409

Apolipoprotein B and a Second Major GeneLocus Contribute to Phenotypic Variation ofSpontaneous Hypercholesterolemia in Pigs

Robert J. Aiello, David N. Nevin, David L. Ebert, Patricia J. Uelmen, Mary E. Kaiser,Jean W. MacCluer, John Blangero, Thomas D. Dyer, Alan D. Attie

Abstract The Lpb5 apolipoprotein B (apoB) allele occurs inpigs with spontaneous hypercholesterolemia. Low-density lipo-protein (LDL) from these pigs binds to the LDL receptor witha lower affinity and is cleared from the circulation more slowlythan control pig LDL. However, the severity of hypercholester-olemia in pigs with the mutant apoB allele is highly variable.This study aimed to determine the metabolic basis for thephenotypic heterogeneity among Lpb5 pigs. Lpb5 pigs weredivided into two groups: those with plasma cholesterol greaterthan 180 mg/dL (Lpb5.1) and those with plasma cholesterol lessthan 180 mg/dL (Lpb5.2). LDL from both Lpb5.1 and Lpb5.2pigs was catabolized in vivo and in vitro at a similarly reducedrate. The difference in plasma cholesterol between the twophenotypic groups was in part due to a higher buoyant LDL

Population genetic studies have demonstrated thatheredity contributes significantly to hyperlipid-emia. In some instances, particular hyperlipid-

emic syndromes have been shown to be caused bysingle-gene mutations. However, the frequency of theknown single-gene mutations is still far below the fre-quency of hyperlipidemia. The population studies there-fore predict that multiple gene loci are involved and thatin some instances there may be interactions betweengene loci to produce a clinical syndrome.

Apolipoprotein B (apoB) may be one such gene locus.ApoB is the primary protein component of low-densitylipoprotein (LDL), which carries the bulk of cholesterolin human plasma, and is a ligand for the LDL receptor;thus, apoB functions in the assembly of the LDLparticle and mediates its clearance from the blood.Given the central role of apoB in LDL metabolism,apoB mutations alone or in combination with otherdefects probably contribute significantly to hypercholes-terolemia in the human population. One recently dis-covered apoB point mutation (Argjjoo—>Gln) results inLDL particles with reduced LDL receptor bindingactivity and in some patients is associated with a re-duced rate of LDL clearance from the plasma.1 Thismutation is relatively common in the human population,

Received July 19, 1993; revision accepted December 16, 1993.From the Departments of Biochemistry and Comparative Bio-

sciences, University of Wisconsin-Madison (RJ.A., D.N.N.,D.L.E., PJ.U., M.E.K., A.D.A.) and the Department of Genetics,Southwest Foundation for Biomedical Research, San Antonio,Tex (J.W.M., J.B., T.D.D.).

Correspondence to Dr Alan D. Attie, Department of Biochem-istry, University of Wisconsin, 420 Henry Mall, Madison, Wl53706.

production rate in Lpb5.1 pigs than in Lpb5.2 pigs. The in vivoLDL receptor status was evaluated by measuring the catabolismof LDL chemically modified to abrogate LDL receptor binding.Approximately 50% of LDL clearance in normal and Lpb5.2pigs was via the LDL receptor; in Lpb5.1 pigs, 100% of LDLclearance was LDL receptor independent. Quantitative pedi-gree analysis of the segregation of the plasma cholesterolphenotype suggested that two major gene loci (the apoB locusand a second apparently unlinked locus) contribute to thedetermination of plasma cholesterol levels in this pig popula-tion. (Arterioscler Thromb. 1994;14:409-419.)

Key Words • LDL • mutation • cholesterol • genetics• hyperlipidemia • animal models • LDL receptor

suggesting that clinically relevant mutations in apoBmay be abundant.

Hypercholesterolemia is also highly heritable in pigs.Rapacz et al2 described a strain of pigs bearing animmunogenetically defined lipoprotein-associatedmarker, designated Lpb5,3 and having marked hyper-cholesterolemia and premature atherosclerosis. Theantiserum used to identify the Lpb5 pigs was shown toreact only with apoB-100,2 and the apoB gene fromLpb5 pigs has a unique 283-bp insertion at intron 28.46

In addition, Lpb5 pigs have LDL with defective bindingto the LDL receptor7 and delayed plasma clearance.8

These observations strongly suggest an association be-tween apoB polymorphisms and hypercholesterolemiain pigs. However, pigs with the Lptr allele of apoB arephenotypically heterogeneous; the more severe pheno-type (designated Lpb5.1) involves twofold to fourfoldincreases in plasma cholesterol, whereas a second phe-notypic group (designated Lpb5.2) has only a moderateplasma cholesterol elevation.9

In this study, we further characterize pigs with theLpb5.1 and Lpb5.2 phenotypes to clarify the metaboliccontributions to variation in plasma cholesterol in pigs.In addition, quantitative pedigree analysis was appliedto suggest a genetic model consistent with the segrega-tion pattern of the plasma cholesterol phenotype.

MethodsDescription of the Pig Population

The study population consisted of pigs representing theLpb5.1, Lpb5.2, and normocholesterolemic phenotypes. Thefounders of this population carrying the Lpb5 allele of apoBwere donated by Dr Jan Rapacz from the University ofWisconsin Immunogenetic Herd. Non-Lpb5 founder pigs were

by guest on June 10, 2018http://atvb.ahajournals.org/

Dow

nloaded from

410 Arteriosclerosis and Thrombosis Vol 14, No 3 March 1994

obtained from the University of Wisconsin Arlington SwineFarm. The pigs were housed at the University of Wisconsin-Madison and maintained on a 0% cholesterol, 5% fat diet(University of Wisconsin Gestation Diet).

The pedigree analyzed in this study included 22 sires, 27dams, and 320 offspring. Blood samples were obtained from alloffspring at 3 to 4 months of age (6 weeks after weaning) fordetermination of plasma cholesterol levels and isolation ofgenomic DNA for apoB genotype determination as describedbelow. In all, plasma cholesterol determinations from a total of337 pigs (91 Lpb5.1 pigs, 160 Lpb5.2 pigs, and 86 non-Lpb5pigs) were included in the pedigree-phenotype dataset.

Previously, Lpb5 pigs were immunologically identified by thepresence of an LDL-borne allotype, designated Lpb5.2 Theseanimals are now identified using the polymerase chain reactionto amplify the target genomic DNA and detect a 283-bpinsertion at intron 28 in the apoB gene unique to the Lpb5

allele.4-5 Lpb5.1 pigs have the apoB intron 28 insertion andsevere hypercholesterolemia (plasma cholesterol >180 mg/dL). Lpb5.2 pigs have the apoB intron 28 insertion andmoderately elevated plasma cholesterol (80 to 130 mg/dL).Non-Lpb5 pigs lack the apoB intron 28 insertion and haveplasma cholesterol levels of 60 to 115 mg/dL.

Lipoprotein PreparationBlood was collected ascptically from overnight-fasted ani-

mals into a phosphate-buffered saline (PBS)/EDTA (1mmol/L) cocktail as described previously.8 Very-low-densitylipoprotein (VLDL, d< 1.006 g/mL), intermediate-density li-poprotein (IDL, d=1.006 to 1.019 g/mL), and LDL (d=1.019to 1.063 g/mL) were isolated and purified as described previ-ously." For some experiments, LDL was subfractionated andthe fractions were pooled into two density ranges: d= 1.019 to1.038 g/mL (buoyant LDL) and d= 1.038 to 1.063 g/mL (denseLDL).10 The density cutoff at 1.038 g/mL was used for allanimal LDL separations and represents the most commondensity at which there was a trough in the mutant LDL densityprofile.8 The protein concentration of each subspecies wasdetermined by the modified Lowry method of Markwell et al.1'Cholesterol concentrations in whole plasma and lipoproteinfractions were determined by enzymatic assay (Sigma Diag-nostic Kit No. 351, Sigma Chemical Co).

