1. Introduction

The detailed understanding of mechanisms involved in the development and growth of mammalian cells is of fundamental importance. In this regard a variety of recent studies have established that cells respond not only to chemical stimuli but, intriguingly, to physical stimuli including laser light which may promote optical guidance. A future vision in the topic of Neurobiology is to control the regeneration of neurons in various neurodegenerative diseases or traumas: the elucidation of how nerve cells grow and develop under various chemical and physically differing environments will underpin this work. For decades it has been known that light can manipulate both multi- and uni-cellular organisms from plants and animals to primitive protozoa. However, in the last decade, it has been observed that individual mammalian cells can also respond to light in some surprising manners. In 1991 experiments showed that migrating Swiss 3T3 cells (an embryonic mouse fibroblast cell) extend pseudopodia (temporary cellular projections) towards infrared light when positioned in close vicinity [1]. It was found that strongest cellular responses were observed using light wavelengths in the range of 800–900 nm, at rates of 30–60 pulses per second. The temperature increase induced by light absorption was reported to be negligible [1]. The mechanism by which the cell detects and responds to the light stimuli is not understood but it is thought the cytoskeleton plays an important role.

More recently it has also been proposed that the cytoskeleton also contributes significantly to the guiding of neuronal growth by light. PC12 cells (a rat pheochromocytoma cell from the adrenal gland), are a neuron precursor cell-line that can be stimulated to produce cellular processes that have neuronal properties [2]. There are many chemical signaling pathways that dictate this process and which may play a key role. A notable experiment in 2002 showed that processes from within these neuronal cells can also respond to light by motion of the growth towards a focused Ti:Sapphire laser beam (800 nm) of relatively low power (<100 mW) [2]. This optical guiding process was not attributed to the physical holding of the growth cone by optical tweezing or temperature induced gradients but rather was because the light was pooling actin monomers and therefore providing nucleation sites for actin polymerization which is known to be the driving force for neuronal growth [2, 3]. Subsequently, the same authors developed a computer controlled system for exploring this process [4]. Other studies, notably using a different cell-type (N1E-115 cells from a neuroblastoma tumor) and a separate optical set-up and at a different temperature (25 °C), have indicated the use of an asymmetrical light field at 1064 nm to also promote this phenomenon, though the efficiency of this process at this wavelength is unknown [5].

In the original work it was hypothesized that heating effects caused by absorption of the laser beam at 800 nm played a minor role because the steady-state temperature increase is estimated to be only a few degrees [2]. However at 1064 nm the opposite was hypothesized but then discounted through use of a symmetric line trap [5]. However, there has been no definitive direct measurement of temperature changes induced at either of these wavelengths to conclusively rule out whether or not changes in temperature could be affecting the rate of actin polymerization and subsequently causing neuronal growth. Furthermore, these experiments would ideally be performed on the same cell-line at the same temperature with exactly the same optical beam shape for excitation. This would contrast to references 2 and 5 which were on different cell-lines at different ambient temperatures and differing optical beam shaping schemes.

Whilst these studies have shown the promise of this new and exciting field, for it to be taken forward and be applicable in a biological context detailed comparison studies using the same cell-type and same set-up and at the same temperature are required. This would reveal the repeatability and robustness of this process and assist in revealing the exact mechanisms responsible for this effect. Also it is essential to elucidate whether the process, and indeed the associated mechanism, has a wavelength bias and whether changes in temperature do occur at different wavelengths. In this paper, we perform the first detailed comparison study of neuronal growth at near infra red wavelengths on one single cell type. In contrast to previous studies we present data for the manipulation of growth of the same cell type at two different laser wavelengths (780 nm and 1064 nm) at which little damage occurs to biological systems. We also followed the example of Ehrlicher et al., as studies were performed on large numbers of cells (samples of 50 cells) for each wavelength. There was no statistical bias to either wavelength. To address the temperature discrepancy in the literature, we also performed the first direct temperature measurements at 780 nm and 1064 nm wavelengths using Q values for trapped objects. This showed that 1064 nm has stronger heating capability than 780 nm. Transmission measurements through the cellular medium further confirmed this observation, though each wavelength gave no more than a few degrees of heating of the sample medium. Temperature is known to affect the rate of actin polymerisation under various conditions [7], but our findings are in agreement with Ehrlicher et al. as the measured increase in temperature from both wavelengths would have little effect on the overall rate of polymerization at 37 °C [2, 7]. Therefore as both wavelengths had the same growth rate capability this demonstrates for the first time we can discount heating as the prime cause of this effect.

