Background: In order to modulate microglial phenotypes in vivo, M1 microglia were depleted by administration of gadolinium chloride and the expression of M2 microglia was induced by IL-4 administration in an animal model of sepsis to better characterize the role of microglial phenotypes in sepsis-induced brain dysfunction. Methods: Wistar rats were submitted to sham or cecal ligation and perforation (CLP) and treated with IL-4 or GdCl3. Animals were submitted to behavioral tests 10 days after surgery. In a separated cohort of animals at 24 h, 3 and 10 days after surgery, hippocampus was removed and cytokine levels, M1/M2 markers and CKIP-1 levels were determined. Results: Modulation of microglia by IL-4 and GdCl3 was associated with an improvement in long-term cognitive impairment. When treated with IL-4 and GdCl3, the reduction of pro-inflammatory cytokines was apparent in almost all analyzed time points. Additionally, CD11b and iNOS were increased after CLP at all time points, and both IL-4 and GdCl3 treatments were able to reverse this. There was a significant decrease in CD11b gene expression in the CLP+GdCl3 group. IL-4 treatment was able to decrease iNOS expression after sepsis. Furthermore, there was an increase of CKIP-1 in the hippocampus of GdCl3 and IL-4 treated animals 10 days after CLP induction. Conclusions: GdCl3 and IL-4 are able to manipulate microglial phenotype in an animal models of sepsis, by increasing the polarization toward an M2 phenotype IL-4 and GdCl3 treatment was associated with decreased brain inflammation and functional recovery.
Sepsis is a life-threatening condition that is often complicated by acute brain dysfunction and long-term cognitive impairment [1,2], is highly prevalent, and affects both early and late morbidity and mortality [3–5].
Studies have demonstrated that microglia have a number of physiological, noninflammatory functions that are crucial for central nervous system (CNS) functioning in the adult brain . Activated microglia are very plastic and may exist in multiple phenotypes that present different responses in accordance with changes in the cerebral microenvironment. They can induce brain repair, or cytotoxicity, as well as immune-regulatory or pro-inflammatory functions [7,8].
Classically, microglia phenotypes use the same M1 and M2 terminology as macrophage polarization in the literature. M1 macrophages not only release high levels of proinflammatory cytokines and free radicals that are crucial for killing microbes, but also induce damage in healthy neighboring tissue, contributing to cell death [9–11]. M2 macrophages display anti-inflammatory features and are involved in parasite containment as well as promoting tissue repair [9–11].
CKIP-1 (casein kinase 2 interacting protein-1) is a pleckstrin homology domain-containing protein, which plays important roles in the regulation of cell apoptosis, cell morphology, and cell differentiation [12–14]. Recently, CKIP-1 has been identified as a molecular toggle manipulating microglia polarization . CKIP-1 act as a scaffold protein, mediating interactions with multiple proteins, including CK2α, CPα, AP-1/c-Jun, Akt, ATM, IFP35/Nmi, and Smurf1 . CKIP-1 functions through different ways, such as plasma membrane recruitment, transcriptional activity modulation, and post-transcriptional modification regulation. As an adaptor protein, CKIP-1 is involved in various important signaling pathways, controlling cell growth, apoptosis, and differentiation .
Gadolinium chloride (GdCl3), a Kupffer cell inhibitor, is a rare earth metal that is taken up by liver macrophages [17,18] and when administered into the brain induces apoptosis of inflammatory microglia/macrophage via competitive inhibition of Ca2+ mobilization and damage to cell membranes [19–21]. In mice, TLR/MyD88 signaling may be involved in the M1 microglia/macrophage-induced necroptosis of astrocytes. Considering that GdCl3 also affects neutrophils [21,22], the beneficial effects of GdCl3 treatment may also be influenced by the inhibition of cells.
The cytokine most commonly used to induce M2 polarization is IL-4 . IL-4 can induce processes leading to potent anti-inflammatory functions, such as arginase-1 up-regulation, inhibition of NF-κB, and expression of phagocytosis receptors [24–26]. A single administration of IL-4 is sufficient to induce the expression of M2 markers in microglia as well as to promote the appearance of a macrophage subset that has a phenotype compatible with resolution-phase macrophages .
