Interleukin‐13 immune gene therapy prevents CNS inflammation and demyelination via alternative activation of microglia and macrophages

Detrimental inflammatory responses in the central nervous system are a hallmark of various brain injuries and diseases. With this study we provide evidence that lentiviral vector‐mediated expression of the immune‐modulating cytokine interleukin 13 (IL‐13) induces an alternative activation program in both microglia and macrophages conferring protection against severe oligodendrocyte loss and demyelination in the cuprizone mouse model for multiple sclerosis (MS). First, IL‐13 mediated modulation of cuprizone induced lesions was monitored using T2‐weighted magnetic resonance imaging and magnetization transfer imaging, and further correlated with quantitative histological analyses for inflammatory cell influx, oligodendrocyte death, and demyelination. Second, following IL‐13 immune gene therapy in cuprizone‐treated eGFP+ bone marrow chimeric mice, we provide evidence that IL‐13 directs the polarization of both brain‐resident microglia and infiltrating macrophages towards an alternatively activated phenotype, thereby promoting the conversion of a pro‐inflammatory environment toward an anti‐inflammatory environment, as further evidenced by gene expression analyses. Finally, we show that IL‐13 immune gene therapy is also able to limit lesion severity in a pre‐existing inflammatory environment. In conclusion, these results highlight the potential of IL‐13 to modulate microglia/macrophage responses and to improve disease outcome in a mouse model for MS. GLIA 2016;64:2181–2200


Introduction
W hile innate immune cells of the central nervous system (CNS) at one hand play a major role in development and maintenance of brain integrity and function (Kettenmann et al., 2011;Nimmerjahn et al., 2005;Ransohoff and Perry, 2009), dysregulation of CNS immune responses on the other hand is highly associated with the pathophysiology of various neurodegenerative diseases (Perry et al., 2010;Prinz et al., 2011), including demyelinating disorders like multiple sclerosis (MS) (Lassmann and van Horssen, 2016;Mahad et al., 2015;Rawji and Yong, 2013;Sriram, 2011). Microglia, wellrecognized as brain tissue-specific macrophages (Chan et al., 2007;Ginhoux et al., 2010Ginhoux et al., , 2013Tambuyzer et al., 2009), have the unique capacity to sense alterations of the brain microenvironment, and in response rapidly migrate, proliferate, and undergo specific differentiation programs to restore brain homeostasis (Ousman and Kubes, 2012). Functionally differentiated microglia and macrophages are commonly subdivided in two categories Mills et al., 2000;Mosser and Edwards, 2008;Murray et al., 2014), the classically activated phenotype, which is characterized by the secretion of pro-inflammatory and cytotoxic mediators, and the alternatively activated phenotype, which generally contributes to extracellular matrix remodelling, progenitor cell differentiation and tissue regeneration. Experimental modulation of these microglia and macrophage polarization states can easily be achieved in vitro through supplementation of microglia/macrophage cultures with lipopolysaccharides (LPS) and interferon gamma (IFN-g) or with interleukin (IL)24/13 to induce the classically activated or the alternatively activated polarization state, respectively (Gordon, 2003;Mills et al., 2000). In this study, our main aim is to investigate whether targeted in vivo delivery of interleukin 13 (IL-13) at the onset of, or during neuroinflammatory processes, can alter degenerative (histo-)pathological disease outcomes.
Activated microglia and macrophages are suggested to co-orchestrate oligodendrocyte death in the cuprizone mouse model of CNS demyelination (Hiremath et al., 1998;Kipp et al., 2009;Matsushima and Morell, 2001;Praet et al., 2014a). In this study, we applied lentiviral vector (LV)-mediated transduction of the splenium of the corpus callosum (CC), that is, the brain area affected most prominent in this neuroinflammation mouse model, to investigate the effect of IL-13 on developing neuroinflammatory lesions and subsequent demyelinating events. While recent in vivo treatment strategies aiming at directing the differentiation of microglia and macrophages toward an alternatively activated state highly rely on histological analyses to demonstrate reduced CNS damage and improved disease outcome (Hu et al., 2012;Ma et al., 2015;Miron et al., 2013;Yang et al., 2015), we here additionally performed magnetic resonance imaging (MRI) as the gold standard for in vivo detection of demyelinating lesions in CNS white matter (Polman et al., 2011;Trip and Miller, 2005). In addition to standard T 2 -weighted MRI, we also used magnetization transfer imaging (MTI) as an emerging technique to provide greater insight into the tissue microstructure in terms of inflammation, gliosis, and myelination (Ropele and Fazekas, 2009). Hence, we here report on multiparametric MRI approaches complementary and correlating to our quantitative histological analyses for an accurate in vivo monitoring of CNS inflammatory and demyelinating pathology, as well as the effect of IL-13 immune gene therapy thereon. Finally, as one of the major unsolved issues in CNS neuro-de/regeneration relates to the specific roles for CNSresident microglia and CNS-invading peripheral macrophages (London et al., 2013), in our experimental setup using eGFP 1 bone marrow (BM) chimeric mice we attempted to clarify potential different and specific roles of both populations in disease and during IL-13 LV-mediated immune gene therapy in the CNS.

Materials and Methods
Mice C57BL/6J-Tyr c-2J /J mice (albino C57BL/6J mice, strain code 000058) and transgenic C57BL/6-eGFP mice (strain code 003291) were obtained via Jackson Laboratories and bred in the animal facility of the University of Antwerp. For all experiments, mice were kept in normal day-night cycle (12/12) with ad libitum access to food and water. All experimental procedures were approved by the Ethics Committee for Animal Experiments of the University of Antwerp (Approval No 2011/13).

The Cuprizone Mouse Model
At the age of eight weeks, mice received a standard rodent chow mixed with 0.2% w/w cuprizone (Sigma-Aldrich) for a period of 4 weeks to induce neuroinflammation and CNS demyelination, as previously described (Guglielmetti et al., 2014;Orije et al., 2015;Praet et al., 2012Praet et al., , 2015a.

Stereotaxic Injections of LV
All surgical interventions were performed according to institutional guidelines. Briefly, mice were anesthetized by intraperitoneal (IP) injection of a ketamine (75 mg/kg, Anesketin; Eurovet NV/ SA) 1 medetomidine (1 mg/kg, Domitor; Pfizer Animal Health S.A.) mixture and positioned in a stereotactic head frame to achieve flat skull position (Stoelting). Stereotactic coordinates to target the splenium of the CC were as follows: AP 21.6 mm, Lat 0.3 mm, and DV 21.1 mm (relative to bregma, Fig. 1a). A midline scalp incision was made to expose the skull, and a hole was drilled in the skull using a dental drill burr (Stoelting). Thereafter, an automatic micro-injector pump (kdScientific) with a 10 lL Hamilton syringe was positioned above the exposed dura. A 32-gauge needle (Hamilton), attached to the syringe, was placed through the intact dura and positioned at the respective depth. After 2 min of pressure equilibration, 2 lL of LV concentrate (1.5 3 10 7 2 3 3 10 7 pg/mL p24) was injected at a speed of 0.5 lL/min. Before final needle retraction, a waiting period of 5 min was kept to allow for pressure equilibration and to prevent backflow of the injected LV suspension. Next, the skin was sutured (Vicryl, Ethicon) and a 0.9% NaCl (Baxter) solution was given subcutaneously to prevent dehydration while mice were placed under a heating lamp to recover. Anesthesia was reversed by an IP injection of atipamezol (Antisedan 5 mg/mL, Pfizer Animal Health S.A.).

