key: cord-0291655-zzy64czb
authors: Opi, D. H.; Boyle, M. J.; McLean, A.; Reiling, L.; Chan, J.-A.; Stanisic, D. I.; Ura, A.; Mueller, I. J.; Fowkes, F. J.; Rogerson, S. J.; Beeson, J.
title: Reduced risk of placental parasitaemia associated with complement fixation on Plasmodium falciparum by antibodies among pregnant women
date: 2021-06-05
journal: nan
DOI: 10.1101/2021.06.02.21258254
sha: d0907eb8d10ece44bd6081624ce9b09aa0c4359f
doc_id: 291655
cord_uid: zzy64czb

Background: The pathogenesis of malaria in pregnancy (MiP) involves accumulation of P. falciparum-infected red blood cells (pRBCs) in the placenta, contributing to poor pregnancy outcomes. Parasite accumulation is primarily mediated by P. falciparum erythrocyte membrane 1 (PfEMP1). Magnitude of IgG to pRBCs has been associated with reduced risk of MiP in some studies, but associations have been inconsistent. Further, antibody effector mechanisms are poorly understood, and the role of antibody complement interactions is unknown. Methods: Studying a longitudinal cohort of pregnant women (n=302) from a malaria-endemic province in Papua New Guinea (PNG), we measured the ability of antibodies to fix and activate complement using placental binding pRBCs and PfEMP1 recombinant domains. We determined antibody-mediated complement inhibition of pRBC binding to the placental receptor, chondroitin sulfate A (CSA) and associations with protection against placental parasitaemia. Results: Some women acquired antibodies that effectively promoted complement fixation on placental-binding pRBCs. Complement fixation correlated with IgG1 and IgG3 antibodies, which dominated the response. There was, however, limited evidence for membrane attack complex activity or pRBC lysis or killing. Importantly, a higher magnitude of complement fixing antibodies was prospectively associated with reduced odds of placental infection at delivery. Using genetically-modified P. falciparum and recombinant PfEMP1 domains, we found that complement-fixing antibodies primarily targeted a specific variant of PfEMP1 (known as VAR2CSA). Furthermore, complement enhanced the ability of antibodies to inhibit pRBC binding to CSA, which was primarily mediated by complement C1q protein. Conclusion: These findings provide new insights into mechanisms mediating immunity to MiP and reveal potential new strategies for developing malaria vaccines that harness antibody-complement interactions.

Malaria in pregnancy (MiP) causes significant maternal, fetal and neonatal mortality 62 and is a major health issue globally (1). Parasite accumulation in the placenta is a key 63 feature of MiP following infection with Plasmodium falciparum (2, 3) but is not 64 prominent with other human infecting Plasmodium species. This largely results from 65 the selective binding of pRBCs to chondroitin sulfate A (CSA) expressed on 66 syncytiotrophoblasts (4, 5), and other binding interactions may play secondary roles 67 (6, 7). This is mediated by a Plasmodium falciparum erythrocyte membrane protein 1 68 protective antibodies directed against placental-binding P. falciparum infected red 72 blood cells (pRBCs)(9). Antibodies to placental-binding pRBCs and VAR2CSA have 73 been associated with improved outcomes in some studies, although associations have 74 not been entirely consistent (10). How antibodies to pRBCs and VAR2CSA function 75 in protective immunity to MiP and improved birth outcomes is not fully understood 76 (10), and these key knowledge gaps, are restricting advancement of MiP vaccines. 77

Existing data shows that antibodies may act via inhibition of placental adhesion of 78 pRBCs (11) and promoting phagocytosis of pRBCs (12). However, a recent 79 systematic review found that data on associations between these antibody functions 80 and protection from the consequences of MiP are limited and variable (10) and data 81 suggest these mechanisms may not fully explain immunity to MiP. VAR2CSA is a 82 leading vaccine candidate for MiP with two VAR2CSA based vaccines having 83 completed phase I trials (13, 14), which highlights the importance of a strong 84 understanding of immunity to inform further vaccine design and development for 85

