Road, Belfast BT12 6BA, UK. 2 Division. ofBiochemistry,. School ofBiology ... high sensitivity and high speci- ficity in differeptiati#{241}g pfe-eta. HDL from other.
‘
CLIN.
CHEM.
38/1 1 , 2273-2277
(1992)
Pre-Beta High-Density Lipoprotein Chemiluminescent Detection J. O’Kane,’
Maurice
G.
Wisdom,2
Brian
Determined
Jane
McEneny,3
We describe a novel assay of pre-beta high-density lipoprotein (HDL), a unique apolipoprotein A-I (apo A-I)containing lipoprotein particle. The pre-beta and alpha
lipoproteins and
are separated
transferred
onto
by electrophoresis
a membrane
in agarose
by capillary
blotting.
The membrane blot is sequentially incubated with sheep anti-human apo A-I antiserum and then with a conjugate of rabbit anti-sheep immunoglobulin and horseradish peroxidase.
Chemiluminescence
catalyzed enhancer
,‘ ‘
beta
-
r ,
1!
.
oxidation is captured
H9L
‘
band
formed
by the peroxidase-
of Iuminol in the presence on photographic film, and
of the
Neil V. McFerran,2 ago by Levy tively little
an pre-
is quantified
by transmission densitomewith standards prepared from a reference serum diluted in 9 mol/L urea. Within-batch r#{233}#{243}siofl (CV) at pre-beta HOL concentrations of 22.1 and 44.3 mg/L Was 7% and 4.9% respectively. Pre-beta HDL contained I .6% (0.65-2.6%, mean and range) of total serum apo A-I in 30 normolipidemic subjects. Keyphrases:
electrophoresis,
agarose
.
apoilpopro-
tein
Apolipoprotein A-I (apo A-I) is the major protein component of high-density lipoproteins (HDLs), and only trace amounts are present in other lipoprotein classes [very-low-density lipoproteins (VLDLs) and chylomicrons].4 A small but significant fraction of total
serum
apo
unique
particle
A-I
(0.5-10.9%)
is present
a separate,
and
refers
contrast
to the alpha ref#{235}isto the supposedly In fact, pre-beta HDL
density, to
which
has
el#{233}ctrophoretic
not
been
mobility,
in
mobility of HDL. Free ape A-I lipid-poor nature ofthe particle. appears to have a molecular mass
of -43 kDa (4), contains apo A-I as the only protein component, and is composed mostly (50-70% by weight) oflipid, mainly phospholipid (5). For this reason free apo A-I is a term best avoided and pre-beta HDL is pre-
an important The has
pre-beta
HDL
was first identified
>20
it has metabolic
(6),
its
determinant
of pre-beta
HDL
concentra-
(7).
measurement
hindered
of pre-beta
HDL
is difficult,
which
its physiological function. Pre-beta as apo A-I, and its ?eoiicentratlOn is loW in absolute terms. A suitable assay muM th&efoie kbibit high sensitivity and high specificity in differeptiati#{241}g pfe-eta HDL from other apo A-I-contAining lipopioteins. Currently available assays are unsatisfactory. Radial immunodiffusion does not
separate
work
in elucidating HDL is measured
the
tire-beta and alpha lipoproteins before and appears to overestimate the pre-beta (1, 4). Isolation of pre-beta HDL in the d
concentration >121 kg/L fraction by ultracentrifugation also appears to overestimate the concentration, probably by loss of apo A-I from HDL into the pro-beta fraction during (8, 9). Crossed
ultracentrifugation
sis and affinity
chromatography
resis
and
in
starch
subsequent
immunoelectrophore-
followed
elution
immunoassay
of the are
both
by electropho-
pre-beta
band
laborious
and
with
cum-
(3, 9). Electrophoresis in agarose followed by immunoblotting has been combined with both radioisotope and colorimetric detection (10, 11). The former has the usual drawbacks associated with radioisotope use, whereas the latter is limited by relatively high imprecision (11).
We describe trophoresis,
years
a novel
approach,
immunoblotting,
using and
agarose-gel
chemiluminescent
elecde-
tection of an enzyme label on photographic film; the pre-beta band is quantified by transmission densitometry. This technique is specific and sensitive, and allows simultaneous analysis of large numbers of specimens. Materials
ferred.
