A number of animal models of F. tularensis infection have been developed, including mice, rats, rabbits, guinea pigs and non-human primates (e.g., cynomolgus and rhesus monkeys). The advantages and disadvantages of these various animal models (and how they compare with the marmoset model) are presented in Table 4 . To the best of our knowledge, we are the only group to report on a marmoset model of F. tularensis infection to-date (Nelson et al., 2009; Nelson et al., 2010a; Nelson et al., 2010b). In this section, the marmoset model of inhalational tularemia will be discussed with a particular emphasis on the immunological features. The reader is directed to the above publications for full details of the model.

The marmoset as an NHP model of tularemia has a number of advantages (see Table 4 ); importantly, the course and progression of disease accurately recapitulated human disease – including the development of ulcers, a feature not observed in any other animal model (Nelson et al., 2010b; Roberts et al., 2018). Evidence of an immune response was demonstrated by the production of pro-inflammatory cytokines with disease progression. For example, at 72 hrs post-challenge, monocyte chemoattractant protein (MCP)-1 (CCL2) was detectable in the spleen, lungs and blood and the level increased until death (Nelson et al., 2010b). Additional cytokines, including macrophage inflammatory protein (MIP-1α; CCL3), MIP-1β (CCL4), interleukin (IL-6), IL-1β and regulated on activation, normal T-cell expressed and secreted (RANTES; CCL5), were upregulated in all organs at 96 hrs post-challenge (Nelson et al., 2010b). Interestingly, MIP-1α and IL-6 were first observed shortly prior to death, akin to the murine model of inhalational tularemia (Conlan et al., 2008; Nelson et al., 2010b). Neutrophils and natural killer (NK) cells were the first cells to arrive at the site of infection (24 hrs post-challenge), followed by macrophages, T-cells and additional influx of NK cells (48 hrs post-challenge) (Nelson et al., 2010b). A decline in the percentage of neutrophils in the lung and blood at 72 hrs post-challenge was observed, raising important questions concerning the role of neutrophils in response to F. tularensis infection. Indeed, studies assessing the importance of neutrophils in the response to F. tularensis infection are conflicting. Using the neutrophil-depleting antibody RB6-8C5, Sjöstedt and colleagues found that mice depleted of neutrophils were vulnerable to otherwise sublethal doses of F. tularensis, delivered either intraveneously or intradermally, suggesting a key role for neutrophils in controlling bacterial replication (Sjöstedt et al., 1994). Meanwhile, KuoLee and colleagues demonstrated that depleting the number of neutrophils had no effect on the bacterial burden or time to death (Kuolee et al., 2011). It has been suggested that the role of neutrophils in response to infection with F. tularensis may be dependent on the site of infection and that, in some cases, excessive neutrophil recruitment may contribute to the over-production of pro-inflammatory cytokines that ultimately lead to sepsis (Malik et al., 2007; Metzger et al., 2013; Steiner et al., 2014). Notably, whilst infection with the type A strain rapidly induced neutrophil recruitment in the marmoset, the type B (but not type A) strain led to neutrophil influx in the mouse, highlighting an important difference between the two species (Hall et al., 2008; Nelson et al., 2010b). By 72 hrs post-challenge, the number of B-cells and T-cells in the spleen and blood increased (Nelson et al., 2010b). By 96 hrs post-challenge, the number of neutrophils in the blood and organs returned to normal levels; a concomitant decline in the number of NK cells, both B- and (CD4+) T-lymphocytes and macrophages in the lungs was also observed (Nelson et al., 2010b). The proportion of CD8+ T-cells and γδ T-cells in the spleen and lung were increased 96 hrs post-challenge (Sumida et al., 1992; Poquet et al., 1998; Kroca et al., 2000; Nelson et al., 2010b). γδ T-cells are thought to play a role in the innate immune response and thought to be important in human infections with F. tularensis (Rowland et al., 2012a; Rowland et al., 2012b). An increase of γδ T-cells in the blood was not observed, consistent with reports in humans, where cells were discerned approximately one week post-infection (Kroca et al., 2000).

