Quantification of apolipoproteins in rat serum and in cultured rat hepatocytes by enzyme-linked immunosorbent assay

ANALY.fICAL

BIOCHEMISTRY

154, 3 16-336

( 1986)

Quantification of Apolipoproteins in Rat Serum and in Cultured Hepatocytes by Enzyme-Linked lmmunosorbent Assay’

Rat

RENEE C. LIN

Received

October

15, 1985

An enzyme-linked immunosorbent assay (ELIS.4) has been developed to measure apolipoproteins in rat serum. Nondelipidated whole serum was heat-treated at 52°C for 3 h in phosphatebuffered saline containing 0.1% Tween-20 before assay. Monospecilic rabbit anti-rat apolipoprotein antibodies were added to 96-well polystyrene microtiter plates which had been coated with purified rat serum apolipoproteins or unknown samples. After incubation and washing. goat anti-rabbit serum antibodies conjugated with horseradish peroxidase were added to the plates and incubated. The bound peroxidase activity was assayed after further washing. Serum apolipoprotein concentrations were calculated by comparison against purihed standards that were assayed simultaneously with the unknown samples. The intraassay coefficients of variation for apolipoprotein Al. E. and AIV (Apo AI. E. and AIV) were 1.3. 4.4. and 5.3%. and interassay coethcients of variation were 6. I. 5.5. and 7.9%. respectively. The ELISA assay is sensitive to nanogram quantities ofrat serum apolipoproteins and the results agree well with those measured by densitometu. The serum concentrations of Apo Al, E. and AIV of a normal fed rat were found to be 504 t 8, 4 I3 t 20. and 262 + 20 pg/ml. respectively. When cultured as monolayers in Waymouth’s medium for I day, rat hepatocytes secreted Apo AI. E. and AIV at rates of 2.5 I. 6 I .8. and 48.9 ng protein/mg cell protein/h. r!s 1986 Academic Press. Inc. KEY WORDS: gel electrophoresis: proteins: immunochemical methods: lipoproteins: lipids.

Liver and intestine are the two major organs that synthesize and secrete plasma lipoproteins (l-4). Apolipoprotein AI, apolipoprotein E. and apolipoprotein AIV with molecular weights of 27.000, 35,000, and 46.000, respectively (5), constitute most of the protein components of rat high-density lipoproteins. Apolipoprotein E is also a major component of rat very low-density lipoproteins (5). Many investigators have reported that cultured rat hepatocytes can synthesize and secrete various apolipoproteins (6-9). We have recently reported (10) that, when cultured as monolayers, rat hepatocytes secreted newly synthesized cholesterol along with some protein factors. presumably apolipoproteins. into culture me-

dium. These cholesterol-binding factors were found to be in the HDL’ fraction when the culture medium was centrifuged according to the method of Have1 et al. (1 1). This demonstrates that cultured liver cells can be used as an experimental model to study the relationship between hepatic synthesis of cholesterol and lipoproteins. However, a sensitive method that can measure nanogram quantities of apo!ipoproteins is necessary in order to study the secretion and synthesis of lipoproteins in cultured hepatocytes. ’ Abbreviations used: HDL. high-density lipoproteins: Apo, apolipoprotein: ELISA. enzyme-linked immunosorbent assay: SDS-PAGE, sodium dodecyl sulfate-polyacrylamide gel electrophoresis: IgG. immunoglobulin G: PBS. phosphate-buffered saline: PBST, phosphate-buffered saline, pH 7.0. 0.1% Tween-90; BSA. bovine serum albumin: ELITB. enzyme-linked immunoelectrotransfer blot.

’ This research is supported by Research Fund from Veterans Administration and Grant-in-Aid from American Heart Association. Indiana Affiliate, Inc. 0003-2697/X6

$3.00

Copyright G 1986 by Academic Press. Inc. All rights of reproduction m any form reserved.

316

IMMUNOASSAY

FOR

Individual enzyme-linked immunosorbent assays (ELI%%) for Apo AI, Apo E, and Apo AIV have been developed in this study to determine apolipoprotein concentrations in rat serum and to measure their secretion rates by cultured rat hepatocytes. MATERIALS

