Malaria is a devastating disease that causes over 400,000 deaths per year, mainly among young children in sub-Saharan Africa [1
parasites are responsible for causing disease, and key gaps exist in our understanding of the parasite’s lifecycle. Following transmission from an infected female anopheline mosquito to the skin of a human during blood feeding, the Plasmodium
parasite makes its way by gliding motility to blood vessels and enters the bloodstream to hone to the first site of invasion and development, the liver [2
]. Unlike the cyclical development of Plasmodium
in erythrocytes, the liver stage (LS) of infection is clinically silent [3
]. The LS has not been as well-studied as many other steps in the parasite life cycle. When studying the LS, it is important to consider not only sporozoite invasion of hepatocytes but also the steps leading up to this event. To gain access to hepatocytes, sporozoites must traverse the sinusoidal barrier, which contains liver endothelial cells and Kupffer cells (KCs). It is estimated that at least 60% of Plasmodium berghei
sporozoites pass through a KC on their way to hepatocytes [4
KCs, also known as the liver-resident macrophages, make up about 35% of the liver non-parenchymal cells in adult mice [5
] and about 30% in humans [6
]. They line the liver sinusoids across the Space of Disse from hepatocytes and rapidly clear bacteria and other foreign particles from the blood stream [7
]. They also play an important role in promoting immune tolerance in the liver to prevent unnecessary inflammation [8
]. However, in cases of high infection levels or liver injury, as demonstrated during Leishmania
infections, KCs can serve as immune activators [12
]. However, in the case of Plasmodium
infection, sporozoites can traverse these KCs without being phagocytosed or killed [15
]. On the other hand, a recent report highlighted that hepatocyte growth factor (HGF) from KCs of infected mice is essential in promoting apoptosis of Plasmodium
-infected hepatocytes [16
], suggesting that the KCs do produce soluble molecules to affect the overall state of the liver during the infection. Additionally, it was previously shown that an innate immune response can be induced during the LS, contributing to host resistance to reinfection [17
], and that leukocytes in the liver can respond to a hepatocyte-propagated type I interferon signal to respond to sporozoite infection [18
]. Furthermore, transmission of Plasmodium
by mosquito bite leads to an increase in the innate immune response when compared to transmission by direct injection of blood stage parasites, suggesting a strong role for the liver’s innate immune system in infection control [19
]. However, the full milieu of proteins secreted from KCs upon sporozoite exposure remains unknown.
When the KC is traversed by the sporozoite, it has been reported that many of the KCs become wounded and succumb to death [4
]. However, signs of collagen secretion and inflammation, which should follow cell wounding and death [21
], have not been noted to take place upon sporozoite traversal and infection of the liver. Hepatocytes, which are similarly traversed by sporozoites, are not largely wounded and killed [4
]. These observations imply that the sporozoite is modulating the cellular responses in its favor through a mechanism that is not well understood.
While previous studies have examined downstream effects of sporozoite exposure on the ability of KCs to mount an immune response against a subsequent LPS challenge and have shown down-modulation of the pro-inflammatory response [20
], few studies have addressed the KC’s immediate response to sporozoite exposure. Therefore, the true fate and activity of the KC upon traversal remains unclear. Here, we determined the innate immunological response of primary rat KCs (PRKCs) to P. berghei
sporozoite exposure, and evaluated whether the PRKCs undergo death following exposure. Our work captured a short-lived KC-cytokine secretion profile that was unique to live sporozoite exposure and waned over time while also providing additional evidence that KCs remain viable following exposure to sporozoites.
