During tissue morphogenesis, cells must precisely coordinate their functions over distance by communicating with secreted paracrine signaling proteins, such as the fibroblast growth factors (FGFs). The spatiotemporal regulations of FGF signaling are critical for development and homeostasis in all organisms. However, the mechanisms that control the location, timing, and level of FGF secretion and ensure its target-specific delivery over distance are poorly understood. This thesis has addressed this fundamental question by using a Drosophila FGF Branchless (Bnl) during tracheal development. Bnl is produced in a small group of wing-disc cells. Bnl regulates the development of a disc-associated tracheal branch, the air-sac-primordium (ASP), a vertebrate lung bud analog. Although FGFs are secreted diffusible signals, previous work has shown that the inter-organ dispersion of Bnl from the source to the ASP is highly polarized, target-specific, and is mediated over distance via long polarized actin-based filopodia, named cytonemes. The contact-dependent cytoneme-mediated Bnl delivery is programmed by the glycosylphosphatidylinositol (GPI) anchoring of Bnl. A GPI anchor tethers Bnl to the outer leaflet of the source cell membrane and inhibits free/random Bnl secretion. Surprisingly, GPI anchoring also facilitates Bnl’s release, but only through the cytoneme contact sites, thereby activating MAPK signaling exclusively in those recipient cells that establish contact with its source. However, the molecular mechanisms that might spatiotemporally activate the release of the membrane-anchored Bnl at the cytoneme contact sites remained unknown. This work addresses this knowledge gap by using a combination of cell biological, biochemical, and genetic analyses. Visualization of the distribution of a fluorescently-tagged Bnl revealed that a proteolytic cleavage within the Bnl backbone is essential for its target-specific release. Phenotypic analyses of an uncleavable Bnl mutant generated by genome editing suggested that the proteolytic event that releases Bnl is essential for Bnl signaling and morphogenesis. An in vitro cell-culture-based screening for candidate sheddases identified five membrane-localized enzymes that specifically shed Bnl from the cell surface. These enzymes include two matrix metalloproteinases, MMP1 and MMP2, a pro-protein convertase, Furin1, and two ADAM family metalloproteases, Meltrin and Tace. An inherently disordered region within the Bnl backbone was identified as the substrate for the proteolytic cleavage. Serial mutagenesis in this region revealed four minimal zones required for enzymes to act on Bnl. Interestingly, one of the minimal regions at the juxta-membrane position harbors the Bnl cleavage site, and the other three zones apparently play regulatory roles. Further work on MMPs using mosaic gain-of-function analyses confirmed MMP-mediated Bnl shedding in vivo. However, MMPs are not expressed in the Bnl source. Consequently, Bnl is not freely released. In contrast, MMPs are expressed in the ASP and are localized on the recipient ASP cytonemes. in vitro cell culture experiments that reconstituted contact-dependent Btl-Bnl interactions between Bnl-source and recipient cells confirmed that MMPs are selectively recruited by the recipient cells to the signaling contact sites and activate Bnl shedding in trans. These results explained why and how Bnl is released only via the cytoneme contact sites. Notably, MMP1 and MMP2 are GPI-anchored proteins but are also catalytically active in secreted soluble forms. Therefore, how is Bnl prevented from non-specific shedding by the extracellular soluble MMPs? in vitro and in vivo experiments using GPI-modified and non-GPI-modified MMPs revealed that only the membrane-tethered MMPs that are retained on the cell surface could efficiently shed cell surface Bnl. These findings supported a working model suggesting that recipient ASP cells extend receptor-containing cytonemes and establish contact with the Bnl-source via the receptor-ligand interaction. Due to the GPI-anchoring, MMPs are presented on the surface of ASP cytonemes and are delivered via these cytonemes to their contact sites, leading to the contact-dependent target-specific Bnl trans-shedding. These results provide new insights into how contact-mediated signal dispersion via cytonemes can be modulated by cell surface sheddases. Additional work on Furin1 and ADAM proteases showed that Tace and Meltrin shed Bnl in a cell-non-autonomous manner, like MMPs. In contrast, Furin1 is required only in the source cells for Bnl shedding. How and under what contexts Furin1 and ADAM proteases are activated to shed Bnl needs to be explored in the future. Collectively, this work characterized a novel enzymatic Bnl shedding mechanism and provided insights into how Bnl sheddases might be controlled in space and time to ensure cytoneme-mediated Bnl exchange.