Diagnostic Test for Mucopolysaccharidosis. II. Rapid Quantificationof Glycosaminoglycan in Urine Samples Collected on a Paper Matrix

Transcription

1 LIN.HEM.35/1, (1989) Diagnostic Test for Mucopolysaccharidosis. II. Rapid Quantificationof Glycosaminoglycan in Urine Samples ollected on a Paper Matrix hester B. Whitley, Karl A. Draper, haryl M. Dutton, PatrIce A. Brown, Suzanne L Severson, and Laura A. France The direct 1,9-dimethylmethyleneblue (DMB) method for quantifyingsulfatedglycosaminoglycan(gag) in urine (lin hem 1989;35:374-9) has been adapted to a convenient meansfor samplecollectionandtransportas a test to identify individualswith mucopolysaccharidosis(mps) storage diseases. Results correlated moderately well (r =.85) with those of a commonly used, but more laborious,quantitative method. In studying factors to maximize differentiationof pathological from normal values, we found that GAG excretion(expressedas milligramsgag per gram creatinine)fitsa logarithmic function with respect to age and vanes markedly below age five years. This mustbe consideredindeveloping normative values and formingdiagnoses. Of 112 separate urine specimensobtainedfrom 41 MPS patientsrepresenting the major MPS diseases, glycosaminoglycanexcretion by allexceeded that for age-matchednormalindividuals.the convenienceof this method allowed us to establishthe first normativevalues for three-week-oldinfants(n = 435) found to have a mean glycosaminoglycan excretion of 179 (SD 86.3) mg of GAG per gram of creatinine. This method improvesthe diagnosticcapabilityfor those MPS diseases that have been particularlydifficultto identify (Sanfilippo s syndrome and Morquio s syndrome), and may also provide a test forotherdisorderswith previouslyunrecognizedabnormal excretionof glycosaminoglycan(e.g., mucolipidosisand acromesomelicdysplasia).most importantly, this MPS diagnostic test is unique in its suitability formass screeningof newborn infants. AdditIonal Keyphrases: heritable disorders screening neonates age-related effects reference interval The mucopolysaccharidosis (MPS) storage diseases are inborn errors of glycosaminoglycan metabolism resulting from deficiency of one of at least 11 different enzymes involved in catabolism of these sulfated heteropolysaccharides (for recent reviews, see 1_3).1 The consequent systemic accumulation of heparan sulfate, dermatan sulfate, and keratan sulfate is associated with specific clinical syndromes. Although reliable epidemiological data are not available, some estimates suggest that the cumulative incidence of MPS diseases may exceed 1 per 26 births. Recent progress in the treatment of these disorders has demonstrated that very early intervention is crucial (4-6)-a strong motivation for diagnosis at birth. However, the normal appearance of affected newborn infants belies Institute of Human Genetics and Department of Pediatrics, University of Minnesota Medical School, 42 Delaware St. S.E., Minneapolis, MN 55455; and Variety lub hildren s Hospital, Harvard St. and East River Road, Minneapolis, MN Nonstandard abbreviations: P, cetylpyridinium chloride; DMB, 1,9-dimethylmethylene blue; GAG, glycosaminoglycan; ML, mucolipidosis; and MPS, mucopolysaccharidosis. Received May 1, 1989; accepted July ii, the underlying metabolic defect. The presence of a serious MPS disease is usually not suspected until nine months to four years of age, when severe physical and (or) neurological manifestations become apparent. Genetic counseling to inform parents of the 25% recurrence risk for future pregnancies is not possible until the initial proband is identified; thus, the formation of families with two or three affected children is an especially unfortunate consequence of the typically late diagnosis of MPS disease. Because systemic accumulation probably begins prenatally, excessive glycosaminoglycans are excreted in urine at birth (7), which could provide a means of early presymptomatic diagnosis. However, technical limitations in measuring urinary glycosaminoglycans are prohibitive, and mass screening of newborns for these conditions has been limited to one early pilot study (8). Existing test methodologies for diagnosis of MPS conditions require large fresh or frozen urine specimens, which are not routinely obtainable from infants. Recently, we have developed a direct method exploiting the metachromatic reaction of 1,9-diniethylmethylene blue (DMB) dye. The method is applicable to specimens as small as.1 ml, provides quantitative results, and obviates the cumbersome and labor-intensive procedures for separating glycosaminoglycans from other constituents of urine (9). However, application of this or any other existing method to test large numbers of newborn infants is currently limited by the difficulties of obtaining, storing, and transporting fresh or frozen urine specimens. In attempting to solve these problems, we discovered that urinary glycosaminoglycans can be dried onto a paper matrix, stored or mailed, and then eluted into an aqueous solution for quantification by the direct DMB method. As demonstrated by comparison of normal and pathological specimens, this procedure surmounts the major problems inherent to methods requiring fresh or frozen urine specimens, thus providing the technical capability for mass screening of newborn infants. MaterIals and Methods DMB dye