Calpains promote neutrophil recruitment and bacterial clearance in an acute bacterial peritonitis model

Authors

  • Vijay Kumar,

    1. Department of Biomedical and Molecular Sciences, Queen's University, Kingston, Ontario, Canada
    Search for more papers by this author
  • Stephanie Everingham,

    1. Department of Biomedical and Molecular Sciences, Queen's University, Kingston, Ontario, Canada
    2. Division of Cancer Biology and Genetics, Queen's Cancer Research Institute, Kingston, Ontario, Canada
    Search for more papers by this author
  • Christine Hall,

    1. Division of Cancer Biology and Genetics, Queen's Cancer Research Institute, Kingston, Ontario, Canada
    Search for more papers by this author
  • Peter A. Greer,

    1. Division of Cancer Biology and Genetics, Queen's Cancer Research Institute, Kingston, Ontario, Canada
    2. Department of Pathology and Molecular Medicine, Queen's University, Kingston, Ontario, Canada
    Search for more papers by this author
  • Andrew W. B. Craig

    Corresponding author
    1. Department of Biomedical and Molecular Sciences, Queen's University, Kingston, Ontario, Canada
    2. Division of Cancer Biology and Genetics, Queen's Cancer Research Institute, Kingston, Ontario, Canada
    • Full correspondence Dr. Andrew W. B. Craig, Department of Biomedical and Molecular Sciences, Queen's University, 18 Stuart St, Botterell Hall, 3rd Fl., Kingston, Ontario K7L 3N6, Canada

      e-mail: andrew.craig@queensu.ca

    Search for more papers by this author

Abstract

Activation of the innate immune system is critical for clearance of bacterial pathogens to limit systemic infections and host tissue damage. Here, we report a key role for calpain proteases in bacterial clearance in mice with acute peritonitis. Using transgenic mice expressing Cre recombinase primarily in innate immune cells (fes-Cre), we generated conditional capns1 knockout mice. Consistent with capns1 being essential for stability and function of the ubiquitous calpains (calpain-1, calpain-2), peritoneal cells from these mice had reduced levels of calpain-2/capns1, and reduced proteolysis of their substrate selenoprotein K. Using an acute bacterial peritonitis model, we observed impaired bacterial killing within the peritoneum and development of bacteremia in calpain knockout mice. These defects correlated with significant reductions in IL-1α release, neutrophil recruitment, and generation of reactive oxygen species in calpain knockout mice with acute bacterial peritonitis. Peritoneal macrophages from calpain knockout mice infected with enterobacteria ex vivo, were competent in phagocytosis of bacteria, but showed impaired clearance of intracellular bacteria compared with control macrophages. Together, these results implicate calpains as key mediators of effective innate immune responses to acute bacterial infections, to prevent systemic dissemination of bacteria that can lead to sepsis.

Introduction

Bacterial infection triggers pattern-recognition receptor (PRR) signaling within host cells designed to allow rapid bacterial clearance and avoid progression to systemic infection [1]. In bacterial peritonitis, resident innate immune cells in the peritoneum (e.g. mast cells, macrophages) orchestrate the immune response via rapid release of preformed and de novo generated vasoactive mediators, cytokines, and chemokines [2-4]. This leads to recruitment of neutrophils and bacterial capture in neutrophil extracellular traps [5]. Bacterial clearance is facilitated by rapid degranulation, production of ROS and reactive nitrogen species, and phagocytosis by neutrophils and macrophages [6]. Balancing this potent immune response is critical to eliminate the pathogens prior to systemic dissemination and avoid tissue damage observed in severe sepsis [3]. Failure to tightly control production of proinflammatory cytokines such as IL-1, are also linked to the development of autoinflammatory disorders [7]. Thus, a better understanding of positive and negative regulation of PRR signaling may inform the development of more effective treatments for these diseases.