Lipoprotein IodinationLDL and LDL subspecies were iodinated with 125I (Du Pont)

or 131I (ICN) as described previously.10 Each preparation wasassayed for proper relative electrophoretic mobility and lack ofbacterial contamination. LDL specific activities were 50 to 450cpm/ng protein. Greater than 95% of the radioactivity in theLDL preparations was precipitable with trichloroacetic acid(TCA; final concentration, 10%), and less than 2% of theradioactivity was chloroform soluble. ApoE was not detected inany of the LDL preparations by sodium dodecyl sulfate-poly-acrylamide gel electrophoresis. All iodinated LDL preparationsto be used for in vivo studies were injected into animals 2 daysafter iodination. An aliquot of each iodinated lipoprotein prep-aration was again subjected to density gradient ultracentrifuga-tion to determine the purity of the injected doses.

LDL Lysine Modifications to Block LDLReceptor Binding

Glucose modification of apoB lysine residues was achievedby incubating LDL (3 to 5 mg/mL) under sterile conditions for5 days at 37°C in PBS, pH 7.4, containing 80 mmol/L glucoseand sodium cyanoborohydride (final concentration, 40mmol/L) as described by Kesaniemi et al.12 The sodiumcyanoborohydride was added as a fresh sterile solution eachday. This method allows incorporation of 40 to 50 nmol ofglucose per milligram of LDL protein, as determined bymeasuring incorporation of [t/^CJglucose into LDL protein(data not shown). Unreacted glucose and reducing agent were

E5,U

350

300

250

200

150

100-

50-

0 i

I Control• Lpb 5.20 Lpb 5.1

Plasma VLDL IDL Buoyant Dense HDLLDL LDL

FIG 1. Bar graph shows plasma and lipoprotein cholesterolconcentrations In Lpb5 and control pigs. Very-low-density lipo-protein (VLDL), intermediate-density lipoprotein (IDL), low-den-sity lipoprotein (LDL), buoyant LDL (d=1.019 to 1.038 g/mL),dense LDL (d= 1.038 to 1.063 g/mL), and high-density lipopro-tein (HDL) were separated by density gradient ultracentrrfuga-tlon. Values are means from control (pig Nos. 8, 9, 10, 11),Lpb5.2 (pig Nos. 15,16,17,18), and Lpb5.1 (pig Nos. 29, 30, 31,32) pigs.

removed by gel filtration on a Sephadex G25 column, followedby dialysis against PBS/EDTA (1 mmol/L). The glucosylatedLDL had enhanced mobility when subjected to agarose gelelectrophoresis (Paragon Lipogel, Beckman) and did notcompete with native LDL for binding to human skin fibroblasts(data not shown).

Fractional Clearance Rates of Lpb5.1 and Lpb5.2Unfractionated LDL

Pigs used for in vivo turnover studies were 12 to 20 weeks oldand weighed 30 to 60 kg. Iodinated LDLs ("unfractionated,"d= 1.019 to 1.063 g/mL, 3 ftCi/kg body wt) were injected intononanesthetized pigs via vena caval puncture. Each animal wasinjected simultaneously with two different LDL preparations(iodinated with either 125I or 131I) isolated from either control,Lpb5.1, or Lpb5.2 pigs (see figure legends for the exact LDLpreparations used for each experiment). The animals werefasted 12 hours before the start of each experiment, fed 6hours after the initial injection, and fed twice daily thereafter.During the 96-hour experimental period, 15 6-mL sampleswere collected from each animal via vena caval puncture intoVacutainer tubes containing EDTA (final concentration, 1.5mg/mL). Plasma samples were precipitated in TCA (10%wt/vol, 4°C), and pellet radioactivities were determined in amodel 5001 gamma counter (Packard).

The plasma disappearance curves for each unfractionatedLDL preparation were analyzed by a two-pool model de-scribed by Matthews,13 and kinetic parameters were estimatedusing the nonlinear least-squares curve-fitting program KINETIC(as modified by G.A. McPherson, Elsevier-BIOSOFT) as de-scribed previously.10 This program uses a multiexponential algo-rithm to calculate the slopes and intercepts of each exponentialcomponent. The inverse of the area of the biexponential fit wasused to estimate the fractional clearance rate, and the F test wasused to determine significant differences between each pair ofcurves generated from each animal.14

Transport Rates Between LDL SubspeciesIodinated autologous LDL subspecies (eg, 123I-buoyant

LDL and l31I-dense LDL) were simultaneously injected intononanesthetized Lpb5.2 pigs, and plasma samples were col-lected as described above. Plasma samples were dialyzedovernight at 4°C against an NaBr solution (d=1.04 g/mL)

by guest on June 10, 2018http://atvb.ahajournals.org/

Dow

nloaded from

Aiello et al Phenotypic Heterogeneity of Spontaneous Hypercholesterolemia in Pigs 411

TABLE 1. Genotypes and Phenotypes of Pigs Used In Metabolic Studies

Pig

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

26

27

28

29

30

31

32

Lpb

Non-5

Non-5

8/8

8/8

Non-5

Non-5

4/4

Non-5

Non-5

Non-5

Non-5

5/5

5/5

5/5

5/5

5/5

5/5

5/5

5/5

5/5

5/5

5/5

5/5

5/5

5/5

5/5

5/5

5/5

5/5

5/5

5/5

5/5

LDL indicates low-densityd= 1.038 to 1.063 g/mL.

Phenotype

Control

Control

Control

Control

Control

Control

Control

Control

Control

Control

Control

Lpb5.2

Lpb5.2

Lpb5.2

Lpb5.2

Lpb5.2

Lpb5.2

Lpb5.2

Lpb5.2

Lpb5.2

Lpb5.2

Lpb5.2

Lpb5.2

Lpb5.2

Lpb5.2

Lpb5.1

Lpb5.1

Lpb5.1

Lpb5.1

Lpb5.1

Lpb5.1

Lpb5.1

lipoprotein. For

PlasmaCholesterol,

mg/dL

104

83

87

83

73

79

85

95

70

72

89

121

115

136

110

116

93

85

95

94

130

117

112

135

110

230

220

260

204

315

362

380

buoyant LDL, d=1.019 to

LDL Protein, mg/dL

Buoyant

12.4

8.4

11.1

14.3

9.7

6.9

8.7

11.8

10.0

8.7

9.0

16.4

20.2

18.6

16.9

16.7

14.9

18.5

21.0

15.1

28.8

18.9

16.7

28.2

20.0

44.6

46.5

49.5

41.5

51.7

71.0

82.3

1.038 g/mL; for

Dense

18.1

20.2

20.0

21.3

31.4

20.6

15.1

17.7

21.0

15.5

18.9

17.8

25.7

23.2

20.7

22.0

17.8

15.5

22.0

26.0

28.7

34.1

21.8

30.0

36.1

21.0

18.5

15.2

28.0

18.9

35.2

24.6

dense LDL,

containing 1 mmol/L EDTA. Four milliliters of plasma was thenlayered between equal volumes of NaBr (d=1.026 and d=1.054g/mL)/EDTA (1 mmol/L) solutions. Each plasma sample wascentrifuged in a Beckman SW 41 rotor (72 hours, 37 000 rpm,15°C) and fractionated into 20 0.56-mL fractions as describedpreviously.10 The density of each fraction was estimated from therefractive index of identical fractions isolated from a blank NaBrgradient run in parallel. Bovine serum albumin (BSA; finalconcentration, 100 fig/mL) was added to each fraction andTCA-precipitable radioactivity determined. Those fractions witha density of 1.019 to 1.038 g/mL (buoyant LDL) were pooled, aswere those fractions with a density of 1.038 to 1.063 g/mL (denseLDL). In experiments in which autologous LDL subspecies wereinjected, specific activity curves were analyzed by a two-accessi-ble-pool noncompartmental model developed by Gurpide et al13

and discussed by Cobelli and Toffolo.16 This model focuses onlyon the physically accessible pools (ie, buoyant and dense LDL)and assumes that there is an undetermined number of compart-ments within each pool. This model has been described previ-ously and used for pig buoyant and dense LDL subspecies10 andprovides estimates for rates of irreversible removal from thebuoyant and dense pools as well as rates of interconversionbetween the two pools. The specific differential equations used toobtain the kinetic parameters have been reported previously.10