Having shown the marmoset model of tularemia effectively recapitulates human disease, a follow-up study by our research group evaluated the efficacy of levofloxacin, a fluoroquinolone shown to be effective against F. tularensis (Hepburn and Simpson, 2008). Fluoroquinolones have a number of advantages over current treatment protocols, including their broad-spectrum activity (important when diagnosis is difficult), bactericidal effects, tolerability and oral administration (Fish, 2003; Hepburn and Simpson, 2008). Further, levofloxacin is effective as a single daily dose which will likely increase compliance (Nelson et al., 2010a). Indeed, levofloxacin is approved for the treatment of inhalational anthrax in both children and adults (Deziel et al., 2005; Li et al., 2010). To achieve licensure of any therapeutic agent for a given disease under the animal rule (discussed earlier), the efficacy and safety profile must first be assessed in a NHP. In our study, all animals that received levofloxacin for ten days post-exposure survived and showed no clinical signs of disease, indicating the efficacy of oral levofloxacin against inhalational tularemia (Nelson et al., 2010a).

In summary, the common marmoset model of tularemia effectively and accurately recapitulates human disease and has numerous advantages over alternative animal models. It will be useful for the evaluation and licensure of medical countermeasures by the FDA.

Despite early studies of experimental melioidosis in rhesus macaques, much of our understanding of the pathogenesis and the effectiveness of therapies against melioidosis and glanders has emerged from small animal models, specifically mice and hamsters (Warawa, 2010; Amemiya et al., 2017). Reports from the 1990s described experimental infections of baboons with B. pseudomallei and B. mallei (Manzeniuk et al., 1999). More recently, our group established and characterized a common marmoset model of B. pseudomallei infection following inhalational challenge (Nelson et al., 2011a). An African green monkey and rhesus macaque model of experimental infection has also been described (Miller et al., 1948; Yeager et al., 2012). The advantages and disadvantages of these various animal models, and how they compare with the marmoset model, are presented in Table 5 . In this section, the marmoset model of melioidosis is discussed with particular emphasis on the immunological features. The reader is directed to the above publications for full details of the model.

Work from our research group has led to the development of a marmoset model of experimental melioidosis caused by three natural routes of exposure to B. pseudomallei, i.e., through broken skin, inhalation and ingestion (Nelson et al., 2011a; Nelson et al., 2014; Nelson et al., 2015; Nelson et al., 2021; Nelson et al., 2022a; Ngugi et al., 2022). Clinically, this is important as the route of exposure, whilst often difficult to determine at disease presentation, is likely to impact on the efficacy of medical countermeasures. Whilst early studies of experimental melioidosis in the marmoset reported limited immunological findings, a recent study by Ngugi and colleagues provided the most complete and comprehensive analysis of the immunological features of acute pneumonic disease resulting from B. pseudomallei exposure to-date (Ngugi et al., 2022). Significantly, features of the marmoset immune response to infection (e.g., neutrophil and macrophage migration and activation, T-cell activation and the production of pro-inflammatory mediators) mimicked acute disease in humans and was associated with disease prognosis, providing additional evidence as to the validity of the model. The proceeding section will focus predominantly on neutrophils, though other immunological components will be noted.

Notably, naïve marmoset neutrophils exhibited a rather different phenotype compared to the human counterpart. Specifically, HLA-DR (MHC II) was constitutively expressed on naïve marmoset neutrophils whereas in humans HLA-DR expression is not typically observed on resting neutrophils (Meinderts et al., 2019; Ngugi et al., 2022). Additionally, expression of the classical marker used to identify human neutrophils, CD16 (the Fc receptor gamma III), was lower on marmoset neutrophils (Silvestre-Roig et al., 2019; Ngugi et al., 2022). Considering that the proportion of circulating cells (and particularly neutrophils) in the marmoset more closely resembles that in humans, the significance of these phenotypic variations is unclear and the marmoset remains a viable model of human disease. Most importantly, both the proportions and cellular phenotypes changed during the course of the disease providing an objective, quantitative metric of disease progress and thus the opportunity to assess the efficacy of therapeutic interventions. In this study, the proportion of circulating neutrophils increased during the first 48 hrs post-challenge, after which the number declined significantly (and below baseline levels) in terminal animals. Meanwhile, the proportion of neutrophils in the lung declined 12 hrs post-challenge which is contrary to the scenario in the mouse, whereby neutrophil influx into the lung is observed post-challenge (Laws et al., 2011). At 36 hrs post-challenge, neutrophil proportion began to recover, returning to near-baseline levels by 48 hrs post-challenge. The authors noted, however, that since cell typing was proportional, it was not clear whether the apparent decline in the number of neutrophils in the lung was the result of neutrophil death [as a result of bactericidal processes (Kaplan and Radic, 2012)] or merely indicative of enhanced lymphocyte infiltration. Concomitantly, the proportion of circulating T (but not B) lymphocytes declined as the disease progressed. As noted, lymphocyte proportions were increased in the lung at 12 hrs post-challenge and continued to increase until 36 hrs post-challenge, after which levels declined. Changes to the proportions of cells in the spleen were similar to those observed in blood. In addition to changes to the proportion of cells in the various tissues, phenotypic changes were observed in neutrophils immediately following challenge. Significantly, expression of HLA-DR (which is constitutively expressed on marmoset neutrophils) dropped as disease progressed in the blood, lung and spleen. In blood, significantly reduced expression of HLA-DR was observed at all-time points post-challenge; in the lung and spleen, a significant decline in the proportion of neutrophil HLA-DR expression was observed by 12 hrs post-challenge and before the onset of clinical signs of disease, e.g., fever. Taken together, these findings provide additional evidence to support the use of the marmoset model of melioidosis for assessing medical countermeasures. Encouragingly, these findings regarding HLA-DR, CD54 and CD16 were also observed in a more recent, related study with B. pseudomallei (Nelson et al., 2022a).

