Article Commentary: The contributions of Louis I Woolf to the treatment,early diagnosis and understanding of phenylketonuria

Abstract

Early diagnosis of phenylketonuria (PKU) became a goal worth pursuing following demonstration of the efficacy of the dietary treatment conceived by Louis I Woolf. This paper narrates the history of this treatment, describes Woolf's role in the establishment of neonatal PKU screening and surveys his other contributions to our understanding of this condition. If Woolf, Centerwall, Baird and Berry had waited until all the scientific evidence about PKU that is now at our disposal had been brought to light, there would still be no neonatal screening programmes.

Introduction

The first inborn error of metabolism to be associated with mental disorder, phenylketonuria (PKU), was first described, as ‘imbecillitas phenylpyruvica’, in 1934.^{1, 2} Ivar Asbjørn Føiling reported the identification of phenylpyruvic acid in the urine of 10 mentally retarded children of similar phenotype and clinical history, six of whom formed pairs of siblings. Over the next few years several authors confirmed that PKU had an inheritance pattern compatible with a recessive Mendelian character.^3–5 Føiling suggested that the defect might relate to phenylalanine metabolism, but there was uncertainty regarding the normal metabolism of this amino acid. In 1944 Mary and Frederick Bernheim showed that in the normal organism phenylalanine is predominantly parahydroxylated to tyrosine. ⁶ Three years later Jervis established that the metabolic flaw in PKU is inability to perform this hydroxylation, ⁷ and provided evidence that an excess of phenylalanine in serum, not a deficit of tyrosine, was responsible for PKU. ⁷

We describe here the contributions of Louis I Woolf to the treatment of PKU, to the development of neonatal PKU screening programmes, and to the elucidation of the epidemiology and genetics of the disorder.

The Dietary Treatment of PKU

The idea that PKU should be treated with a diet low in phenylalanine was already being put forward in the 1930s, before the biochemistry of PKU had been clarified. Lionel Penrose tested a diet composed solely of fruit, sugar, olive oil and vitamins; ⁸ and in 1939 Block and Jervis were planning a less drastic low-phenylalanine diet. ⁹ The abundance of phenylalanine in protein, and the dependence of normal development on the availability of amino acids, meant that normal protein sources had to be eliminated from the diet and replaced by an artificial amino acid source containing a strictly limited amount of phenylalanine. Two strategies presented themselves for the preparation of this amino acid source: formulation of a mixture of pure amino acids in appropriate proportions, and elimination of excess phenylalanine from natural protein. The first strategy was prohibitively expensive as a clinical treatment. How the second might be pursued was first realized around 1949, by Louis I Woolf. ¹⁰

A method for removing phenylalanine and other aromatic amino acids from protein hydrolysates – passage through activated charcoal – had been known since 1943. ¹¹ However, Woolf was the first person to see its relevance to the treatment of PKU, possibly because, before joining the Institute of Child Health at Great Ormond Street Hospital, he had worked on the mass production of casein hydrolysate, with a view to its possible administration as predigested food to severely starved persons in post-war Europe. ¹² In 1949 he began to suggest that supplementation of carbon-treated casein hydrolysate with appropriate quantities of the missing amino acids (including sufficient phenylalanine to avoid phenylalanine deficiency) would produce the desired low-phenylalanine food. ¹⁰

Woolf was – and is – a chemist, not a physician, and the physicians at Great Ormond Street Hospital were unwilling to try his proposed treatment on any phenylketonurie patients. ¹³ In a 1951 paper co-authored with DG Vulliamy he nevertheless revived in print the idea of low-phenylalanine dietary treatment of PKU:

‘If the amount of phenylalanine and its breakdown products could be reduced, normal cerebral function might result. Whether this would lead to recovery or not, would probably depend upon the duration of exposure of the brain to the injurious concentration of this substance.

Two possible methods of achieving this reduction seem worthy of investigation: restricting the phenylalanine intake in the diet to the basic minimum early in life; and increasing the rate of excretion of phenylalanine by administering a substance which would competitively reduce tubular reabsorption’. ¹⁴

Then he was approached by Horst Bickel and Evelyn Hickmans of the Birmingham University Children's Hospital.

