Temporal disruption of neuromuscular communication and muscle atrophy following noninvasive ACL injury in rats

Abstract

Many patients with anterior cruciate ligament (ACL) injuries have persistent quadriceps muscle atrophy, even after considerable time in rehabilitation. Understanding the factors that regulate muscle mass, and the time course of atrophic events, is important for identifying therapeutic interventions. With a noninvasive animal model of ACL injury, a longitudinal study was performed to elucidate key parameters underlying quadriceps muscle atrophy. Male Long-Evans rats were euthanized at 6, 12, 24, or 48 h or 1, 2, or 4 wk after ACL injury that was induced via tibial compression overload; controls were not injured. Vastus lateralis muscle size was determined by wet weight and fiber cross-sectional area (CSA). Evidence of disrupted neuromuscular communication was assessed via the expression of neural cell adhesion molecule (NCAM) and genes associated with denervation and neuromuscular junction instability. Abundance of muscle RING-finger protein-1 (MuRF-1), muscle atrophy F-box (MAFbx), and 45 s pre-rRNA along with 20S proteasome activity were determined to investigate mechanisms related to muscle atrophy. Finally, muscle damage-related parameters were assessed by measuring IgG permeability, centronucleation, CD68 mRNA, and satellite cell abundance. When compared with controls, we observed a greater percentage of NCAM-positive fibers at 6 h postinjury, followed by higher MAFbx abundance 48 h postinjury, and higher 20S proteasome activity at 1 wk postinjury. A loss of muscle wet weight, smaller fiber CSA, and the elevated expression of run-related transcription factor 1 (Runx1) were also observed at the 1 wk postinjury timepoint relative to controls. There also were no differences observed in any damage markers. These results indicate that alterations in neuromuscular communication precede the upregulation of atrophic factors that regulate quadriceps muscle mass early after noninvasive ACL injury.

NEW & NOTEWORTHY A novel preclinical model of ACL injury was used to establish that acute disruptions in neuromuscular communication precede atrophic events. These data help to establish the time course of muscle atrophy after ACL injury, suggesting that clinical care may benefit from the application of acute neurogenic interventions and early gait reloading strategies.

INTRODUCTION

Anterior cruciate ligament (ACL) injury is a common orthopedic injury that occurs annually in over 200,000 individuals in the United States (1). Despite significant time spent in formalized rehabilitation, many patients with ACL tears have persistent quadriceps weakness and atrophy (2, 3). This protracted weakness is not trivial, as it has widespread negative implications for the patient including elevated reinjury risk (4), depressed quality of life (3, 5, 6), early-onset osteoarthritis (7, 8), and lifelong reduced physical activity levels (9). The majority of ACL research has used indirect electrophysiological techniques to examine the neurological pathways of quadriceps muscle weakness (e.g., see large systematic reviews in Refs. 10–12), but few have directly assessed the cellular mechanisms of action that regulate muscle mass (13–15). A significant number of patients with ACL injuries have suboptimal patient-related outcomes and severe muscle weakness (3, 5, 6), and it is, therefore, readily apparent that the pathogenic mechanisms involved in the development of quadriceps muscle atrophy after ACL injury require further study to develop more effective interventions.

Cross-sectional studies have shown that patients with a history of ACL injury demonstrate greater quadriceps muscle collagen content and fibrosis (13, 15), as well as increased abundance of inflammatory markers like tumor necrosis factor-α (TNFα), both in the muscle and circulating serum (16). Alterations in quadriceps muscle volume and cross-sectional area (CSA) have also been well documented after ACL injury (14, 17–20). However, the gap between research and clinical solutions remains, as longitudinal work that studies atrophic mechanisms of action to guide effective interventions is lacking. This gap remains, in part, because of the invasive nature of tissue biopsies that often preclude routine use (13, 15) and the utilization of preclinical models that surgically transect the ACL, confounding the native injury response (21–23).

To advance our basic understanding of the pathogenic mechanisms involved in the development of quadriceps muscle atrophy after ACL injury, we created a preclinical closed-ACL injury model (25) to explore a number of key processes and the time course of events related to quadriceps muscle mass. To do this, vastus lateralis (VL) muscle was systematically characterized using widely accepted markers of muscle atrophy including decreased muscle wet weight and reduced mean muscle fiber CSA. The response of markers of muscle atrophy were also characterized [muscle RING-finger protein-1 (MuRF-1), muscle atrophy F-box (MAFbx)] and coupled with proteolytic activity associated with the ubiquitin proteasome pathway (20S proteasome). Markers associated with ribosome biogenesis (45 s pre-rRNA, total RNA), muscle damage [immunoglobulin G (IgG) infiltration, satellite cell number (Pax7⁺ cells), centralized nuclei] and inflammation [macrophage abundance (CD68 mRNA), tumor necrosis factor-α (TNFα), interleukin-1β (IL-1β), interleukin-6 (IL6)] were also investigated. Finally, because of the strong link between ACL injury and acute neurological inhibition of the muscle (10–12), evidence of impaired neuromuscular communication was assessed via the expression of neural cell adhesion molecule (NCAM) (26) and the targeted analysis of genes associated with neuromuscular junction instability [Arabidopsis thaliana genome (ACR2), growth arrest and DNA damage inducible-α (Gadd45a), run-related transcription factor 1 (Runx1)] (27).

