Differential Glycosylation Expression in Injured Rat Spinal Cord Treated with Immunosuppressive Drug Cyclosporin-A.

Glycosylation is ubiquitous throughout the central nervous system and altered following spinal cord injury (SCI). The glial scar that forms following SCI is composed of several chondroitin sulfate proteoglycans, which inhibit axonal regrowth. Cyclosporin-A (CsA), an immunosuppressive therapeutic, has been proposed as a potential treatment after SCI. We investigated CsA treatment in the spinal cord of healthy, contusion injured, and injured CsA-treated rats. Lectin histochemistry using fluorescently labeled lectins, SBA, MAA, SNA-I, and WFA, was performed to identify the terminal carbohydrate residues of glycoconjugates within the spinal cord. SBA staining decreased in gray and white matter following spinal cord injury, whereas staining was increased at the lesion site in CsA-treated animals, indicating an increase in galactose and N-acetylgalactosamine terminal structures. No significant changes in MAA were observed. WFA staining was abundant in gray matter and observed to increase at the lesion site, in agreement with increased expression of chondroitin sulfate proteoglycans. SNA-I-stained blood vessels in all spinal cord regions and dual staining identified a subpopulation of astrocytes in the lesion site, which expressed α-(2,6)-sialic acid. Glycosylation were altered in injured spinal cord treated with CsA, indicating that glycosylation and alteration of particular carbohydrate structures are important factors to consider in the examination of the environment of the spinal cord after injury.

Introduction

In spinal cord injury (SCI), the primary mechanical trauma is caused by direct physical compression of the spinal cord by fractured and displaced bone fragments. Blood vessels are crushed, causing microhemorrhages, neural cell membranes are broken, and axons are damaged, leading to a loss of functional connections.[1,2] Secondary injury adds to the complexity of SCI. Neuronal destruction causes the release of toxic chemicals that attack the neighboring tissue via excitotoxicity and often leads to a wave of apoptosis and free radical-induced lipid oxidation.[3,4] Inflammation plays an important role in early and late stages of SCI.[5,6] Cells of the immune system migrate from the periphery through the damaged blood–brain barrier and join the resident glial cells within the spinal cord. A fluid-filled cyst forms at the site of injury. Many astrocytes become hypertrophic and adopt a reactive phenotype.[7] Reactive astrocytes begin to secrete inhibitory extracellular matrix (ECM) molecules, including several chondroitin sulfate proteoglycans (CSPGs), and migrate to the lesion site to build a barrier known as the glial scar,[8] which produces a physical and chemical barrier and stops damaged and severed axons from traversing the site of injury to unscarred spinal regions where they could potentially regenerate and reconnect functional spinal circuits.[9,10] The molecular organization of the scar and the production of inhibitory molecules by astrocytes are contributing factors for regenerative failure after SCI.[7,8,11]

Glycosylation is ubiquitous throughout the central nervous system (CNS) and plays critical roles in development and normal cellular function. Sialic acids are charged residues that often terminate mammalian cell surface structures, ECM glycoconjugates, and glycolipids. Sialic acids fulfil important functions in the CNS as critical components of gangliosides and the repeating units of polysialic acid.[12] Terminal galactose (Gal) residues interact with galectins (receptors found throughout the cell) on cell surfaces and in the ECM, and exert biological effects. For example, galectin-1 is expressed by reactive astrocytes and after SCI appears to have a neuroprotective effect mediated by reactive astrocytes.[13] Variously sulfated N-acetylgalactosamine (GalNAc) residues comprise one of the two residues of the repeating units of chondroitin sulfate (CS) chains on CSPGs, with the structures and sulfation pattern of the CS repeating units determining the function. For example, chondroitin-4,6-sulfate (also known as CS E) is potently inhibitory to axonal growth.[14] Previously, we have shown in vitro that neuronal glycosylation in an injured environment returns toward normal with chondroitinase ABC treatment.[15] Thus, a return to healthy ECM and cellular glycosylation after repair would coincide with resumed normal function in the spinal cord.

Many therapeutic strategies have been proposed in the treatment of SCI.[16] Cyclosporin-A (CsA) is an immunosuppressive cyclic peptide that has been shown to have neuroprotective properties by inhibiting calcineurin and the mitrochondrial permeability transition pore, reducing lipid peroxidation, and releasing neurotrophic factors.[17] Additionally, it is known to induce growth-associated protein-43 (GAP-43) expression, which is involved in process extension of neurons.[18] Treatment with CsA leads to improved functional recovery after SCI in rats, and it is frequently combined with neural transplantation to avoid tissue rejection.[19,20] However, the effect of CsA treatment on tissue and cell glycosylation is unknown.

In this study, we investigated glycosylation in vitro using normal (primary) astrocytes and an astrocyte cell line, Neu7, that overexpresses CSPGs. We also examined glycosylation in vivo within uninjured rat spinal cord, injured spinal cord, and injured spinal cord treated with CsA. In particular, we selected a panel of lectins, carbohydrate-binding proteins, to elucidate the presence, localization, and potential alteration of structures containing sialic acid, Gal, and GalNAc residues, including CS, in the tissues of the CsA treatment compared to normal and no treatment. To clarify potential interactions and functions of the altered glycosylation expression and their roles in the documented effects of CsA treatment after SCI, we also explored the relationship between axonal growth, CS expression, astrocytes, and sialylation expression using a combination of lectin histochemistry and immunohistochemical strategies.

