MyD88 represents a miR-669c-3p target after ischemic stroke

Ingenuity Pathway Analysis (IPA, Qiagen Inc.) revealed that miR-669c-3p is likely involved in pathways related to neuroinflammation and, in particular, implicated in Toll-like receptor signaling (Fig. 7). Therefore, to find potential targets for miR-669c-3p after stroke, we carried out a miRNA pulldown assay in BV2 cells and searched for likely targets in the TLR pathway. These efforts revealed the MyD88 transcript downregulated in miR-669c-3p overexpressing BV2 cells (Fig. 8A: -31% comparing

to control BV2s, p=0.0101) and confirmed MyD88 as a direct target of miR-669c-3p in BV2s (Fig. 8B: +5362% increase of pulled down MyD88 transcript in biotinylated miR-669c-3p transfected cells, p=0.0002). The other direct targets revealed by pulldown assay for miR-669c-3p in BV2s were MMP9 (Fig. 8C: +272%, p=0.0013) and TNF-

transcripts (Fig. 8D: +218%, p=0.0016). Even though the levels of MyD88 were reduced in miR-669c overexpressing BV2 cells, the expression of other members of TLR signaling pathway predicted as miR-669c-3p targets, TLR4 and IRAK4, was not altered and they were not targeted by miR-669c-3p (data not shown). MyD88 transcript was also confirmed as a direct target of miR-669c-3p in N2a cells (Fig. 8E:

+786% increase of pulled down MyD88 transcript in biotinylated miR-669c-3p transfected cells, p<0.0001). Confirming our findings, MyD88 expression was reduced in miR-669c overexpressing N2a cells (Fig. 8F: -56% comparing to control N2a cells, p=0.0007).

After identifying MyD88 transcript as miR-669c-3p target in vitro, we determined its expression in brains of stroked mice. MyD88 immunoreactivity, increased in the ipsistriatal area of LV1-GFP animals after stroke in comparison to LV1-GFP sham animals (data not shown, p=0.029), was attenuated in miR-669c overexpressing stroke animals (Fig. 8G-M: -28% comparing to control stroke animals, p=0.0478).

5.2 SYSTEMIC IL-13 TREATMENT PROMOTES ANTI-INFLAMMATORY MICROGLIAL AND MACROPHAGE ACTIVATION, AND PROVIDES NEUROPROTECTION IN MOUSE CEREBRAL ISCHEMIA (II)

Study (II) aimed to elucidate the protective role of IL-13 in cerebral infarction in a mouse model of pMCAo with subsequent IV administration of murine recombinant IL-13. We employed immunohistochemical staining, CBA, qPCRs and behavioral outcome assessments with adhesive removal and CatWalk gait tests to investigate the actions of this anti-inflammatory cytokine in vivo, while in vitro assays included determining IL-13 effects on murine primary microglia cultures and co-cultures of N2a cells with RAW 264.7 macrophages.

5.2.1 IL-13 treatment alleviates ischemic brain damage

To assess the neuroprotective role of IL-13 in cerebral ischemia, mice received either vehicle or IL-13 at a dose of 1, 2 or 5 µg per animal immediately after pMCAo. MRI of brains at 3 dpi revealed a decrease of infarct volume with all tested IL-13 doses in comparison to vehicle-treated mice. IL-13 dose of 1 µg reduced the ischemic lesion size most significantly (Fig. 1A: -12% comparing to vehicle-treated mice, p=0.0091, Fig. 1B: -23%, p=0.0327 and -23%, p=0.0314).

5.2.2 IL-13 treatment reduces leukocyte infiltration to the lesion area without altering peri-ischemic astrogliosis

Analysis of GFAP immunoreactivity confirmed astrogliosis after ischemic stroke in the peri-ischemic area, which was not influenced by IL-13 delivery (Fig. 2A, p=0.2618). Infiltrating leukocytes express high levels of CD45, which is present only at low levels in resident microglial cells[314]. We readily detected CD45⁺ cells, indicative of infiltrating leukocytes, in the infarct core of ischemic mice at 3 dpi, which was markedly decreased after IL-13 injection (Fig. 2F: -17.5% comparing to vehicle-injected mice, p=0.0461).

