Anti‐inflammatory effects of miR‐21 in the macrophage response to peritonitis

Short abstract

MiRNA‐21 is beneficial in survival following induction of peritonitis with LPS by limiting the pro‐inflammatory phase of the innate response.

Abstract

We investigated the role of microRNA‐21 in the macrophage response to peritonitis; microRNA‐21 expression increases in peritoneal macrophages after lipopolysaccharide stimulation but is delayed until 48 hours after cecal ligation and puncture. MicroRNA‐21–null mice and bone marrow–derived cell lines were exposed to cecal ligation and puncture or lipopolysaccharide, and survival, microRNA‐21 levels, target messenger RNAs and proteins, and cytokines were assayed. Macrophages were also transfected with microRNA‐21 mimics and antagomirs, and similar endpoints were measured. Survival in microRNA‐21–null mice was significantly decreased after lipopolysaccharide‐induced peritonitis but unchanged after cecal ligation and puncture compared with similarly treated wild‐type mice. MicroRNA‐21 expression, tumor necrosis factor‐α, interleukin 6, and programmed cell death protein 4 levels were increased after lipopolysaccharide addition in peritoneal cells. Pelino1 and sprouty (SPRY) messenger RNAs were similarly increased early, whereas programmed cell death protein 4 messenger RNA was decreased after lipopolysaccharide, and all microR‐21 target messenger RNAs were subsequently decreased by 24 hours after lipopolysaccharide. Transfection with mimics and antagomirs led to appropriate responses in microRNA‐21 and tumor necrosis factor‐α. Knockdown of microRNA‐21 in bone marrow–derived cells showed increased tumor necrosis factor‐α and decreased interleukin 10 in response to lipopolysaccharide. Target proteins were unaffected by knockdown as was extracellular signal‐regulated kinase; however, the nuclear factor κB p65 subunit was increased after lipopolysaccharide in the microRNA‐21 knockout cells. In contrast, there was little change in these parameters after cecal ligation and puncture induction between null and wild‐type mice. MicroRNA‐21 is beneficial to survival in mice following lipopolysaccharide peritonitis. Overexpression of microRNA‐21 decreased tumor necrosis factor‐α secretion, whereas suppression of microRNA‐21 expression increased tumor necrosis factor‐α and interleukin 6, and decreased interleukin 10 levels after lipopolysaccharide. Protein targets of microRNA‐21 were not different following suppression of microRNA‐21. Nuclear factor κB was increased by suppression of microRNA‐21. These findings demonstrate microRNA‐21 is beneficial in modulating the macrophage response to lipopolysaccharide peritonitis and an improved understanding of the anti‐inflammatory effects of microRNA‐21 may result in novel, targeted therapy against peritonitis and sepsis.

Abbreviations

ceRNA

competing endogenous RNA

CLP

cecal ligation and puncture

C57BL/6

C57 black 6 mice strain

PEC

peritoneal exudate cell

Peli1

pelino1

PDCD4

programmed cell death protein 4

SPRY

sprouty

WT

wild type

Introduction

Peritonitis is regularly seen as a part of clinical surgical practice, and the resulting bacterial infections of the peritoneal cavity can have multiple origins. Patients with peritonitis can have a variable disease course, from a mild, self‐contained infection, through a whole‐body septic response, potentially leading to multiorgan failure and fatality [1]. Although improvements in critical care have led to increased survival, there has been little change in the treatment of peritonitis in the past 50 y [2]. The mainstay remains 1) source control by debridement or drainage of infected and necrotic tissue, 2) antibiotics directed at gram‐negative pathogens including anaerobes, and 3) supportive treatment of homeostatic functions in organs.

Mortality from peritonitis is directly associated with an exaggerated proinflammatory innate response with development of sepsis and multiple organ failure [3]. Several immune modulators have undergone clinical trials in recent years with the aim of manipulating the immune response to the benefit of the patient, but none have successfully been able to change patient outcomes [4, 5–6]. The purpose of this work was to investigate specific mechanisms by which the response of peritoneal macrophages to endotoxin and bacterial infection are regulated and to identify potential routes for therapeutic intervention.

The early innate immune response to invading microbes during peritonitis is largely orchestrated by the peritoneal macrophage. These cells have 3 basic functions: phagocytosis, antigen presentation, and protein secretion. Macrophages have been classed into 2 phenotypes: M1, the classic or proinflammatory macrophage that produces TNF‐α, IL‐1, IL‐6, nitric oxide, and has strong microbial activity; and M2 or alternatively activated macrophages that produce IL‐10, Arginase‐1, and Mrc1 (CD206) and are involved in preventing excessive injury in the host tissue remodeling [7, 8]. The mesothelial lining and peritoneal macrophages are the main source of chemokines and cytokines that attract neutrophils and circulating monocytes to the peritoneum [9]. IL‐10 produced by macrophages inhibits neutrophil recruitment and chemokine generation, which, in turn, inhibits proinflammatory cytokine release [10]. Diffusion of inflammatory mediators TNF‐α and IL‐1β from the circulation has an important role in the pulmonary activation of resident macrophages, which results in the associated lung disease and neutrophil infiltration seen in animal models [11, 12].

