different approachesfor the interpolation of the dose-response curve. J Nucl Med Allied Sci 1982;26:235-48. 15.

Pi1 A, Zucchelli GC, Chiese MR, Fersleghim H. Performances of IRMAs vs RIAs compared through their precision profiles [Abstract]. Nuklearmedizin 1987;26:144-5.

16. Zucchelli GC, Pilo A, Clerico A, et al. The TDx assay for cyclosporine and its metabolites in blood samples compared with HPLC and RIA methods. Drugs Exp Clin Res 1989;15:185-8. 17. Billingham ME. Diagnosis of cardiac rejection by endomyocardial biopsy. J Heart Transplant 1980;1:25-30.

CLIN. CHEM. 36/12, 2109-2113 (1990)

Insufficient Accuracy and Specificity of Polyanion Precipitation Methods for Quantifying Low-Density Lipoproteins
RUdigerSlekmeier,WinfrledM#{228}rz, and WernerGroB1 Recently, polyanion precipitation assays for low-density lipoprotein (LDL)-cholesterol have been found to underestimate their analyte in normolipidemic samples (Siekmeier et al., Clin Chim Acta 1988;177:221-30). Therefore, accuracy, specificity, and interference by nonesterified fatty acids have been studied for three precipitants (obtained by hepann, dextran sulfate, or polyvinyl sulfate precipitation). At normal concentrations of LDL, precipitation is incomplete, whereas it is nearly quantitative at high concentrations of LDL. The polyvinyl sulfate reagent markedly responds to variations in the amount of non-LDL protein present in the precipitation mixture. In the dextran sulfate and the polyvinyl sulfate method, but not in the heparin method, the percentages of LDL precipitated notably increase as the concentration of the polyanion compound is decreased. In either assay, very-lowdensity lipoproteins, but not high-density lipoproteins, are significantly coprecipitated (dextran sulfate 28%, polyvinyl of HDL-C after precipitation of apolipoprotein (ape) B-containing lipoproteins (4-7) and subsequent calculation of LDL-C according to Friedewald et al. (4). In the meantime, methods aimed at the selective precipitation and direct quantification of LDL-C have been designed and are commercially available after
(8-10).

that concentrations precipitation with polyanion compounds are lower than those obtained with a combined ultracentrifugation-precipitation procedure (11), we have evaluated the accuracies and specificities for LDL-C of three assays (810) based on precipitation of LDL with either heparin, dextran sulfate (DS), or polyvinyl sulfate (PVS).
Starting from our recent observation
of LDL-C

Materials and Methods Apparatus. For preparative ultracentrifugation

we used

sulfate and heparin 66%) in a concentration-independent

fashion. increased concentrations of nonesterified fatty acids markedly interfere with the dextran sulfate and polyvinyl sulfate assay, but do not much affect results with the heparin reagent. Additional Keyphrases: intermethod comparison
analytical error Hypercholesterolemia represents one of the primary risk for the premature development of atherosclerosis. Whereas increased concentrations of low-density lipoproteins (LDL) confer a high cardiovascular risk, high-density

cholesterol

factors

lipoproteins (HDL) are thought to exert protective effects (1_3).2 Hence, individual risk proffles should include assays for HDL and LDL. A common approach to the quantification of lipoprotein classes has been the measurement

Gustav Embden-Zentrum der biologischenChemie der J. W. Goethe-Umversitat, Frankfurt/Main, F.R.G. Author for correspondence. 2Nodjd abbreviations: VLDL, LDL, HDL, very-low-, low-, and high-density lipoproteins; VLDL-C, LDL-C, HDL-C, cholesterol in VLDL, LDL, and HDL; DS, dextran sulfate; PVS, polyvinyl sulfate; LDL-C, LDL-CDS, LDL-C8, LDL-C as measured after heparin, DS, or PVS precipitation; LDLUC,PI,T, LDL-C measured after a combination of ultracentrifugation and phosphotungstic acid/MgCl2 precipitation; and FFA, free fatty acids. Received May 24, 1988; accepted October 4, 1990.

