Generation of Pericytic-Vascular Progenitors from Tankyrase/PARP-Inhibitor-Regulated Naïve (TIRN) Human Pluripotent Stem Cells

Abstract

Tankyrase/PARP inhibitor-regulated naïve human pluripotent stem cells (TIRN-hPSC) represent a new class of human stem cells for regenerative medicine that can differentiate into multi-lineage progenitors with improved in vivo functionality. Chemical reversion of conventional, primed hPSC to a TIRN-hPSC state alleviates dysfunctional epigenetic donor cell memory, lineage-primed gene expression, and potentially disease-associated aberrations in their differentiated progeny. Here, we provide methods for the reversion of normal or diseased patient-specific primed hPSC to TIRN-hPSC and describe their subsequent differentiation into embryonic-like pericytic-endothelial “naïve” vascular progenitors (N-VP). N-VP possess improved vascular functionality, high epigenetic plasticity, maintain greater genomic stability, and are more efficient in migrating to and re-vascularizing ischemic tissues than those generated from primed isogenic hPSC. We also describe detailed methods for the ocular transplantation and quantitation of vascular engraftment of N-VP into the ischemia-damaged neural retina of a humanized mouse model of ischemic retinopathy. The application of TIRN-hPSC-derived N-VP will advance vascular cell therapies of ischemic retinopathy, myocardial infarction, and cerebral vascular stroke.

1. Introduction

Human vasculature arises during embryonic development at extraembryonic and embryonic locations from highly proliferative mesoderm-derived vascular progenitors (VP). VP generate pericytes and endothelial progenitor cells and form interconnected primary vitelline, embryonic, and placental circulations [1]. Embryonic-like VP, defined by the cell-surface markers CD31⁺ CD146⁺ CXCR4⁺ can similarly be generated from human embryonic stem cell (hESC) and human induced pluripotent stem cell (hiPSC), and appear to functionally phenocopy the mesodermal lineage that gives rise to embryonic human vasculature [2–7]. Such human pluripotent stem cell (hPSC)-derived VP generated endothelial (CD31⁺ CD34⁺ CD144⁺ CD105⁺) and pericytic (CD140b⁺ CD105^lo) populations (Fig. 1a, b) with highly efficient blood vessel-forming abilities (Fig. 1c–f) that regenerated ischemia-damaged adult retinal blood vessels in a humanized murine ischemia/reperfusion (I/R) model [2, 3].

Fig. 1.

Generation of embryonic pericytic-endothelial VP from primed hPSC. (a) Primed H9 hESC were differentiated to vascular mesoderm. CD31⁺ CD146⁺ CXCR4⁺ VP were further FACS-purified (red gate), and culture-expanded in EGM-2 medium into populations that contained distinct (b) endothelial CD144⁺ CD105+ CD34⁺ (purple-gated) and pericytic CD140b⁺ CD34⁻ CD105^lo (green-gated) populations. Expanded VP were assayed for endothelial function with (c) Ac-Dil-LDL uptake assays, and (d) In vivo Matrigel plug chimeric vascular network formation in NOD/SCID mice. Sections were stained with anti-human CD31 antibodies (brown). (e) Primed cord blood-hiPSC-derived CD31⁺ CD146⁺ VP formed assembled vascular tubes in collagen gel. (f) These tubes were imaged with transmission electron microscopy, which revealed cooperating endothelial-pericytic-like cells with bifurcations. L: lumen, n: nuclei

Although multipotent embryonic-like VP may offer patient-specific vascular therapies [8–11], such VP with regenerative capacities are rare and limited in proliferative expansion in the adult, especially in diabetics [12–17]. Furthermore, primed hPSC lines remain restricted in their ultimate utility for cell therapies by highly variable vascular differentiation efficiencies [3, 6, 18, 19] and poor in vivo functionality of the VP derived from them [2, 3]. This limitation is ultimately due to the developmental similarities of hPSC with primed murine post-implantation epiblast stem cells (mEpiSC) which adopt a spectrum of pluripotent states with highly variable lineage-primed transcriptomes and post-implantation epiblast epigenetic marks [6, 7, 20]. As a result, primed hPSC possess a restricted multi-lineage differentiation potential compared to naïve epiblast-like ESC [20]. Moreover, epigenetic aberrations in diseased states, such as diabetes, inhibit the efficient somatic donor cell reprogramming to a functional primed pluripotent state [12–14, 21, 22].

