The sexually transmitted disease (STD) gonorrhea caused by N. gonorrhoeae is the second most common STD which causes around 86.9 million new infections annually around the world, as estimated for 2016 (Rowley et al., 2019). The rapid increase in antibiotic resistance makes gonorrhea a serious threat to public health worldwide (Unemo and Shafer, 2011; Rubin et al., 2020). The mucosal surface of the female cervix and male urethra, anorectal, pharyngeal, and conjunctival areas are the most common sites of N. gonorrhoeae infection (Miller, 2006). N. gonorrhoeae can pass into the deeper tissue layers and reach the bloodstream, causing systemic disseminated gonococcal infections (DGI) in 0.5% to 3% of patients, which leads to endocarditis, meningitis, peri hepatitis, and permanent joint damage (Knapp and Holmes, 1975; Cannon et al., 1983; Kerle et al., 1992; Hansen et al., 2014).

History of gonococcal infection research traces back to the use of animal models of rabbits, guinea pigs, hamsters, mice, and chimpanzees (Miller et al., 1945; Lucas et al., 1971; Arko, 1972) ( Figure 1 ). Finding a proper model for gonorrhea research, however, has always been a challenge since N. gonorrhoeae is a human-specific pathogen. Several in vitro cell models have been established using various cell sources including cell lines (cancer and immortalized) and primary cells (Shaw and Falkow, 1988; Harvey et al., 1997; Giardina et al., 1998; Scheuerpflug et al., 1999; Christodoulides et al., 2000; Edwards et al., 2000; Hopper et al., 2000; Fichorova et al., 2002). Apart from animal models, a variety of approaches have been used to recreate the 3D structure of the site of infection as well as pathogen penetration and immune cell transmigration. These have included: Transwell^® insert technology (Merz et al., 1996; Edwards et al., 2000; Stevens et al., 2018), decellularized scaffolds (Heydarian et al., 2019), as well as organ cultures derived from various sites within the urogenital tract (Carney and Taylor-Robinson, 1973; Tjia et al., 1988; Mosleh et al., 1997; Timmerman et al., 2005; Yu et al., 2019). The present review focuses on the established in vivo and in vitro tissue models used in investigations of the infection caused by gonococci. The recent advances in gonorrhea disease research using animal and in vitro/ex vivo tissue models are also reviewed. Finally, we discuss the main challenges and obstacles, followed by future perspectives in the field of models used for studying gonococcal infection.

Human male volunteers have been used as experimental subjects (Cohen et al., 1994; Hobbs et al., 2013) and have helped understand the function of various virulence factors of gonococci as well as the immune response to infection, which can support vaccine research [reviewed in (Hobbs et al., 2011)]. Studies using human-derived source material have provided important insights into the pathogenesis of gonococcal infection in men (Ward and Watt, 1972; Apicella et al., 1996). Nevertheless, due to the ethical limitation, in-depth research of the mechanism of gonococcal infection has mostly relied on animal models, which provide a great tool for studying N. gonorrhoeae infection and the corresponding immune response, due to the presence of the intact immune system (Lucas et al., 1971; Arko, 1972; Arko, 1974).

