Complement receptor 3 mediates ruffle-like, actin-rich aggregates during phagocytosis of Leishmania infantum metacyclics

Abstract

The parasitic protozoan Leishmania infantum resides primarily in macrophages throughout mammalian infection. Infection is initiated by deposition of the metacyclic promastigote into the dermis of a mammalian host by the sand fly vector. Promastigotes enter macrophages by ligating surface receptors such as complement receptor 3 (CR3), inducing phagocytosis of the parasite. At the binding site of metacyclic promastigotes, we observed large asymmetrical aggregates of macrophage membrane with underlying actin, resembling membrane ruffles. Actin accumulation was observed at the point of initial contact, before phagosome formation and accumulation of peri-phagosomal actin. Ruffle-like structures did not form during phagocytosis of attenuated promastigotes or during phagocytosis of the intracellular amastigote form of L. infantum. Entry of promastigotes through massive actin accumulation was associated with subsequent delay in fusion of the parasitophorous vacuole (PV) with the lysosomal markers LAMP-1 and Cathepsin D. Actin accumulation was also associated with entry through CR3, since macrophages from CD11b knockout (KO) mice did not form massive aggregates of actin during phagocytosis of metacyclic promastigotes. Furthermore, intracellular survival of L. infantum was significantly decreased in CD11b KO compared to wild type macrophages, although entry rates were similar. We conclude that both promastigote virulence and host cell CR3 are needed for the formation of ruffle-like membrane structures at the site of metacyclic promastigote phagocytosis, and that formation of actin-rich aggregates during entry correlates with the intracellular survival of virulent promastigotes.

Graphical Abstract

Introduction

The Leishmania spp. protozoa cause a diverse spectrum of clinical syndromes in infected humans, collectively called leishmaniasis. Different Leishmania species lead to distinct clinical manifestations, with the common thread that all are obligate intracellular parasites primarily residing in mononuclear phagocytes of the infected mammalian host [1]. Leishmania spp. have a life cycle consisting of two stages, namely, the promastigote, and the amastigote. In the gut of the sand fly vector, promastigote replication is followed by metacyclogenesis, a developmental process in which parasites acquire virulence factors and become infectious. When sand flies take a blood meal, metacyclic promastigotes are inoculated in the dermis of the mammalian host. Promastigotes facilitate their own receptor-mediated phagocytosis by attaching to receptors on mononuclear phagocytes. About 48 hours after phagocytosis, promastigotes transform into the obligate intracellular amastigotes. Thereafter, amastigotes replicate intracellularly, spread to additional phagocytes, and cause disease [2–5].

Many reports document the ability of Leishmania spp. to enter macrophages silently and evade intracellular killing [6, 7]. Promastigotes enter macrophages through facilitated phagocytosis after ligation of surface receptors. Receptor-mediated phagocytosis results in the formation of a membrane-bound phagosome that proceeds through a series of fusions with endocytic compartments, eventually fusing with lysosomes. The phagolysosome can be the site of microbicidal activity and degradation of ingested substances [4, 5]. Desjardins and colleagues have shown that L. donovani promastigotes reside in parasitophorous vacuoles (PVs) that are surrounded by polymerized actin during the first hours following phagocytosis [8, 9]. In contrast to the usual phagosome, fusion of Leishmania-harboring PVs with lysosomes is delayed for hours after phagocytosis, a delay that could allow the parasite to initiate conversion to the more resilient amastigote parasite stage [4, 8–10]. The promastigote and amastigote stages of Leishmania spp. differ morphologically as well as in the molecules displayed on their surface molecules [3]. Promastigote surface molecules include the phosphorylated glycolipid lipophosphoglycan (LPG) and the major surface protease [11] amongst others [12]. MSP facilitates promastigote attachment to macrophages at least in part via ligating the complement receptor 3 (CR3), either directly or indirectly through opsonized C3bi [4, 5]. LPG is required for both peri-phagosomal actin accumulation and for the delay in fusion of promastigote-containing PVs with lysosomes [4, 13]. LPG is downregulated at the amastigote surface, and MSP is present internally rather than retaining high surface expression [4, 5].

Recognizing the differences between parasite life stages, the present study was based on the hypothesis that the pathway of macrophage entry may be uniquely suited to optimize the survival of the different life forms of Leishmania spp. Our studies were conducted with a Brazilian strain of L. infantum, a causative agent of visceral leishmaniasis in Latin America and parts of the Old World [14]. Using confocal microscopy, we documented the transient accumulation of actin-based “ruffle”-like structures induced at the site of phagocytosis of the highly virulent, metacyclic promastigotes. These structures resembled the ruffles reported during phagocytosis of Neisseria gonorrhoeae by cervical cells [15]. Membrane ruffles have also been reported during the uptake of Salmonella enterica in human and avian epithelial cells [16]. Similarly, membrane ruffles have been reported during phagocytosis of Salmonella typhimurium by various cell types including macrophages and intestinal epithelial cells [17]. Furthermore, membrane ruffling has been documented in an increasing number of models including the yeast Cryptococcus neoformans and the protozoa parasites Toxoplasma gondii and Trypanosoma cruzi [18, 19]. Membrane ruffling facilitates microbe uptake through macropinocytosis. The implications of these events in the host cell and the potential effects in infection are just beginning to be explored [20, 21]. In this report we showed that uptake of the highly virulent, metacyclic promastigotes of L. infantum occurred through the induction of massive actin-rich aggregates similar to macrophage membrane ruffles. In contrast, actin aggregates did not form around incoming virulent amastigotes or attenuated promastigotes. Importantly, structures resembling ruffles did not form in the absence of host cell CR3. Entry rates of metacyclic promastigotes were similar in wild type (WT) and CR3 deficient, CD11b knockout (KO), macrophages. However, survival of L. infantum was significantly decreased in cells of CD11b KO origin compared with WT macrophages. Our data suggest that these actin-based structures are triggered by the initial contact between highly virulent promastigotes and the macrophage receptor CR3. Additionally, our results support the notion that entry through these actin-rich structures is exploited by the parasite to facilitate its intracellular survival.