Antibodies to placental-binding pRBCs and VAR2CSA are dominated by 87 MAC formation can lead to cell lysis or death, ii) complement components C3 and C5 94 . CC-BY-NC 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted June 5, 2021. ; https://doi.org/10.1101/2021.06.02.21258254 doi: medRxiv preprint can promote phagocytosis by monocytes and neutrophils through interactions with 95 complement receptors expressed on those cells, iii) binding of complement 96 components to a pathogen surface may also have direct inhibitory or neutralizing 97 activity (18-20). Antibody-mediated complement fixation against P. falciparum has 98 been implicated in immunity in non-pregnant individuals, targeting merozoites (21, 99 22), sporozoites (23, 24) and gametocytes (25). While there is a potential role for 100 antibody-mediated complement fixation in immunity against MiP, this has not been 101 established and RBCs are known to express complement regulatory proteins that can 102 inhibit complement activation and confer resistance to lysis (26). Older studies have 103 reported antibody-mediated complement fixation on the surface of mature pigmented 104 trophozoite stage pRBCs, but these did not assess different functional activities or 105 associations with protection (26, 27). Further, MiP has also been associated with 106 excessive complement activation and production of inflammatory products mediating 107 adverse outcomes (28). Therefore, the role of complement fixation in preventing or 108 reducing placental infection remains unclear, and there has been little investigation of 109 its potential role in immunity. 110

We hypothesised that acquired antibodies against VAR2CSA in some 111 pregnant women may fix complement on placental binding pRBCs, which may 112 contribute to the control or prevention of placental parasitaemia. Using a prospective 113 longitudinal cohort study of malaria-exposed pregnant women from PNG, we 114 investigated antibody-complement interactions in immunity to MiP. We investigated 115 antigenic targets of complement fixation on the surface of pRBCs using genetically-116 modified pRBCs and recombinant VAR2CSA proteins, and evaluated functional 117 mechanisms of complement-fixing antibodies. We examined whether antibody-118 mediated complement fixation on pRBCs is associated with reduced risk of placental 119 infection. 120

. CC-BY-NC 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted June 5, 2021. α+thalassaemia are common in this population and have been associated with 148 protection against severe clinical malaria in some studies ((32, 33)). Therefore, these 149 were included as possible confounders in our study. Molecular typing was conducted 150 as previously described (34). Samples from malaria non-exposed residents of 151

Melbourne, Australia were used as negative controls in all assays. 152 153 154 . CC-BY-NC 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted June 5, 2021. 

The ability of immune antibodies in the presence or absence of complement to inhibit 174 pRBC adhesion to CSA was assessed using a modified version of a static-based 175 binding assay described previously (38). Petri dishes were coated with 2.5μg/ml of 176 CSA diluted in PBS and incubated overnight at 4°C. Plates were subsequently 177 blocked with 1% bovine serum albumin (BSA) in PBS followed by gentle washing. 178

Mature pigmented-trophozoite stage CS2 pRBCs at approximately 3% parasitaemia 179 and 1% haematocrit, in RPMI, were opsonized with a pool of antibodies of serum 180 samples from either PNG pregnant women (N=9) with high IgG reactivity to 181 VAR2CSA-DBL5 (3D7) as determined by ELISA or malaria non-exposed Australian 182 donors (N=10). The pooled samples were tested in assays at a final concentration of 183 10% in RPMI-HEPES. Additionally, to test for complement fixation, samples were 184 concurrently incubated with normal serum (NS; complement active) or heat 185 inactivated serum (HIS; complement inactive) at a final dilution of 25% in RPMI, 186 10μg/ml purified human C1q (Millipore) or C1q-depleted human serum (Millipore) at 187 a final dilution of 10% in RPMI, or RPMI (negative control) for 30 minutes at 37°C at 188 . CC-BY-NC 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted June 5, 2021. ; https://doi.org/10.1101/2021.06.02.21258254 doi: medRxiv preprint a final volume of 50μl. Both NS and HIS were from malaria non-exposed Australian 189 donors. The parasite suspension was then added onto the CSA coated spots and 190 incubated at 37°C for a further 15 minutes. Unbound cells were washed off with 191 gentle agitation and bound pRBCs were fixed in 2% glutaraldehyde in PBS followed 192 by staining with 10% Giemsa. Each sample was tested in duplicate spots and repeated 193 in 3-6 independent assays. Images of adherent pRBCs were captured using an with 2% casein in PBS at 37°C for 2 hours and then incubated with antibodies at a 211 dilution of 1/100 followed by purified C1q (Millipore) at 10μg/ml or C5-depleted 212 serum (Millipore) at a dilution of 1/10 for the detection of C3, at room temperature for 213 1 hour. A combination of goat anti-C1q plus HRP-conjugated rabbit anti-goat 214 antibodies and rabbit anti-C3 plus HRP-conjugated goat anti-rabbit at dilutions of 215 1/2000 were used for the detection of C1q and C3 respectively. Reactivity was 216 determined by measuring the optical density (OD) at 405nm following the addition of 217 ABTS (2,2'-azino-bis(3-ethylbenzothiazoline-6-sulfonic acid)) and stopping of the 218 reaction with 1% sodium dodecyl sulfate (SDS) after 15 minutes to 1 hour. Blank 219 wells, without antibodies added, were used to subtract non-specific signal from each 220 well. IgG to VAR2CSA was measured using goat anti-human IgG-HRP (Millipore) 221

