Mitochondria are essential organelles responsible for energy production and regulation of critical processes including immune activation and signal transduction^[1]. Their dynamic morphology, governed by fission and fusion events, is tightly coupled to cellular homeostasis and adaptation to metabolic demands^[2–5]. In cardiac tissue—the most energy-demanding organ—disruption of mitochondrial dynamics leads to structural abnormalities, functional impairment, and accelerated aging^[6–10]. Age-related stressors, such as oxidative stress^[11], DNA damage^[12], and inflammatory responses interact^[13], contribute to mitochondrial dysfunction, leading to damage in myocardial cells and, ultimately, causing irreversible changes in the structure and function of cardiac tissue^[14]. Therefore, preserving mitochondrial integrity is thus pivotal for maintaining cardiac health and mitigating aging^[15].

Direct observation of mitochondrial fusion and fission dynamics at molecular resolution remains a pivotal challenge in aging research. Conventional approaches each face intrinsic limitations. For example, electron microscopy achieves nanoscale ultrastructural detail but is restricted to static, fixed-cell imaging, rendering dynamic processes invisible^[16]; live-cell fluorescence microscopy enables real-time tracking but is hindered by the diffraction limit, obscuring critical sub organellar interactions^[17]; and biochemical assays indirectly quantify fission and fusion proteins but lack spatial and temporal resolution to map these events in living systems^[18]. This limitation originates from the optical diffraction barrier, which confines conventional microscopy to resolutions near 200 nm^[19], preventing conventional microscopy from resolving sub-mitochondrial structures or capturing rapid nanoscale dynamics of fission and fusion—key processes driving age-related mitochondrial dysfunction. Super-resolution techniques partially address this issue^[20]. For example, stimulated emission depletion microscopy (STED) achieves 50 nm resolution but suffers from high phototoxicity^[21]; photo-activated localization microscopy (PALM) and stochastic optical reconstruction microscopy (STORM) attain sub-20 nm resolution yet require fixed samples^[22,23]; and structured illumination microscopy (SIM) provides 100 nm resolution at low excitation power suitable for live-cell imaging^[24,25]. Additionally, integrating SIM with nonlinear effects can significantly enhance the imaging resolution. Therefore, we employ two-photon nonlinear structured illumination microscopy (TP-SIM) to capture higher frequency information through the nonlinear effects of two photons in this study. This approach not only enhances the resolution but also reduces phototoxicity compared to saturated structured illumination microscopy^[26], making it effective for dynamic observations of live cells^[27].

In this study, we demonstrate the application of super-resolution techniques in mitochondrial imaging and dynamic observation. First, we validated the resolution of TP-SIM. Using COS7 cells, we performed super-resolution imaging of mitochondria with immunofluorescent labeling, confirming that the system’s resolution is sufficient to reveal internal mitochondrial details and enable dynamic monitoring of organelle behaviors. Furthermore, we induced aging in H9C2 cells and utilized super-resolution methods to observe changes in mitochondrial morphology before and after aging induction as well as to analyze the effects of aging on mitochondrial structure. D-galactose-induced cellular aging is a widely used model for studying the aging process^[28]. It is reduced to galactitol by aldose reductase or oxidized to hydrogen peroxide, leading to oxidative stress, cellular damage, and impairment of mitochondrial function and dynamics in cardiomyocytes^[29,30]. Based on this approach, we observed the mitochondrial fission and fusion processes in aging H9C2 cells. Our findings establish the capability of TP-SIM in resolving mitochondrial fusion and fission dynamics.

Figure 1 illustrates the configuration of the system utilized in our study. The system incorporates the Mai Tai HP tunable ultrafast laser from Spectra-Physics, enabling laser output at 810 nm. After attenuation of laser intensity via a polarization beam splitter, the beam is directed into an electro-optic modulator (EOM305, Conoptics) for sinusoidal waveform modulation. The modulated beam is then scanned spatially using a galvo scanner (GS, Cambridge Tech, Model 6210 H). Subsequently, the beam is introduced into a Nikon inverted microscope through a tube lens and is focused onto the sample using a 60× oil objective (CFI Plan Apochromat Lambda). Fluorescence emitted from the sample upon laser excitation is collected through the objective and reflected by a dichroic mirror (DM). Finally, the emitted light passes through a filter before being detected by a photo multiplier tube (PMT, Hamamatsu, H7422-50). In the experiment, we employed LabVIEW to control the waveform of the EOM and utilized a data acquisition card to receive the image signal from the PMT.

In this experiment, the GS was set at 1.75 V, yielding a 28-µm field size at 800×800 scanning points. We employed second-harmonic fringes generated based on the frequency-doubling effect in two-photon microscopy to achieve a threefold increase in resolution. For structured illumination image reconstruction, we implemented a three-direction acquisition scheme with five phase shifts per direction, requiring a total of 15 images to accomplish super-resolution imaging.

