Main

Homocystinuria due to cystathionine β-synthase (EC 4.2.1.22) deficiency is an inborn error of the transsulfuration and one-carbon transfer pathways (Fig. 1). It has ocular, skeletal, vascular, and neurologic complications(1). Biochemically the disorder is characterized by an increase in plasma concentrations of homocystine (and other mixed disulfides of homocysteine) and methionine, with homocystine in the urine. The cofactor for cystathionine β-synthase is pyridoxal phosphate, and approximately half of patients with homocystinuria respond to pharmacologic doses of pyridoxine. Pyridoxine non-responders are treated by a low methionine diet and/or the methyl-donor betaine. Betaine is a substrate for the enzyme betaine-homocysteine methyltransferase (EC 2.1.1.5) that catalyzes the remethylation of homocysteine to methionine. Betaine has previously been shown to reduce plasma concentrations of homocysteine(2–5) and to correct secondary abnormalities in plasma serine concentration(4,5).

Figure 1
figure 1

The transmethylation and transsulfuration pathways. THF, tetrahydrofolate; CH2, methylene; and CH3-R, a methyl group acceptor. 1, S-adenosylhomocysteine hydrolase; 2, cystathionine β-synthase; 3, 5,10-methylenetetrahydrofolate reductase; and 4, homocysteine:5-methyltetrahydrofolate methyltransferase. The block shows the site of interruption of the pathway in cystathionine β-synthase deficiency.

Full size image

The neurologic complications of cystathionine β-synthase deficiency include intellectual deficits, seizures, dystonia and other extrapyramidal motor disorders, rages and other psychiatric disturbances, and cerebrovascular events(6–10). The pathogenesis of these is not clear. Intellectual deficits occur in approximately 50% of cases, and although cerebrovascular disease has been suspected as the cause(1, 11, 12), the absence of focal clinical signs and the partial reversibility of the impairment in response to treatment suggest a metabolic basis(2, 3, 10, 13). There is little information about the metabolic changes in the CNS in cystathionine β-synthase deficiency, although homocystine has been found to be elevated in the cerebrospinal fluid of two untreated patients(1, 14). This suggests that homocysteine may accumulate within the CNS, and may be important in the pathogenesis of the neurologic complications.

To examine whether homocysteine accumulates within the CNS in cystathionine β-synthase deficiency, we have measured total homocysteine and related metabolites in the cerebrospinal fluid of five children before and during treatment with betaine. Because betaine-homocysteine methyltransferase is not present in the brain(15), any effect of betaine upon brain metabolites must be indirect by altering hepatic metabolism of homocysteine. We therefore also measured plasma concentrations of some metabolites in these patients and in five additional children with cystathionine β-synthase deficiency.

METHODS

Five patients were studied aged 6-14 y; each had pyridoxine-nonresponsive cystathionine β-synthase deficiency. The diagnosis was made by the clinical findings, demonstration of a raised methionine and homocyst(e)ine (either homocystine or total homocysteine) concentration in plasma, and the absence of a biochemical response to a trial of pyridoxine. Each child had a lumbar puncture during general anesthesia for eye surgery. Two patients were treated with a methionine-restricted diet and folic acid (10 mg daily) throughout the study, the others were untreated before they received betaine. Each patient was studied before and during treatment with betaine monohydrate (Fluka, Poole, UK) 250 mg/kg/d, the duration of treatment varied between 3 and 6 mo, and compliance was ensured by parental supervision. We also studied changes in plasma metabolites alone in five additional patients (nos. 6-10). Brief patient details are given in Table 1. Approval for the study was obtained from the Research Ethics Committee of the Institute of Child Health and Great Ormond Street Hospital for Children, London.

Table 1 Brief patient details
Full size table

Cerebrospinal fluid was sampled before and during treatment with betaine, and the following metabolites were measured: total homocysteine, methionine, S-adenosylmethionine, serine, glycine, and 5-methyltetrahydrofolate. Cerebrospinal fluid was collected in a standardized manner(16) after a 4-h fast with the last dose of betaine given 12-16 h previously, frozen in liquid nitrogen at the bedside, and stored at -70 °C until analysis. All metabolites were measured on the third milliliter of cerebrospinal fluid. Plasma samples were taken before and during treatment with betaine, and the following metabolites were measured: total homocysteine, methionine, serine, and glycine. Blood was taken into iced lithium heparin tubes after a 4-h fast in all patients, samples were separated within 15 min, and the plasma was stored at -70 °C until analysis.

