The formation of a proper sarcomere network is important for the maturation of iPSC-derived cardiomyocytes. We present a super resolution-based approach that allows for the quantitative evaluation of the structural maturation of stem cell derived cardiomyocytes, to improve culture conditions promoting cardiac development.
The maturation of iPSC-derived cardiomyocytes is a critical issue for their application in regenerative therapy, drug testing and disease modeling. Despite the development of multiple differentiation protocols, the generation of iPSC cardiomyocytes resembling an adult-like phenotype remains challenging. One major aspect of cardiomyocytes maturation involves the formation of a well-organized sarcomere network to ensure high contraction capacity. Here, we present a super resolution-based approach for semi-quantitative analysis of the α-actinin network in cardiomyocytes. Using photoactivated localization microscopy a comparison of sarcomere length and z-disc thickness of iPSC-derived cardiomyocytes and cardiac cells isolated from neonatal tissue was performed. At the same time, we demonstrate the importance of proper imaging conditions to obtain reliable data. Our results show that this method is suitable to quantitatively monitor the structural maturity of cardiac cells with high spatial resolution, enabling the detection of even subtle changes of sarcomere organization.
Cardiovascular diseases (CVD) such as myocardial infarction or cardiomyopathy remain the major cause of death in the western world1. As the human heart possesses only poor regenerative capacity, there is a need for strategies to promote the recovery from CVDs. This includes cell replacement therapies to replenish lost cardiomyocytes (CM), as well as the development of new anti-arrhythmic drugs for efficient and safe pharmaceutical intervention. Induced pluripotent stem cell (iPSC) have been shown to be a promising cell source for the unlimited generation of human CM in vitro, suitable for regenerative therapies, disease modeling, and for the development of drug screening assays2,3,4.
Although many different cardiac differentiation protocols exist, iPSC-derived CM still lack certain phenotypical and functional aspects that impede the in vitro and in vivo application5,6. Beside electrophysiologic, metabolic, and molecular changes, the cardiac maturation process involves the structural organization of sarcomeres, which are the fundamental units required for force generation and cell contraction7. While adult CMs exhibit a well-organized contractile apparatus, iPSC-derived CMs commonly demonstrate disarranged sarcomere filaments, associated with a reduced contraction ability and altered contraction dynamics8,9. In contrast to mature CM that show uniaxial contraction pattern, the disoriented structures in immature CM results in a radial contraction of the whole cell or promote the appearance of contraction focal points9,10.
For improving cardiac maturation, multiple approaches have been applied, including 3D cell culture methods, electrical and mechanical stimulation, as well as the use of extracellular matrices mimicking in vivo conditions11,12,13. To evaluate the success and efficiency of these different culture conditions, techniques are needed to monitor and estimate the degree of the structural maturation of iPSC CM, e.g., by microscopic techniques. In contrast to conventional confocal imaging, the resolution in case of photoactivated localization microscopy (PALM) is approximately 10x higher. This technique in turn allows for a more accurate analysis, detecting even subtle alterations of cellular structures14. Considering the high resolution of PALM-based imaging, the overall goal of this method was the microscopic evaluation of sarcomere maturity in iPSC-derived CMs by precise determination of z-Disc thickness and sarcomere length. In previous studies, these structural features have been shown to be appropriate parameters to assess cardiac maturity15. For example, diseased iPSC-CM lacking full length dystrophin exhibit reduced sarcomere length and z-band width when compared to wild type cells16. Likewise, the length of individual sarcomeres was measured to investigate the impact of topographic cues on cardiac development16. Hence, we applied this approach to evaluate the structural maturation of the sarcomere network in iPSC-CM by quantitatively measuring the sarcomere length and z-disc thickness.
All steps in this protocol involving neonatal and adult mice were performed according to the ethical guidelines for animal care of the Rostock University Medical Centre.
1. Cultivation and dissociation of iPSC-derived cardiomyocytes
2. Adult cardiomyocyte isolation
3. Isolation and cultivation of neonatal cardiomyocytes
4. Immunofluorescence labeling of the α-actinin network
NOTE: For optimal results, cells are cultured in 8 well glass bottom chambers. Labeling should be performed one day before imaging.
5. Preparation of the PALM imaging buffer
NOTE: It is critical to freshly prepare the PALM imaging buffer for each experiment.
6. PALM image acquisition
7. Reconstruction of PALM data
8. Analysis of sarcomere filaments
For estimating the degree of structural maturation of CM, neonatal, fully mature adult, and iPSC CM were initially labeled the CM with α-actinin antibody to visualize the sarcomere network. Following PALM acquisition, images were reconstructed, and z-disc thickness was measured using plugin-based image processing software for the automatic detection of the width of individual filaments. Sarcomere length was calculated by measuring the distance between two adjacent intensity peaks, corresponding to neighboring filaments. Figure 1 shows the evaluation of the sarcomere organization in iPSCs-derived CM.
