Molecular Characterization of Human Retinal Progenitor Cells

Abstract

Purpose.: To examine the molecular profile of fetal human retinal progenitor cells (hRPCs) expanded in vitro and those grown in a co-culture system with mouse retina through the analysis of protein and gene expression and neurotransmitter-stimulated calcium dynamics.

Methods.: hRPCS were isolated from human retina of 14 to 18 weeks gestational age (GA) and expanded in vitro. Immunoblot, microarray, and immunocytochemistry (ICC) assays were performed on undifferentiated hRPCs and those co-cultured with mouse retinas for 2 weeks. Cell function was assessed by using calcium imaging.

Results.: The ICC results showed a gradual decrease in the percentages of KI67-, SOX2-, and vimentin-positive cells from passages (P) 1 to P6, whereas a sustained expression of nestin and PAX6 was observed through P6. Microarray analysis of P1 hRPCs showed the expression of early retinal developmental genes: VIM (vimentin), KI67, NES (nestin), PAX6, SOX2, HES5, GNL3, OTX2, DACH1, SIX6, and CHX10 (VSX2). At P6, hRPCs continued to express VIM, KI67, NES, PAX6, SOX2, GNL3, and SIX6. On co-culture, there was a significant increase in the expression of MKI67, PAX6, SOX2, GNL3, SIX3, and RHO (rhodopsin). Calcium imaging showed a functional response to excitatory neurotransmitters.

Conclusions.: Fetal-derived hRPCs show molecular characteristics indicative of a retinal progenitor state up to P6 (latest passage studied). They show a progressive decrease in the expression of immature markers as they reach P6. These cells are functional, respond to excitatory neurotransmitters, and exhibit changes in expression patterns in response to co-culture with mouse retina.

Stem and progenitor cells have been receiving considerable attention as potential tools for the restorative treatment of degenerative diseases affecting the retina and visual system. A sound understanding of the relationship between human retinal progenitor cells (hRPCs) and the normal development of the human eye and the ability to culture cells in vitro are prerequisites to arriving at an effective stem cell treatment. Molecular expression patterns of hRPCs can be better understood in the context of temporally ordered genetic events that lead to the development of the human eye. Different germ layers are formed during embryogenesis, the eye being derived from the ectoderm. Once committed to this lineage, a series of transcription factors act in a multistep cascade to drive the cells toward retinal tissue development. From this point forward, a number of transcription factors (some eye-specific) are activated to control the progressive restriction of lineage potential. These multipotent embryonic and fetal cells, fated to develop into the adult retina, are known as retinal progenitor cells (RPCs). Because of their ontogenetic role, the developmental population of RPCs must have the potential to differentiate into each of the six neuronal cell types (ganglion, horizontal, amacrine, cone, rod, or bipolar cells) or one glial cell type (Müller cell) present in the mature neural retina. ^1,2  

Retinal progenitor cells have been isolated from both neonate and adult retina from a number of different species: mouse, ^3,4 rat, ^5–7 pig ^8,9 and human. ^10–14 Gene/protein characterization of RPCs has thus far been performed primarily in rodents and pig. ^8,9 Like neural stem cells, RPCs have typically been enriched and characterized based on their ability to proliferate in vitro. ¹⁵ Isolation of RPCs involves removing the full neuroretina, dissociating the cells and plating this suspension of heterogeneous cells in culture. The majority of the post mitotic cells undergo cell death and are discarded from the culture, whereas the actively dividing cells continue proliferating as the passages progress. Successful isolation and expansion of RPCs therefore produces viable populations of cells with the potential to develop into functional retinal cells. 

In this study, we report for the first time the molecular characterization of freshly isolated and cultured human fetal-derived hRPCs. One of the important objectives of this work was to correlate RPC data acquired from animals and humans. It is essential to isolate RPCs at an appropriate developmental/genetic stage as it has been shown that the differentiation process is controlled by temporally ordered intrinsic and extrinsic factors that determine the capacity of the RPCs to differentiate into mature cells. ^16–19 Furthermore, we sought to determine the stability of gene and protein expression from initial isolation through late passage. An in-depth study of this type has not been conducted on fetal-derived hRPCs. In our results, undifferentiated human retinal progenitor cells expressed several known neurodevelopmental markers identified in other species, and this phenotype was retained in cultured hRPCs up to P6. Furthermore, the undifferentiated hRPCs responded to exogenous signals from postnatal mouse retina. In addition, hRPCs exhibited neural characteristics as evidenced by functional responsiveness to the excitatory neurotransmitters glutamate and N-methyl-d-aspartic (NMDA) acid. 

