Plasmodium vivax infection compromises reticulocyte stability

The structural integrity of the host red blood cell (RBC) is crucial for propagation of Plasmodium spp. during the disease-causing blood stage of malaria infection. To assess the stability of Plasmodium vivax-infected reticulocytes, we developed a flow cytometry-based assay to measure osmotic stability within characteristically heterogeneous reticulocyte and P. vivax-infected samples. We find that erythroid osmotic stability decreases during erythropoiesis and reticulocyte maturation. Of enucleated RBCs, young reticulocytes which are preferentially infected by P. vivax, are the most osmotically stable. P. vivax infection however decreases reticulocyte stability to levels close to those of RBC disorders that cause hemolytic anemia, and to a significantly greater degree than P. falciparum destabilizes normocytes. Finally, we find that P. vivax new permeability pathways contribute to the decreased osmotic stability of infected-reticulocytes. These results reveal a vulnerability of P. vivax-infected reticulocytes that could be manipulated to allow in vitro culture and develop novel therapeutics.

T he unique biconcave shape and structural dynamics of red blood cells (RBCs) permit traversal of the circulatory system and the delivery of oxygen to tissues. RBC structural integrity and deformability is determined by the molecular and cellular functions of the RBC cytoskeleton and membrane transporters 1 . Characterizing RBC stability under osmotic stress is a powerful method for probing the structural properties of the cytoskeleton and membrane transporter activity. An example of the relationship between RBC structural integrity and osmotic stability is seen with naturally occurring polymorphisms in molecules such as PIEZO1 and Ankyrin-1, that disrupt cell volume regulation and cytoskeleton, respectively, and are associated with the clinically relevant RBC phenotypes of hereditary xerocytosis (HX) 2 and hereditary spherocytosis (HS) 3 . These mutations manifest in structurally compromised RBCs with abnormal shapes (stomatocytes and spherocytes), premature hemolysis in vivo 4 and exhibit abnormal sensitivity to hypotonic challenge, which is decreased in HX and increased in HS.
During the Plasmodium spp. intra-erythrocytic developmental cycle (IDC), parasite remodeling alters the physiology of the host RBC. In the case of the most well studied of these parasites, P. falciparum, changes include disruption of the cytoskeleton by parasite proteins inserted into the RBC membrane 5 and increased plasma membrane permeability 6,7 that culminate in the infected RBC taking on a spherical shape 8 . These changes compromise the structural integrity of the host RBC, and accordingly decrease the osmotic stability of P. falciparum-infected RBCs 6,7,9 . The consequence of the reduced stability of P. falciparum-infected RBCs is an increased propensity of infected cells to be destroyed in circulation via clearance by the spleen as well as intravascular hemolysis 10 .
Plasmodium vivax, the most globally widespread of the malaria parasites infecting humans 11 , exhibits a strict tropism for reticulocytes [12][13][14] , the youngest of circulating RBCs. This is in contrast to P. falciparum, which infects both reticulocytes and mature normocytes. All P. falciparum osmotic stability studies have been done in the older, more abundant normocyte fraction. Furthermore, recent observations that P. vivax increases reticulocyte deformability 13,15,16 , while P. falciparum decreases normocyte deformability 5,15,17 suggest that the pathophysiology of P. vivax reticulocyte infection is fundamentally different than P. falciparum normocyte infection. The absence of an in vitro culture system and appropriate experimental methods have impeded the study of how P. vivax impacts the structural stability of the host reticulocyte. Moreover, though reticulocytes [18][19][20] and erythroid progenitors 21 are known to be more resistant to hypotonic lysis than normocytes, the dynamics of osmotic stability during erythropoiesis and reticulocyte maturation are unknown. Understanding the impact of P. vivax infection on the structural integrity of the host reticulocyte has the potential to provide important insight into why a long-term in vitro culture system for P. vivax has yet to be established, as well as the in vivo fitness of P. vivax-infected reticulocytes.
