Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You and Your Team.

Learn More →

SMASH – semi-automatic muscle analysis using segmentation of histology: a MATLAB application

SMASH – semi-automatic muscle analysis using segmentation of histology: a MATLAB application Background: Histological assessment of skeletal muscle tissue is commonly applied to many areas of skeletal muscle physiological research. Histological parameters including fiber distribution, fiber type, centrally nucleated fibers, and capillary density are all frequently quantified measures of skeletal muscle. These parameters reflect functional properties of muscle and undergo adaptation in many muscle diseases and injuries. While standard operating procedures have been developed to guide analysis of many of these parameters, the software to freely, efficiently, and consistently analyze them is not readily available. In order to provide this service to the muscle research community we developed an open source MATLAB script to analyze immunofluorescent muscle sections incorporating user controls for muscle histological analysis. Results: The software consists of multiple functions designed to provide tools for the analysis selected. Initial segmentation and fiber filter functions segment the image and remove non-fiber elements based on user-defined parameters to create a fiber mask. Establishing parameters set by the user, the software outputs data on fiber size and type, centrally nucleated fibers, and other structures. These functions were evaluated on stained soleus muscle sections from 1-year-old wild-type and mdx mice, a model of Duchenne muscular dystrophy. In accordance with previously published data, fiber size was not different between groups, but mdx muscles had much higher fiber size variability. The mdx muscle had a significantly greater proportion of type I fibers, but type I fibers did not change in size relative to type II fibers. Centrally nucleated fibers were highly prevalent in mdx muscle and were significantly larger than peripherally nucleated fibers. Conclusions: The MATLAB code described and provided along with this manuscript is designed for image processing of skeletal muscle immunofluorescent histological sections. The program allows for semi-automated fiber detection along with user correction. The output of the code provides data in accordance with established standards of practice. The results of the program have been validated using a small set of wild-type and mdx muscle sections. This program is the first freely available and open source image processing program designed to automate analysis of skeletal muscle histological sections. Keywords: Histological muscle analysis, Standardized quantitative analysis, Image segmentation, mdx mouse Background a widely available image processing software package Skeletal muscle has a robust ability to adapt to the pat- specifically designed for muscle histological analysis. tern of use and to regenerate following injury. These are Altering muscle fiber size is one of the primary methods often quantified using histological techniques. However, in which muscle responds to external stimuli. Muscle the methods for this quantification remain disparate mass may be increased in response to resistance training among investigators and often require painstaking man- [3] or with potential pharmacological agents like myosta- ual procedures [1,2]. The goal of this work is to provide tin inhibitors [4], while muscle atrophy occurs in response to disuse [5] and injuries such as denervation [6]. These conditions primarily reflect hypertrophy or atrophy of * Correspondence: smithlu@dental.upenn.edu Department of Anatomy and Cell Biology, School of Dental Medicine, individual fibers rather than hyper- or hypoplasia [7]. University of Pennsylvania, Philadelphia, PA, USA Muscle fiber size is routinely evaluated using fixed or fro- Pennsylvania Muscle Institute, University of Pennsylvania, Philadelphia, PA, zen tissue sections. Fiber outlines are visualized using a USA © 2014 Smith and Barton; licensee BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly credited. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated. Smith and Barton Skeletal Muscle 2014, 4:21 Page 2 of 15 http://www.skeletalmusclejournal.com/content/4/1/21 variety of techniques, including hematoxylin and eosin standard operating procedures developed by TREAT- staining, laminin immunostaining, dystrophin immuno- NMD when available using sophisticated boundary de- staining, and wheat germ aggluttinin staining [8]. While tection algorithms [2]. The software also includes built-in these techniques enable visualization of fiber boundaries, image editing to manually inspect and manipulate fiber determining fiber cross-sectional area (CSA) is often still boundaries. Fully automated fiber size determination as performed by manual tracing of individual fibers. There well as fiber types and CNFs may be possible with ad- are software programs available to help automate fiber equate image acquisition [9,20]. However, these newly detection, however they are often expensive and are not designed fully automated programs are not yet available specifically designed for muscle histology [9]. [9] and/or have a significant cost [20]. Additionally, Muscle fiber type distributions are often investigated allowing the user to have manual control over some as- in muscle histology as they are known to be altered in pect of image processing allows users to maintain the response to exercise, inactivity, and aging [10]. Fiber type fidelity established by manual techniques. The open na- is primarily determined by the myosin heavy chain iso- ture of this software also allows custom usage and fur- form, which has differential contractile and ATPase ac- ther advancement of the methods. For users that do not tivity. Fiber type is often determined by ATPase staining have access to a MATLAB license or the image process- [11,12] or with immunostaining for specific myosin ing toolbox we have compiled an .exe file that runs heavy chain (MyHC) isoforms individually [13,14]. How- using the freely available MATLAB Runtime Compiler ever, methods to determine fiber type can be subjective (MCR) version 8.3 (http://www.mathworks.com/products/ and tedious when fibers are manually classified. Follow- compiler/mcr/). Automating a large portion of muscle ing fiber segmentation, computing the size distribution histology makes it feasible to analyze full muscle cross- of single fiber types is easily automated. sections, eliminating variability introduced by selecting Muscle fibers also undergo changes in morphology as only a portion of the cross section for analysis. This soft- they develop. In particular, centrally nucleated fibers are ware is validated with muscles from mdx mice, which have often used as a marker for muscle regeneration. While many alterations of muscle fiber morphology compared to fully mature fibers have peripheral nuclei, newly regener- wild-type mice [21]. The purpose of this study is to de- ated fibers have central nuclei [15]. In many muscular velop freely available automatic and standardized image dystrophies, which are characterized by continual cycles segmentation platform and validate the program using of degeneration and regeneration, the number of cen- standard muscle histological analysis. trally nucleated fibers (CNFs) is substantial while CNFs are hardly present in healthy muscle. Although nuclei Implementation are easily stained with DAPI, determination of CNFs is Mice often performed manually. Combined use of automated All animal experiments were approved by the University CNF and fiber size determination allows the size of re- of Pennsylvania Institutional Animal Care and Use Com- generating fibers to be calculated, providing a measure mittee. C57Bl/6 mice were used as wild-type controls of how efficiently regeneration is occurring after acute and mdx mice were used as a dystrophic model. Both injury [16]. animal groups were analyzed at 1 year of age. Skeletal muscle is a highly metabolically active tissue requiring large blood supply. As with fiber type shifts, Immunohistochemistry capillary density of skeletal muscle may be affected by al- Soleus muscles from both groups (n = 4 per group) were tered metabolic demand or in disease [17]. Endothelial dissected, embedded in OCT, and frozen in liquid nitrogen cells and capillaries are frequently stained in skeletal cooled isopentane. Frozen 10 um sections were cut from muscle with Von Willibrand Factor or PECAM [18,19]. muscles and mounted on slides. Sections were washed in Automated determination of capillary density in relation PBS and immunostained using antibodies to laminin to fiber size and number provides a useful parameter of (Thermo Scientific), and either myosin heavy chain I (Devel- skeletal muscle histology. opmental Studies Hybridoma Bank) or platelet endothelial All of the methods discussed above are commonly per- cell adhesion molecule (PECAM; eBiosciences) overnight formed using immunofluorescence, which provides high at 4°C. After PBS wash, fluorescent secondary antibodies contrast in stained and unstained structures. We have (Sigma) were applied for 1 h at room temperature. Nuclei developed MATLAB (MATLAB and Image Processing were labeled with DAPI incorporated into the mounting Toolbox 2014a, MathWorks) scripts bundled into a media (Vectashield). Images were acquired using a Leica MATLAB App (see Availability and Requirements) that DM RBE microscope and DFC350FX camera and Open- automate, or partially automate determination of fiber Lab software. Individual fields were stitched together to size, fiber type, centrally nucleated fibers, and capillary create a composite full view of the muscle cross-section density. These programs are created to comply with using Photoshop (Adobe). Smith and Barton Skeletal Muscle 2014, 4:21 Page 3 of 15 http://www.skeletalmusclejournal.com/content/4/1/21 Image selection function. The Initial Segmentation function uses the water- The software has built in several steps of image processing shed transformation to determine the fiber edges, as it has tools within the same script. Initially the user must select become one of the most common and standard choices for an image file (.bmp, .jpg, .png, and .tif.) to be processed. Fol- image segmentation [22,23]. However, the watershed trans- lowing selection of the appropriate file the user is provided formation often leads to over-segmentation due to local a list of the built in functions (Figure 1A). A representative minima created from noise within an image [24]. Prior to image of a soleus muscle from a 1-year-old mdx mouse is applying the watershed transformation, the image is used which has been immunostained with laminin (red) smoothed with a function that suppresses local minima and slow myosin heavy chain (green) as well as DAPI (blue) whose depth is below a given threshold. Upon selecting (Figure 1B and C). Dystrophic muscle can be more difficult this function users are presented with options for segmen- to process automatically due greater interstitial spaces and tation (Figure 2A). The pixel size measured in μm/pixel is the examples highlight some manual adjustments that may requested in all functions. The segmentation takes place be required. The software includes an Excel (Microsoft) file based on a single RGB color channel, which the user pro- containing default parameter values (Table 1), which may vides designating the color of the fiber outlines. The user be altered to the needs of the user. The details of each par- also provides an initial value for suppressing local minima ameter are discussed in the relevant section below. based on 8-bit channel images (0 to 255). Previous experi- ence shows that a segmentation value of 4 to 10 produces Initial segmentation robust fiber identification, depending on the quality of Prior to any other function being run, a mask file for staining (Figure 3). However, the segmentation value may the image must be created with the ‘Initial Segmentation’ need to be adjusted based on the image quality and Figure 1 Image selection. (A) Dialog box allowing selection of a single function to perform on the selected muscle section. (B) Example muscle section of 1-year-old soleus muscle from mdx mouse. Red is laminin stain, green is slow muscle myosin heavy chain, and blue is DAPI. (C) Enlarged portion of (B). Smith and Barton Skeletal Muscle 2014, 4:21 Page 4 of 15 http://www.skeletalmusclejournal.com/content/4/1/21 Table 1 Default parameters from Excel file convexity. Interstitial space may fit the size requirements Default parameters Default values of a fiber; however, it will not have a grossly circular shape in cross section. The eccentricity takes an ellipse with the Pixel size (μm/pixel) 0.645 same second moments as the objects and is the ratio be- Fiber outline channel (red = 1, green = 2, blue = 3) 1 tween the distance between the foci of the ellipse and the Nuclei channel (red = 1, green = 2, blue = 3) 3 major axis. Thus, circles