5.2.3 IL-13 enhances anti-inflammatory immune responses in the ischemic brain

We found no differences in Iba1 expression between the vehicle and IL-13-treated mice at 3 dpi in the peri-ischemic area of stroke lesions (Fig. 3A-E, p=0.5606).

Microglia commonly acquire a proinflammatory phenotype (M1), but under specific conditions, including exposure to IL-13, polarize towards an anti-inflammatory M2 phenotype. Using Arg1 IHC staining, we determined if the protective effect of IL-13 administration in pMCAo is associated with a phenotypic shift of microglia/macrophages from M1 to M2. Indeed, as shown in Fig. 3F-H, IL-13 promoted an increase of M2-type microglia/macrophages after ischemic stroke in the ipsilateral hemisphere when compared to vehicle treatment (+38%, p=0.0381). IL-13 treatment did not alter cleaved caspase-3 levels in the peri-ischemic area, indicating that it does not exert its beneficial effect by preventing apoptosis (data not shown).

5.2.4 IL-13 increases expression of M2 markers in the ischemic brain and elevates anti-inflammatory cytokine levels in the plasma

qPCR analysis revealed upregulation of alternative activation markers Arg1 (Fig. 4A:

+257% in comparison to PI area in vehicle-treated mice, p=0.0408) and Ym1 (Fig. 4B:

+200%, p=0.0122) after IL-13 administration in peri-ischemic areas at 3 dpi. In addition, in comparison to vehicle-treated ischemic mice, IL-6 levels were increased in mice treated with IL-13 (Fig. 4C: +164%, p=0.0055). On the other hand, IL-10 levels were not altered by the IL-13 administration (Fig. 4D, p=0.1477). Phagocytosis-associated Mer receptor tyrosine kinase (Mertk) expression levels in peri-ischemic areas were also unaffected by IL-13 treatment (Fig. 4E, p=0.0757), however a moderate, but nearly significant increase of Gal3 (Lgals3) transcript levels was found in IL-13 mice (Fig. 4F: +100%, p=0.0594). Gal3 deficiency has been reported to increase neuronal apoptosis and ischemic lesion volume by impairing microglial responses to the injury[120,316]. CBA analysis at 3 dpi revealed upregulation of both 6 and IL-10 protein levels in blood plasma of IL-13 treated mice (Fig. 4G: +75.5% in comparison to vehicle-treated stroke animals, p=0.0448 and 4H: +31.5%, p=0.0341).

5.2.5 IL-13 improves somatosensory and locomotor impairments in mice after pMCAo

The adhesive removal test was used to detect mouth and forepaw sensitivity,

treatment decreased the time needed to remove the adhesive from the contralateral front paw at 14 dpi, the late stage after ischemic stroke (Fig. 5A: -63% in comparison to vehicle-treated stroke mice, p=0.0345). Moreover, the stroke-induced delay of removing the adhesive between contra- and ipsilateral forelimbs was shorter after IL-13 administration (Fig. 5B: -77%, p=0.0517).

Next, we used the CatWalk automated gait analysis system to evaluate locomotor problems following ischemic stroke. At 7 and 14 dpi, Diagonal Support, which is predominant in healthy conditions and accounts for 60–70% of support types[317], was improved in the IL-13 treated group compared to controls (Fig. 5C: +31%, p=0.0015 and +24%, p=0.0094, respectively). Some of the gait parameters of IL-13-treated animals were exclusively improved at 7 dpi. Support Three, a parameter describing the number of paws used to support body weight during the Step Cycle, was decreased in the IL-13 treated group (Fig. 5D: -42%, p=0.0349), and Duty Cycle (Fig. 5E: -12%, p=0.0195), defined as the Stand calculated as a the percentage of Step Cycle[318]. The Initial Dual Stance of the Right Front paw (Fig. 5F: -67%, p=0.0465), Terminal Dual Stance of the Left Front paw (Fig. 5G: -73%, p=0.0253) and Stand of the Left Hind paw (Fig. 5H: -20%, p=0.0242) were also reduced in IL-13-treated animals. Kinetic parameter Swing Speed of the Right Hind paw, defining the velocity when paw is not in contact with the surface, was higher in the group of IL-13-treated mice (Fig. 5I: +14%, p=0.0336). The Support Girdle, an interlimb coordination parameter describing the relative duration of the contact with the surface of two paws at the same time (Fig. 5J: -70%, p=0.0129), and Terminal Dual Stance of the ipsilateral hindlimb (Fig. 5K: -60%, p=0.0086) were decreased at 14 dpi by IL-13 administration. Finally, also at 14 dpi, the Stride Length of the ipsilateral front paw, describing the distance between the consecutive placements of a paw, was increased in the IL-13 group (Fig. 5L: +16%, p=0.0125). In summary, IL-13-treated mice no longer required to compensate abnormal movements of the contralateral limbs and showed improved coordination patterns during the step cycle[318].

