CLIN. CHEM.37/3, 430-434 (1991)
Macromolecular Creatine Kinase Type 1: a Serum Marker Associated with Disease
MurielLaureys,’Jean-PaulSlon,’ HansSlabbynck,2Laurette Steenssens,3 ChrlstaCobbaert,4Marle-PauleDerde,5and FransK.
Gorus’
The prevalence of circulating macromolecular creatine
kinase type 1 (macro CK type 1 or CK-immunoglobulin
complexes) is significantly higher in a patient population
selected for CK isoenzyme assay than in age- and
sex-matched blood donors (n = 1304). In >8000 patients
studied, 49 individuals with macro CK type 1 were identified, yielding an overall prevalence of 0.61%. Macro CK
type 1 complexes occurred more frequently in women and
in patients older than 70 years, and were often associated
with complications of cardiovascular disease, life-threatening conditions, and poor outcome. These latter clinical
associations could arise, at least partly, from the selection
of patients for whom CK isoenzyme analysis was ordered.
Addftlonal Keyphrases:
cardiovascular disease
macroenzymes
autoantibodies
autoimmunity
Circulating enzyme forms with higher-than-normal
molecular mass-so-called
macroenzymes-have
been
described for several clinically
useful enzymes (1-4).
13.8%, depending on the type of enzyme, patient group,
or methodology (1, 2, 12).
Circulating CK-Ig complexes (macro CK type 1) have
been studied most, probably because their presence
readily
causes unexplained
or disproportionate
increases in apparent CK-MB activities as measured with
widely used rapid diagnostic tests, based on immunoinhibition or ion-exchange chromatography (1, 2, 11). In
electrophoresis of CK isoenzymes, the presence of an
atypical enzyme fraction migrating between CK-MM
and CK-MB is highly suggestive of macro CK type 1,
whereas (high-performance) gel-permeation chromatography (HPGPC) can confirm the macromolecular nature
of this enzyme activity and distinguish it from (fluorescent) artefacts or other enzyme activities (1-3).
Several studies have attempted to define the prevalence of macro CK type 1, but only a few involved a large
number of patients (1, 2, 13-15). The largest studies so
far have yielded estimates of 0.9%-1.2% for overall
macro CK type 1 prevalence in patients, with the
Certain macroenzyme species have been clearly associated with disease states, e.g., oligomeric mitochondrial
creatine kinase (CK; EC 2.7.3.2; macro CK type 2) with
deep cellular necrosis (5, 6), and alkaline phosphatase
highest values in females and older subjects (2, 13-16).
No clear-cut association of macro CK type 1 with a
particular
type of disease has as yet been established,
although the macroenzyrne has occasionally been de-
(EC 3.1.3.1)-containing
membrane fragments with
cholestasis (7)7 Binding of circulating enzyme molecules to specific immunoglobulins
(Igs) constitutes yet
another common mechanism for macroenzyme formation. Thus far, the latter forms have not emerged as
specific disease markers, although occasionally their
appearance has paralleled autoimmune disease activity
tected in pediatric patients and even in apparently
healthy individuals (2, 14, 17). However, because comparative data in healthy control populations are largely
missing, it is impossible to decide whether enzyme-Ig
complexes represent disease-associated markers or instead reflect physiological or preclinical events related
to, e.g., aging or variations in natural autoimmunity
or
immune
repertoire (18-21). We therefore compared the
prevalence of the CK-Ig complex, macro CK type 1, in
two populations: a patient population selected on a
(1,
2,
8-10).
The presence of enzyme-Ig
complexes can lead to
misinterpretation
of concomitantly increased (iso)enzyme activities (1, 2, 11) but, even more importantly,
raises questions about their possible pathogenetic or
diagnostic
significance.
The widespread
implementation
of methods for rapid indirect estimation
retic separation
of isoenzymes in clinical
facilitated
a more systematic detection
complexes in recent years (1, 4) and
estimates of their prevalence, ranging
or electrophochemistry
has
of enzyme-Ig
allowed rough
from 0.1% to
Departments of ‘ Clinical Chemistry, 2lnternal Medicine, and
5Biostatistics, Akademisch Ziekenhuis Vrije Universiteit Brussel,
Laarbeeklaan 101, B-1090 Brussels, Belgium.
