Severe adult malaria is associated with specific PNAS PLUS PfEMP1 adhesion types and high parasite biomass

Maria Bernabeua, Samuel A. Danzigera, Marion Avrila, Marina Vazb, Prasad H. Babarc,d, Andrew J. Braziera, Thurston Herricksc,d, Jennifer N. Makic,d, Ligia Pereirac,d, Anjali Mascarenhasc,d, Edwin Gomesb, Laura Cheryc,d, John D. Aitchisona, Pradipsinh K. Rathodc,d, and Joseph D. Smitha,1

aCenter for Infectious Disease Research, Seattle, WA 98109; bDepartment of Medicine, Goa Medical College & Hospital, Bambolim, Goa 403202, India; cDepartment of Chemistry, University of Washington, Seattle, WA 98195; and dDepartment of Global Health, University of Washington, Seattle, WA 98195

Edited by Mats Wahlgren, Karolinska Institutet, Stockholm, Sweden, and accepted by the Editorial Board April 7, 2016 (received for review December 10, 2015) The interplay between cellular and molecular determinants that This subset of PfEMP1s includes mediators of distinct infected lead to severe malaria in adults is unexplored. Here, we analyzed erythrocyte adhesion categories, including “rosetting” and endo- parasite virulence factors in an infected adult population in India thelial protein C receptor (EPCR) binding. Rosetting involves ad- and investigated whether severe malaria isolates impair endothe- hesion to uninfected red blood cells (20, 21), possibly leading to lial protein C receptor (EPCR), a protein involved in coagulation greater microvascular obstruction (22). EPCR binding involves in- and endothelial barrier permeability. Severe malaria isolates overex- fected erythrocyte adhesion to vascular endothelial cells (23). The pressed specific members of the Plasmodium falciparum var gene/ important role of the EPCR-activated protein C (APC) pathway in P. falciparum PfEMP1 ( erythrocyte membrane protein 1) family that regulating coagulation, inflammation, and endothelial barrier prop- var bind EPCR, including DC8 genes that have previously been linked erties (24) has led to the hypothesis that EPCR-binding parasites to severe pediatric malaria. Machine learning analysis revealed that may drive pathogenic mechanisms by inhibiting the APC–EPCR DC6- and DC8-encoding var transcripts in combination with high par- interaction (23, 25–28), thus increasing vascular dysfunction and asite biomass were the strongest indicators of patient hospitalization permeability. Indeed, cerebral swelling is a major risk factor for and disease severity. We found that DC8 CIDRα1 domains from pediatric death (29) and there is loss of EPCR and fibrin depositions severe malaria isolates had substantial differences in EPCR binding MICROBIOLOGY affinity and blockade activity for its ligand activated protein C. at sites of cerebral sequestration in pediatric autopsies (30). How- Additionally, even a low level of inhibition exhibited by domains ever, the extent to which severe malaria isolates disrupt EPCR from two cerebral malaria isolates was sufficient to interfere with function is poorly understood. A better understanding of adhesion- activated protein C-barrier protective activities in human brain based pathogenic mechanisms may inform novel targeted adjunctive endothelial cells. Our findings demonstrate an interplay between drug therapies to improve patient survival and outcomes. parasite biomass and specific PfEMP1 adhesion types in the devel- Another factor that determines malaria disease severity is total opment of adult severe malaria, and indicate that low impairment parasite burden. Plasma levels of P. falciparum histidine rich of EPCR function may contribute to parasite virulence. Significance malaria | Plasmodium falciparum | var | PfEMP1 | EPCR The clinical presentation of severe malaria differs between evere malaria caused by Plasmodium falciparum is responsi- children and adults, but the factors leading to these differences Sble for at least 400,000 deaths every year (1), mainly affecting remain poorly understood. Here, we investigated parasite vir- children younger than 5 y old. However, in areas of low and un- ulence factors in adult patients in India and show that specific stable transmission, severe malaria affects both children and adults endothelial protein C receptor (EPCR)-binding parasites are (2), although disease symptomatology varies according to patient age. associated with severe adult malaria and act together with Whereas severe anemia, metabolic acidosis, and cerebral malaria are parasite biomass in patient hospitalization and disease sever- the major severe syndromes in children, multisystem disease is more ity. We found substantial differences in EPCR binding activity common in adults, including renal impairment, jaundice, respiratory from severe malaria isolates. However, even parasite domains distress, metabolic acidosis, and cerebral malaria (3, 4). In addition, that partially obstructed the interaction between EPCR and its disease mortality sharply increases with the age of the patient (4). ligand activated protein C were sufficient to interfere with The factors that drive age-related differences are unknown. activated protein C-barrier protective activities in human brain A central process in severe falciparum pathology is the seques- endothelial cells. Thus, restoration of EPCR functions may be a tration of infected erythrocytes to microvascular endothelial cells key target for adjunctive malaria drug treatments. (5). Extensive tissue-specific sequestration results in organ pathol- ogy, such as cerebral malaria and placental malaria, and contributes Author contributions: M.B., S.A.D., M.A., and J.D.S. designed research; M.B., M.A., P.H.B., A.J.B., and T.H. performed research; M.V., J.N.M., L.P., A.M., E.G., L.C., and P.K.R. contrib- to metabolic acidosis and endothelial dysfunction (6, 7). Proteins uted new reagents/analytic tools; M.B., S.A.D., M.A., J.D.A., and J.D.S. analyzed data; M.B., of the P. falciparum erythrocyte membrane protein 1 (PfEMP1) S.A.D., and J.D.S. wrote the paper; M.V. and E.G. conducted patient enrollment and data family, encoded by the var genes, are responsible for infected management; J.N.M., A.M., L.C., and P.K.R. designed the clinical study; J.N.M., L.P., A.M., – and L.C. designed the clinical study data management; and P.K.R. was responsible for the red blood cell binding to the microvasculature (8 10). PfEMP1s study site. are classified into three main groups—A, B, and C—basedonup- The authors declare no conflict of interest. stream sequence (UpsA, UpsB, UpsC) and chromosome location This article is a PNAS Direct Submission. M.W. is a guest editor invited by the Editorial (11). The extracellular domain of PfEMP1s presents a modu- Board. lar structure composed of adhesion domains, called Duffy binding- Freely available online through the PNAS open access option. like (DBL) and cysteine-rich interdomain region (CIDR) (12), Data deposition: The sequences reported in this paper have been deposited in the Gen- which sometimes can be found in conserved tandem arrange- Bank database (accession nos. KU843600–KU843604). mentsknownasdomaincassettes(DC)(13).Expressionof 1To whom correspondence should be addressed. Email: [email protected]. – group A PfEMP1 variants (14 18) and PfEMP1 encoding DC8 This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10. (15, 19) have been strongly linked with pediatric severe malaria. 1073/pnas.1524294113/-/DCSupplemental.