Several non-human species studies, which used rabbits (Miller et al., 1945), guinea pigs (Novotny et al., 1978; Wong et al., 1979), hamsters (Arko, 1972), mice (Streeter and Corbeil, 1981; Johnson et al., 1989), chimpanzees (Lucas et al., 1971; Kraus et al., 1975), or recently greater wax moth larvae (Dijokaite et al., 2021) ( Table 1 ) have been performed to study the disease development and for drug discovery. In 1990, for the first time, Taylor-Robinson and colleagues performed successful colonization of mice with N. gonorrhoeae using estradiol-treated germ-free female mice (Taylor-Robinson et al., 1990). The 17β-estradiol treatment promotes the long-term colonization of N. gonorrhoeae by the expansion of the estrus phase. However, as during this treatment the overgrowth of commensal flora of mice can prevent the attachment of the gonococci, usage of germ-free mice or antibiotics enhances the success of the colonization by N. gonorrhoeae (Jerse, 1999; Hobbs et al., 2013; Kim et al., 2019). Depending on the type of estradiol treatment, the gonococci can persist in the murine genital tract from 10 to as long as 40 days, but the success of colonization and host response are dependent on the mouse strain used [reviewed in (Jerse et al., 2011)]. Usually, the infection is performed by intravaginal inoculation (Johnson et al., 1989; Taylor-Robinson et al., 1990; Jerse, 1999), or, for the modeling of the upper reproductive tract infection, by transcervical inoculation (Corbeil et al., 1978; Connolly et al., 2021). These mouse models have contributed to the better understanding of the immune response to gonococci (Song et al., 2008; Feinen et al., 2010), and the role of the MtrC–MtrD–MtrE efflux pump system in infection (Jerse et al., 2003; Warner et al., 2007) to name the few examples.

Certain similarities between the physiology of the human and mouse reproductive tract exist. The average vaginal pH in women is below 4.5, whereas the cervix has a pH in the range of 6.5 to 7.5 (Ng et al., 2018). In comparison, the average vaginal pH of 17β-estradiol-treated mice that are susceptible to N. gonorrhoeae infection is 6.6, similar to the pH of the human cervix, which is the primary site of infection for humans (Jerse, 1999; Muench et al., 2009). However, there are also some anatomical and endocrinological differences between mice and men [reviewed in (Cunha et al., 2019)]. For instance, the vaginal flora in mice and humans are dissimilar (Noguchi et al., 2003). Another reason that the mouse model fails to appropriately represent human physiology is the lack of certain host-specific receptors, which are the binding target of different virulence factors of N. gonorrhoeae (Jerse et al., 2011; Quillin and Seifert, 2018).

To address this issue, a variety of humanized transgenic mice has been generated. These express different receptors of gonococcal virulence factors, such as human (h)CEACAM receptors, to increase the similarity of the mouse model to humans (Eades-Perner et al., 1994; Gu et al., 2010; Muenzner et al., 2010; Li et al., 2011). Such models were used to show that CEACAM-binding gonococci colonize the urogenital tract of genetically modified mice by suppressing the exfoliation of mucosal cells (Muenzner et al., 2010; Muenzner and Hauck, 2020). Transgenic mice for hCEACAM1, hCEACAM5, and CEABAC2 were also successfully used to reproduce the differential expression of CEACAM1 and CEACAM5 on the surface of the epithelial cell monolayer of the upper and lower female genital tract. This enabled the studies of the effect of Opa-CEACAM mediated interactions during the gonococcal infection along the female genital tract (Islam et al., 2018). Another example of humanized mice models is the engrafting of the immunodeficient NSG mice with human CD34+ hematopoietic stem cells. These mice were used to study N. gonorrhoeae and HIV co-infection (Xu et al., 2018).

Although transgenic animals are promising models for studying the early stages of N. gonorrhoeae infection (Wang et al., 2008; Gu et al., 2010; Francis et al., 2018), they cannot fully recapitulate all the features of gonorrhea in humans (Wang et al., 2017a; Quillin and Seifert, 2018; Stevens and Criss, 2018). Nevertheless, only a few non-transgenic animal models such as chimpanzees were successfully infected by N. gonorrhoeae and showed a greater resemblance of the symptoms and host reaction (Kraus et al., 1975). However, the application of primate models is limited due to the low availability and high maintenance costs (DiGiacomo et al., 1977; Jerse, 1999). In conclusion, the most abundant animal models used in the field of gonorrhea research are estradiol-treated wild-type and transgenic mice expressing human receptors ( Table 1 ).