3Blood Transfusion Center Jette, Brussels, Belgium.
4Department
of Clinical Chemistry,
Universitair
Ziekenhuis,
St. Raphael, Leuven, Belgium.
6 Author
for correspondence.
Nonstandard abbreviations:
CK, creatine kinase; Ig, immunoglobulin; and HPGPC, high-performance
gel-permeation
chromatography.
Received September 17, 1990; accepted December ii, 1990.
430
clinical basis for CK-isoenzyme
analysis and an ageand sex-matched population of apparently healthy blood
donors.
CLINICAL CHEMISTRY, Vol. 37, No. 3, 1991
Materials and Methods
Subjects
The patient population studied consisted of 2140 individuals (1204 men and 936 women, including both inand outpatients) for whom a CK isoenzyme analysis was
ordered in our laboratory during 1987 for suspected
cardiac, muscular, or neurological disease and (or) increased total CK activity.The age group between 20 and
70 years comprised 1304 patients (794 men, 510 women). An age- and sex-matched control group consisted of
794 men and 510 women blood donors (ages 20-69
years). For each 10-year age class, the number of men
and women was identical to that of the corresponding
category in the patient population. Belgian blood donors
are selected according to a standardized protocol, including collection of anainnestic data (absence of medication, drug abuse, recent surgery, and various chronic or
acute affections), physical examination (absence of injection marks on body, apyrexia, good general condition,
weight >45 kg, normal blood pressure and pulse), and
blood analysis (negative serology for several bloodborne
infectious diseases, absence of anemia, normal liver
tests, and protein electrophoresis). In addition, we reviewed the medical files of all patients (n = 49) in whom
macro CK type 1 was detected between January 1987
and April 1990 among a population of 8080 patients
selected for CK isoenzyme assay as above.
Blood Samples
Blood was obtained by venipuncture of the antecubital vein, collected in commercial dry-plastic tubes
(Sarstedt, Haasrode, Belgium,
for patients, and Laborimpex, Brussels, Belgium, for blood donors), allowed
to clot at room temperature, and centrifuged within 4 h
after sampling at 1000 x g for 15 mm. Sera were stored
as previously described (3).
Detection of Macro CK Type 1
Macro CK type 1-containing samples were identified
on basis of the following criteria: (a) apparently increased CK-B activity (>15 U/L) determined by immunoinhibition
assay (CK-MB, N-acetyl-L-cysteine-activated; Boehringer Mannheim, Mannheim, F.R.G.) performed at 30#{176}C
with an RA-1000 analyzer (Technicon,
Dublin, Ireland), which detects macro CK with 100%
sensitivity
but low specificity (3, 11); (b) in CK-Bpositive sera, the demonstration of an abnormal CK
band, typically migrating between CK-MM and CK-MB
markers (1,4), determined by ParagonE agarose electrophoresis (Beckman Instruments,
Brea, CA) and quantified by densitometric scanning as previously described
(3); (c) in sera with suggestive isoenzyme profiles, the
presence of a peak of CK activity eluting at higher-thannormal Mr during HPGPC of whole serum (3). All samples
with residual CK-B activity >20% of total CK activity
and with macro CK type 1-like activity of at least 10
U/L, as judged by electrophoresis, proved positive for
macro CK type 1 after HPGPC analysis. Failure to detect
a high-Mi CK peak in a few samples appeared related to
low macro CK type 1 activity by electrophoresis or
pseudo-macro CK type 1 bands attributable to fluorescent artefacts in hyperlipemic sera (3). In samples with
confirmed macro CK type 1, the macroenzyme activities
estimated by HPGPC correlated well with electrophoretic
data (3). No unstable complexes were noted; however,
catalytically inactive complexes would go unnoticed if
present (3).
Statistical Analysis
For differences
in macro CK type 1 prevalence between various subject groups, we evaluated statistical
significance by means of the Fisher exact test (Table 1)
or the chi-square test with Yates correction (Table 2).
The 95% confidence intervals for prevalence values were
Table 1. Macro CK Type 1 Prevalence in Patients and
Blood Donors
n
Ages,
Mean (SD)
Age,
Total Men
Women
years
Patients selected for CK electrophoresis
All
2140
1204
936
61(18)
y
Prevalence
of macro
Total CK,
CK 1, %
U/L
(no.)
