Introduction

The eukaryotic host–microbe interrelationship accompanies multicellular organisms from the origin of eukaryotic cells as a consequence of endosymbiosis. Several hundred billion microorganisms, such as viruses, bacteria, and fungi, populate hosts and are collectively named microbiota. Many species are beneficial to their hosts and keep their immune systems in a state of readiness. To study host–microbe interactions and microbes’ influence on host fitness, the concept of the gnotobiotic animal was introduced. The words “gnotobiont” and “gnotobiotic” were derived from the Greek words “gnostos” and “biota” meaning known living flora and fauna in association with the host. In practice, the gnotobiotic animal is a germ-free or originally germ-free organism artificially populated by a known strain of microorganism. Germ-free organisms (axenic animals) can be obtained by Caesarean section, hysterectomy from their mothers, or by sterile embryo transfer into germ-free foster mothers. Depending on species, precocial animals can thereafter be reared in sterile isolators without their mothers, while altricial animals need germ-free foster mothers. The terms gnotobiotic, germ-free, and mono-associated models often overlap, because all have known microbial status (for gnotobiotic terminology and criteria, see, for example, Gordon and Pesti [1]).

Over the decades, the studies of gnotobiotic animals have demonstrated the appropriateness of differentially mono-associated gnotobionts for analyzing organism’s protective functions. These studies have focused mainly on the immune system [2, 3], structural and metabolic functions of the organism [4–6], and developmental aspects of vertebrate functional systems [7–10]. Furthermore, microbiome models have been adopted for studies on carcinogenesis [11–15]. One of the areas attracting considerable attention and utilizing the gnotobiotic animal model is that of the host–pathogen interrelationship, with studies covering the basic microbiological and immunological questions, vaccine development, and general questions on health–disease transition. The appeal of these gnotobiotic models is that they actually constitute “biologically pure” models, and the influence of external signals on the studied organism can be monitored without “interior noise” caused by that organism’s microbiome.

Such bacterial intracellular pathogens as mycobacteria, brucellae, listeriae, francisellae, and salmonellae are studied assiduously, because their impact within the health care system or safety and security field can be considerable, and especially so if they were to be used for military or terroristic purposes. This brief compendium is, therefore, devoted to the issue of host–pathogen relationship studies conducted using gnotobiotic models, and particularly with reference to intracellular bacterial pathogens.

Historical perspective

The gnotobiotic concept originated in the requirement to work with strictly defined systems in biological experimentation. Various reviews on germ-free studies have reported the original concepts from the 19. Century and credited the ideas of J. B. Boussingault, E. Duclaux, L. Pasteur, and M. Schottelius [1, 16–18]. The initial purpose of using germ-free animals was to prove Pasteur’s assertion that life of an animal host would become impossible in the context of a germ-free experiment and his view that “microbial associates have become synergists which are indispensable in the life of the host” [1]. Meanwhile, Nencki [19] and Metchnikoff [20] had asserted just the opposite and that microbes are antagonistic to the well-being of their host [1]. Actual proof that vertebrates’ normal life is possible in the absence of microorganisms was found in Reyniers’s studies during the 1940s (presented in a summary at the Eighth Annual Meeting of the Animal Care Panel, November 7–9, 1957 in San Francisco, CA, USA) and Gustafsson’s group in the 1950s [21–23].

The technical ins and outs of germ-free technology during its history are famously presented in the paper “Life in a Germ-Free World”: Isolating Life from the Laboratory Animal to the Bubble Boy by Robert G. W. Kirk [24]. This article describes the original ideas and concepts of germ-free technology and the basic principles of germ-free isolator creation. The applications of germ-free technology in the laboratory for the production of laboratory animals, in agriculture for the efficient production of farm animals, and in hospitals for the control and prevention of cross-infection and the protection of individuals from infection are discussed in detail.

Subsequently, to utilize the advantages of gnotobionts for biological experimentation, several gnotobiological laboratories were founded in Europe, the USA, and Japan during the 20 century. A substantial contribution to understanding the nature of physiological differences between germ-free and microbiota-colonized animals was made by studies carried out at the gnotobiological laboratory of the Czechoslovak Academy of Sciences in Novy Hradek.