Methods

Isolation and Standard Cell Culture

Human fetal eyes (14–18 weeks GA) were obtained through therapeutic termination of pregnancy from Advanced Bioscience Resources (Alameda, CA) with informed consent and IRB approval. The eyes were dissected under sterile conditions and the neuroretina separated as described previously. ^3,9,12 Briefly, the eyes were rinsed in cold Hanks' buffered salt solution (HBSS), the sclera was punctured just behind the limbus, and the anterior segment was cut away. The vitreous was removed, and the neuroretina was peeled away from the pigment epithelium. Retinas from both eyes were minced and dissociated in sterile collagenase (0.1%; Sigma-Aldrich, Inc., St. Louis, MO) by gently mixing at 250 rpm for 20 minutes. The dissociated cell suspension was then centrifuged at 1000 rpm for 3 minutes, and any remaining undissociated tissue was processed through another cycle in fresh 0.1% collagenase. Four cycles were performed for each donation, to completely dissociate the retina. After centrifugation of the cell suspensions, the pellets were resuspended in culturing medium (Ultraculture; Lonza, Basel, Switzerland) supplemented with 10 ng/mL epidermal growth factor (EGF; Sigma-Aldrich), 20 ng/mL basic fibroblastic growth factor (bFGF; Invitrogen, Carlsbad, CA), 2 mM l-glutamine (Invitrogen), 50 μg/mL gentamicin (Invitrogen), and 5% fetal bovine serum (FBS, Invitrogen). The number of live and dead cells was counted by using the trypan blue assay (Sigma-Aldrich). The isolated cells were plated onto fibronectin-coated (100μg/mL) tissue culture flasks at a density of 9,000 to 13,000 cells/cm². The cells were cultured at 37°C in the presence of 5% CO₂. A full medium change without FBS was performed 24 hours after plating. 

hRPCs were placed in culturing medium (Ultraculture; Lonza) supplemented with 10 ng/mL EGF, 20 ng/mL bFGF, 2 mM l-glutamine, and 50 μg/mL gentamicin. They were fed every 48 hours by replacing half the volume with fresh medium and were passaged as they reached 70% to 85% confluence (every 3–4 days). Briefly, adherent cells were dissociated with 0.25% trypsin (Sigma-Aldrich) for 5 minutes at 37°C. Trypsin activity was neutralized by using trypsin inhibitor with 1% human serum albumin (HSA; Sigma-Aldrich) in DMEM-F12 (Invitrogen). The cells were centrifuged at 1000 rpm for 3 minutes and the pellet was resuspended in medium. Live and dead cells were counted at the time of each passage by using trypan blue and were plated onto new fibronectin-coated flasks at a constant seeding density of 9,000 to 13,000 cells/cm². 

hRPCs and Early Retina Co-culture

hRPCs from P2 were plated onto fibronectin-coated (100 μg/mL) six-well plates (BD Falcon; Bedford, MA) at a density of 13,000 cells/cm² 24 hours before co-culturing with dissociated postnatal (days 1–3) DS red mouse retinal cells (Jackson Laboratory, Bar Harbor, ME). For primary mouse retinal cell suspensions, the eyes were enucleated, and the retinas were isolated in nonmitogenic culturing medium (Ultraculture; Lonza) and triturated through a flame-polished glass pipette into a single-cell suspension. The dissociated cells were then centrifuged at 1000 rpm for 3 minutes. After centrifugation, the supernatant was aspirated, and the cells were resuspended in 1 mL of medium. The cells were further dissociated by titurating the suspension until it became homogenous and then dividing it into six equal portions and transferring it onto noncoated 0.4-μm pore cell culture inserts (BD Falcon) placed in the six-well plates with the hRPCs. Two six-well plates were used, one containing hRPCs only (control) and the other containing hRPCs adherent to the base of the six-well plate with dissociated mouse retina placed in the insert (co-culture system). Both plates were cultured for 2 weeks and were observed for viability by fluorescence microscopy. These co-cultured cells were then used for genetic profiling by microarray analysis, as described later. The results obtained for both the control and the co-cultured hRPCs were statistically compared by paired t-test. 

