Fabrication of homotypic neural ribbons as a multiplex platform optimized for spinal cord delivery

Cell therapy for the injured spinal cord will rely on combined advances in human stem cell technologies and delivery strategies. Here we encapsulate homotypic spinal cord neural stem cells (scNSCs) in an alginate-based neural ribbon delivery platform. We perform a comprehensive in vitro analysis and qualitatively demonstrate graft survival and injury site retention using a rat C4 hemi-contusion model. Pre-configured neural ribbons are transport-stable modules that enable site-ready injection, and can support scNSC survival and retention in vivo. Neural ribbons offer multifunctionality in vitro including co-encapsulation of the injury site extracellular matrix modifier chondroitinase ABC (chABC), tested here in glial scar models, and ability of cervically-patterned scNSCs to differentiate within neural ribbons and project axons for integration with 3-D external matrices. This is the first extensive in vitro characterization of neural ribbon technology, and constitutes a plausible method for reproducible delivery, placement, and retention of viable neural cells in vivo.


scNSCs differentiate into posterior CNS cells in vitro.
Under continued exposure to RA/Hh-Ag1.5, differentiating cultures expressed pMN-specific transcription factor OLIG2 by day 12 (Fig. 3a). By day 14, the neuronal nuclear antigen NeuN was observed to co-localize in TUJ1+ cells, suggesting an early stage of motor neuron progenitor. From this point, robust neurogenesis was observed, and differentiating neuronal cultures developed a step-wise biomarker profile characteristic of cervical/brachial SMNs (Fig. 3b) 27 . That is, the pMNspecific transcription factor Nkx-6.1, terminal SMN-related ISL-1&2, lateral motor column protein FOXP1, terminal SMN-specific HB9/MNX1, cholinergic neurotransmitter enzyme ChAT, and the nerve growth factor receptor NGFR/CD271/p75NTR. SMN maturation is supported by phosphorylation of high molecular weight (MWT) neurofilaments in axons using the SMI312 antibody cocktail (Fig. 3c), and Peripherin, an intermediate filament marker of the peripheral nervous system. We observed that hiPSC-derived SMNs show high affinity for rodent myotubes and assemble neuromuscular junctions (NMJs) in co-cultures (Fig. 3d) as assessed by colocalization of Synapsin 1 (SYN1; pre-synaptic terminals) with α-Bungarotoxin (αBTX) targeting nicotinic acetylcholine receptors (nAChR, post-synaptic NMJ terminal). We also identified O4+ oligodendrocyte progenitor cells (OPCs) and GFAP+ astrocytes in differentiating cultures (Fig. 3e), validating scNSC ability to produce multiple cell types of the spinal cord.
Co-encapsulation of scNSCs and chABC in multiplex neural ribbons. We established an alginate hydrogel-based neural ribbon system for rapid, reproducible encapsulation of scNSCs and co-encapsulation with www.nature.com/scientificreports/ the microenvironment regulator chondroitinase ABC (chABC; Fig. 4). Alginate chains crosslink via aggregation of guluronate blocks in the presence of divalent Ca 2+ ions allowing moldability into 3-D hydrogel networks 28 . In our system, chABC and scNSCs were combined with 1.5% sodium alginate and co-encapsulated by constrained needle tip extrusion into 100 mM CaCl 2 (Fig. 4a,b,d). Modified hydrogel ribbons were also investigated on the basis of interpenetrating network (IPN) formation with collagen to provide exogenous ECM and RGD-functionalized bioactive peptides for improved cell-cell interactions. We generated three separate ribbon compositions to achieve goal-directed structural and functional outcomes (Fig. 4c) that are ultrapure sodium alginate (ProNova) alone, ProNova alginate-collagen IPNs (NovaCol) and RGD-functionalized alginate-collagen IPNs (RGD-Col; Supplementary Table 3, 4). Hydrogel ribbons with specified cross-sectional area were templated by needle tip inner diameter. Three diameters were compared (60, 100, 150 