www.pnas.org/cgi/doi/10.1073/pnas.1524294113 PNAS Early Edition | 1of10 Downloaded by guest on September 27, 2021 protein 2 (PfHRP2), a surrogate of parasite biomass, can predict target groups A, B, or C (VarA, UpsB1, UpsB2, UpsC1, and disease severity and fatality rates in both children and adults (31, UpsC2) (16) and 40 degenerate primers that target specific ad- 32), the probability of disease deterioration (33), retinopathy-posi- hesion domains (15). Overall, patients in the study presented tive cerebral malaria (34), and whether a fever is caused by malaria a complex population of parasites that transcribed a mixture of (35). Nevertheless, a recent longitudinal study in Tanzanian children A, B, and C var genes (Fig. 2A and Table 2). However, the showed that high PfHRP2 levels did not necessitate severe disease median VarA transcript level was higher than groups B and C in (36), suggesting severe disease requires additional factors. all patient groups, and SM patients had significantly elevated In this study, statistical and machine-learning approaches were VarA and UpsB1 transcripts in comparison with OP (Fig. 2A and used to explore the relationship between PfEMP1 expression, Table 2). parasite biomass, and disease severity in adults with P. falciparum To gain further insight into parasite binding phenotypes as- infections treated at the Goa Medical College, and we investigated sociated with adult SM, domain-specific primers were used to whether severe malaria isolates impair the APC–EPCR pathway. identify adhesion subtypes expressed in patients, also taking advantage of the functional specialization of PfEMP1 proteins Results to infer parasite binding traits. In particular, the N-terminal Characteristics of the Study Population. A total of 59 P. falciparum- PfEMP1 head structure (DBL–CIDR tandem) has diversified infected patients from the Goa Medical College were enrolled in between group A (EPCR binding or rosetting) and groups B and the study. Among them, 26 patients had severe malaria (SM) and C (CD36 binders). Head structures containing CIDRα1 subtypes presented at least one WHO SM criterion (37), 13 patients were bind EPCR, CIDRβ/γ/δ subtypes are associated with rosetting, hospitalized but did not have any severity criterion for SM and CIDRα2–6 subtypes bind CD36 (Fig. S1) (reviewed in ref. (moderately severe malaria, MSM), and 20 were outpatients (OP). 38). Using in silico analysis, we predicted var genes that would be As shown in Table 1, the three groups did not present significant amplified by the 40 domain-specific primers in seven annotated differences in age. Previous studies have found a border between parasite genomes (13) and assigned a predicted binding pheno- pediatric and adult disease symptomatology at 11 y old (4), so four type to each gene (EPCR, CD36, or rosetting) depending on its teenagers were included in this study. Among SM patients, 77% head structure (Fig. S2). Some primers target var domains that presented more than one severity criteria and 57.7% had at least are associated with more than one type of head structure (15). three different severity signs, indicative of multisystem disorders. Using these criteria, we inferred the CD36, EPCR, and rosetting Because peripheral parasitaemia does not reflect the seques- binding phenotype for each primer (Table 2 and Tables S1 and S2). tered parasite population, we measured plasma concentration Compared with outpatients, the SM group presented signifi- of PfHRP2 as a surrogate of total parasite biomass (Fig. 1). SM cantly higher transcript levels of var genes with an expected patients had significantly higher plasma PfHRP2 levels than EPCR-binding phenotype (DC8: DBLα-CIDRα, CIDRα1.1, both MSM (P = 0.004) and OP (P < 0.0001) (Fig. 1A). Fur- DBLβ12 and DBLβ3/5, DBLγ4/6; EPCR binders: CIDRα1.4 and thermore, PfHRP2 concentration increased significantly with CIDRα1) or domains associated with EPCR, rosetting, or CD36 the number of severe criteria and organ dysfunction (Spearman’s binding PfEMP1 (DBLγ of DC6) (Fig. 2 B and C and Table 2). ρ = 0.67, P < 0.0001) (Fig. 1B). The elevated VarA transcription in SM (Fig. 2A) could be of either EPCR or rosetting variants. However, a comparison of SM Isolates Overexpressed EPCR-Binding var Transcripts. To identify domain-specific primers suggested EPCR-binding head structure var genes associated with adult severe disease, we performed domains (DBLα1.1/2/4/7, CIDRα1.4, and CIDRα1) were increased quantitative RT-PCR (qRT-PCR) using a set of 5 primers that in SM patients (Table 2). By comparison, the primers detected few

Table 1. Clinical characteristics of the patients Patient characteristic SM (n = 26) MSM* (n = 13) OP (n = 20)

Age (mean; IQR), y 33; 22–41 33; 22–47 26; 19–28 Minimum–maximum 18–63 15–62 15–55 Male (%) 88.5 84.6 93.3 P. falciparum and P. vivax coinfection 0 1 0 † Parasite density (median), parasites per microliter 41,410 27,103 51,352 IQR 17,898–93,296 0–47,879 31,811–130,536 Hemoglobin (mean ± SD) (g/dL) 9.9 ± 3.1 11.8 ± 2.4 12.2 ± 2 Glasgow coma score <10 (%) 26.9 0 NA Glasgow coma score (median) 14 14 NA Minimum–maximum 3–15 12–15 NA Respiratory distress (%) 57.7 0 NA Respiratory rate (mean ± SD), breaths per minute 26.6 ± 9.3 18 ± 0.1 NA Jaundice (%) 69.2 0 NA Acute renal failure (%) 46.2 0 NA Shock (%) 11.5 0 NA Acidosis (%) 34.6 0 NA Days in the hospital (median) 7 4 NA Minimum–maximum 2–31 3–9NA ‡ Degree of severity (mean) 3.2 NA NA Percentiles 25 and 75 1.75–5NANA Mortality (%) 15.4 0 0

*Patients requiring admission to Goa Medical College who did not fulfill any of the WHO criteria for SM classification. † SM (n = 22), MSM (n = 10), OP (n = 12). ‡ Number of WHO SM criteria. IQR, interquartile range; NA, not applicable.

2of10 | www.pnas.org/cgi/doi/10.1073/pnas.1524294113 Bernabeu et al. Downloaded by guest on September 27, 2021 PNAS PLUS

Fig. 1. Association between PfHRP2 plasma concentrations and disease severity. PfHRP2 levels were compared between patient groups and by number of WHO SM criteria. (A) PfHRP2 plasma concentrations among disease groups. Horizontal lines indicate median for each group. Pairwise comparisons were analyzed using the Mann–Whitney U test. Significant higher concentration is represented by **P < 0.01 and ***P < 0.001. (B) Spearman’s rank correlation coefficient (ρ) and P value for the association between PfHRP2 plasma concentrations and number of severity criteria.

differences in CD36-binding variants between groups. Transcripts multitude of decision trees and measures the mean decrease for two CD36-binding domains were significantly increased in the in classifier accuracy (MDCA) when a particular feature (e.g., SM group (CIDRα3.1–3, DBLα0.1) and for one in the OP group transcript level detected by a var primer) is removed from the (DBLe2ofDC7),butotherCD36-binding transcripts did not differ model. Recent advances in computational techniques have made between patient groups (Table 2). it possible to associate P values with features ranked in this The MSM group presented a very similar transcription profile manner (40, 41). First, we used a RF (39) with 1,000,000 trees to to the SM group and none of the transcripts presented a signifi- select parasite features that could be used to accurately predict cant differential expression after correcting for multiple compar- patient hospitalization and disease severity. Although transcrip-