255 (709) 0.61 (13)
70
793
383
410
79(6)
279 (798) 0.76(6)
20-69
1304
794
510
51(13) 237 (613)8 0Mb (7)
Blood donors
20-69
1304
794
510
51(12)
86(46)
0 (0)
8 P <0.001 vs blood donors.
P <0.016 vs blood donors. According to Bonferroni adjustments, differences in prevalence were significantfor P <0.017.
computed as described (22). The primary
hypotheses
tested in this investigation were (a) Is the prevalence of
macro CK type 1 in a patient population (n = 1304; ages
20-69 years) significantly
different from that in a group
of blood donors, matched for age, sex, and extent (Table
1)?‘and (b) Do differences in macro CK type 1 prevalence
between men and women, and between younger (20-69
years) and older (70
years) subjects occur within a
(larger) patient population (n = 8080; Table 2)? According to Bonferroni adjustments, differences in prevalence
were considered significant if P <0.0 17. Mean values of
age and (macro) CK activity were compared in different
subject groups by the Student’s t-test for independent
measurements (if n >30 for the samples compared) or by
the Mann-Whitney test for smaller samples. All statistical tests were carried out two-sided at the 5% level of
significance, unless indicated otherwise.
Results
Prevalence of Macro CK Type 1 in Patients and Blood
Donors
Figure IA represents the age and sex distribution of
the 2140 patients studied, who were selected for routine
CK-isoenzyme analysis during 1987. Among this population, 13 cases with circulating
macro CK type 1
activity were identified according to the criteria defined
in Materials
and Methods; all patients but one were
older than 50 years and nine were females (Figure lB.).
Patients tested for CK isoenzyme composition were also,
in general, older than 50 years (mean age
SD, 61 ± 18
years) but 58% were men (Figure IA). In contrast, of the
total patient population attending our academic hospital during the same one-year period, only 38% were
older than 50 years and only 44% were men (Figure 1C).
We next investigated whether the relatively low prevalence of macro CK type 1 in the studied patient
population (13 of 2140, or 0.61%) was significantly
higher than that in apparently healthy blood donors
and, if so, whether age- or sex-related variations could
be demonstrated. Because Belgian blood donors are
recruited
exclusively among 20- to 69-year-old subjects,
the control group was matched with the patient group
for age, sex, and numbers (Table 1). In the patient
group, both total CK activity and macro CK type 1
CLINICAL CHEMISTRY, Vol. 37, No, 3, 1991
431
Table 2. Macro CK Type 1 Prevalence According to Age and Sex
Mean (SD)
Subjects
All
Women
Men
20-69
>70 y
n
49
33
16
21
28
Total CK acty, tilL
145 (79)
151 (78)
133(83)
147(86)
143 (75)
Ag., years
71(14)
74(14)
67 (12)
58(10)
81(6)
Macro CK type 1
acty, tilL
56(67)
58(62)
52(67)
44(59)
65(66)
Prevalence, no. (%)
49/8080 (0.61)
33/3531 (0.93)
16/4549 (0.35)8
21/4921 (0.43)
28/2990 (#{216},g4)t
95% confidence
Interval, %
0.45-0.80
0.64-1.31
0.20-0.56
0.26-0.65
0.62-1.30
P <0.002 vs women.
P <0.01 vs age group 20-69 years. Based on Bonferroni adjustments, differences in prevalence are significant for P <0.017.
P.o..,i.
SAl..,.:
..
tk.Lini,.ph.,..,#{149}
-
PS...I.
*th
U.n.
nrfflH
#{149}ti,,i. ..U,
CX
U.Ci.
CX
Ta.
CA
-
LI
I
I-
-‘
rnrri
-
H..p,i.i
P#{149}01s,.:i
I!TirlTfftki
I..
SS
ri
ii
nfl
N..p,i.i
9r
P.pul.,I.,
i!ft-Fftrffri
A..
4..
ri
Fig. 1. Age- and sex-distribution(men, left panels; women: right
panels) of patients selected for CK electrophoresis, patients scored
positive for circulating macro CK type 1 complexes,and patients
attendingour hospitalduring 1987
prevalence were significantly higher than those in blood
donors, for whom no case of macro CK type 1 was
detected (Table 1). The macro CK type 1 prevalence was
not significantly
higher in patients 70 years than in
the 20-69-year age group, nor was the macro CK type 1
prevalence significantly higher in women than in men.
