Mechanism of action of two potent LRRC8/VRAC channel inhibitors
Sara T. Granados, Richard Song, Francisca Añazco, Jose M. Vermejo, Yan Rao, Axel R. Concepcion
TL;DR
This study reveals how two VRAC inhibitors, DCPIB and dicumarol, work and highlights their off-target effects on T cells and calcium signaling.
Contribution
The paper identifies the mechanism by which DCPIB and dicumarol inhibit VRACs and cause off-target effects in T cells.
Findings
DCPIB and dicumarol inhibit VRACs by accumulating in the cell membrane in an albumin-dependent manner.
The compounds disrupt Ca2+ signaling and T cell function in the absence of serum/albumin.
Mitochondrial depolarization and oxidative stress are key mechanisms of their adverse effects on T cells.
Abstract
Volume-Regulated Anion Channels (VRACs), composed of Leucine-Rich Repeat Containing 8 (LRRC8) proteins, are emerging as promising therapeutic targets, but their pharmacology is poorly defined. Small-molecule VRAC inhibitors share lipophilic properties and exhibit a wide range of off-target effects, rendering them unsuitable for physiological studies. Furthermore, the mechanisms of action underlying their on- and off-target effects remain largely unclear. Here, we show that two structurally unrelated small-molecule inhibitors of VRACs, DCPIB and dicumarol, exert their cellular effects by accumulating in and permeating the cell membrane in an albumin-dependent and VRAC-independent manner. In conditions lacking serum/albumin, both compounds not only inhibit VRAC function but also disrupt store-operated Ca2+ entry (SOCE), Ca2+ signaling, and the activation and function of human and mouse T…
Genes, proteins, chemicals, diseases, species, mutations and cell lines named across the full text — each resolved to its canonical identifier and authoritative record.
Click any figure to enlarge with its caption.
Figure 1
Figure 2
Figure 3
Figure 4
Figure 5
Figure 6
Figure 7
Figure 8Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
Taxonomy
TopicsIon Channels and Receptors · Ion channel regulation and function · Calcium signaling and nucleotide metabolism
INTRODUCTION
Volume-regulated anion channels (VRACs) play a crucial role in regulating cell volume in vertebrates (1–4). Canonically, these channels activate in response to prolonged hypotonic stress and facilitate the efflux of chloride ions (Cl^−^) and organic osmolytes, thereby enabling regulatory volume decrease (5, 6). Beyond osmoregulation, VRACs are implicated in numerous key physiological processes as transporters of second messengers, small-molecule drugs, and metabolites under normotonic conditions (7–13). VRACs are heterohexameric channels composed of the obligatory subunit leucine-rich repeat-containing protein 8A (LRRC8A) and at least one of its four paralogs (LRRC8B-E) (14, 15). The functional diversity of VRACs can be attributed in part to the inherent variability in their assembly, as the specific subunit composition of the channel determines its substrate selectivity. For instance, heteromeric LRRC8A/C channels transport potent immunomodulatory cyclic dinucleotides (CDNs), including 2’3’-cGMP-AMP (cGAMP), which regulate immune cell function (10–12). In contrast, LRRC8A/D channels have been shown to transport neurotransmitters such as glutamate, aspartate, and γ-aminobutyric acid (9), as well as the chemotherapeutic drug cisplatin (8, 16).
Pharmacological inhibitors of VRAC are often used to abolish VRAC currents (IVRAC) when investigating the expression and functional roles of these channels. The current best-in-class VRAC inhibitor is the ethacrynic acid derivative 4-[(2-butyl-6,7-dichloro-2-cyclopentyl-2,3-dihydro-1-oxo-1H-inden-5-yl)oxy] butanoic acid (DCPIB), which completely abrogates IVRAC in electrophysiological studies at a concentration of 10 μM (17). The proposed model of VRAC inhibition by DCPIB is based on cryogenic electron microscopy (cryo-EM) structures of homohexameric LRRC8A complexes, which show that the inhibitor binds to the extracellular selectivity filter, blocking the channel pore via a cork-in-bottle mechanism (18). According to this model, the pore blockage is caused by the electrostatic interactions between the negatively charged carboxylic acid group of DCPIB and a ring of arginine residues (R103) at the constriction site of the extracellular selectivity filter (18). A similar study using SN-407, a DCPIB analog, reveals a similar mechanism of inhibition (19). However, a conflicting study has indicated that LRRC8A homomeric channels exhibit poor pharmacological responses to DCPIB (20). Additionally, VRACs with amino acid substitution other than arginine at the constriction site of the extracellular selectivity filter remain inhibited by DCPIB (20), suggesting that DCPIB may inhibit VRACs through a mechanism different from the currently proposed model.
Although DCPIB is effective at inhibiting VRACs, it has been associated with numerous off-target effects on several pathways and membrane proteins (21–30). However, it is unclear whether these off-target effects occur independently or share a common mechanism. Despite these effects, DCPIB remains widely used to inhibit VRAC in both in vitro and in vivo assays (11, 23, 31–36). Furthermore, the concentrations of DCPIB commonly used to inhibit VRACs in these studies often exceed those that fully block the channel in electrophysiological conditions (i.e., >20–100 μM compared to ~10 μM, respectively) (22, 23, 33, 37, 38). In some cases, the lack of DCPIB effects has caused the function of VRACs to be mistakenly dissociated from their known role (32). Consequently, efforts have been made to identify novel VRAC inhibitors, such as pranlukast and zafirlukast (39), and more recently, dicumarol (13). However, these drugs are similarly non-specific: pranlukast and zafirlukast are cysteinyl leukotriene receptor 1 (CysLT1R) antagonists (39, 40), while dicumarol is a natural anticoagulant that inhibits vitamin K epoxide reductase (13). Therefore, the promiscuous nature of VRAC inhibitors remains one of the main pharmacological limitations to studying the functional roles of these channels.
Here, we report that two potent and structurally unrelated VRAC inhibitors, DCPIB and dicumarol, impair Store-Operated Ca^2+^ Entry (SOCE), a critical mechanism for Ca^2+^ influx and signal transduction in T lymphocytes and other cells (41, 42). SOCE is activated by the depletion of Ca^2+^ stores in the endoplasmic reticulum (ER) via inositol trisphosphate receptors (IP_3_Rs) and ryanodine receptors (RyRs) (43, 44), which is sensed by the stromal interaction molecules (STIM1/2). The depletion of ER-Ca^2+^ stores leads to conformational changes and oligomerization of STIM proteins, resulting in their mobilization toward ER-plasma membrane (PM) junctions (45). At these junctions, STIM proteins interact with and activate the Ca^2+^-release-activated Ca^2+^ (CRAC) channel, which is composed of ORAI1/2/3 proteins (46). The Ca^2+^ influx generated by the activation of SOCE in T cells is essential for their metabolic reprogramming, proliferation, differentiation, cytokine production, and adaptive immune response (42, 46–48).
Our findings indicate that both DCPIB and dicumarol deplete intracellular Ca^2+^ stores in T cells through different mechanisms, while simultaneously inhibiting SOCE. Importantly, the suppression of SOCE by these small-molecule inhibitors occurs independently of the expression of LRRC8A and LRRC8C, which are the main components of functional VRACs in T cells. We demonstrate that the off-target effects on SOCE observed with DCPIB and dicumarol result from depolarization of the mitochondrial membrane potential (Δψm), which precedes changes in actin aggregation and apoptosis of T cells. Interestingly, we found that the alterations in Ca^2+^ signaling caused by DCPIB and dicumarol were mitigated by adding albumin to the buffer. Consistently, all observed off-target effects in vitro were completely prevented by the presence of serum in the culture media, suggesting that the cellular permeability of DCPIB and dicumarol is necessary for inducing these effects. Lastly, we discovered that the lack of permeation of these small-molecule inhibitors led to insufficient inhibition of VRAC, which is required for mediating cGAMP transport to induce T cell death. Our study reveals that the effectiveness of DCPIB and dicumarol as VRAC inhibitors is limited by their cell permeability under physiological conditions, as well as their off-target effects on mitochondrial depolarization and Ca^2+^ signaling. Therefore, we recommend that these findings be considered when interpreting previous studies that have used DCPIB and dicumarol, as earlier results may have been influenced by the identified off-target effects. While DCPIB and dicumarol remain valuable tools for short-term electrophysiological experiments, their use in studies of VRAC function under physiological conditions should be limited.
RESULTS
DCPIB inhibits Ca2+ signaling, cytokine production, and survival of T cells in serum-free conditions
We previously reported that LRRC8C deletion in T cells resulted in a complete abrogation of IVRAC function, as well as enhanced proliferation, survival, Ca^2+^ influx, cytokine production, and adaptive immune response (12). To test whether the increased Ca^2+^ influx observed in Lrrc8c^−/−^ T cells was directly associated with their impaired VRAC function, we pre-treated murine activated CD4^+^ T cells with DCPIB, the best-in-class inhibitor of VRAC (IC_50_ ~ 4.1 μM) (17). We then induced SOCE by inhibiting the sarco-endoplasmic reticulum Ca^2+^ ATPase (SERCA) with thapsigargin (TG) in the presence of DCPIB in Ringer’s buffer. Surprisingly, DCPIB not only abolished the ER-Ca^2+^ store depletion by TG but also inhibited Ca^2+^ influx (Fig. 1A, B). As anticipated, Lrrc8c^−/−^ CD4^+^ T cells exhibited enhanced SOCE, consistent with our previous findings (12). However, when challenged with DCPIB, Lrrc8c^−/−^ CD4^+^ T cells also failed to deplete ER-Ca^2+^ stores and to undergo Ca^2+^ influx upon TG treatment, similar to wild-type T cells (Fig. 1A, B). This unexpected result suggests that DCPIB impairs Ca^2+^ signals in T cells independently of LRRC8C expression.