miRNAs modulate gene expression posttranscriptionally by negatively regulating the translation of mRNA to protein [13]. Historically, miRNA‐21 has a significant role in cardiovascular injury, cancer progression, and metastasis through regulating genes involved in apoptosis, cell proliferation, cell migration, and invasion. In particular, miRNA‐21 is up‐regulated in multiple types of cancers [14, 15, 16–17]. Additionally, miRNA‐21 in the plasma of patients with colorectal cancer shows the ability of miRNAs to act on their surrounding environment [18]. The significant role of miRNAs in cancer has led to further investigation into their role in the immune response [13, 19]; for example, miRNA‐21 appears to be pronecrotic in murine pancreatitis models [20, 21].

Currently, there are multiple miRNAs known to target the TLR‐4 signaling pathway [22]. This has been confirmed through the ability of LPS to induce miRNA‐21 via TLR‐4, MyD88, and NF‐κB, which has a binding site in the miRNA‐21 promoter region [23]. Importantly, transcription is stimulated by NF‐κB and knockdown of NF‐κB by small interfering RNA, and inhibitors similarly reduce expression of miRNA‐21 [24, 25]. The role of miRNA‐21 in the immune system is still under active investigation, and it is hypothesized that miRNA‐21 could be involved in the transition between the proinflammatory and anti‐inflammatory phases of the early innate immune response via macrophage polarization. The miRNA‐21 targets that may contribute to this transitional role of the immune response are PDCD4 [25, 26, 27–28], Peli1 [29], and SPRY proteins 1, 2 and 4 [29].

The proposed theory that miRA‐21 has a role in the transition from the proinflammatory to anti‐inflammatory phase of the innate immune response is due to its relationship to the TLR‐4 pathway. TLR‐4 activation leads to the initiation of the innate immune response and release of proinflammatory and anti‐inflammatory cytokines. The expression of miRNA‐21 has been shown to increase in RAW264.7 cells following LPS stimulation [25]. Recent studies have shown a role for miRNA‐21 in the regulation of TLR‐4 pattern recognition receptor signaling, which is present on macrophages [22, 25]. We propose that miRNA‐21 acts to decrease TNF‐α and increase IL‐10 production by a dual mechanism. Two proven targets of miRNA‐21, PDCD4 and Peli1, are positive regulators of the TLR‐4 signaling pathway [25, 30, 31]. Inhibition of PDCD4 and Peli1 by miRNA‐21 decreases the signaling of the TLR‐4 pathway and results in decreased expression of TNF‐α and its translation into protein. In contrast, the SPRY proteins are negative regulators of the TLR‐4 signaling pathway. Inhibition of SPRY expression leads to an increase in ERK activation that leads to an increase in IL‐10 expression and production. Successful modulation of TNF‐α and IL‐10 secretion by manipulation of miRNA‐21 expression may have potential therapeutic applications. To assess the clinical potential in the treatment of peritonitis, we compared the survival of miRNA‐21^+/+ and miRNA‐21^−/− mice in LPS and CLP models of peritonitis.

MATERIALS AND METHODS

Mice

All study protocols were approved by the Institutional Animal Care and Use Committee of the University of Louisville and the Robley Rex Veterans Affairs Medical Center. All mice were housed with food and water ad libitum according to Institutional Animal Care and Use Committee guidelines.

LPS‐induced peritonitis

Male C57BL/6 miRNA‐21^+/+ and miRNA‐21^−/− mice aged 12–18 wk were weighed, single‐housed, and fasted for 3 h before intraperitoneal injection of 25 mg/kg of Escherichia coli LPS (E. coli 055:B5; Sigma‐Aldrich, St. Louis, MO, USA). Surface temperature, daily body weight, and survival were recorded for 5 d. Mice were given food and water ad libitum.

CLP‐induced peritonitis

Male C57BL/6 miRNA‐21^+/+ and miRNA‐21^−/− mice aged 12–18 wk were weighed before undergoing cecal ligation and puncture. Briefly, under isoflurane anesthesia, the abdomen was incised and the cecum exteriorized and ligated 1 cm from the tip. A “through and through” puncture was made with a 20‐gauge needle, and the cecum was returned to the abdomen. The peritoneal layer was closed with 3–0 Ticron (Covidien, New Haven, CT, USA) and 50 μl of 2% lidocaine applied to the suture line. Skin was closed with 3–0 silk. Postoperatively, mice were given 600 μl subcutaneous 0.9% saline and single‐housed. Surface temperature, daily weight, and survival were recorded for 5 d. Mice were given food and water ad libitum.

Wild‐type mice for peritoneal macrophage experiments

Male C57BL/6 mice (6 wk old) were obtained from Jackson Laboratories (Bar Harbor, ME, USA).

