Paper Review — verify claims with raw data

Concise appraisal

The preprint reports that GLI3 has two temporally distinct roles in neural crest development: an early, apparently repressor-biased role required for specification of premigratory NCCs (loss reduces PAX7, SOX10, FOXD3) and a later activator-biased role promoting cranial NCC differentiation into ectomesenchymal (skeletal) lineages (loss lowers Runx2, Sp7, Alpl and causes craniofacial hypoplasia) supported by CUT&RUN peaks near ectomesenchymal genes and H3K27ac enrichment at those peaks Pajanoja et al GLI3 dual role preprint

Main caveats: GLI3A versus GLI3R functional separation is not directly shown (authors acknowledge this), Wnt1-Cre timing may miss earliest specification windows, and cross-species extrapolation (chick/mouse/human) needs careful interpretation Pajanoja et al Limitations

Detailed paper analysis and critique

1. What the paper claims

• GLI3 is expressed in premigratory and cranial neural crest across chick, mouse and human in published and new datasets and by in situ/Western blotting GLI3 expression in NCCs
• Early GLI3 loss (morpholino in chick, siRNA in human NCCs) reduces NCC specifier markers (PAX7, SOX10, FOXD3, TFAP2A) — interpreted as GLI3 necessary for specification stage GLI3 knockdown reduces NCC specifiers
• Late conditional deletion of Gli2 and Gli3 in Wnt1 lineage (Gli2 f/f; Gli3 f/f; Wnt1-Cre) after specification impairs CNCC differentiation into ectomesenchymal lineages (reduced Runx2, Sp7, Alpl, Sox9; craniofacial hypoplasia), while non-ectomesenchymal derivatives increase relatively (Foxd3, Phox2b, Tubb3) Gli2/Gli3 conditional knockout phenotypes
• CUT&RUN for GLI3 (n=3) in E11.5 mandibular prominence identifies ~3895 peaks enriched at promoters, motif GBM GACCACC, and peaks align with H3K27ac (active mark) near ectomesenchymal genes (Runx2, Sox9), supporting direct regulatory roles during differentiation GLI3 CUTandRUN and active marks

2. Strengths

• Multi-model evidence: chick in ovo electroporation, human ES-derived NCCs (siRNA), and conditional mouse mutants provide cross-validation across species rather than reliance on one system Multi species approach
• Integration of epigenomic (CUT&RUN) with transcriptomics (bulk and scRNA-seq) strengthens inference about direct regulatory relationships (GLI3 peaks at active promoters of ectomesenchymal genes) Epigenomic and transcriptomic integration
• Data sharing: GEO accessions and GitHub scripts are provided for reproducibility of sequencing analyses (GSE301704, GSE221188, GSE162035, GSE125145, and GitHub repository links) Data availability and code

3. Key weaknesses, caveats and alternative interpretations

1. GLI3 isoform ambiguity — The paper does not definitively show which isoform (GLI3A activator versus GLI3R repressor) binds specific targets in CUT&RUN or mediates the early versus late roles. The authors explicitly note development of dual-tagged lines to separate GLI3A/GLI3R is future work; without isoform-specific ChIP/CUT&RUN, one cannot ascribe early specification roles to GLI3R versus late differentiation roles to GLI3A unequivocally GLI3 isoform limitation
2. Timing of conditional knockouts — Wnt1-Cre recombines around E8.0 and thus may not remove GLI3 during the earliest NCC induction window; some early phenotypes were assayed with morpholinos/siRNA but the conditional knockout negative results for specification may reflect late recombination rather than true lack of requirement. That weakens causal claims from the knockout alone Timing caveat Wnt1-Cre
3. Compensation and GLI family redundancy — Gli2 and Gli3 were deleted together in some experiments but partial redundancy and network compensation can complicate phenotype interpretation; morpholino/siRNA results risk off-target effects and may differ from genetic deletion phenotypes GLI redundancy caveat
4. Cross-species extrapolation — Patterns seen in chick/human in vitro and mouse in vivo are compelling but species differences in timing and pathway wiring (HH/WNT/BMP crosstalk) mean the dual-role model needs species-specific mechanistic verification rather than assuming full conservation Cross-species caution
5. Mechanistic integration with Wnt and Sufu literature — prior work shows Sufu integrates Hedgehog and Wnt (Xenopus Sufu binds Gli and beta-catenin and modulates cross-talk) and GLI proteins can both repress and activate depending on context; the new GLI3 dual-role model fits conceptually with Sufu-mediated crosstalk but requires explicit mechanistic links (cofactors, phosphorylation, cilia) in NCCs rather than inference Sufu integrates Hh and Wnt

