Intravital imaging of host–parasite interactions in skin and adipose tissues

Abstract Intravital microscopy allows the visualisation of how pathogens interact with host cells and tissues in living animals in real time. This method has enabled key advances in our understanding of host–parasite interactions under physiological conditions. A combination of genetics, microscopy techniques, and image analysis have recently facilitated the understanding of biological phenomena in living animals at cellular and subcellular resolution. In this review, we summarise findings achieved by intravital microscopy of the skin and adipose tissues upon infection with various parasites, and we present a view into possible future applications of this method.


| INTRODUCTION
Over the course of almost a century, the implantation of imaging windows on animals for microscopic observation has been widespread, and within this time frame, it has evolved significantly. Now coined "intravital microscopy" (IVM), the technique of using optical windows to visualise phenomena at cellular or subcellular resolution has come a long way since its first use in 1824, when it was applied to visualise the rolling of leukocytes on the vascular endothelium of living frogs (Dutrochet, 1824;Wagner, 1839). These studies and an additional series of first discoveries were key for understanding endothelial physiology. Various animal models (including worms, fish, insects, amphibians, reptiles, and mammals) have been imaged since by IVM. In parallel, a series of technological breakthroughs over the past few decades in the field of physics, optics, and genetics have transformed IVM from an exotic tool to a commonly used platform to dissect processes of health and disease in living animals. Advances in image analysis have also transformed the use of IVM from mostly qualitative to yield quantitative results (reviewed by Coombes & Robey, 2010).
In parasitology, even though animal models exist for a considerable range of parasites, many questions remain understudied in vivo. In this review, we present findings that have been achieved by IVM focusing on vector-borne parasites. Furthermore, we provide insights into the optimised methodology, surgical procedures, imaging, and anaesthetic techniques that have made these findings possible. We aim to convey the relevance of IVM to the parasitology community, given the significant findings it has already allowed. Equally importantly, it has allowed us to reassess ideas and questions proposed by in vitro studies, which were not possible to observe in vivo, or not possible to quantify. With current technology, we have been able to revisit these questions, and in many cases find unexpected features relevant to pathology, or cell biology. The use of animals and live imaging in particular has been extremely valuable for the parasitology field in the absence of relevant in vitro or in silico models that can faithfully reproduce host-parasite interactions of clinical relevance to human health. Schistosoma spp. where the swimming cercariae can actively penetrate the skin, and Dracunculus spp. (Guinea worms), which is ingested orally, but subsequently invades subcutaneous tissues, from where it slowly egresses upon contact with water. However, the skin is much more than a route of entry into the host, as most parasites spend at least part of their existence there and often initiate a first host response. The skin can also serve as an anatomical reservoir of parasites and is a recurring theme in arthropod-borne human diseases, probably because skin invasion for enhanced transmission to the next host is likely a powerful evolutionary force.