reagent. A stock solution containing.35 mmol of 1,9-dimethylmethylene blue chloride (no ; Polysciences, Inc., Warrington, PA) per liter, the lox stock solution, was made by dissolving 122 mg of dye in 1 ml of 95% ethanol, then diluting to 1 L with sodium formate buffer (ph 3.5,.2 moljl). Stored in an amber-colored bottle at room temperature, this lox stock solution could be used reliably for as long as two months. For analysis of urine specimens, we prepared a working solution of 15 Mmol/L ( 3 x ) dye solution on the day of use by appropriate dilution with the sodium formate buffer. ollection of urine specimens. For preliminary characterization of methodologies, urine specimens were collected from subjects with MPS diseases and from unaffected individuals, and stored at -2 #{176} until assay. Aliquots of urine were applied to a paper matrix, approximately 5 ml of urine per 3 cm2 of paper, either 3MM chromatography 274 LINIAL HEMISTRY, Vol. 35, No. 1, 1989

2 paper (cat. no ; Whatman International Ltd., from Sigma) and 1. ml of a 7.5 g/l solution of sodium Maidstone, U.K.) or Type 93 specimen-collection paper hydroxide. After 2 miii, we determined absorbance at 535 (cat. no. 1575; Schleicher & Schuell, Inc., Keene, NH). nm and compared it with that of creatinine standard Urine-wetted paper was laid flat on clean plastic wrap and solutions [1, 3, and 1 ung/l (no ; Sigma)l. dried overnight under ambient conditions. Specimens dried Eluates with high concentrations of creatinine were diluted on paper could be stored indefinitely in a zip-lock plastic 1-fold and then re-assayed. Duplicate aliquots of each bag at room temperature. However, incompletely dried eluate from paper were assayed, and the result was reported as the mean of these two values. specimens are rapidly discolored by fungal growth and must be discarded; thus complete drying before sealing in Data management arid statistical methods. Results were the bag is crucial. recorded on Lotus version 2A (Lotus Development orp., ambridge, MA), which we programmed to calculate For preparation of specimens at home, subjects were glycosaminoglycan and creatinune concentrations and to instructed to prepare urine specimens in a similar fashion calculate urinary glycosaminoglycan excretion as the ratio by wetting a piece of Type 93 specimen-collection paper of milligrams of glycosamunoglycan to grams of creatinine. with fresh urine, or by blotting the paper against a urinewet diaper. Alternatively, a piece of specimen-collection Lotus was also programmed to review data and to flag results at the extremes, thus indicating insufficient urine paper was inserted between the skin and the front of the sample (absorbance for glycosaminoglycan or creatinine diaper, and later removed when damp with urine (taking test reactions of <.25) and suspiciously high excretion care to avoid contamination with skin creams or stool). (e.g., results of >35 mg/g for children). We evaluated the After overnight drying, the specimen-collection paper was precision of this manual method according to guidelines placed in a plastic bag and returned to the laboratory by intended for comparing clinical chemistry devices (1). first-class U.S. mail. Three separate urine specimens were Lotus and ricket Graph version 1.2 (ricket Software, Malvern, PA) were used for graphic illustrations, and obtained from most subjects. AU MPS patients had received an appropriate diagnostic examination by us or, in a few for calculations of coefficients and other statistical parameters. cases, the parents provided sufficient information to establish the medical diagnosis. Each specimen was accompanied by a list of medications taken during the 24 h preceding urine collection. All urine specimens prepared by sub- Results jects (or parents) appeared by visual inspection to be Developmentof Standard Assay onditions satisfactory for study and were analyzed. As a means of conveniently obtaining, transporting, and Glycosaminoglycan standards. Reference solutions of storing specimens of urine from large numbers of subjects, heparan sulfate from bovine kidney (no. H9637), keratan a method of collecting urine onto a paper matrix was sulfate from bovine cornea (no. K31), chondroitin sulfate examined. To determine conditions that would allow recovery and quantification of urinary glycosaminoglycans on a type A from whale cartilage (no. 4134), dermatan sulfate (i.e., chondroitin sulfate type B from porcine skin; no. routine basis, we initially studied the reactivity of DMB 4259), and chondroitin sulfate type from shark cartilage (no. 4384), all from Sigma hemical o., St. Louis, ery of glycosaminoglycan reference solutions. dye with the paper matrix material, and evaluated recov- MO, were prepared in distilled water and stored at -2 #{176}. Inertness of paper matrix to dye. When virgin paper Elution of urine specimens. To study specimens collected matrix was soaked with 3x DMB dye reagent, both the on paper, we cut the specimen-collection paper to a convenient size of 3 cm2 (typically 1 cm x 3 cm), folded this of unreacted dye and showed no change to the violet color of paper matrix and the aqueous phase retained the blue color accordion-style, and placed it in a 15-mL polypropylene glycosaminoglycan-dmb (GAG-DMB) complexes. However, because the paper was stained blue when removed tube (Falcon