MiRNA‐21 knockout mice and WT controls

The gene targeting of miRNA‐21 in mice was performed by inGenious Targeting Laboratory (Ronkonkoma, NY, USA) as previously described [32]. Genotype was confirmed before experimentation.

Bone marrow macrophage cell lines

Immortalized macrophage cell lines were established by infecting the bone marrow of miRNA‐21^+/+ and miRNA‐21^−/− mice with the murine recombinant J2 retrovirus containing the v‐myc and v‐ref oncogenes, as previously described [33, 34–35]. Cells were cultured in RPMI‐1640 medium with 5% FBS, 1% l‐glutamine, 1% HEPES, and 0.1% gentamicin (all Sigma‐Aldrich) at 37°C in a humidified incubator with 5% CO₂. Absence of miRNA‐21 was confirmed by PCR.

Primary peritoneal cells

Naïve mice were anesthetized with 2–3% isoflurane (Butler Schein, Dublin, OH, USA), and the abdomen was lavaged with 3 ml of heparinized PBS with no calcium or magnesium (Sigma‐Aldrich). Lavage fluid was centrifuged to pellet PECs for macrophage isolation. Magnetic bead isolation products were purchased from Miltenyi Biotec (San Diego, CA, USA). PECs of each mouse were resuspended in autoMACS (Miltenyi Biotec) running buffer and incubated at room temperature for 15 min with CD11b beads. After washing, the cells were resuspended in running buffer and passed through MS columns in a MiniMACS separator magnet per manufacturer's protocol. This macrophage‐enriched population was manually quantified by hemocytometer with the trypan blue exclusion method to determine total macrophage numbers and cell viability. Flow cytometry confirmed that >82% of CD11b⁺ cells isolated from the peritoneum of naïve WT mice were F4/80⁺.

LPS stimulation

For most experiments the macrophages were suspended at 0.5 × 10⁶/ml in RPMI‐1640 medium with 10% FBS, 1% l‐glutamine, and 1% antibiotic/antimycotic (Sigma‐Aldrich) and were plated in a 24‐well plate at 0.5 × 10⁶/ml per well. Cells were rested overnight at 37°C in a humidified incubator with 5% CO₂ before stimulation. Samples were stimulated with LPS (100 ng/ml, E. Coli 011:B4; Sigma‐Aldrich). After stimulation, cells were incubated for the indicated periods. The culture medium was removed and stored at −80°C for cytokine quantification. Macrophages were allocated to either RNA isolation or lysed in 100 μl of radioimmunoprecipitation assay buffer (Sigma‐Aldrich) for protein quantification and Western blot.

RNA isolation and miRNA and mRNA expression

Total RNA was isolated from the macrophages using mirVana miRNA isolation kits (Ambion, Life technologies, Grand Island, NY, USA) following manufacturer's protocol. RNA concentrations were determined and RNA purity was assessed using the NanoDrop 2000 (Thermo Fisher Scientific, Hudson, NH, USA).

MiRNA‐21 expression was measured using Taqman single miRNA assays and Taqman MiRNA Reverse Transcription Kit for miRNA (Applied Biosystems, Life Technologies). U6 was used as a housekeeping gene miRNA for normalization.

For mRNA expression, RNA was transcribed into cDNA using high‐capacity cDNA reverse transcription kit for mRNA (Applied Biosystems). Taqman single‐gene assays for PDCD4, Peli1, SPRY1, and SPRY4; 18s was used as an internal control.

The StepOne Plus (Applied Biosystems) was used to run the real‐time PCR using fast advanced master mix and the fast protocol (all Life Technologies). StepOne Plus software v2.1 (Applied Biosystems) generated cycle threshold values, which were used to calculate the fold changes using the ΔΔ cycle threshold method [36].

Cytokine assays

TNF‐α, IL‐6, and IL‐10 levels were measured in lavage media and culture media supernatants using ELISAs. Murine Ready‐Set‐Go kits were purchased for all ELISAs from e‐Biosciences (San Diego, CA, USA) and then performed according to the manufacturer's instructions. All samples were analyzed in duplicate. Where necessary, samples were diluted in assay diluent to achieve cytokine concentrations within the range of the standard curve.

Protein isolation and Western blot

Protease and phosphatase inhibitors (Thermo Fisher Scientific) were added to proteins in radioimmunoprecipitation assay buffer and sonicated using a Sonifier 250 (Branson Ultrasonics Danbury, CT, USA). Protein quantification was determined by BCA assay (Thermo Fisher Scientific, Rockford, IL, USA) against a standard curve on Dynatech MR4000 using BioLinx 2.0 software.

