Significance:Combining genetically encoded neuronal activity indicators (GENIs), restricted to only one neuron type, with a voltage-sensitive dye (VSD) that reports pan-neuronal activity, could be beneficial for understanding neural circuits. Recently introduced far-red EF-630 may be compatible with green GENIs and could serve as an internal reference of neuronal activity for multiple GENIs. Aim:Here, we assess the EF-630 compatibility with several green-fluorescent-protein-based GENIs, including the calcium indicator GCaMP6f, the glutamate indicator iGluSnFR, and two voltage indicators ASAP2s, and ASAP5-Kv, for recording neuronal aggregate responses. Approach:Mouse brain slices expressing each GENI were stained with EF-630, and then extracellular stimulation and population optical imaging were sequentially performed at two wavelengths. In addition, cre-dependent ASAP5-Kv transgenic mice were generated and characterized. Results:Dual recordings provided population signals in both channels for all combinations. For each indicator pair, we quantified Δ F / F amplitudes and compared ON and OFF kinetics. ASAP2s and ASAP5-Kv displayed faster temporal dynamics and less temporal summation than VSD signals, suggesting the influence of cell-type-specific expression in observed kinetics. Conclusions:These kinetic differences underscore how both the choice of indicator pair and the targeted cell type influence the interpretation of neural population activity. Overall, our work provides the first systematic characterization of paired VSD-GENI measurements, establishing practical considerations and performance benchmarks for dual optical imaging of neuronal populations.

New voltage-sensitive ElectroFluor (EF) dyes that emit across the visible and near-infrared spectrum (e.g., 730 nm) were recently developed. We evaluated EF-530, EF-630, and EF-730p-dyes spectrally orthogonal to green fluorescent protein (GFP)-at excitation wavelengths outside the conventional 470 nm range used for GFP-based indicators. Although previously applied in cardiac voltage imaging, their performance in neuronal tissue remains untested. We performed side-by-side comparisons using population voltage imaging in mouse cerebral cortex slices at optimal excitation wavelengths (530, 630, and 730 nm) and assessed cross-channel signal bleed-through across four excitation wavelengths (475, 530, 630, and 730 nm). All dyes produced robust optical signals at their optimal wavelengths, though non-preferred channels exhibited bleed-through with distinct amplitudes, polarities, and photobleaching patterns. These results provide detailed quantifications of EF dye performance for neuronal population imaging.

We recently introduced a new class of voltage-sensitive dyes (VSDs) incorporating an amino chromene moiety, which to our knowledge also represents a completely new push-pull chromophore system. We substituted this group for the isostructural aminonaphthyl electron donor that has been part of many of our established VSD sensors. The aminochromene donor imparts an 80–100 nm red shift to the chromophores without any significant reduction in fluorescence quantum efficiency or voltage sensitivity. In this work, we introduce several new members of this class of VSDs with varying physical properties, to allow them to be applied to different biological preparations.

Cardiac optical mapping has traditionally been performed in ex-vivo, motion-arrested hearts. Recently, in-situ cardiac optical mapping has been made possible by both motion correction techniques and long-wavelength voltage sensitive dyes (VSDs). However, VSDs have been observed to wash out quickly from blood-perfused in-situ hearts. In this study, we evaluate the performance of a newly developed VSD, di-5-ANEQ(F)PTEA, relative to an earlier VSD, di-4-ANEQ(F)PTEA. We find that di-5-ANEQ(F)PTEA persists over 3 times longer, produces improved signal-to-noise ratio, and does not prolong loading unacceptably.Clinical Relevance—Optical mapping has provided many insights into cardiac arrhythmias, but has traditionally been limited to ex-vivo preparations. The present findings extend the utility of optical mapping in the more realistic in-vivo setting and may eventually enable its use in patients.