Our data demonstrate that PRKCs can mount a rapid and diverse cytokine response to sporozoite exposure. This PRKC response is remarkable since exposure to the well-characterized KC activator LPS resulted in more protracted release kinetics; TNF-α secretion from PRKCs, for example, does not increase until the 4–12-h timeframe (Figure 1
, Supplementary Figure S4
). However, our findings are consistent with the observations that traversal of cells by sporozoites occurs on the scale of minutes [4
] and that just 5 min is long enough for macrophages to begin an immune response [24
]. Additionally, the response to sporozoite exposure was markedly different from the response to uninfected salivary gland extract. While low levels of pro- and anti-inflammatory cytokines were secreted in response to the uninfected salivary gland extract, secretion was statistically significantly higher upon exposure to sporozoites at early timepoints for the M1-associated cytokines IFNγ, IL-12p70, Mip-3α, IL-2, and RANTES; the M2-associated cytokines IL-1α, IL-4, IL-5, IL-13, EPO, and VEGF; and the non-M1-M2 partitioning cytokines IL-7 and IL-17α (Figure 1
and Figure 2
). Some additional cytokines, such as IL-1β, displayed significantly altered secretion levels under only one of the two culture conditions used, preventing conclusive findings for these cytokines.
One important limitation of our study is the variation in the absolute value of the cytokine levels from experiment to experiment. These variations are likely due to culture-specific phenotypes, differences between the donor rats from which the KCs were isolated, and/or sporozoite and mosquito lot to lot variation. It is also important to remember that despite the use of rigorous dissection methods similar to those used in other immune assays like the ILSDA [26
], matched levels of mosquito contaminants will always be present in the sporozoite and uninfected salivary gland pools [27
], which can cause variability in the background KC cytokine secretion from experiment to experiment. Additionally, the in vivo implications of the specific quantities of cytokines produced remain unclear. In spite of these limitations, statistically significant differences between secretion levels from sporozoite-exposed and uninfected salivary gland-exposed PRKCs are robustly replicated between experiments for IFNγ, IL-12p70, Mip-3α, IL-2, RANTES, IL-1α, IL-4, IL-5, IL-13, EPO, VEGF, IL-7, and IL-17α.
The rapid cytokine secretion response is also short-lived, with cytokine levels dropping to control levels within 2–4 h. It has long been known that IL-1β and IL-18 can be stored in the cell as a soluble, inactive form that is activated for rapid secretion and response [29
], but all other cytokines studied herein do not share this phenotype. This may suggest that the KCs do in fact have additional stored cytokine stocks that can be released immediately, as opposed to what has been observed for a targeted, pathogen-associated molecular pattern (PAMP)-activated response. Alternatively, the parasite may be interfering with calcium signaling within the cells, causing a rapid cytokine release that is not maintained as calcium signaling returns to normal over a 2–4-h timeframe following traversal [30
]. If a sustained production of cytokines were occurring, we would expect to see an increase in secretion again by the later time points; however, this was not observed (Figure 1
and Figure 2
). Additionally, the secreted cytokines come from both MyD88-dependent and MyD88-independent pathways [32
]. These data, along with the observation of both M1 and M2 cytokine secretion in our assays, suggest to us that the cytokine response may be non-specific and not reliant on any particular TLR signaling path; further testing of this hypothesis will be needed. Previous studies in Plasmodium
have shown that glycosylphosphatidylinositol anchors can stimulate TLRs and that innate immune cells can recognize TLR ligands [36
], though the exact roles of these processes during the pre-erythrocytic stages have not been well characterized.
Although many of the cytokines and chemokines that we have measured in our assays are not frequently associated with KC responses, they have been shown to be secreted by KCs and/or macrophages in previous studies, i.e. RANTES, EPO, IL-5, VEGF, and M-CSF [33
]. Of particular note, IL-7 is a cytokine that has not been previously associated with secretion by KCs. IL-7 is known to be associated with secretion from hepatocytes [42
], but our KC isolation method excludes the collection of hepatocytes, and hepatocytes were never observed by microscopy in the KC cultures used. We interpret the data as suggesting that KCs may also contribute to IL-7 levels in the liver, though it is important to note that this observation requires further study. Additionally, we cannot completely rule out the possibility that a small number of capsular macrophages, a relatively recently recognized but poorly-described cell population [43
], or perivascular macrophages could have been present in our KC preparations.
