Ciliary cAMP regulates Shh signal interpretation to drive polarisation of differentiating neurons

This article has been Reviewed by the following groups

Read the full article See related articles

Listed in

Log in to save this article

Abstract

Cellular differentiation is characterised by transitions in cell-state through reinterpretation of extracellular signals. However, the mechanisms facilitating context-dependent signal interpretation remain poorly understood. Differentiating neurons remodel a molecularly distinct primary cilium as they switch from canonical to non-canonical Shh signalling. Here, using long-term live-tissue imaging, we demonstrate that the opposing Shh signalling modulators Smo and GPR161 simultaneously accumulate in the remodelled primary cilium. The correct balance of Smo and GPR161 leads to elevated ciliary cAMP levels, which suppresses Gli transcription factor activation and regulates actin dynamics to drive neuron polarisation. Notably, disrupting this balance through Smo hyperactivation or GPR161 depletion results in reduced ciliary cAMP, inappropriate activation of canonical Shh signalling, dysregulated actin dynamics and initiation of multiple unstable axon-like projections. Thus, this study identifies shifts in ciliary cAMP levels as a key regulator of cellular signal interpretation, and links primary cilium-mediated signal transduction to precise control of cytoskeletal organisation.

Article activity feed

  1. Note: This response was posted by the corresponding author to Review Commons. The content has not been altered except for formatting.

    Learn more at Review Commons


    Reply to the reviewers

    __Manuscript number: __RC-2026-03469

    Corresponding author(s): Gabriela Toro-Tapia and Raman M. Das

    [Please use this template only if the submitted manuscript should be considered by the affiliate journal as a full revision in response to the points raised by the reviewers.

    If you wish to submit a preliminary revision with a revision plan, please use our "Revision Plan" template. It is important to use the appropriate template to clearly inform the editors of your intentions.]

    1. General Statements [optional]

    __We thank the reviewers for their thorough and constructive evaluation of our manuscript. __We are encouraged by the reviewers’ positive assessment of the study, including recognition of its careful experimental design, use of complementary functional approaches and signalling reporters, and live imaging of neuronal differentiation within a near-native tissue environment. The reviewers also highlighted the mechanistic insight provided by the work, the strength of our ex vivo slice-culture approach, and its broad relevance to cell biology, neurodevelopmental biology and the study of ciliopathies and neurodevelopmental disorders.____

    ____In response to the reviewers’ thoughtful suggestions, we have substantially strengthened the manuscript through additional experiments, quantitative analyses and statistical evaluation. Most importantly, we have directly manipulated ciliary cAMP using a cilium-targeted cAMP sponge, providing new functional evidence that reducing available ciliary cAMP disrupts neuronal polarisation and establishment of a single axon. This provides direct evidence for a link between ciliary cAMP signalling and neuronal polarisation. We have also carried out quantitative analysis of neuronal protrusive behaviour and Gli activity, added statistical analyses of ciliary Smo/GPR161 and ADCY3 localisation, and clarified the interpretation and limitations of our cADDis measurements.____

    ____We have further revised the discussion and model to more clearly distinguish our experimental findings from proposed mechanisms, particularly regarding the relationship between ciliary cAMP, PKA signalling and actin dynamics, and the potential contributions of ciliary and extraciliary Smo to non-canonical Shh signalling in differentiating neurons. We have also incorporated the reviewers’ suggestions regarding additional ciliary GPCRs, Gli3R and the broader relevance of GPR161 to neuronal development. Finally, the figures have been revised throughout to improve clarity and consistency.____

    ____We believe that these revisions directly address the reviewers’ principal concerns while building on the strengths they identified in the original study. ____

    Below, we address each comment individually and describe the corresponding modifications, with changes identified by line and figure number in the revised manuscript.

    2. Point-by-point description of the revisions

    This section is mandatory. *Please insert a point-by-point reply describing the revisions that were already carried out and included in the transferred manuscript. *

    * *

    Reviewer 1:

    This study by Gabriela Toro-Tapia et al explores the role of the opposing Shh signalling modulators Smoothened and GPR161 during the remodelling of the Shh downstream signalling cascade that the same authors have proposed to occur upon newborn neurons delamination from the apical surface, a phenomenon that entails ciliary "shedding" and regrowth. They explore the possibility that the conversion from a canonical Shh pathway based on Gli TF activation to a non-canonical pathway relies on a shift from a mutually exclusive ciliary localization of the two transmembrane factors Smo and GPR161 to a (frequent) ciliary colocalization that would result in simultaneous ciliary Smo and elevated ciliary cAMP levels, underlying the non-conventional pathway. To explore this hypothesis, the study first characterizes the co-detection of the two receptors during the neuronal delamination and morphogenesis sequence, and then presents a series of functional perturbations centered on the modulation of cAMP levels in the cilium. The authors use the chick embryonic spinal cord as a model of neuronal delamination and axonogenesis; they use in ovo electroporation and live imaging in ex ovo slice culture to characterize cellular changes and differenciation of newborn neurons with a single cell subcellular and cellular resolution in an "intact" 3D cellular environment; except for the dynamic characteriztion of co-detection of GPR161 and Smo, which is performed in mouse embryo sections at a similar development stage for technical reasons, all the study is performed in the chick embryo

    The authors first illustrate an increased co-detection of Smo and GPR161 in delaminated neurons as the apical endfoot retracts, and report on the presence of the adenylyl cyclase ADCY3 in the regrowing/remaining cilium portion of half of these cells. Using a ciliary cAMP sensor cADDI, they observe a transient increase in reporter activity (indicating transient decrease in ciliary AMPc) in the reassembling cilium, followed by an increase in cAMP levels at the end of retraction.

    The authors then use shRNA-mediated downregulation of GPR161, and observe defects in neuronal repolarization; these cells also display reduced ciliary cAMP levels monitored with the cADDIs sensor; similar phenotypes are obtained when increasing Smo accumulation in the cilium by SAG treatment, indicating that the balance between ciliary Smo and GPR161 regulates ciliary levels of AMPc, and that this level may in turn control the cell remodeling associated with neuronal morphogenesis

    Using a photoconvertible reporter of Gli activation (GBS-nls-kikGR) that reports canonical Shh signaling in GPR-161 knock-down cells, the authors then show that GRP161 is required in delaminating cells to switch off canonical signaling; using the reporter in SAG-treated cells again shows that the balance between ciliary Smo and GPR161 is essential.

    The authors then describe the dynamics of actin cytoskeleton reorganization during neuronal polarization, and show accumulations of F-actin at the site of axon emergence and in dynamic protrusions associated with axon formation. They then use Chromophore-Assisted Light Inactivation to destroy the cilium in young neurons undergoing axonal extension, and show that this eliminates actin accumulations and stops and reverts axonal growth, showing a requirement for the cilium in the process; blocking Shh signalling with the Smo antagonist cyclopamine leads to the same failure in polarization and/or reversion of axonal growth. Finally, they show that GPR161 knock-down or SAG treatment lead to similar defects in F-actin accumulation and protrusion formation, and conclude that the balance between Smo and GPR161 in the reformed cilium after delamination is crucial to control cAMP levels in response to Shh signalling, and that this leads to an (as yet not elucidated) non canonical, Gli-activation independant transduction that orchestrates the cytoskeletal remodeling involved in neuronal morphological remodeling and axonal extension

    Major comment:

    The central point of the paper is that cAMP levels in the cilium are key in the transition in Shh signal transduction. However, functionally, cAMP levels are always adressed indirectly; the authors provide several converging lines of evidence, with functional challenges that result in a change in ciliary cAMP level and a simultaneous range of defects (Shh reporter read out, morphological changes). While overall the convergence of multiple approaches supports their claim, one might object that this does not directly prove that ciliary cAMP levels control the downstream defects. For example, SAG treatment may also activate non-ciliary Smo functions; similarly, we cannot formally exclude the (unlikely) hypothesis that GPR161 knock-down may also affect the process in a cAMP-independant manner.

    To provide direct evidence that ciliary cAMP levels are central in the phenomenon that they describe, the authors could express a ciliary cAMP sponge, such as the one used in the 2023 study by Atkins et al (ref 29 in the manuscript); the expectation is that this should reproduce all the phenotypes (e.g. Gli-reporter activity in differentiating neurons, defects in polarization) observed with their different approaches (SAG treatment, GPR161 knock-down). The most direct phenotype to describe, should the authors perform these experiements, would probably be the morphological defects in neurons.

    (Alternatively, using the photoactivable ciliary Adenylyl cyclase used in the same study by Atkins et al to restore high cAMP in the cilium may rescue the phenotypes caused by SAG treatment or GPR161 KD).

    Using a direct modulation of ciliary cAMP would, in my view, reinforce the conclusions of the study.

    We thank the reviewer for this important comment. We have performed an additional experiment to directly test whether reducing available ciliary cAMP results in similar morphological defects during neuronal polarisation.

    As helpfully suggested by the reviewer, we expressed a validated cilium-targeted cAMP chelator (cilium-targeted cAMP sponge; pCX-5HT6-cAMPSponge-mRFP) (Atkins et al., 2023) in chick neuroepithelial cells and confirmed correct localisation to the primary cilium. This produced a phenotype closely resembling that observed following GPR161 knockdown, with differentiating neurons exhibiting unstable axon-like protrusions and defects in the establishment of a single axon. These findings provide direct evidence that perturbation of the ciliary cAMP pool is sufficient to disrupt neuronal polarisation and therefore support our conclusion that elevated ciliary cAMP levels are required for this process.

    Notably, expression of the cilium-targeted cAMP sponge also caused some neurons to form branched protrusions. We speculate that this reflects the broader disruption of ciliary cAMP signalling caused by the cAMP sponge, implying the involvement of multiple factors that regulate cAMP turnover, such as additional GPCRs, rather than GPR161 alone, as noted in the discussion (line 534).

    These new results are described in __lines 240-251 __and presented in Fig. 4, Fig. S3, and Movies S8, 9 and 10):

    ‘To directly assess the importance of ciliary cAMP levels for neuron polarisation and axon initiation, we reduced available ciliary cAMP by expressing a cilium-targeted cAMP chelator fused to mRFP (cilium-targeted cAMPsponge-mRFP) (Atkins et al, 2023), followed by timelapse imaging. The majority of cilium-targeted mRFP expressing control cells initiated a single axon and underwent normal axon extension towards the ventral midline (Figure 4A, Movies S8; 89% 32/36 cells, 10 embryos). In contrast, cells expressing cilium-targeted cAMPsponge-mRFP, similar to GPR161 knockdown cells, often failed to initiate a normal axon and instead exhibited abnormal protrusive activity, generating multiple unstable axon-like projections (Figure 4B and 4C, Movie S9; 63% 17/27, 10 embryos), some of which underwent branching (Figure S3A, Movie S10). Therefore, local buffering of ciliary cAMP disrupts neuronal polarisation and normal axon initiation, resulting in multiple unstable, branched protrusions and impaired axon extension. Taken together, these findings indicate that the appropriate balance of ciliary Smo and GPR161 drives an increase in ciliary cAMP that is required for establishing a single axon.’