METHODS

Study Design and Animal Experimentation

This study was approved by the University of Connecticut Institutional Animal Care and Use Committee and followed all National Institute of Health guidelines for the ethical treatment of animals (IACUC approval number A17-042, approved 01/08/2018). Fifty-six adult male Long Evans rats (aged 16 wk at time of injury) were used for this study. Six rats served as controls. Animals were kept in standard sized cages on a 12-h:12-h light/dark cycle with access to food and water ad libitum and underwent a 1-wk acclimation period before injury. Rats were euthanized at one of seven timepoints after injury: 6, 12, 24, or 48 h (n = 6 rats) or 1, 2, or 4 wk (n = 8 rats).

Closed-ACL Injury and Tissue Collection

Closed-ACL injury was induced by a single overload of tibial compression that has previously been described (25). Briefly, the custom device consists of two custom-built loading platforms. The top knee stage is rigidly mounted to a linear actuator (DC linear actuator L16-63-12-P, Phidgets, Alberta, CA) that positions the right hind limb in 30° of dorsiflexion and 100° of knee flexion. The bottom stage holds the flexed knee and is mounted directly above a load cell (HDM Inc., PW6D, Southfield, MI). For ACL injury, rats were anesthetized (5% induction, 2% maintenance isoflurane/500 mL O₂ via a nose cone) and then the right hind limb was subjected to single load of tibial compression at a speed of 8 mm/s. ACL injury was noted by a release of tensile force during injury that was monitored via a custom program (LabVIEW, National Instruments, Austin, TX). Postinjury, study personnel performed the Lachman test to clinically confirm an ACL rupture had occurred and checked for the presence of gross bone deformity. If no contraindications were present, rats were returned to their cages and allowed to recover. Notably, in accordance with IACUC approval, no opioids or analgesics were administered because of the ability for these drugs to alter the natural biological response. Rats were euthanized (via CO₂ asphyxiation) at indicated times, and the VL muscles were harvested, weighed, flash-frozen, and stored at −80°C for further analysis. The ACL-injured/right hind limb was used for all analyses, except for CSA where both hind limbs were used. Each ACL-injured knee was also dissected to confirm isolated ACL injury and to note the presence of any potential fractures, ruptures, and avulsions of other tissues. The VL muscle was selected for analysis because of its relatively large strength contribution compared with other muscles of the knee extensor mechanism (28). The VL was also selected as this quadriceps muscle is most often characterized in human ACL studies (14, 29).

Immunohistochemistry

Fiber cross-sectional area/central nuclei.

VL muscle was sectioned (8 µm), mounted on a slide glass, and then immunoreacted with dystrophin antibody (dilution 1:200, Cat. No. PA5-32388, Thermo Fischer Scientific, Waltham, MA) for 1 h at room temperature, followed by goat-anti mouse 488-conjugated secondary antibody (dilution 1:250, Cat. No. D-20691, Invitrogen, Carlsbad, CA) overnight at 4°C. Sections were postfixed in 4% paraformaldehyde and stained with 4′,6-diamidino-2-phenylindole (dilution, 1:10,000, DAPI, Cat. No. D9542, Sigma-Aldrich, St. Louis, MO) to visualize the nuclei. Slides were then coverslipped with Vectashield mounting medium (Cat. No. H-1000, Vector Laboratories, Burlingame, CA) and images were acquired using a Zeiss AxioImager M1 upright fluorescent microscope (Carl Zeiss, Göttingen, Germany). Five random fields of muscle at a magnification of ×100 were inspected to reflect the nonhomogeneous nature of the muscle and to ensure that all areas of the muscle were represented. Muscle sections were photographed using Zeiss Axiovision software (v4.8, Zeiss) and subsequently analyzed with MyoVision analysis software (30) to calculate mean fiber CSA. To further define the differences in fiber size between groups, frequency distributions are shown (Supplemental Fig. S1; all Supplemental material is available at https://doi.org/10.6084/m9.figshare.16622953). Central nuclei were determined by counting all DAPI⁺ nuclei with clear separation from the dystrophin border by an experienced blinded assessor (31).

Satellite cells quantification.

To evaluate the number of Pax7⁺ cells (32, 33), air-dried frozen muscle sections (8 µm) were postfixed in 4% paraformaldehyde. Following fixation, sections underwent an epitope retrieval protocol at 92°C using a sodium citrate buffer (triodium citrate and water, pH adjusted with hydrochloric acid). Endogenous peroxidase activity was blocked using 3% hydrogen peroxide in PBS, and sections were reacted with blocking reagent (TSA, Invitrogen, Carlsbad, CA). Sections were then incubated in Pax7 primary antibody (dilution1:100, Developmental Studies Hybridoma Study Bank, Iowa City, IA) followed by the incubation with biotin-conjugated secondary antibody (dilution 1:1,000, Cat. No. 200-002-211, Jackson ImmunoResearch Laboratories, West Grove, PA) and followed by streptavidin-horseradish peroxidase to amplify (dilution 1:1,000, Cat. No. 016-030-084, Jackson ImmunoResearch Laboratories, West Grove, PA). Sections were then reacted with TSA-Alexa Fluor 488 (dilution 1:500, Cat. No. A32723, Invitrogen, Carlsbad, CA) to visualize antibody binding and stained with DAPI (dilution 1:10,000, Cat. No. D9542, Sigma-Aldrich, St. Louis, MO) to visualize nuclei. Pax7⁺/DAPI⁺ nuclei were counted and normalized to fiber number. Images were captured using the Zeiss Axio Imager M1 microscope and visualized using ZEN software (version 4.8, Zeiss).