The Institute of Biology of the Czechoslovak Academy of Sciences in Novy Hradek, Czech Republic was founded in 1953. From its beginning, the laboratory campus had a special security regime and was protected by armed guards. The reasons lay in its studies on anthrax. For these studies, there were developed special isolators (in fact a version of Hansen’s sterile box) connected to a special air-conditioning system equipped with bacteriological filters. From 1958 to 1968, the Institute of Biology was used by the Czechoslovak army as a center for pathogenetic and immunogenetic studies on highly infectious microorganisms. After a lecture tour in the USA, Jaroslav Sterzl, a leading immunologist in the Czech Republic, rebuilt one of the infection laboratories into a gnotobiotic lab in 1960. The Laboratory of Gnotobiology was especially engaged in breeding of germ-free animals [25, 26] and began to utilize a unique experimental animal model allowing to distinguish the innate immune mechanisms of immune reaction that resulted from the interactions with microbiota as well as certain factors determining the course of antibody formation and development of immune response [27, 28]. Over the years, studies were carried out by Jaroslav Sterzl and fellow immunologists Helena Tlaskalova-Hogenova, Ivo Miler, Miloslav Pospisil, and Lydie Jaroskova; the chemists Jiri Rejnek and Jiří Zikan; and the microbiologist Vladimir Dlabac.

The dominant experimental models were rats, pigs, and mice. The model of colostrum-free and germ-free piglets came to be a suitable one for studying effects of the bacterial microbiota and examining innate as well as acquired immunity, where the development of antibodies under the influence of commensal bacteria was distinguished in comparison with a mother’s passively transferred antibodies [28].

Some original studies were devoted to developmental aspects of antibody formation and the processes occurring during differentiation of immunologically competent cells, as well as the role of antigens in the development of humoral immune responses [28–32]; to induction and regulation of secondary immune response to antigen (revaccination) [33, 34]; or to modeling of general aspects of antibody response [33, 35]. Powered by this knowledge, Jaroslav Sterzl and his colleagues demonstrated during the 1980s that the frequencies of antigen-binding B cells are comparable in the model of germ-free and in conventional mice. They further showed that antigen stimulation plays a role in activating antibody production but not in the establishment of B cell repertoire [36–43].

Thus, it occurred that over time gnotobiotic models began to be widely exploited in various branches of biomedical research.

Gut microbiota and microbiome

The coevolution of eukaryota with prokaryota across the epochs brokered for each body a unique set of microorganisms known as its microbiota. Gut microbiota, the most abundant and well-studied microbiota, has also been termed a “forgotten organ” [44] and a “neglected endocrine organ” [45]. Bacterial density in the gut reaches 1011–1012 cells/g in the distal human colon [46], and the sum of gut microbiota genomes, known as a microbiome, contains more than 100-fold more genes than are encoded in the human genome [47].

Gut microbiota has the capacity to influence the body’s physiology and metabolism. A very well-known example is the dependency of normal immune system development and function on the presence of gut microbiota (for example, see [48–50]). Modulatory effects on epithelial cell proliferation, villus architecture, and angiogenesis within the intestine, influence on xenobiotic metabolism, bone mineral density, behavior, and alteration of several metabolic functions have also been documented [51, 52]. Emerging data from humans and mouse models suggest that the gut microbiota plays a role in the development of metabolic diseases [6]. The microbiota seems to be important also for the programming and presentation of normal social behaviors, including social motivation and preference for social novelty. The original gut–brain axis concept has been changed to the microbiota–gut–brain axis, which is an emerging concept in modern medicine [53, 54].

The complex of genes which encompass the genomes of enteric and other bacteria comprises a particular organism’s unique microbiome. Knowledge of the microbiome and its changes, as they relate to the alteration of physiological, developmental, or metabolic processes in the organism offer the possibility to study the association of the microbiome with health and/or disease (i.e., the host–microbe relationship). For this reason, the Human Microbiome Project (HMP) was funded as an initiative of the NIH Roadmap for Biomedical Research (http://nihroadmap.nih.gov). HMP is focused on high-throughput technologies that can characterize the aforementioned changes in the human microbiome that can be associated with health or disease traits. A detailed analysis of molecular relationships between the microbiome and host can contribute to the development of new monitoring, diagnostic, and therapeutic procedures. Moreover, currently, accumulating data testify to the feasibility of manipulating the current health status of an organism through its microbiota [55–59]. Recently, comprehensive reviews have summarized the functions of microbiota in shaping the nutrient environment of the mammalian intestine by contributing to nutrient exchange, carbohydrate metabolism, and manipulation of the host’s metabolic machinery [60]. Other reviews have underscored the importance of the colonic environment for gut immune maturation [61], fully functional innate and systemic immunity [62], and anti-cancer responses [15, 63–65]. Nevertheless, quite limited data exist regarding the mutual interaction of germ-free models or their cells and pathogens. In this respect, the majority of experiments have been carried out using per os application of microbes that exert their pathogenicity predominantly through the gastrointestinal tract.

