Thymidine and deoxyuridine accumulate in tissues of patients with mitochondrial neurogastrointestinal encephalomyopathy (MNGIE)

Abstract

Mitochondrial neurogastrointestinal encephalomyopathy (MNGIE) is an autosomal recessive disease due to ECGF1 gene mutations causing thymidine phosphorylase (TP) deficiency. Analysis of post-mortem samples of five MNGIE patients and two controls, revealed TP activity in all control tissues, but not in MNGIE samples. Converse to TP activity, thymidine and deoxyuridine were absent in control samples, but present in all tissues of MNGIE patients. Concentrations of both nucleosides in the tissues were generally higher than those observed in plasma of MNGIE patients. Our observations indicate that in the absence of TP activity, tissues accumulate nucleosides, which are excreted into plasma.

1. Introduction

Mitochondrial neurogastrointestinal encephalomyopathy (MNGIE) is a clinically recognizable autosomal recessive disease characterized by ptosis and ophthalmoparesis, gastrointestinal dysmotility, peripheral neuropathy, leukoencephalopathy, and mitochondrial alterations. MNGIE is caused by mutations in the ECGF1 gene encoding thymidine phosphorylase (TP)[1-4]. TP is a cytosolic enzyme that catalyzes the reversible phosphorolysis of deoxythymidine (dThd) and deoxyuridine (dUrd) to the corresponding bases, thymine or uracil, and 2-deoxyribose 1-phosphate[5,6]. Typically, MNGIE patients have severely reduced TP activity in buffy coat (<10% of the control mean) causing dramatic elevations of plasma dThd (8.6±3.4μM, mean±standard deviation; normal <0.05) and dUrd (14.2±4.4; normal <0.05)[7-10]. Thus, in this disease, the association of ECGF1 mutations with altered pyrimidine nucleoside metabolism and a distinct clinical phenotype indicate a causal relationship; however, its pathogenesis is not fully understood.

One of the hallmarks of MNGIE is somatic alteration of mitochondrial DNA (mtDNA), namely, depletion, multiple deletions, and site-specific point mutations[1,11-13], which cause variable defects of the mitochondrial respiratory chain[1,14]. We have hypothesized that, in MNGIE, elevated levels of dThd and dUrd cause nucleotide pool imbalances that, in turn, cause mtDNA instability[3,10]. In vitro studies have confirmed that high concentrations of extracellular dThd (10-50 μM) alter nucleotide pools both in cytosol and in mitochondria, and unbalanced nucleotide pools, in turn, cause alterations of mtDNA[15,16].

Curiously, deficiency of TP, a vital enzyme expressed in most tissues, preferentially affects extraocular muscles, gastrointestinal system, peripheral nerves, and central nervous system white matter. To better understand the organ-specificity of MNGIE, we have measured TP activity and levels of deoxynucleosides in post-mortem tissues from MNGIE patients and controls.

2. Methods and Materials

2.1 Patient and control samples

We analyzed several post mortem tissues from five MNGIE patients previously reported as MN2-01, MN3-01, MN4-02, MN7-02, and MN29-01[2,17]. All of the patients manifested the typical clinical features of MNGIE and their diagnoses were confirmed by measurement of buffy coat TP activity and identification of ECGF1 mutations (Table 1). The ages at death were 29-39 years old. Tissues were collected 11-24 hours post–mortem and stored at −80°C. Control autopsy tissues from age-matched subjects with similar postmortem intervals were obtained from the Brain and Tissue Bank for Developmental Disorders (University of Maryland). Autopsy and biopsy tissues were collected and analyzed under Columbia University Institutional Review Board approved protocols.

Table 1.

Molecular genetic and biochemical features in blood of MNGIE patients whose post-mortem tissues were analyzed in this report. Patients are numbered according to previous publications[2,17]. TP activity in nmoles thymine formed/h/mg protein. UND: undetectable.