Muscle damage.

To determine whether muscle fibers from the VL exhibited overt muscle damage as a result of the ACL injury mechanism, fibers were immunoreacted for IgG infiltration (32, 34). Muscle sections were fixed in ice-cold acetone (100%) for 10 min and washed in PBS. Muscle sections were then incubated in fluorescein isothiocyanate-conjugated mouse anti-IgG (dilution 1:500, Cat. No. 62-6511, Invitrogen, Carlsbad, CA) overnight at 4°C. Following incubation, slides were washed twice in 5% bovine serum albumin and once in PBS. To finish, sections were coverslipped using Vectashield mounting medium (H-1000, Vector Laboratories, Burlingame, CA). Images were captured using a Zeiss Axio Imager M1 microscope (Carl Zeiss, Göttingen, Germany). Five fields were taken to reflect the nonhomogeneous nature of the muscle and to ensure that all areas of the muscle were represented. Sections were photographed (×100) and subsequently analyzed using ZEN software (version 4.8, Carl Zeiss). Fluorescent intensity was quantified by the densitometric mean of each fiber (AU, arbitrary units) as previously described (32). IgG density, a marker of overt fiber damage, was counted by an experienced blinded assessor.

Neuromuscular communication.

To determine whether fibers exhibited signs of impaired neuromuscular communication following ACL injury, muscle sections were immunoreacted for NCAM (13). Muscle sections were fixed in ice-cold acetone (100%) for 20 min and washed in PBS. Muscle sections were then incubated in normal horse serum (NHS) (2.5%, Cat. No. S-2012, Vector Laboratories, Burlingame, CA) blocking reagent for 1 h. Following blocking, sections were incubated in anti-rat NCAM (AB5032, Millipore, Burlington, MA) in 2.5% NHS (dilution 1:50) overnight at 4°C. Following primary antibody incubation, sections were incubated in secondary antibody goat-anti rabbit 488 (dilution 1:200, A11034, Life Technologies, Carlsbad, CA) for 1 h at room temperature. To finish, sections were stained with DAPI (dilution 1:10,000, Cat. No. D9542, Sigma-Aldrich, St. Louis, MO) to visualize nuclei and coverslipped with Vectashield mounting medium (Cat. No. H-1000, Vector Laboratories, Burlingame, CA). Images were captured using a Zeiss Axio Imager M1 microscope (Carl Zeiss, Göttingen, Germany). Five fields were taken to reflect the nonhomogeneous nature of the muscle and to ensure that all areas of the muscle were represented. Sections were photographed (100×) and subsequently analyzed using ZEN software (version 4.8, Carl Zeiss). Fiber number and fibers staining positive for NCAM were subsequently analyzed by two, blinded investigators who characterized the fluorescent density of every fiber as either “positive” or “negative” (26). The number of NCAM⁺ fibers was then expressed as a ratio of total fibers per animal. The CSA of “positive” NCAM fibers was analyzed using ZEN software. Alterations in the temporal nature of NCAM expression were further analyzed using custom written MATLAB scripts (version R2019b, MathWorks, Natick, MA) to identify subpopulations of NCAM-expressing fibers within the groups using a bimodality coefficient (BC) (32) and then fit the underlying populations with probability density functions. These analyses were undertaken because unlike the established NCAM expression that occurs following nerve transection or crush injury, we expected NCAM expression in our model to be less predictable, with NCAM expression varying in onset, progression, and resolution occurring simultaneously as muscle fibers respond to the perturbation at the joint level (35, 36). Thus, we quantified the likelihood that groups of muscle fibers of the VL would express NCAM at a given density at any time following ACL injury. Temporal analysis were accomplished by obtaining individual cytoplasmic fluorescence values for each fiber (32), and then entering the continuous fluorescent density data into MATLAB and partitioning it into bins of optimal width using the Freedman–Diaconis rule (37) to fit the histograms to normal probability density functions. In the case of bimodality (where a BC > 0.555; Ref. 38), custom scripts were written for MATLAB to uncover the size, mean, and variance of each subpopulation within the group.

mRNA Isolation, Quantification, cDNA Synthesis, and RT-PCR Determination of MuRf-1, MAFbx, IL6, IL1β, TNFα, CD68, ACR2, Gadd45a, and Runx1