    The reviewer also raises the important question of whether reduced ciliary cAMP activates downstream Gli-dependent transcription. Due to spectral overlap, it is challenging to combine the cAMP-sponge construct with the GBS-NLS-KikGR reporter assay in vivo. Consequently, our new experiments do not directly test the relationship between ciliary cAMP depletion and Gli activity. However, this link has been established previously (Moore et al., 2016). We have clarified this point in the revised manuscript (lines 216-220):

    ‘Smo and GPR161 are suggested to have opposite effects on cAMP production. While Smo ciliary accumulation has been proposed to decrease ciliary cAMP, allowing activation of Gli transcription factors and canonical Shh signalling (Moore et al., 2016), ciliary accumulation of GPR161 is proposed to elevate ciliary cAMP, which negatively regulates canonical Shh signalling by promoting processing of Gli transcription factors to their repressor forms (Hwang et al, 2021; Mukhopadhyay et al., 2013).’

    Finally, we have also considered performing a rescue experiment by restoring ciliary cAMP levels in GPR161-depleted cells. In our experimental model, this would require co-transfection of an additional construct alongside the five plasmids already required, substantially reducing co-transfection efficiency and making interpretation of partial rescue difficult. Moreover, preliminary collaborative experiments using the cilium-targeted photoactivated adenylyl cyclase Arl13b-bPAC indicate that increasing ciliary cAMP is itself disruptive to neuronal differentiation (Frechen F. and Wachten D., unpublished observations). Thus, both decreases and increases in ciliary cAMP appear detrimental, making a straightforward rescue strategy difficult to interpret. We have therefore focused on the cilium-targeted cAMP sponge experiment, which provides a more direct test of the reviewer's central concern and strengthens the conclusion that appropriate levels of ciliary cAMP are required for neuronal polarisation.

    Minor comments

    Fig1B: do the authors have values for colocalization of GPR161 and Smo in attached cilia? The separate values (shown in Toro-Tapia 2022, Figure 4) measured in the chick embryo suggest that co-detection should be close to zero, single GPR161 should be low, and single Smo should be high, but not every reader will have the data from the previous study in mind, so it would be good to remind them. Unless there is a technical impossibity to obtain these values in Tuj1+ attached cells, it would be even better to have the double values in mouse to complement the graph in Figure 1B and illustrate the increase in double labelling from attached -> detached proximal -> detached distal.

    Similarly, in figure 1C, how does ADCY3 look like in Tuj1+ attached cells?

    In this same figure, some statistical validation of the significance of the increase in double positive cells would also be appreciated.

    Thank you for these suggestions. We have now included data for GPR161 and Smo localisation in attached primary cilia and performed a pooled contingency analysis that showed an association between differentiation stages and Smo/GPR161 ciliary accumulation state (chi-square test of independence, p = 0.0025), supporting our conclusion that the distribution of Smo-only, GPR161-only, Smo/GPR161-positive (both), and Smo/GPR161-negative (none) cilia changes during neuronal differentiation. This is presented in line 106-108____; Figure 1B’:


    *‘Contingency analysis of observations across different stages of neuronal differentiation demonstrated an association between differentiation stage and Smo and GPR161 ciliary accumulation state (pooled data from 4 embryos, Figure 1B’)’ *

    We have now also included attached cells in the ADCY3 analysis (line 116____; Figure 1C and Figure S1). The proportion of ADCY3 accumulated in the cilia remained near 50%, with no significant differences between differentiation stages, supporting our conclusion that ADCY3 ciliary localisation remains relatively stable during neuronal differentiation. This is presented in lines 119-122 and Figures 1C and S1:

    ‘This revealed ciliary accumulation of ADCY3 in approximately 50% of remodelled primary cilia at all stages of primary cilium reassembly (Figure 1C’, Figure S1A; attached: 45% n=28/62, proximal: 46%, n= 46/100 cells; distal: 55%, n= 81/146 cells; axonogenesis: 51% n=63/124 cells, 4 embryos).’

    -Figure 1E: could the authors comment on the "transient" nature and overall timing of the cADDIs signal increase during retraction? I find it counter intuitive that the signal is low at the beginning of the time series in Figure 1E (I expected a low cAMP concentration and high signal right from the start of the sequence); besides, the drop in signal intensity also appears quite late in the sequence, when the apical extension is fully resorbed, compared to the timing of the appearance in double GPR161/Smo-positive ciliary remnants in Figure 1B. This does not really fit with the observation that the number of single or double GPR161 cells is lower in cells undergoing axonogenesis that in delaminating cells with a distal ciliary remnant. Is there an explanation for the delay between the increase in the number of double positive cells, and the actual shift in cAMP levels? Could it be linked to the changes in size of the ciliary remnant/cilium itself, which may need to regrow to a threshold size for signal reception and transduction to become effective?

    We thank the reviewer for this insightful observation. We agree that the transient nature and timing of the cADDis signal during apical process retraction are not immediately intuitive and warrant further discussion.

    First, we caution against interpreting the low cADDis signal at the beginning of apical process retraction as a reflection of elevated cAMP levels. As our previous work (Toro-Tapia and Das, 2020) demonstrated, the early stages of apical process retraction are characterised by remodelling of the primary cilium. While a short Arl13b-positive particle is retained, intraflagellar transport is not visibly active during early retraction, suggesting that the retained Arl13b-positive particle is not fully functional. Under these conditions, we cannot exclude the possibility that low cADDis fluorescence reflects the immature state of the retained particle, including incomplete accumulation of the reporter within the reforming ciliary compartment, rather than cAMP levels alone. Importantly, at later stages of apical process retraction, when ciliary remodelling is complete and intraflagellar trafficking has resumed, we consistently observe strong ciliary cADDis fluorescence. As cADDis fluorescence increases when cAMP levels decrease, these observations indicate that recently remodelled cilia of newborn neurons contain relatively low ciliary cAMP levels. We therefore have greater confidence in the interpretation of cADDis signals once ciliary remodelling has been completed.

    The apparent delay between the increase in GPR161/Smo double-positive cilia and the subsequent cADDis response may reflect both biological and technical factors. Ciliary accumulation of GPR161 and Smo is likely to be highly dynamic, and fluctuations in their relative abundance during apical process retraction may delay or gradually shape the resulting change in ciliary cAMP. Indeed, such accumulation may occur before the remodelled cilium has regained full signalling competence. In this respect, we agree with the reviewer's suggestion that the ciliary remnant may need to reach a critical size and/or degree of structural maturity before efficient signal transduction can occur. Thus, accumulation of Smo and GPR161 may precede detectable changes in ciliary cAMP. Technical factors may also contribute, as the cADDis reporter likely requires a threshold change in cAMP concentration before a robust fluorescence response can be detected. We favour the interpretation that receptor redistribution and cAMP signalling are temporally linked but not necessarily simultaneous events. However, the current experiments do not allow us to distinguish conclusively between delayed acquisition of ciliary signalling competence, gradual changes in receptor activity, and limitations in reporter sensitivity.

    We have clarified these points in the revised manuscript (line 142-146):

    ‘At the later stages of apical process retraction, when primary cilium reassembly is complete and intraflagellar trafficking restored (Toro-Tapia and Das, 2020), we observed high levels of green cADDis fluorescence, indicating low ciliary cAMP levels. As the remodelled primary cilium entered the neuronal cell body, and preceding axon initiation, cADDis fluorescence decreased sharply (10/10 cells from 10 embryos), indicating a rapid increase in ciliary cAMP. Notably, this increase follows the ciliary accumulation of GPR161 in neurons whose apical processes had retracted distal to the apical surface (Figure 1B), consistent with a potential link between GPR161 accumulation and increased ciliary cAMP.’

    In figure S4, the images only show heatmap of the F-tractin signal; the "GFP-GPI" and "Arl13b supernova" mentions should be removed from the figure; the color code for F-trctin-GFP may be changed from green to "heatmap-like" color code

    GFP-GPI and Arl13 Supernova have been removed, and we have changed the labelling for F-tractin to a heat-map like colour.

    Similarly, in Fig S6, "luciferase RNAi" and "GPR161 RNAi" are the treatment applied to the cells, and should be indicated in black in a lateral, vertical box (like DMSO or SAG treatments)

    We have modified the text as suggested.

    In figure 6, the color code is confusing; why are both Smo, GliA and Actin dynamics in green, and GPR161 and the the cilium in magenta? Maybe using more colors would help

    We have modified the colours in Figure 6 (now Figure 8) accordingly.

    • In Figure 6: the construction of the visual summary also suggests that the inhibition of GliA regulates actin dynamics, but I do not think that the authors provide functional evidence for this causal link. The data (especially the Gli reporter data) is overall compatible with this notion, and it is certainly the most likely interpretation, but it does not prove it directly, and I would suggest a dashed arrow and a question mark here.

    We have revised Figure 6 (Now Figure 8) based on these comments

    -In Figure 2D", one of the 25 dots in the SAG condition is missing in the AUC graph.

    We have added the additional data point to the AUC graph in Figure 2D’’ (Now Figure 3C’’).

    Reviewer #1 (Significance (Required)):

    Significance:

    This is a very carefully constructed study. The authors use complex manipulation of gain/loss of function with reporters and sensors in the context of live imaging to nicely approach the dynamics of changes in Shh signal reception and transduction during the process of neuronal delamination and re-polarization. The use of subcellular reporters of signalling dynamics and morphological changes is particularly impressive. I command them for this brave and beautiful piece of work.

    In a previous study (Toro-Tapia et al, 2022), the same authors had shown that Smo and GPR161 were both observed in the remodelled cilium of the newborn neuron, and although they had not visualized them simultaneously, they postulated the the presence of GPR161 in the cilium was responsible from a shift from the canonical Shh->Smo->GliActivation pathway to a non-canonical pathway downstream of Smo. Here, they explore the hypothesis that this transition is achieved via the GPR161-dependent maintenance of elevated ciliary cAMP concentration.

    One strength of the study is that they use a combination of complementary functional approaches and sensors that all converge towards the validation of this hypothesis. Its main limitation, in my view, is that it lacks an experiment that directly assesses the specific role of ciliary cAMP concentration.

    Another strength is that it addresses this question in live conditions in a "near" native tissue environment via the authors elegant ex vivo slice culture assay. This is a rare feat, which sets this work apart from many studies on axonal polarization and growth that are performed in dissociated cells.

    One open question for future investigations is why the remodeled cilia after "cilium shedding" tolerates the simultaneous entry of Smo and GPR161, while it does not happen in the cell-cycle regulated ciliary resorptions and regrowth sequence.

    From a clinical perspective, the work opens leads to future studies and interpretation of phenotypes associated with ciliopathies, including Joubert and Meckel syndromes, as pointed out by the authors in their discussion.

    This work should be of interest to cell biologists and developmental biologists alike, and is also relevant to clinicians interested in neurodevelopmental disorders and more broadly in ciliopathies.

    We thank the reviewer for their positive assessment of the significance of our study. We particularly appreciate their recognition of the value of examining dynamic ciliary signalling and neuronal polarisation within a near-native tissue environment, and of the complementary approaches used to interrogate this process. We are also encouraged by the reviewer’s view that the findings have broader relevance to cell and developmental biology, as well as to understanding neurodevelopmental disorders and ciliopathies.

