LRRC8A is dispensable for a variety of microglial functions and response to acute stroke

Abstract Microglia, resident brain immune cells, are critical in orchestrating responses to central nervous system (CNS) injury. Many microglial functions, such as phagocytosis, motility and chemotaxis, are suggested to rely on chloride channels, including the volume‐regulated anion channel (VRAC), but studies to date have relied on the use of pharmacological tools with limited specificity. VRAC has also been proposed as a drug target for acute CNS injury, and its role in microglial function is of considerable interest for developing CNS therapeutics. This study aimed to definitively confirm the contribution of VRAC in microglia function by using conditional LRRC8A‐knockout mice, which lacked the essential VRAC subunit LRRC8A in microglia. We demonstrated that while VRAC contributed to cell volume regulation, it had no effect on phagocytic activity, cell migration or P2YR12‐dependent chemotaxis. Moreover, loss of microglial VRAC did not affect microglial morphology or the extent of ischemic damage following stroke. We conclude that VRAC does not critically regulate microglial responses to brain injury and could be targetable in other CNS cell types (e.g., astrocytes) without impeding microglial function. Our results also demonstrate a role for VRAC in cell volume regulation but show that VRAC is not involved in several major cellular functions that it was previously thought to regulate, and point to other, alternative mechanisms of chloride transport in innate immunity.

including the volume-regulated anion channel (VRAC), are implicated in regulating microglial function in both physiological and pathological settings (Kettenmann et al., 2011).
VRAC is activated in response to osmotic swelling, where it mediates chloride efflux from the cytoplasm, which is followed by osmotically obliged water in a process known as regulatory volume decrease (RVD; Osei-Owusu et al., 2018). Leucine-rich repeat containing 8A (LRRC8A/SWELL1), along with its paralogues LRRC8B-E, are the molecular components of VRAC (Qiu et al., 2014;Voss et al., 2014).
VRAC channels are heterohexameric assemblies, with at least one LRRC8A subunit required for activity (Osei-Owusu et al., 2018).
LRRC8A alone, however, cannot form functional VRACs and requires the presence of at least one of the other four paralogues (Osei-Owusu et al., 2018). Beyond these basic constraints, subunit composition is variable, and differing stoichiometries determine the electrophysiological characteristics and solute specificities of the resulting channel (Lutter et al., 2017;Schober et al., 2017). In addition to chloride, VRAC transports a large repertoire of substrates, which are generally small organic osmolytes. Examples include amino acids such as glutamate and aspartate, as well as nucleosides such as ATP and cyclic GMP-AMP (cGAMP; Lahey et al., 2020;Osei-Owusu et al., 2018;Zhou et al., 2020a). As such, VRACs are capable of influencing not only cell volume, but also various intercellular signaling pathways.
VRAC inhibitors such as tamoxifen, 5-nitro-2-(3-phenylpropyl-amino) benzoic acid (NPPB) and 4-(2-butyl-6,7-dichlor-2-cyclopentylindan-1-on-5-yl) oxobutyric acid (DCPIB) are known not only to inhibit RVD, but also cell migration and phagocytosis, which are critical to microglial and macrophage function (Ducharme et al., 2007;Furtner et al., 2007;Harl et al., 2013;Schwab et al., 2012;Zierler et al., 2008). Additionally, VRAC blockers inhibit ATP-dependent chemotaxis of microglial processes following laser ablation in brain slices and prevent ramification of microglia in vitro (Eder et al., 1998;Hines et al., 2009). Thus, it has been suggested that VRAC-dependent chloride conductance also serves to regulate cell shape changes such as those required for ramification, migration and engulfment of particles. However, interpretation of these results with respect to VRAC is complicated by the limited specificity of chloride channel blockers for VRAC. Genetic studies are therefore required to confirm whether VRAC regulates these critical activities.
Astrocytic VRACs have been shown to act as glutamate release channels, regulating both baseline neural excitability and excitotoxic injury following cerebral ischemia (Yang et al., 2019;Zhou et al., 2020b). Mice with astrocyte-specific or whole-brain deletion of LRRC8A develop significantly smaller infarcts in experimental stroke models (Yang et al., 2019;Zhou et al., 2020b). Thus, VRAC channels could serve as a target for reducing acute neurotoxicity following CNS injury. However, pharmacological VRAC inhibition is likely to be indiscriminate of cell type and could potentially impair protective microglial functions. As such, greater understanding of the role of VRAC in microglial physiology is required, particularly in the context of brain injury.
To determine the true contribution of VRAC to microglial function, we utilized microglia/macrophage-specific LRRC8A knockout mice. In this model, we demonstrate that microglial development, morphology, and function remain intact despite complete loss of volume control in response to hypo-osmotic stress. Moreover, we show that microglial VRAC did not influence infarct volume in cerebral ischemia. Thus, we conclude that VRAC channel activity is dispensable for numerous important aspects of microglial physiology. and Calcein-AM (Biolegend) were obtained as powders dissolved in DMSO to Â200 concentrated stocks and stored at À20 C. pHrodo-SE (ThermoFisher) was dissolved in DMSO to 10 mg/ml and stored in aliquots at À80 C. Human Aβ 1-42 (Eurogentec) was dissolved in hexafluoro-2-propanol, dried into films under a rotary evaporator (Eppendorf) and stored at À80 C. Tomato lectin conjugated to DyLight 594 (Vector Biolabs) was washed twice through a centrifugal filter (10 kDa MWCO) to remove sodium azide, sterile filtered and stored as a 1 mg/ml stock at 4 C.

