Immunohistochemical labeling of ongoing axonal degeneration 10 days following cervical contusion spinal cord injury in the rat

Abstract

STUDY DESIGN:

Experimental Animal Study.

OBJECTIVE:

To continue validating an antibody which targets an epitope of neurofilament light chain (NF-L) only available during neurodegeneration and to utilize the antibody to describe the pattern of axonal degeneration 10 days post-unilateral C4 contusion in the rat.

SETTING:

University of Florida laboratory in Gainesville, USA.

METHODS:

Sprague Dawley rats received either a unilateral 150kdyne C4 contusion (n = 4 females, n = 5 males) or a laminectomy control surgery (n = 2 females, n = 3 males). Ten days following SCI or laminectomy, spinal cords and brainstems were processed for immunohistochemistry. Serial spinal cord and brainstem cross-sections were stained with the degeneration-specific NF-L antibody (MCA-6H63) and dual labeled with either an antibody against the C-terminus portion of NF-L (NF-L-Ct), to label healthy axons, or an antibody against amyloid precursor protein (APP), considered the current “gold standard” for identifying axonal injury. The pattern of ongoing axonal degeneration was assessed.

RESULTS:

Spinal cord and brainstem cross-sections from injured rats had punctate MCA-6H63 positive fibers with a pathological appearance, loss of anti-NF-L-Ct colabeling, and frequent colocalization with APP. Immunopositive fibers were abundant rostral and caudal to the lesion in white matter tracts that would be disrupted by the unilateral C4 contusion. This pattern of staining was not observed in control tissue.

CONCLUSIONS:

The MCA-6H63 antibody labels degenerating axons following SCI and offers a tool to quantify axonal degeneration.

INTRODUCTION

Axonal degeneration begins immediately after spinal cord injury (SCI) [1]. When a portion of an axon is injured beyond its capacity for repair, the distal portion undergoes Wallerian degeneration while the proximal portion undergoes axonal dieback or retraction. Wallerian degeneration is characterized by the formation of axonal spheroids, or irregular swellings of the axon, which creates a “bead on a string” appearance, followed by fragmentation, and ultimately degeneration of the entire distal portion of the axon [2]. During axonal dieback or retraction, the axon undergoes prolonged retrograde degeneration with formation of a retraction bulb followed by stabilization of the axon, and typically a failed attempt to regenerate [2]. SCI causes axonal shearing, compression, and severing that results in an initial wave of axonal degeneration [3]. A subsequent cascade of factors including ischemia, excitotoxicity, oxidative stress, and inflammation contribute to the secondary injury reflected by the extensive collateral damage of additional axons which may have otherwise been spared [4]. Mitigating the secondary injury cascade is a target for neuroprotective strategies [5]. The axonal degeneration that results from both the primary and secondary injury has profound impacts on functional ability, quality and longevity of life, and responsiveness to rehabilitation strategies. Axonal degeneration can lead to sensorimotor impairment, chronic pain, spasticity, autonomic dysfunction, bowel and bladder dysfunction, and respiratory failure.

Axonal degeneration is typically assessed histologically using important outcome measures such as axonal pathology, axonal counts, and stains identifying actively degenerating axons. Identification of axons with pathological features such as axonal spheroids or retraction bulbs provides insight into the process of axonal degeneration and can allow for low-throughput assessments of the extent of axonal degeneration [2, 6, 7]. Immunohistochemical stains for epitopes associated with healthy axons can be quantified to assess axonal loss compared to a control group [1, 8]. Axonal loss as an outcome measure has improved understanding of how different lesion models impact axonal degeneration and the importance of axonal sparing [1, 9]. In a recent publication, we epitope mapped two commercial monoclonal NF-L antibodies used in the Uman Diagnostics NF-Light^™ ELISA, the Quanterix Simoa^™, and similar assays [8]. These assays efficiently detect elevations of NF-L in the blood and CSF in the setting of neurodegeneration and have become widely used in a rapidly growing variety of experimental and clinical studies [10, 11]. We showed that both Uman antibodies bind neighboring epitopes in the center of the α-helical “Coil2” segment of NF-L, amino acids 311–362 of the human sequence [8]. These antibodies did not stain axons in tissue sections from animals without traumatic CNS injury, apart from a few very rare profiles which we interpreted as axons undergoing spontaneous degeneration. In sharp contrast, abundant Uman antibody staining was observed in spinal cord white matter tracts of rats at 1–5 days post-SCI. The immunopositive axons were typically swollen, irregular in shape, discontinuous, and in some cases sinusoidal, consistent with a degenerating or degenerated appearance [8]. We found that several NF-L antibodies that bind to the non-helical C-terminal “tail” segment of NF-L, here referred to as NF-L-ct reagents, labeled non-injured “healthy” axons but not the Uman antibody positive material. Spinal cord sections without SCI but exposed to proteases demonstrated robust axonal immunopositivity with Uman-type antibodies and concomitant loss of binding with NF-L-ct antibodies [8]. We concluded that during axonal degeneration, proteases degrade neurofilaments revealing the Uman-type epitopes while also destroying NF-L-ct epitopes. We noted little overlap between the two staining profiles suggesting that the transition from NF-L-Uman negative, NF-L-ct positive to NF-L-Uman positive, NF-L-ct negative is rapid. A novel panel of antibodies against amino acids 311–362 of human NF-L was created, and these all stained in the degeneration-specific fashion as described for the Uman reagents. One of these, MCA-6H63, was utilized in the present study.