Calpains are intracellular, calcium (Ca++)-dependent cysteine proteases implicated in diverse physiological processes, including innate and adaptive immunity [8, 9]. Calpain-1 and calpain-2 (encoded by Capn1 and Capn2, respectively) are widely expressed proteases that heterodimerize with a common small subunit (encoded by Capns1) that is required for calpain stability [10]. Calpastatin is an endogenous inhibitor of calpains that binds and obstructs the active site of the protease domain [11, 12]. Calpastatin is upregulated in macrophages following activation of TLRs, providing a negative feedback mechanism to limit calpain activity likely invoked by calcium flux in activated innate immune cells [13]. Studies using calpain inhibitors, gene silencing, or gene knockout approaches have led to identification of a vast array of putative calpain substrates linked to regulation of cell growth, survival, and motility (reviewed in [14]). Calpain-mediated cleavage of Talin in focal adhesions has been linked to increased cell motility by promoting trailing edge retraction in fibroblasts and neutrophils [15-17]. Calpains also promote TNF-α-induced neutrophil arrest and ROS production in vitro [18]. In macrophages, calpain activation leads to cleavage of Selenoprotein K (SelK) [13], a transmembrane endoplasmic reticulum protein implicated in regulating Ca++ flux, ROS production, and cell motility [19]. Calpain activation also allows Group B Streptococcus to evade killing by inducing caspase-independent apoptosis, disruption of the cytoskeleton, and phagosome escape in macrophages [20, 21]. Another key calpain substrate relevant to chronic inflammation and immunity is IL-1α, which lacks a signal peptide, undergoes proteolysis in the cytoplasm, and is secreted by both activated and dying cells [22]. Unlike its family member IL-1β that is cleaved by inflammasome-associated Caspase-1 [23], IL-1α is cleaved by calpains or calpain-like proteases [24]. However, Caspase-1 is required for IL-1α secretion from dendritic cells that are treated with soluble activators of inflammasomses, but not with particulate activators like uric acid crystals that induce gout [25]. It is worth noting that Caspase-1 may indirectly enhance calpain activity and IL-1α cleavage or secretion in this model due to its ability to proteolyse Calpastatin [26]. In recent studies, calpain inhibitors were shown to block proteolysis and secretion of IL-1α, and the release of IL-1α was required for neutrophil recruitment to the peritoneum in mice injected with uric acid crystals [25]. Thus, calpains are likely to play multiple roles coordinating innate immune cell recruitment, activation and function in resolving bacterial infections.

Calpain activation promotes transepithelial migration of neutrophils in response to TLR-2 agonists in vitro and in Pseudomonas aeruginosa-infected mice [27]. In Shigella flexneri-infected epithelial cells, calpain activation allows Shigella to avoid genotoxic stress-induced apoptosis via calpain-mediated cleavage of p53, but ultimately triggers necrotic cell death [28]. Thus, calpains are emerging as key regulators of PRR signaling during acute bacterial infection. However, calpains also enhance tissue damage during severe sepsis [29, 30]. Recently, transgenic mice overexpressing Calpastatin were shown to be protected from tissue damage and disseminated intravascular coagulation in the cecal ligation and puncture (CLP) model of sepsis [31]. In this study, Calpastatin was shown to reduce the proinflammatory response, and the levels of procoagulant circulating microparticles. These results suggest that calpain inhibitors may benefit sepsis patients. However, a recent study of muscle wasting in human sepsis patients found no elevation in calpain activity [32]. Thus, a more comprehensive understanding of calpain functions is needed in acute and chronic bacterial infection models. Also, given the potential lack of specificity of some calpain inhibitors, it will be important to compare these findings with calpain deletion or silencing. Although null mutations in Capns1 are embryonic lethal (due to loss of calpain-1 and calpain-2) [33, 34], a conditional Capns1 knockout (KO) model has been developed that allows study of adult mice with inducible or tissue-specific calpain deficiency [35].

In this study, Capns1 was deleted via Cre expression in fes-expressing cells (primarily hematopoietic progenitors, macrophage, neutrophils, and mast cells), and these calpain KO mice were tested in an acute bacterial peritonitis model. Compared with WT mice that effectively cleared bacteria from the peritoneum within 6 h of infection, calpain KO mice showed a significant defect in bacterial clearance and evidence of systemic infection. Within the peritoneum of infected mice, calpains were required for IL-1α cleavage/secretion and enhanced early release of TNF-α. These defects correlated with delayed neutrophil recruitment and reduced ROS production in KO mice, which may explain their impaired response to bacterial infection. Calpains also promoted bacterial killing response of peritoneal macrophages infected with enterobacteria ex vivo, thus implicating calpains in enhancing both neutrophil and macrophage responses to bacterial infection.

Results

Calpain knockout mice have impaired bacterial killing in an acute bacterial peritonitis model

