previously in a pharmacological study,32 had a markedly low elimination of DCA. We show directly in vitro that MAAI containing the p.Val99Met substitution degrades DCA slowly. Clinically, we would be cautious about using DCA in individuals with MHSA. We also avoid the use of chloral hydrate, which is degraded in part to DCA.33
An abstract by Berger et al34 described siblings with fatal hepatic and renal failure, in one of whom MAAI enzymatic deficiency was documented. SA was not detected in urine. This entity was designated tyrosinemia type 1b (OMIM 603758). Unfortunately, no samples are now available from these patients or their parents (R Matalon, personal communication, 2016). Admitting the possibility that a small increase in SA may not have been detected using the assay methods and approaches available in 1988, this abstract implies that liver and kidney failure with absent (or mild) hypersuccinylacetonaemia may be a phenotype of MAAI deficiency. We reason that individuals with clinically severe MAAI deficiency would be expected to have levels of SA at least as high as those of the individuals with MHSA reported in this article. If so, such individuals would be detected by newborn screening in Québec. No such patient has been identified among >1.3 million births in Québec since 1996, when SA was adopted as the primary screening marker for HT1. Furthermore, the presence of a 40-year-old c.295G>A (p.Val99Met) homozygote in the ExAC database suggests that prolonged survival is possible despite reduced MAAI function. Like the first patient described as having tyrosinemia, who retrospectively proved to have a phenotype atypical for HT1 (reviewed in reference),1 the fascinating report of Berger et al remains a challenge to current research about MAAI deficiency. Our observations argue that a causally related triad of MAAI deficiency, infantile liver failure and normal SA is rare in humans, and conversely, that elevated SA and normal liver function occurs in at least some people with MAAI deficiency.
In summary, experience to date supports the decision not to treat individuals with MHSA and normal liver function with nitisinone and dietary phenylalanine and tyrosine restriction. MAAI deficiency should be suspected in such individuals. It cannot be excluded that, under some conditions, human MAAI deficiency may produce clinically relevant hepatic or other signs. For this reason, prolonged medical surveillance of individuals with MHSA is planned.
References
Footnotes
WA-H and DC contributed equally.
Collaborators Active members of the Québec NTBC Study Group who read and approved the manuscript: Fernando Alvarez, Catherine Brunel-Guitton, Daniela Buhas, Josée Dubois, Martyne Gosselin, Ugur Halac, Rachel Laframboise, Bruno Maranda, John Mitchell, Guy Parizeault and Jean-François Turcotte. We thank Martyne Gosselin for coordinating the Québec NTBC Study; Mélanie Beaucage for facilitating the sampling of several participants; Tommy Gagnon and Patrick Bherer for expert analytical contributions; Daniel MacArthur, Harvard Medical School, for help with interrogating the ExAC database; Reuben Matalon for updating patient data and André Imbeau for financial support. Data from this article were presented in part at the TYROSINÉMIE2015 Symposium in Chicoutimi, Québec, Canada, on 25 September 2015.
Contributors HY performed expression experiments and wrote part of the manuscript; WA-H contributed patient data; DC devised and performed SA measurements; RL contributed patient data; GP contributed patient data; SPW cloned MAAI and supervised enzymatic studies; FR interpreted molecular analyses; M-TB and YG reviewed the newborn screening data; PJW supervised analysis of SA and wrote part of the article and GAM supervised clinical, molecular and enzymatic investigations and wrote part of the article. All of the above authors and participating Québec NTBC Study members (Fernando Alvarez, Catherine Brunel-Guitton, Daniela Buhas, Josée Dubois, Martyne Gosselin, Ugur Halac, Bruno Maranda, John Mitchell, Jean-François Turcotte) reviewed, criticized the data and approved the article.
Funding This work was funded in part from account FHSJ-AI-TYROSINEMIE-G. Mitchell.
Competing interests None.
Ethics approval Approved as a clinical report by the Research Ethics Board of CHU Sainte-Justine.
Provenance and peer review Not commissioned; externally peer reviewed.