5.2.6 IL-13 promotes alternative polarization of primary microglia under proinflammatory conditions

We first assessed viability of microglia exposed to either IFN-γ/LPS (referred as M1) alone or in combination with IL-13. Neither the metabolic capacity to reduce MTT (Fig. 6A, p=0.0871) nor LDH enzyme secretion (Fig. 6B, p=0.3374) was altered by IL-13 exposure, indicating that IL-IL-13 treatment does not impact cell viability of activated microglia. qPCR analysis of primary microglial cultures showed a marked increase of M2-type marker Arg1 transcripts (Fig. 6C: +1270% in comparison to M1-treated cells, p=0.0001) and decrease of proinflammatory IL-1β (Fig. 6E: -54.5%, p<0.0001) as response to IL-13. While the expression level of another alternative activation marker Ym1 was significantly decreased with IL-13 as compared to M1 alone (Fig. 6D: -49%, p=0.0001), it was still upregulated in comparison to vehicle-treated cells.

Interestingly, IL-13 combined with M1 exposure increased MerTK (Fig. 6F: +69%, p<0.0001) and Lgals3 (Fig. 6G: +38%, p=0.0255) levels, which have been shown to be involved in the microglial phagocytosis[319,320].

5.2.7 IL-13 protects N2a cells from inflammation-induced death in co-culture with RAW 264.7 macrophages

Finally, in order to evaluate the neuroprotective actions of IL-13 in neuroinflammation, we exposed N2a and RAW 264.7 co-cultures either to vehicle, LPS and IFN-γ (M1), or M1 combined with IL-13 (Fig. 7A-D). As shown by the FACS analysis, 48 h of M1 exposure increased N2a cell death by 26% compared to vehicle-treated cells (p<0.0001), which was attenuated by IL-13 (-12% in comparison to M1 treatment, p<0.0001). Additionally, NO production in N2a and RAW 264.7 co-cultures exposed to proinflammatory conditions was decreased by IL-13 treatment (Fig. 7E: -40%, p<0.0001).

5.3 HX600, A SYNTHETIC AGONIST FOR RXR-NURR1 HETERODIMER COMPLEX, IS NEUROPROTECTIVE IN A MOUSE MODEL OF ISCHEMIC STROKE (III)

In study (III) we tested the hypothesis that the drug HX600 facilitates neuronal survival by inhibiting inflammation via activation of NURR1 receptors. We tested this hypothesis in an in vitro model of inflammation-induced neuronal death and in vivo in mice undergoing pMCAo using several methods, including immunohistochemical staining, ex vivo analyses of microglial morphology, flow cytometry analyses and metabolic profiling using UHPLC combined with mass spectrometry.

5.3.1 HX600 treatment decreases expression of inflammatory mediators in primary microglia and protects neurons against inflammation-induced cell death when co-cultured with microglia

In order to see whether HX600 decreases microglial proinflammatory activation, primary microglia were pretreated with HX600 for 24 h followed by LPS challenge for 3 h. After that, mRNA levels of inflammatory mediators were analyzed by qPCR.

