University of Nebraska - Lincoln DigitalCommons@University of Nebraska - Lincoln

Public Health Resources Public Health Resources

1997

CYSTATHIONINE β-SYNTHASE DEFICIENCY: METABOLIC ASPECTS

S. Harvey Mudd National Institute of Mental Health, Bethesda, MD,

Follow this and additional works at: https://digitalcommons.unl.edu/publichealthresources

Part of the Public Health Commons

Mudd, S. Harvey, "CYSTATHIONINE β-SYNTHASE DEFICIENCY: METABOLIC ASPECTS" (1997). Public Health Resources. 206. https://digitalcommons.unl.edu/publichealthresources/206

This Article is brought to you for free and open access by the Public Health Resources at DigitalCommons@University of Nebraska - Lincoln. It has been accepted for inclusion in Public Health Resources by an authorized administrator of DigitalCommons@University of Nebraska - Lincoln. Published in:

HOMOCYSTEINE METABOLISM: FROM BASIC SCIENCE TO CLINICAL MEDICINE

Editors Ian Graham, MD THE ADELAIDE HOSPITAL TRINITY COLLEGE DUBLIN IRELAND Helga Refsum, MD UNIVERSITY OF BERGEN DEPARTMENT OF CLINICAL BIOLOGY BERGEN NORWAY Irwin H. Rosenberg, MD JEAN MAYER USDA HUMAN NUTRITION RESEARCH CENTER ON AGING AT TUFTS UNIVERSITY BOSTON, MA USA Per Magne Ueland, MD UNIVERSITY OF BERGEN DEPARTMENT OF CLINICAL BIOLOGY BERGEN NORWAY

Scientific Editor: Jill M. Shuman, MS, RD, ELS TUFTS UNIVERSITY SCHOOL OF NUTRITION SCIENCE AND POLICY MEDFORD,MA USA

Kluwer Academic Publishers

BOSTON DORDRECHT LONDON 1997 S. Harvey Mudd National Institute of Mental Health Bethesda, MD, USA 11. CYSTATHIONINE P-SYNTHASE DEFICIENCY: METABOLIC ASPECTS

s. Harvey Mudd

Introduction nonresponders was 56; for responders, 78. Among untreated patients, optic lens dislocation had oc­ Cystathionine ~-synthase (CBS) deficiency was first curred in 50% of nonresponders by age six years, a demonstrated in 1964 in an eight-year-old mentally clinically apparent thromboembolic event had oc­ retarded girl with bilaterally dislocated optic lenses curred in 25% of patients by age 15 years, and spinal who excreted abnormally elevated amounts of homo­ osteoporosis was detected in 50% by age 12. By con­ cystine in her urine {l]. Patients with similar meta­ trast, the corresponding ages for untreated B - bolic abnormalities and clinical findings had first 6 responders were 10,20, and 20 years. been discovered 2 years earlier by Carson and her colleagues during a survey of mentally backward children in Northern Ireland [2}. CBS deficiency Metabolic Aspects has proven to be the most frequently encountered of The primary abnormality in CBS-deficient patients is the human genetic diseases causing homoeystinuria a failure to covert homocysteine to cystathionine, a and severe hyperhomocyst(e)inemia. Worldwide, it is process that normally must occur chiefly in the liver. detected with a frequency of about 1: 344,000 by Mansoor and colleagues [5} have shown that in nor­ screening programs of the newborn, but this is un­ mal plasma, only 2% of the homocysteine-derived doubtedly an underestimate because some indi­ moieties exist as homocysteine itself; 16% occur as viduals are being missed [3]. This chapter will briefly free disulfides (homocystine, homocysteine-cysteine focus on the major clinical manifestations and meta­ mixed disulfide, other mixed disulfides); and 82% are bolic aspects of CBS deficiency. bound to protein as mixed disulfides of homocysteine and protein cysteine [5}. Comparable measurements Major Clinical Features by modern methods have not been reported for in­ Clinically, the most prominent features of CBS defi­ tracellular homocysteine-derived moieties in CBS­ ciency are mental retardation, dislocation of the optic deficient patients. Clearly there is some export of lenses, a tendency toward early thromboembolic the abnormally accumulated homocysteine so that the events, and osteoporosis and other bony abnormali­ total plasma homocyst(e)ine (tHcy) rises dramatically. ties. An international survey published in 1985 [4} The greatest relative rise occurs in homocysteine, with presented data on 629 CBS-deficient patients. Among lesser rises in free disulfides and the protein-bound those so classified, some 44% were judged to be B6- mixed disulfide (up to 400-fold, 90-fold, and 20-fold responsive; that is, when given high doses of pyridox­ above normal, respectively) [6}. Plasma methionine ine they underwent prompt and marked decreases in also rises in most CBS-deficient patients, presumably plasma homocyst(e)ine and methionine concentra­ due to enhanced methylation of abnormally elevated tions. An equal proportion were judged nonrespon­ homocysteine. The source(s) of the methyl group for sive, with little or no decreases in the abnormally such methylation is (are) uncertain. The availability elevated concentrations of these amino acids follow­ of betaine must be limiting under normal dietary ing pyridoxine administration. Although this crude conditions because administration of betaine usually classification is undoubtedly an oversimplification, results in further decreases in tHey. N 5 -Methyltetra­ the results clearly showed that in the absence of spe­ hydrofolate would be expected to provide an cific treatment, as a group the B6-nonresponsive pa­ alternative supply of methyls (see below). tients were more severely affected clinically than were Additional compounds that would be expected to the B6-responsive patients. The median IQ for accumulate abnormally in cells of CBS-deficient pa- 78 I. BIOCHEMISTRY AND GENETIC STUDIES

