www.advhealthmat.de www.MaterialsViews.com
Elisa López-Dolado, Ankor González-Mayorga, María Teresa Portolés, María José Feito, María Luisa Ferrer, Francisco del Monte, María Concepción Gutiérrez, and María Concepción Serrano*
FULL PAPER
Subacute Tissue Response to 3D Graphene Oxide Scaffolds Implanted in the Injured Rat Spinal Cord
of neurotrophic factors, and biomaterials, among others. In this context, materials with advanced properties could offer novel perspectives to serve helping this aim. One of these promising materials is graphene, a free-standing single atomic plane of graphite with excellent characteristics that are strongly impelling these days its application in biomedical areas including bioimaging, drug delivery, and photothermal therapy for cancer.[3–6] More recently, other derivatives such as graphene oxide (GO) or partially reduced GO (rGO) are also attracting significant attention as they combine some of the properties of graphene (e.g., mechanical ductility and stiffness), with other interesting features such as a higher hydrophylicity and versatility than graphene for functionalization with biological moieties.[7,8] As described for graphene,[9,10] composites containing GO nanosheets can enhance neural differentiation[11–13] and promote axonal alignment.[14] Additionally, graphene-derived materials (GDM) do not only serve as a platform to support and direct neural growth, but also to assist the fabrication of interfaces for recording and stimulating neural electrical signals.[15–17] The toxicity and biocompatibility in vivo of GDM remain an open debate.[18] Nonetheless, published results support the existence of a safe range of conditions in which these materials might not initiate any toxic responses in biological systems.[19–21] In this sense, studies in mice focused on the intravenous administration of GO nanomaterials have evidenced a clear dose-dependent toxicity with no pathological changes or damage in major organs (e.g., lung, liver, kidney, spleen) at doses as low as 1 mg kg−1 in 14 d.[22] Results at higher doses are contradictory and depend on the experimental conditions.[23,24] Studies on oral,[25] intravitreal,[26] intraperitoneal,[25] and subcutaneous routes refer no major toxic responses.[27] On the contrary, intratracheal administration typically generates some degree of lung injury.[28,29] Regarding the interaction of GDM with immune cells, some early attempts in Caenorhabditis elegans, an alternative nonmammalian organism for toxicity screening, have evidenced a key role played by the innate immunity in the regulation of the chronic toxicity in vivo of GO.[30] Importantly, it is known
The increasing prevalence and high sanitary costs of lesions affecting the central nervous system (CNS) at the spinal cord are encouraging experts in different fields to explore new avenues for neural repair. In this context, graphene and its derivatives are attracting significant attention, although their toxicity and performance in the CNS in vivo remains unclear. Here, the subacute tissue response to 3D flexible and porous scaffolds composed of partially reduced graphene oxide is investigated when implanted in the injured rat spinal cord. The interest of these structures as potentially useful platforms for CNS regeneration mainly relies on their mechanical compliance with neural tissues, adequate biocompatibility with neural cells in vitro and versatility to carry topographical and biological guidance cues. Early tissue responses are thoroughly investigated locally (spinal cord at C6 level) and in the major organs (i.e., kidney, liver, lung, and spleen). The absence of local and systemic toxic responses, along with the positive signs found at the lesion site (e.g., filler effect, soft interface for no additional scaring, preservation of cell populations at the perilesional area, presence of M2 macrophages), encourages further investigation of these materials as promising components of more efficient material-based platforms for CNS repair.
1. Introduction Lesions at the spinal cord in the central nervous system (CNS) have become a major social concern because of their increasing prevalence and the high sanitary costs associated.[1] These limitations are functioning as a fuel to encourage experts in different fields to explore new avenues for neural regeneration,[2] including pharmacological approaches, cell therapy, delivery
Dr. E. López-Dolado, A. González-Mayorga, Dr. M. C. Serrano Hospital Nacional de Parapléjicos (SESCAM) Finca de La Peraleda s/n, 45071 Toledo, Spain E-mail: [email protected] Prof. M. T. Portolés, Dr. M. J. Feito Department of Biochemistry and Molecular Biology I Universidad Complutense de Madrid Ciudad Universitaria s/n, 28040 Madrid, Spain Dr. M. L. Ferrer, Dr. F. del Monte, Dr. M. C. Gutiérrez Instituto de Ciencia de Materiales de Madrid (ICMM) Consejo Superior de Investigaciones Científicas (CSIC) C/ Sor Juana Inés de la Cruz 3, 28049 Madrid, Spain
DOI: 10.1002/adhm.201500333
Adv. Healthcare Mater. 2015, 4, 1861–1868
© 2015 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
wileyonlinelibrary.com
1861
www.advhealthmat.de
FULL PAPER
www.MaterialsViews.com
that the cross-talk between neural stem and immune cells is critical for inducing reparative responses.[31] To this regard, specific studies in the spinal cord by Popovich and co-workers have revealed the existence of distinct macrophage subsets with pivotal implications in neurotoxic versus neuroregenerative responses.[32] In this sense, the proinflammatory (M1) versus immunomodulatory/reparative (M2) balance has been ascribed to the capacity of macrophages to play both positive and negative roles in disease processes and tissue remodeling after injury.[33,34] Here, we investigate for the first time the subacute tissue responses to 3D rGO scaffolds implanted in the injured rat spinal cord (C6 level). The interest of these structures mainly relies on their excellent mechanical properties. Particularly, mechanical compression studies demonstrated their remarkable flexibility in both longitudinal and transversal directions,[35] thus becoming a soft material interface that is mechanically compatible with the nervous tissue (0.3–1.0 kPa).[36] Values ca. 1 kPa have been reported for comparable GO foams also prepared by freeze-drying when exposed to strains under 10%.[37] This attractive feature, along with their biocompatibility demonstrated with neural cell cultures in vitro,[35] has boosted the exploration of their behavior in vivo to anticipate advantages and/or limitations concerning their potential use as a guiding platform for neural regeneration in the CNS. Any interest from the electrical point of view is abolished by their dramatically low conductivities (ca. 10−3 S cm−1, unpublished data measured as previously described).[38] In this work, subacute responses have been thoroughly investigated both local and systemically. Due to the pivotal relevance of immune cross-talk on neural repair, the expression of macrophage surface markers CD80 and CD163 is also investigated to identify M1 and M2 cells, respectively, in contact with the rGO implant.
2. Results and Discussion Encouraged by the attractive properties of GDM and their interest for neural repair in the shape of 3D substrates,[39,40] we herein investigate the tissue response of the injured rat spinal cord to 3D porous rGO scaffolds after 10 d of implantation, defined as a subacute phase in rat models.[41] Further details on the physicochemical characterization of both the GO used and the resulting rGO scaffolds have been published elsewhere.[35] Figure 1A displays sequential schemes of the spinal cord injury practiced (for further details, please refer to the Experimental Section). Treatment groups included control without injury (C), injury (I) and injury + scaffold (I+SC). Images of a representative rGO scaffold implanted at C6 level and the posterior coverage of the injured area with a thin gelatin hydrogel film (Figure 1B,C, respectively) are also displayed. The average amount of rGO implanted per animal was 0.25 mg, a quantity that is within the safe range of GO concentrations described as nontoxic in previous reports in mice.[22,23] As evidenced from systematic post-operatory (PO) observations, all injured animals no matter the treatment received regained interest for actively exploring the environment and showed an adequate hydration state, absence of pain or distress signs, and proper food intake and defecation after 3–4 d post-surgery. An evident
1862
wileyonlinelibrary.com
Figure 1. Scheme of the surgery model practiced (A). Representative images of an rGO scaffold implanted at the right spinal hemicord at C6 (B), posteriorly covered with a thin gelatin hydrogel (C). Injured rats carrying rGO implants for 7 d displayed two apparent phenotypes: close fist (D) and open hand (E). Ventral view of a representative 10-d explanted spinal cord (F). Set of arrows indicates: C—Caudal, L—Left, R—Right, and Ro—Rostral. Dashed red line in (F) marks the midline of the spinal cord.