All products for Western blots were purchased from Life Technologies (Carlsbad, CA, USA) unless otherwise specified and run according to manufacturer's protocols. Protein samples with 4× Bolt LDS sample buffer and 1:100 2‐mercaptoethanol were heated to 95°C for 5 min and loaded into a Bolt 4–12% Bis–Tris Plus mini gel. Gels were immersed in Bolt MES SDS running buffer in a Bolt Mini Gel Tank and run on a Bio‐Rad Power Pac 1000 (Hercules, CA, USA) at 165 mV for 35–60 min with SeeBlue Plus2 (Thermo Fisher) prestained standard. Proteins were transferred onto nitrocellulose membrane using iBlot Gel Transfer stacks and an iBlot transfer device (Thermo Fisher). Nonspecific binding was blocked with 7% nonfat milk in TBST for 1 h at room temperature. The membranes were incubated with rabbit anti‐mouse primary antibodies (Cell Signaling, Danvers, MA, USA, unless otherwise specified). All antibodies were suspended in 5% nonfat milk made with TBST for 1–2 h at room temperature or overnight at 4°C. Peli1 1:500 (Thermo Fisher), SPRY1 1:200 (Santa Cruz Biotechnology, Dallas, TX, USA), and SPRY4 1:500 (Thermo Fisher). Following incubation with each antibody, membranes were washed and subsequently incubated with rabbit anti‐mouse primary antibody for housekeeping protein β‐actin (1:10,000) or GAPDH (1:10,000) and then with goat anti‐rabbit secondary antibody (1:5,000). Pierce ECL Western blotting substrate (Thermo Fisher) was applied to the membrane as per the manufacturer's instructions. CL‐XPosure Film (Thermo Fisher) was used to expose the membrane for 15 min. The resultant image was developed, fixed, scanned, and quantified using ImageJ (U.S. National Institutes of Health, Bethesda, MD, USA).

MiRNA transfection

Primary peritoneal macrophages from naïve WT C57BL/6 mice were suspended in media. Cells were immediately transfected. MirVana mimic, inhibitor, and negative control to mmu–miR‐21‐5p (Life technologies) were assembled with N‐TER peptides (Sigma Aldrich) per manufacturer's protocol. Macrophages were incubated in a concentration of 120 nM of mimic, antagomir, or negative control for 24 h in at 37°C with 5% CO₂ in cell suspension. After 24 h, the samples were diluted with further media and stimulated with LPS 100 ng/ml. Six or 24 h after LPS stimulation, the culture media was collected for cytokine measurement, and the macrophages were processed for RNA isolation and subsequent analysis of miRNA expression.

Phosflow pathway analysis

MiRNA‐21^+/+ and miRNA‐21^−/− immortalized macrophages were resuspended in media. Cells in suspension were rested overnight before stimulation. After stimulation with LPS (200 ng/ml), cells were incubated for the indicated periods. Following the indicated time, the cells were fixed immediately with 3.2% formaldehyde. After washing, cells were permeabilized with chilled 100% methanol. Cell samples were split for analysis of NF‐κB or ERK activation. Intracellular staining occurred in the dark for 1 h with phospho‐NF‐κB p65 or phospho‐ERK AlexaFluor 488 antibodies (Cell Signaling). After staining, cells were washed and resuspended in FBS‐stain buffer (2% FBS in PBS, both Sigma Aldrich). For baseline pathway activation determination, similarly cultured, unstimulated miRNA‐21^+/+ and miRNA‐21^−/− macrophages were used. A FACS Calibur (BD Bioscience, San Jose, CA, USA) was used to acquire the samples. An LSR II (BD Bioscience) was used to acquire the ERK samples. We counted 10,000 events for each sample. CellQuest Pro (BD Bioscience) was used for analysis of NF‐κB, and http://www.cytobank.org (Cytobank, Mountain View, CA, USA) for ERK.

Statistical analysis

Data are shown as the means ± se. Sigma Plot 13.0 (Systat Software Inc, San Jose, CA, USA) was used to perform the Student's t test, ANOVA, and linear regression to determine statistical significance among appropriate groups. A P value of <0.05 was considered significant.

Results

MiRNA‐21 confers a significant survival benefit in response to LPS‐induced but not CLP‐induced peritonitis

To assess the protective role miRNA‐21 has in peritonitis, we compared survival of miRNA‐21^+/+ and miRNA‐21^−/− mice in 2 models of peritonitis: LPS‐ and CLP‐induced peritonitis. Following LPS‐induced peritonitis, significantly more miRNA‐21^−/− mice died compared with miRNA‐21^+/+ mice ( Fig. 1A ; P = 0.002). In contrast, there was no difference in survival between the miRNA‐21^+/+ and miRNA‐21^−/− mice in CLP‐induced peritonitis (Fig. 1B). Because of the very different nature of these models (sterile LPS vs. polymicrobial live infection), this contrast also has been observed with other potential immune modulators [37]. The combination of ischemia and live polymicrobial infection following CLP leads to a bacterial infection with a multitude of exogenous mediators, which signal through multiple pathways to initiate a more complex immune response [38, 39–40]. Following CLP, McClure et al. [41] also observed no difference in survival when BALB/c mice were injected with miRNA‐21 antagomirs; they required the combination with miR‐181b antagomirs to elicit a difference in survival. Notably, miRNA‐21 is induced by and acts on the TLR‐4 pathway [42], the crucial pathway activated following LPS stimulation, and implicates this mechanism in the observed difference in survival between miRNA‐21^+/+ and miRNA‐21^−/− mice after LPS‐induced peritonitis. Thus, we studied in more depth the LPS‐elicited responses.