Materials and methods

Parasites

The Brazilian L. infantum strain (MHOM/BR/00/1669) used in this study was isolated from a patient in Northeast Brazil with visceral leishmaniasis. Parasites were serially passed in male Syrian hamsters to maintain virulence. Promastigotes were cultivated in hemoflagellate-modified minimal essential medium (HOMEM), supplemented with 10% heat-inactivated fetal calf serum (GIBCO, Carlsbad, CA), pH 7.4 at 26 °C [11]. To ensure virulence, promastigotes were used for experiments within 3 weeks of isolation from a hamster.

Metacyclic L. infantum promastigotes were isolated from late stationary growth phase cultures (7-9 days of culture), and separated by density on a Ficoll density gradient, modified from the original procedure developed for L. major metacyclic promastigotes [22, 23]. The number of metacyclic promastigotes in L. chagasi (aka L. infantum) cultures progressively decreases to less than 2% of the population after just 8 serial passages [24, 25]. The attenuated L5 strain is a clonal line derived from the original L. infantum isolate that has been attenuated by serial passage in liquid culture for more than 10 years and was used as described [26]. Amastigotes were isolated from the spleens of infected Golden Syrian male hamsters, and incubated overnight in growth medium containing 20% FCS (GIBCO) at 37°C and 5% CO₂, pH 5.5 as described [27, 28]. Amastigotes were pelleted at 2060 x g. This method allows spontaneous detachment of host PV membranes and the isolation of pure amastigote cultures [29]

Mice

WT BALB/c and C57BL/6 mice were purchased from Harlan Laboratories (Indianapolis, IN). Animals were housed at the Animal Care Facility of the Iowa City VAMC and procedures followed protocols approved by the VA IACUC. Bones from CD11b KO or WT control mice were kindly harvested at the University of Notre Dame, and sent to the University of Iowa via overnight courier. The CD11b KO line on each background was generated from a CD11b heterozygote cross at the University of Notre Dame as described [30].

All work with vertebrate animals was performed according to the Guide for Care and Use of Laboratory Animals of the NIH. Animal use protocols were approved by the Iowa City VA Medical Center Animal Care Committee (ACORP 1390501 and 1190302), and the University of Iowa Animal Care and Use committee (ACURF 1107161 and 1390502). Bone marrow for isolation of murine macrophages was collected after euthanasia with CO₂ using approved methods to reduce animal suffering.

Murine macrophage isolation and infection

Bone marrow derived macrophages (BMMs) were cultured at 37°C, 5% CO₂ in RP-10 (RPMI with 10% heat-inactivated fetal calf serum, 2 mM L-glutamine, 100 U of penicillin/ml, and 50 μg of streptomycin/ml (GIBCO, Carlsbad, CA) containing 20% L929 cell culture supernatant (American Tissue Type Collection, Manassas, VA) as a source of macrophage colony-stimulating factor. After 7 to 9 days, differentiated adherent BMMs were detached from Petri dishes with 2.5 mg trypsin/ml plus 1 mM EDTA (GIBCO), and cultivated overnight on glass coverslips prior to use.

Metacyclic L. infantum, attenuated L5 promastigotes, hamster-derived L. infantum amastigotes, virulent S. typhimurium or GFP-latex beads were opsonized with 5% C5-deficient A/J mouse serum for 30 min at 37°C. Five x 10⁵ BMMs adherent to glass coverslips were infected with L. infantum parasites as described. Briefly, L5 or metacyclic L. infantum promastigotes were used at a multiplicity of infection [MOI] of 5:1; amastigotes were used at a MOI of 2:1. Infections were synchronized by centrifugation (3 min, 330 x g, 4 °C), and macrophages incubated at 5% CO₂, 37°C. After 30 minutes, extracellular particles were removed by rinsing and macrophages were returned to 37°C, 5% CO₂. At specified time points, triplicate coverslips were fixed and stained with Wright Giemsa (Diff-Quik, Fisher Scientific) [26, 28].

Confocal microscopy

Metacyclic L. infantum promastigotes, attenuated L5 promastigotes or L. infantum amastigotes were stained with carboxy-fluorescein diacetate succinimide ester (CFSE) as described [31]. Macrophages on 12 mm coverslips were infected with CFSE-stained parasites as described above. Control particles for the induction of membrane ruffling included green fluorescent protein (GFP)-labeled virulent S. typhimurium or GFP latex beads as positive and negative controls, respectively.

