Discoveries go full circle in the hunt for an elusive protein’s function

From the Strong Lab, Basic Sciences and Vaccine and Infectious Disease Divisions, and the Olson Lab, Seattle Children’s Research Institute

Down a nondescript hallway in the Weintraub building on the Fred Hutch campus, there is an ordinary-looking side room filled with sparkling crystals.

Before you go looking for a ski mask and hatch plans for the heist of the century, let me tell you that these crystals are microscopic, and that they’re worth virtually nothing to all but a very small group of people on this planet. Dr. Peter Rupert, a structural biologist in the Strong Laboratory in the Fred Hutch Basic Sciences Division, is one of those few people. On what I like to imagine was a dark and stormy Seattle night some time ago, Dr. Rupert sat in that small room, fervently examining small, buffer-filled trays for any signs of crystal growth. Unlike the crystals you may be picturing, the ones that Rupert was hunting for are made of proteins—in this case, they would be made of one particular protein called mesothelin. Why was he so intent on making crystals of mesothelin? Because—if he could obtain sufficiently large and stable crystals—he could send them to a particle accelerator at Lawrence Berkeley National Lab called the Advanced Light Source (ALS), where the crystals would be bombarded with precisely calibrated X-rays to produce a diffraction pattern. This diffraction pattern would be unique to mesothelin, and Dr. Rupert could analyze it to obtain the structure of the protein in what amounts to a molecular yearbook photo fit for a sci-fi film.

Let’s back up a second—what is mesothelin, and why do we care about its structure? “Mesothelin is a truly enigmatic little protein,” begins Dr. Roland Strong, who’s been thinking about protein structures for longer than I’ve been alive. “It’s a cell surface protein normally expressed in a tissue called the mesothelium (which is also where the cancer mesothelioma gets its name), but it’s overexpressed on the surface of a variety of solid tumors. While we can detect it all over the place, we still have little idea what this protein does or why it’s on these tumors.” Rupert and Strong thus sought to address this knowledge gap by applying the time-honored credo of structural biology: form dictates function. If they figured out what mesothelin looked like, they could compare its structure to other proteins of known function. There was one problem, however: they couldn’t get the dang thing to form crystals. As Dr. Strong notes, “crystallography (the process of using a protein crystal to derive its structure) is an exact science. Crystallization (imagine him waving his arms in emphasis), however, is a black box.” Indeed, for a variety of biophysical reasons incompletely understood, some proteins are simply less apt to form crystals—mesothelin was one of these proteins.

At the same time, Dr. Colin Correnti—a former graduate student of Dr. Strong’s and co-founder of Link Immunotherapeutics down the road from the Hutch—was interested in a different aspect of mesothelin. Correnti was intent on designing mesothelin-binding antibodies (which, surprisingly, doesn’t require knowing its structure) to leverage for cancer immunotherapy. If he could produce effective antibodies against this cancer-specific surface protein, perhaps those antibodies could be used to design so-called bispecific engagers: biologics which bind both a tumor antigen and a T-cell antigen to bring the two into close proximity for therapeutic effect (see here for a distinct but related approach). Two mesothelin antibodies already existed, but they performed poorly in clinical trials—Dr. Correnti sought to produce novel mesothelin antibodies that bound the protein closer to the cell membrane (which the team reasoned would improve its performance due to several technical reasons described in the study). Using a clever trick involving dosing mice with a chimeric mesothelin protein consisting of a human-derived (immunogenic) membrane proximal region fused to a mouse-derived (non-immunogenic) membrane-distal region, Correnti was able to produce such an antibody, which the team named 1A12. Novel antibody in hand, Dr. Correnti used it to design two different bispecific reagents targeting mesothelin and CD3 (A conserved T-cell antigen). Remarkably, the team showed that these reagents bound and activated T-cells in vitro and were more potent than analogs using the previously discovered mesothelin antibodies!

And now, we go full circle: one ‘trick of the trade’ available to structural biologists trying to crystallize a protein is to add antibodies, which are thought to stabilize its structure and increase its proclivity to form crystals. Dr. Rupert wondered if he could use 1A12 (and the two previous mesothelin antibodies) to get his protein structure—he mixed purified mesothelin with the antibodies, and after many rounds of optimizing the conditions… a crystal! Some X-rays and computation later, and the team had the most complete structures of mesothelin (comprising its entire extracellular portion, or ectodomain) ever obtained. When Rupert and colleagues went to compare mesothelin’s structure to an exhaustive database of known protein structures, they were astonished to find a completely novel fold: there was only one weak match in the database, belonging to an intracellular yeast protein involved in DNA repair. “For us, this was exhilarating and frustrating—we had discovered a fold not yet seen in nature, but that also meant that the structure’s utility in inferring function was limited,” Strong noted.

an image of a crystal as seen under a microscope, with a computer rendering of a protein strcture superimposed on top
A depiction of the 1A12 antibody (blue and white, left) interacting with the mesothelin ectodomain (green, right), superimposed on a photo of the protein crystals used to solve these structures. Image courtesy of Dr. Peter Rupert.

Despite the uniqueness of the structure leaving mesothelin’s function still shrouded in mystery, the team is excited to use their structures to better understanding mesothelin’s interactions with its binding partners and are poised to continue exploiting mesothelin to design better cancer immunotherapies. It’s a tale full of suspense and unexpected findings set against the backdrop of an ever-evolving mystery—or, as we come to know, just another day in the lab.

The spotlighted research was funded by the Washington Research Foundation, St. Baldrick’s Foundation, the Children’s Oncology Group Foundation, and the Fred Hutchinson Cancer Research Center.

Fred Hutch/University of Washington/Seattle Children’s Cancer Consortium members Drs. Roland Strong, James Olson and Soheil Meshinchi contributed to this study.

Lin, I., Rupert, P. B., Pilat, K., Ruff, R. O., Friend, D. J., Chan, M. K., Clarke, M., Hoffstrom, B. G., Carter, J., Meshinchi, S., Bandaranayake, A. D., Mehlin, C., Olson, J. M., Strong, R. K., & Correnti, C. E. (2023). Novel mesothelin antibodies enable crystallography of the intact mesothelin ectodomain and engineering of potent, T cell-engaging bispecific therapeutics. Frontiers in Drug Discovery, 3, 1216516.