107

Atherosclerosis, 24 (1976) 107-118 @ Elsevier Scientific Publishing Company, Amsterdam - Printed in the Netherlands

SOLID PHASE RADIOIMMUNOASSAY IN NORMAL HUMAN PLASMA

OF APOLIPOPROTEIN

GILBERT R. THOMPSON *, MARIEL E. BIRNBAUMER, ANTONIO M. GOTTO, JR.

B (APO B)

ROBERT I. LEVY AND

Division of Atherosclerosis and Lipoprotein Research, Department of Medicine, Baylor College of Medicine and The Methodist Hospital, Houston, Texas, and The National Heart and Lung Institute, National Institutes of Health, Bethesda, Md. (U.S.A.) (Received 28th July, 1975) (Revised, received 24th October, 1975) (Accepted 24th October, 1975)

Summary

A solid phase radioimmunoassay (RIA) has been developed for apolipoprotein B (apoB), a major constituent of very low density lipoprotein (VLDL) and low density lipoprotein (LDL) in man. Antisera were prepared against apoB in goats and rabbits, coupled to bromoacetyl cellulose, and the complex was incubated with [ 12’1] LDL. The RIA was based on the displacement of [ 12’1] LDL by unknown samples, as compared with unlabeled LDL standards, using a logit transformation to calculate results. The RIA was found to be satisfactory in terms of precision, sensitivity, reproducibility and specificity. Control subjects had mean apoB levels of 94 mg/lOO ml in whole fasting plasma, of which 3.6 mg/lOO ml plasma were in the VLDL, while 86 mg/lOO ml plasma were in the LDL. Both the triglyceride and apoB content of VLDL, and the cholesterol and apoB content of LDL were positively correlated. It is concluded that the solid phase radioimmunoassay described in the present report provides a rapid and relatively simple means of quantitating apoB in normal human plasma.

Key words:

APO-B -Lipoproteins

-- Radioimmunoassay

-Solid

phase

This work was supported in part by the Lipid Research Contract, NIH-NHLI-71-2156 from the National Heart and Lung Institute. * Present address: Medical Research Council Lipid Metabolism Unit, Hammersmith Hospital. Ducane Road, London W12 OHS (Great Britain).

108

Introduction Apo-lipoprotein B (apoB or apoLDL), the major apoprotein of low density lipoprotein (LDL), is now known to be identical to one of the main apoprotein components of very low density lipoprotein (VLDL) [l].Bilheimer et al. [2] have shown that a significant proportion of the apoB circulating as LDL in human plasma is derived from the partial catabolism of VLDL. To facilitate further studies of the metabolism of VLDL and LDL in man, we have developed a specific and sensitive solid phase radioimmunoassay (RIA) for quantitating apoB. The solid phase method was chosen because of its speed and simplicity. Previously described radioimmunoassays for apoB require the use of a second precipitating antibody [3,4,5], but this has been eliminated in the present method by coupling anti-apoB antibodies to bromoacetyl cellulose [6]. This modification halves the overall incubation time and thus doubles the number of samples which can be assayed in any given period. Methods Preparation of lipoproteins and apoproteins Human plasma lipoproteins were isolated from normal donors or from subjects with defined plasma transport disorders [ 71 as previously described [ 81. Immunochemical purity of the isolated density fractions was determined by double diffusion techniques. Lipid-free VLDL was fractionated on Sephadex G-100 and DEAE-cellulose to yield apoB and the apoC proteins, as previously described [9]. The major apoproteins of HDL (apoA-I and apoA-II) were isolated on Sephadex G.-l50 as described by Scanu et al. [lo]. LDL were delipidated and the apoprotein solubilized in 100 mM sodium decyl sulfate [ll]. The solubilized protein was dialyzed against 100 mll4 Tris HCl buffer, pH 8.0, containing 0.01% EDTA, 0.01% Na azide and 0.1 m&Ysodium decyl sulfate. ApoB was isolated by chromatography on Sephadex G-150 [ 111. After dialysis against 100 mM Tris-HCl buffer, pH 8.0, containing 0.01% EDTA, 0.01% Na azide and 2 mM sodium decyl sulfate, the apoB preparation was stored at 4°C. Preparation and specificity of antisera Antisera to apoB were prepared by injecting either male New Zealand white rabbits or male hybrid goats. The rabbits (2 kg) were given an injection into each toe pad of 0.25 ml of a mixture containing 1 ml of apoB (3 mg/ml), and 0.5 ml of complete Freund’s adjuvant (Hyland). The injections were repeated after 3 weeks. After an additional 3 weeks the animals were exsanguinated by cardiac puncture. Goats (20 kg) were given intradermally 10 ml of a solution containing equal volumes of apoB (0.5 mg/ml) and complete Freund’s adjuvant, together with 2 ml of Pertussis vaccine by intramuscular injection [12]. Booster doses of apoB were given every 3 to 6 months. The antisera selected for use in the RIA were first absorbed with the d > 1.063 fraction of plasma; they then gave a single precipitin line with LDL, VLDL and delipidated