Gnotobionts as a model organism for study of host–pathogen interaction

Very soon after implementation of germ-free technology, it became evident that gnotobiotic animals constitute excellent tools for investigating the initial interactions of microorganisms with their hosts and that those studies can elucidate host immune responses to infections.

Initial studies tested mono-associated animals’ resistance to infection. Rapid colonization of the organs of germ-free animals infected with Listeria monocytogenes (L. monocytogenes) and higher sensitivity to this infection was obvious from these original trials. Nevertheless, in parallel, it was also reported that the conventionalization of L. monocytogenes mono-associated rats or di-association of mice with some representative of indigenous flora protects these animals against systemic L. monocytogenes infection [66, 67]. A similar conclusion concerning the importance of indigenous flora on the host’s immune status was made based on models of Salmonella typhimurium and Vibrio cholerae (V. cholerae) model infections [68, 69]. Thus, the extreme sensitivity of gnotobionts to pathogenic bacteria is a result of rapidly invasive bacterial infection and the slow development of immune response.

The onset timing, intensity, and composition of gnotobionts’ immune responses are critical for their resistance to pathogenic bacteria. Using the model of L. monocytogenes infection, it was demonstrated that lack of an intestinal microbiota impairs early innate immunity but enhances activation of memory T cells [70]. Although the T cells of gnotobionts are primed normally by listerial antigens, their trafficking to inflamed sites is severely impaired, and this may lead to increased susceptibility to infection with L. monocytogenes [71].

The absence of intestinal microbiota is also accompanied by a state of inflammatory hyporesponsiveness actively mediated by IL-10. IL-10 restrains proinflammatory mediator production and neutrophil recruitment and favors Klebsiella pneumoniae growth and dissemination in gnotobiotic mice. A parallel study demonstrated that transient toll-like receptor (TLR) activation restores the inflammatory response and ability to control pulmonary bacterial infection [72]. All these studies conclusively confirmed the importance of microbiota for the immune system’s effective response to pathogenic bacteria.

For the sake of completeness, it must be said that bacterial load of commensals alone is insufficient to protect against infection; rather, certain bacterial species correlate with protection [73]. Mice colonized with mouse microbiota, orally infected with Salmonella enterica (S. enterica) serovar Typhimurium, had a significantly lower Salmonella load in feces and less dissemination to the spleen in comparison to mice colonized with human microbiota. Moreover, ceca of mice colonized with mouse microbiota appeared healthy on histology, whereas ceca of mice colonized with human microbiota, similarly as original germ-free mice, had severe gross pathological changes characterized by thickening of the cecal wall, inflammation, and edema [61].

Gut microbiota have been demonstrated to be critical for mucosal protection from bacterial invasion and disease [74]. The precise role of microbiota in modulation of host immune response to pathogenic bacteria remains unknown, and however, comparative studies of host defence in gnotobionts and specific-pathogen-free (SPF) animals with normal microbial colonization are required.

The data demonstrated that the presence of commensal microbiota leads to competition for space and nutrients or regulates the production of intestinal mucins, which consequently inhibits the adherence of numerous pathogenic bacteria to intestinal epithelial cells [75–77]. The modulated immunological parameters have been demonstrated the induction of IgA as a means to reduce pathogen colonization and, by opsonizing the commensal bacteria, to interact with mucosal dendritic cells. This interaction is essential for the maintenance of intestinal homeostasis due to the tolerogenic profile of dendritic cells [78–80].

Commensal bacteria also influence the function of Th17 cells producing IL-17 and IL-22, which promote antibody class switching in B cells, induce production of anti-microbial peptides, and contribute to neutrophil recruitment [81, 82]. Another modulatory response can be the expression of pattern recognition receptors, including TLRs. One of these, TLR9, recognizes unmethylated CpG sequences in DNA. Its expression, along with the intracellular compartment of immune system cells, was demonstrated on the surface of intestinal epithelial cells. TLR9 was expressed on the colonic apical surface in wild-type mice but not in germ-free mice. Thus, intestinal epithelial cells responding to normal microflora colonization by surface expression of TLR9 may recognize pathogenic bacterial DNA and respond with an inflammatory response to pathogenic DNA [83]. Moreover, TLR9 engagement by commensal DNA promotes IL-17 and IFN-γ-producing T cells and reduces Treg frequency, which could facilitate immunity to pathogens [84].

