Fertilized specific pathogen-free (SPF) white leghorn chicken eggs (Charles River Laboratories, Wilmington, MA, USA) were maintained at room temperature (25°C) for 2 h before being placed in an incubator at 37.5°C with 60% humidity. The eggs were incubated for 96–100 hr (~4 days) to procure chick embryos at developmental stage embryonic day 1–4 (E1–E4) or Hamburger-Hamilton stages (HH) 12–22.

Subretinal injections and in ovo electroporation of embryos at E4 were performed as per the literature [35, 36] and shown in Figure 2. Briefly, the amnion and vitelline membranes were carefully removed with fine, sterile forceps to expose the embryo. For retinal injection, capillary glass pipettes (1.4 mm diameter; Drummond Scientific Company, Broomall, PA, USA, 2-000-001) were pulled with a microelectrode puller to get a 20 mm taper tip. The tip was broken under a dissecting microscope using a tweezer to obtain a tip opening approximately 0.1 μm in diameter. The microneedle was loaded with a mixture of 500 μM of the Top2b inhibitor ICRF-193 (Santa Cruz Biotechnology, Dallas, TX, USA, 21416-68-2) dissolved in HIBERNATE^® media (Gibco, Thermo Fisher Scientific, Waltham, MA, USA, A1247501). Many contemporary studies have applied the commercial molecule tested to inhibit both Top2b and Top2a [37, 38]. However, our study focused on cell migration, which is a cell process known to be regulated by Top2b rather than Top2a [39, 40]. Further, Top2a expression is mainly observed in proliferating cells, rather than in the post-mitotic cells used in this study [41]. Hence, the inhibitory effect of ICRF-193 in postmitotic cells that undergo migration is largely specific to Top2b. This commercial inhibitor was chosen because of its widespread usage in both in vivo and in vitro projects [40, 42, 43]. Further, the concentration used in testing was selected to corroborate published studies [40, 42, 43].

The reporter pCAG-GFP plasmid solution with a concentration ranging 3–6 μg/μL and 0.2 μL of fast green (0.025%) (MilliporeSigma, Burlington, MA, USA, 2353-45-9) was attached to a 0.1 mL syringe (Hamilton Company, Reno, NV, USA, Gastight Model 1710) for visual confirmation of solution delivery and mounted onto a manual micromanipulator (WPI, Sarasota, FL, M3301-M3-R) with a small piece of masterflex silicone tubing on the syringe needle. The mixture was injected directly into the embryonic subretinal space, and the embryo was electroporated with 5 square pulses of 15 V for 50 ms with 950 ms intervals using a pulse generator (BTX Harvard Apparatus, Boston, MA, USA, ECM 830). GFP expression was used as a marker to locate retinal tissue affected by the Top2b inhibitor ICRF-193.

Chick embryos post-E4 retinal injections were harvested from the shell and euthanized by decapitation at E6 and E12. The eyes were harvested and transferred to a petri dish of cold 1× PBS (phosphate buffer saline; Thermo Fisher Scientific, Waltham, MA, USA, BP3994). The retinal pigmented epithelium [9] was carefully dissected using fine forceps, and the cornea, lens, and vitreous were removed. The retinal tissue was then fixed in 4% paraformaldehyde (PFA) (Thermo Fisher Scientific, Waltham, MA, USA, AA433689L) in 1× PBS overnight at 4°C, followed by three times cold PBS wash for 10 min at 4°C and then infiltrated overnight in 30% sucrose (in 1× PBS). Embryos injected with the GFP-reporter plasmid constructs were verified for successful transfection via retinal inspection with a fluorescent dissection microscope (Leica Microsystems, Wetzlar, Germany, Leica MZ16FA), before embedding the retinal tissues in a solution of optimal cutting temperature compound (OCT, Electron Microscopy Sciences, Hatfield, PA, USA). Tissues embedded in OCT were stored at –80°C until ready for sectioning. Embedded tissue samples were sectioned using a cryostat to obtain sections of 12-μm-thickness (Thermo Fisher Scientific, Waltham, MA, USA, 0620E) and mounted on super-frost slides (Thermo Fisher Scientific, Waltham, MA, USA, 12-550-15S24), followed by air-drying. Immunohistochemistry using cell-specific markers was performed immediately afterwards.

