1. Introduction

Near-infrared (NIR) ultrashort pulsed laser irradiation has been used for cellular surgery [1,2] and cellular stimulation [3–6] due to its long penetration depth, lower scattering, and nonlinear absorption [7]. We have demonstrated that the localized irradiation effects in living cells can lead to photo-induced intracellular Ca²⁺ waves [4,8]. In previous work done by our group, the laser source repetition rate was fixed at 82 MHz and we were only able to observe a single type of Ca²⁺ response (a Ca²⁺ wave). Even when the laser power was attenuated, the probability of observing a Ca²⁺ wave decreased, while the observed concentration of laser-stimulated Ca²⁺ release was largely unaffected. The binary-type Ca²⁺ elevation is caused by the self-amplification of the intracellular Ca²⁺ release and is a known feature of cell signaling [9]. However, self-amplification notwithstanding, cells naturally exhibit a variety of calcium responses. Sparks, puffs, and lower level calcium elevations [10,11] can be observed in cells in vivo and in vitro, depending on the type and amount of stimulation. For laser stimulation to be most useful as a tool to provoke and study calcium dynamics, the invoked calcium response could ideally be varied, depending on the laser parameters.

Here we demonstrate the effects of living cell stimulation using NIR pulsed laser irradiation with repetition rates of several hundred kHz. We found repetition rate dependent intracellular Ca²⁺ responses in laser-irradiated cells and observed that intracellular Ca²⁺ waves can be induced by lower repetition rates with a wave speed much lower than that observed in previous 82 MHz results. This finding allows a method of cell stimulation which can modulate the strength and speed of stimulation. The variation in the rate of pulse energy deposition in the sample also allows us to compare time-varying and non time-varying laser-cell interactions, and determine diffusion-related effects contributing to laser-induced Ca²⁺ waves. Previous experiments to determine the role of Reactive oxygen species (ROS) in the laser-cell interaction have been hampered by the fact that ROS scavengers are limited in diffusion speed and can be depleted at the laser focal zone. By lowering the repetition rate, and allowing more time for diffusion in the cell, ROS are shown to be a significant contributor to the mechanism of laser-induced Ca²⁺ waves.

2. Sample preparation and optical setup

Cells used in experiments were cultured HeLa (cancerous epithelial) cells, grown in Dulbecco’s Modified Eagle’s Medium (DMEM) which contained 10% fetal calf serum, 2 mM glutamine, 100 unit/ml penicillin and 100 μg/ml streptomycin, in a humidified atmosphere (5% CO₂) at 37 °C. HeLa cells were grown for 3 or 4 days until they became confluent on the glass-bottomed dish. They were immersed in a saline solution which contained CaCl₂ (1 mM), NaCl (145 mM), KCl (4 mM), MgCl₂ (1 mM), glucose (10 mM), and N-2-hydroxyethylpiperazine N’-2-ethanesulphonic acid (HEPES, buffer, pH=7.4 NaOH modified, 10 mM), and Fluo-4AM (0.018 mM) for 30 minutes in order to infuse the samples with the Ca²⁺ indicator [12]. The solution was then replaced by the saline solution without Ca²⁺ indicator to prevent extracellular fluorescence emission. All experiments were carried out at room temperature (~23 °C). To test for the production of free radicals, or other reactive oxygen species (ROS) by laser stimulation, 200 μM Trolox (a cell-permeable ROS scavenger and water soluble model for vitamin E), was added to the saline and loaded with HeLa cells for 10minutes at 23 °C before cell stimulation experiments.

 

Fig. 1. The effect of leakage pulses on average power at each repetition rate. Dots are measured laser powers. The solid line is the curve fit to experimental data shown by dots (1.52 nJ pulse energy, 0.0033 EOM extinction ratio). Dotted lines are calculated curves with constant pulse energy (1.52 nJ) and each extinction ratio (0, 0.1, 0.5, 1.0, 5.0, and 10%)

Download Full Size | PDF

The laser beam (wavelength 780 nm) from a mode-locked Ti:Sapphire laser (Tsunami, Spectra Physics) emitting 80 fs pulses was introduced into a Olympus BX50WI microscope. A dichroic mirror reflected the laser beam into the microscope while allowing simultaneous fluorescence imaging by using excitation light from a mercury lamp. The laser beam was focused into a spot of ~0.7 μm in diameter (FWHM) by a 60×/0.9 water-immersion objective lens (LUMPlan Fl, Olympus).

