May 6, 1985 - Roy, Riviera Beach, FL 33404), and a fluorometric detector (cat. no. SPF; Aminco-Bowman,. Silver. Spring, Ml) 20910) with a 25-FL flow cell ...
CLIN. CHEM. 31/11, 1841-1845 (1985)
Measurementof CreatineKinaseMM Sub-typeby Anion-ExchangeLiquid Chromatography Alan H. B. Wu and Terrie G. Gornet Electrophoreticand isoelectric focusing studieshave demonstrated the presence of multiple sub-forms of the major creatine kinase isoenzymes, and preliminarywork has suggested that measurement of the MM sub-forms provides early diagnostic information concerning acute myocardial infarction. We developed an isocratic method based on liquid chromatography to examine further the clinical potential of measuring these sub-forms. An anion-exchange column is used coupled with the fluorometric detection of NADPH after reaction of the sub-forms with CK reagents, which are added through a post-column pump. We examined various chro-
matographic variables in producing the best separations. Results for blood collected from normal individuals and patientswith acute myocardialinfarctionand skeletalmuscle diseases are compared. We conclude that for routine CK-MM
sub-formanalysis,this chromatographicprocedureis better suited than isoelectric focusing because of the faster turnaroundand lower costs. Wevers et al. (1) were the first to demonstrate multiple sub-forms of creatine kinase (CK, EC 2.7.3.2) isoenzymes in human serum.1 Specifically, three MM and two MB subforms can be observed either by electrophoresis on agarose gel or by isoelectric focusing (2). Because only the MM3 subform can be demonstrated in purified extracts of human muscle, itisthoughttobe the pure gene product (3). During cellular necrosis, MM3 is released into the circulation, where it is converted first to MM2 and then to MM1 by a conversion factor (4). Recently, this factor has been identified as a carboxypeptidase (5). The cleavage of the Cterminal lysine on each M subunit of MM3 alters the isoelectric point of the enzyme and provides the basis for its analysis by isoelectric focusing. The clinical utility of sub-MM isoenzyme evaluation was demonstrated in a preliminary study of eight patients with myocardial infarction (MI). Morelli et at. (6) found that the ratio of MM3 to MM1 was increased before CK-MB and may provide for early diagnosis of MI. They also found that measuring these sub-forms may help determine whether or not coronary re-perfusion isachievedin MI patientsbeing treated with streptokinase (7). Direct CK-M chain analysis may also be useful in determining how long MM isoenzymes have been in blood after MI (8), which may have applications for infarct sizing (9). Annesley et at. (10) used subisoenzyme analysis to examine the prognosis for patients with polymyositis. These reports evidence the desirability of a rapid method for determining MM sub-forms. of Pathology and Laboratory Medicine, University of Medical School, Houston, TX 77025. Presented atthe37thnational meeting of the AACC, Atlanta, GA, July 1985. ‘Nonstandard abbreviations: CK, creatine kinase (ATP:creatine N-phosphotransferase, EC 2.7.3.2); AK, adenylate kinase (ATP:AMP pho8photransferase, EC 2.7.4.3); Ml, myocardial infarction. Received May 6, 1985; accepted August 14, 1985. Department Texas Health
Measurement of the major isoenzymes of creatine kinase by use of”high-pressure” anion-exchange liquid chromatography (HPLC) has been described by several investigators. Methods of detection exploit the fluorescence of CK-coupled NADH (11, 12) and bioluminescence of CK-coupled ATP (13). In none of these reports was analysis of the various subforms optimized and the MM isoenzyme was eluted as a single peak. Here we describe an isocratic HPLC method whereby the resolution is expanded to allow for quantification of individual MM sub-types. Anion-exchange chromatography isfollowedby post-column reaction, and fluorescence detection. We present data to illustrate the potential use of these measurements in patients with acute Ml and skeletal-muscle diseases. Analysis time is considerably shorter than for isoelectric focusing, and thus the procedure is more amenable for routine analysis.