isons [false-discovery rate (FDR) ≤ 0.2] (Table S1). Furthermore, tion of domains DBLα-CIDRα of DC8 and DBLγ of DC6 were MICROBIOLOGY similar differences in var expression were found when both groups the strongest factors to predict patient hospitalization (SM + were combined in a hospitalized group and compared with OP MSM vs. OP), the level of PfHRP2 was the most important (Table S2). Transcripts from DC8 (DBLα-CIDRα,CIDRα1.1, feature for patient severity (SM vs. OP) (Fig. 3A). Moreover, the β β γ α DBL 12 and DBL 3/5, DBL 4/6) and EPCR binders (CIDR 1.4 analysis suggested that transcription of both DC8 and DC6 do- α α α and CIDR 1, DBL 2/ 1.1/2/4/7) presented a higher expression mains plays a role in patient severity as they ranked second and in hospitalized patients. In addition, expression of transcripts third in the prediction of SM [mProbes familywise error rates γ amplified by primers DBL of DC6 (EPCR, CD36, rosetting) and (FWER) ≤ 0.2]. In addition, all four primers that target the DC8 γ DBL of DC5 (rosetting) was increased among hospitalized pa- domains ranked in the top 10 of the severity model, highlighting tients (Table S2). Taken together, these data show that a higher the importance of DC8-containing PfEMP1 in disease severity. prevalence of predicted EPCR binding and rosetting parasites was PfEMP1 variants could contribute to disease severity by dis- detected by multiple primers targeting Group A, DC8, DC6, and tinct mechanisms, including by promoting the rapid multiplica- DC5 PfEMP1 variants in severe and hospitalized patients (sum- tion of parasites (biomass) or by encoding dangerous adhesion marized in Table S3). traits (impairing critical endothelial functions). Univariate cor- relation analysis showed an association between PfHRP2 plasma Var Transcript Abundance, Parasite Biomass, and Clinical Disease. Plasma levels and transcription of domain DBLγ of DC6. To investigate PfHRP2 concentration presents a close correlation with disease severity in both children and adults (31, 32). To better understand the possibility that the correlation of the DC6 domain in disease the interplay between parasite biomass and parasite adhesion severity was simply because of an association with parasite bio- types in adult SM, we performed correlation analysis. Across all mass, we applied conditional mutual information (CMI) algo- rithms. CMI is an information theory technique to test if one patients, there was a positive correlation between PfHRP2 plasma ’ concentration and transcript abundance of three var domains: variable s predictive power is weakened by the presence of oth- DBLγ of DC5 (rosetting) (Spearman’s ρ = 0.3, P = 0.04), DBLγ of ers. After PfHRP2 filtration (i.e., filtering out those primers that DC6 (EPCR, rosetting, and CD36) (Spearman’s ρ = 0.33, P = 0.02), were uninformative after PfHRP2 levels were accounted for), γ α and CIDRα3.1–3(CD36)(ρ = 0.32, P = 0.03), suggesting that transcription of domains DBL of DC6 and CIDR 1.1 of DC8 parasites expressing these variants might lead to higher parasite remained important to disease severity and patient hospitalization P ≤ load. In addition, elevated transcription of domains DBLα-CIDRα (CMI PfHRP2 filtration 0.05), showing that the virulence of of DC8 (ρ = 0.3, P = 0.04) and DBLγ of DC6 (ρ = 0.46, P = 0.001) these DCs is not simply explained by a major contribution in par- was correlated with the number of severity criteria. Conversely, asite biomass (Fig. 3A). transcripts targeted by primer DBLe2ofDC7(CD36)(ρ = −0.33, To provide a visualization of the logic used by decision trees, P = 0.03) and DBLe12 of DC12 (CD36) (ρ = −0.31, P = 0.03) were we generated evolutionary decision trees to understand how negatively associated with PfHRP2. Taking these data together, this parasite factors might interact to determine disease severity in analysis suggests that specific PfEMP1 domains may contribute to adults (Fig. 3B). In the tree that tests patient hospitalization higher parasite biomass, whereas other PfEMP1 domains may act in (MSM + SM vs. OP), a combination of high PfHRP2 levels and concert with parasite biomass to increase disease severity when they elevated transcription of domains CIDRα1.1 of DC8 and DBLγ pass a certain threshold. of DC6 were sufficient to correctly classify 91% of patients as To test this hypothesis, we built machine-learning models to hospitalized or not. Furthermore, a combination of these three investigate disease causation and performed statistical analysis. features was sufficient to classify 100% of SM patients (SM vs. Multivariate, threshold-based logic maps nicely on to a decision OP). Taken together, machine-learning approaches determined tree structure, and random forests (RF) are powerful tools for that both parasite biomass and transcription of DC8- and DC6- ranking feature importance (39). This methodology generates a containing PfEMP1 variants are critical for adult malaria severity.

Bernabeu et al. PNAS Early Edition | 3of10 Downloaded by guest on September 27, 2021 The machine-learning analysis implicated both an N-terminal PfEMP1 domain (DC8 cassette: DBLα2-CIDRα1.1/1.8-DBLβ12- DBLγ4/6) and a C-terminal PfEMP1 domain (DC6 cassette: DBLγ14-DBLζ5-DBLe4) in severe disease. DC6 can be present in combination with all four types of PfEMP1 head structures: group A rosetting, group A EPCR-binding, DC8 EPCR-binding, or CD36 binding (Fig. S1) (13). To investigate a possible struc- tural linkage between DC8 and DC6 in SM patients, we per- formed correlation analysis. As expected, the four DC8 domains were highly correlated (Spearman’s ρ = 0.51–0.71 for all pairwise comparisons). By comparison, the DBLγ of DC6 domain was equally correlated to other DC6 domains (ρ = 0.34) and DC8 domains and group A variants (ρ > 0.3) (Fig. S3). These results suggest that DC8 and DC6 may be linked in the same PfEMP1, but DC6 may also be present in non-DC8 PfEMP1. Thus, it is possible that different subsets of DC6-containing PfEMP1 vari- ants may contribute to SM. Many of the var primers match different parts of the same DC (e.g., DC8) or share similar functional annotations (e.g., EPCR binding). Consequently, var primers with the same annotation may artificially lower MDCA values in the RF analysis by act- ing as proxies for each other. To account for this effect, set- enrichments were performed to combine primers targeting the same var group or adhesion types (Fig. S2 and Table S4). For this analysis, a Mann–Whitney U test was used to compare the MDCA values of set-enriched primer groups versus primers not targeting that annotation (SI Materials and Methods). In both the patient hospitalization and disease-severity models, different combinations of DC8 primers (DC8 pure and DC8 all) presented enrichment P values lower than 0.005 that remained significant after FDR correction (FDR ≤ 0.2) (Fig. 3C). The association with patient hospitalization and disease severity was still signif- icant but reduced in the group that contained all EPCR binders (DC8 EPCR + group A EPCR). Despite playing an important role in the univariate analysis, VarA transcriptional levels were not necessary to predict hospitalization or disease severity. An enrichment group that excluded DC8 and incorporated both types of UpsA adhesion traits (group A = EPCR and rosetting variants combined) failed to reach significance, reinforcing the importance of DC8. Rosetting, IgM, or intercellular adhesion molecule-1 (ICAM-1) binding domains groups, which have been previously associated with SM (reviewed in ref. 42) did not present any significance in disease predictability of adult SM. As expected, the CD36 binding group showed no significance for predicting either hospitalization or severity. Taken together, these data show that machine-learning approaches highlight the importance of DC8 in adult SM, and determine that in combi- nation with the DBLγ of DC6 domain and elevated parasite biomass, is strongly associated with disease severity.

DC8 CIDRα1 Domains Expressed by SM Isolates Inhibit the APC–EPCR Interaction. DC8 CIDRα1 differ in sequence, binding affinity, and ability to inhibit the APC–EPCR interaction (25, 27, 28), but there have been no in-depth functional characterizations of DC8 CIDRα1 from SM isolates. To perform a deeper phenotypic characterization of DC8 variants in severe isolates, we amplified α Fig. 2. Transcription of UpsA, DC6, and DC8 var is elevated in SM pa- and sequenced the full-length DC8 CIDR 1 domain (Fig. S4) tients. The transcript abundances of var gene subtypes were investigated from five patients who met WHO SM criteria: patient 24 had among patients. (A) Transcript levels of A, B, and C var gene groups and anemia, patient 25 had jaundice, and patients 62, 87, and 95 (B) domain subtypes of DC8 and DC6 in SM and OP groups. Horizontal presented multiorgan complications, including cerebral malaria lines indicate median for each group. Differences among groups were in patients 62 and 87. In total, seven CIDRα transcripts were compared by using the Mann–Whitney U test. Significant higher tran- amplified from the five patients. All of the sequences clustered ≤ ≤ ≤ scription is represented by *P 0.05 and FDR 0.2, ***P 0.005 and together with domains from subgroups CIDRα1.1 and CIDRα1.8, FDR ≤ 0.05. (C) Heat map showing transcription levels of DC8 and DC6 domain A subtypes and VarA genes. Maximum transcription levels are represented in which are diagnostic for DC8-like (Fig. 4 ). All five patients pre- sented a single DC8 sequence except for patient 62, who presented red, minimum transcription in blue, and median transcription levels α in white. Color equivalents were set by comparing each primer tran- three distinct DC8 CIDR 1 domains. Notably, an identical DC8 script among all patients analyzed. FDR, Benjamini–Hochberg adjust- CIDRα1 transcript was amplified from three patients with multi- ment for false discovery rate. organ complications, including both cerebral malaria patients