In order to modulate the repetition rate of the mode-locked Ti:Sapphire laser we inserted an electro-optical modulator (350-160, Conoptics. Inc.) into the optical path between the Ti:Sapphire laser and up-right microscope. We tuned the laser repetition rate from 100 Hz to 82 MHz, while a mechanical shutter (F77-4 & F77-6, Suruga Seiki ltd.) controlled the time of exposure to laser irradiation and therefore the total number of pulses incident on the cells. The laser power transmitted through the objective lens was measured in air by a diode-based power meter (PD-300, Ophir) positioned under the lens. The water-immersion objective angle of refraction changes from 43 to 64 degrees when used out of water, and the measured values were compensated for the inefficiency of the diode-based power meter at high light collection angles. Compensation was done by combining information provided by the manufacturer and calibration by a thermal power meter with negligible measured power dependence on collection angle. For an EOM with a perfect extinction ratio, the measured average power for a train of pulses is linearly dependent on the repetition rate. The extinction ratio of the EOM in the experiment was 300:1 and the measured average power therefore became constant at the lower repetition rates used in experiments. The contribution of leakage pulses at these lower rates dominates the average power measurement, and is shown in Fig. 1.

3. Experimental results

3.1 Repetition rate dependence of intracellular Ca²⁺ wave generation

We found that intracellular Ca²⁺ waves can be induced by NIR pulsed laser stimulation with repetition rates as low as several hundred kHz but the exposure time required to induce intracellular Ca²⁺ wave became much longer when compared to previous results [4]. The repetition rate was set by the EOM to 10 kHz, 100 kHz, 400 kHz, or 1 MHz. Figure 2 shows the probability of intracellular Ca²⁺ wave generation as a function of the number of exposed laser pulses at 2 different pulse energies. In both cases, the pulse energy was kept constant when the repetition rate was changed. The pulse energies used in Fig. 2(a) and Fig. 2(b) were 1.3 nJ and 0.99 nJ, respectively, as calculated by fitting the measured laser power at each repetition rate to laser powers calculated theoretically by the following equation (Eq. 1).

where P_ave the average power, RR is the repetition rate, E_p is the pulse energy, and E_r is the extinction ratio (obtained from measurements shown in Fig. 1). Since the average laser power changes considerably with repetition rate, repetition rate comparisons were carried out by fixing the laser pulse energy and varying the laser exposure time. At repetition rates below 1MHz, the number of laser pulses required to induce intracellular Ca²⁺ wave was similar. The number of pulses required for 50% Ca²⁺ wave probability was: ~280000 pulses at 10, 100, and 400 kHz (400 kHz not shown), falling to ~90000 pulses at 1MHz (Fig. 2(a)). Similarly, when the pulse energy was lowered to 0.99nJ, the number of pulses required for 50% Ca²⁺ wave probability was: ~160000 pulses at 10, 100, and 400 kHz (10kHz not shown), and only ~40000 pulses at 1 MHz (Fig. 2(b)).

The laser-induced calcium release mechanism undergoes a transition from repetition rate dependent to repetition rate independent as the repetition rate at approximately 1 MHz. There was, however, a lower limit on the usable repetition rate to stimulate a Ca²⁺ response. If the number of irradiated laser pulses at a given pulse energy and at low repetition rates is the only factor to influence whether an intracellular Ca²⁺ wave is induced or not, very low repetition rates such as 100 Hz may induce Ca²⁺ waves if the exposure time is sufficiently long. To test this, target cells were exposed to laser pulses at 100 Hz for up to 100 s, resulting in no observation of laser-induced Ca²⁺ waves (data not shown). The lack of Ca²⁺ response at very low repetition rates is interesting for two reasons. Firstly, it confirms the lack of effect of EOM leakage pulses on the laser-induced Ca²⁺ response. Even though the contribution of leakage laser pulses at a repetition rate of 100 Hz is 100 times higher than the effect at a repetition rate of 10 kHz, no laser-induced Ca²⁺ waves were induced by leakage pulses. Secondly, at very long exposures, the natural dynamics of the cell must be considered. Cell motility (movement of the cell) and cytoplasmic streaming (movement within the cell) during a long exposure time are both contributing to the lack of intracellular Ca²⁺ response. Cytoplasmic streaming is significantly faster than cell motility, with a reported speed in a HeLa cell of 2~3 μm/s [13]. By cytoplasmic streaming and the subsequent variance in the precise location of intracellular organelles over time, the effects of pulsed irradiation are not always arriving at the same position and the cumulative effects of the irradiation become different. The endogenous Ca²⁺ pumps are also simultaneously pumping cytosolic Ca²⁺ back into the endoplasmic reticulum and extracellular space, and therefore competing with laser-induced Ca²⁺ release over longer periods of time. Due to these factors, 10 kHz may represent the lowest practical repetition rate by which the laser-induced Ca²⁺ response can be reliably produced.