a Model L 8-70 (Beckman Instruments, Fullerton, CA) or a Model TGA 75 (Kontron AG, Analytical Division, Zurich, Switzerland) ultracentrifuge with fixed-angle rotors (Kontron types TFF 50.38 and TVF 45.6). Densities were monitored with a DMA 55 digital precision density meter (A. Paar KG, Gras, Austria). Reagents and other materials. Reagent kits for cholesterol (Monotest CHOD-PAP) and triglycerides (GPO-PAP) were purchased from Boehringer Mannheim, Mannheim, F.R.G. The NEFAC test for the enzymatic determination of nonesterified (free) fatty acids was obtained from Wako Chemicals, Neuss, F.R.G. Celite 545 (analytical grade), palmitic acid, stearic acid, and Tris were from Serva Feinbiochemica, Heidelberg, F.R.G.; other chemicals were from E. Merck, Darmstadt, F.R.G. Blood collection. Blood samples from ostensibly healthy donors, mostly women, ages 20 to 30 years, were obtained by venipuncture after an overnight fast and drawn into tubes containing K2EDTA at a final concentration of 1.5-2 g/L (Sarstedt, NUmbrecht, F.R.G.). Plasma was recovered by centrifugation (1500 x g, 30 min) and stored at 4#{176}C. Lipid analyses were completed within no more than five days after blood collection. Quantification of LDL-C. LDL-C was determined either after precipitation with heparin (LDL-C), dextran sulfate (LDL-Cns), polyvinyl sulfate (LDL-C8), or by a combined ultracentrifiigation and phosphotungstic acidl MgC12 method (LDL-CUc,phT). For LDL-C (8), we added 100 uL of sample to 1.0 mL of precipitant (sodium citrate, 64 mmol/L, pH 5.04, and heparin, 50 000 U/L), swirled, and incubated the mixture CLINICAL CHEMISTRY, Vol. 36, No. 12, 1990 2109

for 10 mm at room temperature. The precipitate was pelleted by centrifligation at 3500 x g for 15 mm. We determined the cholesterol in 0.5 mL of the supernate within 1 h. For LDL-CDS (9), we added 1.0 mL of precipitant (Quantolip LDL cholesterol; Immuno AG, Vienna, Austria) containing Tris HC1, 0.2 mmol/L, pH 7.4, and dextran sulfate (Mr 500 000) to 100 L of sample. After 10 mm of incubation at room temperature and centrifugation at 3500 X g for 10 mm, we measured the cholesterol in the supernate. For LDL-CPVS (10), we mixed 200 aL of sample and 100 pL of precipitant (LDL cholesterol PVS; Boehringer Mannheim) containing, per liter, 5 mmol of Na2EDTA, pH 4.2, 1 g of PVS, and 169 g of polyethylglycol methyl ether as accelerator. After 15 mm of incubation at room temperature and 15 mm of centrifugation at 1500 x g, we removed 150 pL of the supernate for the cholesterol determination. LDL-CHEP, LDL-CDS, and LDL-CPVS were calculated as the differences of total cholesterol minus the cholesterol in the supernates after LDL precipitation. For LDL-CUchT (7), we overlayered 2 mL of plasma with 1 mL of a solution containing 100 mg of Na2EDTA and 11.42 g of NaCl per liter (solvent density, d20 = 1.0063 kg/L) in thick-walled polycarbonate tubes and centrifuged the gradient in the TFF 45.6 rotor (2.35 x 10 g mm, 15 #{176}C). The supernatant VLDL were collected by aspiration with a hypodermic syringe and the fraction volume was made up to 2 mL with isotonic saline (NaCl 150 mmol/L) before cholesterol was determined. HDL-C was quantified after precipitation of apolipoprotein B-containing lipoproteins with a solution of 3.6 g of sodium phosphotungstate and 45 mmol of MgC12 per liter (supplementary reagent for HDL cholesterol, Boehringer Mannheim). LDL-CUC,PhT was calculated as the difference of total plasma cholesterol minus the sum of VLDL-C and HDL-C. Preparation of lipoproteiris. Lipoprotein fractions were isolated from pooled sera of normolipidemic and healthy donors of both sexes, ages between 20 and 30 years. Before fractionation, the sera had been supplemented with Na2EDTA and NaN3 to final concentrations of 0.4 and 0.5 g/L, respectively. Trace amounts of chylomicrons and particles were removed after centrifugation for 30 mm at 30 000 x g (d20 c = 1.0063 kg/L). Subsequently, VLDL, LDL, and HDL were prepared by sequential ultracentrifugation at solvent densities (d ) of 1.0063 kg/L (2.3 108 g mm), 1.063 kg/L (2.6 108 g mm), and 1.21 kg/L (3.4 108 g mm), respectively. Lipoprotein preparations were washed once by recentrifugation under identical conditions, stored at 4#{176}C, and dialyzed against freshly prepared Krebs-Ringer solution (per liter: 119 mmol of NaCl, 4.7 mmol of KC1, 2.5 mmol of CaC12, 25 mmol of NaHCO3, 1.2 mmol of KH2PO4, and 1.2 mmol of MgSO4, pH 7.4) just before use. Unless otherwise indicated, this solution was
.

