Differential expression of plasma proteins in pregnant women exposed toPlasmodium falciparummalaria

27 In sub-Saharan Africa, pregnant women are at greater risk of malaria infection than non -pregnant 28 adult women. The infection may lead to pregnancy-associated malaria (PAM) because of the sequestration 29 of Plasmodium falciparum-infected erythrocytes in the placental intervillous space. Although there are 30 several tools for diagnosing malaria infection during pregnancy , including blood smear microscopic 31 examination, rapid diagnostic tests, and PCR, there are no tools for detecting placental infection and, by 32 extension, any dysfunction associated with PAM. Thus, PAM, specifically placental infection, can only be 33 confirmed via postnatal placental histopathology. Therefore, there is an urgent need for s pecific serum 34 biomarkers of PAM. Here, we used the high throughput proximity extension assay to screen plasma from 35 malaria-exposed pregnant women for differentially expressed proteins that can predict PAM or adverse 36 malaria outcomes. Such biomarkers may also elucidate the pathophysiology of PAM. We observed that the 37 IgG Fc receptor IIb (Uniprot ID P31994) and HO-1 (P09601) are consistently highly expressed in malaria-38 positive samples compared to samples from malaria-negative pregnant women. On the contrary , NRTN 39 (Q99748) and IL-20 (Q9NYY1) were differentially expressed in the malaria -negative women . IL-20 40 exhibited the highest discriminatory power (AUC = 0.815), indicating a strong association with malaria 41 status. These proteins should be considered for further evaluation as biomarkers of malaria-induced 42 placental dysfunction in pregnant women. 43 44

Pregnancy-associated malaria, malaria in pregnancy, biomarkers, falciparum m alaria, 45 proteomics, proximity extension assay 46 . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted May 3, 2024. ; https://doi.org/10.1101/2024.05.01.24306614doi: medRxiv preprint 3

47 Pregnant women in most tropical areas are at an increased risk of malaria because of the placental 48 sequestration of Plasmodium falciparum (P. falciparum )-infected red blood cells, which can lead to 49 stillbirth and low birth weight [1]. Despite the availability of numerous diagnostic tools for detecting 50 malaria during pregnancy, there is a lack of a definitive test for determining actual placental infection and 51 dysfunction within the context of pregnancy -associated malaria (PAM) [2]. PAM has several similarities 52 with pre-eclampsia, for which several biomarkers of placental stress and dysfunction have been identified 53 [3,4]. While some of these biomarkers are inflammatory markers that lack specificity, others , such as s -54 endoglin and sFlt1, which are elevated in peripheral blood during pre -eclampsia, appear more specific to 55 placental dysfunction [5]. The predictive value of such biomarkers should be evaluated in the context of 56 PAM to determine their effectiveness as indicators of PAM-associated placental infection and dysfunction 57 [6]. This is crucial for enhancing the diagnosis of typically sub-microscopic malaria infections and malaria-58 induced placental dysfunction, thereby facilitating timely intervention in pregnant women. Indeed, 59 comprehension of disease biology and the interaction between the malaria parasite and the human host is 60 crucial for developing diagnostic, prognostic, and uncovering predictive biomarkers for PAM [7]. 61 Elucidating the expression patterns of plasma proteins can provide insights into how PAM affects placental, 62 maternal, or fetal wellbeing [6,8]. 63 64 The rapid development of proteomics platforms , such as multiplex proximity extension assay 65 (PEA), offers highly effective, high-throughput strategies for studying malaria -associated molecular 66 pathophysiological changes [9]. PEA technology enables simultaneous measurement of a large number of 67 proteins using minute amount of sample [7,10]. Studies using this approach have demonstrated its clinical 68 utility in identifying important biomarkers for various diseases, including cancer, long COVID, ischemic 69 stroke, and early pregnancy bleeding [10–15]. In this study, we used PEA to screen sera from pregnant, 70 malaria-exposed women to identify differentially expressed proteins. Using this strategy, we identified four 71 proteins significantly differentially expressed in samples from malaria-infected pregnant women compared 72 with non-infected pregnant counterparts , suggesting that their alteration is P. falciparum -driven. Such 73 proteins can potentially develop into biomarkers of PAM or adverse malaria outcomes. 