Voltage-sensitive dyes (VSDs) are used to image electrical activity in cells and tissues with submillisecond time resolution. Most of these fast sensors are constructed from push-pull chromophores whose fluorescence spectra are modulated by the electric field across the cell membrane. It was found that the substitution of naphthalene with chromene produces a 60 to 80 nm red-shift in absorption and emission spectra while maintaining fluorescence quantum efficiency and voltage sensitivity. One dye was applied to ex vivo murine heart with excitation at 730 nm, by far the longest wavelength reported in voltage imaging. This VSD resolves cardiac action potentials in single trials with 12% ΔF/F per action potential. The well-separated excitation spectra between these long-wavelength VSDs and channelrhodopsin (ChR2) enabled monitoring of action potential propagation in ChR2 hearts without any perturbation of electrical dynamics. Importantly, by employing spatially localized optogenetic manipulation, action potential dynamics can be assessed in an all-optical fashion with no artifact related to optical cross-talk between the reporter and actuator. These new environmentally sensitive chromene-based chromophores are also likely to have applications outside voltage imaging.

Optical mapping has been widely used in the study of cardiac electrophysiology in motion-arrested, ex vivo heart preparations. Recent developments in motion artifact mitigation techniques have made it possible to optically map beating ex vivo hearts, enabling the study of cardiac electromechanics using optical mapping. However, the ex vivo setting imposes limitations on optical mapping such as altered metabolic states, oversimplified mechanical loads, and the absence of neurohormonal regulation. In this study, we demonstrate optical electromechanical mapping in an in vivo heart preparation. Swine hearts were exposed via median sternotomy. Voltage-sensitive dye, either di-4-ANEQ(F)PTEA or di-5-ANEQ(F)PTEA, was injected into the left anterior descending artery. Fluorescence was excited by alternating green and amber light for excitation ratiometry. Cardiac motion during sinus and paced rhythm was tracked using a marker-based method. Motion tracking and excitation ratiometry successfully corrected most motion artifact in the membrane potential signal. Marker-based motion tracking also allowed simultaneous measurement of epicardial deformation. Reconstructed membrane potential and mechanical defor-mation measurements were validated using monophasic action potentials and sonomicrometry, respectively. Di-5-ANEQ(F) PTEA produced longer working time and higher signal/noise ratio than di-4-ANEQ(F)PTEA. In addition, we demonstrate potential applications of the new optical mapping system including electromechanical mapping during vagal nerve stimulation, fibrillation/defibrillation. and acute regional ischemia. In conclusion, although some technical limitations remain, optical mapping experiments that simultaneously image electrical and mechanical function can be conducted in beating, in vivo hearts.

Cancer cells feature a resting membrane potential ( V m ) that is depolarized compared to normal cells, and express active ionic conductances, which factor directly in their pathophysiological behavior. Despite similarities to ‘excitable’ tissues, relatively little is known about cancer cell V m dynamics. Here high-throughput, cellular-resolution V m imaging reveals that V m fluctuates dynamically in several breast cancer cell lines compared to non-cancerous MCF-10A cells. We characterize V m fluctuations of hundreds of human triple-negative breast cancer MDA-MB-231 cells. By quantifying their Dynamic Electrical Signatures (DESs) through an unsupervised machine-learning protocol, we identify four classes ranging from "noisy” to “blinking/waving“. The V m of MDA-MB-231 cells exhibits spontaneous, transient hyperpolarizations inhibited by the voltage-gated sodium channel blocker tetrodotoxin, and by calcium-activated potassium channel inhibitors apamin and iberiotoxin. The V m of MCF-10A cells is comparatively static, but fluctuations increase following treatment with transforming growth factor- β 1, a canonical inducer of the epithelial-to-mesenchymal transition. These data suggest that the ability to generate V m fluctuations may be a property of hybrid epithelial-mesenchymal cells or those originated from luminal progenitors.