| Animals
The Lrrc8a fl/fl line was generated in-house using CRISPR-Cas9 as described previously (Green et al., 2020) and crossed with a Cx3cr1 Cre -expressing line (Yona et al., 2013) to generate littermates homozygous for the floxed LRRC8A allele and either null (Lrrc8a fl/fl : Cx3cr1 +/+ ) or heterozygous (Lrrc8a fl/fl :Cx3cr1 +/Cre ) for Cx3cr1 Cre , corresponding to WT and KO groups, respectively. The Creexpressing line (Lrrc8a wt/wt :Cx3cr1 +/Cre ) were also used in some studies as controls for Cre. All animals were housed under standard conditions (21 ± 2 C; 55% ± 5% humidity) in individually ventilated cages, under 12-h light/dark cycles with ad libitum access to food and water. Unless otherwise specified, all animals were at least 8 weeks old and of mixed sex. All in vivo work was performed in accordance with the Animals (Scientific Procedures) Act 1986 under relevant UK Home Office licenses and approved by the local Animals Welfare and Ethical Review Board (University of Manchester, UK).

| Cell culture
For microglia cultures, adult mice (8-44 weeks old) were perfused with ice-cold saline and the brains collected, the cerebellum and olfactory bulb were discarded, and the remaining tissue diced into small (1-2 mm) chunks using a scalpel. Papain and DNAse digestions were performed using a neural tissue dissociation kit (Miltenyi) according to the manufacturer's instructions. Following digestion, a dounce homogenizer with a loose pestle was used to dissociate the brain into a single-cell suspension, which was then pelleted at 300g (5 min) and myelin removed by centrifugation through 30% Percoll Plus (GE Healthcare). Cells were then resuspended in MACS buffer (PBS, 0.5% BSA, 2 mM EDTA) with anti-CD11b magnetic beads (Miltenyi) and incubated on a roller for 15 min. Microglia were enriched using a quadroMACS separator with LS columns (Miltenyi) according to the manufacturer's instructions. After counting, 1-2 x 10 4 cells were spotplated onto 96-well Cell+ anionic/cationic plates (Sarstedt) and allowed to adhere for 10 min before flooding with media. Cells were cultured for at least 7 days in DMEM/F12 (Gibco) supplemented with 1% penicillin/streptomycin (Sigma), 2 mM glutamine, 10% FBS (Gibco), 50 ng/ml TGF-β2 (Peprotech) and 20 ng/ml IL-34 (R&D Systems).
Membranes were washed three times in PBS-T before incubating with HRP-conjugated rabbit anti-mouse (Dako) in block buffer for 1 h at room temperature. After washing three further times, membranes were imaged on a G:Box Chemi XX6 (Syngene) using Amersham ECL Prime chemiluminescence reagent (GE Healthcare).
For loading controls, membranes were washed and incubated with HRP-conjugated mouse anti-β-actin (Sigma) in block buffer for 1 h at room temperature, followed by washing and imaging as above.