The equations for the buoyant and dense LDL specific activitycurves were determined by nonlinear regression using the SAAMmodeling program,17 and areas under these curves were calcu-lated by integration. The same methodology was used for theLpb5.2 pigs in this study as was used for Lpb5.1 and non-Lpb5pigs examined previously,10 allowing comparisons of kinetic pa-rameters among these three phenotypic groups.

by guest on June 10, 2018http://atvb.ahajournals.org/

Dow

nloaded from

412 Arteriosclerosis and Thrombosis Vol 14, No 3 March 1994

II

20

10

| Control

• Lpb5.20Lpb5.1

1.022 1.032 1.042 1.052 1.062

Density (g/ml)

FIG 2. Graph shows representative distributions of low-densitylipoprotein (LDL) protein after density gradient ultracentrifuga-tion of control (No. 8), Lpb5.2 (No. 17), and Lpb5.1 (No. 30) pigplasma.

Competition and Degradation ExperimentsNormal human fibroblasts (TJ-6F) were kindly provided by

Dr Lynn Allen-Hoffman (University of Wisconsin-Madison).Pig skin fibroblasts were obtained from a control pig by punchbiopsy from the hind thigh. The cells were maintained on

^ 400-

•og 300-2 a.

I*j ° 200-

— • — Control LDL

— o — 5.1 LDL

— • — 5.2 LDLy

4

Time (hr)

100H

75-i

J* 50-

25-

0-

— • — Control LDL

— o — 5.1 LDL

— • — 5.2 LDL

vv--^•g$^__n-6

10 15 20

LDL (ng/ml)

FIG 3. Top, Une graph shows degradation of control, Lpb5.1,and Lpb5.2 low-density lipoprotein (LDL). Pig fibroblasts weregrown in Dulbecco's modified Eagle medium (DMEM)+ 10%lipoprotein-defident serum (LDS) for 48 hours before experi-ments. Medium was changed to DMEM+10% LDS containing2.0 /ig/mL of each 125I-LDL, and cells were Incubated at 37 °C forIndicated times. Precipitation and quantitation were performedas described in "Methods." Bottom, Line graph shows compe-tition for binding of control, Lpb5.1, and Lpb5.2 LDL to humanfibroblasts. Cells were prepared similarty and incubations carriedout at 4°C in DMEM+25 mmol/L HEPES+5 mg/mL bovineserum albumin containing 2.0 M9/mL control 125I-LDL plus vary-ing concentrations of unlabeled competing LDLs. Bound LDLwas released by dextran sulfate and counted; results are ex-pressed as percent of maximum bound.

Dulbecco's modified Eagle medium (DMEM; GIBCO) sup-plemented with 10% fetal bovine serum (HyClone), 100 U/mLpenicillin G, and 100 fig/mL streptomycin (P/S, GIBCO) asdescribed previously.17 Cells were grown to subconfluence inCorning six-well plates and washed with PBS, and the mediumwas replaced with 2 mL DMEM containing 5% lipoprotein-deficient serum plus P/S 48 hours before experiments. Forcompetition experiments, fibroblasts were incubated in thepresence of l25I-labeled LDL (2 jig/mL) and increasing con-centrations of unlabeled LDL in DMEM, 25 mmol/L HEPES,and BSA (5 mg/mL), pH 7.4, for 4 hours at 4°C on a rotaryshaker. Cell surface-bound LDL was released with dextransulfate and quantified according to the method of Goldstein etal18 as described previously.7 Data were analyzed using theEBDA program (Elsevier-BIOSOFT) for determination of theconcentration at which 50% of radiolabeled LDL binding wasinhibited (=IQo). For degradation experiments, cells wereincubated at 37°C with lz3I-labeled control LDL. Proteolyticdegradation was quantified as described7 except that silvernitrate was used to precipitate free iodide. The cells werewashed and dissolved in 2N NaOH, and cell-associated LDLwas measured by gamma counting.

Quantitative Pedigree Analysis ofPlasma Cholesterol

Complex segregation analysis was performed on plasmacholesterol levels using a modification of the Pedigree AnalysisPackage19 as previously described.20 Four models were evalu-ated (Table 5). Before analysis, data were corrected for age.There was an age effect in males only, with a decrease in totalcholesterol of 2 mg/dL per month. Segregation analysis resultstherefore were standardized to age 6 months. Cholesterollevels are 37 mg/dL higher in nursing animals. In all segrega-tion analysis models, cholesterol levels were allowed to beinfluenced by sex, genotype at the apoB locus, multiple geneswith small effects (polygenes), and random environmentalfactors. The simplest model (last line of Table 5) assumed thatall variation in cholesterol levels was attributable to thesefactors alone. The remaining three models all included aneffect of a major gene in addition to the apoB locus butdiffered in their assumptions about the linkage of the majorgene to the apoB locus. In one, the loci were constrained to beunlinked (recombination fraction 9=0.5); in the second, theloci were assumed to be very tightly linked or identical (6=0);and in the third, 6 was estimated.

For each model, we estimated the parameter values thatmaximized the likelihood of observing the pedigree data.Parameters estimated included allele frequencies at the majorlocus and the apoB locus, parameters describing the meansand standard deviations of the cholesterol distributions corre-sponding to each genotype, the polygenic heritability h2 (theproportion of the variance around each genotypic mean that isattributable to polygenes), gametic disequilibrium (the degreeof association between specific alleles at the apoB locus andthe second major locus), and a sex effect. Of the four modelsevaluated, the most general is model 1, the two-locus modelwith 8 estimated. Other models were compared with this oneby using likelihood ratio statistics to compute A; (where idenotes degrees of freedom), obtained as minus twice thedifference between the log, likelihoods of the models. Thisdifference is approximately distributed as x2 w ' m degrees offreedom equal to the difference in the numbers of parametersestimated in the two models. The best model is the onerequiring the fewest estimated parameters and exhibiting alikelihood that is not significantly smaller than that for model 1.

ResultsPhenotypes of Lpb5.1 and Lpb5.2 Pigs

In this study, total plasma cholesterol levels wereelevated by 200% to 300% in the Lpb5.1 pigs relative to

by guest on June 10, 2018http://atvb.ahajournals.org/

Dow

nloaded from

Aiello et al Phenotypic Heterogeneity of Spontaneous Hypercholesterolemia in Pigs 413

1001 100

10

£a.