ABSTRACTCancer cells feature a resting membrane potential (Vm) that is depolarized compared to normal cells, and express active ionic conductances, which factor directly in their pathophysiological behavior. Despite similarities to ‘excitable’ tissues, relatively little is known about cancer cell Vm dynamics. With high-throughput, cellular-resolution Vm imaging, we characterized Vm fluctuations of hundreds of human triple-negative breast cancer MDA-MB-231 cells and compared to non-cancerous breast epithelial MCF-10A cells. By quantifying their Dynamic Electrical Signatures (DESs) through an unsupervised machine-learning protocol, we identified four classes ranging from “noisy” to “blinking/waving”. The Vm of MDA-MB-231 cells exhibited spontaneous, transient hyperpolarizations that were inhibited by the voltage-gated sodium channel blocker tetrodotoxin. The Vm of MCF-10A cells was comparatively static, but fluctuations increased following treatment with transforming growth factor-β1, a canonical inducer of the epithelial-to-mesenchymal transition. These data suggest that the ability to generate Vm fluctuations is acquired during transformation and may participate in oncogenesis.

The 2-component FRET-based voltage sensing mechanism was reported two decades ago by the late Roger Tsien and his colleagues. It involves a FRET pair consisting of an immobile chromophore that stains one side of the membrane and a membrane-permeant anionic chromophore that translocates across the membrane in a voltage dependent manner. However, the membrane permeant anion component has been limited to only 2 dyes: bis-(1,3-dialkyl-2-thiobarbiturate)-trimethineoxonol (DiSBACn(3)) and dipicrylamine (DPA). We set out to design additional dyes that are anionic at physiological pH, are rapidly membrane permeant, sample a variety of spectral regions and are photostable. This was achieved via structural modification of fluorescein, DiSBAC, and JPW1132. We characterized the voltage sensitivity in artificial membranes with di-4-ANEPPS as the immobile FRET component. Several of our new anions displayed high voltage sensitivity with fast kinetics in response to voltage steps. We anticipate these new chromophores can also be valuable as a basis for the design of other fluorescent sensors and labels. (Supported by NIH grants EY030712 and MH116830. Conflict of Interest statement: LML, CDA, and PY are founders and owners of Potentiometric Probes LLC.)

Differentiation of human induced pluripotent stem cells (iPSCs) can be used to develop assays for therapeutics and their off target effects. Differentiated iPSCs also provide a means to study mechanisms of genetic disorders that are patient specific and an avenue for the development and application of “personalized medicine”. iPSC-derived electrically excitable cell cultures including neurons as well as cardiomyocytes (iPSC-CMs), which beat spontaneously in the dish, have been reported. Voltage-sensitive dyes (VSDs) are a powerful tool to study these cultures and offer important advantages for assay design. To make such recordings as sensitive, robust, and informative as possible, we are working to optimize electrochromic VSDs, using in-vitro test assays, as well as mouse and human cell lines. Combined patch clamp and imaging studies can derive the optimal imaging parameters for maximum signal-to-noise. The fluorinated VSD Di-4-AN(F)EP(F)PTEA was especially sensitive in neuronal or cardiac culture systems. We used it to characterize performance and possible “toxicity” effects including changes in spontaneous firing, and altered AP widths, or waveforms. For testing in human cells, iPSCs were differentiated to iPSC-CMs using sequential activation and inhibition of the Wnt signaling pathway. Spontaneously beating cells were imaged using an inexpensive, global reset CMOS camera synchronized to a Lumincor SpectraX LED light source for dual excitation wavelength ratiometric imaging. Ratio imaging eliminated motion and uneven staining artifacts to permit long-term high fidelity recordings of APs. Single sweep optical AP traces were obtained at 200 frames/s with signal-to-noise as high as 20, which allows recording of asynchronous APs necessary to study neuronal network activity. Disclosures: CA, PY and LML are cofounders of Potentiometric Probes, which develops and distributes voltage-sensitive dyes. (Supported by NIH grants R43MH116830, R01HL142787 and R43MH121236.)