Naïve-like hPSC with pre-implantation epiblast phenotypes, decreased lineage priming, improved epigenomic stability, and higher functionality of differentiated progenitors may solve these obstacles. Several groups have reported various small molecule approaches that putatively captured human naïve epiblast-like pluripotent states that were more primitive than those exhibited by primed hPSC (reviewed in [20]). However, many of these human naïve states exhibited karyotypic instability, global loss of parental genomic imprinting, and impaired multi-lineage differentiation performance that was attributed to the small molecule cocktails that were employed. In contrast, Zimmerlin and colleagues demonstrated that these caveats could be avoided by culturing primed hESC and hiPSC in a cocktail termed “LIF-3i” comprising of the small molecule tankyrase/PARP (poly ADP ribose polymerase) inhibitor XAV939, the GSK3β inhibitor CHIR99021, the MEK inhibitor PD0325901, and LIF [6, 7].

The LIF-3i method rapidly reverted primed, conventional hiPSC and hESC to a stable Tankyrase/PARP inhibitor-regulated naïve hPSC (TIRN-hPSC) state that adopted biochemical, transcriptional, and epigenetic features of the human pre-implantation naïve epiblast. TIRN reversion conferred a broad repertoire of normal and diabetic conventional hiPSC with molecular characteristics unique to naïve pluripotency, including increased phosphorylated STAT3 signaling, decreased ERK phosphorylation, global 5-methylcytosine CpG hypomethylation, genome-wide CpG demethylation at ESC-specific gene promoters, and dominant distal POU5F1 (OCT4) enhancer usage. TIRN-hiPSC maintained normal karyotypes and were devoid of systematic loss of imprinted CpG patterns or irreversible demethylation defects reported in other naïve reversion systems that were attributed to prolonged culture with MEK inhibitors [6, 7].

More importantly, reversion of primed hPSC to TIRN-hPSC improved multi-lineage potency, eliminated interline variability of differentiation among primed hPSC, and, significantly, erased primed hPSC donor cell “epigenetic memory” (Fig. 2a) [2, 6, 7]. The improved functional pluripotency of TIRN-hPSC was directly potentiated by the inclusion of XAV939 to the classical 2i-based culture media. Interestingly, inclusion of XAV939 into other systems has now identified stem cells with expanded pluripotency capable of generating both embryonic and extra-embryonic lineages [23, 24]. Thus, TIRN-hPSC represent a new class of human stem cells with improved multi-lineage functionality and potentially high impact for regenerative medicine.

Fig. 2.

(a) Waddington landscape model for TIRN reversion (red) of primed hiPSC (blue), adapted from [2]. Epigenetic obstacles to VP differentiation may be overcome with molecular reversion to a TIRN epiblast-like state (red) with a developmentally more primitive epigenetic configuration. (b) Schema of paradigm for generating cGMP-grade, HLA-defined TIRN-hiPSC (from reprogrammed CD34⁺ hematopoietic progenitors) with improved multi-lineage vascular, neural, and retinal pigmented epithelial (RPE) differentiation for comprehensive regeneration of ischemic neural retina

In a further advance of this human naïve stem cell system, Park and colleagues reported that the functionality of VP generated from both normal and diabetic primed hiPSC was significantly improved following reversion to a TIRN epiblast-like state [2]. Primed diabetic hiPSC were reprogrammed from type-1 diabetic donor fibroblasts and stably reverted to diabetic TIRN-hiPSC. Embryonic diabetic naïve VP (N-VP) differentiated from diabetic TIRN-hiPSC expanded more efficiently, possessed more stable genomic integrity, and displayed higher in vitro vascular functionality than VP generated from isogenic primed diabetic hiPSC. Moreover, human diabetic N-VP survived, migrated, and engrafted in vivo into the deep vasculature of the neural retinal layers with significantly higher efficiencies than isogenic primed diabetic VP in a humanized murine model of ischemic retinopathy.