20 40 60Time (hr)

80 100

— * — 5.2 LDL—o— 5.1 LDL

20 40 60Time (hr)

80 100

FK3 4. Line graphs show representative plasma disappearance curves of control and Lpb5.2 low-density Iipoprotein (LDL) in a controlpig (left) and Lpb5.2 and Lpb5.1 LDL in a control pig (right).

non-Lpb5 pigs, whereas the Lpb5.2 pigs displayed amean 30% increase in total plasma cholesterol (Fig 1and Table 1). VLDL was elevated twofold in bothLpb5.2 and Lpb5.1 pigs, but this made only a minorcontribution to total plasma cholesterol (Fig 1). In boththe Lpb5.1 and Lpb5.2 pigs, the primary contributor tothe plasma cholesterol increase was LDL. However, thedistribution of LDL cholesterol and protein within theLDL density range differed between the phenotypicgroups. In normal pigs, most LDL was in the dense LDLdensity range, whereas the Lpb5.1 pigs manifested theincrease in LDL predominantly in the lower-densitybuoyant LDL subpopulation (Table 1 and Fig 2). Thebuoyant LDL concentration in the Lpb5.2 pigs was asmuch as twofold higher than in the control pigs but only25% to 50% of the level found in Lpb5.1 pigs (Table 1).

Relative Affinities for the LDL ReceptorLDL from the hypercholesterolemic Lpb5.1 pigs has a

reduced affinity for the LDL receptor.7 Because theLpb5.2 pigs apparently have the same apoB allele asLpb5.1 pigs (see "Discussion"), we investigated thepossibility that Lpb5.2 LDL also has a reduced affinityfor the LDL receptor. As previously reported,7 Lpb5.1LDL was degraded more slowly than control LDL incultured fibroblasts. The Lpb5.2 LDL degradation ratewas similar to that of Lpb5.1 LDL (Fig 3, top). In vitro

competition binding experiments demonstrated thatLpb5.1 and Lpb5.2 LDLs were equally poor competitorsfor the LDL receptor on fibroblasts when comparedwith control pig LDL; IQo values for Lpb5.1 and Lpb5.2LDL were not significantly different from each other(Lpb5.1, 2.39±0.15 fig/mL; Lpb5.2, 2.83±1.01 fig/mL),whereas control pig LDL was a significantly bettercompetitor than either Lpb5.1 or Lpb5.2 LDL(IQo= 1-25±0.70 fig/mL; P<.O16) (values are the aver-age of at least four determinations per pig type; repre-sentative competition curves are shown in Fig 3, bot-tom). Thus, both Lpb5.2 and Lpb5.1 pigs have LDL witha similarly reduced affinity for the LDL receptor invitro.

Plasma Clearance of Lpb5.2 and Lpb5.1 LDLsPrevious experiments showed that LDL from Lpb5.1

pigs was cleared from the circulation more slowly thanLDL from non-Lpb5 pigs.8 To ascertain whether theLDL particles from Lpb5.2 pigs also had a decreasedplasma clearance rate, the in vivo catabolism of control,Lpb5.2, and Lpb5.1 LDL particles was compared. WhenLpb5.1 and Lpb5.2 LDLs were labeled with 131I and 125Iand simultaneously injected into control pigs, theirplasma clearance rates were essentially identical (Fig 4,right). When simultaneously injected into a control pig,Lpb5.2 LDL was cleared from the plasma approxi-

TABLE 2.

Recipient

Plasma

Pig

Clearance

Donor

of Lpb5.2 Low-Density Upoprotein in Control

Plasma Clearance (Pools/h)

Lpb5.2 LDL Control LDL

Pigs

FTest

2

3

4

5

6

12

12

13

13

14

14

0.032

0.032

0.034

0.037

0.040

0.023

0.047

0.047

0.054

0.050

0.053

0.039

P<.001

P<.001

P<.05

P<.001

P<.001

P<.05

LDL indicates low-density Iipoprotein. Each control pig (recipient) was injected simultaneously withautologous control LDL (d=1.019 to 1.063 g/mL) and LDL from an Lpb5.2 pig (donor). Fisher's F testwas used to determine significant differences between plasma disappearance curves for Lpb5.2 andcontrol LDL within each animal. Mean fractional clearance rate for Lpb5.2 LDL was significantlydifferent (P<.05) from the mean control LDL fractional clearance rate by paired Student's f test.

by guest on June 10, 2018http://atvb.ahajournals.org/

Dow

nloaded from

414 Arteriosclerosis and Thrombosis Vol 14, No 3 March 1994

TABLE 3. Plasma Clearance of Lpb5.2 Low-Density LJpoproteln In Lpb5.2 and Lpb5.1 Pigs

Donor Pig and Radlolabel

Recipient Pig

19

20

26

27

Genotype

5/5 (Lpb5.2)

5/5 (Lpb5.2)

5/5(Lpb5.1)

5/5(Lpb5.1)

PlasmaCholesterol,

mg/dL

95

94

230

220

19, 1 20, 111I 26 , 125I 27, 1J1I FTest

0.030

0.018

0.022

0.010

0.024

0.012

0.026

0.015

NS

NS

NS

NS

lodinated Lpb5.1 and Lpb5.2 low-density lipoproteins (LDLs, d=1.019 to 1.063 g/mL) were injected simultaneously into each pig.Rsher's F test was used to determine significant differences between plasma disappearance curves within each animal. Mean fractionalclearance rates for autologous Lpb5.2 and Lpb5.1 LDLs were significantly different (P<.05) as detennined by Student's f test.

mately 30% more slowly than control LDL (Fig 4, left,and Table 2).

LDL from four Lpb515 littermates, two Lpb5.2 and twoLpb5.1 pigs, was isolated and differentially iodinatedwith either 13II or 123I. Each of the four animals wassimultaneously injected with autologous LDL and LDLfrom a littermate of the opposite phenotype. In theLpb5.2 pigs, Lpb5.1 and Lpb5.2 LDLs were cleared atthe same rate. Similarly, in the Lpb5.1 pigs, the clear-

100

75 100

Time (hr)FIG 5. Line graphs show representative plasma disappearancecurves of native and modified (glucosyiated) low-density lipopro-tein (LDL) in a control pig (top), Lpb5.2 pig (middle), and Lpb5.1pig (bottom). Included Is the best line of fit. Pigs were simulta-neously injected with 12Sl-autologous native and 131l-modifiedLDL (d=1.019 to 1.063 g/mL). All pigs were injected with thesame preparation of control modified LDL.

ance rates of the two types of LDL were not significantlydifferent (Table 3).

This defective behavior of Lpb5.1 and Lpb5.2 LDL invivo paralleled the abnormal behavior of these particlesin their ability to compete for binding to the LDLreceptor in cultured fibroblasts. Thus, particle differ-ences between Lpb5.1 and Lpb5.2 LDL cannot accountfor the difference in plasma LDL levels between the twophenotypic groups.

In Vivo LDL Receptor-Mediated ClearanceAlthough Lpb5.1 and Lpb5.2 LDL particles behaved

identically in vivo and in vitro, Lpb5.1 pigs cleared LDLparticles (regardless of their source) more slowly thandid Lpb5.2 pigs (Table 3). Moreover, whereas normaland Lpb5.2 pigs cleared Lpb5.1 and Lpb5.2 LDL moreslowly than normal LDL, Lpb5.1 pigs cleared LDL fromall three types of pigs at the same rate (Table 3 andReference 8), suggesting that they lacked the ability torecognize the structural defect in Lpb5 apoB.

0.06

0.05

0.04

0.03

0.02

0.01

0.00 x

T

k b

• NrtveLDL0 Modified LDL

1Control Lpb 5.2 Lpb 5.1

Fra 6. Bar graph shows clearance of native and modified(glucosyiated) tow-density lipoprotein (LDL) in control (pig Nos.8, 9, 10, 11), Lpb5.2 (pig Nos. 15, 16, 17, 18), and Lpb5.1 (pigNos. 29, 30, 31, 32) pigs. Pigs were injected simultaneously with125!-autologous native and control pig (pig No. 7) 131l-modifiedLDLs (d=1.019 to 1.063 g/mL). All pigs received the samepreparation of control modified LDL Fractional clearance ratesrepresent total clearance from plasma. Data are plotted as theaverage of values obtained from four anlmals±SD. Differentletters represent significantly different means (P<.05) as deter-mined by Student's f test: lowercase letters designate nativeLDL; uppercase letters designate modified LDL

by guest on June 10, 2018http://atvb.ahajournals.org/

Dow

nloaded from

Aiello et al Phenotypic Heterogeneity of Spontaneous Hypercholesterolemia in Pigs 415

TABLE 4. Low-Density Upoproteln Fluxes in Lpb5.2 Pigs

Pig No.

