Skeletal muscle differentiation is a highly dynamic process, which particularly relies on nuclear positioning. Here, we describe a method to track nuclei movements by live cell imaging during myoblast differentiation and myotube formation and to perform a quantitative characterization of nuclei dynamics by extracting information from automatic tracking.
Nuclear positioning within cells is important for multiple cellular processes in development and regeneration. The most intriguing example of nuclear positioning occurs during skeletal muscle differentiation. Muscle fibers (myofibers) are multinucleated cells formed by the fusion of muscle precursor cells (myoblasts) derived from muscle stem cells (satellite cells) that undergo proliferation and differentiation. Correct nuclear positioning within myofibers is required for the proper muscle regeneration and function. The common procedure to assess myoblast differentiation and myofiber formation relies on fixed cells analyzed by immunofluorescence, which impedes the study of nuclear movement and cell behavior over time. Here, we describe a method for the analysis of myoblast differentiation and myofiber formation by live cell imaging. We provide a software for automated nuclear tracking to obtain a high-throughput quantitative characterization of nuclear dynamics and myoblast behavior (i.e., the trajectory) during differentiation and fusion.
Skeletal muscle is the largest tissue in the human body, totaling 35%-40% of body mass1. Satellite cells are muscle stem cells, anatomically characterized by their position (juxtaposed to the plasma membrane, underneath the basal lamina of muscle fibers), that give rise to proliferating myoblasts (myogenic progenitor cells), which eventually differentiate and integrate into existing myofibers and/or fuse to form new myofibers2,3,4. Their discovery and the progress in the study of their biology has led to significant insights into muscle development and regeneration.
Protocols to isolate and differentiate myoblasts into myotubes have been developed many years ago and are still widely used to study skeletal muscle differentiation5,6,7. However, most of these methods represent static procedures that rely on the analysis of fixed cells and, consequently, do not allow scientists to fully explore highly dynamic processes, such as myoblast fusion and myofiber maturation. The most striking example is nuclear positioning, which is tightly regulated, with nuclei initially in the center of the myofiber and, then, located at the periphery after myofiber maturation8,9. Live imaging is the most appropriate technique to obtain further insights into such a peculiar phenomenon.
Here, we describe a method that enables scientists to record myoblast differentiation and myotube formation by time-lapse microscopy and to perform quantitative analyses from the automatic tracking of myoblast nuclei. This method provides a high-throughput quantitative characterization of nuclear dynamics and myoblast behavior during differentiation and fusion. The protocol is divided into four different parts, namely (1) the collection of muscles from the hindlimbs of mice, (2) the isolation of primary myoblasts that consists in mechanical and enzymatic digestion, (3) myoblast proliferation and differentiation, and (4) live imaging to track nuclei within the first 16 h of myoblast differentiation.
In the following procedure, myoblasts are isolated from H2B-GFP mice and treated with 1 µg/mL of doxycycline to induce H2B-GFP expression, as previously described10. Alternatively, it is possible to isolate the myoblasts from other transgenic mice that express a fluorescent protein in the nucleus or to transfect the cells isolated from wild-type mice to express a fluorescent protein in the nucleus, as described by Pimentel et al.9.
All procedures involving animal subjects were approved by the San Raffaele Institutional Animal Care and Use Committee.
1. Dissection of mouse hindlimb muscles
2. Isolation of primary myoblasts
NOTE: All the procedures for cell culture are done in sterile conditions.
3. Myoblast proliferation and differentiation assays
NOTE: When myoblast density is high, some cells initiate elongation, and then, it is necessary to split the cells. Generally, it is possible to split cells 2x–3x without affecting their phenotype.
4. Live-imaging of myoblast differentiation and nuclear tracking
To automatically follow nuclear movement during myoblast differentiation in live imaging, the nuclei should preferentially be fluorescently labeled. It is important to note that using DNA-intercalating molecules is not feasible because these molecules interfere with the proliferation and differentiation of primary myoblasts13. As an example, proliferation and differentiation have been analyzed in primary myoblasts cultured with or without Hoechst (Figure 1). It is evident that proliferation (Figure 1A, B) and differentiation (Figure 1C, middle panel) are strongly impaired in myoblasts cultured with Hoechst 33342. Conversely, myoblasts isolated from H2B-GFP mice and cultured with doxycycline have nuclei with green fluorescence and differentiate similarly to myoblasts isolated from wild-type mice (Figure 1C, right and left panels, respectively).