4of10 | www.pnas.org/cgi/doi/10.1073/pnas.1524294113 Bernabeu et al. Downloaded by guest on September 27, 2021 Table 2. Transcript levels of var domain subtypes in severe patients and outpatients PNAS PLUS † Binding phenotype* Group Primer SM (IQR) n = 24 OP (IQR) n = 19 P value FDR

Unknown binding A DBLα1.1 of DC1 9.2 (0-13.3) 2.8 (0-10.2) 0.23 0.36 A DBLe8 of DC3 6.8 (0.4–36.1) 2.1 (0.1–17.6) 0.27 0.38 EPCR A DBLα2/α1.1/2/4/7 67.2 (19.5–271.3) 33 (3.7–152.7) 0.06 0.18 A DBLα1.4 0.1 (0-1.8) 0.1 (0-2) 0.11 0.28 A CIDRα1.6 0 (0-0.1) 0 (0-0.1) 0.34 0.40 A CIDRα1.7 2.5 (0-9.6) 0 (0-17.6) 0.28 0.38 A DBLβ3 1.5 (0-20.4) 0.2 (0-7.1) 0.14 0.30 A DBLα1.7 of DC13 0.5 (0.1–23) 0.2 (0.1–3.8) 0.2 0.35 A CIDRα1.4 of DC13 0.9 (0-2.1) 0.5 (0-3.1) 0.32 0.40 A CIDRα1.4 and CIDRα1 2.9 (0.2–22.8) 0 (0-3.6) 0.006 0.11 B/A DBLα-CIDRα of DC8 4.5 (1.1–9.6) 0.1 (0-3.1) 0.02 0.13 B/A CIDRα1.1 of DC8 9.5 (2.8–22) 2 (0.2–11) 0.04 0.16 B/A,A DBLβ12 and DBLβ3/5 45.3 (6.2–89.8) 11.8 (0.3–36.3) 0.01 0.11 B/A DBLγ4/6 of DC8 7.5 (2.6–30.6) 1.8 (0.6–12.8) 0.03 0.16 EPCR, CD36, and rosetting B (A,C) DBLγ of DC6 19.7 (4.1–66.7) 1.4 (0.3–3.1) 0.0003 0.01 B (A,C) DBLζ5 of DC6 5 (0.2–15.6) 0.9 (0.1–5.4) 0.1 0.25 B DBLζ4 of DC9 20.9 (2.2–50.6) 20.1 (2.3–60.8) 0.49 0.50 A,B,C DBLγ9 0 (0-0) 0 (0-0) 0.46 0.49 CD36 B (C) DBLe2 of DC7 1.2 (0.2–3.5) 4.3 (0.4–23) 0.04 0.16 B (A,C) DBLe3 of DC7 1.9 (0.2–9.4) 3.2 (0.1–6.3) 0.5 0.50 B DBLγ of DC9 8.7 (0.4–76.5) 24.8 (1.2–71.5) 0.18 0.35 B DBLζ6 of DC10 1.5 (0.3–10.6) 1.2 (0.2–4) 0.18 0.35 B DBLα0.16 of DC19 4.9 (0.5–32.4) 1 (0.4–22.3) 0.12 0.28 B,C CIDRα3.4 of DC19 21.8 (8.2–57.8) 10.9 (4.6–35) 0.06 0.18 α – B DBL 0.9 of DC20 7.3 (1-29) 2.9 (0.5 10.6) 0.13 0.29 MICROBIOLOGY B DBLα0.1 1(0.3–4.6) 0.4 (0.3–1.3) 0.05 0.18 B DBLα0.6/9 39.3 (10.6–139.1) 42.5 (1.2–93.1) 0.25 0.38 B CIDRα2.2 1.2 (0-40.5) 5.3 (0-48.5) 0.32 0.40 B CIDRα2.3/5/6/7/9/10 30.7 (9.9–97.7) 38.3 (9.1–57.7) 0.41 0.46 B,C CIDRα3.1–3 6.7 (1.6–28) 2.2 (0.6–7.7) 0.02 0.13 B CIDRγ2/9 0.4 (0-3.9) 0.2 (0-4.2) 0.45 0.49 B (A,C) DBLβ5 0 (0-4.3) 0 (0-0.3) 0.23 0.36 B,C CIDRγ 82.9 (34-140.8) 138.1 (24-315.2) 0.21 0.35 CD36 and rosetting B,A DBLe12 of DC12 1.9 (0-15.3) 3 (0.2–9.5) 0.33 0.40 B CIDRγ1/2 6.5 (1-30.2) 4.4 (0.6–19.8) 0.32 0.40 Rosetting A DBLγ of DC5 0 (0-0.2) 0 (0-0) 0.06 0.18 A DBLβ7 & 9 of DC5 0 (0-0) 0 (0-0.1) 0.19 0.35 A DBLα1.5/6b of DC16 0.6 (0.1–1.7) 0.6 (0.1–4.4) 0.21 0.35 A DBLα1.5/6a of DC16 3.1 (0.3–20.1) 8.2 (0.4–16.5) 0.28 0.38 A CIDRδ of DC16 1.5 (0.1–13) 2.9 (0.2–14.7) 0.37 0.43 VarA 118.9 (27.7–298) 39.8 (3.6–112.2) 0.02 0.13 UpsB1 36 (18-68.4) 21.4 (5.7–47.2) 0.04 0.16 UpsB2 18.8 (8.1–30) 13.5 (10.5–19) 0.09 0.24 UpsC1 5.4 (3.3–18.8) 9.8 (3-15.6) 0.49 0.50 UpsC2 4.7 (2.4–17.1) 2.2 (1.2–7.8) 0.08 0.23

Median Tu level and IQR of var subtypes in SM and OP groups. P values comparing the SM and OP group were calculated using a one-tailed Mann–Whitney U test. *Predicted binding phenotype of PfEMP1 head structure. † Benjamini–Hochberg adjustment for FDR. Differential expression of transcripts with P ≤ 0.05 and FDR ≤ 0.2 is represented in boldface and considered significant.

(62 and 87). By comparison, the five distinct CIDRα1 sequences inhibition of APC binding to CHO745-EPCR cells coincubated had 32% sequence identity. with either the negative-control PFE1640w CIDRα1.3 domain or To investigate whether SM isolates interfere with the APC– the weak binder domain r62-2-22 (Fig. 4B). By comparison, two EPCR interaction, the five unique Indian CIDRα1 domains were CIDR domains from Indian isolates (r25-2-4 and r62-2-1) pre- expressed as recombinant proteins (Fig. S4 and Table S5). Four sented low inhibition of APC binding and the two remaining of five bound EPCR with low to moderate nanomolar binding CIDR domains from Indian isolates (r62-2-23 and r24-2-4) constants (Kds = 3.81 nM to 63 nM) (Fig. 4B and Fig. S5), similar presented moderate APC inhibition (Fig. 4B). However, all of in range to other CIDRα1.1/8 domains (27, 29). Although r62-2-22 the DC8 domains expressed by SM isolates competed less effec- CIDRα1.1 did not bind recombinant EPCR, this domain exhibited tively for the APC–EPCR interaction than a positive control, group dose-dependent binding to Chinese hamster ovary (CHO) cells A variant (IT4var07 CIDRα1.4 domain), which strongly blocks the expressing EPCR in a flow-cytometry binding assay (Fig. S5), APC–EPCR interaction (25, 26, 28). suggesting it also possesses low EPCR binding activity. As To study if SM isolates interfere with the endothelial bar- expected, in binding competition studies, there was limited or no rier protective activity of APC, we investigated the ability of

Bernabeu et al. PNAS Early Edition | 5of10 Downloaded by guest on September 27, 2021 Fig. 3. Machine-learning approach to understand disease severity. Parasite factors associated with a higher risk of patient hospitalization and disease severity were revealed by machine-learning approaches. (A) Summary of RF feature selection strategy to identify parasite virulence factors that discriminate between hospitalized patients (SM + MSM) (Left), SM patients (Right), and OP. The top 10 parasite factors with the highest MDCA are shown. Positive correlation with disease severity is shown with a 1, negative with a −1, and no association with a 0. To adjust for false discovery, familywise error rates (RF mProbes FWER) were estimated using mProbes algorithm and values ≤ 0.2 were considered significant. The predicted binding phenotype was determined as described in Fig. S2. The CMI P value is used to find primers that are significantly informative even after PfHRP2 is accounted for. Var features with a P ≤ 0.05 presented virulence not explained by parasite biomass. (B) evTrees illustrate disease pathways to patient hospitalization (Left) or severe disease (Right) after PfHRP2 filtration (P ≤ 0.20). The percentages in the boxes represent the probability of each pathway to classify patients into hospitalized (H), severe malaria (S), or outpatients (OP). The number of patients in each pathway is indicated below. The percentages beneath the lines show the proportion of the total severe patients classified by each pathway. (C) Var primers were grouped according to binding phenotype or var group (Fig. S2 and Table S4) and ranked by MDCA (15, 16). The association with patient hospitalization (Left) and disease severity (Right) was determined using a Mann–Whitney U. P ≤ 0.05 and FDR ≤ 0.2 are considered significant. FDR, Benjamini–Hochberg adjustment for FDR.

