Abstract

Background

   Neural stem cell (NSC)-based therapies hold great promise for treating
   diseases of the central nervous system (CNS). However, several
   fundamental problems still need to be overcome to fully exploit the
   clinical potential of NSC therapeutics. Chief among them is the limited
   survival of NSC grafts within hostile microenvironments.

Methods

   Herein, we sought to engineer NSCs in an effort to increase graft
   survival within ischemic brain lesions via upregulation of global
   SUMOylation, a post-translational modification critically involved in
   mediating tolerance to ischemia/reperfusion.

Findings

   NSCs overexpressing the SUMO E2-conjugase Ubc9 displayed resistance to
   oxygen-glucose-deprivation/restoration of oxygen/glucose (OGD/ROG) and
   enhanced neuronal differentiation in vitro, as well as increased
   survival and neuronal differentiation when transplanted in mice with
   transient middle cerebral artery occlusion in vivo.

Interpretation

   Our work highlights a critical role for SUMOylation in NSC biology and
   identifies a biological pathway that can be targeted to increase the
   effectiveness of exogenous stem cell medicines in ischemic stroke.

Fund

   Intramural Research Program of the NINDS/NIH, the Italian Multiple
   Sclerosis Foundation (FISM), the Bascule Charitable Trust,
   NIH-IRTA-OxCam and Wellcome Trust Research Training Fellowships.

   Keywords: Neural stem cells (NSCs), SUMOylation, Ubc9,
   Ischemia/reperfusion, Stroke, Cell therapy, Regenerative medicine,
   Cellular engineering

Graphical abstract

   Unlabelled Image
   [51]Open in a new tab

Highlights

     * •
       Ubc9-overexpressing NSCs demonstrate enhanced neuronal
       differentiation.
     * •
       Upregulating SUMOylation in NSCs increases resistance to
       ischemia/reperfusion in vitro.
     * •
       Ubc9-overexpressing NSC grafts robustly integrate within the brain
       of mice post-stroke.
     __________________________________________________________________

Research in context.

Evidence before this study

   Several studies have demonstrated the preclinical efficacy of NSC
   therapies with regard to their ability to protect and/or repair the
   ischemic brain. Transplanted NSCs exert beneficial effects not only via
   the structural replacement of dysfunctional and/or damaged endogenous
   cells, but also via immunomodulatory and neurotrophic actions.
   Unfortunately, the successful translation of such promising cell-based
   approaches remains elusive, in part due to the fact that grafted NSCs
   often die in the hostile ischemic/post-ischemic microenvironment.

Added value of this study

   Ischemic stroke continues to pose a massive burden of disease and is a
   leading cause of morbidity and mortality throughout the world. However,
   beyond thrombolysis/mechanical thrombectomy we possess few effective
   therapies that are able to modulate the pathogenesis of evolving
   ischemic brain injury. The paucity of therapeutic options stands in
   stark contrast to the intensity of research efforts/number of clinical
   trials that have been performed to date. Further, there are, as of yet,
   no effective treatments that improve functional recovery in
   post-ischemic patients (i.e. regenerative therapies).

Implications of all available evidence

   While NSC-based therapy for stroke holds great promise fundamental
   questions/problems persist and in so doing hinder the clinical
   translation of such approaches. As such, the work presented herein
   sought to explore the effects of the recruitment and optimization of a
   global neuroprotective modality (i.e. SUMOylation) on the efficacy of
   NSC cell-based therapy. Such work may ultimately find utility not only
   in the realm of stroke therapy, but could also extend to a wide variety
   of other degenerative/inflammatory neurological disorders that share
   components of stroke pathobiology (e.g. Alzheimer's, multiple
   sclerosis, spinal cord injuries, etc.).

   Alt-text: Unlabelled Box

1. Introduction

   In ischemic stroke a crisis in energy availability due to the sudden
   reduction of oxygen and glucose leads to the depletion of cellular
   energy stores, aberrant neurotransmitter release, and intense
   cytotoxic/inflammatory responses [[52]1]. The intrinsic complexity of
   this phenomena accounts for the continuing failure of clinical
   approaches targeting single pathogenic mechanisms. As such, there is
   currently no effective treatment that improves functional recovery and
   replaces lost neural tissue in stroke patients.

   Compelling evidence exists that the transplantation of neural stem
   cells (NSCs) exerts beneficial effects in multiple CNS disorders,
   including stroke, through both the structural replacement of endogenous
   neural cells as well as immunomodulatory and tissue trophic effects
   [[53][2], [54][3], [55][4]]. However, the metabolic demands of cellular
   grafts are only partially met, due to poor perfusion of post-ischemic
   brain tissue [[56]5]. Accordingly, very few transplanted NSCs survive
   and integrate in the hostile post-ischemic microenvironment, thus
   hindering the therapeutic potential and successful translation of
   NSC-based approaches in stroke patients [[57]6]. Counteracting cell
   death, the predominant fate of the NSC graft, therefore provides a
   clear target for interventions that aim at increasing the integration
   of cellular grafts [[58]7].

   A candidate pathway involved in resistance to the ischemic stress is
   the post-translational modification of proteins by the Small
   Ubiquitin-like MOdifier (SUMO), a process called SUMOylation [[59][8],
   [60][9], [61][10]]. We first described a striking increase in levels of
   global SUMOylation during hibernation torpor in 13-lined ground
   squirrels (Ictidomys tridecemlineatus) [[62]8,[63]11]. These animals
   showed no evidence of cellular damage or functional deficits despite
   prolonged exposure to low perfusion levels similar to those observed
   within the “ischemic core” [[64]11]. Of interest, recent work has even
   come to suggest a role for SUMOylation within the human penumbra
   [[65]12].

