MKK3 regulates mitochondrial biogenesis and mitophagy in sepsis-induced lung injury

Abstract

Sepsis is a systemic inflammatory response to infection and a major cause of death worldwide. Because specific therapies to treat sepsis are limited, and underlying pathogenesis is unclear, current medical care remains purely supportive. Therefore targeted therapies to treat sepsis need to be developed. Although an important mediator of sepsis is thought to be mitochondrial dysfunction, the underlying molecular mechanism is unclear. Modulation of mitochondrial processes may be an effective therapeutic strategy in sepsis. Here, we investigated the role of the kinase MKK3 in regulation of mitochondrial function in sepsis. Using clinically relevant animal models, we examined mitochondrial function in primary mouse lung endothelial cells exposed to LPS. MKK3 deficiency reduces lethality of sepsis in mice and by lowering levels of lung and mitochondrial injury as well as reactive oxygen species. Furthermore, MKK3 deficiency appeared to simultaneously increase mitochondrial biogenesis and mitophagy through the actions of Sirt1, Pink1, and Parkin. This led to a more robust mitochondrial network, which we propose provides protection against sepsis. We also detected higher MKK3 activation in isolated peripheral blood mononuclear cells from septic patients compared with nonseptic controls. Our findings demonstrate a critical role for mitochondria in the pathogenesis of sepsis that involves a previously unrecognized function of MKK3 in mitochondrial quality control. This mitochondrial pathway may help reveal new diagnostic markers and therapeutic targets against sepsis.

sepsis, a systemic inflammatory reaction to infection, is the leading cause of death globally. The incidence of sepsis worldwide is 18 million every year with 30% mortality. The economic impact of sepsis is substantial with costs of up to $50,000/patient and $17 billion annually in United States alone (30, 43). Death from sepsis occurs due to multiorgan failure, and biological therapies do not exist. Prevailing theories attribute multiple organ failure in sepsis to an uncontrolled inflammatory response, apoptosis, or disorders in the coagulation (38). Unfortunately, therapies against these responses, such as anti-inflammatory agents and activated protein C, have been unsuccessful. Since current medical care remains purely supportive, there is an urgent need to develop targeted therapies.

Mitochondria are mediators of inflammatory responses (19, 54) and become dysfunctional in sepsis and lung injury (4, 6), suggesting they are involved in the observed pathology. Mitochondria are essential hubs of innate immune signaling and inflammation in sepsis (48, 49) and are also major sites of reactive oxygen species (ROS) production in the cells. Mitochondria are constantly exposed to ROS, and hence ongoing biogenesis and turnover are needed to maintain a functional network. Dysfunctional mitochondria are removed through selective degradation via autophagy by the lysosomal machinery (32), a process known as mitophagy. We hypothesized herein that disrupted mitochondrial homeostasis, owing to altered biogenesis and/or mitophagy, contributes to sepsis pathology.

MAP kinase kinase 3 (MKK3) is an upstream kinase of p38 (15). MKK3 and p38 are ancient ancestral components of innate immune responses and may have predated the canonical Toll-like receptor signaling pathways (20). We have recently shown that MKK3 is a critical mediator of lethal murine endotoxemia and organ failure (29). Since LPS-exposed MKK3-deficient mice and cells were protected against death with less inflammation and mitochondrial ROS compared with wild-type (WT) mice and cells, we hypothesized that MKK3 deficiency protects against sepsis by restoring mitochondrial homeostasis.

MATERIALS AND METHODS

Mice.

We have previously described the MKK3^−/− mice (55). PINK1^−/− mice were generated as described (21), and wild-type mice were purchased from the National Cancer Institute. All strains were backcrossed more than 10 times into C57BL/6 background. All of the protocols were reviewed and approved by the Animal Care and Use Committee at Yale University School of Medicine.

Sepsis models.

Mice received intraperitoneal injections of 40 mg/kg and 5 mg/kg LPS (Escherichia coli 055:B5, Sigma Aldrich). A dose of 1 × 10⁷ CFU/mouse E. coli (ATCC25922) was injected intraperitoneally in 500 μl of sterile PBS. Cecal ligation and puncture was performed as described previously (39, 46). Briefly, a midline incision was made in the peritoneum, and the cecum was exteriorized. Eighty percent of the cecum was ligated and pierced through with a 21-G needle, and then a small drop of cecal contents was extruded. The cecum was returned to the peritoneal cavity, and the abdomen was closed in two layers. Blood pressure was measured by a noninvasive tail-cuff method (CODA System, Kent Scientific), and body surface temperature was measured with an Infrascan infrared thermometer (LaCrosse Technologies). The Animal Care and Use Committee at Yale University School of Medicine approved all of the protocols.

Human sepsis samples.

We collected blood from critically ill patients within 24 h of admission to the Medical Intensive Care Unit (ICU) at Yale-New Haven Hospital. Patients were enrolled and consented as part of a broad biorepository of newly admitted, critically ill patients. Septic patients were identified by the American College of Chest Physicians/Society of Critical Care Medicine (ACCP/SCCM) Consensus Criteria (3) as those with presence of infection and the presence of at least two of the four features of the systemic inflammatory response syndrome. Nonseptic ICU patients were those who did not meet ACCP/SCCM criteria for any of the sepsis syndromes. We excluded patients receiving dialysis and patients having received blood transfusions within 48 h of blood draw because in a pilot analysis we found that these conditions elevated the levels of MKK3 activity considerably. Peripheral blood mononuclear cells (PBMCs) were isolated from blood by histopaque gradient method (12). Cells were counted, and 3 million were lysed in 15 μl of 1× vendor supplied AlphaScreen lysis buffer and frozen until assayed. MKK3/6 activity was detected by using an AlphaScreen SureFire phospho-MKK3/6 kit (Perkin Elmer). Lysates were thawed and spun down, and 3 μl were mixed with 5 μl of acceptor mix and incubated for 2 h at room temperature before addition of 2 μl of donor mix followed by another 2 h incubation. The mixture was then read on an Envision reader (Perkin Elmer). Vendor-supplied positive and negative control HeLa cell lysates (1.2 μl) were included and used to normalize the results. Western blot protein was extracted with RIPA from PBMCs. The antibody used was anti p-MKK3 (Novus Biologicals, NB100-82048). All protocols were approved by the Institutional Review Board at the Yale University School of Medicine.

