|Home | About | Journals | Submit | Contact Us | Français|
The Plasmodium liver stage is an attractive target for the development of anti-malarial drugs and vaccines, as it provides an opportunity to interrupt the life cycle of the parasite at a critical early stage. However, targeting the liver stage has been difficult. Undoubtedly, a major barrier has been the lack of robust, reliable and reproducible in vitro liver stage cultures. Here, we establish the liver stages for both Plasmodium falciparum and Plasmodium vivax in a microscale human liver platform composed of cryopreserved, micropatterned human primary hepatocytes surrounded by supportive stromal cells. Using this system, we have successfully recapitulated the full liver stage of P. falciparum including the release of infected merozoites and infection of overlaid erythrocytes, and also the establishment of small forms in late liver stages of P. vivax. Finally, we validate the potential of this platform as a tool for medium-throughput anti-malarial drug screening and vaccine development.
Despite major advances in the prevention and treatment of malaria, this disease continues to be a major global health problem in human populations with ~250 million cases and nearly 1 million deaths every year (World Health Organization. Global Malaria Programme, 2010). Malaria is transmitted by Plasmodium sporozoites after they are injected by an infected mosquito. Uninucleate sporozoites travel to and invade liver hepatocytes, where they mature and multiply to form liver stage schizonts. These schizonts eventually release pathogenic merozoites into the blood, where they invade erythrocytes and lead to the major clinical symptoms, signs, and pathology of the disease. Human malaria is primarily caused by four species of Plasmodium parasites. Plasmodium falciparum (P. falciparum) is the most virulent and causes the vast majority of deaths. Plasmodium vivax (P. vivax) is less deadly but highly disabling. Notably, some P. vivax sporozoites develop into dormant hypnozoites, which remain in the liver and serve as a P. vivax reservoir that gives rise to a chronic, relapsing infection and causes significant added clinical and financial burden (Price et al., 2007). Currently, there is a renewed interest and focus on global malaria eradication, and it is now widely recognized that existing tools are insufficient to meet this goal (Alonso, 2011a, 2011b).
In particular, the clinical options that target the liver stages of the parasite life cycle are inadequate. There is only one licensed drug that eliminates hypnozoites only a few drugs that target liver stage parasites, and no licensed malaria vaccines. This problem has been exacerbated by the emergence of drug resistance, and the inability to treat some populations with primaquine, the only currently approved drug with anti-hypnozoite activity (Wells et al., 2010). The Plasmodium liver stage is an attractive therapeutic target for the development of both anti-malarial drugs and vaccines, as it provides an opportunity to interrupt the life cycle of the parasite at a critical early stage. Therefore, screening platforms that model the in vivo Plasmodium liver stage could be used to advance the pipeline for anti-malarial drug development and also to validate promising liver stage vaccine candidates (Epstein et al., 2011; Plowe et al., 2009).
Studies of rodent Plasmodium pathogens (P. berghei and P. yoelii) have provided important insights, due in part to the capacity to conduct both in vitro and in vivo assays (Hoffman et al., 1989; Rodrigues et al., 2008; Silvie et al., 2007). Nonetheless, there are essential differences between the rodent and human parasites, such as their antigenic variation and mechanisms of host cell invasion (Carlton et al., 2002; McCutchan et al., 1985). To date, our understanding of the liver stage of human malaria, mainly P. falciparum and P. vivax, is based in large part on the infection of human hepatoma cell lines (Chattopadhyay et al., 2010; Epstein et al., 2011; Hollingdale et al., 1983; Karnasuta et al., 1995; Sattabongkot et al., 2006). These cell lines, however, display abnormal proliferation, aberrant signaling, dysregulated gene expression, altered host responses to infection, and inadequate CYP450 and drug metabolism activity, and thus do not accurately recapitulate human hepatocyte biology. Furthermore, in situ observation of pathogen development in liver cell lines is typically obscured after 6 days in culture, due to the continued proliferation of infected cells (Yokoo et al., 2004). Primary cultured human hepatocytes that exhibit more physiologic human liver functions have been studied previously, albeit less frequently than cell lines, and can support the development of the liver forms of P. falciparum and P. vivax (Mazier et al., 1985; Mazier et al., 1984; Rodrigues et al., 2008; van Schaijk et al., 2008; Yalaoui et al., 2008). Nonetheless, in the 25 years since these findings were first published, primary hepatocyte systems are rarely employed and difficult to translate to screening platforms due to limited cell availability, and challenges in maintaining their functional phenotype over extended periods of time in vitro (Bhatia et al., 1999).
