The Immune
System

The human body is a warm, nutrient-rich, chemically stable environment operating at 37 degrees Celsius with a continuous supply of oxygen, glucose, and amino acids. From the perspective of a bacterium, a virus, a fungus, or a parasitic worm, it is among the most attractive environments on Earth. The surface area of the gut alone, unfolded, covers approximately 32 square metres. The respiratory tract processes approximately 10,000 litres of air every day, every litre of which carries thousands of microbial particles. The skin covers approximately 2 square metres and is breached by injury on a continuous basis throughout a lifetime.

The immune system exists to solve a problem that is simultaneously simple to state and extraordinarily difficult to execute: identify and destroy what should not be here, without destroying what should. This is the problem of immunological self-non-self discrimination, and its difficulty arises from a fundamental asymmetry. The things that should not be here, pathogens of every category, are extraordinarily diverse, mutate rapidly, and have been evolving to evade detection for as long as immune systems have existed. The thing that should be here, the self, must be protected with absolute reliability: any immune response that attacks the body's own tissues is potentially as dangerous as the pathogen it was meant to eliminate.

The solution evolution arrived at is not a single mechanism but a layered architecture of extraordinary sophistication, incorporating barrier defences, rapid non-specific responses, and a system of specific recognition and memory that generates targeted responses against threats it has never previously encountered, retains the memory of those responses for decades, and does all of this while maintaining precise tolerance of the body's own molecular landscape. No human engineer has produced anything approaching its capability.

The immune system is conventionally divided into two broad arms. The innate immune system is the older and faster arm: a set of mechanisms present in essentially the same form in all multicellular organisms, operating within minutes to hours of infection, using recognition receptors that detect broad classes of microbial features rather than specific pathogens. The adaptive immune system is the more recently evolved arm, present only in vertebrates, operating over days to weeks, and capable of generating exquisitely specific responses tailored to the precise molecular features of any pathogen it encounters. The two systems are not independent. They communicate continuously and the innate response provides the signals that shape and direct the adaptive response.

The innate system's recognition strategy depends on pattern recognition receptors (PRRs) that detect conserved molecular features of pathogens, structures that are present in bacteria, viruses, or fungi but absent from host cells. These features are called pathogen-associated molecular patterns (PAMPs). The lipopolysaccharide (LPS) in the outer membrane of gram-negative bacteria is a PAMP. The double-stranded RNA produced during viral replication is a PAMP. The beta-glucan in fungal cell walls is a PAMP. The innate system does not need to have encountered a specific bacterium before to respond to it. It responds to the entire class of bacteria simultaneously, because all gram-negative bacteria have LPS.

The discovery of the receptor family responsible for much of innate immune recognition, the Toll-like receptors (TLRs), was made by Jules Hoffmann in fruit flies and by Charles Janeway and Ruslan Medzhitov in mammals. Bruce Beutler identified in 1998 that a specific TLR, TLR4, was the mammalian receptor for LPS, solving a decades-old puzzle about why LPS caused the dramatic inflammatory response known as septic shock. Hoffmann and Beutler shared the Nobel Prize in Physiology or Medicine in 2011, together with Ralph Steinman who had discovered dendritic cells.

When a pattern recognition receptor detects its PAMP, the cell initiates a response that escalates rapidly. The cell activates transcription factors that drive the production of cytokines, small signalling proteins that are broadcast to surrounding tissues and the bloodstream, recruiting immune cells to the site of infection, increasing vascular permeability to allow immune cells to leave the blood and enter tissues, and raising body temperature, because many pathogens replicate less efficiently at elevated temperatures. The swelling, heat, redness, and pain that constitute the visible signs of inflammation are the macroscopic consequences of these molecular events. Inflammation is not a malfunction of the immune response. It is the immune response, the mobilisation of resources toward a site of infection or tissue damage, trading the disruption of local tissue function for the benefits of concentrated antimicrobial activity.

At the site of infection, the innate immune response deploys two particularly important mechanisms. The first is complement: a system of approximately 30 serum proteins that circulate in inactive forms and are activated in a cascade triggered by recognition of microbial surfaces. Activated complement proteins coat pathogens in a process called opsonisation, marking them for destruction. They recruit additional immune cells by releasing chemotactic fragments. In some cases they directly kill bacteria by assembling a structure called the membrane attack complex, which punches pores in the bacterial outer membrane and causes the bacterium to lyse.