Isolation and FACS analysis of total lung cells and inflammatory cells.

Lungs of mice were dissected out and rinsed in ice-cold PBS. They were then minced and incubated with 4.9 ml HEPES buffer with 100 μl Collagenase D (Roche Applied Science) and 10 μl DNase I (Roche Applied Science). The lungs were kept at 37°C for 30 min until disaggregation was complete. The resulting cell suspensions were filtered through 70-μm cell strainers (BD Biosciences). Cells were counted and 1 × 10⁶ cells were stained with anti-Ly6G-PE (BD Biosciences) and GR1-APC (BD Biosciences) to identify neutrophils, whereas anti-F4/80-PE (eBioscience) and anti-CD45-APC (BD Biosciences) were used to identify macrophages. Stained cells were fixed with 4% paraformaldehyde and analyzed via a BD FACSCaliber machine. For analysis the cells were gated on live cells by forward and side scatter and analyzed with FlowJo 8.7 software.

Isolation of primary lung endothelial cells.

We have previously described primary lung endothelial cells isolation (56). Briefly, lungs were extracted, minced, and digested for 1 h at 37°C in 0.1% collagenase (Roche Diagnostics) in RPMI-1640 with 100 U/ml penicillin G and 100 μg/ml streptomycin. The digest was passed through a 100-μm cell strainer to remove undigested tissue fragments. Cells were centrifuged at 200 g for 5 min; pellets were resuspended in endothelial medium containing 20% FBS, 40% DMEM, and 40% F12 with 100 U/ml penicillin G and 100 μg/ml streptomycin and plated onto 0.1% gelatin-coated T75 flasks. Cells were washed after 24 h and cultured for 2–4 days. Cultured cells were trypsinized with 2 ml trypsin/EDTA, PBS was added, and the cells were spun for 5 min at 200 g to remove the supernatant. Cells were resuspended in 2% FBS containing 10 μl biotin-labeled rat anti-mouse CD31 (PECAM-1) antibody (BD Biosciences-Pharmingen). After incubation on ice for 30 min, the cells were washed with streptavidin magnetic beads (New England Biolabs). Cells were washed with 2% FBS, resuspended in 5 ml of 2% FBS, and incubated on ice for 30 min. The cells were then placed on the magnet for 5 min; unbound cells were removed, whereas bound cells were resuspended in medium and plated onto a 0.1% gelatin-coated T25 flask. Staining of CD31 and flow cytometry analysis confirmed that more than 95% of the cells were endothelial cells. Cells were maintained in 40% DMEM and 40% F12 tissue culture medium supplemented with 20% FBS.

Western blot analysis.

Protein was extracted from cells by using RIPA buffer with protease and phosphatase inhibitors (Roche), and 50 μg of protein was loaded per lane of a 5–20% gel, transferred to a PVDF membrane, and immunoblotted with primary antibodies. Detection was performed via a horseradish peroxidase detection system (Cell Signaling Technology). Equivalent sample loading was confirmed by stripping membranes with Blot Restore Membrane Rejuvenation solution (Thermo Scientific) and probing for β-actin. Antibodies used were Pink (Abgent), Parkin (Santa Cruz Biotechnology), MAP1 light chain 3 (LC3B; Cell Signaling), sirtuin-1 (Sirt1; Abcam), peroxisome proliferator-activated receptor γ coactivator 1 (PGC-1; Santa Cruz Biotechnology), nuclear respiratory factor 1 (Nrf1; Santa Cruz Biotechnology), acetylated lysine (Cell Signaling), and β-actin (Santa Cruz Biotechnology). For coimmunoprecipitation studies 50 μg of lung lysate was precleared with Protein A/G PLUS (Santa Cruz Biotechnology). Then the lysate was incubated with PGC-1 antibodies overnight and then precipitated with Protein A/G as per the manufacturer's protocol and run on Western blots as above. Relative densitometry was measured with ImageJ software (1).

Mitochondrial/oxidant assays.

Mitosox Red, MitoTracker Green, and MitoTracker Deep red (Invitrogen) were used to assess mitochondrial function in endothelial cells. Cells were seeded onto six-well plates 1 day before the experiment. The next day, 1 μg/ml of LPS was added to the cells for 6 h. Cells were washed and exposed to 2.5 μM Mitosox, 50 nM MitoTracker Green, or 50 nM MitoTracker Deep red in regular media at 37°C for 25 min. The cells were then washed with PBS, detached gently with 0.4 M EDTA, and analyzed by use of a BD FACSCaliber machine. For analysis the cells were gated on live cells by forward and side scatter. Mean fluorescent intensity was calculated by using FlowJo 8.7 software.

siRNA knockdown.

Endothelial cells were seeded onto six-well plates 1 day prior to transfection with 40% DMEM and 40% F12 tissue culture medium supplemented with 20% FBS, without antibiotics. At the time of transfection with the specific siRNA, the cells were 50–60% confluent. Lipofectamine 2000 Reagent (Invitrogen) was used as the transfection agent. The cells were collected and analyzed 48 h after transfection. siRNAs used were scrambled siRNA and MKK3 (ON-TARGETplus SMARTpool, Dharmacon), Pink1, Parkin, and Sirt1 (Santa Cruz Biotechnology).

Histology and microscopy.

