“We send your T cells to boot camp.”

That is how hematologist-oncologist and Stanford Cancer Institute member Melody Smith, MD, describes CAR-T cell therapy to her patients, a simple metaphor for a treatment that is changing the way doctors approach certain blood cancers.

“We arm the patient’s T cells with a receptor that recognizes a marker on the tumor so that, when the cells are reinfused, they can specifically target and kill the cancer.”

These “trained” cells, known as CAR-T cells, come from a patient’s own white blood cells, specifically T lymphocytes. CAR-T cells target specific markers on a tumor. It is FDA-approved for several types of blood cancer, including leukemia, lymphoma, and multiple myeloma.

“It has really revolutionized the treatment,” Smith said. About 40% of lymphoma patients receiving CAR-T cells achieve long-term, durable remission. The responses we have seen are remarkable.”

At the same time, the therapy doesn’t keep the cancer from coming back long-term for everyone. About 60% of patients still relapse, and some experience serious side effects, including neurological complications. Why do some patients experience life-saving remissions while others relapse? Smith wondered whether the answer might be hidden in an unexpected place: the gut.

The gut microbiome connection

Smith began to wonder if gut microbiome factors influence the divergent outcomes of CAR-T therapy. Research in related areas offered a clue. Studies on immune checkpoint therapy and in patients undergoing bone marrow transplants show that the composition of the gut bacteria can significantly impact treatment responses and complications. 

The gut microbiome refers to the vast community of organisms living in the digestive tract. These microbes are key in regulating immunity, producing metabolites, small chemicals made by bacteria that influence immune cells, help digest nutrients, and support overall health.

For the microbiome to function properly, it needs a stable and diverse community. Treatments such as chemotherapy and antibiotics can disrupt that balance, resulting in a condition known as dysbiosis, meaning an imbalance in gut bacteria. Cancer patients often arrive at CAR-T therapy with compromised microbial communities.

“Among medications, antibiotics induce the most disruption to the microbiome,” Smith said.

Many patients receive antibiotics because their immune systems are weakened after chemotherapy, leaving them susceptible to infections. Neutropenic fever, a common side effect of chemotherapy that results from a low white blood cell count, requires rapid antibiotic treatment prior to CAR-T cell infusion. Ironically, the medications that protect them could also interfere with the treatment designed to save their lives.

Antibiotics' double-edged sword

During her postdoctoral work at Memorial Sloan Kettering Cancer Center, Smith and colleagues at the University of Pennsylvania conducted the first study to examine how the gut microbiome influences CAR-T cell outcomes. They found that patients who received PIM antibiotics, an acronym for piperacillin–tazobactam, imipenem, and meropenem, which target anaerobic bacteria important for maintaining gut microbial balance, had significantly worse outcomes.

“Exposure to PIM in the four weeks prior to CAR-T cell therapy correlated with decreased progression-free and overall survival, as well as increased neurologic toxicity,” she said.

To ensure that these worse outcomes were not simply due to patients receiving antibiotics being sicker to begin with, Smith and her team adjusted for clinical factors, such as disease status and prior treatments. Even after accounting for these factors, exposure to anaerobe-targeting antibiotics remained strongly associated with lower survival rates.

Multiple independent studies have since confirmed these associations. The research indicates that the specific type of antibiotic matters, rather than antibiotic use itself. 

“It is not antibiotic exposure per se,” Smith explained, “but which bacteria the antibiotics target.”

Smith emphasized that antibiotics remain essential.

“The take-home message is not that antibiotics should not be given when needed,” she said, “but that we should be more judicious and practice antibiotic stewardship.”

Stanford has begun to adjust its approach. The antibiotic cefepime, commonly used to treat neutropenic fever, was not associated with worse outcomes.

“When patients have a fever during their CAR-T cell treatment, we administer cefepime instead of broader-spectrum anaerobe-targeting antibiotics,” she explained.

Cefepime treats infections effectively and better preserves the gut microbiome. The team intends to analyze how this shift affects outcomes as more data becomes available.

Uncovering microbiome clues

Smith’s lab is now digging into exactly how gut bacteria affect CAR-T cells.

“I am most excited about understanding what the microbiome is doing, particularly how bacterial metabolites influence CAR-T cell efficacy and toxicity,” she said.

Her team uses mouse models to investigate the effects of antibiotics on gut bacteria and CAR-T therapy, including tumor growth. These models enable them to observe how changes in gut microbes affect therapy outcomes with greater clarity than can be achieved in patients alone.

Smith’s lab is now testing various bacterial metabolites, byproducts of bacterial breakdown of food, to see whether any can enhance CAR-T cells or help them fight cancer more effectively and for longer periods. She notes that dozens of candidate metabolites are being screened, and early results suggest several may influence CAR-T cell activity.

Rewriting cancer care from the inside out

Smith envisions a future in which microbiome information is routinely integrated into clinical care.

“In an ideal scenario, maybe ten years from now, microbiome sequencing could be much faster and significantly less expensive,” she said. “We could collect a sample, sequence it in a cost-effective way, assess a patient’s microbiome before treatment, and determine interventions based on their microbiome profile.”

These interventions could include dietary changes, administration of a defined consortium of bacteria, targeted antibiotics, specific microbial metabolites, or adjustments to the CAR-T cell manufacturing processes. Even modest improvements could lead to meaningful benefits.

If we can improve remission rates by just 10 to 20% through these interventions, that would be hugely beneficial for patients.

“If we can improve remission rates by just 10 to 20% through these interventions,” Smith said, “that would be hugely beneficial for patients. They would need less treatment, enjoy a better quality of life, and require fewer subsequent lines of therapy. This potential benefit motivates us as we investigate these questions.”

CAR-T cell therapy has already transformed the outlook for patients with difficult-to-treat blood cancers. Smith emphasized that the gut microbiome may hold some of the most important clues to making CAR-T therapy safer and more effective. The next significant advancement in cancer immunotherapy may not come from engineered cells, but from the trillions of microbes already living inside us.