Temporal profiling of Kv1.3 channel expression in brain mononuclear phagocytes following ischemic stroke

Background Microglia and CNS-infiltrating monocytes/macrophages (CNS-MPs) perform pro-inflammatory and protective anti-inflammatory functions following ischemic stroke. Selective inhibition of pro-inflammatory responses can be achieved by Kv1.3 channel blockade, resulting in a lower infarct size in the transient middle cerebral artery occlusion (tMCAO) model. Whether beneficial effects of Kv1.3 blockers are mediated by targeting microglia or CNS-infiltrating monocytes/macrophages remains unclear. Methods In the 30-min tMCAO mouse model, we profiled functional cell-surface Kv1.3 channels and phagocytic properties of acutely isolated CNS-MPs at various timepoints post-reperfusion. Kv1.3 channels were flow cytometrically detected using fluorescein-conjugated Kv1.3-binding peptide ShK-F6CA as well as by immunohistochemistry. Quantitative reverse-transcriptase polymerase chain reaction (qRT-PCR) was performed to measure Kv1.3 (Kcna3) and Kir2.1 (Kcnj2) gene expression. Phagocytosis of 1-μm microspheres by acutely isolated CNS-MPs was measured by flow cytometry. Results In flow cytometric assays, Kv1.3 channel expression by CD11b+ CNS-MPs was increased between 24 and 72 h post-tMCAO and decreased by 7 days post-tMCAO. Increased Kv1.3 expression was restricted to CD11b+CD45lowLy6clow (microglia) and CD11b+CD45highLy6Clow CNS-MPs but not CD11b+CD45highLy6chigh inflammatory monocytes/macrophages. In immunohistochemical studies, Kv1.3 protein expression was increased in Iba1+ microglia at 24-48 h post-tMCAO. No change in Kv1.3 mRNA in CNS-MPs was observed following tMCAO. Conclusions We conclude that resident microglia and a subset of CD45highLy6clow CNS-MPs are the likely cellular targets of Kv1.3 blockers and the delayed phase of neuroinflammation is the optimal therapeutic window for Kv1.3 blockade in ischemic stroke. Electronic supplementary material The online version of this article (10.1186/s12974-019-1510-8) contains supplementary material, which is available to authorized users.


Introduction
Ischemic stroke is a leading cause of mortality and disability, and there is an urgent need for therapies that can limit ischemic brain injury beyond the acute phase following ischemic stroke [1][2][3][4]. Neuroinflammation mediated by innate immune cells of the CNS (microglia and CNS-infiltrating macrophages collectively called CNS mononuclear phagocytes or CNS-MPs) comprises of complex pro-inflammatory responses that promote neuronal injury as well as anti-inflammatory/neuroprotective responses [3,5]. Potentially protective antiinflammatory CNS-MP phenotypes predominate in the first 1-2 days following ischemic injury whereas detrimental pro-inflammatory phenotypes increase after 48 h [6][7][8]. Identifying the key regulators of detrimental and protective CNS-MP responses can facilitate discovery of novel targets for ischemic stroke. The recently proven efficacy of endovascular thrombectomy for acute ischemic stroke has created exciting opportunities to translate neuro-immunomodulatory and neuroprotective strategies that target disease mechanisms beyond the acute phase [9,10].
The diverse transcriptional profiles adopted by CNS-MPs have been recently revealed by transcriptomic studies which found an emergence of unique disease-associated microglial (DAM) profiles and downregulation of homeostatic genes in mouse models of chronic neuroinflammation, aging, and neurodegeneration [11][12][13]. More recently, we identified molecular heterogeneity within DAM which is comprised of independently regulated proinflammatory and anti-inflammatory DAM sub-profiles [13]. The potassium channel Kv1.3 was identified as a marker and regulator of pro-inflammatory DAM [13], and inhibition of Kv1.3 channels was found to be beneficial in Alzheimer's disease models as well as in the transient middle cerebral artery occlusion (tMCAO) ischemic stroke model [13][14][15]. Although increased expression of Kv1.3 channels have been confirmed in acutely isolated CNS-MPs by electrophysiology, it is unclear whether microglia and/or CNS-infiltrating macrophages are the primary cell type with increased Kv1.3 channel expression [14].
We have applied a novel flow cytometric assay of functional cell-surface Kv1.3 channels coupled with parallel phagocytic profiling of acutely isolated CNS-MPs [16] at early and delayed timepoints following tMCAO, to define the temporal profiles of Kv1.3 channel expression within CNS-MP subsets.