Figure 1.

Survival after LPS and CLP peritonitis. (A) MiRNA‐21 confers a significant survival benefit in response to 25 mg/kg LPS peritonitis (MiRNA‐21^+/+ n = 17, miRNA‐21^−/− n = 9). (B) MiRNA‐21 does not convey a survival benefit in response to peritonitis by 20G CLP (n = 10). 20G, 20‐gauge.

MiRNA‐21 expression increases after LPS stimulation in WT primary peritoneal macrophages

The expression of miRNA‐21 has been shown to increase in RAW264.7 cells following LPS stimulation [25]. In cultured primary peritoneal macrophages, we confirmed that miRNA‐21 expression was increased in a linear fashion following LPS stimulation (R ² = 0.82; Fig. 2A and B ).

Figure 2.

Peritoneal macrophage miRNA‐21 and cytokine expression after LPS. (A) Peritoneal macrophage miRNA‐21 expression after LPS stimulation. (B) Linear regression, R ² = 0.82, P < 0.001. TNF‐α (C), IL‐6 (D), and IL‐10 E) concentrations from media of WT peritoneal macrophages after LPS stimulation, n = 8. Data are shown means ± sem.

There was a concurrent rise in TNF‐α, IL‐6 and IL‐10 following LPS stimulation of macrophages (Fig. 2C–E). More recent perceptions of the interactions between pro‐ and anti‐inflammatory responses to LPS stimulation have noted that these responses occur simultaneously, though the downstream effects maybe more delayed in the classic anti‐inflammatory response [40].

The proven or postulated mRNA targets of miRNA‐21 respond to LPS by one of 2 distinct patterns: 1) a sharp decrease, followed by a prolonged period of down‐regulation (PDCD4; Fig. 3A ); or 2) an immediate robust rise and subsequent return to baseline (Peli1, Fig. 3B; and SPRY1, Fig. 3C) with a sustained period of down‐regulation (SPRY4, Fig. 3D). Because miRNA‐21 is induced steadily over time (Fig. 2A), it will more likely contribute greater influence to delayed responses. For example, before induced miRNAs can act to influence mRNA levels or affect protein translation, they need to be incorporated into the RISC (RNA‐induced silencing complex) [43].

Figure 3.

Peritoneal macrophage target mRNA expression after LPS. WT peritoneal macrophage PDCD4 (A), Peli1 (B), SPRY1 (C), and SPRY4 (D) mRNA fold change after LPS stimulation, n = 8. Data are shown means ± sem.

A delay in miRNA action is, therefore, expected. Because of this complexity, we needed to test the regulatory effects of miRNA‐21 on TLR‐4 signaling by modulating miRNA‐21 expression.

miRNA‐21 affects cytokine production in macrophages after LPS stimulation

Using miR‐21 mimics and antagomirs, the modulation of the expression of miRNA‐21 in primary peritoneal macrophages confirmed an anti‐inflammatory effect of miRNA‐21 on TNF‐α and IL‐10 production in WT primary peritoneal macrophages in culture. Transfection of both miRNA‐21 mimics and antagomirs was successful, as confirmed by appropriate directional changes in miRNA‐21 gene expression relative to the negative control ( Fig. 4A ). Manipulation of miRNA‐21 expression with antagomirs and mimics led to statistically significant changes in TNF‐α concentration after LPS stimulation (Fig. 4B), which is directly linked to survival in sepsis [44, 45]. Overexpression of miRNA‐21 at 24 h did not result in a significant change in TNF‐α levels at 24 h. This may be due to ceRNAs [46, 47, 48–49], which contain multiple binding sites for the miRNA of interest and can act as effectively as artificial antisense oligonucleotides [50]. Therefore, the artificial up‐regulation of miRNA‐21 in peritoneal macrophages by mimics could be counteracted by a known or unknown mRNA target acting as a ceRNA. A ceRNA for miRNA‐21 has not yet been identified; however, PTEN (phosphatase and tensin homolog), a target of miRNA‐21 in tumor cells has been implicated in other miRNA networks [51].

Figure 4.

MiRNA‐21 transfection. (A) Transfection with MiRNA‐21 mimic in WT peritoneal macrophages led to significant overexpression of miRNA‐21 at both 6 and 24 h after LPS stimulation (P < 0.001 and P < 0.029, respectively). The use of antagomirs led to suppressed miRNA‐21 expression at both times, although only statistically at 6 h (P < 0.001). (B) Modulation of miRNA‐21 expression showed the expected results. Suppression of miRNA‐21 significantly increased TNF‐α at 6 and 24 h (P = 0.037 and P = 0.012, respectively). Overexpression significantly reduced TNF‐α at 6 h after LPS stimulation (P = 0.037) but not after 24 h. (C) Modulation of miRNA‐21 expression was not successful in modulating IL‐10 secretion in primary peritoneal macrophages at either time after LPS stimulation, n = 9. Data are shown means ± sem. *, P < 0.05 vs. negative control.