Patient	Age at death	Gender	Age at onset	TP mutations	Buffy coat TP	Plasma dThd (μM)	Plasma dUrd (μM)
MN2-01	33	M	18	IVS4 +2 T>C E289A	UND	Not determined	Not determined
MN3-01	37	M	18	G145R G145R	UND	Not determined	Not determined
MN4-02	29	F	15	K222S frameshift	UND	6.4	Not determined
MN7-02	39	F	26	E289A IVS9 −1 G>C	UND	5.7	Not determined
MN29-01	32	M	Infancy	G153S G153S	UND	5.5	Not determined

2.2 Measurements of nucleosides

dThd and dUrd levels were assessed by a gradient-elution HPLC method as previously described [8]. Between 15-280 mg of tissue was homogenized in lysis buffer (50 mM Tris-HCL, pH 7.2, containing 1% Triton X-100, 2mM phenylmethylsulfonyl fluoride, and 0.02% 2-mercaptoethanol), sonicated, and centrifugated at 2,500g for 10 minutes. Supernatants were deproteinized with 0.5 M perchloric acid and injected into an Alliance high-performance liquid chromatography (HPLC) apparatus (Waters Corporation) with an Alltima C₁₈ NUC reversed–phase column (Alltech) at a constant flow rate of 1.5 ml/min using two buffers: eluent A (20 mM potassium-phosphate, pH 5.6) and eluent B (20mM potassium phosphate-60% methanol, pH 5.6). Samples were eluted over 110 min with a gradient as follows: 0-5 minutes, 100% eluent A; 5-92 minutes, 100% to 0% eluent A; 92-93 minutes, 0% to 100% eluent A; and 93-110 minutes, 100% eluent A. Absorbance of the eluates was monitored at 267nm and dThd and dUrd peaks were quantitated by comparing their peak areas with a calibration curve obtained with aqueous standards. For definitive identification of dThd and dUrd peaks for each sample, we used a second aliquot treated with excess of purified E. coli TP (Sigma) to specifically eliminate dThd and dUrd. The detection limit of this method was 0.05 μmol/L for both deoxynucleosides.

2.3 Thymidine phosphorylase activity

TP activity was measured using an HPLC detection method as described[18]. Between 30-120 mg of frozen tissue was homogeneized in lysis buffer and briefly sonicated. Samples were centrifugated at 13,000 g × 30 min at 4°C and 100 μg of protein were incubated with 10 mM dThd in reaction buffer (0.1 M Tris-arsenate, pH 6.5) , at 37°C for 1 h, in a total volume of 100 μl. For each sample a blank was also processed to which dThd was not added. The reaction was stopped by the addition of 1 ml 0.55 M perchloric acid in all the samples and 10 mM dThd in all the blanks. The supernatants were injected into the HPLC apparatus described above and eluted, over 20 min, at a constant flow rate of 1.5 ml/min, with 20 mM potassium phosphate-10% methanol, pH 5.6. The absorbance of the eluate was monitored at 267 nm and thymine peaks were identified and quantified using a calibration curve. The protein content of the supernatants was determined using the bicinchoninic acid method[19]. Due to limited amounts of post-mortem tissues, TP activity was measured in tissues from only two patients.

3. Results

3.1 Levels of thymidine and deoxyuridine are elevated in all tissues of MNGIE patients

We measured levels of nucleosides in five clinically affected tissues (small intestine, peripheral nerve, muscle, occipital white matter and liver [MN2-01 had cirrhosis, MN4-02 had liver steatosis]) and two unaffected tissues (kidney and heart) in patients and controls (Table 2). In MNGIE patients, all tissues showed accumulation of dThd and dUrd whereas control tissues had no detectable pyrimidine nucleosides (Table 2). The highest concentrations of dThd and dUrd were detected in four of the five affected tissues: peripheral nerve, small intestine, occipital white matter, and liver while kidney and heart, two unaffected organs and skeletal muscle, a less severely affected tissue, had lower levels of nucleosides (Table 2).

Table 2.

Levels of deoxythymidine (dThd) and deoxyuridine (dUrd) in tissues of MNGIE patients and controls.

Tissues		dThd				dUrd				dThd and dUrd
		pmoles / mg tissue		nmoles / g prot		pmoles / mg tissue		nmoles / g prot
	Patients	Mean	Range	Mean	Range	Mean	Range	Mean	Range	Controls
Small intestine	n=4	21	1.9 - 40	471	38 - 926	18	9.2 - 27	379	136 - 630	n=3	UND
Occipital white matter	n=4	17	10 - 33	469	93 - 889	22	10 - 41	498	217 - 694	n=4	UND
Peripheral nerve	n=5	14	5.0 - 28	507	177 - 1532	18	6.5 - 33	467	351 - 724	n=3	UND
Muscle	n=4	5.5	2.0 - 14	102	17 - 201	8.0	5.2 - 16	109	44 - 223	n=4	UND
Muscle autopsy*	n=1	14	-	201	-	16	-	223	-	-	-
Muscle biopsy*	n=1	4.8	-	49	-	5.4	-	55	-	-	-
Liver	n=4	33	9.7 - 80	384	61 - 943	24	11 - 48	264	88 - 566	n=4	UND
Kidney	n=5	18	4.0 - 30	243	64 - 396	17	5.2 - 35	252	102 -464	n=4	UND
Heart	n=5	8.3	3.2 - 13	116	35 - 278	9.6	3.0 - 17	138	32 - 374	n=2	UND

^*Samples from patient MN7-2.

