Calcium/calmodulin-dependent protein kinase (CaMK) Iα mediates the macrophage inflammatory response to sepsis

CaMKIα regulates systemic inflammation and organ dysfunction during sepsis by regulating secretory lysosomal HMGB1 release and IL-10 production in macrophages exposed to LPS.

Abstract

Dysregulated Ca²⁺ handling is prevalent during sepsis and postulated to perpetuate the aberrant inflammation underlying subsequent organ dysfunction and death. The signal transduction cascades mediating these processes are unknown. Here, we identify that CaMKIα mediates the Mφ response to LPS in vitro and the inflammation and organ dysfunction of sepsis in vivo. We show that LPS induced active pThr¹⁷⁷-CaMKIα in RAW 264.7 cells and murine peritoneal Mφ, which if inhibited biochemically with STO609 (CaMKK inhibitor) or by RNAi, reduces LPS-induced production of IL-10. Transfection of constitutively active CaMKIα (CaMKI293), but not a kinase-deficient mutant (CaMKI293^K49A), induces IL-10 release. This production of IL-10 is mediated by CaMKIα-dependent regulation of p38 MAPK activation. CaMKIα activity also mediates the cellular release of HMGB1 by colocalizing with and regulating the packaging of HMGB1 into secretory lysosomes. During endotoxemia, mice receiving in vivo CaMKIα^RNAi display reduced systemic concentrations of IL-10 and HMGB1 in comparison with mice receiving NT^RNAi. These data support the biological relevance of CaMKIα-dependent IL-10 production and HMGB1 secretion. In a CLP model of sepsis, CaMKIα^RNAi mice display reduced systemic concentrations of IL-10, IL-6, TNF-α, and HMGB1 in comparison with NT^RNAi mice, which correlate with reductions in the development of renal dysfunction. These data support that CaMKIα signaling is integral to the Mφ responding to LPS and may also be operant in vivo in regulating the inflammation and organ dysfunction consequent to sepsis.

Introduction

Ca²⁺ signaling, usually through the Ca²⁺ receptor protein CaM, is a ubiquitous event, integral to nearly all cellular functions, including death and survival [1, 2]. Mφ signal transduction, consequent to a variety of inflammatory stimuli, is dependent on and can be modulated by intracellular and extracellular Ca²⁺. In Mφ, LPS-induced TNF-α production is inhibited by Ca²⁺ chelation and CaM inhibition, and iatrogenically elevating Mφ intracellular Ca²⁺ augments proinflammatory cytokine production [3–6]. Recently, HMGB1, an architectural chromatin-binding factor that bends DNA and directs protein assembly on specific DNA targets, has been demonstrated to function as a late mediator of mortality in murine endotoxemia and sepsis [7–9]. Mφ have been demonstrated to be a primary source of HMGB1 [9], and evidence is accumulating that production of this inflammatory mediator is Ca²⁺-dependent and regulated by the family of CaMKs [10–12].

The multifunctional CaMKs (CaMKI, -II, and -IV) are a family of serine/threonine kinases sensitive to changes in intracellular Ca²⁺, which coordinate a variety of cellular functions, including gene expression, cell cycle, differentiation, and ischemic tolerance [13, 14]. Each member has a catalytic domain adjacent to a regulatory region that contains an overlapping AID and the CaM-binding domain. Binding of Ca²⁺–CaM induces conformational alterations in the AID, such that it no longer interferes with substrate engagement and activates the kinase. Phosphorylation of a kinase-specific Thr residue [auto-pCaMKII (Thr²⁸⁶) and pCaMKIα (Thr¹⁷⁷), pCaMKIβ (Thr¹⁷⁵), pCaMKIδ (Thr¹⁸⁰), pCaMKIγ (Thr¹⁷⁸), and pCaMKIV (Thr¹⁹⁶) by the upstream CaMKK α/β] markedly augments kinase activity and generates Ca²⁺-independent activity. Whereas CaMKIV has a relatively restricted tissue distribution (testes, brain, bone marrow), CaMKI, comprised of four isoforms (α, β, δ, γ), appears to be expressed in all mammalian cells [15].

Recently, we have begun to characterize a role of the CaMK cascade in mediating the inflammatory phenotype of the monocyte/Mφ. The broad CaMK inhibitor KN62 prevented LPS-induced MAPK activation and TNF-α production in human monocytes [16]. We have subsequently shown that CaMKIV mediates the nucleocytoplasmic shuttling of HMGB1, in part, by performing the prerequisite step of HMGB1 serine phosphorylation [12]. Other studies support our observations and identify additional functions of CaMKIV in TLR4-mediated signaling, including DC survival during endotoxemia and the inflammation consequent to hepatic ischemia/reperfusion [11, 17]. In contrast, few physiological roles of CaMKI have been identified, and whether these kinases are operant in the inflammatory response to sepsis in vivo is unknown. Here, we characterize the distinct role of CaMKIα in the Mφ response to LPS. We extend these observations into in vivo models of surgical sepsis, the results of which emphasize the biological relevance of CaMKIα in regulating the inflammatory response to sepsis and the subsequent sequelae of organ dysfunction.

MATERIALS AND METHODS

Reagents

Ultra Pure LPS (Escherichia coli 0111:B4) was obtained from List Biologicals (Campbell, CA, USA). KN62, obtained from Calbiochem (San Diego, CA, USA), was dissolved in sterile DMSO and used at a concentration of 20 μM. STO609 was obtained from Calbiochem and used at concentrations of 1–20 μM. STO609 is highly specific for CaMKK: it has an in vitro IC₅₀ of 0.13–0.38 μM for CaMKK and 32 μM for CaMKII, with little or no inhibition of CaMKI, CaMKIV, PKA, PKC, ERK, or myosin light-chain kinase [18]. SB203580, a specific inhibitor for p38α and p38β MAPK (IC₅₀=0.6 μM in vitro) [19], with no inhibition of ERK1/2, JNK, MEK3, or MEK6, was purchased from Assay Designs (Ann Arbor, MI, USA) and used at a concentration of 5 μM. PD98059, which indirectly blocks the activation of p44/42 MAPK via inhibition of MEK1 activation by c-Raf with an IC₅₀ = 4 μM [20, 21], was purchased from Calbiochem and used at a concentration of 10 μM, far below the concentration at which untargeted effects have been observed [20, 22, 23]. Phospho-specific antibodies for dual-pp38 and -pERK1/2 were obtained from Cell Signaling (San Diego, CA, USA) and for dual-pJNK from Promega (Madison, WI, USA). mAb for Thr^177/180 pCaMKI were the generous gift of Dr. Naohito Nozaki (Kanagawa Medical College, Kanagawa, Japan) [24]. This mAb has been shown to recognize the active, Thr-pCaMKIα (Thr¹⁷⁷), -pCaMKIβ (Thr¹⁷⁵), and -pCaMKIδ (Thr¹⁸⁰) [24]. Antibodies for total ERK1/2, p38, JNK, CaMKI (α, β, δ, γ), HMGB1, LAMP1, and histone 3 were obtained from Abcam (Cambridge, MA, USA).