Two recent advances may help overcome the deficiencies of existing in vitro liver models. First, several groups have developed methods and novel culture platforms that support the maintenance of primary human hepatocytes (Bhatia et al., 1999; Guguen-Guillouzo and Guillouzo, 2010; LeCluyse et al., 2012). In particular, our group has developed a microliver platform, which leverages bioengineering techniques to organize primary human hepatocytes amongst supportive stromal cells (Khetani and Bhatia, 2008). Hepatocytes in these micropatterned cocultures (MPCC) exhibit human-specific drug metabolism, retain drug responsiveness and hepatic energy metabolism, secretion of liver-specific proteins, polarize, and do not proliferate. Importantly, the hepatocytes in MPCC maintain a functional phenotype for up to 4-6 weeks, and are compatible with medium-throughput drug screening methods and automated data collection. This platform has been found to be more predictive than existing in vitro liver models for generating and identifying human drug metabolites, drug-induced liver toxicity, and supports the persistent replication of hepatitis C virus (Jones et al., 2010; Ploss et al., 2010; Wang et al., 2010). Furthermore, the availability of large cryopreserved lots of human primary hepatocytes means that donor dependent inter-experimental variability can be minimized. Nonetheless, reproducible access to sporozoites is also critical to achieve a practical system. Cryopreservation of large batches of aseptic sporozoites has also now been established for both P. vivax (Chattopadhyay et al., 2010) and P. falciparum (Epstein et al., 2011) sporozoites.
In this resource report, we demonstrate the feasibility of integrating cryopreserved human hepatocytes in MPCC with cryopreserved P. falciparum and P. vivax sporozoites to form an in vitro platform that supports the liver stages of human malaria infection. This platform offers the potential for automation based on pre-selection of cryopreserved human hepatocyte and sporozoite batches to standardize infection rate, a machine-learning algorithm that enables an automated imaging-based readout of immunofluorescent staining, and generates a positive Z-factor in response to drug exposure. The 96-well, medium-throughput format requires less reagents (drugs, sporozoites) than larger footprint in vitro assays or in vivo assays such as humanized mice (Vaughan et al., 2012). Collectively, our data document the development and characterization of a highly-reproducible, medium-throughput microscale human liver platform that may aid in the development of safe and efficacious antimalarial drugs and liver-stage vaccines.
In order to establish an MPCC in vitro culture of primary human hepatocytes (Figure 1A) suitable for Plasmodium infection, we screened cryopreserved hepatocyes from several individual patient donors to identify those that met the following criteria: a) selective adhesion to collagen type I; b) maintenance of a functional hepatocyte phenotype for up to 3 weeks, as assessed by albumin expression, urea production and CYP450 activity; and c) expression of the previously identified Plasmodium host entry factor, CD81 (Silvie et al., 2003). We identified eight donor sources of cryopreserved human hepatocytes that were plateable on collagen I and that displayed typical hepatocyte morphology, including the presence of bile canaliculi (Figure S1). Seven of these sample sets demonstrated functional capacity in culture, as quantified by their production of albumin and urea, and also exhibited CYP450 activity for up to 3 weeks (Figure 1B). We next quantified the expression of the host entry factor CD81 on hepatocytes from each of the seven donors at the day of infection. Four of these samples (Donors 2, 3, 7 and 8) expressed high levels of CD81 by immunofluorescence (Figure 1D).
