Salla disease

ABSTRACT

An automated ultraviolet (UV) enzymatic assay for urine total sialic acid (SA), performed on a Cobas Fara analyser, is described and compared with the colorimetric Warren method, which is used widely to determine urine SA. Intra-assay coefficient of variation (CV) for urine total SA determination was 0.83% for the UV assay and 3.5% for the Warren method. Inter-assay CVs were 1.8% and 5.6%, respectively. Recovery of urine total SA ranged from 89% for the UV assay to 61 % for the Warren method. Both were linear over a range of urine SA from 20 to 240 mg/L The UV assay was automated, took approximately 20 min to produce a result and avoided the need for solvent extraction; however, the reagents were expensive in comparison to those required for the Warren method. Urine samples with a creatinine concentration > 14 mmol/L were diluted with distilled water to optimise SA recovery by the UV method. Urine SA:creatinine ratios for normals were 4.7 (+/-1.7) g/mol with the Warren method and 4.5 (+/-1.0) g/mol for the UV method. Similarly, in type-2 diabetic patients, urine SA:creatinine ratios were 7.6 (+/-2.3) g/mol (P

KEY WORDS: Enzyme tests. Sialic acids. Urine analysis.

Introduction

Sialic acid (SA) is a major component of glycoproteins and glycolipids, in which it commonly forms the terminal constituent of saccharide side chains.1 The term 'sialic acid' is a generic one that describes a group of derivatives of neuraminic acid - a nine-carbon sugar. SA is highly negatively charged, giving rise to active or repulsive forces between different cell surfaces, and can influence membrane receptor and glycoprotein conformational states. Sialic acids often form part of antigenic determinants of glycolipids or glycoproteins.1

Urine total SA (TSA) is increased in diabetic patients, possibly due to alterations in neuraminidase and sialytransf erase activities,2 and also in individuals with chronic glomerulonephritis.3 In addition, there are a number of inherited inborn errors of SA metabolism, such as sialidosis, galactosialidosis, Kanzaki disease and mucolipidosis II and III, in which an increase in urine total SA is found.4-6

The majority of urine SA is bound to glycopeptides and oligosaccharides; however, some 40% is unbound and exists in free form. This free fraction is filtered by glomeruli but not reabsorbed by the renal tubules. Urine SA also can be increased (predominantly as free SA) in other inborn errors of metabolism such as infantile sialic acid storage disease, sialuria and Salla disease.7-9

The purpose of this study is to establish a simple, automated and relatively quick assay to measure urine TSA using commercially available reagents on a Cobas Fara discrete analyser. This ultraviolet (UV) enzymatic method is based upon that reported for serum samples by Arakai et al." In addition, we compare this with the Warren method," as modified by Roboz,12,13 which is a colorimetric method using thiobarbituric acid and solvent extraction used traditionally to assay urine SA. A rapid automated assay to determine urine SA would be useful clinically in facilitating the diagnosis of some of the inborn errors of SA metabolism mentioned above.

Materials and methods

The assays used for urine SA were performed using reagents (including all enzymes) supplied by the Sigma Chemical Company, Poole, Dorset, UK. The standards (primary) used for both assays were prepared in-house from N-acetylneuraminic acid (NANA, type VIII; Sigma) over a range from 20 to 500 mg/L. Calibration was checked with SA standard material supplied by Roche Diagnostics, Lewes, Sussex, UK.

Ultraviolet enzymatic assay

Briefly, the UV automated enzymatic assay of Arakai et al." is a coupled enzyme assay, incorporating neuraminidase (Clostridium perfringens), N-acetylneuraminic acid aldolase and lactate dehydrogenase (LDH). Liberated NADH is converted to NAD', which can be measured kinetically by following the reaction absorbance at 340 nm. The method was set up for use on a Cobas Fara analyser (Roche, Welwyn, Herts, UK). In this study we prepared three reagent mixtures.

Reagent 1: 10 mmol/L Tris buffer (pH 7.5) in 2 mL distilled water containing 40 (mu)L LDH at 1000 units/mL and 2mg NADH.

