Annals of Clinical Biochemistry >
Volume 45, Number 6 >
Pp. 575-584
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
1 Biochemical Genetics Unit, Box 247, Addenbrookes Hospital, Hills Road, Cambridge CB2 0QQ, UK;
2 Biochemical Genetics Laboratory, Greenwood Genetic Center, 125 Gregor Mendel Circle, Greenwood SC 29646, USA
Corresponding author: Dr R S Carling. Email: rachel.carling{at}nhs.net
 | Abstract |
|---|
 Top Abstract IntroductionMaterials and methodsResultsDiscussionACKNOWLEDGEMENTSREFERENCES |
|---|
Methods: Analysis was performed by liquid chromatography-tandem mass spectrometry in positive ionization mode. Samples were analysed underivatized using a rapid ‘dilute and shoot’ approach. Chromatographic separation of the three compounds was achieved. Stable isotope internal standards were used for quantification.
Results: Creatine, creatinine and guanidinoacetate were measured with a 2.5 minute run time. For guanidinoacetate, the standard curve was linear to at least 5000 µmol/L and for creatine and creatinine it was linear to at least 25 mmol/L. The lower limit of quantitation was 0.4 µmol/L for creatine and guanidinoacetate and 0.8 µmol/L for creatinine. Recoveries ranged from 86% to 106% for the three analytes. Intra- and inter-assay variation for each analyte was <10% in both urine and plasma.
Conclusion: A tandem mass spectrometric method has been developed and validated for the underivatized determination of guanidinoacetate, creatine and creatinine in human urine and plasma. Minimal sample preparation coupled with a rapid run time make the method applicable to the routine screening of patients with suspected CCDS.
| Introduction |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionACKNOWLEDGEMENTSREFERENCES |
|---|
|
The most common of the CCDS is CRTR deficiency which is an X-linked disorder resulting from mutations in the SLC6A8 gene and causes mild to severe MR, absent or delayed speech and epilepsy. Biochemically, these patients have increased urinary creatine:creatinine ratios and plasma creatine is usually within the reference range. Diagnosis can be confirmed using mutation analysis and creatine uptake in fibroblasts.8 CRTR deficiency is thought to be a relatively common cause of X-linked MR (XLMR) and its prevalence in this population has been investigated by several authors. Rosenberg et al.13 investigated 290 patients with XLMR and found that the prevalence of CRTR deficiency, as demonstrated by SLC6A8 mutation analysis, to be 2.1% which is similar to that of fragile-X syndrome. A similar study of 114 boys with MR reported an even higher prevalence of 3.5%,14 whereas two other studies reported slightly lower prevalences of 1% and 0.25% in populations of 478 males with MR15 and 1600 males with MR and/or autism,16 respectively. These studies demonstrate the importance of screening for CRTR in patients with XLMR.
GAMT deficiency is an autosomal recessive disorder with a more complex phenotype than CRTR deficiency. It is associated with MR, epileptic seizures, speech delay, autism and extrapyramidal movement disorder.17 The pathogenesis of GAMT deficiency is thought to involve the toxicity of guanidinoacetate as well as the deficiency of creatine. In GAMT deficiency, there are increased concentrations of guanidinoacetate in urine, plasma and cerebrospinal fluid (CSF) with low concentrations of creatine in urine and plasma. The diagnosis can be confirmed using mutation analysis15 or enzyme analysis in lymphocytes or cultured amniocytes.16,17 The prevalence of GAMT deficiency is not yet known but in the region of 30 patients had been reported by 2005.17–21
AGAT deficiency is an autosomal recessive disorder characterized by mild MR and autistic behaviour. It is the least common of the CCDS with only four cases, from two unrelated families, being reported to date.7,22,23 Biochemically, these patients have decreased concentrations of guanidinoacetate and creatine in both urine and plasma. Mutation and enzyme analysis are available for confirmation of the diagnosis.7
Several analytical methods for the determination of guanidinoacetate and creatine in biological samples have been validated, including gas chromatography-mass spectrometry (GCMS),24,25 liquid chromatography (LC)/atmospheric pressure chemical ionization-MS26 and LC tandem MS (LC-MSMS).27–29 However, without exception these methods require some form of sample pretreatment such as solid-phase extraction or derivatization that makes sample preparation time-consuming, labour-intensive and more expensive. Furthermore, the GCMS assay does not allow the simultaneous determination of creatinine so a separate assay is required. High-performance LC (HPLC) with fluorescence detection has also been used,30 but the inability to separate creatine from creatinine is a major limitation.31 We have developed and validated a rapid, robust method for the determination of creatine and guanidinoacetate in urine and plasma, suitable for the routine investigation of patients with suspected CCDS. The LC-MSMS assay described here does not require sample derivatization or clean up but still provides adequate sensitivity for the analysis of patient samples. One other method employing a similar ‘dilute and shoot’ approach has been reported32 but exists only in abstract form with no published method validation. Although to date relatively few patients with CCDS have been reported this is likely to be a reflection of the limited availability of quantitative assays, resulting in significant under diagnosis.
