Abstract
Genetic disorders such as neurofibromatosis type 1 (Nf1) increase vulnerability to cognitive and behavioral disorders, such as autism spectrum disorder and attention-deficit/hyperactivity disorder. Nf1 results from mutations in the neurofibromin gene that can reduce levels of the neurofibromin protein. While the mechanisms have yet to be fully elucidated, loss of Nf1 may alter neuronal circuit activity leading to changes in behavior and susceptibility to cognitive and behavioral comorbidities. Here we show that mutations decreasing Nf1 expression alter motor behaviors, impacting the patterning, prioritization, and behavioral state dependence in a Drosophila model of Nf1. Loss of Nf1 increased spontaneous grooming in male and female flies. This followed a nonlinear spatial pattern, with Nf1 deficiency increasing grooming of certain body parts differentially, including the abdomen, head, and wings. The increase in grooming could be overridden by hunger in foraging animals, demonstrating that the Nf1 effect is plastic and internal state dependent. Stimulus-evoked grooming patterns were altered as well, suggesting that hierarchical recruitment of grooming command circuits was altered. Yet loss of Nf1 in sensory neurons and/or grooming command neurons did not alter grooming frequency, suggesting that Nf1 affects grooming via higher-order circuit alterations. Changes in grooming coincided with alterations in walking. Flies lacking Nf1 walked with increased forward velocity on a spherical treadmill, yet there was no detectable change in leg kinematics or gait. These results demonstrate that loss of Nf1 alters the patterning and prioritization of repetitive behaviors, in a state-dependent manner, without affecting low-level motor functions.
- attention-deficit/hyperactivity disorder
- autism spectrum disorder
- foraging
- kinematics
- Nf1
- perseveration
- ras
- repetitive behavior
- walking
Significance Statement
Neurofibromatosis type 1 (NF1) is associated with an increased risk of cognitive and behavioral disorders, yet the underlying neuronal mechanisms remain poorly understood. Our study utilizes a Drosophila model to demonstrate that loss of neurofibromin (Nf1) expression impacts motor behavior and the prioritization of repetitive actions, such as grooming, in a hunger state-dependent manner. Our experiments also suggest that alterations in neuronal circuit activity due to the loss of Nf1 influence behavior without impairing motor coordination. Understanding how Nf1 loss affects motor function can reveal the broader neuronal mechanisms contributing to cognitive impairment, providing valuable insights for developing therapeutic strategies.
Introduction
Human genetic disorders can alter brain circuits and behavior, affecting processes ranging from motor function to cognition. The mutations underlying such disorders impact cellular processes such as neuronal growth and differentiation, synaptic transmission and synaptic plasticity, and neuronal excitability (Costa et al., 2002; Wang et al., 2005; Anastasaki et al., 2022). These cellular changes cascade through the nervous system, altering complex behaviors. Nervous system function is altered in fragile X syndrome, Rett syndrome, Angelman syndrome, and Williams syndrome, among others, impacting cognition and behavior (Meyer-Lindenberg et al., 2005; Contractor et al., 2015; Ip et al., 2018; Sun et al., 2019). Given the heterogeneity of neuronal subpopulations—excitatory versus inhibitory, etc.—as well as the plethora of connectivity architectures across brain regions, a given mutation could produce a range of effects across cells and circuits. Unraveling the complex contributions to behavioral symptoms will require understanding both how genetic mutations alter the function of individual neurons and how they interact at the circuit and systems levels.
Neurofibromatosis type 1 (Nf1) is a genetic disorder that results from mutations in the NF1 gene that encodes the neurofibromin protein. This disorder causes tumor formation and predisposition to cognitive and behavioral symptoms. In addition to its core (mainly cutaneous) symptoms, patients frequently experience comorbidities such as attention-deficit/hyperactivity disorder (ADHD; ∼50%; Hyman et al., 2005) and autism spectrum disorder (ASD; ∼25%; Garg et al., 2015), along with seizures, poor visuospatial skills, executive function deficits, disrupted sleep, repetitive behaviors, and/or pain (Hyman et al., 2005; Hyman et al., 2006; Diggs-Andrews and Gutmann, 2013; Walsh et al., 2013; Constantino et al., 2015; Garg et al., 2015; Morris et al., 2016; Plasschaert et al., 2016; Eijk et al., 2018). Multiple genetic mutations underlying Nf1 result in loss of expression and/or function in Nf1. How Nf1 deficiency alters neuronal activity through effects on interacting circuits across the nervous system is not well understood.
As Nf1 is associated with a wide range of cognitive and behavioral symptoms, understanding how loss of NF1 function alters repetitive, sequenced behaviors may provide insight into how genetic disorders impact nervous system function. Alterations in complex motor function may contribute to the ADHD and ASD symptoms in Nf1. Such motor functions include grooming and locomotion, which are repetitive, sequenced behaviors that follow hierarchical syntax rules. Animals as diverse as rats and flies groom their body parts in a hierarchical sequence, beginning at the head and proceeding caudally down the body (Berridge et al., 1987; Seeds et al., 2014; Hampel et al., 2015; Kalueff et al., 2016). The sequence is dynamically modulated by sensory feedback from each body part (Seeds et al., 2014; Hampel et al., 2015, 2017). Self-grooming is a useful model of behavioral dysregulation in animal models for disorders such as ASD—multiple ASD risk factor genes alter grooming when mutated in animal models (Kalueff et al., 2016; King et al., 2016, 2020). Similarly, walking patterns are the result of sequential, coordinated leg movements that modulate forward walking, backward walking, turning, speed changes, and stopping (Bidaye et al., 2018). Children with ADHD exhibit alterations in gait (Meachon et al., 2023), suggesting that disease-driving mutations can affect motor coordination while also driving behavioral symptoms.
Drosophila expresses a Nf1 ortholog, and deficiency in this protein leads to behavioral and physiological phenotypes reminiscent of those in mammals. Drosophila nf1 mutants exhibit learning and memory deficits (Guo et al., 2000; Buchanan and Davis, 2010; Gouzi et al., 2011; Georganta et al., 2021), circadian rhythm and sleep disruption (Williams et al., 2001; King et al., 2016; Botero et al., 2021; E. B. Brown et al., 2022), reduced body size (The et al., 1997), increased grooming (King et al., 2016, 2020), impaired jump reflex habituation (Fenckova et al., 2019), social (courtship) alterations (Moscato et al., 2020), altered metabolism (Maurer et al., 2020; Botero et al., 2021; E. B. Brown et al., 2022), and tactile hypersensitivity (Dyson et al., 2022). Little is known about how behavioral sequences are altered in Nf1 and whether the different phenotypes influence one another. To approach these questions, we examined how neurofibromin deficiency alters the temporal organization and prioritization of sequenced behaviors in Drosophila, including grooming, locomotion, and feeding.
