Cancer immunotherapy
Cancer immunotherapy treats a cancer by modulating the immune response against it, usually by reactivating responses to cancer-cell antigens that the tumour has switched off. The anti-tumour response is generated in steps and then suppressed inside the tumour, each class of agent intervenes at one of those steps, and the adverse events follow from the same mechanism that gives the efficacy. Therapeutic vaccines, from an autologous dendritic-cell product to mRNA vaccines, are one such class. The agents are combined with chemotherapy and radiotherapy, and the field grew from a nineteenth-century observation of tumours regressing after infection.
What immunotherapy is
Immunotherapy may be defined as the treatment of disease by modulating the immune response, and on that definition it could be argued that some vaccines, therapeutic vaccines for example, are a form of immunotherapy. Immunotherapies harness the body's own immune system to target and attack a disease. Unlike chemotherapies, which essentially assault all rapidly growing cells, and unlike targeted therapies such as tyrosine kinase inhibitors, which interrupt cellular signalling to restrict tumour growth, immunotherapies target the immune system instead of the disease. Most immunotherapies are biological agents, such as cytokines, inhibitory or activating monoclonal antibodies, engineered lymphocytes, and allergens. Many of the drugs approved in recent decades have been immunotherapies and considerably more were advancing through pharmaceutical and biotechnology pipelines, with unprecedented efficacy against several cancers including melanoma and lymphoma, and trials under way in many others. Cancer immunotherapy was one of the most active areas of research and development in 2018. Vaccines and cancer immunotherapies were among the largest categories of pharmaceutical research and development products in 2017.
An immunotherapy may be designed to amplify an immune response or to reduce it, and either can be achieved by targeting various stages in innate and adaptive immunity — neutralising specific inflammatory or immunosuppressive cytokines, for instance, or inhibiting regulatory molecules involved in antigen presentation to T lymphocytes. Cancer immunotherapy usually involves reactivating immune responses against cancer-cell antigens that the immunosuppressive environment within tumours has switched off; immunotherapies for autoimmune diseases are usually designed to do the opposite and suppress immune-mediated destruction of host tissues. Monoclonal antibodies neutralise or block certain proteins or may deliver a radioisotope or chemicals to mediate targeted cell killing; examples are anti-tumour-necrosis-factor antibody for rheumatoid arthritis and immune checkpoint inhibitors for boosting T-cell responses against tumours. Cell-based therapies inject modified activated immune cells that fight threats such as malignant cells; chimeric antigen receptor T cells, for example, eliminate leukaemic cells. Suppressive immunotherapies, such as allergen desensitisation and immunosuppressive drugs, suppress the immune system or induce tolerance, and cytokines, such as interleukin-2 and interferon-α, are immune signalling molecules that modulate immune responses.
In one account, immunotherapies divide into two main types: active ones, which engage the host's immune response, and passive ones, in which the agent directly neutralises the target and induces no immune response. In another, cancer immunotherapies were in the past classified as passive, usually defined as those stimulating a patient's own immune response, or active, those inducing a de novo response to attack tumour cells directly; this definition is often misused in the literature and not always relevant, and the terms can be misleading and should probably be revised. An alternative classification sorts agents by whether they modulate an existing immune response, or provoke a de novo response or replace a missing one.
The anti-tumour immune response
The immune response to a tumour follows three main successive steps. The first is tumour recognition: tumour-associated proteins, or antigens, released by dead or dying tumour cells are captured by specific immune cells, mainly dendritic cells, which process them and present them on their surface — which is why dendritic cells are also known as antigen-presenting cells. When dendritic cells process and present the antigens, they also need an activation or maturation signal, which can come through a number of immune-stimulating pathways. The typical signals, sometimes called endogenous adjuvants in contrast to the exogenous adjuvants given therapeutically, are pro-inflammatory cytokines, co-stimulatory CD40 and CD40 ligand proteins, factors released by dying tumour cells such as adenosine triphosphate (ATP) or high-mobility group box 1 protein (HMGB1), and toll-like receptor ligands. In the cancer-immunity cycle, after an initial skirmish between neutrophils and the tumour antigen, antigen-presenting cells, typically dendritic cells, engulf the necrotic remains and carry the digested fragments to inform T cells.