Immunostaining was performed using Shandon Slide Rack (Thermo Fisher Scientific, Waltham, MA, USA). Slides containing tissue sections were placed on a warming plate for 15 min to remove excess OCT medium (Fisher Healthcare, Waltham, MA, USA, 23-730-571). Sections were then fixed with drops of 4% PFA, followed by 3 washes of 1× PBS for 10 min. Retinal sections were incubated for 1 hr in blocking solution [0.05% Triton X-100, 3% BSA, 10% goat serum or donkey serum (VWR, Radnor, PA, USA, 80054-446), in 1× PBS] at room temperature, followed by primary antibody overnight. Primary antibodies and dilutions used were as follows: mouse anti-Brn3a (1:200; MilliporeSigma, Burlington, MA, USA, AB5945) mouse, mouse Tuj1 (1:1,000; Abcam, Cambridge, UK, ab18207), mouse Calretinin (1:2,000; Chemicon, Rolling Meadows, Illinois, USA, MAB1568) and mouse Parvalbumin (1:1,000; Millipore, Burlington, MA, USA, MAB1572). Opposing control slides were not applied with any primary antibody. All slides were then washed with 1× PBS followed by the application of a secondary antibody carrying fluorescence from the appropriate host (mouse 594 nm, 1:300 dilutions; Jackson Immuno Research, West Grove, PA, USA, 715-585-150) for 1 hr. The slides were washed with 1× PBS and air dried, and a coverslip was placed on top to seal the samples using Cytoseal-60 (Thermo Fisher Scientific, Waltham, MA, USA, 50-252-72) for image analysis.

Retinal progenitor cells were commercially purchased (R28; Kerafast, Boston, MA, USA, EUR201) and cultured for testing. Growth media was prepared under sterile conditions using 89% Dulbecco’s Modified Eagle’s Medium (DMEM; ATCC, Manassas, VA, USA, 30-2002), 10% fetal bovine serum (FBS; Thermo Fisher Scientific, Waltham, MA, USA, 26140087), and 1% Pen-Strep (Gibco, Thermo Fisher Scientific, Waltham, MA, USA, 15-140-122). Cells were grown in ventilated, T-25 flasks (VWR, Radnor, PA, USA, 29185-300) for 4–5 days until near confluency. Growth medium was then aspirated, and cells were incubated with 3 mL of Trypsin (Gibco, Thermo Fisher Scientific, Ontario, CA, 25200-072) at 37°C for 5 min to dislodge adherent cells. Trypsin was then neutralized with 8 mL of R28 media and cells were pipetted into a 15 mL centrifuge tube and centrifuged at 1,500 rpm for 5 min. The supernatant was aspirated, and the cell pellet was resuspended in 3 mL of fresh R28 growth medium. The cells were then re-seeded in a new T-25 flask at 10⁶ cells per mL and incubated at 37°C and 5% CO₂ until confluency.

A series of 24-well plate (VWR, Philadelphia, PA, USA, 29442-044) were coated with 15 μg/mL laminin (LM) (Corning, NY, USA, 354239) and incubated for 24 hr. Confluent cells were plated at a density of 0.5 × 10⁵ cells/mL with MEM media. Cells were incubated for 24 hr and allowed to attach. Transductions were performed using a lentiviral vector (Origene, Rockville, MD, USA, TL516385V, labeled with GFP) with a multiplicity of infection (MOI) of 1.5. The lentiviral vector was mixed with the medium, added to the cells, and incubated for 24 hr. The media was replaced after 24 hr and cells were fixed with 4% PFA (Thermo Fisher Scientific, Waltham, MA, USA, AA433689L) and stained for the desired receptor expression. Transfections were performed using the Lipofectamine™ 3000 Transfection Kit (Thermo Fisher Scientific, Carlsbad, CA, USA, L3000015, labeled with GFP). The Lipofectamine 3000 agent was diluted in MEM media. The master mix was prepared by diluting the DNA (VectorBuilder, Chicago, IL, USA, VB900007-6027zju) with MEM media and adding the P3000 reagent and was incubated for 10–15 min. The master mix was added to the cells and incubated for 4 days. Live transfected cells were then selected for antibiotic resistance using a 2.5 mg/mL solution of puromycin dihydrochloride (Gibco, Thermo Fisher Scientific, Waltham, MA, USA, A1113803), as per Figure S1.