µm) that are suitable for injection into a hemi-contusion-induced SCI cavity in rat models, and that each permit access of gases, nutrients and extracellular signals to cells at the neural ribbon core 29,30 . To enable ribbon visualization for transplantation studies, co-encapsulation of fluorescent indicators was done and signal retention monitored ( Supplementary Fig. 1). Biodegradation of alginate hydrogel was tested by exposure to artificial cerebrospinal fluid (aCSF) with daily solution changes and determination of the  www.nature.com/scientificreports/ Hydrogel ribbons containing chABC regulate external CSPG ECM. To incorporate chABC encapsulation and delivery capability into our neural ribbon technology, we used in-house purified recombinant P. vulgaris chABC enzyme generated using two bacterial systems that are gram-negative E. coli and gram-positive Bacillus megaterium. By UV-Vis digestion assays with chondroitin sulfate (CS) substrates, both purified enzymes performed as well as the standard commercial chABC product (Supplementary Table 7). We focused on Bacillus-purified recombinant chABC that is endotoxin-free for further analysis to extend chABC to animal studies. chABC activity was tested using multiple in vitro assays exploiting the endogenous production of CSPGs by scNSCs ( Supplementary Fig. 2) as well as acellular CSPG gradient spot assays for glial scar modeling ( Fig. 5) 31,32 . Despite the endogenous production of CSPG ECM, scNSCs are largely prohibited from penetrating the concentrated rim of CSPG inverse gradient spots. The CS-56 antibody enables visualization of both the CSPG spot and punctate intracellular CSPG signal. In the presence of chABC, CSPG signal is restricted to the perinuclear region of neural cells but does not impact scNSC proliferation or viability. Notably, expression of the functional inhibitory CSPG receptor LAR and PTPσ are similarly expressed during differentiation from scNSC to SMN ( Supplementary Fig. 2). We investigated in vitro modulation of the scNSC CSPG microenvironment by encapsulated release of chABC from neural ribbons ( Fig. 5; chABC-NR). CSPG inverse gradient spots were formed to mimic the glial scar, as previously described 31,32 . CSPG spots were completely degraded by 24 h after incubation with chABC-NR (0.1 U/ ml working concentration) in aCSF (Fig. 5a). By seeding scNSCs onto CSPG spots, we observed that the inverse concentration gradient is inhibitory to scNSCs despite endogenous CSPG production. Occasionally we observed a spontaneous cell-crossing event (~ 2 events/spot). The frequency of such events was significantly increased by incubation with chABC-NR (~ 22 events/spot; Fig. 5b, d; two-tailed t test, t = 9.87, df = 16, ****P < 0.0001; N = 9 CSPG spots counted per condition). Seeding of scNSC-derived SMNs onto CSPG spots also revealed a marked inability of axons to penetrate the CSPG inhibitory rim, while incubation with chABC-NR enabled axon migration and navigation within this region (Fig. 5c). Again, axon crossing events were significantly increased from 0 to ~ 12 events/spot by addition of chABC-NR ( Fig. 5e; t = 7.68, df = 8, ****P < 0.0001; N = 9 CSPG spots counted per condition). This effect of chABC-NR was replicated using plated SMN neurospheres that undergo robust axonal outgrowth (Fig. 5f). Finally, we investigated the ability of encapsulated chABC to modulate the inhibitory CSPG microenvironment using an scNSC-based assay ( Fig. 5g-i). Single Dextran-labeled chABC-NR were incubated with confluent scNSC cultures and embedded in Type 1 collagen matrix for immobilization (Fig. 5g). As a negative control for microenvironment modulation, the chABC-NR condition was compared to ribbons without chABC (Fig. 5h). chABC-NR co-culture resulted in a dramatic attenuation of extracellular scNSC CSPGs (0.1 U/ml working concentration) versus the negative control, and was quantified by CS-56 fluorescence intensity ( Fig. 5i; t = 9.63, df = 14, ****P < 0.0001).