In the current work, we studied spinal tissues from rats that had a mid-cervical (C4) contusion SCI to accomplish two major goals. The first was to continue the validation of the MCA-6H63 antibody for its ability to specifically identify degenerating axons after SCI. This was accomplished through comparison to traditional methods of identifying degenerating axons. The second goal was to utilize the MCA-6H63 antibody to assess the pattern of degenerating axons at a subacute time point (10 days post-injury), relevant for ongoing secondary injury and the antibody’s potential use as an outcome measure in neuroprotection studies.

METHODS

Experimental animals

All procedures were approved by the University of Florida’s Animal Care and Use Committee and in compliance with the National Institute of Health guidelines. A total of 14 Sprague Dawley rats (n = 6 females, n = 8 males; Sprague Dawley^® SD, Envigo Indianapolis, IN) were used in this experiment. Rodents were housed on a 12:12 light-dark cycle with ad libitum access to food and water.

Laminectomy and spinal cord injury surgeries

Rats (12 weeks old and 218 ± 32 g) were anesthetized with ketamine (80 mg/kg) and xylazine (10 mg/kg; IP). Contusion spinal cord injuries were performed (n = 9, 4 females, 5 males) as previously described [12, 13]. The skin was incised above the cervical spinal cord to expose the muscles over C2–C6. The muscles were reflected and the vertebrae were exposed using the curette. A laminectomy was performed on the right side of the fourth cervical vertebrae (C4). In the contusion group, stabilization clamps were placed on C3 and C6 and the Infinite Horizons impactor (Precision Systems and Instrumentation, Lexington, KY) with a 2.5 mm tip was used to deliver a 150kdyne contusion to the right side of the spinal cord. The stabilization clamps were removed and the overlying muscle layers were sutured with sterile 4-0 vicryl. The skin was closed with 9 mm wound clips. Post-operative medications, including three days of buprenorphine(0.03 mg/kg every 12 hours) and 48 hours of carprofen (5 mg/kg, q.d), baytril (5 mg/kg, q.d.), and lactated ringer’s solution (10 mL/day, q.d), were administered. Rats were monitored daily for signs of distress, dehydration, and weight loss with appropriate veterinary care given as needed. A group of rats (n = 5, 2 females, 3 males) received only a laminectomy, with removal of the overlying vertebrae with no injury to the spinal cord.

Immunohistochemistry

Ten days post-contusion SCI or laminectomy surgery rats were injected with 1 mL of intraperitoneal Beuthanesia-D Special (0.39 mg/mL pentobarbital, 0.05 mg/mL phenytoin). Rats were transcardially perfused with saline followed by 4% paraformaldehyde. The spinal cords were harvested and then post-fixed in 4% paraformaldehyde overnight. The tissue was then cryoprotected with 30% sucrose until the tissue sank, embedded into Optimal Cutting Temperature media, frozen, and stored at −80°. Spinal cords were cryosectioned in 20 μm thick sections into a seven-section series to generate slides with sections 120 μm apart. The slides used in the contained study were used as part of a previous study [12]. A full set of slides from 6 rats (n = 3 females, n = 3 males) were randomly selected from the original 10 rats generated as part of an injured, untreated control group. The researcher who randomly selected these slides was blinded to the individual rat data from previous study. The sample size was selected to ensure reproducibility of observations across several female and male rats. The same process was completed for 5 (n = 2 females, n = 3 males) uninjured control rats. These slides were used for immunofluorescent staining with the degeneration-specific mouse monoclonal antibody to NF-L Degenotag^™ peptide (EnCor Biotechnology Inc., Gainesville, FL, USA; MCA-6H63; 1:5000; RRID: AB_2923484) and rabbit polyclonal antibody to NF-L-ct (EnCor Biotechnology RPCA-NF-L-ct; 1:5000; RRID: AB_2861179). MCA-6H63 binds to an epitope centered on amino acids 311–315 of the human NF-L sequence, while RPCA-NF-L-ct was raised against amino acids 515–543 of the rat NF-L sequence. These slides were used to map MCA-6H63 staining throughout the cervical spinal cord. An additional set of slides in these animals was used for cresyl violet staining as per previously described methods [13]. To compare APP and MCA-6H63 staining, we selected two slides from 5 (n = 2 females, n = 3 males) randomly selected spinally injured rats. Slides spanned the caudal medulla to approximately C8 and were dual stained with MCA-6H63 and rabbit polyclonal antibody to the C-terminus of amyloid precursor protein (APP; clone Y188; Abcam, Waltham, MA, USA; ab32136; 1:200; RRID: AB_2289606). The APP antibody is a marker of disturbed axonal transport which accompanies axonal degeneration. It is considered the gold standard for the detection of injured axons after traumatic brain injury. Secondary antibodies were goat anti-Mouse AF488 (Invitrogen, Waltham, MA, USA; a32723; 1:1000; AB_2633275) and goat anti-rabbit AF 594 (Invitrogen, a32740; 1:1000; AB_2762824). Injured spinal tissue was also incubated with isotype and concentration matched control Ig instead of exposure to primary antibody to provide another negative control. Importantly, the MCA-6H63 antibody staining is not compatible with antigen retrieval methods that can cause degradation of neurofilament polypeptide and produce positive labeling not due to the injury or disease state.

