A Quantitative Assay for Ca2+ Uptake through Normal and Pathological Hemichannels

Connexin (Cx) hemichannels (HCs) are large pore hexameric structures that allow the exchange of ions, metabolites and a variety of other molecules between the cell cytoplasm and extracellular milieu. HC inhibitors are attracting growing interest as drug candidates because deregulated fluxes through HCs have been implicated in a plethora of genetic conditions and other diseases. HC activity has been mainly investigated by electrophysiological methods and/or using HC-permeable dye uptake measurements. Here, we present an all-optical assay based on fluorometric measurements of ionized calcium (Ca2+) uptake with a Ca2+-selective genetically encoded indicator (GCaMP6s) that permits the optical tracking of cytosolic Ca2+ concentration ([Ca2+]cyt) changes with high sensitivity. We exemplify use of the assay in stable pools of HaCaT cells overexpressing human Cx26, Cx46, or the pathological mutant Cx26G45E, under control of a tetracycline (Tet) responsive element (TRE) promoter (Tet-on). We demonstrate the usefulness of the assay for the characterization of new monoclonal antibodies (mAbs) targeting the extracellular domain of the HCs. Although we developed the assay on a spinning disk confocal fluorescence microscope, the same methodology can be extended seamlessly to high-throughput high-content platforms to screen other kinds of inhibitors and/or to probe HCs expressed in primary cells and microtissues.

Deregulated Ca 2+ flux through HCs has been implicated in several diseases linked to Cx mutations [18,[38][39][40]. In addition, the opening of wild type (wt) Cx HCs is associated with an impressive list of pathological conditions that affect a major proportion of the world population, including ischemia/stroke, Alzheimer's disease, epilepsy, liver fibrosis and cirrhosis, nonalcoholic steatohepatitis, inflammation and others [41][42][43][44][45][46]. Of note, inhibition of Cx HCs in animal models in vivo exerted anticonvulsant effects [47], lowered the degree of liver fibrosis induced by administration of thioacetamide [48] and improved recovery from spinal cord injury [49].
Cx HC activity has been probed indirectly by the release of ATP and other metabolites as mentioned above [21][22][23][24][25][26]. Other classical methods include electrophysiology (patch clamp) and dye uptake or leakage assays. The latter are based on passive diffusion of fluorescent molecules through open HCs in the cell plasma membrane [50]. Experiments of this type have been mainly conducted in model cell systems but occasionally also in native tissues [24,25] or in vivo [26]. The aim of this work was to establish a complementary, all-optical method to investigate Cx HC activity based on the uptake of Ca 2+ .
We focused on Cx26, which is expressed at extremely low, oftentimes undetectable levels in epidermal keratinocytes, yet, paradoxically, mutant Cx26 leaky/hyperactive HCs cause genodermatoses, including keratitis-ichthyosis-deafness syndrome (KID) and palmoplantar keratoderma with deafness (PPKDFN) that may drastically compromise quality of life and life-long health [40,51,52] (Table 1). To broaden the impact of this study and demonstrate the wider applicability of the methodology, we also investigated HCs composed of Cx46 [11,12], which is an important target for cancer therapy [53][54][55][56][57] and, when mutated, is involved in cataract formation [58].
The Ca 2+ uptake assay we developed provides a highly sensitive readout of Cx HC activity, as detailed hereafter. Furthermore, the assay is also suitable to evaluate the efficacy of different HC inhibitors that are attracting a growing attention as potential therapeutics [41][42][43][44][45][46][47][48], including antibodies targeting the extracellular domain of HCs [49,[59][60][61][62]. Table 1. Rare genodermatoses caused by leaky/hyperactive HCs. (P = prevalence; C = number of cases described in the literature; and E = estimated number of people with this disease).

Type of Disease (Acronym/OMIM § No.)
P C E Cx (Gene)

