Urinary galactitol and galactonate quantified by isotope-dilution gas chromatography-mass spectrometry

Clinica Chimica Acta 366 (2006) 216 – 224 www.elsevier.com/locate/clinchim

Urinary galactitol and galactonate quantified by isotope-dilution gas chromatography-mass spectrometry Claire Yager a, Suzanne Wehrli b, Stanton Segal a,c,* a

Metabolic Research Laboratory, The Children’s Hospital of Philadelphia, Philadelphia, PA 19104, USA b NMR Facility, The Children’s Hospital of Philadelphia, Philadelphia, PA 19104, USA c Department of Pediatrics, University of Pennsylvania School of Medicine, Philadelphia, PA 19104, USA Received 30 August 2005; received in revised form 11 October 2005; accepted 11 October 2005 Available online 5 December 2005

Abstract Background: Measurements of urine galactitol have been used to monitor the adequacy of diet therapy in the treatment of galactosemia. We have devised a gas chromatographic mass spectrometry (GC/MS) isotope-dilution method for the simultaneous quantification of urine galactitol and another alternate pathway product, galactonate. Methods: We prepared trimethylsilyl (TMS) derivatives and used d-[UL-13C]galactitol and d-[UL-13C]galactonate as the internal standard for GC/MS. Results obtained with this method were compared with those determined by the established GC method for galactitol and the NMR method for galactonate. Thirty-three normal urine specimens were analyzed by the isotope dilution technique for galactitol and galactonate. Results of galactitol in 6 of these urine specimens along with 18 from classic galactosemics and 19 variant galactosemics were compared with the established GC method. Results for galactonate in 15 urine specimens from galactosemics were compared to the established NMR technique. Results: The method was linear up to 200 nmol with lower limits of detection of 1.1 nmol (1.75 mmol/mol creatinine) (Cr) and 0.8 nmol (1.28 mmol/mol Cr) for galactitol and galactonate, respectively. Intra- and Interassay imprecision ranged from 2.1 – 6.7% for galactitol and 3.5 – 8.0% for galactonate. The excretion of both metabolites was age dependent in both normal and galactosemics. In 12 normal urines from subjects under 1 year, values for galactitol ranged from 8 – 107 mmol/mol Cr, and in 7 over age 6, ranged from 2 – 5 mmol/mol Cr. Under 1 year, the range for galactonate was non-detectable to 231 and in the over 6 years group non-detectable to 25 mmol/mol Cr. In galactosemics under 1 year, the value for galactitol ranged from 397 – 743 and for galactonate 92 – 132 mmol/mol Cr while in nine patients over age 6 the range was 125 – 274 mmol/mol Cr for galactitol and 17 – 46 mmol/mol Cr for galactonate. Conclusions: The GC/MS method enables the simultaneous determination of urine galactitol and galactonate and is precise and useful over the wide range of concentrations needed to assess the galactose burden in patients with galactosemia. D 2005 Elsevier B.V. All rights reserved. Keywords: Galactosemia; Galactitol; Galactonate; Gas chromatography mass spectrometry; Urine galactitol; Isotope dilution

1. Introduction

Abbreviations: GALT, galactose-1-phosphate uridyltransferase; TMS, trimethylsilyl; GC/MS, gas chromatographic mass spectrometry; Cr, creatinine; RBC, red blood cell. * Corresponding author. Metabolic Research Laboratory, The Children’s Hospital of Philadelphia, Philadelphia, PA 19104, USA. Tel.: +1 215 590 3372; fax: +1 215 590 3364. E-mail address: [email protected] (S. Segal). 0009-8981/$ - see front matter D 2005 Elsevier B.V. All rights reserved. doi:10.1016/j.cca.2005.10.015

When the normal pathway of galactose metabolism involving its conversion to glucose via the Leloir enzyme sequence of galactokinase, galactose-1-phosphate uridyltransferase (GALT) and uridine diphosphate galactose-4epimerase is impaired, alternate routes of galactose disposition become prominent. In human galactosemia due to GALT deficiency as well as in galactokinase deficiency [1], two pathways emerge to handle galactose. The first involves