Animals
Male C57BL/6 J mice (JAX 000664) aged 8-12 weeks were housed in the Department of Animal Resources at Emory University under standard conditions. Institutional Animal Care and Use Committee approval was obtained, and all in vivo studies were performed in strict accordance with the Guide for the Care and Use of Laboratory Animals of the National Institute of Health, and in compliance with the ARRIVE guidelines.

Transient middle cerebral artery occlusion (tMCAO) model
After anesthesia (isoflurane) and exposure of the carotid artery, a 6-0 silk suture was advanced from the external carotid artery into the internal carotid artery until the origin of the middle cerebral artery on the ipsilateral side [19,20]. The suture was then withdrawn after 30 min of cerebral ischemia. Laser Doppler flow was used to monitor cerebral perfusion at baseline, at time of MCAO, immediately after recanalization, and 5 min postrecanalization. Animals with > 75% decreased middle cerebral blood flow and > 75% reperfusion were included. In total, 32 mice were included for flow cytometric analyses (30 min: n = 8, 24 h: n = 7, 48 h: n = 8, 72 h: n = 4, 7 days: n = 5). Sham surgery controls were performed (n = 3 per condition). Transcranial Doppler flow data demonstrating effective MCAO and complete reperfusion for mice included in the study are shown in Additional file 1: Figure S1.

CNS-MP isolation
After tMCAO, mice were euthanized at 30 min, 24 h, 48 h, 72 h, and 7 days followed by cardiac perfusion. CNS-MPs were acutely isolated from ipsilateral and contralateral hemispheres using Percoll density (70%/35%) centrifugation as previously described [16,18]. After myelin removal, mononuclear cells were re-suspended in saline or Trizol. From each hemisphere, half was used for RNA extraction and the other half was used for flow cytometry.

Flow cytometry
Functional cell-surface Kv1.3 channel expression in acutely isolated CNS-MPs was measured using a validated fluorescein-conjugated ShK-F6CA assay [13,18,21]. Cells were incubated with 10 nM ShK-F6CA along with fluorophore-conjugated CD11b (APC-Cy7), CD45 (PE-Cy7), and Ly6c (PE) antibodies for 30 min and then washed twice with cold PBS to remove any unbound antibodies. Compensation controls were run along with appropriate negative/isotype controls [18]. Live mononuclear cells were first gated based on forward and side scatter profiles, and then single cells were gated and then further gated for CD11b + CNS-MPs. CD11b + CNS-MPs were gated into CD45 low Ly6c low , CD45 high Ly6c low , and CD45 high Ly6c high subpopulations. Each CNS-MP subpopulation [22,23] was evaluated for ShK-F6CA labeling.

Immunohistochemical studies
Following 30 min MCAO, mice were euthanized at 1 h, 24 h, and 48 h timepoints and brains were post-fixed in 4% PFA for 24 h, then transferred to 30% sucrose for another 24 h, and then sectioned (30 μm) and preserved in cryoprotectant solution. Free-floating sections were blocked with 10% horse serum for 1 h, then incubated with primary antibodies (anti-Kv1.3 mAb 1:100 UC Davis/NIH NeuroMab clone L23/27, anti-Tmem119 rabbit mAb 1:100 #ab209064, or anti-Iba1 rabbit mAb 1:300 #ab178846) at 4°C overnight. Sections were then washed and incubated with fluorophore-conjugated secondary antibodies (1:500) for 30 min and then mounted on slides and dried. Hard-mounting medium (with DAPI-VectorLabs # H1500) was used to mount sections which were then imaged on an immunofluorescence microscope (Microscope: Olympus BX51 and camera: Olympus DP70) at 20× and 60× (oil immersion) magnifications. The ipsilateral (stroke) hemisphere was marked prior to sectioning. At least five fields in the peri-infarct region ipsilaterally and corresponding region on contralateral hemisphere were imaged per mouse (n = 2-3 mice per timepoint). All immunofluorescence images were processed using ImageJ software (version 1.52a).

Statistical considerations
GraphPad Prism version 7.0 and SPSS Version 24 were used for statistical analyses. Pairwise comparisons were performed using T-tests (independent sample, two-tailed, unequal variance). Non-parametric Mann-Whitney U test was used for pairwise comparisons of flow cytometric data. The statistical significance level was set at p < 0.05.