Bone Marrow Transplantation
To discriminate microglia and macrophages, eGFP 1 BM chimeric mice were generated as previously described (Le Blon et al., 2014). Briefly, four to six weeks old mice received 10Gy total body irradiation using an XRAD320 small animal irradiation device (Precision X-Ray). For this, groups of five nonanesthetised mice were placed in a single cage within the whole irradiation field (without head protection). Next, a single intravenous injection of total eGFP 1 BM cells (1.5 3 10 6 cells in 100 mL PBS) was administered via the tail vein within 6 h postirradiation. Total eGFP 1 BM cells were isolated from 8-week old C57BL/6-eGFP mice by flushing dissected femurs and tibias with sterile PBS. Before administration, total BM cells were filtered over a 70 mm sterile mesh (Becton Dickinson), centrifuged, and suspended in PBS. During recovery, mice were treated with Enrofloxacin (1 mL/mL; Baytril 10%; Bayer) added to the drinking water for 4 weeks postirradiation.

MRI Acquisition
In vivo imaging experiments were conducted at 400 MHz on a 9.4T Bruker Biospec system (Biospec 94/20 USR, Bruker Biospin) using a standard Bruker cross coil setup, with a quadrature volume coil for excitation and quadrature mouse surface coil for signal detection. During imaging, mice were anesthetized using 2% isoflurane (Iso-floV R , Abbot Laboratories) in a mixture of 30% O2 and 70% N2O at a flow rate of 600 mL/min. Mice were fixed in an animal restrainer with ear bars and a tooth bar. Respiratory rate was continuously monitored and body temperature was measured and maintained constant at 378C using a feedback coupled warm air system (MR compatible Small Animal Monitoring and Gating System, SA instruments). First, anatomical RARE-T 2 images were acquired to determine the position of the mouse in the magnet and enable a uniform position of the coronal slices for every T 2 and MTI experiments [repetition time (TR) 5 2500 ms; echo time (TE [effective]) 5 33.44 ms; RARE factor 5 8; number of slices (NS) 5 16 with a slice thickness of 0.4 mm; field of view (FOV) 5 20 3 20 mm; matrix size 5 256 3 256; in-plane resolution of 0.078 3 0.078 mm; scan time 5 1 min 20 s). T 2 values were acquired with the Multi-Slice, Multi-Echo sequence that is based on the Carr-Purcell-Meiboom-Gill sequence, where transverse magnetization of a 908 pulse is refocused by a train of 1808 pulses generating a series of echoes. The following imaging parameters were used: number of averages (NA) 5 1; NS 5 6 with a slice thickness of 0.4 mm and an interslice thickness of 0.4 mm; number of echoes 5 10 with echo spacing 5 8.5 ms (TE) being 8.5; 17; 25.5; 34; 42.5; 51; 59.5; 68; 76.5; 85); TR 5 4000 ms; FOV 5 20 3 20 mm; matrix size 5 256 3 256; in-plane resolution 5 0.078 3 0.078 mm, scan time 5 12 min 48 s). For magnetization transfer contrast (MTC) MRI, an off-resonance RF pulse was incorporated into a RARE T 2 sequence (TR 5 4100 ms, TE [effective] 5 33.44 ms; RARE factor 5 8; NS 5 1 with a slice thickness of 1 mm, FOV 5 20 3 20 mm, matrix size 5 192 3 192, in-plane resolution of 0.104 3 0.104 mm, NA 5 3, scan time 5 4 min 55 s). RARE T 2 images were acquired without off-resonance RF pulse (unsaturated) and with off-resonance RF pulse (saturated) applied (pulse strength 5 3 lT; pulse length 5 4000 ms, saturation time 5 0.244 ms) at the offset frequency of 2000 Hz. MTC MRI was acquired at two different location of the mouse brain, first at the level of the splenium of the CC, and second at the level of the external capsule.

MRI Processing
T 2 maps were generated with custom-built programs written in MAT-LAB (MATLAB R2011b, The MathWorks) using a monoexponential fit function [y 5 A 1 C*exp (2t/T 2 )], where A 5 Absolute bias, C 5 signal intensity, T 2 5 transverse relaxation time. Regions of interest (ROIs) were drawn manually on the T 2 -weighted images (Fig. 1b), according to a mouse brain atlas, with AMIRA software (Mercury Computer systems) and regional average T 2 values were calculated. ROIs included the external capsule and the splenium of the CC. For MTI experiments, saturated images were realigned to the unsaturated image using AMIRA, for each animal separately. The realigned images were used to calculate magnetization transfer ratio (MTR) 5 (unsaturated 2 saturated)/unsaturated). ROIs were manually delineated on the MTR images (Fig. 1b) and the mean MTR were calculated for the external capsule and the splenium of the CC.

Histological Quantification
Quantitative analyses of macrophage and/or microglia responses were performed using TissueQuest immunofluorescence analysis software (TissueGnostics GmbH, v3.0), as previously described by us Le Blon et al., 2014;Praet et al., 2015b;Reekmans et al., 2013). The following parameters were determined: the cellular density of F4/80 1 microglia/macrophages, the cellular density of Arg-1 1 microglia/macrophages, the cellular density of MHCII 1 microglia/macrophages and the cellular density of eGFP 1 macrophages. Additionally, the cellular density of CC-1 1 oligodendrocytes was determined using the same software. Using NIH ImageJ analysis software (v1.46r), the degree of glial fibrillary acidic protein (GFAP) astrogliosis and myelin basic protein (MBP) was determined based on the image-covering staining and expressed as percentage of the total area, as previously described by us (Praet et al., 2015a).

qRT-PCR
Mice were perfused with ice-cold 0.9% NaCl solution, directly followed by removal of the brain and dissection of the CC area around the injection site. The extracted tissue sections were immediately placed in RNALater solution (Ambion) and following overnight incubation at 48C stored at 2208C until further processing. Total RNA was extracted using the Purelink RNA Kit (Invitrogen). RNA quantity and purity were determined using a ND-1000 micro-spectrophotometer (NanoDrop Technologies). One-lg total RNA was reverse-transcribed using a mixture of random primers (High Capacity cDNA Reverse Transcription kit, Applied Biosystems). PCR primers (provided in Supp. Info. Table S1) were designed with Primer3 software to bridge the exon-intron boundaries within the gene of interest to exclude amplification of contaminating genomic DNA. Primers were purchased from IDT (Laboratorios Conda S.A., Torrejon de Ardoz, Spain). Real-time quantitative RT-PCR analysis was carried out using SYBR green I dye detection (#11761500, Invitrogen) using the iCycler iQTM Multicolor Real-Time Detection System (Bio-Rad, Hercules, CA). Optimized thermal cycling conditions were: 1 min at 508C, 8 min, and 30 s at 958C and 40 cycles of 15 s at 958C and 30 s at 608C. Data were collected after each cycle and graphically displayed (iCycler iQTM Real-time Detection System Software, version 3.1, Bio-Rad). Melt curves were performed upon completion of the cycles to ensure specificity of the product amplification. Housekeeping gene for normalization was succinate dehydrogenase complex subunit A. Data were analysed with the 2-DDCT method. For comparison purposes, values of the samples of each group (n 5 8-10 mice per group) are expressed as fold versus the mean (n 5 7 mice) value of the control noninjected group.

