Keywords

These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.

1 Introduction

Borrelia burgdorferi is the causative agent of Lyme disease, a multisystemic disorder affecting primarily skin, nervous system, and joints. Borrelia belongs to the phylum of spirochetes, is characterized by a double membrane and an elongated helical morphology (Li et al. 2000), and can grow to lengths of 10–40 µm (Aberer and Duray 1991). Typical for spirochetes, borreliae feature a set of flagellae that run lengthwise along the cell body, in the periplasmic space between the inner and outer membrane, and enable considerable motility of the bacterium, with velocities of up to 4 µm/sec (Goldstein et al. 1994; Moriarty et al. 2008).

Borrelia burgdorferi sensu stricto is part of the B. burgdorferi sensu lato complex, which also encompasses further genospecies such as Borrelia afzelii, Borrelia garinii and others, many of which are pathogenic to humans. Borreliae typically propagate in rodents, deer, or birds and are transmitted by ticks of the Ixodidae family, with humans being inadvertent hosts (Lane et al. 1991). Once borreliae are transmitted by a blood meal, they can disseminate throughout the skin, which is often accompanied by the formation of Erythema migrans, a prominent rash that spreads from the center of infection and is enriched in neutrophils, dendritic cells, and macrophages (Salazar et al. 2003). These cells represent the first line of the host innate immune system, and their interaction with infecting borreliae is thus decisive for the outcome of a respective infection. In particular, uptake and elimination of borreliae by macrophages has been shown to be crucial to prevent dissemination of borreliae (Carrasco et al. 2015). Conversely, if an infection with B. burgdorferi proceeds unchecked, Lyme disease can also enter a chronic stage, leading to the development of neuroborreliosis or cardiac arrhythmia.

Successful uptake and elimination of infecting borreliae by macrophages involves a succession of tightly choreographed steps that are based on fine-tuned restructuring of the cytoskeleton, in particular of actin microfilaments. Macrophages form several actin-based structures during capturing and uptake of borreliae (Figs. 1 and 2). Molecular regulators of these structures, and particularly actin-regulatory factors such as formins and Arp2/3 complex, have been shown to critically influence effective phagocytosis of borreliae (Linder et al. 2001; Hoffmann et al. 2014; Naj et al. 2013). Subsequently, regulators of vesicular trafficking such as small GTPases of the Rab family, steer the internalized spirochetes towards a degradative compartment (Naj and Linder 2015). Here, we discuss the current knowledge about actin-based uptake structures formed by macrophages during phagocytosis of borreliae, the intracellular processing of internalized spirochetes, as well as the respective molecular regulators of these processes.

Fig. 1
figure 1

Uptake of borreliae by human macrophages involves dynamic restructuring of the host actin cytoskeleton. Still image from confocal time-lapse video showing a primary human macrophage expressing RFP-Lifeact (red) internalizing several GFP-expressing spirochetes (green) with actin-rich cell protrusions (white arrows). Video can be openly accessed at http://www.linderlab.de/movies. Scale bar: 10 µm

Full size image
Fig. 2
figure 2

Borrelia induces filopodia formation in human macrophages. Z-stacks of primary macrophages (stained for F-actin) without stimulation (a) or after 1 h of coincubation with Borrelia burgdorferi (b). Note formation of several filopodial protrusions after coincubation of macrophage with borreliae in (b). Scale bar: 5 µm

Full size image

2 Borrelia and the Stages of Lyme Disease

Lyme disease, also known as Lyme borreliosis, was first described in 1976 in Lyme, Connecticut, where an epidemic of juvenile rheumatoid arthritis occurred (Steere et al. 1977). In 1982, Willy Burgdorfer isolated spirochete bacteria from ticks of the Ixodes complex, which were abundant in this area. After the successful cultivation of the same spirochetes from patients, the causative agent of Lyme disease was called, based on its discoverer, B. burgdorferi (Burgdorfer et al. 1982).

To date, 18 different Borrelia genospecies are known (Margos et al. 2011). They are commonly referred to as the B. burgdorferi sensu lato complex. Among these 18 genospecies, seven have been identified to be infectious for humans, with three species being most frequently detected in patients. While B. burgdorferi sensu stricto is the most prevalent genospecies causing Lyme disease in North America, Borrelia afzelii and Borrelia garinii are the most commonly isolated human pathogenic species in Europe and Asia.

Lyme disease is a multisystemic disease that is considered to mainly originate from inflammatory responses to the Borrelia infection. The progression of Lyme disease is divided into three stages: (i) the early localized infection, (ii) the early disseminated infection, and (iii) the persistent infection (Zajkowska et al. 2012). However, not all three stages become necessarily apparent during the course of an infection.