Considering the role of neutrophils as first-responders to injury and insult, and their documented significance in early melioidosis (Easton et al., 2007; Laws et al., 2011),the fact that neutrophils showed the most significant variation of all cellular parameters assessed is not surprising. In the mouse, neutrophils play a central role in the acute response to aerosol infection (Easton et al., 2007). Though susceptibility to infection is largely pre-determined depending on the specific mouse strain (Warawa, 2010), marmosets are considered to demonstrate enhanced sensitivity to (particularly) aerosol challenge and this may be due to the tendency for a decline in the proportion of neutrophils in the lung during the early stages of infection (Nelson et al., 2022a; Ngugi et al., 2022). Alternative explanations should not be disregarded. These include the possibility that early neutrophil influx into the lung does occur, yet neutrophils are not detectable by flow cytometry because they are infected and degraded. In this scenario, subsequent neutrophil recruitment and activation occurs too late to counteract an already rapidly escalating bacterial burden. Encouragingly, the pattern of neutrophil recruitment in the marmoset mirrors that observed in other NHP models in the rhesus macaque and African green monkey (Yeager et al., 2012). Additional evidence implicating neutrophils as key players in early melioidosis include the association between excessively high or low neutrophil counts and poorer outcomes in humans, and the increased susceptibility of individuals with certain conditions (e.g., diabetes) associated with suboptimal neutrophil function (Chanchamroen et al., 2009; Saengmuang et al., 2014; Jenjaroen et al., 2015).