When we compared the concentrations of nucleosides in biopsy and autopsy muscle of the same MNGIE patient, we found lower concentrations in the biopsy tissue (dThd 49 nmoles/g prot, dUrd 55 nmoles/g prot) than in the post-mortem (dThd 201, dUrd 223). Nevertheless, the levels in biopsy samples were within the range of the values found in all the autopsy muscles (dThd range 17-201, dUrd range 44-223) (Table 2). The difference in nucleoside levels in biopsy and post-mortem samples may be due to random fluctuations, increased levels in the terminal phase of the illness, or post-mortem accumulations of nucleosides.

3.2 Thymidine phosphorylase activity is deficient in all tissues of MNGIE patients

TP activity was not detectable in all the autopsy tissues from two individuals with MNGIE (Table 3), confirming the drastic reduction of the enzymatic activity found in buffy coats of patients[2,3]. TP activity was detected at low levels in control muscle, as previously reported[20] and also in kidney a organ reported to lack this enzyme[21].

Table 3.

TP activity in nmoles thymine formed · h⁻¹· (mg protein)⁻¹, in autopsy and biopsy tissues from MNGIE patients and controls. UND=undetectable

	patients	TP		controls	Mean TP (range)
Small intestine	n=2	UND		n=2	1,121 (911 , 1331)
Occipital white matter	n=2	UND		n=2	16 (16 , 16)
Peripheral nerve	n=2	UND		n=2	34 (33 , 35)
Muscle	n=2	UND		n=3	55 (51 - 60)
Muscle biopsy	n=1	UND		n=3	21 (16 - 27)
Liver	n=2	UND		n=2	1,817 (1642 , 1992)
Kidney	n=2	UND		n=2	234 (226 , 242)
Heart	n=2	UND		n=2	40 (36 , 44)

4. Discussion

MNGIE is a rare autosomal recessive disorder caused by mutations in the nuclear gene encoding TP[1-3]. Absence of TP activity causes a generalized metabolic defect characterized by a dramatic accumulation of TP substrates, dThd and dUrd, in plasma of MNGIE patients[8-10]. The absence of these nucleosides in plasma of healthy controls and in asymptomatic TP mutation carriers[8-10] suggests that the disease develops due to the toxic effect of dThd and dUrd overload on mtDNA, as in vitro experiments have demonstrated that high concentrations of these nucleosides cause depletion and multiple deletions of mtDNA[15,16].

To date, accumulations of dThd and dUrd in MNGIE patients have been observed in plasma and urine of MNGIE patients[7-10]. In one deceased patient, TP dysfunction was confirmed through demonstration of elevated dThd and dUrd in skeletal muscle, the only available tissue[22]. Here, we have measured nucleosides in clinically affected and unaffected tissues from autopsies of five MNGIE patients; levels of dThd and dUrd were dramatically elevated in all tissues, indicating that, in MNGIE, nucleoside imbalance is ubiquitous.

In five of the seven tissues analyzed in patients, concentrations of these nucleosides were higher in tissues than in blood consistent with the notion that thymidine and deoxyuridine are cellular by-products, which, in the absence of TP catabolism, are excreted into extracellular space. We reported mean plasma concentrations of dThd and dUrd in MNGIE patients were 8.6 and 14.2 μmoles/l[8] and observed mean plasma dThd concentration of 5.9 μM in the subgroup of patients in this study, while, in contrast, most of the tissues studied here showed average content of dThd and dUrd higher than 14 pmoles/mg tissue (units similar to μmoles/l, assuming the density of the tissues close to 1 g/ml) (Table 2). This difference in nucleoside levels may be underestimated, because recovery of nucleosides from homogenized tissues is likely to be less than 100%, while nucleosides are measured directly in plasma and therefore completely recovered. We have observed up to 2-fold fluctuations in plasma thymidine levels in clinically stable MNGIE patients indicating that nucleoside levels vary significantly. Furthermore, quantitation of nucleosides in autopsy tissues may be inaccurate due to post-mortem deoxynucleotide degradation that might have artifactually exaggerated the increase of nucleoside levels in these samples and contributed to the wide variation in the results. In fact, when muscle biopsy and post-mortem muscle samples from a single patient were analyzed, both dThd and dUrd levels were 4-fold higher in the autopsy, compared to the biopsy suggesting that nucleoside measurements in post-mortem tissues may overestimate levels. Nevertheless, this interpretation is based on studies of only one patient and is therefore tenuous. This difference between the biopsy and autopsy analysis could be due to intra-individual variation. Analyses of additional muscle biopsies, which is not possible in this study due to limited tissue availability, would be important.

