Impact of Plasmodium falciparum small-sized extracellular vesicles on host peripheral blood mononuclear cells

Background: Exagerated immune activation has a key role in the pathogenesis of malaria. During blood-stage infection, Plasmodium falciparum can interact directly with host immune cells through infected red blood cells (PfiRBCs), or indirectly by the release of extracellular vesicles (EVs). Here, we compared the impact of PfiRBCs and P. falciparum small-sized EVs (PfsEVs, also known as exosomes) from a Kenyan clinical isolate (PfKE12) adapted to short-term laboratory culture conditions on host peripheral blood mononuclear cells (PBMC). Methods: PfsEVs were isolated from cell-free culture-conditioned media by ultracentrifugation while mature trophozoite PfiRBCs were purified by magnetic column separation. The PfsEVs and the PfiRBCs were co-cultured for 18 hours with PBMC. Cellular responses were quantified by cell surface expression of activation markers (CD25, CD69) and cytokine/chemokine levels in the supernatant. Results: Relative to negative control conditions, PfsEVs induced CD25 expression on CD4+, CD19+ and CD14+ cells, while PfiRBCs induced on CD19+ and CD14+ cells. Both PfsEVs and PfiRBCs induced CD69 on CD4+, CD8+ and CD19+ cells. In addition, PfiRBCs induced higher expression of CD69 on CD14+ cells. CD69 induced by PfiRBCs on CD4+ and CD19+ cells was significantly higher than that induced by PfsEVs. Secretion of MIP1α, MIP1β, GM-CSF, IL-6, IL-8, and TNFα were significantly induced by both PfsEVs and PfiRBCs whereas MCP-1, IL-10, IL-17α  were preferentially induced by PfsEVs and IP-10 and IFN-γ by PfiRBCs. Prior exposure to malaria (judged by antibodies to schizont extract) was associated with lower monocyte responses to PfsEVs. Conclusions: PfsEVs and PfiRBCs showed differential abilities to induce secretion of IL-17α  and IFN-γ, suggesting that the former are better at inducing Th17, whilst  the latter induce Th1 immune responses respectively. Prior exposure to malaria significantly reduces the ability of PfsEVs to activate monocytes, suggesting immune tolerance to PfsEVs may play a role in naturally acquired  anti-disease immunity.


Introduction
Plasmodium falciparum causes in the region of half a million deaths per year 1 . In the human host, the parasite infects both the liver and red blood cells, but it is the parasite-host interaction during the blood stage that is responsible for pathology. P. falciparum-infected red blood cells can interact with the host cells directly, for example through endothelial protein C receptor (EPCR) to induce inflammation 2 , and also indirectly through secreted parasite factors [3][4][5][6][7] .
One way secreted effector molecules are released from cells is through extracellular vesicles (EVs). EVs are double-layered membrane-bound nanoparticles that are released by cells. They are usually classified into two major sub-groups, small-sized and medium-sized extracellular vesicles 8 often termed as exosomes and microvesicles respectively 9,10 . Small-sized extracellular vesicles (sEVs) are vesicles with a diameter of 30-150 nm. They are generated through inward budding of the limiting membrane of late endosomes resulting in the formation of intraluminal vesicles (ILVs). Late endosomes containing many ILVs are called multivesicular bodies (MVBs) which then fuse with cell membranes and release ILVs into the extracellular space as exosomes or sEVs 8,10 . On the other hand, the diameter of the medium-sized EVs (mEVs) range between 100 and 1000 nm and are formed when part of the cell plasma membrane 'pinches-off' with part of the cytoplasm and surface receptors/proteins and are released into the extracellular milieu 10 .
EVs have the ability to transfer their packaged signaling competent molecules (including proteins, lipids, nucleic acids, and metabolites) from one cell to another, consequently modifying the properties of the recipient target cell(s) 9,[11][12][13] . Research in the pathophysiology of several diseases such as cancer 12,14,15 and infectious diseases 16-19 provide evidence for a role of EVs in mediating intercellular interactions 9,[16][17][18][19][20] . Tumor-and pathogen-derived EVs have been shown to have the ability to abrogate the host immunological defense mechanisms as a way of evading immune responses within the host 19,21-24 .
