There is little information on the prevalence and clinical presentation of inborn errors of purine and pyrimidine (PnP) metabolism in developing countries.1 The diagnosis of PnP disorders remains problematic owing to the low level of awareness of these diverse groups of disorders, the existence of disease variants and non-specific clinical features, the lack of expertise and laboratories to make and confirm the diagnosis, and numerous other competing health priorities in these countries.2,3 Nevertheless, with the reduction in infant mortality in many developing countries, these inherited disorders have increasingly become an important paediatric condition for early diagnosis so that appropriate genetic counselling, treatment and supportive treatment can be provided.1,4 Over 30 enzyme defects of PnP metabolism have been described in the literature to date but only 17 are known to cause human diseases.5 The clinical features of these disorders may be broadly divided into four main groups of presentation: patients with unexplained psychomotor or cognitive retardation (including ophthalmological and neonatal presentations), immunodeficiency or haematological conditions, renal diseases and arthropathy.6
PnP are essential constituents of nucleic acids and they play important roles in the storage of genetic information, gene expression, energy metabolism, cell signalling, regulatory molecules, and as the precursors of many coenzymes.7–9 Any defects in the pathways for the biosynthesis and degradation of PnP nucleotides will lead to a number of diverse clinical features. On the other hand, some of these enzyme defects are relatively benign and non-disease-causing. Some of these disorders can be easily diagnosed through analysing these abnormal metabolites or by identifying the disease-specific markers present in the urine by reversed-phase high-performance liquid chromatography (RP-HPLC).
The aim of the study was to establish a simple and cost-effective method for the diagnosis of inborn errors of PnP metabolism among Malaysian children who are suspected to have these disorders. The secondary aim was to study the epidemiology and biochemical phenotype in our patients. In addition, the age-related reference ranges and urinary uric acid–creatinine ratio in normal children and adults was determined to aid in diagnostic interpretation related to this group of disorders.
Materials and methods
To estimate the incidence rate of inborn errors of PnP metabolism in Malaysia, children who were referred from paediatricians nationwide for suspected inborn errors of metabolism (IEM) were selected in this study. Criteria of selection were infants or children who presented with non-specific neurological symptoms including epilepsy, psychomotor retardation, hypotonia, mental and growth retardation, microcephaly, muscle weakness, autism, aggressive or self-mutilation behaviour, immunodeficiency states, rheumatological conditions, renal stones and chronic renal failure. Known cases of IEM other than PnP disorders were excluded from this study. Parents of these children were informed about the study and a urine sample was collected after obtaining parental written consent. A total of 1499 urine samples were collected and stored at −20°C until analysis. To establish the normal reference ranges for Malaysian subjects, urine samples were collected from 556 healthy Malaysians with ages ranging from 0–20 years to > 21 years. All urine samples obtained were subjected to a urine dipstick examination to exclude contamination, and contaminated samples were excluded from the study. This study was approved by the Ethics Committee of the Ministry of Health, Kuala Lumpur, Malaysia.
The standard mixtures contained orotic acid, orotidine, uracil, pseudouridine, creatinine, uric acid, xanthine, hypoxanthine, thymine, uridine, adenine, inosine, guanosine, deoxyinosine, deoxyguanosine, thymidine, adenosine and deoxyadenosine, while 5-bromouridine was used as an internal standard. All standards used were high-performance liquid chromatography (HPLC) grade and obtained from Sigma Aldrich (Sigma Aldrich Inc., St Louis MO, USA). The above standards were weighed and dissolved in 5 ml of saturated Li2CO3, transferred into a 50 ml volumetric flask and filtered Milli-Q water was added to the mark of the flask. The 5 mmol/l standard mixtures were further diluted to 1 mmol/l with filtered Milli-Q water. The internal standard (5-bromouridine) was prepared at 1 mmol/l stock standard and 100 μl was added to the sample as well as to the external standards.