Although
R. Trimbie”3
bersome
known as either pre-beta HDL or free The term pro-beta HDL classifies the the HDL group solely by apo A-I content
apo A-I (1-3). particle within rather than by hydrated established
as
Elisabeth
and Fredrickson attention and
measurement
AddItIonal
and
with
received relarole remains unclear. However, evidence now emerging suggests that it plays an important role in lipoprotein metabolism, particularly with regard to reverse cholesterol transport and the metabolism of triglyceride-rich lipoproteins (5, 7). Lecithin:cholesterol acyltransferase (LCAT) may be tions
assay is calibrated
-try. The
by lmmunoblotting
and Methods
Agarose Type I, bovine serum albumin, 5,5’-dithio-bis(2-nitrobenzoic acid) (DTNB) were purchased from Sigma Chemical Co. (Poole, Dorset, UK). Gel Bond was obtained from Pharmacia-LKB Ltd. MateriaLs.
and 1
Department
ofClinical
Biochemistry,
Royal
Grosvenor Road, Belfast BT12 6BA, UK. 2 Division ofBiochemistry, School ofBiology
Victoria and
Hospital,
Biochemistry,
and Department of Clinical Biochemistry, The Queen’s Universit1’ of Belfast, Belfast, UK. Nonstandard abbreviations: apo A-I, apolipoprotein A-I; DTNB, 5,5’-dithio-bis(2-mtrobenzoic acid); HDL, LDL, VLDL, high-, low-, and very-low-denzitylipopmteins, respectively; LCAT, lecithin: cholesterol acyltransferase; PVDF, polyvinylenedifluoride; and TBS, tris-buffered saline. Received April 21, 1992; accepted July 21, 1992.
(Milton
Polyvinylenedifluoride (PVDF) from Millipore (Bedford, MA) and 0.2-sm pore size mtrocellulose membranes from Scheicher and Schuell(Dassel, FRG). Sheep anti-human ape A-I antiserum was supplied by Immuno Ltd. (Vienna, Austria). Rabbit anti-sheep immunoglobulinhorseradish peroxidase conjugate was obtained from
membranes
Keynes,
were
CLINICAL
UK).
obtained
CHEMISTRY,
Vol. 38, No. 1 1 1992 ,
2273
Bio-Rad Laboratories (Richmond, CA). The reagents for chemiluminescent detection, obtained in kit form (ECL Western blotting detection system), and photographic film (Hyperfilm ECL) were from Amersham International (Bucks, UK). Reagents for the enzymatic determination of cholesterol and triglyceride were obtained in kit form from Boehringer Mannheim (Sussex, UK). Specimens. Blood was collected from 30 normal volunteer subjects (17 men, 13 women, ages 22-36 years) after they had fasted for 12 h overnight. It was placed on ice and into either an EDTA-containing collection tube or plain glass tube. All subjects were healthy and had no history of cardiovascular disease. Serum or plasma was separated by centrifugation and maintained at 4 #{176}C until electrophoresis, which was performed within 4 h of venipuncture. Electrophoresis. An agarose gel (225 x 115 x 1.5 mm) with precast wells (4 x 1 mm) was made on a sheet of Gel Bond: 0.5% agarose and 0.25% albumin in Tris barbital buffer (per liter: Tris 5.78 g, barbital 2.47 g, Na barbital 9.76 g; pH 8.6). Plasma samples or standards (7 SAL) were placed in the wells and horizontal electrophoresis was performed by using a Multiphor System (Pharmacia-LKB) at 5 #{176}C and 30 V/cm until the bromphenol blue marker had migrated 7 cm. Immunoblotting. PVDF membrane was wet in 100% methanol and washed in distilled water followed by isotonic saline (0.15 mollL NaCl). Nitrocellulose was washed in saline alone. The gel was overlaid for 30 mm at 4 #{176}C by the blotting membrane, covered by eight sheets ofdry Whatman no. 1 filter paper, two sheets of medium-thick blotting paper, a glass plate, and a 1-kg weight. The blotting membrane was removed from the gel and incubated for 30 mm at room temperature in a 50 g/L low-fat powdered-milk solution in 100 mL of Ths-buffered isotonic saline (TBS; 0.1 molIL Tris, pH 7.4, containing 0.15 mol/L NaC1). After a single 5-mn wash in TBS, the membrane was incubated with gentle shaking for 1 h at 37 #{176}C with sheep anti-human apo A-I (2500-fold dilution in 30 mL of TBS). The membrane was washed three times (5 mm each) in 50 mL ofTBS containing 0.5 mL of Tween 20 per liter and then incubated with gentle shaking for 30 mm at 37 #{176}C with rabbit anti-sheep immunoglobulin-horseradish peroxidase conjugate (3000-fold dilution in 30 mL of TBS). The membrane was washed three times (5 mm each) in TBS containing Tween and stored for s2 h at 4 #{176}C in isotonic saline before being developed. Ckemiluminescent detection. The membrane was developed with the ECL Western blotting detection system according to the manufacturer’s instructions. Briefly, the membrane was immersed in detection reagents (containing luminol, a peroxide, and an enhancer) for 1 min, then wrapped in Cling Film and placed on photographic film for 10-20 s. The photographic film was developed conventionally and was subsequently scanned by transmission densitometry at 600 nm on a Model 620 densitometer (Bio-Rad Laboratories). 2274
CLINICAL
CHEMISTRY,
Vol. 38, No. 11,
1992
Apo A-I standards. Sera from five fasting subjects were pooled and the apo A-I concentration (1.61 g/L) was determined by immunonephelometry. The serum pool was divided into portions and frozen at -70 #{176}C. A series of standards was prepared by diluting the pooled serum in 9 mol/L urea reagentjust before assay to give a range ofapo A-I concentrations from 8.1 to 180 mg/L. Usually three standards per gel (8.1, 35.8, and 108 mgIL) were
run
in duplicate. Stability
assessed
studies.