4. Recommended experiments to strengthen conclusions

1. Isoform specific occupancy and function: generate GLI3 isoform-specific CUT&RUN/ChIP (GLI3A vs GLI3R) via epitope-tagged knockin (dual-tagged lines the authors plan) and test which isoform binds ectomesenchymal promoters and which binds early specification targets.
2. Temporal conditional deletion with earlier Cre drivers (e.g., Pax3-Cre earlier or inducible Cre with narrow time windows) to remove GLI3 before NCC induction and compare to morpholino/siRNA phenotypes to resolve timing discrepancies.
3. Rescue experiments: express GLI3R-only or GLI3A-only isoforms in GLI3-deficient background to test sufficiency for specification versus differentiation roles.
4. Assess Sufu/GLI3/Wnt interactions in NCCs directly: co-IP of GLI3 with SUFU, ZIC proteins, beta-catenin in NCCs, and reporter assays (Gli and TOP reporters) in NCCs to test pathway crosstalk in-context (building on Sufu literature) Sufu coIP and reporter assays

5. How convincing is the claim that GLI3 has a dual role?

Overall the evidence is suggestive and internally consistent: early knockdowns reduce specification markers and later conditional loss reduces ectomesenchymal gene programs and craniofacial skeleton; GLI3 occupancy at promoters of ectomesenchymal regulators and H3K27ac enrichment support a direct role in differentiation. But because isoform-specific binding/activity and precise temporal genetic loss data are incomplete, the mechanistic claim (GLI3R for specification and GLI3A for differentiation) remains provisional rather than proven Dual role evidence and caveats

6. Broader context and prior literature

GLI3 is a bifunctional transcription factor with well-established activator and repressor isoforms critical for limb and craniofacial development; mouse Gli3 mutants alter Fgf8 and apoptosis in neural tube and face (regulatory precedent) and sources show SUFU coordinates GLI-WNT cross-talk — the current paper extends these themes to neural crest specification/differentiation but needs to place isoform dynamics and ciliary regulation front-and-center as prior work demonstrates cilia and SUFU heavily influence GLI processing and crosstalk GLI3 regulates Fgf8 and apoptosisSufu mediates Hh-Wnt crosstalk

7. Data and reproducibility assessment

• Reproducibility is supported by deposit of sequencing data in GEO and links to analysis code/GitHub; methods use standard and current pipelines (STAR, DESeq2, MACS, HOMER, Seurat) which enables reanalysis Methods and data deposition
• Key reproducibility gap: absence of isoform-specific reagents and quantitative isoform ChIP/CUT&RUN makes the mechanistic claims harder to falsify without additional targeted experiments.

8. Minor technical/method points to check

• Inspect CUT&RUN peak annotation distribution and blacklist filtering to ensure promoter bias is not driven by artifact (ENCODE blacklist regions) — authors mention ChIPseeker/ChIP analysis but raw peak lists should be cross-checked CUTandRUN peak distribution
• Quantitative reporting: provide effect sizes and statistics for qPCR/IF reductions (fold-change means, SD/SEM, p-values, replicate counts) in main figures rather than summary statements to permit meta-analysis.

9. Overall judgement (concise)

The manuscript presents a well-executed, multi-modal data package supporting a plausible model where GLI3 contributes at two developmental windows to neural crest biology: early for specification and later for ectomesenchymal differentiation. The central experimental shortfalls are lack of isoform-resolved mechanistic data and the timing ambiguity of conditional deletions. With isoform-specific occupancy/function assays and narrower temporal genetics the conclusion could be elevated from plausible to convincing.

10. Useful references cited here

• Sufu Hh/Wnt integration mechanistic precedent Sufu integrates Hh and Wnt
• GLI3 developmental roles and Fgf8 regulation precedent GLI3 regulates Fgf8 and apoptosis
• Perspective on enhancer redundancy relevant to interpreting promoter/enhancer binding of GLI3 at developmental genes Shadow enhancers and robustness