Analyses of CpG DNA methylation and histone configurations at developmental promoters of diabetic TIRN-hPSC and diabetic N-VP revealed tight lineage-specific gene expression, and a de-repressed embryonic state with high epigenetic plasticity. TIRN-hPSC generated N-VP from a diseased state with improved epigenetic plasticity and superior vascular regenerative function [2, 6, 7]. Thus, reprogramming of patient donor cells to TIRN-hPSC may erase acquired epigenetic aberrations sustained from dysfunctional donor cell or disease-associated epigenetic memory (Fig. 2a) [2].

Here, we describe how to generate prolific human CD31⁺ CD146⁺ CXCR4⁺ N-VP from TIRN-hPSC, and detail the methods for transplanting these cells therapeutically in vivo into a humanized murine model of ischemic retinopathy. Endothelial-pericytic N-VP possess superior capacities for restoring blood flow to ischemia-damaged neural retina, relative to primed VP that are differentiated from conventional, primed hPSC. Clinical-grade, patient-specific or HLA-banked N-VP with improved functionalities could be used for regenerating, stabilizing, or even reversing end-stage blinding complications. Moreover, efficient differentiation of N-VP in parallel with neural retinal progenitors from the same TIRN-hiPSC line could comprehensively regenerate multiple cell types rendered defective in an adult diseased ocular environment [25, 26] (Fig. 2b). This N-VP system could be further optimized to establish a paradigm for future comprehensive multi-lineage vascular repair of other diseased ischemic tissues (e.g., cerebrovascular stroke and myocardial ischemia) using either patient-specific or banked HLA-defined TIRN-hiPSC [27, 28].

2. Materials

2.1. TIRN-hPSC Cell Culture and N-VP Differentiation

1. Essential 8 (E8) medium (see Note 1): DMEM/F-12 with 2.5 mM L-Glutamine, 15 mM HEPES and 14 mM sodium bicarbonate, supplemented with 50–100 ng/mL recombinant human FGF2, 2 ng/mL recombinant human TGF-β1, 64 μg/mL L-ascorbic acid-2-phosphate magnesium, 14 ng/mL sodium selenite, 10.7 μg/mL recombinant human transferrin, and 20 μg/mL recombinant human insulin. Sterile filter. Store at 4 °C for up to 3 weeks. Prewarm before use.

2. Vitronectin-XF solution (10 μg/mL).

3. Versene or phosphate-buffered saline (PBS)-based enzyme-free cell dissociation buffer.

4. LIF-3i TIRN medium (see Note 2): DMEM/F-12 with GlutaMAX, supplemented with 20% KnockOut Serum Replacement (KOSR), 0.1 mM MEM non-essential amino acids, 1 mM L-Glutamine, 0.1 mM β-mercaptoethanol, 20 ng/mL recombinant human LIF, 3 μM CHIR99021, 1 μM PD0325901, and 4 μM XAV939. Sterile filter the complete medium. Store at 4 °C for up to 3 weeks. Prewarm before use.

5. LIF-5i TIRN adaptation medium: supplement LIF-3i TIRN medium with 10 μM Forskolin, 2 μM purmorphamine, and 10 ng/mL recombinant human FGF2. Sterile filter. Store at 4 °C for up to 3 weeks. Prewarm before use.

6. Murine embryonic fibroblast (MEF) medium: DMEM supplemented with 10% heat-inactivated fetal bovine serum (FBS), 1× MEM non-essential amino acids, 0.1 mM β-mercaptoethanol, 1 mM L-glutamine and 0.5% penicillin/streptomycin. Sterile filter the complete medium.

7. LIF-3i/5i TIRN hESC and hiPSC plate preparation (see Note 3): 6-well culture-treated plates pre-coated overnight with sterile 0.1% gelatin solution, and pre-seeded with mitotically inactive MEF. Rinse wells twice with PBS just before use.

8. Accutase.

9. TIRN cryopreservation medium (10% DMSO, 40% KOSR, 50% TIRN medium, and 5 μM Y-27632).

10. Vascular mesoderm specification medium (see Note 4): Sterile-filter Albumin Polyvinyl-alcohol Essential Lipids (APEL) low insulin medium [29] and supplement with 25 ng/mL human Activin A, 50 ng/mL human VEGF₁₆₅, 30 ng/mL human BMP4, and 1.5 μM CHIR99021. Store at 4 °C for up to 3 weeks. Prewarm before use.

