Ror2 signaling regulates Golgi structure and transport through IFT20 for tumor invasiveness

Picture invasion not as a mad dash, but as logistics. A cancer cell that breaks away and bores through tissue has to reorganize its internal shipping department so that degradative enzymes arrive precisely where the cell is pushing forward. That push happens at invadopodia, those tiny, actin-rich drilling platforms that concentrate matrix metalloproteinases to chew through the extracellular matrix. At the center of this logistics trick sits the Golgi, which has to form a ribbon-like structure, sprout its own microtubules, and funnel cargo to the front. As Nishita and colleagues show, a receptor called Ror2 rewires all of that. The twist is who Ror2 recruits to do it: IFT20, a protein famous for cilia, pressed into a very non-ciliary job. Here's the setup. In the osteosarcoma line SaOS2, which under these conditions lacks primary cilia, knocking down Ror2 drops IFT20 expression hard. The authors saw IFT20 messenger RNA fall to about 40 percent of control by microarray and quantitative polymerase chain reaction, a result that matched the protein. Silencing Wnt5a didn't budge IFT20, hinting that this arm of regulation sits at least partly outside the usual Wnt5a and Ror2 relationship. They verified the non-ciliated context directly: SaOS2 didn't make cilia under their conditions, while mesenchymal stem cells did. So whatever IFT20 is doing here, it's not about building a cilium. Does this matter for behavior? Yes. When they suppressed Ror2 or IFT20, invasion through Matrigel fell, with a p-value below 0.01. Invadopodia formation shrank too, scored as F-actin puncta coinciding with gelatin degradation, and that effect was even stronger statistically, with a p-value below 0.001. The smoking gun was rescue: add back an siRNA-resistant IFT20, and invadopodia and invasion return—even in cells where Ror2 was knocked down. That places IFT20 downstream of Ror2 in the path that licenses invasion. Where does IFT20 work in the cell to do this? It parks on the cis side of the Golgi. Using GM130, a cis-Golgi marker, and Golgin-97 for the trans side, the team showed that IFT20 colocalizes more with GM130 and sits close to invadopodia marked by cortactin. When Ror2 or IFT20 was silenced, the elegant Golgi ribbon dispersed into fragments. You could see the polarity fray too: in edge cells asked to invade across a boundary, roughly 70 percent of controls pointed their centrosomes toward the invasion front. With Ror2 or IFT20 knockdown, that dropped to about 40 to 50 percent. Restoring IFT20 reassembled the ribbon and put the centrosome back on target. This wasn't a one-cell-line quirk. In BT549 and U2OS cells, which also don't build cilia under these conditions, silencing Ror2 or IFT20 reduced invasion and scattered the Golgi, and the siRNA-resistant IFT20 put things back together. By contrast, mesenchymal stem cells, ciliated and in a very different context, didn't show Golgi dispersion under the same knockdowns. That contrast reinforces the core point: this is a non-ciliary role for IFT20, revealed in tumor cells that have let the cilium go. Nishita's team also separated this Golgi polarity axis from canonical Wnt signaling. In a TCF and LEF luciferase assay, Ror2 knockdown reduced canonical Wnt activity, as expected for this receptor. IFT20 knockdown didn't. So IFT20 is not just a generic effector of Ror2; it's carrying the Golgi-polarity torch while Ror2 does other things on the Wnt side. Mechanistically, the key is how the cis-Golgi scaffolds a microtubule-building platform. IFT20 physically associates with GM130 and AKAP450, the anchoring factor that seeds microtubules on Golgi membranes. Co-immunoprecipitation pulled down GM130 and AKAP450 with IFT20, and proximity ligation assays lit up when any two of these were probed together—signals that vanished when IFT20 was silenced. In control cells, AKAP450 and GM130 huddle at the Golgi; knock down Ror2 or IFT20 and the partnership loosens. Put back siRNA-resistant IFT20 and the pair reunites. Notably, AKAP450's position at the centrosome was unchanged by these perturbations, pointing to a Golgi-specific disruption. Take AKAP450 itself out of the picture and you get the same phenotypes: the ribbon breaks and invadopodia formation falters, but IFT20's expression and Golgi localization stay intact. That puts AKAP450 functionally downstream of IFT20 for ribbon integrity. Another nice nuance emerged when they compared cis- and trans-Golgi markers. Even when the ribbon fell apart in Ror2- or IFT20-silenced cells, GM130 and Golgin-97 still traveled together in mini-stacks. The stacks weren't gone; it was their higher-order assembly into a continuous ribbon that failed. In a time-course classification of Golgi morphologies, the transition from intermediate clusters to a mature perinuclear ribbon was selectively stalled, with that step showing a significant impairment with a p-value below 0.001. Microtubule dynamics explained why. After dissolving microtubules with nocodazole and letting them regrow, the group scored three kinds of fibers: Golgi-anchored, centrosomal, and everything else. Control cells readily nucleated microtubules from Golgi fragments. Ror2 or IFT20 knockdown