Levels of dThd and dUrd in different tissues did not correlate with the clinical involvement of organs. For example, skeletal muscle, which is affected clinically and histologically in most MNGIE patients, showed the lowest levels of nucleosides and below levels found in kidney, which is unaffected. Liver, which is inconsistently affected in patients, had dThd and dUrd levels similar to small intestine, peripheral nerve and white matter, which are invariably affected in MNGIE. In line with these observations, previous reports showed no correlation between severity of the symptoms and levels of circulating dThd and dUrd, when above 3 μM[2,8]. Only patients with less deleterious mutations in the ECGF1 gene, which cause partial TP dysfunction (∼10-15% of control mean), accumulate only modestly increased plasma dThd and dUrd levels (0.5 – 1.5 μM), resulting in a later onset of the disease[18].

In MNGIE, post-mitotic cells appear to be preferentially affected. For example, enteric smooth muscle degeneration causes gastrointestinal dysmotility and skeletal muscle involvement causes proximal limb weakness[1,2,23]. The tissue-specificity in MNGIE may be attributable to the dependence of mtDNA replication on the mitochondrial nucleoside salvage pathway in quiescent cells[15,24,25], energy requirements of specific cells, or a combination of both factors.

Biochemical analysis of MNGIE tissues confirmed the absolute or virtually complete absence of TP activity found in buffy coat samples. TP activity in these organs was analyzed in samples from several controls, resulting in variable activities for different tissues. In contrast to several reports that failed to find TP protein in skeletal muscle and kidney in humans[20,21,26,27], we could clearly detect and quantitate TP activities in these tissues. The discrepancies between our detection of TP activity and prior immunohistochemical studies may be related to the lower sensitivity of antibodies compared to biochemical assays. It is unlikely that uridine phosphorylase (UP) could have interfered our enzymatic analysis, because human UP poorly catabolizes thymidine, the substrate used in our TP assay[28] and, in MNGIE patients, we were unable to detect TP activity in liver and kidney, tissues that express UP[29].

We could not identify correlations between the magnitude of normal TP activities and elevations of nucleosides in tissues of MNGIE patients. For example, normal small intestine has very high TP activity (1121 nmoles thymine formed/h/mg prot) while loss of TP function in this tissue caused moderate increases in dThd (21 pmol/mg tissue) and dUrd (18 pmol/mg tissue) which are similar to levels in peripheral nerve of MNGIE patients, dThd (14 pmol/mg tissue) and dUrd (18 pmol/mg tissue), despite the relatively low TP activity (34 nmol thymine formed/h/mg prot) in normal peripheral nerve (Table 2). This lack of correlation suggests that the function of high TP activity in organs is not merely catabolism of locally produced nucleosides, but rather systemic clearance of dThd and dUrd. In support of this notion, we and others have observed high TP activities in highly vascularized tissues, liver, small intestine, kidney and placenta, which are exposed to large volumes of blood[30]; therefore, the role of TP, at least in some tissues, may be systemic clearance of dThd and dUrd.

Pharmacokinetics studies indicate that in humans, dThd is distributed in the total body water volume[31]. The rapid equilibration of dThd in tissues may be facilitated by nucleoside transporters, which have been identified in plasma membrane of many cells types[32]. Furthermore, studies of cultured cells have demonstrated a rapid passage of dThd across both plasma and mitochondrial membranes[33], which may be mediated by human equilibrative nucleoside transporter 1 (hENT1), which is present in plasma membranes of liver, spleen, adipose cells, brain, colon and other tissues, and has also been identified in the inner membrane of mitochondria [34,35]. Consistent with these findings, our results show that, in MNGIE, dysfunction of TP produces a generalized increase of its substrates, dThd and dUrd, in all organs. Further studies are necessary to determine the intramitochondrial levels of these nucleosides and their potential effects on mitochondrial nucleotide pools and mtDNA maintenance.

List of Abbreviations

dUrd

deoxyuridine

MNGIE

mitochondrial neurogastrointestinal encephalomyopathy

dThd

thymidine

TP

thymidine phosphorylase

UP

uridine phosphorylase