Annals of the Royal College of Surgeons of England (I975) vol 56
Immunity
in
acute bacterial infections
A A Glynn MD FRCP FRCPathi Professor in Bacteriology, Wright-Fleming Institute, St Mary's Hospital Medical School, London
Summary A brief survey of the ways in which humoral and cellular immune responses deal with bacteria and their products shows that a variety of effector mechanisms is involved. Their efficiency varies, partly in relation to the kind of bacterial attack with which they have to deal. Some bacterial factors though not crudely toxic can interfere with specific host defences. Introduction 'Find the germ of the disease; prepare from it a suitable anti-toxin; inject it three times a day quarter of an hour before meals; and what is the result? The phagocytes are stimulated; they devour the disease; and the patient recovers-unless, of course, he's too far gone.' SIR RALPH BLOOMFIELD BONINGTON Although it is over sixty years since Bernard Shaw1 wrote his accurate though wholly misleading account, immunization and stimulation of the phagocytes are still topical, while the possibilities of renal transplantation and home dialysis have made 'The Doctor's Dilemma' even more acute. Not many of the infections I am going to talk about have any direct surgical significance, but since Erasmus Wilson, who was a Vice-President of the Royal College of Surgeons in I879
and i 88o and President in i 88 I, was in fact a dermatologist you will perhaps excuse a broadly medical and somewhat academic approach. Surgeons, of course, have always been concerned with infection. Infection may follow obstruction which needs surgical relief; abscesses may need drainage; foreign bodies in tissues are a common precipitating cause and may have to be removed by the surgeon, even if he put them there in the first place. The ways in which these and other factors encourage infection are not always known, but some interference with host immune defences is common. Later on I shall discuss some experiments with staphylococci which depend on the realization that a stitch abscess is due to a more sophisticated impairment of host defences than a simple hole allowing the mechanical entry of bacteria. In recent years deliberate suppression of immune mechanisms in order to prevent graft rejection has increased the extent of the problem. The range of host defences, though finite, is wide, and it only makes sense to analyse them in relation to the sort of attack they may have to withstand. One of the things which has come out of the past few years' preoccupation with immune deficiences, whether they are congenital or iatrogenic, is the realization that the failure of any one defence mechanism does not give rise simply to a general increased risk of infection but usually
Based oni an Erasmus Wilson demonstration given on 12th November 1973
Immunity in acute bacterial infections to characteristic infections with specific organisms or groups of organisms2. For example, children with hypogammaglobulinaemia are usually infected with organisms such as the pyogenic cocci or Haemolphilus influenzae, which need antibodies if they are to be efficiently phagocytosed. Patients with deficient cell-mediated immunity are more likely to suffer from chronic intracellular infections. It is convenient to group acute bacterial infections by site. Thus in a surface infection such as cholera there may be defence mechanisms involved which are not found in a tissue infection such as a staphylococcal abscess. Such unlikely bedfellows as the gonococcus34, dysentery bacillus, and Bordetella pertussis5 all attack mucous membranes and all give the defences the problem of dealing with adhesion and penetration'. Septicaemias or a systemic infection like typhoid bring yet other defences into action. Whatever the site of what Miles7 has called the primary lodgement, prevention of spread and finally killing of the bacteria is desirable though perhaps not essential. Complete elimination is fine from a short-term clinical point of view, although if no immunity develops repeated infection is a possibility; look at recurrent urinary tract infection or gonorrhoea, for example. From a long-term evolutionary standpoint the development of symptomless carriage and eventually symbiosis between host and microbe may well be a better solution. Although you need to understand how immune responses are called forth, in what follows I shall concentrate on the effector side of the immune system-that is, on the processes by which cells and antibodies deal with infectious agents.