In the context of malaria, P. falciparum EVs (PfEVs) have a role in intercellular communication as well as in inducing sexual commitment 16,17 . Extracellular vesicles reflect the molecular phenotype of the cells releasing them 12 . Analysis of the impact of host-parasite interactions mediated by PfEVs may improve our understanding of the pathogenesis of severe malaria and the mechanisms through which the parasite modulates the host immune response. Furthermore, proteomic analysis revealed that PfEVs are enriched in parasite proteins involved in interaction with the host cells 16,25 and have been shown to induce inflammation 16 and endothelial activation 26 . However, to date, all studies on the impact of Plasmodium EVs on host immune response have been performed using either rodent malaria 27,28 or using long-term laboratory-adapted P. falciparum isolates 16,29 . Our study is the first to use P. falciparum clinical isolates. The quantity and repertoire of the content of PfEVs from clinical isolates appears to be different from that of long-term laboratory-adapted parasite isolates 25 and this may affect their functional impact on host immune cells. In this study, we compared the functional impact of PfsEVs and the autochthonous PfiRBCs of a clinical isolate adapted to short-term culture (<70 cycles) on human peripheral blood mononuclear cells (PBMC).

Methods
Plasmodium falciparum isolate A Kenyan P. falciparum clinical parasite isolate (unique lab identifier, 9215) was used in this study. This isolate was obtained from a child admitted at Kilifi County Hospital with respiratory distress 30 and was adapted to in vitro culture (<70 cycles) and used in previous studies 31,32 . Isolate 9215 was renamed to PfKE12 in unpublished genome data 33 .
Harvesting and processing of parasite culture media for isolation of PfsEVs PfKE12 was grown under standard culturing conditions 34 . Mycoplasma contamination was routinely monitored using PCR. The parasite cultures were tightly synchronized using D-sorbitol (Sigma) treatment and bulked up to six flasks, each containing 500 µl packed cell volume at 7% parasitemia. These cultures were grown in 40 ml of complete culture media, (RPMI 1640 + L-glutamine + Hepes + D-glucose + gentamicin + sodium-hypoxanthine) (all from Gibco) supplemented with Albumax-II (Gibco) that had previously been depleted of sEVs by ultracentrifugation at 150,000×g for 2 hours. The culture media added at an early ring stage was harvested after 24 hours when the parasites were in the mature trophozoite stage (herein referred to as the rings-to-trophozoite or the RT sample). In this study, the culture-conditioned media (CCM) from the RT sample was harvested as previously described (Abdi et al., 2017) and is shown in Figure 1A. Briefly, the parasite culture was transferred to a 50-ml Falcon tube and centrifuged at 440×g for 5 minutes to pellet down RBCs and the supernatant transferred to a new 50-ml falcon tube. The supernatant (CCM) was then centrifuged once at 440×g for 5 minutes to remove any remaining RBCs and the supernatant transferred into a new 50-ml Falcon tube. This was followed by centrifugation twice at 2,000×g for 10 minutes, once at 3,600×g for 10 minutes and once at 15,000×g for 30 minutes to pellet out mEVs.The mEVs pellet was then resuspended in 1×PBS and stored at -80°C until use (this pellet will be referred to as PfmEVs if from P. falciparum CCM). The resultant supernatant from the final centrifugation at 15,000×g above was filtered through a 0.2-µm filter (FiltoSpurS, Sarstedt) and stored at -80°C until use. Uninfected red blood cells (uRBCs) freshly obtained from a donor or stored for a month at 4°C were also both incubated at 37°C in culture media for either 24 or 48 hours, the CCM harvested and processed as described for the P. falciparum-infected red blood cells cultures. The uRBCs used for CCM harvesting were of the same batch as the one used in maintaining the iRBCs cultures for P. falciparum CCM harvesting.