while IgG subclasses to VAR2CSA were measured using sheep anti-human IgG1, 222 . CC-BY-NC 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted June 5, 2021. high IgG reactivity to VAR2CSA by ELISA during assay optimisation. Seropositive 227 samples were classified as having an OD greater than the mean + 3 standard 228 deviations of malaria non-exposed Australian donors (n=15). 229 230 Flow cytometry complement fixation assays with intact pRBCs: Pigmented 231 trophozoite stage parasites at ~10% parasitaemia and 0.5% haematocrit were 232 incubated with 25% NS and 10% antibodies from PNG pregnant women or Australian 233 malaria non-exposed donors for 1 hour at 37°C. pRBCs were then washed in 0.5% 234 BSA in PBS and C1q, C3 and C5-9 detected by staining with polyclonal rabbit anti-235 human C1q, goat anti-human C3 or monoclonal mouse anti-human C5-9 antibodies 236 followed by respective polyclonal Alexa-488 conjugated goat anti-rabbit, rabbit anti-237 goat or goat anti-mouse antibodies. pRBCs were detected by staining with ethidium 238 bromide (Bio-Rad) at 1/1000 dilution. Levels of complement fixation on pRBCs were 239 expressed as the Alexa-488 geometric mean fluorescent intensity of ethidium bromide 240 stained pRBC populations. It is reported that aged RBCs can bind complement factors 241 (46), and in assay work-up we found it was important to use fresh (recently collected) 242

RBCs in assays. To prepare parasites for these assays, mature-pigmented trophozoites 243 were purified by magnet filtration, mixed with fresh RBCs, and cultured for 48 hours 244 to obtain a parasitaemia of ~10% parasitaemia. 245 246 Detection of complement fixation on pRBC surface by western blot: Magnet-247 purified pigmented trophozoite pRBCs (>95% purity) were incubated with a 10% 248 pool of antibodies (N=9) from PNG pregnant women with high IgG reactivity to 249 VAR2CSA-DBL5 (3D7) or a pool of antibodies (N=10) from malaria non-exposed 250

Australian donors, and 25% NS as a source of complement, for 15 minutes at 37°C. 251 pRBCs were then washed in cold PBS containing protease inhibitors. Samples were 252 resuspended in reducing SDS sample buffer, heated at 95°C for 5 minutes and 253 proteins separated on 4%-12% Bis-Tris gels (Invitrogen). Proteins were then 254 transferred to nitrocellulose membranes (Invitrogen). C1q (29 Kd) was detected by 255 staining with a rabbit anti-C1q antibody. C3b was detected using an anti-C3 antibody 256 . CC-BY-NC 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted June 5, 2021. ; https://doi.org/10.1101/2021.06.02.21258254 doi: medRxiv preprint that labels both C3b and iC3b components that comigrate at ~68kDa (21). Labelling 257 of heat shock protein 70 (HSP-70) was used as a loading control. 258 259

Mature pigmented trophozoite stage CS2 parasites at 0.5% parasitaemia and 3% 261 haematocrit were cultured in complete medium in 96 well plates in the presence of 262 10% antibodies from PNG pregnant women (N=7) alone or with additional 25% NS 263 or HIS. A negative control with no PNG or Australian malaria non-exposed donor 264 antibodies and no NS or HIS added was also included in the assay. After heterozygous or homozygous). We tested for evidence of effect modification by P. 287

falciparum peripheral infection at enrolment or gravidity on birth outcomes using log-288 likelihood ratio tests, with and without an interaction term. For all analysis P values 289 less than 0.05 were considered statistically significant. 290 . CC-BY-NC 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted June 5, 2021. ; https://doi.org/10.1101/2021.06.02.21258254 doi: medRxiv preprint