C. Yager et al. / Clinica Chimica Acta 366 (2006) 216 – 224

reduction of galactose to the sugar alcohol, galactitol, which has been identified in tissues of affected individuals as well as urine and plasma [2]. Since there appears to be no further metabolism of galactitol [3], it is excreted in urine as a ‘‘dead end’’ metabolite of galactose. The large urinary excretion of galactitol in classic galactosemic patients has prompted its measurement as a monitoring procedure in the dietary management of patients and as assessment of their galactose burden [4,5]. The methods used for the quantitation have involved high-performance liquid chromatography [6], gas chromatographic (GC) procedures with trimethylsilyl (TMS) [4,7] or methoxylamine-acetate derivatives [5] and a GC/MS method using a hexacetate derivative [8]. In the second alternate pathway galactose is oxidized to galactonate which may be further metabolized by liberation of carbon one as CO2 with pentose formation as described by Cuatrecasas and Segal [9,10]. Bergren et al. [11] identified galactonate in the urine of both normals and galactosemia patients after giving a large load of galactose and Gitzelmann, Wells and Segal [12] quantitated its excretion in a galactokinase deficient patient by GC with TMS derivatives. Measurement of urinary galactonate in a group of GALT deficient patients on a galactose-restricted diet was reported by Wehrli et al. by NMR of underivatized urine [13]. That technique which established that the urine excretion of galactitol was several fold higher than galactonate was, however, too costly for routine analysis. Urinary galactonate quantitation of a butylgalactonamide pentaacetate method by GC/MS has been reported [14]. In devising a direct method for measurement of erythrocyte galactose-1-phosphate another metabolite accumulating in galactosemia by isotope dilution mass spectrometry utilizing TMS derivatization and [213C]galactose-1-phosphate as the internal standard [15], we showed in the resulting chromatography that both galactitol and galactonate were identified [16]. This resulted in the development of methods for the simultaneous quantitation of red blood cell (RBC) galactitol and galactonate by the isotope dilution method employing [UL13C]galactitol and [UL13C]galactonate as internal standards [17]. Recognizing the clinical need for the simultaneous measurement of galactitol and galactonate in urine for galactosemic subjects, we have devised a method, which for the first time, permits their measurement in urine by a single accurate and precise procedure. Its description forms the basis of this report.

2. Materials and methods 2.1. Materials d-galactitol and [UL-13C6]galactitol 99 atm.% excess (APE) were purchased from Pfanstiehl and Omicron Biochemicals, Inc, respectively. Potassium galactonate and

217

[UL-13C]galactonate were synthesized by iodine oxidation of galactose as reported by Blair and Segal [18]. The dgalactose and d-[UL-13C6]galactose, 99% APE, used in the synthesis of galactonate were also purchased from Pfanstiehl and Omicron Biochemicals, Inc, respectively. The purity of the d-[UL-13C6]galactonate was determined by GC/MS and chromatographic analysis with an APE of 97%. N,O-bis(trimethylsily)trifluoroacetamide with trimethylchlorosilane (BSTFA + 1% TMCS) was obtained from Pierce. Other reagents were obtained from Sigma and Fisher Scientific. 2.2. Sample preparation and derivatization Random samples of urine were collected and frozen at  80 -C until assayed. A 100 Al sample of urine as well as the standards prepared in H2O were first mixed with 30 U of urease in 20 Al and heated to 37 -C for 30 min to eliminate a large urea peak in the chromatography. After cooling, the 13 C internal standards were added. Ethanol, 0.9 ml, was added to deproteinize the samples and standards. They were centrifuged for 10 min and the protein was discarded. The supernatant was evaporated to dryness under a stream of nitrogen gas. The residue was trimethylsilylated with 100 Al of BSTFA + 1% TMCS and 100 Al of pyridine at 90 -C for 45min. A 1 Al portion of each derivatized sample was injected for GC/MS analysis. 2.3. GC/MS The analysis was performed on a Hewlett Packard GC/ MS 6890/5973 by electron ionization at 70 eV using an HP5MS column 30m  250 Am  0.25 Am cross linked with 5% phenylmethylsiloxane. The helium flow was kept in a constant flow mode at 0.6 mL/min. The oven ramp was set at initial temperature 125 -C for 1 min; at ramp 1, it was increased to 175 -C at a rate of 20 -C/min, then at ramp 2 to 230 -C, at a rate of 4 -C /min. Once the temperature reached 230 -C, it was kept at that temperature for 3 min. Finally, the oven temperature was increased to 300 -C at a rate of 40 -C/min. The temperature was kept at 300 -C for 5 min. The inlet temperature was set to 250 -C, and the splitless mode was applied. The purge flow was started at a rate of 50 mL/min 0.7 min after the 1-AL sample injection. The mass selective detector transfer line heater temperature was set to 280 -C. The source temperature was 230 -C. Galactitol and galactonate were identified by retention times chromatographically and by their fragmentation patterns [16,17]. Quantitation for galactitol uses the shift of the m/z 217 ion fragment of derivatized d-[12C]galactitol to the corresponding m/z 220 fragment of d-[UL-13C6]galactitol resulting from an increase of 3 mass units. For galactonate, the m/z 319 fragment of d-[12C]galactonate shifted to the corresponding m/z 323 fragment of the labeled compound with an increase of 4 mass units [17].