Generation of Stable Cell Pools Overexpressing the Cx of Interest and GCaMP6s
HaCaT cells are spontaneously immortalized line of aneuploid human keratinocytes [63] widely adopted to model the epidermis in vitro [64,65]. We and others previously overexpressed Cx HCs in HaCaT cells [59,61,66]. Here, given that uncontrolled expression of leaky HCs causes toxicity and rapidly leads to cell death (see, e.g., [66][67][68][69][70][71][72][73]), we designed and produced lentiviruses (LVs) for Tet-on bicistronic expression of the Cx of interest and the genetically encoded indicator GCaMP6s that permits optical tracking of [Ca 2+ ] cyt changes with high sensitivity [74] (see Materials and Methods).
We inserted the Cx coding sequence (CDS) upstream of an internal ribosome entry site (IRES) sequence, in turn, followed by the CDS of GCaMP6s. Next, by puromycin selection of virally transduced cells, we obtained stable cell pools expressing human Cx26 (referred to as HaCaT-Cx26-GCaMP6s cells) or the lethal KID-related mutant Cx26G45E that generates leaky HCs [68,70,[75][76][77] (referred to as HaCaT-Cx26G45E-GCaMP6s cells). The expression levels of both IRES and GCaMP6s mRNAs increased over 350-fold within 24 h of doxycycline (dox) addition (2 µg/mL) to the culture medium ( Figure 1). The non-zero levels of both mRNAs in the absence of dox can be attributed to the minimal basal leakiness of the TRE promoter in the lentiviral construct [78].
Analyses of HaCaT-Cx26-GCaMP6s cells by live microscopy showed a monotonic growth of GCaMP6s fluorescence levels with time, reaching a steady state after~30 h from the onset of dox exposure ( Figure 1C). In HaCaT-Cx26-GCaMP6s cells pre-exposed to dox for 24 h, we confirmed overexpression of Cx26 by qPCR ( Figure 2A) as well as by immunofluorescence staining with a commercial antibody that recognizes an intracellular epitope ( Figure 2B). Note that Cx26 was detected at very low levels also in the absence of dox. Immunoreactivity against Cx26 produced fluorescent puncta at points of contacts between adjacent cells, corresponding to IGJCs ( Figure 2C).
To confirm that functional HCs were present in the plasma membrane of live HaCaT-Cx26-GCaMP6s cells, we performed a standard dye uptake assay ( Figure 3) using 4 ,6diamidino-2-phenylindole (DAPI, 5 µM) that is non fluorescent in the medium but becomes fluorescent upon penetrating into the cell and binding to nucleic acids [50]. Validation of Tet-on bicistronic lentiviral constructs used to generate stable cell pools. mRNA expression levels of IRES (A) or GCaMP6s (B) assayed by qPCR (mean values ± standard deviation) in HaCaT-Cx26-GCaMP6s and HaCaT-Cx26G45E-GCaMP6s cells exposed to doxycycline for 24 h (+dox), or not exposed to doxycycline (−dox). p-values (p) were calculated using the two-tailed t-test. (C) GCaMP6s fluorescence levels in live HaCaT-Cx26-GCaMP6s cells vs. time after addition of dox (2 µg/mL, filled symbols) or vehicle (empty symbols) to the culture medium.
The mean values ± standard error of the mean (s.e.m.) for Δ = − 0 were computed for each condition as average fluorescence signal ( ) of pixels that exceeded an arbitrary threshold minus the average value obtained before dox administration ( 0 , t = 0). For each condition, values of n = 4-7 fields of view from one cell culture were averaged, and the statistical significance was calculated using a two-tailed t-test; *, P < 0.05; **, P < 0.01; ***, P < 0.001; and n.s., not significant. qPCR assay of Cx26 mRNA expression levels (mean values ± standard deviation) in HaCaT-Cx26-GCaMP6s cells not exposed (−dox) or exposed (+dox) to dox (2 µ g/mL, 24 h). p-values (P) were calculated using a two-tailed t-test.  Validation of Tet-on bicistronic lentiviral constructs used to generate stable cell pools. mRNA expression levels of IRES (A) or GCaMP6s (B) assayed by qPCR (mean values ± standard deviation) in HaCaT-Cx26-GCaMP6s and HaCaT-Cx26G45E-GCaMP6s cells exposed to doxycycline for 24 h (+dox), or not exposed to doxycycline (−dox). p-values (p) were calculated using the twotailed t-test. (C) GCaMP6s fluorescence levels in live HaCaT-Cx26-GCaMP6s cells vs. time after addition of dox (2 µg/mL, filled symbols) or vehicle (empty symbols) to the culture medium. The mean values ± standard error of the mean (s.e.m.) for ∆F = F t − F 0 were computed for each condition as average fluorescence signal (F t ) of pixels that exceeded an arbitrary threshold minus the average value obtained before dox administration (F 0 , t = 0). For each condition, values of n = 4-7 fields of view from one cell culture were averaged, and the statistical significance was calculated using a two-tailed t-test; *, P < 0.05; **, P < 0.01; ***, P < 0.001; and n.s., not significant. Figure 1. Validation of Tet-on bicistronic lentiviral constructs used to generate stable cell pools. mRNA expression levels of IRES (A) or GCaMP6s (B) assayed by qPCR (mean values ± standard deviation) in HaCaT-Cx26-GCaMP6s and HaCaT-Cx26G45E-GCaMP6s cells exposed to doxycycline for 24 h (+dox), or not exposed to doxycycline (−dox). p-values (p) were calculated using the two-tailed t-test. (C) GCaMP6s fluorescence levels in live HaCaT-Cx26-GCaMP6s cells vs. time after addition of dox (2 µg/mL, filled symbols) or vehicle (empty symbols) to the culture medium.
Cells pre-exposed to dox for 24 h showed a significant increase of DAPI fluorescence emission when bathed in an extracellular medium (ECM) containing a low [Ca 2+ ] ex (60 µM).
In the experiments of Figure 3, DAPI uptake was also inhibited by flufenamic acid (FFA, 100 µM), a well-known and widely used aspecific blocker of HCs [77] ( Figure 3B). To corroborate these findings, we used abEC1.1m, a previously characterized antagonist mAb, formatted as a single chain fragment variable (scFv)-fragment constant (Fc) diabody [59] with human scFv and mouse Fc domain, which specifically binds to the extracellular domain of Cx26, Cx30 and Cx32 HCs, however, is not effective on Cx43 HCs [60]. The effect of abEC1.1m (1 µM) was like that of FFA ( Figure 3B). The simplest explanation for these results is that DAPI uptake occurred primarily through Cx26 HCs, whose contribution prevailed over that of HCs formed by other residual Cx isoforms expressed in dox-induced HaCaT-Cx26-GCaMP6s cells.

