Introduction: Traditionally, peripheral sensory neurons were seen as the sole transducers of external stimuli. However, the contribution of non-neuronal cells is increasingly recognized. In the epidermis, keratinocytes act as sensors and transmitters of somatosensory and nociceptive input. They form intimate connections with afferent intraepidermal nerve fibers at the neuro-cutaneous unit. A conserved mechanism of nerve fiber ensheathment was described up to human skin. Via release of signaling molecules like ATP, keratinocytes can activate nerve endings, induce action potentials, and thereby tune peripheral input. Material & Methods: We obtained primary human fibroblasts and keratinocytes from diagnostic skin punch biopsies. Induced pluripotent stem cell-derived sensory neurons were generated from fibroblasts. After four-week maturation, neurons were transferred into a two-comparted well chamber system and primary keratinocytes seeded in the opposite well. We tracked the formation of neurite-keratinocyte contacts via live-imaging and subsequently investigated these interactions by confocal and super-resolution fluorescence microscopy. Results: Upon barrier removal, neurites form the neuronal compartment actively grew towards keratinocytes and formed contacts within few days. Via discriminative labeling of keratinocyte and neuronal membranes, we identified several types of interactions. Neurites tunneled through keratinocyte cytoplasm and formed bouton-like structures in contact with keratinocytes. Furthermore, connexin 43, a hemichannel and gap junction forming pore protein expressed in keratinocytes, accumulated at neurite-keratinocyte contact sites. Discussion: We successfully established a fully human co-culture model of sensory neurons and keratinocytes maintaining viability over several days. Our 2D model reduces the multilayered complexity of native human skin, yet conserves ensheathment of sensory neurites as a feature of the NCU and showed Cx43 plaques as a potential signaling route for NCU communication. This approach inherits a high potential for translational investigations, bypassing challenges of heterologous animal co-culture models and facilitating the integration of patient derived neurons and keratinocytes.