Immune responses: cellular effector mechanisms An outline of the processes involved in the
2I3
immune response to bacterial infection is
given in the accompanying figure. Bacteria, or rather bacterial antigens, directly or after processing by macrophages stimulate the central immune mechanisms to produce clones of the appropriate immunocompetent cells. Some of these will be B cells producing specific antibodies of the various classes. Others will be T cells of at least three functional types. The 'helper' cells co-operate with B cells in the antibody response. The 'memory' cells are long-lived and provide the mechanism for augmented secondary responses to subsequent infection by the same organisms. Other T cells on stimulation by antigen produce a variety of factors, the lymphokines, which attract, hold, and activate macrophages. This is the basis of cell-mediated immunity, one manifestation of which is delayed hypersensitivity. Although the stimulation of lymphocytes is antigenically specific, the increased phagocytic and bactericidal properties of the resulting activated macrophages are non-specific. A fuirther specific element in phagocytosis may, however, be given by the presence of cytophilic antibodies. The lymphocytes involved in this reaction are short-lived but have a great capacity for homing to areas of acute inflammation-that is, to areas where they are most needed9. 'Killer' lymphoorted
Lymphoctes
I I
AI Bacteriai
Immune response to bacterial infection.
2I4
A A Glynn
cytes which liberate soluble products capable of destroying other cells have not so far been shown to have any effect on bacteria. Although cellular immunity is of paramount importance in chronic intracellular infections such as tuberculosis, it is also the major mechanism of resistance in some acute bacterial infections, the best example being typhoid. As many chronic infections show, phagocytosis is not an end in itself. What is important is the subsequent intracellular fate of the bacteria, whether they are killed or multiply, destroy the phagocyte or live in an tunstable equilibrium with it. The nature of intracellular bactericidal systems is beyond the scope of this talk, but their failure is appreciated in, for example, chronic granulomatous disease of children, where the polymorphs are deficient in the myeloperoxidase system and fail to kill catalase-positive organisms such as staphylococci10. Cell-mediated immune mechanisms are not invariably beneficial to the host; witness the destructive effects of delayed hypersensitivity in tuberculosis. Delayed hypersensitivity is also responsible for much of the tissue damage in acute staphylococcal infections. Thus in mice injected subcutaneously with a virulent strain of Staphylococcus aureus a local necropurulent lesion develops in 24-48 h. The lesion is easily prevented by first depleting the mouse of T cells. In contrast very severe lesions can be produced by injecting additional immune T cells taken from a syngeneic mouse already infected with staphylococci".
were injected subcutaneously in guineapigs, Miles and his colleagues12 were able to show that suppression of the early inflammatory response by adrenaline resulted in a larger lesion for a given infecting dose. They calculated in fact that the early inflammation was responsible for a 99%/0 kill of the inoculum. It is likely that some kind of inhibition of inflammation by certain types of suture is the basis of stitch abscesses. Elek'3 demonstrated that the minimum infective dose of Staph. aureuis injected subcutaneously in human volunteers was reduced iooo-fold if the bacteria were put in by means of an infected thread. Using a modification of this technique in mice Agarwal14 could distinguish between virulent and avirulent staphylococci. The significant feature of the virulent strains was that they suppressed fluid and polymorph exudation over the first few hours of infection. Hill15 localized the factor involved to the staphylococcal cell wall. It acts early on in the sequence of inflammatory mediators by inhibiting the release of kinins"6. This staphylococcal factor is a good example of a whole range of bacterial products which interfere with host defence mechanisms and are hence sometimes known as impedins17. Thus the capsules of pneumococci and the K antigens of some strains of Escherichia coli prevent phagocytosis. While inflammation enables cells and antibodies to be delivered to an appropriate site, there have now been a lot of experiments showing that the inflammatory response is itself stimulated in part by immune processes. The role of inflammation A variety of reactions at the site of a lesion, Although it is easy to accept the idea that such as antigen-antibody combinations or tisinflammation is a mechanism by which phago- sue damage with release of proteases, may cytic cells and humoral factors such as anti- result in the activation of complement and body are concentrated at the site of a noxious the release of anaphylatoxins from C3 and stimulus, it is more difficult to come by C5 which cause local inflammation and are evidence of their value. In a series of experi- chemotactic for leucocytes. Where there is ments in which known numbers of bacteria a deficiency in circulating complement, as in
Immunity in acute bacterial infections some strains of mice18 or following the injection of some antisera'", then the inflammatory reaction is reduced not only to obvious potentially immunologgical stimuli such as bacteria but also to purely irritant ones such as turpentine. For those who find the intricacies of complement an inadequate challenge I can recommend a study of the complex interactions of the complement, kinin, and clotting systems, which have been reviewed by Ratnoff20.