As presented in Figure 2A, iPSC-CM and neonatal cells were found to exhibit a similar α- actinin pattern with irregular, disarranged sarcomere structures. Likewise, quantitative assessment demonstrated that the length and thickness of α- actinin filaments were almost identical which indicates a premature developmental state of iPSC CM. More precisely, the average sarcomere length was about 1.83 µm (adult vs. iPSC CM vs. neonatal CM: 1.91 ± 0.02 1.83 ± 0.049 µm vs. 1.82±0.03 µm, n=20 cells), while z-Disc thickness was approximately 74 nm (adult vs. iPSC CM vs. neonatal CM: 71.30 ± 1.64 vs. 73.95 ± 0.86 nm vs. 74.08 ± 0.12 nm, n=20 cells) (Figure 2B). In contrast, adult mature CM showed a regular sarcomere network with slightly increased sarcomere length and reduced z-Disc thickness.
Comparison of conventional confocal imaging and PALM demonstrated no significant difference of sarcomere length (Figure 2C) (confocal vs. PALM: 1.75 ± 0.02 vs. 1.70 ± 0.02, n=10 cells). However, a profound reduced z-Disc thickness was detected when iPSC-CM were subjected to PALM imaging (Figure 2C) (confocal vs. PALM: 224.71 ± 4.31 vs. 73.91 ± 1.31, n=10 cells). Representative images highlight the gain in resolution when PALM was applied, supported by corresponding intensity plots (Figure 2D,E). Calculated full width at half maximum of fluorescence intensity revealed that α-actinin structures are ~3-fold thinner in PALM images if compared to standard confocal microscopy (Figure 2E).
Single molecule localization microscopy, like PALM, enables the detection of intracellular structures far below the diffraction limit. To ensure maximum spatial resolution, appropriate imaging conditions are required for precise detection of single molecule localization. The used imaging buffer system is critical for this acquisition process as it influences the photophysical properties of the fluorescent dye and, thus, has a significant impact on the overall resolution and accuracy of the final PALM image. Comparison between high quality buffer and buffer prepared one day before imaging revealed a profound difference in filament thickness (Figure 3A,B). Sarcomere structures acquired with low quality buffer appear to be thicker when compared to optimal imaging conditions (Figure 3A). Indeed, quantitative evaluation showed that z-Disc thickness was increased by ~65% (high quality vs. low quality: 73.87 ± 1.02 nm vs. 113.9 ± 1.33 nm, n=155 filaments) (Figure 3B). This lack of data accuracy is due to reduced blinking properties of the fluorophore that results in less detected photons per localization event (high vs. low buffer quality: 29689 photons/event vs. 16422 photons/events) (Figure 3C). Moreover, localization precision is decreased under deteriorated imaging conditions (high quality vs. low quality: 14.25 ± 5.85 nm vs. 19.56 ± 6.7 nm), which lowers overall resolution of the reconstructed PALM image (Figure 3C).
In addition, sample drift, e.g., caused by thermal instability, can affect precise localization of fluorescent molecules and results in blurry images, as presented in Figure 3D. While optimal PALM imaging gives clear and well-defined intensity peaks, excessive sample drift produces an irregular intensity pattern that makes it difficult to accurately determine the distance between two adjacent filaments. These data highlight the importance of tightly controlled imaging conditions in single molecule localization microscopy as even subtle changes during the acquisition process can dramatically decrease image quality and data accuracy.
Figure 1: Evaluation of the sarcomere organization in iPSCs-derived CM.
The sarcomere network was fluorescently labeled with an α-actinin antibody, followed by PALM image acquisition. Subsequent reconstruction leads to the final PALM image that was used for quantitative analysis. The length of individual sarcomeres was determined by measuring the distance between two intensity peaks corresponding to neighboring sarcomere filaments (1-5). Z-Disc thickness was automatically calculated by a plugin-based image processing tool. Scale bar 10 µm. Please click here to view a larger version of this figure.
Figure 2: Quantitative comparison of z-disc thickness and sarcomere length in CM derived from iPSCs, adult, and neonatal heart tissue.
