Pulsed laser activated impulse response encoder (PLAIRE): sensitive evaluation of surface cellular stiffness on zebrafish embryos.

Mechanical properties of cells and tissues closely link to their architectures and physiological functions. To obtain the mechanical information of submillimeter scale small biological objects, we recently focused on the object vibration responses when excited by a femtosecond laser-induced impulsive force. These responses are monitored by the motion of an AFM cantilever placed on top of a sample. In this paper, we examined the surface cellular stiffness of zebrafish embryos based on excited vibration forms in different cytoskeletal states. The vibration responses were more sensitive to their surface cellular stiffness in comparison to the Young's modulus obtained by a conventional AFM force curve measurement.


Introduction
Cells use both intrinsic and extrinsic mechanical stimuli to regulate their fundamental functions [1,2]. In a tissue, cells are mechanically interacted with the extracellular matrix or the neighboring cell through anchoring junctions composed of multiprotein complexes, and anchoring junctions are linked to the cytoskeleton under the membrane [3]. The cytoskeletal contraction generates a tension and transmits force as mechanical stimuli to the surrounding cells, which affects biological responses such as migration and proliferation [1][2][3]. Moreover, collective cell movements arisen from coordinated contractile processes are one of the key mechanisms for tissue shaping and morphogenesis [4]. For example, a relief of overall mechanical tension is involved in an embryonic transition from a spherical to an elongated form [5]. In this context, measuring mechanical properties and evaluating transmitted mechanical forces within a tissue are of importance to elucidate how the mechanical forces are translated into complicated cellular processes.
Tensile and indentation characteristics are often used to measure the elastic modulus of biological objects. Especially, nanoindentation using a cantilever of an atomic force microscope (AFM) are widely applied to small biological samples [6,7]. This approach directly detects sample mechanical responses, however indentation methods only reflect average elasticity of the entire sample under the cantilever. As other options, several types of optical elastography such as Brillouin microscopy and optical coherence elastography (OCE) have also been advanced over the last decade in biological and biomedical fields [8][9][10][11]. Brillouin microscopy is based on the frequency shift of the Brillouin scattered light which is related to Young's modulus of the sample, and this technique enables to acquire three-dimensional elasticity distributions at a submicron resolution [9,10]. OCE visualizes mechanical properties of the soft tissue from sample displacement based on optical coherence tomography with axial resolution of typically 5 to 10 microns [8]. So far, many OCE configurations has been reported in terms of sample displacement measurement and force loading methods [8,11,12]. Among them, spectral-domain OCE is promising for medical applications because of its high image acquisition rates and displacement sensitivity [13].
Recently, we developed a technique that we have named the pulsed laser activated impulse response encoder (PLAIRE), which permits analysis of mechanical properties of submillimeter scale soft objects. In PLAIRE, an elastic wave is excited on the target by adding an impulsive force exploited by focusing a femtosecond laser beam in water. When an intense near-infrared (NIR) femtosecond laser beam is focused into an aqueous solution through a microscope objective lens, efficient non-linear ionization represented by multiphoton ionization leads to a micro explosive phenomenon at the laser focal point. Along with the explosion, rapid expansion and contraction of a cavitation bubble generates hydrodynamic flow to the surrounding area which acts as an impulsive force [14,15]. As the laser pulse is not directly focused on biological samples, it is able to excite the elastic vibrations without damaging.
The loading force and area of the femtosecond laser impulse can be tuned from a few micrometers to tens of micrometers, and from kPa order to a tissue destructive level by varying an input laser pulse energy. The temporal profile of the loading force induced by a single laser pulse is in tens of microseconds that corresponds to 50-100 kHz. Its loading frequency can be also chosen from a single pulse up to a repetition rate of the laser system. Such spatiotemporal and force loading controllability is one advantage of this technique to measure responses to various mechanical stimulation.
By the action of the impulse, deformation and different types of vibrations such as axialpressure waves, transverse-shear waves and surface acoustic waves (SAWs) should be excited within a sample. Those deformation and vibration amplitudes are directly detected with a nm scale sensitivity as a motion of the cantilever of an atomic force microscope (AFM) placed on the sample. PLAIRE gives a spatial distribution of mechanical properties of the sample depending on the excited vibration modes and wave propagating pathways although PLAIRE is basically not an imaging technique and it has much less spatial resolution compared to Brillouin microscopy and OCE. Another aspect of the AFM based vibration detection is that it allows easy comparison between vibration and static indentation responses of the sample. The relation between vibration and static indentation responses would be important to connect local and global behaviors against mechanical stimulations. In our previous demonstration of PLAIRE, we detected SAWs on calcium alginate (CaAlg) spheres as a phantom for small biological objects, and succeeded in extracting surface elasticity of CaAlg from phase velocities of SAWs [16,17].
In this paper, we applied PLAIRE to evaluate surface cellular stiffness of zebrafish embryos. We confirmed excited vibration forms on zebrafish embryos were altered after modification of the cytoskeletal structure of the surface cell layer. Higher sensitivity of PLAIRE signal to the surface cellular stiffness were demonstrated compared to a conventional force curve measurement made by the indentation method with the AFM.