drawn stirred

fasting

plasma,

and the suspension

was carefully

for 30 mm. Finally, the Celite was removed by two centriftigations at 2 #{176}C and 15 000 x g for 10 mm, and the supernate was passed through a membrane filter (pore size 1.2 pm). Adsorption of LDL to the Celite particles was ruled out by a control experiment in which 600mg (hexanetreated) Celite without FFA was added to 4 mL of serum: LDL-C measured with either precipitation method was not affectedby exposure to Celite particles alone. Statistics. Nonparametric linear regression (13) and nonlinear curve-fitting (14) were carried out as described.The
relationship
recovered

between

the

percent

(LDL-C)

and LDL-C actually

proportion of LDL-C present (LDL-

Cp,.es) was assumed to follow the function LDLCrec

=

LDLCpres

ki/(LDLCpre8

+ k2)

so that LDL-C measured in plasma samples (LDLCme) could be corrected for LDL recoveries (%) by the following equation: LDLCmea8 100 2 k1

LDL-C8

1!

/ (LDLCmeas

100)2
+

LDLCmeas

100

4#{149}k12

Results
percentages

In experiments with ultracentrifugally purified LDL, the of LDL precipitated by DS and PVS depended on the LDL-C concentration (cf. Figure 1). For LDL-C at 3.5 mmolJL, both PVS and DS precipitated no more than 70%

of the initial
(>95%) mmoliL

LDL content. Quantitative

precipitation
<2.0 purirange we obthe phospho-

was obtained for LDL-C concentrations >12 (about threefold the normal concentration). By

contrast, mmol/L),

even at low LDL concentrations (LDL-C

heparin

almost quantitatively

precipitated

fied LDL. Within the same concentration tained complete precipitation of LDL with tungstate/MgCl2 reagent.

To study the influence of plasma proteins on the precip-

a) a
CC. C a)
U)

a)

0
U -J -J

also used to adjust the concentrations of lipoprotein fractions.

Enrichment of plasma with fatty acids. Fatty acids were

added to ordinary plasma according to Spector and Hoak (12). In brief, 0.5 mmol of stearic acid and 0.5 mmol of
palmitic acid were dissolved in 10 mL of hexane. We spread 10 g of Celite in a beaker, the thickness of the solid phase not exceeding 5 mm. The FFA-hexane solution was added, and the organic solvent was evaporated under a stream of N2. Weighed amounts (ranging from 25 to 650 mg) of dry FFA-coated Celite particles were added to 4 mL of freshly

1216
[mmol/L]

20

LDL-cholesterol

Fig. 1. Precipitation of ultracentnfugally prepared LDL with polyanions: percentage of LDL-C precipitated vs LDL-C (U) heparin, (A) DS, (#{149}) PVS. The insert shows the effect of adding LOL-depleted serum to the precipitation mixture:the differences between the reagents disappear, but the percentage of LDL precipitated remains dependent on LDL concentration