The first workshop on Novel Optics-based approaches for Cardiac Electrophysiology (NOtiCE) was held in Florence Italy in 2018. Here, we learned how optical approaches have shaped our basic understanding of cardiac electrophysiology and how new technologies and approaches are being developed and validated to advance the field. Several technologies are being developed that may one day allow for new clinical approaches for diagnosing cardiac disorders and possibly intervening to treat human patients. In this review, we discuss several technologies and approaches to optical voltage imaging with voltage-sensitive dyes. We highlight the development and application of fluorinated and long wavelength voltage-sensitive dyes. These optical voltage sensors have now been applied and well validated in several different assays from cultured human stem cell-derived cardiomyocytes to whole hearts in-vivo. Imaging concepts such as dual wavelength ratiometric techniques, which are crucial to maximizing the information from optical sensors by increasing the useful signal and eliminating noise and artifacts, are presented. Finally, novel voltage sensors including photoacoustic voltage-sensitive dyes, their current capabilities and potential advantages, are introduced.

Voltage sensitive dyes (VSDs) are used for in vitro drug screening and for in vivo imaging of patterns of electrical activity. However, wide application of this technology is limited by poor sensitivity. A promising approach uses a 2-component system consisting of charged membrane permeable quenchers together with fluorophores labeling one side of the membrane; this produces voltage-dependent fluorescence quenching. However, to achieve good sensitivity, the quencher compound must be used at high concentrations, which can perturb the membrane capacitance or have other pharmacological effects. By developing tethered bichromophoric fluorophore quencher (TBFQ) dyes, where the fluorophore and quencher are covalently connected by a long hydrophobic chain, the concentration required is minimized, and the sensitivity is maximized. A series of 13 TBFQ dyes based on the AminoNaphthylEthenylPyridinium (ANEP) fluorophore and the dipicrylamine anion (DPA) quencher have been synthesized and tested in an artificial lipid bilayer apparatus. The best one from the screening, TBFQ1, shows a 2.5 fold change in fluorescence per 100mV change in membrane potential, and the response kinetics is in 10-20 ms range. This sensitivity is an order of magnitude better than commonly used fluorescent voltage sensors. The design principles for TBFQ VSDs described here can be readily extended to other spectral regions and promise to greatly enhance our ability to monitor electrical activity in cells and tissues.

In cortical pyramidal neurons, backpropagating action potentials (bAPs) supply Ca2+ to synaptic contacts on dendrites. To determine whether the efficacy of AP backpropagation into apical tuft dendrites is stable over time, we performed dendritic Ca2+ and voltage imaging in rat brain slices. We found that the amplitude of bAP-Ca2+ in apical tuft branches was unstable, given that it varied from trial to trial (termed "bAP-Ca2+ flickering"). Small perturbations in dendritic physiology, such as spontaneous synaptic inputs, channel inactivation, or temperature-induced changes in channel kinetics, can cause bAP flickering. In the tuft branches, the density of Na+ and K+ channels was sufficient to support local initiation of fast spikelets by glutamate iontophoresis. We quantified the time delay between the somatic AP burst and the peak of dendritic Ca2+ transient in the apical tuft, because this delay is important for induction of spike-timing dependent plasticity. Depending on the frequency of the somatic AP triplets, Ca2+ signals peaked in the apical tuft 20-50 ms after the 1st AP in the soma. Interestingly, at low frequency (<20 Hz), the Ca2+ peaked sooner than at high frequency, because only the 1st AP invaded tuft. Activation of dendritic voltage-gated Ca2+ channels is sensitive to the duration of the dendritic voltage transient. In apical tuft branches, small changes in the duration of bAP voltage waveforms cause disproportionately large increases in dendritic Ca2+ influx (bAP-Ca2+ flickering). The stochastic nature of bAP-Ca2+ adds a new perspective on the mechanisms by which pyramidal neurons combine inputs arriving at different cortical layers.NEW & NOTEWORTHY The bAP-Ca2+ signal amplitudes in some apical tuft branches randomly vary from moment to moment. In repetitive measurements, successful AP invasions are followed by complete failures. Passive spread of voltage from the apical trunk into the tuft occasionally reaches the threshold for local Na+ spike, resulting in stronger Ca2+ influx. During a burst of three somatic APs, the peak of dendritic Ca2+ in the apical tuft occurs with a delay of 20-50 ms depending on AP frequency.