74 . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted May 3, 2024. ; https://doi.org/10.1101/2024.05.01.24306614doi: medRxiv preprint 4

75 76 Study design & population. 77 78 This study used serum samples collected from women between 12 and 18 weeks of gestation. These 79 women were enrolled in the study during their routine antenatal clinic visits at Webuye County Hospital in 80 Bungoma County, Kenya, a region characterized by high malaria transmission [16]. The study protocol and 81 the use of this well-characterized biobank of samples were approved by the Independent Ethics Research 82 Committee (IERC) of Mount Kenya University (No# MKU/IERC/0543, MKU/IERC/2461). The enrolled 83 cohort consisted of women undergoing routine intermittent preventive malaria treatment in pregnancy 84 (IPTp) with sulfadoxine -pyrimethamine (IPTp -SP) [17]. All study participants gave w ritten informed 85 consent before joining the study. The study protocol complied with the International Conference on 86 Harmonization Good Clinical Practices and the Declaration of Helsinki guidelines [18]. Serum samples 87 were collected 4 weeks after IPTp administration and frozen immediately, and then transported on dry ice 88 to the Centre for Malaria Elimination laboratories at Mount Kenya University and immediately stored in a 89 -80°C deep freezer without cold-chain interruption. Each serum sample was matched with corresponding 90 anonymized demographic and pregnancy outcome data. Data on pregnancy progress and outcomes were 91 collected during the scheduled and unscheduled visits. Malaria diagnosis was performed using both 92 microscopy and a ParaCheck Rapid Diagnostic Kit (Orchid Biomedical Systems, India). 93 94 Protein abundance measurement by multiplex proximity extension assay 95 Multiplex PEA technology was used to analyze protein expression in the sera samples using the 96 Olink Inflammation and Cardiovascular II protein panels (Olink ™ Proteomics, Uppsala Sweden, 97 https://olink.com), loading 1µL sample/panel. These panels were selected since they target inflammatory 98 and vascular pathophysiological proteomic signatures, which were hypothesized to be altered by PAM. In 99 PEA, each target protein is recognized by two antibodies attached to single-stranded DNA oligonucleotide 100 [11]. When a pair of antibodies binds to the same target protein, the conjugated DNA oligonucleotides are 101 brought in proximity and hybridize to each other. Subsequently, the DNA molecules are extended to form 102 a double-stranded DNA as template for quantitative PCR amplification [11]. The results are presented as 103 normalized protein expression (NPX) values, which are arbitrary units on a log2 scale , where a one-unit 104 increase in NPX corresponds to a two-fold increase in protein concentration. The limit of detection (LOD) 105 for each protein was determined based on three times the standard deviation (SD) above the NPX value of 106 the negative controls in each run. In this study, a total of 183 different proteins were measured, where each 107 . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted May 3, 2024. ; https://doi.org/10.1101/2024.05.01.24306614doi: medRxiv preprint 5 panel included 92 proteins involved in diverse biological processes and 4 technical controls. Analyses were 108 performed at the Clinical Biomarker Facility at SciLifeLab, Uppsala University. 