no. 295; Becton Dickinson Labware, Oxnard, A). We then eluted the urine sample by adding 3.5 ml of from dye solution, it was presumed that nonspecific binding distilled water and gently rocking the sample for 1-2 could potentially interfere with quantitative GAG-DMB miii. The eluate from paper was assayed for glycosaminoglycan and for creatinune. When samples were found to method was abandoned and elution of glycosauninoglycans formation; therefore, development of a quantitative in situ have glycosaminoglycan or creatinine concentrations exceeding the limits of linearity for the test reaction, we Virgin paper matrix material was soaked in 3.5 ml of from the paper matrix was investigated. eluted an appropriately smaller piece of paper (typically distilled water, and then aliquots of the eluate were mixed 3-fl cm2) with 3.5 ml of distilled water and assayed in the with 1 ml of DMB dye reagent. There was no visible color same manner. change and no increase in absorbance (at 535 nm) over the appropriate blank. Thus, although DMB dye appeared to adsorb onto paper, the paper matrix was otherwise inert to Dye-binding assay for glycosaminoglycan. For quantffication of glycosaminoglycan, we added.5 ml of the paper matrix eluate to.25 ml of 3x dye reagent. The absorbance of the test reaction was compared with those for standard solutions of chondroitin 6-sulfate (1, 2, 3,... 1 zg/ml). Eluates with high concentrations of glycosaminoglycan were diluted 1-fold with distilled water, then assayed in the usual test reaction. Duplicate aliquots of each eluate from paper were assayed, and the result was reported as the mean of these two values. reatinine assay. reatinine was determined according to our modification of the method of Jaff#{233}. We mixed.25 ml of paper eluate with 1. ml of 2% saturated picric acid (fivefold-diluted saturated picric acid solution, no , DMB dye and did not produce a change in absorbance at the appropriate wavelength. Time course of elution of solutes from paper matrix. Further to characterize recovery of glycosaminoglycan and creatinine from a paper matrix, we applied solutions of chondroitun 6-sulfate and creatinine to 3 x 1 cm pieces of specimen-collection paper, dried them overnight, and then eluted with distilled water for specified times. reatinune dissolved quickly, being maximally eluted within 3 s, whereas chondroitin 6-sulfate was more slowly removed into solution. Both glycosaminoglycan and creatinine were maximally eluted within 1 to 2 mm, which was chosen as LINIALHEMISTRY, Vol. 35, No. 1,

3 a convenient elution interval for subsequent studies. Quantitative recovery of glycosaminoglycan from paper matrix. Quantitative recovery of different glycosaminoglycan species from specimen-collection paper was studied. Glycosaminoglycan reference solutions were applied to 3 x 1 cm paper strips, dried overnight, and then eluted with 3.5 ml of distilled water. Recovery from Type 93 specimen-collection paper was found to be proportional to the amount applied, although recovery (quantified against the usual chondroitin 6-sulfate reference solution) varied with the GAG (dermatan sulfate 14%, heparan sulfate 86%, and keratan sulfate 73%). Similar results were obtained when the same heparan sulfate, dermatan sulfate, and keratan sulfate reference solutions were applied to and eluted from Whatman 3MM laboratory chromatography paper, suggesting that quantitative recovery of glycosaminoglycan from paper matrix is a generalized phenomenon. As a means of providing an internal physiological standard, we studied the extraction of creatimne reference solutions in a similar fashion and found recovery of creatmine to be complete. Precision evaluation. Reproducibility of this method of collection, elutiom, and manual quantification was assessed by adaptation of the precision performance evaluation protocol as described for comparison of clinical chemistry instrumentation (1). For this evaluation, we prepared three concentrations of urinary glycosaminoglycan that extended over the range of anticipated normal and pathological specimens: a single, large urine sample from a normal adult was analyzed unaltered (i.e., low glycosaminoglycan) or after supplementation with additional chondroitin 6-sulfate in moderate (i.e., intermediate ) or high concentration. Multiple specimens of each concentration were dried on the paper matrix and serially tested in 4 separate repetitions or runs dome during two months (Table 1). Under these standard assay conditions the coefficient of variation (V) was found to range from 1.2% to 2.1%. It was not surprising that reproducibility was best for the intermediate and high values, insofar as our preliminary studies (not described here) adjusted reaction conditions for the range of glycosaminoglycan excretion by children and newborn infants. For other applications, reaction conditions could be readily adjusted to maximize reproducibility at other levels of glycosaminoglycan excretion as required. Table 1. Summary of PrecIsIon Performance EvaluatIon Glyco.amlnoglyc.n sxcrmlon en Low Intermediate High Msan, mg GAG/g crestinine Total Sob V, % a A precision performance evaluation was accomplished by adaptation of the protocol described for companson of clinical chemistry Instrumentation (.eference 1, with modifications according toappendix ). FortyIndependent