Recent data obtained using mouse models of colonisation have emphasised the importance of Th17 responses for protection against subsequent re-infection of the nasopharynx, and have shown that colonisation also protects against subsequent invasive infection. The development of protection against invasive infection required both CD4 cells and antibody but the precise mechanisms involved have not been clearly defined 
. In this study, we have investigated in detail potential mechanisms of protection against subsequent pneumonia associated with S. pneumoniae
nasopharyngeal colonisation. Mortality due to S. pneumoniae
is mainly related to severe infections with septicaemia. Hence, to identify protective mechanisms that are effective against severe infection we have deliberately chosen a disease model of fulminant infection requiring a large inoculum with rapid spread of bacteria from the lungs to the blood and a high mortality. Several mechanisms through which colonisation may impact on subsequent disease progression were identified.
Firstly, prior nasopharyngeal colonisation was associated with a stronger mucosal inflammatory response 4 h after pneumonia challenge, with higher levels of some cytokines and a more rapid influx of neutrophils. The rapidity of onset of this difference between colonised and control mice suggests that colonisation may lead to alteration of the ‘innate immune rheostat’, priming for a more robust response to subsequent pneumonia challenge 
. During the time that the nasopharynx remains colonised, small numbers of bacteria are likely to be aspirated into the lungs. These could have effects on innate immune cells such as macrophages and gamma-delta T-cells as shown for non-bacteraemic pneumonia 
. This could lead to more robust cytokine production early in subsequent infection. Alternatively, mucosal antibody induced through prior colonisation may facilitate the interaction between bacteria and host cells such as alveolar macrophages during early pneumonia challenge and enhance cytokine responses. However, the effects of prior colonisation on the early inflammatory response in the lungs did not limit the development of disease in this model. There were no significant differences between colonised and control mice in in vivo
phagocytosis of bacteria by alveolar macrophages, in bacterial CFU in BALF or lung, or in the histological severity of pneumonia. Potentially, the effects of colonisation on early inflammatory responses may have protective effects with other less fulminant models of pneumonia using lower inocula and / or less virulent S. pneumoniae
Nasopharyngeal colonisation also primed for an enhanced lung and systemic IL-17 response following in vivo
challenge. This was dependant on CD4-cells, strongly suggesting that it derives from a Th17-cell response to colonisation similar to that found in the nasopharynx following re-colonisation 
. However in contrast to the importance of a Th17-cell response for preventing re-colonisation of the nasopharynx, colonised mice depleted of CD4 cells were still protected against bacteraemia and still had reduced lung CFU at 18 hours. Hence, Th17 responses are not required for colonisation-induced protection in this model. The inability of previous colonisation to protect MHCII deficient mice against lethal challenge 
is likely to reflect a need for CD4 T-cells in supporting the development of a mature antibody response to colonisation, rather than an effector role at the time of challenge. Rapid neutrophil recruitment occurs within the lungs even in uncolonised control mice, and bacteraemia is established relatively early in infection, probably eclipsing any benefits from a Th17-cell mediated increase in phagocyte recruitment. In contrast, recruitment of phagocytes to the nasopharynx during S. pneumoniae
colonisation is delayed unless supported by an adaptive Th17 cell response induced by a previous colonisation event 
. Hence a Th17 response is critical for adaptive immunity against colonising S. pneumoniae
but not for the rapidly invasive model of pneumonia described in this mansucript.
In addition, there is some evidence that Th17 cells actually might be deleterious during S. pneumoniae
, possibly explaining why we found CD4 depletion was associated with a trend towards fewer bacteria in the blood at 18 h. Vaccination-induced Th17-cell responses may still be beneficial for protection against a less fulminant pneumonia in which bacteria remain within the lung.
The strongest effect of colonisation observed during challenge in our model was the protection against bacteraemia. This suggests protective responses either prevent spread of S. pneumoniae from the lungs to the blood or that on reaching the blood bacteria are rapidly cleared. Nearly all mice developed serum IgG responses to S. pneumoniae D39 antigens, and experiments with antibody-deficient mice demonstrated that protection against bacteraemia was highly dependent on antibody to S. pneumoniae. Following passive transfer of serum by intraperitoneal injection, there was a significant reduction in the number of bacteria present in the lungs 18 h following infection. This suggests that systemic antibody can assist control of infection within the lungs as well. Antibody to S. pneumoniae is mainly thought to cause protection by opsonising bacteria for phagocytosis, and in vitro phagocytosis assays and in vivo IV clearance data confirmed that colonisation induced antibody responses improved S. pneumoniae phagocytosis. The strength of the effect on IV clearance was particularly striking, reducing bacterial CFU by a factor of 3 to 4 log10, readily explaining why colonised mice do not develop septicaemia. The observation that colonisation prevents the development of systemic infection provides one potential explanation for the rapid fall in the incidence of S. pneumoniae septicaemia in older children after a period of recurrent colonisation as infants.
Which S. pneumoniae
antigens are the targets for protective responses in our model is not clear. Both capsular and protein antigens may induce antibody responses after murine nasopharyngeal colonisation 
. However, colonisation with S. pneumoniae
mutant strains not able to express the capsule or seemingly immunodominant protein antigens were still able to induce protective responses 
suggesting antigens inducing protective antibody responses are partially redundant. We are currently identifying which antibodies dominate the immune response to colonisation in this model, and which are the important antibodies in effecting protection.
To conclude, we have investigated the mechanisms of protection against subsequent pneumonia induced by nasopharyngeal colonisation with S. pneumoniae. In our model although colonisation results in a more rapid inflammatory response during early lung infection and a significant CD4-dependent IL-17 response, neither was necessary for the powerful protection against fulminant pneumonia provided by prior colonisation. Instead, protection was due to serum antibody responses that promoted rapid clearance of S. pneumoniae from the blood.