(A) Reconstructed PALM images of the α-actinin network of iPSC and neonatal CM. (B) Quantitative assessment revealed that all neonatal and share high similarity in sarcomere length (adult vs. iPSC CM vs. neonatal CM: 1.91 ± 0.02 1.83 ± 0.049 µm vs. 1.82±0.03 µm) and thickness of individual filaments (adult vs. iPSC CM vs. neonatal CM: 71.30 ± 1.64 vs. 73.95 ± 0.86 nm vs. 74.08 ± 0.12 nm), indicating the premature phenotype of iPSC CM. (C) Comparison of sarcomere length and z-Disc thickness between conventional confocal imaging and PALM-based data acquisition. As sarcomere length was determined by measuring the peak-to-peak distance, the impact of increased resolution on data accuracy was less pronounced (confocal vs. PALM: 1.75 ± 0.02 vs. 1.70 ± 0.02). In contrast, a significantly lower z-Disc thickness was detected when PALM was applied (confocal vs. PALM: 224.71 ± 4.31 vs. 73.91 ± 1.31). (D) Representative microscopic images of iPSC-CM obtained by confocal and PALM imaging. (E) Fluorescence intensity plots corresponding to the red line shown in (D). Values represent full width at half maximum of fluorescence intensity, indicating a profound increase of resolution in PALM images. Data are presented as mean ± SEM, n=10-20 cells, 20 sarcomeres per cell were evaluated. Statistical significance was determined using students t-Test. **p<0.005, Scale bar 10 µm. Please click here to view a larger version of this figure.
Figure 3: Impact of buffer quality and sample drift on data accuracy and reliability.
(A) Representative PALM images of iPSC-derived CM acquired under different imaging conditions. Sarcomere filaments appear thicker in samples that have been imaged with buffer of low quality. Red lines indicate representative measurement of sarcomere length, while green filament structures were included into z-Disc analysis. (B) Quantitative evaluation confirmed a significant difference in z-Disc thickness between low- and high-quality buffer (high vs. low buffer quality: 73.87 ± 1.02 nm vs. 113.9 ± 1.34 nm) (C) This difference in image accuracy was based on a reduced blinking capability of the fluorophore, leading to a reduced number of detected photons per molecule (high vs. low buffer quality: 29689 photons/event vs. 16422 photons/events). At the same time, localization precision decreased which in turn lowered overall resolution (high vs. low buffer quality: 14.25 ± 5.85 vs. 19.56 ± 6.7) (D) Likewise, excessive sample drift can impair image quality. While proper image acquisition without sample drift results in well-defined intensity peaks of α-actinin filaments, increased sample drift provokes an irregular intensity pattern, which strongly affects accurate analysis of sarcomere length. Data are presented as mean ± SEM. 155 filaments were analyzed. Statistical significance was determined using students t-Test. ****p<0.0001. Scale bar 10µm. Please click here to view a larger version of this figure.
The generation of functional iPSC-derived CM in vitro is important for regenerative therapies, disease modeling and the development of drug-screening platforms. However, insufficient maturity of these CM is a major obstacle in cardiovascular research20. In this regard, high resolution imaging techniques are needed that enable monitoring of the structural maturation state of iPSC-derived CM. At the same time, super resolution microscopy can be a valuable tool to precisely analyze the function of specific proteins, required for proper sarcomere formation as recently demonstrated for titin and troponin10,21,22.
In the current protocol, we present a PALM-based approach to quantitatively evaluate the structural maturation of iPSC CM by analyzing the α-actinin sarcomere network.
Compared to the conventional light microscopy, PALM enables visualization of cellular structures with a resolution of ~20-50 nm23. Thus, it allows detecting even subtle alterations of the cardiac sarcomere network that are barely or not even detectable by classical light microscopy approaches. Applying PALM, we have measured an average sarcomere length of ~1.84 µm in iPSC and neonatal CM (Figure 2). This is in line with several previous reports showing that the size of individual sarcomeres is ~1.7-2.0 µm24,25,26. Compared to sarcomere length, precise estimation of the thickness of z-lines is more complicated as its size is far below the classical resolution limit of light microscopy. Using electron microscopy, previous studies revealed that z-Disc thickness ranges from 50 to 80 nm in iPSC CM, which is similar to our PALM-based detection of ~73 nm (Figure 2).
Comparison with standard confocal microscopy demonstrated the dramatic increase in resolution when PALM was applied. We found a 3-fold lower z-Disc thickness, following PALM imaging (Figure 2). However, no significant difference was measured for the sarcomere length. Since this parameter is measured by detecting the peak-to-peak distance of intensity plots, the effect of increased resolution is less pronounced.
To achieve this high spatial resolution, PALM requires well-defined imaging conditions that need to be carefully addressed by the user. For an accurate localization of single molecules, fluorophores are needed that possess special photophysical properties, enabling rapid switching between fluorescent and dark state, called blinking27. As previously demonstrated, the selection of fluorophores can strongly affect image quality28. A profound blinking capability was described for Alexa 647, one of the best and widely used dyes for single molecule localization microscopy29,30,31. However, since numerous PALM fluorophores are available, users will have high flexibility in terms of sample labelling27,28.
