Covalently Immobilizing Interferon-γ Drives Filopodia Production through Specific Receptor-Ligand Interactions Independently of Canonical Downstream Signaling

Abstract

Immobilizing a signaling protein to guide cell behavior has been employed in a wide variety of studies. This approach draws inspiration from biology, where specific, affinity-based interactions between membrane receptors and immobilized proteins in the extracellular matrix guide many developmental and homeostatic processes. Synthetic immobilization approaches, however, do not necessarily recapitulate the in vivo signaling system and potentially lead to artificial receptor-ligand interactions. To investigate the effects of one example of engineered receptor-ligand interactions, we focus on the immobilization of interferon-γ (IFN-γ), which has been used to drive differentiation of neural stem cells (NSCs). To isolate the effect of ligand immobilization, we transfected Cos-7 cells with only interferon-γ receptor 1 (IFNγR1), not IFNγR2, so that the cells could bind IFN-γ but were incapable of canonical signal transduction. We then exposed the cells to surfaces containing covalently immobilized IFN-γ and studied membrane morphology, receptor-ligand dynamics, and receptor activation. We found that exposing cells to immobilized but not soluble IFN-γ drove the formation of filopodia in both NSCs and Cos-7, showing that covalently immobilizing IFN-γ is enough to affect cell behavior, independently of canonical downstream signaling. Overall, this work suggests that synthetic growth factor immobilization can influence cell morphology beyond enhancing canonical cell responses through the prolonged signaling duration or spatial patterning enabled by protein immobilization. This suggests that differentiation of NSCs could be driven by canonical and non-canonical pathways when IFN-γ is covalently immobilized. This finding has broad implications for bioengineering approaches to guide cell behavior, as one ligand has the potential to impact multiple pathways even when cells lack the canonical signal transduction machinery.

Figure 1.

NSCs produce filopodia in response to immobilized IFN-γ. A: Representative TIRF images of live primary rat NSCs stained with DiO-C18 (a membrane incorporating dye) following incubation on control (no IFN-γ) surfaces (n=100) or immobilized IFN-γ surfaces (n=160). NSCs on immobilized IFN-γ produce a large amount of filopodia compared to control surface. See Figure S1 for additional images. Scale bars represent 10 μm. B: Close-up of cell edge showing filopodia in detail. Filopodia on control surfaces are sparse and incubation on immobilized IFN-γ surfaces produces overlapping structures.

Figure 2:

IFNγR1-expressing Cos-7 cells generate filopodia in response to IFN-γ immobilization. A: TIRF imaging of live Cos-7 transiently transfected with IFNγR1-GFPSpark following incubation on control (no IFN-γ) surfaces, with soluble ligand, or on immobilized IFN-γ surfaces. Exposure to immobilized IFN-γ induces a pronounced increase in filopodia production. Scale bars represent 10 μm. B: Close-up of cell edge on a surface containing immobilized IFN-γ, showing filopodia in detail. Scale bar represents 10 μm. C: Number of filopodia on individual cells quantified using Filoquant^, and represented by a bee swarm plot and box and whisker plot overlay, the red line indicates the median of the data. Exposure to immobilized IFN-γ yields a significantly larger number of filopodia per cell. **** represents p < 0.0001 as determined by Student’s t-test in pairwise comparisons with the control group. D: Cell edge length (in μm) of individual cells quantified and represented in the same manner as C. Both soluble and immobilized IFN-γ produce a longer cell edge than cells on control surfaces, with immobilized IFNγ being the highest. ** represents p < 0.01, **** represents p < 0.0001 as determined by Student’s t-test in pairwise comparisons with the control group. The grey numbers above each plot indicate the number of individual cells analyzed. Additional data can be found in Figure S2.

Figure 3:

Immobilizing IFN-γ causes IFNγR1-expressing Cos-7 to retain their individual filopodia for longer time scales. TIRF time-lapse imaging of Cos-7 expressing IFNγR1-GFPSpark grown on control (no IFN-γ), adsorbed IFN-γ, or immobilized IFN-γ surfaces. White arrows show filopodia of interest, which form and retract in the adsorbed control, but are retained with immobilization of IFN-γ. Each time-lapse is obtained at the same magnification and cropped to the region of interest, with scale bars in each panel representing 10 μm.

Figure 4:

IFNγR1 and azide-tagged IFN-γ interact specifically. A: Representative Cos-7 cell epifluorescence when transiently transfected with IFNγR1-GFPSpark and stimulated with soluble IFN-γ-rhodamine. Scale bars represent 10 μm. B: Average diffusion coefficient of IFNγR1 for control and soluble IFN-γ-stimulated Cos-7 cells. During stimulation, diffusion decreases, indicating larger molecular size and complex formation. The grey numbers above each plot represent the number of individual cells analyzed. **** represents p < 0.0001 as determined by Student’s t-test. C: Typical auto- and cross-correlation curves for azide-tagged IFN-γ-stimulated cells. Average $f\text{c}$ value for IFNγR1-soluble IFN-γ cross-correlation is 0.34 (n=31) indicating formation of receptor-ligand complex. See Figure S3 for additional correlation curves.

Figure 5:

Exposure of IFNγR1-expressing Cos-7 to IFN-γ does not activate canonical signaling pathways. Confocal imaging at the basal surface of Cos-7 cells (left) and NSCs (right) following incubation on control (no IFN-γ) surfaces, with soluble ligand, or on immobilized IFN-γ surfaces. Representative Cos-7 cells show no apparent change in STAT1p levels based on treatment. Typical stimulated NSCs show STAT1p expression has nuclear localization (Hoechst 33342, yellow) compared to control cells, as expected. Scale bars represent 10 μm.

Figure 6:

Immobilized IFN-γ does not generate focal adhesions. After exposing IFNγR1-expressing Cos-7 cells to IFN-γ in soluble and immobilized forms, there was no observed change in focal adhesion generation when compared to control cells, as seen in representative confocal images of the basal surface. Cells grown on collagen (bottom row) generated many FAs. White arrows indicate vinculin expression at the ends of actin stress fibers. Scale bars represent 10 μm.