Targeted intracerebral delivery of the anti-inflammatory cytokine IL13 promotes alternative activation of both microglia and macrophages after stroke

Wild type C57BL/6J mice were obtained via Charles River Laboratories (strain code 027) and were used for in vivo detection of IL13 mRNA produced by grafted IL13-MSCs. In total, 24 mice were included in the experiment, divided over the following three groups: (i) a non-injected control mice (n = 8), (ii) MSC-grafted mice (n = 8), and (iii) IL13-MSC-grafted mice (n = 8).

Transgenic CX₃CR1^eGFP/eGFP mice (strain code 005582) and CCR2^RFP/RFP mice (strain code 017586) were obtained via Jackson Laboratories. CX₃CR1^eGFP/+ CCR2^RFP/+ mice were obtained by breeding CX₃CR1^eGFP/eGFP mice with CCR2^RFP/RFP mice. During the entire study, mice were kept in the animalarium of the University of Antwerp (UA) under normal day-night cycle (12/12) with free access to food and water. All animal experimental procedures were approved by the Ethics Committee for Animal Experiments of the UA (Approval No 2015-84 and 2018-36).

Ischemic stroke was induced in 32 male CX₃CR1^eGFP/+ CCR2^RFP/+ mice (11–13 weeks, 22–26 g) by middle cerebral artery occlusion (MCAO), as described below. Stroke lesions were checked by magnetic resonance imaging (MRI) scans 48 h after MCAO, before transplantation of mesenchymal stem cells (MSCs). Two mice died after stroke induction (stroke-only group), one mouse before stroke during anesthesia. Four mice were excluded due to too small or absent stroke lesions. Eight mice died after cell implantation (n = 3 for MSC group; n = 5 for IL13-MSC group). In total, 17 mice were included in the study, resulting in the following three groups: (i) control group subjected to stroke only (n = 6), (ii) MSC-treated group, which received MSC transplantation 2 days after MCAO (n = 5), and (iii) IL13-MSC-treated group, which received a transplantation of IL13 producing MSCs 2 days after MCAO (n = 6). All mice were perfused at day 14 after stroke induction. An overview of the experimental protocol is presented in Fig. 1.

In this study, we used a previously established and characterized C57BL/6 mouse bone marrow-derived MSC line [22] and a derivative thereof, genetically engineered to express murine interleukin-13 (further named as IL13-MSCs) [23]. The latter MSC line was generated by transduction of parental MSCs with the pCHMWS-IL13-IRES-Pac lentiviral vector, according to previously optimized procedures [24, 25]. For expansion, both MSC lines were cultured in standard cell culture plastic ware in “complete expansion medium” [24] consisting of Iscove’s modified Dulbecco’s medium (IMDM; Lonza) supplemented with 8% fetal bovine serum, 8% horse serum, 200 U/mL + 100 μg/mL penicillin/streptomycin, and 1 μg/mL amphotericin B (all products from Thermo Fisher Scientific). Culture medium for IL13-MSCs was further supplemented with 5 μg/mL puromycin (InvivoGen). MSC cultures were split 1:5 twice a week using 0.05% trypsin-EDTA (Thermo Fisher Scientific) for cell detachment.

Wild type MSCs and IL13-MSCs were plated at a concentration of 2 × 10⁵/well in a six-well plate and allowed to adhere during overnight incubation. After the following 24 h of culture, supernatant was harvested and analyzed for the presence of IL13 protein by means of ELISA, according to the manufacturer’s instructions (Peprotech).

All surgical experiments were performed under sterile conditions, as previously described [26‐28]. Briefly, mice were anesthetized by an intraperitoneal injection of a ketamine (80 mg/kg, Pfizer) + xylazine (16 mg/kg, Bayer Health Care) mixture in 0.9% NaCl solution (Baxter) and placed in a stereotactic frame (Stoelting). A midline scalp incision was made and a hole was drilled in the skull using a dental burr drill (Stoelting). Stereotactic coordinates were as follows (relative to bregma): AP 0 mm, Lat. 2.3 mm and − 2.3 mm, and DV − 2.3 mm. Next, an automatic microinjector pump (kdScientific) with a 10 μl Hamilton Syringe was positioned above the exposed dura. A 30-gauge needle (Hamilton), attached to the syringe, was lowered through the intact dura and positioned at the respective depth. After 2 min of tissue pressure equilibration, a suspension of 5 × 10⁴ MSCs in a volume of 0.4 μl was injected. Following 4 min to allow pressure equilibration and to prevent backflow of the injected cell suspension, the needle was fully retracted. The exact same procedure was performed at both left and right side of the brain. The skin was sutured (Vicryl, Ethicon), and 100 μl of a 0.9% NaCl solution was administered subcutaneously in order to prevent dehydration while mice were placed under a heating lamp to recover.

