Quantile-specific heritability of total cholesterol and its pharmacogenetic and nutrigenetic implications

Abstract

Background: "Quantile-dependent expressivity" occurs when the effect size of a genetic variant depends upon whether the phenotype (e.g. cholesterol) is high or low relative to its distribution. We have previously shown that the effect of a 52-SNP genetic-risk score was 3-fold larger at the 90th percentile of the total cholesterol distribution than at its 10th percentile. The objective of this study is to assess quantile-dependent expressivity for total cholesterol in 7006 offspring with parents and 2112 sibships from Framingham Heart Study.

Methods: Quantile-specific heritability (h²) was estimated as twice the offspring-parent regression slope as robustly estimated by quantile regression with nonparametric significance assigned from 1000 bootstrap samples.

Results: Quantile-specific h² increased linearly with increasing percentiles of the offspring's cholesterol distribution (P = 3.0 × 10^-9), i.e. h² = 0.38 at the 10th percentile, h² = 0.45 at the 25th percentile, h² = 0.52 at the 50th, h² = 0.61 at the 75th percentile, and h² = 0.65 at the 90th percentile of the total cholesterol distribution. Average h² decreased from 0.55 to 0.34 in 3564 offspring who started cholesterol-lowering medications, but this was attributable to quantile-dependent expressivity and the offspring's 0.94 mmol/L average drop in total cholesterol. Quantile-dependent expressivity likely explains the reported effect of the CELSR2/PSRC1/SORT1 rs646776 and APOE rs7412 gene loci on statin efficacy. Specifically, a smaller genetic effect size at the lower (post-treatment) than higher (pre-treatment) cholesterol concentrations mandates that the trajectories of the genotypes cannot move in parallel when cholesterol is decreased pharmacologically.

Conclusion: Cholesterol concentrations exhibit quantile-dependent expressivity, which may provide an alternative interpretation to pharmacogenetic and nutrigenetic interactions.

Keywords: Cholesterol; Diet; Gene-environment interactions; Heritability; Nutrigenetics; Obesity; Pharmacogenetics; Precision medicine; Quantile regression.

Figure 1

A) Offspring-parent regression slopes (β_OP) for selected quantiles of the offsprings’ total cholesterol concentrations, with corresponding estimates of heritability (2β_OP=h²). The slopes became progressively greater (i.e., steeper) with increasing quantiles of the cholesterol distribution. B) Quantile-specific heritability function formed by combining the selected quantile-specific regression slopes from above with those of other quantiles. Significance of the linear, quadratic and cubic trends and the 95% confidence intervals (shaded region) were determined by 1000 bootstrap samples. Figure 1C compares the offspring-parent (β_OP) and full-sib (β_FS) regression slopes for total cholesterol concentrations by quantiles of the offspring and sibling distribution. Theoretically, 2β_OP and 2β_FS both equal h² if assortative mating, dominance, and sibling shared environmental effects are negligible, as suggested by their overlap.

Figure 2.

A) Quantile-specific heritability plots before and after starting cholesterol medication when the two plots are displayed together by the percentile of the offspring’s total cholesterol concentrations. B) The heritability plots when they are matched by the offsprings’ total cholesterol concentrations in mmol/L using P-P plots [19]. Note that the heritability difference on and off medications (0.19) in the upper panel is essentially eliminated when matched by cholesterol levels in the lower panel.

Figure 3.

Precision medicine perspective of genotype-specific cholesterol change (histogram inserts) vs. quantile-dependent expressivity perspective of larger genetic effect size when average cholesterol concentrations were high (line graphs) for: A) Hopewell et al.’s 2013 report [20] on LDL-cholesterol before and after being treated with 40 mg/d simvastatin in 185 carriers of two ε2, 2794 carriers of one ε2, and 12,305 carriers of zero ε2 alleles of APOE rs7412 (P_interaction 2.7×10⁻³⁰); B) Hopewell et al.’s 2013 report [20] on LDL-cholesterol before and after simvastatin treatment in 11,305 TT homozygotes, 6137 CT heterozygotes, and 847 CC homozygotes of CELSR2/PSRC1/SORT1 rs646776 (P_interaction =9.5×10⁻⁵); C) Dreon et al.’s 1995 report [21] on total cholesterol change in switching from a high-fat to a low-fat diet in 28 APOE ε4-carriers, 65 ε3ε3, and ten ε3ε2 (P_interaction=0.02), D) Friedlander et al.’s 1995 report [22] on the LDL-cholesterol reduction in switching from basal to treatment diets in four APOE ε2ε3, forty-nine ε3ε3, and three ε3ε4 (P_interaction≤0.01); E) Tikkanen et al’s 1990 report [23] on total cholesterol change due to switching from a high-fat high-cholesterol diet (average of baseline and switchback diets) to a low-fat low-cholesterol diet in eight APOE ε4ε4 and 102 non-ε4ε4 (P_interaction=0.01); F) Moreno et al.’s 2004 report [24] on LDL-cholesterol change when switched from a 47% carbohydrate/20% SFA diet to a 55% carbohydrate/<10% SFA diet in 48 carriers of the T allele and seven GG homozygotes of the APOE −219G>T promoter (P_interaction<0.05).

Figure 4.

Precision medicine perspective of genotype-specific cholesterol change (histogram inserts) vs. quantile-dependent expressivity perspective of larger genetic effect size when average cholesterol concentrations were high (line graphs) for: A) Humphries et al.’s 1996 report [25] on the total-cholesterol response to a high SFA fat and high PUFA fat diets in 45 H+ and 10 H− genotypes of the Hindlll LPL gene loci (P=0.03); B) Jansen et al.’s 1997 report [26] on the LDL-cholesterol response of switching from an NECP type 1 diet to a high SFA diet in 25 APOA4 347Thr homozygotes and 16 carriers of the 347Ser allele (P_interaction=0.05); C) Herron et al.’s 2006 reported [27] on the total cholesterol response to switching from egg substitute to one egg/d. for 30 days in 68 CC homozygotes and 23 G-carriers of the ABCG5 polymorphism (P_interaction<0.05); and D) Perez-Martinez et al.’s 2003 report [28] on LDL-cholesterol concentrations in switching from a SFA-rich diet to high-carbohydrate diet in sixty-five 1/1 homozygotes and thirty-two ½ heterozygotes of SRB-I exon 1 (P_interaction =0.007); E) Guevara-Cruz et al.’s 2013 report [29] of 23 AA and 18 AG and GG genotypes of rs12449157 in the glucose-fructose oxidoreductase domain containing 2 (GFOD2) locus on pre-treatment basal diet followed by three-month low saturated fat diet with 25 g soy protein and 15 g soluble fiber supplement (P_interaction=0.0001); F) Jemaa et al.’s 2004 report [30] showing greater total cholesterol reduction with low-calorie weight loss diet in 110 carriers of the Del allele of the APOB Ins/Del signal peptide polymorphism vs. 121 Ins/Ins homozygotes (P_interaction=0.001).