2. Experiment setup

The experimental setup for neuronal growth is shown in Fig. 1. The system was a home-built inverted microscope system with a variable optical telescope to focus the beam to a desired spot size. The lasers employed in these studies were a violet diode laser operating at 405 nm (Toptica Photonics, 50 mW), a circularized 780 nm diode laser (Pro-lite PS110, 80 mW), and a 1064 nm fibre laser (Spectra Physics, 100 mW). A dichroic mirror directed the beam into the back aperture of a microscope objective where the beam was focused to an appropriate spot size. The objective used for this study was a Newport M-60X air (NA=0.85).

Typically the 780 nm laser was used at a power of 9–25 mW at the focus, with a 1.5 micron beam waist. The transmission at this wavelength was 65% through the optics and 77% through the objective, giving a system transmission of approximately 50%. The 1064 nm laser was used at a power of 8–22 mW at the focus, again with a 1.5 micron beam waist. The system transmission was approximately 68% at this wavelength. Light at 405 nm had a transmission again of around 50% through the system, but as it caused cell damage and induced no growth following our first stage experiments we did not pursue further investigations at this wavelength. Imaging was performed in a bright field geometry using white light illumination with the sample plane imaged through the final dichroic mirror onto a CCD camera. Video data was recorded for the neuronal growth over a period of up to 1 hour.

We performed our studies using the NG108 cell line. NG108s are a rat/mouse hybrid neuroblastoma cell line which produce active growth cones. They were cultured in Dulbecco’s Modified Eagle Medium (DMEM), with 10% foetal bovine serum (FBS). The NG108s were seeded on to glass-bottomed culture dishes which had been previously coated with 7 µg/cm² of collagen (Sigma). In the dishes the cells were kept in two-thirds DMEM with 1% FBS and one-third pre-conditioned DMEM (media removed from the culture flask). Cells were placed on a sample stage, heated by an electric heating element powered by Marlow SE5000 temperature controller. The temperature was stabilised to 37±0.5 °C.

 

Fig. 1. The optical setup. The system used laser wavelengths of 780 nm and 1064 nm each focused to approximately the same beam spot in the sample plane.

Download Full Size | PDF

The laser power variation noted above was employed to ensure the sample was irradiated with sufficient power. For every cell sample dish investigated, the laser power was increased until the neuron retraction was observed. The power was then reduced to below this retraction threshold to begin the growth process. The focused laser spot was positioned at the tip of the growth cone and dithered back and forth as explained in reference [2].

3. Results

Our studies were designed to explore if we could monitor neuronal growth cone development under our designed set-up. Rather than observing whether we could increase or decrease the rate of growth cone development in the same direction of the original growth cone, we opted to measure whether we could deviate the direction of the growth cone by at least 30 degrees away from the original direction of neuron growth travel. This therefore provides a more accurate and objective measurement of the manipulation of growth cone formation by light. We also opted to perform this on a large number of cells due to the inherent variability from use of biological material.

Examples of the changes observed when influencing growth cone formation and growth are shown in Fig. 2 (A – 1064 nm; B – 780 nm). In both examples the laser point was able to initiate a new direction of growth as the original light source was always placed to one side of the original growth cone. After 30 min of monitoring, the laser point can be seen to have moved from its origin indicating that the laser initiated a change in growth direction and that it was not the cell moving away from a fixed point giving the illusion of new growth.

These experiments were then performed on much larger numbers of cells than has been previously attempted in the literature. The results are summarized in Table 1. Out of 50 cells studied for each condition in separate monitored experiments, 31–33 cells had actively growing growth cones. If the growth cone rate was less than 2 microns/hour then they were diagnosed as being “comatose”. In approximately half of the actively growing growth cones there was a statistically significant influence of the use of a laser source as compared to a no laser control (p<0.05, Mann-Whitney test). Indeed there appeared to be no appreciable difference between either the 780 nm or 1064 nm wavelengths of light. We also noted that this phenomenon appeared to be balanced as approximately a third of the growth cones would also undergo a period of retraction after a period of attraction. When experiments were attempted with a 405 nm laser then this resulted every time in the immediate retraction of the growth cone and so was not pursued further.

 

Fig. 2. Examples showing the influence of laser light (circled) of (A) 1064 nm or (B) 780 nm wavelength on growth cone direction. Each image is 6 minute intervals for a total of 30 min of observed growth. The black circle denotes the position of the laser spot which was moved over time. The arrow in the top images indicates the original direction of neuronal growth. The arrows in the bottom images indicate the final direction, thus showing the change in direction of growth caused by the influence of the laser light.

