Search
Close this search box.

Real-time study of spatio-temporal dynamics (4D) of physiological activities in alive biological specimens with different FOVs and resolutions simultaneously – Scientific Reports

Tracheal branching in Drosophila melanogaster

Figure 1
figure 1

The projection images at (11.11times) and (22.22times) magnification on XY plane. Some of the observed branches are pointed by arrowheads of different colours for different types of branches. DLT: Dorsal longitudinal trunk; TC: Transverse connective; aLLT: Anterior lateral longitudinal branch; pLLT: Posterior lateral longitudinal branch; TC: Transverse connective; GB: Ventral ganglionic branch. Line-plot—showing the variation of normalized intensity along a randomly chosen line from the same region of interest in both magnifications (as they marked in the figures)—is included as an inset. Scale bar: 50 (mu m).

Figure 2
figure 2

Mapping of tracheal branches in Stage 15 of embryonic development of Drosophila melanogaster (fixed sample) at two different magnifications ((11.11times) ((a) and (c)) and (22.22times) ((b) and (d)). Figures depict the tracing of tracheal systems that are observed at two different magnifications or resolutions ((c) and (d)). Scale bar: 50 (mu m).

Figure 1 presents the experimental results for experiments being conducted in a fixed sample (more specifically, Drosophila embryos at Stage 15 of embryonic development). Figure 1a and b give projected images (in xy-plane) corresponding to magnifications obtained at (11.11times) (detection Arm-I) and (22.22times) (detection Arm-II) respectively. For this magnified view, the detection arm (Arm-II) was kept fixed over the area as indicated by dashed lines shown in Fig. 1a). These projected images were generated from a sequence of 2D images covering the entire volume of the specimen while each stack of the 2D images was undertaken a sequence of image processing for improving the obtainable image quality (before taking the projection). The details with the associated results are provided separately in Supplementary 5. In the figure (Fig. 1), one observes the structures of the specimens with better clarity and higher resolution in higher magnification image that are clearly observable from the line-plot included as an inset in the figure (Fig. 1a). The line plot from the same region of interest from both magnifications shows that the (22.22times) magnification contains more information, especially from the narrow tracheal branches, than the (11.11times) magnification. There is a slight difference in the two images as they are corresponding to two different sides of the specimens that are not perfectly symmetric. The different branches (which are marked by different colour arrowheads) can be seen very clearly in Fig. 1 but for a better understanding of the tracheal branches they have been traced using the NeuronJ plugin in ImageJ. Hence in the figures (Fig. 2a and b), we depict the tracing of tracheal branches being overlaid on the projected images of the specimen. Figure 2c and d give the tracheal branching without overlying in the background. It is observed that the image obtained at magnification (sim 11.11times) can depict the entire embryo (but at a lower resolution) and we can discern the different transverse connectives of the tracheal branches from the dorsal longitudinal trunk, namely, DLT, TC, aLLT, and pLLT. The image at (22.22times) magnification gives better resolution (but at lower FOV). With this higher magnification, one can distinctly observe the fine branches or structures. Figure 3 presents the corresponding 3D views of the sample obtained at two separate magnifications ((11.11times) and (22.22times)) simultaneously at an instant of time. The arrowheads in Fig. 3 show the branches spread in the volume. The (360^circ) panoramic views of the 3D reconstructed images at both magnifications are given separately in Supplementary 6. These experimental results demonstrate the capability of our imaging system for imaging the entire volume of a given specimen at two separate fields of view (FOVs) and resolutions simultaneously or at one instant of time.

Figure 3
figure 3

3D reconstructed images of the tracheal system of Drosophila melanogaster (at Stage 15). The arrows indicate the dorsal branches spread in the volume.

Figure 4
figure 4

Tracing of tracheal branching of embryonic development for Sample I: Stage 11 to Stage 12 (I(aj)) and Sample II: Showing progression from late Stage 13 to Stage 15 (II(aj)). The magnified images of Sample II at t0 and (t0+60) min shows the different branches with labels where DLT: Dorsal Longitudinal trunk, DT: Dorsal tracheal branch, TC: Transverse connective, aLLT: Anterior lateral longitudinal branch, pLLT: Posterior lateral longitudinal branch. Scale bar: 50 (mu m).

