henhaobio
Sign in

bioRxivNeurosciencePreprint

Organization, development, and plasticity of the serotonergic dorsal raphe nucleus projectome

Lucía Jiménez-Fernández, Zoltán Kristóf VargaiD, Klara Thurner, Florence KermeniD

PDFDOI

Abstract

The dorsal raphe nucleus (DRN) is the main source of serotonin (5-HT) in the central nervous system. 5-HT DRN neurons modulate a wide range of physiological functions and brain states via their widespread and heterogeneous projections. Yet the organizational logic, developmental dynamics, and experience-dependent plasticity of these projections remain poorly understood. By reconstructing the whole brain projections of 81 individual 5-HT DRN neurons across developmental stages in the larval zebrafish, we find that these neurons exhibit diverse projection patterns, ranging from broadly distributed to spatially confined. We identify two previously uncharacterized, projection-defined subpopulations within the larval dorsal raphe nucleus, which selectively target either the telencephalon or the spinal cord. This organization is consistent with the existence of anatomically distinct serotonergic pathways that may differentially influence forebrain and sensorimotor circuits. Longitudinal analysis revealed that the gross architecture of 5-HT neurons remained stable during larval development, although region-specific remodeling occurred in all neurons. Importantly, stress exposure during the first week of development, but not thereafter, altered the density of 5-HT DRN projections to telencephalic regions, suggesting a critical window of increased plasticity during early life. Collectively, our findings clarify the developmental trajectory, maturation, and stress-induced plasticity of the serotonergic circuitry, providing important insights for deciphering the pathophysiology of stress-related mental disorders.

In-depth summary

Claude Opus 5.5 · xhigh effort · 7 Nov 2026 · from the full text of version 2 (11 Sept 2026) · CC BY 4.0

In larval zebrafish, 68% of 81 traced serotonergic dorsal raphe neurons target only one of telencephalon and spinal cord; stress in week 1, not week 2, raised telencephalic fiber density.

How single serotonergic dorsal raphe nucleus (DRN) neurons build and refine their brain-wide projections in early life, and whether early-life stress reshapes them only within certain windows, was unresolved. Earlier single-neuron work in mice sampled under 0.6% of DRN serotonergic cells; the authors present theirs as the first whole-brain reconstruction of single neuromodulatory neurons compared across a one-week developmental interval.

Key results

  1. ShownDRN serotonergic neurons split by target into two largely non-overlapping subsets, one reaching the telencephalon and one the spinal cord, alongside wide variation from confined to brain-wide reach.

    Of 81 neurons, 11% reach both telencephalon and spinal cord and 68% only one (46% spinal cord without telencephalon, 22% the reverse). Hierarchical clustering with principal component analysis of density in five CNS subdivisions gave three clusters; cluster 1 was spinal cord-high and telencephalon-low, cluster 2 the reverse. Over 60% of neurons reach three or more subdivisions, 20% one or two. · Organizational logic of 5-HT DRN neurons’…; Figure legends

  2. SuggestedStress in the first week of life raised serotonergic fiber density in telencephalic regions, whereas chronic unpredictable stress (CUS) in the second week left projection density unchanged even though it still increased anxiety-like behavior.

    Cortisol at 2-6 days post-fertilization (dpf): area × treatment p<0.0001 (n=12, 10, 10 fish); density rose in olfactory bulbs at 5 µM and in olfactory bulbs, pallium, subpallium and dorsomedial telencephalon (Dm) at 10 µM; DRN area unchanged (p=0.422). Week-1 CUS: subpallium only (one-sided p=0.027; n=18 vs 19). Week-2 CUS: area × stress p=0.597 (n=13 vs 14), though anxiety score rose (p=0.039). · Early life stress reshapes the developing 5-HT…; A critical window of sensitivity to early-life…; Statistical analysis; Figure legends

  3. ShownBetween 1 and 2 weeks of age, traced neurons kept their broad reach, but every one was locally remodeled: growth exceeded loss, projections became more bilateral, and the telencephalon, diencephalon and mesencephalon changed most.

    Same 19 neurons imaged at both ages: projection score unchanged (paired t-test, n=18, p=0.331); laterality score decreased (n=19, p=0.021); new neurites 27% vs lost 18% (p=0.001). Neurite change differed across subdivisions (mixed-effects model, p=0.0002); loss was uniform (p=0.503) while growth varied (p<0.0001). · Stability and regional remodeling of the…; Figure legends

  4. ShownTarget regions differ both in how many neurons reach them and in how densely they are innervated; a zero-inflated model sorted 11 CNS regions into three innervation classes.