Humoral effector mechanisms Although antibodies can neutralize toxins, particularly exotoxins, they do not by themselves generally do any great harm to bacteria. The final attack is either by complement or by phagocytosis and intracellular killing. Diphtheria and tetanus, the bacterial infections in which protective immunization has been most successful, are special cases. Both are purely local infections in which the local lesion is overshadowed by the widespread effects of a diffusible toxin. Although cholera and diarrhoea due to enteropathogenic E. coli are other diseases in which the infection is reasonably local and the pathogenic effects largely due to diffusible toxins, antitoxic immunization or therapy has not so far proved of any use. Antibodies are frequently themselves opsonic and play an important role in activating complement with the production of chemotactic and opsonic factors. Agglutination of bacteria by antibody could conceivably be useful in preventing spread of infection in vivo, but direct evidence of its value is lacking. Certainly flagellar antibodies which inhibit motility have never been shown to have any protective value in salmonella or coliform infections. Two interesting aspects of the role of antibodies in acute infections are the functional differences between antibodies of different classes and the blocking effects of antibodies in surface in-
I5
21
fections. A useful but certainly over-simplified view is that IgG antibodies are particularly efficient at neutralizing toxins, IgM antibodies at initiating complement killing and lysis of cells, and IgA antibodies at protecting mucous surfaces. IgM antibodies appear first in response to antigenic stimulation and would therefore be expected to be significant early in an infection. They are particularly efficient at activating complement because one molecule of the first component of complement, Ci, needs to attach to the Fc ends of 2 antibody molecules themselves modified in some way because the antibody of which they are part has become bound to antigen. Since one IgM molecule may contain 5 Fc pieces, even one molecule once bound to antigen can activate complement. In contrast, activation by IgG depends not only on the concentration of IgG antibody present but also on whether the density of the specific antigenic determinants on the surface of a cell is sufficient for there to be a reasonable chance of 2 IgG molecules becoming attached close together. Nevertheless, IgG antibodies can mediate complement opsonization or bactericidal action. Because an antibody is measured in vitro in a certain way does not necessarily mean that that is the way it acts in vivo. Forty years ago Fothergill and Wright2" related susceptibility to H. in/luenzae meningitis to the absence of bactericidal activity in the blood. Susceptibility to meningococcal meningitis has now been shown to follow a similar pattern22. Both types of meningitis occur particularly in infants from about 6 months of age and in very young children. Infants up to 6 months are protected by persistent maternal antibody and older children by antibody to non-clinical infections or to cross-reacting antigens in the normal throat flora. Studies in army camps showed that meningococcal meningitis oc-
2I6
A A Glynn
curred among those with no bactericidal antibody. Now although in these surveys antibody was measured by its bactericidal activity (in conjunction with complement), the same antibody is also opsonic and it is arguable which function is the most important. Whichever it is, protective antibodies to some serotypes of meningococci can now be induced by purified polysaccharide vaccines. Perhaps there is most controversy about IgA, sometimes called 'antiseptic paint' because of its presumed function in the surface secretions where it predominates. Although IgA was earlv on found not to fix complement, it can under suitable conditions, which include the presence of lysozyme, initiate the complement killing of E. coli23. This observation, like those of others attributing opsonic activity to IgA, has been the subject of much debate and was often disbelieved because it conflicted with classical complement dogma. The discovery24 of an alternative pathway for complement activation via C3 has broken tip the more rigid complement theory and left these problems again open to experimental testing. Whatever the final decision it would appear at first unlikely that either complement killing or phagocytosis could occur in the unpromising milieu of the surface of, say, the intestinal mucosa. However, we should not be too ready to rule out the possibility of special microenvironments, especially when there is local inflammation. IgA antibodies can certainly inhibit toxins, but how far this is important in natural infections is not yet clear. More striking is the ability of IgA antibodies to prevent the adherence of bacteria to surface epithelial cells. Such adherence is essential if newly arrived pathogens are to establish themselves on a site washed by secretions or by food and drink. A protective role for IgA acting in this way has been shown in experimental cholera25. IgA in saliva prevents the adhesion of various types
of streptococci and could be important in the prevention of caries2"'. It would be possible to go on multiplying examples of immunity to acute bacterial infections, but I have said enough to indicate some of the major mechanisms involved. There has been a continuing argument since the end of the igth century on the relative merits of cellular and humoral immunity. It should be clear by now that the two are mutually dependent, though which particular aspect of which is the most significant needs to be decided independently for individual infections.