Focal cerebral ischemia was induced by transient occlusion of the right middle cerebral artery (MCA) in all animals, using the intraluminal filament model as described previously [29, 30]. Briefly, mice were anesthetized with 1–2% isoflurane in a gas mixture of O₂/N₂O (30:70%) and received a subcutaneous injection of 4.0 mg/kg Carprofen (Pfizer, Karlsruhe, Germany) for analgesia after the surgical interventions. With a neck incision, the common carotid artery (CCA) and its proximal branches were exposed. The internal carotid artery (ICA) was occluded for a short time with a metal microvessel clip. A silicon rubber-coated filament with a tip diameter of 170 μm (7017PK5Re, Doccol Corporation, Sharon, MA, USA) was inserted into the ICA. The filament was advanced through the ICA until it blocked the blood flow to the middle cerebral artery. Animals were allowed to recover during the 30 min occlusion in a temperature stable box (MediHeat, Peco Services Ltd., Brough, UK). Afterwards, the animals were re-anesthetized, the filament was carefully removed to initiate reperfusion and the CCA was permanently ligated. During 1 week after the surgery, body weight was daily monitored in all animals. Animals were randomly assigned to the three different experimental groups.

All surgical experiments were performed as described earlier under section cell transplantation in healthy mouse brain. The needle was positioned at a depth of 2.5 mm into the striatum close to the ischemic lesion (as determined following MRI analysis). First, a suspension of 2 × 10⁴ MSCs in a volume of 0.4 μl was injected. Following 4 min of pressure equilibration, the needle was retracted to a depth of 1.0 mm and a second cell infusion (2 × 10⁴ MSCs in a volume of 0.4 μl) was performed.

To observe the different aspects of neurological functions, a modified neurological deficit score (mNDS) was performed before and every 2 days after MCAO, based on a modification of a previous report [31]. In brief, the modified NDS consists of a set of motor tests (muscle status and abnormal movement), sensory tests (tactile, and proprioceptive), and reflex tests on a scale of 0–18. Increasing score indicates the severity of the stroke damage [32].

At 2 days post induction of stroke, mice were subjected to MRI using a 7T Pharmascan MR scanner with a 16-cm-diameter horizontal bore (Bruker, Ettlingen, Germany). This system is equipped with a standard Bruker cross-coil setup, using a quadrature volume coil for excitation and an array mouse surface coil for signal detection. The system was interfaced to a Linux PC running Topspin 2.0 and Paravision 5.1 software (Bruker). Mice anesthesia was induced using 2% isoflurane (Forane Abbott, UK) in a gas mixture of 30% O₂ and 70% N₂ at a flow rate of 600 ml/min. During MRI acquisition, isoflurane concentration was set at 2%, and the respiration rate was continuously monitored using a pressure sensitive pad. In addition, body temperature was monitored via a rectal probe and was held constant between 37.0 and 37.3 °C using warm air coupled to a feedback unit (SA instruments, NY, USA). Both respiration and body temperature control systems were controlled by PC-Sam monitoring software (SA instruments, NY, USA). Following MR image acquisition, mice were left to recover separately under a heating lamp before returning to their respective cages. Following scout scans, we acquired an axial proton density weighted Rapid Acquisition and Relaxation Enhancement (RARE) image using the following parameters: repetition time (TR) = 3000 ms, echo time (TE) = 13.3 ms, matrix size (256 × 256), field of view (FOV) = (17.5 × 17.5) mm², resolution = (0.068 × 0.068) mm², 10 coronal slices, slice thickness = 0.8 mm, RARE factor = 8, and averages = 2. Next, for determination of the T2 values, we applied a multi-slice, multi-echo sequence based on the Carr-Purcell-Meiboom-Gill sequence. The following parameters were used: TR = 3000 ms, number of echoes = 10, echo spacing = 11 ms, matrix size (128 × 128), field of view (FOV) = (17.5 × 17.5) mm², resolution = (0.137 × 0.137) mm², 10 slices, and slice thickness = 0.8 mm.