Clinical and Biological Findings in 49 Patients with Macro
CK Type 1
To eliminate the absence of sex- and age-linked differences in macro CK type 1 prevalence because of the
limited number (n = 13) of patients with macro CK type
1, we analyzed clinical and biological data from 49
patients (ages 30-1 00 years) with detectable amounts of
circulating macro CK type 1, identified in our laboratory between January
1987 and April 1990. For this
period the overall prevalence of macro CK type 1 in
patients was also 0.61% (95% confidence intervals: 0.450.80%) (Table 2). These subjects with macro CK type 1
were predominantly
women and older than 70 years
(Table 2). The men and women did not differ signifi432
CLINICALCHEMISTRY, Vol. 37, No. 3, 1991
nor did older
years) and younger patients (20-69 years) differ
with respect to enzyme activities (Table 2). Finally, the
prevalence of macro CK type 1 in the selected patient
population was significantly
higher in females and in
patients
older than 70 years (Table 2).
Review of the patients’ medical files revealed that the
LI
i
cantly in age and in macro CK activity,
(70
9
presence of macro CK type 1 could be associated with
various pathologies, but most frequently with cardiac,
vascular, and neurological complications of cardiovascular disease (51%) (Table 3). Although no clinical
information
was available for four patients, at least 19
of the remaining 45 subjects had very severe illnesses,
as evidenced within a two-year follow-up, or biological
markers of deep necrosis or cancer, such as CK-BB,
mitochondrial
CK, or monoclonal components (1,2,5,6)
(Table 3). Total (macro) CK activity was, however, not
significantly higher in these patients, compared with
the other macro CK type 1-positive patients. About 15%
(seven of 49) of the macro CK type 1-positive patients
had other clinical or biological evidence of autoimmune
Analysis of the referring clinical units showed
that 42% of all CK isoenzyme assays were ordered by
disease.
the departments of cardiology, neurology, or intensive
care, the latter unit alone accounting for 38% of the
demands, primarily for suspected cardiac or neurological problems.
Discussion
First described in 1979 (23), circulating
Ig-CK complexes (macro CK type 1) have been observed in pa-
Table 3. ClInical Findings In 49 Macro CK Type
1-Positive Patients
ClinIcal FindIngs
Complications of cardiovascular disease
Suspected (pre)cancerous lesions
Gastrointestinaldisease
Traumaor diseaseof muscles,bones,or joints
No. of patIents
25
10
Lung disease
10
8
7
4
No relevant clinical information available
Severeillness;poor outcome8
4
19
Markers of autoimmune disease
Withintwo years of follow-up:death, cachexia, precancerouslesions,local
or disseminated cancers, or biological markers of deep necrosisor cancer,
such as CK-BB, mitochondnalCK, or monoclonal components.
tients, including children, with various diseases, as well
as in apparently
healthy individuals
comparing
the prevalence of macro
(1, 2, 13-17). By
CK type 1 in a
with that of age- and sex-matched
patient population
blood donors, we document here that such Ig-CK complexes are indeed preferentially associated with disease
states.
As in previous
reports (1, 2), macro CK type 1 was
observed most often in women, despite the predominance of men in the selected patient group. Autoimmune diseases and circulating autoreactive antibodies
are generally more frequent in women, perhaps in part
because of the immuno-enhancing
effects of estrogens
(24). Also in agreement with earlier data (1, 2, 12, 14,
16, 25), most macro CK-positive patients were older
than 50 years. The prevalence of macro CK type 1 in
patients >70 years exceeded that in the 20-69-year age
group. However, the occurrence of macro CK type 1
cannot be envisaged as a mere consequence of immune
senescence, because no such complexeswere observed in
older blood donors.
Why CK or other self-molecules might occasionally
become antigenic is at present unclear. Several hypotheses have been proposed, including (1, 2, 18-20,26) (a)
release during cell destruction of self-antigens
in an
altered physicochemical form or from sequestered sites,
(b) occurrence of molecular mimicry, (c) imbalance in
the idiotype-anti-idiotype
network as a consequence of
dysregulation of tolerance or, conversely, as a defense
mechanism against autoimmunity.