109 110 Data analysis 111 Data analyses were performed using R software (www.r -project.org). Normalized protein 112 expression (NPX) values were used for analysis because they tend to follow a normal distribution. Samples 113 from malaria infected group and the controls were randomized within the plates to minimize variability 114 between plates. The inter-plate variation was further normalized for each protein in each plate by adding a 115 z-score factor, which was calculated as follows: factor = (actual value - median of all samples) / standard 116 deviation. To compare malaria cases and controls, NPX values were adjusted if a significant effect (adjusted 117 p value < 0.05) f or age on protein levels was found by linear regression in both the control and malaria 118 groups. 119 For univariate analysis, the difference in protein levels between the malaria and non-malaria groups 120 was examined using linear regression, using age and gravida as covariates. The dependent variable was 121 NPX, and the independent variables were malaria status. The likelihood ratio test was used to assess the 122 significance of the difference. The statistical significance of the differences between two groups was 123 determined for continuous variables using the Mann–Whitney–Wilcoxon test. For categorical variables, the 124 Chi-square or Fisher's exact test was used. The analysis of variance (ANOVA) test was used to compare 125 more than two groups. The paired Mann–Whitney–Wilcoxon test was used to compare diff erent malaria 126 groups. Spearman's rank rho was used to test the correlation coefficients or collinearity between each 127 protein marker and other continuous variables. The Bonferroni–Dunn post hoc test was used to correct for 128 multiple testing by adjusting the p-values for false discovery rate. A difference was considered significant 129 if the q-value (adjusted p-value) was less than 0.05. 130 Additionally, elastic-net penalized logistic regression (ENLR) was used to identify a combination 131 of analytes that could improve the discrimination between cases and controls. ENLR applies a penalty to 132 the regression coefficients and finds groups of correlated variables. The optimal penalization proportion α 133 was determined using grid search with a 10-fold cross-validation. The optimal tuning parameter λ was 134 determined as the mean value of 100 iterative lambda values that minimized the model's deviance. The 135 regression coefficients were used to assess the contribution of individual proteins to the case -control 136 discrimination. The ENLR model was estimated using the R package glmnet [19]. The regression 137 coefficients for the selected proteins were then calculated by rerunning ENLR with only proteins with non-138 zero coefficients after 10 repetitions. To determine the number of proteins required to predict PAM, ROC 139 . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted May 3, 2024. ; https://doi.org/10.1101/2024.05.01.24306614doi: medRxiv preprint 6 curves were generated while adding one more protein at a time and then compared with the first ROC curve. 140 This process was repeated until none of the ROC curves showed significant improvement. 141

142 143 Protein reactivity 144 PEA is optimal for quantification of the levels of soluble plasma proteins in small sample volumes. 145 Here, to identify proteins that might indicate P. falciparum infection during pregnancy, we probed the Olink 146 Target 96 Inflammation and Olink Target 96 Cardiovascular II Panels (Table S1) using serum samples 147 collected from 50 women, who were sampled at three different timepoints during pregnancy (Table 1). The 148 initial sampling was performed at 12–18 weeks of gestation and two more samples were collected one and 149 two months before delivery. Twelve (12) women (median age: 24 years, range: 20–29 years) had malaria 150 infection that was detectable using a rapid diagnostic test. The women in the different groups did not differ 151 significantly in terms of age, gravidity, hemoglobin levels, and gestational age. 152 153 154 Table 1. Characteristics of the women assessed in this study. 