| Materials and methods |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionACKNOWLEDGEMENTSREFERENCES |
|---|
Creatine, creatinine, guanidinoacetate, bovine serum albumin (BSA) and formic acid were supplied by Sigma Chemicals (Poole, Dorset, UK). HPLC grade methanol (AnalaR) was purchased from BDH (Poole, England). [2H3]-creatine (99.0% isotopic purity) and [2H3]-creatinine (99.8% isotopic purity) were obtained from CDN isotopes (Quebec, Canada) and [13C2]-guanidinoacetate was supplied by Dr HJ Ten Brink (VU University Medical Centre, Amsterdam, the Netherlands).
Standards, internal standards and quality control
Aqueous stock standards of each analyte (5, 50 and 100 mmol/L for guanidinoacetate, creatine and creatinine, respectively) and internal standard (4, 10 and 25 mmol/L for [13C2]-guanidinoacetate, [2H3]-creatine and [2H3]-creatinine, respectively) were prepared. Formic acid (50 µL 98%/50 mL) was added to aid dissolution of guanidinoacetate. Mixed working standards were prepared by further dilution of the stock standards to produce a series of calibrators. A set of mixed working standards were also prepared in phosphate-buffered saline (PBS), pH 7.4 containing 4.5% BSA, for the determination of the LOQ. Combined working internal standard solution was prepared by further dilution of the stock internal standards to give concentrations of 100 µmol/L for [13C2]-guanidinoacetate and [2H3]-creatine and 1 mmol/L for [2H3]-creatinine for analysis of urine and concentrations of 5 µmol/L [13C2]-guanidinoacetate and [2H3]-creatine for analysis of plasma. All calibrators and internal standards were stored at –80°C.
Urine and plasma samples were aliquoted and stored in polypropylene microfuge tubes at –80°C for use as quality control (QC) material.
Urine
Samples, standards or QC (50 µL) were added to polypropylene microfuge tubes (1.5 mL). Internal standard (50 µL) was added and the sample vortex mixed. Distilled water was added (500 µL), the sample vortex mixed and then microfuged at 13,000 rpm for 10 minutes. Two hundred microlitres were transferred to a 96-well microtitre plate for analysis.
Plasma
Samples or QC samples (50 µL) were added to polypropylene microfuge tubes (1.5 mL). Internal standard (50 µL) was added and the sample vortex mixed. Methanol was added (500 µL), the sample vortex mixed and then microfuged at 13,000 rpm for 10 minutes. Two hundred microlitres of supernatant were transferred to a 96-well microtitre plate for analysis.
High-performance liquid chromatography
Chromatography was performed on a Perkin Elmer Series 200 pump and autosampler. Creatine, creatinine and guanidinoacetate were separated isocratically by injecting sample extract (5 µL) onto a C18 Synergi Fusion RP column (4 µm 150 x 3.0 mm) fitted with a Security Guard C18 column (4.0 x 3.0 mm), both from Phenomonex (Macclesfield, UK). The column assembly was at ambient temperature. The mobile phase was composed of methanol and 0.025% formic acid (5:95 v/v) and delivered at a flow rate of 500 µL/minute. The total run time was 2.5 minutes.