CNS infiltration by peripherally derived monocytes/ macrophages at 48-72 h post-tMCAO
Effective MCAO (18% baseline) and effective reperfusion (108% of baseline) were achieved in our 30-min tMCAO model (Additional file 1: Figure S1a). At 24 h, 2,3,5-triphenyltetrazolium chloride (TTC) staining demonstrated cerebral infarction as expected in the ipsilateral hemisphere (Additional file 1: Figure S1b). We characterized CD11b + CNS-MPs in the ipsilateral and contralateral hemispheres (excluding the cerebellum) following tMCAO at 30 min to 7 days timepoints (Fig. 1a). As compared to 30 min following tMCAO, we observed a significant increase in the proportions of CD11b + CD45 high CNS-MPs at 48 h (p = 0.02), 72 h (p = 0.00014), and 7 days (p < 0.00001) following tMCAO in the ipsilateral hemisphere but not in the contralateral hemisphere ( Fig.  1b and Additional file 2: Figure S2a). Among all CD11b + CNS-MPs, we also observed a significant increase in the proportions of CD45 high Ly6c high CNS-infiltrating macrophages/inflammatory monocytes at 48 h (p = 0.023) and 72 h (p = 0.01) following tMCAO, and in proportions of CD11b + CD45 high Ly6c low non-inflammatory monocytes/ activated microglia at 48 h to 7 days (p = 0.04) ipsilaterally following tMCAO [26,27] (Fig. 1c). As compared to CD45 high Ly6c low CNS-MPs, the increase in proportions of CD45 high Ly6c high CNS-MPs post-tMCAO was more robust (Fig. 1c). No changes in CNS-MP subset proportions were observed on the contralateral side. There was also an early increase in CNS CD11b neg CD45 high lymphocytes ipsilaterally at 30 min which rapidly decreased by 24 h (Additional file 2: Figure S2b). These data confirm that a wave of CNS infiltration by inflammatory monocytes occurs between 48 h to 7 days post-tMCAO [26,27].

Immunohistochemical validation of increased Kv1.3 protein expression following tMCAO
To validate our flow cytometric findings, we performed immunofluorescence imaging to characterize Kv1.3 channel protein expression in the brain at 1 h, 24 h, and 48 h following tMCAO ( Fig. 3 and Additional file 4: Figure S4). We observed a change in microglial morphology from a ramified state to a more rounded, ameboid, and activated morphology in the ipsilateral hemisphere following tMCAO at 24 and 48 h timepoints (Fig. 3a). At 1 h post-tMCAO, minimal Kv1.3 immunoreactivity was noted in both ipsilateral and contralateral hemispheres and very low-level Kv1.3 positivity was observed in microglia (Additional file 4: Figure S4). At 24 h and 48 h post-tMCAO, we observed increased Kv1.3 immunoreactivity in Iba1 + CNS-MPs in the peri-infarct regions (Fig. 3a). We also performed immunofluorescence imaging to measure Kv1.3 expression and Tmem119 post-tMCAO to confirm microglial specificity of Kv1.3 expression (Fig.  3b). At 1 h post-MCAO, Tmem119 immunoreactive microglia exhibited an expected ramified morphology and low levels of Kv1.3 protein expression. At 48 h post-tMCAO, we observed a dramatic decrease in Tmem119 immunoreactivity indicating downregulation of Tmem119 (Fig. 3b). These findings are consistent with downregulation of homeostatic genes/ proteins such as Tmem119 [11,13] following ischemic injury. Based on our flow cytometric and immunohistochemical results as well as the recently  [14], the pattern of functional Kv1.3 channel expression following tMCAO indicates that microglia that highly express Kv1.3 channels at 24-72 h post-tMCAO, but not infiltrating monocytes/macrophages, are likely to be the targets of Kv1.3-blocking therapies in stroke.

Transcriptional alterations in potassium channel genes following tMCAO
Since we observed significant alterations in functional Kv1.3 channel expression on the surface as well as at the protein level in microglia at 24-72 h following tMCAO, we determined whether changes in Kv1.3/Kcna3 gene expression in CNS-MPs can account for this increased functional channel expression. In qRT-PCR studies performed on acutely isolated CNS-MPs (≈ 60% CD11b + CNS-MPs [28]) from ipsilateral and contralateral hemispheres following tMCAO, we observed no significant changes in Kcna3 (Kv1.3) mRNA expression in ipsilateral or contralateral hemispheres (Fig. 4a). Interestingly, we observed a significant downregulation in gene expression of an inward-rectifying K channel (Kir2.1/Kcnj2) in CNS-MPs from both ipsilateral (48 h p = 0.02, 72 h p = 0.004) and contralateral hemispheres (24 h p = 0.025, 72 h p = 0.03) post-tMCAO (Fig. 4b) [29]. Downregulation  Fig. 4b). It has been previously reported that increased Kv1.3 channel expression in microglia that is induced by proinflammatory stimuli is associated with a reciprocal decrease in Kir2.1 channels while anti-inflammatory stimuli increase Kir2.1 currents [29]. Furthermore, Kv1.3 channels may represent regulators of pro-inflammatory microglial responses while Kir2.1 may regulate antiinflammatory and protective responses [13,28]. Based on the minimal changes in Kv1.3/Kcna3 mRNA levels in CNS-MPs following tMCAO, the observed robust changes in functional Kv1.3 channels in microglia post-tMCAO may indicate post-transcriptional regulation of Kv1.3 channels in microglia.

