Article Figures and data Abstract eLife digest Introduction Results Discussion Materials and methods Data availability References Decision letter Author response Article and author information Metrics Abstract Experience alters brain structure, but the underlying mechanism remained unknown. Structural plasticity reveals that brain function is encoded in generative changes to cells that compete with destructive processes driving neurodegeneration. At an adult critical period, experience increases fiber number and brain size in Drosophila. Here, we asked if Toll receptors are involved. Tolls demarcate a map of brain anatomical domains. Focusing on Toll-2, loss of function caused apoptosis, neurite atrophy and impaired behaviour. Toll-2 gain of function and neuronal activity at the critical period increased cell number. Toll-2 induced cycling of adult progenitor cells via a novel pathway, that antagonized MyD88-dependent quiescence, and engaged Weckle and Yorkie downstream. Constant knock-down of multiple Tolls synergistically reduced brain size. Conditional over-expression of Toll-2 and wek at the adult critical period increased brain size. Through their topographic distribution, Toll receptors regulate neuronal number and brain size, modulating structural plasticity in the adult brain. eLife digest Everything that you experience leaves its mark on your brain. When you learn something new, the neurons involved in the learning episode grow new projections and form new connections. Your brain may even produce new neurons. Physical exercise can induce similar changes, as can taking antidepressants. By contrast, stress, depression, ageing and disease can have the opposite effect, triggering neurons to break down and even die. The ability of the brain to change in response to experience is known as structural plasticity, and it is in a tug-of-war with processes that drive neurodegeneration. Structural plasticity occurs in other species too: for example, it was described in the fruit fly more than a quarter of a century ago. Yet, the molecular mechanisms underlying structural plasticity remain unclear. Li et al. now show that, in fruit flies, this plasticity involves Toll receptors, a family of proteins present in the brain but best known for their role in the immune system. Fruit flies have nine different Toll receptors, the most abundant being Toll-2. When activated, these proteins can trigger a series of molecular events in a cell. Li et al. show that increasing the amount of Toll-2 in the fly brain makes the brain produce new neurons. Activating neurons in a brain region has the same effect, and this increase in neuron number also depends on Toll-2. By contrast, reducing the amount of Toll-2 causes neurons to lose their projections and connections, and to die, and impairs fly behaviour. Li et al. also show that each Toll receptor has a unique distribution across the fly brain. Different types of experiences activate different brain regions, and therefore different Toll receptors. These go on to trigger a common molecular cascade, but they modulate it such as to result in distinct outcomes. By working together in different combinations, Toll receptors can promote either the death or survival of neurons, and they can also drive specific brain cells to remain dormant or to produce new neurons. By revealing how experience changes the brain, Li et al. provide clues to the way neurons work and form; these findings may also help to find new treatments for disorders that change brain structure, such as certain psychiatric conditions. Toll-like receptors in humans could thus represent a promising new target for drug discovery. Introduction Structural brain plasticity and neurodegeneration reveal generative and destructive processes operating in the brain. Plasticity reflects adaptations of the brain to environmental change, involving adult neurogenesis, growth of neurites and synapses, which correlate with learning, experience, physical exercise and anti-depressant treatment (Holtmaat and Svoboda, 2009; Deng et al., 2010); conversely, neuroinflammation, neurodegeneration, loss of neurons, neurites and synapses, correlate with ageing, stress, depression and disease (Wohleb et al., 2016). Structural