2110 CLINICAL CHEMISTRY, Vol. 36, No. 12, 1990

itation of LDL, we supplemented reaction mixtures with plasma fractions of d20 >1.063 kg/L (insert, Figure 1). Thereby, we accounted for coprecipitation of Lp(a) and lipoprotemns of d20 c >1.063 kg/L by assaying controls that lacked LDL. This confirmed the dependency of LDL recoveries on the initial concentrations of LDL. However, the presence of LDL- and VLDL-depleted serum markedly improved LDL recoveries with the DS and the PVS reagent: about 90% of LDL was precipitated from LDL-C at 3.5 mmol/L. Consequently, we precipitated LDL either in the absence or in the presence of bovine serum albumin (3.6 or 7.2 g/L), but omitted the d20 c >1.063 kg/L fraction. Intriguingly, the addition of albumin led to an increase in the percentage of LDL precipitated by PVS but not by DS. To rule out the possibility that this difference between PVS and DS was due to the higher dilution of the sample in the DS method, we added albumin directly to the DS reagent. This indeed brought about minor increases in the percentage of LDL precipitated. However, these differences were far too slight to account for the differences observed between precipitations in purified LDL and after supplementation of the d >1.063 kg/L fraction. Hence, factors in plasma other than albumin may favor LDL precipitation, at least by DS. To verify that the polyanion to LDL ratio also governs the accuracy of the precipitation methods, we divided the polyanion concentrations in half by diluting the heparin, DS, and PVS reagents with citrate buffer (64 mmol/L, pH 5.04), doubly distilled water, and Na2EDTA (5 mmol/L, pH 4.2), respectively. Whereas the heparmn method was virtually unaffected by the change in polyanion concentration, decreasing the DS and PVS considerably enhanced the precipitation of purified LDL (Figure 2). Coprecipitation of VLDL was assessed by adding various amounts of VLDL to a pool of normolipidemic samples with known VLDL-C content (final concentrations 0.2-2.0 mmol/L). Within this range roughly two-thirds of the added VLDL was precipitated with LDL by heparin and PVS, whereas DS coprecipitated only 28%. The percentage of coprecipitated VLDL was independent of the initial VLDL-C concentration. Experiments in which sera with low concentrations of HDL were enriched in HDL disclosed that none of the reagents discernibly precipitated HDL. Figure 3 summarizes the effect of FFA on the polyanion

-S
7
-J 0

E E
0 a)
U)

5 4

a) 0 U

3 2

a,
cc
C a) C.

0 0 2 4 6 Non-esterified fatty acids [mmol/L]

Fig. 3. Influence of free fatty acids on LDL precipitation (#{149}, 0) heparin, (A, ) DS, (#{149}, 0) PVS, and (*) phosphotungstate/M9CI2. Open and closed symbols distinguish twodifferentexperiments precipitation of LDL. Evidently, FFA concentrations exceeding 1-1.5 mmol/L decreased the amounts of LDL precipitable with DS and PVS, almost complete inhibition being attained at FFA concentrations of about 4 mmol/L. Although a discernible influence was still present, the interference by FFA with the heparin method was far less pronounced than that observed with DS or PVS. Finally, FFA do not at all affect the PhT/MgCl2 precipitation of apolipoprotein B-containing lipoproteins. We determined LDL-C in 113 samples from ostensibly healthy donors with each precipitation method and our combined ultracentrifugation and precipitation method (11). Tables 1 and 2 summarize the results. Both means and regression lines indicate that the precipitation assays underestimate LDL-C. Given the assumption that this was attributable to incomplete precipitation of LDL, we corrected raw LDL-C scores for concentration-dependent recovery, and recalculated the regression lines (bottom half, Tables 1 and 2). For this purpose we estimated k1 and k2 from the curves shown in the insert to Figure 1. Because of the low VLDL content in these samples, we disregarded the coprecipitation of VLDL. This transformation did, in fact, eliminate or at least considerably decrease the negative intercepts of the original regression lines.

100
C

Table 1. LDL-Cholesterol as Determined by Polyanlon Precipitation and Combined Ultracentrlfugation/

Preclpitationa
LDL, mmol/L

80
C.

.60 Raw scores 40

0 C., -J -J

Mean LDL-Cuc,phT LDL-CHEP 3.23

SD

MedIan 68%#{176} 2.93 0.68

0.98
1.10 0.87
1.04 0.77 0.82

2.54 2.81
2.90

2.33
2.35 2.67

0.84
0.68

20
0

LDL-CDS LDL-C5 Corrected for LDL recovery 0

2 4 LDL-cholesterol 6 [mmol/L] 8 LDL-CHEP LDL-CDS

2.45 0.82

0.67
0.82 0.65

2.71
2.69

LDL-C5
n
a
=

2.78 3.22

3.09 0.64

Fig. 2. Effect of polyanion concentrations on percentages of LDL-C precipitated (U) hepann, (A) OS, and (#{149}) PVS. Broken lines indicate reagents modified by diluting the polyanion compound twofold with the appropriate buffer

113 samples with triglycerides < 1.71 mmol/L Part of the data are reproduced with permission of Clinics Chimica Acta

(11). b 68th percentile of distances from the median.

