60 PREDICTING FROZEN BOAR SPERM FERTILITY USING NOVEL AND CONVENTIONAL ANALYSES

Abstract

Cryopreserved boar sperm is seldom used for AI because fertility is reduced. Despite many potential advantages of frozen-thawed sperm for AI, lack of reliable fertility estimation of frozen ejaculates before AI limits the application of frozen sperm. Conventional post-thaw evaluation of sperm does not accurately estimate fertility. Identifying sperm traits that predict fertility would help select ejaculates that produce adequate litter sizes. Our objective was to identify traits of cryopreserved sperm that are related to boar fertility for AI through the use of novel and traditional laboratory analyses. Semen from 14 boars of several breeds was cooled to 15°C for shipping before freezing. Post-thaw motility was evaluated using a microscope and confirmed with computer-automated sperm analysis. Sperm viability and acrosome integrity were measured at 0, 30, and 60 min post-thaw. In addition to traditional analyses, each sperm sample was tested by IVF to record fertilization, cleavage, and blastocyst development. A sperm-oviduct binding assay was used to compare the number of sperm bound to epithelial aggregates harvested from the isthmus. Additionally, a competitive zona binding assay using 2 distinct fluorophores for boar identification was used to count the number of sperm from each boar bound to the zona. Frozen sperm from the same ejaculates subjected to laboratory analyses were used to determine actual boar fertility. Fertility was measured by AI of mature gilts using 4.0 × 10⁹ total sperm from one boar at 24 h and a second boar at 36 h after the onset of oestrus, and AI order was reversed in consecutive replicates. Fertility was expressed as the percentage of the litter sired by each boar. Reproductive tracts were harvested at 32 days after AI, and fetal paternity was identified using microsatellite markers. The actual boar fertility was regressed against the mean of each laboratory evaluation by boar, and the assays that best predicted fertility were identified using stepwise logistic regression. The model generated was highly predictive of fertility (P < 0.001, r² = 0.87) and included 5 traits: acrosome compromised sperm (0 and 30 min), percent live sperm (0 min), percent total motility (30 min), and the number of zona bound sperm. An additional model in which fertility was assessed by the number of piglets sired by boar also predicted fertility (P < 0.05, r² = 0.57) and shared many of the same traits. These models were highly accurate when used to predict actual fertility of cryopreserved boar sperm. This approach may be used to screen ejaculates before AI and advance the use of frozen boar sperm by the swine industry.

Research was supported by Agriculture and Food Research Initiative Competitive Grant no. 2010-85112-20620 from the USDA National Institute of Food and Agriculture.