Reagent 2: 100 units neuraminidase in 10 mL Tris buffer (pH 7.5)

Reagent 3: NANA aldolase (25 units) dissolved in 2.5 mL Tris buffer (pH 7.5).

Warren method

Briefly, the Warren method consists of the oxidation of NANA with periodate, which is terminated by the addition of arsenite. This is followed by the addition of thiobarbituric acid, resulting in formation of a red colour that can be extracted in cyclohexanone and read at 549 nm.

The method used was as follows: 400 (mu)L 0.063mol/L sulphuric acid was added to 100 (mu)L urine and incubated at 80 deg C for 1 h. The tubes were cooled in iced water for 5 min and then 150 (mu)L 0.2mol/L sodium periodate was added and the tubes allowed to stand for 20 minutes. The reaction was terminated by adding 1 mL 10% (w/v) sodium arsenite and each tube was vortex-mixed.

Subsequently, 3 mL 0.6% (w/v) thiobarbituric acid made up in 0.5 mol/L sodium sulphate was added and the mixture boiled for 5 min. After cooling in iced water, the pH of the reaction mixture was adjusted to 5.6-6.0 with 5 moVL sodium hydroxide (300 (mu)L). Cyclohexanone (4 mL) was added, the tubes were centrifuged at 900 xg for 5 min at room temperature and the upper organic layer discarded. The pH of the aqueous layer was readjusted to 1.75-1.95 with 5 mol/L hydrochloric acid (300 (mu)L), a further 4 mL cyclohexanone was added and the tubes centrifuged again.

Absorbance of the upper organic layer was measured at 549 nm in a Phillips PU870 scanning spectrophotometer. Urine creatinine was measured by the Jaffe reaction in a Vitros 250 (Johnson and Johnson, Amersham, Bucks) using reagents supplied by the manufacturer. The intra-assay coefficient of variation (CV) for this assay was

Samples

Thirty-six fresh urine samples were collected in the early morning from 18 healthy individuals and 18 type-2 diabetic patients. These two subject groups were chosen in order to study urine TSA across a wide range of concentrations. To our knowledge, none of the individuals included in the study had suffered recent urinary tract infection, malignant or inflammatory disease, or had an inborn error of SA metabolism.

In total, the two groups comprised 16 females and 20 males with a mean age (+/-SD) of 30 (+/-10) years, range 22-55 years. Urine was collected into a plastic universal container without preservative and stored at 4 deg C for a couple of day before analysis. Prior to assay, the samples were mixed gently and returned to room temperature.

Intra-assay CVs were calculated on at least 10 determinations. Inter-assay CVs for urine TSA were also calculated on 10 different SA concentrations on three consecutive days. Linearity for urine TSA assay was studied up to a concentration of 240 mg/L using serial dilutions with distilled water.

Recovery for the urine TSA assay was studied by 'spiking' urine samples with 50 mg/L NANA. Before-addition and after-addition urine SA was determined for each sample and the percentage recovery calculated.

Statistics

Results are presented as the mean (+/-SD) or, if the data were not normally distributed, the mean (range). CV was calculated as follows: SD of assay x 100%/mean of assay results. Comparison of the two urine SA assays was made using the method of Altman and Bland. This is recommended for assay comparison where CVs cannot be assumed to be similar.14 Differences between data means were compared using the Student's t-test. Statistical significance was taken as P

Results

Intra-assay CVs for urine TSA determination using the UV assay and the Warren method were 0.83% and 3.5%, respectively. Inter-assay CVs were 1.8% and 5.6%, respectively. Recoveries for the urine TSA assays ranged from 88% to 108% for the UV assay and 50% to 61% for the Warren method. However, it was noted that recoveries for the UV method decreased to between 65% and 88% when urine creatinine concentration was >14 mmol/L. Assay linearity for the UV assay is shown in Figure 1 and for the Warren method in Figure 2. An Altman-Bland plot comparing the two urine SA assay methods is shown in Figure 3. Urine SA:creatinine ratio for normals was 4.7 (+/- 1.7) g/mol with the Warren method and 4.5 (+/- 1.0) g/mol with the UV method. Similarly, in type-2 diabetic patients, urine SA:creatinine ratios were 7.6 (+/- 2.3) g/mol (P