Upon antigen recognition, the stimulation of the T cell receptor (TCR) triggers multiple signaling pathways that are essential for the activation, proliferation, survival, and differentiation of T cells (Fig. 1C). To investigate the effect of DCPIB on Ca^2+^ signaling, we leveraged our published RNA-Seq data on wild-type CD4^+^ T cells treated with DCPIB during T cell stimulation (GSE163679) (12). In that experiment, wild-type CD4^+^ T cells were pre-treated with 20 μM DCPIB (or vehicle control) in serum-free RPMI media for 30 min and then activated with anti-CD3+CD28 antibodies for 1 and 2 days in the presence or absence of DCPIB in standard culture media (12). We analyzed the expression of Ca^2+^-dependent genes such as Myc transcription factor, glycolytic enzymes, and Il2 cytokine (47, 48), and found that DCPIB treatment resulted in an overall suppression in the induction of these Ca^2+^-dependent genes upon anti-CD3+CD28 stimulation (Fig. 1D). To assess the broader impact of DCPIB on T cell activation, we performed comparative pathway analysis using normalized enrichment scores (NES) of specific, well-defined biological states or processes from Hallmark gene sets (49). In wild-type CD4^+^ T cells, TCR stimulation triggers multiple signaling pathways associated with metabolic reprograming, proliferation, and survival (Supplementary Fig. 1). These pathways were further enhanced in T cells lacking LRRC8C expression (Fig. 1E), supporting our previous findings that implicate LRRC8C as a negative regulator of T cell function (12). LRRC8C deletion or DCPIB treatment in T cells further downregulated p53 signaling and TNFα signaling via NFκB, as previously reported (12). However, unlike LRRC8C deletion, DCPIB treatment largely abrogated the induction of most TCR-induced pathways (Fig. 1E). Collectively, these data strongly suggest that DCPIB suppresses Ca^2+^ signaling and T cell activation, potentially through mechanisms independent of VRAC inhibition.
SOCE in T cells is essential for cytokine production (42). We therefore analyzed how DCPIB treatment and reduced SOCE affects cytokine production in activated CD4^+^ T cells. Unexpectedly, under standard culture conditions, treatment with 20 μM or 40 μM DCPIB during the 5h cytokine production assay had no effect on the production of IL-2 and TNF-α (Fig. 1F, G). However, when we pre-treated T cells with 20 μM or 40 μM DCPIB for 30 min in serum-free RPMI and then repeated the cytokine production assay in full RPMI medium, we observed a dose-dependent reduction in IL-2 and TNF-α production (Fig. 1F, G). These results suggest that the effect of DCPIB on suppressing cytokine production may depend on its cell permeability, as the presence of serum seems to mitigate this effect. To test the broader impact of DCPIB on T cell function, we stimulated mouse splenocytes for 2 days with anti-CD3+CD28 and then treated them with 20 μM or 40 μM DCPIB in both serum-containing and serum-free RPMI media for additional 24h. In this experiment, we observed that the withdrawal of serum from the culture media had a minor impact on the survival of both CD4^+^ and CD8^+^ T cells treated with the vehicle control (Fig. 1H, I). While the treatment with DCPIB in serum-containing RPMI medium had a negligible effect on the survival of CD4^+^ T cells and no effect on CD8^+^ T cells, treatment with DCPIB in serum-free medium induced significant apoptosis in both T cell subsets, regardless of the DCPIB concentration (Fig. 1H, I). Together, these data demonstrate that DCPIB suppresses T cell survival and function, and these effects may depend on DCPIB permeability, as they are completely abrogated when serum is present in the culture medium.
Cell permeability of DCPIB induces a passive depletion of intracellular Ca2+ stores
Given the relevant role of Ca^2+^ as a second messenger in cell signaling and function, we decided to investigate the profound yet uncharacterized effect of DCPIB on suppressing Ca^2+^ signals in T cells in more detail. To this purpose, we used Jurkat cells, which are an immortalized line of human T cell leukemia widely used to study T cell signaling (50). First, we validated that Jurkat cells exhibit similar sensitivity to DCPIB as observed in murine T cells. Treatment with 20 μM or 40 μM DCPIB resulted in a time-dependent and dose-dependent apoptosis of Jurkat cells in a serum-dependent manner, similar to murine T cells (Supplementary Fig. 2). Next, we assessed the presence of functional VRACs and their sensitivity to DCPIB. As expected, whole-cell patch-clamp recordings of Jurkat cells displayed robust swelling-activated Cl^−^ currents (IVRAC) when exposed to ~210 mOsm hypotonic solution. These IVRAC were characterized by moderate outward rectification and DCPIB sensitivity (Fig. 2A), as previously reported (51). We then tested the depletion of intracellular Ca^2+^ stores by challenging Jurkat cells with the Ca^2+^ ionophore ionomycin (positive control) or DCPIB in Ca^2+^-free Ringer’s buffer. Interestingly, DCPIB triggered a modest Ca^2+^ store depletion that contrasts with the rapid depletion induced by ionomycin (Fig. 2B). This depletion of intracellular Ca^2+^ stores caused by DCPIB may explain the lack of Ca^2+^ responses observed in DCPIB pre-treated T cells (Fig. 1A). Furthermore, the Ca^2+^ store depletion upon DCPIB treatment, unlike that from ionomycin, was completely abrogated by the presence of albumin in the Ca^2+^-free Ringer’s buffer at a concentration equivalent to standard culture conditions (0.2%) (Fig. 2B, C). This observation is consistent with the well-established role of albumin in sequestering diverse drugs and small molecules (52). This suggests that DCPIB induced intracellular Ca^2+^ signaling may depend on drug permeation cross the cell membrane. To directly investigate this observation, we performed a parallel artificial membrane permeability assay (PAMPA) to evaluate the ability of DCPIB to passively diffuse across a synthetic lipid membrane (Fig. 2D) (53, 54). In this assay, a solution of 50 μM DCPIB was placed in a donor compartment and incubated for 5h in a humid chamber. The concentration of DCPIB was then determined in both the donor and acceptor wells using liquid chromatography-mass spectrometry (LC-MS). Interestingly, we observed ~60% accumulation of DCPIB in the lipid membrane, confirming its lipophilic nature (Fig. 2E). As expected, DCPIB passively permeated across the lipidic membrane and was detectable in the acceptor well in an albumin-dependent manner (Fig. 2E). Notably, the presence of 0.2% albumin reduced drug retention in the membrane, permeability, and accumulation in the acceptor well (Fig. 2E). These findings suggest that the effects of DCPIB in cells may be initiated by its absorption and accumulation in the membrane, as well as its intracellular permeation in the absence of serum, while albumin appears to mitigate these effects.
To identify the intracellular Ca^2+^ stores depleted by DCPIB upon permeation, we blocked IP_3_Rs and RyRs, which are the Ca^2+^ release pathways located at the ER, the main reservoir of intracellular Ca^2+^. When we treated Jurkat cells with CD3 monoclonal antibodies (OKT3) to activate the TCR/CD3 complex, we observed a transient release of intracellular Ca^2+^, which was completely abrogated by 50 μM 2-APB, a potent inhibitor of IP_3_Rs. In contrast, ryanodine, a strong RyR inhibitor, had no effect (Fig. 2F, G). TG treatment led to a robust depletion of Ca^2+^ stores, which was strongly inhibited by 2-APB and partially reduced by ryanodine (Fig. 2F, G). Interestingly, while 2-APB significantly reduced DCPIB-induced Ca^2+^ store depletion, inhibition of RyRs produced a greater decrease Ca^2+^ release. Together, these results suggest that DCPIB permeation induces Ca^2+^ release from intracellular ER-Ca^2+^ stores through both IP_3_R and RyR, operating independently of SERCA pump inhibition.
DCPIB depletes different intracellular Ca2+ stores but simultaneously inhibits SOCE
The depletion of ER-Ca^2+^ stores is the main mechanism that activates CRAC channels and SOCE (45). We have shown that DCPIB not only affects ER-Ca^2+^ store depletion but also inhibits Ca^2+^ influx in murine T cells (Fig. 1A). To determine whether the ER-Ca^2+^ store depletion caused by DCPIB has an impact on SOCE in Jurkat cells, we depleted intracellular Ca^2+^ stores in Ca^2+^-free Ringer’s buffer with either ionomycin or DCPIB. We then supplemented the Ringer’s buffer with 1 mM Ca^2+^ to assess Ca^2+^ entry. Using this protocol, we observed that while ionomycin treatment induced a robust SOCE upon the addition of extracellular Ca^2+^, this Ca^2+^ influx was absent in the DCPIB-treated cells (Fig. 3A, B). This result suggests two potential mechanisms: i) the store depletion triggered by DCPIB is insufficient to activate SOCE, or ii) DCPIB inhibits SOCE. To directly evaluate whether DCPIB suppresses SOCE, we fully depleted ER-Ca^2+^ stores with ionomycin and then treated the cells for 10 min with either DCPIB or ionomycin before inducing SOCE. Using this modified protocol, we observed that ionomycin treatment resulted in a robust SOCE activation, as expected, while the acute treatment with DCPIB prior to SOCE induction was sufficient to suppress Ca^2+^ influx (Fig. 3C, D). Similar results were observed in Jurkat cells challenged with TG or in primary human T cells treated with ionomycin (Supplementary Fig. 3A-D). Surprisingly, we noticed that after ionomycin-induced ER-Ca^2+^ store depletion, treatment with DCPIB consistently displayed a small but significant amount of intracellular Ca^2+^ release (Fig. 3E, F), suggesting that DCPIB may act on an additional intracellular Ca^2+^ store other than the ER (Fig. 3G). These findings indicate that although DCPIB causes depletion of various intracellular Ca^2+^ stores, this small-molecule VRAC inhibitor simultaneously inhibits SOCE.
