Developmental regulation of regenerative potential in Drosophila by ecdysone through a bistable loop of ZBTB transcription factors

Abstract

In many organisms, the regenerative capacity of tissues progressively decreases as development progresses. However, the developmental mechanisms that restrict regenerative potential remain unclear. In Drosophila, wing imaginal discs become unable to regenerate upon damage during the third larval stage (L3). Here, we show that production of ecdysone after larvae reach their critical weight (CW) terminates the window of regenerative potential by acting on a bistable loop composed of two antagonistic Broad-complex/Tramtrack/Bric-à-brac Zinc-finger (ZBTB) genes: chinmo and broad (br). Around mid L3, ecdysone signaling silences chinmo and activates br to switch wing epithelial progenitors from a default self-renewing to a differentiation-prone state. Before mid L3, Chinmo promotes a strong regenerative response upon tissue damage. After mid L3, Br installs a nonpermissive state that represses regeneration. Transient down-regulation of ecdysone signaling or Br in late L3 larvae enhances chinmo expression in damaged cells that regain the capacity to regenerate. This work unveils a mechanism that ties the self-renewing and regenerative potential of epithelial progenitors to developmental progression.

Fig 1. Chinmo is up-regulated in the blastema of damaged wing imaginal discs and promotes efficient regeneration.

(A) Schematic representation of the rn^ts>egr ablation system used to induce wing pouch ablation. Strong wg expression at R0 is observed in response to damage when ablation is initiated at d7 for 40 hours. wg expression is drastically reduced when ablation is initiated at d9. From [9]. (B) Distribution of degrees of wing regeneration in rn^ts>egr,yw adults after d7 ablation (n = 617 wings), rn^ts>egr,yw adults after d9 ablation (n = 216 wings), and rn^ts>egr,chinmo^RNAi adults after d7 ablation (n = 344 wings). p = 2.2 × 10⁻⁵⁷ and p = 0.0013 (rn^ts>egr,yw at d7 compared to rn^ts>egr,yw at d9 and rn^ts>egr yw at d7 compared to rn^ts>egr,chinmo^RNAi at d7, respectively). (C) Anti-Dcp-1 (gray), anti-Wg (green), and anti-Chinmo (magenta) stainings at R0 in an rn^ts>egr wing disc after d7 ablation. Blow-up shows that wg and chinmo are highly coexpressed in wing pouch cells. (D) Anti-Dcp-1 (gray), anti-Wg (green), and anti-Chinmo (magenta) stainings at R0 in an rn^ts>egr wing disc after d9 ablation. Blow-up shows that Wg and Chinmo are low in the wing pouch. (E) Anti-Dcp-1 (gray), anti-Wg (green), and anti-Chinmo (magenta) stainings at R0 in a rn^ts>egr,chinmo^RNAi wing disc after d7 ablation. Blow-up shows that both Wg and Chinmo are low in wing pouch cells. (F) Volume of anti-Wg staining over total wing disc volume at R0 upon d7 ablation in rn^ts>egr larvae (n = 11 wing discs, m = 0.146 ± 0.011), upon d9 ablation in rn^ts>egr larvae (n = 13 wing discs, m = 0.043± 0.004), and upon d7 ablation in rn^ts>egr,chinmo^RNAi larvae (n = 10 wing discs, m = 0.039 ± 0.007). p = 8.0 × 10⁻⁷ and p = 5.7 × 10⁻⁶ (rn^ts>egr,yw at d7 compared to rn^ts>egr,yw at d9 and rn^ts>egr,yw at d7 compared to rn^ts>egr,chinmo^RNAi at d7, respectively). Scale bars: 30 μm. Underlying data for Fig 1 can be found in S1 Data. d, day; Dcp-1, Death Caspase-1; egr, eiger; RNAi, RNA interference; rn^ts, rotund-GAL4, tubulin-GAL80^{thermo-sensitive}; R0, beginning of the recovery period; vol, volume; Wg, Wingless; yw, yellow,white.

Fig 2. Chinmo down-regulation and Br up-regulation occur shortly after the CW has been reached.