The second step generates the response: antigen-presenting dendritic cells travel to the lymph nodes and elicit an antigen-specific T-cell response. Without a co-stimulatory maturation signal, "immature" dendritic cells suppress the response to the tumour by promoting immunosuppressive regulatory T cells or by inducing T-cell depletion or anergy — the absence of a response to an antigen — a phenomenon of immune suppression also known as immune tolerance. With the maturation signal, "matured" dendritic cells provoke or stimulate a T-cell response, mainly of effector cytotoxic T cells. That response also depends on specific interactions between dendritic cells and T-cell co-stimulatory molecules: interaction of CD80 or CD86 with CD28, or of OX40 ligand with OX40, stimulates, while interaction of CD80 or CD86 with CTLA-4, or of PD-L1 or PD-L2 with PD-1, suppresses T-cell responses. T-cell priming and activation is therefore a critical stage that determines the nature of the immune response.
In the third step the cytotoxic T cells leave the lymph node together with other tumour-antigen-specific lymphocytes — B cells, natural killer cells and natural killer T cells — reach the bloodstream and head for the tumour site, where they enter the tumour bed and the cytotoxic T cells recognise and kill the cancer cells. The primed T cell divides and proliferates, dispersing T cells into the bloodstream to patrol for and eliminate tumour cells. The dead and dying cells then provide a fresh source of tumour antigens, also called neo-antigens, which initiate a new immunity cycle.
Described another way, an effective immune response requires four steps, whether the targets are conventional pathogens or cancer cells: foreign antigens must be recognised and presented to cytotoxic T cells to activate them, which antigen-presenting cells of the innate immune system, particularly dendritic cells, do; the activated T cells must travel to the targets; they must kill them; and memory T cells must develop to complete the eradication and maintain control against further incursions.
Immune surveillance and immune escape
It is well established that the immune system recognises and eliminates cancer cells, a process known as immune surveillance, and a failure in it, immune escape, is associated with cancer initiation and progression. Because the same rules apply to cancer cells and pathogens, it is puzzling that the immune system does not prevent cancer. In the long run, surveillance may act as an evolutionary force that promotes more problematic cancer cells, which can avoid T cells or induce changes in the tumour microenvironment that inhibit T-cell infiltration and activity. Several immunotherapies were being developed to deal with these problems.
The immune system's maintenance of homeostasis has been compared to driving a car with gas and brake pedals: it must mount attacks to keep pace with new mutations or other malignant antigens, and it must also suppress autoimmune momentum to avoid damaging the body. The brakes, the immune checkpoints, are the vulnerable aspect of the cancer-immunity cycle. Cancer cells can bypass these regulatory pathways by passing as self instead of non-self. Tumour cells may enter immune-editing, co-opting the mechanism that protects against autoimmune assault; once they evade elimination they can remain seemingly dormant, in equilibrium, while mutating and developing neoantigens, and may then escape altogether, resulting in metastasis or recurrence. Immune-editing describes the interaction between tumour and immune system as a three-phase process of elimination, equilibrium and evasion.
Immunosuppression in the tumour
Killing cancer cells is not that simple. Inside the tumour the cytotoxic T cells meet an immunosuppressive environment. Tumour cells and the myeloid-derived suppressor cells infiltrating the tumour tissue use a variety of strategies to suppress cytotoxic T-cell function: tumour cells release T-cell suppressor molecules such as prostaglandin E2 and indoleamine 2,3-dioxygenase (IDO), while the suppressor cells produce inhibitory molecules such as arginase and nitric oxide synthase. More importantly, tumour cells express programmed cell death protein ligands PD-L1 and PD-L2 on their surface, which engage PD-1 receptors on the cytotoxic T cells and cause their anergy. Reduced oxygenation within the tumour tissue, a common feature of rapidly growing tumours, releases immunosuppressive molecules including vascular endothelial growth factor A, adenosine and CCL28 that add to the local immunosuppression. Tumours have also developed mechanisms to interfere with antigen presentation, suppressing the initial step of the response. All these mechanisms make it challenging for cytotoxic T cells to target and kill cancer cells efficiently.