by measuring
The
stability
concentrations
of pro-beta
HDL
in plasma
was
stored
for various times at 4 and -70 #{176}C. The role of LCAT in regulating pro-beta HDL concentration was assessed by incubating plasma at 37 #{176}C with or without the LCAT inhibitor DTNB (1.4 mmollL). Other assays. Cholesterol and triglyceride were measured enzymatically with a Cobas Bio centrifugal arialyzer (Hoffmann-La Roche, Basel, Switzerland). HDL cholesterol was measured after precipitation ofLDL and VLDL by manganese-heparin (final MnCl2 concentration 92 mmol/L, final heparin concentration 183 UIL) (12). Apo A-I and apo B were measured by rate immunonephelometry with a Beckman Array Protein System (Beckman Instruments, Brea, CA) and Beckman reagents, according to the manufacturer’s instructions (13). The Beckman calibrators for apo A-I and apo B are traceable to the World Health Organization’s international reference materials. Results Immunoblotting and detection. Blotting with PVDF membrane instead of nitrocellulose membrane consistently resulted in complete removal of protein from the gel, as indicated by the absence of stainable protein after blotting (not shown). The PVDF membrane was also preferred because it was more robust (easier to handle) and because it was associated with a lower background signal in the photographic film. Figure 1
shows
the
results
of electrophoresis
and
blotting
of a
typical plasma sample. The alpha and pie-beta bands are clearly demarcated and there is minimal background signal. The pre-beta band therefore can be easily and specifically quantified by transmission densitometry (Figure 2). The optimum exposure time ofthe blot to the photographic film was established in pilot experimerits. Too little exposure results in faint bands and poor analytical precision; too much exposure results in both increased background staining and excessive absorbance of pro-beta bands on densitometry, thereby decreasing the assay specificity and sensitivity. An exposure time of 10-20 s was optimal for the conditions described. To ensure that the anti-apo A-I antiserum was not exhibiting cross-reactivity with other apolipoproteins, we investigated its specificity. Electrophoresis of serum was performed in a sodium dodecyl sulfate polyacrylamide gel. Protein was transferred onto a mtrocellulose membrane by Western blotting and then probed with the antiserum. Significant reactivity was present at only one position, corresponding to a molecular mass of
30’
E E
Alpha
a, 20
a, .0
.‘c a, a,
10
I-
a 0
Pre -beta
0 0
40
50
120
Pro-beta
200
160
HDL (mg/L]
Fig. 3. Relationship between ape A-I concentration densitometric area (absorbance units mm) The best-fit line was obtained by linear regression (r = 0.97)
Origin-
Standardization.
(mg/L)
of the serum
Treatment
pool
and
with
9
molJL urea causes all the apo A-I to migrate in a discrete band to the pre-beta region (Figure 1), as noted preyously (4). The standard can thus be easily quantified by transmission densitometry. We attempted to prepare a urea-free standard by dialysis ofthe urea-treated serum against phosphate-buffered saline. Samples taken be-
Pre-beta
fore and
Origin-,
.