Materials and Methods
Drosophila maintenance
Flies were raised on a cornmeal/agar food medium, housed in incubators at 25°C and 60% relative humidity, and kept on a 12:12 h light/dark cycle according to standard protocols. Behavioral assays were performed using 3–8-d-old flies. Except where indicated, males were used (to prevent egg accumulation). The nf1P1 mutation was backcrossed six generations into the wCS10 genetic background. The Nf1 RNAi line was obtained from the Vienna Drosophila Resource Center (VDRC #109637). UAS-dicer2 was coexpressed to potentiate the RNAi effect (Pfeiffer et al., 2012) and was included with lines that contain UAS-Nf1RNAi (except where noted in control experiments to isolate and test UAS-dcr2 independently). The empty attP control line (VDRC #60100) was used in Gal4/+ control crosses to match the genetic background across groups. The following lines used in this study were obtained from the Bloomington Drosophila Stock Center (BDSC): R57C10-Gal4 (39171), nSyb-Gal4 (51635), repo-Gal4 (7415), R81E10-Gal4 (BDSC #48367), R71D01-Gal4 (39579), R23A07-Gal4 (49010), R18C11-Gal4 (48808), R31H10-Gal4(48104), R30B01-Gal4 (49517), and R50B07-Gal4 (38729).
Grooming analysis
Grooming was quantified as previously described (King et al., 2016, 2020). Flies were placed into an open-field arena, 25.4 mm in diameter and 2.85 mm in height, consisting of an opaque (white) PLA lateral boundary covered on the top and bottom with two clear polycarbonate sheets. The apparatus was illuminated from below with white light-emitting diodes that were filtered through a sheet of white acrylic; light intensity was measured at 720 lm/m2 in the location of the fly. A camera (FLIR Teledyne Blackfly S) fitted with a 25 mm lens (Edmund Optics) was mounted above each arena. A single male fly was placed in the arena with an aspirator and recorded at 7.5 fps, 1,616 × 1,240 with lossless Motion JPEG 2000 compression. Five-minute videos were recorded at several intervals, starting at 0, 30, 60, and 150 min following introduction into the open-field arena. Grooming was manually scored via frame-by-frame analysis, recording the start and stop frames for each grooming event. Grooming events were categorized according to body part: front legs, head/eye, abdomen, wings, or hind legs. The percentage of time spent grooming, bout count, and bout duration was calculated. Grooming was graphed in aggregate or by individual body parts, as appropriate. Ethograms were generated with a custom Python script (https://github.com/sethtomchik/) and heat maps were plotted using GraphPad Prism. Stimulus-evoked grooming was performed as previously described (Seeds et al., 2014). Each fly was coated in Reactive Yellow 86 dust and transferred to an open-field area with a mesh floor to allow dust to fall through. Flies were allowed to clean for a period of time and then anesthetized with CO2, and the head, thorax, abdomen, and wings were separated and photographed individually. Using the color selection tool in Adobe Photoshop, yellow pixels were selected and then manually corrected as needed to ensure that dust particles were accurately selected. Yellow pixels were quantified, and their number was divided by the total number of pixels to calculate the percentage of dust coverage on each body part. The percentage coverage for each body part/fly/time point was normalized to the median value of the dust coverage at Time 0.
Open-field locomotion tracking
Locomotion was tracked (in videos collected as above) using MATLAB with DLTdv8 (Hedrick et al., 2023). Accuracy of the tracking was visually confirmed, and any errors in the xy position tracking were manually corrected. Total distance and mean walking speed were calculated over the duration of the video from the frame-by-frame xy positions.
High-resolution 3D leg kinematics analysis
Locomotor gait and kinematics were examined using a tethered fly preparation as previously described (Seelig et al., 2010; Berendes et al., 2016; Loesche and Reiser, 2021; Sapkal et al., 2024). Seven- to ten-day-old male flies were tethered to a 34 gauge needle with ultraviolet light-cured glue and placed on a spherical treadmill (6 mm diameter) suspended in a stream of compressed air. The flies were placed on the ball with a minimum wait time after tethering and allowed up to 5 min of recovery before starting the experiment. The compressed air was passed through an in-line heating element to bring the local temperature on the ball up to 32°C, inducing high-speed spontaneous walking. Rotational velocity of the ball was monitored in real time using two orthogonally placed motion sensors at 50 Hz. Each trial was triggered in a closed loop fashion when the forward velocity crossed a threshold empirically determined to signify sustained walking. Each fly could trigger a maximum of 10 trials, but flies that triggered >5 trials were included in the final dataset for analysis. Each trial was 7 s long. Eight cameras (FLIR BFS-U3-16S2M-CS) fitted with InfiniStix 194100 lenses and near-infrared bandpass filters (Midopt BP850) were placed surrounding the ball so that all legs were continuously visible from at least one pair of cameras. The fly was illuminated with a custom infrared ring emitting focused light to the plane of the ball. The cameras, infrared light source, and the ball tracker were all synchronized and triggered at 200 Hz by an Arduino microcontroller, with camera exposure time set to 200 µs. The fly was recorded with a resolution of 1,440 × 1,072 pixels. Each fly was left on the ball for a maximum of 20 min.
Capillary feeding assay
Feeding was quantified in adult male flies using a capillary feeding assay (Ja et al., 2007). Individual flies were placed into chambers (46 × 7 mm) containing 1% agar at the bottom and a glass capillary at the top. The glass capillaries contained an aqueous food solution (5% sucrose + 5% yeast extract) and dye to track food consumption. One group of 10 flies was fed for 150 min, one group of 10 flies was starved for 60 min, and one group of 10 flies was starved for 150 min prior to providing food for the experiment. Total feeding was measured in a 30 min window and analyzed in Fiji.
Immunohistochemistry
Five- to seven-day-old R18C11-Gal4 adult brains were dissected in 1% paraformaldehyde in S2 medium and processed as previously described (Jenett et al., 2012). Samples were stained with primary antibodies for 3 h at room temperature and at 4°C overnight, followed by secondary antibodies for 3 h at room temperature and 5 d at 4°C. All antibody incubations were performed in a PAT3 blocking solution, consisting of 0.5% Triton X-100 and 0.5% BSA in 1× PBS, supplemented with 3% normal goat serum. Samples were mounted in VECTASHIELD (Vector Laboratories) for analysis. The following antibodies were used: rabbit anti-GFP (1:1,000, Invitrogen), mouse anti-nc82 (1:50, DSHB), goat anti-rabbit IgG, and goat anti-mouse IgG (1:800, Alexa Fluor 488 or Alexa Fluor 633, respectively, Invitrogen). Images were obtained using a Leica SP8 Laser Scanning Confocal microscope with the LAS X software.
