Endometriosis is characterized by the presence of uterine stroma and glands outside the uterine cavity, predominantly on the pelvic peritoneum and ovaries. It affects approximately 5–10% of women of reproductive age and represent a significant disease burden (Shah and Jagani, 2018; Simitsidellis et al., 2018; Zondervan et al., 2018). While pelvic and abdominal pain symptoms (e.g., chronic pelvic pain, dysmenorrhea, dyspareunia) and subfertility are the most common clinical signs of endometriosis (Mechsner, 2016), many relevant co-morbidities with affective disorders like depression or anxiety are described and have severe impact on quality of life of patients (Brandes, 2007; Facchin et al., 2015, 2017). The association between endometriosis and mental illnesses has been reported by several studies and confirms a significant incidence of anxiety, depression, and psychopathological symptoms among women with endometriosis [for review, (Laganà et al., 2017)].

Various animal models have been established for the identification and development of drugs for the treatment of affective disorders. In such models, a behavioral characterization of complex cognitive and//or emotional processes is required. For example, the forced swimming test and the sucrose preference test have been widely used to evaluate depression-like behaviors, and the open field test, the novelty suppressed feeding, and the elevated plus maze (EPM) test, have been used to evaluate anxiety-related behaviors. Outcomes such as behavioral despair, immobility responses, anhedonia; as well as locomotion, thigmotaxis, latency to approach food and/or food intake, and time spent in open/illuminated vs. closed/dark areas are quantified. The antidepressant and anxiolytic activities of potential therapeutic agents can be evaluated by the degree of reduction of such behavioral responses upon agent administration (Belovicova et al., 2017).

In endometriosis research it is crucial to understand the complex pathophysiology of endometriosis as well as the identification and development of new therapeutic options [for review: (Simitsidellis et al., 2018)]. The availability of rodent models, their validity for translational research (e.g., construct, face, and predictive validity) and the quality of their characterization are of evident importance. A recent literature survey of preclinical models for endometriosis shown that the most common outcome measures are biological and histopathological evaluations of the lesions and identification of inflammatory markers. In strong contrast, the behavioral outcomes related with pain have been rarely studied and are fundamental to improve the validity of these models for the effective development of new therapies for endometriosis (Nunez-Badinez et al., 2021). In the present study, we aim to characterize the behavioral phenotype resulting from surgery-induced endometriosis in mice, using established anxiety and pain assays.

Endometriosis, characterized by the presence and growth of endometrial tissue outside the uterine cavity, often cause pelvic and abdominal pain in patients, subfertility and a marked reduction in health-related quality of life (Imboden and Mueller, 2018; Zondervan et al., 2018). Moreover, co-morbidities with affective disorders like depression or anxiety have also been described (Laganà et al., 2017). Several studies using endometriosis rodent models are demonstrating biological and histopathological similarities of these models with the human situation, but their behavioral characterization was never performed. To elucidate how the induction of experimental endometriosis might affect pain sensation and anxiety in mice, we performed two studies: In the first study both pain- and anxiety-related behaviors were assessed; and the second study focused on anxiety-related behaviors. Our main findings are that EIM versus sham group do not show significant signs of generalized mechanical- nor thermal-induced pain states; and second, despite this, EIM do display anxiety-related behaviors.

The quality of an animal model for translational research relies on three different criteria: construct-, face-, and predictive validity (Vervliet and Raes, 2013), meaning that (1) the etiology (2) the symptoms/hallmarks observed, and (3) the effect of treatment with standard of care should be, ideally, directly translatable to the original disease. In endometriosis, many theories about its pathogenesis have been proposed (Zondervan et al., 2018) and therefore, the construct validity of the animal models relies on them. Retrograde menstruation is a physiological process in which menstrual tissue flows backwards via fallopian tubes and deposits tissue in the intraperitoneal space, providing a source for the development of endometriotic lesions. It is hypothesized that this process contributes to the development of endometriosis. Additionally, factors causing a high endometrial flow (e.g., early menarche, no pregnancies, late menopause) are discussed as risk factors in woman to develop endometriosis. Transplanting endometrial tissue samples onto peritoneum and mesenteric arteries intends to model the consequences of retrograde menstruation and high endometrial flow. In the syngeneic transplantation model, this is achieved by mostly two means: suture or intraperitoneal inoculation of endometrial biopsies from donor into recipient mice. The latter has been proposed as having a higher construct validity as the intervention is less invasive and mimics retrograde menstruation (Nunez-Badinez et al., 2021). Nevertheless, both methods induce endometriotic-like lesions with dense vascularization, adhesions and secretion of cytokines in the implanted areas and surroundings (Somigliana et al., 1999; Bordignon et al., 2013; Simitsidellis et al., 2018). Furthermore, antiangiogenic therapy eradicates the lesions (Dabrosin et al., 2002), supporting predictive validity (Edwards et al., 2013). However, other important symptoms/hallmarks of endometriosis, like anxiety or depression (Laganà et al., 2017) have not yet been evaluated; therefore, the face validity is not well characterized. We will discuss studies that have measured pain and anxiety in syngeneic models of endometriosis and compare their results with ours.

