Plasmodium vivax: restricted tropism and rapid remodeling of CD71-positive reticulocytes

Abstract

Plasmodium vivax merozoites only invade reticulocytes, a minor though heterogeneous population of red blood cell precursors that can be graded by levels of transferrin receptor (CD71) expression. The development of a protocol that allows sorting reticulocytes into defined developmental stages and a robust ex vivo P vivax invasion assay has made it possible for the first time to investigate the fine-scale invasion preference of P vivax merozoites. Surprisingly, it was the immature reticulocytes (CD71(+)) that are generally restricted to the bone marrow that were preferentially invaded, whereas older reticulocytes (CD71(-)), principally found in the peripheral blood, were rarely invaded. Invasion assays based on the CD71(+) reticulocyte fraction revealed substantial postinvasion modification. Thus, 3 to 6 hours after invasion, the initially biomechanically rigid CD71(+) reticulocytes convert into a highly deformable CD71(-) infected red blood cell devoid of host reticular matter, a process that normally spans 24 hours for uninfected reticulocytes. Concurrent with these changes, clathrin pits disappear by 3 hours postinvasion, replaced by distinctive caveolae nanostructures. These 2 hitherto unsuspected features of P vivax invasion, a narrow preference for immature reticulocytes and a rapid remodeling of the host cell, provide important insights pertinent to the pathobiology of the P vivax infection.

Figure 1

Reticulocyte tropism of P vivax is restricted to CD71⁺ reticulocytes. (A) Flow cytometry dot plot showing the CD71 and TO staining used to define the different erythrocyte subsets: TO⁺/CD71⁺ for the immature reticulocytes (green box), TO^low/CD71⁻ for the mature reticulocytes (red box), and TO⁻/CD71⁻ for the normocytes (black box). The insets on the right show the NMB-stained reticulocytes sampled from these gates by flow cytometry sorting. (B) Histogram showing the fold increase reinvasion, normalized to that measured for normocytes (TO⁻/CD71⁻) (mean 2.17 ± 1.23 standard deviation [SD] vs 9.47 ± 4.81 SD), for the 3 reticulocyte subsets harvested by cell sorting. Bar represents 8 μm. (C) Flow cytometry dot plot showing the reticulocyte sampling based on the CD71 expression (CD71⁻, CD71^low, CD71^medium, CD71^high). The top insets show each population stained by NMB. Note the red arrows in the right insets showing the newly reinvaded erythrocytes stained with Hoechst. (D) Histogram showing the fold increase reinvasion, normalized to that measured for the CD71⁻ population and for the 3 younger reticulocyte populations (mean CD71^low: 5.09 ± 1.75 SD; CD71^medium: 6.53 ± 2.41 SD; CD71^high: 7.94 ± 3.14 SD; n = 3). Note the red arrows in the right insets showing the newly reinvaded erythrocytes stained with Hoechst. Bar represents 8 μm. (E) Representative flow cytometry dot plot showing Hoechst and ethidium levels observed at the start (0 hours) and after 24 hours of the reinvasion assay with CD71⁻ erythrocytes, and CD71⁺ reticulocytes treated or not with anti-Duffy antibodies. For the 0-hour data, the black thick-lined squares show the percentage of schizonts; the black thin-lined squares show the percentage of rings before invasion. For the 24 hours of data, the black thin-lined squares at 24 hours show the percentage of unburst schizonts; the orange thick-lined squares show the percentage of rings after invasion. (F) The percentage of rings after reinvasion (n = 5). The value observed with CD71⁺ reticulocytes incubated with anti-Duffy antibodies was taken as the threshold of reinvasion (0.74%). hr, hour. *P < .05; **P < .01.