Development and Validation of the Ca 2+ Uptake Assay
Based on the outcome of the experiments described above, we deemed it worthwhile to conduct Ca 2+ uptake experiments in confluent cultures of HaCaT-Cx26-GCaMP6s cells using a custom-built spinning disk confocal fluoresce microscope (described in [86]) equipped with a 20× water-immersion objective. The wide field of view (FOV) achieved with this objective (630 µm ∅) allowed us to collect and average images of GCaMP6s fluorescence emission from more than 100 cells per culture. Experiments were carried out in accordance with the following procedure ( Figure S1).
First, we pre-incubated the cells under the microscope in ECM for about 15 min to allow [Ca 2+ ] cyt to equilibrate with the low [Ca 2+ ] ex (60 µM) of the medium. Next, we excited GCaMP6s using a 488 nm laser and started to collect fluorescence images at 1 frame/s for 20 s of baseline (pre-stimulus) recording. Then, we added a bolus of CaCl 2 (2 µL) to the medium in the recording chamber (1 mL volume) at different concentrations in different experiments, aiming to achieve final estimated [Ca 2+ ] ex of 100 µM, 260 µM, 560 µM, 1 mM and 2 mM and continued to acquire images at 1 frame/s for a total of 10 min. By the end of each recording interval, we applied a bolus of the Ca 2+ ionophore ionomycin (200 µM, 2 µL and 0.4 µM estimated final concentration) to confirm that GCaMP6s signals were not saturated ( Figure S1). Thereafter, the experiment was terminated, and the culture was discarded.
For each experiment, GCaMP6s fluorescence emission was quantified pixel-by-pixel and frame-by-frame as ∆F = F − F 0 , where F represents emission at time t, and F 0 represents emission averaged over all baseline (pre-stimulus) frames [87]. We then spatially averaged ∆F pixel signals over all cells in the FOV and produced a single ∆F trace for each experiment. In the first 20 s after CaCl 2 addition, HaCaT-Cx26-GCaMP6s pre-exposed to dox for 24 h responded with a transient peak of [Ca 2+ ] cyt followed by a delayed and broader secondary maximum the height of which depended on CaCl 2 bolus concentration (see Discussion) ( Figure 4A). Ionomycin generated a third response peak that largely exceeded the two previous ones ( Figure S1). Cells not exposed to dox showed minimal responses due to the basal leakiness of the TRE promoter ( Figure S1). For statistical analyses, we normalized each ∆F trace for the peak response to ionomycin, ∆F max ( Figure S1). We computed average ∆F/∆F max traces from n independent experiments (see figure legends) in each condition, using a different culture for each experiment and displayed them as the mean (solid line) ± standard error of the mean (s.e.m., dashed lines, Figure 4A).
Finally, we quantified the [Ca 2+ ] cyt load in terms of the area subtended by the ∆F/∆F max traces. The data analysis showed that the sensitivity of the assay plateaued between 1 mM and 2 mM final [Ca 2+ ] ex ( Figure 4B), therefore, we adopted 2 mM as the standard final [Ca 2+ ] ex for all the other experiments. Reducing the initial [Ca 2+ ] ex from 60 µM to 20 µM abolished the delayed and broader secondary maximum, resulting in a dramatically reduced assay sensitivity ( Figure S2).  (1 µM, orange, n = 7) and ECM plus abEC1.1h-IgG (1 µM, red, n = 3). (D) Quantification of cytosolic Ca 2+ load for the results in (C) computed as described in (B). (E) Dose-dependent effect of the abEC1.1h-IgG for concentrations in the range 0.001 nM-10 µM (data are quoted as the mean ± s.e.m. for each condition). (F) Immunofluorescence labeling of HaCaT-Cx26-GCaMP6s cells with abEC1.1h-IgG; the antibody (15 nM) was applied to live cells that were then fixed but not permeabilized and counterstained with a red-fluorescent secondary antibody selective for the human Fc domain of the IgG. The green signal is due to GCaMP6s fluorescence. Scale bar: 20 µm.
We next tested the effects of FFA and abEC1.1m on Ca 2+ uptake quantified as explained above. Based on these analyses, abEC1.1m (1 µM, in ECM) exerted an inhibitory effect comparable to that of FFA (100 µM, in ECM; P = 1, ANOVA; Figures 4C and S3), suggesting that Cx26 HCs provided the dominant uptake route for Ca 2+ . In this set of experiments, we tested also a new fully human immunoglobulin G (IgG) format of this mAb, henceforth referred as abEC1.1h-IgG (1 µM, in ECM). Although abEC1.1h-IgG modified the kinetics of Ca 2+ uptake differently from abEC1.1m ( Figure 4C), the overall blockade effect exerted by the two mAbs was undistinguishable (P = 1, ANOVA; Figure 4D).
To better characterize the new IgG, we constructed a dose-inhibition curve ( Figures 4E and S4) and determined that the half-maximal effective concentration (EC 50 ) was comprised between 1 and 10 nM. At saturating concentrations (>1 µM), the inhibitory effect of abEC1.1h-IgG was about 80%, comparable to that the ScFv-Fc format previously estimated by the patch-clamp technique [59,61]. Immunolabeling with abEC1.1h-IgG produced fluorescent puncta at cell boundaries but not at points of contact between adjacent cells ( Figure 4F; compare to Figure 2C) consistent with binding of abEC1.1h-IgG to unpaired HCs in the cell plasma membrane.