Anticancer immunotherapies and the steps they target
The aim of anticancer immunotherapy is to initiate, stimulate or restore anti-tumour immunity without disrupting the self-tolerance mechanisms, which would result in pathological autoimmune inflammatory responses.
Immunotherapies can intervene at three critical stages of the anti-tumour response: by stimulating antigen processing and presentation by dendritic cells; by stimulating the T-cell response in the lymph nodes; and by overcoming the immunosuppressive mechanisms within the tumour microenvironment. At the first stage the intervention may be a therapeutic cancer vaccine introducing tumour-specific antigens from outside the body to stimulate the T-cell response. At the second, it may be the delivery of exogenous adjuvants that initiate dendritic-cell maturation, such as toll-like receptor ligands or CD40 antibodies that activate receptors on the dendritic cell, or even the provoking of tumour-cell lysis to release the adjuvant molecules ATP or HMGB1. At the third, immunosuppression at the tumour site can be antagonised by blocking the PD-L1 and PD-1 interaction between tumour cells and cytotoxic T cells, by counteracting immunosuppressive molecules such as IDO or adenosine, or by enhancing recognition of cancer cells by cytotoxic T cells.
The agents that modulate an existing response include immunomodulatory monoclonal antibodies, whether co-stimulatory or blocking; immunostimulatory cytokines such as interleukin-2; and small molecules such as inhibitors of immunosuppressive metabolism, which boost the immune response or block immunosuppressive T-cell functions. The blocking antibodies anti-PD-1 (nivolumab and pembrolizumab) and anti-CTLA-4 (ipilimumab), known as checkpoint inhibitors or checkpoint blockers, were FDA-approved drugs as of 2018 and represent a major breakthrough in immuno-oncology. They can restore anti-tumour T-cell function and showed clinical benefit to some cancer patients. The agents that provoke a de novo response or replace a missing one include cell therapy, known as adoptive cell transfer, anticancer vaccines, oncolytic viruses and bispecific T-cell engagers. Conventional adoptive cell therapy isolates tumour-infiltrating T cells from the patient, expands them in vitro and reintroduces them; two recently described and very promising methods, T-cell receptor and chimeric antigen receptor, under development and clinical evaluation in 2018, isolate T cells from the patient's blood and manipulate them in vitro to redirect their specificity toward tumour-specific antigens before reintroduction. Anticancer vaccines — dendritic-cell, whole-tumour, DNA or peptide based — deliver tumour-specific antigens to initiate a response; oncolytic viruses selectively infect and kill tumour cells; and bispecific T-cell engagers are antibody-based recombinant molecules that force cytotoxic T cells to recognise tumour-associated antigens and so activate anti-tumour cytotoxic activity.
Cancer produces mutations and thus new antigens, neoantigens, which promote T-cell activation via dendritic cells when major histocompatibility complex proteins present them on the cell surface. Neoantigens could be used to develop anticancer vaccines, but efforts with them have been problematic because they are usually unique to each patient; technological advances were making patient-specific vaccines plausible. T cells isolated from melanoma patients' tumours can be expanded and stimulated in vitro and returned to the patient, where they often successfully attack the remaining tumours, and this adoptive cell transfer also underlies chimeric antigen receptor T cells. T cell–engaging bispecific antibodies bind CD3 on the T-cell surface with one arm, which directly activates the T cell, and a cancer-cell-specific surface protein with the other, bringing the activated T cell into contact with the cancer cell so that it can kill it. As with chimeric antigen receptor T-cell therapy, finding an appropriate cancer-cell surface protein is a problem.