CLINICAL CHEMISTRY, Vol. 36, No. 12, 1990

2111

Table 2. Regression Parameters for LDL-C Determinations by Various Procedures

Mean (and upper/lower lImits, 95% confIdence Interval)
Methods

Slope 1.16 (i.06/i.27)t 1.01 (0.91/1.10) 0.99 (0.92/1.06)

Intercept

Raw scores LDL-CHEP vs UC LDL-C05vs UC LDL-C5 vs UC

-1.09 (-1 .44/_0.79)c -0.71 (_0.96/_0.43)6 -0.34 (_0.55/_0.14)c

(_0.86/_0.23)C (-0.46/0.06) (-0.03/0.44) 113 samples with

Corrected for LOL recovery LDL-CHEP vs UC 1.09 (1.00/1 20)b -0.51 LDL-CDS VS UC -0.22 0.95 (0.86/1.04) LDL-C5 vs UC 0.25 0.94 (0.87/1.01) Slopes and intercepts have been calculated for n
=

triglycerides < 1.71 mmol/L. a Part of the data are reproduced with permission of Clinics Chimica Acta
(11).
b
C

Same samples as in Table 1. Slope significantly different from unity (P <0.05). Intercept significantly different from zero (P <0.05).

In keeping with earlier reports (8-10), we found that neither polyanion reagent coprecipitated H])L. However, all reagents significantly captured VLDL, which is at variance with the results of other investigators (8-10). At present, there is no final explanation for this inconsistency, but coprecipitation of VLDL, e.g., with the heparin reagent, may be due to the well-documented interaction of heparin with the apolipoprotein B or apolipoprotein E moiety (or both) of triglyceride-rich lipoproteins (19,20,25,26), which may occur even in the absence of divalent cations (16-18, 22). Interferences from FFA have recently been pointed out for the DS reagent (9, 18), but no systematic investigations had been available for the heparin and PVS method. Here we have shown that FFA concentrations above a threshold of 1-1.5 mmolfL significantly influence precipitation with DS and PVS; a moderate effect was seen for the heparin method, and none for the phosphotungstate/MgC12 reagent. In conclusion, none of these methods for the direct determination of LDL proved specific for this analyte nor were they accurate in terms of analytical recovery. The formation of insoluble lipoprotein-polyanion complexes depends on various conditions: ionic strength, pH, sulfation grade and molecular mass of the polyanion, and stoichiometric relationships between lipoproteins and polyanion (9, 16, 19, 20). Furthermore, this formation may be modified by the presence of other macromolecular matrix constituents. Perhaps modifications of the present precipitation methods for LDL with respect to these factors will bring sufficient improvements to justify their application in the clinical laboratory.
We thank Mrs. Sabine Cezanne, Mrs. Bettina Donnerhak, and Mrs. Christine von Hayn for their excellent technical assistance. Part of this study was supported by grants from the Riese-Stiftung and the Scheidel Stiftung, Frankfurt am Main.

Discussion
Interactions between lipoproteins and sulfated polyanions in the presence or absence of divalent cations have been known for many years and their nature has comprehensively been characterized by several investigators (1520). Making use of these interactions has allowed the development of precipitation methods for quantitative and selective precipitation of LDL (8-10). Ample agreement among the precipitation assays for LDL-C (21-23) as well as between LDL-C after precipitation and composite methods similar to ours (8-JO) has been reported. By contrast, we have recently observed that procedures for the precipitation of LDL underestimate LDL-C (11, cf. Tables 1 and 2). Regression lines relating the precipitation methods to an established method showed pronounced negative intercepts, which became even more severe when hypertriglyceridemic samples were included (11). Here we have demonstrated that this is caused by incomplete precipitation of LDL within the concentration range found in healthy subjects. Consistently, adjustment of the raw data for incomplete recovery eliminated the negative intercepts from the regression lines. As could be demonstrated for low and normal concentrations of LDL, reducing the PVS and DS concentrations increased the percentage of LDL precipitated, whereas dilution of heparin had no effect. Along with the concentration-dependent behavior of LDL recoveries, this suggests that the assay accuracies, at least of the DS and PVS procedures, are subject to a stoichiometric relationship between LDL and polyanion compound. In LDL solutions the LDL-CDs and LDL-CPVS values were lower than in mixtures of LDL with the ultracentrifugal d >1.063 kgfL bottom fraction. Addition of albumin to an LDL preparation improved percentage of LDL precipitated with PVS, but recoveries with DS remained low. This underscores that the matrix in which precipitations are performed may be of crucial importance and that plasma proteins and (or) other plasma constituents may exert adjuvant effects on the LDL-polyanion interaction. Our findings are in strong disagreement with observations (24) suggesting that, in general, polyanion precipitations of lipoproteins are more readily attained with purified lipoproteins than in the presence of plasma proteins.
2112 CLINICAL CHEMISTRY, Vol. 36, No. 12, 1990

