Control of variance in experimental studies of hyperlipidemia using the WHHL rabbit.

Tm Donnelly,Dm Levine,Ts Parker,Sf Kelsey

doi:10.1016/s0022-2275(20)41972-8

Tm Donnelly, Dm Levine + Show 2 more

Open Access

https://doi.org/10.1016/s0022-2275(20)41972-8

Copy DOI

Abstract

Between-animal variability has frustrated many experimental studies in outbred animal models of human disease. Variability that arises from genetic heterozygosity can be minimized by use of experimental designs that match littermates (polyzygotic twins) across control and treatment groups. Poor breeding vigor has prevented use of this experimental design in the WHHL rabbit model of hyperlipidemia and atherosclerosis. A comparison of reproduction in WHHL and normal rabbits demonstrated that litter size is limited by functional deficits at ovulation, implantation, and gestation in WHHL females. Superovulation of females reliably produced expanded litters of WHHL rabbits. Plasma lipids were measured in expanded litters of Japanese White WHHL (JW-WW) and English Half-lop WHHL (EHL-WW) rabbits. The variance of plasma cholesterol within sibships was two- to three-fold less than that between-litters. Intraclass correlation of total cholesterol within litters of EHL-WW was 0.72 and within litters of JW-WW was 0.67. These data provide evidence of genetic modulation of hypercholesterolemia in WHHL rabbits and demonstrate that experimental designs in which littermates are paired across groups can decrease the number of animals needed or increase the sensitivity of hypothesis tests by two- to threefold.

Highlights

Between-animal variability has frustrated many experimental studies in outbred animal models of human disease
Japanese White rabbit (JW)-WW females gave birth at 29-31 days compared to 30-32 days for New Zealand White rabbit (NZW)-ww females (30.6 k 1.2 days, n = 13 vs. 31.8 k 0.4 days, n = 13; P < 0.005)
The smaller number of live-born pups in the W H H L litters together with the large number of spontaneous abortions and stillborn young suggested that the W H H L females were losing embryos early in pregnancy

Summary

Introduction

Between-animal variability has frustrated many experimental studies in outbred animal models of human disease. Plasma lipids were measured in expanded litters of Japanese White W H H L OW-WW) and English Half-lop W H H L (EHL-WW) rabbits. Cholesterol-fed rabbits develop aortic lesions that contain foam cells predominantly [4,5], unless the diet is carefully adjusted to match plasma lipid concentrations in W H H L rabbits [6,7,8]. Experimental studies that use the W H H L rabbit must be designed with sufficient numbers of animals to cope with the problem of between-animal variability It is a fundamental principle of statistics that the number of samples (animals) needed to test an hypothesis is directly proportional to the variance of the measure of outcome [19].

Methods

Results

Discussion

Conclusion