Live cell imaging with primary myoblasts expressing the H2B-GFP protein allows the tracking of nuclei during differentiation (Figure 2 and Supplementary Video 1). Merged images of transmission and GFP channels at the initial (Figure 2A) and final time points (Figure 2B) during the differentiation of H2B-GFP myoblasts allow scientists to identify myotubes and, consequently, nuclei that end up integrating into a myotube (e.g., the nucleus in the blue circle) or nuclei that do not fuse into a myotube (e.g., the nucleus in the red circle).
To extract information on nuclei/cell movements from the live cell imaging data, it is possible to use the provided software to track the trajectories of the nuclei. The software uses the image from the H2B-GFP channel (the nuclei; Figure 3A) to create a mask and to segment the nuclei in each frame. For nuclear segmentation in a frame, a "conservative" threshold is selected using Otsu's method on the image after Gaussian filtering14. Next, objects are roughly segmented; to get a finer segmentation, each object is masked again using a threshold based on the average in its bounding box. A watershed transform is then used to separate nearby objects (Figure 3B). In our hands, one watershed transform has been unable to separate most nearby nuclei (Figure 3B, *inset). Thus, we used the area to select the larger objects in the image and apply a new watershed transformation (Figure 3B, **inset), which is able to separate additional nearby nuclei. With this approach, most of the nuclei are identified in the image (Figure 3C).
To track the nuclei, we improved the routines by incorporating the tracking routines published by Tinevez15. In short, the tracking algorithm links nuclei detected in consecutive frames with the aim of minimizing the sum of the distances between the position of each nucleus in a frame and the position of the same nucleus in the next one, using the "Hungarian algorithm" for such minimization16. The step is repeated in order to link unlinked nuclei that are up to a certain maximum number of frames away. A threshold for the maximum distance between the position of an object in a frame and the next one is also established. For the data presented here, a maximum number of five frames and a maximum distance of 20 pixels per step give a high number of correct tracks (Figure 3D). Importantly, we can use these routines to check the quality of the tracking. It is also possible to use this script to find cells that have a given behavior, for example forming (or not) a myofiber, and subsequently analyze the dynamical features of the set of cells of interest.
The applied software generates the tracks with the x- and y-coordinates for each cell in frame n, , for hundreds of cells (Figure 3D). We can use such information to extract information about nuclei motion. As an example, the total displacement between the initial frame (n = 0) and the final frame N is defined as follows.
Displacement of a trajectory= || ||
Here, "||·||" stands for the norm of the vector (Figure 4A).
The velocity of each nucleus in frame n is as follows.
Then, we can define the length of a trajectory for a given cell as follows.
Length of trajectory=
As an example of how these simple parameters can already give relevant information, we computed the length of nuclei trajectories of myoblasts incubated with or without Hoechst. It appears that the total length of the trajectories is slightly higher for cells stained with Hoechst, although the difference is not significant (Figure 4B). However, when we computed the total displacement during a time lapse, we observed that the total displacement of cells stained with Hoechst was significantly smaller than that of unstained cells (Figure 4C). This indicates that the movement of stained cells has a lower directionality. To further strengthen this hypothesis, we quantified the directionality of the motion of a cell by computing the angle of the velocity in each frame (Figure 4A).
We also quantified the variation of the angle between frames.
The total angle variation along a trajectory can then be defined as follows.
Total angle variation=
As predicted, the total angle variation for the unstained cells is significantly smaller compared to cells stained with Hoechst (Figure 4D). Hence, this simple measure of directionality, obtained from a straightforward calculation, indicates that the myoblasts stained with Hoechst are less prone to maintain the direction of their displacement, which reflects their impaired ability in forming myotubes (Figure 1).