Osmotic stability is traditionally assessed by measuring release of hemoglobin from bulk RBC preparations in increasingly hypotonic solutions 22 . This methodology cannot resolve the osmotic stability of specific RBC subpopulations within characteristically heterogeneous reticulocyte and P. vivax samples. Furthermore, using free hemoglobin as the readout for RBC osmotic stability is problematic as hemoglobin levels are in flux in both parasite-infected RBCs and uninfected reticulocytes, because Plasmodium spp. digest hemoglobin as they mature 23 , and reticulocytes are actively synthesizing hemoglobin 24 . Flow cytometry, with its ability to examine discrete cell subsets within heterogeneous populations, has the potential to better define RBC osmotic stability. Finally, though previous studies have demonstrated the capacity of flow cytometry to detect hemolysis [25][26][27][28] , the current strategies have not been used to examine the lysis of discrete RBC populations and are furthermore not amendable to collecting sufficient data on rare cell populations.
In this work, we have developed a flow cytometric osmotic stability assay to assess the osmotic stability dynamics within inherently heterogeneous reticulocyte and malaria-infected RBC populations. We observe that osmotic stability steadily decreases during erythroid differentiation and reticulocyte maturation. Of enucleated RBCs, the youngest CD71+ reticulocytes which P. vivax preferentially invades using the CD71/transferrin receptor molecule as an invasion ligand 13,14,29 , are the most osmotically stable. Examination of clinical P. vivax samples showed P. vivax infection destabilized the host reticulocyte to levels close to those observed for RBCs from individuals with hemolytic anemia. Furthermore, P. vivax-infected reticulocytes are significantly less stable than P. falciparum-infected normocytes. Finally, we find that the decreased stability of P. vivax-infected reticulocytes corresponded with the appearance of P. vivax new permeability pathways (NPPs). The observation that reticulocyte osmotic stability is reduced by P. vivax infection suggests that P. vivaxinfected reticulocytes may be prone to pre-mature hemolysis in vivo. These results reveal a key vulnerability of the parasite that if corrected may facilitate the adaptation of P. vivax to in vitro culture, and conversely if targeted could yield novel therapeutics to treat blood stage P. vivax infection.

Results
Flow cytometry analysis of RBC osmotic stability. While flow cytometry has been employed to study osmotic stability, previous studies have not harnessed the capacity of flow cytometry to analyze multiple cell populations within a single specimen [25][26][27][28] . Therefore, we developed a flow cytometric osmotic stability assay that yields the same hemolysis curves as the traditional hemoglobin absorbance assay and that by quantifying the individual cells that survive hemolysis challenge is able to assess osmotic stability dynamics within heterogeneous RBC populations (Fig. 1a). The analysis of hemolysis by flow cytometry revealed typical RBC forward scatter (FSC) and side scatter (SSC) profiles for non-lytic isotonic conditions, and the subsequent appearance of a low-FSC/low-SSC population in hypotonic, lytic conditions ( Fig. 1b and Supplementary Fig. 1).
To confirm that the low-FSC/low-SSC population were RBC ghosts, fluorescent phalloidin, which is excluded from intact cells but binds actin in the cytoskeleton of permeabilized cells was included in the osmotic stability assay. We found that phalloidin stained the FSC-low population that appears in lytic conditions, indicating that the flow lysis assay is sensitive to RBC ghosts ( Fig. 1c and d). Considering only intact RBCs, the lysis curves produced by the flow lysis assay were indistinguishable from those of the hemoglobin absorbance lysis assay (Fig. 1e). No difference in the lysis 50 values (mOsm at which 50% of RBCs lyse) obtained for normal RBCs using the two lysis assays confirmed the accuracy of the flow lysis assay. To test the limits of the flow lysis assay, we measured the osmotic stability of HX RBCs that are resistant to hypotonic lysis. With HX RBCs as well, we observed no difference in lysis 50 values generated by the two assays (Fig. 1f).
Finally, as P. vivax osmotic stability studies included cryopreserved samples and were performed on RBCs maintained in vitro, we examined the effect of (i) cold storage and (ii) P. vivax in vitro culture conditions on RBC osmotic stability. As expected 30,31 , we observed variation in the osmotic stability of normal RBCs taken from 12 different donors with lysis 50 values ARTICLE NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-021-21886-x ranging from 93.9 to 137.9 mOsm (variation of 11.2%) (Fig. 1g). Subsequent storage of RBCs from six different donors at 4°C for 2 weeks resulted in no appreciable change (mean slope −0.063 ± 0.24 SEM, Pearson r = 0.15) in osmotic stability (Fig. 1h). However, when cryopreserved RBCs and RBCs stored at 4°C were transferred to P. vivax in vitro culture conditions, osmotic stability increased by 13 ±2.6% and 14.6 ± 0.8% after 24 h of incubation (Fig. 1I).