AND

METHODS

Isolation (f rat serum apolipoproteins. Rat serum was prepared from pooled fresh blood obtained by cardiac puncture. The density of rat serum was adjusted to 1.35 g/ml by adding solid potassium bromide, then overlayered with l/10 volume of 0.15 M NaCl containing KBr to a density of 1.2 1 g/ml. After 34 h centrifugation at 105.000~ with a Beckman 40 rotor, the top layer (d < 1.21) was recovered and used as the source for apolipoproteins. and the bottom layer (d > 1.2 1) was used to prepare lipoprotein-deficient serum. Apolipoproteins were separated by sodium dodecyl sulfatepolyacrylamide gel electrophoresis (SDSPAGE) without delipidation. The lipoprotein fraction containing 3 mg protein was applied to a 140 X 1 l.O-mm gel slab (7.5% polyacrylamide) and electrophoresis was performed at 40 mA/gel for 2 h. A small strip of the gel was cut and stained with Coomassie Blue for protein. Apolipoprotein AI. E, and AIV were identified by their reported molecular weights (5). The unstained portion of the gel slab was then aligned with the stained strip. Bands of apolipoproteins were cut using the stained gel as a guide. Sl.rips of gels were placed in individual flasks and extracted with a buffer containing 0.05 !vl Tris. 0.15 M glycine. and 0.1%) SDS at pH 8.13.The extracts were concentrated with an Amicon Ultrafiltration Stirred Cell. The purity of each extract was checked with SDS-PAGE. Preparations showing a homogeneous protein band corresponding to each apolipoproteln were pooled, then aliquoted and stored at -70°C. Apolipoproteins AI, E, and AIV purified in this manner were used for antibody preparations and as standards in ELISA. With this procedure. approximately 12. 9. and 5 mg of purified Apo AI, E, and

RAT

APOLJPOPROTEINS

317

AIV, respectively, were obtained from 100 ml of rat serum. Preparation yf‘antibodies to apolipoproteim. Each rabbit (New Zealand White, female) was immunized with 200 gg of a purified apolipoprotein (AI, E. and AIV) emulsified with complete Freund‘s adjuvant. followed by two booster shots with incomplete Freund’s adjuvant at 3-week intervals. Serum was prepared from blood collected by cardiac puncture. Approximately 30-40 ml serum was obtained from each rabbit. The IgG fraction was isolated by chromatography on protein ASepharose CL-4B. Tests with Ouchterlony double immunodiffusion (Fig. 1) showed these IgG fractions to form single bands between purified apolipoproteins and the proteins in serum. Monospecificities of these antibodies were further demonstrated by enzyme-linked immunoelectrotransfer blotting ( 12). Apolipoproteins from the rat serum lipoprotein fraction were lirst resolved on SDS-PAGE. then transferred electrophoretically onto a

FK. 1. Ouchtcrlony double immunodiffusion of rabbit antibodies to rat serum apolipoproteins against their antigens. Wells contained 90 pg IgG. 1.2 pg puriiied apolipoproteins. or 7 ~1 rat serum (well I of panel B contains 3.5 ~1 rat serum). (A) Anti-APO AI IgG: (B) Anti-Apo E IgG: (C) Anti-APO AIV IgG: ( 1) rat serum. (2) Apo Al, (3) 4~0 E. (4) Apo AIV.

318

RENEE

FIG. 2. Enzyme-linked immunoelectrotransfer blot (ELITB) of rat serum apolipoproteins. Apolipoproteins in rat serum lipoprotein fraction (d < 1.2 1) were separated by SDS-polyacrylamide gel electrophoresis and then electrotransferred to a piece of nitrocellulose paper as described under Materials and Methods. Strips of the paper were exposed to rabbit anti-rat apolipoprotein antibodies followed by peroxidase-conjugated goat anti-rabbit serum IgG. The bound enzyme-linked antigen-antibody complex was visualized by incubating with a peroxidase substrate containing 4-chloro- I naphthol. Lane I. molecular weight standard proteins (Bio-Rad) stained with Amido Black. Lane 7. rat serum lipoproteins stained with Amido Black. Lane 3. rat strum lipoproteins ELITB with rabbit antiApo AI I@. Lane 4. rat serum lipoproteins ELITB with rabbit anti-APO E IgG. Lane 5. rat serum lipoproteins ELITB with rabbit anti-APO AIV IgG.

piece of nitrocellulose paper. Nonspecific binding sites of the paper were blocked by incubating with 0.01 M phosphate-buffered saline (PBS, pH 7.0) containing 1% bovine serum albumin (BSA). The paper was then rinsed with PBS and cut into strips. Each strip was incubated with PBS containing individual rabbit anti-rat apolipoprotein IgG. After a careful washing, the paper strips were incubated with goat anti-rabbit serum IgG conjugated with horseradish peroxidase. Finally bands were visualized by incubating with a peroxidase substrate system containing 4chloro- 1-naphthol and hydrogen peroxide (Kirkegaard and Perry Laboratories, Gaithersburg, Md.). Figure 2 shows that our IgG

C.