References
1. Levy RI. Cholesterol, lipoproteins, apolipoproteins, and heart disease: present status and future prospects. Clin Chem 1981;27:653-62. 2. Gordon T, Kannel WB, Castelli WB, Dawber TR. Lipoproteins, cardiovascular disease and death. The Framingham study. Arch Intern Med 1981;141:ll28-31. 3. Steinberg D. Lipoproteins and atherosclerosis. A lookback and a look ahead. Arteriosclerosis 1983;3:283-301. 4. Friedewald WT, Levy RI, Fredrickson DS. Estimation of the concentration of low-density lipoprotein cholesterol in plasma without use of the preparative ultracentrifuge. Clin Chem 1972;13:499-502. 5. Drager B, Wahlefeld AW, Ziegenhorn J. A practical method for HDL cholesterol quantitation. Lab Med 1982;6:198-202. 6. Warrnck R, Nguyen T, Albers AA. Comparison of improved precipitation methods for quantification of high-density lipoprotein cholesterol. Clin Chem 1985;31:217-22. 7. MSrz W, Gross W. Analysis of plasma lipoproteins by ultracentrifugation in a new fixed angle rotor: evaluation of a phosphotungstic acidIMgCl2 precipitation and a quantitative lipoprotein electrophoresis assay. Clin Chim Acts 1986;160:l-18.

8. Wieland H, Seidel D. A of low density lipoproteins. 9. Armstrong VW, Seidel the determination of LDL

simple specific method for precipitation J Lipid Res 1983;24:904-9. D. Evaluation of a commercial kit for cholesterol in serum based on precipi-

tation of LDL with dextran sulfate. Arztl Lab 1985;31:325-30. 10. Assmann G, Jabs HU, Kohnert U, Nolte W, Schriewer H. LDL cholesterol determination in blood serum following precipitation of LDL with polyvinylsulfate. Clin Chim Acts 1984;140:77-83. 11. Siekmeier R, Mfirz W, Gross W. Precipitation of LDL with sulfated polyanions: three methods compared. Clin Chim Acts
1988;177:221-30.

12. Spector AA, Hoak JC. An improvedmethodfor the additionof long-chain free fatty acid to protein solutions. Anal Biochem 1969;32:297-302. 13. Passing H, Bablock W. A new biometrical procedure for testing the equality of measurements from two different analytical methods.J Clin Chem Clin Biochem 1983;21:709-20. 14. McIntosh JEA, McIntosh RP. Mathematical modelling and computers in endocrinology. Berlin-Heidelberg-New York: Springer, 1980. 15. Bernfield P, Nisselbaum JS, Berkeley BJ, Hanson RW. The influence of chemical and physicochemical nature of macromolecular polyanions on their interaction with human serum (3-lipoproteins. J Biol Chem 1960;235:2852-9. 16. Cornwell DG, Kruger FA. Molecular complexes in the isolation and characterization of plasma lipoproteins. J Lipid Res
1961;2:110-34. 17. Burstein M, Scholnick HR. Lipoprotein-polyanion-metal interactions. In: Paoletti R, Kritchevsky D, eds. Advances in lipid research. Vol. 11. New York: Raven Press, 1973:67-108. 18. Nishida T. Effect of phospholipase A treatment of low density lipoproteins on the dextran sulfate-lipoprotein interaction. J Lipid Res 1968;9:627-35. 19. Jackson RL, Socorro L, Fletcher GL, Cardin AD. Heparin binding to lipoprotein lipase and low density lipoproteins. FEBS

