The body remembers its insults. After a viral infection, memory T cells scatter throughout the body, ready to launch an immediate attack against reinfection in the gut, lungs, skin and other sites. These peripheral T cells stay at their posts for months; sometimes years. We now have a better idea of what keeps them there.

In the February issue of Immunity, Ray et al.1 show that the adhesion molecule very late antigen-1 (VLA-1; also called α1β1-integrin) enables the retention and survival of CD8+ T cells directed against influenza A in the lung. Their results have implications for designing vaccines against this and other hard-to-beat viruses. The findings should also spur thinking on how to hinder peripheral T cells in diseases where memory T cells are out of control, such as arthritis or diabetes.

There are two sides to T-cells. Acute T-cell responses during viral infection help clear the virus, but they also mediate lethal pathology, such as pneumonia in the case of influenza A or meningitis in the case of lymphocytic choriomeningitis virus (LCMV)2,3,4.

Rapidly mobilized memory T-cell responses can be particularly important in mediating protective immunity upon reexposure to heterosubtypic influenza virus strains—strains with different coat proteins that evade neutralizing antibody responses. Humoral immune responses target the viral surface proteins neuraminidase and hemagglutinin, sites of frequent mutations that are not cross-protective and that define the new heterosubtypic strains of influenza that develop each year. Memory CD8+ T-cell responses, on the other hand, target the more conserved internal viral epitopes—matrix and nucleoprotein—and thus can protect against heterologous strains.

After infection, T cells disperse throughout the host not only into lymphoid organs but also, at an even greater frequency, into nonlymphoid organs5,6,7,8. Most memory T cells in peripheral sites, unlike T cells in lymphoid organs, retain the characteristics of effector lymphocytes, such as immediate cytotoxic function and cytokine production, suggesting that they are instantly active against an invading pathogen. Thus, peripheral memory T cells are an ideal early first line of defense against an invading pathogen. In fact, studies have shown that immunity to a heterosubtypic strain of influenza A, which has no cross-neutralizing antibodies but identical T-cell epitopes, disappears about six months after influenza A infection, after antigen-specific memory T cells in the lung have declined9. Other studies on the LCMV infection model have shown that after the decline of peripheral effector-type memory T cells functional memory is also compromised to a peripheral challenge, despite intact central lymphoid memory10.

Very little is known about the factors that retain these activated memory T cells in the peripheral compartments. Peripheral memory T cells seem to be more resistant to apoptosis than lymphoid memory cells11, and they do not readily leave peripheral sites such as the lung and return to the circulation12.

Earlier work on VLA-1 showed that it may be intricately involved in T-cell immunity13,14. For instance, viral infections can increase the expression of VLA-1 and other adhesion molecules on many leukocyte populations, and can influence T-cell activation. The ligands for VLA-1 are collagens I and IV, which populate the extracellular matrix (ECM) of all peripheral organs. For example, collagen fibrils make up 15% of the lung by dry weight. Thus, T cells migrating through peripheral tissues are constantly in contact with collagen. Histological examination shows that lymphocytes localize near epithelial or endothelial boundaries near the basement membrane, which is rich in type IV collagen.

Interestingly, the ECM is separated from the T-cell compartments by the fibroblastic reticular network, so cells in the spleen and lymph nodes do not come in contact with the ECM. This could affect their interactions with antigen-presenting cells. In the presence of collagen, as in peripheral tissues, T cells form only transient interactions with antigen-presenting cells. This would favor the activation of memory T cells over naive T cells, which require prolonged periods of contact. Thus, it is possible that naive cells are best activated in the lymphoid compartment, because in the absence of collagen contact they are free to interact with dendritic cells for the prolonged periods required for T-cell activation.

In vitro studies have shown that VLA-1 modulates T-cell activation, as VLA-1 interaction with collagens I and IV enhances cytokine production and proliferation triggered through the T-cell receptor. Antibody blocking studies implicate VLA-1 in several in vivo models of inflammation mediated by T cells, including LCMV-induced delayed-type hypersensitivity responses and collagen-induced arthritis13,14.

Ray et al. showed that during acute influenza A infection, virus-specific CD8+ T cells increased expression of VLA-1 and were preferentially retained in peripheral organs, including regions of the lung airways and tissue (bronchioalveolar lymphocytes and lung parenchyma; Fig. 1). VLA-1 expression on T cells, however, was not required for the initial trafficking into peripheral tissues. Histological staining revealed that VLA-1+CD8+ T cells congregated at sites of high ECM content near bronchioles and vessels, suggesting that these cells were interacting with their ligands, collagens I and IV.

Figure 1: The role of VLA-1 in retention and survival of virus-specific memory CD8+ T cells in the lung.
figure 1

Renee Lucas

After activation of naive, virus-specific CD8+ T cells in the lymphoid compartment, Ray et al. show that VLA-1 expression is enhanced on T cells. These VLA-1+ T cells migrate into peripheral sites such as the lung and bind to their ligands, collagens I and IV, in the extracellular matrix (ECM) on epithelial cells of the lung airways. This interaction results in retention and increased survival of virus-specific memory T cells in the lung. These cells are then readily available as a first line of defense against reinfection with the virus. APC, antigen-presenting cell; TCR, T-cell receptor.

Full size image

These T cells in contact with ECM through VLA-1 seemed to have features of an effector phenotype, expressing high CD69 and low CD62L. They were also less prone to apoptosis than VLA-1− antigen-specific cells, as shown by TUNEL staining. These results suggest that the interaction of VLA-1 on memory T cells with its ligand in the ECM enhances both survival and retention of these cells in the lung.

The authors went on to show that VLA-1 is crucial for protective immunity to heterologous influenza virus infection. That was made clear by using a blocking antibody to VLA-1 and by analyzing mice genetically lacking VLA-1. Interestingly, although VLA-1 knockout mice had more influenza A–specific CD8+ memory T cells in the spleen than normal mice, they had no protective immunity.

What is not yet clear is exactly how VLA-1 enhances cell survival and maintains the effector phenotype. Does VLA-1 initiate an intracellular signaling process, or does attachment to collagen somehow protect these cells from other, potentially lethal, cellular interactions? The mechanism behind the decline of memory T cells over time also remains to be identified. Memory T cells may ultimately lose VLA-1 expression and return to the circulation, or the VLA-1/collagen interaction may provide only limited protection from apoptosis.

This study could enhance efforts to develop immunotherapy against T-cell-mediated diseases and create new vaccines. Treatments that interfere with the interaction of T-cell VLA-1 with collagen may diminish the immunopathological symptoms of diseases like asthma and rheumatoid arthritis.

Using the newly developed, intranasally administered, live attenuated influenza vaccine that also stimulates cellular immune responses directly in the mucosa of the respiratory tract could have added benefits over more conventional vaccines. The conventional killed influenza vaccine is administered intramuscularly and relies on the induction of humoral immune responses; the vaccine does not provide protection against new heterosubtypic influenza strains.

The live attenuated vaccine, only recently introduced into clinics, could make the difference between life and death for some individuals. A live vaccine has the benefit of more conserved, cross-protective CD8+ T cell responses that could provide at least partial protection against new emerging strains. The new avian influenza is one such strain; it has the potential to cause the next influenza pandemic, and is the focus of intensive vaccine development efforts.

A note of caution is in order: repeated doses of a vaccine that maintains high numbers of memory T cells in the lung could also amplify immunopathology after repeated infection by heterologous viruses15.