Malaria parasites target the hepatocyte receptor EphA2 for successful host infection

See allHide authors and affiliations

Science  27 Nov 2015:
Vol. 350, Issue 6264, pp. 1089-1092
DOI: 10.1126/science.aad3318

How malaria parasites infect the liver

Early in infection, malaria parasites establish themselves within hepatocytes in the liver. Inside these cells, the parasites occupy a so-called parasitophorous vacuole. Kaushansky et al. show that malaria parasites prefer to create vacuoles within hepatocytes that express the EphA2 receptor. Hepatocytes with low levels of this receptor were less conducive to malaria infection.

Science, this issue p. 1089


The invasion of a suitable host hepatocyte by mosquito-transmitted Plasmodium sporozoites is an essential early step in successful malaria parasite infection. Yet precisely how sporozoites target their host cell and facilitate productive infection remains largely unknown. We found that the hepatocyte EphA2 receptor was critical for establishing a permissive intracellular replication compartment, the parasitophorous vacuole. Sporozoites productively infected hepatocytes with high EphA2 expression, and the deletion of EphA2 protected mice from liver infection. Lack of host EphA2 phenocopied the lack of the sporozoite proteins P52 and P36. Our data suggest that P36 engages EphA2, which is likely to be a key step in establishing the permissive replication compartment.

Malaria infections place a tremendous burden on global health (1). Their causative agents, Plasmodium parasites, are transmitted to mammals as sporozoites by the bite of Anopheles mosquitoes. After entry into a capillary, sporozoites are carried to the liver, where they pass through multiple cells before recognizing and invading hepatocytes. During invasion, the sporozoite forms a protective parasitophorous vacuole made of hepatocyte plasma membrane, which ensconces the parasite, establishes the intrahepatocytic replication niche, and supports successful infection. Highly sulfated proteoglycans are known to provide a signal to sporozoites to invade the liver parenchyma (2, 3), and hepatocyte CD81 and scavenger receptor B1 are important for hepatocyte infection (46). Beyond this, the molecular mechanisms underlying infection remain poorly understood.

Hepatocytes exhibit differential susceptibility to infection. Sporozoites preferentially enter polyploid hepatocytes (7). Also, BALB/cByJ mice are more susceptible than BALB/cJ mice to Plasmodium yoelii sporozoite infection (8). To identify potential host receptors that might contribute to differential susceptibility, we used an antibody array to assess the levels of 28 activated receptors in the livers of BALB/cJ and BALB/cByJ mice. Nine receptors, including EphA2, were present at significantly (P < 0.01) and substantially elevated levels in highly susceptible BALB/cByJ mice (table S1). Polyploid hepatocytes also expressed higher levels of EphA2 (fig. S1).

In metazoans, Eph receptors and their cognate Ephrin ligands mediate cell-cell contact (9), making EphA2 a candidate to mediate the hepatocyte-sporozoite interaction. Furthermore, an Ephrin-like fold is present in the parasite’s 6-Cys protein family (10). Although Hepa1-6 cells (a murine hepatocyte line) expressed EphA2 consistently across passages, variation within a culture was substantial (fig. S2). We therefore postulated that if EphA2 mediates sporozoite invasion, susceptibilities might vary within a culture of Hepa1-6 cells.

We infected Hepa1-6 cells with P. yoelii sporozoites; after 24 hours, we assessed parasites in hepatocytes that expressed high levels of EphA2 (Fig. 1A). We also observed this by flow cytometry 1.5 hours after infection (Fig. 1B and fig. S3A), and parasite-infected cells exhibited significantly increased levels of both total (Fig. 1C) and surface (fig. S3, B to D) EphA2. Similarly, the frequency of infection in cells in the top 50% of EphA2 expression (EphA2high) was elevated compared with infection frequency in cells in the bottom 50% (EphA2low) (Fig. 1D). When we included only the top 40, 30, 20, or 10% of EphA2-expressing cells in the EphA2high gate, the preference was even more pronounced (fig. S3E).

Fig. 1 Plasmodium sporozoites invade hepatocytes with high EphA2 expression.