Statistical Analyses
Statistical analyses of the T 2 relaxation times and MTR were performed separately on the mean value of the splenium using (i) a one-way ANOVA for group comparison prior to cuprizone administration (W0) and (ii) using a (random effects) linear regression model for group comparison after four weeks of cuprizone administration, as described by the following equations: with Y 1 and Y 2 defined as the difference between the mean T 2 relaxation times or the mean MTR after four weeks of cuprizone administration (W4) and W0. For the EC, statistical analyses of the T 2 relaxation times and MTR were performed separately on the mean value of the left and right EC using a linear mixed model. The obtained P-values were corrected for multiple testing using the Tukey HSD post-hoc test.
For the results obtained from histological analyses the following statistical analyses were applied: (i) Comparison between experimental groups in cellular density for CC-1 1 cells and F4/80 1 cells as well as the coverage percentage of MBP and GFAP were analyzed using one-way ANOVA, and the given P-values were corrected for multiple testing using the Tukey HSD post-hoc test. (ii) The appearance of MHCII-and Arg1-expressing F4/80 cells was analyzed using one sample t test. (iii) Comparison between groups in cellular density of eGFP 1 macrophages were analyzed using generalized estimating equations (GEE), with the given P-values being corrected for multiple testing using the false discovery rate method. (iv) The appearance of MHCII-and Arg1-expressing eGFPmicroglia and eGFP 1 macrophages was analyzed using GEE, with the given P-values being corrected for multiple testing using the false discovery rate method. (v) The dependence of MBP and GFAP coverage percentage, and cellular density for CC-1 1 and F4/80 1 between T 2 relaxation times or MTR was measured using the Spearman's correlation coefficient. Correlation coefficients and P-values are stated in the results section. For the results obtained with qRT-PCR, group comparison was performed using a Wilcoxon rank sum test with FDR correction. All of the above data and/or analyses are presented either in dot plots or in graphs showing mean 6 standard deviation (SD) or standard error of the mean (SEM), as indicated in the figure legends. A P-value of <0.05 was considered statistically significant.

Results
Noninvasive MRI Reveals Less Severe T 2 and MTR Deviation in the Splenium of Cuprizone Treated Mice Following IL-13 LV Injection Before initiating therapeutic experiments, correct stereotaxic targeting of the LV injections to the splenium was optimized using the BFP LV and validated by means of histological analysis (Fig.  1a). To investigate the effect of IL-13 expression in the splenium on cuprizone induced inflammation and demyelination, noninjected control mice (No inj.), blue fluorescent protein (BFP) LV injected mice and IL-13 LV injected mice were randomly assigned to a subgroup that received a 4-week regular rodent diet (control) or a subgroup that received a 4-week cuprizone supplemented diet (cuprizone; Fig. 2a). This 4-week time point was chosen as it corresponds to the optimal time window in the cuprizone mouse model comprising the highest level of microglia/macrophage activation with extensive demyelination in the splenium (Praet et al., 2014a). In this part of the study, noninvasive MRI was performed on all mice prior to LV injection and/or cuprizone administration and at the end of the 4-week study period. For analysis of MRI data, ROI were delineated on obtained T 2 and MTR maps at the level of the splenium (Fig. 1b  upper row). No significant differences in T 2 or MTR values were observed between the 6 different groups at the beginning of the experiment (data not shown). After 4 weeks, while noninjected mice, BFP LV injected mice and IL-13 LV injected mice that received a regular rodent diet displayed T 2 and MTR values comparable to baseline values (respectively, Fig. 2b-e, CONTROL No inj., BFP and IL-13), mice that received the cuprizone supplemented diet without LV injection or with injection of the BFP LV displayed significantly increased T 2 values and significantly FIGURE 2. decreased MTR values at the level of the splenium (respectively, Fig. 2b-e, CONTROL versus CUPRIZONE No inj. and BFP). These observations are suggestive for inflammation and demyelination in the splenium. Interestingly, cuprizone treated mice expressing IL-13 in the splenium displayed a significantly lower deviation of the normal T 2 and MTR values at the level of the splenium as compared with noninjected or BFP LV injected cuprizone treated mice (Fig. 2b-e, CUPRIZONE IL-13 vs. CUPRIZONE No inj. and CUPRIZONE BFP). Lastly, we noted a strong inverse correlation between T 2 and MTR values (Fig.  2f), implying that T 2 and MTR changes may be indicative of similar or associated pathological alterations. In summary, the presented data show that IL-13 immune gene therapy in the splenium strongly decreases cuprizone induced pathological alterations as assessed by noninvasive MRI measurements.