For electron microscopy of lung tissue, mice were anesthetized and a cannula was placed in the pulmonary artery. Next, 50 units of heparin was injected into the lungs and the vasculature was perfused with HBSS for 20 min. During the perfusion, the lungs were ventilated through a tracheal cannula at a rate of 120 strokes/min and tidal volume of 0.8 ml to eliminate any bias in the imaging that could occur from lung collapse or blood clots. The lungs were then fixed in situ [20 min at room temperature (RT)] by injecting a mixture of 3% formaldehyde + 1.5% glutaraldehyde in 0.1 M sodium cacodylate buffer, pH 7.3, through the pulmonary artery cannula. Lungs were then dissected out and fixed in the same mixture for 1 h at RT, then in 2% osmium tetroxide (OsO₄) in 0.1 M sodium-cacodylate buffer, pH 6.8 for 1 h on ice. The samples were then placed in the dark for 1 h at RT with 1% uranyl acetate, dehydrated through submersion in graded ethanol (70–100%), and embedded in Epon 812 resin blocks. Tissue blocks were cured for 72 h at 90°C prior to being cut into 60-nm sections with a Leica microtome. Sections were counterstained with uranyl acetate and lead citrate and examined and photographed in a FEI Tenai Biotwin transmission electron microscope (14). All materials were obtained from Electron Microscopy Sciences, except uranyl acetate (Sigma Aldrich). For fluorescent microscopy analysis, endothelial cells were attached to sterile coverslips overnight. Next day they were exposed to LPS 1 μg/ml for 4 h. Cells were then fixed with 4% paraformaldehyde and permeabilized with 0.3% Triton X-100 for 10 min. They were then incubated overnight with primary antibodies LC3B (Cell Signaling) and Hsp60 (Santa Cruz Biotechnology) at 1:100 concentrations. Next day the cells were washed and incubated with appropriate fluorescent secondary antibodies (Invitrogen) at 1:200 concentration for 2 h. The coverslips were then mounted with Prolong Gold antifade reagent with DAPI (Invitrogen) and photographed in Olympus BX61 microscope fitted with Olympus Q Color 5 camera. Images were analyzed with ImageJ software (1) and the colocalization coefficient was calculated to measure the fraction of overlap of signals.

Real-time RT-PCR.

Total RNA was extracted from lungs or cells by using TRIzol reagent according to the manufacturer's protocol (GIBCO BRL, Carlsbad, CA). First-strand cDNA was synthesized by using Superscript II Reverse Transcriptase (Invitrogen) with random hexamers; conditions were 10 min at 25°C, 30 min at 48°C, and 5 min at 95°C. Real-time RT-PCR reactions were carried out in Power SsoFast EvaGreen Supermix (Bio-Rad) and an ABI Prism 7000 Sequence Detection System (Applied Biosystems). Actin was amplified as a control. Primers used are specified in Table 1. Real-time PCR conditions were 95°C for 10 min and 40 cycles of 95°C for 15 s, followed by 60°C for 1 min. The relative values of gene expression were calculated from the accurate threshold cycle (Ct), which is the PCR cycle at which an increase in reporter fluorescence from dye can first be detected above a baseline signal. The CT values for β-actin were subtracted from the CT values for candidate genes in each well to calculate ΔCT. The ΔCT values for each sample were averaged. To calculate the fold induction over controls (ΔΔCT), the average ΔCT values calculated for WT tissue/cells were subtracted from ΔCT values calculated for MKK3^−/− tissue/cells. Next, the fold induction for each well was calculated by the 2^−ΔΔCT formula. The fold induction values for replicate wells were averaged, and data were presented as means ± SE of triplicate wells. For mitochondrial quantification DNA was prepared from serum, lungs, and endothelial cells by Gentra Puregene kit (Qiagen). PCR protocols were similar to those used above.

Table 1.

Gene	Primer sequence 5′–3′
Pink1	Forward ACCAGCATGTTGGCCTGCGGCT
	Reverse ATGGGCTGTGGACACCTCAGGGGC
Parkin	Forward TGGGAGGTGTGCTGTGCCCCCG
	Reverse AAAACAAACCCGCAGCCCAGGCCGT
LC3B	Forward GCAGCTGCCCGTCCTGGACAA
	Reverse TGAGCTGCAAGCGCCGTCTGA
Sirt1	Forward CAGTGTCATGGTTCCTTTGC
	Reverse CACCGAGGAACTACCTGAT
PGC-1α	Forward GCCGTAGGCCCAGGTACGACAGC
	Reverse CCCGCTTCTCGTGCTCTTTGCGGT
Nrf1	Forward GCAGGCTATGGAAGTAAA
	Reverse TGTAGTTGGTGCTAAGAGGGT
β-Actin	Forward GTGGGCCGCTCTAGGCACCA
	Reverse TGGCCTTAGGGTTCAGGGGG
NADH dehydrogenase subunit 1	Forward ACCGGGCCCCCTTCGACCTGAC
	Reverse AACGCGAATGGGCCGGCTGCG
Cytochrome b	Forward GCCACCTTGACCCGATTCTTCGCT
	Reverse AGGTGAACGATTGCTAGGGCCGCG
Cytochrome c oxidase subunit I	Forward CACTACCAGTGCTAGCCGCAGGCA
	Reverse TTGGGTCCCCTCCTCCAGCGGGA
ATP synthase F0 subunit 6	Forward GCTCACTTGCCCACTTCCTTCCACA
	Reverse AGCCGGACTGCTAATGCCATTGGT
18S	Forward CGCGGTTCTATTTTGTTGGTTT
	Reverse GCGCCGGTCCAAGAATTT

Statistics.

Murine data were analyzed by Student's t-tests and are expressed as means ± SE. Human data were analyzed by Wilcoxon (Exact) rank-sum tests. Differences among survival curves were evaluated by log-rank tests. For all two-sided tests, P < 0.05 defines statistical significance. Statistical analyses were performed with GraphPad Prism 5.03 (GraphPad) software and SAS v9.3 (Cary, NC).

RESULTS

MKK3^−/− mice are resistant to sepsis.

To assess whether MKK3 is a determinant of survival, we induced sepsis in mice by three clinically relevant models of sepsis. We administered lethal doses of intraperitoneal lipopolysaccharide (LPS), or intraperitoneal E. coli, or we performed cecal ligation and puncture. We found that survival of MKK3^−/− mice was significantly higher than survival of WT mice (Fig. 1, A–C). Consistent with frequently used clinical parameters for sepsis, MKK3^−/− mice showed less physiological derangement, such as less hypotension and hypothermia, than WT mice (Fig. 1D).

Fig. 1.

