Here, we present a protocol for rapid muscle fiber analyses, which allows improved staining quality, and thereby automatic acquisition and quantification of fiber populations using the freely available software ImageJ.
Quantification of muscle fiber populations provides a deeper insight into the effects of disease, trauma, and various other influences on skeletal muscle composition. Various time-consuming methods have traditionally been used to study fiber populations in many fields of research. However, recently developed immunohistochemical methods based on myosin heavy chain protein expression provide a quick alternative to identify multiple fiber types in a single section. Here, we present a rapid, reliable and reproducible protocol for improved staining quality, allowing automatic acquisition of whole cross sections and automatic quantification of fiber populations with ImageJ. For this purpose, embedded skeletal muscles are cut in cross sections, stained using myosin heavy chains antibodies with secondary fluorescent antibodies and DAPI for cell nuclei staining. Whole cross sections are then scanned automatically using a slide scanner to obtain high-resolution composite pictures of the entire specimen. Fiber population analyses are subsequently performed to quantify slow, intermediate and fast fibers using an automated macro for ImageJ. We have previously shown that this method can identify fiber populations reliably to a degree of ±4%. In addition, this method reduces inter-user variability and time per analyses significantly using the open source platform ImageJ.
Skeletal muscle composition undergoes profound changes during physiological processes such as aging1,2, exercise3,4,5,6,7, or pathophysiological processes such as disease8,9,10 or trauma11. Hence, several fields of research concentrate on the structural effects of these processes to understand functional changes. One of the key aspects determining muscle function is the composition of muscle fibers. Muscle fibers express different myosin heavy chain (MHC) proteins and are thereby classified into slow, intermediate, or fast fibers7,12,13,14,15,16,17. Physiologically, muscles have different muscle fiber compositions depending on their function in the body. Using muscle fiber typing, fiber populations can be quantified to identify adaption to physiological or pathophysiological processes7,17. Historically, a number of time-consuming methods have been applied to differentiate between muscle fiber types. For this purpose, muscle fibers were classified either by reactivity of myosin ATPase at various pH levels or muscle enzyme activity. As different fiber qualities could not be assessed in a single section, multiple cross sections were required to identify all muscle fibers and allow manual quantification14,16,17,18,19,20,21,22. In contrast, recent publications used immunohistochemistry (IHC) against myosin heavy chain protein to rapidly stain multiple fiber types in a single cross sections. Based on the advantages of this procedure, it is now considered the gold standard in muscle fiber population analysis19,23,24. Using improved IHC staining protocols, we were recently able to show that the fully automatic acquisition of whole muscle cross sections and subsequent automatic muscle fiber quantification is feasible using the open source platform ImageJ. Compared to manual quantification, our procedure provided a significant decrease in time (approximately 10% of manual analyses) required per slide while being accurate to ±4%25.
The overall goal of this method is to describe a rapid, reliable, user-independent guide to automatic muscle fiber quantification in whole rat muscles using an open source platform. In addition, we describe potential modifications that would permit its use for other specimens such as mice or human muscles.
All procedures including animal subjects were conducted in compliance with the principles of laboratory animal care as recommended by FELASA26. Approval was obtained prior to the study by the institutional review board of the Medical University of Vienna and the Austrian Ministry for Research and Science (BMWF: Bundesministerium fuer Wissenschaft und Forschung, reference number: BMWF-66.009/0222-WF/II/3b/2014).
1. Muscle Harvest
NOTE: A previous publication by Meng et al.27 is available describing the correct freezing of muscle specimens in great detail.
2. Staining
NOTE: A large number of antibodies against myosin heavy chain proteins are available; however, high quality antibodies are essential for automated acquisition and analyses. For dilutions and reference to antibodies, see Table 1. For a spreadsheet file to calculate the correct dilutions and see the number of required solution quantities, see Supplemental File 1.
3. Microscopy
4. Automated Fiber Analysis
NOTE: The macro can be obtained from the following web page: https://www.meduniwien.ac.at/hp/bionicreconstruction/macro/
Whole rat muscle cross sections were stained rapidly using immunohistochemistry to identify MHC I, IIA and IIB muscle fibers. Using a fluorescent microscope slide scanner, entire cross sections were then automatically acquired for automated muscle fiber analyses with ImageJ. The concept of the procedure is based on providing a simple, reliable and time-efficient workflow for quantification of muscle fibers.