The analytical method was adopted from Vidotto et al.,10 and for this study, we used the rapid resolution HPLC system (Agilent Technologies 1200 series) equipped with binary pump, robotic auto-sampler, reverse phase column (Supelcosil LC18–5; Supelco-analytical) with multi-wavelength diode array detector. The system is fully automated and controlled by an HP desktop computer with Agilent Chemstation software (version A.06.01). In preparation for the chromatography, two high-aqueous elution buffers were used for separation of the compounds. Solvent A was 25 mmol/l KH2PO4 and solvent B was a mixture of solvent A and methanol (75 : 25 by volume). Both were adjusted to pH 4.6 using 5 mM glacial acetic acid, filtered and degassed before use. A flow rate of 1.1 ml/minute was applied. The following elution gradient was used: 0–10 minutes, 100% A; 10–16 minutes, 60% A; 16–26 minutes, 0% A; 26–28 minutes, 100% A. Between runs, the system was re-equilibrated with 100% eluent A for 2 minutes. The analysis time for one sample was 28 minutes. Urine samples were determined for the creatinine by the modified Jaffe method (alkaline-picrate), while uric acid concentration was measured by the enzymatic colorimetric (uricase) method using the Modular P chemistry analyzer (Roche Diagnostics GmbH, Mannheim, Germany) and Roche reagents (Roche Diagnostics GmbH, Mannheim, Germany). Prior to the analysis of PnP metabolites, the urine samples were gently warmed at 56°C for 30 minutes and adjusted to pH 6–7 with 10 mol/l NH4OH. The urine sample was then diluted with solvent A to give a final concentration of 1 mmol/l of creatinine. The injection samples mixtures consisted of 300 μl of buffer A, 100 μl of urine sample and 100 μl of internal standard which were filtered through a 0.2 μm Millipore filter before injecting 10 μl of filtered urine into the HPLC system.
The analytical method was validated according to the Eurochem Group (1988),11 which includes linearity, limit of detection (LOD), limit of quantitation (LOQ), accuracy and precision (intraday and interday variation). The linearity and detection limits were determined by seven repeat injections of three different concentrations of calibration mixture. The calibration curves were generated from the external standard peak area versus standard concentration. The LOD was calculated from the mean of the blank sample plus three times the standard deviation (SD) obtained on the blank sample. LOQ was established as the mean of the spiked sample plus 10 times the SD obtained from the spiked sample. Intraday precision (repeatability) was determined by analysing a blank urine sample plus three urine samples spiked with low (31.25 μmol/l), medium (125 μmol/l) and high (500 μmol/l) concentrations of standard mixtures which were analysed for 10 times within a day. For determination of interday precision (reproducibility), the same urine samples were analysed on 10 different days. To evaluate the recovery, 10 normal control urine samples were analysed in duplicate before and after being spiked with low (40 μmol/l), medium (160 μmol/l) and high (400 μmol/l) concentrations of standard mixtures.
Results are presented as arithmetic means and SDs. Paired t-test was used for statistical comparison of differences in the means. For comparison among several groups, ANOVA (analysis of variance) was used with subsequent post hoc least square difference (LSD) test; P < 0.05 was considered as statistically significant. Deviations from Gaussian distribution were tested using the Kolmogrov–Smirov test. All the statistical testing was analysed using SPSS software [V11.5, Institute for Medical Research (IMR), Malaysia].
Performance of reversed-phase high-performance liquid chromatography
The method established was able to separate up to 18 PnP metabolites in a single analytical run time of 28 minutes. The chromatogram illustrates good resolution and well-separated standard mixtures of PnP metabolites (Figure 1). The retention time of the standard mixtures fell within the SD (mean ± SD < 1.0) and a coefficient variation (CV) of < 2% was observed for all PnP standard mixtures, except for adenine which had a CV of 2.17% (Table 1).
CV, coefficient variation.
Detected at 254 nm wavelength.
Standard, retention time was determined from a mean of 10 injections.
The linearity ranges were found to be linear up to 2000 μmol/l with correlation coefficients (r2) > 0.999, except for orotidine, uridine, pseudouridine and thymine which had a linearity range of 500–1000 μmol/l. The LOD and the LOQ were in the range of 2.18–12.5 μmol/l and 4.28 –21.7 μmol/l, respectively.
Performance specification showed good precision with CV < 2%. The mean CVs for intraday precision of low (urine sample spiked with standard mixtures of 31.25 μmol/l), medium (125 μmol/l) and high (500 μmol/l) concentrations were 1.62%, 0.5% and 2.02%, respectively. The interday mean CVs were 3.92% (low), 3.43% (medium) and 3.53% (high). The overall recovery of PnP bases and nucleosides was found within the range of 99.8–108.4% in all three levels of low, medium and high concentration, as depicted in Table 2.