Homotypic scNSCs survive and retain differentiation potential in neural ribbons. We confirmed and optimized the biocompatibility of hydrogel ribbons with scNSCs in terms of cell viability, proliferation, and multipotent differentiation potential (Fig. 6). Dissociation of day 10 scNSC rosettes into single cell suspensions and subsequent encapsulation (1 × 10 7 cells/ml) yields a disperse distribution of cells within the ribbons, particularly at 150 µm diameter (Fig. 6a). The encapsulated scNSCs retained viability (CellTracker) and metabolic activity (MTT assay) at 24 h (Fig. 6b), and dividing cells with mitotic figures were observed by DAPI and the Ki-67 antibody (Fig. 6c). A live-cell dual apoptosis reporter kit (488 NucView Caspase-3 substrate co-localization with TexasRed-Annexin V) was used to quantify apoptotic cells in adherent scNSC cultures (day 10) exposed to 100 mM CaCl 2 for durations relevant to ribbon initial formation and handling (no exposure, 1, 2, 5 min). The percentage of apoptotic cells was slightly reduced over time, although insignificantly ( Fig. 6d; one-way ANOVA, F = 0.058, R2 = 0.021, n.s. P = 0.98; n > 500 total cells counted per condition, average of three separate fields per condition). A Ca 2+ -mediated lift-off process may disproportionally affect loosely bound apoptotic cells. We next analyzed the effect of ribbon diameter on viability using Trypan blue exclusion on cells recovered from 60, 100 and 150 µm neural ribbons. Single cell suspension scNSCs encapsulated at high density (1 × 10 8 cells/ml; day 10) were recovered 24 h post-encapsulation by dissolving alginate ribbons in 1.6% sodium citrate ( Fig. 6e; see also Supplementary Table 3). RGD-Col neural ribbons yielded the highest viability of scNSCs, presumably due to improved cell adhesion and migration, but varied with diameter (one-way ANOVA, F = 19.6, **P = 0.0024; n = 300 cells from three ribbons analyzed per condition). Viability remained high in ProNova and NovaCol ribbons, but with no significant variation within groups by one-way ANOVA. Viability increased with decreasing ribbon diameter independent of ribbon composition. Using the higher cell loading concentration, neural ribbon quality and cell health improved overall. As well, loading of neural aggregates (~ 50-100 µm) that retain cell-cell interactions (Fig. 6f) provided superior results compared to single cell suspensions. Easy handling in culture medium allows for efficient transfer of neural ribbons to cryovials for long distance shipping (> 1,750 miles; 20-24 h laboratory-to-laboratory) in a portable incubator (Supplementary Fig. 1) enabling offsite in vivo studies. Finally, neural ribbons (1 × 10 8 cells/ml) cultured in suspension for five days in scNSC maintenance medium 5 retain multipotency and produce TUJ1+ neurons and GFAP+ astrocytes (Fig. 6g).

Neural ribbons extend neurites for 3-D matrix microenvironment integration.
We tested the ability of scNSCs to differentiate within NovaCol ribbons and integrate into external collagen matrices in vitro (Fig. 6h, i). We developed a neural ribbon analysis platform by immobilization in collagen gel in the viewing area of a glass bottom dish (Fig. 6h). This strategy optimizes imaging efficiency and statistical analysis. Encapsulated scNSCs were visualized by live cell imaging with Fluo-4 AM (NovaCol, 60 µm diameter), along with Cadherin IF (Fig. 6i). Incubation of the neural ribbon in CPM drives differentiation to MNPs and neuritogenesis. Robust lateral extension of neurites is visible by day 7 after encapsulation, extending orthogonally out of the ribbon body www.nature.com/scientificreports/ and into the surrounding collagen matrix (Fig. 6i). Average neurite length at day 7 as measured from the edge of the ribbon using projected 2-D images was 223 ± 34 µm (n = 54 neurites from N = 6 fields). Pan-Cadherin antibody staining reveals maintenance of extensive calcium-dependent cell-cell connections within the NovaCol hydrogel network even as early as day 1 after extrusion.