IL-13 LV Injection in the Splenium Prevents Cuprizone Induced Inflammation, Oligodendrocyte Death, and Demyelination
To further validate our imaging observations, quantitative histological analyses were performed. As expected, mice that received the cuprizone supplemented diet without LV injection or with injection of the BFP LV displayed a significantly lower number of CC-1 1 oligodendrocytes in the splenium, which further resulted in severe demyelination, as demonstrated by the significant loss of MBP expression in the splenium (Fig. 3a,b, CON-TROL vs. CUPRIZONE No inj. and CUPRIZONE BFP). Additionally, significant recruitment of F4/80 1 activated microglia/macrophages was observed in mice that received the cuprizone supplemented diet without LV injection or with injection of the BFP LV (Fig. 3c, CONTROL vs. CUPRIZONE No inj. and CUPRIZONE BFP). Similarly, significant astrocyte activation, as demonstrated GFAP 1 astrogliosis, was observed in mice that received the cuprizone supplemented diet without LV injection or with injection of the BFP LV (Fig. 3d, CONTROL vs. CUPRIZONE No inj. and CUPRIZONE BFP). In contrast, as already suggested by the MRI analyses, cuprizone treated mice expressing IL-13 in the splenium displayed significantly higher CC-1 1 oligodendrocyte numbers and less demyelination as compared with noninjected or BFP LV injected cuprizone treated mice (Fig. 3a,b, CUPRIZONE IL-13 vs. CUPRIZONE No inj. and CUPRIZONE BFP). Interestingly, cuprizone treated mice expressing IL-13 in the splenium also displayed significantly reduced recruitment of F4/80 1 activated microglia/macrophages (Fig. 3c, CUPRIZONE IL-13 vs. CUPRIZONE No inj. and CUPRIZONE BFP). In addition, also GFAP 1 astrogliosis was significantly decreased as compared with noninjected or BFP LV injected cuprizone treated mice (Fig. 3d, CUPRIZONE IL-13 vs. CUPRIZONE No inj. and CUPRIZONE BFP). In agreement with our conclusion from the MRI analyses, histological data thus confirms that IL-13 LV injection in the splenium decreases cuprizone induced inflammatory responses and subsequent-at least in part-oligodendrocyte death and demyelination. Furthermore, given the large dataset obtained, we performed a correlation analysis between MRI metrics and histological metrics at the level of the splenium for all experimental groups. At one hand, Spearman correlation coefficients revealed an inverse correlation between T 2 values and the oligodendrocyte/myelin markers CC-1 and MBP (Supp. Info. Figure S1, q 5 20.8154 and q 5 20.7997, respectively) and a positive correlation between T 2 values and the inflammation-associated markers F4/80 and GFAP (Supp. Info. Figure S1, q 5 0.8506 and q 5 0.7716, respectively). However, MTR values display a positive correlation with the oligodendrocyte/myelin markers CC-1 and MBP (Supp. Info. Figure S1, q 5 0.7990 and q 5 0.8318, respectively) and an inverse correlation with the inflammation-associated markers F4/80 and GFAP (Supp. Info. Figure S1, q 5 20.8929 and q 5 20.8229, respectively). Concluding, and in agreement with preceding studies (Acs et al., 2009;Boretius et al., 2012;Chandran et al., 2012;Harsan et al., 2008;Thiessen et al., 2013;Torkildsen et al., 2009), both T 2 and MTR values can be used as predictive norms for monitoring cuprizone induces inflammatory and demyelinating events, including therapeutic intervention thereon.
FIGURE 2: Noninvasive MRI reveals less severe T 2 and MTR deviation in the splenium of cuprizone treated mice following LV mediated expression of IL-13.(a) To induce inflammation and demyelination, mice were fed a cuprizone supplemented diet for a period of four weeks, while mice of the control groups were fed a regular rodent chow. On the first day of cuprizone administration mice received an injection of a BFP LV or IL-13 LV or no injection (No inj.) and were assigned to a cuprizone or control group. MRI was performed prior to LV injection and after four weeks of cuprizone feeding, immediately followed by histological analysis. (b) Panel shows a representative T 2 map for each experimental group at the level of the splenium of the CC. Hyper-intense contrast can be visualized at the level of the splenium after cuprizone administration, reflecting inflammation and/or demyelination, but not for mice that also received the IL-13 LV injection (white arrows indicate the location of the CC of cuprizone treated mice). Scale bar 5 1 mm. (c) In a similar manner, panel shows a representative color-coded MTR image for each experimental group at the level of the splenium of the CC. Disappearance of the CC (color-coded in red) can be observed after cuprizone administration, indicating inflammation and demyelination, but not for mice that also received the IL-13 LV injection (white arrows indicate the location of the CC of cuprizone treated mice). Scale bar 5 1 mm. (d) IL-13 LV Injection in the Splenium Protects Locally During Cuprizone Induced Inflammation, Oligodendrocyte Death, and Demyelination Given the potential of IL-13 to interfere with the development of inflammatory and demyelinating events following cuprizone treatment, we additionally investigated whether the beneficial effect of IL-13 is restricted to the LV injection site (i.e., the splenium), or whether protection can be observed at cuprizone affected sites distinct from the splenium. For this, additional ROI were delineated on obtained T 2 and MTR maps in the external capsule at the level of the genu (Fig. 1b, lower row). Here, all cuprizone treated mice display significantly increased T 2 values and significantly decreased MTR values as compared with control groups indicating extensive demyelination and inflammation in the external capsule (Fig. 4a,b, CONTROL versus CUPRIZONE). Interestingly, MTR, but not T 2 was able to detect a lower deviation of the normal MTR for mice that received the cuprizone diet and the IL-13 LV injection as compared to mice that received cuprizone and BFP LV injection or no injection (Fig. 4b CUPRIZONE IL-13 vs. CUPRIZONE No inj. and CUPRI-ZONE BFP). Further histological analyses, as shown by the representative images provided in Figure 4, clearly showed a loss of CC-1 1 oligodendrocytes (Fig. 4c), demyelination (Fig. 4d), microglia/macrophage recruitment (Fig. 4e) and astrogliosis (Fig.  4f) at the level of the external capsule in cuprizone treated mice as compared with the control mice. Correlation analyses confirmed the findings reported for the splenium of the CC, that is, an inverse correlation between T 2 and MTR (q 5 20.7666, P < 0.0001) and similar association of MRI metrics with histological markers at the level of the external capsule (Supp. Info. Figure S2). Altogether these data suggest that expression of IL-13 in the splenium alters the response to cuprizone diet, but is not sufficient to rescue cuprizone-induced pathology in regions distinct from the LV injection site.

IL-13 LV Injection in the Splenium Alters
Phenotypical Properties of Activated F4/80 1 Microglia/Macrophages During Cuprizone Treatment To further understand the observed protective effect of IL-13 LV injection in the splenium, we investigated the expression of major histocompatibility complex class II (MHCII) and Arginase-1 (Arg-1) by activated F4/80 1 microglia/macrophages. As already described above (Fig. 3c), mice that received the cuprizone supplemented diet without LV injection displayed a significant increase of F4/80 1 microglia/macrophages at the level of the FIGURE 3: LV mediated expression of IL-13 in the splenium prevents cuprizone induced inflammation, oligodendrocyte death, and demyelination.Histological analyses performed at the level of the splenium for CC1 1 oligodendrocytes, MBP myelination, F4/80 1 activated microglia/macrophages and GFAP 1 astrogliosis are depicted in panels a-d, respectively, and confirm decreased oligodendrocyte cell loss and demyelination associated with a decreased microgliosis and astrogliosis in IL-13 LV injected cuprizone treated mice. *P < 0.05, **P < 0.01, ***P < 0.001, One-way ANOVA and Tukey post hoc test (n 5 3 -6 mice). Scale bar 5 100 lm. [Color figure can be viewed at wileyonlinelibrary.com] splenium compared with control mice (Fig. 5a,b, CUPRIZONE No inj. versus CONTROL No inj., BFP LV and IL-13). Interestingly, a significant subpopulation of F4/80 1 activated microglia/ macrophages in cuprizone treated mice displays MHCII expression upon BFP LV or IL-13 LV injection (Fig. 5a, CUPRIZONE No inj. vs. CUPRIZONE BFP and IL-13). Note that this subpopulation of MHCII 1 F4/80 1 activated microglia/macrophages also significantly appears following BFP LV injection in control mice (Fig. 5a, CONTROL BFP). Furthermore, a significant subpopulation of F4/80 1 activated microglia/macrophages displays Arg-1 expression upon IL-13 LV injection only in cuprizone treated mice (Fig. 5b CUPRIZONE IL-13 vs. CUPRIZONE No inj. and BFP). Next, we investigated whether the observed F4/80 1 MHCII1 and F4/80 1 Arg-11 cells are independent cell populations or whether they are the same cell population. Interestingly, upon costaining of MHCII and Arg-1, it is clear that cuprizone treated mice expressing IL-13 present three distinct populations of cells, which either are MHCII 1 , Arg-1 1 , or Arg-1 1 -MHCII 1 , with the latter two being the most prominent (Fig. 5c, CUPRIZONE IL-13). Summarizing, we here demonstrate that LV injection, either BFP LV or IL-13 LV, can modulate the activation state of F4/80 1 activated microglia/macrophages as demonstrated by the expression of MHCII. Additionally, expression of IL-13 was able to induce Arg-1 expression in activated F4/80 1 and F4/80 1 MHCII 1 microglia/macrophages, which might account for the beneficial effects observed on inflammation and demyelination.

