LAUSR

We aimed to compare three methods to detect pregnancy based on interferon-tau stimulated gene expression in peripheral blood polymorphonuclear cells and Doppler ultrasonography (Doppler-US) 20 days after timed AI and the concentrations of pregnancy associated glycoproteins (PAGs) 25 days after timed AI. Our second objective was to compare interferon-tau stimulated genes and PAGs as early embryo loss (EEL) predictors. Nelore cows (n = 144) and heifers (n = 103) were submitted to timed AI (Day (D) 0). On D20, polymorphonuclear cells were isolated from blood samples by Ficoll gradient (GE Healthcare), and the RNA was extracted. Expression of the interferon-tau stimulated gene (OAS-1) was quantified by qPCR and normalized to reference genes (GAPDH and ACTB). On D25, blood was collected from the jugular vein and PAG concentrations were measured by an in-house enzyme-linked immunosorbent assay. Pregnancy diagnoses were performed on D20 by luteolysis detection with Doppler-US and on D30 by detection of an embryo by heartbeat. Animals were classified as pregnant (P; fetus on D30), non-P (NP; no active corpus luteum on D20), and EEL (active corpus luteum on D20 but NP on D30). Expression of OAS-1 and PAG concentrations were analysed by analysis of variance using PROC MIXED of SAS considering the effects of group, category, and their interaction. Receiver operating characteristic curves were created, and the area under the curve (AUC), accuracy, specificity, and sensitivity were calculated for pregnancy predictions on D20 (OAS-1 and Doppler-US) and D25 (PAGs) compared with the standard diagnosis method on D30. Expression of OAS-1 and PAG concentrations were greater (P < 0.01) in the P (2.4 ± 0.2 and 4.8 ± 0.2 ng mL−1, respectively) compared with the NP (0.50 ± 0.1 and 1.2 ± 0.2 ng mL−1, respectively) and EEL (0.8 ± 0.2 and 1.0 ± 0.2 ng mL−1, respectively) groups. Receiver operating characteristic analysis indicated that OAS-1, Doppler, and PAGs were significant (P < 0.01) predictors of pregnancy in heifers (AUC = 0.86, 0.92, and 0.94, respectively) and cows (AUC = 0.82, 0.94, and 0.95, respectively). The PAGs and Doppler-US presented higher accuracy on diagnosing pregnancy (92 and 89% for heifers, 94 and 93% for cows, respectively) than OAS-1 (81% for heifers and 75% for cows). Doppler-US was 100% sensitive to detect nonpregnant females, which avoided false negative results, whereas PAG concentrations presented a higher specificity (86% for heifers and cows) when compared with Doppler-US and OAS-1 (76 and 65% for heifers, 85 and 75% for cows, respectively). On D25, PAGs were 90% accurate to detect heifers and cows that would experience EEL, whereas on D20, OAS-1 was 50% accurate. In conclusion, Doppler-US can be used as early as D20 with high accuracy and sensitivity. Likely, the use of PAG concentrations on D25 is highly accurate and is more effective at detecting EEL. Pregnancy diagnosis through OAS-1 expression is not as accurate as the other methods and could not efficiently predict EEL. This research was supported by FAPESP (2015/106069; 2017/134729).