Reduced satellite cell population may lead to contractures in children with cerebral palsy

Abstract

AIM

Satellite cells are the stem cells residing in muscle responsible for skeletal muscle growth and repair. Skeletal muscle in cerebral palsy (CP) has impaired longitudinal growth that results in muscle contractures. We hypothesized that the satellite cell population would be reduced in contractured muscle.

METHOD

We compared the satellite cell populations in hamstring muscles from participants with CP contracture (n=8; six males, two females; age range 6–15y; Gross Motor Function Classification System [GMFCS] levels II–V; 4 with hemiplegia, 4 with diplegia) and from typically developing participants (n=8; six males, two females, age range 15–18y). Muscle biopsies were extracted from the gracilis and semitendinosus muscles and mononuclear cells were isolated. Cell surface markers were stained with fluorescently conjugated antibodies to label satellite cells (neural cell adhesion molecule) and inflammatory and endothelial cells (CD34 and CD4 respectively). Cells were analyzed using flow cytometry to determine cell populations.

RESULTS

After gating for intact cells a mean of 12.8% (SD 2.8%) were determined to be satellite cells in typically developing children, but only 5.3% (SD 2.3%;p<0.05) in children with CP. Hematopoietic and endothelial cell types were equivalent in typically developing children and children with CP (p>0.05) suggesting the isolation procedure was valid.

INTERPRETATION

A reduced satellite cell population may account for the decreased longitudinal growth of muscles in CP that develop into fixed contractures or the decreased ability to strengthen muscle in CP. This suggests a unique musculoskeletal disease mechanism and provides a potential therapeutic target for debilitating muscle contractures.

Cerebral Palsy (CP) is a group of diseases that affect the upper motor neurons in the developing brain.¹ CP is the most common movement disorder in children, with a prevalence of 3.6 cases per 1000 births in the USA.² Spastic CP is the most common form, and describes a velocity-dependent resistance to muscle stretch. Although the primary insult to upper motor neurons in CP is not progressive, the secondary dysfunction of the musculo-skeletal system often progresses.³ Patients with spastic CP often develop debilitating fixed muscle contractures that restrict the normal range of motion around a joint, limit mobility, and may be painful. For the purpose of this study, muscle contracture is defined as an adaptation of muscle that limits the range of motion around a joint by increasing passive stiffness, excluding any active force production component of the muscle. Although contractures account for the major disability of patients with CP, the mechanism of contracture formation is not well understood.

Contractured muscle is often considered “shortened” as it has a reduced functional range. The functional range is known to be related to the normalized fiber lengths within the muscle.⁴ Previous studies have shown increased sarcomere lengths in CP,^5,6 which would predict fewer sarcomeres in series and thus a shorter, normalized fiber length. It has been shown in cat soleus muscle that, as bony growth occurs during development, normal muscle has the ability to add sarcomeres in series to maintain sarcomere length in the desired range.⁷ The extent to which human muscle changes its serial sarcomere number during development is not known, but it has been shown that human muscle can add serial sarcomeres during distraction osteogenesis.⁸ However, this process may be interrupted by the multifactorial adaptations present in muscle contracture.⁹ Transcriptionally, pathways are altered for muscle growth, metabolism, and extracellular matrix production.¹⁰ Contractured muscle has been shown to be stiffer, regardless of length, as a result of both stiffer fibers¹¹ and stiffening of the extracellular matrix.⁵ This stiffening has also recently been demonstrated at the whole-muscle level in the medial gastrocnemius, which bears more passive tension and less active force around the range of motion of the ankle.¹² Increased extracellular matrix content and muscle fibrosis is a consistent finding within contractured muscle.^5,13,14 This leads to higher stresses and limited strain in muscle, but also suggests that there may be a biological or structural barrier to growth in contractured muscle.