To test whether primary hepatocytes stabilized by culturing in MPCCs can be infected with P. falciparum, we exposed MPCCs to cryopreserved P. falciparum sporozoites (NF54) (Figure 2A,B). We confirmed productive infection in all 7 donors by staining for HSP70 expression, however, a wide range of infection rates were observed between lots (Figure 1C, left panel). Specifically, at day 3 post-infection, the number of HSP70+, P. falciparum infected hepatocytes was much higher in donor 7 and 8 samples relative to the other tested lots. Hepatocytes from donors 7 and 8 were selected for further characterization of the model. We also examined the susceptibility of the same donor samples to productive infection by two different rodent species of Plasmodium. As shown in Figure 1C, middle panel, donor 7 and 8 also exhibit the highest levels of P. yoelii infection. However, this correlation was not observed in P. berghei (Figure 1C, right panel). Overall, infection with P. berghei was higher compared with P. falciparum and P. yoelii, with donor 1 showing slightly higher infection with P. berghei than the other donors. Interestingly, high CD81 expression was necessary but not sufficient to support robust infection by P. falciparum (Figure 1C and 1D). We next evaluated three different batches of cryopreserved P. falciparum sporozoites. As seen in Figure 1F, this experiment confirmed productive infection using sporozoites from three cryopreserved batches, however, infection efficiencies varied across batches. Based on higher infection levels, sporozoites from batch 2 were selected for further characterization of the model. Importantly, we observed higher infection rates when hepatocytes were cultured in the MPCC format as opposed to standard, unpatterned (randomly distributed) monocultures (Figure 1E, Hep Random). MPCCs remained susceptible to P. falciparum infection for many weeks after they were patterned; however, infection rates are optimal at day 2 after patterning (data not shown).
Having established that P. falciparum can infect MPCCs, we next sought to establish whether the entire liver stage could be reproduced in vitro. The viability of the cryopreserved P. falciparum sporozoites was evaluated by assessing their gliding motility (Figure 2C), and by their capacity to traverse across cells using a cell wounding assay (Mota et al., 2001). On average, the addition of 37,000 motile cryopreserved P. falciparum sporozoites to 10,000 patterned hepatocytes resulted in 3% rhodamine+, traversed cells (Figure 2D). To quantify the ability of the P. falciparum sporozoites to invade hepatocytes, P. falciparum-treated MPCCs were fixed 3 hours post-infection(Renia et al., 1988). On average, 10% of hepatocytes contained intracellular cryopreserved P. falciparum sporozoites (Figure 2E). The human hepatocyte infection rate, based on the percentage of cells containing HSP70-expressing parasites at day 3 post-infection was 0.2% (Figure 2J). Notably, at this same time point, fresh sporozoites achieved infection rates that were 7- to 13-fold higher (Figure 5A). Representative images of the HSP70-expressing parasites at day 3 and day 5 are shown in Figures 2F and 2G.
The maturation of parasites derived from cryopreserved P. falciparum was assessed using three parameters: a) immunostaining the infected cultures with antibodies against two proteins first expressed during late liver stages, PfEBA-175 (erythrocyte-binding antigen, 175 kDa) and PfMSP-1 (merozoite surface protein 1), which indicate full development of the respective schizonts inside the hepatocytes into merozoites; b) the size of HSP70-expressing schizonts, relative to previous reports of P. falciparum schizont size during liver stage development; and c) the progression rate, calculated as the percentage of schizont-bearing cells at day 5 relative to day 3. At day 5 post-infection, cryopreserved sporozoites express both PfEBA-175 and PfMSP-1 (Figure 3C). The parasites we observed were similar in size to what has been reported in other in vitro settings (approx.10 to 15μm at day 5) (Mazier et al., 1985; van Schaijk et al., 2008), but smaller than those reported in vivo (Shortt et al., 1951; Shortt and Garnham, 1948; Vaughan et al., 2012), (Figure 2L and and3B).3B). Importantly, we demonstrated that the rate at which parasites progressed to the schizont stage between day 3 to day 6 post-infection was higher in infected MPCCs (33%) relative to the commonly-used hepatoma line, HC04 (14%) (Figure 2K).