The modulation of microglial phenotype by IL-4 or GdCl3 can induce alternative M2 differentiation and inhibit the classical M1 pathway, respectively [9–11,28,29]. In order to modulate microglial phenotypes in vivo, we depleted M1 microglia by administration of GdCl3  and we induced the expression of M2 markers by IL-4 administration in an animal model of sepsis. We hypothesized that the in vivo modulation of microglial phenotypes improves brain inflammation and long-term cognitive outcomes in an animal model of severe sepsis.
GdCl3 (439770) and recombinant IL-4 (I3650) were purchased from Sigma-Aldrich.
Two-month-old Wistar rats were subjected to CLP as previously described . Briefly, animals were anesthetized using a mixture of ketamine (80 mg/kg) and xylazine (10 mg/kg) given intraperitoneally. Under aseptic conditions, a 3-cm midline laparotomy was performed to expose the cecum and adjoining intestine. The cecum was ligated with a 3.0 silk suture at its base, below the ileocecal valve and was perforated once with a 14-gauge needle. The cecum was then squeezed gently to extrude a small amount of feces through the perforation site. The cecum was then returned to the peritoneal cavity, and the laparotomy was closed with 4.0 silk sutures. Animals were resuscitated with regular saline (30 ml/kg) subcutaneously (s.c.) immediately and 12 h after CLP. All animals received antibiotics (ceftriaxone at 30 mg/kg) every 6 h s.c. for 3 days and dipyrone sodium (80 mg/kg) (s.c.) immediately and 12 h after CLP. To minimize variability between different experiments, the CLP procedure was always performed by the same investigator. We extensively characterized acute and long-term cognitive impairment and brain inflammation using this animal model [32–35].
Animals were submitted to sham or CLP and treated immediately after surgery with IL-4 or GdCl3. For IL-4, 1 μl saline or 100 ng of recombinant rat IL-4 (in 1 μl saline) was injected i.c.v (intracerebroventricular). For GdCl3, 20 μg per rat or the same volume of saline was injected i.c.v. Ten days after surgery, animals were submitted to behavioral tests. In a separated cohort of animals at 24 h, 3 and 10 days after surgery, animals were killed, and hippocampus were processed to different measures (see below).
Perfusion fixation of tissue for immunohistochemistry
Animals were anesthetized with ketamine and xylazine (30 and 10 mg/kg i.p., respectively) and were perfused with 0.9% sterile saline for 10 min (flow rate 20 ml/min) followed by 10 min with paraformaldehyde solution 4% in PBS (pH 7.4) (flow rate 20 ml/min). The brains were extracted and maintained in PFA 4% for 24 h, then placed in sucrose 15% for 24 h and placed in sucrose 30% for 24 h. Brains were slightly dried and frozen in −20°C.
To further characterize microglial phenotypes, immunohistochemistry (IH) was performed. Briefly, 40-μm sections from the total brain were incubated in 0.5% hydrogen peroxide in 0.1 M PBS containing 0.3% Triton-X100 to block endogenous peroxidase activity. After, sections were incubated with 2% bovine serum albumin to block non-specific protein binding. Sections were then incubated overnight at 4°C with a rabbit monoclonal IgG antibody against IBA-1 (1:4000; Abcam) or CD11b (1:4000; Abcam) or iNOS (1:100; Abcam), or IL-10 (1:200; Abcam), or arginase-1 (1:200; Invitrogen). Sections were incubated with biotinylated anti-rabbit IgG (1:100 dilution; Abcam) and after with 3,3′-diaminobenzidine (DAB) (Spring Bioscience). Sixteen random images per brain section were acquired at ×100 magnification, hippocampus region has been priorities, and the immune-positive area was expressed as a percent of total area analyzed. Positive controls were used according to the datasheet of each antibody.