Quantitative polymerase chain reaction (qPCR)
Heads were collected from adult flies (5–10 d post-eclosion; 50–100 heads per genotype). Each sample was added to QIAzol Lysis Reagent (Qiagen, catalog #79306) and homogenized using a mortar and pestle. RNA was isolated with the Qiagen RNeasy Lipid Tissue Mini Kit. Complimentary DNA library was constructed using LunaScript RT Master Mix (New England Biolabs, catalog #E3025L) with the Random Primer Mix (New England Biolabs, catalog #S1330S). Nf1 gene expression was quantified in fly head samples using Luna Universal qPCR Master Mix (New England Biolabs, catalog #M3003L). Primer sequences were as follows: 5′-CTTTTGGCACGTTTCGAGGAT-3′ (Nf1_F), 5′-GGTAGCGCGATATGTGGATCA-3′ (Nf1_R), 5′-ATGCTAAGCTGTCGCACAAATG-3′ (Rpl32_F), and 5′-GTTCGATCCGTAACCGATGT-3′ (Rpl32_R). Three biological replicates, with three technical replicates each, were analyzed for each genotype. Analysis of gene expression was performed using standard ΔΔCt analysis. Gene expression was normalized to Rpl32, and log2 fold change was calculated.
Statistical analysis
Normality of data was assessed with the Shapiro–Wilk normality test. In figures, box plots graph the median as a line and the interquartile range (IQR) as a box, and whiskers extend to the min/max values. Hypothesis testing was carried out using t tests or ANOVA followed by Šidák's multiple-comparison tests (parametric) or Wilcoxon rank-sum test or Kruskal–Wallis omnibus test followed by Dunn multiple-comparison tests (nonparametric). Two-way comparisons were carried out with a two-way ANOVA followed by Šidák's multiple-comparison tests. For RNAi analysis, the experimental group was compared with both heterozygous genetic controls and considered positive if it significantly differed from both controls in the same direction. Statistics and graphing were carried out with GraphPad Prism, version 10.1.1.
Results
Neurofibromin deficiency drives distinct spatiotemporal grooming patterns across the body
Neurofibromin deficiency increases the time that flies spend grooming (King et al., 2016, 2020), potentially reflecting alterations in the activity of command circuits that regulate this motor behavior (Seeds et al., 2014; Hampel et al., 2015, 2017; Zhang and Simpson, 2022). To dissect the dynamic role that neurofibromin plays in the nervous system, we first carried out behavioral analysis of grooming over time. The nf1P1 genomic mutant harbors a large deletion encompassing most of the Nf1 locus [including the catalytic GAP-related domain (GRD)], along with several nearby E(spl) genes (The et al., 1997). We first confirmed that loss of Nf1 in the nf1P1 mutant increases grooming in both males and females (Fig. 1A). This was due specifically to Nf1, as pan-neuronal rescue of wild-type Nf1 via Gal4/UAS-mediated expression in the nf1P1 background restored normal grooming levels (Fig. 1B). To examine how Nf1 affects the temporal evolution of grooming patterns, we placed individual flies into open-field arenas and quantified grooming over 5 min time windows at 0, 30, 60, and 150 min (Fig. 1C). When placed into the arena, control flies exhibited an initial burst of grooming (0 min) that decreased by 30 min (Fig. 1D,F). In contrast, nf1P1 mutants groomed at elevated levels for at least 60 min (Fig. 1E,F). Depending on the time, mean grooming in the nf1 mutants increased 72–403% relative to controls (Extended Data Table 1). By 150 min, grooming in both control and nf1 mutants had decreased and were no longer significantly different from one another (Fig. 1F).
Nf1 deficiency alters grooming frequency across time. Box plots: Median, line; box, IQR; whiskers, min/max values; individual data points: circles. A, Loss of Nf1 increased grooming in both males and females. B, Pan-neuronal rescue of the wild-type Nf1 transgene in the nf1P1 background [nSyb-Gal4>UAS-Nf1(wt);nf1P1/nf1P1] and controls. The no-transgene control (wCS10) and nf1P1 mutants are from different animals than plotted in panel A. C, Time course of video collection and example of data (5 min grooming ethograms, replicated from panel D) visualized at two different time points. D, Ethograms of grooming in control (wCS10) flies, showing each grooming bout across animals, with the groomed body part color-coded. E, Ethograms of grooming in nf1P1 mutants. F, Grooming time in nf1P1 deletion mutants and wCS10 controls. G, Grooming time in nf1E1 nonsense point mutants and iso2,3 controls. H, Grooming time with pan-neuronal Nf1 knockdown (R57C10-Gal4>UAS-Nf1RNAi) and controls. *p < 0.05; *p < 0.01; ***p < 0.001 re: control(s) (Šidák). Extended Data Figure 1-1 for more details.
Figure 1-1
(A) Nf1 mRNA expression in fly heads quantified by qPCR. Each point is the mean of three technical replicates. *p < 0.05 (Šidák; n = 3 biological replicates). (B) Pan-neuronal overexpression of UAS-dcr2, compared to heterozygous controls and knockdown of UAS-Nf1RNAi (including UAS-dcr2). ***p < 0.001 (Šidák). (C) Effect of knocking down Nf1 in glia on grooming. ***p < 0.001 (Šidák). Download Figure 1-1, TIF file.
Tables - Statistical analysis
Download Tables - Statistical analysis, XLSX file.
To further examine whether the elevation in grooming occurs via Nf1 loss of function, we tested a nonsense nf1 mutant and RNAi. The nf1E1 mutant harbors a point mutation upstream of the catalytic GRD, resulting in loss of protein expression (Walker et al., 2006). This mutant exhibited increases in grooming relative to its genetic background control (iso2,3), with a quantitative reduction in the difference at the 150 min time point (Fig. 1G). This pattern was similar to the nf1P1 mutant (Fig. 1F). To isolate the neuronal contributions to this effect, we knocked down Nf1 pan-neuronally with RNAi. This RNAi line reduced Nf1 mRNA in fly heads (including both neurons and non-neuronal cells) by 61% (Extended Data Fig. 1-1A, Table 1). Upon introduction of these flies to the open-field arena (0 min), there was significantly more grooming in the experimental group compared with the heterozygous driver (Gal4/+) and effector (UAS/+) genetic controls (Fig. 1H). Flies with pan-neuronal Nf1 knockdown displayed elevated grooming levels at all time points, though the time spent grooming decreased quantitatively as time progressed (Fig. 1H; Extended Data Table 1). In Nf1 RNAi experiments, dicer-2 was overexpressed to enhance the knockdown efficacy (Dietzl et al., 2007). To validate the specificity of the RNAi, we examined the effect of expressing dicer-2 in isolation. This revealed no effect on grooming when dicer-2 was pan-neuronally overexpressed (Extended Data Fig. 1-1B, Table 1). Furthermore, knockdown of Nf1 in glia (via repo-Gal4) did not affect grooming frequency (Extended Data Fig. 1-1C, Table 1). Overall, these results indicate that loss of Nf1 increased spontaneous grooming for at least 1 h due to neuronal effects, with the effect decreasing over time in the open field.