Isolation of sEVs from culture-conditioned media The 0.2-µm-filtered CCM supernatant stored at -80°C was thawed on ice and transferred under sterile conditions into quick-seal ultracentrifuge tubes (Beckman Coulter cat# 343322) that were then heat-sealed. The sealed tubes were next ultracentrifuged using Optima XE-90 ultracentrifuge in a pre-cooled (4°C) 70.1Ti fixed-angle rotor (Beckman Coulter) at 150,000×g for 2 hours at 4°C ( Figure 1A). This approach is likely to enrich for sEVs, based on their size and the 0.2-µm filter cut-off and henceforth, referred to as PfsEVs. The pellet was then washed twice by re-suspending in ice-cold 1× PBS followed by ultracentrifugation at 150,000×g for 2 hours at 4°C between washes. The final PfsEVs pellet was re-suspended in 400µl of ice-cold 1× PBS, aliquoted and stored at -80°C until use ( Figure 1A). The Bradford protein assay was used to determine the protein concentration in the PfsEVs following the manufacturer's instructions ( Figure 1A). The isolated PfsEVs were also tested for mycoplasma contamination. The processed uRBCs' CCM was similarly subjected to the above isolation process to obtain the equivalent of PfsEVs from uninfected RBCs (uRBCs-sEVs).

Isolation of infected red blood cells (iRBCs)
Plasmodium falciparum-infected RBCs (PfiRBCs) were isolated from the P. falciparum culture using magnetic-activated cell sorting (MACS) columns (Miltenyi Biotec). Tightly synchronized culture at 7% parasitemia was taken through MACS separation columns during the late trophozoite stages. Glycerolyte was added to the MACS-purified PfiRBCs at 3:1vol/vol ratio, stored at -80°C in small aliquots until use. Uninfected RBCs (uRBCs) that were cultured for CCM harvesting (as explained above) were washed twice in incomplete culture media (after CCM harvesting) and stored in appropriate volumes of glycerolyte in -80°C until use.

PBMC isolation
PBMC were obtained from 20 adult Kenyan volunteers. Fresh heparinized whole blood (~20-30 mls) was obtained from each consenting adult donor for plasma and PBMC isolation. The blood was centrifuged at 440×g for 5 minutes to remove plasma that was then stored at -80°C. The cells were topped-up with a wash buffer, R0 (RPMI 1640 supplemented with L-glutamine and Penicillin/Streptomycin [Gibco]) equivalent to 5 volumes of the cell pellet the plasma volume aspirated out. The blood was then layered on Lymphoprep™ (Stemcell Technologies) at a Lymphoprep to blood ratio of 1:2 and centrifuged at 440×g for 20 minutes at room temperature. The PBMC layer between the Lymphoprep and media was aspirated out into a new 50ml Falcon tube and washed twice in R0 at 360×g for 7 minutes and 4°C. The washed PBMC were resuspended in ice-cold freezing medium (10% dimethyl sulfoxide (DMSO) in fetal calf serum (FCS)) and stored overnight at -80°C in Mr. Frosty (Thermo Scientific) before being transferred to liquid nitrogen until use. Prior to storage or use, PBMC number and viability The bar graphs show the crude protein concentration of medium EVs (mEVs) and small EVs(sEVs) isolated from P. falciparum (Pf) infected and uninfected RBC (uRBCs) culture conditioned media; PfmEVs, uRBCs-mEVs, PfsEVs and uRBC-sEVs using Bradford assay. Each bar represent data from three biological replicates. *Significant difference using Mann-Whitney U-Test.
was determined by Trypan blue exclusion using the Vi-CELL XR 2.03 counter (Beckman Coulter, USA) and/or hemocytometer chamber.
PBMC stimulation assays A 1×10 6 PBMC per stimulation condition were co-cultured with PfiRBC or PfsEV in 96-well U-bottomed cell culture plates (Greiner Bio-One).
