Production and certification of an enzyme reference material for pancreatic α-amylase (CRM 476)

ELSEVIER

Clinica Chimica Acta 251 (1996) 145-162

Production and certification of an enzyme retqerence material for pancreatic s-amylase (CRM 476) G e m m a Gubern a, Francesca Canalias a, F. Javier Gella *a, Elizabeth Colinet b, Christos Profilis b, Derek H. Calam c, Ferruccio Ceriotti d, J. Dufaux e, Anthony G. Hadjivassiliou f, Jean Marc Lessinger g, Klaus Lorentz h, Anne Vassault i "Departament de Bioqu~nica i Biologia Molecular, Unitat de Bioqufmica-Medicina, Universitat Autonbma de Barcelona, Edifici M, 08193-Bellaterra, Barcelona, Spain bBCR, European Commission, 200, rue de la LoL 1049-Brussels, Belgium CNational Institute for Biological Standards and Control Potters Bar, London, UK dlstituto Scientifico, H.S. Raffaele, Milano, Italy eAssociation Pharmaceutique Beige, Brussels, Belgium I Metaxas Cancer Hospital Piraeus, Greece gUniversitb Louis Pasteur, lllkirch, France hMedizinische Hochschule Lfibeck, Lfibeck, Germany iHbpital Necker-Enfants Malades, Paris, France

Received 9 January 1996; accepted 29 January 1996

Abstract

We describe the preparation of a lyophilized material containing purified human pancreatic a-amylase and the certification of its catalytic concentration. The enzyme was purified from human pancreas by ammonium sulphate precipitation and chromatography successively on DEAE-Sephacel, CM-Sepharose and Sephadex G-75. The purified enzyme Nonstandard abbreviations: IFCC, International Federation of Clinical Chemistry; BCR, Bureau Communautaire de Reference; CRM, certified reference material; NIBSC, National Institute for Biological Standards and Control; PNP-G7-B, 4,6-benzylidene-4-nitrophenyl-~-o-maltoheptaoside; CNP-G3, 2-chloro-4-nitrophenyl-:c-D-maltotrioside; MES, 2-morpholinoethanesulfonic acid; PIPES, piperazine-N,N'-bis[2-ethanesulfonic acid]; SDS-PAGE, sodium dodecyl sulphate-polyacrylamide gel electrophoresis; PAGE, polyacrylamide gel electrophoresis; PMSF, phenylmethylsulfonyl fluoride; TPCK, N-tosyl-L-phenylalanine chloromethyl ketone; TLCK, N-tosyI-L-lysine chloromethyl ketone. *Corresponding author, Tel.: 35 81 15 75; Fax: 35 81 15 73.

0098-8981/96/$15.00 © 1996 Elsevier Science B.V. All rights reserved SSDI S0009-8981(96)06302-4

146

G. Gubern et al. / Clinica Chimica Acta 251 (1996) 145-162

had a specific activity of 52.9 kU/g protein and was >99% pure on polyacrylamide gel electrophoresis. Only trace amounts of lipase and lactate dehydrogenase were detected in the purified fraction. The purified pancreatic a-amylase had a molar mass of 57 500 g/mol and an isoelectric point at 7.1. The material was prepared by diluting the purified or-amylase in a matrix containing PIPES buffer 25 mmol/1, pH 7.0, sodium chloride 50 mmol/l, calcium chloride 1.5 mmol/1, EDTA 0.5 mmol/1 and human serum albumin 30 g/l, dispensing in ampoules and freeze-drying.The ampoules were homogeneous and the yearly loss of activity on the basis of accelerated degradation studies was less than 0.01% at -20°C. The certified value for or-amylase catalytic concentration in the reconstituted reference material is 555 U/1 _+ 11 U/I when measured by the specified method at 37°C. The material can be used to verify the comparability of results from different laboratories, for intra-laboratory quality control or for calibration of a-amylase catalytic concentration measurements.