Characterization of phagocytic properties of CNS-MP subsets following tMCAO
Microglia and CNS-infiltrating macrophages are involved in phagocytic clearance following CNS injury [3,30]. We recently showed that higher phagocytic activity is observed in CNS-MPs with higher Kv1.3 channel expression and that phagocytic properties can be augmented by Kv1.3 blockade [13,28]. Therefore, we characterized the phagocytic capacity of acutely isolated CD11b + CNS-MPs following tMCAO at various timepoints using a validated flow cytometric assay [16]. We observed that microsphere phagocytosis by CD11b + CNS-MPs peaked at 48 h post-tMCAO in both ipsilateral (p = 0.008) and contralateral (p = 0.03) hemispheres although this augmentation was more pronounced in the ipsilateral hemisphere (Fig. 5). Among CNS-MP subsets, phagocytic augmentation was observed at 48 h in both CD45 low and CD45 high CNS-MPs although CD11b + CD45 high cells were more avid phagocytosers at all timepoints. We also observed an augmentation of phagocytic activity, both ipsilaterally and contralaterally (Fig. 5c). There was no correlation between phagocytic activity and Kv1.3 channel expression in our study (Additional file 5: Figure S5). These results suggest that CNS-MPs globally augment phagocytic activity in the ischemic brain and CNS-MP phagocytic activity in the tMCAO model does not appear to be coupled to Kv1.3 channel expression.