Grooming comprises a set of distinct cleaning movements that clear different body parts (e.g., legs, head, thorax/abdomen, wings; Szebenyi, 1969; Phillis et al., 1993; Seeds et al., 2014). Loss of Nf1 likely exerts behavioral effects via actions on distributed neuronal circuits (King et al., 2016, 2020), raising the question of how homogeneous or heterogeneous the neuronal effects of Nf1 deficiency are across neuronal circuit elements. To approach this, we investigated how Nf1 deficiency affects the temporal evolution of grooming across different body parts. Grooming was scored in nf1P1 mutants, nf1E1 mutants, and pan-neuronal Nf1 RNAi. nf1P1 mutants exhibited significantly increased grooming of the abdomen at 30, 60, and 150 min following introduction to the arena (Fig. 2A,G). The front and back legs exhibited a small Nf1-dependent increase relative to the control, at only one time point each (Extended Data Fig. 2-1). Grooming of other body parts, including the head and wings, was not statistically significant from controls (Fig. 2A,D,J; Extended Data Fig. 2-1, Table 2).
Nf1 deficiency alters grooming in a body part-specific manner. A, Heat map of grooming time across body parts and time in controls (wCS10) and nf1P1 mutants. B, Heat map of grooming time across body parts and time in controls (iso2,3) and nf1E1 mutants. C, Heat map of grooming time across body parts and time with pan-neuronal Nf1 knockdown (R57C10-Gal4>UAS-Nf1RNAi) and controls. D, Head grooming across time in controls and nf1P1 mutants. Box plots: median, line; box, IQR; whiskers, min/max values; individual data points, circles. E, Head grooming across time in controls and nf1E1 mutants. F, Head grooming with pan-neuronal Nf1 knockdown (R57C10-Gal4>UAS-Nf1RNAi). G, Abdomen grooming across time in controls and nf1P1 mutants. H, Abdomen grooming across time in controls and nf1E1 mutants. I, Abdomen grooming with pan-neuronal Nf1 knockdown (R57C10-Gal4>UAS-Nf1RNAi). J, Wing grooming across time in controls and nf1P1 mutants. K, Wing grooming across time in controls and nf1E1 mutants. L, Wing grooming with pan-neuronal Nf1 knockdown (R57C10-Gal4>UAS-Nf1RNAi). Extended Data Figures 2-1 and 2-2 for more details.
Figure 2-1
Nf1 deficiency produces little effect on leg grooming. Box plots: median = line, box = interquartile range; whiskers = min/max values, individual data points: circles. *p < 0.05, ***p < 0.001 (Šidák). (A) Front leg grooming in controls (wCS10) and in nf1P1 mutants across time. (B) Back leg grooming in controls (wCS10) and in nf1P1 mutants across time. (C) Front leg grooming in controls (iso2,3) and in nf1E1 mutants across time. (D) Back leg grooming in controls (iso2,3) and in nf1E1 mutants across time. (E) Front leg grooming over time with pan-neuronal Nf1 knockdown (R57C10-Gal4 > UAS-Nf1RNAi). (F) Front leg grooming over time with pan-neuronal Nf1 knockdown (R57C10-Gal4 > UAS-Nf1RNAi). Download Figure 2-1, TIF file.
Figure 2-2
Nf1 deficiency altered grooming initiation and duration, increasing bout count and bout duration. Box plots: median = line, box = interquartile range; whiskers = min/max values, individual data points: circles. *p < 0.05, ***p < 0.001 (Šidák). (A) Abdomen grooming bout count across time in controls and nf1P1 mutants. (B) Abdomen grooming bout duration across time in controls and nf1P1 mutants. Download Figure 2-2, TIF file.
To further probe Nf1 deficiency effects across grooming circuits, we examined nf1E1 mutants and pan-neuronal Nf1 knockdown. Upon introducing the flies to the open-field arena, nf1E1 mutants exhibited increases in grooming of the head, abdomen, and wings (Fig. 2B,E,H,K; Extended Data Fig. 2-1, Table 2). The head and abdomen grooming remained significantly elevated for at least 30 min before falling, while wing grooming was elevated at the first time point (0 min) only. There was a significant increase in front leg grooming at 0 min, along with increased back leg grooming at 0, 30, and 150 min (Extended Data Fig. 2-1). When knocking down Nf1 pan-neuronally, the pattern was similar to nf1E1 mutants across the head, abdomen, and wing. There was a significant increase in head grooming compared with the Gal4/+ and UAS/+ genetic controls which persisted for 30 min before decreasing (Fig. 2C,F; Extended Data Fig. 2-1). Similarly, there were increases in abdomen and wing grooming (Fig. 2C,I,L) that persisted for 30–60 min. There were no consistent changes in front or back leg grooming (Extended Data Fig. 2-1). Overall, these data demonstrate a consistent increase in abdomen grooming with loss of Nf1, with head and wing grooming often elevated as well. Leg grooming was not consistently elevated, with only some genotypes exhibiting differences at some time points. Thus, loss of Nf1 produces heterogeneous effects on grooming across different body parts over time.
Changes in total grooming time result from changes in either frequency of grooming initiation or duration of grooming bouts. To determine which of these underlies the Nf1 effect, we analyzed individual grooming bouts in nf1P1 mutants, quantifying the bout count and bout duration in mutants and controls. This analysis focused on abdomen grooming, as it is significantly elevated in nf1P1 mutants and consistent across loss-of-function alleles. There were significant increases in both bout count and bout duration in nf1P1 mutants (Extended Data Table 3), which occurred across multiple time points (Extended Data Fig. 2-2). This suggests that loss of Nf1 enhances the activation of grooming circuits, resulting in both more frequent grooming initiation and increased perseveration (i.e., sustained grooming).
Neurofibromin deficiency modulates a state-dependent behavioral switch from grooming- to foraging-dominant behavioral modes
As Nf1 deficiency increases grooming, grooming command circuits are aberrantly engaged. To test how this behavior is prioritized in the face of competing stimuli, we examined the hunger state dependence of spontaneous grooming. When food is withheld for some time, flies increase their locomotion to forage for food, a phenomenon known as starvation-induced hyperactivity (Lee and Park, 2004; Landayan et al., 2018; Yurgel et al., 2019). When they are walking, they cannot be grooming. Thus, we wondered whether the reduction in grooming over time in the open field (Fig. 1C; which lacks food) could be attributed to altered hunger state. To test this, we first compared flies in arenas without food to those with ad libitum food access throughout the experiment (Fig. 3A). In the presence of food, nf1P1 mutants displayed significantly higher grooming compared with controls at all time points, including the longest time point (150 min; Fig. 3A; Extended Data Table 3). Thus, nf1 mutant flies continue to groom significantly more than controls if food is available. This suggests that when food is absent, grooming may decrease in a hunger state-dependent manner.