2.5 | Regulatory volume decrease assay 5 x 10 4 BMDM were seeded onto black-walled 96-well plates (Corning), whereas microglia were spot-plated onto Cell+ plates. Cells were loaded with 10 μM calcein-AM (BioLegend) for 1 h, washed three times with media and then rested for 30 min to allow calcein to equilibrate. After three further washes with isotonic buffer (132 mM NaCl, 2.5 mM KCl, 2 mM CaCl 2 , 2 mM MgCl 2 , 10 mM Glucose, 20 mM HEPES, pH 7.4, 312 mOsm/L) cells were imaged on an Eclipse Ti microscope (Nikon) equipped with a stage incubator maintaining 37 C and 5% CO 2 . Calcein fluorescence was captured in the GFP channel using low laser power to minimize phototoxicity and bleaching. Live imaging was conducted using point-visiting to image all conditions simultaneously. One image was captured every 2 min, with hypotonic shock induced at 5 min by adding distilled water. For quantification, fluorescent images were processed in Fiji, using a rolling-ball background subtraction (50 pixels ball size) to remove non-cell associated fluorescence. The average pixel intensity of each frame was then measured and normalized to the value of the first frame, yielding F/F 0 curves. For statistical testing, AUC values of the curves were calculated in Prism.

| Phagocytosis assays
Phagocytosis assays were performed using an IncuCyte ZOOM (Essenbio) time-lapse microscope housed in a humidified incubator maintaining 37 C and 5% CO 2 . Prior to imaging, cells were incubated in Opti-MEM (containing drugs or vehicle as specified) for 15 min.
Human Aβ 1-42 (Eurogentec) was dissolved to 10 mM in DMSO, then further diluted to 1 mM using sterile water and incubated at 37 C for 7 days. Fibrils were pelleted by centrifugation at 18,000g for 15 min, labeled with 25 μM pHrodo red-SE (ThermoFisher) in PBS at room temperature for 1 h, washed twice with PBS and added to cells at a final concentration of 10 μM.
For RBC assays, whole mouse blood was collected by cardiac puncture and centrifuged briefly at 500g to pellet cells. The pellet was then resuspended in PBS, layered onto a discontinuous percoll gradient containing 65% and 35% layers and centrifuged at 1000g for 15 min.
RBCs were recovered from the pellet, washed once with PBS, and 10 8 cells were labeled with 50 μM pHrodo-SE for 1 h, followed by two washes with PBS. Labeled RBCs were then opsonized with 1 μg/ml rabbit anti-mouse RBC (34-3C, Hycult) for 30 min, washed two further times with PBS, and added to cells at 10 6 RBC per well. Un-opsonised controls were included to verify that uptake was IgG-dependent.
Following addition of particles, three images per well were captured every 15 min for a total of 3 h, and automatically analyzed using the IncuCyte software (Essenbio). Custom scripts were created which gave optimal detection for each phagocytic substrate and were run identically on all conditions. For each well, the pHrodo-positive area was normalized to the area covered by cells (determined from the baseline image using phase confluence function) to give the phagocytic index.