RNA was isolated from a frozen section of VL tissue (∼30–50 mg) that was homogenized using a handheld homogenizer in 1 mL of TRIzol (15596026, Invitrogen, Carlsbad, CA) (33, 39). The homogenate was spun at 12,000 g for 15 min at 4°C, the supernatant was removed, and 200 µL chloroform was added. The mixture was centrifuged again at 12,000 g for 15 min at 4°C to isolate the upper aqueous layer that was removed for precipitation. To precipitate the RNA pellet, the aqueous layer was isolated, mixed with isopropanol, and left for 20 min at room temperature. After incubation, samples were spun at 12,000 g for 15 min at 4°C. Once precipitation occurred, the pellet was washed and reconstituted using RNAse free water. RNA concentration was measured using a Nano Drop (ND-2000, Thermo Fisher Scientific, Foster City, CA). For each sample, cDNA was synthesized from 1 µg of total RNA using iScript Reverse Transcriptase according to the manufacturer’s protocol (1708840, Bio-Rad, Hercules, CA). Predesigned PrimeTime quantitative PCR assays (Integrated DNA Technologies, San Diego, CA) for 45-s pre-rRNA, MAFbx (assay ID: Rn.PT.58.351 116353), MuRF-1 (assay ID: Rn.PT.58.7441296), ACR2 (assay ID: Rn.PT.58.9283625), Gadd45a (assay ID: Rn.PT.58.7466023), Runx1 (assay ID: Rn.PT.58.35019962), CD68 (assay ID: Rn.PT.58.37733352), IL6 (assay ID: Rn.PT.58.13840513), IL1β (assay ID: Rn.PT.58.38028824), and TNFα (assay ID: Rn.PT.58.11142874) were used for mRNA analyses. Quantitative PCR was performed using PowerUp SYBR Green Master Mix (Thermo Fisher Scientific, Waltham, MA). The quantitative PCR data were normalized to the expression of Tubulin, NADH, and β2-microglobulin using the ΔCt methods (40).

Proteasome Activity

Alterations in the activity of the 20S proteasome after injury were determined using a commercially available kit (Cat. No. APT280, Millipore, Burlington, MA) (41–43). Specifically, proteolytic action of the 20S proteasome, chymotrypsin-like activity, was measured by quantifying the activity against the fluorogenic substrate LLVY (44). Briefly, frozen VL tissue (∼30–50 mg) was homogenized, using a handheld homogenizer, with homogenate buffer containing 50 mM Tris, 150 mM sodium chloride, 5 mM magnesium chloride, 1 mM ethylenediaminetetraacetic acid, and 1 mM dithiothreitol at a pH of 7.5. As previously described (45), samples were spun at 10,000 g for 15 min at 4°C, and the supernatant was removed. Activities were measured using 130 µg of protein, fluorescently tagged substrates, and a total well volume of 100 µL, according to the manufacturer’s instructions. Each sample was run in duplicate and in the presence and absence of the specific proteasome inhibitor Lactacystin. Proteasome inhibitor was used to subtract the nonspecific activity from specifically tagged fluorescence activity between two timepoints (time 0 and 2 h). The inhibited activity was then subtracted from the total activity allowing for the calculation of proteasome activity.

Statistical Analysis

Data were analyzed with a one-way analysis of variance with a Dunnett’s post hoc test using PRISM 8 software (San Diego, CA) to determine statistically significant differences between ACL-injured and control rats. For within-group comparisons, paired t tests were used. All values reported are means ± SE, and the statistical significance was set at P < 0.05.

RESULTS

Closed-ACL Injury Assessment

Knee dissection confirmed a proximal-to-mid-substance ACL tear in all knees (n = 48) that underwent the noninvasive ACL injury procedure. There was no evidence of any bony fracture or gross deformity to the knee joint and no obvious macroscopic damage to the articular cartilage that could be observed during dissection. An intra-articular hematoma was present in all of the ACL injured knees.

Muscle Wet Weight and Immunohistochemical Analyses of Fiber CSA

There was significant difference in VL muscle wet weight at 1 wk compared with control (P = 0.001, Fig. 1A). The normalized muscle weight to body weight ratios at 1-wk (P = 0.001) and 2-wk (P = 0.025) timepoints were also found to be significantly lower in the ACL-injured rats than in controls (Fig. 1B). Muscle fiber CSA of the ACL-injured limb was lower at 1 wk but was not significantly different from control (P = 0.113, Fig. 2C). To confirm that the differences in muscle wet weight at 1 wk were due to the injury, and not due to external factors, we further explored interlimb difference of mean fiber CSA. The right ACL-injured hind limb mean fiber CSA was significantly lower than the left hind limb at 1 wk (P = 0.011, Fig. 2D), confirming the loss of muscle mass due to injury itself. The fiber distribution curve for week 1 after injury is shifted to the left confirming the muscle weight results (Supplemental Fig. S1) and indicating that there are more small fibers in the 1-wk timepoint than all other groups.

Figure 1.

Lower muscle weight 1 wk after anterior cruciate ligament (ACL) injury. Muscle weight (A) and muscle to body weight ratio (B) for vastus lateralis muscles from control rats and ACL-injured rats at 6, 12, 24, and 48 h and 1, 2, and 4 wk after ACL injury. Values are means ± SE. $P < 0.05, difference from control.

Figure 2.

Lower fiber cross-sectional area (CSA) 1 wk after anterior cruciate ligament (ACL) injury. Representative images of vastus lateralis muscle stained with dystrophin (green) for control (A) and 1 wk (B). Mean cross-sectional area between groups (C) and within groups (D) for vastus lateralis muscles of rats without injury (control) and at 6, 12, 24, and 48 h and 1, 2, and 4 wk after ACL injury. Values are means ± SE. Con, control; L, left limb. #P < 0.05, different from 1 wk.