   Mechanistically, SUMOylation requires a cascade of enzymatic steps that
   involves an E1 activating enzyme, an E3 SUMO ligase, and the sole E2
   conjugase Ubc9, which is coded by the UBE2I gene [[66]9]. Transgenic
   mice overexpressing Ubc9 showed increased global SUMOylation, with
   levels of SUMOylation directly proportional to the degree of
   cytoprotection [[67]13,[68]14].

   Herein, we sought to engage SUMOylation in an effort to engineer NSC
   grafts capable of surviving within unhospitable ischemic
   microenvironments. We established Ubc9-overexpressing NSCs and
   characterized their global gene/protein expression profiles, phenotype,
   and function, both in vitro and in vivo. The upregulation of neurogenic
   pathways in Ubc9 NSCs led to enhanced neuronal differentiation in
   vitro. Ubc9 NSCs also displayed a strikingly-increased resistance to
   oxygen-glucose deprivation and restoration of oxygen/glucose (OGD/ROG)
   in vitro. Finally, when transplanted directly into the ischemic lesion
   of mice with middle cerebral artery occlusion (MCAO), Ubc9 NSC grafts
   showed a significant, 2-fold higher level of survival and integration
   in vivo as compared to control NSC grafts.

   Increasing SUMOylation in NSCs represents a novel approach to graft
   preconditioning that has the potential to overcome some of the current
   limitations of stem cell medicines within regenerative neurology.

2. Materials and methods

2.1. Contact for reagent and resource sharing

   Further information and requests for resources and reagents should be
   directed to and will be fulfilled by the Lead Contacts: Joshua D.
   Bernstock ([69]bernstockjd@ninds.nih.gov); Luca Peruzzotti-Jametti
   ([70]lp429@cam.ac.uk); John M. Hallenbeck ([71]hallenbj@ninds.nih.gov);
   and Stefano Pluchino ([72]spp24@cam.ac.uk).

2.2. Intraluminal (temporary) middle cerebral artery occlusion (MCAO), and
stereotaxic NSC transplantation

   Adult male C57NL/6 mice (10–12 weeks old) were purchased from Charles
   River. Animals were anesthetized with 1–1.5% isoflurane. Temperature
   was maintained between 36.5 °C and 37.0 °C, and laser Doppler flow was
   monitored during the procedure and for up to 15 min post-reperfusion to
   ensure the return of cerebral blood flow (CBF). Focal cerebral ischemia
   of the left MCA was induced with a silicon-coated 6–0 nylon filament
   (Doccol Corporation). The filament was retracted after 45 min allowing
   reperfusion. At 72 h after ischemia, animals were randomized to receive
   intraparenchymal transplantation of either GFP-labelled WT NSCs or
   GFP-labelled Ubc9 NSCs (both at passage ≤15). A 26S gauge small hub RN
   needle 2 in point style 3 (Hamilton Company, Reno, NV) was inserted at
   the following coordinates, with the origin at bregma:
   anterior-posterior [+] 0.5 mm, mediolateral [−] 1.8 mm, dorsoventral
   [−] 2.6 mm. A 0.5 mm reservoir was created using the needle immediately
   prior to cell injection. A total of 100,000 cells in a total volume of
   2 μl was injected slowly and continuously over a period of 5 min.
   Post-injection, the needle was withdrawn continuously at a rate of
   0.1 mm per 30 s and the surgical incision was sutured. The equipment
   used in the transplantation procedure is as follows: 5 μl Model 75 RN
   syringe (Hamilton Company), Benchmark Stereotaxic Frame (Leica
   Biosystems), Benchmark Digital Stereotaxic Control Panel (Leica
   Biosystems). All animal experiments conformed to the guidelines set
   forth by the NIH/NINDS Animal Care and Use Committees (ACUC) (protocol
   #1268-15).

2.3. Genotyping of transgenic animals and cell lines

   Transgenic animals and derived cell lines were genotyped via touchdown
   PCR to confirm the presence of the UBE2I transgene. Part of the CAG
   promoter sequence (5′-gcgccggcaggaaggaaatg-3′) was used as a forward
   primer and part of the Ubc9 coding sequence
   (5′-ggtgatagctggcctccagtcc-3′) was used as a reverse primer. Only the
   transgene (CAG-UBE2I) is amplified as a 650-base pair DNA fragment.
   SHSY5Y cell extracts (total DNA preps) from trans gene inserted plasmid
   are used as positive control.

