Assessment of IgG3 as a serological exposure marker for Plasmodium vivax in areas with moderate–high malaria transmission intensity

A more sensitive surveillance tool is needed to identify Plasmodium vivax infections for treatment and to accelerate malaria elimination efforts. To address this challenge, our laboratory has developed an eight-antigen panel that detects total IgG as serological markers of P. vivax exposure within the prior 9 months. The value of these markers has been established for use in areas with low transmission. In moderate–high transmission areas, there is evidence that total IgG is more long-lived than in areas with low transmission, resulting in poorer performance of these markers in these settings. Antibodies that are shorter-lived may be better markers of recent infection for use in moderate–high transmission areas. Using a multiplex assay, the antibody temporal kinetics of total IgG, IgG1, IgG3, and IgM against 29 P. vivax antigens were measured over 36 weeks following asymptomatic P. vivax infection in Papua New Guinean children (n = 31), from an area with moderate–high transmission intensity. IgG3 declined faster to background than total IgG, IgG1, and IgM. Based on these kinetics, IgG3 performance was then assessed for classifying recent exposure in a cohort of Peruvian individuals (n = 590; age 3–85 years) from an area of moderate transmission intensity. Using antibody responses against individual antigens, the highest performance of IgG3 in classifying recent P. vivax infections in the prior 9 months was to one of the Pv-fam-a proteins assessed (PVX_125728) (AUC = 0.764). Surprisingly, total IgG was overall a better marker of recent P. vivax infection, with the highest individual classification performance to RBP2b1986-2653 (PVX_094255) (AUC = 0.838). To understand the acquisition of IgG3 in this Peruvian cohort, relevant epidemiological factors were explored using a regression model. IgG3 levels were positively associated with increasing age, living in an area with (relatively) higher transmission intensity, and having three or more PCR-detected blood-stage P. vivax infections within the prior 13 months. Overall, we found that IgG3 did not have high accuracy for detecting recent exposure to P. vivax in the Peruvian cohort, with our data suggesting that this is due to the high levels of prior exposure required to acquire high IgG3 antibody levels.


Supplementary
. The kinetics of total IgG against 30 P. vivax antigens following asymptomatic P. vivax infections in PNG children.
Total IgG antibody responses against 30 P. vivax antigens following asymptomatic P. vivax infections in PNG children (n=31) over 36 weeks were measured using a multiplex assay. Data is presented in RAU (in log 10). Each dot point shows the individual antibody response to each P. vivax antigen measured at each timepoint. Locally estimated scatterplot smoothing (LOESS) lines and 95% confidence intervals were plotted over time. Negative control (grey line) was the average antibody response of naïve individuals (n=274) and the grey ribbon was set as 2 SD from the average negative control. Anti-antigen antibody with LOESS-smoothed line that was seropositive at week 0 and became seronegative within 9 months are indicated by an asterix (*). Anti-antigen antibody with LOESSsmoothed line that was seropositive for at least 36 weeks is indicated by an up arrow (^). Figure S2. The kinetics of IgM against 32 P. vivax antigens following asymptomatic P. vivax infections in PNG children.

Supplementary
IgM antibody responses against 32 P. vivax antigens following asymptomatic P. vivax infections in PNG children (n=31) over 36 weeks, measured using a multiplex assay. Data are presented in RAU (in log 10). Each dot point shows the individual IgM antibody response to P. vivax antigens measured at each timepoint. LOESS-smoothed line and 95% confidence intervals were plotted over time. Negative control (grey line) was the average antibody response of naïve individuals (n=260) and the grey ribbon shows 2 SD from average negative control. No anti-antigen antibody with LOESS-smoothed line that was seropositive at week 0 and became seronegative within 9 months. Figure S3. The kinetics of IgG1 against 29 P. vivax antigens following asymptomatic P. vivax infections in PNG.