Bickel had gained experience of metabolic diseases in Zürich, where under Guido Fanconi's direction he had discovered Fanconi–Bickel syndrome.^{15, 16} In 1950 he moved to Birmingham to further his research into congenital metabolic disorders, and it was at his suggestion that Følling's ferric chloride test for PKU began to be applied at the Birmingham Children's Hospital to children with mental retardation. ¹³ On 13 March 1951, the third patient to be tested, a 17-month-old girl of Irish descent called Sheila, gave a positive result that was confirmed by quantitation of phenylpyruvic acid in urine and of phenylalanine in urine and plasma. ¹⁷

Bickel did not originally intend to develop a treatment for PKU. However, pressed by the child's mother, he eventually accepted the challenge. He initially tried to lower plasma phenylalanine concentration by administration of glutamic acid, which had been the treatment attempted, unsuccessfully, in the study described in Woolf's 1951 paper. ¹⁴ It was when this approach failed that he and Evelyn Hickmans, the head of the Children's Hospital biochemistry laboratory, asked Woolf how a low-phenylalanine protein source might be prepared;^{13, 18} Woolf obliged. Although Gerrard ¹³ and Bickel and Grueter ¹⁹ give one to understand that Woolf attributed his knowledge of the technique to ‘a recent report’ or ‘suggestion’ by Block and Boiling, in their 1955 paper ²⁰ Woolf et al. cited Schramm and Primosigh's paper of 1943. ¹¹ Block and Boiling mentioned Schramm and Primosigh's paper in the second edition of their book The Amino Acid Composition of Proteins and Foods (1951), ²¹ but not in the first edition (1945). Woolf probably mentioned Block and Bolling's book simply as the most convenient source for the Bickel group to consult.

Bickel was not convinced that the proposed diet had much chance of success, ²² but he nevertheless set about the laborious task of preparing sufficient quantities of low-phenylalanine hydrolysate.^{18, 23} Substitution of this preparation for protein in Sheila's diet began in December 1951, when she was 26 months old; in Woolf's words, ¹² ‘it must have taken great courage for the doctor to give her a completely unproven diet’. Within a few months, however, ‘she learnt to crawl, to stand, and to climb on chairs; her eyes became brighter; her hair grew darker; and she no longer banged her head or cried continuously’; ¹⁷ ‘roughness and eczema of the skin disappeared, as did the mouse-like smell’, ²⁴ and she ‘began to develop good interaction with her mother and the nurses’. ¹⁸ Transitory surreptitious replacement of phenylalanine in the diet gave rise to an almost immediate reversion to her former state that was in turn reversed by renewed limitation of phenylalanine.

It is probably because Woolf's 1951 paper, in which the idea of a low-phenylalanine diet was relaunched, contained neither results on the application of such a diet nor details of how to prepare one, that this paper is not cited by Bickel and co-workers in their early reports and later descriptions of their treatment of Sheila.^13,17,18,24 Nor did the Birmingham team's own preliminary report in The Lancet ¹⁷ describe the diet in detail, although it did of course describe the low-phenylalanine hydrolysate. Full details of the development of the complete diet were not published until 1954 ²⁴ (Woolf's team at Great Ormond Street, who began to treat children in 1953, gave full details of the diet in 1955 ²⁰ and of subsequent improvements in 1958, ²³ and in 1962 Woolf reviewed the nutritional requirements of phenylketonurics in a paper to the Nutrition Society ²⁵ ). Bickel et al.^{17, 24} do thank Woolf for drawing their attention to how phenylalanine could be removed from casein hydrolysate, but the early reports do so in terms that might suggest mere technical assistance, without explicit acknowledgement that Woolf had himself been attempting to have such a diet tested. This led to some confusion among outsiders. Pena, ²⁶ for example, in his 1958 review paper in Revista Española de Pediatría, wrote that ‘the application of this diet occurred to Bickel’ without any mention of Woolf. For his part, in his 1955 paper ²⁰ Woolf describes his relationship with the Birmingham group in terms that are appropriate to the context of his paper, but might suggest that it was Woolf who initiated the relationship:

‘It seemed likely that a suitable amino-acid mixture could be obtained at a reasonable cost by passing an acid hydrolysate of casein through charcoal and adding tyrosine and tryptophan. This was suggested by one of us (L.I.W.) to Dr. H. Bickel, who at that time had a suitable phenylketonuric patient under his care. Bickel carried out the preparation of the suggested diet, fed his patient on it, and reported a very marked improvement in the child's mental condition (Bickel, Gerrard, and Hickmans, 1953).’

