ProAgio is a laboratory-designed protein that latches onto a receptor called integrin αvβ3 and pushes the cell carrying it into programmed cell death. Researchers at Georgia State University, led by molecular biologist Zhi-Ren Liu, described it in Nature Communications on 31 May 2016. In mice, it almost halted the growth of human prostate tumours and cut the number of blood vessels feeding them, without obvious harm to healthy organs. It has since moved into early human trials, but it is not an approved treatment, and nobody yet knows whether it works in people.
Here is what the protein does, why its target matters, what the evidence actually shows, and where the project stands today.
Why integrin αvβ3 interests cancer researchers
Integrins are pairs of proteins that sit across the cell membrane. They act a little like hands and ears at once: they grip the surrounding tissue and pass signals about that grip into the cell. Humans have many integrin pairs, each with a different job.
The αvβ3 pair is unusual because healthy adult tissue carries very little of it. It turns up mainly where something is growing or being remodelled: in the lining of new blood vessels, in certain immune cells called macrophages, and in osteoclasts, the cells that break down bone. It is also abundant on many tumour cells, especially those that spread, and on the new vessels that tumours recruit to feed themselves. Liu summed up the appeal simply: "This integrin pair, αVβ₃, is not expressed in high levels in normal tissue."
That pattern makes αvβ3 an attractive target. A drug that recognises it should, in theory, concentrate on diseased tissue and leave most healthy cells alone. Besides cancer, abnormal vessel growth is involved in eye diseases such as diabetic retinopathy and in some fibrotic and inflammatory conditions, which is why the Georgia State team described their protein as a possible tool for several diseases.
Why earlier αvβ3 drugs fell short
Drug makers have chased this receptor for decades. Most earlier attempts worked by blocking the spot where αvβ3 normally grabs its partners in the surrounding tissue, a site that recognises a short protein motif known as RGD. The best-known example is cilengitide, a small cyclic peptide developed at the Technical University of Munich with Merck KGaA.
Cilengitide reached a large phase III trial in glioblastoma, the aggressive brain cancer. The trial, run by the European cancer research group EORTC, enrolled more than 500 patients in 23 countries, and adding the drug to standard treatment did not help them live longer. Antibodies aimed at the same receptor were also tested in patients without becoming approved treatments. One lesson researchers drew was that simply blocking the receptor's grip is not enough; in some settings it can even send the cell confusing signals.
How ProAgio works differently
Liu's group took another route, which he described as taking "a unique angle". Instead of plugging the usual binding pocket, they used computer modelling to redesign part of a human protein called CD2 so that it would attach to a separate groove on the β3 half of the receptor. Their paper reports that ProAgio binds tightly, and that cilengitide and a classic αvβ3 antibody did not stop it from binding, which supports the idea that it docks somewhere else.
The bigger surprise was what happened next. Once attached, ProAgio did not just block the receptor. It caused the receptor to recruit an enzyme called caspase 8 to the inside of the cell membrane. Caspase 8 is one of the starter switches for apoptosis, the orderly self-destruct routine cells use when they are damaged or no longer needed. When the researchers added a caspase 8 inhibitor, most of the cell death disappeared, which points to that enzyme as a key step.
In plain terms, the protein turns a marker of trouble into a trigger for the cell's own shutdown. The authors were careful to say the exact mechanism is still partly speculative, and their map of the binding site came from modelling and chemical cross-linking rather than a solved three-dimensional structure.
What the 2016 experiments showed
Most of the evidence in the original paper comes from cells in dishes and a small number of mouse experiments, with groups of about six animals.
- Blood vessel growth: in lab tests, ProAgio almost completely stopped human vessel-lining cells from forming tube networks, and in a gel-plug test in mice it held back new vessel growth more strongly than the approved drug Avastin.
- Tumour growth: in mice carrying human prostate cancer cells, a long-acting version of the protein nearly stopped tumours from growing, and some tumours disappeared. It also slowed tumour growth in a mouse breast-cancer model.
- Existing vessels: new tumour vessels were reduced, while established vessels in normal organs looked unaffected.
- Safety signals: mice given repeated high doses showed no obvious damage to the liver, lungs, heart, spleen or kidneys, and blood markers of organ injury stayed normal over the observation period.
These are encouraging results, but they come with limits the authors themselves listed: a short observation window, one dosing plan for the toxicity tests, and weaker binding to a related integrin, αvβ5, which means the protein is not perfectly specific.
The short animation below, released alongside the research, shows how the protein is thought to work.
From mouse studies to the first patients
The work did not stop in 2016. In February 2021, Georgia State reported two further mouse studies. In pancreatic cancer, ProAgio killed cancer-associated fibroblasts, the support cells that make pancreatic tumours dense and hard for drugs to penetrate. That reopened collapsed blood vessels inside the tumour, helped chemotherapy reach its target and extended survival in the animals. In triple-negative breast cancer, it reduced the leaky, disorganised vessels typical of such tumours.
ProAgio is licensed to ProDa BioTech, a company Liu founded, which received a $2 million grant from the US National Cancer Institute in 2018 for the toxicology and drug-handling studies that regulators require before human testing. A first-in-human phase 1 trial, titled "ProAgio in Previously Treated Advanced Pancreatic Cancer and Other Solid Tumor Malignancies", was registered as NCT05085548. A second study, NCT06182072, began in September 2023: an open-label phase I/Ib trial combining ProAgio with the standard chemotherapy pair gemcitabine and nab-paclitaxel in pancreatic cancer, aiming to enrol 46 people and running to an estimated 2028.
Phase 1 trials are mainly about safety and finding a tolerable dose. They can hint at benefit, but they are not designed to prove that a drug works.
Reading “cancer cells self-destruct” headlines with care
Stories about this protein were often headlined as if a cure were close. A few points help keep such news in proportion:
- Mice are not people. Many compounds that shrink tumours in mice never reach patients, often because of side effects or because human tumours behave differently.
- "Kills cancer cells" is the start, not the finish. Plenty of substances kill cancer cells in a dish. The hard part is doing it inside a body, safely, at doses people can tolerate.
- Its target is not unique to tumours. Bone-remodelling cells, immune cells and healing tissue also carry αvβ3, so long-term effects on bone and wound repair are questions human trials will have to answer.
- Trials take years. Even with good phase 1 results, larger randomised trials would be needed before any approval.
If you or someone close to you has cancer and is curious about trials like these, the best first step is to talk to a doctor, who can say whether any open study fits your situation.
What makes ProAgio worth following is the idea behind it: instead of starving tumours or blocking a receptor, it borrows a molecule that tumours rely on and uses it to switch on their own self-destruct programme. Whether that idea survives contact with human biology is what the current trials should begin to show.
Sources: Turaga et al., Nature Communications 7:11675 (2016): Rational design of a protein that binds integrin αvβ3 outside the ligand binding site; Georgia State University news, 31 May 2016; Futurity (Georgia State University), 12 February 2021; ClinicalTrials.gov: NCT05085548; NCT06182072 trial summary; Wikipedia: Cilengitide; animation via YouTube; first reported by ScienceAlert.
Related:
COMMENTS