tients include S-adenosylhomocysteine (AdoHcy) (if The metabolic consequences of CBS deficiency adenosine is available) and S-adenosylmethionine might be expected to include abnormally low rates of (AdoMet). The former is expected to accumulate be­ synthesis of compounds distal to the enzyme block, cause the equilibrium of AdoHcy hydrolase favors including cystathionine and cysteine. Because a rela­ AdoHcy formation, and the latter because of the el­ tively small residual activity of CBS may support a evated methionine concentration in the presence of substantial flux of homocysteine into cystathionine intact methionine adenosyltransferase that (in the (see below), the extent of the abnormalities in liver) is normally operating far below saturation with cystathionine and cysteine in a given patient will respect to that substrate. Neither of these compounds probably depend upon how much residual CBS activ­ has been measured in liver or other tissue of CBS­ ity is present. Cystathionine is present in normal deficient patients by use of methods that are now human brain at high concentrations (12), but was known to be necessary to avoid rapid changes after present at most only in trace amounts in brains tissue is removed from the body. Indirect evidence obtained at postmortem examination of three homo­ for an abnormal elevation of AdoHcy is provided cystinuric individuals {l3,14). In the absence of data by the fact that Perry and coworkers {7,8) found on the B6-responsiveness of these patients, it is uncer­ unusual amounts of this compound and of 5-amino-4- tain whether or not they had any residual CBS activ­ imidazolecarboxamide-5'-S-homocysteinylriboside, a ity. Reliable assays of tissue cysteine concentrations compound structurally (and presumably metaboli­ have not been reported, but Mansoor et al. (6) found cally) closely related to AdoHcy, in the urines of their that in seven patients with homocystinuria and CBS-deficient patients. hypermethioninemia consistent with CBS deficiency The intracellular AdoMet/AdoHcy ratio, often (but not specified with respect to B6-responsiveness), termed the "methylation index," will determine plasma concentrations of total cyst(e)ine ranged from the relative activities of most AdoMet-dependent 52 to 167~, below the mean normal value of ap­ methyltransferases (9). It can be inferred that this proximately 250~ (6). ratio may be relatively undisturbed in CBS-deficient patients by the fact that they seem not to suffer from the neurologic abnormalities associated with The Mechanism of B6-responsiveness defective myelination that are common in patients As discussed above, a major determinant of the clini­ with methylenetetrahydrofolate reductase (MTHFR) cal prognosis for CBS-deficient patients is whether or deficiency {lO) who, because of their low, or low not they are B6-responsive. In considering possible normal, methionine and elevated tHcy concentra­ mechanisms of response it is important to realize that tions, may be expected have low AdoMet/AdoHcy although B6 administration may bring about marked ratios. Jencks and Matthews stated that the physi­ drops in plasma tHcy and methionine in responsive ologic activity of MTHFR "must ... be determined patients together with rises in abnormally low plasma by competition between AdoMet and AdoHcy for cyst(e)ine concentrations, these patients are not re­ a common ligand-binding site and would therefore stored to biochemical normalcy. They usually con­ be expected to be very sensitive to the AdoMet/ tinue to have some elevation of homocystine in AdoHcy ratio" (11). The inferred normality of this plasma and urine, and respond abnormally to a me­ ratio in CBS-deficient patients would mean that there thionine load as judged by the extent and/or duration is no interference with the biosynthetic pathway to of elevations of plasma methionine and of plasma and N l -methyltetrahydrofolate, and that there should be urinary homocystine (15,16). sufficient amounts of this compound to support me­ In principle, the B6-induced response could be due thylation of the abnormally elevated homocysteine. either to stimulation of an alternative pathway for The increased input of methyl groups through catabolism of methionine or homocysteine, or to homocysteine methylation raises the problem of stimulation of residual CBS activity. No evidence has how these methyl groups are to be disposed of. If been reported indicating stimulation of an alternative utilization for homocysteine methylation keeps N l _ pathway, and indeed there is a possibility that the methyltetrahydrofolate at a low concentration, gly­ methionine transamination pathway, which is the cine methyltransferase will be active and methyl major known alternative route for methionine degra­ groups will flow to sarcosine as a step in their dation, is inhibited in CBS-deficient patients (17). ultimate oxidation. It would thus be of interest to Furthermore, the genetic evidence now available assess quantitatively the rates of homocysteine proves that B6-responsiveness or nonresponsiveness is methylation and sarcosine formation in CBS-deficient determined by some property of the mutant CBS patients. itself. This was first indicated by the fact that B6- 11. CYSTATHIONINE ~-SYNTHASE DEFICIENCY: METABOLIC ASPECTS 79