Figure 4 presents the experimental results with experiments being performed in alive biological specimens (Drosophila melanogaster). Experiments were conducted in separate groups of live Drosophila embryos (more specifically, at different stages of embryonic development). Figure 4(I(a–j)) present images of the sample at embryonic development Stages 11 to 12 (say, Sample I) while Fig. 4(II(a–j)) depict images corresponding to the sample progressing from late Stage 13 to Stage 15 (say, Sample II). The embryo’s activity was recorded at an interval of 5 min for an entire duration of 60 min (for Sample I) and 95 min (for Sample II). In Fig. 4, we present only certain selective images—that are recorded at the time interval of 15 min— while the complete information of images corresponding to recording at an interval of 5 min is given in Fig. 5. Figure 4(I(a–e)) presents images obtained at (11.11times) magnification while Fig. 4(I(f–j)) corresponds to (22.22times) magnification (Sample 2). Similarly, Fig. 4(II(a–e)) depict images obtained at (11.11times) magnification while Fig. 4(II(f–j)) correspond to (22.22times) magnification (Sample II). The corresponding time-resolved spatio-temporal dynamics (i.e., video) of the tracheal branching for both samples are given separately in Supplementary 7. Over the duration of one-hour imaging of tracheal development (in Sample II), it is observed that the Drosophila embryo develops from Stage 10 to 12 where we can see all of the ten tracheal pits (or placodes) are re-positioned inside the embryo. The tracheal pits that are invaginating also begin to elongate giving rise to a dorsal and a ventral stem. The stem towards the ventral side also shows a slight bifurcation giving rise to a small posterior branch and a small anterior branch growing into branches while re-arranging to cover the entire region of the whole embryo. This is in good agreement with reported studies (see Supplementary I). Here, the image at (11.11times) gives the complete view of the whole embryo where the ten tracheal pits are clearly observed to be re-positioned or re-arranged. At the same time, in the image with (22.22times) magnification that corresponds to the magnified view of the region marked by the square box indicated in Fig. 4(I-a), we observe each tracheal pit also developing the bifurcating stems. Similarly, live imaging of the development of the tracheal system from late Stage 13 to Stage 15 was conducted with (11.11times) and (22.22times) magnifications for a duration of 95 min (images at 15 min intervals for 1 h shown in Fig. 4(II(a–e)) and Fig. 4(II(f–j)) respectively). The lower magnification image covers the entire embryo (but at lower resolution) while higher magnification gives the higher resolved images of branch formations (but at the cost of FOV). In the time interval from t0 to (t0+95) min, we observe the formation of some new tracheal branches as well as the continuation of growth and movements of the existing branches. It is observed that the photo-bleaching over the time of imaging is relatively low (See Supplementary 4). We characterize quantitatively the growth in length of the tracheal branches with respect to time.

Figure 5
figure 5

Variation of the lengths of various tracheal branches with time for Drosophila melanogaster (Sample II).