Anticancer immunotherapeutics also include tumour-targeting as well as immunomodulatory monoclonal antibodies, pattern recognition receptor agonists such as toll-like receptor agonists, immunogenic cell death inducers — radiation therapy and chemotherapy — and inhibitors of immunosuppressive metabolism such as IDO or adenosine inhibitors.
Checkpoint inhibitors
Checkpoint inhibitors are monoclonal antibodies designed to disrupt the regulatory intersection where cancer can prevent its own elimination by manipulating the adaptive immune system. Checkpoint blockade antibodies block molecular regulators of T-cell activation, resulting in enhanced anti-tumour immune responses. They can release the brakes by keeping the receptors and ligands that suppress immunity from binding together. As of 2019, the approved checkpoint inhibitors worked by blocking one of two checkpoint pathways: CTLA-4 binding the B7 ligands CD80 and CD86, which impedes immunity early in the process, for instance in the lymph nodes, and PD-1 binding PD-L1, which applies the brakes later in peripheral areas such as where a solid tumour sits. At least 23 other inhibitory and stimulatory pathways were being studied in about 95 early-phase trials. The CTLA-4 checkpoint, formed when CTLA-4 binds CD80 and CD86, also called B7-1 and B7-2, sends an inhibitory signal to T cells, and PD-1 on cytotoxic T cells, when bound to PD-L1, applies the brakes that keep the immune system from destroying healthy cells. Activation of a T cell requires antigen bound to its receptor plus costimulation through CD28. The activated T cell soon expresses CTLA-4, which prevents further costimulation by competing with CD28 and markedly reduces T-cell activity, and maintained production of PD-1 and its binding to PD-L1 on tumour cells also reduces T-cell activity. Antibodies against these checkpoints often rescue T cells from exhaustion, restore antitumour activity and cure cancers.
Among the attempts to strengthen the immune system against cancer, checkpoint inhibitors had gained the widest attention by 2019 for improving efficacy and reducing toxicity, with FDA approvals for indications from non-small cell lung cancer to bladder cancer. James Allison developed the first approved checkpoint inhibitor, a CTLA-4 monoclonal antibody, and the approval of ipilimumab for advanced or metastatic melanoma began the cascade of later approvals, most of which sprang from Tasuku Honjo's discovery of the PD-1 and PD-L1 checkpoint. In October 2018 the two were awarded the Nobel Prize in Physiology or Medicine "for their discovery of cancer therapy by inhibition of negative immune regulation". Checkpoint inhibitors deliver previously unseen rates of long-term durable response and stability, but as monotherapies they still reach only a portion of patients, with objective response rates in clinical trials typically 35% to 40%. CTLA-4 blockade, also known as checkpoint inhibition, can enhance anti-tumour immunity and be used to fight cancer.
Adverse events
The same mechanisms of action that give immunotherapies their efficacy can also cause adverse events. Stimulating an immune response, as in cancer immunotherapy, can produce immune-related adverse events such as gastrointestinal, hepatic and skin inflammation, ranging from mild to lethal; suppressive immunotherapies can also cause adverse events such as increased susceptibility to infections.
Immunotherapies have different toxicity profiles from cytotoxic agents, and their toxicity does not follow the pattern seen with traditional chemotherapy. The toxicity profile depends on the agent, its mode of action and the type of tumour it targets. Toxicities divide broadly into infusion reactions and immune-related adverse events. Infusion reactions, allergic or non-allergic, are immune reactions that most frequently occur during the first administration; immunotherapies in general have a low incidence of them, but some have a non-negligible incidence of non-allergic reactions resulting from cytokine release, which causes symptoms including fever, nausea, chills, hypotension, tachycardia and fatigue. Among immune-related events the most frequently affected organs are the skin, colon, endocrine organs, liver and lungs, and the commonest events are diarrhoea, rashes and fatigue. In contrast to chemotherapy, the onset of immune-related adverse events is often delayed, some beginning as long as a year after treatment. Overall, immunotherapeutics are well tolerated, but severe and life-threatening toxicities do occur.