HX600 exposure reduced NOS2 (iNOS), Marco, Il1b (IL-1β), Il6 (IL-6) and Mmp9 levels in treated primary microglial cells (Fig. 2A: -65% comparing to LPS-treated microglia, B: -63%, C: -88%, D: -72%, E: -41%). Next, to assess whether HX600 is able to convey neuroprotection, we exposed in vitro co-cultures of primary microglia or BV2 cells with primary neurons to LPS and IFN-γ. Pretreatment of co-cultures with HX600 prevented inflammation-induced neuronal death as determined by quantification of MAP2 immunoreactivity (Fig. 3A, B: +37% in viability comparing to LPS/IFN-γ-treated co-cultures). To verify that HX600 beneficial effect is mediated through NURR1 induction, BV2 cells were transfected with Nr4a2 siRNA, which resulted in 40% knockdown of NURR1 expression levels (Fig. 3C). Nr4a2 siRNA abolished HX600-mediated protection against inflammation-induced neuronal death in BV2-neuron co-cultures (Fig. 3D), while mock transfected cultures were still protected (Fig. 3E: +107% in viability comparing to LPS/IFN-γ-treated co-cultures).

5.3.2 HX600 treatment alleviates brain damage and provides functional improvement after pMCAo

HX600 was administered to stroked mice immediately after ischemia induction and thereafter every 24 h. Quantification of infarct volumes at 48 h post-stroke showed that HX600 administration decreased the lesion size by 21% in comparison to the vehicle-treated mice (Fig. 4A-C). Ischemic animals displayed impaired locomotor activity at 1 dpi, since pMCAo caused the latency to move time to double, comparing to sham-operated animals. In mice treated with HX600 these impairments were no longer evident (Fig. 4D).

5.3.3 HX600 administration decreases Iba1, phospho-p38 and TREM-2 levels in the infarcted brain

By immunohistochemistry we evaluated the effect of HX600 treatment on post-stroke inflammatory activation. HX600 reduced Iba1 by 21% (a marker for activated microglia), phospho-p38 by 27% (activated by e.g. inflammatory cytokines) and TREM2 by 52% (phagocytosis marker) immunoreactivities in the ischemic area at 1 dpi (Fig. 5). Next, we carried out quantitative analysis of microglial morphology, which demonstrated that HX600 administration increases the Iba1⁺ cell perimeter by 15% and reduces the Area/Perimeter ratio by 6% (Fig. 6).

5.3.4 HX600 normalizes the proportion of CD45^Hi CD11b^Low and Ly6C^Hi CD45^Low cells in the ischemic brain

To more specifically investigate immune cell populations in ischemic brain tissue, we evaluated the proportions of different immune cell types using FACS analyses from freshly isolated cells at the peak of macrophage infiltration (3 dpi)[304]. In comparison to sham operated mice, ischemic stroke induced a 59% increase in infiltrating CD45^Hi CD11b^Low lymphocytes (Fig. 7A,B) and 53% increase in Ly6C^Hi CD45^Low cells, which can be either infiltrating monocytes or activated microglia (Fig.

7C). HX600-treated animals exhibited a decreased level of these infiltrating cell populations. Of note, cerebral ischemia caused a small, yet significant loss of resident CD45^Low F4/ 80^Hi Ly6C^Low microglia, which was prevented by HX600 administration (Fig. 7D: +6% of microglia in comparison to vehicle-treated stroke animals).

5.3.5 HX600 treatment normalizes the metabolic profile in the peri-ischemic area

We carried out high-throughput metabolic profiling of the peri-ischemic areas and the corresponding contralateral brain samples of HX600 and vehicle-treated ischemic mice at 1 dpi (Fig. 8 and 9). HX600 treatment normalized the stroke-affected levels of several metabolites (Fig. 9). HX600 specifically alleviated the ischemia-evoked increase in levels of acylcarnitine species: 10:0, 16:0, 16:3, 18:0, 18:1, 18:2 and 18:3 (Fig.

9). Similarly, the stroke-associated increase in levels of lysoPC species 14:0, 16:0, 16:1, 18:0 and 18:1, as well ADP-ribose in HX600-treated mice was normalized to levels in the contralateral hemisphere.