References I
2
3 4
5 6
7 8 9
1o iI
12
Shaw, G B (I91i) The Doctor's Dilemma, London, Constable. Good, R A, Finstad, J, and Gatti, R A (1970) in Infectious Agents and Host Reactions, ed. S Mudd, p. 76. Philadelphia, Saunders. Ward, M E, and Watt, P J (1972) Journal of Infectious Diseases, I 26, 6oi. Swanson, J, Kraus, S J, and Gotschlich, E C (I97I) Journal of Experimental Medicine, 134, 886. Holt, L B (1972) Journal of Medical Microbiology, 5, 407. Savage, D C (I972) in Microbial Pathogenicity in Man and Animals; Symposium of the Society for General Microbiology, vol. 22, p. 25. Miles, A A (I955) Lectures on the Scientific Basis of Medicine, 3, 235. Mackaness, G B, and Blanden, R V (I967) Progress in Allergy, ii, 89. Koster, F T, and McGregor, D D (I97I) Journal of Experimental Medicine, 133, 864. Klebanoff, S J (I97i) Annual Review of Medicine, 22, 39. Easmon, C S F, and Glynin, A A Immunology. In press. Miles A A, Miles, E M, and Burke, J (I957) British Journal of Experimental Pathology, 38, 79.
Immunity in acute bacterial infections 13 Elek, S D, and Conen, P E (I957) British Journal of Experimental Pathology, 38, 573. 14 Agarwal D S (I967) British Journal of Experimental Pathology, 48, 436, 468. 15 Hill, M J (I968) Journal of Medical Microbiology, I, 33. i6 Easmon, C S F, Hamilton, I, and Glynn, A A (I973) British Journal of Experimental Pathology, 54, 638. '7 Glynn, A A (1972) in Microbial Pathogenicity in Man and Animals; Symposium of the Society for General Microbiology, vol. 22, p. 75. i8 Medhurst, F A, Hill, M J, and Glynn, A A (I969) Journal of Medical Microbiology, 2, 147. 19 Willoughby D A, Cook, E, and Turk, J L (I969) Journal of Pathology, 97, 295.
17
20 Ratnoff, 0 D (I969) Advances in Immunology, 10, 145.
2I Fothergill, L D, and Wright, J (I933) Journal of Immunology, 24, 28I. 22 Goldschneider, 1, Gotschlich, E C, and Artenstein, M S (I969) Journal of Experimental Medicine 129, I307.
23 Adinolfi, M, Glynn, A A, Lindsay M, and Milne, C M (I966) Immunology, IO, 5I7.
24 Osler, A G, and Sandberg, A A in Allergy, 17, 51. 25 Fubara, E S, and Freter, R Immunology, III, 395.
(I973) Progress
(I973) Journal of
26 Williams, R C, and Gibbons, R J (I972) Science,
'77, 697.