For a better understanding of branch growth, the quantification of an increase in branch length over time is studied for (several) branches over certain regions of interest (for Sample II). The graphical plot of branch length vs. time is given in Fig. 5. In the figure, we can distinctly observe that the branch lengths are increasing and saturated, i.e., tracheal branches are growing with time and attaining maturities at different time intervals. The DLT branch length is almost saturated from the beginning. During embryonic development of Drosophila melanogaster, around late Stage 13 (early Stage 14), the individual portions of the dorsal trachea originating from the 10 tracheal placodes fuse completely to give rise to a single dorsal longitudinal trunk (DLT) of the tracheal system on each side of the body. Although minor cellular movements continue within the DLT, at late Stage 13 (where the imaging of Sample II begins), the DLT is a fully formed continuous tube. Therefore, we do not observe any further visible longitudinal growth in it. Meanwhile, during Stages 13-15, several new tracheal branches emerge from the DLT sequentially. The transverse connectives (TC) and the dorsal tracheal branches (DT) emerge simultaneously from the DLT towards the transverse and dorsal sides of the embryo respectively. Towards the later part of Stage 15, the anterior and posterior lateral longitudinal branches (aLLT and pLLT) extend to fuse with the lateral longitudinal branches of the adjacent placode, along with the emergence of tracheal terminal cells (TTC—not depicted in the images). The LLTs are smaller branches compared to both DTs and TCs. Sample II depicts an embryo between the Stages 13-15. Therefore, we observe the most noticeable and significant growth in the transverse connectives (e.g., TC07 branch), followed by the dorsal tracheal branches (e.g., DT09). We also observe the emergence of the lateral longitudinal branches (e.g., aLLT08 and pLLT07) and capture a portion of their growth dynamics within the time frame of our imaging session. In short, in the figure, we observe that TC07 increases significantly while the DLT shows almost no increase. A comparison of branch lengths of all the branches in Sample II is given separately in Supplementary 8 (see Supplementary Fig. S5-II). For the quantitative analysis of Sample I, the areas of the placodes (at the initial and final time of imaging) are calculated, and the comparison is given in Supplementary 8 (see Supplementary Fig. S5-I).

In a true sense, simultaneous imaging of a single frame/slice in complex three-dimensional biological samples such as the live Drosophila embryo is of great importance but remains as a technological challenge. The inclusion of a higher magnification objective in the setup suggests the potential to scrutinize developmental processes at the cellular level, deep within the tissues, with precise spatio-temporal kinetics. The ability to image the same 3-D sample simultaneously using an objective of lower magnification allows us to continuously monitor the overall status of the live sample. This can allow biologists to track the developmental progress, viability of the embryo, morphological (physical) changes at the surface of the embryo, etc. while tracking in parallel the corresponding cellular changes at the deeper layers inside the embryo. Our set-up, for simultaneous imaging of Drosophila embryos (and other 3-D live biological samples) at two different magnifications, therefore, is necessary for biologists during experimentations and is also advantageous, allowing parallel data collection of the cellular processes as well as the status of the experimental sample. This is to note that, even though we present images obtained at (11.11times) and (22.22times) (in the present study), objective lenses adapted in the two separate or individual detection arms (Arm I and Arm II) are not limited to the two magnifying powers. Instead, the objective lenses can be changed independently and thus, one can obtain images of specimens not only at any magnifications or resolutions but also with any possible combinations that can be achieved by adapting appropriate objective lenses of interest (and the combinations) in the detection arms.

Dynamics of mitochondria in HeLa cells

Figure 6
figure 6

Images of mitochondria of HeLa cells that are obtained at (11.11times) (a) and (44.44times) (b) magnifications but simultaneously. (c) Zoomed-in views for detection of various mitochondrion (marked in coloured boxes) in each frame at four different regions. Comparison of (d) average speed ((bar{v})) of mitochondria in clustered cells, moderately clustered cells and isolated cells and (e) average speed ((bar{v})) of mitochondria at the periphery of the cell and near the nucleus. Scale bar: 20 (mu m).