Fig. 1 Film images of ape A-I blots Top: five normal plasma samples with different pre-beta HDL concentrations (1 1.1 , 24.6, 42.6, 33.0, and 15.3 mg/L). Bottom: a series of standards (in duplicate) prepared by diluting serum in 9 mol/L urea (180, 125, 35.8, and 8.1 mq/L)
Abs. 1.2
0.4
40
Migration Fig. 2. Densitometric and pro-beta
-27 28.3
kDa. kDa
Five
profile of a plasma
50
60
70
80
90
(mm) sample,
showing
the alpha
bands
This agrees closely with the molecular mass reported for apo A-I by Baker et al. (14).
serum
samples
(VLDL
cholesterol
both
by crossed immusets of samples
line
from
stretching
cathodally
tory urea-free studies.
standard
Analytical
performance.
the
position, precipitin
pre-beta
arc
(not
tration sample 44.3 conditions
to a concentration
mgfL or described
of 8.1 mgfL
(CV 10.2%). The within-batch precision (CV) was 7% at a mean concentration of 22.1 mgfL and 4.9% at 44.3 mgfL (n = 12 each). Because of the instability of probeta HDL during storage, it was not possible to determine between-assay precision. However, the assay performance was monitored by measuring the size of the pre-beta band in samples prepared daily by diluting portions of the serum pool in 9 mol/L urea at two
different
dilutions.
In
10 sequential
batches
the
be-
tween-assay precision (CV) was 8.5% at a mean measured concentration of 24.4 mg/L and 5.5% at a mean measured concentration of 35.8 mgfL. Specimen stability. Serum and EDTA plasma samples were obtained simultaneously from each offive subjects. The samples were stored at 4 #{176}C and the pre-beta HDL concentration was measured within 4 h of venipuncture.
There
was no significant CLINICAL
difference
CHEMISTRY,
(paired
f-test)
Vol. 38, No. I 1 1992 ,
in the 2275
Table
I Stability
of Pro-Beta HDL In Plasma 4and -70#{176}C
.
Pre-bata Storag.
time, h
4#{176}C
0
36.4 33.0
6 24
35.1
48
40.7 72.9
168
Stored
at
Table 2. Reference
HDL cone, mg/L
-70#{176}C 36.4 52.3 52.5 61.7
54.6
(0
40
.0
0 0
2
4
6
Time
8
10
12
14
[h]
The effect on pre-beta HDL concentration of incubatin9 plasma samples at 37 #{176}C with () or without (#{149}) 1 .4 mmol/L DTNB; 4.
mean±SE
concentration
in serum
SD 7.7, vs 30.2
mg/L,
and
plasma
(26.7
no special are rapidly
48, and 168 concentration
mg/L,
SD 8.1).
%
1 .6 (0.65-2.6)
h (Table 1). There was no increase at 4 #{176}C for 24 h, but concentrations
thereafter. In contrast, there was a significant after 6 h storage at -70 #{176}C. The increase
was
measured
at set times
during
the
itor of LCAT. Reference interval for
in
in
incuba-
for normal
subjects.
The
healthy
reference
normolipidemic subjects is shown in Table 2. Pro-beta HDL had a range of 10.144.3 mg/L (mean 25.7); this accounted for 1.6% (0.652.6%; mean, range) of total serum apo A-I. CLINICAL
30
CHEMISTRY,
Vol. 38, No. 11, 1992
handling obtained
for pre-beta
pita
are
lipoproteins
and
transferred
The
apo A-I is probed
by
a
precautions and the HDL,
separated
to a membrane
horseradish
globulin
tified
tion. For each subject there was an initial decrease in concentration for 2 h followed by a gradual increase to a value greater than the starting concentration at 14 h (Figure 4). The initial decrease was prevented by incubation in the presence ofDTNB (1.4 mmol/L), an inhib-
2276
1 .64 (1.2-2.25) 25.7(10.1-44.3)
with
the
and
by capillary
a specific
is detected
alpha
and
by electrophoresis antibody
peroxidase-labeled
conjugate
(18). Fitechniques
blotting. followed
anti-immunoby
the
peroxidase-
catalyzed oxidation of luminol in the presence of an enhancer (19). This reaction results in chemiluminescence, which is captured on photographic film and quan-
pre-beta HDL concentration in frozen specimens was related to the freezing-thawing process. Repetitive freeze-thaw cycles in a plasma sample showed a 57% increase after one cycle, a 92% increase in concentration after three cycles, and a 112% increase after six cycles. Plasma samples from four subjects were incubated at 37 #{176}C for 14 h. The pre-beta HDL concentration of the
interval
1.57(1.0-2.6)
are relatively simple. In this novel assay
by transmission
A pool of freshly collected plasma was prepared, Diluting serum divided into portions, and stored at both 4 and -70 #{176}C. a lipid-apolipoprotein Pro-beta HDL concentration was measured at 0, 6, 24, tical to pre-beta
samples
1.13 (0.39-1.7)
0.61 (0.48-0.82)
and requires nally, results
20 0
increased increase
(rang.)
