What we know that could influence future treatment of phenylketonuria.

Phenylketonuria (PKU), a Mendelian autosomal recessive phenotype (OMIM 261600), is an inborn error of metabolism that can result in impaired postnatal cognitive development. The phenotypic outcome is multifactorial in origin, based both in nature, the mutations in the gene encoding the L-phenylalanine hydroxylase enzyme, and nurture, the nutritional experience introducing L-phenylalanine into the diet. The PKU story contains many messages including a framework to appreciate the complexity of this disease where phenotype reflects both locus-specific and genomic components. This knowledge is now being applied in the development of patient-specific therapies.

The birth prevalence of PKU in populations of European, South Asian and sub-Saharan African ancestry living in South East England.

Phenylketonuria (PKU) is an autosomal recessive inborn error of metabolism (OMIM 261600). Treatment with a low-phenylalanine diet following early ascertainment by newborn screening prevents impaired cognitive development, the major disease phenotype in PKU. The overall birth prevalence of PKU in European, Chinese and Korean populations is approximately 1/10,000. Since the human PAH locus contains PKU-causing alleles and polymorphic core haplotypes that describe and corroborate an out-of-Africa range expansion in modern human populations, it is of interest to know the prevalence of PKU in different ethnic groups with diverse geographical origin. We estimated PKU prevalence in South East England, where a sizeable proportion of the population are of Sub-Saharan African or South Asian ancestry. Over the period 1994 to 2004 167 children were diagnosed with PKU. Using birth registration and census data to derive denominators, PKU birth prevalence per 10,000 live births (95% Bayesian credible intervals) was estimated to be 1.14 (0.96-1.33) among white, 0.11 (0.02-0.37) among black, and 0.29 (0.10-0.63) among Asian ethnic groups. This suggests that PKU is up to an order of magnitude less prevalent in populations with Sub-Saharan African and South Asian ancestry that have migrated to the UK.

Recommendations for locus-specific databases and their curation.

Expert curation and complete collection of mutations in genes that affect human health is essential for proper genetic healthcare and research. Expert curation is given by the curators of gene-specific mutation databases or locus-specific databases (LSDBs). While there are over 700 such databases, they vary in their content, completeness, time available for curation, and the expertise of the curator. Curation and LSDBs have been discussed, written about, and protocols have been provided for over 10 years, but there have been no formal recommendations for the ideal form of these entities. This work initiates a discussion on this topic to assist future efforts in human genetics. Further discussion is welcome.

CpG methylation accounts for a recurrent mutation (c.1222C>T) in the human PAH gene.

The human PAH gene (GenBank: U49897.1 (cDNA), AF404777 (gDNA)) harbors alleles that either cause or are associated with hyperphenylalaninemia and phenylketonuria (http://www.pahdb.mcgill.ca). Mutation analysis has identified approximately 500 alleles of which approximately 30 produce polymorphic core haplotypes. The c.1222C>T allele (p.R408W) is the most prevalent and widely encountered PKU-causing allele. Because it occurs on multiple locus-specific polymorphic haplotypes, it is probably not identical by descent in different populations. This mutation involves a CpG dinucleotide in a so-called "hypermutable" codon suggesting that c.1222C>T could be a recurrent allele following spontaneous methylation-mediated deamination of 5 mC. This concept is widely assumed and accepted but the 5mC status of hypermutable codons has seldom been confirmed. We show that the PAH c.1222C nucleotide is indeed methylated (c.1222 mC) in somatic genomes (leukocyte and brain) of H. sapiens. Examination of a representative region in exon 12 (and also in exon 7) in the PAH gene shows that 5 mC is restricted to cytosines in CpG dinucleotides in the hypermutable codons. The methylation pattern seen in human PAH exon 12 was also observed in the corresponding codon in three nonhuman primates. The finding offers at least one explanation for the high relative frequency of the c.1222C>T (p.R408W) allele in the human population.

After the genome--the phenome?

