The importance of the diestrus stages (early, middle, late) to the diagnosis of the endometrial disease in the bitch

The endometrium is the uterine layer wherein embryo implantation and development take place. Complete knowledge of the endometrium’s normal histology is critical for advancing the diagnosis of subclinical endometrial diseases. Several researchers studied subclinical endometritis (END) in different species, but the cause/s remains unknown. Furthermore, simple and reliable diagnostic tools are still necessary to diagnose subclinical END. In the bitch (Canis lupus familiaris), diestrus has been selected as the stage of the reproductive cycle to perform the diagnosis of END. This work aimed to perform morphometric and cytological studies in bitches with normal endometrium (NE) and END in diestrus’ stages. Forty-five mixed-breed clinically healthy bitches, in diestrus were used in this study. Biopsies and cytology were obtained from uterine horns, processed, and examined using a microscope. Samples from NE and END were selected, images acquired, and software for image analysis was used to measure the endometrial thickness (ET), external glandular diameter (EGD), internal glandular diameter (IGD), glandular density, and glandular cell height (CG). Significant differences were observed in GC height, EGD, IGD, and serum progesterone (P4) concentrations along diestrus. In END, both EGD and IGD were smaller than NE. The current study showed histomorphometric characteristics that appear to be influenced by P4 concentrations. In our opinion, it is relevant to consider how far the bitch had progressed into diestrus at the moment of the diagnosis of endometrial disease. Data recorded in the present study could be helpful to clarify cases of bitches with unexplained infertility or subfertility.


Introduction
Histologically, the uterus comprises of three layers, the endometrium, myometrium, and perimetrium. The endometrium plays a significant role in the reproductive efficiency of a female, and its normal structure and function are essential to achieving optimal implantation rates (Dimitriadis et al. 2005). This fact makes a healthy uterus and its endometrial integrity the basis for high pregnancy rates and litter size. Several uterine conditions, which are usually difficult to identify, can affect the endometrial integrity and impact implantation. Histological endometritis (END) has been defined as inflammation of the all-endometrial layers visible on microscopic examination, but without apparent clinical signs of disease (Praderio et al. 2019(Praderio et al. , 2021Bukowska et al. 2020). It is well known that histological END could be related to subfertility or infertility in the bitch (Fontaine et al. 2009;Schlafer 2012;Mir et al. 2013;Gifford et al. 2014;Mitacek et al. 2020). A uterine biopsy as a diagnostic tool has been widely used in Veterinary Medicine and has been helpful for the diagnosis of END in mares, cows, and bitches (Kenney and Doig 1986;Mir et al. 2013;Madoz et al. 2014;Gifford et al. 2014;García Mitacek et al. 2017;Praderio et al. 2019Praderio et al. , 2022. In the same way, endometrial cytology has been widely used to diagnose END in cows and mares (Overbeck et al. 2013;Madoz et al. 2014). However, a low degree of agreement was observed between the diagnostic outputs of uterine biopsy and endometrial cytology (Praderio et al. 2019). In this way, Bukowska et al. (2020) found * María Alejandra Stornelli astornel@fcv.unlp.edu.ar a low degree of correlation between the results obtained by endometrial cytology and the biopsy in subclinical uterine diseases. The morphometry allows objective evaluation of the variations of the cell morphology. The importance of quantitative pathological analysis of cells and tissues has been reported to provide important diagnostic and prognostic information (Baak 1987). Uterine morphometric studies could elucidate endometrial changes related to uterine diseases. Although in canines, there are several studies on morphometry in different tissues, there are few studies on uterine morphometry (Monteiro et al. 2009;Groppetti et al. 2010;Salinas et al. 2017;Gossler et al. 2017), and there are no morphometric studies on END. In this sense, histomorphometric studies on the endometrium could offer exciting data to use as a tool in endometrial diagnosis. The histomorphology of the canine endometrium changes with estrus cycle, according to progesterone (P 4 ) and estrogen (E 2 ) concentrations (Galabova et al. 2003). In the bitch, the follicular phase is followed by a prolonged luteal phase in the nonpregnant bitch. The luteal phase lasts approximately the same time in nonpregnant and pregnant females, and up to now, a luteolytic mechanism is not known in the nonpregnant bitch (Concannon 2011). Therefore, knowing that diestrus is prolonged in non-pregnant bitch, it could be interesting to study the endometrial changes in the different diestrus stages (early, middle, and late). Studying the endometrial cytology in the different diestrus stages could improve the END diagnosis. This work aimed to perform morphometric and cytological studies in normal endometrium (NE) and END in the early, middle, and late diestrus in the bitch. We hypothesized that: (1) normal endometrium shows morphometric differences compared with END in the bitch, and (2) the endometrial morphometric features and the number of polymorphonuclear neutrophilic leukocytes (PMNL) on the endometrial cytology change during diestrus in NE and END.