mAbs Targeting the Extracellular Domain of Cx26 Inhibit Ca 2+ Uptake through Leaky HCs Implicated in KID Syndrome
Having successfully established working conditions to conduct Ca 2+ uptake assays, we turned to cells overexpressing Cx26G45E. The percentage of pyknotic nuclei in HaCaT-Cx26G45E-GCaMP6s cells exposed to dox for 24 h ( Table 2) was not significantly different from that of cultures not exposed to dox (Table 3). Therefore, we used cells pre-exposed to dox for 24 h to compare the effect of ECM alone or ECM supplemented with FFA (100 µM), abEC1.1m (1 µM) or abEC1.1h-IgG (1 µM). All three inhibitors significantly reduced Ca 2+ uptake ( Figure 5), suggesting that Ca 2+ uptake was mediated by Cx26G45E HCs rather than by passive leakage through a compromised plasma membrane. Based on patch-clamp experiments conducted on Cx26G45E mutant HCs [59] and in vivo experiments on a mouse model of a related genodermatosis [61], these results suggest that the tested mAbs are candidate therapeutics for KID syndrome.

Effects of Uncharacterized mAbs, Targeting the Extracellular Domain of Cx46, on Ca 2+ and DAPI Uptake
For the final set of experiments, we generated a stable pool of HaCaT-Cx46-GCaMP6s cells and used them to test new uncharacterized mAbs targeting the extracellular domain of Cx46 ( Figure 6). The mean ± s.e.m areas subtended by normalized Ca 2+ uptake traces in the time interval between t = 20 s and t = 590 s. Statistical significance was calculated using the Bonferroni post hoc test for pairwise comparisons after the ANOVA test; *, P < 0.05; and n.s., not significant.

Effects of Uncharacterized mAbs, Targeting the Extracellular Domain of Cx46, on Ca 2+ and DAPI Uptake
For the final set of experiments, we generated a stable pool of HaCaT-Cx46-GCaMP6s cells and used them to test new uncharacterized mAbs targeting the extracellular domain of Cx46 ( Figure 6).