In short, live cell imaging data generated as described here provide a quantitative link between nuclear dynamics and the ability to form myotubes and can be used as input for a high-throughput quantitative characterization of nuclear dynamics and myoblast behavior during differentiation and fusion.
Figure 1: Incubation with Hoechst interferes with the proliferation and differentiation of myoblasts. (A) Representative images of primary myoblasts stained with Hoechst after 24 h in proliferation medium (left panel) or cultured for 24 h in proliferation medium with Hoechst (right panel), and (B) the relative quantification. Data are the mean ± SEM, and the statistical significance was assessed with Student's t-test. *P < 0.05. (C) Representative images of primary wild-type myoblasts, cultured for 24 h in differentiating medium without (left panel) or with Hoechst (middle panel), and of H2B-GFP myoblasts cultured for 24 h in differentiating medium (right panel) and stained with Hoechst (blue) and an anti-myosin heavy chain antibody (MyHC; red). The scale bars = 500 µm. Please click here to view a larger version of this figure.
Figure 2: Live imaging of H2B-GFP myoblasts during differentiation. Merged images of transmission and GFP channels at (A) the initial (t = 0 h) and (B) final time points (t = 16 h) during the differentiation of H2B-GFP myoblasts (20x objective). Some examples of myotubes are highlighted with black arrows in panel B. The scale bars = 50 µm. (C) Magnified dotted areas from panels A and B in which it is possible to identify a single nucleus that ends up integrating into a myotube (in blue) or not (in red) at t = 0 h, t = 8 h, and t = 16 h. The scale bars = 50 µm. Please click here to view a larger version of this figure.
Figure 3: Segmentation of H2B-GFP myoblasts tracking during differentiation. (A) Example of a frame of the time lapses showing the nuclei of approximately 400 cells. (B) Example of nuclei masking and segmentation. Bright objects are segmented using both a global and a local threshold, and a watershed transformation is applied to separate nearby nuclei. Nearby nuclei might be difficult to segment (*inset), so a second watershed-based segmentation is performed in objects of large areas to segment a higher number of nuclei (**inset). (C) Detected nuclear positions (red crosses), that are later used in the tracking algorithm. (D) Tracks generated using the tracking algorithm that minimizes the sum of the distances between each object's position in two consecutive frames. A maximum possible value of such distance for each object is provided as an input that allows scientists to link nuclei separated by more than one frame. The scale bars = 50 µm. Please click here to view a larger version of this figure.
Figure 4: Quantitative analyses of nuclei dynamics uncover the impaired ability of myoblasts incubated with Hoechst to maintain cell directionality. (A) Description of parameter measurements.It is possible to define the velocity of a nucleus by considering the position in two consecutive frames. The trajectory angle and the angle variation can then be readily calculated from the positions as indicated in the scheme. (B) Distribution of the trajectories length, (C) total displacement, and (D) variation of the trajectories angle of myoblasts incubated without Hoechst (green line) or with Hoechst (blue line). The two distributions are not significantly different, while the total displacement is significantly higher, and the variation of the angle is significantly lower for unstained cells (one-sided Kolmogorov-Smirnov test, *p < 0.05 and **p < 0.005). Please click here to view a larger version of this figure.
Supplementary Video 1: Live imaging of H2B-GFP myoblasts during differentiation. Imaging of H2B-GFP myoblasts cultured in differentiating medium from the initial (t = 0 h) to final time points (t = 16 h) using transmission and GFP channels (20x objective). Please click here to download this file.
Muscle fibers (myofibers) are multinucleated cells that are formed by the fusion of muscle precursors cells (myoblasts) derived from muscle stem cells (satellite cells) that undergo proliferation and differentiation2,3,4. To assess myoblast differentiation, the common procedure consists of culturing myoblasts in differentiating medium and fixing the cells at different time points to perform immunofluorescence staining for MyHC, a marker of differentiation, and staining of nuclei with DNA-intercalating molecules (e.g., Hoechst)5. This method is useful to evaluate myoblast differentiation by measuring different parameters, such as the differentiation index (number of MyHC-positive cells/total cell number) or the fusion index (number of myotubes with at least two nuclei/total cell number). However, myoblast fusion and myotube formation are highly dynamic processes that rely on the motility of the nuclei8. Indeed, nuclear positioning within myofibers is required for proper muscle regeneration and function17. The mobility of myoblasts and of their nuclei might turn out to be a critical parameter to evaluate myoblast differentiation and to uncover novel defects in muscular disorders. Hence, it is essential to develop novel procedures to study cell behavior and nuclear movement during myoblast differentiation and myofiber formation.