    We also appreciate the reviewer highlighting direct manipulation of ciliary cAMP as the principal experiment needed to strengthen the study. As detailed in our response to the major comment above, we have now performed this experiment and incorporated the resulting data into the revised manuscript. We believe this directly addresses the principal limitation identified in the reviewer’s assessment.

    Finally, we agree that understanding how the remodelled cilium of a differentiating neuron permits simultaneous accumulation of Smo and GPR161 is an intriguing question that emerges from this work and as acknowledged in the discussion (lines 524-526), will be important to address in future studies.

    Reviewer 2:

    * *

    Summary: This manuscript expands upon a model proposed in a previous publication from the same authors showing that primary cilium remodeling corresponds to a shift in SHH signaling activity during neuronal differentiation (https://doi.org/10.1126/sciadv.abb0601 [doi.org]). The current manuscript describes a series of embryo slice culture experiments that are performed to evaluate ciliary localization of the SHH-regulatory GPCRs Smoothened (SMO) and GPR161 and to determine how ciliary cAMP is affected by GPR161 knockdown and/or SMO activation or repression. Based upon the results of these experiments, the authors propose that differentiation-associated changes in SHH signaling activity are controlled by modulation of ciliary cAMP in response to shifts in SMO vs. GPR161 primary cilium enrichment during neuronal ciliary remodeling. Failure of cilia to increase cAMP during differentiation prevents axon elongation, prompting the authors to propose a model whereby ciliary cAMP signals to control actin dynamics in the growth cone. Although this is an interesting hypothesis, most of the experimental results provided are largely correlative. The model would be better supported if additional quantitative experimental data were included.

    * *

    Major comment:

    A key limitation of the study is that the authors do not evaluate whether changes in ciliary cAMP during neuronal differentiation are stimulated by mechanisms other than GPR161 or SMO signaling. This is important because it seems that during all stages of neuronal maturation, a significant proportion of primary cilia have neither SMO nor GPR161 (at least according to the gray bars in the graph in Figure 1A). As such, I'm not confident you can conclude that the observed effects on cAMP are due only to signals from SMO and GPR161. Other cell biological processes that influence ciliary elongation may contribute to the observed cAMP changes as cilia remodel. Ciliary elongation can be controlled through cAMP elevation that is driven by prostaglandin signaling through primary cilium-localized, Gαs-coupled EP4 (https://doi.org/10.1038/ncb3029 [doi.org]), which may be relevant if cilia are regressing and regrowing during apical abscission.

    We thank the reviewer for this important comment. We agree that changes in ciliary cAMP cannot be attributed exclusively to GPR161 and Smo, and we did not intend to imply that these proteins are the sole regulators of ciliary cAMP during neuronal differentiation. As discussed in our response to Reviewer 1, the ciliary localisation of GPR161 and Smo is likely dynamic, and changes in their relative abundance may contribute to cAMP regulation at different stages of neuronal differentiation. However, other ciliary GPCRs and signalling proteins not investigated in this study may also contribute to the regulation of ciliary cAMP.

    We note that our experiments involving GPR161 depletion and SAG-mediated activation of Smo demonstrate clear effects on ciliary cAMP, linking these Shh modulators to regulation of ciliary cAMP levels in differentiating neurons. Importantly, we have now directly tested the functional significance of ciliary cAMP levels independently of manipulation of GPR161 or Smo. We reduced available ciliary cAMP by expressing a cilium-targeted cAMP chelator (cAMP-Sponge; pCX-5HT6-cAMPSponge-mRFP, Atkins et al., 2023), as described in our response to Reviewer 1 and further elaborated in response to this reviewer’s later comment below. Sequestration of ciliary cAMP produced defects in neuronal polarisation that closely resembled those observed following GPR161 knockdown. These findings demonstrate that reducing available ciliary cAMP is itself sufficient to disrupt neuronal polarisation, independently of the specific upstream mechanism responsible for altering cAMP levels.

    Notably, ciliary cAMP sequestration produced a broader phenotype than GPR161 knockdown, including increased protrusion branching. While this does not formally demonstrate the involvement of additional ciliary regulators, it is consistent with ciliary cAMP integrating inputs from multiple signalling pathways during neuronal differentiation.

    We also agree that changes in ciliary morphology may contribute to the observed cAMP dynamics. The primary cilium undergoes substantial remodelling during neuronal delamination, and we observed fluctuations in cADDis intensity during this process. These observations are compatible with contributions from additional pathways associated with ciliary assembly, elongation, or maturation. Importantly, however, the major shift in ciliary cAMP levels occurs after the remodelled cilium has been delivered to the cell body, at which point ciliary remodelling is complete and intraflagellar trafficking has resumed (Toro-Tapia & Das, 2020). This temporal separation argues that ciliary remodelling alone is unlikely to account for the observed increase in cAMP.

    The reviewer's suggestion that prostaglandin signalling via ciliary EP4 receptors may contribute to ciliary cAMP regulation is interesting and plausible. However, whether EP4 or other cAMP-regulating receptor candidates localise to the primary cilium of differentiating neurons in the developing spinal cord remains unknown and is beyond the scope of the current study. Determining how additional ciliary GPCRs contribute to cAMP regulation during neuronal differentiation will therefore be an important avenue for future investigation.

    We now discuss the above points in the revised manuscript (lines 527-536):

    ‘Furthermore, the contribution of other ciliary GPCRs to cAMP regulation also cannot be excluded. Indeed, our observation that ciliary cAMP sequestration produces a more severe phenotype than GPR161 knockdown, including increased protrusion branching (Figure 4) is consistent with the possibility that multiple signalling pathways contribute to the maintenance of ciliary cAMP homeostasis during neuronal differentiation. For example, the cilium-localised prostaglandin receptor EP4 has been proposed to promote ciliary elongation by elevating ciliary cAMP (Jin et. al, 2014), raising the possibility that cAMP may contribute to ciliary reassembly during neuronal delamination. However, the marked increase in ciliary cAMP that we observe occurs only after the remodelled cilium enters the neuronal cell body, when ciliary reassembly and elongation are complete. Future studies will therefore be required to determine how GPR161, and other ciliary GPCRs, cooperate to establish distinct cAMP dynamics during neuronal differentiation.’

    -It would be helpful to discuss whether (or not) the SHH-mediated cytoskeletal regulation in the growth cone requires ciliary SMO signaling because some noncanonical signaling can occur from outside ciliary membrane.

    We thank the reviewer for this important comment. Our previous work and other studies support a role for non-canonical Shh signalling in axon extension and guidance (Toro-Tapia & Das, 2020; Yam et al., 2009). In the present study, we observed co-accumulation of GPR161 and Smo in the primary cilium and propose that their relative abundance contributes to ciliary cAMP regulation, thereby suppressing Gli activity and influencing cytoskeletal remodelling.

    Importantly, disrupting primary cilium function in neurons extending axons (Figure 6C’’) produced effects similar to those observed following Smo inhibition with cyclopamine (Figure 6D). These findings support a requirement for a functional primary cilium for the appropriate cytoskeletal response to Shh. However, they do not establish that the Smo activity driving this response occurs exclusively within the cilium. We therefore cannot exclude a contribution from extraciliary Smo, or the possibility that ciliary Smo and extraciliary Smo signalling is coordinated to generate cytoskeletal responses, including within the growth cone. We have expanded the Discussion to address the possible contributions of ciliary and extraciliary Smo signalling (lines 504-518):

    ‘These findings reveal an unexpected link between ciliary non-canonical Shh signalling and regulation of the spatial organisation of actin, which leads to the maturation of a single actin accumulation into an axon. However, the precise mechanisms through which ciliary non-canonical Shh signalling modulates actin dynamics and its intermediate regulators remain unclear. Furthermore, given that current models of Smo-mediated non-canonical Shh signalling are derived mainly from contexts in which signalling occurs outside the primary cilium (Bijlsma et al, 2012; Ho Wei et al, 2018), a contribution from cilium-independent non-canonical Shh signalling during neuronal differentiation cannot be excluded. Intriguingly, our finding that disruption of primary cilium function in neurons extending axons (Figure 6C’’) produced effects on actin organisation similar to those observed following Smo inhibition with cyclopamine (Figure 6D) support a requirement for an intact primary cilium for the appropriate cytoskeletal response to Shh. However, they do not establish that the Smo activity driving this response occurs exclusively within the cilium. We therefore cannot exclude a contribution from extraciliary Smo, or the possibility that ciliary Smo and extraciliary Smo signalling is coordinated to generate cytoskeletal responses. Future studies exploring the role of ciliary and extraciliary Smo and actin-modulating proteins in this process will delineate the intermediate signalling steps linking the primary cilium and non-canonical Shh signal transduction to the actin cytoskeleton.’

    -Please perform statistical analysis of the graphs shown in Figure 1 that show quantification of ciliary localization of SMO, GPR161, and ADCY3. Please indicate that ADCY3 is being scored in the graph in Figure 1C - from the labels "Absent" and "Present", it's not immediately clear what protein is being tracked.

    Thank you for highlighting these points. In response to reviewer 1, we have now also included data for GPR161, Smo and ADCY3 ciliary localisation in primary cilia of Tuj1 expressing cells that are attached to the apical process. We have also performed a contingency analysis to determine whether there is an association between relative frequencies of ciliary Smo/GPR161 across stages of neuronal differentiation. This revealed a significant association between differentiation stage and ciliary localisation state (chi-square test of independence, p = 0.0025), supporting our conclusion that the distribution of Smo-only, GPR161-only, Smo/GPR161-positive, and Smo/GPR161-negative cilia changes during neuronal differentiation. The proportion of ADCY3 accumulated in the cilia remained near 50%, with no statistically significant differences between differentiation stages, indicating that ADCY3 ciliary localisation remains relatively stable during neuronal differentiation. The new analysis for Smo and GPR161 is described in lines 106-108 and displayed in Figure 1B:

    ‘Contingency analysis of observations across different stages of neuronal differentiation demonstrated an association between differentiation stage and Smo and GPR161 ciliary accumulation state (pooled data from 4 embryos, Figure 1B’)’

    The new analysis for ADCY3 is described in lines 119-122 and Figures 1C’ and S1:

    ‘This revealed ciliary accumulation of ADCY3 in approximately 50% of remodelled primary cilia at all stages of primary cilium reassembly (Figure 1C’, Figure S1A; attached: 45% n=28/62, proximal: 46%, n= 46/100 cells; distal: 55%, n= 81/146 cells; axonogenesis: 51% n=63/124 cells, 4 embryos).’

    We have also modified the labelling in Figure 1C’ to make it clear that ADCY3 is being tracked.

    Lines 113-115 read as though the authors are suggesting that ADCY3 localizes to cilia because of SMO or GPR161 enrichment in the organelle. I'm not aware of evidence that the presence of SHH signaling components in the cilium impacts ADCY3 ciliary localization. Moreover, the quantification provided in Figure 1C suggests there is no significant difference in ADCY3 localization as neurons mature. Please rephrase to clarify.

