We observed nonenzymic peaks when serum isoenzymes of lactate dehydrogenase. (EC 1.1.1.27; LD) and creatine kinase (EC 2.7.3.2; CK) were separated by ...
CLIN. CHEM. 26/6, 707-711(1980)
Interferences Appearing in Fluorometrically Measured LiquidChromatographic Profiles of Creatine Kinase Isoenzymes in Serum Timothy D. Schlabach,1 Joe A. Fulton,2 Peter B. Mockridge, and E. Clifford Toren, Jr.3 We observed nonenzymic peaks when serum isoenzymes of lactate dehydrogenase (EC 1.1.1.27; LD) and creatine kinase (EC 2.7.3.2; CK) were separated by “high-performance” liquid chromatography and detected by continuously monitoring the column effluent for enzyme activity. Such background peaks were particularly apparent in CK isoenzyme profiles obtained from human sera. We observed two nonenzymic peaks with fluorescence detection, one in the CK-MB region, the other in the CK-BB region. Serum albumin was a major component in the artifactual CK-MB peak, with lipoprotein as a minor component. We present evidence that the material responsible for the other peak fluoresced quite strongly and is mostly pre-albumin. AddItIonal Keyphrases: “high-performance” tography lactate dehydrogenase
liquid chroma-
Creatine kinase (EC 2.7.3.2; CK)4 isoenzymes have been separated electrophoretically and by liquid chromatography. In either technique, evaluation of the proportions of the isoenzymes requires determination of their activity after the separation. However, kinetic assays that rely on a single reading are susceptible to interference from endogenous materials, especially when their absorbance and fluorescent properties are similar to those of the enzymic product. Because human serum fluoresces under near-ultraviolet light (1, 2), reports (3,4) of nonreacting, interfering peaks in the patterns for (profiles of) CK isoenzymes are not surprising. The fluorescing materials have not yet been completely identified and further elucidation is needed. An “albumin-like” artifact with an electrophoretic mobility that may be similar to that of CK-BB, depending on the medium, has been reported (5, 6). There are preliminary indications that this is an albumin complex (7), and bilirubin (8) has been tentatively identified as the partner in the complex. Early reports focused on the presence of a nonenzymic artifact in the CK-BB region (3, 4, 9), but a second, more cathodic artifact was described later (10). Reports (11, 12) on the separation of some other serum isoenzymes by “high-performance” liquid chromatography (HPLC) also describe such interferences. HPLC separation of lactate dehydrogenase (EC 1.1.1.27; LD) isoenzymes in serum revealed a nonenzymic fluorescent material, appearing as a peak eluting after LD-1 (13). This peak seemed identical to the artifact observed by McKenzie and Henderson (14), a peak appearing in the fluorometric assay of serum LD-isoDepartment
of Pathology,
University
of South
Alabama,
Mobile,
enzymes after their electrophoretic separation on agarose. Fluorescent background peaks were also seen in the LD-2 to LD-1 region when these isoenzymes were separated by HPLC and measured with and without post-column detection of enzyme activity in the effluent (15). Addition of both albumin and bilirubin to a preparation of LD isoenzymes produced a greater background peak than did the addition of albumin alone, when LD isoenzymes were separated by HPLC with post-column detection of activity (16). A nonenzymic fluorescent peak was also seen in the CK-MB region when CK isoenzymes in serum were separated by HPLC (17). We also observed nonenzymic peaks, which became increasingly more pronounced as we improved the sensitivity (18). We have recently described (19) the two-point kinetic method and the computer-algorithm used in this work to distinguish between background peaks (interferences, nonenzymic peaks) and activity profiles (enzymic peaks); the latter increase in area with time and the former remain constant. The method was subsequently demonstrated for LD activity profiles (12). Here we use this method to investigate the properties of the nonenzymic peaks that appear when serum samples are separated by HPLC, then examined in the presence and absence of detection of CK activity. We present chromatographic data confirming that proteins account for much of these nonenzymic peaks, and we tentatively identify them.