Modulation of miRNA‐21 levels did not have any effect on the levels of IL‐10 at 6 or 24 h after LPS stimulation (Fig. 4C). At 24 h, there was a trend toward increased IL‐10 in the mimic‐transfected macrophages; however, the effects were not consistent among mice. In these experiments, miRNA‐21 expression in peritoneal macrophages peaks at about 10‐fold up‐regulation at 120 h after LPS stimulation (Fig. 2A). In contrast, we observed a 1000‐fold up‐regulation in the mimic‐transfected macrophages (Fig. 4A), but this was still insufficient to elicit a significant change in IL‐10 production. We propose that the effects of miRNA‐21 on IL‐10 production are regulated through an ERK signaling cascade. MiRNA‐21 inhibits the Sprouty proteins, which in turn, inhibit the ERK cascade. There are 4 Sprouty isoforms; all of which are theoretical targets of miRNA‐21. Higher levels of miRNA‐21 may be required to sufficiently inhibit the protein levels to see a change in IL‐10 production, even at 24 h.

Confirmation of the results seen with transfection was made using miRNA‐21^+/+ and miRNA‐21^−/− bone marrow–derived macrophage cell lines regarding the pattern of cytokine production following LPS stimulation. Inhibition of endogenous miRNAs with transfection is often incomplete because transfection efficiency can vary from cell to cell within a treated sample.

Knockout of miRNA‐21 expression did not affect the overall trend of TNF‐α but resulted in an overall elevation of TNF‐α concentration from 6 h after LPS stimulation ( Fig. 5A ). IL‐6 continually trended upward and was significantly higher at 24 and 48 h (P < 0.05) in the miRNA‐21^−/− macrophages relative to WTs (Fig. 5B).

Figure 5.

Bone marrow derived macrophage cytokine expression after LPS. Knockout of miRNA‐21 led to significant increase in TNF‐α (A) at 6 and 24 h (P < 0.05) and trending toward significance at 48 h (P = 0.059). IL‐6 levels (B) were elevated, and IL‐10 levels (C) decreased following LPS stimulation in miRNA‐21 knockout cells, n = 8. Data are shown means ± sem. *, P < 0.05.

The miRNA‐21^−/− macrophages had significantly lower levels of IL‐10 at early time points after LPS stimulation (Fig. 5C) and did not reach miRNA‐21^+/+ levels until 24 h (P < 0.05). MiRNAs act by inhibiting protein translation; therefore, if a miRNA is constitutively turned off, as in knockout cells, we propose that there will be higher baseline levels of that protein expressed within the cell, leading to the increase IL‐10 concentrations seen from 30 min after LPS stimulation.

The observed changes in cytokine production between miRNA‐21^+/+ and miRNA‐21^−/− cells led to an expectation that we would detect associated differences in the miRNA‐21 target proteins, with higher levels present in the miRNA‐21^−/− cell lines. However, we were unable to detect a difference in miRNA‐21 target proteins (PDCD4, Peli1, and SPRY1, 2, and 4) at baseline before stimulation or at 24 h after LPS stimulation between miRNA‐21^+/+ and miRNA‐21^−/− macrophages ( Fig. 6 ).

Figure 6.

Bone marrow–derived macrophage protein expression after LPS. Knockdown of miRNA‐21 did not lead to a significant change in target proteins 24 h after LPS stimulation. PDCD4 representative Western blot (A) and relative abundance (B). Peli1 representative Western blot (C) and relative abundance (D). SPRY2 representative Western blot (E) and relative abundance (F), n = 4 Data are shown means ± sem.

PDCD4 protein has been shown to be reduced following LPS stimulation at 24 h in RAW264.7 cells, mouse bone marrow–derived macrophages, and human peripheral blood monocytes, with the same dose as used in these experiments [25]. Knockdown of PDCD4 protein levels was shown with PROmiRNA‐21 in RAW264.7 cells with and without LPS stimulation. This is contrary to what we saw in our results. This may be due to differences in cell culture concentrations, or the difference in protein levels may be more subtle than we are able to detect.

Knockout of MiRNA‐21 increases NF‐κB signaling in bone marrow macrophages after LPS stimulation

Changes in TNF‐α production have been shown to be associated with NF‐κB activation [52, 53], and we have demonstrated a higher level of NF‐κB p65 activation in miRNA‐21^−/− 15 min after LPS stimulation but no difference in total NF‐κB p65 ( Fig. 7A–C ). There was a maximal increase in the levels of NF‐κB 10 min after LPS stimulation (Fig. 7D).

Figure 7.