(A) Chinmo (magenta) protein level is high in the wing disc during early L3 (5–10 hours) but rapidly decreases shortly after the CW when cut starts to be expressed (arrowhead, mid L3, 15–20 hours). Chinmo is absent when Sens starts to be expressed (arrowhead, mid L3, 25–30 hours) and remains absent in late L3 stage (late L3, 40 hours). (B) Chinmo (magenta) and Br-Z1 (green) protein levels in the wing disc during early L3, after the CW (mid L3), and in late L3. (C) Wing imaginal discs of early L3 larvae transferred in sucrose 5% for 48 hours before the CW maintain high levels of Chinmo and no Br. In contrast, discs of control larvae of the same age maintained on normal food exhibit an absence of Chinmo but high Br. (D) Schematic outline of the above experiment. Scale bars: 30 μm. br, broad; CW, critical weight; eL3, early L3; hr, hours; L3, last larval stage; sens, senseless.

Fig 3. Ecdysone signaling cell-autonomously induces a Chinmo-to-Br switch.

GAL4 expression and Flip-out clones are marked with GFP and outlined in yellow. (A) A chinmo-lacZ enhancer trap exhibits transcriptional silencing in late L3 larvae. (B) Misexpression of EcR^DN in the posterior compartment of the wing disc using en-GAL4 prevents chinmo silencing (magenta) and triggers br and br-Z1 repression (blue) in late L3. (C) Misexpression of EcR^DN using nab-GAL4 leads to ectopic chinmo-lacZ expression in the wing pouch of late L3 larvae. (D) Flip-out clones misexpressing EcR^RNAi show decreased anti-Chinmo staining (magenta, 21/22 clones, n = 7 discs) and increased anti-Br staining (blue, 21/22 clones, n = 7 discs) in mid L3. (E) Schematic outline of the above experiments. Dotted lines may represent direct or indirect regulatory interactions. Scale bars: 30 μm. br, broad; Br-Z1, Broad-Z1; β-Gal, β-Galactosidase; CW, critical weight; EcR^DN, dominant negative form of ecdysone receptor; eL3, early L3; en, engrailed; FO, Flip-out; GFP, green fluorescent protein; L3, third larval stage; RNAi, RNA interference.

Fig 4. Chinmo and Br form a bistable loop during L3.

Flip-out clones and MARCM clones are marked with GFP and outlined in yellow. (A) Flip-out clones misexpressing chinmo exhibit a down-regulation of Br (magenta, 7/7 clones, n = 4 discs) and ectopic Chinmo (blue) during late L3. (B) MARCM clones mutant for chinmo exhibit increased br expression (magenta, 31/31 clones, 7 discs) during eL3 stage. (C) Misexpression of br-Z1 in Flip-out clones during early L3 leads to repression of chinmo (magenta, 16/17 clones, n = 8 discs). (D) Chinmo (magenta) is ectopically expressed during late L3 in br^RNAi Flip-out clones (11/11 clones, n = 5 discs). (E) Schematic outline of the above experiments. Scale bars: 30 μm. br, broad; CW, critical weight; eL3, early L3; FO, Flip-out; GFP, green fluorescent protein; L3, third larval stage; RFP, red fluorescent protein; RNAi, RNA interference.

Fig 5. Chinmo prevents differentiation during L3 stages.

Flip-out clones and MARCM clones are marked with GFP and outlined in yellow. (A–B) Anti-Cut (A, magenta) and anti-Sens (B, magenta) stainings are lost in late L3 Flip-out clones misexpressing chinmo (5/5 clones, n = 3 discs and 10/10 clones, n = 3 discs, respectively). (C) Relative cell area of wild-type and misexpressing chinmo Flip-out clone cells compared to wild-type surrounding cells in late L3. Wild-type Flip-out cells (n = 6 clones, 3 discs, 462 cells, m = 0.99 ± 0.03) and chinmo misexpressing cells (n = 9 clones, 3 discs, 245 cells, m = 1.60 ± 0.07). p-value is 3.5 × 10⁻¹³. (D–E) Anti-Cut (blue) and anti-Sens (magenta) stainings appear precociously in MARCM clones that are mutant for chinmo (5/5 clones, n = 5 discs and 3/3 clones, n = 3 discs, respectively). (F) Schematic outline of the above experiments. Scale bars: 30 μm. br, broad; CW, critical weight; eL3, early L3; FO, Flip-out; GFP, green fluorescent protein; hr, hours; L3, third larval stage; MARCM, Mosaic Analysis with a Repressible Cell Marker sens, senseless.

Fig 6. Ecdysone signaling and Br cooperate to induce differentiation.