Results and Discussion

Cells are coated with a layer of carbohydrates of complex structure, which facilitates the interaction of the cells with other cells and with their environment. Changes in carbohydrate expression can be examined in different tissues (e.g., injured versus uninjured spinal cord tissue) using lectins. Lectin histochemistry for both cells and tissues was also carried out in the presence of haptenic sugars. A reduction in intensity of lectin staining was observed for all lectin staining (not shown), which confirmed the carbohydrate-mediated binding of lectins.[21] This study encompasses a multidisciplinary chemical biology approach to studying tissue regeneration after SCI.

Lectin Staining of Astrocytes in Vitro

To initially screen for differences in carbohydrate expression related to CSPGs and sialylation, protein extracts from two types of astrocytes, primary and Neu7, were made, total cell lysates and extracts enriched for membrane (hydrophobic) and cytosolic (hydrophilic) proteins. Neu7 cells overproduce the inhibitory CSPGs including NG2, the CS-56 antigen, and versican and is used to simulate a reactive astrocyte phenotype.[22] These protein preparations were fluorescently labeled and incubated on lectin microarrays presenting a panel of five lectins with specificity for terminal Gal and GalNAc residues (SBA and WFA) and sialic acid (MAL-I, MAL-II, and SNA-I) (Figure ).

Figure 1

Intensity of primary astrocyte and Neu7 cell protein extracts binding to lectins on microarray. Bar chart representing the differences in binding of fluorescently labeled protein extracts to printed lectins on a microarray surface where Ast_Lys is the total lysate from primary astrocytes, Ast_Cyt is the cytosolic (hydrophilic) protein-enriched extract from primary astrocytes, Ast_Mem is the membrane (hydrophobic) protein-enriched extract from primary astrocytes, Neu7_Lys is the total lysate from Neu7 cells, Neu7_Cyt is the cytosolic (hydrophilic) protein-enriched extract from Neu7 cells, and Neu7_Mem is the membrane (hydrophobic) protein-enriched extract from Neu7 cells. Graph represents the mean of three replicate experiments (except for duplicate experiments for Neu7_Mem binding to MAL-II), with each experiment the median of six individual replicates. Error bars represent ±1 standard deviation of the mean of the three experiments.

From comparison of astrocyte and Neu7 cell lysates, it is clear that there was an increased expression of Gal and GalNAc terminal residues in Neu7 cell lysates as indicated by increased SBA and WFA binding compared to primary astrocytes (Table ). This is in agreement with the overproduction of CSPGs by Neu7. Similarly, MAL-I and MAL-II binding to Neu7 cell lysates was increased compared to primary astrocyte lysates, which indicated an increased expression of α-(2,3)-linked sialic acid (Table ). However, binding to SNA-I was similar for both cell lysates, which indicated a similar expression of α-(2,6)-linked sialic acid (Table ). Relative lectin-binding intensities of the cells’ cytosolic protein-enriched extracts were largely in agreement with the cells’ lysates, but the lectin-binding intensities of the membrane protein-enriched extracts differed (Figure ). Binding to MAL-I was increased for Neu7 cells compared to primary astrocytes, whereas binding intensities to MAL-II was similar for the membrane protein-enriched extracts for both cell types, indicating a potentially modest increase of α-(2,3)-linked sialic acid expression in Neu7 cells compared to primary astrocytes. On the other hand, the binding intensity to SNA-I was greater for the membrane protein-enriched extracts of primary astrocytes compared to Neu7 cells.

Table 1

Lectins, Their Origin, Their Corresponding Carbohydrate Binding Specificity and Haptenic Sugars

abbreviation	lectin	origin	binding specificity	haptenic sugar
MAL-I	Maackia amurensis leukoagglutinin (MAL)	Maackia amurensis	Neu-α-(2 → 3)-Gal-β-(1 → 4)-GlcNAc-R, SO42–-3-Gal-β-(1 → 4)-GlcNAc-R	Lac
MAL-II	Maackia amurensis hemagglutinin (MAH)	Maackia amurensis	Neu-α-(2 → 3)-Gal-β-(1 → 3)- (±Neu-α-(2 → 6)-GalNAc), SO42–-3-Gal-β-(1 → 3)-(±Neu-α-(2 → 6)-GalNAc), SO42–-3-Gal-β-(1 → R)	Lac
SBA	soybean agglutinin	Glycine max	terminal α- or β-linked GalNAc, Gal	Gal
MAA	Maackia amurensis agglutinin	Maackia amurensis	Neu-α-(2 → 3)-Gal-β-(1 → 4)-GlcNAc-R, SO42–-3-Gal-β-(1 → 4)-GlcNAc-R	Lac
SNA-I	Sambucus nigra isolectin-I	Sambucus nigra (elderberry)	Neu-α-(2 → 6)-Gal(NAc)-R	Lac
WFA	Wisteria floribunda agglutinin	Wisteria floribunda (Japanese wisteria)	terminal α- or β-linked GalNAc, lactose, Gal, chondroitin sulfate	Gal

For histochemical inhibition experiments, lectins were co-incubated with 100 mM of the haptenic sugar and for lectin microarray printing, lectins were co-incubated in 1 mM of the haptenic sugar.

The enrichment process used here for cytosolic and membrane protein fractionation is based on relative hydrophobicity and hydrophilicity. Therefore, cell surface bound hydrophilic proteins such as CSPGs and mucins can be extracted to the “cytosolic” fractions, so the protein membrane fractionation in this case can only be taken as an enrichment and may not be truly completely representative of the cell surface protein and proteoglycan population.