To validate experimentally the feasibility of our M(lambda)-sMx-SPIM to study spatio-temporal dynamics at the cellular level, we conducted experiments on HeLa cells as the imaging specimen, more specifically, to study the spatio-temporal dynamics of organelles (mitochondria). Figure 6 presents images obtained at the initial time ((t0=0) s). Figure 6a and b present the corresponding images obtained at (11.11times) magnification (detection Arm I) and (44.44times) magnification (detection Arm II). The marked region (indicated in Fig 6a) shows the field of view corresponding to Fig. 6b ((44.44times) magnification). The lower magnification view enables us to locate selectively the desired cells from a cluster or large number of cells that is in agreement with the arguments presented in Supplementary 2. Imaging was performed for 98 seconds, and 222 frames were recorded (in the detection Arm II) while 863 frames were recorded in the detection Arm I. The lower magnification ((11.11times)) fails to provide enough resolution to observe the movements of organelle (mitochondria). But, at higher magnification ((44.44times)), one can clearly observe and detect the mitochondria and their spatio-temporal dynamics. For the quantification of the mitochondrial spatio-temporal dynamics, speeds of different mitochondria—as indicated by marked boxes of different colours (various regions are zoomed-in in Fig. 6c)— were calculated by using a custom-made MATLAB program. The underlying algorithm with flowchart is given in Supplementary 9 (see Supplementary Fig. S6). Video showing the time-resolved dynamics of mitochondria and its tracking is given separately in Supplementary 10. It is observed that the speed of mitochondria near the nucleus (distance from the nucleus (<2 mu m) ) and the periphery of the cell (distance from the nucleus (>2 mu m)) were tabulated separately for different cells. Mean and standard deviation in the measurements of speed were calculated for correlation studies.

As it is clearly visible in Fig. 6b, some cells are observed to be clustered together while some are isolated. The mean speed of mitochondria—both near the nucleus and periphery of the cell—for each of the cells are plotted in Fig. 6d. The cells are marked and labeled separately as clustered, moderately clustered, and isolated. The labeling of the cells and the mitochondria are presented in Supplementary 11. The average speed ((bar{v})) of mitochondria in the clustered cells ((sim 8.03 pm 1.05) (mu m/min.)) is found to be comparatively lower than that of the moderately clustered cells ((sim 8.68 pm 0.64) (mu m/min.)). The speed is measured to be much higher for isolated cells ((sim 11.19 pm 1.30) (mu m/min.)), as it is expected that mitochondria have a higher degree of freedom as far as dynamics is concerned. This is due to the higher interaction of organelles when the parent cells are clustering, i.e., cells are highly interactive when they are in a cluster in comparison to when they are isolated. Figure 6e shows the comparison of the average speed ((bar{v})) of mitochondria near the nucleus and the periphery of the cells which are estimated to be (sim 7.32 pm 0.74) (mu m/min.) and (sim 8.62 pm 1.42) (mu m/min.) respectively. The mean speed of mitochondria is higher at the periphery of the cell than that at the distance close to the nucleus. Again, we estimated the average speed of the mitochondria—comprising of all cases, namely, close to the nucleus, the periphery of the cells (far away from the nucleus), clustered cells, moderately clustered, and isolated— that is found to be 8.08 (mu m/min.) with standard deviation of (16.65%). This clearly shows that not only cells are not correlated but also mitochondria are not correlated corresponding to the different regions of the cells as well as the nature of the cells. Again, spatio-temporal dynamics of organelles (mitochondria) are similar for similar types of cells but highly uncorrelated with respect to that of different types of cells.

Table 1 Comparison of average speed ((bar{v})) of mitochondria in the periphery of the cell and near the nucleus.

Statistical analysis was performed by the unequal variance t-test (Welch’s test)31—which is a statistical test that is used to compare the means of two groups and to determine whether two groups are different from one another—for quantitative characterization of spatio-temporal dynamics (average speed) of organelles (mitochondria) and thus, to check the significance of the difference in average speed of mitochondria near the nucleus and at the peripheral of the cell. The result is presented in Table 1. The obtained P-value ((sim 0.00066)) confirms the statistical significance ((P < 0.05) which was used to denote significance). Close to the cell nucleus, mitochondria density is observed to be higher than that away from the nucleus and thus, they will be more free to move (at the periphery of the cells). This shows that there exists a higher interaction for organelle (mitochondria) close to the nucleus as it is compared to that in the periphery of the cells. In other words, one may say that the lower interaction of cellular organelles (among themselves and with the nucleus) can be the reason for the higher speed of mitochondria at the periphery of the cells.