The procedure's workflow (Figure 1) starts with the correct removal and freezing of muscle samples, as previously described in a previous tutorial by Meng et al.27. This is essential for adequate staining quality. In the next step, the muscle samples are cut and stained using primary antibodies against myosin heavy chain proteins class I, IIA and IIB and appropriate secondary fluorescent antibodies. For optimal staining quality, the manufacturer's recommended concentrations for the secondary antibodies were doubled and Triton X used for the antibody cocktails to provide even staining quality. The immunofluorescent staining of muscle fibers is quick with minor cross-reactivity and can be accomplished in approximately 5 h for batches of up to 26 slides in our setup. Resulting images showed high contrast between positive fibers and surrounding tissue and a strong distinction between different fiber types (Figure 2). Next, image acquisition is performed using a slide scanner that permits automatic scanning of entire muscle cross sections. Here, up to 12 slides are scanned automatically, for example, overnight. Hereby, overview images of the cross sections are created in high detail with a resolution that permits inspection of single fiber morphology (Figure 2).
In the final step, individual channels of each cross section are exported for quantification of muscle fiber populations using a specifically designed macro for ImageJ. The automatic analysis takes approximately 5-10 min for a rat biceps cross section containing between 3,000-6,000 fibers, whereas the manual analyses took approximately 60 min25. Automatic analyses depend strongly on computing power of the workstation and file size. As previously shown, the macro is able to detect relative muscle fiber populations with an accuracy of ± 4% compared to manual analysis (Figure 3). Overall absolute counts are higher compared to manual analyses, however, relative counts remained similar within a ±4% range, as previously shown.
Figure 1: Workflow of the Muscle Fiber Quantification Protocol. First, muscle samples are obtained from the animal and frozen correctly (see JoVE video by Meng et al.27). Then cut muscle cross sections and conduct immune staining. Acquire images of stained cross sections using a slide scanner for detailed whole cross sections images. Run the quantification macro for ImageJ for muscle fiber population analyses. Please click here to view a larger version of this figure.
Figure 2: Rat Biceps Muscle Cross Section. Overview of a stained rat biceps muscle cross section: The cross section shows slow MHC I fibers in yellow, intermediate MHC IIA fibers in green and fast MHC IIB fibers in red. Please click here to view a larger version of this figure.
Figure 3: Muscle Fiber Analyses Results of Rat Cross Section in Figure 2. Analyses of muscle MHC I, MHC IIA and MHC IIB fibers. Top: Absolute counts of muscle fibers by manual or automatic count. Middle: Relative portions of muscle fibers populations after manual quantification. Bottom: Relative portions of muscle fibers populations after absolute quantification. Please click here to view a larger version of this figure.
Table 1: Immunohistochemical Staining of MHC Fibers. Description of primary and secondary antibodies including dilutions used for the staining procedure. (Modified from Bergmeister et al.25)
Here, we demonstrate a widely accessible methodology to study and automatically quantify the muscle fiber populations of rat cross sections via immunohistochemistry in a time efficient manner. For reproducibility, we present a detailed step by step description and potential modifications for applications in other species not described in this study. Furthermore, we discuss the advantages of the procedure, prerequisites for optimal function and its limitations.
Currently, a number of staining methods exist to identify muscle fiber populations and a number of methods have been described for manual or semi-automatic quantification of muscle fibers17. However, no protocols exist that combine both and describe a standardized approach to reliably stain muscle fiber populations and automatically acquire and quantify muscle fibers, using freely available analyses software. Based on previous publications24,29, we optimized existing myosin heavy chain immunohistochemistry staining procedures for automated analyses of fiber populations in murine muscle cross sections. Thereby, we recently showed that the automated analysis of whole cross sections is feasible, reliable compared to manual analyses, and more time-efficient27. In addition, the method is user-independent thus reducing human error and thereby providing an objective comparison as for example between different treatment groups.
The whole protocol is based on providing detailed and high-quality images of whole cross sections for the automated quantification process. ImageJ as an open source software provides a number of powerful yet simple commands, which arranged in a macro allow muscle fiber identification. However, quality differences within or between cross section need to be minimized to achieve reliable results. We have identified some critical steps that are essential to achieving the required quality. This includes the use of optimal primary antibodies. Similarly, the use of appropriate secondary antibodies providing distinct fluorescent signals was essential, and we doubled concentrations for improved signal quality. Additionally, we observed a higher and more even staining quality with the use of Triton X in the antibody cocktails. Importantly, this was not observed with the use of Tween instead. Images with suboptimal staining quality were manually adjusted for better contrast to allow automatic image acquisition. Most important for the acquisition process was the correct function of the autofocus system. For this purpose, we used a DAPI-kit to stain the cell nuclei of satellite cells, which provided a reliable landmark for the autofocus system.