LIN

preparations do not cross-react with the other antigens. .-lnimu/. Adult male Sprague-Dawley rats (Harlan Sprague-Dawley, Inc.. Indianapolis, Ind.), 250-300 g. were used in all experiments and as blood donors. Rats were housed in an air-conditioned, windowless room equipped with an automatic timing device that maintained lights on from 6 AM to 6 PM, and had free access to laboratory rat chow and tap water. T/w ELIS.4 procedure ,/iv Ape A I uncl :lpo E. Sera prepared from blood samples were stored at 0-4°C in the presence of 0.1 %I sodium azide and assayed within 3 weeks. Neither serum samples nor purified apolipoproteins used as standards were delipidated. Purified apo AI and apo E were diluted with 0.0 I M phosphate-buffered saline, pH 7.0, containing 0.1% Tween-20 (PBST) and 0.1% bovine serum albumin (Sample Buffer) to 1000 and 500 rig/ml, respectively. Unknown sera were diluted 500-fold with Sample Buffer. Diluted standards and sera were heated at 52°C for 3 h in a water bath to expose their immunoreactive sites ( 13.14). The heat-treated standards were then further diluted to a l-500 ng/ ml range. Eight serial dilutions ranging from lo- to 2000-fold were performed on heated unknown sera with Sample Buffer. Antigens were adsorbed to wells of polystyrene microtiter plates (Costar No. 3590. Cambridge, Mass.) by pipetting 200 ~1 of diluted standards or unknown samples into each well. The microtiter plates were covered and incubated at room temperature overnight. Overnight incubation appears to be sufficient for maximal adsorption because prolonging incubation to 48 h did not enhance color formation for even the most dilute standards (0.1 rig/well) or unknown serum samples ( 1 X 1O6dilution). The antigen solutions were aspirated the following day and wells were washed five times with 0.0 1 M phosphate-buffered saline, pH 7.0, with 0.1% Tween-20. Rabbit anti-rat apolipoprotein IgG was diluted to a predetermined optimal concentration (Fig. 3) with PBST. A 200~1 aliquot of diluted rabbit IgG was added to

IMMUNOASSAY

0.8

0.7

FOR

RAT

319

APOLIPOPROTEINS

///+xks~ ...*_ .,.A r ::-....A’ : : :- .A’ ::r.; A.’ : : .’ *.“... ...A-:e.. *o..” .::.m.. \:.:*.**. :-\\&,e j 1.2

1.0

0.8

2 -

0 *

0.6 E 0.4

0.2

0

0.1

0.5

1

fig

2

5

10

20

IgG well

FIG. 3. Titration curves of the amounts of antibodies bound to antigens measured by ELISA. Total binding (closed symbol) was determined by adding various dilution of rabbit antibodies to wells of microtiter plate containing a near saturated amount of respective purihed rat serum apolipoproteins. then followed by adding peroxidase-conjugated goat anti-rabbit serum IgG and measured for the bound enzyme activity. Nonspecific binding (open symbol) was measured by iirst coating wells with apolipoproteinfree Sample Butler. Half symbol represents the net difference between total and nonspecific bindings. (m. q . 0) Anti-APO E IgG: (0, 0. ~3) anti-APO AI IgG: (A. A. A) anti-APO AIV IgG.

each well and incubated at room temperature for 2 h. After the plates had been washed five times with PBST. each well then received 200 ~1 of goat anti-rabbit serum IgG conjugated with horseradish peroxidase (1: 1000 dilution, Sigma Chemical Co.. St. Louis, MO.). The plates were washed in the same manner after 2 h incubation at room temperature. Peroxidase substrate containing 125 pg/ml o-dianisidine. 3HCI and 0.006% Hz02 in 0.1 M phosphate-citrate buffer. pH 5.0, was prepared freshly according to Koritnik and Rude1 (14). Then 200 ~Lllof the substrate solution was added to eacih well. The color was developed in the dark for 30 min and the enzyme reaction was terminated by adding 50 ~1 of 2 N HCI into each well. Optical density ofthe color was read at 4 10 nm with a Dynatech Microplate Reader. Model MR600 (Dynatech Instruments. Inc., Torrance. Calif.). Concentrations of the apolipomproteins in the unknown samples

were calculated from their respective standard curves and corrected for dilution factors. For each sample, concentrations calculated from values within the working range are linearly proportional to their dilutions. Results reported in this study are the means of three to six values from eight dilutions of each sample that fell within the working range of each apolipoproteins and expressed as micrograms of apolipoprotein per milliliter of serum. The sample was reassayed at different serial dilutions if less than three values were within the working range. The ELIS.4 procechrejbr .+w NJ’. Purified rat serum Apo AIV was diluted to 3000 ng/ ml with PBST containing 0.3% BSA, incubated at 52°C for 3 h. then further diluted with the same buffer containing 0.3%~BSA to obtain a working range (50 to 1000 rig/well) for the standard curve. Because rat serum was reported to have an albumin concentration of 30 mg/ml (15) sera were diluted IO-fold with PBST containing no BSA to make up an albumin concentration of 0.3%. After heat treatment at 52°C for 3 h serum samples were further diluted 5- to lOO-fold (8 dilutions) with PBST containing 0.3% BSA. ELlSA determination of Apo AIV shows better proportionality as well as reproducibility when the sample buffer contains 0.3%’ rather than a lower (0.1%) concentration of BSA. Other procedures for Apo AIV were identical to those for Apo Al and Apo E. except that the concentration of o-dianisidine * 2HCl in the peroxidase substrate had been increased to 200 pg/ml for Apo AIV. Quunt$cation