The reading of the blank sample was substrate from the spiked urines which are not presented in this table.
Table 3 shows the reference ranges of the urinary uric acid – creatinine ratio and PnP metabolites in Malaysian children and adults. Out of the 18 PnP metabolites studied, only seven were detected in the urine of the normal subjects. The metabolites detected were orotic acid, creatinine, uracil, pseudouridine, uric acid, hypoxanthine, xanthine and 7-methylguanine. Orotic acid was detected at very low levels in all age groups and the average mean ranged from 0.9 to 1.17 μmol/mmol of creatinine. Uracil was not detected in newborn infants under 5 days old.
P < 0.05 in comparison between groups and within group. n, total number of samples; D, day; yr, year; UA/creatinine ratio, uric acid : creatinine ratio.
The ratio was determined using modified Jaffe (alkaline picrate) and enzymatic uricase methods.
The study subjects consisted of 554 (37%) children aged 6–10 years old; 300 (20%) aged 1–3 years, and 270 (18%) aged 4–5 years. Other age groups were 11–15 years (12%), < 1 year (8%), > 21 years (3%) and 16–20 years (2%). Of the study subjects 989 (66%) were male and 510 were female; 660 were Malay and 615 were Chinese, while 195 patients were Indian and the remaining 29 patients were of other ethnicities. Only 77 (5.14%) patients came from consanguineous parents. About 87% of the patients presented with neurological symptoms, 7% had renal manifestation, 3% had haematological and/or immunological manifestation, and 3% had joint involvement.
Newly diagnosed cases by reversed-phase high-performance liquid chromatography
We identified 12 positive cases or a 0.8% incidence rate of PnP metabolism disorders over the four-year study period. The 12 identified cases were adenylosuccinate lyase (ADSL) deficiency (one case), molybdenum cofactor deficiency (MoCD; four cases), isolated sulphite oxidase deficiency (ISOD; three cases), thymidine phosphorylase (TP) deficiency (two cases) and dihydropyrimidine dehydrogenase (DPD) deficiency (two cases). Parental consanguinity was noted in one family with TP deficiency. The urine chromatogram of these 12 patients revealed characteristic biochemical markers, which were consistent with PnP metabolism disorders.12,13,14,15 Mutation analysis was performed for patients with ADSL, TP and ISO deficiency to further confirm the diagnosis.
Four other patients showing increased excretion of pyrimidine metabolites, including orotic acid, uracil and uridine, and these patients were later confirmed to have urea cycle defects of argininosuccinate lyase (ASA) and ornithine transcarbamylase (OTC) deficiency.
Adenylosuccinate lyase deficiency
The urine sample from our first reported ADSL-deficient patient revealed the characteristic biochemical markers of succinyladenosine (S-Ado) and succinyl-aminoimidazaole carboximide riboside (SAICAr), which are strongly suggestive of ADSL deficiency. The patient presented with frequent convulsions and severe myoclonic jerk within the first few days of life as well as severe psychomotor retardation. Mutation analysis of the ADSL gene revealed two novel mutations: c.445C>G (p.R 14G) and c.774_776insG (p.A260GfsX24).12
Molybdenum cofactors deficiency and isolated sulphite oxidase deficiency
The first Malaysian MoCD case was reported by Ngu et al. in 2009;13 subsequently we have diagnosed another three patients with MoCD. Biochemical study showed hyperuricaemia, positive sulphite dipstick, elevated urinary sulphocysteine, hypoxanthine and xanthine, which were all very suggestive of MoCD. Mutation analysis is still in progress. Another three patients whose urine samples were tested were found to test positive for sulphite and elevated level of sulphocysteine, but were suspected of having isolated sulphite oxidase deficiency from MoCD due to normal serum and urinary uric acid, hypoxanthine and xanthine. Analysis of the SUOX gene showed a homozygous 1029C>G (Y343X) mutation in one patient, and two heterozygous mutations, 1029C>G (Y343X) and 478G>A (R16oQ), in another patient.14
Thymidine phosphorylase deficiency