We tested whether acquired antibodies from pregnant women from a malaria endemic 294 PNG region could fix and activate complement on pRBCs. Trophozoite stage pRBCs 295 of the placental binding P. falciparum isolate CS2 were opsonized with antibodies 296 from 302 PNG women or 15 malaria non-exposed Australian donors and tested for 297 their ability to fix complement factors C1q or C3 by flow cytometry. Antibodies from 298 malaria-exposed pregnant PNG women mediated significantly higher levels of 299 complement C1q ( Figure 1A ) and C3 ( Figure 1B ) fixation to the surface of CS2 300 pRBCs, compared to malaria non-exposed control donors (both P<0.001) 301

(characteristics of the 302 women included in this study data are shown in Table 1 ). 302

The prevalence of complement-fixing antibodies among pregnant women was 66% 303 (200/302) for C1q and 51% (155/302) for C3 ( Figure 1A -B). C1q fixation on pRBCs 304 was moderately positively correlated with C3 fixation on pRBCs (Spearman r=0.36, 305 P<0.001). C3 fixation was abolished when using heat-inactivated serum (HIS) 306 (complement inactive) as a control, compared to normal serum (NS) (complement 307 active) (P<0.001) ( Figure 1C ). C1q fixation on the surface of pRBCs was confirmed 308 by Western blot in the presence of antibodies from PNG pregnant women compared 309 to malaria non-exposed antibody pool ( Figure 1D ). Some C3 fixation on pRBCs was 310 detected when using malaria non-exposed antibodies or unopsonized controls, but 311 levels were much lower than those seen with the PNG antibodies pool ( Figure 1D) . A 312 subset of samples was tested for complement fixation using an isolate from PNG 313 (XIE), which is genetically distinct from CS2 (47). There was significantly higher C3 314 fixation among PNG pregnant women's samples (N=35) compared to non-exposed 315 Figure S2A -B)) we tested for evidence of C5-9 322 fixation and MAC activity on the surface of CS2 PRBCs. We observed no evidence of 323 . CC-BY-NC 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted June 5, 2021. ; https://doi.org/10.1101/2021.06.02.21258254 doi: medRxiv preprint elevated C5-9 fixation on pRBCs by flow cytometry when these PNG antibodies were 324 used to opsonize pRBCs in the presence of normal serum (Figure 2A) . Additionally, 325

we did not observe any significant correlation between C5-9 fixation and C1q or C3 326 fixation on the surface of CS2 pRBCs ( Figure 2B ). To assess whether there was 327 significant MAC activity leading to pRBC lysis or impacting on pRBC viability, we 328 tested whether incubation of CS2 pRBCs in the presence of PNG women's antibodies 329 (N=7) with active complement (NS) compared to heat-inactivated serum (inactive 330 complement) impacted parasite replication in vitro. There was no evidence of reduced 331 in vitro growth of pRBCs that were exposed to antibodies and active complement 332 ( Figure 2C ). Therefore, these data suggest that while there is fixation of C1q and 333 activation of C3 on the surface of pRBCs, there is limited formation of MAC or 334 complement-mediated lysis of pRBCs. 335 336

PfEMP1 has been implicated as a major antigen on the surface of pRBCs (48). 338 Therefore, we quantified its significance as a target of complement fixing antibodies. 339

We first established that antibody samples from PNG pregnant women fixed C1q on 340 recombinant VAR2CSA DBL5 ( Figure 3A ) and DBL3 domains ( Figure 3B ) at 341 significantly higher levels compared to malaria non-exposed donors (both P<0.001); 342 57% (171/302) and 56% (170/302) of PNG samples were classified as positive for 343

C1q fixation against DBL5 and DBL3 respectively. C1q fixation on DBL5 and DBL3 344

were moderately correlated with each other (Spearman r=0.63, P<0.001) and with 345