Effects of Uncharacterized mAbs, Targeting the Extracellular Domain of Cx46, on Ca 2+ and DAPI Uptake
For the final set of experiments, we generated a stable pool of HaCaT-Cx46-GCaMP6s cells and used them to test new uncharacterized mAbs targeting the extracellular domain of Cx46 ( Figure 6).  FFA (100 µM) strongly reduced Ca 2+ uptake in HaCaT-Cx46-GCaMP6s cells preexposed to dox for 24 h. In contrast, we observed a significant, clear-cut inhibitory effect of about 63% only with Ab17396 at 10 µM concentration. At two lower concentrations (0.1 and 1.0 µM), we could not estimate the [Ca 2+ ] cyt load from the area subtended by the ∆F/∆F max traces because, surprisingly, the traces dropped below the pre-stimulus levels after CaCl 2 bolus delivery (see Discussion).
For a complementary estimate of Ab17396 activity, we performed DAPI uptake experiments in ECM alone or supplemented with Ca 2+ (2 mM), FFA (100 µM) and Ab17396 (1.0 or 10 µM) (Figure 7). DAPI uptake was significantly inhibited by both Ca 2+ and FFA, whereas the reductions provided by Ab17396 were not statistically significant (P = 0.085 between ECM and 10 µM Ab17396 conditions). However, these results were negatively affected by the limited number of experimental replicates that we were able to collect due to the shortage of Ab17396.
Finally, we also performed DAPI uptake experiments with other anti-Cx46 mAbs from the pool that led to the selection of Ab17396 (Figure 8). Together, based on the results of this preliminary screening, two of them, Ab17396 and Ab17398, appear worthy of further analyses to determine their potency, selectivity and usefulness as Cx46 HC inhibitors.
For a complementary estimate of Ab17396 activity, we performed DAPI uptake experiments in ECM alone or supplemented with Ca 2+ (2 mM), FFA (100 µ M) and Ab17396 (1.0 or 10 µ M) (Figure 7). DAPI uptake was significantly inhibited by both Ca 2+ and FFA, whereas the reductions provided by Ab17396 were not statistically significant (P = 0.085 between ECM and 10 µ M Ab17396 conditions). However, these results were negatively affected by the limited number of experimental replicates that we were able to collect due to the shortage of Ab17396. Finally, we also performed DAPI uptake experiments with other anti-Cx46 mAbs from the pool that led to the selection of Ab17396 (Figure 8). Together, based on the results of this preliminary screening, two of them, Ab17396 and Ab17398, appear worthy of further analyses to determine their potency, selectivity and usefulness as Cx46 HC inhibitors.

Discussion
In this work, we developed a novel all-optical assay to probe Ca 2+ uptake through normal or pathological HCs, which are emerging drug targets for a variety of pathological conditions [41][42][43][44][45][46][47][48][49]61,62], using cytosolic GCaMP6s fluorescence intensity variations as a readout. This choice was dictated by different considerations. On the one hand, the study of Ca 2+ dynamics in cells expressing pathological HCs is of utmost importance, given that deregulated Ca 2+ influx through HCs has been implicated in different disorders, including but not limited to those affecting the skin, such as KID syndrome [40] and Clouston

