Loading field map
Immunotherapy is the branch of medicine that treats disease by deliberately altering the activity of a patient's immune system. Rather than attacking a pathogen or tumor directly with a drug, immunotherapy aims to change the relationship between the immune system and the disease—either by strengthening an insufficient immune response, by suppressing an excessive one, or by redirecting immune cells toward a target they were ignoring. The term is used most often in oncology, where the goal is usually to make a patient's own T cells recognize and destroy cancer cells. But the same underlying logic applies across medicine: in autoimmune diseases, the goal is to dampen an immune attack on healthy tissue; in allergy, to blunt an inappropriate response to a harmless substance; in chronic infections, to revive an exhausted response; and in transplantation, to prevent rejection without destroying protective immunity entirely.
The immune system is a distributed network of cells and molecules that distinguishes self from non-self and eliminates threats. Its main effectors include antibodies, which neutralize extracellular pathogens, and T cells, which kill infected or abnormal cells directly. The system also maintains memory, so that a second encounter with the same pathogen produces a faster, stronger response.
For most of medical history, the only way to exploit this system was vaccination: exposing a healthy person to a harmless form of a pathogen so that their immune system would generate protective memory before the real threat arrived. Immunotherapy differs from vaccination in a crucial way. Vaccination primes a response in advance, in a person whose immune system is otherwise functioning normally. Immunotherapy intervenes in an ongoing disease, in a patient whose immune system has either failed to respond adequately, is responding in the wrong way, or is actively causing harm.
The central difficulty of immunotherapy is that the immune system's power is tightly regulated. A T cell that attacks cancer cells with full force could, if misdirected, attack healthy tissue. The same signaling molecules that drive an effective anti-tumor response also drive the inflammation of autoimmune disease. Every immunotherapeutic strategy therefore faces a version of the same trade-off: increasing immune activity risks collateral damage to healthy tissue, while decreasing it risks leaving the patient vulnerable to infection or cancer. Much of the field's history is the story of learning to manipulate specific molecular switches in the hope of widening the therapeutic window between benefit and harm.
The conceptual roots of immunotherapy lie in the late nineteenth century, before the term existed. In 1890, Emil von Behring and Kitasato Shibasaburō showed that serum from animals immunized against diphtheria could cure other animals infected with the disease. This was passive immunotherapy: transferring ready-made antibodies from an immune donor to a sick recipient. The approach worked because the antibodies neutralized the toxin directly, and it became a standard treatment for diphtheria and tetanus. The principle survives today in the form of monoclonal antibody drugs and convalescent plasma, though the modern versions use laboratory-made antibodies rather than animal serum.
A second early thread came from observations that some cancer patients improved after bacterial infections. In the 1890s, William Coley, a New York surgeon, injected killed bacteria into patients with inoperable tumors, hoping to provoke an immune reaction that would also attack the cancer. Coley's results were inconsistent and his mechanism was unknown, and the approach was abandoned with the rise of radiation and chemotherapy. But the underlying idea—that the immune system could be stimulated to fight cancer—never entirely disappeared.
These early efforts were empirical. Practitioners knew that serum could neutralize toxins and that inflammation sometimes accompanied tumor regression, but they had no molecular understanding of how the immune system recognized targets or controlled its own intensity. The field could not advance beyond trial and error until immunology itself became a molecular science.
The modern era of immunotherapy began when researchers identified the specific molecules that immune cells use to communicate and to recognize their targets. Two discoveries in particular created the framework for everything that followed.
The first was the identification of cytokines—small signaling proteins that immune cells release to activate, recruit, or suppress one another. In the 1970s and 1980s, researchers learned to produce these proteins in the laboratory and began testing them as drugs. Interferon-alpha, a cytokine that activates antiviral and anti-tumor responses, became an approved treatment for certain leukemias and melanomas. Interleukin-2, which drives T cell proliferation, produced dramatic regressions in some patients with metastatic melanoma and kidney cancer—the first time anyone had shown that stimulating the immune system could reliably shrink solid tumors. But the responses were unpredictable, and the toxicity was severe, because interleukin-2 activates all T cells, not just the ones that recognize the tumor.
