Atypical soluble guanylyl cyclases control brain size in Drosophila

Abstract

Hypoxia-induced proliferation of neural stem cells has a crucial role in brain development. In the brain of Drosophila melanogaster , the optic lobe exhibits progressive hypoxia during larval development. Here, we investigate an alternative oxygen-sensing mechanism within this brain compartment, distinct from the canonical hypoxia signaling pathway mediated by HIF. Using genetic tools, immunostaining, and confocal microscopy, we demonstrate that the loss of the atypical soluble guanylyl cyclase (asGC) subunit Gyc88E , or the ectopic expression of Gyc89Db in neural stem cells leads to increased optic lobe volume. We propose the existence of a link between cGMP signaling and neurogenesis in the developing brain.

Figure 1. Effect of atypical soluble guanylyl cyclase loss- and gain-of-function in brain development of Drosophila larvae .

(A) Volume of brain hemispheres at third larval instar. Loss-of-function Gyc88E ^-/- mutant larvae (88E ^-/- ) displayed increased brain hemisphere volume, (median±SD: 1.93e+06 ± 3.54e+05 µm ³ ), whereas Gyc89Da ^-/- Db ^-/- double mutant (89Da ^-/- Db ^-/- ) did not (median±SD: 7.29e+05 ± 4.34e+05 µm³). Increased hemisphere volume was also observed in larvae ectopically expressing Gyc89Db in optic-lobe neuroepithelial cells under control of GAL4 ^c855a (c855a>89Db, median±SD: 1.84e+06 ± 1.53e+05 µm ³ ) or overexpresing it in neuroblasts under control of insc-Gal4 (insc>89Db, median±SD: 1.88e+06 ± 1.85e+05 µm ³ ), compared to w ¹¹¹⁸ (median±SD: 9.86e+05 ± 5.70e+05 µm ³ ). (B) Volume of central brain. Loss-of-function mutant larvae displayed increased central brain volume, Gyc88E ^-/- (median±SD: 1.08e+06 ± 2.46e+05 µm ³ ), Gyc89Da ^-/- Db ^-/- (median±SD: 8.53e+05 ± 1.81e+05 µm ³ ) as did ectopic expression lines driving Gyc89Db in optic-lobe neuroepithelial cells (median±SD: 8.53e+05 ± 6.18e+04 µm³). Overexpression in neuroblasts did not alter volume occupied by neuroblasts (median±SD: 8.41e+05 ± 1.37e+05 µm ³ ), compared to w ¹¹¹⁸ (median±SD: 5.05e+05 ± 1.70e+05 µm ³ ) . (C) Volume of the optic lobe (OL). Gyc88E ^-/- mutant larvae did not show increased OL volume (median±SD: 8.04e+05 ± 2.01e+05 µm³), nor did Gyc89Da ^-/- Db ^-/- larvae (median±SD: 6.84e+05 ± 2.46e+05 µm³). Larvae ectopically expressing Gyc89Db in neuroepithelial cells showed increased OL volume (median±SD: 1.04e+06 ± 4.21e+05 µm³) whereas those overexpressing Gyc89Db in neuroblasts did not (median±SD: 8.72e+05 ± 1.95e+05 µm³) when compared to w ¹¹¹⁸ (median±SD: 4.81e+05 ± 4.23e+05). (D) Assessment of neuroblast number using anti-Deadpan immunostaining. Optic lobe neuroblasts were counted in the brain of larvae from each experimental group. The number of neuroblasts was increased in Gyc88E ^-/- (median±SD: 230.00 ± 67.56) and in Gyc89Da ^-/- Db ^-/- mutant larvae (median±SD: 211.50 ± 56.10). Larvae expressing Gyc89Db in neuroepithelial cells exhibited a 5.8-fold increase in their neuroblast number (median±SD: 425.50 ± 39.99). Overexpression in neuroblasts did not alter their number (median±SD: 85.00 ± 26.06) compared to w ¹¹¹⁸ (median±SD: 73.00 ± 79.18)larvae. (E) Quantification of proliferating optic lobe neuroblasts, identified by double staining for anti-Deadpan and anti-PH3. The number of proliferating neuroblasts was low, without statistically significant differences in Gyc88 E ^-/- (median±SD: 1.00 ± 2.76) or Gyc89Da ^-/- Db ^-/- double mutant larvae (median±SD: 5.00e-01 ± 8.16e-01) compared to w ¹¹¹⁸ (median±SD: 1.00 ± 7.63). Brains expressing Gyc89Db under the neuroepithelial driver GAL4 ^c855a (median±SD: 17.5 ± 3.32) showed a 2.5-fold increase in the number of mitotically active neuroblasts, while those expressing Gyc89Db under the neuroblast driver insc - gal4 showed a 3.4-fold increase (median±SD: 24.00 ± 16.76). (F) Quantification of total proliferating cells, measured by anti-PH3 immunostaining. Gyc88E ^-/- larvae showed increased numbers of total proliferating cells (1.4-fold, median±SD: 154.00 ± 74.42) whereas Gyc89Da ^-/- Db ^-/- did not (median±SD: 79.50 ± 74.46). Larvae expressing Gyc89Db in neuroepithelial cells under control of the GAL4 ^c855a driver exhibited a 4.3-fold increase in the number of mitotic cells (median±SD: 459.50 ± 44.14), whilst larvae expressing Gyc89Db under the neuroblast driver insc-gal4 showed a 1.5-fold increase (median±SD: 164.50 ± 65.58) compared to w ¹¹¹⁸ larvae (median±SD: 64.50 ± 51.07). Mann-Whitney U test, * p<0.05, ** p<0.01, *** p<0.001. Mutant, ectopic expression, and overexpression N=6, w ¹¹¹⁸ N=12.