Over the past fifty years, a great majority of research on N. gonorrhoeae pathogenesis has been conducted using 2D cell culture of cell lines and primary cells ( Table 2 ). In these in vitro models, cells are normally cultured on cell culture well plates, with and without coating with extracellular matrix (ECM) proteins such as collagen and Matrigel (Harvey et al., 1997; Christodoulides et al., 2000; Simons et al., 2005). Commonly used epithelial cell lines include HeLa cells (Brodeur et al., 1977; Gray-Owen et al., 1997; Naumann et al., 1997), human endometrial adenocarcinoma cells (HEC-1-B) (Shaw and Falkow, 1988), epidermoid carcinoma cervix cells ME180 (Naumann et al., 1997), human colorectal carcinoma cells (T84) (Merz et al., 1996), human conjunctiva epithelial cells (Chang) (Grassmé et al., 1996), and human Asian endometrial adenocarcinoma cells (Ishikawa) (Rodríguez-Tirado et al., 2012).

In 1988, Shaw and Falkow reported the first in vitro monoculture model based on HEC-1-B cells for studying the invasion of N. gonorrhoeae. Here, the authors were able to document the invasion of gonococci, which was inhibited by cytohalasin D and was not affected by the state of piliation of the bacteria (Shaw and Falkow, 1988). N. gonorrhoeae can also be transmitted from infected mothers to their infants during birth, causing neonatal blindness (Rees et al., 1977; Woods, 2005; Quillin and Seifert, 2018). Therefore, adherence and invasion of gonococci were studied in the cultured human corneal epithelial cells, showing that PilC plays a central role in pilus-mediated adherence of the bacteria (Rudel et al., 1992). Human Chang cells were also used for studying the Opa-mediated invasion of N. gonorrhoeae (Grassmé et al., 1996). It is important to note that the Chang cell line, originally thought to be derived from human conjunctiva, was later shown to have been established through contamination with HeLa cells (Lavappa, 1978; Capes-Davis et al., 2010) similar to the Hep-2 cell line (Brodeur et al., 1977). Nevertheless, Chang cells are an important tool for studying the interaction of gonococci with epithelial cells if one is not exploring tissue type-specific effects. For example, the Chang cell line has been used by our department to show the necessity of neutral sphingomyelinase 2 activity for the PorB_IA-dependent invasion of N. gonorrhoeae under phosphate-free conditions (Faulstich et al., 2015).

Despite the advantages of immortalized cell lines, which include availability, easy handling, and expansion (Pan et al., 2009), they have their drawbacks, such as being derived from tumor tissues or showing significant changes from the original cell type caused by immortalization and prolonged cultivation. Alternatively, different in vitro models using primary cells or immortalized primary cells have been established. These include early attempts using human sperm (James-Holmquest et al., 1974) or the cells derived from the urethral epithelium (Harvey et al., 1997; Giardina et al., 1998; Zenni et al., 2000; Harvey et al., 2001; Edwards and Apicella, 2005), as well as the immortalized primary urethral cells (Harvey et al., 2002). Primary urethral epithelial cells were mostly used to provide a model system similar to the native tissues for studying the attachment and invasion of gonococci (Harvey et al., 2001), or cytokine response to infection (Harvey et al., 2002) but in some cases were used to confirm the validity of the observations made in tumor cell lines (Giardina et al., 1998). Primary cells derived from different parts of the female reproductive tract were also used in various studies. Among them are those of cervix (Edwards et al., 2000; Swanson et al., 2001; Edwards et al., 2002; Steichen et al., 2008; Edwards, 2010; Jennings et al., 2011), endometrium (Christodoulides et al., 2000; Timmerman et al., 2005), and fallopian tube (Swanson et al., 2001; Maisey et al., 2003; Morales et al., 2006; Reyes et al., 2007). Additionally, immortalized human cervical epithelial cells were introduced in 2009 and used in studies of gonococcal biofilm formation (Falsetta et al., 2009) as well as studies of N. gonorrhoeae interaction with cell surface glycans (Semchenko et al., 2019). Primary cell culture systems proved to be useful for the N. gonorrhoeae host cell interaction studies where the role of certain receptors was investigated or where the induction of immune response was studied, as well as for the ultrastructural analysis of infected cells. For example, some cancer cell lines such as HEC-1-B and HeLa were shown not to express CEACAM molecules, whereas primary cervical cells, but also ME180 cell line, did (Swanson et al., 2001). Primary cells are also a model of choice when various signaling or cell death pathways are investigated or in gene expression studies (Reyes et al., 2007), because tumor cell lines might be too divergent from the original tissues for these purposes. From the technical view, the fact that most primary cells can be passaged for only a few passages represents a significant drawback. The introduction of immortalized primary cells can offer a solution to this problem while providing a model system that is still closer to the native tissue than tumor cell lines.