Discussion

Data showing the performance characteristics of a urine TSA UV enzymatic assay for use on a Cobas Fara analyser are presented. The UV assay was found to be of good analytical precision with intra-assay CVs

To our knowledge, this is the only reported study to have looked at a urine SA assay based on an enzymatic method using commercially available reagents in a Cobas analyser system. In our hands, the Warren method performed poorly compared to the UV method in the recovery experiments; however, the number of manual steps in the former may have contributed to this finding. The Altman-Bland plot shows a positive bias with the Warren method, which may be not too surprising given that the UV assay utilised an enzyme system supposedly specific for SA.

The majority of work on the UV assay method has been performed with serum or plasma samples and we are unaware of a comparative study with the Warren method a urine TSA assay used commonly in the routine clinical laboratory. The UV assay gave better CV values and recoveries; however, it did give lower recoveries when urine creatinine concentration was > 14 mmol/L. Thus, we would advocate caution when using concentrated urine samples and suggest that they be diluted with distilled water to reduce the creatinine concentration. In view of this, it may be useful to express SA as a creatinine ratio in urine samples.

Both assays were linear over the urine SA concentration ranges found in our samples. Urine SA:creatinine ratio was significantly higher in the type-2 diabetic patients by both methods; however, there were no significant differences between the two methods. It is known that increased glomerular permeability occurs in diabetes mellitus and the results may reflect leakage of sialylated proteins into the urine.

Many workers measure urinary SA by modifications of the Warren method,12,13 which is a colorimetric assay that utilises thiobarbituric acid. However, there has been criticism of the method because of interference with 2-deoxyribose and therefore the pH correction procedure devised by Roboz et al.13 should be applied. The SA UV enzymatic assay used in the present study is not prone to this interference and is thought to provide greater specificity for SA. This would be important because of possible interference from drugs or from other 'sugar' molecules.

It should be noted, however, that the two methods compared here may give different results not only because of differences in specificity but also because of methodological variations in the assay procedure and the liberation of bound SA. The Warren method uses acid hydrolysis and heating of the sample at 80 deg C, whereas the UV assay employs the enzyme neuraminidase to liberate bound SA. Interestingly, however, not all sialoglyco-- conjugates are cleaved by neuraminidase. It may also be possible to measure urine free SA using the UV assay by measuring SA before and after treatment of urine with neuraminidase. Proteinuria may result in increased urine total SA because some urine proteins are sialylated.

Previously, we described a colorimetric assay for the determination of serum TSA using a Cobas Bio analyser15 - a method that incorporated pyruvate oxidase and peroxidase with a colorimetric dye.16,17 However, the manufacturer of the manual kit for this method does not recommend its use with urine samples because of interfering substances, and was one of the reasons we used the urine UV SA method reported here. Furthermore, there are a number of highperformance liquid chromatography techniques to assay SA18-20 but these are considered to be mainly for research purposes, being too tedious for routine clinical use.12

In conclusion, the UV assay reported here is quick to perform and is automated. Assay time is considerably shorter (approximately 20 min to allow the assay reaction to complete) than the Warren method, which requires heating at 80 deg C for one hour and subsequent organic solvent extraction. Furthermore, the sodium arsenite used in the Warren method is toxic and can be difficult to purchase. Conversely, the reagents used in the UV assay are expensive (approximately L1 per test). Therefore, we developed the assay using the Cobas Fara analyser, an automated machine which requires only small reagent volumes.

Finally, urine samples with a creatinine concentration >14 mmol/L should be diluted to ensure adequate SA recovery when using the UV assay, and the higher SA:creatinine ratio observed in our type-2 diabetic patients merits further study in a larger cohort.

We are grateful to Prof. Swaminathan for use of laboratory facilities.

References

1 Schauer R. Characterisation of sialic acids. Methods Enzymol 1978; 50: 64-89.

2 Baggio B, Briani G, Cicerello E, Gambaro G, Bruttomesso D, Tiengo A. Urinary glycoaminoglycans, sialic acid and lysosomal enzymes increase in non-albuminuric diabetic patients. Nephron 1986; 43: 187-90.