Cells were seeded in 24-well plates (VWR, Philadelphia, PA, USA, 29442-044) at a concentration of 2.5 × 10⁵ cells/mL and allowed to attach for 24 hr. Solutions of media from each well were removed, wells were washed 3 times with Dulbecco’s phosphate-buffered saline (DPBS; Sigma-Aldrich, Allentown, PA, USA, D8537), and cells were fixed with cold PFA (4%) for 5 min. Then, wells were washed with DPBS for 5 min twice at room temperature. Blocking buffer solution (0.05% Triton X-100, 2% donkey serum, and 3% BSA in DPBS) was added to each well for 15 min at room temperature. Following, wells were washed twice with DPBS for 2 min, and a primary antibody for the receptors was added to each well for overnight incubation: CXC chemokine receptor 4 (CXCR4; Invitrogen, Carlsbad, CA, USA, PA5-19856), fibroblast growth factor receptor (FGFR1; Invitrogen, Carlsbad, CA, USA, PA5-104788), and vascular endothelial growth factor receptor (VEGFR1; Proteintech, Rosemont, IL, USA, 13687-1-AP). The next day, each well was washed 3 times with 1× DPBS for 2 min, followed by the addition of secondary antibodies (Alexa Fluor 647 AffiniPure; ImmunoResearch Laboratories, Inc., West Grove, PA, USA, 711-605-152) 1 hr at room temperature. Wells were washed 3 times with 1× DPBS for 2 min before addition of DAPI (1:1,000; Thermo Fisher Scientific, Carlsbad, CA, USA, D1306) into each well for 5 min at room temperature. Each well was washed with DPBS 3 times for 2 min.

Boyden chamber assays (Sigma-Aldrich, St. Louis, MO, USA, ECM506) were used to examine the motility of each cell group toward a given concentration of different growth factors. Cells were seeded in the upper chamber of the well assay and allowed to settle on the upper surface of the membrane overnight. Growth factor solutions were then added to the media in the lower reservoir, and cells were left to migrate for 6 hr to through the transwell membrane, as described previously by our group and many others (Reviewed in [44]).

Transwell inserts used a polyester (PET) membrane (VWR, Radnor, PA, USA, 89235-020) of 10 μm thickness and 8.0 μm pore size. Approximately 600 μL of each growth factor vascular endothelial growth factor (VEGF; R&D Systems, Minneapolis, MN, USA, 564-RV), fibroblast growth factor-8 (FGF-8; Invitrogen, Carlsbad, CA, USA, PHG0184), and stromal derived factor 1-a (SDF-1a; Sigma Aldrich, St. Louis, MO, USA, SRP3276) was added to the bottom of separate transwell assays. Cells were seeded on the membrane at a concentration of 2.5 × 10⁵ cells/mL within media at 5% CO₂ and 37°C and left to incubate for 6 hr. After incubation, cells remaining on the top membrane were removed using a wet cotton swab. Control cells were stained with calcein-AM (Thermo Fisher, Burlington, ONT, CA, C1430) and reconstituted in PBS. Only live, fluorescent cells validated via optical microscopy were used for experiments.

A parameter called the Cell Migration Index, CMI, was defined to numerically compare the results of motile cells in response to different chemoattractants and control experiments (medium only). CMI is defined as the ratio of average numbers of motile cells in response to an exogenous growth factor and the average number of motile cells in control wells, as shown:

(1) CMI i = n i N control

where the subscript i denotes the reagent in a specific well measured, n_i represents the number of motile cells in that well, and “N_control” is the average number of motile cells in control wells, per experiment.