Imaging and quantification

Microscopy was conducted using the Keyence microscope (BZ-X700, Keyence Corporation of America, Itasca, IL). Before imaging, all sections in each animal were screened and assessed for the presence and pattern of staining. Representative images were selected for their resemblance with the composite pattern of staining observed across animals in each group (injured vs laminectomy control). Whole section images were created by stitching 10X images using Adobe Photoshop Photomerge. The intensity of fluorescent images was increased using levels to enhance visibility. Cresyl violet sections were white balanced. All adjustments made to images were applied to the entire image and standardized across images in both experimental and control groups. A 3D reconstruction of the spinal cord was created for one rat to demonstrate a representative lesion. To create the 3D reconstruction, serial 2X images of cresyl violet stained sections were captured using the Keyence microscope. Adobe Illustrator was then used to trace the section, the white/gray matter junction, and the lesion. The traced images were converted into 3D images and stacked. While creating the representative 3D reconstruction, sections with sufficient damage, such as rips or folds, which prevented accurate tracing, were eliminated (13 out of 78 total sections were eliminated). These sections were skipped during the creation of the 3D reconstruction. No more than 2 consecutive sections were eliminated. The example 3D reconstruction was chosen based on the best cresyl violet staining with the minimal number of excluded sections to depict the most accurate reconstruction.

The percentage of fibers labeled with APP only, MCA-6H63 only, or both APP and MCA-6H63 was assessed in specific regions of interest. These regions included the ipsi- and contralesional rostral dorsal fasciculus, rostral and caudal ventrolateral white matter, and caudal lateral and ventral white matter as well as the rostral dorsal corticospinal tract. These areas were chosen based on where abundant MCA-6H63-positive fibers were observed during the MCA-6H63 mapping portion of this study. The putative identity of these tracts was determined using axonal tract anatomy [14, 15]. Regions were imaged on a Keyence microscope and consistently positioned across spinal cord cross-sections using the relative location of the gray matter as an anatomical landmark. These regions were selected to include axonal tracts rostral (~C3), caudal (~C7), ipsilateral, and contralateral to the lesion as well as tracts presumably undergoing Wallerian degeneration versus axonal dieback. A total of 84 40X images were evaluated from 12 randomly selected sections (n = 5 rostral, n = 7 sections caudal) spanning the n = 5 injured spinal cords. Immunopositive fibers were manually counted and characterized as APP-positive, MCA-6H63-positive, or colocalized. Fibers were considered colocalized if both APP and MCA-6H63 staining were present within the same fiber. The percentage of immunopositive fibers that were APP-positive, MCA-6H63-positive, or colocalized were calculated for all regions of interest and plotted via GraphPad Prism^® along with the median and interquartile range. The average percent of colocalized fibers relative to the total number of immunopositive fibers was compared between tracts presumably undergoing Wallerian degeneration versus axonal dieback. Tracts presumably undergoing axonal dieback included the rostral dorsal corticospinal tract and the caudal ventrolateral white matter (putative spinothalamic tract and spinocerebellar tract) [14, 15]. Tracts presumably undergoing Wallerian degeneration included the rostral dorsal fasciculi and the rostral ventrolateral white matter [14, 15]. Because the ventral and lateral white matter contains overlapping ascending and descending axonal tracts, these regions were not analyzed as part of the comparison between axonal dieback and Wallerian degeneration [14, 15]. The average percentage of APP-positive, MCA-6H63 positive, and colocalized fibers were also compared between locations rostral vs caudal and ipsilateral vs contralateral to the lesion. Comparison between axonal dieback versus Wallerian degeneration and the locations of regions of interest were plotted using the median and interquartile range using GraphPad Prism^®. Normality of distribution was tested using the Shapiro-Wilk test. Statistical analysis was conducted using a two-way ANOVA with Tukey’s multiple comparison using GraphPad Prism^®.