    We apologise for not being clear here. The reviewer is correct that our data show no significant variation in ciliary ADCY3 localisation across the stages examined, with approximately 50% of primary cilia displaying detectable ADCY3. Our intention was not to suggest that ciliary localisation of ADCY3 is regulated by GPR161 or Smo. Rather, these experiments were included to establish that ADCY3 is present within the remodelled primary cilium of differentiating neurons and could therefore contribute to the regulation of ciliary cAMP. We have rephrased the relevant text to make this distinction clear and avoid implying a causal relationship between GPR161/Smo accumulation and ADCY3 localisation (lines 117-124):

    ‘To investigate the potential for ciliary cAMP production, we first fixed and labelled E10.5 mouse embryos for the cAMP-producing enzyme adenylyl cyclase III (ADCY3), which localises to neuronal primary cilia (Bishop et al, 2007) and Tuj1 (Figure 1C). This revealed ciliary accumulation of ADCY3 in approximately 50% of remodelled primary cilia at all stages of primary cilium reassembly (Figure 1C’; attached: 45% n=28/62, proximal: 46%, n= 46/100 cells; distal: 55%, n= 81/146 cells; axonogenesis: 51% n=63/124 cells, 4 embryos). Therefore, the persistent accumulation of ADCY3 in remodelled primary cilia, together with increased GPR161 accumulation as neuronal differentiation progresses, suggests that ciliary GPR161 accumulation may promote elevated ciliary cAMP levels.’

    Inclusion of a ciliary marker should be considered for the cADDis experiments shown in Figures 1 and 2. This will help to clearly identify cilia in cells with low ciliary cADDis expression. As presented, it's unclear if the cADDis signal is reduced due strictly to a ciliary cAMP response, or if the signal is reduced due to the cilia being smaller. By the end of movie 1, the cADDis cilium is invisible” Figure 2B - how is cADDis intensity normalized? Is this the radiometric sensor? If so, why isn't the mCherry cADDis signal visible in the cilium?”

    We agree that an independent ciliary marker would help distinguish differences in cADDis fluorescence resulting from changes in cAMP from those associated with changes in cilium size during remodelling. However, we are unable to do this due to technical constraints. The 5-HT6-mCherry-cADDis construct used in these experiments is a ratiometric sensor, with mCherry providing a cAMP-independent reference signal. However, we were unable to obtain sufficiently robust mCherry fluorescence within the cilium to perform reliable ratiometric measurements. Moreover, the mCherry reference fluorophore spectrally overlaps with mKate2, which we use as a cell membrane marker to monitor progression through neuronal differentiation, and whose expression is more abundant than 5-HT6-mCherry. We therefore could not simultaneously use mKate to track cellular morphology and mCherry for ratiometric normalisation. We also attempted using a BFP-labelled ciliary membrane marker, but the intensity of fluorescent signal was insufficient for reliable tracking, and prolonged 405-nm excitation was unsuitable for live-tissue imaging due to phototoxicity. These technical limitations prevented simultaneous monitoring of neuronal differentiation and ratiometric cADDis measurements.

    For the experiments presented in Figures 1 and 2, we therefore quantified the green cADDis fluorescence rather than calculating the cADDis /mCherry ratio. Importantly, during validation of the sensor using forskolin (Figure S1), analysis of green fluorescence alone and the cADDis/mCherry ratio produced a sharp decline in the cADDis green fluorescence response, supporting the use of the green channel as a readout of changes in cAMP under our imaging conditions.

    We now explicitly state this technical limitation in the revised manuscript (lines 140-142):

    ‘As the mCherry reference fluorophore of cilium-targeted cADDis overlaps spectrally with mKate2, which displays more robust expression, we were unable to perform ratiometric imaging and instead relied on cADDis green fluorescence.’

    Figure 2A - is there a way to more specifically quantify the effect of GPR161 knockdown on axon extension - rather than stating the percentage of cells that had axons? Is the change in presence or absence of an axon following GPR161 knockdown statistically significant? What happens to axon extension if SMO is knocked down?

    We thank the reviewer for this important suggestion. To provide a more robust assessment of the effect of GPR161 depletion, we have now quantified the number of protrusive events following GPR161 knockdown compared to control cells, rather than relying solely on the proportion of cells undergoing normal axon extension. Consistent with our observation that GPR161 knockdown results in the initiation of multiple transient protrusions, this analysis demonstrated a statistically significant increase in the number of protrusive events following GPR161 depletion. This data is presented in Figure 2A″ and described in lines 194-196:

    ‘Consistent with the formation of multiple transient projections, GPR161-knockdown cells exhibited significantly more protrusive events than control cells (Figure 2A’’).’

    Regarding Smo knockdown, our in ovo electroporation technique for introducing expression constructs predominantly transfects neural progenitor cells. Therefore, sustained depletion of Smo throughout neuronal differentiation could affect earlier Shh-dependent developmental processes, making it difficult to distinguish the specific requirement for Smo during neuronal polarisation. Instead, we used acute pharmacological manipulation of Smo function at defined stages of neuronal differentiation, allowing us to examine its requirement more specifically during neuronal polarisation. We have clarified this technical consideration in the manuscript (lines 227-228):

    ‘Acute pharmacological manipulation allowed us to assess Smo function specifically during neuronal polarisation, while minimising effects on Smo-dependent processes at earlier developmental stages.’

    -Figure 2D" - Why does the cADDis signal slowly increase, quickly decrease to untreated levels, and then rise again over the SAG time course? Is there a biological explanation for what might be going in here?

    Thank you for highlighting this interesting feature of the response. The initial increase in cADDis fluorescence may coincide with increased ciliary accumulation of Smo following exposure to SAG (Figure 1C″). However, we do not currently have sufficient evidence to determine the mechanism underlying the subsequent decrease and second increase in cADDis fluorescence following SAG treatment. One possibility is that sustained Smo activation produces a dynamic rather than monotonic change in ciliary signalling, potentially involving adaptation or compensatory regulation through Smo trafficking. Such regulation could generate transient changes in cAMP despite continued exposure to SAG. However, our experiments were not designed to distinguish between changes in Smo activity or trafficking and downstream mechanisms regulating ciliary cAMP and are beyond the scope of the current manuscript. We therefore refrain from assigning a specific mechanism to the biphasic response and have explicitly described this response in the revised manuscript (lines 232-233):

    ‘In contrast, SAG-mediated activation of Smo produced a non-monotonic response, with ciliary cAMP initially decreasing, subsequently increasing, and then decreasing again’

    -Can axon extension metrics be included for SAG treated cells?

    As with GPR161 knockdown, we have now quantified the number of protrusive events in SAG-treated cells for each experiment (Figure S4D and Figure S7B’’).

    -What happens to axon extension if you manipulate ciliary cAMP levels independently of GPR161 knockdown or SMO activation? This would provide nice supporting evidence that the observed effects are governed through cAMP modulation and not some other signal controlled by SMO or GPR161. What happens if dBcAMP is added? This would increase ciliary cAMP in a manner independent of GPR161

    -Can the GPR161 knockdown effect be rescued by adding back a cAMP analog like dBcAMP?

    We thank the reviewer for these important suggestions. Although dBcAMP could increase cAMP independently of GPR161 or Smo, it would elevate cAMP throughout the cell rather than specifically within the primary cilium. This is an important consideration in differentiating neurons, as global elevation of cAMP can itself influence neuronal differentiation and morphology. Indeed, dBcAMP has been reported to promote neuronal differentiation, neurite formation, elongation, and branching in neuronal cell models (Kovalevich & Langford, 2013) and has been widely used to enhance neuronal regeneration (Qiu et al., 2002). Thus, any rescue or morphological phenotype following dBcAMP treatment would be difficult to attribute specifically to restoration of the ciliary cAMP pool.

    To address the reviewer's central question more directly, we instead manipulated cAMP specifically within the primary cilium, independently of GPR161 or Smo, using a cilium-targeted cAMP chelator (cilium-targeted cAMP sponge; pCX-5HT6-cAMPSponge-mRFP) (Atkins et al., 2023), as also described in our response to Reviewer 1. Sequestration of ciliary cAMP phenocopied key aspects of GPR161 depletion, resulting in unstable axon-like protrusions and impaired establishment of a single axon. These findings provide independent evidence that reducing available cAMP specifically within the primary cilium is sufficient to disrupt neuronal polarisation and support the conclusion that the effects of GPR161 depletion are mediated, at least in part, through altered ciliary cAMP signalling rather than exclusively through another GPR161-dependent pathway.

    Notably, related to the reviewer’s earlier point about multifactorial regulation of ciliary cAMP, expression of cilium-targeted cAMP sponge also caused neurons to frequently form branched protrusions. We speculate that this reflects the broader disruption of ciliary cAMP signalling caused by the sponge, which is expected to affect cAMP generated or regulated by multiple ciliary GPCRs rather than GPR161 alone. As we note in the discussion (line 526), this observation is consistent with the possibility that additional ciliary GPCRs contribute to the regulation of ciliary cAMP during neuronal differentiation.

    These new results are described in lines 239-250 and presented in Figures 4 , S3 and Movies S8, 9 and 10):

    ‘To directly assess the importance of ciliary cAMP levels for neuron polarisation and axon initiation, we reduced available ciliary cAMP by expressing a cilium-targeted cAMP chelator fused to mRFP (cilium-targeted cAMPsponge-mRFP) (Atkins et al, 2023), followed by timelapse imaging. The majority of cilium-targeted mRFP expressing control cells initiated a single axon and underwent normal axon extension towards the ventral midline (Figure 4A, Movies S8; 89% 32/36 cells, 10 embryos). In contrast, cells expressing cilium-targeted cAMPsponge-mRFP, similar to GPR161 knockdown cells, often failed to initiate a normal axon and instead exhibited abnormal protrusive activity, generating multiple unstable axon-like projections (Figure 4B and 4C, Movie S9; 63% 17/27, 10 embryos), some of which underwent branching (Figure S3A, Movie S10). Therefore, local buffering of ciliary cAMP disrupts neuronal polarisation and normal axon initiation, resulting in multiple unstable, branched protrusions and impaired axon extension. Taken together, these findings indicate that the appropriate balance of ciliary Smo and GPR161 drives an increase in ciliary cAMP that is required for establishing a single axon.’

    -Figure 4 - It looks like cyclopamine was used at a concentration of 15 µM, which is on the high side. Cyclopamine demonstrates off-target cytotoxicity when used at high concentration. I suggest using a SMO antagonist like LDE225, which can inhibit SMO at lower concentrations (~10 nM in vitro) that do not trigger cytotoxicity.

    We thank the reviewer for raising this important concern. We agree that 15 µM cyclopamine is relatively high compared with concentrations used in some cell-culture systems and that high concentrations can produce off-target effects. However, inhibitor potency is highly assay and system dependent, and application to tissue-based assays generally requires higher concentrations. We have previously validated the use of cyclopamine at this concentration in ex vivo chick embryonic slices (Toro-Tapia and Das, 2020). During time-lapse imaging, we monitored overall tissue morphology and spinal cord integrity throughout the experiment and observed no overt deterioration or loss of tissue integrity for at least 8 hours following treatment. Thus, although we cannot completely exclude concentration-dependent off-target effects, we have no evidence of overt cytotoxicity under the conditions and time window used in these experiments. We therefore retained cyclopamine at 15 µM because this concentration has been validated in our experimental model and allows direct comparison with previous experiments. We agree that an alternative Smo antagonist such as LDE225 would provide an additional means of inhibiting Smo, but introducing a different inhibitor would require additional independent optimisation and validation in the ex vivo chick spinal cord assay.