Materials
and Methods
Apparatus Figure 1 shows our apparatus. We used a gradient HPLC (Model 8500; Varian Instrument Division, Palo Alto, CA 95303) with a reagent pump (Model 3500; Spectra-Physics, Sante Clara, CA 95051). Sample injections were made with Rheodyne Model 7120 valves (Alitech Associates, Houston, TX 77055). Absorbance was measured with a dual-channel detector, equipped with 340- or 280-nm filters (Model 440; Waters Associates, Milford, MA 0175!), fluorescence with Aminco FluoroMonitors (Model J4-7461; American Instrument Co., Silver Spring, MD 20910), equipped with Corning 7-60 primary and Wratten 2A secondary filters. We acquired detector data with a DEC LAB 11/V03 computer (Digital Equipment Corp., Maynard, MA 01754) and processed and plotted these data with the DEC GT-46 system described earlier (15). We made post-column reaction coils from 508 zm (i.d.) stainless-steel tubing (Ailtech Associates). We have described the chromatographic column earlier (12).
AL 36688.
Reagents
Present address: Varian Instrument, 2700 Mitchell Dr., Walnut Creek, CA 94598. 2 Present address: Western Electric ERC, P. 0. Box .900, Princeton, NJ 08540. To whom correspondence should be addressed. Nonstandard abbreviations used: CK, creatinine kinase (EC 2.7.3.2); HPLC, high-performance liquid chromatograph(yi; LD, lactate dehydrogenase (EC 1.1.1.27). Received Nov. 19, 1979; accepted Feb. 6, 1980.
Two buffers were used for gradient elution. The “weak” buffer was 20 mmol/L tris(hydroxymethyl)aminomethane, adjusted to pH 7.8 (buffer A); the “strong” buffer contained 0.30 mol of sodium chloride per liter of buffer A. Prepare the post-column CK assay reagent from a CPK-10 reagent kit (Biodynamics, Indianapolis, IN 46250). Combine four such vials after reconstituting each with 50 mL of the CK buffer (per liter, 400 mmol of triethanolamine, 40 mmol of
1
CLINICAL CHEMISTRY, Vol. 26, No. 6, 1980
707
70.
Fig. 1. Schematic diagram of the system for separating CK isoenzymes Gadlent HPLC (A) was used for chrornatoaphy. The pre-column injection valve (C) was equipped with a 216.7-zL loop; the post-column injection valve (E) was not used. The chromatographic column (D) measured 4.2 X 180 mm. The re agent pump (B) introduced the CK assay reagent or buffer A through the 1.6-mm tee connection (F) after reagent passage through the 3.05-m heating coil (0). The 15.25-rn pre-incubation coil (l led into the first absorbance and fluorescence detectors (Ji and K1. respectively). The 30.5-rn incubation coil (0 separated the first detectors from the second absorbance and fluorescence detectors (J2 and K2, respectively). The response from each detector was multiplexed to an analog-to-digital converter (L) that was part of our computer. The dotted line represents a water bath maintained at 37 #{176}C
A
B 0.00
D-glucose, and 20 mmol of magnesium acetate; mixture adjusted to pH 7.0), and add a vial of hexokinase (EC 2.7.1.1) and glucose-6-phosphate dehydrogenase (EC 1.1.1.49) coupling enzymes (cat. o. H8629; Sigma Chemical Co., St. Louis, MO 63178). Keep this reagent in an ice bath for the duration of the experiment. Rabbit antiserum to human pre-albumin and goat antiserum to rabbit immunoglobulins were obtained from Behring Diagnostics, Somerville, NJ 08876. The CK isoenzyme control was obtained from Sigma (cat. no. C0300). Frozen serum samples were obtained from the University of South Alabama Medical Center. Sera from a healthy 30-year-old man were also used and assayed on the same days collected. MM
I-
z H
-J
U
S.
A
B
TIME Fig. 2. Fluorescence profile of CK control Trace
B was recorded
(MIN) activity
at the first fluorescence
in
detector.
