Emmanuel S. Antonarakis, MD, recently spent thousands of dollars profiling the immune cells of a single patient whose prostate cancer disappeared after immunotherapy — an outcome that once seemed impossible. The experiment wasn’t meant to show that immunotherapy had worked in prostate cancer. It was meant to explain why, against all expectations, it already had.

In the early days of cancer immunotherapy, experts had good reason to believe prostate cancer would respond. In 2010, the FDA approved the therapeutic cancer vaccine sipuleucel-T for men with metastatic castration-resistant prostate cancer after randomized trials revealed an overall survival benefit of about 4 months. Though only a modest benefit, experts were excited. The findings indicated that the immune system could be harnessed against prostate cancer, which fueled optimism that more powerful immunotherapies would follow.

When the immune checkpoint inhibitor era arrived, it transformed treatment for many cancers — melanoma, non-small cell lung cancer, renal cell carcinoma, bladder cancer, and head and neck cancers — but not prostate cancer.

“The optimism was quickly replaced with skepticism,” said Antonarakis, a prostate cancer specialist at the University of Minnesota Masonic Cancer Center in Minneapolis.

Immune checkpoint inhibitors repeatedly failed in large clinical trials. Six phase 3 trials testing checkpoint inhibitors in prostate cancer all came back negative, reinforcing the idea that prostate tumors were simply too “cold” — meaning the tumors had few functional tumor-infiltrating T cells and minimal preexisting antitumor immune activity — for checkpoint inhibitors to work effectively.

“That really threw a huge damper on the field,” said Lawrence Fong, MD, a medical oncologist at the Fred Hutchinson Cancer Center in Seattle, who worked on the target antigen that made sipuleucel-T possible.

Many considered prostate cancer one of immunotherapy’s biggest disappointments. But buried in those averages and medians were some patients who defied them. Some achieved complete remissions that lasted years.

By studying those rare patients — and the immune cells inside their tumors — researchers are starting to understand why most prostate cancers resist immunotherapy and what biological features may distinguish the rare patients who experience extraordinary responses.

Cracks in the Data

In virtually every failed trial, a small subset of men had striking, durable responses to immunotherapy.

“If you’re a medical oncologist like me, you tend to remember the rare remarkable outliers, not the tens or hundreds of patients that did not respond well,” Antonarakis said.

Fong saw the same pattern in his own early-phase work: a small cohort of exceptional responders who benefited from immunotherapy.

Many of the best-characterized exceptional responders had a telltale feature — mismatch repair deficiency. Found in about 2%-4% of prostate cancers, these tumors are often characterized by microsatellite instability and a high tumor mutational burden, features strongly associated with response to immunotherapy.

However, the existing trials were never designed to find those patients. In all negative phase 3 trials, “we were just taking all comers,” Fong said. “We had some exceptional responders,” but “they were just diluted out.”

Rather than conclude that immunotherapy simply does not work in prostate cancer, researchers began asking a different question: Why do a small number of patients experience dramatic, durable responses, while most derive little benefit?

“Nobody has been able to solve that,” Antonarakis said.

Decoding a Cold Tumor

To first understand why most prostate cancers resist immunotherapy, Fong’s lab turned to single-cell profiling to trace how the immune landscape shifts as the disease advances.

Using tumor biopsies from 43 patients across three disease stages — early localized disease, hormone-sensitive metastatic cancer on androgen deprivation therapy, and castration-resistant disease progressing despite treatment — the team sequenced nearly 150,000 individual cells and mapped 14 distinct myeloid subsets within the tumor microenvironment.

“We were thinking, oh, the difference is going to be in the T cells,” he said. But “it really was in the myeloid cells.”

Published in 2024, Fong’s single-cell atlas revealed a specific macrophage population — a myeloid cell subset marked by elevated SPP1 transcripts — that became increasingly dominant as the disease progressed. These macrophages were potently immunosuppressive, capable of shutting down T-cell activity and promoting immune exhaustion. The macrophages carried a molecular feature that explained one of the field’s most puzzling clinical failures: They barely expressed CSF1R, the receptor that the entire myeloid-targeting drug strategy had been built around.