CIDR domains to inhibit APC protective activity in thrombin- inhibited (∼30–45% reduction relative to APC), and IT4var07 and induced endothelial barrier permeability assays. APC diminished r24-2-4 strongly inhibited the APC protective pathway in both thrombin-induced barrier disruption by 24% in EA.hy926 human EA.hy926 cells (40–48% reduction) and primary brain endothelial umbilical vein endothelial cells and 35% in cultured primary hu- cells (60–75% reduction). In general, CIDR domains tended to man brain endothelial cells (Fig. 4C). In agreement with the have slightly higher inhibition on brain endothelial cells than binding competition studies, r62-2-22 and r25-2-4 did not inhibit EA.hy926 cells (Fig. 4C), although the difference did not reach APC barrier protective function, r62-2-1 and r62-2-23 partially statistical significance. Taking these data together, this analysis

6of10 | www.pnas.org/cgi/doi/10.1073/pnas.1524294113 Bernabeu et al. Downloaded by guest on September 27, 2021 PNAS PLUS MICROBIOLOGY

Fig. 4. Inhibition of the APC–EPCR interaction by DC8 CIDRα from severe isolates. DC8 CIDRα domains expressed from SM isolates were analyzed for EPCR binding affinity and blocking the interaction with its ligand APC. (A) Neighbor-joining tree (bootstrap n = 100) of 66 previously classified CIDRα1sequences(14)

and 7 Indian CIDRα1 transcripts amplified from adult SM patients in this study (black dots). (B) The first column shows the dissociation constant (Kd)forrCIDRα1.1- EPCR measured by biolayer interferometry (see Fig. S5 for detailed kinetics). Histograms show APC binding to CHO-EPCR cells in the presence or absence of 250 μg/mL + Indian CIDRα1 domains. The vertical line shows the primary and secondary antibody background used to set the gate for APC cells. Red: strong inhibition; blue: medium; green: low; light gray: no inhibition. The bar graphs show the percentage of APC binding in the presence of CIDR domains relative to APC alone (mean and SD, n = 4 independent experiments). (C) Inhibition by rCIDRα1.1 of APC-dependent protection of endothelial barrier properties. (C, Left)KineticsshowingAPC (50 nM)-mediated protection of thrombin (2 nM) induced barrier disruption in human brain endothelial cell monolayers, and examples of rCIDR1.1 that do or do not inhibit APC barrier protection activity. (Right) Bar graph showing the barrier protection (%) activity of APC on human brain endothelial cells and HUVEC cells (EA. hy926 cells) pretreated with rCIDRα1.1 (mean and SD, n = 6 independent experiments for all CIDR domains, except n = 3 for rPFE1640wCIDRα1.3). P values were calculated using a one-way ANOVA and Dunnet’s multiple comparison test. Significant values are represented by *P ≤ 0.05, **P ≤ 0.01, and ***P ≤ 0.001.

Bernabeu et al. PNAS Early Edition | 7of10 Downloaded by guest on September 27, 2021 indicates that DC8 CIDRα1 domains expressed from SM isolates However, the low parasite biomass in this patient (plasma PfHRP2 = bind EPCR and may inhibit the APC–EPCR pathway. 15.86 ng/mL) might explain the lack of a life-threatening symp- tomatology. Therefore, the relation between CIDRα1 phenotypes Discussion and disease symptoms reinforces the notion that a certain threshold Studies to understand the role of PfEMP1 in SM pathogenesis of parasite biomass in combination with virulent PfEMP1 variants is have been mainly focused on children in Africa (15, 18, 19) and associated with overlapping severe symptomatology in adults. information on adult severe patients still remains scarce (43). The clear association of DC8 with both children (15) and adult However, in areas of unstable transmission, SM occurs across age SM spotlights the CIDRα–EPCR interaction. However, DC8 groups. The different symptomatology in adult SM, the presence of variants encode multiple endothelial binding domains (53), and multiorgan complications, and the higher fatality rate urge research it is possible that other undefined coadhesion traits may also in- on this population. Here, we investigated the relationship between crease the risk for SM. Furthermore, our study, to our knowledge parasite biomass and PfEMP1 in adult SM. for the first time, implicates specific C-terminal PfEMP1 domains Although adult malaria patients presented a complex mixture in SM. DC6 is characterized by DBLγ14-DBLζ5-DBLe4domains of parasites in peripheral blood, parasites expressing group A, and can be present in combination with rosetting, EPCR, or DC8-, and DC6-containing var transcripts were elevated in SM CD36-binding head structures (13). Thus, it will be important to patients. Prior studies suggest group A and DC8-containing var explore if both rosetting and nonrosetting PfEMP1 variants are genes are preferentially expressed in young African children with contributing to adult SM. Although DC6 binding properties are limited immunity (15, 17, 44) and nonimmune European travelers uncharacterized, recent studies have shown that DBLζ and DBLe (45), suggesting these variants confer a parasite growth advantage mediate binding to IgM and α2-macroglobulin, and it has been in malaria naïve hosts, and in some circumstances increase the risk hypothesized that binding to these serum factors can cross-link for SM (14, 15, 17, 23). Machine-learning prediction models sug- multiple PfEMP1 to increase the binding affinity of N-terminal gest that high transcription of DC8 and DC6 domains in combi- domains (54, 55). Further sequencing of DC6 variants will be nation with high parasite biomass is associated with adult patient necessary to assess whether DC6 and DC8 are part of the same or hospitalization and severity. Conversely, the importance of group different proteins in severe isolates, and to determine the binding A dropped in the RF analysis, possibly because group A transcripts properties of DBLζ and DBLe Indian domains to serum factors. were highly expressed in all patient groups, as might be expected A limitation of our study is the relatively small population for a lower transmission setting. It also remains possible that group studied after 3 y of patient recruitment. Future machine-learning A may be more closely linked to specific adult disease syndromes. approaches with larger sample groups and examining different Previously, logistic regression analysis was used to assign different geographic settings will be important to understand whether binding variants to respiratory distress and impaired consciousness differences in var expressions are responsible for distinct adult in pediatric malaria (46), but to our knowledge, this is the first time severe symptoms. A second limitation is that the primers used for that more advanced machine-learning approaches have been used var profiling and sequencing (15) might fail to recognize some var to understand parasite factors associated with disease progression domains important for disease severity or might be biased toward and malaria severity. We expect that powerful machine-learning certain variants. For example, the DC6 cassette is found in both algorithms, such as those presented here, would be useful for rosetting and nonrosetting PfEMP1 variants and rosetting variants analyzing var expression data from African children (15, 18, 19) may have been underestimated by the PCR typing approach. It will and may shed light into pathogenesis mechanisms that drive pe- be valuable to use independent methodologies to investigate the diatric cerebral malaria, respiratory distress, or anemia. contribution of rosetting parasite variants in our adult India pop- Endothelial dysfunction is thought to play an important role in ulation. Other parasite adhesins expressed at the surface of infected SM pathology. Indeed, microthrombi are a common finding in erythrocytes, such as RIFINs and STEVOR, play a role in rosetting pediatric cerebral malaria autopsies (47, 48), and cerebral swelling (56, 57) but were not studied here. In the future, it would be in- is a major risk factor for death (29). Although fibrin deposits and teresting to study the interplay between var and other parasite cerebral swelling are more variable in adult cerebral malaria au- adhesins in disease severity. Nevertheless, the degenerate primers topsies (49, 50), alterations in blood–brain barrier integrity have against var adhesion domains remain a sensitive and cost-effective been associated with infected erythrocyte sequestration in adult tool to cover the var geographical diversity and, combined with Southeast Asian and pediatric African autopsies (51, 52). It has machine-learning approaches, provide a powerful methodology to been hypothesized that EPCR binding parasites may drive disease investigate pathogenic mechanisms. pathogenesis by blocking the anticoagulation, anti-inflammatory, In summary, our data show that elevated DC8 and DC6 var and barrier protective functions of the APC–EPCR pathway (24). transcripts, along with high parasite biomass, promote disease However, recent studies of CIDRα1 domains from long-term progression in adult SM. In addition, our findings raise the pos- cultured adapted parasites have revealed large differences in their sibility that DC8 CIDRα1 domains with low or moderate APC ability to inhibit the APC–EPCR interaction (25–28). The con- blockade activity interfere with APC–EPCR protective pathways, sequences of these differences for disease severity remain un- highlighting attention on this pathway for disease interventions and known. Here, we showed that DC8 CIDRα1 domains expressed by the future development of SM adjunctive therapies. SM isolates possess differential activity to disrupt the EPCR–APC protective pathway. Some DC8 domains had low activity and Materials and Methods others had nearly equivalent activity to a group A CIDRα1do- Ethical Approval. Informed consent was obtained from all study participants. main that strongly blocks the APC–EPCR interaction (25, 26, 28). The study was approved by the ethics boards at Goa Medical College and Notably, an identical DC8 CIDRα1 domain (r62-2-1) isolated Hospital, the University of Washington, the Western Institutional Review from two patients with cerebral malaria and a third patient with Board used on behalf of the Center for Infectious Disease Research, as well as multiorgan complications produced a small and significant in- by the Government of India Health Ministry Screening Committee. hibition of APC barrier protective activity on brain endothelial cells. This finding raises the possibility that even parasite domains Patient Recruitment and Samples Collection. Subjects were recruited between April 2012 to October 2014 from the hospital admission or outpatient wards with weaker APC inhibition activity may interfere with this impor- from patients presenting at Goa Medical College and Hospital (Goa, India). tant host regulatory pathway, especially in microvascular beds where Subjects were enrolled by project staff who explained the project. Following there is loss of EPCR expression associated with parasite seques- informed consent, blood samples from P. falciparum-positive patients were tration (30). Unexpectedly, a DC8 CIDRα1 domain (r24-2-4) iso- collected in acid citrate dextrose vacutainers and separated into plasma and lated from an anemic patient had the highest APC blockade activity. red blood cells in RNALater. Samples were stored at –80 °C. Infections were