NF‐κB activation after lipopolysaccharide in bone marrow derived macrophages. Activation of NF‐κB p65 subunit was higher in miRNA‐21^−/− compared with miRNA‐21^+/+ bone marrow derived macrophage cell lines after lipopolysaccharide (LPS) stimulation. A) Phosphorylated NF‐κB p65, B) Total NF‐κB p65 both at 15 min after LPS and C) Relative Abundance of Phosphorylated NF‐κB p65. D) Mean channel fluorescence of Phosphorylated NF‐κB p65 as measured by Phosflow. N = 6. Data are shown means ± sem.

We proposed that knockout of miRNA‐21 would lead to decreased ERK activation, which correlates to lower levels of IL‐10; however, we did not detect a difference in total or phosphorylated ERK between the miRNA‐21^+/+and miRNA‐21^−/− macrophages.

MiRNA‐21 does not affect peritoneal macrophages after CLP‐induced peritonitis

CLP is among the most representative models of human peritonitis. We saw no difference in survival between miRNA‐21^+/+ and miRNA‐21^−/− mice following 20‐gauge needle CLP; however, we proposed, that despite the absence of a difference in survival, we would see a difference in the levels of miRNA‐21 in the macrophage population and the intraperitoneal cytokine profile. miRNA‐21 expression after CLP was initially down‐regulated until 24 h, but then increased by 48 h ( Fig. 8A ). This is in direct contrast to LPS peritonitis wherein a linear up‐regulation of miRNA‐21 was observed after stimulation. The early down‐regulation in the macrophages may be due to a delay in the stimulation of TLR‐4 by bacteria migrating from the lumen of the bowel into the peritoneum. Recent data show an increase of miRNA‐21 in early (day 3) and late (day 12) sepsis after CLP [41], but the immediate production of miRNA‐21 after CLP (days 1–3) has not previously been demonstrated. Because early up‐regulation of miR‐21 may be critical to differential protection in a setting of LPS, any delay in miR‐21 response could result in a “too little, too late” scenario as seen in our CLP model.

Figure 8.

Peritoneal macrophage expression after CLP. (A) Peritoneal macrophage miRNA‐21 expression after 20‐gauge CLP. There is no difference in the number of PECs (B), macrophages (C), or macrophages as a percentage of the total peritoneal population (D) between miRNA‐21^+/+ and miRNA‐21^−/−mice following CLP. Peritoneal cytokine expression was similar in TNF‐α (E), IL‐6 (F), and IL‐10 (G) between miRNA‐21^+/+ and miRNA‐21^−/−mice following CLP, n = 7–8. Data are shown means ± sem.

In fact, we observed no differences in PEC cell count, total macrophage count, or macrophages as a percentage of the total number of PEC cells following CLP in miRNA‐21^+/+ and miRNA‐21^−/− mice (Fig. 8B–D). All cytokines were similar regardless of the presence or absence of miRNA‐21 (Fig. 8E–G). These data provide a partial explanation of why there was no difference in survival observed between miRNA‐21^+/+ and miRNA‐21^−/− mice following induction of peritonitis with a 20‐gauge CLP, indicating perhaps a crucial therapeutic potential of early miR‐21 administration in polymicrobial peritonitis, which will need to be tested in further studies.

DISCUSSION

Macrophages have the plasticity to transform from one phenotype into another. The key transcription factors that control the transformation of macrophage phenotype include: signal transducer and activator of transcription, NF‐κB, and camp‐responsive element‐binding protein. Therapeutic targeting to change the polarization of macrophages is still in the primordial phase of development, but the microenvironment and anatomic location of the cells are important contributing factors [8, 54].

In this study, we focused on how miRNA‐21 could contribute to the macrophage microenvironment and in turn effect macrophage polarization following LPS stimulation. We saw that an increase in miRNA‐21 expression in macrophages following LPS stimulation resulted in a cytokine profile consistent with an M2, or anti‐inflammatory macrophage, as there was a decrease in TNF‐α and IL‐6 production and an increase in IL‐10. The mechanism by which the expression of these cytokines was modulated is not clear from these experiments. PDCD4 has emerged as the most likely target protein to be effecting the cytokine modulation, and changes in NF‐κB activation were observed by Western blot analysis. Overall, we have shown that miRNA‐21 modulates the cytokine output of peritoneal macrophages, from a proinflammatory to anti‐inflammatory profile, suggesting that miRNA‐21 is involved in macrophage polarization. Further characterization of whether there is any difference in the M2 expressed genes, such as Arginase‐1, chitinase 3‐like 3 (Ym1), and Mrc1 (CD206) between miRNA‐21^−/− and miRNA‐21^+/+ cells should yield novel data to clarify the active mechanisms involved in the transition of the early innate immune response to peritonitis. The role of macrophage polarization clinically has yet to be determined, but preconditioning with xenon in a similar LPS sepsis model increased miRNA‐21 expression [55].

The PDCD4 protein has been shown to be reduced following LPS stimulation at 24 h in RAW264.7 cells, mouse bone marrow–derived macrophages, and human peripheral blood monocytes, with the same dose as used in these experiments [25]. Knockdown of PDCD4 protein levels was shown with PROmiRNA‐21 in RAW264.7 cells with and without LPS stimulation. This is contrary to what we saw in our results. This may be due to differences in cell culture concentrations because we increased our cell concentration to increase the protein yield for Western blot samples. The difference in protein levels may be more subtle than we are able to detect.