TABLE 11-1. Cystathionine ~-synthase activities in livers of cysathionine ~-synthase-deficient patients

B6-responsive B6-nonresponsive

B6 administration

no yes no yes

Liver CBS activity, % of mean control* Citation

12.3** 35 30.5** 0 36 2.0 3.0 15 1.2 4.6 3.7 16.5 0 o 37 9.4 11.7 0 o 5.9 22.6 5.6 24.7 15.3** 55.1 7.1 6.7 38 6.2 10.7 4.0 8.4

* All in vitro assays with pyridoxal-5'-phosphate added. ** Patient with less marked homocystinuria: genetic status uncertain.

responsiveness or nonresponsiveness is the same for all by the addition of cofactor, as would be expected

CBS-deficient sibs within a given sibship {4}, sug­ were these simply Km mutations. By contrast, no B6- gesting that the same (or a closely linked) genetic nonresponsive patient, whether or not given B6, had mutation that makes an individual CBS-deficient de­ detected hepatic activity. * termines whether he will be B6-responsive. More The physiologic importance of the relatively small recent studies of mutations in the CBS gene have activities of CBS in the B6-responsive patients has firmly supported this conclusion. Striking examples been indicated by measurements of the abilities of include the findings that "the 'Celtic' mutation these patients to convert the sulfur of an oral load of G307S appears to be incompatible with B6 respon­ methionine to inorganic sulfate, and by determina­ siveness whether it is present in one or two copies in tion by nitrogen balance studies of their capacities to the patient" {I8}, and that the prevalent mutation form cysteine from methionine. The normal human I278T is associated in homozygotes with B6 respon­ has the capacity to convert methionine sulfur to inor­ siveness and, in compound heterozygotes, with some ganic sulfate at a rate of at least 84 millimole/day. In degree of B6 responsiveness {l9}' their basal states, B6-responsive CBS-deficient pa­ A few assays of CBS activity in livers of CBS­ tients had a mean capacity of 8.3 millimole/day, and deficient patients were performed before the discovery this capacity was increased by B6-treatment to 16.1 that this activity is expressed in normal cultured skin millimole/day. Although each of these rates is small fibroblasts later obviated the need for studies of liver compared to normal, they are sizable compared to the for diagnostic purposes. Those carried out with suffi­ normal methionine intake of some 8-12 millimole/ cient sensitivity to detect relatively small residual day {l5}. Finally, in contrast to normal humans, B6- CBS activities are summarized in table 11-1. Each B6- nonresponsive CBS-deficient patients do not maintain responsive patient had detected residual activity, and treatment of the patient with therapeutic doses of B6 * The results of assays of CBS activities in cultured skin fibroblasts almost invariably led to at least some increase in the generally support the suggestion that B6-responsive patients have hepatic activity. Pyridoxal 5'-phosphate was added in small residual activities of CBS, whereas B6-nonresponsive patients vitro for these assays. This fact, in addition to other do not {33]. However, experience has shown that such results are fallible, at least in the sense that individuals who must have CBS available evidence {3}, makes it clear that in no in­ activity in their liver may fail to manifest such activity in cultured stance was the deficit in activity overcome merely fibroblasts. See, for example, de Franchis et at. {34}. 80 I. BIOCHEMISTRY AND GENETIC STUDIES