The first Malaysian case of TP deficiency was identified in a 27-year-old Bayau male when his urine and plasma chromatograms showed marked elevation of thymidine and 2-deoxyuridine. The biochemical findings were well correlated with suggestive clinical features of classical mitochondrial neurogastrointestinal encephalomyopathy (MNGIE) including ptosis, ophthalmoparesis, cachexia and gastrointestinal dysmotility. Urinary and plasma analysis of his 12-year-old sister also showed elevation of thymidine and 2-deoxyuridine. Direct DNA sequence of the TP (thymidine phosphorylase) gene showed both a novel and a previously described mutation. The novel mutation of ECGF1 was identified in exon 5 (NG_011860.1:g.7387C>T, p.179 Q>X), resulting in a premature stop codon and premature truncation of TP activity. The second mutation occurred at exon 10 (NG_011860.1:g.9279C>T) and is a missense homozygous mutation at protein position 471, which was previously described.15
Dihydropyrimidine dehydrogenase deficiency
We described here an 8-year-old Indian female who presented with developmental delay and microcephaly, and a one-month-old Malay male with early-onset encephalopathy and seizure. The diagnosis of DPD deficiency was based on the characteristic biochemical markers of thymine-uraciluria and 5-OH-methyluracil present in their urine samples. The profile was consistent with gas chromatography/mass spectrometry findings of elevated thymine and uracil with absence of dihydrouracil and dihydrothymine. Mutation analysis is still in progress to confirm the diagnosis.
Urea cycle defects
A two-year-old Pakistani male whose urinary profile showed marked elevation of uracil, orotic acid and uridine was further investigated. A urea cycle defect was suspected due to intermittent episodes of hyperammonaemia during the newborn period. Analysis of his urine and plasma amino acids revealed marked elevation of argininosuccinate acid and a moderate elevation of citrulline. The amino acid profile obtained was consistent with argininosuccinate lyase deficiency.
Three infants with neonatal coma due to severe hyperammonaemia were identified as OTC deficient when their plasma amino acid showed undetectable levels of citrulline and low arginine with marked elevation of glutamine (> 2000 μmol/l). Marked elevation of orotic acid (> 600 μmol/mmol creatinine), uracil and uridine was detected in their urine samples, which was consistent with OTC deficiency.
There are over 30 defects of PnP metabolism which have been described in literature to date. These disorders are characterized by a wide variety of symptoms ranging from relatively benign to severe presentation that may lead to early death. These non-specific symptoms may include mental retardation, developmental delay, seizures, muscular hypotonia, learning difficulties, self-mutilation, renal disease, recurrent infections, unexplained anaemia and sometimes joint involvement. As there is no laboratory in Malaysia that offers diagnostic testing for these, there is a need to establish a method to screen and diagnose these groups of disorders. RP-HPLC remains the method of choice because of its simplicity, cost-effectiveness and capability of comprehensive detection of the whole range of PnP metabolites with good efficiency.
Several techniques have been applied in the past for the measurement of various PnP compounds in urine samples. Earlier studies by Klinbenberg et al.16 and Simmonds et al.17 had successfully used enzymatic spectrophometric and column chromatography to separate some of the purine compounds, but the main limitation of these methods was their inability to separate hypoxanthine and xanthine. Thin-layer chromatography and gas chromatography/mass spectrometry had also been used, but these methods required time-consuming derivatization of the samples.18,19 RP-HPLC has been the most widely used technique for the analysis of PnP metabolites.2,20,21 Duran et al.22 too suggested that RP-HPLC is still the method of choice for detection of PnP metabolites in urine samples. Vidotto et al.10 described RP-HPLC for separation of the PnP metabolites in urine samples. This method is simple and able to separate up to 16 urinary PnPs in a single run.
Although screening for inborn errors of PnP metabolism using HPLC-electrospray-tandem mass spectrometry has gained popularity in recent years, this method is still not widely used in most diagnostic laboratories in the country. The technique requires the use of expensive stable isotope-labelled internal standards, and the need for specialized skills for the operation, maintenance and interpretation of results and the high instrumentation costs are among the limiting factors for their applications. Thus, this study still used the conventional method of RP-HPLC to detect PnP metabolites in our sample.