Neural ribbon low-dose scNSC survival and injury site integration in vivo. Day  www.nature.com/scientificreports/ of our extrusion system as well as of the contusion injury size. Two animals received ~ 5,000 cells injected in suspension with vehicle (1,000 cells/µl). Nine days after grafting (injury age day 23, scNSCs day 21), spinal cords were harvested and processed for histological analysis ( Fig. 7b-g). Serial longitudinal sections (sagittal plane) revealed that cells injected in suspension outside of neural ribbons were scarcely detected (Fig. 7b) and signal in the spared parenchyma resembled cellular debris (Fig. 7d). However, by injecting single neural ribbon segments, we observed that STEM121 immunopositive viable human scNSCs were retained within the contused injury cavity and began to extend processes (> 500 µm) along the cavity border (Fig. 7c,e,f). scNSCs that merged with host tissue at the cavity border also co-localized with Synapsin 1 nine days after grafting (Fig. 7g). These qualitative data demonstrate a neuroprotective function of the hydrogel platform, supporting the use of neural ribbons to position and retain cells in the SCI cavity even at low delivery doses.

Discussion
Neural ribbons containing homotypic, cervically-patterned NSCs constitute a protective multifunctional platform for cell delivery, retention, and positioning in SCI (Fig. 8). We apply advanced NSC differentiation protocols that focus on spinal cord identity, and along with the neural ribbons, result in enhanced cell survival and retention at the injury site. Several studies in rat, mini-pigs and non-human primates demonstrate the promise of NSC therapeutics but recognize challenges posed by the injury setting [33][34][35][36][37][38][39] . An ongoing human clinical trial with isolated human spinal cord-derived NSC line NSI-566 supports clinical safety 40 . However, in these studies, high-dose NSC integration into the injured spinal cord without a delivery platform shows mixed ability to reestablish lost locomotor modalities 35,40 . Contusion injury is the most prevalent form of SCI in humans. Here using a relevant rat C4 hemi-contusion model 16 , we demonstrate that minimally invasive delivery of scNSC neural ribbons by syringe injection into the contused injury cavity reproducibly achieves cell survival and retention with as few as 5,000 cells. Traditional NSC methods in rat models typically deliver cells in a dose range of 200,000 to 2 million cells. Here, the transplanted scNSCs survive and project long processes that co-localize with Synapsin with greater efficiency compared to scNSCs injected outside of the ribbons. While higher cell delivery numbers may be required to fill the cavity of larger injuries, the survival of our small grafts is a proof-of-principle for the neuroprotective function of this platform. Building on foundational work in human stem cell models, 17,18 Kumamaru et al. 5 developed a method to stabilize NSCs over several passages while retaining spinal cord identity. We use this early NMp induction method to induce Hox gene collinear activation during the first 5 days of differentiation, but converted to neuroectoderm by next modulating RA and Shh signaling pathways to generate scNSCs with cervical regional identity. The approaches described in this work offer strategies that apply to a variety of CNS pathological conditions requiring cell delivery to specified anatomic compartments.