have an eccentricity of 0 and line Fiber type channel (red = 1, green = 2, blue = 3) 2 segments have an eccentricity of 1. Further, interstitial Object channel (red = 1, green = 2, blue = 3) 2 space may have a stellate shape between fibers while main- Segmentation smoothing factor 5 taining high eccentricity. The stellate shape creates regions of concavity, while polygonal fibers have minimal concav- Nuclear smoothing factor 5 ity and high convexity. The convexity is determined by the Object smoothing factor 10 ratio of the object area and the smallest convex polygon Minimum fiber area (μm ) 100 that can contain the entire region. Thus, a completely con- Maximum fiber area (μm ) 5,000 vex image will have a convexity of 1. After the filtering pa- Maximum eccentricity 0.95 rameters have been selected, a new mask is created with Minimum convexity 0.8 only the objects passing the filter (Figure 4B) [25]. At this point the user has the option to inspect the mask and se- Nuclear distance from boarder (μm) 10 2 lect any objects that may have passed the filter but are not Minimum nuclear size (μm)5 fibers (Figure 4C). This may include areas of interstitial Fiber properties output folder C:\SMASH\Output space, vessels, or other structures the user does not want Fiber type output folder C:\SMASH\Output to include (Figure 4D, yellow arrow). The zoomed ex- Central nuclei output folder C:\ SMASH\Output ample image shows fibers that were filtered because of Objects output folder C:\ SMASH\Output high eccentricity and high convexity due to improper sep- aration of fiber from interstitial space (Figure 4D, pink exposure (Figure 3). For example, a suboptimal segmenta- arrows). Following user removal of non-fiber objects the tion can generate false objects (Figure 3A, E) or merge ob- mask file is overwritten. jects (Figure 3C, E). Once initial parameters have been selected, the initial segmentation takes place with the Fiber properties segmented fiber edges overlaid on the original image After the mask has been filtered to contain only objects (Figure 2B) [25]. After inspecting the image the user can that are fibers the properties of the fibers may be deter- decide to re-run the function with alternative parameters. mined. The user is prompted for pixel size and the Once proper segmentation has been achieved, the user has folder location that the output data should be exported the option to separate any merged fibers that may have to. The output file is in Excel (.xls) format with the same arisen from poor staining. The user implements a freehand name as the image file appended by ‘_Props’ (Figure 5A). drawing tool to separate any fibers that are merged with The Feret diameter properties are obtained from an other fibers (red dotted line) or interstitial space (yellow incorporated publically available script on MATLAB dotted line) (Figure 2C). After the user has finished separ- Central [26]. The function displays histograms for fiber ating any fibers a mask file is saved. Initial segmentation minimal Feret diameter and fiber CSA for inspection will often produce many small segments in interstitial (Figure 5B). The Feret properties are defined by the space; however, non-fiber areas will be filtered in the ‘Fiber smallest rectangle that bounds the fiber [27]. The mini- Filter’ function. All fibers in the mask should be separated mum Feret diameter has been shown to be a more ro- prior to running the fiber filter function to prevent in- bust measure of fiber size. Feret diameter is much less appropriate filtering of fibers. sensitive to oblique muscle section compared to fiber CSA [2]. For each fiber the properties exported are: cen- Fiber filter troid in x and y in pixels, the maximum Feret diameter Following the initial segmentation the fiber filter function (μm), minimum Feret diameter (μm), Feret direction (ra- removes objects based on user-defined criteria (Figure 4A). dians), and fiber area (μm ). The output Excel file is then Fibers that touch the edge of the image are removed so written with all fiber properties as well as the mean, partial fibers are not included in further calculations. To standard deviation (SD), and standard error of the mean provide the proper scale, the pixel size is entered by the (SEM) for each property (Figure 5C). user in μm/pixel. Any fibers below the minimum fiber area or above the maximum fiber area entered in μm are Fiber typing removed from the mask and from further analysis. More The fiber typing function uses the fiber mask to inspect advanced filtering is also included with eccentricity and each fiber area for determination of fiber staining. The Smith and Barton Skeletal Muscle 2014, 4:21 Page 5 of 15 http://www.skeletalmusclejournal.com/content/4/1/21 Figure 2 Initial segmentation. (A) Input box with options for fiber segmentation. (B) The example image with white lines drawn along fiber boarders based on segmentation parameters. (C) Dialog box that appears after separating a fiber. (D) Zoomed in of B. Red dotted lines shows where two fibers are separated by the user. Yellow dotted lines shows where a fiber is not separated from intestinal space. function examines the average intensity of staining within positively stained fibers. Intermediate values may represent the individual fiber region determined by the mask to cal- mixed fibers. A histogram with the CSA of positive fibers culate an average intensity. A threshold value is used to de- (red, front) and all fibers (blue, back) is presented to exam- termine if the fiber is positive or negative based on the ine any fiber type-dependent size changes. After inspecting average intensity. The user inputs the pixel size, the RGB the figure the user is able to adjust the threshold value and channel which includes fiber type staining, and the output thefigureisupdated. Oncethe user is satisfiedwiththe folder location (Figure 6A). The initial threshold value for analysis they select accept and are prompted to create an average fiber type intensity is obtained using Otsu’s Excel file with each fiber’s average intensity, fiber area, and method, which calculates the threshold to minimize intra- if it is classified as positive (1) or negative (0) (Figure 6B). class variance of the thresholded binary image [28]. The Summary data including average fiber area, average fiber program then performs an initial calculation of fiber types. intensity and percent of positive fibers are included for all, A figure is displayed with the original image and an image positive, and negative fibers. where positive fibers are white and negative fibers are grey (Figure 6D). The figure includes a histogram of staining in- Centrally nucleated fibers tensities along with the threshold. The intensity histogram The CNF function again requires prior creation of a fiber should have two peaks, with the lower and sharper peak mask to analyze the centers of fibers for nuclei. The user being negative fibers and high intensity peak representing is prompted to input pixel size as well as the RGB channel Smith and Barton Skeletal Muscle 2014, 4:21 Page 6 of 15 http://www.skeletalmusclejournal.com/content/4/1/21 Figure 3 Segmentation examples. Example muscle section of 1-year-old soleus muscle from mdx mouse. Red is laminin stain, green is slow muscle myosin heavy chain, and blue is DAPI. (A) Segmentation filter value of 2 produces over segmentation of original image. (B) Segmentation filter value of 5 produces appropriate segmentation. (C) Segmentation filter value of 12 produces under segmentation of original image. (D) Reducing the brightness of laminin staining by 50% (representing poor staining) produces under segmentation. (E) Enhancing the brightness of laminin by 400% (representing over exposure) produces over segmentation. (F) Overexposure can be compensated for by increasing the segmentation filter value to 12. Yellow arrow heads show areas of over segmentation. Pink arrow heads show areas of under segmentation. that corresponds to nuclear staining (Figure 7A). To re- inspecting the image the user may adjust the threshold strict the inspection to central nuclei the nuclear staining value as desired. Mousing over smoothed nuclear image must be greater than a user-defined distance from the may provide useful information for manual threshold- border of the fiber input in μm. The user is also allowed ing. Once a proper threshold is determined, the user is to specify the minimum area of nuclei to count so that prompted if to determine if the results should be output to random staining is minimized. To minimize the effect of Excel. If the answer is yes, a file is created that contains the streaking that can be present with poor DAPI staining the fiber area, the area of positive nuclei staining in the center image can be smoothed using the same suppression of region, and if the fiber is classified as a CNF (1) or not (0) local minima used for fiber segmentation. An automatic (Figure 7B). Summary data include the average fiber area, threshold is applied after nuclei smoothing using Otsu’s area of central nuclei, and percent of CNFs for all fibers, method. Finally, the user also specifies the output folder CNFs, and non-CNFs. If parameter selection was not suffi- for the Excel data to be written to. After input selection a cient, the function can be rerun with new parameters. figure is displayed that contains the original image and an image with the borders. The distance not inspected is in Capillary density or non-fiber objects red and the nuclei are in blue, with peripheral nuclei It is often useful to identify objects within muscle sections appearing purple (Figure 7C, bottom left). The figure also that are not within fibers. The object counter was designed contains images of only the smoothed nuclear image to identify capillary density using a PECAM staining to and an image with only CNFs. In some situations the identify endothelial cells, however other cell types could be automatic thresholding of nuclei may be unsuitable. After analyzed with the same techniques, such as macrophage Smith and Barton Skeletal Muscle 2014, 4:21 Page 7 of 15 http://www.skeletalmusclejournal.com/content/4/1/21 Figure 4 Fiber filter. (A) Input box with options for filtering fibers. (B) Image generated in which fibers passing the filter are randomly colored to more easily distinguish adjacent fibers. (C) Zoomed in of (B). Yellow arrows point to possible user selections as non-fiber objects. Pink arrows show fibers that did not pass the filter. Lower left pink arrow did not pass due to improper segmentation as depicted in Figure 2D. (D) After selection of fibers the user is asked if they would like to indeed delete them or complete the function operation. infiltration [29]. The user is prompted for pixel size and the SMASH data verification RGB channel of the objects to be identified (Figure 8A). Ini- To verify the data generated by SMASH we compared tial values for smoothing the objects, identical to that for the same images used in previously published data [30]. nuclei and fiber type, with initial threshold automatically In this case legacy methods were done using Openlab generated using Otsu’s method. The user also identifies the software (Improvision, PerkinElmer) with simple thresh- folder location of the Excel file output. After initial parame- olding of individual RGB channels for fiber size and fiber ters are selected a figure is presented with a figure contain- type. For central nucleation fibers were manually deter- ing the original image, the object channel image, smoothed mined to be CNF or PNF and a marker placed using object channel image, and thresholded object image Openlab to enable counting. (Figure 8C). After inspection, the user is then prompted to change the threshold if desired. Once the user has achieved Statistical analysis a suitable threshold the user selects whether to output an The C57 and mdx output data were compared using a Excel file with the data. Each object has its area and cen- Student’s t-test. For comparing CNFs and PNFs a paired troid listed along with summary data including the total Student’s t-test was used. All data analyses were per- number of objects in the image. formed using PRISM (Graphpad Software). Smith and Barton Skeletal Muscle 2014, 4:21 Page 8 of 15 http://www.skeletalmusclejournal.com/content/4/1/21 Figure 5 Fiber properties. (A) Input box with options for fiber properties. (B) Histograms showing minimum Feret diameter (top) and fiber CSA (bottom). (C) Truncated example of Excel output from running fiber properties function. Results multitude of previous reports demonstrate that mean indi- In order to evaluate the suitability of the software described, vidual fiber cross-sectional area in mdx muscle is not skeletal muscle morphologies from a selection of wild-type consistently altered. However, as the mdx muscle is under- C57BL6 and mdx mouse soleus muscles were analyzed going continual cycles of degeneration and regeneration the (n = 4 per group). The mdx mouse is a model of Duchenne variability in fiber CSA is