1. Introduction

The cytoskeleton of eukaryotic cells is fundamental to the spatial organization of the cell components. Cytoskeletal systems are dynamic and adaptable; they can change or persist, according to need. Cellular structures must be assembled, disassembled and reorganized during the life of the cell; thus the individual macromolecular components that make up these structures are in a constant activity. Cells are small and complex, it is hard to see their structure and movements, and what we can learn about cells depends on the tools available, and from the introduction of different techniques [1].

PC 12 cells originate from rat pheochromocytoma cell line from adrenal medulla; when grown in a serum-containing medium, PC12 cells divide and resemble precursors of adrenal chromaffin cells and sympathetic neurons. Upon addition of 50~100 ng/ml Nerve Growth Factor (NGF), PC12 cells extend long, branching neuronal-like processes, rapidly onsetting sequential changes in their surface architecture within 1 minute that leads to sympathetic neurons after 2–4 days of exposure (differentiated PC 12-like neuronal cells) [2–5]. The cells of a multicellular organism are member of a highly organized community; their number is tightly regulated not simply by controlling the rate of cell division, but also by controlling the rate of cell death. Cells that die by necrosis, as a result of acute injury, typically swell and burst, and they spill their contents all over their neighbours, causing a potentially damaging inflammatory response [1].

The SNOM system [6, 7] is a purely non-invasive, non-contact method therefore the natural life activity of the sample is unperturbed. During the latest years these cells have been very well characterized both morphologically and biologically [2–5], therefore they represent the ideal substrate for our studies with the SNOM. Our intent is in fact to identify a specific SNOM signal associated to the morphological modification of PC12 due to NGF treatment and necrotic stimulus, thus revealing the high sensibility of SNOM.

2. Experimental setup and procedure

Detection apparatus was a customized SNOM [8, 9], the sample holder was modified in order to set the plastic dish containing the cells and culture media. The system was fitted with a shear force (SF) control system [10] in order to calibrate the probe-substrate distance, however all measures were done in free running conditions. At first approach the probe on the bare glass, SF controller will give us a reference to zero, then we move electronically to a 0.5 micron probe-substrate distance and switch off the SF device. With the aid of a CCD camera, we located the target cell and moved the probe on it, further sample-probe distance minor adjustment are done here, see Fig. 1 for a sketch of the system. Illumination is given from underneath through a transparent cell holder with a λ=442 nm, 200mW Laser. The probe, a pencil-type 100 nm apertured optical fiber [11], is placed over the chosen target point on the cell membrane. The probe substrate distance is fixed at 500 nm, since the cell thickness in culture should not exceed 200–300 nm [3], the probe is located about 200–300 nm off the cell membrane. The light in the proximity of the sample is coupled to the optical fiber and then it is detected by a photomultiplier at the other end of it. Refraction index of the cell material and surrounding liquid are both very similar to water, therefore optical contrast is reduced. The optical mechanisms involved in the sample-probe coupling are complex. As far as we know, a rigorous complete optical model does not exist in literature, however, it is accepted [8–10] that optical signal is strongly dependent with sample-probe distance. This distance dependence is responsible for the optical contrast that we are observing.

 

Fig. 1. Schematic representation of the core of the test system. Laser is a He-Cd 442 nm, the objective lens is a Nikon ELWD 40x with 0.6 Numerical Aperture focusing on the fiber tip.

Download Full Size | PDF

Blue light is effective in treating some diseases [12, 13]. The susceptibility of cells to blue light depends on several factors, and less susceptibility is mostly related to their healthy conditions. During our experiments we monitored the illuminated area with the non irradiated regions of the sample without observing differences or particular damage of the cells, thus we assume that the laser light does not affect the cell conditions that are relevant to the measurements.

Undifferentiated rat pheochromocytoma PC12 cell line were routinely cultured in 100 mm-diameter plates in D-MEM/F-12 (Dulbecco’s modified eagle medium: nutrient mixture F-12 Ham 1:1, GIBCO) with L-glutamine, sodium bicarbonate and pyridoxine hydrochloride, supplemented with 10% heat-inactivated fetal bovine serum (FBS, JRH Biosciences, Lenexa) and 5% horse serum (GIBCO) in a humidified atmosphere of 5% carbon dioxide and 95% air at 37°C. For the experiments, PC12 cells were plated onto 35 mm-diameter collagen-coated dishes in fresh growth medium and maintained in the humidified incubator until they reached ~ 90% confluence. Medium was changed every 3 days. Then the culture was transferred to the optical setup for examination. The area surrounding the sample was maintained at ~ 37°C in a humidified atmosphere by circulation of warm humid air under a dark plastic curtain that surrounded the optical system to guarantee the cell survival in optimum conditions during the experiments. Cell viability was assessed by trypan blue dye exclusion prior to all experiments. Cells with compromised cell membrane appeared blue due to accumulation of dye, and were considered as dead. Healthy cells with ~ 96.0% viability were employed in all the experiments here reported.

3. Experimental results and discussion

Nanomechanical activity of live cells wall was demonstrated successfully using AFM [14]. The main difference between AFM measurements and our approach with a SNOM setup is that the latter is a purely optical method; as mentioned above, we work in free running conditions, therefore there are no contact forces between probe and cell membrane, and measurements are possible even on very soft material in their natural state. The ability of PC12 to extend long, branching neuronal-like processes, that yeast doesn’t possess, could provide important evidences and diagnosis on the conditions of mammal living cells that could be of interest in bio-medical field.