The early, localized infection typically starts one to four weeks after transmission of borreliae with a painless skin rash spreading from the tick bite in a characteristic double ring shaped morphology called Erythema migrans (EM). EM develops through the response of immune cells, such as macrophages, neutrophils, and dendritic cells, which secrete inflammatory cytokines concomitantly with the spreading of the bacteria within the skin (Steere et al. 1983). Early infection with borreliae is often accompanied by flu-like symptoms, such as fever, malaise, and headache. If Lyme disease is diagnosed during that stage, successful treatment with antibiotics such as doxycycline or amoxicillin has a good prognosis (Jares et al. 2014). However, considering that EM occurs only in ~60–80 % of all cases, infection with borreliae remains often unrecognized and can thus progress into the stage of the early disseminated infection.

During early disseminated infection, borreliae transmigrate from the skin into blood vessels, from where they spread through the blood stream to various organs (Kumar et al. 2015; Petzke and Schwartz 2015). Depending on the site of Borrelia dissemination, the infection results in different symptoms. Therefore, beyond the originally identified manifestation in joints in the form of rheumatoid arthritis (Steere et al. 1977), patients can also develop cardiac symptoms like arrhythmia, skin lesions known as Acrodermatitis atrophicans, or Neuroborreliosis, which is accompanied by symptoms like facial palsy, meningitis, and encephalitis (Zajkowska et al. 2012; Steere et al. 2004).

Also at this later stage, patients can be cured by a prolonged treatment with antibiotics. However, in many cases symptoms persist even beyond such a regime. This stage is referred to as post-Lyme disease. It is under debate whether the symptoms are based on the presence of bacteria that persist despite the antibiotics therapy necessitating further or additional courses of antibiotics treatment (Aguero-Rosenfeld and Wormser 2015). The more widely accepted assumption, however, is that these posttreatment symptoms occur even in the absence of any remaining borreliae and are rather a consequence of damaged tissue, or of ongoing inflammation processes and autoimmune disorders (Berende et al. 2010; Pearson 2014).

To clearly answer the question whether disseminated borreliae are able to persist despite prolonged antibiotics treatment, successful cultivation of spirochetes from patient samples would be necessary. However, this is very inefficient and thus not reliable. Therefore, post-Lyme disease is still not well understood, and a clear definition of the causes as well as specific diagnostic tools are missing. The published guidelines of the “Infectious Diseases Society of America” (IDSA) advise against repeating courses of antibiotics treatment if symptoms reappear after a first antibiotic course (Wormser et al. 2006). Still, some uncertainty remains, as several studies detected persistent bacteria despite a long-term treatment with antibiotics (Stricker and Johnson 2013; Berndtson 2013). In rhesus macaques, which were infected with borreliae, low numbers of intact spirochetes were successfully recovered, despite an aggressive long-term antibiotics treatment (Embers et al. 2012). Due to these controversies, the existence of a post-Lyme syndrome, whether it results from a persistent infection, as well as its potential treatment, are still points of debate (Borgermans et al. 2014; Aguero-Rosenfeld and Wormser 2015).

3 Phagocytic Uptake of Borrelia by Macrophages

At the stage of the early localized infection, the immune system of the host can still prevent the spreading of bacteria. Thus, interaction of Borrelia with cells of the immune system, and especially with phagocytes, is critical for the outcome of the infection. EM biopsies have shown that T cells, macrophages and dendritic cells locally infiltrate the skin (Ziuzia Iu et al. 1999; Salazar et al. 2003; Duray 1989). In this review, we focus in particular on the interactions of Borrelia with macrophages, which are professional phagocytes and capable to efficiently eliminate bacteria from infected tissue.

Macrophages are part of the innate immune system, the first line of defense against infecting pathogens. At the same time, they also play a role as activators of the adaptive immune system. As professional phagocytes, macrophages are able to take up and efficiently eradicate a large number of bacteria per individual cell, through a process called phagocytosis. Phagocytosis is defined as the uptake of a particle >0.5 μm in diameter. It is a multistep process that involves detection of a phagocytic target, its capturing or immobilization, with subsequent internalization, followed by intracellular degradation.