nitrogen balance when given methionine-adequate, ments are likely to have hyperhomocyst(e)inemia only cystine-free diets [20,21]. However, B6-responsive following intake of methionine (i.e., at times when patients, even on normal B6 intakes, did remain in the methionine concentration is also high), rather nitrogen balance on such diets. Titration of the than fasting hyperhomocyst(e)inemia [29,30]. Again, amount of cystine required to maintain balance in if the transsulfuration pathway is normally not opera­ a nonresponsive patient indicated that the B6- tive in endothelial cells and vascular tissues (a point responders must be synthesizing each day at least upon which present evidence conflicts [31,32]), one 2.5-4.0 millimole of cysteine, a minimal value that is might expect the pathophysiologic effects of a given sizable compared to the normal methionine intake extent of hyperhomocyst(e)inemia to be less severe [21] and that, if stimulated by no more than two- to with impairments of CBS than of homocysteine me­ threefold, would become sufficient to metabolize such thylation (JD Finkelstein, personal communication). an intake. These possible mitigating effects are not mutually exclusive. They may merit being borne in mind as work progresses on hyperhomocyst(e)inemia and its Are the Vascular-damaging Effects of role in cardiovascular disease. Perhaps nature's experi­ Hyperhomocysteinemia Somehow Partially ments with CBS-deficiency have lessons to teach us Mitigated in CBS-deficient Patients and in worthy of our attention. Heterozygotes for CBS Deficiency? The results of long-term treatment of CBS-deficient References patients are now beginning to be reported [22-24]. 1. Mudd SH, Finkelstein JD, Irreverre F, Laster 1. B6-nontesponsive patients have been free of throm­ Homocystinuria: An enzymatic defect. Science 143: boembolic events, both those in Australia treated 1443-1445, 1964. with betaine, B6, folate, and B12 [22,23], and those in 2. Carson NAJ, Cusworth DC, Dent CE et al. Ireland given low-methionine diets (E Naughten, Homocystinuria: A new inborn error of metabolism personal communication). This is surprising in view associated with mental deficiency. Arch Dis Child of the fact that, for the Australian patients, medica­ 38:425-436, 1963. tions were given in doses that maintained plasma free 3. Mudd SH, Levy HL, Skovby F. Disorders of homocyst(e)ine at "usually ... less that 60 J.1M" [22], transsulfuration. In: Scriver CR, Beaudet At, Sly WS, Valle D (eds) The Metabolic and Molecular Bases 0/ Inher­ so that plasma tHcy may have been at least twice this ited Disease, 7th ed. New York: McGraw-Hill, Inc., value [6]-i.e., at least 120J.1M. Compared with the 1995, pp 1279-1327. levels of hyperhomocyst(e)inemia reported in the ab­ 4. Mudd SH, Skovby F, Levy HL et al. The natural history stracts and papers for this meeting to confer signifi­ of homocystinuria due to cystathionine /3-synthase cant relative risks of thromboembolic phenomena, deficiency. Am] Hum Genet 37:1-31, 1985. 120JlM is extremely high. Why have these patients 5. Mansoor MA, Svardal AM, Ueland PM. Determination been free of thromboembolism? of the in vivo redox status of cysteine, cysteinylglycine, The same question arises with regard to heterozy­ homocysteine, and glutathione in human plasma. gotes for CBS deficiency. Such individuals are not Analyt Biochem 200:218-229, 1992. overrepresented among individuals with premature 6. Mansoor MA, Ueland PM, Aarsland A, Svardal A. Redox status and protein binding of plasma vascular disease [25,26], and are at little, if any, homocysteine and other aminothiols in patients increased relative risk for such disease [27,28]. Could with homocystinuria. Metabolism 42:1481-1485, it be that the vascular-damaging effects of hyper­ 1993. homocyst(e)inemia are somehow mitigated in indi­ 7. Perry n, Hansen S, MacDougall L, Warrington PD. viduals with partial or severe lack of CBS activiry? Sulfur-containing amino acids in the plasma and urine Perhaps methionine acts as a competitive structural of homocystinurics. Clin Chim Acta 15:409-420, inhibitor of this damage, and the elevated concentra­ 1967. tions of this amino acid in CBS-deficient individuals, 8. Perry n, Hansen S, Bar H-P, MacDougall 1. in contrast to its normal or low concentrations in Homocys~inuria: Excretion of a new sulfur-containing those with homocysteine methylation difficulties, is amino acid in urine. Science 152:776-778, 1966. 9. Cantoni GL, Richards HH, Chiang PK. Inhibitors of protective. Other differences exist between individu­ S-adenosylhomocysteine hydtolase and their role in the als with impaired CBS and impaired homocysteine regulation of biological methylation. In: Usdin E, methylation-the former have higher concentrations Borchardt RT, Creveling CR (eds) Transmethylation. of AdoMet and lower concentrations of cystathionine New York: ElsevierlNorth-Holland, 1979, pp 155- and, perhaps, cysteine; and those with milder impair- 172. 11. CYSTATHIONINE ~-SYNTHASE DEFICIENCY: METABOLIC ASPECTS 81