brain plasticity affects the brain topographically, influencing the specific regions involved in experience-dependent processing. These manifestations suggest that brain function is encoded in physical changes to cells. Structural plasticity occurs in the Drosophila brain (Sugie et al., 2018). Breeding adult flies in constant darkness decreases, and in constant light increases brain volume (Barth and Heisenberg, 1997; Barth et al., 1997). Breeding adult flies in isolation vs. crowded conditions, or in single sex vs. mixed groups, also causes brain volume changes (Technau, 1984; Heisenberg et al., 1995). The affected modules include the optic lobe, the mushroom body calyx and central complex (Technau, 1984; Heisenberg et al., 1995; Barth and Heisenberg, 1997; Barth et al., 1997). Changes in brain volume are prominent in a critical period spanning from adult eclosion to day 5, and correlate with changes in fiber number (Technau, 1984; Heisenberg et al., 1995; Barth and Heisenberg, 1997; Barth et al., 1997). The molecular mechanisms underlying structural brain plasticity are unknown, and discovering them is crucial to understand the normal functionality of the brain as well as its pathological responses to disease. Primary candidates to regulate brain plasticity are the neurotrophins. In the mammalian brain, neurotrophins (BDNF, NGF, NT3, NT4) regulate cell proliferation, cell survival, circuit connectivity, synaptic transmission and potentiation (Lu et al., 2005). Alterations in neurotrophins underlie brain disease, and anti-depressants increase the levels of the neurotrophin BDNF (Krishnan and Nestler, 2008; Wohleb et al., 2016). NTs have dual functions, as they promote plasticity via p75NTR activating NF-κΒ, and via Trk receptors activating AKT, ERK and CREB downstream, and they promote neurodegeneration via p75NTR and JNK signalling (Lu et al., 2005). Drosophila neurotrophins (DNTs) also regulate neuronal survival and death, connectivity and synaptic structural plasticity (Zhu et al., 2008; Sutcliffe et al., 2013; McIlroy et al., 2013; Foldi et al., 2017; Ulian-Benitez et al., 2017). However, there are no canonical tyrosine-kinase-Trk and p75NTR receptors in Drosophila, and instead, DNTs are ligands for the Kekkons, kinase-less members of the Trk family, and Tolls (McIlroy et al., 2013; Foldi et al., 2017; Ulian-Benitez et al., 2017). Drosophila Toll and mammalian Toll-Like-Receptors (TLRs) are best known for their universal function in innate immunity (Leulier and Lemaitre, 2008), but also have non-immune functions in development and in the central nervous system (CNS)(Anthoney et al., 2018). In neurons, Tolls and TLRs can promote neuronal survival via MyD88 and neuronal death via Sarm, both in flies and mammals (Kim et al., 2007; McIlroy et al., 2013; Mukherjee et al., 2015; Foldi et al., 2017). In humans, alterations in TLR function underlie brain diseases from stroke and neurodegeneration to multiple sclerosis and neuroinflammation (Okun et al., 2011; Hanamsagar et al., 2012). Most attention has focused on TLR functions in microglia, their response to damage or infection, and in neuroinflammation (Fiebich et al., 2018). However, TLRs are also in neurons, but functions in neurons and neural progenitor cells are largely unknown. Importantly, TLRs can influence neurogenesis, neuronal survival and death, neurite growth, synaptic transmission and behaviour, including learning and memory (Ma et al., 2006; Rolls et al., 2007; Okun et al., 2010b; Okun et al., 2011; Qi et al., 2011; Okun et al., 2012; Madar et al., 2015; Liu et al., 2016b; Patel et al., 2016; Hung et al., 2018; Min et al., 2018). These findings suggest that TLRs could regulate structural brain plasticity, but this remains little explored. Tolls regulate cell number plasticity in the Drosophila ventral nerve cord (VNC) through a three-tier mechanism (Foldi et al., 2017). In embryos and larvae, Toll-6 and Toll-7 maintain neuronal survival via MyD88 and NF-κB (McIlroy et al., 2013; Foldi et al., 2017). However, in pupae, they can also promote apoptosis via Weckle (Wek), Sarm and JNK (Foldi et al., 2017). Furthermore, different Tolls lead to different outcomes, for instance, Toll-1 is more pro-apoptotic than Toll-6 (Foldi et al., 