Extrusion-based techniques with alginate and disease-specific requirements for cell types are being explored to develop a higher level of control over therapeutic cells, including incorporation of factors to direct or drive differentiation 29,41 , and ability to engineer structures with increasing complexity 42 , and have seen success in animal models of disease, particularly of diabetes mellitus 42,43 . Here we advance alginate microfiber technology in the form of multiplex neural ribbons applied for the first time to in vitro and in vivo models of SCI. scN-SCs were encapsulated as high-density single cells or pre-established neural aggregates at day 10 according to neural rosette cell number and size parameters. scNSCs retained metabolic activity, cell cycling, viability, and multipotent differentiation potential. These parameters were improved by increasing encapsulated cell density tenfold, and further by the loading of scNSC aggregates with pre-established 3-D cell-cell connections versus single cells. It was previously shown that cell-cell interactions and survival are bolstered by including natural ECM components 42 or chemical modifications such as conjugation with the integrin-binding domain peptide RGD that together more fully recapitulate a natural microenvironment. We therefore compared neural ribbons formed with alginate alone (ProNova) to an alginate-type 1 collagen IPN (NovaCol) and RGD-conjugated alginate with type 1 collagen IPN (RGD-Col) ( Fig. 4; see also Table S3). scNSC viability was highest using the RGD-Col composition, and encapsulated cell-cell connectivity was promoted by smaller 60 µm ribbon diameter versus 100 µm or 150 µm diameters. Since the RGD modification may exacerbate potential fibrotic responses of host cells in vivo, 60 µm diameter NovaCol neural ribbons were chosen for in vivo grafting studies. Importantly, we show the ability of scNSCs to form robust calcium-dependent adherens junctions within NovaCol neural ribbons, and to simultaneously differentiate and penetrate into surrounding 3-D ECM in vitro on a time scale similar to alginate biodegradation.
We demonstrate neural ribbon multifunctional customization by co-encapsulating chABC in vitro. However, chABC was excluded in our in vivo pilot study due to ongoing debate of the role of the glial scar in regeneration and repair. Our data support the premise that younger NSCs do not require chABC for initial navigation into host tissue, but acknowledge that chABC may be required for transplanting specialized postmitotic neurons and that this need may vary between subjects.
The ability to ship long-distance intact and viable neural constructs that retain stemness has not been previously demonstrated in the stem cell community and further highlights the broad clinical applicability of this approach to facilitate future GMP SCI therapeutic schemes. The ability to overcome multiple central barriers to SCI cell therapy remains a challenge, but is expected to be achievable by advancing knowledge of the appropriate cell type and reproducible delivery method.

Methods
Human hiPSC culture and maintenance. The self-designated African American human iPSC line F3.5.2 was previously developed by the Paluh and Cibelli labs using lentivirus transduction of donor fibroblasts with Yamanaka factors and extensively characterized 19,20 . Here, F3.5.2 was maintained in mTeSR1 or mTeSR Plus complete growth medium (STEMCELL Technologies) with 1 × penicillin-streptomycin (P-S; Gibco) on hESC- www.nature.com/scientificreports/  www.nature.com/scientificreports/ www.nature.com/scientificreports/ In vitro CSPG glial scar models and chABC activity. We modeled aspects of the glial scar inhibitory microenvironment in SCI using the well-established Laminin-CSPG gradient spot assay 31,32 . Briefly, 2 µl CSPG spots (0.4 mg/ml Aggrecan, 10 µg/ml Laminin in HBSS) were plated onto PDL-Nitrocellulose coated chambers of Lab-Tek II cover glass (Nunc #155382; 2-5 CSPG spots/chamber) and allowed to air dry. Dried spots were stable in HBSS or cell culture media for > 1 week at 4 °C. To visualize the inverse CSPG gradient, spots on coverslips were incubated overnight in 1% BSA (1 × PBS) with the anti-CS-56 monoclonal antibody against CSPG (1:1,000), rinsed 3X in HBSS and incubated 1:1,000 with donkey anti-mouse AlexaFluor 594 secondary antibody for microscopy. We grew day 10 scNSCs as well as differentiating neuronal cultures on CSPG spots coated with fresh Matrigel to test the efficacy of the CSPG gradient to exclude neural cells. The outer rim of CSPG spots was sufficient to exclude both scNSCs and neuronal axon. CSPG spots were incubated with 0.1 U/ml purified chABC in aCSF for 3 h and analyzed for enzymatic digestion (CS-56 IF) to test chABC activity. Relative fluorescence intensities of treated spots compared to controls were measured in ImageJ using nearly identical fields of view. aCSF and cell culture media both provide Ca 2+ cation that is a necessary cofactor for chABC enzymatic activity. chABC activity was also tested in multiple assays exploiting the endogenous production of CSPG ECM by scNSCs. To determine the time to achieve full chABC effect, day 10 scNSC rosettes were seeded into chambered

