2.1.

DHM

DHM^{26, 27, 28, 29, 30} is an innovative interferometric microscopy technology enabling complete complex optical wavefront retrieval from a specimen, illuminated either in reflection or in transmission. The working principle relies on the acquisition of a hologram, generated by the interference between a so-called object wave $\left(O\right)$ and a reference wave $\left(R\right)$ , by a digital camera, such as a charge-coupled device (CCD). Thereafter, the complex wavefront in the acquisition plane can be entirely retrieved by software processing,^{31, 32, 33, 34} both in amplitude and noticeably in quantitative phase,^{22, 23} and eventually digitally propagated to the in-focus plane by common Fresnel, convolution, or angular spectrum algorithms.^{28, 30, 35}

Thanks to its noninvasive and noncontact nature, DHM has already been widely applied in transmission configuration to investigate transparent samples in real-time, mostly regarding life sciences applications under $1\text{-}\mathrm{g}$ laboratory conditions, like monitoring neurons under hypotonic stress,¹⁷ red blood cells,²¹ cancerous cells,^{18, 19} dynamic cytoskeletons changes,²⁰ or cell-division dry mass evolution.³⁶ In transmission, the measured quantitative phase signal $\phi$ is linked to the so-called integrated OPL (i.e., the optical thickness) through the sample as in

The study of living cells often requires their submergence in culture medium either inside a closed immersion or open perfusion chamber as illustrated in Fig. 1 . In such a configuration, the measured phase data correspond to the phase difference $\Delta \phi$ between ${\phi }_1$ and ${\phi }_2$ (see Fig. 1) as expressed in Eq. 2, where it is proportional to the refractive index mismatch between the sample and the immersion medium

Fig. 1

Schematic of the usual transmission DHM sample configuration with a closed or open perfusion chamber, and the corresponding measured phase signals ${\phi }_i$ with $i=\mathrm{1,2}$ . $n_{\mathrm{m}}$ , culture medium refractive; $n_{\mathrm{ic}}$ , mean intracellular refractive index; $h_{\mathrm{c}}$ , cellular thickness.

As can be observed in Eq. 2, the DHM-measured phase information gives rich insights about both intracellular refractive index, proportional to intracellular content density (water, proteins, etc.), and the actual cytomorphology. It is typical to reach $3\times {10}^{-4}$ precision on refractive index and about $10\text{-}\mathrm{nm}$ axial accuracy with transmission DHM,¹⁷ thanks to the interferometric precision of the technique. The lateral resolution remains diffraction limited by the microscopy-objective (MO) numerical aperture (NA), as in classical microscopy. If needed, so-called decoupling procedures to independently retrieve both refractive index and topographical values from the phase measurement exist using for instance, sequential culture medium exchange,¹⁷ fixed chamber channel height squeezing the cells,¹⁹ or recently true real-time decoupling by dual wavelength using dispersion with a dye-enhanced surrounding medium.³⁷[ME1]

2.2.

DHM Setup

The RPM-DHM prototype used in this study was a transmission setup based on a Mach–Zehnder interferometer configuration, as depicted in Fig. 2 . The laser source was a $650\text{-}\mathrm{nm}$ wavelength semiconductor laser diode with an optical output power of $1\mathrm{mW}$ , with a coherence length of $\sim 300\mu \mathrm{m}$ . Briefly, at first the beam was collimated and divided by a first beamsplitter (BS) into two interferometer arms, the object wave $O$ , and the reference wave $R$ . A neutral density filter (NF) was inserted into the object arm to balance the intensity ratio between the arms. A condenser lens focuses the object beam onto the sample, where diffracted light was collected by a $60\times 0.85\mathrm{NA}$ MO to be recombined with the reference wave in off-axis configuration (slight angle between the propagation vectors) generating the hologram recorded by a digital CCD camera (Basler A101f, Basler Vision Technologies, Ahrensburg, Germany ). The delay line (DL) in the reference arm was designed to match the OPL of both arms in order to get interferences with the reduced coherence length. The use of low temporal coherence light discards most parasitic interferences contributions and enables it to reach a single-shot axial accuracy in phase of roughly $2–3\mathrm{deg}$ . The reference beam was then spatially filtered and expanded within the beam-expander (BE) section.

Fig. 2

Custom-designed transmission DHM for the RPM: (a) Optical setup. M, mirror; L, lens; BS; NF BE; O, object wave; R, reference wave; DL; CL, condenser lens; and MO. (b) Actual experimental microgravity configuration, with the DHM mounted on the inner frame of the RPM and a built-in minicomputer fixed to the platform as well, partially counterbalancing the weight of the DHM.