LDL fluxes

25

21

22

23

24

Mean

SD

LDL fluxes

Mean

SD

LDL fluxes

Mean

SD

PlasmaCholesterol,

mg/dL

in Lpb5.2 pigs

110

130

117

112

135

121

10

in Lpb5.1 pigs*

218

25

In control pigs*

89

15

LDL Protein,

Buoyant

20.0

26.2

18.9

16.6

28.2

21.9

4.4

63.2

15.9

8.6

2.8

mg/dL

Dense

37.1

29.3

34.1

21.8

30.0

30.4

5.3

20.5

5.6

22.0

6.2

Production

F,o

64

129

163

123

117

119

32

326

143

14

30

Fa,

447

500

639

392

581

511

89

251

117

449

135

Flux, (fifl/kg)/h

Conversion

F,,

12

34

11

14

43

22

36

124

60

32

26

F12

129

71

208

108

150

133

46

117

23

75

53

Catabollsm

Foi

181

157

337

197

225

219

63

319

151

57

27

Foi

330

472

565

318

473

431

94

259

141

406

141

FractlorCons

Poo

"01

0.020

0.013

0.028

0.031

0.022

0.022

0.006

0.011

0.005

0.015

0.003

lal Ratetants,Is/h

0.031

0.040

0.035

0.037

0.046

0.037

0.005

0.028

0.011

0.043

0.012

LDL indicates low-density lipoproteln. Autologous buoyant (d= 1.019 to 1.038 g/mL) and dense (d= 1.038 to 1.063 g/mL) LDLs wereinjected simultaneously into Lpb5.2 pigs. Fractional rate constants (k0, and k^ were calculated by dividing LDL flux by correspondingLDL concentration (eg, F0] /buoyant LDL=fco,). These rates represent direct clearance from the pool and not total plasma clearance. F10

and F20 represent direct input into buoyant and dense LDL, respectively. F2, and F12 represent flux from buoyant to dense LDL and fluxfrom dense to buoyant LDL, respectively. Fo, and Fm represent total catabolic rate of buoyant and dense LDL, respectively. /c0, and frmrepresent fractional rate constants for the clearance of buoyant and dense LDL, respectively.

*Values for Lpb5.1 and control pigs are means of n=5 reported in Checovich et al.'°

It therefore was likely that the Lpb5.1 pigs were notexpressing normal levels of the LDL receptor in vivodespite the fact that fibroblasts from these pigs appearto express normal LDL receptor activity.2 To test thepossibility that the amount of receptor-mediated LDLclearance differed among the three phenotypic groups,we measured LDL receptor and nonreceptor in vivoclearance in control, Lpb5.2, and Lpb5.1 pigs. The pigswere simultaneously injected with native LDL and LDLchemically modified to abolish LDL receptor binding.Control and Lpb5.2 pigs cleared modified LDL at 50%of the rate at which they cleared native LDL (Figs 5 and6), suggesting that 50% of LDL clearance in theseanimals occurs via the LDL receptor pathway. Lpb5.1pigs cleared native and modified LDL at the same rate,suggesting that essentially all LDL clearance in theLpb5.1 pigs is LDL receptor independent (Figs 5 and 6).Thus, Lpb5.1 pigs have decreased in vivo expression ofLDL receptor activity relative to control and Lpb5.2

Pigs-It is interesting to note that although all pigs were

injected with the same preparation of modified LDL,the plasma clearance rate of the modified LDL stilldiffered in each phenotypic group. Modified LDL, likenative LDL, was cleared most rapidly in control pigs andslowest in the Lpb5.1 pigs. The fractional clearance rateof both the native and modified LDL therefore ap-peared to be inversely proportional to LDL pool size.

Production Rates of LDL SubspeciesThe Lpb5.1 pigs overproduce buoyant LDL.10 We

investigated whether buoyant LDL overproduction also

contributed to the moderate hypercholesterolemia as-sociated with the Lpb5.2 pig. Compared with valuespreviously obtained in control and Lpb5.1 pigs,10 theLpb5.2 pigs produced significantly more buoyant LDLthan did control pigs and significantly less buoyant LDLthan did Lpb5.1 pigs (Table 4). In the Lpb5.2 pigs,direct input into dense LDL (Fm) was greater thandirect input into buoyant LDL (F0)). Because Lpb5.1pigs produced less dense LDL than either control orLpb5.2 pigs, the total LDL production (input intobuoyant and dense LDL, Foi + F^) was actually similarfor Lpb5.1 and Lpb5.2 pigs.

Quantitative Pedigree Analysis of PlasmaCholesterol Phenotype

The existence of severely hypercholesterolemicLpb5.1 pigs and moderately hypercholesterolemicLpb5.2 pigs that apparently share the same apoB alleleis inconsistent with a genetic model whereby hypercho-lesterolemia in these pigs is solely due to a single mutantallele at the apoB locus. We therefore subjected ourpedigree-phenotype dataset, consisting of plasma cho-lesterol levels and known genotypes at the apoB locus,to quantitative pedigree analysis.

Four models were evaluated for their ability to de-scribe the inheritance pattern of plasma cholesterolphenotypes. As shown in Table 5, the model that doesnot include a second major gene in addition to apoB(model 4) can be rejected when compared with the mostgeneral model considered (model 1), with P< .000001.We can also reject a model that has a major locus tightlylinked to the apoB locus (model 2, 0=0, P<.001). The

by guest on June 10, 2018http://atvb.ahajournals.org/

Dow

nloaded from

416 Arteriosclerosis and Thrombosis Vol 14, No 3 March 1994

TABLE 5. Comparison of Models for Inheritance of Plasma Cholesterol Level

Model Gene Loci Recombination Parameters

(1)

(2)

(3)

(4)

Two loci

Two loci, tightly linked

Two loci, unlinked

One locus

apoB

MG, PG

apoB

MG, PG

apoB

MG, PG

apoB

PG

e estimated

0=0

6=0.5

12

11

11

2926.54

2937.42

2929.96

<.0O1

.0645

2992.85 <.000001

apoB indicates apolipoprotein B; MG, major gene; PG, polygenic effects; and 9, recombination fraction.*For comparison with the two-locus model wtth 6 estimated (model 1).

model that includes a major locus unlinked to the apoBlocus (model 3) cannot be rejected (F=.O645) and is thebest-fitting (most parsimonious) model for our pedi-gree-phenotype dataset. The parameters of this modelinclude (Table 6) allele frequencies at the apoB locusand the major locus. The major locus appears to beunlinked to apoB, although power to estimate 0 is low;the estimated recombination fraction in model 1 is 0.21,but with a log of the odds score of only 0.74. The majorlocus has a nearly recessive allele determining highcholesterol levels, with an effect that is approximatelyequal to the effect of the apoB locus. The frequency ofthis allele is 0.37, and the frequency of the Lpb5 apoBallele is 0.42. These are maximum likelihood estimatesdetermined from founders (individuals whose parentsare not included in the pedigree). The two loci togetheraccount for 64% of the variance in total cholesterol, andpolygenes account for an additional 24% (data notshown).