At specified times points, infected or control cells were fixed in 2% paraformaldehyde (30 minutes) in PBS, permeabilized in 0.2% Triton X-100 (15 minutes), incubated in 50 mM glycine (15 minutes), and blocked in 5% non-fat dry milk/PBS (30 minutes). Macrophages were incubated with one of the following primary antibodies, diluted 1:50 in 5% non-fat dry milk/PBS overnight at 4°C: rat 1D4B anti-LAMP-1 (developed by J. Thomas August, Developmental Studies Hybridoma Bank, University of Iowa), goat polyclonal anti-Cathepsin D (Santa Cruz SC-6486), or goat anti-Talin (Santa Cruz SC-7534). After two rinses, macrophages were incubated in 1:200 secondary antibodies for 1 hour at room temperature. Secondary antibodies were Alexa Fluor 568 (red) goat anti-rat (LAMP-1), Alexa Fluor 568 (red) donkey anti-goat (Cathepsin D) or 1:200 Alexa Fluor 647 (pseudo blue) donkey anti-goat (Talin), all from Molecular Probes. Actin was stained with Alexa Fluor 647 (pseudo red) Phalloidin at 1:50 (Molecular Probes Inc., Eugene, OR). Slides were rinsed in PBS, mounted with Vectashield H-1000 (Vector Labs, Burlingame, CA), and examined on a Zeiss 510 laser scanning confocal microscope. Images were captured using LSM 510 version 3.2 software. Confocal optical sections were analyzed using the LSM 5 image browser (Carl Zeiss). Microscopic studies were performed at the University of Iowa Central Microscopy Research Facility.

Quantification of actin-rich structures

Quantitative analysis of actin-rich structures was performed by examining confocal images of more than 200 BMMs with attached or intracellular parasites at each time point. BMMs lacking ruffle-like structures were defined as cells in which phalloidin-stained actin conformed to the shape of the invading parasite at the site of phagosomes formation, and actin was symmetrically distributed around the macrophage periphery. BMMs with ruffle-like structures were defined as macrophages in which the density of phalloidin-stained actin at the cell periphery was not symmetrical, and in which actin surrounding the incoming parasite extended beyond the shape of the parasite.

Light microscopy

At varied time points after infection, coverslips were blow dried, fixed with methanol, and stained with Diff Quik (Wright Giemsa) stain. Parasite loads were quantified by light microscopy examination of more than 200 macrophages in triplicate coverslips per condition at each time point. Percent of infection was expressed as the percent of BMMs containing intracellular or attached parasites. Parasite load was expressed as the total number of intracellular plus attached parasites per 100 total BMMs.

Statistical Analysis

Statistical analyses were performed using T-test as implemented in GraphPad Prism 8 (GraphPad Software, Inc, CA USA).

Results

Virulent promastigotes and amastigotes replicate in macrophages, but attenuated promastigotes do not

As our purpose was to contrast macrophage entry by L. infantum parasites of differing stages and virulence, it was necessary to first characterize the virulence of these forms in vitro. Through metacyclogenesis, promastigotes acquire virulence factors and become highly infectious [2, 23]. As such, we used metacyclics as a model of highly virulent promastigotes. The attenuated L5 line, which is defective in its ability to infect mice was used as a model for attenuated promastigotes [26]. Promastigotes and amastigotes differed in the expression of the virulence markers LPG and GP63 [32]. In addition, compared to metacyclics, L5 promastigotes showed lower expression of LPG and GP63 (Fig 1A, left and right panels, respectively).

Fig 1. Virulence characteristics of L. infantum parasites.

(A) Whole cell lysates of metacyclic L. infantum promastigotes (M), stationary phase attenuated L5 promastigotes (L5) or hamster-derived amastigotes (AM) were subjected to immunoblot analysis with either antiserum CA7AE which detects the unmodified phosphoglycan repeats in LPG, or with polyclonal antiserum to L. infantum MSP (GP63). Blots were re-probed with antibody to β-tubulin as a loading control. (B) BALB/c BMMs on glass coverslips were incubated with opsonized L5 or metacyclic L. infantum (Li) promastigotes (MOI 5:1) or amastigotes (MOI 2:1). Extracellular parasites were removed after 30 minutes, and macrophages were incubated at 37°C, 5% CO₂ for the indicated times. Coverslips were stained (Wright-Giemsa) and the number of intracellular parasites was quantified using light microscopy. Data represent the mean ± SE of the percent of infected macrophages (left panel) or the macrophage-associated parasites per 100 macrophages (right panel). Values were assessed by counting 200 cells in triplicate coverslips for each condition, in three separate experiments. Statistical analysis showed no difference in parasite entry (1hour time point) between all the tested groups. P-values (t-test) compare the numbers of intracellular parasites in macrophages infected with metacyclic (Li) promastigotes versus attenuated (L5) promastigotes (**p<0.005), (***p<0.001).

BMMs were incubated with metacyclic L. infantum promastigotes, attenuated L5 promastigotes, or virulent amastigotes isolated from infected hamsters. The initial rate of parasite uptake by macrophages, shown either as the percent of macrophages infected or the number of parasites per 100 macrophages was similar for all three parasite types (Fig 1B, left and right panels, respectively). Although some L. infantum metacyclic promastigotes were cleared from macrophages after 24 hours of infection, intracellular replication of both metacyclic promastigotes and amastigotes was apparent over 72 hours. In contrast, despite a similar rate of entry, attenuated L5 promastigotes were rapidly destroyed over the initial 48 hours and the parasite load was lower than 4% by 72 hours after infection.

