Research in the Campbell Lab is focused on understanding cell communication on a molecular level. In order to respond to their environment, cells must be able to relay external signals into the cell (outside-in signaling) as well as transmit internal signals out (inside-out signaling). Integrins are a diverse family of membrane proteins that signal bidirectionally and are critical for accurate cell communication. These uniquely shaped and versatile receptors undergo large-scale conformational changes to relay signals across the cell membrane. The integrin family is made up of 24 different — though structurally related — proteins that have varying tissue distributions and functions, allowing them to take part in a vast array of cellular processes. Integrin dysregulation has been associated with a myriad of pathologies including autoimmune, cardiac, pulmonary and blood diseases as well as cancer and infectious diseases; however few approved therapeutics exist. We utilize a broad range of methods in structural biology, biophysics, protein engineering, biochemistry, and cell biology to shed light on the molecular details of integrin signaling. Concurrently, we are developing cryo-electron microscopy methods—both computational and biochemical—to boost structural detail and more thoroughly characterize the dynamics of integrin receptors.
Integrin αvβ8 plays important roles in fibroinflammatory processes and anti-tumor immunity. In particular, the immunosuppressive function of regulatory T cells, which is a major mechanism of tumor immune evasion, is determined by αvβ8-mediated activation of transforming growth factor (TGF-β). Although TGF-β has therapeutic potential, its ubiquitous expression has hindered drug development due to off-target effects. Targeting αvβ8 integrin-mediated TGF-β activation provides a specific therapeutic strategy without the accompanying toxicities.
Integrin activation. Using cryoEM we defined the architecture and dynamics of avb8 integrin in extended conformations, suggesting the flexibility contributes to a ligand surveillance mechanism. We showed this conformation is consistent with αvβ3 integrin highlighting its importance towards establishing a general integrin activation mechanism at a molecular level.
TGF-β activation. We revealed how integrin αvβ8 recognizes latent TGF-β to understand the underpinnings of integrin binding specificity and TGF-β activation. This led us to find that integrin αvβ8 can activate TGF-β even when TGF-β remains within the confines of its prodomain. This structural insight and unexpected activation mechanism provide a rational basis for improved therapeutic strategies to inhibit αvβ8-mediated TGF-β activation.
Over the past decade, cryo electron microscopy (cryoEM) has been established as a leading method for structural determination. Macromolecules ranging from small proteins to large multi-component assemblies can now be visualized in atomic detail using this technique. CryoEM enables studying protein dynamics (which is an integral part of their function), receptor/drug interactions, and macromolecular complexes in their native cellular environments. Ongoing method development bringing in collaborative insight from computer vision, chemistry, and material science will continue to propel the limits of cryoEM forward.
Motion Correction. Microscope stage instabilities and irradiation with a high energy-electron beam cause proteins to move while in suspended vitreous ice. This blurs and significantly compromises image quality. We showed that by recording multi-frame movies instead of a single static image, motion can be computationally 'corrected' leading to significant gains in resolution and interpretability.
Image Denoising. CryoEM images have an extremely low signal to noise ratio making accurate identification and alignment of individual particles a challenge. We can ‘denoise’ images using a convolutional neural network algorithm trained on a subset of each specific dataset. This greatly boosts the image contrast and enables us to more rigorously process and interpret data.
Sample Preparation. For accurate structural determination using single-particle cryoEM, a comprehensive range of protein views is necessary. Traditional methods of sample preparation oftentimes yield a limited subset of projection orientations, limiting our ability to interpret structures. By functionalizing graphene oxide and using it as a substrate, the range of orientations can be enriched, resulting in better resolved cryoEM maps.