We also tested the hypothesis of mendelian transmis-sion at the major locus by evaluating a model in whichthe parameter values were set equal to those of thetwo-locus mendelian model, but the transmission prob-abilities associated with the three major locus genotypeswere estimated rather than being restricted to their

mendelian values (results not shown). The two-locusmendelian model could not be rejected compared withthis more general model, and the estimated transmis-sion probabilities were not significantly different frommendelian values. We further tested an environmentalheterogeneity model, with parameter values equal tothose of the two-locus mendelian model but with geno-type of offspring independent of parental genotype(results not shown). This model was rejected comparedwith the more general model in which transmissionprobabilities were estimated. These tests provide fur-ther evidence that a major locus, in addition to the apoBlocus, influences total cholesterol in our pig population.

DiscussionWe previously characterized spontaneously hyper-

cholesterolemic pigs carrying a defined allele for apoB(Lpb5). The animals had been selected on the basis ofapoB allotypes,21 and those allotypes correlated withspecific markers in apoB genomic DNA.4-6 The hyper-cholesterolemia phenotype always segregated with theapoB genomic markers; ie, all hypercholesterolemicpigs carried the Lpb5 apoB allele and all normocholes-terolemic pigs carried non-Lpb5 alleles. This led us tohypothesize that a structural defect encoded by the Lpb3

TABLE 6. Parameter Estimates for Model 3 and Mean Estimated Plasma CholesterolConcentration for the Nine Two-Locus Genotypes

Parameter Estimates

Frequency of the Lpb5 apoB allele

Frequency of the "high" second locus allele

Linkage disequilibrium

Polygenic heritability (h2)

Variation around apoB and second locus genotypes

Sex effect (males higher than females)

0.42

0.37

0.61

0.68

38.0 mg/dL

11.0 mg/dL

Major Locus Allele

"Low" homozygote

Heterozygote

"High" homozygote

Mean Estimated Plasma

Non-5/Non-5

93

85

207

Cholesterol, mg/dL

ApoB Allele

5/Non-5

121

113

235

5/5

149

142

264

apoB indicates apolipoprotein B.

by guest on June 10, 2018http://atvb.ahajournals.org/

Dow

nloaded from

Aiello et al Phenotypic Heterogeneity of Spontaneous Hypercholesterolemia in Pigs 417

allele was responsible for the hypercholesterolemia ofthose pigs. This hypothesis was supported by studiesthat revealed functional defects in LDL particles frompigs carrying the Lpb5 allele; LDL from the mutanthypercholesterolemic Lpb5 pigs has a twofold to sixfoldlower affinity for the pig LDL receptor7 and is catabo-lized more slowly in vivo8 compared with LDL fromnon-Lpb5 pigs.

In addition to being biologically defective, LDL par-ticles from the Lpb5 pigs are also structurally unusual.8

Buoyant cholesterol ester-enriched LDL particles arepresent at four to eight times the normal concentrationin pig plasma and comprise approximately 70% of thetotal plasma LDL protein in Lpb5 pigs. In vivo turnoverstudies showed that the accumulation of buoyant LDLwas primarily due to a 14-fold increase in the produc-tion rate of these particles.10 We therefore concludedthat there were two distinct phenotypes associated withthe severe hypercholesterolemia of Lpb5 pigs: slowlycatabolized LDL particles and overproduction of buoy-ant LDL.

However, the Lpb5 pigs displayed phenotypic hetero-geneity: some had severe hypercholesterolemia with aprominence of buoyant LDL (Lpb5.1 pigs), and othershad very modest hypercholesterolemia with only a smallincrease in buoyant LDL (Lpb5.2 pigs).9 Because thesame polymorphic markers used to distinguish Lpb5

apoB from the seven other known pig apoB alleles arepresent in Lpb5.1 and Lpb5.2 apoB, we presume thatthese two groups of pigs share the same apoB allele,although complete sequence analysis of the Lpb5.1 andLpb5.2 apoB genes is required before this can beverified. If polymorphism within the Lpb5 apoB gene isresponsible for the variation in plasma cholesterol levelsamong Lpb5 pigs, it is not manifested by altered catab-olism of the resulting LDL particles; we have shown thatdifferences in the catabolic properties of the LDLparticles could not account for the differences in plasmacholesterol between Lpb5.1 and Lpb5.2 pigs.

One of the more striking observations arising fromthis study was the apparently complete absence of LDLreceptor activity in Lpb5.1 pigs. This may explain whythese animals are unable to distinguish Lpb5 LDL fromnormal pig LDL.8 In addition, it raises the possibilitythat buoyant LDL overproduction might be related tothe absence of LDL receptor activity. For example, adefect in an enzyme associated with cholesterol metab-olism might increase the metabolically active hepaticcholesterol pool. This pool may simultaneously affectLDL receptor expression and influence the cholesteroland/or fatty acid composition of de novo lipoproteinparticles. Alternatively, a lack of LDL receptor activitymay cause the more extensive production of buoyantLDL by delaying VLDL remnant clearance, resulting inan increased conversion of VLDL to LDL. This wouldbe analogous to the situation in the Watanabe heritablehyperlipidemic rabbit, which displays overproduction ofLDL due to slow clearance of IDL.22 However, Huff andTelford23 determined that 80% of LDL in miniaturepigs is directly synthesized by pathways not involving theVLDL cascade. In Lpb5.1 pigs, we could account foronly 20% of buoyant LDL as having been derived fromplasma VLDL, suggesting that "direct" production ofbuoyant LDL was occurring.9 As explained by Shamesand Havel,24 one cannot rule out the possibility that

there is a rapidly turning over pool of VLDL that isconverted to LDL. Because the VLDL pool size in pigsis very small, it is impossible to carry out VLDLturnover studies in pigs in a manner that accuratelymodels the heterogeneity of VLDL.

LDL heterogeneity has been observed in humans.25

Austin et al26 proposed that LDL size might be deter-mined by a single major gene locus. Recently Nishina etal27 reported linkage of the LDL size phenotype to theLDL receptor locus, although recent evidence suggeststhat the linked locus is not the LDL receptor structuralgene.28 Although variation in LDL receptor activity maycertainly occur in this pig population, two observationsare incompatible with the LDL receptor being thesecond major cholesterol-influencing locus. First, cul-tured fibroblasts from Lpb5.1 pigs appear to expressnormal LDL receptor activity.2 Second, the expressionof the second major locus is apparently recessive,whereas the known LDL receptor mutations are ex-pressed in a codominant fashion. Taken together, theobservations in the mutant pigs and in humans areconsistent with multiple factors affecting LDL hetero-geneity, either through an effect on LDL receptorexpression (in pigs) or perhaps as a consequence of thepolymorphism near the LDL receptor structural gene(in humans).

The quantitative pedigree analysis performed on ourpig population indicated that the distribution of plasmacholesterol levels was most consistent with the apoBgene and a second major locus, probably unlinked toapoB, contributing to the hypercholesterolemia. Thesetwo loci account for a substantial portion (64%) of thevariance in plasma cholesterol concentration in thispedigree. Even so, the high polygenic heritability(h2=0.68, Table 6) indicates that further genetic effectsremain to be identified. The remaining genetic contri-bution could be either multiple genes with small effects,an additional major gene, or a combination of the two.

A surprising result of the genetic analysis was highgametic disequilibrium, indicating an association be-tween alleles at the two major loci —surprising becausethe prediction of the same analysis was that the twogenes are unlinked. Gametic disequilibrium impliesnonrandom transmission of alleles at different loci andnot necessarily physical proximity of the two loci. Thus,specific haplotypes occur with higher than randomfrequency in our pig population. This could result froma bias in the selection of animals for breeding. Indeed,because we had generally bred for Lpb5 pigs with highcholesterol, we most likely selected for pigs carrying thehigh-cholesterol alleles at each of the two major loci.

