|Home | About | Journals | Submit | Contact Us | Français|
An explosion of discovery between 1870 and the early 1900s identified the major bacterial pathogens of Homo sapiens, and in the 1930s Domagk's work on sulpha compounds and Fleming's luck with Penicillium notatum initiated our love affair with antibiotics. The 1950s brought apparent victory in our war with bacteria, and so research in microbiology switched to other sources of human infection, principally the virus; but the rapid escalation of antibiotic resistance in the past two decades has led to a renaissance in bacteriological research and the development of a new science, cellular microbiology1,2,3,4—a fusion of classic microbiology and cell biology with major inputs from molecular and structural biology.
Cellular microbiology has moved the focus from the study of bacteria, or of host cells, in isolation to the study of the interactions between these cells that cause bacterial diseases. Many of our most successful pathogens target the gut, and one of the most interesting of these pathogens is enteropathogenic Escherichia coli or EPEC. This pathogen does what all such organisms must do to infect—it adheres to the walls of the intestine. The interaction involves one of the paradigms of modern cellular microbiology—the type III secretion system. In simple terms this is a molecular syringe with which the bacterium can pierce the host cell plasma membrane and inject proteins that alter the behaviour of the injected cell. In EPEC, as in the other bacteria possessing type III secretion systems, the genes encoding this syringe (injectisome) are found on what is termed a pathogenicity island. These islands are large segments of the bacterial chromosome which are associated with virulence (strains lacking these islands are generally avirulent), and have a different nucleotide composition from the host bacterial chromosome. In EPEC the pathogenicity island is called the ‘locus of enterocyte effacement’ and consists of those genes encoding the injectisome and those proteins secreted by it. The basic trick of EPEC is to manipulate the actin cytoskeleton of the host enterocytes to induce loss of the microvilli and produce a raised area on the cell surface (the pedestal) to which the bacterium attaches, via a receptor. The most remarkable facet of the interaction of EPEC with enterocytes is that the receptor to which the bacteria bind—the translocated intimin receptor—is not a host protein but a bacterial protein that has been injected into the host cell5.
Bacteria have evolved many strategies for interacting with key elements of eukaryotic cell behaviour in order to infect their hosts or to live with them in harmony as commensals. We are only beginning to understand the complexity of the interactions which occur between bacterial pathogens and their infected hosts and there is an enormous amount still to be learned. Much of our ignorance concerns the interactions of oral bacteria with the many cells and tissues that constitute the oral cavity.
As the accompanying article6 reveals, the human oral cavity is awash with bacteria, and the numerous habitats within this anatomical site contain many hundreds of bacterial species. There are two related questions about our interactions with these oral bacteria. The first concerns the enormous amount of oral bacterial disease that Homo sapiens endures5. The whole population is subject to the ravages of caries, gingivitis and periodontitis. Although the gut contains probably as many bacterial species as does the mouth, we do not have to take daily prophylactic measures to limit the bacterial content of the gut. The very common oral diseases are due to a small number of Gram-positive or Gram-negative bacteria. The related question is—what prevents us responding to all of our oral microflora? Our understanding of the interaction of oral bacteria with host cells and surfaces is still in its infancy, but the discipline of oral microbiology has begun to show how bacterial biofilm—a complex organization of bacteria adherent on a surface—contributes to the various oral diseases.
One example will be given of how cellular microbiology is being used to address the behaviour of oral pathogens. Actinobacillus actinomycetemcomitans is a Gramnegative bacterium which is either a member of the Haemophilus—Actinobacillus—Pasteurella family or closely related to it7. This family contains major human and animal pathogens such as Haemophilus influenzae, causing meningitis in man, and Pasteurella multocida, causing a chronic inflammatory condition of pigs in which the bones of the snout are destroyed. A. actinomycetemcomitans is implicated as a causative organism of periodontitis, particularly localized juvenile periodontitis8. This organism is said to produce several unusual virulence factors9,10 and to invade host cells10,11.
Amongst the reported putative virulence factors of this bacterium are a 116 kDa leukotoxin which targets only human neutrophils and monocytes12, a secreted chaperonin 60 protein which stimulates bone resorption13 by acting as a growth factor for osteoclasts14 and promotes cytolysis of epithelial cells15, and a small peptide with unusual cytokine-network-inducing properties16. The organism possesses the three contiguous genes (cdtA, cdtB and cdtC) which encode an activity known as cytolethal distending toxin (CDT)17. This toxin blocks cycling cells in the latter part of the cell cycle (termed G2)—an action that seems to be due to the nuclease activity exhibited by CdtB18. Our own work suggests that CdtC is required to allow CdtB to enter into cells (Akifusa, Henderson, Stenbeck and Henderson, Unpublished). The role of CDT in pathogenesis is unclear. One possibility is that it blocks lymphocyte proliferation and is therefore immunosuppressive19.