of apolipoproteins

/I!, dmsi-

tometrj’. Protein concentration of purified apolipoproteins was determined by the method of Lowry et al. ( 16) using bovine serum albumin as the standard. Purified apolipoproteins and serum lipoprotein fractions (cr’< 1.2 1) were loaded onto a SDS-polyacrylamide gel (7.5%) slab. After electrophoresis at 40 mA/gel for 2 h, the gel was stained with Coomassie Blue R-250 for protein bands. It was then destained and dried between two sheets of dialysis membrane backing (Bio-Rad

320

RENEE

Chemicals, Richmond, Calif.). Intensity of protein bands was monitored at 560 nm using a Gilford Multi Media densitometer attached to a Hewlett-Packard plotter-integrator. A standard curve was obtained by measuring peak areas under the densitometric response curve of known quantities of each purified apolipoprotein. A range from 1 to 4 pg for each lane used in this experiment yielded linear lines for all three apolipoproteins. Duplicate of two different dilutions of three rat serum lipoprotein preparations (d < 1.2 1) were analyzed with the standard curve obtained from the purified apolipoprotein on the same gel slab. This was done to ensure that concentrations of proteins used were within the linear range of densitometry for each apolipoprotein. Isolation and culture of hepatoc~~tes. Hepatocytes were isolated by perfusing a rat liver with collagenase and were cultured as monolayers in Waymouth’s medium (GIBCO, MB 752/l, Grand Island, N.Y.) in collagen-coated plastic Petri dishes (20 X 100 mm) according to the method described before (17). Fetal calf serum (10%) was added to the culture medium for the first 4 h in order to facilitate cell attachment to the dish. After 4 h incubation cell monolayers were washed and changed to serum-free Waymouth’s medium, returned to a humidified incubator and continued incubation at 37°C for 16 h. Preparation oj‘culture medium .fbr ELI&l. Culture medium of rat hepatocytes was collected for determination of apolipoprotein concentrations with ELISA method after 16 h incubation. The medium was centrifuged at lO,OOOg for 15 min to remove cell debris. Concentrated sodium azide was added to the medium to a final concentration of 0.1%. The medium was then stored at 0°C and assayed within 1 week. On the day of assay, 0.9 ml of the medium was mixed with 0.1 ml of a buffer containing 0.1 M sodium phosphate, pH 7.0, I .5 M NaCl, 1% Tween-20, and either 1% BSA (for Apo AI and Apo E) or 3% BSA (for Apo AIV). The mixtures were heated at 52°C for 3 h in a water bath. The heated medium was then made into eight dilutions, ranging from

(‘.

LIN

I - to 200-fold for Apo Al, lo- to 2000-fold for Apo E, and l- to ‘O-fold for AIV. using the same buffers as for serum samples. Diluted medium (200 ~1) was added to wells of microtiter plates. Similar to serum samples, we found that overnight incubation of medium samples at room temperature was sufficient for maximal antigen adsorption. Other ELISA procedures were identical to those for the rat serum. After the medium had been removed, cell monolayers were carefully washed and rinsed with saline. then collected by scrapping with a rubber policeman and determined for protein by the method of Lowry et al. ( 16). Secretion rate of hepatic apolipoprotein is expressed as nanograms of apolipoprotein per milligram of cellular protein per hour. RESULTS

Potencies qf frirzti-apolipoprotein

Antibodies

To test for the potency of anti-apolipoprotein IgG, a near saturated amount of purified rat serum apolipoprotein (200 ~1 each of Apo AI, Apo E, and Apo AIV at 50, 50, and 1000 rig/ml, respectively, as determined in Figs. 4 and 5) was added to wells of a microtiter plate and incubated overnight. Various amount of antibodies were then added to wells after washing five times with PBST. This step was followed by the addition and incubation of peroxidase-conjugated goat anti-rabbit serum IgG. Figure 3 shows that peroxidase activity increased with an increased amount of antibody. Concomitantly, nonspecific binding occurred at high concentrations of antibodies in the absence of antigens. For best ELISA results, the amount of antibody used must yield the highest enzyme activity with minimal background color caused by nonspecific binding. By this criterion, the optimal amounts of anti-APO AI and anti-APO E antibodies for ELISA were 2 to 10 pg/well, and was 5 to 10 pg/well for anti-APO AIV. Subsequently 2 pg/ well of rabbit IgG’s was chosen in all assays. At this concentration, the assay for Apo AIV

IMMUNOASSAY

FOR

RAT

APOLIPOPROTEINS

321

trations of the pooled sera, stored in the presence of 0.1% sodium azide at 0°C. and measured by ELISA method did not systematically decrease in 7 weeks. Nonetheless, results are best compared when samples are assayed simultaneously.