Rapacz and Hasler-Rapacz29 previously determinedthat three separate immunologically identified genescontribute to hypercholesterolemia in the pig popula-tion: Lpb, Lpu, and Lpr. The combination of the Lpb5,Lpu', and Lpr1 alleles was associated with the mostseverely hypercholesterolemic pigs, and the originalnomenclature of the Lpb5.1 and Lpb5.2 pigs was de-rived from their haplotypes at the Lpb and Lpu loci(Lpb5.1 pigs were Lpb5,Lpu'; Lpb5.2 pigs wereLpb5,Lpu2).3 The Lpr gene encodes an apolipoproteinunique to pigs termed apoR,30-31 and subsequent to theinitial observation, we were unable to confirm that thegenotype at the Lpr locus is associated with variation inplasma cholesterol levels.32 Thus, we do not suspect that

by guest on June 10, 2018http://atvb.ahajournals.org/

Dow

nloaded from

418 Arteriosclerosis and Thrombosis Vol 14, No 3 March 1994

the second major locus indicated by the pedigree analy-sis is Lpr. The protein product of the Lpu locus hasnever been determined, although it is known to bepresent on pig LDL. A thorough investigation by Lee etal33 of the effect of homozygosity and heterozygosity atthe Lpu locus on lipoprotein composition and concen-tration in Lpb5 pigs revealed that Lptf'^Lpu'" pigsdisplay the most elevated total plasma cholesterol andbuoyant LDL levels. Lpu may be either a separate alleleof apoB or tightly linked to the apoB gene, as norecombination between Lpb5 and Lpu' occurred ingenetic crosses over 18 generations.29-33 Because Lpu isan LDL-borne allotype, it may represent epitopesunique to Lpb5.1 buoyant LDL, thus accounting for itslinkage disequilibrium with apoB and its correlationwith severe hypercholesterolemia.

Our finding that a second major gene influencesplasma cholesterol levels in pigs is of considerableinterest in light of recent findings in humans. Severalstudies have linked polymorphisms within the apoBgene to variation in plasma triglyceride, cholesterol, orapoB levels3438 as well as to the development of athero-sclerosis.39"" However, these correlations greatly de-pend on the population being studied, such that apoBgene polymorphisms may be associated with levels ofnone, some, or all of the lipoprotein components.Clearly, other genetic factors contribute to hyperlipid-emia in humans. Coresh et al42 recently found thatcertain apoB restriction fragment length polymorphismsdid not appear to contribute to variation in plasma apoBconcentration and that apoB is not the major geneinfluencing plasma apoB levels in 198 individuals from23 families, as determined by pedigree and sib-pairlinkage analysis.42 Although variation in the apoB geneclearly contributes to apoB and plasma cholesterollevels in pigs, the evidence that another major gene orgenes determine cholesterol concentration supports theuse of these Lpb5 pigs as a model for human hypercho-lesterolemia and as a tool for identifying other gene lociinvolved in the determination of plasma cholesterollevels.

AcknowledgmentsThis work was supported by National Institutes of Health

(NIH) grant HL-37251 and the Council for Tobacco ResearchGrant No. 2077 to Dr Attie and by NIH grants HL-28972 andHL-45522 to Dr MacCluer and Dr Blangero. Dr Aiello and DrEbert were supported by NIH postdoctoral fellowships. DrNevin was supported by NIH Physician Scientist Award HL-02162. Dr Attie is an Established Investigator of the AmericanHeart Association. The authors wish to thank the staff of theUW-Madison School of Veterinary Medicine, Charmany Fa-cility, and the staff of the UW-Madison Department of Vet-erinary Science for their excellent care of the pigs used in thisstudy. We are also grateful to Dr Lynn Allen-Hoffman forproviding us with human fibroblasts and to Kurt A.A. Grun-wald for outstanding technical assistance.

References1. Innerarity TL, Mahley RW, Weisgraber KH, Bersot TP, Krauss

RM, Vega GL, Grundy SM, Friedl W, Davignon J, McCarthy BJ.Familial defective apolipoprotein B-100: a mutation of apolipo-protein B that causes hypercholesterolemia. J Upid Res. 1990;31:1337-1349.

2. Rapacz J, Hasler-Rapacz J, Taylor KM, Checovich WJ, Attie AD.Lipoprotein mutations in pigs are associated with elevated plasmacholesterol and atherosclerosis. Science. 1986;234:1573-1577.

3. Rapacz J, Hasler-Rapacz J, Kuo WH. Immunogenetic poly-morphism of lipoproteins in swine, 2: five new allotypic specificities(Lpp6, Lppll, Lppl2, Lppl3, and Lppl4) in the Lpp system.Immunogenencs. 1978;6:405-424.

4. Maeda N, Ebcrt DL, Doers TM, Newman M, Hasler-Rapacz J,Attie AD, Rapacz J, Smithies O. Molecular genetics of the apo-lipoprotein B gene in pigs in relation to atherosclerosis. Gene.1988;70:213-218.

5. Kaiser M, Nevin DN, Sturley SL, Purtell C, Attie AD. Determi-nation of pig apolipoprotein B genotype by gene amplification andrestriction fragment length polymorphism analysis. Anim Genet.1993;24:117-120.

6. Purtell C, Maeda N, Ebert DL, Kaiser M, Lund-Katz S, Sturley SL,Kodoyianni V, Grunwald K, Nevin DN, Aiello RJ, Attie AD.Nucleotide sequence encoding the carboxyl-terminal half of apo-lipoprotein B from spontaneously hypercholesterolemic pigs.J Upid Res. In press.

7. Lowe SW, Checovich WJ, Rapacz J, Attie AD. Defective receptorbinding of low density lipoprotein from pigs possessing mutantapolipoprotein B alleles. J Biol Chem, 1988;263:15467-15473.

8. Checovich WJ, Fitch WL, Krauss RM, Smith MP, Rapacz J, SmithCL, Attie AD. Defective catabolism and abnormal composition oflow-density lipoproteins from mutant pigs with hypercholester-olemia. Biochemistry. 1988;27:1934-1941.

9. Cooper ST, Aiello RJ, Checovich RJ, Attie AD. Low densitylipoprotein heterogeneity in spontaneously hypercholesterolemicpigs. Mol Cell Biochem. 1992;113:133-140.

10. Checovich WJ, Aiello RJ, Attie AD. Overproduction of a buoyantlow density lipoprotein in spontaneously hypercholesterolemicmutant pigs. Artenoscler Thromb. 1991;11:351-361.

11. Markwell MK, Haas SM, Bieber LL, Tolbert NE. A modificationof the Lowry procedure to simplify protein determination inmembrane and lipoprotein samples. Anal Biochem. 1978;87:206-210.

12. Kesaniemi YA, Witztum JL, Steinbrecher UP. Receptor-mediatedcatabolism of low density lipoprotein in man: quantitation usingglucosylated low density lipoprotein. J Clin Invest. 1983;71:950-959.

13. Matthews CME. The theory of tracer experiments with I3'I-labelled plasma proteins. Phys Med Biol. 1957;2:36-53.

14 Snedecor GW, Cochran WG. Statistical Methods. Ames, Iowa: IowaState University Press; 1980.

15. Gurpide E, Mann J, Sandberg E. Determination of kineticparameters in a two pool system by administration of one or moretracers. Biochemistry. 1964;3:1250-1255.

16. Cobelli C, Toffolo G. Compartmental vs. noncompartmentalmodeling for two accessible pools. Am J Physwl. 1984;247:R488-R496.

17. Berman M, Beltz WF, Greif PC, Chabay R, Boston RC. Conver-sational SAAM (CONSAM) User's Guide. Washington, DC- Gov-ernment Printing Office; 1983.