218

C. Yager et al. / Clinica Chimica Acta 366 (2006) 216 – 224

2.4. Calibration curve Calibrations were prepared in HPLC grade water at amounts ranging from 2.5– 40 nmol of [12C]galactitol and galactonate to which was added 1 nmol of the 13C labeled compounds. The calibrators were treated by the same procedure described above under ‘‘Sample Preparation and Derivatization.’’ The ratio of m/z 217 to m/z 220 for galactitol and m/z 319 to m/z 323 for galactonate were plotted vs the prepared galactitol and galactonate amounts. 2.5. Precision and recovery The intra- and interassay imprecision was examined in two ways. In one, a standard solution of either low or high concentration was added to a normal urine. In the second, urine of a normal subject with a low concentration and another from a galactosemic with a high concentration of galactitol and galactonate was determined. The intra-assay imprecision was calculated on the basis of 12 injections of both the low and high standards and the patient’s urine. The interassay imprecision was based on two replicate analyses performed on 16 different days over a 20-day period. Recoveries were evaluated by assay of urine before and after the addition of 2.5 and 25 nmol of [12C]galactonate and galactitol. 2.6. Calculation The calculation of galactitol and galactonate by the isotope dilution method is similar to that described in our publication for quantitation of RBC galactose-1-phosphate [15] and RBC galactitol and galactonate using the same 13 C internal standards employed here [17]. The set ratio R n of m/z 217/220 for galactitol and m/z 319/323 for galactonate (corrected for 13C in 12C compounds and 12C in the standard 13C compounds) is used to calculate amount by using the linear regression equation of the standard calculation: nmol ¼

Rn  b K

where K and b are the slope and intercept of the line relating R n to amount. Conversion to mmol/L of sample is: mmols=L ¼

nmol  103 : mil urine derivatized

2.7. Comparison with existing methods The isotope dilution method for galactitol and galactonate was compared to methods previously published from this laboratory [4,13]. The galactitol was compared to the GC method of the TMS derivative based on internal ribitol

and perseitol standards currently used at The Children’s Hospital of Philadelphia clinical laboratory [4]. Galactonate was assayed in an underivatized urine by NMR spectroscopy using a Brucker AM 400 MH wide-bore spectrometer according to our previous publication [13]. Urine was analyzed simultaneously by the GC/MS, GC and NMR methods. The levels in urine are reported as mmol/mol of Creatinine (Cr), the latter determined by the Jaffe method or by NMR analysis [13]. 2.8. Statistics Linear regression analysis was used to determine the relationship of the m/z 217/220 ratio of galactitol and the m/z 319/323 ratio of galactonate to the amount of the calibrator as well as to that of the results by the isotope dilution method to that of the established GC [4] and NMR [13] methods. The Bland – Altman analysis [19] was also used to compare the GC/MS method with the GC and NMR methods. Data are presented as the meanTSD. 2.9. Participants Random non-fasting urine samples were obtained from the clinical laboratory of The Children’s Hospital of Philadelphia. Thirty-three urines from 22 day- to 32 year-old individuals were specimens sent to the laboratory for analysis of amino or organic acids and found to have no abnormalities. They were considered ‘‘normal’’ with regard to galactose metabolism and were analyzed for galactitol and galactonate by the isotope dilution technique. Six of these specimens were analyzed by the GC method also for comparison to the new GC/MS procedure. Urine from eighteen classic galactosemic patients ranging in age from 18 days to 41 years sent to the clinical laboratory for monitoring of their galactose-restricted diet therapy were examined by both the GC [4] and GC/MS methods for galactitol and fifteen of these specimens were used to compare the isotope dilution procedure with our previously published NMR technique [13]. Urine from nineteen variant galactosemics, ages 3 days –3 years, were also compared for galactitol levels by both the GC and GC/MS methods. Fifteen were Duarte/galactosemic compounds, one was homozygous for the Duarte mutation and three were galactosemic heterozygotes. All of the galactosemics had absent RBC GALT and variants had RBC GALT levels consistent with their genotype. All had appropriate RBC isoelectric focusing results and were confirmed by mutational analysis. All of the classic galactosemics, as well as the variants, were on galactoserestricted diets. Duarte/galactosemia compounds under age 1 year were also on galactose restriction while others were on normal diets. The research was carried out with the approval of The Children’s Hospital of Philadelphia Institutional Review Board.