Muscle Atrophy Markers and Proteasome Activity

MAFbx mRNA abundance was significantly higher at 48 h after injury compared with control (P = 0.003, Fig. 3B). There were no differences in MuRF-1 mRNA abundance (P = 0.228, Fig. 3A). In addition, there were no significant differences in RNA concentration (P = 0.296, Fig. 4A) or 45 s pre-rRNA (P = 0.368, Fig. 4B). Chymotrypsin-like proteasome activity was also found to be significantly higher at 1 wk postinjury compared with control (P = 0.001, Fig. 5).

Figure 3.

Muscle atrophy F-box (MAFbx) mRNA abundance is higher at 48 h after anterior cruciate ligament (ACL) injury. Muscle RING-finger protein-1 (MuRF-1) mRNA abundance (A) and MAFbx mRNA abundance (B) for vastus lateralis muscles of rats without injury (control) and at 6, 12, 24, and 48 h and 1, 2, and 4 wk after ACL injury. Values are means ± SE. $P < 0.05, difference from control.

Figure 4.

No differences in RNA concentration and pre-ribosomal RNA after anterior cruciate ligament (ACL) injury. RNA concentration (A) and 45-s pre-rRNA abundance (B) for vastus lateralis muscles of rats without injury (control) and at 6, 12, 24, and 48 h and 1, 2, and 4 wk after ACL injury. Values are means ± SE.

Figure 5.

Proteasome activity is higher 1 wk after anterior cruciate ligament (ACL) injury. Proteasome activity for vastus lateralis muscles of rats without injury (control) and at 6, 12, 24, and 48 h and 1, 2, and 4 wk after ACL injury. Values are means ± SE. $P < 0.05, difference from control; &P < 0.05, difference from 48 h.

Neuromuscular Communication

Representative NCAM staining is depicted in Fig. 6A. Although the percentage of fibers displaying NCAM⁺ staining in ACL-injured rats was significantly higher than controls at 6 h postinjury (P = 0.019, Fig. 6B), there were no differences in CSA of NCAM⁺ fibers between any groups (P > 0.05, Fig. 6C). NCAM expression revealed a normal distribution in control muscle fibers (Fig. 7); however, at 48 h after injury, muscle fibers exhibited a wide range of fluorescent intensities that resulted in a rightward shift in mean fluorescence (273.18 AU, n = 1,318 fibers) and the emergence of a subpopulation of fibers with greater mean fluorescent densities of NCAM expression (681.03 AU, n = 1,145 fibers). Two distinct populations of NCAM-expressing fibers continued into the second week postinjury. Although the mean fluorescence of each population remained relatively stable (231.17 AU and 638.98 AU), the number of fibers in the subpopulation had decreased to n = 629 (Fig. 7 and Table 1). The probability distribution functions of the control group predicted a population of fibers that exhibited a normal distribution with a 95% confidence interval (CI) of 10.75–730.75 AU, spread around a mean NCAM density of 370.75 ± 179.70 AU in uninjured muscles (Fig. 7 and Table 1). At 6 h following ACL injury, mean NCAM density increased to 555.28 ± 285.53 AU (95% CI = 15.78–1,126.34) but maintained a normal distribution (Fig. 7 and Table 1). By 4 wk postinjury, the population of muscle fibers once again exhibited a normal distribution, with a mean fluorescent density and 95% CI that was very similar to the control group (Fig. 7 and Table 1), indicating a return to a more stabilized neuromuscular connection. To further evaluate indices of abnormal quadriceps neuromuscular communication, we conducted a targeted analysis of specific genes associated with neuromuscular instability and denervation. We observed that Runx1 was elevated 1 wk after ACL injury relative to controls (Table 2, P = 0.038). Gene expression of ACR2 and Gadd45a were not altered (Table 2).

Figure 6.

Signs of impaired neuromuscular communication at 6 h after anterior cruciate ligament (ACL) injury. Representative images of vastus lateralis muscles stained with neural cell adhesion molecule (NCAM; A). NCAM⁺ cell abundance (B) for vastus lateralis muscles of rats without injury (control) and at 6, 12, 24, and 48 h and 1, 2, and 4 wk after ACL injury. Mean fiber cross-sectional area (CSA; C) for NCAM⁺ fibers (bars) compared with mean fiber CSA of all fibers (red lines) for rats without injury (control) and at 6, 12, 24, or 48 h or 1, 2, or 4 wk following ACL injury. Values are means ± SE. $P < 0.05, difference from control.

Figure 7.

Differential subpopulations of fibers expressing neural cell adhesion molecule (NCAM) at 48 h or 1 or 2 wk after anterior cruciate ligament (ACL) injury. Fluorescent density of fibers expressing NCAM for vastus lateralis (VL) muscles of rats without injury (control) and at 6, 12, 24, and 48 h and 1, 2, and 4 wk after ACL injury. Results of the distribution analysis can be found in Table 1. AU, arbitrary unit.

Table 1.