Usually, recognition of bacteria as phagocytic targets takes place through either deposited opsonins, including factors of the complement system and antigen-targeting antibodies, or through conserved surface exposed proteins, so-called pathogen-associated molecular patterns (PAMPs). Opsonins and PAMPs are recognized by several cell surface receptors. Once a ligand–receptor interaction is established and the bacteria immobilized, macrophages develop localized protrusions, which engulf and help to internalize the pathogen. These steps of immobilization and internalization require a highly fine-tuned and localized regulation of the actin cytoskeleton. In the case of Borrelia, it is known that their phagocytosis is mediated by several different receptors, including opsonic receptors like FcɣR (Benach et al. 1984; Montgomery et al. 1994) and the complement receptor 3 (CR3) (Garcia et al. 2005; Hawley et al. 2012; Cinco et al. 1997). Furthermore, internalization of borreliae can also be mediated by the non-opsonic toll like receptor 2 (TLR2) (Salazar et al. 2009), with downstream signaling involving both myeloid differentiation factor 88 (MyD88)-dependent but also -independent pathways. (Shin et al. 2009). It has been shown that knock-out mice, which are not able to express either TLR2, Fc receptor common gamma chain (FcεRγ) or CR3 develop higher Borrelia burdens and more pronounced symptoms (Wang et al. 2004; Lawrenz et al. 2003; Liu et al. 2004). In vitro, both opsonized and unopsonized borreliae were shown to attach to macrophages. However, opsonization of the spirochetes by serum containing factors of the complement system, or by Borrelia-targeting antibodies, enhances their attachment to macrophages 4–5 fold (Linder et al. 2001). Collectively, Borrelia phagocytosis is mediated by several receptors, which act in concert to allow efficient clearance of spirochetes.

Internalization of pathogens is accompanied by their uptake into a specific intracellular compartment, the phagosome, whose coat is derived from the pathogen-engulfing membrane, which is closed upon final internalization and pinched off from the plasma membrane (Fairn and Grinstein 2012). Phagosomes then undergo a process of maturation, which is based on their fusion with endosomes and lysosomes, resulting in their acidification and the acquisition of lytic enzymes (Vieira et al. 2002). Collectively, these processes result in the degradation of the internalized pathogen. After macrophages internalize and degrade bacteria, they present antigenic peptides on their major histocompatibility complex II (MHCII) and present it to T helper cells, which in turn activate B-cells to produce antigen-targeting antibodies (Hsieh et al. 1993a, b; Kahlert et al. 2000; Unanue and Askonas 1968; Hoffman et al. 2016).

4 Actin-Rich Uptake Structures: Filopodia and Coiling Pseudopods

To initiate phagocytosis, macrophages need to establish close physical contact with pathogens. Until recently, contact formation between an immune cell and its phagocytic target was seen as a more passive event, constituting a direct consequence of chemotactic immune cell migration and being followed by surface receptor clustering (Michl et al. 1983).

However, more recent work demonstrated that immune cells also actively enhance the probability of securing a target at their surface, by probing their environment with filopodia, receptor-containing cellular protrusions (Flannagan et al. 2010). Filopodia are elongated, finger-like protrusions of cells that contain bundles of linear actin filaments and can extend from the cell surface up to several tens of microns (Svitkina et al. 2003; Mallavarapu and Mitchison 1999). They are highly dynamic and constantly extend and retract, which is based on actin filament dynamics. Considering that macrophages mostly migrate in a three-dimensional environment and are embedded in the network of the extracellular matrix, an array of actively probing filopodial protrusions allows these cells to scan a much larger volume of space, compared to their actual cell body.

Filopodia dynamics are tightly regulated. Accordingly, receptor–ligand interactions established upon capturing a target are believed to induce a signaling cascade that leads to reduction of filopodia elongation, and instead favoring their retraction (Romero et al. 2011). However, clear evidence for the existence of such a signaling cascade is missing, and potentially involved molecular regulators remain to be identified.

Importantly, filopodia are also able to exert forces in the range of several hundreds of piconewtons (pN). This allows them to pull on attached particles, thereby bringing them into close contact with the surface of the host cell (Heidemann et al. 1990; Vonna et al. 2007; Kress et al. 2007). It was demonstrated that filopodia are able to pull persistently on objects during their retraction. Accordingly, filopodia of RAW (Abelson leukemia virus-transformed murine macrophage-like) cells that pulled on IgG-coated beads (Kress et al. 2007) or sheep red blood cells (Flannagan et al. 2010) were able to resist applied counterforce by an optical trap. Interestingly, the retraction speed of the bead-pulling filopodia slowed down in relation to the force applied by the optical trap (Kress et al. 2007). In contrast, macrophages that were treated with the F-actin stabilizing agent jasplakinolide, thus inhibiting actin turnover (Cramer 1999), failed to maintain this interaction (Flannagan et al. 2010). This latter experiment emphasizes the essential role of actin cytoskeleton dynamics for filopodia-mediated capturing of phagocytic targets.

The ability to maintain the attachment to objects despite applied counterforce becomes especially important in situations when phagocytes have to capture highly motile pathogens at their surface and prevent them from detaching. B. burgdorferi, equipped with periplasmic flagella, is an excellent example for such a highly motile pathogen. Intra-vital imaging showed the spirochetes are able to move at a speed of 4 μm/sec in murine ear tissue (Moriarty et al. 2008), similar to the speed of 4.25 μm/sec measured in vitro (Goldstein et al. 1994).