Keywords: Referencematerial; Enzyme activity; Standardization

1. Introduction Considerable efforts have been devoted in the last decade to develop standardized conditions for the measurement of enzyme catalytic concentration. The reference methods specified by the Expert Panel on Enzymes of the International Federation of Clinical Chemistry (IFCC) are intended to provide criteria of accuracy against which other methods may be judged [1]. However, experience has shown that even though reference methods were described in detail, reference enzyme preparations would be required to harmonize the enzyme measurements [2,3]. In 1980, the Bureau C o m m u n a u t a i r e de Reference (BCR) of the European Communities supported a working group of clinical enzymologists in preparing and evaluating reference preparations of enzymes of diagnostic importance [4]. U p to now six enzyme reference materials have been prepared and are available with certified values for their catalytic concentrations [5-10]. The measurement of the catalytic concentration of s-amylase (EC 3.2.1.1) in serum is the most frequently used enzymatic test for the diagnosis of pancreatic disease [11,12]. Elevated serum s-amylase has also been reported in patients with mumps, renal diseases and abdominal disorders such as cholecystitis [13]. Usually, total serum s-amylase activity is measured without discriminating the organ sources, although during the last years much interest has been focused on the possible diagnostic use of s-amylase isoenzyme measurement, especially of the pancreatic isoenzyme [14]. The existing wide variety of methods for the determination of the catalytic concentration of serum s-amylase and the absence of a reference

G. Gubern et al. I Clinica Chimica Acta 251 (1996) 145-162

147

method c,auses a wide interlaboratory dispersion of values, as demonstrated by external quality assurance surveys [15]. A reference preparation of human pancreatic a-amylase could assist in efforts to harmonize the results of determinations carried out with different methods, in developing new assay methods (especially a reference method) and in assessing interlaboratory performance. We de,;cribe here the development of a pancreatic a-amylase reference material and the certification of its catalytic concentration in the reconstituted lyophilized material. This Certified Reference Material is now available as CRM 476. 2. Materials and methods

2.1. Reagents Monoclonal antibody specific for human pancreatic a-amylase (clone 6103) was from Oy Medix Biochemica (Kauniainen, Finland); DEAESephacel, CM-Sepharose CL 6B and Sephadex G-75, Mono-Q HR 5/5 and Superose 12 HR 10 /30 were from Pharmacia (Uppsala, Sweden). Human serum albumin, HEPES, 2-morpholinoethanesulfonic acid (MES), piperazine-N,N'-bis[2-ethanesulfonic acid] (PIPES), benzamidine, phenylmethylsulfonyl fluoride (PMSF), N-tosyl-L-phenylalanine chloromethyl ketone (TPCK), N-tosyl-L-lysine chloromethylketone (TLCK), 4-nitrophenyl-a-glucopyranoside, protease from Streptomyces type IV, pepsin and azocasein were purchased from Sigma Chemicals (St. Louis, MO, USA). 2-Chloro-4-nitrophenyl-a-D-maltotrioside was from Genzyme Diagnostics (Kent, UK) and twice crystallized 2-chloro-4-nitrophenol with a melting point of 110.7°C was kindly provided by Dr. K. Lorentz (Medizinische Hochschule Liibeck, Liibeck, Germany). Pepsin was from Boehringer Mannheim (Mannheim, Germany) and chymotrypsin from Merck (Darmstadt, Germany). 2.2. Assay methods a-Amylase activity was determined with 4,6-benzylidene-4-nitrophenyla-o-maltoheptaoside as substrate and a-glucosidase (EC 3.2.1.20) and glucan 1.,4-a-glucosidase (EC 3.2.1.3) as auxiliary enzymes using a commercial kit from Knickerbocker (Barcelona, Spain). a-Amylase catalytic concentration was measured in the certification exercise with 2-chloro-4-nitrophenyl-a-o-maltotrioside as substrate. The method used was a modification of that described by Winn-Deen et al. [16]. a-Amylase acts on the substrate liberating 2-chloro-4-nitrophenol which can be measured at 405 nm. The reagent composition was as follows: 50 mmol/1 MES buffer pH 6.28 (37°C), 2.25 mmol/1 CNP-G3, 900

148

G. Gubern et al. / Clinica Chimica Acta 251 (1996.) 145-162

mmol/1 potassium thiocyanate, 300 mmol/l sodium chloride and 5 mmol/1 calcium chloride. The sample, 0.01 ml, was mixed with 1.0 ml of reagent. The measuring temperature was 37°C. After a delay of 30 s, the absorbance of the mixture was monitored at 405 nm for 120 s. The purified enzyme was examined for the following possible contaminating enzymes: aspartate aminotransferase (EC 2.6.1.1), alanine aminotransferase (EC 2.6.1.2), creatine kinase (EC 2.7.3.2), alkaline phosphatase (EC 3.1.3.1), L-lactate dehydrogenase (EC 1.1.1.27) and v-glutamyltransferase (EC 2.3.2.2) using commercial kits from Biosystems (Barcelona, Spain); triacylglycerol lipase (EC 3.1.1.3), trypsin (EC 3.4.21.4) and chymotrypsin (EC 3.4.21.1) were measured using kits from Boehringer Mannheim; a-glucosidase with 4-nitrophenyl-~-glucopyranoside as substrate [17]; and protease activity with azocasein as substrate [18]. Protein was determined by the method of Bradford [19] with bovine serum albumin as standard.