11. N-VP differentiation medium (see Note 4): supplement sterile-filtered APEL low insulin medium with 50 ng/mL human VEGF and 10 μM SB431542. Store at 4 °C for up to 3 weeks. Prewarm before use.

12. Magnetic activated cell sorting (MACS) system for N-VP isolation (see Note 5): human CD31 microbead kit, LS columns, and MidiMACS separator. MACS wash buffer consists of PBS supplemented with 0.04% Bovine Serum Albumin (BSA).

13. N-VP expansion medium: supplement endothelial growth medium-2 (EGM-2; commercially available from Lonza) with an additional 25 ng/mL of human VEGF₁₆₅. Store at 4 °C for up to 3 weeks. Prewarm before use.

14. N-VP expansion plate preparation: 6-well culture-treated plates coated overnight with 10 μg/mL fibronectin solution.

2.2. Ocular Ischemic Reperfusion (I/R) Injury Mouse Model Reagents

1. 6–8-week-old male NOD/Shi-scid/IL-2Rγ^null (NOG) mice (see Note 6).

2. Deep anesthesia injectable solution: 0.9% sodium chloride solution, 50 mg/kg ketamine, 10 mg/kg xylazine. Anesthesia is maintained by two additional 50 μL intramuscular injections of 20 mg/mL ketamine solution.

3. Animal pupil dilation solution: 2.5% phenylephrine hydrochloride ophthalmic solution.

4. Ophthalmic topical anesthetic solution: 0.5% tetracaine hydrochloride.

5. Anterior chamber cannulation (see Note 7): ophthalmic surgical microscope, 30-gauge needle, silicone infusion line, sterile irrigating balanced salt solution.

6. Ophthalmic topical antibiotic ointment: Bacitracin zinc and Polymyxin B sulfate (AK-Poly-Bac).

7. Digital pico-injector (PLI-100; purchase from Harvard Apparatus, Holliston, MA, USA).

2.3. Whole Mount Retinal Preparation and Immunostaining Reagents

1. Tris-Buffered Saline (TBS), adjusted to pH 7.4 with HCl.

2. Fixative solution (see Note 8): 2% paraformaldehyde in TBS.

3. Permeabilization solution: 0.1% Triton X-100 in TBS.

4. Blocking solution: TBS, 2% goat serum, 1% BSA.

5. Antibodies (see Note 9): Cy3-conjugated anti-human nuclear antigen (HNA), rabbit anti-mouse Collagen IV, rat anti-mouse CD31, Cy3-conjugated goat anti-rabbit secondary antibody, and Alexafluor-647-conjugated goat anti-rat secondary antibody.

6. Fluorescent nuclear dye (e.g., 4′,6-diamidino-2-phenylindole (DAPI) or Hoechst 33342).

7. Glass slides and coverslips.

8. Confocal microscope with 10× objective and tile-scan feature.

2.4. Cross-Section Retinal Preparation and Immunostaining Reagents

1. Fixative solution: 0.1 M Sodium Phosphate, 2% paraformaldehyde, 5% sucrose, adjust to pH 7.2 with HCl.

2. Cryopreservation wash solution: 0.1 M Sodium Phosphate, 5% sucrose, adjust to pH 7.2 with HCl.

3. Cryopreservation high sucrose solution: 0.1 M Sodium Phosphate, 20% sucrose, adjust to pH 7.2 with HCl.

4. Cryopreservation embedding solution: 33% Optimal Cutting Temperature (OCT) compound diluted in cryopreservation high sucrose solution.

5. Liquid nitrogen-cooled Isopentane.

6. Cryostat microtome machine.

7. Glass slides and coverslips.

8. Wash buffer (see Note 10): PBS, 0.05% Tween 20, 50 mM Tris, 150 mM NaCl, adjust pH to 7.6 with HCl.

9. Blocking solution: PBS, 5% goat serum, 0.05% Tween 20.

10. Antibodies (see Note 9): Cy3-conjugated anti-human nuclear antigen (HNA), mouse anti-human CD34, mouse anti-human CD31, rabbit anti-mouse collagen type IV, rat anti-mouse CD31, Alexa488 conjugated goat anti-mouse secondary antibody, Alexa647-conjugated goat-anti rabbit secondary antibody, and Alexa647-conjugated goat-anti rat secondary antibody.