 

Fig. 2. The probability of intracellular Ca²⁺ wave generation at each number of pulses. The laser (780 nm wavelength, (a) 10 kHz, 100 kHz, and 1 MHz repetition rate, 1.31 nJ pulse energy, 0.0024 EOM extinction ratio, (b) 100 kHz, 400 kHz, and 1 MHz repetition rate, 0.99 nJ, 0.0036 EOM extinction ratio) was focused into the cytoplasm of the HeLa cell. n=30.

Download Full Size | PDF

3.2 ROS generation

Figure 2 shows that the total number of pulses required to induce Ca²⁺ waves did not change significantly at lower repetition rates. This implies that the effect of each laser pulse on the cell combines to produce the larger observed response. We considered the generation of diffusing free radical or other reactive oxygen species production in triggering the Ca²⁺ response. The presence of laser-induced radicals near the cellular calcium stores can cause lipid peroxidation [14] and direct inhibition of the Ca²⁺ATPase [15]. Hydrogen peroxide has been implicated in increased Ca²⁺ activity in cultured endothelial cells [16]. The generation of at least some amount of laser-induced ROS is probable, and has been observed in other experiments using mode-locked infrared lasers to generate H₂O₂ and other ROS in a variety of cell types [5,17,18]. The combination of HEPES buffer (used in our culture media) and ultraviolet-A (320-400nm) illumination (which is close to the two-photon absorption wavelength) has recently been reported and may also contribute to intracellular ROS generation [19]. We applied Trolox, a water-soluble and cell-permeable analogue for vitamin E and powerful scavenger of cytosolic ROS, to obtain a 200 μM concentration in the extracellular solution. We observed whether the Ca²⁺ waves could still be induced when the cellular antioxidant cascades are enhanced. Trolox can be expected to disperse evenly throughout the cellular cytosol. In contrast, laser-induced ROS was expected to be localized, existing in a narrow region surrounding the laser focus. The types of generated ROS differ in their chemical activity, as well as lifetime and therefore diffusion radius. The lifetime of singlet oxygen reported in water or biological tissue is on the order of several μs or lower [20,21], leading to a mean diffusion distance of less than 0.5 μm. The hydroxyl radical has a mean diffusion distance of only ~30 Å, while the hydrogen peroxide radical is stable enough that it can cross cell membranes [22]. Some types of ROS, such as singlet oxygen and hydroxyl radicals (highly reactive), can therefore be expected to be confined to the focal zone by localization due to both two-photon absorption and inherent ROS lifetime. Since the generation of ROS was expected to be localized at the focal spot, Trolox must be able to diffuse into the focal spot in order to be effective. We therefore tested ROS inhibition at different repetition rates, where the ability of Trolox to diffuse into the focal region and inhibit generation of ROS was expected to change with repetition rate.

With a 100 kHz pulse train, and a pulse energy of 1.3 nJ, cells were stimulated at 100 kHz, then the intracellular Ca²⁺ wave generation rate was measured 10 minutes after application of 200 μM Trolox to the cell culture dish. The probability of intracellular Ca²⁺ wave generation was measured again. The application of 200 μM Trolox caused a reduction of the probability of intracellular Ca²⁺ wave generation. Partial recovery of the Ca²⁺ wave generation was then observed after removing Trolox (Fig. 3). This result was unique to the low repetition rate experiments.

