Skip to main content
Figure 6 | Genome Medicine

Figure 6

From: Transcriptional profiling defines dynamics of parasite tissue sequestration during malaria infection

Figure 6

Dynamics of gametocyte sequestration. (A) Gametocyte ring and mature gametocyte clusters. Left panel: gametocyte ring clusters 13, 44 and 110 show expression in the same subset of isolates from two patient cohorts. The commitment-specific transcription factor ap2-g (PFL1085w) is active in the same individuals, demonstrating presence of circulating gametocyte rings in this subset of patients. Right panel: genes from the mature gametocyte cluster 49, including the gold standard gametocyte marker Pfs25, are co-expressed in the same seven patients. Red arrowheads mark isolates with slide positivity for gametocytes (data available for Milner et al. [26] only). (B) Relative expression of gametocyte clusters in vivo. Expression dynamics in samples from cerebral malaria patients in Malawi [26]. Gametocyte clusters were separated into four categories based on the minimum q-value resulting from gene enrichment for: i) committed schizont, ii) gametocyte ring, iii) immature gametocyte (young and developing gametocytes), and iv) mature gametocyte markers (Additional file 1). Two clusters (36 and 262, marked with crosses) were manually reassigned due to significance in multiple categories and greater marker expression in a non-minimum category. Within these categories, clusters are horizontally sorted by peak time, as indicated above the heatmap: asexual peak time was used to map clusters of committed schizonts and gametocyte rings; for immature and mature gametocyte clusters the NF54 gametocyte time course by Young et al. [30] was used as reference. Transcriptional analysis of clusters indicated expression of four gametocyte ring clusters, two immature gametocyte clusters, and seven mature gametocyte clusters in a subset of patient blood samples. Predicted circulation and sequestration properties during gametocyte development are indicated above the cluster peak times. Slide positivity for each sample is indicated on the right. Numbers represent number of gametocytes per microliter based on smears. The heat map at the bottom highlights clusters significantly enriched (q-value ≤0.05) in putative gametocyte-specific genes, identified by four independent approaches, in black. (C) Expression dynamics of circulating and sequestering parasite clusters. To compare global expression levels of circulating and sequestering parasite populations, we analyzed mean transcript expression across clusters. For asexual parasites, mean transcript expression across circulating (peak time <22 hours post-invasion) versus sequestering (peak time >22 hours post-invasion) clusters was compared in samples from Milner et al. [26] and Joice et al. [19]. Mean transcript expression from clusters representing circulating asexual parasites was significantly higher than from sequestering ones (P < 10−15). For gametocytes mean transcript expression was determined from the subset of patient samples with detectable transcript levels. These are the six samples marked with black arrowheads in (A) (left) for clusters representing young circulating parasites (as defined in (B)), and the seven samples marked with black or red arrowheads in (A) (right) for clusters representing mature circulating gametocytes (as defined in (B)). For mean expression of clusters representing commitment and sequestering (immature) gametocytes the combined 13 samples from above were analyzed. Transcripts from both circulating young and mature gametocytes are detectable at significantly higher levels than the genes from immature gametocytes (P < 10−10), but they are not significantly different from each other (P = 0.3939). (D) Gametocyte marker quantification in blood samples by qRT-PCR showing 78 samples from a cerebral malaria cohort in Blantyre, Malawi. Sentinel markers for gametocyte ring cluster 44 (PF14_0744, PF14_0748) and immature gametocyte cluster 36 (Pfs48/45) and the constitutive marker Ubiquitin conjugating enzyme (UCE) [19] are shown. Overall comparisons of all groups showed a significant difference in at least two of these genes (P < 10−15, Kruskal-Wallis) and post-test pair-wise comparisons showed significant differences (P < 0.0001, Kruskal-Wallis) when comparing the transcript levels of PF14_0744 and PF14_0748 against that of Pfs48/45. Data are normalized by RNA input and values are shown as arbitrary units.

Back to article page