runs, performedovera two-monthperiod,determined GAG/creatinineratio in two specimens prepared from each of the three controls (low, Intermediate, high). The measurement on each specimen was the mean of duplicateassays described In MateriaLsand Methods. b Total precision standard deviation (SD,-) as defined (1). Application to Unne Specimens omparison with P/carbazole methodology. The direct DMB method applied to paper-matrix eluate, i.e., the paper matrix/dmb method, was compared with the commonly used quantitative procedure of Bitter and Muir (11) as modified by Di Ferrante (12), i.e., overnight precipitation with cetylpyridinium chloride (P) and quantification of uronic acid with carbazole reagent. For this comparison, we analyzed by each method 33 urine specimens covering the range of normal and pathological glycoaaminoglycan concentrations. As shown in Figure 1, results by the two methods agreed closely with respect to the glycosaminoglycan/creatmnine ratio (r =.85). Rapid quantification of un nary glycosaminoglycan excretion in MPS. As shown in Table 2, a relatively large number of normal subjects and MPS patients could be studied, owing to the convenience of obtaining and transporting specimens with the paper-matrix system. Glycosaminoglycan excretion in normal individuals was highly variable with respect to age. As illustrated in Figure 2, trials of several curve-fitting calculations suggested that the data best fit a logarithmic function, with a marked inflection near age five years. Inspection of the results for measured glycosaminoglycan excretion by the group of MPS patients (Table 2) suggests a similar trend, which is somewhat masked by the very high concentrations of glycosaminoglycan. We presume that the trend in both groups (normal subjects and MPS patients) is attributable to age-related differences in creatinine excretion and to increased proteoglycan turnover during early bone growth. This graphic illustration serves to emphasize the need to rely upon age-specific normative values, especially for individuals younger than five years, for whom normal values vary greatly. Of 112 separate urine specimens obtained from 41 MPS patients representing each of the major MPS diseases, all had values markedly greater than age-appropriate controls (Table 2). Urinary glycosaminoglycan excretion was typically five- to 1-fold above that for age-matched normal individuals, thus pathological values were readily distinguished. m to #{149} ( S a). s I U A = P/arbazole Method (mg Uronic Acid/g reatinine) FIg. 1. omparison of the new paper matiix/dmb method with the standard P/carbazole method Thirty-threespecimens (fromnormalsubjects and frompatientsaffected with MPS diseases) were tested by both methods 276 LINIALHEMISTRY,Vol.35, No. 1, 1989

4 Table 2. Rapid QuantIfIcation of Glycosaminoglycan Excretion In UrIne Samples ollected, DrIed, and MaIled on a Paper MatrIx Giycosaminoglycan excretion, mg GAG/g creatinine Patient Age at test, y Individual samples Mucopolysaccharidosis type I, Hurler s syndrome (and variantsp 1 NB.59 2 AN.62 3 ED TN JD E BB JM ST KM AS ED BO ZH DK (type I-S) HF AW JW TK (type I-HS) HF (type I-HS) GD (type I-HS) SD (type I-HS) PR (type I-S) Mucopolysaccharidosis type II, Hunter s syndrome 1ML 2.4 2AS BE 3.7 4BD LO JB 7.27 Mucopo!ysaccharidosis type Ill, San filippo s syndrome 1 MS (type Ill-B) 2 RR (type Ill-A) 3 BL (type Ill-A) 4 D (type Ill-) 5 JD (type Ill-B) 6 EN 7 B (typeill-b) 8 AM (type Ill-A) Mucopolysaccharidosis type IV, Morquio s syndrome 1 NV (type IV-A) 2 KD Mucopolysaccharidosis type VI, Maroteaux-Lamy syndrome 1TW 2 JW Mucopolysaccharidosis type VII, Sly s syndrome 1 RE Non-mucopolysaccharidosis skeletal dysplasiat 1 SK, acromesomelicdysplasia 2 O, mucolipidosis type III Normal children and adults 1 SK 2 SK 3 AE 4 ED , , 1517, , 961, ,354, ,568, , 1813, , 1613, , 1133, , 6, ,77, , 864, , 4246, , 19,11 29, 55, , 1493, , 682, , ,4,41 113,263, ,226, ,143, , 5, , 595, 482k 454, 331, ,423 76, 619,416 56, 425, b 484, ,246, , 193,16 119,344, ,247,32 163,276, ,314, ,97,19 26,32,21 97, 745, , 733, , 1111, , 89, 43 67, 5, 55 87, ,195 99, 134, Mean continued on nextpage LINIALHEMISTRY, Vol.35,No. 1,

5 Patient Normal children and adults (continued) 5 MO 6 RD 7 BW 8 AK 9 W 1 JW 11 Sw 12 MA 13 TA 14 SA 15 SS 16 MM 17 W Individual samples Patients with MPS type I had Hurter s syndrome (MPS type I-H), except forthose with Scheies syndrome (type l-s) and Hurler-Scheie syndrome (type I-HS). Aliquot of a 24-h urine collection was dried onto paper matrixin the laboratory (in contrast to all other specimens, which were preparedat home from a single-void grab urine specimen, dried overnight, and then transported to the laboratory by mall). Two subjects with nori-mps types of skeletaldysplasiahad somewhat greater glycosaminoglycan excretion but were readily distinguishable from the more markedly increased amounts characteristic of MPS disease. Table 2. Age at test, y ontinued Glycosaminoglycan excretion, mg GAG/gcreatinlne ,4,36 21,24,23 39, 36 19, 2 33, 32 21, , 7 5, 6 6, 9 8, 15 Normalmean ± SD Normalrange (n) Mean ± (17) Urinary glycosaminoglycan excretion by normal infants. To evaluate the applicability of this method for testing large numbers of infants, and to establish the range of glycosaminoglycan excretion in this population, we applied the