Microenvironment Transformation
Injury site CSPGs

SCI Site Modulators hiPSC Technologies
Spinal cord patterning Organotypic targets  www.nature.com/scientificreports/ cover glass at high density and grown to confluency. chABC was then added at 0.1 U/ml in culture medium and returned to 37 °C. Cells were fixed at 30 min intervals to 120 min and compared to controls (no chABC treatment). CS-56 fluorescence intensity (a.u.) was quantified over time (mean ± s.e.m.). n = 7 fields were quantified for each time point (N = 35 total measurements). scNSCs were seeded identically for chABC addition and washout studies. First, cells were fixed at three-hour intervals to determine CS-56 baseline fluorescence intensity. At 6 h, 0.1 U/ml chABC was added and samples were fixed at three-hour intervals to 12 h. The remaining samples were then washed 4 × in cell culture medium to dilute chABC and again fixed at three-hour intervals to 18 h, at which point samples were fixed at six-hour intervals to up to 36 h. n = 4 separate fields were measure per time point (N = 40 total measurements). We combined Ki-67 IF with the dual apoptosis kit described previously to quantify the effect of chABC on scNSC proliferation and apoptosis. Cells were fixed at t = 24 and 48 h after addition of 0.1 U/ml in culture medium. The percent of nuclei expressing Ki-67 versus dual apoptotic Casp3S-AnnV markers were quantified and compared to no chABC treatment controls (n = 4 fields averaged per condition). Vybrant MTT (ThermoFisher) chromogenic assays were also used to measure effect of chABC on cell proliferation. Acellular culture medium alone was used as a negative control. scNSCs were seeded into Matrigel-coated 96-well plates and grown to ~ 60% confluency, at which point 50% of wells containing scNSCs received 0.1 U/ml chABC. MTT measurements were performed at t = 24 and 48 h after chABC addition according to manufacturer's instructions. Briefly, 10 µl of 12 mM MTT stock was added and incubated in the dark at 37 °C for 3 h. Insoluble formazan was then dissolved in SDS-HCl at 37 °C and absorbance was measured at 570 nm using an automated plate-reader. The experiment was performed in triplicate and averaged.

In Vivo
Neural ribbon scNSC and chABC encapsulation. Ultrapure sodium alginates tested in this study are summarized in Supplementary To encapsulate scNSC single cell suspensions, rosettes were dissociated using Accutase or TrypLE for 5 min at 37 °C, distributed to 1.5 ml Eppendorf tubes 1:1 with DMEM/F-12 and counted. scNSCs were centrifuged at 350×g for 5 min followed by complete aspiration of supernatants. The scNSC pellet was resuspended in alginate or alginate-collagen mixtures to concentrations of 1 × 10 7 or 1 × 10 8 cells/ml by gentle static mixing with the pipette tip, avoiding bubbles. The scNSC-hydrogel suspension was then loaded directly into the Nordson tip and extruded into CaCl 2 using a syringe as described above. CaCl 2 was aspirated immediately after crosslinking with a 1,000 µl pipet tip. Ribbons were suspended in culture medium, inspected, and returned to the incubator. To load scNSC aggregates (50-100 µm), GCDR with mechanical passaging was used in place of Accutase/TrypLE dissociation. The same number of wells at 100% confluency (ave. 3.5 × 10 6 cells/well in 12-well plate) were used to generate neural aggregate ribbons as for the 1 × 10 8 cells/ml condition. For chABC encapsulation, purified enzyme was added to alginate suspensions prior to extrusion for a final working concentration of 0.1 U/ml. scNSC dose, proliferation, and viability in neural ribbons. Cell dose (i.e. cell number per unit ribbon length) was determined experimentally at SUNY Polytechnic Institute and compared to values obtained by Houston Methodist after shipping. scNSC neural ribbons were formed directly on glass slides, rinsed in deionized water and manually severed into to 3 mm segments. Ribbon segments were individually dissolved in 20 µl 1.6% sodium citrate (Ca 2+ chelation) and counted both by hemocytometer and automated cell counting. Only ribbons formed using 1 × 10 8 cell/ml single cell suspension or corresponding neural aggregates (50-100 µm) were subject to this analysis. For neural aggregates, only segments containing cells along the entire 5 mm length were used. After alginate dissolution, aggregates were dissociated completely prior to counting. n = 5 neural ribbon segments were counted for each loading configuration (N = 10 segments total).