A triple quadrupole mass spectrometer, Sciex API 2000 (Applied Biosystems) was used for analysis. The instrument was operated in positive ionization mode and sample introduction was via the HPLC system described above. To tune the mass spectrometer, 10 µmol/L solutions of creatinine, creatine and guanidinoacetate in mobile phase were infused directly into the ion source. For each analyte, the source conditions were optimized to maximize the intensity of the MH+ precursor ions for the analytes and internal standards (m/z 114, 117, 118, 120, 132 and 135). The optimized settings are summarized in Table 1. The collision energy was then adjusted to optimize the signal for the most abundant product ions (m/z 44, 47, 76, 78, 90 and 93). The common instrument settings were as follows: capillary voltage, 5.5 kV; source temperature 400°C; curtain gas 20 pounds per square inch (psi); collision-activated dissociation gas 3; nebulizing gas 50 psi; heater gas 50 psi; interface heater on.
|
Linearity
Standard curves were prepared in triplicate and analysed in a single batch. Analyte to internal standard peak area ratio was plotted against analyte concentration. All calibration curves were linear (correlation coefficient, r2 > 0.99).
Ion suppression
Ion suppression was investigated by direct infusion of internal standard solution into the mass spectrometer while an extracted urine sample was simultaneously injected via the autosampler. This process was repeated with a plasma extract.
Precision
Intra-batch precision was determined at various concentrations by randomized analysis of patient samples, 10 times within one batch. Inter-batch precision was determined using the analysis of the same patient sample, 10 times, over a four-week period. A coefficient of variation (CV) of <10% was considered acceptable.
Sensitivity
Lower limit of quantitation (LOQ) was determined by serial dilutions of matrix-matched standards (PBS containing 4.5% BSA). Samples were analysed 10 times within a batch and the LOQ was taken to be the lowest concentration of each analyte that was measurable with a %CV < 20, bias < 20% and a signal-to-noise ratio > 10. The limit of detection (LOD) was determined using the same set of samples and was defined as the analyte concentration that gives a 3:1 signal-to-noise ratio.
Recovery
Recovery was determined by comparison of each analyte in patient samples, before and after the addition of a known amount of each analyte. Urine samples (n = 6) were spiked to give additional concentrations of guanidinoacetate, creatine and creatinine of 0, 250, 1000 and 5000 µmol/L, 0, 2500, 10,000 and 20,000 µmol/L and 0, 5 and 20 mmol/L, respectively. Plasma samples (n = 6) were spiked to give additional concentrations of guanidinoacetate and creatine of 0, 2,10 and 40 µmol/L, and 0, 10, 50 and 200 µmol/L respectively.
Stability
The effect of storage temperature on the stability of the three analytes was investigated. Six urine samples were aliquoted. One aliquot of each was immediately frozen while the others were left at room temperature and 4°C for 1, 2, 3, 4, 5, 6, 8 and 12 hours. Stability was also assessed by analysis of urine samples after one to seven freeze thaw cycles. The stability in lithium heparin plasma, blood and serum was investigated. Six plasma samples were aliquoted and left at room temperature. One aliquot of each was immediately frozen while others were left at room temperature for 3, 6, and 24 hours. This was repeated for serum except with the serum left in the gel tube and an aliquot was removed at the appropriate time. Six blood samples were aliquoted and left at room temperature for 1, 2, 3, 4, 5 and 6 hours prior to separation and freezing. In each case, samples were judged to be stable if the variation in results was less than the within-batch CV. For samples where the variation was greater than the within-batch CV, an F test was performed and P < 0.05 was considered to indicate a significant difference.
Accuracy
The accuracy of the method was assessed by the comparison of urine and plasma assays with the all-laboratory trimmed mean (ALTM) of the European Research Network Inherited Disorders of Metabolism (ERNDIM) quality assurance samples.