All immunofluorescence analyses were performed according to previously described procedures [26, 28]. Briefly, mice were transcardially perfused with 0.9% NaCl solution followed by 4% paraformaldehyde (PFA). Next, brains were isolated and post-fixed with 4% PFA for 3 h, then dehydrated through a sucrose gradient of 5, 10, and 20%. Brains were snap-frozen in liquid nitrogen and kept at − 80 °C until further processing. Ten-micrometer-thick cryosections were made using a microm HM500 cryostat. Immunofluorescent staining was performed on brain slides using the following antibody combinations: a primary rabbit anti-RFP antibody (2.5 μg/ml, Abcam, ab62341) with a secondary donkey anti-rabbit Alexa Fluor 555 antibody (2 μg/ml, Thermo Fisher Scientific, A31572), a primary rat anti-F4/80 antibody (4 μg/ml, Bio-Rad, MCA497R) with a secondary goat anti-rat Cy5 antibody (10 μg/ml, Thermo Fisher Scientific, A10525), a primary rat anti-MHC-II antibody (2.5 μg/ml, eBioscience, 14–5321-82) with a secondary goat anti-rat Alexa Fluor 350 antibody (10 μg/ml, Thermo Fisher Scientific, A21093), and a primary goat anti-Arg1 antibody (4 μg/ml, Santa Cruz, sc-18354) with a secondary donkey anti-goat Alexa Fluor 350 (10 μg/ml, Thermo Fisher Scientific, A21081) and a primary rabbit anti-Ym1 antibody (1:50 dilution, StemCell Technologies 01404) with a secondary goat anti-rabbit Alexa Fluor 647 antibody (10 g/ml, Thermo Fisher Scientific, A21245). Before mounting with Prolong Gold antifade reagent, nuclear staining was performed using TOPRO-3 (1/200, Thermo Fisher Scientific). Images were acquired using a BX51 fluorescence microscope equipped with an Olympus DP71 digital camera. Olympus cellSens dimension software was used for image processing. In each brain section, five different regions of interest (ROIs) were selected with the constant exposure time: the border and the core region of the ischemic hemisphere in the striatum, the core region of the ischemic hemisphere in the cortex, and two ROIs in the cortex and striatum of the intact contralateral hemisphere.

For quantitative phenotypic analysis of microglia and macrophage activities after stroke, immunofluorescence images were analyzed using NIH ImageJ analysis software (ImageJ) and TissueQuest 4.0 (TissueGnostics, Vienna, Austria). An entire picture of all ROIs, taken at × 20 magnification, was used for quantification. Cells were recognized based on the nuclei staining (nuclei size, staining intensity, and discrimination by area was optimized manually), followed by the analysis of specific staining. The cellular densities of eGFP⁺RFP⁻ microglia (CX₃CR1^eGFP/+), eGFP⁻RFP⁺ macrophages (CCR2^RFP/+), and eGFP⁺RFP⁺ double-positive microglia/macrophages (CX₃CR1^eGFP/+ CCR2^RFP/+) were quantified. To achieve optimal cell detection, the background threshold was defined. To discriminate false signals due to the coverage of CX₃CR1^eGFP/+ cell ramification with total nuclei intensity, the cut-off was defined for each ROI. Scattergrams were generated to visualize the corresponding positive cells in the source ROIs through the real-time back gating component. Mean intensity and the relative number of the co-expressed CX₃CR1^eGFP/+ or CCR2^RFP/+ with pro- or anti-inflammatory markers F4/80, Arg-1, and MHC-II were obtained. Afterwards, the mean values were determined from analyses of at least three brain sections per mouse. Based on the above cell density analysis, the proportion of microglia or macrophages expressing F4/80, Arg-1, or MHC-II was calculated. To observe anti- and pro-inflammatory representative markers in double staining with eGFP⁺ or RFP⁺, the above-defined five different ROIs provided the following number of cells/mm³: (i) total cell density according to the nuclei staining of TOPRO-3, (ii) eGFP⁺ microglia cell density, (iii) RFP⁺ macrophage cell density, (iv) eGFP⁺RFP⁺ double-positive microglia/macrophages cell density, (v) F4/80⁺/eGFP⁺ or RFP⁺ cell density, (vi) Arg-1⁺/eGFP⁺ or RFP⁺ cell density, and (vii) MHC-II⁺/eGFP⁺ or RFP⁺ cell density.

For statistical analyses of qRT-PCR and ELISA data, GraphPad Prism software was used. For all other experiments, statistical analyses were performed with SPSS version 22 (IBMSPSS statistics, Ehningen, Germany). The Normality test and homogeneity of variances were assessed for all data. The nonparametric analysis approach, Kruskal-Wallis H, was performed for analysis of behavioral scores (mNDS). Immunohistochemistry (IHC) and qRT-PCR data from in vivo experiments were analyzed for significant changes between the three groups using one-way analysis of variance (ANOVA) with Bonferroni-corrected post hoc comparisons. qRT-PCR data and ELISA data for comparison of in vitro cultured MSCs and IL13-MSCs unpaired t tests were performed. Differences were considered statistically significant at a p value < 0.05. Data shown represent mean value per group ± standard deviation.