IL-13 LV Injection in the Splenium Induces a Unique Phenotype and Cytokine Profile During Cuprizone Treatment
To further extend our hypothesis that IL-13 is capable of modulating inflammatory responses in cuprizone treated mice, we determined the mRNA expression levels of multiple FIGURE 4: LV mediated expression of IL-13 in the splenium protects locally against cuprizone induced inflammation, oligodendrocyte death, and demyelination.Panel a shows a representative T 2 map for each experimental group at the level of the external capsule. Hyperintense contrast can be visualized at the level of the external capsule, reflecting inflammation and/or demyelination. Scale bar 5 1 mm. T 2 values of individual mice are displayed in the graph with mean 6 SD and reveal an increase of T 2 value in all cuprizone treated mice at the level of the external capsule (n 5 5-14 mice). ***P < 0.001, Linear mixed model, Tukey post hoc test. In a similar manner, panel b shows a representative color-coded MTR image for each experimental group at the level of the external capsule. Disappearance of the contrast from the external capsule (color-coded with red) can be visualized in groups that received the cuprizone supplemented diet, reflecting inflammation and/or demyelination. Scale bar 5 1 mm. MTR of individual mice are displayed in the graph with mean 6 SD and indicated a decrease of MTR in all cuprizone treated mice. *P < 0.05, ***P < 0.001, Linear regression model, Tukey post hoc test (n 5 5-11 mice). Histological analyses performed at the level of the external capsule for CC1 1 oligodendrocytes, MBP myelination, F4/80 1 activated microglia/macrophages and GFAP 1 astrogliosis are depicted in panels c-f, respectively, and confirm oligodendrocytes cell loss and demyelination associated with the presence of microgliosis and astrogliosis and in all cuprizone treated mice at the level of the external capsule, independently of LV injection (n 5 3-5 mice). Scale bar 5 100 lm. [Color figure can be viewed at wileyonlinelibrary.com] cytokines and immune phenotype markers at the level of the CC in different experimental groups by means of qRT-PCR. At first, it is clear that mRNAs encoding Arg-1, YM-1, and Gal-3, which are associated with alternative microglia/macrophage activation, are significantly upregulated in cuprizone treated mice injected with IL-13 LV, as compared with healthy control and noninjected cuprizone treated mice (Fig.  6a-c). While the pro-inflammatory cytokines TNF-a, IL-1b, and iNOS were significantly upregulated in cuprizone treated mice injected with the BFP LV mice, but not in noninjected cuprizone treated mice, cuprizone treated mice injected with IL-13 LV were able to counteract TNF-a, but not IL-1b expression (Fig. 6d-f ). Unexpectedly, a small increase of iNOS mRNA was detected in cuprizone treated mice injected with IL-13 LV (Fig. 6f ). Concluding, our results demonstrate that in vivo LV mediated IL-13 expression is capable to induce mRNA expression associated with alternative microglia/macrophage activation, which might lie at the basis of above-described inflammatory/neuroprotective effects in the cuprizone mouse model, but also indicates that LV mediated gene transfer on its own can shift neuroinflammatory responses toward a more proinflammatory character. Nevertheless, downstream IL-13 events are capable-at least in part-to counteract LV mediated reinforcement of inflammatory responses.