155 Characteristics Malaria No Malaria Overall P value (n=12) (n=38) (n=50) Age in years at enrollment, mean (mean, range) 24.0 (20.0 - 29.0) 27.3 (19.0 - 42.0) 26.9 (19.0 - 42.0) 0.7775 Gestation in weeks (mean, range)a 18.7 (10.0 - 26.0) 14.2 (4.0 - 28.0) 21.8 (4.0 - 28.0) 0.005 Gravida (range) 1.8 (1.0 - 3.0) 2.5 (1.0 - 5.0) 2.3 (1.0 - 5.0) 0.07 Hemoglobin (mean, IQR, g/dL) a 11.5 (7.6 -13.6) 12.8 (9.6 - 15.8) 12.5 (7.6 - 15.8) 0.011 Random sugar Levels (mean, IQR, mmol/L) 5.4 (4.7 - 6.4) 4.9 (4.0 - 6.5) 5.0 (4.0-6.5) 0.10 156 a Statistically significant between the groups (P<0.05 Mann-Whitney-Wilcoxon test; followed by 157 Bonferroni–Dunn post hoc test); =malaria was defined as any parasitemia with or without fever; 158 IQR = interquartile range; All participants were of the same ethnicity.159 . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted May 3, 2024. ; https://doi.org/10.1101/2024.05.01.24306614doi: medRxiv preprint 7 Protein reactivity and the characteristic of the sampled women 160 Protein reactivity to the proteomic panels was measured at three different visits per participant. 161 The pattern of expression profiles varied considerably among the volunteers, with most proteins being 162 present at different time points (first, second and third trimester) for most women. However, in about 163 10% of the women, less than 20 proteins were detected (Figure 1). 164 165 166 167 Figure 1: The pattern of plasma proteins expression among pregnant women volunteers 168 assessed from western Kenya. 169 170 To assess if there was any relationship between serum protein levels and participant 171 characteristics, such as age, gravida, and gestation al age, we conducted a correlation analysis to each 172 of these parameters against the NPX for each protein. Although there was no correlation between 173 protein levels and participant s’ age, overall plasma protein levels were higher in samples from 174 primigravida (first-time pregnant) women when compared with multigravida women (who have had 2–175 5 pregnancies), although the difference did not reach statistical significance (Table S2). 176 We further assessed the relationship between markers of morbidity, including hemoglobin and 177 random blood sugars, and expressed plasma proteins. We observed that 19 proteins were significantly 178 negatively correlated with hemoglobin levels , and 9 proteins with random blood sugar levels 179 (unadjusted P < 0.05). However, the correlations did not remain significant after False Discovery Rate 180 (FDR) adjustments. Our findings suggest that these proteins may have a role in malaria infection 181 or that they may be involved in the development and/or progression of pregnancy 182 complications, such as anemia and/or gestational diabetes. 183 184 Serum reactivity [0%, 10%](10%, 20%](20%, 30%](30%, 40%](40%, 50%](50%, 60%](60%, 70%](70%, 80%](80%, 90%](90%, 100%] Number of proteins 0 20 40 60 80 100 120 140 160 Olink Target 96 Cardiovascular II and Inflammation . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted May 3, 2024. ; https://doi.org/10.1101/2024.05.01.24306614doi: medRxiv preprint 8 Differential protein expression in malaria vs no malaria women 185 Next, we sought to identify differential protein expression in malaria vs non -malaria women. 186 This analysis showed that 48 p roteins were significantly differentially expressed in malaria vs no 187 malaria women, at least in one of the three clinic visits ( Table S2 ). Notably, IgG Fc receptor IIb 188 (Uniprot ID : P31994) and HO -1 (P09601) were consistently highly expressed in malaria -positive 189 samples when compared with malaria-negative samples, while NRTN (Q99748) and IL-20 (Q9NYY1) 190 were consistently highly expressed in sera from malaria-negative women (Figure 2). This data suggest 191 that malaria infection may downregulate or upregulate various proteins. 192 193 194 Figure 2: Differential protein expression in malaria-infected vs non malaria-infected women. 195 The title represent protein Uniprot acronym and ID. Relative PEA counts are derived from 196 NPX/ LOD to allow protein-to-protein comparison. 197 198 We therefore sought to assess if the levels of these proteins fluctuated during the antepartum 199 period using samples collected at different pregnancy timepoints. Except for IL-20, whose serum levels 200 reduced during the follow-up period, there was no significant change in the other proteins, suggesting 201 that the initial induction by malaria infection did not alter the levels of these proteins during the follow-202 up period. The top differentially expressed proteins are presented in Figure 3, and the full protein list 203 is presented in Supplementary Tables S3. 