Altered Folate Homeostasis in Children with Down syndrome: A Potential Basis for Enhanced Methotrexate Toxicity

Abstract

Methotrexate (MTX) is used to treat autoimmune and oncologic diseases in children with Down syndrome (DS). However, increased MTX-related toxicity is reported in this population. We evaluated differences in the concentrations and distribution of erythrocyte folates in children with DS as a potential basis for this enhanced toxicity.

Down syndrome, or trisomy of chromosome 21, is the most frequent genetic cause of intellectual disability and occurs at an estimated incidence of 14.5 per 10,000 individuals in the United States (1, 2). Children with DS are known to have an increased incidence of oncologic and autoimmune conditions, specifically juvenile idiopathic arthritis (JIA), acute lymphoblastic and acute myeloid leukemia, hypothyroidism, and celiac disease, among other conditions that often require chronic drug therapies (3). In particular, inflammatory arthritis, observed at a 3-fold higher incidence in children with DS compared with the general population (4) can be a diagnostic and therapeutic challenge in children with DS. Children with DS who develop arthritis have a delay in diagnosis of their arthritis, coupled with a more severe polyarticular course, which puts them at increased risk for long-term joint damage and physical disability (5, 6). Moreover, therapeutic challenges unique to children with DS compound this risk, including enhanced sensitivity to the toxic effects of commonly used drugs to treat juvenile arthritis, including methotrexate) (5).

As a potent anti-folate drug, MTX has several enzymatic targets in the folate pathway, leading to the accumulation of anti-inflammatory agents (such as adenosine) and the interruption of purine and pyrimidine synthesis, which are pivotal for DNA synthesis (7, 8). Reduction in serum and erythrocyte folate are observed after MTX administration (9,10), and folic acid supplementation is thus often used to ameliorate common clinical side effects to the drug (11).

Folates are water-soluble B vitamins consisting of a reduced pteroic acid ring system conjugated to a polyglutamate chain. Ingested folates are deglutamated to their folate monoglutamte (folateGlu₁) form in the gastrointestinal tract to facilitate absorption (12). Circulating folateGlu1 is transported into the cell where glutamyl chain accumulation (folate polyglutamation) occurs (folateGlu_n). Polyglutamation is an important process as it facilitates intracellular retention (13) and leads to bioactivation, by enhancing the interaction between folate and folate metabolizing enzymes (7). Once within the cell, folates convert between intermediates that differ by oxidation state and one-carbon substitution to participate in downstream biochemical reactions. The distributions of folate intermediates within the cell maintain an important balance of one carbon derivatives for use in DNA synthesis and methylation. A deficiency or disruption in cellular folate distribution results in aberrant DNA methylation, genome instability, and abnormal gene expression. There are data to suggest there is cellular adaptation to folate deficiency by an increased expression of FPGS (14) and higher concentrations of long chain folateGlu_n in in vitro and animal models (15, 9).

Individuals with DS have a well-documented clinical sensitivity to high-dose MTX for the treatment of leukemia (16) as well as to low-dose MTX for the treatment of arthritis (5). These sensitivities may be related to unique metabolic alterations associated with the extra chromosome 21 that may influenceMTX biotransformation. For example, several enzymes in the folate pathway are encoded on chromosome 21, some highly suspected to affect MTX metabolism. SLC19A1 is a transmembrane protein that transports intracellular reduced folates as well as MTX (17, 18). It has been documented that in DS patients, hyperdiploid acute lymphoblastic leukemia (ALL) cells with extra copies of chromosome 21 generate higher levels of active MTX polyglutamates, proposed to be due to increased intracellular transport of MTX via SLC19A1 (18). Additional folate enzymes encoded on chromosome 21 include cystathionine beta synthase (CBS) (19) whose overexpression in DS results in altered homocysteine metabolism and a subsequent “folate trap” phenomenon, and phosphoribosylglycinamide transformylase (GART), a multifunctional protein integral for purine synthesis (20).

Although serum and total concentrations of erythrocyte folates have been characterized in subjects with DS, prior research has not characterized the distribution of erythrocyte folate intermediates and the polyglutamated states of erythrocyte folate within this population of patients. We hypothesize that differences in intracellular folate homeostasis may contribute to the MTX sensitivity observed clinically in patients with DS. Thus, in this pilot study we compared intracellular folate isoforms and their polyglutamated states in a sample of pediatric patients who have DS with a previously-collected cohort of pediatric patients with JIA, with neither group receiving MTX therapy at the time of sampling (21). By establishing improved methods for folate analysis we aimed to clarify our knowledge of folate metabolism and folate intermediates in patients with DS to enable future larger studies to quantify the effects of folate deprivation or supplementation in children with DS.