MKK3^−/− mice are resistant to lethal LPS. A: a lethal dose of intraperitoneal LPS (40 mg/kg) was given to wild-type (WT) mice and MKK3^−/− mice and survival was assessed (n = 15, *P < 0.001). B: a lethal dose of intraperitoneal live Escherichia coli (1×10⁷ CFU/mouse) was given to WT mice and MKK3^−/− mice and survival was assessed (n = 10, *P < 0.05). C: WT and MKK3^−/− mice underwent either cecal ligation and puncture (n = 10 per group) or sham surgery (n = 4 per group) and survival was assessed (*P < 0.05). D: a sublethal dose of intraperitoneal LPS (5 mg/kg, 6 h) was administered to WT and MKK3^−/− mice, and temperature and blood pressure were measured (n = 9, horizontal lines mark means, *P < 0.001). All results are representative of at least 3 independent experiments.

MKK3 deficiency protects against LPS-induced lung injury.

We used electron microscopy (EM) to examine lungs of MKK3^−/− and WT mice exposed to LPS. Representative sections showed higher inflammatory cell infiltration in the lungs of WT mice than in those of MKK3^−/− mice (Fig. 2A). EM lung sections of WT mice treated with LPS also showed more evidence of diffuse injury, including endothelial injury, than those of MKK3^−/− mice. In addition, WT cells showed more damage and dying mitochondria than MKK3^−/− cells, suggested by swollen mitochondria with loss of normal architecture and distortion of cristae (Fig. 2B). There was also more evidence of endothelial damage, as depicted by swelling, vacuolization, and widening of gap junctions in WT lungs than in MKK3^−/− lungs (Fig. 2C).

Fig. 2.

MKK3^−/− mice were protected against lung injury after LPS. A: representative electron microscopy images of lungs of mice after intraperitoneal LPS (40 mg/kg, 12 h). Arrows point to inflammatory cells. Bar = 10 μM. B: representative electron microscopy images of lungs of mice after intraperitoneal LPS (40 mg/kg, 12 h). WT lungs show more mitochondrial damage characterized by pale swollen mitochondria with loss of normal architecture and distortion of cristae. Arrows indicate mitochondria. Bar = 1 μM. C: representative electron microscopy images of lungs of mice after intraperitoneal LPS (40 mg/kg, 12 h). Endothelial damage is shown by increased swelling, vacuolization and disruption of the inter-endothelial gap junctions in WT (arrows) compared with MKK3^−/− mice (arrowheads). Top, bar = 1 μM; bottom, bar = 200 nm. D: bar graph of absolute number of neutrophils and macrophages were calculated as a fraction of the total number of cells in lung digests. WT and MKK3^−/− mice were given intraperitoneal LPS (40 mg/kg, 6 h), the lungs were digested and analyzed by flow cytometry for neutrophils (Ly6G+, GR-1+) and macrophages (F4/80+, CD45+) (mean ± SE, n = 3, *P < 0.05). All results are representative of at least 3 independent experiments.

Because less cellular infiltrate existed in the MKK3^−/− lungs than in WT lungs after LPS exposure, we next determined the populations of inflammatory cells recruited to the lung parenchyma. We performed whole lung digests and analyzed the cells using flow cytometry. Ly6G- and GR-1-positive cells were identified as neutrophils, and F4/80- and CD45-positive cells were identified as macrophages. We detected more neutrophils and macrophages in the lungs of WT mice than in MKK3^−/− mice after LPS treatment (Fig. 2D). There was no difference in the number of lymphocytes between WT and MKK3^−/− mice. These results show that MKK3^−/− mice are more resistant to death, lung injury, and inflammation resulting from LPS-induced sepsis.

Mitochondrial injury and ROS production are lower in MKK3^−/− endothelial cells.

Mitochondrial dysfunction and ROS appear to be driving forces in sepsis because they cause cellular damage and inflammation (11). Circulating mitochondrial DNA appears to mediate systemic inflammatory responses after injury (54). Because mitochondria in the lungs of WT mice differed from those of MKK3^−/− mice in appearance (Fig. 2), we assessed the levels of mitochondrial DNA (mtDNA) in the serum of LPS-treated mice. We found increased mtDNA copy number in the serum of WT mice increased after LPS exposure, indicating a loss of mitochondrial and cellular integrity. In MKK3^−/− mice, however, the levels of serum mtDNA remained unchanged before and after LPS (Fig. 3A).

Fig. 3.

Mitochondrial injury and reactive oxygen species (ROS) production were lower in MKK3^−/− mice and endothelial cells after LPS. A: bar graphs of mitochondrial DNA in serum of septic mice. WT and MKK3^−/− mice given intraperitoneal LPS (40 mg/kg, 6 h) and serum was checked for the presence of mitochondrial DNA by quantitative PCR (qPCR). Mitochondrial genes detected were cytochrome b (CYTB), NADH dehydrogenase subunit 1 (ND1), ATPase 6 (AP6) and cytochrome c oxidase subunit I (COX1) (*P < 0.05). B: representative fluorescence quantification of LPS-exposed cells. WT and MKK3^−/− lung endothelial cells were exposed to LPS (1 μg/ml, 6 h). C, control. Cells were stained with Mitosox red, which detects mitochondrial ROS, and levels were measured by flow cytometry. Values are expressed as mean fluorescent intensity ± SE (P < 0.05). All results are representative of at least 3 independent experiments.

The endothelium is central to the pathogenesis of sepsis through effects on inflammation, leukocyte recruitment, vascular tone, coagulation, and thrombosis (34). On the basis of our finding that LPS treatment disrupts endothelial cells in the lung, we focused our study on the endothelium. The release of mtDNA from cells is dependent on ROS production (52, 53). Given that mitochondria are an important source of ROS, we determined whether there were differences in the production of mitochondrial ROS between MKK3^−/− and WT endothelial cells using Mitosox red, a fluorescent dye that detects mitochondrial superoxide (31). We found lower levels of mitochondrial ROS in MKK3^−/− than in WT cells (Fig. 3B). Interestingly, we also found that levels of mitochondrial ROS in MKK3^−/− cells were lower at baseline as well as after LPS exposure, indicating that MKK3^−/− mice and lung endothelial cells have less mitochondrial injury and ROS production.

MKK3^−/− endothelial cells have stress-resistant mitochondria.