much greater than wild-type muscular dystrophy (DMD) that is commonly studied. A muscle [2,31]. The results produced here clearly reflect the Smith and Barton Skeletal Muscle 2014, 4:21 Page 9 of 15 http://www.skeletalmusclejournal.com/content/4/1/21 Figure 6 Fiber type. (A) Input box with options for fiber type function. (B) Truncated example of Excel output from running fiber type function. (C) Portion of immunofluorescent image. Red is laminin stain, green is slow muscle myosin heavy chain, and blue is DAPI (top left). Image showing negative fibers in gray and positive fibers in white (top right). Histogram of average fiber intensity for fiber type stain with threshold value (bottom left). Histogram of fiber CSA with positive fiber in red and all fibers in blue (bottom right). previous literature on mean CSA and variability of CSA increase in the standard deviation of Feret diameter within (Figure 9A and B). The mean fiber CSA was not different an mdx muscle (Figure 9C and D). between the C57 and mdx muscles; however the standard Fiber type determination is commonly performed in deviation (SD) of CSA within a muscle was much greater muscle histological analysis. The muscles studied here in mdx than C57. While measurements of fiber CSA have have been stained for slow myosin heavy chain (Type I). commonly been used, the fiber CSA is highly dependent on The soleus muscle has a large portion of type I fibers of the angle of the section relative to the axis of the fiber. Use approximately 40% in wild-type mice [32]. This proportion of minimal Feret diameter provides a robust measure of is close to the value obtained with the software (Figure 9E), fiber size largely independent of the angle of the section however there was a large increase in the percentage of andisendorsedas a standard operating procedure [2]. Our slow fibers in the mdx soleus muscle. This is expected results on Feret diameter align well with that of CSA show- based on the advanced age of these mice of 1 year and that ing no change in the mean Feret diameter, but a substantial type I fibers are more resistant to damage in mdx muscle Smith and Barton Skeletal Muscle 2014, 4:21 Page 10 of 15 http://www.skeletalmusclejournal.com/content/4/1/21 Figure 7 Centrally nucleated fibers. (A) Input box with options for CNF function. (B) Truncated example of Excel output from running CNF function. (C) Portion of immunofluorescent image. Red is laminin stain, green is slow muscle myosin heavy chain, and blue is DAPI (top left). Image showing filtered nuclei (top right). Image with boarder regions in red and nuclei above threshold in blue (bottom left). Image in which only CNFs are depicted in white (bottom right). [33]. Determination of the size of various fiber types is an CNFs. The percentage of CNFs in dystrophic muscle is a automatic feature of the software. Assuming all unlabeled commonly measured histological marker of disease [34]. fibers are type II fibers the results show that C57 type I fi- The automatic detection demonstrated the expected re- bers are smaller than type II and the relationship is not sults with C57 soleus muscle showing very rare CNFs different in mdx soleus muscles (Figure 9F). while >30% of mdx soleus muscle fibers were CNFs Due to the continual cycles of degeneration and regen- [33,35] (Figure 9G). While CNF percentage is routinely eration, dystrophic muscles have a high prevalence of performed, often using manual methods, reporting on CNFs, while uninjured wild-type muscles have very few the size of the CNFs is less common. The software Smith and Barton Skeletal Muscle 2014, 4:21 Page 11 of 15 http://www.skeletalmusclejournal.com/content/4/1/21 Figure 8 Object counter. (A) Input box with options for object counter function. (B) Truncated example of Excel output from running object counter function. (C) Portion of immunofluorescent image. Red is laminin stain, green is PECAM, and blue is DAPI (top left). Image showing only the objects of interest, here PECAM from the green channel (top right). Image showing objects of interest after smoothing filter is applied (bottom left). Binary image showing discrete objects that pass threshold value (bottom right). makes it trivial to measure the size of the CNF popula- process each function in SMASH compared to legacy tion. The data collected show that within mdx muscle methods is reported in Table 3. Smash demonstrates CNFs are larger than peripherally nucleated fibers greatly reduced time of analysis for the muscles tested. (PNFs) (Figure 9H). This relationship has been previ- ously observed in the EDL muscle of mdx mice [16,36]. Discussion The difference between data output from legacy The manual analysis of skeletal muscle immunofluores- methods and SMASH for fiber size, fiber size variability, cence is often a laborious task. The Semiautomatic fiber type percentage, and CNF percentage are presented Image Processing of Skeletal Muscle Histology Software in Table 2. There was little discrepancy between values described and tested provides researchers with a valuable obtained for fiber type percentage or CNF percentage. tool for measuring multiple facets of muscle histology. However, the calculated fiber areas were consistently While image analysis software is available to conduct higher using SMASH compared to legacy methods. This many of these features, it carries a high monetary cost is due to the greater and more variable fiber border re- and is not specifically designed for skeletal muscle ana- gion using simple thresholding in legacy systems com- lysis. In contrast, this program is available on the widely pared to SMASH (Figure 10). The time required to used MATLAB software and is designed to investigate Smith and Barton Skeletal Muscle 2014, 4:21 Page 12 of 15 http://www.skeletalmusclejournal.com/content/4/1/21 Figure 9 Comparison of output data from 1-year-old C57 and mdx mouse soleus muscle. (A) Mean of cross-sectional area of fibers in C57 and mdx muscle sections. (B) Standard deviation of CSA of fibers. (C) Mean of minimum Feret diameter (MFD) of fibers. (D) Standard deviation of MFD diameter. (E) Percentage of Type I fibers in C57 and mdx muscle sections. (F) The ratio of mean CSA of type I fibers to type II fibers. (G) The percentage of fibers with centrally nucleated fiber (CNF)s. (H) The mean CSA of mdx peripherally nucleated fibers (PNF) and CNFs. *P <0.05 for mdx compared to C57, † P <0.05 for CNFs compared to PNFs within the same muscle. specific features of muscle histology, and is open to cus- with much improved reliability over standard threshold- tom modifications for advanced users. ing. However, the software maintains the manual ability All of the functions in the program rely and are based to correct image segmentation. Furthermore, if the built- on the segmentation of skeletal muscle fibers in an in manual editing features are not suitable, the user may image. This is commonly done with manual tracing, edit the created mask file with software of his choice, which is very time-consuming, or basic thresholding of such as Adobe Photoshop. After the mask has been fi- fiber outlines, which requires extensive manual correc- nalized, running the analysis functions is expedient. The tion on all but the most pristine sections. Using a com- user is required to enter parameters for analysis and bination of the image smoothing fiber and the watershed may adjust them based on the resulting image guides. transform fiber boundaries are automatically produced To analyze the entire cross-section of 1-year-old soleus muscles for fiber size, fiber type, and CNFs from the Table 2 Data comparison from legacy methods to SMASH same image took approximately 15 min, substantially Image Legacy Method SMASH Difference less than manual methods. Increased time efficiency al- Fiber type - myosin IIa (%) lows analysis of the whole muscle cross-section, avoiding the sampling issues of only using select fields of view. A59 58 -1 Using the whole muscle cross-section is required to B60 57 -3 meet TREAT-NMD standards of practice [37]. Analyzing C58 61 3 more fiber and eliminating regional differences in cross- Fiber size - fiber area (μm ) section improves the consistency of the results. Highly A 708 898 189 significant results were obtained with low variability in B 745 934 189 our investigation with only four muscles per group. The output of fiber size data in Excel format permits C 622 819 198 the use of many graphing tools to create plots of fiber Fiber size - fiber area SD (μm ) size commonly used. Ensuring the ability to measure A 303 415 112 Feret diameter ensures compliance with the latest stand- B 265 346 81 ard operating procedures for fiber size analysis [2]. C 219 339 121 Automation of fiber type data with defined boundaries Centrally nucleated fibers (%) has advantages over using thresholding to determine fiber type and size. Instead of determining the fiber size A71 66 -5 by the area above a threshold value of staining, the area B58 55 -3 is determined entirely by the fiber outline so the positive C61 60 -1 fiber size will not be affected by blotchiness in fiber type Smith and Barton Skeletal Muscle 2014, 4:21 Page 13 of 15 http://www.skeletalmusclejournal.com/content/4/1/21 Figure 10 Comparison of mask file from legacy method and SMASH. (A) Original image of a laminin stained muscle section. (B) Fibers are colored randomly from SMASH output mask with dark regions corresponding to fiber area using a simple threshold. Dark gray areas are interstitial in SMASH and fibers using legacy methods while white areas are interstitial in both masks. Figure demonstrates the larger fiber area obtained with SMASH compared to the legacy method of using a simple threshold for fiber area. staining. The fiber type function may also be used for done in SMASH provides signal that is more robust to additional signals internal to fibers such as Evan’s Blue these parameters. SMASH provides a mask more con- Dye permeability or IgG infiltration of necrotic fibers sistent with manual tracing of fibers than applying a [2,38]. Automatic determination of CNFs provides a use- threshold. While SMASH reduces the border region be- ful measure of skeletal muscle histology. Although it is tween adjacent fibers, it is also clearly capable of delin- not frequently studied, combining fiber size with CNF eating interstitial space between adjacent fibers when determination may provide a more informative marker there is an appreciable separation as evidenced by the of muscle health [16]. The computation of this add- white and grey areas in Figure 10. Thus, we attest that itional parameter is trivial using this software. Analysis SMASH generally provides a more robust and accurate of structures outside of fibers themselves in skeletal fiber size as well as requiring less manual editing than muscle is also important. The object counter function was using simple thresholds. It is also noteworthy that fiber designed to automate analysis of capillary density, as is area is generally more variable than Feret diameter as done in muscle histological analysis for capillaries per fiber for perfectly round areas it is squaring the difference and capillaries per area [19]. However, it may also be used and the more elongated the fiber the greater the propor- to analyze a multitude of other structures within a muscle, tion of border region that may influence the results. such macrophage infiltration or matrix proteins. In addition the gains in robust analysis Table 3 demon- The data generated by SMASH are validated against strates that SMASH greatly reduces the time required to legacy methods showing largely consistent results be- analyze images. Manual editing of the fiber mask is still tween methods for fiber type and CNF percentages required in the majority of muscle sections for both ini- (Table 2). However, the discrepancy in fiber area illus- tial segmentation and fiber filtering and take the major- trates the importance of using a single analytical method ity of the processing time. However, with SMASH this for a given study and highlights the variability between manual editing is reduced to just a few minutes in the different approaches. Using a simple threshold to separ- case of the soleus muscles tested. The time gains are es- ate fibers creates highly variable borders between adja- pecially significant when doing multiple analyses on the cent fibers that are dependent on parameters such as same image, as manually editing the mask is the major exposure time and focus during image acquisition. Fil- time consumer and additional functions are able to be tering the signal and using the watershed function as is processed in just seconds. Table 3 Run time (mm:ss) comparisons from legacy methods to SMASH Image Legacy method - functions SMASH - functions Diff a a FP FT CNF TOT IS FF FP FT CNF TOT TOT A 30: 10: 16:30 56:30 1:50 1:49 0:06 0:17 0:12 4:14 52:16 B 30: 10: 13:50 53:50 2:46 3:42 0:05 0:26 0:15 7:14 46:36 C 30: 10: 13:30 53:30 1:45 2:39 0:07 0:14 0:10 4:55 48:35 Times are approximate. CNF functions were