| In vitro motility assay
Microglia cultured in 96-well plates were labeled with Hoechst (1 μg/ml) in growth media for 45 min, washed with serum-free media and incubated in Opti-MEM (containing vehicle or drugs as stated) for 15 min prior to imaging. Cells were imaged on a Nikon Eclipse Ti widefield microscope with a stage incubator maintaining 37 C/5% CO 2 using both phase contrast and DAPI channels with low UV laser power to prevent toxicity. Images were obtained every 5 min for 3 h.
The fluorescent (nuclear) channel was then analyzed for cell movement in ImageJ by first performing a rolling-ball background subtraction (ball radius 20) and loading the resulting images into the TrackMate plugin (Tinevez et al., 2017). To identify nuclei, LoG detection was used with a spot size of 8 μm, thresholds were empirically determined for each experiment and applied equally across all conditions. No spot filtering was necessary. Tracking was performed using the simple LAP tracker with max linking distance of 40 μm, and max gap closing distance of 70 μm over two frames.

| RNA sequencing
Microglia were purified from 10-12 week-old female WT, KO and Cre mice via MACS and RNA isolated immediately using Purelink RNA miniprep kits (Invitrogen). Total RNA was submitted to the genomic technologies core facility (GTCF) at the University of Manchester.
Quality and integrity of the RNA samples were assessed using a 2200 TapeStation (Agilent Technologies) and then libraries generated using the TruSeq ® Stranded mRNA assay (Illumina, Inc.) according to the manufacturer's protocol. Briefly, total RNA (0.1-4 ug) was used as input material from which polyadenylated mRNA was purified using poly-T, oligo-attached, magnetic beads. The mRNA was then fragmented using divalent cations under elevated temperature and then reverse transcribed into first strand cDNA using random primers. Second strand cDNA was then synthesized using DNA Polymerase I and Rnase H. Following a single "A" base addition, adapters were ligated to the cDNA fragments, and the products then purified and enriched by PCR to create the final cDNA library. Adapter indices were used to multiplex libraries, which were pooled prior to cluster generation using a cBot instrument. The loaded flow-cell was then paired-end sequenced (76 + 76 cycles, plus indices) on an Illumina HiSeq4000 instrument. Finally, the output data was demultiplexed (allowing one mismatch) and BCL-to-Fastq conversion performed using Illumina's bcl2fastq software, version 2.20.0.422.

| RT-qPCR
RNA was isolated using PureLink kits (Invitrogen) and cDNA transcribed using SuperScript III first-strand synthesis kits (Invitrogen). RT-qPCR was performed using 5 ng cDNA along with 200 nM forward and reverse primers ( Gapdh values as a loading controls, and Gapdh-normalized expression of exon 3 was further normalized to expression of exons 1-2 in the same sample to account for varying baseline Lrrc8a expression levels.
2.10 | Acute brain slices Acute brain slices were prepared according to published methods (Etienne et al., 2019

| Lectin injection and cranial window implantation
Cranial windows were implanted as previously described (Goldey et al., 2014). Animals were anesthetized with 2.5% isoflurane in 100% O 2 and the scalp was removed over the animals left and right hemisphere. A metal head plate was mounted (Narishige CP-2, Japan) using

T A B L E 1 Primers used for RT-qPCR analysis
Lrrc8a (exon 1-2) Forward GAGCAAAAGGAATGTCAGGGC

Reverse CAATCTCCACTTTGCCACTGC
Note: All sequences are given in the 5 0 -3 0 direction. dental cement (Sun Dental, Japan) to allow stereotaxic fixation under the 2-photon microscope. A circular piece of bone with a diameter of 3 mm was then removed using a dental drill. The dura was left intact for the whole experiment. For this experiment, animals with body weights between 20 and 25 g were used-generally males were 8-12 weeks old and females were 8-20 weeks old.
Prior to window implantation, tomato lectin conjugated to DyLight-594 was diluted to 50 μg/ml in sterile PBS, and 200 nl was injected through the dura into the cortex using a borosilicate microcapillary tube pulled to an outer tip diameter of 5 μm. Injections were performed over 10 min (UltraMicroPump III, World Precision Instruments, USA), followed by a 5 min rest before the needle was withdrawn. Injection depths were between 150 and 250 μm below the dura. Once the injection was finished, a circular coverslip (Warner Instruments, USA) was glued in place of the removed bone using dental cement (Sun Dental, Japan).
Mice were imaged immediately after cranial window implantation over a period of 4 h. During this time, they were maintained under 2% isoflurane in 100% O 2 and the body temperature was maintained at 37.5 C via a heating blanket controlled with a temperature probe (Harvard Apparatus, Kent, UK).

