Triclosan Disrupts Immune Cell Function by Depressing Ca²⁺ Influx Following Acidification of the Cytoplasm

Abstract

Triclosan (TCS) is an antimicrobial agent that was effectively banned by the FDA from hand soaps in 2016, hospital soaps in 2017, and hand sanitizers in 2019; however, TCS can still be found in a few products. At consumer-relevant, non-cytotoxic doses, TCS inhibits the functions of both mitochondria and mast cells, a ubiquitous cell type. Via the store-operated Ca²⁺ entry mechanism utilized by many immune cells, mast cells undergo antigen-stimulated Ca²⁺ influx into the cytosol, for proper function. Previous work showed that TCS inhibits Ca²⁺ dynamics in mast cells, and here we show that TCS also inhibits Ca²⁺ mobilization in human Jurkat T cells. However, the biochemical mechanism behind the Ca²⁺ dampening has yet to be elucidated. Three-dimensional super-resolution microscopy reveals that TCS induces mitochondrial swelling, in line with and extending the previous finding of TCS inhibition of mitochondrial membrane potential via its proton ionophoric activity. Inhibition of plasma membrane potential (PMP) by the canonical depolarizer gramicidin can inhibit mast cell function. However, use of the genetically encoded voltage indicators (GEVIs) ArcLight (pH-sensitive) and ASAP2 (pH-insensitive), indicates that TCS does not disrupt PMP. In conjunction with data from a plasma membrane-localized, pH-sensitive reporter, these results indicate that TCS, instead, induces cytosolic acidification in mast cells and T cells. Acidification of the cytosol likely inhibits Ca²⁺ influx by uncoupling the STIM1/ORAI1 interaction that is required for opening of plasma membrane Ca²⁺ channels. These results provide a mechanistic explanation of TCS disruption of Ca²⁺ influx and, thus, of immune cell function.

Introduction

Triclosan (TCS) is a formerly widespread antimicrobial agent: it was estimated that 75% of the US population in 2008 was exposed to TCS (Calafat et al., 2008). Recently, the U.S. Food and Drug Administration effectively banned TCS from hand soaps in 2016 (Kux, 2016), from hospital products in 2017 (Kux, 2017), and from over-the-counter hand sanitizers in 2019 (Gottlieb, 2019). TCS was also removed by the Colgate-Palmolive company from its top-selling toothpaste product (Kary, 2019) in 2019. Despite these stoppages, TCS remains in some antibacterial household products that are not regulated by the FDA and in a few remaining personal care products (www.ewg.org). TCS is readily absorbed into the skin (Queckenberg et al., 2010) and oral mucosa (Gilbert, 1987) where it remains for a significant time prior to metabolism and clearance (Moss et al., 2000), thus allowing for a constant chronic exposure upon periodic re-exposure. In this study, non-cytotoxic, micromolar TCS doses, that model human exposure levels to TCS following personal care product application (Weatherly and Gosse, 2017; Weatherly et al., 2018), are utilized. Added to consumer products for its antimicrobial properties, TCS, ironically, inhibits the functioning of mammalian immune cells whose physiological purpose is to fight microbial infections (Udoji et al., 2010; Palmer et al., 2012; Hurd-Brown et al., 2013).

Clinically, TCS is used for its antimicrobial (Daoud et al., 2014) and anti-gingivitis properties (Rover and Leu-Wai-See, 2014). Additionally, previous studies provided evidence that TCS could potentially be used to treat atopic dermatitis (Sporik and Kemp, 1997; Tan et al., 2010). Despite these positive clinical effects, within the past few years a panoply of triclosan epidemiology studies have emerged, reporting various adverse TCS health effects (Weatherly and Gosse, 2017). Adverse TCS-linked effects on the reproductive system include increased spontaneous abortion rate (Wang et al., 2015), abnormal sperm morphology (Jurewicz et al., 2018; Zamkowska et al., 2018), and decreased fecundity (Vélez et al., 2015; Zhu et al., 2019). Additionally, newborn infants’ weight, length, and head circumference decreased due to TCS exposure (Etzel et al., 2017). Cognitive effects have also been observed; TCS has been linked to lower cognitive test scores in children (Jackson-Browne et al., 2018) and to higher behavior problem scores in 8-year old boys (Jackson-Browne et al., 2019). TCS causes metabolic effects including decreased BMI (Li et al., 2015), increased risk of gestational diabetes (Ouyang et al., 2018), changes in thyroid hormone levels in blood (Koeppe et al., 2013; Wang et al., 2017) and increased risk of type 2 diabetes in women (Xie et al., 2020). Triclosan exposure is also associated with decreased bone mass density and increased osteoporosis (Cai et al., 2019). Additionally, allergic rhinitis has been associated with TCS exposure (Kim and Kim, 2019). However, the cellular and molecular mechanisms underlying these epidemiological observations are not fully understood. Triclosan inhibition of mast cells (Palmer et al., 2012; Weatherly et al., 2013) and mitochondria (Ajao et al., 2015; Shim et al., 2016; Weatherly et al., 2016; Weatherly et al., 2018; Weatherly et al., 2020), in human cells (Weatherly et al., 2016) and in other species, may be related to these human health effects.

One of the underlying mechanisms of adverse human health effects caused by TCS may be its mitochondrial toxicity. An uncoupler due to its ionizable proton (Ajao et al., 2015; Weatherly et al., 2016), triclosan inhibits adenosine triphosphate (ATP) production and increases oxygen consumption rate in multiple cell types (Weatherly et al., 2016) and in live zebrafish (Shim et al., 2016). Additionally, TCS deflates mitochondrial membrane potential (MMP), thwarts mitochondrial translocation and Ca²⁺ dynamics, and induces mitochondrial fission and deformation (as assessed in two-dimensional [2D], super-resolution, live-cell microscopy (Weatherly et al., 2018). These TCS effects on MMP and mitochondrial morphology have been recapitulated in an in vivo mouse study (Weatherly et al., 2020). Critically, mitochondrial toxicity by TCS has been directly connected to its instigation of inflammation and immunotoxicity: TCS increases Drp1 and decreases Opa1 expression levels, effects which both increase mitochondrial fission and activate the NLRP3 inflammasome (Weatherly et al., 2020).

Separate from its mitochondrial toxicity, TCS disrupts other cellular signal transduction processes, such as Ca²⁺ mobilization and cytoskeletal remodeling, which are shared by numerous cell types (Weatherly et al., 2018). For example, TCS disrupts the functioning of the immunological/neurological cell type mast cells. TCS mitotoxicity explains an estimated ~10% this inhibition (Weatherly et al., 2018); the complete biochemical mechanism underlying TCS inhibition of mast cell function remains unknown and is explored further in this manuscript. Mast cells are found in a large variety of tissues (Dvorak, 1986; Theoharides and Sant, 1991; Farrell et al., 1995; Blank et al., 2007) and are involved in defense against parasites (Metcalfe et al., 1997), bacteria (Johnzon et al., 2016), and cancer (Hempel et al., 2017). Mast cell dysfunction is associated with neurological diseases (Silver and Curley, 2013; Girolamo et al., 2017) such as multiple sclerosis (Elieh-Ali-Komi and Cao, 2017). Of course, mast cells are also major effector cells of allergy and asthma (Galli et al., 2005), and previous clinical reports of atopic disorder alleviation by TCS, noted above (Sporik and Kemp, 1997; Tan et al., 2010), align with our findings of mast cell inhibition by TCS (Palmer et al., 2012).

Mast cells undergo degranulation, the release of bioactive mediators such as histamine and serotonin, in a process that begins with multivalent antigen (Ag) mediated cross-linking of immunoglobulin E (IgE)-primed FcεRI receptors on the cell surface. This binding initiates a phosphorylation cascade in which phospholipase C gamma (PLCγ) is activated (Kinet, 1999). PLCγ subsequently cleaves phosphatidylinositol 4,5-bisphosphate (PIP₂) into diacylglycerol (DAG) and phosphatidylinositol 4,5-bisphosphate (IP₃). The latter is then able to bind to its receptor on the endoplasmic reticulum (ER) membrane and initiate the release of ER Ca²⁺ stores into the cytosol (Berridge, 1993). Reduced ER Ca²⁺ concentration results in altered conformation of ER protein STIM1, causing it to bind to and change the conformation of the ORAI1 subunit of the Ca²⁺ release-activated Ca²⁺ (CRAC) channel, opening it to a flood of external Ca²⁺ flowing into the cytosol (Vig et al., 2006; Hogan et al., 2010); this process is termed store-operated Ca²⁺ entry, or SOCE (Putney, 1986). SOCE is required for mast cell degranulation (as reviewed in (Holowka et al., 2012) because increased cytosolic Ca²⁺ levels allow for events required for the movement of granules to the plasma membrane for exocytosis, such as protein kinase C (PKC) activation (Ozawa et al., 1993), phospholipase D (PLD) activation (Chahdi et al., 2002), and microtubule polymerization (Guo et al., 1998; Smith et al., 2003).

These signaling processes are important in a variety of cell types, including in the immune system and brain. SOCE mechanisms are largely conserved across cell types (Prakriya and Lewis, 2015). For example, T-cell signaling (Marano et al., 1993; Trebak and Kinet, 2019) also begins with T-cell receptor (TCR) cross-linking, as with FcεRI. In T-cells (part of the adaptive immune system) in vivo, this crosslinking is primarily done by an antigen presenting cell (APCs; innate immune cells such as macrophage or a dendritic cells), presenting antigen on the major histocompatibility complex (Punt et al., 2019). The TCR/APC interaction can be emulated in vitro through the use of an anti-TCR antibody (Marano et al., 1993). Such cross-linking leads to PLCγ activation and subsequent generation of IP₃ and SOCE, via the same process outlined above for mast cells. The subsequent increase in cytosolic Ca²⁺ concentration allows for calmodulin/calcineurin binding and subsequent activation of the nuclear factor of activated T-cells (Punt et al., 2019; Trebak and Kinet, 2019), an important Ca²⁺-dependent transcription factor. Thus, signaling upstream of SOCE is very similar between mast cells and T-cells; whereas signaling downstream of SOCE is largely different between mast cells and T-cells. While TCS is known to interfere with bodily distribution of T cells (Anderson et al., 2016), cytokine production (Barros et al., 2010; Marshall et al., 2015; Anderson et al., 2016; Marshall et al., 2017), and expression of calcium-binding biomarkers of inflammation (Marshall et al., 2017), triclosan effects on intracellular signal transduction, functional outcomes, and SOCE within T cells have not yet been examined.

