Natural Genetic Variation Screen in Drosophila Identifies Wnt Signaling, Mitochondrial Metabolism, and Redox Homeostasis Genes as Modifiers of Apoptosis

Abstract

Apoptosis is the primary cause of degeneration in a number of neuronal, muscular, and metabolic disorders. These diseases are subject to a great deal of phenotypic heterogeneity in patient populations, primarily due to differences in genetic variation between individuals. This creates a barrier to effective diagnosis and treatment. Understanding how genetic variation influences apoptosis could lead to the development of new therapeutics and better personalized treatment approaches. In this study, we examine the impact of the natural genetic variation in the Drosophila Genetic Reference Panel (DGRP) on two models of apoptosis-induced retinal degeneration: overexpression of p53 or reaper (rpr). We identify a number of known apoptotic, neural, and developmental genes as candidate modifiers of degeneration. We also use Gene Set Enrichment Analysis (GSEA) to identify pathways that harbor genetic variation that impact these apoptosis models, including Wnt signaling, mitochondrial metabolism, and redox homeostasis. Finally, we demonstrate that many of these candidates have a functional effect on apoptosis and degeneration. These studies provide a number of avenues for modifying genes and pathways of apoptosis-related disease.

Keywords: Drosophila; apoptosis; genetic variation; modifier genes.

Figure 1

Activation of apoptosis through p53 and rpr-associated pathways. Apoptosis is primarily initiated through either p53 or Jun-induced (JNK) transcriptional activation of the Inhibitor of Apoptosis (IAP, in Drosophila Diap) inhibitors hid, rpr and grim. While p53 is primarily activated by DNA damage and disruption of the cell cycle, JNK signaling is activated downstream of cellular stress, such as endoplasmic reticulum (ER) stress, through Ire1 and Cdk5. ER stress occurs when misfolding proteins, like the rhodopsin mutant Rh1^G69D, accumulate in the ER (Chow et al. 2016). Expression of rpr, grim, and hid leads to inhibition of Diap1, releasing the inhibition on initiator caspases and allows for the activation of effector caspases and downstream apoptosis. Models used in this or previous studies of retinal degeneration in the DGRP are indicated in white.

Figure 2

Apoptosis levels vary across genetic background in p53 and rpr models of apoptosis-induced degeneration. Apoptosis induced by overexpression of p53 (A) or rpr (B), as measured by adult eye size, varies across different genetic backgrounds. DGRP strains are arranged along the X-axis from smallest to largest. Each point represents the median eye size of an individual DGRP strain, while the error bars represent the standard deviation of eye size for that strain. Representative images of GMR > p53 eyes (C) or GMR-rpr eyes (D) in different DGRP backgrounds demonstrate the phenotypic variation quantified in panels A and B.

Figure 3

Eye size is correlated between GMR-rpr and both GMR > p53 and GMR > Rh1^G69D models of degeneration. Correlation in mean eye size between the GMR-rpr, GMR > p53, and GMR > Rh1^G69D models across the DGRP. A. Eye size is significantly correlated in the same DGRP strains expressing GMR-rpr and GMR > p53 (r = 0.19, P = 0.0071). B. Eye size is significantly correlated in the same DGRP strains expressing GMR-rpr and GMR > Rh1^G69D (r = 0.25, P = 0.001). C. Eye size is not correlated in the same DGRP strains expressing GMR > p53 and GMR > Rh1^G69D (r = 0.12, P = 0.13). * P < 0.05, ** P < 0.005.

Figure 4

rpr modifiers are enriched for neuronal function, Wnt signaling, and metabolic pathways. A. rpr modifier network, as plotted by the GeneMANIA plugin in Cytoscape (Shannon et al. 2003; Montojo et al. 2010). Significant candidate modifiers are indicated in red, with physical interactions shown in green, genetic interactions shown in blue, and predicted interactions shown in gray. Thicker lines indicate stronger evidence for the association. B. Top 20 significant ontological categories as identified by GSEA. Categories are arranged from most significant on top to least significant along the y-axis. P-values are indicated by red-to-blue gradient, with red the lowest p-values and blue the highest P-values. Gene number identified in each category is indicated along the y-axis.

Figure 5

Knockdown of candidate rpr modifiers significantly alters apoptosis-induced degeneration. RNAi against candidate modifiers was expressed under the control of GMR-GAL4 in the GMR-rpr model. The genetically matched attP2 line was crossed into the GMR-rpr line as a control (blue). Eye size in pixels was quantified for N = 11-15 flies per strain and plotted with the 25^th-75^th percentile of measurements in the central box. Measurements lying outside of 1.5 × interquartile range are indicated as points. Representative images of each line are found above the data for that line. Knockdown of LIMK1 or swim significantly reduces eye size in the GMR-rpr model of degeneration compared to controls. Loss of Sema1a, MED16, or hay results in a significant increase in eye size compared to controls. Loss of CG1907 does not significantly alter eye size, but changes in pigmentation are similar in the presence or absence of GMR-rpr (Fig S4C). Loss of αManA1, LysRS, or CG3032 do not produce a significant effect. RNAi lines with significant changes in eye size are indicated in red, while those that are not significantly changed are indicated in white. * P < 0.05, *** P < 0.0005.