Osmotic stability decreases during RBC differentiation. Having established the capacity of flow cytometry to measure RBC osmotic stability, we next examined the osmotic stability dynamics within the RBC fractions that harbor P. vivax infection. To this end, we took advantage of the capacity of flow cytometry to track discrete cell populations, and separately assess the osmotic stability of nucleated RBC precursors, reticulocytes, and normocytes in bone marrow aspirates ( Fig. 2a and Supplementary 2A). We found that fluorescent labeling of reticulocytes and nucleated RBC precursors present within bone marrow aspirates allowed us to quantitate the osmotic stability dynamics of each of these RBC subpopulations simultaneously (Fig. 2b). For nucleated RBC precursors we  additionally found that lysed cells were also identifiable with a live/dead stain (Supplementary 2B). Importantly, as both osmotic stability studies with bone marrow aspirates and clinical P. vivax samples relied on Percoll enrichment to raise reticulocytemia and P. vivax parasitemia to reliably measurable levels, we found that Percoll had no effect on the osmotic stability of reticulocytes and nucleated precursors from bone marrow aspirates (Supplementary 2C). The lysis 50 values obtained for RBC precursors and reticulocytes from bone marrow aspirates revealed that nucleated RBC precursors (DNA+ CD71+) were the most osmotically stable (Lysis 50 70.1 ± 5.8) followed by the youngest CD71+ RNA+ DNA− reticulocytes (Lysis 50 87.3 ± 8.0), older CD71− RNA+ DNA− reticulocytes (Lysis 50 107.4 ± 7.3) and CD71− RNA− DNA− normocytes (Lysis 50 114.5 ± 5.4) ( Fig. 2b and c). To establish osmotic stability dynamics during erythropoiesis, we assessed the osmotic stability of erythroid progenitors differentiated in vitro from CD34+ stem cells 19,32 . This study revealed a strong trend of decreasing osmotic stability as cells progressed in vitro from majority basophilic and polychromatic normoblasts at day 9 (Lysis 50 73.8 ± 3.6) to a majority polychromatic and orthochromatic normoblast population at day 11 (Lysis 50 102.7 ± 15.6), and then further decreased as cells progressed through final orthochromatic normoblast maturation occurring between day 14 (Lysis 50 102.2 ± 2.0) day 17 (Lysis 50 126.2 ± 11.2), and day 20 (Lysis 50 143.4 ± 15.1) (Fig. 2d). Parallel examination of the osmotic stability of CD71+ reticulocytes differentiated in vitro showed no significant difference in osmotic stability as a function of day reticulocytes were sampled ( Supplementary Fig. 2D). Furthermore, consistent with a previous study 19 , CD71+ reticulocytes generated in vitro (Lysis 50 110.1 ± 3.7) were less stable than CD71+ reticulocytes from bone marrow samples (Lysis 50 87.3 ± 8.0) ( Supplementary Fig. 2E). Together these results demonstrate the capacity of flow cytometry to assess the osmotic stability of discrete RBC subsets within heterogeneous populations and clearly shows that erythroid development and reticulocyte maturation are associated with significant changes in osmotic stability.  c Normocyte (purple data points), old reticulocytes (green data points), young reticulocytes (blue data points), and RBC precursors (orange data points) from bone marrow aspirates lysis 50 values (n = 6). Unique symbol indicate biological replicates. Horizontal lines and error bars represent mean ± SEM. Asterisks denote significant differences in lysis 50 values (p = 0.0001) using RM one-way ANOVA analysis. d RBC progenitor lysis 50 values during course of CD34+ differentiation in vitro (n = 2). Unique symbols indicate RBC progenitor lysis 50 values at days 9, 11, 14, 17, and 20 from two independent differentiations. n.s. indicates no significant difference in lysis 50 during course of differentiation (p = 0.2) using Friedman test. Inset are representative photos of RBC progenitors during in vitro differentiation. Scale bar, 10 μm. Colored bars below images represent proportion of RBC progenitor developmental stages (gray -proeythroblast, green-basophilic normoblast, blue-polychromatic normoblast, purple-orthochromatic normoblast, red-reticulocyte) present at day-9, -11, -14, -17, and -20 of differentiation.