Finally, to demonstrate that hepatic schizonts derived from cryopreserved P. falciparum sporozoites could achieve full maturity and release infective merozoites, red blood cells were added to the hepatocyte cultures at day 6 post-infection. After 10 days, Giemsa staining revealed infection of erythrocytes by P. falciparum merozoites (Figure 2H and 2I).
The ability of the MPCCs to recapitulate the entire liver stage of P. falciparum in vitro highlights the potential to use this platform to study the biology of P. falciparum infected hepatocytes. For example, this capability should enable the assessment of candidate pre-erythrocyte malarial vaccines that are based on live attenuated parasites (Annoura et al., 2012; Epstein et al., 2011; Mueller et al., 2005; van Schaijk et al., 2008).
To illustrate the potential use of MPCCs in vaccine development, we compared the infection capacity of a pre-erythrocytic malaria vaccine candidate comprised of cryopreserved, live-attenuated P. falciparum sporozoites to that of cryopreserved non-attenuated (wild-type) P. falciparum sporozoites. Entry of the attenuated parasites was evaluated relative to the function of non-attenuated parasites. As seen in Figure 3A, entry by both groups of parasites was similar. Late liver-stage development was evaluated at day 5 post-infection. Immunofluorescence staining with EBA-175 and MSP-1 antibodies detected mature schizonts only in MPCCs infected with the non-attenuated sporozoites, whereas MPCCs infected with the live-attenuated sporozoites were positive only for the early liver-stage antigen, LSA-1 (Figure 3C). As expected, schizonts that were established by the non-attenuated sporozoites were larger than the immature forms established by the attenuated sporozoites, where the hepatic stage development is arrested. Figure 3B shows a scatter plot of the range of EEF sizes generated by each group.
We next explored the utility of the MPCCs as a potential anti-malarial drug screening platform. The hepatocyte serves both as the site of antimalarial drug metabolism (or bioactivation) and the host for the parasite. Thus, phenotypic stability of the hepatocyte phenotype and a full drug metabolism repertoire has the potential to capture a full range of drug responses such as efficacy, drug-drug interaction and toxicity (Ploss et al., 2010). MPCCs were established in a 96 well format, infected with P. falciparum, and treated with the canonical malaria drug, primaquine. The impact of the drug was evaluated based on its ability to reduce parasite infection relative to control cultures in a multi-day dose. As seen in Figure 4A and 4B, the IC50 for primaquine was approximately 1μM for P. falciparum during 3 days in culture. The IC50 for P. falciparum in the MPCC was lower compared to the IC50 obtained using unpatterned, monocultures of the hepatoma cell line, HC04 (Figure 4A and 4B). Primaquine is known to act via bioactivation to active daughter metabolites in the liver (Pybus et al., 2012). Thus, the clearance of the parent compound can be used as a proxy for the level of bioactivation in culture. We monitored depletion of primaquine in MPCCs, patterned monocultures of primary hepatocytes only (Hep MP), and the HC04 cell line over two days by high-performance liquid chromatography and found that MPCCs most efficiently cleared primaquine over this time frame (Figure 4C). In order to correlate these macroscopic differences in depletion of a parent compound with the drug metabolism repertoire of these model systems, we then compared the expression levels of 83 human-specific drug metabolism genes. In general, the bulk of hepatocyte drug metabolism genes were expressed at higher levels in the MPCCs (Figure 4E). Of the three major enzymes involved with primaquine metabolism, 2 of the 3 were expressed at higher levels in MPCC than the other systems. In particular, Monoamine oxidase A (MAO-A) which is thought to account for over 75% of primaquine metabolism, was expressed at 4-fold higher than hepatoma cell lines, suggesting that the platform may offer a more predictive system with which to assay candidate drug performance, including studies of mechanism of action.