paresis of the anterior right limb was referred in all cases, with rats displaying two slightly different phenotypes, “close fist” (Figure 1D) and “open hand” (Figure 1E), thus indicating a small variation in the lesion epicenter size that, in the second phenotype, comprised finger and wrist muscles besides elbow extensors. When sacrificed at 10 PO days (POD), rGO scaffolds were found properly allocated at the injured site inside the spinal cord where initially placed, with evident scaffold adhesion to both the adjacent neural tissue in the spinal cord and deep muscle layers on top. Macroscopic examination of the explants also revealed the accomplishment of nearly complete hemisections at the right C6 spinal segment, as confirmed in ventral views of the injured spinal cord (Figure 1F). We next focused on the specific tissue response of the injured spinal cord. Conventional hematoxylin-van Giesson (HvG) staining confirmed a complete adhesion of the rGO scaffolds implanted to the surrounding neural tissue inside the injured spinal cord, without significant cavities (Figure 2, right column) and with an evident tissue interface. The initial criterion to define the borders of such an interface was based on these HvG-stained sections, posteriorly confirmed as distinguishable on immune-stained ones. Prior to implantation, these 3D rGO scaffolds typically present a porosity of ca. 80%, as estimated from scanning electron microscopy (SEM) images.[35] This feature, along with their characteristic softness, supported the infiltration of cells and collagen fibers (Figure 2, bottom right). As characteristic in CNS lesions,[42] most of these cells probably migrated by chemotactic phenomena from surrounding tissues and blood vessels. It is worth noting that GO could also partially contribute to this effect due to its described role as a nonspecific enhancer of mammalian cell growth.[43] Remaining free space in the implanted scaffolds will permit further colonization by cells and extracellular matrix components, thus compelling with the prerequisite of tissue perfusion
© 2015 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Healthcare Mater. 2015, 4, 1861–1868
www.advhealthmat.de www.MaterialsViews.com
FULL PAPER
I+SC groups. Figure 5 shows similar stainings for the interface around the scaffold in I+SC. Quantification of all these markers in gray matter areas is included in Figure 6 and 2S and Table S1 (Supporting Information). Positive stained area for cell nuclei in the proximities of the lesion (at 0–1 and 1–2 mm from the lesion borders) in both I and I+SC groups was similar to that found on healthy nervous tissue in C rats (Figure 6A). At the injury site in I and I+SC, the amount of cells significantly increased compared to that of perilesional and no lesion areas likely indicating the activation of both cell migration and proliferation from surrounding tissues, which is characteristic of subacute CNS lesions.[42] The interface displayed a similar Figure 2. Histological examination of the implantation site at 10 POD by HvG staining. Images at the bottom row represent zoom-in details of areas marked with orange squares in cell density to that found at the lesion in I top images. Spinal cords are oriented in all cases as indicated by the set of arrows: C—Caudal, group, both significantly inferior to scafD—Dorsal, Ro—Rostral, and V—Ventral. folds in I+SC maybe related to the ability of GO to promote cell proliferation.[43] Markers for neurons (e.g., map-2 and tau) were highly present outside that is pursued in biomaterials aimed to have scaffold functions the lesion epicenter (map-2 mainly in gray matter areas, tau in (i.e., supportive and guiding properties).[2] The tissue forming both white and gray matter ones) and slightly at the interface, the interface in the I+SC group resembled that found filling while totally absent inside the injured area (Figure 6B and S2, the cavity in I (Figure 2, middle column), with some small Supporting Information). On the contrary, the positive reacpieces of damaged neural tissue not properly extracted after tivity found for vimentin (potentially contributed by glial cells, lesion found in the injured area in both I and I+SC groups. pericytes, and connective tissue cells, among others) revealed Morphometric analyses demonstrated that the average lesion a significant distance-dependent gradient, thus being very low volume was 7.6 ± 2.2 mm3, with 3.7 ± 1.6 mm3 contributed by in control and perilesional nervous tissue at 1–2 mm, sigthe scaffold and 3.9 ± 1.2 mm3 by the interface tissue formed nificantly higher at 0–1 mm from the lesion and even higher around the scaffold (percentage of interface at the lesion site: in injured areas of both I and I+SC (scaffold and interface) 52.8 ± 14.7 %) (Figure 3A). Interestingly, interface width around the scaffold did not show significant differences on the different groups. Interestingly, at closer perilesional areas (0–1 mm), regions (i.e., rostral, ventral, and caudal, p = 0.393) (Figure 3B). vimentin expression was higher in rostral than caudal zones HvG images confirmed that this interface tissue was composed likely indicative of a higher retrograde than anterograde of both cells and collagen fibers (Figure 3C). Finally, no indicaneural damage. Posterior immunofluorescence studies also tions of scaffold degradation were observed at 10 POD. Further showed highly positive staining for ED1 (typically expressed studies at the chronic state will necessarily serve to address this by macrophages) and platelet-derived growth factor receptor β aspect at longer time points. (PDGFRβ), recognized as an essential regulator of early hematTo better characterize the effect of rGO scaffold implantaopoiesis and blood vessel formation,[44] at the injured areas of tion at the lesion site, immunofluorescence studies of specific both I and I+SC groups. These two markers were rare in the markers were carried out to identify the nature of both the cells control and perilesional nervous tissue, being PDGFRβ+ cells infiltrating the scaffold and those forming the interface tissue mainly grouped at the perivascular regions (i.e., pericytes). The in comparison to no lesion areas. Figure 4 illustrates representdramatically positive reactivity for PDGFRβ+ cells at the injury ative images of no lesion areas in C and lesion sites for I and site is not surprising as this receptor is a typical marker of type
Figure 3. Examination of the interface tissue in contact with the implanted rGO scaffold at 10 POD. Percentage of interface versus scaffold in the injured spinal cord (A), in which Y axis indicates consecutive sections (S) of a representative I+SC spinal cord sample. The interface tissue formed displayed a similar width at rostral, ventral, and caudal regions (B). A representative HvG-stained section illustrating the interface tissue formed around the rGO scaffold implanted (C). Yellow dashed lines limit the borders of the interface.
Adv. Healthcare Mater. 2015, 4, 1861–1868
© 2015 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
wileyonlinelibrary.com
1863
www.advhealthmat.de
FULL PAPER
www.MaterialsViews.com
Figure 4. Representative immunofluorescence images of the lesion site at 10 POD. Control samples are displayed on top for comparison. Bright-field images are included to confirm location of the stained cells inside the scaffold. Cell nuclei were stained with Hoechst (hoe). Scale bars: 100 µm.
sent at the interface or infiltrating the 3D scaffold structure, being slightly more abundant in the first one. This result is not surprising as both neurotoxic and neuroregenerative effects can be elicited concurrently at the injury site.[46] As can be noticed, only a reduced fraction of the ED1+ population previously identified at the injury site (Figure 4, 5) stained positive for these markers, likely due to the fact that macrophage phenotype occurs in vivo along a continuum between M1 and M2 ends.[47] Contrarily to these results, no evidence of these cells was found in perilesional areas. The slight reduction of macrophage population inside the scaffold might be related to some specific toxic effects driven by rGO on this cell type, as recent work has revealed that GO is capable of inducing macrophage necrosis via Toll-like receptor 4 even in the absence of GO phagocytosis.[48] A significant percentage of actively proliferating cells was also identified at the injury site in both I (12.2 ± 5.4%) and I+SC (12.3 ± 3.5% in the scaffold and 15.4 ± 5.6% at the interface) groups, in contrast with the reduced amount found in C rats (1.6 ± 2.8%, p = 0.025) (Figure 7, bottom row). This finding indicates that some of the cells filling the cavity created at the injury site and infiltrating the scaffold, which we hypothesized were recruited from circulation and surrounding tissues by chemotactic effects, retained an actively proliferative phenotype. Motivated by the local effects found at the injured spinal cord and alerted by previously reported work on graphene toxicity, we then examined the major organs (e.g., brain, heart, intestine, kidney, liver, lungs, and spleen), searching for any signs of systemic toxicity induced by rGO scaffold implantation. After a gross inspection, organs did not show any Figure 5. Representative immunofluorescence images of the interface tissue in contact with major macroscopic abnormalities in size, color, morphology, or texture. Organ weight the scaffold at 10 POD. Cell nuclei were stained with Hoechst (hoe). Bright-field images are in I and I+SC groups remained unaltered at included to confirm location of the stained cells relative to the scaffold. A pericytes. These cells are involved in the formation of the core of the spinal cord scar tissue,[45] as well as in the growth of connective tissue cells and collagen production characteristic of wound healing processes,[44] all these phenomena commonly taking place at CNS injuries.[42] The expression of glial fibrillary acidic protein (GFAP), a specific marker of astrocytes, also followed a distance-dependent gradient as previously indicated for vimentin (i.e., higher GFAP expression at locations closer to the lesion). However, and contrary to vimentin, this marker was almost negligible at the lesion epicenter in both I and I+SC. Despite this border effect in the injury site, a discrete internalization of GFAP+ cells into the scaffold structure was observed in three out of six animals, with migration distances that overpassed 200 µm (up to 325 µm in one case) (Figure S3, Supporting Information). Figure 7 illustrates positive staining for macrophages expressing either CD80 (top row) or CD163 (middle row) from a representative I+SC rat. Both types of macrophages were pre-
1864
wileyonlinelibrary.com
© 2015 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Healthcare Mater. 2015, 4, 1861–1868
www.advhealthmat.de www.MaterialsViews.com
10 POD when compared to C rats (Figure S4, Supporting Information). However, the percentage of total weight loss per animal was found highly variable among individuals, but independent from the type of treatment received (I: 15.06 ± 7.20 %; I+SC:
Figure 7. Immunofluorescence studies of macrophages expressing either CD80 (top) or CD163 (middle) at 10 POD in perilesional areas, at the interface or inside the scaffold in I+SC rats. Scale bars: 25 µm. Bottom row images illustrate the presence of proliferating cells stained with Ki67 in C (left), I (middle) and I+SC (right) rats at 10 POD. Staining with Hoechst (blue) is merged to facilitate proper cell identification. Scale bars: 100 µm.
Adv. Healthcare Mater. 2015, 4, 1861–1868
© 2015 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
wileyonlinelibrary.com
FULL PAPER
Figure 6. Quantitative analyses of positive stained areas for A) cell nuclei and B) specific immunofluorescence markers in the injured spinal cord at 10 POD. Statistically significant differences are indicated in comparison to: a) no lesion (control) and perilesional areas, b) I, c) interface in I+SC (IF), and d) scaffold in I+SC (SC) groups.