selectively reduced those Golgi-derived microtubules, leaving the centrosomal and non-Golgi pools essentially unchanged. The drop in Golgi-anchored fibers was robust and highly significant statistically, and it was reversed when IFT20 was restored. That selective deficit maps cleanly onto the idea that IFT20 stabilizes the GM130 and AKAP450 complex to let the Golgi make its own microtubules and keep the ribbon intact. Trafficking through the Golgi was the other half of the story. Think of cargo flow in two lanes: forward, from the endoplasmic reticulum to the surface, and backward, from the Golgi back to the endoplasmic reticulum. IFT20 pushes the forward lane faster without gumming up the return. Using the temperature-sensitive VSVG cargo, the team did the classic 15-degree block to pile VSVG at the cis-Golgi, then warmed cells to 32 degrees to let it run toward the trans-Golgi network. With Ror2 or IFT20 silenced, that cis-to-trans transit lagged; re-expressing IFT20 fixed it. They saw the same pattern in a broader endoplasmic reticulum-to-surface readout: slower delivery when IFT20 was down and rescue with siRNA-proof IFT20. What about the return lane? For retrograde traffic, they used a VSVG-KDELR chimera that cycles back to the endoplasmic reticulum. IFT20 knockdown didn't meaningfully affect that pathway. Endoplasmic reticulum-to-Golgi entry—before cargo even reaches the Golgi stacks—was largely intact too. So the choke point sits inside the Golgi, on the forward path. That's exactly where you'd expect a ribbon-dependent, microtubule-guided conveyor to run faster or slower. Cargo matters as well. The authors probed a chimera that carries the cytoplasmic tail of MT1-MMP, a collagen-degrading protease that literally arms invadopodia. Intra-Golgi transport of this VSVG–MT1-MMP cargo slowed when IFT20 was suppressed. Put differently, the very packets that need to be delivered to the drilling platform are the ones whose journey is most sensitive to the state of the Golgi and its microtubule tracks. All of this adds up to a clean causal chain. Ror2 signaling elevates IFT20 in tumor cells that have shed their cilia. IFT20 enriches at the cis-Golgi and fortifies a GM130 and AKAP450 hub that lets the Golgi nucleate its own microtubules. Those microtubules stitch fragmented stacks into a ribbon and aim the cell's secretory apparatus toward the front. The ribbon, in turn, speeds anterograde traffic through the Golgi—especially the kind of cargo that builds and feeds invadopodia—while leaving retrograde, COPI-mediated return traffic alone. The end result is polarized secretion and a cell that can degrade matrix and move with purpose. There are a few anchor numbers to keep in your head as you connect the dots. Ror2 knockdown pushes IFT20 messenger RNA to about 40 percent of normal; invasion and invadopodia formation fall significantly when either gene is silenced, and those functional hits are rescued by siRNA-resistant IFT20. Centrosome polarization at the leading edge drops from around 70 percent in controls to roughly 40 to 50 percent with Ror2 or IFT20 knockdown, and Golgi-anchored microtubule nucleation is selectively and significantly reduced. Those are the quantitative fingerprints of a trafficking system reprogrammed for or away from invasion. One more clarification before we look ahead. Ror2 is a player in multiple Wnt pathways, and it can dampen canonical Wnt signaling. In these experiments, that canonical arm is separable. The TCF and LEF reporter fell with Ror2 knockdown but didn't change with IFT20 knockdown. So the Golgi-centric mechanism runs in parallel to, not through, Ror2's canonical Wnt effects. That separation matters if you're thinking about targeting one axis without blindly hitting the other. Step back, and the picture is strikingly specific: a cilium protein, repurposed at the Golgi in non-ciliated tumor cells, becomes the linchpin of invasive polarity. The Golgi isn't just a mailroom here; it's the compass and the highway. Ror2, by dialing up IFT20, flips the switch from a diffuse, indecisive network to a focused, front-directed machine. Where does that leave us? Two quick thoughts, clearly marked as future-facing. First, the context dependence is a feature, not a bug. The fact that ciliated mesenchymal stem cells didn't show the same Golgi fragility when IFT20 was knocked down suggests a therapeutic window: you might be able to disrupt this Ror2 and IFT20 and Golgi axis in tumors that have lost their cilia without derailing healthy, ciliated tissues. Second, the mechanism gives you handles. You can imagine screening for small molecules that destabilize the GM130 and AKAP450 interaction at the cis-Golgi or that selectively slow Golgi-derived microtubule nucleation, then asking whether they phenocopy the clean, anterograde-specific slowdown and the loss of invadopodia seen here. But let's end where the data are strongest. As Nishita and colleagues showed, invasion is a trafficking problem solved by polarity. Ror2 solves it by recruiting IFT20 to the Golgi, building a ribbon, laying down microtubule tracks, and opening the fast lane for the cargo that equips a cancer cell to cut its way forward. If you want to stop the march, you could do worse than to start at that ribbon.