Next, to further adapt the microliver platform for medium-throughput screens, we characterized several critical parameters. As seen in Figures 5A and 5B, we found that day-to-day variability of hepatocyte infection using the same batch of cryopreserved sporozoites could be minimized (CV 10%) when compared to the range of infection rates obtained using 3 independent batches of fresh sporozoites (CV 42%). Next, a positive Z-factor (Z’>0), which indicates confidence in separating two normally distributed populations was obtained when primaquine-treated wells (10 μM) were compared to control cultures (Figure 5C). In addition, the reproducibility of a phenotypic assay was assessed by establishing consistent dose-dependent inhibitory effects of two different drugs at three different concentrations in 2 independent experiments (atovaquone and primaquine) (Figure 5D). Furthermore, the magnitude of the inhibitory impact of primaquine was consistent across two independent human hepatocyte donors (Figure S2). Finally, we established and validated an automated image analysis read-out using this platform, following the infection of MPCCs with two different doses of sporozoites. As shown in Figure 5E and Figure S3, the number of parasites detected by automated image acquisition and analysis closely matches the counts obtained using conventional manual fluorescence microscopy.
Plasmodium vivax differs from P. falciparum in several important ways. A key feature of P. vivax that underlies its persistence in the population is that the liver acts as a reservoir for dormant hypnozoites or ‘small forms’ (Cogswell, 1992; Krotoski et al., 1982). These hypnozoites can reactivate after weeks, months, or even years, depending on the strain of P. vivax (Dao et al., 2007; Durante Mangoni et al., 2003; Garnham et al., 1975). However, due to a lack of model systems available to investigate this elusive organism, the biology of the dormant form of P. vivax is underexplored.
Based on our ability to establish the liver stage of P. falciparum in MPCCs, to maintain the hepatocyte phenotype for 4-6 weeks, and the observation that some strains of P. vivax can reactivate over a similar time scale, we next explored the feasibility of establishing the liver stages of P. vivax in MPCC over time. As for the P. falciparum experiments, the viability of the cryopreserved P. vivax sporozoites was evaluated by assessing their gliding motility prior to each experiment (Figure 6A). We explored several strains of P. vivax including Chesson, a strain known to efficiently form dormant forms and to reactivate at shorter time scales (Hollingdale et al., 1986; Krotoski et al., 1986). Cultures were infected and then fixed at various time points and stained for CSP. Using immunofluorescence microscopy, we analyzed the size distribution and localization of the P. vivax forms over a three week period. We readily observed P. vivax liver forms in MPCC, including mature, liver-stage schizonts larger than 20μm found at day 6 post-infection (Figure 6C, 6B and Figure S4A-B). In addition, a population of smaller PvCSP-positive forms (<5μM) was detected within MPCCs at all time points, particularly identifiable at 6 and 21 days post-infection (Figure 6A, 6B and Figure S4A-C). The infection efficiency, based on the percentage of hepatocytes containing large and small PvCSP+ forms at day 6 post-infection was 0.013% and 2%, respectively (Figure S4D). Similar numbers of small forms were observed up to day 21 (data not shown). Importantly, 85% of the small forms observed were intracellular, based on immunostaining performed before and after permeabilization (Figure S4C). In contrast, cultures infected with P. falciparum contained very few small forms after 15 days in culture, and those were predominantly extracellular (Figure S4E). These P. vivax small forms may represent the dormant hypnozoite stage of the parasite life cycle, which are responsible for clinical relapses in P. vivax malaria patients; however, further characterization will be required to substantiate this hypothesis. Both small and large forms were observed when other strains of fresh and frozen P. vivax sporozoites were examined (Figure S4A and data not shown). Finally, we demonstrated that the MPCC system can support maturation of hepatic P. vivax schizonts, based on the detection of the late stage antigen, PvMSP-1, at day 12 post-infection (Figure S4B).