10. Rosenblatt DS. Inherited disorders of folate transport 24. Ward P, Naughten E, Cahalane S, Murphy D, and metabolism. In: Scriver CR, Beaudet AL, Sly WS, Mayne P. Clinical outcome in a neonatal screened Valle D (eds) The Metabolic and Molecular Bases of Inher­ population with homocystinuria (HCU) 1971- ited Disease, 7th ed. New York: McGraw-Hill, Inc.,- 1995 Ireland (Abstr). Ir ] Med Sci 164 (Suppl 15): 1995, pp 3111-3128. 24, 1995. 11. Jencks DA, Marrews RG. Allosteric inhibition of 25. Kozich V, Kraus E, de Franchis R et al. Hyper­ methy lenetetrahydrofolate reductase by adeno­ homocysteinemia in premature arterial disease: sylmethionine: Effects of adenosylmethionine and Examination of cystathionine ~-synthase alleles at NADPH on the equilibrium between active and inac­ the molecular level. Hum Mol Genet 4:623-629, tive forms of the enzyme and on the kinetics of ap­ 1995. proach to equilibrium.] Bioi Chem 262:2485-2493, 26. Whitehead AS, Ward P, Tan S et al. The molecular 1991. genetics of homocystinuria, hyperhomocysteinaemia, 12. Tallan HH, Moore S, Stein WHo L-Cystathionine in and premature vascular disease in Ireland. In: Mato human brain.] Bioi Chem 230:707-716, 1958. JM, Caballero A (eds) Methionine Metabolism: Molecular 13. Gerritsen T, Waisman HA. Homocystinuria: Absence Mechanisms and Clinical Implications. Madrid, Spain: of cystathionine in the brain. Science 145:588, 1964. Cosejo Superior de Investigationes Cientfficas, 1994, 14. Brenton DP, Cusworth DC, Gaull GE. Homocys­ pp 81-83. tinuria: Biochemical stuies of tissues including a com­ 27. Mudd SH, Havlik R, Levy HL, McKusick VA, parison with cystathioninuria. Pediatrics 35:50-56, Feinleib M. A study of cardiovascular risk in heterozy­ 1965. gotes for homocystinuria. Am] Hum Genet 33:883- 15. Mudd SH, Edwards W A, Loeb PM, Brown MS, Laster 893, 1981. 1. Homocystinuria due to cystathionine synthase defi­ 28. de Valk HW, van der Griend R, van Eeden MKG ciency: The effect of pyridoxine.] Gin Invest 49: 1762- et al. Absence of premature atherosclerosis in adults 1773, 1970. with heterozygosity for cystathionine-~-synthase 16. Boers GHJ, Smals AGH, Drayer JIM et al. Pyridoxine deficiency (Abstr). Ir] Med Sci 164(Suppl 15):6, treatment does not prevent homocystinemia after me­ 1995. thionine loading in adult homocystinuria patients. 29. Brattstrom L, Israelsson B, Norrving B et al. Impaired Metabolism 32:390-397, 1983. homocysteine metabolism in early-onset cerebral and 17. Blom HJ, Boers GHJ, Trijbels JMF, van Roessel peripheral occlusive arterial disease: Effects of pyridox­ JJM, Tangerman A. Cystathionine-synthase­ ine and folic acid treatment. Atherosclerosis 81:51-60, deficient patients do not use the transamination 1990. pathway of methionine to reduce hypermethionine­ 30. Miller JW, Nadeau MR, Smith D, SelhubJ. Vitamin mia and homocystinemia. Metabolism 