The second mechanism is phagocytosis: the engulfment and destruction of pathogens by specialised immune cells. Ilya Mechnikov, working in Messina in 1882, observed that mobile cells in the transparent larvae of a starfish would surround and engulf foreign particles he introduced, including rose thorns, and argued that this cellular eating, which he called phagocytosis from the Greek for devouring cell, was a fundamental mechanism of defence. He was essentially correct, and he shared the Nobel Prize in Physiology or Medicine in 1908 with Paul Ehrlich, whose work on antibodies had developed the competing humoral theory of immunity. History vindicated both of them. The two mechanisms they championed, cellular and humoral immunity, turned out to be two components of a single integrated system.

Neutrophils and macrophages are the principal phagocytic cells of the innate response. Neutrophils are the most abundant white blood cells in the human body, constituting approximately 50 to 70 percent of all circulating white blood cells, and they are the first cells to arrive in large numbers at a site of bacterial infection. They engulf bacteria, destroy them with a combination of toxic reactive oxygen species, antimicrobial peptides, and digestive enzymes within specialised intracellular compartments called phagolysosomes, and die in large numbers in the process. Pus is largely the accumulated debris of dead neutrophils and bacteria.

Dendritic cells, discovered by Ralph Steinman in 1973, are the critical bridge between innate and adaptive immunity. Dendritic cells sample their environment constantly, phagocytosing material and processing the proteins they encounter into short peptide fragments. When a dendritic cell encounters strong innate immune activation signals, it undergoes maturation: it migrates to the nearest lymph node, presenting the peptide fragments from the pathogen it encountered on molecules called major histocompatibility complex (MHC) proteins on its surface. At the lymph node, these presented peptides are the information that activates the adaptive immune response.

Steinman was awarded the Nobel Prize in 2011. He died of pancreatic cancer three days before the award was announced, having participated in an experimental dendritic cell-based immunotherapy against his own cancer. He was unaware that he had won.

The adaptive immune system accomplishes something that has no parallel in any other biological or technological system: it generates the capacity to recognise any molecular structure it might ever encounter, including structures that have never existed in the history of life, before it has encountered any of them. The lymphocyte repertoire, the full set of B cells and T cells in the human body, contains an estimated 10¹² cells, each one carrying a receptor with a unique binding specificity. The diversity of this repertoire is so vast that for essentially any molecular structure that can be synthesised, there exists a B cell or T cell in the body whose receptor will bind it with some affinity.

How is this diversity generated? Not by having 10¹² different receptor genes. The genome is not large enough. Instead, receptor diversity is generated by a process of controlled DNA rearrangement called V(D)J recombination, discovered by Susumu Tonegawa, who received the Nobel Prize in Physiology or Medicine in 1987 for this work. In each developing lymphocyte, the receptor gene is assembled from a random combination of gene segments drawn from three pools designated V (variable), D (diversity), and J (joining). The human antibody heavy chain gene has approximately 40 functional V segments, 25 D segments, and 6 J segments. Any combination is possible. The joining process is also imprecise: nucleotides are randomly added and removed at the junctions between segments, adding a further layer of diversity. The result is that each lymphocyte carries a unique receptor sequence that was assembled de novo in that cell and in no other, and that did not exist in the genome of the organism's parents.

B cells produce antibodies: soluble proteins that circulate in the blood and lymph, binding to their target antigens with the same specificity as the receptor from which they derive. T cells do not secrete their receptors but instead act through direct cell-to-cell contact. Cytotoxic T cells (CD8+ T cells) recognise and kill infected cells that are displaying fragments of pathogen proteins on their surface. Helper T cells (CD4+ T cells) produce cytokines that amplify both B cell and cytotoxic T cell responses, acting as coordinators of the adaptive response.