C1q fixation on pRBCs (Spearman r=0.54 and r=0.50 respectively, P<0.001). C3 346 fixation on DBL5, tested for a subset of 35 randomly selected PNG antibody samples, 347 was also significantly higher compared to malaria non-exposed donor samples (n=10) 348 (P <0.001) ( Figure 3C ) and strongly correlated with C1q fixation (Spearman r=0.89, 349 P<0.001) ( Figure 3D ). C3 fixation on DBL5 versus pRBCs were moderately 350 correlated (Spearman r=0.48, P=0.003). Using a subset of 35 randomly selected PNG 351 samples and 10 Australian malaria non-exposed donor samples, we confirmed that 352 acquired antibodies fixed complement to a different allele of VAR2CSA, DBL5 353 Figure S3) . The correlation between complement C3 354 fixation to the 2 different alleles DBL5 (3D7) and DBL5 (7G8) was 0.61, p<0.001. 355

We next evaluated antibody-mediated complement fixation on the pRBC 356 surface using a CS2 P. falciparum isolate that was genetically-modified to reduce 357 . CC-BY-NC 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted June 5, 2021. ; https://doi.org/10.1101/2021.06.02.21258254 doi: medRxiv preprint PfEMP1 surface expression; CS2 skeleton-binding protein 1 knockout (CS2-358 SBP1KO) (49). We found that C1q ( Figure 3E ) and C3 ( Figure 3F ) fixation were 359 greatly reduced on the surface of CS2-SBP1KO pRBCs compared to CS2-parental 360 pRBCs (P=0.016), indicating PfEMP1 as the major target of antibodies. Median C1q 361 ( Figure 3E ) and C3 ( Figure 3F) Figure S4) and 368 moderately correlated with C1q fixation on DBL5 and DBL3 (Table 2) . (Tables S2 & S3) . However, 392

it is important to note that statistical power is reduced in these sub-group analyses. 393

Among women who were not parasitemic at enrolment, there was no significant 394 association between antibody complement fixation on pRBCs (C1q or C3) and risk of 395 placental infection. C1q is the major factor important in enhancing binding inhibition by antibodies 417 ( Figure 5B) . 418 . CC-BY-NC 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted June 5, 2021. ; https://doi.org/10.1101/2021.06.02.21258254 doi: medRxiv preprint

Antibodies are thought to mediate protection against placental infection by P. 420 falciparum. However, the mechanisms mediating immunity are not fully understood. 421

Here, we reveal a new mechanism in immunity to malaria in pregnancy, which may 422 contribute to reducing the risk of placental parasitemia. We show that acquired 423 antibodies among pregnant women can mediate complement fixation on placental-424 binding P. falciparum pRBCs. Using genetically-modified P. falciparum pRBCs and 425 recombinant antigens, we found that antibody-mediated complement-fixation 426 predominantly targeted PfEMP1 (VAR2CSA) expressed on the surface of pRBCs. 427

Importantly, higher complement-fixing antibodies were associated with a reduced risk 428 of placental parasitaemia and resulted in enhanced inhibition of pRBC binding to 429 CSA suggesting this mechanism contributes to immunity to MiP. 430

We demonstrated the ability of acquired antibodies to fix and activate 431 complement using several approaches. Antibodies could fix C1q, the first step in the 432 classical pathway and C3, indicating complement activation. This was demonstrated 433 by labelling complement components on pRBCs by flow cytometry, and by western 434 blotting of pRBC protein extracts, as well as using recombinant VAR2CSA domains. 435

Complement fixation among antibodies from malaria-exposed pregnant women was 436 significantly higher than antibodies from non-exposed donors and correlated with IgG 437 reactivity. Furthermore, complement-fixing antibodies were generally higher among 438 multigravid than primigravid women, consistent with the reported acquisition of 439 immunity to MiP. Consistent with data from Africa (15, 16), IgG1 (80% 440 seropositivity) and IgG3 (37% seropositivity) dominated responses in PNG women 441 and moderately correlated with complement-fixing antibodies. We observed some 442 complement C3 fixation when using malaria non-exposed donor antibodies or non-443 opsonized pRBCs, suggesting some activation via the antibody-independent alternate 444 pathway; however, complement fixation was always much higher in the presence of 445 antibodies from pregnant women supporting a greater role for the classical pathway. 446