2.3. Purification of pancreatic a-amylase Pancreatic a-amylase was purified from human pancreas according to a modification of the procedure described by Sampson et al. [20]. Human pancreas was obtained at autopsy from subjects free from pancreatic and transmissible diseases and stored at - 20°C until used. Sera of the subjects were found to be free of hepatitis B antigen and anti-HIV antibodies. The purification procedure was as follows: Tissue (64.8 g) was homogenized in 650 ml of 200 mmol/l phosphate buffer pH 6.5 containing protease inhibitors (0.1 mmol/l PMSF, 0.5 mmol/1 benzamidine, 0.1 mmol/l TPCK and 0.1 mmol/1 TLCK) in a Polytron homogenizer (Kinematica, Littau, Switzerland) for 5 min at 4°C. The homogenate was centrifugated at 6000 x # for 20 min at 4°C and the supernatant obtained was adjusted to pH 7.0. The supernatant was then precipitated with 2.14 mol/l ammonium sulphate and centrifugated at 6000 x 9 for 20 min at 4°C. After centrifugation the precipitate was redissolved in 85 ml of 40 mmol/l phosphate buffer pH 6.5, and dialyzed overnight against the same buffer. The dialyzate was then chromatographed on a 2.6 x 70 cm column of DEAE-Sephacel equilibrated with the same buffer. The active fractions were concentrated to a final volume of 30 ml by ultrafiltration using an Amicon PM10 membrane (Amicon, Danvers, MA, USA) and applied to a 2.6 x 70 cm column of CMSepharose-CL 6B equilibrated with the same buffer, a-Amylase was not retained and eluted as a single peak. Active fractions were pooled and concentrated as above to 8 ml. Finally, the concentrate was chromatographed on a 1.6 × 100 cm column of Sephadex G-75 gel equilibrated with the same phosphate buffer, a-Amylase eluted as a single peak that was

G. Gubern et al. I Clinica Chimica Acta 251 (1996) 145-162

149

pooled and concentrated by ultrafiltration to 9 ml. The final solution of the purified pancreatic ct-amylase had a specific catalytic activity of 52.9 kU/g of protein.

2. 4. Preparation and characterization of the reference material The purified pancreatic or-amylase was diluted 400-fold in a matrix containing 25 mmol/l PIPES buffer, pH 7.0, 50 mmol/l sodium chloride, 0.5 mmol/1 EDTA, 1.5 mmol/l calcium chloride and 30 g/1 human serum albumin. The 2-1 batch of material was frozen and shipped in dry ice to the National Institute for Biological Standards and Control (NIBSC, Potters Bar, UK), where the container was thawed at 4°C. The preparation was filtered at the same temperature through a sterile 0.22 #m filter. The filtrate was dispensed in a continuous process at 4°C into neutral clear-glass ampoules at a nominal volume of 1.0 ml per ampoule, which were then chilled to - 5 0 ° C and lyophilized 1-21]. After desiccation, the ampoules were filled with pure, dry nitrogen, sealed by fusing the glass and stored at -20°C. The precision and accuracy of dispensing the solution were monitored by weighing 27 ampoules taken at random after dispensing the 1.0 ml aliquot. "]['he dry weight after lyophilizing was also determined in six ampoules. The residual moisture content of the lyophilized material in three ampoules was measured by the Karl Fischer method 1-22]. All these determinations were performed at the NIBSC. The homogeneity of the batch was assessed by the ampoule-to-ampoule variation in the a-amylase catalytic concentration (measured with the 4,6-benzylidene-4-nitrophenyl-ct-D-maltoheptaoside method at 37°C). Measurements on the reconstituted material were carried out in 20 ampoules taken at random from the batch. Each ampoule was assayed in triplicate, on two successive days. Data were evaluated by analysis of variance. The stability of the pancreatic c~-amylase catalytic concentration in the lyophilized material was examined by a temperature-accelerated degradation study as described by Tydeman and Kirkwood 1-23]. Ampoules of the material 'were stored at -20°C, 4°C, 20°C, 37°C, 45°C, and 56°C and catalytic activities were determined in duplicate after 31, 184 and 857 days. A predicted loss of activity was obtained from the data on basis of application of the Arrhenius law, which relates rate of degradation to temperature by the method of maximum likelihood 1-23]. 2. 5. Certification procedure Participants (nine laboratories) were supplied with ampoules of the lyophilized a-amylase material. They reconstituted one ampoule on each