Both Microglia and Macrophage Phenotype Is Modulated Through LV Mediated Expression of IL-13 in the Splenium
As the initiation and modulation of neuroinflammation is a complex interplay between brain resident microglia and peripheral immune cells, we here investigated whether any of the observed MHCII 1 , Arg-1 1 , or MHCII 1 Arg-1 1 cell populations can be specifically linked to microglia and/or macrophage cell populations. For this, we generated 8-week old eGFP 1 BM chimeras that were subjected to BFP LV or IL-13 LV injection followed by 4 weeks of cuprizone administration. Noninjected cuprizone treated eGFP1 BM chimeras served as control in this experimental setup (Fig. 7a). Cuprizone induced pathology and IL-13 LV protection was first verified by means of T 2 -weighted imaging (Fig. 7b). In agreement with the above-described results, MHCII 1 cells were detected in both BFP LV and IL-13 LV injected mice, while Arg-1 1 cells were only detected in IL-13 LV injected mice (Fig. 7c, second 1 third column). Based on eGFP expression, which discriminates infiltrating eGFP 1 peripheral macrophages from brain resident eGFP 2 microglia, it can be noted that a significantly higher number of eGFP 1 peripheral macrophages contributes to-or at least are present-within cuprizone induced inflammatory lesions in the splenium upon BFP LV or IL-13 LV injection (Fig. 7c, first column;  Fig. 7d, No inj. Vs. BFP and IL-13). Further colocalization studies revealed that MHCII is expressed by both eGFP 1 macrophages and eGFP 2 microglia upon control BFP LV (Fig. 7c, fourth column middle image; Fig. 7e, BFP). Similarly, and highly interesting, colocalization studies reveal that expression of IL-13 in the splenium by means of LV injection and subsequent cuprizone administration leads to the significant appearance of 4 distinct inflammatory cell populations: MHCII 1 , Arg-1 1 , and Arg-1 1 MHCII 1 eGFP 1 macrophages and Arg-1 1 MHCII 1 eGFP 2 microglia (Fig. 7c, fourth column lower image; Fig. 7e, IL-13). These data indicate that both infiltrating macrophages and brain-resident microglia have the capacity to display Arg-1 expression in vivo upon stimulation with IL-13, and thereby potentially contribute to the beneficial effects observed on inflammation and demyelination.
Therapeutic LV Mediated Expression of IL-13 in the Splenium Protects against Severe Cuprizone induced Demyelination and Is Associated with an Altered Microglia/Macrophages Phenotype Finally, we further investigated whether IL-13 LV injection may also exert immunomodulatory and/or neuroprotective effects in a therapeutically relevant setup. For this, mice received a cuprizone supplemented diet for a period of 2.5 weeks to trigger a neuroinflammatory response and oligodendrocyte metabolic impairment. After 2.5 weeks of cuprizone treatment, mice received an injection of the IL-13 LV and were kept under cuprizone diet for an additional 1.5 weeks, after which MRI evaluation and histological analyses were performed (Fig. 8a). Although calculated T 2 values did not show a major difference between noninjected cuprizone treated mice and IL-13 LV injected mice, MTR values in contrast revealed a significantly lower deviation of the normal MTR in IL-13 LV injected cuprizone treated mice as FIGURE 5: LV mediated expression of IL-13 in the splenium alters phenotypic properties of activated F4/80 1 microglia/macrophages during cuprizone treatment.Phenotypic properties of microglia/macrophages at the level of the splenium of the CC was assessed after injection of LV followed by four weeks of cuprizone administration. (a) Representative images of each experimental group stained for F4/80 (red) and MHC-II (green). Scale bar 5 100 lm. Graph displays MHCII staining expressed as percentage relative of F4/80 1 cells of noninjected cuprizone treated mice 6 SEM. Downward error bars for white coded bars. Upward error bars for black coded bars. Significant appearance of MHCII expression is indicated by the * sign. **P < 0.01, ***P < 0.001, One sample t-test (n 5 3-6 mice). (b) Representative images of each experimental group stained for F4/80 (red) and Arg-1 (green). Scale bar 5100 lm. Graph displays Arg-1 staining expressed as percentage relative of F4/80 1 cells of noninjected cuprizone treated mice 6 SEM. Downward error bars for white coded bars. Upward error bars for black coded bars. Significant appearance of Arg-1 expression is indicated by the # sign. #P < 0.05, One sample t-test (n 5 3-6 mice). (c) Representative images of each experimental group stained for Arg-1 (red) and MHCII (green). Scale bar 5100 lm. Graph shows the density of Arg-1 1 and MHCII 1 cells 6 SEM. Significant appearance of MHCII, Arg-1 and MHCII/Arg-1 expression is indicated by the *, # and ‡ symbols, respectively. */#/ ‡ P < 0.05, **P < 0.01, One sample t-test (n 5 3-6 mice). [Color figure can be viewed at wileyonlinelibrary.com] compared to noninjected cuprizone treated mice, indicative for preservation of the splenium of the CC following IL-13 LV injection (Fig. 8b,c). Subsequently, histological analyses (Fig. 8d) confirmed a significant reduction F4/80 1 inflammatory cells (Fig. 8e), with the remaining inflammatory cells being Arg-1 1 MHC-II 1/2 (Fig. 8f ). Furthermore, quantitative analysis of GFAP and MBP expression indicated significantly reduced astrogliosis and demyelination (Fig. 8g,h).
Most importantly, CC1 1 oligodendrocyte survival (Fig. 8i) was significantly higher in the IL-13 LV injected CPZ group as compared with the noninjected CPZ control group. In summary, these results suggest that LV mediated expression of IL-13 can also be applied to a pre-existing neuroinflammatory environment to force microglia/macrophages into an alternative state of activation, with subsequent neuroprotective effects. SEM. *P < 0.05, **P < 0.01, ***P < 0.001, One-way ANOVA, Dunnett's multiple comparison test (n 5 7-10 mice).

