Penn researchers find a way through the labyrinth keeping CAR-T from solid tumors
Blood vessels are supposed to act like trees, pumping in oxygen tissues need to grow and immune cells required to clear out pathogens. But in tumors, the forest can go a bit haywire. Vessels grow prodigiously and bulge and twist at abrupt points, making it difficult to even tell what’s a vein and what’s an artery. It starts to look less like a forest and more like a gnarled root floor. “A disorganized labyrinth,” one oncologist has called it.
For cancer, chaos is a virtue. That gnarled root floor insulates solid tumors from immune cells and, in recent years, has flustered drug developers’ best attempts at developing therapies meant to rev up the immune system and direct it toward the tumors.
Researchers at the University of Pennsylvania, however, think they may have stumbled onto a solution, a way of whipping the blood vessels back into proper shape. If it works, experts say, it could pave the way for CAR-T treatments that attack solid tumors and potentially improve the effectiveness for more traditional approaches, such as radiation and chemotherapy.
“It’s a really novel and potentially important approach,” Patrick Wen, a neuro-oncologist at Dana-Farber who was not involved in the work, told Endpoints News. “They really did good work. This is a very different way of improving immunotherapy.”
Yi Fan, a radiation oncologist at Penn’s School of Medicine, has been working for the last few years to understand why the labyrinth appears in the first place. Researchers had previously circled in on the so-called growth factors that stimulate blood vessel formation. Attempts to block these factors, though, disappointed; Avastin, an antibody against the factor VEGF, became a blockbuster but has continually failed to improve survival on a range of malignancies.
Scientists would have to go more fundamental. In a pair of 2018 papers, Fan showed that part of the problem is a process called “endothelial cell transformation.” Cells lining the blood vessels around the tumor acquire stem cell-like properties that allow them to reproduce and expand rapidly, as stem cells do.
“There’s a genetic reprogramming,” Fan told Endpoints. “They’ll become really aggressive.”
But how did that reprogramming happen? If Fan could pin down the pathway, he figured he could then devise a way to block it. He started knocking out kinases — the cellular engines that can drive epigenetic change, or “reprogramming” — one by one in endothelial cells isolated from patients with an aggressive brain cancer called glioblastoma. Out of 518, 35 prevented transformation and one did so particularly well: PAK4.
Then they injected tumors into mice, some who had PAK4 and some who had the kinase genetically removed: Eighty percent of the mice who had PAK4 removed lived for 60 days, while all of the wild-type mice died within 40. Fan’s team also showed that T cells infiltrated the tumors more easily in the PAK4-less mice.
It was a fortuitous finding: Drug companies had developed several PAK inhibitors a decade ago, when kinase inhibitors were the flashiest thing in pharma. Many had been abandoned, but Karyopharm had recently brought a PAK4 blocker into Phase I.
To see whether drug developers could exploit this finding, Fan and his team removed T cells from mice and developed a CAR-T therapy to attack the tumors.
They gave mice three different regimens. The CAR-T therapy on its own failed to reduce tumor size, apparently unable to reach through the vessels. The Karyopharm drug also had little effect on its own. But combined, they managed to reduce tumor size by 80% after five days. They published the results in Nature Cancer this week.
“It is a really eye-opening result,” Fan said. “I think we see something really dramatic.”
That, of course, is just in mice, but Fan already has strong supporting evidence for PAK4’s role in cancer. Last December, while Fan was still completing his experiment, Nature Cancer published a paper from Antoni Ribas’ UCLA lab suggesting that PAK4 inhibitors can help T cells infiltrate around various solid tumors. They showed that the same Karyopharm inhibitor could boost the effects of PD-1 inhibitors in mice, allowing activated T cells to better reach tumors.
That work has already translated into the clinic; weeks after it came out, Karyopharm added an arm to their Phase I study of the drug that will look at the PAK4 inhibitor in combination with the PD-1 blocker Opdivo.
Ribas said that Fan’s work is compelling and helps confirm the role of PAK4, but he said a CAR-T therapy would face a much longer path to the clinic. It’s simply much easier to combine an approved drug with an experimental one than to devise a new CAR-T therapy, mix it with the unapproved inhibitor (and all the other things, such as bone marrow-clearing chemotherapy, CAR-T recipients receive) and then deduce what effect each is having.
“It will a take a while,” Ribas told Endpoints. “But I hope this is right and it’s developed clinically.”
There are also other unresolved obstacles for CAR-T in solid tumors, Wen said. Developers still struggle to find targets that won’t also send the super-charged T cells after healthy tissue. And tangled blood vessels are just one of several mechanisms tumors have of defending themselves. They can, for example, turn tumor-eating immune cells into tumor-defending ones.
Still, Wen said, in the short term, the approach offered a path toward boosting the efficacy of radiation, chemotherapy and other small molecule drugs. Although Fan focused on glioblastoma, researchers agreed PAK4 likely plays the same vessel-warping role in many other solid tumors.
“There’s a lot of things you could look at,” he said.
In a January review, Jessica Fessler and Thomas Gajewski at the University of Chicago said Ribas’ paper pointed towards a path for improving PD-1 and overcoming resistance in some tumors. But they also raised questions about the Karyopharm drug, noting that it hits other proteins besides PAK4. That could mean other mechanisms are also at play and that the drug could affect other tissues in humans.
Ribas agreed that Karyopharm’s drug might not be the perfect molecule but said others could be on their way. He serves as a scientific advisor to Arcus, the Terry Rosen startup that is now working on developing its own PAK4 inhibitor.
“If they can develop a very selective PAK4 inhibitor,” he said, “it may be a more direct way of testing the role of PAK4.”
Tests with that drug, in turn, could help clear up a biological mystery that emerged out of Fan’s and Ribas’ papers. Although both investigators zeroed in on PAK4, each of them suggested very different mechanisms by which PAK4 kept immune cells out of the tumor. Ribas suggested it directly suppresses T cells, while Fan found it led to those transformations inside the blood vessels near the tumor.
Kinases are versatile proteins and both researchers said it’s possible that PAK4 is doing both. It’s also possible, they said, that one is more important than the other, or simply that one of them is just wrong.
“When you start with completely new biology, it’s hard to get it right the first time,” Ribas said.