Materials and Methods Apparatus. We used a Model 955 dual-piston HPLC pump (Tracor Analytical, Austin, TX 78721), a six-port rotary injector with a 100-juL sample loop (cat. no. 7125; Rheodyne, Cotati, CA 94928), a 4.1 mm x 2 cm guard column of weak anion-exchanger and a 4.1 mm x 25 cm analytical column (SynChropak AX300; both from SynChrome, Inc., Linden, IN 47955), a single-piston post-column reagent pump (cat. no. 396; LDC/Milton Roy, Riviera Beach, FL 33404), and a fluorometric detector (cat. no. SPF; Aminco-Bowman, Silver Spring, Ml) 20910) with a 25-FL flow cell (Hellina, Jamaica, NY 11424) set at 350 and 450 nm for excitation and emission, respectively. Figure 1 shows a block diagram of the system. Total CK was measured in a Hitachi 705 analyzer (Boehringer Mannheim Diagnostics, Houston, TX 77063), at 37 #{176}C. For isoenzyme assay we used the “Paragon” agarose electrophoresis system (Beckman Instruments, Brea, CA 92621). CK-MB was measured by using Isomune CK-MB (Roche Diagnostics, Nutley, NJ 01770) in the 705 analyzer after immunoprecipitation (14). CK-MM sub-form assay. The serum sample is diluted with buffer (per liter, 20 mmol of Tris and 5 mmol of ED’FA, pH 7.80 ± 0.02) according to its total CK activity. EDTA is added to stabilize the sub-forms and prevent in vitro conver-
Science Center,
Fig. 1. Block diagramof the HPLC systemforthe analysisof CK-MM sub-forms CLINICALCHEMISTRY, Vol. 31, No. 11, 1985 1841
sions (15). We diluted
samples as follows: twofold dilution CK ranging from 25 to 200 UIL, fourfold for 200-500 UIL, sixfold for 500-1000 UIL, 10-fold for 1000-2000 U/L, and 20-fold for activity >2000 UIL. A 100-jL portion is injected into the chromatograph. Before use, the mobile phase (Tris buffer, 20 mmoJJL, pH 7.80 ± 0.02) is passed through a ifiter of 0.22-gm average pore size (Millipore Corp., Bedford, MA 01730).To maintain precise ionic strengths, we adjusted the pH by mixing stock 20 mmol/L solutions of Tris acid and base. For the reagent used to detect the CK activity, we used CK-NAC (UV) (Boehringer) with the Boehringer Mannheim 8700 analyzer, the stock reagents being prepared according to the manufacturer’s instructions. The working reagent was prepared by combining 30 mL of the substrate/enzyme with 3.5 mL of the starter (creatine phosphate). Flow rates for both the mobile phase and post-column reagent were set at 0.5 mL/min, which results in an incubation time of about 3 mm. The separation and incubation temperature was 45#{176}C, except where noted. Quality control. Serum samples containing high concen-, trations of total CK were pooled, ifitered, and assayed. To this pool, EDTA and mercaptoethanol were added to a final concentration of 5 and 10 mmol/L, respectively. We dispensed 0.5-mL aliquots into 1.5-mL polyethylene tubes, and FIg. 2. The effect of mobile-phase pH on the chromatographic resolustored them frozen at -70 #{176}C. In addition,the contentsof tionof CK-MM sub-forms some of these tubes were lyophiuized and stored at -70 #{176}C. (A) pH 6.9. ( pH 7.5. (C) pH 7.8. Aow rate:0.5 mL/mln for both the mobile At the time of assay, samples are thawed at 37#{176}C for 3 miii phase (20 mmol/L Tris) and post-column reagent (CK reagent with creatine phosphate). Temperature 45 ‘C. Injection size: 100 pL, of a 1:4 dIlution (with and used as soon as possible. buffer)of the quality control sample Sub-type nomenclature. We have elected to adopt the nomenclature established by Wevers et al. (1,8) and used by others (4-6) whereby the most cathodic sub-form, as measured by agarose electrophoresis is designated as MM3. In our system, the order of elution is: MM3, MM2, MM1.
for samples
with
total
Results Effect of pH and ionic strength. We optimized the chromatographic conditions for separating the sub-forms of CKMM on the weak anion-exchange column. The most critical factors are the pH and ionic strength of the mobile phase. Retention of proteins on the ion-exchanger depends on the surface charge oftheproteins and the extent of ionization of the support. If the mobile phase has a pH that exceeds the p1 of the protein, an anion-exchange column (16) must be used. Sometimes proteins are eluted from the column at pHs as high as 1-2 units above the p1(17). The isoelectric point of the three major CK-MM sub-forms ranges from 6.90 for MM3 to 6.40 for MM1 (6), thus we separately tested mobile phases with pHs ranging from 6.9 up to 8.0, the maximum pH that the weak anion-exchange column can support. As shown in Figure 2, the MM sub-forms were co-eluted as a single band at pH 6.9. As the pH increases, resolution improves,
the best
separation
being
obtained
at
pH 7.80.