Concentrations of 15 μg/mL LM (Corning, Corning, NY, USA, 354239), 20 μg/mL poly-L-lysine (PLL) (Sigma Aldrich, St. Louis, MO, USA, P4707), and 20 μg/mL collagen IV (CIV) (Sigma Aldrich, St. Louis, MO, USA, C6745-1ML) were used to coat inner surfaces of 24-well plates (Falcon Corning, Corning, NY, USA, 353047). Uncoated polystyrene wells (VWR, Philadelphia, PA, USA, 29442-044) were used as a control. The plate was incubated at 37°C for 1 hr. Then, cells were seeded at 45,000 cells/mL in a 24-well plate. Five images were taken of each well using brightfield. This was repeated every 4 hr for 24 hr. A parameter called Cell Shape Index (CSI) was used to evaluate cell morphology, defined as:

(2) CSI = 4 P 2 A

where P is the cell perimeter and A is the cell area. A CSI value of 0 represents a purely circular cell while a value of 1 represents a purely elongated cell [45].

Expression levels of four genes encoding adhesion molecules were measured using quantitative, real time polymerase chain reaction (qT-PCR): cadherin 2 (Cdh2), cadherin 6B (Cdh6B), cadherin 7 (Cdh7), and cadherin 8 (Cdh8). Primer specificity was verified using Basic Local Alignment Search Tool (BLAST), which confirmed the selected forward and reverse primers listed. A FASTA sequence of desired species specific (Gallus gallus) mRNA was acquired from NCBI nucleotide database as shown in Table 1.

Primer pairs were selected based on PCR product size (200–300 bp) and analyzed using IDT’s oligo analyzer tool. Primers were examined for hairpin loop structures, high ∆G values and annealing temperature range T_m values for specificity and reliability. Relative gene expression data were analyzed using the conventional 2^–ΔΔCt (DDCT) method [46].

Microscopic images were analyzed using an upright fluorescence microscope (Zeiss Axio Imager A1) with a monochrome digital camera Axiocam MRM (Zeiss, Oberkochen, Germany). Images were acquired to confirm cell-specific markers in 549 nm, 594 nm, and 647 nm channels. Images were captured at the same exposure and threshold, and at the same intensity per condition. Cells that were modified to overexpress or knockdown Top2b expression were evaluated for CSI, surface area, and migration if they displayed a GFP tag after 24 hr of incubation. Each well was imaged five times, and ImageJ software was used to evaluate 5 cells per image for CSI, surface area, and migration distance. Fluorescing cells were counted using ImageJ.

Each in vitro data set was collected from replicates using n = 4 per condition, with five images collected per well containing at least five cells per image. A Shapiro-Wilks test was performed to assess the normality of the data. Cell migration was analyzed with two-way ANOVA with Tukey’s post-hoc test. Surface area index and CSI were analyzed with two-way ANOVA with Dunnett’s post-hoc test. Cell receptor expression was evaluated using Kruskal-Wallis rank test with Dunn’s multiple comparisons post-hoc test. Significance was denoted as (*) p < 0.05, (**) p < 0.01, (***) p < 0.001, and (****) p < 0.0001.

The first set of tests examined the effects of pharmacological Top2b inhibition using the well-established chick model of retinal embryonic development [47]. Tests combined our previous methodology of in ovo retinal injection with electroporation to deliver ICRF-193, a known potent topoisomerase II inhibitor [48]. Figure 3 depicts histological sections of developing chick retina at embryonic day E6 and E12, treated and untreated with the Top2b inhibitor ICRF-193. Images of E6 illustrate the nuclei (DAPI staining in blue) of neuronal cells within the outer (ONBL) and inner neuroblastic layer (INBL), which are early developmental stages of the ONL, INL, alongside the GCL. Tests first examined differences in the neuronal marker beta tubulin III (Tuj1), a structural component of the microtubule network expressed during the earliest stages of neuronal differentiation in the chick retina [49]. As shown (Figure 3A, 3B), both the number and intensity of Tuj1⁺ retinal cells (red) in the INBL were significantly reduced via Top2b inhibition compared to control retina. Top2b inhibition additionally disrupted cellular arrangement of the INBL layer at E6, as its thickness was significantly reduced compared to control and INBL cell arrangement depicted wide spacings not observed in control retina. Images gathered at the later differentiation stage of E12 illustrate nuclear staining of neuronal cells (DAPI in blue) within the developed retinal ONL, INL, and GCL. In addition, OPL and IPL separate the cell layers at this developmental stage, as marked. Data (Figure 3C, 3D) illustrate no significant differences observed in the number of Tuj1⁺ cells between Top2b-treated and control retinas at E12. Moreover, differences in the thickness of GCL layers were insignificant (as shown by arrows).