 

Fig. 3. The probability of intracellular Ca²⁺ wave generation with or without Trolox. The laser (780 nm wavelength, 100 kHz repetition rate, 520 μW average power, 10 s exposure time) was focused into the cytoplasm of the HeLa cell. Error bar is calculated from standard deviation in binominal distribution. (n=40)

Download Full Size | PDF

Figure 4 shows the effect of 200 μM Trolox application at repetition rates of 100 kHz, 400 kHz, and 82 MHz. These results show that the generation of ROS in the focal zone is a contributing factor to laser-induced Ca²⁺ release of Ca²⁺ from stores such as the endoplasmic reticulum, although they only demonstrate the effect at repetition rates below 1 MHz. From this, we must conclude either that effects resulting from ROS occur only at low repetition rates, or that Trolox is not an effective inhibitor at high repetition rates. Since photochemical effects leading to ROS formation cannot be expected to decrease when the repetition rate is raised, we suggest that Trolox scavenging is ineffective at the higher repetition rates. If, as previously mentioned, several strongly reactive types of ROS are confined to the focal zone, Trolox may become depleted in the focal zone and the effectiveness of Trolox on ROS inhibition will then be determined by the rate of diffusion into the focal zone.

To determine the rate at which Trolox can diffuse into the focal zone, the 1D diffusion equation in spherical coordinates with angular symmetry is:

where U represents (arbitrary) concentration and κ is the diffusion constant for the molecule. Based on the molecular weight, the diffusion constant of Trolox can be taken to be approximately 5×10^-10 m²s^-1. The recovery time for depleted Trolox in a 0.5 μm radius sphere centered on the laser focal spot is shown in Fig. 5. If Trolox is fully depleted by one laser pulse, the concentration will recover to reach 50% of the original value within 100 μs. At the repetition rate of 10 kHz, 100 μs is the arrival time of the next pulse. If Trolox is not completely depleted by one laser pulse, then the repetition rate can be higher than 10 kHz without losing the ROS scavenging effect. As shown in Fig. 4, the effect of Trolox is still observable at repetition rates up to 400 kHz, indicating that the effect of a single laser pulse of up to 1.75 nJ does not produce enough ROS to deplete locally available Trolox scavengers.

 

Fig. 4. The probability of intracellular Ca²⁺ wave generation at each average power with or without Trolox. The laser (780 nm wavelength) was focused into the cytoplasm of the HeLa cell. Error bar is calculated from standard deviation in binominal distribution. (n=20) (a) 100 kHz repetition rate, 10 s exposure time (b) 400 kHz repetition rate, 10 s exposure time (c) 82 MHz repetition rate, 13 ms exposure time

Download Full Size | PDF

It is worth noting that even though ROS generation was suppressed by the addition of 200 μM Trolox, 100% probability of intracellular Ca²⁺ wave generation at sub-MHz repetition rates could still be observed by compensating for the Trolox effect with higher laser pulse energy. Additionally, the shape of the Ca²⁺ wave probability vs. laser power relationship (Figs. 4(a) and (b)) does not change significantly under the influence of Trolox. The decrease in Ca²⁺ wave probability attributable to Trolox observed at 400 kHz was similar to that at 100 kHz, though the time available for Trolox diffusion between pulses was 4 times longer. This implies that ROS generation is not the only mechanism involved in intracellular Ca²⁺ wave generation, but instead works in combination with direct photoionization of other targets, including endogenous and exogenous fluorophores in the cytosol.

 

Fig. 5. Simulation of diffusion-based recovery of an ROS scavenger into a depleted region inside the focal zone. The graph is based on complete depletion of Trolox by a single laser pulse in a sphere with a 0.5 μm radius. After complete depletion, 50% recovery is achieved within 100 μs, corresponding to a laser repetition rate of 10 kHz.

Download Full Size | PDF

The overall Ca²⁺ response probability dependence on an increase in laser pulse energy was remarkably similar for repetition rates varying over three orders of magnitude (Fig. 4). The shape of the power dependence curve is similar to those previously measured at high repetition rates [4]. At all repetition rates, the average power (corresponding to pulse energy) at the onset of laser-induced Ca²⁺ waves was approximately half of the average power required to generate laser-induced Ca²⁺ waves with 100% probability. Given the femtosecond pulse width, and combining the results shown in Figs. 2 and 4, NIR laser pulse energy and number of exposed pulses are the most important parameters governing the laser-induced Ca²⁺ release. The interaction is mediated by ROS, but the observed inhibition of ROS effects by Trolox does not preclude other mechanisms of laser-induced Ca²⁺ release, such as photoionization. As the repetition rate decreases, the total number of pulses required to generate Ca²⁺ waves becomes constant (Fig. 2). This demonstrates that low repetition rate laser-induced Ca²⁺ release is based on photo-induced changes that are not strongly diffusing in the cell.