To study roles of the ubiquitous calpains in innate immunity, we crossed Capns1fl/fl mice [35] with transgenic mice expressing Cre recombinase under the control of the human fes gene [36]. Fes is most highly expressed in hematopoietic progenitors, macrophages, neutrophils, and mast cells [37-39]. The resulting capns1fl/fl:fes-Cre mice (hereafter referred to as KO) had no overt defects and no differences in cellularity of bone marrow (BM), blood, or resident peritoneal cells (PCs; data not shown). PCR genotyping was performed to assess the efficiency and specificity of Capns1 deletion and Cre expression on genomic DNA from BM, PC, and neutrophils (PMNs). As expected, only the floxed capns1 allele was detected in capns1fl/fl (fes-Cre-negative) mice, which served as WT controls (Fig. 1A). However, in calpain KO mice the capns1 null allele was the major product in BM, PCs, and PMNs (Fig. 1A). At the protein level, we observed a significant reduction in both Capns1 and calpain-2 in peritoneal macrophage (PMac) cell lysates by immunoblot (Fig. 1B; actin served as a loading control). Deletion of Capns1 was previously shown to cause loss of calpain-1 and calpain-2 catalytic subunits due to protein instability associated with loss of this common regulatory subunit [33]. Next, we tested for effects of Capns1 deletion on calpain substrate levels in PMac cell lysates. SelK was recently identified as a substrate of calpain-2 downstream of TLR's in macrophages, neutrophils, and dendritic cells (DCs), but not in lymphoid cells [13]. Indeed, we observed increased levels of SelK in KO PC lysates compared with WT, along with mostly full length SelK in KO compared with primarily cleaved SelK in WT PC lysates (Fig. 1C, see arrows; actin served as a loading control). Taken together, this novel model of calpain deficiency shows differences in a calpain substrate that is known to regulate innate immune responses [19].

Figure 1.

Fes-cre driven calpain knockout mice are defective in bacterial clearance during acute bacterial peritonitis. (A) Genomic DNA was isolated from bone marrow (BM) stromal cells, peritoneal cells (PCs), and BM-derived neutrophils (PMNs; isolated by FACS using CD11bhi/Gr1hi) from WT and calpain KO mice, and subjected to PCR analysis of capns1 flox and null alleles (top panel). In separate PCR reactions, the presence of Fes-Cre transgene was analyzed for the same samples. (B) Lysates from WT and KO peritoneal macrophages (PMac) were analyzed by immunoblot (IB) for Capns1, Capn2, and β-actin (loading control). Positions of molecular mass markers are shown on the left. (C) Lysates from WT and KO PMac were subjected to IB analysis of calpain substrate SelK (actin served as a loading control). Note that SelK antisera detects both the full-length SelK and its cleavage product (indicated by closed and open arrows). Positions of molecular mass markers are shown on the left. (D) Viable enterobacteria colonies were counted for PLF isolated from sham or FIP-treated WT and KO mice at 2 and 6 h postinjection. Graph represents mean ± SD for 3–6 mice/genotype for each treatment (** indicates a significant difference (p < 0.01) between genotypes using paired t-test). (E) Viable enterobacteria colonies were counted for peripheral blood isolated from sham or FIP-treated WT and KO mice at 6 h postinjection. Graph represents mean ± SD for three mice/genotype for each treatment. ** indicates a significant difference (p < 0.01) between genotypes genotypes using paired t-test. (A–E) Data shown are representative of results from three independent experiments.

To assess the role of calpains in the innate immune response in vivo, we subjected WT and calpain KO mice to the feces-in-peritoneum (FIP) model of acute bacterial peritonitis [40]. Colonic enterobacteria were cultured overnight prior to injection in the peritoneal cavity of WT and calpain KO mice (106 CFU). In WT mice, the peritoneal lavage fluid (PLF) was cleared of viable enterobacteria within 6 h of infection, whereas calpain KO mice showed significant defects in bacterial killing compared with WT (Fig. 1D). As expected, no viable bacteria were detected in PLF from sham treated mice (Fig. 1D). Interestingly, calpain KO mice subjected to FIP showed evidence of enterobacteria dissemination into peripheral blood at 6 h postinfection, while no bacteria were observed in FIP-treated WT mice (Fig. 1E). Thus, calpains are required for effective clearance of enterobacteria during acute peritonitis to prevent systemic infections, and potential progression to sepsis.

Calpain KO mice have impaired production of mature IL-1α in the FIP model

Activation of PRR's on host cells leads to both rapid release of preformed mediators and upregulation of cytokines and chemokines that coordinate the immune response [1]. To identify molecular mechanisms for the observed bacterial killing defects in calpain KO mice, we tested for defects in cytokine production. Both WT and calpain KO mice subjected to FIP showed elevated IL-6 production compared with sham controls (Fig. 2A). Although calpain KO mice produce similar levels of IL-6 at early times, we observed a significant reduction at 6 h postinfection compared with WT mice (Fig. 2A). In contrast, early release of TNF–α was reduced in calpain KO compared with WT mice subjected to FIP (1 h) but was similar between genotypes at later times (Fig. 2B). The most dramatic differences in calpain-dependent cytokine production we observed were for IL-1α, which was undetectable at early times and significantly reduced throughout the FIP challenge in calpain KO mice compared with WT (Fig. 2C). Interestingly, we observed no differences in IL-1b levels in PLF from calpain KO mice subjected to FIP (Fig. 2D). While our observed differences in TNF-α and IL-6 production in calpain KO mice is likely due to calpains signaling downstream of PRR's, IL-1α production defects are consistent with previous reports of calpain-mediated cleavage of pro-IL-1α (p33) to mature IL-1α (p17) [24, 25]. To test this, PMac lysates from FIP-treated WT and KO mice (1 h) were subjected to immunoblot for IL-1α. Interestingly, calpain KO cells showed increased levels of pro-IL-1α and reduced levels of mature IL-1α, compared with WT cells (Fig. 2E; actin served as a loading control). Since both pro-IL-1α and mature IL-1α can be secreted, we examined their relative levels in PLF from sham or FIP-treated WT and calpain KO mice. As expected, no IL-1a was detected in sham PLF, whereas both forms of IL-1a were detectable at 1 h post-FIP treatment (Fig. 2F). Interestingly, WT PLF contained primarily mature IL-1a, whereas calpain KO PLF contained mostly pro-IL-1a (Fig. 2F). These results are consistent with calpains promoting IL-1α production during bacterial infection via direct cleavage of pro-IL-1α in innate immune cells.