RESULTS

Lesion histology after unilateral C4 contusion

Cresyl violet staining allowed for the assessment of lesion histopathology. Within the lesion (solid black line Fig. 1A) there was destruction of normal tissue architecture, loss of neurons, and the presence of blood, cellular debris, and vacuoles. The lesion extended contralesionally in 5 of 6 rats, primarily in the center of the spinal cord and the dorsal white matter (Fig. 1A). The lesion extended rostrally and caudally in the ipsilesional white matter approximately one to two spinal segments. A representative three-dimensional reconstruction of the lesion depicts the area of injury (Fig. 1B, C).

Fig. 1. Lesion description and MCA-6H63 validation.

A Cresyl violet stained cross section at the lesion epicenter. The solid line indicates the lesion epicenter while the dashed lines indicate the border of the white matter and the gray/white matter border. Example 3D reconstruction of the cervical spinal cord following C4 unilateral contusion (B dorsolateral perspective of the ipsilesional cord; C dorsal view). Ventral spinal cord ipsilateral and caudal to the lesion (D, F) or laminectomy (E). D Injured spinal cord stained with MCA-6H63 (green). E Uninjured (laminectomy) spinal cord stained with MCA-6H63 (green). F Injured spinal cord incubated with concentration matched, isotype control Ig. Dual staining of an example ipsilesional cervical ventral root from an injured spinal cord with MCA-6H63 (green; G, I) and NF-L C-terminus (red; H, I). Dual staining of ipsilesional injured spinal cord cross sections with MCA-6H63 (green) and APP (red) in the rostral dorsal white matter (J–L) and caudal lateral white matter (M–O). P The percent of positive fibers labeled with APP, MCA-6H63, or both in 40X regions of interests (n = 84 regions of interests across 5 animals). Q The percentage of fibers with APP, MCA-6H63, or colocalized staining rostral, caudal, contralateral, and ipsilateral to the lesion as well as in tracts presumably undergoing Wallerian degeneration (WD; rostral dorsal fasciculi and rostral ventrolateral white matter) and axonal dieback (AD; rostral dorsal corticospinal tract and caudal ventrolateral white matter). Data represented as median and interquartile range.

MCA-6H63 antibody stains degenerating axons

Spinal cord sections stained with MCA-6H63 had abundant small and irregularly shaped punctate immunopositive profiles in the white matter rostral and caudal to the lesion. This pattern of staining was consistent in all rats in the contusion group (n = 6/6; Fig. 1D) and was not observed in the laminectomy-only group (n = 5/5; Fig. 1E). Spinally injured sections incubated with an isotype and concentration matched control Ig did not have positive staining rostral or caudal to SCI (Fig. 1F). MCA-6H63 positive fibers were also observed in the ventral roots (Fig. 1G). MCA-6H63 positive axons in the ventral roots had a beaded, pathological appearance demonstrated by the irregular width, shape, and contour of the axon (Fig. 1G, I). Axons stained with NF-L-ct had a more regular width, shape, and contour indicative of healthy axons (Fig. 1H, I). As noted in our previous study [8], axons labeled with the MCA-6H63 antibody were generally not positive for NF-L-ct (Fig. 1G–I). Cross sections of injured spinal cords dual stained with APP and MCA-6H63 (rostral dorsal white matter, Fig. 1J–L and caudal lateral white matter, Fig. 1M–O) revealed frequent colocalization (examples: white arrows). However, there were fibers interspersed which were only positive for MCA-6H63 or APP. The MCA-6H63 antibody typically labeled defined, irregularly shaped puncta (Fig. 1J, M). In fibers with colocalized staining (white arrows, Fig. 1J–O), APP positive labeling (Fig. 1K, N) typically either wrapped around an MCA-6H63 core or appeared as a cluster of punctate APP positive labeling surrounding the MCA-6H63 (Fig. 1L, O). There were also fibers in which both MCA-6H63 and APP-labeled fibers demonstrated a more complete overlap of pixels, but the shape and contour of the profiles typically differed slightly (Fig. 1L, O). Disorganized whisp- or burst-like shaped MCA-6H63 fibers were visualized that, if colabeled, typically had intermingled APP labeling (Fig. 1M–O). Each morphological variation in MCA-6H63 and APP-labeled fibers was present in colocalized or solely labeled fibers. APP and MCA-6H63 staining were present in the same putative white matter tracts. Across all regions of interest, 40.0% (IQR: 30.1–52.5%) of fibers stained for APP only, 13.3% (IQR: 7.8–21.2%) of fibers stained for MCA-6H63 only, and 44.5% (IQR: 30.1–51.5%) of fibers stained both with APP and MCA-6H63 per region (Fig. 1P). The average percentage of APP-positive, MCA-6H63-positive, or colocalized fibers did not differ depending on location ipsilateral or contralateral to injury or as a result of the presumed process of degeneration, axonal dieback versus Wallerian degeneration (Fig. 1Q). There was a statistically significant increase in APP positive fibers (p = 0.006) and decrease in colocalized fibers (p = 0.008) rostral to the injury relative to caudal to the injury with no differences in the percentage of solely MCA-6H63 positive fibers.