18. Goldstein JL, Basu SK, Brown MS. Receptor-mediated endocytosisof low-density lipoprotein in cultured cells Methods Enzymot 1983;98:241-260.

19. Hasstedt SJ, Cartwright PE. PAP: Pedigree Analysis Package. SaltLake City, Utah: Department of Medical Biophysics and Com-puting, University of Utah; 1989. Technical report 13.

20. Blangero J, MacCluer JW, Kammerer CM, Mott GE, McGill HCJr. Genetic analysis of apolipoprotein A-I in two environments. AmJ Hum Genet. 1990;47:414-428.

21. Rapacz J. Lipoprotein immunogenetics and atherosclerosis. Am JMed Genet. 1978;l:377-405.

22. Kita T, Brown MS, Bilheimer DW, Goldstein JL. Delayedclearance of very low density and intermediate density lipo-proteins with enhanced conversion to low density lipoprotein inWHHL rabbits. Proc Natl Acad Sci USA. 1982;79:5693-5697.

23. Huff MW, Telford DE. Direct synthesis of low-density lipoproteinapoprotein B in the miniature pig. Metabolism, 1985;34:36-42.

24. Shames DM, Havel RJ. De novo production of low density lipo-proteins: fact or fancy. / Upid Res. 1991^32:1099-1112.

25. Austin MA, King MC, Vranizan KM, Krauss RM. Atherogeniclipoprotein phenotype: a proposed genetic marker for coronaryheart disease risk. Circulation, 1990;82:495-506.

26. Austin MA, Brunzell JD, Fitch WL, Krauss RM. Inheritance oflow density lipoprotein subclass patterns in familial combinedhyperlipidemia. Arteriosclerosis. 199O;10:520-530.

27. Nishina PM, Johnson JP, Naggert JK, Krauss RM. Linkage ofatherogenic lipoprotein phenotype to the low density lipoproteinreceptor locus on the short arm of chromosome 19. Proc Natl AcadSci US A. 1992;89:708-712.

by guest on June 10, 2018http://atvb.ahajournals.org/

Dow

nloaded from

Aiello et al Phenotypic Heterogeneity of Spontaneous Hypercholesterolemia in Pigs 419

28. Austin MA, Stephens K, Wijsman E, Walden C, Loretz C.Evidence against the linkage of small dense lipoproteins (LDL) tothe LDL receptor locus on chromosome 19. Circulation. 1993;88(suppl I):I-268. Abstract.

29. Rapacz J, Hasler-Rapacz J. Investigations on the relationshipbetween immunogenetic polymorphisms of /3-lipoproteins and the0-lipoprotein and cholesterol levels in swine. In: Lenzi S, Des-covich GC, eds. Atherosclerosis and Cardiovascular Diseases.Bologna, Italy: Editrice Compositori; 1984:99-108.

30. Rapacz J, Hasler-Rapacz J, Kuo WH. Immunogenetic poly-morphism of lipoproteins in swine: genetic, immunological andphysiochemical characterization of the two allotypes lprl and Ipr2.Genetics. 1986;113:985-1007.

31. Cooper ST, Attie AD. Pig apolipoprotein R: a new member of theshort consensus repeat (SCR) family of proteins. Biochemistry.1992;31:12328-12336.

32. Cooper ST. Pig apolipoprotein R: a new member of the shortconsensus repeat (SCR) family of proteins. Madison, Wis: Uni-versity of Wisconsin; 1992. Dissertation.

33. Lee DM, Mok T, Hasler RJ, Rapacz J. Concentrations and com-positions of plasma lipoprotein subfractions of Lpb5-Lpulhomozygous and heterozygous swine with hypercholesterolemia.J Lipid Res. 1990;31:839-847.

34. Berg K. Genetics of coronary heart disease and its risk factors. In:Gock G, Collins GM, eds. Molecular Approaches to HumanPotygenic Disease. Chichester, England: Wiley & Sons; 1987:14-28.

35. Aalto SK, Gylling H, Helve E, Kovanen P, Miettinen TA, TurtolaH, Kontula K. Genetic polymorphism of the apolipoprotein B genelocus influences serum LDL cholesterol level in familial hypercho-lesterolemia. Hum Genet. 1989;82:305-307.

36. Xu C, Tikkanen MJ, Huttunen JK, Pietinen P, Butler R,Humphries S, Talmud P. Apolipoprotein B signal peptide inser-tion/deletion polymorphism is associated with Ag epitopes andinvolved in the determination of serum triglyceride levels. J LipidRes. 1990;31:1255-1261.

37. Hansen PS, Gerdes LU, Klausen IC, Gregerson N, Faergeman O.Polymorphisms in the apolipoprotein B-100 gene contribute tonormal variation in plasma lipids in 464 Danish men born in 1948.Hum Genet. 1993;91:45-50.

38. Visvikis S, Cambou JP, Arveiler D, Evans AE, Parra HJ, AguillonD, Fruchart JC, Siest G, Cambien F. Apolipoprotein B signalpeptide polymorphism in patients with myocardial infarction andcontrols. Hum Genet. 1993;90:561-565.

39. Hegele RA, Huang LS, Herbert PN, Blum CB, Buring JE, Hen-nekens CH, Breslow JL. Apolipoprotein B-gene DNA polymor-phisms associated with myocardial infarction. N EnglJ Med. 1986;315:1509-1515.

40. Rajput-Williams J, Wallis SC, Yarnell J, Bell GI, Knott TJ,Sweetnam P, Cox N, Scott J. Variation of apolipoprotein-B gene isassociated with obesity, high blood cholesterol levels, andincreased risk of coronary heart disease. Lancet. 1988;52:1442-1446.

41. Hixson JE, McMahan CA, McGill HC Jr, Strong JP. Apo B inser-tion/deletion polymorphisms are associated with atherosclerosis inyoung black but not young white males. Arterioscler Thromb. 1992;12:1023-1029.

42. Coresh J, Beaty TH, Kwiterovich PJ, Antonarakis SE. Pedigreeand sib-pair linkage analysis suggest the apolipoprotein B gene isnot the major gene influencing plasma apolipoprotein B levels. AmJ Hum Genet. 1992;50:1038-1045.

by guest on June 10, 2018http://atvb.ahajournals.org/

Dow

nloaded from

and A D AttieR J Aiello, D N Nevin, D L Ebert, P J Uelmen, M E Kaiser, J W MacCluer, J Blangero, T D Dyer

spontaneous hypercholesterolemia in pigs.Apolipoprotein B and a second major gene locus contribute to phenotypic variation of

Print ISSN: 1079-5642. Online ISSN: 1524-4636 Copyright © 1994 American Heart Association, Inc. All rights reserved.

Avenue, Dallas, TX 75231is published by the American Heart Association, 7272 GreenvilleArteriosclerosis, Thrombosis, and Vascular Biology

doi: 10.1161/01.ATV.14.3.4091994;14:409-419Arterioscler Thromb Vasc Biol. 

http://atvb.ahajournals.org/content/14/3/409World Wide Web at:

The online version of this article, along with updated information and services, is located on the

  http://atvb.ahajournals.org//subscriptions/

at: is onlineArteriosclerosis, Thrombosis, and Vascular Biology Information about subscribing to Subscriptions:

  http://www.lww.com/reprints

Information about reprints can be found online at: Reprints: 

document.Permissions and Rights Question and AnswerFurther information about this process is available in theis being requested is located, click Request Permissions in the middle column of the Web page under Services.Clearance Center, not the Editorial Office. Once the online version of the published article for which permission

can be obtained via RightsLink, a service of the CopyrightArteriosclerosis, Thrombosis, and Vascular Biology Requests for permissions to reproduce figures, tables, or portions of articles originally published inPermissions:

by guest on June 10, 2018http://atvb.ahajournals.org/

Dow

nloaded from