2017). Whether a neuron lives or dies in the CNS depends on the ligand and its cleavage state it receives, the Toll or combination of Tolls it expresses, and the downstream adaptors available for signalling (Foldi et al., 2017). Thus, cell number control is context dependent. The ability of DNTs and Tolls to regulate cell number by promoting both cell survival and cell death is crucial for the modulation of structural brain plasticity, homeostasis and neurodegeneration. Here, we asked whether Toll receptors influence developmental and structural plasticity in the Drosophila brain. Results A Toll receptor map in the Drosophila brain To find out whether Toll receptors are expressed in the brain, we looked for Toll transcripts in embryos and dissected CNSs from larvae to adult brains, using reverse-transcription PCR (RT-PCR) (Figure 1—figure supplement 1). Toll-3 transcripts were absent from larval L2 CNSs; Toll-4 and −9 mRNAs were barely detected in all sample types; whereas Toll-1,–2, −5,–6, −7,–8 were expressed in embryos, larval (L2, L3) CNSs, and pupal and adult fly heads (Figure 1—figure supplement 1). Thus, all Tolls are expressed in pupal and adult brains, with Toll-1,–2, −5,–6, −7,–8 most prominently. To visualise the spatial distribution of Tolls in the brain, we generated GAL4 reporter lines for the Tolls. Using CRISPR/Cas9-accelerated homologous recombination to insert a pTV cassette (Baena-Lopez et al., 2013), we generated a knock-in/knock-out Toll-2pTV allele, and we used the pTV-attP landing site to generate a Toll-2pTV-GAL4 driver line. Toll-4GAL4 and Toll-5GAL4 were generated by CRISPR/Cas9, inserting T2AGAL4 immediately upstream of the start codon. Unfortunately, we could not get transformants for Toll-9. Toll-3GAL4, Toll-6GAL4 and Toll-7GAL4 were made using Recombinase-Mediated Cassette Exchange (RMCE) of MIMIC insertions into the intronless coding regions of these genes. Toll-8GAL4 is TolloMD806, which has a P-element insertion just 180 bp upstream of the start codon, within the 5'UTR of Toll-8. The GAL4 driver lines were used to visualise membrane tethered FlyBow (for Toll-2,–4, −5,–7) and tdTomato (for Toll-3,–6, −8) reporters, and all necessarily reproduced the endogenous expression patterns of the Toll genes. Toll-1 was visualised using commercially available and previously validated anti-Toll-1 antibodies (Lund et al., 2010; Khadilkar et al., 2017). In the adult brain, Toll-1 was found in all photoreceptor cells (Figure 1A,G). Toll-2,–5, −6,–7 and −8 were all expressed in Kenyon cells, with Toll-2 and −6 comprising most cells (Figure 1A,B,C). Toll-5 and −7 were expressed in the protocerebral bridge (Figure 1A,B). Toll-2,–5, −6,–7 and −8 were differentially expressed in the antennal lobes (Figure 1D). Toll-1,–2, −3,–5, −6,–7, −8 were expressed in distinct and overlapping fan shaped body neuropile layers (Figure 1E); Toll-1,–2 and −7 in distinct ellipsoid body neuropile rings (Figure 1F), and Tol-1,–2, −6 and −8 in the sub-esophageal ganglion (SOG) (Figure 1A). Toll-2,–3, −5,–6, −7,–8 were expressed in optic lobes, with Toll-3 and −6 having a prominent expression in the lamina (Figure 1A,G) and Toll-2 a broad expression throughout the optic lobes (Figure 1A,G). In summary, these patterns revealed: (1) a map of Toll expression profiles coincident with anatomical brain domains; (2) profiles specific to each Toll; (3) complementary patterns in neuropiles of the visual system and central complex (fan shaped body, ellipsoid body and protocerebral bridge); (4) overlapping distributions in optic lobes, antennal lobes, Kenyon cells and mushroom bodies. Tolls could influence brain structure and connectivity by virtue of their topographic profiles (Figure 1I). Figure 1 with 1 supplement see all Download asset Open asset Expression of Tolls demarcates the anatomical map of the adult brain. (A) Toll receptor expression visualised with: Anti-Toll-1 antibodies, in retinal photoreceptors; CRISPR/Cas9 generated Toll-2pTVGAL4 > FlyBow, throughout the brain; MIMIC-RMCE generated Toll-3GAL4 > tdTomato restricted to the lamina; CRISPR/Cas9 generated Toll-4GAL4 > FlyBow did not reveal any signal; CRISPR/Cas9 generated Toll-5GAL4 > FlyBow was prominent in medulla, Kenyon cells and