Figure 7
figure 7

Quantitative evaluation spatio-temporal dynamics (velocity both magnitude and direction) of mitochondria, more specifically, vector representation of time-resolved displacement of three randomly selected mitochondria indicated by marking shown in (a). (b-i) to (b-iii) show the displacements for the entire 222 frames while (c-i) to (c-iii) present the displacements corresponding to randomly chosen 10 consecutive frames. In figures ((b) and (c)), the background images depict the selected mitochondria corresponding to the first frame (at t0), i.e., the initial points are marked at the centers of the mitochondria. Scale bar for (a): 20 (mu m). Scale bar for (b) and (c): 0.4 (mu m).

Figure 7 depicts a typical pictorial representation of the spatio-temporal dynamics of the mitochondria, more specifically, instantaneous displacements or velocities are denoted by vectors. Here, each vector represents the displacement of the mitochondria in sequential (image) frames. In the figure, the length of the arrows and the size of the arrowheads indicate the measures of magnitude of displacement, i.e., the distances through which mitochondria are traversing. The figure depicts the movements of three randomly selected mitochondria (as they are marked in Fig. 7a)—that include one near the nucleus, one from the peripheral of a cell (in a cluster of cells), and one from the peripheral of an isolated cell—for the entire 222 frames (collected at the time interval of 98s) are given in Fig.7(b-i) to Fig.7(b-iii). Similarly, Fig.7(c-i) to Fig.7(c-iii) show the mitochondrial movements of the three mitochondria for ten consecutive frames that are randomly picked up (for the sake of better clarity). The starting points of the initial vectors corresponding to t0 are marked at the center of the corresponding mitochondria which are in turn represented by the background images in the figures. From the experimental results, the displacements of the mitochondria close to the nucleus (Fig. 7(b-i)) are relatively low in comparison to that of mitochondria at the peripheral of the cell (both for clustered and isolated cells (Fig. 7(b-ii) and (b-iii)). Shortly, we may conclude that mitochondria close to the nucleus move shorter distances or are more highly interacted and thus, tightly bound to the nucleus. We may note that we acquire the sequential images at the same frame rate ((sim 2.27) frames/s), i.e., acquisition time period ((sim 0.44) s). With the given same acquisition time of image frames, the depicted displacement vectors indicate the corresponding velocity—in both magnitudes (represented by lengths) and directions (represented by arrowheads), and consequently, we may conclude that mitochondria close to the nucleus move at lower speed in comparison to that of mitochondria at peripheral of cells. This may be because of the highly interactive nature of mitochondria close to the nucleus. Again, the mitochondria in isolated cells undergo movement or spatio-temporal dynamics not only at a higher speed but also with higher randomization. This may be because of loose in interaction of the parent cells with other (neighbouring) cells. Lastly, from the figures, it is clear that all of the mitochondria (either isolated or clustered) undergo spatio-temporal dynamics about their thermodynamic equilibrium position.

As illustrated in Fig. 6, lower magnification imaging allows us to observe and study specimens at a wider field of view, encompassing a substantial number of cells. This capability enables the selective identification of specific regions of interest (such as certain clusters of cells/tissue for a selective or focused study). Consequently, correlations among different cells and/or organelles, that include physiological activities (say, spatio-temporal dynamics (velocities), as demonstrated in Fig. 6) and parameters (like the number of clusters and cells per cluster) can be established. This stands in contrast to the conventional techniques where imaging is limited to certain pre-specified cells or organelles, i.e., with a restricted field of view at a given instant. Our proposed dual-arm imaging modality uniquely allows the observation of selective cells or organelles at higher magnification or resolution, akin to conventional techniques, without sacrificing generality. Prior to our studies, the simultaneous acquisition of multiple image frames—at different magnifications, spatial resolutions, and fields of view— presented a technological challenge, even with conventional fluorescence microscopy.