paper matrix/dmb method to a single specimen collected from each of 435 infants. These specimens were obtained at three weeks of age. As seen from the frequency distribution (Figure 3), this population of presumed-normal three-week-old infants had a mean glycosaminoglycan excretion of 179 (SD 86.3) mg of GAG per gram of creatinine. Discussion Inadequacies of existing methods. For any screening test, sample collection must be simple and the analytical technique must be inexpensive, sensitive, and reliable (13). Despite numerous studies describing putative screening tests for MPS diseases, no such test currently is routinely applied to any large population because, in part, no test fulfills these requisites. A review of available methods for testing for MPS diseases is characterized by inadequacies (14), especially as they might be applied to the presymptomatic evaluation of large numbers of newborn infants (Table 3). Most existing tests require a relatively large urine specimen, 1-1 ml, which cannot be routinely obtained from infants (13). For most tests, results are either qualitative or semiquantitative and are subject to limited interpretation. Only a few methods yield quantitative results from which positive/negative limits can be assigned, thus permitting convenient definition of acceptable falsepositive and false-negative error rates based on population data and cost-effectiveness considerations (2). Existing quantitative methods are labor-intensive and correspondingly expensive and impractical for mass screening of newborns. Only one test, the direct DMB method we previously reported (9), surmounts all of these problems. Importantly, all existing tests, including the direct DMB method, require a fresh or frozen-stored liquid urine specimen. This remains a paramount impediment to actual implementation of a mass newborn screening program. However, as developed in this study, application of the direct DMB method to a convenient sample collection system provides a solution to this problem. Glycosaminoglycan-binding to paper matrix. The convenience in obtaining, storing, and transporting urine specimens on a paper matrix was first demonstrated by Berry et al. (21). The system was subsequently adapted to screening newborn infants for other metabolic diseases (22) and is now used routinely in the few existing urine-based newborn screening programs for metabolic diseases (13, 2). The difficulty in extending this strategy to testing for MPS diseases is related to recovering glycosaminoglycan from the paper matrix, and to obtaining quantitative results. In reviewing the Berry system for collecting urine specimens on paper, Scriver (13) suggested that urine-impregnated paper could be subjected to a variety of in situ spot tests on specimen-collection paper. He also suggested that urine might be eluted from urine-impregnated paper for analysis in solution. Several disorders can be identified from such dried urine specimens, but to the best of our knowledge, quantitative elution of glycosaminoglycan from specimencollection paper has never been reported. Whereas the ability to recover amino acids, organic acids, creatinine, and other metabolites from paper matrix has been well established, the difficulties in eluting glycosaminoglycan are implicit in several existing laboratory methods that actually depend upon tight binding of glycosaminoglycan to paper matrix material. For example, the toluidine blue spot test (7, 23) or the Berry spot test requires that measured aliquots of urine be applied to a paper matrix, dried, and then reacted in situ with toluidine blue. Glycosaminoglycan-dye complexes remain bound to the paper matrix to permit satisfactory visual observations. Analogously, the alcian blue spot test (24) requires that urinary glycosarninoglycan is similarly applied to and dried on a paper matrix. The bound glycosaminoglycan is then reacted in situ with dye. Positive tests are those for 278 LINIAL HEMISTRY, Vol.35,No. 1, 1989

6 3.c X E ) 5 (5 ).) > Subject Age y Fig. 2. Logarithmiccurve-fit of glycosaminoglycanexcretionwith respectto age (17 normalindmduals) { 5 E 3 3 U- 2 been impregnated with azure A dye. After excess dye and other soluble components are washed away, the colored glycosaminoglycan-dye complexes that remain bound to the paper matrix are detectable as a positive result. On the basis of these observations, it was apparent from the outset of our work that urinary glycosaminoglycans are readily bound to paper matrix (possibly by the same cellulosic hydrogen-bonding that is responsible for the basic integrity of paper matrix). It seemed unlikely that urinary glycosaminoglycans dried on a paper matrix could be recovered for quantification in the direct DMB method. To investigate the potential for measuring urinary glycosaminoglycan excretion with this system, we initially attempted to measure glycosaminoglycans in situ on paper matrix by directly adding dye reagent solution. However, we were discouraged from pursuing an in situ test when we observed the blue staining of paper matrix material, which we presumed would interfere with the quantification. Elation of glycosaminoglycan from paper matrix. As an alternative approach to developing an in situ method, we investigated the potential for eluting dried glycosamino- glycan back into solution for quantification