Actin-rich structures during phagocytosis of Leishmania: differences between parasite forms

We previously showed that L. infantum promastigotes and amastigotes exploit distinct host membrane microdomains to enter macrophages and establish infection [28]. Considering that differences in parasite uptake could influence intracellular survival, we compared macrophages during phagocytosis of each parasite form. CFSE-labeled parasites were incubated with BMMs for 15 minutes, fixed, and processed for confocal microscopy. GFP-labeled virulent S. typhimurium or GFP-latex beads served as the respective positive or negative controls for phagocytosis-induced actin ruffling (Fig 2A).

Fig 2. Virulent L. infantum promastigotes induced ruffle-like actin structures during phagocytosis.

BMMs were incubated with opsonized metacyclic L. infantum (Li) promastigotes, L5 promastigotes, or Li amastigotes as described. Parasites were labeled with CFSE (green). GFP-labeled virulent S. typhimurium (GFP-St) was used as a positive control for phagocytosis-induced membrane ruffling whereas GFP latex beads were used as a negative control. After 15 minutes at 37°C, 5% CO₂, extracellular parasites/particles were rinsed off, cells were fixed in 2% paraformaldehyde and stained, and BMMs were visualized by confocal microscopy. (A) Macrophages were stained with phalloidin (red) to visualize actin after phagocytosis of various particles (green). Yellow: co-localization of actin and CFSE-parasites. Images shown are representative of 3 independent experiments. Scale bar = 10 μm. (B) The number of macrophages with ruffle-like aggregates of actin (defined in the Materials and Methods section) were quantified microscopically 15 minutes after infection. More than 200 macrophages were enumerated per condition, and triplicate experiments were quantified. Statistical comparisons between Li promastigotes and other conditions were performed by t-test (**p<0.005), (***p<0.001). (C) Macrophages containing metacyclic Li promastigotes (green) were stained with phalloidin (red) and antibody to the actin-binding protein Talin. Images shown are representative of 3 independent experiments. Pink: co-localization of Talin and phalloidin. Scale bar =10 μm.

Phagocytosis of virulent metacyclic L. infantum induced an intense network of macrophage membrane extensions and projections which contained a high concentration of filamentous (F)-actin, as detected by phalloidin staining. In contrast, very few, if any, actin extensions were observed in the membranes of BMMs incubated with attenuated L5 promastigotes or L. infantum amastigotes (Fig 2A). The proportions of macrophages with parasite-induced ruffle-like actin aggregates were quantified using the criteria outlined in the Materials and Methods section. Fifteen minutes after exposure to metacyclic L. infantum promastigotes, 75% of BMMs exhibited actin-rich aggregates near the site of phagocytosis. This compared to the formation of actin-rich structures in 4% of BMMs incubated with attenuated L5 promastigotes, or 12% of BMMs incubated with hamster-derived amastigotes (Fig 2B). Talin is an actin-binding protein and a component of the functional module linking actin to cell-matrix adhesion receptors (integrins) [33, 34]. Consistent with this role, aggregated Talin colocalized with phalloidin-stained, actin-rich, ruffle-like structures during phagocytosis of metacyclic L. infantum promastigotes (Fig 2C).

Requirements for Leishmania induced actin-rich aggregates

Because actin polymerization seemed to initiate at the point of promastigote contact with the macrophage membrane, we investigated the requirement for CR3, one of the main molecules facilitating receptor-mediated phagocytosis of Leishmania spp. promastigotes [35, 36]. Mice deficient in CD11b, the unique subunit of CR3, were backcrossed onto a BALB/c background [30, 37]. We then examined Leishmania phagocytosis by BMMs obtained from either WT or CD11b KO BALB/c mice. Phagocytosis of L. infantum metacyclic promastigotes resulted in a significant accumulation of transient actin-rich aggregates in macrophages derived from WT versus CD11b KO mice (Figs 3A and 3B, respectively). The percentage of WT BMMs with actin aggregates was 84% by 15 minutes of infection, but by 2 hours decreased to 15%. In contrast, CD11b KO BMMs expressed few actin aggregates at 15 minutes (16%) and 2 hours (12%) (Fig 3C).

Fig 3. Actin remodeling during phagocytosis of L. infantum promastigotes required complement receptor 3.

BMMs derived from WT or CD11b KO (CR3 deficient) BALB/c mice were incubated with opsonized L. infantum metacyclic promastigotes at a MOI of 5:1. The infection was synchronized by centrifugation, and macrophages were incubated for the indicated times at 37°C, 5% CO₂. Samples were fixed, stained, and examined by confocal microscopy. (A) The arrow marks an example of a parasite associated with an actin ruffle-like formation. (B) The arrowhead marks a promastigote that is not associated with actin aggregates. Red phalloidin: actin, Green: CFSE-stained parasites, Yellow: Actin + parasite co-localization. Scale bar = 10 μm. (C) The graph shows the percent of macrophages with actin-rich, ruffle-like formations at the indicated times. Results were obtained from counting at least 200 macrophages per condition at each time point in 3 replicate experiments. Statistical comparisons between WT and CD11b KO macrophages were performed by t-test (**p<0.005).