To further confirm that the effects of DCPIB on intracellular Ca^2+^ responses depend on its ability to permeate the cell membrane, we conducted additional experiments in Jurkat cells with and without albumin in the Ringer’s buffers to prevent DCPIB permeation. In these conditions, we depleted ER-Ca^2+^ stores using ionomycin and then added DCPIB before SOCE induction. As expected, we observed that in albumin-free Ringer’s buffer, DCPIB suppressed SOCE activation (Fig. 3H, I). However, this effect was completely dampened in albumin-containing Ringer’s buffer (Fig. 3H, I). We also noticed that albumin not only eliminated the effect of DCPIB in suppressing SOCE but also reduced the overall Ca^2+^ influx response, even in the absence of DCPIB (Fig. 3H, I). To further demonstrate that the loss of DCPIB response in suppressing SOCE was due to its lack of permeation, we increased the extracellular Ca^2+^ concentration 2-fold, from 1 mM to 2 mM, to increase the Ca^2+^ gradient. Under these conditions, compared with 1 mM Ca^2+^ Ringer’s buffer, we observed an increment in Ca^2+^ influx in the presence of 0.2% albumin, albeit with faster inactivation kinetics (Fig. 3H, I). Nonetheless, DCPIB completely lost its capacity to suppress SOCE. In summary, these data demonstrate that the ability of DCPIB to alter intracellular Ca^2+^ signals and suppress SOCE is critically dependent on its extracellular availability and permeation into the cell.
DCPIB inhibits SOCE through mitochondrial depolarization
Previous studies have shown that DCPIB inhibits the electron transport chain (ETC) complexes I-III, which leads to dissipation of the mitochondrial membrane potential (Δψm), a reduction in ATP production, and increased oxidative stress, independent of LRRC8A expression (22). Some of these effects have been previously reported to inhibit SOCE (45, 55–60). Our RNA-Seq analysis in murine T cells shows that DCPIB suppresses oxidative phosphorylation (Fig. 1E). Moreover, we have found that, beyond ER-Ca^2+^ stores, DCPIB may also act on additional intracellular stores to release Ca^2+^ (Fig. 3E–G). We hypothesized that mitochondrial depolarization induced by DCPIB causes the release of Ca^2+^ from this organelle due to the ETC impairment, uncoupling, and proton (H^+^) accumulation in the matrix. To test this hypothesis, we first confirmed that chemical mitochondrial uncouplers, such as carbonyl cyanide-p-trifluoromethoxyphenylhydrazone (FCCP), can achieve comparable effects in suppressing SOCE. When we challenged Jurkat cells with 1 μM FCCP for 10 min before inducing SOCE, we observed a suppression of Ca^2+^ influx, similar to the effects of DCPIB (Fig. 4A, B). To assess the effect of DCPIB on Δψm, we measured changes in the resting Δψm using the dye tetramethylrhodamine ethyl ester (TMRE) in Jurkat cells. We found that DCPIB causes rapid depolarization of the Δψm, comparable to FCCP (Fig. 4C, D). To investigate the mechanism by which FCCP and DCPIB inhibit SOCE, we examined whether these small molecules interfere with the interaction between the ER-resident protein STIM1 and the ORAI1 Ca^2+^ channel at the plasma membrane. We analyzed the formation of STIM1-ORAI1 puncta following depletion of ER-Ca^2+^ stores using total internal reflection fluorescence microscopy (TIRFM). We constitutively overexpressed YFP-ORAI1 in HEK293 cells and transfected them with mCherry-STIM1. In non-stimulated HEK293 cells with intact ER-Ca^2+^ stores, YFP-ORAI1 was diffusely expressed across the entire plasma membrane, while mCherry-STIM1 formed puncta, likely due to its overexpression (Fig. 4E, F). After ER Ca^2+^ stores were depleted with TG, YFP-ORAI1 aggregated and colocalized with mCherry-STIM1 (Fig. 4E, F). In contrast, both FCCP and DCPIB reduced YFP-ORAI1 aggregation and its colocalization with mCherry-STIM1 puncta (Fig. 4E, F). These results suggest that depolarization of mitochondria by DCPIB, similar to that caused by FCCP, may contribute to the suppression of SOCE by interfering with the interaction between STIM1 and ORAI1.
In T cells, prolonged mitochondria depolarization has been linked to excessive production of reactive oxygen species (ROS), actin aggregation, and cell death (61). To further investigate whether mitochondrial depolarization induced by DCPIB leads to oxidative stress, we measured mitochondrial superoxide (O_2_^−^) accumulation using MitoSOX. We observed that acute treatment of Jurkat cells with FCCP or DCPIB triggered a rapid increase in mitochondrial O_2_^−^ production compared to the vehicle control or ionomycin-treated cells. Notably, DCPIB induced 2.3-fold more mitochondrial O_2_^−^ than FCCP after 5 min of treatment (Fig. 4G, H). Additionally, we observed that both DCPIB and FCCP caused time-dependent actin aggregation in Jurkat cells (Fig. 4I, J). DCPIB treatment resulted in ~85% actin aggregation upon 3h of treatment and complete actin aggregation after 6h, while FCCP caused ~50% reduction in soluble actin at both 3 and 6h of treatment. As expected, the induction of actin aggregation was completely prevented in the presence of serum, strongly suggesting that the effect of both DCPIB and FCCP depend on the membrane permeation of these small molecules. In summary, these data demonstrate that DCPIB permeation alters mitochondrial membrane potential, leading to an accumulation of mitochondrial ROS and actin aggregation, which accompanies apoptosis observed in both primary T cells (Fig. 1H, I) and Jurkat cells (Supplementary Fig. 2).
Dicumarol, a novel VRAC inhibitor, impairs SOCE and T cell survival through a mechanism similar to that of DCPIB
To identify new VRAC inhibitors with fewer off-target effects than DCPIB, a recent study screened the NIH drug library containing ~3000 FDA-approved drugs and identified dicumarol (dicoumarol) as a potent VRAC inhibitor (IC_50_ ~3.8 μM) (13). We tested whether dicumarol exhibits similar effects on Ca^2+^ signaling and T cell function as we have found for DCPIB. First, we confirmed that dicumarol inhibits IVRAC when Jurkat cells are exposed to a hypotonic solution (~210 mOsm). Similar to DCPIB (Fig. 2A), dicumarol completely abolished IVRAC in Jurkat cells following a hypotonic challenge (Fig. 5A). We then tested dicumarol permeability using the PAMPA assay and found that dicumarol shares similar pharmacological properties with DCPIB in terms of membrane retention, lipid membrane permeation, and albumin-dependent sequestration, confirming its lipophilic nature (Fig. 5B). Next, we examined whether dicumarol permeation triggers intracellular Ca^2+^ store depletion in Ca^2+^-free Ringer’s buffer. Surprisingly, we found that, like DCPIB, dicumarol promoted a passive release of intracellular Ca^2+^, and this effect was abolished by the addition of 0.2% albumin in Ca^2+^-free Ringer’s buffer (Fig. 5C). In contrast to DCPIB, the intracellular Ca^2+^ store depletion induced by dicumarol was independent of ER-Ca^2+^ release, as the inhibitors of IP_3_R (2-APB) and RyR (ryanodine) failed to prevent the increase in intracellular Ca^2+^ (Fig. 5D). Similarly to DCPIB, acute treatment with dicumarol for 10 min before inducing SOCE in ionomycin-treated cells significantly suppressed Ca^2+^ influx in both Jurkat cells (Fig. 5E) and primary human T cells (Supplementary Fig. 4A, B). Dicumarol also exhibited a similar pattern in depleting non-ER-Ca^2+^ stores in Jurkat cells that were pre-treated with ionomycin (Supplementary Fig. 4C, D). In line with its lack of Ca^2+^ store depletion in albumin-containing Ca^2+^-free Ringer’s buffer, dicumarol failed to inhibit SOCE in the presence of 0.2% albumin in 2 mM Ca^2+^-containing Ringer’s buffer (Supplementary Fig. 4E-G). Together, these data strongly suggest that the specificity of dicumarol to inhibit VRAC may be compromised by both its cell permeability and effects on altering Ca^2+^ signals, similar to DCPIB.