with the direct DMB method. It was quickly determined that various salts alter the formation of GAG-DMB complexes (data not 6 i:ir,t shown), and that the stability of GAG-DMB complexes in 1 kir solution would confound attempts at glycosaminoglycan quantification if variations in ionic strength were used to remove glycosaminoglycans from the paper matrix. How- 4 ever, we found that simple distilled water is a satisfactory 2 P1ElEl / eluent. Glycosaminoglycan reference solutions that were applied to paper and dried in the usual fashion, then eluted in distilled water, could be measured with consistent and 411E1I acceptable analytical recovery. reatinine was also quantitatively accounted for in distilled water and, as demon- El 1ElEl12 strated by numerous other studies (9, 19, 26-28), could Jj 1J jj serve as an internal physiological standard against which t ;663 Glycoeasninoglycan Excretion (mg GAG/g reattnine) Fig. 3. Frequencydistributionof glycosaminoglycanexcretionby newborn infants (n = 435) which glycosaminoglycan-dye complexes are visibly retained on the paper matrix after the paper matrix is thoroughly rinsed with glacial acetic acid and then with water. For Ames MPS Papers spot test (25), a urine specimen is added to squares of paper matrix that have to assess glycosaminoglycan excretion. To evaluate the reproducibility of this method, and to assess the sources of methodologic variation, each of two technicians (one experienced in the technique and one a novice) accomplished 2 runs of the test on three control urine solutions. The variation was acceptable, the V s12% in the range of values expected for children and newborn infants. Application to normal and MPS subjects. Most quantitative tests, including the commonly used method of Bitter and Muir (11, 12) and a recent adaptation involving the Table 3. UrIne Screening Tests for ompared False-positive FsIse-n.gatlve rats rate Smen rsqulrsment, Testing methodology % ml Type of results DirectDMB method Adjustable Adjustable.1 Quantitative Toluldineblue spot(berry spottest) 1 (15) 31.7 (16).4 Semiquantitatlve Alcianblue spot test 25 (17).(17) 1. Qualitative Acidalbuminturbidity test, gross 6.3 (15) 1 (17) 1. Semiquantitatlve Acid albuminturbiditytest, quantitative 17.2 (15) 1 (17) 1. Quantitative etyltrimethylammoniumbromide(tab) test 42 (16). (18) 5. Qualitative etylpyridiniumchloridecitrate(p) turbiditytest 32 (17). (17) 2. Quantitative MPS Papers spot test 34 (16). (18) 1 drop Qualitative MPSPapers spot test, semiquantitativemodification 2.9 (19) 3.4 (19) 1 drop Semiquantitative a Reference nos. given in parentheses. LINIALHEMISTRY, Vol. 35, No. 1,

7 DMB reagent (29, 3), rely upon quantification of precipitable glycosaminoglycans. In contrast, the direct DMB method (and this adaptation to the paper matrix collection system) should measure all sulfated glycosaminoglycans, including any relatively small species that might escape precipitation. This would be especially important if some MPS patients excrete smaller species that might not be detected by precipitation methods. For this reason, the direct DMB method may better discriminate normal from MPS specimens as suggested by our earlier comparison of the two methods (see reference 9, Figure 3). Nevertheless, our comparison of the present method with the cetylpyridimum chioride/carbazole method by simultaneous testing of 33 specimens (Figure 1) showed a good correlation of quantitative results (r =.85). In a pilot study, results obtained with the new method revealed that urinary excretion of GAG, when assessed in terms of the ratio of GAG to creatinine, is highly age dependent, both in normal subjects and in patients with MPS diseases. Variation with respect to age has been noted by other investigators using different test methods (19,28) but was presumed to be due to the age-dependent excretion of glycosaminoglycan; this was described as a simple linear relationship with respect to age (28). In contrast, onr attempts at curve-fitting suggested that this variation in normal individuals best fits a logarithmic function (Figure 2), probably owing to changes in creatinine excretion as well as to increased proteoglycan turnover during growth. Because the relationship is logarithmic, with a dramatic inflection during early childhood, it must be recognized that attempts at newborn screening would be especially dependent upon age-specific normative values. This report presents the first such data for a large number of newborn infants (Figure 3), a study that was feasible only owing to development of the paper matrixfdmb method of specimen collection, elution, and quantification. The use of single-void grab urine specimens from normal individuals has been criticized because of sampleto-sample changes in the type of glycosaminoglycan excreted (31) and circadian periodicity in excretion (32). The ease of sample collection and analysis with the present method permitted us to study multiple single-void specimens to address this issue. Despite this sample-to-sample variation in normal individuals, results for MPS patients were consistently different from the age-specific normal reference interval. False-negative results for patients with Sanfilippo s syndrome may be a relatively common problem with existing methods (33). The present test detected somewhat lower values of