6 DISCUSSION

6.1 OVEREXPRESSION OF miR-669c IN THE BRAIN PROTECTS MICE AGAINST ISCHEMIC STROKE BY ENHANCING MICROGLIAL ANTI-INFLAMMATORY RESPONSES

The purpose of the study (I) was to elucidate how miR-669c, a member of C2MC, is implicated in ischemic stroke pathology, and further to assess the role of this miRNA in regulation of neuroinflammatory events. We demonstrate that miR-669c-3p is ischemia-inducible, intracerebral overexpression of this miRNA molecule modulates microglial polarization, and that it targets MyD88, a key adaptor protein involved in toll-like receptor (TLR) signaling. Lentiviral-mediated overexpression of miR-669c provided protection in a mouse transient ischemia model through reduction of microglia proinflammatory responses and concurrent microglial phenotypic switch towards M2a activation.

Cerebral ischemia provokes a rapid inflammatory response, which is initiated and aggravated by brain-resident microglia, but also involves the recruitment of circulating peripheral immune cells into the injured brain. A growing body of literature indicates that modulation of inflammatory responses is protective in cerebral ischemia and this protection is often mediated by an enhancement of microglial/macrophage alternative M2a polarization[209,260,321,322]. One of the typical features of the M2a phenotype is the increased expression of Arg1[207,260,323–325], which is involved in endogenous tissue repair processes[326,327] by promoting extracellular matrix remodeling[313], as well as supporting axonal growth and neuronal survival[328]. The majority of Arg1⁺ cells in pMCAo have been demonstrated to be infiltrating macrophages[329], which may play a crucial role in neuroprotective responses early post-stroke[330]. In addition, in tMCAo miRNA-mediated Arg1 induction in microglia and macrophages is neuroprotective and provides functional improvement[260]. In our study, the beneficial effects in miR-669c overexpressing animals were also linked to an increase of Arg1 levels in the ipsilateral hemisphere. As evaluated by the Arg1/CD45 double immunostaining, the Arg1⁺ cells appeared to be mainly CD45^low, indicating that they primarily are brain-resident immune cells, microglia.

Some studies have pinpointed miRNAs ability to regulate neuroinflammatory responses for therapeutic benefit[331]. Either inhibition or overexpression of miRNAs can modulate microglial activation and/or leukocyte infiltration, as demonstrated by antagomirs for miR-22[332] and miR-181a[247] and overexpression of miR-203[250]. Inhibition of miR-3473b[254] or upregulation of let-7c-5p[245] were shown to decrease microglial activation in vitro and in vivo, and provide protection in a mouse model of cerebral ischemia. Our data indicate that even though miR-669c overexpression did not prevent OGD/R-induced cell death in neuronal N2a cells,

which is likely attributed to lack of glia-mediated protection in pure neuronal cultures, overexpression of miR-669c decreased microglial proinflammatory responses. We also detected enhanced levels of alternative activation markers, denoting microglial skewing towards a M2a neuroprotective phenotype both in cell culture and mice. These observations were accompanied by a decrease in lesion volume and improvement of neurological deficits in mice.

669c is understudied in the brain, and we are the first to demonstrate that miR-669c-3p is increased upon ischemic stroke and has a role in regulation of neuroinflammatory events. Kuypers et al. found that the Sfmbt2 miRNA cluster participates in regulation of oligodendrocyte proliferation and remyelination[265].