The incidence of anxiety and depression traits in endometriosis patients has been reported in many clinical studies (Sepulcri and Amaral, 2009; Facchin et al., 2015, 2017; Friedl et al., 2015; Laganà et al., 2015; Chen et al., 2016) and reviewed (Pope et al., 2015; Laganà et al., 2017). Chen et al. (2016) reported that endometriosis patients have an increased risk of developing major depression, any depressive disorder, and anxiety disorders in later life compared to age and sex-matched controls. Additionally, depressive and anxiety disorders accounted for the most common comorbid conditions in endometriosis (Pope et al., 2015).

In rodents, the EPM, the light-dark box and the open field test belong to the most used anxiety-related defense behavior assays (Mohammad et al., 2016). To our knowledge, there is only one study that has measured anxiety-related behaviors in EIM using the syngeneic transplantation model (Li et al., 2018). Li and colleagues have observed that endometriosis mice show a decreased time and distance traveled at 6, 8, 10, and 12 weeks after surgery compared to controls in the open field test, while locomotor function was unaffected. These observations demonstrate anxiety-like behavior in EIM and concur with our observations on a decreased time spent in the open arms of the EPM 5 weeks post-surgery without alterations in locomotive function. However, some differences in the methodology need to be highlighted: we performed two independent studies assessing anxiety at the same time post-surgery. In contrast, Li and colleagues measured anxiety in the same cohort of animals every 2 weeks post-surgery, which may lead to habituation and in consequence, a decrease in explorative behavior. Still, the distance traveled by both groups appeared stable along the time points measured. In our study, the percentage of time spent in the exposed region fully coincides with the average observed in a meta-analysis of preclinical models of anxiety, being 18% of total time in the exposed area a standard measure of anxiety (Mohammad et al., 2016).

Dysregulation of food consumption is another maladaptive response in anxiety disorders, being hypophagia, an increase in latency to begin feeding and decrease in the amount of ingested food, one trait in rodent models of anxiety that can be tested by assays like NEIFS (Cryan and Sweeney, 2011). In our observations, EIM tended to higher latencies to approach a palatable food compared to sham, which may argue for a mild anxiety-related behavior. However, this effect was maintained at day 5, when animals were back in their home cages, which may also argue for a generalized alteration of food motivation. Up to date we have not found other studies assessing NEIFS in endometriosis research. Since pharmacological administration of anxiolytics successfully reduces the latency to approach palatable food in anxiety-induced mice (Rex et al., 1996), it would be interesting to see if anxiolytic treatment would reduce these behaviors in EIM.

It is unclear which mechanisms may explain the occurrence of anxiety-related behavior in EIM. Sustained local inflammation in the endometriotic lesions and/or peritoneal cavity as described for women suffering from endometriosis is discussed as main driver for activation of peripheral nerve endings leading to chronic pelvic pain (Laux-Biehlmann et al., 2015b; McKinnon et al., 2015). Additionally, in the current ICD framework, endometriosis is characterized as “chronic visceral pain from persistent inflammation in the pelvic region” (Aziz et al., 2019), and low-grade chronic inflammation has been attributed to anxiety and depression in patients (Leonard, 2018). Kalfas et al. (2022), described a strong association between anxiety and pelvic pain in endometriosis patients. While we were not able to show generalized pain, it cannot be ruled out that EIM suffered from low chronic pelvic pain. It was shown recently, that induction of endometriosis in a syngeneic inoculation model leads to chronic abdominal pain (Greaves et al., 2017). The observed effects were associated with an over-representation of the PGE2 signaling pathway and nociceptive ion channels in endometriosis lesions of the animals suggesting ongoing inflammation. In context of other pain models it was shown that chronic pain conditions lead to anxiety-like behavior in rodents quantified in EPM (Caspani et al., 2014). Taken together, our model might connect both lines of evidence, showing consequently that sustained inflammation in the mouse endometriosis model associates with increased anxiety and might provide an approach with relevant face validity for this endometriosis-related co-morbidity.