A RP-HPLC method suggested by Vidotto el al.10 has been adopted and successfully established for the separation of the PnP metabolites in urine samples. The method was capable of separating up to 18 PnP metabolites in a single run with a relatively short analysis time. As noted, an earlier study by Valik and Jones2 required about 45 minutes for a single analysis run. The established method was found to be rapid, requiring an analysis time of only 28 minutes and an equilibration period of 2 minutes between the runs, hence making it possible to analyse 40 samples per day. Moreover, the established method required only a minimal amount of sample preparation. Despite the much shorter analysis time, this method was able to maintain a good resolution between the PnP metabolites and the urinary interference from drug metabolites.
The reference ranges and PnP metabolites in Malaysian children obtained by our RP-HPLC method showed some variation with those reported by previous authors.2,4,10 Only 8 out of 18 metabolites were detected in 556 normal control subjects. These metabolites were orotic acid, creatinine, uracil, pseudouridine, uric acid, hypoxanthine, xanthine and 7-methylguanine. The metabolites which were not detected in our normal subjects included inosine, guanosine, thymine and adenosine. Significant differences in urinary excretion of uracil, pseudouridine, uric acid, hypoxanthine, xanthine and 7-methylguanine were observed in all age groups; the exceptions were orotic acid and creatinine. The newborn group of patients (< 5 days old) had higher amounts of uric acid, creatinine and pseudouridine than other age groups. However, uracil was not detected at all in newborn babies < 5 days of age (n = 14). The increased level of these metabolites in the newborn was probably due to the carryover from the mother. The significant increase of urinary PnP excretion in children compared with adults may be due to high cell turnover and rapid child development. Interestingly, our study showed that Malaysian children have a lower excretion of urinary PnP but higher excretion of uric acid than the Western population.10 The general finding of higher uric acid excretion among the Malaysian population may be due to the varying dietary purine intakes.23 The ratio of uric acid – creatinine was also found to be age dependent among children < 10 years old and the ratio decreased as the age increased. The decreased uric acid – creatinine ratio was caused by the decrease in uric acid excretion in children aged 3–14 years.24 Duran et al.22 have emphasized the significance of measurement and assessment of uric acid and urine uric acid – creatinine ratio for unexplained metabolic diseases; the establishment of normal reference intervals of this parameter is beneficial as it may be used as a first-line screening for purine defects in our local Malaysian patients.
The implementation of the RP-HPLC method for the measurement of PnP metabolites at our centre in early 1997 led to the diagnosis of several rare cases of inborn errors of PnP metabolism among Malaysian children, which had never been reported before. These cases include the ADSL deficiency in Malaysian patients with a novel mutation by Chen et al.12 and the first confirmed case of MNGIE in Malaysian patients with a novel mutation in exon 5 of the TP gene.15 Ngu et al.13 reported the first Malaysian with molybdenum cofactor deficiency, and isolated sulphite oxidase deficiency in a Malaysian patient was reported by Shanti et al.14 In this paper, we also described two newly diagnosed cases of dihydropyrimidine dehydrogenase (DPD) deficiency in our patients. To the best of our knowledge, there are no previous confirmed reports of DPD disorders in the literature on Malaysian patients and, therefore, this is the first report from this region. Indeed, in addition to providing the method to detect them fairly rapidly and with minimal expertise, this study provided epidemiological and baseline data on these disorders in the Asian population. It is important to diagnose these conditions early as genetic counselling and prenatal diagnosis can be provided for the affected families.
In conclusion, the rapid HPLC method which we have established has proven to be efficient, reliable and sensitive enough to be applied and adopted in our clinical laboratory for the diagnosis of inborn errors of PnP metabolism in our local population. This study further demonstrated that there are significant inborn errors of PnP metabolism among Malaysian children. Assessment of PnP analysis should be considered for patients with unexplained neurological symptoms, particularly seizures, mental retardation and hypotonia, which were found to be the most common and predominant neurological features (87%) among the patients within this study.
The present study indicated that more advanced investigations are needed to further improve the diagnosis of disorders of PnP metabolism in the local population. Further urine samples from patients with various manifestations should be analysed to look for other inborn errors of PnP metabolism.