Discussion
In this work, we developed a novel all-optical assay to probe Ca 2+ uptake through normal or pathological HCs, which are emerging drug targets for a variety of pathological conditions [41][42][43][44][45][46][47][48][49]61,62], using cytosolic GCaMP6s fluorescence intensity variations as a readout. This choice was dictated by different considerations. On the one hand, the study of Ca 2+ dynamics in cells expressing pathological HCs is of utmost importance, given that deregulated Ca 2+ influx through HCs has been implicated in different disorders, including but not limited to those affecting the skin, such as KID syndrome [40] and Clouston syndrome [61].
On the other hand, fluorometric assays based on intracellular Ca 2+ variations are industry standards widely used in drug screening for membrane channels and G-protein coupled receptors (GPCRs) [88,89]. As shown here, the assay provided a sensitive indication for the effect of HC inhibitors, considering that it allowed us to detect changes in the uptake of Ca 2+ with as little as 0.1 nM of the abEC1.1h-IgG mAb in HaCaT-Cx26-GCaMP6s cells.
In these experiments, GCaMP6s signals exhibited complex kinetics with an early peak in close temporal proximity to the CaCl 2 bolus application, followed by a delayed response that generated a second relative maximum with a delay of 2 to 3 min. The reason for this biphasic response is unclear. We speculate that the sudden increase of the [Ca 2+ ] ex had the contrasting effect of initiating Ca 2+ uptake, due to the increased driving force for the cation, while also promoting HC closure. This would account for the short-duration first peak.
As for the delayed response producing the second peak (which was detectable at the final [Ca 2+ ] ex > 500 µM, Figure 4A), it is worth considering that keratinocytes (and several other cell types) sense changes in the [Ca 2+ ] ex via the G-protein coupled calciumsensing receptor (CaSR) [90]. Activation of phospholipase C β isoforms downstream of CaSR-dependent G-protein stimulation results in the breakdown of phosphatidyl inositol-4,5-bisphosphate (PIP 2 ), leading to the formation of inositol 1,4,5-trisphosphate (IP 3 ) and diacylglycerol (DAG). DAG remains associated with the plasma membrane and activates protein kinase C, whereas IP 3 mediates Ca 2+ release from the endoplasmic reticulum (ER), by activating IP 3 receptors in a ligand-dependent manner [91]. Note that IP 3 receptors are co-regulated by Ca 2+ that promotes their opening at low [Ca 2+ ] cyt and inhibits them at higher [Ca 2+ ] cyt , depending on the IP 3 level [92,93]. Therefore, the amplitude of this delayed, IP 3 -mediated response would depend strongly also on Ca 2+ influx through HCs. In other words, the second peak that develops after CaCl 2 bolus addition may be seen as and IP 3 -dependent (nonlinear) amplification mechanism for the signal generated by Ca 2+ influx through HCs. This would explain its sensitivity to HC inhibitors in HaCaT-Cx26-GCaMP6s and HaCaT-Cx26G45E-GCaMP6s cells. The data in Figures 4A and S2 support this conclusion. However, dedicated experiments, beyond the scope of this article, are required to validate this hypothesis (e.g., siRNA-mediated knock down of the CaSR and/or pre-empting of the intracellular Ca 2+ stores with the SERCA pump inhibitor thapsigargin).
In the experiments with HaCaT-Cx46-GCaMP6s cells, Ca 2+ uptake traces dropped below pre-stimulus levels in the recovery phase in the presence of Ab17396 at 0.1 µM and 1.0 µM but not with Ab17396 at 10 µM or with FFA at 100 µM. In contrast, in HaCaT-Cx26-GCaMP6s and HaCaT-Cx26G45E-GCaMP6s cells, the traces never dropped below baseline. Thus, we suspect that Ab17396 may affect, in a concentration-dependent manner, also other molecular players involved in the homeostasis of cytosolic Ca 2+ . In particular, low concentrations Ab17396 might facilitate the export of Ca 2+ from the cytosol to the extracellular milieu, e.g., by potentiating the activity of plasma membrane Ca 2+ ATPases or sodium-calcium exchangers. At higher concentrations, the prevailing effect of Ab17396 appears to be inhibition of Cx46 HCs. Further experiments are required to address this issue.
A stronghold of the Ca 2+ uptake assay is the all-in-one Tet-on bicistronic system for the simultaneous, controlled expression of the Cx of interest and GCaMP6s in model cells. GCaMP6s provided not only an indirect fluorescent reporter for the successful expression of the Cx in HaCaT cells, but also a key element for the sensitive detection of [Ca 2+ ] cyt variations at the core of the method. Although we used LVs, other systems can be contemplated, including the piggyBac transposon Tet-on vector system [94]. Furthermore, the methodology can be readily extended to different Cx isoforms or pathological variants in other popular cell lines that express low levels of endogenous Cxs, e.g., HeLa cells [95]. At the opposite extreme, the assay could be used to probe native HCs in cells, organoids, native tissues (e.g., slices), or even in vivo. In all these cases, the requirements are (i) achieving adequate levels of the Ca 2+ indicator (e.g., by in vivo viral transduction or electroporation) and (ii) being able to rapidly modify the [Ca 2+ ] ex . Indeed, biasing (overexpressed or native) HCs towards the open state in a low [Ca 2+ ] ex medium prior to switching to a high [Ca 2+ ] ex medium to promote Ca 2+ uptake is key for these experiments.
The strengths of this approach for the study of HCs can be summarized as follows. The all-in-one Tet-on system allows for inducible bicistronic expression of both the Cx isoform/mutant of interest and the biosensor required to test HC functionality (GCaMP6s). After Cx-expressing stable cell pools are established, no further manipulations are required to run the assay, ensuring homogeneity across samples. The assay can be run on large cell populations to rapidly obtain statistically robust results. It is sensitive enough to construct dose-inhibition curves for drug candidates, including mAbs, without the need for electrophysiological recordings of HC-mediated currents, which may be challenging in the case of leaky HCs. Last but not least, the assay can be extended to high-throughput high-content platforms and/or primary cells and microtissues.
Whereas dye uptake assays provide a useful complement, as also demonstrated here, they may lack the required sensitivity to discriminate among different experimental conditions (e.g., concentrations and/or inhibitors), chiefly due to charge/size of permeating fluorescent molecule in relation to HC pore size and electrostatic profile [96,97]. On the other hand, the time course of the fluorescence changes caused by dye accumulation is typically monotonic and thus comparatively easier to analyze and interpret.