The older classification system for myeloid cells — M1 vs M2 — would have missed this population entirely, Fong said. In other words, earlier myeloid-targeting drugs were likely missing a major target: SPP1-high macrophages.

The analysis also pointed to a different vulnerability. Rather than relying on CSF1R signaling, the SPP1-high macrophages appeared to depend on adenosine signaling. Fong’s team tested this and found about 20%-30% of men with metastatic castration-resistant prostate cancer responded to an adenosine-blocking drug combined with checkpoint blockade.

Looking more closely at the data, “it was the ones who had high levels of these SPP1 myeloid cells that actually responded,” Fong said.

One Patient, One Expensive Experiment

If Fong’s work largely highlighted a key reason immunotherapy fails in prostate cancer — the immune-suppressive myeloid cells that can keep T cells from functioning — Antonarakis’ work explored the other side of the equation: the rare immune features that allow the immune system to mount a response powerful enough to eradicate a tumor.

In a paper published in June, Antonarakis and first author Alexander K. Tsai, MD, PhD, reported a case that reads like a textbook example of an exceptional responder.

A 67-year-old man presented with an 8-cm prostate mass that was invading the bladder and rectum. He had the highest-grade disease, Gleason 5 + 5 = 10. Genetic testing revealed mismatch repair deficiency with an ultrahigh tumor mutation burden of 10 times the median.

Combined hormone therapy failed within 3 months, but when the patient received pembrolizumab, his prostate-specific antigen (PSA) fell to undetectable levels in two cycles. In four cycles, imaging showed no residual tumor. After a radical prostatectomy, the pathologist found no cancer at all.

“I was stunned,” Antonarakis said. “It’s one thing to see a PSA level going to zero. It’s another thing to get a prostatectomy report that says that no viable tumor cells were seen in the entire prostate gland.”

Antonarakis then looked for all tumor mutations. The patient had thousands of them prior to starting immunotherapy. All were gone after treatment: a complete molecular response.

“This patient appears to be absolutely cured,” Antonarakis said.

To understand what drove that response, Antonarakis, Tsai, and colleagues profiled the patient’s immune cells before and after immunotherapy treatment. What emerged were two unconventional T-cell populations: natural killer (NK)-like T cells (CD8+ cells that had acquired NK cell markers and cytotoxic machinery) and “double-positive” T cells expressing both CD4 and CD8, a phenotype normally confined to immature cells in the thymus.

A single NK-like T-cell clone expanded nearly 50-fold immediately after the first dose of pembrolizumab and remained expanded at every subsequent timepoint. Both cell populations were present in the tumor before treatment began.

“We think it was the NK-like T cells that eradicated the cancer,” Antonarakis said. “The double-positive T cells might also have been important here.”

When the team examined other cancer cohorts, including men with prostate cancer receiving ipilimumab, they found expansion of the same two T-cell populations in responders but not in nonresponders.

“This patient cost our lab a lot of money because those single-cell experiments can be pricy,” said Antonarakis. “But I feel so gratified that I did that because we uncovered something important.”

The New Frontier

Now, a new generation of treatments is attempting to either bypass or dismantle the barriers those exceptional responders overcame naturally.

The most clinically advanced are bispecific T-cell engagers, which physically bridge T cells to tumor cells, bypassing the need for these “cold” tumors to naturally attract T cells in the first place.

“The new frontier in prostate cancer immunotherapy is not going to be PD-1 or CTLA-4 inhibition,” Antonarakis said. “It’s going to be CD3-targeting T-cell engagers.”

Three agents — xaluritamig, pasritamig, and VIR-5500 — are the main ones progressing forward. In phase 1 testing, xaluritamig produced PSA declines in roughly half of the heavily pretreated patients and tumor shrinkage in about 40% at the target dose, pasritamig produced PSA responses in about 42% of patients, and VIR-5500 led to PSA declines of 50% or greater in 82% of the 17 PSA-evaluable patients at the highest doses tested.