Triclosan disrupts mast cell function by strongly inhibiting Ag-stimulated Ca²⁺ influx into the cytosol via the CRAC channels (Weatherly et al., 2018). This depressed cytosolic Ca²⁺ concentration leads, as expected, to inhibition of downstream events, such as reduced PLD activity (Shim et al., 2019) and microtubule polymerization (Weatherly et al., 2018). Thus, inhibition of SOCE by TCS explains its inhibition of mast cell function. The question remains: how does TCS inhibit Ca²⁺ entry through CRAC channels?

One explanation could be TCS disruption of any of the upstream signaling events leading to SOCE: FcεRI crosslinking, any link in the phosphorylation cascade, PIP₂ signaling function, or IP₃ binding to its ER membrane receptor. However, TCS does not hinder Ca²⁺ efflux from the ER (Weatherly et al., 2018) and does not seriously interfere with PIP₂ (Shim et al., 2019)— evidence that TCS does not inhibit any of these events upstream of ER Ca²⁺ release.

Because TCS does not hamper the signaling events culminating in ER Ca²⁺ release, an alternative mechanism underlies TCS inhibition of CRAC channel Ca²⁺ entry. Chemical depolarization of plasma membrane potential (PMP) reduces Ca²⁺ entry into the cytosol of mast cells (Mohr and Fewtrell, 1987a) and also of T cells (Sarkadi et al., 1990). Integrity of PMP is important for SOCE due to its contribution to the free energy, ΔG_transport, available to import Ca²⁺ down its electrochemical gradient into the cytoplasm. In this study, we hypothesized that TCS, as a proton ionophore (Weatherly et al., 2016), inhibits Ca²⁺ influx by depolarizing PMP similarly to how TCS deflates membrane potential of mitochondria (Weatherly et al., 2018) and of artificial bilayers (Popova et al., 2018)—by acting as a “Trojan horse” to provide safe passage for protons through lipid membranes.

The driving force for Ca²⁺ influx through the CRAC channel is calculated as (Nelson and Cox, 2017)

$$\Delta G_{Transport}=RTln\left(\frac{C_2}{C_1}\right)+nZ\text{F}\Delta \Psi$$

Using physiological temperature of 310K, the gas and Faraday’s constants, the number of charges on the transported ion (n=2), the sign of the charge of the transported ion (Z=+), resting PMP ΔΨ of −82.5 mV (Lindau and Fernandez, 1986; Wischmeyer et al., 1995), C₁ = extracellular Ca²⁺ concentration in the buffer (1.8 mM), and C₂ = average cytosolic Ca²⁺ concentration in an Ag-stimulated mast cell (1 μM) (Millard et al., 1988; Chandra et al., 1994), the driving force (Gibbs free energy) for Ca²⁺ influx through open CRAC channels is the highly exergonic value −35 kJ/mol. If TCS were to completely depolarize the mast cell PMP, the term nZFΔΨ would collapse to zero, the driving force behind SOCE would be reduced ~45%, to −19 kJ/mol, and, hence, 45% less Ca²⁺ would flood into the cell. In fact, this PMP-knockout-predicted ~45% reduction in integrated Ca²⁺ influx was observed in mast cells following 20 μM TCS exposure (Weatherly et al., 2018).

Thus, in this study, the first aim was to measure TCS effects on PMP in mast cells. The mast cell model rat basophilic leukemia, clone 2H3 (RBL), which are functionally similar to mature human mast cells, rodent mucosal mast cells, and basophils (Fewtrell et al., 1979; Metzger et al., 1982; Seldin et al., 1985; Metcalfe et al., 1997; Abramson and Pecht, 2007; Lee et al., 2012), was used. RBL cells are effective for toxicological studies due to their ability to respond to exogenous agents in a similar fashion as primary bone-marrow derived mast cells (Zaitsu et al., 2007; Thrasher et al., 2013; Alsaleh et al., 2016).

Conventionally, PMP is measured through the use of patch clamping or of voltage-sensitive fluorescent organic dyes. However, traditional patch clamping is a low-throughout process. Also, TCS is a chemical quencher of many unprotected fluorescent chromophores (such as those found in organic voltage sensitive dyes) (Weatherly, 2017). Thus, genetically encoded voltage indicators (GEVIs), reporter protein constructs with a β-barrel protein structure that protects the internal fluorophore from TCS fluorescence interference (Weatherly, 2017; Weatherly et al., 2018) were used. Targeted to the plasma membrane, GEVIs communicate changes in PMP through changes in fluorescence intensity. To our knowledge, this study represents the first use of a GEVI in an immune cell model or in the field of toxicology.

An additional novel use of a biophysical technique in toxicology, three-dimensional (3D) Fluorescence Photoactivation Localization Microscopy (FPALM) super-resolution microscopy was also employed to detail TCS effects, via membrane depolarization, on mitochondrial morphology in mast cells (Parent and Hess, 2019). Conventional light microscopy, including widefield and confocal imaging, cannot resolve objects that are less than ~250 nm apart, due to the diffraction limit, but FPALM breaks this diffraction limit (Hess et al., 2006). These 3D studies augment, by providing information on mitochondrial volume and surface area, previous 2D work showing that TCS disrupts mitochondrial nanostructure, causing “donut” shapes or fragmentation (Weatherly et al., 2018).

In this study, we test the hypothesis that TCS inhibits mast cell function by depolarizing the plasma membrane. However, careful experimentation with two GEVI constructs, which function via disparate molecular machinery, indicate a wholly different mechanism of TCS action: cytoplasmic acidification. Furthermore, TCS replicates this mechanism of action and abrogation of function in another immune cell type, T cells.

Methods

Chemicals and Reagents

Triclosan (TCS; 99% purity; Sigma-Aldrich) was prepared in BT (Tyrodes buffer containing bovine serum albumin [BSA]) without use of organic solvent (Weatherly et al., 2013) and diluted to deliver non-cytotoxic concentrations (Palmer et al., 2012) checked by UV-Vis spectrophotometry (Weatherly et al., 2013). Gramicidin (Sigma- Aldrich) was dissolved in 100% DMSO (Sigma-Aldrich) and diluted with BT.

Cell Culture

RBL-2H3 mast cells were cultured as in (Hutchinson et al., 2011).

Human Jurkat T cells, clone E6–1, were obtained from ATCC and maintained in suspension in phenol red-containing RPMI-1640 medium (ATCC) supplemented with 10% fetal bovine serum (Atlanta Biologics) and 100 U/ml penicillin-100 μg/ml streptomycin (Sigma-Aldrich). Cells were passaged once a week with thorough trituration technique to break up clumps, seeded at 35,000 cells/mL (for 6–7 days until harvest) to 125,000 cells/mL (for 3 days), and grown at 37°C and 5% CO_2. Supplement fresh media was added after 3 days in culture. These culturing conditions resulted in maximal cell density of 1–2 million cells/mL.

Fluorescence Photoactivation Localization Microscopy (FPALM) Imaging and Processing

Three-dimensional FPALM mitochondrial imaging was performed as in (Parent and Hess, 2019). RBL cells were transfected with an expression vector for Dendra2-Tom20 (Weatherly et al., 2018) using an Amaxa transfection kit (Lonza), then plated in μ-Slide 8-well plates with polymer coverslip (ibidi) at 100,000 cells/well in 200 F06DL/well phenol red-free RBL media. The next day, cells were exposed to 20 μM TCS or BT for 1 hour and fixed with 4% paraformaldehyde (Sigma Aldrich) before imaging. Imaging was performed using a 558 nm laser (Crystalaser) for Dendra2-Tom20 excitation, and fluorescence was captured using an Olympus IX-71 microscope with 60X 1.45NA oil lens, 2X telescope, and an EMCCD camera (Andor iXon DU-897 #BV). Custom MATLAB analysis software was used to obtain localized data points (Hess et al., 2006; Gudheti et al., 2013; Curthoys et al., 2019); details of microscopy are in Supplemental Methods.

FPALM Mitochondrial Analysis

In MATLAB, raw data was localized first by fitting each Point Spread Function (PSF) to a two-dimensional (2D) Gaussian, which was then drift corrected. After drift correction, following the methods in (Huang et al., 2008; Parent and Hess, 2019), the z-coordinate of each localized point was obtained from the measured calibration curve connecting the x- and y- widths of the PSF as a function of z- position. After obtaining the z-position for each localized point, the data set was processed through another custom MATLAB script for cluster identification, which compares the distances between each localization and all nearby localizations: localizations that lie within d_max (75 nm) of each other are considered to be in the same cluster. Clusters with a minimum number of 50 localizations were analyzed further. The nearest-neighbor single linkage cluster analysis (SLCA) method (Sneath, 1957; Gudheti et al., 2013) used here has been extended to three dimensions (Parent and Hess, 2019). Each cluster is then analyzed using the MATLAB convex hull and alpha shape functions, in order to quantify the local mean curvature for all localizations within an individual mitochondrion. Histograms and averages of the curvature, area, and volume of the convex and alpha hulls are then determined.

Degranulation Assay

Degranulation assays were performed as in (Weatherly et al., 2013), adapted for use with gramicidin in 0.003% DMSO vehicle. This fluorescence-based assay, performed in the 96-well format in a microplate reader, detects substrate cleavage by granule marker enzyme β-hexosaminidase in the supernatant from degranulating mast cells.

DiSC₃(5) Fluorescence Interference Assay

Triclosan effects on DiSC₃(5) (TCI America) organic voltage sensitive dye fluorescence were assayed. The dye was dissolved in 100% DMSO. In a 96 well black bottom plate (Greiner), DisC₃5 dye and TCS were mixed to achieve a final dye concentration of 1.02 μM in 0.01% final DMSO percentage, with various TCS concentrations (1–15 μM) in BT. The mixture was allowed to equilibrate for 10 min at 37°C. Fluorescence measurements were taken using a microplate reader (Synergy 2, Biotek) with 530 ± 20 nm excitation and 645 ± 15 nm emission.