Immunity
in
acute bacterial infections
A A Glynn MD FRCP FRCPathi Professor in Bacteriology, Wright-Fleming Institute, St Mary's Hospital Medical School, London
Summary A brief survey of the ways in which humoral and cellular immune responses deal with bacteria and their products shows that a variety of effector mechanisms is involved. Their efficiency varies, partly in relation to the kind of bacterial attack with which they have to deal. Some bacterial factors though not crudely toxic can interfere with specific host defences. Introduction 'Find the germ of the disease; prepare from it a suitable anti-toxin; inject it three times a day quarter of an hour before meals; and what is the result? The phagocytes are stimulated; they devour the disease; and the patient recovers-unless, of course, he's too far gone.' SIR RALPH BLOOMFIELD BONINGTON Although it is over sixty years since Bernard Shaw1 wrote his accurate though wholly misleading account, immunization and stimulation of the phagocytes are still topical, while the possibilities of renal transplantation and home dialysis have made 'The Doctor's Dilemma' even more acute. Not many of the infections I am going to talk about have any direct surgical significance, but since Erasmus Wilson, who was a Vice-President of the Royal College of Surgeons in I879
and i 88o and President in i 88 I, was in fact a dermatologist you will perhaps excuse a broadly medical and somewhat academic approach. Surgeons, of course, have always been concerned with infection. Infection may follow obstruction which needs surgical relief; abscesses may need drainage; foreign bodies in tissues are a common precipitating cause and may have to be removed by the surgeon, even if he put them there in the first place. The ways in which these and other factors encourage infection are not always known, but some interference with host immune defences is common. Later on I shall discuss some experiments with staphylococci which depend on the realization that a stitch abscess is due to a more sophisticated impairment of host defences than a simple hole allowing the mechanical entry of bacteria. In recent years deliberate suppression of immune mechanisms in order to prevent graft rejection has increased the extent of the problem. The range of host defences, though finite, is wide, and it only makes sense to analyse them in relation to the sort of attack they may have to withstand. One of the things which has come out of the past few years' preoccupation with immune deficiences, whether they are congenital or iatrogenic, is the realization that the failure of any one defence mechanism does not give rise simply to a general increased risk of infection but usually
Based oni an Erasmus Wilson demonstration given on 12th November 1973
Immunity in acute bacterial infections to characteristic infections with specific organisms or groups of organisms2. For example, children with hypogammaglobulinaemia are usually infected with organisms such as the pyogenic cocci or Haemolphilus influenzae, which need antibodies if they are to be efficiently phagocytosed. Patients with deficient cell-mediated immunity are more likely to suffer from chronic intracellular infections. It is convenient to group acute bacterial infections by site. Thus in a surface infection such as cholera there may be defence mechanisms involved which are not found in a tissue infection such as a staphylococcal abscess. Such unlikely bedfellows as the gonococcus34, dysentery bacillus, and Bordetella pertussis5 all attack mucous membranes and all give the defences the problem of dealing with adhesion and penetration'. Septicaemias or a systemic infection like typhoid bring yet other defences into action. Whatever the site of what Miles7 has called the primary lodgement, prevention of spread and finally killing of the bacteria is desirable though perhaps not essential. Complete elimination is fine from a short-term clinical point of view, although if no immunity develops repeated infection is a possibility; look at recurrent urinary tract infection or gonorrhoea, for example. From a long-term evolutionary standpoint the development of symptomless carriage and eventually symbiosis between host and microbe may well be a better solution. Although you need to understand how immune responses are called forth, in what follows I shall concentrate on the effector side of the immune system-that is, on the processes by which cells and antibodies deal with infectious agents.