Results of the neural cell adhesion molecule distribution analysis

					48 h		1 wk		2 wk
					Subpopulation (BC = 0.5589)		Subpopulation (BC = 0.6503)		Subpopulation (BC = 0.7429)
	Control	6 h	12 h	24 h	1	2	1	2	1	2	4 wk
Mean (NCAM expression), AU	370.75	555.28	432.15	256.14	273.18	681.82	309.16	810.31	231.17	638.98	428.94
Standard deviation	179.70	285.53	162.18	69.74	61.03	183.10	105.19	172.51	54.45	182.79	236.19
95% confidence interval	10.75–730.75	15.78–1,126.34	110.12–756.38	116.16–395.62	151.12–395.24	334.52–1,048.02	98.78–519.54	460.26–1,155.33	140.27–322.07	273.33–1,004.56	43.44–901.32
N	3,569	2,535	2,758	3,336	1,318	1,145	3,425	774	2,197	629	3,355

Values are means, standard deviations, 95% confidence intervals, and bimodality coefficients (BC) for each group. Subpopulation indicates a significant bimodality coefficient and bimodal distribution of fiber fluorescent intensity (b > 0.5555). AU, arbitrary units; NCAM, neural cell adhesion molecule.

Table 2.

mRNA abundance for markers related to denervation and neuromuscular junction dysfunction for vastus lateralis of rats

	Control	6 h	12 h	24 h	48 h	1 wk	2 wk	4 wk
Runx1	1.0 ± 0.44	1.2 ± 0.27	3.3 ± 1.3	1.3 ± 0.28	3.1 ± 0.68	4.6 ± 0.85*	3.2 ± 1.1	1.3 ± 0.44
Gadd45a	1.0 ± 0.33	0.41 ± 0.08	1.1 ± 0.26	0.62 ± 0.14	1.1 ± 0.20	1.8 ± 0.57	0.72 ± 0.11	0.36 ± 0.09
ACR2	1.0 ± 0.40	3.17 ± 1.20	2.4 ± 0.97	1.2 ± 0.43	0.92 ± 0.21	2.9 ± 1.1	1.4 ± 0.43	3.6 ± 1.9

Values are means ± SE of rats without injury (control) and at 6, 12, 24, and 48 h and 1, 2, and 4 wk after anterior cruciate ligament injury. ACR2, Arabidopsis thaliana genome; Gadd45a, growth arrest and DNA damage inducible-α; Runx1, run-related transcription factor 1. *P < 0.05, difference from control.

Muscle Damage Markers

Representative IgG, dystrophin, and Pax7 staining is depicted in Fig. 8, A, C, and E, respectively. Pax7⁺ cell abundance and IgG infiltration were not different between the control and ACL-injured rats (Fig. 8, B and F), indicating that the VL muscle did not undergo overt muscle damage in response to ACL injury. There were also no differences in the number of central nuclei per fiber (P = 0.693, Fig. 8D), further reinforcing the lack of muscle damage. CD68 mRNA abundance was not different with ACL injury (P = 0.363, data not shown) and mRNA abundance of IL6, IL1β, and TNFα was not detectable in our muscle tissue, demonstrating that there was likely not an inflammatory response in the VL following ACL injury.

Figure 8.

Anterior cruciate ligament (ACL) injury does not cause overt muscle damage. Representative images of vastus lateralis (VL) muscle stained with immunoglobulin G (A), dystrophin (C), and Pax7 (E). IgG mean intensity (B), central nuclei number (D), and Pax7+ cell abundance (F) for VL muscles of rats without injury (control) and at 6, 12, 24, and 48 h or 1, 2, and 4 wk after ACL injury. AU, arbitrary unit; Con, control. Values are means ± SE.

DISCUSSION

The primary purpose of our work was to use a preclinical closed-ACL injury model to assess the cellular mechanisms of action that regulate quadriceps muscle mass and the time course of events after injury. Our most significant finding is that we observed signs of disrupted neuromuscular communication at just 6 h postinjury, which preceded cellular determinants of muscle atrophy and declines in muscle weight. We also observed a partial recovery of muscle size by 2 wk postinjury that likely speaks to the early reloading strategy that was used by the rodents. Outcomes from the current study indicate that the loss of muscle mass after injury was not related to direct muscle damage but likely driven by a link between early neural disruption and increased protein degradation. This work represents a significant advancement for the field, as understanding which atrophic factors to target, and when, will help to substantiate and guide postinjury rehabilitation.

Persistent quadriceps atrophy following ACL injury has long been attributed to the alterations in neurological function, as patients are often unable to fully contract their muscle after injury. Although the exact mechanisms driving this response are unknown, altered activity after ACL injury has been well-documented using electrophysiological techniques at the spinal (10, 11) and cortical levels (11, 46, 47), negatively influencing the ability to generate a muscle contraction. Although the nervous system and skeletal muscle are known to have many inherent physical and biological links, the extent to which alterations in neural activity after ACL injury directly contribute to muscle atrophy is unclear, as neither its time course, nor true prevalence, in human or animal ACL injury models has been determined. The nonsurgical ACL injury model that was used in this study is an important step forward to better understand the muscle injury response.