To demonstrate the capability of TP-SIM for super-resolution imaging of mitochondria and visualizing their internal details, we utilized this system to observe mitochondria in the COS7 cells. The mitochondrial imaging results of the COS7 cells are depicted in Fig. 2. Using MitoTracker Orange CMTMRos dye, the inner membranes of the mitochondria were successfully visualized. Figure 2(A) shows a two-photon mitochondrial image obtained through two-photon excitation, while Fig. 2(B) presents the reconstructed image using a two-photon nonlinear structured illumination microscope.

Compared to the two-photon image, the two-photon super-resolution system provides a significantly higher spatial resolution, along with enhanced image contrast and edge sharpness. The reconstructed super-resolution image of the mitochondria achieved a resolution of 82 nm full width at half-maximum (FWHM), which is three times higher than traditional methods. This demonstrates the capability of our system to provide more reliable structural information about mitochondria. Furthermore, during the resolution measurement experiments, we assessed the excitation power of the system’s objective lens, which can be minimized to 2.4 mW. Our system enables clearer observation of the microstructure and dynamic changes inside mitochondria, establishing its potential as a powerful tool for imaging biological samples with high precision and minimal phototoxicity.

An in-depth study of mitochondria requires not only high-resolution imaging techniques to resolve their intricate structures but also real-time dynamic observation capabilities to capture their dynamic changes. Therefore, we utilized TP-SIM to dynamically observe mitochondria in COS7 cells.

To enhance the temporal resolution of our point-scanning system while maintaining consistent pixel dimensions, we optimized the acquisition parameters by reducing the number of scanning points. This optimization enabled an imaging speed of 0.13 s per frame, allowing continuous observation of mitochondrial dynamics over 4 min and 30 s.

The results of our dynamic observation are presented in Fig. 3, where we performed continuous imaging of mitochondria within COS7 cells. Figure 3(A) depicts a two-photon mitochondrial image acquired via two-photon excitation, showcasing the initial state of the mitochondria. Figure 3(B), on the other hand, exhibits the mitochondrial image after super-resolution reconstruction, highlighting the enhanced resolution achieved with this technique.

Figure 3(C) specifically focuses on the dynamic imaging of the mitochondrial fusion process, confined within the white dashed boxes marked in Figs. 3(A) and 3(B). In this figure, the upper panel illustrates the two-photon image of the mitochondrial fusion event, while the lower panel presents the same event after super-resolution processing. A comparative analysis between the two images reveals that the super-resolution technique significantly enhances the visualization of the mitochondrial membrane details and the fusion process, which remain indistinct in the two-photon image. Notably, the entire duration of the mitochondrial fusion process was recorded for over 58 s, demonstrating the system’s proficiency in capturing dynamic cellular events in real time.

To explore the impact of cellular aging on mitochondrial morphology, we cultured H9C2 cells in vitro and induced them into a senescent state using D-galactose. We performed specific staining of mitochondria using MitoTracker Deep Red, followed by imaging with TP-SIM for observation, and conducted a detailed analysis of mitochondrial images.

The mitochondrial imaging results of the H9C2 cells, as depicted in Fig. 4(A), compare the control and senescent groups, showcasing mitochondrial morphology at both standard and super-resolution levels using two-photon microscopy. In the control group, H9C2 cells exhibit a complex, elongated mitochondria network, indicating healthy mitochondrial dynamics. In contrast, the senescent group shows a noticeable transition to fragmented and dispersed mitochondria, with a significant reduction in filamentous structures.

Comparative image analysis clearly highlights the profound impact of senescence on mitochondrial morphology, characterized by an increase in mitochondrial count, a decrease in mitochondrial size, and a transition from elongated to punctate morphology. The combination of super-resolution and two-photon images provides a more detailed perspective on mitochondrial structure, greatly facilitating a comprehensive analysis of age-related morphological changes.

To quantitatively analyze the quality of mitochondria in the images, we employed the Mitochondrial Analyzer plugin in ImageJ for further analysis. Figure 4(B) shows the mitochondrial data of the H9C2 cells before and after super resolution, including the count, area, aspect ratio, and branch data of the mitochondria. The analysis demonstrated that following super-resolution imaging, the mitochondrial count increased, the area decreased, and the mitochondrial morphology was enhanced. These findings suggest that utilizing higher-resolution imaging for mitochondrial analysis allows for more precise and detailed characterization. The results shown in Fig. 4(C) focus on key metrics, such as mitochondrial count, area, average aspect ratio, and branch junctions, as well as comparisons between the senescent and control groups. The analysis revealed a significant increase in the mitochondrial count following cellular aging, accompanied by a noticeable decrease in the average size of mitochondria after aging. Additionally, the average aspect ratio of the mitochondria and the branch junctions of mitochondria post-aging was lower than that of pre-aging, indicating a clear trend of mitochondrial length reduction and fragmentation during the aging process and causing a reduction in the connectivity of the mitochondrial network. These aging-related changes in mitochondrial morphology were consistent across samples. In addition, the application of super-resolution technology underscores its value in capturing subtle changes in mitochondria with greater precision.