What next? The Human Genome Project signifies complexity rather than simplification in the relationship between genotype and phenotype. Genotypes are embedded in genomes. Individuality in phenotypes is embedded in components of the phenome (transcriptome, metabolome, proteome, etc.). The phenome, its layers, and its nodes, links and networks, require elucidation; there is a need for a Human Phenome Project (Freimer and Sabatti 2003). Biology has largely been a reductive science in the recent past; integrative biology lies ahead. Clinician-scientists (including human biochemical geneticists) will be recognized as key participants in the 'medical' Phenome Project as it reveals components of individuality, and their contributions, in simple or combinatorial fashion, to Mendelian and complex traits; better ways to treat 'genetic disease' will be by-products of the project. Although the Word is common to all, most men live as if each had a private wisdom of his own.Herakleitos

Translating knowledge into practice in the "post-genome" era.

The Human Genome Project is "completed", but it is only a beginning in the understanding of genomic structure and function. A "human phenome project" is waiting in the wings. The complexity involving a phenotype can be glimpsed, for example, if one enquires into the relationships between mutant phenylalanine hydroxylase (PAH) genotypes and the clinical disorders called PKU/Hyperphenylalaninemia-so called lessons from PKU genotypes and phenotypes. Since genomes speak biochemistry, not phenotype (said RHA Plasterk), for genomics to penetrate medicine, biochemistry and biology must be allies. The ideal translators and ambassadors of the knowledge that must cross the gap between laboratory and bedside are the clinician scientists; restoration of that attenuated community of colleagues is a necessary step in the implementation of genomic medicine.

Effect of gene dose and parental origin on bone histomorphometry in X-linked Hyp mice.

X-linked hypophosphatemia (XLH) is characterized by rickets and osteomalacia and arises from mutations in the Phex and PHEX genes in mice (Hyp) and humans, respectively. The present study was undertaken to examine the effect of gene dose on the skeletal phenotype using a histomorphometric approach. Metrical traits (vertebral length, growth plate thickness, cancellous osteoid volume per bone volume, and cancellous, endocortical, and periosteal osteoid thickness) were compared in caudal vertebrae of mutant female (Hyp/+, Hyp/Hyp) and male (Hyp/Y) mice and their normal female (+/+) and male (+/Y) littermates. Mutant animals had trait values that differed significantly from those of normal animals. However, with the exception of vertebral length and cancellous osteoid thickness, values were not significantly different between the three mutant genotypes. We also examined the effect of gamete-of-origin on histomorphometric parameters in obligate Hyp/+ females derived from male or female transmitting parents. The metrical trait values in both groups of Hyp/+ mice were similar, with the exception of vertebral length and cancellous osteoid volume per bone volume. In summary, we demonstrate that the amount of osteoid per bone volume is similar in the three mutant genotypes and conclude that the extent and magnitude of the mineralization defect is fully dominant and likely not affected by gene dose. The differences in vertebral length in the mutants suggest that rickets and osteomalacia are not the only causes of decreased vertebral growth in Hyp mice and that Phex protein may influence bone growth and mineralization by distinct pathways.

Does hereditary metabolic disease modulate senescence and ageing?

Hereditary metabolic diseases in the context of evolutionary biology elicit interesting questions about ageing and senescence: Will persons successfully treated for inborn errors of metabolism, age and die prematurely because of compromised longevity? Because some unhealthy longevity has its origins in germline and somatic mutational processes, and in an inability to withstand metabolic stress, are there lessons to be learned about senescence from hereditary metabolic disease? Why are ageing, senescence and death necessary for Homo sapiens and how do they happen? These questions form the theme upon which several variations are played during the course of this essay. The theory of the disposable soma recognizes genomic and environmental events, well-seasoned by Chance, as determinants of ageing and senescence. Together, they cause the somatic damage that results in death. Genomics will reveal genes involved in longevity, both healthy and unhealthy. There will be schedules of gene expression behind our life-history traits. As in the field of hereditary metabolic disease, analogous genetic enquiries about ageing can be formulated. For example, how will heterozygotes age? Will association studies in centenarians reveal 'longevity genes'? Will disparate longevity in sib pairs reveal genetic factors? If there are 'ageing' mutations, of what types and with what effects? Will these initiatives lead to healthier longevity? A deeper question yet remains: why has human biology invested so greatly in grandparenthood?