2.4. NSC derivation and cell culture

   Adult neurospheres were generated from the SVZ of both WT C57BL/6 and
   Ubc9 transgenic 4–8-month-old mice. Briefly, mice were anesthetized by
   an intraperitoneal injection of pentobarbital and sacrificed via
   cervical dislocation. The brains were removed and placed in artificial
   cerebrospinal fluid (aCSF) (124 mM NaCl, 5 mM KCl, 1.3 mM MgCl[2],
   0.1 mM CaCl[2], 26 mM NaHCO[3], and 10 mM d-glucose, pH 7.4) aerated
   with 95%, O[2], 5% CO[2] at room temperature. SVZ neural tissue was
   isolated after coronal sectioning and cut into 1 mm^3 pieces. Pieces
   were transferred into 30 ml of aCSF containing 1.3 mg/ml trypsin,
   0.67 mg/ml hyaluronidase, and 0.2 mg/ml kynurenic acid (Sigma-Aldrich)
   and incubated, under continuous oxygenation and stirring, for 90 min at
   32–34 °C. Tissue sections were then rinsed in aCSF for 10 min,
   transferred to DMEM/F12 (Thermo Fisher Scientific) medium containing
   0.7 mg/ml ovomucoid (Sigma-Aldrich), and carefully triturated using a
   fire-polished Pasteur pipette. Cells were collected by centrifugation
   and resuspended in growth factor-free, chemically-defined DMEM/F12
   medium containing 2 mM l-glutamine, 0.6% glucose, 9.6 mg/ml putrescine,
   6.3 ng/ml progesterone, 5.2 ng/ml sodium selenite, 0.025 mg/ml insulin,
   0.1 mg/ml transferrin, and 2 mg/ml heparin (Sigma-Aldrich). Cells were
   then cultured in NS-A medium (Euroclone) containing 20 ng/ml of
   epidermal growth factor (EGF) and 10 ng/ml fibroblast growth factor
   (FGF)-II (Peprotech). The number of primary spheres was counted after
   7–12 days in vitro. For cell amplification, 8000 cells/cm^2 were plated
   at each sub-culturing passage in untreated tissue culture flasks. NSC
   Proliferation Medium which consisted of 10% NeuroCult Proliferation
   Supplement (Stem Cell Technologies), 1% Pen/Strep, 0.1% v/v Heparin
   (Sigma), 0.01% v/v EGF (Peprotech), 0.01% v/v bFGF (Peprotech) in
   NeuroCult Basal Medium (Stem Cell Technologies) was used for
   growth/cell maintenance and incubated at 37 °C and 5% CO[2]. After ~
   4 days neurospheres were harvested, enzymatically dissociated using
   Accumax (Stem Cell Technologies) counted using trypan blue and
   ultimately re-plated under the same culture conditions.

2.5. NSC growth curves, viability, clonogenicity and cell cycle in vitro

   The linear growth curve was generated by calculating the total number
   of cells by multiplying the growth rate (i.e. number of live cells
   divided by the number of seeded cells) by the total number of cells
   present at the previous time point. The mean of the total number of
   cells per time point ± standard error of the mean (SEM) was reported to
   build the linear trend line. The daily growth rate was obtained by
   dividing the growth rate by the number of days per passage. Viability
   was defined as the percentage of viable cells over dead cells ±SEM. For
   clonogenicity assays, single cells were sorted via flow cytometry into
   100 μL of NSC proliferation media/well in flat bottom 96-well plates.
   7 days later the number of neurospheres were counted and scored (i.e.
   size and total number). For cell cycle analysis, 350,000 WT or Ubc9
   NSCs per well were plated into 6-well plates pre-coated with Matrigel
   Matrix (Corning). After the incubation period, cells were fixed,
   permeabilized, and stained using the Click-iT EdU Plus Flow Cytometry
   Assay Kit (ThermoFisher Scientific) as per the manufacturer's
   instructions. This was followed by counterstaining with 1 μg/ml
   4′,6-diamidino-2-phenylindole (DAPI) to assay total DNA content per
   cell. Cells were then analysed by flow cytometry using a MoFlo Astrios
   cell sorter with Summit Acquisition software (Beckman Coulter);
   5 × 10^4 events were acquired per sample. Data analysis was completed
   with Kaluza software (Beckman Coulter).

2.6. NSCs differentiation in vitro

   For differentiation assays in vitro, cells were seeded on 13 mm glass
   coverslips pre-coated with Matrigel™ (Corning) at a density of 8 × 10^4
   cells/coverslip and cultured in 400 μl differentiation medium
   (NeuroCult™ basal medium) (Stem Cells Technologies) with 10% mouse
   differentiation supplement (Stem Cells Technologies), 1% pen/strep].
   Half of the medium was replaced with fresh differentiation medium after
   3 days. After 3 more days (6 days in total), coverslips were washed
   with PBS and fixed with 4% paraformaldehyde (Sigma-Aldrich) and 2%
   sucrose in PBS. For immunofluorescence staining, cells were rinsed with
   PBS, and then blocked for 1 h at room temperature in blocking buffer
   (0.1% Triton X100 and 10% secondary antibody species serum in PBS).
   Primary antibodies were incubated at 4 °C overnight. The following
   primary antibodies diluted in blocking buffer were used: anti-nestin
   (1:200) (Abcam), anti-SOX2 (1:100) (Abcam), anti-vimentin (1:100)
   (Abcam), anti-glial fibrillary acidic protein (GFAP) (1:500) (Abcam),
   anti-β-tubulin-III (1:500) (BioLegend), anti-O4 (1:400) (R&D Systems).
   Cells were then washed in PBS with 0.1% Triton X100 and incubated with
   the appropriate fluorescent secondary antibodies (1:1000 Alexa-Fluor
   405, 488, 555, 647, Invitrogen) 1 h at room temperature. After washing
   in PBS, nuclei were counterstained with 1 μg/ml DAPI (Invitrogen) for
   3 min and then mounted with Dako mounting kit (Sigma-Aldrich).
   Nonspecific staining was observed in control incubations in which the
   primary antibodies were omitted. For quantification, images were
   acquired using a CCD camera (DC 480) (Leica) equipped fluorescence
   microscope (Olympus BX51) (Olympus) with a 40× objective on 6 regions
   of interest (ROI) per coverslip. Images were analysed and prepared
   using ImageJ software (NIH, Bethesda, MD). Data are presented as the
   percentage of positive cells over the total of DAPI positive cells ±
   SEM.