Flip-out clones, GAL4 expression, and MARCM clones are marked with GFP and outlined in yellow. (A) Anti-Cut staining (magenta) is absent during late L3 when br is knocked down in the posterior compartment using en-Gal4. (B) cut (magenta) is ectopically expressed in Flip-out clone cells misexpressing br-Z1 during mid L3. (C) cut (magenta) is not expressed in MARCM clones mutant for both chinmo and br (chinmo¹;br^RNAi cells) during late L3 (11/12 clones, n = 7 discs). (D) br-Z1 (magenta) is silenced in Flip-out clones expressing both EcR^DN and chinmo^RNAi transgenes in late L3 (52/52 clones, n = 4 discs). (E) cut (magenta) and sens (blue) are not expressed in Flip-out clones expressing both EcR^DN and chinmo^RNAi transgenes during late L3 (15/15 clones, n = 7 discs). (F) cut (magenta) is not expressed in Flip-out clones expressing both br-Z1 and EcR^DN transgenes during late L3 (4/4, n = 3 discs). (G) Schematic outline of the above experiments. Scale bars: 30 μm. br, broad; CW, critical weight; EcR^DN, dominant negative form of ecdysone receptor; eL3, early L3; en, engrailed; FO, Flip-out; GFP, green fluorescent protein; L3, third larval stage; MARCM, Mosaic Analysis with a Repressible Cell Marker; RNAi, RNA interference; sens, senseless.

Fig 7. The Chinmo-to-Br switch restricts regenerative capacity.

(A) Before d7 ablation, Chinmo (magenta) is highly expressed and Br (green) is low, whereas Chinmo is low and Br is highly expressed at d9. (B) Anti-Chinmo and anti-Br stainings (color-coded relative to staining intensity) at R0 in a rn^ts>egr wing disc after d7 and after d9 ablations. Blow-up shows that Chinmo (magenta) is high in the wing pouch while Br is low (green) at d7. In contrast, Chinmo is low in the wing pouch while Br is strong at d9. (C) Anti-Chinmo and anti-Br stainings (both color-coded relative to staining intensity) and anti-Dcp-1 (gray), anti-Wg (green), and anti-Chinmo (magenta in the blow-up) stainings in an rn^ts>egr,br-Z1 wing disc at R0 after d7 ablation. Chinmo and Wg are low while Br is high in the wing pouch. (D) Volume of anti-Wg staining over total wing disc volume at R0 upon d7 ablation in rn^ts>egr larvae (n = 11 wing discs, m = 0.145 ± 0.011) and rn^ts>egr,br-Z1 larvae (n = 13 wing discs, m = 0.094 ± 0.008). p = 9.2 × 10⁻⁴. (E) Distribution of degrees of wing regeneration upon d9 ablation in rn^ts>egr adults (n = 122 wings) and rn^ts>egr adults grown on sucrose 5% from d7 to d9 (n = 162 wings). p = 5.1 × 10⁻²¹. Scheme depicting the timing of starvation and ablation procedure. Scale bars: 30 μm. Underlying data for Fig 7 can be found in S1 Data. br, broad; d, day; Dcp-1, Death Caspase-1; egr, eiger; eL3, early L3; L3, third larval stage; rn^ts, rotund-GAL4, tubulin-GAL80^{thermo-sensitive}; R0, beginning of the recovery period; vol, volume; Wg, Wingless; yw, yellow,white.

Fig 8. Preventing the Chinmo-to-Br switch can restore regenerative potential in late L3.

