Optogenetic Stimulation Array for Confocal Microscopy Fast Transient Monitoring

Optogenetics is an emerging discipline with multiple applications in neuroscience, allowing to study neuronal pathways or serving for therapeutic applications such as in the treatment of anxiety disorder, autism spectrum disorders (ASDs), or Parkinson's disease. More recently optogenetics is opening its way also to stem cell-based therapeutic applications for neuronal regeneration after stroke or spinal cord injury. The results of optogenetic stimulation are usually evaluated by immunofluorescence or flow cytometry, and the observation of transient responses after stimulation, as in cardiac electrophysiology studies, by optical microscopy. However, certain phenomena, such as the ultra-fast calcium waves acquisition upon simultaneous optogenetics, are beyond the scope of current instrumentation, since they require higher image resolution in real-time, employing for instance time-lapse confocal microscopy. Therefore, in this work, an optogenetic stimulation matrix controllable from a graphical user interface has been developed for its use with a standard 24-well plate for an inverted confocal microscope use and validated by using a photoactivable adenyl cyclase (bPAC) overexpressed in rat fetal cortical neurons and the consequent calcium waves propagation upon 100 ms pulsed blue light stimulation.


I. INTRODUCTION
O PTOGENETICS is a set of strategies that includes a way to excite or inhibit neural cells modified to artificially express light-gated ion channels but also to activate other lightsensitive proteins [1] developed during the last two decades. This set of tools follows the emergence of protein-coding opsin genes that respond to illumination at certain optical wavelengths. From this point, optogenetic techniques have seen increasing development and interest, being chosen in 2010 as Method of the Year by Science [2]. This neuromodulation method makes it possible, among other things, to control the behavior of certain cells present in a system with a level of precision never seen and event triggering in a non-invasive way and with a high temporal resolution.
The response of these biosystems is mainly characterized by triggering events after the stimulus and hyperpolarizing and depolarizing currents [3]. We recently showed that blue-light stimulation of neural precursor cells (NPC) engineered to artificially express the excitatory light-sensitive protein channelrhodopsin-2 (ChR2-NPCs) prompted an influx of cations and a subsequent increase in proliferation and differentiation into oligodendrocytes and neurons and the polarization of astrocytes from a pro-inflammatory phenotype to a pro-regenerative/ anti-inflammatory phenotype [4], therefore, prompting a more pro-regenerative profile for further NPC therapeutic applications such as spinal cord injury or stroke [5], [6].
Hand in hand with the development of the field, electronic equipment has been emerging to perform ex vivo experimentation, generally with highly specific characteristics in terms of temporal parameters and simultaneous experimentation (i.e., the application of different protocols to different wells in the same well matrix). The use of laser [5] and LED [7] aimed over the culture dish through an optic fiber are the simplest setups.
In the early days of this discipline, an astable NE555-based optostimulation platform was developed for in vitro application [8], offering certain configuration capabilities through a hardware potentiometer and stating the future development of the next generation optostimulation platforms based on microcontroller. In [9] a data acquisition card is used to control a LED source from a PC. The setup is assembled to allow This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see https://creativecommons.org/licenses/by/4.0/ the simultaneous monitoring of the cardiac electrophysiology perturbation induced by the optogenetic stimulation through an optic microscope.
Nowadays, there are some devices available in the market that offer great flexibility capabilities in terms of wavelength and configuration of the stimulus sequences: The Light Plate Apparatus (LPA) [10] is an open-hardware optogenetics and photobiology stimulation device, with instructions for its assembly and calibration, and offers a great deal of configuration flexibility through the LED selection from a calibrated list and a web-based configuration Graphic User Interface. The LPA is configurated for a standard 24-well plate, displaying up to 2 LEDs per well.
The optoPlate-96 [11] displays a geometrical conformation to insert in the standard 96-well and 384-well plates. Each cell is configured to be able to stimulate with far-red, red, and blue lights. A Graphic User Interface has been added to the system [12].
Nevertheless, both outstanding optostimulation platforms in the state of art are developed to fit over them the well-plates limiting the optical resolution for high magnification in vivo assays. Then, the assessment of the results is performed not in real-time but afterward by harvesting the cells for flow cytometry or immunofluorescence analysis. Here we have developed a platform that allows real-time high-resolution microscopy acquisition at the time of stimulation as indicated for fast calcium waves imaging evaluation with requires the use of higher resolution instrumentation, such as confocal microscopy.
Calcium imaging has been extensively used to interrogate cellular activity determined by a calcium dependent-signaling thanks to the use of newly developed fluorescent calcium indicators such as Fluo-4 [13]. Calcium imaging employing such as fluorescent sensors is monitored by real-time confocal acquisition from NPCs or mature neuronal cells to evaluate their functional responses since intracellular calcium plays critical roles, for instance, in cell cycle control, cell death or gene expression [14]. Calcium imaging in mature neuronal cultures serve i.e., as an indicator of synaptic plasticity in post-synaptic in a neuronal network [15].