C. Yager et al. / Clinica Chimica Acta 366 (2006) 216 – 224

3. Results 3.1. GC analysis of galactitol and galactonate The chromatographic separation of the TMS derivatives of galactitol and galactonate when 25 nmol of each were added to a normal urine is shown in Fig. 1A. Fig. 1B shows a chromatogram of urine from a galactosemic patient demonstrating the presence of the two metabolites. The peaks were also identified by their fragmentation pattern and the corresponding position of the 13C labeled standards added to the urine. The separation corresponds to that previously observed when TMS derivatives of RBC extracts from galactosemic patients were chromatographically analyzed by a similar method [16,17]. 3.2. Linearity of calibration curve Shown in Fig. 2A and B are typical calibration curves of galactitol and galactonate over a 40 nmol range which

219

encompass concentrations known to exist in urine of normal individuals and galactosemic patients maintained on a galactose-restricted diet. The curves were, indeed, linear to 200 nmol. The linearity of the curves was reproducible for each substrate and was observed in curves constructed with each set of unknown samples over a 3month period. The r 2 of the ratio m/z 217 / 220 to galactitol and m/z 319 / 323 to galactonate was 1 T 0.01 (M T SD) and 0.99 T 0.004, respectively. The slope of the composite curve for galactitol was 0.90 T 0.05 (M T SD) and the y intercept was  0.15 T 0.42. For galactonate, the composite slope was 0.75 T 0.03 (M T SD) and the y intercept 1.40 T 0.40. The CV of the average values of each standard in the curve for galactitol varied from 2.6% to 6.4% and for galactonate from 2% to 12%. The lower limit of detection was calculated from the composite equations using the y intercept and  2 SD divided by the slopes. The resulting value for galactitol was 1.1 nmol (1.75 mmol/mol Cr) and galactonate 0.8 nmol (1.28 mmol/ mol Cr).

A

650000 600000 550000 500000 450000 400000

Galactonate

Galactitol

Abundance

350000 300000 250000 200000 150000 100000 50000

Time

10.00 11.00 12.00 13.00 14.00 15.00 16.00 17.00 18.00 19.00 20.00

Galactitol

B Abundance 8000000 7000000 6000000

Galactonate

5000000 4000000 3000000 2000000 1000000

Time

9.00

10.00 11.00 12.00 13.00 14.00 15.00 16.00 17.00 18.00 19.00 20.00 21.00

Fig. 1. Total-ion chromatograms showing the separation of TMS derivatized galactitol and galactonate. (A) A urine sample of a normal subject with the addition of 25 nmol of galactitol and galactonate. (B) A urine sample of a galactosemic patient.

220

C. Yager et al. / Clinica Chimica Acta 366 (2006) 216 – 224

A

Table 2 Reproducibility of the GC/MS method

Ratio (217/220)

40 30 20 10 0 0

10

20

30

40

Galactitol (nmoles)

Galactitol (mmol/mol Cra)

Galactonate (mmol/mol Cra)

25 50 75 100 Mean CV

230 212 213 215 217.5 T 8.43b 3.90%

163 171 148 154 159 T 10.1 6.40%

a

B

b

40

Ratio (319/323)

Urine volume (Al)

Each value is the average of 3 determinations. Mean T SD.

30 20 10 0 0

10

20

30

40

Galactonate (nmoles) Fig. 2. (A) Relationship between the ratio of the peak area of ion m/z 217 of galactitol to the peak area of m/z 220 of the internal standard and the amount of galactitol. The symbols represent single determinations. The data are representative of 12 such curves. The curve for the mean of the 12 curves is represented by the equation: y = (0.40 T 0.05)x  (0.152 T .415), where the values in parenthesis represent the mean T SD. (B) Relationship between the ratio of the peak area of ion m/z 319 of galactonate to the peak area of m/z 323 of the internal standard and the amount of galactonate. The symbols represent single determinations. The data are representative of 12 such curves. The curve for the mean of the 12 curves is represented by the equation: y = (0.75 T 0.034)x  (1.404 T .39), where the values in parenthesis represent the mean T SD.

3.3. Precision and accuracy Table 1 shows the intra-assay and interassay results. The intra-assay CV’s (calculated from results in millimole per mole (mmol/mol) of urine Cr for galactitol added to urine at low and high concentrations was 2.08% and 2.98%, respectively. For a normal urine sample at low concentration

it was 6.67%, and for a urine sample from a galactosemic patient with a high concentration it was 3.83%. For the low and high galactonate standards added to urine, the CV’s were 3.50% and 8.00%, respectively. In the normal urine sample studied, galactonate was not detectable. In the urine of the galactosemic patient, the CV for galactonate was 3.47%. In the interassay the CV for galactitol at low and high concentrations added to urine was 9.60% and 6.72%, respectively, and for galactonate it was 11.17% and 8.50%, respectively. For normal urine the CV for galactitol was 10.93% and for galactosemic urine at the high concentration it was 6.18%. The CV for galactonate was 6.91% in the galactosemic urine. Table 2 shows the reproducibility when 25 to 100 AL of urine from a known galactosemic was examined. The results for both substances were similar at all the volumes used with a CV of 3.90% for galactitol and 6.40% for galactonate. In recovery studies, when 2.5 nmol of galactitol were added to urine on 22 occasions, the average recovery was 98% T 12 (M T SD), and for 25 nmol the average recovery on 21 occasions was 105% T 8. For galactonate at 2.5 nmol the recovery in 25 assessments was 103 T 12% and for 25 nmol 111 T12% in 23 determinations.