    Cerebellum and hypothalamus received fibers from over 70% of the 81 neurons, olfactory bulbs and optic tectum from under 10%. A zero-inflated negative binomial model, chosen by Bayesian information criterion, grouped telencephalic areas and preoptic area (dense, few neurons); pons-medulla, cerebellum, spinal cord and hypothalamus (dense, many); thalamus and optic tectum (sparse, few). · 5-HT DRN neurons provide highly heterogeneous…; Statistical analysis; Figure legends

  5. ShownSoma position within the DRN partly predicts targets: anterior neurons send more neurites ventrally and to the hypothalamus, dorsal neurons send more to the cerebellum, and neurons co-expressing Gad1b sit mostly in the anterior DRN.

    n=81: ventrality index vs anterior-posterior soma position rho=-0.558, p<0.0001; no trend with dorso-ventral position (rho=0.041, p=0.712). Hypothalamic density vs anterior-posterior position rho=-0.427, p=0.002; cerebellar density vs dorso-ventral position rho=0.390, p=0.005. Tph2+Gad1b+ neurons were 36% (30 vs 51 Gad1b-negative), skewed anterior (rho=-0.302, p=0.006). · Organizational logic of 5-HT DRN neurons’…; Figure legends

Methods, data and reproducibility

Single DRN serotonergic neurons, sparsely labeled in Tg(tph2:Gal4;UAS:Caax-GFP; gad1b:DsRed) larvae, were confocal-imaged, traced in SNT and registered with CMTK to Z-brain or new 2-week templates; some were re-imaged a week later. Broadly labeled larvae given cortisol or chronic unpredictable stress in week 1 or 2 were tested in a swimming plus maze, and Tph2 fiber density was mapped.

  • Sparsely labeled single DRN serotonergic neurons traced at 6-7 dpf (1 week old) in Tg(tph2:Gal4;UAS:Caax-GFP; gad1b:DsRed) larvae · zebrafish (nacre mifta-/- background) · n = 81 neurons (only 2 from the same animal)
  • Same single neurons re-imaged and traced at 1 and 2 weeks old · zebrafish · n = 19 neurons at both ages; 21 neurons at 2 weeks
  • Broadly labeled DRN projection maps used as whole-population reference · zebrafish · n = 12 fish (1 week), 13 fish (2 weeks)
  • Cortisol exposure 2-6 dpf (0, 5, 10 µM), swimming plus maze at 7 dpf and fiber-density mapping · zebrafish · n = 12, 10, 10 fish imaged; n=29 in behavioral analysis
  • Chronic unpredictable stress 2-6 dpf, behavior at 7 dpf, fiber density and DRN soma counts · zebrafish · n = 30 vs 30 behavior; 18 vs 19 imaged; 17 vs 19 soma counts
  • Chronic unpredictable stress 6-13 dpf, behavior at 14 dpf and fiber-density mapping · zebrafish · n = behavior n(control,CUS)=24,25; 13 vs 14 imaged
  • Z-brain atlas regions and reference/template brains used for 1-week registration · zebrafish · n = 294 regions; 11 CNS regions analyzed
  • 2-week template brains built in this study from Tg(elavl3:H2b:GCaMP6s) scans, plus line-specific templates · zebrafish · n = four best-registered scans averaged
Methods & tools
Simple Neurite Tracer (SNT), FIJI ImageJ, Computational Morphometry Toolkit (CMTK), Z-brain atlas, MATLAB (custom scripts), Ethovision XT 17, R, Graphpad Prism 10, Zeiss LSM 700 confocal microscope, swimming plus maze, zero-inflated negative binomial model, Basler ace2 camera
Reagents & environment
Ethovision XT 17 and Graphpad Prism 10 versions stated; R version not given; CMTK registration command-line parameters printed in Methods.
Reproducibility gaps
No code or data repository is given: the custom MATLAB thresholding and density scripts, traced neurons and the new 2-week template brains are not linked. Numerical data behind the figures are said to be in a supplementary file 'Numerical data.xls'.
Limitations (6)
  • Saturating the DRN to reveal faint ventral fibers prevented counting somata in the cortisol and week-2 CUS experiments, so Tph2+ DRN area served as a proxy for cell number. · stated by the authors
  • Pons-medulla and spinal cord were left out of the stress analyses because they contain tph2+ cell bodies, so stress effects on those projections were not tested. · stated by the authors
  • Fish that still registered poorly after a second, manually rotated attempt were excluded; the spinal cord mask was shortened to match scan coverage, and one longitudinal fish lacked spinal cord in its 2-week scan. · stated by the authors
  • The authors note that rostral raphe projections to the spinal cord have been proposed not to persist into adulthood in zebrafish, so the larval organization mapped here may not hold in adults. · stated by the authors
  • The critical-window claim rests on a single null week-2 experiment (CUS only, n=13 vs 14 fish); cortisol was not tested in week 2, so timing and stressor type are not fully crossed. · noted in this summary
  • The week-1 CUS subpallium effect comes from one-sided t-tests across four telencephalic regions (p=0.027), with no multiple-comparison correction described for that test set. · noted in this summary

Read this if you study development or plasticity of neuromodulatory circuits, map single-neuron projections in larval zebrafish, or want anatomical readouts of early-life stress on serotonergic wiring.

Citations
1

Reproducibility ledger

Nobody has reported trying to reproduce this paper yet. Did you replicate an experiment or run the code? Your report saves the next lab a week.

Sign in to file a reproducibility report.

Similarity Map
SelectedSame fieldCross-field
THISVarga '26Kisner '24Orav '24Ren '18Hamada '24Chang '21
80% semantic matchNeuroscience
Zoltán Kristóf Varga, Archana Golla, Florence Kermen·Communications Biology·Mar 20, 2026· 3 citations