Plasma Membrane Potential and Cytosolic pH Assays in RBL-2H3 Mast Cells

In order to measure changes in plasma membrane potential, RBL cells were transfected with genetically-encoded voltage indicator (GEVI) protein construct ArcLight-A242 (a gift from Vincent Pieribone; Addgene plasmid # 36857; http://n2t.net/addgene:36857; RRID:Addgene_36857) (Jin et al., 2012) or ASAP2 (pcDNA3.1/Puro-CAG-ASAP2s was a gift from Francois St-Pierre; Addgene plasmid # 101274; http://n2t.net/addgene:101274; RRID:Addgene_101274) (Chamberland et al., 2017). To investigate TCS effects on cytosolic pH, RBL cells were transfected with plasma membrane-targeted Lyn-tailed mCherry-SuperEcliptic (SE) pHluorin protein construct (Lyn-tailed mCherry-SEpHlourin was a gift from Sergio Grinstein; Addgene plasmid # 32002; http://n2t.net/addgene:32002; RRID:Addgene_32002) (Koivusalo et al., 2010). RBL cells were transiently transfected with ArcLight A242, ASAP2, or Lyn-tailed mCherry-SuperEcliptic (SE) pHluorin through electroporation using RBL-specific Amaxa Nucleofector transfection kit T (Lonza), then plated in 8-well plates (ibidi) at ~150,000 cells/well in 200 μL/well phenol red-free RBL media. Cells were grown for 16–24 hrs at 37°C/ 5% CO₂. Before imaging, cells were washed with BT and the media was replaced with 200 μL BT. After the initial, untreated, image was taken, cells were treated with 200 μL of either BT (for control) or various 2X concentrations of TCS or gramicidin, depending on the experiment. See “Confocal Microscopy” below for imaging details.

Plasma Membrane Potential Assay in Jurkat T Cells

Ibidi 8-well plates were coated with 150 μL of human fibronectin (VWR) prepared at 166 μg/ml in phosphate buffered saline (PBS) (Lonza) (Wang et al., 2019) overnight in the tissue culture incubator (37°C/5% CO₂); these wells were washed the next day with 200 μL PBS/well before use. Jurkat cells were transiently transfected with ArcLight-A242 through electroporation using Jurkat-specific Amaxa Nucleofector transfection kit T (Lonza), then plated in the fibronectin pre-coated plates at ~1million cells/well in 300 μL phenol red-free Jurkat media. Cells were grown for 16–24 hours at 37°C/ 5% CO₂. Next day, cell media was removed carefully with a transfer pipette to avoid cell detachment and replaced with 200 μL control BT before imaging. After the initial, untreated, image was taken, cells were treated with 200 μL of either additional BT (for control) or various 2X concentrations of TCS. See “Confocal Microscopy” below for imaging details.

Confocal Microcopy

For plasma membrane potential and cytosolic pH assays, an Olympus FV-1000 confocal microscope, with an Olympus IX-81 inverted microscope and a 30 milliwatt multi-argon laser, was used to collect images. ArcLight-A242 and ASAP2 plasmid-transfected cells were imaged using an oil immersion 100x objective with NA 1.4 and 488 nm excitation, 505–605 nm band pass emission filter. Lyn tailed mCherry-SEpHluorin transfected cells were imaged using oil immersion 60x objective with NA 1.4 and 488 nm excitation, 505–525 nm emission filter. All images were taken at 37°C using ibidi plate heating system.

Manual Image Analysis

All image-analysis figures in this paper were generated with this manual method except for Figure 8 and Supplement Figure 3, which were analyzed with the automated image analysis procedure noted below. Confocal microscopy images of RBL and Jurkat cells transfected with ArcLight and ASAP2 were analyzed manually using Fiji ImageJ software (NIH). Well-transfected cells (i.e. visible expression of reporter and well-targeted to the plasma membrane) were identified and a Region of Interest (ROI) was drawn on the plasma membrane using the free-hand tool. The area of the ROI (in units of square pixels), integrated density (= sum of fluorescence intensity values of all pixels in the ROI), mean intensity (= integrated density divided by area of the ROI), and length of the ROI were subsequently measured at various specified time points before and after treatment.

Figure 8.

Triclosan effects on fluorescence of Lyn-tailed mCherry-SuperEcliptic (SE) pHluorin in RBL mast cells. RBL cells were transiently transfected with Lyn-tailed mCherry-SE pHluorin, washed with BT, and exposed to BT (N=35) or 20 μM TCS (N=52) for 15 minutes. At each time point, the average fluorescence of each plasma membrane was measured, background-subtracted, and normalized to the 0 min timepoint, as described in Image J Automated Methods. Values presented are means ± SEM of 3 independent days of experiments per treatment. Statistically significant results at each timepoint, as compared to the appropriate control (0 μM TCS or 0 μM gramicidin + DMSO vehicle), are represented by ***p < 0.001, as determined by one-way ANOVA followed by Tukey’s post-hoc test.

Mean background fluorescence intensity (=integrated density of ROI divided by area of ROI) for a field of view, was obtained by either drawing a square shape ROI in the cytoplasmic region of transfected cells or by drawing an ROI around the plasma membrane of untransfected cells. Background mean fluorescence intensity obtained through either of these methods was subtracted from mean ArcLight fluorescence intensity of cells present in that field of view, and the resulting value at each time point was then normalized to (divided by) its own 0 min timepoint value of mean intensity. Both of these background subtraction methods resulted in similar ArcLight fluorescence changes due to 20 μM TCS: data for 7 cells selected randomly at the 15 min with 20 μM TCS, the cytosolic square shape method of background subtraction results in a calculated decrease of 36% ± 7% SEM in Arclight fluorescence while, for the same cells, the background subtraction method using a plasma membrane trace of untransfected cells results in a calculated decrease of 36% ± 8% SEM.

Photobleaching or drifting effect during imaging was calculated from BT-treated (control) cells by calculating a change in fluorescence values from 0 min. This change value was added/subtracted back to the treatment groups to account for these non-treatment effects and thus allowing measure of the true effect of a drug. Normalized values from multiple cells were averaged and used to generate line plots.

Automated Image Analysis

In FIJI image J, individual images at different time points (0, 2, 5, 7, and 15 min) were converted to 8-bit stack. Background fluorescence was subtracted from the stack using pseudo flatfield correction. Next, by applying threshold, binary masks of the entire cell and only cytosol were created. Both of these masks were applied to transfected cells to obtain area of the ROI, integrated density, and mean intensity values. To find the area of the ROI, mean intensity and integrated density of plasma membrane, values of cytosol were subtracted from the area of the ROI and integrated density of whole cell. The calculated integrated density was divided by the calculated area of ROI. This will give mean fluorescence intensity of the plasma membrane. Mean intensity at different time points of an individual cell was normalized to (divided by) its own 0 min time point value of mean intensity. Fluorescence was adjusted for drifting or photobleaching as described in “Manual Image Analysis” Methods above. This automated method was utilized to generate Figure 8 and Supplement Figure 3.

Analysis of Triclosan Effects on ArcLight-A242 Fluorescence Intensity as a Function of Initial ArcLight-A242 Expression Level

Initial (0 min) mean plasma membrane fluorescence intensity of ArcLight for each individual cell was noted. Next, for each individual cell, the percentage drop in ArcLight fluorescence mean intensity incurred by 15 min exposure to 20 μM triclosan (or to control BT) was then plotted as a function of that cell’s initial mean intensity of ArcLight (its expression level of the construct). Linear regression analysis was performed, and the slope of the best-fit line and R² value was determined for each plot.

Apparent Plasma Membrane Potential Percentage Decrease Calculation

Apparent PMP percentage decrease was calculated for the treatments, cell types, and GEVIs utilized in this study. Each normalized and photobleaching-/drift-corrected GEVI fluorescence value was subtracted from 1 to determine change in fluorescence due to treatment. This result was subsequently multiplied by its respective reporter’s PMP to fluorescence change ratio to calculate mV change as reported by each construct. For ArcLight-A242, this ratio is 100 mV for every 35% decrease in fluorescence (100 mV/0.35) (Jin et al., 2012). For ASAP2, this ratio is 100 mV for every 39% decrease in fluorescence (100 mV/0.39) (Chamberland et al., 2017). This change, in units of mV, was then divided by the resting PMP of the corresponding cell type, RBL mast cell (−82.5 mV is the average value from two patch-clamp studies (Lindau and Fernandez, 1986; Wischmeyer et al., 1995)) or T cell (−60 mV)(Sarkadi et al., 1990) to determine the (unitless) change in total resting cell PMP. Finally, this result was subsequently subtracted from 1 and multiplied by 100 to obtain the apparent PMP percentage.

TCS Cytotoxicity on Jurkat Cells

Trypan blue exclusion and lactate dehydrogenase (LDH) cytotoxicity assays were used to assess TCS cytotoxicity on Jurkat cells. In the trypan blue exclusion assay, 1 million cells were plated into each of 3 wells of a 24-well, flat-bottom cell culture plate (Costar) for TCS (Sigma) treatment and 3 wells for the control. Immediately after plating, BT and 2X TCS doses were added to respective wells. Next, cells were incubated for 30 minutes at 37°C/5% CO₂. After the incubation, a small sample was taken from each well and mixed 1:1 with trypan blue dye (0.4%, Lonza) evenly. Cells were counted using a hemocytometer for viability and TCS-treated cells were normalized to the control. LDH cytotoxicity methods were those of Hutchinson et al., 2011 using a cytotoxicity detection kit (Roche). However, the lysis solution was added to the “high control” in the final 15 minutes of the 1-hour TCS exposure instead of an additional 15 minutes after the 1-hour TCS exposure (Palmer et al., 2012).