Figure 2.

Calpains promote TNF-α and IL-1α release during acute bacterial peritonitis. (A) IL-6 levels were measured in PLF from sham and FIP-treated WT and KO mice at 1, 2, and 6 h postinjection by ELISA. ** indicates a significant difference (p < 0.01) between genotypes genotypes using paired t-test. (B) TNF-α levels were measured in PLF from sham and FIP-treated WT and KO mice at 1, 2, and 6 h postinjection by ELISA. * indicates a significant difference (p < 0.05) between genotypes genotypes using paired t-test. (C) IL-1α levels were measured in PLF from sham and FIP-treated WT and KO mice at 1, 2, and 6 h postinjection by ELISA. ** indicates a significant difference (p < 0.01) between genotypes using paired t-test. (D) IL-1β levels were measured in PLF from sham and FIP-treated WT and KO mice at 3 h postinjection by ELISA. (A–D) Graphs represent mean ± SD for 3–6 mice/genotype for each treatment and represent two independent experiments performed with similar results. (E) Lysates from PMac isolated from FIP-treated WT and KO mice were subjected to IB analysis for pro-IL-1α (p33) and mature IL-1α (p17; actin served as loading control). Positions of molecular mass markers are shown on the left. (F) PLF from sham or FIP-treated WT and KO mice (1 h) were subjected to IB analysis for IL-1α as described above. Positions of molecular mass markers are shown on the left. (E–F) Data shown are representative of results from two independent experiments.

Calpain KO mice show defects in neutrophil recruitment and ROS production in the FIP model

Neutrophil recruitment to the peritoneal cavity is an early response to acute bacterial peritonitis [1], and defects in this response may explain defects we observed in bacterial killing response of calpain KO mice. Also, our observed defects in IL-1α production in calpain KO mice may lead to defects in neutrophil recruitment given recent studies of IL-1α function in promoting neutrophil recruitment to the peritoneum during sterile inflammation [25]. Although both p33 and p17 forms were detected in PLF, a more recent study demonstrates that the calpain-cleaved form of IL-1α (p17) is much more active than p33 in eliciting neutrophils [41]. To test whether calpains are required for neutrophil recruitment during FIP, we compared the kinetics of neutrophil recruitment between WT and KO mice by FACS staining of Gr-1hi/CD11bhi cells in the peritoneum. As expected, neutrophils were elicited in FIP-treated WT mice compared with sham (Fig. 3A). Interestingly, calpain KO mice showed significant defects in absolute numbers of neutrophils recruited to the peritoneum in the FIP model (Fig. 3B). This defect in neutrophil recruitment may contribute to the impaired bacterial killing response in calpain KO mice.

Figure 3.

Calpains promote neutrophil recruitment and ROS production during acute bacterial peritonitis. (A) PCs were isolated from sham and FIP-treated WT and KO mice and stained with CD11b and Gr1 antibodies. Flow cytometry was performed to detect neutrophils (CD11bhiGr1hi) and representative histograms are shown for indicated treatments, genotypes and time points (hours). (B) Absolute neutrophil numbers are shown as mean ± SD for three mice/genotype for each treatment and are pooled from three independent experiments performed with similar results. * indicates a significant difference (p < 0.05) between genotypes using paired t-test. (C) PCs were isolated in sham or FIP-treated WT and KO mice (1 h postinjection) and ROS generation was detected following incubation of samples with CM-H2DCFDA and flow cytometry. A representative histogram is shown. (D) Graph represents the fold change in CM-H2DCFDA MFI in PCs from FIP treated mice relative to sham controls for six mice/genotype (mean ± SD is shown along with individual data for each mouse) and are pooled from two independent experiments performed with similar results. * indicates a significant difference (p < 0.05) between genotypes using paired t-test.

