Marsupials and Multi-Omics: Establishing New Comparative Models of Neural Crest Patterning and Craniofacial Development

Abstract

Studies across vertebrates have revealed significant insights into the processes that drive craniofacial morphogenesis, yet we still know little about how distinct facial morphologies are patterned during development. Studies largely point to evolution in GRNs of cranial progenitor cell types such as neural crest cells, as the major driver underlying adaptive cranial shapes. However, this hypothesis requires further validation, particularly within suitable models amenable to manipulation. By utilizing comparative models between related species, we can begin to disentangle complex developmental systems and identify the origin of species-specific patterning. Mammals present excellent evolutionary examples to scrutinize how these differences arise, as sister clades of eutherians and marsupials possess suitable divergence times, conserved cranial anatomies, modular evolutionary patterns, and distinct developmental heterochrony in their NCC behaviours and craniofacial patterning. In this review, I lend perspectives into the current state of mammalian craniofacial biology and discuss the importance of establishing a new marsupial model, the fat-tailed dunnart, for comparative research. Through detailed comparisons with the mouse, we can begin to decipher mammalian conserved, and species-specific processes and their contribution to craniofacial patterning and shape disparity. Recent advances in single-cell multi-omics allow high-resolution investigations into the cellular and molecular basis of key developmental processes. As such, I discuss how comparative evolutionary application of these tools can provide detailed insights into complex cellular behaviours and expression dynamics underlying adaptive craniofacial evolution. Though in its infancy, the field of "comparative evo-devo-omics" presents unparalleled opportunities to precisely uncover how phenotypic differences arise during development.

Keywords: GRN; NCC; constraint; evolution; heterochrony; mammal; skull.

FIGURE 1

Neural crest and craniofacial development between therian mammals. Craniofacial heterochrony between therian mammals arises from altered neural crest cell behaviours. (A) In placental mammals, the neural crest forms in the neural folds and delaminates from the neural tube to migrate throughout the embryo. In marsupials however, the neural crest forms and delaminates from the neural plate border, leading to accelerated migration in the early embryo—redrawn from (Martik and Bronner, 2021). (B) Neural crest migration pathways are shared between therian mammals, though are accelerated in marsupials relative to developmental stage. Note marsupials display rapid development of the facial complex and forelimbs, while the CNS and hindlimbs are rudimentary. (C) The facial prominences of newborn marsupials resemble those observed in an embryonic mouse (credit FaceBase.org) (Samuels, B. D., 2020). (D) Comparative images of newborn mouse and dunnart, demonstrating the altriciality of the dunnart at birth. The adult dunnart superficially resembles a mouse. Image credits: dunnart newborn—Laura Cook; Dunnart Adult—David Paul—Museums Victoria; mouse pup—created with BioRender.com. Abbreviations: fl, forelimb; FNP, frontonasal process; hl, hindlimb; ht, heart; md, mandibular process; mx, maxillary process; pa, pharyngeal arch; pm, paraxial (head) mesoderm.

FIGURE 2

Workflow for comparative craniofacial single-cell multiomics. (A) Single cells can be isolated from dunnart and mouse NCCs and developing craniofacial prominences for multiplexed isolation and sequencing of RNA and open chromatin. Facial tissue sections can be processed for in situ spatial transcriptomics or MERFISH (Xia et al., 2019). Single-cell RNA and ATAC-seq data can be readily integrated using pipelines such as Seurat (Butler et al., 2018). (B) Dunnart and mouse datasets can be individually clustered or integrated to generate an atlas of homologous cell type populations—adapted from (Stuart et al., 2019). Cell transcriptomic and epigenetic profiles can be further mapped back to their spatial organization in the embryo. (C) The combination of these methods allows sophisticated downstream workflows to examine differential expression between species-specific clusters (pseudobulk) or cell lineage differentiation (pseudotime), inference of cell-cell signalling relationships (Jin et al., 2021), or construction of GRNs and species-specific patterns of gene regulation (Martik and Bronner, 2021).