Lett 1985;190:297-300. 20. Shelburne FA, Quarfordt SH. The interaction of heparin with an apoproteinofhuman very low density lipoprotein. J Clin Invest 1977;60:944-50. 21. Demacker PN, Humans AG, Brenninkmejjer BJ, Jansen AP, vant Laar A. Five methodsfor determining low-densitylipoprotein cholesterolcompared. Clin Chem 1984;30:1797-800. 22. Hoffman GE, Hiefinger R, Weiss L, Poppa W. Five methodsfor measuring low-density lipoprotein cholesterol concentration in serum compared. Clin Chem 1985;31:1729-30. 23. Mertz DP, Thuilot G. Einfaches Verfahren zur quantitativen Bestimmung von LDL-Cholesterin. J Clin Chem Clin Biochem 1986;24:355-7. 24. Bachorik PS, Albers JJ. Precipitation methods for quantification of lipoproteins.In: Albers JJ, Segrest JP, eds.Plasma lipoproteins. Part B. Characterization, cell biology, and metabolism. Methods Enzymol 1986;129:78-100. 25. Cardin AD, Nobuyoshi H, Blankenship DT, et al. Binding of a high reactive heparin to human apolipoproteinE: identificationof two heparin-binding domains. Biochem Biophys Res Commun 1986;134:783-9. 26. Iverius PH. The interaction between human plasma lipoproteins and connective tissue glycosaminoglycans. J Biol Chem
1972;247:2607-13.

CLIN. CHEM. 36/12,2113-2117(1990)

Determination of Carrier Status in Duchenne and Becker Muscular Dystrophies by Quantitative Polymerase Chain Reaction and Allele-Specific Oligonucleotides
ThomasW. Prior,1AudreyC. Papp,1PamelaJ. Snyder,1 W. EdwardHlghsmfth, Jr.,2KennethJ. Friedman,2 Tenly R. Perry,2 Lawrence N. Silverman,2 and Jerry R. MendeIl3
Detection of carriers of Duchenne muscular dystrophy (DM0) affects -1 of 3500 to 5000 newborn males (1). Both DMD and the milder allelic Becker muscular dystrophy (BMI)) and Becker muscular dystrophy (BMD), in the deletion cases, result from mutations in the dystrophin gene. The 14-kb involves calculating gene dosage from Southern blots. We coding sequence of the dystrophin gene has been cloned (2). show that the analysis of dosage can also be made from the Use of cDNA probes has shown that molecular deletions polymerase chain reaction (PCR) withuse ofallele-specific account for -65% of the mutations in DM1) and BMD (3-6). oligonucleotides (ASO5). The deletion-prone exons are amCarrier detection in the deletion cases involves assessing plified, transferred to a membrane, and hybridized with ASOs gene dosage by quantitative Southern-blot analysis, complementary to the exons; the autoradiographic bands are whereby one determines whether the female at risk exhibthen quantified with a densitometer. After determining the its no reduction (noncarrier status) or 50% reduction (carquantitative conditions of the amplification reaction, we were rier status) in hybridization intensity in those bands that able to identify deletions in a DMD/BMD carrier female. The are deleted in the affected male (7). The dosage determinadetermination of carrier status via PCR removes several of tions permit direct analysis for carrier status and eliminate the technical limitations of Southern analysis and is also costthe inherent problems of the restriction-fragment-lengthand labor-effective. polymorphism technique (e.g., recombinations, noninformative meioses, unavailability of family members, and sporadic mutations). To further increase the accuracy of the AdditionalKeyphrases:gene probes heritable disorders dosage analysis, we quantify the autoradiographic bands by scanning with a densitometer (8). Duchenne muscular dystrophy (DMD) is an X-linked Although dosage analysis has significantly improved carrier studies, particularly in the isolated cases of the disease, there are technical limitations. Dosage analysis of recessiveprogressivemuscle-wastinggenetic disorder that Southern blots requires optimal conditions; very good qual1 Division of Molecular Pathology, Department of Pathology, ity blots are necessary, with uniform transfer and hybridand3 Department of Neurology,Ohio State University, North 305 Doan Hall, 410 West 10th Ave., Columbus,OH 43210. 2Division of Molecular Pathology, Departments of Laboratory Medicine and Pathology, North Carolina Memorial Hospital and 4Nonstandard abbreviations:DMD, Duchenne muscular dystroUniversity of North Carolina, 1071 Patient Support Tower, Chapel phy; BMD, Becker muscular dystrophy; PCR, polymerase chain reaction; ASO, allele-specific oligonucleotide; and SDS, sodium Hill, NC 27599. Received July 23, 1990; accepted September 27, 1990. dodecyl sulfate.
.

CLINICAL CHEMISTRY, Vol.36, No. 12, 1990 2113