PRR signaling in innate immune cells leads to ROS production and bacterial killing [42, 43]. Since calpains were reported to enhance ROS production in neutrophils in vitro [18], we investigated whether calpain KO mice exhibit defects in ROS production during acute bacterial peritonitis in vivo. DCFDA was used to measure levels of ROS by FACS for peritoneal cells isolated from sham and FIP treated WT and KO mice. As expected, at 1 h post-FIP treatment, a dramatic increase in ROS levels were observed in WT mice compared with sham controls (Fig. 3C). In contrast, calpain KO mice failed to produce a dramatic increase in ROS during FIP response (Fig. 3C). Quantification of these results for six separate WT and KO mice revealed significant defects in ROS levels in calpain KO mice during bacterial infection (Fig. 3D). These results are consistent with calpains playing a key role in innate immune cells that produce ROS to enhance bacterial killing. In the absence of calpains, defects in ROS production may facilitate bacterial survival and systemic dissemination.

Calpain KO macrophages show delayed bacterial killing response

Calpain activation downstream of TLRs in macrophages has implicated these proteases in regulating apoptosis and phagocytosis [13, 20, 21]. To test whether defects in FIP response of calpain KO mice relate to defects in macrophages, we isolated peritoneal macrophages from WT and KO mice, and subjected them to entero-bacterial infection ex vivo (using bacterial cultures employed in FIP model). At various time points, we measured viable extra-cellular and intracellular bacteria CFU. Extracellular bacteria counts were significantly higher at early times of infection in KO peritoneal macrophages compared with WT (Fig. 4A), suggesting that calpains enhance either bacteria uptake or extracellular killing response. Since viable intracellular bacteria counts were detected within 15 minutes of infection in KO peritoneal macrophages, as in WT cells, we conclude that calpains are not required for bacterial uptake (Fig. 4B). While the levels of intracellular bacteria diminished between 15 and 60 min in WT macrophages, this intracellular bacterial killing response was significantly impaired in KO macrophages (Fig. 4B). Similar results were observed in bone marrow derived macrophages (data not shown), thus implicating calpains in enhancing bacterial killing via macrophages, which may involve enhancing ROS production, or phagolysosome maturation (Fig. 4C). Since more dramatic differences in bacterial killing were observed in vivo, it is likely that combined roles of calpains in macrophages and neutrophils (and possibly other cell types) contribute to an effective innate immune response that protects the host from systemic dissemination of bacteria during acute infections (Fig. 4C).

Figure 4.

Calpains promote bacterial killing in macrophages. (A) Peritoneal macrophages from WT and KO mice (in antibiotic-free media) were infected ex vivo with FIP enterobacterial culture as described in the Materials and methods. Conditioned media was collected to score viable extracellular bacteria counts at 0, 15, 30, and 60 min postinfection. (B) To assess bacteria uptake and intracellular killing of bacteria, extracellular bacteria were killed prior to permeabilization of peritoneal macrophages infected ex vivo, as described in the Materials and methods. (A, B) Graph represents bacterial count for each time point and data are shown as mean ± SD for triplicate wells/genotype and are representative of two independent experiments performed with similar results. * indicates a significant difference p < 0.05; ** indicates p < 0.01 between genotypes using paired t-test. (C) A schematic model illustrating roles of calpains within innate immune cells responding to bacterial infection. TLR signaling in macrophages leads to calcium flux and calpain activation. This promotes cleavage of pro-IL-1α, and subsequent release of IL-1α promotes neutrophil recruitment. Early production of TNF-α was also enhanced by calpains, and may contribute to neutrophil recruitment and activation. Calpains also enhanced ROS production in neutrophils during acute bacterial peritonitis, which contributes to bacterial clearance. Calpains also modulate the rate of bacterial phagocytosis and intracellular killing in macrophages.

Discussion

Coordination of the innate immune response during acute bacterial peritonitis is crucial to prevention of systemic dissemination and development of sepsis [44]. Here, we report novel insights into the roles of calpain proteases in mounting an effective innate immune response to bacterial infection in mice. Deletion of Capns1 in fes-expressing cells (i.e. macrophages, neutrophils, mast cells) led to defects in proteolysis of calpain-1/calpain-2 substrates in peritoneal cells, and defective innate immunity. Although FIP-treated WT mice effectively cleared enterobacteria within 6 h, calpain KO mice had impaired bacterial killing and developed bacteremia. These defects correlated with a severe reduction in IL-1α cleavage and release into the peritoneum of FIP-treated calpain KO mice, leading to impaired neutrophil recruitment. In addition, calpains functioned in bactericidal responses of innate immune cells, including enhancing ROS production and bacterial killing in macrophages and neutrophils (see Fig. 4C for schematic model). Overall, our findings identify calpain proteases as central players in bacterial immunity.