P. vivax infection reduces reticulocyte osmotic stability. Having established that the youngest CD71+ reticulocytes that are preferred by P. vivax for invasion 13,14 are the most osmotically stable of all enucleated RBCs, we next assessed the impact of P. vivax infection on reticulocyte osmotic stability. In the absence of continuous P. vivax in vitro culture, cryopreserved clinical P. vivax samples are an invaluable resource for investigating P. vivax biology [33][34][35][36] . Cognizant of the decreased stability of cryopreserved RBCs, however (Fig. 1i), we first directly compared the in vitro survival and osmotic stability of cryopreserved and noncryopreserved P. vivax clinical samples as parasites progressed through the IDC ( Fig. 3a and b and Supplementary Fig. 3A). Consistent with previous studies 33-36 , we observed a 78 ± 3.3% and 63 ± 30.8% loss of P. vivax-infected reticulocytes prior to completion of the IDC in vitro (44-h cultures) for Brazilian cryopreserved and Indian non-cryopreserved clinical samples, respectively (Fig. 3c). Parallel osmotic stability measurements of the uninfected RBCs in P. vivax clinical samples revealed that, as observed previously (Fig. 1I), the stability of cryopreserved uninfected RBC populations increased upon transfer to in vitro culture while non-cryopreserved uninfected RBCs remained steady during the course of culture (Supplementary Fig. 3B and C). For P. vivax-infected reticulocytes, we observed that cryopreserved stage I ring-infected reticulocytes were less stable than non-cryopreserved stage I rings. For subsequent time points we observed (i) no difference in the osmotic stability of cryopreserved and non-cryopreserved P. vivax-infected reticulocytes and (ii) a decrease in the stability of P. vivax-infected reticulocytes as they progressed through the IDC (Fig. 3d). Moreover, when we consider the clinical laboratory cutoff for normal RBCs (lysis 50 154 mOsm or 0.45% NaCl) 31,37,38 , the osmotic stability of 88% of P. vivax-infected reticulocytes (cryopreserved and non-cryopreserved) in 24-and 44-h cultures fell into the range of osmotic instability associated with RBC disorders that cause hemolytic anemia such as HS.
We next examined the degree of instability P. vivax induced in the host reticulocyte by comparing the osmotic stability of a.   Fig. 3D and 3E) is longer than previously observed for CD71+ cord blood reticulocytes invaded in vitro by P. vivax 13 . The reason for this discrepancy is not immediately evident and therefore subject for future investigation. We found infected reticulocytes were less stable than uninfected CD71+ reticulocytes at all-time points assessed (8-, 16-, and 24-hour cultures), and the progression of P. vivax through the IDC further decreased reticulocyte osmotic stability with the appearance of stage III late trophozoite forms in 24-h cultures corresponding to P. vivax-infected CD71+ and CD71-reticulocytes being 71.4 ± 12.0% and 74.0 ± 14.6% less stable than uninfected CD71+ reticulocytes (Fig. 3d). Finally, no difference in the stability of CD71+ and CD71− infected-reticulocytes indicated that P. vivax infection and not reticulocyte age is the primary determinate of the stability of P. vivax-infected reticulocytes ( Supplementary Fig. 3F).
P. vivax-infected reticulocyte are less stable than P. falciparuminfected normocytes. Considering previous observations that P. vivax and P. falciparum-infected RBCs exhibit different rheological properties 5,13,15-17 , we next compared P. vivax destabilization of reticulocytes with P. falciparum destabilization of normocytes. To account for the influence of cryopreservation on our P. vivax osmotic stability studies, we assessed the in vitro survival and osmotic stability of cryopreserved P. falciparum clone 3D7 P2G12 40 . We observed similar progression of P. vivax and P. falciparum through the asexual IDC but a greater frequency of sexual gametocyte forms for P. vivax. In vitro survival, however, was markedly different, with a 77.3 ± 6.5% survival rate observed for P. falciparum infected-normocytes at 44-h of culture compared to a 22.4 ± 0.03 survival rate for P. vivax-infected reticulocytes (Figs. 3c and 4a). We subsequently assessed the osmotic stability of uninfected and P. falciparum trophozoite and schizonts stage-infected normocytes in 24-and 44-h cultures (the points at which P. vivax-infected reticulocytes were most destabilized). This analysis revealed no difference in the osmotic stability of P. falciparum infected (majority trophozoite) and uninfected-normocytes in 24h cultures and a reduction in P. falciparum infected-normocyte (majority schizonts) stability of 23.6 ± 6.9% compared to uninfected normocytes in 44-h cultures (Fig. 4b). This is in contrast to P. vivax, which had decreased reticulocyte stability by 74.0 ± 14.6% by the time parasites had matured to the trophozoite form in 24-h cultures. Furthermore, direct comparison of the osmotic stability of P. falciparum-infected normocytes and P. vivaxinfected reticulocytes revealed that P. vivax-infected reticulocytes (max Lysis 50 184.1 ± 8.3, 24-h culture) are significantly less stable, p < 0.0003, than P. falciparum-infected normocytes (max Lysis 50 107.8 ± 4.7, 44-h culture) (Fig. 4c).