Cytosolic Ca²⁺ assay

Prior to performing Ca²⁺ assays, widefield fluorescence microscopy was used to assess transfection efficiency of GCaMP6 Ca²⁺ reporter construct in Jurkat T cells, and Ca²⁺ assays were performed only on samples of highly-transfected cells. First, 8-well ibidi plates were coated with 150 μL/well fibronectin (166 μg/ml) prepared in PBS and incubated overnight in tissue culture incubator. The next day, Jurkat T cells were transfected with pGP-CMV-GCaMP6f (a gift from Douglas Kim & GENIE Project; Addgene # 40755; http://n2t.net/addgene:40755; RRID: Addgene_40755) (Chen et al., 2013) using Jurkat specific Amaxa Nucleofector transfection kit T (Lonza). The electroporated cells were plated in phenol red-free Jurkat cell media at ~1 million cells/well with phenol red-free media in the fibronectin pre-coated 8-well ibidi plates. Cells were grown for 16–24 hours at 37°C/ 5% CO₂. The next day, cell media was removed carefully with a transfer pipette to avoid cell detachment and replaced with 200 μL BT, and images were taken with a wide-field IX83 (Olympus, Waltham MA) microscope with a Prime 95B CMOS Camera (Photometrics) and HLD117 stage (Prior Scientific) controlled by a Proscan III. Fluorescence excitation was provided by an Xcite 120 LEDBoost (Excelitas). Images were taken at 60x (Olympus-APON-60X-TIRF objective) using standard excitation and emission filters for GFP (Semrock). The microscope is controlled by CellSens software v1.18 (Olympus).

For assessment of Jurkat T cell cytosolic Ca²⁺ levels following anti-T cell receptor (anti-TCR) stimulation, a plate reader assay was performed similarly to that in (Weatherly et al., 2018), with the following modifications. First, a 96-well, black-walled, clear-bottom plate (Grenier) was coated overnight with 50 μL/well fibronectin (166 μg/ml) prepared in PBS (Lonza). The next day, cells were transfected with the GCaMP6 construct as noted above and plated in 200 μL/well phenol red-free media in the fibronectin pre-coated 96-well plate at 330,000 cells/well. Cells were cultured overnight at 37°C and in 5% CO₂. Next day, media was carefully aspirated from wells with a micropipette to avoid cell detachment, and cells were exposed to 100 μL/well 0.2 μg/ml (Holowka et al., 2018) Anti-TCR OKT3 monoclonal antibody (Thermo Fisher Scientific) in combination with control BT or TCS treatments for 1 hour. Fluorescence was measured with 485 ± 10 nm excitation and 528 ± 10 nm emission during this 1 hour. Area under the curve was determined as per methods in Weatherly et al., 2018.

Statistical Analyses

All analyses were performed in Graphpad Prism. In most figures, raw values for treatment groups were normalized to its appropriate untreated control. Biological replicates from at least three independent experiments were averaged, and standard error of the mean (SEM) was calculated across those independent experiments. Raw (non-normalized) values were analyzed for mitochondria swelling and LDH release experiments (Figures 1, 6D). One-way ANOVA followed by Tukey’s post-hoc test (α=0.05) was used to determine the significance level of most experiments (Figures 2, 3B, 5, 6, 7B, 8, Supplement Figures 2–6). The significance level of the FPALM mitochondrial volume and surface area were assessed using an unpaired one-tailed t-test (Figure 1). The significance level of the cytosolic Ca²⁺ level area under the curve (AUC) of Jurkat cells was assessed via paired t-test (Figure 9B). Significance is represented by ***p<0.001, **p<0.01, and *p<0.05.

Figure 1.

Triclosan effects on mitochondrial volume and surface area in RBL mast cells. Super-resolution FPALM 3D images of Dendra2-TOM20, which labels outer mitochondrial membranes, were processed through custom-built MATLAB code in which an algorithm identifies individual mitochondria. As detailed in the Methods, (A) average mitochondrial volume was calculated by the convex hull method, and (B) average mitochondrial surface area was calculated by the alpha shape method. Values presented are mean SEM for total of 62 cells in three independent experiments, where each independent experiment had 8–11 cells for each condition (0 vs. 20 μM triclosan). Statistically significant results are represented by **p<0.01 and ***p<0.001 as compared to control (0 μM TCS) for volume and surface area, respectively, as determined by unpaired t-test.

Figure 6.

Triclosan effects on fluorescence of ArcLight-A242 in human Jurkat T cells, triclosan apparent effects on Jurkat plasma membrane potential (PMP), and cytotoxicity determination. (A) Jurkat cells were transiently transfected with ArcLight-A242, washed with BT, exposed to BT (0 μM TCS) (N=20) or 20 μM TCS (N=17) for 15 minutes. At each time point, the average fluorescence of each plasma membrane was measured, background-subtracted, and normalized to the 0 min timepoint, as described in Methods. (B) Data from Figure 6A were utilized to calculate percentage of PMP (with each 0 min timepoint defined as 100%) as a function of time (min), as described in Methods. (C) Jurkat T cell cytotoxicity to TCS was assessed by trypan-blue exclusion assay and by (D) lactate dehydrogenase (LDH) detection kit from Roche. “High control” is a sample treated with lysis solution provided by the kit. Values presented are mean ± SEM of at least three independent experiments. Statistically significant results, as compared to appropriate control, are represented by **p<0.01, ***p < 0.001, as determined by one-way ANOVA followed by Tukey’s post-hoc test.

Figure 2.

Relative degranulation response of antigen-stimulated RBL mast cells exposed to micromolar doses of the canonical plasma membrane depolarizer gramicidin (A and B) or to triclosan (C and D). IgE-sensitized cells were stimulated for 1 h with either 0.0004 μg mL⁻¹ Ag (A and C) or 0.001 μg mL⁻¹ Ag (B and D). All samples (A - D) contained 0.003% DMSO, which was the vehicle required for gramicidin dissolution. For spontaneous release (“Spont” on x-axes), cells were incubated for 1 h in BT with 0.003% DMSO (with no IgE, Ag, TCS, or gramicidin). Values represent the mean ± SEM of 3 independent experiments; three replicates per treatment type were used each experimental day. Statistically significant results, as compared to the appropriate control (0 μM TCS + DMSO vehicle) or (0 μM gramicidin + DMSO vehicle), are represented by *p < 0.05, **p < 0.01, ***p<0.001 as determined by one-way ANOVA followed by Tukey’s post-hoc test.

Figure 3.

Triclosan and gramicidin effects on fluorescence of ArcLight-A242 in RBL mast cells. (A) One representative live-cell confocal microscopy image of an RBL mast cell transiently transfected with ArcLight construct, prior to gramicidin or TCS treatment. Scale bar, 10 μm. (B) RBL cells were transiently transfected with ArcLight, washed with BT, exposed to control (BT) (N = 52), 10 μM TCS (N = 22), 20 μM TCS (N = 28), or 1 μM gramicidin (N = 24) for 15 minutes. At each time point, the average fluorescence of each plasma membrane was measured, background-subtracted, and normalized to the 0 min timepoint, as described in Methods. Values presented are means ± SEM of at least 3 independent days of experiments per treatment. Statistically significant results at each timepoint, as compared to the appropriate control (0 μM TCS or 0 μM gramicidin + DMSO vehicle), are represented by **p < 0.01, ***p < 0.001, as determined by one-way ANOVA followed by Tukey’s post-hoc test.

Figure 5.

Triclosan and gramicidin apparent effects on plasma membrane potential (PMP) of RBL mast cells. Data from Figure 3B were utilized to calculate percentage of PMP (with each 0 min timepoint defined as 100%) as a function of time (min), as described in Methods. Values presented are means ± SEM of at least 3 independent days of experiments per treatment. Statistically significant results at each timepoint, as compared to the appropriate control (0 μM TCS or 0 μM gramicidin + DMSO vehicle), are represented by ***p < 0.001, as determined by one-way ANOVA followed by Tukey’s post-hoc test.

Figure 7.

Triclosan effects on fluorescence of ASAP2 in RBL mast cells. (A) One representative live-cell confocal microscopy image of an RBL mast cell transiently transfected with ASAP2 construct, prior to TCS treatment. Scale bar, 10 μm. (B) RBL cells were transiently transfected with ASAP2, washed with BT, exposed to control (BT) (N=18) or 20 μM TCS (N=21) for 15 minutes. At each time point, the average fluorescence of each plasma membrane was measured, background-subtracted, and normalized to the 0 min timepoint, as described in Methods. Values presented are means ± SEM of at least 6 independent experiments per treatment. Analysis by one-way ANOVA followed by Tukey’s post-hoc test found no statistical significance.

Figure 9.

Triclosan effects on cytosolic Ca^{2 +} levels as activated by anti-T Cell Receptor (TCR) antibody in Jurkat T Cells. Jurkat cells were transiently transfected with cytosolic Gcamp6 overnight, then exposed in BT Control or 20 μM TCS, each with 0.2 μg/ml anti-TCR (“Ab”) for 1 hour. “Unstimulated” contained no TCS and no anti-TCR. Raw fluorescence was measured using a microplate reader as described in Methods. (A) Representative figure (no error bars) shows fluorescence obtained following background fluorescence subtraction (mock-transfected cells’ fluorescence) from treatment groups at each time point. (B) Area under the curve (AUC) was obtained from the fluorescence curves and values were normalized to the control of each day for 3 replicates per treatment group per day. AUC values presented are means ± SEM of 3 independent days of experiments. Statistically significant results, as compared to the appropriate control (0μM TCS), are represented by *p < 0.05, as determined by unpaired t-test.