II. SYSTEM OVERVIEW
The problem addressed in this work is the development of a tool for in-vitro optogenetic stimulation while allowing the simultaneous monitoring of the induced fast transient responses. For this purpose, the Optogenetic Stimulation for In Vitro Experimentation (OSIVE Slim) is designed to fit in a commercial confocal microscope, which will be used to obtain such real-time feedback. A 24-well plate it's chosen in order to maximize the experimentation throughput: a total of eight different experimentation groups, with three replicates each, can be seeded, stimulated, and monitored simultaneously. The optogenetic stimulation array, with 24 photoemitters matching the geometrical configuration of the well plate array, will be configured and controlled through a USB interface with a PC. The general block diagram of the system is shown in Fig. 1.
The inverted confocal microscope Leica TSP-SP8 (in service in the Centro de Investigación Príncipe Felipe (CIPF) Confocal Microscopy Service) was employed for real-time acquisition and therefore, served for device configuration design.
A synthetic excitable cell model was generated by a genetic modification of primary rat cortical neurons to induce artificial expression of a photo-stimulated adenyl-cyclase (bPAC) able to respond to brief pulses of blue light. Previous reports using hippocampal neurons in vitro have shown that in response to saturating light pulses bPACs induce large photocurrents sufficient to induce action potential firing neurons. bPAC-induced currents peaked in the order of in the order of hundreds of milliseconds (bPAC, one-half peak after 723 ± 101 ms) [16].
bPAC increases intracellular cyclic adenosine monophosphate (cAMP) [17] and in turn, the cAMP opens a cAMP-gated ion channel and whole-cell Ca 2+ transients and voltage changes Optogenetic Stimulation for In-Vitro Experimentation (OSIVE), Slim version, over inverted confocal microscope setup. OSIVE Slim fits inside the space between the P-24 in-vitro well array and the microscope's enclosure, allowing the real-time measurement of the fast transients after a stimulation shot or burst. 1 OSIVE Slim, 2 laser safety lid, 3 P-24 well array, 4 objective lens, 5 color splitter, 6 laser array (collimated excitement pinhole), 7 adjustable prism, and 8 detectors. [18]. This system then serves as a platform for calcium sensors evaluation upon optogenetics stimulation. Monolayer, adherent cortical neuronal cultures were monitored as shown below in Fig. 2. Hence, the optoelectronic system was placed over the plate, so the result of irradiation can be measured at the moment of application. This leaves a constraint in the size and form design of the system to fit the microscope chamber in the remaining 15 mm height. Satisfying this constraint will allow the enclosure of the microscope's laser safety lid. A limitation to bear in mind is regarding the cables that need to be brought outside from this chamber.
As per the bibliography visited, the minimum and maximum irradiance power selected for the design are 0.2 and 2.0 mW/cm 2 . The irradiance applied can be controlled with pulse-width modulation (PWM): a widely used technique to achieve a mean value with high resolution with the weighted switching between two states, high (e.g., 3.3V) and low (0V). The stimulation protocol is pulsed: sequencing bursts of pulses (stimulation) with rest (no stimulation) phases. All, the number of bursts, the number of pulses or shots per burst, and the rest between bursts from seconds to hours, shall be configurable. Each shot has a certain cadence (period) and pulse width, which shall be configurable with millisecond resolution, allowing the fast switch between excitation and refractory phases within a burst. In Fig. 3.a graphical representation of the configurable stimulation protocol is shown. Consequently, the system shall be designed to be capable of driving each photoemitter at a selected level of power, applying pulses at high frequency, with different duty cycles, and on-off long time control.
To configure all the beforementioned parameters, a Graphic User Interface (GUI) is implemented. These parameters will be sent through USB to a microcontroller, which will handle all the necessary operations.