Description

Oxygen availability is a powerful driver of evolutionary novelty, and metazoans have developed sophisticated hypoxia sensing systems that are intricately related to the control of stem cell niches and development (Simon and Keith 2008; Rytkönen et al. 2011; Hammarlund et al. 2018) . Increased atmospheric oxygen levels during evolution had profound effects on insects enabling them, for instance, to develop larger body size and the ability to fly (Stamati et al. 2011) .

The atypical soluble guanylyl cyclase (asGC) subunits Gyc88E, Gyc89Da and Gyc89Db are expressed from embryo to adult stages in Drosophila melanogaster (Langlais et al. 2004; Morton et al. 2005) . They are regulated by O ₂ levels and when activated by hypoxia generate cGMP, functioning as molecular O ₂ sensors (Vermehren et al. 2006) with faster responses than the canonical hypoxia pathway (de Lima et al. 2021) .

The neural stem cells within the optic lobe (OL) of the Drosophila larval brain are hypoxic relative to the central brain (Misra et al. 2017) , with different cell types experiencing varying degrees of hypoxia (Baccino-Calace et al. 2020) . There is evidence that optic lobe progenitor cells might not exhibit a canonical hypoxia response (Baccino-Calace et al. 2020) . In situ hybridization data (Langlais et al. 2004) suggests that asGC subunits are expressed within the OL, potentially linking O ₂ sensing to an alternative signaling pathway activated by hypoxia. Furthermore, transcriptomic data analysis (Yang et al. 2015) , revealed that Gyc89Db is expressed in L3 neuroblasts. The combination of different asGC subunits might provide different sensitivity thresholds to hypoxia to Drosophila neurons (Lu et al. 2024) .

Here, we used genetic tools to investigate the role of asGC in larval brain development. We explored asGC loss-of-function using a Gyc88E ^-/- mutant and a Gyc89Da ^-/- Db ^-/- double mutant. Additionally, we performed targeted gain-of-function with drivers specific to ectopically express Gyc89Db in neuroepithelial cells or overexpressing it in neuroblasts.

As the hypoxic OL remains neurogenic during larval life, we hypothesized that this hypoxia might activate asGC. Given that cGMP derived from nitric oxide signaling can activate mammalian neural stem cell proliferation (Santos et al. 2014) , if hypoxia-driven cGMP signaling were to activate neural stem cell proliferation, loss of asGC function would abolish it. Contrary to our expectation, the Gyc88E ^-/- loss-of-function mutant exhibited a 1.9-fold increase in brain hemisphere volume ( Figure 1A ). Interestingly, the double mutation Gyc89Da ^-/- Db ^-/- had not the same effect. Notably, ectopic expression in NE or overexpression of Gyc89Db in NB lead to an increase in overall brain size ( Figure 1A ). The larval brain comprises two functionally and developmentally different compartments, the central brain (CB) which is formed mostly during embryonic stages and contains the neuronal synapses and the optic lobe (OL), which lack fully differentiated neurons with synapses and remains neurogenic throughout larval stages (Hartenstein et al. 2008; Baccino-Calace et al. 2020) .

When we measured and compared CB and OL volumes separately, we observed that the macrocephalous phenotype observed in Gyc88E ^-/- mutants can be explained by a specific effect on the CB. Gyc88E ^-/- loss-of-function mutants showed a 2.1-fold increase in CB volume ( Figure 1B ), while the OL volume remains unaffected ( Figure 1C ). The double mutation Gyc89Da ^-/- Db ^-/- also elicited an increase in CB volume but limited to a lesser extent (1.7-fold, Figure 1B ).

Ectopic expression of Gyc89Db under the control of the OL-specific neuroepithelial driver GAL4 ^c855a (Egger et al. 2007) resulted in increased volume of both CB ( Figure 1B ) and OL ( Figure 1C ). Overexpression under the pan-neuroblast driver inscuteable -GAL4 reproduced the macrocephalous phenotype only in the CB ( Fig. 1B ). Provided the OL is the major neurogenic region within the larval brain, experiencing a dramatic volume increase between 24-72 hs after larval hatching (Baccino-Calace et al. 2020) , we reasoned that, in the larva, an eventual increase in proliferation would only be observable in this hypoxic brain region.