To investigate whether the mechanism by which dicumarol inhibits SOCE resembles that of DCPIB, we used TIRFM to test whether dicumarol interferes with the interaction between STIM1 and ORAI1 in HEK293 cells after ER-Ca^2+^ store depletion. Like DCPIB, pre-treatment with dicumarol for 10 min led to a significant reduction in YFP-ORAI1 aggregation and colocalization with mCherry-STIM1 puncta following ER-Ca^2+^ store depletion with TG (Fig. 5F, G). We then tested whether dicumarol triggers mitochondrial depolarization and actin aggregation in Jurkat cells. Thus, acute treatment with dicumarol caused significant mitochondrial depolarization, albeit with slower dynamics than DCPIB, resulting in a 37% reduction in the TMRE fluorescence (Fig. 5H) compared with the 73% reduction observed with DCPIB after ~5 min of treatment (Fig. 4D). Consistent with its weaker, yet significant, effect on depolarizing the mitochondrial membrane potential compared with DCPIB, dicumarol treatment led to substantial actin aggregation after 6 h, although with slower kinetics (Fig. 5I, J and Fig. 4I, J). Notably, these effects of dicumarol were completely prevented by the addition of serum in the Ringer’s buffer, supporting the notion that VRAC inhibitors are sequestered by albumin (Fig. 5I, J). To determine whether the effects of dicumarol on mitochondrial depolarization and actin aggregation are accompanied by T cell apoptosis and dependent on cell permeability, we evaluated the survival of Jurkat cells and primary murine T cells treated with 20 μM or 40 μM dicumarol in both serum-containing and serum-free media. We found that, in serum-free conditions, both concentrations of dicumarol induced apoptosis in ~50% of Jurkat cells compared to ~25% in the vehicle control (Supplementary Fig. 4H). In line with our findings in the human T cell leukemia line, dicumarol also induced a dose-dependent increase in apoptosis of murine activated CD4^+^ and CD8^+^ T cells after 24h of treatment in serum-free medium (Supplementary Fig. 4I, J). Taken together, these data demonstrate that the novel VRAC inhibitor dicumarol induces T cell apoptosis similarly to DCPIB. The effects of both VRAC inhibitors seem to depend on their permeation through the cell membrane, which may compromise their efficacy in functional in vitro experiments using standard culture conditions.
DCPIB inhibits VRAC independently of its effects on mitochondrial depolarization and Ca2+ signaling
Given that both DCPIB and dicumarol, which are structurally unrelated molecules, exhibit similar effects on mitochondrial depolarization and Ca^2+^ signaling while inhibiting VRACs, we were tempted to determine whether these effects contribute to the suppression of VRACs. To test this hypothesis, we used the mitochondrial uncoupler FCCP, which suppresses SOCE and induces actin aggregation similarly to DCPIB and dicumarol. Jurkat cells exposed to hypotonic solution displayed a robust outwardly rectifying IVRAC that was sensitive to DCPIB (Supplementary Fig. 5A, B). However, treatment with 1 μM FCCP in the bath hypotonic solution did not affect the IVRAC (Supplementary Fig. 5A, B). Current densities recorded in whole-cell configuration over time exhibited the delayed activation of IVRAC when exposed to hypotonic challenge (Supplementary Fig. 5A, B). Although a 10-min exposure to 1 μM FCCP in the hypotonic bath solution was insufficient to inhibit IVRAC, a subsequent application of 20 μM DCPIB completely eliminated these currents (Supplementary Fig. 5C). Additional properties of IVRAC, such as voltage-dependent inactivation at large positive membrane potentials, were comparable between the hypotonic bath solutions with and without 1 μM FCCP (Supplementary Fig. 5D, E). However, subsequent application of 20 μM DCPIB completely inhibited IVRAC across all tested voltages (Supplementary Fig. 5D, E). These findings indicate that the inhibition of VRAC by DCPIB occurs independently of its effects on mitochondrial depolarization and Ca^2+^ signaling, as FCCP, which mimics those effects, failed to suppress IVRAC.
The suppression of SOCE by DCPIB and dicumarol is independent of LRRC8A/C expression
In T cells, functional VRACs are mainly composed of LRRC8A and LRRC8C paralogs (12, 33). In this study, we have shown that DCPIB alters intracellular Ca^2+^ signaling in T cells, regardless of LRRC8C expression (Fig. 1A, B). Other studies have found that DCPIB also affects the mitochondrial membrane potential, independent of LRRC8A expression, in both the HAP-1 human myeloid leukemia cell line and HEK293 human embryonic kidney cells (22). Although it has been proposed that DCPIB interacts directly with the R103 residues of LRRC8A homomeric complexes (18, 19), conflicting studies indicate that LRRC8A homomeric channels exhibit poor pharmacological responses to DCPIB (20). Additionally, homomeric LRRC8C-8A(IL1^25^) chimeric channels, which lack R103 residues and instead have a leucine residue at the homologous position within the extracellular selectivity filter, are still inhibited by DCPIB (20). Therefore, we cannot rule out the possibility that DCPIB may interact with other regions of the channel not resolved in the Cryo-EM structures or with paralogs other than LRRC8A (18, 22). Since both DCPIB and dicumarol exhibit similar effects while targeting VRAC, we aimed to investigate whether the suppression of SOCE in T cells by these small molecules depends on LRRC8/VRAC expression. To this end, we targeted LRRC8A and LRRC8C expression in T cells by generating mice lacking LRRC8A specifically in T cells (Lrrc8a^fl/fl^Cd4^Cre+^, referred to as Lrrc8a^CD4^ mice) in a LRRC8C knockout background (Lrrc8a^CD4^Lrrc8c^−/−^ mice). Thymocytes from Lrrc8a^CD4^Lrrc8c^−/−^ mice lack expression of both LRRC8A and LRRC8C (Fig. 6A, B). Moreover, while purified CD4^+^ T cells from wild-type mice exhibited the DCPIB-sensitive outwardly rectifying IVRAC upon hypotonic challenge, Lrrc8a^CD4^Lrrc8c^−/−^ CD4^+^ T cells completely failed to elicit IVRAC under hypotonic conditions (Fig. 6C–E). Next, we performed intracellular Ca^2+^ measurements in CD4^+^ T cells treated with ionomycin to deplete the ER-Ca^2+^ stores and subsequently challenged them for 10 min with either DCPIB or dicumarol before SOCE induction. As observed in Jurkat cells (Fig. 3C, D) and primary human T cells (Supplementary Fig. 3C, D), murine CD4^+^ T cells from wild-type mice also displayed a significant decrease in SOCE induction when treated with DCPIB or dicumarol (Fig. 6F). Interestingly, SOCE in Lrrc8a^CD4^Lrrc8c^−/−^ CD4^+^ T cells was similarly suppressed by DCPIB and dicumarol (Fig. 6G). These results strongly suggest that the effect of DCPIB and dicumarol in altering Ca^2+^ signaling in T cells is completely independent of the presence of functional VRACs (i.e., off-target).
DCPIB and dicumarol treatments fail to replicate the effects of T cells lacking functional VRACs
The canonical function of LRRC8/VRACs to regulate cell volume is facilitated by the efflux of Cl^−^ and other osmolytes under hypotonic conditions. However, growing evidence indicates that these channels also have non-canonical functions involving the influx of substrates that support cell signaling and physiology under normal tonicity (3, 4). We and others have previously shown that both LRRC8C and LRRC8A facilitate the transport of cyclic dinucleotides (CDNs) into T cells (12, 33, 62). The influx of naturally occurring CDNs, such as the eukaryotic-derived 2’3’-cGAMP, as well as the bacteria-derived 3’3’-cGAMP, c-di-AMP, and c-di-GMP, activates the Stimulator of Interferon Genes (STING) protein located in the ER (12, 63, 64). In innate immune cells and other non-immune cells, STING plays a critical role in regulating type-I IFN response to cytosolic DNA (63). However, in T cells, strong STING activation leads to proliferation arrest and apoptosis (12, 65–70). Given the significant role of LRRC8A/C-containing VRACs in cGAMP transport, we utilized this function of VRACs to assess the ability of DCPIB and dicumarol to inhibit the VRAC-dependent CDN-mediated cell death in T cells (Fig. 7A). First, we validated that deleting the functional VRAC subunits in T cells, LRRC8A and LRRC8C, conferred protection against CDN-induced cell death. When we stimulated wild-type CD4^+^ T splenocytes with anti-CD3+CD28 in the presence of 2’3’-cGAMP or 3’3’-cGAMP for 3 days, we observed the expected increase in apoptosis (Fig. 7B, C), as previously reported (12, 70). However, deletion of LRRC8A and LRRC8C in T cells completely prevented CDN-induced apoptosis (Fig. 7B, C). To confirm that Lrrc8a^CD4^Lrrc8c^−/−^ T cells possess functional STING signaling, we bypassed VRACs requirement and directly induced STING activation by using the small-molecule STING agonist DMXAA (12). Treatment with DMXAA resulted in complete cell death in both wild-type and Lrrc8a^CD4^Lrrc8c^−/−^ CD4^+^ T cells after 3 days of anti-CD3+CD28 stimulation (Fig. 7B, C). These results confirm that LRRC8A/C-containing VRACs facilitate CDN transport in T cells, leading to T cell death. We next tested the ability of DCPIB and dicumarol to prevent CDN-induced T cell death by challenging CD4^+^ T splenocytes with 2’3’-cGAMP or 3’3’-cGAMP in the presence of 20 μM or 40 μM of DCPIB or dicumarol in standard culture conditions (i.e., supplemented with 10% FBS). We hypothesized that albumin in the serum would prevent the membrane accumulation and/or permeation of DCPIB and dicumarol, as in these conditions they lack off-target effects on CD4^+^ T cell apoptosis (Fig. 1H and Supplementary Fig. 4I). We also speculate that the extracellular concentration of the effective free inhibitors may be sufficient to inhibit VRACs at the extracellular selectivity filter, as it has been proposed (18, 19). In the absence of STING agonists, no significant changes in apoptosis were observed in the presence of DCPIB or dicumarol at the two concentrations tested (Fig. 7D, E). As expected, 2’3’-cGAMP and 3’3’-cGAMP induced apoptosis in CD4^+^ T cells in the absence of VRAC inhibitors (Fig. 7D, E). Surprisingly, treatment with either DCPIB or dicumarol at 20 μM or 40 μM failed to prevent CDN-induced apoptosis in CD4^+^ T cells (Fig. 7D, E). As anticipated, DMXAA resulted in complete apoptosis, regardless of DCPIB or dicumarol treatment (Fig. 7D, E). These findings suggest that CDN-induced T cell death is not prevented when VRAC inhibitors are applied in standard culture conditions, likely due to their lack of membrane absorption and cell permeability.