150

G. Gubern et al. I Clinica Chimica Acta 251 (1996) 145-162

of 3 days and made duplicate measurements of catalytic concentration of a-amylase on each ampoule on the day of reconstitution. Participants also measured the absorbance of a solution of 2-chloro-4-nitrophenol 5.0 mmol/1 on each day on which catalytic concentration was measured. Each participating laboratory was also provided with procedures for reconstituting the ampoules, calibrating pipettes, specifications for reagents and procedure for the preparation of solutions, specifications for the measurement conditions, procedure and calculations. The catalytic concentration of a-amylase in the reconstituted material was calculated on the basis of the molar absorption coefficient of 2-chloro-4-nitrophenol (1549 m2/mol at 405 nm and 37°C) and correcting for the increase in absorbance per min of the reagent blank and for the volumetric errors in the reconstitution of the material and pipetting procedure. The certified value and uncertainty were based on careful technical evaluation and statistical calculations [24]. 3. Results

3.1. Purification and characterization Fig. 1 shows the profile of the Sephadex G-75 chromatography. The enzyme eluted as a single peak retarded by the interaction of a-amylase with the gel matrix. The purification procedure is summarized in Table 1. The purity of the preparation was checked by polyacrylamide gel electrophoresis in the presence of sodium dodecyl sulphate (SDS-PAGE) showing a major band with a relative molar mass of 57 500 and a trace contaminant with a relative molar mass of 63 000 (Fig. 2). Polyacrylamide gel electrophoresis (PAGE) showed a single band, which was identified to be pancreatic a-amylase by immunoblotting with a specific antibody Table 1 Pancreatic a-amylase purification from human pancreas (64.8 g of fresh material) Stage

Protein (g)

Activity (U)

Crude extract Ammonium sulphate precipitation DEAE-Sephacel CM-Sepharose Sephadex G-75

13.8 6.3

14 600 10600

1.1 1.7

100 73

1.0 1.6

4000 1900 1600

14.6 35.9 52.9

28 13 11

13.8 34.1 50.2

0.28 0.052 0.030

Specific activity (kU/g)

Recovery (%)

Purification (fold)

151

G. Gubern et al. I Clinica Chirnica Acta 251 (1996) 145-162

'I

35

30 0.8

25

..J ~

0.6 20

._e o

E

o..

15 0.4

10

0.2

0

10

20

VO

30

40

50

60~ 70 Vt

80

90

Fraction

Fig. 1. Elution pattern of pancreatic ~-amylase from Sephadex G-75 chromatography. (11)

Indicates or-amylase catalytic concentration (kU/l) and (~) indicates protein concentration

(g/l).

against human pancreatic or-amylase (Fig. 3). Purity of the preparation was also assessed by FPLC chromatography on Mono-Q HR 5/5 and Superose 12 HR 10/30 columns. Only the ~-amylase peak was observed. Isoelectric focusing on acrylamide gel of the purified enzyme showed a major band of pI 7.1 (Fig. 4). Measurements of several possible contaminating enzymes in the purified preparation showed only trace amounts of triacylglycerol lipase and L-lactate dehydrogenase (0.07 U and 0.1 U per 100 U of ~-amylase, respectively).

3.2. Reference material preparation The reference material was prepared as described in Materials and

152

G. Gubern et al. / Clinica Chimica Acta 251 (1996) 145-162 1

94 000

2

3

4

5

6

)

67 000

43 000

D,

30 000

20 100

I~

Fig. 2. SDS-PAGE of different steps in the purification procedure (16/~g was applied to each lane). (1) Molecular mass markers, (2) crude extract, (3) ammonium sulphate precipitate, (4) DEAE-Sephacel peak, (5) CM-Sepharose peak, and (6) Sephadex G-75 peak.