Above this, the ionization ofthe supportbeginstodiminish, and the exchanger is unable to retain the protein. Resolution with the weak anion-exchanger was adequate, so we did not investigate the use of mobile phase at pus >8, which requires the use of a strong exchanger (18). Chromatograms similar to those in Figure 2 were obtained when we varied the ionic strength of the mobile phase from 100 to 5 mmol/L. A high salt concentration decreases retention time and peak resolution; a low salt concentration in the mobile phase produces broad peaks with long retention times. Accordingly, for routine applications, we used 20 mmol/L Tris, which gives a good balance between required resolution and reasonable retention time. Temperature. Increasing the temperature of the column performs two functions, as shown in Figure 3A-C. First, 1842 CLINICALCHEMISTRY, Vol. 31, No. 11, 1985
10
--
TIME (mm) Fig. 3. The effect oftemperatureandflow rate:(A) 25#{176}C at 0.5 mL/min; ( 37#{176}C at 0.5 mLlmin; (C) 45#{176}C at 0.5 mL/min; (L 45#{176}C at 1.0 mlJmin; (E) 45#{176}C at 0.25 mL/min Otherconditionsare as in Figure2 resolution is improved. Second, sensitivity is improved because the reaction rate is increased. The sub-types are not affected by the temperature change because their relative proportions are unchanged for temperatures between 25 and 45#{176}C. Indeed, many electrophoretic methods for CK are
based on the incubation of the isoen.zymes with the substrate at 45#{176}C. Flow rate. Figure 3C-E shows the effect of the flow rate of the mobile phase. Since the incubation-loop volume is fixed, decreasing the flow rate increases the incubation time, and thus the amount of NADPH produced. In addition, resolution is also slightly improved. The optimum flow rate is a trade-off between desired analysis time and degree of resolution. With a new column, we begin with 0.5 mL/min. As the column ages, the flow rate must be decreased to obtain the same degree of resolution. Validation of MM subunit analysis. The MM isoenzyme from myocardial tissue consists entirely of MM3 (3,19). To validate the HPLC analysis for CK-MM3, we processed human myocardial tissue according to a previously described procedure (20), modified to include 5 mmol of EDTA per liter. We added serial dilutions of the myocardial extract to human serum and analyzed for CK-MM3. Table 1 shows the analytical recoveries. To validate the MM1 sub-form, we converted all of the MM3 and MM2 sub-forms in a serum sample to MM1 by incubating it at 37#{176}C for 96 h. In the absence of stabilizers, endogenous serum carboxypeptidase N (arginune carboxypeptidase, EC 3.4.17.3) will convert the MM sub-types from the tissue form (pure gene product, MM3) to the form found in serum (MM1) at a rate dependent on the existing in vitro peptidase activity. For example, in one specimen, MM3 was converted from an initial 54% of total MM content to 31%, 21%, and 0% after 24, 48, and 96 h, respectively, while the MM1 increased from 5% to 100% during over this time. To determine the analytical recovery of MM1, we supplemented a fresh serum sample with varying amounts of “converted” MM1. These results are also shown in Table 1. We were not able to obtain control material consisting of only MM2 for similar testing; however, we presume that its recovery from serum would be equivalent to that found for MM1 and MM3. Effect of EDTA on sub-form conversions in vitro. To prevent the in vitro conversion of MM during storage, we added 5 mmol of EDTA per liter to serum samples and checked its stability by incubating the specimen for 48 h at 37 #{176}C. As demonstrated by Chapelle et al. (15), the composition of the MM sub-forms is not significantly altered in the presence of EDTA, and we add it to all subsequent specimens. Precision. The peak areas of each CK-MM sub-form is a direct function of the incubation time with the substrate, thus we were not able to calibrate the chromatographic signal directly to enzyme activity. The precision is therefore reported as percent of total CK-MM activity, as shown in Table 2. Of the two control materials tested, we found that the lyophilized control has higher stability than the frozen material and can be used for at least four months. Reference range. We selected 27 normal individuals and non-MI patients who had normal values for CK, in the establishment of a reference range. For total CK (reference
Table 1. Analytical Recovery of MM3 and MM, Added
Sub-form a
iiia
3
MM3 MM,b MM1 MM1 ‘MM
Found U/L
120 148
104
336
328
412
364 552
504
156
sub-formproportionsbefore supplementation:MM3
87 105 98 88 109 19.5%, MM2
52.4%, MM, 28.1%, total CK 424 U/L. bMM sub-form proportions before 31.3%, MM, 62.9%, total CK 450 tilL.