New permeability pathways increase P. vivax-infected reticulocyte permeability. New permeability pathways (NPPs) in related malaria parasites, P. falciparum and P. knowlesi, increase the permeability of the infected RBC to certain solutes 6,7 . To determine whether P. vivax possesses NPPs that are contributing to the decreased stability of the host reticulocyte, we assessed the sensitivity of Percoll-enriched cryopreserved Brazilian clinical P. vivax samples to the NPP antagonists D-sorbitol and L-alanine. For P. vivax-infected reticulocytes, we found that early stage parasites present in 8-h cultures were resistant to D-sorbitol and Lalanine lysis, while the appearance of stage III late trophozoite parasites in 16-h cultures corresponded with P. vivax-infected reticulocytes lysing in isotonic D-sorbitol (16-h-47.9 ± 9.0%, 24-h-45.2 ± 9.2%, and 44-h-57.5 ± 7.7%) and L-alanine solutions (16-h-60.7 ± 10.5%, 24-h-61.0 ± 9.2%, and 44-h-55.8 ± 2.3%). Finally, the NPP inhibitor, furosemide 41 , protected P. vivax-infected reticulocytes from D-sorbitol and L-alanine lysis ( Fig. 5a and b). The incomplete lysis of P. vivax-infected cells by D-sorbitol and L-alanine along with variation in the sensitivity of different P. vivax clinical isolates to D-sorbitol and L-alanine lysis are potentially driven by (i) the high frequency of more stable gametocytes (42.4 ± 9.5% of P. vivax-infected reticulocytes in  42,43 (Fig. 3c), (ii) variable NPP activity in different P. vivax isolates, or (iii) P. vivax being less sensitive to D-sorbitol lysis than P. falciparum 44,45 .

Discussion
Structural stability of P. vivax-infected reticulocytes, critical for the successful propagation of blood-stage malaria infections, has not been thoroughly assessed due to lack of appropriate methodology compatible with P. vivax's unique and challenging biology. To study the impact of P. vivax infection on the structural integrity of the host reticulocyte, we adapted the osmotic stability assay to be analyzed by flow cytometry. This permitted us to work with the limited cell numbers and heterogeneity characteristic of both the reticulocyte compartment to which P. vivax infection is restricted as well as P. vivax-infected clinical samples. In the proceeding studies, we observed that the youngest CD71+ reticulocytes, which P. vivax preferentially invades, were the most osmotically stable of enucleated RBC subsets and that P. vivax-infected reticulocytes were less stable than uninfected CD71+ reticulocytes. These results indicate that P. vivax severely compromises the structural integrity of the host reticulocyte. In addition to its use in determining the osmotic stability of erythroid precursor and malaria infected-RBC samples, the single-cell resolution of the flow cytometry osmotic stability assay provides a means for tracking the impact of erythropoiesis perturbations on overall RBC osmotic stability and examining osmotic stability dynamics within heterozygous hematological diseases such as sickle cell and G6PD deficiency.
The decreasing osmotic stability we observed in the later stages of erythropoiesis and then continuing through reticulocyte maturation is likely driven by three major changes occurring to erythroid cells during their differentiation: (i) intracellular solutes changes due to changing membrane transporter abundance and activity 46,47 , (ii) cytoskeleton remodeling 48,49 , and (iii) the progressive reduction in the surface to volume ratio driven by membrane loss 50 . The observation by us and others 18-20 that reticulocytes are more resistant to osmotic lysis than normocytes is seemingly incongruent with previous reports of membrane instability in reticulocytes 51 . This discrepancy is likely due to either (i) the factors that determine osmotic stability, like higher surface area to volume ratio and greater abundance and activity of membrane transporters in reticulocytes, are more dominant than reticulocyte membrane instability, or (ii) the biology underlying membrane instability and osmotic stability are independent of one another. Ultimately, the factors that determine the stability of erythroid precursors and a. b.