Table 1 Assessment of the analytical precision of the isotope-dilution method Urine standards Mean T SD

Urine samples Mean T SD

CV%

CV%

Intra-assay precision Low conc (n = 12) High conc (n = 12)

35.15 T .73 429 T 12.81

2.08 2.98

Low conc (n = 12) High conc (n = 12) Interassay precision

34 T 1.19 455 T 36.55

3.50 8.00

Low conc (n = 20) High conc (n = 21)

39.71 T 3.82 429.20 T 28.86

9.60 6.72

Low conc (n = 14) High conc (n = 16)

38.14 T 4.26 475.38 T 40.35

11.17 8.50

a

Not detectable.

Galactitol (mmol/mol Cr) Low conc (n = 12) High conc (n = 12) Galactonate (mmol/mol Cr) Low conc (n = 12) High conc (n = 12) Galactitol (mmol/mol Cr) Low conc (n = 24) High conc (n = 24) Galactonate (mmol/mol Cr) Low conc (n = 24) High conc (n = 23)

5.83 T .118 254.22 T 9.74

6.67 3.83

Nda 144 T 5

3.47

6.70 T .73 249.33 T 15.4

10.93 6.18

Nda 146.39 T 10.13

6.91

C. Yager et al. / Clinica Chimica Acta 366 (2006) 216 – 224

500 400 300 200 100 0 100

200

300

400

500

mmol/mol Cr galactitol (GCMS) Difference between methods mmol/mol Cr (GC/MS-GC)

B 100 75 50 25 0 -25 -50 -75 -100

+2 SD Mean -2 SD

0

100 200 300 400 Average of two methods, mmol/mol Cr

500

Fig. 3. Relationship between the new GC/MS method and the established GC method for galactitol. (A) Linearity of values obtained by the GC/MS method vs. the GC method. (B) Bland – Altman plot.

3.4. Comparison of methods Fig. 3 shows the comparison of the GC/MS method with the GC method with ribitol and perseitol as internal standards currently used by The Children’s Hospital Metabolic Diagnostic Laboratory for urine quantitation of galactitol. Fig. 3A indicates the linear correlation of the two methods with an r 2 of 0.99. Fig. 3B is a Bland – Altman plot of the averages of the two values plotted against the differences between them. There is scatter about the mean with only two assays having a difference beyond T2 SD of the mean. Eighteen of 43 assayed by GC/MS values were more than 10% higher than the GC values while 22 differed by less than 10%. Fig. 4A shows the correlation between the GC/MS and the NMR methods for urine galactonate. This was done for 15 urine samples from galactosemic patients since the NMR method is not sensitive enough to measure galactonate in the normal or near-normal range. For values ranging from 56 to 663 mmol/mol Cr, the r 2 was 0.78. Fig. 4B is a Bland – Altman plot of the averages of the two methods plotted against the differences which shows reasonable scatter about the mean difference. In general, the NMR values tended to be higher than the levels by the GC/MS method. 3.5. Reference values The values based on randomly obtained urine specimens are shown in Table 3 for both galactitol and galactonate by the GC/MS method. The values are presented in the age groups that were delineated in our previous publication of urine galactitol values by the GC technique [4]. This has been done since the excretion in normals as well as galactosemics were known to be age dependent [4,5]. This

4. Discussion The alternate pathway metabolites of galactose, galactitol and galactonate accumulate in tissues [20,21] and are excreted in the urine of individuals with inherited defects in galactose metabolism [4,5,12 – 14]. The most frequent of these abnormalities is the deficiency of GALT with an incidence of 1 in 40,000 births. The measurement of urinary galactitol excretion has achieved some currency for assessing the biochemical results of initiating galactose-restricted diets in newborns diagnosed with GALT deficiency galactosemia and for monitoring the adherence to the diet during later life [4,5]. In contrast to galactitol, which is not metabolized (3), galactonate is an intermediate in an alternate route of galactose disposition [9,22,23]. Indeed, the excretion of these metabolites reflects two very different

A mmol/mol Cr galactonate (NMR)

0

is clearly seen in Table 3 for galactitol where the averages in normal urine decrease from 46 to 29 and 6.7 mmol/mol Cr in the <1, 1– 6 and >6 years groups. This occurs in the classic galactosemics from an average of 521 to 291 and 177 mmol/mol Cr in the same three age groups, respectively. The same age dependence is seen for galactonate excretion in normals where the averages decrease from 66 mmol/mol Cr in the < 1 year group to 41 and 11 mmol/mol Cr in the 1– 6 and >6 years group, respectively, although there is an overlap in the ranges. The same age related decrease is also observed in the averages of galactosemic urine values with apparent overlap of the ranges. A distinguishing feature of the data is the difference in the galactitol to galactonate ratio, which is below 1 in nongalactosemic urine of all age groups, while it is between 3 and over 5 in galactosemic groups.

