Betaine rescue of an animal model with methylenetetrahydrofolate reductase deficiency

Animal husbandry and experiments

All mice were produced in our own breeding facility and housed in adequate shoebox cages with free access to food and water. The animal experiments were approved by the Animal Care Committee of the Montreal Children's Hospital and complied with the guidelines of the Canadian Council for Animal Care.

Mice, heterozygous for a disruption of the Mthfr gene (Mthfr+/−) and from F6 or F7 generations of backcrosses to the BALB/cAnNCrlBR strain, were used for breeding. Female Mthfr+/− mice on regular rodent chow (Purina laboratory rodent diet 5001; Purina Mills) were supplemented with anhydrous betaine (Sigma) at 2% (w/v) in the drinking water before mating them to Mthfr+/− males. Water intake was monitored. Betaine supplementation was continued until weaning of the pups at 3 weeks of age. Untreated matings served as controls. Cages were checked daily. The day of birth was designated as P0 (postnatal day 0). Newborn mice were examined daily for viability and body weight. The body weights of homozygous mutant (Mthfr−/−) mice were compared with those of their wild-type (Mthfr+/+) and heterozygous (Mthfr+/−) littermates. A few mice who died before P2 were excluded from weight and mortality calculations. This mortality, due to non-specific causes, was gender- and genotype-independent.

A total of 15 Mthfr+/+ and Mthfr+/− mice of betaine-supplemented mothers were supplemented with 2% betaine in the drinking water from weaning until 7–9 weeks of age for measurements of betaine and DMG (dimethylglycine) in plasma and tissues. Results were compared with those from 20 untreated mice of the same age whose mothers had not been supplemented with betaine.

At different prenatal [E18 (embryonic day 18)] and postnatal (P6, P9, P12 and adult) time points, mice were killed. The blood of older mice was collected by heart puncture, anticoagulated with EDTA (MicrovetteR 500; Sarstedt) and immediately placed on ice. Plasma was separated by centrifugation at 10000 g for 5 min and frozen immediately at −70 °C until analysis. Livers and brains were dissected and weighed. They were then frozen on solid CO₂ and stored in aliquots at −70 °C, or fixed by immersion in 10% (v/v) formalin and processed for 5 μm paraffin sections and haematoxylin/eosin staining.

Genotyping

Mice were genotyped for the Mthfr disruption by PCR of genomic DNA extracted from tail biopsies, as described previously [9].

Metabolites

Plasma total homocysteine concentrations were measured after reduction by a HPLC method, as described previously [11]. Betaine and DMG concentrations in plasma and tissue extracts were analysed by HPLC as reported in [12]. Betaine and DMG were extracted from tissues by adding 1 ml of 100 mM KH₂PO₄ (pH 4.5) to 100–150 mg of freeze-dried tissues in 2 ml Eppendorf tubes. After addition of an equal volume of clean glass beads, the tubes were mixed vigourously to homogenize the contents. The mixture was incubated for 10 min with intermittent mixing and vortexing. At the end of the incubation time, the tubes were centrifuged (5000 g, 10 min) and the supernatant was collected. The extraction procedure was repeated twice and the combined supernatants were frozen until analyses.

Tissue aminothiols were measured as reported [13], and tissue SAM and SAH (S-adenosylhomocysteine) concentrations were measured according to the method described by Melnyk et al. [14]. Global DNA methylation was assessed by a cytosine-extension assay as described by Pogribny et al. [15]. Adenosine in tissues was measured according to Nithipatikom et al. [16] after adaptation for electrochemical detection.

BHMT activity

BHMT activity in crude liver extracts was analysed as reported previously [17].

Statistical analyses

Standard software was used for all analyses. Arithmetric means±S.E.M. are provided. Genotype and gender distributions were compared using the χ² test. Weights or metabolite levels were compared between groups using one-way ANOVA and, if a significant difference was found, single parameters were compared with the unpaired, two-sided t test. If values were not normally distributed, the Kruskal–Wallis test and the non-parametric Wilcoxon test were used as appropriate. Linear correlations between two parameters were calculated and expressed using Spearman's coefficient. Throughout all analyses, a P value of <0.05 was considered significant.

Betaine supplementation increases the survival rate of Mthfr−/− mice

A total of 314 mice in 50 litters (mean litter size 6.28±0.31) were born after mating of untreated Mthfr+/− mice. Mthfr genotype frequencies were 25.1% Mthfr+/+, 48.0% Mthfr+/− and 26.9% Mthfr−/−. This genotype distribution was not significantly different from expected values using the χ² test. Gender proportions showed a small but significant preponderance of female (56.4%) compared with male (43.6%) mice for the Mthfr+/+ and Mthfr+/− genotypes. The genders of the 45 Mthfr−/− mice were evenly distributed (23 female, 22 male; with five surviving females and six surviving males). One-third of the Mthfr−/− mice could not be classified by gender because of early death and/or incomplete recovery of their bodies. Of the homozygous mutants, 83% died, at an average age of 9.1±0.7 days (median 7.0 days; n=53). Only 11 out of 64 (17%) Mthfr−/− mice survived (Figure 1).