2.7. XF assays

   Oxygen consumption rate (OCR) and extracellular acidification rate
   (ECAR) were measured using the Seahorse XF24 Extracellular Flux
   Analyser (Seahorse Bioscience). Undifferentiated WT and Ubc9 NSCs were
   dissociated and seeded into a 24 well XF24^e cell culture microplate
   (3.3 × 10^5 cells/cm^2) pre-coated with Matrigel™ (Corning) at a final
   density of 1 × 10^5 cells/well in normal media. The assays were
   performed 18 h after seeding. For differentiated WT and Ubc9 NSCs we
   adapted the protocol described above (NSCs differentiation in vitro) to
   the XF24^e cell culture microplate. Briefly, undifferentiated NSCs were
   seeded after dissociation on a 24 well XF24^e cell culture microplate
   pre-coated with Matrigel™ (Corning) at a final density of 3.3 × 10^5
   cells/cm^2 and cultured in 400 μl differentiation medium [NeuroCult™
   basal medium (Stem cells Technologies), 10% mouse differentiation
   supplement (Stem Cells Technologies), 1% pen/strep]. 200 μl of
   differentiation medium were replaced with fresh medium after 3 days.
   Differentiation was continued until day 6, and assays were performed
   the following day. Before XF analysis, culture media were removed,
   cells were washed with sterile PBS, and XF medium pH 7.35–7.45 was
   added and allowed to equilibrate for 30 min at 37 °C. Data were
   normalized on total protein content per well evaluated using a BCA
   assay kit (Thermo Fisher Scientific) and are expressed as pmoles/min
   (OCR) or milli-pH units (mpH)/min (ECAR) mean ± SEM.

2.8. Oxygen/glucose deprivation (OGD)/restoration of oxygen/glucose (ROG) and
FACS analysis of live/dead cells

   We subjected NSCs (both differentiated and undifferentiated) to OGD for
   5 h followed by ROG for 18 h. RNA for the microarrays was harvested 1 h
   into reperfusion. To induce OGD, the basal culture medium was replaced
   with a glucose-free, balanced salt solution (BSS), which contained (in
   mmol/L) the following: NaCl, 116; CaCl[2], 1.8; MgSO[4], 0.8; KCl, 5.4;
   NaH[2]PO[4], 1; NaHCO[3], 14.7, and HEPES, 10 (pH 7.4). Cell death was
   assessed via nuclear staining with Hoechst 33342 and propidium iodide
   (PI) followed by FACS analysis. Typically, 1 × 10^5 cells were
   analysed. The percentages of vital or apoptotic/necrotic cells in the
   OGD condition were normalized to their corresponding non-OGD sample to
   control for baseline death rate before deriving the fold difference vs
   the WT average.

2.9. Cleaved caspase-3 activity

   Apoptosis in undifferentiated WT and Ubc9 NSCs after OGD/ROG was also
   assessed with an assay for cleaved caspase-3. Cells were collected and
   analysed according to the manufacturer's instructions (R&D Systems).
   Cleaved caspase-3 levels in Ubc9 NSCs were normalized as fold
   differences relative to their corresponding passage-matched WT.

2.10. Immunoprecipitation (IP) and western blot analyses

   For the global SUMO-1 pulldown, after removal of the media and washing
   with phosphate buffered saline (PBS), cells were lysed in an IP buffer
   (50 mM HEPES pH 7.3, 100 mM NaCl, 1.5 mM MgCl[2], 1% NP40, 0.1% SDS,
   1 mM PMSF, 20 mM NEM, MS-SAFE protease/phosphatase inhibitor cocktail).
   The cell lysates were then incubated for 1 h on ice and then
   centrifuged for 20 min at ~12,000 x g at 4 °C. After pre-clearing with
   Dynabeads Protein G (Thermo Fisher Scientific), protein concentrations
   were measured with a Pierce BCA Protein Assay (Thermo Fisher
   Scientific). Equal amounts of protein among the relevant samples were
   ultimately used for IP. After incubation with the primary antibody for
   2 h at 4 °C, Dynabeads Protein G were added and incubated overnight at
   4 °C. The washing and eluting of IP products that followed conformed
   entirely to the manufacturer's (Thermo Fisher Scientific) protocol.
   Total cell lysates for western blots from NSCs were prepared and the
   antibodies used throughout the course of this study were as follows:
   rabbit polyclonal anti-SUMO-1 and anti-SUMO-2/3 antibodies (both
   developed in-house), rabbit monoclonal anti-UBE2I/Ubc9 (Abcam), and
   anti-β-actin (Sigma-Aldrich). Protein expression levels were determined
   via densitometric analysis of the corresponding protein bands of
   interest using ImageJ (NIH, Bethesda, MD). To measure SUMO-conjugation
   levels, regions corresponding to molecular weights above 100 kDa in
   each lane were cropped and the total intensity analysed. All densities
   were normalized to the corresponding actin levels and expressed as a
   fold difference vs WT NSCs.

   For the Sox2 IP, after removal of the media and washing with phosphate
   buffered saline (PBS), cells were lysed in Pierce IP Lysis Buffer
   (Thermo-Fisher Cat: 87788) with Halt Protease and Phosphatase Inhibitor
   Cocktail (Thermo-Fisher Cat: 78440). IP was performed with Dynabeads
   Protein G (Thermo-Fisher Cat: 10003D) conjugated with a mouse anti-Sox2
   primary antibody (Abcam Cat: 171380), following manufacturer's
   instructions. Samples were then loaded on Mini-PROTEAN TGX Stain-Free
   Gels (Biorad Cat: 456–8086), transferred and probed with the mouse
   anti-Sox2 primary antibody (Abcam Cat: 171380), a rabbit anti-SUMO-1
   primary antibody (Cell Signaling Technology Cat: 4930) and relevant
   anti mouse/rabbit HRP secondary antibodies (GE Health Care Cat:
   NXA931and Thermo-Fisher Cat: 31402 respectively).