| Multiphoton imaging
Imaging was conducted using a Leica SP8 multiphoton microscope equipped with a MaiTai Ti:Sapphire MP laser (Spectra-Physics) and an HC Fluotar L 40Â water dipping lens. 1024 x 1024 pixel z-stacks were captured at Â1.8 confocal zoom, corresponding to approximate dimensions of 180 x 180 μm, with Z-planes spaced 2 μm apart. For each series, 15-20 z-planes were captured, corresponding to 15-20 μm above and below the lesion. For imaging GFP, the 2-photon laser was tuned to 880 nm, and to 800 nm for imaging tomato lectin. In each case, imaging laser power was kept below 5% and Â3 line averaging was used. Lesions were induced using the point bleach function in the LAS X software to illuminate a single pixel for 500 ms with laser power set to 20-30x that used for imaging. Image acquisition was then automatically started, and z-stacks acquired 1 min apart for 10 (for Cx3cr1-eGFP brain slices) or 15 min (in vivo imaging).
For analysis, xyzt hyperstacks were first registered in Fiji using the turboreg plugin, and then cropped to a 120 μm x 120 μm square centered around the lesion. A median filter with a radius of 2 pixels was then applied, and background subtraction with a rolling ball radius of 50 pixels performed. Maximum-intensity z-projections were then created and a further median filter applied to smooth noise. eGFP images were then thresholded using the Huang method. For tomato lectin images, thresholds were empirically determined due to variability in the staining intensity and presence of other bright objects (e.g., blood vessels), which complicated automatic thresholding.
Thresholded time-series were then analyzed in MATLAB using the MGPtracker script (Gyoneva et al., 2014; available from the original authors upon request), which quantifies the area of a polygon in each image surrounding the lesion which is free of microglial processes.
The resulting area-over-time data were then normalized to the value of the first frame for each video.

| Middle cerebral artery occlusion
Thrombi formation and cerebral ischemia were performed using the 2.14 | Immunohistochemistry  3.2 | LRRC8A does not affect microglial density and minimally affects gene expression It has been reported that loss of LRRC8A can affect development and severely restrict the survival of certain cell types, including T and B lymphocytes (Kumar et al., 2014;Platt et al., 2017). To assess whether lack of LRRC8A affects microglial development or homeostasis, we first quantified microglial density in the hippocampus and cortex of WT and KO mice, as well as Cre mice to control for Cx3cr1 hemizygosity in KO mice. All three genotypes displayed comparable numbers of microglia, indicating that LRRC8A is dispensable for microglial number and survival (Figure 2a . All data are presented as mean ± SEM. Phagocytic index traces are means of two biological replicates from a single experiment. AUC data points represent biological replicates and are normalized to the mean of the WT control. All experiments contain data from 3 to 6 mice. Statistics in B-C are from two-tailed Student's t-tests, g is from one-way ANOVA with Dunnett's post hoc and h-j are from two-way ANOVA. **p < .01, ****p < .0001 be enriched for any particular pathway or biological process (data not shown). Moreover, expression of the microglial homeostatic signature genes Tmem119, P2ry12, Sall1, Tgfbr1 and Egr1, as well as Trem2 and the inflammatory markers C3 and ApoE was consistent between the three genotypes ( Figure 2h). Cx3cr1 expression displayed the expected halving of expression intensity in KO and Cre microglia, but did not appear to be affected by loss of LRRC8A. Thus, lack of LRRC8A does not appear to have any profound influence on microglial survival or homeostasis at baseline.
We also considered the possibility that loss of LRRC8A might induce compensatory upregulation of other chloride channels.