performed on different sections than other functions. CNF: Centrally nucleated fibers function; FF: Fiber filter function; FP: Fiber properties function, including fiber size; FT: Fiber typing function; IS: Initial segmentation function; TOT: Total time for analysis. Smith and Barton Skeletal Muscle 2014, 4:21 Page 14 of 15 http://www.skeletalmusclejournal.com/content/4/1/21 While this software provides many advantages there previously incorporated functions. Automated functions are notable limitations. The analysis is currently limited improve the speed and consistency of skeletal muscle to immunofluorescence and is not compatible with com- histological analysis. Although it requires a MATLAB li- mon stains such as hematoxylin and eosin. The software cense, this is the only freely available software designed for does often require manual segmentation and filtration of the analysis of skeletal muscle histology. fibers based on the current algorithms. These manual adjustments allow the user more control over the ana- Availability and requirements lysis, but also increase the time required for analysis and Project name: SMASH - Semiautomatic Image Process- introduce a degree of subjectivity. Manual adjustments ing of Skeletal Muscle Histology: a MATLAB Application. are more frequently required in analysis of diseased Project homepage: http://dx.doi.org/10.6084/m9.figshare. muscle as well. While these algorithms are advanced compared to many of the techniques currently in use, Operating System: Platform Independent. extending fiber segmentation algorithms may provide Programming Language: MATLAB. more reliable boundaries. The proposed method of Other requirements: SMASH Stand Alone (SMASH_ measuring fiber type is limited to investigation of a sin- Installer.exe) requires MATLAB Compiler Runtime (MCR) gle fiber type per image, or per color in an RGB image. version R2014a (8.3) which is freely available from Mathworks Thus to measure each fiber type individually requires (http://www.mathworks.com/products/compiler/mcr/). multiple image segmentation masks from serial sections. SMASH App (SMASH_App.mlappinstall) requires Alternatively, using distinguishable fluorophors for each MATLAB version R2014a (8.3) or later with the Image fiber type [39] would permit the analysis of two fiber Processing Toolbox. types using the same mask using this software. Using License: CC-BY. serial sections for fiber typing is standard procedure in Any restrictions to use by non-academics: None. many labs and it is not enhanced by this software. The Abbreviations analysis of myonuclei is limited to CNFs in this software, CNF: Centrally nucleated fiber; CSA: Cross-sectional area; DAPI: 4′,6-diamidino- as opposed to providing a measure of myonuclear dens- 2-phenylindole; EDL: Extensor digitorum longus; PECAM: Platelet endothelial cell adhesion molecule; PNF: Peripherally nucleated fiber; RGB: Red, Green, and ity with peripheral nuclei [9,20]. The method of measur- Blue image format. ing CNFs is recommended by TREAT-NMD, however it causes an issue with very small fibers as the entire CSA Competing interests may be in the defined border region, making it impos- The authors declare that they have no competing interests. sible to be labeled as a CNF. The measurement of non- Authors’ contributions fiber objects designed for capillary density is currently LRS developed the code, performed histology, conducted analysis, and preliminary. There is no filtering of objects or a method drafted the manuscript. ERB contributed to the experimental design and revision of the manuscript. Both authors read and approved the final to select an object of interest. However, the output of manuscript. object size does allow filtering based on the objects CSA within Excel. This program currently is designed for use Acknowledgements We would like to acknowledge our funding sources including support from with muscle cross-sections and not designed to analyze the Muscle Core Facilities and Training from the Wellstone Muscle images from longitudinal muscle sections. As an open Physiology Core (AR052646) and support from the Pennsylvania Muscle source program, users may address any of these limita- Institute Training Fellowship (AR053461). We would like to acknowledge members of the Barton Lab for beta testing the software. We would also like tions as they see fit within the framework of the software to acknowledge Dr. Gretchen Meyer for providing technical expertise in platform and MATLAB. MATLAB programming. Received: 24 August 2014 Accepted: 15 October 2014 Conclusions Published: 27 November 2014 The software package based in MATLAB provides image processing tools to analyze immunofluorescent muscle References cross-sections. The semi-automatic fiber segmentation 1. Papadopulos F, Spinelli M, Valente S, Foroni L, Orrico C, Alviano F, Pasquinelli G: Common tasks in microscopic and ultrastructural image functions provide advanced algorithms for fiber segmenta- analysis using ImageJ. Ultrastruct Pathol 2007, 31:401–407. tion as well as provide an interface for users to manually 2. Briguet A, Courdier-Fruh I, Foster M, Meier T, Magyar JP: Histological parameters correct any errors. The histological analysis includes func- for the quantitative assessment of muscular dystrophy in the mdx-mouse. Neuromuscul Disord 2004, 14:675–682. tions for fiber CSA, fiber Feret diameter, fiber typing, 3. Andersen JL, Aagaard P: Effects of strength training on muscle fiber types CNFs, and capillary density. These functions produced ex- and size; consequences for athletes training for high-intensity sport. pected results comparing wild-type and dystrophic mouse Scand J Med Sci Sports 2010, 20(Suppl 2):32–38. 4. McPherron AC, Lawler AM, Lee SJ: Regulation of skeletal muscle mass in muscle. These functions may be purposed for other ana- mice by a new TGF-beta superfamily member. Nature 1997, 387:83–90. lyses. This open source platform provides users a frame- 5. Hortobagyi T, Dempsey L, Fraser D, Zheng D, Hamilton G, Lambert J, Dohm L: work to create their own functions or modification of Changes in muscle strength, muscle fibre size and myofibrillar gene Smith and Barton Skeletal Muscle 2014, 4:21 Page 15 of 15 http://www.skeletalmusclejournal.com/content/4/1/21 expression after immobilization and retraining in humans. JPhysiol 2000, 31. Coulton GR, Morgan JE, Partridge TA, Sloper JC: The mdx mouse skeletal 524:293–304. muscle myopathy: I. A histological, morphometric and biochemical 6. Ansved T, Larsson L: Effects of denervation on enzyme-histochemical and investigation. Neuropathol Appl Neurobiol 1988, 14:53–70. morphometrical properties of the rat soleus muscle in relation to age. 32. Denies MS, Johnson J, Maliphol AB, Bruno M, Kim A, Rizvi A, Rustici K, Acta Physiol Scand 1990, 139:297–304. Medler S: Diet-induced obesity alters skeletal muscle fiber types of male 7. Taylor NA, Wilkinson JG: Exercise-induced skeletal muscle growth. but not female mice. Physiol Rep 2014, 2:e00204. Hypertrophy or hyperplasia? Sports Med 1986, 3:190–200. 33. Carnwath JW, Shotton DM: Muscular dystrophy in the mdx mouse: 8. Mula J, Lee JD, Liu F, Yang L, Peterson CA: Automated image analysis of histopathology of the soleus and extensor digitorum longus muscles. skeletal muscle fiber cross-sectional area. J Appl Physiol (1985) 2013, J Neurol Sci 1987, 80:39–54. 114:148–155. 34. Torres LF, Duchen LW: The mutant mdx: inherited myopathy in the mouse. Morphological studies of nerves, muscles and end-plates. 9. Liu F, Fry CS, Mula J, Jackson JR, Lee JD, Peterson CA, Yang L: Automated Brain 1987, 110:269–299. fiber-type-specific cross-sectional area assessment and myonuclei 35. Terada M, Kawano F, Ohira T, Nakai N, Nishimoto N, Ohira Y: Effects of counting in skeletal muscle. J Appl Physiol (1985) 2013, 115:1714–1724. mechanical over-loading on the properties of soleus muscle fibers, with 10. Schiaffino S, Reggiani C: Fiber types in mammalian skeletal muscles. or without damage, in wild type and mdx mice. PLoS One 2012, 7:e34557. Physiol Rev 2011, 91:1447–1531. 36. Karpati G, Carpenter S, Prescott S: Small-caliber skeletal muscle fibers 11. Brooke MH, Kaiser KK: Some comments on the histochemical do not suffer necrosis in mdx mouse dystrophy. Muscle Nerve 1988, characterization of muscle adenosine triphosphatase. J Histochem 11:795–803. Cytochem 1969, 17:431–432. 37. Nagaraju K, Willmann R: Developing standard procedures for murine and 12. Guth L, Samaha FJ: Qualitative differences between actomyosin ATPase canine efficacy studies of DMD therapeutics: report of two expert of slow and fast mammalian muscle. Exp Neurol 1969, 25:138–152. workshops on “Pre-clinical testing for Duchenne dystrophy”: Washington 13. Schiaffino S, Gorza L, Sartore S, Saggin L, Ausoni S, Vianello M, Gundersen K, DC, October 27th-28th 2007 and Zurich, June 30th-July 1st 2008. Lomo T: Three myosin heavy chain isoforms in type 2 skeletal muscle Neuromuscul Disord 2009, 19:502–506. fibres. J Muscle Res Cell Motil 1989, 10:197–205. 38. Faber RM, Hall JK, Chamberlain JS, Banks GB: Myofiber branching rather 14. Gauthier GF, Lowey S: Polymorphism of myosin among skeletal muscle than myofiber hyperplasia contributes to muscle hypertrophy in mdx fiber types. J Cell Biol 1977, 74:760–779. mice. Skeletal Muscle 2014, 4:10. 15. Ferrari G, Cusella-De Angelis G, Coletta M, Paolucci E, Stornaiuolo A, Cossu 39. Gregorevic P, Meznarich NA, Blankinship MJ, Crawford RW, Chamberlain JS: G, Mavilio F: Muscle regeneration by bone marrow-derived myogenic Fluorophore-labeled myosin-specific antibodies simplify muscle-fiber progenitors. Science 1998, 279:1528–1530. phenotyping. Muscle Nerve 2008, 37:104–106. 16. Narita S, Yorifuji H: Centrally nucleated fibers (CNFs) compensate the fragility of myofibers in mdx mouse. Neuroreport 1999, 10:3233–3235. doi:10.1186/2044-5040-4-21 17. Duscha BD, Kraus WE, Keteyian SJ, Sullivan MJ, Green HJ, Schachat FH, Cite this article as: Smith and Barton: SMASH – semi-automatic muscle Pippen AM, Brawner CA, Blank JM, Annex BH: Capillary density of skeletal analysis using segmentation of histology: a MATLAB application. muscle: a contributing mechanism for exercise intolerance in class Skeletal Muscle 2014 4:21. II-III chronic heart failure independent of other peripheral alterations. J Am Coll Cardiol 1999, 33:1956–1963. 18. Miyazaki D, Nakamura A, Fukushima K, Yoshida K, Takeda S, Ikeda S: Matrix metalloproteinase-2 ablation in dystrophin-deficient mdx muscles reduces angiogenesis resulting in impaired growth of regenerated muscle fibers. Hum Mol Genet 2011, 20:1787–1799. 19. Qu Z, Andersen JL, Zhou S: Visualisation of capillaries in human skeletal muscle. Histochem Cell Biol 1997, 107:169–174. 20. Pertl C, Eblenkamp M, Pertl A, Pfeifer S, Wintermantel E, Lochmuller H, Walter MC, Krause S, Thirion C: A new web-based method for automated analysis of muscle histology. BMC Musculoskelet Disord 2013, 14:26. 21. Bulfield G, Siller WG, Wight PA, Moore KJ: X chromosome-linked muscular dystrophy (mdx) in the mouse. Proc Natl Acad Sci U S A 1984, 81:1189–1192. 22. Khajehpour H, Dehnavi AM, Taghizad H, Khajehpour E, Naeemabadi M: Detection and segmentation of erythrocytes in blood smear images using a line operator and watershed algorithm. J Med Signals Sensors 2013, 3:164–171. 23. Saez A, Acha B, Montero-Sanchez A, Rivas E, Escudero LM, Serrano C: Neuromuscular disease classification system. JBiomedOpt 2013, 18:066017. 24. Liu F, Mackey AL, Srikuea R, Esser KA, Yang L: Automated image segmentation of haematoxylin and eosin stained skeletal muscle cross-sections. JMicrosc 2013, 252:275–285. 25. Doke J: flattenMaskOverlay. In MATLAB Central. 2011. Submit your next manuscript to BioMed Central 26. Legland D: Feret diameter and oriented box. In MATLAB Central. 2011. 27. Legland D, Beaugrand J: Automated clustering of lignocellulosic fibres and take full advantage of: based on morphometric features and using clustering of variables. Ind Crops Prod 2013, 45:253–261. • Convenient online submission 28. Otsu N: A threshold selection method from gray-level histograms. • Thorough peer review Syst Man Cybern IEEE Trans on 1979, 9:62–66. 29. Malm C, Nyberg P, Engstrom M, Sjodin B, Lenkei R, Ekblom B, Lundberg I: • No space constraints or color figure charges Immunological changes in human skeletal muscle and blood after • Immediate publication on acceptance eccentric exercise and multiple biopsies. J Physiol 2000, 529:243–262. • Inclusion in PubMed, CAS, Scopus and Google Scholar 30. Park S, Brisson BK, Liu M, Spinazzola JM, Barton ER: Mature IGF-I excels in promoting functional muscle recovery from disuse atrophy compared • Research which is freely available for redistribution with pro-IGF-IA. J Appl Physiol (1985) 2014, 116:797–806. Submit your manuscript at www.biomedcentral.com/submit http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Skeletal Muscle Springer Journals

SMASH – semi-automatic muscle analysis using segmentation of histology: a MATLAB application

Skeletal Muscle , Volume 4 (1) – Nov 27, 2014

Loading next page...