III. ELECTRONIC DESIGN
Given the 470 nm optogenetic stimulation wavelength, the HLMP-CB2A-VW0DD [19] LED is chosen for the OSIVE Slim optogenetic stimulation array. The procedure shown below can be repeated for a different dominant wavelength LED according to the parameters indicated in its datasheet to assess whether it meets the irradiance requirements for the target application.
The LED's datasheet states that at an angle of view α of 23°C, its luminous intensity is between 4200 and 7200 mcd when 20 mA is applied. Knowing the bottom of the wells has a surface of 1.8 cm 2 and the lens outputs the light at a distance d of 15 mm height of it, we can apply a simple calculus to get the luminous flux in lumens (see representation in Fig. 4).
Given that 1 candela is equivalent to 1 lumen per steradian (sr), we first calculate the solid angle Ω produced by our LED given an angle of view α of 23°C (1) [20].
Compared with the area of the well, of 180 mm 2 , this suggests that just a certain part of the well will be irradiated A IR : calculating r with simple trigonometry, the irradiated area ratio AR is obtained (2): Now we apply the candela to lumen conversion (3): To the minimum and maximum margins, we find that at 20 mA the luminous flux (Φ) at the bottom of the wells will be between 3360 and 5760 mlm.
As the lumen is a subjective measurement of the luminosity, dependent on the wavelength and the human eye, we need to convert lumen to watt of irradiated power (radiant power, Φ e ), which is given by equivalence tables [21]: at 470 nm, 62.139 lm = 1 W (4).
Therefore, the minimum and maximum radiant fluxes are obtained: Φ e min = 54.07 mW, and Φ e max = 92.69 mW.
Irradiance flux density E e is the radiant power per area, A IR = 127.4 mm 2 , considering its minimum and maximum values obtained before, the minimum and maximum irradiance are obtained (E e min and E e max ) (5).
With the purpose of assessing the adequacy of the former ca lculations, we prepared an experimental setup setting the current to 20 mA through a HLMP-CB2A-VW0DD LED. The voltage drop over the LED was 2.9 V, equivalent to electrical power P e of 58 mW. Regarding that the LED has a luminous efficiency η V of 78 lm/W, luminous flux (Φ) and irradiance (E e ) are obtained (6).
Resulting in a luminous flux Φ of 4.524 lm, and an irradiance E e of 57 mW/cm 2 , which fits with the typical value. Even though, we should not forget that the irradiance given by a LED might vary between its minimum and maximum values. To provide a wide range of irradiance values to the experiments, considering spatial and power supply restrictions, we design the next control circuit (Fig. 5).
Considering the BC337-25 will be working in cutoffsaturation regimes and the most restrictive situation with V BE = 1.2 V, h FE = 100 and V CEsat = 0.7 V, and given a 3.2 V diode voltage drop, the characteristic values given by its datasheet, the current at saturation I Csat can be obtained (7).
Experimentally, we measured a current of 55 mA through the diode branch, a V D of 3.14 V and a V CEsat of 24 mV. Given these values, the theoretical calculations concur with the normal experimental values V BE = 0.7 V, h FE = 100 and V CEsat = 0.05 V @ I C = 50 mA (8).  Moreover, the circuit is simulated in LTSpice, the results are shown in Fig. 6, which concur with our prototype results and calculations.
Given these results and given the LED forward current vs relative luminous intensity curve (see Fig. 3 in [19]), we determine that at 55 mA current the luminous intensity is 2.2 times the normalized intensity at 20 mA, which irradiance has been calculated before (6), so the maximum irradiance of this system is of 126 mW/cm 2 .
To avoid functioning over the absolute maximum ratings of the LED, the average power is diminished through PWM, limiting the irradiance to 30 mW/cm 2 , the full scale irradiance E FS . This is done with the PWM modules of the Atmel SAM3X8E 32-bit ARM Cortex-M3 core, configured to output a 20 kHz PWM with a varying duty cycle between 0 and 34.5%. This PWM is also modulated to output configurable shots of light up to 20 Hz (50 ms period) so a time for excitation/refractory period can be estimated and set, and a rest time up to 72 hours between trains of pulses is also added. The general block diagram for the LED control is shown in Fig. 7. The exact output values for the power, excitation and refractory period times, the number of pulses and repetitions are set by the user in the Graphical User Interface, connected to the microcontroller by USB to serial conversion Regarding the power consumption and supply considerations, OSIVE Slim can be powered directly from a USB port from the computer hosting the Graphic User Interface. USB 2.0 ports supply up to 500 mA, and 24 LEDs at maximum power will draw 250 mA (9), adding up to the maximum of 130 mA of the Arduino DUE consumption.
Otherwise, the 5V USB micro-B power connector or the ∅ 2.1 mm 7-12 V power jack connector can be used to supply the necessary power to OSIVE Slim from any commercial power supply.
To satisfy the control and height constraints, OSIVE Slim is designed within two subsystems: r A Control Shield, based on Arduino DUE board (with an Atmel SAM3X8E microcontroller as core) is used as the controller of the system and interface with the PC.