In this report, we describe an in vitro cell-based platform that recapitulates the human liver stage of P. falciparum and P. vivax infection. Although some attempts to infect cryopreserved human primary hepatocytes have been described in the past (Meis et al., 1985; Silvie et al., 2004), this source of hepatocytes has not been routinely adopted by the field, to date. The advantages of screening individual donor-derived cryopreserved hepatocytes is paralleled by the successful production of purified, vialed, cryopreserved sporozoites (Chattopadhyay et al., 2010; Epstein et al., 2011; Hoffman et al., 2010). These cryopreserved resources overcome the donor-to-donor variability seen in primary cultured human hepatocyes, as well as infectivity rate variability introduced by different batches of mosquitoes or sporozoites. In addition, the use of cryopreserved components in this platform allows for a reliable source of reagents for use in longitudinal experiments, including screening and subsequent validation. Furthermore, since patterned hepatocyte culture using the MPCC platform can maintain individual patient samples for between 4 and 6 weeks (Khetani and Bhatia, 2008), long-term monitoring of even blood-stage Plasmodium infections, analysis of genetic changes acquired inside Anopheles mosquitoes, safety assessment of attenuated sporozoite vaccine candidates, and in vitro characterization studies of P. vivax hypnozoites are all feasible. Notably, development of vaccines and drugs against this stage of P. vivax has been identified as a critically important goal for research to eradicate malaria (Alonso et al., 2011a; Alonso et al., 2011b). Recent publications have highlighted that existing candidate vaccines continue to underperform in clinical trials, and significant “blood breakthrough” of presumably attenuated parasites formulations has been observed (Annoura et al., 2012).
In our experience with disease modeling of drug-induced liver injury and hepatitis C infection, establishing platforms that better reflect host biology is an important first step to determining where existing model systems were lacking (e.g. P450 activity and interferon signaling, respectively). In this case, we have already observed three advantages over in vitro hepatoma cultures: 1) sporozoites appear to progress through the parasite life cycle more efficiently in MPCC relative to infected hepatoma cells, offering the potential to improve studies of drugs and sporozoite attenuation strategies that act in the second half of the liver life cycle; 2) MPCC are able to predict differences in infection rates of sporozoites in vivo that result from cryopreservation, likely reflecting the presence of critical host factors that are altered in hepatoma cells (Albuquerque et al., 2009; Epiphanio et al., 2008; Prudencio et al., 2008; Rodrigues et al., 2008; Silvie et al., 2003); and 3) MPCCs fabricated from different human hepatocyte donors enable direct comparison of host factors that impact entry of different sporozoite species and strains (e.g. CD81 in P. yoeli, P. berghei, P. falciparum), whereas hepatoma cells are typically limited to one or very few donor genotypes. We demonstrated that maintenance of hepatocyte function and expression of the host entry factor CD81 (Silvie et al., 2003) were necessary but not sufficient to obtain adequate levels of infection by P. falciparum. This finding suggests the existence of as-yet uncharacterized molecular differences among donors that determine their permissiveness to Plasmodium infection. Furthermore, most antimalarial drug development leads identified in red blood cell-based high throughput screens do not require metabolic activation, thus in such cases, screening via the MPCC format might yield the same or higher/lower IC50 predictions, should the candidate compound be cleared rapidly or bioactivated via metabolic pathways (Gamo et al., 2010; Guiguemde et al., 2010)
The infection rates reported in this study using cryopreserved sporozoites are comparable or higher than the ones previously documented with fresh sporozoites, and they are an additional ~10-fold higher when we used fresh sporozoites (Mazier et al., 1985; Mazier et al., 1984; Sattabongkot et al., 2006; van Schaijk et al., 2008). The infection rates of hepatocytes in MPCC is similar in comparison to the infection rates recently reported in HC04 cells which range from 0.4% to 0.06% (Epstein et al., 2011; Sattabongkot et al., 2006); however, in our system the progression rate from one stage of the life cycle to the other was much higher than in HC04, offering the potential for studying later stages of the liver life cycle more efficiently. Nonetheless, our MPCC infection rates remain low relative to those recorded in in vivo settings (Shortt et al., 1951; Shortt and Garnham, 1948). Experiments done in mice with P. yoelii (Conteh et al., 2010), and nonhuman primates with P. knowlesi (Jiang et al., 2009) have demonstrated that intravenous inoculation of only a few non-cryopreserved sporozoites (10 spz) can lead to a productive malaria infection that results in detectable parasitemia in the blood stage. Varying hypotheses have been put forward to explain the discrepancy between model systems. For example, our in vitro systems do not provide the host hepatocytes with potentially necessary, physiologically relevant cellular components, such as Kupffer cells or sinusoidal endothelial cells. Further, the MPCC platform conformational cues may be important for EEF growth (e.g. 2D versus 3D). These hypotheses will be explored further.