 
/lp/springer-journals/smash-semi-automatic-muscle-analysis-using-segmentation-of-histology-a-5DBrq0zcWQ
Publisher
Springer Journals
Copyright
Copyright © 2014 by Smith and Barton; licensee BioMed Central Ltd.
Subject
Life Sciences; Cell Biology; Developmental Biology; Biochemistry, general; Systems Biology; Biotechnology
eISSN
2044-5040
DOI
10.1186/2044-5040-4-21
pmid
25937889
Publisher site
See Article on Publisher Site

Abstract

Background: Histological assessment of skeletal muscle tissue is commonly applied to many areas of skeletal muscle physiological research. Histological parameters including fiber distribution, fiber type, centrally nucleated fibers, and capillary density are all frequently quantified measures of skeletal muscle. These parameters reflect functional properties of muscle and undergo adaptation in many muscle diseases and injuries. While standard operating procedures have been developed to guide analysis of many of these parameters, the software to freely, efficiently, and consistently analyze them is not readily available. In order to provide this service to the muscle research community we developed an open source MATLAB script to analyze immunofluorescent muscle sections incorporating user controls for muscle histological analysis. Results: The software consists of multiple functions designed to provide tools for the analysis selected. Initial segmentation and fiber filter functions segment the image and remove non-fiber elements based on user-defined parameters to create a fiber mask. Establishing parameters set by the user, the software outputs data on fiber size and type, centrally nucleated fibers, and other structures. These functions were evaluated on stained soleus muscle sections from 1-year-old wild-type and mdx mice, a model of Duchenne muscular dystrophy. In accordance with previously published data, fiber size was not different between groups, but mdx muscles had much higher fiber size variability. The mdx muscle had a significantly greater proportion of type I fibers, but type I fibers did not change in size relative to type II fibers. Centrally nucleated fibers were highly prevalent in mdx muscle and were significantly larger than peripherally nucleated fibers. Conclusions: The MATLAB code described and provided along with this manuscript is designed for image processing of skeletal muscle immunofluorescent histological sections. The program allows for semi-automated fiber detection along with user correction. The output of the code provides data in accordance with established standards of practice. The results of the program have been validated using a small set of wild-type and mdx muscle sections. This program is the first freely available and open source image processing program designed to automate analysis of skeletal muscle histological sections. Keywords: Histological muscle analysis, Standardized quantitative analysis, Image segmentation, mdx mouse Background a widely available image processing software package Skeletal muscle has a robust ability to adapt to the pat- specifically designed for muscle histological analysis. tern of use and to regenerate following injury. These are Altering muscle fiber size is one of the primary methods often quantified using histological techniques. However, in which muscle responds to external stimuli. Muscle the methods for this quantification remain disparate mass may be increased in response to resistance training among investigators and often require painstaking man- [3] or with potential pharmacological agents like myosta- ual procedures [1,2]. The goal of this work is to provide tin inhibitors [4], while muscle atrophy occurs in response to disuse [5] and injuries such as denervation [6]. These conditions primarily reflect hypertrophy or atrophy of * Correspondence: smithlu@dental.upenn.edu Department of Anatomy and Cell Biology, School of Dental Medicine, individual fibers rather than hyper- or hypoplasia [7]. University of Pennsylvania, Philadelphia, PA, USA Muscle fiber size is routinely evaluated using fixed or fro- Pennsylvania Muscle Institute, University of Pennsylvania, Philadelphia, PA, zen tissue sections. Fiber outlines are visualized using a USA © 2014 Smith and Barton; licensee BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly credited. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated. Smith and Barton Skeletal Muscle 2014, 4:21 Page 2 of 15 http://www.skeletalmusclejournal.com/content/4/1/21 variety of techniques, including hematoxylin and eosin standard operating procedures developed by TREAT- staining, laminin immunostaining, dystrophin immuno- NMD when available using sophisticated boundary de- staining, and wheat germ aggluttinin staining [8]. While tection algorithms [2]. The software also includes built-in these techniques enable visualization of fiber boundaries, image editing to manually inspect and manipulate fiber determining fiber cross-sectional area (CSA) is often still boundaries. Fully automated fiber size determination as performed by manual tracing of individual fibers. There well as fiber types and CNFs may be possible with ad- are software programs available to help automate fiber equate image acquisition [9,20]. However, these newly detection, however they are often expensive and are not designed fully automated programs are not yet available specifically designed for muscle histology [9]. [9] and/or have a significant cost [20]. Additionally, Muscle fiber type distributions are often investigated allowing the user to have manual control over some as- in muscle histology as they are known to be altered in pect of image processing allows users to maintain the response to exercise, inactivity, and aging [10]. Fiber type fidelity established by manual techniques. The open na- is primarily determined by the myosin heavy chain iso- ture of this software also allows custom usage and fur- form, which has differential contractile and ATPase ac- ther advancement of the methods. For users that do not tivity. Fiber type is often determined by ATPase staining have access to a MATLAB license or the image process- [11,12] or with immunostaining for specific myosin ing toolbox we have compiled an .exe file that runs heavy chain (MyHC) isoforms individually [13,14]. How- using the freely available MATLAB Runtime Compiler ever, methods to determine fiber type can be subjective (MCR) version 8.3 (http://www.mathworks.com/products/ and tedious when fibers are manually classified. Follow- compiler/mcr/). Automating a large portion of muscle ing fiber segmentation, computing the size distribution histology makes it feasible to analyze full muscle cross- of single fiber types is easily automated. sections, eliminating variability introduced by selecting Muscle fibers also undergo changes in morphology as only a portion of the cross section for analysis. This soft- they develop. In particular, centrally nucleated fibers are ware is validated with muscles from mdx mice, which have often used as a marker for muscle regeneration. While many alterations of muscle fiber morphology compared to fully mature fibers have peripheral nuclei, newly regener- wild-type mice [21]. The purpose of this study is to de- ated fibers have central nuclei [15]. In many muscular velop freely available automatic and standardized image dystrophies, which are characterized by continual cycles segmentation platform and validate the program using of degeneration and regeneration, the number of cen- standard muscle histological analysis. trally nucleated fibers (CNFs) is substantial while CNFs are hardly present in healthy muscle. Although nuclei Implementation are easily stained with DAPI, determination of CNFs is Mice often performed manually. Combined use of automated All animal experiments were approved by the University CNF and fiber size determination allows the size of re- of Pennsylvania Institutional Animal Care and Use Com- generating fibers to be calculated, providing a measure mittee. C57Bl/6 mice were used as wild-type controls of how efficiently regeneration is occurring after acute and mdx mice were used as a dystrophic model. Both injury [16]. animal groups were analyzed at 1 year of age. Skeletal muscle is a highly metabolically active tissue requiring large blood supply. As with fiber type shifts, Immunohistochemistry capillary density of skeletal muscle may be affected by al- Soleus muscles from both groups (n = 4 per group) were tered metabolic demand or in disease [17]. Endothelial dissected, embedded in OCT, and frozen in liquid nitrogen cells and capillaries are frequently stained in skeletal cooled isopentane. Frozen 10 um sections were cut from muscle with Von Willibrand Factor or PECAM [18,19]. muscles and mounted on slides. Sections were washed in Automated determination of capillary density in relation PBS and immunostained using antibodies to laminin to fiber size and number provides a useful parameter of (Thermo Scientific), and either myosin heavy chain I (Devel- skeletal muscle histology. opmental Studies Hybridoma Bank) or platelet endothelial All of the methods discussed above are commonly per- cell adhesion molecule (PECAM; eBiosciences) overnight formed using immunofluorescence, which provides high at 4°C. After PBS wash, fluorescent secondary antibodies contrast in stained and unstained structures. We have (Sigma) were applied for 1 h at room temperature. Nuclei developed MATLAB (MATLAB and Image Processing were labeled with DAPI incorporated into the mounting Toolbox 2014a, MathWorks) scripts bundled into a media (Vectashield). Images were acquired using a Leica MATLAB App (see Availability and Requirements) that DM RBE microscope and DFC350FX camera and Open- automate, or partially automate determination of fiber Lab software. Individual fields were stitched together to size, fiber type, centrally nucleated fibers, and capillary create a composite full view of the muscle cross-section density. These programs are created to comply with using Photoshop (Adobe). Smith and Barton Skeletal Muscle 2014, 4:21 Page 3 of 15 http://www.skeletalmusclejournal.com/content/4/1/21 Image selection function. The Initial Segmentation function uses the water- The software has built in several steps of image processing shed transformation to determine the fiber edges, as it has tools within the same script. Initially the user must select become one of the most common and standard choices for an image file (.bmp, .jpg, .png, and .tif.) to be processed. Fol- image segmentation [22,23]. However, the watershed trans- lowing selection of the appropriate file the user is provided formation often leads to over-segmentation due to local a list of the built in functions (Figure 1A). A representative minima created from noise within an image [24]. Prior to image of a soleus muscle from a 1-year-old mdx mouse is applying the watershed transformation, the image is used which has been immunostained with laminin (red) smoothed with a function that suppresses local minima and slow myosin heavy chain (green) as well as DAPI (blue) whose depth is below a given threshold. Upon selecting (Figure 1B and C). Dystrophic muscle can be more difficult this function users are presented with options for segmen- to process automatically due greater interstitial spaces and tation (Figure 2A). The pixel size measured in μm/pixel is the examples highlight some manual adjustments that may requested in all functions. The segmentation takes place be required. The software includes an Excel (Microsoft) file based on a single RGB color channel, which the user pro- containing default parameter values (Table 1), which may vides designating the color of the fiber outlines. The user be altered to the needs of the user. The details of each par- also provides an initial value for suppressing local minima ameter are discussed in the relevant section below. based on 8-bit channel images (0 to 255). Previous experi- ence shows that a segmentation value of 4 to 10 produces Initial segmentation robust fiber identification, depending on the quality of Prior to any other function being run, a mask file for staining (Figure 3). However, the segmentation value may the image must be created with the ‘Initial Segmentation’ need to be adjusted based on the image quality and Figure 1 Image selection. (A) Dialog box allowing selection of a single function to perform on the selected muscle section. (B) Example muscle section of 1-year-old soleus muscle from mdx mouse. Red is laminin stain, green is slow muscle myosin heavy chain, and blue is DAPI. (C) Enlarged portion of (B). Smith and Barton Skeletal Muscle 2014, 4:21 Page 4 of 15 http://www.skeletalmusclejournal.com/content/4/1/21 Table 1 Default parameters from Excel file convexity. Interstitial space may fit the size requirements Default parameters Default values of a fiber; however, it will not have a grossly circular shape in cross section. The eccentricity takes an ellipse with the Pixel size (μm/pixel) 0.645 same second moments as the objects and is the ratio be- Fiber outline