3 Ozben T Elevated serum and urine sialic acid levels in renal disease. Ann Clin Biochem 1991; 28: 44-8.

4 Parkinnen J, Finne J. Isolation and structural characterisation of five major sialyloligosaccharides and a sialylglycopeptide from normal human urine. Eur J Biochem 1983; 136: 355-61.

5 Kanzaki T, Yoketa M, Mizuno N, Matsumoto Y, Hirabayashi Y. Novel lysosomal glycoaminoacid storage disease with angiokeratoma corporis diffusum. Lancet 1989; i: 875-7.

6 van Pelt J, van Bilsen DGJL, Kamerling JP, Vliegenthart JFG. Structural analysis of 0-glycosidic type of sialyloligosaccharide

alditols derived from urinary glycopeptides of a sialidosis patient. Eur J Biochem 1988; 174: 183-7.

7 Seppala R, Renlund M, Bernardini I, Tieze F, Gahl WA. Renal handling of free sialic acid in normal humans and patients with Salla disease or renal disease. Lab Invest 1990; 63: 197-202,

8 Baumkotter J, Cantz M, Mendla K, Baumann W, Friebolin H, Gehler J. N-acetylneuraminic acid storage disease. Hum Genet 1985; 71: 155-9.

9 Stevenson RE, Lubinsky M, Taylor HA, Wenger DA, Schroer RJ, Olmstead PM. Sialic acid storage disease with sialuria: clinical and biochemical features in the severe infantile type. Paediatrics 1983; 72: 441-9.

10 Araki H, Yamada M. Sialic acid. In: Bergmeyer HU, ed. Methods of enzymatic analysis. Vol VI. 3rd edn. Weinheim: Verlag Chemie, 1984: 80-90.

11 Warren L. The thiobarbituric acid assay of sialic acids. I Biol Chem 1959; 234: 1971-5.

12 Waters PJ, Lewry E, Pennock CA. Measurement of sialic acid in serum and urine: clinical applications and limitations. Ann Clin Biochem 1992; 29: 625-37.

13 Roboz J, Suttajit M, Bekesi JG. Estimation of 2-deoxyribose interference in the thiobarbituric acid determination of N-acetylneuraminic acid in tumor cells by pH-dependent extraction with cyclohexanone. Ann Biochem 1981; 110: 380-8.

14 Bland JM, Altman DG. Statistical methods in assessing agreement between two methods of clinical measurement. Lancet 1986; i: 307-10.

15 Crook M. The determination of serum or plasma sialic acid. Clin Biochem 1993; 26: 31-37.

16 Sugahara K, Sugimoto K, Nomura 0, Usui T Enzymatic assay of serum sialic acid. Clin Chim Acta 1980; 108: 493-8.

17 Taniuchi K, Chifu K, Hayashi N, Nakamachi Y, Yamaguchi N, Miyamoto Y A new enzymatic method for the determination of sialic acid in serum and its application for a marker of acutephase reactants. Kobe J Med Sci 1981; 27: 91-102.

18 Karamanos NK, Wikstrom B, Antonopolous CA, Hjerpe A. Determination of N-acetyl and N-glycolyl-neuraminic acids in glycoconjugates by reversed-phase HPLC with UV detection. J Chromatogr 1990; 503: 421-9.

19 Wang WT, Erlansson K, Lindh F, Lundgren T, Zopf D. Highperformance liquid chromatography of sialic acid containing oligosaccharides and acidic monosaccharides. Anal Biochem 1990; 190: 182-7.

20 Manzi AE, Diaz S, Varki A. HPLC of sialic acid on a pellicular resin anion-exchange column with pulsed amperometric detection: a comparison with six other systems. Anal Biochem 1990; 503: 421-9.

M.A. CROOK, S. KARGBO and P. LUMB

Clinical Chemistry Department, 5th Floor Tower. Guy's Hospital, London SE1 9RT, UK.

Accepted: 9 November 2001

Correspondence to: Dr Martin Crook Email: martin.crook@gstt.sthames.nhs.uk

Copyright Royal Society of Medicine Press Ltd. 2002
Provided by ProQuest Information and Learning Company. All rights Reserved