Only live, whole sporozoites triggered this increased secretion of the diverse cytokine profile. Cytokine secretion in response to lysed sporozoites was even lower than secretion in response to uninfected salivary gland extracts, likely because any other immune stimuli present underwent the same lysis procedure as the sporozoites, altering their stimulatory capabilities. Similarly, exposure to the traversal-deficient SPECT2−
mutant sporozoites produced slightly lower cytokine secretion than did wild type sporozoites; however, this trend did not achieve significance for most cytokines (Supplementary Figure S1
), leaving open the possibility that traversal of KCs is not required for their innate immune activity. Overall, these data highlight the importance of parasite factors that can actively influence the PRKC innate immune response. For example, it has been previously shown that circumsporozoite protein binds the low-density lipoprotein receptor-like protein 1 on KCs to induce a cAMP-dependent signaling pathway to suppress the oxidative burst normally used to kill pathogens [44
]. It has also been shown that the Plasmodium
sporozoite protein essential for cell traversal (SPECT) putatively interacts with KC surface proteins to facilitate cell traversal [45
]. It is likely that many other poorly characterized Plasmodium
proteins have roles that strongly affect KC function and activation.
Finally, in contrast to previous reports [20
], our data suggest that KCs are not wounded significantly or killed by sporozoite exposure (Figure 4
, Supplementary Table S1
). It is important to note that the live-dead imaging was carried out with cryopreserved PRKCs (Figure 4
B, Supplementary Table S1
), which have been previously noted to have ~60% recovery after thawing [46
], affecting the baseline levels of cell death observed. Two major differences between previous work on KC wounding and death in response to sporozoites and our study are the species and number of sporozoites used to interact with the KCs. In our study, we used P. berghei
at a ratio of 1 sporozoite to 1 KC, while previous work on KC response to sporozoites used P. yoelii
at a ratio of 1 sporozoite to 1 KC and 3 sporozoites to 1 KC [20
]. Memory CD8 T cell effector paths for targeting P. berghei
and P. yoelii
sporozoites are quite different [47
], so it is likely there would also be species to species variation in how KCs respond to Plasmodium
sporozoites. Additionally, the higher numbers of sporozoites interacting with and traversing each KC may have caused more damage, resulting in cell death; in fact, previous work has shown that the use of a 1:1 ratio results in notably lower rates of cell death compared to the 3:1 ratio [20
Overall, our study suggests that the KC–sporozoite interaction is not silent. KCs mount a rapid, short-lived cytokine secretion response without being largely wounded or killed. This work opens the door to answering further questions about this essential step in the parasite’s life cycle. Finding ways to tailor the KC’s response towards a more effective activation of the body’s immune system could potentially eliminate the parasite’s ability to enter the liver so discreetly and aid in malaria eradication efforts.
4. Materials and Methods
All animals and experimental protocols used in this study were approved by the Johns Hopkins Animal Care and Use committee and the IACUC, and the methods were carried out in accordance with IACUC and institutional guidelines and regulations. PRKCs were obtained commercially (cryopreserved rat KCs, ThermoFisher Scientific, Waltham, MA, USA) for live-dead cell imaging assays or were freshly isolated from rats using a protocol adapted from Dr. Zhaoli Sun [48
] for cytokine secretion and LDH assays. It has been previously demonstrated that cryopreserved KCs maintain their phenotypic features and can be successfully used as surrogates for freshly isolated cells [46