Pattern of axonal degeneration 10 days following C4 contusion

Rostral to the lesion (Fig. 2A–H), staining was most abundant in the ventrolateral (n = 6/6) and dorsal (n = 6/6) white matter ipsilateral to the lesion. Dorsal white matter staining appeared most robust in the dorsal fasciculi with some staining located in the region of the dorsal corticospinal tract. Dorsal white matter staining was bilateral in most rats (n = 5/6). The rat without contralesional dorsal column staining did not have significant extension of the lesion across midline on cresyl violet analysis of the lesion. Immunopositive fibers were also present in the contralesional ventrolateral white matter (n = 6/6) but were more sparse than the ipsilesional staining. Caudal to the lesion (Fig. 2I–P), ipsilesional immunopositive fibers were robust in the ventral and lateral white matter (n = 6/6). Contralesional staining was observed in the ventral, ventrolateral, and lateral white matter (n = 6/6) but was most abundant in the ventral white matter (n = 5/6). The rat that did not exhibit robust staining in the contralesional ventral white matter did not have contralesional extension of the lesion across midline on cresyl violet lesion analysis. Occasional immunopositive fibers were observed in the dorsal column caudal to the lesion (n = 6/6).

Fig. 2. Pattern of MCA-6H63 staining in the cervical spinal cord 10 days following unilateral C4 contusion.

Representative images (A–C, E–G, I–K, M–O) and composite drawings (D, H, L, P) of the pattern of MCA-6H63 (green) staining across rats with a C4 unilateral contusion. Far rostral (A–D approximately spinal level C1-C2) and rostral (E–H; approximately spinal level C3-C4) to the lesion staining was most robust in the ipsilesional ventrolateral white matter (B, F) and the dorsal column (C, G). Composite drawings depict the typical pattern of MCA-6H63 staining in the far rostral (D) and rostral (H) spinal cord. Caudal (I–L approximately spinal level C6-C7) and far caudal (M–P approximately spinal level C8-T1) to the lesion staining was most robust in the lateral (J, N) and ventral (K, O) white matter. Composite drawings depict the typical pattern of MCA-6H63 staining in the caudal (L) and rostral (P) spinal cord. Scale bars = 200 μm.

MCA-6H63 positive fibers were also observed in the dorsal (Fig. 3A, B, E) and ventral (Fig. 3A, C, D, E) medulla. Staining in the medulla was sparser than in the cervical spinal cord. The most abundant staining in the medulla was in the ipsilesional ventral medulla (n = 6/6). Sparse contralesional staining was also observed in the ventral medulla (n = 6/6; Fig. 3C). Sparse immunopositive fibers in the ipsilesional dorsal medulla (Fig. 3B) were observed in all six rats although were more rarely observed in one of those rats. Contralesional staining in the dorsal medulla was rare but was observed at least intermittently, in all 6 rats. The qualitative relative differences in the amount of medulla MCA-6H63 labeling between rats could not be explained based on the size or extent of the lesion shown on cresyl violet analysis.

Fig. 3. MCA-6H63 staining in the caudal medulla.

A Example image of MCA-6H63 immunopositive staining in the caudal medulla 10 days following C4 contusion. Asterisk indicates artifactual staining caused by debris trapped beneath the coverslip. Immunopositive fibers were observed in the ipsilesional dorsal (B), ipsilesional ventral (C), and contralesional ventral (D) medulla. Rare and sparse staining was also observed in the contralesional dorsal medulla (high magnification not shown). E Composite drawing of the MCA-6H63 staining superimposed on an anatomical atlas [14].