16.01 ± 7.23 %; p = 0.824). However, some degree of correlation was found with the size of the lesion (y = 0.27x + 2.78; R2 = 0.6955), being larger lesions prompted to cause a higher weight loss. We then proceeded with a thorough inspection of those organs typically involved in material degradation, elimination, and accumulation (i.e., kidney, liver, lung, and spleen).[22,24] In agreement with previous work on GO-based nanomaterials inoculated oral, intraperitoneal or intravenously[24,25,49] and as expected from the reduced amount implanted,[22,23] no significant damage was observed in any of the major organs evaluated, including those of the reticuloendothelial system (i.e., liver and spleen). Transmission optical microscopy analyses revealed no alteration of normal tissue structure, inflammation, fibrosis, atrophy, or atypical cell responses in any analyzed section of the target viscera at either 10× images (Figure 8) or higher magnification ones (40×) (Figure S5, Supporting Information). Taking together, these results indicate that the implantation of rGO scaffolds at the injured rat spinal cord does not induce any significant subacute toxic responses either at any of the major organs investigated or perilesionally. At the lesion site, these structures have demonstrated: 1) A clear filler effect that allows to regain tissue integrity after injury at earlier time points, 2) A soft interface that facilitates accommodation into the injured site preventing additional scaring by friction,[50] 3) 3D scaffold colonization by a majority of PDGBRβ+ cells contributing to the stabilization of the injury and the prevention of further damage, and 4) Infiltration of reparative M2 macrophages, which could eventually prompt suitable repair responses. In this particular sense, the potential modulation of immune cell function by regulating the M1/M2 balance suggests the possibility of designing biomaterials capable of eliciting appropriate immune responses at the implantation site.[51] Future work will aim to direct macrophage differentiation towards the M2 phenotype by, for instance, releasing interleukin-4,[32] as already explored for peripheral nerve repair.[52] Studies targeting PDGBRβ+ cells present at the lesion site will be also of enormous interest due to their potential relationship with pericytes and ependymal cells and their implication in spinal cord scar tissue.[45] For all these studies, it will be critical to achieve the correct redox state of the GO used, as its degree of stacking and electrostatic/ hydrogen bonding is responsible for growth factor binding and its posterior influence on cell responses including differentiation.[53] From the mechanical point of view, it is known that the glial response around CNS implanted devices strongly depends on their mechanical properties—the more compliant the material is, the less inflammatory glial response encapsulating the device is formed.[54] Therefore, further studies at longer implantation times will take advantage of the mechanical compliance and ability to absorb compressions of these scaffolds to better guide neural growth by defining optimized guidance channels decorated with both topographical and biological cues. Since the topographical and structural features of these 3D rGO structures appeared insufficient, effective cocktails for their biofunctionalization will be explored to stimulate pro-regenerative elements at the scar, benefit neural cell growth and axonal sprouting through the lesion site and then aid the re-establishment of functional connections. Alternatively, 3D porous structures composed of biocompatible and biodegradable polymers such as poly(lactic-co-glycolic acid) and polycaprolactone, both
1865
www.advhealthmat.de
FULL PAPER
www.MaterialsViews.com
1866
immersion. Scaffold morphology was characterized by SEM (Hitachi S-4700 SEM microscope) and confirmed similar to that previously described.[35] Gelatin (Batch#1189632-52205137, porcine skin) hydrogel was dialyzed in distilled water and further dissolved at 50 °C. The final solution was then freeze-dried and stored at 4 °C until used. Fresh hydrogel films (9 mm2, ca. 100 µm in thickness) were prepared by dissolving the freeze-dried product in sterile distilled water (2 wt%) at 37 °C. Finally, the resulting films were sterilized under UV radiation and hydrated with saline solution 5 min prior implantation. Surgical Procedures: Adult male wistar rats were provided by a commercial supplier (Harlan Ibérica, Spain) and used at the age of ca. 30 weeks (n = 16, 451 ± 67 g). The use of aged adult rats was motivated by two reasons: 1) A larger size facilitating surgical procedures and 2) the incidence of spinal cord injury rising in older adults and the Figure 8. Histological examination of major organs at 10 POD. Representative 10× optical elderly human population nowadays.[57] Although microscopy images of kidney, liver, lung, and spleen after HE staining. Control samples are the role of age in recovery and mortality after spinal displayed for comparison. cord injury is not completely clarified yet, some differences had been shown in the inflammatory [58] morphometric CNS tissue size after lesion[59] and cell response, materials extensively explored for nervous tissue regeneration, functional locomotor deficits at the chronic states.[60] The lesion model and benefitted from GDM coatings could be also an interesting of choice was a right lateral hemisection of approximately 8 mm3 avenue to explore in this scenario. (2 × 2 × 2) at the C6 segment, rostral to the bulk of triceps brachii motoneurons, previously demonstrated as a suitable model to evaluate therapeutic strategies aimed at promoting neural plasticity and repair. As it causes focal foreleg impairments attributable to the neural damage 3. Conclusion caused, both the impairments and their eventual recovery can be evidenced and measured in detail during locomotion by using careful We have implanted for the first time 3D porous and flexible functional and anatomical techniques.[61] Some other advantages of rGO scaffolds at the injured rat spinal cord and explored the this model include: a) higher clinical relevance, since most human subacute tissue response generated. Results obtained evidence spinal cord injuries occur in the cervical area and restoration of arm the capacity of these structures to facilitate regaining tissue and hand function is a top priority for improving the quality of life of tetraplegic patients;[62] b) the proximity of cervical motor nuclei to the integrity after spinal cord injury as early as 10 d and to prelesion, so even short distance axonal growth could improve forelimb vent the extension of the lesion, although unable to guide function; and c) regarding ethical considerations and requirements neural growth through it. Absence of local and systemic toxic of animal care, this type of lesion typically causes damage in the responses, along with promising features at the injury site such ipsilateral forelimb with minimal contralateral forelimb and hindlimb as the establishment of a soft interface preventing additional sensorimotor impairments or autonomic dysfunction. All experimental scaring and the presence of reparative cells such as M2 macprotocols adhered to the regulations of the European Commission (directives 2010/63/EU and 86/609/EEC) and the Spanish government rophages, encourage further investigation of GDM as prom(RD53/2013) for the protection of animals used for scientific purposes. ising components of more efficient platforms for the treatment Specific details on the surgical procedures carried out are provided in of spinal cord injury. the Supporting Information. In those animals carrying rGO scaffolds, monoliths of appropriate dimensions were implanted at the lesion site to fill in the cavity created. All lesions, with or without scaffolds, were covered with a thin gelatin hydrogel film to serve as a biological and mechanically compatible interface between the injured nervous 4. Experimental Section tissue and the deep muscle layers. From the surgical point of view, this Additional experimental details are available in the Supporting hydrogel film assisted the preservation of the scaffold position inside Information. the cavity practiced in the spinal cord during the first stages postMaterial: Chemical reagents and antibodies were purchased from surgery without causing mechanical mismatch with the surrounding Sigma–Aldrich and used as received, unless otherwise indicated. native tissues. The use of this gelatin hydrogel is supported by previous Preparation of 3D Porous rGO Scaffolds and Gelatin Hydrogel Films: work demonstrating the interest of this polymer as part of regenerative 3D rGO scaffolds were fabricated by using the ice segregation-induced strategies at the injured spinal cord.[63,64] Treatment groups included: [ 38,55 ] [ 35 ] self-assembly (ISISA) technique as previously described. The C (n = 3), I (n = 6) and I+SC (n = 6). An exhaustive PO care protocol ISISA procedure is a simple and versatile, bottom-up self-assembly was applied after surgery, with major attention placed into signs of pain, methodology based on the unidirectional freezing of gels or colloidal distress, dehydration, intestinal obstruction, and respiratory failure. suspensions and their subsequent liophilization.[56] Details on the To assess the early tissue response, rats were sacrificed at 10 POD by physicochemical characterization of the GO powder used as well as a standard perfusion-fixation protocol (further details in Supporting [ 35 ] of the obtained rGO scaffolds were published elsewhere. Prior to Information). implantation, monoliths (4.5 mm in diameter, ca. 2 mm in height) Necropsy Protocol for Thoracoabdominal Viscera and CNS Tissue: were sterilized under UV radiation and then infiltrated with tissue Immediately after the perfusion-fixation protocol, rats were placed culture-grade water under vacuum conditions until complete scaffold on their back with the thoracic cavity opened and the different
wileyonlinelibrary.com
© 2015 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Healthcare Mater. 2015, 4, 1861–1868
www.advhealthmat.de www.MaterialsViews.com
Supporting Information Supporting Information is available from the Wiley Online Library or from the author.
Adv. Healthcare Mater. 2015, 4, 1861–1868
Acknowledgements E.L.D. and A.G.M. contributed equally to this work. This work was supported by the Instituto de Salud Carlos III (ISCIII) and Ministerio de Economía y Competitividad (MINECO) (Grant CP13/00060), co-financed by FEDER funding, and MINECO (Grants MAT2011-25329 and MAT2013-43229-R). AGM and MCS acknowledge ISCIII-MINECO for respective contracts associated to project CP13/00060. Authors express their gratitude to Dr. Collazos-Castro for fruitful discussions.
FULL PAPER
thoracoabdominal viscera progressively extracted (further details in Supporting Information). Then, rats were bent over to obtain the CNS tissue (i.e., brain and spinal cord). After extraction, all organs were weighted, carefully examined looking for any signs of gross damage and properly prepared for subsequent histopathological and immunofluorescence procedures. Organ and Neural Tissue Processing: After the perfusion, all samples were placed in paraformaldehyde 4% at 4 °C overnight and then 3 d in sucrose (30% in phosphate buffer saline, PBS) at 4 °C for cryoprotection. C5–C7 spinal cord segments were selected for further analysis. Tissue pieces were mounted on plastic containers, quick-frozen in optimal cutting temperature compound (Tissue Tek, Hatfield, PA) and cut in horizontal sagittal sections of 10-µm by using a Microm HM550 cryostat with an angle of 10°. In the particular case of the spinal cords, cuts were carried out throughout the entire C5–C7 fragment following the sagittal direction from right to left. Histology and Morphometric Analyses: Spinal cords and organs were initially examined after conventional HvG and hematoxylin-eosin (HE) stainings, respectively. Digital images were collected with an Olympus BX51 fluorescence microscope with a digital camera coupled. All images were calibrated and processed by using ImageJ (1.47v, National Institutes of Health, USA). For each animal, the size and extension of the lesion in the rostrocaudal direction were measured by drawing the contours of the total lesion, the interface and the scaffold on calibrated 2× images of entire HvG series. The 3D reconstruction of the lesion was completed by estimating the contribution from the remaining nine series in the right–left direction. Specifically, area values (µm2) for total lesion, interface, and scaffold in each section were multiplied by 10 µm (i.e., section thickness) and 10 (i.e., total number of series for each spinal cord). The total size of the lesion was then obtained as the summation of all these volumes and expressed in mm3. Immunofluorescence Studies: Spinal cord samples were examined for the presence of the following markers: 1) map-2 for somas and dendrites in neurons, 2) tau for axons in neurons, 3) vimentin for non-neuron cells including glial cells and connective tissue cells, 4) GFAP for astrocytes, 5) ED1 for macrophages, 6) PDGFRβ for pericytes, precursors of oligodendrocytes and connective tissue cells such as fibroblasts and smooth muscle cells,[44,45] 7) CD80 for M1 macrophages, 8) CD163/ M130 for M2 macrophages, and 9) Ki67 for actively proliferating cells. Appropriate secondary antibodies were selected. Further details of the specific antibodies and conditions used are provided in the Supporting Information. Cell nuclei were labeled with Hoechst (1 mg mL−1). Digital images were collected with a fluorescence microscope with a digital camera coupled by using appropriate filters. Figure S1 (Supporting Information) illustrates the different areas of study, including control in C rats (no lesion), perilesional areas at 1–2 and 0–1 mm from the lesion site in I and I+SC groups (both rostral and caudally) and lesion areas in I and I+SC. Additionally, bright-field optical microscopy images were captured to define cell location relative to the scaffold. Images were calibrated and processed by using the ImageJ software. For each fluorescent marker, positive stained areas in at least 10 images were measured, expressed as a percentage of the total image area and averaged. Statistics: Values were expressed as mean ± standard deviation of at least three different animals per group (n ≥ 3). Statistical analysis was performed by using the Statistical Package for the Social Sciences software (SPSS, version 17.0). Comparisons among groups were done by analysis of variance (ANOVA) and either post-hoc Scheffé or GamesHowell tests (for either homogeneous or heterogeneous variances, respectively). The significance level was defined as p < 0.05 in all statistical evaluations.