Intriguingly, coincubation of primary human macrophages with borreliae strongly enhanced filopodia formation per cell as compared to control cells (3.0 ± 0.2 in borrelia stimulated cells compared to 1.2 ± 0.2 in control cells) (Naj et al. 2013) (Fig. 2). Also, Borrelia-induced filopodia were longer compared to those of unstimulated cells [6.2 μm ± 0.4 μm for filopodia in borreliae stimulated macrophages, 3.2 μm ± 0.3 μm for unstimulated macrophages (Naj et al. 2013)]. Moreover, quantification of filopodia formation in cells stimulated solely with Borrelia culture supernatant revealed no difference compared to unstimulated cells, supporting the hypothesis that induction of filopodia is based on direct interaction of macrophages with pathogens, and not on soluble factors (Hoffmann et al. 2014; Naj et al. 2013). However, the respective molecular mechanism responsible for Borrelia-induced upregulation of filopodia formation is currently unclear.

Filopodia-dependent capturing by host cells has also been demonstrated for other bacteria such as enteroinvasive Shigella flexneri (Romero et al. 2011). In this case, entry of Shigella triggers opening of connexin-dependent hemichannels, resulting in enhanced levels of extracellular ATP, which in turn increased filopodia-mediated attachment of Shigella to Hela cells. This was accompanied by enhanced Erk1/2 activity, which was shown to be important for efficient filopodia retraction. It is an intriguing speculation that similar events might also play a role in Borrelia capturing by macrophage filopodia.

Comparable to the described experiments using latex beads, motile borreliae that were contacted by macrophage filopodia often remained in direct contact with the protrusions, reinforcing the notion that filopodia indeed constitute cellular organelles for capturing of pathogens, which are able to adhere continuously to a contacted bacterium. Moreover, multiple filopodia were often observed to surround captured borreliae at the cell surface of macrophages (Fig. 3), which might also point to a role for filopodia as a physical barrier that further hinders the escape of surface-attached spirochetes.

Fig. 3
figure 3

Macrophage filopodia facilitate contact between borreliae and the host cell. 3D reconstruction using confocal Z-stacks of a macrophage stained for F-actin using Alexa 568 phalloidin (green) and expressing EGFP-mDia1 (red), with DNA of borreliae stained by Hoechst 33258 (gray). 3D reconstruction shows part of the macrophage surface that is in contact with several spirochetes, which are encaged by actin-rich filopodia. Note punctate enrichment of EGFP-mDia1 at the tips of filopodia. Scale bar: 5 µm

Full size image

After borreliae are captured and brought close to the cell surface, they are preferentially internalized by a specific mechanism called coiling phagocytosis (Rittig et al. 1992) During this process, spirochetes are progressively enwrapped by a long actin-rich cell protrusion that arises from the cell surface at the Borrelia-contact site, surrounding the spirochete in multiple whorls (Fig. 1). Experiments using live or heat-inactivated borreliae (Rittig et al. 1998b; Rechnitzer and Blom 1989) and also other spirochetes showed that coiling phagocytosis is a host cell-driven process, which is probably based on the specific morphology of spirochetes and does not depend on the viability of bacteria.

Indeed, the phenomenon of coiling phagocytosis has been known for decades and was described for the uptake of several pathogens, including Legionella pneumophila (Horwitz 1984), Trypanosoma cruzi, Leishmania spp. as well as several fungal cells (Rittig et al. 1998c). It was found that coiling pseudopods are induced to the same extent by live as by killed pathogens (Rittig et al. 1998b; Rechnitzer and Blom 1989). Furthermore, neither supernatant of bacteria culture nor by sonification fragmented bacteria triggered pseudopod formation (Rittig et al. 1998b). Moreover, it has not been possible to relate this mechanism to any specific type of phagocytic receptor (Rittig et al. 1992), though coiling phagocytosis of borreliae shows characteristics of both CR3- and Fcγ-dependent phagocytosis (Linder et al. 2001). This again suggests that the helix-like structure of the coiling pseudopod is most probably a result of the spirochete morphology rather than being based on a particular ligand–receptor interaction (Rittig et al. 1998a, c).

This coiling pseudopod is highly flexible and contains multiple bending nodes, and is thus clearly distinct from the rather stiff filopodia (Figs. 1, 2, 3 and 4). Its flexible morphology allows the pseudopod to closely align along the spirally shaped body of the Borrelia cell (Naj et al. 2013; Hoffmann et al. 2014). Individual observations by live-cell imaging have shown that full internalization can require longer than 40 min (Naj and Linder 2015).