The stimulation conditions included: PfsEVs added at a determined concentration of 20µg/ml based on consideration from a previous experiment by Mantel et al. 16 , and 1×10 6 MACS-purified iRBCs (PfiRBCs). The positive controls were Staphylococcal enterotoxin B (SEB) at 2.5 µg/ml as a polyclonal activator and CpG-ODN (2.5 µg/ml). Lipopolysaccharide (LPS) at 200 ng/ml was included as an additional positive control for a subset of experiments that involved PBMC from 12 donors. Co-culture wells with (PBMC + cell growth medium only) and (PBMC + 1×10 6 uRBC) were included in each experiment as negative controls. These uRBCs were of the same batch as that used to culture the parasites. The PBMC were co-incubated with the stimulants for 18 hours in a humidified incubator at 37°C and 5% CO 2 . The cells were stained for 30 minutes at 4°C and washed twice before being re-suspended in 300 µl of FACS-flow buffer (BD Biosciences). Cells were acquired on the LSRFortessa™ cell analyzer (BD Biosciences). At least 100,000 events were acquired per stimulation condition and at least 170,000 events for the PfiRBCs and uRBCs conditions since these had an extra cell lysis step prior to staining. Data were analyzed using FlowJo® software version 10.0 (Tree Star).

ELISA and LUMINEX assays
Cell free culture supernatants from the stimulation assays (above) were aspirated and stored at -20°C prior to cytokine analysis. Supernants used for this assay were all from the subset of experiments where LPS was included as an additional positive control condition. Supernants were thawed on ice and used to quantify a total of 29 analytes; IFN-α2, IFN-γ, IL-1α, IL-1β, IL-1ra, IL-2, IL-3, IL-4, IL-5, IL-6, IL-7, IL-10, IL-13, IL-15, IL-17A, TNF-α, TNF-β, GM-CSF, G-CSF, IL-12 (p40), IL-12 (p70) , IL-8, EGF, IP-10 (CXCL10), MCP-1 (CCL2), MIP-1α (CCL3), MIP-1β (CCL4), VEGF, and Eotaxin (CCL11). Analytes were measured using the MILLIPLEX Human Cytokine/Chemokine Magnetic bead 29-Plex assay (catalogue #HCYTMAG-60K-PX29) from Merck-Millipore following the manufacturer's instructions. A total of 25 µl of the culture supernatant was diluted 1:5 times in assay medium and incubated with 25 µl of anti-cytokine antibody-coupled magnetic beads for 2 hours at room temperature while shaking at 500 rpm in the dark. The beads were then washed twice and incubated with 25 µl of biotinylated detector antibody for 1 hour at room temperature, before addition of streptavidin R-phycoerythrin and further incubation for 30 minutes (between each washing step, the beads were retained in the plate using a magnetic separator). After a final wash, beads were re-suspended in 150 µL of LUMINEX Drive Fluid and 100 beads counted for each cytokine in a MAG-PIX reader running on MAGPIX xPOTENT 4.2 software (Luminex Corporation). Analyte concentrations were calculated (via Milliplex Analyst v5.1 [VigeneTech]) from the mean fluorescence intensity expressed in pg/mL using standard curves with known concentrations of each analyte. In addition to the 29 analytes measured using the Luminex platform, TGF-β levels were measured using an ELISA kit (ThermoFisher; cat# BMS249-4) following the manufacturer's protocol.
Anti-schizont antibody ELISA Frozen plasma samples from the PBMC donors were thawed and used in an anti-schizont ELISA to determine prior exposure to malaria. ELISA plates were coated with 100 µl of 1:6000 diluted crude schizont extract/lysate and incubated overnight at 4°C. The plates were then aspirated and blocked with 1% skimmed milk and incubated for 5 hours at room temperature with washes between each step. This was followed by addition and an overnight incubation with 100µl of the 1:1000 diluted plasma, 3 hours incubation at room temperature with 100 µl of HRP-conjugated Rabbit anti-human IgG (Thermo Scientific) and final incubation with 100µl of the o-Phenylenediamine dihydrochloride (OPD) substrate for 15 minutes. The reaction was stopped with 25 µl of 2 M H 2 SO 4 and the plates read at 492 nm on a Synergy 4 (Bio Tek) plate reader, recording the samples' optical densities (OD).