Here, we describe a method that enables scientists to follow myoblast differentiation and myotube formation by live cell imaging and to perform quantitative analyses from the automatic tracking of nuclei. The advantage of this method is the possibility to track during the differentiation of fluorescent nuclei of myoblasts, isolated from H2B-GFP mice, without any cell transfection or additional staining. The H2B-GFP mice are commercially available and, consequently, easily accessible. Alternatively, it is possible to isolate the myoblasts from other transgenic mice that express a fluorescent protein in the nucleus or to transfect the cells isolated from wild-type mice to express a fluorescent protein in the nucleus, as previously described9. These different options further extend the potential applications of the method presented here.
The current protocol relies on the culture of primary myoblasts, and consequently, a high variability might be observed in the following experimental procedures, also due to the potential presence of nonmyogenic cells after the isolation from muscles18. To overcome this limitation, it is possible to isolate myoblasts by cell sorting as previously described19. Another critical point in the method described here is the difficulty to generate a perfect tracking of nuclei when cells are close to each other, such as during myotube formation. However, the software routines discussed here allow scientists to visually inspect the quality of the tracking and, if necessary, to select a subset of cells of interest for a subsequent analysis. Further improvement of the software might be required to provide a complete characterization of nuclear dynamics during myofiber formation.
In conclusion, this method is useful to provide quantitative insight into nuclei dynamics during myoblast differentiation by comparing different conditions, as we reported with Hoechst incubation. This example illustrates the ability to extract meaningful dynamic data from time-lapse experiments.
The authors have nothing to disclose.
This work was supported by the AFM-Telethon to E.V. (#21545) and by the Ospedale San Raffaele (OSR) Seed Grant to S.Z. (ZAMBRA5X1000). Dr. Jean-Yves Tinevez from the Image Analysis Hub of the Institut Pasteur is acknowledged for publicly sharing his "Simple Tracker" MATLAB routines.
chicken embryos extract | Seralab | CE-650-J | |
collagenase | Sigma | C9263-1G | 125units/mg |
collagen from calf skin | Sigma | C8919 | |
dispase | Gibco | 17105-041 | 1.78 units/mg |
doxyciclin | Sigma | D1515 | |
DMEM | Sigma | D5671 | |
fetal bovine serum | Life technologies | 10270106 | |
gentamicin | Sigma | G1397 | |
Horse serum | Invitrogen | 16050-098 | |
Hoechst | life technologies | 33342 | |
IMDM | Sigma | I3390 | |
L-glutamine | Sigma | G7513 | |
Matrigel | Corning | 356231 | |
penicillin-streptomycin | Sigma | P0781 | |
red blood cells lysis medium | Biolegend | 420301 | |
Digestion medium | |||
collagenase | 40 mg | ||
dispase | 70 mg | ||
PBS | 20 ml | ||
filtered 0.22um | |||
Blocking medium | |||
DMEM | |||
Fetal bovine serum | 10% | ||
L-glutammine | 1% | ||
penicillin-streptomycin | 1% | ||
gentamicin | 1 ‰ | ||
filtered 0.22um | |||
proliferation medium | |||
IMDM | |||
Fetal bovine serum | 20% | ||
L-glutammine | 1% | ||
penicillin-streptomycin | 1% | ||
gentamicin | 1 ‰ | ||
chichen embryo extract | 3% | ||
filtered 0.22um | |||
differentiation medium | |||
IMDM | |||
Horse serum | 2% | ||
L-glutammine | 1% | ||
penicillin-streptomycin | 1% | ||
gentamicin | 1 ‰ | ||
chichen embryo extract | 1% | ||
filtered 0.22um |