Human genetics: lessons from Quebec populations.

The population of Quebec, Canada (7.3 million) contains approximately 6 million French Canadians; they are the descendants of approximately 8500 permanent French settlers who colonized Nouvelle France between 1608 and 1759. Their well-documented settlements, internal migrations, and natural increase over four centuries in relative isolation (geographic, linguistic, etc.) contain important evidence of social transmission of demographic behavior that contributed to effective family size and population structure. This history is reflected in at least 22 Mendelian diseases, occurring at unusually high prevalence in its subpopulations. Immigration of non-French persons during the past 250 years has given the Quebec population further inhomogeneity, which is apparent in allelic diversity at various loci. The histories of Quebec's subpopulations are, to a great extent, the histories of their alleles. Rare pathogenic alleles with high penetrance and associated haplotypes at 10 loci (CFTR, FAH, HBB, HEXA, LDLR, LPL, PAH, PABP2, PDDR, and SACS) are expressed in probands with cystic fibrosis, tyrosinemia, beta-thalassemia, Tay-Sachs, familial hypercholesterolemia, hyperchylomicronemia, PKU, oculopharyngeal muscular dystrophy, pseudo vitamin D deficiency rickets, and spastic ataxia of Charlevoix-Saguenay, respectively) reveal the interpopulation and intrapopulation genetic diversity of Quebec. Inbreeding does not explain the clustering and prevalence of these genetic diseases; genealogical reconstructions buttressed by molecular evidence point to founder effects and genetic drift in multiple instances. Genealogical estimates of historical meioses and analysis of linkage disequilibrium show that sectors of this young population are suitable for linkage disequilibrium mapping of rare alleles. How the population benefits from what is being learned about its structure and how its uniqueness could facilitate construction of a genomic map of linkage disequilibrium are discussed.

Henry Friesen Award Lecture. Work, the clinician-scientist and human biochemical genetics.

The pursuit of human biochemical genetics has allowed us to understand better how the person with the (genetic) disease differs from the disease the person has and to develop the concept that genetics belongs in all aspects of health care. It is a perspective that comes quite readily to the clinician-scientist, and the restoration of that "species" in the era of functional genomics is strongly recommended. Garrod, the initial founder of human "biochemical genetics" belonged to the clinician-scientist community. Archibald Edward Garrod introduced a paradigm, new for its day, in medicine: biochemistry is dynamic and different from the static nature of organic chemistry. It led him to think about metabolic pathways and to recognize that variation in Mendelian heredity could explain an "inborn error of metabolism." At the time, Garrod had no idea about the nature of a gene. Genes are now well understood; genomes are being described for one organism after another (including Homo sapiens) and it is understood that genomes "speak biochemistry (not phenotype)." Accordingly, in the era of genomics, biochemistry and physiology become the bases of functional genomics, and it is possible to appreciate why "nothing in biology makes sense without evolution" (and nothing in medicine will make sense without biology). Mendelian, biochemical and molecular genetics together have revealed what lies behind the 4 canonical inborn errors described by Garrod (albinisn, alkaptonuria, cystinuria and pentosuria). Both older and newer ideas in genetics, new tools for applying them (and renewed respect for the clinician-scientist) will enhance our understanding of the human biological variation that accounts for variant states of health and overt disease. A so-called monogenic phenotype (phenylketonuria) is used to illustrate, in some detail, that all disease phenotypes are, in one way or another, likely to be complex in nature. What can be known and what ought to be done, with knowledge about human genetics, to benefit individuals, families and communities (society), is both opportunity and challenge.

Homomeric and heteromeric interactions between wild-type and mutant phenylalanine hydroxylase subunits: evaluation of two-hybrid approaches for functional analysis of mutations causing hyperphenylalaninemia.