In our previous studies related to in vivo grafting of IL13-MSCs, we assumed transgenic IL13 mRNA (and protein) to be effectively produced as determined by the appearance of alternatively activated microglia and macrophages [23, 33, 34]. In this first part of our study, we aimed to determine whether IL13-MSCs effectively produce transgenic IL13 mRNA after in vivo transplantation. At first, results obtained from cultured MSCs and IL13-MSCs demonstrate that both mRNA and protein expression levels of IL13 are only detected in IL13-MSCs and not in MSCs (respectively; p < 0.0001 and p = 0.0003) (Fig. 2a, b). Next, following transplantation of MSCs and IL13-MSCs in healthy mouse brain, we were able to detect significant levels of transgenic IL13 mRNA expression in five out of eight IL13-MSC-injected mice at day 7 post-implantation, as compared to control non-injected (p = 0.0216) and MSC-injected mice (p = 0.0214) (Fig. 2c). Overall, these data demonstrate that IL13-MSC can effectively produce transgenic IL13 mRNA in situ following grafting in the CNS.

Occlusion of the right middle cerebral artery with a silicone rubber-coated filament resulted in cerebral ischemia. Lesion size and location were assessed by MRI using quantitative T2 maps. Ischemic territory shows up as area of increased T2 values on T2 maps, representing the area with augmented tissue water content. In the hemisphere ipsilateral to the lesion, vasogenic edema increased T2 values and a clear lesion was detectable in the right ischemic hemisphere on T2 maps. In the intact contralateral hemisphere, T2 maps showed no signs of lesion. Two days after MCAO, two characteristic lesion types were detected: (1) in 5 mice, lesions were restricted to the striatum and (2) in 12 mice, lesions involved both striatum and cortex. Subsequently, mice were divided over three experimental groups: (1) control mice with untreated stroke, (2) mice with stroke receiving a control MSC graft, and (3) mice with stroke receiving an IL13-MSC graft. Representative lesions in each experimental group are displayed in Fig. 3.

We determined whether the procedure of cell grafting (MSC or IL13-MSC) into the ischemic brain influences the motor and sensory performance after stroke in mice. In all three experimental groups, the neurological deficit scores increased at day 2 after MCAO (up to score 7 in mNDS) compared to baseline (score 0 in mNDS), indicating the impaired motor and sensory functions due to the stroke. No differences were observed between the three experimental groups within the two-week observation period. Nevertheless, our findings clearly indicate that the additional surgical procedure, intracranial MSC transplantation at 2 days after MCAO, did not worsen the neurological deficits after stroke (Additional file 1: Figure S1).

In order to histologically investigate brain trafficking of recruited microglia and macrophages after cerebral ischemia, we took advantage of CX₃CR1^eGFP/+ CCR2^RFP/+ double transgenic mice. eGFP and RFP expression in the brain sections of these mice enables us to simultaneously track resident eGFP⁺ microglia (CX₃CR1^eGFP/+) and infiltrating RFP⁺ monocytes/macrophages (CCR2^RFP/+). Using these transgenic mice, we observed differences in displacement of migration between microglia and peripheral monocyte-derived macrophages. Histological analysis revealed a noticeable infiltration of both microglia and macrophages throughout the ischemic territory at day 14 after stroke induction, (Fig. 4a–c). From the representative immunofluorescence images in Fig. 4b, we noted that eGFP and RFP expression shows a spatially different distribution in the cortex and the striatum. In the cortical part of the ischemic lesion, CCR2^RFP/+ macrophages were located in the infarct core and CX₃CR1^eGFP/+ microglia were mainly accumulated at the infarct border. In contrast, in the striatal lesion area, CX₃CR1^eGFP/+ microglia and CCR2^RFP/+ macrophages were diffusely distributed across the whole territory. The only population within the intact contralateral hemisphere contained resident CX₃CR1^eGFP/+ microglia (Fig. 4c). The microglia density on the ischemic hemisphere was significantly higher than on the contralateral hemisphere, in all groups and analysis (Fig. 4d–f). No obvious difference in this distribution pattern was observed between the three experimental groups.