| BIOLOGICAL RELEVANCE OF THE SKIN AND ADIPOSE TISSUE FOR PARASITOLOGY
The skin is the largest organ of the human body, and it represents the first line of immunological defence against many infections, with extensive crosstalk between epithelial, stromal, and immune cells to ensure homeostasis. Most parasites have developed mechanisms to evade detection and successfully establish an infection either in the skin itself or elsewhere in the host. Anatomically, the skin can be divided into three distinct compartments: the epidermis, which is an avascular layer mostly composed of keratinocytes and Langerhans cells; the dermis, which is highly perfused by blood and draining lymphatic vessels ( Figure 1); and the subcutaneous adipose tissue. The structure of the skin provides an interface between the vascular and lymphatic circulations, as well as the interstitial space. The latter is a fluid-filled anatomical compartment defined by a complex lattice of collagen bundles, found within and between tissues including the dermis (Benias et al., 2018). Until recently, the physiological importance and extent of the interstitium had been largely understudied, yet this compartment is very likely to be of relevance for host-pathogen interactions defining phenomena such as extravasation and sequestration of different parasites.
Direct in vivo and in situ imaging of the skin has facilitated the understanding and identification of host and pathogen factors relevant for effective parasite transmission upon vector biting (or parasite traversal itself, as is the case of, e.g., Schistosoma spp.). They have also allowed the study of general aspects of parasite development including proliferation, migration, and interactions with the host immune system.
A tissue that together with the skin has gradually gained momentum and interest for its relevance as a parasite reservoir is the adipose tissue (reviewed by Tanowitz, Scherer, Mota, & Figueiredo, 2017), although its exact function(s) have not yet been established.
In various fields of research, the adipose tissue went from being regarded as an inert site of energy storage or "fat," to being considered now as a complex, multicomponent site of paramount relevance to metabolism and systemic immunity (reviewed by Tilg & Moschen, 2006). This importance has permeated into parasitology, and it is possible that the adipose tissue modulates/impacts on parasite biology. It might serve as a source of nutrients, allow modulation of immunity and pathology, and/or support transmission (in the case of subcutaneous adipose tissue). Coincident with the change in conception of the relevance of the adipose tissues, observations of some parasites at this anatomical site were already reported decades ago but few functional, dynamic, or IVM studies were pursued, or reported until recently. The potential of IVM to this tissue is further discussed in this review.  Table 1. Skin at distinct anatomical sites exhibits important differences in, for example, thickness, composition, vascularization, and cytokine profiles. In recent years, several groups have investigated the relevance of cellular heterogeneity across skin sites with tools including IVM (Driskell et al., 2013;Jain & Weninger, 2013;Tong et al., 2015). This led the generation of a threedimensional immune cell atlas of mouse skin (Tong et al., 2015), dem- sites, impacting T cell responses, and immunisation outcomes (Wang et al., 2008). Importantly, these and other findings of skin cell heterogeneity have helped reconcile disparate results across studies in other Leunig, Menger, Nolte, & Messmer, 1993;Sandison, 1924). Modern optical windows are made of titanium (Menger, Laschke, & Vollmar, 2002). Alternatively, they can be made of non-metallic polymer materials that allow their use with radiation and magnetic resonance imaging (Erten et al., 2010;Gaustad, Brurberg, Simonsen, Mollatt, & Rofstad, 2008) and are lighter in weight thus causing less discomfort to the mice. The method for generating an optical window for skin IVM consists of placing two symmetrical frames to "sandwich" an extended double layer of skin. The skin is covered with a removable glass coverslip, which can be incorporated into one of the frames, allowing visualisation with an objective. This set-up allows for repetitive analysis over a period of 2-4 weeks (Lehr et al., 1993), and it requires a 2-3-day recovery period following implantation. Finally, skin flank IVM is among the least explored methods in the context FIGURE 2 Optical windows and imaging chambers for skin and adipose tissue visualisation by IVM. Non-invasive methods include (a) ear pinna imaging and (b) foot-pad imaging. A semi-invasive method (c) includes the dorsal skinfold chamber, which requires the surgical implantation of two titanium or polymer frames that can hold a ring with a glass coverslip through which imaging is performed. An invasive method is the generation of a skin flap (d), whereby a skin flap is generated, exposing a large imaging area. This procedure is invasive and terminal. A less commonly used method for IVM imaging is the skin flank (e) which requires the generation of an incision at a dorsolateral location, and either direct imaging or mounting on a stainless steel disc for stable image acquisition. For adipose tissue imaging, various types of window exist to visualise various depots (marked by X). To image the perigonadal adipose tissue, a terminal lower abdominal window (    Excision of the bite site showed that mosquito-injected Plasmodium sporozoites remained in the skin for at least 5 to 15 min, before entering the blood stream. It was hypothesised at the time that sporozoite migration to the blood vessels following inoculation was delayed either because of anti-sporozoite antibodies or a cutaneous hypersensitivity reaction to the mosquito bite (Sidjanski & Vanderberg, 1997).
Interestingly, this consistently resulted in fewer inoculated sporozoites in the ear, but the kinetics of migration were conserved between sites.
IVM also revealed that up to 15-20% of inoculated sporozoites are transported via lymphatic vessels to the lymph nodes (Amino et al., 2006;Yamauchi et al., 2007), and that a small fraction of parasites that remain in the skin can begin developing within this organ (Coppi et al., 2011;Gueirard et al., 2010). The remaining majority of sporozoites are either eliminated or successfully travel to the liver endothelium via blood vessels.
Finally, IVM has allowed studying and quantifying the speeds of sporozoite migration at different bite sites of the dermis, the dynamics of sporozoite contact with host blood vessels, and the subsequent migration of sporozoites from the skin into the bloodstream (Amino et al., 2006;Hellmann et al., 2011;Hopp et al., 2015; Figure 1a,b).
These studies revealed that the sporozoite migration path is determined by the environment (Hellmann et al., 2011); that sporozoites migrate through cells using a secreted perforin-like protein (Amino et al., 2008), and slow down when migrating in close proximity to blood capillaries (Hopp et al., 2015).
These results also informed in vitro studies at the intersection of physics, and biology (Hellmann et al., 2011;Muthinja et al., 2017), that use patterned environments to address questions that could not be answered by IVM. For example, the different migration path in the skin of ear or tail could be assigned to the different architecture of the environment (reviewed in Muthinja et al., 2018). Small differences in sporozoite migration in the skin were observed by IVM in mice lacking a specific integrin, possibly suggesting that this molecule interacts with a sporozoite surface protein (Dundas et al., 2018). Finally, a recent elegant study revealed that the sporozoite surface protein CSP protects the parasites from the action of its own perforins, which are necessary for migration through cells in the skin (Aliprandini et al., 2018).