109

Fig. 1. lmmunoreactivity of goat anti-human ape-B, which was placed in each central well. with the following antigens: A: 1 = plasma, 2 = LDL. 3 = HDL. 4 = albumin. 5 = VLDL. B: 1 = apoA-I, 2 = apoA-II, 3 = apoC-III, 4 = apoC-I, 5 = apoC-II. C: 1 = plasma. 2 = LDL, 3 = porcine LDL, 4 = porcine VLDL, 5 = VLDL.

LDL, but did not react with HDL, albumin, apoC-I, apoC-II, apoC-III, apoA-I or apoA-II. They did, however, react with porcine LDL (Fig. 1). Binding of antisera to bromoacetyl cellulose Rabbit or goat antisera to apoB were concentrated by precipitation with an equal volume of 28% sodium sulfate at room temperature. After centrifugation the precipitate was dissolved in a volume of 0.15 M NaCl equal to the original volume of plasma and was dialyzed against several changes of 0.15 M NaCl. An equal amount of anti-apoB was added to bromoacetyl cellulose to give a 1 : 1 (w/ w) ratio of protein to adsorbant [ 6,131; the bromoacetyl cellulose was previously sonicated to increase the uniformity and fineness of the particle size. Identical procedures were used to couple non-immune rabbit and goat globulins to bromoacetyl cellulose, which were then stored in 10 volumes of 0.15 M NaCl at 4°C. Prior to use the stock suspensions were diluted in 3% human serum albumin/borate buffer. Iodination of LDL LDL was iodinated with ‘*‘I using,a modification of the iodine monochloride method of McFarlane [14]. The LDL was first concentrated to approximately 10 mg protein/ml by dialysis against solid polyethylene glycol (mw 20,000, Fisher Scientific). After dialysis against 0.1 M glycine buffer, pH 10 [15], 1 mg of LDL-protein in 0.1 ml glycine buffer was mixed with 3 mCi of Na “‘1 (Amersham-Searle), in 50 ~1 glycine buffer; 25-50 ~1 of iodine monochloride were then jetted into the mixture. The iodine monochloride was prepared by diluting a 0.02 M stock solution with 2 N sodium chloride [16] so as to provide one atom of iodine per molecule of LDL-protein, assuming a molecular weight of 100,000 for apoB and an average efficiency of iodination of 33%. After 1 min the mixture was dialyzed extensively against 0.15 M NaCl and, finally, overnight against borate buffer, pH 8.0. Paper electrophoresis of [ “‘I] LDL gave one band of radioactivity that coincided in migration with unlabeled LDL. A single peak of radioactivity was ob-: tained when [ 125I] LDL was chromatographed on a column of 2% agarose. By the addition of anti-apoB sera, 94% of the total counts was precipitated. A

