Plasmodium gametocytes display homing and vascular transmigration in the host bone marrow

Abstract

Transmission of Plasmodium parasites to the mosquito requires the formation and development of gametocytes. Studies in infected humans have shown that only the most mature forms of Plasmodium falciparum gametocytes are present in circulation, whereas immature forms accumulate in the hematopoietic environment of the bone marrow. We used the rodent model Plasmodium berghei to study gametocyte behavior through time under physiological conditions. Intravital microscopy demonstrated preferential homing of early gametocyte forms across the intact vascular barrier of the bone marrow and the spleen early during infection and subsequent development in the extravascular environment. During the acute phase of infection, we observed vascular leakage resulting in further parasite accumulation in this environment. Mature gametocytes showed high deformability and were found entering and exiting the intact vascular barrier. We suggest that extravascular gametocyte localization and mobility are essential for gametocytogenesis and transmission of Plasmodium to the mosquito.

Fig. 1. Parasite distribution in rodent tissues.

(A to C) Bioluminescence and intravital imaging and ex vivo quantification in ice infected with transgenic P. berghei coexpressing mCherry and luciferase reporters (mCherry_Hsp70-FLuc_ef1α) (36). (A) Representative images show bioluminescence signal in the BM (femur and pelvic bones marked by arrow), in mice at (i) 65 hours following sporozoite infection, namely, after completion of the pre-erythrocytic stages, immediately following egress of merozoites from the liver to begin the first blood stage cycle; (ii) at 22 hours following infection with nonsequestering (N.S.) schizonts (that is, thereby, the only population detected at this time point are schizonts of the immediate next cycle) (9); and at day 7 (d7) postinfection (pi) with mixed blood stage parasites without (iii) or with (iv) pretreatment with PH-S (PH-S will select for gametocytes while killing the asexual populations of parasites). Note that the signal from the BM is potentially obscured by that of the adipose tissue in (iii), but not in (ii) because of the lack of sequestration phenotype. hpi, hours post-invasion. (B) Bioluminescence-based quantification of individual extracted organs from 15 mice (5 in triplicate experiments) infected with either synchronous schizonts (black bars) or after PH-S treatment (red bars). Bars indicate luminescence (γ/s), normalized to and expressed as a ratio of schizont signal in the brain. Arrows show that the largest relative gametocyte signal occurs in spleen and BM. AT (ab), abdominal adipose tissue. (C) Representative intravital and/or ex vivo images of mCherry_Hsp70-FLuc_ef1α P. berghei gametocytes (red) in various tissues of transgenic mice treated with PH-S (green) expressing either UBC-GFP (top panels) or Flk1-GFP (bottom panels). Lung, abdominal adipose tissue, heart, and brain show gametocytes in vasculature, whereas liver, spleen, and BM also show extravascular gametocytes. Scale bars, 20 μm. (D) CD31- and Flk1-based immunofluorescence quantification of vascular versus extravascular mCherry_Hsp70-FLuc_ef1α P. berghei distribution in mice after asexual blood stage infection. Highest extravascular parasite levels are observed in the spleen, BM, and liver (white stacked bars). Error bars represent SDs of quantifications in 25 fields of view across 18 separate mice. (E) Schematic showing specific fluorescence of gametocyte producer (Con and EG) or nonproducer (GNP) P. berghei lines, under constitutive (Con-CFP and GNP-CFP) or early gametocyte–specific promoters (EG-CFP). (F) Distribution of gametocyte producer and nonproducer lines within the parenchyma (P), sinusoids (S), and intravascular (IV) spaces of BM, spleen, and liver of mice at day 4 pi. Con-CFP parasites are equally distributed across BM and spleen compartments. In contrast, EG-CFP parasites are enriched in parenchyma and sinusoids, whereas GNP parasites are enriched in the vasculature. Error bars represent SDs of quantification in 25 fields of view across 30 separate mice. (G) Extravascular parasites in the BM localize to sinusoids and the parenchyma, as defined by ICAM-1 labeling (yellow). The inset in the ICAM-1 image shows gametocytes in close distance to the sinusoidal lining (left panel). Scale bars, 20 μm. For all experiments, three to five mice were used per triplicate experiment, and at least 25 fields of view were quantified per mouse.

Fig. 2. Parasite homing and extravascular BM localization during mouse infection.