Hippocampal dependent behavior tests
The inhibitory avoidance (IA) procedure was described in a previous report . The apparatus was an acrylic box (50 × 25 × 25 cm) whose floor consisted of parallel-caliber steel bars (1-mm diameter) spaced 1 cm apart and a platform that was 7 cm wide and 2.5 cm high. Animals were placed on the platform, and their latency to step down on the grid with all four paws was measured with an automatic device. Training sessions were performed 10 days after surgery. Immediately after, animals received a foot shock of 0.3 mA for 2s. In test sessions carried out 24 h after training, no foot shock was given, and the step-down latency (maximum of 180 s) was used as a measure of retention.
Behavior was also assessed in an open field (OF) apparatus to evaluate both locomotor and exploratory activities. The apparatus is a 40 cm × 60 cm open field surrounded by 50 cm high walls. The floor of the OF is divided into nine rectangles. The animals were placed on the left rear quadrant and left to explore the arena for 5 min (training session). Twenty-four hours after training session, animals were submitted again to open-field session (test session). The decrease in the number of crossings and rearings between the two sessions was taken as a measure of the retention of habituation memory .
Concentrations of tumor necrosis factor alpha (TNF-α), interleukin (IL)-1β and IL-6 in the hippocampus were determined using a commercially available sandwich ELISA kit (R&D Systems) in a microplate reader.
Analysis of M1 and M2 gene expression by real-time PCR (RT-qPCR)
Total RNA was isolated with Trizol® reagent (Invitrogen, U.S.A.) in accordance with the manufacturer’s instructions. The total RNA was quantified by spectrophotometry (A260/280 nm) and then treated with Deoxyribonuclease I (Invitrogen) to eliminate genomic DNA contamination. The cDNA was synthesized with ImProm-II™ Reverse Transcription System following the manufacturer’s instruction. Quantitative PCR was performed using SYBR®Green (Invitrogen) in a 7500 Fast Real-Time System Software v.2.0.5 (Applied Biosystems). Relative mRNA expression levels were determined using the target/GAPDH method. Primers are in Supplementary Material S1. Previous results from our laboratory demonstrated that the expression of these genes was different when compared with Sham and CLP animals [38,39]. Thus, based in the 3R's Principle a sham group was not included in gene analysis, and all comparisons were performed between CLP treated and not treated animals.
Recently, CKIP-1 has been identified as a molecular toggle manipulating microglia polarization. To perform immunoblotting of CKIP-1, tissue samples of hippocampus were homogenized in Laemmli buffer (62.5 mM Tris-HCl, pH 6.8, 1% (w/v) sodium dodecyl sulfate (SDS), 10% (v/v) glycerol) and equal amounts of protein (30 µg/well) were fractionated by polyacrylamide gel electrophoresis-sodium dodecyl sulfate (SDS-PAGE) and electro-transferred to nitrocellulose membranes. The efficiency of the electro-transfer was verified by Ponceau staining, and the membrane was then blocked in Tris-Tween buffer saline (TTBS: 100 mM Tris-HCl, pH 7.5, containing 0.9% NaCl and 0.1% Tween 20) with 5% albumin. The membranes were incubated overnight at 4°C with rabbit monoclonal anti-CKIP-1 (1:1000). Secondary anti-rabbit IgG was incubated with the membrane for 2 h (1:10,000). The immunoreactivity was detected by chemiluminescence using ECL. Densitometry analysis of the films was performed using the ImageJ v.1.34 software. All results were expressed as a relative ratio between anti-CKIP-1 and GAPDH.
Molecular data were analyzed with Expression Suite Software expressed as target/GAPDH and analyzed by two-way ANOVA, followed by Tukey’s post hoc test. For IH, the immune-positive area was expressed as a percent of total area. Biochemical data were analyzed by one-way ANOVA, followed by Tukey’s post hoc test. Data from the inhibitory avoidance task were reported as median and interquartile ranges, and comparisons among groups were performed using Mann–Whitney U tests. All tests were analyzed with SPSS or GraphPad Prism 4.0. Software. In all comparisons P < 0.05 indicated statistical significance. Immunohistochemistry images were analyzed by ImageJ v1.34, U.S.A.