2.11. Proteomic identification & pathway enrichment analysis

   Protein identification of excised Coomassie Blue-stained gel bands was
   performed by in-gel tryptic digestion, analysis of the resulting digest
   by liquid chromatography coupled to mass spectrometry LC/MS, and
   database searching. Briefly, samples were digested with trypsin
   overnight at 37 °C. Tryptic peptides were desalted first and analysed
   on an Orbitrap Elite mass spectrometer (Thermo Scientific). Peptides
   were separated on an Ultimate 3000 HPLC (Thermo-Dionex) using a ES802
   Easy-Spray column (75-μm inner diameter, 25 cm length, 3 μm C[18]
   beads; Thermo Scientific). A 55-min linear gradient of 2–27% mobile
   phase B (mobile phase A: 2% acetonitrile, 0.1% formic acid; mobile
   phase B: 98% acetonitrile, 0.1% formic acid) was used. The Thermo
   Scientific Orbitrap Elite mass spectrometer was operated in positive
   nano-electrospray mode. The CID MS/MS data were acquired in data
   dependent mode. The top ten most abundant ions were selected for
   product ion analysis. Mass spectrometry-identified proteins for two
   separate SUMO1 pull-down fractions representing the WT NSC condition
   and the Ubc9 NSC condition were filtered to retain proteins having at
   least 2 significant peptide matches in at least one fraction
   representing WT NSC condition and/or in at least one fraction
   representing the Ubc9 NSC condition. For these proteins, three lists
   were organized by symbol: detected in both WT NSC and Ubc9 NSC
   conditions, detected in WT NSC condition only, and detected in Ubc9 NSC
   condition only. For proteins detected in the Ubc9 NSC condition only,
   symbols were import into IPA ([73]www.ingenuity.com) and the
   corresponding enriched functions and pathways identified.

2.12. Microarray analysis (Affymetrix Mouse Gene 2.0 ST Arrays)

   Total RNA was extracted by samples using the RNeasy Midi Kit (QIAGEN)
   then labelled according to the manufacturer's guidelines for use with
   the Mouse Gene 2.0 Array (Affymetrix). Labelled cRNA were hybridized to
   these arrays in a blinded interleaved fashion. The Scanner 3000
   (Affymetrix) was used in conjunction with the GeneChip Operation
   Software (Affymetrix) to generate one CEL file per hybridized cRNA. The
   CEL files were then loaded into R, RMA was normalized with the oligo
   package, and filtered to only retain probes annotated as “main”.
   Differential expression testing was performed using limma and the
   resulting p.values were corrected with the Benjamini-Hochberg method.
   To test the interaction between OGD/ROG response and genotype, we added
   the following contrasts to the contrasts.fit function of limma:
   (Ubc9_NDiff_OGDROG-Ubc9_NDiff_CNTL)-(NPC_Ndiff_OGDROG-NPC_Ndiff_CNTL)
   and (Ubc9_Diff_OGDROG-Ubc9_Diff_CNTL)-(NPC_diff_OGDROG-NPC_diff_CNTL).
   GO enrichment analyses were performed using the topGO package
   ([74]https://bioconductor.org/packages/release/bioc/html/topGO.html)
   with the classic algorithm and Kolmogorov-Smirnov statistic. Clustering
   for microarray heatmaps was realized using Euclidean distance and the
   hclust algorithm (complete method). The KEGG pathway enrichment
   analysis were performed with the CAGE packages using lists of genes
   sorted by log2 fold change.

2.13. Lentiviral GFP tagging

   Cells used for transplantation studies were transduced in vitro using a
   third-generation lentiviral carrier (pRRLsinPPT-hCMV) coding for the
   enhanced farnesylated green fluorescent protein (GFP), which targets
   the fluorescent protein to the inner plasma membrane of transduced
   cells. Neurospheres were harvested, dissociated to a single cell
   suspension, and seeded at high density [1.5 × 10^6 in a T75 cm^2 flask
   (Sigma-Aldrich)] in fresh medium. After 12 h, 3 × 10^6 T.U./ml of
   lentiviral vectors were added. 6 h later, fresh medium was added. 72 h
   after viral transduction, cells were harvested and re-seeded at a
   normal concentration. Transgene expression was measured by FACS
   analysis before transplantation and > 98% of cells were found to be
   labelled in both lines (data not shown).