Establishment of Stable Cell Pools
Twenty-four hours post infection, the LV-containing medium was replaced with fresh medium. Forty-eight hours post infection, 1 µg/mL of puromycin (Cat. No. P9620, Merck KGaA) was added to the medium and kept until a selected homogeneous cell pool was established (typically 3 days). The cell pools thus obtained were denoted as HaCaT-Cx( . . . )-GCaMP6s. For precise gene regulation of Tet-on pools, we replaced the ordinary FBS (see above) with Tet-approved FBS (Cat. No. A4736401, TFS). All cell pools were routinely tested for mycoplasma contamination by staining with DAPI (Cat. No. D1306, TFS) followed by visual inspection at the fluorescence microscope.
For imaging experiments or immunofluorescence analyses, 5 × 10 5 cells were plated onto round glass coverslips (12 mm ∅) in a 35 mm ∅ dish two days before the assay. The next day and at least 24 h before experiments, dox (2 µg/mL, Cat. No. D1822, Merck KGaA) was added to the culture medium to promote Cx and GCaMP6s expression. To evaluate time-dependent dox-induced GCaMP6s expression in HaCaT-Cx26-GCaMP6s cells ( Figure 1C), samples were washed and imaged with a spinning disk fluorescence microscope adapted for live-cell imaging [86] at different time points after dox administration (3, 6, 10, 24, 30 and 48 h).
To estimate the percentage of pyknotic cells in HaCaT-Cx26-GCaMP6s and HaCaT-Cx26G45E-GCaMP6s pools (Tables 2 and 3), cell-plated coverslips exposed to dox or vehicle for 24 h were fixed with 4% paraformaldehyde (PFA; Cat. No. 158127, Merck KGaA) for 15 min and permeabilized with 0.1% Triton X-100 (Cat. No. 1.08603, Merck KGaA) for 5 min. Next, samples were washed with PBS and nuclei were stained with DAPI (1 µM, 5 min). After visualization at the fluorescence microscope, cells that exhibited chromatin condensation and DNA fragmentation were counted in six different FOVs (n = 88-219 cells/FOV) for each condition, and the results were divided by the total number of cells in each FOV.

qPCR
The total RNA was extracted from HaCaT-Cx26-GCaMP6s and HaCaT-Cx26G45E-GCaMP6s cells exposed or not exposed to dox for 24 h using TRIzol Reagent (Cat. No. 15596018, TFS). 10 µg of the extracted total RNA was digested with DNase I (Cat. No. 18068015, TFS) and then purified using Rneasy mini Kit (Cat. No. 74104, Qiagen, Hilden, Germany) according to the manufacturer's instructions. A NanoDrop spectrophotometer was used to quantify the RNA by measuring the absorbance and reverse transcription. (RT)-PCR was performed from 1µg of the purified RNA using high-capacity cDNA kit (Cat. No. 4368814, TFS), according to the manufacturer's instructions but with a slight modification in which Oligo(dt) (Cat. No. 18418012, TFS) was combined with the cDNA kit random primers at a ratio of 1:5.
Quantitative PCR (qPCR) was performed on the cDNA with specific primers (see below). Samples were analyzed in triplicate and gene expression levels were estimated by the 2 −∆∆CT method, using glyceraldehyde-3-phosphate dehydrogenase (GAPDH) as internal reference gene and Sybr green (