    Figure 5A' - can axon collapse be quantified in a way that statistical analysis can be applied (i.e. Scoring number of failed axon extension attempts? Number of axon-like projections? Lengths? Retractions of failed axons?)? Rigor would be improved by including a quantitative way of comparing DMSO vs. SAG treated conditions.

    We have included these quantifications (protrusive events per cell) and comparisons in the revised manuscript (Figure 2, Figure 4, Figure S4, Figure S5 and Figure S7).

    Discussion comments:

    Line 325-227 states "Furthermore, our findings suggest that elevated ciliary cAMP levels regulate actin dynamics to facilitate the morphological polarisation of newborn neurons through the initiation of a clearly defined axon." How do you envision that ciliary cAMP translates into control of actin dynamics? What is the signal being transduced? This is briefly discussed around lines 345-350, but the discussion mainly states what is not known.

    We thank the reviewer for raising this important mechanistic question. We agree that the molecular pathway linking ciliary non-canonical Shh signalling to actin remodelling remains unresolved. Most studies investigating non-canonical Shh-dependent regulation of the actin cytoskeleton have focused on signalling components localised at or acting within the growth cone, including the Dock/Elmo complex (Makihara et al., 2018) and ZBP1 (Lepelletier et al., 2017). These mechanisms may therefore be more directly relevant to extraciliary Smo signalling. To our knowledge, a direct molecular pathway linking ciliary Shh/cAMP signalling to actin remodelling has not yet been established.

    Our findings nevertheless suggest a potential mechanism involving cAMP-dependent PKA signalling. GPR161 contains an AKAP motif that interacts with the type I regulatory subunit of PKA, providing a potential mechanism through which changes in ciliary GPR161 and cAMP could influence PKA ciliary localisation and activity. Changes in this ciliary signalling state could, in turn, influence downstream phosphorylation-dependent pathways regulating actin organisation and neuronal morphology. However, our experiments do not identify the relevant PKA substrates or establish how such a signal is transmitted from the primary cilium to the actin cytoskeleton.

    We have therefore revised the manuscript to distinguish our experimental conclusion that ciliary cAMP is required for normal neuronal polarisation from the proposed mechanism linking ciliary cAMP to actin remodelling. We use the phrase ‘ciliary cAMP levels influence F-actin dynamics’ instead of ‘ciliary cAMP levels regulate F-actin dynamics’ throughout and discuss cAMP-PKA signalling as one potential molecular link, while acknowledging that determining how this ciliary signal is communicated to the actin cytoskeleton will require further investigation (lines 537-543):

    ‘Nevertheless, GPR161 is particularly relevant in this context as it contains an AKAP (A-kinase anchoring protein) motif that specifically binds to the PKA regulatory subunit PKAI-RI, linking GPR161 to the regulation of ciliary PKA signalling, and likely to PKA ciliary accumulation (Bachmann et al, 2016; May et al, 2021). Thus, perturbation of GPR161 may influence both local cAMP signalling and downstream PKA-dependent responses, potentially coupling ciliary cAMP to phosphorylation-dependent regulation of actin organisation and neuronal morphology. Defining this mechanistic link between ciliary cAMP and actin architecture will be an important avenue for future investigation.’

    • Lines 341-342 refer to "primary cilium-mediated non-canonical Shh signalling...". Noncanonical SMO signalling is generally thought to occur from outside the primary cilium. Do you have direct evidence that the signal controlling actin dynamics occurs specifically through ciliary SMO? Please clarify.

    We thank the reviewer for this important clarification. We agree that our data do not establish that the signal controlling actin dynamics arises specifically from ciliary Smo. As discussed in the revised manuscript (lines 500-508), we cannot exclude contributions from extraciliary Smo signalling or coordinated signalling between ciliary and extraciliary Smo pools.

    Nevertheless, several findings reported in the revised manuscript support a functional requirement for the primary cilium in this process. SAG, which promotes Smo accumulation in the primary cilium (Rohatgi et al., 2007), increased Gli-dependent activity in differentiating neurons and was associated with altered actin dynamics and axon extension, phenocopying aspects of GPR161 depletion (Figure 5). In addition, CALI-mediated disruption of primary cilium function altered actin dynamics and produced effects comparable to those observed following Shh inhibition with cyclopamine (Figure 6). Finally, we have now performed additional experiments to sequester ciliary cAMP, which disrupts neuronal polarisation and prevents establishment of a single axon, providing independent evidence that perturbation of ciliary signalling is sufficient to affect neuronal polarisation (Figure 4). Together, these findings support a functional role for the primary cilium and ciliary cAMP signalling in regulating neuronal polarisation. However, they do not demonstrate that Smo signals directly from the ciliary membrane to regulate actin dynamics. We therefore cannot distinguish whether the cytoskeletal response is mediated directly by ciliary Smo, by extraciliary Smo, or through coordination between ciliary and extraciliary signalling compartments. We have explicitly clarified this distinction to avoid implying that non-canonical Smo signalling controlling actin dynamics occurs exclusively within the primary cilium (lines 502-518):

    ‘Our results indicate that optimal ciliary cAMP levels, maintained by balanced ciliary accumulation of GPR161 and Smo, define a transient cell state in which differentiating neurons become competent to polarise and initiate a nascent axon. These findings reveal an unexpected link between ciliary non-canonical Shh signalling and regulation of the spatial organisation of actin, which leads to the maturation of a single actin accumulation into an axon. However, the precise mechanisms through which ciliary non-canonical Shh signalling modulates actin dynamics and its intermediate regulators remain unclear. Furthermore, given that current models of Smo-mediated non-canonical Shh signalling are derived mainly from contexts in which signalling occurs outside the primary cilium (Bijlsma et al, 2012; Ho Wei et al, 2018), a contribution from cilium-independent non-canonical Shh signalling during neuronal differentiation cannot be excluded. Intriguingly, our finding that disruption of primary cilium function in neurons extending axons (Figure 6C’’) produced effects on actin organisation similar to those observed following Smo inhibition with cyclopamine (Figure 6D) support a requirement for an intact primary cilium for the appropriate cytoskeletal response to Shh. However, they do not establish that the Smo activity driving this response occurs exclusively within the cilium. We therefore cannot exclude a contribution from extraciliary Smo, or the possibility that ciliary Smo and extraciliary Smo signalling is coordinated to generate cytoskeletal responses. Future studies exploring the role of ciliary and extraciliary Smo and actin-modulating proteins in this process will delineate the intermediate signalling steps linking the primary cilium and non-canonical Shh signal transduction to the actin cytoskeleton.’

    Minor comments:

    • Line 89 - what is meant by "improved fidelity of multi-label immunofluorescence....

    We apologise for the lack of clarity. The original phrase referred to the greater reliability of multi-label immunofluorescence in mouse tissue compared with our chick system. Many of the primary antibodies used in these experiments were raised against mouse proteins and show lower affinity for their chicken orthologues, resulting in a reduced signal-to-background ratio in chick tissue. This limitation is particularly pronounced for directly conjugated primary antibodies, where signal amplification using secondary antibodies is not possible. We therefore performed these experiments in mouse embryos to achieve clearer and more reliable simultaneous detection of all four markers. We have revised the text to clarify this point (lines 90-92):

    ‘To investigate this, we utilised multi-label immunofluorescence in mouse embryos, which share key mechanisms of neuronal differentiation with chick (Kicheva et al, 2014) and enabled higher-resolution simultaneous detection of multiple markers.’

    Diagram figures are too small.

    Text on graphs is too small and is usually inconsistent with font size in the rest of the figure (ex. Figure 1 bar graphs).

    We have increased the size of the schematic diagrams and standardised the font size across all graphs to improve readability and consistency throughout the revised figures.

    Reviewer #2 (Significance (Required)):

    General assessment:

    A strength of the manuscript is that in vivo slice cultures are used to evaluate SMO, GPR161, and cAMP activity in differentiation neurons.

    Limitations include the lack of statistical analysis for many of the assays and the correlative nature of the results. Contributions by other ciliary GPCRs to ciliary cAMP control are not considered.

    Advance: The advance is that the study adjusts a prior model proposed by this same group in an earlier paper. As written, it represents an incremental advance in understanding of how neuronal differentiation proceeds.

    Audience: This work will be of interest to specialists in the field.

    We thank the reviewer for their assessment and for recognising the strength of our live imaging approach in embryonic slice cultures. We have sought to address the principal limitations identified by the reviewer through substantial additional experimentation and quantitative analysis. In particular, we have added statistical analyses across the relevant datasets and directly manipulated ciliary cAMP using a cilium-targeted cAMP sponge, demonstrating that disruption of the ciliary cAMP pool is sufficient to impair neuronal polarisation and single-axon initiation. We have also revised the discussion to consider the potential contribution of additional ciliary GPCRs to cAMP regulation. We believe these additions substantially strengthen the evidence supporting the central model and extend the study beyond the correlative observations in the original submission.

    Reviewer #3


    Evidence, reproducibility and clarity (Required):

    The current manuscript examines how ciliary cAMP levels regulated by Smo and GPR161 influence Shh signal interpretation and neuronal polarization during neural development. The authors utilize embryo slice culture, time-lapse imaging of Ftractin-GFP, CALI based ciliary ablation, and ciliary cAMP measurements in their studies. They show that primary cilium remodeling and the regulation of ciliary cAMP levels are essential for the suppression of Gli activity, ensuring the precise control of cytoskeletal organization and the initiation of a single, stable axon during neuronal differentiation. Furthermore, the dynamic balance between GPR161 and Smo in primary cilia is essential for proper F-actin organization, thereby regulating normal neuronal polarization and axon extension in embryonic development. Overall, the work is interesting, mechanistically insightful, and very well done. The study will be of general interest to readers from neuro-developmental biology and cell biology fields.

    I have a few comments that the authors might consider in discussing, interpreting the results and in strengthening the manuscript.

    First, the actual mechanisms underlying cAMP changes in cilia regulating actin dynamics in neurite projections is not clear. In this case, it would be useful for the readers to know the authors thoughts.

    We thank the reviewer for this important comment. In response to the reviewers’ feedback, we have expanded the Discussion to consider potential mechanisms linking ciliary cAMP signalling to actin remodelling. These possibilities are also addressed in our response to Reviewer 2 above and in the revised manuscript (lines 500-541).