38:577-582, B-6 deficiency vs folate deficiency: Comparison of re­ 1989. sponses to methionine loading in rats. Am] Clin Nutr 18. Kraus JP. Molecular basis of phenotype expression 59:1033-1039, 1994. in homocystinuria. ] Inher Metab Dis 17:383-390, 31. Jacobsen DW, Savon SR, DiCorleto PE. Metabolism of 1994. homocysteine by vascular cells and tissues: Is the 19. Shih VE, Fringer JM, Mandell R et al. A missense transsulfuration pathway active? (Abstr). Ir] Med Sci mutation (I278T) in the cystathionine ~-synthase gene 164(Suppl 15):10, 1995. prevalent in pyridoxine-responsive homocystinuria and 32. Wang J, Dudman NPB, Wilcken DEL, Lynch JF. associated with mild clinical phenotype. Am] Hum Homocysteine metabolism: Levels of 3 enzymes in Genet 57:34-39, 1995. cultured human vascular endothelium and their 20. Brenton DP, Cusworth DC, Dent DE, Jones EE. relevance to vascular disease. Atherosclerosis 97:97-106, Homocystinuria: Clinical and dietary studies. Q] Med 1992. 35:325-346, 1966. 33. Mudd SH, Levy HL, Skovby E. Disorders of 21. Poole JR, Mudd SH, Conerly EB, Edwards W A. transsulfuration In: Scriver CR, Beaudet AL, Sly WS, Homocystinuria due to cystathionine synthase defi­ Valle D (eds) The Metabolic Basis of Inherited Disease, ciency: Studies of nitrogen balance and sulfur excre­ 6th ed. New York: McGrwa-Hill, Inc., 1989, pp 693- tion.] Gin Invest 55:1033-1048, 1975. 734. 22. Wilcken B, Wilcken D. The long-term outcome 34. de Franchis R, Kozich V, McInnes RR, Kraus JP. in homocystinuria: Treatment and thrombo-embolic Identical genotypes in siblings with different events (Abstr). SSIEM: 32nd annual symposium, homocystinuric phenotypes: Identification of three Edinburgh, 6th-9th September 1994. mutations in cystathionine ~-synthase using an im­ 23. Wilcken DEL, Wilcken B. Cystathionine ~-synthase proved bacterial expression system. Hum Mol Genet deficiency: Clinical outcomes. In: Graham I, Refsum 3:1103-1108, 1994. H, Rosenberg IH, Ueland P (eds) International Confer­ 35. Finkelstein JD, Mudd SH, Irreverre F, Laster 1. ence or Homocysteine Metabolism: From Basic Science Homocystinuria due to cystathionine synthetase defi­ to Clinical Medicine. Norwell, MA: Kluwer Academic ciency: The mode of inheritance. Science 146:785-787, Publishers, 1997, pp 51-56. 1964. 82 I. BIOCHEMISTRY AND GENETIC STUDIES

36. Tada K, Yoshida T, Yokoyama Y et al. Cysta­ and ultrastructural studies. ] Pediatr 84:381-390, thioninuria not associated with vitamin B6 depen­ 1974. dency: A probably new type of cystathioninuria. 38. Longhi RC, Fleisher LD, Tallan HH, Gaull GE. TohokuJ Exp Med 95:235-242, 1968. Cystathionine ~-synthase deficiency: A qualitative ab­ 37. Gaull GE, Sturman ]A, Schaffner F. Homocystinuria normality of the deficient enzyme modified by vitamin due to cystathionine synthase deficiency: Enzymatic B6 therapy. Pediatr Res 11:100-103, 1977.