Based on the above, dynamic observations of mitochondria in the senescent H9C2 cells were conducted using our two-photon non-linear structured illumination microscope system, as depicted in Fig. 5. Figure 5(A) illustrates two-photon images of the H9C2 cell mitochondria acquired through two-photon excitation, while Fig. 5(B) shows the super-resolution reconstructed mitochondria images. A detailed analysis revealed that the senescent mitochondria exhibit a pronounced preference for fusion events, accompanied by a notable reduction in fission activity. This observation shows a contrast with the relatively balanced mitochondrial dynamics observed over extended periods in non-senescent cells, highlighting the functional and structural changes associated with cellular aging. The green dashed box [Fig. 5(C)] and white dashed box [Fig. 5(D)] from Fig. 4 highlight the dynamic images of the mitochondria. Through super-resolution imaging, we visualized complex steps involved in mitochondrial fusion with unprecedented detail. The gradual approach, contact, and eventual fusion of mitochondrial membranes into a unified structure, which are processes that remain indistinct under two-photon microscopy limitations, were clearly visualized using our system. This enhanced resolution is further exemplified in Fig. 5(D), revealing intricate details of mitochondrial dynamics, including the formation of contact points and subsequent fusion events.

These experiments successfully reveal the morphological and dynamic changes of mitochondria before and after aging, verifying that aging leads to more frequent mitochondrial fusion activities. The experimental results indicate a close relationship between aging and mitochondrial morphology and dynamics, characterized by enhanced fusion activity and weakened division activity of mitochondria in aging cells. Super-resolution imaging uniquely captures more detailed structural changes during mitochondrial fusion and apoptosis, which are beyond the capabilities of two-photon microscopy. In addition, our system has successfully achieved continuous dynamic photography of mitochondria in aging cells for up to 5 min, without inducing photobleaching interfering with mitochondrial viability during this process, highlighting its exceptional capabilities for prolonged observation of live-cell dynamics.

The H9C2 cells and COS7 cells were cultured in the DMEM medium supplemented with 10% fetal bovine serum (FBS) and 1% penicillin–streptomycin. The cells were maintained in a humidified CO2 incubator at 37°C to provide suitable growth conditions. Passaging was performed when the cells reached 60%–70% confluency at a 1:2 split ratio. Experiments were conducted when the cells were at their optimal state.

During passaging, a portion of the cells was seeded in a confocal cell culture dish. After the cells adhered to the dish, 10 g/L D-galactose was added to the DMEM/FBS medium to induce cell aging. The cells were incubated with D-galactose for 72 h before proceeding with subsequent observations.

For live-cell imaging, H9C2 cells and COS7 cells were seeded in a confocal cell culture dish. After reaching the desired cell density, the cell culture medium was carefully removed, followed by gently washing the dish twice with PBS to eliminate dead cells. MitoTracker™ Deep Red was added to the DMEM/FBS medium to achieve a final concentration of 200 nmol/L for H9C2 cell staining. MitoTracker™ Orange CMTMRos was added to the DMEM/FBS medium to achieve a final concentration of 200 nmol/L for COS7 cell staining. The cells were then incubated in a cell culture incubator for 30 min. Following incubation, the cells were washed three times with PBS to remove any excess dye. Subsequently, a serum-free culture medium was added prior to visualizing the cells using a microscopic system.

In this study, we developed a TP-SIM system to address the limitations of two-photon conventional microscopy in capturing mitochondrial dynamics during cellular aging. Traditional approaches—constrained by diffraction-limited resolution and phototoxicity—fail to resolve transient mitochondrial fission and fusion events or track organelle dynamics in live cells over extended periods. Our TP-SIM system overcomes these barriers by combining structured illumination with two-photon nonlinear excitation, achieving 82-nm resolution at ultralow laser power, enabling continuous 5-min live-cell imaging with minimal photodamage.

Using D-galactose-induced senescent cardiomyocytes, we demonstrate that TP-SIM captures aging-associated mitochondrial fragmentation, characterized by increased mitochondrial number and reduced average size compared to normal cells. Dynamic imaging further revealed imbalanced mitochondrial fusion activity in aged cells, directly linking structural changes to senescence. The system’s low phototoxicity preserved cellular viability during imaging, allowing precise correlation between mitochondrial morphological remodeling and functional decline. Compared to conventional microscopy, TP-SIM significantly improves image clarity, resolving sub-mitochondrial details critical for aging research.

TP-SIM provides dual super-resolution and deep-tissue imaging capabilities. The combination of two-photon optical sectioning and structured illumination enables volumetric imaging of intact tissues, offering more physiologically relevant data than monolayer cultures. With its high temporal resolution, the system opens up new possibilities for studying mitochondrial dynamics across different tissue layers and understanding their spatiotemporal relationships during aging processes. Future optimization for in toto tissue imaging could transform studies of age-related tissue degeneration.