Our finding that mitochondrial ROS was lower in MKK3^−/− endothelial cells compared with WT cells, even at baseline, suggests that mitochondrial biogenesis or function is altered in MKK3-deficient cells. We used qPCR of mtDNA and fluorescent imaging to assess mitochondrial abundance in the lungs and in endothelial cells of WT and MKK3^−/− mice. On the basis of these assays, we concluded there are more mitochondria in lungs and endothelial cells of MKK3^−/− mice than in WT mice (Fig. 4A). As an estimate of the functional capacity of the mitochondrial network, we costained cells with MitoTracker Green and MitoTracker Deep red, which estimate mitochondrial mass and membrane potential, respectively. The percentage of cells that were high for both Mitotracker Deep Red and Mitotracker Green, indicating mitochondria with preserved membrane potential, was significantly higher in MKK3^−/− than in WT cells (Fig. 4, B and C) at baseline and in response to LPS.

Fig. 4.

MKK3^−/− endothelial cells have higher mitochondrial mass and potential compared with WT cells. A: bar graphs of relative mitochondrial DNA. Mitochondrial mass was measured in lungs and endothelial cells as relative copy number of mitochondrial gene (cytochrome b) compared with nuclear gene (18s RNA gene) (*P < 0.05). B: flow cytometry of WT and MKK3^−/− endothelial cells left untreated or stimulated with LPS (1 μg/ml, 6 h) or carbonyl cyanide m-chlorophenyl hydrazine (CCCP, 50 μM) or ATP (5 mM) for 15 min. Cells were stained with Mitotracker Deep Red and Mitotracker Green before analysis. Data are representative of 3 experiments. C: quantitative representation of mean fluorescent intensity of Mitotracker staining and % of cells positive for both Mitotracker Deep Red and Green. Data are representative of 3 experiments (*P < 0.05 compared with WT cells, #P < 0.05 compared with untreated controls). All results are representative of at least 3 independent experiments.

Given that MKK3^−/− cells have more intact mitochondria after LPS challenge, we asked whether they are protected against specific mitochondrial stress. Sepsis is associated with mitochondrial depolarization and production of ROS (2, 11). We exposed WT and MKK3^−/− endothelial cells to carbonyl cyanide m-chlorophenyl hydrazine and ATP to simulate sepsis-induced mitochondrial uncoupling and ROS generation (7, 32) and found that MKK3^−/− endothelial cells again maintained more functional mitochondria than WT cells (Fig. 4, B and C), indicating that MKK3^−/− endothelial cells have mitochondria that are resistant to LPS and other mitochondrial stressors, or that the increase in biogenesis is protective.

Sirt1 expression and activity are increased in MKK3^−/− lungs and endothelial cells.

Given the differences in mitochondrial integrity and mitochondrial mass we observed in MKK3^−/− endothelial cells, we investigated two key processes affecting mitochondrial abundance and quality: mitochondrial biogenesis and mitophagy. We found that levels of Sirt1, a regulator of mitochondrial biogenesis (36), are elevated in MKK3^−/− lungs and endothelial cells compared with WT cells (Fig. 5, A and B).

Fig. 5.

Sirt1 expression and activity are increased in MKK3^−/− lungs and endothelial cells. A: levels of Sirt1 checked by Western blots of lung lysates and endothelial cells after LPS exposure (40 mg/kg, 6 h and 1 μg/ml, 6 h respectively). Densitometric quantification of Sirt1 is shown in the graph. (*P < 0.05 compared with WT). B: Sirt1 levels were measured in WT and MKK3^−/− lungs by qPCR (*P < 0.05). C: levels of PGC-1α checked by Western blots for lung lysates in WT and MKK3^−/− mice after LPS exposure (40 mg/kg, 6 h). Densitometric quantification of PGC-1α is shown in the graph. (*P < 0.05 compared with WT). D: PGC-1α levels were measured in WT and MKK3^−/− lungs by qPCR (*P < 0.05). E: lung lysates from control and LPS-exposed mice (40 mg/kg, 6 h) were immunoprecipitated (IP) first with a PGC-1α-specific antibody followed by Western immunoblotting (IB) with a acetylated lysine antibody. − control, negative control in which lung lysates were processed without the immunoprecipitation antibody. Densitometric quantification of acetylated lysine is shown in the graph. (*P < 0.05 compared with WT). F: nuclear extracts from lungs of control and LPS-exposed mice (40 mg/kg, 6 h) were checked by Western blots for levels of transcription factor Nrf1. Lamin B was used as a nuclear loading control. Densitometric quantification of Nrf1 is shown in the graph. (*P < 0.05 compared with WT). G: Nrf1 levels were measured in WT and MKK3^−/− lungs by qPCR (*P < 0.05). All results are representative of at least 3 independent experiments.

Sirt1 activates the transcription factor PGC-1α, a key regulator of mitochondrial biogenesis. Because we saw higher Sirt1 levels in MKK3^−/− than in WT tissue we sought to determine whether PGC-1α levels and activity were altered. Our results show that MKK3^−/− lungs had increased PGC-1α mRNA and protein (Fig. 5, C and D).

Sirt1 interacts with and deacetylates PGC-1α, leading to its activation (33, 40). Therefore, we measured the acetylation status of PGC-1α in WT and MKK3^−/− lungs using immunoprecipitation and subsequent probing with acetylated lysine antibody (Fig. 5E). We show that MKK3^−/− lungs have lower PGC-1α acetylation, consistent with increased activation by Sirt1.

We next determined whether the increased activation of PGC-1α by Sirt1 in MKK3^−/− lungs leads to activation of a downstream transcription factor, Nrf1, which promotes mitochondrial biogenesis in collaboration with PGC-1α (50). MKK3^−/− lungs had higher levels of Nrf1 compared with WT lungs (Fig. 5, F and G), suggesting that the observed increase in mitochondrial biogenesis in MKK3^−/− cells is due, at least in part, to Sirt1-dependent activation of PGC-1α and coactivation of nuclear-encoded mitochondrial genes with Nrf1.

MKK3^−/− endothelial cells have increased mitophagy.