Animals
Of the total of 84 bitches sampled, 45 mixed-breeds clinically healthy bitches (8-72 months; 5-27 kg) in diestrus, without a history of progestogen administration, lack of vaginal discharges, without macroscopically changes observed in the uterus obtained at the ovariohysterectomy, and with the histologic image of normal endometrium or END were included in this study. Bitches with a history of progestogen administration (n = 8), without data about progestogen administration (n = 22), or with pathologic changes different to END (n = 9) were excluded from the study. All bitches belonged to a breeding control program at the municipal pet public shelter in La Plata.

Confirming of diestrus
The diestrus was confirmed by vaginal cytology, corpora lutea observed macroscopically after ovariohysterectomy, and serum P 4 concentrations. Based on the date of the end of the last heat, diestrus was divided into three stages: early diestrus (EDI, until 20 days after the end of the heat [DEH]), middle diestrus (MDI, between 21 and 40 days after DEH), and late diestrus (LDI, > 40 days after DEH; Fig. 1). Serum P 4 concentrations were measured by chemiluminescence immunoassay (Elecsys ® , Progesterone II; Roche, Mannheim, Germany).

Endometrial samples (biopsy and cytology)
One endometrial cytology sample and a biopsy were taken from the middle part of each uterine horn. Cytology samples were collected with a cytobrush rolling into the endometrium, rotated on a slide, stained (T15; Biopur S.R.L., Rosario, Argentina), and examined under a microscope at X400 to determine the percentages of PMNL in 200 cells. Biopsies were fixed in 10% buffered formalin solution, paraffin-embedded, sectioned, stained with H&E, examined under a microscope, and categorized as indicated Praderio et al. (2019). Then, samples with END and NE were selected. Normal endometrium samples were divided into three groups (G) based on whether it was early, middle or late diestrus; GN1: samples from bitches in EDI; n = 4, GN2: in MDI; n = 4, GN3: in LDI; n = 6. The same criteria were used to classify END samples; GEND1: samples from bitches in EDI; n = 5, GEND2: in MDI; n = 9, GEND3: in LDI; n = 17.

Computerized morphometric analysis
Endometrial samples were examined with light microscopy (Nikon Eclipse 50i), and photographs were taken using the same microscope equipped with a Nikon DS-Fi2 digital camera. Additionally, the ImageJ program (National Center for Microscopy and Imaging Research: ImageJ Mosaic Plugins, RRID: SCR_001935) was used to perform measurements on endometrial features (Fig. 2). In each sample, 10 endometrial thickness measurements (ET) were performed at X40 magnifications (Acuña et al. 2020). Additionally, observing at X600 magnifications, randomly selected, the height of 10 glandular cells (GC) per field in 5 fields of the endometrium, was recorded. Also, in an average of 7 crosssection glands per sample, the external glandular diameter (EGD) and internal glandular diameter (IGD) were recorded, each in two perpendicular measurements. In addition, the number of glands in 5 fields was registered for evaluating glandular density, expressed as glands per field.

Statistical analysis
Linear regression models with spatial repeated measures were fitted to assess the effect of endometrium category (NE or END) and diestrus categories on GC height, EGD, IGD, glandular density, and ET (Proc GLIMMIX, SAS ® 9.4). The results of fitted models are reported as Least Square Means ± SE. The level of significance was set at p < 0.05.