Thus, the two cell types were then assessed for more specific surface glycosylation changes by cytohistochemisty on the intact cells using the fluorescein isothiocyanate (FITC)-conjugated lectins SBA, MAA, WFA, and SNA-I (Table ). The Maackia amurensis agglutinin (MAA) contains both MAL-I and MAL-II lectins and, as both have binding specificity for terminal α-(2,3)-linked sialic acid, MAA was used in place of MAL-I and -II for histochemistry experiments.[23] Although the differences in secreted CSPGs between primary astrocytes and Neu7 cells have been characterized, to our knowledge, the cell surface glycosylation has not been previously profiled. Lectin histochemistry revealed a greater expression of terminal GalNAc (SBA staining) and/or Gal residues and α-(2,3)-linked sialylation (MAA staining) on Neu7 cells compared to primary astrocytes (Figure A–E). The greater SBA and MAA binding of Neu7 cells compared to primary astrocytes was in agreement with the findings from the lectin microarray profiling of the cell protein extracts. However, there was equivalent expression of α-(2,6)-linked sialic acid on primary astroctyes and Neu7 cells as indicated by SNA-I binding (Figure A,F,G), which was in agreement with the SNA-I binding of cell lysates on the lectin microarray. WFA binding in vitro was the same in primary astrocytes and Neu7 cells (Figure A,H,I), in contrast to the findings of the protein extracts on the lectin microarray. However, as has been noted above, protein extractions are not completely representative of the molecules actually present on the cell surface, so the cytochemistry observations are more indicative of the cell surface expression. It is notable that the lectins SBA and WFA did not have the same binding pattern to Neu7 cells and primary astrocytes, which indicated that the lectins favored binding to different carbohydrate structures or presentations. Both SBA and WFA have been previously characterized as having similar binding specificities and affinities for terminal α- and β-linked GalNAc and Gal residues.[24] Although it is known that WFA additionally binds to CS and is frequently used as a histochemical marker for perineuronal nets (PNNs), the exact target structure(s) and sulfation pattern(s) to which this lectin binds in CS is not currently known.[25,26] Thus, it is likely that the additional structures recognized by WFA on the primary astrocytes cell surface are components of CS. Expression of α-(2,6)-linked sialic acid is greater compared to α-(2,3)-linked sialic acid on the primary astrocyte surface.[27] Apart from α-(2,8)-linked polysialic acid, α-(2,3)-linked sialic acid is typically predominant in the nervous system, and there is very little α-(2,6)-sialylation.[27] The presence of α-(2,6)-sialylation on the astrocyte cell surface may be a characteristic of this cell type or the cell type under certain conditions, such as in culture.

Figure 2

Intensity of lectin staining in primary astrocytes and Neu7 astrocytes in vitro. Graph shows average intensity of SBA, MAA, SNA-I, and WFA in primary astrocytes and Neu7 astrocytes (A). Mean ± standard error of the mean (SEM). *P < 0.05. Photomicrographs show SBA (B, C), MAA (D, E), SNA-I (F, G), and WFA (H, I) lectin staining in primary astrocytes and Neu7 astrocytes, respectively. Scale bar = 30 μm.

Lectin Staining of Spinal Cord Cryosections

Lectin histochemistry of the spinal cord tissue from the three animal groups, uninjured, injured, and injured treated with CsA, were examined. The gray and white matter of the uninjured group had a higher intensity of SBA binding overall compared to the same two regions in the injured and CsA-treated groups (Figure A–E,G,H), which indicated a decreased expression of nonsulfated terminal Gal and/or GalNAc residues in the injured and treated tissues compared to healthy tissues. In addition, the SBA-binding intensity of the healthy gray matter was approximately 3 times that of the uninjured white matter. At the lesion site, a slight increase in SBA intensity was observed in the gray matter and lesion site of the CsA-treated group compared to the injured group (Figure A,F,I). Lozza et al. (2009) showed that SBA stains neuronal cell bodies and not glial cells in young and aged rat spinal cords.[28] Although the binding of SBA was significantly higher in Neu7 astrocytes in vitro compared to normal primary astrocytes, the same binding profile for this lectin was not observed in the injured spinal cord tissue compared to uninjured. This finding highlights the importance of cautious interpretation when relying on in vitro data alone for glycosylation analysis.

Figure 3

Integrated density of SBA lectin staining in spinal cord slices. Graph shows integrated density of SBA lectin staining in the white matter, gray matter, and lesion site of uninjured, injured, and injured spinal cords treated with CsA (A). Mean ± SEM. *P < 0.05. Photomicrographs show SBA histochemical staining in the white matter (B, D, G), gray matter (C, E, H), and lesion site (F, I) of uninjured, injured, and injured spinal cords treated with CsA. Scale bar = 50 μm.