700 600 500 400 300 200 100 0 0

100 200 300 400 mmol/mol Cr galactonate (GCMS)

500

B Difference between methods (GC/MS-NMR) mmol/mol Cr

mmol/mol Cr galactitol (GC)

A

221

300 200 100 0 -100 -200 -300

+SD Mean -SD

0

100 200 300 400 500 Average of the two methods, mmol/mol Cr

600

Fig. 4. Relationship between the GC/MS method and the established NMR method for galactonate. (A) Linearity of values obtained by the GC/MS method vs. the NMR method. (B) Bland – Altman plot.

222

C. Yager et al. / Clinica Chimica Acta 366 (2006) 216 – 224

Table 3 Galactitol and galactonate excretion in random urine samples by normals and galactosemics Age Normals <1 year (12) 1 – 6 years (14) >6 years (7) Galactosemics <1 year (3) 1 – 6 years (6) >6 years (9)

Galactitol (mmol/mol Cr)

Galactonate (mmol/mol Cr)

8 – 107b 46 T 37d 7 – 74 29 T 23 2 – 15 6.7 T 5.2

Ndc – 231 66 T 71 Nd – 124 41 T 43 Nd – 25 11 T 9.4

397 – 743 521 T193 215 – 382 291 T 70 125 – 274 177 T 48

92 – 132 113 T 20 45 – 129 88 T 35 17 – 46 32 T 10

Ra

0.70 0.62 0.56

4.55 3.12 5.56

a

R is ratio of the averages of galactitol to galactonate. Range. c Nd is <2. For calculations of the average 2 mmol/mol creatinine was used. d M T SD. b

aspects of galactose disposition. Although the focus has been on galactitol excretion as a key metabolite [4], little attention, until recently, has been paid to galactonate excretion [13]. Indeed, the monitoring of both metabolites in urine may be important for enhancing the assessment of the galactose burden in affected individuals. Various methods have been utilized to measure galactitol including HPLC [6], GC [4,5,7] and GC/MS [8]. We have routinely employed a GC method utilizing a trimethlsilyl derivative with internal standards of ribitol and perseitol for urinary analysis which, however, did not permit the simultaneous quantitation of galactonate [9]. Using TMS derivatization we devised a GC program for separating sugar phosphates for quantitation of galactose-1-phosphate in RBCs [15]. The GC program also permitted the identification of galactitol and galactonate in galactosemic RBCs [16]. With the commercial availability of UL 13C labeled galactitol, we devised a method for the accurate and sensitive quantitation of galactitol by isotope dilution mass spectrometry [17]. From the mass fragmentation pattern we were able to select a target ion m/z 217 which shifted 3 mass units to m/z 220 because of the presence of the isotope carbons. To take advantage of the chromatographic identification of galactonate in the extracts for quantification by isotope dilution MS we synthesized [UL-13C]galactonate by mild iodine oxidation of [UL-13C]galactose which is a relatively simple procedure [18]. From the mass fragmentation pattern we selected a target ion m/z 319 which shifted 4 mass units to m/z 323 due to the isotope carbon. This has permitted the simultaneous quantitation of urinary galactitol and galactonate. The procedure requires removal of urea by incubation with urease as the only aid to the chromatography without other clean up procedures. There is no interference from galactose itself since there is little

excretion of galactose by normal people ingesting milk or galactosemics on a lactose-restricted diet. Gluconate does not interfere with the galactonate quantitation. The method replaces the costly NMR method for galactonate we have utilized as a research procedure [13]. The urine isotope dilution method for both metabolites is accurate and reproducible with a CV within 10% (Table 1). Standard curves were reproducible. The reproducibility is also demonstrated in Table 2 where different volumes of urine were analyzed. The new GC/MS method for urine galactitol compares well with the current GC method used in The Children’s Hospital Metabolic Laboratory as shown by the linearity of the values and by the Bland – Altman plots. In the series of normal urine that were analyzed by the new method alone, the range for infants under age 1 year was 8 to 107 and < 2 to 78 mmol/mol of Cr by the current GC method. In the > 6 year old group, the range of the new method was 2– 15 mmol/mol Cr while by the current method it was <2 to 19 mmol/mol Cr [4]. The new GC/MS method for galactonate compared well with the samples analyzed by NMR used in our research procedure for urine from galactosemics with elevated galactonate values [13]. The normal values discerned by GC/MS cannot be compared with the NMR procedure which will not detect galactonate in the normal range. The main usefulness of quantifying urinary galactitol and galactonate is in monitoring the adequacy of diet therapy and the detection of breaks in the diet of galactosemics. As such, they are a supplement to the commonly used measurement of RBC galactose-1-phosphate [15] and may provide cogent evidence of whole body galactose metabolism. In the new GC/MS procedure the measurements were made in randomized urine specimens from non-fasting individuals, a practice we use in our clinic to monitor patients. Indeed, a fasting requirement would be difficult in view of varying times of patients’ clinic visits. Schadewaldt et al. report that in a healthy individual, there is an average two-fold difference in the galactitol excretion between a fasting and a non-fasting urine specimen [8]. This would be expected in normals depending on the milk and lactose intake in the non-fasting state. However, the measurement of galactitol is not routinely made in healthy individuals but is useful for monitoring the low lactose diet of galactosemics. We have established a range of values in random nonfasted healthy individuals. All studies of urinary [4,5,8] galactitol in normals and galactosemics have shown an agerelated decrease which appears to be satisfied by a logarithmic function [5]. In Table 3 we have used the age distribution of normals first published by Jansen et al. [7] for galactitol. The decreasing average value and range for galactitol in the normals are evident. The values compare favorably with those reported by Jansen et al. [7] and Palmieri et al. [4] using GC methods. Schadewaldt et al. using a GC/MS method [8] report a level of 3 – 14 mmol/