110

similar percentage of the radioactivity was precipitated by the addition of an equal volume of 10% trichloro-acetic acid. Only 2-3s of the radioactivity was extracted with chloroform methanol (2 : 1). [12’1] LDL (10 pg/ml) was stored at 4°C in 3% albumin/borate buffer pH 8.0, containing 0.01% sodium azide, and was redialyzed and passed through a 7 /J Millipore filter before use, on the morning of each assay. Standard solutions of LDL (50-10,000 ng/O.l ml) were prepared and quantitated as described above, and were stored at 4” in the 3% albumin/borate buffer. Since LDL isolated between d 1.006-1.063 gave virtually identical results to LDL isolated between d 1.025-1.050, the former was routinely prepared for use as a standard. Radioimmunoassay

procedure

Reagents were always added in the following order: labeled antigen, buffer, unknown or standard solution, and antiserum. By including 3% human serum albumin (Fraction V, Sigma) in the assay system and by using polypropylene tubes, non-specific binding was reduced to 2.5% of the counts added. Using an automatic dispenser (Micromedic), 0.1 ml of [‘2sI] LDL (lo-20 ng of protein containing approximately 10,000 cpm), and 0.1 ml of borate buffer, pH 8.0, containing 0.01% EDTA, 0.01% sodium azide and 3% albumin were added to each assay tube; 0.1 ml of a standard or unknown solution was then added manually with immediate mixing. Finally, 0.1 ml of the appropriate bromoacetyl celluloseantibody complex, diluted such that 0.1 ml bound 50% of the [ 12’1] LDL in the absence of unlabeled antigen, plus 0.1 ml of the buffer were added. After mixing, the tubes were incubated without further agitation for 3 h at 25” followed by 12 h at 4”) or for 24 h at 4”. After incubation, 2 ml of 3% albumin borate buffer were added with mixing, and each sample was centrifuged for 30 min at 5000 X g. The supernatant solution was decanted and the residual fluid was removed by vacuum aspiration with a Pasteur pipette. Radioactivity in the precipitate was determined in an autogamma counter (Packard). Corrections were made for any non-specific uptake by substituting a non-immune globulin-bromoacetyl cellulose complex. Bound to free ratios of radioactivity and the logit of the % of radioactivity bound [ 171 were calculated and plotted against the log of the apoB concentration of the standard. The values of the unknown samples were determined by mathematical interpolation on the regression line, using a Wang Programming Calculator. The logit of the % of “‘1 bound was calculated as follows: CPM __-CPM bound in presence of sample-control CPM boundinthe absence of sample-control CPM Logit = 2.303 X log,, 1 _ CPM bound in presence of sample-control CPM CPM bound in the absence of sample-control CPM Other methods

Protein was determined by the method Gustafson et al. [ 191. Plasma cholesterol formed using the autoanalyzer techniques Clinics Program.

of Lowry et al. [ 181 as modified by and triglyceride analyses were perof the collaborative Lipid Research

111

Results Standard curves A typical standard curve for the displacement of [‘*‘I] LDL by lo-10,000 ng of LDL exhibited a characteristic sigmoid appearance when the log of the LDL-protein concentration was plotted against the bound to free ratio (B/F) (Fig. 2). A plot of the logit of the amount of radioactivity bound against LDLprotein yielded a similar curve, but increased the range of linearity. Maximum sensitivity of the assay was in the range 50-100 ng of LDL-protein, using one batch, of rabbit antiserum, and 100-1000 ng with goat antiserum, as shown in Fig. 2. Immunoassay controls When unlabeled LDL-protein was omitted, more than 95% of the total radioactivity was recovered in the precipitate after incubation of [**‘I] LDL with anti-apoB. More than 90% of bound radioactivity could be displaced by addition of 10,000 ng LDL-protein. By contrast, when a non-immune globulin fraction was coupled to bromoacetyl cellulose, less than 5% of the total counts was precipitable. Effect of storage on [lzsI] LDL and on LDL standards To determine the effects of storage on [ 1251]LDL and unlabeled LDL standards (see Methods), freshly prepared and stored [‘*‘I] LDL were incubated with LDL standards. After 15 to 20 days of storage, [1251]LDL exhibited abnormalities of uptake by antibody and displacement by a fresh LDL standard. [‘*‘I] LDL was, therefore, prepared freshly every 2 weeks. Incubation of