MAA lectin staining exhibited no significant change in intensity in the gray and white matter of the injured and CsA-treated group compared to the uninjured group (Figure A–E,G,H), and no overall difference in MAA intensity was observed between the lesions sites of the injured and CsA-treated groups (Figure A,F,I). However, consistent with the greater MAA binding to Neu7 cells compared to primary astrocytes, there was an observed trend of slightly increased MAA binding to the injured and CsA-treated tissue overall compared to uninjured, with the intensity of gray matter and lesion site of CsA-treated tissues approximately twice that of the same areas in injured tissues. There is an increased expression of sialylation, sulfation, and fucosylation in tissue and on cell surfaces overall associated with inflammation, which facilitates interactions with selectins, although detailed studies of structural changes over time have not been undertaken in neural tissue.[29] Previously, it has been shown that neurite outgrowth from cerebellar neurons was enhanced by the presence of α-(2,3)-sialylation on glial CD24 via lectin-like binding to the L1 adhesion molecule.[30]

Figure 4

Integrated density of MAA lectin in spinal cord slices. Graph shows integrated density of MAA lectin staining in the white matter, gray matter, and lesion site sampling regions within uninjured, injured, and injured spinal cords treated with CsA (A). Mean ± SEM. *P < 0.05. Photomicrographs show MAA histochemical staining in the white matter (B, D, G), gray matter (C, E, H), and lesion site (F, I) of uninjured, injured, and injured spinal cords treated with CsA. Scale bar = 50 μm.

There was also no statistically significant difference in SNA-I staining intensity for the white and gray matter between groups (Figure A–E,G,H). Although an increase in intensity was observed at the lesion site of the CsA-treated group compared to the injured group, which indicated increased expression of α-(2,6)-linked sialic acid, this was also not a significant change (Figure A,F,I). The morphology of the cells in SNA-I-stained tissue in the gray and white matter of CsA-treated group appeared to resemble glial cells (particularly astrocytes). For this reason, dual SNA-I lectin and glial fibrillary acidic protein (GFAP) immunohistochemical staining was carried out on the injured and CsA-treated group to determine if co-localization could be observed between SNA-I and astrocytes. In the tissue, SNA-I bound to the endothelial cells of blood vessels in all regions examined, as expected,[31] and to only a particular subpopulation of astrocytes in the lesion of the injured (Figure A,B) and CsA-treated groups (Figure C,D) at the lesion site, which indicated that a subpopulation of astrocytes expressed α-(2,6)-sialylation. It is well known that GFAP expression increases throughout the lesion border in response to injury.[32,33] The number of astrocytes have been previously reported to increase at the lesion site of CsA-treated animals compared to untreated injured animals at 3 weeks after injury.[18] Previously in a study of rat brain injury, both MAA and SNA-I staining was increased in the injured tissue compared to sham-operated control, but neither lectin staining co-localized with astrocyte immunoreactivity. However, polysialic acid did co-localize with both microglia and astrocytes, whereas SNA-I and MAA both co-localized with Iba-1 immunoreactivity.[34] In our study, SNA-I also stained other glial cells aside from astrocytes within the lesion areas, as many stained cells were observed to surround those co-localized with GFAP. Since GFAP does not stain all of the cellular processes of astrocytes, this may be the reason for the SNA-I staining surrounding GFAP-positive astrocytes.[35] SNA-I also stained neurons in the lesion. The expression of the abnormal α-(2,6)-linked sialic acid motif has been reported on the neuronal cell surface in an in vitro injury model environment.[15] Reactive astrocytes are more abundant in the spinal cord lesion at later survival times of 7–28 days and mainly surround and enclose tissue with phagocytic macrophages and activated microglia.[36] Although we cannot confirm that the astrocyte subpopulation expressing α-(2,6)-linked sialylation are reactive astrocytes, this subpopulation was more frequent at the lesion border in the untreated tissue (Figure A), whereas this subpopulation was more dispersed in the CsA-treated lesion area (Figure B). In an effort to identify which structure on the astrocyte subpopulation surface may have been modified with α-(2,6)-sialic acid, dual staining by SNA-I and βI-integrin was carried out. There appeared to be some sparse evidence of co-localization observed (Figure ). Different glycosylation of receptors alters the signaling response induced in cells after binding to ligands and may serve to regulate signaling functions in vivo. Sialylation and galactosylation of the N-linked oligosaccharides on intercellular adhesion molecule-1 (ICAM-1) enhanced the signaling response in mouse astrocytes.[37] The expression of α-(2,6)-sialylation has been previously observed on necrotic and apoptotic cells.[38] The β1 integrin is expressed ubiquitously and can pair with at least 12 different α subunits to bind different ligands and induce a signaling response. Variable glycosylation of β1 integrin alters the binding and signaling responses[39] and α-(2,6)-sialylation of β1 integrin has been shown to block binding to Gal-3 and protect colon carcinoma cells against apoptosis.[40] However, in our study, we were unable to co-localize SNA-I and β1 integrin staining, suggesting that a different receptor or structure was modified on the astrocyte subpopulation expressing α-(2,6)-sialylation in the lesion.

Figure 5

Integrated density of SNA-I lectin in spinal cord slices. Graph shows integrated density of SNA-I lectin staining in the white matter, gray matter, and lesion site sampling regions within uninjured, injured, and injured spinal cords treated with CsA (A). Mean ± SEM. Photomicrographs show SNA-I histochemical staining in the white matter (B, D, G), gray matter (C, E, H), and lesion site (F, I) of uninjured, injured, and injured spinal cords treated with CsA. Scale bar = 50 μm.

Figure 6

Dual staining of astrocytes and SNA-I in injured spinal cord slices. Photomicrographs show GFAP (red), SNA-I (green), and 4′,6-diamidino-2-phenylindole (DAPI) (blue) staining in the lesion site of spinal cord slices of injured and CsA-treated animals (A, C). Scale bar = 50 μm. The boxed areas in (A) and (C) are magnified in (B) and (D), respectively. Scale bar = 20 μm.