Phenylketonuria (PKU) is caused by mutations in the phenylalanine hydroxylase gene (PAH), while mutations in genes encoding the two enzymes (dihydropteridine reductase, DHPR, and pterin-4-alpha-carbinolamine dehydratase, PCD) required for recycling of its cofactor, tetrahydrobiopterin (BH(4)), cause other rarer disease forms of hyperphenylalaninemia. We have applied a yeast two-hybrid method, in which protein--protein interactions are measured by four reporter gene constructs, to the analysis of six PKU-associated PAH missense mutations (F39L, K42I, L48S, I65T, A104D, and R157N). By studying homomeric interactions between mutant PAH subunits, we show that this system is capable of detecting quite subtle aberrations in PAH oligomerization caused by missense mutations and that the observed results generally correlate with the severity of the mutation as determined by other expression systems. The mutant PAH subunits are also shown in this system to be able to interact with wild-type PAH subunits, pointing to an explanation for apparent dominant negative effects previously observed in obligate heterozygotes for PKU mutations. Based on our findings, the applications and limitations of two-hybrid approaches in understanding mechanisms by which PAH missense mutations exert their pathogenic effects are discussed. We have also used this technique to demonstrate homomeric interactions between wild-type DHPR subunits and between wild-type PCD subunits. These data provide a basis for functional studies on HPA-associated mutations affecting these enzymes.

Garrod's foresight; our hindsight.

Archibald Edward Garrod introduced a paradigm, new for its day, in medicine: Biochemistry is dynamic and different from the static nature of organic chemistry. It led him to think about metabolic pathways and to recognize that variation in Mendelian heredity could explain an 'inborn error of metabolism'. At the time, Garrod had no idea about the nature of a gene. Genes are now well understood, genomes are being described for one organism after another (including H. sapiens) and it is understood that genomes 'speak biochemistry (not phenotype)'. Accordingly, in the era of genomics, biochemistry and physiology become the bases of functional genomics and it is possible to appreciate why 'nothing in biology makes sense without evolution' (and nothing in medicine will make sense without biology). Mendelian, biochemical and molecular genetics together have revealed what lies behind the four canonical inborn errors described by Garrod (albinism, alkaptonuria, cystinuria and pentosuria). Both older and newer ideas in genetics, new tools for applying them, and renewed respect for the clinician-scientist will enhance our understanding of the human biological variation that accounts for variant states of health and overt disease; an 'unsimple' phenotype (phenylketonuria) is used to illustrate in some detail. What can be known and what ought to be done with knowledge about human genetics to benefit individuals, families and communities (society) is both opportunity and challenge.

The clinical phenotype and outcome of mitochondrial acetoacetyl-CoA thiolase deficiency (beta-ketothiolase or T2 deficiency) in 26 enzymatically proved and mutation-defined patients.

Mitochondrial acetoacetyl-CoA thiolase (T2 enzyme) deficiency (MIM 203750) is an autosomal recessive disorder of isoleucine and ketone-body metabolism. We determined the molecular basis of T2 enzyme deficiency in 26 patients at the levels of skin fibroblast enzyme activity, protein integrity, and DNA nucleotide sequence. Thirty different disease-associated alleles were identified. From these data we predicted that T2 in 6 of the 26 patients would have a mild effect on the enzyme protein and 20 would have a severe effect from their mutant genotypes. The corresponding clinical data were collected (by interviews and questionnaires) for the patients in the two groups. We found that genotype does not predict clinical severity and mutant sibs can have different clinical phenotypes; there were no consistent differences in clinical severity between patients with null-conferring or residual-conferring genotypes for T2 activity; only the absence of or a low urinary excretion of tiglyglycine during ketoacidosis correlated with a mild genotype. In general, T2 deficiency has a favorable outcome and 23 of 26 patients developed normally; one died during the first ketoacidotic episode and two have developmental delay. The median age at onset for the first ketoacidotic episode is 15 months (range 3 days to 48 months). The frequency of attacks falls with age, the last in our series occurring at 10 years of age; 11 patients had only one episode and 3 patients had none. We conclude that clinical consequences of T2 deficiency are avoidable with early diagnosis, appropriate management of ketoacidosis, and modest protein restriction.

Measurement of phenyllactate, phenylacetate, and phenylpyruvate by negative ion chemical ionization-gas chromatography/mass spectrometry in brain of mouse genetic models of phenylketonuria and non-phenylketonuria hyperphenylalaninemia.