11. Fluorescent nuclear dye: 4′,6-diamidino-2-phenylindole (DAPI).

12. Mounting antifade medium (see Note 9).

13. Confocal microscope with 20× objective (see Note 9).

3. Methods

3.1. Chemical Reversion of Primed hPSC to TIRN-hPSC

Because the genetic background of hPSC lines strongly contributes to interline variability, rigorously assess isogenic cultures at matched culture time points when comparing primed hPSC to TIRN-hPSC (see Note 11).

1. Prepare all cell culture media within a sterile biological safety hood cabinet. Incubate all cells at 37 °C in a normoxic, humidified 5% CO₂ incubator.

2. Maintain primed hESC and hiPSC in feeder-free E8 medium on 6-well culture-treated dishes pre-coated with Vitronectin-XF solution (Fig. 3a; see Note 12). Replace medium daily. Passage cells every 5–6 days by either mechanical selection or non-enzymatic reagents, such as Versene or PBS-based enzyme-free cell dissociation buffer.

3. Replace E8 medium with LIF-5i TIRN adaptation medium (see Note 13) in 4- to 5-day-old (since the last passage) hPSC cultures. The cells should be at 60–70% confluence. The next day, wash the cultures with PBS and dissociate the cells using Accutase. Transfer all of the cells at a 1:1 ratio onto a PBS-washed, gelatin- and MEF-coated plate in LIF-5i TIRN adaptation medium. Maintain culture in LIF-5i TIRN adaptation medium for one passage (4–5 days) by replacing medium daily.

4. After 4–5 days, bulk passage the LIF-5i TIRN cells into LIF-3i TIRN conditions. Add 2 mL LIF-3i TIRN medium to each well of a gelatin- and MEF-coated PBS-washed 6-well plate. Dissociate LIF-5i TIRN adapted cells into a single cell suspension using Accutase. Resuspend the cells in LIF-5i TIRN medium and count them. Wash and transfer 0.5–1 × 10⁶ LIF-5i TIRN-adapted single cells into each well containing LIF-3i medium. Replace LIF-3i TIRN medium daily.

5. Passage TIRN-hESC/hiPSC in LIF-3i TIRN conditions for at least 6–7 passages using Accutase prior to use in functional or differentiation experiments (see Note 14). For each LIF-3i TIRN passage, transfer 0.2–0.5 × 10⁶ cells per well every 3–4 days in LIF-3i TIRN medium on gelatin- and MEF-coated plates (Fig. 3a). Replace LIF-3i TIRN medium daily. Cryopreserve excess TIRN-hPSC that are not used in functional assays (see Note 15).

Fig. 3.

Stepwise transition of isogenic primed hPSC to Tankyrase/PARP inhibitor-regulated naïve (TIRN)-hPSC culture conditions, and subsequent vascular directed differentiation protocol to N-VP. (a) Schematic of stepwise protocol for LIF-3i TIRN reversion of primed hPSC. (b) Schematic of APEL vascular progenitor (VP) differentiation system. AA: Activin A; B: BMP4; CHIR: CHIR99021; V: VEGF. Differentiated CD31-expressing N-VP are enriched with magnetic-activated cell sorting (MACS)

3.2. In Vitro Vascular Differentiation of N-VP from TIRN-hPSC

Efficient differentiation of TIRN-hESC/hiPSC into CD31⁺ CD146⁺ N-VP (see Note 16) has been achieved using either embryoid body-based or monolayer hemato-vascular differentiation systems [2, 6, 7]. This approach uses an improved monolayer vascular differentiation method to generate N-VP from TIRN-hPSC (summarized in Fig. 3b) and is based on a modified, previously published [2] version of the APEL medium system [29].