Betaine supplementation in the drinking water was tolerated without apparent side effects. A mean daily betaine intake in water of approx. 3 g/kg body weight was observed during mating and pregnancy, which increased to a variable extent during lactation. Matings of betaine-treated heterozygotes yielded a total of 249 mice in 38 litters. The litter size (mean 6.38±0.44) was not different from that of untreated matings. Genotype frequencies were 27.3% Mthfr+/+, 44.9% Mthfr+/− and 27.8% Mthfr−/−. Gender (51.1% female, 48.9% male) and genotype distributions were not significantly different from expected values. The mortality of homozygous mutant pups from betaine-treated mothers decreased significantly compared with that of pups from untreated mothers, to 26% (average day of death 16.3±3.4; median 11.0 days; n=16). Of 62 Mthfr−/− mice, 46 (74%) survived. Withdrawal of betaine supplementation after weaning of homozygous mutant pups, at the age of 3 weeks, had no adverse effects on survival (Figure 1).

Betaine supplementation improves the somatic development of Mthfr−/− mice

Biochemical consequences of betaine supplementation

Betaine supplementation improves the survival and somatic development of MTHFR deficient mice

Severe MTHFR deficiency due to homozygous disruption of the mouse Mthfr gene leads to a distinct pathology that was described in the original study of these mice when they were on a mixed genetic background [9]. In contrast with the lower mortality reported in that study (survival of 76% of nullizygotes), the majority of Mthfr−/− mice in the present study, after multiple rounds of backcrosses to the BALB/cAnNCrlBR strain, died during the first two postnatal weeks. The mixed genetic background of the initial colony presumably alleviated some of the deleterious effects of severe MTHFR deficiency in the homogeneous BALB/c background.

The expected genotype distribution of offspring from Mthfr+/− matings was observed in the present study, as well as in the study of Chen et al. [9], indicating an absence of fetal losses. However, the lower birth weights of the Mthfr−/− pups point to some degree of intrauterine onset of pathology. Nonetheless, we were not able to distinguish surviving from non-surviving nullizygotes during the first few days of extrauterine life. The most reliable indicator of imminent death was poor weight gain over a few days and eventually weight loss, due to insufficient food intake as indicated by empty stomachs. Another typical feature of nullizygotes was pronounced deceleration of weight gain after weaning.

In another dietary animal model, severe folate deficiency during pregnancy and lactation led to increased fetal resorption, small litter size, low birth weight, postnatal death and depressed food consumption and weight gain, with significantly lower body and brain weights at weaning [18]. Some of these symptoms resemble those of the nullizygous pups of our present study, although we did not observe any intrauterine mortality.

The major pathology in Mthfr deficiency appears to occur after birth, suggesting that maternal folate pools may protect mutant pups in utero. Standard rodent chow contains very high levels of folate and preformed methyl groups, which are transferred from maternal to fetal tissues. However, the early postnatal period, with its increased need for methyl groups to maintain rapid cellular proliferation and differentiation, presents a particular challenge for Mthfr-deficient pups. Primary pathogenic factors in MTHFR deficiency may be toxicity of elevated homocysteine or its derivatives, impaired methylation or a disruption in the balance of folate species. Since betaine has not been reported to correct imbalances in folate derivatives, and betaine supplementation clearly improved survival and growth in Mthfr-deficient pups, the primary pathogenic factors presumably relate to the accumulation of homocysteine or its metabolite, SAH, and the consequent inhibition of transmethylation reactions.

Betaine supplementation of dams had a dramatic effect on the survival and growth of their nullizygous pups. With this simple nutritional intervention, most of the nullizygotes could be rescued from death. Nullizygotes of treated mothers had significantly higher body weights compared with untreated nullizygotes during the first two postnatal weeks, but not at birth. The lack of adverse effects of withdrawal of betaine 3 weeks after birth also demonstrates the importance of an adequate supply of transferable methyl groups during the critical first 2 weeks of postnatal life. The primary reason for the high mortality of untreated Mthfr nullizygotes may therefore be an insufficient intake of preformed methyl groups, leading to developmental disturbances and, beyond a point of no return, to starvation and death.