On molecular level, both suture and inoculation models of endometriosis induced a pro-inflammatory state with increased concentrations of pro-inflammatory cytokines IL-1β and/or TNF-α (Cao et al., 2004; Zhou et al., 2010; Laux-Biehlmann et al., 2015a; Boyken et al., 2016; Mishra et al., 2020). These inflammatory processes might be the driver of the observed anxiety-related behavior, given that the pro-inflammatory cytokine IL-1β has demonstrated anxiogenic effects in mice (Munshi et al., 2019), and inflammatory processes including upregulation of IL-1b was described in peritoneal fluid of endometriosis patients (Mori et al., 1991). Another possibility is that endometriosis-pain chronically activate the hypothalamus-pituitary-adrenal (HPA) axis, leading to permanently increased cortisol levels and anxiety, as observed in endometriosis patients (Masciullo et al., 2021).

Interestingly, our results indicate the existence of anxiety-related behaviors in EIM without observing differences on generalized thermal- nor mechanical-pain. Other in vivo studies in surgically induced syngeneic models of endometriosis have observed generalized hyperalgesia, determined by decreases in hind paw withdrawal latencies to heat stimulation (Guo et al., 2013, 2015; Long et al., 2016, 2019; Ding et al., 2018) or thresholds to mechanical stimulation (Greaves et al., 2017), contrasting our observations. One possibility for this discrepancy is that the surgery was too invasive to distinguish pain thresholds between groups, lowering MPT in sham mice as well, and masking the pain-related behavior from the EIM. Another possibility is that EIM only develop local hyperalgesia in the abdominal/pelvic area. In this study, abdominal thresholds were only measured at the end of the experiment, and no differences were observed between groups (data not shown). We do not exclude that other type of endometriosis models would develop generalized hyperalgesia in EIM mice, or longer lasting local hyperalgesia in the abdominal area.

Others have observed local abdominal hyperalgesia to noxious mechanical (Greaves et al., 2017) or chemical (Tian et al., 2018) stimulation in preclinical models of endometriosis that can be attenuated following treatment (Guo et al., 2015; Greaves et al., 2017; Ding et al., 2018; Tian et al., 2018; Long et al., 2019). For example, Greaves et al. (2017) observed abdominal mechanical hyperalgesia in a syngeneic inoculation model 21–22 days post-inoculation. Authors performed an intraperitoneal injection of tissue derived from decidualized donor mice into an ovariectomized recipient mice supplemented with estradiol. Our study used surgical sutures of uterine horn biopsies into the mesentery and abdominal wall in normal cycling mice.

Measures of spontaneous, observer-independent pain behaviors are recommended to complement evoked-pain behaviors (Tappe-Theodor and Kuner, 2014). Although we did not observe differences on locomotor activity in the time points tested, it would be interesting to measure this in a longitudinal study. Furthermore, analysis of ultrasonic vocalizations or continuous monitoring of rodents’ behavior in their home cages have emerged as attractive alternatives to reveal the affective/emotional state of rodents as well as behaviors otherwise impossible to observe and have to potential to increase reproducibility and internal validity in translational research.

In summary, in a mouse model of endometriosis we observed increased anxiety-like behaviors, suggesting that the presence of endometriotic lesions in the abdominal cavity could, similarly to patients, induce psychopathological impairments in mice behavior. These endpoints might support identification of mechanisms relevant for the development of endometriosis-related comorbidities.

The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

The animal study was reviewed and approved by American Preclinical Services Institutional Animal Care and Use Committee (IACUC).

JN and TZ contributed to the conceptual design of experiments and data analysis. MH and OB contributed to the execution of the experiments. AL-B and PN-B contributed to data analyses and creation of figures. All authors contributed to manuscript writing and revision and approved the submitted version.