c. reticulocytes are likely multifactorial and highly dynamic. Finally, we and others 19 have found that in vitro differentiated reticulocytes are less stable than ex vivo reticulocytes. Interestingly the addition of cholesterol improved the stability of in vitro cultured reticulocytes 19 , suggesting that approaches pursuing the determinants of erythroid progenitor and reticulocyte osmotic stability may yield strategies for generating more stable and consequently more viable RBCs from in vitro cultures. Our study revealed that P. vivax infection decreased the osmotic stability of the host reticulocyte. Though consistent with what has been observed for other Plasmodium spp. 6,7,9,52,53 , P. vivax notably decreased reticulocyte osmotic stability to a significantly greater degree than the more extensively studied P. falciparum decreased normocyte osmotic stability. Variation in the instability induced by Plasmodium spp. parasites has also been observed for P. knowlesiinfected normocytes, that are similarly less stable than P. falciparum-infected normocytes 44 . Comparing this data set with ours however, we estimate P. vivax-infected reticulocytes are the least stable of the three Plasmodium spp., being 70% less stable than P. falciparum-infected normocytes whereas P. knowlesi-infected normocytes are 30% less stable than P. falciparum-infected normocytes. Moreover, P. vivax-infected reticulocyte instability is on par with RBC instability observed for hemolytic anemias like HS. Together these findings identify a new parameter of malaria blood stage pathophysiology that differs between Plasmodium spp. that has the potential to reveal avenues to species-specific therapeutics.
The observation that P. vivax compromises the osmotic stability of the host reticulocyte is consistent with the observation that the majority of late stage P. vivax-infected reticulocytes lysed when passed through 2 μm microfluidic channels 16 , and suggest that the instability induced by P. vivax significantly increases the risk of premature hemolysis of infected reticulocytes. The destabilization of reticulocytes by P. vivax infection is also consistent with the precipitous loss of P. vivax-infected reticulocytes from in vitro culture observed in this as well as previous studies [33][34][35][36] . That said, as P. vivax is clearly capable of propagating its blood stage infection in vivo, P. vivax likely has strategies for coping with the compromised structural integrity of the host reticulocyte. Future studies will provide important insight into (i) the in vivo consequences of the compromised structural integrity of infected-reticulocytes and how P. vivax is able to circumvent the reduced fitness of its host cell to propagate successful blood stage infections and (ii) strategies for stabilizing P. vivax-infected reticulocytes in vitro to help facilitate culture adaptation of P. vivax.
Intriguingly, clinical observations that P. vivax infection results in 4 times higher clearance of uninfected RBCs than P. falciparum 54,55 suggests that P. vivax infection not only more severely compromises the host reticulocyte it infects but surrounding RBCs as well. As a consequence of the global destabilization of host RBCs, the frequency and severity of P. vivax associated anemia is on par with P. falciparum 55-57 associated anemia despite the reticulocyte tropism of P. vivax resulting in a lower parasite biomass than P. falciparum infections 58 . This anemia is likely the consequence of intravascular hemolysis, opsonophagocytosis driven extravascular hemolysis and dyserythropoiesis 59,60 . The degree to which each of these contributes to P. vivax associated anemia remain to be determined.
In our study we found that P. vivax NPPs are associated with reduced stability of host reticulocytes. This raises several questions including: (i) are P. vivax NPPs required for the acquisition of essential nutrients from the host serum and (ii) what host and or parasite transporters are responsible for the observed NPP activity. The presence of NPP activity in P. vivax-infected reticulocytes suggests some overlap between the underlying factors responsible for the reduced osmotic stability of P. falciparum and P. vivax-infected cells. In P. falciparum, Clag3.1 and Clag3.2 have been identified as parasite mediators of NPP activity in infected RBCs 61,62 , and P. vivax possesses a clag family of genes 63 . This raises the possibility of therapeutically targeting P. vivax-infected reticulocytes by inhibiting Clag function, as has been done for P. falciparum 64,65 .