In conclusion, the ability to support the liver stages of P. falciparum and P. vivax parasites in a medium-throughput format offers promise to improve our fundamental understanding of the liver stages of human malaria as well as accelerate the development of drugs and vaccines to aid eradication.
12mm coverslips that were placed into tissue culture polystyrene 24-well plates or glass-bottomed 96-well plates were coated homogenously with rat tail type I collagen (50 μg/ml) and subjected to soft-lithographic techniques (Ploss et al., 2010) to pattern the collagen into micro-domains (islands of 500 μm) that mediate selective hepatocyte adhesion. To create MPCCs, cryopreserved primary human hepatocytes were pelleted by centrifugation at 100× g for 6 min at 4 °C, assessed for viability using Trypan blue exclusion (typically 70–90%), and then seeded on collagen-micropatterned plates. The cells were washed with medium 2–3 h later and replaced with human hepatocyte culture medium. 3T3-J2 murine embryonic fibroblasts were seeded (40,000 cells in each well of 24-well plate and 10,000 cells in each well of 96-well plate) in fibroblast medium 3 days later. Fibroblast medium was replaced with human hepatocyte medium 24 h after fibroblast seeding and subsequently replaced daily (Khetani and Bhatia, 2008).
Mosquitoes were fed on Pf- and Pv- infected blood as previously described (Chattopadhyay et al., 2010; Epstein et al., 2011). Briefly, P. falciparum and P. vivax sporozoites were extracted from infected mosquitoes by dissection of their salivary glands and passing the glands back and forth through a 26K G needle fitted to a 1 mL syringe. Following extraction, sporozoites were purified from mosquito salivary gland material contamination and cryopreserved in liquid nitrogen vapor phase (LNVP) (Epstein et al., 2011; Hoffman et al., 2010). Live-attenuated P. falciparum sporozoites were attenuated by exposure of PfSPZ-infected mosquitoes to 150 Gy (Epstein et al., 2011; Hoffman et al., 2010).
Typically, infection of MPCCs is conducted two days after hepatocyte seeding, but can also be initiated after longer culture periods. Cryovials containing P. falciparum or P. vivax were warmed for 30 sec at 37°C, 200 μl of human hepatocyte medium was added and the cryovials were centrifuged for 2 min at 14000× g. 200 μl of the supernatant was aspirated, and the sporozoite pellet was resuspended and diluted accordingly. Each well was infected with a ratio of 3:1 (infective sporozoites:hepatocytes). After incubation at 37°C and 5% CO2 for 3h, the wells were washed once and fresh medium was added. Media was replaced daily. Samples were fixed on day 3 and 5 post-infection with P. falciparum, and days 3, 6, 12, and 21 post-infection with P. vivax.
Sporozoite migration through cells can be quantified by the detection of sporozoite-wounded hepatocytes using a cell-impermeant fluorescent tracer macromolecule as previously described (Mota et al., 2001). Briefly, MPCCs were infected with P. falciparum in the presence of 1 mg/ml tetramethylrhodamine dextran lysine fixable, 10,000 Da (Sigma). At 3h post-infection, MPCCs were washed thrice with PBS, fixed with 1% paraformaldehyde at room temperature for 20 min, and mounted on glass slides. Migration of sporozoites through cells is quantified by the number of dextran-positive hepatocytes per island.
At 3 h post-infection, primary human hepatocytes or MPCCs were fixed and stained using a double-staining protocol as previously described (Renia et al., 1988). Briefly, to label extracellular sporozoites, the samples were first fixed with 4% paraformaldehyde for 10min at room temperature, blocked with 2% BSA in PBS, incubated with a primary mouse anti-PfCSP (Sanaria, 1:100) (Nardin et al., 1982), washed thrice in PBS and incubated with a secondary goat anti-mouse AlexaFluor488 conjugate. This was followed by a permeabilization with −20°C methanol for 10min at 4°C, incubation with the same primary mouse anti-PfCSP, washing thrice with PBS, and incubation with a secondary goat anti-mouse AlexaFluor594 conjugate. This second step labels both intracellular and extracellular sporozoites. In case of MPCCs, the samples were counterstained with Hoechst and mounted on glass slides as described above. The number of invaded sporozoites (only green) in primary human hepatocytes were counted using Acumen Explorer.