Significant polymorphisms occur in VAR2CSA, which can impact on binding of 447 acquired IgG (47, 50, 51). Previously we established that the CS2 isolate used in this 448 study is well recognised by malaria-exposed pregnant women in PNG, and IgG 449 reactivity to CS2 pRBCs strongly correlated with IgG reactivity to a placental-binding 450 clinical isolate from PNG (37). 451 . CC-BY-NC 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted June 5, 2021. Two earlier studies showed evidence of complement fixation by antibodies 461 among non-pregnant individuals on pRBCs with known binding to CD36 and ICAM-462 1 endothelial receptors (26, 27). However, the acquisition and targets of these 463 antibodies were not assessed, nor were associations with protective immunity. In 464 contrast to our findings, one recent study evaluated purified IgG from a pool of 465 plasma samples from malaria-exposed pregnant Ghanaian women and a VAR2CSA 466 monoclonal antibody (52). Complement fixation was detected on recombinant 467

VAR2CSA, but not on the surface of pRBCs. However, the acquisition and functions 468 of antibodies, or associations with clinical features or outcomes were not assessed. 469

The detection of limited complement fixation on pRBCs in that study, compared to 470 our results, may reflect differences in the sensitivity of different assays and the 471 reagents used, and in that study the authors used 1% fresh serum as a source of 472 complement. Additionally, using a pool of samples would contribute to a reduced 473 ability to detect complement-fixing antibodies if the prevalence of such antibodies is 474 low in the population. is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted June 5, 2021. ; https://doi.org/10.1101/2021.06.02.21258254 doi: medRxiv preprint transmission is heterogenous only a proportion of the population are exposed to 486 infection. One approach to address this is to stratify the analysis of associations with 487 protection by infection status at baseline (53). Other studies investigating associations 488 between antibodies and protection against malaria in children (54, 55) or pregnant 489 women (56) have used this approach, revealing protective associations only in the 490 infected group who have documented exposure to malaria. A potential limitation of 491 our study is that women commenced in the study in mid-pregnancy (most were 492 second trimester), even though we recruited pregnant women at their first antenatal 493 clinic visit. Presenting to first antenatal clinic in the second trimester is commonly 494 observed in PNG (57, 58). This means that we cannot fully account for malaria 495 exposure history in their pregnancy. A further limitation is that placental tissue was 496 available only for 77% women. However, we did not observe major difference in 497 demographic or clinical features of women who did or did not have placental tissue 498 collected (29). Ideally, a future study would enrol women early in pregnancy with 499 frequent follow-up during pregnancy and employ a larger sample size to further 500 investigate associations we observed here, including the relative roles of different 501 antibodies and the influence of gravidity. 502

We found that complement significantly enhanced inhibition of pRBC 503 adhesion to CSA by acquired antibodies, suggesting a potential mechanism mediating 504 protective effects of complement-fixing antibodies. Additionally, it is well established 505 that complement enhances phagocytosis through interactions with complement 506 receptors expressed on monocytes and neutrophils, particularly via C3b (19, 59). 507 Therefore, higher complement fixation by antibodies is likely to contribute to 508 enhanced clearance of pRBCs. Anti-VAR2CSA antibodies have been proposed to 509 protect against MiP possibly through inhibition of P. falciparum binding in the 510 placenta along with enhanced phagocytosis (11, 15, 60, 61) . Previous studies, 511 however, have largely examined these functional mechanisms in the absence of 512 complement. The initial step of the classical pathway of complement activation 513 involving C1q appeared essential and sufficient in mediating inhibition enhancement. 514

However, it should be noted that there is evidence that C1q can bind directly to CSA 515 is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted June 5, 2021. ; https://doi.org/10.1101/2021.06.02.21258254 doi: medRxiv preprint neutralization, without the need for additional complement components (20, 63). In 520 contrast to protective roles of antibody-complement interactions, widespread 521 complement activation has also been implicated in MiP pathology. Higher plasma 522

C5a was associated with adverse birth outcomes including low birth weight, fetal 523 growth restriction and preterm birth (28). Excess C5a is thought to mediate 524 pathogenesis via inflammation and skews angiogenic profiles critical for normal 525 placental vascularization and development towards an anti-angiogenic profile that 526 promotes fetal growth restriction (28). Therefore, there is a balance between antibody-527 mediated complement activity on the surface of pRBCs that may be protective, versus 528 widespread systemic complement activation that may be detrimental (64). We 529

propose that acquisition of antibodies with the right specificity and properties leads to In conclusion, we have generated significant new evidence supporting 551 antibody-complement interactions in immunity against MiP. These findings provide 552 . CC-BY-NC 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted June 5, 2021. ; https://doi.org/10.1101/2021.06.02.21258254 doi: medRxiv preprint novel insights into the mechanisms mediating immunity in pregnancy and inform 553 approaches to develop vaccines or other interventions against MiP. 554 . CC-BY-NC 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted June 5, 2021. 