Animal experimentation

We performed all animal experiments in accordance with the National Institutes of Health guidelines, under protocols approved by the Institutional Animal Care and Use Committee of the University of Pittsburgh (PA, USA). We randomly grouped 6- to 8-week-old male C57BL/6J mice and assigned them to a specific experiment. Investigators who treated animals knew the treatment groups and collected samples, which were then analyzed by other investigators blinded to the specific treatment.

For our model of endotoxemia, mice were anesthetized with isofluorane (2–4% induction), and LPS (List Biologicals), dissolved in sterile, normal saline, was injected i.p. at a dose of 1.5 mg/kg. We performed CLP by anesthetizing mice with isofluorane (2–4% induction) and ketamine (50 mg/kg, i.p. injection). The surgical site was shaved and sterilely prepped and draped. A 1-cm midline laparotomy incision was made, and the cecum was identified, devascularized, and ligated tightly 5 mm from its base with a 4.0 silk suture, without obstructing the bowel. The cecum was punctured once with a sterile 21-gauge needle on the antimesenteric border, and gentle pressure was applied to the cecum to extrude a small amount (1 mm) of feces. The bowel was returned to the peritoneal cavity, and the abdominal incision was closed in two layers with a 4.0 silk suture. Saline (3 cc/100 g) was then injected s.c. to resuscitate and to prevent dehydration. Imipenem/Cilastin (0.5 mg/Kg) was administered s.c., 12 h postoperatively. Following surgery, animals were injected intramuscularly with an analgesic (Buprenorphine, 0.10 mg/Kg) and every 6 h thereafter. Sham-operated controls underwent laparotomy and bowel manipulation without CLP; they too received saline and antibiotic.

To study CaMKIα in vivo, we subjected mice to in vivo CaMKIα RNAi using siSTABLE CaMKIα siRNA (Dharmacon, Lafayette, CO, USA), for which measurable concentrations persist in human serum for up to 5 days. Animals received CaMKIα or scrambled NT^RNAi (50 ug), which was dissolved in (animal mass/10 mL) Ringer's lactate and administered via hydrodynamic tail-vein injection, 72 h prior to experimentation. Animals were killed at 4 h and 16 h after endotoxemia or surgery with isoflurane (2% induction), followed by pentobarbitol (75 mg/kg, i.p.). Blood, liver, lung, kidney, and intestine were harvested. A survival study was performed in which separate cohorts of mice (n=23) underwent CLP, as described, and then were observed for 96 h.

Cell isolation and treatment

Murine Mφ cell line RAW 264.7 (American Type Culture Collection, Manassas, VA, USA) was grown in DMEM (BioWhittaker, Walkersville, MD, USA), supplemented with 10% FCS (Sigma-Aldrich, St. Louis, MO, USA), 50 U/mL penicillin, and 50 μg/mL streptomycin (Cellgro Mediatech Inc., Kansas City, MO, USA). Peritoneal Mφ were isolated from C57BL/6J mice by lavaging the peritoneal cavity with 5–2 mL aliquots 4°C PBS. The lavage was centrifuged at 300 g for 5 min, and the cells were resuspended in RPMI supplemented with 10% FCS, 50 U/mL penicillin, and 50 μg/mL streptomycin. Selected cells were pretreated with varying concentrations of KN62 (20 μM) or STO609 (1–20 μM) for 30 min or transfected with CaMKI^RNAi. DMSO, at concentrations similar to comparative experiments, was used as an appropriate control and never exceeded 0.2%. Selected cells were then treated with LPS (100 ng/mL).

siRNA

RAW 264.7 Mφ (2×10⁴) or murine peritoneal Mφ (1×10⁵) were plated in 0.5 ml growth medium (without antibiotics) in each well of a 24-well plate, resulting in 30% or 80% confluence, respectively. Fluorescein-labeled cyclophylin control siRNA, NT^RNAi, CaMKKα, CaMKKβ, CaMKIα, CaMKIβ, and CaMKIδ siRNA, obtained from Dharmacon, was added to 50 μl serum-free DMEM in a final concentration of 25 nM. We used the Smartpool siRNA from Dharmacon, which incorporates four separate siRNA sequences for each protein (Table 1).

Table 1. Sequences of siRNA.