Nf1 deficiency modulated a state-dependent behavioral switch from grooming to locomotion. Box plots: median, line; box, IQR; whiskers, min/max values; individual data points, circles. *p < 0.05; ***p < 0.01; ***p < 0.001 (Šidák). A, Quantification of grooming (% time) in an open-field arena when solid food was provided ad libitum (+Food), comparing control (wCS10) flies with nf1P1 mutants. B, A screen capture of fly locomotion tracking, with xy position tracks over 5 min for a representative control fly and nf1P1 mutant at 0 and 150 min. C, Total distance traveled for control flies and nf1P1 mutants. D, Mean walking speed for control flies and nf1P1 mutants. E, Diagram of the capillary feeding assay and protocol to test homeostatic feeding. Food was withheld at starting at t = 0, and food consumption was measured immediately (0 min) or following 60 or 150 min of starvation. F, Feeding in controls and nf1P1 mutants, comparing feeding across time. ***p < 0.001 re: 0 min (Šidák). Extended Data Figure 3-1 for more details.
Figure 3-1
Locomotion in representative controls and nf1P1 mutants over time. (A) xy position tracks for a control (wCS10) fly (same as in Figure 4B) over five-min videos at 0 min, 30 min, 60 min, and 150 min following introduction to the open-field arena. (B) xy position tracks for an nf1P1 mutant (same as in Figure 4B). (C) Grooming ethograms (top) and locomotor activity traces for a control fly at 0 min, 30 min, 60 min, and 150 min following introduction to the open field arena. (D) Grooming ethograms (top) and locomotor activity traces for an nf1P1 mutant. Download Figure 3-1, TIF file.
To investigate whether the decrease in grooming frequency at longer time points could be driven by foraging, we quantified locomotion. Locomotion was tracked over time, analyzing distance traveled and walking speed. In open-field arenas, flies exhibit an initial stage of high walking activity that gradually decreases over time (considered to be exploratory behavior; Connolly, 1967; Liu et al., 2007). In control flies, we observed this as higher locomotor activity upon introduction to the arena, which dropped within 30 min (Fig. 3B,C; Extended Data Table 3). This was reflected in both a significant decrease in the distance traveled and speed from the 0 to 30 min time points (Fig. 3C,D). While the median locomotor activity decreased in nf1P1 mutants across these time points, the effect did not reach statistical significance (Fig. 3C,D; Extended Data Fig. 3-1). However, from 60 to 150 min, a significant increase in locomotion was observed in the nf1P1 mutants for both distance and speed (Fig. 3C,D; Extended Data Fig. 3-1). These data are consistent with the interpretation that the reduction in grooming frequency results from an increase in locomotion, providing additional evidence that neurofibromin deficiency prompts a behavioral shift toward foraging behavior.
Grooming decreased over time in the open field, while locomotion increased, raising the question of whether hunger state drives the behavioral changes. To determine whether hunger was a driver of increased locomotion, we examined food intake for signs of homeostatic feeding, characterized by a rebound in feeding driven by negative energy balance [50, 51]. Food intake was measured after periods of food deprivation using a capillary feeding assay (Ja et al., 2007). Flies deprived of food for 60 or 150 min were compared with flies provided food ad libitum. These time points corresponded to the longer two time points of food deprivation experienced in the open-field experiments, when putative negative energy balance began to trigger starvation-induced hyperactivity. In control flies, we did not observe a significant difference in the amount of liquid food consumed between the fed and starved groups at the 60 min time point (Fig. 3F; Extended Data Table 3). However, at 150 min, starved control flies consumed significantly more food compared with the fed controls (Fig. 3F). In the case of nf1P1 mutants, homeostatic feeding was observed by 60 min (Fig. 3F), demonstrating that food deprivation increased hunger faster in Nf1-deficient flies. This increase in hunger corresponds to a higher metabolic rate in nf1P1 mutants (Botero et al., 2021) and could drive starvation-induced hyperactivity along with decreasing grooming.
Loss of Nf1 in sensory and grooming command neurons modulates grooming in a circuit-specific manner
Grooming of each body part is controlled by discrete sensory neurons and grooming command neurons (Seeds et al., 2014; Hampel et al., 2015, 2020; L. Guo et al., 2022; Zhang and Simpson, 2022), raising the question of whether loss of Nf1 affects the sensory neurons and/or command neurons directly. To test this, we employed a targeted approach involving the knockdown of Nf1 in selected neuronal subsets (King et al., 2020). We considered a range of plausible circuit architectures interconnecting the command circuits. Nf1 could affect the sensory input to the command neurons, the command neurons themselves, and interactions between the sensory and command neurons (Fig. 4A) or via interconnected neurons across more broad swaths of the central nervous system. Based on these hypothetical network architectures, we generated four testable models: (1) sensory input modulation, (2) command circuit modulation, (3) sensory + command circuit modulation, and (4) systems-level modulation. These models were challenged with a series of circuit-specific Nf1 manipulations.
Knocking down Nf1 in sensory neurons and/or grooming command circuits shifted the pattern of grooming without affecting total grooming time. Box plots: median, line; box, IQR; whiskers, min/max values; individual data points, circles. *p < 0.05; **p < 0.01; n.s., not significant (Šidák). A, Simplified diagram of sensory neurons and antennal grooming command neurons, potential sites of modulation by Nf1 deficiency. RNAi was targeted to sensory neurons, command neurons, or both. B, Effect of Nf1 knockdown in sensory neurons on head grooming (R81E10-Gal4>UAS-Nf1RNAi). Experimental flies were compared with heterozygous Gal4/+ and UAS/+ controls. C, Effect of Nf1 knockdown in eye/head grooming command neurons on head grooming (R23A07-Gal4>UAS-Nf1RNAi). D, Effect of Nf1 knockdown in wing grooming command neurons on wing grooming (R31H10-Gal4>UAS-Nf1RNAi). E, Effect of Nf1 knockdown in antennal grooming command neurons on head grooming (R18C11-Gal4>UAS-Nf1RNAi). F, Expression pattern of neurons labeled by R18C11-Gal4, focusing on the central brain. The box highlights the inset shown in panel G. GFP:green; brp:magenta. G, Expanded view of the boxed region from panel F, including the somata of antennal descending neurons (white arrowheads). H, Effect of Nf1 knockdown in antennal descending neurons on head grooming using two different drivers (R71D01-Gal4 or R26B12-Gal4). I, Effects of Nf1 knockdown in sensory neurons (R30B01-Gal4), wing grooming command neurons (R50B07-Gal4), and both, on wing grooming. J, Effects of Nf1 knockdown in sensory neurons (R30B01-Gal4), eye/head command neurons (R23A07-Gal4), and both, on head grooming. See Extended Data Figure 4-1 for more details.