Retinal slices were next examined for differences in expression of the marker Brn3a, a transcription factor critical to the differentiation of dendritic arbors that regulate function in retinal ganglion cells [50, 51]. Images at E6 (Figure 3E, 3F) reveal that Brn3a⁺ cells (red) were located uniformly along the INBL in control retinas but observed to be unevenly dispersed along the INBL of Top2b-inhibited retina. Moreover, Brn3a⁺ cells were highly clustered in ICRF-193 treated retinas, but not in control (denoted by arrows). By E12, however, Top2b-inhibited retinas and illustrated no significant differences in cell clustering within the GCL compared to control (Figure 3G, 3H).

Control and Top2b-inhibited retina were additionally studied for changes in the expression of calretinin and parvalbumin, two calcium-binding proteins that regulate calcium transport and neuronal excitability [52, 53]. Figure 4 illustrates decreases in the expression of calretinin (red) across cells of the INBL between control and ICRF-193 treated retina at E6. As shown, decreased levels were observed along a significantly thinner INBL (Figure 4A, 4B). However, no difference in calretinin expression or GCL thickness was observed by E12 (Figure 4C, 4D). Retinas stained for parvalbumin similarly showed lowered expression in cells of the INBL between control and ICRF-193 treated retina at E6. Moreover, arrows point to significant reduction of INBL thickness with lower expression of parvalbumin (Figure 4E, 4F). By E12, however, changes in parvalbumin expression within the GCL were insignificant across control and treated retina (Figure 4G, 4H).

Our study next examined differences in the gene expression of adhesion molecules known to be significant to cell to cell signaling and GCL formation [54]. Tests applied qRT-PCR between developing retina treated and untreated with Top2b inhibitor ICRF-193 to illustrate relative changes in the gene expression of four critical cadherin molecules in E6–E12 retina: Cdh2, Cdh6B, Cdh7, and Cdh8. Figure 5 illustrates that levels of gene expression for all cadherin molecules were lower for Top2b-treated retina than control from E6–E10. However, the relative expression of Cdh2, Cdh6B, Cdh7, and Cdh8 was insignificant (p > 0.05) between control and Top2b-inhibited retina by E12.

Experiments next manipulated a heterotypic, retinal progenitor cell line previously reported to contain a majority of retinal ganglion cells [55] to develop two distinct cultured cell groups with different profiles of Top2b expression. Top2b-KD were produced through in vitro transfection with a Top2b-shRNA (shTop2b) plasmid construct previously established by our group [29], and Top2b-OE were developed via transfection with a plasmid encoding Top2b cDNA. Retinal progenitor cells with unaltered Top2b expression are denotes as wildtype or Top2b-WT. Figure 6 illustrates Top2b-induced differences measured in the expression of 3 chemotactic receptors of interest to retina: VEGFR (receptor for VEGF ligand), FGFR (receptor for FGF-8 ligand), and CXCR4 (receptor for SDF-1a, also known as CXCL12). As seen, no significant differences (p > 0.05) were measured in the expression of VEGFR between Top2b-WT, Top2b-KD, and Top2b-OE cell groups. However, expression of FGFR was significantly higher for Top2b-WT than Top2b-KD (p < 0.01), while FGFR expression of Top2b-KD cells was not significantly different against Top2b-OE (p > 0.05). By contrast, Top2b-OE cells exhibited dramatic increases in the expression of CXCR4, with average intensity values that were 2.5× higher than Top2b-WT (p < 0.05). VEGFR and FGFR expressions in Top2b-OE cells were only slightly higher than Top2b-WT (p > 0.05).

