Optogenetic actuators are light-sensitive proteins that change their conformation and activity when stimulated by a particular wavelength of light. Since light can be controlled with sub-micrometer, and sub-millisecond precision this toolbox delivers high spatiotemporal selectivity. As the tools are genetically encoded, they can be targeted within cells by signal tags, or restricted to certain cell-types by promoter controls. They, therefore, complement the optical probes used to visualize intracellular physiology discussed above. Since they utilize the same microscopy infrastructure simultaneous experiments to control and observe responses in single cells can be delivered on a time and space scale unimaginable with other techniques.
Although single cell applications of optogenetic tools in other systems have been used to activate a variety of transmembrane receptor signaling pathways, activate cyclic nucleotide second messenger systems, regulate transcriptional and/or even epigenetic states [117
], this extended toolbox is currently underutilized in cardiac cells. To date, cardiovascular applications have almost exclusively used optically-controlled ion channels to depolarize (e.g., Channelrhodopsin, ChR2, a class I bacterial opsin isolated from Chlamydomonas reinhardtii
which requires light in the blue-green spectrum) the cell and thereby activate contraction. Typically, these have been undertaken in whole animals [118
], explanted tissues [119
], or 2D cultures [120
] as the field investigates alternatives to traditional pacing, or pharmacotherapy to control heart rate or rhythm [121
]. Hyperpolarizing tools (e.g., NpHR, the chloride pump Halorhodopsin isolated from Natronomonas pharaonis
which requires light in the green-red spectrum) are also described but, to date, have not been widely used in cardiac applications. A schematic of current activating, and inactivating optogenetic tools is presented in Figure 3
The tandem cell unit strategy overcomes this [123
] as the transgene (in this case ChR2) is stably expressed in an immortal cell line which can be co-cultured with cardiomyocytes to allow coupling between the two cell types. Alternative cell autonomous strategies for primary ventricular cardiomyocytes include transgenic ChR2 mice [124
], in vivo myocardial injection of adenoviral expression systems with cardiomyocyte isolation after a week for mice [68
], or in vitro infection for the relatively long-lived vGPCM [68
]. vGPCM is perhaps the best rodent model of the physiological system in humans, demonstrating in vitro viability up to 72 h in our hands. A lentiviral gene delivery strategy of a more light-sensitive, and more divalent cation-selective ChR2 variant [125
], has been demonstrated in the neonatal atrial cardiomyocyte model [126
In summary, triggered depolarization of virtually all contemporary biological models of the heart/cardiomyocyte can be successfully achieved using a variety of gene delivery methods, and a number of ChR2 variants. Studies comparing the complete spectrum of ChR2 variants in particular cardiomyocyte models are lacking, therefore, currently there is no dominant strategy for a given system. At the single cell level, optical pacing has been demonstrated in iPSc/ESC cardiomyocytes, neonatal cardiomyocytes, and adult cardiomyocytes.
Like any toolbox with a decade of enabling research there are inevitable pitfalls for those wishing to adopt these approaches without guidance. Since the greatest immediate application of optogenetics in cardiac research appears to be the capability to replace traditional electrical stimulation, we will focus on the considerations relevant to excitatory (AP generating) tools, the prototype for which is ChR2. ChR2 turns on (<5 ms) and off (<50 ms) quickly enough [127
] to make it a suitable tool for cardiomyocyte stimulation at frequencies up to 10 Hz. Subsequent protein engineering efforts have improved key parameters, including light sensitivity, and speeds of activation and inactivation, in addition to the magnitude and ion selectivity of the currents evoked [68
]. Community resources to catalogue developments and guide the researcher have been proactively developed in this field, e.g., openoptogenetics.org or the optogenetics resource center maintained by the Deisseroth lab. To ensure that as projects are devised/funded the most appropriate tool is utilized and any caveats anticipated before peer review exposes errors, periodic attention to these repositories is encouraged. A Twitter bot for community questions (@Optoquestion) has also been established.