channel (red = 1, green = 2, blue = 3) 1 tween the distance between the foci of the ellipse and the Nuclei channel (red = 1, green = 2, blue = 3) 3 major axis. Thus, circles have an eccentricity of 0 and line Fiber type channel (red = 1, green = 2, blue = 3) 2 segments have an eccentricity of 1. Further, interstitial Object channel (red = 1, green = 2, blue = 3) 2 space may have a stellate shape between fibers while main- Segmentation smoothing factor 5 taining high eccentricity. The stellate shape creates regions of concavity, while polygonal fibers have minimal concav- Nuclear smoothing factor 5 ity and high convexity. The convexity is determined by the Object smoothing factor 10 ratio of the object area and the smallest convex polygon Minimum fiber area (μm ) 100 that can contain the entire region. Thus, a completely con- Maximum fiber area (μm ) 5,000 vex image will have a convexity of 1. After the filtering pa- Maximum eccentricity 0.95 rameters have been selected, a new mask is created with Minimum convexity 0.8 only the objects passing the filter (Figure 4B) [25]. At this point the user has the option to inspect the mask and se- Nuclear distance from boarder (μm) 10 2 lect any objects that may have passed the filter but are not Minimum nuclear size (μm)5 fibers (Figure 4C). This may include areas of interstitial Fiber properties output folder C:\SMASH\Output space, vessels, or other structures the user does not want Fiber type output folder C:\SMASH\Output to include (Figure 4D, yellow arrow). The zoomed ex- Central nuclei output folder C:\ SMASH\Output ample image shows fibers that were filtered because of Objects output folder C:\ SMASH\Output high eccentricity and high convexity due to improper sep- aration of fiber from interstitial space (Figure 4D, pink exposure (Figure 3). For example, a suboptimal segmenta- arrows). Following user removal of non-fiber objects the tion can generate false objects (Figure 3A, E) or merge ob- mask file is overwritten. jects (Figure 3C, E). Once initial parameters have been selected, the initial segmentation takes place with the Fiber properties segmented fiber edges overlaid on the original image After the mask has been filtered to contain only objects (Figure 2B) [25]. After inspecting the image the user can that are fibers the properties of the fibers may be deter- decide to re-run the function with alternative parameters. mined. The user is prompted for pixel size and the Once proper segmentation has been achieved, the user has folder location that the output data should be exported the option to separate any merged fibers that may have to. The output file is in Excel (.xls) format with the same arisen from poor staining. The user implements a freehand name as the image file appended by ‘_Props’ (Figure 5A). drawing tool to separate any fibers that are merged with The Feret diameter properties are obtained from an other fibers (red dotted line) or interstitial space (yellow incorporated publically available script on MATLAB dotted line) (Figure 2C). After the user has finished separ- Central [26]. The function displays histograms for fiber ating any fibers a mask file is saved. Initial segmentation minimal Feret diameter and fiber CSA for inspection will often produce many small segments in interstitial (Figure 5B). The Feret properties are defined by the space; however, non-fiber areas will be filtered in the ‘Fiber smallest rectangle that bounds the fiber [27]. The mini- Filter’ function. All fibers in the mask should be separated mum Feret diameter has been shown to be a more ro- prior to running the fiber filter function to prevent in- bust measure of fiber size. Feret diameter is much less appropriate filtering of fibers. sensitive to oblique muscle section compared to fiber CSA [2]. For each fiber the properties exported are: cen- Fiber filter troid in x and y in pixels, the maximum Feret diameter Following the initial segmentation the fiber filter function (μm), minimum Feret diameter (μm), Feret direction (ra- removes objects based on user-defined criteria (Figure 4A). dians), and fiber area (μm ). The output Excel file is then Fibers that touch the edge of the image are removed so written with all fiber properties as well as the mean, partial fibers are not included in further calculations. To standard deviation (SD), and standard error of the mean provide the proper scale, the pixel size is entered by the (SEM) for each property (Figure 5C). user in μm/pixel. Any fibers below the minimum fiber area or above the maximum fiber area entered in μm are Fiber typing removed from the mask and from further analysis. More The fiber typing function uses the fiber mask to inspect advanced filtering is also included with eccentricity and each fiber area for determination of fiber staining. The Smith and Barton Skeletal Muscle 2014, 4:21 Page 5 of 15 http://www.skeletalmusclejournal.com/content/4/1/21 Figure 2 Initial segmentation. (A) Input box with options for fiber segmentation. (B) The example image with white lines drawn along fiber boarders based on segmentation parameters. (C) Dialog box that appears after separating a fiber. (D) Zoomed in of B. Red dotted lines shows where two fibers are separated by the user. Yellow dotted lines shows where a fiber is not separated from intestinal space. function examines the average intensity of staining within positively stained fibers. Intermediate values may represent the individual fiber region determined by the mask to cal- mixed fibers. A histogram with the CSA of positive fibers culate an average intensity. A threshold value is used to de- (red, front) and all fibers (blue, back) is presented to exam- termine if the fiber is positive or negative based on the ine any fiber type-dependent size changes. After inspecting average intensity. The user inputs the pixel size, the RGB the figure the user is able to adjust the threshold value and channel which includes fiber type staining, and the output thefigureisupdated. Oncethe user is satisfiedwiththe folder location (Figure 6A). The initial threshold value for analysis they select accept and are prompted to create an average fiber type intensity is obtained using Otsu’s Excel file with each fiber’s average intensity, fiber area, and method, which calculates the threshold to minimize intra- if it is classified as positive (1) or negative (0) (Figure 6B). class variance of the thresholded binary image [28]. The Summary data including average fiber area, average fiber program then performs an initial calculation of fiber types. intensity and percent of positive fibers are included for all, A figure is displayed with the original image and an image positive, and negative fibers. where positive fibers are white and negative fibers are grey (Figure 6D). The figure includes a histogram of staining in- Centrally nucleated fibers tensities along with the threshold. The intensity histogram The CNF function again requires prior creation of a fiber should have two peaks, with the lower and sharper peak mask to analyze the centers of fibers for nuclei. The user being negative fibers and high intensity peak representing is prompted to input pixel size as well as the RGB channel Smith and Barton Skeletal Muscle 2014, 4:21 Page 6 of 15 http://www.skeletalmusclejournal.com/content/4/1/21 Figure 3 Segmentation examples. Example muscle section of 1-year-old soleus muscle from mdx mouse. Red is laminin stain, green is slow muscle myosin heavy chain, and blue is DAPI. (A) Segmentation filter value of 2 produces over segmentation of original image. (B) Segmentation filter value of 5 produces appropriate segmentation. (C) Segmentation filter value of 12 produces under segmentation of original image. (D) Reducing the brightness of laminin staining by 50% (representing poor staining) produces under segmentation. (E) Enhancing the brightness of laminin by 400% (representing over exposure) produces over segmentation. (F) Overexposure can be compensated for by increasing the segmentation filter value to 12. Yellow arrow heads show areas of over segmentation. Pink arrow heads show areas of under segmentation. that corresponds to nuclear staining (Figure 7A). To re- inspecting the image the user may adjust the threshold strict the inspection to central nuclei the nuclear staining value as desired. Mousing over smoothed nuclear image must be greater than a user-defined distance from the may provide useful information for manual threshold- border of the fiber input in μm. The user is also allowed ing. Once a proper threshold is determined, the user is to specify the minimum area of nuclei to count so that prompted if to determine if the results should be output to random staining is minimized. To minimize the effect of Excel. If the answer is yes, a file is created that contains the streaking that can be present with poor DAPI staining the fiber area, the area of positive nuclei staining in the center image can be smoothed using the same suppression of region, and if the fiber is classified as a CNF (1) or not (0) local minima used for fiber segmentation. An automatic (Figure 7B). Summary data include the average fiber area, threshold is applied after nuclei smoothing using Otsu’s area of central nuclei, and percent of CNFs for all fibers, method. Finally, the user also specifies the output folder CNFs, and non-CNFs. If parameter selection was not suffi- for the Excel data to be written to. After input selection a cient, the function can be rerun with new parameters. figure is displayed that contains the original image and an image with the borders. The distance not inspected is in Capillary density or non-fiber objects red and the nuclei are in blue, with peripheral nuclei It is often useful to identify objects within muscle sections appearing purple (Figure 7C, bottom left). The figure also that are not within fibers. The object counter was designed contains images of only the smoothed nuclear image to identify capillary density using a PECAM staining to and an image with only CNFs. In some situations the identify endothelial cells, however other cell types could be automatic thresholding of nuclei may be unsuitable. After analyzed with the same techniques, such as macrophage Smith and Barton Skeletal Muscle 2014, 4:21 Page 7 of 15 http://www.skeletalmusclejournal.com/content/4/1/21 Figure 4 Fiber filter. (A) Input box with options for filtering fibers. (B) Image generated in which fibers passing the filter are randomly colored to more easily distinguish adjacent fibers. (C) Zoomed in of (B). Yellow arrows point to possible user selections as non-fiber objects. Pink arrows show fibers that did not pass the filter. Lower left pink arrow did not pass due to improper segmentation as depicted in Figure 2D. (D) After selection of fibers the user is asked if they would like to indeed delete them or complete the function operation. infiltration [29]. The user is prompted for pixel size and the SMASH data verification RGB channel of the objects to be identified (Figure 8A). Ini- To verify the data generated by SMASH we compared tial values for smoothing the objects, identical to that for the same images used in previously published data [30]. nuclei and fiber type, with initial threshold automatically In this case legacy methods were done using Openlab generated using Otsu’s method. The user also identifies the software (Improvision, PerkinElmer) with simple thresh- folder location of the Excel file output. After initial parame- olding of individual RGB channels for fiber size and fiber ters are selected a figure is presented with a figure contain- type. For central nucleation fibers were manually deter- ing the original image, the object channel image, smoothed mined to be CNF or PNF and a marker placed using object channel image, and thresholded object image Openlab to enable counting. (Figure 8C). After inspection, the user is then prompted to change the threshold if desired. Once the user has achieved Statistical analysis a suitable threshold the user selects whether to output an The C57 and mdx output data were compared using a Excel file with the data. Each object has its area and cen- Student’s t-test. For comparing CNFs and PNFs a paired troid listed along with summary data including the total Student’s t-test was used. All data analyses were per- number of objects in the image. formed using PRISM (Graphpad Software). Smith and Barton Skeletal Muscle 2014, 4:21 Page 8 of 15 http://www.skeletalmusclejournal.com/content/4/1/21 Figure 5 Fiber properties. (A) Input box with options for fiber properties. (B) Histograms showing minimum Feret diameter (top) and fiber CSA (bottom). (C) Truncated example of Excel output from running fiber properties function. Results multitude of previous reports demonstrate that mean indi- In order to evaluate the suitability of the software described, vidual fiber cross-sectional area in mdx muscle is not skeletal muscle morphologies from a selection of wild-type consistently altered. However, as the mdx muscle is under- C57BL6 and mdx mouse soleus