Cell metabolic activity was assessed using CellTracker dye 24 h post-encapsulation. The Vybrant MTT kit (ThermoFisher) was additionally used according to manufacturer's instructions and insoluble Formazan crystals in ribbons were imaged by phase contrast. For cell proliferation within ribbons, we combined Ki-67 IF with Cell-Tracker and DAPI counterstain and acquired images of mitotic figures using high-magnification fluorescence microscopy ( Supplementary Tables 1, 2). We investigated scNSC apoptosis after CaCl 2 incubation for 1, 2 and 5 min versus control (no CaCl 2 ) to test the effect of 100 mM CaCl 2 on scNSC viability over a timespan relevant to ribbon formation. Culture medium was aspirated from adherent scNSCs seeded in chambered cover glass (Nunc #150680) and rinsed once in warm 1 × PBS prior to addition of sterile 100 mM CaCl 2 . At an appropriate timepoint, CaCl 2 was aspirated and cells were rinsed twice with 1 × PBS and returned to the incubator in Neural ribbon-encapsulated chABC in vitro assays. Using similar in vitro glial scar and CSPG ECM assays as described in the previous section, we tested the ability of neural ribbon-encapsulated chABC to ameliorate the inhibitory CSPG microenvironment. For these experiments, we formed single 40 µl ribbons containing enough chABC to establish a working concentration of 0.1 U/ml in 0.5 ml of medium, assuming 100% release.
As an initial test, single ribbons were incubated with CSPG spots in aCSF and fixed for CS-56 fluorescence at 24 h. For cellular assays, day 10 scNSCs were seeded onto Matrigel-coated CSPG spots (1,500 cells/mm 2 ) and grown to confluency. scNSCs were then incubated with empty (negative control) or chABC-containing ribbons in 0.5 ml CPM for 24 h. scNSC crossing events per CSPG spot were quantified by manual counting. The same assay was repeated with differentiating neuronal cultures (1,500 cells/mm 2 ) or plated neurospheres grown in suspension (5-8 neurospheres/well). For neuronal cultures, the experiment was initiated when sufficient axon outgrowth was observed by phase contrast 3-4 days after seeding. n = 9 full CSPG spots quantified per condition for scNSCs and neurons (N = 36 total spots). For microscopy, ribbons were removed from chambers prior to fixation and ICC with 1 × PBS-based solutions. We embedded neural ribbons in scNSCs and Type 1 collagen gel. scNSCs were grown to 100% confluency on Matrigel-coated glass bottom dishes (12 mm viewing area; Nunc, #150680). TexasRed-Dextran ribbons were formed +/− chABC, rinsed and manually positioned onto scNSCs. The scNSC/NR combined culture was immediately embedded in 2.5 mg/ml PureCol EZ gel and incubated overnight at 24 h (0.5 ml culture medium), at which point, cultures were co-stained for DAPI/CS-56 and quantified for CS-56 immunoreactivity (fluorescence intensity; n = 9 fields for no chABC, n = 7 fields for +chABC). For the latter assay, ICC was performed in the PIPES/HEPES/CaCl 2 buffer to preserve neural ribbon integrity.