0.6 8 PY 0.4

I

I

I

I

I

50

lcm

am

MO

500

II ml ml

ZE

I SW0

1

ngProtem Fig. 2. Standard curves for radioimmunoassay of LDL. The concentration of LDL-protein is plotted on a log scale vs. the amount of [ 12sII LDL specifically bound to antibody. The latter is expressed both as the bound to free ratio of radioactivity and as the logit of the I bound radioactivity, The vertical scales are not directly comparable.

112

fresh some came were

and aged LDL standards with freshly prepared [ 1251]LDL also revealed loss of potency of the standards with time. However, this effect only beapparent after two months storage. Consequently, fresh LDL standards prepared every 4-6 weeks.

Criteria of assay performance

The precision of the RIA was calculated by determining the mean *SD of 10 estimates of the logit of the % bound at each point on the standard curve. Over the concentration range of 100-2000 ng of LDL protein, there was a linear correlation between the logit of the % bound and the amount of LDLprotein added; the average standard deviation of the logit % ‘*‘I bound was kO.13. Using the goat anti-apoB antiserum, the sensitivity of the assay was 50 ng of apoB. The within assay variability, expressed as the mean standard deviation of multiple samples, each assayed in duplicate, was 24.75 for the d > 1.006 fraction of plasma, and 20.29 for the d < 1.006 fraction. Between assay variability, expressed as the mean standard deviation of several samples, each assayed in duplicate on 5 separate occasions, was +10.9 for the d > 1.006 fraction, +0.46 for the d < 1.006 fraction of plasma, and +13.4 for whole plasma. These values represent between assay coefficients of variation of 14,lO and 15%, respectively. The specificity of the assay was assessed by observing the effect of the substance in question on the uptake of [ ‘251] LDL by antibody and nonimmune globulin complexes in both the presence and absence of unlabeled LDL. All the apoproteins known to be present in normal VLDL and HDL were tested. Native HDL, apoA-I, apoC-I or apoC-II had no effect when 100-1000 ng of each was added to the assay system. Both apoC!-III and to a lesser extent, apoA-II, significantly reduced the uptake of [ 1251]LDL by anti-apoB antibody when added alone but not if previously mixed with an excess of phospholipid vesicles [20]. This indicates that under normal circumstances, where these apoproteins exist complexed to phospholipids, neither will interfere with the immunoassay of LDL. Neither the d > 1.21 fraction of normal plasma nor the plasma from a patient with abetalipoproteinemia [21] caused any significant displacement of [‘*‘I] LDL. Immunoreactivity

of apoB, LDL and VLDL

Although the antiserum used in these assays was prepared to apoB, the lipidfree apoprotein (solubilized in sodium decyl sulfate) was only one-eighth as potent as native LDL in the displacement of [ ‘251] LDL. Analogous results were obtained when VLDL was assayed, displacement of [‘*‘I] LDL again being less than with native LDL but greater than with delipidated apoB. On the basis of seven separate assays, the mean ? 1 SD content of apoB in VLDL was 25 * 5.5%, expressed as a percentage of the total protein present as determined by the Lowry procedure [ 181. Assay of apoB in plasma and its sub-fractions

Repetitive d < 1.006

estimations of the apoB content of whole plasma, and of the and d > 1.006 fractions thereof. were carried out in 8 male sub-