Received: May 1, 2015 Revised: June 1, 2015 Published online: June 25, 2015
[1] V. Y. Ma, L. Chan, K. J. Carruthers, Arch. Phys. Med. Rehabil. 2014, 95, 986. [2] J. K. Hyun, H. W. Kim, J. Tissue Eng. 2010, 1, 650857. [3] D. Bitounis, H. Ali-Boucetta, B. H. Hong, D. H. Min, K. Kostarelos, Adv. Mater. 2013, 25, 2258. [4] Y. Yang, A. A. Asiri, Z. Tang, D. Du, Y. Lin, Mater. Today 2013, 16, 365. [5] H. Zhang, G. Grüner, Y. J. Zhao, Mater. Chem. B 2013, 1, 2542. [6] L. Feng, L. Wu, X. Qu, Adv. Mater. 2013, 25, 168. [7] X. Tan, L. Feng, J. Zhang, K. Yang, S. Zhang, Z. Liu, R. Peng, ACS Appl. Mater. Interfaces 2013, 5, 1370. [8] B. J. Hong, O. C. Compton, Z. An, I. Eryazici, S. T. Hguyen, ACS Nano 2012, 6, 63. [9] S. Y. Park, J. Park, S. H. Sim, M. G. Sung, K. S. Kim, B. H. Hong, S. Hong, Adv. Mater. 2011, 23, H263. [10] Y. Wang, W. C. Lee, K. K. Manga, P. K. Ang, J. Lu, Y. P. Liu, C. T. Lim, K. P. Loh, Adv. Mater. 2012, 24, 4285. [11] O. Akhavan, E. Ghaderi, E. Abouei, S. Hatamie, E. Ghasemi, Carbon 2014, 66, 395. [12] D. Yang, T. Li, M. Xu, F. Gao, J. Yang, Z. Yang, W. Le, Nanomedicine 2014, 9, 2445. [13] S. Shah, P. T. Yin, T. M. Uehara, S. T. D. Chueng, L. Yang, K. B. Lee, Adv. Mater. 2014, 26, 3673. [14] A. Solanki, S. T. D. Chueng, P. T. Yin, R. Kappera, M. Chhowalla, K. B. Lee, Adv. Mater. 2013, 25, 5477. [15] X. Duan, T. M. Fu, J. Liu, C. M. Lieber, Nano Today 2013, 8, 351. [16] E. K. Wujcik, C. N. Monty, Wiley Interdiscip. Rev. Nanomed. Nanobiotechnol. 2013, 5, 233. [17] T. H. Kim, K. B. Lee, J. W. Choi, Biomaterials 2013, 34, 8660. [18] A. B. Seabra, A. J. Paula, R. de Lima, O. L. Alves, N. Durán, Chem. Res. Toxicol. 2014, 27, 159. [19] A. M. Pinto, I. Gonçalves, F. D. Magalhães, Colloid. Surf. B 2013, 111, 188. [20] X. Hu, Q. Zhou, Chem. Rev. 2013, 113, 3815. [21] C. Bussy, H. Ali-Boucetta, K. Kostarelos, Acc. Chem. Res. 2013, 46, 692. [22] X. Zhang, J. Yin, C. Peng, W. Hu, Z. Zhu, W. Li, C. Fan, Q. Huang, Carbon 2011, 49, 986. [23] K. Wang, J. Ruan, H. Song, J. Zhang, Y. Wo, S. Guo, D. Cui, Nanoscale Res. Lett. 2011, 6, 8. [24] K. Yang, J. Wan, S. Zhang, B. Tian, Y. Zhang, Z. Liu, Biomaterials 2012, 33, 2206. [25] K. Yang, H. Gong, X. Shi, J. Wan, Y. Zhang, Z. Liu, Biomaterials 2013, 34, 2787. [26] L. Yan, Y. Wang, X. Xu, C. Zeng, J. Hou, M. Lin, J. Xu, F. Sun, X. Huang, L. Dai, F. Lu, Y. Liu, Chem. Res. Toxicol. 2012, 25, 1265. [27] A. Sahu, W. I. Choi, G. Tae, Chem. Commun. 2012, 48, 5820.
© 2015 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
wileyonlinelibrary.com
1867
www.advhealthmat.de
FULL PAPER
www.MaterialsViews.com
1868
[28] M. C. Duch, G. R. Budinger, Y. T. Liang, S. Soberanes, D. Urich, S. E. Chiarella, L. A. Campochiaro, A. Gonzalez, N. S. Chandel, M. C. Hersam, G. M. Mutlu, Nano Lett. 2011, 11, 5201. [29] B. Li, J. Yang, Q. Huang, Y. Zhang, C. Peng, Y. Zhang, Y. He, J. Shi, W. Li, J. Hu, C. Fan, NPG Asia Mater. 2013, 5, e44. [30] Q. Wu, Y. Zhao, J. Fang, D. Wang, Nanoscale 2014, 6, 5894. [31] Z. Kokaia, G. Martino, M. Schwartz, O. Lindvall, Nat. Neurosci. 2012, 15, 1078. [32] K. A. Kigerl, J. C. Gensel, D. P. Ankeny, J. K. Alexander, D. J. Donnelly, P. G. Popovich, J. Neurosci. 2009, 29, 13435. [33] B. N. Brown, B. D. Ratner, S. B. Goodman, S. Amar, S. F. Badylak, Biomaterials 2012, 33, 3792. [34] P. M. Kou, J. E. Babensee, J. Biomed. Mater. Res. A 2011, 96, 239. [35] M. C. Serrano, J. Patiño, C. García-Rama, M. L. Ferrer, J. L. Fierro, A. Tamayo, J. E. Collazos-Castro, F. del Monte, M. C. Gutiérrez, J. Mater. Chem. B 2014, 2, 5698. [36] Y. B. Lu, K. Franze, G. Seifert, C. Steinhäuser, F. Kirchhoff, H. Wolburg, J. Guck, P. Janmey, E. Q. Wei, J. Käs, A. Reichenbach, Proc. Nat. Acad. Sci. U.S.A. 2006, 103, 17759. [37] J. Kuang, L. Liu, Y. Gao, D. Zhou, Z. Chen, B. Han, Z. Zhang, Nanoscale 2013, 5, 12171. [38] M. C. Serrano, S. Nardecchia, C. García-Rama, M. L. Ferrer, J. Collazos-Castro, F. del Monte, M. C. Gutiérrez, Biomaterials 2014, 35, 1543. [39] N. Li, Q. Zhang, S. Gao, Q. Song, R. Huang, L. Wang, L. Liu, J. Dai, M. Tang, G. Cheng, Sci. Rep. 2013, 3, 1604. [40] S. R. Shin, B. Aghaei-Ghareh-Bolagh, X. Gao, M. Nikkhah, S. M. Jung, A. Dolatshahi-Pirouz, S. B. Kim, S. M. Kim, M. R. Dockmeci, X. S. Tang, A. Khademhosseini, Adv. Funct. Mater. 2014, 24, 6136. [41] R. van den Brand, J. Heutschi, Q. Barraud, J. DiGiovanna, K. Bartholdi, M. Huerlimann, L. Friedli, I. Vollenweider, E. M. Moraud, S. Duis, N. Dominici, S. Micera, P. Musienko, G. Courtine, Science 2012, 336, 1182. [42] J. W. Fawcett, R. A. Asher, Brain Res. Bull. 1999, 49, 377. [43] O. N. Ruiz, K. A. Fernando, B. Wang, N. A. Brown, P. G. Luo, N. D. McNamara, M. Vangsness, Y. P. Sun, C. E. Bunker, ACS Nano 2011, 5, 8100. [44] P. H. Chen, X. Chen, X. He, Biochim. Biophys. Acta 2013, 1834, 2176. [45] C. Göritz, D. O. Dias, N. Tomilin, M. Barbacid, O. Shupliakov, J. Frisén, Science 2011, 333, 238.
wileyonlinelibrary.com
[46] J. C. Gensel, S. Nakamura, Z. Guan, N. van Rooijen, D. P. Ankeny, P. G. Popovich, J. Neurosci. 2009, 29, 3956. [47] D. M. Mosser, J. P. Edwards, Nat. Rev. Immunol. 2008, 8, 958. [48] G. Qu, S. Liu, S. Zhang, ACS Nano 2013, 7, 5732. [49] H. Ali-Boucetta, D. Bitounis, R. Raveendran-Nair, A. Servant, J. van den Bossche, K. Kostarelos, Adv. Healthcare Mater. 2013, 2, 433. [50] K. A. Potter, M. Jorfi, K. T. Householder, E. J. Foster, C. Weder, J. R. Capadona, Acta Biomater. 2014, 10, 2209. [51] S. Franz, S. Rammelt, D. Scharnweber, J. C. Simon, Biomaterials 2011, 32, 6692. [52] N. Mokarram, A. Merchant, V. Mukhatyar, G. Patel, R. V. Bellamkonda, Biomaterials 2012, 33, 8793. [53] W. C. Lee, C. H. Y. X. Lim, H. Shi, L. A. L. Tang, Y. Wang, C. T. Lim, K. P. Loh, ACS Nano 2011, 5, 7334. [54] P. Moshayedi, G. Ng, J. C. F. Kwok, G. S. H. Yeo, C. E. Bryant, J. W. Fawcett, K. Franze, J. Guck, Biomaterials 2014, 35, 3919. [55] A. Abarrategi, M. C. Gutiérrez, C. Moreno-Vicente, M. J. Hortigüela, V. Ramos, J. L. López-Lacomba, M. L. Ferrer, F. del Monte, Biomaterials 2008, 29, 94. [56] M. C. Gutiérrez, M. L. Ferrer, F. del Monte, Chem. Mater. 2008, 20, 634. [57] M. J. Hooshmand, M. D. Galvan, E. Partida, A. J. Anderson, Immun. Ageing 2014, 11, 15. [58] H. Kumamaru, H. Saiwai, Y. Ohkawa, H. Yamada, Y. Iwamoto, S. Okada, J. Cell. Physiol. 2012, 227, 1335. [59] A. C. Laing, R. Cox, W. Tetzlaff, T. Oxland, Anat. Rec. 2011, 294, 1326. [60] Y. S. Gwak, B. C. Hains, K. M. Johnson, C. E. Hulsebosch, J. Neurotrauma 2004, 21, 983. [61] E. López-Dolado, A. M. Lúcas-Osma, J. E. Collazos-Castro, J Neurotrauma 2013, 30, 191. [62] NSCISC National Spinal Cord Injury Statistical Center. 2013 Annual Statistical Report. University of Alabama, https://www.nscisc.uab.edu/reports.aspx (accessed: April 2015). [63] H. Altinova, S. Moellers, T. Fuhrmann, R. Deumens, A. Bozkurt, I. Heschel, L. H. H. O. Damink, F. Schugner, J. Weis, G. A. Brook, Brain Res. 2014, 1585, 37. [64] S. Rochkind, A. Shahar, D. Fliss, D. El-Ani, L. Astachov, T. Hayon, M. Alon, R. Zamostiano, O. Ayalon, I. Biton, Y. Cohen, R. Halperin, D. Schneider, A. Oron, Z. Nevo, Eur. Spine J. 2006, 15, 234.