(A) Synchronous infections of mice with mCherry_Hsp70-FLuc_ef1α merozoites, rings, trophozoites, schizonts, or gametocytes. Data show that merozoites, rings, and gametocytes are significantly enriched in BM and spleen (red arrows, ^***P < 0.001) compared to mean across all organs (dashed line). Data are expressed relative to parasite numbers quantified in the brain (normalized to 100) and correspond to intravascular and extravascular compartments in all organs. (B) Representative diagram of hematopoietic development. Hematopoietic stem cells (HSCs) undergo a series of developmental steps until formation of the final nucleated stage termed orthochromatic erythroblast (OCE), which upon enucleation gives rise to reticulocytes (Retic). Reticulocytes intravasate into circulation and rapidly mature to terminal normocytes. Extravascular RBC precursor cells are marked by CD44, whereas CD71 is lost shortly after intravastion. (C) Presence of stage-specific reporter parasites in CD44⁺ RBCs following injection of merozoites. At 24 hpi, EG-CFP parasites, but not GNP-CFP and Con-CFP parasites, are largely confined to parenchyma and sinusoids where most are present in CD44⁺ RBCs (left panel), reflecting distribution of CD44⁺ RBCs (mid left panel). No parasites are detected in CD44⁺ cells in peripheral blood (mid right panel). Representative images of CD44⁺ cells in a BM section (magenta) are shown, upon injection of fluorescein isothiocyanate (FITC)–dextran (cyan) to mark vasculature (top panel). Bottom image: Close-up of an infected CD44⁺ RBC in BM. (D) Presence of uninfected UBC-GFP RBCs in the BM extravascular niche. UBC-GFP RBCs were obtained from transgenic donor mice and injected into naïve, uninfected recipient mice and quantified in BM. A small proportion of RBCs accumulates in BM extravasculature but not in the brain extravascular niche (left panel). Representative image is shown (left) and a control with mCherry_Hsp70-FLuc_ef1α P. berghei parasites (right). (E) Inhibition of parasite extravasation using receptor antibodies. Mice were pretreated for 24 to 48 hours with P-selectin antibodies before injection of mCherry_Hsp70-FLuc_ef1α merozoites and compared to untreated and nonviable (4% paraformaldehyde-fixed) control parasites (left panel). Parasite abundance in BM but not in brain or peripheral circulation is reduced upon P-selectin pretreatment compared to controls. Injection of merozoites from stage-specific reporter parasites upon P-selectin pretreatment reveals greatest reduction in BM parenchyma with EG-CFP and Con-CFP parasites, whereas intravascular localization is increased (right panel). Error bars represent SEs from five separate experiments in three mice per condition. (F) Quantification of peripheral parasitemia upon treatment with receptor antibodies. UBC-GFP mice were pretreated for 24 to 48 hours with P-selectin before injection of mCherry_Hsp70-FLuc_ef1α merozoites (right panel) and compared to untreated control parasites (left panel). Peripheral parasitemia and tissue burden were quantified before and after cardiac perfusion to remove circulating parasites—leaving only those in the extravascular spaces. Extravascular parasite load decreases significantly in BM and spleen compared to control animals, whereas peripheral parasitemia (obtained before perfusion) increases. P values are shown as **P < 0.05 and ***P < 0.005. All error bars represent SDs of the mean.

Fig. 3. Asexual and gametocyte stages in the BM parenchyma of human malaria cases.

(A) NanoString analysis of BM, brain, and heart tissue from patients who died of clinical cerebral malaria. Asexual P. falciparum parasite stages are predominantly sequestered in the brain and heart (left panel). All gametocyte-specific transcripts are significantly enriched in BM compared to the other tissues (right panel). Data were normalized on the basis of background subtraction and expression of housekeeping genes. **P < 0.01 and ***P < 0.001. (B and C) Quantification of immunohistochemistry (IHC) studies on human autopsy cases from BM tissue (n = 22). (B) Quantification of IHC with KAHRP antibodies to mark asexual P. falciparum parasites and CD31 to mark vasculature (HPF, high-powered fields). Histogram demonstrated that presence of parasites in BM is highest in the parenchyma compared to sinusoids and vessels. (C) Patients succumbing in absence of cerebral malaria (CM) had higher numbers of gametocytes (Pfs16) in parenchyma than asexual parasites (KAHRP), whereas these numbers were similar in CM patients. Representative IHC images of CD31 (brown) and KAHRP (red) from a human autopsy BM sample are shown. Arrows indicate intravascular asexual parasites, and arrow heads indicate extravascular asexual parasites. *P < 0.05. Scale bars, 20 μm.

Fig. 4. Gametocyte transmigration dynamics and vascular leakage.