Bickel's co-author Gerrard, recounting these events in 1994 without the stylistic constraints of an academic paper, ¹³ was less cursory:

‘While we in Birmingham were wondering how to prepare a low phenylalanine diet, Louis Woolf, a biochemist at The Hospital for Sick Children in Great Ormond Street, London, […] was trying to convince his colleagues to at least put one child with PKU on a diet low in phenylalanine, for he thought that it was the high level of phenylalanine in the blood which was responsible for the mental retardation. He knew that it was possible to prepare a diet low in phenylalanine, for he had seen a recent report by Block and Boiling […].

Our problem, the preparation of a formula low in phenylalanine, was discussed by Bickel and Hickmans with Louis Woolf. The latter very kindly shared his information with us […].’

Writing at around the same time as Gerrard,^{27, 28} Charles Scriver also grants Woolf a key role in the demonstration of the feasibility of a low-phenylalanine diet, though he dates this idea to 1951, the year of Woolf and Vulliamy's paper, ¹⁴ rather than two years earlier:

‘Louis Woolf had translated an idea he first had in 1951 […] into practice, and with Bickel and colleagues in the United Kingdom and Armstrong and Tyler in the United States, the chemical phenotype of PKU was shown to respond to dietary restriction of phenylalanine.’

(Armstrong and Tyler ²⁹ used a mixture of amino acids, though they were aware of the possibility of producing a low-phenylalanine protein hydrolysate.)

Similarly, Centerwall and Centerwall ³⁰ are careful to mention Woolf in describing Bickel, ‘who, in consultation with British scientist Dr. L.I. Woolf, was the first person to make a low-phenylalanine food and treat a child with PKU’.

The value of low-phenylalanine dietary treatment did not meet with immediate universal recognition. Penrose, for example, in spite of having himself attempted a dietary treatment in the 1930s, ⁸ had become convinced that in PKU ‘mental retardation was only associated with the metabolic derangement, and was not caused by it’. ¹³ Bickel and coworkers appear to have had some difficulty in publishing their experience with Sheila due to their results being regarded as too speculative, ¹³ and indeed, a few years later a two-year-old treated by Woolf's team at Great Ormond Street completely failed to respond even to the limited extent expected for such late-started therapy.^{12, 23} On the other hand, many later criticisms would have been more appropriately directed, not at the diet, but at those who applied it without due caution. In his 1955 ²⁰ and 1958 ²³ papers, Woolf not only described in considerable detail how to prepare the diet, but also stressed both the need for careful monitoring and the possible need for the diet to be adhered to throughout life. He later recalled that he repeatedly ‘warned that some of the amino acid would have to be given […] in the form of normal food, otherwise the child wouldn't be able to grow. In the first few years of this treatment for PKU this advice either wasn't known or wasn't followed by some other doctors and the results were disastrous’. ¹² Such incidents encouraged fears that low-phenylalanine diets would inevitably lead to malnutrition (see, for example, the round-table Discussion at the end of Ref. 31). As late as 1967 Woolf ³² was obliged to rebut the conclusions of a paper ³³ that, after not unreasonably criticizing a rather weak study, ³⁴ suggested that low-phenylalanine dietary treatment required validation by a clinical trial; and in the same year a phenylketonuric newborn was treated with tyrosine supplement alone (a treatment continued for 4 years in spite of evident mental decline!) by physicians who apparently told his parents that this might be an effective alternative to a low-phenylalanine diet. ³⁵