IL-1 family cytokines and their receptor IL-1R1 play key roles in immune responses and their dysregulation occurs in metabolic and inflammatory diseases [45]. Therefore, a detailed understanding of mechanisms leading to IL-1 production and IL-1R1 signaling are of critical importance. Compared with the pathway leading to IL-1β production downstream of inflammasomes and Caspase-1, molecular mechanisms leading to IL-1α cleavage and secretion remain controversial [46]. Until recently, IL-1α was thought to be passively released by necrotic cells to serve as a danger signal [47]. However, recent studies show that non-necrotic cells release IL-1α via a regulated pathway involving both the inflammasome and calpain proteases [25, 48]. An early study had identified a calcium-dependent pathway of processing pro-IL-1α (p33) to mature IL-1α (p17) involving calpain proteases in macrophages [24]. More recently, calpain inhibitors were shown to block IL-1α processing and release downstream of activators of the inflammasome, and IL-1α was required for neutrophil recruitment to the peritoneum in a sterile inflammation model [25]. Although release of both p33 and p17 were detected during sterile inflammation, the exact contributions of these cytokines to IL-1R1 signaling was not reported [25]. However, Clarke et al. recently demonstrated that cleavage of p33 to p17 by calpains enhances IL-1R1 binding and elicits neutrophils more efficiently than p33 [41]. They also report a soluble receptor (IL-1R2) that prevents cleavage of p33 until inflammasome activation leads to Caspase-1-mediated proteolysis of IL-1R2, thus allowing subsequent cleavage of p33 by calpains [41]. Our results provide further support for calpain-dependent IL-1α production and neutrophil recruitment during acute bacterial infection in vivo. However, the fact that some mature IL-1a was detected in calpain KO mice likely reflects the contributions of other proteases such as Granzyme B and Chymase in IL-1α processing during FIP [49]. It is worth noting that IL-1R1 signaling is required for neutrophil recruitment and killing of Staphylococcus aureus in a skin infection model [50]. Testing the contributions of calpains to bacterial immunity in skin or other organs will be quite feasible using additional tissue-specific calpain KO models.

Although our results identify calpains as protective during acute bacterial infection, other studies implicate calpains in promoting tissue damage. In a ventilator-induced lung injury model, calpain inhibitors, or gene silencing, led to reduce neutrophil infiltration and edema [51]. In endotoxemia models, calpain activation also leads to damage of cardiac, lung, and muscle tissues in mice [29, 30, 52]. In the CLP model of sepsis, calpain activation peaked within the first 6 h of the immune response [31]. This study also showed that transgenic mice with ubiquitous over-expression of Calpastatin, showed reduced calpain activation and tissue damage following CLP [31]. Interestingly, similar defects in TNF-α and IL-1α production were noted in Calpastatin transgenic mice subjected to CLP [31], as we have observed here in FIP-treated calpain KO mice. While differences in bacterial killing were not observed in the Calpastatin transgenic mice following CLP [31], it is important to note that this may reflect differences in the levels of calpain activity required in chronic (CLP) versus acute (FIP) infection models. Our findings suggest that inhibiting calpains may not improve outcomes for cases of peritonitis, due to enhanced risk of bacteremia and developing sepsis. However, in severe sepsis patients, calpain inhibitors may indeed improve outcomes by reducing tissue damage. Future studies using our calpain KO model in CLP assays will allow direct testing of the contributions of calpains to early and late stages of sepsis pathophysiology.

In this study, we observed defects in early release of TNF-α in FIP-treated calpain KO mice compared with WT controls. Although this defect was not observed at later times of infection, it likely contributes to the delayed neutrophil recruitment and bacterial killing in calpain-deficient mice. Prior studies have implicated mast cells as responsible for early TNF-α release in peritonitis and CLP models leading to protection from organ damage [53-55]. However, mast cell-derived TNF-α can also exacerbate mortality during severe sepsis [56]. Indeed, the role of mast cells and their mediators in sepsis remains controversial, with findings ranging between protective and destructive roles [57]. The reliance on Kit mutant mice (e.g. KitW/Wv, KitWsh/Wsh) as models of mast cell deficiency have likely contributed to these discrepancies, since defects extend beyond mast cell deficiency in these models. With the recent description of Cre/LoxP-driven models of mast cell deficiency [58, 59], it is now possible to more accurately identify the roles of mast cells and their mediators in bacterial immunity and sepsis. It will be interesting to test the contributions of calpain signaling in mast cells to mediator release using our calpain KO mice. Reconstitution of mast cell deficient mice with WT and calpain KO mast cells will allow for direct testing of the contributions of calpain proteases in mast cell responses during peritonitis or sepsis.