Target	Sequence
CaMKIα:	sense GAACGAGAUUGCCGUCUUAUU, antisense UAAGACGGCAAUCUCGUUCUU
	sense CGGAAGACAUUAGGGAUAUUU, antisense AUAUCCCUAAUGUCUUCCGUU
	sense GGAGAGCUGUUUGACCAGGUU, antisense UUCGGUCAAACAGCUCUCCUU
	sense AUACAGCUCUGGAUAAGAAUU, antisense UUCUUAUCCAGAGCUGUAUUU
CaMKIβ:	sense AAUACAAGCUGGCAACAUGUU, antisense CAUGUUGCCAGCUUGUAUUUU
	sense GGACGCCAGCCACCUUGUAUU, antisense UACAAGGUGGCUGGCUCCUUU
	sense GAGCAGAAACCCUACGGGAUU, antisense UCCCGUAGGGUUUCUGCUCUU
	sense GACAAAGCCCAGAGGGUGAUU, antisense UCACCCUCUGGGCUUUGUCUU
CaMKIδ:	sense GGAGAAAGACCCAAAUAAAUU, antisense UUUAUUUGGGUCUUUCUCCUU
	sense GAAAAUGACUCGAAGCUGUUU, antisense ACAGCUUCGAGUCAUUUUCUU
	sense GGUGAUCGCCUAUAUCUUGUU, antisense CAAGAUAUAGGCGAUCACCUU
	sense GCUCGACAGGGAUGGAUUGUU, antisense CAAUCCAUGGGUGUCGAGCUU
CaMKKα:	sense GGGCUCAAGUUGGGCUUAUUU, antisense UAAGGCCCAACUUGAGCCCUU
	sense GCGUGUGUAUCAUGACAUUUU, antisense AAUGUCAUGAUACACACGCUU
	sense GGAAGUGCCCUGCGACAAGUU, antisense CUUGUCGCAGGGCACUUCCUU
	sense GGAAGUGCCCGUUCAUUGAUU, antisense UCAAUGAACGGGCACUUCCUU
CaMKKβ:	sense GGCUGAGAAUUCAGUCAAAUU, antisense UUUGACUGAAUUCUCGACCUU
	sense GGACUCUCAUCCUUAAGUAUU, antisense UACUUAAGGAUGAGAGUCCUU
	sense GAACGAAUCAUGUGUUUGCUU, antisense GCAAACACAUGAUUCGUUCUU
	sense UCACACCACGUCUCCAUUAUU, antisense UAAUGGAGACGUGGUGUGAUU

In a separate tube, 3 μl Hiperfect was diluted in 50 μl serum-free DMEM and incubated at room temperature for 5 min. These two solutions were combined, and the final transfection mixture was incubated for 20 min at room temperature. This transfection mixture was applied to each well and incubated for 6 h, after which, it was replaced by 500 μl cell medium and incubated for 72 h. Transfection efficiency was determined at 24 h by fluorescence microscopy. For each experiment, at least three microscopic visual fields (200× magnification) were counted to facilitate calculation of the ratio of fluorescent cyclophilin-expressing cells to nonfluorescent cells. Inhibition of each targeted protein was determined by immunoblot or RT-PCR. All experiments and cell number determinations were performed in triplicate.

Plasmid construction and transfection

Plasmids encoding a constitutively active CaMKI (CaMKI293) or a kinase-inactive CaMKI293^K49A mutant were the generous gifts of Dr. Marinna Picciotto (Yale University, New Haven, CT, USA). CaMKI293 contains a C-terminally truncated version of the CaMKI-encoding gene truncated to 293. CaMKI293^K49A was constructed by changing lysine 49 to alanine in CaMKI293, which negatively affects ATP binding at the catalytic site [25, 26]. For transient transfection, RAW 264.7 cells were seeded in a 24-well plate at 3 × 10⁵ cells/well. After 2 h of adhesion, Mφ were transfected with 1 ug plasmid CaMKI293 or CaMKI293^K49A using the Lipofectamine 2000 reagent, according to the instructions specified by the manufacturer (Life Technologies, Carlsbad, CA, USA). Following transfection, cells were handled as detailed in the figure legends.

Cellular protein extraction

Total cellular protein was extracted at 4°C in 500 μL lysis buffer [20 mM Tris, 137 mM NaCl, 2 mM EDTA, 10% glycerol, 1% Triton X-100, 1 μM sodium orthovanadate, 100 μM DTT, 200 μM PMSF, 10 μg/mL leupeptin, 0.15 U/mL aprotinin, 50 mM sodium fluoride, 10 mM sodium pyrophosphate, 2.5 μg/mL pepstatin A, 1 mM benzamidine, and 40 mM (α)-glycerophosphate]. Protein concentration was determined using a BCA protein assay (Pierce, Rockford, IL, USA).

Nuclear and cytoplasmic protein isolation

Cells were harvested and washed with PBS, followed by centrifugation at 300 g for 10 min. The cell pellet was lysed with NE-PER nuclear and cytoplasmic extraction reagent, according to the instructions specified by the manufacturer (Thermo Scientific, Rockford, IL, USA).

Secretory lysosome isolation

Lysosome isolation was performed using the lysosme enrichment kit (Thermo Scientific), according to the manufacturer′s instructions. Briefly, the pellets of cells were collected by centrifuging at 850 g for 2 min. Enrichment buffer A (800 mL) was added to the pellets, then vortexed to suspend the cells, and incubated on ice for 2 min. The cell suspension was sonicated shortly prior to the addition of 800 mL lysosme enrichment buffer B. The cell suspension was centrifuged (500 g) for 10 min, yielding a supernatant prepared for discontinuous gradient centrifugation. The cell extract was mixed with Optiprep^TM cell separation media to yield a final concentration of 15% Optiprep^TM media and was overlaid on the density gradients. After ultracentrifugation at 145,000 g for 2 h at 4°C, the lysosome layer was located as the top layer of the gradient. To remove the Optiperp^TM media, the isolated lysosome fraction was mixed with PBS and gradient dilution buffer separately, followed by centrifuging at 18,000 g for 30 min at 4°C.The lysosome pellet was collected and subjected to immunoblot analysis.

Western blots

Total cellular protein was electrophoresed in a 10% SDS-PAGE gel and was transferred to a Hybond-ECL nitrocellulose membrane (Amersham Pharmacia Biotech, Piscataway, NJ, USA). The membrane was blocked for 1 h at room temperature with 1% BSA, 5% BSA, or 5% milk and was then incubated with primary antibody for 12 h at 4°C. Blots were then incubated in a HRP-conjugated secondary antibody against the primary antibody at room temperature for 1 h. The blot was developed using the SuperSignal chemiluminescent substrate (Pierce) and was exposed on KAR-5 film (Eastman Kodak, Rochester, NY, USA). Densitometry was performed using the NIH Image program (National Institutes of Health, Bethesda, MD, USA) to quantitate OD. All gels were reblotted for total ERK1/2, p38, JNK/SAPK, or isoform-specific CaMK to confirm equal loading or to assess knockdown after RNAi.

Cytokine production

After the treatments described previously, supernatants were harvested under all conditions after 4 h of stimulation. IL-10, TNF-α, and IL-6 concentrations were quantitated by an enzyme immunoassay kit (Assay Designs), which is based on a coated-well, sandwich enzyme immunoassay.