Fig. 4
figure 4

Macrophage coiling pseudopods enwrapping borreliae are enriched in actin regulators. (a) Enrichment of GFP-Daam1∆C50 during phagocytosis of borreliae. Confocal micrographs of primary human macrophage expressing GFP-Daam1∆C50, a non-autoinactivated mutant (green) which is accumulated at the uptake structure of a Borrelia cell visualized by Hoechst 33342 staining of DNA (blue). (b) Isosurface reconstruction of confocal Z-stack, using Volocity software. Borrelia cell stained using antibody specific for OspA surface antigen (red), macrophage protrusion stained for F-actin (green) and Arp2/3 complex (blue). Macrophage cell body not shown. Note dot-like enrichment of Arp2/3 complex along the Borrelia cell, typical for coiling phagocytosis. Scale bars: 5 µm

Full size image

Considering the more recently discovered involvement of actin-based filopodia prior to the formation of the coiling pseudopod during Borrelia phagocytosis, it was unclear whether both structures are formed independently or whether the coiling pseudopod arises from enhanced lateral growth of the already existing filopodia. Importantly, live-cell imaging experiments showed that coiling pseudopods indeed constitute separate structures that are formed de novo after borreliae are captured by filopodia, and are not developed by further growth of filopodia that are already in contact with spirochetes This is also reflected by the distinct requirement for different actin regulators during the formation of either structure.

5 Actin Dynamics During Borrelia Phagocytosis: The Roles of Formins and Arp2/3 Complex

Formation and restructuring of actin filaments in cells is exquisitely controlled on many levels, to ensure exact formation of the required structures in time and space. Filaments can be formed de novo by nucleation or through fragmentation of existing ones. Elongation of filaments can be promoted or stopped by respective regulators, and dissolution is driven by processive disassembly or by severing into smaller filaments. Finally, individual filaments can be associated into higher ordered structures by bundling or crosslinking factors (Mellor 2010; Mattila and Lappalainen 2008).

Formation of filopodia involves the nucleation and elongation of unbranched actin filaments, as well as their connection in higher ordered bundles, to achieve the required stiffness. Of note, the formins FMNL1 and mDia1 have been localized to borreliae-induced filopodia, and their activity was shown to be required for filopodia formation in response to macrophage contact with borreliae, and also for subsequent internalization (Naj et al. 2013). Proteins of the formin family, 15 of which are expressed in human tissues (Schonichen and Geyer 2010), are important regulators of unbranched filaments. In principle, they are able to regulate all of the actin-related activities described above, including nucleation, elongation, capping, depolymerization, severing, and bundling with the individual set of abilities varying widely between the different isoforms (Schonichen and Geyer 2010; Grikscheit and Grosse 2016; Bohnert et al. 2013). In vitro, FMNL1 displays actin severing activity, thus giving rise to free barbed ends that can be used for the growth of new actin filaments (Harris et al. 2004), while mDia1 shows actin elongation and crosslinking activity (Li and Higgs 2003; Esue et al. 2008).

Furthermore, FMNL1 was localized along the whole shaft of borreliae-induced filopodia, whereas mDia1 showed a dot-like accumulation at the tips of filopdia (Naj et al. 2013) (Fig. 3). Combining both sets of observations, it is thus likely that mDia1 at the tips of filopodia regulates the growth of these structures by elongation of actin filaments, while FMNL1 might be involved in filopodia growth through the generation of free barbed ends along the shaft of the structure (Fig. 5). Structural stability of filopodia is supplied by bundling of actin filaments through fascin, which, comparable to FMNL1, localizes at the filopodial shaft (Naj et al. 2013). Of note, filament bundling is probably further supported through the activity of yet another formin, Daam1 (Hoffmann et al. 2014). Like FMNL1, endogenous and overexpressed forms of Daam1 were found to localize along the whole shaft of borreliae-induced and fascin-positive filopodia (Fig. 5).

Fig. 5
figure 5

Model of formin- and Arp2/3 complex dependent actin regulation in coiling phagocytosis of Borrelia. (a) Upon stimulation with borreliae, macrophages form filopodial protrusions that arise from the cortical network. Filopodia are enriched in the formins mDia1 (localized at tips) and FMNL1 (localized at tips and shaft), which probably contribute to longitudinal growth of filopodia, and Daam1 (localized in filopodial shaft), which is probably involved through its actin-bundling activity. (b) Upon capturing of a Borrelia cell by filopodia, the spirochete is enwrapped by a coiling pseudopod. Until recently, it was unclear whether coiling pseudopods develop from filopodia or constitute independent structures. Live-cell experiments showed that Daam1-positive coiling pseudopods arise as a second independent structure from the macrophage surface and enwrap borreliae. The flexibility of coiling pseudopods that enables them to enwrap the spiral-shaped borreliae is probably due to dot-like accumulations of Arp2/3 complex, which lead to formation of small branched actin networks and probably act as “hinges” at coiling nodes