One of the first in vivo clues of the relevance of adipose tissue to
Plasmodium asexual blood stages came from a study using bioluminescence to characterise Plasmodium berghei sequestration in rodents

| Trypanosoma brucei
Human African trypanosomiasis (HAT), or sleeping sickness, is caused by two main subspecies of T. brucei that are exclusively transmitted through the bite of Glossina tsetse flies. Within the fly, T. brucei undergoes a complex developmental cycle culminating in the production of salivary metacyclic forms that can infect mammalian hosts (Rotureau & Van Den Abbeele, 2013). Like Plasmodium and other vector-borne pathogens, the presence of T. brucei changes the composition of the saliva and modifies the feeding behaviour of the tsetse fly in a way that enhances the chances of parasite transmission to the host (van den Abbeele, Caljon, de Ridder, de Baetselier, & Coosemans, 2010). Tsetse flies are pool feeders and lacerate the skin of their host rather than inserting a proboscis directly into the vasculature. After causing significant local damage and inflammation at the bite site that often result in a transient chancre, the insects feed on the resultant pool of capillary blood and lymph (Bouchet & Lavaud, 1999;Goodwin, 1970). However, our knowledge on the early interface between the parasites and the host skin in vivo remained limited until recently.
Various studies have documented skin reactions to the parasite in different animal models by histology, fluorescence-based imaging, electron microscopy (Akol & Murray, 1982;Dwinger, Rudin, Moloo, & Murray, 1988;Goodwin, 1971Goodwin, , 1970Ikede & Losos, 1972;Mwangi, Hopkins, & Luckins, 1995;Mwangi, Hopkins, & Luckins, 1990;Sbarbati et al., 2010;Thuita et al., 2008), thermographic imaging, or molecular and flow cytometric methods (Caljon et al., 2016).  (Capewell et al., 2016). This imaging approach consisted of combining whole animal bioluminescence imaging, spinning-disc confocal microscopy on the ear pinna, and two-photon microscopy on the abdominal flank, in the same animals. In parallel, a functionally adapted population of parasites was shown to occupy adipose tissues in a mouse infection (Trindade et al., 2016), and it is possible that the adipocyte-rich hypodermal layer of the skin may attract, host, and/or maintain at least part of the skin-dwelling trypanosome population by providing them an immunological niche and/or a stable lipid-rich nutritive environment (Tanowitz et al., 2017).
Although the existence of a significant extravascular population of T. brucei parasite is not novel (Goodwin, 1970) (Büscher et al., 2018). The presence of parasites in extravascular sites (skin, adipose tissue, and brain) for extended periods raises several fundamental questions on parasite biology.
Indeed, parasites in the blood and adipose tissue appear functionally specialised as they can catabolise fatty acids only in the fat (Trindade et al., 2016). This in turn raises several questions regarding how parasites proliferate, how differentiation is triggered preferentially in transmission-compatible tissues, how parasites migrate, and whether and how they sequester.

| Trypanosoma cruzi
Triatomines are the haematophagous insect vectors of T. cruzi-the causative agent of Chagas disease. Triatomines are vessel feeders, and after piercing, the host skin will probe via rapid whip-like intradermal movements of the maxillae until a vessel is found. Like mosquitoes and tsetse flies, triatomines possess various molecules in the saliva, which reduce haemostasis and have anti-inflammatory properties (Ribeiro, 1995;Pereira et al., 1996;Ribeiro, Schneider, & Guimaraes, 1995 (1); Ribeiro & Francischetti, 2003). both focused on the patterns of triatomine feeding and salivation at the mouse skin, rather than vector-to-host parasite transmission (Soares et al., 2006;Soares, Araújo, Carvalho-Tavares, Gontijo, & Pereira, 2014). The first study used the ear pinna of mice, together with triatomine saliva labelled with acridine orange. Using epifluorescence microscopy, it was found that salivation occurs throughout all feeding phases (probing and engorgement) of the triatomine Rhodnius prolixus and measured the frequency of saliva emissions (Soares et al., 2006).
The second study investigated the interface between the feeding process of the triatome and the host's response at the vascular endothelium ( Figure 1h). This work used epifluorescence microscopy, as well as stereomicroscopy and an electromyogram. The main findings included (a) the induction of vascular permeability alterations following a triatomine bite via injection of dyes, (b) the immediate platelet and leukocyte aggregation at the venular endothelium following a bite, and (c) the integration of imaging and analysis methods in vector and host, which enabled monitoring vessel wall pulsations to register movements during blood pumping, as well as the evaluation of blood flow through the triatomine's head (Soares et al., 2014). Although there are no IVM studies yet that investigate parasite transmission at the skin, one ex vivo study addressed the kinetics of skin penetration of T. cruzi trypomastigotes in mice (Schuster & Schaub, 2000). A drop of vector faeces or urine containing trypomastigotes was placed on the bite site of a triatomine, and the skin was surgically removed after different periods of time. This showed that the minimal exposure period necessary for infection was as little as 5 min with longer periods of exposure time correlating with higher infection rates. This showed that T. cruzi can rapidly invade the host, and that some parasites can be carried away from the bite site immediately (Schuster & Schaub, 2000) and sets the stage for similar studies as those conducted in T. brucei and Plasmodium to reveal the dynamics of infection.