Supplementary
IgG1 antibody response against 29 P. vivax antigens following asymptomatic P. vivax infections in PNG children (n=31) over 36 weeks, measured using a multiplex assay. Data are presented in RAU (in log 10). Each dot point shows the individual antibody response to P. vivax antigens measured at each timepoint. LOESS-smoothed lines and 95% confidence intervals were plotted over time. Negative control (grey line) was the average antibody response of naïve individuals (n=248) and the grey ribbon shows 2 SD from average negative control. Anti-antigen antibody with LOESS-smoothed line that was seropositive at week 0 and became seronegative within 9 months are indicated by an asterix (*). Antiantigen antibody with LOESS-smoothed line that was seropositive for at least 36 weeks is indicated by an up arrow (^). Figure S4. The kinetics of IgG3 against 29 P. vivax antigens following asymptomatic P. vivax infections in PNG.

Supplementary
IgG3 antibody response against 29 P. vivax antigens following asymptomatic P. vivax infections in PNG children (n=31) over 36 weeks, measured using a multiplex assay. Data are presented in RAU (in log 10). Each dot point shows the individual antibody response to P. vivax antigens measured at each timepoint. LOESS-smoothed lines and 95% confidence intervals were plotted over time. Negative control (grey line) was the average antibody response of naïve individuals (n=256) and the grey ribbon shows 2 SD from average negative control. Anti-antigen antibody with LOESS-smoothed line that was seropositive at week 0 and became seronegative within 9 months are indicated by an asterix (*). Figure S5. Total IgG responses to 28 P. vivax antigens in Peruvian cohort based on time since last infected with P. vivax.

Supplementary
Total IgG responses to 28 P. vivax antigens were measured at the end of study of a Peruvian cohort (n=590)). Antibody responses were presented in relative antibody unit (log 10) in median with interquartile range. Individuals were grouped based on the last P. vivax infection status as measured by qPCR. Antigen-specific IgG were ordered based on decreasing seropositivity to IgG3 to reflect Fig  3. Anti-antigen antibody that was significantly higher in individuals with infections in the prior 9 months compared to individuals with no infections or infections >9 months ago are indicated by an asterix (*) (student's t-test, p<0.05). Anti-antigen antibody with consistent IgG levels above background, regardless of recency of infection are indicated by ^. Figure S6. Correlation of total IgG antibodies to 20 P. vivax antigens measured in Peruvian cohort using magnetic and non-magnetic bead in multiplex assay Total IgG antibodies to 20 P. vivax antigens were measured in Peruvian cohort using magnetic and non-magnetic beads in the Peruvian cohort (n=590). Antibodies level is represented in relative antibody unit. Correlation is analysed using a Spearman's rank correlation. Figure S7. Correlation of total IgG antibodies to 19 P. vivax antigens measured in negative controls using magnetic and non-magnetic bead in multiplex assay Total IgG antibodies to 19 P. vivax antigens were measured in negative controls using magnetic and non-magnetic beads in the Peruvian cohort (n=590). Antibodies level is represented in relative antibody units. Correlations were analysed using a Spearman's rank correlation. Figure S8. Comparison of RBP2b1986-2653 classification performance for identifying individuals with P. vivax infections in the last 9 months generated using non-magnetic beads and magnetic beads in Peruvian individuals.

Supplementary
(a) ROC curves for RBP2b1986-2653 generated using non-magnetic beads and magnetic beads; (b) Bioplex RBP2b1986-2653 classification performance; (c) Magpix RBP2b1986-2653 classification performance. The size of the bar corresponds to the number of samples in each category. The darker colour represents the proportion of correctly classified individuals, and the lighter colour represents the proportion of mis-classified individuals. Supplementary Table S1. List of P. vivax antigens used for detecting antibodies in multiplex assay. Protein amount signifies the amount of protein added to a bulk coupling of magnetic beads as described in the methods. In the current study double the optimisied amount of protein was added compared to usual (see Mazhari et al PLOS One 2020 in main reference list), as a modification required to measure the presence of functional antibody responses. Log-linear standard curves were still achieved for all proteins. Abbreviations: aa=amino acid; WGCF= wheat germ cell-free system. Annotations: a =PlasmoDB.

Protein
PlasmoDB ID