Apart from colonizing epithelia, N. gonorrhoeae can cause systemic infections, due to its ability to reach deeper tissue layers and blood stream. During its traversal across tissues, it encounters a variety of non-epithelial cells; therefore, cells residing beneath the epithelium such as fibroblasts and endothelial cells have also been used to investigate host-pathogen interaction (Scheuerpflug et al., 1999; Fichorova et al., 2002). Fibroblasts are responsible for the structure and shape of the connective tissue by producing and secreting the ECM components such as collagen, glycosaminoglycans (GAGs), and proteoglycans (Le and Brown, 2012). In addition to epithelial cells, which are capable of innate responses to infection, including the production of complement components, prostaglandins, and cytokines, and in some cases are hormone-responsive (Schleimer et al., 2007; Hickey et al., 2011), fibroblasts also play a role in pathological and immune responses by producing and responding to various cytokines (Delves and Roitt, 1998; LeBleu and Neilson, 2020). In particular, cervical fibroblasts can change their biomechanical properties in response to different stimuli such as pro-inflammatory cytokines (e.g. IL-1β) and hormones (e.g. progesterone), resulting in differential remodeling of the ECM (Shukla et al., 2018). Studying N. gonorrhoeae infection in a monoculture model of murine fibroblasts showed an increase in the number of viable bacteria recovered from the infected cells up to 48 hours post-infection. This effect was not inhibited by washing, indicating that N. gonorrhoeae was attached to and/or internalized by fibroblast cells (Waitkins and Flynn, 1973). Furthermore, it has been shown that the activation of the phosphatidylcholine-specific phospholipase C (PC-PLC) and acid sphingomyelinase (ASM) by N. gonorrhoeae is crucial for the pathogen entry into the human fibroblasts (Grassmé et al., 1997).

Infection with N. gonorrhoeae triggers a potent innate immune response, which involves neutrophils (also known as polymorphonuclear leukocytes or PMNs). Several reports showed the presence of viable bacteria in gonorrheal exudates, indicating that PMNs are not able to completely clear gonococci (Johnson and Criss, 2011). Various approaches were used to study the N. gonorrhoeae-neutrophil interaction ( Table 3 ). A variety of neutrophil-like cell lines, such as differentiated human leukemia (HL-60) (Bauer et al., 1999; Pantelic et al., 2004), and human monocytic (THP-1) have been developed. These cell lines retain most of the features of neutrophils. For example, HL-60 cells induced by retinoic acid express CEACAM1 and can phagocyte gonococci (Pantelic et al., 2004) and have been shown to be an appropriate model for N. gonorrhoeae-PMN interaction (Chen and Seifert, 2011). However, HL-60 cells also lack specific granules and fail to show the antimicrobial activity associated with primary cells (Le Cabec et al., 1997). Human neutrophils isolated from blood as primary cell models helped understand the role of virulence factors in pathogen invasion and disruption of immune cell functions (Watt, 1970; Thongthai and Sawyer, 1973; Simons et al., 2005; Stohl et al., 2013; Johnson et al., 2015; Reimer et al., 2016). They were used to show that gonococci are resistant to engulfment by neutrophils and can multiply inside them (Simons et al., 2005), as well as that they can inhibit neutrophil apoptosis and activate the secretion of proinflammatory cytokines (Chen and Seifert, 2011).