(A) Hepa1-6 cells were infected with P. yoelii sporozoites and visualized by immunofluorescence 24 hours after infection. The scale bar is 5 μm. (B to D) Hepa1-6 cells were infected with 105 P. yoelii sporozoites. (B) shows the distribution of EphA2 1.5 hours after infection (mEphA2, mouse EphA2). In (C), EphA2 levels are compared between parasite-infected and uninfected cells. (D) shows parasite-infection rates in EphA2high and EphA2low cells (Py, P. yoelii). The numbers in the bars are the percentages of infected cells within each subset. (E to G) BALB/c mice were infected with 106 P. yoelii sporozoites by intravenous injection. Hepatocytes were analyzed as in (B) to (D). (H to J) HC-04 cells were infected with 105 P. falciparum sporozoites (hEphA2, human EphA2). Analyses were performed as in (B) to (D). (K) Hepa1-6 cells were incubated with EphA2-blocking antibody (D4A2) or immunoglobulin G (IgG) as a control 30 min before infection with 105 P. yoelii sporozoites. The infection rate was normalized to the infection rate in the presence of IgG. Each figure represents at least three independent experiments. The bar graphs show means with standard deviations.

We next challenged BALB/c mice with 106 P. yoelii sporozoites and isolated hepatocytes after 3 hours. We again observed a strong parasite preference for EphA2high hepatocytes (Fig. 1, E to G). Finally, we tested whether the preference for infection of EphA2high hepatocytes is also present in the human parasites by infecting HC-04 hepatocytes with P. falciparum. We observed elevated levels of EphA2 in infected cells and a higher proportion of sporozoite-containing cells in the EphA2high population (Fig. 1, H to J).

EphA2 has an extracellular ligand-binding region and an intracellular kinase domain, which mediates downstream signaling. To assess whether interaction with the extracellular portion of EphA2 is critical for Plasmodium infection, we infected hepatocytes in the presence of an antibody that binds extracellular EphA2. The presence of the antibody reduced sporozoite infection in a dose-dependent manner (Fig. 1K). In contrast, inhibiting the kinase domain of EphA2 did not inhibit infection (fig. S4). Thus, the extracellular portion of EphA2 facilitates Plasmodium invasion of hepatocytes.

To test whether EphA2 levels are important forliver-stage parasite survival and development, we measured infection rates in EphA2high and EphA2low cells over the course of 48 hours, normalizing each infection rate to the rate at 1.5 hours after infection. Whereas the number of EphA2high infected cells was maintained throughout the course of infection, the number of EphA2low infected cells decreased over time (Fig. 2A). This difference could not be accounted for by division rates, because we observed lower levels of host cell division among EphA2low cells. Thus, our results may in fact underestimate the impact of EphA2 on infected cell survival (fig. S5). When we infected EphA2(−/−) and wild-type mice with 105 P. yoelii sporozoites, we observed a large decrease in liver-stage burden after 42 hours in EphA2(−/−) mice (Fig. 2B). EphA2(−/−) mice also exhibited a delay in the onset of blood-stage infection by 1 to 3 days (Fig. 2C). Thus, without EphA2, the host is far less susceptible to productive parasite liver infection.

Fig. 2 EphA2 affects PVM formation.

(A) Time sequence showing the maintenance of P.yoelii infection in Hepa1-6 cells. Maintenance of infection was defined as the infection rate at a given point divided by the infection rate at 1.5 hours after infection. Error bars indicate the standard deviation of biological replicates. (B) EphA2(−/−) or age-matched wild-type (WT) mice were infected with 105 P. yoelii sporozoites. Infection was assessed by quantitative polymerase chain reaction 42 hours after infection (Py18s, P. yoelii 18S ribosomal RNA; mGAPDH, mouse glyceraldehyde-3-phosphate dehydrogenase). Error bars show SEM. (C) EphA2(−/−) or strain-matched WT mice were infected with 102 P. yoelii sporozoites. Patency was monitored daily by thin smear. The horizontal bar indicates the median. (D) A PyUIS-Myc parasite in a Hepa1-6 cell 24 hours after infection. The scale bar is 5 μm. (E and F) PyUIS4-Myc parasites were used to infect Hepa1-6 cells, and the cells were analyzed for the presence of the PVM (UIS4pos) after 24 hours (MFI, median florescence intensity). (G) PyUIS4-Myc infected Hepa1-6 cells were assessed for permeability after 48 hours. Each figure represents at least three independent experiments. (E) to (G) show means with standard deviations.

The parasitophorous vacuole membrane (PVM) is critical for liver-stage development. One liver-stage PVM-resident protein, UIS4, is highly expressed after invasion when it is exported to the PVM (11), making it a useful marker. We constructed a P. yoelii parasite line, PyUIS4-Myc, which expressed a UIS4-Myc fusion protein driven by the endogenous UIS4 promoter (Fig. 2D). This allowed us to monitor PVM prevalence (UIS4pos) in infected cells by flow cytometry. Most of the UIS4pos infected host cells were in the EphA2high category (Fig. 2E). Similarly, the level of EphA2 expression was higher in UIS4pos infected cells than in UIS4neg infected cells (Fig. 2F). Thus, sporozoites not only preferentially entered EphA2high cells, but invasion accompanied by PVM formation was far more effective in these cells. UIS4neg infected hepatocytes suffered a higher frequency of cell death (Fig. 2G).