Calpains also regulate phagocytosis and bacterial killing in macrophages. A recent study of Group B Streptococcus (GBS) infected macrophages showed that calpain activation led to proteolysis of a number of key regulators of the actin cytoskeleton and microtubules [21]. In GBS-infected macrophages, calpain inhibitors, or gene silencing, led to 50-fold higher levels of viable intracellular GBS compared with control cells. These results suggest that calpain activation leads to enhanced bacterial killing, likely by promoting phagolysosome maturation. Calpains are also required for phagosomal escape of Listeria monocytogene infected macrophages, and enhance release of both IL-1α and IL-1β [60]. Although we did not observe this in our FIP model in vivo, it will be important to define the contributions of calpains to phagocytosis of a number of bacterial pathogens using this calpain KO model in future. Although calpains contributed to bacterial killing in macrophages infected ex vivo, it is worth noting the phenotype was more pronounced in vivo. Thus, calpains likely contribute to bacterial immunity via functioning within multiple arms of the innate immune system (Fig. 4C).

Our results identify calpains as key for ROS production in peritoneal cells isolated at early times post-FIP treatment. At this time point, macrophages and neutrophils are likely to play major roles in ROS production and clearance of bacteria. Thus, calpain KO macrophages and neutrophils failed to mount comparable ROS production to WT control cells, which may explain the delayed bacterial killing we observed. Indeed, ROS production by NADPH oxidase 2 (NOX2) is required to protect mice from spontaneous and recurrent bacterial infections [61]. Our results are also in agreement with a recent study that showed calpain inhibitors dampen ROS production in neutrophils treated with TNF-α [18]. Although the molecular mechanisms for calpain regulation of ROS has not been established, it is worth noting that loss of the calpain substrate SelK also results in neutrophil migration and FcγR-mediated ROS production in macrophages [19]. Although it is clear that calpain cleaves SelK in macrophages [13], it remains to be determined whether the accumulation of uncleaved SelK in our calpain KO mice is relevant to the defects we observed in the FIP assay. Future studies will be required to address the relative contributions of calpain substrates IL-1α or SelK in neutrophil recruitment and bacterial killing (Fig. 4C). In addition to delayed recruitment, calpain KO neutrophils may also possess defects in directed chemotaxis toward their target since neutrophils treated with calpain inhibitors had defects in firm adhesion and polarized motility toward a TNF-α stop signal [18].

Taken together, our study identifies calpain proteases as necessary for rapid response to a localized bacteria infection. Loss of calpains leads to delayed bacterial killing and systemic dissem-ination, which could lead to development of sepsis. Although calpain inhibitors may protect from tissue damage in severe sepsis, gout, and ventilator-associated acute lung injury [30, 31, 52], these inhibitors may have unintended, negative effects on the immune system. Indeed, the results of this study suggest that inhibiting calpains may compromise the host immune response to acute bacterial infection.

Materials and methods

Transgenic mice

Capns1fl/fl mice [35] were crossed with Fes-Cre transgenic mice, which were previously reported to delete floxed alleles in hematopoietic cells [36]. Littermates of homozygous Capns1fl/fl mice with or without the Fes-Cre transgene (129/SvJ background) were used as calpain KO or WT mice, respectively. To characterize the degree of capns1 deletion in relevant cell types, genomic DNA was isolated from bone marrow, peritoneal cells and bone marrow neutrophils (CD11b+Gr1+ isolated by FACS). Samples were analyzed for the presence of capns1 flox and deleted alleles as previously described [62]. Genotyping was conducted using Capns1 allele primers were Ex8 for: GAA CTT CCA GGG GCC TTT GAG, Ex9 rev: GTT TGG TCT CAG GGC CCC AGC, Ex11 rev: GGT GGG GTG ACC CTT CAG TAG. Detection of Fes-Cre transgene was conducted using CreF: GAC GGA AAT CCA TCG CTC GAC CAG and CreR: GAC ATG TTC AGG GAT CGC CAG GCG primers.

Acute bacterial peritonitis model

Acute bacterial peritonitis was induced using a slightly modified FIP model, which was described previously [40]. Enterobacteria were obtained from the colon of a mouse and incubated overnight at 37°C in Luria broth (LB), bacteria were pelleted and resuspended in saline at 0.1 od (600 nm). Intraperitoneal injections of 1 mL saline (sham) or 0.1 od bacterial culture (FIP) were performed on WT and calpain KO mice (1–4 months, age and sex matched). Mice were sacrificed at 1, 2, and 6 h postinjections (3 mice/genotype for each timepoint). PLF was collected at each time point (in 10 mL saline), and peripheral blood was also collected by cardiac puncture on deeply anesthetized mice at 6 h. All procedures with mice were performed in accordance with Canadian Council for Animal Care guidelines with approval from Queen's Animal Care Committee.