Figure 7

Dual staining of βI-integrin and SNA-I in injured spinal cord slices. Photomicrographs show βI-integrin (red) and SNA-I (green) staining in the lesion site of spinal cord slices of injured and CsA-treated animals (A, C). Scale bar = 50 μm. The boxed areas in (A) and (C) are magnified in (B) and (D), respectively. Scale bar = 20 μm.

Although similar binding of WFA was observed in vitro between normal astrocytes and Neu7 cells in our study, WFA lectin staining was significantly increased in the gray matter compared to the white matter of the uninjured tissues, at approximately 10 times greater magnitude. The uninjured gray matter was significantly greater than the gray matter in both the injured and CsA-treated animal groups, whereas the WFA-stained uninjured white matter was significantly less compared to the white matter in the injured and CsA-treated animal group (Figures A–E and 9). A significant increase in WFA binding was also observed at the lesion site compared to gray and white matter in both the injured and CsA-treated groups, with less binding in the gray matter and lesion site of the CsA-treated group compared to the injured group (Figure A,D–I). Increased expression of CSPGs has been shown within perineuronal nets (PNNs) in CNS-lesioned tissue.[41] In a conditional Sox9 knockout, reduction of CSPGs within PNNs has been shown.[42] The CSPGs are also present in the white matter of the spinal cord.[43] After SCI, these CSPGs are elevated and cause the inhibition of axonal growth.[44] Previously, ChABC was injected into the spinal cord near the lesion, resulting in degradation of the CSPGs, which promoted the sprouting of the injured dorsal column axons and functional recovery and also reduced the WFA staining of PNNs.[45] In our study, WFA staining in the lesion site of the CsA-treated tissue was higher in comparison to the gray and white matter. WFA staining at the injury was much lower in the CsA-treated group compared to the injured group with no treatment, indicating that CsA may have an effect on CS expression. WFA staining is used to stain PNNs, which wrap around neuronal cell bodies or neurons, as well as other CSPGs present.[25] Dual WFA lectin and GAP-43 immunohistochemical staining was carried out on the injured (Figure A,B) and CsA-treated groups (Figure C,D) at the lesion site to assess the relationship between neurons and WFA. Although both markers stained adjacent and closely related cells, no co-localization of WFA and GAP-43 was observed. The dual lectin and immunohistochemistry images indicated tissue remodeling in the lesion of the CsA-treated tissue compared to injured tissue (Figures –10), supporting the idea that CsA treatment reduced the formation of the glial scar at the lesion 3 weeks after injury.

Figure 8

Integrated density of WFA lectin in spinal cord slices. Graph shows integrated density of WFA lectin staining in the white matter, gray matter, and lesion site sampling regions within uninjured, injured, and injured spinal cords treated with CsA (A). Mean ± SEM. *P < 0.05. Photomicrographs show WFA histochemical staining in the white matter (B, D, G), gray matter (C, E, H), and lesion site (F, I) of uninjured, injured, and injured spinal cords treated with CsA, respectively. Scale bar = 50 μm.

Figure 9

Dual staining of GAP-43 and WFA in injured spinal cord slices. Photomicrographs show GAP-43 (red), WFA (green), and DAPI (blue) staining in the lesion site of spinal cords of injured and CsA-treated animals (A, C). Scale bar = 50 μm. The boxed area in (A) and (C) are magnified in (B) and (D), respectively. Scale bar = 20 μm.

Figure 10

Dual staining of neurogranin and WFA in injured spinal cord slices. Photomicrographs show neurogranin (red) and WFA (green) staining in the lesion site of spinal cords in the injured and CsA-treated animals (A, D). Scale bar = 50 μm. (A) An overview of white and gray matter. (B, C) White and gray matter in injured spinal cord slice, respectively. The boxed area from CsA-treated animal in (D) is magnified in (E). Scale bar (B), (C), and (E) = 20 μm.

As the removal of PNNs is associated with a return to plasticity of the neurons and functional recovery, both CsA-treated and injured tissue groups were stained for the presence of neurogranin.[46] Neurogranin (RC3, BICKS) is a neuron-specific calmodulin-binding protein kinase C substrate found in the neuronal cell body, dendrites, and axons and can be associated with neuronal plasticity.[47] Although dual staining with WFA and antineurogranin antibody did not reveal any co-localization in the injured tissue group (Figure A–C), there did appear to be some co-localization in the CsA group (Figure D,E), providing further evidence of the beneficial treatment with CsA.

Many studies have been carried out where CsA administration caused functional recovery. The neuroprotective effects of CsA have been well documented[17,18] and this drug inhibits oxidative free radicals and stabilizes the injury environment. The best functional improvement was seen in animals that had CsA administered 6 h after injury,[48] and CsA treatment has been shown to increase neuronal survival and inhibit demyelination when administered within a day after SCI.[17,19] When CsA was administered 4 days after injury, after the primary wave of injury had passed, functional improvement was also observed 3 weeks after injury.[18] However, CsA treatment does not always show functional recovery due to calcineurin present in the lesion, which inhibits CsA treatment.[49]

In this study, we have shown that glycosylation was altered in injured rat spinal cord with CsA treatment compared to injured tissue with no treatment. Although the overall tissue glycosylation did not return to normal conditions, CsA treatment from 4 days to 3 weeks after injury did appear to significantly lower the CSPG expression in the CsA-treated lesion compared to injured tissue and also indicated tissue remodeling. In addition, a subpopulation of astrocytes localized in the lesion expressed α-(2,6)-linked sialylation, and these were also present in the CsA-treated lesion tissue. These alterations in glycosylation may contribute to the functional improvements previously observed for CsA treatment. This study supports the inclusion of glycosylation, a critical component of the CNS, as an important aspect in future research on CNS injury and regeneration to provide a more complete perspective on successful molecular repair and remodeling strategies.