4.2. Primary Cell Isolation and Purification
To obtain cells for cytokine secretion and LDH assays, male Lewis rats (150–250 g) were used. Rat KCs were chosen over those of mice due to the large number that can be obtained per animal. Rats were anesthetized with isoflurane; the abdomen was dissected, and the portal vein was cannulated. The liver was then perfused with 100 mL Hank’s Buffered Saline Solution (HBSS) for 5 min. The inferior vena cava was cut to allow the fluid to drain. The liver was then perfused with 100 mL 0.05% collagenase (Type IV from Clostridium histolyticum; Sigma-Aldrich, St. Louis, MO, USA) solution in HBSS over 7–10 min. The liver was then removed from the body cavity and washed with HBSS + 30 mM HEPES + 0.1% calcium chloride. The tissue was mashed in HBSS + 30 mM HEPES + 0.1% calcium chloride with 0.05% collagenase, 0.1 mg/mL DNase, and 0.2 mg/mL Pronase. Connective tissue was removed, and the liver cell suspension was incubated at 37 °C for 15 min with agitation. Cells were then spun at 300× g for 5 min 3 times and washed with HBSS + 30 mM HEPES + 0.1% calcium chloride + 0.05 mg/mL DNase + 1000 units/mL penicillin + 1000 µg/mL streptomycin after each spin. Cells were then spun at 100× g for 1 min to pellet hepatocytes; the pellet was discarded. The supernatant was collected and spun at 300× g for 5 min. The cell pellets were resuspended in HBSS + 30 mM HEPES + 0.1% calcium chloride and applied to a Percoll gradient (15 mL 50% Percoll solution, 15 mL 25% Percoll solution, 15 mL cell suspension). The gradients were spun at 800× g for 15 min. The top fraction of the spun gradient was removed, and PRKCs were collected from the second fraction and washed twice with HBSS + 30 mM HEPES + 0.1% calcium chloride. The final cell pellets were then resuspended in complete RPMI media (cRPMI), defined as RPMI + 10% heat-inactivated fetal bovine serum (HIFBS) + 1 × MEM amino acids solution (ThermoFisher Scientific, Waltham, MA, USA) + 1000 units/mL penicillin + 1000 µg/mL streptomycin, and used for plating following two different methods. (1) Cells were plated on T75 flasks that had been coated for 48 h with HIFBS. After cells had adhered in a humidified chamber at 5% CO2 and 37 °C for an hour, the flasks were washed with HBSS + 30 mM HEPES + 0.1% calcium chloride 3 times, and adherent cells were lifted from the flasks in cold PBS on ice for 1–2 h. This cell suspension was then used to plate cells in cRPMI for 12–16 h before assays. (2) Cells were plated directly for assays on 24- or 48-well plates previously coated for 48 h with HIFBS; cells were allowed to adhere for 2 h in a humidified chamber at 5% CO2 and 37 °C and washed 3 times with HBSS + 30 mM HEPES + 0.1% calcium chloride before assays. All assays were performed in cRPMI.
Primary rat T cells were obtained from the matched PRKC donor rat. Rat spleens were homogenized in cRPMI. Cells were then passed through a 21-guage needle in a 5 mL syringe to filter out tissue clumps. Cells were spun at 300 × g, 7 min. The cell pellet was resuspended in red blood cell lysis buffer (ammonium chloride buffer) and incubated at room temperature for 10 min with agitation. Cells were spun at 300 × g, 5 min, and the cell pellet was resuspended in cRPMI. This cell suspension was then allowed to adhere to plates to remove adherent cells. The supernatant was collected, and cells were pelleted and resuspended in MACs buffer (Miltenyi Biotec). Pan-T cell MicroBeads (Miltenyi Biotec) were added to the cell suspension and incubated on ice for 15 min. The cells were then suspended in 500 µL MACs buffer and run through an LS column on a MACs magnet (Miltenyi Biotec). T cells were collected from the elution of cells that bound to the column in the magnet.