204 205 206 1.5 2.0 2.5 IgG.Fc.receptor.II.b_P31994 Malaria in pregnancy status PEA counts Malaria No malaria 8.5 9.0 9.5 10.0 10.5 HO.1_P09601 Malaria in pregnancy status PEA counts Malaria No malaria 0.5 1.0 1.5 NRTN_Q99748 Malaria in pregnancy status PEA counts Malaria No malaria 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 IL.20_Q9NYY1 Malaria in pregnancy status PEA counts Malaria No malaria . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted May 3, 2024. ; https://doi.org/10.1101/2024.05.01.24306614doi: medRxiv preprint 9 207 Figure 3: Kinetics of protein expression in malaria-infected vs non-malaria-infected women at 208 three time points during the pregnancy – 1st (red), 2nd (green) and 3rd Visit (blue). * Represent 209 statistical difference between hospital visits, (Kruskal Wallis, P < 0.005). 210 211 Protein levels could distinguish malaria cases vs controls 212 We also assessed whether a combination of serum proteins could distinguish malaria cases vs 213 controls. This analysis revealed varying predictive performance among the four proteins we examined 214 in relation to malaria status. IL-20 (Q9NYY1) protein exhibited the highest discriminatory power (AUC 215 = 0.815), indicating a strong association with malaria status (Figure 4) . This suggests that IL -20 216 expression levels serve as a promising biomarker for malaria diagnosis or prognosis. IgG Fc receptor 217 IIb (Uniprot ID: P31994) protein demonstrated moderate discriminatory ability (AUC = 0.717), 218 suggesting a notable but less pronounced association with malaria status compared to IL-20. Similarly, 219 HO-1 (P09601) protein also displayed moderate discriminatory power (AUC = 0.729), indicating some 220 association with malaria status, although not as strong as IL -20. NRTN (Q99748) protein, with the 221 lowest AUC value (AUC = 0.635), showed weaker discriminatory ability compared to the other proteins. 222 Although NRTN expression levels may still have some association with malaria status, the association 223 is less pronounced compared to IL-20, IgG Fc receptor IIb, and HO-1. 224 225 malaria no malaria 1.0 1.5 2.0 2.5 3.0 IgG.Fc.receptor.II.b_P31994 Malaria in Pregnancy status PEA counts 1 2 3 malaria no malaria 8.0 8.5 9.0 9.5 10.0 10.5 HO.1_P09601 Malaria in Pregnancy status PEA counts malaria no malaria 0.0 0.5 1.0 1.5 NRTN_Q99748 Malaria in Pregnancy status PEA counts malaria no malaria 0.0 0.2 0.4 0.6 0.8 IL.20_Q9NYY1 Malaria in Pregnancy status PEA counts * . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted May 3, 2024. ; https://doi.org/10.1101/2024.05.01.24306614doi: medRxiv preprint 10 226 227 Figure 4: ROC curves for different protein expression levels in malaria- vs non-malaria-228 infected women. Each protein has been assigned a colour code, as indicated in the legend. 229 230 231

232 Accurate tools for diagnosing malaria during pregnancy are urgently needed to guide effective clinical 233 interventions [20]. In this study, we aimed to identify potential markers of malaria infection by using 234 high-throughput PEA technology. We observed that IgG Fc receptor IIb (Uniprot ID, P31994), HO -1 235 (P09601), NRTN (Q99748) and IL-20 (Q9NYY1) were differentially expressed in sera from malaria-236 infected pregnant women vs from non-malaria-infected controls. Moreover, these proteins may also 237 highlight processes that underlie malaria pathophysiology during pregnancy. 238 239 IgG Fc receptor IIb (FCGR2B, Uniprot ID, P31994), a low affinity immunoglobulin gamma Fc 240 region receptor, has multiple isoforms in macrophages, lymphocytes, and IgG -transporting placental 241 epithelium [21]. It is the only inhibitory Fc receptor that controls many aspects of immune and 242 inflammatory responses [22]. Variations in the gene encoding this protein , such as loss -of-function 243 polymorphism, have implications for autoimmunity and infection, including severe malaria. FCGR2B 244 is involved in several processes, including Fc receptor-mediated immune complex endocytosis, nervous 245 system development, and immune responses [23]. Other studies have identified FCGR2B as a marker 246 of human metastatic melanoma, where its expression impairs the tumor susceptibility to FcgammaR -247 dependent innate effector responses [24]. With the availability of monoclonal antibodies against 248 FCGR2B [25], studies are needed to evaluate its potential as a biomarker of PAM. 249 250 . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted May 3, 2024. ; https://doi.org/10.1101/2024.05.01.24306614doi: medRxiv preprint 11 We also identified HO -1 (heme oxygenase-1) as abundantly expressed in PAM cases. HO-1 251 plays a crucial role in heme degradation and cellular responses to stress [26]. During pregnancy, HO-1 252 maintains a balanced immune response and supports healthy placental development [27]. Evidence 253 from a mouse model [27] indicates that it helps protect against oxidative stress and inflammation, which 254 are often elevated in pregnancy-associated conditions. We hypothesize that in the context of pregnancy-255 associated malaria, HO-1 may play a role in mitigating the adverse effects of the infection by reducing 256 oxidative stress and inflammation. 257 258 Two other proteins that decreased in PAM cases, Neurturin (NRTN) and Interleukin-20 (IL-259 20), deserve further evaluation. NRTN has been implicated in supporting healthy placental growth and 260 development. It is also involved in the regulation of blood vessel formation and trophoblast invasion, 261 which are critical for successful pregnancy outcomes [28]. IL-20 plays a critical role in the development 262 and maintenance of the placenta, as well as in the regulation of trophoblast invasion, angiogenesis and 263 vascular remodeling [29]. Thus, IL-20 downregulation during PAM might contribute to poor pregnancy 264 outcomes. Indeed, disruptions in IL -20 signaling have been associated with adverse pregnancy 265 outcomes [30]. In the context of pregnancy-associated malaria, IL-20 may modulate the inflammatory 266 response and placental immune tolerance, potentially impacting the severity and outcomes of the 267 infection. These molecules need further evaluation. 268 269 The choice of two Olink panels, Inflammation and Cardiovascular II, provided a wide repertoire 270 of proteins with a critical role during pregnancy or with diagnostic capability for assessing malaria 271 infection of the placenta. However, the panels could have excluded other potentially important 272 proteomic targets, and future studies should aim to include a more diverse range of proteins. This study 273 involved a relatively small sample size and was based on a single study center. Hence, these 274 observations warrant further validation using larger, multicenter malaria-in-pregnancy cohorts. Because 275 this study is correlative, future studies will also seek to experimentally validate the involvement of the 276 identified factors in malaria during pregnancy and the underlying mechanisms, which may uncover 277 additional diagnostic and therapeutic strategies. 278 279 In conclusion, this study identified four proteins that were differentially regulated between 280 malaria and non-malaria women. Through further investi gation, expression of these proteins could 281 potentially be validated for diagnosis of pregnancy-associated malaria. 282 . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted May 3, 2024. ; https://doi.org/10.1101/2024.05.01.24306614doi: medRxiv preprint 12

283 We appreciate the study volunteers from Webuye County Hospital and thank the research teams 284 from Mount Kenya University for their technical assistance in obtaining and processing the field samples. 285 286 Funding 287 This work was funded by InDevelops u-landsfond InDevelops u-landsfond grant to MKM and JG, 288 African Academy of Sciences to JG , and the Swedish Research Council under the grant 2020 -02258 to 289 MKM. BNK is an EDCTP Fellow under EDCTP2 programme supported by the European Union grant 290 number TMA2020CDF-3203-EndPAMAL. BNK has also received support from Terumo Life Science 291 Foundation. The funders had no role in study design, data collection and analysis, decision to publish, or 292 preparation of the manuscript. 293 294 Competing interests 295 The authors declare that the research was conducted in the absence of any commercial or financial 296 relationships that could be construed as a potential conflict of interest. 297 298 Authors contribution 299 BNK, MKM, MV and JG conceived and designed experiments. FMK, HW, JM, and RG 300 conducted experiments. BNK, RG, FMK, and HW analyzed the data. BNK, FMK, and JG wrote the 301 manuscript. All authors discussed and edited the manuscript. 302 303