Data analysis
The flow cytometry data from the FlowJo® analysis and ELISA data analyses were performed using Prism 6.01 (GraphPad). Mann-Whitney U-test was used to compare continuous variables between two conditions. The chemokine/cytokine dataset was normalized using Yeo Johnson transformation and t-test was used to compare between two conditions. The unstimulated PBMC cutlture medium (media) was used as the background negative control condition for both PfsEVs and PfiRBCs. For PfiRBCs, relative comprison to uRBCs condition was also shown. To test the impact of prior malaria exposure on PBMC response to PfsEVs and PfiRBCs stimulation, the anti-schizont IgG response in the contemporaneous plasma sample of each PBMC sample was related,using Spearman's rank correlation, to the induced expression of the activation markers on CD4 + , CD8 + , CD19 + , CD14 + cells following stimulation with either PfsEVs or PfiRBCs. Due to small sample size (N=8), the same test was not done for the cytokine/chemokine data. For all tests, P values were considered significant if <0.05.

Ethical statement
Ethical approval was obtained from Kenya Medical Research Institute Scientific and Ethical Review Unit (KEMRI/SERU/ CGMRC/022/3149), and written informed consent was obtained from the PBMC sample donors. The study methods were carried out in accordance with the approved guidelines.

Results
Infected RBCs release a greater quantity of small-sized extracellular vesicles (sEVs) than uninfected RBCs mEVs and sEVs were isolated from culture conditioned media (CCM) of PfiRBCs from a Kenyan clinical isolate (PfKE12) and uRBCs as shown in Figure 1a. We have previously demonstrated successful isolation of sEVs from PfiRBCs CCM by transmission electron microscopy 25 using the protocol described in the methods and schematically represented in Figure 1a. In this study, we used the isolated mEVs and sEVs protein concentration as a proxy for EV abundance 8 . As shown in Figure 1B, the mean protein concentration of mEVs from PfiRBCs CCM (PfmEVs) was ~4.9-fold higher than that of uRBCs (uRBCs-mEVs), as would be expected when parasite proteins are packaged in the mEVs.
The mean protein concentration of the sEV fraction from fresh or aged uRBCs CCM was negligible ( Figure 1B), consistent with previous reports that show mature RBCs primarily release mEVs (microvesicles) but not sEVs (exosomes) 35 .
Both PfiRBCs and PfsEVs induced expression of at least one of the activation markers on T cells, B cells and monocytes PBMC samples from 20 healthy Kenyan adult donors were co-cultured with PfiRBCs and PfsEVs from the same isolate (PfKE12). In addition to the test conditions we included unstimulated control (PBMC' culture medium) and uRBCs as negative controls; and CpG/SEB as positive control conditions. For 12 of the 20 PBMC samples, LPS was also included as an additional positive control condition. Our inability to detect meaningful protein content in the sEV fraction isolated from uRBCs CCM equivalent to the PfsEVs used in this study excluded the possibility of using uRBC-sEVs as a negative control in the PBMC stimulation experiments.
Following co-culturing of the PBMC with the stimulants, the expression of the surface activation markers, CD25 and CD69 on T cells, B cells and monocytes was assessed by flow cytometry and analysed using FlowJo Software v.10.6.2. Extended data, Figure S1 36 shows the FlowJo gating strategy to enumerate different cell populations. Raw flow cytometry files are available as Underlying data 36 .
PfsEVs induced significantly higher expression of CD25 on CD4 + T cells, CD19 + B cells and CD14 + compared to the background media condition (Figure 2A-C). In contrast, the effect of PfiRBCs was apparent only on the antigen-presenting cells, CD19 + B-cells and CD14 + monocytes ( Figure 2B, C) with a significantly higher CD25 expression relative to the background condition (media or uRBCs).
Both PfsEVs and PfiRBCs induced a higher expression of CD69 on CD4 + , CD8 + and CD19 + relative to the background condition ( Figure 2D-F). Additionally, PfiRBCs induced significant CD69 expression on CD14 + cells ( Figure 2G). Notably, PfiRBCs induced a higher CD69 expression on CD4 + and CD19 + cells relative to that induced by PfsEVs ( Figure 2E, F). Taken together, the above result shows that both PfsEVs and PfiRBCs can activate T and B cells and monocytes; but PfsEVs tended to induce higher expression of the activation marker, CD25 while PfiRBCs preferentially induced higher CD69.