With a marmoset-specific candidate biomarker indicative of infection (a reduction in neutrophil HLA-DR expression), our research group recently evaluated the efficacy of co-trimoxazole using the marmoset model of experimental melioidosis (Nelson et al., 2022a). In this study, animals were challenged by one of three exposure routes: inhalational, ingestion or subcutaneous. Once fever had developed, a proportion of the animals were administered oral co-trimoxazole; all remaining animals received a placebo. A second-dose was administered 12 hrs after the first, followed by one dose every 12 hrs up until a total of 28 doses was delivered. With respect to the immunological perturbations, the proportion of neutrophils increased at the onset of fever, yet there was a drop in the level of HLA-DR expression that continued until animals succumbed to disease. HLA-DR expression was at a normal level by day 15 post-challenge in those animals that received oral co-trimoxazole. In addition to validating the observation of decreased HLA-DR expression with the onset of fever in an independent study, the immunophenotyping panel was also expanded and incorporated markers for CD16 (Fc gamma receptor III, expressed on NK cells, macrophages and neutrophils, plays a role in the internalization of exogenous antigens by binding the Fc portion of IgG immune complexes),CD66b (an activation marker on granulocytes), CD80 (a co-stimulation marker used by professional phagocytes to aid in MHC to T-cell receptor interactions) and CD54 (intracellular adhesion molecule-1 (ICAM-1), an adhesion molecule involved in lymphocyte homing and activation). Expression of all these markers decreased in the placebo group; meanwhile, neutrophil CD16 expression returned to normal levels in the co-trimoxazole treatment group. Upon treatment cessation, animals either survived, relapsed and succumbed to disease or exhibited abnormal immunological perturbations indicative of subclinical disease. Importantly, those animals that survived without relapse maintained normal levels of HLA-DR expression on neutrophils. A decline in neutrophil HLA-DR expression was observed in those animals that would later relapse and succumb to disease; likewise, elevated circulating IFN-γ was detectable and indicative of relapse up to three days prior to death. At post-mortem, a reduced proportion of neutrophils in the blood was the only indicator of fatal disease. Minor immunological changes were observed between those animals that succumbed, recovered and later relapsed and those that survived. For example, there was a somewhat increased proportion of CD69+ CD8+ T-cells and decreased expression of CD40, CD16 and CD64 on macrophages. Interestingly, whereas neutrophil influx into the lung was a feature of those animals that received the placebo, there was no evidence for this in animals that received treatment and later relapsed. Akin to the situation in humans (Jenjaroen et al., 2015; Nithichanon et al., 2018), there was evidence of T-cell activation (indicated by expansion of the cytotoxic T-cell proportion and expression of CD16 and CD69) in animals that survived until the study end. The population of γδ T-cells was also expanded in survivors, providing additional evidence to support an important role for this cell type in the response to infection (Haque et al., 2006; Andreu-Ballester et al., 2013; Laws et al., 2013; Kronsteiner et al., 2019). Notably, a re-stimulation assay of splenic T-cells taken from those animals that survived revealed enhanced IFN-γ production compared with the negative control (Nelson et al., 2022a). In those animals that survived to the study end, high antibody titers were observed. Yet the relative protective value of the humoral response in humans is limited, despite the importance of vaccine-induced humoral immunity having been demonstrated in animal studies (Burtnick et al., 2018; Khakhum et al., 2019; Chaichana et al., 2020; Chaichana et al., 2021).

In summary, the common marmoset model of melioidosis has been well characterized and shown to recapitulate human disease and exhibit a higher degree of similarity to human disease compared with other animal models. It will no doubt have value in the evaluation and licensure of medical countermeasures.

The search for a more robust animal model of human HCV infection, particularly one permitting testing of vaccine efficacy, is important and remains a pressing unmet need in hepatitis C research. Whilst highly effective treatments for HCV infection exist, these are often prohibitively expensive and, consequently, are unavailable to those most at-risk individuals (Etzion and Ghany, 2015; Chahal et al., 2016). The development of preventative measures (like vaccines) is therefore key.

Development of a surrogate common marmoset model (Parks et al., 1969; Lanford et al., 2003; Bright et al., 2004; Kyuregyan et al., 2005; Haqshenas et al., 2007) of human HCV infection (with the NHP-specific GBV-B and, later, HCV chimera) followed earlier studies performed in tamarins (Deinhardt et al., 1967; Beames et al., 2000; Beames et al., 2001) which, compared to marmosets, are difficult and costly to breed in captivity. Though tamarins are susceptible to GBV-B infection, the utility of the tamarin model (and indeed monkey models more generally) of HCV infection was highly debated given the inability to establish chronic infection, a hallmark of human HCV infection (Lanford et al., 2003; Weatherford et al., 2009). The usefulness of the tamarin model was also limited by the availability of animals (Weatherford et al., 2009). Early studies in the marmoset revealed the susceptibility of the species to GBV-B infection, with animals developing acute viraemia (albeit to a lower level compared with that seen in tamarins) (Parks et al., 1969; Lanford et al., 2003; Bright et al., 2004). Interestingly, the level of viraemia in the marmoset was similar to that seen in chimpanzees (10⁷ copies/mL or less) which have been shown to develop persistent infections (Fernandez et al., 2004; Bukh et al., 2008). Thus, it has been suggested that lower viral loads in the acute phase of the infection may actually support viral persistence and the development of chronic inflammation (Iwasaki et al., 2011). Indeed, Iwasaki and colleagues were the first to show that infection of the marmoset with GBV-B produced a chronic and progressive disease similar to human hepatitis C, as indicated by fibrosis and recurrent increases of the liver enzyme alanine transaminase (ALT) (Iwasaki et al., 2011). Further, one marmoset experienced piecemeal necrosis and elevated ALT levels four years post-infection, indicative of an acute exacerbation associated with chronic hepatitis (Iwasaki et al., 2011), itself a feature of human viral hepatitis (Perrillo, 1997). Notably, marmosets infected with GBV-B were shown to exhibit two distinct phenotypes: susceptible and partially resistant (Weatherford et al., 2009). In contrast, HCV chimera (carrying core, E1, E2 and p7 structural proteins of HCV) causes persistent infection in marmosets (Li et al., 2014b). Since long-term viral persistence was established in animals with lower viral loads during the acute phase on infection (i.e., within the first 2 weeks post-infection), it seems reasonable to conclude that animals with the partially-resistant phenotype (where viral growth is restricted) will support the development of chronic infection. Viral persistence in those animals with lower viral loads may be the result of diminished early antiviral immune responses (Iwasaki et al., 2011). Data concerning the innate and adaptive immune response to infection in animals exhibiting acute disease compared with those that progress to develop chronic disease are still lacking and will prove critical in deciphering the mechanisms responsible for the establishment of chronic infection.