protocerebral bridge (pb). (Unfortunately, we could not get CRISPR/Cas9 data for Toll-9). MIMIC-RMCE generated Toll-6GAL4 > tdTomato is prominent in lamina, Kenyon cells and central brain interneurons. MIMIC-RMCE generated Toll-7GAL4 > FlyBow was prominent in optic lobe and Kenyon cells. Toll-8GAL4MD806 > tdTomato was prominent in central brain and Kenyon cells. (B–G) Higher magnification views to show signal in: (B) Kenyon cells (KCs); (C) mushroom bodies (mb); (D) antennal lobes (AL); (E) fan shaped body (fsb); (F) ellipsoid body (eb); (G,H,J) optic lobes (OL). (I) Drawing illustrating the brain domains of KCs, central brain (CB) and OLs used for the functional analysis. La: lamina; me: medulla; lo: lobula; lp: lobula plate; no: noduli; pcb: protocerebral bridge. Scale bars: A,E,F,G: 25 μm; B,C,D,H,J: 50 μm. For genotypes and sample sizes see Materials and methods and Supplementary file 2. See Figure 1—figure supplement 1. Toll-2 is neuro-protective in the brain To ask whether Tolls may influence brain development and/or adult structural brain plasticity, we focused on Toll-2, as it is most broadly expressed. In neurons, Drosophila Tolls and mammalian TLRs promote neuronal survival via MyD88 and neuronal death via Sarm (Kim et al., 2007; McIlroy et al., 2013; Mukherjee et al., 2015; Foldi et al., 2017; Figure 2C). Thus, to investigate whether Toll-2 is required for cell survival in brain development, we first verified whether MyD88 was expressed in the brain. MyD88NP6394 flies bear a GAL4 insertion within the transcribed 5'UTR exon, and thus it necessarily represents the endogenous expression pattern of the gene (from now on called MyD88GAL4). MyD88GAL4 >tdTomato revealed MyD88+ cells throughout the optic lobes, central brain, and mushroom bodies (Figure 2A). Toll-2 appears to be expressed in all Kenyon cells, whereas MyD88 is only in the subset that projects along the core α,β lobes (Figure 2A). Using the nuclear reporter histone-YFP (his-YFP) revealed that more cells expressed Toll-2 than MyD88 (Figure 2B). In the optic lobes, MyD88 >hisYFP includes large, sparsely distributed cells, that may or may not also be Toll-2+ (Figure 2B). To identify the Toll-2+ and MyD88+ cells, Toll-2 >hisYFP and MyD88 >hisYFP adult brains were labelled with pan-neuronal anti-Elav and pan-glial anti-Repo. There were many MyD88+ Elav+ neurons as well as MyD88+ Repo+ glia (Figure 2D). By contrast, none of the Toll-2+ cells were Repo+, whilst most Toll-2+ cells were Elav+ (Figure 2E). Thus, MyD88+ cells comprise both neurons and glia, that are most likely regulated by multiple Tolls, and Toll-2+ cells are mostly neurons. Figure 2 with 1 supplement see all Download asset Open asset Toll-2 knock-down caused neuronal loss in the pupal and adult brain. (A,B) Overlapping but distinct expression of Toll-2 and the adaptor MyD88, visualised with MyD88 >tdTomato, Toll-2pTVGAL4 > UASFlyBow, MyD88 >histone YFP and Toll-2pTVGAL4 > UAShistone YFP. Toll-2 is expressed in more Kenyon cells (arrowheads) than MyD88. Note the large MyD88+ cells in the optic lobes (B, arrowhead). (C) Diagram of signalling pathways downstream of Toll receptors regulating cell death and cell survival (adapted from Foldi et al., 2017). (D,E) The pan-neuronal marker anti-Elav co-localises with His-YFP in both MyD88+ and Toll-2+ cells, whereas the pan-glial marker anti-Repo only co-localises with MyD88 >His-YFP+ cells (arrowheads). (F) Drawings showing in green the central brain region of interest (ROI), and Kenyon cells (KCs), used for automatic cell counting with DeadEasy. (G) Toll-2 RNAi knock-down increased the number of anti-Dcp1+ apoptotic cells, in day one pupal central brains (dashed line indicates ROI); cells quantified automatically in the ROI in 3D throughout the stack of images, with DeadEasy software. Left: full projection; right: projection of five optical sections only (5 μm). Quantification in box-plot graph: Student t-test p=0.0295. (H) Toll-2 RNAi knock-down decreased MyD88+ cell number in pupal and adult brains, latter using two independent RNAi lines; MyD88 RNAi knock-down also decreased cell number in the pupal brain (left), but not in the adult brain (right). Dashed lines in (H) indicate central brain ROI used for automatic counting