Checkpoint inhibitors are generally less toxic than chemotherapy, whose toxicities are predictable and intense and whose side effects often interfere with completing adequate treatment in older patients through discontinuation or dose reduction. Checkpoint inhibitors can still have unwelcome side effects, occasionally severe and even deadly (in under 1% across the class). Their immune-related adverse events are primarily inflammatory reactions triggered by immune stimulation and can affect any organ system: skin (itching, rash), eyes, intestines (diarrhoea, colitis), lungs (interstitial pneumonitis), endocrine systems (thyroid, adrenal) and nerves (peripheral neuropathy), among others, with pneumonitis, colitis and pancreatitis among the life-threatening ones. A meta-analysis of more than 11,000 patients in 73 checkpoint-inhibitor trials found any-grade immune-related events considerably more frequent with CTLA-4 treatment (53.8%) than with PD-1 (26.5%) or PD-L1 treatment (17.1%). Patients with preexisting autoimmune disorders are predictably vulnerable to these events, but even grade 3 and 4 events prove manageable with immunosuppressive treatment such as corticosteroids.
Checkpoint inhibitors are generally considered as well tolerated by older as by younger patients, although the understanding of their pharmacodynamic effect and toxicity in older patients is not yet solid, primarily because of underrepresentation in clinical trials and possibly because of suboptimal reporting, and for those over 75 the literature conflicts on rates of toxicity. Older patients receiving them will likely need closer monitoring than standard — for dehydration and renal insufficiency with diarrhoea, and for an increased risk of bone fracture with extended corticosteroid use. As of 2019, whether checkpoint inhibitors offer substantial promise for an ageing population had not received focused investigation.
Autologous dendritic-cell therapy
Cvac, produced by the Australian biotechnology company Prima BioMed and trialled in epithelial ovarian cancer, is an immunotherapeutic product that uses modified autologous dendritic cells to prime the patient's own immune system to attack the cancer cells. It is designed to treat patients while they are in remission, in the hope of prolonging remission: once the immune system is primed to attack the cancer cells, recurrences would be destroyed before they grow enough to cause trouble. The product relies on the fact that many ovarian cancer cells over-express a surface protein, mucin 1; it is normally extensively glycosylated in healthy cells but much less so in ovarian cancer cells, and the more exposed protein is an easier target for immunological attack.
Because it is an autologous cellular product it has to be prepared individually for each patient: the patient undergoes leukapheresis, her mononuclear cells are harvested and differentiated into dendritic cells, and those are cultured with a fusion protein of mannan, acting as an adjuvant, and modified mucin 1. The product must be injected back into the same patient the cells came from. The primed dendritic cells can then activate T cells to recognise the modified mucin 1 and kill any cells expressing it in any quantity, which it is hoped includes ovarian cancer cells.
As of March 2012, recruitment into a phase II study of about 60 patients had just finished, with a year or two still to run; a study that size was not expected to say much about efficacy, though it might hint at an effect if the product works well. The trial's statistician reported that initial safety results seemed very promising, with no sign so far of the toxicity that might be expected from conventional cancer chemotherapy, and called this the great advantage of a precisely targeted therapy over the sledgehammer approach of cytotoxic chemotherapy. A phase III study designed with advice from both the European Medicines Agency and the Food and Drug Administration was about to start. Whether the product prolongs remission, or works at all, was not yet known and awaited the results of the trials.
mRNA cancer vaccines
The mRNA vaccine has emerged as a novel weapon against cancer on the strength of preclinical studies and notable promise, and its success in the COVID-19 pandemic has been extrapolated to cancer treatment. An mRNA vaccine carries mRNA encoding a tumour antigen, packaged in lipid nanoparticles for cellular uptake; the host's cells produce the antigen, and the adaptive immune response then targets the cells expressing it. Administering such a vaccine relies on accurate antigen selection, so that treatment attacks the tumour precisely while side effects stay under control. Non-viral carriers have shown greater performance than viral ones. One study found that combining lipid nanoparticles with layers of fat can compactly store large numbers of mRNA molecules within each particle, the particles mimicking the size and delivery of viruses. When the vaccine is administered, the innate immune response is triggered, causing a rapid response by the body to destroy the cancer cells. Self-amplifying mRNA platforms aim to enhance the potency of tumour-antigen-encoding mRNA and sustain the immune response at lower doses.