Aiming to further investigate the cellular basis of the macrocephalous phenotype, we assessed neuroblast number and proliferation within the OL. asGC loss-of-function mutants and ectopic expression of Gyc89Db in the OL neuroepithelium increased neuroblast number in this region, while overexpression in neuroblasts did not ( Figure 1D ).

Despite the increased OL volume, no differences in the number of proliferating neuroblasts were observed in the loss-of-function mutants ( Figure 1E ), whereas Gyc89Db ectopic expression in neuroepithelial cells or overexpression in neuroblasts increased the number of mitotically active neuroblasts by 2.5-fold and 3.4-fold, respectively ( Figure 1E ).

Furthermore, larvae with Gyc88E ^{-/ -} loss of function or ectopic expression of Gyc89Db in the neuroepithelium or overexpression in neuroblasts show increased total cell number in mitosis in the optic lobe ( Figure 1F ), suggesting a link between oxygen-dependent cGMP signaling and the control of proliferation in neural progenitor cells.

Previous research has shown that Gyc88E is active in the absence of additional subunits, while Gyc89Da and Gyc89Db enhanced the activity of Gyc88E when co-expressed, suggesting that these enzymes acted likely as heterodimers (Morton and Vermehren 2007; Morton 2011) . It is important to mention that Gyc89Db is expressed in the brain during larval development (Langlais et al. 2004) and produces cyclic GMP in response to low oxygen in vitro (Morton 2004; Vermehren et al. 2006) . Thus, asGCs may fine tune the cellular response promoting proliferation through hypoxia-driven cGMP signaling.

Methods

Fly strains

Flies were raised on cornmeal medium at constant temperature 25°C and under 12:12 h light:darkness cycles as previously described (Ferreiro et al. 2018) . Loss-of-function stocks Gyc88E ^-/- (Vermehren-Schmaedick et al. 2010) (Z3-1083, from the Seattle tilling project) and the double mutant Gyc89Da ^-/- Db ^-/- (Vermehren-Schmaedick et al. 2010) (BDSC #93108) were a kind gift of David Morton (Oregon Health & Science University, Oregon, USA). The following driver and responder lines were used: c855a-GAL4 (BDSC stock #6990), insc-GAL4 (BDSC stock #8751), UAS-gyc89Db (Vermehren-Schmaedick et al. 2010) (BDSC stock #93116) was a kind gift from David Morton (Oregon Health & Science University, Oregon, USA). w ¹¹¹⁸ was used as control as mutant and transgenic lines were built on a white background.

Immunostaining and image acquisition

All experiments were performed on wandering third-instar larvae kept in normoxia conditions. Brains were dissected, fixed and immunostained as previously described (Baccino-Calace et al. 2020) . Primary antibodies included a monoclonal mouse anti-Discs large antibody to outline all brain cells and allow for precise identification of the OL boundaries (1:20, DSHB #4F3, Developmental Studies Hybridoma Bank (DSHB), Iowa, USA), guinea pig anti-Deadpan to identify neuroblasts (1:2500, kind gift from Jürgen Knoblich, Eroglu et al. 2014) and rabbit anti-phosphorylated H3 histone antibody as a mitotic marker (anti-PH3, 1:200, Cell Signaling Technologies #9713). Fluorescent conjugated secondary antibodies Alexa488, Alexa568 and Cy5 were used (Thermo-Fisher). DNA was stained with 1 µg/ml Hoechst 33342 (Thermo-Fisher). Images were acquired with a Zeiss LSM800 Airyscan confocal microscope and processed with FIJI (Schindelin et al. 2012) . Semi-automated OL segmentation was performed with TrakEM2 (Cardona et al. 2012) . Analyses and illustrations were made using R version 4.4.0 on RStudio version 2023.06.01.

Reagents

Stocks: Stocks obtained from the Bloomington Drosophila Stock Center (BDSC) or kindly provided by colleagues as stated under the Methods section.

GAL4 ^c855 ( w[1118]; P{w[+mW.hs]=GawB}C855a )

insc-GAL4 ( w[*]; P{w[+mW.hs]=GawB}insc[Mz1407] )

UAS-gyc89Db ( w[*]; P{w[+mC]=UAS-Gyc89Db.V}2 )

Gyc88E ^-/- (point mutation V474M)

Gyc89Da ^-/- Db ^-/- (w[*]; PBac{w[+mC]=RB}Gyc89Da[e01821] Mi{GFP[E.3xP3]=ET1}Gyc89Db[MB03197])

Funding Statement

DP was a post-doctoral researcher at the Instituto de Investigaciones Biológicas Clemente Estable, Ministerio de Educación y Cultura, Uruguay. Grants FVF-2019-05 (Dirección Nacional de Innovación, Ciencia y Tecnología) awarded to DP and FCE_1_2019_1_156160, awarded to RC and support from Agencia Nacional de Investigación e Innovación to DP and RC are acknowledged.