of MyD88 >hisYFP+ cells with DeadEasy Central Brain software. Box-plots: Left: One Way ANOVA p<0.001, and right p<0.0001, post-hoc Dunnett tests. (I) Neither Toll-2 over-expression nor RNAi knock-down altered KC number, in pupal nor adult brains, but Toll-2pTV/Tollp2Δ7-35 mutants had more KCs. Dashed lines in (I) indicate ROI counted automatically with DeadEasy Kenyon Cells software, box-plot graphs on right: Kruskal Wallis ANOVA, both p>0.1. Scale bars: A,B,G left,H: 50 μm; D,E: 10 μm; G right, I: 25 μm. For genotypes, sample sizes and statistical details, see Supplementary file 2. *p<0.05, **p<0.01, ***p<0.001. See Figure 2—figure supplement 1. To ask whether Toll-2 might regulate cell survival in brain development, we visualised apoptotic cells using anti-Dcp1 antibodies upon Toll-2 knock-down. In brain development, the peak of cell death occurs 24 hr after puparium formation (Hara et al., 2018). So, we quantified apoptosis in day one pupal brains, using purposely developed DeadEasy Central Brain software. Toll-2 RNAi knock-down in MyD88+ cells increased apoptosis in the pupal central brain (Figure 2G), meaning that Toll-2 is required to maintain cell survival. To verify whether apoptosis resulted in cell loss, we counted automatically MyD88 >hisYFP cells in the central brain. Using two independent UAS-Toll-2 RNAi lines of flies, Toll-2 knock-down with MyD88GAL4 decreased cell number in the central brain of both pupae and adult flies (Figure 2H). Thus, Toll-2 loss of function in MyD88+ cells increased apoptosis and caused cell loss. On the other hand, sustained over-expression of Toll-2 with MyD88GAL4 throughout development did not affect cell number in the pupal or adult brains (Figure 2H). In larvae and pupae, Tolls can also induce apoptosis via Sarm, and different Tolls have distinct pro-apoptotic drive (Foldi et al., 2017). As Toll-2 gain of function did not reduce cell number, this meant that Toll-2 does not induce apoptosis in the pupal or adult brain. Together, these data showed that Toll-2 maintains the survival of MyD88+ neurons during brain development. To test if Toll-2 maintains neuronal survival via the MyD88 pathway, we knocked-down MyD88 with MyD88GAL4. Similarly to Toll-2 loss of function, MyD88 knock-down also resulted in cell loss in the pupal central brain (Figure 2H). However, by the adult stage, cell number was restored vs. controls (Figure 2H). This was in contrast to the persistent cell loss caused by Toll-2-RNAi into the adult, suggesting that MyD88 carries out further functions too. These data showed that MyD88 is required to maintain cell survival during brain development, downstream of at least Toll-2. To analyse the effect in Kenyon Cells (KCs), we developed another plug-in - DeadEasy KCs- to count KCs labelled with Toll-2 >hisYFP. Toll-2pTV/Toll-2Δ7-35 mutations increased KC number in pupal brains, but neither sustained gain nor loss of Toll-2 function with Toll-2 >Toll-2RNAi knock-down or Toll-2pTV/Toll-2Δ7-35 mutations affected KC number in adult brains (Figure 2I). Over-expression of Toll-2 with another mushroom body driver, MBGAL4, did not affect KC number in pupal or adult brains either (Figure 2—figure supplement 1B,C). Thus, KCs are resilient to alterations in Toll-2 function alone. To further test whether Toll-2 is neuroprotective, we induced Toll-2pTV homozygous mutant MARCM clones. Genetic complementation tests over the previously described null allele 18wΔ7-35 (18 w is a synonym of Toll-2, thus hereby will be referred to as Toll-2Δ7-35) and a deficiency for the locus, Df(2R)BSC594, showed that Toll-2pTV is a strong hypomorphic loss of function allele (Figure 3—figure supplement 1A). Toll-2pTV mutant clones were induced from dividing cells in the pupal brain, where Toll-2 is widely expressed (Figure 3A). They were induced using hsFlp, and resulting mutant neurons were visualised in adult brains with elav >mCD8 GFP (Figure 3B). Loss of Toll-2 function caused extensive neuronal loss, neuronal atrophy, loss of neurites - axons and dendrites- and axonal misrouting (Figure 3C–H). Loss of dendrites could be clearly observed in the lamina (Figure 3C); axonal degeneration and misrouting in medulla and lobula (Figure 3D); and loss of entire axonal