“Between the three, I bet at least one is going to have a positive phase 3 trial result, just because the clinical activity that we’re seeing in the clinic is unequivocal,” said Fong.

Earlier-stage work is testing whether the suppressive environment itself can be reprogrammed.

Michelle Bradbury, MD, PhD, a professor of radiology at Weill Cornell Medicine in New York City, published a study this year showing how precisely engineered ultrasmall core-shell silica nanoparticles, known as Cornell dots (C’ dots), could reprogram the immunologically “cold” prostate tumor microenvironment and improve the activity of immune checkpoint blockade in clinically relevant mouse models that are otherwise resistant to these therapies.

When combined with checkpoint blockade and a myeloid-targeting therapy, the treatment produced complete tumor remissions in up to half of the animals tested. Trial designs for a first-in-human therapeutic study are underway.

The idea here is to “determine whether a fundamentally different therapeutic platform could address aspects of tumor immune suppression that existing checkpoint inhibitors do not,” Bradbury said.

Targeted radiation could offer another approach to boost the effectiveness of immunotherapy. Radioligand therapy delivers radiation directly to tumor cells, and when those cells die, they release the kind of inflammatory signals that can wake up a suppressed immune system.

In one early-stage 2023 trial, Fong and colleagues gave 25 patients with microsatellite stable disease a single dose of 177-lutetium (¹⁷⁷Lu)-PSMA-617 before starting pembrolizumab. More than half saw their tumors shrink. Meanwhile, the PRINCE trial in Australia, published earlier this year, used multiple cycles of ¹⁷⁷Lu-PSMA-617 along with pembrolizumab and reported PSA declines in 76% of patients.

Both trials were small and lacked control arms, so the effect of the checkpoint inhibitor remains unclear; however, a pattern emerged where breaking the immune suppression first enabled immunotherapy to exert a stronger effect.

These newer approaches will take time to reach the clinic. The bispecifics are several years away from potential approval, and the radioligand and nanoparticle combinations are earlier still.

But Fong and Antonarakis both argued that the field has reached a genuine turning point, built on the wreckage of the checkpoint failures.

After two decades of watching immunotherapy fail to achieve response in his patients with prostate cancer, Fong is finally feeling hopeful. “Immunotherapy is coming of age in prostate cancer,” he said.

But even as the new treatments advance, a smaller subset of exceptional responders remains unexplained — patients who achieved dramatic, durable responses without mismatch repair deficiency or any other known biomarker.

“These are the ones that keep me up at night,” Antonarakis said.

Tumor mutation burden, the best-known predictor of checkpoint response, is really a numbers game, he said. “It’s like winning the lottery. If you buy 1000 lottery tickets, you’ve got a higher chance of winning than if you buy one ticket, but you can still win the lottery if you buy a single ticket if it’s the lucky one.”

The question the field still hasn’t answered: “In those patients that don’t have the super-high tumor mutation burden, how can we identify the rare gentleman who does have that highly immunogenic tumor?”

Fong reported being involved in clinical trials of T-cell engagers from Amgen (xaluritamig), Johnson & Johnson (pasritamig), and Vir Biotechnology (VIR-5500), with support from those companies. Antonarakis reported receiving research grants to his institution from Actinium Pharmaceuticals, Clovis, Constellation Pharmaceuticals, MacroGenics, Merck, Novartis, Orion, and Seagen; receiving consulting fees from Johnson & Johnson, GSK, and others; receiving advisory board honoraria from Astellas, AstraZeneca, Bayer, Merck, Pfizer, Sanofi, Vir Biotechnology, and others; and having a licensing agreement with Qiagen. Bradbury reported being an inventor on patents related to nanomedicine, cancer immunotherapy, tumor microenvironment, and molecular imaging. She reported having no financial relationships with commercial entities relevant to the subject matter discussed.