Experimental Section

Materials

Culture plastics were from BD Biosciences (San Jose, CA). Fluorescein isothiocyanate (FITC)-conjugated and unlabeled lectins (Table ) were purchased from EY Labs (San Mateo, CA) except for unlabeled MAL-I and MAL-II, which were from Vector Laboratories Inc. (Burlingame, CA). Bovine serum albumin (BSA, ≥99%), rhodamine-conjugated antirabbit IgG antibody, mouse monoclonal antigrowth-associated protein-43 (GAP-43) antibody, and 4′,6-diamidino-2-phenylindole (DAPI) were from Sigma-Aldrich Co. (Dublin, Ireland). The BSA was periodate-treated (pBSA)[54] and used for all histochemical blocking. The complete protease inhibitor cocktail, ethylenediaminetetraacetic acid (EDTA)-free (Cat. no. 11873580001) and PhosSTOP phosphatase inhibitor cocktail were from Roche, Inc. (Basel, Switzerland). The Pierce bicinchoninic acid (BCA) assay kit and Mem-PER Plus membrane protein extraction kit were from Thermo Fisher Scientific (Waltham, MA). The carboxylic acid succinimidyl ester Alexa Fluor 555 (AF555) fluorescent label, ProLong Gold antifade, and Alexa Fluor 594-conjugated goat antimouse IgG antibody were from Life Technologies (Grand Island, NY). Amicon Ultracel 3 kDa molecular weight cutoff (MWCO) centrifugal ultrafiltration units were supplied by Millipore (Cork, Ireland). The rabbit polyclonal antiglial fibrillary acidic protein (GFAP) antibody was from DakoCytomation (Dublin, Ireland), the rabbit polyclonal antineurogranin antibody from Abcam (Cambridge, U.K.), and the mouse monoclonal anti-βI-integrin antibody were from Chemicon International Inc. (Temecula, CA). Nexterion Slide H microarray slides were supplied by Schott AG (Mainz, Germany). The 8-well gasket slide and incubation cassette system was from Agilent Technologies Ireland, Ltd. (Cork, Ireland). All other reagents were from Sigma-Aldrich Co. (Dublin, Ireland) unless otherwise indicated, and were of the highest grade available.

Cell Culture

Primary cerebral astrocytes for immunocytochemistry were obtained from P2 Sprague–Dawley rat pups and were purified and cultured as previously described.[15,50] For protein extracts, primary astrocytes were prepared from spinal cords isolated from 3–4-day-old postnatal Sprague–Dawley rats. Spinal cords were isolated by minor modification of the “Ejection method”.[51] Briefly, after decapitation, a transverse cut was made at the lower lumbar region of the spine and spinal cords were flushed through the spinal canal using 19-gauge needle attached to a syringe filled with Hanks’ balanced salt solution. The spinal cords were transferred immediately to Dulbecco’s modified Eagle’s medium (DMEM) with high glucose supplemented with 1% penicillin and streptomycin (P/S) on ice. Spinal cords were transferred to a Petri dish containing the same media and meninges were gently peeled under microdissection microscope. Spinal tissues were chopped into fine (approximately 1 mm) pieces and digested using 1% trypsin–EDTA solution at 37 °C for 15–20 min. Trypsin activity was inhibited using DMEM with high glucose supplemented with 10% fetal bovine serum and 1% P/S and digested tissue was triturated by passing it through various sizes of needles (18–23 gauge). Digested tissue was expelled through a cell strainer filter (70 μm mesh size, Falcon) to eliminate clumped cells or undigested tissue.

Neu7 astrocytes (generously provided from Professor James Fawcett[52]) were cultured in DMEM supplemented with 10% horse serum, 1% l-glutamine, and 1% P/S at 37 °C in a 5% humidified CO2 atmosphere. Primary astrocytes and Neu7 astrocytes were seeded at a density of 10 000 cells onto sterile coverslips in a 12-well tray and grown for 4 days in vitro.

Spinal Cord Injury

Female Sprague–Dawley rats (Charles River UK Ltd, Margate, U.K.) weighing between 220 and 225 g were used in this study. The rats were housed with a 12 h light/dark cycle in a temperature-controlled room. Food and water were provided ad libitum. All animal experiments were carried out in accordance with the Council Directive 2010/63EU of the European Parliament. All housing and surgical procedures carried out in this study were approved by the Animal Care Research Ethics Committee at the National University of Ireland, Galway, and the Health and the Health Products Regulatory Authority. Nine female Sprague–Dawley rats were used in this study. Three animals were used as control (uninjured) rats and six rats received a SCI. Prior to surgery, the rats were weighted and preoperative analgesia, Buprenorphine (0.1–0.025 mg/kg, FortDodge Animal Health Ltd), was delivered intraperitoneally (IP). The rats were anesthetized by IP injection of ketamine and xylazine (100 and 10 mg/kg, respectively), following which a laminectomy was performed at T8–T10. The injured animals received a 200 kilodyne moderate contusion injury at T9 using an Infinite Horizon Impactor Device (Precision Systems and Instrumentation, Lexington, KY). The muscle and skin was sutured with absorbable suture material (Vicryl, 4 metric) and animals kept warm on a heated blanket until fully recovered from intervention. Each animal received a subcutaneous injection of 5–10 mg/kg Enrofloxacin (Baytril 5%, Bayer) antibiotic once daily for a minimum period of a week. Pain relief was provided by administering buprenorphine (Torbugesic, FortDodge Animal Health Ltd.) at 0.1–0.25 mg/kg twice daily for 7 days after surgery. Saline solution (3.5 mL) was administered subcutaneously for 3 days following surgery. Bladders were manually expressed twice daily from the day of injury.