To exclude the possibility of low-level Mycoplasma contamination that may be enriched during PfsEVs isolation, PfsEV-DNA extract was used in a PCR-based test. No contamination was detected (Extended data, Figure S2) 36 .
Pfs EVs and PfiRBCs induced common and unique cytokine/chemokine profile For 8 out of the 20 PBMC samples , levels of 29 different cytokines/chemokines in the culture supernatant induced by each of the stimulants were measured using Luminex. Additionally, TGF-β was measured by ELISA. IL-3, IL-4, and IL-5 levels were below the limit of detection with both PfsEVs and PfiR-BCs (Extended data, Figure S3) 36 . The levels of 16 cytokines/ chemokines induced by PfsEVs and by PfiRBCs were not significantly greater than the levels seen in the negative controls using two-sided t-test, however, on one-sided t-test, IL-15 and G-CSF induced by both PfsEVs and PfiRBCs were significantly higher relative to the background conditions (Extended data, Figure S3) 36 . Output ELISA files are available as Underlying data 36 .
Of the remaining 11 cytokines/chemokines, significant induction relative to the negative control background was observed with; a) both PfsEVs and PfiRBCs for MIP1α, MIP1β, GM-CSF, IL-6, IL-8, and TNFα ( Figure 3A-F), b) PfsEVs only for MCP1, IL-10 and IL-17α ( Figure 3G-I) and c) PfiRBCs only for IFNγ and IP-10 ( Figure 3J-L). Notably, the levels of IFNγ and IP-10 induced by PfiRBCs were significantly higher than the levels induced by PfsEVs ( Figure 3J-K), while the concentration of IL-17α induced by PfsEVs tended to be higher than that induced

Figure 2. Proportion of T cells, B-cells and monocytes expressing the activation markers CD25 and CD69 under the different stimulation conditions. (A-G) Proportions of CD4 + T-cells, CD8 + T-cells, CD19 + B-cells, and CD14 + Monocytes expressing the activation markers CD25 (upper panel) and CD69 (lower panel) following PBMC co-culture with
PfsEVs or PfiRBCs. unstimulated PBMC (media) and uRBCs were included as negative control conditions while CpG and SEB were included as positive control conditions. Both PfsEVs and PfiRBCs conditions was compared to the background media condition. PfiRBCs was also compared with uRBCs condition. (*P-value = 0.01-0.05; **P-value = 0.001-0.01; ***P-value = 0.0001-0.001; ****P-value <0.0001; the red horizontal lines indicates the median. P-value was calculated using Mann-Whitney U-test). by PfiRBCs (p=0.07, Figure 3I). Interestingly, in contrast to most cytokines/chemokines, the level of MCP-1 induced by both PfiRBCs and PfsEVs was uniquely higher than that induced by LPS ( Figure 3G). Additionally, the level of IP-10 induced by PfiRBCs was also significantly higher than that induced by LPS ( Figure 3K). Overall, both PfsEVs and PfiRBCs were able to induce various chemokines and cytokines but PfsEVs seems to be better in inducing secretion of IL-17α while PfiRBCs was better in inducing IFNγ and IP-10 suggesting that they might have differential abilities in inducing Th17 and Th1 T cell responses, respectively.
Exposure to malaria tends to tolerise monocytes responses to PfsEVs Given that malaria exposure has been shown to induce immunological tolerance 37-39 , we related the cell surface activation data to each donor's IgG response to crude schizont extract using Spearman's rank correlation. As shown in Figure 4, CD25 expression on CD14 + cells (CD14 + CD25 + ) following stimulation of the PBMC samples with PfsEVs or PfiRBCs decreased with the level of pre-existing anti-schizont IgG in the plasma of each PBMC donor and this association reached significance for PfsEVs (rho=-0.51, p=0.02, N=20). Furthermore, anti-schizont IgG level explained 19% of the variation in CD25 expression on CD14 + monocytes when anti-schizont was used as an explanatory variable in a linear regression model predicting CD14 + CD25 + (coeff(95%CI)=-0.28(-0.52,-0.03), p=0.03, adjusted R 2 =19%, N=20). This result suggest malaria exposure tolerizes the host innate immune response to PfsEVs. The cytokine/chemokine data was not subjected to the same analysis due to small sample size (N=8).

