This study demonstrates that the composition of the colonizing flora may be affected by competition between multiple microbial species through the innate host responses they induce. Either of the two species analyzed in this report persisted on the mucosal surface of mice when given individually, but in combination one species quickly and consistently became predominant. This competitive relationship between species, moreover, was the result of the host's response to co-colonization and was not predicted by in vitro investigation of direct bacterial–bacterial interactions [4
In the example of microbial interference described in this study, clearance of S. pneumoniae
required both complement and the recruitment of neutrophil-like cells to the mucosal surface. A central role of these host factors was not unexpected since complement-mediated opsonization followed by ingestion and killing by professional phagocytes such as neutrophils or macrophages is a major host defense against this encapsulated gram-positive pathogen [22
]. A further consideration in defining the contribution of these components of innate immunity is that inflammatory responses to polymicrobial colonization may be markedly different from those to a single type of organism. In an earlier report, we described the synergistic proinflammatory responses of respiratory epithelial cells in vitro and in the nasal mucosa in vivo when exposed to H. influenzae
and S. pneumoniae
in combination [20
]. Discrete signals from each species contribute to levels of the neutrophil-recruiting chemokine MIP-2 (or IL-8 in human epithelial cells) significantly greater that seen with either organism alone. This synergistic response of the epithelium correlates with an influx of neutrophil-like cells into the nasal spaces. In the current study, which examines the outcome of these inflammatory responses, it was not practical to obtain adequate numbers of neutrophils directly from the nasal spaces to address whether these were sufficient for the clearance of S. pneumoniae
. When tested ex vivo, murine neutrophils enriched from PECs did not kill S. pneumoniae Sp
1121 efficiently. This suggested that the recruitment of neutrophils may not be sufficient to account for the competitive effect described here. Rather, stimulation of neutrophil-like cells with bacterial components of H. influenzae
was required for efficient clearance of S. pneumoniae
. Killing in these assays was dependent on active complement, consistent with its role as an opsonin promoting phagocytosis. If similar events occur in the local environment of the nasopharynx, innate immune responses, consisting of complement and enhanced neutrophil recruitment and activation through recognition of microbial products, may underlie the host's role in clearance of colonizing bacteria from the mucosal surface.
Activation and enhanced opsonophagocytosis of neutrophil-like cells ex vivo was found in response to products of H. influenzae
but not equivalent doses of products from S. pneumoniae
. The selective innate responses of inflammatory cells such as neutrophils to products from one microbe but not another, therefore, may provide a mechanism whereby one species induces clearance of a competitor. The LPS of H. influenzae
and multiple other cellular components, including peptidoglycan, lipoproteins, phosphorylcholine, and an incompletely characterized soluble cytoplasmic fraction, have been implicated in promoting inflammatory responses [23
]. Purified LPS from other species has been shown to trigger an increase in the migration, life span, and activity of neutrophils [28
]. The molecular nature of the signal(s) from H. influenzae
mediating the recruitment and activity of neutrophils is the topic of ongoing investigation. Results to date do not suggest that purified LPS of H. influenzae
is sufficient to stimulate killing of S. pneumoniae
by neutrophil-enriched PECs (data not shown). Although components of S. pneumoniae,
including cell wall fragments and lipoteichoic acid and its toxin (pneumolysin), have been shown to be potent inducers of inflammation, these appeared to be at least 100-fold less active on a per cell basis in generating the neutrophil responses described here [31
]. H. influenzae
products, furthermore, stimulated killing of another species (S. pneumoniae),
but had no effect in opsonophagocytic killing assays on the same species and strain from which these products were derived. Thus, our findings demonstrate that one species may compete with another through selective induction of host responses and may benefit from the differences in its susceptibility to the antimicrobial host factors it induces.
This study shows the importance of specific innate immune responses in dictating the initial success of a species in becoming established within a competitive niche such as the mucosal surface of the nasopharynx. Selective microbial pattern recognition, as demonstrated here for phagocytic activity, may act in the setting of a complex milieu of organisms that differ in their ability to trigger these host-specific responses. This process ultimately selects for the persistence of those species best able to evade the local host clearance factors induced by polymicrobial stimulaton of the innate immune system. The role of innate immunity in colonization described here is distinct from its more extensively studied role in infection.
An additional consideration is that the clearance of S. pneumoniae
in co-colonized SCID mice demonstrates that the effects of complement and neutrophil-like cells were independent of adaptive immunity and the presence of antibody. Antibody-independent clearance was also demonstrated by in vitro assays in which we observed efficient killing in the presence of serum lacking anti-phagocytic antibodies. Antibody-independent opsonophagoctic killing of S. pneumoniae,
as previously recognized in the classic studies of Wood et al., may be important in protection during the critical period prior to the acquisition of specific anti-capsular antibody [35
]. Activation of phagocytic cells, however, has not been a feature of standardized opsonophagocytic killing assays for S. pneumoniae
Colonization of mucosal surfaces is often the first step in the development of disease for many important pathogens. This study demonstrates that the presence of one species may impact the ability of another to persist in the same microenvironment on a mucosal surface. The focus of this report is on bacteria that commonly colonize and potentially infect the respiratory tract of humans. There may be clinical relevance to our observations that in a model of dual colonization H. influenzae
was able to induce responses that caused the complete elimination of S. pneumoniae,
a leading opportunistic pathogen. In regard to colonization, numerous surveys have described carriage rates for H. influenzae
and S. pneumoniae,
although, to our knowledge, none appear to have examined the effect of colonization by one species on the density of the other in a quantitative manner. In regard to disease involving the respiratory tract, however, some reports suggest that antagonism between these two species may occur in the natural host [37
]. Most H. influenzae
disease is currently caused by non-typeable strains, which were not tested in in vivo experiments, because of their less efficient colonization of either SCID or immunocompetent mice compared to the type b strain used in our study (data not shown). Nonetheless, an isolate of non-typeable H. influenzae
was shown to be equally effective in stimulating neutrophil-mediated killing.
The competitive interactions described in this report may also be applicable to other combinations of microbes where there is evidence of antagonism in vivo. For example, a previously unrecognized competitive interaction between S. pneumoniae
and Staphylococcus aureus
could explain recent reports that children who receive the pneumococcal conjugate vaccine have lower rates of vaccine-type S. pneumoniae
carriage, but higher rates of Sta. aureus
nasal colonization as well as otitis media caused by Sta. aureus
]. In this regard, the composition of the normal flora is generally regarded as a critical factor in protection from potentially more virulent opportunistic organisms. Our study provides an initial mechanistic understanding of how manipulation of the colonizing flora could have unexpected consequences on competitors. Since an expanding number of medical interventions impact the composition of the microflora, it would seem prudent to more fully appreciate the scope of competitive interactions on mucosal surfaces.
Our findings also demonstrate that the success of an organism in initiating carriage may depend on its ability to resist innate clearance mechanisms of mucosal surfaces generated in the setting of polymicrobial stimulation. Since characteristics that enhance evasion of innate immunity are often critical determinants of microbial pathogenicity, competition between species may promote the selection for virulence among species such as S. pneumoniae and H. influenzae that must first establish a niche on heavily colonized surfaces.