1. For comparative studies (see Note 17), maintain simultaneous isogenic cultures of parental primed hPSC (in E8 medium) and TIRN-hPSC in parallel (see Note 18). One to two days after passaging TIRN-hPSC or primed hPSC (see Note 19), start the differentiation directly (without re-priming steps; see Note 20) by switching the culture media to vascular mesoderm specification medium. Maintain the culture in mesoderm specification medium for 2 days (Fig. 3b).

2. Switch the culture medium to N-VP differentiation medium and differentiate the cells for 5–6 days (Fig. 3b). Replace N-VP differentiation medium every 2 days (see Note 21).

3. Dissociate adherent cells differentiated for 7–8 days (i.e., following 2 days vascular mesoderm specification plus 5–6 days N-VP differentiation) using Accutase. Remove cell clumps with a 40 μm cell-strainer, count the cells, and resuspend them at a concentration of 1 × 10⁷ cells/mL in N-VP expansion medium.

4. Label the N-VP cells using the human CD31 microbead kit (incubate 20 μL FcR blocking reagent followed with 20 μL CD31 microbeads for 15 min at 4 °C). Enrich the differentiation cultures for CD31-expressing N-VP using LS columns and a MidiMACS separator. Prepare the columns according to the manufacturer’s instructions. After applying labeled cells to the column, wash the column 3 times with MACS wash buffer. Centrifuge the cells at 200 × g for 5 min and resuspend in N-VP expansion medium. Count the cells (see Note 22).

5. Transfer the CD31⁺ N-VP cells onto fibronectin-coated plates in N-VP expansion medium at 0.1–0.2 × 10⁶ cells per well of a 6-well plate. Grow N-VP cells to 80–90% confluency for 7–9 days prior to in vivo injections into I/R-injured murine NOG eyes, or in vitro analyses (Fig. 4a) (see Notes 23 and 24).

Fig. 4.

In vitro differentiation of N-VP from TIRN-hPSC and in vivo engraftment into a humanized murine model of ischemic retinopathy. (a) Representative images of primed VP versus N-VP from a fibroblast-hiPSC after 1 passage (1 week) in EGM-2 medium post-CD31⁺ MACS purification, demonstrating higher human CD31 (hCD31) and Ki-67 proliferation antigen expression. (b) Humanized murine model of ischemic retinal disease for quantifying N-VP engraftment (schematic adapted from [2]). Left panel, I/R location at anterior chamber and site of human cell injections into the vitreous body. Right panel, timeline of in vivo engraftment analysis. Immunodeficient NOG mice with retinal ischemia-reperfusion (I/R) injury can be leveraged for testing human N-VP therapies by injection directly into the vitreous body. Intraocular pressure (IOP) is elevated to 120 mm Hg for 90 min in the anterior chamber, followed by vascular reperfusion, which results in retinal vasculature loss, and apoptotic death of retinal neurons ~7 days following I/R injury. Although this model does not exactly simulate the sequence of events of diabetic retinopathy, it produces the same pathology seen in all retinopathies: acellular capillaries and ischemic neuronal degeneration. (c) Representative stitched image of a whole-mount retina demonstrating survival of HNA⁺ N-VP cells in the superficial vascular layers of the I/R damaged retina at 7 days following intra-vitreal injection

3.3. Ocular I/R Injury and N-VP Injection into NOG Mouse Eyes

The humanized I/R ocular model utilizes transient high intraocular pressure to induce retinal ischemia-reperfusion injury (Fig. 4b), and was previously described [2, 3]. Set up an appropriate number of animals for statistically relevant analyses of retinal engraftment of human cells.

1. Perform deep anesthesia on NOG mice by intraperitoneal (IP) injection of deep anesthesia injectable solution following procedures in line with approved local institutional and national animal use and care requirements, and in accordance with the statement on Use of Animals in Ophthalmic and Vision Research by Association for Research in Vision and Ophthalmology.

2. Dilate pupils with 2.5% phenylephrine hydrochloride ophthalmic solution followed by 0.5% tetracaine hydrochloride eye drops as a topical anesthetic solution.

3. Cannulate the anterior chamber of one eye under microscopic guidance (see Note 25) with a 30-gauge needle connected to a silicone infusion line providing sterile balanced salt solution. Avoid injury to the corneal endothelium, iris, and lens.