second 708
CLINICAL CHEMISTRY, Vol. 26. No. 6, 1980
an isoenzyme Trace
A at the
TIME
(MIN)
Fig. 3. Fluorescence profile of CK isoenzymes in serum from a patient who had suffered a myocardial infarction Trace A was recorded at the first fluorescence
detector,
Trace
B at the
second
Procedures Divide the serum sample into two 1 .0-mL aliquots. Dialyze one aliquot for 24 h vs isotonic saline containing 10 mmol of tris(hydroxymethyl)aminomethane buffer, pH 7.4, per liter. Store the other aliquot at 4 #{176}C. Divide serum obtained from the same donor into two 1.0-mL aliquots. To one aliquot add 0.1 mL of a 40 g/L phosphotungstate solution containing 0.5 mol of magnesium chloride per liter; this solution was obtained from an HDL-Cholesterol kit ( Biodynamics Inc., Indianapolis, IN 46250). To the other add 0.1 mL of buffer A. Vortexmix both samples for 5 mm and then centrifuge at 1500 X g for 30 mm. Apply the supernates to the column. To one of two 100-tL aliquots of serum obtained from the same source as above, add 200 zL of anti-human pre-albumin. To the other add 200 1zL of buffer A. Incubate each aliquot for 1 h at 37 #{176}C and allow to equilibrate overnight at room temperature. To each add 100 L of anti-rabbit immunoglobulins and 500 tL of 120 g/kg polyethylene glycol (Mr 6000) in distilled water. Vortex-mix the samples and centrifuge at 2100 X g for 1 h. Chromatography. Chromatograph serum samples with a linear gradient progressing from 50 to 1000 mL of strong buffer per liter in 24 mm. Separate serum proteins and CK isoenzymes with the chromatographic column (D in Figure 1). To detect CK isoenzymes, combine the separated isoenzymes with CK reagent (F), monitoring the NADPH produced before and after the incubation coil (I). Measure both absorbance (J1, J2) and fluorescence (K1, K2). (We included the pre-incubation coil (H) to allow the CK reaction to pass through the lag phase.) Determine CK activity by subtracting the peak areas recorded at the first detector (J1 or K1) from those at the second (J2 or K2). To obtain serum protein profiles, substitute buffer A for the CK reagent, then record the background absorbance profile of serum proteins at 280 nm. We used the profiles recorded at the first absorbance or fluorescence detector because they exhibited better resolution.
0.400
20.0
MM
I-
z
1#{149}
H
I
U)
MS
0.000
TIME
(MIN)
TIME
i.6
Fig. 4. Absorbance (340-nm) profile of CK isoenzymes in the same serum as used in Fig. 3 Trace B was recorded at the first absorbance detector, Trace A at the second
Results Figure 2 shows the profile of CK isoenzymes in a control sample having about 200 U of CI( activity per liter. The lower profile, Trace B, shows the response of the first fluorescence detector; the upper profile, Trace A, the response of the second. About 90% of the applied activity was recovered. No background peaks (peaks that remain the same in area) are apparent. Replacing the CK reagent with buffer A also did not produce background fluorescence peaks. Figure 3 shows the fluorescence profiles of CK isoenzymes in a serum sample from a patient who had recently suffered a myocardial infarction. Subtracting the two fluorescence profiles revealed about 120 U/L CK-MM activity and about 15 U/L CK-MB activity. Subtraction also revealed the presence of a background fluorescence peak on the front of the CK-MB peak in the 8- to 10-mm region and one after CK-MB in the 12- to 14-mm region. Replacing the CK reagent with buffer A eliminated the activity peaks but not the background peaks. Absorbance (340 nm) profiles of CK activity, obtained simultaneously with the fluorescence data shown in Figure 3, revealed comparable amounts of CK-MM and CK-MB activity (Figure 4). However, the background peak just before CK-MB appears much larger, and that after CK-MB much smaller, than observed in the fluorescence data. These background peaks also remained when buffer A was substituted for CK reagent. Figure 5 (Trace A) shows the background fluorescence profile of a serum sample. The background profile was similar to that found with other serum samples when the CK reagent was replaced with buffer A. The first background peak, in the 4- to 6-mm region, consistently appears when more than 100 tL of serum are applied to the column; that peak is likely caused by column overloading. Trace B in Figure 5 demonstrates the effect of dialysis on the background fluorescence profile. Dialysis decreased the area of the background peaks in the 8- to 10- and 12- to 14-mm regions by less than 30%;
OlIN)
Fig. 5. Effect of dialysis on the fluorescence
16 profile of serum
Serum was chromatographed under the same conditions as before, except that the CX reagent was replaced with buffer A. The background fluorescence profile of serum before dialysis is shown in Trace A; after dialysis, Trace B
therefore, less than 30% of the background fluorescence can be directly attributed to small molecules. Precipitation of serum protein with the phosphotungstate reagent dramatically decreases the fluorescent background peaks in the 7- to 10- and 12- to 14-mm regions (Figure 6). Trace A represents the background fluorescence profile before precipitation, Trace B, the profile after precipitation of serum 2S00
I-
z
H
I
U
B
TIME
(MIN)
Fig. 6. Effect of precipitating serum proteinon the background fluorescence profile of serum: A, before precipitation; B, after precipitation CLINICAL
CHEMISTRY,
Vol. 26. No. 6, 1980
709
Trace
4.20
I-
zH -J U
A B
0.00,
TIME
16
(MIN)
Fig. 7. Comparison of the absorbance (280 nm) ( Trace A) and fluorescence (Trace B) profiles of serum proteins protein. The peak in the 8- to 10-mm region that remains after precipitation may represent high-density lipoprotein, which reportedly remains in the supernatant fluid (20). The large peak in the 4- to 6-mm region in Trace B is caused by the release of bound material from the column by the precipitating reagent. Clearly, the substitution of buffer A for the enzyme reagent reveals primarily the fluorescence profile of serum proteins. Figure cence
A, shows the immunoglobulmns
resolved lipoproteins which is, as expected,
7 compares profiles
the
of serum
absorbance proteins.