To confirm these results, we directly assessed the efficacy of DCPIB in regulating the activation of STING upon CDN challenge. Thus, we treated wild-type CD4^+^ T cells for 3h and 6h with 2’3’-cGAMP, with and without DCPIB, in Ringer’s buffer lacking serum or albumin. Interestingly, DCPIB completely suppressed STING phosphorylation compared to the vehicle control treatment (Supplementary Fig. 6A). However, the lack of STING phosphorylation was accompanied by the off-target effect of DCPIB, characterized by actin aggregation after 3h and 6h of treatment (Supplementary Fig. 6A). To determine whether these effects of DCPIB on suppressing STING phosphorylation were due to its off-target effects rather than its role in inhibiting VRAC-mediated CDN transport, we challenged CD4^+^ T cells with the STING agonist DMXAA, which does not require VRAC for its transport. We observed a similar suppression of STING phosphorylation by DCPIB following DMXAA treatment as compared with 2’3’-cGAMP treatment (Supplementary Fig. 6B). We concluded that DCPIB-mediated suppression of STING phosphorylation is likely due to its off-target effect on actin aggregation rather than its role in inhibiting CDN transport, as i) a similar effect was observed with DMXAA treatment, and ii) STING trafficking from the ER to the Golgi, required for its phosphorylation (71), dependents on an intact actin cytoskeleton (72). Finally, to support this conclusion, we repeated the experiments in CD4^+^ T cells challenged with 2’3’-cGAMP and DMXAA under standard culture conditions. We found that both 2’3’-cGAMP and DMXAA triggered STING phosphorylation regardless of the presence of DCPIB (Supplementary Fig. 6C, D). However, when we treated CD4^+^ T cells with the STING inhibitor H-151 as a control (73), we observed a robust reduction in STING phosphorylation after stimulation with 2’3’-cGAMP and DMXAA (Supplementary Fig. 6C, D). Taken together, our results strongly suggest that DCPIB fails to suppress VRAC-mediated CDN transport when its permeation is hindered by serum, indicating that DCPIB requires access to the cell membrane and intracellular compartment to effectively inhibit VRAC transport.
DISCUSSION
We here show that both the on-target and off-target effects of two structurally unrelated VRAC inhibitors, DCPIB and dicumarol, are dependent on their accumulation in membranes and permeability into the cell. Both DCPIB and dicumarol deplete intracellular Ca^2+^ stores, with DCPIB promoting the release of intracellular Ca^2+^ from ER and non-ER sources and dicumarol only triggering Ca^2+^ release from non-ER sources. Additionally, both drugs simultaneously suppress SOCE and induce mitochondrial depolarization. We speculate that the loss of mitochondrial membrane potential following treatment with either DCPIB or dicumarol results in the release of Ca^2+^ from mitochondria, contributing to an increase in cytosolic Ca^2+^ levels. These variations in intracellular Ca^2+^ ultimately culminate in the suppression of SOCE and a reduction in cytokine production by T cells. Our findings are consistent with previous research indicating that alterations in mitochondrial Ca^2+^ signaling, resulting from depolarization of the inner mitochondrial membrane, can suppress SOCE (45, 55–60, 74). Moreover, the sustained mitochondrial depolarization induced by both small-molecule VRAC inhibitors leads to time-dependent mitochondrial damage, accumulation of actin aggregates, and apoptosis in both human and mouse T cells. Such shared off-target effects render these inhibitors unsuitable for investigations focused on intracellular signaling regulated by VRAC-dependent transport.
DCPIB is a non-diuretic derivative of ethacrynic acid (75), an FDA-approved drug primarily used to treat edema resulting from heart failure, liver disease, or kidney disorders (76). In addition to DCPIB, other derivatives of ethacrynic acid (e.g., IAA-94, DIOA, and the SN-40X series) have also been characterized as VRAC inhibitors (19, 77–85), underscoring the significance of the chemical structure of ethacrynic acid in mediating this inhibitory action on these channels. DCPIB was first identified as a selective blocker of VRACs in calf pulmonary artery endothelial cells and guinea-pig atrial cardiomyocytes (17). Its selectivity has been systematically tested against several Cl^−^ channels, including CFTR, CaCC, CLCs, Maxi-CL, and PAC, as well as various K^+^ channels, including Kv1.5, Kv4.3, minK and HERG (17, 85, 86). Nonetheless, emerging evidence has illustrated that DCPIB has multiple off-targets effects on signaling pathways and membrane proteins across a diverse range of cell types. These off-target effects include glutamate transport pathways (21), mitochondrial respiration (22), vascular endothelial growth factor receptor 2 (VEGFR2) signaling (23), the H^+^/K^+^ ATPase (24), inward-rectifier K^+^ (K_ir_) channels (25), two-pore domain K^+^ (K_2P_) channels (26–28), large-conductance Ca^2+^-activated K^+^ (BK) channels (29), and the pannexin 2 channel (30). To date, no evidence has been reported regarding off-target effects on SOCE following treatment with DCPIB. Interestingly, many of the off-target effects observed with DCPIB also apply to its parent compound, ethacrynic acid. For instance, ethacrynic acid has been demonstrated to inhibit mitochondrial Complex III, alter mitochondrial morphology, induce oxidative stress, and trigger apoptosis (87–90). Notably, some of these effects manifest several times faster in buffer conditions than in cell culture media (89). Therefore, it is plausible that the off-target effects of DCPIB arise from its chemical nature as a derivative of ethacrynic acid.
On the other hand, dicumarol is a naturally derived anticoagulant originating from the fungal spoilage of coumarin in sweet clover. It was developed as a pharmaceutical agent and approved by the FDA for the prevention of blood clotting through the inhibition of vitamin K metabolism. More recently, dicumarol was identified as a novel inhibitor of VRACs (13). Furthermore, additional coumarin-derived molecules, such as V-116 (91), the rodenticide bromadiolone (92), and naturally occurring sesquiterpene coumarins like galbanic acid and karatavic acid (93), have also exhibited potent VRAC inhibition, suggesting that coumarin derivatives may share properties as VRAC inhibitors. However, dicumarol has also been shown to influence mitochondrial metabolism in cancer cells by inhibiting pyruvate dehydrogenase kinase 1 (PDK1), leading to ROS accumulation, depolarization of mitochondrial membrane potential, and apoptosis (94). Thus, similar to ethacrynic acid-derivatives, coumarin-based small molecules also display multiple off-target effects causing mitochondrial damage and apoptosis, a reason for their repurposing as anti-cancer agents (94–97).
Considering the shared off-target effects of DCPIB and dicumarol reported here, we explored whether changes in mitochondrial homeostasis contributed to the suppression of VRACs, similar to their effects on SOCE. We found that the mitochondrial uncoupler FCCP, which mimics the off-target effects of DCPIB and dicumarol, did not inhibit IVRAC, suggesting that the inhibition of VRACs by these small molecules is indeed a targeted effect. This conclusion is consistent with previous studies, which demonstrate: i) the suppression of swelling-activated IVRAC is reversible upon removal of these inhibitors in hypotonic solutions (13, 17), and ii) DCPIB inhibits single-channel currents in reconstituted VRAC complexes in lipid bilayers (6, 98). Interestingly, earlier studies using both whole-cell and single-channel recordings from cell-attached patches of rat pancreatic β-cells established that the modulatory effects of DCPIB on VRACs are concentration-dependent and influenced by exposure duration, regardless of the experimental configuration tested (99). Consequently, it was concluded that the kinetics of VRAC inhibition by DCPIB likely occur when DCPIB permeates the cell membrane and reaches a sufficient intracellular concentration, which explains the delay in both its inhibitory and recovery effect on IVRAC after addition and washout maneuvers in hypotonic solution (99). This process may depend on the binding and accumulation of DCPIB in cellular lipid membranes due to its strong hydrophobicity, (99), similar to other less-selective VRAC channel inhibitors (100). Supporting this notion, our membrane permeability assays indicate that both DCPIB and dicumarol are markedly retained within lipid membranes.