© 2015 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Healthcare Mater. 2015, 4, 1861–1868
Elisa López-Dolado, Ankor González-Mayorga, María Teresa Portolés, María José Feito, María Luisa Ferrer, Francisco del Monte, María Concepción Gutiérrez, and María Concepción Serrano*
FULL PAPER
Subacute Tissue Response to 3D Graphene Oxide Scaffolds Implanted in the Injured Rat Spinal Cord
of neurotrophic factors, and biomaterials, among others. In this context, materials with advanced properties could offer novel perspectives to serve helping this aim. One of these promising materials is graphene, a free-standing single atomic plane of graphite with excellent characteristics that are strongly impelling these days its application in biomedical areas including bioimaging, drug delivery, and photothermal therapy for cancer.[3–6] More recently, other derivatives such as graphene oxide (GO) or partially reduced GO (rGO) are also attracting significant attention as they combine some of the properties of graphene (e.g., mechanical ductility and stiffness), with other interesting features such as a higher hydrophylicity and versatility than graphene for functionalization with biological moieties.[7,8] As described for graphene,[9,10] composites containing GO nanosheets can enhance neural differentiation[11–13] and promote axonal alignment.[14] Additionally, graphene-derived materials (GDM) do not only serve as a platform to support and direct neural growth, but also to assist the fabrication of interfaces for recording and stimulating neural electrical signals.[15–17] The toxicity and biocompatibility in vivo of GDM remain an open debate.[18] Nonetheless, published results support the existence of a safe range of conditions in which these materials might not initiate any toxic responses in biological systems.[19–21] In this sense, studies in mice focused on the intravenous administration of GO nanomaterials have evidenced a clear dose-dependent toxicity with no pathological changes or damage in major organs (e.g., lung, liver, kidney, spleen) at doses as low as 1 mg kg−1 in 14 d.[22] Results at higher doses are contradictory and depend on the experimental conditions.[23,24] Studies on oral,[25] intravitreal,[26] intraperitoneal,[25] and subcutaneous routes refer no major toxic responses.[27] On the contrary, intratracheal administration typically generates some degree of lung injury.[28,29] Regarding the interaction of GDM with immune cells, some early attempts in Caenorhabditis elegans, an alternative nonmammalian organism for toxicity screening, have evidenced a key role played by the innate immunity in the regulation of the chronic toxicity in vivo of GO.[30] Importantly, it is known
The increasing prevalence and high sanitary costs of lesions affecting the central nervous system (CNS) at the spinal cord are encouraging experts in different fields to explore new avenues for neural repair. In this context, graphene and its derivatives are attracting significant attention, although their toxicity and performance in the CNS in vivo remains unclear. Here, the subacute tissue response to 3D flexible and porous scaffolds composed of partially reduced graphene oxide is investigated when implanted in the injured rat spinal cord. The interest of these structures as potentially useful platforms for CNS regeneration mainly relies on their mechanical compliance with neural tissues, adequate biocompatibility with neural cells in vitro and versatility to carry topographical and biological guidance cues. Early tissue responses are thoroughly investigated locally (spinal cord at C6 level) and in the major organs (i.e., kidney, liver, lung, and spleen). The absence of local and systemic toxic responses, along with the positive signs found at the lesion site (e.g., filler effect, soft interface for no additional scaring, preservation of cell populations at the perilesional area, presence of M2 macrophages), encourages further investigation of these materials as promising components of more efficient material-based platforms for CNS repair.
1. Introduction Lesions at the spinal cord in the central nervous system (CNS) have become a major social concern because of their increasing prevalence and the high sanitary costs associated.[1] These limitations are functioning as a fuel to encourage experts in different fields to explore new avenues for neural regeneration,[2] including pharmacological approaches, cell therapy, delivery
Dr. E. López-Dolado, A. González-Mayorga, Dr. M. C. Serrano Hospital Nacional de Parapléjicos (SESCAM) Finca de La Peraleda s/n, 45071 Toledo, Spain E-mail: [email protected] Prof. M. T. Portolés, Dr. M. J. Feito Department of Biochemistry and Molecular Biology I Universidad Complutense de Madrid Ciudad Universitaria s/n, 28040 Madrid, Spain Dr. M. L. Ferrer, Dr. F. del Monte, Dr. M. C. Gutiérrez Instituto de Ciencia de Materiales de Madrid (ICMM) Consejo Superior de Investigaciones Científicas (CSIC) C/ Sor Juana Inés de la Cruz 3, 28049 Madrid, Spain
DOI: 10.1002/adhm.201500333
Adv. Healthcare Mater. 2015, 4, 1861–1868
© 2015 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
wileyonlinelibrary.com
1861
www.advhealthmat.de
FULL PAPER
www.MaterialsViews.com
that the cross-talk between neural stem and immune cells is critical for inducing reparative responses.[31] To this regard, specific studies in the spinal cord by Popovich and co-workers have revealed the existence of distinct macrophage subsets with pivotal implications in neurotoxic versus neuroregenerative responses.[32] In this sense, the proinflammatory (M1) versus immunomodulatory/reparative (M2) balance has been ascribed to the capacity of macrophages to play both positive and negative roles in disease processes and tissue remodeling after injury.[33,34] Here, we investigate for the first time the subacute tissue responses to 3D rGO scaffolds implanted in the injured rat spinal cord (C6 level). The interest of these structures mainly relies on their excellent mechanical properties. Particularly, mechanical compression studies demonstrated their remarkable flexibility in both longitudinal and transversal directions,[35] thus becoming a soft material interface that is mechanically compatible with the nervous tissue (0.3–1.0 kPa).[36] Values ca. 1 kPa have been reported for comparable GO foams also prepared by freeze-drying when exposed to strains under 10%.[37] This attractive feature, along with their biocompatibility demonstrated with neural cell cultures in vitro,[35] has boosted the exploration of their behavior in vivo to anticipate advantages and/or limitations concerning their potential use as a guiding platform for neural regeneration in the CNS. Any interest from the electrical point of view is abolished by their dramatically low conductivities (ca. 10−3 S cm−1, unpublished data measured as previously described).[38] In this work, subacute responses have been thoroughly investigated both local and systemically. Due to the pivotal relevance of immune cross-talk on neural repair, the expression of macrophage surface markers CD80 and CD163 is also investigated to identify M1 and M2 cells, respectively, in contact with the rGO implant.
2. Results and Discussion Encouraged by the attractive properties of GDM and their interest for neural repair in the shape of 3D substrates,[39,40] we herein investigate the tissue response of the injured rat spinal cord to 3D porous rGO scaffolds after 10 d of implantation, defined as a subacute phase in rat models.[41] Further details on the physicochemical characterization of both the GO used and the resulting rGO scaffolds have been published elsewhere.[35] Figure 1A displays sequential schemes of the spinal cord injury practiced (for further details, please refer to the Experimental Section). Treatment groups included control without injury (C), injury (I) and injury + scaffold (I+SC). Images of a representative rGO scaffold implanted at C6 level and the posterior coverage of the injured area with a thin gelatin hydrogel film (Figure 1B,C, respectively) are also displayed. The average amount of rGO implanted per animal was 0.25 mg, a quantity that is within the safe range of GO concentrations described as nontoxic in previous reports in mice.[22,23] As evidenced from systematic post-operatory (PO) observations, all injured animals no matter the treatment received regained interest for actively exploring the environment and showed an adequate hydration state, absence of pain or distress signs, and proper food intake and defecation after 3–4 d post-surgery. An evident
1862
wileyonlinelibrary.com
Figure 1. Scheme of the surgery model practiced (A). Representative images of an rGO scaffold implanted at the right spinal hemicord at C6 (B), posteriorly covered with a thin gelatin hydrogel (C). Injured rats carrying rGO implants for 7 d displayed two apparent phenotypes: close fist (D) and open hand (E). Ventral view of a representative 10-d explanted spinal cord (F). Set of arrows indicates: C—Caudal, L—Left, R—Right, and Ro—Rostral. Dashed red line in (F) marks the midline of the spinal cord.