To test the hypothesis that modulating M1/M2 polarization determines behavioral outcomes in sepsis survivors, animals were submitted to sepsis and treated with IL-4 or GdCl3 during the acute phase of sepsis development, and the cognitive function in sepsis survivors was determined 10 days after surgery.
In the test section of the IA, there was a significant increase in latency time for the sham-treated group, but not septic animals, and this was prevented by IL-4 and GdCl3 treatment (Figure 1A). In addition, the number of crossings and rearings was decreased in the test section in sham animals, which was prevented by IL-4 and GdCl3 treatment (Figure 1B). These results indicate that microglia activation and polarization are important factors implicated in the long-term cognitive impairment associated with sepsis and this can be improved by IL-4 and GdCl3 treatment.
Effect of IL-4 and GdCl3 treatment on long-term cognitive impairment after sepsis
Cytokines levels were measured at 24 h, 3 and 10 days after sepsis in hippocampus in animals submitted to sepsis and treated with IL-4 or GdCl3 (Figure 2). All cytokines increased at all times, except for IL-6 at 10 days, when compared with sham and CLP animals. IL-4 and GdCl3 treatment was able to decrease TNF levels when compared with CLP animals at all times (Figure 2A). This was different when analyzing IL-1 levels, since IL-4 did not decrease IL-1 levels at 24 h and 3 days (Figure 2B). A different behavior was also noted to IL-6 levels (Figure 2C). At 24 h both treatments decreased IL-6 levels when compared with CLP group. At 3 days, only IL-4 decreased IL-6 levels, and, interestingly, at 10 days GdCl3 increased IL-6 levels (Figure 2C). To further characterize brain inflammation after sepsis, it was determined microglial activation by IBA-1 (Figure 3). IBA-1+ cells were highly expressed in hippocampus of CLP animals at all time points when compared with sham animals. IL-4 and GdCl3 treatments reduced IBA-1+ cells at 3 and 10 days after CLP induction.
Effect of IL-4 and GdCl3 treatment on brain cytokine levels after sepsis
Effect of IL-4 and GdCl3 treatment on microglia activation after sepsis
To better characterize the effect of IL-4 and GdCl3 on microglia phenotype the expression of CD11b, iNOS (M1 markers), arginase-1 and IL-10 (M2 markers) were measured (Figure 4 and Supplementary Material S2). CD11b and iNOS were increased after CLP at all time points, and IL-4 and GdCl3 treatments were able to reverse this at all time points (Figure 4). Arginase-1 decreased in the CLP group at all time points, and both IL-4 and GdCl3 treatment reverted this only at 10 days after sepsis (Figure 4). Moreover, IL-10 did not change over time in any group (Figure 4). Additionally, gene expression of iNOS was decreased by IL-4 treatment at 3 and 10 days (Supplementary Material S2). On the other hand, GdCl3 treatment only decreased CD11b gene expression at 10 days (Supplementary Material S2). IL-4 increased gene expression of arginase-1 only at 10 days (Supplementary Material S2) and GdCl3 increased arginase-1 at 3 and 10 days and IL-10 at 10 days after sepsis induction (Supplementary Material S2).
Effect of IL-4 and GdCl3 treatment on microglia phenotypes after sepsis
To identify one possible mechanism that drives the effects of IL-4 and GdCl3 on microglia phenotype, it was measured the levels of CKIP-1 (Supplementary Material S3). There was an increase in CKIP-1 after GdCl3 and IL-4 treatment 10 days after CLP induction (see Supplementary Material S3).
It was here demonstrated that the up-regulation of brain inflammation that persists up to 10 days after sepsis could be attenuated by IL-4 and GdCl3 treatment, probably by their effect on microglia phenotype, and this was related to improvement of long-term cognitive dysfunction.