Method comparison was performed on patient samples that had previously been analysed by an established GCMS method at VU University Medical Center, Amsterdam, Netherlands (urine, n = 22, plasma n = 44) or LC-MSMS method at Greenwood Genetics Centre, SC, USA (urine, n = 93). The results were compared using the Bland–Altman plots.
Samples from four patients with CCDS confirmed by mutation analysis were analysed.
| Results |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionACKNOWLEDGEMENTSREFERENCES |
|---|
|
Ion suppression experiments demonstrated that in plasma the main area of ion suppression occurs between 1.1 and 1.3 minutes, with no significant ion suppression occurring when the creatine and guanidinoacetate elute (Figure 3). In urine ion suppression experiments demonstrated that while creatine and guanidinoacetate were separated from the main area of ion suppression, creatinine partially co-eluted. However, sensitivity is not an issue with the analysis of creatinine and it was considered more important to maintain its separation from guanidinoacetate and creatine, which may be present in concentrations up to two orders of magnitude lower.
|
|
|
In plasma, the mean recoveries for guanidinoacetate and creatine were 99.9% (range 89.4–107.2) and 101.6% (range 89.8–104.6), respectively. In urine, the mean recoveries for guanidinoacetate, creatine and creatinine were 98.5% (range 93.7–106.0), 89.4% (range 85.5–96.3) and 87.7% (range 82.5–90.8), respectively.
F tests demonstrated that creatinine and guanidinoacetate were stable in urine for up to 12 hours at room temperature. For creatinine, the average variation was 2.2% (range 1.0–4.2) and for guanidinoacetate the average variation was 2.5% (range 1.7–4.1). However, at room temperature urinary creatine was found to be unstable exhibiting an average increase of 27.1% over 12 hours (range 3.4–66.5, n = 6) as shown in Figure 4. F tests showed the increases in creatine to be significant in four of the six samples (P < 0.05) and they also demonstrated that although creatinine and guanidinoacetate were stable in urine for up to seven freeze thaw cycles, creatine was not. The %CVs for the seven cycles were 3.9 and 2.7 for creatinine and guanidinoacetate, respectively, whereas creatine had increased by 22.4% after only two freeze thaw cycles and by 152.4% after seven cycles which was significant (P < 0.05).
|
For both plasma and urine, all analytes were within standard deviation (2SD) of the ERNDIM ALTM (n = 10), as shown in Figure 5. In urine, the average SDs from the ERNDIM mean were 0.20, 0.05 and –0.07 for creatinine, guanidinoacetate and creatine, respectively. In plasma, the average SDs from the ERNDIM mean were 0.23 and 0.97 for guanidinoacetate and creatine, respectively.
|
9 and demonstrate reasonable agreement for each analyte in both urine and plasma. The correlation coefficient, regression equations and mean bias for the GCMS comparison were y = 0.9867x – 35.786 (r2 = 0.9869), –56 µmol/L for urinary creatine, y = 0.9205x + 15.711 (r2 = 0.9890), –9 µmol/L for urinary guanidinoacetate, y = 0.9769x – 1.8858 (r2 = 0.9690), –4 µmol/L for plasma creatine and y = 0.9318x + 0.4779 (r2 = 0.9604), 0.4 µmol/L for plasma guanidinoacetate. The correlation coefficient, regression equations and mean bias for the LC-MSMS comparison were y = 0.9575x + 135.2 (r2 = 0.9893), 58 µmol/L for creatine, y = 0.9325x + 9.5127 (r2 = 0.9793), –17 µmol/L for guanidinoacetate and y = 0.8616x + 408.4 (r2 = 0.9624), 0.5 mmol/L for creatinine.
|
|
|
|
|
| Discussion |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionACKNOWLEDGEMENTSREFERENCES |
|---|
In the study by Arias et al.16 investigating the prevalence of CRTR deficiency in a population of 1600 males with MR and/or autism, 29 false-positive creatine:creatinine ratios were obtained. The authors postulated that this was a dietary effect and subsequently investigated 13 adult volunteers, analysing urinary creatine:creatinine ratios in first morning urine samples collected after low protein (salad, pasta) and protein-rich (beef and oily fish) meals. While this appeared to show a protein-rich diet resulted in a creatine:creatinine ratio above the reference range (460 µmol/mmol) in four of the 13 volunteers, the paper did not indicate how promptly the urine samples were frozen so it is possible that the lack of stability of creatine also contributed to the false-positive rate.