SPRY proteins have not previously been directly linked to TLR‐4 signaling, although they have been shown to inhibit p38 in the MAPK pathway of interferon receptor signaling in mouse embryonic fibroblasts [56]. The experimental verification of SPRY proteins as targets of miRNA‐21 has occurred in other nonmacrophage cell types [32, 57]. SPRY proteins are induced by ERK activation [58]. The similarities between the abundance in both miRNA‐21^+/+ and miRNA‐21^−/− macrophage cell lines may be because the SPRY proteins do not have an important role in TLR‐4 signaling. Further experiments using either silencing RNA (small interfering RNA) or SPRY protein knockouts will add data to inform us as to whether SPRY proteins modulate TLR‐4 signaling.

We have corroborated our previous data that miRNA‐21 changes the macrophage cytokine profile in response to LPS stimulation.

Collectively our data suggest that TLR‐4 signaling has a definite role in the innate immune response to peritonitis, although the redundancy of multiple pathways stimulated in following induction of peritonitis with CLP may minimize the effects of miRNA‐21 in this model.

This study has shown that miRNA‐21 has a role in regulating the immune response by contributing to the switch from proinflammatory to anti‐inflammatory phase of the early innate response to LPS peritonitis. Its role in the CLP model is difficult to interpret because miRNA‐21 did not increase until 48 h after CLP, and little difference was observed in several outcome parameters. MiRNA‐21 could potentially have a very important role in sepsis because recent reports in the literature have postulated an early/acute hyperinflammatory and a late/chronic hypoinflammatory phase in response to sepsis [41, 59, 60]. This immunosuppression phase in late sepsis leads to mortality from persisting primary infection or an opportunistic new infection, otherwise known as the 2‐hit model. Some data suggest that altered myeloid maturation or differentiation may be important. Adoptive transfer of CD34⁺ hematopoietic stem‐progenitor cells after CLP improved survival by 65%. Cell transfer resulted in no change in the early hyperinflammatory response but corrected the later immunosuppression, increasing circulating proinflammatory cytokines, enhancing phagocytic activity, and clearing bacterial peritonitis. It is likely that these new cells differentiated into competent immune cells in the blood and tissue, reversing or replacing hyporesponsive endotoxin‐tolerant cells in the peritoneum [61]. There is initial data from McClure et al. [41] that miRNA‐21 is still elevated in this late stage of sepsis, suggesting that inhibition of miRNA‐21 to change to a more proinflammatory response may be potentially beneficial.

MiRNA‐21 is present in plasma and peritoneal fluid [18, 26, 62]. There are currently no data on the presence and pattern of exosomal or free miRNA‐21 in peritoneal fluid, in either the normal or infected abdomen. Isolation of miRNA‐21 from perioperative abdominal washings undergoing abscess drainage or drain fluid in postoperative surgical patients would improve our understanding of the course of miRNA‐21 secretion in response to clinical peritonitis. Future in vivo experiments could determine whether the addition of exogenous miRNA‐21 to the peritoneum during peritonitis would be able to modulate the local immune response. We hypothesized that the outcome of peritonitis can be modulated by miRNA‐21 through down‐regulation of the NF‐κB signaling that affects macrophage cytokine release, to diminish local inflammation. These findings demonstrate that miRNA‐21 is beneficial in modulating the macrophage response to LPS. The improved understanding of the anti‐inflammatory effects of miRNA‐21 in macrophage response to peritonitis may result in a targeted therapy to intervene in clinical peritonitis.

There are certain limitations to this work. The isolation process may have activated macrophages, although we tried to counteract any effects by using a gentle process and resting the cells for 24 h before use to alleviate some of these effects on subsequent stimulation. Bone marrow–derived macrophage cell lines were used to evaluate the absence of miRNA‐21; however, these macrophages are an immortalized cell line, which are not primary peritoneal macrophages. These cells are likely to be phenotypically or genetically different from primary peritoneal macrophages and, therefore, are only able to approximate what is occurring in the peritoneum. The evaluation of the individual cell type responses in vitro also need to be related to what in occurring in vivo because cells do not respond alone; multiple cells types communicate and respond to infectious stimuli together.

AUTHORSHIP

R.E.B. provided manuscript preparation, concepts, design, and performance. D.J.C. provided manuscript edit, concepts, design, data analysis, and performance. L.R. provided concepts, experimental design, performance of multiple assays, and data analysis. R.C.K. provided manuscript edit, experimental design, performance of multiple assays, and data analysis. V.R. performed multiple assays and provided data analysis. E.A.S. developed the experimental design, performed multiple assays, and analyzed data. S.S. provided concepts, experimental design, and data analysis. A.B. provided concepts, experimental design, and data analysis. W.G.C. provided concepts, experimental design, and manuscript preparation and edit.

DISCLOSURES

The authors declare no conflicts of interest.