Next, we assessed whether any differences in cell density existed between resident and infiltrating cells 2 weeks after stroke. When evaluating the whole lesion site (i.e., the cortical and/or striatal region), the vast majority of recruited cells within the lesion are CX₃CR1^eGFP/+ microglia in all three experimental groups. Statistical analysis confirmed that CX₃CR1^eGFP/+ cell density, being of microglial origin, is significantly higher than CCR2^RFP/+ cell density, being of macrophage origin, in all three experimental groups at day 14 after stroke (stroke-only group p = 0.013, MSC-treated group p = 0.038, IL13-MSC-treated group p = 0.017) (Fig. 4d). When analyzing the cell density of microglia and macrophages, separately for striatal and for cortico-striatal lesions, the statistical significance between both cell populations was preserved in all cases, except for the two MSC groups in the cortical analysis (Fig. 4e), where only a trend was reached (MSC group p = 0.06; IL13-MSC group p = 0.063). Despite the difference in spatial distribution of microglia and macrophages between the striatal and cortical lesion area, further quantitative histological analysis revealed closely comparable numbers of infiltrating macrophages and of resident microglia between cortical (Fig. 4e) and striatal (Fig. 4f) lesions over the three experimental groups.

CX₃CR1^eGFP/+ microglia are visible in both the intact and the ischemic hemisphere, but are found in different numbers, distribution patterns (see above), and morphological appearance. In this study, we could observe at least three distinct morphological appearances of CX₃CR1^eGFP/+ microglia at day 14 after stroke induction (Fig. 5, left column): (i) cells with long thin arborized processes and a small cell body, defined as ramified cells (top row), (ii) cells with swollen processes and elongated amorphous larger cell body, classified as intermediate cells (middle row), and (iii) cells with round shape and no plasmalemmal processes, specified as amoeboid cells (bottom row). Ramified CX₃CR1^eGFP/+ cells were mostly observed in the healthy periphery of the lesion and in the contralateral hemisphere, but were not seen in the core of the lesion. Intermediate type CX₃CR1^eGFP/+ cells were closely associated with peri-infarct regions. Amoeboid CX₃CR1^eGFP/+ cells were mainly localized in the ischemic core. In contrast, recruited CCR2^RFP/+ macrophages displayed a uniform oval, kidney form, or round shape without apparent processes (Fig. 5, central column). Some CCR2⁺ cells showed cellular processes, suggesting that these cells may acquire an intermediate phenotype in the damaged tissue. Examples in Fig. 5 were taken from the stroke-only group, but no obvious differences between the appearance of microglia and macrophage morphology was observed between the three experimental groups.

Not only the recruitment of microglia and macrophages is a key feature of neuroinflammation, but also the subsequent phenotypic alterations associated with their activation status. Here, we first investigated the appearance of F4/80 expression, a general marker of activation, on microglia and macrophages following stroke (Fig. 5, right column). F4/80 expression was not detected contralateral to the lesion site (top row), while both microglia and macrophage populations (central and bottom row) within the lesion area, at least in part, express this F4/80 activation marker. Notably, morphological inspection of F4/80-expressing cells shows that these cells display an intermediate or amoeboid shape. Next, we assessed whether microglia and/or macrophage activation, based on F4/80 expression, was altered upon transplantation of IL13-producing MSC, as compared to the situation after transplantation of MSC alone or without any transplantation after stroke induction (Fig. 6). As shown by the representative immunofluorescent images in Fig. 6a, and further supported by the quantitative analysis provided in Fig. 6b (relative proportion of F4/80 expression, stacked bars) and Fig. 5c (absolute proportion of F4/80 expression, pie charts), F4/80 expression was observed on both microglia and macrophages in all three experimental groups. Interestingly, we observed a significantly increased expression of F4/80 by CCR2^RFP/+ macrophages upon IL13-MSC transplantation as compared to F4/80 expression by CCR2^RFP/+ macrophages following stroke or stroke + control MSC grafting (p = 0.008 and p = 0.007). The increased number of F4/80⁺ microglia detected following IL13-MSC grafting was not significantly different from other groups.

To further analyze the phenotype of recruited microglia and macrophages within the ischemic stroke lesion, as well as the effect of IL13-MSC thereon, we first performed additional immunofluorescence staining for Arg-1. Notably, we wanted to investigate whether IL13-producing MSC were able to convert stroke-associated neuroinflammatory immune responses into an alternatively activated inflammatory response. As shown by the representative immunofluorescent images in Fig. 7a, and further supported by the quantitative analysis provided in Fig. 7b (relative proportion of Arg1 expression, stacked bars) and Fig. 6c (absolute proportion of Arg1 expression, pie charts), significant increase in Arg1 expression was observed within and surrounding the lesion area on both microglia and macrophages following grafting of IL13-MSC (for microglia, stroke + IL13-MSC vs. stroke + MSC and stroke only, p = 0.002 and p = 0.005 respectively; for macrophages, stroke + IL13-MSC vs. stroke + MSC and stroke only, both p < 0.001). Furthermore, the induction of this alternative activation program, as characterized by Arg1 expression, was significantly higher in macrophages than in microglia in the IL13-MSC group (p = 0.002).

