E. Farr, E. Kritikaki, M. Chroscik, C. Admane, E. Graves, C. Tudor, H. M. Chan, J. Boccacino, J. McWilliam, F. Torabi, K. Chakala, D. Basurto-Lozada, T. Li, A. Binkevich, A. Predeus, M. Prete, M. Panamarova, D. Adao, K. Evans, K. Stewart, L. Steele, E. Winheim, N. H. Gopee, E. Stephenson, M. Patel, C. Hale, L. Gambardella, B. Harpur, C. Smith, D. Horsfall, V. Shanmugiah, L. Parts, D. J. Adams, M. Kasper, A. Dugourd, J. Saez-Rodriguez, A. R. Foster, M. Haniffa
Single-cell technologies have generated cell censuses of tissues, however, how tissue geometry reflects functional needs remains poorly characterized. The human pilosebaceous unit offers a tractable model, a prenatally-formed complex mini-organ combining hair and sebum production with a stem cell reservoir. Using histomorphology, spatial transcriptomics, and single-cell multiomics on the same human prenatal scalp skin samples (8-19 post-conception weeks), integrated and analyzed using machine learning approaches, we built a spatiotemporal map of pilosebaceous unit development. We demonstrate that epithelial-mesenchymal interactions coordinate cellular fate and organogenesis, using an in vitro hair-bearing skin organoid model to validate this tissue-patterning. In addition, we show sebaceous gland developmental programmes are overcome during tumor formation. Our large-scale multi-modal analysis provides a unique framework for understanding form and function of tissues with applications in tissue engineering and pathology.