4.3. FACS Analysis for T Cell Quantification
PRKCs and primary rat T cells were used fresh after isolation; PRKCs were isolated and enriched using method (1) above with cells being taken for FACS analysis after lifting in PBS on ice and before the final plating step. PRKCs and T cells were blocked with anti-FcɣR II/III to prevent nonspecific binding. Cells were then stained with anti-F4/80-PE-Cy7 and anti-CD3-AF488. Propidium iodide was used to determine cell viability. Cells were analyzed on a DakoCytomation MoFlo (Beckman Coulter) with the following detection filters: FL1/AF488 (530/30), FL2 (580/30), and FL3/PE-Cy7 (740 L). Wavelengths are reported in nm. Gating was performed to exclude cellular debris by using the forward scatter and side scatter gates. From the population of cells, viable cells were selected by gating for propidium iodide negative cells. Due to the high autofluorescence of KCs, a gating strategy using FL1/AF488 vs FL2, instead of FL3/PE-Cy7, was used for subsequent gating to identify the T cell population. For more detail on the gating strategy, see Supplementary Figure S2
4.4. Sporozoite Generation and Collection and Uninfected Salivary Gland Extract Collection
(day 6–10) mosquitoes were fed on a mouse infected with P. berghei
mCherry parasites or P. berghei
parasites (a mutant deficient in cell traversal) exhibiting 0.5–2 exflagellations per field under a 40× objective. Fully fed mosquitoes were dissected 18–24 days post-feed to collect salivary gland sporozoites, based on the standard method for inhibition of liver stage development assays [26
]. Unfed mosquitoes reared in parallel were dissected simultaneously to collect uninfected salivary gland extracts to use as a control to account for mosquito proteins and other potential mosquito-derived contaminants that cannot be separated from sporozoites despite the most rigorous washing and purification steps [27
]. Mosquitoes were collected in 70% ethanol, then transferred to 1 × PBS in a petri dish on ice; salivary gland pairs were dissected and collected into 500 µL cRPMI on ice; the glands were lightly spun for 3 min at 1200× g
and then crushed by hand with a plastic, sterile pestle; this 500 µL crushed preparation was then filtered through glass wool before use. Both sporozoites and infected salivary gland extracts were then diluted identically at least 10-fold in sterile cRPMI before being added to cells in a volume of 500 µL in 24-well plates or 200 µL in 48-well plates. Sporozoites were used at a ratio of 1 sporozoite:1 PRKC, and the analogous volume of uninfected salivary gland extract was used. Although we noted that direct salivary gland dissections (as opposed to commonly used thorax-dissections) produced sporozoites without any large mosquito cellular debris, these sequential “dilution-washing” steps were included to “dilute” the potential gland-derived contaminants from the sporozoites; especially since low rcf centrifugation does not effectively pellet sporozoites that have been released from salivary glands, and higher rcf negatively impacts sporozoite viability/activity. Lysis of sporozoites was achieved by rapid freezing and thawing of sporozoites in liquid nitrogen (1 min) and a 37 °C water bath (4 min) 5 times. Fixation-mediated killing of sporozoites was achieved by incubating the released sporozoites in 4% paraformaldehyde for 20 min at room temperature. The sporozoites were then further washed in 1 mL cRPMI and spun to remove excess paraformaldehyde that could otherwise affect the KC response.
4.5. Bio-Plex Cytokine Assays
Cells were plated in 24- or 48-well plates following methods (1) or (2) outlined in the “Primary Cell Isolation and Purification” section. The same size plate was used for all samples within a given experiment, and in experiments where secretion from T cells and PRKCs are compared, the same number of cells were plated for each cell type. Cells were exposed to no stimuli (naïve), uninfected mosquito salivary gland extracts (sg), P. berghei mCherry sporozoites at a ratio of 1 sporozoite to 1 cell (Pb), P. berghei SPECT2− sporozoites at a ratio of 1 sporozoite to 1 cell (SPECT2−), 1 µg/mL LPS from E. coli (LPS), lysed P. berghei mCherry sporozoites at a ratio of 1 sporozoite to 1 cell (lys.), or P. berghei mCherry paraformaldehyde-fixed sporozoites at a ratio of 1 sporozoite to 1 cell (fPb) and incubated in a humidified chamber at 5% CO2 and 37 °C. The ratio of 1 sporozoite to 1 cell was chosen to most closely mimic the natural in vivo conditions of the KC-sporozoite interaction. After the appropriate amount of time, the supernatant from the culture was removed and spun at 12,000× g for 10 min to remove any cellular debris before being used for analysis. Supernatants were analyzed using the Bio-Plex cytokine array platform following manufacturer’s instructions (Bio-Rad). Briefly, the cytokine assay standard(s) were reconstituted on ice for 30 min. A fourfold dilution series of the standard was made. The Bio-Plex magnetic beads were added to the assay plate and washed twice. Standards, blanks, and supernatants were then added to the plate and allowed to bind the beads for 1 h. The plate was washed three times, and then detection antibodies were added for 30 min. The plate was washed three times, and then Streptavidin-PE was added for 10 min. The plate was washed three times again, and then the beads were resuspended in assay buffer to be read on a Bio-Plex 200 instrument (Bio-Rad) using high PMT or the Luminex MAGPIX instrument. Standard curves for each cytokine were analyzed and optimized in the Bio-Plex Manager software. PRKC supernatants were run in biological triplicate with technical duplicates. T cell supernatants were run with technical duplicates.