Likewise, we investigated the degree of MHC-II expression within and surrounding the lesion area over the different experimental conditions. As shown by the representative immunofluorescent images in Fig. 8a, MHC-II expression is observed mainly inside the ischemic lesion in all three experimental groups at day 14 after stroke induction. Further quantitative analyses provided in Fig. 8b (relative proportion of MHC-II expression, stacked bars) and 7C (absolute proportion of MHC-II expression, pie charts), indicated that MHC-II expression is significantly higher in CCR2^RFP/+ macrophages, compared to CX₃CR1^eGFP/+ microglia (p = 0.008, p = 0.007 and p = 0.002, respectively, in each group). Following IL13-MSC grafting, no significant difference in MHC-II expression was observed on microglia. Even so, infiltrating macrophages seem to display less MHC-II expression following IL13-MSC grafting, albeit this decrease was not statistically significant.

Similar to our preceding studies [23, 33‐35], MSC and IL13-MSC grafts were able to survive in the pro-inflammatory stroke environment and displayed a similar remodeling pattern, MSC graft-infiltrating CCR2^RFP/+ monocytes/macrophages at the core of the MSC grafts and brain-resident CX3CR1^eGFP/+ microglia and astrocytes surrounding the MSC grafts . Expression of the two M2 markers Arginase1 and Ym1 by microglia and monocytes/macrophages, as a direct result of stimulation by IL13, was only detected in IL13-MSC grafts, but not in MSC grafts (Additional file 2: Figure S2).

The phenotypic distinction between CNS-infiltrating macrophages and brain-resident microglia is a major immunohistochemical concern. It is not only important in recognizing the origin of these cell populations, but it is also of fundamental importance in assessing the pathogenic and therapeutic significance of immune cells within the damaged brain. The lack of a single specific membranous and/or biochemical marker allowing definitive and discriminating identification of these cells is still a puzzling issue in neurobiology. Fortunately, alternative methods have been developed to overcome this hurdle, including the generation of bone marrow chimeric mice, which has proven to provide a powerful tool to distinguish microglia and infiltrated macrophages following ischemic stroke [36, 37]. In this study, however, we made use of the CX₃CR1^+/GFPCCR2^+/RFP transgenic mouse model, which is based on the fact that chemokine receptor CX₃CR1 is predominantly expressed by CNS resident microglia and that CCR2 is upregulated in activated infiltrated macrophages and therefore allows distinction of eGFP-expressing microglia from RFP-expressing infiltrated macrophages [5]. Through the use of this transgenic mouse model and immunofluorescence microscopy, we were able to demonstrate a region-dependent distribution of microglia and macrophages at day 14 after stroke. In the cortical ischemic lesion, CX₃CR1⁺ microglia distinctly accumulated at the border of the lesion, while CCR2⁺ macrophages were localized at the core of the lesion. At day 14, after ischemic stroke, CX₃CR1^+/GFP microglia were the major contributors of recruited cells to the ischemic lesion. These observations are in line with our earlier findings following MSC transplantation in mouse brain, which demonstrated a similar distinct distribution of microglia and macrophages, where infiltrated macrophages invaded the hypoxic/ischemic MSC transplant core, while active microglia remained in the surrounding border [23, 25, 38]. In contrast to their orderly distribution in the cortex, both CX₃CR1⁺ microglia and CCR2⁺ macrophages were found to be randomly distributed in striatal lesions. In agreement with our own observations, others have also reported that resident microglia and blood-derived macrophages localize in the ischemic brain with different temporal and spatial patterns. Garcia-Bonilla et al. describe that at the first week after stroke, mostly diffused CCR2⁺ macrophages were observed throughout the ischemic lesion, while during the second and third week after stroke, they saw that CX₃CR1⁺ cells were localized in the peri-infarct area encircling the ischemic lesion core filled with CCR2⁺ macrophages [39]. Indeed, in a rather complex study design, these latter authors demonstrated that CCR2⁺CX3CR1⁻ monocytes may, in a time-dependent fashion after infiltration, turn into CCR2⁻CX3CR1⁺ macrophages. Thus, we must caution that our assignment of green fluorescent, CX3CR1-positive microglia may to some extent also contain macrophages. This, however, does not affect our assignment of CCR2⁺ red cells as a pure macrophage population. Another study showed that infiltrated blood-borne monocytes were exclusively located at the ischemically injured striatum at days 3, 7, and 14 after MCAO [40]. Their results represent a random distribution of brain-resident microglia and infiltrated monocyte within the striatum after ischemic stroke, with microglia being the vast majority of cells invading the ischemic lesion. Moreover, in a mouse model of traumatic spinal cord injury, CX₃CR1^+/GFP and CCR2^+/RFP cells were randomly distributed around and inside the lesion, with CCR2^+/RFP cells constituting the greater number of accumulated cells in the lesion area [41].