| Schistosoma
Various Schistosoma species are the causative agents of schistosomiasis. The sexual reproduction of the parasite occurs in humans as well as other hosts, whereas asexual reproduction occurs in freshwater snails (Biomphalaria spp.). Schistosoma are motile in all life stages, and this motility is relevant to their capacity to search for, and invade the host skin, and to circulate within the host following invasion.
Schistosomes are phototrophic, which leads the cercariae to preferentially localise to the surface of shallow waters, where they can maximise their contact with humans (Stirewalt & Dorsey, 1974).
Cercariae then respond to thermal gradients in order to find the skin of the host and then to chemical cues that allow them to complete invasion (Fishelson et al., 1992;Gordon & Griffiths, 1951;Haas, Diekhoff, Koch, Schmalfuss, & Loy, 1997;Lewert & Lee, 1954;Salter, Lim, Hansell, Hsieh, & McKerrow, 2000 showed that as cercariae move through the skin, CFDA-SE-labelled material is released via the oral sucker. This material is thought to be a mixture of digestive proteases that aid in the migration of the parasite. Ex vivo work investigating the dynamics of ES molecule uptake by macrophages and dendritic cells showed that, depending on the relative abundance of each cell type, and differential rates of antigen processing by these cells, this might be key to the success of adaptive immune priming in response to Schistosoma infection (Paveley et al., 2009).

| FUTURE DIRECTIONS
Considering its ease, IVM of parasites in the skin has focused mainly on the use of the ear pinnae of mice. Although this has advantages in terms of accessibility and non-invasiveness, the few studies in Plasmodium and T. brucei using other skin sites to image parasite dynamics have shown differences between locations. This is particularly important as the skin is physiologically different between anatomical sites and parasites may adapt to the respective environment. The use of skinfold chambers, which allow visualisation within a much larger surface area, for a much longer period of time will be a useful tool to study new tissue locations. Interestingly, most IVM studies have focused on vector-to-host transmission, with only one having imaged host-to-vector transmission after a bite. The importance of the skin and adipose tissue as reservoirs and for disease pathology remains largely obscure and requires more attention in the future.
Technological advances including the increasing range of available fluorescent reporter mice, and transgenic parasites expressing different fluorescent proteins alongside tagged genes potentially involved in transmission will lead to further advances in our understanding of parasite biology. These technologies, in combination with the use of single-cell transcriptome analysis and immune profiling, may reveal new tissue-specific interfaces between parasites and their hosts that have not been observed previously.
An exciting area of IVM that is becoming more popular for parasite research is the use of humanised models. Although conventional animal models are invaluable in providing insights into parasite behaviour, they cannot fully replicate human disease due to differences in parasite strains, cell morphology, and humoral immune responses. As a result, there has been a drive in recent years to examine human parasites directly. One way to achieve this is through the use of humanised models. By incorporating human tissues and stem cells into immunocompromised animals, these models successfully mimic human physiology allowing human parasites to be studied in an in vivo environment. They have been particularly popular for studying P. falciparum infection as they allow certain aspects of malaria to be studied in a way not possible using murine Plasmodium spp. (reviewed by Minkah, Schafer, & Kappe, 2018).
One of these aspects is the sequestration of the P. falciparum asexual stages in the microvasculature. Although this also occurs in certain murine malaria infections, the binding ligand of P. falciparum is unique, limiting our abilities to study this behaviour closely. This has been addressed in a number of studies by utilising mice implanted with human skin grafts. Through IVM, they have observed the rolling and adherence behaviour of parasitized red blood cells in both postcapillary venules and arterioles and demonstrated that this behaviour could be reversed using antibodies that blocked the parasites binding ligands (Ho et al., 2000;Yipp et al., 2007). Similar models using human subcutaneous fat, a more prominent site of sequestration that closely mimics brain endothelium (Moxon et al., 2013) have yet to be fully realised but are in development (Meehan et al., unpublished).
Aside from these limited studies using P. falciparum, little IVM research has been carried out on parasites using humanised animals, but as this technology develops further, it will likely be utilised by an increasing number of researchers in the future.