Results

Triclosan increases mitochondrial volume and surface area in RBL-2H3 mast cells: Indicators of inhibited mitochondrial membrane potential

Previously, an organic dye fluorescence method (without imaging) revealed that TCS inhibits MMP (Weatherly et al., 2018). To test this finding at the nanoscale, three-dimensional super-resolution FPALM imaging (Huang et al., 2008; Parent and Hess, 2019) was employed to assess TCS effects on 3D mitochondrial morphology, changes in which are linked to inhibition of MMP (Guillery et al., 2008; Giedt et al., 2012; Weatherly et al., 2018). RBL cells were transiently transfected with Dendra2-TOM20 (Weatherly et al., 2018), a construct used to label the outer membrane of mitochondria. The next day, cells were exposed with TCS or BT control for 1 hour, then fixed using paraformaldehyde. While the TCS effects on mitochondrial volume and surface area were not dramatic visually (a no-TCS control figure is shown in Supplement Figure 1), they were statistically significant when analyzed quantitatively. Triclosan statistically significantly increases the mean volume, as calculated by the convex hull method, by 17% compared to control (Figure 1A). Also, triclosan increases the mean mitochondrial surface area, as calculated by the alpha shape method, by 19% (Figure 1B). Thus, TCS modulates the surface area and volume of individual mitochondria, as assessed with 3D FPALM super-resolution microscopy.

Plasma membrane potential depolarizer gramicidin potently inhibits the degranulation of RBL-2H3 mast cells

To determine the effects of the canonical plasma membrane depolarizer, gramicidin, on RBL mast cell degranulation, an adapted fluorescence-based β-hexosaminidase release assay (Weatherly et al., 2013) was employed. A multivalent DNP (dinitrophenol)-BSA Ag was used to laterally crosslink IgE-bound FcεRI receptors, thus initiating an allergic/pro-inflammatory signal transduction that ends with the release of granules containing bioactive substances including β-hexosaminidase, which is monitored fluorometrically. Gramicidin depolarizes cells by forming an ion channel at the plasma membrane, allowing for free passage monovalent ions such as H⁺, NH₄⁺, K⁺, Na⁺, and Li⁺ down their concentration and electrochemical gradients, and, thus, depolarizing the plasma membrane (Myers and Haydon, 1972; Mohr and Fewtrell, 1987b).

Figure 2A presents the results for IgE-sensitized RBL cells incubated for 1 h in Tyrodes—BSA (BT) buffer containing a 0.0004 μg/ml Ag dose, with gramicidin or DMSO vehicle. All tested gramicidin doses, as low as 0.1 μM, inhibit degranulation, reducing the response down to the same level as the spontaneous group (unstimulated with Ag). These data indicate a potent, plasma membrane depolarization-mediated inhibition of degranulation. A similar, but slightly less potent, gramicidin-induced depression of degranulation can be seen in the group treated with the higher Ag dose, 0.001 μg/ml Ag (Figure 2B). In this case, statistically significant inhibition of degranulation does not begin until 0.5 μM gramicidin and appears to proceed in a dose-responsive fashion. Such data coincide with the previously reported observation that pharmacologically induced inhibition of degranulation occurs at lower toxicant doses when cells are stimulated at a low level (0.0004 μg/ml Ag) than when stimulated at a high level (0.001 μg/ml Ag) (Palmer et al., 2012).

A dose comparison study for gramicidin and TCS was performed as part of testing the hypothesis that TCS inhibits degranulation via PMP inhibition, by repeating the above experiments with TCS in place of gramicidin at equivalent antigen and DMSO concentrations. While DMSO is not needed/used here to dissolve the TCS, it was included to match the conditions of the gramicidin experiments. Statistically significant, dose-responsive inhibition of mast cell degranulation by TCS is reported in Figures 2C and 2D, beginning at 10 μM (Figure 2C). This observation recapitulates previously observed TCS-mediated inhibition of degranulation (Palmer et al., 2012), but the presence of DMSO reduces its potency: 10 μM TCS inhibits 0.0004 μg/ml Ag-stimulated degranulation by about one-half in the absence of DMSO (Weatherly et al., 2013) and only by about one-third in the presence of 0.003% DMSO (Figure 2C). Also, the TCS inhibitory effect lessens in the presence of higher Ag dose (0.001 μg/ml Ag; Figure 2D), as noted with gramicidin Figure 2B. Overall, depolarizer-mediated inhibition of degranulation (Figure 2A and 2B) proceeds similarly as that of TCS, though gramicidin is more potent than TCS, especially in the presence of DMSO. The result strengthens the connection between PMP inhibition and degranulation inhibition.

Gramicidin and triclosan depress ArcLight-A242 fluorescence in RBL-2H3 mast cells.

Gramicidin, a known depolarizer of PMP, strongly inhibits degranulation (Figs. 2A and 2B), and TCS also reduces degranulation (Palmer et al., 2012). Also, TCS depolarizes the MMP (Weatherly et al., 2018) (Figure 1). Thus, we hypothesized that TCS also inhibits the PMP, as its underlying mechanism of degranulation inhibition. To do so, an organic voltage-sensitive fluorescent dye, DiSC₃(5) was employed. However, when TCS (micromolar levels) and DiSC₃(5) (1 μM) (Te Winkel et al., 2016) are mixed together in the absence of cells, TCS chemically quenches the fluorescence of DiSC₃(5) in a dose-response manner (Supplement Figure 2); thus, DiSC₃(5) cannot be used to accurately measure TCS effects on PMP. Triclosan quenching of unprotected fluorophores (such as those found in various voltage-sensitive, fluorescent organic dyes) has previously been observed (Weatherly, 2017; Weatherly et al., 2018). Thus, to measure TCS effects on PMP, genetically encoded voltage indicators (GEVIs), which contain β-barrel protein structures which may protect their fluorophores from chemical quenching or aggregation effects (Chalfie and Kain, 2005), were used. Triclosan does not interfere with the fluorescence of the fluorescent protein within the reporter construct GCaMP (Weatherly et al., 2018).

To measure PMP of cells using a GEVI, RBL cells were transiently transfected with ArcLight-A242 (ArcLight) (Jin et al., 2012) (Figure 3A). The next day, confocal images were collected for each cell at different time points, up to 15 minutes, before (Figure 3A) and after (Figure 3B) addition of control (BT), TCS, or gramicidin. Fluorescence obtained was quantified using FIJI image J (Manual Image Analysis Methods) and normalized to the 0 min time point for each condition. Gramicidin, a known PMP depolarizer, statistically significantly reduces ArcLight fluorescence within 2 min of exposure and by 23% (± 2% SEM) at the end of 7 minutes. TCS, at 20 μM, statistically significantly lowers ArcLight fluorescence within 5 min of exposure and by 25% (± 3% SEM) at the end of 15 min. This result was further confirmed by an automated image analysis method (Supplement Figure 3). The 10 μM TCS data suggest an inhibition of ArcLight fluorescence by 15 min but were not statistically significant. The decrease in ArcLight fluorescence suggests that both gramicidin and TCS inhibit mast cell PMP.

Since gramicidin treatment contained DMSO vehicle (0.003%), it was important to determine whether gramicidin’s effects on ArcLight fluorescence are modulated by the presence of DMSO. Thus, a control experiment was performed by repeating the gramicidin-ArcLight experiments in the presence of varying DMSO concentrations. Gramicidin’s ability to inhibit ArcLight fluorescence is unaffected by increasing DMSO concentrations, ranging from 0.003% to 0.01% (Supplement Figures 4 and 5).

After experiments with transiently-transfected cells were successfully conducted, ArcLight stably-transfected cell lines were created, with the goal of enabling high-throughput testing of numerous concentrations of triclosan and other toxicants, as well as additional time points. However, unfortunately, clones of RBL cells stably transfected with ArcLight display construct aggregation and poor plasma membrane localization (Supplement Figure 6).

To check whether plasmid expression level affects the change in GEVI fluorescence in response to triclosan, the initial, 0 min (pre-triclosan-exposure), background-subtracted mean fluorescence intensity of each individual cell from Figure 3’s data was analyzed. The percentage drop in ArcLight fluorescence by 15 minutes time of exposure (control buffer-treated in 4A, TCS-treated in 4B), for each individual cell, was plotted as a function of that particular cell’s initial mean intensity of ArcLight. There is no correlation between plasmid expression level and the magnitude of TCS inhibition of ArcLight fluorescence (Figure 4B), as analyzed by linear regression. Thus, TCS effects on GEVI fluorescence are unaffected by the “brightness” (GEVI expression level) of the cell.

Figure 4.

Effect of cellular expression level of ArcLight-A242 on TCS inhibition of ArcLight-A242 fluorescence. Data from Figure 3B, ArcLight-A242 transfected cells analyzed by confocal imaging and image analysis comparing control (0μM TCS) and 20μM TCS treatments, were plotted as individual cells’ data. The percentage drop in fluorescence by 15 minutes time of exposure, for each individual cell, is plotted as a function of that cell’s initial mean intensity of ArcLight-A242. Statistical results per plot are represented by equation of linear regression and R² values.

Plasma membrane depolarization causes ArcLight fluorescence intensity to decrease, in a linearly proportional fashion (Jin et al., 2012). PMP percentage decline can be calculated from the measured change in ArcLight fluorescence, as outlined in Methods. This apparent mast cell PMP, as a percentage of initial, 0 min, pre-triclosan-exposure value, is plotted as a function of time following TCS or gramicidin exposure (Figure 5). Within 7 min of exposure to 1 μM gramicidin, RBL mast cell PMP appears to decrease by 79% (± 7% SEM) of initial resting PMP. Similarly, within 15 min of exposure to 20 μM TCS, RBL cell PMP appears to decrease by 90% (± 11% SEM). There appears to be a modest, but not statistically significant, decline in PMP within 15 min exposure to 10 μM TCS.

Triclosan inhibits ArcLight fluorescence, and, thus, apparently inhibits RBL cell PMP. Thus, the next experiments tested whether these results are extendable to another immune cell type which is also dependent on PMP and SOCE for its function, T-cells.

Non-cytotoxic doses of triclosan depress ArcLight-A242 fluorescence in Jurkat T cells.

To measure triclosan effects on PMP of T cells, human Jurkat T cells were transiently transfected with ArcLight. The next day, confocal images were collected for each cell at different time points, up to 15 minutes, before and after addition of control (BT) or TCS (Figure 6A). Fluorescence obtained was quantified using FIJI image J (Manual Image Analysis Methods) and normalized to the 0 min time point for each condition. TCS, at 20 μM, lowers ArcLight fluorescence within 2 min of exposure and by 31% (± 3% SEM) at the end of 15 min (Figure 6A). The decrease in ArcLight fluorescence suggests that TCS inhibits T cell PMP.