In patients with CP, the cross-sectional area, the major determinant of force production in skeletal muscle, is typically reduced in contractured muscle. This reduction has been demonstrated in numerous studies using a variety of methods.¹⁵ It results in muscle weakness, which is exacerbated by reduced neuronal drive to the muscle,¹⁶ and is another major inhibitor to muscle growth. These data suggest that muscle growth is inhibited both in parallel with and longitudinal to the muscle fibers by the pathologic environment in CP.

The cells responsible for muscle growth are the satellite cells. Satellite cells represent a stem cell population that resides between the sarcolemma of muscle fibers and the surrounding basal lamina. In response to growth stimuli, satellite cells proliferate to maintain the pool; they migrate to the site of damage and fuse to existing fibers or other satellite cells to form multinucleated myotubes and, eventually, mature fibers.¹⁷ The impact of the milieu of altered stimulus, including spasticity, on satellite cells in patients with CP is unknown. Fibrotic muscle tissue has been shown to limit the ability of satellite cells to participate in the generation of new muscle tissue. The classical marker for satellite cells is Pax7, a transcription factor that stops expression after differentiation.¹⁸ The neural cell adhesion molecule (NCAM) has also been used as a satellite cell marker in previous work.^19,20 Other cell populations are also important to muscle regeneration and function. Chronic infiltration of immune cells has been shown to initiate the fibrotic tissue response. Understanding these various stem cell populations in human muscle and how they are altered in response to diseases such as CP is fundamental to understanding the mechanism of this and perhaps other muscle disorders.

Flow cytometry is routinely used in stem cell research to identify distinct populations from a cellular milieu based on multiple antigen labeling. The technique is most often commonly used for cellular fluids, but is also amenable to cell suspensions from soft tissue. Flow cytometry has recently been used to investigate mononuclear cell types present in muscle.^21–23 Although this technique does not allow the counting of muscle fibers themselves, owing to their size, it provides a highly quantitative, representative, and high-throughput method to compare mononuclear cell distributions. Flow cytometry studies of human muscle are rare, but at least one study showed reliable quantification of satellite cell number compared with the traditional immunohistochemical technique.¹⁹ This work showed that satellite cell number increased in response to muscle exercise. The current study was designed to quantify the cell populations of contractured muscle in CP, to provide evidence as to how they may create or contribute to the fibrotic muscle tissue observed. This work was aimed at providing insight into how cell types respond to diseased conditions and could provide targets for needed therapies to treat muscle contracture.

METHOD

Muscle biopsy collection

This study met the ethical standard of the Declaration of Helsinki and was approved by the institutional review board of the University of California, San Diego, Human Research Protection Program. Participants were recruited into this study after obtaining age-appropriate assent from children and consent from parents. Children with CP (n=8; six males, two females; age range 6–15y; Gross Motor Function Classification System [GMFCS] levels II–V; 4 with hemiplegia, 4 with diplegia) underwent hamstring lengthening surgery involving the gracilis and semitendinosus muscles whereas typically developing children (n=8; six males, two females; age range 15–18y) were recruited as a comparison group because they were undergoing anterior cruciate ligament reconstructive surgery using a hamstring autograft and had no history of neurological disorders. Surgery for typically developing patients took place approximately 4 months after the initial injury of their anterior cruciate ligament. At this time point, they were able to walk freely and continue daily living with the exception of participating in cutting sports. This made these patients well suited as ‘controls’ in our study, since their muscle was about as normal as a child's muscle can be while still being accessible through surgery. The autograft was made using gracilis and semitendinosus tendons, which were excised along with a portion of distal muscle that could be obtained as it was trimmed from the tendon. Biopsies from the children with CP and typically developing children were obtained from the distal portion of the muscle near the musculotendinous junction. (We thus have no information on any differences in muscle properties that may occur across the entire muscle.) All patients with CP had developed a contracture requiring surgery, despite receiving conservative treatment. Patients were classified based on clinical measures of the GMFCS²⁴ and popliteal angle (Table I). Patients had not had previous hamstring lengthening or undergone botulinum toxin injections within the previous 6 months.

Table I.