Given the harsh experimental conditions onto the RPM, it is clear that the use of a dry-configuration MO was mandatory. Indeed, although the reachable resolution is theoretically better with high-NA oil-immersion MOs, the platform for the DHM on the RPM was obviously going to continuously rotate, which prevented using immersion oil or any other liquid. The retained microscopy option was therefore a dry-configuration $60\times 0.85\mathrm{NA}$ MO, providing a field-of-view of about $65\times 65\mu {\mathrm{m}}^2$ on a $512\times 512\text{pixels}$ region of interest on the CCD camera, with a lateral resolution in the $500\text{-}\mathrm{nm}$ range.

A built-in minicomputer integrated in the DHM was in charge of controlling the camera acquisition parameters, as well as to acquire and store the recorded holograms during the experiment. The onboard computer was configured with a low-consumption CPU (Intel Core Soleo $1\mathrm{GHz}$ ; SDRAM $667\mathrm{MHz}$ ) and a compact flash memory of $8\mathrm{GB}$ at $20\mathrm{MB}∕\mathrm{s}$ , enabling a theoretical recording rate of $78\text{images}∕\text{second}$ for $512\times 512\text{pixels}$ holograms (the used CCD camera was limited to $25\text{images}∕\text{second}$ ). However such a high frame rate was not necessary for the experiments, and effective framerate for both $1\text{-}\mathrm{g}$ control and microgravity recorded holograms was chosen as $1\text{image}∕\text{second}$ . At the end of experimental sessions, all locally stored data was uploaded to an external workstation for hologram reconstruction and all experimental data postprocessing. Hologram file size was set to $512\times 512\text{pixels}$ , recorded in $8\text{-bit}$ TIFF format. An overview of the actual experimental configurations onboard the RPM is depicted in Fig. 2. It is important to note that the DHM position on the RPM platform was designed so that the sample position was exactly located at the RPM rotation center, in order to fulfill microgravity conditions for the specimen and eliminate centrifugal acceleration. Finally, the DHM-investigated C2C12 cells were cultivated at $37°\mathrm{C}$ in special custom-designed flasks ( $\mu$ -slide I, Ibidi GmbH, Munich, Germany), which purposely had the same thickness as standard coverslips $\left(0.17\mathrm{mm}\right)$ .

2.3.

Random Positioning Machine (RPM)

The RPM or three-dimensional clinostat is a laboratory instrument to simulate microgravity condition. Originally, the machine was developed by Hoson ³⁸ and manufactured by Dutch Space, Leiden, The Netherlands. Samples mounted on a platform randomly change the position in the 3-D space on the machine, controlled by dedicated software running on a personal computer. The movement of the experimental platform suspended in the center of two cardanic frames is realized by two independently running engines. These engines are controlled by feedback signals from encoders mounted on the motor axes. The RPM was operated as 3-D clinostat in a random walk (basic mode) with a rotation speed of $60\mathrm{deg}∕\mathrm{s}$ in a temperature controlled room $(37\pm 1°\mathrm{C})$ .

2.4.

Fluorescence Microscopy Investigations on Cells Exposed to Microgravity

As already mentioned, all experiments were carried out on actin-GFP transfected C2C12 mouse myoblasts. Thus, actin filament structures were visible under fluorescence microscope (blue excitation, green emission) without additional labeling. The C2C12 cells were myogenic progenitor cells, capable of differentiating into myotubes in vitro on serum withdrawal.³⁹ The fusion of myoblasts with existing myotubes is what gives rise to the adult skeletal muscle fibers observed in the animal.^{40, 41}

C2C12 myoblast cells were cultured in an incubator at $37°\mathrm{C}$ , 5% $\mathrm{C}{\mathrm{O}}_2$ in RPMI 1640 (Cellgro, Mediatech, Inc., Manassas, Virginia ) supplemented with 10% fetal calf serum (PAA, Cölbe, Germany), 1% penicillin/streptomycin (PAA, Cölbe, Germany), and 0.6% geneticin (GIBCO, Auckland, New Zealand). Antibiotic treatment ensures that only GFP-tagged cells are cultivated because they also express the antibiotic resistance. This way, a loss of fluorescent signal can be reduced. The cells were fed twice a week. Before seeding the cells in different cell flasks (Nunc, Roskilde, Denmark) for the investigation on the RPM, they were detached from the culture flask by exposure to 0.25% trypsin/EDTA (Invitrogen, Eggenstein, Germany) for $2\min$ at confluency of maximum 80%. The RPM rotation velocity was then set to $60\mathrm{deg}∕\mathrm{s}$ and sample images were taken after various time points of simulated microgravity exposure. Ground controls $\left(1\mathrm{g}\right)$ were placed on the frame of the RPM in order to subject all the samples to the same temperature variations and mechanical vibrations as the microgravity samples.