Full size image

This requirement for more than one bundling factor is surprising, but fascin has been shown to stabilize Daam1 at filopodia in B16F1 mouse melanoma cells, and silencing of Daam1 in these cells led to a decrease in the number of filopodia and also defects in their architecture (Jaiswal et al. 2013), pointing to a cooperative role of Daam1 and fascin in both formation and stabilization of filopodia. Similarly, siRNA-mediated knockdown of Daam1 in human macrophages resulted in a two-third reduction of filopodia formed upon contact with borreliae (Hoffmann et al. 2014), comparable to the effect of a fascin knockdown in these cells (Hoffmann et al. 2014). Interestingly, knockdown of Daam1 had a more pronounced effect on the number of filopodia than knockdown of either FMNL1 or mDia1 (40–50 % reduction each), which may point to the relative importance of Daam1 in filopodia formation or stabilization. Combined knockdown of all three formins, however, had no additive effect, showing that these formins work in the same pathway that ensures efficient filopodia formation (Hoffmann et al. 2014). Of note, the requirement for specific formins in filopodia regulation has been shown to vary between cell types, and especially between adherent and suspension cells (Young et al. 2015). Thus, also the relative importance for FMNL1, mDia1, and Daam1 for formation of Borrelia-capturing filopodia may vary, depending on the type of immune cell involved, and also on the two- or three-dimensional context in which borreliae are encountered by immune cells in the body.

Inside–outside stainings of formin-depleted macrophages also showed that reduction of filopodia resulted in a ~50 % decrease of internalized borreliae, demonstrating that capturing of spirochetes by filopodia is an important step for efficient internalization by macrophages (Hoffmann et al. 2014). This is probably based on the effects that 1) filopodia allow cells to scan a larger volume of space and 2) immobilization of the highly motile spirochetes allows more time for the development of the coiling pseudopod, which is in most cases the decisive surface structure mediating phagocytosis of borreliae (Rittig et al. 1992; Naj et al. 2013).

Interestingly, live-cell imaging revealed that Daam1 apparently plays a dual role during uptake of borreliae by macrophages: not only through stabilization of borreliae-capturing filopodia, but also through formation of the coiling pseudopod itself (Hoffmann et al. 2014) (Fig. 4). This observation also resolved the question whether the coiling pseudopod forms independently from filopodia or through lateral growth of these structures. As the primary biochemical function of Daam1 is bundling of actin filaments, it is likely that this formin also works as an actin-bundling factor in the coiling pseudopod (Fig. 5).

This leaves the question how actin filaments in coiling pseudopods are nucleated or elongated. A partial answer to this lies in the localization of actin-nucleating Arp2/3 complex and its activator WASP (Wiskott-Aldrich Syndrome protein) at coiling pseudopods of macrophages (Linder et al. 2001). Interestingly, Arp2/3 complex has been shown to form dot-like accumulations along the coiling pseudopod, which often coincide with helical turns of the spirochete body (Fig. 4). It is thus tempting to speculate that Arp2/3 complex, as a generator of branched actin filament networks (Amann and Pollard 2001) provides nodes of branched actin that may alternate with sections of unbranched actin filaments, thus bringing the necessary flexibility to the coiling pseudopod structure that has to closely follow the helical spirochete morphology (Fig. 5). Of note, Arp2/3 complex has also been shown to be important for a subset of filopodia, by providing localized actin networks as a structural basis for their longitudinal extension (Young et al. 2015). However, the potential impact of Arp2/3 complex on borreliae-induced filopodia has not been tested yet.

An important upstream activator of WASP is Cdc42, a small GTPase of the Rho family (Mullins 2000). Accordingly, microinjection of dominant negative Cdc42 strongly reduced coiling pseudopod formation (Linder et al. 2001). It is thus very likely that an activation cascade Cdc42-WASP-Arp2/3 complex regulates formation of borreliae-induced coiling pseudopods. Furthermore, Cdc42 and other RhoGTPases, most notably RhoA or Rac1, may also be involved in the regulation of formin-dependent activities during Borrelia capturing and internalization. Formins are usually autoinhibited by backfolding of an inhibitory DID domain (diaphanous inhibitory domain) to a regulatory DAD domain (diaphanous autoregulatory domain), and only binding of RhoGTPases and other factors leads to release of this autoinhibition and to full activity of formins (Kuhn and Geyer 2014; Higgs 2005). Accordingly, many of the experiments regarding the involvement of formins in borreliae-induced filopodia formation were performed using non-autoinhibited constructs such as Daam1∆C50 that lack the respective DID domains and thus circumvent the need for RhoGTPase-dependent activation. An important role for RhoGTPases in borreliae-capturing filopodia is thus highly likely, although the involvement of specific RhoGTPases in activation of respective formins has not been tested yet.