CsA Administration

The six injured animals were randomly divided into two groups: CsA-treated group (n = 3) and control group (n = 3). A subcutaneous injection of 5 mg/kg CsA (Sandimmun, Sandoz) was administered 4 days after injury to the CsA-treated group and every day thereafter for the duration of the experiment.

Tissue Processing

Three weeks from the time of injury, all animals were deeply anesthetized by IP injection of sodium pentobarbital (50 mg/kg) and perfused transcardially with saline followed by 4% paraformaldehyde in 0.01 M phosphate-buffered saline, pH 7.4 (PBS). Spinal cords were dissected out, postfixed overnight with 4% paraformaldehyde, immersed in 30% sucrose overnight, and frozen in liquid nitrogen-chilled isopentane. Spinal cords were cryosectioned transversely at 20 μm thickness in a rostral to caudal direction.

Cell Protein Extractions and Fluorescent Labeling

Primary astrocytes and Neu7 cells (5 × 106 cells/mL) were lysed to make total protein lysate in a radioimmunoprecipitation assay (RIPA) buffer (50 mM Tris–HCl, 150 mM NaCl, 0.02% sodium azide, 0.1% sodium dodecyl sulfate, 1% Nonidet P-40, 0.5% sodium deoxycholate, pH 8.0) with complete protease inhibitor cocktail, EDTA-free (1:100), phenylmethylsulfonylfluoride (1:50), and PhosSTOP phosphatase inhibitor cocktail (1:10). The cells were also fractionated using the Mem-PER Plus membrane protein extraction kit (Thermo Fisher Scientific) into membrane (hydrophobic) and cytosolic (hydrophilic) protein-enriched fractions according to manufacturer’s instructions. The protein concentrations of the total cell lysate, cytosolic protein fraction, and membrane protein fraction was determined using a Pierce BCA assay kit and a BSA standard. Protein preparations were then aliquoted and stored at −80 °C until further use.

Protein extracts (200 μg) were labeled with AF555 (λex = 555 nm, λem = 580 nm) in 250 mM sodium borate, pH 8.3, in the dark, essentially as previously described.[53] Briefly, 1 mg of AF555 was dissolved in 100 μL dimethylsulfoxide and 5 μL of the dissolved dye was added to each sample in a final volume of approximately 300 μL and incubated at 25 °C for 2 h in the dark at room temperature. Labeled protein samples were then purified and buffer exchanged in PBS using a 3 kDa MWCO centrifugal filters. These protein samples were quantified for protein content and substitution according to manufacturer’s instructions and stored in the dark at 4 °C until further use.

Lectin Microarray Construction, Incubation, and Data Extraction

A panel of five unlabeled pure lectins, MAL-I, MAL-II, SBA, WFA, and SNA-I, were prepared at 0.5 mg/mL in PBS, pH 7.4, supplemented with 1 mM of the appropriate haptenic sugar (Table ). The lectins were printed at approximately 1 nL per feature on Nexterion Slide H microarray slides using a sciFLEXARRAYER S3 piezoelectric printer (Scienion AG, Berlin, Germany) as previously described.[53] Each microarray slide contained eight replicate subarrays, with each lectin spotted in replicates of six per subarray. To ensure complete conjugation, these slides were then incubated in a humidity chamber overnight at room temperature. Residual functional groups were deactivated by incubation in 100 mM ethanolamine in 50 mM sodium borate, pH 8.0, for 1 h at room temperature. Each slide was washed with PBS, pH 7.4, containing 0.05% Tween-20 three times for 3 min per wash, once with PBS, centrifuged dry (450g, 5 min), and stored at 4 °C with desiccant until use.

Prior to use, the lectin microarray slides were allowed to equilibrate to room temperature for 30 min with desiccant. The microarrays were protected from light throughout the procedure. Fluorescently labeled protein samples were diluted in Tris-buffered saline (TBS) supplemented with Ca2+ and Mg2+ ions (TBS; 20 mM Tris–HCl, 100 mM NaCl, 1 mM CaCl2, 1 mM MgCl2), pH 7.2, with 0.05% Tween-20 (TBS-T) for incubation on the microarray slides. Initially, two fluorescently labeled protein samples were titrated (1–10 μg/mL) to determine the optimal concentration for signal-to-background ratio. For triplicate experiments, the optimal concentration of 2 μg/mL of each labeled sample in TBS-T was incubated on three separate microarray slides. For incubations, 70 μL of each diluted sample was applied to each well of the gasket slide, sandwiched with the lectin microarray in an incubation cassette system (Agilent Technologies), and incubated in the dark (1 h, 23 °C, 4 rpm) essentially as previously described.[53] Following incubation, the microarrays were washed twice in TBS-T and once in TBS for 3 min per wash. Finally, the microarrays were dried by centrifugation and imaged in an Agilent G2505B (Agilent Technologies) microarray scanner using the green channel (532 nm excitation, 90% photomultiplier tube, 5 μm resolution). Images were stored as high-resolution.tif files.