After activation, microglia can acquire different phenotypes that may contribute to neuroinflammation. Although little is known about the dynamics of microglial polarization in sepsis , modulation of microglial function can mitigate sepsis-related brain injuries, thus promoting tissue repair and functional recovery. In this context, IL-4 and GdCl3 were used here trying to reach this objective.
IL-4 is a pleiotropic regulator of numerous immune/inflammatory pathways [41,42]. Recognized initially as a product of activated T cells stimulating B-cell proliferation [43,44], it has become evident that IL-4 is a central initiator and effector of Th2 immune responses, M2 macrophage polarization, and IgE secretion [44–46]. The results of the present study show an important role for IL-4 in the recovery of brain damage caused by sepsis. The activity of IL-4 in the brain has recently received considerable attention in relation with its anti-inflammatory and pro-resolution effects that are directly mediated by microglia. Expression and secretion of IL-4 are induced by neuroinflammatory insults in activated immune cells and damaged neurons, and IL-4-activated microglia produce a series of mediators, including neurotrophic and matrix remodeling, factors, and proteolytic enzymes that help reduce neuroinflammation and promote tissue repair, as shown in experimental models of neurodegenerative diseases [27,47–49]. IL-4 administration represents a valuable experimental procedure to study the signaling pathways that control M2 polarization in microglia . Activation of the M2 phenotype in the brain has been previously observed in animal models of neurological diseases and associated with suppression of inflammation, tissue remodeling and matrix deposition in a time- and environment-specific manner [51,52]. M2 responses observed following central administration of IL-4 represent advantageous features in estimating the capability of microglia cells to acquire the M2 phenotype among and within different brain areas, in physiological as well as pathological conditions .
Recent studies have revealed that pre-treatment of rodents with gadolinium chloride (GdCl3, an inhibitor of Kupffer cells) prevents LPS-induced mortality, liver, and lung injuries and attenuates pro-inflammatory cytokine production and increased production of anti-inflammatory cytokines in these tissues [53,54]. Here, it was demonstrated that a single dose of GdCl3 decreased production of pro inflammatory cytokines, besides decreasing the expression of markers M1 and increased M2 markers mainly at 10 days after sepsis. Strande et al.  showed that GdCl3 treatment decreased circulating monocytes and neutrophils, and leukocyte infiltration into injured myocardium, and this seems be the case in the brain as well. Evidence suggests that GdCl3 interferes with macrophage and neutrophil function in the liver by decreasing macrophage secretion of inflammatory cytokines and toxic oxygen radicals  and by inhibiting neutrophil infiltration . Kishta and cols  showed that the pretreatment with GdCl3 significantly attenuated LPS-induced neutrophil infiltration, and pulmonary cell apoptosis. GdCl3 is a lanthanide that is commonly used to evaluate the functional roles of liver macrophages in several processes including LPS-, ozone-, and hyperoxia-induced damage [58–60]. While the exact mechanisms of action of GdCl3 are not yet clear, it has been proposed that it inhibits Mϕ phagocytosis by competitive blockade of K-type Ca2+ channels  and by reducing cellular expression of cytochrome P450s, thereby interfering with Mϕ metabolism .
Several authors have already reported [34,35,63,64] an increase in cytokines levels after CLP in the hippocampus, both early (24 h) and late (3 and 10 days) after sepsis induction, and this is consistent with our findings. When animals were treated with IL-4 and GdCl3, the reduction of pro-inflammatory cytokines was apparent at almost all time points after sepsis. Cytokines are one of main factors that regulate the functional phenotype of microglia and macrophages [9,65]. These data demonstrate that a single acute administration of IL-4 after sepsis drives some aspects of M2 polarization in microglia, and treatment with GdCl3 may improve brain function probably via apoptosis of infiltrating monocytes and activated microglia. From a mechanistic point of view, several different aspects of brain function could improve by decreasing brain inflammation. Using minocycline to decrease microglia activation, it was demonstrated an improvement of blood–brain barrier function , an important step involved in brain dysfunction during sepsis . Despite the fact that it was not measured blood–brain barrier function in the present study, minocycline decreased microglial M1 phenotype during sepsis evolution, supporting this hypothesis . Additionally, neutralizing IL-1 could improve long-term cognitive impairment by improving blood–brain barrier function, decreasing inflammation and preventing synaptic deficit observed in sepsis survivors [68,69].