Immune responses: cellular effector mechanisms An outline of the processes involved in the
2I3
immune response to bacterial infection is
given in the accompanying figure. Bacteria, or rather bacterial antigens, directly or after processing by macrophages stimulate the central immune mechanisms to produce clones of the appropriate immunocompetent cells. Some of these will be B cells producing specific antibodies of the various classes. Others will be T cells of at least three functional types. The 'helper' cells co-operate with B cells in the antibody response. The 'memory' cells are long-lived and provide the mechanism for augmented secondary responses to subsequent infection by the same organisms. Other T cells on stimulation by antigen produce a variety of factors, the lymphokines, which attract, hold, and activate macrophages. This is the basis of cell-mediated immunity, one manifestation of which is delayed hypersensitivity. Although the stimulation of lymphocytes is antigenically specific, the increased phagocytic and bactericidal properties of the resulting activated macrophages are non-specific. A fuirther specific element in phagocytosis may, however, be given by the presence of cytophilic antibodies. The lymphocytes involved in this reaction are short-lived but have a great capacity for homing to areas of acute inflammation-that is, to areas where they are most needed9. 'Killer' lymphoorted
Lymphoctes
I I
AI Bacteriai
Immune response to bacterial infection.
2I4
A A Glynn
cytes which liberate soluble products capable of destroying other cells have not so far been shown to have any effect on bacteria. Although cellular immunity is of paramount importance in chronic intracellular infections such as tuberculosis, it is also the major mechanism of resistance in some acute bacterial infections, the best example being typhoid. As many chronic infections show, phagocytosis is not an end in itself. What is important is the subsequent intracellular fate of the bacteria, whether they are killed or multiply, destroy the phagocyte or live in an tunstable equilibrium with it. The nature of intracellular bactericidal systems is beyond the scope of this talk, but their failure is appreciated in, for example, chronic granulomatous disease of children, where the polymorphs are deficient in the myeloperoxidase system and fail to kill catalase-positive organisms such as staphylococci10. Cell-mediated immune mechanisms are not invariably beneficial to the host; witness the destructive effects of delayed hypersensitivity in tuberculosis. Delayed hypersensitivity is also responsible for much of the tissue damage in acute staphylococcal infections. Thus in mice injected subcutaneously with a virulent strain of Staphylococcus aureus a local necropurulent lesion develops in 24-48 h. The lesion is easily prevented by first depleting the mouse of T cells. In contrast very severe lesions can be produced by injecting additional immune T cells taken from a syngeneic mouse already infected with staphylococci".
were injected subcutaneously in guineapigs, Miles and his colleagues12 were able to show that suppression of the early inflammatory response by adrenaline resulted in a larger lesion for a given infecting dose. They calculated in fact that the early inflammation was responsible for a 99%/0 kill of the inoculum. It is likely that some kind of inhibition of inflammation by certain types of suture is the basis of stitch abscesses. Elek'3 demonstrated that the minimum infective dose of Staph. aureuis injected subcutaneously in human volunteers was reduced iooo-fold if the bacteria were put in by means of an infected thread. Using a modification of this technique in mice Agarwal14 could distinguish between virulent and avirulent staphylococci. The significant feature of the virulent strains was that they suppressed fluid and polymorph exudation over the first few hours of infection. Hill15 localized the factor involved to the staphylococcal cell wall. It acts early on in the sequence of inflammatory mediators by inhibiting the release of kinins"6. This staphylococcal factor is a good example of a whole range of bacterial products which interfere with host defence mechanisms and are hence sometimes known as impedins17. Thus the capsules of pneumococci and the K antigens of some strains of Escherichia coli prevent phagocytosis. While inflammation enables cells and antibodies to be delivered to an appropriate site, there have now been a lot of experiments showing that the inflammatory response is itself stimulated in part by immune processes. The role of inflammation A variety of reactions at the site of a lesion, Although it is easy to accept the idea that such as antigen-antibody combinations or tisinflammation is a mechanism by which phago- sue damage with release of proteases, may cytic cells and humoral factors such as anti- result in the activation of complement and body are concentrated at the site of a noxious the release of anaphylatoxins from C3 and stimulus, it is more difficult to come by C5 which cause local inflammation and are evidence of their value. In a series of experi- chemotactic for leucocytes. Where there is ments in which known numbers of bacteria a deficiency in circulating complement, as in
Immunity in acute bacterial infections some strains of mice18 or following the injection of some antisera'", then the inflammatory reaction is reduced not only to obvious potentially immunologgical stimuli such as bacteria but also to purely irritant ones such as turpentine. For those who find the intricacies of complement an inadequate challenge I can recommend a study of the complex interactions of the complement, kinin, and clotting systems, which have been reviewed by Ratnoff20.