These findings suggest an unknown mechanism by which temporally and spatially distinct ischemic regions can differentially signal to microglia and macrophages. Region-specific differences between cortex and striatum in regard to vascular volume [30], neurogenic potential [42] and neuroinflammatory responses [43] may cause the region-dependent distribution of microglia and macrophages after ischemic stroke.

In our experimental setup, considering CX₃CR1^+/GFP cells as brain-resident microglia, it might be argued that perivascular macrophages, supraependymal macrophages, epiplexus cells of the choroid plexus, and meningeal macrophages express CX₃CR1 too [44]. In addition, we cannot neglect the presence of patrolling macrophages Ly6c⁻ CX₃CR1^high CCR2^low in the ischemic brain with disrupted blood-brain barrier [41, 45]. Interestingly, in an attempt to observe chemokine receptor expression changes in microglia and monocytes/macrophages in development and during inflammatory condition, Mizutani and colleagues have indicated that CCR2 is absent in adult naïve and inflamed CNS resident microglia [5]. Importantly, they have demonstrated that CX₃CR1 expression by microglia is significantly higher compared to CX₃CR1^low CCR2^high or CX₃CR1^high CCR2^low monocytes/macrophages. Therefore, we hypothesize that CX3CR1^+/GFP cells are mostly of brain-resident microglial origin and double-positive cells for CX₃CR1^+/GFP and CCR2^+/RFP are blood-derived macrophage populations [33, 41, 46]. It should be emphasized that the presence of double-positive cells at day 14 after ischemic stroke accounts for a very small proportion of infiltrated macrophages in our experiment. Further assessment of earlier or later time point of macrophage recruitment following CNS damage should also be evaluated. Nevertheless, the discovery of more microglia-specific markers would contribute to more efficient discrimination between CNS resident microglia and infiltrated macrophages in future studies.

In this study, we are the first to determine the immunomodulatory potential of MSC and IL13-expressing MSC engraftment in adjusting the polarization of brain-resident microglia and infiltrated macrophages in an ischemic stroke model. We focused on addressing important applications of IL13-expressing MSCs as a new potential therapeutic approach for ischemic stroke. The main goal of our study was to enhance the therapeutic effects of MSC transplantation following ischemic stroke by boosting the immunomodulatory properties of MSCs. For this reason, we used MSCs as a biodegradable delivery system for the immune modulating cytokine IL13. With this method, we achieved long-term effect on immune function until the second week after stroke induction, during the pro-inflammatory peak of stroke [6]. Earlier investigations by us, on the use of MSCs as a delivery system for the anti-inflammatory cytokine IL13, demonstrated that continued secretion of IL13 by MSCs is able to limit microgliosis, oligodendrocyte loss, and demyelination in cuprizone-treated mice, a model for neuroinflammation and demyelination [23], and promotes histopathological and functional recovery following spinal cord injury in mice [33]. In these studies, we also showed that transplantation of IL13-expressing MSCs induces alternative activation in both MSC graft-invading macrophages and in MSC graft-surrounding microglia, characterized by the expression of Arg-1 and Ym1 as anti-inflammatory-associated markers.