Some patients with
macro CK type 1 showed other clinical and biological
signs of autoimmunity,
which may point to a more
global dysregulation of the inunune system as an underlying cause for the occurrence of anti-CK antibodies.
However, the association of the latter with various
severe disease states indirectly
favors the view that
macro CK type 1 arises as a consequence of cellular
damage in patients, mostly women and elderly people,
with an increased tendency towards immune autoreactivity (24, 25). There is at present no indication that
anti-CK antibodies actually cause cell destruction.
Although no clear correlation between severity of
illness and titer of macro CK type 1 was observed, the
occurrence of the latter often coincided with important
clinical and biological findings,
including life-threatening complications of cardiovascular disease, deterioration of general condition, death within less than two
years after blood sampling, (suspicion of) generalized or
local cancer, and biochemical markers of deep necrosis
such as CK-BB and mitochondrial CK (6, 27). The
associated increase in total CK activity may be due to
decreased clearance of CK-Ig complexes and (or) increased CK release from damaged tissues (1,2). Biological and clinical findings in macro CK type 1-positive
patients could, at least in part if not completely, be
explained by patient selection. In this respect, the frequent association of macro CK type 1 positivity with
cardiovascular
disease could largely result from the
recruitment of patients for CK isoenzyme analysis in
cardiology, neurology, and intensive care units, whereas
poor outcome could also be linked to patient selection as
well as to old age.
Although the possible clinical and prognostic value of
macro CK type 1, if any, remains uncertain, its systematic detection by screening tests or electrophoresis is
still warranted
for the correct interpretation
of increased CK (isoenzyme)
activities
and as a putative
marker of associated (autoimmune) disease activity.
Further investigations on the Ig-components involved
in CK binding (e.g., IgA vs IgG) may yield new clues
regarding the mechanism of immune-complex formation or possibly autoimmune
disease in general. The
design of class-specific radio-binding
assays for the
quantification of CK autoantibodies, e.g., through monitoring the binding of radiolabeled CK-BB tracer to
serum
after dissociation of immune complexes in acid
medium, by analogy to insulin autoantibody assays (26),
could provide more-sensitive markers of CK autoimmunity that would remain independent of CK activity
measurements
and immune-complex formation. This
approach could give more reliable estimates of macro
CK type 1 complexes or CK autoantibodies; favor the
detection of high-affinity CK antibodies, which are more
likely to be clinically relevant; and disclose associations
between certain high-affinity
CK-antibody subclasses
and specific (autoimmune?)
diseases (1,2,8-10,
12,26).
In conclusion, macro CK type 1 is associated with
to occur as a marker or a consequence of cellular damage in a minority of possibly
predisposed individuals, predominantly women and elderly people.
disease and is likely
The efficient secretarial help of L. Vermeir and the dedicated
technical assistance of L. De Pree, M. De Winter, P. Goubert, F.
Lebleu, and E. Verkest is gratefully acknowledged. Part of the
blood donor samples were kindly provided by Dr. L. Muylle (Blood
Transfusion Center, Edegem, Belgium) and information on patient
populations by M. Duyck (Department of Informatics, Akademisch
Ziekenhuis Vrije Universiteit Brussel).
References
1. Remaley AT, Wilding P. Macroenzymes: biochemical characterization, clinical significance, and laboratory detection [Review].
Clin Chem 1989;35:2261-70.
2. Sturk A, Sanders GTB. Macroenzymes:
prevalence, composition, detection and clinical relevance [Review]. J Clin Chem Clin
Biochem 1990;28:65-81.
3. Sion JP, Laureys M, Gerlo E, Gores F. Detection of macroenzyines in serum by high-performance
gel permeation chromatography. J Chromatogr 1989;496:91-100.
4. Gorus FK, De Pree L. Effect of isoenzyme composition on Kodak
Ektachem test results for creatine kinase, lactate dehydrogenase,
alkaline phosphatase, and amylase. Clin Chem 1990;36:685-6.
5. Stein W, Bohner J, Renn W, Maulbetsch R. Macro creatine
kinase type 2: results of a prospective
study in hospitalized
patients. Clin Chem 1982;28:19-24.