Nevertheless, the data on the impact of changes induced by intestinal microbiota in the intestinal cell surfaces and, consequently, the impact on local immune responses to peripheral immune responses are ambiguous. On the one hand, commensals influence the mucosal and systemic immune responses to viruses and some bacteria [85–88]. On the other hand, there are data showing that germ-free mice intragastrically infected with S. enterica serovar Typhimurium have higher bacterial burden in the mesenteric lymph nodes compared to conventionally raised animals. Salmonella penetration into the lamina propria of the small intestine and splenic bacterial burden were not altered. Contrary to similar inflammatory phagocyte recruitment, intragastrically infected germ-free mice displayed a higher frequency of IFN-gamma-producing NK, NKT, CD4+, and CD8+ T cells in the mesenteric lymph nodes in comparison with conventionally raised mice. All these differences among germ-free and conventional mice were abrogated when the mice were infected intraperitoneally [89]. To explain the cellular and molecular consequences, infection caused by administrating bacteria to hosts via different routes will require further experiments.

Along with studies on immune response of gnotobionts to pathogens, germ-free animals have been used in analyzing bacterial pathogenicity. The pathogenicity of Escherichia coli [90], Mycoplasma pneumoniae [91], and S. enterica serovar Typhimurium [92] was demonstrated in germ-free animals. Moreover, gnotobionts seem to be useful for studies confirming individual bacterial components as virulence factors [93, 94], and they also have been used to monitor pathogen–host cell interactions [95]. The model of gnotobiotic animals does allow the evaluation of systemic and mucosal immune responses to antigens expressed by V. cholerae in vivo [69]. These data may be useful for designing a conjugate vaccine for cholera. Thus, gnotobionts may be used in multiply focused studies within bacteriology, immunology, or infection biology.

Even as the contribution of commensal gut bacteria and their microbiomes has been well established over the decades, the gut also contains an enteric virome that can modulate the body’s physiological parameters either alone or with a contribution from bacterial microbiota. The possible role of viruses present under homeostatic conditions in the gastrointestinal tract is not well understood [96–98]. Human virome tested on fecal samples from healthy children revealed a complex community of enteric viruses, including viruses of the families Picobirnaviridae, Adenoviridae, and Astroviridae and such species as bocaviruses, enteroviruses, rotaviruses, and sapoviruses [99]. Viruses that can create the gut virome include also pathogens, and these viruses are ubiquitously detected also in healthy individuals. Namely, these are members of the Anelloviridae that can cause chronic human viral infections or Caliciviridae, which cause viral gastroenteritis in humans. A substantial part of the gut virome is not characterized despite that there is limited sequence homology with known viruses [100–103].

As similarly true of bacteria, viruses can exert modulatory effects on an organism’s structural, metabolic, and defence functions. For example, infection of germ-free or antibiotics-treated mice with murine norovirus restores intestinal morphology, induces transcriptional changes leading to the development of immune status, and modulates lymphocyte functions without inducing inflammation or disease [104]. The data suggest that eukaryotic viruses (of course those that do not harm their host) have the capacity to modulate intestinal homeostasis and mucosal immunity in a manner analogous to that of commensal bacteria.

In respect of viral and other nonbacterial infections, gnotobiotic models offer a useful tool for the study of mutual interrelationships between bacterial microbiota, other infection agents, and the host. The host’s microbiota affects the replication and transmission of a diverse array of viral pathogens in both positive and negative senses. Changes in host microbiota have been suggested as a potential target for therapeutic intervention [105].

Nevertheless, the modulatory effect of microbiota on viral infections cannot be generalized. In contrast to orthomyxoviruses and arenaviruses, retrovirus-resistant mice control retroviral infection independently of commensal microbiota [106]. On the other hand, as was mentioned above, enteric viruses can replace the microbiota in supporting intestinal homeostasis and functioning of mucosal immunity, similarly as do commensal bacteria [104]. The gnotobiotic model also reveals a more complex interrelationship among bacterial, viral, and parasitic agents and their host [107].