Humoral effector mechanisms Although antibodies can neutralize toxins, particularly exotoxins, they do not by themselves generally do any great harm to bacteria. The final attack is either by complement or by phagocytosis and intracellular killing. Diphtheria and tetanus, the bacterial infections in which protective immunization has been most successful, are special cases. Both are purely local infections in which the local lesion is overshadowed by the widespread effects of a diffusible toxin. Although cholera and diarrhoea due to enteropathogenic E. coli are other diseases in which the infection is reasonably local and the pathogenic effects largely due to diffusible toxins, antitoxic immunization or therapy has not so far proved of any use. Antibodies are frequently themselves opsonic and play an important role in activating complement with the production of chemotactic and opsonic factors. Agglutination of bacteria by antibody could conceivably be useful in preventing spread of infection in vivo, but direct evidence of its value is lacking. Certainly flagellar antibodies which inhibit motility have never been shown to have any protective value in salmonella or coliform infections. Two interesting aspects of the role of antibodies in acute infections are the functional differences between antibodies of different classes and the blocking effects of antibodies in surface in-
I5
21
fections. A useful but certainly over-simplified view is that IgG antibodies are particularly efficient at neutralizing toxins, IgM antibodies at initiating complement killing and lysis of cells, and IgA antibodies at protecting mucous surfaces. IgM antibodies appear first in response to antigenic stimulation and would therefore be expected to be significant early in an infection. They are particularly efficient at activating complement because one molecule of the first component of complement, Ci, needs to attach to the Fc ends of 2 antibody molecules themselves modified in some way because the antibody of which they are part has become bound to antigen. Since one IgM molecule may contain 5 Fc pieces, even one molecule once bound to antigen can activate complement. In contrast, activation by IgG depends not only on the concentration of IgG antibody present but also on whether the density of the specific antigenic determinants on the surface of a cell is sufficient for there to be a reasonable chance of 2 IgG molecules becoming attached close together. Nevertheless, IgG antibodies can mediate complement opsonization or bactericidal action. Because an antibody is measured in vitro in a certain way does not necessarily mean that that is the way it acts in vivo. Forty years ago Fothergill and Wright2" related susceptibility to H. in/luenzae meningitis to the absence of bactericidal activity in the blood. Susceptibility to meningococcal meningitis has now been shown to follow a similar pattern22. Both types of meningitis occur particularly in infants from about 6 months of age and in very young children. Infants up to 6 months are protected by persistent maternal antibody and older children by antibody to non-clinical infections or to cross-reacting antigens in the normal throat flora. Studies in army camps showed that meningococcal meningitis oc-
2I6
A A Glynn
curred among those with no bactericidal antibody. Now although in these surveys antibody was measured by its bactericidal activity (in conjunction with complement), the same antibody is also opsonic and it is arguable which function is the most important. Whichever it is, protective antibodies to some serotypes of meningococci can now be induced by purified polysaccharide vaccines. Perhaps there is most controversy about IgA, sometimes called 'antiseptic paint' because of its presumed function in the surface secretions where it predominates. Although IgA was earlv on found not to fix complement, it can under suitable conditions, which include the presence of lysozyme, initiate the complement killing of E. coli23. This observation, like those of others attributing opsonic activity to IgA, has been the subject of much debate and was often disbelieved because it conflicted with classical complement dogma. The discovery24 of an alternative pathway for complement activation via C3 has broken tip the more rigid complement theory and left these problems again open to experimental testing. Whatever the final decision it would appear at first unlikely that either complement killing or phagocytosis could occur in the unpromising milieu of the surface of, say, the intestinal mucosa. However, we should not be too ready to rule out the possibility of special microenvironments, especially when there is local inflammation. IgA antibodies can certainly inhibit toxins, but how far this is important in natural infections is not yet clear. More striking is the ability of IgA antibodies to prevent the adherence of bacteria to surface epithelial cells. Such adherence is essential if newly arrived pathogens are to establish themselves on a site washed by secretions or by food and drink. A protective role for IgA acting in this way has been shown in experimental cholera25. IgA in saliva prevents the adhesion of various types
of streptococci and could be important in the prevention of caries2"'. It would be possible to go on multiplying examples of immunity to acute bacterial infections, but I have said enough to indicate some of the major mechanisms involved. There has been a continuing argument since the end of the igth century on the relative merits of cellular and humoral immunity. It should be clear by now that the two are mutually dependent, though which particular aspect of which is the most significant needs to be decided independently for individual infections.