Immunofluorescence
Cells from stable pools grown on glass coverslips and exposed to dox or vehicle for 24 h were fixed with 4% PFA for 15 min, permeabilized with 0.1% Triton X-100 for 5 min and incubated for 1 h in blocking buffer containing 0. For immunofluorescence labeling with abEC1.1h-IgG ( Figure 4F), cells from stable pools, grown on glass coverslips and exposed to dox for 24 h, were incubated at 37 • C with ECM supplemented with 15 nM abEC1.1h-IgG for 1 h. Cells were then fixed with 4% PFA for 15 min and incubated for 1 h with a goat anti-human IgG Alexa Fluor 594 secondary antibody (1:800, Cat. No. 109-585-170, Jackson ImmunoReasearch, Ely, Cambridgeshire, UK). Samples were washed three times with PBS between all steps.
Fluorescence images (pixel size 96 nm) were acquired with a confocal microscope (TCS SP5 Leica, Wetzlar, Germany) equipped with a 63× oil immersion objective (HC PL Apo, UV optimized, NA 1.4, Leica). Alexa Fluor 488 and GCaMP6s fluorescence was excited by a 488-nm Argon laser and collected between 500 nm and 540 nm. Alexa Fluor 555 and Alexa Fluor 594 fluorescence was excited by a 543-nm HeNe laser and collected between 568 nm and 680 nm and between 600 and 670 nm, respectively. For visualization purposes, a σ = 0.8 Gaussian filter was applied on the green channel of fluorescence images. For experiments with HaCaT-Cx26-GCaMP6s pools, coverslips with plated cells were washed with ECM and imaged using the two-photon microscope of Ref. [79] before and after 5 min incubation with one of the mentioned DAPI-containing solutions at room temperature. The microscope was equipped with a 25× water immersion multiphoton objective (XLPlanN25XWMP2, NA = 1.05, WD = 2 mm, Olympus Corporation, Tokyo, Japan). DAPI was excited at 780 nm and fluorescence emission signals were collected through a blue emission filter (460/50 nm, ET460/50m, Chroma Technology Corp, Bellow Falls, VT, USA). For each sample, we acquired multiple z-stacks (through-focus image sequences obtained with a 2 µm step size) of at least three different FOVs. FFA and abEC1.1m were pre-incubated at 37 • C for 30 min.
For experiments with HaCaT-Cx46-GCaMP6s pools, the day before the experiment cells were seeded into a black 96-well microplate for high content imaging (2 × 10 4 cells/well, #4580, Corning Inc., Corning, NY, USA) and dox (2 µg/mL) was added to the culture medium. The glass bottom of each well was pre-treated with laminin (2 µg/cm 2 dissolved in water, 30 min at 37 • C; Cat. No. 11243217001, Merck KGaA).
On the day of the experiment, cells were washed in ECM supplemented by 2 mM CaCl 2 to remove FBS and mean GCaMP6s fluorescence value was measured at the plate reader (Varioskan LUX, Cat. No. N16045, TFS) for future normalization of DAPI uptake results to the number of cells present in each well. Then, wells were pre-incubated with one of the mentioned media for 20 min at 37 • C without DAPI (n = 3-6 wells/condition) to allow the blockers to act on HCs. Next, DAPI-deprived media were replaced with analogous DAPI-containing solutions and the microplate was moved to the plate reader to evaluate DAPI uptake rate.
The mean DAPI fluorescence value for each well was measured every 10-15 min starting from 5 min after the onset of DAPI incubation (t = 0 in the graphs) and proceeded for 2 h. We ruled out any possible contribution of dead cells in the cultures (for which DAPI uptake is much faster than HC-dependent uptake in live cell) because for each well we treated the mean value of the first DAPI fluorescence reading (5 min after the onset of dye incubation) as a baseline and subtracted it from all the following measurements.

Ca 2+ Uptake
On the day of the experiment, cell-plated coverslips with confluent cell population were transferred to the stage of the spinning disk microscope described in Ref. [86] and equipped with water immersion objective (20×, N.A. = 0.95, XLUMPlan Fl, Olympus Corporation). Cells were incubated at room temperature for 15 min in ECM, thereafter fluorescence images were acquired at 1 frame/s for 12 min. GCaMP6s fluorescence was excited by a 488 nm diode laser (Cat. No. COMPACT-150G-488-SM, World Star Tech, Markham, Ontario, Canada) and emission signals were collected though a 535/30 nm band-pass filter (Cat. No. ET535/30M, Chroma Technology Corp., Bellows Falls, VT, USA).
After 20 s of baseline recordings, a 2 µL bolus of 1 M CaCl 2 solution was delivered through a manually operated pipette held at the edge of the FOV, and GCaMP6s fluorescence variations were monitored for 10 min. Finally, a 2 µL bolus of ionomycin solution (200 µM, Cat. No. I3909, Merck KGaA) was delivered to normalize the Ca 2+ influx response to the maximum achievable fluorescence variation (i.e., to the level of biosensor expression throughout experiments or experimental sessions). We used each cell-plated coverslip only once. To assay the efficacy of HC inhibitors, coverslips were pre-incubated at 37 • C with ECM supplemented with one of the following HC compounds: 100 µM FFA (40 min), 1 µM abEC1.1m (30 min); abEC1.1h-IgG at a concentration varying from 0.001 nM to 10 µM for the construction of dose-inhibition curve (30 min).

Selection of Anti-Cx46 mAbs
A synthetic, phage-displayed antibody library [99] was selected for binding to Cx46 protein expressed on the surface of Virus-Like Particles (Cx46-VLPs). Briefly, Expi293 cells (Cat. No. A14527, TFS) were co-transfected with plasmids expressing either HA-tagged Cx46 together with HIV Gag (Cx46 VLPs) or HIV Gag alone (empty VLPs), and VLPs were isolated from supernatant 2 days post transfection. Cx46 VLPs (100 µL at 10 µg/mL) were immobilized on Maxisorp Immuno plates (Cat. No. 12-565-135, TFS) and used for positive binding selections with library phage pools that were first exposed to similarly immobilized empty VLPs to deplete non-specific binders. After four rounds of binding