muscles were analyzed going continual cycles of degeneration and regeneration the (n = 4 per group). The mdx mouse is a model of Duchenne variability in fiber CSA is much greater than wild-type muscular dystrophy (DMD) that is commonly studied. A muscle [2,31]. The results produced here clearly reflect the Smith and Barton Skeletal Muscle 2014, 4:21 Page 9 of 15 http://www.skeletalmusclejournal.com/content/4/1/21 Figure 6 Fiber type. (A) Input box with options for fiber type function. (B) Truncated example of Excel output from running fiber type function. (C) Portion of immunofluorescent image. Red is laminin stain, green is slow muscle myosin heavy chain, and blue is DAPI (top left). Image showing negative fibers in gray and positive fibers in white (top right). Histogram of average fiber intensity for fiber type stain with threshold value (bottom left). Histogram of fiber CSA with positive fiber in red and all fibers in blue (bottom right). previous literature on mean CSA and variability of CSA increase in the standard deviation of Feret diameter within (Figure 9A and B). The mean fiber CSA was not different an mdx muscle (Figure 9C and D). between the C57 and mdx muscles; however the standard Fiber type determination is commonly performed in deviation (SD) of CSA within a muscle was much greater muscle histological analysis. The muscles studied here in mdx than C57. While measurements of fiber CSA have have been stained for slow myosin heavy chain (Type I). commonly been used, the fiber CSA is highly dependent on The soleus muscle has a large portion of type I fibers of the angle of the section relative to the axis of the fiber. Use approximately 40% in wild-type mice [32]. This proportion of minimal Feret diameter provides a robust measure of is close to the value obtained with the software (Figure 9E), fiber size largely independent of the angle of the section however there was a large increase in the percentage of andisendorsedas a standard operating procedure [2]. Our slow fibers in the mdx soleus muscle. This is expected results on Feret diameter align well with that of CSA show- based on the advanced age of these mice of 1 year and that ing no change in the mean Feret diameter, but a substantial type I fibers are more resistant to damage in mdx muscle Smith and Barton Skeletal Muscle 2014, 4:21 Page 10 of 15 http://www.skeletalmusclejournal.com/content/4/1/21 Figure 7 Centrally nucleated fibers. (A) Input box with options for CNF function. (B) Truncated example of Excel output from running CNF function. (C) Portion of immunofluorescent image. Red is laminin stain, green is slow muscle myosin heavy chain, and blue is DAPI (top left). Image showing filtered nuclei (top right). Image with boarder regions in red and nuclei above threshold in blue (bottom left). Image in which only CNFs are depicted in white (bottom right). [33]. Determination of the size of various fiber types is an CNFs. The percentage of CNFs in dystrophic muscle is a automatic feature of the software. Assuming all unlabeled commonly measured histological marker of disease [34]. fibers are type II fibers the results show that C57 type I fi- The automatic detection demonstrated the expected re- bers are smaller than type II and the relationship is not sults with C57 soleus muscle showing very rare CNFs different in mdx soleus muscles (Figure 9F). while >30% of mdx soleus muscle fibers were CNFs Due to the continual cycles of degeneration and regen- [33,35] (Figure 9G). While CNF percentage is routinely eration, dystrophic muscles have a high prevalence of performed, often using manual methods, reporting on CNFs, while uninjured wild-type muscles have very few the size of the CNFs is less common. The software Smith and Barton Skeletal Muscle 2014, 4:21 Page 11 of 15 http://www.skeletalmusclejournal.com/content/4/1/21 Figure 8 Object counter. (A) Input box with options for object counter function. (B) Truncated example of Excel output from running object counter function. (C) Portion of immunofluorescent image. Red is laminin stain, green is PECAM, and blue is DAPI (top left). Image showing only the objects of interest, here PECAM from the green channel (top right). Image showing objects of interest after smoothing filter is applied (bottom left). Binary image showing discrete objects that pass threshold value (bottom right). makes it trivial to measure the size of the CNF popula- process each function in SMASH compared to legacy tion. The data collected show that within mdx muscle methods is reported in Table 3. Smash demonstrates CNFs are larger than peripherally nucleated fibers greatly reduced time of analysis for the muscles tested. (PNFs) (Figure 9H). This relationship has been previ- ously observed in the EDL muscle of mdx mice [16,36]. Discussion The difference between data output from legacy The manual analysis of skeletal muscle immunofluores- methods and SMASH for fiber size, fiber size variability, cence is often a laborious task. The Semiautomatic fiber type percentage, and CNF percentage are presented Image Processing of Skeletal Muscle Histology Software in Table 2. There was little discrepancy between values described and tested provides researchers with a valuable obtained for fiber type percentage or CNF percentage. tool for measuring multiple facets of muscle histology. However, the calculated fiber areas were consistently While image analysis software is available to conduct higher using SMASH compared to legacy methods. This many of these features, it carries a high monetary cost is due to the greater and more variable fiber border re- and is not specifically designed for skeletal muscle ana- gion using simple thresholding in legacy systems com- lysis. In contrast, this program is available on the widely pared to SMASH (Figure 10). The time required to used MATLAB software and is designed to investigate Smith and Barton Skeletal Muscle 2014, 4:21 Page 12 of 15 http://www.skeletalmusclejournal.com/content/4/1/21 Figure 9 Comparison of output data from 1-year-old C57 and mdx mouse soleus muscle. (A) Mean of cross-sectional area of fibers in C57 and mdx muscle sections. (B) Standard deviation of CSA of fibers. (C) Mean of minimum Feret diameter (MFD) of fibers. (D) Standard deviation of MFD diameter. (E) Percentage of Type I fibers in C57 and mdx muscle sections. (F) The ratio of mean CSA of type I fibers to type II fibers. (G) The percentage of fibers with centrally nucleated fiber (CNF)s. (H) The mean CSA of mdx peripherally nucleated fibers (PNF) and CNFs. *P <0.05 for mdx compared to C57, † P <0.05 for CNFs compared to PNFs within the same muscle. specific features of muscle histology, and is open to cus- with much improved reliability over standard threshold- tom modifications for advanced users. ing. However, the software maintains the manual ability All of the functions in the program rely and are based to correct image segmentation. Furthermore, if the built- on the segmentation of skeletal muscle fibers in an in manual editing features are not suitable, the user may image. This is commonly done with manual tracing, edit the created mask file with software of his choice, which is very time-consuming, or basic thresholding of such as Adobe Photoshop. After the mask has been fi- fiber outlines, which requires extensive manual correc- nalized, running the analysis functions is expedient. The tion on all but the most pristine sections. Using a com- user is required to enter parameters for analysis and bination of the image smoothing fiber and the watershed may adjust them based on the resulting image guides. transform fiber boundaries are automatically produced To analyze the entire cross-section of 1-year-old soleus muscles for fiber size, fiber type, and CNFs from the Table 2 Data comparison from legacy methods to SMASH same image took approximately 15 min, substantially Image Legacy Method SMASH Difference less than manual methods. Increased time efficiency al- Fiber type - myosin IIa (%) lows analysis of the whole muscle cross-section, avoiding the sampling issues of only using select fields of view. A59 58 -1 Using the whole muscle cross-section is required to B60 57 -3 meet TREAT-NMD standards of practice [37]. Analyzing C58 61 3 more fiber and eliminating regional differences in cross- Fiber size - fiber area (μm ) section improves the consistency of the results. Highly A 708 898 189 significant results were obtained with low variability in B 745 934 189 our investigation with only four muscles per group. The output of fiber size data in Excel format permits C 622 819 198 the use of many graphing tools to create plots of fiber Fiber size - fiber area SD (μm ) size commonly used. Ensuring the ability to measure A 303 415 112 Feret diameter ensures compliance with the latest stand- B 265 346 81 ard operating procedures for fiber size analysis [2]. C 219 339 121 Automation of fiber type data with defined boundaries Centrally nucleated fibers (%) has advantages over using thresholding to determine fiber type and size. Instead of determining the fiber size A71 66 -5 by the area above a threshold value of staining, the area B58 55 -3 is determined entirely by the fiber outline so the positive C61 60 -1 fiber size will not be affected by blotchiness in fiber type Smith and Barton Skeletal Muscle 2014, 4:21 Page 13 of 15 http://www.skeletalmusclejournal.com/content/4/1/21 Figure 10 Comparison of mask file from legacy method and SMASH. (A) Original image of a laminin stained muscle section. (B) Fibers are colored randomly from SMASH output mask with dark regions corresponding to fiber area using a simple threshold. Dark gray areas are interstitial in SMASH and fibers using legacy methods while white areas are interstitial in both masks. Figure demonstrates the larger fiber area obtained with SMASH compared to the legacy method of using a simple threshold for fiber area. staining. The fiber type function may also be used for done in SMASH provides signal that is more robust to additional signals internal to fibers such as Evan’s Blue these parameters. SMASH provides a mask more con- Dye permeability or IgG infiltration of necrotic fibers sistent with manual tracing of fibers than applying a [2,38]. Automatic determination of CNFs provides a use- threshold. While SMASH reduces the border region be- ful measure of skeletal muscle histology. Although it is tween adjacent fibers, it is also clearly capable of delin- not frequently studied, combining fiber size with CNF eating interstitial space between adjacent fibers when determination may provide a more informative marker there is an appreciable separation as evidenced by the of muscle health [16]. The computation of this add- white and grey areas in Figure 10. Thus, we attest that itional parameter is trivial using this software. Analysis SMASH generally provides a more robust and accurate of structures outside of fibers themselves in skeletal fiber size as well as requiring less manual editing than muscle is also important. The object counter function was using simple thresholds. It is also noteworthy that fiber designed to automate analysis of capillary density, as is area is generally more variable than Feret diameter as done in muscle histological analysis for capillaries per fiber for perfectly round areas it is squaring the difference and capillaries per area [19]. However, it may also be used and the more elongated the fiber the greater the propor- to analyze a multitude of other structures within a muscle, tion of border region that may influence the results. such macrophage infiltration or matrix proteins. In addition the gains in robust analysis Table 3 demon- The data generated by SMASH are validated against strates that SMASH greatly reduces the time required to legacy methods showing largely consistent results be- analyze images. Manual editing of the fiber mask is still tween methods for fiber type and CNF percentages required in the majority of muscle sections for both ini- (Table 2). However, the discrepancy in fiber area illus- tial segmentation and fiber filtering and take the major- trates the importance of using a single analytical method ity of the processing time. However, with SMASH this for a given study and highlights the variability between manual editing is reduced to just a few minutes in the different approaches. Using a simple threshold to separ- case of the soleus muscles tested. The time gains are es- ate fibers creates highly variable borders between adja- pecially significant when doing multiple analyses on the cent fibers that are dependent on parameters such as same image, as manually editing the mask is the major exposure time and focus during image acquisition. Fil- time consumer and additional functions are able to be tering the signal and using the watershed function as is processed in just seconds. Table 3 Run time (mm:ss) comparisons from legacy methods to SMASH Image Legacy method - functions SMASH - functions Diff a a FP FT CNF TOT IS FF FP FT CNF TOT TOT A 30: 10: 16:30 56:30 1:50 1:49 0:06 0:17 0:12 