paresis of the anterior right limb was referred in all cases, with rats displaying two slightly different phenotypes, “close fist” (Figure 1D) and “open hand” (Figure 1E), thus indicating a small variation in the lesion epicenter size that, in the second phenotype, comprised finger and wrist muscles besides elbow extensors. When sacrificed at 10 PO days (POD), rGO scaffolds were found properly allocated at the injured site inside the spinal cord where initially placed, with evident scaffold adhesion to both the adjacent neural tissue in the spinal cord and deep muscle layers on top. Macroscopic examination of the explants also revealed the accomplishment of nearly complete hemisections at the right C6 spinal segment, as confirmed in ventral views of the injured spinal cord (Figure 1F). We next focused on the specific tissue response of the injured spinal cord. Conventional hematoxylin-van Giesson (HvG) staining confirmed a complete adhesion of the rGO scaffolds implanted to the surrounding neural tissue inside the injured spinal cord, without significant cavities (Figure 2, right column) and with an evident tissue interface. The initial criterion to define the borders of such an interface was based on these HvG-stained sections, posteriorly confirmed as distinguishable on immune-stained ones. Prior to implantation, these 3D rGO scaffolds typically present a porosity of ca. 80%, as estimated from scanning electron microscopy (SEM) images.[35] This feature, along with their characteristic softness, supported the infiltration of cells and collagen fibers (Figure 2, bottom right). As characteristic in CNS lesions,[42] most of these cells probably migrated by chemotactic phenomena from surrounding tissues and blood vessels. It is worth noting that GO could also partially contribute to this effect due to its described role as a nonspecific enhancer of mammalian cell growth.[43] Remaining free space in the implanted scaffolds will permit further colonization by cells and extracellular matrix components, thus compelling with the prerequisite of tissue perfusion
© 2015 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Healthcare Mater. 2015, 4, 1861–1868
www.advhealthmat.de www.MaterialsViews.com
FULL PAPER
I+SC groups. Figure 5 shows similar stainings for the interface around the scaffold in I+SC. Quantification of all these markers in gray matter areas is included in Figure 6 and 2S and Table S1 (Supporting Information). Positive stained area for cell nuclei in the proximities of the lesion (at 0–1 and 1–2 mm from the lesion borders) in both I and I+SC groups was similar to that found on healthy nervous tissue in C rats (Figure 6A). At the injury site in I and I+SC, the amount of cells significantly increased compared to that of perilesional and no lesion areas likely indicating the activation of both cell migration and proliferation from surrounding tissues, which is characteristic of subacute CNS lesions.[42] The interface displayed a similar Figure 2. Histological examination of the implantation site at 10 POD by HvG staining. Images at the bottom row represent zoom-in details of areas marked with orange squares in cell density to that found at the lesion in I top images. Spinal cords are oriented in all cases as indicated by the set of arrows: C—Caudal, group, both significantly inferior to scafD—Dorsal, Ro—Rostral, and V—Ventral. folds in I+SC maybe related to the ability of GO to promote cell proliferation.[43] Markers for neurons (e.g., map-2 and tau) were highly present outside that is pursued in biomaterials aimed to have scaffold functions the lesion epicenter (map-2 mainly in gray matter areas, tau in (i.e., supportive and guiding properties).[2] The tissue forming both white and gray matter ones) and slightly at the interface, the interface in the I+SC group resembled that found filling while totally absent inside the injured area (Figure 6B and S2, the cavity in I (Figure 2, middle column), with some small Supporting Information). On the contrary, the positive reacpieces of damaged neural tissue not properly extracted after tivity found for vimentin (potentially contributed by glial cells, lesion found in the injured area in both I and I+SC groups. pericytes, and connective tissue cells, among others) revealed Morphometric analyses demonstrated that the average lesion a significant distance-dependent gradient, thus being very low volume was 7.6 ± 2.2 mm3, with 3.7 ± 1.6 mm3 contributed by in control and perilesional nervous tissue at 1–2 mm, sigthe scaffold and 3.9 ± 1.2 mm3 by the interface tissue formed nificantly higher at 0–1 mm from the lesion and even higher around the scaffold (percentage of interface at the lesion site: in injured areas of both I and I+SC (scaffold and interface) 52.8 ± 14.7 %) (Figure 3A). Interestingly, interface width around the scaffold did not show significant differences on the different groups. Interestingly, at closer perilesional areas (0–1 mm), regions (i.e., rostral, ventral, and caudal, p = 0.393) (Figure 3B). vimentin expression was higher in rostral than caudal zones HvG images confirmed that this interface tissue was composed likely indicative of a higher retrograde than anterograde of both cells and collagen fibers (Figure 3C). Finally, no indicaneural damage. Posterior immunofluorescence studies also tions of scaffold degradation were observed at 10 POD. Further showed highly positive staining for ED1 (typically expressed studies at the chronic state will necessarily serve to address this by macrophages) and platelet-derived growth factor receptor β aspect at longer time points. (PDGFRβ), recognized as an essential regulator of early hematTo better characterize the effect of rGO scaffold implantaopoiesis and blood vessel formation,[44] at the injured areas of tion at the lesion site, immunofluorescence studies of specific both I and I+SC groups. These two markers were rare in the markers were carried out to identify the nature of both the cells control and perilesional nervous tissue, being PDGFRβ+ cells infiltrating the scaffold and those forming the interface tissue mainly grouped at the perivascular regions (i.e., pericytes). The in comparison to no lesion areas. Figure 4 illustrates representdramatically positive reactivity for PDGFRβ+ cells at the injury ative images of no lesion areas in C and lesion sites for I and site is not surprising as this receptor is a typical marker of type
Figure 3. Examination of the interface tissue in contact with the implanted rGO scaffold at 10 POD. Percentage of interface versus scaffold in the injured spinal cord (A), in which Y axis indicates consecutive sections (S) of a representative I+SC spinal cord sample. The interface tissue formed displayed a similar width at rostral, ventral, and caudal regions (B). A representative HvG-stained section illustrating the interface tissue formed around the rGO scaffold implanted (C). Yellow dashed lines limit the borders of the interface.
Adv. Healthcare Mater. 2015, 4, 1861–1868
© 2015 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
wileyonlinelibrary.com
1863
www.advhealthmat.de
FULL PAPER
www.MaterialsViews.com
Figure 4. Representative immunofluorescence images of the lesion site at 10 POD. Control samples are displayed on top for comparison. Bright-field images are included to confirm location of the stained cells inside the scaffold. Cell nuclei were stained with Hoechst (hoe). Scale bars: 100 µm.
sent at the interface or infiltrating the 3D scaffold structure, being slightly more abundant in the first one. This result is not surprising as both neurotoxic and neuroregenerative effects can be elicited concurrently at the injury site.[46] As can be noticed, only a reduced fraction of the ED1+ population previously identified at the injury site (Figure 4, 5) stained positive for these markers, likely due to the fact that macrophage phenotype occurs in vivo along a continuum between M1 and M2 ends.[47] Contrarily to these results, no evidence of these cells was found in perilesional areas. The slight reduction of macrophage population inside the scaffold might be related to some specific toxic effects driven by rGO on this cell type, as recent work has revealed that GO is capable of inducing macrophage necrosis via Toll-like receptor 4 even in the absence of GO phagocytosis.[48] A significant percentage of actively proliferating cells was also identified at the injury site in both I (12.2 ± 5.4%) and I+SC (12.3 ± 3.5% in the scaffold and 15.4 ± 5.6% at the interface) groups, in contrast with the reduced amount found in C rats (1.6 ± 2.8%, p = 0.025) (Figure 7, bottom row). This finding indicates that some of the cells filling the cavity created at the injury site and infiltrating the scaffold, which we hypothesized were recruited from circulation and surrounding tissues by chemotactic effects, retained an actively proliferative phenotype. Motivated by the local effects found at the injured spinal cord and alerted by previously reported work on graphene toxicity, we then examined the major organs (e.g., brain, heart, intestine, kidney, liver, lungs, and spleen), searching for any signs of systemic toxicity induced by rGO scaffold implantation. After a gross inspection, organs did not show any Figure 5. Representative immunofluorescence images of the interface tissue in contact with major macroscopic abnormalities in size, color, morphology, or texture. Organ weight the scaffold at 10 POD. Cell nuclei were stained with Hoechst (hoe). Bright-field images are in I and I+SC groups remained unaltered at included to confirm location of the stained cells relative to the scaffold. A pericytes. These cells are involved in the formation of the core of the spinal cord scar tissue,[45] as well as in the growth of connective tissue cells and collagen production characteristic of wound healing processes,[44] all these phenomena commonly taking place at CNS injuries.[42] The expression of glial fibrillary acidic protein (GFAP), a specific marker of astrocytes, also followed a distance-dependent gradient as previously indicated for vimentin (i.e., higher GFAP expression at locations closer to the lesion). However, and contrary to vimentin, this marker was almost negligible at the lesion epicenter in both I and I+SC. Despite this border effect in the injury site, a discrete internalization of GFAP+ cells into the scaffold structure was observed in three out of six animals, with migration distances that overpassed 200 µm (up to 325 µm in one case) (Figure S3, Supporting Information). Figure 7 illustrates positive staining for macrophages expressing either CD80 (top row) or CD163 (middle row) from a representative I+SC rat. Both types of macrophages were pre-
1864
wileyonlinelibrary.com
© 2015 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Healthcare Mater. 2015, 4, 1861–1868
www.advhealthmat.de www.MaterialsViews.com
10 POD when compared to C rats (Figure S4, Supporting Information). However, the percentage of total weight loss per animal was found highly variable among individuals, but independent from the type of treatment received (I: 15.06 ± 7.20 %; I+SC:
Figure 7. Immunofluorescence studies of macrophages expressing either CD80 (top) or CD163 (middle) at 10 POD in perilesional areas, at the interface or inside the scaffold in I+SC rats. Scale bars: 25 µm. Bottom row images illustrate the presence of proliferating cells stained with Ki67 in C (left), I (middle) and I+SC (right) rats at 10 POD. Staining with Hoechst (blue) is merged to facilitate proper cell identification. Scale bars: 100 µm.
Adv. Healthcare Mater. 2015, 4, 1861–1868
© 2015 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
wileyonlinelibrary.com
FULL PAPER
Figure 6. Quantitative analyses of positive stained areas for A) cell nuclei and B) specific immunofluorescence markers in the injured spinal cord at 10 POD. Statistically significant differences are indicated in comparison to: a) no lesion (control) and perilesional areas, b) I, c) interface in I+SC (IF), and d) scaffold in I+SC (SC) groups.