1

2

3

4

5

6

i

Fig. 3. PAGE of purified pancreatic s-amylase analyzed as follows: lanes 1-2, immunoblotted with specific antibody for human pancreatic s-amylase; lanes 5-6, with amido black staining. Lanes 3 and 4 correspond to purified salivary c~-amylase immunoblotted with specific antibody for human pancreatic s-amylase (lane 3) and stained with amido black (lane 4).

G. Gubern et al. / Clinica Chimica Acta 251 (1996) 145-162

1

2

153

3

Fig. 4. lsoelectric focusing on acrylamide gel (pH range from 3.0 to 9.0) of purified pancreatic c~-amylase. Lane 1, isoelectric pH markers; lane 2, a pilot pancreatic purification; lane 3, reference material.

methods (Section 2.4). The purified enzyme required a dilution of the order of 400-fold to produce a catalytic concentration within the range of the methods of measurement performance of which it is intended to assess. A target level of activity of about four times the upper reference limit of such metlhods was chosen, so that the reference material could also be used to prepare calibration curves. The dispensing mean mass in the ampoule filling procedure was 1.0141 g (range from 1.0118 to 1.0164 g, n = 27 ampoules). There was no evidence for any trend in the variation of mass throughout the filling procedure. The residual moisture content of the lyophilized material was 0.103% (n = 3). The results confirmed that adequate drying had taken place. The dry weight was 0.0401 g (n = 6). Table 12 shows the variability of pancreatic ~-amylase activity measured in reconstituted specimens. No significant between-ampoule variation was detected by analysis of variance (~ = 0.01). Results of the statistical evaluation indicated that the batch was sufficiently homogeneous to be used for certification.

3.3. Stability studies In preliminary trials, stability of purified pancreatic a-amylase was

154

G. Gubern et al. I Clinica Chimica Acta 251 (1996) 145-162

Table 2 Homogeneity study Components of variability

S.D. (U/I)

CV (%)

Between-ampoule Within-ampoule

3.6 8.1

0.8 1.9

Components of variance of pancreatic o-amylase catalytic concentration measured in the reference material (mean 429 U/l). Analysis of variance of the results (six determinations per ampoule in two different days on 20 ampoules). S.D., standard deviation. CV, coefficient of variance. tested in a liquid matrix containing different stabilizers. Fig. 5 shows ~-amylase stability at 37°C in three of the tested matrices. Experiments indicated that pancreatic ~-amylase activity was stabilized in the matrix containing albumin. Increasing the concentration of h u m a n albumin to 30 g/1 led to a protection of the enzyme also upon lyophilization. The effect of possible traces of contaminating proteases on the integrity of the pancreatic or-amylase molecule was studied. We incubated the purified preparation for 2 h at 30°C with up to 20 mg/1 of chymotrypsin,

120

8O m

~60

o m nm

20

0 I

]

I

I

I

I

I

I

I

I

0

2

4

6

8

10

12

14

16

18

20

Dmym at 37*(:

Fig. 5. Pancreatic or-amylase stability at 37°C in the following matrices: (.) PIPES buffer 50 mmol/l pH 7.4; (*) PIPES buffer 50 mmol/l pH 7.4, NaC1 50 mmol/l, EDTA 0.5 mmol/l and CaCI2 1.5 mmol/l; (11) PIPES buffer 50 mmol/1 pH 7.4, NaCI 50 mmol/i, EDTA 0.5 mmot/l, CaC12 1.5 mmol/i and human albumin 5 g/l.

G. Gubern et al. I Clinica Chimica Acta 251 (1996) 145-162

155

trypsin, pepsin and protease from Streptomyces type IV. Only trypsin was found to affect a-amylase activity. Albumin notably protected the purified a-amylase from trypsin proteolysis, so that a decrease of only 5% of ~-amylase activity in the reference material (containing 30 g/1 of albumin) was observed after incubation at 30°C for 2 h with trypsin 0.2 g/1. To assess a-amylase stability in the lyophilized material, an accelerated thermal degradation study was carried out as described in Materials and methods (Section 2.4). F r o m all the combined results the predicted yearly relative loss of activity was 0.35% at 4°C and 0.03% at - 2 0 ° C . The results were reasonably consistent with the Arrhenius model. After reconstitution, no loss of activity was observed for 15 days at 37°C.