DISCUSSION

We recently described a panel of anti-NF-L-Coil2 antibodies, including the MCA-6H63 antibody studied here, which were based on the UMAN-type antibodies used in peripheral neurofilament biomarker assays. These antibodies appear to specifically label degenerating axons [8]. Building upon that work, the current paper accomplishes three major goals. First, we reproduced and expanded upon the validation of an anti-NF-L-Coil2 degeneration-specific antibody, MCA-6H63, in the setting of SCI. Second, we compared the pattern of MCA-6H63 and anti-APP antibody staining which demonstrated that the majority of immunopositive fibers were either colabeled or solely labeled with APP while a minority of the fibers were solely labeled with MCA-6H63. Third, the current paper described the extent of axonal degeneration using the MCA-6H63 antibody at 10 days post-injury and showed abundant degenerating fibers present at this time point. Collectively, the data presented here show that the MCA-6H63 antibody effectively labels degenerating axons at 10 days following SCI, and accordingly will be useful for studying the time course of axonal degeneration and/or the impact of therapeutics.

Existing methods for identifying degenerating axons

Immunohistochemical markers for axons undergoing active degeneration after primary and secondary injury provide a tool for studying the time course of degeneration and for testing the efficacy of neuroprotective interventions. The time course of axonal degeneration is fundamentally important when determining when neuroprotective interventions could or should be initiated. However, traditional methods of labeling actively degenerating axons such as the Marchi method [16], modified silver stain [17], fluoro jade [18], and antibodies against APP [19, 20] work through different mechanisms and have limitations that have impaired reliable and reproducible study of actively degenerating axons. The Marchi method uses osmium tetroxide and an oxidizing agent to stain breakdown products of myelin but is finicky and prone to artifact [16, 21]. Additionally, this method of staining is only useful in myelinated neurons and the timing of myelin breakdown does not mimic axonal pathology [22]. The modified silver stains use pretreatments to reduce the argyrophilic nature of normal neuronal structures so that only degenerating neurons and their processes are impregnated with silver. However, these protocols suffered from the presence of artifactual or background staining, a tug of war between specificity and sensitivity, the complexity and fragility of the protocols, and difficulty staining degenerating axons at early time points after axonal injury [23–27]. Fluoro jade staining was proposed as a simpler, less capricious, and faster method of labeling degenerating axons as compared to the silver stains [28]. However, fluoro jade does not appear to be an effective tool in the identification of degenerating axons after SCI due to a lack of specificity and sensitivity post-injury with no significant difference in staining from uninjured spinal cords [29–31].

Antibodies against APP are considered the current gold standard in research and clinical autopsies for identifying damaged axons, particularly after diffuse axonal injury in TBI. The APP antibody has also been used to assess the extent of axonal injury after SCI and other neurodegenerative diseases that have disruptions in axonal transport caused by axonal injury. APP is transported from the soma down the axon via rapid anterograde transport. As a result, when an axon is injured and transport is disrupted, APP rapidly accumulates in axonal spheroids and retraction bulbs [19, 20]. However, antibodies against APP only label the portion of the axon where transport is disrupted leading to the accumulation of transported molecules, not the entire degenerating portion of the axon [20]. It is also unclear whether APP is a reliable marker of Wallerian degeneration which occurs in a portion of the axon disconnected from the cell body where APP is made. Recently, APP accumulation was also found in oligodendrocytes after TBI in mice, preventing the assumption that APP-positive fibers are degenerating axons [32].

The MCA-6H63 antibody

In the current study, robust MCA-6H63 staining was present rostral and caudal to the lesion in axonal tracts which would be expected to be disrupted by the unilateral C4 contusion. Fibers that were immunopositive for MCA-6H63 had a swollen and irregularly shaped punctate appearance consistent with pathological axons in cross-section. Immunopositive axons in ventral roots demonstrated the beaded appearance indicative of an axon undergoing Wallerian degeneration [2] whereas the MCA-6H63 negative, NF-L-ct positive axons had a healthier appearance with a more regular shape and contour. Axons labeled with MCA-6H63 typically did not co-label with NF-L-ct consistent with our prior report [8]. The presence of axonal pathology and loss of NF-L-ct staining confirms that the MCA-6H63 positive axons in the spinal cord and ventral roots were degenerating. The absence of MCA-6H63 staining in tissues from spinal-intact (no injury) rats supports the specificity of the MCA-6H63 antibody for degenerating axons. Additionally, immunopositivity was not present in injured spinal cord cross sections rostral or caudal to the lesion in sections incubated with concentration matched, isotype control Ig instead of primary antibody. This supports that MCA-6H63 staining was not non-specific Ig binding, autofluorescence, or artifact.

