Traditionally, most experiments in the field of equine reproduction have been performed in living animals [1]. Given the fact that the horses do not serve as a model, but are the target animal species supposed to benefit from this research, this procedure has had many advantages. Above all, it has allowed researchers to investigate physiological and pathological processes in the complex environment of an individual organism. In parallel, the application of the 3R principles (reduction, refinement and replacement of animal experimentation) [2] and new technological advancements have accelerated the development of other research models in past years. Among those, extracorporeal organ perfusion models have been established, including ex vivo models of human [3], bovine [4], porcine [5] and, only recently, equine [6] uteri. Thus far, no studies have investigated the stability and correlation to protein abundance in isolated perfused uteri.

In equine samples from in vivo experiments, quantitative real-time PCR (RT-qPCR) and immunohistochemistry have been used concurrently for assessment of endometrial stability and functionality for almost 20 years [7]. In ex vivo perfusion of transgenic swine uteri, both techniques have successfully been employed for determining gene expression [8]. Endometrial morphology is largely affected by the cycle stage-dependent presence of progesterone and estrogen [9,10], the presence and functionality of two nuclear steroid hormone receptors - progesterone receptor (gene: PGR; protein: PR) and estrogen receptor-α (gene: ESR1; protein: ERα) - are essential for the physiological function of the endometrium. Hence, their expression throughout the estrous cycle has been studied extensively in endometrial biopsies of living animals [[11], [12], [13]]. Both hormones affect the expression of endometrial oxytocin receptor (OXTR), which is important for the endometrial secretion of PGF2α during luteolysis [14]. In addition to receptors, the expression levels of two other genes, coding for Ki-67 and Caspase 3, have been assessed for determination of proliferation and vitality [11,15,16]. While Ki-67 is a well-known marker of cell proliferation, being expressed in all but the G0 stage of the cell cycle [17], the activated form of Caspase 3 is a well-established marker for apoptosis [16,18]. The balance of both processes (proliferation and apoptosis) is important for the homeostasis of endometrial tissue.

Recent advances in digital histopathology provide the possibility to automatically detect, quantify and characterize cells with software tools [19]. The open source software QuPath [20] has been used for semi-automatic identification of cells and staining patterns in various equine tissues, including granulation [21] and lamellar tissue [22]. While there are no current examples of using QuPath in equine reproductive studies, it has successfully been employed in human gynecology for the computer-assisted evaluation of endometriosis lesions [19], placental tissue [23] and endometrial biopsies [24]. Incorporation of this open source software into pathohistological examination of equine endometrial biopsies thus provides potential for a faster and possibly more objective tissue sample analysis.

The aims of the present study were (i) to evaluate the morphological stability and functionality of isolated-hemoperfused equine uteri at the molecular level for further validation of the ex vivo model and (ii) to compare the aptness of the open source software QuPath and to traditional visual analysis in identifying endometrial cells and scoring hormone receptor (PR, ERα) as well as marker of proliferation (Ki-67) and apoptosis (Casp3) staining.