Morbidity and mortality from infectious diseases can be caused either by direct damage to the host by the pathogen or by collateral damage to host tissues by the immune response to the pathogen. This collateral damage is referred to broadly as immunopathology and can result from overproduction of inflammatory signals by immune cells.

Mammalian hosts employ two interconnected systems—innate and adaptive immunity—to protect themselves from infection while minimizing immunopathology. We are only beginning to understand how these two systems are coordinated to maintain this delicate balance. It is generally thought that innate immunity combats infection immediately, whereas adaptive immunity reacts only after a delay of several days. This suggests that adaptive immunity should not influence the early innate response. In this issue of Nature Medicine, however, Kim et al.1 reveal that T cells of the adaptive immune system actively suppress the cells of the innate immune system to prevent an overzealous early innate response and severe immunopathology.

Unlike invertebrates, which rely exclusively on innate immunity, mammals require both innate and adaptive immunity for an effective host response to infection. As the first line of defense, the innate immune system senses infection through pattern-recognition receptors, which recognize conserved molecular features of pathogens that are unique to microbial life forms2. These pattern-recognition receptors, such as the Toll-like receptors (TLRs), trigger a variety of antimicrobial responses to combat the infection. When the innate immune system is unable to contain an infection, the cells of the adaptive immune system step in as a second line of defense.

T and B lymphocytes of the adaptive immune system use randomly generated antigen receptors and, once activated, maintain a long-term memory of previously encountered pathogens3. These lymphocytes, however, cannot reliably distinguish 'self' from 'non-self', and so they rely on the innate immune system for instructions on when and how to respond to infection2. In turn, activated T and B cells further activate and direct innate defenses: T helper 1 (Th1) cells activate macrophages, Th2 cells activate eosinophils, and antibodies produced by B cells activate the complement pathway, phagocytosis and mast-cell degranulation.

When either the innate or the adaptive immune system is compromised, the host is unable to combat microbial infection or control endogenous microflora4,5,6. Thus, maximal immunity is achieved only when innate and adaptive immunity work together to combat infection. For instance, mice unable to mount an adaptive immune response, such as nude mice that lack T cells or Rag-deficient mice that lack all lymphocytes, succumb rapidly to infections that would normally be cleared in wild-type animals4. It was long-assumed that these mice died because of unchecked microbial growth in the absence of adaptive defenses. Kim et al.1 set out to test this assumption and found that it does not always hold true.

Kim et al.1 inoculated nude mice with a normally sublethal dose of the coronavirus mouse hepatitis virus (MHV) and observed the expected high rate of lethality. However, upon examination, these mice had only negligible increases in viral load and virus-induced tissue pathology, suggesting that they did not die from an overwhelming infection. Instead, when the authors measured cytokine levels in these mice, they found that the levels of interferon-γ (IFNγ) and tumor necrosis factor-α (TNFα) were drastically increased, suggesting that the mice died from damage caused by the high amounts of inflammatory cytokines (cytokine storm) released by cells of the innate immune system.

To confirm that this lethality was caused by a cytokine storm and not by the infection per se, the authors1 stimulated the antiviral immune response without introducing virus into the animals by injecting the synthetic TLR3 ligand poly(I:C), which mimics viral double-stranded RNA. This ligand can induce lethal immunopathology via a cytokine storm in wild-type mice, but in Rag-deficient mice, even a normally sublethal dose of poly(I:C) caused a very rapid death. Antibodies to TNFα prevented the poly(I:C)-induced death, indicating that the cytokine storm caused the lethality.

Because Rag-deficient mice lack both T and B cells, Kim et al.1 went further to show that it is the T cells that suppress inflammatory cytokine production. Mice that had been depleted of T cells also showed high levels of cytokines after having been given poly(I:C), and nude mice into which lymphocytes had been adoptively transferred had reduced levels of cytokines. In total, these results revealed an unexpected negative regulation of the early innate response by the adaptive immune system and suggested that T lymphocytes are necessary and sufficient to suppress an overzealous innate immune response (Fig. 1).

Figure 1: Conventional T cells suppress overzealous early innate responses, thus preventing severe immunopathology.
figure 1

Kim Caesar

In response to infection or to purified pathogen-associated molecular patterns, TLRs on macrophages and dendritic cells (DCs) of the innate immune system are activated, inducing the production of inflammatory cytokines, such as TNFα. (a) In wild-type mice, conventional T cells of the adaptive immune system suppress early inflammatory cytokine production by innate cells in a contact- and MHC class II–dependent manner; regulatory T cells can also suppress innate cytokine production similarly (not shown). The precise mechanism of suppression, however, is unclear. (b) Nude mice, which are deficient in T cells, or Rag-deficient mice, which lack all adaptive immunity, are unable to control the early innate response to infection or to pathogen-associated molecular patterns. In the absence of T-cell–mediated regulation of innate immunity, an overzealous early innate response characterized by the overproduction of TNFα can lead to severe immunopathology and death.

Regulatory T cells (Treg cells) inhibit both innate and adaptive immune responses7 and are the obvious candidates for the suppressors of the lethal cytokine storm. However, both Treg cells and conventional T cells were able to repress poly(I:C)-induced cytokine production by cells of the innate immune system in vitro1. This suppression was dependent on direct contact between T cells and cells of the innate immune system, as well as on the antigen-presenting molecule major histocompatibility complex (MHC) class II (Fig. 1). Nonetheless, the precise mechanism by which conventional T cells suppress the lethal cytokine storm, including whether or not they use the same mechanisms used by Treg cells to regulate innate immune responses, will need to be investigated further.

The realization that adaptive suppression of early innate immunity is necessary to maintain the balance between immunity and immunopathology has important implications for our understanding of immune regulation in health and disease. In spite of this, some of the most interesting implications of the study by Kim et al.1 relate not only to the current mammalian immune system but also to the evolutionary history of modern mammalian immunity.

Innate immunity alone is sufficient for host defense in invertebrates, yet mammals require both innate and adaptive immunities, indicating that the advent of adaptive immunity may have altered the innate immune system in several ways. The evolution of an adaptive immune response has allowed vertebrate animals to minimize immunopathology by specifically targeting host defenses to pathogens, and it has prevented repeated infection with commonly encountered pathogens through the formation of immune memory4.

Because it provided these distinct advantages, the vertebrate development of adaptive immunity probably caused drastic changes in the way immune tasks were both delegated and executed. The study by Kim et al.1 suggests that one such change is the tempering of the early innate response by adaptive lymphocytes—an alteration that might have arisen to maximize the benefits gained from engaging the adaptive immune system.

Kim et al.1 have unmasked one reason why mice lacking adaptive immunity do not survive a normally sublethal pathogen challenge, revealing a new regulatory relationship between adaptive and innate immunity. Because the innate immune response precedes the adaptive immune response to infection by several days, one would assume that adaptive immunity should not affect the early innate response, but the findings of Kim et al.1 show that even the earliest innate response requires adaptive regulation.

It seems that the coevolution of innate and adaptive immunity is a story that began with cooperation and has ended in codependence. The adaptive immune system appears to have evolved ways to regulate the early innate immune response in an effort to minimize immunopathology and maximize host defense. Now accustomed to this level of control, the innate immune system can no longer properly regulate its own response in the absence of adaptive suppression. In the future, it will be fascinating to learn how conventional T cells suppress innate immunity and to determine the importance of this suppression in the proper regulation of immune response and resolution of pathogen threats.