neuropiles in the medulla, SOG and fan shaped body (Figure 3D,E,H). Whether mushroom bodies were affected was less clear (Figure 3F,G), perhaps because our heat shock regime missed mushroom body neuroblast divisions, as we could not observe many mushroom body projections in control brains either (Figure 3G). Dramatic neuronal deficits could be found throughout many brain domains (Figure 3C–H). The loss of neurons in mutant clones was consistent with Toll-2 maintaining neuronal survival, but could also reflect a function promoting progenitor cell proliferation; neurite atrophy meant that Toll-2 loss of function prevented neuronal differentiation or caused neurodegeneration. Figure 3 with 1 supplement see all Download asset Open asset Loss of Toll-2 function caused neurodegeneration and impaired behaviour. (A) Toll-2 is expressed in pupal brains, most prominently in optic lobes, mushroom bodies and SOG. (B–H) Toll-2pTV mutant MARCM clones - labelled with GFP - induced in pupa cause neuronal loss throughout the adult brain (genotype of clones: elavGAL4, UASmCD8GFP, hsFlp; neo FRT42D Toll-2pTV/neo FRT42D Toll-2pTV). Compare controls with Toll-2 mutant clones in: (B) whole brains; (C) loss of neurons and dendrites (arrow) in lamina (la); (D) neuronal loss, misrouted and/or aberrant axons in optic lobe; (E) loss of sub-esophageal ganglion (SOG) neuropile; (F,G) Calyx (ca) and mushroom bodies (mb) were less affected; (H) loss of fan shaped body (fsb) neuropiles. me: medulla; lo: lobula; lp: lobula plate. (I–M) Buridan arena behavioural assay revealed impaired locomotion with Toll-2pTV loss of function. (I) Representative trajectories of single flies filmed for the same amount of time, walking in a lit-up arena, with two diametrically opposed dark stripes. Phenotypes were divided into three categories – between stripes, random, little walking - and percentages indicate how many flies per genotype showed each phenotype (penetrance); no statistically significant differences were found. (J–M) Automatic measurement of locomotion parameters with purposely written software. Toll-2 mutant flies walked less than controls: they achieved equal speeds as controls, but they had more walking bouts that were brief, and thus walked shorter distances. Box-plots, quantifications: (J, K, L) One Way ANOVA: (K) p<0.0001; (L) p<0.01; (M) Kruskal-Wallis ANOVA p<0.0001; stars indicate post-hoc (J,K,L) Dunnett and (M) Dunn's test comparisons to fixed controls. Scale bars: A,B,D,E,F: 50 μm; C,G,H: 25 μm. *p<0.05; **p<0.01; ***p<0.001; ****p<0.0001. For genotypes, sample sizes and statistical details, see Supplementary file 2. See Figure 3—figure supplement 1. Toll-2 mutants are semi-lethal, have reduced lifespan and impaired climbing (Figure 3—figure supplement 1B,C) – phenotypes commonly associated with neurodegeneration. Toll-2 is expressed in the visual system, ventral nerve cord and central complex, which is the higher control center for locomotion and spatial navigation in the brain (Strauss and Heisenberg, 1993). Thus, we tested the performance of Toll-2 mutants in the Buridan arena, which could reveal whether loss of Toll-2 affected vertical vs. horizontal locomotion, visual processing, or motivation to walk. Wild-type flies free to walk in a circular lit-up arena walked back and forth between two diametrically opposed dark stripes (38.5%), but most often wondered randomly (61.5%, Figure 3I). Toll-2pTV/Df(2R)BSC22 and Toll-2Δ7-35/Df(2R)BSC22 mutants also most often walked randomly or along the perimeter (63.7–92%), but overall walked less than controls, and some walked very little (7.7–18.1%). Adult specific Toll-2 knock-down in neurons, with tubulinGAL80ts to switch on GAL4 and drive elav >Toll-2RNAi after adult fly eclosion, reproduced the behavioural phenotypes of the mutants (Figure 3I). Importantly, this shows that Toll-2 is required in adult neurons. Both wild-type and mutant flies could walk between the black stripes, meaning that loss of Toll-2 function does not impair vision. Quantitative analysis of the flies' walking behaviour did not reveal significant differences between the genotypes in their preference to walk between the dark stripes, or away from the centre of the arena. Hence, we cannot conclude