2.14. Ex vivo pathology and stereological analysis

   Five- or 30-days post-transplantation (dpt) of WT and Ubc9 NSCs, mice
   were deeply anesthetized with isoflurane and transcardially perfused
   with PBS for 15 min, followed by a solution of 4% PFA in PBS for
   15 min. Brains were then exposed, isolated, post-fixed overnight in 4%
   PFA and ultimately cryopreserved in 10–30% sucrose in PBS at 4 °C for
   24–48 h. Brains were then embedded in optimum cutting temperature (OCT)
   medium, frozen using liquid nitrogen, cryo-sectioned coronally at a
   thickness of 20 μm and stored at −80 °C. Quantification of total
   ischemic volume was performed using cresyl violet (CV) stained brain
   sections. Briefly, sections were first dehydrated with increasing
   dilutions of ethanol in water, washed in xylene (Merck), and rehydrated
   with decreasing dilutions of ethanol in water. Sections were then
   incubated with the CV solution for 10 min at room temperature. Stained
   tissues were then dehydrated with 75%, 95% and 100% ethanol diluted in
   water, washed in xylene, and mounted using a synthetic mounting medium
   (EUKITT). Data are expressed as percentage of contralateral hemisphere
   volume ± SEM of n = 12 sections per mouse. For quantification of WT and
   Ubc9 NSC graft survival, 20 μm coronal sections were pre-treated with
   peroxidase 3% for 15 min, incubated in a blocking solution which
   consisted of PBS + 10% normal goat serum (NGS, Sigma-Aldrich) for 1 h
   at room temperature, and then incubated with chicken anti-GFP (1:500)
   diluted in PBS and 1% NGS at 4 °C overnight. Sections were washed with
   PBS and incubated with goat anti-chicken biotinylated secondary
   antibody (1:500, Sigma-Aldrich) diluted in PBS and 1% NGS for 1 h at
   room temperature. Components “A” and “B” of the Vectastain kit (Vector
   Laboratories) were pre-mixed for 45 min and sections were incubated
   with this solution for 1 h at room temperature. The stain was developed
   by means of 3,3′-diaminobenzidine (DAB) following the manufacturer
   instructions. After 30 s the reaction was stopped by dipping the
   section into distilled water. Sections were then counterstained with
   haematoxylin for 30 s at room temperature and mounted using Dako
   mounting medium (Fluka). The mean thickness of analysed sections was
   17.64 μm (shrinkage to 88.2% of the original thickness). Transplanted
   GFP^+ cells were quantified according to the optical fractionator
   method on n = 12 equally spaced sections with the assistance of an
   Olympus BX53 microscope with motorized stage and Neurolucida software
   (11.07 64-bit, Microbrightfield); the motorized stage of the microscope
   controlled by the software allowed precise and well-defined movements
   along the x-, y- and z-axes. The counting was performed using a 100×
   oil immersion lens. Data are expressed as number of counted cells ±
   SEM.

   For the detection of apoptotic cells within coronal brain sections a
   Click-iT™ Plus TUNEL Assay for in situ apoptosis detection (Alexa
   Fluor™, ThermoFisher) was used in combination with primary antibodies
   (Ki67 at 1:500) as per the manufacturer's instructions. Briefly, after
   washing, sections were incubated for 1 h at room temperature with the
   appropriated combinations of secondary antibodies (1:1000 Alexa-Fluor
   488, 555, 647, Invitrogen). Nuclei were counterstained with DAPI
   (1:10,000, Invitrogen) for 3 min at room temperature. For
   quantification of NSC differentiation at 30 dpt, 20 μm coronal brain
   sections were washed with PBS and blocked for 1 h at room temperature
   in 0.3% Triton X100 and 10% secondary antibody species serum in PBS. A
   Fab fragment affinity-purified IgG anti-mouse was used if anti-mouse
   antibody were incubated with sections (1:10, Jackson ImmunoResearch).
   The following primary antibodies were then incubated overnight at 4 °C:
   anti-GFP (1:250) (Invitrogen), anti-GFAP (1:500) (Abcam), anti-NeuN
   (1:200) (Chemicon), anti-Olig2 (1:500) (Millipore). Differentiation
   data are expressed as % GFP^+ cells that are reactive for a specific
   marker.

   To determine the percentage of WT and Ubc9 NSCs that have
   differentiated into neurons and to study the synapse formation of these
   neurons, cryo-sections of the brain were stained for markers specific
   for microtubule-associated protein 2 (anti-MAP2) (1:200) (Abcam) and
   for postsynaptic density protein 95 (anti-PSD95) (1:300) (Abcam) while
   concurrently staining for the GFP (1:200) (Invitrogen) labelled NSCs.
   Sections were briefly rinsed in distilled water and washed with PBS.
   After blocking of the sections in blocking buffer (PBS + 0.1% Triton
   X100) with 10% NGS for 1 h at room temperature, and blocking with 1:10
   goat anti-mouse IgG (in blocking buffer, the primary antibodies were
   diluted in blocking buffer with 10% NGS and incubated overnight at
   4 °C. Sections were then washed, incubated with the appropriate
   secondary antibodies (1:400 Alexa-Fluor 488, 546, 647, Invitrogen) for
   2 h at 4 °C, and washed again. Nuclei were counterstained with DAPI
   (1:10,000) (Invitrogen) for 3 min at room temperature. Sections were
   washed and mounted using Dako fluorescence mounting medium (Agilent).
   Images of the staining were acquired with an Andor dragonfly confocal
   microscope (Oxford Instruments) and analysed using Fiji software. Data
   is expressed as % GFP^+ area that is positive for MAP2 or PSD95.

2.15. Quantification and statistical analysis

   Data are mean values ± SEM. The 2-tailed Student's t-test was used for
   single comparisons. Protein mass spectrometry data were analysed with
   Fisher's Exact Test. p-value<.05 was considered significant. Gene
   expression profiling assays were realized with Mouse Gene 2.0 Arrays
   (Affymetrix). See specific sections for the detailed experimental and
   analytical procedures.

2.16. Data and software availability

   The microarray data have been deposited on GEO with the accession
   number E-MTAB-7428.

3. Results

3.1. Overexpression of Ubc9 in NSCs induces genes and proteins governing cell
cycle and neurogenesis in vitro

   We established NSC lines from the subventricular zones (SVZ) of adult
   Ubc9-overexpressing mice (Ubc9 NSCs) [[75]13] and control wild-type
   mice (WT NSCs) ([76]Fig. 1A and B).