3.4. Catalytic properties The apparent Michaelis-Menten constant (Km) of a-amylase for the substrates 2-chloro-4-nitrophenyl-~-D-maltotrioside and 4,6-benzylidene4-nitrophenyl-~-D-maltoheptaoside were determined using the reconstituted reference material and selected human serum samples containing the pancreatic isoenzyme (Table 3). The pH effect profile was also studied using the purified pancreatic isoenzyme and human serum. The effect of pH (Fig. 6) was very similar for both specimens.

3.5. Certification procedure All the participants met the specifications for spectrophotometers. The reaction temperature was controlled in the reaction mixture. All the participants used a reference thermometer for temperature calibration. The mean volume of the 1.0 ml pipette was 1.0006 ml (S.D. 0.0054 ml) and it was 0.1)099 ml (S.D. 0.00013 ml) for the 0.01 ml pipette. No laboratory reported any difficulty in following the specified procedure for the measurement. Table 3 Apparent Michaelis-Mentenconstant (Kin; mmol/1)of human pancreatic ~-amylase for PNP-G7-B and CNP-G3 (n = 5) at 37°C (mean _ S.D.)

Reference material Human seruma

PNP-G7-B

CNP-G3

0.207 +_0.061 0.200 _+0.065

0.302 __+0.090 0.554 + 0.071

"Human serum with pancreatic ~-amylase isoenzyme. S.D., standard deviation.

156

G. Gubern et al. I Clinica Chimica Acta 251 (1996) 145-162

120

100 \

a/

' %

80

\

._>

"5 a

60

>. E

40

/

20

0-

5

55

6

65

7

7.5

8

pH

Fig. 6. pH Effect on c~-amylase activity as measured in the reference material ( ) and in human serum with pancreatic e-amylase isoenzyme ( - - - ) with P N P - G 7 - B (119 or C N P - G 3

(11) as substrate.

On each day that catalytic concentration was measured, 0.010 ml of the 2-chloro-4-nitrophenol solution was mixed with 1.00 ml of the e-amylase reagent, in duplicate and using the same pipettes employed for the measurement of the catalytic concentration. The absorbance of the mixture was read at 405 nm and at 37°C against the e-amylase reagent. The results obtained are shown in Table 4. Table 5 gives the results (mean and S.D.) for e-amylase catalytic concentration obtained by each participating laboratory. One outlier was identified among the individual values reported by laboratory number 6. The Gaussian distribution of data was confirmed. One outlying mean, that of laboratory 9, was identified. That laboratory reported only three values, showed the highest volume error in the pipette calibration and modified the measurement procedure by doubling the pipetted volumes. Taking into account all these considerations, laboratory number 9 was discarded in the calculation of the mean of means. The analysis of variance showed that the between-laboratory variation was significantly different from zero. As a consequence, the laboratory mean values were used for the calculation of the certified value and the uncertainty (0.95 confidence interval). A two-way analysis of variance (nested model) was carried out to estimate between- and within-day contribution to the variance. Table 6 summarizes the results of the statistical evaluation. There was no significant between-day difference (~ = 0.05).

G. Gubern et al. I Clinica Chimica Acta 251 (1996) 145-162

157

Table 4 Absorbance at 405 nm (37°C) of the standard solution of 2-chloro-4-nitrophenol Lab code

Mean

S.D.

1 2 3 4 5 6 7 8 9

0.788 0.784 0.762 0.770 0.774 0.769 0.771 0.760 0.787

0.017 0.021 0.007 0.003 0.003 0.007 0.030 0.003 0.036

M e a n o f means

0.774

0.010

Mean and standard deviation (S.D.) obtained by each .laboratory of six measurements of absorbance.

T h e a r i t h m e t i c m e a n of the m e a n s (the certified v a l u e o f the c~-amylase c a t a l y t i c c o n c e n t r a t i o n in C R M 476) w a s 555 U / I (9.25 #kat/1) a n d the u n c e r t a i n t y c o r r e s p o n d e d to 11 U/1 (0.17 #kat/1).

4. Discussion T h e p r e p a r a t i o n of a suitable e n z y m e reference m a t e r i a l involves the selection of a n a p p r o p r i a t e s o u r c e of the e n z y m e . T h e m a i n r e q u i r e m e n t s Table 5 ~t-Amylase catalytic concentration (U/l) as measured by each laboratory Lab code

Mean

S.D.

1 2 3 4 5 6 7 8 9

560.1 551.6 575.6 555.8 539.2 538.8 561.5 555.0 527.3a

6.5 10.5 4.9 5.6 2.7 26.4 24.6 2.2 12.6

Mean and standard deviation (S.D.) of six replicates, except for lab 9 (reported only the mean value obtained each day). aOutlier value (Box-Whisker test).