Here, we present for the first time that transplantation of IL13-expressing MSCs at 48 h after ischemia shifts major players of the immune reaction, namely microglia and macrophages, towards an anti-inflammatory, neuroprotective phenotype at 14 days after ischemia. Although not investigated in detail in this study, our previous data with regard to grafting IL-13 expressing MSCs in muscle of brain tissue have shown that the anti-inflammatory M2 phenotype can be induced already at 5 and 7 days post grafting, respectively [35]. The transplantation of IL13-expressing MSCs at day 2 after ischemia caused a noticeable increase in the expression of the general activation marker F4/80, especially by infiltrated macrophages. Most importantly, MSC graft-associated microglia and infiltrated macrophages were strongly driven towards an Arg-1⁺ alternative activation status at day 14 after stroke. Further distinction between recruited brain-resident microglia and infiltrated systemic macrophages following stroke and MSC transplantation in CX₃CR1^eGFP/+ CCR2^RFP/+ double transgenic mice revealed that about two thirds of Arg-1⁺ cells are infiltrated macrophages. Evidently, continued secretion of the cytokine IL13 by MSCs was able to boost the immunomodulatory properties of MSCs by increasing the number of Arg-1⁺ cells, thus acting as a potent neuroprotective mediator [47]. Several groups have reported that Arg-1 expression by different cell types, including astrocytes, neurons, and immune cells, notably decreases during the first few days after cerebral ischemia, after an early peak at day 3 [6, 48]. By continuous secretion of IL-13 via MSCs, we could enhance and prolong the expression of neuroprotective and anti-inflammatory marker Arg-1 until week 2 after stroke induction. In an earlier study by our laboratory, in which we attempted to modulate the polarization of microglia and macrophages by microRNA-124 after ischemic stroke, we achieved an upregulation of Arg-1 and CD-206 expression by microglia and macrophages until day 6 after stroke. In this study, the increased polarization of microglia and macrophages towards the more anti-inflammatory phenotype was positively correlated to increased neuronal survival and functional recovery at day 6 after stroke [32, 49].

Another important observation to be taken into account here is the expression and accumulation of major histocompatibility complex class II (MHC-II) following MSC and IL13-expressing MSC engraftment after stroke. MHC-II expression was predominantly observed on infiltrated macrophages, localized at the core of the ischemic lesion. MHC-II is considered to be a pro-inflammatory mediator, which is overexpressed under a wide range of pathological conditions [50]. However, some studies have highlighted that MHC-II expression can also provide protective and neurotrophic functions and is involved in regeneration and axon integrity in neurodegenerative disorders [51, 52]. As reported previously, accumulation of CD11b⁺ MHC-II⁺ cells in the ipsilateral ischemic hemisphere after transient MCAO provides trophic support and is effective in the remyelination process after stroke [53]. With regard to the expression of Arg-1 and MHC-II, mostly by macrophages at the core of ischemic lesion in our study, we could argue a substantial role of infiltrated macrophages for the regeneration and repair process of injured brain. We propose that infiltrated pro-inflammatory CCR2⁺MHC-II⁺ macrophages may exert some pro-neurogenic functions such as clearance of cell debris after brain injury, which could be efficient at neurogenesis and regeneration process after ischemic stroke. Based on our data, transplantation of IL13-expressing MSCs does not broadly suppress M1 phenotype of microglia and macrophages, but it rather adjusts the balance between pro- and anti-inflammatory immune cell processes. As M1-activated cells typically resolve injured and/or infected tissue, which can lead to detrimental exacerbated injury of healthy tissue, they can also assist in the repair/conservation of synaptic connectivity and axonal regeneration [54, 55]. When considering M2 activated immune cells, mainly anti-inflammatory and regeneration-inducing functions are described, but it is worth noting that long-term maintenance of anti-inflammatory phenotype may adversely affect the immune response, which leads to serious negative effects, such as tumorigenesis [56]. Induction of an M1/M2 phenotype switch to improve natural repair processes in neurodegenerative disorders should be considered with caution. In stroke, our earlier studies on modulation of the pro-/anti-inflammatory balance by microRNA-124 demonstrated enhanced neuronal survival and behavioral improvement [32, 49].

In our present investigation, we have not analyzed the influence of continuous secretion of the anti-inflammatory cytokine IL-13 and/or M2 polarized microglia and macrophages on the survival of autologous MSC grafts after stroke. But in an earlier study, we have demonstrated that transplantation of IL13-expressing MSCs leads to a notable decrease of direct and indirect immune recognition and rejection of cell grafts [35]. Further support of functional interaction between alternative activation status of surrounding inflammatory milieu at the graft site and survival of MSCs comes from a study by Yu et al. They have recently demonstrated that M2 macrophage-secreted OA/GPNMB, osteoactivin/glycoprotein in non-metastatic melanoma protein B, positively contributes to viability, proliferation, and migration of MSCs, through ERK and AKT signaling pathways [57]. Similar MSC survival improvement was achieved via an ischemic microenvironment in vitro study, in which macrophage migration inhibitory factor (MIF), a cytokine expressed by activated monocytes/macrophages, protects MSCs from apoptosis via a CD74-dependent Akt-FOXO3a-related pathway [58]. Enhancing MSC survival in the face of oxygen and nutrient deprivation that naturally occur in the ischemic stroke and cell graft procedure clearly needs further investigation.