More recently, an independent study using the VRAC-mediated iodide quenching assay in the human glioma cell line LN215 demonstrated that both the coumarin derivative V-116 and DCPIB exhibit a ~5-fold decrease in their IC_50_ values for suppressing VRAC transport in the presence of serum (91). In our study involving human and mouse T cells, we found that all observed effects from DCPIB and dicumarol treatment, regardless of whether they influence VRAC transport or manifest off-target effects, are exacerbated in serum-free or albumin-free conditions, and are nearly entirely abolished by the presence of albumin in the buffer or by serum under standard culture conditions. This phenomenon can be attributed to the well-established role of albumin, the predominant serum protein, in binding, transporting, and facilitating the removal of drugs and small molecules (52). Therefore, we propose that VRAC inhibition by DCPIB and dicumarol may arise from their accumulation in the plasma membrane or from their intracellular permeation due to their lipophilic characteristics, effects that are mitigated by the presence of albumin in our membrane permeability studies.
Recent investigations have revealed that two structurally distinct cysteinyl leukotriene receptor antagonists, zafirlukast and pranlukast, inhibit VRACs independently of their role as CysLT1R antagonists (39), and such inhibition occurs through direct interaction with VRACs (40). Interestingly, they found that the drug inhibition involves the disruption of protein-lipid interactions and destabilization of the pore in the LRRC8C-8A(IL1^25^) homomeric chimera and LRRC8A/C heteromeric channels, thereby proposing a novel and conserved mechanism of VRAC inhibition by lipophilic small molecules. Intriguingly, as highlighted in that study, the wide variety of chemical scaffolds with inhibitory properties on VRACs share a common high partition coefficient (LogP) due to their lipophilic profiles (40, 101). A comprehensive list of inhibitors that block VRACs in a voltage-independent manner includes the estrogen receptor antagonist tamoxifen (102, 103), ethacrynic acid derivatives (DCPIB, DIOA, and IAA-94) (17, 77, 78, 85), unsaturated fatty acids (oleic and arachidonic acids) (104) and glycyrrhetinic acid derivatives (carbenoxolone and 18 β-glycyrrhetinic acid) (105, 106), coumarin-derivatives (bromadiolone, tioclomarol, and dicumarol) (13, 92), CysLT1 antagonist (zafirlukast and pranlukast) (39, 40), fenamate derivatives (flufenamic acid, mefenamic acid, fenamic acic, and niflumic acid) (85, 104, 107, 108), Ferula sesquiterpenes (galbanic acid, karatavic acid, and ferutinin) (93), sulfonylurea and diaryl urea derivatives glibenclamide and N-[3,5-bis(Trifluoromethyl)phenyl]-N’-[4-bromo-2-(2H-tetrazol-5-yl)phenyl]urea (NS3728), respectively (102, 103, 109, 110), stilbene derivatives (DIDS and SITS) (102–104, 111), 5-nitro-2-(3-phenylpropylamino)benzoic acid (NPPB) (103, 104, 111), verapamil (112), the quinoline derivative mefloquine (113), the naturally occurring alkaloids quinine and quinidine quinidine (114), and the flavonoid phloretin (115). The LogP values of these molecules range from 7.1 for tamoxifen to 2.6 for phloretin. While none of these inhibitors are specific to VRACs, it is plausible that they share a similar mechanism of inhibition due to hydrophobic interactions with the cell membrane, which disrupts the interactions between lipid-transmembrane domain of the channel. In fact, emerging evidence based on the structural analysis of homomeric and heteromeric VRAC complexes suggest that lipid-channel interactions may play a role in channel gating (116–118). However, additional properties of these chemicals, beyond their hydrophobicity, may be necessary for effective VRAC inhibition.
We demonstrate that genetic deletion of LRRC8A/C abolishes swelling-activated IVRAC and mitigates the deleterious effect of cGAMP transport in T cells. However, both DCPIB and dicumarol were unable to reproduce the effect seen with the genetic deletion of VRACs under standard culture conditions, even when using concentrations 2 to 4-fold higher than those required to inhibit the channel in electrophysiological studies. We hypothesize that albumin in serum neutralizes the small-molecule inhibitors extracellularly, allowing the hydrophilic cGAMP to permeate VRACs and activate STING, leading to T cell apoptosis. These findings challenge the current cork-in-bottle mechanism of VRAC inhibition by DCPIB, as the concentration of free inhibitor never reaches a level sufficient to inhibit the channel extracellularly. This is due to the lipophilic nature of DCPIB which makes it more likely to bind to albumin than the channel itself. Therefore, it is plausible to speculate that VRAC inhibition occurs due to hydrophobic interactions at the plasma membrane or through intracellular interactions once the small-molecule inhibitor gains access to the cytosol, rather than through hydrophilic interactions between the carboxylic group of DCPIB and the R103 residues located at the extracellular selectivity filter of VRACs. These findings present a significant challenge to studies attempting to inhibit VRACs in physiological settings, as the inhibitory effects observed in electrophysiological studies will not necessarily translate to effective inhibition of VRACs in culture conditions. Consequently, the removal of small-molecule scavengers may come at the cost of the off-target effects trigger by the permeation of these inhibitors.
Our findings show that while DCPIB and dicumarol are effective inhibitors of VRACs in short-term electrophysiological studies, their effectiveness in other contexts is limited due to their cell permeability and off-target effects on altering mitochondrial homeostasis and Ca^2+^ signaling. Therefore, the genetic elimination of LRRC8 proteins is currently the only recommended approach to validate the role of VRAC-mediated transport in regulating intracellular signaling. Our kinetic experiments highlight the off-target effects of these small-molecule inhibitors on Ca^2+^ signaling, mitochondrial depolarization, mitochondrial superoxide accumulation, actin aggregation, and apoptosis, and illustrate the cellular consequences of removing small-molecule scavengers such as albumin (Fig. 8). These effects need to be considered when testing and interpreting the impact of VRAC inhibition by these compounds. Furthermore, we suggest the need for better blockers that can be utilized for functional studies of VRACs in physiological settings.
MATERIALS AND METHODS
Mice.
Lrrc8c^−/−^ mice (B6.129S2-Fad158<tm1Maim>/7Maim, RBRC05143, Riken) were generated by crossing Lrrc8c^+/−^ mice, as previously described (12, 119). Lrrc8a^fl/fl^Cd4^Cre+^ (referred to as Lrrc8a^CD4^) mice were generated by crossing Lrrc8a^fl/fl^ mice (kindly provided by Dr. Rajan Sah) with Cd4^Cre^ mice (The Jackson Laboratory strain 022071). The Lrrc8a^CD4^ mice were then crossed with Lrrc8c^−/−^ to obtain Lrrc8a^CD4^Lrrc8c^−/−^ double-mutant mice. Wild-type mice (Lrrc8a^fl/fl^) were obtained from breeding of Lrrc8a^fl/fl^Cd4^Cre+^ and Lrrc8a^fl/fl^Cd4^Cre−^ mice, and co-housed with Lrrc8a^CD4^Lrrc8c^−/−^ mice. All mice were maintained on a C57BL/6 genetic background, and male and female mice were used between 8 and 20 weeks of age. Mice were maintained under specific pathogen-free conditions with a 12 h dark/light cycle, at 22–25 °C and 50–60% humidity with water and food provided ad libitum. All experiments were conducted in accordance with protocols approved by the Institutional Animal Care and Use Committee at the University of Chicago.
Human samples.
T cells from deidentified healthy donors were isolated from whole blood by density gradient centrifugation using Ficoll-Paque plus (GE Amersham) and expanded in vitro as previously described (120). Informed consent for the studies was obtained from healthy donors in accordance with the Institutional Review Board approval at the University of Chicago.
Cell lines, recombinant DNA, and production of pseudotyped retrovirus.
Jurkat cells (kind gift from Dr. Erin J. Adams) were cultured in RPMI 1640 media (Corning) supplemented with 10% FBS (Corning) and 1% penicillin plus streptomycin (Gibco). HEK293 cells were cultured in DMEM media (Corning) supplemented with 10% FBS and 1% penicillin plus streptomycin (full DMEM). Platinum E (Plat-E) packaging cells were cultured in full DMEM medium supplemented with 10 μg/ml Blasticidin S (Invivogen) and 1 μg/ml Puromycin (Sigma). For overexpression of YFP-ORAI1 (generously provided by Dr. Murali Prakriya) in HEK293 cells, the YFP-ORAI1 DNA was subcloned into the pMSCV-hygro retroviral vector (Addgene). Retroviral supernatant was produced in the Plat-E retroviral packaging cell line following their transfection with retroviral expression plasmids and the amphotropic packaging vector pCL-10A1 using lipofection (GenJet, SignaGen). Retroviral supernatant was collected 36 and 60h after transfection and used to spin-infect HEK293 cells in the presence of 8 μg/ml polybrene. After 3 days, HEK293 cells were selected with 200 μg/ml hygromycin B (Thermo Fisher Scientific) for additional 3 days, and transduction efficiency was confirmed by detecting YFP^+^ cells via flow cytometry. All cell lines were cultured at 37°C with 5% CO_2_ and passaged twice a week.
Isolation, activation, and culture of murine T cells.