4. Induce retinal ischemia by raising intraocular pressure (IOP) of cannulated eyes to 120 mmHg for 90 min by elevating the saline reservoir height (see Note 26).

5. Confirm ischemia by iris whitening and loss of retinal red reflex. Maintain anesthesia with two doses of intramuscular ketamine solution for up to 90 min. Withdraw the needle to normalize IOP, and confirm reperfusion of retinal blood flow by evaluating reappearance of the red reflex.

6. Leave the contralateral eye of each animal unmanipulated to serve as a non-ischemic negative control.

7. Apply ophthalmic topical antibiotic ointment.

8. Two days after I/R injury, inject expanded human N-VP cells into the vitreous body of I/R-treated eye. Use 50,000 cells in a maximum of 2 μL per eye in a 1:1 sterile mixture of EGM-2/PBS using a digital pico-injector.

3.4. Harvest of NOG Mouse Retinae

Human N-VP cell engraftment in the NOG mouse retinae can be detected directly with anti-human nuclear antigen (HNA) immunohistochemistry, along with murine vascular marker co-localization using anti-murine CD31 and collagen IV antibodies.

1. Euthanize animals by cervical dislocation (see Note 27) for retinal harvests according to approved local institutional and national animal use and care protocols (see Note 28). Prepare the appropriate number of retinae for downstream analyses of HNA-positive cell quantitation in whole-mount retinae and human vascular cell quantitation by cross-section evaluations at desired post-treatment intervals (e.g., 1, 4, 8, 12 weeks) following human N-VP (or primed VP) injections (i.e., 2 days post-I/R injury).

2. Enucleate eyes, remove cornea and lens, and carefully separate the retinae from the choroid and sclera with a blunt cyclodialysis spatula. Fix retinae in fixative solution (Subheading 2.3) overnight at 4 °C. Transfer tissues to sterile PBS before processing eyes for whole-mount retinae or cross-sections.

3.5. Preparation and Imaging of Flat Whole-Mounted NOG Mouse Retinae for Human Engraftment Analysis by HNA Expression

1. For whole-mount retinae, permeabilize cells by incubation with permeabilization solution in an Eppendorf tube for 15 min at 4 °C.

2. Thoroughly wash the tissues with TBS. Block free-floating retinae with blocking solution (Subheading 2.3) for 6–8 h. Incubate the tissues overnight at 4 °C in primary antibody solutions diluted in blocking solution: rabbit anti-mouse Collagen IV (1:100) and/or rat anti-mouse CD31 (1:50) to label basement membrane and endothelial cells of blood vessels, respectively.

3. On the next day, wash retinae with wash buffer, and incubate the tissues with secondary antibodies for 6 h at 4 °C. A goat anti-rabbit Cy3-conjugated secondary antibody can be used to detect the collagen IV primary antibody, and a goat anti-rat Alexafluor-647-conjugated secondary antibody can be used to detect the anti-CD31 primary antibody. Human cells can be detected using directly Cy3-conjugated anti-HNA (1:100). Counterstain nuclei using a fluorescent nuclear dye.

4. After removing excess antibody and nuclear dye with wash buffer, pie-cut retinae and place flattened retinae on glass slides (see Note 29). Mount the slides with a coverslip (see Note 30). Image flat-mount retinae with a confocal microscope (see Note 31). For each individual eye, the entire retina can be imaged by tile-scanning and stitching (7 × 7 frames, 10% overlapping).

5. Quantitate human (HNA⁺) cell vascular engraftment in stitched photomicrographs of whole-mount murine retinae using a microscopy analysis software (Fig. 4c) (see Note 32). First, conservatively delineate the whole retina by creating a region of interest (ROI) using the nuclear dye channel (see Note 33). Exclude any background in the Cy3 channel at the edges of the retina tissue by adjusting the ROI (see Note 34). Process the Cy3 HNA channel with the “smooth” function and create a mask by thresholding. Exclude non-specific areas outside the ROI. Quantitate the number of HNA⁺ nuclei in the Cy3 channel using the “analyze particle” plugin (if using ImageJ) by using standardized sequential corrections (see Note 35). Identify HNA⁺ nuclei as particle objects by automated counting using fixed size and circularity parameters.