References I
2
3 4
5 6
7 8 9
1o iI
12
Shaw, G B (I91i) The Doctor's Dilemma, London, Constable. Good, R A, Finstad, J, and Gatti, R A (1970) in Infectious Agents and Host Reactions, ed. S Mudd, p. 76. Philadelphia, Saunders. Ward, M E, and Watt, P J (1972) Journal of Infectious Diseases, I 26, 6oi. Swanson, J, Kraus, S J, and Gotschlich, E C (I97I) Journal of Experimental Medicine, 134, 886. Holt, L B (1972) Journal of Medical Microbiology, 5, 407. Savage, D C (I972) in Microbial Pathogenicity in Man and Animals; Symposium of the Society for General Microbiology, vol. 22, p. 25. Miles, A A (I955) Lectures on the Scientific Basis of Medicine, 3, 235. Mackaness, G B, and Blanden, R V (I967) Progress in Allergy, ii, 89. Koster, F T, and McGregor, D D (I97I) Journal of Experimental Medicine, 133, 864. Klebanoff, S J (I97i) Annual Review of Medicine, 22, 39. Easmon, C S F, and Glynin, A A Immunology. In press. Miles A A, Miles, E M, and Burke, J (I957) British Journal of Experimental Pathology, 38, 79.
Immunity in acute bacterial infections 13 Elek, S D, and Conen, P E (I957) British Journal of Experimental Pathology, 38, 573. 14 Agarwal D S (I967) British Journal of Experimental Pathology, 48, 436, 468. 15 Hill, M J (I968) Journal of Medical Microbiology, I, 33. i6 Easmon, C S F, Hamilton, I, and Glynn, A A (I973) British Journal of Experimental Pathology, 54, 638. '7 Glynn, A A (1972) in Microbial Pathogenicity in Man and Animals; Symposium of the Society for General Microbiology, vol. 22, p. 75. i8 Medhurst, F A, Hill, M J, and Glynn, A A (I969) Journal of Medical Microbiology, 2, 147. 19 Willoughby D A, Cook, E, and Turk, J L (I969) Journal of Pathology, 97, 295.
17
20 Ratnoff, 0 D (I969) Advances in Immunology, 10, 145.
2I Fothergill, L D, and Wright, J (I933) Journal of Immunology, 24, 28I. 22 Goldschneider, 1, Gotschlich, E C, and Artenstein, M S (I969) Journal of Experimental Medicine 129, I307.
23 Adinolfi, M, Glynn, A A, Lindsay M, and Milne, C M (I966) Immunology, IO, 5I7.
24 Osler, A G, and Sandberg, A A in Allergy, 17, 51. 25 Fubara, E S, and Freter, R Immunology, III, 395.
(I973) Progress
(I973) Journal of
26 Williams, R C, and Gibbons, R J (I972) Science,
'77, 697.