Motility of cryopreserved sporozoites was determined in each batch to define the number of infective sporozoites. Sporozoite gliding was evaluated with 20,000 sporozoites for 40 minutes in complete DMEM, at 37°C on glass cover slips covered with anti-circumsporozoite protein (CSP) monoclonal antibody (clone 2A10 for P. falciparum and 210 clone NSV3 for P. vivax). Sporozoites were subsequently fixed in 4% paraformaldehyde (PFA) for 10 minutes and stained with anti-CSP. The percentage of sporozoites associated with CSP trails was visualized by fluorescence microscopy. Quantification was performed by counting the average percentage of sporozoites that perform at least one circle.
Infected MPCCs were incubated with media containing primaquine diphosphate (Sigma) ranging from 0.5-20 μM. Fresh primaquine-containing medium was added daily until the samples were fixed at days 3 and 5 post-infection.
Total RNA from three wells per condition was purified using Trizol (Invitrogen) and Mini-RNeasy kit (Qiagen) and pooled for analysis. LMA-Luminex procedures and probes are previously described (Chen et al., 2011). Briefly, data for triplicate loadings, expressed in mean fluorescent intensity of at least 100 beads per sample, were scaled to the human transferrin gene and row-normalized for heatmap representation using Gene Pattern open-source software (Broad Institute).
Images of 96 well plates were acquired using high content screening microscopes (Molecular Devices IXM), and were then analyzed by Cell Profiler and Cell Profiler Analyst (Broad Institute) (Carpenter et al., 2006; Jones et al., 2008). Parasites were visualized through immunofluorescent staining of the HSP70 protein, and can be distinguished from imaging artifacts by their proximity to a hepatocyte nucleus (within 30 pixels), and lack of auto fluorescence (no signal in unlabeled channels). We developed an automated image analysis pipeline to identify every infection in every image and measure hundreds of features (e.g. shape, area, texture) of each parasite. These features are then used to train machine learning algorithms to identify and count the number of parasites in each image using CellProfiler Analyst.
We thank the Malaria Research and Reference Reagent Resource Center (MRA-16, PvMSP-1/19 rabbit antiserum), deposited by (JH Adams, USF); New York University for the monoclonal antibody, 2A10; NIAID, NIH for R217; Dr. R. Wirtz, Centers for Disease Control and Prevention, and Dr. F. Zavala, Johns Hopkins University for PvCSP and HSP70 monoclonal antibodies, respectively. We are grateful to the Sanaria Manufacturing Team for the production of PfSPZ and PvSPZ. R. Schwartz for confocal microscopy help, S. Suresh and M. Diez for aid in establishing RBC cocultures, and A. Rodriguez (NYU), D. Wirth and E. Lund (HSPH) for providing mosquitoes infected with P. yoelii and berghei, J. Prachumsri (Mahidol Vivax Research Center) and J. Adams (University South Florida) for providing fresh P. vivax, H. Green (Harvard) for providing J2-3T3 fibroblasts, and H. Fleming for manuscript editing. PvSPZ production was supported by a grant from Medicines for Malaria Venture, and PfSPZ production was supported by a Phase II NIAID, NIH Small Business Innovative Research Grant (2R44A1055229) awarded to SLH. Software improvements were supported by an NIH grant (R01GM089652) to AEC. This work was supported by the Bill & Melinda Gates Foundation (Award #51066). SN is supported by an A*STAR (Agency for Science, Technology and Research, Singapore) NSS. SNB is an HHMI Investigator. The authors wish to dedicate this paper to the memory of Officer Sean Collier, for his caring service to the MIT community and for his sacrifice.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Supplemental data include four figures