The lack of stability of creatine may also bring into question the validity of existing reference ranges quoted in the literature. However, any discrepancies are almost certainly compounded by the relatively small numbers of subjects used to determine the reference ranges. Almeida et al.33 reported the following age-related reference ranges for urinary creatine:creatinine ratios: <4 years: 6–1200 µmol/mmol, 4–12 years: 17–720 µmol/mmol and >12 years: 11–240 µmol/mmol. Although the total number of subjects is stated to be 140, the number in each age group is not specified and while the paper does state that the samples were stored frozen, it does not specify how promptly. Likewise, Valongo34 determined age-related urinary creatine:creatinine reference ranges which are similar but perhaps somewhat broader: <4 years: 40–1510 µmol/mmol (n = 70), 5–11 years: 40–1070 µmol/mmol (n = 71) and 12–16 years: 40–560 µmol/mmol (n = 28), whereas Cognat28 quotes a narrower reference range of 12–585 µmol/mmol, although this is based only on 47 subjects with an age range of 5–23 years. We are in the process of collecting reference range data but as yet have insufficient samples to produce age-related reference ranges, given the potential problem of stability. Preliminary data gives urine reference ranges (non-parametric, n = 152) of 5–1645 µmol/mmol for creatine and 9–175 µmol/mmol for guanidinoacetate and plasma reference ranges (non-parametric, n = 72) of 10–154 µmol/L for creatine and 0.7–3.9 µmol/L for guanidinoacetate. However, pending collection of sufficient in-house data, we have provisionally adopted the plasma and urine reference ranges determined by the Greenwood Genetics Centre (Table 4). For CSF the only available data is that of Almeida et al.33 which is based only on 25 subjects: 0.02–0.56 µmol/L guanidinoacetate and 17–87 µmol/L creatine. Evidently, there is a need to establish further reference range data for urine, plasma and CSF, based on a greater number of subjects and with clearly defined preanalytical requirements.
Preliminary investigations of CSF guanidinoacetate have been performed using the LC-MSMS method described here with sample preparation identical to that of plasma. Ion suppression studies demonstrated that in CSF, guanidinoacetate elution is well removed from the region of ion suppression. Recovery experiments (n = 6) showed mean recoveries of 97.6% (range 77.0–112.4%). However, analysis of CSF guanidinoacetate lacked sensitivity at normal concentrations with intra- and inter-assay CVs of 17.7 and 20.1%, respectively (n = 10). While normal CSF contains negligible quantities of guanidinoacetate,27 at the concentrations seen in GAMT deficiency, which are one to two orders of magnitude higher, the intra- and inter-assay CVs were <10% indicating that potentially the method can differentiate GAMT patients from controls.
The method described here is clearly capable of differentiating CRTR and GAMT patients from controls but the diagnostic sensitivity of this assay with respect to AGAT deficiency is not known. However, AGAT deficiency is extremely rare (four cases reported to date) and accounts for a very small minority of the known CCDS patients so we believe that the speed and simplicity of this method make it worthwhile for screening large numbers of patients for CRTR and GAMT deficiencies, for which it has adequate sensitivity.
In summary, we have developed and validated a rapid, simple and inexpensive LC-MSMS method for the determination of guanidinoacetate, creatine and creatinine in urine and plasma. Samples from the four patients with confirmed CCDS demonstrate the ability of this method to differentiate them from controls and the method comparisons, and ERNDIM results give us confidence in the accuracy of our results.
This method has the advantages of speed, simplicity, linearity and accuracy over a wide analytical range and has now been offered as a routine service for over a year.
| ACKNOWLEDGEMENTS |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionACKNOWLEDGEMENTSREFERENCES |
|---|
| REFERENCES |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionACKNOWLEDGEMENTSREFERENCES |
|---|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