113 TABLE 1 apoB CONTENT FROM HEALTHY

OF WHOLE PLASMA,

AND ITS d < 1.006

AND d > 1.006

SUBFRACTIONS,

TAKEN

SUBJECTS

For the d < 1.006 fraction. the sample was diluted 1 was 1 : 500. Plasma

Single samples (8 subjects) Multiple samples: subject 1 2a 3 4

94 * 33 1

: 25

before assaying. For the d > 1.006, the dilution

apoB (mg/lOO ml) d < 1.006

d > 1.006

3.6 3.7 2.6 2.9 2.1

88 97 IO 84 73

+ 1.4 c 1.5 ?r0.8 + 0.6 f 0.7

f f k f +

33 16 10 8 10

__--

Results give the mean + 1 SD. a Female.

jects, each value being the mean of 5 separate assays. In addition, multiple samples of plasma were taken on 6-12 occasions over a period of up to 6 months from another 4 subjects. These samples were subjected to ultracentrifugation at d 1.006 and the apoB content of the 2 sub-fractions determined in at least 2 separate assays. The results are shown in Table 1. From the results it is evident that the majority of the apoB in normal plasma is located in the d > 1.006 fraction. The average recovery of apoB in the d < 1.006 plus the d > 1.006 fractions after ultracentrifugation at d 1.006 averaged 95% of that found in whole plasma. Variability of results within samples from each subject was less than between subjects, as might be expected. Correction

between

lipid and protein

content

of lipoprotein

fractions

Samples of plasma were obtained from normal subjects and lipoprotein fractions were isolated by ultracentrifugation. The amount of apoB in the d < 1.006 and d > 1.006 fractions was determined by RIA, whereas the total protein content of the d 1.006-1.063 fraction was measured by the Lowry method. In addition, the d < 1.006 fraction was analyzed for triglyceride

TABLE 2 CONCENTRATION OF LDL PROTEIN AND CHOLESTEROL FRACTIONS OF 25 DIFFERENT SAMPLES OF PLASMA Plasma fraction

Protein

IN THE d 1.006-1.063

Cholesterol

(mg/lOO ml)

(mg/lOO ml)

d 1.006-1.063 d > 1.006

60.5 f 8.8 a 78.7 ? 10.8 b

78.8 + 18.8 ’ 102.4 ?: 16.8 d

Paired t-test

P < 0.01

P < 0.01

Results show the mean k 1 SD. a Determined by the Lowry procedure. b Determined by immunoassay. ’ Measured after Folch extraction. d Determined as LDL-cholesterol.

Cholesterol

AND d > 1.006

: protein

1.3 1.3 ~__

-

114 80 z

70-

g

60-

r = 0.69

; 50E ti uo3 ; 30s A

20-

:

IO-

.

0’

I

1

1

1

2

3

I 5

I

4 d 1.006 fraction was estimated after

the LDL cholesterol content of the heparin and manganese precipitation, as previously described [22]. As shown in Table 2, the concentration of LDL was approxiprotein and cholesterol in the fraction between d 1.006-1.063 mately 20% less than that in the d > 1.006 fraction of plasma. Since the LDL-cholesterol to protein ratio was identical in isolated LDL and in the d > 1.006 fraction of plasma, this suggests that considerable losses of LDL occurred during ultracentrifugation at d 1.063. (This was confirmed by measuring the distribution of ‘*‘I in ultracentrifugal fractions of plasma obtained 10

130 120 L

0

a9 0

8”

Cl

0

/

,”

0

i

A 40 30 2“, 50 60 6D

10 III1

0

:

III 60

40 20 80 d 1.006-1.063 PROTEIN ~LowrylmgilOO ml

Fig. 4. Correlation between the cholesterol between the LDL cholesterol and apoLDL

I

0

I1

I1 20

I 40

I1 60

I 80

I

I’ 100

d>l.W6apoLDLlRIAImgl100ml

and protein content of LDL (d 1.006-1.063). r = 0.67, and (apoB) content of the d > 1.006 fraction of plasma, r = 0.75.

