Transcriptional plasticity of virulence genes provides malaria parasites with greater adaptive capacity for avoiding host immunity

Abstract

Chronic, asymptomatic malaria infections contribute substantially to disease transmission and likely represent the most significant impediment preventing malaria elimination and eradication. Plasmodium falciparum parasites evade antibody recognition through transcriptional switching between members of the var gene family, which encodes the major virulence factor and surface antigen on infected red blood cells. This process can extend infections for up to a year; however, infections have been documented to last for over a decade, constituting an unseen reservoir of parasites that undermine eradication and control efforts. How parasites remain immunologically "invisible" for such lengthy periods is entirely unknown. Here we show that in addition to the accepted paradigm of mono-allelic var gene expression, individual parasites can simultaneously express multiple var genes or enter a state in which little or no var gene expression is detectable. This unappreciated flexibility provides parasites with greater adaptive capacity than previously understood and challenges the dogma of mutually exclusive var gene expression. It also provides an explanation for the antigenically "invisible" parasites observed in chronic asymptomatic infections.

Keywords: antigenic variation; gene expression; immune evasion; single cell RNAseq; var genes.

Figure 1.. Single-cell analysis by Drop-Seq reveals different var expression states

(A) var expression profiles of the four populations analyzed in the Drop-Seq experiments, represented as histograms and pie charts. Expression of each gene is determined by quantitative RT-PCR and is represented as relative to seryl-tRNA synthetase (PF3D7_0717700). (B) Cumulative var expression of the four populations resulting from Drop-Seq represented as pie charts. The size of each slice of the pie is proportional to the number of UMI detected per individual var gene. (C) Number of total var UMIs relative to total UMI per individual cell in the two “low-many” populations (blue, N=1048) compared to the two “high-single” populations (red, n=873). The mean ± standard deviation is shown, and an unpaired t-test indicates a ****p < 0.0001. (D) Representative examples of individual cells in three var states: Single (pink), cells expressing a single var gene at a high-level; Null (black), cells not expressing any or a very low-level of var transcripts; Multiple (Blue), cells expressing two or more var genes at the same time. Expression is shown as number of UMI detected per individual gene. (E) Percentage of individual cells in the Single state (pink), Null state (black) or Multiple (blue) in each of the populations. Genes were considered expressed with at least 2 UMIs.

Figure 2.. Decreased histone methylation disrupts mutually exclusive var gene expression in individual cells

(A) Total var expression levels as determined by quantitative real-time RT-PCR (qRT-PCR) for two “low-many” populations (blue), two “high-single” populations (red) and two SAMS-KD populations (green). The horizontal line shows median value. (B,C) var expression profiles of the SAMS-KD Population A (B) and Population B (C) determined by quantitative RT-PCR and represented as relative to seryl-tRNA synthetase (PF3D7_0717700). (D) Number of total var UMI relative to total UMI per individual cell in the two “low-many” populations (blue, n=1048) compared to the two “high-single” populations (red, n=873) and the two SAMS-KD populations (green, n=1199). The mean ± standard deviation is shown, and a one-way ANOVA test indicates a ****p < 0.0001. (E) Percentage of individual cells in the Single state (pink), Low-Null state (black) or Multiple (blue) in each of the SAMS-KD populations. Genes were considered expressed with at least 2 UMIs. (F, G) var expression profiles of 5 individual cells from Population A (F) and Population B (G) determined by Drop-Seq and represented as number of UMIs.

Figure 3.. var-enrichment probes allow deeper var transcript detection in scRNA-Seq

(A) Schematic representation of the enrichment procedure. (B) UMI percentage over total UMI for var genes (blue), control genes (pink) and genes not targeted in the enrichment process (grey) in “high-single” and “low-many” populations before and after enrichment. (C) Number of UMI over total UMI for each control gene in “high-single” and “low-many” populations before and after enrichment. (D) var expression profiles after Drop-Seq and enrichment of “high-single” clone A and “low-many” clone A represented as number of UMIs. (E) Pie charts representing var expression profiles after Drop-Seq before and after enrichment of “high-single” clone A and “low-many” clone A.

Figure 4.. var-enrichment probes allow detection of multiple var transcripts in individual cells

var gene expression is displayed for the top-50 cells for “high-single” (A,C) or “low-many” (B,D) populations according to total UMI detected by Drop-Seq. var gene expression is shown either before (A,B) or after (C,D) enrichment. Each color in a bar represents a single var gene and each bar represents an individual cell. (E, F) Relative UMI counts are shown as the percentage of total var UMI in “high-single” (E) and “low-many” (F) from Drop-Seq after enrichment.

Figure 5.. var genes are the main cluster-drivers in HIVE scRNA-Seq

(A) Number of cells recovered with a minimum of 25 UMI per cell in the Drop-Seq experiments (black) compared to HIVE experiments (pink). (B) Average number of UMI per cell in the Drop-Seq experiments (black) compared to HIVE experiments (pink). (C) UMAP of the HIVE single-cell transcriptomes obtained from the four parasite populations with cells colored according to their clustering (see Supplemental Table 4 for cluster information). (D) UMAP of the HIVE single-cell transcriptomes with cells colored according to the parasite population that was sampled. (E) UMAP of the HIVE single-cell transcriptomes obtained from the four parasite populations excluding clonally-variant genes from the analysis. Cells are colored according to the parasite population that was sampled. (F) UMAP graphs as in (C, D) with cells colored according to expression level of different var genes. (G) UMAP graphs as in (C, D) with cells colored according to expression level of different ring-expressed genes.

Figure 6.. HIVE scRNA-Seq confirms multiple var transcripts in individual cells

var gene expression is displayed for the top-100 cells for “high-single” (A,C) or “low-many” (B,D) populations according to total UMI detected by HIVE scRNA-Seq. Each color in a bar represents a single var gene and each bar represents an individual cell. UMI counts are displayed for individual cells obtained from “high-single” (A) and “low-many” (B) populations. Relative UMI counts are shown as percentage of total var UMI in cells obtained from “high-single” (C) and “low-many” (D) populations.

Figure 7.. Parasites in the “low-many” state exhibit reduced immunogenicity

(A) Total var expression levels for IT4 clones determined by quantitative RT-PCR, with transcripts for each var gene shown in a different color. Values are shown as relative to seryl-tRNA synthetase (PfIT_020011400). (B) Example of flow-cytometry with hyperimmune IgG on a “high-single” IT4 clone (blue, gated infected RBC), one “low-many” (red, gated infected RBCs) and uninfected RBCs (orange). Histogram shows normalized cell count over FITC intensity. (C) Correlation between total var expression determined by qRT-PCR (blue) and mean FITC intensity (pink) quantified by flow-cytometry for each IT4 clone. FITC intensity of infected RBCs is normalized to FITC intensity of uRBCs.