4:14 52:16 B 30: 10: 13:50 53:50 2:46 3:42 0:05 0:26 0:15 7:14 46:36 C 30: 10: 13:30 53:30 1:45 2:39 0:07 0:14 0:10 4:55 48:35 Times are approximate. CNF functions were performed on different sections than other functions. CNF: Centrally nucleated fibers function; FF: Fiber filter function; FP: Fiber properties function, including fiber size; FT: Fiber typing function; IS: Initial segmentation function; TOT: Total time for analysis. Smith and Barton Skeletal Muscle 2014, 4:21 Page 14 of 15 http://www.skeletalmusclejournal.com/content/4/1/21 While this software provides many advantages there previously incorporated functions. Automated functions are notable limitations. The analysis is currently limited improve the speed and consistency of skeletal muscle to immunofluorescence and is not compatible with com- histological analysis. Although it requires a MATLAB li- mon stains such as hematoxylin and eosin. The software cense, this is the only freely available software designed for does often require manual segmentation and filtration of the analysis of skeletal muscle histology. fibers based on the current algorithms. These manual adjustments allow the user more control over the ana- Availability and requirements lysis, but also increase the time required for analysis and Project name: SMASH - Semiautomatic Image Process- introduce a degree of subjectivity. Manual adjustments ing of Skeletal Muscle Histology: a MATLAB Application. are more frequently required in analysis of diseased Project homepage: http://dx.doi.org/10.6084/m9.figshare. muscle as well. While these algorithms are advanced compared to many of the techniques currently in use, Operating System: Platform Independent. extending fiber segmentation algorithms may provide Programming Language: MATLAB. more reliable boundaries. The proposed method of Other requirements: SMASH Stand Alone (SMASH_ measuring fiber type is limited to investigation of a sin- Installer.exe) requires MATLAB Compiler Runtime (MCR) gle fiber type per image, or per color in an RGB image. version R2014a (8.3) which is freely available from Mathworks Thus to measure each fiber type individually requires (http://www.mathworks.com/products/compiler/mcr/). multiple image segmentation masks from serial sections. SMASH App (SMASH_App.mlappinstall) requires Alternatively, using distinguishable fluorophors for each MATLAB version R2014a (8.3) or later with the Image fiber type [39] would permit the analysis of two fiber Processing Toolbox. types using the same mask using this software. Using License: CC-BY. serial sections for fiber typing is standard procedure in Any restrictions to use by non-academics: None. many labs and it is not enhanced by this software. The Abbreviations analysis of myonuclei is limited to CNFs in this software, CNF: Centrally nucleated fiber; CSA: Cross-sectional area; DAPI: 4′,6-diamidino- as opposed to providing a measure of myonuclear dens- 2-phenylindole; EDL: Extensor digitorum longus; PECAM: Platelet endothelial cell adhesion molecule; PNF: Peripherally nucleated fiber; RGB: Red, Green, and ity with peripheral nuclei [9,20]. The method of measur- Blue image format. ing CNFs is recommended by TREAT-NMD, however it causes an issue with very small fibers as the entire CSA Competing interests may be in the defined border region, making it impos- The authors declare that they have no competing interests. sible to be labeled as a CNF. The measurement of non- Authors’ contributions fiber objects designed for capillary density is currently LRS developed the code, performed histology, conducted analysis, and preliminary. There is no filtering of objects or a method drafted the manuscript. ERB contributed to the experimental design and revision of the manuscript. Both authors read and approved the final to select an object of interest. However, the output of manuscript. object size does allow filtering based on the objects CSA within Excel. This program currently is designed for use Acknowledgements We would like to acknowledge our funding sources including support from with muscle cross-sections and not designed to analyze the Muscle Core Facilities and Training from the Wellstone Muscle images from longitudinal muscle sections. As an open Physiology Core (AR052646) and support from the Pennsylvania Muscle source program, users may address any of these limita- Institute Training Fellowship (AR053461). We would like to acknowledge members of the Barton Lab for beta testing the software. We would also like tions as they see fit within the framework of the software to acknowledge Dr. Gretchen Meyer for providing technical expertise in platform and MATLAB. MATLAB programming. Received: 24 August 2014 Accepted: 15 October 2014 Conclusions Published: 27 November 2014 The software package based in MATLAB provides image processing tools to analyze immunofluorescent muscle References cross-sections. The semi-automatic fiber segmentation 1. Papadopulos F, Spinelli M, Valente S, Foroni L, Orrico C, Alviano F, Pasquinelli G: Common tasks in microscopic and ultrastructural image functions provide advanced algorithms for fiber segmenta- analysis using ImageJ. Ultrastruct Pathol 2007, 31:401–407. tion as well as provide an interface for users to manually 2. Briguet A, Courdier-Fruh I, Foster M, Meier T, Magyar JP: Histological parameters correct any errors. The histological analysis includes func- for the quantitative assessment of muscular dystrophy in the mdx-mouse. Neuromuscul Disord 2004, 14:675–682. tions for fiber CSA, fiber Feret diameter, fiber typing, 3. Andersen JL, Aagaard P: Effects of strength training on muscle fiber types CNFs, and capillary density. These functions produced ex- and size; consequences for athletes training for high-intensity sport. pected results comparing wild-type and dystrophic mouse Scand J Med Sci Sports 2010, 20(Suppl 2):32–38. 4. McPherron AC, Lawler AM, Lee SJ: Regulation of skeletal muscle mass in muscle. These functions may be purposed for other ana- mice by a new TGF-beta superfamily member. Nature 1997, 387:83–90. lyses. This open source platform provides users a frame- 5. Hortobagyi T, Dempsey L, Fraser D, Zheng D, Hamilton G, Lambert J, Dohm L: work to create their own functions or modification of Changes in muscle strength, muscle fibre size and myofibrillar gene Smith and Barton Skeletal Muscle 2014, 4:21 Page 15 of 15 http://www.skeletalmusclejournal.com/content/4/1/21 expression after immobilization and retraining in humans. JPhysiol 2000, 31. Coulton GR, Morgan JE, Partridge TA, Sloper JC: The mdx mouse skeletal 524:293–304. muscle myopathy: I. A histological, morphometric and biochemical 6. Ansved T, Larsson L: Effects of denervation on enzyme-histochemical and investigation. Neuropathol Appl Neurobiol 1988, 14:53–70. morphometrical properties of the rat soleus muscle in relation to age. 32. Denies MS, Johnson J, Maliphol AB, Bruno M, Kim A, Rizvi A, Rustici K, Acta Physiol Scand 1990, 139:297–304. Medler S: Diet-induced obesity alters skeletal muscle fiber types of male 7. Taylor NA, Wilkinson JG: Exercise-induced skeletal muscle growth. but not female mice. Physiol Rep 2014, 2:e00204. Hypertrophy or hyperplasia? Sports Med 1986, 3:190–200. 33. Carnwath JW, Shotton DM: Muscular dystrophy in the mdx mouse: 8. Mula J, Lee JD, Liu F, Yang L, Peterson CA: Automated image analysis of histopathology of the soleus and extensor digitorum longus muscles. skeletal muscle fiber cross-sectional area. J Appl Physiol (1985) 2013, J Neurol Sci 1987, 80:39–54. 114:148–155. 34. Torres LF, Duchen LW: The mutant mdx: inherited myopathy in the mouse. Morphological studies of nerves, muscles and end-plates. 9. Liu F, Fry CS, Mula J, Jackson JR, Lee JD, Peterson CA, Yang L: Automated Brain 1987, 110:269–299. fiber-type-specific cross-sectional area assessment and myonuclei 35. Terada M, Kawano F, Ohira T, Nakai N, Nishimoto N, Ohira Y: Effects of counting in skeletal muscle. J Appl Physiol (1985) 2013, 115:1714–1724. mechanical over-loading on the properties of soleus muscle fibers, with 10. Schiaffino S, Reggiani C: Fiber types in mammalian skeletal muscles. or without damage, in wild type and mdx mice. PLoS One 2012, 7:e34557. Physiol Rev 2011, 91:1447–1531. 36. Karpati G, Carpenter S, Prescott S: Small-caliber skeletal muscle fibers 11. Brooke MH, Kaiser KK: Some comments on the histochemical do not suffer necrosis in mdx mouse dystrophy. Muscle Nerve 1988, characterization of muscle adenosine triphosphatase. J Histochem 11:795–803. Cytochem 1969, 17:431–432. 37. Nagaraju K, Willmann R: Developing standard procedures for murine and 12. Guth L, Samaha FJ: Qualitative differences between actomyosin ATPase canine efficacy studies of DMD therapeutics: report of two expert of slow and fast mammalian muscle. Exp Neurol 1969, 25:138–152. workshops on “Pre-clinical testing for Duchenne dystrophy”: Washington 13. Schiaffino S, Gorza L, Sartore S, Saggin L, Ausoni S, Vianello M, Gundersen K, DC, October 27th-28th 2007 and Zurich, June 30th-July 1st 2008. Lomo T: Three myosin heavy chain isoforms in type 2 skeletal muscle Neuromuscul Disord 2009, 19:502–506. fibres. J Muscle Res Cell Motil 1989, 10:197–205. 38. Faber RM, Hall JK, Chamberlain JS, Banks GB: Myofiber branching rather 14. Gauthier GF, Lowey S: Polymorphism of myosin among skeletal muscle than myofiber hyperplasia contributes to muscle hypertrophy in mdx fiber types. J Cell Biol 1977, 74:760–779. mice. Skeletal Muscle 2014, 4:10. 15. Ferrari G, Cusella-De Angelis G, Coletta M, Paolucci E, Stornaiuolo A, Cossu 39. Gregorevic P, Meznarich NA, Blankinship MJ, Crawford RW, Chamberlain JS: G, Mavilio F: Muscle regeneration by bone marrow-derived myogenic Fluorophore-labeled myosin-specific antibodies simplify muscle-fiber progenitors. Science 1998, 279:1528–1530. phenotyping. Muscle Nerve 2008, 37:104–106. 16. Narita S, Yorifuji H: Centrally nucleated fibers (CNFs) compensate the fragility of myofibers in mdx mouse. Neuroreport 1999, 10:3233–3235. doi:10.1186/2044-5040-4-21 17. Duscha BD, Kraus WE, Keteyian SJ, Sullivan MJ, Green HJ, Schachat FH, Cite this article as: Smith and Barton: SMASH – semi-automatic muscle Pippen AM, Brawner CA, Blank JM, Annex BH: Capillary density of skeletal analysis using segmentation of histology: a MATLAB application. muscle: a contributing mechanism for exercise intolerance in class Skeletal Muscle 2014 4:21. II-III chronic heart failure independent of other peripheral alterations. J Am Coll Cardiol 1999, 33:1956–1963. 18. Miyazaki D, Nakamura A, Fukushima K, Yoshida K, Takeda S, Ikeda S: Matrix metalloproteinase-2 ablation in dystrophin-deficient mdx muscles reduces angiogenesis resulting in impaired growth of regenerated muscle fibers. Hum Mol Genet 2011, 20:1787–1799. 19. Qu Z, Andersen JL, Zhou S: Visualisation of capillaries in human skeletal muscle. Histochem Cell Biol 1997, 107:169–174. 20. Pertl C, Eblenkamp M, Pertl A, Pfeifer S, Wintermantel E, Lochmuller H, Walter MC, Krause S, Thirion C: A new web-based method for automated analysis of muscle histology. BMC Musculoskelet Disord 2013, 14:26. 21. Bulfield G, Siller WG, Wight PA, Moore KJ: X chromosome-linked muscular dystrophy (mdx) in the mouse. Proc Natl Acad Sci U S A 1984, 81:1189–1192. 22. Khajehpour H, Dehnavi AM, Taghizad H, Khajehpour E, Naeemabadi M: Detection and segmentation of erythrocytes in blood smear images using a line operator and watershed algorithm. J Med Signals Sensors 2013, 3:164–171. 23. Saez A, Acha B, Montero-Sanchez A, Rivas E, Escudero LM, Serrano C: Neuromuscular disease classification system. JBiomedOpt 2013, 18:066017. 24. Liu F, Mackey AL, Srikuea R, Esser KA, Yang L: Automated image segmentation of haematoxylin and eosin stained skeletal muscle cross-sections. JMicrosc 2013, 252:275–285. 25. Doke J: flattenMaskOverlay. In MATLAB Central. 2011. Submit your next manuscript to BioMed Central 26. Legland D: Feret diameter and oriented box. In MATLAB Central. 2011. 27. Legland D, Beaugrand J: Automated clustering of lignocellulosic fibres and take full advantage of: based on morphometric features and using clustering of variables. Ind Crops Prod 2013, 45:253–261. • Convenient online submission 28. Otsu N: A threshold selection method from gray-level histograms. • Thorough peer review Syst Man Cybern IEEE Trans on 1979, 9:62–66. 29. Malm C, Nyberg P, Engstrom M, Sjodin B, Lenkei R, Ekblom B, Lundberg I: • No space constraints or color figure charges Immunological changes in human skeletal muscle and blood after • Immediate publication on acceptance eccentric exercise and multiple biopsies. J Physiol 2000, 529:243–262. • Inclusion in PubMed, CAS, Scopus and Google Scholar 30. Park S, Brisson BK, Liu M, Spinazzola JM, Barton ER: Mature IGF-I excels in promoting functional muscle recovery from disuse atrophy compared • Research which is freely available for redistribution with pro-IGF-IA. J Appl Physiol (1985) 2014, 116:797–806. Submit your manuscript at www.biomedcentral.com/submit

Journal

Skeletal MuscleSpringer Journals

Published: Nov 27, 2014

References