Exposure to S. aureus
has a wide variety of outcomes, and there is strong evidence that host genetics play an important role (von Kockritz-Blickwede et al. 2008
). We therefore characterized 13 strains of inbred mice for four measures of infection severity and utilized HAM (Pletcher et al. 2004
) to identify a single region on chromosome 7 significantly associated with colonization of the kidney by S. aureus
and containing a family of candidate genes.
Across the 13 mouse strains, we observe a wide variety of phenotype values that are consistent with what is known about S. aureus
pathogenesis and the host immune response. High bacterial counts in the kidneys are indicative of bacterial dissemination and are correlated with kidney dysfunction in mouse strains susceptible to S. aureus
infection (Deshmukh et al. 2009
; von Kockritz-Blickwede et al. 2008
). High bacterial counts in the peritoneal fluid are indicative of deficient host clearing of bacteria (von Kockritz-Blickwede et al. 2008
). Similarly, low levels of IL-6 may be indicative of a deficient host immune response, as IL-6 is an important mediator of inflammation and activator of neturophils and has been shown to be required for successful defense against bacterial pathogens, such as Streptococcus pneumoniae
(van der Poll et al. 1997
) and Listeria monocytogenes
(Kopf et al. 1994
). Taken together, these data indicate that the highly susceptible strains do not effectively clear S. aureus
from the site of infection, do not mount an effective IL-6–mediated inflammatory response, and are subject to extensive bacterial dissemination.
We applied HAM as described in Pletcher et al. (2004)
to the four sets of phenotype data and detected one significant association, which was between bacterial colonization of the kidney and a single haplotype block, B7
, on chromosome 7. There is strong experimental evidence that B7
is in fact linked to one or more causal variants: in a separate study conducted by our group, the presence of a causal variant on chromosome 7 was demonstrated using consomic mice created from the highly susceptible A/J strain and the highly resistant C57BL/6J strain. Mice from the consomic mouse strain created by replacing the C57BL/6J chromosome 7 with the A/J chromosome 7 (Nadeau et al. 2000
) were more susceptible than C57BL/6J to S. aureus
infection, with a median survival time of 2.5 days (Ahn et al. 2010
does not contain any genes but lies within the gene cluster of the 26-member extended kallikrein gene family. The kallikrein proteins (KLK) have well-established roles in the degradation of extracellular matrix (ECM), the generation of antimicrobial peptides, and the regulation of immune responses, particularly inflammation [reviewed in Morizane et al. (2010)
, Sotiropoulou and Pampalakis (2010)
, and Sotiropoulou et al. (2009)
]. Degradation of ECM facilitates the infiltration by immune cells of the skin and other tissues. Thus, reduced ECM degradation by KLK enzymes could inhibit infiltration of the site of infection by host immune cells. Similarly, enhanced ECM degradation by KLK enzymes could facilitate S. aureus
dissemination. The proteolytic activities of KLKs are important for the generation of antimicrobial peptides, particularly cathelicidins and defensins, which directly kill microbes as well as influence innate immune response processes. In this case also, reduced KLK activity would weaken the host immune response to S. aureus
KLKs play an important role in the regulation of inflammation, particularly through activation of the IL-1β precursor and the potent vasoactive peptides bradykinin and kallidin (Moreau et al. 2005
). Thus, alterations of KLK activity could result in dysregulation of the inflammatory response. We and von Kockritz-Blickwede et al. (2008)
observed increased kidney infiltration by S. aureus
in susceptible mice relative to resistant mice. In addition, von Kockritz-Blickwede et al. (2008)
observed increased lung infiltration by S. aureus
and erythrocytes, as well as evidence of extensive lung hemorrhage in A/J mice. These observations are all consistent with severe, increased microvascular permeability in susceptible mice in response to S. aureus
infection. Further, von Kockritz-Blickwede et al. (2008)
observed increased levels of serum bradykinin in A/J mice as well as decreased activated partial thromboplastin time. Both observations provide evidence that susceptible mice experience increased microvascular permeability as a result of increased activation of the kallikrein-kinin or contact system.
Our gene expression microarray results provide further evidence for the role of these genes in mediating host susceptibility to S. aureus infection. Given the large number of Klk genes, however, the varied and wide-ranging functions of their gene products, and the complicated patterns of coregulation via reciprocal- and auto-proteolysis, a series of gene-specific experiments at the nucleic acid and protein levels are required for each of the 26 genes to disentangle their precise role in host susceptibility to S. aureus.
Our study has some limitations. First, we only detected a significantly associated haplotype block for one of four phenotypes. This is likely due to the fact that the HAM approach assumes that phenotypic similarities between mouse strains result from shared underlying genetic variants. Our study phenotypes may result from interactions between many different genetic variants, each of which is shared by only a subset of the strains exhibiting similar phenotypes. In addition, the current study may lack sufficient power to detect the corresponding causal variants. The mouse strain panel used in this study is relatively small. Although some studies have used a similar number of strains and detected significant associations using HAM (Yang et al. 2009
), many other studies have used much larger numbers of strains (Bopp et al. 2010
; Kirby et al. 2010
). It is thus likely that with a larger mouse strain panel, additional associations would be detected.
Finally, survival times were recorded for only five days, resulting in an underestimate of the median survival times, particularly for the resistant strains. This results in an underestimate of the F-statistic sum of squares for the within- and between-haplotype group variability. Although underestimates of the within-group variability may result in false-positive associations, underestimates of the between-group variability may result in false-negative associations.
Despite these limitations, we were able to identify a genomic region significantly associated with susceptibility to S. aureus infection, for which there is strong supporting experimental evidence and which implicates a large gene family whose members are promising candidate genes for future biological validation. Future studies in the mouse will identify specific members of the family with a role in S. aureus pathogenesis in the murine host and will elucidate the specific mechanisms by which the gene products confer susceptibility. These studies in the mouse will be followed by studies to evaluate the role of the corresponding orthologous genes in human susceptibility to S. aureus.