The system generates diversity first and tests utility afterward. Out of 10¹² possible receptors, only a tiny fraction will ever encounter a pathogen their receptor can recognise. The vast majority of lymphocytes live and die without ever being activated. This is an extraordinarily wasteful strategy from an engineering perspective. From an evolutionary perspective, it is the only strategy that works, because the diversity of potential pathogens is so vast that no system that generated receptors in response to specific pathogens could be fast enough.

The theory of clonal selection, proposed by Frank Macfarlane Burnet in 1957, is the conceptual foundation of adaptive immunity. Burnet proposed that each lymphocyte carries receptors of a single specificity, that encounter with an antigen selects those lymphocytes whose receptors bind it, causing them to proliferate and differentiate into effector cells, and that some of these activated cells persist as memory cells that provide faster and stronger responses upon subsequent encounters with the same antigen. Burnet shared the Nobel Prize in Physiology or Medicine in 1960 with Peter Medawar, whose work on transplantation tolerance demonstrated the complementary principle of self-tolerance.

The clonal selection mechanism explains vaccination. When a vaccine introduces an antigen, whether a killed pathogen, a protein subunit, or an mRNA encoding a pathogen protein, the adaptive immune system responds as it would to a genuine infection: it selects and expands the lymphocytes whose receptors bind the antigen, generates effector cells that produce antibodies or cytotoxic T cell responses against the antigen, and then, after the antigen is cleared, retains a population of long-lived memory cells carrying the same receptor specificity. When the vaccinated individual subsequently encounters the actual pathogen, the memory cells respond within hours rather than the days required for a primary response, at a magnitude orders of magnitude greater, and typically eliminate the infection before it can establish itself.

The affinity of antibodies produced during an immune response increases over time through a process called affinity maturation. Activated B cells in lymph nodes undergo rapid cell division and, during each division, introduce mutations into their antibody genes at a rate approximately 10⁶ times higher than the background genomic mutation rate. This process, called somatic hypermutation, generates variants of the antibody with slightly different binding properties. B cells whose mutant antibodies bind the antigen more tightly are preferentially selected for survival and further division. B cells whose mutations reduce antigen binding die. The result is a progressive increase in the affinity of the antibody response over the course of an infection, a Darwinian process of selection operating within the immune system on a timescale of days.

Every nucleated cell in the body continuously samples its own protein content, degrading some fraction of its proteins into short peptide fragments and displaying them on MHC class I molecules on the cell surface. This display is the mechanism by which cytotoxic T cells survey the interior of cells they cannot enter directly. A cell infected by a virus will have viral proteins among its displayed peptides. A cancerous cell may have mutant proteins. The cytotoxic T cell patrols the surface of every cell it encounters, reading the displayed peptides through its T cell receptor. If it finds a non-self peptide, it kills the displaying cell.

This system requires that cytotoxic T cells reliably ignore self peptides while responding to non-self peptides with lethal force. This tolerance is established during T cell development in the thymus, in a process called central tolerance. Developing T cells are first tested for the ability to bind self-MHC molecules at all. Those that cannot bind self-MHC are eliminated: they would be useless. Those that bind self-MHC appropriately are then tested for the tendency to respond strongly to self-peptides. Those whose receptors bind self-peptides too strongly are eliminated by negative selection through a process called clonal deletion. Only cells that bind self-MHC appropriately but do not strongly react to self-peptides survive thymic selection and are exported to the periphery.

The gene encoding AIRE (autoimmune regulator), identified in 1997, is expressed in thymic epithelial cells and drives the promiscuous expression of tissue-specific proteins in the thymus, proteins normally found only in the pancreas, the eye, the thyroid, and elsewhere in the body. This ectopic expression allows the thymus to present peptides from essentially every tissue to developing T cells, ensuring that T cells reactive to any self-protein are deleted before they reach the periphery. Mutations in AIRE cause autoimmune polyendocrinopathy syndrome type 1, a severe multi-organ autoimmune disease, demonstrating that continuous AIRE function is required to maintain self-tolerance in humans.

The MHC gene region on chromosome 6 is the most polymorphic region of the human genome. There are hundreds of different alleles at each MHC locus across the human population, each allele encoding an MHC protein with slightly different peptide-binding properties. Population-level MHC diversity is a collective immune resource: no single pathogen is likely to evade recognition in all individuals simultaneously.