Preparation of zebrafish embryos
Zebrafish embryos were obtained by mating of wild-type zebrafish (Fig. 1). All zebrafish experiments were performed with the approval of the Animal Studies Committee of Nara Institute of Science and Technology.
The entire embryo at the 90% epiboly stage, the yolk part and a single surface cell were used to characterize Young's modulus of the complex embryo based on that of its individual components. A two-cell stage embryo was used in place of the yolk part in the 90% epiboly stage. These two stages of zebrafish embryos were dechorionated and fixed along with a small amount of embryo  The protocol for the single cell separation from the 90% epiboly embryo has been reported elsewhere [18]. In brief, the dechorionated 90% epiboly was mechanically dissociated into single cells by mild shaking in 1 mL cell culture medium of CO 2 -independent DMEM/F12. Yolk proteins were removed by washing twice with 1 mL fresh medium followed by centrifugation at 1000 rpm for 2 min. Dissociated cells were seeded on a glass bottom culturing dish (Matsunami).

Modification of the surface stiffness of zebrafish embryos
The embryo was treated with 10 µM cytochalasin D (Wako) in embryo buffer containing 0.1% of dimethyl sulfoxide (DMSO, vehicle) for 60 min to modify the mechanical conditions of the surface cell layer of the zebrafish embryos. Cytochalasin D inhibits polymerization of actin proteins which are the main components of cytoskeleton [19]. DMSO treated embryos were used as negative controls. After the cytochalasin D treatment, cell architectures based on polymerized actin proteins were visualized with Alexa-488-phalloidin (Thermo Fisher Scientific). The embryo was fixed with 4% paraformaldehyde for 10-16 h and washed with 0.1% Triton X-100 in phosphate buffer saline (PBT). The embryo was then incubated with 0.5% PBT for 60 min, and with Alexa-488-phalloidin (1:100) in 0.1% PBT at 4℃ for 24 h, and rinsed with PBT several times. The embryos were flat-mounted in Vectashield mounting medium (Vector Laboratories) to avoid fluorescence bleaching. Fluorescence images of the surface cell layer were acquired by a confocal microscope (Zeiss, LSM700) with 488 nm excitation.

AFM force curve measurements
An AFM force curve was obtained using the force-distance curve mode of the AFM (Pacific Nanotechnology, Nano-R2). The glass slide substrate or glass bottom culturing dish was set on an inverted microscope (Olympus, IX71) coupled with the AFM (Fig. 2(b)). An upright low-magnification microscope (Olympus, MVX10) was also installed on the microscope system to adjust a position of the zebrafish embryo under AFM cantilever. A glass sphere with a diameter of 20 µm was glued on top of a tipless AFM cantilever (NANOSENSORS, qp-SCONT; k = 0.006). The spherical contact tip design of the cantilever avoids damaging the sample surface and simplifies calculating Young's modulus by Hertz's contact theory given by where E is Young's modulus of the sample to be determined, F is loaded force on the sample, d is indentation depth, R is the radius of a glass bead bonded to the cantilever (10 µm) and v is Poisson's ratio of the sample. F and d were obtained from the force curve measurements, and v of 0.49 was used as a general value of soft materials [20].