Microarray data extraction from image files was performed essentially as previously described[53] using GenePix Pro v6.1.0.4 (Molecular Devices, Berkshire, U.K.). The data were then exported as text to Excel (version 2010, Microsoft). Local background-corrected median feature intensity data (F532median-B532) values were selected and the median of six replicate spots per subarray was handled as a single data point for graphical analysis. Binding-intensity data are represented in bar charts as the mean intensity with single standard deviation of all like experimental replicates.

Lectin Cytochemistry

Lectin cytochemistry was performed at room temperature on primary astrocytes and Neu7 astrocytes. The cells were fixed with 4% paraformaldehyde in PBS for 10 min and washed four times in Tris-buffered saline (TBS) supplemented with 1 mM each of the divalent cations Ca2+ and Mg2+ necessary for lectin function for 2 min per wash. pBSA was used for all histochemical blocking.[54] The cells were blocked with 2% pBSA in TBS for 30 min, washed four times in TBS, and then incubated with 20 μg/mL in TBS of the FITC-conjugated SNA-I, MAA, WFA, or SBA lectins (Table ) for 1 h in the dark. Inhibitory controls were also carried out in parallel to ensure that lectin binding was carbohydrate-mediated.[21] Inhibition was done by pre-incubating lectins with 100 mM of appropriate haptenic carbohydrates in TBS for 1 h prior to cell staining (SNA-I and MAA in lactose and WFA and SBA in GalNAc), and lectin incubation was also carried out in the presence of the appropriate haptenic sugar. After lectin incubation, the cells were washed twice in TBS and incubated with DAPI for 5 min (1 μg/mL in TBS). The cells were washed four times in TBS and mounted on glass slides with a drop of ProLong Gold antifade reagent (Life Technologies, Grand Island, NY). Images were captured on a Nikon Eclipse E400 fluorescent microscope at 40× magnification and stored digitally for further image analysis.

Lectin Histochemistry of Tissues

Lectin histochemistry was carried out on three slides from all animals within the three experimental groups. The frozen sections were rehydrated in TBS containing 0.05% Triton X-100 (TBS-T2) and washed twice in TBS-T for 2 min per wash. The sections were blocked with 2% pBSA in TBS-T2 for 1 h at room temperature and then washed twice. From this point all, all staining procedures were carried out in the dark at room temperature. The sections were then incubated with 20 μg/mL of FITC-conjugated SNA-I, MAA, WFA, or SBA in TBS-T2 for 2 h. The sections were washed three times in TBS-T2 for 2 min each, followed by a final wash in TBS, and the slides were then cover slipped with ProLong Gold antifade reagent. Inhibitory controls were carried out in parallel as described above. Images were captured on a Nikon Eclipse E400 fluorescent microscope at 40× magnification and stored digitally for subsequent image analysis.

Dual Lectin and Immunohistochemistry of Tissues

For lectin and immunohistochemistry double staining, immediately after the final TBS wash step in the lectin histochemistry method (as above) and continuing the staining procedures in the dark at room temperature, the spinal cord sections were rehydrated in 0.01 M PBS and then blocked with 20% normal goat serum (NGS) (Sigma-Aldrich Co., Dublin, Ireland) in PBS containing 0.2% Triton X-100 for 20 min. The primary antibodies rabbit polyclonal anti-GFAP, 1:300 dilution (DakoCytomation, Dublin, Ireland), rabbit polyclonal antineurogranin (1:100), mouse monoclonal anti-βI-integrin (1:100), and rabbit polyclonal antigrowth-associated protein-43 (GAP-43, 1:100) were diluted in PBS containing 2% NGS and 0.02% Triton X-100 and sections were incubated with the primary antibody for 2 h. The sections were washed three times in PBS and the appropriate secondary antibody, antirabbit IgG conjugated to rhodamine, or antimouse conjugated to Alexa Fluor 594 (Life Technologies, Grand Island, NY) was diluted 1:500 in PBS and incubated for 1 h. The sections were again washed in PBS and cover slipped with ProLong Gold antifade reagent. A negative control was carried out for each antibody by substituting PBS for the primary antibody. Images were captured on an Olympus IX81 fluorescent microscope at 20× and 40× magnifications and stored digitally for further image analysis.

Image Analysis

Three images were captured at the same exposure time from each coverslip containing primary astrocytes and Neu7 astrocytes for each lectin stain. The images were then analyzed for the integrated density of green fluorescence (lectin binding) using Image Pro Plus software (Media Cybernetics, Silver Springs, MD). The quantity of green fluorescence was compared to the number of cells present using the DAPI images to count cell numbers and give the integrated density reading per cell.

Three images were captured from both the white matter and gray matter of the uninjured animal group and from the white matter, gray matter, and lesion site of the injured animal groups for all three lectins. Images were captured at the same exposure time for each region of interest and animal group examined. For each region of interest, the images were randomly chosen. Image Pro Plus was again used to acquire the integrated density of lectin binding from the digital images. For some of the images, the brightness and contrast have been enhanced using Adobe Photoshop CS2 9.0 to allow for easier identification of positively stained tissue.

Statistical Analysis

The average integrated density of lectin staining in vitro and in vivo for each sampling region was calculated using Microsoft Excel v. 2007 and standard error of the mean (SEM) was calculated. Statistical calculations were performed using Minitab 16 software (Minitab Ltd., Coventry, U.K.). A two-way analysis of variance was performed to examine differences between regions of interest and treatment groups. Post-hoc comparisons were undertaken by Fisher’s test. Differences were considered to be statistically significant at a probability value (P) ≤ 0.05.