FIG. 4. Dose response curves for purified rat serum Apo Al and Apo E by ELISA. Various dilutions of purified rat serum apolipoproteins were coated to wells of microtiter plates. IgG (2 pgjwell) against either Apo Al or Apo E was added to the microtiter plates and assayed for the specific binding and the cross-reaction. Solid lines indicate antiApo Al IgG. and dashed lines anti-APO E IgG added to wells containing ((0) Apo AL (B) Apo E. and (A) Apo AIV.

is 90% maximal with a considerable saving of anti-APO AIV IgG.

Figure 4 shows dose response curves for purified rat serum Apo AI and Apo E using optimal amounts of rabbit IgG’s as determined in Fig. 3 for ELISA. The working ranges of standard curves were between 0.5 and 10 ng/ well for both apolipoproteins. No significant cross-reaction of antibodies with other antigens was observed within these ranges. The working range for Apo AIV was from 50 to 600 rig/well (Fig. 5). Neither Apo AI nor Apo E interfered with Apo AIV ELISA up to 2000 ng protein per well. The intraassay coefficients of variation of Apo AI, Apo E, and Apo AIV were 2.3. 4.4, and 5.3% (n = lo), respectively. The interassay coefficients of variation were 6.1, 5.5, and 7.9% for AI, E, and AIV. respectively, over .a period of 7 weeks using eight pools of serum. The apolipoprotein concen-

Because apolipoproteins in serum or culture medium are in complex forms binding to lipids and other proteins. it is necessary to verify whether isolated apolipoproteins can be used as ELISA standards for these solutions. Figure 6 showed that dilution curves of purified antigens, sera and culture media were parallel throughout the working ranges of these apolipoproteins. This indicates that serum lipids and proteins do not interfere with apolipoproteins for ELISA under the conditions described in this report. Furthermore. we mixed purified apolipoproteins with whole serum. lipoprotein-deficient serum or the culture medium and assayed for apolipoprotein concentrations by ELISA. In all cases, the concentrations of the mixtures were nearly equal to the sum of

FIG. 5. Dose response curves for purified rat serum apolipoproteins against anti-APO AIV IgG by ELBA. Anti Apo AIV IgG (2 pg/well) was added to wells of the mlcrotiter plate containing (A) Apo AIV. (0) Apo AI. and (w) Apo E. and assayed for binding according to procedures described under Materials and Methods.

322

RENEE

FIG. 6. Dilution curves of antigens from different and Methods. (A) anti-APO AI IgG: (BJ anti-APO whole rat serum: (W) hepatocyte culture medium.

C.

LIN

sources. ELISA procedures E IgG; (C) anti-APO AIV

concentrations of two components when assayed separately (Table 1). These mixtures contained serum proteins which far exceeded (90- to 3300-fold) purified antigen proteins. These results thus demonstrate that isolated apolipoproteins behave similarly to those in rat serum and in the culture medium in this ELBA procedure.

Quantijication of Apolipoproteins h.11 Densitomettyl vs ELIS.4 Apolipoprotein AI, E, and AIV of three preparations of serum lipoprotein fraction (d < 1.2 1) were determined by densitometry and by ELISA. Table 2 shows that results obtained by both methods were in agreement, differing by a range of 2 to 24%.

were described under Materials IgG. (0) Purified antigen; (A)

Concentrations c?ffpolipoproteins in Ral Serwn and the Culture ,~~ediurw sf’Rat NepatoqTes When measured with the ELBA method, serum concentrations of normal fed SpragueDawley rats were 504 k 8,4 13 + 20, and 262 k 20 (n = 8) pg/ml for Apo AI, Apo E, and Apo AIV, respectively (mean t SEM). A comparison of results obtained by ELISA. electroimmunoassay, and radioimmunoassay from several laboratories is presented in Table 3. Cultured hepatocytes secreted very little Apo AI compared to Apo E and Apo AIV. When hepatocytes were cultured in serumfree, hormone-free Waymouth’s medium for 1 day, secretion rates for Apo AI, E, and AIV

IMMUNOASSAY

FOR

RAT

TABLE INTERFERENCE

OF SERUM

FACTORS

I

ON ELBA

MEASIJREMENT

Apolipoprotein Purified apolipoprotein

II

Apo AI Apo E Apo AIV

5 5 3

’ Mean

323

APOLIPOPROTEINS

Whole serum

OF APOLIPOPROTEINS

by ELISA

when

mixed

Lipoprotein-deficient serum’