304 [1] Rogerson SJ, Mwapasa V, Meshnick SR. Malaria in pregnancy: linking immunity and 305 pathogenesis to prevention. Am J Trop Med Hyg 2007;77:14–22. 306 [2] Brabin BJ. An analysis of malaria in pregnancy in Africa. Bull World Health Organ 307 1983;61:1005–16. 308 [3] Brabin BJ, Johnson PM. Placental malaria and pre-eclampsia through the looking glass 309 backwards? J Reprod Immunol 2005;65:1–15. 310 https://doi.org/https://doi.org/10.1016/j.jri.2004.09.006. 311 [4] Roberts JM, Lain KY. Recent Insights into the Pathogenesis of Pre-eclampsia. Placenta 312 2002;23:359–72. https://doi.org/https://doi.org/10.1053/plac.2002.0819. 313 [5] Jena MK, Sharma NR, Petitt M, Maulik D, Nayak NR. Pathogenesis of Preeclampsia and 314 . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted May 3, 2024. ; https://doi.org/10.1101/2024.05.01.24306614doi: medRxiv preprint 13 Therapeutic Approaches Targeting the Placenta. Biomolecules 2020;10. 315 https://doi.org/10.3390/biom10060953. 316 [6] Gueneuc A, Deloron P, Bertin GI. Usefulness of a biomarker to identify placental 317 dysfunction in the context of malaria. Malar J 2017;16:1–7. 318 https://doi.org/10.1186/s12936-016-1664-0. 319 [7] Acharya P, Garg M, Kumar P, Munjal A, Raja KD. Host-parasite interactions in human 320 malaria: Clinical implications of basic research. Front Microbiol 2017;8:1–16. 321 https://doi.org/10.3389/fmicb.2017.00889. 322 [8] Unger HW, Ome-Kaius M, Wangnapi RA, Umbers AJ, Hanieh S, Suen CSNLW, et al. 323 Sulphadoxine-pyrimethamine plus azithromycin for the prevention of low birthweight in 324 Papua New Guinea: A randomised controlled trial. BMC Med 2015;13:1–16. 325 https://doi.org/10.1186/s12916-014-0258-3. 326 [9] Wik L, Nordberg N, Broberg J, Björkesten J, Assarsson E, Henriksson S, et al. Proximity 327 Extension Assay in Combination with Next-Generation Sequencing for High-throughput 328 Proteome-wide Analysis. Mol Cell Proteomics 2021;20:100168. 329 https://doi.org/10.1016/j.mcpro.2021.100168. 330 [10] Dayon L, Cominetti O, Affolter M. Proteomics of human biological fluids for biomarker 331 discoveries: technical advances and recent applications. Expert Rev Proteomics 332 2022;19:131–51. https://doi.org/10.1080/14789450.2022.2070477. 333 [11] Assarsson E, Lundberg M, Holmquist G, Björkesten J, Thorsen SB, Ekman D, et al. 334 Homogenous 96-plex PEA immunoassay exhibiting high sensitivity, specificity, and 335 excellent scalability. PLoS One 2014;9. https://doi.org/10.1371/journal.pone.0095192. 336 [12] Allan-Blitz L-T, Akbari O, Kojima N, Saavedra E, Chellamuthu P, Denny N, et al. Unique 337 immune and inflammatory cytokine profiles may define long COVID syndrome. Clin Exp 338 Med 2023;23:2925–30. https://doi.org/10.1007/s10238-023-01065-6. 339 [13] Angerfors A, Brännmark C, Lagging C, Tai K, Månsby Svedberg R, Andersson B, et al. 340 Proteomic profiling identifies novel inflammation-related plasma proteins associated with 341 ischemic stroke outcome. J Neuroinflammation 2023;20:1–11. 342 https://doi.org/10.1186/s12974-023-02912-9. 343 [14] Davies MPA, Sato T, Ashoor H, Hou L, Liloglou T, Yang R, et al. Plasma protein 344 biomarkers for early prediction of lung cancer. EBioMedicine 2023;93:1–11. 345 https://doi.org/10.1016/j.ebiom.2023.104686. 346 [15] Guterstam YC, Acharya G, Schott K, Björkström NK, Gidlöf S, Ivarsson MA. Immune 347 cell profiling of vaginal blood from patients with early pregnancy bleeding. Am J Reprod 348 Immunol 2023;90:1–13. https://doi.org/10.1111/aji.13738. 