Download Full Size | PDF

Table 1. Summary of the results from 150 separate experiments. *Significantly different from non-laser control (Mann-Whitney test, p<0.05 where p denotes the 95% confidence level). #Non significant difference between these two wavelengths (Mann-Whitney test, p<0.05). For every cell sample dish investigated, the laser power was increased until the neuron retraction was observed. The power was then reduced to below this retraction threshold to begin the growth process.

View Table

4. Discussion and role of temperature for neuronal growth

The control and manipulation of nerve cell growth is an actively researched area because of its fundamental biological importance but also its potential clinical applications. Most biological experiments have centered on the use of chemical signals and from this work has resulted an understanding of the biochemical pathways that are involved in this dynamic process [3]. In general, neuronal growth is thought to be driven by the fluctuation of two forms of the cytoskeletal protein, actin. Actin can exist as either a monomer termed G-actin or a polymerized form termed F-actin. In neuronal growth the amount of F-actin remains constant with the G-actin at the leading edge of the growth cone polymerizing into F-actin. At the distal end of the F-actin, depolymerisation occurs with the production of the G-actin monomer which is then transported back to the leading edge of the growth cone to complete the cycle. Therefore the critical starting point is the polymerization of G-actin into F-actin at the leading edge and this occurs only if a suitable signal is detected by the leading edge filopodia. This signal is typically in the form of a chemical signal from the surrounding medium or substratum. This ‘actin tread-milling’ is also influenced by a wide range of supporting proteins and biochemical signaling pathways [3].

Light mediated cell guidance is a very new field of bio-physical research and as such very little is known about how this light induced phenomenon relates to the known chemical induced pathways or whether it relies on a completely new signal transduction pathway. For these studies to progress it is imperative to determine how reproducible these experiments are, and to determine some of the bio-physical parameters that govern this phenomenon, as these will influence future biological experimentation.

Here we have presented the first comparison of this novel physical event on the same optical set-up, the same cell-type and at two different wavelengths on a large number of cells. Proteins in general are known to absorb light and hence their concentration can be commonly measured by means of a spectrophotometer at particular wavelengths e.g. natively at 280 nm or in conjunction with a dye (Bradford’s reagent) at 580 nm. The infra-red region of light was not damaging to cells, and was able to manipulate cellular growth. Intriguingly, there appeared to be no difference between 780 nm and 1064 nm wavelengths of light in inducing this phenomenon with similar beam shapes at both wavelengths. This suggests that the light detection mechanism within the cell is not due to a single protein with a defined activity wavelength as occurs for example with the photo-receptor opsin proteins in the mammalian eye [6].

As was previously found [2], it was interesting to note that the cells that responded were those which were already actively growing. The point of light was adjusted manually to the tip of the growth cone, suggesting that it is not necessary for the light source to be kept constantly at the leading tip of the growth cone. This would imply that the light trigger once stimulated, allows a period of continued growth in this new direction. As previously stated, in neuronal growth it is the polymerization of G-actin to F-actin which is important in directing neuronal growth and not solely the transport of G-actin [3]. However, the actual mechanism of light induced growth remains elusive but one suggested mechanism has been the involvement of localized heating [5]. However, a discrepancy exists in the literature when this is applied to neuronal growth. Previously at 800 nm it is was suggested that temperature was not important because it was estimated that the steady state increase in temperature was negligible [2]; however, at 1064 nm then an increase in temperature was suggested as a possible mechanism [5], again noting that each of these experiments was conducted at a different base temperature and different cell-types. It has been previously shown by measuring the actin polymerization of purified proteins, that this can be affected by temperature but this in turn is affected by both actin concentration and local ionic strength [7]. As an example (see Fig 2(c) in Ref [7]) a change in temperature of ~±5 °C change in temperature results in ~10% change in the polymerization rate at a base temperature of 37 °C.

To address this issue, we performed a comparative study of the localized heating effects in water near the focus of 780 nm and 1064 nm laser wavelengths. Based on the established method reported by Mao and co-workers [8], we constructed a dual-beam optical tweezers system consisting of an AlGaInP laser diode (660 nm) and ×60 objective lens for trapping, together with either a 780 nm or 1064 nm Nd:YVO₄ laser (Photonics Innovation Centre, St. Andrews, Scotland, U.K.).