NPPs are unlikely to be the single variable contributing to the reduced stability of P. vivax-infected reticulocytes. Previously reported shedding of microvesicles from P. vivax-infected reticulcoytes 66 and 100 nm "holes" in P. vivax-infected reticulocytes 13 could each reduce osmotic stability by decreasing surface to volume ratio and increasing permeability of P. vivaxinfected reticulocytes. Determining the biology underlying the compromised structural integrity of infected reticulocytes has the potential to yield targets against which therapeutics could be developed. Finally, considering the compromised stability of P. vivax-infected reticulocytes in the context of the protection from Plasmodium infection afforded by RBC polymorphisms such as HS, thalassemia, and G6PD deficiency, it is possible that a mechanism underlying protection from P. vivax may be premature hemolysis of infected RBCs.
Together our data support a model in which P. vivax referred CD71+ reticulocytes are more osmotically stable than CD71− reticulocytes and normocytes, but, upon P. vivax infection and subsequent maturation, the infected reticulocyte undergoes a precipitous loss of osmotic stability. Moreover, the magnitude of instability induced by P. vivax is close to the instability observed for hemolytic anemias like HS. Additionally, the onset of NPP activity corresponds with the decreasing osmotic stability of infected reticulocytes (Fig. 5c). As a result, P. vivax-infected reticulocytes likely exhibit increased rates of intravascular hemolysis and premature clearance from circulation similar to what is observed for RBCs from individuals with hemolytic anemias. Future studies will focus on the mechanisms underlying these changes in osmotic stability that impact the structural integrity of the host reticulocyte and the consequences of this instability on in vivo survival. Finally, identifying strategies for stabilizing P. vivax-infected reticulocyte could prove key to culture-adapting P. vivax, while strategies to further destabilize P. vivax-infected reticulocytes could conversely yield novel blood stage therapies for P. vivax.

Methods
Ethics approval. For HX, bone marrow aspirate, and Brazilian and Indian clinical P. vivax samples, we confirm that all relevant ethical regulations were complied with. For HX and human bone marrow samples negative for blasts or diserythropoietic conditions, informed consent was obtained for all patients and samples were collected under the approval of the Boston Children's Hospital Institutional Review Board (IRB04-02-017R) and were de-identified for use in this study. We obtained human IRB waivers for the use of the de- Bone marrow and parasite enrichment. Bone marrow aspirates, cryopreserved Brazilian clinical P. vivax samples and P. falciparum 3D7 P2G12 were enriched on 1.080 g/mL KCl high Percoll gradients 33 . Briefly, 2 mL of re-suspended cells (up to 25% hematocrit) were layered on 3 mL 1.080 g/mL KCl high Percoll gradient and spun for 15 min at 1200×g. Subsequently, the interface was removed, washed, and applied to assays. For P. falciparum 3D7 P2G12 cultures, the interface and pellet were washed and recombined before being applied to assays. CD34+ in vitro RBC differentiation. CD34+ hematopoetic stem cells (HSCs) purchased commercially (Lonza) were differentiated in vitro following the established three-step differentiation protocol 19,32 in Iscove's liquid medium (Biochrom) with 5% human plasma (Octapharma). Giemsa-stained cytospins were prepared and a minimum 200 cells per slide (×1000 magnification) were called as either proerythroblast, basophilic normoblast, polychromatic normoblast, orthochromatic normoblast, or reticulocyte. Images were taken at ×1000 magnification using an Excelis HD Camera attached to an Olympus BX40 microscope.
P. vivax and P. falciparum in vitro culture. Indian and Brazilian clinical P. vivax samples and P. falciparum 3D7 P2G12 were cultured at 100 × 10 6 cells per mL in IMDM (Gibco) containing 10% AB + heat-inactivated sera and 50 μg/mL gentamicin at 37°C in 5% CO 2, 1% O 2 , and 94% N 2 33 . Hemacolor-stained cytospins were prepared and 200 cells per slide (×1000 magnification) were called as either asexual stage (I-V) or gametocyte 67 . Images of cytospins were taken at ×1000 magnification using an Excelis HD Camera attached to an Olympus BX40 microscope. Osmotic stability assays for non-cryopreserved clinical Indian P. vivax samples were performed locally at Goa Medical College where samples were collected while cryopreserved clinical Brazilian P. vivax samples were shipped from São Paulo Brazil to Boston, MA where they were thawed and osmotic stability assays were performed. falciparum samples at 37°C for 20 min. Prior to analysis by flow cytometry, AccuCheck counting beads (Thermo Fisher Scientific) were added to each sample. All flow cytometry experiments were acquired on a Miltenyi MACSQuant instrument equipped with MACSQuantify version 2.11 and 405, 488, and 638-nm lasers and a minimum of 200 events were collected for each cell population being analyzed for the no lysis control sample. All subsequent gating and analysis was based on the no-lysis control sample. As lysis in subsequent samples resulted in ultimate destruction of the cell populations, this minimum event count was not able to be achieved for all samples. For all conditions analyzed, a minimum of 2000 events were collected for cell counting beads. Samples with parasitemia and or reticulocytemia below 0.05% were considered below the limit of detection. Data were analyzed using FlowJo (Version 10.4).