3.6. Preparation and Imaging of NOG Mouse Neural Retinal Cross-Sections for Quantitation of Human Vascular Engraftment

Human CD34⁺ and CD31⁺ cell engraftment can be detected and quantified within the defined neural layers of the murine retina by analyzing retinal cross-sections with the following methods (Fig. 5).

Fig. 5.

Method of human CD31⁺ engraftment quantitation in the vasculature of the neural layers (ILM, GCL, IPL, INL, OPL, ONL) of ischemia-injured mouse retinae. Shown is a representative example of ROI defining the neural retinal layers evaluated for human chimeric vascularization with murine blood vessels. A method of regional separation of neural retinal layers was employed for quantification of human CD31⁺ (hCD31) or alternatively CD34⁺ cells via automated counts (Analyze particles function) using the Fiji distribution of ImageJ software

1. Dissect the anterior eye (cornea/iris) free by a circumferential cut at the limbus. Fix eyecups in fixative solution (Subheading 2.4) for 1–2 h at room temperature. Wash 3 times for 10 min in cryopreservation wash solution. Prepare for cryopreservation by immersion in gradients of sucrose using successive 2:1, 1:1, 1:2 mix ratios of cryopreservation wash solution and cryopreservation high sucrose solution [30]. Keep tissues in cryopreservation high sucrose solution overnight at 4 °C. The next day, hemisect eyes through the optic nerve, and embed the two halves in cryopreservation embedding solution. Freeze in isopentane cooled with liquid nitrogen. Serially cryosect retinal hemi-sections (8 μm thickness) using a cryostat microtome. Transfer sections onto glass slides and store at −80 °C.

2. Label equally interspaced microsections with either anti-human CD34 or CD31 antibodies (see Note 36). Let the cryopreserved slides stabilize at room temperature and wash 3 times for 5 min at room temperature in wash buffer (Subheading 2.4). Incubate retinal sections for 1 h at room temperature in blocking solution (Subheading 2.4). Remove excess blocking solution, and without washing, incubate the retinal sections overnight at 4 °C with either mouse anti-human CD34 (1:50) or mouse anti-human CD31 (1:50) antibodies that are diluted in blocking solution. Wash tissues 3 times in wash buffer. Incubate tissue with Alexa 488-conjugated secondary goat anti-mouse antibody (1:200) for 1 h at room temperature (see Note 37). Wash 3 times in wash buffer. Label sections with either rabbit anti-mouse collagen type IV (1:200) or rat anti-mouse CD31 (1:50) diluted in blocking solution for 1 h at room temperature. Wash 3 times in wash buffer. Incubate with Alexa 647-conjugated goat-anti rabbit (1:200) or goat anti-rat secondary antibodies (see Note 38) diluted in blocking solution for 1 h at room temperature. Counterstain nuclei with DAPI.

3. Mount retinal sections with a glass coverslip in mounting antifade medium and cure overnight in the dark. Acquire images using a 20× objective with a confocal microscope.

4. Quantitate the number of human CD34 or CD31 blood vessels within the defined neural layers of the murine retina using microscopic image analysis software (Fig. 5) (see Note 32). Create three ROIs using the DAPI channel as a template, and delineate the ganglion cell layer (GCL), inner nuclear layer (INL), and outer nuclear layer (ONL). The other regions are defined as intercalated around and between these three DAPI-defined ROIs. Switch to the Alexa 488 (either human CD34 or CD31) channel and process the image for quantitation using a fixed sequential series of defined parameters. We recommend applying the “smooth”, “despeckle”, and “filter (median)” functions (see Note 39). Proceed with thresholding the images (pixel values >50). Count the number of thresholded objects within and between ROIs (see Note 40). The number of human CD34⁺ or CD31⁺ blood vessels can be automatically captured and enumerated using the “Analyze particles” plugin in ImageJ (see Note 41). Cross-validate each automatic count by manually counting the objects within the thresholded images and exclude any duplicate objects (see Note 42). To increase the precision of human CD34 or CD31 expression quantitation, check the Alexa647 channel and score only the regions where the Alexa488⁺ signal is detected at chimeric human-murine blood vessels that also express either murine collagen IV (mColIV) or murine CD31 (mCD31).