Characteristics of study participants

Participant no.	Group	Sex	Age (y)	GMFCS level	Popliteal angle (deg)
81	CP	M	13	II	135
82	CP	M	15	II	120
83	CP	F	10	III	120
85	CP	M	10	V	100
91	CP	M	6	II	100
94	CP	M	10	IV	90
95	CP	M	8	III	90
92	CP	F	14	III	95
84	TD	M	18	NA	NA
88	TD	M	17	NA	NA
89	TD	F	18	NA	NA
90	TD	M	16	NA	NA
86	TD	M	17	NA	NA
87	TD	M	18	NA	NA
93	TD	F	15	NA	NA
96	TD	M	18	NA	NA

GMFCS, Gross Motor Function Classification System; CP, cerebral palsy; M, male; F, female; TD, typically developing; NA, not applicable.

Flow cytometry

Biopsies from gracilis and semitendinosus muscles were combined and prepared for flow cytometry analysis adapted from standard methods and criteria.²⁵ Briefly, muscles were immediately placed in a 15mL conical vial containing 6mL of digestion buffer (Dulbecco's modified Eagle medium [DMEM; Gibco/Invitrogen, Carlsbad, CA, USA]; 20mg/mL collagenase type 1 [Sigma, St. Louis, MO, USA]; 6mg/mL dispase II [Roche, Mannheim, Germany]; 6ku/mL penicillin [Gibco/Invitrogen]; 5mg/mL streptomycin [Gibco/Invitrogen]) and transported on ice for approximately 30 minutes. After transfer, tissue was placed in an incubator for 30 minutes (5% CO₂, 37 °C). Samples were transferred to 3.5cm tissue culture plates and mulched using surgical scissors in a sterile tissue culture hood. After the tissue was adequately mulched, it was returned to the incubator for an additional 50 minutes. The plate and the mixture were then manually triturated through a 5mL plastic serological pipette, to further disrupt the tissue, and incubated for an additional 5 minutes. Following incubation, the mixture was filtered through a 70lm mesh cell strainer (BD Falcon, San Jose, CA, USA) into a 50mL conical-bottom tube. The plate was subsequently washed with 15mL of DMEM and 10mL of fetal bovine serum (FBS; Gibco/Invitrogen) and passed through the filter. The final volume was brought to 40mL using DMEM. The mixture was then centrifuged at 600g for 10 minutes at 4 °C to obtain a pellet containing mononuclear cells. The pellet was resuspended in 1mL of fluorescence-activated cell sorting (FACS) buffer (2.5% goat serum; 1mmol/l ethylenediaminetetraacetic acid in phosphate-buffered saline [PBS] at pH 7.4) and a 15lL sample was taken for analysis of total mononuclear cell concentration and 60lL samples were taken for fluorescence minus one (FMO) controls. Cell-surface markers were then labeled by adding the following fluorophore conjugated antibodies to the cell suspension: NCAM (R-phycoerythrin [PE]; Santa Cruz Biotechnology, Santa Cruz, CA, USA, 1:50), CD34 (PE-Cy7; eBiosciences, San Diego, CA, USA, 1:200), CD45 (eFluor 450; eBiosciences, 1:100), ER-TR7 (PerCP; Santa Cruz Biotechnology, 1:50), and platelet-derived growth factor receptor a (APC; R&D Systems, Minneapolis, MN, USA, 1:100). After 20 minutes of incubation on ice, cells were pelleted and washed in 3mL of FACS buffer. Cells were then pelleted and resuspended in 1mL of FACS buffer and the cell suspension was added dropwise to dry ice-cooled 70% ethanol under gentle agitation for fixation and stored at )20 °C. Before analysis, fixed cells were pelleted for 5 minutes and resus-pended in 1mL of blocking solution (2% bovine serum albumin [BSA], 5% FBS, 0.2% Triton X-100, 0.1% sodium azide, in PBS). Cells undergoing intracellular Pax7 labeling were then pelleted and resuspended in 1mL of FACS buffer containing Pax7 antibody (rabbit immunoglobulin G; Abcam, Cambridge, MA, USA). Cells were then washed in 10mL of FACS buffer and then incubated in secondary antibody (Texas Red, anti-rabbit immunoglobulin G [Abcam]) and incubated on ice for 20 minutes. Finally, cells were washed in 3mL of FACS buffer and resuspended in 1mL of FACS buffer for analysis.