Discussion
Microglial activation and CNS infiltration by peripheral myeloid cells following ischemic stroke can impact infarct expansion, edema, and neuronal survival [31]. Disease-modifying roles for CNS-MPs in ischemic stroke are suggested by microglia depletion studies resulting in exacerbation of neuroinflammation and brain injury in focal ischemic stroke models [32]. In neurodegeneration, microglia progressively transition from homeostatic to DAM states [12,33,34] and within DAM, we have identified distinct pro-inflammatory and anti-inflammatory DAM sub-profiles with potentially opposing functional roles [9,13]. Although the relevance of the homeostatic and DAM paradigm to ischemic stroke is unclear, antiinflammatory CNS-MP phenotypes have been observed in the first 1-2 days after which pro-inflammatory phenotypes predominate in rodent MCAO models [7]. Therefore, selective neuro-immunomodulatory therapies for stroke are needed that inhibit pro-inflammatory CNS-MP responses and shift CNS-MPs towards protective anti-inflammatory and homeostatic states rather than non-specific anti-microglial strategies [6,32,35,36]. The Kv1.3 potassium channel has been identified as a promising therapeutic target to inhibit pro-inflammatory CNS-MP responses in neurodegenerative disease [15,24,28] as well as ischemic stroke models [14], resulting in improved neuropathological outcomes. Kv1.3 is highly co-expressed with pro-inflammatory DAM genes (Il1b, Tlr2, Hif1a, Ptgs2) and is a key regulator of proinflammatory CNS-MP responses [13,28,29,37]. Kv1.3 channels regulate membrane potential, calcium flux, immune signaling, and effector functions of effector memory T cells, memory B cells, and subsets of activated microglia and macrophages [38,39]. Neuronal Kv1.3 expression is limited to olfactory and cortical neurons as hetero-tetramers with other Kv1-family channels while immune cells exclusively express the homo-tetrameric channel [38,39]. This pattern of expression allows selective blockade of immune Kv1.3 channels by highly selective small molecule blockers (Pap1) and sea anemone toxin-based peptides (ShK analogs) without undesired off-target effects [40,41]. An early phase study of Kv1.3 channel-blocking ShK analog (ShK-186/ dalazatide) in humans has demonstrated safety, supported by multiple pre-clinical studies showing safety and efficacy of Kv1.3 blockers in models of systemic autoimmunity, obesity, CNS demyelinating disorders, and neurodegeneration [15,18,28,[42][43][44], suggesting that rapid translation of Kv1.3 blockers to ischemic stroke is feasible [45]. To further build on the preclinical rationale of Kv1.3 channel blockade as a therapeutic approach in ischemic stroke, the key CNS immune cells targeted by Kv1.3 blockers and the optimal therapeutic window for Kv1.3 blockers need to be determined.
We have utilized a rapid flow cytometric assay of functional cell-surface Kv1.3 channels [13,18,21] to phenotype acutely isolated CNS-MPs following tMCAO, and found that functional Kv1.3 channel expression is increased specifically in CD11b + CD45 low Ly6c low resident microglia as well as in the subset of CD11b + CD45 high-Ly6c low CNS-MPs, but not in CD11b + CD45 high Ly6c high inflammatory monocytes. Therefore, it is highly likely that microglia are the primary cell types targeted by Kv1.3 blockers and that the beneficial effects of Kv1.3 blockers in stroke models are unlikely to be mediated via modulation of peripheral immune responses. We also found increased Kv1.3 expression in the CD11b + CD45 high Ly6c low population of CNS-MPs which may represent an activated population of microglia or non-inflammatory patrolling monocytes that are recruited to the brain [26]. We also observed that not all microglia increase Kv1.3 expression, suggesting that additional immune sub-profiling of microglia may provide novel biological insights into the regulation and roles of Kv1.3 channel expression in microglial activation [46]. Increased functional cell-surface Kv1.3 channel expression at 48 h post-tMCAO was not accounted for by mRNA-level changes in Kv1.3, implicating posttranscriptional and post-translational processes in channel regulation in microglia [46]. However, this interpretation is limited by the large variance in our qRT-PCR data, insufficient mRNA yields from CNS-MP subsets and the use of all CNS-MPs (which contain 60% CD11b + CNS-MPs), rather than enriched CD11b + CNS-MPs within the scope of this study. Based on increased Kv1.3 channel expression between 24 and 72 h followed by a decrease by 7 days post-tMCAO, we also define the therapeutic window for Kv1.3 blockers as the subacute/ delayed phase of ischemic stroke rather than hyperacute/acute timeframes.
Our tMCAO studies were limited to 30 min of vessel occlusion followed by reperfusion. This model was specifically selected to minimize mortality observed with longer durations of occlusion especially beyond 48 h post-tMCAO.
Although we did not characterize Kv1.3 channel expression by flow cytometry in longer MCAO models, other groups have shown in the 60-min MCAO model that Kv1.3 protein expression was indeed increased between day 2 and day 8 post-MCAO, suggesting that our findings regarding delayed Kv1.3 channel expression by microglia seems to be consistent regardless of the duration of vessel occlusion [14]. As compared to traditional electrophysiological methods such as a whole-cell patch clamp which has very limited sampling capabilities (< 40-50 cells per mouse), our flow cytometric approach to characterize Kv1.3 channel expression provides a rapid and comprehensive strategy which also allows us to assess Kv1.3 expression within CNS-MP subsets. Another interesting finding that we report is the downregulation of Tmem119 protein in microglia at 48 h post-tMCAO, which suggests that microglia downregulate homeostatic genes/ proteins and may potentially transition towards activated phenotypes following MCAO [11][12][13]28]. Whether these delayed activated phenotypes share any similarities with DAM phenotypes observed in neurodegeneration remains to be explored. We also did not note any relationship between phagocytic activity in CNS-MPs and Kv1.3 channel expression. The overall increase in phagocytic activity in both hemispheres at 48 h post-tMCAO may indicate a global response although the mechanisms or implications of this observation remain unexplored. Our results with phagocytic uptake of latex microspheres also do not necessarily reflect the ability of CNS-MPs to clear cellular debris in the ischemic core or to phagocytose healthy neuronal synapses as these processes involve specific receptors including complement receptors [3,47]. Ongoing studies investigating specific mechanisms of phagocytosis in post-stroke CNS-MPs, as well as transcriptomic and proteomic efforts to characterize CNS-MPs in stroke models, will provide additional clarification.
From a clinical perspective, our findings in the transient 30-min MCAO model are relevant to acute largevessel occlusion (LVO) stroke, which represent 25-40% of acute ischemic stroke [1,48]. With the advent of mechanical thrombectomy as a highly effective reperfusion strategy, acute LVO stroke patients can be treated up to 24 h post-onset of stroke symptoms. However, nearly 50% of patients still suffer significant disability [9] and very few effective therapies exist beyond the 24-h window [10,49]. Since reperfusion can be established in LVO patients, the stroke field is ideally poised to revisit neuroprotective and neuro-immunomodulatory strategies such as Kv1.3 blockade to target disease mechanisms most relevant in the post-acute phases of ischemic stroke.