Discussion
In this study we investigated 1) whether PfsEVs can induce PBMC activation in vitro, specifically by measuring cell surface activation markers and cytokines/chemokines secreted into the culture media following co-culture with PBMC; 2) how the PBMC activation induced by PfsEVs compares with that of PfiRBCs; and 3) whether prior exposure to malaria among the PBMC donors influences the level of the induced PBMC activation markers.
We showed that both PfsEVs and PfiRBCs induced T-cells, B-cells, and monocytes to express at least one of the surface activation markers examined relative to the negative control condition. PfsEVs showed relatively stronger induction of CD25 expression while PfiRBCs preferentially induced CD69 expression, particularly on B-cells. At the cytokine/chemokine level, both induced secretion of several cytokines/chemokines but they also showed differential ability to induce secretion of some cytokines. Notably, PfsEVs induced secretion of higher levels of IL-17α relative to the background media and tended to be higher than that induced by PfiRBCs ( Figure 3I). IL-17α is known to be secreted by Th17 CD4+ T cells and the cytokines, IL-6 and TGF-β have been shown to be able to induce differentiation of naïve CD4 T-cells into Th17 cells in vitro 40 . Interestingly, PfsEVs also induced significantly higher levels of IL-6 ( Figure 3E) but not TGF-β (Extended data, Figure S3) 36 . On the other hand, PfiRBCs showed superior ability to induce secretion of IFNγ and IP-10 relative to the background uRBCs and even to PfsEVs condition when co-cultured with PBMC ( Figure 3J-K. IP-10 secretion is driven by a pro-inflammatory cytokine milieu including IFN-γ 41 ; thus, it is plausible that high levels of IFN-γ led to the IP-10 secretion that we identified in response to PfiRBCs stimulation.
Hence our interpretation is that PfsEVs and PfiRBCs can induce differentiation of CD4+ T cells into Th17 and Th1 cells respectively. In a recent study 42 , PfiRBCs were shown to induce NK cells to secrete IFNγ while PfmEVs could not, which is consistent with the observation made in this study with PfsEVs and PfiRBCs co-cultured with PBMC. This finding warrants further characterization of the T cell subsets activated by PfsEVs, including γδ T-cells known to be activated by phosphoantigens in P. falciparum culture medium 43 .
We observed PfiRBCs could induce significantly higher MCP-1 (CCL2) and IP-10 (CXCL10) compared to LPS when co-cultured with PBMC ( Figure 3C, K). This result is consistent with observation previously made with co-culturing of PfiRBCs with purified dendritic cells 44 . However, in the previous study 44 , co-culture of PfiRBCs with purified dendritic cells could not induce secretion of inflammatory cytokines such as TNF-α and IFN-γ 44 . This difference might be explained by the presence of cells such as T cells in PBMC, which could be the source of the TNF-α and IFN-γ.
Previously, persistent exposure to malaria infection have been shown to tolerizes T and B cell response to malaria antigens 38,39 . In this study we showed, albeit with a small sample size, that the degree of previous malaria exposure (determined by the level of IgG response to crude schizont extract in the plasma of each PBMC donor) was negatively associated with monocyte's response to PfsEVs (Figure 4). PfsEV interaction with monocytes in vitro have been shown to induce inflammatory response 26 , potentially contributing to malaria pathogenesis. Therefore tolerance to PfsEVs following frequent malaria infection may be part of the naturally acquired anti-disease immunity 37 .