115

,

Ii

5

6

7

I

I

I

I

II

8

9

IO

11

12

1

13

14

I

15

I

16

DAYS AFTER INJECTION

Fig. 5. Comparison of specific activity of [ 1z511 LDL in the d > 1.006 fraction of plasma, expressed as cpm/mg of immunoassayable apoLDL (apoB), and in LDL. expressed as cpm/mg of calorimetrically determined LDL-protein.

min after i.v. injection of [ 1251]LDL into 6 normal subjects. Although a mean of 94 + 1.8% of the injected radioactivity was present in the d > 1.006 fraction only 73 + 7.3% was recovered in the d 1.006-1.063 fraction.) In the d < 1.006 fraction, there was a good correlation between VLDLtriglyceride and the amount of immunoassayable apoB (Fig. 3). Similar correlations were found between total cholesterol and protein in LDL and between LDL-cholesterol and apoB in the d > 1.006 fraction of plasma (Fig. 4). Determination

of LDL specific

activity in turnover

studies

One potential use for the RIA is the determination of apoB specific activity in studies of the turnover of [ 12’I]LDL. Samples of plasma were obtained during the course of a study of [ 12’I]LDL turnover in a normal subject, and the specific activity of the d > 1.006 fraction of each sample measured by counting the radioactivity present in 1 ml aliquots and quantitating the apoB present in 0.1 ml of a 1 : 500 dilution. (The large dilution factor and small volume of sample used reduces the interference by added ‘?‘I to negligible amounts.) The results were compared with those obtained by isolating LDL from the same samples by ultracentrifugation and then counting the radioactivity present (using an internal standard to correct for quenching by KBr) and quantitating protein con-

116

tent by the Lowry procedure. It is evident methods gave very similar results.

as shown

in Fig. 5 that

both

Discussion This paper describes the development of an immunoassay for quantitating human apoB, the major apoprotein in LDL and VLDL [l]. The major differences between the present report and those described previously is that antisera were raised to purified apoB, rather than to LDL, and that the antisera were coupled to bromoacetyl cellulose thus eliminating the need for a second precipitating antibody. The reason for using apoB antisera is that antibodies to apoB form precipitin lines of complete identity between LDL, VLDL and apoB, whereas this does not always occur with antibodies to LDL [ 11. However, it is probable that coupling anti-LDL antibodies to bromoacetyl cellulose would give very similar results in the RIA to those obtained with anti-apoB antibodies. Various methods are available for expressing the displacement of labeled antigen from antibody by unlabelled antigen. The advantage of the method used in this study, namely logit transformation, is that it gives a linear relationship over a wider range of LDL concentration than the B/F ratio [17] and also simplifies quality control [ 231. The statistical procedures used to assess the reliability and performance characteristics of the assay were those described by Midgley et al. [24]. In general, the precision, sensitivity, specificity and reproducibility of the present assay compare favorably with other immunoassays, as discussed by Galskov [ 251 and also recently reviewed by Skelley et al. [ 261. Comparison of the displacement of [ “‘I] LDL from anti-apoB antibody showed that LDL was considerably more effective on a weight basis than apoB. The relative inefficiency of apoB may have been due to the presence of the detergent necessary to maintain it in solution or alternatively it is possible that some of the antigenic sites become masked in the absence of lipid, due to aggregation. Estimates of the apoB content of VLDL obtained by immunoassay during the present study were somewhat lower than those in two previous reports based on chromatographic separation and calorimetric quantitation [27,28], but similar to those reported by Schonfeld et al. [4] who used both the chromatographic and immunochemical methods of quantitation. Discrepancies between the two methods could have been due to selective losses of smaller peptides during chromatographic procedures or delipidation, leading to an overestimate of apoB content. Alternatively, it is possible that immunoassay methods underestimate the apoB content of VLDL, due to incomplete exposure of antigenic sites [ 31. Further studies will be required before this issue can be satisfactorily resolved. One of the major advantages of radioimmunoassay is that it permits quantitation of apoB in situations where the latter occurs in lower concentrations than are present in plasma. These include peripheral and intestinal lymph, and certain tissues such as the arterial wall. Also, in view of the current interest in tissue culture techniques as a means of studying lipoprotein metabolism, it is possible that radioimmunoassay could eventually be used to measure apoB synthesis in vitro. Another use of radioimmunoassay is in the mea-