that Toll-2 mutant flies have impaired visual processing. Interestingly, all wild-type flies walked more than Toll-2 mutants. In fact, loss of Toll-2 function significantly affected locomotion: all genotypes could walk at the same speed (Figure 3J), but Toll-2 mutants spent less time walking than controls, thus overall travelled shorter distances (Figure 3K), and although they had as many or more walking bouts, these were brief (Figure 3L,M). Importantly, these phenotypes were consistent across different Toll-2 mutant alleles, and also when Toll-2 was conditionally knocked-out in adult post-mitotic neurons only (Figure 3J–M). As Toll-2 mutants could achieve the same speeds as wild-type flies, but walked less, either motor circuit function and/or the motivation to walk were impaired. To conclude, Toll-2 loss of function resulted in neurodegeneration and impaired behaviour. Thus, Toll-2 is required for the formation and integrity of brain neural circuits. Toll-2 can increase cell number in the adult brain To ask whether Toll-2 might affect structural plasticity in the adult brain, we altered its function at the adult critical period (i.e. day 0–5 post-eclosion), when the brain is most plastic (Technau, 1984; Heisenberg et al., 1995; Barth and Heisenberg, 1997; Barth et al., 1997). We used tubulinGAL80ts to silence GAL4, then switched on GAL4 at post-eclosion adult day 0 and analyzed the brains two days later. Over-expression of Toll-2 with MyD88GAL4 increased the number of histone-YFP+ cells in the central brain (Figure 4A), meaning that Toll-2 can regulate cell number at the adult critical period. Conditional Toll-2-RNAi-knock-down also caused a mild increase in cell number (Figure 4A), which could be due to compensatory adjustments by other Tolls. In mushroom body KCs, neither Toll-2 knock-down nor over-expression with Toll-2ptvGAL4 restricted to the critical period had any effect on KC number (Figure 4B). The optic lobes are particularly susceptible to structural plasticity (Heisenberg et al., 1995; Barth et al., 1997), thus we drove conditional over-expression with tubulinGAL80ts and Toll-2 >histone YFP, and automatically counted YFP+ cells with purposely adapted DeadEasy Optic Lobes software (Figure 4C). Conditional Toll-2-RNAi knock-down had no effect, whereas over-expression of Toll-2 increased the number of YFP+ medulla neurons (Figure 4C). Thus, Toll-2 can increase cell number in the adult optic lobes. The increase in cell number by Toll-2 gain of function in the central brain and optic lobes is consistent with a neuroprotective function, but could also involve cell proliferation. Either way, these data showed that Toll-2 is not pro-apoptotic in the adult brain, and instead can positively regulate cell number during the adult critical period. Figure 4 with 1 supplement see all Download asset Open asset Cell number is regulated by Toll-2 and neuronal activity at the adult critical period. At adult days 0–2 post-eclosion, within the critical period: (A) Conditional over-expression of Toll-2 increased MyD88 >hisYFP+ cell number in the central brain. Cells were counted automatically in 3D throughout the whole stack with DeadEasy Central Brain, dashed line in all figures indicates ROI quantified. Box-plots, Kruskal-Wallis p<0.0001, post-hoc Dunn test; (B) neither conditional over-expression nor knock-down of Toll-2, altered Toll-2 >hisYFP+ Kenyon cell number, counted automatically with DeadEasy Kenyon Cells, box-plots; (C) conditional over-expression of Toll-2 increased Toll-2 >hisYFP+ cell number in the optic lobe medulla, counted automatically with DeadEasy Optic Lobe. Box-plots, One Way ANOVA p<0.0001, post-hoc Dunnett; (D) pulses of neuronal activation with TrpA1 increased Toll-2 >hisYFP+ cell number in the medulla, and this could be rescued with Toll-2 RNAi knock-down. Box-plots: Left: Un-paired Student t-test, p=0.0058; Right: Un-paired Student t-test, p=0.0225. (E) Knocking-down JNK and cactus and over-expressing activated PI3K (UAS-Dp110CAAX), alone or in combination, in MyD88+ cells with tubGAL80ts, MyD88GAL4 increased cell number in the central brain, consistent with pro-survival signalling downstream of Toll-2. However, over-expressing either wek or MyD88 RNAi knock-down in