A reference population of 27 closely age-matched (0.4-12.5 y) patients was also studied. Each had a neurologic disease not expected to interfere with the transmethylation or transulfuration pathways (10 nonspecific learning difficulties, 8 congenital lactic acidosis, 5 epilepsy, 2 Miller-Fisher syndrome, 1 albinism, 1 benign neonatal hypotonia). Each had simultaneous plasma and cerebrospinal fluid samples taken for diagnostic purposes, and these were treated as above.

Plasma and cerebrospinal fluid total homocysteine were measured using HPLC with UV detection, modified from a previously published method(17). The assay was linear up to 1 mM with a recovery of 93%. The lower limit of detection using spiked cerebrospinal fluid and a signal-to-noise ratio of 3:1 was 0.1 μM. The within-day variation was 5%, and the between-day variation 9%. Plasma and cerebrospinal fluid methionine, serine, and glycine were measured using a Waters HPLC PicoTag system(18). Cerebrospinal fluid S-adenosylmethionine and 5-methyltetrahydrofolate were measured by HPLC with electrochemical detection as previously described(19, 20).

The results were grouped into those before betaine treatment, during treatment with betaine, and the reference range and analyzed by ANOVA after logarithmic transformation to equalize the variances. Post hoc significance testing used Tukey's HSD test; p values <0.05 were regarded as significant. Because of the relatively small number of cerebrospinal fluid samples, these results were also analyzed using nonparametric statistical methods. ANOVA by rank sum was performed with the Kruskal-Wallis test, and post hoc significance testing used an anolog of the Tukey test(21). Statistical analyses were performed using SPSS for Windows software.

RESULTS

The results for both plasma and cerebrospinal fluid are summarized in Tables 2 and 3, and the results for cerebrospinal fluid are shown in Figure 2. There was no difference in the findings using either parametric or nonparametric statistical methods, and the results using parametric methods are shown.

Table 2 Cerebrospinal fluid concentrations of the metabolites before and during treatment
Full size table
Table 3 Plasma concentration of the metabolites before and during betaine treatment
Full size table
Figure 2
figure 2

Metabolite concentrations in cerebrospinal fluid. Concentrations is shown on the y axis, and the results are arranged on the x axis in three groups: reference range to the left, before betaine treatment in the center, and during betaine treatment to the right. Results from each patient are joined by a line. (A) total homocysteine, (B) methionine, (C) S-adenosylmethionine, (D) serine, (E) glycine, and (F) 5-methyltetrahydrofolate.

Full size image

Cerebrospinal fluid. Cerebrospinal fluid total homocysteine was significantly raised above the reference range before and during treatment with betaine, but treatment caused a significant reduction. Cerebrospinal fluid methionine was also significantly raised before treatment, and although the mean rose during treatment with betaine, the increase was not statistically significant. However, treatment did cause a significant rise in the cerebrospinal fluid methionine to homocysteine ratio. Betaine treatment caused the cerebrospinal fluid S-adenosylmethionine concentration to increase significantly above the reference range. Cerebrospinal fluid serine was significantly depressed before treatment, but normalized with treatment. Cerebrospinal fluid glycine was also significantly reduced before treatment and fell further with treatment. Although the mean 5-methyltetrahydrofolate concentrations fell with betaine treatment, there were only three posttreatment values, and the results were not statistically significant.

Plasma. In all the patients the plasma total homocysteine was significantly raised above the reference range before treatment. Treatment with betaine caused a significant fall in homocysteine concentrations, but these remained significantly above the reference range. Plasma methionine was significantly raised above the reference range both before and during treatment with betaine; but there was no significant difference between treated and untreated concentrations. The plasma methionine to homocysteine ratio was significantly increased before and during betaine therapy, but although treatment further increased the ratio, this was not statistically significant. Although there was no significant difference in the plasma concentrations of serine nor glycine before or during treatment compared with the reference range, treatment significantly raised plasma serine concentrations and reduced glycine.

Two children (nos. 1 and 10) were treated for 1 mo with a higher dose of betaine (500 mg/kg/d), which did not cause a further decrease in plasma total homocysteine concentrations.

DISCUSSION

We were unable to detect homocysteine in cerebrospinal fluid from our reference population, indicating that the normal concentration is less than 100 nM. Two reference ranges for cerebrospinal fluid total homocysteine concentrations have recently been published and give conflicting results. One group found concentrations of 7-20 nM in six adult control subjects(22), supporting the findings here. However, reference values in nine adult control subjects have also been reported to range from 280 to 660 nM(23), values that fall into the range we found in cystathionine β-synthase-deficient children receiving betaine. These latter authors used a different HPLC methodology, described adult patients, and gave no details of the collection method and the subsequent handling of the cerebrospinal fluid. It is possible that these factors may explain the discrepancies.