Intracellular voltage-sensitive dyes are used to monitor membrane potential changes from neuronal compartments not readily accessible to glass electrodes, such as basal dendritic segments more than 140 mu m away from the cell body. Optical imaging is uniquely suitable to reveal voltage transients occurring simultaneously in two or more dendritic branches, or in two or more locations along the same dendritic branch (simultaneous multi-site recordings). Voltage-sensitive dye recordings can be combined with bath application of drugs that block membrane conductances as well as with focal application of neurotransmitters. The results of dendritic voltage-sensitive dye measurements are naturally incorporated into computational models of neurons with complex dendritic trees. The number of model constraints is notably heightened by a multi-site approach. An interaction between multi-site voltage-sensitive dye recording (wet experiment) and multicompartmental modeling (dry experiment) constitutes one of the most insightful combinations in quantitative neurobiology. This chapter discloses disadvantages associated with voltage-sensitive dyes. It brings useful information for deciding whether voltage-sensitive dye imaging is an appropriate method for your experimental question, and how to determine if a student is ready to work with intracellular voltage-sensitive dyes. Our chapter describes the most important, previously unpublished, practical issues of loading neurons with voltage-sensitive dyes and obtaining fast optical signals (action potentials) from thin dendritic branches using equipment at half price of a standard confocal microscope.

Dendritic spines are specialized structures in the brain that receive excitatory synaptic inputs. Activation of synaptic receptors on spines leads to current influx that depolarizes the activated spine, the parent dendrite and the neuron's somatic and axonal compartments. Local membrane potential fluctuations affect voltage-dependent channels producing calcium influx in an interplay of electrical and biochemical signaling. We combine 2-photon glutamate uncaging with 2-photon voltage-sensitive dye (VSD) imaging to study changes in membrane potential in individual spines. We compare two methods of calibrating these signals in spines that may be spatially distant from the somatic compartment. Subthreshold calibration signals have limited amplitude, and actual amplitudes can fall short of the expected values due to patch clamp limitations and space clamp effects, both leading to over-estimated EPSP amplitudes. Backpropagating action potentials have large amplitude but an average attenuation decay rate must be incorporated leading to some uncertainty. We measure the small but significant nonlinearity in dye response that would lead to under-estimated EPSP amplitudes if not accounted for. Calibrated evoked spine EPSP (spEPSP) amplitudes, along with somatic EPSP (soEPSP) amplitudes are input to a compartmental model (NEURON) to predict spine neck resistance (Rneck) values that are consistent with our measurements. While highly variable, predicted Rneck is centered on 200MOhm and is consistent with independent estimates based on FRAP experiments using a cytosolic dye in the absence of the VSD.