    To summarise, our experiments do not identify the molecular intermediary that couples changes in ciliary cAMP to actin dynamics. However, one plausible mechanism involves cAMP-dependent regulation of PKA activity within the primary cilium. GPR161 contains an AKAP motif that interacts with the type I regulatory subunit of PKA, providing a potential mechanism through which changes in ciliary GPR161 could influence ciliary localisation and activity of PKA. Changes in ciliary cAMP could thereby alter local PKA activity and subsequent downstream phosphorylation-dependent signalling pathways that regulate the actin cytoskeleton. The identity and localisation of the relevant downstream effectors remain unknown, and it is also possible that multiple cAMP-dependent pathways contribute to the cytoskeletal response. We therefore present cAMP-PKA signalling as a potential mechanistic link rather than a conclusion of the present study and have revised the Discussion accordingly (lines 537-543):

    ‘Nevertheless, GPR161 is particularly relevant in this context as it contains an AKAP (A-kinase anchoring protein) motif that specifically binds to the PKA regulatory subunit PKAI-RI, linking GPR161 to the regulation of ciliary PKA signalling, and likely to PKA ciliary accumulation (Bachmann et al, 2016; May et al, 2021). Thus, perturbation of GPR161 may influence both local cAMP signalling and downstream PKA-dependent responses, potentially coupling ciliary cAMP to phosphorylation-dependent regulation of actin organisation and neuronal morphology. Defining this mechanistic link between ciliary cAMP and actin architecture will be an important avenue for future investigation.’

    • Second, the authors mention a switch from canonical to noncanonical HH signaling in neuronal polarization. The strict definition of noncanonical is unclear. There have been recent reports of GLI3R signaling, a result of ciliary GPR161 signaling in neuroepithelial cells, during cranial neural fold elevation and closure. The GLI3R levels might rather dictate the polarization switch.

    Thank you for noting this. We agree that the term “non-canonical” can encompass several distinct modes of Hedgehog signalling. We have therefore revised the manuscript to clarify that, in this study, we use “non-canonical Shh signalling” specifically to refer to Shh signalling characterised by suppression of Gli-mediated transcription (lines 55-58):

    ‘Following this, delaminating newborn neurons remodel a new primary cilium and undergo a switch to Smo-dependent non-canonical Shh signal interpretation characterised by suppression of Gli-activity, henceforth referred to as non-canonical Shh signalling’

    As the reviewer notes, GPR161-dependent regulation of the Gli3 repressor could control this switch. However, we did not directly examine Gli repressor forms in this study. Our GBS-NLS-KikGR reporter measures overall Gli transcriptional activity and therefore allows us to detect pathway activation or suppression but does not distinguish the specific contribution of Gli3R. We have discussed this possibility in the revised manuscript (lines 552-562):


    ‘Consistent with an important role for ciliary GPR161 in neuronal development, GPR161 has been implicated in cranial closure and neural precursor patterning through regulation of Gli3R levels (Brooks et al, 2025), highlighting the importance of its ciliary localisation for the correct downstream signalling. Determining whether this mechanism contributes to the suppression of Gli activity that we observe in differentiating neurons will be an important direction for future investigation. GPR161 and other GPCRs have also been reported to regulate cellular sensitivity to Shh, suppressing responses to low levels and attenuating responses to high levels of Shh in neural progenitor cells (Pusapati et al, 2018). These observations are consistent with our model in which GPR161 contributes to fine-tuning ciliary cAMP levels and, consequently, appropriate interpretation of Shh signalling during neuronal differentiation. Moreover, disruption of GPR161 has been associated with defects in neuronal migration and ectopic accumulation of differentiating neurons, leading to periventricular heterotopia in mice (Shimada et al, 2019), further illustrating how altered GPR161-mediated Shh signalling can disrupt neural architecture.’

    *Other points: *

    -Comments on quantitative analysis and statistical evaluation: A substantial portion of the data lacks proper quantification and statistical analysis. For example: In Figure 1E, how much does the cADDis fluorescence decrease over time following FSK treatment? In Figure 2A, what proportion of cells show defects in axon initiation after GPR161 knockdown? In Figure 3, what percentage of cells are Gli-positive? These results should be presented with clear quantitative plots along with appropriate statistical analysis to demonstrate significance.

    We agree that quantitative analysis and appropriate statistical evaluation are important for interpreting these experiments. We have therefore added extensive quantitative analyses to the revised manuscript, as also detailed in our responses to referees 1 and 2 above.

    For the forskolin validation experiment associated with Figure 1E (Figure S1), images were acquired at 5-minute intervals for 45 minutes following treatment with 10 µM forskolin. A decrease in green cADDis fluorescence was detectable immediately following forskolin addition, reaching approximately 50% of the initial fluorescence intensity. We have also quantified this response using area-under-the-curve (AUC) analysis, which demonstrated a statistically significant response to forskolin and confirms the responsiveness of the cADDis biosensor under our experimental conditions (Figure S1C’’).

    For Figure 2A, 71% of cells following GPR161 knockdown exhibited abnormal axon extension (line 191). To provide a more quantitative assessment of this phenotype, we have additionally quantified the number of protrusive events per cell in timelapses (Figure 2A″). This analysis demonstrated a statistically significant increase in protrusive events following GPR161 knockdown compared with control cells (lines 195-197), consistent with our observation that GPR161 depletion results in the formation of multiple transient projections rather than the stable extension of a single axon.

    For Figure 3 (now Figure 5), we have added the fluorescence recovery measurements for individual cells in each experiment, allowing the extent of Gli activation to be visualised directly. We also report the proportion of cells exhibiting Gli activation for each condition: 19% following control pRFPRNAi-Luc treatment compared with 81% following GPR161 knockdown (line 298), and 23% following DMSO treatment compared with 62% following SAG treatment (line 311). The corresponding statistical analyses are now included in the revised Figure 5 and manuscript.

    Together, these additions provide quantitative and statistical support for the cADDis, neuronal polarisation, and Gli-activity phenotypes described in the manuscript.

    Comparison between SMO inhibition and activation: In Figure 4D and Figure 5B', the authors used a Smo antagonist and agonist, respectively. It would strengthen the study to directly compare these conditions by quantifying the number of axon-like protrusions per cell and a comparison between these two treatments. Such analysis would help clarify how modulation of Smo activity differentially affects neuronal morphology.

    We thank the reviewer for this helpful suggestion. We have now added quantitative analyses of protrusion number under both conditions, together with the corresponding statistical analyses (Figures S3D and S4D″). Smo activation with SAG significantly increased the number of protrusions compared with the DMSO control, consistent with enhanced maturation of F-actin accumulations into protrusive structures. In contrast, Smo inhibition with cyclopamine did not significantly alter protrusion number compared with the ethanol control. Following cyclopamine treatment, F-actin was redistributed and overall F-actin accumulation decreased over time, suggesting that Smo inhibition prevents the maturation of F-actin accumulations that drives protrusive behaviour. These analyses indicate that activation and inhibition of Smo produce reciprocal effects on actin organisation, which produce distinct effects on protrusion dynamics. This comparison has been added to the revised manuscript in lines 445-448:

    ‘Notably, in contrast to cyclopamine treatment, which dispersed actin accumulations but did not significantly alter the number of protrusive events (Figure S5D’’, SAG treated cells were characterised by increased protrusive activity along with aggregated actin accumulations (Figure S7B’’), demonstrating that inhibition and activation of Smo differentially regulate actin organisation and neuronal morphology’

    -3. Varying time points in imaging data: In Figure 4A and 4A', the time points shown are not consistent (Figure 4A starts at 1 h 30 min, whereas Figure 4A' starts at 5 h 50 min). The authors should clarify the rationale for this discrepancy or present comparable time points to ensure consistency and facilitate interpretation.

    The different time points indicate the absolute time elapsed from the start of each continuous time-lapse acquisition, rather than equivalent stages of neuronal differentiation. Although cells of interest were generally identified at the start of imaging, individual neurons initiated differentiation or entered the field of view at different times during the time-lapse sequence. For example, the cell shown in Figure 7A′ entered the field of view approximately 5 hours after imaging began. Thus, the different timestamps do not reflect differences in treatment or imaging conditions but rather the different times at which individual neurons became available for observation. We have retained the original acquisition times to accurately represent the time-lapse experiments and clarified this point in the revised Methods (lines 631-634):

    ‘Although cells of interest were generally identified at the start of imaging, individual neurons initiated differentiation or entered the field of view at different times during the time-lapse sequence. Timestamps on time-lapse imaging datasets indicate the absolute time elapsed from the start of each continuous time-lapse acquisition, rather than equivalent stages of neuronal differentiation.’

    • Finally, studies before showed role of lack of Gpr161 in periventricular nodular heterotopia, which could be discussed along with ciliopathies with neuronal migration defects.

    We have added this to the discussion (lines560-563):

    ‘Moreover, disruption of GPR161 has been associated with defects in neuronal migration and ectopic accumulation of differentiating neurons, leading to periventricular heterotopia in mice (Shimada et al., 2019), further illustrating how altered GPR161-mediated Shh signalling can disrupt neural architecture.’

    Reviewer #3 (Significance (Required)):

    During spinal cord development, neuroepithelial progenitor cells undergoing neuronal differentiation delaminate from the neuroepithelium via apical abscission, a regulated process that coincides with disassembly of the primary cilium. Upon cell cycle exit, delaminating neurons subsequently reassemble a new primary cilium; however, the mechanistic basis by which this ciliary remodeling suppresses Gli transcription factor activity remains poorly understood. Similarly, how context-dependent Sonic Hedgehog (Shh) signaling shapes the cytoskeletal organization of newborn neurons to drive the initiation and maintenance of a single, stable axon is largely unresolved. This work identifies ciliary cAMP as a central orchestrator of Shh signal interpretation, establishing a mechanistic link between primary cilium-mediated signaling and the precise cytoskeletal regulation underlying axon formation. They show that primary cilium remodeling and the regulation of ciliary cAMP levels are essential for the suppression of Gli activity, ensuring the precise control of cytoskeletal organization and the initiation of a single, stable axon during neuronal differentiation. Furthermore, the dynamic balance between GPR161 and Smo in primary cilia is essential for proper F-actin organization, thereby regulating normal neuronal polarization and axon extension in embryonic development.The study will be of general interest to readers from neuro-developmental biology and cell biology fields.

    We thank the reviewer for this very positive and thoughtful assessment of the significance of our study. We greatly appreciate their recognition of the mechanistic insight provided by our findings, the importance of ciliary cAMP in shaping Shh signal interpretation during neuronal polarisation. We are also encouraged by the reviewer’s view that our work reveals an important functional link between primary cilium signalling, cytoskeletal organisation and axon formation, and that these findings will be of broad interest to the neurodevelopmental and cell biology communities.