We assessed muscle-cell behavior using NCAM, an adhesion molecule, ubiquitously expressed in both neurites and myotubes during embryonic development due to its essential role in initiating and establishing neuromuscular communication (48). In adult muscle, NCAM is localized to the junctional, postsynaptic membrane to provide connectivity, stability, and maintenance to the neuromuscular junction (49–51), but its expression can be rapidly upregulated. When needed, the reexpression of NCAM in adult muscle involves redistribution outward from the neuromuscular junction to extra-junctional areas along the cell surface (50, 52, 53). Simultaneously, polysialylation of NCAM produces the highly soluble PSA-NCAM that disperses throughout the cytoplasm of the cell where it engages in intracellular and intercellular signaling, gene expression, and calcium homeostasis (29, 49, 50, 54). The synthesis and extra-junctional insertion of new NCAM molecules create a large area of attraction that promotes synaptogenesis or remodeling and improved connectivity of the existing neuromuscular junction (29, 49, 50, 54). Synaptogenesis permits a denervated muscle fiber to restore a synapse at the neuromuscular junction and recover its own function (55), whereas polysynaptogenesis permits an innervated muscle fiber with high NCAM expression to erroneously add another nerve to the neuromuscular junction, temporarily (51, 56). The rapid reexpression of NCAM in adult muscle and its association with synaptogenesis has led some to speculate a recapitulation of embryonic muscle development and a causal association between NCAM and denervation/reinnervation (57–59). However, NCAM staining is not a measure of neuromuscular junction innervation or denervation but has been a marker of inactivity by itself since 1986 (50). NCAM has been shown to be associated with tenotomy (26), paralysis (60), and hind limb suspension (61, 62), which are all disuse muscle states without denervation. The results in our study, therefore, show that indeed NCAM is upregulated after ACL injury under circumstances where there is impaired neuromuscular communication and remodeling over the course of recovery but not denervation. Specifically, our analyses show that disrupted neuromuscular communication is evident at just 6 h postinjury (Fig. 6) and precedes markers of muscle atrophy that first peak at 48 h (Fig. 3), and reductions in muscle size that occurred at 1 wk postinjury (Fig. 1). The fiber-to-fiber NCAM fluorescent densities (Fig. 7) reveal irregularities in the neuromuscular communication states that continued for weeks after injury. Most prominently, underlying fiber subpopulations of NCAM expression were found within 48 h of the ACL rupture, with 50% of muscle fibers likely to express higher than normal NCAM. This subpopulation was found to be smaller at 1 and 2 wk postinjury, as the fiber NCAM expression shifts back to the left and demonstrates recovery within 4 wk. The targeted analysis of neuromuscular genes associated with denervation reinforce the notion that elevation of NCAM in our study was associated with impaired neuromuscular communication and remodeling as only Runx1 was elevated 1 wk after ACL relative to controls (27).

Neuromuscular instability events promote muscle atrophy via activation of FoxO transcription factors (63). Upregulation of the FoxO family of transcription factors is known to stimulate gene coding for the E₃-ubiquitin ligases that mediate protein breakdown (64). Indeed, we found higher levels of markers of disrupted neuromuscular communication hours after injury, which was followed by an upregulation of MAFbx and increased 20S proteasome activity, part of the ubiquitin proteasome system. These early data can be used to help substantiate and guide empirically driven solutions that emphasize the use of early neurogenically targeted interventions after ACL injury.

We also observed that pathways involved in muscle protein degradation were involved as early as 48 h postinjury, where an upregulation of MAFbx, an E₃-ubiquitin ligase that directs the polyubiquitination of proteins for proteolysis by the 26S proteasome (65), was observed. This observation of an early upregulation of MAFbx after a stressor or injurious event is common (66), and our work agrees with that of others that have used ACL transection models to study atrophy pathways and observed this at the 3-day postinjury timepoint (21, 22). In addition to the upregulated E₃-ubiquitin ligase, we also observed increased 20S proteasome activity at 1 wk postinjury (Fig. 6). The power of our model is that the noninvasive mechanism of injury provides a controlled system to study the effects of ACL rupture [unlike ACL transection models (21–23)] and is also not confounded by the use of postinjury analgesics known to interfere with the native biological response and signaling pathways (67, 68). Hence, it is clear that ACL injury triggers an early response by the ubiquitin proteasome pathway that precedes the onset of muscle atrophy. MAFbx is a widely regarded factor in muscle atrophy, and the overexpression of MAFbx, and subsequent increase in proteasome activity, in some cases constitutes a series of events in the atrophying process (66). It has been shown that the deletion of MAFbx can preserve muscle mass in a variety of atrophy-inducing conditions (66). Additional studies are required to identify the primary role of MAFbx (and MuRF-1) in atrophic conditions and the benefits of therapeutically blunting their effect.

A surprising finding in our study is that muscle mass was recovered by 2 wk postinjury, but we did not observe any differences in markers associated with ribosomal turnover (Fig. 5). This was intriguing, because it has generally been argued that to promote skeletal muscle hypertrophy, an increase in muscle protein synthesis is necessary (69). To this point, our data show that markers of skeletal muscle atrophy were only transiently upregulated after injury, but that blocking protein degradation may be substantial enough to restore muscle size (Fig. 4, A and B, and Fig. 5). This may speak to the benefit of early reloading after ACL injury, as this strategy was used by the rodents, but is unlike routine clinical care that often consists of early off-loading after injury (70). Thus, it seems plausible that early reloading of the ACL-injured limb may help to promote a net positive protein metabolism and, therefore, help to mitigate the muscle atrophy processes that plague many patients. Studying this question directly, the role of early reloading after ACL injury on muscle atrophy, would be an important addition to future studies.