The advantages claimed are rapid development, with mass production by high-yield in vitro transcription once the key sequence is identified; strong immunogenicity and high specificity for cancer cells; no risk of accidental infection and no need for viral vectors, plasmids or recombinant factors; sequences modifiable to encode unique antigens or cytokines for customised therapy, with molecular refinement able to reduce unwanted side effects and improve safety; and immunological memory against resurgent tumours. Against the alternatives, mRNA cancer vaccines are described as more easily preserved than DNA vaccines, much more stable than viral or bacterial vaccines, cheaper to produce than peptide-based vaccines and far more versatile across cancer types than dendritic cell vaccines.
The limitations include an extremely expensive manufacturing process and the logistics of ultracold storage, and to resolve them production methods must be simplified and account for temperature-stable designs to make the vaccines significantly more affordable. Inside the tumour microenvironment, regulatory T cells and cytokines inhibit the immune response and reduce the vaccine's efficacy, and combining the vaccine with immune checkpoint inhibitors might bypass that barrier. An optimal anticancer defence needs several immune cells working in concert, and because the microenvironment's immunosuppression encourages tumour growth and evasion of immune regulation, more than one dose may be needed to reach the desired potency. As of 2026, no route of administration (intravenous, intramuscular, subcutaneous or other) had been identified as best, and more preclinical trials were needed.
Combination with other treatments
Understanding of anticancer immunity gained over the five decades before 2018 has allowed a more educated design of immunotherapies and their use as monotherapy or, more recently, in combination with chemotherapy, radiotherapy and surgery. Patients with a pre-existing immune response tend to respond better to immunotherapies, which provides a rationale for combining them with more classical therapies such as chemotherapy and radiotherapy. These therapies kill tumour cells, liberating antigens that prime the immune system, and radiotherapy also induces an abscopal effect, an immune response outside the irradiated field. Combining immunotherapeutic agents that target distinct steps of the immune response, and at different time points, might also prove beneficial.
Immuno-oncology was in its infancy in 2018. Despite their proven efficacy, immunotherapies were effective only for certain patients, and which patients will benefit most and how to combine immunotherapeutics with traditional therapies remained open questions. Combination therapies and individualised therapy based on patients' predicted responses were among the challenges ahead.
History of cancer immunotherapy
Cancer immunotherapy may seem to have sprung from nowhere in the decade before 2019, yet it is the result of more than a century's intense study of immunology. It began in the late nineteenth century with the New York surgeon and cancer researcher William Coley, who noted cases of spontaneous remission of sarcoma in patients who had developed acute Streptococcus pyogenes infections and hypothesised that stimulating the immune system with a bacterial infection might be associated with a bystander anti-tumour activity that would result in tumour regression. From 1891 he treated mainly inoperable sarcoma patients with intra-tumoural injections of initially live and then inactivated Streptococcus pyogenes and Serratia marcescens, the so-called Coley's toxins. His 1891 injections of a mixture of live and inactivated bacteria into patients' tumours achieved complete remission of several types of malignancy. The approach fell into disuse, hampered by a cure rate of 10%, the absence of standardised toxin manufacturing, the lack of prospective clinical trials to evaluate safety and efficacy, and the arrival of radiation therapy and chemotherapy. One exception is still in use: intravesical injection of live bacille Calmette-Guérin as an efficient treatment for superficial bladder cancer.
Since Coley's procedure the concept of immune surveillance has emerged, and immuno-oncology originated as an approach to stimulate or restore the patient's immune response to cancer. Modern immunotherapy made its way gradually over the 50 years before 2018 as anti-cancer immunity came to be better understood. The editors of Science chose active immunotherapies as the breakthrough of the year for 2013.