2 µm silica spheres (Duke Scientific) were first optically trapped in water by the 660 nm laser. The associated trapping efficiency was then determined at three laser focal powers (P₆₆₀=13, 27 and 36 mW) by measuring the maximum average lateral trapping velocity, v, where v=(Qn/3πηdkc)P and Q is the Q-value, n and η are respectively the refractive index and viscosity of the surrounding medium, d is the diameter of the sphere, k is a correction factor for the position of the sphere in the sample chamber (k=0.979 in this case), and c is the speed of light [9]. By then plotting P versus v, the gradient of the straight line fit was then used to determine a ratio of Q/η. The 780 nm laser was then focused on the same sample plane as the trapped sphere using a ×60 objective lens, and positioned at a lateral distance of 5 µm away from the trap. This distance of 5 µm was close enough to allow detection of any heating at the trap site, but at a sufficient distance to avoid any contributions by the 780 nm and 1064 nm lasers towards the optical gradient force of the 660 nm trap. The power at focus of the 780 nm laser was set to 15 mW, which represents a realistic power level for neuron growth experiments. The trapping efficiency of the 660 nm trap was then repeated at the same trapping powers (P₆₆₀=13, 27 and 36 mW) in the presence of the nearby 780 nm laser spot. Finally, the 1064 nm laser (15 mW at focus) was positioned in place of the 780 nm laser, and trapping efficiency of the 660 nm trap was repeated once more.

We then calculated the fractional rise in the ratio of Q/η for the presence of each of the 780 nm and 1064 nm laser spots, thereby determining any changes in medium viscosity, and therefore in temperature [10], resulting from the presence of the 780 nm and 1064 nm wavelengths. At a room temperature of 25 °C for the case of using only the 660 nm trapping laser, we measured a temperature of 26.4±1.5 °C and 28.1±2.3 °C in the presence of the 780 nm and 1064 nm lasers respectively. The associated inaccuracies are fractional errors of the best-fit line gradients, Q/η. This represents the first direct temperature measurements at 780 nm and 1064 nm wavelengths using Q values for trapped objects, and demonstrates that 1064 nm has stronger heating capability than 780 nm, as expected. Similar amounts of heating have been reported in the literature at nearby wavelengths – e.g. negligible heating at 830 nm [8] (where the degree of water absorption is very similar to 780 nm, as seen in Fig. 3) and 1.9 °C per 100 mW at 1064 nm [11, 12].

 

Fig. 3. Measured transmittance of water and Dulbecco’s Modified Eagles Medium (DMEM), the neuron growth culture media. Data was acquired using an ellipsometer.

Download Full Size | PDF

Our observations of stronger heating at 1064 nm than 780 nm can be explained by greater absorption at this wavelength, as is confirmed by our measured transmittances of water and DMEM (Fig. 3). These transmission measurements were made using a variable-angle spectroscopic ellipsometer (M-2000DI, Woollam Co. Ltd., Lincoln NE, USA), with the source and detector arms aligned for zero deflection angle. Each sample was placed in a 10 mm light-path optical silica cuvette (Hellma UK), and transmission measurements were acquired over a range of wavelengths from 190 nm to 1700 nm using both a deuterium lamp (UV-visible) and quartz tungsten halogen bulb (visible-IR). It is interesting to note that 1064 nm light experiences almost 10% more absorption in water (and DMEM) than 780 nm light. The water-dominant transmittance of DMEM above 700 nm is evident, as is the absorption of phenol red component of DMEM at wavelengths below 650 nm. However, despite these apparent differences in heating capability these results show that neither wavelength causes sufficient heating to affect the rate of actin polymerisation within cells held at 37 °C [2, 7] and thus in turn affect growth.

Conclusion

Neuronal growth on a single cell type has been investigated as a function of applied laser wavelength. Light assisted growth can occur equally well at both 780 nm and 1064 nm as supported by measurements on over 150 cells with equivalent setups. We also performed the first direct temperature measurements at 780 nm and 1064 nm wavelengths using Q values for trapped objects. This showed that 1064 nm has stronger heating capability than 780 nm. However at both wavelengths the increases in temperature would not be enough to affect the rate of polymerization at 37 °C. This therefore demonstrates that despite this difference in temperature heating both wavelengths under physiologically relevant conditions had the same growth rate capability. We thus can conclusively discount heating as the prime cause of this effect for both wavelengths. Our experiment is the first comparison study of a light mediated neuronal growth process on the same cell type at 37 °C and with the same optical excitation scheme and is a step towards a thorough and detailed understanding of this process.