158

G. Gubern et al. I Clinica Chimica Acta 251 (1996,) 145-162

Table 6 Summary of the statistical evaluation in the certification campaign Sources of variation

S.D. (U/l)

CV (%)

Between-lab Within-lab Within-day

10.7 13.8 6.3

1.9 2.5 1.1

SD, standard deviation. CV, coefficient of variation.

for an enzyme reference material are assured stability within defined periods of storage and use, and kinetic properties of the reference enzyme as close as possible to those of the corresponding human enzyme in serum [25]. While a high catalytic activity content is also desirable for an enzyme reference material, an equally important criterion is the absence of potentially interfering enzymes or other proteins. Even though the catalytic properties of an enzyme mainly depend on its origin and type, they can be altered by procedures during purification, by the matrix added and by the lyophilization process. The diagnosis of acute pancreatitis is mainly based on the increase of serum and urine a-amylase and especially of its pancreatic isoenzyme. Therefore, the ability to choose an enzyme that is identical with the target analyte outweighs the disadvantages, either ethical or in terms of the risks of infection, that are associated with the use of human tissues. Furthermore, large qualitative differences in a-amylase derived from malt, bacteria, porcine and human sources have been demonstrated [20,26,27], making it necessary that the a-amylase used as a reference material must be of human origin. An additional reason for choosing human pancreas was that the reference material might serve not only to standardize the measurements of "total" a-amylase catalytic concentration but also the specific measurement of the pancreatic isoenzyme. Our preparation proved to be free from contaminating enzymes. Only traces of triacylglycerol lipase and L-lactate dehydrogenase were detected. SDS-PAGE and isoelectric focusing showed one major band corresponding to pancreatic a-amylase. Although purification, freeze-drying and addition of preservatives may alter the properties of the native enzyme, our reference material showed unchanged molecular and kinetic characteristics. Pancreatic a-amylase was notably stable in matrices containing NaC1 50 mmol/1, EDTA 0.5 mmol/1, CaC12 1.5 mmol/1, human albumin 5 g/1 and PIPES buffer at pH 7.0 to 7.5. To protect the enzyme from denaturation, the concentration of

G. Gubern et al. I Clinica Chimica Acta 251 (1996) 145-162

159

human albumin was increased to 30 g/1. A predicted annual relative loss of activity of 0.35% at 4°C and the consistency of the data with the Arrhenius model corroborated our preliminary results in stability studies. We concluded that the pancreatic ~-amylase reference material was extraordinarily stable, even more than other enzyme reference materials [5,6,8]. The candidate reference preparation was submitted to trials to study its behaviour in comparison with that of human serum samples; such studies are essential in establishing the commutability of the material. Commutability is defined as the ability of a material to show inter-assay properties comparable to those of human serum [28]. Furthermore, commutability is additional evidence that an enzyme purified from a tissue source and stabilized in a defined matrix is catalytically similar to the enzyme occurring in human serum. Apparent Michaelis-Menten constant and optimum pH of the ~-amylase reference material were close to those of human :serum pancreatic isoenzyme. The reference material has been found to be commutable with human sera both by its inter-method behaviour in a previous study [26] and by its kinetic characteristics, in the present work. The uniformity of the batch material was controlled throughout the filling and lyophilization procedures and gave satisfactory results. The batch was found to be sufficiently homogeneous to be used for the certification procedure. Because no reference method has been defined for ~-amylase catalytic concentration measurement, we had to choose a method for the certification campaign. A review of the methods of current use in the European Community for ~-amylase catalytic concentration measurement indicated that methods with maltooligosaccharides as substrates are mainly employed. Recently, a new ~-amylase substrate has been introduced [16] which uses 2-chloro-4-nitrophenyl-~-D-maltotrioside as substrate. This compound acts as substrate and chromophore, allowing the continuous monitoring of the 0~-amylase activity without auxiliary enzymes. Besides that, it is freely available to individual laboratories and manufacturers of diagnostic reagents. We selected a method using that substrate for the certification exercise. The majority of methods of measuring catalytic activity that are in current use depend on the knowledge of the molar absorption coefficient of the reaction product to allow instrumental readings to be transformed into units of catalytic concentration. The catalytic concentrations of enzymes in CRM 426 [7], CRM 299 [8] and CRM 404 [10] were certified on this basis. Previous experience in the BCR enzyme reference materials programme has shown that variation in the performance of spectrophoto-