supplementatIon: MM3
5.8%, MM2
Table 2. Within-Run and Between-Run Precision CV. % MM sub-type MM3
Composition, 17.4
MM2
38.5
MM,
44.0
Within-run Between-run 1.0 6.3 3.5 4.2 3.7 3.9 452 U/L, n = 6 foreach determinationfor
%
Frozenpool (-70 #{176}C), totalCK = both within and between run precision. a
range 8-130 UIL): Si = 63 ± 27, range 16-129 U/L. For MM3 expressed as % of total MM: = 16.4 ± 5.2, range 10.0-25.9%, MM2: 51 = 35.4 ± 9.5, range 29.2-52.9%, and MM1: = 48.5 ± 12.2, range 23.9-68.8%. For the ratio of IvIM1fMM3: SI = 3.3 ± 1.4, range 1.0-6.7. Using data on 95% of the population, our reference interval for the MM1IMM3 ratio is 1.4-5.3. These values are slightly lower than those reported by Annesley et al. (10), probably because of the difference in methods. We estimate a sensitivity limit for each of the MM sub-forms to be about 5 UIL (at an S/N of 2), as determined from the sample with a total activity of 8 U/L. This sensitivity is adequate for routine analysis, because the MM isoenzyme is usually the one found in the highest concentration in serum. Effect of other isoenzymes and atypical CKs. We injected several samples containing high concentrations of CK-MB as determined by Isomune CK-MB (Roche Diagnostics). With use of a low salt concentration, 20 mniol/L Tris buffer, in the mobile phase, CK-MB does not elute off the weak anion-exchange column. MB will be eluted with a gradient in which the salt concentration is increased to 300 mmol/L (11). We were not interested here in measuring CK-MB, so we maintained an isocratic mobile phase such that CK-MB was retained on the column. We also examined specimens containing substantial amounts of CK-BB and of the atypical macroCK types 1 and 2 (21), as confirmed by examining the electrophoretic pattern before and after CK-MM and CK-MB were removed by immunoprecipitation (22). When we injected these samples after removing CK-MM and CK-MB, no chromatographic peaks were observed. We suggest that these forms are not eluted under these isocratic conditions. Macro CK type 2 is thought to be a stable, high-molecular-mass polymeric aggregate of mitochondrial CK. In contrast, climeric mitochondrial CK is very labile, and its p1 is near that for CK-MM. Although we were not able to test the chromatographic migration of mitochondrial CK in our system, it is not usually observed in high concentrations in serum, even after MI (23). Effect of AK. Adenylate kinase (AK, EC 2.7.4.3) catalyzes the conversion of ADP to ATP, thereby enhancing the apparent CK activity (24) in samples with high AK activity. It is co-eluted with CK-MM by anion-exchange chromatography (25). We studied the effect of AK by examining a normal sample before and after adding AK from lysed erythrocytes. The original CK activity of the sample, 231 U/L, increased to 267 UIL after the extract (final hemoglobin 7 g/L) was added. As shown in Figure 4, no artifactual peaks are produced by AK when the CK substrate, phosphocreatine, is removed. The presence of diadenosine pentaphosphate (11.7 mol/L) and adenosine monophosphate (5.8 mmolJL) in the substrate formulation may sufficiently inhibit the AK activity to bring it below detection limits. CK-MM sub-forms in MI. We selected and assayed blood from several patients with acute MI for CK-MM sub-types. MI was diagnosed on the basis of the clinical history, including chest pain, specific electrocardiographic changes, and abnormal values for enzymes and isoenzyme values such as CK, CK-MB, and lactate dehydrogenase and its CLINICAL CHEMISTRY, Vol. 31, No. 11, 1985
1843
B
TIME (mm) The effect of AK Agrosslyhemolyzedspecimenwas processed with (A) and without( the CKsubstratecreatinephosphate.Other conditionsare as In Figure2 4.