Contrasting APP and MCA-6H63 staining

As mentioned earlier, APP antibodies can be considered the “gold standard” for detecting axons with disturbed axonal transport and axonal injury, and accordingly, we contrasted staining with MCA-6H63. Colocalization of APP and MCA-6H63 staining was clearly identified in approximately 45% of fibers per region of interest. However, we observed that approximately 40% of fibers were labeled only for APP and 15% were labeled only for MCA6H63. APP, a marker of disturbed axonal transport, and MCA-6H63, a marker of neurofilament cytoskeletal degradation, are indicative of different pathological processes, which may or may not always cooccur. Prior work investigating the colocalization of APP staining and a marker for neurofilament side arm compaction has revealed little overlap [33, 34]. Neurofilament side arm compaction, which occurs as a result of axonal injury and makes the neurofilament cytoskeleton more likely to degrade, can be labeled using the RMO14 antibody [34]. Studies investigating RMO14 and APP staining after traumatic axonal injury have found limited colocalization [34–38]. Additionally, therapeutics capable of reducing APP staining did not reduce RMO14 staining [38, 39]. Collectively, these studies suggested that axonal transport dysfunction and cytoskeletal damage are independent pathological processes. In the current study, we found a much higher percentage of colocalized fibers but, in accordance with previous research, found a large percentage of solely APP-positive axons with a minority of solely MCA6H63-positive axons. A portion of the solely APP-positive staining may represent oligodendrocytes, not axons [32]. Future dual staining studies will be necessary to confirm the identity of APP-positive, MCA-6H63 negative fibers.

We expected a portion of axons to be labeled solely with MCA-6H63. The ability of APP to label the portion of the axon undergoing Wallerian degeneration is unclear. Several studies in TBI and stroke models have not observed evidence of APP accumulations in axons with pathology indicative of Wallerian degeneration [32, 40, 41]. In the current study, APP-positive axons were found in the same regions of white matter as MCA-6H63 positive axons, and, as in Li et al. [42], were found in both tracts expected to undergo Wallerian degeneration and axonal dieback. The percentage of fibers labeled with both APP and MCA-6H63 did not differ depending on the presumed mechanism of degeneration (Wallerian versus axonal dieback). If APP was not labeling axons undergoing Wallerian degeneration, we would have expected a lower percentage of dual labeled fibers in tracts presumably undergoing Wallerian degeneration, such as the rostral dorsal fasciculi and rostral ventrolateral white matter which are the putative locations of ascending sensory, spinothalamic, and spinocerebellar tracts [14, 15]. Because APP accumulates in the retraction bulb and axonal spheroids [43], whereas neurofilament pathology is not confined to these regions of the axon [34], a cross-section of an MCA-6H63 positive axon may have APP accumulation in a different portion of the axon. There is currently no perfect method to identify degenerating axons. The APP antibody is largely considered the best marker of axonal injury and transport dysfunction which often, but not always, coincides with axonal degeneration. The large percentage of fibers which were colabeled with MCA-6H63 and APP suggests promise for MCA-6H63’s ability to label degenerating axons. Because neurofilament is specific to neurons, it is unlikely that the MCA-6H63 antibody is labeling other cell types, which is a common limitation of other mechanisms of identifying degenerating axons such as fluoro jade and APP. However, extensive dual-labeling studies could further confirm the specificity of the antibody. Additionally, time course studies or in vivo imaging experiments aimed at improving understanding of axonal transport dysfunction and neurofilament degradation may elucidate how these processes occur during axonal degeneration and occasions in which they may not cooccur. These studies may also help confirm the sensitivity of MCA-6H63 and similar antibodies as a marker for axonal degeneration.

Pattern of degeneration at 10 days post cervical contusion injury

Continued degeneration observed even 10 days post-injury highlights the continued importance of neuroprotective therapeutics even in the subacute period after injury. There were qualitative differences in the amount of MCA-6H63 positive fibers between rats. Rats without significant extension of the lesion to the contralesional dorsal and ventral white matter visualized on cresyl violet analysis did not have substantial contralesional MCA-6H63 staining in the dorsal or ventral white matter. There was also variability in the amount of MCA-6H63 positive fibers which could not be explained by differences in lesion volume or extension of the lesion contralesionally. Specifically, while MCA-6H63 positive fibers were observed in the ipsilesional dorsal medulla in all rats, these fibers were relatively rare in one of these rats. This variability in MCA-6H63 staining may be occurring as a result of expected differences in the extent of spinal cord pathology across experimental animals. The amount of axonal degeneration is known to correlate with lesion severity [1, 9]. If the differences in MCA-6H63 staining are a product of variations in lesion severity, this serves to emphasize the promise of this antibody in providing an outcome measure for studying neuroprotective strategies and their ability to reduce the amount of axonal degeneration.