(A) Time series of mCherry_Hsp70-FLuc_ef1α gametocyte crossing the vascular endothelium in an Flk1-GFP (top panel) or UBC-GFP (bottom panel) transgenic mouse. [Time lapse, 5 frames/s; scale bar, 10 μm.] P, parasite; IV, intravascular; EV, extravascular; C1,2, contact site; T1,2, transmigration. In the top panel series, a gametocyte is exiting the BM vasculature (marked as dotted lines), whereas the bottom panel series are showing a gametocyte entering the BM vasculature. (B) Quantification of observed transmigration events across the BM vascular barrier. Upon injection of synchronized parasite stages, only merozoites and gametocytes show significant levels of transmigration events. (C) Sphericity and deformability of P. berghei parasites. Left: Representative in vivo images across stages of mCherry_Hsp70-FLuc_ef1α parasites within UBC-GFP RBCs in static and flow conditions (scale bar, 5 μm). Right: Deformability indices across stages defined as the ratio of maximum cellular diameter before (static) and during (contracted) transmigration (n = 100; bars represent mean, and error bars represent SD. ***P < 0.001). (D and E) Transmigration and vascular leakage. Mice are injected with synchronous mCherry_Hsp70-FLuc_ef1α parasite stages followed by FITC-dextran inoculation, and images are taken at a rate of 5 fps. Representative images of FITC-dextran fluorescence intensity are shown (D). Quantification of FITC-dextran leakage (representative of peripheral blood including iRBCs and uRBCs), demonstrating a unique transient leakage pattern upon injection of mature gametocytes (E). Note that merozoites/rings (M/R) values are normalized by schizont values because of the observed low level of sustained leakage pattern likely caused by co-injected hemozoin. MFI, mean fluorescence intensity. (F) Gametocyte treatment with Cytochalasin D (C.D.) abrogates extravascular accumulation. mCherry_Hsp70-FLuc_ef1α gametocytes are purified ex vivo and incubated with C.D. (or vehicle control) before washout and intravenous injection into naïve mice. Relative abundance and distance from vessels in the extravascular BM compartment are significantly reduced. IVS, intravascular space; EVS, extravascular space. (G) Treatment of infected mice with sildenafil citrate results in gametocyte accumulation in BM and spleen. Mice treated with phenylhydrazine were infected with mCherry_Hsp70-FLuc_ef1α parasites and treated with sulfadiazine–sildenafil citrate for 3 days (or sulfadiazine only, control). Treatment leads to significantly increased accumulation of gametocytes in the BM and spleen of mice compared to control. In all panels, error bars represent SDs of the mean. *** P < 0.001 and **P < 0.01. N.A., not applicable.

Fig. 5. Gametocytes display mobility behavior in tissues of the reticuloendothelial system.

(A) Gametocytes are mobile in the BM and spleen. Early (EG) and mature (MG) gametocyte behavior is quantified across intravascular and extravascular compartments. Most mobile gametocytes are mature; most of the mature gametocytes are found passively circulating in the vasculature, whereas a subset is present in all compartments. Early gametocytes are confined to the extravascular niche and static. (B) Mature gametocytes move at speeds between 0.01 and 1 μm/s in the parenchyma of BM, similar to leukocytes. A subset of gametocytes and leukocytes (black box) show contact-dependent movements in the sinusoids and small vessels of the same organs at a speed between 4 and 10 μm/s. In contrast, uninfected erythrocytes and asexual parasite stages are static in the parenchyma and passively move with the flow in sinusoids and small vessels at a much higher speed. (C and D) Time-lapse intravital images of gametocyte behavior in the BM or spleen. P. berghei mCherry gametocytes are imaged in Flk1-GFP mice. Top panel: Mature gametocyte circulating within the vasculature (arrow). Second panel: Mobile (arrow) and static (arrow head) gametocytes in the BM extravascular space of Flk1-GFP mice. Third panel: Motile behavior of leukocytes in uninfected UBC-GFP mice. Bottom panel: Mobile behavior of gametocytes in infected UBC-GFP mice. (D) Mobility pattern of one gametocyte dynamically moving in the BM. Mature gametocytes show multiple mobility patterns in the spleen and BM: multiple contact points in parenchyma (represented by 0 in the y axis) followed by forward motion (left panel) and rapid translocation upon contact with sinusoidal endothelium. Compared to ring stages, gametocytes contact the parenchyma two to four times more frequently (middle panel). At every contact point leading to a stop phase, gametocytes display a lag in velocity of up to 35 s (right panel). (Graphs show median and interquartile ranges of lag time following contact.) Error bars show SEM. ***P < 0.001. (E) Mature gametocytes display various behaviors in relation to crossing the vascular endothelium. Most of the parasites recorded were in the vasculature and not crossing, whereas a subset is either crossing once or twice. (F) Hypothesized model of parasite dynamics in the BM. (I) Distribution of Plasmodium gametocyte stages in the BM. Early gametocyte stages accumulate in the parenchyma and sinusoids of BM (and spleen), whereas mature gametocytes are found both intra- and extravascularly. (II) Vascular leakage and its relation to asexual parasite accumulation in BM. Disease progression coincides with vascular leakage in sinusoids of the reticuloendothelial system but not in the vasculature of other tissues. This leakage results in increased accumulation and replication of asexual parasite stages in the BM parenchyma. (III) Mobility and vascular migration in the BM. A subset of merozoites, including most of the sexual merozoites, invades BM-resident RBC precursor cells in sinusoids and parenchyma. The latter requires merozoite extravasation across the sinusoidal barrier via specific receptor-ligand interactions. Early gametocytes are located on the sinusoidal lining and attached to erythroblastic islands in the parenchyma, where they develop until maturity (represented by red arrow). Mature gametocytes become mobile, presumably coinciding with a switch in deformability, and enter circulation from their extravascular niche.