β-Hexosaminidase assay

Supernatant (10 μL) was added to 100 μL reaction mixture [5 ml 0.4 M sodium acetate, pH 4.4; 5 ml 8 mM 4-methylumbelliferyl-N-acetyl-B-D-glucopyranoside (Sigma-Aldrich) in water; 0.25 mL Triton X-100; in 9.75 ml water] in a 96-well plate incubated at 37°C for 40 min. The reaction was stopped by addition of 75 μL 2 M Na₂CO₃, and fluorescence was measured in a spectrofluorimeter at excitation 370 nm, emission 450 nm [27].

RNA preparation and RT-PCR

Total RNA was extracted from RAW 264.7 cells, using TRIzol (Invitrogen, Carlsbad, CA, USA), according to the instructions specified by the manufacturer. The mRNA level was determined by quantitative RT-PCR. Briefly, 1 ug total RNA was used for the first-strand cDNA synthesis, which was carried out by using a RT system kit, according to the instructions of the manufacturer (Promega). cDNA (1 μL) was used as a template in PCR reactions, and the PCR reaction mixture at a 25 μL volume contained 12.5 μL 2× ReactionReady HotStart “Sweet” PCR master mix (SuperArray Bioscience, Frederick, MD, USA), 1 μL cDNA, and 1 μL sense and antisense primer sets (SuperArray Bioscience). After denaturing at 95°C for 15 min, the amplification protocol consisted of 30 cycles of denaturing at 95°C for 30 s, annealing at 55°C for 30 s, and extension at 72°C for 30 s. Expression levels of mRNA were normalized by the housekeeping gene β-actin.

Cell viability and morphologic features

Representative cell populations from each condition were examined under light microscopy. Cell viability was also confirmed by MTT assay. Cells were incubated in 96-well plates (Costar, Corning, NY, USA). After a 24-h incubation in 100 μl RPMI medium containing the stimulus, 50 μL 5 mg/ml MTT (Sigma-Aldrich) solution in PBS was added, and cells were incubated at 37°C for 2 h. The cells were then lysed by the addition of 100 μL/well extraction buffer [20% (w/v) SDS, 50% (v/v) N,N-dimethyl formamide, 2% (v/v) acetic acid, pH 4.7]. After overnight incubation with extraction buffer, the OD at 562 nm was measured.

Statistical analysis

Statistical analyses were performed using Stata 11 software (StataCorp LP, College Station, TX, USA). Values are expressed as means ± sem. Groups are compared by ANOVA. A P < 0.05 was considered statistically significant.

RESULTS

LPS activates CaMKIα in Mφ

CaMKI is expressed in all mammalian cells. Four isoforms of CaMKI have been identified (α, β, δ, and γ), and we focused on the dominant isoform, CaMKIα. We initially examined the activation and Thr^177/180 pCaMKI using an antibody that has been shown to recognize the active Thr-pCaMKIα (Thr¹⁷⁷), -pCaMKIβ (Thr¹⁷⁵), and -pCaMKIδ (Thr¹⁸⁰) [24]. As shown in Fig. 1A, LPS induced pCaMKI, which was first noted within 5 min and remained elevated above baseline at 2 h after LPS exposure. We observed three bands migrating between 37 KD and 43 KD, which correlated with three of the four isoforms of CaMKI: CaMKIα (42KD), CaMKIδ (43KD), and CaMKIβ (37KD). We did not observe any pCaMKI band in the predicted MW region of CaMKIγ (53 KD). Inhibiting the upstream CaMKK with STO609 reduced pCaMKI at all time-points (Fig. 1A). CaMKIα was strongly expressed, and the absolute expression did not change during LPS exposure (Fig. 1A). We observed a similar pattern of CaMKI activation in primary peritoneal Mφ (Fig. 1B). Unlike the nuclear localization of active pCaMKIV, active pCaMKI remained within the cytoplasm after LPS stimulation (Fig. 1C) [12].

Figure 1. LPS activates CaMKI.

(A) RAW 264.7 cells were exposed to DMSO or STO609 (5 μM) and stimulated with LPS (100 ng/mL). Total cell protein was isolated, and pCaMKI Thr^177/180 was determined by immunoblot (IB) using pThr^177/180-CaMKI mAb (representative blot of four independent experiments). Densitometry was performed for each experiment, and each bar represents the mean ± sem response at each time-point. (B) Primary peritoneal Mφ were exposed to LPS (100 ng/mL); total cell protein was isolated and subjected to immunoblot using pThr^177/180-CaMKI mAb (representative blot of three independent experiments). (C) RAW 264.7 cells were exposed to LPS (100 ng/mL); nuclear and cytoplasmic protein was isolated and analyzed by immunblot using pThr^177/180-CaMKI mAb. Purity of nuclear and cytoplasmic fractions was assessed by probing for histone 3 (representative blot of three independent experiments).

We focused on the predominant isoform CaMKIα, as it is widely distributed and expressed. Thus, we subjected Mφ to RNAi. CaMKIα siRNA inhibited CaMKIα expression specifically and effectively by 63% (P<0.05) in comparison with a NT scrambled siRNA (Fig. 2A and B). We then assessed the degree of pCaMKI in Mφ exposed to LPS after CaMKIα RNAi. As shown in Fig. 2C, CaMKIα RNAi eliminated the largest, slowest migrating pThr^177/180-CaMKI band and also inhibited, although to a lesser degree, the two more rapidly migrating bands (Fig. 2C). These data suggest that CaMKIα is activated by LPS. These data, combined with the specificity of CaMKIα siRNA (Fig. 2A), suggest that CaMKIα may regulate the other pCaMKI isoforms (i.e., CaMKIδ).

Figure 2. LPS activates CaMKIα in Mφ.

(A) RAW 264.7 cells were subjected to isoform-specific CaMKI (α, β, and δ) RNAi; total cell protein was isolated and analyzed for the specificity of CaMKI knockdown by immunoblot using isoform-specific CaMKI (α, β, and δ) antibodies (representative blot of three independent experiments). (B) Densitometry was performed for each CaMKIα experiment, and each bar represents the mean ± sem response for each condition. (C) RAW 264.7 cells were subjected to CaMKIα RNAi and then exposed to LPS (100 ng/mL); total cell protein was isolated and analyzed by immunoblot using pThr^177/180-CaMKI mAb (representative blot of three independent experiments).