This apparent T cell PMP was calculated (see Methods) as a percentage of initial, 0 min, pre-triclosan-exposure value and is plotted as a function of time following TCS exposure (Figure 6B). Within 15 min of exposure to 20 μM TCS, T cell PMP appears to decrease to −150% (± 15% SEM) of initial resting PMP, apparently a more-than-complete dampening of T cell PMP.

In order to determine if the results obtained above were truly functional changes or due to cytotoxicity from TCS, the effect of micromolar doses of TCS (0 μM, 10 μM, and 20 μM) on Jurkat T cell viability were assessed using a trypan blue exclusion assay. Results indicate that TCS does not cause a decrease in cell viability within a 30 min exposure (Figure 6C). In order to confirm these cytotoxicity results, a more sensitive microplate reader-based assay, LDH assay was also performed. LDH release was measured in response to varying doses of TCS for 1 hour. There is no significant release of LDH from TCS-treated cells as compared to control; the “high control” sample is a positive control detecting LDH release from lysed cells (Figure 6D). These data from cytotoxicity experiments indicate that TCS dosage timing and concentration used in this study do not cause Jurkat T cell cytotoxicity.

Triclosan does not depress the fluorescence of the genetically encoded voltage indicator ASAP2 in RBL-2H3 mast cells.

To check the results obtained using ArcLight, an alternative GEVI, called ASAP2, which utilizes a different voltage-sensing mechanism from that of ArcLight (Jin et al., 2012; Chamberland et al., 2017) was used. RBL mast cells were transiently transfected with ASAP2 (Figure 7A). The next day, confocal images were collected for each cell at different time points, up to 15 minutes, before (Figure 7A) and after (Figure 7B) addition of control (BT) or TCS. Fluorescence obtained was quantified using FIJI image J (Manual Image Analysis Methods) and normalized to 0 min time point for each condition. Triclosan (20 μM, up to 15 min) does not alter fluorescence of ASAP2 when compared to the 0 μM TCS control (Figure 7B). In stark contrast to the clear triclosan dampening of fluorescence observed with the ArcLight reporter (Figure 3), these ASAP2 results (Figure 7B) indicate that TCS does not change the PMP of RBL mast cells.

However, TCS inhibits the fluorescence of ArcLight, which contains a pH-sensitive super ecliptic pHlourin on the cytoplasmic side of the plasma membrane. Thus, the next investigation centered on whether TCS-induced depression of ArcLight fluorescence is due to a change in cytosolic pH instead of a change in PMP.

Triclosan depresses fluorescence of a plasma membrane-targeted pHlourin in RBL mast cells, indicating triclosan reduction of cytosolic pH.

To identify whether TCS affects cytosolic pH, plasma membrane-targeted fluorescence reporter called Lyn-tailed mCherry-SEpHlourin (Koivusalo et al., 2010), which measures sub-plasma membrane cytosolic pH of a cell, was employed. The transfection and imaging procedures used for the GEVI experiments were repeated, along with an automated image analysis technique (described in Methods). Triclosan, at 20 μM, decreases mCherry-SEpHlourin fluorescence intensity within 5 min of exposure and by 43% (± 2% SEM) at the end of 15 min (Figure 8). This result is similar to the magnitude of triclosan’s effect on ArcLight fluorescence (Figure 3), suggestive of a pH change in response to TCS rather than a PMP change.

TCS inhibits cytosolic Ca²⁺ response to Anti-T Cell Receptor stimulation in Jurkat T cells

TCS affects Ca²⁺ dynamics and inhibits SOCE in mast cells (Weatherly et al., 2018). Due to the similar triclosan depression of ArcLight fluorescence in Jurkat T cells (Figure 6A) as in RBL mast cells (Figure 3B), we hypothesized that a similar calcium effect also occurs in Jurkat T cells. To test this hypothesis, Jurkat T cells were transfected with Ca²⁺ reporter construct GCaMP6 (Chen et al., 2013). The next day, cells were stimulated with 0.2 μg/ml Anti-T Cell Receptor (TCR) OKT3 ± TCS (Holowka et al., 2018), and cytosolic Ca²⁺ was measured throughout the duration of 1 hour. Cells were transfected at a highly efficient rate (Supplement Figure 7). Following anti-TCR stimulation, an initial rise in Ca²⁺ is seen within the first few minutes, followed by a plateau Ca²⁺ level above unstimulated basal level (compare “0 μM TCS + Ab” to “Unstimulated” curves in Figure 9A), as expected (Holowka et al., 2018). TCS, at 20 μM, inhibits Ca²⁺ levels in comparison to the control in anti-TCR activated cells, in particular by flattening the plateau region (Figure 9A). Triclosan minimally affects the initial Ca²⁺ rise but heavily dampens calcium in the plateau region, similar to its effects in RBL mast cells (Weatherly et al., 2018). AUC analysis reveals an average decrease of 70% (± 17%) integrated Ca²⁺ levels in anti-TCR activated cells with TCS treatment compared to that of activated control (Figure 9B).These data indicate that TCS inhibits cytosolic Ca²⁺ signaling in T cells.

Discussion

Triclosan disrupts mast cell function, rapidly (within tens of minutes) and at non-cytotoxic concentrations relevant to consumer product exposure (Palmer et al., 2012; Shim et al., 2019), but the underlying biochemical mechanism was unknown. In this study, we deduced the mode of action of TCS in mast cells and replicated these TCS effects in another immune cell type, T cells.

The process of deciphering the mechanism began with two observations: TCS is a mitochondrial uncoupler (Weatherly et al., 2016; Weatherly et al., 2018) and also a disruptor of the central signal for mast cell degranulation—cytosolic Ca²⁺ rise (Weatherly et al., 2018). As an uncoupler, TCS depresses the MMP, which results in mitochondrial fragmentation, observed in live cells with super-resolution microscopy (Weatherly et al., 2018). Mitochondrial toxicity and deformation are important markers of TCS toxicity in numerous cell types and species (Shim et al., 2016; Weatherly et al., 2018; Weatherly et al., 2020). Disruption of mitochondrial shape plays a role in diseases such as cognitive decline (Hara et al., 2014), Parkinson’s disease (Cui et al., 2010; Bhandari et al., 2014), insulin resistance (Jheng et al., 2012), immune dysfunction (Weatherly et al., 2020), inflammation (Compan et al., 2012), and disrupted embryonic development (Chen et al., 2003). In this study (Figure 1), we have augmented these findings by utilizing 3D super-resolution FPALM to reveal TCS enhancement of mitochondrial surface area and volume, indications of mitochondrial swelling and MMP disruption (Guillery et al., 2008; Giedt et al., 2012). This increase in surface area and volume has also been observed in neuronal mitochondria due to exposure to carbonyl cyanide-4 (trifluoromethoxy)phenylhydrazone (FCCP), a known proton ionophore and MMP depolarizer (Safiulina et al., 2006). The connections between mitochondrial deformation and disrupted MMP are extended to PMP depolarization, inhibited store-operated calcium entry (SOCE), and, subsequently, dampened mast cell degranulation via studies with a similar proton ionophore mitochondrial uncoupler, carbonyl cyanide m-chlorophenylhydrazone (CCCP), which also inhibits MMP, PMP, SOCE, and mast cell degranulation (Mohr and Fewtrell, 1987b). Triclosan’s enhancement of mitochondrial surface area and volume led to a PMP collapse hypothesis as the mode of action of immune cell dysfunction.

Thus, we hypothesized that TCS inhibits not just MMP but also PMP. Direct evidence for this hypothesis was provided by reports that TCS induces an electrical current (evidence of its proton ionophore nature) across artificial membranes (Popova et al., 2018) and depolarizes neuronal PMP (Arias-Cavieres et al., 2018; Popova et al., 2018) at micromolar doses within tens of minutes. As calculated in the Introduction, a complete depolarization of PMP in mast cells would lead to a ~45% reduction in the driving force for cytosolic Ca²⁺ influx into the cell through activated CRAC channels; indeed, Weatherly et al., 2018 reported ~45% decrease in cytosolic calcium levels after TCS exposure in RBL mast cells. This study also reports (Figure 9) a robust decrease in cytosolic calcium levels after TCS exposure in Jurkat T cells. In turn, the decrease in cytosolic Ca²⁺ inhibits degranulation (Holowka et al., 2012). Additionally, the known PMP depolarizer gramicidin inhibits mast cell degranulation in parallel conditions to TCS mast cell disruption (Figure 2), lending further support for the PMP collapse hypothesis for the TCS mode of action.

To measure TCS effects on PMP, two genetically encoded voltage indicator (GEVI) reporter constructs, ArcLight-A242 (ArcLight) (Jin et al., 2012) and ASAP2 (Chamberland et al., 2017), were utilized. ArcLight senses changes in PMP due to its two specialized subunits: a transmembrane voltage-sensing domain which, upon changes in PMP, undergoes conformational changes that cause conformational changes in the attached green fluorescent protein (GFP)-derived, super ecliptic pHlourin fluorescent domain which is localized intracellularly, in the cytosol. ASAP2 also contains a transmembrane voltage-sensing domain but, in contrast to ArcLight, ASAP2 contains an extracellularly-located, circularly-permuted (non-pHlourin-based) fluorescent GFP (cpGFP) domain. For both of these GEVIs, their fluorescence intensity decreases upon PMP depolarization. These experiments yielded drastically different results: Triclosan inhibits ArcLight fluorescence in RBL mast cells (Figure 3) and in Jurkat T cells (Figure 6), suggesting TCS-induced collapse of PMP in these cell types (Figures 5 and 6). Note that the effect was not statistically significant with 10 μM TCS but did suggest a modest inhibition at that dose/timeframe. In contrast, TCS had no effect on ASAP2 fluorescence (Figure 7). While both ArcLight and ASAP2 are plasma membrane voltage-sensitive, there is a crucial difference in their structures: the location and pH sensitivity of their fluorescent protein domains. ArcLight’s fluorescent domain is a pHlourin, which is strongly pH-sensitive in the physiological pH range (Jin et al., 2012) (Supplement Figure 2B from this article shows that a pH reduction from 7.5 to 6.5 would drastically reduce this pHlourin’s intensity, by ~50% or more). It is located in the cytosol, so it reports changes in cytosolic pH, with acidification leading to fluorescence intensity decline. ASAP2 cannot report cytosolic pH changes because its fluorescent domain is located outside the cell, unless those changes affect the PMP (personal communication, Dr. Francois St-Pierre). Interpretation of the ASAP2 data is somewhat confounded by the fact that its cpGFP fluorescent domain is also somewhat pH sensitive (Miyawaki and Niino, 2015; Kostyuk et al., 2019). While TCS mitochondrial toxicity could lead to extracellular acidification which could affect ASAP2 fluorescence, TCS-induced extracellular acidification likely occurs on a longer timescale (Ajao et al., 2015) than the 15 min exposure assessed in the current study. Also, if TCS were acidifying the extracellular space around the RBL mast cells and Jurkat T cells within the experimental exposure time (15 min), overcoming the buffering of the BT solution, a decrease in ASAP2 fluorescence would have occurred (Stepanenko et al., 2008). Together, these data suggest that TCS does not affect mast cell or T cell PMP but, instead TCS lowers cytosolic pH.