(A) Anti-Chinmo and anti-Br wing disc stainings at R0 after d9 ablation in various genetic conditions (color-coded relative to staining intensity). (B) Anti-Dcp-1 (gray) and anti-Wg (green) wing disc stainings at R0 after d9 ablation in various genetic conditions. Blow-up shows that wg and chinmo (magenta) are highly coexpressed in the blastema of rn^ts>egr,EcR^DN; rn^ts>egr,chinmo; and rn^ts>egr,br^RNAi discs, whereas Chinmo and Wg are lower in the blastema of rn^ts>egr and rn^ts>egr,EcR^DN,chinmo^RNAi discs. (C) Anti-Mmp1 (gray) staining at R0 after d7 and d9 ablation in various genetic conditions. Dilp8-GFP (gray) is poorly expressed in d9 rn^ts>egr yw wing discs, whereas it is highly expressed in d7 rn^ts>egr,yw and d9 rn^ts>egr,chinmo wing discs. (D) Volume of anti-Wg staining over total wing disc volume at R0 upon d9 ablation at in rn^ts>egr larvae (n = 13 wing discs, m = 0.043 ± 0.004); rn^ts>egr,EcR^DN larvae (n = 9 wing discs, m = 0.092 ± 0.005); rn^ts>egr,EcR^DN,chinmo^RNAi larvae (n = 8 wing discs, m = 0.047 ± 0.003); rn^ts>egr,chinmo larvae (n = 10 wing discs, m = 0.165 ± 0.015); and rn^ts>egr,br^RNAi larvae (n = 11 wing discs, m = 0.089 ± 0.011). p = 8.0 × 10⁻⁶, p = 8.2 × 10⁻⁵, p = 1.7 × 10⁻⁶, and p = 8.0 × 10⁻⁷ (rn^ts>egr compared to rn^ts>egr,EcR^DN; rn^ts>egr,EcR^DN compared to rn^ts>egr,EcR^DN,chinmo^RNAi; rn^ts>egr compared to rn^ts>egr,chinmo; and rn^ts>egr compared to rn^ts>egr,br^RNAi, respectively). (E) Volume of anti-Mmp1 staining over total wing disc volume at R0 upon d7 ablation in rn^ts>egr larvae (n = 10 wing discs, m = 0.150 ± 0.014) and upon d9 ablation in rn^ts>egr larvae (n = 14 wing discs, m = 0.060 ± 0.008), rn^ts>egr,chinmo larvae (n = 11 wing discs, m = 0.230 ± 0.029), rn^ts>egr,EcR^DN larvae (n = 8 wing discs, m = 0.136 ± 0.013), and rn^ts>egr,br^RNAi larvae (n = 12 wing discs, m = 0.147 ± 0.005). p = 1.9 × 10⁻⁵, p = 89.0 × 10⁻⁷, p = 1.9 × 10⁻⁴, and p = 6.2 × 10⁻⁶ (d7 rn^ts>egr compared to d9 rn^ts>egr and d9 rn^ts>egr compared to rn^ts>egr,chinmo, to rn^ts>egr,EcR^DN, and to rn^ts>egr,br^RNAi, respectively). (F) Volume of anti-dilp8-GFP staining over total wing disc volume at R0 upon d7 ablation in rn^ts>egr larvae (n = 10 wing discs, m = 0.160 ± 0.011) and upon d9 ablation in rn^ts>egr larvae (n = 10 wing discs, m = 0.057 ± 0.008) and in rn^ts>egr,chinmo larvae (n = 11 wing discs, m = 0.200 ± 0.016). p = 2.2 × 10⁻⁵ and p = 5.7 × 10⁻⁶ (d7 rn^ts>egr compared to d9 rn^ts>egr and d9 rn^ts>egr compared to d9 rn^ts>egr,chinmo, respectively). (G) Distribution of degrees of wing regeneration in rn^ts>egr,yw adults (n = 216 wings) and rn^ts>egr,br^RNAi adults (n = 172 wings) after d9 ablation. p = 4.4 × 10⁻³³. Scale bars: 30 μm. Underlying data for Fig 8 can be found in S1 Data. br, broad; d, day; Dcp-1, Death Caspase-1; Dilp8, Drosophila Insulin-like peptide 8; EcR^DN, dominant negative form of ecdysone receptor; egr, eiger; GFP, green fluorescent protein; L3, third larval stage; Mmp1, Matrix metalloprotease 1; RNAi, RNA interference; rn^ts, rotund-GAL4, tubulin-GAL80^{thermo-sensitive}; R0, beginning of the recovery period; vol, volume; Wg, Wingless; yw, yellow,white.

Fig 9. Ecdysone coordinates self-renewal, differentiation, and regenerative potential with developmental progression via the Chinmo/Br bistable loop.

During early development, Chinmo represses differentiation programs while promoting self-renewal state in wing disc epithelial cells. Upon ecdysone production after the CW, br becomes activated and promotes differentiation by repressing Chinmo and possibly other genes. The Chinmo-to-Br switch induced by ecdysone also causes restriction of regenerative potential. br, broad; CW, critical weight; EcR, ecdysone receptor; eL3, early L3; L3, third larval stage.