CaMKIα mediates IL-10 release in Mφ stimulated with LPS

In the context of these data, we hypothesized that a CaMKIα-dependent mechanism regulated Mφ cytokine production. As shown in Fig. 3A, RAW 264.7 cells transfected with NT^RNAi released IL-10 when stimulated with LPS, which was reduced by 81% (1205 vs. 206 pg/mL; P<0.001) in CaMKIα^RNAi Mφ. By real-time PCR, CaMKIα^RNAi Mφ also expressed 63% less IL-10 mRNA than NT^RNAi (P=0.02) after exposure to LPS, suggesting that CaMKIα may regulate IL-10 at the level of IL-10 mRNA (Fig. 3B). Similarly, the CaMKK-specific inhibitor STO609 inhibited LPS-induced IL-10 production in a dose-dependent manner (Fig. 3C). At concentrations of 5 μM and 20 μM, IL-10 production was inhibited by 53% (1223 vs. 573 pg/mL; P=0.13) and 85% (1223 vs. 180 pg/mL; P<0.01). These data suggest that a CaMKK–CaMKI pathway regulates IL-10 production.

Figure 3. CaMKIα mediates IL-10 and IL-6 release in Mφ stimulated with LPS.

RAW 264.7 cells were subjected to RNAi using NT or CaMKIα siRNA. After 72 h, cells were exposed to LPS (100 ng/mL); supernatant was collected and assayed by ELISA for the following cytokines: (A) IL-10; (D) IL-6; (E) TNF-α (n=6 independent experiments). (B) RAW 264.7 cells were subjected to RNAi with NT or CaMKIα siRNA and exposed to LPS (100 ng/mL); mRNA was isolated and analyzed for IL-10 mRNA by RT-PCR (representative blot of four independent experiments). (C) RAW 264.7 Mφ were exposed to DMSO (media), the CaMKK-specific inhibitor STO609, or the broad CaMK inhibitor KN62 and then stimulated with LPS (100 ng/mL). Supernatant was harvested, and IL-10 concentration was determined by ELISA (n=4 independent experiments). (F) RAW 264.7 cells were subjected to RNAi using NT or CaMKIα siRNA. After 72 h, cells were exposed to LPS (100 ng/mL), and cell viability was assessed by MTT assay (n=4 independent experiments).

We observed a modest, yet nonsignificant reduction in IL-6 production (1980 vs. 1187 pg/mL; P=0.07; Fig. 3D), which was similarly associated with a nonsignificant 25% (P=0.28) decrease in IL-6 mRNA (data not shown). There was no significant reduction in TNF-α production (7186 vs. 7478 pg/mL) in CaMKI^RNAi in comparison with NT^RNAi cells (Fig. 3E). Cell viability was similar for all conditions (Fig. 3F).

CaMKIα regulates LPS-induced p38 activation in mediating IL-10 production

The MAPK p38 has been shown to be important in IL-10 production in other cell models [28, 29]. Hence, we investigated whether CaMKIα regulated p38 activation in Mφ exposed to LPS. As shown in Fig. 4A, LPS induced pp38, pERK1/2, and pJNK. RNAi of CaMKIα inhibited pp38 at all time-points; there was no significant inhibition of pJNK or pERK1/2 (Fig. 4A).

Figure 4. CaMKIα regulates LPS-induced p38 activation in mediating IL-10 production.

(A) RAW 264.7 Mφ were subjected to NT and CaMKIα RNAi and exposed to LPS (100 ng/mL); total cell protein was isolated, and p38, ERK1/2, and JNK activation was analyzed by immunoblot (representative blot of four independent experiments). Densitometry for p38 activation was performed for each experiment, and each bar represents the mean ± sem pp38 response at each time-point. (B) RAW 264.7 cells were transfected with a constitutively active CaMKIα (CaMKI293) or a kinase-deficient mutant (CaMKI293^K49A). Total cell protein was isolated and analyzed for pp38 by immunblot (representative blot of four independent experiments). Densitometry for p38 activation was performed for each experiment, and each bar represents the mean ± sem pp38 response at each time-point. *P < 0.05. (C) RAW 264.7 cells were incubated with PD98059 (PD; 10 uM) or SB203580 (SB; 5 uM) and then transfected with CaMKI293 or CaMKI293^K49A. Supernatant was collected and assayed by ELISA for IL-10, IL-6, and TNF-α concentration (n=5 independent experiments).

These data support a mechanism in which pCaMKIα induces pp38, thereby inducing elevations in IL-10 mRNA and IL-10 release. To investigate whether activated CaMKIα is sufficient for p38 activation and IL-10 release, we used a constitutively active species of CaMKIα (CaMKI293) and a kinase-deficient CaMKIα (CaMKI293^K49A) [30]. As shown in Fig. 4B, transfection of CaMKI293, in contrast to the kinase-deficient CaMKI293^K49A, induced pp38, which was noted at 2 h after transfection and remained elevated 8 h later. We did not observe any activation of ERK1/2 or JNK (data not shown).

In addition to inducing pp38, CaMKI293 induced a greater degree of IL-10 release than that observed with CaMKI293^K49A (127 vs. 42 pg/mL; P=0.08; Fig. 4C). This IL-10 production was mediated by p38, as pretreatment with the p38 inhibitor SB203580 reduced IL-10 production significantly (127 vs. 41 pg/mL; P=0.04; Fig. 4C). We did not observe any inhibition with PD98059, an inhibitor of the MEK/ERK cascade. As published previously, p38 inhibition reduced LPS-induced IL-10 production (Fig. 4C).

We also observed higher IL-6 production with CaMKI293 than CaMKI293^K49A (180 vs. 55 pg/mL; P=0.007), which was not inhibited with SB203580 or PD98059 (Fig. 4C). There was no significant increase in TNF-α in CaMKI293- in comparison with CaMKI293^K49A-treated cells. These data support our hypothesis that CaMKIα mediates LPS-induced IL-10 production by regulating p38 function. Furthermore, data suggest that CaMKIα may also regulate IL-6 production, through mechanisms independent of p38 and ERK1/2 MAPK.