(280 The
nm)
absorbance
and
first, followed
that overlap with the albumin the largest peak. The shoulder
by poorly peak, peak at
about 12 mm is probably pre-albumin, the portion of serum proteins that migrates ahead of albumin during electrophoresis (21). Comparing this profile with the fluorescence profile, Trace B, reveals that the background peak in the 8- to 10-mm region is probably albumin. The background fluorescence peak in the 12- to 14-mm region corresponds to the small peak appearing after albumin, tentatively identified as pre-albumm. The absence of a peak in the 4- to 6-mm region in Trace B was attributable to the use of a smaller injection volume, 102.8 L instead of 216.7 1zL. Figure 8 depicts the background fluorescence profiles of serum, Trace A, and a 40 g/L solution of human serum albumin, Trace B. The position of the albumin peak in Trace B coincides with the peak in the 8- to 10-mm region in Trace A. This further substantiates the identification of the protein peak eluting in that region as albumin. Removal of pre-albumin from serum by immunoprecipitation with antiserum produces the fluorescence profile shown in Trace A (Figure 9). The control serum sample is shown in Trace B. The fluorescence background peak in the 8- to 10mm region is larger in Trace A because of the contribution of the albumin in the antiserum. The antiserum was specific for human pre-albumin. The conspicuous absence of a background peak in the 12- to 14-mm region in Trace A but not Trace B strongly suggests that this protein is pre-albumin.
Discussion Background fluorescence or absorbance peaks are not usually observed in profiles of CK isoenzymes obtained from control materials in our system, probably because of their higher CK activity and lower bulk protein content than human serum. Although these routine interferences in serum profiles of CK isoenzymes are removed by our two-point measurement of enzyme activity (19), we sought to characterize the background peaks. We demonstrated by the simple
fluoresprofile,
5.00
20.00
U
CO
z 0 a.
I-
z
H
U)
U
I
U
A
B
0.00 TIME
(MIN)
1.6
Fig. 8. Comparison of the fluorescence profile of serum proteins (Trace A) with that of human serum albumin (Trace B) 710
CLINICAL
CHEMISTRY,
Vol. 26, No. 6, 1980
TIME
(MIN)
16
Fig. 9. Comparison of the fluorescence profile of serum proteins before (Trace B) and after(Trace A) removal of pre-albumin by immunoprecipitation
Gun. Chem. 24, 1084-1085 (1978). Letter.
method of dialysis and protein precipitation that these background peaks were probably nonenzymic proteins. By comparing the serum protein profile with an albumin standard, we showed that one of the peaks was probably albumin. By comparing the fluorescence profile of serum proteins before and after removal of pre-albumin, we demonstrated that the other background peak was probably pre-albumin. Although the absorbance at 280 nm of the background peak in the 12- to 14-mm region was fairly low, the fluorescence intensity of that peak was quite high. This observation also suggests that the background peak in the 12- to 14-mm region is pre-albumin because pre-albumin is rich in tryptophan (21), the most fluorescent of all the amino acids (22). We have found that the major interferences-i.e., background peaks-in CK isoenzyme profiles of serum are albumin and pre-albumin, with a minor contribution possibly from lipoprotein. Although our system differs from simpler electrophoretic systems, we suggest that the background fluorescent artifact commonly observed in the CK-BB region on
chronic
electrophoresis
14. McKenzie, D., and Henderson, A. R., An artifact in lactate dehydrogenase isoenzyme patterns, assayed by fluorescence, occurring in the serum of patients with end-stage renal disease requiring maintenance dialysis. Gun. Chim. Acta 70, 333 -336 (1976).
may
be pre-albumin.
We thank Dr. Dorothy Vacik for assistance in preparing this manuscript, Dr. William Dean of Corning Glassworks for the column support material, and Dr. Paul Schneider of Biodynamics bmc, for the CK assay reagents. This work was supported by a grant from the NIGMS, NIH, grant no. GM 24452.