Flow cytometry was conducted using an LSR Fortessa (BD Biosciences, San Jose, CA, USA) instrument at the Sanford Burnham Medical Research Institute Flow Cytometry Core (La Jolla, CA, USA; http://www.sanfordburnham.org/Pages/Splash.aspx). Optical alignment and fluidics of the cytometer were verified daily by a trained technician using BD Cytometer Setup and Tracking Software (BD Biosciences). The excitation and emission wavelengths used were NCAM (PE) excitation=532nm, emission=478nm; Pax7 (Texas Red) excitation=565nm, emission=613nm; CD45 (eFluor 450) excitation=405nm, emission=455nm; CD34 (PE-Cy7) excitation=743nm, emission=767nm; ER-TR7 (PerCP) excitation=490nm, emission=675nm; platelet-derived growth factor receptor a (APC) excitation=650nm, emission=660 nm.

Gating and analysis

Because the human cell sorting gates have not been unambiguously defined, a complete compensation matrix was created using rat immunoglobulin G compensation beads (BD Biosciences) labeled with a single fluorophore. Gating strategies were optimized using multiple experiments that included various unstained and FMO controls. Initial gating was set based on a two-dimensional plot of forward and side scatter to target intact cells while limiting cellular debris, which is often obtained when isolating cells from solid tissue (Fig. 1a). Satellite cell gating was performed with a one-dimensional gate placed such that fewer than 1% of the cells in the FMO were positive (Fig. 1b,c).²⁶ Gating for satellite cells was done initially, as they may also be CD34+.²⁷ Gating for endothelial cells and inflammatory cells was performed on a two-dimensional plot of CD34 and CD45, with CD34+ and CD45–cells designated as endothelial and CD45+ and CD34–designated as inflammatory (Fig. 1d,e).^27,28 Attempts were made to measure fibroblasts and fibro/adipogenic progenitors using ER-TR7 and platelet-derived growth factor receptor a respectively, but no samples produced positive signal, suggesting poor binding of these antibodies to human muscle. All samples were run in the same session as a full set of controls, including FMOs and compensation beads. Significant differences in population size between groups were determined using a Student's t-test with significance set at <0.05 and data are reported as mean and the standard error of the mean.

Figure 1.

Gating protocol used to define mononuclear cell populations in human muscle. (a) Sample of isolated muscle mononuclear cells plotted with forward and side scatter. The enclosed region shows the events that passed through the cell gate. (b) Histogram of cells against the satellite cell marker neural cell adhesion molecule (NCAM; R-phycoerythrin) in which all antibodies were added except for anti-NCAM The gate is drawn at the lowest point where <1% of cells are positive. (c) Application of the fluorescence minus one gate from (b) to the fully labeled cells, showing the percentage of satellite cells present in the cell population. (d) Two-dimensional plot of CD45 (e450) and CD34 (PE-Cy7) showing sample placement of gates for CD45+/CD34–inflammatory cells and CD45–)/CD34+ endothelial cells. FSC-A, forward scatter area of measurement; PE-A, Phycoerthrin- area of measurement.

RESULTS

The isolated cell suspension from muscle biopsies contained a variety of cells and debris. The initial gate was set to include predominantly whole mononuclear cells and included 41.8% (SD 6.0%) of gating events from control tissue and 39.0% (SD 5.4%) of events from CP muscle. Cells were then analyzed for fluorescent antibody labeling to determine their classification.