Betaine supplementation to lactating dams increases betaine availability in their pups and ameliorates biochemical anomalies in nullizygotes

Betaine-supplemented dams and their suckling offspring had elevated plasma betaine levels, demonstrating increased betaine availability in pups from treated lactating mothers. These mothers are heterozygotes for the Mthfr gene disruption, a condition that is associated with elevated plasma homocysteine [9,10], decreased 5-methyltetrahydrofolate [9] and decreased liver choline and betaine [10]. These anomalies are likely to be aggravated by pregnancy and lactation, conditions that are associated with choline and betaine depletion in rats [19].

In a previous study on the effects of 2 weeks of betaine supplementation to adult Mthfr-deficient mice [10], we observed some disturbances in choline metabolism in untreated animals. We suggested that the fatty infiltration of livers in adult nullizygous mice might be due to choline deficiency. In the present study, most of the livers of nullizygotes at age P6 also had clear fatty infiltration. Neonatal rats have a 2-fold higher intake of choline than weanling animals. Rat and human milk do not contain measurable amounts of betaine [20,21]. However, since choline can be oxidized to betaine, choline can replace betaine as a source of transferable methyl groups. Consequently, betaine supplementation could increase choline availability in the livers of nullizygous mice either by a choline-sparing effect or by provision of methyl groups for the methylation of phosphatidylethanolamine to phosphatidylcholine, a precursor of choline. However, betaine supplementation did not decrease the extent of fatty infiltration of the liver in mutant pups in the present study. This could be a dose-dependent effect, particularly if the preformed methyl groups are utilized preferentially in maintaining brain methylation reactions at the expense of the liver in early postnatal life. On the other hand, since SAH levels were still elevated compared with control levels despite a betaine-lowering effect, phosphatidylethanolamine methyltransferase might have been inhibited.

When we investigated betaine metabolism and the consequences of high-dose betaine supplementation in adult Mthfr+/+ and +/− mice in the present study, we identified betaine depletion in the plasma, liver and brain of Mthfr+/− animals. Depletion of betaine in liver is attributable to the decreased capacity of Mthfr+/− mice to meet the demands for folate-dependent homocysteine remethylation, which leads to lower SAM concentrations. The compensatory increase in betaine catabolism by BHMT to restore homocysteine remethylation results in a decrease in betaine levels in liver. Betaine concentrations in other tissues are dependent on betaine export by hepatocytes and are probably reduced because betaine availability is decreased through lowered plasma concentrations. Betaine supplementation leads to greatly enhanced betaine storage and further induction of betaine metabolism, as indicated by increased concentrations of its metabolite DMG and increased BHMT activity (Table 1). These findings are in agreement with our previous study [10].

BHMT activity in fetal rat liver is only approx. 25% of that found in adult liver, whereas the activity in the neonate is equal to or greater than that in adult rats [22]. In agreement with these earlier findings, fetal mice in our study had low BHMT activity a few days before birth, with a 3-fold increase at P12, up to adult levels. The dramatic effect of betaine on the survival of Mthfr−/− mice provides strong evidence for an important role for BHMT activity during the first days of postnatal life in mice.

SAH is the product of transmethylation reactions, which occur primarily in liver [23]. When homocysteine is elevated, the SAH hydrolase reaction (EC 3.3.1.1), which usually catalyses the breakdown of SAH to homocysteine and adenosine, can operate in the reverse direction, to generate SAH [2]. Betaine supplementation caused marked decreases in both homocysteine and SAH, and an increase in methionine, in heterozygotes and nullizygotes. In contrast, the administration of betaine to wild-type mice caused an increase in SAM and SAH levels, an observation that was also reported in another rodent study [24] and that could be interpreted being due to increased turnover of homocysteine by enhanced remethylation.

Adenosine was increased in the livers in heterozygotes and nullizygotes. This increase was diminished by betaine supplementation. Adenosine levels appeared to correlate strongly with the concentrations of homocysteine and SAH in liver; this finding may relate to the fact that SAH generates homocysteine and adenosine, and that adenosine elimination by adenosine deaminase may not have been enhanced.

Homocysteine can also be eliminated by trans-sulphuration, in which homocysteine condenses with serine to form cystathionine and subsequently cysteine. In the livers of nullizygous P6 pups, the most striking anomalies were elevations of homocysteine and SAH. In addition, cystathionine was markedly increased, whereas cysteine was only moderately elevated. This demonstrates the failure of co-ordinate regulation of homocysteine metabolism and, in particular, the inability of the two remaining pathways, betaine-dependent remethylation and homocysteine trans-sulphuration, to cope with a disruption of folate-dependent remethylation.