4.87(3.6-6.0)
Over the past decade there has been considerable interest in the use of chemiluminescent labels and chemiluminescent detection in immunoassay (1 7, 18). Chemiluminescence is the release of energy as light during a chemical reaction, and its use offers several distinct advantages over radioisotopic labels or enzyme labels coupled to colorimetric reactions. It exhibits very high sensitivity and, in contrast to the situation for radioisotopic labels, chemiluminescent detection is safe
60
mean
Subjects
DIscussion
=
Fig.
Cholesterol, mmol/L Triglycerides, mmol/L HDL cholesterol, mmol/L Apo A-I, g/L Pre-beta HDL, mg/L Total apo A-I in pre-beta fraction, APOB,g/L
a, E
6)
for 30 Healthy
Mun
100
-J
Values
densitometry. in 9 molJL urea converts the apo A-I to complex that appears to be idenHDL (9). However, different antisera
may show variable apo A-I (20). The reactive in serum fore, tion
reactivity
with
native
and denatured
antiserum used in this study was less with apo A-I in native HDL than with apo A-I treated with 9 molIL urea. It was not, there-
possible simply
pro-beta HDL concentrathe fraction of total area under the pro-beta HDL peak multiplied by the total serum apo A-I concentration. Instead, the assay was calibrated with standards from a serum pool diluted in 9
molfL
to quantify by computing
the
urea.
This special
assay is not requirements
sitometer. It achieves tivity, and linearity.
technically difficult, and the only are darkroom facilities and a den-
high analytical The assay offers
specificity, considerable
sensiad-
vantages over existing assays for pre-beta HDL. The run-time is only 5 h, and it is less laborious than either crossed immunoelectrophoresis or electrophoresis combined with elution and immunoassay (3, 4). Unlike crossed immunoelectrophoresis, where the number of samples run per electrophoresis tank is strictly limited, this technique allows 28 samples to be analyzed per gel. The assay sensitivity and precision are better than
can
be achieved
by
immunoblotting
and
colorimetric
detection (11). The high sensitivity means that much less is antiserum is required (crossed immunoelectrophoresis uses up to 300-fold more antiserum per sample analyzed). We found that a mean of 1.6% (range 0.65-2.6%) of
total
serum
apo A-I was present
in the pre-beta (n = 30). This amount
fraction is lower
in normolipemic subjects than reported for other assays: 15% (10-30%) by radial immunodiffusion (1), 9.8% (4.2-15.4%) by ultracentrifugation (8), and 14.4% (5-25%) by affinity chromatography and electrophoresis followed by elution and immunoassay (3). However, all these other techniques may overestimate pre-beta HDL concentration (4). With radial immunodiffusion, two precipitin rings are visible in the gel plate: an outer ring produced by apo A-I in HDL and an inner ring produced by apo A-I in pro-beta HDL. Because these two species are not separated before the immunochemical reaction occurs, quantification of probeta HDL by using the diameter of the inner ring may overestimate the concentration (4). In the assays using ultracentrifugation and affinity chromatography, apo A-I may be lost from HDL into the pro-beta fraction during analysis (4, 9). The value obtained by crossed immunoelectrophoresis is 5.8% (3.5-8%), again higher than that obtained in this study. This difference is perhaps explained in part by the use of different antisera and apo A-I standards. Our stability data confirm the findings of others that pro-beta HDL is formed in vitro at 4, 37, and -70 #{176}C and on freeze-thaw cycling (4, 14). With incubation at 37 #{176}C, there is an initial decrease in pre-beta HDL concentra-
tion
that
can be prevented
by adding
the LCAT-inhibi-
tor DTNB. This was previously shown for mouse (11) and human plasma (7), suggesting that the reduction is LCAT-mediated (7). We recommend that samples be stored at 4 #{176}C and analyzed as soon as possible (within 24 h). There is now sufficient evidence to suggest that prebeta HDL is an important intermediate in lipoprotein metabolism. These particles appear to be produced from both HDL and triglyceride-rich lipoproteins and may be an important substrate for LCAT (7). Other work showed that pro-beta HDL is an initial acceptor of labeled membrane cholesterol from cultured fibroblasts, indicating a direct role in reverse cholesterol transport (5). Much remains to be elucidated before a comprehensive understanding of the role of pro-beta HDL can be achieved. In conclusion, chemiluminescent also have more
ment
of proteins
this assay detection widespread
on blotted
of pre-beta HDL, based on on photographic film, may application for the measure-
membranes.