Some microgravity-induced morphological and cytoskeletal change measurements under simulated microgravity were carried out on fixed cells. At particular exposure periods to microgravity and $1\mathrm{g}$ , cells were harvested and immediately fixed by 4% paraformaldehyde for $15\min$ before DAPI staining. The following protocol was used: $5\text{-}\min$ incubation in Triton X-100 (0.1%) and then $5\text{-}\min$ exposure to DAPI dye $(50\mathrm{ng}∕\mathrm{mL})$ . Rinsing of the cells with phosphate buffered saline (PBS) was carried out in between all the steps described. At the end, the slides were embedded with Aqua Poly/Mount (Polysciences Inc., Warrington, Pennsylvania). Samples were observed thereafter by using an inverse epifluorescence microscope equipped with a $40\times 0.85\mathrm{NA}$ UplanApo MO and a $60\times 1.45\mathrm{NA}$ oil-immersion Ph PlanApo MO. All the fluorescence images were recorded in gray scale on both channels (DAPI and GFP filters) separately, then colored and merged together.

The perinuclear actin concentration was further investigated by extracting pixel intensity profiles from GFP images. Each profile had a fixed length of $80\text{pixels}$ (which is about the diameter of a nucleus) and was drawn from the edge of the nucleus in the opposite direction. In each time series, two profiles per image were taken from different directions (at least $20\mathrm{deg}$ apart). Four cells per group (microgravity or $1\mathrm{g}$ for a specific time series) were investigated.

2.5.

Protein Content Verification Procedure

C2C12 cells were cultured until sub-confluence had been reached. At this stage, the cells were cultured after exposure to normal and simulated microgravity conditions for 0, 5, 10, 30, and $60\min$ . Immediately afterward, the cells were trypsinised, centrifuged ( $1500\mathrm{rpm}$ , $5\min$ ), washed with PBS, and centrifuged again. At the end, pellets were stored at $-80°\mathrm{C}$ until further processing. For the analysis, the pellets were resuspended in lysis buffer (CellLytic-M Cell Lysis Reagent, Sigma, C2978) containing freshly done protease inhibitor cocktail (Complete Mini, Roche, ref 11 836 153 001). After a $30\text{-}\min$ incubation period at $4°\mathrm{C}$ , insoluble cells debris were removed by centrifugation ( $4°\mathrm{C}$ , full speed, $30\min$ ). Protein concentrations were determined afterwards by using the BC Assay Protein Quantification Small Kit (Uptima, UP40840B) according to manufacturer’s instructions. The quantification was performed using Microplate Manager Software (measurement filter: $540\mathrm{nm}$ , mix time $5\mathrm{s}$ ). BSA was used as a standard. Statistical analyses were carried out using the software SPSS (SPSS Inc., Chicago, Illinois).

3.1.

Morphological and Cytoskeletal Changes Observation by Fluorescence Microscopy

In comparison to control cells shown in Fig. 3 , exposure of C2C12 cells to simulated microgravity stimulated the growth of lamellipodia as depicted by a weblike organization of the actin cytoskeleton at the mobile edge of the cell in Fig. 3. An increase in lamellipodial outgrowth was observed after only $10\min$ of exposure to microgravity. The proportion of cells exhibiting such actin cytoskeletal changes increased with time exposure to microgravity up until $\sim 2\mathrm{h}$ , where a steady state in cell morphological changes seemed to be reached (data not shown).

Fig. 3

Widefield images of C2C12 cells at (a) $10\min$ $1\text{-}\mathrm{g}$ control condition, and (b) $10\min$ $0\text{-}\mathrm{g}$ exposure, with examples of profiles (red arrows) for determination of the microgravity-induced accumulation of actin around the nucleus. (Color online only.)

In addition to lamellipodial outgrowth, the actin distribution within the soma also changed in response to simulated microgravity; stress fibers became denser and more randomly organized and nonfilamentous actin began to accumulate around the nucleus [Fig. 3]. To evaluate the time course of actin accumulation around the nucleus, pixel intensity profiles were extracted and normalized to the maximum value. Two profiles were taken at different orientations (at least 20° between the two segments) as depicted by the red arrows in Fig. 3 at $10\min$ , $1\mathrm{h}$ , $2\mathrm{h}$ , and $4\mathrm{h}$ of exposure to simulated microgravity exposure. Indeed, given the time frame of fluorescent sample preparation described in Section 2.4, measurements of microgravity exposures of $<10\min$ make little sense and risk of induced bias are high. Significant differences in perinuclear complexity were detected (The Mann-Whitney U-test) until up to $2\mathrm{h}$ , as is demonstrated in Fig. 4 . The rate of decrease in actin concentration with distance from the cell nucleus was greater in $1\text{-}\mathrm{g}$ control cells than in microgravity. After $10\min$ of exposure to simulated microgravity, the difference between the two groups ( $1\mathrm{g}$ and microgravity) was significant in Fig. 4. At $1\mathrm{h}$ , the same significance was computed for in Fig. 4 and after $2\mathrm{h}$ in Fig. 4. At longer durations $(4–25\mathrm{h}),$ the cells seemed to adapt themselves and no noticeable difference was calculated as depicted in Fig. 4.