MCA-6H63-positive fibers were generally most abundant in axonal tracts likely undergoing Wallerian degeneration. For example, rostral to the lesion, MCA-6H63 positive fibers were abundant in the ventrolateral white matter, the putative locations of the spinothalamic tract and ventral spinocerebellar tract, and the dorsal fasciculi. These tracts are ascending tracts that carry sensory information to the brain or cerebellum [14, 15]. MCA-6H63 positive fibers in putative ascending sensory and spinocerebellar tracts were also visualized in the medulla [14]. Wallerian degeneration causes the eventual loss of the entire distal portion of the axon and is thus a farther-reaching process than axonal dieback [2]. Interestingly, staining in the dorsal corticospinal tract was largely absent caudal to the lesion where Wallerian degeneration would be expected to occur. Because MCA-6H63 fibers were found in the rostral dorsal corticospinal tract, it is unlikely that the lack of staining represents a failure to stain these types of axons. Additionally, the presence of rostral dorsal corticospinal tract staining and the significant impact of the lesion on the dorsal corticospinal tract, confirmed by lesion analysis using cresyl violet, indicates that the lack of staining is not a result of the lesion model not affecting the dorsal column. It is possible that Wallerian degeneration in the dorsal corticospinal tracts occurs at an earlier time point than 10 days post-injury, a finding which would have implications on the necessary timing of neuroprotective interventions aimed at improving motor function. A time course is necessary to determine differences in the pattern and extent of axonal degeneration over time. Caudal to the lesion, MCA-6H63 positive fibers were abundant in the lateral and ventral white matter in the location of the putative rubrospinal, reticulospinal, and vestibulospinal tracts. These descending tracts would presumably undergo Wallerian degeneration caudal to the SCI. However, ascending tracts such as the dorsal spinocerebellar and ventral spinothalamic tracts mix and overlap with the descending tracts in these regions [14], preventing assumptions regarding whether the MCA-6H63 positive axons in these regions are undergoing axonal dieback or Wallerian degeneration. Further axonal tracing studies would be necessary to definitively confirm the identity of these axons. However, given the presence of MCA-6H63 positive fibers in both the rostral dorsal corticospinal tract and dorsal fasciculi, which are relatively homogonous and anatomically distinct, we can conclude that the MCA-6H63 antibody is labeling fibers undergoing both Wallerian degeneration and axonal dieback.

Staining was also visualized in white matter tracts likely undergoing axonal dieback. Rostral to lesion there were MCA-6H63 immunopostive axons in the dorsal corticospinal tract. Caudal to the lesion, MCA-6H63 staining was present in the ventrolateral white matter in the putative spinothalamic tract and spinocerebellar tracts. However, staining in the dorsal fasciculi caudal to the lesion was rare and when observed was very sparse and in segments closer to the lesion. There is evidence that by one-week post-injury, proximal portions of ascending sensory axons in the dorsal white matter stabilize and discontinue axonal dieback whereas dieback in the corticospinal tracts may continue for a longer duration of time and more extensive distance [44]. Although axonal regeneration typically fails endogenously in the central nervous system, one area of SCI research focuses on improving axonal regeneration. Understanding the timing and extent of axonal dieback in different white matter tracts and the discovery of neuroprotective mechanisms to ameliorate dieback could assist in improving regenerative ability by decreasing the distance in which axons would need to regrow.

CONCLUSION

We have continued to validate the MCA-6H63 antibody as a marker of degenerating axons after SCI, compared the pattern of the MCA-6H63 antibody with the well-established APP antibody, and described the pattern of axonal degeneration using the MCA-6H63 antibody at 10 days post-injury. Collectively these data support that the MCA-6H63 antibody labels degenerating axons following SCI. In APP and MCA-6H63 colocalization analysis, we found the majority of immunopositive fibers were either solely labeled with APP or colocalized with MCA-6H63 and APP while a minority of fibers were labeled solely with MCA-6H63. Although future research is necessary to determine why a large percentage of fibers are APP positive and MCA-6H63 negative, the larger percentage of colocalized fibers supports the utility of MCA-6H63 as a marker for degenerating axons. The MCA-6H63 antibody revealed extensive ongoing degeneration 10 days following SCI and suggests that the MCA-6H63 antibody provides a useful tool for studying the time course of axonal degeneration as well as the impact of therapeutics, particularly those aimed at ameliorating secondary degeneration after neural injury.

DATA AVAILABILITY

The data supporting the described findings can be made available from the corresponding author upon reasonable request.