Pax7 is the standard for labeling of satellite cells from muscle tissue.¹⁸ However, NCAM has also been used for satellite cell labeling in muscle tissue showing a high level of concordance with Pax7. In our flow cytometry experiments, use of a cell-surface marker (NCAM) simplified the protocol compared with using an intracellular protein (Pax7), negating the requirement for cell permeabilization and also permitting future culturing of these cells.¹⁹ To confirm that Pax7 and NCAM co-labeled the same cells, cells were analyzed for both labels. The results produced a high degree of correlation between Pax7 (Texas Red) and NCAM (PE) signal (Fig. 2). This allowed us to use NCAM as the satellite cell marker in subsequent experiments.

Figure 2.

Demonstration of the equivalence of Pax7 (Texas Red) and neural cell adhesion molecule (NCAM, R-phycoerythrin) labeling of presumed satellite cells. Intracellular labeling of Pax7, the traditional satellite cell marker is shown to be equivalent to NCAM labeling, a replacement extra-cellular satellite cell marker. Using the NCAM surface marker, it is not necessary to first permeabilize the cells.

Satellite cells were our primary cell of interest and we used the FMO gating strategy to identify satellite cells definitively (Fig. 1b,c). In control muscle tissue, 12.8% (SD 2.8%) of the cells were determined to be satellite cells, compared with a significantly smaller proportion of satellite cells (5.3%; SD 2.3%) in CP muscle (p<0.05; Fig. 3). Populations of cells determined to be inflammatory cells (CD45+/CD34)) and endothelial cells (CD45)/CD34+) were also examined. There was no significant difference between the inflammatory cell population in comparison tissue (7.5%; SD 1.6%) and CP tissue (6.7%; SD 1.0%). A similar result was seen in the mononu-clear cell population of endothelial cells, of which 17.8% (SD 1.6%) were positive in control tissue compared with 17.1% (SD 2.4%) in CP muscle tissue. Based on equivalent levels of inflammatory cells and endothelial cells in both experimental populations, it is likely that mononuclear cells are isolated to equal extents in both groups.

Figure 3.

Cell population quantification determined from the gating protocol established in Figure. 1 Satellite cells are significantly reduced in patients with cerebral palsy (CP) compared with typically developing participants TD; *p<0.05). Inflammatory and endothelial cell populations are not different between groups.

There was also a substantial degree of between-participant variability in muscle cell populations. Although the percentage of satellite cells was significantly different in typically developing children and children with CP, there was some overlap with typically developing populations. In the case of CP populations, data were clustered below 5% of cells, with the exception of one obvious outlier. This could represent a patient who had a unique mechanism of contracture. Endothelial cells were generally present in larger abundance and had less variability with a covariance of 33%. Fewer inflammatory cells were present than other cell types and the covariance was 52%.

DISCUSSION

Mononuclear cells, beyond the adult multinucleated muscle fibers that make up the majority of muscle tissue, play an important role in muscle function and adaptation to disease or injury. Here, we investigated populations of satellite cells, inflammatory cells, and endothelial cells and how they are altered in response to CP in spastic muscle contractures. This was conducted using biopsies from patients undergoing surgery for muscle lengthening in the case of CP or anterior cruciate ligament reconstruction for typically developing patients. The use of flow cytometry in human skeletal muscle is an emerging technique that permits quantitative investigation of cell populations; however, the limited use of the technique in humans makes comparison with other studies difficult. Importantly, we found that significantly fewer satellite cells were present in CP muscle than in muscle from typically developing individuals. As a control for our isolation methods and to rule out the role of non-muscle tissue, similar levels of inflammatory and endothelial cells were found in CP muscle and comparison samples.