The role of other immune cells in gonococcal infection has also been addressed in various studies ( Table 3 ). Macrophage models include murine RAW 264.7 and human U937 and THP-1 cell lines, as well as human monocyte-derived primary macrophages (MDMs), which can be differentiated in vitro towards different phenotypes, M1 or M2 [reviewed in (Escobar et al., 2018)]. U937, THP-1, and MDMs have been used to address the interaction of gonococci with macrophages and show that gonococcal porin can arrest phagosome maturation within the macrophages (Mosleh et al., 1998), which might be connected to the further observations that macrophages represent a niche for gonococcal replication (Château and Seifert, 2016), where bacteria occupy distinct subcellular niches when colonizing these immune cells (Ivanov et al., 2021). A series of works using mouse RAW 264.7 macrophage cell line (Escobar et al., 2013), MDMs (Ortiz et al., 2015), as well as dendritic cells (Zhu et al., 2012) has shown that gonococci induce tolerogenic phenotype in macrophages and dendritic cells so that they cannot induce the proliferation of T cells. More specifically, the role of different receptors on human dendritic cells, which included C-type lectins MGL and DC-SIGN, in the interaction with N. gonorrhoeae and induction of T cell response has been addressed using human monocyte-derived dendritic cells (van Vliet et al., 2009). Additionally, MDMs were used to demonstrate inflammatory pyroptosis induced by N. gonorrhoeae (Ritter and Genco, 2018), and human dendritic cells were used in studies of the connection between gonococcal and HIV-1 infection (Zhang et al., 2005). Lymphocyte models include the usage of chicken DT40 B cells, Jurkat cell line, and primary mouse and human lymphocytes ( Table 3 ). The results stemming from these studies mostly pointed in the direction of gonococcal suppression of T cell proliferation and B cell humoral response (Boulton and Gray-Owen, 2002; Pantelic et al., 2005; Lee et al., 2007; Lee et al., 2008; Liu et al., 2012). Gonococcal interaction with memory B cells, which results in a broad, T cell-independent immunoglobulin response (So et al., 2012) has also been recorded. Regarding immune cells, the findings obtained with cell lines were often confirmed using primary cells, which was necessary to validate the results. However, the availability of primary immune cells is greater than the one of primary epithelial cells. They are therefore more often used in modeling of the gonococcal interaction with the host immune system, their disadvantage being donor variability.

Although primary cells enable us to better mimic the site of infection, their application has been limited due to their short life span, low proliferation rate, cross-contamination during cell isolation, and heterogeneity (Unger et al., 2002; Quillin and Seifert, 2018). Another source of primary cells could be induced pluripotent stem cells (iPSCs). They have already been used to study the related Neisseria species, N. meningitidis, and have proved useful in providing novel insights into meningococcal pathogenesis (Martins Gomes et al., 2019). All in all, the monoculture (2D) cell models provide the basic characteristics of the site of infection and are useful for addressing certain questions. They, however, cannot fully replicate the cellular complexity of a 3D model (Duval et al., 2017; Hoarau-Véchot et al., 2018).

Many efforts have been undertaken to develop a suitable model for studying N. gonorrhoeae ( Figure 1 and Tables 1 - 4 ). Progress in the development of animal models and in vitro tissue models from simple 2D monoculture to complex 3D co-culture models has helped us to further our understanding of the pathogenesis of N. gonorrhoeae. Since the cells cannot form a multi-dimensional structure in 2D cell culture, 3D cell culture has emerged as an alternative to improve the cell microenvironment. Driven by the drawbacks of the animal models such as the lack of human receptors and anatomical differences, transgenic and hormone-treated animal models have been developed. Even though the ex vivo and in vitro models of tissues showed promising results in recapitulating the main characteristics of the site of infection in humans, each of these models has also its disadvantages. The transition between rather simple 2D monoculture to complex biomimetic co-culture and triple co-culture tissue models can assist us in further comprehending the crosstalk between epithelial cells and immune cells during gonococcal infection. In addition to the epithelial and immune cells, stromal cells have been shown to contribute to the growth of epithelial cells as well as to the tissue response to hormones (Arnold et al., 2001; Bläuer et al., 2005), indicating the importance of co-culturing models of epithelial cells with fibroblasts. Moreover, fibroblasts appeared to be a niche for N. gonorrhoeae internalization (Waitkins and Flynn, 1973) and were shown to play a crucial role in the long-term infection of the 3D co-culture model of epithelial/fibroblast cells (Heydarian et al., 2019), which further emphasizes the necessity of their presence when modeling tissues.