16.01 ± 7.23 %; p = 0.824). However, some degree of correlation was found with the size of the lesion (y = 0.27x + 2.78; R2 = 0.6955), being larger lesions prompted to cause a higher weight loss. We then proceeded with a thorough inspection of those organs typically involved in material degradation, elimination, and accumulation (i.e., kidney, liver, lung, and spleen).[22,24] In agreement with previous work on GO-based nanomaterials inoculated oral, intraperitoneal or intravenously[24,25,49] and as expected from the reduced amount implanted,[22,23] no significant damage was observed in any of the major organs evaluated, including those of the reticuloendothelial system (i.e., liver and spleen). Transmission optical microscopy analyses revealed no alteration of normal tissue structure, inflammation, fibrosis, atrophy, or atypical cell responses in any analyzed section of the target viscera at either 10× images (Figure 8) or higher magnification ones (40×) (Figure S5, Supporting Information). Taking together, these results indicate that the implantation of rGO scaffolds at the injured rat spinal cord does not induce any significant subacute toxic responses either at any of the major organs investigated or perilesionally. At the lesion site, these structures have demonstrated: 1) A clear filler effect that allows to regain tissue integrity after injury at earlier time points, 2) A soft interface that facilitates accommodation into the injured site preventing additional scaring by friction,[50] 3) 3D scaffold colonization by a majority of PDGBRβ+ cells contributing to the stabilization of the injury and the prevention of further damage, and 4) Infiltration of reparative M2 macrophages, which could eventually prompt suitable repair responses. In this particular sense, the potential modulation of immune cell function by regulating the M1/M2 balance suggests the possibility of designing biomaterials capable of eliciting appropriate immune responses at the implantation site.[51] Future work will aim to direct macrophage differentiation towards the M2 phenotype by, for instance, releasing interleukin-4,[32] as already explored for peripheral nerve repair.[52] Studies targeting PDGBRβ+ cells present at the lesion site will be also of enormous interest due to their potential relationship with pericytes and ependymal cells and their implication in spinal cord scar tissue.[45] For all these studies, it will be critical to achieve the correct redox state of the GO used, as its degree of stacking and electrostatic/ hydrogen bonding is responsible for growth factor binding and its posterior influence on cell responses including differentiation.[53] From the mechanical point of view, it is known that the glial response around CNS implanted devices strongly depends on their mechanical properties—the more compliant the material is, the less inflammatory glial response encapsulating the device is formed.[54] Therefore, further studies at longer implantation times will take advantage of the mechanical compliance and ability to absorb compressions of these scaffolds to better guide neural growth by defining optimized guidance channels decorated with both topographical and biological cues. Since the topographical and structural features of these 3D rGO structures appeared insufficient, effective cocktails for their biofunctionalization will be explored to stimulate pro-regenerative elements at the scar, benefit neural cell growth and axonal sprouting through the lesion site and then aid the re-establishment of functional connections. Alternatively, 3D porous structures composed of biocompatible and biodegradable polymers such as poly(lactic-co-glycolic acid) and polycaprolactone, both
1865
www.advhealthmat.de
FULL PAPER
www.MaterialsViews.com
1866
immersion. Scaffold morphology was characterized by SEM (Hitachi S-4700 SEM microscope) and confirmed similar to that previously described.[35] Gelatin (Batch#1189632-52205137, porcine skin) hydrogel was dialyzed in distilled water and further dissolved at 50 °C. The final solution was then freeze-dried and stored at 4 °C until used. Fresh hydrogel films (9 mm2, ca. 100 µm in thickness) were prepared by dissolving the freeze-dried product in sterile distilled water (2 wt%) at 37 °C. Finally, the resulting films were sterilized under UV radiation and hydrated with saline solution 5 min prior implantation. Surgical Procedures: Adult male wistar rats were provided by a commercial supplier (Harlan Ibérica, Spain) and used at the age of ca. 30 weeks (n = 16, 451 ± 67 g). The use of aged adult rats was motivated by two reasons: 1) A larger size facilitating surgical procedures and 2) the incidence of spinal cord injury rising in older adults and the Figure 8. Histological examination of major organs at 10 POD. Representative 10× optical elderly human population nowadays.[57] Although microscopy images of kidney, liver, lung, and spleen after HE staining. Control samples are the role of age in recovery and mortality after spinal displayed for comparison. cord injury is not completely clarified yet, some differences had been shown in the inflammatory [58] morphometric CNS tissue size after lesion[59] and cell response, materials extensively explored for nervous tissue regeneration, functional locomotor deficits at the chronic states.[60] The lesion model and benefitted from GDM coatings could be also an interesting of choice was a right lateral hemisection of approximately 8 mm3 avenue to explore in this scenario. (2 × 2 × 2) at the C6 segment, rostral to the bulk of triceps brachii motoneurons, previously demonstrated as a suitable model to evaluate therapeutic strategies aimed at promoting neural plasticity and repair. As it causes focal foreleg impairments attributable to the neural damage 3. Conclusion caused, both the impairments and their eventual recovery can be evidenced and measured in detail during locomotion by using careful We have implanted for the first time 3D porous and flexible functional and anatomical techniques.[61] Some other advantages of rGO scaffolds at the injured rat spinal cord and explored the this model include: a) higher clinical relevance, since most human subacute tissue response generated. Results obtained evidence spinal cord injuries occur in the cervical area and restoration of arm the capacity of these structures to facilitate regaining tissue and hand function is a top priority for improving the quality of life of tetraplegic patients;[62] b) the proximity of cervical motor nuclei to the integrity after spinal cord injury as early as 10 d and to prelesion, so even short distance axonal growth could improve forelimb vent the extension of the lesion, although unable to guide function; and c) regarding ethical considerations and requirements neural growth through it. Absence of local and systemic toxic of animal care, this type of lesion typically causes damage in the responses, along with promising features at the injury site such ipsilateral forelimb with minimal contralateral forelimb and hindlimb as the establishment of a soft interface preventing additional sensorimotor impairments or autonomic dysfunction. All experimental scaring and the presence of reparative cells such as M2 macprotocols adhered to the regulations of the European Commission (directives 2010/63/EU and 86/609/EEC) and the Spanish government rophages, encourage further investigation of GDM as prom(RD53/2013) for the protection of animals used for scientific purposes. ising components of more efficient platforms for the treatment Specific details on the surgical procedures carried out are provided in of spinal cord injury. the Supporting Information. In those animals carrying rGO scaffolds, monoliths of appropriate dimensions were implanted at the lesion site to fill in the cavity created. All lesions, with or without scaffolds, were covered with a thin gelatin hydrogel film to serve as a biological and mechanically compatible interface between the injured nervous 4. Experimental Section tissue and the deep muscle layers. From the surgical point of view, this Additional experimental details are available in the Supporting hydrogel film assisted the preservation of the scaffold position inside Information. the cavity practiced in the spinal cord during the first stages postMaterial: Chemical reagents and antibodies were purchased from surgery without causing mechanical mismatch with the surrounding Sigma–Aldrich and used as received, unless otherwise indicated. native tissues. The use of this gelatin hydrogel is supported by previous Preparation of 3D Porous rGO Scaffolds and Gelatin Hydrogel Films: work demonstrating the interest of this polymer as part of regenerative 3D rGO scaffolds were fabricated by using the ice segregation-induced strategies at the injured spinal cord.[63,64] Treatment groups included: [ 38,55 ] [ 35 ] self-assembly (ISISA) technique as previously described. The C (n = 3), I (n = 6) and I+SC (n = 6). An exhaustive PO care protocol ISISA procedure is a simple and versatile, bottom-up self-assembly was applied after surgery, with major attention placed into signs of pain, methodology based on the unidirectional freezing of gels or colloidal distress, dehydration, intestinal obstruction, and respiratory failure. suspensions and their subsequent liophilization.[56] Details on the To assess the early tissue response, rats were sacrificed at 10 POD by physicochemical characterization of the GO powder used as well as a standard perfusion-fixation protocol (further details in Supporting [ 35 ] of the obtained rGO scaffolds were published elsewhere. Prior to Information). implantation, monoliths (4.5 mm in diameter, ca. 2 mm in height) Necropsy Protocol for Thoracoabdominal Viscera and CNS Tissue: were sterilized under UV radiation and then infiltrated with tissue Immediately after the perfusion-fixation protocol, rats were placed culture-grade water under vacuum conditions until complete scaffold on their back with the thoracic cavity opened and the different
wileyonlinelibrary.com
© 2015 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Healthcare Mater. 2015, 4, 1861–1868
www.advhealthmat.de www.MaterialsViews.com
Supporting Information Supporting Information is available from the Wiley Online Library or from the author.
Adv. Healthcare Mater. 2015, 4, 1861–1868
Acknowledgements E.L.D. and A.G.M. contributed equally to this work. This work was supported by the Instituto de Salud Carlos III (ISCIII) and Ministerio de Economía y Competitividad (MINECO) (Grant CP13/00060), co-financed by FEDER funding, and MINECO (Grants MAT2011-25329 and MAT2013-43229-R). AGM and MCS acknowledge ISCIII-MINECO for respective contracts associated to project CP13/00060. Authors express their gratitude to Dr. Collazos-Castro for fruitful discussions.
FULL PAPER
thoracoabdominal viscera progressively extracted (further details in Supporting Information). Then, rats were bent over to obtain the CNS tissue (i.e., brain and spinal cord). After extraction, all organs were weighted, carefully examined looking for any signs of gross damage and properly prepared for subsequent histopathological and immunofluorescence procedures. Organ and Neural Tissue Processing: After the perfusion, all samples were placed in paraformaldehyde 4% at 4 °C overnight and then 3 d in sucrose (30% in phosphate buffer saline, PBS) at 4 °C for cryoprotection. C5–C7 spinal cord segments were selected for further analysis. Tissue pieces were mounted on plastic containers, quick-frozen in optimal cutting temperature compound (Tissue Tek, Hatfield, PA) and cut in horizontal sagittal sections of 10-µm by using a Microm HM550 cryostat with an angle of 10°. In the particular case of the spinal cords, cuts were carried out throughout the entire C5–C7 fragment following the sagittal direction from right to left. Histology and Morphometric Analyses: Spinal cords and organs were initially examined after conventional HvG and hematoxylin-eosin (HE) stainings, respectively. Digital images were collected with an Olympus BX51 fluorescence microscope with a digital camera coupled. All images were calibrated and processed by using ImageJ (1.47v, National Institutes of Health, USA). For each animal, the size and extension of the lesion in the rostrocaudal direction were measured by drawing the contours of the total lesion, the interface and the scaffold on calibrated 2× images of entire HvG series. The 3D reconstruction of the lesion was completed by estimating the contribution from the remaining nine series in the right–left direction. Specifically, area values (µm2) for total lesion, interface, and scaffold in each section were multiplied by 10 µm (i.e., section thickness) and 10 (i.e., total number of series for each spinal cord). The total size of the lesion was then obtained as the summation of all these volumes and expressed in mm3. Immunofluorescence Studies: Spinal cord samples were examined for the presence of the following markers: 1) map-2 for somas and dendrites in neurons, 2) tau for axons in neurons, 3) vimentin for non-neuron cells including glial cells and connective tissue cells, 4) GFAP for astrocytes, 5) ED1 for macrophages, 6) PDGFRβ for pericytes, precursors of oligodendrocytes and connective tissue cells such as fibroblasts and smooth muscle cells,[44,45] 7) CD80 for M1 macrophages, 8) CD163/ M130 for M2 macrophages, and 9) Ki67 for actively proliferating cells. Appropriate secondary antibodies were selected. Further details of the specific antibodies and conditions used are provided in the Supporting Information. Cell nuclei were labeled with Hoechst (1 mg mL−1). Digital images were collected with a fluorescence microscope with a digital camera coupled by using appropriate filters. Figure S1 (Supporting Information) illustrates the different areas of study, including control in C rats (no lesion), perilesional areas at 1–2 and 0–1 mm from the lesion site in I and I+SC groups (both rostral and caudally) and lesion areas in I and I+SC. Additionally, bright-field optical microscopy images were captured to define cell location relative to the scaffold. Images were calibrated and processed by using the ImageJ software. For each fluorescent marker, positive stained areas in at least 10 images were measured, expressed as a percentage of the total image area and averaged. Statistics: Values were expressed as mean ± standard deviation of at least three different animals per group (n ≥ 3). Statistical analysis was performed by using the Statistical Package for the Social Sciences software (SPSS, version 17.0). Comparisons among groups were done by analysis of variance (ANOVA) and either post-hoc Scheffé or GamesHowell tests (for either homogeneous or heterogeneous variances, respectively). The significance level was defined as p < 0.05 in all statistical evaluations.