We questioned whether actin-rich aggregates are induced by a different species of Leishmania or in macrophages from a different mouse strain. Similar to our observations in BALB/c mice which have a non-self-healing d/d MHC haplotype, ruffle-like structures were also formed in macrophages from C57BL/6 which have a self-resolving b/b MHC haplotype [38]. Specifically, actin-rich structures were formed upon incubation of C57BL/6 macrophages with either L. donovani or L. infantum metacyclic promastigotes (S figs 1A and 1B). Taken together, these results suggest that uptake through ruffle-like structures at the macrophage membrane is a mechanism utilized by visceralizing species of Leishmania across different host genetic backgrounds and healing phenotypes.

Maturation of parasitophorous vacuoles containing metacyclic versus attenuated L. infantum promastigotes

The work of Desjardins and Descoteaux has shown that PV fusion with lysosomes is delayed for 3-4 hours after phagocytosis of L. donovani promastigotes [8, 9]. In the current study, we observed actin-rich aggregates in the macrophage membrane during the phagocytosis of metacyclic promastigotes. However, actin aggregates were not observed during the uptake of attenuated promastigotes. Variations in uptake may affect PV development and intracellular survival. As such, we then examined how metacyclic and attenuated parasites compared in the kinetics of phagosome maturation. BALB/c BMMs were infected with attenuated L5 promastigotes or the highly virulent, L. infantum metacyclic promastigotes. Confocal microscopy was used to assay the association of PVs with the endosome/lysosome system. Cathepsin D was used as a lysosome marker and Lysosomal Associated Membrane Protein-1 (LAMP-1) as a marker of late endosomes/early lysosomes. Bright patches of focused staining surrounding or co-localizing with PVs would suggest fusion with late endosomes/early lysosomes (LAMP-1) or lysosomes (Cathepsin D). Overall, 8 hours after infection, few metacyclic-containing PVs colocalized with LAMP-1 or Cathepsin D (Figs 4A and 4B, upper panels).

Fig 4. Differential maturation kinetics of parasitophorous vacuoles containing virulent versus attenuated promastigotes.

BMMs were infected with either CFSE-labeled virulent metacyclic L. infantum promastigotes (Li) or with attenuated L5 promastigotes (L5) at a MOI of 5:1. After 30 minutes, extracellular parasites were removed by rinsing twice in PBS, and macrophages were returned to 37°C, 5% CO₂ in fresh medium. Confocal microscopy was used to assess co-localization of more than 200 PVs with the phagosomal maturation markers LAMP-1 (A) or Cathepsin D (B). Micrographs show macrophages after 8 hours of infection. Stains are CFSE-labeled promastigotes (green), LAMP-1(red; panel A), and Cathepsin D (red; panel B). The micrographs shown are representative of 3 replicate experiments. Scale bar = 10 μm. (C) Data were quantified for the numbers of promastigotes in compartments positive for LAMP-1 or Cathepsin D, counting 100-200 macrophages from each condition in 3 separate experiments. Statistical analysis (t-test) compares PVs containing highly virulent metacyclic L. infantum versus attenuated L5 promastigotes that were LAMP-1 or Cathepsin D positive (*p<0.05).

In contrast to metacyclic promastigotes, most attenuated L5 parasites localized in compartments that stained positive for LAMP-1 as well as Cathepsin D (Figs 4A and 4B, lower panels). Eight hours after infection, quantitative analyses indicated that 71% of attenuated parasites co-localized with LAMP-1. In contrast, only 16% of metacyclic parasites colocalized with LAMP-1 during the same 8-hour infection period (p<0.05) (Fig 4C). In a similar pattern, 8 hours after infection, most PVs containing L5 promastigotes acquired Cathepsin D (81%). This contrasts with PVs surrounding metacyclic promastigotes, amongst which only 13% were Cathepsin D positive after 8 hours of infection (p<0.05) (Fig 4C).

Maturation of amastigote-containing parasitophorous vacuoles

The expression of surface molecules in amastigotes differs from that of promastigotes (e.g., LPG, MSP; see Fig 1), and the morphology of the pseudopod surrounding each of the parasites during phagocytosis also differs [22]. The current work showed that uptake of metacyclic promastigotes resulted in actin-rich aggregates at the macrophage membrane followed by delayed PV maturation. In contrast, amastigote phagocytosis barely induced actin-rich aggregates. We hypothesized that there would be a corresponding difference in the kinetics of phagosome maturation between the two parasite stages. As such, we assayed the acquisition of maturation markers in amastigote-containing PVs. The percentage of LAMP-1 positive PVs steadily increased from 46% (2 hours), to 56% (8 hours) to 74% (24 hours) (Figs 5A and 5C). Similarly, the percentage of Cathepsin D positive PVs increased from 39% (2 hours), to 50% (8 hours) to 67% (24 hours) (Figs 5B and 5C). Hence, contrary to metacyclic promastigotes, amastigote-containing PVs rapidly accumulated endosomal/lysosomal markers after phagocytosis.

Fig 5. Recruitment of LAMP-1 and Cathepsin D to amastigote-containing parasitophorous vacuoles early after infection.

BMMs from WT BALB/c mice were incubated with L. infantum hamster-derived amastigotes at an MOI of 2:1. After 30 minutes, extracellular parasites were removed by rinsing twice in PBS, and macrophages were returned to 37°C, 5% CO₂ in fresh RP-10. The samples were fixed with 2% paraformaldehyde and processed for confocal microscopy at the indicated times after infection. (A) LAMP-1 (red), (B) Cathepsin D (red), Green: CFSE-stained parasites, Yellow: parasite co-localization with LAMP-1 or Cathepsin D (Panels A and B, respectively). Scale bar = 10 μm. (C) Graphs show quantitative analyses of the numbers of amastigotes in compartments positive for LAMP-1 or Cathepsin D. Data are derived from 3 separate experiments in which at least 200 macrophages were quantified per condition at each time point of each experiment.