Neural ribbon long-distance shipping and receiving. The Houston Methodist Neuroregeneration lab is > 1,750 miles from SUNY Polytechnic CNSE in Albany, NY. For long-distance shipping of scNSC neural ribbons encapsulated at day 10 (FedEx overnight, ~ 20-24 h laboratory-to-laboratory), we used a battery-powered portable incubator with removable block (CryoLogic BioTherm INC-12V). Prior to shipping, scNSC neural ribbons were formed directly in 12-well plates and resuspended in CPM supplemented with 10 µM ROCK inhibitor. Ribbon suspensions were then transferred to 1.8 ml cryovials using wide bore 1,000-µl pipette tips. Neural ribbon transport stability depends both on physical and chemical factors, including cell-cell contacts that are promoted by high densities, and prevention of necrosis due to reduced diffusion of gases and nutrients. 30 Upon receipt, neural ribbons were resuspended in fresh medium. Cell recovery after shipping was tested by dissolving alginate in 1.6% sodium citrate and plating of recovered cells onto freshly-coated Matrigel cultureware. In parallel, a group of neural ribbons were kept intact for in vivo injection in rodent C4 hemi-contusion injury models.
Rat cervical spinal cord hemi-contusion injuries and grafting. Lateral right-sided C4 hemi-contusion lesions were induced with moderate displacement (0.8 mm) in 4-month-old female Long-Evans rats (N = 6) as previously described 16 . Four animals received neural ribbons and two animals received scNSCs delivered in suspension. Briefly, the C4 laminae was exposed for right-sided hemi-laminectomy in anesthetized animals. The animal was then placed in a custom spinal frame, clamping the lateral processes of C3 and C5. An electromagnetic probe (Ling Inc.) was positioned at the surface of the dura with an initial sensing force followed by rapid 0.8 mm displacement with a dwell of 14 ms. Afterward, muscles were sutured in layers and the skin was clipped. Contused animals were immunosuppressed with cyclosporine-A (CSA; 10 mg/kg, i.p.) two days prior to the human neural ribbon transplantation. The CSA regimen was maintained until euthanasia. For grafting studies, the contused spinal cord was re-exposed 14 days after injury in anesthetized animals and the lateral process of C3 was clamped. 60 µm diameter NovaCol ribbons containing human scNSC aggregates and Fluorescein Dextran were manually severed into 3-4 mm segment lengths. Single segments containing approximately 5,000 cells were drawn into a 26G curved needle Hamilton syringe with a 45° bend ~ 2 mm proximal to the bevel. An Hamilton syringe was placed in a Kopf micromanipulator fixed to the spinal frame and positioned 0.5 mm lateral to the midline ipsilaterally. Dura was punctured by the syringe tip and advanced to 0.7 mm depth. A single ribbon segment was injected over 1 min in 5 µl of HBSS vehicle per animal. The needle was maintained in position for 2 min dwell, followed by 100 µm retraction and another 2 min dwell. The needle was slowly retracted while monitoring for efflux of fluid or the ribbon segment onto the surface of the dura. Finally, muscle and skin layers were sutured and the animals received standard post-operative care. All rat experimental protocols were approved by Houston Methodist Research Institute IACUC and carried out in accordance with relevant ethical guidelines and regulations. This includes animal comfort, veterinary care, methods and reasons for euthanasia, and materials and hiPSC-derived neural injections.
Tissue processing and immunohistochemistry. Nine days after neural ribbon injection, rats were sedated with isoflurane and transcardially perfused first with cold 0.1 M PBS with 10,000 U of heparin followed by 4% paraformaldehyde in 0.1 M PBS (pH 7.4). The spinal cord was harvested and stored in 4% paraformaldehyde at 4 °C