117

surement of apoB specific activity during LDL turnover studies, as illustrated in this paper. Quantitation of the apoB content of the d > 1.006 fraction of plasma eliminates the necessity to perform the two extra ultracentrifugal spins which are required if the protein content of LDL is to be determined by the Lowry procedure. Values obtained for the concentration of apoB in human plasma and its sub-fractions in the present study were very similar to those reported by Schonfeld et al. [4] and Bautovich et al. [ 51, who used double immunoassay procedures. The solid phase method, however, is far less timeconsuming; one technician can analyse up to 40 samples in duplicate in 24 hours. It is of interest that Aubert [29] has suggested that solid phase techniques will eventually replace the double antibody procedures which are currently used to assay polypeptide hormones in plasma. Acknowledgements We are grateful to Dr. D. Bilheimer for his advice on the iodination of LDL, to Mr. Mauro Nava for his unstinting efforts and technical expertise, to Mr. Richard Plumlee for his help with the computation of results, and to Dr. Richard Jackson for his helpful criticism. References 1 Gotto, A.M., Brown. W.V., Levy. R.I.. Bimbaumer, M.R. and Fredrickson. D.S., Evidence for the identity of the major apoprotein in low density and very low density lipoproteins in normal subjects and patients with familial hyperlipoproteinemia, J. Clin. Invest., 51 (1972) 1486. 2 BiIheimer, D.W.. Eisenberg, S, and Levy, R.I., The metabolism of very low density lipoproteins. Part 1 (Preliminary in vitro and in viva observations), Biochim. Biophys. Acta. 260 (1972) 212. of beta lipoprotein-protein of rat serum. J. Clin. 3 Eaton, R.P. and Kipnis, D.M., Radioimmunoassay Invest., 48 (1969) 1387. 4 Schonfeld, G., Lees, R.S.. George, P.K. and Pfleger, B., Assay of total plasma apolipoprotein B concentration in human subjects, J. Clin. Invest., 53 (1974) 1458. 5 Bautovich, G.J., Simons. L.A.. Williams, P.F. and Turtle, J.R., Radioimmunoassay of human plasma apolipoproteins, Part 1 (Assay of apolipoprotein-B). Atherosclerosis, 21 (1975) 217. 6 Jagendorf. A.T., Patchornik, A. and &la, M., Use of antibody bound to modified cellulose as an immunospecific adsorbent of antigens, Biochim. Biophys. Acta. 78 (1963) 516. In: J.B. Stanbury. J.B. Wyngaarden 7 Fredrickson, D.S. and Levy. R.I., Familial hyperlipoproteinemia. and D.S. Fredrickson (Eds.), The Metabolic Basis of Inherited Disease, McGraw-Hi& New York, N.Y., 1972, P. 545. 8 Havel, R.J., Eder, H.A. and Bragdon, J.H. The distribution and chemical composition of ultracentrifugally separated lipoproteins in human serum, J. Clin. Invest., 34 (1955) 1345. 9 Brown, W.V.. Levy, R.I. and Fredrickson. D.S.. Studies of the proteins in human plasma very low density lipoproteins, J. Biol. Chem., 244 (1969) 5687. 10 Scanu, A., Toth. J., Edelstein, C., Koga, S. and Stiller, E., Biochemistry, 8 (1969) 3309. A.S., Birnbaumer, M.E. and Fredrickson. D.S., The structure 11 Gotto, A.M., Levy, R.I., Rosenthal, and properties of human betalipoprotein and betaapoprotein, Biochim. Biophys. Research Communications, 31 (1968) 699. 12 Vaitukaitis, J.. Robbins. J.B., Nieschlag, E. and Ross, G.T.. A method for prodwin: specific antisera with smaIl doses of immunogen, J. Clin. Endocrin., 33 (1971) 988. 13 Mann, D.. Granger. H. and Fahey, J.L., Use of insoluble antibody for quantitative determination of small amounts of immunoglobulin, J. Immunol.. 102 (1969) 618. 14 McFarlane, A.S., Munro, H.N. and Allison, J.B. (Eds.), Metabolism of Plasma Proteins in Mammalian Protein Metabolism, Academic Press. New York and London, 1964, p. 331. of proteins by the iodine mono15 Hehnkamp. R.W., Contreras. M.A. and Bale, W.F., I 13Qabelling chloride method, Int. J. Appl. Rad. Isotop., 18 (1967) 737. 16 Helmkamp. R.W.. Goodland, R.L., Bale, W.F., Spar, I.L. and Mutschler. L.E., High specific activity iodination of y-globulin with iodine-131 monochloride. Cancer Res.. 20 (1960) 1495.