Mouse splenocytes were isolated as previously described (121). Briefly, spleens were placed in a 70 μm strainer over a 50 ml conical tube, rinsed, ground in ACK lysis buffer (in mM: 150 NH_4_Cl, 1 KHCO_3_, 0.1 Na_2_EDTA, pH 7.2). The resulting cell suspension was passed through the strainer, collected in the 50 ml conical tube, and centrifuged at 750 x g for 5 min. The supernatant was discarded, and cell pellets were resuspended in FACS buffer for the isolation of CD4^+^ T cells or in full RPMI media for further activation and treatments. CD4^+^ T cells were purified using the MagniSort Mouse CD4^+^ T cell Enrichment Kit (Thermo Fisher Scientific) according to the manufacturer’s protocol. CD4^+^ T cells were stimulated in flat bottom 12-well plates (1×10^6^ cells/ml per well) with 1 μg/ml plate-bound anti-CD3 (clone 2C11) and 1 μg/ml anti-CD28 antibodies (clone 37.51; both Bio X cell) in full RPMI media (Cellgro) containing 10% FBS, 1% L-glutamine, 1% penicillin-streptomycin and 0.1% β-mercaptoethanol. For patch-clamp electrophysiology experiments, murine CD4^+^ T cells stimulated for 2 days were diluted twice with full RPMI media supplemented with 20 IU/ml hr-IL-2 and 20 ng/ml mr-IL-7 (both from Peprotech). Then, cells were detached and cultured in new 12-well plates. Alternatively, T cells were stimulated in round-U bottom 96-well plates (5×10^5^ cells/ml per well) with Dynabeads Mouse T-Activator CD3/CD28 (Thermo Fisher Scientific) at a 1:2 bead-to-cell ratio. For cytokine expression, activated T cells were re-stimulated with 1 μM ionomycin (Cayman) plus 20 nM phorbol myristate acetate (PMA, Sigma) for 5h in the presence of brefeldin A (eBioscience) and analyzed by flow cytometry as described below. Additionally, activated T cells with anti-CD3+CD28 were treated with 20 μM or 40 μM DCPIB (Tocris), 20 μM or 40 μM dicumarol (Sigma), 10 μM H-151 (InvivoGen), 20 μg/ml 2’3’-cGAMP, 10 μg/ml 3’3’-cGAMP, (both from InvivoGen), and 3 μg/ml DMXAA (Cayman).
Intracellular Ca2+ [Ca2+]i measurements.
Human T cells, Jurkat cells, or purified activated murine CD4^+^ T cells were loaded with 1 μM Fura-2-AM (Invitrogen) for 30 min in serum-free RPMI1640 medium at 37°C and 5% CO_2_. Fura-2 emissions at 510 nm were measured following excitation at 340 nm and 380 nm (F340/380 ratios) using the FlexStation 3 Multi-Mode Microplate Reader (Molecular Devices), as previously described (121). Briefly, Fura-2-AM-loaded T cells were plated on poly-L-lysine-coated 96-well plates (Falcon) at a density of 1×10^6^ cells/well and incubated for 15 min in Ringer’s buffer containing (in mM): 155 NaCl, 4.5 KCl, 2 CaCl_2_, 1 MgCl_2_, 10 D-glucose, and 5 HEPES (pH 7.4). Then, T cells were washed twice with Ca^2+^-free Ringer’s solution, which contained (in mM): 155 NaCl, 4.5 KCl, 3 MgCl_2_, 10 D-glucose, and 5 HEPES (pH 7.4). Store depletion was monitored in Ca^2+^-free Ringer’s solution by stimulating the cells with 1 μM thapsigargin (TG, Cayman), 1 μM ionomycin, 5 μg/ml CD3 monoclonal antibody (OKT3) to activate TCR signaling in human T cells, 20 μM DCPIB, 20 μM dicumarol, or 1 μM FCCP (Sigma). In some experiments, cells were pre-treated with 50 μM 2-APB or 400 μM ryanodine to block ER-Ca^2+^ store mechanisms prior to conducting the experiments, and these treatments were maintained during the recordings. Ca^2+^ influx was induced by adding Ca^2+^-containing Ringer’s solution to reach a final concentration of 1 mM or 2 mM Ca^2+^, as indicated in each experiment. [Ca^2+^]i, was represented as the area under the curve (AUC) for the store depletion and Ca^2+^ influx as specified in each experiment.
Flow cytometry.
Cells from tissue culture or isolated from mouse spleens were washed in ice-cold PBS containing 3% FBS and 2 mM EDTA (FACS buffer). For apoptosis detection, both Jurkat cells and mouse T cells were labeled in the dark for 10 min with Annexin V-Alexa-Fluor 647, using Annexin V binding buffer (all from BioLegend). Additionally, murine T cells were co-stained with a combination of fluorescently labelled antibodies: anti-CD4-PE-Cy7 and anti-CD8-APC-Cy7. Intracellular cytokine staining was performed after fixing the cells with intracellular (IC) fixation buffer (BioLegend) for 30 min at room temperature. The cells were then stained for 45 min with anti-CD4-PE-Cy7, anti-CD8-PercP-Cy5.5, anti-IL2-FITC, anti-TNFα-APC in FoxP3 permeabilization buffer (all from BioLegend). For measuring mitochondrial membrane potential (Δψm), Jurkat cells were incubated for 30 min with 100 nM tetramethylrhodamine ethyl ester (TMRE, Invitrogen) dye in serum-free RPMI media, then washed and resuspended in Ringer’s solution. To detect mitochondrial superoxide (O_2_^−^), Jurkat cells were loaded with 1μM MitoSOX (Invitrogen) for 15 min in serum-free RPMI media at 37°C and 5% CO_2_, then washed and resuspended in Ringer’s solution. Changes in Δψm and mitochondrial O_2_^−^ were monitored using specific excitation/emission wavelengths: for Δψm, λ Ex/Em = 561/574 nm; for mitochondrial O_2_^−^, λ Ex/Em = 405/650 nm. Time traces were recorded 90 s to establish a baseline; at this point, treatments were added to the cells, and the recordings continued for an additional ~5 min. Samples were acquired on a BD FACSymphony^™^ A3 cell analyzer (BD Biosciences) and analyzed using FlowJo software (FlowJo 10.10.0).
Patch-clamp electrophysiology.
Jurkat cells or primary murine CD4^+^ T cells were seeded in poly-lysin coated rounded cover glass. Patch-clamp experiments were conducted in the whole-cell configuration using the PatchMaster software and the EPC9 amplifier (HEKA), and recordings were sampled at 2 kHz and filtered at 1 kHz. Only isolated individual cells were picked. All recordings were performed at room temperature (~23 °C) using an isotonic bath solution containing (in mM): 150 NaCl, 6 KCl, 2 MgCl_2_, 1.5 CaCl_2_, 10 HEPES and 10 D-glucose (pH 7.4, ~310 mOsm). Hypotonic bath solution contained (in mM): 105 NaCl, 6 CsCl, 1 MgCl_2_, 1.5 CaCl_2_, 10 HEPES and 10 D-glucose (pH 7.4, (~210 mOsm). The pipette solution contained (in mM): 100 Cs-methanesulfonate, 1.9 CaCl_2_, 40 CsCl, 1 MgCl_2_, 4 Na_2_-ATP, 5 EGTA, and 10 HEPES (pH 7.4, ~310 mOsm). Borosilicate pipettes were fabricated using a P-97 puller (Sutter), fired polished with a micro forge (Narishigue) and had a 4–6 MΩ resistance in isotonic solution. Cells were held at −70 mV and baseline traces were recorded with a ramp protocol from −80 to +90mV in isotonic solution, then cells were perfused with a hypotonic solution and ramp protocols were applied every 30 s until currents reached the maximum amplitude. Voltage steps protocols were then applied from −70 mV to +140 mV at +10 mV voltage intervals for 40 ms. VRAC inhibitors DCPIB and dicumarol (both at 20 μM), and mitochondrial uncoupler FCCP (1 μM) in hypotonic solution were directly perfused into the recording chamber.
Compound permeability assay.
In vitro membrane permeability of VRAC inhibitors was performed using parallel artificial membrane permeability assay (PAMPA) (53, 54). Briefly, lipid-oil-lipid tri-layer membranes were assembled on a 96-well polycarbonate filter (donor) plate and a matching 96-well plate (acceptor) was used as the receiver plate. Compound solutions were prepared in assay buffer (2% DMSO in PBS, pH 7.4) by diluting 25 mM DMSO stock solutions in PBS to a final concentration of 50 μM, and 200 μL were added to each well of the donor plate. Separately, 300 μL of assay buffer was added to each well of the acceptor plate. The donor plate was carefully placed on top of the acceptor plate, ensuring no gap or bubbles between the bottom of filter plate wells and the liquid surface of accepter wells. This sandwich setup was incubated in a humid chamber for 5 h. After the incubation period, the concentrations of compounds in both the donor wells and the acceptor wells were quantified by liquid chromatography-mass spectrometry (LC-MS). Total drug concentration in the donor wells solutions was determined after treatment with 50 % acetonitrile to release the drug bound to BSA. The apparent permeability was calculated using the following equation:
where P_e_ is permeability in the unit of cm/s. A = effective filter area (0.28 cm^2^); V_D_ = donor well volume (0.2 mL); V_A_ = acceptor well volume (0.3 mL); t = incubation time (18,000 s); C_A_ = compound concentration in acceptor well after incubation, and C_eq_ is the equilibrium concentration as if there was no membrane, calculated with following equation:
where C_D_ = compound concentration in donor well after incubation.
Mass retention (R) was calculated with the following equation:
where C_0_ = initial donor concentration (50 μM). Propranolol and Furosemide were used as high- and low-permeability controls, respectively.