Apart from the cell type and origin, bioreactors and microfluidic platforms also play a key role in the increase of the degree of biomimicry of the tissue models, from static to dynamic. Until now, most of the investigations in the field of Neisseria infection have been performed under static conditions, which ignores the dynamics of the natural environment where infection takes place, such as the urethra of men, or blood stream. Recently, a microfluidic culture model of the human reproductive tract has been introduced, which simulated the endocrine loops between organ modules of the ovary, fallopian tube, uterus, cervix, and liver (Xiao et al., 2017). In addition, a variety of microfluidic systems have been developed to study neutrophil migration, neutrophil extracellular traps, and reactive oxygen species production (Agrawal et al., 2008; Boneschansker et al., 2014; Moussavi-Harami et al., 2016; Yang et al., 2017). Microfluidic devices work well for studying the immune response as they can be customized, require fewer reagents and material, and are ideal for working with primary human cells. One can control the spatiotemporal presentation of signaling molecules and the pathogen added to the system, making them applicable for studying complex interactions. Current developments in the microfluidics models allow one to incorporate live and intact pathogens into the models. This not only increases the significance of neutrophil response but also allows studying the direct interaction between neutrophils and a pathogen. The major advantage of microfluidics devices lies in their ability to allow single-cell analysis amidst neutrophil heterogeneity. Continuous development of such devices will allow extensive insights into signals controlling neutrophil function (Richardson et al., 2021).

Further developments of the tissue models would also include the introduction of the microbiome, which is present in various parts of the human female reproductive system (Chen et al., 2017). A correlation between the microbiome, menstrual cycle, and immune responses of the genital tract has been reported (Rönnqvist et al., 2006; Mirmonsef et al., 2011; Chen et al., 2017). Exploring host-microbiome interactions during the infection would help reveal the protective role of commensal microorganisms against the infection.

Recently developed 3D cell culture models such as organoids of the fallopian tube (Kessler et al., 2015), human endometrium (Boretto et al., 2017; Turco et al., 2017), and cervix (Maru et al., 2020; Chumduri et al., 2021) have opened new avenues in the field of investigations of the pathology of the female reproductive tract, including infection research (Alzamil et al., 2021). Human fallopian tube organoids have been successfully used for studying Chlamydia trachomatis infection (Kessler et al., 2019). Many organoids, however, are formed with the epithelial/luminal surface on the interior, which requires the infection to be performed either by microinjection, by reversion of the organoids, or by disruption of the organoids with reseeding after infection [reviewed in (Aguilar et al., 2021)]. In addition to being used as infection models, organoids can also be a source of primary cells for scaffold-based 3D tissue models, which helps to overcome the problems coupled with the infection of the organoids. This will allow us to take one more step towards obtaining high-fidelity tools for studying gonococcal infection under close to natural conditions. Considering the latest advancements in the generation and development of 3D tissue models, we can say that we are on a good way to achieving that goal.

All authors have written parts of the manuscript. MH, RR, and VK-P have generated tables and the figure. All authors contributed to the article and approved the submitted version.

This work was funded by the Deutsche Forschungsgemeinschaft (DFG) GRK 2157 “3D Tissue Models for Studying Microbial Infections by Human Pathogens” to VK-P. This publication was funded by the German Research Foundation (DFG) and the University of Wuerzburg in the funding program Open Access Publishing.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.