96 * 5% a 89 k 6% I10 lr 4%

withb Hepatocyte culture medium 94 k 13% 94 + 3% 107 k 4%

102 t 8% 95 + 6% I27 t 5@6

t SEM.

h Concentration Concentration

(mixture of purified Apo and serum factor) (purified Apo) + Concentration (serum factor)

x ,ool ‘:

Protein ratio of the mixture: whole serum:purified Apo AI = 90: I; whole serum:purified Apo E = 90: I: whole serum: purilied Apo AIV = 300: I. Lipoprotein-deficient serum:purified Apo 41 = 900: I: lipoprotein-deficient serum:purified Apo E = 900: I: lipoprotein-deficient serum:putified Apo AIV = 300: I. Culture medium:purified Apo AI = 3300: 1: culture medium:purified Apo E = 400: I: culture medium:purilied Apo AIV = 100: I. ’ Lipoprotein-deficient serum was prepared by ultracentrifugation as described under Materials and Methods. After the lipoprotein fraction was removed from rat serum, the bottom layer fraction (d > 1.2 1) was dialyzed against PBS, pH 7.0. exhaustively before use.

were 2.5 1, 68.8, and 48.9 ng protein/mg protein/h, respectively.

cell

DISCUSSION

The measurement of plasma apolipoproteins is of major interest. They exist in complex mixtures with lipids. are relatively insoluble, and cannot be assayed on the basis of their functions. Various types of immunoassays are currently used in different laboratories, in-

cluding radial immunodiffusion (18) electroimmunoassay ( 19.20), and radioimmunoassay (13.2 1). However, normal values for each of the apolipoproteins vary from laboratory to laboratory. The difficulty arises in part from some intrinsic properties of apolipoproteins. e.g., instability, self-association, multiple antigenic sites, and possible masked antigenic sites. Considerations on the quantification of apolipoproteins have been discussed in detail (22.23).

TABLE COMPARISON

Apo Sample

ELISA

OF APOLIPOPROTEIN

2

MEASUREMENT

AI

Apo

Densitometry

ELISA

BY ELISA

AND DENSITOMETRY

E Densitometry

Apo AIV ELISA

Densitometry

(mgiml? I 2 3

3.22 i 0.07” 2.93 i 0.37 4.31 f 0.32

3.38 k 0.13 3.62 f 0.07 4.10 f 0.10

3.30 + 0.06 2.64 * 0. IO 2.76 + 0.16

2.89 + 0.08 2.60 -+ 0.05 2 .69 f 0.09

2.22 -+ 0.07 2.62 I 0.07 3.59 +- 0.22

Norm. Samples were three different preparations of rat serum lipoproteins. Quadruplicates of each assayed with ELISA. Results for densitometry were the means of duplicates of two different dilutions. a Mean f SEM. n = 4.

2. I3 k 0.02 2.78 + 0.17 3.28 -c 0.05 sample

were

324

RENEE TABLE

C.

LIN 3

CONCEKTRATIONSOFAPOLIPOPROTEINSINRATSERUM Concentration. Quantitation ELISA EIA EIA RIA RIA

method

Reference This

study (19) (24) (25)

(26)

Rat strain

II

Sprague-Dawley Hood Long-Evans Sprague-Dawley Wistar

8 9 37 I 5

No/e. All rats were male and were fed regular ’ Mean + SEM.

laboratory

rat food.

Apo AI

Body

504k

Apo 8”

909 i 18 480 -t 23 614 + 83 weights

Kg/ml

413 247 245 240 416

were between

I! t -t * i-

serum

E

Apo AI\’

20 11 9 30 30

262 + 70

200 and 300 g.