Influence of CR3 on parasitophorous vacuole maturation and intracellular survival

The data herein suggest that ruffle-like actin aggregates play a role in parasite uptake and trafficking. As such, it is reasonable to speculate that the intracellular survival of L. infantum could be affected in the absence of CR3, a condition in which ruffle-like structures do not form (Fig 3). To investigate this possibility, we incubated BMMs from WT or CD11b KO of BALB/c origin with the highly virulent, metacyclic L. infantum promastigotes. Surprisingly, the absence of the surface receptor CR3 did not alter the rate of initial phagocytosis of opsonized metacyclics by macrophages (Fig 6A, left and right panels). However, there was a significant defect in the ability of parasites to replicate in CD11b KO compared to WT (Fig 6A, right panel). L. infantum infection of WT and CD11b KO macrophages from C57BL/6 mice showed a similar pattern in which initial uptake was not compromised but intracellular survival was impaired in the absence of CR3 (S fig 1C).

Fig 6. Effects of CR3 on maturation of parasitophorous vacuoles and intracellular survival.

(A) BMMs derived from either WT or CD11b KO BALB/c mice were incubated with opsonized L. infantum metacyclic promastigotes at a 5:1 MOI for the indicated number of hours. Extracellular parasites were removed by rinsing after 30 minutes, and incubated for the indicated time points as described. Coverslips were fixed, stained with Wright-Giemsa and examined by light microscopy. Graphs show the mean ± SE of the percent of infected macrophages (left panel) or the mean ± SE of intracellular parasites per 100 macrophages (right panel) in 3 replicate experiments, each with triplicate conditions. Statistics were done by t-test (*p <0.05). (B) BMMs derived from either WT or CD11b KO BALB/c mice were incubated with opsonized metacyclic L. infantum promastigotes at a 5:1 MOI (left panel) or opsonized hamster-derived amastigotes at a 2:1 MOI (right panel). Samples were fixed, stained, and examined by confocal microscopy either 2 or 8 hours after infection. Data represent the mean ± SE percentage of promastigotes or amastigotes residing in compartments positive for LAMP-1 at each time point. Numbers of parasites surrounded by a LAMP-1 positive compartment were enumerated in 100 macrophages from each coverslip. Three replicate coverslips per condition were counted in each experiment, and 3 separate experiments were performed. P values were determined by t-test (*p<0.05).

Corresponding to the differences in survival, maturation of PVs containing metacyclic promastigotes was delayed in WT compared to CD11b KO BMMs. Quantitative analysis of confocal micrographs showed that the proportions of LAMP-1+ PVs increased between 2 and 8 hours after infection in both WT and CD11b KO BMMs. However, the proportion of LAMP-1 + PVs was significantly higher in CD11b KO compared to WT BMMs (p<0.05) (Fig 6B, left panel). In contrast, LAMP-1 fusion with amastigote-containing PVs was comparable between WT and CD11b KO macrophages (Fig 6B, right panel). These results suggest that CR3 does not affect the fusion of amastigote-containing PVs with the endosome/lysosome system. In addition, the levels of LAMP-1 fusion with amastigote-containing PVs remained at similar levels during the 2 to 8 hours post-infection period (Fig 6B, right panel). Overall, in the initial 8 hours of infection, LAMP-1 fusion was higher in amastigote-than promastigote-containing PVs (Fig 6B, right versus left panel).

Discussion

Obligate intracellular microbes employ a variety of mechanisms that enable them to evade the microbicidal responses of the host cell [39]. The Leishmania spp. protozoa encounter mammals first as an extracellular promastigote inoculated by the sand fly vector and then develop into an obligate intracellular amastigote that survives inside the macrophage. Both parasite stages must avoid inducing phagocyte microbicidal responses. Extracellular Leishmania spp. promastigotes enter macrophages by facilitated phagocytosis after ligation of macrophage surface receptors [4, 14]. Accumulating evidence suggest that the earliest steps during pathogen uptake are important determinants in intracellular survival and hence, the outcome of infection [26, 28, 40].

The current study is based on the hypothesis that the mechanisms activated by parasite entry may reveal critical strategies for intracellular survival in the host macrophage. As a means of addressing this hypothesis, we contrasted macrophage entry between the intracellular amastigote and the extracellular promastigote stages. In addition, we contrasted macrophage entry between highly virulent metacyclics and attenuated promastigotes forms of L. infantum. Using confocal microscopy, we found that metacyclic L. infantum promastigotes were taken up through large aggregates of polymerized actin. These actin aggregates resembled the membrane “ruffles” reported to surround the phagocytosis sites of S. typhimurium, S. enterica or N. gonorrhea [16, 17, 41]. In contrast, attenuated promastigotes as well as virulent amastigotes entered through phagocytic mechanisms that did not induce the accumulation of actin-rich structures.