8of10 | www.pnas.org/cgi/doi/10.1073/pnas.1524294113 Bernabeu et al. Downloaded by guest on September 27, 2021 confirmed by study staff using Giemsa-stained thin and thick smears for Biolayer Interferometry Analysis. Binding of the CIDR domains to biotinylated PNAS PLUS parasitemia determination and Plasmodium species identification. Rapid EPCR was determined on the Octet QKe instrument (ForteBio), as described

diagnostic test (Zephyr Biomedicals) was additionally used for the diagnosis previously (28). Mean Kon and Koff and apparent Kd values were determined of parasite species. Parasites per milliliter was calculated from thin film from double-reference subtracted data from three concentrations that were smears (count/1,000 RBC/125.6/Hematocrit). SM was defined as: (i) coma fitted globally to a 1:1 Langmuir binding model using the data analysis software. (Glasgow Coma Score < 10), (ii) severe anemia (Hb < 7 g/dL), (iii) jaundice > > (bilirubin 3 mg/dL), (iv) renal compromise [serum creatinine 3 mg/dL or rCIDRα Binding Titration to CHO745-EPCR and APC Competition Assay. For the > < (blood urea nitrogen 17 mmol/L), (v) shock (systolic blood pressure 80 rCIDRα binding titration to CHO745-EPCR and APC competition assay, 105 mmHg with cold extremities), (vi) metabolic acidosis (peripheral venous bi- CHO745-EPCR cells were lifted and washed in complete HBSS (HBSS with carbonate < 15 mmol/L), (vii) respiratory distress (respiratory rate > 20 3mMCaCl, 0.6 mM MgCl , 1% BSA) (28). A five-point titration curve was de- breaths per minute or PaO < 75 mmHg), and (viii) hypoglycemia (blood 2 2 2 termined with 1–250 μg/mL of recombinant CIDR for 30 min. CIDR binding was glucose < 40 mg/dL). Patients admitted to the hospital without any of these detected with a rabbit polyclonal anti-StrepII tag antibody followed by a goat criteria were considered as MSM and nonadmitted patients were considered as OP. Patients with mono P. falciparum infections were treated with oral anti-rabbit Alexa488-coupled antibody. Control samples were labeled with pri- artesunate and mefloquine, and intravenous artesunate was used for SM mary and secondary antibodies alone to set gates. APC competition assays were 5 patients. One study patient was coinfected with P. falciparum and Plasmo- done as described previously (28). Briefly, 10 CHO745-EPCR cells were coincu- dium vivax and treated with oral artesunate, mefloquine, and primaquine. batedwith50μg/mL APC (Sigma) and 50 or 250 μg/mL CIDR recombinant pro- All patient care was managed according to hospital standard procedures. teins for 30 min on ice. APC binding was detected with a goat anti-APC mAb (1:100; Affinity Biologicals) followed by a chicken anti-goat Alexa488 coupled PfHRP2 Plasma Quantification. PfHRP2 was quantified using double-site antibody. The percentage (%) APC binding was calculated relative to the value of sandwich ELISA according to published methodologies (31, 36). In brief, cells incubated with APC alone. Labeled cells were analyzed by flow cytometry plates were coated overnight with mouse anti-PfHRP2 IgM antibody (MPFM- using LSRII (Becton Dickinson) and data were analyzed by FlowJO 10 software 55A, ICL) at 1 μg/mL in PBS and blocked for 4 h with 2% (wt/vol) BSA-PBS. (Tree Star). Patient plasma samples were diluted to the desired detectable dilutions (1:10–1:200) and tested in triplicate for 1 h. For detection, mouse anti- Measurement of the Monolayer Permeability. Barrier function was moni- PfHRP2 IgG antibody (MPFG-55P, ICL) was added at 0.2 μg/mL in 2% (wt/vol) tored using a real-time cell analyzer (xCELLigence System, ACEA Biosciences). BSA-1% Tween 20-PBS for 1 h, incubated for 5 min with TMB reaction This system measures electrical impedance across the cell monolayer, cell im- substrate, and measured spectrophotometrically at 450 nm. Positive and pedance (CI), via gold microelectrodes integrated on the bottom of a 96-well negative controls were included in each plate. A standard curve was plate. Next, 10,000 EA.hy926 HUVEC (ATCC) or primary human brain endo- established using purified recombinant PfHRP2 protein (kindly donated by thelial cells (Cells Systems) were seeded in each well. Cell proliferation was