More recently, Koritnik and Rude1 ( 14) tionally, antigens or antibodies are first bound employed a noncompetitive. sandwiched en- to the bare plastic surface for ELISA. The exzyme-linked immunoassay to measure apolicess binding sites are then blocked by further poprotein AI concentration in nonhuman incubation with a buffer containing a high primate serum. In their procedure, the antigen concentration (0.5%) of BSA. We choose to (Apo AI) of the serum samples was attached include BSA with the antigen in the initial into purified. monospecific anti-monkey Apo AI cubation for two reasons. (1) We found that antibodies adsorbed on a polystyrene microthe quantity of BSA (0.1%) present in the titer plates, It was sandwiched then with more sample buffer for Apo AI and E did not interof the same antibodies conjugated with horse- fere with the adsorption of apolipoproteins to radish peroxidase. Their method was reported plastic plates. Yet, at this concentration. BSA to be as sensitive as radioimmunoassay (13). sufficiently blocks the excess binding sites of Results of this ELISA method were reported the plate as evident by very little nonspecific to be highly reproducible and compared fa- binding of antibody which was added subsevorably with those obtained by radial immuquently (Fig. 3): this saves time since one innodiffusion. A simple ELISA method for Apo stead of two sequential incubations is needed. AI, E, and AIV is reported here utilizing com(2) The presence of 0.3% BSA for Apo AIV mercially available goat anti-rabbit serum IgG assay appeared to be necessary by stabilizing conjugated with horseradish peroxidase. this apolipoprotein, because difficulty was exInstead of carbonate buffer which is comperienced in obtaining consistent and repromonly used for many ELISA, PBST containducible results when BSA was maintained at ing BSA was used in this method for the non0.1% level. The higher concentration of BSA specific binding of antigens to microtiter (0.3%) used in the preparation of Apo AIV plates. Assay had been compared using these samples, however, did compete with Apo AIV two buffers. Results for rat serum Apo AI and molecules and resulted in a weaker color forApo E were found to be identical. However, mation when peroxidase activity was assayed binding of purified Apo AIV was approxi(Fig. 7). This can be compensated by increasmately 50%, and the concentration of serum ing substrate (o-dianisidine) concentration to Apo AIV assayed by ELISA in 0.2 M NaHC03, 200 pg/ml to enhance its color formation. pH 9.6, was only 20% of that obtained in The sensitivity of ELISA method for Apo PBST. Thus, although ELISA for Apo AI and AI was comparable with that of Koritnik and Apo E worked well in carbonate buffer, results Rude1 ( 14). The intraassay coefficients of varifor Apo AIV were not satisfactory. Convenation were 2.3, 4.4, and 5.3% and the inter-

IMMUNOASSAY

FOR

% BsA

FIG. 7. Effect of bovine serum albumin on ELISA of apohpoprotein AIV. Purified rat serum apolipoprotein AIV was heat-activated and then serially diluted in PBST with or without BSA at various concentrations. Other ELISA procedures were as described under Materials and Methods.

assay coefficients of variation were 6.1. 5.5, and 7.9% for .4po AI, E, and AIV. respectively. Thus, this ELISA method is quite reproducible especially when samples are assayed simultaneously. Most importantly, we have demonstrated that apolipoprotein concentrations measured by ELISA were in good agreement with those obtained by densitometry. Quantification by densitometry is totally independent of immunoreactivity of apolipoproteins. it also differs IOOO-fold in sensitivity from ELBA. The agreement therefore confirms that ELISA determination of apolipoproteins is accurate. The serum concentrations of Apo AI, Apo E, and Apo .4IV of normal fed rats were determined by ELBA method. These concentrations were in the same order of magnitude as those reported for rat serum Apo AI and Apo E measured by electroimmunoassay (19,24) or radioimmunoassay (25,26). Some discordance among these laboratories may be due to differences in methods, protein standards or strains of rats used. Therefore it is important that some common standards be established in order to compare data among laboratories. Results of Table 3 and Fig. 6 indicate that both Apo AI and Apo E are approximately l/80 to l/l 00 of the protein in

RAT

APOLIPOPROTEINS

325

rat serum. Thus the serum concentrations of these two apolipoproteins in the rat are very different from those in the human (27.28). Although the concentration of Apo AIV in rat lipoprotein fraction (ci < 1.2 1) had been reported by Roheim and his co-workers (29.30). its concentration in the whole serum was expressed only as arbitrary units with electroimmunoassay. They found that Apo AIV in rat serum was approximately 50 to 90% that of Apo AI. Values obtained with ELISA method are consistant with their findings. Therefore our method to quantify rat serum lipoproteins by ELISA is comparable to radioimmunoassay in sensitivity and precision, without the expense and the hazard involved in the handling and the disposal of radioisotopes. The apolipoprotein concentrations of sera measurable by ELISA remained relatively constant over a period of 6 weeks. as long as the samples had been stored in presence of 0.1% sodium azide at 0°C and heat-treated before assay. Finally, we have proved that our ELISA method is sensitive enough to measure these apolipoproteins (AI, E, and AIV) secreted by hepatocytes cultured as monolayers without a need to concentrate the culture media beforehand. However. secretion rates for both Apo AI and Apo E of cultured hepatocytes reported here are lower than the ones reported by some investigators (3 1,328). This is probably due to differences in culture conditions as well the standards used for assay. Secretion of Apo AIV by cultured hepatocytes has not been reported previously. In conclusion, an ELISA method that can accurately measure apolipoprotein AI, E, and AIV in both rat serum and hepatocyte culture medium has been established. This tool will enable the synthesis and secretion of apolipoproteins to be studied in cultured hepatocytes. ACKNOWLEDGMENTS The author is indebted to William Kossmann Hahn for excellent technical assistance.

