Immune cells can reach head and neck tumors, but once there, many stop functioning. This helps explain why immunotherapy with immune checkpoint inhibitors — which have transformed treatment for some cancers — benefit only 17%–23% of patients with head and neck squamous cell carcinoma. These are cancers that arise from cells lining the mouth, throat and voicebox.

Researchers at the Johns Hopkins University School of Medicine and Johns Hopkins Sidney Kimmel Comprehensive Cancer Center and its Bloomberg~Kimmel Institute for Cancer Immunotherapy have now measured what those immune cells run on (how each cell generates and uses energy) in tumor tissue taken from patients. They found that the cells that have stopped working share a metabolic weak point, and that blocking that weak point in laboratory experiments changes how those cells behave, reactivating them. This identifies a potential target for future treatment, but should not be considered a treatment itself.

A description of the work was published Oct. 28 in Cancer Immunology Research, a journal of the American Association for Cancer Research.

Head and neck squamous cell carcinoma is the seventh most common cancer worldwide, with about 890,000 new cases each year. Growing evidence suggests that metabolic reprogramming can influence whether a T cell mounts an effective response or becomes exhausted, but researchers have rarely examined this link directly in human tumor tissue.

“Take a piece of head and neck tumor tissue, and you will find immune cells inside it, cells whose sole purpose is to recognize something abnormal and destroy it. The question is: What has happened to them?” says senior author Martin Alphonse, Ph.D., an assistant professor of dermatology at the Johns Hopkins University School of Medicine.

Many have become exhausted — a state in which a cell, activated too long and continuously, stops responding, Alphonse explains. Checkpoint immunotherapy, such as anti–PD-1 antibodies, releases one of the brakes on those cells. “For some patients, it works remarkably well. For many others, it does not, and the reasons have remained unclear,” he says.

Metabolism is a plausible place to look, the researchers say, because an immune cell’s capacity to act is constrained by how it generates energy. However, asking that question in human tumors has been difficult. The standard laboratory method for measuring a cell’s metabolism requires a large, pure cell population, often difficult to obtain from biopsy tissue.

“What arrives is a mixture; there isn’t much of it, and the cells you most want to study are often the rarest ones. So, most of what we know about immune cell metabolism comes from mouse tissue and cells grown in dishes,” says the study’s first author, Sujeetha A. Rajakumar, Ph.D., a former research associate in the Department of Otolaryngology–Head and Neck Surgery at Johns Hopkins Medicine. 

In this study, the team measured each cell’s metabolic machinery alongside the surface markers that identify it. This let them read a single cell’s identity, functional state and metabolic profile at once. They applied this method to different types of tumor-infiltrating immune cells including mucosal-associated invariant T (MAIT) cells, CD8+ T cells, and innate lymphoid cells (ILCs).

The tissue samples were collected from patients who participated in two phase II trials at Johns Hopkins Medicine in which immunotherapy was given before surgery. Participants received the anti-PD-1 antibody nivolumab alone or in combination with an antibody against interleukin-8, a signaling protein that directs immune cells to areas of inflammation or infection. Tumor tissue samples were collected before the start of immunotherapy and again four weeks later, at the time of surgery, allowing investigators to study changes within the cells.

Three findings stood out. First, they identified a population of immune system T cells within these tumors that proved to be metabolically fitter and less exhausted than neighboring cells. Second, the more exhausted cells showed elevated levels of the enzyme glucose-6-phosphate dehydrogenase (G6PD), which marks a metabolic pathway those cells rely on — and therefore a possible point of intervention. Third, a subset of ILCs carried a metabolic profile associated with immune suppression, suggesting they may pose an obstacle to checkpoint therapy.

To test the finding that the more exhausted cells rely on the G6PD pathway, the researchers turned to laboratory cultures of tumor-infiltrating CD8+ T cells from a separate group of untreated patients. They sorted the cells by expression of a protein called CD39, which can contribute to immunosuppression by limiting the activation of immune system T cells. Then, they treated the cells either with a G6PD inhibitor or an anti–PD-1 antibody, or gave both treatments or neither treatment. The combination treatment produced the largest gene expression changes in the most exhausted cells and increased their secretion of CD27, a marker of T-cell activation.

“When we blocked this metabolic pathway alongside checkpoint inhibitor treatment, the cells we expected to be least reachable changed the most and became reactivated,” Rajakumar says. “That tells us the pathway is worth pursuing.”

“These types of interdisciplinary, collaborative studies are critical to understanding how to render therapies more effective in the future,” says study co-author Carole Fakhry, M.D., M.P.H., senior associate dean for clinical affairs at the Johns Hopkins University School of Medicine, and Charles W. Cummings, M.D., professor of otolaryngology–head and neck surgery, and professor of oncology. “We are lucky at Johns Hopkins Medicine to have a rich environment that leverages the translational capabilities of clinical trials to provide critical insights.”

“These findings need confirmation in larger patient groups before anything can be built on them,” Alphonse says. The next steps, he says, are to reproduce the findings in independent cohorts and to test whether interfering with this metabolic step has any effect in a living system, not just in a dish. 

Additional co-authors are Namya Nanda, Chloe Kim, Dustin Dikeman, Tanguy Y. Seiwert, Zubair Khan and Sewon Kang, all affiliated with Johns Hopkins Medicine.

The work was supported by the Johns Hopkins Department of Dermatology; a Career Development Award from the Dermatology Foundation; the Johns Hopkins Department of Otolaryngology–Head and Neck Surgery; and the Bloomberg~Kimmel Institute for Cancer Immunotherapy.

Alphonse and Rajakumar are inventors on two provisional patent applications assigned to The Johns Hopkins University related to the work described herein. The Johns Hopkins University manages these relationships in accordance with its conflict-of-interest policies.