Our study further demonstrates the quantitative characterization of spatio-temporal dynamics (velocity, both amplitude and direction) of organelles (mitochondria) and their mutual correlations. Notably, mitochondria close to the nucleus (or in clustered cells) exhibit a lower degree of freedom compared to those at the cellular periphery (or isolated cells), as highlighted in the Abstract. In addition to velocity/speed, it will be interesting to look at other parameters like mitochondrial membrane potential which influence the energy production capacity of mitochondria various across different clusters of cells or within cell populations32. Consequently, the quantitative analysis of mitochondrial membrane potential in different cell clusters holds valuable information, especially in tissue samples where variations in mitochondrial movement and energy production capacity exist across cell populations. Tissues exhibit larger heterogeneity in their cell population both in terms of density of cells as well as type of cells. Hence, it is imperative to study how the nearby cells influence the activity of internal organelles like mitochondria. Lower magnification aids in classification like relative positioning and density, while higher magnification provides the necessary resolution for quantitative measurements. Furthermore, our microscope facilitates studying interactions between mitochondria and the microbiome. Mitochondria, with evolutionary origins linked to bacteria, exhibit similarities with certain bacterial structures. Investigating these interactions in a larger context may provide insights into the co-evolution and symbiotic relationships between mitochondria and bacteria. The movement and energy parameters of mitochondria can be measured while cells interact with surrounding microbes. It will be interesting to have quantifiable data on how the position of microbes influences various organelles in the cells33,34,35. Another intriguing phenomenon is organelle transfer between cells like mitochondrial transfer, and it would be fascinating to quantify various parameters during this process. Questions about which mitochondria are chosen for transfer from donor to recipient cells, their movement relative to the rest of the mitochondrial population, and their energy levels could be answered through a holistic view provided by lower FOV and a magnified view for accurate measurement. While these experiments were beyond the scope of this manuscript due to various constraints, the potential for future investigations is evident36,37.

Multi-spectral imaging

Mouse embryo

Figure 8
figure 8

(a) and (b) shows images (obtained at (11.11times)) of AF594 tagged CD31 positive endothelial cells and AF488 tagged HCs respectively. (c) shows the merged or fused image for (a) and (b). (d), (e), and (f) show the corresponding images obtained with (22.22times) magnification. DA marks the dorsal aorta. Arrowheads mark HCs, somites, and blood vessels/vascular networks. Scale bar: 50 (mu m).

In the E10.5 mouse embryo, Hematopoietic cells (HCs) are seen associated with the endothelial lining of major blood vessels such as the dorsal aorta (DA). Here, we mainly focus on such regions to locate the HCs. The laser was tuned at an optical wavelength ((lambda _{1}=594) nm) to image endothelial cells. Images are recorded simultaneously at (11.11times) and (22.22times) magnifications and they are depicted in Fig. 8a and d respectively. Further, the images of HCs are acquired by tuning the wavelength at (lambda _{2}=488) nm. Figure 8b and e show the images obtained at (11.11times) and (22.22times) magnifications respectively. The individual images of endothelial cells and HCs are merged to get the multi-spectral image at both magnifications, which are given in Fig. 8c and f. From the merged image, we can locate HCs within the DA, in which the DA, somites, and vascular networks are indicated by arrowheads in Fig. 8a.

HeLa cells

Figure 9
figure 9

(a), (d), and (g) show images of microtubules, and (b), (e), and (h) show the nucleus of HeLa cells for the magnification (5.56times), (11.11times), and (22.22times) respectively. (c), (f), and (i) are the merged images of the nucleus and microtubule. Scale bar: 40 (mu m).

Figure 9 shows the multispectral imaging of HeLa cells at three different magnifications, namely, (5.56times), (11.11times), and (22.22times). The laser source was first tuned to a wavelength (lambda _{1}=488) nm to image the microtubules and then tuned to (lambda _{2}=590) nm to image the nuclei of the cell line. The individually obtained images were merged (as mentioned above) to obtain the complete image of HeLa cells. The localization of nuclei and microtubules is well observable from the multispectral image. This is to note that one can select optical excitation at any wavelength in the range from (420-670) nm (even though the experiments were conducted only for two wavelengths in this present study). Again, our proposed M(lambda)-sMx-SPIM imaging technology facilitates giving any optical excitation wavelength by the change in the tuneable pulsed laser source.