Besides selection of suitable fluorophore, the imaging buffer system is another critical point in PALM imaging as it dictates the photophysical properties of the fluorescent dye. We have applied pyranose oxidase as an oxygen scavenger system that is superior to glucose oxidase as it provides an increased pH stability, hence, allowing long-term imaging without a significant decrease of fluorophore blinking over time32,33. Yet, as the lifetime of the imaging buffer is limited to several hours, it must be freshly prepared for each experiment to ensure high reproducibility. Our results demonstrate a dramatic decrease of data accuracy following PALM imaging with low quality buffer, leading to impaired blinking capability and reduced localization precision (Figure 3A-C).
Moreover, thermal stability of the imaging system is mandatory to avoid increased sample drift. While moderate drift can be corrected by computational analysis, excessive sample movement leads to miscalculated localization of detected fluorophores and reduces image quality and data accuracy. This can be prevented by using microscopes with heating chambers and sufficient thermal equilibration of the respective sample before starting PALM imaging. In addition, labeling density and efficiency as well as the use of antibody fragments or nanobodies should be considered to optimize imaging conditions34,35,36.
The acquisition process of large cellular structures may take 10-30 min, depending on the imaging parameters (field of view, fluorescent probe, imaging buffer etc.). This long acquisition time is a drawback of PALM which makes it less appropriate for live cell imaging. Typically, 5.000-10.000 frames are captured to achieve an adequate number of blinking events with high localization precision. Also, the imaging depth is usually limited to few hundred nanometers, which makes it difficult to investigate thicker samples. However, enhanced resolution in z-direction can be obtained with a 3D PALM setup37.
We have selected two parameters to characterize the α-actinin network of iPSC CM, including sarcomere length and z-Disc thickness. Additional features could be collected to get a more comprehensive view on the sarcomere scaffold, such as filament orientation. Likewise, 3D PALM can help to quantitatively analyze the whole sarcomere structure by estimation of present nodes and branches within the network.
Although PALM enables a very detailed analysis of the sarcomere structure, this method does not allow the acquisition of functional parameters like cellular contractility, which is another important feature to evaluate cardiac maturity. Former reports have shown that microscopy-based assessment of sarcomere structure can be combined with video analysis to obtain functional data38,39. However, since the authors have used conventional fluorescence microscopy obtained data about the sarcomere structure are less accurate if compared with PALM. Our approach also provides the possibility to correlate PALM data with previous time lapse recordings and therefore allows the acquisition of contraction measurements.
In summary, this protocol provides a method to quantitatively evaluate the structural maturation of the sarcomere network in CM. Using this super resolution-based approach, strategies can be establish targeting an improved cardiac development of iPSC-derived CM to obtain a more adult-like phenotype.
The authors have nothing to disclose.
This study was supported by the EU structural Fund (ESF/14-BM-A55-0024/18). In addition, H.L. is supported by the FORUN Program of Rostock University Medical Centre (889001 and 889003) and the Josef and Käthe Klinz Foundation (T319/29737/2017). C.I.L. is supported by the Clinician Scientist Program of the Rostock University Medical Center. R.D is supported by the DFG (DA1296/6-1), the DAMP foundation, the German Heart Foundation (F/01/12) and the BMBF (VIP+ 00240).
We thank Madeleine Bartsch for her technical support in iPSC cell culture and cardiac differentiation.
human iPSC cell line | Takara | Y00325 | |
µ-Slide 8 Well Glass Bottom | ibidi | 80827 | |
0.5ml eppendorf tube | Eppendorf | 30121023 | |
Bovine serum albumin | Sigma Aldrich | A906 | |
Cardiomyocyte Dissociation Kit | Stem Cell Technologies | 05025 | |
Catalase | Sigma Aldrich | C40-1G | |
Cyclooctatetraene | Sigma Aldrich | 138924-1G | |
Cysteamine | Sigma Aldrich | 30070-10g | |
Dulbecco's phosphate-buffered saline without Ca2+ and Mg2+ | Thermo Fisher | 14190169 | |
F(ab')2-Goat anti-Mouse IgG Alexa Fluor 647 | Thermo Fisher | A-21237 | |
Fiji image processing software (Image J) | |||
Glucose | Carl Roth | X997.2 | |
Hydrochloric acid | Sigma Aldrich | H1758 | |
LSM 780 ELYRA PS.1 system | Zeiss | ||
Paraformaldehyde | Merck | 8187150100 | |
Pyranose oxidase | Sigma Aldrich | P4234-250UN | |
sarcomeric α-actinin antibody [EA-53] | Abcam | ab9465 | |
Sodium chloride | Sigma Aldrich | S7653 | |
sterile water | Carl Roth | 3255.1 | |
Triton X-100 | Sigma Aldrich | X100 | |
Trizma base | Sigma Aldrich | T1503 | |
β-Mercaptoethanol | Sigma Aldrich | 63689 |