Despite the increasing incidence of kidney-related diseases, we are still far from understanding the underlying mechanisms of these diseases and their progression. This lack of understanding is partly because of a poor replication of the diseases in vitro, limited to planar culture. Advancing towards three-dimensional models, hereby we propose coaxial printing to obtain microfibers containing a helical hollow microchannel. These recapitulate the architecture of the proximal tubule (PT), an important nephron segment often affected in kidney disorders. A stable gelatin/alginate-based ink was formulated to allow printability while maintaining structural properties. Fine-tuning of the composition, printing temperature and extrusion rate allowed for optimal ink viscosity that led to coiling of the microfiber’s inner channel. The printed microfibers exhibited prolonged structural stability (42 days) and cytocompatibility in culture. Healthy conditionally immortalized PT epithelial cells and a knockout cell model for cystinosis (CTNS-/-) were seeded to mimic two genotypes of PT. Upon culturing for 14 days, engineered PT showed homogenous cytoskeleton organization as indicated by staining for filamentous actin, barrier-formation and polarization with apical marker α-tubulin and basolateral marker Na+/K+-ATPase. Cell viability was slightly decreased upon prolonged culturing for 14 days, which was more pronounced in CTNS-/- microfibers. Finally, CTNS-/- cells showed reduced apical transport activity in the microfibers compared to healthy PT epithelial cells when looking at breast cancer resistance protein and multidrug resistance-associated protein 4. Engineered PT incorporated in a custom-designed microfluidic chip allowed to assess leak-tightness of the epithelium, which appeared less tight in CTNS-/- PT compared to healthy PT, in agreement with its in vivo phenotype. While we are still on the verge of patient-oriented medicine, this system holds great promise for further research in establishing advanced in vitro disease models.