Discussion
A previous study was the first to perform morphometry on normal and pathologic uterus in different cycle stages and shows its relevance in diagnosing endometrial diseases (Groppetti et al. 2010). However, to our knowledge, this is the first histomorphometric study to report findings on the endometrium in the early, middle, and late diestrus in clinically healthy bitches with and without END. Our results showed by linear regression analysis that GC height, EGD, IGD, and serum P 4 concentration change along diestrus in NE and END. However, ET did not change with the diestrus stages in NE, according to Gossler et al. (2017) but changed in the bitches with END (Fig. 4). As in previous studies, serum P 4 concentration decreased along diestrus (Galabova et al. 2003;Reusche et al. 2018;Stornelli et al. 2020). In bitches, strong individual differences were reported in serum P 4 concentrations (Concannon et al. 1975;Luz et al. 2006;Stornelli et al. 2020). In the current study, in the normal group, the serum P 4 concentrations in EDI and MDI were higher than in other studies (Galabova et al. 2003;Stornelli et al. 2020). These differences could be due to the different hormonal determination methods used in other studies and the individual differences between bitches. Galabova et al. (2003) previously described morphologic changes in canine NE during diestrus. The current study showed that in NE, GC height and EGD decrease when P 4 decreases, while IGD increases ( Fig. 1B-D). Several studies reported the action of P 4 and progestogen on the canine uterus (Sokolowski and Zimbelman 1974;Noakes et al. 2001;Hagman 2018). Salinas et al. (2017) observed differences in the ET between nulliparous treated with medroxyprogesterone and untreated bitches, thus confirming the effect of the progestogen on ET. In the same way, we observed differences in GC height, EGD, and IGD along diestrus with different serum P 4 concentrations. In the END group, the endometrium was thinner in late diestrus than in early and middle, likewise that serum P 4 concentrations. However, this difference in the ET was not observed in the normal group. This observation could be related to the fact that endometrial alterations result in the deregulation of apoptosis, producing insufficient endometrial regeneration (Reusche et al. 2018). In this sense, studying endometrial apoptosis in bitches with END could be relevant as a future line of investigation. In addition, a study on cows showed that ET measured by ultrasonography increased when P 4 concentrations decreased (Sugiura et al. 2018). Similarly, a recent study on cattle evaluating ET to improve the detection of nonpregnant females post-AI showed that ET is thicker in nonpregnant cows because they return to estrus and have lower P 4 concentrations (Madoz et al. 2022). Is possible that the influence of P 4 on ET be different in normal than pathologic endometrium. The deliveries, age, and uterine size influence the NE´s ET in mammals (Suguria et al. 2018;Binder et al. 2019). Similar to feline and in contrast to bovine in canine endometrium, characteristics of both, endometrial cells and glands, increase when P 4 increases (Chatdarong et al. 2005;Martí et al. 2021).
Although several studies have been carried out on END, there is no complete understanding of the etiopathogenesis of this disease (England et al. 2021). A study reported that epithelial heigh, glandular diameter, and density were greater in susceptible mares than in resistant to persistent post-breeding END during estrus (Herrera et al. 2018). In contrast, in the current study, ET, GC height, and glandular density did not change between NE and END. Also, the EGD and IGD were lower in END than NE. The differences observed between the current study and Herrera et al. could be due to stages of the reproductive cycle and the species studied (estrus in mares vs. diestrus in bitches). The differences reported between NE and END indicate that endometrium is histologically different and, thus, could probably be functionally distinct. In the END, the endometrial remodeling could be altered, affecting canine fertility. So, endometrial biopsy has a relevant role in advancing the etiopathogenesis of END and should be used by clinicians, and not only in an investigation study. In this way, the reproductive clinical practice could be improved, and the reproductive efficiency of the bitches too.
Concerning endometrial cytology, our findings showed no significant differences in the percentage of PMNL along diestrus, but a tendency was observed. In the endometrium are immune system cells such as lymphocytes, macrophages, plasma cells, and PMNL (Segerson et al. 1991;Gottshall and Hansen 1992). The role of endometrial cytology as a valuable tool for diagnosing END in the bitch is controversial (Fontaine et al. 2009;Praderio et al. 2019;Bukowska et al. 2020). Fontaine et al. (2009) reported END by endometrial cytology in bitches at different cycle stages. Another study showed no agreement between endometrial cytology and uterine biopsy to diagnose END in bitches in diestrus (Praderio et al. 2019). The authors compared the percentage of PMNL between NE and END groups in diestrus and concluded that endometrial cytology would not seem helpful in diagnosing END in the bitch. However, the diestrus stages (early, middle, late) were not considered in that study. The originality of the current research about endometrial cytology is that the percentage of PMNL along diestrus´ stages (early, middle, and late) was studied. It is known that E 2 increases uterine blood flow (Dickson et al. 1969) and might contribute to the increasing number of cells infiltrating the endometrium. In contrast, P 4 is responsible for to be able slowing the infiltration of leukocytes into the uterus at diestrus (Matsuda et al. 1985). In this sense, it is possible that, as occurs in vaginal cytology, PMNL decrease as diestrus advances. In this investigation, we could not show significant differences in the number of PMNL along the diestrus. However, a tendency was observed between early and late diestrus. It is for this reason that more research is needed.
One of the limitations of this study is the low number of animals in some groups. Future studies which include a greater number of animals could demonstrate with greater precision the effect of the P 4 and END on the glands, and too could evidence the differences in the percentage of PMNL on endometrial cytology in early, middle and late diestrus. Finally, our work was conducted with samples obtained of bitches after ovariohysterectomy, and this was another limitation since we could not study fertility in those cases. So, to carry out studies on this topic in vivo and then evaluate the fertility of the bitches studied would be a great advance in our area to really know the impact of histological END on canine fertility.
In summary, the endometrial morphometry change when the diestrus advance and serum P 4 concentration decreases. Also, EGD and IGD are lower in END than NE. More studies on histological END are necessary to understand its pathogenesis and perform therapeutic protocols. So, data from this study provide new information on canine endometrial morphology. This study's contribution is essential at scientific and clinical levels since it confirmed that morphometry could be a valuable tool for comparing NE and END in infertile bitches. Also, this investigation shows that it is essential not to consider diestrus as a single stage in the bitch since it is long and very different at the beginning and end.

Conclusions
In conclusion, based on our results, the authors suggest considering the moment of diestrus (early, middle, late) to diagnose endometrial disease since the endometrium changes along diestrus. So, regarding to diestrus, without considering these differences could hinder the diagnosis of endometrial disease.