PLAIRE setup
A single laser pulse from the amplified femtosecond Ti:sapphire laser (Spectra-Physics, Solstice-Ace; 800 nm, 100 fs, 20 Hz) was introduced into the inverted microscope and focused through a 10X objective lens (Olympus; NA, 0.25) at a distance of 5 µm from the embryo in the biggest cutting plane parallel to the glass slide, (Figs. 2(a) and (b)). The single laser pulse was extracted using a mechanical shutter with a 50 ms gate time. The pulse energy was tuned to 1 µJ/pulse using a neutral density filter. A small part of the laser pulse was reflected by a glass plate inserted into the optical path, and detected with a silicon photodiode to record a laser irradiation timing. A temporal profile of vibrational amplitude on the embryo excited by the femtosecond laser impulse was detected as oscillations of the cantilever (NANOSENSORS, TL-CONT; k = 0.4 N/m) placed on the top ( Fig. 1(a)). The voltage difference between the upper and bottom sides of a quadrant photodiode inside the AFM was monitored with an oscilloscope (GW instek, GDS-3154), and the signal was converted to the cantilever displacement using a linear coefficient of 1.7 mV/nm. Duration of the impulsive force action was roughly estimated within 5 µs from the pressure wave directly reaching the AFM cantilever through the water without any sample present.

Young's modulus of zebrafish embryos
Young's modulus as an elastic property of the zebrafish embryo was first obtained by AFM force curve measurements which represent the relationship between an indentation depth and reaction force from the sample. Figure 3 shows representative force curves measured on the top of the zebrafish embryo at the 90% epiboly stage, at the two-cell stage and for the separated single surface cell. From slopes of their tangents, we see the yolk has the smallest elasticity, and the single cell from the surface cell layer has the biggest among the three embryo components. Using Hertz's contact theory, we calculated the Young's modulus values and they are listed in Table 1. Young's modulus of the single cell corresponds to the reported value, which indicates the experimental operations were done properly [21]. Interestingly, Young's moduli of the 90% epiboly and the two-cell stages are two orders of magnitude smaller than that of the single cell. The surface of the 90% epiboly embryo is covered with a surface cell layer, however information about the surface stiffness is hidden by the small Young's modulus of the yolk. Thus, the force curve measurement gives only the combined Young's modulus but not the modulus of the individual components. Moreover, standard deviations of the obtained Young's moduli for different measurements are about half of their average values. This could be an inherent problem of point-by-point measurements when using the small AFM tip for inhomogeneous samples.