The datasets used and/or analysed during the current study are available from the 563 corresponding author on reasonable request. 564 565

The authors declare no competing interests. CC-BY-NC 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted June 5, 2021. ; https://doi.org/10.1101/2021.06.02.21258254 doi: medRxiv preprint

The authors thank the mothers and their families for participation in this study, the 589 staff of Alexishafen Health Centre for their enthusiastic cooperation, and staff of the 590 PNG Institute for Medical Research, particularly Francesca Baiwog (now deceased), 591

Prof. Peter Siba and Prof. Willie Pomat. We also thank Prof. Joseph Smith (Seattle 592

Children's Research Institute, Seattle, USA) for his generous contribution of all the 593 VAR2CSA recombinant proteins and Dr. Paul Gilson for the anti-HSP70 antibody. 594 . CC-BY-NC 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted June 5, 2021. ; https://doi.org/10.1101/2021.06.02.21258254 doi: medRxiv preprint absence of either PNG or Australian malaria non-exposed antibodies and NS or HIS. 629

Comparisons were made using Mann-Whitney U-test. CS2 pRBCs binding to CSA was tested following opsonisation of CS2 pRBCs with 682 10% PNG antibody pool in the presence of NS, purified human C1q or C1q-depleted 683 human serum. A malaria non-exposed antibody pool was used as a negative control. 684

Plots represent median and interquartile range of pRBCs bound to CSA per mm 2 . 685

Each data point represents an independent assay with 3-6 independent assays 686 performed for each condition. 687 688 . CC-BY-NC 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted June 5, 2021. 

. CC-BY-NC 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted June 5, 2021. is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted June 5, 2021. ; https://doi.org/10.1101/2021.06.02.21258254 doi: medRxiv preprint . CC-BY-NC 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted June 5, 2021. ; https://doi.org/10.1101/2021.06.02.21258254 doi: medRxiv preprint . CC-BY-NC 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted June 5, 2021. ; https://doi.org/10.1101/2021.06.02.21258254 doi: medRxiv preprint . CC-BY-NC 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted June 5, 2021. ; https://doi.org/10.1101/2021.06.02.21258254 doi: medRxiv preprint 2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1 

Selective accumulation of 704 mature asexual stages of Plasmodium falciparum-infected erythrocytes in the 705 placenta

Placental pathologic changes in malaria. A 707 histologic and ultrastructural study

Chondroitin sulfate 709 A is a cell surface receptor for Plasmodium falciparum-infected erythrocytes

Adherence of Plasmodium falciparum to chondroitin 712 sulfate A in the human placenta

Adhesion of Plasmodium falciparum-infected erythrocytes to hyaluronic acid in 715 placental malaria

Role 717 of nonimmune IgG bound to PfEMP1 in placental malaria

720 Evidence for the involvement of VAR2CSA in pregnancy-associated malaria

Pregnancy-725 specific malarial immunity and risk of malaria in pregnancy and adverse birth 726 outcomes: a systematic review

HIV impairs 730 opsonic phagocytic clearance of pregnancy-associated malaria parasites

733 PRIMVAC vaccine adjuvanted with Alhydrogel or GLA-SE to prevent placental 734 malaria: a first-in-human, randomised, double-blind, placebo-controlled study

First-in-human, Randomized, Double-blind Clinical Trial of 738 Differentially Adjuvanted PAMVAC, A Vaccine Candidate to Prevent Pregnancy-739 associated Malaria

Placental malaria induces variant-specific antibodies of the cytophilic 742 subtypes immunoglobulin G1 (IgG1) and IgG3 that correlate with adhesion 743 inhibitory activity

Effects of pregnancy 745 and intensity of Plasmodium falciparum transmission on immunoglobulin G 746 subclass responses to variant surface antigens

IgG subclasses and allotypes: from 749 structure to effector functions

751 Complement in malaria immunity and vaccines

Complement 754 System Part I -Molecular Mechanisms of Activation and Regulation

757 Complement protein C1q reduces the stoichiometric threshold for antibody-758 mediated neutralization of West Nile virus

Human antibodies fix complement to inhibit Plasmodium falciparum invasion of 761 erythrocytes and are associated with protection against malaria