Considerations for Optogenetic Application in Cardiomyocytes
Successful optical stimulation requires a light-induced current large enough to trigger depolarization. This depends on many experimental variables: gene expression, power density, and wavelength of stimulating light energy, and in the case of syncytial approaches, the magnitude of conductance between the cell expressing the optical tool and the cardiomyocyte target. Considerations for the successful application of ChR2-based stimulation include:
Since transfection of ventricular or stem cell-derived cardiomyocytes is generally a thankless task, viral delivery systems are needed. Where control and phenotyping tools are desired, the compact size of ChR2 (737 amino acids, 77 kDa) means it can be expressed with a 2A peptide and a phenotyping tool, particularly if the typical YFP fusion used to follow ChR2 infection and membrane localization is replaced by a small peptide tag, like the myc epitope [68
]. Since gene expression, and auto-catalytic 2A peptide cleavage take time, we usually do not try to run an experiment until 48 h post infection. Unfortunately, the more mature the cardiomyocyte is the harder it is to infect. Typically, we need multiplicities of infection four orders of magnitude more for adult vGPCM compared to iPS-CMs [30
]. Transient transfection typically generates low μM expression levels [60
]. By contrast knock-in approaches with endogenous promoters may be three orders of magnitude lower. This may ultimately prove insufficient or run the risk of altering cell behavior due to insertional mutagenesis. Stable long lived RNA’s of ChR2 are commercialized by Ncardia as an alternative to viral delivery for iPSc cardiomyocytes to circumvent this issue.
Excitation (light sensitivity):
ChR2 has a broad excitation spectrum that extends from UV to green parts of the visible spectrum [121
], with optimal stimulation at 480 nm. Generally, this forces the use of phenotyping tools with excitation/emission into the red end of the visible spectrum to prevent inadvertent activation of ChR2 (see cross talk
, below). The power density of stimulating light is an important, and often ignored, parameter. In simple terms low gene expression can be offset by application of more light (either by higher power, longer duration, or wider area) to boost the probability that sufficient channels will open and depolarize the cell. Full activation of ChR2 requires 1018
at 480 nm [125
]. Typical power densities for single cell application in the range of 0.1 to 10 mW/mm2
10 ms, which can be generated by most light-sources equipped on standard microscopes. The choice of light source should depend on the needs of the user rather than what is available. For example, the advantage of coherent light sources (lasers) is that confocal platforms are designed to deliver these with subcellular precision so individual cells can be targeted within a field, or even regions within a single cell [68
], but lasers produce higher power than LED sources. This can reduce viability and experimental duration, particularly if confocal imaging is employed during phenotyping, as this is both high power and slow. Our solution to this was to develop a microscope using the lower power of LED light combined with the high sensitivity and speed of the electron-multiplying charge-coupled device (EMCCD) camera to provide continuous imaging while a laser power source can operate independently to control the cell on millisecond and subcellular space scales [68
Alternative solutions to the problem are described; for example, the EMCCD chip architecture limits the speed of imaging due to the dead-time needed to discharge information from the detector. Since this is longer (~40 ms) than the time needed to stimulate the optical control tool, a very elegant solution from the Nagai lab has been to hide the optical stimulation pulse inside the camera dead-time [128
] so that imaging is not interrupted by stimulation at all.