Total internal reflection fluorescence (TIRF) microscopy.
HEK293 cells with stable overexpression of YFP-hORAI1 were plated onto 35 mm glass bottom dishes with 10 mm wells (#1.5 coverslip, Cellvis) coated with poly-L-lysine (Sigma) and then transfected with mCherry-hSTIM1. Eight hours after transfection, cells were treated with 20 μM DCPIB, 20 μM dicumarol, or 1.5 μM FCCP for 10 minutes followed by 1 μM thapsigargin for another 10 minutes, then fixed using 4% PFA. All treatments were performed in Ca^2+^-free Ringer’s solution. TIRF images were acquired using a Leica SR GSD 3D localization microscope, a 160x / N.A. 1.43 oil-immersion objective (Leica), and an iXon Ultra 897 EMCCD camera (Andor). Two laser lines (488 and 543 nm) were used to acquire fluorescent images. The colocalization of YFP-hORAI1 with mCherry-hSTIM1 was measured by using Fiji’s Coloc 2 plugin to calculate the Pearson’s correlation coefficient for each cell.
Immunoblotting.
Murine thymocytes and activated CD4^+^ T cells were used to prepare whole cell lysates using NP-40 lysis buffer (20 mM Tris-Cl, 150 mM NaCl, 5 mM EDTA, 1% Nonidet P40, 1 mM PMSF, 1 mM Na_3_VO_4_, 1% protease inhibitor cocktail, 1X PhosStop). For experiments using Jurkat cells, 1×10^6^ cells were seeded in 96-well plates in Ringer’s solution supplemented with or without 10% FBS. Jurkat cells were treated with 20 μM DCPIB, 20 μM dicumarol, or 1.5 μM mitochondrial uncoupler FCCP for 0, 0.5, 1, 3, or 6 h, and were kept at 37 °C with 5% CO_2_. Cells were washed twice with ice cold PBS and used to prepare whole cell lysates with NP-40 lysis buffer. The detergent soluble and detergent insoluble protein fractions were separated by centrifugation at 13,000 x g for 5 min. Detergent insoluble proteins were solubilized by heating the pellet in SDS lysis buffer (2% SDS, 50 mM Tris-HCl, 10 mM EDTA, 10% glycerol, 1x PhosStop, 1% protease inhibitor cocktail) at 50 °C for 1 hour; remaining cell debris was removed by centrifuging at 13,000 x g for 5 min. Protein extracts were treated with 4X Laemmli sample buffer (Bio-Rad) supplemented with β-mercaptoethanol (Thermo Fisher Scientific) and heated at 95 °C for 5 min (except for LRRC8A and LRRC8C protein detection). Equal protein contents (250 μg per lane) were then separated by SDS-PAGE using Novex^™^ WedgeWell^™^ 4–20% tris-glycine gel (Invitrogen) and transferred to polyvinylidene difluoride (PVDF) membranes. Membranes were blocked in 5% BSA in TBST for 2 h at room temperature and then incubated overnight at 4°C with the following primary antibodies: HSP60 rabbit mAb (D6F1, 1:1000), STING rabbit mAb (D2P2F, 1:1000), phospho-STING-Ser365 rabbit mAb (D8F4W, 1:1000, all from Cell Signaling Technology), a custom-made rabbit anti-mouse LRRC8A polyclonal antibody that recognizes the last 14aa at the C terminus of LRRC8A (1:1000) and a custom-made rabbit anti-mouse LRRC8C polyclonal antibody that recognizes the last 17aa at the C terminus of LRRC8C (1:1000, both from YenZym Antibodies), and a mouse anti-β-Actin mAb (1:5000, 66009–1-Ig, proteintech). Immunoreactive bands were detected after incubation with secondary antibodies coupled with horseradish peroxidase (A9044, A9169, both from Sigma-Aldrich) or Alexa-Fluor 647 (A21236, Invitrogen) and visualized by enhanced chemiluminescence or fluorescence (Amersham ImageQuant 800), respectively. Band densities were quantified and analyzed using Fiji.
RNA-seq data analysis.
RNA-seq datasets for purified CD4^+^ T cells from wild-type and Lrrc8c^−/−^ mice, and wild-type CD4^+^ T cells treated or not with DCPIB, were downloaded from NCBI Gene Expression Omnibus (GEO) under accession GSE163679 (12) using SRA Toolkit v.2.11.3. Technical replicate FASTQ files for each biological sample were concatenated prior to quantification. Transcript abundances were quantified using Salmon v.1.4.0 against the mouse transcriptome GRCm39 obtained through Ensembl. Transcript-level counts were imported into R v4.5.1 and aggregated to raw gene-level counts using tximport v1.36.1 (122) with transcript-to-gene mapping extracted from biomaRt v2.64.0 (123). Ensembl gene IDs were converted to gene symbols using biomaRt. Gene counts were normalized using the median-of-ratios method on DESeq2 v.1.48.1 (124). Differential gene expression analysis was performed using DESeq2 with a grouped experimental design. Pairwise comparisons at the same time points of treatment were generated for Lrrc8c^−/−^ and DCPIB-treated groups versus wild-type samples stimulated with anti-CD3+CD28. Gene lists were ranked on Wald statistic and Gene Set Enrichment Analysis (GSEA) was performed using clusterProfiler v4.16.0 referencing Hallmark pathways from mouse gene sets sourced from the Molecular Signatures Database (MSigDB) (49) via msigdbr v25.1.1. Comparative pathway analysis calculated normalized enrichment scores (NES) for each group independently and pathways were ranked by the magnitude of NES differences between Lrrc8c^−/−^ and DCPIB effects. Statistical significance was determined after a Benjamini-Hochberg correction for multiple testing (adjusted p < 0.05). Baseline T cell activation was assessed by comparing wild-type stimulated samples (24h and 48h) versus unstimulated 0h controls using Hallmark pathways.
Data analysis and statistics.
Data are expressed as mean ± SEM. No statistical methods were used to predetermine sample sizes, but our sample sizes are similar to those reported in previous publications. Data distribution was assumed to be normal, but this was not formally tested. Unless when restricted by the genotype, animals and cell plates were assigned randomly to experimental conditions. Data collection and analysis were not blinded to the conditions of the experiments. No data exclusion was performed. Statistical significance was performed using GraphPad Prism software (v.10.6.1) and determined by using two-tailed, unpaired Student’s t-test, and 2-way ANOVA corrected for Dunnett’s multiple comparisons tests, as indicated in the figure legends. Statistical analysis of RNA-Seq data was determined after a Benjamini-Hochberg correction for multiple testing using RStudio (v.4.5.1). A value of two-tailed P < 0.05 was considered statistically significant.
Supplementary Material
This is a list of supplementary files associated with this preprint. Click to download.
The reference list from the paper itself. Each links out to its DOI / PubMed record.
- 1Osei-Owusu J, Yang J, Vitery MDC, Qiu Z. Molecular biology and physiology of volume-regulated anion channel (VRAC). Curr Top Membr. 2018;81:177–203.30243432 10.1016/bs.ctm.2018.07.005PMC 6604840 · doi ↗ · pubmed ↗
- 2Strange K, Yamada T, Denton JS. A 30-year journey from volume-regulated anion currents to molecular structure of the LRRC 8 channel. J Gen Physiol. 2019;151(2):100–17.30651298 10.1085/jgp.201812138 PMC 6363415 · doi ↗ · pubmed ↗
- 3Thone FMB, Polovitskaya MM, Jentsch TJ. LRRC 8/VRAC chloride and metabolite channels in signaling and volume regulation. Trends Biochem Sci. 2025;50(10):873–91.40769835 10.1016/j.tibs.2025.07.001 · doi ↗ · pubmed ↗
- 4Yanushkevich S, Zieminska A, Gonzalez J, Anazco F, Song R, Arias-Cavieres A, Recent advances in the structure, function and regulation of the volume-regulated anion channels and their role in immunity. J Physiol. 2025;603(15):4255–91.39709525 10.1113/JP 285200 PMC 12182595 · doi ↗ · pubmed ↗
- 5Cannon CL, Basavappa S, Strange K. Intracellular ionic strength regulates the volume sensitivity of a swelling-activated anion channel. Am J Physiol. 1998;275(2):C 416–22.9688595 10.1152/ajpcell.1998.275.2.C 416 · doi ↗ · pubmed ↗
- 6Syeda R, Qiu Z, Dubin AE, Murthy SE, Florendo MN, Mason DE, LRRC 8 proteins form volume-regulated anion channels that sense ionic strength. Cell. 2016;164(3):499–511.26824658 10.1016/j.cell.2015.12.031PMC 4733249 · doi ↗ · pubmed ↗
- 7Lee CC, Freinkman E, Sabatini DM, Ploegh HL. The protein synthesis inhibitor blasticidin s enters mammalian cells via leucine-rich repeat-containing protein 8D. J Biol Chem. 2014;289(24):17124–31.24782309 10.1074/jbc.M 114.571257 PMC 4059153 · doi ↗ · pubmed ↗
- 8Planells-Cases R, Lutter D, Guyader C, Gerhards NM, Ullrich F, Elger DA, Subunit composition of VRAC channels determines substrate specificity and cellular resistance to Pt-based anti-cancer drugs. EMBO J. 2015;34(24):2993–3008.26530471 10.15252/embj.201592409 PMC 4687416 · doi ↗ · pubmed ↗