The main criticism of the institution of dietary treatment for all children with high serum levels of phenylalanine was that it was not certain whether all such children would necessarily become retarded if left untreated. No studies had been carried out to determine directly how many positive-testing individuals of normal intelligence there might be in the general population. Indeed, by the mid-1960s a number of children with normal intelligence and high serum phenylalanine had been identified, and the discrepancy between the prevalence of high serum phenylalanine estimated from extensive testing in asylums and the prevalence estimated from the results of mass screening of newborns suggested that as many as 50% of persons with high serum phenylalanine might be of normal intelligence. ³¹ Even so, however, as Woolf pointed out, ³¹ treatment of the phenylketonuric at birth was ‘the only course open to us since … [he or she] stands at least a 50 percent chance of being mentally retarded’ if untreated. Furthermore, it soon became clear that at least some of the discrepancy was due to children with serum phenylalanine levels that were high at birth and later declined, but who would likewise become mentally retarded if not treated during the high-phenylalanine period. ³⁶ Today, in the light of another 40 years’ experience and the information made accessible by vastly more powerful analytical methods, it is known that, leaving aside cases of transient hyperphenylalaninaemia of newborns or co-factor deficiency, serum phenylalanine levels range almost continuously from normal to levels typical of classical PKU depending on the particular alleles inherited, and that the threat to intelligence varies more or less accordingly. Dietary restriction of phenylalanine is tailored to the needs of each patient, is lifelong, and is accompanied by continual monitoring and adjustment – exactly as Woolf suggested might be necessary 50 years ago. ²³

Screening

Though devising a low-phenylalanine treatment was Woolf's most critical contribution to ‘the PKU story’, it required far less work on his part than his subsequent efforts to promote, establish and improve neonatal screening programmes in the United Kingdom.

The natural history of PKU clearly indicated that, in the absence of treatment, developmental retardation began within the first few weeks of life; and early experience of low-phenylalanine dietary treatment suggested that the earlier treatment began, the more likely it was to be effective. ³⁷ Nevertheless, neither Bickel nor Woolf initially proposed the mass screening of newborns. Bickel ²⁴ suggested that phenylketonuric babies ‘could be detected in families where a sibling is known to have the disease’, and Woolf ²⁰ emphasized that ‘the urine of every baby or young child in whom there is the slightest suspicion of mental retardation must be tested for phenylpyruvic acid’. Even in this 1955 paper, however, Woolf et al ²⁰ presented a rough analysis suggesting that society would be economically better off if all phenylketonurics were detected and began to be treated at a sufficiently early age:

‘It has been estimated that the incidence of phenylketonuric amentia is about 4 for every 100,000 of the population […] and without treatment most of these will eventually reach an institution, where the cost each week is said to average at least £5. The special dietary constituents […] cost 10s 9d a day at present prices. ³⁸ [….] The grades of many of these patients might be raised so that the ineducable become educationally subnormal and the E.S.N, group might get accepted at ordinary schools, thus saving the cost of institutional care. Later on, society may gain productive members.’

By July 1957, Woolf was not only satisfied that the evidence in favour of the benefits and long-term feasibility of treatment was sufficient to make it possible to recommend screening, but was also taking steps to initiate screening programmes based on Følling's ferric chloride test: ²³

‘We advocate that every child's urine should be tested at the age of 21 days. In the light of further experience it may be possible to shorten this period. […]

The local health authorities should […] see that health visitors collect and test the urine of every baby. A coloured card with instructions as to testing is being prepared. Already one large local authority has asked one of us to talk on the subject to health visitors and midwives in the area.’

The possibility that screening at birth might result in unnecessary treatment of a large number of healthy individuals was implicitly argued against by appeal to Penrose's 1951 data on the distribution of IQ among institutionalized phenylketonurics, which suggested that only about 0.1% of phenylketonurics could be in the normal IQ range. ³⁹