118

17

Rodbard, D.. Rayford, P.L., Cooper, J.A. and Ross, G.T.. Statistical quality control of radioimmunoassays, J. Clin. Endocrin., 28 (1968) 1412. 18 Lowry. O.H., Rosebrough. N.J., Farr, A.L. and Randall, R.J.. Protein measurement with the Folin phenol reagent. J. Biol. Chem., 193 (1951) 265. 19 Gustafson, A.. Alaupovic, P. and Furman. R.H., Studies of the composition and structure of serum lipoproteins: isolation, purification and characterization of very low density lipoproteins of human serum, Biochemistry, 4 (1965) 596. Huang, C.. Studies on phosphatidylcholine vesicles - Formation and physical characteristics, Biochemistry. 13(1969) 344. 21 Fredrickson. D.S., Gotto, A.M. and Levy, R.I., Familial alpha and beta lipoprotein deficiency. In: J.B. StanbuN, J.B. Wyngaarden and D.S. Fredrickson (Eds.), The Metabolic Basis of Inherited Disease. McGraw-Hill, New York. N.Y., 1972, p. 493. 22 Fredrickson. D.S., Levy, R.1. and Lees. R.S.. Fat transport in lipoproteins - An integrated approach to mechanisms and disorders, New Engl. J. Med., 276 (1967) 148. 23 Rodbard, D. and Lewald. J.E.. Computer analysis of radioligand assay and radioimmunoassay data, Acta Endocrin., Suppl.. 147 (1970) 79. 24 Midgley, A.R., Niswender, G.D. and Rebar, R.W.. Principles for the assessment of the reliability of radioimmunoassay methods (precision, BCCWPCY. sensitivity and specificity), Acta Endocrin. (Copenhagen), Suppl. 142 (1969) 163. corticotrophin determination, Acta Endocrin. (Copenhagen), 25 Galskov. A.. Radioimmunochemical Suppl. 162 (1972) 5. 26 Skelley, D.S., Brown. L.P. and Besch. P.K., Radioimmunoassay, Clin. Chem., 19 (1973) 146. of the human 27 Brown, V.W., Levy, R.I. and Fredrickson, D.S., Further separation of the apoproteins plasma very low density lipoproteins, Biochim. Biophys. Acta, 200 (1970) 573. 28 Eisenberg. S.. Bilheimer. D., Lindgren, F. and Levy. R.I., On the apoprotein composition of human plasma very low density lipoprotein subfractions. Biochim. Biophys. Acta, 260 (1972) 329. assay for the dosage of the polypeptide her29 Aubert. M.L.. Critical study of the radioimmunological mows in plasma, J. Nucl. Biol. Med.. 14 (1970) 1. 20