Figure 4-1
Knocking down Nf1 in sensory neurons, command neurons, and sensory + command neuron combinations does not affect grooming frequency (total grooming time). Box plots: median = line, box = interquartile range; whiskers = min/max values, individual data points: circles. *p < 0.05, **p < 0.01, n.s. = not significant (Šidák). (A) Effect of Nf1 knockdown in sensory neurons on grooming time (R81E10-Gal4> UAS-Nf1RNAi). (B) Effect of Nf1 knockdown in eye/head grooming command neurons on grooming time (R23A07-Gal4 > UAS-Nf1RNAi). (C) Effect of Nf1 knockdown in antennal grooming command neurons on grooming time (R18C11-Gal4 > UAS-Nf1RNAi). (D) Effect of Nf1 knockdown in wing grooming command neurons on grooming time (R31H10-Gal4 > UAS-Nf1RNAi). (E) Effects of Nf1 knockdown in sensory neurons (R30B01-Gal4), wing grooming command neurons (R50B07-Gal4), and both, on grooming time. (F) Effects of Nf1 knockdown in sensory neurons (R30B01-Gal4), eye/head command neurons (R23A07-Gal4), and both, on grooming time. Download Figure 4-1, TIF file.
To test whether loss of Nf1 in sensory neurons alone increased grooming frequency, we knocked Nf1 down in sensory neurons with RNAi using two Gal4 drivers (R81E10 and R30B01-Gal4; Pfeiffer et al., 2012; Hampel et al., 2017). These drivers label multiple subsets of sensory neurons including mechanosensory bristles, the Johnston's organ, chemosensory receptors, chordotonal organs, and campaniform sensillae (Hampel et al., 2017). There was no significant effect on grooming frequency when Nf1 was knocked down in the sensory neurons labeled by either of these drivers (Fig. 4B,I,J; Extended Data Fig. 4-1). Overall, these data suggested that Nf1 deficiency in sensory neurons did not account for the increased grooming when Nf1 was reduced across the nervous system. Next, we tested the impact of Nf1 knockdown in command circuits on grooming frequency and temporal structure, utilizing six different Gal4 drivers (R23A07, R18C11, R31H10, R71D01, R26B12, and R50B07-Gal4) that encompass grooming command neurons. These drivers include command neurons for circuits that control grooming of the eye/head (R23A07), antennae [R18C11, R71D01 (aBN1), and R26B12 (aBN2)], and wings (R31H10 and R50B07; Seeds et al., 2014; Hampel et al., 2015, 2017; Zhang and Simpson, 2022). No significant changes in overall grooming frequency were observed when Nf1 was knocked down using any of these drivers (Fig. 4C–E,H; Extended Data Fig. 4-1, Table 4). Examining grooming of each body part individually, we found that knocking down Nf1 in antennal descending neurons (via R18C11) modestly biased grooming toward the head (Fig. 4E). This driver labels three pairs of antennal descending neurons that modulate antennal grooming (Fig. 4F,G; Hampel et al., 2015). The behavioral effect of Nf1 knockdown in these neurons suggests that loss of Nf1 in one component of the neuronal circuits driving grooming may affect grooming patterning. In contrast, knocking down Nf1 in aBN1 and aBN2 antennal grooming command neurons did not detectably alter head grooming (Fig. 4H). Thus, there is a large increase in grooming with broad knockdown likely representing either additive effects across multiple components of the grooming command circuits and/or their inputs.
Next we tested whether Nf1 functions additively in pairs of neuron types within the grooming circuitry, knocking Nf1 down in both command neurons and and their sensory inputs. This was done using one Gal4 driver and/or in grooming command neurons with a second Gal4 driver. Two combinations of sensory + command neuron drivers were tested: R30B01 + R50B07-Gal4 and R30B01 + R23A07-Gal4. The R30B01-Gal4 driver provides broad coverage of sensory neurons innervating mechanosensory bristles across the eye, head, abdomen, wing, notum, and leg, as well as sensory neurons innervating the Johnston's organ, chemosensory receptors, chordotonal organs, and campaniform sensillae (Hampel et al., 2017). Thus, the R30B01 + R50B07-Gal4 combination includes both mechanosensory inputs to the grooming command circuits (R30B01) and the wing grooming command circuit (R50B07). Similarly, the R30B01 + R23A07 covers mechanosensory neurons and an eye/head grooming command circuit. Neither of these two driver pairs significantly altered grooming frequency when used to knock down Nf1 (Fig. 4H–J). These data suggest that Nf1 deficiency in sensory circuits does not interact additively with command neurons and that it is required more broadly and/or in higher-level neurons.
Loss of Nf1 alters the pattern and prioritization of stimulus-evoked grooming behaviors
Loss of neurofibromin increased the frequency of spontaneous grooming, but its effect on stimulus-evoked grooming is unknown. When flies are covered with a fine layer of dust, they vigorously groom to remove the dust (Phillis et al., 1993; Seeds et al., 2014). This stimulus-evoked grooming follows a temporal progression (approximately cephalocaudal) resulting from hierarchical command circuit recruitment (Seeds et al., 2014). To test whether neurofibromin alters the pattern and prioritization of sensory-evoked grooming movements, we dusted flies and quantified dust removal. Within 35 min of dusting, controls removed much of the dust from their head, thorax, wings, and abdomen (Fig. 5; Extended Data Table 5). nf1P1 mutants similarly removed most of the dust from their head, thorax, and abdomen, but wing cleaning was significantly reduced—i.e., the dust was removed from the wings more slowly in mutants. This represented a departure from the normal cephalocaudal grooming sequence, wing grooming shifted down in priority. Loss of Nf1 affected spontaneous grooming and stimulus-evoked grooming in different ways. Spontaneous grooming was elevated in the abdomen, head, and wings (Fig. 2). Yet the deficit in stimulus-evoked grooming was focused on the wings. Overall, these data suggest that both the patterning and prioritization of stimulus-evoked grooming was altered by loss of neurofibromin. Furthermore, since different body parts were affected in different conditions, loss of Nf1 affects multiple circuits, potentially including those upstream of the grooming command neurons.
Loss of Nf1 altered the temporal evolution of stimulus-evoked grooming. Box plots: median, line; box, IQR; whiskers, min/max values; individual data points, circles. A, Diagram of the experimental protocol. Flies were dusted, and the amount of dust remaining on each body part was imaged at t = 0, 8, 25, and 35 min. B, Reference images of different body parts immediately after dusting, with images showing dust coverage after 0, 8, 25, and 35 min. C, Dust removal in control (wCS10) flies. Dust coverage is the fraction of dust relative to those imaged at time 0. *p < 0.05; **p < 0.01, ***p < 0.001 re: time 0 (Šidák). D, Dust removal in nf1P1 mutants, plotted as in panel C. E, Time course of dust removal (same data as in panels C, D) comparing controls and nf1P1 mutants at each time point. *p < 0.05; **p < 0.01; ***p < 0.001; comparing controls and nf1P1 mutants (Šidák). Error bars, SEM.