CaMKIα mediates LPS-induced release of HMGB1 and secretory lysosomal contents

We also explored the role of CaMKIα in HMGB1 (29 KD) release. LPS induced the release of HMGB1 by NT^RNAi Mφ, which was attenuated by ∼73% (P<0.05) in CaMKIα^RNAi Mφ (Fig. 5A). We have shown that CaMKIV regulates the nucleocytoplasmic shuttling of HMGB1; however, CaMKIα is predominantly cytoplasmic (Fig. 1C). In accordance with this spatial restriction, we did not observe any inhibition of nucleocytoplasmic shuttling of HMGB1 with CaMKIα RNAi (data not shown).

Figure 5. CaMKIα mediates LPS-induced release of HMGB1 and secretory lysosomal contents.

(A) RAW 264.7 cells were subjected to RNAi using NT or CaMKIα siRNA. After 72 h, cells were exposed to LPS (100 ng/mL); supernatant was collected and assayed by immunoblot for HMGB1 (29 KD; representative blot of four independent experiments). Densitometry for HMGB1 concentration was performed for each experiment, and each bar represents the mean ± sem HMGB1 concentration for each condition. (B) RAW 264.7 cells were subjected to RNAi with NT, CaMKIα, or CaMKK siRNA and exposed to LPS (100 ng/mL). Supernatant was harvested and assayed for β-hexosaminidase activity (n=8 independent experiments). *P < 0.001 compared with LPS-stimulated cells. (C) RAW 264.7 cells were incubated with DMSO, the broad CaMK inhibitor KN93, or the CaMKK-specific inhibitor STO609 and then stimulated with LPS (100 ng/mL). Supernatant was harvested and assayed for β-hexosaminidase activity (n=8 independent experiments). #P = 0.008 compared with LPS-stimulated cells.

HMGB1 has been shown to be released by a nonclassical secretory lysosomal pathway, and we studied whether CaMKIα regulates secretory lysosomes [31]. As shown in Fig. 5B and C, LPS induced the release of lysosomal contents in NT^RNAi and control Mφ, with a respective 2.3- and 2.8-fold increase in supernatant β-hexosaminidase activity 24 h after LPS exposure. In contrast to NT^RNAi Mφ, CaMKI^RNAi Mφ released 65% less secretory lysosomal contents (0.80 vs. 2.3; P=0.008; Fig. 5B). Similar inhibition was observed with CaMKKα/β RNAi (0.98 vs. 2.3; P=0.008; Fig. 5B). Biochemical CaMKK inhibition with STO609 (0.98 vs. 2.8; P=0.008) and broad CaMK inhibition with KN93 (1.0 vs. 2.8; P=0.008) also reduced the release of secretory lysosomal contents in comparison with similarly treated control cells (Fig. 5C).

The prior observations suggested that CaMKIα governs the packaging and/or release of secretory lysosomal contents as a mechanism regulating HMGB1 release. Thus, we explored whether CaMKIα regulated the translocalization of HMGB1 to secretory lysosomes. As shown in Fig. 6, the concentration of HMGB1 was higher in secretory lysosomes isolated from Mφ exposed to LPS, in contrast to control, unstimulated Mφ. In addition, CaMKIα colocalized to these LAMP1-positive secretory lysosomes. Similarly, HMGB1 localized to secretory lysosomes in NT^RNAi Mφ exposed to LPS. However, this was not observed in similarly treated CaMKIα^RNAi Mφ (Fig. 6). Collectively, these data suggest that CaMKIα colocalizes with HMGB1 and regulates HMGB1 localization to secretory lysosomes.

Figure 6. CaMKIα regulates the localization of HMGB1 to secretory lysosomes.

RAW 264.7 cells were subjected to RNAi using NT or CaMKIα siRNA. After 72 h, cells were exposed to LPS (100 ng/mL) for 16 h. Total cell protein or secretory lysosomal protein was isolated, and the expression of HMGB1, CaMKIα, and the secretory lysosome marker protein LAMP1 was assessed by immunoblot (repesentative blot of three independent experiments).

Inhibition of CaMKIα reduces systemic IL-10 and HMGB1 concentrations during endotoxemia

We explored the role of CaMKIα in regulating systemic inflammation following LPS administration in vivo. The administration of CaMKIα siRNA effectively reduced CaMKIα expression by an average of 56% in the kidney and 90% in the lung; there was more variable (∼38%) inhibition in the liver (Fig. 7A). Mice subjected to NT^RNAi displayed significant elevations of systemic HMGB1, IL-10, IL-6, and TNF-α concentrations after LPS (Fig. 7B–E). Similar to what was observed in vitro, CaMKIα^RNAi inhibited the elevation in systemic HMGB1 concentration by 74% (P=0.04; Fig. 7B) and IL-10 concentration by 49% (1322.6 vs. 671 pg/mL; P=0.01; Fig. 7C). There was no significant inhibition of IL-6 (15,625 vs. 15,125 pg/mL; P=1.0; Fig. 7D) or TNF-α (285 vs. 285; P=0.83; Fig. 7E). Collectively, our findings show that CaMKIα mediates IL-10 and HMGB1 production in response to LPS in vitro and in vivo.

Figure 7. CaMKIα regulates the systemic release of IL-10 and HMGB1 during endotoxemia.

(A) Mice were subjected to in vivo RNAi. After 72 h, mice were killed, kidney, lung, and liver were harvested, and tissue lysate was probed for CaMKIα (42 KD) or actin (43 KD) expression by immunoblot (representative blot of six independent experiments). (B–E) Mice were subjected to in vivo RNAi using NT or CaMKIα siRNA. After 72 h, mice were subjected to i.p. injection of LPS (1.5 mg/kg). Mice were killed at 4 h and 16 h after LPS, and blood was isolated by cardiac puncture. (B) HMGB1 concentration was determined by immunoblot (representative blot of three independent experiments). (C–E) Serum IL-10, IL-6, and TNF-α concentrations were determined by ELISA (n=6 independent experiments).