The induction of type I interferons represents one of the first responses to infection with HCV. HCV utilizes a NS3/4A protease to inactivate these early antiviral responses, possibly leading to viral persistence (Kaukinen et al., 2006). An interferon-inactivating NS3/4A protease is also present in GBV-B (Li et al., 2014b). In humans and chimpanzees, both CD4+ and CD8+ T cells play an important role in the response to HCV infection (Cooper et al., 1999; Lechner et al., 2000; Day et al., 2002; Woollard et al., 2003). The generation of virus-specific T cells that recognize multiple viral epitopes is crucial for viral clearance. Indeed, the accumulation of HCV-specific CD4+ and CD8+ T cells (recognizing multiple viral epitopes) in the liver is associated with acute resolving infection (He et al., 1999; Grabowska et al., 2001; Woollard et al., 2008). Conversely, a weaker T cell response against a limited number of viral epitopes is associated with viral persistence and chronic disease (Woollard et al., 2008). In the marmoset, IFN-γ production was first detectable five weeks post-infection, coinciding with a 1000-fold reduction in viral load (Woollard et al., 2008). A T cell response against NS3/N54A epitope (but no other viral epitope) was observed predominantly in the liver at week seven post-infection, coinciding with the clearance of viraemia (Woollard et al., 2008). At this point, virus-specific T cells appear in peripheral blood (Woollard et al., 2008). Akin to the situation in humans and chimpanzees, virus-specific T cells are present in higher frequencies in the liver than in the blood, suggesting the accumulation of T cells in the liver at the site of viral replication (He et al., 1999; Grabowska et al., 2001; Woollard et al., 2008). It is currently unclear whether the anti-HCV adaptive immune response is mediated by CD4+ or CD8+ T cells. Recently, the role of regulatory T cells (Tregs) in the response to HCV infection has gained increasing attention. Tregs, a unique type of CD4+ T cell with suppressor functions, are important in maintaining immune tolerance (Sakaguchi et al., 2008). In the context of an infection, Tregs can modulate effector T cell responses and, by inhibiting the anti-viral functions of specific T cells, may permit viral persistence (Boer et al., 2015; Liu et al., 2023). In chronically infected individuals, Treg populations are maintained, whereas the suppressor function of Tregs was diminished in individuals with acute resolving infection (Liu et al., 2023). The phenotype and role of Tregs in the marmoset is yet to be determined.

Another important aspect of the immune response against HCV is memory. In chimpanzees, virus-specific memory cells are essential for protection against reinfection (Grakoui et al., 2003; Shoukry et al., 2003). Marmosets were also protected from reinfection for several months after clearance of primary infection, pointing to the existence of virus-specific memory cells (Woollard et al., 2008). Consistently, T cell responses were both greater in magnitude and occurred faster following secondary infection, indicating recall of memory T cells (Bright et al., 2004; Woollard et al., 2008). In comparison to cell-mediated mechanisms of immunity, the humoral response to HCV infection is less well defined and requires further investigation.

In summary, the marmoset is susceptible to infection with both GBV-B and HCV chimeras and develops a hepatitis C-like disease, the pathology of which mirrors that of human HCV infection. Varying susceptibility phenotypes are likely genetically-determined, with some animals more likely to exhibit viral persistence and therefore chronic infection. In this sense, the marmoset may represent a valuable surrogate model of human hepatitis C.