Nf1 deficiency alters forward walking velocity without any major defects in leg kinematics
Given that Nf1-deficient flies show altered locomotion, we wondered whether their walking gait was impaired (a common feature of genetic disorders affecting motor function). To test this, we used a 3D leg kinematics analysis pipeline (Sapkal et al., 2024) to compare the fine structure of leg kinematics during walking in controls and flies with genomic Nf1 deletion (nf1P1). For this analysis, flies were tethered to a pin and allowed to walk on a spherical treadmill (Fig. 6A). Nf1 mutants exhibited increased walking speed in the tethered preparation compared with K33 (w+) controls (Fig. 6B), similar to untethered flies in the open field (Fig. 3B–D). Pan-neuronal knockdown of Nf1 produced a similar effect (Fig. 6B). Despite the change in walking speed, overall gait was not altered compared with controls. Leg placement was similar in nf1 mutants and control flies (Fig. 6C,D). Stance, step, and swing durations were also similar between genotypes across the instantaneous velocity range (Fig. 6E–G). This suggests that flies lacking Nf1 increased their walking speed in the same way as control flies—by changing the frequency of stepping (decreasing stance duration and increasing stance length; Fig. 6E–G; Bidaye et al., 2020). Furthermore, interleg coordination was normal in mutants—there was no major difference in coupling of leg movements between either legs that move in phase (e.g., left front leg and right middle; Fig. 6K,L) or antiphase (e.g., left and right front legs) (Fig. 6I,J) during tripod gait. Taken together, these results indicate that Nf1 deficiency altered walking speed without affecting gait. Loss of Nf1 is unlikely to affect lower-level motor control areas like proprioceptors and motor neurons, as alterations in those would lead to major anomalies in leg kinematics. Therefore, Nf1 likely impacts locomotion via higher-order walking control centers.
Nf1 deficiency increased walking speed without altering gait or kinematics. A, Diagram of the experimental setup. Fly drawing modified from biorender.com. B, Forward walking speed of controls (K33) versus nf1P1 mutants [***p < 0.001 (Mann–Whitney); n = 70–90] and with pan-neuronal Nf1 knockdown [R57C10-Gal4>UAS-Nf1-RNAi; *p < 0.05; ***p < 0.001 (Kruskal–Wallis/Dunn); n = 80–124]. C, Stance trajectory in control flies. Three hundred individual points plotted (randomly selected from >1,000 steps). The dots indicate touch down locations, which are connected to the liftoff with a line. The black line is the mean of all trajectories. D, Stance trajectory in nf1P1 mutants, plotted as in panel C. E, Stance duration for the L1 leg, comparing K33 controls and nf1P1 mutants. Probability density is graphed as a heat map. F, Step period for the L1 leg. G, Swing duration for the L1 leg. H, Diagram of leg coordination phase comparisons shown in panels I–L. I, L1–R1 leg movement phase plot, comparing K33 controls and nf1P1 mutants. J, L1–L2 leg movement phase plot. K, L1–R2 leg movement phase plot. L, L1–L3 leg movement phase plot.
Discussion
The present data demonstrate that loss of Nf1 alters motor pattern structure and prioritization of motor behaviors in Drosophila. To ensure survival, animals must prioritize one behavior and inhibit another based on internal and external cues (Real, 1991). We found that Nf1 is required to maintain normal activity levels across multiple behaviors over time, as Nf1-deficient flies exhibit altered grooming and walking. The grooming can be overridden by hunger, which shifts the balance toward walking to prioritize foraging. These findings suggest that loss of Nf1 alters neuronal activity, increasing activation of spontaneous grooming and walking speed, as well as changing the prioritization of stimulus-evoked grooming. Importantly, motor coordination is not notably altered, and grooming remains plastic and state dependent.
Grooming is a structured and sequenced set of motor patterns that are modulated by sensory inputs and prioritized according to body part (Seeds et al., 2014). In both flies and mammals, grooming follows an approximately anterior to posterior sequence. In flies, grooming of each body part is controlled by discrete command circuits. Each command circuit comprises multiple elements: sensory neurons that provide the input to command neurons that initiate grooming and send output to descending neurons to execute the motor patterns (Seeds et al., 2014; Hampel et al., 2015, 2017; L. Guo et al., 2022; Zhang and Simpson, 2022). Hierarchical regulation ensures that, when sensory stimuli drive the need to clean multiple body parts, grooming occurs in a prioritized sequence, with certain body parts (such as the eyes and the antennae) taking precedence over others (such as the wings and thorax; Hampel et al., 2017).
Loss of Nf1 increases spontaneous grooming frequency (King et al., 2016), likely through effects on distributed circuits (King et al., 2020). This created two likely a priori scenarios regarding the impact of Nf1 on grooming; loss of Nf1 could (1) increase the grooming across all body parts or (2) increase the grooming of body parts at the pinnacle of the grooming hierarchy, which would then be expected to suppress the grooming of other body parts via inhibitory circuit effects. However, our data revealed a third, unexpected outcome: Nf1 deficiency predominantly increased spontaneous grooming of the abdomen (with knockdown also increasing head and wing grooming). Leg grooming was not consistently elevated. This suggests that the effects of loss of Nf1 could be heterogeneous across different command circuits/neurons, leading to distinct alterations in grooming patterns for different body parts. The pattern of grooming alterations differed between the nf1P1 genomic mutation and nf1E1/pan-neuronal knockdown with RNAi. This may be due to loss of additional e(Spl) genes in the nf1P1 deletion—those additional genes could partially suppress the Nf1 grooming phenotype. A range of additional factors could modulate grooming in an Nf1-dependent manner, particularly in the RNAi condition. Firstly, the expression level of Nf1 could be important, with possible functional differences between reduction (RNAi) and elimination of Nf1. Secondly, variations in the expression level of the R57C10-Gal4 driver (or efficacy of the RNA-induced silencing complex) across different sets of neurons may drive distinct grooming patterns. Thirdly, the expression pattern of Gal4 lines can shift over the course of development (Li et al., 2014), potentially introducing heterogeneity of Nf1 knockdown during critical periods (King et al., 2020). Finally, non-neuronal cells could contribute to the observed phenotype; while R57C10 drives relatively selective Gal4 expression in neurons (Pfeiffer et al., 2012), there are some scattered non-neuronal cells in its expression pattern (Holsopple et al., 2022). In a slightly different arena/environment, head grooming was favored (rather than abdomen) with loss of Nf1 (King et al., 2020), suggesting that environmental context may influence the pattern as well.
Effects of Nf1 loss may be due to excitatory and/or inhibitory circuits. Increased grooming is due to loss of Nf1 in cholinergic circuits (King et al., 2020), which represent the major excitatory circuits in the fly brain. Relatedly, Nf1 deficiency causes alterations in synaptic transmission at the larval neuromuscular junction that originate, at least in part, from central cholinergic circuits (Dyson et al., 2022). In contrast, learning and memory alterations are driven by Nf1 effects in inhibitory GABAergic circuits (Georganta et al., 2021). The number and connectivity of excitatory versus inhibitory neurons in each command circuit may influence how loss of Nf1 affects grooming of each body part. Future studies are needed to address the heterogeneity of cell-autonomous effects of Nf1 deficiency, as well as the contributions of variations in neuronal architecture. Alterations in excitatory:inhibitory (EI) balance are hypothesized to contribute to cognitive and behavioral alterations across a range of genetic disorders (Sohal and Rubenstein, 2019). Given the position of Nf1 as modulator of central neuronal signaling pathways including Ras, cAMP, and G-protein–coupled receptor signaling and its effects on excitatory and inhibitory circuits, it is plausible that EI balance is altered in ways that affect cognition and behavior.