Methods

Fifteen pediatric patients were recruited as a convenience sample from the multidisciplinary Down syndrome clinic at Children’s Mercy, Kansas City. None of the patients were on MTX therapy at the time of sampling. If clinical laboratory tests were required at the visit, parental written informed consent was obtained to draw an additional blood sample for research purposes. Age in months and folate supplementation/multivitamin (MVI) use were recorded for the DS cohort at the time of the sample collection. Erythrocyte folate concentrations in patients with DS were compared with a cross-sectional cohort of juvenile arthritis patients reported previously (21). Work described was carried out in accordance with the Code of Ethics of the World Medical Association and written informed consent was obtained from the parents or guardians of the children who served as subjects and, when appropriate, assent from the subjects themselves. Studies were conducted under the approval of the Children’s Mercy, Kansas City Institutional Review Board.

Preparation of erythrocyte lysates and folate analyses:

During phlebotomy for routine laboratory monitoring, a 6 mL sample of blood for folate assessment was collected and processed immediately upon receipt. During processing, two 300 μL aliquots of blood were immediately frozen at −80°C, and the remaining sample was centrifuged at low speed (600g) to pellet the erythrocytes. The plasma and buffy coat were removed and the remaining erythrocytes were suspended 4:1 in Phosphate Buffered Saline, and respun. The wash procedure was repeated two more times. Packed erythrocytes and plasma were divided into aliquots and stored at −80°C. Concentrations of 5-methyl-tetrahydrofolate (5-CH₃-THF) and 5,10-methenyl- tetrahydrofolate (5,10-methenyl-THF) in whole blood and plasma were determined by LC-MS/MS utilizing the endogenous plasma enzyme gamma-glutamylhydrolase to deglutamate the folates into their monoglutamate form as described by Huang et al. (22) Commercially available, isotopically labeled folate monoglutamates were used as internal standards. The folate concentration in erythrocytes was calculated from whole blood folate values, using hematocrit levels and correcting for plasma folate. As a complete set of standards for 5-CH₃-THFGlu_n is not commercially available, we measured the relative proportion of each polyglutamate 5-CH₃-THFGlu₃ to 5-CH₃-THFGlu₁₀. Relative concentrations of 5-CH₃-THF polyglutamates (5-CH₃-THFGlu_n) were measured in erythrocyte lysates using an ion-pair reversed phase chromatography (IPRPC) method modified from our previously published method (23).

Statistical Analysis:

Descriptive statistics were employed to characterize variability in the concentrations of each individual folate isoform and polyglutamate pattern. Univariate analyses were performed between clinical variables, including age and folate supplementation, and intracellular erythrocyte folate isoforms 5-CH₃-THF and 5,10-methenyl-THF, using two-sample t-test, analysis of variance, or simple linear regression as appropriate. For data that were not normally distributed (ie, 5,10 methenyl-THF concentrations), Wilcoxon rank-sum analysis or Spearman correlation analysis was used as appropriate. In the multivariable regression analyses, we chose age, folate supplementation, and presence of DS as explanatory variables for variation in total erythrocyte folate levels. Analyses were performed using JMP 11 (Cary, NC).

Results

For the 15 children with DS recruited to this study, median (IQR) age was 52 (24, 63) months. Six (40%) were on a MVI that contained folate, with doses that ranged from 200–400 mcg/day.

Samples from the previously reported cohort of 93 juvenile arthritis patients had been collected in a cross-sectional manner and included all children who presented with arthritis not currently receiving anti-folate medications such as methotrexate (MTX) (21). None of the patients carried the diagnosis of DS. The median (IQR) age of this cohort was 144 (100.5, 192) months. Twenty-five (27%) were on a MVI that contained folate, with doses that ranged from 200–400 mcg/day.

Mean total erythrocyte folate concentrations (ie, sum of erythrocyte 5-CH₃-THF and 5,10-methenyl-THF) in patients with DS were 25% lower than the JIA comparator group (p=0.03) (Table). The lower erythrocyte folate concentrations represented disproportionately lower erythrocyte 5,10-methenyl-THF concentrations (P = .005) compared with erythrocyte 5-CH₃-THF concentrations (p=0.05), and represented 65% and 23% reductions in the DS population, respectively.

Table 1.

Erythrocyte folate levels in subjects with DS compared with a cohort of 93 JIA patients.