CaMKIα^RNAi attenuates systemic cytokine concentrations and renal dysfunction during CLP sepsis

To determine whether CaMKIα alters systemic cytokine response in a clinically relevant model, we studied the effect of in vivo CaMKIα^RNAi in a CLP model of sepsis. NT^RNAi mice displayed significant elevations of systemic IL-10, IL-6, TNF-α, and HMGB1 concentrations, which were attenuated in CaMKIα^RNAi mice: 79% reduction in IL-10 concentration (2041 vs. 369 pg/mL; P=0.03; Fig. 8A), 88% reduction in IL-6 concentration (5101 vs. 593; P=0.02; Fig. 8B), 79% reduction in TNF-α production (168 vs. 35 pg/mL; P=0.02; Fig. 8C), and 74% reduction in HMGB1 concentration (P=0.04; Fig. 8D). Thus, in a polymicrobial sepsis model, inhibition of CaMKIα had a broader effect with reductions in systemic TNF-α and IL-6 concentrations, which may be related to CaMKIα regulation of non-TLR4 pathways.

Figure 8. CaMKIα regulates the systemic release of IL-10, IL-6, TNF-α, and HMGB1 during CLP sepsis.

(A–C) Mice were subjected to in vivo RNAi, as denoted on the abscissa of each graph. After 72 h, mice were subjected to CLP. Mice were killed at 16 h after CLP, and blood was isolated by cardiac puncture. Serum IL-10, IL-6, and TNF-α concentrations were determined by ELISA (n=6 independent experiments). (D) HMGB1 concentration was determined by immunoblot (representative blot of six independent experiments). Densitometry was performed for each HMGB1 experiment, and each bar represents the mean ± sem response for each condition.

Inflammation is postulated to underlie the development of multiple organ dysfunction, and renal dysfunction is typically an early event. We studied whether CaMKIα^RNAi might influence the development of renal dysfunction by measuring the BUN and creatinine concentrations in mice subjected to CLP. As shown in Fig. 9, CLP induced marked elevations of BUN and mild elevations of creatinine in NT^RNAi-treated mice. The disproportionate elevation in BUN relative to creatinine suggests a prerenal state of diminished circulating volume. In contrast, there was no significant elevation in BUN or creatinine in CaMKIα^RNAi mice subjected to CLP (Fig. 9).

Figure 9. CaMKIα RNAi attenuates renal dysfunction during CLP sepsis.

Mice were subjected to in vivo RNAi as denoted on the abscissa of each graph. After 72 h, mice were subjected to CLP. Mice were killed at 16 h after CLP, and blood was isolated by cardiac puncture. Serum BUN and creatinine were analyzed by ELISA (representative blot of six independent experiments).

In light of the reduction in inflammation and renal dysfunction with CaMKIα inhibition, we hypothesized that CaMKIα inhibition may also impart an early survival advantage. CaMKIα^RNAi mice exhibited improved survival in contrast to NT^RNAi mice (41.9% vs. 26.9%; hazard ratio for death 0.67, 95% confidence interval 0.36–1.29; P=0.24); however, this did not attain statistical significance (data not shown). Thus, CaMKIα inhibition resulted in profound reductions in systemic IL-10, IL-6, TNF-α, and HMGB1 release and attenuation in renal dysfunction following CLP sepsis but contributed to a 15% reduction in absolute mortality.

DISCUSSION

Prior investigations have established that monocyte/Mφ function is dependent on Ca²⁺/CaM signaling, although the specific Ca²⁺/CaM-dependent mechanisms orchestrating the monocyte/Mφ response to LPS have been difficult to define [4, 5, 12]. In this study, we characterize a novel, biological role of CaMKIα in regulating the inflammatory response to septic insult. CaMKIα activity mediates Mφ IL-10 production, using mechanisms involving the regulation of p38 MAPK signal transduction. CaMKIα also regulates the cellular release of HMGB1, colocalizing with and regulating the translocation of HMGB1 to secretory lysosomes [31, 32]. Of clinical significance is that the in vivo response to septic inflammation is dependent on CaMKIα activity, and modulating CaMKIα with in vivo RNAi reduces systemic cytokine concentrations and the development of renal dysfunction.

The combined observations of our in vitro and in vivo studies highlight that the CaMKK/CaMKIα cascade is essential and sufficient for Mφ IL-10 production. Prior studies support the importance of p38α signaling for IL-10 production in monocyte/Mφ responding to LPS [33–39]. Our data now link the CaMK and MAPK cascades and characterize a CaMKIα-dependent pathway induced by LPS, which regulates p38 signal transduction necessary for elevated IL-10 production. The α/β isoform specificity of the p38 inhibitors and the lack of expression of p38β in Mφ support a CaMKIα-p38α signaling pathway [19, 21]. These events occurred independent of ERK1/2 signaling, which is in agreement with other studies [38, 39].

Although CaMKIα partly regulated IL-6 release, it did so independent of p38. Our observations agree with other in vitro monocyte/Mφ studies, which observed that LPS-induced IL-6 production was less sensitive to p38 inhibition with SB203580 [40, 41]. Similarly, ERK1/2 inhibition did not alter CaMKIα-dependent IL-6 production. The expression of IL-6 in monocytes and Mφ may be regulated by JNK/SAPK. JNK1^−/− and JNK2^−/− mice have shown low levels of IL-6 in response to LPS stimulation, and the JNK pathways were reported to play a role in LPS-induced IL-6 production in Mφ [42, 43].

Inhibition of CaMKIα did not alter LPS-induced TNF-α release, despite inhibiting pp38. Similarly, transfection of active CaMKIα failed to induce TNF-α, despite inducing pp38. We used a low dose of SB203580, which is highly selective for p38α, p38β1, and p38β2, carries little significant effect on the activities of ERK1/2 and JNK, and does not inhibit p38δ or p38γ. Our data provide evidence against a regulatory role of p38α in LPS-induced TNF-α production. Similar observations have been reported by other investigators [44, 45] using similar doses of LPS and SB203580; in these studies, TNF-α was not inhibited significantly. Other investigators have observed a heightened degree or inhibition of TNF-α in RAW Mφ, using similar LPS and SB203580 doses [46, 47]. Studies demonstrating a regulatory role of p38 in Mφ TNF-α production tend to study monocytes and use high concentrations of SB203580 (30 μM), raising concern about the selectivity of p38 inhibition [48]. Our data do not exclude the possibility of other p38 isoforms being involved, however. Alternatively, CaMKIα might induce other mechanisms that serve to counter and inhibit p38-dependent TNF-α production.