Collectively, these data lead to the following multistep model of Borrelia phagocytosis by macrophages: (1) physical contact of borreliae with macrophages leads to increased formation of filopodia that are able to contact and bind borreliae, thus leading to immobilization of the highly motile spirochets on the bacterial surface. Filopodia formation depends on the concerted activity of three formins, FMNL1, mDia1, and Daam1, which respectively regulate actin filament formation, elongation, and bundling, with further bundling activity provided by fascin. In a second phase, a filopodia-independent structure, the coiling pseudopod, arises from the macrophage surface. It closely follows the helical spirochete morphology and thus tightly enwraps captured borreliae (Fig. 5). Actin within coiling pseudopods is probably present alternatingly as elongated unbranched filaments that are bundled by Daam1 and as nodes of branched actin networks formed by Cdc42-WASP-Arp2/3-dependent actin nucleation. This architecture would allow the necessary flexibility that is required for this structure that enwraps the helical spirochete. Finally, the coiling pseudopod has to contract, to be brought in close contact with the macrophage surface, and the captured borreliae have to be internalized. These steps likely involve regulators of actin-based contractility such as myosin II and also disassembly of actin filaments, possibly necessitating further formin activity. This should prove to be a fertile field for future research. Of note, due to their elongated morphology, intracellular processing of captured borreliae can already be in progress, as outlined below, even when extracellular parts of the spirochetes are still being enwrapped by actin-driven coiling pseudopods.

6 Intracellular Processing of Borrelia—A Central Role for RabGTPases

During phagocytosis, bacteria enter the cell in a membrane-delimited compartment termed the phagosome. Apart from the internalized target, the phagosome is initially filled with fluids that derive from the extracellular space. Subsequent alteration in phagosome composition proceeds through highly coordinated exchange of material with vesicles of the endomembrane system (Fairn and Grinstein 2012). Ultimately, these steps lead to the maturation of the phagosome into an acidic, oxidative compartment that is enriched in hydrolytic enzymes. According to the enrichment and/or loss of respective marker proteins and also to the progressive drop of the intraluminal pH, phagosomes are classified into distinct stages: (i) early phagosome, (ii) late phagosome, and (iii) phagolysosome. The fully matured phagolysosome is able to digest lipids, proteins, and carbohydrates and thus neutralizes infecting bacteria.

Key regulators of this maturation process include members of the Rab (Ras-related proteins in brain) GTPase family, which are molecular switches cycling between their active GTP-bound and inactive GDP-bound state (Vieira et al. 2002; Stenmark et al. 1994). In their active state, RabGTPases interact with their respective effector proteins that control various processes including phagosomal membrane fusion and fission events and motor-dependent transport (Gautreau et al. 2014; Hutagalung and Novick 2011).

In this respect, it could be shown that borreliae are initially internalized into a Rab22a positive phagosome. Subsequently, the phagosome is contacted by Rab5a positive vesicles. Quickly after internalization, the elongated spirochetes are compacted into globular, dense structures. This striking compaction of borreliae is accompanied by repeated tubulation of membrane from phagosomes, suggesting that reduction of the phagosomal surface could be a driving force for spirochete compaction. Interestingly, fission of membrane tubules occurs preferentially at sites where Rab5a positive vesicles contact the Rab22a positive phagosomal membrane, indicating that the concerted activity of both RabGTPases is necessary for this process (Naj and Linder 2015) (Fig. 6).

Fig. 6
figure 6

Model of intracellular processing of borreliae by macrophages. Borreliae captured by macrophages via coiling pseudopods are internalized through uptake into Rab22a-positive phagosomes. Phagosomes are subsequently contacted by Rab5a positive vesicles mediated by the endoplasmic reticulum. Subsequent membrane tubulation causes reduction of the phagosome surface, leading to visible compaction of borreliae. Further maturation of this compartment leads to its development into a degradative phagolysosome, indicated by the presence of the lyosomal marker protein LAMP1, and resulting in elimination of spirochetes. In contrast, escape from Rab22a-/Rab5a-dependent processing can lead to enhanced intracellular survival of borreliae. Modified from Naj and Linder (2015), with permission

Full size image

Previous reports showed that the endoplasmic reticulum (ER) contacts Rab5 positive early endosomes at subdomains just before fission occurs at those sites (Rowland et al. 2014). Consistent with the notion that similar processes could be involved in the membrane fission events from the phagosomes during Borrelia compaction, it was observed that the ER forms a network around Borrelia-containing phagosomes and that Rab5a vesicles contact the Rab22a-positive phagosomal membrane along the ER (Naj and Linder 2015) (Fig. 6).