The dependence on relatively high light power requirements for ChR2 activation can be exploited. As outlined above, the power-density of bioluminescent probe emissions is low, typically >5 orders of magnitude (~nW/mm2
range) less than that needed to activate ChR2. This means that bioluminescent probes cannot inadvertently activate the optical control tool even though they may occupy the same part of the visible spectrum. As bioluminescence eliminates the need for extrinsic light for imaging, and by virtue of low-power light emission liberating the full range of the visible spectrum for phenotyping tools, they should become increasingly preferred as the brightness of bioluminescent tools improves allowing production of calcium [87
] and voltage indicators [95
]. However, caution may be required if the tandem cell unit strategy is used for activation as minimal irradiance needed in that system was 6 μW/mm2
Crosstalk can be minimized by the choice of reporter, reduction of light power intensity, or spatial restriction of excitation light in multicellular applications. Indicator strategies compatible with ChR2 are highlighted within this article [42
]; but combining optical stimulation with imaging probes or proteins needs a little more consideration than multi-color imaging for two reasons. Firstly, ChR2 has a broad excitation spectrum of 400–520 nm (peak 480 nm). In general, it is recommended to choose indicators with an excitation spectrum as far from ChR2 as possible to avoid inadvertent activation of the cell during dynamic imaging. If this happens the measured signal, the intensity will drift upwards [130
], whereas typically slow dye photobleaching should be expected to cause signal baseline intensity to drift downwards. Secondly, the user must be conscious of the potential for the light used for ChR2 stimulation to cause photoactivation or photoconversion of the reporter probe. Many of the current genetically encoded red indicators that give rise to the RCAMP’s hail from proteins with complex excitation spectra with chromophore maturation through non-red intermediates. Residual properties of the parental protein can be retained in its progeny, but only unmasked by application of high energy blue:green light. This complicates the application of these tools with optogenetics. Furthermore, many of the leading fluorescent GEVI’s appear to have undesirable properties when combined with optical control strategies [95
When mixed cell preparations are used, conductance for successful depolarization between the ChR2-expressing activating cell and the cardiomyocyte must exceed a minimum value of about 2 nS, similar to that when the activating cell expresses HCN2, the channel that normally produces the physiological If
current underlying biological pacing [131
]. Since coupling depends on the physical contacts established between two cell types it will be influenced by the methods to isolate and singularize cells (particularly the extracellular proteases). This process may be additionally vulnerable to how cells are handled following isolation, for example, shaking, or differential rates of sedimentation of large cardiomyocytes compared to smaller spark cells. Fortunately, as depolarization is an all-or-none phenomenon this approach appears not to be exquisitely sensitive to the ratio of cell types present.
The active chromophore in ChR2 is all-trans-retinal which is produced from vitamin A. In cultured cells it is expected that this co-factor may become limiting, encouraging some to supplement media accordingly [132
]. This may not be without biological consequence for stem cell-derived models in particular.
ChR2 exhibits inward monovalent cation selectivity (H+
) with minimal outward current. Its reversal potential is close to 0 mV. Single-channel conductance estimates have been in the 0.04–2 pS range [133
] approximately one tenth (or less) of the sodium channel. This may make the tool unsuited to traditional single cardiomyocyte experiments that are conducted in Na+
Although not widely reported ChR2 has a voltage dependent activation and inactivation. At more negative potentials (−80 mV), currents are 10 times higher than at −20 mV [123
]. This factor can become limiting in iPSc cardiomyocyte populations where historical methods of differentiation generate cells with resting membrane potentials in a range from −40 mV to −60 mV [134
] perhaps due to the relative lack of dominant rectifier current in humans (IK1
). This, therefore, will make certain cells within iPSc cardiomyocyte differentiations harder to excite than others.
ChR2 can exist in a lower conductance, light-adapted, state meaning that chronic fast optical pacing trains may cause partial inactivation of ChR2. Experimentally, 100 s of 2 Hz stimulation using 10 ms of 480 nm light in ventricular myocytes leads to a 22% reduction in current [134
ChR2 conductance can be affected by temperature where, in general, higher temperatures lead to larger currents [127
]. Although the energy in electrical pacing may heat the perfusion media; no such practical concerns are apparent for light-based stimulation where typical irradiances in the low mW/mm2
range are theoretically predicted to have a minimal temperature change (0.013 °C for a 5.5 mW/mm2
pulse for 90 ms) [135