Prior studies have revealed the capacity for members of the CaMK cascade (i.e., CaMKI and CaMKIV) to regulate the activation of JNK, p38, and ERK1/2 MAPK, transduction cascades integral to immune cell function. Insight into the mechanisms and biological relevance of these interactions in the context of sepsis is just beginning to be acquired [49]. Previously, we reported that broad CaMK inhibition attenuated ERK1/2 activation in adhering monocytes and reduced TNF-α production consequent to LPS challenge [16]. Here, we detail that a CaMKIα-p38 pathway, induced by LPS, is required for full Mφ IL-10 production. This regulatory mechanism may involve regulation of the canonical MEK3/6-p38α cascade or a recently described, alternate pathway, in which CaMKIα facilitates auto-pp38 in a complex orchestrated by TGF-β-activated kinase-binding protein 1 [50].

In addition to cytokine regulation, CaMKIα appears important for HMGB1 release, an architectural binding protein that has been shown to function as a late mediator of mortality in murine endotoxemia and sepsis [7–9]. In our prior study, CaMKIα did not regulate HMGB1 in hepatocytes exposed to H/R, an event that is TLR4-dependent [11]. Collectively, the current study and our prior data suggest that the mechanisms governing HMGB1 secretion are cell (Mφ vs. hepatocyte)- and stimulus (LPS vs. H/R)-specific.

LPS-stimulated Mφ release HMGB1 via an active process that necessitates shuttling the protein from nucleus to cytoplasm and then secretion from the cell [31]. Our collective data support a paradigm in which the coordinated functions of CaMKIV and CaMKIα mediate the cellular release of HMGB1 from Mφ via mechanisms reflective of their subcellular distribution. As we have shown previously, CaMKIV regulates the nucleocytoplasmic shuttling of HMGB1 by translocating to the nucleus upon LPS exposure and mediating the prerequisite step of serine phosphorylation [12]. Unlike CaMKIV, CaMKIα is localized predominantly to the cytoplasm and remains cytoplasmic after activation. Our data suggest that CaMKIα governs this process by colocalizing to secretory lysosomes and regulating the translocation of HMGB1 to secretory lysosomes [31]. Currently, we can only speculate as to how this may occur. Neuronal studies have provided data linking CaMKK and CaMKI with rearrangements in the actin cytoskeleton during dendritic spine morphogenesis and synaptogenesis [51–53]. Thus, CaMKI-dependent regulation of actin may be one mechanism mediating the release of secretory lysosomes containing HMGB1. Alternatively, the upstream CaMKKα mediates cellular autophagy, a regulated process that sequesters cytoplasmic proteins and targets them to lysosomal digestion [54]. Future studies are needed to determine whether these or alternate mechanisms are used by CaMKIα in coordinating the cytoplasmic release of HMGB1 [55].

In addition to our in vitro findings, we evaluated the in vivo relevance of CaMKIα following i.p. administration of LPS. We show that CaMKIα siRNA attenuates systemic IL-10 and HMGB1 concentrations, results nearly mirroring those of our in vitro studies. Such data suggest that the CaMKIα signaling induced by LPS and regulating Mφ function is involved in the production of IL-10 and HMGB1 in vivo. When we evaluated the role of CaMKIα in a clinically relevant model of CLP sepsis, however, CaMKIα siRNA also correlated with reductions in systemic concentrations of TNF-α and IL-6 in addition to IL-10 and HMGB1. This apparent discrepancy may be attributed to the additional microorganisms, processes, cells, and organs involved in the in vivo response to polymicrobial infection, in contrast to LPS alone. These additional events, distinct to CLP, may be regulated by CaMKIα.

In addition to attenuating systemic inflammation, CaMKIα RNAi attenuated the magnitude of renal dysfunction. Recent studies highlight a causal role of exaggerated inflammation and HMGB1 in mediating this pathophysiology. Methods to reduce systemic HMGB1 concentrations, such as ethyl pyruvate, have been associated with reductions in the development of organ dysfunction in experimental models of pancreatitis [56]. Similar mechanisms may be operant in renal ischemia and reperfusion injury [57, 58]. HMGB1 has also been shown to mediate inflammation and remote organ injury following trauma and hepatocellular injury induced during ischemia/reperfusion [11]. We have shown that this latter event, including the release of HMGB1, appears to involve Ca²⁺ and CaMK signaling [11]. These studies lend credence to the idea that CaMKIα-dependent regulation of inflammation, specifically HMGB1, underlies the development of organ dysfunction. Whether this regulatory mechanism is systemic or localized to the organ requires further investigation. Similarly, additional studies are needed to determine the specificity of our observations to CaMKI of Mφ or parenchymal origin.

These reductions in inflammation and organ dysfunction correlated with improved, albeit not significant, survival during CLP sepsis in mice receiving CaMKIα RNAi. One possible explanation for why we did not observe a larger difference is survival in mice following CaMKIα siRNA is that IL-10 is considered essential to quiesce the inflammatory response; thus, concomitant inhibition of IL-10 may negate any benefit imparted by reduced TNF-α, IL-6, and HMGB1 release. Prior studies of TNF-α blockade during CLP sepsis observed worse survival in contrast to the benefits noted in models of endotoxemic shock, which could also explain our observations [59]. In addition, our study is underpowered to significantly detect the observed 33% reduction in mortality, and any variability in the efficacy of siRNA to inhibit CaMKIα expression in the critical cell/organ mediating this adverse outcome would introduce bias away from detecting a difference.

In summary, CaMKK/CaMKIα mediates LPS-induced IL-10 and HMGB1 production through respective mechanisms that involve regulating p38 MAPK activation and secretory lysosomes. CaMKIα-dependent processes appear to be operant in vivo by regulating systemic cytokine production and HMGB1 release and contributing to the renal dysfunction of sepsis. Subsequent studies are needed to explore and distinguish the contributions of other CaMKI isoforms to sepsis-induced inflammation. These observations support further investigation into the clinical use of CaMK modulation during inflammatory states such as sepsis.

AUTHORSHIP

X.Z. was involved in the conception and design of the study, the performance of experimentation and interpretation of the data, and final revision of the manuscript. L.G., R.D.C., J.L.S., and A.T. were involved in the performance of experimentation, interpretation of the data, and final revision of the manuscript. J.S.L. was involved in the conception and design of the study, the interpretation of the data, and manuscript preparation and final revision. M.R.R. was involved in the conception and design of the study, the performance of experimentation and interpretation of the data, and manuscript preparation and final revision.