Phenylketonuria (PKU) (OMIM 261600) is the first Mendelian disease to have an identified chemical cause of impaired cognitive development. The disease is accompanied by hyperphenylalaninemia (HPA) and elevated levels of phenylalanine metabolites (phenylacetate (PAA), phenyllactate (PLA), and phenylpyruvate (PPA)) in body fluids. Here we describe a method to determine the concentrations of PAA, PPA, and PLA in the brain of normal and mutant orthologous mice, the latter being models of human PKU and non-PKU HPA. Stable isotope dilution techniques are employed with the use of [(2)H(5)]-phenylacetic acid and [2,3, 3-(2)H(3)]-3-phenyllactic acid as internal standards. Negative ion chemical ionization (NICI)-GC/MS analyses are performed on the pentafluorobenzyl ester derivatives formed in situ in brain homogenates. Unstable PPA in the homogenate is reduced by NaB(2)H(4) to stable PLA, which is labeled with a single deuterium and discriminated from endogenous PLA in the mass spectrometer on that basis. The method demonstrates that these metabolites are easily measured in normal mouse brain and are elevated moderately in HPA mice and greatly in PKU mice. However, their concentrations are not sufficient in PKU to be "toxic"; phenylalanine itself remains the chemical candidate causing impaired cognitive development.

A heteroallelic mutant mouse model: A new orthologue for human hyperphenylalaninemia.

Hyperphenylalaninemias (HPA) are Mendelian disorders resulting from deficiencies in the conversion of phenylalanine to tyrosine. The vast majority are explained by a primary deficiency of phenylalanine hydroxylase (PAH) activity. The majority of untreated patients experience irreversible impairment of cognitive development. Although it is one of the best known hereditary metabolic disorders, mechanisms underlying the pathophysiology of the disease are still not fully understood; to this end, the availability of an orthologous animal model is relevant. Various mutant hyperphenylalaninemic mouse models with an HPA phenotype, generated by N-ethyl-N'-nitrosourea (ENU) mutagenesis at the Pah locus, have become available. Here we report a new hybrid strain, ENU1/2, with primary enzyme deficiency, produced by cross breeding. The ENU1, ENU1/2, and ENU2 strains display mild, moderate, and severe phenotypes, respectively, relative to the control strain (BTBR/Pas). The Pah enzyme activities of the various models correlate inversely with the corresponding phenylalanine levels in plasma and brain and the delay in plasma clearance response following a phenylalanine challenge. The maternal HPA effect on the fetus correlates directly with the degree of hyperphenylalaninemia, but only the ENU2 strain has impaired learning.

Characterization of phenylketonuria missense substitutions, distant from the phenylalanine hydroxylase active site, illustrates a paradigm for mechanism and potential modulation of phenotype.

Missense mutations account for 48% of all reported human disease-causing alleles. Since few are predicted to ablate directly an enzyme's catalytic site or other functionally important amino acid residues, how do most missense mutations cause loss of function and lead to disease? The classic monogenic phenotype hyperphenylalaninemia (HPA), manifesting notably as phenylketonuria (PKU), where missense mutations in the PAH gene compose 60% of the alleles impairing phenylalanine hydroxylase (PAH) function, allows us to examine this question. Here we characterize four PKU-associated PAH mutations (F39L, K42I, L48S, I65T), each changing an amino acid distant from the enzyme active site. Using three complementary in vitro protein expression systems, and 3D-structural localization, we demonstrate a common mechanism. PAH protein folding is affected, causing altered oligomerization and accelerated proteolytic degradation, leading to reduced cellular levels of this cytosolic protein. Enzyme specific activity and kinetic properties are not adversely affected, implying that the only way these mutations reduce enzyme activity within cells in vivo is by producing structural changes which provoke the cell to destroy the aberrant protein. The F39L, L48S, and I65T PAH mutations were selected because each is associated with a spectrum of in vivo HPA among patients. Our in vitro data suggest that interindividual differences in cellular handling of the mutant, but active, PAH proteins will contribute to the observed variability of phenotypic severity. PKU thus supports a newly emerging paradigm both for mechanism whereby missense mutations cause genetic disease and for potential modulation of a disease phenotype.