Fig. 1.

   [77]Fig. 1
   [78]Open in a new tab

   Overexpression of Ubc9 induces genes and proteins related to cell
   signaling, cell metabolism and cell cycle.

   (A) Representative confocal images of WT NSCs and Ubc9 NSCs growing in
   vitro as SOX2^+/Vimentin (Vim)^+ neurospheres. Nuclei are stained with
   DAPI. See also Fig. S1. (B) Genomic PCR for CAG-UBE2I transgene in Ubc9
   NSCs (+: positive control). (C) Representative immunoblots and
   densitometries normalized to corresponding actin levels and expressed
   as a fold induction (FI) relative to WT NSCs. Data are means ± SEM.
   N ≥ 3 per group, * p-value<0.05 (2-tailed Student's t-test). (D) Global
   SUMO-1 pulldown in WT NSCs and Ubc9 NSCs. Input = precleared lysate,
   IP = immunoprecipitated lysate. See also Data S1. (E) Venn diagram of
   protein mass spectrometry data showing that 125 proteins
   co-immunoprecipitated with SUMO-1 in WT NSCs only, 815 proteins in Ubc9
   NSCs only, and 759 were common to both. Pathways and functions
   significantly upregulated in Ubc9 NSCs compared to WT NSCs based on
   protein mass spectrometry results are reported. N = 2 per group
   (Fisher's Exact Test). See also Data S1 and Fig. S1. (F) Heatmap of
   genes differentially expressed in Ubc9 NSCs vs WT NSCs. N = 4 per
   group, adjusted p-value<0.05. See also Data S2. (G) Bar chart of the
   fold change (log2) in Ubc9 NSCs vs WT NSCs of the 10 most upregulated
   and downregulated genes. Adjusted p-value<0.05. See also Data S2. (H)
   Plot of the GO enrichment results for genes differentially expressed in
   Ubc9 NSCs vs WT NSCs. The x-axis shows the enrichment ratio, the color
   indicates the enrichment p-value (Kolmogorov–Smirnov statistic) and the
   size of each point indicates the number of significant genes in the
   corresponding GO category. See also Data S2.

   We found that global SUMO-1 and SUMO-2/3 conjugation, as well as Ubc9
   protein levels, were increased in Ubc9 NSCs ([79]Fig. 1C). We then
   performed global SUMO-1 immunoprecipitation (IP) ([80]Fig. 1D) and
   protein pathway analysis ([81]Fig. 1E) confirming that, beyond
   SUMOylation and ubiquitination, protein pathways related to cell
   signaling, cell metabolism, and cell cycle were prevalent in Ubc9 NSCs
   (Data S1). Ubc9 NSCs also displayed significantly higher growth rates
   and self-renewal in vitro, with an increased proportion of cells in the
   S-phase of the cell cycle (Fig. S1).

   We next performed a microarray expression analysis of Ubc9 NSCs
   ([82]Fig. 1F and Data S2) and found that 497 and 922 genes were
   significantly downregulated and upregulated, respectively, in Ubc9 NSCs
   (p-value<0.05).

   On one hand, Ubc9 NSCs demonstrated a significant reduction in the
   expression levels of genes related to apoptosis, including the death
   receptor Tumor Necrosis Factor Receptor Superfamily (Tnfrsf)10b, the
   regulator of autophagy Integral Membrane Protein 2A (Itm2a), and the
   cholinergic receptor nicotinic alpha 4 subunit (Chrna4) ([83]Fig. 1G).

   On the other hand, the cell cycle regulator Protein Phosphatase 2
   Regulatory Subunit B delta (Ppp2r2d) [[84]15], and the regulator of
   glycolytic metabolism
   6-phosphofructo-2-kinase/fructose-2,6-biphosphatase 2 (Pfkfb2)
   [[85]16], were significantly upregulated in Ubc9 NSCs ([86]Fig. 1G).
   Notably, the majority of the upregulated genes in Ubc9 NSCs were
   related to enhanced neurogenesis and neuronal function [e.g. actin
   alpha 2 smooth muscle aorta (Acta2); the regulator of neuronal
   intracellular trafficking Kx Antigen (Xk), the hormone Relaxin-1
   (Rln-1) and the sialic acid transporter Solute Carrier (Slc) 17a5]
   ([87]Fig. 1G).

   Gene Ontology (GO) enrichment analysis of the expression data showed
   that the six most enriched pathways affected by the Ubc9 transgene were
   related to neural development, function, or differentiation ([88]Fig.
   1H).

   These data suggest that the presence of the Ubc9 transgene in NSCs not
   only increases global SUMOylation levels, cellular growth rates, and
   self-renewal in vitro, but also has effects on gene expression,
   reducing transcripts related to apoptosis and increasing the expression
   of neurogenic genes.

3.2. Ubc9 NSCs display a quiescent metabolic profile upon differentiation and
are predisposed to neuronal differentiation in vitro

   To study the role of Ubc9 overexpression in the differentiation of
   NSCs, we cultured NSCs in the absence of epidermal growth factor (EGF)
   and fibroblast growth factor-2 (FGF-2) and examined their
   transcriptional profiles ([89]Fig. 2A and Data S2).

Fig. 2.