David Sullivan, Johns Hopkins Bloomberg School of Public Health, Baltimore) MICROBIOLOGY assessed for 72 h, at which time the cells reached a sustained maximum CI value. diluted 0.25–55 ng/mL in PBS. Five patient samples that presented a lower For the experiment, cells were incubated with rCIDRα1 (0.05 mg/mL) or culture concentration than the detection limit were excluded from the analysis. medium. After 30 min, 100 nM of human APC (Haematologic Technologies) was added and incubated for 2 h. Barrier disruption was induced with 5 nM Determination of var Transcription by qRT-PCR. Thawed red blood cells in thrombin (Sigma) and compared with untreated cells (baseline). Cells were RNAlater were dissolved in 12 volumes of TRIzol and RNA was extracted using treated in triplicate and CI was measured every minute up to 120 min, then an RNeasy micro kit (Qiagen). cDNA was synthesized using random hexamers every 5 min up to 245 min after thrombin challenge. The baseline-normalized and a MultiScribe reverse transcriptase (Thermo Fisher). qRT-PCR was per- formed using QuantiTect SYBR in a Mastercycler Realplex2 following pub- cell index was calculated by comparing the CI values of treated cells to the CI lished amplification conditions (15, 16). Data were acquired after the values for baseline control wells of untreated cells at the time point immedi- elongation step of each cycle. Absence of DNA in RNA samples was con- ately before the thrombin challenge. The level of barrier protection achieved firmed by running a reverse-transcriptase negative sample with the house- by APC + thrombin treatment relative to thrombin treatment alone was set to keeping gene primer adenylosuccinate lyase (ASL) (PFB0295w). Levels of var 100% to calculate the binding inhibitory activity of CIDR domains. gene expression were determined by relative quantification of the average expression of ASL and seryl-tRNA synthetase housekeeping genes (ΔCt Machine-Learning Models. Decision trees for understanding parasite factors var_primer = Ct var_primer − Ct average_housekeeping primers). The level associated with patient hospitalization and disease severity were analyzed by of var expression was represented as Transcript units (Tu) and calculated as RF (39) and evolutionary Trees (59) or by CMI analysis (60). Detailed methods (5−ΔCt) Tu = 2 (15). Samples were included in the univariate statistical analysis can be found in SI Materials and Methods. only if the Ct average of the housekeeping genes was below 30. Statistical Analysis. Univariate analyses were performed using GraphPad CIDRα1 Sequencing and Recombinant Protein Expression. DC8 transcripts were Prism v5.02 for Windows. Correlations between variables were tested using the amplified from patients using the strategy depicted in Fig. S4. Bands of the Spearman’s rank correlation coefficient. Differences between groups were expected size were excised, purified, and cloned in a Zero Blunt TOPO vector. At evaluated using the Mann–Whitney U test with a Benjamini–Hochberg adjust- least 15 different colonies were sequenced per patient and analyzed using ment for FDR or the one-way ANOVA with a Dunnett’s multiple comparison test. Geneious (7.1.7). Sequences were deposited in GenBank with accession numbers KU843600–KU843604. Recombinant CIDRα1 were synthesized as GBlocks gene – ACKNOWLEDGMENTS. We thank the patients who participated in this study; fragments (IDT) and produced as His6-MBP-TEV-PfEMP1 insert-StrepII tagged Dr. David Sullivan for PfHRP-2 recombinant protein; and Profs. Panda and proteins, as previously described (58). Recombinant proteins were purified in a Patankar for the use of their ELISA reader at IIT, Bombay. This work was two-step process using an amino-terminal His tag and a carboxyl-terminal StrepII supported by funds from NIH Grants U19 AI 089688 (to P.K.R. and J.D.S.) and tag and analyzed by SDS/PAGE according to standard procedures (28). P41 GM109824 (to J.D.A.).

1. World Health Organization (2015) WHO World Malaria Report 2015. Avaliable 8. Baruch DI, et al. (1995) Cloning the P. falciparum gene encoding PfEMP1, a malarial at www.who.int/malaria/publications/world-malaria-report-2015/report/en/. Accessed variant antigen and adherence receptor on the surface of parasitized human eryth- February 2016. rocytes. Cell 82(1):77–87. 2. Anonymous (2014) Severe malaria. Trop Med Int Health 19(Suppl 1):7–131. 9. Smith JD, et al. (1995) Switches in expression of Plasmodium falciparum var genes 3. White NJ, et al. (2014) Malaria. Lancet 383(9918):723–735. correlate with changes in antigenic and cytoadherent phenotypes of infected 4. Dondorp AM, et al. (2008) The relationship between age and the manifesta- erythrocytes. Cell 82(1):101–110. tions of and mortality associated with severe malaria. Clin Infect Dis 47(2): 10. Su XZ, et al. (1995) The large diverse gene family var encodes proteins involved in 151–157. cytoadherence and antigenic variation of Plasmodium falciparum-infected erythro- 5. Miller LH, Baruch DI, Marsh K, Doumbo OK (2002) The pathogenic basis of malaria. cytes. Cell 82(1):89–100. Nature 415(6872):673–679. 11. Lavstsen T, Salanti A, Jensen AT, Arnot DE, Theander TG (2003) Sub-grouping of 6. Dondorp AM, et al. (2008) Direct in vivo assessment of microcirculatory dysfunction in Plasmodium falciparum 3D7 var genes based on sequence analysis of coding and non- severe falciparum malaria. J Infect Dis 197(1):79–84. coding regions. Malar J 2(1):27. 7. Yeo TW, et al. (2008) Angiopoietin-2 is associated with decreased endothelial nitric 12. Smith JD, Subramanian G, Gamain B, Baruch DI, Miller LH (2000) Classification of oxide and poor clinical outcome in severe falciparum malaria. Proc Natl Acad Sci USA adhesive domains in the Plasmodium falciparum erythrocyte membrane protein 105(44):17097–17102. 1 family. Mol Biochem Parasitol 110(2):293–310.