and James

326

RENEE

REFERENCES 1. Hamilton. R. L. (I 972) .+I& E-x-p. :\I& R/o/. 26, 724. 2. Marsh. J. B. (1974) J LtptdRe.7 15, 544-550. 3. Windmueller. H. G.. and Spaeth, A. E. ( 1972) J Lipid Rc~s. 13. 92-105. 4. Wu. A. L.. and Windmueller. H. G. (1979) J Blol. Chetn. 254, 73 16-7322. 5. Swaney, J. B.. Braithwaite. F.. and Eder. H. A. (1977) Biochetnistr~~ 16, 271-278. 6. Bell-Quint, J., and Forte. T. (1981) Biochitn. Biophyc. Acta 633, 83-98. 7. Pa&h. W.. Franz, S.. and Schonfeld, G. (1983) J Clin. Invest. 71, 1161-l 174. 8. Melin. B.. Keller. G., Glass, C.. Weinstein. D. B., and Steinberg. D. (1984) Biochitn. Bioplqx .kta 795, 594-588. 9. Davis. R. A.. and Malone-McNeal, M. (1985) Biochetn. J. 227, 29-35. IO. Lin. R. C. (1984) Biochitn Bwp/~~~s. .-k/a 793. l93201. I I Havel, R. J., Eder, H. A., and Bragdon, J. H. ( 1955) J. Clin. Invest. 34, 1345-1353. 12. Tsang, V. C. W., Peralta. J. M.. and Simons, A. R. (1983) in Methods in Enzymology (Langone. J. J.. and Vanvunakis. H.. eds.). Vol. 92. pp. 377-391, Academic Press, New York. 13. Karlin. J. B., Juhn. D. J.. Starr. J. I., Scanu. A. M., and Rubenstein. A. H. (1983) J. LipidRes. 17,3037. 14. Koritnik, D. L., and Rudel. L. L. (1983) J. LipidRe.7. 24, 1639-1645. 15. Katz. J.. Bonorris. G.. and Sellers, A. L. ( 1963) J. Lab. Clin. Med. 62, 9 10-934. 16. Lowry, 0. H.. Rosebrough, N. J.. Farr. A. L.. and Randall. R. J. (1951) J. Biol. Chem. 193.265-275.

C.

LIN 17. Lin. R. C.. and Sondgrass. P. J. ( 1975) Bioc~hcvtz. Biophyv. REV Comtnwt. 64, 715-734. 18. Albers. J. J.. Wahl. P. W.. Cabana. V. G.. Hazzard. W. R.. and Hoover. J. J. (1976) Meudwli.vn 25. 633-644. 19. Wang. L., and Rubinstcin. D. (1978) C‘unad J Bbtc~hctn. 56, I6 I - 166. 20. Roheim. P. S.. Edelstein. D., and Pinter. G. G. ( 1976) Proc. :Vuf. kad. Sci. ITS,-i 73, 1757-l 760. 21. Fainaru. M.. Havel. R. J.. and Imaizumi. K. (1977) Biochem. Btctphys. kta 490, 144-I 55. 22. Steinberg, K. K.. Cooper, G. R., Gaiser, S. R., and Rousseneu. M. ( 1983) C/in. Chetn. 29, 415-426. 23. Lippel, K.. Tyroler. H. A.. Gotto. A.. Brewer, H. B.. Albers. J. J.. Scanu, A.. and Alaupovic. P. (1983) .4rtprictscl~~ro.~is 3, 452-464. 24. Dolphin. P. J., and Forsyth. S. J. ( 1983) J. Lipid Rex 24, 541-551. 25. Cole. T. G., Kuisk, I., Patsch, W.. and Schonfeld, G. ( 1984) J. Lipid Re.y. 25, 593-603. 26. Calandra. S., Gherardi. E.. Fainaru. M.. Guaitani. A., and Bartosek, 1. ( 198 I) B&hem. Biophw. .kto 665, 331-338. 27. Fielding. C. J., Reaven, G. M.. and Fielding. P. E. (1982) Proc. Nat. .-lead. Sci. US.1 79, 6365-6369. 28. Hollenback. C. B.. Connor. W. E.. Riddle, W. C.. Alaupovic. P.. and Leklen, J. E. (I 985) Afetuho/istn 34, 559-566. 29. VanLenten. B. J., and Roheim. P. S. (1982) J. Lipid Res 23, 1187-I 195. 30. DeLamatre. J. G.. and Roheim, P. S. (1983) Bioc%etn. BiophJ,s. .kta 751, 2 1O-2 17. 31. Krul, E.. and Dolphin P. J. (I 982) Biochetn. Rioph?s. .-lcta 713. 609-62 I 32. Patsch. W., Tamai. T., and Schonfeld, G. (1984) J. Clin. Itwest. 72, 371-378.