| LRRC8A does not contribute to microglial phagocytosis
Previous studies have indicated that VRAC blockers applied to macrophages suppress phagocytosis of latex beads and E. coli, implying a role for LRRC8A-dependent volume regulation in phagocytosis (Ducharme et al., 2007;Furtner et al., 2007;Harl et al., 2013). To

| LRRC8A does not regulate microglial migration and P2RY12-dependent chemotaxis
Volume regulatory processes are proposed to contribute to cell motility and migration by facilitating chloride and water efflux at the trailing edge, and influx at the leading edge of the migrating cell (Schwab et al., 2012).
VRAC inhibitors have also been shown to suppress migration and chemotaxis in microglia, neutrophils and a variety of other cell types (Hines et al., 2009;Schwab et al., 2012;Volk et al., 2008). Thus, VRAC has been suggested as a regulator of cell movement. To assess whether LRRC8A deficiency affected cell migratory capacity, we utilized time-lapse imaging of primary adult microglia which, when cultured with serum, exhibit rapid, spontaneous migration (Bohlen et al., 2017). despite an inhibitory quality of chloride channel blockers, VRAC did not appear to contribute to whole-cell migration in this model.
The VRAC inhibitors NPPB and tamoxifen have also been reported to suppress microglial P2YR12-dependent chemotactic responses elicited by focal laser ablations in brain tissue (Hines et al., 2009). To verify this effect on ATP-dependent chemotaxis, we performed laser ablation injuries in acute brain slices obtained from CX3CR1-eGFP mice via two-photon microscopy. We observed consistent chemotactic reactions in control slices, which were inhibited by the presence of high concentrations (20 μM) of DCPIB, but unaffected by moderate concentrations (10 μM; Figure 5a-c). In order to determine whether LRRC8A-containing VRACs mediated this effect, we performed in vivo ablations in WT and LRRC8A KO mice through implanted cranial windows. Prior to imaging, tomato lectin conjugated to DyLight-594 was injected into the parenchyma to label microglia.
We observed no significant difference in chemotactic response between WT and KO microglia, as quantified by the reduction in process-free area following laser injury (Figure 5d-f). Thus, both stochastic migration and P2RY12-dependent microglial chemotaxis appear to occur independently of LRRC8A.
3.5 | Loss of microglial VRAC does not affect ischaemic damage or microglial morphology after stroke LRRC8A-containing VRACs in astrocytes have been shown to contribute to excitotoxic damage in ischemic stroke models by facilitating glutamate release (Yang et al., 2019;Zhou et al., 2020b). We therefore hypothesized that microglial VRACs might also influence ischemic  (Figure 6a). VRAC has also been suggested to support morphological transformation of microglia (Eder et al., 1998). As such, we also assessed microglial morphology in the contralateral and periinfarct brain regions (Figure 6b) via confocal microscopy coupled to an automated MATLAB algorithm for extraction of morphological parameters (Heindl et al., 2018). Principal component analysis of the 1409 total microglia analyzed (905 peri-infarct, 504 contralateral) confirmed that strong differences in microglial morphology were detected between the injured and non-injured regions ( Figure 6c). Specifically, microglia in the peri-infarct region exhibited significantly higher median sphericity, node volume and significantly fewer skeleton nodes per cell than those in the contralateral hemisphere (Figure 6d-f), corresponding to the expected decrease in process complexity and ramification. However, when analyzed separately, neither the contralateral nor peri-infarct microglia displayed appreciable genotype-dependent variability in PCA (Figure 6g,k). Moreover, no significant differences were observed in sphericity, total skeleton nodes per cell, or median node volume between the three genotypes in either the peri-infarct region ( Figure 6h-j) or contralateral cortex (Figure 6l- prior evidence to suggest a differential contribution of VRAC to phagocytosis between in vitro and in vivo assays. Nonetheless, given the potential for phenotypic differences between microglia in vivo and in vitro, this remains a potential caveat to the current work.