    Atkins, M., Wurmser, M., Darmon, M., Roche, F., Nicol, X., & Metin, C. (2023). CXCL12 targets the primary cilium cAMP/cGMP ratio to regulate cell polarity during migration. Nat Commun,* 14*(1), 8003. https://doi.org/10.1038/s41467-023-43645-w

    Kovalevich, J., & Langford, D. (2013). Considerations for the use of SH-SY5Y neuroblastoma cells in neurobiology. Methods Mol Biol,* 1078*, 9-21. https://doi.org/10.1007/978-1-62703-640-5_2

    Lepelletier, L., Langlois, S. D., Kent, C. B., Welshhans, K., Morin, S., Bassell, G. J., Yam, P. T., & Charron, F. (2017). Sonic Hedgehog Guides Axons via Zipcode Binding Protein 1-Mediated Local Translation. J Neurosci,* 37*(7), 1685-1695. https://doi.org/10.1523/JNEUROSCI.3016-16.2016

    Makihara, S., Morin, S., Ferent, J., Cote, J. F., Yam, P. T., & Charron, F. (2018). Polarized Dock Activity Drives Shh-Mediated Axon Guidance. Dev Cell,* 46*(4), 410-425 e417. https://doi.org/10.1016/j.devcel.2018.07.007

    Moore, B. S., Stepanchick, A. N., Tewson, P. H., Hartle, C. M., Zhang, J., Quinn, A. M., Hughes, T. E., & Mirshahi, T. (2016). Cilia have high cAMP levels that are inhibited by Sonic Hedgehog-regulated calcium dynamics. Proc Natl Acad Sci U S A,* 113*(46), 13069-13074. https://doi.org/10.1073/pnas.1602393113

    Qiu, J., Cai, D., Dai, H., McAtee, M., Hoffman, P. N., Bregman, B. S., & Filbin, M. T. (2002). Spinal axon regeneration induced by elevation of cyclic AMP. Neuron,* 34*(6), 895-903. https://doi.org/10.1016/s0896-6273(02)00730-4

    Rohatgi, R., Milenkovic, L., & Scott, M. P. (2007). Patched1 regulates hedgehog signaling at the primary cilium. Science,* 317*(5836), 372-376. https://doi.org/10.1126/science.1139740

    Shimada, I. S., Somatilaka, B. N., Hwang, S. H., Anderson, A. G., Shelton, J. M., Rajaram, V., Konopka, G., & Mukhopadhyay, S. (2019). Derepression of sonic hedgehog signaling upon Gpr161 deletion unravels forebrain and ventricular abnormalities. Dev Biol,* 450*(1), 47-62. https://doi.org/10.1016/j.ydbio.2019.03.011

    Toro-Tapia, G., & Das, R. M. (2020). Primary cilium remodeling mediates a cell signaling switch in differentiating neurons. Sci Adv,* 6*(21), eabb0601. https://doi.org/10.1126/sciadv.abb0601

    Yam, P. T., Langlois, S. D., Morin, S., & Charron, F. (2009). Sonic hedgehog guides axons through a noncanonical, Src-family-kinase-dependent signaling pathway. Neuron,* 62*(3), 349-362. https://doi.org/10.1016/j.neuron.2009.03.022

  2. Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.

    Learn more at Review Commons


    Referee #3

    Evidence, reproducibility and clarity

    The current manuscript examines how ciliary cAMP levels regulated by Smo and GPR161 influence Shh signal interpretation and neuronal polarization during neural development. The authors utilize embryo slice culture, time-lapse imaging of Ftractin-GFP, CALI based ciliary ablation, and ciliary cAMP measurements in their studies. They show that primary cilium remodeling and the regulation of ciliary cAMP levels are essential for the suppression of Gli activity, ensuring the precise control of cytoskeletal organization and the initiation of a single, stable axon during neuronal differentiation. Furthermore, the dynamic balance between GPR161 and Smo in primary cilia is essential for proper F-actin organization, thereby regulating normal neuronal polarization and axon extension in embryonic development. Overall, the work is interesting, mechanistically insightful, and very well done. The study will be of general interest to readers from neuro-developmental biology and cell biology fields.

    I have a few comments that the authors might consider in discussing, interpreting the results and in strengthening the manuscript.

    First, the actual mechanisms underlying cAMP changes in cilia regulating actin dynamics in neurite projections is not clear. In this case, it would be useful for the readers to know the authors thoughts.

    Second, the authors mention a switch from canonical to noncanonical HH signaling in neuronal polarization. The strict definition of noncanonical is unclear. There have been recent reports of GLI3R signaling, a result of ciliary GPR161 signaling in neuroepithelial cells, during cranial neural fold elevation and closure. The GLI3R levels might rather dictate the polarization switch.

    Other points:

    1. Comments on quantitative analysis and statistical evaluation: A substantial portion of the data lacks proper quantification and statistical analysis. For example: In Figure 1E, how much does the cADDis fluorescence decrease over time following FSK treatment? In Figure 2A, what proportion of cells show defects in axon initiation after GPR161 knockdown? In Figure 3, what percentage of cells are Gli-positive? These results should be presented with clear quantitative plots along with appropriate statistical analysis to demonstrate significance.
    2. Comparison between SMO inhibition and activation: In Figure 4D and Figure 5B', the authors used a Smo antagonist and agonist, respectively. It would strengthen the study to directly compare these conditions by quantifying the number of axon-like protrusions per cell and a comparison between these two treatments. Such analysis would help clarify how modulation of Smo activity differentially affects neuronal morphology.
    3. Varying time points in imaging data: In Figure 4A and 4A', the time points shown are not consistent (Figure 4A starts at 1 h 30 min, whereas Figure 4A' starts at 5 h 50 min). The authors should clarify the rationale for this discrepancy or present comparable time points to ensure consistency and facilitate interpretation.
    4. Finally, studies before showed role of lack of Gpr161 in periventricular nodular heterotopia, which could be discussed along with ciliopathies with neuronal migration defects.

    Significance

    During spinal cord development, neuroepithelial progenitor cells undergoing neuronal differentiation delaminate from the neuroepithelium via apical abscission, a regulated process that coincides with disassembly of the primary cilium. Upon cell cycle exit, delaminating neurons subsequently reassemble a new primary cilium; however, the mechanistic basis by which this ciliary remodeling suppresses Gli transcription factor activity remains poorly understood. Similarly, how context-dependent Sonic Hedgehog (Shh) signaling shapes the cytoskeletal organization of newborn neurons to drive the initiation and maintenance of a single, stable axon is largely unresolved. This work identifies ciliary cAMP as a central orchestrator of Shh signal interpretation, establishing a mechanistic link between primary cilium-mediated signaling and the precise cytoskeletal regulation underlying axon formation. They show that primary cilium remodeling and the regulation of ciliary cAMP levels are essential for the suppression of Gli activity, ensuring the precise control of cytoskeletal organization and the initiation of a single, stable axon during neuronal differentiation. Furthermore, the dynamic balance between GPR161 and Smo in primary cilia is essential for proper F-actin organization, thereby regulating normal neuronal polarization and axon extension in embryonic development.The study will be of general interest to readers from neuro-developmental biology and cell biology fields.

  3. Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.

    Learn more at Review Commons


    Referee #2

    Evidence, reproducibility and clarity

    Summary: This manuscript expands upon a model proposed in a previous publication from the same authors showing that primary cilium remodeling corresponds to a shift in SHH signaling activity during neuronal differentiation (https://doi.org/10.1126/sciadv.abb0601). The current manuscript describes a series of embryo slice culture experiments that are performed to evaluate ciliary localization of the SHH-regulatory GPCRs Smoothened (SMO) and GPR161 and to determine how ciliary cAMP is affected by GPR161 knockdown and/or SMO activation or repression. Based upon the results of these experiments, the authors propose that differentiation-associated changes in SHH signaling activity are controlled by modulation of ciliary cAMP in response to shifts in SMO vs. GPR161 primary cilium enrichment during neuronal ciliary remodeling. Failure of cilia to increase cAMP during differentiation prevents axon elongation, prompting the authors to propose a model whereby ciliary cAMP signals to control actin dynamics in the growth cone. Although this is an interesting hypothesis, most of the experimental results provided are largely correlative. The model would be better supported if additional quantitative experimental data were included.

    Major comments: A key limitation of the study is that the authors do not evaluate whether changes in ciliary cAMP during neuronal differentiation are stimulated by mechanisms other than GPR161 or SMO signaling. This is important because it seems that during all stages of neuronal maturation, a significant proportion of primary cilia have neither SMO nor GPR161 (at least according to the gray bars in the graph in Figure 1A). As such, I'm not confident you can conclude that the observed effects on cAMP are due only to signals from SMO and GPR161. Other cell biological processes that influence ciliary elongation may contribute to the observed cAMP changes as cilia remodel. Ciliary elongation can be controlled through cAMP elevation that is driven by prostaglandin signaling through primary cilium-localized, Gαs-coupled EP4 (https://doi.org/10.1038/ncb3029), which may be relevant if cilia are regressing and regrowing during apical abscission.

    It would be helpful to discuss whether (or not) the SHH-mediated cytoskeletal regulation in the growth cone requires ciliary SMO signaling because some noncanonical SMO signaling can occur from outside ciliary membrane.

    Please perform statistical analysis of the graphs shown in Figure 1 that show quantification of ciliary localization of SMO, GPR161, and ADCY3. Please indicate that ADCY3 is being scored in the graph in Figure 1C - from the labels "Absent" and "Present", it's not immediately clear what protein is being tracked.

    Lines 113-115 read as though the authors are suggesting that ADCY3 localizes to cilia because of SMO or GPR161 enrichment in the organelle. I'm not aware of evidence that the presence of SHH signaling components in the cilium impacts ADCY3 ciliary localization. Moreover, the quantification provided in Figure 1C suggests there is no significant difference in ADCY3 localization as neurons mature. Please rephrase to clarify.

    Inclusion of a ciliary marker should be considered for the cADDis experiments shown in Figures 1 and 2. This will help to clearly identify cilia in cells with low ciliary cADDis expression. As presented, it's unclear if the cADDis signal is reduced due strictly to a ciliary cAMP response, or if the signal is reduced due to the cilia being smaller. By the end of movie 1, the cADDis cilium is invisible.

    Figure 2A - is there a way to more specifically quantify the effect of GPR161 knockdown on axon extension - rather than stating the percentage of cells that had axons? Is the change in presence or absence of an axon following GPR161 knockdown statistically significant? What happens to axon extension if SMO is knocked down?

    Figure 2B - how is cADDis intensity normalized? Is this the radiometric sensor? If so, why isn't the mCherry cADDis signal visible in the cilium?

    Figure 2D" - Why does the cADDis signal slowly increase, quickly decrease to untreated levels, and then rise again over the SAG time course? Is there a biological explanation for what might be going in here?

    Can axon extension metrics be included for SAG treated cells?

    What happens to axon extension if you manipulate ciliary cAMP levels independently of GPR161 knockdown or SMO activation? This would provide nice supporting evidence that the observed effects are governed through cAMP modulation and not some other signal controlled by SMO or GPR161. What happens if dBcAMP is added? This would increase ciliary cAMP in a manner independent of GPR161.

    Figure 4 - It looks like cyclopamine was used at a concentration of 15 µM, which is on the high side. Cyclopamine demonstrates off-target cytotoxicity when used at high concentration. I suggest using a SMO antagonist like LDE225, which can inhibit SMO at lower concentrations (~10 nM in vitro) that do not trigger cytotoxicity.

    Figure 5A' - can axon collapse be quantified in a way that statistical analysis can be applied (i.e. Scoring number of failed axon extension attempts? Number of axon-like projections? Lengths? Retractions of failed axons?)? Rigor would be improved by including a quantitative way of comparing DMSO vs. SAG treated conditions.

    Can the GPR161 knockdown effect be rescued by adding back a cAMP analog like dBcAMP?