Indeed, TCS acidifies the cytosol, as measured with lyn-tailed mCherry-SEpHluorin, a pH-sensitive, PMP-insensitive reporter construct (Figure 8). ArcLight and lyn-tailed mCherry-SEpHluorin share the same pH-sensitive fluorophore, the super ecliptic pHlourin (Koivusalo et al., 2010; Jin et al., 2012). A decrease in lyn-tailed mCherry-SEpHluorin’s fluorescence is equivalent to a cytosolic pH decrease (Koivusalo et al., 2010). This intracellular pH effect may either be localized, or be most drastic, near the plasma membrane itself as both of these pH-sensitive reporters are plasma membrane-localized (Koivusalo et al., 2010; Jin et al., 2012). As noted in Methods, a 35% decrease in the fluorescence intensity of ArcLight corresponds with a 100 mV PMP decrease (Jin et al., 2012); thus, a 29% decrease in fluorescence intensity would correspond with a complete depolarization of the RBL cell PMP (~82.5 mV) (Lindau and Fernandez, 1986; Wischmeyer et al., 1995). Thus, if the ArcLight response were measuring PMP changes in response to TCS, the maximum response of any cell should have been ~29%. On the contrary, nearly half of cells responded to TCS exposure with a greater-than-29% drop in ArcLight fluorescence (Figure 4). Some of this variation is likely due to cell-to-cell variation in signaling, even in this clonal cell line (Millard et al., 1988).

However, because ArcLight’s fluorescence can also communicate changes in pH and can be depressed nearly 100% within a 1 pH unit acidification starting at normal physiological pH (Jin et al., 2012), the cell-to-cell variation of TCS response in Figure 4 provides further evidence for TCS acidification of the cytosol.

Taking into account the pH sensitivity of its fluorophore (Jin et al., 2012), the ArcLight data (Figure 3) can be re-interpreted as a 20 μM TCS-induced pH decrease of −0.23 pH units (Table 1). The pH sensitivity of the fluorescence intensity of ArcLight’s super ecliptic A227D pHlourin (plotted in the Supplement of (Jin et al., 2012)) is approximately linear between pH 6.5 to 7.5. This relationship can be approximated using the following: F = 0.6pH – 3.8 where F is the fraction of maximal fluorescence, and the equation is valid over a range of ± 0.5 pH units. Thus, at a pH of 7.2, F = 0.52, meaning that ArcLight will exhibit 52% of its maximal fluorescence intensity at the starting, physiological pH. The 26% ± 3% (SEM) decline in ArcLight fluorescence due to 15 min of 20 μM TCS exposure (Figure 3) translates to a 26% reduction in this 0.52 value, bringing the final, normalized, ArcLight fluorescence intensity to 0.38 ± 0.02; solving for the corresponding pH value returns a final pH value of 6.97 ± 0.02 following this TCS exposure. This yields the −0.23 pH unit decrease estimate from the ArcLight data (Table 1). Additionally, the same data interpretation can be employed with lyn-tailed mCherry-SEpHluorin, resulting in a pH decrease to 6.83 ± 0.02, yielding the −0.37 pH unit decrease estimate (Table 1). These pHlourin-measured decreases in pH are similar in magnitude, providing corroborating evidence that TCS acidifies the cytosol.

Table 1.

Summary of estimated changes in cytosolic pH following TCS exposure.

Method	Magnitude of pH Change	Citation
Calculated estimate	−0.3	Discussion
ArcLight experiments	−0.23	Figure 3
pHlourin experiments	−0.37	Figure 8
Average	−0.3	Table 1

These measured pH changes were confirmed with theoretical calculations of TCS acid-base chemistry. Taking into account the 400 μL volume of 20 μM TCS in each ibidi well containing ~400,000 cells on each experimental day (see Methods), along with an estimated 10% absorption rate (Moss et al., 2000; Weatherly and Gosse, 2017), each cell received ~5.3 × 10⁻¹⁵ moles, or 3.2 × 10⁹ molecules of TCS. Considering its pKa of 7.9 (PubChem), a resting cell cytosolic pH of 7.2 (Johnson et al., 1980; Lodish et al., 2000; Beck et al., 2014), and the Henderson-Hasselbalch equation, 16.6% of TCS is deprotonated when it encounters the pH 7.2 resting cytosol. This 16.6% value equates to 5.3 × 10⁸ deprotonated TCS molecules per cell, and hence, 5.3 × 10⁸ excess H⁺ delivered to each cell as this monoprotic weak acid dissociates. While this excess proton dose, from TCS, far exceeds the proton concentration of the cell, it rapidly encounters the cellular buffering system. Carbonate (CO²/HCO₃⁻, pKa 6.1) buffering (Nelson and Cox, 2017) accounts for approximately two-thirds of a cell’s total buffering power (Boron, 2004), and average mammalian cells contain 12 mM of HCO₃⁻ (Lodish et al., 2000), the conjugate base of this buffering system (A⁻); these values and the Henderson-Hasselbalch equation can be used to calculate the concentration of the weak acid (HA) in the untreated cell: 0.953 mM. The average mammalian cell has a volume of 1766 μm³ (Barrandon and Green, 1985), and the cytosol itself accounts for 70% cellular volume (Luby-Phelps, 2000). Thus, the volume of the cytosol is 1236 μm³ or 1.236 ×10⁻¹² L. Multiplying this volume by the HA and A⁻ concentrations noted above and by Avogadro’s number yields the number of HA and A⁻ molecules found within the cytosol of an untreated cell: 7.1 × 10⁸ HA molecules and 8.9 × 10⁹ A⁻ molecules. The 5.3 × 10⁸ excess H⁺ delivered to the cell by TCS exposure interact 1:1 with these A⁻ molecules, thereby producing 5.3 × 10⁸ new HA molecules added to the original pool of HA molecules and subtracted from the original pool of A⁻ molecules: resulting in 1.24 × 10⁹ molecules of HA and 8.4 × 10⁹ A⁻ molecules. Plugging these values into the Henderson-Hasselbalch equation leads to a cytosolic pH of TCS-treated cells of 6.9: a −0.3 pH depression caused by TCS exposure (Table 1). In fact, the average of the pH depressions captured in the ArcLight and lyn-tailed mCherry-SEpHluorin measurements is −0.3, in agreement with this calculated value.

Acidification of the cytosol has been shown to decrease Ca²⁺ release-activated Ca²⁺ current (I_CRAC), the final step in SOCE, in several cell types including RBL and Jurkat T cells expressing stromal interaction molecule 1 (STIM1) and ORAI1 (Beck et al., 2014). In the Beck study, the cytosol of each assayed cell was acidified via direct introduction of NaOH or HCl via pipette. In the current study, the cytosol was acidified by exposure to TCS (a weak acid, pKa 7.9) via the extracellular buffer (Table 1), rather than by direct injection of strong acid or base (Beck et al., 2014). A stepwise effect, reduction of I_CRAC, occurred upon decreasing the pH from 7 to 6 (Beck et al., 2014). Similarly, TCS reduces Ca²⁺ flux into the cytosol of RBL and Jurkat cells (Weatherly et al., 2018) (Figure 9). Triclosan-mediated acidification of the cytosol may thus explain the previously observed inhibition of Ca²⁺ dynamics (Weatherly et al., 2018) and, thus, of degranulation (Holowka et al., 2012).

Another proton ionophore mitochondrial uncoupler, CCCP, also inhibits Ca²⁺ influx into RBL mast cells (Mohr and Fewtrell, 1987b). While this effect is partly caused by CCCP’s depolarization of PMP (Mohr and Fewtrell, 1987b), it can also be reversed by increasing the pH (i.e. alkalization) of the surroundings (Mohr and Fewtrell, 1987b) and, hence, of the cytosol because H⁺ becomes plasma membrane-permeant when CCCP is incorporated into the membrane (McLaughlin and Dilger, 1980). These findings suggest that CCCP inhibition of SOCE is partly caused by its pH modulation. Additionally, experiments on isolated neurons have shown that another proton ionophore mitochondrial uncoupler, FCCP, also acidifies the cytosol (Tretter et al., 1998). Therefore, TCS is acting as expected, from its proton ionophore mitotoxicant nature, in acidifying the immune cell cytosol and, then, inhibiting SOCE.

