PREreview of REL2 overexpression in theAnopheles gambiaemidgut causes major transcriptional changes but fails to induce an immune response

In Hoermann et al., they examine the role of over-expressing the IMD regulatory protein, REL2, in Anopheles gambiae mosquitoes to aid in elimination of malaria parasites. The over-expression of REL2 has been previously shown to be effective in blocking Plasmodium transmission and altering reproductive mosquito behavior in favor of gene drive dynamics in Anopheles Stephensi. Thus, this work presented here addresses an important gap in our knowledge in A. gambiae mosquitoes towards generating additional control tools for an important malaria vector. Using previously established Integral Gene Drive (IGD) lines, they integrate REL2 into the midgut-specific bloodmeal-inducible zinc carboxypeptidase (CP) A1 locus and examine the impact on Plasmodium transmission and vector fitness. As a control, they knock out the CP locus and subsequently measure the same fitness parameters. Finally, they perform RNA-seq analysis to examine transcriptional changes resulting from REL2 overexpression and examined transcriptional changes that contained a REL2 DNA-binding site in the promoter region. 

The introduction provides adequate background on the Immune Deficient Pathway and the role of Relish as transcriptional regulator for immunity in the previously published work in Anopheles stephensi. The authors then provide the conceptional framework for their experiments using their previously published Integral Gene Drive (IGD) system with REL2 in A. gambiae. The final statements of the introduction section regarding the major findings should be moved to the discussion and further expanded on in that section. The experimental methods (transgenic mosquitoes, parasite infections and RNA-seq analysis) used are appropriate in the study and a few concerns regarding the statistical analysis for certain experiments are detailed in the points below. Both the results section and figure legends could use improvements to improve clarity on the described experiment. Much of the discussion reads as a re-statement of the results section and could be replaced with a short summary of the overall findings and then expanding upon why the authors think there may be differences between the previously published findings using a different mosquito species and potential future directions to overcome this challenge.

Major Points:   A major gap in the study is the lack of confirmation for REL2 function in the IMD pathway for the over-expression line. One of the key findings from the previous work done in Anopheles Stephensi is the impact of REL2 on the microbiome of the vector, which is subsequently missing from this study. As REL2 is part of the anti-microbial pathway, this should be thoroughly examined through either 16S quantitative PCR or plating of midguts for CFU counts after a blood meal. This is essential as the over-expression of REL2 cannot be traced back to the specific upregulation of antimicrobial peptides (and as the RNA-seq data seems to indicate leads to the down-regulation of multiple AMPs) as it does in the Drosophila system.   A large focus of the manuscript (almost half) is based on RNA-seq data generated from the CP-REL2 line and then subsequently examining potential gene candidates regulated by REL2 by computation analysis. However, for this data to justify such a significant portion of the manuscript, additional experiments (luciferase assay, EMSA DNA binding) should be done to support REL2 binding to target promoter regions for gene activity. Minor Points: The methodology behind the generation of the REL2-S over-expression line is confusing for readers who are not familiar with the laboratories' previous publication on generating the IGD lines. A more detailed figure, including a map of the HDR donor plasmid pD-REL2-CP, should be included in Figure 1. It would be beneficial to measure REL2 gene expression levels between the WT line and REL2-CP lines to confirm the construct results in enhancing REL2 expression. For the phenotypic characteristics of the CP-REL2 line in Figure 3, it would better to exclude the zero’s (egg numbers and oocysts infection intensity) and plot them as prevalence %. The zero’s can easily skew the mean/median of the data set and thus should be analyzed separately. For example in Figure 3A, the mean of eggs / female by REL2-CP group ~50 eggs; however, that is because ~20% of mosquitoes fail to produce any eggs, lowering the overall group average. Thus, if a female end up producing eggs, it will be ~70 eggs / female, which closer to the WT group than is represented in the current graph.

For the survival curves in Figure 3C, A log-rank (Mantel–Cox) test is the appropriate test to be used to compare survival distribution. The infection plots (Fig. 3D) should display the median (oocysts intensity is a non-normal distribution) and a generalized linear model should be used to pool the data and consider replicate effects between the groups for statistical analysis. It seems like all three replicates have a small reduction in oocyst numbers in the CP-REL2 line, and if analyzed this way it would possibly reach significance.  Several studies have correlated a reduction in egg numbers with an increase in parasite size/development. Since there was a reduction in egg numbers in the CP-REL2 line, was there any change to parasite size as noted in Figure 4 with the KO-CP line? Is it possible that the female mosquitoes that died before examination are the ones that display fewer oocysts? If mosquitoes were examined at day 4-5 rather than day 7 would there be a difference in oocysts intensity?

Would an alternative to over-expressing REL2 be to over-express the individual AMPs genes of interest under the CP Promoter? This could be discussed as a future potential project in the discussion.

Competing interests

The author declares that they have no competing interests.