Based on a previous transcriptional profiling study, we suggested that muscle in CP has a reduced growth potential, leading to overstretching of muscle, increased sarcomere length and tissue damage.⁵ This reduced growth potential could be the result of fewer satellite cells being available to fuse with myofibers and maintain or increase muscle mass. Our demonstration of decreased satellite cell number in contractured muscle is consistent with this idea. The consequence would be impaired longitudinal growth and muscle excursion as well as smaller cross-sectional area and muscle force production. Large muscle strains also lead to muscle damage, which has been shown to induce satellite cell proliferation in order to repair the damage.¹⁹ However, this was not observed in our tissue samples. While this study lacks the sample size necessary to determine whether differences in satellite cell populations correlate with function in the CP patients, it is important to note that this is a non-homogeneous group in terms of motor function. Further studies are required to elucidate how mobility and function may play a role in altering the satellite cell population and vice versa. It should be noted that all patients who participated in this study were undergoing surgery to correct a muscle contracture that was inhibiting range of motion. Studies conducted earlier in the disease process, prior to the formation of static contractures, may demonstrate increased satellite cell number as seen with injury, but the prolific response of satellite cells may lead to exhaustion of the supply and thus the clinically significant muscle contracture that results. Longitudinal studies that would best address this question are very difficult to perform in a human patient population, and no animal model has been shown to develop as a model of CP.²⁹ In the present study we were unable to distinguish whether decreased satellite cell population contributes to muscle contracture or if it is a secondary response to the muscle contracture itself.

Satellite cells have also been shown to alter their cell fate from myoblasts to myofibroblasts.³⁰ Depletion of satellite cells along this lineage is consistent with the increased fibrosis observed in CP muscle.^5,13,14 The clinical significance of decreased satellite cell population in CP would be supported by a positive correlation with clinical parameters such as range of motion or functional scores; however, this study was under-powered to reveal any such relationship. Although this study was conducted exclusively in pediatric patients, the comparison patients were older (mean 17y 1mo, SD 1y 1mo) compared with the patients with CP (mean 10y 10mo, SD 3y 1mo). The difference in age between groups is a confounding factor within this study as it was impossible to age match the two groups. However, importantly, the lower population of satellite cells in the younger population of children with CP is actually an underestimate of the difference between groups as younger children would be expected to have an increased satellite cell population based solely on age.³¹ Our results thus suggest a mechanism whereby satellite cell number limits muscle growth and repair in patients with CP.

Mononuclear cells other than satellite cells may also play a role in the adaptation that takes place in CP. During muscle damage there is infiltration of inflammatory cells, which assist in directing the repair and regeneration of muscle tissue. However, a persistent presence of inflammatory cells in muscle is often associated with a fibrotic response in animal models.³² Since CP muscle has been shown to be fibrotic,^5,13,14 we hypothesized that, owing to muscle damage, CP biopsies would contain a higher population of inflammatory cells. However, there was no significant increase in inflammatory cells present in the biopsies tested. There are, however, many types of inflammatory cells that coordinate the response in muscle, which this study does not capture.

Skeletal muscle tissue is very energetically active and capable of adapting to provide increased muscle blood perfusion. In adapting to CP, the change in muscle fibers to or from fast glycolytic fibers that require less perfusion and slow oxidative fibers that require an extensive supply of oxygen from the blood has mixed results in the literature.³³ The number of endothelial cells would be expected to be increased in muscles that have a higher proportion of slow fibers. This study, however, showed no change in the endothelial cell population which suggests that there was little or no fiber type shift. Similar to inflamma-tory cells, though, the classification of endothelial cells was not ideal. CD34 was used to eliminate many hematopoietic progenitors from the inflammatory pool, but a more specific marker, such as CD31, could be used to target endothelial cells.³³

The present study looked at select cell populations in debilitating muscle contractures in patients with CP. While inflammatory and endothelial cell populations did not change in spastic muscle contractures, satellite cell number was decreased. These essential cells for muscle growth and repair suggest that the ability of muscle in CP to adapt is impaired in contracture and could directly lead to contracture. While further studies are required to elucidate the mechanism for this decrease, these findings offer a new way of understanding the debilitating contractures experienced by patients with CP. They also point to further studies designed to enhance the satellite cell pool to reverse this decrease through traditional techniques or cell transplantation.

What this paper adds

• Muscle tissue in hamstring contractures has fewer satellite cells than typically developing muscle.

• Endothelial and inflammatory cell populations are not altered in muscle contracture.

ABBREVIATIONS

FACS

Fluorescence-activated cell sorting

FMO

Fluorescence minus one

NCAM

Neural cell adhesion molecule