IV. APPLICATION IMPLEMENTATION
The application consists of two related programs: the Graphic User Interface, implemented in Processing, allowing the global configuration of OSIVE Slim and its real-time control, and  the Atmel SAM3X8E firmware, in C, programmed with Atmel Studio (currently renamed as Microchip Studio).
Both flux diagrams, for each program, are shown in Fig. 9. The GUI, after startup, handles the events coming from the user manual input. That is to say: the connection and disconnection to the USB interface (to the COM virtual port), the configuration of all the parameters for an upcoming experiment and the start and stop commands, sending the necessary data to the Control Shield to run an experiment, or sending the high priority stop message. All messages implement checksum to ensure transmission integrity of the message received, sending an acknowledge message from OSIVE Slim which triggers the corresponding GUI indicator. The GUI is shown in Fig. 10. Fig. 8 show an exploded view of both subsystems of OSIVE Slim.
On the other hand, the microcontroller's firmware is always running a thread that checks for new messages, upon arrival verifies the data chain integrity, and executes one of two options:  over stop command, turns off every photoemitter, restart and hold the timers; and over start command, stores the received information, configures the timers, and sets the "experiment on" flag (EOF). In both cases it sends the corresponding to acknowledge message. Once the message has been processed or there is no message, it will check the EOF to assess, and if it proceeds, update, every timer, with the corresponding action (turn on or off, at the specified power, the group of LEDs). Once timers have been configured, automatically each of the PWM modules will generate the control signals. After this, it will start from the beginning.

V. EXPERIMENTAL PROCEDURE
Primary cultures of cortical neurons were obtained from fetuses at 14.5 days of gestation (E14.5). Briefly, the brain cortex was mechanically triturated and once homogeneous, filtered through a nylon net with 90 μm pores and centrifuged at 500 rcf for 5 min. Cell pellets were first seeded into DMEM high glucose supplemented with 10% fetal bovine serum (FBS) and 1% P/S (plating medium) into poly-L-lysine (PLL)-coated bottom-glass 24 well plates (Cellvis; Mountain View, CA; #P24-0-N). After 2 h of incubation at 37°C, 5% CO2, the plating medium was replaced by Neurobasal medium supplemented with 1X B27 supplement (Gibco), 50 mM Glutamax (Gibco), and 1X P/S. Expression of bPAC protein in cortical neurons was achieved by viral infection using a AAV9 vector (pAAV-CamkII-bPac(WT)-mCherry-minWPRE) three days after plating at 10 5 transducing units/cell..sbd; Addgene #118278). The DNA construct delivered to the cells carries a codon-optimized gene encoding the BL-activated photo-stimulated adenylated cyclase (bPAC) to generate intracellular cAMP [22] fused to a red reporter (mCherry) under a neuronal promoter (CamkII). One week after the infection, the successful expression of bPAC was verified by observation of red fluorescence conferred by mCherry. Before the time-lapse recording, the cells were incubated with 3 μM Fluo-4 AM (Thermo Fisher) following the manufacturer's recommendations. For the calcium recordings we have used a confocal microscope (Leica TSP-SP8) equipped with resonant scanning mirrors to allow the adequate imaging speed required for real-time recordings of calcium in living cells [23]. Moreover, our confocal system allows us to maintain adequate temperature and CO 2 levels for live imaging. During the recording, we have acquired 1 image every 25 ms (40 frames/s) for 120 s. Forty seconds after the beginning of the recording, photostimulation was applied using the OSIVE Slim device with blue light (100 ms pulses at 470 nm every 10 s), which efficiently induced Ca 2+ influx assayed using a Fluor4 probe).
Next, we measured the timing of cAMP activity in vitro. After a light flash of 4 s, cAMP continued to rise in the dark with a time constant 23 s.

VI. EXPERIMENTAL RESULTS
Calcium dynamics are studied based on the increment of fluorescent given by Fluo-4 (green, fluorescent probe for calcium) upon real-time optogenetic stimulation and image acquisition as shown in Fig. 11. Optostimulated cells consist of transduced cortical neurons expressing the optoactivatable adenylate cyclase protein (bPAC) under the control of a neuronal promoter (CAMKKII), and stimulated with 100 ms blue light on-10 s blue light off pulses at 0.2 mW/cm 2 starting after 40 s of the beginning of the recording (as indicated within the blue dotted line, Fig. 11). The non-stimulated group corresponds with not exposed wild cells. The increment in fluorescent can be appreciated in stimulated cells but not in non-stimulated indicating that the blue light stimulation induces calcium accumulation in the transfected cells.
The developed platform, OSIVE Slim, has made possible the optogenetic stimulation with real-time measurement of fluorescence. The device, fully configurable in terms of irradiance and temporal parameters, opens the door to the study of optogenetically-induced transient phenomena in a flexible and economical way, using a confocal inverted confocal microscope. In turn, optostimulation with other wavelengths is within reach by changing the photoemitter, following the comprehensive step-by-step procedure described above.