Actin-rich, ruffle-like structures formed at the contact site of metacyclic promastigotes with the macrophage membrane before phagosome formation. Thus, these asymmetrical actin accumulations are an earlier event than the peri-phagosomal accumulation of actin previously reported to delay lysosome fusion in L. donovani infection [42]. In our study, we observed actin-rich aggregates during uptake of either L. infantum or L. donovani by macrophages from C57BL/6 mice. In addition, we showed that ruffle-like structures formed upon phagocytosis of L. infantum metacyclics by either self-healing C57BL/6 or non-healing BALB/c mice [38]. Taken together, these results suggest that massive accumulation of actin is an early phagocytosis event induced by either visceralizing parasite species regardless of the ultimate healing or non-healing phenotype of the host.

Multiple macrophage surface receptors are implicated in the uptake of Leishmania parasites. Promastigote entry is mediated by CR1, CR3, Toll-like receptors (TLR)-1, -2, -4 and -6 as well as the fibronectin, scavenger, and mannose receptors [43–48]. In addition, Fc-gamma and phosphatidylserine (PS) receptors are implicated in amastigote entry [49]. The outcome of infection is influenced by the receptors ligated by the parasites and the ensuing signaling cascades in the phagocyte. Correspondingly, to facilitate their intracellular survival, Leishmania spp. modulate the expression of macrophage receptors involved in pathogen recognition and phagocytosis, as well as the signaling cascades triggered by these receptors [47, 48, 50].

Various models have documented the CR3-mediated phagocytosis of Leishmania promastigotes by macrophages [4, 5]. In human macrophages, entry of non-virulent, logarithmic L. infantum promastigotes is preferentially mediated by mannose receptor over CR3. In contrast, entry of highly virulent, metacyclic L. infantum promastigotes is overwhelmingly mediated by CR3 [36]. The CR3 entry pathway may facilitate promastigote “silent” entry into macrophages, ultimately leading to parasite survival [4, 14, 39, 51, 52].

In this study, using WT versus CD11b KO macrophages, we extend these results by showing that CR3 was necessary for the formation of actin-rich aggregates at the macrophage membrane during Leishmania phagocytosis. In addition, we showed that uptake of metacyclic promastigotes occurred through CR3-mediated ruffle-like structures. These observations are reminiscent of N. gonorrhoeae entry into human cervical cells. In that system, ruffle formation requires functional CR3 and does not occur in cell lines negative for the CR3 receptor [15, 41]. However, our study showed that during L. infantum uptake, formation of CR3-mediated, ruffle-like structures is more complex. For example, entry of attenuated promastigotes or virulent amastigotes did not lead to actin-rich aggregates. Formation of CR3-mediated, ruffle-like structures, only occurred during uptake of the highly virulent metacyclic promastigotes. Thus, the data showed that parasite entry through CR3-mediated, actin-rich aggregates is stage- and virulence-dependent. Furthermore, the cytosolic protein Talin, which provides a link between integrins and the cytoskeleton during integrin mediated adhesion, was also observed in nascent ruffle-like structures at the site of incoming metacyclics [53]. Taken together, these results suggest that CR3 contact by virulent promastigotes trigger events reminiscent of macrophage membrane ruffles.

Overall, phagocytosis of metacyclic promastigotes induced actin-rich, ruffle-like structures and was associated with PVs that experienced delayed maturation. This was observed as the limited accumulation of the late endosome and lysosome markers LAMP-1 and Cathepsin D during the initial 8 hours of infection. Neither virulent amastigotes nor attenuated L5 promastigotes induced actin aggregates during phagocytosis and their PVs rapidly acquired LAMP-1 and Cathepsin D. The localized actin-rich, ruffle-like structures that we observed at the site of macrophage attachment were transient and formed at an earlier step than the retention of peri-phagosomal F-actin reported to surround L. donovani PVs [42]. Aggregates of polymerized F-actin surrounding L. amazonensis promastigotes after phagocytosis have also been reported by Courret et al. [54]. Both a delay in phagosome maturation and accumulation of peri-phagosomal actin have been attributed to the parasite surface glycolipid LPG [4, 5, 39, 42]. It is likely that the CR3-mediated, actin-rich aggregates observed in this study also require parasite ligands for this receptor, possibly bridged by opsonization with the CR3 ligand iC3b.

Collectively, our work and that of others, suggest that there are two steps of Leishmania promastigote phagocytosis in which there are unusual actin accumulation. First, actin-based, ruffle-like structures form around the site of metacyclic attachment to the macrophage surface, mediated through the phagocytic receptor CR3. Second, after entry, virulent promastigotes reside in a PV that becomes enveloped by polymerized F-actin, in a process dependent on parasite LPG. Both events are associated with the subsequent delay in PV maturation and lysosomal fusion [4, 39]. Furthermore, Leishmania infection could also have long term effects in actin reorganization. For example, infection with L. amazonensis leads to alterations in F-actin dynamics that impairs migration of infected macrophages. Specifically, compared with non-infected controls, 24 hours of L. amazonensis infection resulted in reduced macrophage roaming and directional chemotaxis. These events were caused by decreased ruffling at the leading edge and increased F-actin turnover [55].