Immunocytochemistry

hRPCs were grown in fibronectin-coated (100 μg/mL) 16-well chamber slides at a density of 9,000 to 13,000 cells/cm². Undifferentiated cells were maintained in culturing medium (Ultraculture; Lonza) with mitogens (EGF and bFGF) for 1 day. The cells were fixed for 30 minutes in 4% paraformaldehyde in 0.1 M phosphate-buffered saline (PBS) and were blocked with 10% goat serum and 0.1% Triton-x (Sigma-Aldrich) for 1 hour. The cells were then immunolabeled overnight at 4°C with blue opsin (1:300; Chemicon, Temecula, CA), CRX (1:100; Santa Cruz Biotechnology; Santa Cruz, CA), GFAP (1:500; Chemicon), KI67 (Sigma-Aldrich; 1:100), nestin (1:200; BD Transduction Laboratories, Lexington, KY), PAX6 (DSHB; 1:250), recoverin (1:1000; Millipore, Billerica, MA), SOX2 (1:250; Chemicon) or vimentin (1:200; AbCam). Cells were subsequently incubated in Cy3-conjugated secondary antibodies (1:100; Jackson ImmunoResearch Laboratories, West Grove, PA) for 1 hour. The slides were later rinsed with PBS and mounted in antifade mounting agent with DAPI (Vectashield; Vector Laboratories, Burlingame, CA) and observed under a fluorescence microscope (Nikon, Tokyo, Japan). 

Immunoblot Analysis

Cells were cultured as just described. At 70% to 85% confluence, they were dissociated by using trypsin and were spun at 1000 rpm for 3 minutes. After centrifugation, the supernatant was discarded, and the cells were washed in 10 mL of sterile PBS. They were respun at 1000 rpm for 3 minutes, the supernatant was discarded, and the cells were homogenized in lysis buffer (50 mM Tris-HCl [pH 7.6], 150 mM NaCl, 10 mM CaCl₂, 1% Triton X-100, and 0.02% NaN₃) followed by centrifugation. The proteins in the supernatant were isolated, and their concentrations were determined by using a BCA protein assay (Pierce Chemicals, Rockford, IL). Equivalent amounts of protein (50 μg) were subjected to SDS-PAGE (8%–10% acrylamide), transferred to nitrocellulose, and probed with the following antibodies: β-actin (1:5000, used as a loading control; AbCam), CHX10 (1:1000; Chemicon), nestin (1:1000; BD Transduction Laboratories), and SOX2 (1:1000; Chemicon). Blots were cut and reprobed sequentially, visualized with ECL reagents (NEN, Boston, MA) and exposed to x-ray film (Kodak/Carestream Health; Bio Max Light Film, Rochester, NY). 

Microarray Analysis

To perform a comprehensive microarray analysis, we froze samples of isolated hRPCs at P0 (isolation), P1, P3, and P6 in 10% DMSO in medium at −150°C. The collected samples were revived, washed in 10 mL of sterile PBS, and repelleted. Total RNA from the hRPCs pellets was purified (RNeasy Kit; Qiagen, Valencia, CA). Biotin-tagged cRNA was synthesized, used to probe custom-printed oligo arrays via hybridization for 12 to 16 hours (SuperArray Bioscience Corp., Frederick, MD), and subsequently imaged by using chemiluminescent detection in combination with x-ray film detection (Eastman Kodak). The scanned films were imported to Web-based software (GEarray Suite, ver. 2.0; SuperArray Biosciences) for analysis. Each array included a set of internal housekeeping genes used to adjust and correct for loading. Qualitative absent/present values were evaluated based on 10-minute exposures. A signal that was visually above the background was defined as present. Quantification was assessed by using averaged values obtained from the Web-based software, and t-tests were performed (Prism, ver. 4; GraphPad, San Diego, CA). 

Reverse Transcription–PCR

Total RNA was extracted from cultured cells according to the manufacturer's instructions (RNeasy mini kit; Qiagen) followed by in-column treatment with DNase I (Qiagen). Reverse transcription was performed with reverse transcriptase (Omniscriptase; Qiagen) and random primers (Sigma-Aldrich). Amplification of RPS27A (ribosomal protein) served as the internal control. Amplification conditions were 40 seconds/94°C, 40 seconds/55°C, and 1 minute/72°C for 35 cycles. 