Received: May 1, 2015 Revised: June 1, 2015 Published online: June 25, 2015
[1] V. Y. Ma, L. Chan, K. J. Carruthers, Arch. Phys. Med. Rehabil. 2014, 95, 986. [2] J. K. Hyun, H. W. Kim, J. Tissue Eng. 2010, 1, 650857. [3] D. Bitounis, H. Ali-Boucetta, B. H. Hong, D. H. Min, K. Kostarelos, Adv. Mater. 2013, 25, 2258. [4] Y. Yang, A. A. Asiri, Z. Tang, D. Du, Y. Lin, Mater. Today 2013, 16, 365. [5] H. Zhang, G. Grüner, Y. J. Zhao, Mater. Chem. B 2013, 1, 2542. [6] L. Feng, L. Wu, X. Qu, Adv. Mater. 2013, 25, 168. [7] X. Tan, L. Feng, J. Zhang, K. Yang, S. Zhang, Z. Liu, R. Peng, ACS Appl. Mater. Interfaces 2013, 5, 1370. [8] B. J. Hong, O. C. Compton, Z. An, I. Eryazici, S. T. Hguyen, ACS Nano 2012, 6, 63. [9] S. Y. Park, J. Park, S. H. Sim, M. G. Sung, K. S. Kim, B. H. Hong, S. Hong, Adv. Mater. 2011, 23, H263. [10] Y. Wang, W. C. Lee, K. K. Manga, P. K. Ang, J. Lu, Y. P. Liu, C. T. Lim, K. P. Loh, Adv. Mater. 2012, 24, 4285. [11] O. Akhavan, E. Ghaderi, E. Abouei, S. Hatamie, E. Ghasemi, Carbon 2014, 66, 395. [12] D. Yang, T. Li, M. Xu, F. Gao, J. Yang, Z. Yang, W. Le, Nanomedicine 2014, 9, 2445. [13] S. Shah, P. T. Yin, T. M. Uehara, S. T. D. Chueng, L. Yang, K. B. Lee, Adv. Mater. 2014, 26, 3673. [14] A. Solanki, S. T. D. Chueng, P. T. Yin, R. Kappera, M. Chhowalla, K. B. Lee, Adv. Mater. 2013, 25, 5477. [15] X. Duan, T. M. Fu, J. Liu, C. M. Lieber, Nano Today 2013, 8, 351. [16] E. K. Wujcik, C. N. Monty, Wiley Interdiscip. Rev. Nanomed. Nanobiotechnol. 2013, 5, 233. [17] T. H. Kim, K. B. Lee, J. W. Choi, Biomaterials 2013, 34, 8660. [18] A. B. Seabra, A. J. Paula, R. de Lima, O. L. Alves, N. Durán, Chem. Res. Toxicol. 2014, 27, 159. [19] A. M. Pinto, I. Gonçalves, F. D. Magalhães, Colloid. Surf. B 2013, 111, 188. [20] X. Hu, Q. Zhou, Chem. Rev. 2013, 113, 3815. [21] C. Bussy, H. Ali-Boucetta, K. Kostarelos, Acc. Chem. Res. 2013, 46, 692. [22] X. Zhang, J. Yin, C. Peng, W. Hu, Z. Zhu, W. Li, C. Fan, Q. Huang, Carbon 2011, 49, 986. [23] K. Wang, J. Ruan, H. Song, J. Zhang, Y. Wo, S. Guo, D. Cui, Nanoscale Res. Lett. 2011, 6, 8. [24] K. Yang, J. Wan, S. Zhang, B. Tian, Y. Zhang, Z. Liu, Biomaterials 2012, 33, 2206. [25] K. Yang, H. Gong, X. Shi, J. Wan, Y. Zhang, Z. Liu, Biomaterials 2013, 34, 2787. [26] L. Yan, Y. Wang, X. Xu, C. Zeng, J. Hou, M. Lin, J. Xu, F. Sun, X. Huang, L. Dai, F. Lu, Y. Liu, Chem. Res. Toxicol. 2012, 25, 1265. [27] A. Sahu, W. I. Choi, G. Tae, Chem. Commun. 2012, 48, 5820.
© 2015 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
wileyonlinelibrary.com
1867
www.advhealthmat.de
FULL PAPER
www.MaterialsViews.com
1868
[28] M. C. Duch, G. R. Budinger, Y. T. Liang, S. Soberanes, D. Urich, S. E. Chiarella, L. A. Campochiaro, A. Gonzalez, N. S. Chandel, M. C. Hersam, G. M. Mutlu, Nano Lett. 2011, 11, 5201. [29] B. Li, J. Yang, Q. Huang, Y. Zhang, C. Peng, Y. Zhang, Y. He, J. Shi, W. Li, J. Hu, C. Fan, NPG Asia Mater. 2013, 5, e44. [30] Q. Wu, Y. Zhao, J. Fang, D. Wang, Nanoscale 2014, 6, 5894. [31] Z. Kokaia, G. Martino, M. Schwartz, O. Lindvall, Nat. Neurosci. 2012, 15, 1078. [32] K. A. Kigerl, J. C. Gensel, D. P. Ankeny, J. K. Alexander, D. J. Donnelly, P. G. Popovich, J. Neurosci. 2009, 29, 13435. [33] B. N. Brown, B. D. Ratner, S. B. Goodman, S. Amar, S. F. Badylak, Biomaterials 2012, 33, 3792. [34] P. M. Kou, J. E. Babensee, J. Biomed. Mater. Res. A 2011, 96, 239. [35] M. C. Serrano, J. Patiño, C. García-Rama, M. L. Ferrer, J. L. Fierro, A. Tamayo, J. E. Collazos-Castro, F. del Monte, M. C. Gutiérrez, J. Mater. Chem. B 2014, 2, 5698. [36] Y. B. Lu, K. Franze, G. Seifert, C. Steinhäuser, F. Kirchhoff, H. Wolburg, J. Guck, P. Janmey, E. Q. Wei, J. Käs, A. Reichenbach, Proc. Nat. Acad. Sci. U.S.A. 2006, 103, 17759. [37] J. Kuang, L. Liu, Y. Gao, D. Zhou, Z. Chen, B. Han, Z. Zhang, Nanoscale 2013, 5, 12171. [38] M. C. Serrano, S. Nardecchia, C. García-Rama, M. L. Ferrer, J. Collazos-Castro, F. del Monte, M. C. Gutiérrez, Biomaterials 2014, 35, 1543. [39] N. Li, Q. Zhang, S. Gao, Q. Song, R. Huang, L. Wang, L. Liu, J. Dai, M. Tang, G. Cheng, Sci. Rep. 2013, 3, 1604. [40] S. R. Shin, B. Aghaei-Ghareh-Bolagh, X. Gao, M. Nikkhah, S. M. Jung, A. Dolatshahi-Pirouz, S. B. Kim, S. M. Kim, M. R. Dockmeci, X. S. Tang, A. Khademhosseini, Adv. Funct. Mater. 2014, 24, 6136. [41] R. van den Brand, J. Heutschi, Q. Barraud, J. DiGiovanna, K. Bartholdi, M. Huerlimann, L. Friedli, I. Vollenweider, E. M. Moraud, S. Duis, N. Dominici, S. Micera, P. Musienko, G. Courtine, Science 2012, 336, 1182. [42] J. W. Fawcett, R. A. Asher, Brain Res. Bull. 1999, 49, 377. [43] O. N. Ruiz, K. A. Fernando, B. Wang, N. A. Brown, P. G. Luo, N. D. McNamara, M. Vangsness, Y. P. Sun, C. E. Bunker, ACS Nano 2011, 5, 8100. [44] P. H. Chen, X. Chen, X. He, Biochim. Biophys. Acta 2013, 1834, 2176. [45] C. Göritz, D. O. Dias, N. Tomilin, M. Barbacid, O. Shupliakov, J. Frisén, Science 2011, 333, 238.
wileyonlinelibrary.com
[46] J. C. Gensel, S. Nakamura, Z. Guan, N. van Rooijen, D. P. Ankeny, P. G. Popovich, J. Neurosci. 2009, 29, 3956. [47] D. M. Mosser, J. P. Edwards, Nat. Rev. Immunol. 2008, 8, 958. [48] G. Qu, S. Liu, S. Zhang, ACS Nano 2013, 7, 5732. [49] H. Ali-Boucetta, D. Bitounis, R. Raveendran-Nair, A. Servant, J. van den Bossche, K. Kostarelos, Adv. Healthcare Mater. 2013, 2, 433. [50] K. A. Potter, M. Jorfi, K. T. Householder, E. J. Foster, C. Weder, J. R. Capadona, Acta Biomater. 2014, 10, 2209. [51] S. Franz, S. Rammelt, D. Scharnweber, J. C. Simon, Biomaterials 2011, 32, 6692. [52] N. Mokarram, A. Merchant, V. Mukhatyar, G. Patel, R. V. Bellamkonda, Biomaterials 2012, 33, 8793. [53] W. C. Lee, C. H. Y. X. Lim, H. Shi, L. A. L. Tang, Y. Wang, C. T. Lim, K. P. Loh, ACS Nano 2011, 5, 7334. [54] P. Moshayedi, G. Ng, J. C. F. Kwok, G. S. H. Yeo, C. E. Bryant, J. W. Fawcett, K. Franze, J. Guck, Biomaterials 2014, 35, 3919. [55] A. Abarrategi, M. C. Gutiérrez, C. Moreno-Vicente, M. J. Hortigüela, V. Ramos, J. L. López-Lacomba, M. L. Ferrer, F. del Monte, Biomaterials 2008, 29, 94. [56] M. C. Gutiérrez, M. L. Ferrer, F. del Monte, Chem. Mater. 2008, 20, 634. [57] M. J. Hooshmand, M. D. Galvan, E. Partida, A. J. Anderson, Immun. Ageing 2014, 11, 15. [58] H. Kumamaru, H. Saiwai, Y. Ohkawa, H. Yamada, Y. Iwamoto, S. Okada, J. Cell. Physiol. 2012, 227, 1335. [59] A. C. Laing, R. Cox, W. Tetzlaff, T. Oxland, Anat. Rec. 2011, 294, 1326. [60] Y. S. Gwak, B. C. Hains, K. M. Johnson, C. E. Hulsebosch, J. Neurotrauma 2004, 21, 983. [61] E. López-Dolado, A. M. Lúcas-Osma, J. E. Collazos-Castro, J Neurotrauma 2013, 30, 191. [62] NSCISC National Spinal Cord Injury Statistical Center. 2013 Annual Statistical Report. University of Alabama, https://www.nscisc.uab.edu/reports.aspx (accessed: April 2015). [63] H. Altinova, S. Moellers, T. Fuhrmann, R. Deumens, A. Bozkurt, I. Heschel, L. H. H. O. Damink, F. Schugner, J. Weis, G. A. Brook, Brain Res. 2014, 1585, 37. [64] S. Rochkind, A. Shahar, D. Fliss, D. El-Ani, L. Astachov, T. Hayon, M. Alon, R. Zamostiano, O. Ayalon, I. Biton, Y. Cohen, R. Halperin, D. Schneider, A. Oron, Z. Nevo, Eur. Spine J. 2006, 15, 234.
© 2015 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Healthcare Mater. 2015, 4, 1861–1868