In support of our data, C2MC miRNAs were shown to be upregulated in a rat model of tMCAo[333]. Interestingly, miR-669c-3p interacts with a network of circular RNAs in transient brain ischemia in mice[334], however a specific role of miR-669c-3p in ischemic stroke has not been investigated. Druz et al. showed that the Sfmbt2 gene and miR-669c-3p are upregulated upon glucose deprivation-induced oxidative stress and suggested that C2MC plays a role in diseases characterized by increased oxidative stress[335,336]. Similar to Druz et al., our data demonstrated that miR-669c-3p is upregulated in both primary neurons and N2a cells exposed to glutamate or OGD/R, as well as during cerebral ischemia in mice. However, miR-669c overexpression did not aggravate OGD/R-induced neuronal death, suggesting that upregulation of miR-669c may have other functions in isolated neurons. Indeed, our study indicates that rather than being an intrinsic protective effect, miR-669c regulates microglial inflammatory responses to then in turn impact neuronal survival. Overexpression of miR-669c in microglial BV2 cells increased levels of M2a activation-specific markers Arg1, Chil3 (Ym1) and Pparg (PPAR-γ). Ym1 interacts with heparin/heparan sulfate proteoglycans and CD206a C-type lectin carbohydrate-binding protein present on alternatively polarized microglia or macrophages, facilitating extracellular matrix remodeling and repair processes[337]. Peroxisome proliferator activated receptor-γ activation has already been extensively described to promote M2 polarization in multiple diseases with neuroinflammatory components[338–341]. Interestingly, PPAR-γ directly binds to a distal enhancer in the Arg1 genomic region and thereby participates in regulation of alternative macrophage activation[342]. It is possible that this mechanism partially accounts for the beneficial effects observed in our model. On the other hand, we detected a decrease of several proinflammatory mediators in miR-669c overexpressing, LPS-challenged BV2 cells, Mmp9, Tnfa (TNF-α), Il6 (IL-6), Il1b (IL-1β), and Ccl2 (MCP-1), which are all upregulated in ischemic stroke[343,344]. Hence, the miR-669c-3p-mediated change of microglial responses may also be a result of an endogenous protective mechanisms through inhibition of the MyD88 transcript.

MyD88 acts as key downstream adaptor molecule in TLR/NF-κB and IL-1/IL-1R1 signaling pathways[345]. It is crucial for induction of the proinflammatory gene expression by IL-1, IL-18 and all TLRs, except TLR3, by activation of transcription factors NF-κB and AP-1. MyD88 oligomerization leads to the recruitment of IRAK

family members through death domain interactions[346]. It has been shown in the literature that exaggerated or prolonged TLR-mediated inflammatory responses can lead to worsened pathology in sepsis, myocardial ischemia and ischemic stroke[347–

349]. MyD88 signaling regulates leukocyte infiltration after brain injury and plays an important role in modulation of early proinflammatory gene expression in this pathology[350]. Disruption of the TLR/MyD88/NF-κB signaling pathway protects from myocardial injury by attenuating NLRP3 inflammasome activation[351].

Inhibition of MyD88 signaling is beneficial in a number of CNS pathologies, e.g.

EAE[352], neuropathic pain[353], traumatic brain injury[354,355], epilepsy[356,357], Alzheimer's disease[358], hypoxic neonatal brain injury in LPS-sensitized mice[359], subarachnoid hemorrhage[360], and cerebral ischemia[361,362]. Further, hematopoietic cells have a MyD88-dependent neuroprotective function after ischemic stroke [363]. In contradiction to that, MyD88 knock-down in mice subjected to permanent cerebral ischemia did not limit brain infarction in one study[364]. Such results may indicate some specific, not yet well described role for MyD88 in the acute stress during the early phase of innate response to cerebral ischemia[364]. The differential expression of receptors involved in MyD88-dependent signaling in various cell types involved in inflammatory processes increases the complexity of MyD88 role in neuroinflammation. Since it is difficult to develop inhibitors for adaptor proteins[365], miR-669c-3p-mediated inhibition of MyD88 would serve as a beneficial tool to control overactive TLR signaling in neuroinflammation. Of note, miR-669c-induced increase of M2a microglial activation both in vitro and in vivo is partially an effect of MyD88-independent mechanisms, as our preliminary in silico analysis revealed an interaction of miR-669c-3p with Arg1 promoter region.

miR-669c overexpression affected reactive microglia morphology. In healthy conditions, microglial cells adopt a highly ramified shape, which enables them to dynamically inspect the brain microenvironment[89]. After brain injury, microglia

miR-669c overexpression affected reactive microglia morphology. In healthy conditions, microglial cells adopt a highly ramified shape, which enables them to dynamically inspect the brain microenvironment[89]. After brain injury, microglia