We gratefully George Allen.
acknowledge
the export
technical
of
assistance
References
1. Borreson implications
A, Berg
apoproteins.
Artery
K. Presence of “free” for immunological quantification
apo Al in serum: of HDL and its
1980;7:139-.60.
2. Gebhardt DOE, Schicht B!, Paul LC. The immunochemical determination of apolipoprotein A, total apolipoprotein Al and ‘free’ apolipoprotein A! in serum of patients on chronic haemodialysis. Ann Clin Biochem 1984;21:301-5. 3. Kunitake ST, La Sala KJ, Kane JP. Apolipoprotein Al-containinglipoproteins with pro-beta electrophoretic mobility. J Lipid Res 1985;26:549-55. 4. Neary RH,
Gowland E. Stability of free apolipoprotein A-i concentration in serum and its measurement in normal and hyperlipidemic subjects. Clin Chem 1987;33:1163-9. 5. Castro GR, Fielding CJ. Early incorporation of cell derived cholesterol into pro-beta migrating high density lipoprotein. Biochemistry 1988;27:25-9. 6. Levy RI, Fredrickson DS. Heterogeneity ofplasma high density lipoproteins. J Clin Invest 7. Neary R, Bhatnagar D, Durrington P, Isola M, Arrol 5, Mackness M. An investigation of the role of lecithin:cholesterol acyltransferase and triglyceride-rich lipoproteins in the metabolism of pro-beta high density lipoproteins. Atherosclerosis 1991;89:35-48. 8. Duval F, Frommher ZK, Atger V, Druecke T, Lacour B. Influence of end stage renal failure on concentrations of free apolipoprotein A-I in serum. Clin Chem 1989;35:963-6. 9. Neary RH. Implications of measurement of pro-beta high density lipoprotein (free apo Al) by ultracentrifugation. Clin Chem 1965;44:426-41.
1989;35:2336. 10. Ishida BY, Frolich J, Fielding CJ. Prebeta-migrating high density lipoprotein: quantit.ation in normal and hyperlipidemic plasma by solid phase immunoassay following electrophoretic transfer. J Lipid Res 1987;28:778-86. 11. Ishida BY, Albee D, Paigen B. Interconversion of pro-beta migrating lipoproteins containing apolipoprotein A! and HDL. J Lipid Res 1990;31:227-36. 12. Albers JJ, Warnick GR, Wiebe D, et al. Multilaboratory comparison of three heparin-Mn2 precipitation procedures for estimating cholesterol in high density lipoprotein. Clin Chem 1978;24:853-6.
13. Maciejko JJ, Levinson SS, Markyvech L, Smith MP, Blevins RD. New assay of apolipoproteins Al and B by rate nephelometry evaluation.
Clin
14. Baker
NH, of human high Natl Acad &i 15. Miller JP, ofapolipoprotein
Chem 1987;33:2065-9. T, Jackson RL. The primary structure density apolipoprotein glutamine I (apo Al). Proc USA 1974;71:3631-2. Mao SJT, Patsch JR, Gotto AM. The measurement A-i in serum by electroimmunoassay. J Lipid Res Delahunty
1980;21:775-80.
16. Johnson
HA, Deegan T. Loss of A-apoprotein high-density lipoprotein separation. Ann Clin Biochem 1981;18:308-13. 17. Kricka LaJ, Thorpe GHG. Photographic detection of chemiluminescent and bioluminescent reactions. Methods Enzymol 1986; immunoreactivity
P, Muirhead
during
133:404-20. 18. Kricka
LI. Chemiluminescent and bioluminescent techniques [Review]. Clin Chem 1991;37:1472-81. 19. Thorpe GWG, Kricka U. Enhanced chemiluminescent reactions catalyzed by horseradish peroxidase. Methods Enzymol 1986; 133:331-54.
20. Steinberg KK, Cooper GE, Graiser SR Rosseneu M. Some considerations of methodology and standardization of apolipoprotein Al immunoassays. Clin Chem 1983;29:415-26.
CLINICAL
CHEMISTRY,
Vol. 38, No. 11, 1992
2277