Loss of Nf1 dramatically elevates grooming (72–403%), yet this grooming remains plastic and can be overridden by stimuli that are prioritized over grooming (e.g., foraging). In open-field arenas without food, the flies became hungry and accumulated detectable negative energy balance after 60–150 min (as assayed by homeostatic feeding). Increasing locomotion is a foraging strategy to locate food under starvation conditions (Lee and Park, 2004; Landayan et al., 2018; Yurgel et al., 2019). By 150 min in the open field without food, grooming frequency dropped and locomotion increased, reflecting a state-dependent switch from a grooming- to foraging-dominant behavioral mode. Thus, grooming circuits are engaged by Nf1 deficiency until hunger increases and overrides the grooming drive. The increase in hunger preceded the increase in locomotor activity, suggesting that flies accumulate negative energy balance for a period before foraging is engaged. This is likely a strategy to maximize energy utilization, as increasing locomotor activity under energy-restricted conditions accelerates starvation; it represents a high-risk, last-resort strategy to find food. Overall, this suggests that loss of Nf1 alters network dynamics, reshaping the activation and perseveration of specific motor behaviors in a state-dependent way but sparing some behavioral plasticity.
Nf1 mutants exhibited increased walking speed on a spherical treadmill, yet they exhibited normal leg kinematics and gait patterns. This suggests that Nf1 affects the macroproperties of locomotion like walking speed without causing defects in microproperties like interleg coordination or stepping direction. Other mutations that impact motor function, such as Parkinson's disease (PD; α-synuclein expression, parkin mutants) and spinocerebellar ataxia (SCA) models (mutant SCA3 expression; Stahl and Tomchik, 2024), affect locomotion differently. For instance, PD and SCA disease models exhibit markedly altered gait, generating uncoordinated movements and erratic foot placement (Wu et al., 2019). NF1 patients exhibit difficulties with complex motor coordination, including deficits in fine motor control (Krab et al., 2011). Similar motor dysfunction is observed in animal models such mice and pigs (Rahn et al., 2021; Swier et al., 2024). In flies, our data suggest that alterations in motor behavior result from dysfunction in high-level motor areas. Nf1 is unlikely to affect low-level motor control neurons (central pattern generators, proprioceptors, and descending neurons). Rather it may act on walking control neurons (Bidaye et al., 2014, 2020; Sapkal et al., 2024), directly and/or indirectly (e.g., via metabolic changes; Botero et al., 2021).
Nf1 exerts effects on neuronal function during a critical period of development (King et al., 2020). Specifically, there is a critical period for Nf1 effects on grooming in Drosophila spanning the late third instar larval phase and first half of the pupal phase. This developmental period encompasses a series of neurodevelopmental steps, including cell proliferation, migration, differentiation, dendritic remodeling, axon guidance, synaptogenesis, and survival (Truman and Bate, 1988; Tissot and Stocker, 2000; Urbach and Technau, 2004; Doll and Broadie, 2014; Miyares and Lee, 2019). Therefore, Nf1 effects on the neuronal circuitry regulating grooming could result from alterations in one or more of these processes. Ultimately, the behavioral effects likely emanate from altered adult neuronal composition, connectivity, excitability, or synaptic transmission. Some Nf1 mutations alter neuronal excitability in rodents (Anastasaki et al., 2022). Although this is associated with tumor formation, it suggests that changes in neuronal excitability may be a conserved mechanism. At the molecular level, the behavioral effect of Nf1 requires an intact Ras GTPase-activating domain, suggesting that altered Ras signaling could underlie (or contribute to) the phenotype. Ras influences multiple developmental processes, including cell growth, migration, cytoskeletal integrity, survival, and differentiation (Rajalingam et al., 2007). Though many of the described effects of Nf1 deficiency involve Ras signaling (Costa et al., 2002; Walker et al., 2006; Xie et al., 2016; Cattaneo et al., 2020; Botero et al., 2021; Georganta et al., 2021), loss of Nf1 also alters cAMP signaling, which contributes to some neuronal phenotypes (Guo et al., 2000; Buchanan and Davis, 2010; J. A. Brown et al., 2012; Walker et al., 2013; Machado Almeida et al., 2021; Anastasaki et al., 2022). Therefore, this pathway could contribute to grooming phenotypes, potentially downstream of Ras (Anastasaki and Gutmann, 2014). Future studies will be necessary to delineate the specific contributions of each pathway, as well as dissect the key downstream signaling molecules and interactions.
Overall, the present data reveal that loss of Nf1 affects the neuronal network activity regulating temporally sequenced behaviors in a circuit-specific and state-dependent manner. Changes in grooming were nonuniform across body parts, suggesting heterogeneity of effects on the command circuits that drive grooming of each body part. Loss of Nf1 exerted strong behavioral effects, but those could be overridden by stimuli that are prioritized for survival. In addition, the activation of locomotor behavior was altered, with changes in walking speed (increased forward walking speed on a spherical treadmill), but no detectable change in kinematics or gait. Thus, the loss of Nf1 alters the frequency, patterning, prioritization, and execution of sequenced motor behaviors. These data lay the groundwork for future studies to address the molecular mechanisms of Nf1 effects on neurons, circuit alterations, and the prioritization and sequencing of behaviors following mutation of genes that drive neurodevelopmental disorders.
Footnotes
We thank Timothy Stelmat for designing and constructing lighted enclosures for the open-field experiments and Luis Vivas for the assistance with Python and MATLAB code. Flies obtained from the Bloomington Drosophila Stock Center (NIH P40OD018537) and the Vienna Drosophila Resource Center (VDRC, www.vdrc.at) were used in this study. Imaging was carried out at the University of Iowa Central Microscopy Research Facility (CMRF). Acquisition of the CMRF Leica SP8 Laser Scanning Confocal microscope with STED capability was made possible by a generous grant from the Roy J. Carver Charitable Trust; additional CMRF funding was provided by the University of Iowa Office of the Vice President for Research, the Carver College of Medicine, and the College of Liberal Arts and Sciences. This research was supported by NIH/NINDS R01 NS097237, R01 NS126361, R01 NS114403, R21 NS124198, and DOD CDMRP NF230039. We thank Lisa Ringen, Linda Buckner, Rob Svetly, Kathleen O’Brien, and Melissa Benilous for their administrative support.
The authors declare no competing financial interests.
- Correspondence should be addressed to Seth M. Tomchik at seth-tomchik{at}uiowa.edu.