While these so-called virulence factors have effects on cells which could give rise to tissue disease, their actual role in the diseases caused by A. actinomycetemcomitans is uncertain. The clearest way of determining this is to inactivate the gene of interest and assess whether the bacterium has lost its pathogenicity. An organism related to A. actinomycetemcomitans, Haemophilus ducreyi (which causes the sexually transmitted disease, chancre) likewise produces a leukotoxin and CDT. The assumption that both toxins contributed to its pathogenicity was shown to be incorrect when inactivation of the genes did not block the pathogenicity20,21.
A major goal of modern cellular microbiology is the identification of the complete set of likely virulence genes. With bacteria that kill host cells various mutagenesis techniques can be used to identify genes that contribute to survival in the host and are likely to be involved in virulence. One way of doing this is to introduce mobile genetic elements called transposons that, on random insertion into the bacterial genome, inactivate individual genes. It is only very recently that methods have become available to do this in A. actinomycetemcomitans22 and this methodology has identified flp-1 as a pilin gene subfamily required for the non-specific adherence of this organism23. Many of the virulence proteins produced by bacteria are secreted, and one global method of identifying genes encoding secreted proteins is to make a gene library in a vector which has a truncated version of the gene for alkaline phosphatase. DNA fragments of the chromosome of the bacterium of interest containing signal sequences required to export proteins to the bacterial periplasm are fused to the alkaline phosphatase gene. The exported protein, when it reaches the periplasm, can then be identified by assay for alkaline phosphatase; and the responsible genes can be determined by screening of the genome libraries for clones containing the AP gene. Two recent papers describe identification, by this method, of several genes that could be important in bacterial virulence. Another strategy for identifying secreted proteins is termed proteomics. In this technique bacterial secreted proteins are separated on two-dimensional gels on the basis of differences in their charge and mass. Individual proteins can then be identified by direct sequencing or by in-gel proteolysis and peptide mass fingerprinting by mass spectrometry26.
Antibodies have been used for many years to identify gene products produced by genomic DNA fragments in expression libraries. These antibodies can come from the blood of patients with specific bacterial infections. A recent and very exciting development of this technique, termed IVIAT (in-vivo induced antibody technology), uses plate-grown bacteria to adsorb sera from patients who are infected with the bacterium of interest. In theory, this removes all antibodies directed to the components that bacteria produce when they are not infecting their host, leaving behind those antibodies that recognize only those gene products that bacteria generate when they are producing disease in the host. This simple technique is now being applied to a range of oral bacteria27 and has identified in A. actinomycetemcomitans a substantial number of genes that may be responsible for virulence (Hillman JD, Personal communication).
We have long known that certain organisms, Mycobacterium tuberculosis being a good example, are obligate intracellular bacteria. One of the major discoveries of cellular microbiology has been that many pathogenic bacteria can invade several sorts of host cell. Some strains of A. actinomycetemcomitans can enter both vascular endothelial cells28 and epithelial cells29. The receptor involved in invasion appears to be that which binds to the pro-inflammatory lipid mediator platelet-activating factor28. Most of the bacteria that invade human cells utilize the actin cytoskeleton for movement within and between cells30. Seemingly, A. actinomycetemcomitans in culture enters and moves through epithelial cells by a novel mechanism involving microtubules31. Again, we do not know how important cell invasion is for the pathogenesis of diseases caused by A. actinomycetemcomitans. However, it is clear that bacteria within cells are no longer directly exposed to the immune system or to some antibiotics, and that they gain distinct advantages from hiding within cells.
To fully understand bacterial disease we must grapple with the nature of the two-way communication that exists between bacteria and the cells of the host. Such communication is, as far as we can tell, normally beneficial for both parties and this is why, with perhaps 1000-3000 bacterial species existing in and on our bodies, only a few dozens cause disease. To understand this bacteria-host communication we need to study the interactions in situ. The most obvious arena for such study is the human mouth. This accessible organ, with its multiple habitats, rich microflora and rapid responsiveness to bacterial overgrowth, should be the perfect test-bed for identifying the rich complexity of the conversations that must be occurring between our own bacteria and our tissues. The genetic and cellular techniques needed for this study exist, and the need is for clinicians who will take up the challenge. The importance of this challenge is highlighted by the finding that A. actinomycetemcomitans and Porphyromonas gingivalis were present within buccal epithelial cells of 23 out of 24 individuals examined32. The suggestion that periodontopathogens are normally present inside oral epithelial cells adds a new dimension to the study of host—bacterial interactions in the mouth.