We showed that uninfected red blood cells (uRBCs) do not produce sEVs (exosomes) containing quantifiable amount of proteins using a Bradford assay. During the isolations, we used culture conditioned media (CCM) from varied uRBCs ranging from fresh (processed for culture within 1 hour after phlebotomy) to 2 weeks old cultures. By contrast, the mEV fraction from uRBC CCM repeatedly contained quantifiable amount of proteins. This might indicate that uRBCs release sEVs, but with very low levels of packaged proteins; however, it could alternatively mean that uRBCs primarily release mEVs (microvesicles). The latter interpretation is consistent with a previous study that showed developing red blood cells release sEVs during earlier stage of haematopoiesis, but mature RBCs do not 35 , as they have lost the endocytic pathway that is essential for the biogenesis of sEVs. Other studies have described exosomes from uRBCs 29 , but this discrepancy might be explained by a difference in methodology as our protocol involves a filtration step at 0.2 µm that excludes the majority of mEVs from the sEV fraction, a step that was omitted from the earlier reported work.
While we used multiple PBMC donors to generate these results, this study used only one P. falciparum isolate and, therefore, we cannot conclusively determine if the results we obtained will remain similar if the number of isolates was increased for diversity. Different parasites have been shown to display a difference in virulence with studies demonstrating that this virulence can be transferred to non-virulent parasite phenotypes via secreted EVs 18,19 . Drawing analogy from these experiments, it would be interesting to see if the phenomena we observe are consistent across isolates of P. falciparum with different levels of virulence, and whether any differences can be correlated with differences in protein or RNA content within the sEVs. More functional, proteomic and transcriptomic analysis of the PfEVs is clearly needed. File 'supplementary_material_PfEV_PBMC_paper.pdf' contains the following extended data:
• Figure S2: Gel electrophoresis image after PCR to test for Mycoplasma contamination in P. falciparum cultures and isolated PfEVs.
• Figure S3: Cytokines/chemokines not significantly induced following PBMC co-culture with PfsEVs or PfiRBCs . Another issue: did not the authors find odd that they could not isolate sEVs from uRBCs given the reports that show how they obtain them, even after filtration with a 0.2 um membrane? (check Correa et al, 2019 2 ). Why did they not use silver stain to visualize the proteins in uRBCs as they did with iRBCs or sEVs, but rather chose the much less sensitive Bradford stain for that visualization? Would they have found sEVs from uRBCs had they stained with silver? I believe so. expertise to confirm that it is of an acceptable scientific standard, however I have significant reservations, as outlined above.
Reviewer Report 10 November 2020 https://doi.org/10.21956/wellcomeopenres.17707.r41182 © 2020 Mbagwu S. This is an open access peer review report distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

Smart Ikechukwu Mbagwu
Anatomy Unit, Department of Oncology, Microbiology and Immunology, Faculty of Science and Medicine, University of Fribourg, Fribourg, Germany I have read the article titled Impact of Plasmodium falciparum small-sized extracellular vesicles on host peripheral blood mononuclear cells. The key findings of this work is that there is differential induction of immune response on peripheral blood mononuclear cells by extracellular vesicles derived from malaria infected red blood cells and the plasmodium falciparum infected red blood cells. This differential immune response induced is exerted by the Th17 and Th1 actions which is yet to be fully characterized. In addition the paper details the decline in the capacity of the extracellular vesicles to activate monocytes.
The introduction of the paper was well written with relevant references cited. However, the opening sentence seems to be missing some texts. The authors need to check this out. The methods used were well described but since the aim of the work was to compare the functional impact of PfsEVs and the autochthonous PfiRBCs of a clinical isolate adapted to shortterm culture (<70 cycles) on human peripheral blood mononuclear cells (PBMC). Were there no other methods that could have been employed in the study eg rt-qPCR? I am not sure it is clear what the authors meant by "probably enriched".
The results are clear but it could be more interesting if the authors could demonstrate why infected RBCs release more extracellular vesicles.
The discussion provides a summary of the findings of the study. However, the authors did not provide a clarity on why the difference in the induction of immune response as shown by the level of expression of the activation markers was observed and a justification for the observation. The significance of their observation is not clearly detailed and how this can be provide an insight to an in vivo condition. Do the authors think that different isolates of the Plasmodium falciparum could give a different observation from their study findings?

Is the work clearly and accurately presented and does it cite the current literature? Yes
Is the study design appropriate and is the work technically sound? Yes