Abstract EPSPs occur when the neurotransmitter glutamate binds to postsynaptic receptors located on small pleomorphic membrane protrusions called dendritic spines. To transmit the synaptic signal, these potentials must travel through the spine neck and the dendritic tree to reach the soma. Due to their small size, the electrical behavior of spines and their ability to compartmentalize electrical signals has been very difficult to assess experimentally. In this study, we developed a method to perform simultaneous two-photon voltage-sensitive dye recording with two-photon glutamate uncaging in order to measure the characteristics (amplitude and duration) of uncaging-evoked EPSPs in single spines on the basal dendrites of L5 pyramidal neurons in acute brain slices from CD1 control mice. We were able to record uncaging-evoked spine potentials that resembled miniature EPSPs at the soma from a wide range of spine morphologies. In proximal spines, these potentials averaged 13.0 mV (range, 6.5–30.8 mV; N = 20) for an average somatic EPSP of 0.59 mV, whereas the mean attenuation ratio (spine/soma) was found to be 25.3. Durations of spine EPSP waveforms were found to be 11.7 ms on average. Modeling studies demonstrate the important role that spine neck resistance (Rneck) plays in spine EPSP amplitudes. Simulations used to estimate Rneck by fits to voltage-sensitive dye measurements produced a mean of 179 MΩ (range, 23–420 MΩ; N = 19). Independent measurements based on fluorescence recovery after photobleaching of a cytosolic dye from spines of the same population of neurons produced a mean Rneck estimate of 204 MΩ (range, 52–521 MΩ; N = 34).

Our lab has been developing and improving the optical and molecular tools and techniques to record EPSPs (excitatory postsynaptic potentials) in dendritic spines. With a dual laser 2-photon microscope we are able to record from voltage-sensitive dye (VSD) labeled spines from cortical pyramidal neurons while evoking unitary EPSPs using glutamate uncaging. By varying the uncaging laser intensity we observe a nonlinear response to glutamate when the resulting EPSPs amplitudes are observed at the soma. By optically recording EPSPs in spines we can determine that the observed nonlinearity is often present even when spine EPSPs are relatively small, i.e. < 10 mV. Pharmacology is used to block NMDA receptors and determine the recruitment of these receptors and their contribution to both voltage and calcium influx. Calcium imaging reveals that NMDA receptors are recruited in a very linear fashion according to the spine depolarization. This calcium influx however, does not appear to play a significant role in the spine depolarization. Finally, modeling is used to reconcile the observed phenomena with detailed models of AMPA and NMDA receptor gating, and glutamate binding.

We developed a 2-photon imaging system that allows us to simultaneously perform glutamate uncaging and voltage sensitive dye (VSD) imaging from dendritic spines, which are the postsynaptic targets of excitatory inputs. Using this system and the intracellular fluorinated VSD di-2-AN(F)EPPTEA (Yan et al, 2012), we were able to record MNI-glutamate uncaging evoked excitatory postsynaptic potentials (EPSPs) from single spine heads that resemble mEPSPs (<1 mV at the soma) in the basal dendrites of LV pyramidal neurons. Our data suggests that in the spine heads, these EPSPs do not exceed more than 25 mV, and are attenuated by a mean factor of ∼20 when they reach the soma. Interestingly, we have found no correlation between the EPSP amplitude in the spines and at the soma, which could be due to variability in the spine neck resistances. Based on this, we combined these experiments with FRAP of Alexa488, in order to estimate the spine neck resistance from the spines where we measured the EPSPs amplitudes. The time constant of equilibration of the cytosolic Alexa488, combined with the spine head volume, can be used to estimate the spine neck resistance. We used an image processing algorithm to determine the spine head volume from the 3D fluorescence distribution of Alexa488; from this, we have estimated spine neck resistances that range between 75-500 MΩ. By measuring the EPSP from a neighboring spine, we show experimentally that a spine with a neck resistance of ∼75 MΩ does not fully compartmentalize the EPSP, and it can be seen by a neighboring spine ∼5 um away with an amplitude > 5mV. NIH grants R01 EB001963, P41 GM103313.

In this protocol, we describe the procedures we have developed to optimize the performance of voltage-sensitive dyes for recording changes in neuronal electrical activity. We emphasize our experience in finding the best dye conditions for recording backpropagating action potentials from individual dendritic spines in a neuron within a brain slice. We fully describe procedures for loading the dye through a patch pipette and for finding excitation and emission wavelengths for the best sensitivity of the fluorescence signal to membrane voltage. Many of these approaches can be adapted to in vivo preparations and to experiments on mapping brain activity via optical recording.