An antibody, or immunoglobulin, is a Y-shaped protein consisting of two identical heavy chains and two identical light chains, held together by disulfide bonds. The two tips of the Y-shaped molecule are the antigen-binding domains, each formed by the variable regions of one heavy and one light chain folding together into a structure whose precise shape is determined by the sequence generated by V(D)J recombination. The stem of the Y, the Fc region, is constant within each antibody class and mediates effector functions by binding to Fc receptors on other immune cells and to the complement system.

There are five classes of antibody, designated IgM, IgG, IgA, IgE, and IgD, each with different structural features and different biological roles. IgM is the first antibody produced in a primary immune response, a large pentameric structure particularly effective at activating complement. IgG is the most abundant antibody in serum, the antibody class responsible for long-term protection and for the maternal-to-foetal antibody transfer that provides newborns with immune protection during their first months. IgA is secreted across mucosal surfaces, including in breast milk, saliva, tears, and the gut lining, providing immune protection at the body's interfaces with the external environment. IgE mediates allergic responses and protection against parasitic worms, binding with extraordinarily high affinity to its antigen and triggering the release of inflammatory mediators from mast cells.

Antibodies neutralise pathogens by binding to the molecular features the pathogen uses to enter cells or cause damage, physically blocking those interactions. An antibody bound to the surface protein a virus uses to dock with its target cell prevents the virus from infecting that cell. The same antibody simultaneously neutralises, opsonises, and activates complement.

Paul Ehrlich's concept of the magic bullet, a molecule that seeks out and destroys a pathogen with the precision of a targeted weapon while leaving the host unharmed, was his vision for the future of medicine. Antibodies are the immune system's own magic bullets, and monoclonal antibody therapies have become one of the dominant classes of therapeutic agents in modern medicine, used to treat cancers, autoimmune diseases, infectious diseases, and inflammatory conditions.

Self-tolerance is not perfect. The same mechanisms that enable the immune system to generate vast receptor diversity and then select for useful specificities while eliminating self-reactive ones cannot achieve complete elimination of all self-reactive lymphocytes in all individuals under all circumstances. When self-tolerance fails, the immune system attacks the body's own tissues. This is autoimmunity.

Autoimmune diseases collectively affect approximately 5 to 8 percent of the population in developed countries, with prevalence rising over recent decades in ways that cannot be explained by genetic change and must therefore reflect environmental influences. Rheumatoid arthritis involves immune attack on the synovial tissue of joints. Multiple sclerosis involves attack on the myelin sheaths of neurons in the central nervous system. Type 1 diabetes involves destruction of the insulin-producing beta cells of the pancreas. Systemic lupus erythematosus involves antibodies against the body's own DNA and nuclear proteins. In each case the immune system is doing precisely what it is designed to do: mounting a specific, sustained, adaptive response against a target. The catastrophic difference is that the target is the self.

The genetic basis of autoimmunity is substantially concentrated in the MHC region. Multiple autoimmune diseases are associated with specific MHC alleles: ankylosing spondylitis with HLA-B27, rheumatoid arthritis with HLA-DR4, type 1 diabetes with HLA-DR3 and HLA-DR4. These associations reflect the fact that different MHC alleles present different peptide repertoires to developing T cells during thymic selection, with some alleles apparently permitting the survival or activation of self-reactive T cells that other alleles would delete or fail to activate.

The hygiene hypothesis, proposed by David Strachan in 1989 based on epidemiological observations about the inverse relationship between childhood infections and later development of allergic diseases, has been substantially developed into the old friends hypothesis by Graham Rook. The core proposition is that the immune system evolved in an environment containing a rich array of microbial exposures that have been largely eliminated in high-income countries over the past century. The regulatory circuits of the immune system, particularly those mediated by regulatory T cells (Tregs), may require these exposures to be properly calibrated. In their absence, the immune system may be inadequately regulated and prone to misdirected responses against self-antigens or harmless environmental substances.

The human immunodeficiency virus (HIV) is the most extensively studied virus in the history of medicine, partly because of the scale of the epidemic it caused and partly because of what it reveals about the architecture of the immune system by systematically dismantling it.