Osmotic stability assays. Phosphate buffer saline lysis solutions 22 ranging from 300 to 0 mOsm (distilled water) were made by serial dilution and final osmolarity measured by vapor pressure osmometer (Wescor Vapro 5520). For flow cytometric osmotic stability assays, 1 × 10 6 cells (stained with appropriate fluorescent dyes and or antibodies) were incubated at 5 × 10 6 cells/mL in lysis solutions ranging from 300 to 0 mOsm (distilled water) for 10 min at 37°C. Lysis was stopped with 4× volume of quenching solution (1:10 AccuCheck Counting Beads (Thermo Fisher Scientific) to flow buffer). Only cells with normal flow cytometric FSC/SSC profiles (as defined by the control 300 mOsm condition) were considered to be intact. For assays containing nucleated cells, nuclei and dead cells were additionally excluded from the intact cell population. The ratio of beads to intact cells was used to calculate the absolute number of cells remaining in each lytic condition and percent lysis was determined by normalization to the 300 mOsm control condition (0% lysis). For hemoglobin absorbance osmotic stability assays, cells were lysed as described above before 4× volume of flow buffer was added to stop lysis. Cells were pelleted and the absorbance (380, 415, 450, and 540 nm) of supernatants measured on a Spectramax M5 plate reader. The amount of hemoglobin present in supernatants was calculated by the Harboe method 68 . Percent lysis was determined by normalization to the 300 mOsm control condition (0% lysis).
RBC ghost assay. For RBC ghost detection by flow cytometry, FITC conjugated phalloidin was added to all lysis and quenching solutions. For RBC ghost detection by immunofluorescence microscopy, RBCs suspended in PBS containing FITC-Phalloidin were allowed to settle on a coverslip before a hypotonic solution (57 mOsm) containing FITC-Phalloidin was added and RBC lysis and ghost formation examined by live video microscopy on a Zeiss AxioObserver.Z1. Images were taken at ×630 magnification on a Zeiss AxioObserver.Z1 inverted fluorescent microscope coupled to an AxioCam MRm camera. Images were processed using Fiji ImageJ 1.52p software.
Sorbitol and alanine hemolysis assay. P. vivax cultures were incubated at 5 × 10 6 cells/mL for 30 min at 37°C in flow buffer, 280 mM sorbitol with 1% BSA, or 280 mM alanine with 1% BSA in the presence or absence of 100 μM furosemide. Following incubation, cells were washed twice with flow buffer, and stained for flow cytometry analysis.
Statistics and reproducibility. Lysis 50 values derived from least-squares regression fit curves of normalized data were used to assess the reproducibility of results from biological replicates. Biological replicates shown for (i) normal RBCs, (ii) HX RBCs, (iii) bone marrow aspirates, (iv) CD34+ erythroid culture, and (v) Brazilian and Indian P. vivax samples are biologically independent samples, and experiments were performed in batches with 1-3 samples being assessed in a given experiment. For P. falciparum 3D7 (P2G12 clone) biological replicates are five independently cryopreserved P. falciparum 3D7 (P2G12 clone) samples thawed sequentially so that experiments were performed independently. Prism (GraphPad Prism8) and Stata (version 16.0, StataCorp LLC., College Station, TX, USA) were used to test for statistical significance.
Reporting summary. Further information on research design is available in the Nature Research Reporting Summary linked to this article.

Data availability
All data generated or analyzed during this study are included in this published article (and its Supplementary information files). A reporting summary for this article is available as a Supplementary information file. The P. falciparum 3D7-P2G12 clone used in this study will be made available upon request. All additional samples were clinical ex vivo samples for which we do not have additional material to supply. Source data are provided with this paper.