The second discovery was more precise. In the 1990s, researchers identified the molecules on the surface of T cells that act as brakes on activation. These "checkpoint" molecules—most notably CTLA-4 and PD-1—normally prevent T cells from attacking healthy tissue by limiting how strongly they respond to a signal. Many tumors exploit this system by expressing ligands that engage PD-1 on tumor-specific T cells, effectively switching the brakes on and rendering the T cells inactive. Antibodies that block these checkpoints—called checkpoint inhibitors—release the brakes and allow T cells to attack the tumor.
The first checkpoint inhibitor, targeting CTLA-4, was approved in 2011 for metastatic melanoma. It was followed by drugs targeting PD-1 and its ligand PD-L1, which proved effective across a much wider range of cancers, including lung, kidney, bladder, and Hodgkin lymphoma. These drugs do not cure most patients, and they can cause severe autoimmune side effects, sometimes fatal. But in a subset of patients—sometimes a substantial subset—they produce durable responses that last for years, something rarely seen with chemotherapy for metastatic solid tumors. The success of checkpoint inhibitors transformed oncology and made immunotherapy the most active area of cancer drug development.
Contemporary immunotherapy is best understood not as a single technique but as a family of strategies that intervene at different points in the immune response. They can be grouped by their mechanism of action.
Checkpoint inhibition works by removing the brakes on T cells that already recognize the tumor but have been silenced. Its advantage is that it does not require knowing which antigen the T cells recognize; it simply allows whatever anti-tumor T cells exist to become active. Its limitation is that it only works when such T cells are present. Tumors with few mutations, or tumors that have evolved to exclude T cells entirely, tend not to respond. The field has therefore spent considerable effort trying to identify biomarkers—such as the number of mutations in the tumor or the level of PD-L1 expression—that predict which patients will benefit.
Adoptive cell transfer takes a different approach. Instead of releasing existing T cells, it removes T cells from the patient, selects or engineers the ones that recognize the tumor, expands them in large numbers in the laboratory, and infuses them back into the patient. The most successful version of this is CAR-T therapy, in which a patient's T cells are genetically engineered to express a chimeric antigen receptor—a synthetic protein that combines an antibody-like recognition domain with the signaling machinery that activates T cells. CAR-T cells are directed against a specific molecule on the surface of cancer cells, most commonly CD19 on B cells. This approach has produced remarkable results in certain blood cancers, including acute lymphoblastic leukemia and diffuse large B-cell lymphoma, where a substantial fraction of patients who had failed all other treatments achieved long-term remission. The limitations are equally striking: CAR-T therapy requires a complex, individualized manufacturing process; it can cause a dangerous cytokine release syndrome as the engineered cells activate en masse; and it has so far worked poorly in solid tumors, partly because solid tumors lack a single surface molecule that is both universally expressed and absent from healthy tissue.
Cancer vaccines attempt to prime an anti-tumor response from scratch, much as an infectious disease vaccine primes a response to a pathogen. The challenge is identifying the right antigen. Early vaccines targeted proteins that are overexpressed on tumor cells but also present on normal tissue, producing weak responses. More recent efforts have focused on neoantigens—mutated proteins that are unique to a patient's tumor and therefore recognized as foreign by the immune system. Personalized vaccines made from a patient's own tumor mutations have shown promise in early trials, particularly in combination with checkpoint inhibitors, but they remain experimental and logistically demanding.
Monoclonal antibodies that are not checkpoint inhibitors form a separate category. Some work by directly killing cancer cells: they bind to a surface molecule on the tumor and either trigger antibody-dependent cellular cytotoxicity, in which natural killer cells destroy the antibody-coated cell, or deliver a toxic drug attached to the antibody. Others work by blocking growth signals. The first monoclonal antibody approved for cancer, rituximab, targets CD20 on B cells and is a standard treatment for B-cell lymphomas. These antibodies are sometimes classified as immunotherapy because they engage immune effector mechanisms, though they do not alter the patient's immune system in the way that checkpoint inhibitors or CAR-T cells do.
Cytokine therapy—the direct administration of interleukins or interferons—was the earliest form of modern immunotherapy and remains in use, though it has been largely superseded by more targeted approaches. Its main limitation is its lack of specificity: cytokines activate broad populations of immune cells, producing systemic inflammation and toxicity.