Discussion
Halting the progression of brain damage while preventing and limiting further pathological alterations are critical steps for the treatment of progressive neurodegenerative diseases such as MS (Giunti et al., 2014;Goldmann and Prinz, 2013;van Noort et al., 2011). In this study, we used the cuprizone mouse model of MS to induce highly inflammatory demyelinating lesions in the white matter (Gudi et al., 2009(Gudi et al., , 2014Orije et al., 2015;Remington et al., 2007;Steelman et al., 2012) and subsequently demonstrate that targeted delivery of IL-13 to the lesion site, prior and during lesion induction, induced the expression of markers characteristic for alternative activation both in microglia and macrophages. The latter is here shown to be associated with a protective effect on oligodendrocyte survival and subsequently less severe demyelination during cuprizone administration. These results are fully FIGURE 7: Both microglia and macrophage phenotype is modulated through LV mediated expression of IL-13 in the splenium.(a) eGFP 1 BM chimeric mice were generated to differentiate infiltrating macrophages (green) from brain-resident microglia. At eight week of age mice received an injection of BFP LV, IL-13 LV, or no injection (No inj.) and were fed a cuprizone diet for a period of four weeks to induce inflammation and demyelination. (b) Representative T 2 -weighted images show hyperintense contrast at the level of the splenium of the CC indicative of neuroinflammation and demyelination following four weeks of cuprizone diet, but not in the case of mice that also received the IL-13 LV injection (Arrow indicates the site of injection of IL-13 LV and associated protection against cuprizone). (c) Representative images of each experimental group stained for MHCII (blue) and Arg-1 (red). Scale bar 5 100 lm. (d) Graph shows the mean density 6 RMSD (root mean square deviation) of eGFP 1 infiltrating macrophages into the splenium. *P < 0.05, GEE and Bonferroni post hoc test (n 5 3 mice). (d) Graph shows the mean density 6 RMSD of Arg-1 and MHCII cells. Significant appearance of MHCII 1 , Arg-1 1 and MHCII/Arg-1 1 expressing cells is indicated by the *, # and ‡ symbols, respectively. */#/ ‡ P < 0.05, **P < 0.01, ‡ ‡ ‡P < 0.001, GEE, FDR corrected (n 5 3 mice). [Color figure can be viewed at wileyonlinelibrary.com] FIGURE 8: Therapeutic LV mediated expression of IL-13 in the splenium protects against severe cuprizone induced demyelination and is associated with an altered microglia/macrophage phenotype.(a) Eight week old mice were assigned to a control or cuprizone treated group. To test whether IL-13 LV injection may exert its protective effect during cuprizone induced CNS neuroinflammation and demyelination, a subgroup of mice that received the cuprizone diet for a period of 2.5 weeks were injected with the IL-13 LV. MRI evaluation and histological analysis were performed after 4 weeks of cuprizone administration. (b) T 2 values of individual mice are displayed in the graph with mean 6 SD and reveal a significant increase of the normal T 2 for IL-13 LV injected cuprizone treated mice (n 5 5-15 mice). **P < 0.01, ***P < 0.001, Wilcoxon rank sum test, FDR corrected (n 5 5-15 mice). (c) MTR of individuals are displayed in the graph with mean 6 SD and reveal a lower deviation of the normal MTR of IL-13 LV injected cuprizone treated mice (n 5 5-15 mice). **P < 0.01, ***P < 0.001, Wilcoxon rank sum test, FDR corrected (n 5 5-15 mice). (d) Histological analyses performed at the level of the splenium in mice that received the IL-13 LV injection after 2.5 weeks of cuprizone diet. Representative images showing the presence of F4/80 1 inflammatory cells (red), the MHCII/Arg-1 phenotypic properties of inflammatory cells (green and red, respectively), the presence of astrogliosis (GFAP, red) the presence of myelin (MBP, red) and the presence of CC11 oligodendrocytes (red). n 5 5-8 mice analyzed per group. Scale bar 5 100 lm. Quantitative analysis confirming significantly reduced F4/80 1 inflammatory responses (e), alternative activation of microglia/macrophages (f), reduced astrogliosis (g), reduced demyelination (h), and oligodendrocyte survival (i) at the level of the splenium in mice that received the IL-13 LV injection after 2.5 weeks of cuprizone diet. Graphs e, g, h, and I display mean 1/2 SD, *P < 0.05, ***P < 0.001, One-way ANOVA and Tukey post hoc test (n 5 5 -8 mice). Graph (f) displays mean 1/-SD for the density of Arg-1 1 and MHCII 1 cells. Significant appearance of Arg-1 and MHCII/Arg-1 expression is indicated by the # and ‡ symbols, respectively. ## P < 0.01, ‡ ‡ ‡P < 0.001, One sample t-test (n 5 5-8 mice). [Color figure can be viewed at wileyonlinelibrary.com] in line with current understandings of neuro-inflammatory processes in health and disease (Cherry et al., 2014;Murray et al., 2014;Rawji and Yong, 2013), which suggest that modulating, rather than suppressing, microglial activation may be a more adequate therapeutic approach (David and Kroner, 2011;Tang and Le, 2016). Because microglia share functional similarities with peripheral macrophages, microglial activation and polarization state is generally classified using the established macrophage's nomenclature, that is, classically activated, also referred as M1-like phenotype, or alternatively activated, also referred as M2-like phenotype, which in turn comprises several subtypes of cell activation (David and Kroner, 2011;Hu et al., 2012). In this study, we report on the in vivo induction of arginase1 expressing alternatively activated microglia and macrophages following LV mediated delivery of IL-13, similar to IL-4 a well described inducer of this specific subset of alternatively activated microglia and macrophages (Franco and Fernandez-Suarez, 2015;Orihuela et al., 2016). Of note, although not described in detail in this manuscript, a similar set of experiments using an IL-4 encoding LV also demonstrates similar protection against demyelination upon injection into the splenium of cuprizone treated mice (Supp. Info. Figure S3). Following IL-13 LV administration, as well as following IL-4 LV administration, the alternatively activated state was associated with the prevention of the cuprizone induced pathology. The latter may not be unexpected as IL-4 and IL-13 share common receptor components (Kelly-Welch et al., 2003;Van Dyken and Locksley, 2013), but may display divergent effects on microglia and macrophages. While macrophages generally are directed toward the M2 activation state (Dhakal et al., 2014;Murray et al., 2014;Orihuela et al., 2016), microglia may additionally be forced into apoptotic death under certain conditions following IL-4/ 13 stimulation (Won et al., 2013;Yang et al., 2002). At the time mentioned, we did not observe obvious differences between IL-13 and IL-4 mediated neuroprotection, and therefore decided to narrow down this study to the administration of IL-13, without specific preference.
In-situ delivery of compounds can be achieved through several means, including direct intracerebral injection of the molecule of interest (Kawahara et al., 2012), inoculation of nano-sized vehicles such as liposomes or nanoparticles containing the selected compound (Malam et al., 2009), direct engraftment of engineered cells producing the protein of interest (Stuckey and Shah, 2014;Tan et al., 2005), or gene transfer using altered machinery of viruses (Houghton et al., 2015). Following our approach using LV injections in brains of healthy and cuprizone treated mice, several interesting observations were made. As observed on Figure 3, LV injections resulted in a slight activation of microglia and macrophages (Fig. 3c, F4/80) and upregulation of MHCII expression in case of BFP LV injections in control mice (Fig. 5a,c,MHCII), most likely illustrating the response to the viral particles and the attempt to eliminate them. Remarkably, in case of cuprizone administration, the vast majority of microglia and infiltrating macrophages at the level of the LV injection showed upregulation of MHCII, both in case of BFP or IL-13 LV injections (Figs. (5 and 7)c, 7e, and 8e), while only very little MHCII expression was observed in case of cuprizone treated mice that did not received a LV injection. These alterations in the proinflammatory phenotype following cuprizone administration in BFP LV injected animals was further confirmed by the significant upregulation of TNF-a mRNA (Fig. 6d) and, albeit at a lower degree, the significant upregulation of IL-1b and iNOS mRNA (Fig.  6e,f ). As the latter qRT-PCR analyses were performed on the dissected splenium as a whole, we cannot distinguish whether the detected TNF-a, IL-1b, or iNOS mRNA transcripts were produced by microglia, macrophages, or astrocytes. Nevertheless, these results are in line with previous studies showing that LV transduction can elicit a transient pro-inflammatory response against the LV particles (Kay, 2011), mediated by the activation of toll-like receptor 7 and or toll-like receptor 9, which, respectively, recognize single stranded RNA and unmethylated CpG (Matrai et al., 2010). Consequently, the exacerbated immune response observed following cuprizone administration in BFP LV injected mice can be explained by a possible priming effect of innate immune cells following LV infection and/or by the high infiltration of peripheral macrophages, caused by a blood brain barrier breakdown at the moment of the intracerebral surgery (Fig. 7d), thereby modifying the lesion environment as compared to mice that did not received a LV injection (no mechanical blood-brain barrier damage and limited recruitment of peripheral macrophages). Likewise, we cannot exclude that priming of innate immune cells following LV infection may be the key to successful IL-13 mediated deviation from classically activated towards alternatively activated microglia and macrophages, the latter upregulating, next to MHCII (Zurawski and de Vries, 1994), also arginase1, YM-1 and galectin-3 (Figs. 5b,c, 6a-c, 7c,e, and 8e), which we associate with the beneficial effect on oligodendrocyte survival and myelin preservation. Interestingly, IL-13 was also able to counteract the upregulation of TNF-a mRNA as a consequence of LV injection. This is highly interesting in view of the observation that in TNF-a knock-out mice cuprizone administration results in a higher survival of oligodendrocytes and preservation of myelin up to 4 weeks after the start of the cuprizone diet (Arnett et al., 2001). Furthermore, our observations that IL-13 (and IL-4 immune gene therapy, Supp. Info. Figure S3) is able to exert beneficial effects on neuro-inflammation and demyelination in the cuprizone mouse model is also further supported by a recent publication by Janssens et al. showing that LV mediated delivery of oncostatin M (OSM) can induce alternatively activated microglia and macrophages via the upregulation of IL-4 expression, resulting in protection against cuprizone induced demyelination (Janssens et al., 2015).