4.6. Lactate Dehydrogenase (LDH) Assay
The lactate dehydrogenase assay (Pierce LDH cytotoxicity assay kit) was performed following the manufacturer’s protocol using freshly isolated PRKCs. We determined in our initial pilot experiments that the dextran assay, commonly used to determine hepatocyte cell fate following sporozoite cell traversal [49
], is inappropriate for our system. We observed that naïve PRKCs take up dextran without any stimulus, which would significantly compromise qualitative/quantitative measures of cell viability. We collected supernatants from PRKCs exposed to no stimuli (naïve), uninfected salivary gland extracts (sg), P. berghei
sporozoites (Pb), or lysed P. berghei
sporozoites (lys.) for use in the LDH assay. The reaction mix from the kit was added to the supernatants and incubated for 30 min. Stop solution was then added, and the absorbance at 490 nm and 680 nm were measured. The background 680 nm absorbance was subtracted from each read, and the output was normalized to the kit positive control (cells + lysis buffer) at 1.0.
4.7. Live/Dead Cell Imaging Assay
Due to the existence of multiple forms of cell death that can involve many different pathways [50
], we included an orthogonal assay to measure KC viability following sporozoite exposure. The LDH assay typically detects cells undergoing necrosis, but not necessarily other forms of cell death such as apoptotic, quasi-apoptotic, and nonapoptotic mechanisms [51
]. Additionally, the LDH assay focuses on a population of cells instead of single-cell death events, which can provide additional information about the cell death process [53
]. Therefore, to supplement the LDH assay, we included an additional live/dead assay that allows single-cell analysis and is not necrosis-specific through the use of a cell permeable calcein AM that measures both intracellular esterase activity and membrane integrity. When a cell is alive and has an intact membrane, calcein AM enters the cell and is converted to fluorescent calcein by intracellular esterase activity, while a dead cell with a damaged membrane takes up the cell impermeable nuclear dye. The ThermoFisher Live/Dead cell imaging kit was used following the manufacturer’s protocol. Commercially purchased cryopreserved rat KCs (ThermoFisher Scientific, Waltham, MA, USA; cell purity ≥ 90%) were thawed and plated directly into 24-well plates on collagen-coated coverslips overnight before beginning the assay. KCs were then exposed to no stimuli (naïve), uninfected salivary gland extracts (sg), P. berghei
mCherry sporozoites (Pb), or 387.5 ng/mL listeriolysin O, a pore forming toxin (LLO). LLO was used as a positive control to confirm that the assay could appropriately identify dead cells and differentiate live cells that had taken up the cell impermeable dye in a phagosome from dead cells that had the dye in the nucleus. After 15 min, the live/dead imaging reagent mix was added to the cells. Cells were imaged, and the number of live (fluorescing green) and dead (fluorescing red in the nucleus) cells was counted at 30 min, 1.5 h, and 3 h following cell exposure to stimuli. Cells with red fluorescence in phagosomes but not in the nucleus were not counted as dead cells. Morphology to examine cell size and shape, key distinguishing features between PRKCs and other commonly contaminating liver cells such as hepatocytes and stellate cells, was used to confirm cell purity. For each stimulus, three biological replicates were analyzed, and five microscope fields of each replicate were counted at each time point using the 10× objective on the EVOS Cell Imaging System (ThermoFisher Scientific, Waltham, MA, USA).