This work establishes an animal model to study the function of Lcn2 in the upper respiratory tract. As well as being an important human pathogen, K. pneumoniae
is a tractable model organism that colonizes the nasopharynx and can be genetically manipulated to produce siderophore mutants that do or do not interact with Lcn2. Colonization with wild-type K. pneumoniae
persists for at least seven days, and induces a robust acute inflammatory response. The fact that intranasal inoculation of awake mice produces colonization, and not pneumonia, is likely because the mice do not aspirate nasopharyngeal contents into their lungs 
Similarly to pneumonia, persistence of K. pneumoniae
in the upper respiratory tract requires ongoing iron acquisition 
. Either Gly-Ent or Ybt are required for maximal colonization. This confirms that the nasal mucosa is an iron-limited environment for bacteria, presumably due to the high lactoferrin concentration in nasal secretions 
. Furthermore, this suggests that colonization by K. pneumoniae
requires bacterial replication, and not simply persistence of the originally inoculated organisms. Having established the need for K. pneumoniae
to acquire iron in the nasopharynx, this model can be exploited to examine how host Lcn2 and the bacterial iroA
locus affect the level of colonization.
By using Lcn2−/−
mice and K. pneumoniae
siderophore mutants, this study demonstrates that Lcn2 inhibits nasopharyngeal colonization by bacteria producing unmodified Ent. Although a seemingly narrow target for antimicrobial activity, Ent is produced by many members of the gram-negative Enterobacteriaceae
. Therefore, Lcn2 may contribute to the tropism of enteric commensals for the gut rather than the respiratory tract. The respiratory tract secretes Lcn2 at basal levels and rapidly upregulates Lcn2 expression during colonization 
. In contrast, the large intestine does not express basal Lcn2 despite the presence of huge numbers of colonizing bacteria 
. However, Salmonella
infection induces Lcn2 production in the intestine, and the ability to utilize Gly-Ent confers a competitive advantage over an iroN
mutant during intestinal inflammation 
. Therefore, pathogenic Enterobacteriaceae
such as Salmonella
and K. pneumoniae
can use Gly-Ent to obtain iron in an otherwise restricted environment.
In the respiratory tract, Lcn2 may provide selective pressure for Ent-independent methods of iron acquisition. Many clinical isolates of K. pneumoniae
produce either aerobactin or Ybt 
in addition to Ent. Lcn2 binds bacillibactin of Bacillus anthracis
, but B. anthracis
also produces the unusual siderophore petrobactin that Lcn2 cannot bind 
. Likewise, pathogens such as S. pneumoniae
and H. influenzae
appear to acquire iron in the respiratory tract without producing siderophores 
In the pathogenic K. pneumoniae strain used here, either Gly-Ent or Ybt can support colonization of the nasopharynx. However, Ybt cannot support growth in serum. This defect could be due to an inability of Ybt to strip iron off of transferrin, or a serum component other than Lcn2 that inhibits Ybt-mediated iron acquisition. In contrast, the iroA locus supports robust growth in the presence of Lcn2. Therefore, Ybt and Gly-Ent are not functionally redundant in K. pneumoniae, and likely reflect adaptation to growth in the disparate environments of the respiratory, urinary, and intestinal tracts and the bloodstream.
K. pneumoniae nasal colonization induces a robust inflammatory response characterized by a rapid influx of neutrophils. Neutrophils limit colonization of wild-type K. pneumoniae and prevent hematogenous spread to the spleen. Despite producing Lcn2, neutrophils likely inhibit K. pneumoniae by predominantly Lcn2-independent mechanisms based on the following observations. In the presence of neutrophils, colonization by wild-type K. pneumoniae is the same in Lcn2+/+ and Lcn2−/− mice. In contrast, depletion of neutrophils causes a 10-fold increase in wild-type colonization. Finally, mucosal Lcn2 is able to inhibit colonization of iroA ybtS K. pneumoniae despite depletion of neutrophils. Together, these data indicate that iron sequestration by mucosal Lcn2 is complementary to the antimicrobial actions of neutrophils.
Lcn2 also appears to induce a pro-inflammatory response from respiratory cells when bound to Ent. In vitro, Ent combined with Lcn2 causes a synergistic release of IL-8 from human respiratory cells, but Gly-Ent combined with Lcn2 does not. IL-8 is a neutrophil attracting chemokine, suggesting that signaling by Ent-Lcn2 leads to the recruitment of neutrophils. Accordingly, Ent-producing K. pneumoniae induce nasopharyngeal neutrophil influx in an Lcn2-dependent manner.
The potential signaling pathway between Lcn2, Ent and neutrophil recruitment is unknown. There is no direct IL-8 (CXCL-8) homologue in the mouse 
, although there are several CXC chemokines such as Mip-2, KC, and LIX that have been shown to be induced by K. pneumoniae
respiratory infections 
. Studies to determine the effects of Ent and Lcn2 on murine CXC chemokine production from the murine respiratory mucosa are underway.
The data from this study suggest the following model in which Lcn2 monitors iron utilization by bacteria and activates the immune response when iron stores become depleted: At low bacterial density, secreted Ent binds Fe and Lcn2 in turn binds Fe-Ent to prevent delivery of iron to bacteria. The Fe-Ent-Lcn2 complex is internalized by respiratory epithelial cells 
, and could serve as both a signal of controlled colonization and a mechanism of iron recycling. We hypothesize that as bacterial density increases Lcn2 becomes a pro-inflammatory signal. When bacterial growth outpaces Fe availability, an increased proportion of Ent will be aferric. This aferric Ent-Lcn2 complex could be internalized and serve as a signal of uncontrolled bacterial replication to the respiratory epithelium. In vitro
, respiratory cells respond to Ent-Lcn2 by induction of chemokines, an effect that could explain the increased neutrophil influx seen in response to Ent-producing K. pneumoniae in vivo
The combined data from neutrophil and bacterial counts are consistent with the hypothesis that Lcn2 has a continuum of iron-sequestering and pro-inflammatory activities. Specifically, Ent-producing K. pneumoniae
elicit neutrophils in Lcn2+/+
mice without showing a defect in colonization. The bacterial density of colonization by K. pneumoniae
is relatively low (~1e4
CFU/ml) compared to counts from the pneumonia model 
. If Fe is not depleted during colonization, then Lcn2 may bind primarily Fe-Ent. The small percentage of aferric Ent-Lcn2 could be sufficient to produce a modest increase in the number of neutrophils elicited by K. pneumoniae
colonization. Whereas total depletion of neutrophils leads to increased bacterial numbers (), this incremental increase in neutrophils is not sufficient to affect the density of colonizing organisms. If a perturbance in the microenvironment caused a large increase in bacterial levels, we would predict a greater pro-inflammatory response elicited by Ent-Lcn2. Alternatively, if K. pneumoniae
reaches the lower respiratory tract where it can replicate to high numbers (>109
, there may be a dramatic increase in Ent-Lcn2 formation with a more significant effect of Lcn2 on the immune response. Consistent with this model, Chan and colleagues report that Lcn2 limits growth of K. pneumoniae
ATCC 43816 (the parent strain of our wild type KPPR1) during pneumonia 
. Since Ent is dispensable for growth during pneumonia from KPPR1 
, iron sequestration by Lcn2 cannot account for the observed growth inhibition. This suggests that Lcn2 controls bacterial growth by an additional mechanism such as neutrophil recruitment. Whether Ent production stimulates Lcn2-dependent inflammation during pneumonia remains to be determined. Studying the pro-inflammatory effects of Lcn2 on the mucosal surface could reveal a new paradigm of innate immune signaling in response to bacterial metabolism.