349 [16] Platt A, Obala AA, MacIntyre C, Otsyula B, Meara WPO. Dynamic malaria hotspots in an 350 open cohort in western Kenya. Sci Rep 2018;8:1–11. https://doi.org/10.1038/s41598-017-351 . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted May 3, 2024. ; https://doi.org/10.1101/2024.05.01.24306614doi: medRxiv preprint 14 13801-6. 352 [17] Peters PJ, Thigpen MC, Parise ME, Newman RD. Safety and toxicity of 353 sulfadoxine/pyrimethamine: implications for malaria prevention in pregnancy using 354 intermittent preventive treatment. Drug Saf 2007;30:481–501. 355 https://doi.org/10.2165/00002018-200730060-00003. 356 [18] Dixon JRJ. The International Conference on Harmonization Good Clinical Practice 357 guideline. Qual Assur 1998;6:65–74. https://doi.org/10.1080/105294199277860. 358 [19] Fredriksson S, Gullberg M, Jarvius J, Olsson C, Pietras K, Gústafsdóttir SM, et al. Protein 359 detection using proximity-dependent DNA ligation assays. Nat Biotechnol 2002;20:473–360 7. https://doi.org/10.1038/nbt0502-473. 361 [20] Friedman JH, Hastie T, Tibshirani R. Regularization Paths for Generalized Linear Models 362 via Coordinate Descent. J Stat Softw 2010;33:1–22. https://doi.org/10.18637/jss.v033.i01. 363 [21] Ding XC, Incardona S, Serra-Casas E, Charnaud SC, Slater HC, Domingo GJ, et al. 364 Malaria in pregnancy (MiP) studies assessing the clinical performance of highly sensitive 365 rapid diagnostic tests (HS-RDT) for Plasmodium falciparum detection. Malar J 366 2023;22:1–15. https://doi.org/10.1186/s12936-023-04445-1. 367 [22] Stuart SG, Simister NE, Clarkson SB, Kacinski BM, Shapiro M, Mellman I. Human IgG 368 Fc receptor (hFcRII; CD32) exists as multiple isoforms in macrophages, lymphocytes and 369 IgG-transporting placental epithelium. EMBO J 1989;8:3657–66. 370 https://doi.org/10.1002/j.1460-2075.1989.tb08540.x. 371 [23] Smith KGC, Clatworthy MR. Europe PMC Funders Group Fc γ RIIB in autoimmunity 372 and infection : evolutionary and therapeutic implications 2014;10:328–43. 373 https://doi.org/10.1038/nri2762.Fc. 374 [24] Bournazos S, Woof JM, Hart SP, Dransfield I. Functional and clinical consequences of Fc 375 receptor polymorphic and copy number variants. Clin Exp Immunol 2009;157:244–54. 376 https://doi.org/10.1111/j.1365-2249.2009.03980.x. 377 [25] Cassard L, Cohen-Solal JFG, Fournier EM, Camilleri-Broët S, Spatz A, Chouaïb S, et al. 378 Selective expression of inhibitory Fcγ receptor by metastatic melanoma impairs tumor 379 susceptibility to IgG-dependent cellular response. Int J Cancer 2008;123:2832–9. 380 https://doi.org/10.1002/ijc.23870. 381 [26] Veri MC, Gorlatov S, Li H, Burke S, Johnson S, Stavenhagen J, et al. Monoclonal 382 antibodies capable of discriminating the human inhibitory Fcγ-receptor IIB (CD32B) from 383 the activating Fcγ-receptor IIA (CD32A): Biochemical, biological and functional 384 characterization. Immunology 2007;121:392–404. https://doi.org/10.1111/j.1365-385 2567.2007.02588.x. 386 [27] Ozen M, Zhao H, Kalish F, Yang Y, Jantzie LL, Wong RJ, et al. Inflammation-induced 387 alterations in maternal-fetal Heme Oxygenase (HO) are associated with sustained innate 388 . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted May 3, 2024. ; https://doi.org/10.1101/2024.05.01.24306614doi: medRxiv preprint 15 immune cell dysregulation in mouse offspring. PLoS One 2021;16:1–18. 389 https://doi.org/10.1371/journal.pone.0252642. 390 [28] Schumacher A, Zenclussen AC. Effects of heme oxygenase-1 on innate and adaptive 391 immune responses promoting pregnancy success and allograft tolerance. Front Pharmacol 392 2015;5:1–9. https://doi.org/10.3389/fphar.2014.00288. 393 [29] Morel L, Domingues O, Zimmer J, Michel T. Revisiting the Role of Neurotrophic Factors 394 in Inflammation. Cells 2020;9. https://doi.org/10.3390/cells9040865. 395 [30] Menon R, Ismail L, Ismail D, Merialdi M, Lombardi SJ, Fortunato SJ. Human fetal 396 membrane expression of IL-19 and IL-20 and its differential effect on inflammatory 397 cytokine production. J Matern Neonatal Med 2006;19:209–14. 398 https://doi.org/10.1080/14767050500440986. 399 401 . CC-BY-NC-ND 4.0 International licenseIt is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted May 3, 2024. ; https://doi.org/10.1101/2024.05.01.24306614doi: medRxiv preprint