Since VRAC-dependent chloride conductance has also been postulated as a mechanism of cell migration, we also assessed stochastic migration of microglia in vitro. LRRC8A-KO cells exhibited no intrinsic defect in migratory capacity and, importantly, were equally sensitive to the suppressive effect of chloride channel blockers. Taken together, these results contradict the idea that VRAC serves a dual purpose, supporting both osmotic cell shape changes and those required for migration and phagocytosis. Instead, these two processes appear to occur via an entirely different mechanism to RVD. While it is clear that chloride transport plays a significant role in both migration and phagocytosis, the precise channels responsible remain mysterious due to lack of drugs selective enough to target one particular channel, or family of channels.
Due to its importance for microglial-mediated neuroprotection, we also investigated the involvement of VRAC in P2YR12-dependent chemotaxis of microglial processes. This mode of directed motility is a critical mechanism by which microglia react to changes in neuronal activity, and can regulate the formation of specialized junctions between microglial processes and neuronal somata, which reduce neuronal calcium load and promote survival under conditions of CNS injury (Cserép et al., 2020). In line with previously published data (Hines et al., 2009), our results confirm that P2YR12-dependent chemotaxis is abolished by VRAC inhibitors in ex vivo brain slices. However, we demonstrate that VRAC conductance is in fact dispensable for P2YR12-dependent chemotaxis in vivo. We did not assess whether DCPIB is able to block microglial chemotaxis in the absence of LRRC8A. However, given that LRRC8A-KO itself did not alter chemotaxis and that DCPIB clearly displays off-target activity in other assays, it is likely that inhibition of chemotaxis by DCPIB is also VRAC-independent.
It is possible that chloride channel inhibitors suppress microglial process extension in this model via the same mechanism by which they inhibit stochastic migration and phagocytosis-that is, by producing a widespread block on cell shape changes and movement. However, previous research has suggested that microglia are still capable of extending filopodia during random surveillance whilst exposed to tamoxifen and NPPB at concentrations which inhibit directed motility (Hines et al., 2009). It may therefore be the case that ATP-induced chemotaxis requires additional chloride transporters distinct from those which support general cell motility.
One potential further question not addressed here is whether or not VRAC contributes to microglial surveillance movements in a similar manner to the two-pore domain family K + channel THIK-1 (Madry et al., 2018). Previous studies did not observe any effect of VRAC inhibitors on microglial filopodia extension during undirected surveillance (Hines et al., 2009), but did not assess changes in surveillance behavior on the level of whole cells. Thus, it may be possible that VRAC could contribute to microglial surveillance, though at present there is no direct evidence to suggest this.
Our findings are encouraging in light of recent papers highlighting glutamate-releasing astrocytic VRAC as a potential therapeutic target for stroke and other CNS injuries due to its role in promoting excitotoxicity (Yang et al., 2019;Zhou et al., 2020b). As with any such strategy, it is important to explore the potential off-target effects of VRAC inhibition on other CNS cells, particularly microglia, due to their critical role in damage responses. Our results demonstrate that microglia are fully capable of migration, phagocytosis and P2YR12-dependent chemotaxis in the absence of VRAC, and that abolishing microglial VRAC does not affect the extent of ischemic damage in a mouse model of stroke. Thus, it should be possible to target glutamate transport via astrocytic VRAC in CNS injuries without compromising potential neuroprotective microglial activity. However, another key point raised by our findings is that the current array of VRAC inhibitors are not sufficiently selective for this purpose. Indeed, we and others have demonstrated extensive undesirable off-target activity in even the current gold-standard inhibitor, DCPIB (Afzal et al., 2019;Lv et al., 2019;Minieri et al., 2013). As such, the discovery of compounds with greater selectivity for VRAC over other chloride channels would be of great benefit, both to better understand VRAC's roles in physiology and to explore its value as a therapeutic target.