HIV targets CD4+ helper T cells with the specificity that follows from the virus using the CD4 protein as its primary receptor for cell entry, along with a co-receptor, either CCR5 or CXCR4, depending on the viral strain. Helper T cells are the coordinators of the adaptive immune response. They provide the signals that activate cytotoxic T cells, that drive B cell maturation and antibody class switching, and that amplify the innate response. When helper T cells are lost, every arm of the adaptive immune response is progressively weakened.

HIV replicates using a reverse transcriptase enzyme that converts the viral RNA genome into DNA, which is then integrated into the host cell's genome by the viral enzyme integrase. The integrated viral DNA, or provirus, can remain transcriptionally silent in resting T cells for years, invisible to the immune system, while the virus persists in the body as a reservoir that cannot be cleared by current therapies. This latent reservoir is the central obstacle to HIV cure.

The discovery that individuals with homozygous loss-of-function mutations in the CCR5 gene are strongly resistant to HIV infection provided the rationale for a remarkable therapeutic event: the Berlin Patient, Timothy Ray Brown, who was treated for leukaemia in 2007 with a bone marrow transplant from a donor homozygous for the CCR5 deletion, and whose HIV infection was subsequently undetectable without antiretroviral therapy. The reconstitution of his immune system from cells lacking CCR5 removed the HIV reservoir along with his leukaemia. As of 2020 he was considered the first person cured of HIV. He died later that year from a recurrence of his leukaemia.

The immune system recognises and destroys cancerous cells as part of its normal function. Cytotoxic T cells that detect mutant proteins, the neoantigens produced by cancer-specific mutations, on the surface of tumour cells can kill those cells before a tumour becomes clinically detectable. This process of immune surveillance is one reason that immunosuppressed individuals have substantially elevated risks of certain cancers.

But cancers that establish themselves have typically evolved mechanisms to evade immune recognition or suppress immune responses. One of the most important of these mechanisms involves immune checkpoints: regulatory signals on T cell surfaces that normally prevent T cells from becoming overactivated and causing collateral damage to normal tissue. The PD-1 receptor on T cells, when bound by its ligand PD-L1, sends an inhibitory signal that reduces T cell activity. Many tumour cells overexpress PD-L1, effectively displaying a molecular stop signal that prevents the cytotoxic T cells that recognise them from killing them.

James Allison at the University of Texas and Tasuku Honjo at Kyoto University identified, independently, the CTLA-4 and PD-1 checkpoints respectively, and proposed that blocking these checkpoints with antibodies would release the brakes on the immune response against tumours. Allison pursued the development of anti-CTLA-4 antibodies through clinical trials in the face of substantial institutional scepticism. The first checkpoint inhibitor drug, ipilimumab, was approved for the treatment of melanoma in 2011. Durable long-term remissions were observed in patients with metastatic melanoma who had previously had a median survival measured in months. Allison and Honjo shared the Nobel Prize in Physiology or Medicine in 2018. Checkpoint inhibitor therapies are now used to treat more than a dozen cancer types and have produced durable remissions in patients with previously untreatable advanced cancers.

The immune system, understood precisely, redefines what the self is.

The everyday sense of self is continuous, bounded, and stable. The self is here and everything else is not here. The immune system enforces a different and more precise version of this boundary, and in doing so reveals how contingent, how molecularly specified, and how actively maintained the distinction between self and non-self actually is.

The cells of the body display a continuous molecular census of their protein content, peptide by peptide, on their surface. Every cell announces what it is made of. The T cells that patrol these displays are themselves the products of a developmental selection process that eliminated any cells that would attack the displays they are reading. The tolerance of self is not passivity. It is an active, continuously maintained state, dependent on the continued function of molecular mechanisms that can fail, that evolve, and that vary between individuals in ways that have measurable consequences for health and disease.

The 100 trillion synaptic connections of the human brain do not know what the immune system knows. The brain's sense of a unified continuous self has no access to the 10¹² lymphocytes quietly patrolling the bloodstream, each one carrying a unique receptor, each one part of a surveillance apparatus of extraordinary sophistication that has been evolving for over 500 million years. The self the immune system maintains is not the same self that consciousness experiences. It is older, more precise, and considerably harder to fool.