The present study showed that highly virulent L. infantum promastigotes, entered macrophages through CR3-mediated, actin-rich, ruffle-like structures. Interestingly, there was no difference in the rate of initial phagocytosis of L. infantum metacyclic promastigotes by WT or CD11b KO macrophages from a BALB/c genetic background, similar to what has been demonstrated for L. major and L. donovani [56]. Likewise, there was no difference in the uptake of L. infantum metacyclics by WT or CD11b KO macrophages from a C57BL/6 background. This suggests that CR3-mediated, ruffle-like structures are not a requisite for macrophage entry and that Leishmania uptake can occur through redundant mechanisms. While the absence of CR3 did not abrogate promastigote uptake, PV fusion with LAMP-1 and Cathepsin D was faster in CD11b KO than WT BMMs.

In addition, our data in CD11b KO macrophages showed decreased metacyclic survival despite similar rates of initial uptake. These results support that multiple receptors and pathways are involved in parasite entry, however, they would lead to various outcomes of infection. Recently, Kumar et al. showed that uptake through the caveolin-mediated endocytosis pathway increases the intracellular survival of L. donovani [40]. Related work by our group showed that L. infantum entry through cholesterol-rich macrophage microdomains led to PV association with caveolae markers, delayed lysosome fusion, and increased parasite survival. Entry through this pathway was only utilized by virulent promastigotes, thus, it is virulence- and stage-dependent. Transient cholesterol depletion before infection abrogated PV association with caveolae markers, accelerated lysosome fusion, and decreased parasite survival [26, 28, 57]. Interestingly, cholesterol is also required for the formation of membrane ruffles. In particular, actin reorganization and translocation of Rac1 to the cell membrane are dependent on membrane cholesterol [58]. Together, our previous work and this study showed that virulent L. infantum promastigotes facilitate their intracellular survival by entering macrophages through cholesterol-rich microdomains and CR3-mediated actin aggregates. Further experiments would be necessary to determine if the observed actin-rich aggregates lead to the formation of membrane ruffles.

Membrane ruffles are precursors for macropinosome formation and cellular uptake through macropinocytosis. Because membrane ruffling, cell signaling, and cell metabolism are interconnected, macropinocytosis impacts macrophage activation [20, 21]. In particular, higher levels of macropinocytosis have been reported in anti-inflammatory (M2) macrophages, which are characterized by lower antimicrobial responses [20, 59]. Clearance of intracellular Leishmania requires an antimicrobial response characteristic of pro-inflammatory (M1) macrophages [47, 59]. However, L. infantum infection drives macrophage gene expression away from pro-inflammatory responses and toward an anti-inflammatory profile resembling M2 macrophages [47]. Hence, the potential links between membrane ruffling, macropinocytosis, and the activation phenotype of Leishmania-infected macrophages are intriguing avenues of research that merit additional studies.

Membrane ruffling could also impact Leishmania infection through the molecular components of the resulting macropinosome. In particular, the lipid and protein makeup of macropinosomes differ qualitatively and quantitatively from those of the endosomes formed upon receptor mediated phagocytosis, which is a major pathway of Leishmania uptake [4, 19, 21]. Several studies have shown that the lipid and protein composition of the PV membrane affects its maturation and fusion with lysosomes [60, 61]. Thus, pathogen uptake through CR3-mediated membrane ruffling and the potential formation of pathogen-containing macropinosomes could have major implications in intracellular trafficking and the generation of macrophage microbicidal responses, or lack thereof. Dissection of the signaling cascades initiated by diverse entry pathways could provide targets for the design of therapies that trigger microbicidal responses. As such, the interactions between cholesterol-rich domains, membrane ruffling, and the outcome of Leishmania infection deserves further exploration.

Supplementary Material

1

S fig 1. Actin remodeling occurs during phagocytosis of different Leishmania spp. and in different mouse strains. BMMs derived from C57BL/6 mice were incubated with opsonized L. donovani (A) or L. infantum (B) promastigotes at a 5:1 MOI. The infection was synchronized by centrifugation, and macrophages were incubated for 15 minutes at 37°C, 5% CO₂. Samples were fixed, stained, and examined by confocal microscopy. Red phalloidin: actin, Green: CFSE-stained parasites, Yellow: Actin + parasite co-localization. Scale bar = 10 μm. (C) BMMs derived from either WT or CD11b KO C57B1/L/6 mice were incubated with opsonized L. infantum metacyclic promastigotes (MOI 5:1). Extracellular parasites were removed by rinsing after 30 minutes, and incubated as described for the indicated time points. Coverslips were fixed, stained with Wright-Giemsa and examined by light microscopy. Graphs show the mean ± SE of the percent of infected macrophages (left panel) or the mean ± SE of intracellular parasites per 100 macrophages (right panel) in 3 replicate experiments, each with triplicate conditions. Statistics were done by t-test (*p <0.05).

• Uptake of L. infantum induced ruffle-like structures at the macrophage membrane.

• L. infantum entry through ruffle-like structures is stage- and virulence-dependent.

• Actin-rich, ruffle-like structures were dependent on complement receptor 3 (CR3).

Abbreviations:

BMM

bone marrow derived macrophage

CFSE

carboxy-fluorescein diacetate succinimidyl ester

CR3

complement receptor 3

F-actin

filamentous actin

GFP

green fluorescent protein

HOMEM

hemoflagellate-modified minimal essential medium for parasite culture

LAMP-1

Lysosomal Associated Membrane Protein-1

Li

Leishmania infantum

KO

knockout

MOI

multiplicity of infection

PV

parasitophorous vacuole

RP10

RPMI-based macrophage tissue culture medium

WT

wild type