C. Yager et al. / Clinica Chimica Acta 366 (2006) 216 – 224

mol Cr which corresponds to our 2– 15 mmol/mol Cr in older subjects. Our values for galactonate in normals of different ages is the first to be reported. Our previous publications with the NMR method dealt only with galactosemic values. Schadewaldt et al. [14] reported values in normal fasting adults of 0.4 to 11.6 mmol/mol Cr compared to our range of <2 to 25 in random specimens. The slightly higher galactonate range in random samples of non-fasting normals may be due to prior dietary galactose which would decrease with fasting. Table 3 also shows values for both galactitol and galactonate in the urine of galactosemic subjects of various ages. The excretion of both metabolites is age-dependent as in normals. The values for galactitol reported by the new method correspond well to the values determined by GC methods reported by us [4] and others [5] and by the GC/ MS method of Schadewaldt et al. [8]. It is clear that there is an easily apparent distinction between the normal and galactosemic values. The galactonate values correspond well to those determined by NMR [13] and more recently by GC/MS where the N-(1-butyl) galactonamide pentaacetate derivative was used [14]. In the latter procedure the values for adults ranged from 24 –66 mmol/mol Cr with an average of 42 mmol/mol Cr. Our range in over age 6 years is 17 –46 mmol/mol Cr with an average of 32 mmol/mol Cr. In contrast to galactitol there is overlap in the values for galactosemics and normals. On the other hand, there is a marked difference in the ratio of galactitol to galactonate which is under 1 in normals and well above 1 in galactosemics. The reason for their divergence is unclear. When normal subjects were given an oral galactose load, galactonate excretion was greater than galactitol with the ratio of galactitol to galactonate consistently being under 1. It appears that the high ratio of galactitol to galactonate is characteristic of the galactosemic state since this occurs not only for urine values but also in RBC levels [17]. The diagnosis of galactosemic is made by measurement of RBC GALT and secondarily by DNA mutational analysis [24]. The measurement of urine galactitol and galactonate is not a diagnostic tool. Their excretion appears, however, to be a measure of the patient’s galactose burden. Their excretion is elevated even when the patients are on the best galactose-restricted diet due to a significant endogenous production of galactose [25]. Some variants such as patients who are compound heterozygotes for the Duarte and a galactosemia gene have galactitol excretion above normal and are detected in newborn screening programs of RBCs. Urines from such patients that were sent for urine galactitol analysis by the Children’s Hospital laboratory were used in this study to show the correspondence of two methods. If not placed on a galactose-restricted diet Duarte/galactosemic subjects excrete increased levels of both metabolites but not in the high range of classic galactosemics [4,26]. In conclusion, we have described an isotope dilution mass spectrometry method for the simultaneous quantitation of urine galactitol and galactonate. The method is precise

223

and accurate and would be useful in biochemical genetic laboratories for monitoring the galactose burden and diet adherence in classic galactosemic. It is also applicable to determine the galactose metabolic abnormalities in individuals variant for the GALT genotype.

Acknowledgements This work was supported by Grants DK 60768 and DK 065641 from the National Institutes of Health. We also thank Janice Malseed for her assistance in the preparation of this manuscript.