Moreover, an F-actin coat was detected to surround the early, Borrelia-containing phagosome (Naj and Linder, unpublished). It is therefore conceivable that F-actin plays a role during Borrelia-compaction and the concomitant membrane tubule fission events during this process. Indeed, previous studies showed that WASH, a vesicle-localized Arp2/3 complex activator of the WASP family (Gautreau et al. 2014) regulates membrane fission from Rab5-positive early endosomes (Duleh and Welch 2010). It is thus tempting to speculate that WASH-mediated actin polymerization at phagosomes could contribute to the observed membrane fission events that occur during Borrelia compaction. However, this concept requires further evaluation.

Furthermore, Rab22a and Rab5a knockdown not only inhibited the compaction of the spirochetes, but also the maturation of Borrelia-containing phagosomes. This was demonstrated by decreased (~20–40 %) lysosomal associated membrane protein1 (LAMP1) acquisition at the phagosomal membrane (Naj and Linder 2015) and also by decreased (~15–45 %) phagosomal colocalization with DQ-BSA (Naj and Linder 2015), a marker for proteolytic activity (Fig. 6). Moreover, knockdown of Rab22a and Rab5a in macrophages led to enhanced (~6 fold) intracellular survival of the spirochetes (Naj and Linder 2015).

Of note, the specific subset of RabGTPases can vary between phagosomes that contain different bacteria. Moreover, several bacteria have evolved strategies to influence RabGTPase recruitment and/or activity as part of an escape strategy to influence phagosome maturation and thus avoid being degraded (Smith and May 2013). It is thus noteworthy that a small subpopulation of borreliae (~5 %) colocalized only transiently with Rab22a and Rab5a, did not undergo compaction and retained their elongated morphology (Naj and Linder 2015). Indeed, former electron microscopy studies demonstrated the presence of elongated borreliae localized within the cell cytoplasm without any clearly detectable phagosomal membrane (Filgueira et al. 1996; Ionescu et al. 1997; Hechemy et al. 1992). It is thus conceivable that this could be a subpopulation, which escapes the phagosome. Supporting this notion, heat killed borreliae were detected consistently surrounded by a phagosomal membrane in Vero cells (Hechemy et al. 1992). In contrast, borreliae are not known to harbor any secretion system or to express any virulence factors (Fraser et al. 1997). Closer investigation of this hypothetical subset of borreliae, their potential to persist in human immune cells, as well as the possible molecular mechanisms involved, should thus be an interesting challenge for the future.

7 Conclusions

Efficient uptake and elimination of borreliae by immune cells is crucial for countering the development of Lyme disease. In particular, macrophages form an important part of the initial defense line that prevents dissemination of B. burgdorferi within the host through a carefully orchestrated succession of capturing, internalization, and degradation of spirochetes.

In this context, recent research has highlighted the role of local restructuring of the macrophage actin cytoskeleton, which enables the formation of specific surface structures that interact with infecting spirochetes. First, macrophages respond to the presence of borreliae by forming filopodia, long, rigid protrusions that contain a core of linear, bundled actin filaments. Filopodia enable the capturing and immobilization of the highly motile spirochetes. Second, a filopodia-independent structure is formed, the coiling pseudopod, which enwraps captured borreliae and promotes their internalization. Closely following the helical morphology of the spirochete, this structure requires a more flexible arrangement of actin filaments.

Accordingly, these different structures have been shown to depend on different subsets of actin regulators, with Borrelia-induced filopodia depending the formins FMNL1 and mDia1, regulators of unbranched actin filaments, while the coiling pseudopod apparently contains nodes of Arp2/3 complex-generated branched actin networks. In addition, both structures depend on the activity of the formin Daam1.

Due to the elongated morphology of the Borrelia cell, intracellular processing of internalized parts of the spirochete can happen concomitantly with the uptake of still extracellular parts of the spirochete. Internalized borreliae have been shown to enter phagosomes, which is accompanied by successive compaction of borreliae. Further maturation of phagosomes into degradative lysosomes involves the RabGTPases Rab22a and Rab5a. The activities of both RabGTPases are coordinated by the endoplasmic reticulum, which closely enwraps the internalized parts of the spirochete, thus forming an intracellular counterpart to the coiling pseudopod that enwraps the extracellular parts of the spirochete.

Collectively, these novel insights into the subcellular and molecular regulation of Borrelia capturing, uptake, and degradation illustrate the highly efficient mechanisms that macrophages have developed to counter respective infections. On the other hand, it will be highly interesting to determine if and to which extent B. burgdorferi is able to counter these mechanisms and to thus support its dissemination in the human host.

Finally, the unique spirochete morphology of Borrelia has enabled the detection of subcellular mechanisms during uptake and processing by macrophages that would be difficult to visualize using more globular bacteria. B. burgdorferi is thus also emerging as a useful tool for the detailed study of organelle interactions during phagocytosis in general.