160

G. Gubern et al. / Clinica Chimica Acta 251 (1996) 145-162

meters may contribute significantly to between-laboratory variation in the results of enzyme measurements. Therefore, a standard prepared in one laboratory was provided in the certification exercise to verify the photometric performance of the participants. The results (Table 4) showed a satisfactory between-laboratory variation. The molar absorption coefficient for 2-chloro-4-nitrophenol obtained by the participants using this solution was (mean of means) 1563 m2/mol (S.D. = 21 m2/mol), which is not significantly different from that used in the calculation of the catalytic concentration, 1549 m2/mol (S.D. = 12 m2/mol). In a few particular cases, the certification is traceable to the product of the enzyme reaction that is photometrically measured. Although a standard of 2-chloro-4-nitrophenol was used in the certification procedure of the pancreatic or-amylase material, we decided to consider this solution more as a photometric control than a calibration standard because, (1) data about purity of the substance were not available, (2) the same solution, prepared in only one laboratory, was used by all the participants, and (3) the standard will not be available to future users of CRM 476. Taking into account all the pre-analytical factors of variability, the calculations of or-amylase catalytic concentration (U/I) for the accepted set of mean laboratory data, gave small within-laboratory (2.5%) and a between-laboratory (1.9%) coefficients of variation that are evidence of the excellent reproducibility and transferability achievable by the method used. Those values are similar to those found for other enzyme reference materials and can be taken to represent the current state of the art in inter-laboratory enzyme activity measurements.

Acknowledgements This work was supported by grants 5013/1/6/318/88/ll-BCR-E (10), 5430/1/6/318/91/01-BCR-UK (10), 5521/1/6/318/91/08-BCR-UK (10) and 5506/1/6/318/91/7-BCR-E (10) from the Bureau Communautaire de Reference (BCR) of the European Commission. We thank R. Gaines Das (National Institute for Biological Standards and Control, Potters Bar, UK) for calculating the predicted degradation rates of the reference material and B. Hilger (Boehringer Mannheim, Penzberg) for participating in the inter-laboratory evaluation.

References [1] Bowers GN, Bergmeyer HU, Horder M, Moss DW. Approved recommendation (1978) on IFCC methods for the measurementof catalytic concentrationof enzymes. Part I. General considerationsconcerningthe determination of the catalyticconcen-

162

G. Gubern et al. / Clinica Chirnica Acta 251 (1996) 145-162

[18] Cohen JO. Effect of culture medium composition and pH on the production of M protein and proteinase by group A Streptococci. J Bacteriol 1969;99:737-744. [19] Bradford MM. A rapid and sensitive method for the quantification of microgram quantities of protein utilizing the principle of protein-dye binding. Anal Biochem 1976;72:248-254. [20] Sampson EJ, Duncan PH, Fast DM, Whitner VS, McKneally SS, Baird MA et al. Characterization and intermethod relationships of materials containing purified human pancreatic and salivary amylase. Clin Chem 1981;27:714-720. [21] Campbell PJ. International biological standards and reference preparations. II. Procedures used for the preparation of biological standards and reference preparations. J Biol Stand 1974;2:259-267. [22] Seaman WM, McComas WH, Allen GA. Determination of water by Karl Fischer reagent stoichiometric iodometric method. Anal Chem 1949;21:510-512. [23] Tydeman M, Kirkwood Y. Design and analysis of accelerated degradation tests for the stability of biological standards. J Biol Stand 1984;12:195-224. [24] Marchandise H, Colinet E. Assessment of methods of assigning certified values to reference materials. Fresenius Z Anal Chem 1983;316:669-672. [25] NRSCL 3-T. Guidelines for the development of certified reference materials for use in clinical chemistry for The National Reference System for the Clinical Chemistry. National Committee for Clinical Laboratory Standards 1984;1-20. [26] Gubern G, Canalias F, Gella FJ. Determination of a-amylase activity: methods comparison and commutability study of several control materials. Clin Chem 1995;41:435-438. [27] Bretaudiere JP, Rej R, Drake P, Vassault A, Bailly M. Suitability of control materials for determination of or-amylase activity. Clin Chem 1981;27:806-815. [28] Rej R. Accurate enzyme activity measurements. Two decades of development in the commutablity of enzyme quality control materials. Arch Pathol Lab Med 1993; 117:352- 364.