Recently, CKIP-1 has been identified as a molecular toggle manipulating microglia polarization . Here, it was demonstrated an increase in its protein level 10 days after sepsis induction. An increase in CKIP-1 levels inhibited the expression of proinflammatory cytokines, reduced brain edema, and improved neurological outcomes in an animal model of intracerebral hemorrhage . CKIP-1 is induced by M1 stimuli such as LPS and interferon and suppressed by M2 stimuli such as IL-4 , and this is in accordance with our findings. Future studies should focus in the understanding of the role of CKIP-1 modulation in brain dysfunction associated with sepsis.
Hippocampus is one of brain structures more sensitive to sepsis-induced dysfunction [71,72]. In this context, we here measured two different cognitive tasks: IA and OF. IA is a classical fear memory hippocampus-dependent cognitive task . In this paradigm, plasticity in the hippocampus is fundamental to memory formation , besides other structures such as the amygdala complex works in parallel to the hippocampus . The hippocampal CA1 region is central in memory formation receiving sensorial information from reticular–entorhinal connections or medial–reticular septal connections that reach the dentate gyrus (DG) and are transmitted to the CA3 and CA1 region . Furthermore, OF is also dependent on hippocampus , but differently from IA that does not involves fear, but habituation to a novel environment, and this reflects in important biochemical differences in memory consolidation processes of the two tasks . Some other cognitive task could also be used to assess hippocampal function, such as the novel object task or the Barnes maze analysis, which were not evaluated here. We had previously demonstrated that sepsis survivors presented deficits in the novel object task , and strategies aimed to decrease hippocampus inflammation improves the performance in this task , reinforcing the results presented here.
Some limitations should be kept in mind when interpreting our results. First, we can not ascertain that IL-4 and GdCl3 would act only in microglia, since they can have also an effect on infiltrating monocytes and macrophages, and other cell types even when administered to the brain. Second, IL-4 and GdCl3 were ICV administrated and this limits the clinical translation of our results. However, from a mechanistic point of view it decreases the probability of a systemic, non-specific effect of the treatment.
These data show that a single acute administration of IL-4 after sepsis drives some aspects of M2 polarization in microglia, and treatment with GdCl3 may improve brain function by decreasing M1 phenotype. GdCl3 and IL-4 are important modulators of microglial phenotype, increasing the expression of M2 markers at least by CKIP-1 activation, and the changes in microglial phenotype triggered by IL-4 and GdCl3 treatment are associated with improved tissue protection and functional recovery.
The modulation of microglial phenotype improves outcomes in some models of primary brain injury, but the effect of this strategy on sepsis-associated brain dysfunction is not known.
The in vivo modulation of microglial phenotypes with interleukin-4 and gadolinium chloride improves brain inflammation and long-term cognitive outcomes in an animal model of severe sepsis.
The pharmacological modulation of microglia activation and phenotype could be of potential interest to decrease the disease burden associated with sepsis-induced brain dysfunction.
The authors declare that there are no competing interests associated with the manuscript.
CAPES, CNPQ and UNESC
M.M. contributed to experimental planning, data collection, data analysis, and writing of paper; M.A. contributed to experimental planning and data analysis; A.V. contributed to experimental planning and data analysis; P.A. contributed to experimental planning and data analysis; A.I.G. contributed to data analysis; H.B. contributed to data analysis; E.C. contributed to experimental planning and data analysis; D.D. contributed to experimental planning, and data analysis; T.B. contributed to experimental planning and writing of paper; F.D-P. contributed to experimental planning, data analysis and writing of paper.
The experimental procedures involving animals were performed in accordance with Brazilian law of animal welfare and with the approval of our institutional ethics committee (protocol number: 036/2016-2.) and were performed in Laboratory of Experimental Pathophysiology, University of Southern Santa Catarina.