In children with cystathionine β-synthase deficiency, our findings in plasma are similar to those previously reported(3, 5). Before betaine therapy, plasma total homocysteine and methionine were raised. Betaine significantly reduced total homocysteine, but did not significantly alter plasma methionine, although concentrations increased in six patients. The effect of betaine on plasma methionine concentrations has previously been shown to be variable(3), and the present findings do not alleviate earlier fears that methionine toxicity might limit its use, because high concentrations are observed in some patients(24). During betaine therapy, however, plasma total homocysteine concentrations remain significantly raised, and increasing the dose of betaine did not further reduce these in the two patients studied. Unlike previous reports, plasma serine was not significantly reduced before treatment, but betaine did cause a significant rise in serine concentrations and a reciprocal significant fall in glycine concentration. This suggested that betaine decreased the flux through folate-dependent remethylation pathways as previously reported(5). The decreased flux through folate-dependent remethylation pathways might be caused by decreased activity of homocysteine:5-methyltetrahydrofolate methyltransferase (EC 2.1.1.13; methionine synthase)(5). However, our findings in cerebrospinal fluid that betaine treatment is associated with a fall in 5-methyltetrahydrofolate and a significant increase in S-adenosylmethionine suggest that the mechanism might be due to the inhibition of 5,10-methylenetetrahydrofolate reductase (EC 1.5.1.20) by S-adenosylmethionine(25). Most of our patients were not receiving supplemental folic acid treatment, which could limit the flux through folate-dependent pathways, and this may explain why plasma serine concentrations were not significantly decreased before treatment(5).

Total homocysteine and related metabolite concentrations in the cerebrospinal fluid from patients with cystathionine β-synthase deficiency have not previously been described. We found that these, in general, replicated the findings in plasma. The exception was glycine, which was significantly reduced in cerebrospinal fluid before treatment and fell further with betaine therapy.

The primate brain normally contains the enzyme cystathionine β-synthase(26), and its deficiency would be expected to lead to an accumulation of homocysteine within the cells of the CNS. In other tissues, intracellular accumulation of homocysteine causes export of homocysteine into the extracellular space(27). In the CNS, cerebrospinal fluid is in continuity with the extracellular space, and accumulation of homocysteine in this fluid is likely to reflect accumulation within the brain. The mammalian brain does not contain the enzyme betaine-homocysteine methyltransferase(15). So the effect of betaine is most probably mediated through changes in the plasma homocysteine concentration. Because we found that cerebrospinal fluid homocysteine concentrations decreased with betaine treatment, it suggests that the reduction in plasma total homocysteine concentration caused by betaine is responsible for the decrease in cerebrospinal fluid total homocysteine concentration. It is possible that a reduction in plasma concentration of homocysteine facilitates the active removal of homocysteine from the brain, thereby reducing its accumulation in the cerebrospinal fluid.

Although accumulation of homocysteine within the CNS seems to be one likely pathogenic mechanism leading to neurologic damage, it is not clear how this is mediated. The enzyme S-adenosylhomocysteine hydrolase (EC 3.3.1.1), which catabolizes S-adenosylhomocysteine (Fig. 1) is reversible and has kinetics that favor S-adenosylhomocysteine formation(28). S-Adenosylhomocysteine is a potent inhibitor of biologic methylations, and the ratio of S-adenosylmethionine to S-adenosylhomocysteine concentration determines the rate of biologic methylation(29). Here we have shown that the accumulation of homocysteine in the CNS is greatly in excess to the accumulation of S-adenosylmethionine, and it is likely that a wide variety of biologic methylations are disturbed. Another potential mechanism might involve the oxidation of homocysteine to homocysteic acid, which is an excitotoxin that activates the glutamate N-methyl-D-aspartate receptor(30). N-Methyl-D-aspartate receptor activation can cause neuronal cell death and is a suggested mechanism for the pathogenesis of many neurologic disorders(31).

We also showed a reduction in cerebrospinal fluid concentrations of serine and glycine before treatment and a further reduction in glycine with betaine treatment. These amino acids are readily interconverted and enter into a large number of synthetic reactions via the single carbon pool. Recently, Jaeken et al.(32) described 3-phosphoglycerate dehydrogenase deficiency, which causes deficiency of serine and glycine within the CNS and a severe neurologic disorder; although the reductions in serine and glycine are greater than those observed here.

The findings that homocysteine accumulated in the CNS in children with cystathionine β-synthase deficiency, and serine and glycine are significantly reduced, lend support to a biochemical pathogenesis of the neurologic complications(9). Although betaine has been shown to improve patient well being, this is the first demonstration that it can reduce the accumulation of homocysteine within the CNS. But, like the findings in plasma, betaine does not reduce cerebrospinal fluid homocysteine to normal, and the long-term outcome of this treatment is not known.