Homocysteine metabolism in the developing mouse brain and the impact of betaine supplementation

Studies in different species have shown that the prenatal brain utilizes homocysteine preferentially for folate-dependent remethylation and not for trans-sulphuration. Remethylation enzymes have higher activities during fetal life than postnatally [22,25,26]. During the first 2 weeks of postnatal life, there is an exponential decrease of folate in the mouse brain, with lower levels of polyglutamylated folates, indicating high turnover and demand for folates in early extrauterine life [27]. In contrast with the postnatal decreases in activity of the remethylation enzymes, the rate-limiting enzyme of the trans-sulphuration pathway, cystathionine β-synthase, shows decreased activity in the cerebrum during late fetal stages, increases after birth and peaks during cerebellar development [28]. The activity of the last enzyme of the trans-sulphuration pathway from methionine to cysteine, cystathionase, increases gradually in rat brain over 4 weeks after birth [29]. As a consequence of these temporal regulatory mechanisms, trans-sulphuration in fetal brain is limited.

In contrast with the liver, the brain does not use betaine as a methyl donor. There are low concentrations of betaine in brain [30], and BHMT is not expressed in this tissue [31]. When folatedependent remethylation is disturbed, the brain becomes dependent on preformed methyl groups delivered as SAM. Furthermore, the only means of homocysteine elimination is trans-sulphuration, a pathway that is limited in brain, as discussed above. Consequently, we found more severe biochemical disruption in the brain than in the liver of 6-day-old pups from untreated dams. Most prominent were dramatic increases of homocysteine and clear elevations of cystathionine and cysteine; these findings demonstrate the activation of the trans-sulphuration pathway, but also its inability to cope with the increased homocysteine load.

Brain GSH was unchanged in nullizygotes, despite the elevation of its precursor cysteine, whereas cysteinylglycine was decreased in heterozygotes and nullizygotes. Cysteinylglycine is a product of extracellular GSH degradation, but also provides cysteine and glycine as precursor molecules for GSH synthesis in neurons [32]. Cysteinylglycine in brain was inversely correlated with homocysteine and cysteine concentrations, which could indicate a preference for cysteinylglycine over cysteine for GSH synthesis.

With betaine supplementation, biochemical parameters in brain indicated increased remethylation of homocysteine to methionine and markedly decreased flux through the trans-sulphuration pathway. As a consequence, global DNA hypomethylation was decreased. Decreased homocysteine and SAH levels in mutant pups may have contributed to the phenomenon; DNA methylation status in another model of hyperhomocysteinaemia, the cystathionine β-synthase-deficient mouse, was largely determined by SAH concentrations [33].

The cerebella of Mthfr−/− mice are abnormally small and cerebellar development is disrupted, as first shown by Chen et al. [9]. We also identified some hypotrophy in the hippocampus, with abnormal formation of the neuroepithelium in the dentate gyrus. Betaine supplementation had a beneficial effect on the pups in both areas of the brain, but could not prevent residual structural abnormalities in the cerebellum in adult nullizygotes. Of note is the fact that a compartmentation defect (i.e. a difference in the degree of disturbance between the anterior and posterior cerebellar regions) was still evident.

The cerebellum is one of the late developing brain regions. Shortly after birth until the second week of life, the cerebellar granule neuroblasts proliferate extensively in the external granular layer of the cerebellar cortex, and undifferentiated Purkinje cells migrate to form a single cell layer underneath the external granular layer [34]. This postnatal proliferative and migratory period might be particularly vulnerable to the disruption in folate-dependent remethylation and supply of methyl groups. Other early postnatal insults to rodent brain also result in more severe cerebellar damage than later exposure and produce a pathological pattern that is similar to that in Mthfr−/− mice. Irradiation and methylazoxymethanol have been shown to be especially harmful before P8 and P4 respectively. Both cause foliation abnormalities, misalignment of Purkinje cells and continued multiple innervation of climbing fibres [35].

Intrauterine choline deficiency has been shown to result in structural defects in the hippocampus of rats [36] and mice [37–39]. Choline availability altered the timing, genesis, migration and differentiation of progenitor neuronal-type cells in the fetal hippocampus. Since we have shown previously that choline metabolites are decreased in brains of adult Mthfr−/− mice, and that this decrease can be ameliorated by betaine supplementation [10], it is likely that choline or betaine depletion may have played a role in the structural brain defects in the Mthfr−/− pups.

Conclusions

In summary, our findings demonstrate the importance of epigenetic factors, such as the intake of preformed methyl groups, for a favourable clinical outcome in severe MTHFR deficiency in mice. Very few humans with severe MTHFR deficiency have been identified prenatally; these were siblings of previously identified index patients. One such published case is remarkable in demonstrating that immediate postnatal betaine treatment may prevent symptoms until adulthood ([40]; F. Skovby, personal communication). Our results in mice emphasize the importance of early (prenatal) betaine supplementation in homocystinuria and suggest that newborn screening for hyperhomocysteinaemia would allow for the immediate therapeutic use of betaine to improve the clinical course.