Immune suppression is the mirror image of these strategies. In autoimmune diseases such as rheumatoid arthritis, inflammatory bowel disease, and multiple sclerosis, the immune system attacks healthy tissue, and immunotherapy means suppressing that attack. The most targeted versions are monoclonal antibodies that neutralize specific inflammatory cytokines, such as tumor necrosis factor (TNF) inhibitors, which have transformed the treatment of rheumatoid arthritis and Crohn's disease. Other approaches block the activation of specific immune cell types or deplete the B cells that produce autoantibodies. The trade-off here is the same one that limits checkpoint inhibitors, but in reverse: suppressing the immune system increases the risk of serious infections and, in some cases, cancer.
Allergen immunotherapy is a distinct and much older form of immune modulation. It involves exposing an allergic patient to gradually increasing doses of the allergen—pollen, bee venom, peanut protein—with the goal of shifting the immune response from an allergic type to a tolerant one. The mechanism is not fully understood, but it appears to involve the generation of regulatory T cells and a change in antibody class from IgE to IgG. This is the only immunotherapy that aims to permanently alter the immune system's response to a specific harmless substance, and it remains the only disease-modifying treatment for allergy.
These strategies are not competing schools in the sense of mutually exclusive theories. They are complementary interventions at different stages of the immune response, and they are increasingly used in combination. Checkpoint inhibitors are often combined with chemotherapy, which kills tumor cells and releases antigens that can then be presented to T cells. CAR-T therapy is sometimes followed by checkpoint inhibition to prevent the engineered cells from being silenced. Cancer vaccines are being tested with checkpoint inhibitors to provide both the priming signal and the removal of brakes. The field has moved toward the view that effective immunotherapy for solid tumors will require addressing multiple barriers simultaneously: generating T cells that recognize the tumor, getting them into the tumor, keeping them active once there, and preventing the tumor from evolving resistance.
The relationship between cancer immunotherapy and immune suppression is more complicated. The same molecular pathways are often involved in both. PD-1, for example, is a brake on T cell activation; blocking it enhances anti-tumor immunity but can precipitate autoimmune disease. Conversely, drugs that block inflammatory cytokines in autoimmune disease may increase cancer risk. The field is unified by the recognition that the immune system's intensity is regulated by a balance of stimulatory and inhibitory signals, and that therapeutic intervention means shifting that balance in a particular direction for a particular patient.
Immunotherapy has permanently changed the treatment of cancer and autoimmune disease, but its limits are as instructive as its successes. Checkpoint inhibitors produce durable responses in a minority of patients with most solid tumors; they are not a universal cure. CAR-T therapy is transformative for some blood cancers but has not yet been made to work reliably in solid tumors. Cancer vaccines remain largely experimental. The field's central unsolved problem is specificity: how to direct a powerful immune response against a disease target without also attacking healthy tissue, and how to do so reliably across a patient population that is genetically and immunologically diverse.
The most active areas of current research reflect this problem. One is the search for better biomarkers to predict which patients will respond to which therapy, since the current predictors—PD-L1 expression, tumor mutational burden—are imperfect. Another is the development of combination regimens that address multiple barriers simultaneously. A third is the effort to understand resistance: why some tumors initially respond to checkpoint inhibition and then relapse, and how the tumor's microenvironment evolves to suppress immune attack. A fourth is the extension of CAR-T and related engineered cell therapies to solid tumors, which will require identifying target antigens that are truly tumor-specific or developing ways to control the engineered cells' activity after infusion.
Immunotherapy has also raised conceptual questions that remain unresolved. Why do some patients with highly mutated tumors not respond to checkpoint inhibition, while some patients with few mutations do? What determines whether the immune system mounts a productive anti-tumor response or becomes tolerant of the tumor? How much of the variability in response is due to the tumor's genetics, how much to the patient's inherited immune system, and how much to the composition of the bacteria living in the patient's gut, which has been shown in some studies to influence checkpoint inhibitor efficacy? These questions are active areas of investigation, and the answers will shape the next generation of therapies.
What is durable is the framework itself. Immunotherapy rests on the insight that the immune system is not a fixed defense that either works or fails, but a dynamic, regulated system whose activity can be measured, adjusted, and redirected. The field's history is the progressive refinement of that adjustment—from crude serum transfers, to the blunt stimulation of cytokines, to the molecular precision of checkpoint blockade and engineered cells. Each refinement has expanded the range of diseases that can be treated and has made the trade-off between efficacy and toxicity more manageable, but none has eliminated it. The central challenge of immunotherapy remains what it has always been: harnessing a system whose power is matched by its capacity for harm.