Calcium Imaging

hRPCs were rinsed with Ringer's solution containing (in mM) NaCl 119, KCl 4.16, CaCl₂ 2.5, MgCl₂ 0.3, MgSO₄ 0.4, Na₂HPO₄ 0.5, NaH₂PO₄ 0.45, HEPES 20, and glucose 19 (pH 7.4). The cells were then incubated in Ringer's solution containing 0.5 μm fura-2 tetra-acetoxymethyl ester, 10% pluronic F127, and 250 μm sulfinpyrazone for 40 minutes at 22°C. Fura-2 was excited by alternating 340- and 380-nm light with the use of a filter changer, under the control of commercial software (InCytIM-2; Intracellular Imaging Corp. Cincinnati, OH) paired with a phase-contrast microscope (Eclipse T5100; Nikon). A new ratio (340/380) image was obtained every 0.35 seconds as a measure of Ca²⁺ concentration. hRPCs were then stimulated with either 1 mM l-glutamate (Sigma-Aldrich) or 1 mM NMDA + 1 mM glycine (Sigma-Aldrich), and increases in cytosolic fura fluorescence were analyzed. 

Results

Microarray

We developed a focused and customized microarray composed of several retinal fate-specific genes. Among these were markers for immature retinal, mature retinal, and glial cells. In each of the arrays, the housekeeping gene RPS27A (a ribosomal protein that is normally present at a consistent level within each cell) was present and acted as the internal control. Of the 85 genes present on the microarray, we focused our analysis on 14 genes that were most relevant and showed a significant consistency or change in expression. Also, these genes have been widely accepted to specifically represent states of stemness, or differentiation, among retinal cells, and include CHX10, SIX6, SIX3, DACH1, OTX2, GNL3, HES5, HS1, SOX2, PAX6, NES, KI67, VIM, and RPS27A. These genes were used to evaluate six undifferentiated P1 samples and two undifferentiated donations at P1, P3, and P6. To facilitate interpretation, the blotted arrays were analyzed to identify the presence or absence of each genetic marker. 

To assess genetic homogeneity between different tissue donations, microarray analysis of donations identified as GS013, GS019, GS024, GS027, GS029, and GS030 were compared (Table 1A). Based on genetic expression patterns, these donations were assigned to two distinct groups. Group 1 (GS013, GS019, GS024, GS027, and GS030) donations expressed the following genes: VIM, MKI67, NES, PAX6, SOX2, OTX2, SIX6, and CHX10. In addition, GNL3, HES5, and DACH1 were detected in two of the samples within group 1. The microarray results for group 1 correlated with positive immunocytochemistry (ICC) labeling of KI67, SOX2, NES, VIM, and PAX6 (see Fig. 2). Group 2 consisted of the single sample population GS029. These cells failed to show the expression of the early developmental transcription factors SOX2 or CHX10. 

Table 1.

 

Summary of Two Groups of Microarray Data

 

 Symbols, gene is visually present; blank, gene is visually absent. (A) P1 (n = 6) primary cultures of hRPCs. Excluding donation GS029 (outlier), 100% of P1 hRPCs showed consistent expression of VIM (vimentin), K167, NES (nestin), PAX6, SOX2, OTX2, SIX6, and CHX10. P1 hRPCs showed variability in HES5, GNL3, and DACH1. (B) Undifferentiated hRPCs from two separate donations characterized at P1, P3 and P6. VIM, MKI67, NES, PAX6, SOX2, and GNL3 remained present from P1 to P6.

To address the phenotypic stability of hRPCs in culture, we compared two tissue donations from group 1 (GS024 and GS027) at P0, P1, P3, and P6 (Table 1B). The expression pattern of vimentin, KI67, NES, PAX6, and SOX2 remained consistent between the two populations (GS024 and GS027) across all the passages tested. The expression of SIX6 was not lost until P6 in the GS024-derived culture. On the other hand, the expression of OTX2, DACH1, CHX10, and HES5 was lost during the culture process. Of interest, the expression of GNL3, which was not present in P0 samples, was consistently maintained in both the cultured donations from P1 to P6. In addition, a table of complete microarray results indicates the total qualitative absence or presence of gene expression derived from microarray blot analysis (Table 2). 

Table 2.