glycosaminoglycan excretion in patients with Sanfilippo s syndrome compared with other MPS conditions. However, values were markedly increased compared with those for age-appropriate controls, and therefore were easily distinguished from normal. The importance of age-specific comparison is particularly acute for detection of Morquio s syndrome. Humbel and Etringer (34) reported a visual color-change screening test involving the use of DMB dye, but they observed that patients with Morquio s syndrome gave many false-negative results. In contrast to their observations, we found the DMB dye to be highly reactive with keratan sulfate. Furthermore, our new quantitative method should identi1r most, if not all, patients with Morquio s syndrome when age-specific normative data are used. As seen in Table 2, patients with Morquio s syndrome had values that were not distinguishable from the overall normal range. However, compared with age-specific values, affected patients were readily distinguished. One affected child, 3.26 years of age, had values exceeding the age-specific upper limit of normal by severalfold. Similarly, a 27-year-old patient with Morquio s syndrome had a mean value of 26 mg/g (n = 3) that exceeded by more than twofold the age-specific normal range (for age >2 years, the normal reference interval is 6-12 mg/g). The quantitative distinction between some patients with Morquio s syndrome and normal individuals may be less than for other MPS conditions, but our test does permit differentiation that is not possible by other methods. Diagnosis of non-mps skeletal dysplasias. Application of this method may also extend to other conditions previously undetectable by urine testing. We have recently studied one patient with mucolipidosis (ML) type III who, by usual methods (i.e., negative Berry spot test), was found to have normal urinary glycosaminoglycan excretion. However, by our method, this patient showed somewhat increased GAG excretion (mean of three determinations at 4.14 years of age: 57 mg/g). On the basis of this observation, we propose that I-cell disease (MLII) and pseudo-hurler polydystrophy (ML III), which present with MPS clinical features, might be considered partial defects in glycosaminoglycan catabolism resulting in minimally increased glycosaminoglycan excretion-not previously appreciated owing to the limitations of existing test methods. Analogously, a 2.9-year-old patient with acromesomelic dysplasia (35, 36) had values twofold those for age-appropriate norms (mean of three specimens was 7 mg/g; the range for normal subjects two to five years of age was mg/g). At least one patient with this rare non-mps disorder has had a positive Berry spot test (31). Our patient studied here had normal glycosaminoglycan excretion as determined by the method of Bitter and Muir (11). Thus, the present test may permit earlier recognition and differentiation of non-mps diseases whose presumed aberrant proteoglycan metabolism is revealed by minimally increased GAG excretion. Normal glycosaminoglycan excretion by newborns. Other studies have demonstrated that urinary glycosaminoglycan excretion is increased in MPS patients even at birth (7). Thus, the approach of evaluating newborn infants at three weeks of age, which has become a general practice for other metabolic diseases (13,2), might be feasible for MPS diseases. Age-specific normative data will be required to diagnose newborns. For the first time, such data were obtained by testing 435 infants with this new method (Figure 3). In summary, the unique contribution of this study was the development of a method for recovering and precise quantification of very small amounts of glycosaminoglycan in a fashion applicable to the paper matrix collection system. The method of analysis is rapid with a specimen turnaround time of <2 h. ommon laboratory instrumentation is used, and the reagents are readily available, inexpensive, and stable at room temperature without special precautions. In combination, these features make this method readily adaptable to automation. Nevertheless, manual techniques were used in pilot evaluations to test a variety of normal subjects and patients affected with each of the MPS diseases. As demonstrated in this study, urine specimens can be readily collected by nontechnical personnel (e.g., parents) and conveniently transported to a central testing facility by regular mail. In contrast to most existing 28 LINIAL HEMISTRY, Vol. 35, No. 1, 1989

8 screening methods, which are semiquantitative at best, this is a precisely quantitative technique and thereby provides a means for defining specific positive/negative limits based on population data, and the technique can be varied depending upon the requirements and facilities of any particular screening program. With this novel method, we anticipate that all newborns could be tested for MPS disease on a regular basis. Identification of newborn infants with MPS conditions will obviate the diagnostic conundrums presented by MPS disease (e.g., recurrent ear infections, airway obstruction, macrocephaly, linear growth failure, etc.), facilitate more efficient and specific clinical management, enable genetic counseling before the next pregnancy, and facilitate presymptomatic treatment. This work was supported by grants from the Preston Arnold MPS Research Fund of the Minnesota Medical Foundation (with special thanks to hris Padden and Apollo Jr. High School), from the National MPS Society through a contribution by the family of Kirby Lastinger, and from the Bone Marrow Transplantation Research Fund, and by a special equipment grant from the Minnesota Medical Foundation. For contribution of urine specimens we thank our patients and their families, and thank our colleagues Drs. Peter occia, Robert J. Gorlin, P. Jean Henslee, John H. Kersey, Martin Klemperer, William Krivit, Leonard. Langer, Norma K.. Ramsay,. Ronald Scott, Michael Y. Tsai, Phyllis I. Warkentin, and Ronald Whitley. References 1. Neufeld EF, MuenzerJ. The mucopolysaccharidoses. In: Scriver R, Beaudet AL, Sly WS, Valle D, eds. The metabolic basis of 1inherited disease. 