Figure 2

P vivax infection induces a rapid remodeling of the cell surface and the cytoplasmic compartments of immature reticulocytes. (A) Flow cytometry of P vivax field isolates stained with Hoechst and for CD71 using a gating strategy to identity CD71⁺ (pink) and CD71⁻ (purple) P vivax–infected cells. (B) Proportion of CD71⁻ (pink bars) and CD71⁺ (purple bars) P vivax (Pv)-infected cells from 13 clinical samples. The mean percentage of uninfected CD71⁺ cells is 1.29% ± 1.31% SD. (C) Gating strategy for CD71⁺ and CD71⁻ P vivax rings flow cytometry sorting; the lower Hoechst signal is associated with ring-stage parasites. The insert on the right shows the SEM scans of the 2 populations (scale bars represent 1 μm) and the blue box indicates the area shown at higher magnification (scale bars represent 100 nm). (D) The exclusion reticular material (subvitally stained with Giemsa) in uninfected and P vivax–infected CD71⁺ reticulocytes over a period of 40 hours. Scale bar represents 1 μm. (E) Rapid disappearance of CD71 in an ex vivo P vivax–invaded reticulocyte. These freshly invaded CD71⁺ reticulocytes with P vivax rings were counterstained with Alexa 546 and Hoechst to stain parasite DNA and were observed under a fluorescent microscope during the first 3 hours postinvasion. Scale bar represents 1 μm. (F) Frequency of CD71⁺ (pink) and CD71⁻ (purple) uninfected or P vivax–infected reticulocytes at invasion 0 and 3 hours after reinvasion.

Figure 3

Comparative phenotypic characterization of clathrin pits and caveolae in uninfected reticulocytes and P vivax IRBCs. (A) Representative AFM scans that illustrate the presence and density of clathrin pits on uninfected reticulocytes and of caveolae on P vivax IRBCs; the bottom inserts show the shape of these 2 nanostructures as revealed by transmission electron microscopy. (B) The range of the diameters of the 2 nanostructures observed on uninfected reticulocytes and on P vivax IRBCs and (C) changes in the distribution of caveolae and clathrin pits with the ex vivo maturation of the P vivax IRBCs (both observations derived from the same 13 independent isolates [staged ex vivo maturation], with total of 485 IRBCs examined). The P vivax isolates we received for this staged ex vivo maturation were >6 hours old (n = 10), thus we needed to conduct 3 invasion assays (n = 3) to get AFM data on the missing 0 to 6 hour postinvasion window (tiny ring stage). (D) Representative AFM scans of P vivax IRBCs at various stages of development and the directly corresponding Giemsa-stained picture. This set of scans was taken from staged ex vivo invasion and maturation assay. Bar represents 10 μm.

Figure 4

Evolution of cell deformability during the maturation of P vivax. (A) Representative micropipette aspiration images of CD71⁺ and CD71⁻ reticulocytes and of P vivax IRBCs at the different maturation stages (based on the size of the parasitophorous vacuole and arrangement of hemozoin). (B) Key examples of single-cell measurements of cell length aspirated with different pressure intensities. Also provided is a diagram (upper left corner) and the equation (lower left corner) to show how the shear flow modulus for each cell was calculated. ΔL, aspirated length; ΔP, aspirating pressure; D, micropipette inner diameter; dL/dP, slope of the linear region of the aspiration graph as shown in this figure. (C) The median shear flow modulus (± interquartile range) of each population. Note that the uninfected CD71⁺ reticulocytes and tiny ring forms have the highest shear modulus (ie, the most rigid cells), which was significantly higher than any of the other normocytes and other infected red cells. n.s., not significant.

Figure 5

Hypothetical interactions between P vivax (P.v.) and bone marrow reticulocytes. The reticulocytes (blue) display a gradient of surface CD71 (from high CD71 expression on the erythroblasts [0] to the CD71⁻ stage IV reticulocytes). Stage 0 erythroblasts and stages I/II reticulocytes are present in a steady state in the primary hematopoietic sinus of the red bone marrow. Stage III reticulocytes egress by diapedesis to the sinusoidal capillary lumen and then join the peripheral blood. It is presently considered that this is where the reticulocytes are infected by P vivax, after which rapid host cell remodeling is induced and is associated with microvesicle formation and an accelerated loss of various surface markers, including CD71 and reticular matter. Our in vitro observations indicated that the stage I/II reticulocytes (and possibly the erythroblasts) are the preferred targets for the merozoite. This suggests that P vivax merozoites, or indeed P vivax IRBCs, might be able to enter the red bone marrow compartment and invade their preferred host cells there.