Top2b-KD, Top2b-OE, and Top2b-WT cell groups were next exposed to external growth factors to examine the effects of Top2b expression on chemotaxis. Cell migration was evaluated using the Cell Migration Factor, CMI (Eq. 1), which normalized the number of motile cells of the Top2b-WT group towards the media only condition (no added growth factors, CMI = 1.0). Figure 7 shows that Top2b-WT cells migrated with a lower CMI value to FGF than to media control (p < 0.05), with a larger CMI value to VEGF relative to media control (p < 0.05). Cells of the Top2b-WT group exhibited migration with comparable CMI between SDF-1a and the media control (p > 0.05). Motile Top2b-KD cells produced lower CMI values than Top2b-WT cells, with a CMI value of 0.02 ± 0.058 compared to media control. Top2b-KD cells migrated in lower numbers to all growth factors studied, with a CMI of 0.13 ± 0.155 to FGF (p < 0.05), CMI of 0.15 ± 0.295 to VEGF (p < 0.05), and CMI of 0.08 ± 0.185 to SDF-1a (p = 0.05). Lastly, smaller total numbers of Top2b-OE cells migrated toward media control than those of the Top2b-WT cell group, with a CMI value of 0.56 ± 0.42. Top2b-OE cells migrated in higher numbers than Top2b-KD cells but in lower numbers than Top2b-WT cells to all growth factors relative to media control. Motile Top2b-OE cells displayed an average CMI of 0.19 ± 0.009 to FGF (p < 0.01), CMI of 0.32 ± 0.216 to VEGF (p < 0.05), and a CMI of 0.35 ± 0.151 to SDF-1a (p < 0.05).

The final set of experiments measured morphological differences between Top2b-KD, Top2b-OE, and Top2b-WT cell groups attached upon LM, PLL, CIV, and polystyrene (PST, used as control). These materials were selected for study because they (a) have each been well studied in mechanisms of cell adhesion [56], (b) are components of many contemporary biomaterials [26], and (c) have been examined previously by many groups with retinal progenitors for applications in transplantation [15, 18, 57].

Data represents cell morphology using the parameter of CSI commonly applied to denote varying levels of cell elongation significant to functional matrix adhesion [58]. Figure 8A shows that average CSI values of Top2b-WT cells were observed to decrease over 24 hr for all surface polymers used. As seen (Figure 8A-1), values of CSI upon PST (as control) were highest immediately after plating, with an initial value of 0.95 ± 0.06 that decreased to a minimum value of 0.49 ± 0.09 at 24 hr. CSI of Top2b-WT cells adhered upon CIV followed a similar pattern, with a maximum value of 0.87 ± 0.05 after plating and a minimum CSI of 0.44 ± 0.11 at 24 hr. By contrast, cells of the Top2b-WT group adhered upon LM with an initial CSI of 0.86 ± 0.09, but then decreased sharply after 12 hr. The CSI of these cells upon LM reached a minimum of 0.36 ± 0.15 after 24 hr. Lastly, the CSI values of Top2b-WT cells adhered upon PLL illustrated the largest and sharpest decrease in values, with a maximum of 0.77 ± 0.09 upon plating and a minimum of 0.36 ± 0.08 after 24 hr. In all cases, groups of Top2b-WT cells exhibited large changes in adhesion within the first 12 hr of overnight attachment, and illustrated small changes in the proceeding 12 hr. Statistical significance is shown per time point against PST (used as control).

The adhesion of Top2b-KD cells exhibited similar patterns of adhesion upon PST (control) surfaces with average CSI values of ~0.95 for all conditions. CSI values decreased upon each matrix substrate over 24 hr, with the lowest CSI of 0.42 ± 0.15 measured upon PLL (p < 0.05). In contrast to the group of Top2b-WT cells, adhesion of Top2b-KD cells upon LM illustrated slower changes over time. Further, adhesion of Top2b-OE cells displayed similarly decreasing values of average CSI upon matrix substrates but demonstrated a dramatic decrease for adhesion upon LM. As seen, while all CSI values approached 1.0 upon initial plating, the CSI of Top2b-OE cells upon LM decreased sharply after only 8 hr and reached a minimum value of 0.47 ± 0.09 after 24 hr (p < 0.05). Top2b-OE cells also exhibited increased CSI during adhesion upon PLL and CIV compared to PST (control).

Lastly, cell adhesion was assessed via changes in the average surface area of each cell group, as shown in Figure 8B. Top2b-WT cells exhibited the largest average surface area upon LM (p < 0.05), with statistically insignificant changes (p > 0.05) upon PLL or CIV (Figure 8B-1). By contrast, the highest average surface area of Top2b-KD cells was measured upon PLL (p < 0.05) (Figure 8B-2) and highest upon PST for Top2b-OE cells (Figure 8B-3), with insignificant differences on all other surface matrices (p > 0.05).