    Discussion comments:

    Line 325-227 states "Furthermore, our findings suggest that elevated ciliary cAMP levels regulate actin dynamics to facilitate the morphological polarisation of newborn neurons through the initiation of a clearly defined axon." How do you envision that ciliary cAMP translates into control of actin dynamics? What is the signal being transduced? This is briefly discussed around lines 345-350, but the discussion mainly states what is not known.

    Lines 341-342 refer to "primary cilium-mediated non-canonical Shh signalling...". Noncanonical SMO signalling is generally thought to occur from outside the primary cilium. Do you have direct evidence that the signal controlling actin dynamics occurs specifically through ciliary SMO? Please clarify.

    Minor comments:

    Line 89 - what is meant by "improved fidelity of multi-label immunofluorescence...."

    Diagram figures are too small.

    Text on graphs is too small and is usually inconsistent with font size in the rest of the figure (ex. Figure 1 bar graphs).

    Significance

    General assessment:

    A strength of the manuscript is that in vivo slice cultures are used to evaluate SMO, GPR161, and cAMP activity in differentiation neurons. Limitations include the lack of statistical analysis for many of the assays and the correlative nature of the results. Contributions by other ciliary GPCRs to ciliary cAMP control are not considered.

    Advance: The advance is that the study adjusts a prior model proposed by this same group in an earlier paper. As written, it represents an incremental advance in understanding of how neuronal differentiation proceeds.

    Audience: This work will be of interest to specialists in the field.

  4. Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.

    Learn more at Review Commons


    Referee #1

    Evidence, reproducibility and clarity

    Summary:

    This study by Gabriela Toro-Tapia et al explores the role of the opposing Shh signalling modulators Smoothened and GPR161 during the remodelling of the Shh downstream signalling cascade that the same authors have proposed to occur upon newborn neurons delamination from the apical surface, a phenomenon that entails ciliary "shedding" and regrowth. They explore the possibility that the conversion from a canonical Shh pathway based on Gli TF activation to a non-canonical pathway relies on a shift from a mutually exclusive ciliary localization of the two transmembrane factors Smo and GPR161 to a (frequent) ciliary colocalization that would result in simultaneous ciliary Smo and elevated ciliary cAMP levels, underlying the non-conventional pathway. To explore this hypothesis, the study first characterizes the co-detection of the two receptors during the neuronal delamination and morphogenesis sequence, and then presents a series of functional perturbations centered on the modulation of cAMP levels in the cilium. The authors use the chick embryonic spinal cord as a model of neuronal delamination and axonogenesis; they use in ovo electroporation and live imaging in ex ovo slice culture to characterize cellular changes and differenciation of newborn neurons with a single cell subcellular and cellular resolution in an "intact" 3D cellular environment; except for the dynamic characteriztion of co-detection of GPR161 and Smo, which is performed in mouse embryo sections at a similar development stage for technical reasons, all the study is performed in the chick embryo The authors first illustrate an increased co-detection of Smo and GPR161 in delaminated neurons as the apical endfoot retracts, and report on the presence of the adenylyl cyclase ADCY3 in the regrowing/remaining cilium portion of half of these cells. Using a ciliary cAMP sensor cADDI, they observe a transient increase in reporter activity (indicating transient decrease in ciliary AMPc) in the reassembling cilium, followed by an increase in cAMP levels at the end of retraction. The authors then use shRNA-mediated downregulation of GPR161, and observe defects in neuronal repolarization; these cells also display reduced ciliary cAMP levels monitored with the cADDIs sensor; similar phenotypes are obtained when increasing Smo accumulation in the cilium by SAG treatment, indicating that the balance between ciliary Smo and GPR161 regulates ciliary levels of AMPc, and that this level may in turn control the cell remodeling associated with neuronal morphogenesis Using a photoconvertible reporter of Gli activation (GBS-nls-kikGR) that reports canonical Shh signaling in GPR-161 knock-down cells, the authors then show that GRP161 is required in delaminating cells to switch off canonical signaling; using the reporter in SAG-treated cells again shows that the balance between ciliary Smo and GPR161 is essential. The authors then describe the dynamics of actin cytoskeleton reorganization during neuronal polarization, and show accumulations of F-actin at the site of axon emergence and in dynamic protrusions associated with axon formation. They then use Chromophore-Assisted Light Inactivation to destroy the cilium in young neurons undergoing axonal extension, and show that this eliminates actin accumulations and stops and reverts axonal growth, showing a requirement for the cilium in the process; blocking Shh signalling with the Smo antagonist cyclopamine leads to the same failure in polarization and/or reversion of axonal growth. Finally, they show that GPR161 knock-down or SAG treatment lead to similar defects in F-actin accumulation and protrusion formation, and conclude that the balance between Smo and GPR161 in the reformed cilium after delamination is crucial to control cAMP levels in response to Shh signalling, and that this leads to an (as yet not elucidated) non canonical, Gli-activation independant transduction that orchestrates the cytoskeletal remodeling involved in neuronal morphological remodeling and axonal extension

    Major comment:

    The central point of the paper is that cAMP levels in the cilium are key in the transition in Shh signal transduction. However, functionally, cAMP levels are always adressed indirectly; the authors provide several converging lines of evidence, with functional challenges that result in a change in ciliary cAMP level and a simultaneous range of defects (Shh reporter read out, morphological changes). While overall the convergence of multiple approaches supports their claim, one might object that this does not directly prove that ciliary cAMP levels control the downstream defects. For example, SAG treatment may also activate non-ciliary Smo functions; similarly, we cannot formally exclude the (unlikely) hypothesis that GPR161 knock-down may also affect the process in a cAMP-independant manner. To provide direct evidence that ciliary cAMP levels are central in the phenomenon that they describe, the authors could express a ciliary cAMP sponge, such as the one used in the 2023 study by Atkins et al (ref 29 in the manuscript); the expectation is that this should reproduce all the phenotypes (e.g. Gli-reporter activity in differentiating neurons, defects in polarization) observed with their different approaches (SAG treatment, GPR161 knock-down). The most direct phenotype to describe, should the authors perform these experiements, would probably be the morphological defects in neurons. (Alternatively, using the photoactivable ciliary Adenylyl cyclase used in the same study by Atkins et al to restore high cAMP in the cilium may rescue the phenotypes caused by SAG treatment or GPR161 KD). Using a direct modulation of ciliary cAMP would, in my view, reinforce the conclusions of the study.

    minor comments

    Fig1B: do the authors have values for colocalization of GPR161 and Smo in attached cilia? The separate values (shown in Toro-Tapia 2022, Figure 4) measured in the chick embryo suggest that co-detection should be close to zero, single GPR161 should be low, and single Smo should be high, but not every reader will have the data from the previous study in mind, so it would be good to remind them. Unless there is a technical impossibity to obtain these values in Tuj1+ attached cells, it would be even better to have the double values in mouse to complement the graph in Figure 1B and illustrate the increase in double labelling from attached -> detached proximal -> detached distal. Similarly, in figure 1C, how does ADCY3 look like in Tuj1+ attached cells? In this same figure, some statistical validation of the significance of the increase in double positive cells would also be appreciated. Figure 1E: could the authors comment on the "transient" nature and overall timing of the cADDIs signal increase during retraction? I find it counter intuitive that the signal is low at the beginning of the time series in Figure 1E (I expected a low cAMP concentration and high signal right from the start of he sequence); besides, the drop in signal intensity also appears quite late in the sequence, when the apical extension is fully resorbed, compared to the timing of the appearance in double GPR161/Smo-positive ciliary remnants in Figure 1B. This does not really fit with the observation that the number of single or double GPR161 cells is lower in cells undergoing axonogenesis that in delaminating cells with a distal ciliary remnant. Is there an explanation for the delay between the increase in the number of double positive cells, and the actual shift in cAMP levels? Could it be linked to the changes in size of the ciliary remnant/cilium itself, which may need to regrow to a threshold size for signal reception and transduction to become effective? In figure S4, the images only show heatmap of the F-tractin signal; the "GFP-GPI" and "Arl13b supernova" mentions should be removed from the figure; the color code for F-trctin-GFP may be changed from green to "heatmap-like" color code;Similarly, in Fig S6, "luciferase RNAi" and "GPR161 RNAi" are the treatment applied to the cells, and should be indicated in black in a lateral, vertical box (like DMSO or SAG treatments) In figure 6, the color code is confusing; why are both Smo, GliA and Actin dynamics in green, and GPR161 and the the cilium in magenta? Maybe using more colors would help ; In Figure 6: the construction of the visual summary also suggests that the inhibition of GliA regulates actin dynamics, but I do not think that the authors provide functional evidence for this causal link. The data (especially the Gli reporter data) is overall compatible with this notion, and it is certainly the most likely interpretation, but it does not prove it directly, and I would suggest a dashed arrow and a question mark here. In Figure 2D", one of the 25 dots in the SAG condition is missing in the AUC graph.

    Significance

    This is a very carefully constructed study. The authors use complex manipulation of gain/loss of function with reporters and sensors in the context of live imaging to nicely approach the dynamics of changes in Shh signal reception and transduction during the process of neuronal delamination and re-polarization. The use of subcellular reporters of signalling dynamics and morphological changes is particularly impressive. I command them for this brave and beautiful piece of work. In a previous study (Toro-Tapia et al, 2022), the same authors had shown that Smo and GPR161 were both observed in the remodelled cilium of the newborn neuron, and although they had not visualized them simultaneously, they postulated the the presence of GPR161 in the cilium was responsible from a shift from the canonical Shh->Smo->GliActivation pathway to a non-canonical pathway downstream of Smo. Here, they explore the hypothesis that this transition is achieved via the GPR161-dependent maintenance of elevated ciliary cAMP concentration. One strength of the study is that they use a combination of complementary functional approaches and sensors that all converge towards the validation of this hypothesis. Its main limitation, in my view, is that it lacks an experiment that directly assesses the specific role of ciliary cAMP concentration. Another strength is that it addresses this question in live conditions in a "near" native tissue environment via the authors elegant ex vivo slice culture assay. This is a rare feat, which sets this work apart from many studies on axonal polarization and growth that are performed in dissociated cells. One open question for future investigations is why the remodeled cilia after "cilium shedding" tolerates the simultaneous entry of Smo and GPR161, while it does not happen in the cell-cycle regulated ciliary resorptions and regrowth sequence. From a clinical perspective, the work opens leads to future studies and interpretation of phenotypes associated with ciliopathies, including Joubert and Meckel syndromes, as pointed out by the authors in their discussion. This work should be of interest to cell biologists and developmental biologists alike, and is also relevant to clinicians interested in neurodevelopmental disorders and more broadly in ciliopathies.

    Reviewers expertise:

    Developmental neurosciences, vertebrate neurogenesis, ex vivo live imaging.

  5. Ciliary cAMP regulates Shh signal interpretation to drive polarisation of differentiating neurons https://www.biorxiv.org/content/10.64898/2026.03.11.711079v1

  6. Ciliary cAMP regulates Shh signal interpretation to drive polarisation of differentiating neurons https://www.biorxiv.org/content/10.64898/2026.03.11.711079v1