References [1] Segal S, Berry GT. Disorders of galactose metabolism. In: Scriver AL, Beaudet AL, Sly WS, Valle DThe metabolic and molecular bases of inherited diseases, vol. I. New York’ McGraw-Hill; 1995. p. 967 – 1000. [2] Jaco BSC, Sweitzer S, Dorland B. Galactitol in galactosemics. Eur J Pediatr 1995;154(Supp 1):S50 – 2. [3] Weinstein AN, Segal S. The metabolic fate of [1-14C]galactitol in mammalian tissue. Biochim Biophys Acta 1968;156:9 – 16. et al. [4] Palmieri M, Mazur A, Berry GT, et al. Urine and plasma galactitol in patients with galactose-1-phosphate uridyltransferase deficiency galactosemia. Metabolism 1999;48:1294 – 302. [5] Hutchesson AC, Murdoch-Davis C, Green A, et al. Biochemical monitoring of treatment for galactosemia: biological variability in metabolite concentrations. J Inherit Metab Dis 1999;22:139 – 48. [6] Dethy JM, Callaeri-Deveen B, Jansens M, Lansera A. Determination of sorbitol and galactitol at the nanogram level in biological samples by high performance liquid chromatography. Anal Biochem 1984;143:119 – 24. [7] Jansen G, Maskiet FAJ, Schierbeck A, Berger R, Slip W. Capillary gas chromatographic profiling of urinary, plasma, and erythrocyte sugars and polyols as their trimethylsilyl derivatives, preceded by simple and rapid prepurification methods. Clin Chim Acta 1986; 157:277 – 94. [8] Schaderwaldt P, Killius S, Damalanthan L, Hammeri H-W, Strasburger K, Webdel U. Renal excretion of galactose and galactitol in patients with classic galactosemia, obligate heterozygous parents and healthy subjects. J Inherit Metab Dis 2003;26:459 – 79. [9] Cuatrecasas P, Segal S. Galactose conversion to d-xylulose: an alternate route of galactose metabolism. Science 1966;153:549 – 51. [10] Segal S, Cuatrecasas P. The oxidation of C14 galactose by patients with congenital galactosemia: evidence for a direct oxidative pathway. Am J Med 1968;44:340 – 7. [11] Bergren WR, Ng WG, Donnell GN, Markey SP. Galactonic acid in galactosemia: identification in urine. Science 1972;176:683 – 4. [12] Gitzelmann R, Wells HJ, Segal S. Galactose metabolism in a patient with hereditary galactokinase deficiency. Eur J Clin Invest 1974; 4:79 – 84. [13] Wehrli SL, Berry GT, Palmieri MJ, Mazur A, Elsas LJ, Segal S. Urinary galactonate in patients with galactosemia: quantitation by NMR spectroscopy. Pediatr Res 1997;42:855 – 61. [14] Schadewaldt P, Hammen H-W, Stolpmann S, Kamalanathan L, Wendel U. Galactonate determination in urine by stable isotope dilution chromatography-mass spectrometry. J Chromatogr B 2004; 801:249 – 55. [15] Chen J, Yager C, Reynolds R, Palmieri M, Segal S. Erythrocyte galactose-1-phosphate quantified by isotope-dilution gas chromatography-mass spectrometry. Clin Chem 2002;48:604 – 12.

224

C. Yager et al. / Clinica Chimica Acta 366 (2006) 216 – 224

[16] Chen J, Yager CT, Reynolds RA, Segal S. Identification of galactitol and galactonate in red blood cells by gas chromatography/mass spectrometry. Clin Chim Acta 2002;322:37 – 41. [17] Yager CT, Chen J, Reynolds R, Segal S. Galactitol and galactonate in red blood cells of galactosemic patients. Mol Genet Metab 2003; 80(3):283 – 9. [18] Blair A, Segal S. The isolation of blood glucose as potassium gluconate. J Lab Clin Med 1960;55:959 – 64. [19] Bland JM, Altman DG. Statistical methods for assessing agreement between two methods of clinical measurement. Lancet 1986:307 – 10. [20] Quan-Ma R, Wells HJ, Wells WW, Sherman F, Egan TJ. Galactitol in the tissues of a galactosemic child. Am J Dis Child 1966; 112:477 – 8. [21] Rancour NJ, Hawkins ED, Wells WW. Galactose oxidation in liver. Arch Biochem Biophys 1979;193:232 – 41.

[22] Cuatrecasas P, Segal S. Mammalian galactose dehydrogenase: I. Identification and purification in rat liver. J Biol Chem 1966; 241:5904 – 9. [23] Cuatrecasas P, Segal S. Mammalian galactose dehydrogenase: II. Properties, substrate specificity, and developmental changes. J Biol Chem 1966;241:5910 – 8. [24] Dobrowolski SF, Banas RA, Suzow JG, Berkley M, Naylor EW. Analysis of common mutations in the galactose-1-phosphate uridyltransferase gene. J Mol Diagn 2003;5:42 – 50. [25] Berry GT, Moate PJ, Reynolds RA, et al. The rate of de novo galactose synthesis in patients with galactose-1-phosphate uridyltransferase (GALT) deficiency. Mol Genet Metab 2004;81:22 – 30. [26] Ficicioglu C, Yager CT, Segal S. Galactitol and galactonate in red blood cells of children with the Duarte/galactosemia (D/G) genotype. Mol Genet Metab 2005;84:152 – 9.