   [90]Fig. 2
   [91]Open in a new tab

   Ubc9 NSCs show a quiescent metabolic profile upon differentiation and
   are predisposed to neuronal differentiation in vitro. (A) Heatmap of
   genes differentially expressed in Ubc9 Diff vs WT Diff cells. N = 4 per
   group, adjusted p-value<0.05. See also Data S2. (B) Bar chart of the
   fold change (log2) in Ubc9 Diff vs WT Diff cells of the 10 most
   upregulated and downregulated genes. All genes have an adjusted
   p-value<0.05. See also Data S2. (C) Plot of the GO enrichment results
   for genes differentially expressed in Ubc9 Diff vs WT Diff cells. The
   x-axis shows the enrichment ratio, the color indicates the enrichment
   p-value (Kolmogorov–Smirnov statistic), and the size of each point
   indicates the number of significant genes in the corresponding GO
   category. The table at the bottom shows the 6 KEGG pathways showing the
   strongest enrichment (GAGE analysis p-value<0.05 for all pathways, see
   methods) in the same comparison. See also Data S2. (D) OCR vs ECAR
   graph of undifferentiated and differentiated WT and Ubc9 NSCs. Data are
   expressed as mean pmoles/min (OCR) and mean milli-pH units (mpH)/min
   (ECAR) ± SEM. N = 7 per group from 2 independent experiments.

   (E) Quantification and representative immunofluorescence-staining of WT
   Diff and Ubc9 Diff NSCs. Data are means ± SEM. N ≥ 6 per group from 2
   independent experiments, **p-value<0.01 (2-tailed Student's t-test).
   Nuclei are stained with DAPI, scale bars: 50 μm.

   Several genes were significantly downregulated in differentiated Ubc9
   NSCs (Ubc9 Diff), including genes with known oncogenic potential [e.g.
   the Serpin Family B Member 3 (Srpinb3d), the ETS Proto-Oncogene 1
   (Ets1), and the Transducin-like enhancer protein 2 (Tle2)] and genes
   playing a role in olfactory neuron signaling [vomeronasal receptors
   (Vmnr) and major urinary proteins (Mup)] ([92]Fig. 2B).

   Conversely, we found that genes primarily involved in cell metabolism
   were upregulated in Ubc9 Diff, including the high affinity carnitine
   transporter Slc22a21, the Cytochrome C Oxidase Subunit 7A1 (Cox7a1) and
   the Heme Binding Protein 2 (Hebp2) ([93]Fig. 2B). A GO enrichment and
   Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment
   analysis confirmed that metabolic pathways were most-highly enriched in
   Ubc9 Diff ([94]Fig. 2C).

   When we analysed the metabolic profile of Ubc9 Diff, we observed a
   remarkably low oxygen consumption rate (OCR) and extracellular
   acidification rate (ECAR) ([95]Fig. 2D), which were both suggestive of
   quiescent metabolism.

   Phenotypically, spontaneous in vitro differentiation of Ubc9 NSCs
   yielded similar fractions of glial fibrillary acid protein (GFAP)^+ and
   O4^+ glia compared to WT Diff. However, the fraction of Tuj1^+ neurons
   was significantly higher (2.3-fold increase) in Ubc9 Diff ([96]Fig.
   2E).

   Critically, we demonstrate that Sox2 – a key transcription factor
   essential for maintaining self-renewal of stem cells - is present in
   NSCs and that SUMOylation of Sox2 is upregulated in Ubc9 NSCs, compared
   to wild-type control NSCs (Fig. S2). As such, its inhibited function
   via SUMOylation might account for the enhanced neuronal phenotypes seen
   in Ubc9 NSCs in vitro [[97]17].

   Thus, upon differentiation in vitro, Ubc9 NSCs display increased
   neuronal differentiation with changes in genes related to cell
   metabolism, which may underlie the homeostatic metabolic profile of
   these cells.

3.3. Ubc9 overexpression induces gene expression changes that protect NSCs
against OGD/ROG in vitro

   To determine if any of the biological differences induced by the Ubc9
   transgene would result in enhanced survival within the ischemic
   microenvironment, we exposed NSCs to OGD/ROG, and analysed gene
   expression as well as functional responses.

   OGD/ROG induces drastic changes of genes belonging to several metabolic
   and biosynthetic pathways, including the hypoxia-inducible factor
   (HIF)-1α and the mitogen-activated protein kinase (MAPK) pathways (Fig.
   S3 and Data S2). Hence, to identify only those genes that responded
   differently to OGD/ROG in Ubc9 NSCs only, we applied an interaction
   model that tested the null hypothesis of no difference in fold changes
   after OGD/ROG ([98]Fig. 3A).

Fig. 3.

   [99]Fig. 3
   [100]Open in a new tab

   Ubc9 gain of function NSCs are protected against OGD/ROG in vitro. (A)
   Heatmap showing expression profile (z-score normalized intensities) of
   genes with a statistically-significant interaction (adjusted
   p-value<0.05) between the effect of OGD/ROG vs normoxia (control) and
   the effect of Ubc9 overexpression vs WT NSCs. The colored dendrogram
   shows the three expression clusters described in the text. See also
   Fig. S2, S3, S4 and Data S3. (B) Representative density plot and
   quantification of WT and Ubc9 undifferentiated NSCs post-OGD/ROG for
   live/dead analysis. Data are means ± SEM. N = 11 per group from 3
   independent experiments, ***p-value<0.001 (2-tailed Student's t-test).

   (C) Representative density plot and quantification of WT and Ubc9
   differentiated NSCs post-OGD/ROG for live/dead analysis. X axis
   (Hoechst blue), Y axis (PI). Data are means ± SEM. N ≥ 8 per group from
   3 independent experiments, *p-value<0.05 (2-tailed Student's t-test).
   (For interpretation of the references to color in this figure legend,