Previous reports [7] clearly demonstrated the applicability of SNOM on detecting tiny and sudden vertical vibrations on the membrane of undifferentiated PC12 cells that could be associated to the physiological activity of the cell, such as dynamic of cytoskeletal system and/or cyclic duplication and division. PC12 in their physiological status were used as a control sample representing alive and healthy cells. As in other optical microscopes, we are aware that the optical signal we record does not contain pure morphological information. Differences in optical coupling could arise also by other phenomena in the cell body, as local modifications in the index of refraction or movement of inner cellular structures. In the hypothesis that the morphological displacement takes the major role in our measurements, we can estimate the vertical sensitivity of our recordings. As demonstrated in previous tests [6, 7] using contractions of cardiac myocytes that are known to extend vertically approximately half a micrometer [15, 16], we could observe the beating by a non calibrated voltage signal. The strongest peaks observed where around 5 volts base to peak. Considering a minimum discrimination of 2 millivolts, due to electronic analog to digital conversion, the minimal cell displacement is well below one nanometer.

These considerations are intended to be only as estimation, particularly because they assume linearity between signal and contraction displacement. This linearity is not guarantee, due to the complexity of the phenomena involved. However, we believe this is a reasonable approximation of the order of magnitude of the vertical sensitivity of our instrument.

Because of this extreme sensitivity in near-field conditions, we observed minute vibrations in the optical signal corresponding to 1 to 10 nanometer [6, 7] maximum membrane displacement. In Fig. 2, a sample of signals for the three cell conditions is reported: control, NGF and necrosis [2(a) 2(b) 2(c)]. On the right of each graph the corresponding Fourier transformation is plotted [2(d) 2(e) 2(f)]. Treatment of undifferentiated PC12 with 100 ng/ml NGF for 1 day, resulted in the dynamics plotted in 2(b) and 2(e). This dynamic is characterized by tiny and sudden vertical vibration of the cells, different from what observed in controls [2(a) 2(d)]. This can be associated to the dynamic of PC12 undergoing to neuronal differentiation. Fourier spectrum is revealing a shift to higher frequencies and also a more complex and lively structure than control cells even at frequencies less than one Hertz (figure insets).

Actually, the optical configuration in which we work do not allow us to know if the optical variations that we observe are generated by movements inside the cell body or solely by cell membrane displacements, or a combination of both phenomena. As noticeable in Figs. 2(a), 2(b) and 2(c), residual high frequency optical signal is present superimposed to the lower frequency signal. This high frequency signal appears to persist in necrotic conditions [Fig. 2(c)], and we suspect that it is mainly due to background noise presumably due to Brownian motion of impurities in the culture media.

Exposure of neurons to high concentration of hydrogen peroxide (H₂O₂) induce necrotic cell death [17]. In our system, confluent PC12 treated with 1 mM H₂O₂ for 4 hours (necrotic stimulus) resulted in a approximately 50% of cell death, as measured by tetrazolium bromide (MTT) assay (data not shown). Thus, this concentration at this incubation time was used for our tests with the SNOM. As visible in the time profile (c) after the treatment the membrane vibrations look less active and Fourier spectrum (f) also confirm a shift to lower frequencies revealing that the necrotic process is proceeding.

 

Fig. 2. Time profile of cell recordings for three different cell conditions: normal, NGF and necrosis. (a) control; (b) 24 hours NGF; (c) 4 hours H₂O₂. Time frame is of 100 seconds total for each recording, sampled at 0.01 sec (10.000 points per plot). Using a numerical algorithm we derived the Fourier spectrum for each sample: (d) control; (e) 24 hours NGF; (f) 4 hours H₂O₂. Lower frequencies are plotted in the smaller insets for clarity. Obvious differences are visible in the time profile and in the Fourier spectrum. Vertical scale is Volts for the time profiles and Volts/Frequency (Hz) for the Fourier plots.

Download Full Size | PDF

Since these nano-scaled minute dynamic membrane movements can be thought as a sort of voice of the living cell, we were curious to hear the acoustic component of the signal; we interfaced the photomultiplier to a conventional speaker and we could actually hear these sounds in real time from the live samples during measurements. As shown in Fig. 3, the acoustical part of the spectrum (over about 5 Hz) presents fewer differences compared with lower part of the spectrum. Treatment with NGF (b) was acoustically recognizable, probably due to stronger Fourier peaks at about 8 Hz and 27 Hz compared to the control.

 

Fig. 3. The acoustical part of the spectrum of the membrane dynamics studied. (a) control; (b) 24 hours NGF; (c) 4 hours H₂O₂. Actual sound was generated by means of a transducer. Frequencies from 5 Hz were perceivable and different cell conditions could be distinguished by simply hearing the sound, see text for discussion. Vertical scale is Volts for the time profiles and Volts/Frequency for the Fourier plots.

Download Full Size | PDF

4. Summary and conclusion

Our results show the high-sensibility of a SNOM system to detect tiny vertical movement on the cell membrane of PC12 in which membrane movement is not well known and/or detectable. We demonstrated the ability of SNOM to discriminate dynamics associated to different crucial cell phenomena of the cell’s life, such as cell cycle and cell death, analyzing the cells unperturbed in their liquid.

The different characteristics observed in different physiological status, clearly shows that these minuscule membrane movements are actually associated to the cell physiological condition. These signals were observed along the culture, independently of the cell observed and for the majority of the time. The temporal profile of the signals was also reproducible.

The clear differences detected in different physiological circumstances, demonstrate that nanometer scale membrane movement do actually convey information related to the life conditions and current activities of the cell. We could measure these vibrations instrumentally and we were also able to hear their acoustical components. If these “voices” are simply the mechanical result of internal cytoskeleton rearrangements or if there is a more deep “language” to be discovered, is beyond the aim of this report.

Our observations offer a unique approach to investigate live samples, and they may provide important evidence and diagnosis on the conditions of living cells.