Bernabeu et al. PNAS Early Edition | 9of10 Downloaded by guest on September 27, 2021 13. Rask TS, Hansen DA, Theander TG, Gorm Pedersen A, Lavstsen T (2010) Plasmodium 39. Liaw A, Wiener M (2002) Classification and regression by randomForest. R News 2(3): falciparum erythrocyte membrane protein 1 diversity in seven genomes—Divide and 18–22. conquer. PLOS Comput Biol 6(9):e1000933. 40. Huynh-Thu VA, Saeys Y, Wehenkel L, Geurts P (2012) Statistical interpretation 14. Kyriacou HM, et al. (2006) Differential var gene transcription in Plasmodium falci- of machine learning-based feature importance scores for biomarker discovery. parum isolates from patients with cerebral malaria compared to hyperparasitaemia. Bioinformatics 28(13):1766–1774. Mol Biochem Parasitol 150(2):211–218. 41. Finney OC, et al. (2014) Predicting antidisease immunity using proteome arrays and 15. Lavstsen T, et al. (2012) Plasmodium falciparum erythrocyte membrane protein sera from children naturally exposed to malaria. Mol Cell Proteomics 13(10): 1 domain cassettes 8 and 13 are associated with severe malaria in children. Proc Natl 2646–2660. Acad Sci USA 109(26):E1791–E1800. 42. Rowe JA, Claessens A, Corrigan RA, Arman M (2009) Adhesion of Plasmodium falci- 16. Rottmann M, et al. (2006) Differential expression of var gene groups is associated parum-infected erythrocytes to human cells: Molecular mechanisms and therapeutic with morbidity caused by Plasmodium falciparum infection in Tanzanian children. implications. Expert Rev Mol Med 11:e16. Infect Immun 74(7):3904–3911. 43. Subudhi AK, et al. (2015) Plasmodium falciparum complicated malaria: Modulation 17. Warimwe GM, et al. (2009) Plasmodium falciparum var gene expression is modified by and connectivity between exportome and variant surface antigen gene families. Mol – host immunity. Proc Natl Acad Sci USA 106(51):21801 21806. Biochem Parasitol 201(1):31–46. 18. Warimwe GM, et al. (2012) Prognostic indicators of life-threatening malaria are as- 44. Cham GK, et al. (2009) Sequential, ordered acquisition of antibodies to Plasmodium sociated with distinct parasite variant antigen profiles. Sci Transl Med 4(129):129ra45. falciparum erythrocyte membrane protein 1 domains. J Immunol 183(5):3356–3363. 19. Bertin GI, et al. (2013) Expression of the domain cassette 8 Plasmodium falciparum 45. Bachmann A, et al. (2011) Highly co-ordinated var gene expression and switching in erythrocyte membrane protein 1 is associated with cerebral malaria in Benin. PLoS clinical Plasmodium falciparum isolates from non-immune malaria patients. Cell One 8(7):e68368. Microbiol 13(9):1397–1409. 20. Carlson J, et al. (1990) Human cerebral malaria: Association with erythrocyte rosetting 46. Abdi AI, et al. (2014) Plasmodium falciparum antigenic variation: Relationships be- – and lack of anti-rosetting antibodies. Lancet 336(8729):1457 1460. tween widespread endothelial activation, parasite PfEMP1 expression and severe 21. Ghumra A, et al. (2012) Induction of strain-transcending antibodies against Group malaria. BMC Infect Dis 14(1):170. A PfEMP1 surface antigens from virulent malaria parasites. PLoS Pathog 8(4): 47. Dorovini-Zis K, et al. (2011) The neuropathology of fatal cerebral malaria in Malawian e1002665. children. Am J Pathol 178(5):2146–2158. 22. Kaul DK, Roth EF, Jr, Nagel RL, Howard RJ, Handunnetti SM (1991) Rosetting of 48. Milner DA, Jr, et al. (2014) The systemic pathology of cerebral malaria in African Plasmodium falciparum-infected red blood cells with uninfected red blood cells en- children. Front Cell Infect Microbiol 4:104. hances microvascular obstruction under flow conditions. Blood 78(3):812–819. 49. MacPherson GG, Warrell MJ, White NJ, Looareesuwan S, Warrell DA (1985) Human 23. Turner L, et al. (2013) Severe malaria is associated with parasite binding to endo- cerebral malaria. A quantitative ultrastructural analysis of parasitized erythrocyte thelial protein C receptor. Nature 498(7455):502–505. sequestration. Am J Pathol 119(3):385–401. 24. Mosnier LO, Zlokovic BV, Griffin JH (2007) The cytoprotective protein C pathway. 50. Spitz S (1946) The pathology of acute falciparum malaria. Mil Surg 99(5):555–572. Blood 109(8):3161–3172. 51. Brown H, et al. (1999) Evidence of blood-brain barrier dysfunction in human cerebral 25. Gillrie MR, et al. (2015) Diverse functional outcomes of Plasmodium falciparum liga- malaria. Neuropathol Appl Neurobiol 25(4):331–340. tion of EPCR: Potential implications for malarial pathogenesis. Cell Microbiol 17(12): 52. Brown H, et al. (2001) Blood-brain barrier function in cerebral malaria in Malawian 1883–1899. children. Am J Trop Med Hyg 64(3-4):207–213. 26. Lau CK, et al. (2015) Structural conservation despite huge sequence diversity allows 53. Avril M, Brazier AJ, Melcher M, Sampath S, Smith JD (2013) DC8 and DC13 var genes EPCR binding by the PfEMP1 family implicated in severe childhood malaria. Cell Host associated with severe malaria bind avidly to diverse endothelial cells. PLoS Pathog Microbe 17(1):118–129. 9(6):e1003430. 27. Petersen JE, et al. (2015) Protein C system defects inflicted by the malaria parasite 54. Stevenson L, et al. (2015) α 2-Macroglobulin can crosslink multiple Plasmodium fal- protein PfEMP1 can be overcome by a soluble EPCR variant. Thromb Haemost 114(5): ciparum erythrocyte membrane protein 1 (PfEMP1) molecules and may facilitate 1038–1048. adhesion of parasitized erythrocytes. PLoS Pathog 11(7):e1005022. 28. Sampath S, et al. (2015) Plasmodium falciparum adhesion domains linked to severe 55. Jeppesen A, et al. (2015) Multiple Plasmodium falciparum erythrocyte membrane malaria differ in blockade of endothelial protein C receptor. Cell Microbiol 17(12): protein 1 variants per genome can bind IgM via its Fc fragment Fcμ. Infect Immun 1868–1882. – 29. Seydel KB, et al. (2015) Brain swelling and death in children with cerebral malaria. 83(10):3972 3981. N Engl J Med 372(12):1126–1137. 56. Niang M, et al. (2014) STEVOR is a Plasmodium falciparum erythrocyte binding pro- – 30. Moxon CA, et al. (2013) Loss of endothelial protein C receptors links coagulation and tein that mediates merozoite invasion and rosetting. Cell Host Microbe 16(1):81 93. inflammation to parasite sequestration in cerebral malaria in African children. Blood 57. Goel S, et al. (2015) RIFINs are adhesins implicated in severe Plasmodium falciparum – 122(5):842–851. malaria. Nat Med 21(4):314 317. 31. Dondorp AM, et al. (2005) Estimation of the total parasite biomass in acute falcipa- 58. Avril M, et al. (2012) A restricted subset of var genes mediates adherence of Plas- rum malaria from plasma PfHRP2. PLoS Med 2(8):e204. modium falciparum-infected erythrocytes to brain endothelial cells. Proc Natl Acad Sci – 32. Hendriksen IC, et al. (2012) Diagnosing severe falciparum malaria in parasitaemic USA 109(26):E1782 E1790. African children: A prospective evaluation of plasma PfHRP2 measurement. PLoS Med 59. Grubinger T, Zeileis A, Pfeiffer K-P (2014) evtree: Evolutionary learning of globally 9(8):e1001297. optimal classification and regression trees in R. Journal of Statistical Software, 61(1): – 33. Fox LL, et al. (2013) Histidine-rich protein 2 plasma levels predict progression to ce- 1 29. rebral malaria in Malawian children with Plasmodium falciparum infection. J Infect 60. Wyner AD (1978) A definition of conditional mutual information for arbitrary en- Dis 208(3):500–503. sembles. Inf Control 38(1):51–59. 34. Seydel KB, et al. (2012) Plasma concentrations of parasite histidine-rich protein 61. Fisher R (1921) On the «probable error» of a coefficient of correlation deduced from a 2 distinguish between retinopathy-positive and retinopathy-negative cerebral ma- small sample. Metron 1(Pt 4):1–32. laria in Malawian children. J Infect Dis 206(3):309–318. 62. Bengtsson A, et al. (2013) A novel domain cassette identifies Plasmodium falciparum 35. Hendriksen IC, et al. (2013) Defining falciparum-malaria-attributable severe febrile PfEMP1 proteins binding ICAM-1 and is a target of cross-reactive, adhesion-inhibitory illness in moderate-to-high transmission settings on the basis of plasma PfHRP2 antibodies. J Immunol 190(1):240–249. concentration. J Infect Dis 207(2):351–361. 63. Howell DPG, et al. (2008) Mapping a common interaction site used by Plasmodium 36. Gonçalves BP, et al. (2014) Parasite burden and severity of malaria in Tanzanian falciparum Duffy binding-like domains to bind diverse host receptors. Mol Microbiol children. N Engl J Med 370(19):1799–1808. 67(1):78–87. 37. World Health Organization (2000) Severe falciparum malaria. World Health Organi- 64. Oleinikov AV, et al. (2009) High throughput functional assays of the variant antigen zation, communicable diseases cluster. Trans R Soc Trop Med Hyg 94(Suppl 1):S1–S90. PfEMP1 reveal a single domain in the 3D7 Plasmodium falciparum genome that binds 38. Smith JD, Rowe JA, Higgins MK, Lavstsen T (2013) Malaria’s deadly grip: Cytoadhesion ICAM1 with high affinity and is targeted by naturally acquired neutralizing of Plasmodium falciparum-infected erythrocytes. Cell Microbiol 15(12):1976–1983. antibodies. PLoS Pathog 5(4):e1000386.

10 of 10 | www.pnas.org/cgi/doi/10.1073/pnas.1524294113 Bernabeu et al. Downloaded by guest on September 27, 2021