Steroid Hormone Function Controls Non-competitive Plasmodium Development in Anopheles

Abstract

Transmission of malaria parasites occurs when a female Anopheles mosquito feeds on an infected host to acquire nutrients for egg development. How parasites are affected by oogenetic processes, principally orchestrated by the steroid hormone 20-hydroxyecdysone (20E), remains largely unknown. Here we show that Plasmodium falciparum development is intimately but not competitively linked to processes shaping Anopheles gambiae reproduction. We unveil a 20E-mediated positive correlation between egg and oocyst numbers; impairing oogenesis by multiple 20E manipulations decreases parasite intensities. These manipulations, however, accelerate Plasmodium growth rates, allowing sporozoites to become infectious sooner. Parasites exploit mosquito lipids for faster growth, but they do so without further affecting egg development. These results suggest that P. falciparum has adopted a non-competitive evolutionary strategy of resource exploitation to optimize transmission while minimizing fitness costs to its mosquito vector. Our findings have profound implications for currently proposed control strategies aimed at suppressing mosquito populations.

Keywords: 20E signaling; Anopheles-Plasmodium interactions; EIP; extrinsic incubation period; lipid transport; trade-offs.

Figure 1:. Parasite and egg development are linked.

(A) There is a positive correlation between the number of An. gambiae eggs and P. falciparum oocysts developing in females (Spearman’s correlation. Line shows data trend). (B) A transgenic vasa2-Cas9 mosquito line was crossed with a U6₅₇-Zpg gRNA transgenic line to generate Δzpg mosquitoes. (C–D) Blood-fed Δzpg females (C) fail to produce eggs (scale bar = 500 μm) and (D) support fewer parasites compared to Zpg/+ controls (Cntrl) (Mann-Whitney), while there is no effect on infection prevalence (P) (Chi-Squared). (E) Δzpg females produce lower levels of ecdysteroids at 26 h and 36 h pBM (Kruskal-Wallis, Dunn’s correction). N = sample size.

Figure 2:. The steroid hormone 20E regulates both egg and parasite numbers.

(A) Injecting wild type females with the ecdysteroid oxidase E22O reduces ecdysteroid levels in the female 26 h pBM compared to BSA-injected controls (Cntrl) (unpaired t-test). (B–C) E22O-injected females (B) produce fewer eggs (Mann-Whitney) and (C) have fewer oocysts at 8 d pIBM (GLM, Poisson distribution). There is no effect on infection prevalence (P) (Chi-Squared). (D) Egg and oocyst numbers are positively correlated in control and E22O-injected females (Spearman’s correlations. Lines show data trends). N = sample size.

Figure 3:. 20E controls the egg-parasite correlation via its nuclear receptor.

(A–B) Females injected with dsEcR (EcR) and fed on a P. falciparum infected blood meal (A) produce fewer eggs (Mann-Whitney) and (B) develop fewer oocysts at 7 d pIBM (GLM, Poisson distribution) compared to dsGFP-injected females (Cntrl). There is no effect on infection prevalence (P) (Chi-Squared). (C) The positive correlation between egg and oocyst numbers is lost in dsEcR-injected females (Spearman’s correlation). (D) dsEcR females have lower Vg expression at 24 h pBM compared to controls (means ± SEM, unpaired t-test, FDR corrected). NBF = non-blood fed. (E) The positive egg-oocyst correlation is also lost in dsUSP-injected females (USP) (Spearman’s correlation). N = sample size. Lines show data trends.

Figure 4:. 20E signaling affects parasites during the ookinete-oocyst transition.

(A) dsEcR injections (EcR) have no effect on the number (GLM, Poisson distribution) or prevalence (P) (Fisher’s Exact) of ookinetes traversing the midgut at 24 h pIBM, detected by immunofluorescent labeling of Pfs25 after blood bolus removal. Control = dsGFP-injected (Cntrl). (B–C) By 2 d pIBM, dsEcR females have fewer oocysts compared to controls (Log(y) transformed, GLM, normal distribution), as detected by Pfs25 (green) labeling. Red = actin (phalloidin) (scale bar = 50 μm). There remains no effect on infection prevalence (P) (Fisher’s Exact). (D) TEP1 silencing does not rescue oocyst numbers in dsEcR females. Females injected with dsEcR or co-injected with dsEcR and dsTEP1 (EcR/TEP1) have the same number of oocysts at 7 d pIBM, and both groups have reduced oocyst numbers compared to controls. There is no effect of dsTEP1 (TEP1) alone on oocyst numbers (Kruskal-Wallis, Dunn’s correction). Infection prevalence (P) is unaffected (Fisher’s Exact). N = sample size.

Figure 5:. Disrupting 20E signaling promotes faster parasite development.

(A) Oocysts are larger in dsEcR females (EcR) compared to dsGFP controls (Cntrl) at 8 d pIBM (scale bar = 50 μm). (B) The mean oocyst area per midgut is higher in dsEcR females from 5–10 d pIBM. By 12 d pIBM, as oocysts burst, there is no difference between dsEcR and controls (Least Squares model). (C–D) Salivary glands of dsEcR females show (C) a higher prevalence of sporozoites at 10 d pIBM (Chi-Squared) and (D) significantly more sporozoites at 12 d pIBM (Mann-Whitney). No difference is observed at later time points (Mann-Whitney). (E) Salivary gland sporozoites collected from dsEcR females at 12 d pIBM are as infectious as controls when applied to primary human hepatocytes in vitro (medians ± IQR, unpaired t-test) (EEFs = exoerythrocytic forms). (F) The EIP₅₀ (time to 50% infectious) of dsEcR females is significantly reduced compared to controls, as determined from sporozoite prevalence data shown in (C) and (D) (means ± 95% CI, unpaired t-test). N = sample size with dsGFP listed first. See Figures S1–S2 and Table S1.

Figure 6:. Faster parasite development is mediated by Lp-transported lipids with no additional cost to fecundity.

(A) Lp expression is elevated in dsEcR females (EcR) after blood feeding relative to dsGFP controls (Cntrl) (means ± SEM, unpaired t-test, FDR corrected). NBF = non-blood fed. (B) Total lipids, TAGs, and PCs are significantly higher in dsEcR midguts at 48 h pIBM while DAGs are increased independent of time (means ± SEM, Least Squares model). (C) Co-injecting females with both dsEcR and dsLp (EcR/Lp) largely returns mean oocyst size to control levels at 8 d pIBM (Kruskal-Wallis, Dunn’s correction). (D) In dsGFP controls, there is a negative correlation between the number of eggs and the mean size of oocysts in individual females at 8 d pIBM. This correlation is lost in dsLp females (Spearman’s correlations. Lines show data trends). (E) dsEcR females develop the same number of eggs after either an infectious or inactivated P. falciparum blood meal (Mann-Whitney). N = sample size with dsGFP listed first. See Figures S3–S4.

Figure 7:. Silencing the TAG lipase TL2 induces faster parasite growth.

(A) Total lipids, TAGs, DAGs, and PCs are higher in dsTL2 midguts (TL2) at 24 h and 48 h pBM compared to dsGFP controls (Cntrl) (means ± SEM, Least Squares model). (B) Midguts of dsTL2 females stained with the neutral lipid dye LD540 (red) show an accumulation of lipid droplets at 48 h pBM (scale bar = 20 μm). Blue = DAPI. (C) Lipophorin expression is elevated in dsTL2 fat bodies at 24 h pBM (means ± SEM, Least Squares model). (D) dsTL2 females develop larger oocysts from 5–10 d pIBM (Mann-Whitney). N = sample size with dsGFP listed first. See Figures S4–S5.