The ACL injury model used here is fundamentally different from others that have used ACL transection models, as surgical transections (or even sham transections) violate the joint capsule creating a cascade of negative neuromuscular events that confound initial biological observations (21–23). The transection procedure is also known to cause an inflammatory response (including joint swelling and increases in macrophage abundance) and induce upregulations in atrophic markers (MuRF-1 and atrogin-1) (21, 22). Because of the invasive mechanism of injury, it is impossible to untangle the effect of ACL injury from the surgery on muscle health in a transection model. Our model overcomes these experimental barriers, as it is able to consistently produce an isolated noninvasive ACL injury. To confirm that our injury model did not confound observations by causing overt muscle damage, we found no differences in markers of damage in any of the experimental groups (Fig. 8). We also found no signs of IgG infiltration indicating that our model did not disrupt the surrounding musculature (Fig. 8B). We further confirmed that our model did not introduce any confounding effects by measuring CD68 mRNA abundance, where we saw no differences and TNFα, IL1β, and IL6, which were not detectable in our muscle samples, indicating there were no gross muscle differences in the inflammatory response between our injury model and controls. The power of our study is that our noninvasive mechanism of ACL injury isolates our observations to the injury and is also not confounded by the use of postinjury analgesics known to interfere with the native biological response.

Longitudinal, live animal experiments are powerful scientific tools for systematically detecting mechanistic factors that give rise to challenging clinical conditions. The similar anatomical makeup of the rodent knee and function to that of humans makes the rat knee a useful model to study neuromuscular adaptations after injury. Our preclinical data are most relevant to those suffering from ACL rupture, as we did not directly study the influence of an open surgical reconstructive method on outcomes. We have previously established that much of our data, like impaired gate, are similar to what is frequently observed in ACL-injured patients, who after several weeks of injury can ambulate with reduced knee flexion angles to help stabilize an unstable knee (25) and as a means to improve functionality and mitigate the dramatic losses in muscle strength and volume that occur immediately after ACL rupture. However, it should be noted that there are several limitations to the study. First, we primarily explored the ubiquitin proteasome system. Although this system is responsible for the majority of protein degradation, it is not solely responsible. Other pathways (including cathepsins, calpains, caspase-3, and autophagy) are known to be involved in the breakdown of muscle proteins and would be important additions to future studies. Emerging evidence indicates that fibrotic tissue formation may also interfere with muscular health after ACL injury and that quadriceps muscle may be vulnerable to selective fiber type atrophy after ACL injury (13–15). Thus, the evaluation of other key components of muscular health is important. Future studies should consider extending the time period of observation and evaluating muscle health and atrophy when posttraumatic osteoarthritis has developed. Despite these limitations, this study provides important insight into the biology of adaptations after injury that are key to directing muscle atrophy in patients with ACL injury.

Conclusions

Characterizing the key parameters that regulate muscle mass, and the time course of events, is important for identifying therapeutic targets to mitigate the muscle atrophy that plagues many of those that suffer from ACL injury. Using our preclinical rodent model of ACL injury, we observed signs of abnormal muscle-nerve communication that preceded determinants of muscle atrophy and declines in muscle size. We also observed a partial recovery of muscle by 2 wk postinjury that likely speaks to the early reloading strategy that was used by the rodents. The loss of muscle mass was not found to be related to markers of overt muscle damage caused by the injury model. All together, these data help to establish the time course of muscle atrophy after ACL injury, suggesting that clinical care may benefit from the application of acute neurogenic interventions and early gait reloading strategies.

SUPPLEMENTAL DATA

Supplemental Fig. S1: https://doi.org/10.6084/m9.figshare.16622953.

GRANTS

This work was supported by National Institute of Arthritis and Musculoskeletal and Skin Diseases Grant K01AR071503 (to L. K. Lepley).

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

E.R.H., S.M.D., C.A.J., C.L., E.E.D.-V., T.A.B., and L.K.L. conceived and designed research; E.R.H., S.M.D., C.N.P., K.A.B., and L.K.L. performed experiments; E.R.H., C.N.P., K.C., D.W.V.P., A.L.C., K.A.B., and L.K.L. analyzed data; E.R.H., E.E.D.-V., K.C., D.W.V.P., A.L.C., and L.K.L. interpreted results of experiments; E.R.H. prepared figures; E.R.H. and L.K.L. drafted manuscript; E.R.H., S.M.D., C.N.P., K.C., D.W.V.P., A.L.C., K.A.B., C.A.J., C.L., E.E.D.-V., T.A.B., and L.K.L. edited and revised manuscript; E.R.H., S.M.D., C.N.P., K.C., D.W.V.P., A.L.C., K.A.B., C.A.J., C.L., E.E.D.-V., T.A.B., and L.K.L. approved final version of manuscript.