Targets 764 of complement-fixing antibodies in protective immunity against malaria in 765 children

Human 767 antibodies activate complement against Plasmodium falciparum sporozoites, and 768 are associated with protection against malaria in children

The complement system contributes to functional antibody-772 mediated responses induced by immunization with Plasmodium falciparum 773 malaria sporozoites

Complement-mediated lysis of Plasmodium falciparum gametes by malaria-776 immune human sera is associated with antibodies to the gamete surface antigen 777 Pfs230

Host cell factor CD59 restricts complement lysis of Plasmodium falciparum-780 infected erythrocytes

Complement activation by the 782 surface of Plasmodium falciparum infected erythrocytes

Complement activation and the resulting placental vascular insufficiency drives 786 fetal growth restriction associated with placental malaria

Risk 789 factors for malaria and adverse birth outcomes in a prospective cohort of 790 pregnant women resident in a high malaria transmission area of Papua New 791 Guinea

Placental 793 malaria. I. Pathological classification

Placental 795 malaria. II. A semi-quantitative investigation of the pathological features

A human complement receptor 1 polymorphism that reduces Plasmodium 799 falciparum rosetting confers protection against severe malaria

Human red blood cell polymorphisms and malaria

Minimal 804 association of common red blood cell polymorphisms with Plasmodium 805 falciparum infection and uncomplicated malaria in Papua New Guinean school 806 children

808 Broad analysis reveals a consistent pattern of var gene transcription in 809 Plasmodium falciparum repeatedly selected for a defined adhesion phenotype

Clustering of integral membrane 847 proteins of the human erythrocyte membrane stimulates autologous IgG binding, 848 complement deposition, and phagocytosis

Evaluation of the antigenic diversity of placenta-binding Plasmodium falciparum 851 variants and the antibody repertoire among pregnant women

Targets of antibodies against Plasmodium falciparum-infected erythrocytes in 855 malaria immunity

Skeleton-857 binding protein 1 functions at the parasitophorous vacuole membrane to traffic 858 PfEMP1 to the Plasmodium falciparum-infected erythrocyte surface

Structural polymorphism and diversifying selection on the pregnancy malaria 862 vaccine candidate VAR2CSA

Antigenic differences and conservation among placental Plasmodium falciparum-865 infected erythrocytes and acquisition of variant-specific and cross-reactive 866 antibodies

Evasion of Classical Complement Pathway Activation on Plasmodium 869 falciparum-Infected Erythrocytes Opsonized by PfEMP1-Specific IgG. Frontiers in 870 immunology

Breadth 872 and magnitude of antibody responses to multiple Plasmodium falciparum 873 merozoite antigens are associated with protection from clinical malaria

Identifying Immune Correlates of Protection Against Plasmodium falciparum 877 Through a Novel Approach to Account for Heterogeneity in Malaria Exposure

886 Risk factors and knowledge associated with high unintended pregnancy rates 887 and low family planning use among pregnant women in Papua New Guinea

Mycoplasma genitalium and Other Reproductive Tract Infections in Pregnant 891 Women

CD14(hi)CD16+ monocytes phagocytose antibody-opsonised Plasmodium 894 falciparum infected erythrocytes more efficiently than other monocyte subsets, 895 and require CD16 and complement to do so

Antibodies to variant surface antigens of Plasmodium falciparum-infected 898 erythrocytes are associated with protection from treatment failure and the 899 development of anemia in pregnancy

Antibodies that inhibit Plasmodium falciparum 901 adhesion to chondroitin sulfate A are associated with increased birth weight and 902 the gestational age of newborns

The C1q inhibitor in serum is a 904 chondroitin 4-sulfate proteoglycan

Complement Factors in COVID-19 Therapeutics and 906 Vaccines

Loss of complement regulatory proteins on uninfected erythrocytes in 909 vivax and falciparum malaria anemia

A 911 single point in protein trafficking by Plasmodium falciparum determines the 912 expression of major antigens on the surface of infected erythrocytes targeted by 913 human antibodies

Induction of strain-916 transcendent antibodies to placental-type isolates with VAR2CSA DBL3 or DBL5 917 recombinant proteins

Preclinical immunogenicity and safety of the cGMP-grade placental malaria 920 vaccine PRIMVAC

is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted June 5, 2021. ; https://doi.org/10.1101/2021.06.02.21258254 doi: medRxiv preprint