Mitophagy involves the selective degradation of dysfunctional mitochondria (e.g., low membrane potential) through engulfment by vesicles that are coated with the autophagosome marker LC3B. Hence, mitophagy regulates mitochondrial “quality control” by removing defective mitochondria, including those that have the potential to produce excessive ROS (51). Because we found that mitochondrial ROS was diminished in MKK3^−/− endothelial cells, we analyzed expression of the mitophagy marker LC3B and found that it was higher in MKK3^−/− endothelial cells than in WT endothelial cells (Fig. 6A). LC3B was mainly localized to the mitochondria in MKK3^−/− cells whereas in WT cells LC3B was present in both the mitochondrial and cytoplasmic compartments (Fig. 6, B and C). This indicates that, in MKK3^−/−-deficient endothelial cells, most of the LC3B localizes to the mitochondria suggesting upregulated mitophagy. As elevated levels of LC3B could result from defects in the targeted destruction of the autophagic vacuoles rather than from autophagy, we sought to determine whether autophagic flux is intact in MKK3^−/− cells. For this purpose, we used a lysosomal inhibitor, leupeptin, to arrest the autophagic pathway downstream of LC3B accumulation in the autophagosome. We found that LC3B accumulated in MKK3^−/− cells after exposure to leupeptin, indicating that autophagic flux occurs in MKK3^−/− cells (Fig. 6D).

Fig. 6.

MKK3^−/− endothelial cells have higher levels of mitophagy than WT endothelial cells. A: Western blots of endothelial cells. LC3B, a marker of mitophagy, was measured in control and LPS-exposed (1 μg/ml, 6 h) endothelial cells. Actin is the loading control. B: fluorescent microscopy of control and LPS exposed (1 μg/ml, 6 h) endothelial cells to show localization of LC3B with the mitochondria. LC3B is stained red and mitochondria are stained with antibodies against Hsp60 (green). Magnification ×100. C: quantification of colocalization is represented by the colocalization coefficient, the fraction of LC3B that colocalizes with Hsp60 (*P < 0.05 compared with WT cells, #P < 0.05 compared with untreated control). D: Western blots of MKK3^−/− endothelial cells. Autophagy flux was checked in MKK3^−/− endothelial cells by measuring LC3B levels after treatment with leupeptin (250 μM for 14 h). All results are representative of at least 3 independent experiments.

Pink1 and Parkin are known initiators of mitophagy (32). Because mitophagy is increased in MKK3^−/− endothelial cells and lungs compared with WT, we asked whether Pink1 and Parkin levels were altered in MKK3^−/− tissues. We found higher levels of both Pink1 and Parkin in the MKK3^−/− lung and endothelial cell lysates than in WT (Fig. 7). These results suggest that mitophagy as well as mitochondrial biogenesis increase in MKK3^−/− endothelial cells to optimize clearance of defective mitochondria and ensures maintenance of a functional mitochondria network.

Fig. 7.

MKK3^−/− mice have higher levels of Pink1 and Parkin than WT. Pink1 and Parkin (markers of mitophagy) were measured by Western blot analysis in WT and MKK3^−/− lungs and endothelial cells. Densitometric quantifications of Pink1 and Parkin are shown in the graphs. (*P < 0.05 compared with WT). All results are representative of at least 3 independent experiments.

Inhibition of Sirt1 and mitophagy decreases mitochondrial integrity in MKK3^−/− endothelial cells.

To confirm that Sirt1 and mitophagy mediate the mitochondrial phenotype of MKK3^−/− endothelial cells we used siRNAs to inhibit Sirt1, LC3B, Pink1, and Parkin as well as 3-methyladenine, which specifically inhibits autophagy by blocking autophagosome formation (41). We observed a reduction in the number of polarized mitochondria in MKK3^−/− endothelial cells after inhibition of Sirt1 and mitophagy (Fig. 8A). In addition, when MKK3 was inhibited in WT endothelial cells by use of siRNA, Sirt1, LC3B, Pink1, and Parkin mRNA increased, indicating that MKK3 is upstream of these mediators of mitochondrial biogenesis and mitophagy (Fig. 8B).

Fig. 8.

Inhibition of Sirt1 and mitophagy reverses the mitochondrial integrity of MKK3^−/− endothelial cells. A: quantitative representation of mean fluorescent intensity of Mitotracker staining. Sirt-1 was inhibited by using siRNA and mitophagy was inhibited in MKK3^−/− endothelial cells with LC3B, Pink1, and Parkin siRNA and 3-methylademine (3-MA, 5 mM for 4 h). Cells were stained with Mitotracker Deep Red and Mitotracker Green before analysis (*P < 0.05). B: qPCR of Sirt-1, LC3B, Pink1, and Parkin mRNA were measured after MKK3 silencing by using siRNA in WT endothelial cells with siRNA (*P < 0.05). All results are representative of at least 3 independent experiments.

Inhibition of mitophagy increased susceptibility of mice to sepsis.

To determine the role of mitophagy in susceptibility to LPS-induced sepsis and lung injury in vivo, we generated mice that are heterozygous for Pink1, MKK3, or both, and exposed them to LPS. We found that Pink1^+/− mice were more susceptible to death after a lethal dose of LPS than WT mice. MKK3^+/− Pink1^+/− mice had significantly higher survival than Pink1^+/− mice, indicating that loss of MKK3 can rescue the deleterious effects of Pink1 during sepsis (Fig. 9A). We did not generate adequate numbers of homozygous Pink null mice on our background of interest, C57BL/6. In addition, use of heterozygous Pink1 null mice enabled us to determine whether MKK3 inhibition can increase Pink1 expression, which was not feasible in complete Pink1 null mice (21). PCR analysis of lung tissue from MKK3^+/− Pink1^+/− mice shows that MKK3 deficiency restored mRNA expression of Pink1 (Fig. 9B), confirming that MKK3 regulates Pink1 and that the resistance of MKK3-deficient mice to LPS is, at least partly, due to increased mitophagy. We also show that MKK3 heterozygous mice are partially protected against acute sepsis indicating a protective dose-response of MKK3 deficiency (Fig. 9A).

Fig. 9.