3.3 Slow Ca²⁺ wave stimulation

One fundamental difference observed in Ca²⁺ waves after low repetition rate laser stimulation was the reduction in the speed of Ca²⁺ response (Fig. 6). When the laser (780 nm wavelength, 100 kHz repetition rate, 10 s exposure time) was focused into the cytoplasm of a HeLa cell, the intracellular Ca²⁺ concentration can be seen to gradually increase during the laser irradiation period taking a maximum value 17 s after the shutter is closed. For a group of cells, the average time to reach the maximum value from the onset of laser irradiation was 24.2 ± 6.1 s (n=4) at a repetition rate of 100kHz. For the 82 MHz repetition rate, the maximum value was reached in 10.3 ± 3.5 s (n=5) from the onset of laser irradiation. In addition to the slower rise in Ca²⁺ concentration, the concentration gradient also differs from that observed in high repetition rate laser stimulation.

 

Fig. 6. (a) Fluorescence images of 100 kHz laser-induced intracellular Ca²⁺ wave (false color). Arrow indicates the position of laser irradiation. The time shown below each figure is counted from the start of laser irradiation. The pulse energy was 1.36 nJ and exposure time was 10 s. White scale bar = 25 μm. (b) Time course of fluorescence intensity averaged in a whole laser-irradiated cell. The fluorescence intensity corresponds to Ca²⁺ concentration becomes maximum in 30.7 s after laser irradiation. Black bars above each figure show the laser irradiation.

Download Full Size | PDF

Figure 7 shows the difference in Ca²⁺ gradient. In both cases, to compare the slowest possible Ca²⁺ response at different repetition rates, we used the minimum laser power required to induce intracellular Ca²⁺ waves. The pulse energies at 100 kHz and 82 MHz were 1.37 nJ and 0.29 nJ, respectively, and the irradiation times were 10 s for both cases. The Ca²⁺ wave produced by 82 MHz repetition rate irradiation exhibits a strong gradient, and originates from the laser focus. The 100 kHz Ca²⁺ wave shows no strong gradient, and does not clearly originate from the laser focus. This is in part due to the slower Ca²⁺ concentration rise, allowing more time for the intracellular Ca²⁺ to diffuse evenly throughout the cell, but since the exposure time is the same in both cases, the lack of obvious gradient in the case of 100 kHz indicates that the calcium level is rising throughout the entire cell with no strong dependence on the initial position of the laser focus. The fluorescence signal actually appears to originate from the central region of the cell, since the detection of fluorescence was non-confocal and the cell is thicker in the central region than at the edges. We considered that the difference between fast and slow induced Ca²⁺ waves was due to different types of ROS. Notably, H₂O₂ with its μm scale diffusion radius should be more susceptible to scavenging by Trolox as it diffuses through the cell. However, no differences in Ca²⁺ gradient (see Fig. 7(d)) were observed with or without Trolox, and the Trolox efficacy (Fig. 4) did not change between 100 kHz and 400 kHz repetition rates.

We conclude that the slow Ca²⁺ wave response is the observable result of a several different processes that cause increased cytosolic Ca²⁺. The initial energy deposition occurs via multiphoton absorption, since the Ca²⁺ response cannot be produced without laser mode-locking. Once absorption occurs, the processes leading to cytosolic Ca²⁺ elevation include laser-induced ROS, direct photoionization of lipids or other molecules including water [23,24], ROS-induced Ca²⁺ release [25], and Ca²⁺-induced Ca²⁺ release [9]. This increase is balanced by processes that decrease cytosolic Ca²⁺ levels such as: Ca²⁺ATPase (Ca²⁺ pumps) in the plasma membrane, endoplasmic reticulum, and mitochondria, which are in turn affected by ROS levels in the cytosol [16].

 

Fig. 7. Time course of 100 kHz (a,b) or 82 MHz (c,d) laser-induced intracellular Ca²⁺ wave (false color). A straight yellow line represents the axis through which the video sequence was resliced, to produce the images shown in (b) and (d). The pulse energy was 1.36 nJ and exposure time was 10 s. Scale bar = 2 μm (a, c) or 1s (b, d).

Download Full Size | PDF