Two members of the 6-Cys family of parasite proteins (12, 13), P52 and P36, are expressed in sporozoites, are important for the invasion of hepatocytes (1416), and are critical for PVM formation (14). In mouse livers, parasites without P52 or P36 were almost entirely eliminated within 3 hours after infection (fig. S6). We tested whether the lack of P52 and P36 phenocopies the lack of host EphA2 and found that p52/p36 P. yoelii sporozoites exhibited a reduced preference for EphA2high cells (Fig. 3A). The related 6-Cys protein P12 shows structural similarity to the mammalian ligand for EphA2, EphrinA1 (10).

Fig. 3 P36 interacts with EphA2.

(A) 2 × 105 WT or p52/p36 P. yoelii parasites were used to infect Hepa1-6 cells. Levels of EphA2 were monitored in infected and uninfected cells. (B) P. yoelii sporozoites were used to infect Hepa1-6 cells 30 min after treatment with IgG or EphA2-blocking antibody (αEphA2). Infection levels were normalized to the infection rate in the presence of IgG. (C and D) Hepa1-6 cells were incubated with recombinant P52 or P36, alone or in combination with 10 min of EphrinA1 treatment. Immunoblots show levels of pEphA2 (pY772). (E) 2 × 105 WT or p52/p36/sap1 P. falciparum sporozoites were used to infect 6 × 105 HC-04 cells. EphA2 levels were measured in infected and uninfected cells. Each figure represents at least three independent experiments. The bar graphs show means with standard deviations.

We showed that an interaction in the extracellular region of EphA2 was required for sporozoite entry using an EphA2-blocking antibody (Fig. 1K). Therefore, we next asked whether the presence of P36 and P52 was required for the antibody to block sporozoite entry. The EphA2 antibody blocked infection for wild-type P. yoelii sporozoites, but p52/p36 sporozoite entry was not affected (Fig. 3B). These data suggest that P36 or P52 engages EphA2 at the point of host cell invasion. We next tested whether P52 or P36 could directly impede the interaction between EphrinA1 and EphA2 on the hepatocyte surface, which results in EphA2 activation. When we added EphrinA1 in the presence of P36 to Hepa1-6 cells, P36 blocked the activation of EphA2 (Fig. 3, C and D). P52, however, did not block EphrinA1-mediated activation of EphA2 (Fig. 3, C and D). To determine whether the interaction between EphA2 and P36 also occurs in human parasites, we assessed levels of EphA2 in P. falciparum wild-type or p52/p36/sap1 parasite-infected HC-04 cells. The P52-P36–deficient P. falciparum sporozoites exhibited partially reduced selectivity for EphA2high HC-04 cells compared with P. falciparum wild-type sporozoites (Fig. 3E). Thus, P36 engages EphA2 but does not trigger its activation in rodent and human parasites.

We have shown that both host EphA2 and parasite 6-Cys proteins have a role in sporozoite invasion of hepatocytes and the establishment of the growth-permissive intracellular niche. Without either component, the parasite can still enter hepatocytes, but it does so without a PVM, which can result in death of the infected hepatocyte. The convergence of infection-permissive phenotypes is best explained by an interaction between parasite P36 and hepatocyte EphA2 when the PVM is formed. This role for EphA2 in hepatocyte infection does not preclude the possibility that additional hepatocyte receptors may be critical for infection. Interventional strategies aimed at either EphA2 or sporozoite 6-Cys proteins might block parasite infection before the onset of clinical malaria.

Supplementary Materials

Materials and Methods

Figs. S1 to S7

Table S1


  1. ACKNOWLEDGMENTS: We are grateful to W. Betz for mosquito and sporozoite production. We thank the vivarium staff of the Center for Infectious Disease Research for their work with mice. This work was funded by a NIH Research Project Grant (R01) to S.H.I.K. and A.K. (grant no. 1R01GM101183-01A1). A.K. is also a recipient of a Transition to Independence Award (award no. 1K99AI111785-01A1), which partially funded this work. A.K., N.A., A.N.D., V.V., N.D., H.K., and L.S.A. performed the experiments. A.K., D.N.S., and S.H.I.K. supervised the research. A.K. and S.H.I.K. wrote the paper. A provisional U.S. patent application (application no. 62/110018) has been filed by the Center for Infectious Disease Research, covering interventions that exploit the described host-parasite interaction. The data are included in the main manuscript and in the supplementary materials.
View Abstract

Navigate This Article