To summarize our understanding of host defense mechanisms against influenza (Table 1):
Over 50 years ago, Clayton Loosli showed that serum antibody protected mice against influenza viral pneumonia. Flu shots in humans are designed to maximize serum antibody. They greatly reduce mortality but are variable in their ability to protect against illness. In order to understand this variability, our laboratory has long tried to clarify the relative roles of local and systemic antibody in preventing and promoting recovery from both upper and lower respiratory tract infection with the influenza virus.
The role of systemic humoral and cell-mediated immunity in preventing influenza was examined in a ferret study done by W. Henry Barber. Ferrets are susceptible to the flu. They also have a long trachea, which makes it surgically possible to create an anatomically isolated tracheal pouch. The pouch enabled us to study responses to infection of the nose or trachea separately. In the critical experiment, ferrets were infected in the pouch with a strain of A (H1N1) influenza virus. The infection did not spread beyond the pouch. Three weeks after recovery from the infection, when the animals exhibited peak serum antibody levels against the virus, they were rechallenged with the identical virus in both the pouch and nose. The pouch was immune from infection, but rhinitis developed in all animals and they shed virus from the nose. The duration of nasal infection, however, was half that seen in previously uninfected animals. The results are consistent with the finding that passive immunization of mice with specific antisera before infection with a lethal influenza virus challenge prevents pneumonitis and death but does not prevent rhinotracheitis.
The chief conclusion to be drawn from those observations is that serum antibody and (presumably) cell-mediated immunity do not protect the upper respiratory tract from influenza infection in ferrets or mice. One would not expect T cells to neutralize virus in the nose or lungs but rather to destroy infected cells - which, by definition, is recovery and not prevention. In fact, it has been shown that adoptive immunization with anti-influenza T cells does not prevent influenza infection in mice.
We next sought to determine whether such observations were relevant to humans. An epidemic of H1N1 influenza occurred in the United States in 1979, after being absent for 26 years. Therefore, mothers less than 26 years old had no antibody against H1N1 virus, whereas most of those over 26 did. This experiment of nature made it possible for Peter Reuman, Elia Ayoub, and me to study neonates during the epidemic.
Babies with passive maternal antibody to influenza virus had an infection rate that was indistinguishable from those born of mothers without antibody-although babies with maternal antibody had less severe involvement of the lower respiratory tract. The number of babies studied was small. but the data were consistent with those in the ferret and mouse experiments: Serum antibody and cell-mediated immunity have no role in preventing upper respiratory tract infection with influenza virus. But if serum antibody does not prevent upper respiratory infection, what does?
Secretory IgA has been postulated for that role. The postulate is based on observations by (among others) Robert Chanock, at the National Institute of Allergy and Infectious Diseases, of a correlation between respiratory tract IgA concentrations and protection against infection with various pathogens. including influenza virus.
There are only two ways to establish a cause-and-effect relationship between a component of the immune system and prevention of, or recovery from, an infection. One is selectively to add an immune-system component to a nonimmune animal and show that the component confers immunity to subsequent challenge with the infectious agent. The other is selectively to ablate an immune-system component and show that a challenged host cannot develop immunity or that a previously immune host has lost that immunity.
Kathryn Renegar in the University of Florida labs has recently demonstrated a causal relationship, not just a correlation, between IgA antibody against influenza and protection. She showed that a monoclonal polymeric anti-influenza IgA was selectively transported into nasal secretions when injected intravenously into mice. The antibody prevented nasal infection with the influenza virus, whereas comparable amounts of IgG virus-neutralizing antibody did not. She then showed that the protective effect was ablated by antibody against IgA (i.e., "anti-antibody") administered in nose drops along with the virus. Her most important observation for understanding host defense was that immunity acquired by prior infection could be ablated in convalescent mice by administration of nose drops containing anti-IgA antibody but not anti-IgG, anti-IgM antibodies, or both. Thus, secretory IgA is the primary murine host defense against upper respiratory influenza infection.
Immune mechanisms in recovery are often different from those in prevention. Cell-mediated immunity is not relevant to prevention but is essential for permanent recovery. Conversely, serum antibody helps with recovery of the upper respiratory tract, even though it is not involved in preventing infection of that site. These facts were established by two sets of experiments in nude mice, animals unable to generate cell-mediated immunity.
In one experiment done by Richard Kris in the University of Florida labs, the effect of passive serum antibody on influenza infection in nude mice was examined. The animals were infected with a nonlethal Influenza virus. Their tracheas desquamated, the mice looked sick, and they shed virus continually for at least two months. They were unable to recover. When infected nude mice were given serum antibody, virus shedding decreased, and there was complete regeneration of the tracheal epithelium. However, a month later, when serum antibody was no longer detectable, large amounts of virus were again shed from nose, trachea, and lungs, and the tracheas once again desquamated. We concluded that serum antibody contributed to temporary recovery but was not sufficient to cure the infection. On the other hand, Walter Gerhard and his group at the Wistar Institute have found that IgG antibody can lead to permanent recovery.
Definitive experiments in a number of laboratories (by Brigitte Askonas at John Radcliffe Hospital, Oxford. England; Thomas Braciale at Washington University, St. Louis; Gordon Ada at the John Curtin School of Medical Research, Australian National University. Canberra; and Francis Ennis at the University of Massachusetts) showed that adoptive transfer of virusspecific cloned T cells in influenza-infected nude mice cleared virus from the lungs and induced recovery. Such studies have established that cell-mediated immunity is essential for recovery from influenza.
An age-associated decline in cell-mediated immunity may contribute to the high mortality associated with influenza infection in the elderly. Although not a universal phenomenon, cell-mediated immunity does tend to wane with age, in mice as well as in humans. Several investigators have shown that cytotoxic-T-cell activity is decreased in influenza-infected senescent mice, whereas antibody responses remain relatively intact.
We found that following influenza challenge, aged mice had prolonged viral shedding and that this shedding was correlated with lower anti-influenza (cytotoxic T-lymphocyte) CTL activity. Because influenza A viruses go through periodic antigenic shifts, an influenza virus a person is infected with when old is has different surface antigens than the influenza virus with which he was infected with while young. To test the hypothesis that increased severity of influenza could be due to loss of heterotypic immunity with aging, mice made heterotypically immune with H3N2 virus were then challenged under anesthesia with H1N1 virus. Viral clearance was delayed by at least two days in the aged mice. Naive and heterotypic immune mice were next challenged with H1N1 virus while awake. Under these conditions, initial infection was restricted to the nose in all animals. The virus spread to the lungs of the young and aged naive mice but not of heterotypic immune young mice. The lower heterotypic immunity of aged mice allowed the spread of the virus to the lungs. The longer duration and increased spread to the lungs again correlated with decreased anti-influenza CTL activity.
Although the study showed a correlation between prolonged viral shedding and poor cell-mediated immunity in the elderly, a causal relationship has not yet been proved. Nevertheless, the results suggest that poor cell-mediated immunity is associated with prolonged influenza infection, which in turn is associated with an increased risk of lung infection by both influenza virus and bacteria. A similar phenomenon may be induced in patients whose cell-mediated immunity is compromised by HlV infection.
To summarize our understanding of host defense mechanisms against influenza (Table 1):