Inhibition of mitophagy in mice increased susceptibility to sepsis and lung injury. A: a lethal dose of intraperitoneal LPS (40 mg/kg) was given to WT mice and heterozygous MKK3^+/−, Pink1^+/−, and Pink^+/−, MKK3^+/− mice and survival assessed (n = 7, *P < 0.05 compared with WT and Pink^+/− mice). B: Pink1 mRNA levels were checked by qPCR in lungs of WT mice and heterozygous MKK3^+/−, Pink1^+/− and Pink^+/−, MKK3^+/− mice (*P < 0.05).

MKK3 and Pink1 levels were higher in septic patients.

To determine the clinical relevance of our findings we conducted a study of PBMCs from 22 patients admitted to the medical ICU. Eight of the 22 patients were septic on admission and had significantly higher levels of p-MKK3/6 (phosphorylated or activated MKK3/6) than nonseptic controls (n = 14) (Fig. 10A and Table 2). The Alphascreen assay used here cannot distinguish between activity of MKK3 and that of a closely related kinase, MKK6. Therefore, to determine MKK3 activity we performed Western blots using an antibody with putative selectivity to p-MKK3. Subsequently we analyzed six additional ICU patients (two nonseptic and four septic). Both the nonseptic patients had a diagnosis of fluid overload and congestive heart failure. We found a measurable increase in total, p-MKK3 (phosphorylated or activated MKK3), and Pink1 levels in septic patients compared with nonseptic controls. Remarkably, one patient with severe septic shock with gram-negative rod bacteremia and requiring vasopressors had the highest MKK3 and p-MKK3 levels and the lowest Pink1 levels (Fig. 10B). These results suggest that high MKK3 activity is associated with worse sepsis and lower Pink1 levels.

Fig. 10.

MKK3 and Pink1 levels were higher in septic patients. A: p-MKK3/6 levels were measured in peripheral blood mononuclear cells (PBMCs) of septic and nonseptic intensive care unit patients by AlphaScreen SureFire assay. Relative expression is normalized to HeLa negative controls (*P < 0.05). B: MKK3, p-MKK3, and Pink1 levels were measured in PBMCs of septic and nonseptic patients by Western blots (black arrow). Relative expression was normalized to HeLa negative controls and β-actin levels. Gray arrows in bottom panels indicate the relative expression of MKK3, p-MKK3, and Pink1 levels in a severely ill patient with gram-negative rod bacteremia and shock (*P < 0.05).

Table 2.

Clinical data of nonseptic and septic patients

Patient No.	Sex	Age	Diagnosis	Shock	Pressors	Source of Sepsis	Organism
Nonsepsis
1	Male	66	Heart failure	+	+	−	−
2	Female	56	COPD exacerbation	−	−	−	−
3	Male	46	Alcohol withdrawal	−	−	−	−
4	Female	70	Seizures	−	−	−	−
5	Female	73	Heart failure	−	−	−	−
6	Female	70	COPD exacerbation	−	−	−	−
7	Male	69	Blast crisis	−	−	−	−
8	Male	52	Nephrolithiasis	−	−	−	−
9	Female	83	NSTEMI	−	−	−	−
10	Male	38	Alcohol withdrawal	−	−	−	−
11	Male	43	Syncope, stroke	−	−	−	−
12	Male	52	Drug ingestion	+	+	−	−
13	Female	72	DKA	−	−	−	−
14	Female	75	Atrial fibrillation	+	−	−	−
Sepsis
1	Female	54	Sepsis	+	+	Pneumonia	NA
2	Male	57	Sepsis	−	−	Pneumonia	NA
3	Female	65	Sepsis	+	−	UTI	Entrococcus
4	Male	48	Sepsis	+	+	UTI	Escherichia coli
5	Male	67	Sepsis	+	+	Abdominal abscess	Group B Streptococcus
6	Male	60	Sepsis	−	−	Pneumonia	NA
7	Male	76	Sepsis	−	−	UTI	Proteus
8	Male	87	Sepsis	+	+	Pneumonia	E. coli

COPD, chronic obstructive pulmonary disease; NSTEMI, non-ST elevation myocardial infarction; DKA, diabetes ketoacidosis; UTI, urinary tract infection.

DISCUSSION

Our findings indicate that MKK3 mediates susceptibility to sepsis through the regulation of mitochondrial biogenesis and mitophagy. We show that MKK3 deficiency is protective against sepsis and MKK3 enhances mitochondrial function and turnover. Mitochondrial dysfunction is associated with increased severity and poor outcomes in patients with sepsis and lung injury (4, 6). A recent study of the metabolome and proteome in septic patients found mitochondrial dysfunction as a signature that determines sepsis outcomes (27). We postulate that regimens that improve mitochondrial homeostasis may offer new effective treatment options for sepsis and that pharmacological activators of biogenesis and mitophagy specifically may be a new class of drug target in sepsis. The endothelial bed plays an important role in preserving vascular homeostasis. Endothelial dysfunction appears to be important in sepsis and multiorgan failure. The endothelium is both the target and source of ROS in sepsis (17), and mitochondria are an important source of ROS in the endothelium. Specifically, autophagy and mitophagy in the endothelium are thought to be protective processes against vascular dysfunction (22). In addition Sirt1-dependent PGC-1α activation and mitochondrial biogenesis are important for endothelial protection against oxidative stress (44, 45). We have identified MKK3 as a regulator of mitochondrial quality control in the endothelial cells through the modulation of both mitochondrial biogenesis and mitophagy.

MKK3 is the upstream kinase of p38 MAPK and activates specific isoforms p38α, p38γ, and p38δ, but not p38β, under conditions of stress (8). During hypoxia mitochondrial ROS activate p38 through MKK3 (24); mitochondrial ROS also appear to activate MKK3 and p38α through the Ask1 signalasome (16). Our results show that MKK3 activity modulates mitochondrial ROS, indicating that mitochondrial signaling is both upstream and downstream of MKK3 and may involve a feedback loop. Intriguingly, p38 has been reported to be present in mitochondria, where it regulates apoptotic responses to ceramide (23). This may offer a potential mechanism by which MKK3 and p38 can affect mitochondria. Since MKK3-p38 and Pink1 belong to the same family of serine-threonine kinases, it is also plausible that MKK3-p38 phosphorylates Pink1 directly, although this remains to be tested.