6th ed. New York: McGraw-Hill, 1989: Spranger J. The mucopolysaccharidoses. In: Emery AEH, Rimom DL, eds. Principles and practice of medical genetics. New York: hurchill Livingstone, 1983: Muenzer J. Mucopolysaccharidoees [Review]. Adv Pediatr 1986;33: Hobbs JR, Hugh-Jones K, Barrett AJ, et al. Reversal of clinical features of Hurler s disease and biochemical improvement after treatment by bone marrow transplantation. Lancet 1981;ii: Krivit W, Pierpont ME, Ayaz K, eta!. Bone marrow transplantation in the Maroteaux-Lamy syndrome (mucopolysaccharidoeis type VI): biochemical and clinical status 24 months after transplantation. N Engl J Med 1984;311: Whitley B, Ramsay NK, Kersey JH, Krivit W. Bone marrow transplantation for Hurler syndrome: assessment of metabolic correction. Birth Defects 1986;22: Berry HK. Screening for mucopolysaccharide disorders with the Berry spot test. lin Biochem 1987;2: Sabater J, Villalaba M, Maya A. Mass screening newborns for mucopolysaccharidoses. lin Genet 1973;4: Whitley B, Ridnour MD, Draper KA, Dutton M, Neglia JP. Diagnostic test for mucopolysaccharidosis. I. Direct method for quantifying excessive urinary glycosaminoglycan excretion. lin hem 1989;35: NLS Tentative Guideline EP5-T. User evaluation of precision performance of clinical chemistry devices. Vilanova, PA: National ommittee for linical Laboratory Standards, June Bitter T, Muir HM. A modified uronic acid carbazole reaction. Anal Biochem 1962;4: Di Ferrante NM. The measurement of urinary mucopolysaccharides. Anal Biochem 1967;21: Scriver R. Screening newborns for hereditary metabolic disease. Pediatr lin North Am 1985;12: Pennock A. A review and selection of simple laboratory methods used for the study of glycosaminoglycan excretion and the diagnosis of the mucopolysaccharidoaes. J lin Pathol 1976; 29: Shih VE. Mucopolysaccharidoses. In: Shih V, ed. Laboratory techniques for the detection of hereditary metabolic disorders. leveland: R Press, 1973: Rezvani I, ollipp PJ, Di George AM. Evaluation of screening tests for urinary mucopolysaccharides. Pediatrics 1973;57: Pennock A, Mott MG, Batstone GF. Screening for mucopolysaccharidoees. lin him Acta 19727: Rezvani I, ollipp PJ, Di George AM, Punnet HH. Screening tests for the detection of mucopolysaccharide (MPS) disorders: evaluation of a new test. Pediatr Rae 1972;6: Huang K-, Sukegawa K, Orii T. Screening test for urinary glycosaminoglycans and differentiation of various mucopolysaccharidoses. lin hixn Acta 1985;151: Scriver R. Population screenings. report of a workshop. Prog lin Biol Rae 1985;163B: Berry HK, Sutherland B, Guest GM, Warkany J. Simple method for detection of phenylketonuria. J Am Med Assoc 1958;167: Berry HK. Procedures for testing filter paper. lin hem 1959;5:63-8. urine specimens dried on 23. Berry HK, Spinanger J. A paper spot test useful in study of Hurler s syndrome. J Lab lin Med 1987;55: arson JAJ, Neil DW. Metabolic abnormalities detected in a survey of mentally backward individuals Arch Dis hild 1962;37: of Northern Ireland. 25. Berman ER, Vered J, Bach G. A reliable spot test for mucopolysaccharidoais. lin hem 1971;17: Manley G, Severn M, Hawksworth J. Excretion patterns of glycosaminoglycans and glycoproteins in normal human urine. J lin Pathol 1968;21: Pennock A. A modified screening test for glycoaaminoglycans. J lin Pathol 1969;22: Huang K-, Sukegewa K, Orii T. Glycosaminoglycan excretion in random samples of urine. lin him Acta 1985;151: Panin G, Naia 5, Dall Amico R, et a!. A simple spectrophotometric method for the quantitative determination of urinary glycosaminoglycan excretion. lin hem 1986;32: oppa GV, atassi, Gabrielli, eta!. linical application of a new simple method for the identification of mucopolysaccharidoses. Helv Paediatr Acts 1987;42: Di Ferrante N, Lipecomb 115. Urinary glycosaminoglycans versus creatinine excretion: a used and abused parameter. lin him Acts 197;3: Scott JE, Newton Ill. The recovery and characterization of acid glycosaininoglycans in normal human urine. Influence of a circadian rhythm. onnect Tissue Rae 1975;3: Rattenbury JM, Worthy E, Allen J. Screening tests for glycosaminoglycans in urine: experience from regional interlaboratory surveys. J lin Pathol 1988;41: Humbel R, Etringer S. A colorimetric method for the determination of sulfated glycoaaminoglycans. Rev Roum Biochim 1974;11: Langer LO, Beals RK, Solomon IL, et al. Acromesomelic dwarfism: manifestations in childhood. Am J Med Genet 1977;1: Langer LO, Garrett RT. Acromesomelic dysplasia. Pediatr Radiol 198;137: LINIAL HEMISTRY, Vol. 35, No. 1,