A pilot scheme on the lines suggested by Woolf, in which mothers obtained bulk urine samples from their three-week-old children, added a little preservative, and took it to the infant welfare clinic for analysis, was set up in Cardiff in 1958 and reported on a year later. ⁴⁰ Although it was now recognized that some three-week-old phenylketonurics would not yet have begun to excrete phenylpyruvic acid, an age of three weeks was considered a reasonable compromise between screening procedure sensitivity and the need to institute treatment as early as possible in order to prevent permanent brain damage; a second test at age two months was recommended, but was not routine. This pilot scheme detected one case of PKU (confirmed by paper-chromatographic evaluation of phenylalanine in serum), but only 26% of the target population were actually screened, basically due to the difficulty of obtaining a sample of bulk urine from a young baby. Because of this, in 1959 the Cardiff programme switched to using Phenistix, ³¹ paper strips that turned blue-green within 60 seconds of being wetted with urine containing phenylpyruvic acid. Phenistix had been developed by Ames and Co. following initial work by Baird. ⁴¹ When used to test wet napkins (a procedure that favoured high coverage of the target population, since it allowed health visitors to apply the test in the babies’ homes ⁴² ), Phenistix were both more convenient and more reliable than ferric chloride solution; also, they did not give false-positive results when exposed to p-hydroxyphenylpyruvic acid, the compound responsible for all seven false-positives afforded by the ferric chloride test in the first year of the Cardiff programme.

By the mid-1960s, mass screening for PKU with Phenistix was operative or being set up in almost all the local authorities of England and Wales. However, feedback revealed that the Phenistix test had an unacceptably low sensitivity, some 23% of all phenylketonurics born in 1964 having had false-negative tests. ⁴³ Fortunately, a number of more sensitive tests had by then been developed, and at the time of the 1966 Washington Conference on Phenylketonuria, Woolf was taking part in a large study comparing the all-round PKU-screening performance of Phenistix with the Guthrie inhibition test of blood, the Guthrie test of urine, Woolf's own paper-chromatographic test for o-hydroxyphenylacetic acid (the presence of which in urine was by then known to be as characteristic of PKU as phenylpyruvic acid ⁴⁴ ), and, as reference method, the determination of phenylalanine in serum. ³¹ All these methods except Phenistix required the reception of samples in central laboratories, to facilitate which they adopted Helen K Berry's^{45, 46} concept of collecting samples in absorbent paper (which in the case of urine solves the problem of how to obtain samples from young babies). It is indicative of the success and social impact of the pilot programmes that it was now considered feasible to establish such laboratories all around the UK, whereas at the beginning of the decade they had been insufficient. ⁴⁷

One of the advantages of screening samples in a central laboratory is that they can be screened not only for PKU but also for a wide range of other congenital hereditary or non-hereditary conditions. Berry was herself the first to apply this principle of ‘one sample, one analysis, multiple diagnoses’, incorporating spot tests for protein and for galactose and other sugars into the Cincinnati screening programme, and separating sugars from positive samples by means of what would now be called a ‘multiplexing’ method based on paper chromatography. Woolf was a firm adherent of this strategy, insisting, for example, on the desirability of screening for multiple disorders in his correspondence with Federico Mayor-Zaragoza when the latter was planning to introduce neonatal screening to Spain. ⁴⁸ In 1966 he was involved in a large pilot study screening simultaneously for PKU, histidinaemia, tyrosinosis, homocystinuria, cystinuria, galactosaemia, glucosuria and renal tubular dysfunction by paper chromatography and spot tests of urine samples. ³¹ Including a test for glucosuria, as Berry ⁴⁶ and Woolf ³¹ did, amounts to screening for neonatal diabetes. His paper-chromatographic method for screening for PKU by detection of o-hydroxyphenylacetic acid in urine was used at the Oxford laboratory until 1973. ⁴