Fig.
isoenzyme 1. As shown in Figure 5, acute MI is characterized by an initial release of the MM3 sub-type followed by its conversion to MM2 and MM1 in the ensuing hours. As shown in Figure 5, the ratio of MM1IMM3 for the early samples (8 and 12 h) is below the reference range in this and other MI cases. After the peak release and when the value for total CK has returned to normal (at 84 h) the ratio of MM1IMM3 isalsowithinthe referencelimit. CK-MM sub-forms in muscle disease. We examined the sera from two patients with skeletal muscle diseases. As shown in Figure 6, these patients exhibited remarkably different MM sub-type patterns. Evidently the patient whose pattern is shown in Figure 6A is in a stable or improving state, whereas the patient whose pattern is shown in Figure fiB isin a deteriorating state (10). Routine measurement of the MM sub-forms may aid in monitoring the progress of patients with neuromuscular disease. We are currently attempting to further correlate the significance of MM sub-form analysis for a larger number of patients with neuromuscular
disease.
Discussion Our initial attempts to separate CK-MM sub-forms involved gradient elution. The elution of components at high salt concentration will greatly extend the life of ion-exchange columns. Unfortunately, gradient elution also prolongs chromatography time, because after each run columns must be re-equilibrated to starting conditions. Furthermore, the time between injections isa critical factor in reproduc-
TIME
(mm)
5. CK-MM sub-type analysis in a patient
with MI (A) 8 h after onset of chest pain: total CK 252 U/L, MB 26(10.3%), MM340.8%, MM2 45.0%, MM1 14.1%, MM,/MM3 0.3 ( 12 h: CX 520, MB 70(13.5%), MM3 26.9%, MM2 51.0%, MM, .0%, MM,IMM30.8. ( 84 h: CX 48, MB 2(4.3%), MM3 19.6%, MM231.6%, MM, 48.8%, MM1/MM32.5.Samples dikited accordhig Ag.
to
the protocol. Chromatographicconditions are as in Figure2
1844 CLINICAL CHEMISTRY, Vol. 31, No. 11, 1985
0 TIME
5
10
1!
(mm)
Fig. 6. CK-MM sub-type analysis in skeletal musde diseases (A) patient withmyotonicdystrophy:total CX 720 U/L MM3 3.0%, MM3 24.8%, MM1 72.7%, MMI/MM3 242.
(
patientwith allhrogrypostsmultiplex and
possiblecongenitalmyopathy:total CX 4600 U/L MM3 57.6%, 1.9%, MM,/MM3 0.03
MM2 40.5%, MM,
ing retention times and resolution. Because we were able to elute the MM sub-types isocratically, the useful life of the column was extended through frequent changes of the guard column. The guard column should be replaced after a total of 20 mg of protein has been injected. We use different sample dilutions, but in general this is equivalent to about 10 normal samples (25-200 U/L), 20-30 mildly high-value samples (300-1000 U/L), and 30-50 grossly high-value samples (>1000 UIL). Initially we attempted to monitor the bioluminescence produced by the reactionofATP and luciferin with luciferase. Although this method is very sensitive, we found it too expensive for routine use. We therefore chose to monitor the NADPH produced in the Rosalki method fluorometrically (26). With use of commercial reagents, the current costs for consumables are less than $2.00/sample. MM sub-type studies with electrophoresis and isoelectric focusing techniques have necessarily all been retrospective, because of the length of time required for such analyses and their high cost for single analyses. it is therefore unlikely that sub-isoenzyme analysis by these techniques will be available on an emergency basis, and results will not be directly useful to the MI patient. In a survey conducted by Johnson (27), laboratory personnel found good clinical potential in MM sub-type analysis but questioned the practicality of isoelectric focusing in routine hospital laboratories. In contrast, the present method requires