Clearly, induction of a phenotypic and/or functional state of alternative activation in microglia and macrophages is of high relevance in the process of inhibiting and/or modulating neuroinflammation. However, to date it remains to be investigated in depth how IL-4/13 mediated alternative activation of microglia and macrophages displays protective features in vivo. In this context, we recently demonstrated that in vivo IL-13 priming of microglia and macrophages, with subsequent induction of F4/80, MHCII, arginase1, YM1 and Fizz1 expression, leads to highly reduced downstream effector function in terms of (i) direct in vivo recognition and elimination of allogeneic cellular grafts, and (ii) in vivo induction of allogeneic T-cell immune responses (Hoornaert et al., 2016). As such, it is clear that in vivo IL-13 primed microglia and macrophages display reduced and/or distinct effector functions as compared to pro-inflammatory microglia and/or macrophages phenotypes (in our studies defined as F4/801 MHCII1 without expression of arginase1). Most widely suggested in literature is the competitive action of arginase1 with iNOS, thereby resulting in decreased NO production subsequently being less neurotoxic. Although this is certainly true for in vitro experiments in which microglia and macrophages can be primed towards the spectral ends of M1 and M2 activation by high concentrations of LPS/IFNg and IL-4/13, respectively, qRT-PCR experiments performed in this study (Fig. 6) did not reveal significant iNOS expression in the CC of CPZ-treated mice in the absence of LV injection, thereby ruling out this mode-of-action in vivo for IL-13 mediated alternative activation of microglia and macrophages in the cuprizone model. Again this underscores our limited functional understanding of in vivo alternative activation of microglia and macrophages and warrants in depth in vitro and in vivo phenotypic and functional investigation of M1/ M2 intermediate subtypes.
Although the main aim of this study was to investigate the influence of IL-13 immune gene therapy on inflammation and demyelination, several important aspects still remain to be elucidated. First, electron microscopy studies can provide additional information with regard to myelin integrity following IL-13 LV administration.
Although not yet performed, we do not expect to observe normal myelin structure as still a significant degree of demyelination is present, as well as CC1 1 oligodendrocyte death as compared to healthy control mice. Secondly, further investigation should also reveal whether alternatively activated microglia and macrophages stimulated by the addition of IL-13 play a beneficial role in the repair process and remyelination once cuprizone administration is halted. While in our hands no classical M1or M2-associated phenotypic markers can be found in the cuprizone model, a few studies have shown that microglia and macrophages display a distinct phenotype supportive of remyelination (Gudi et al., 2011;Morell et al., 1998;Olah et al., 2012;Voss et al., 2012). Along with the secretion of anti-inflammatory factors, growth factors and neurotrophic factors, there is accumulating evidence that phagocytosis of myelin debris is an essential step for a successful remyelination and that this step is ensured by alternatively activated polarized microglia and macrophages (Cantoni et al., 2015;Lampron et al., 2015;Poliani et al., 2015;Skripuletz et al., 2013). Furthermore, in the past years, numerous studies have reported the beneficial effect operated by the shift from M1to M2-polarized immune cells in other preclinical models of MS and other models of neuroinflammation (Cash et al., 1994;Janssens et al,. 2015;Liu et al., 2013;Miro-Mur et al., 2016;Miron et al., 2013;Morganti et al., 2015;Ponomarev et al., 2007;Yu et al., 2015;Zhao et al., 2015). Specifically, narrowing down on the distinct roles of microglia and macrophages in natural or therapeutic neuroprotection and/or repair, it will be of utmost importance to carefully design future studies allowing discrimination between both cell populations. Especially for the cuprizone mouse model, prior studies using BM chimeric mice have reported that the population of monocyte-derived macrophages recruited to the CC following cuprizone administration represents 5-30% of the total number of microglia/macrophages (Lampron et al., 2015;Remington et al., 2007;Voss et al., 2012) in the absence of BBB disruption (Arnett et al., 2001;Matsushima and Morell, 2001). Similarly, we used BM chimeric mice to discern whether resident microglia and infiltrating monocytederived macrophages were equivalently affected by IL-13 stimulation and found that both populations displayed an alternatively activated phenotype (Fig. 7c,e). Recently, it has, however, been clearly shown that radiation BM chimeric mice display an altered BBB permeability to systemic monocytes (Morganti et al., 2014), suggesting that certain results need caution during interpretation as the function and recruitment of peripheral macrophages may slightly differ from nonirradiated noninjected animals. Nevertheless, our results were already able to demonstrate that IL-13 act on both microglia and infiltrating macrophages, while future experiments to be performed in the CX3CR1 eGFP/1 CCR2 RFP/1 transgenic mouse model would surely help in further addressing these issues (Saederup et al., 2010).
Finally, we would like to elaborate a little further on our choice to noninvasively detect white matter lesions using multiparametric MRI. To date, MRI represents the method of choice for the evaluation and monitoring of demyelinating lesions over time in patients suffering from MS (Polman et al., 2011). We used T 2 relaxometry, a method routinely used in clinical settings (Poloni et al., 2011;Wattjes et al., 2015), in combination with a newly described method, MTI that is sensitive to the integrity of macromolecules, notably myelin in the CNS (Grossman, 1994;Ou et al., 2009;Turati et al., 2015). We demonstrated that both T 2 and MTR measures are highly correlated to the brain microstructure and underlying histopathological events following cuprizone administration (Supp. Info. Figures S1 and S2). Although a few studies have associated MTR values to oligodendrocytes numbers, myelin, and GFAP positive astrocytes (Boretius et al., 2012;Fjaer et al., 2013;Merkler et al., 2005;Thiessen et al., 2013;Zaaraoui et al., 2008), we were unable to ascribe a specific MRI metric to a specific histological event. This is likely due to the fact that we used a four week cuprizone regime in which neuroinflammation and demyelination are both reaching maximal values while other studies followed cuprizone induced changes over time, hence enabling to differentiate between sequential cellular events. Additionally, we here demonstrate that both T 2 and MTR measures are suitable to detect alterations in neuroinflammation and associated demyelination following IL-13 LV therapeutic intervention. Moreover, we also show that MTI improved the detection of tissue alterations as compared with T 2 , as MTI values reached significance when T 2 displayed a trend (Figs. 4a,b and 8c,d), indicating a lower sensitivity threshold for MTI and highlighting its potential for future use in clinical practice. Moreover, by correlating T 2 and MTI measurements with quantitative histological analyses, the multimodal character or this study clearly contributes to the validity of the observed therapeutic effect of IL-13 immune gene therapy.
In conclusion, the results presented in this study reveal a neuroprotective role of IL-13 in demyelinating lesions through the induction of alternatively activated microglia and macrophages. The ability of IL-13 to trigger the switch from a proinflammatory environment towards an antiinflammatory environment, and thus limit CNS-induced pathology, is of particular interest for the design of new drug based therapies aiming at modulating microglial function to preserve tissue homeostasis and support remyelination in MS and other demyelinating diseases. In addition, these findings emphasize the beneficial action of alternatively activated immune cells that certainly deserve to be investigated in other proinflammatory-driven CNS pathological conditions, such as neurotrauma, spinal cord injury, or stroke.