References 1. Soini, E., and Hemmila, I., Fluoroimmunoassay: andkey problems. Clin. Chem. 25, 353-361 (1979).
Present
status
2. Coolen, R. B., Herbstman, R., and Hermann, P., Spurious brain creatine kinase in serum from patients with renal disease. Chn. Chem. 24, 1636-1638 (1978). 3. Van Lente, F., and Galen, R. S., Electrophoretic identification of the brain isoenzyme of creatine kinase following treatment with anti-BB antisera. Curt. C/tim. Aeta 87, 211 -217 (1978).
4. Aleyassine,
H., Tonks, D. B., and Kaye, M., Natural
fluorescence
in serum of patients with chronic renal failure not to be confused with creatine kinase-BB isoenzyme. Gun. Chem. 24, 492-494 (1978). 5. Galen, R. S., Tips on technology. Med. Lab. Observer 10, 17-18 (Feb. 1978). 6. Chuga, D. .1., and Bachner, P., Creatine kinase isoenzyme BB in the serum of renal-disease patients, distinct from an albumin-like artifact. Gun. Chem. 24, 1286 (1978). Letter. 7. Vladutiu, A. 0., Cunningham, E. E., and Waishe, .J., Attempts to characterize the fluorescent compound(s) in serum of patients with
renal
failure.
8. Chuga, D. J., Fluorescent albumin-bilirubin complex on cellulose acetate electropherograms, which may confuse creatine kinase isoenzyme assay. Clin. Chem. 25, 494 (1979). Letter. 9. Aleyassine, H., and Tonks, D. B., Albumin-bound fluorescence: A potential source of error in fluorometric assay of creatine kinase BB isoenzyme. Clin. Chem. 24, 1849-1850 (1978). Letter. 10. Gerson, B., and Petersen, K., Creatine kinase isoenzyme BB and a fluorescent artifact in hemodialysis patients’ sera. Clin. C/tern. 25, 1518-1519 (1979). 11. Schroeder, R. R., Kudirka, P. J., and Toren, E. C., Jr., Enzymeselective detector systems for high-performance liquid chromatography. J. Chromatogr. 134, 83-90 (1977). 12. Fulton, J. A., Schlabach, T. D., Kerl, J. E., and Toren, E. C., Jr., Dual-detector-post-column reactor system for the detection of isoenzymes separated by high-performance liquid chromatography: II. Evaluation and application to lactate dehydrogenase isoenzymes. J. Chromatogr. 175, 283-291 (1979). 13. Schlabach, T. D., Alpert, A. J., and Regnier, F. E., Rapid assessment of isoenzymes by high-performance liquid chromatography. Gun. Chem. 24, 1:351-1360 (1978).
15. Schlabach,
T. D., Ph.D.
dissertation,
Purdue
University,
1978.
16. Schlabach, T. D., Fulton, J. A., Mockridge, P. B., and Toren, E. C., Jr., New developments in analysis of isoenzymes separated by high-performance liquid chromatography. Gun. Chem. 25, 1600-1607 (1979). 17. Denton, M. S., Bostick, W. D., Dinsmore, S. R., and Mrochek, J., Chromatographic separation and continuously referenced, on-line monitoring of creatine kinase isoenzymes by use of an immobilized enzyme reactor. Clin. C/tern. 24, 1408-1413 (1978). 18. Schlabach, T. D., Fulton, J. A., Mockridge, P. B., and Toren, E. C., Jr., Serum isoenzyme profiles by high performance liquid chromatography. Anal. Chem. 52 (in press, scheduled for April 1980). 19. Fulton, J. A., Schlahach, T. D., Kerl, .J. E., et al., Dual.detectorpost-column reactor system for the detection of isoenzymes separated by high-performance liquid chromatography: I. Description and theory. J. Chromatogr. 175, 269-281 (1979). 20. Burstein, M., Scholnick, H. R., and Morfin, the isolation of lipoproteins from human serum polyanions. J. Lipid Res. II, 583-595 (1970).
R., Rapid method for by precipitation with
21. Antoniades, H. N., Circulating hormones. In The Plasma Proteins, 3, F. W. Putnam, Ed., Academic Press, New York, NY, 1977, p 395. 22. Guilbault, C. G., Practical Huorescence: Theory, Methods, and Techniques, Marcel Dekker, New York, NY, 1973, p107.
CLINICAL CHEMISTRY, Vol. 26, No. 6. 1980
711
