Screening for the ovarian toxicity of industrial chemicals - ovarian follicles and corpora lutea counts in rats, an inter- and intra-laboratory comparison.

The aim of this study was to identify the methodologies currently used by contract research organizations (CROs) for counting ovarian follicles and corpora lutea (CL) in the OECD Test Guideline (TG) 443 (Extended One-Generation Reproductive Toxicity, EOGRT) studies. The current study further sought to characterize variability and its sources in follicle and CL counts. A database of 73 EOGRT studies with information on methodology, reporting and numerical data of follicular and CL counts was assembled and analyzed, revealing that quantitative follicle enumeration was conducted in 97% of studies, while quantitative CL evaluation was performed in 59%. Only 27% of studies provided details on sampling procedures related to location within the ovary, section thickness, step length, number of steps as well as type of histological staining. The percentage coefficient of variation (%CV) for follicle counts in control females ranged from 7.7% to 78.5%, indicating high variability. The median study-specific %CVs for primordial follicles and for combined primordial and growing follicles were 36% and 33%, respectively, while the median %CV for CL was 29%. The intra‑laboratory %CV decreased, indicating improved precision of the follicle‑count method, when a greater number of sections were evaluated but was also influenced by the number of studies performed within the laboratory, based on the dataset included in this evaluation. While the ratios of primordial to growing follicles are not a requirement in the OECD TG 443, they can reveal differences in follicle identification if the variability in ratios between the studies is high. Accordingly, the high variability in the ratios across studies demonstrated that well-defined criteria for follicle identification and validation of antibodies are necessary. The overall conclusion is that the lack of information on follicle identification criteria, and inconsistencies in terminology and reporting practices further complicate the interpretation of results of follicle and CL counts in the EOGRT studies. Our results highlight that clarification of the ovarian follicle terminology and identification criteria at the OECD level as well as clarity on best practices for follicle and CL counting are essential to improve the consistency and reliability of ovarian toxicity testing in EOGRT studies.