Folate	DS	JIA	Difference (%)	p-value
5-CH3-THF, Mean (±SD)	790.1 (±404.2)	1028.9 (±487.6)	−23%	0.05
5,10-methenyl-THF, Median (IQR)	23.1 (15.6, 66.4)	66.9 (36.1, 102.0)	−65%	0.005
Total folate, Mean (±SD)	845.5 (±403.5)	1121.6 (±512.3)	−25%	0.03

Because the age of the subjects was quite different between groups, we assessed the impact of age upon erythrocyte folate concentrations (5-CH₃-THF, 5,10-methenyl-THF, and the total erythrocyte folate concentrations) and no significant association was observed statistically between age and cellular folate concentrations within the group as a whole nor specifically in the cohort of patients with DS. We also assessed if the consumption of MVI containing folic acid would affect erythrocyte folate concentrations, and in the combined group, there was a non-statistically-significant trend towards an increase in erythrocyte folate concentrations in patients who concurrently took MVI containing folic acid (p=0.07).

In light of the differences in baseline characteristics between groups and existing data to suggest that intracellular folate concentrations may decrease with age (24), we adjusted for the presence of DS, age, and folate supplementation in a multivariable linear regression model using total erythrocyte folate concentrations as the dependent variable. Controlling for these additional variables in the model, the presence of DS (p=0.02) and the consumption of MVI with folate supplementation (p=0.05) remained statistically significantly associated with total erythrocyte folate concentrations.

In previous work, regression analysis and hierarchical clustering demonstrated an internal correlation among folate polyglutamates that revealed long chain and short chain folate polygluatamates clustered together and prompted us to evaluate whether the observed reductions in erythrocyte folates in children with DS was associated in a change in the folate polyglutamation profile (21). Examination of the relative contributions of individual 5-CH₃-THFGlu_n revealed a proportional shift towards longer chain 5-CH₃-THF polyglutamates in subjects with DS, resulting from a predominantly lower relative proportion of short chain 5-CH₃-THFGlu_3–6 (Figure). As a result, subjects with DS had a decrease in mean % (±SD) short chain 5-CH₃-THFGlu_3–6 (84.2% [± 3.4] vs. 86.5% [± 2.8], p= 0.03) and a corresponding increase in longer chain 5-CH₃-THFGlu_7–10 (13.5% [± 2.8] vs. 15.8% [± 3.4], p= 0.03).

Figure 1.

Erythrocyte 5-CH₃-THF polyglutamate distribution in patients with DS. Concentrations of 5-CH₃-THF-Glu_3–10 in circulating erythrocytes from patients with DS are presented and compared with the reference JIA cohort. Data represents the mean concentration of 5-CH3-THFGlu_n in nmol/L based on the number of attached glutamic acid residues (n). Error bars represent the standard error of the mean. Statistical significance was evaluated by Student’s t-test analysis (*,p≤0.05; ***,p≤0.0001).

Discussion

Patients with DS have often not been included in clinical trials, yet they have an increased risk of comorbid conditions that require management with chronic pharmacologic therapies.

With respect to methotrexate, one main target of the drug is the endogenous folate pathway, and in DS specifically, this pathway has been investigated on several levels. Alterations in maternal folate homeostasis have been interrogated as potential risk factors for fetal chromosome 21 nondisjunction with inconsistent results (25) and folate supplementation has been investigated as a therapeutic intervention to improve cognition (26).

In this study, we demonstrated the ability to quantify specific cellular redox states of intracellular folate and the relative polyglutamation status of 5-CH₃-THF, the most abundant reduced form. We are able to illustrate the folate depleted state of children with DS, manifested as both a decline in intracellular folate isoforms and a shift in polyglutamation status in favor of longer chain polyglutamates. This may indicate a compensatory effect to retain cellular folate, which plays an important role as a one-carbon donor for many critical downstream cellular activities. The relative greater decline in 5,10-methenyl-THF in our DS cohort is potentially meaningful as this folate isoform is a critical substrate for enzymes involved in de novo purine and pyrimidine biosynthesis. The introduction of MTX could result in further folate depletion in this key folate isoform, which could explain the extreme drug toxicity in children with DS treated with MTX. In fact, a maximal rescue effect in vitro has been shown in trisomy 21 fibroblast cell lines after supplementation with a combination of reduced folate isoforms compared with the oxidized isoform, folic acid. (27) Quantifying this degree of detail in folate homeostasis may prove useful as a pharmacodynamic tool to understand better MTX disposition and response at the level of an individual patient. Because of its widespread use, future studies have the potential to investigate the alterations of erythrocyte folate concentrations subsequent to MTX in several other conditions that use the drug chronically including psoriasis, inflammatory bowel disease, and various additional rheumatologic and oncologic disorders.

This pilot study describes the detailed folate phenotype in a small population of children with DS, and shows a potential role for the folate pathway as a pharmacodynamic focus in future studies. In DS specifically, the endogenous folate pathway is particularly vulnerable to metabolic alterations, and cellular folate phenotyping offers a quantifiable folate signature that could potentially be a biomarker of MTX response or toxicity. A detailed folate phenotype can be quantified and may be a starting point for future research to distiguish the effects of folate deprivation or supplementation in children with Down syndrome.