Thus, the concept of germ-free animals has recently been exploited successfully in many branches of laboratory animal research [74–76]. Their almost naïve innate immune system actually provides a unique model for studying the cellular and molecular events that control host–pathogen interactions during the early stages of infection. These processes subsequently participate in the outcome of these interactions as seen in the emergence of states of health or disease.

Concluding remarks and open questions

Research utilizing gnotobiotic animal models has elucidated some questions concerning the influence of gut microbiota on the initiation and progression of such diseases as type 1 diabetes [108], metabolic syndromes [109, 110], obesity, autoimmune arthritis, inflammatory bowel disease, and irritable bowel syndrome [111–114], as well as the role of the microbiome in cancer [63, 64]. Other studies have investigated the effects of microbiota on the development and function of the immune system [6, 9, 10, 13, 115–117]. Studies on the basic problems of host–pathogen interaction utilizing gnotobiotic models are nevertheless only beginning. Although there have been studies concerning the interaction between mammals and bacteria, the majority of these have been focused on bacteria that colonize or infect the gastrointestinal tract [18]. Studies focused on pathogens that infect their host in a manner other than per os are practically nonexistent. According to our experience, germ-free mice infected subcutaneously with Francisella tularensis reacted differently to attenuated and virulent strains in comparison with specific-pathogen free mice. Thus, the ontogeny of these mice having no contact with bacteria can disclose the unique primary reaction of their immune system against pathogenic bacterial strains and can help in understanding the processes that lead to the establishment of full-fledged protective immunity. In the case of intracellular bacterial pathogens, the tools for inducing such effective immune reactions are still lacking.

Data collected from gnotobiotic as well as SPF animal models of microbial pathogenesis clearly demonstrate the decisive role of primary interaction between microorganism and the host cell that the microbe first encounters. In addition to the intrinsic characteristics of microorganism and host, a dominant role is played by the innate immune recognition process. The hierarchy of immune response functional modules based on the epigenetic reprogramming of innate immune cells after intercellular communication by cytokines and chemokines is underscored by the spatiotemporal aspect of host–microbe interaction.

From this point of view, the cell, with its functional and secretion profile, rather than the host organism in its entirety, seems to be the primary microbe host. The signals generated by this primary host after interaction with a microbe comprise a signaling window that controls all consecutive interactions. The emerging concept of innate immune response based upon a hierarchy of signaling windows inside the social network of immune cells (first presented during Discussion Forum 2016—Host Pathogen Interaction [118]) slightly modifies the damage-response framework (DRF) of microbial pathogenesis [119] by redefining the host. DRF had postulated that: (1) microbial pathogenesis requires a microbe and a host, (2) the microbe and the host must interact, and (3) the relevant outcome of host–microbe interaction is damage within the host whereby damage results from microbial or host factors or both. The main impact of this framework was integration of host and microbe roles in microbial pathogenesis. A new problem emerged, however, in association with defining just what constitutes a host. The revised DRF of microbial pathogenesis defines a host as an entity which houses an associated microbiome/microbiota and interacts with microbes such that the outcome results in damage, benefit, or indifference, thus resulting in the states of symbiosis, colonization, commensalism, latency, and disease [120]. Moreover, the emerging concept of signaling windows does not necessarily assume that the outcome of host–microbe interaction will be damage to the cellular host. The interaction might influence the primary signaling pathways modulating gene expression, secretion profile, and, subsequently, expression of function in a microbe-specific way.

Thus, in contradiction to the unifying tendency of DRF regarding the definition of a host, we postulate the individual cell as host. This highlights the relevance of an individual interacting cell type and its functional profile in relation to an interacting microbe through the process of host–microbe interaction. This concept is more dynamic, is more useful in understanding the role of microenvironment on the host–microbe interaction in the compartmentalized immune system, and can better depict the dynamic processes during innate immune response against infection.

Germ-free animal models in combination with mono-associated gnotobionts and SPF animals can serve as a useful tool in analyzing the host–pathogen interaction, and that is in accordance with the original, pioneering work of Jaroslav Sterzl. A gnotobiont based on the interaction between a germ-free animal and pathogen can help to verify the existing models of innate immune responses induction and clarify the role of individual cell types and their receptors in intracellular and intercellular signaling during their mutual interaction which leads to expression of protective immunity against pathogenic microbes. The basic data obtained in using such an approach may influence the view as to the construction of effective and safe vaccines, especially in such cases as immunoprophylaxis fails or does not yet exist.