Individual and Population Genetics of PKU

Woolf realized from an early date that the phenylalanine hydroxylase gene might have alleles producing semifunctional enzymes giving rise to syndromes intermediate between normality and classical PKU, i.e. what is now known as hyperphenylalaninaemia. In 1957 he suggested this in relation to a six-year-old boy with high serum phenylalanine, low phenylpyruvic acid concentration in urine, and normal intelligence. ⁴⁹ A year later, in discussing Penrose's calculation that only about 0.1% of phenylketonurics could be in the normal IQ range (an estimate based on the distribution of IQ among institutionalized phenylketonurics), ³⁹ he probably had a semifunctional enzyme in mind when he pointed out that this estimate would not be valid if the distribution of IQ among all individuals excreting phenylpyruvic acid were bimodal. However, it was not until 1966 that he was able to obtain good evidence of the existence of such an enzyme by characterizing the kinetics of phenylalanine metabolism following intravenous phenylalanine injections (a procedure initially developed as a more discriminating test of heterozygosity than the use of oral phenylalanine loading), which with some subjects suggested the presence of a phenylalanine hydroxylase that was abnormally susceptible to inhibition by phenylalanine.^50–52 He also found evidence that a fourth variety of phenylalanine hydroxylase might be responsible for atypical cases of PKU that he had seen. ⁵² These atypical cases included those of two sisters seen in 1961. ⁵³ The elder had an IQ of 102, no history of mental or neurological problems, and five children of normal intelligence; the younger had an IQ of 83, a history of epileptic fits and declining intelligence between ages nine and 23 years, and one son of normal intelligence. The younger was described as a case of ‘occult phenylketonuria’ in that, although paper chromatography of her urine revealed phenylpyruvic acid levels higher than normal, she tested negative in all the then-standard screening tests, i.e. the ferric chloride, Phenistix and 2,4-dinitrophenylhydrazine ⁵⁴ urine tests. Since their parents gave phenylalanine loading test results consistent with their being heterozygotes, and since phenylalanine loading of the sisters themselves failed to induce an increase in their serum tyrosine levels, Woolf was inclined at the time to rule out the possibility that the sisters might have an allele for a semifunctional enzyme. However, his most serious and most surprising misjudgement with respect to these sisters concerned what was later to be called ‘maternal PKU’: since their children were of normal intelligence in spite of having been exposed to high phenylalanine concentrations in utero, he concluded that there was ‘no biochemical reason why any phenylketonuric woman should not have normal children’. This was not the first time Woolf had supported hypotheses that subsequently proved to be mistaken, but his mistakes had never been unreasonable in the light of the available evidence. Here, however, it would surely have been more reasonable to limit his conclusions to the children of mothers with atypical PKU of the kind found in these two patients.

During the 1960s Woolf began to study the population genetics of PKU. In a 1967 paper bearing on the ethics of neonatal screening, he published calculations showing that the increased fertility of treated phenylketonurics would not swamp the population with phenylketonurics needing treatment, at least in the next millennium. ⁵⁵ However, his main line of research in population genetics stemmed from Jervis’ conclusion that the defective allele must be especially common in Ireland.^{4, 56} In support of this conclusion, in 1961 Woolf reported that a disproportionately large number of the parents and grandparents of phenylketonurics in SE England came from Ireland or western Scotland, a finding that could not be explained by consanguineous marriages, and therefore had to be due to a high frequency of the defective allele in the population. ⁵⁷ In the chapter ⁵⁸ he contributed to the book ⁵⁹ on PKU that in 1971 he co-edited with Bickel and FP Hudson, he pointed out that, given the virtually zero reproductive rate of homozygotes, maintenance of such a high allele frequency implied that the heterozygote must enjoy some reproductive advantage over individuals who are homozygous for the wild-type allele. In subsequent papers he filled in gaps in this argument, ruling out mutation, gamete selection and various kinds of genetic drift as possible causes of the high PKU allele frequency; and he presented data gathered by his Irish and Scottish collaborators that showed heterozygotic mothers to have lower miscarriage rates than wild-type homozygotes in the first six months of pregnancy. ^60–62 Since the only known effect of the defective allele was to cause its heterozygotic bearer to have slightly higher serum phenylalanine levels than the wild-type homozygote, and since in this respect the influence of maternal metabolism would far outweigh that of fetal metabolism, it seemed that it must be higher maternal serum phenylalanine levels that afforded some advantage to the developing fetus; and since there were no differences between PKU heterozygotes and controls as regards the weight of their newborn babies, and these data had been collected at a time in which there was no evolutionary pressure due to famine (which in any case would have been expected to have similar effects in geographical areas in which the prevalence of PKU was in fact not particularly high), the advantage to the fetus was unlikely to be nutritional. ⁶⁰ Furthermore, analysis of the quantitative difference between the reproductive advantage observed in the study data and the considerably smaller advantage required for genetic equilibrium suggested that the environmental hazard to the fetus, against which the PKU allele afforded protection, must either have arisen in the relatively recent past or have acted only intermittently. ⁶² In 1986, in a letter to the American Journal of Human Genetics, ⁶³ he suggested that the hazard in question might be ochratoxin A, a mycotoxin produced by moulds that infest certain foods, and which, at doses that are harmless to the mother, can lead to fetal death by competing with phenylalanine for phenylalanyl-tRNA synthetase, thus bringing protein synthesis to a halt. Ochratoxin A might be favoured by the mild, wet climate of Ireland and western Scotland; would be not unlikely to be ingested in times of famine; and its competition for phenylalanyl-tRNA synthetase would be less deleterious for heterozygotic mothers, with their higher serum phenylalanine levels, than for mothers who were homozygotic for the normal allele.