Neutrophil infiltration detection by flow cytometry

Peritoneal cells were isolated by centrifugation of PLF from sham and FIP-treated mice, and after blocking of CD16/32 (IgG receptors) were incubated with phycoerythrin (PE)-labeled CD11b (Cedarlane Biotech) and FITC-labeled Gr-1 antibodies or their isotype controls (Cedarlane Biotech). Up to 104 events were collected on a Cytomics FC500 (Beckman Coulter). Data were analyzed using FlowJo software (Tree Star) with gating set to exclude dead cells or cell debris (low forward and side scatter). Isotype control samples were used to set threshold values for CD11bhiGr-1hi neutrophils.

ROS generation assays

Generation of ROS within peritoneal cells isolated from sham or FIP-treated WT and calpain KO mice (1 h) was measured as described (West et al., 2011). At 1 h postsham or FIP injection of WT and calpain KO mice, PLF was collected and cells isolated by centrifugation. Peritoneal cells were washed with PBS and incubated with CM-H2DCFDA (Invitrogen) at 2.5 μM final concentration in serum-free DMEM (Sigma-Aldrich) for 15–30 min at 37°C in a 12 well plate. Cells were washed with warmed PBS (37°C), removed from plates with cold PBS containing 1 mM EDTA by pipetting. Cells were resuspended in cold PBS and up to 104 events were collected on a Cytomics FC500 (Beckman Coulter). Data were analyzed using FlowJo software (Tree Star) with gating set to exclude dead cells or cell debris (low forward and side scatter). Unstained cells were used to set appropriate thresholds, and the MFI values for CM-H2DCFDA staining were reported for FIP relative to WT sham control (fold change).

Cytokine production during FIP

The amounts of IL-1α, TNF–α, and IL-6 in PLF collected from sham and FIP-treated mice were measured by ELISA according to manufacturer's instruction (e-Biosciences). All assays were performed in triplicate, with 3 mice/genotype at each time post-FIP treatment.

Measurement of viable bacteria in PLF and blood

Serial dilutions were prepared for 0.1 od enterobacteria culture, and for PLF and peripheral blood from sham and FIP-treated mice (101–106 dilutions in sterile PBS) and plated on AC Agar (All Culture Agar, US Biological). Plates were incubated at 37°C overnight, and bacterial colonies were counted.

Peritoneal macrophage cultures and bacterial killing assays

Peritoneal macrophage cultures were established by removing nonadherent peritoneal cells from WT and KO mice after 1 h culture in complete macrophage media, as previously described [63]. WT or KO peritoneal macrophages in antibiotic-free media were incubated with 100 μL of 0.1 od enterobacterial culture (prepared as described above) for 0–60 min (triplicate wells). At each time point, media was removed for counting of viable extra-cellular bacteria as described above. For intracellular bacteria counts, cells were washed three times with PBS and then treated with gentamicin (5 μg/mL in DMEM supplemented with 10% FBS) for 20 min to eliminate extracellular bacteria. Cells were then lysed in (500 μL vol.) lysis buffer (PBS/0.1% SDS), and then viable bacteria counts performed as described above.

Immunoblotting

Lysates from WT and KO peritoneal macrophages and BMMs were prepared and immunoblotting performed for Capns1 and Capn2 as previously described (Ho et al., 2012). In addition, immuno-blotting for calpain substrates was performed with Selenoprotein K (Sel K) antisera (Abgent; 1:500), HIF-2α antisera (Novus Biological; 1:1000), and IL-1α antisera (Clone ALF-161; eBioscience; 1:1000). Horseradish peroxidase-conjugated goat anti-rabbit IgG (1:10 000; GE Healthcare) or rabbit anti-hamster IgG (for IL-1α; 1:1000) were used for detection by enhanced chemiluminescence (Applied Biological Materials). Note that for analysis of IL-1α, peritoneal macrophages were isolated from WT and KO mice at 1 h post-FIP treatment.

Statistical analysis

Data were analyzed with commercial software (Microsoft Excel and GraphPad). For pairwise comparisons between WT and calpain KO mice, two-tailed Student's t-test was used. Results were considered statistically significant with p values of less than 0.05. Unless stated otherwise, all experiments were repeated at least three times and representative data are presented.

Acknowledgments

We thank Jalna Meens and staff of the Queen's Cytometry and Imaging facility for their assistance. This work was supported by operating grants from Canadian Institutes for Health Research (MOP82882) to A.W.B.C, and P.A.G (MOP 81189).

Conflict of interest

The authors declare no financial or commercial conflict of interest.

Abbreviations
BM

bone marrow

CLP

cecal ligation and puncture

FIP

feces-in-peritoneum

GBS

Group B Streptococcus

KO

knockout

PLF

peritoneal lavage fluid

PRR

pattern-recognition receptor

SelK

Selenoprotein K

Ancillary