Effects of surface stiffness modification on the AFM force curves and PLAIRE measurements of zebrafish embryos
To evaluate influences of surface cellular states of zebrafish embryos to the AFM force curves and PLAIRE measurements, we performed a cytochalasin D treatment. Figure 4(a) shows confocal fluorescence images of actin structures of surface cell layer at the 90% epiboly embryos treated by cytochalasin D and its control (treated only by DMSO). The actin fragments become smaller and spread over the cytosol. This is a typical effect of cytochalasin D caused by an inhibition of actin polymerization [19]. In previous reports, effects of cytochalasin D on cellular stiffness was investigated by AFM force curve measurements. For many types of cultured cells such as fibroblast and osteoblast, cellular Young's modulus was decreases by a factor of two to three after cytochalasin D treatments [19,22,23]. Thus, prior to PLAIRE experiments, we examined the effects of the surface elasticity modification with cytochalasin D on the force curve of the embryo. As compared in Fig. 4(b), there is no significant difference in the average Young's modulus between the control and the cytochalasin D treated embryo. In addition, the variation of Young's modulus is large and it is difficult to distinguish two surface states. As discussed in section 3.1, the Young's modulus obtained by the force curve measurement did not reflect the stiffness of the surface cell layer but the yolk part, therefore the force curve measurement was not sensitive to the surface modification with cytochalasin D for the embryo. Next, we applied PLAIRE to evaluate the effects of the surface elasticity modification on excited vibrations in the embryos. Temporal profiles of AFM amplitude after femtosecond laser irradiation are presented in Fig. 5(a). A sharp peak appears at t = 0 because of direct scattering of the femtosecond laser beam to the photodiode in the AFM. In our previous PLAIRE experiments using CaAlg spheres with a size of about 1 mm, we observed a broad, indistinct peak in the range between 30 to 100 µs which we attributed to SAW [16,17]. In contrast, here we cannot see a clear signature of propagating waves such as axial-pressure waves, transverse-shear waves and SAW for either of the embryo samples. One possible reason is the inhomogeneous surface cell layer containing many boundaries causes multiple reflections and rapid decay of propagating waves. Nevertheless, there is a clear difference between the vibration forms with and without the cytochalasin D treatment. There is no specific feature for the control but a sine curve-like vibration of about 10 µs cycle (100 kHz) is seen for the embryo with the cytochalasin D treatment.
This mode continues long as elastic waves through the embryo, and its amplitude gradually decays.
The acquired vibrations in a frequency domain are shown in Fig. 5(b). For the embryo with the cytochalasin D treatment, a large band from 50 to 100 kHz appears and is reproducible; it corresponds to the frequency of the oscillation in Fig. 5(a), and higher frequency components from 200 to 700 kHz decrease relative to the control. In fact, appeared bands in 50-100 kHz and 200-700 kHz are similar to the third and higher order bending modes of the natural frequency of used AFM cantilever in water although there are frequency shifts because of the glass sphere attached on the tip of cantilever (See Table S1 in Supplement 1).
From these observations, we regard the bands in 50-100 kHz and 200-700 kHz as the bending modes of the AFM cantilever coupled with spheroidal mode vibrations of the entire embryo. It has been reported that the oscillation frequency of a viscous layer or droplet in contact with a substrate decreased when its surface tension became smaller [24,25]. Thus, it is considered that surface tension of the embryo was relaxed when the cytoskeleton of the surface cell layer was fragmented by cytochalasin D. Consequently, the resonant frequency between the bending modes of the AFM cantilever and the spheroidal modes of the entire embryo would shift from the range in 200-700 kHz to 50-100 kHz. In addition, the time scale of a rise and decay of the laser induced impulse is also in the range from 50 to 100 kHz that would excite this coupled resonant mode efficiently. The presence of the resonant oscillation would be another factor why propagating waves are not clearly detected in this system.
We also note here that higher frequency components in 200-700 kHz are not always same regardless of whether the embryo is treated by the cytochalasin D or not. This unreproducible responses in this frequency range could be due to heterogeneous distribution of the mechanical properties on the embryo surface although it still needs for discussion.

Sensitivity of the AFM force curves and PLAIRE measurements to the surface stiffness of zebrafish embryos
In section 3.2, we discussed PLAIRE signal of the zebrafish embryo is altered as a function of surface cellular stiffness. Hence, we propose to use a relative amplitude of the frequency band in the PLAIRE signal as an indicator of the surface stiffness transformation of zebrafish embryos. In this sense, a comparison of the signal sensitivity of PLAIRE and force curve measurements to surface stiffness will be helpful for the practical use of PLAIRE. For PLAIRE, we calculated the relative amplitude as A 50−100kHz /A 200−700kHz , where A 50−100kHz is the average amplitude from 50 to 100 kHz and A 200−700kHz is that from 200 to 700 kHz. Then, the ratio of the relative amplitudes between control embryo and cytochalasin D treated embryo is 2.6 ± 0.8. For the AFM force curve, the ratio of the average value of Young's modulus between control embryo and cytochalasin D treated embryo is 1.4 ± 0.8. Based on these two signal ratios, we conclude that PLAIRE is roughly two times more sensitive than the conventional AFM force curve measurements regarding the change of surface cellular states.

Conclusion
In conclusion, we applied PLAIRE to evaluate surface stiffness of a zebrafish embryo, and compared its sensitivity with the Young's modulus obtained by conventional AFM force curve measurements. Separately measured individual Young's modulus values of zebrafish components proved the indentation method did not reflect surface mechanical properties but was rather an average of the properties of the components of the embryo. Therefore, the indentation method was not sensitive to the surface stiffness when it had been modified by chemical treatment. In contrast, when the femtosecond laser impulse was applied to the embryo, excited vibration forms were clearly distinguished after modification of the surface stiffness because the amplitude of the excited oscillation would be directly linked to surface tension of the embryo.
Our present results showed the possibility to analyze elastic properties of the samples with core-shell like structures. Therefore, application of this technique is expected to provide better understandings of complex mechanical properties of small biological objects such as embryos and tissues.