6. Berk CJ. Mitochondrial creatine kinase. A poor prognostic sign.
J Am Med Assoc 1975;243:2058-60.
7. DeBroe ME, Borgers M, Wieme RJ. The separation and characterization
of liver plasma membrane fragments circulating in
the blood of patients with cholestasis. Clin Chirn Acts
1975;31:1952-8.
8. Gores F, Aelbrecht W, Van Camp B. Circulating
IgG-LD
complex, dissociable
by addition of NAD.
Clin Chem 1982;
28:236-9.
CLINICAL CHEMISTRY, Vol. 37, No. 3, 1991
433
9. Peters 0, Gores F, Van Camp B. NAD dissociable macromolecular lactate dehydrogenase [Letter]. Clin Chem 1982;28:1862.
10. Biewenga J, Feltkamp TEW. Lactate hydrogenase (LDH)IgG3 immunoglobulin complexes in human serum. Clin Chim Acts
1975;64:1O1-16.
11. Gorus F, Claessens V, Goubert P, Laureys M. A sensitive
bioluminescent immunoinhibition test for CK-B subunit
and a CK-MB specific ELISA compared: correlation with
electrophoresis and influence of CK-isoenzyine profile on
Clin Chem 1988;34:1174-8.
12. Delanghe J, De Scheerder I, De Buyzere M, Algoed
brecht
J. Macro CK type 1 as a marker
activity
agarose
results.
L, Robfor autoimznunity
in
coronary heart disease. Atherosclerosis 1986;60:215-9.
13. Wu AHB, Bowers GN Jr. Evaluation and comparison of
immunoinhibition and immunoprecipitation
methods for differentiating MB from BB, and macro forms of creatine kinase in
patients and healthy individuals. Clin Chem 1982;28:2017-21.
14. Wu AHB, Herson VC, Bowers GN Jr. Macro creatine kinase
types 1 and 2: clinical significance in neonates and children as
compared with adults. Clin Chem 1983;29:201-4.
15. Bohner J, Stein W, Steinhart
R, Wurzburg U, Eggstein M.
Macro creatine kinases: results of isoenzyme electrophoresis and
differentiation
of the imnmunoglobulin-bound
type by radioassay.
Clin Chem 1982;28:618-23.
16. Muffin TE, Bruns DE, Wrotnoski U, et al. Macroamylase,
macro creatine kinase, and other macroenzymes [Case Conference]. Clin Chem 1985;31:1743-8.
434 CLINICALCHEMISTRY,Vol.37, No.3, 1991
17. Whelan PV, Malkus H. A macro creatine kinase isoenzyme in
serum of apparently
healthy
individuals.
Clin Chem 1983;
29:1411-4.
18. Holmbergh D, Coutinho A. Natural antibodies and autoixninunity [Review]. Iminunol Today 1985;6:356-7.
19. Guilbert B, Dighiero G, Avrameas S. Naturally occurring
antibodies against nine common antigens in human sera. 1.
Detection,
isolation
and
characterization.
J
Immunol
1982;128:2779-87.
20. Cohen IR, Cooke A. Natural autoantibodies
might prevent
autoimmune
disease [Review]. Iminunol Today 1986;7:363-4.
21. Frazer CG, Cummings ST, Wilkinson SP, et al. Biological
variability of 26 clinical chemistry analytes in elderly people. Clin
Chem 1989;35:783-6.
22. Comniissariat a l’#{233}nergie
atomique. Statistique appliqu#{233}e
a
l’exploitation des mesures. Paris: Masson, 1973:79-80.
23. Urdal P, Landaas S. Macro creatine kinase BB in serum, and
some data on its prevalence. Clin Chem 1979;25:461-5.
24. Talal N. Autoimmunity and sex revisited [Editorial]. Clin
Immunol Immunopathol 1989;53:355-7.
25. Tomer Y, Shoenfeld Y. Ageing and autoantibodies [Review].
Autoimmunity
1988;1:141-9.
26. Palmer JP. Insulin autoantibodies: their role in the pathogenesis of IDDM [Review]. Diabetes Metab Rev 1987;3:1005-15.
27. Tsung JSH. Creatine kinase and its isoenzymes in neoplastic
disease. Crit Rev Clin Lab Sci 1986;23:65-75.