For the ImageJ quantification process, optimal values were determined by comparing manual to automated results and consecutive fine-tuning of the parameters25: Threshold and Analyze particles. The resulting values were sufficient for biceps and lumbrical muscles and therefore potentially other rat muscles. Additionally, Bloemberg et al.24 have shown that the staining is applicable for other species such as humans or mice. Therefore, the entire protocol can theoretically be adapted for the use in other species, but would, however, require identifying the necessary parameters. The relevant commands are highlighted in the source code.
Further possible modifications of the procedure include the setup of the slide scanner. In our setup, we used a slide scanner system, with a 2.5X lens for the preview and 20X lens for the actual acquisition. Whereas changing the lens for the preview process will provide no significant benefit in terms of quality or speed, the acquisition can also be conducted using a 10X lens to increase the speed (and consequently decrease the resolution) or instead a 40X lens for increasing the resolution and consequently acquisition time. As is shown in Figure 2, increasing the resolution is most likely not necessary for rat muscles but may be considered for other species. Other modifications that are suitable in this setup are staining for MHC-IIX fibers or laminin to visualize the muscle's extracellular matrix. These modifications, however, require changing the acquisition setup for two additional reflectors (for example Cy7) and modifying the ImageJ macro to include additional analyses with the correct parameters to the script code.
A limitation of this technique is that the absolute fiber counts were significantly higher compared to manual counts in a previous analysis, although relative counts remained similar25. This is most likely the result of inadequate separation and thus repeated quantification of fibers. This effect is more present in fibers that are inadequately stained or show other types of artifacts. However, the number of relative fiber counts and thus fiber populations remained highly accurate. In comparison to the common practice of quantifying a defined sub-portion of a cross section, e.g. 40%, we believe that analyses of the entire cross section by the ImageJ are more accurate as a result of the heterogenic muscle fiber architecture (Figure 2). A potential limitation of the procedure is the requirement to use expensive slide scanners for automated acquisition of entire cross section, which are not generally available. An alternative approach is to acquire single field of views and quantify each individually by the macro.
In conclusion, we believe this protocol is widely accessible and allows the fast and reliable quantification of whole muscle cross sections using freely available analyses software. Therefore, the introduced procedure provides a useful contribution to the analyses of muscle fiber morphology.
The authors have nothing to disclose.
This study was supported by the Christian Doppler Research Foundation. We would like to thank Sabine Rauscher from the Core Facility Imaging at the Medical University of Vienna, Austria for support throughout the project. Primary antibodies were developed by Schiaffino, S., obtained from the Developmental Studies Hybridoma Bank, created by the NICHD of the NIH and maintained at The University of Iowa, Department of Biology, Iowa City, IA.
O.C.T compound | Tissue-Tek, Sakura, Netherlands | For embedding of muscle tissue | |
Isopentane | for adequate freezing of muscle tissue | ||
Superfrost Ultra Plus slides | Thermo Scientific, Germany | 1014356190 | adhesive slides |
phosphate buffered saline | |||
Triton X-100 | Thermo Scientific, Germany | 85112 | Detergent Soluation |
Goat serum | Thermo Scientific, Germany | 50197Z | Goat Serum |
DAKO Fluorescent Mounting Medium | Dako Denmark | S3023 | |
Dako pen | Dako Denmark | S200230-2 | |
TissueFAXSi plus | TissueGnostics, Vienna, Austria | ||
Primary antibodies | |||
MHC-I (Cat# BA-F8, RRID: AB_10572253) | Developmental Studies Hybridoma Bank (DSHB, Iowa, USA) | Supernatant | |
MHC-IIa (Cat# SC-71, RRID: AB_2147165) | Developmental Studies Hybridoma Bank (DSHB, Iowa, USA) | Supernatant | |
MHC-IIb (Cat# BF-F3, RRID: AB_2266724) | Developmental Studies Hybridoma Bank (DSHB, Iowa, USA) | Supernatant | |
Secondary antibodies | |||
Alexa Fluor 633 Goat Anti-Mouse IgG2b | Thermo Scientific, Germany | A-21146 | |
Alexa Fluor 488 Goat Anti-Mouse IgG1 (γ1) | Thermo Scientific, Germany | A-21121 | |
Alexa Fluor 555 Goat Anti-Mouse IgM (µ chain), | Thermo Scientific, Germany | A-21426 | |
NucBlue Fixed Cell ReadyProbes Reagent | Thermo Scientific, Germany | R37606 |