While under some conditions acidification of the cytoplasm can inhibit SOCE by blocking the binding of IP₃ to its receptor on the ER (Tsukioka et al., 1994), release of Ca²⁺ from the ER is actually enhanced in TCS-treated mast cells (Weatherly et al., 2018). Following anti-TCR activation of SOCE in T cells, the initial rise of Ca²⁺ release from the ER is largely unaffected by TCS whereas the plateau region representing SOCE via CRAC channels is heavily reduced by TCS (Figure 9), further evidence that IP₃ receptor interference is not the key mechanism of TCS inhibition in T cells, as well. Instead, it is likely that TCS-induced cytosol acidification blocks the proper interaction of the STIM and ORAI1 machinery that is required for CRAC channel opening (Thompson et al., 2009; Mancarella et al., 2011). Histidine 155, found between transmembrane domains (TM) 2 and 3 of ORAI1, plays an important role in sensing intracellular pH (Tsujikawa et al., 2015). Upon cytoplasmic acidification, histidine 155 becomes protonated and affects intermolecular interaction of other components in the loop between TM2 and TM3 of the ORAI1 that may result in CRAC channel closing (Tsujikawa et al., 2015). Based upon the Henderson-Hasselbalch equation and the pKa of histidine (6.0; Nelson and Cox 2017), 5.9% of these histidines are protonated at pH 7.2, whereas 11.2% are protonated at pH 6.9. (Note: if the protein environment surrounding histidine 155 changes its pKa, there could be a greater change in its charge upon acidification). Thus, the level of acidification induced by TCS (−0.3 pH unit, Table 1) likely induces a doubling of the concentration of ORAI1 proteins containing positively-charged histidines in this key cytoplasmic motif. Regardless of the mechanism of acidification (direct injection of strong acid as in Beck et al., 2014, or exposure to weak acid TCS dissolved in the extracellular buffer as in this study), the resulting acidification will cause this critical histidine to be more likely positively charged. This histidine is conserved in rat (as in RBL cells) and human (as in the Jurkat T cells), according to an NCBI Blast multiple amino acid sequence alignment. Overall, protonation and closing of CRAC channels is likely the mechanism of TCS-induced reduction of Ca²⁺ influx into mast cells (Weatherly et al., 2018) and T cells (Figure 9), leading to inhibition of mast cell function (Palmer et al., 2012). Future experiments utilizing the fluorescence resonance energy transfer techniques of Mancarella et al., 2011 or other approaches should be able to directly test the hypothesis that TCS disrupts the STIM1/ORAI1 interaction.

Future research will examine inhibition of T cell functions downstream of SOCE, such as release of essential cytokines (Punt et al., 2019). In addition to the known TCS inhibition of mast cell degranulation and other functions (Palmer et al., 2012), TCS also inhibits the lytic function of natural killer cells (Udoji et al., 2010; Hurd-Brown et al., 2013), which are important defenders against cancer and viral infections.

In addition to acidifying the cytosol by its direct proton ionophore mechanism of providing a pathway for charged protons to flow into the cell, TCS may also be acidifying the cell contents via other mechanisms. TCS increases the production of reactive oxygen species (ROS) in mast cells (Weatherly et al., 2018) and in other cell types (Binelli et al., 2009; Riva et al., 2012; Tamura et al., 2012; Yueh et al., 2014; Lv et al., 2016; Weatherly et al., 2018), which impair the Na/H⁺ exchanger, leading to reduced intracellular pH (Kaufman et al., 1993; Nakamura et al., 2006). TCS also causes mitochondrial fission/fragmentation (Weatherly et al., 2018), processes associated with reduced intracellular pH as a result of lactic acid buildup attributed to increased glycolysis (Johnson and Nehrke, 2010; Schurr, 2014). While the timeframes in which these processes acidify the cell are likely longer than the rapid (within 15 min) acidification reported in this study, these mechanisms suggest that TCS will continue to acidify the cell over longer exposure times and possibly at lower doses; ROS stimulation and mitochondrial dysfunction occur in primary human keratinocytes, mast cells, and other cell types at lower doses (starting ~1 μM) than the 10–20 μM used in the current study.

If this TCS acidification occurs in the mitochondria to a similar degree (~0.3 pH unit acidification of the matrix), the driving force (Gibbs free energy) available for producing ATP on ATP synthase in the inner mitochondrial membrane will be also reduced, as previously observed (Weatherly et al., 2016). This effect can be estimated by calculating the driving force both in the presence and absence of 20 μM TCS, via the equation

$${\mathrm{\Delta }G}_{\text{Transport}}=2.3\text{RT}\mathrm{\Delta }\text{pH}+\text{F}\mathrm{\Delta }\mathrm{\Psi }$$

(Nelson and Cox, 2017). Using the 37°C temperature used experimentally, the gas and Faraday’s constants, pH = −0.75, and ΔΨ = −0.2 V (Nelson and Cox, 2017), the free energy released by protons flowing through ATP synthase into the matrix is a robust ~−24 kJ/mol in a healthy cell. Taking into account a 0.3 pH unit acidification of the matrix (such that ΔpH = −0.45) and a 40% reduction in ΔΨ (Weatherly et al., 2018) due to 20 μM TCS exposure, the free energy released by protons flowing through ATP synthase into the matrix is a weaker ~−14 kJ/mol in a TCS-treated cell. This result implies a ~40% reduction in the free energy available for making ATP. While imperfect, this value is a reasonable match to the >50% reduction in ATP production caused by 20 μM TCS in RBL cells (Weatherly et al., 2016).

TCS inhibits MMP (Weatherly et al., 2018); however, TCS does not inhibit PMP. This apparent contradiction can be explained by the inherent cellular mechanisms in charge of creating and then maintaining either the MMP or the PMP. The MMP is primarily generated by the action of the electron transport chain (ETC) proton pumps (Nelson and Cox, 2017); MMP is entirely reliant on this segregation of protons. TCS is a proton ionophore mitochondrial uncoupler, thus, acting in direct opposition to the proton pumps of the ETC (Weatherly et al., 2016). Such direct opposition, with no alternative means of MMP maintenance, serves to explain why TCS can significantly depress the MMP. In contrast to the MMP’s generation by proton pumping, PMP is primarily generated by the action of the Na⁺/K⁺ ATPase, (Nelson and Cox, 2017), which is located nearly exclusively on the plasma membrane (Bertorello et al., 2003) and is not present on the mitochondrial membrane. In RBL mast cells, the Na⁺/K⁺ ATPase contributes to PMP (Bronner et al., 1989) and resides on the plasma membrane of RBL cells, and its inhibition leads to dampening of Ag-stimulated degranulation (Gentile and Skoner, 1996). Thus, TCS modulation of proton concentrations could affect cytosolic pH without altering PMP. TCS does cause PMP depolarization of artificial membranes which do not contain the Na⁺/K⁺ ATPase (Popova et al., 2018). TCS-mediated changes in proton concentrations across these membranes would therefore not be counteracted by the Na⁺/K⁺ ATPase. TCS also causes PMP depolarization of neuronal models (Arias-Cavieres et al., 2018; Popova et al., 2018) at micromolar doses within tens of minutes. Neurons heavily rely on the PMP for their function and, while the Na⁺/K⁺ ATPase is still the primary source of this voltage, neurons possess additional PMP regulation mechanisms (Bean, 2007), which may explain triclosan’s differential effects on PMP of immune cells vs. neuronal cells. Interestingly, TCS inhibits Na⁺/K⁺ ATPase activity in Labeo rohita gills (Hemalatha et al., 2019), a hint that, over longer exposure periods than those used in the current study, TCS may interfere with cellular PMP.

In conclusion, we report the mechanism of TCS inhibition of mast cells and T cells (Figure 10). Three-dimensional super-resolution microscopy shows that TCS causes mitochondrial swelling in mast cells, further evidence for its depolarization of the mitochondrial membrane. However, TCS does not inhibit immune cells by dampening their plasma membrane potential. Using genetically encoded voltage indicators coupled with pH-indicating reporters, we have identified that TCS acidifies the cytosol but does not affect PMP in immune cells. Cytosolic acidification by TCS likely disrupts the Stim1-ORAI1 interaction, causing CRAC channel closing and collapse of Ca²⁺ influx into mast cells and T cells. TCS-induced reduction of Ag-stimulated cytosolic Ca²⁺ influx results in inhibited enzymatic activity and decreased microtubule polymerization, leading to suppression of mast cell degranulation. Collapse of SOCE in T cells also likely leads to inhibition of T cell function. Any cell type that depends on Ca²⁺ signaling or mitochondrial function is susceptible to TCS toxicity (Feske, 2007; Hill-Eubanks et al., 2011). In summary, TCS, a mitochondrial toxicant, is also an immunotoxicant via its modulation of immune cell signal transduction.

Figure 10.

Effects of triclosan on mast cell signaling. Antigen crosslinking of IgE receptors, early phosphorylation events, and calcium release from the ER are uninhibited by TCS. However, TCS brings protons into the cytosol, decreasing cytosolic pH, thereby inhibiting CRAC channel activation, cytosolic calcium levels, microtubule polymerization, and, subsequently, degranulation.

• Antimicrobial triclosan inhibits mast cell and T cell signal transduction

• Three-dimensional super-resolution microscopy reveals mitochondrial swelling

• Triclosan inhibits crosslinker-stimulated Ca²⁺ influx into mast cells and T cells

• Triclosan does not affect plasma membrane potential in mast cells and T cells

• Triclosan acidifies the cytosol, thus impairing plasma membrane Ca²⁺ channels

Abbreviations:

Ag

antigen

APC

antigen presenting cell

ATP

adenosine triphosphate

AUC

area under the curve

BSA

bovine serum albumin

BT

Tyrodes-bovine serum albumin

CCCP

carbonyl cyanide 3-chlorophenylhydrazone

CRAC

Ca²⁺ release-activated Ca²⁺

DMSO

dimethyl sulfoxide

ER

endoplasmic reticulum

ETC

electron transport chain

FCCP

carbonyl cyanide-p-trifluoromethoxyphenylhydrazone

FPALM

fluorescence photoactivation localization microscopy

GEVI

genetically encoded voltage indicator

IgE

immunoglobulin E

IP₃

inositol 1,4,5-triphosphate

LDH

lactate dehydrogenase

MMP

mitochondrial membrane potential

PBS

phosphate buffered saline

PIP₂

phosphatidylinositol 4,5-bisphosphate

PLCγ

phospholipase C gamma

PLD

phospholipase D

PMP

plasma membrane potential

PSF

point spread function

RBL

rat basophilic leukemia cells, clone 2H3

ROI

region of interest

ROS

reactive oxygen species

SEM

standard error of the mean

SOCE

store-operated Ca²⁺ entry

STIM1

stromal interaction molecule 1

TCR

T-cell receptor

TCS

triclosan

TM

transmembrane domain