PKU-Related Miscellanea

Much of Woolf's work centred on the development of methods for the detection or determination of biochemical analytes. Of immediate clinical relevance to PKU at the time of their publication were his relatively interference-free methods for the determination of phenylpyruvic acid in urine (1952); ⁶⁴ his improvement of La Du and Michael's method for monitoring phenylalanine in serum (1962); ⁶⁵ and his paper-chromatographic method for detection of o-hydroxyphenylacetic acid (1967). ³¹ His improvements to the phenylalanine hydroxylase assay^{66, 67} contributed to basic research on PKU, and on phenylalanine metabolism in general. One of his most notable achievements was the first purification of human phenylalanine hydroxylase, ⁶⁸ a feat that allowed subsequent investigation of its metal content and other properties. ⁶⁹

Woolf also played a role in several studies that may have saved much research effort from being wasted investigating false leads. One showed that defective myelination in the brain of the phenylketonuric was not a phenomenon mediating between high phenylalanine levels and mental deficiency, but a result of prior neuron death; ⁷⁰ others that mice that were homozygous for the dilute lethal (d¹) allele were not, as had sometimes been assumed, a model of PKU.^71–73

Also of interest is the conclusion of what appears to have been Woolf's very first incursion into the field of phenylalanine and tyrosine metabolism, that excretion of phenolic acids by newborns following administration of 1-tyrosine and 1-phenylalanine is reduced by concurrent daily administration of ascorbic acid. ⁷⁴ It is partly due to this result that when high urine phenolic acids levels are detected in neonatal screening laboratories that like ours test for these analytes, the paediatrician is commonly urged to prescribe massive doses of vitamin C, investigate protein intake, and send the laboratory further paper-borne samples of blood and urine a week after institution of treatment. ⁷⁵

Finally, it should not be forgotten that in the course of his career Woolf wrote numerous review papers on PKU and related entities, many of them aimed more or less directly at promoting awareness of the benefits of screening and surveying current advances in that field. Among others, in 1962 he published a brief review of recent advances in PKU and maple syrup urine disease, and contributed a chapter on the detection and quantification of PKU-related biochemical abnormalities to the book on PKU edited by FL Lyman; ⁷⁶ for Vol. 6 of Advances in Clinical Chemistry, which appeared in 1963, he wrote a chapter on errors of phenylalanine and tyrosine metabolism in which, besides PKU, he considered tyrosinosis, tyrosyluria, alkaptonuria and albinism; ⁷¹ in 1968 he wrote three more brief reviews;^36,77,78 and in 1979 he wrote an editorial on PKU and its variants for The Western Journal of Medicine. ⁷⁹

Conclusion

Woolf played a major part in the establishment of neonatal screening programmes, and made significant contributions to our understanding of PKU. But his most crucial contribution in the history of PKU – and thus in the history of neonatal screening in general, which stemmed from screening for PKU – has generally been overlooked because it was others who were in a position to apply his ideas. For it to be worthwhile to screen for PKU, it was first necessary to develop a feasible treatment; and for the development of a feasible treatment it was necessary for awareness of the need for a low-phenylalanine diet to meet knowledge of how such a diet might be prepared at reasonable cost. That nexus was Louis I Woolf.