Fig. 4

Accumulation of actin around the nucleus at $t_{0\mathrm{g}}=\left(\mathrm{a}\right)$ $10\min$ , (b) $1\mathrm{h}$ , (c) $2\mathrm{h}$ , and (d) $4\mathrm{h}$ of simulated microgravity exposure times. Significant differences between 1 and $0\mathrm{g}$ cells were detected up to $2\mathrm{h}$ (Mann–Whitney U-test) with actin distribution around the nucleus apparently adapting after $4\mathrm{h}$ .

3.2.

In Situ DHM Monitoring of Cell Changes under Microgravity Condition

The main purpose of using DHM was to verify if microgravity-induced changes in the cytoskeleton commonly observed by fluorescence microscopy on fixed C2C12 cells (e.g., Figs. 3 and 4) could also be detected by DHM in real time on the RPM. This could enable the investigation of the cellular mechanisms responsible for the observed dynamic changes in the cytoskeleton. In addition to this, DHM potentially provides a much finer temporal resolution to investigate quick responses to microgravity compared to fluorescence microscopy, as in Fig. 4.

Changes in perinuclear optical thickness with time at microgravity were confirmed by DHM analysis. As can be seen in Fig. 5 , when compared to the epifluorescence micrograph of Fig. 3, the phase measurements changed immediately on exposure to microgravity. No significant changes occurred around the nucleus during the $1\text{-}\mathrm{g}$ control period $\left(8\min \right)$ , whereas microgravity readily initiated cytosolic modifications. Switching on the RPM caused a continuous accumulation of proteins around the nucleus, translated into a phase increase (Fig. 5). The first apparent modifications were already detected after $2\min$ of exposure to microgravity and persisted to accumulate throughout the measured period of $1.5\mathrm{h}$ . The current DHM platform does not distinguish between actin or other proteins. Therefore, the series of DHM images shown in Fig. 5 represent a total accumulation process from potentially different sources. Because of technical reasons, noticeably RPM onboard power supply instability during extended operation times, DHM investigation for a period of longer than $1.5\mathrm{h}$ were not possible with this experimental prototype.

Fig. 5

DHM quantitative phase images in perspective 3-D view, showing the cell thickening process around the nucleus under simulated microgravity: Microgravity was switched on after a control period of $8\min$ . FOV: $51\times 58\mu {\mathrm{m}}^2$ .

Growth of lamellipodia under simulated microgravity conditions was also monitored with DHM, as shown in Fig. 6 . During the initial test period ( $10\min$ at $1\mathrm{g}$ ) no changes of leading-edge morphology or OPL were apparent. There was no detectable change in lamellipodia area, and the edges of the lamellipodia retain their interaction zone with the substrate. After switching on the RPM, on the other hand, much of the lamellipodial interactions were retracted and filopodia extended in their place. After $2\min$ of microgravity, the phase-shift measurement doubled in this area, as can be clearly seen in Fig. 6 time-series images. The structure continued to grow until the end of the recording period of $20\min$ . Obviously, the microgravity environment induced a reorganization of the cell’s leading edge. These observations are in good agreement with the results obtained with the traditional fluorescence microscope (Fig. 3) in that actin was shown to organize into dense bundles at the leading edge of the cell (typical filopodial actin scaffold) when exposed to microgravity.

Fig. 6

Quantitative phase images of the evolution of lamellipodium growth initiated by simulated microgravity, monitored in real time by DHM. Vertical scale in degrees, FOV: $32\times 16\mu {\mathrm{m}}^2$ .

3.3.

Estimation of Total Protein Amount

Epifluorescence microscopy, as well as DHM quantitative phase investigations of the C2C12 cells under simulated microgravity, reveals substantial changes of cell morphology in comparison to the $1\text{-}\mathrm{g}$ control group. These changes occur mainly within the first hour of microgravity exposure. In order to verify whether the changes are due to reorganization or de novo synthesis of cytoskeletal proteins, it is useful to measure the total amount of proteins after normal $1\text{-}\mathrm{g}$ and microgravity exposure. The protein content of C2C12 are determined after 5, 10, 30, and $60\min$ . As plotted in Fig. 7 , analysis of the probes reveal no statistically significant difference between all the different groups.

Fig. 7

Total protein amount of C2C12 cells under control ( $1\mathrm{g}$ , gray) and simulated microgravity conditions ( $0\mathrm{g}$ , white), indicating that no differences between the two different conditions are detected, even after $60\min$ . Values are expressed as mean protein $\text{concentration}\pm \mathrm{SD}$ .