Activation of autophagy has several beneficial effects, including extended lifespan, better motor and metabolic function, and resistance to oxidative stress (37). Our laboratory has recently reported that endothelial autophagy is an important determinant of survival against lethal hyperoxia (57). Activation of autophagy is also protective against sepsis and multiorgan failure and mitochondrial biogenesis is increased in the livers of septic mice through a TLR4- and autophagy-dependent pathway (5). Parkin-induced mitophagy promotes recovery of myocardial and mitochondrial dysfunction in LPS models of sepsis (35). Notably, autophagy in the lung appears to confer the highest protection against sepsis, and upregulation of autophagic pathways in the lung alone is sufficient to protect against death and multiorgan failure (10). This suggests that, as one of the most common organs to fail in sepsis, lung responses may drive the dysfunction in other target organs, such as liver and kidney. Developing lung-protective therapies may be critical in improving mortality and multiorgan failure in sepsis. We show that autophagy, as measured by LC3B, and mitophagy, as measured by Pink1 and Parkin levels, are increased in the lungs of MKK3-deficient mice, thus providing significant protection against lethal sepsis.

To maintain a healthy mitochondrial population, there must be constant equilibrium between mitophagy and mitochondrial biogenesis. Unchecked mitophagy in the absence of biogenesis would result in depletion of mitochondria. Thus mitophagy and mitochondrial biogenesis need to be maintained in a fine balance. How cells coordinate biogenesis and mitophagy is still largely unknown and may occur through multiple mechanisms. We have shown that inhibition of MKK3 causes a simultaneous increase in both biogenesis and mitophagy. This increase in both processes appears to require an increase in Sirt1 activity and Pink1-Parkin. Sirt1 mediates biogenesis through the action of PGC-1α and also increases levels of autophagy and mitophagy (18, 28). On the other hand studies have shown that Pink1 and Parkin also increase mitochondrial biogenesis (9, 13, 25, 42). PGC-1α regulation of mitochondrial biogenesis is activated by a variety of mediators at the translational level and through posttranslational modifications. Transcription of PGC-1α mRNA is increased by the mammalian target of rapamycin (mTOR) pathway and it is activated through deacetylation by Sirt1 or phosphorylation by AMP-activated protein kinase (AMPK) or p38 kinases (47). We show here that there is increased deacetylation of PGC-1α in MKK3-deficient tissues; however, it is possible that MKK3 regulates PGC-1α through other mechanisms as well. Further study is needed to determine precisely how MKK3 regulates these processes.

Mitochondrial biogenesis and mitophagy maintain mitochondrial integrity in the face of oxidant stress and other insults. This process may be called “reactive biogenesis and mitophagy” (26) wherein the cell shows a defensive response to an existing extrinsic stress. However, there are some environmental conditions, such as caloric restriction or exercise, in which there is an increase in biogenesis and mitophagy. This is beneficial and protects the organism and cells against degradation and stresses that it may encounter in the future. We call this “proactive” mitochondria quality control, a term that encompasses conditions of protective baseline increases in both biogenesis and mitophagy. In this study, we describe a condition in which MKK3 deficiency at baseline appears to result in proactive mitochondrial quality control through upregulation of both biogenesis and mitophagy that we propose is the basis of protection upon exposure to acute stressors such as sepsis.

In summary, our findings show that MKK3 deficiency increases mitophagy in lung and endothelial cells and that elevated Sirt1 promotes mitochondrial biogenesis through PGC-1α and Nrf1. Thus MKK3 deficiency increases both removal of defective mitochondria and formation of new mitochondria, thereby increasing the number of functional mitochondria and maintaining a healthier network (Fig. 11). This environment leads to less mitochondrial ROS production in MKK3-deficient tissues and cells, resulting in decreased inflammation and injury in response to sepsis. Using an Alphascreen kinase assay we show that that the activity of MKK3 is significantly higher in septic patients compared with nonseptic ICU patients and that MKK3 activity may serve as a marker of sepsis severity in the ICU. These new insights into MKK3 function, mitophagy, mitochondrial biogenesis, and sepsis suggest that MKK3 inhibitors may become novel targets in the treatment of sepsis.

Fig. 11.

Summary of MKK3 effects on mitochondrial health. We postulate that MKK3 deficiency increases both biogenesis (through action of Sirt1, PGC-1α, and Nrf1) and mitophagy (through the actions of Pink1 and Parkin), leading to a larger pool of healthy mitochondria (gray) and less ROS-producing defective mitochondria (brown). The net effect is protection against lethal sepsis.

GRANTS

P. Mannam is supported by American Heart Association grant, AHA 09FTF2090019. A. M. Ahasic is supported by American Heart Association grant AHA 10FTF 3440007. P. J. Lee is supported by R01 HL090660, R01 HL071595, and FAMRI 82384. G. S. Shadel is supported by R01 AG047632 and a joint grant from the United Mitochondrial Disease Foundation and Mitocon. A. P. West is supported by a Postdoctoral Fellowship, PF-13-035-01-DMC, from the American Cancer Society. P. Mannam and M. Trentalange were supported by the Yale Claude D. Pepper Older Americans Independence Center (P30 AG021342). P. Mannam was also supported by NIA R03 AG 042358-02 (GEMSSTAR).

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the author(s).

AUTHOR CONTRIBUTIONS

P.M., R.F.N., W.E.W., M.B., J.M., M.-J.K., C.S.D.C., A.M.A., M.A.P., M.T., A.P.W., G.S.S., J.A.E., and P.J.L. conception and design of research; P.M., A.S.S., A.S., and R.F.N. performed experiments; P.M., M.T., and P.J.L. analyzed data; P.M., W.E.W., M.B., J.M., M.-J.K., C.S.D.C., A.M.A., M.A.P., M.T., A.P.W., G.S.S., J.A.E., and P.J.L. interpreted results of experiments; P.M. prepared figures; P.M. drafted manuscript; P.M., M.B., J.M., A.M.A., and P.J.L. edited and revised manuscript; P.M., A.S.S., R.F.N., W.E.W., M.B., J.M., M.-J.K., C.S.D.C., A.M.A., M.A.P., M.T., A.P.W., G.S.S., J.A.E., and P.J.L. approved final version of manuscript.