 

Complete Array of hRPC(GS027) from P1, P3, and P6

Gene	027-P1	027-P3	027-P6	Gene	027-P1	027-P3	027-P6
RPS27A	X	X	X	VTN	–	–	–
VIM	X	X	X	DPYSL3	X	X	X
MKI67	X	X	X	RBP3	X	–	–
NES	X	X	X	PROM1	–	–	–
PAX6	X	X	X	UNC119	X	X	X
SOX2	X	X	X	TULP1	–	–	–
HES1	–	–	–	PDC	–	–	–
HES5	–	–	–	USH2A	–	–	–
DCX	–	–	–	SAG	–	–	–
LHX2	–	–	–	CRX	–	–	–
GNL3	X	X	X	RTN4	X	X	X
MSI1	–	–	–	RPGR1P1	–	–	–
OTX2	X	–	–	RTBDN	–	–	–
DACH1	–	–	–	PCDH21	X	–	–
SIX3	–	–	–	PLEKHB1	X	X	X
SIX6	X	X	X	PCDH15	–	–	–
PRKCA	–	–	–	AIPL1	–	–	–
CHX10	X	–	–	ENPP5	–	–	–
CHRNB3	–	–	–	ARR3	–	–	–
STMN2	X	X	X	HIST1H2BN	–	–	–
NRN1	X	X	X	GNB3	X	X	X
NEFL	X	X	X	AMPD2	X	X	X
STMN2	X	X	X	CLTB	X	X	X
POU4F1	–	–	–	CABP5	–	–	–
CPLX1	–	–	–	ADRB1	–	–	–
EPPK1	–	–	–	PDE6C	–	–	–
KIF5A	–	–	–	Blank	–	–	–
ATOH7	–	–	–	CLCA3	–	–	–
PROX1	–	–	–	VEGFA	–	–	–
RLBP1	X	X	X	VEGFB	X	X	X
GFAP	–	X	X	VEGFC	X	X	X
MBP	–	–	–	Blank	–	–	–
NRL	–	–	–	Blank	–	–	–
RHO	–	–	–	Blank	–	–	–
NR2E3	–	–	–	PUC18	–	–	–
REEP6	X	–	X	Blank	–	–	–
FSCN2	–	–	–	Blank	–	–	–
PPAP2C	–	–	X	AS1R2	–	–	–
PDE6A	–	–	–	AS1R1	–	–	–
GUCA1B	–	–	–	AS1	–	–	–
RDH12	–	–	–	GAPDH	X	X	X
RGS9BP	–	–	–	B2M	X	X	X
RCVRN	X	–	–	HSP90AB1	X	X	X
KCNB1	–	–	–	HSP90AB1	X	X	X
GNGT1	–	–	–	ACTB	X	X	X
PDE6G	–	–	–	ACTB	X	X	X
KPNA1	X	X	X	BAS2C	X	X	X
KPNA2	X	X	X	BAS2C	X	X	X

 Qualitative absence (−) or presence (X) of gene expression derived from analysis of microarray blots. Most of the genes show continued expression from P1 to P6. House keeping genes are indicated by RPS27A, APDH, B2M, HSP90AB1, and ACTB. Control (nongene) areas are indicated by Blank, AS1R2, AS1R1, AS1, and BAS2C.

Immunoblot

Western blot analysis was performed on various samples, and the protein was probed for PAX6, CHX10, SOX2, and nestin (Supplementary Fig. S1), all being positive in the samples studied. The results further support the findings obtained through microarray and ICC, with the exception of CHX10, which persisted through P3 on Western blot, but not on the microarray. 

Reverse Transcription–PCR

To further evaluate the presence of message detected using microarray, RT-PCR was used to evaluate the presence of message in hRPCs across passages (Supplementary Fig. S2). The genes evaluated included (2) DACH1, (3) SOX2, (4) OTX2, (5) GNL3, (6) HES5, (7) NES, (8) PAX6, (9) SIX6, (10) KI67, (11) VIM, and (D) RPS27A (control). Newly designed primers revealed a moderate to high level of correlation with array results. hRPCs isolated from P0, P1, P3, and P6 showed 70%, 77%, 57%, and 86% correlation, respectively, with message detected in the microarray. The SuperArray primer sequences are not publicly available, and the difference in percent detection may be attributable to a basic difference in primer sequences. 

Immunocytochemistry

ICC was performed to verify the results obtained from microarray and Western blot analyses. Undifferentiated P1 hRPCs were labeled for KI67, nestin, GFAP, PAX6, CRX, recoverin, SOX2, and vimentin. These cells stained positively for a range of proteins, indicating an immature retinal cell state. The proteins included the cell cycle marker KI67 (90.89% ± 2.93% [SEM]), the nuclear transcription factors SOX2 (97.62% ± 1.83%) and PAX6 (91.07% ± 1.87%), and the intermediate filaments vimentin (97.29% ± 1.57%) and nestin (100% ± 0%; Figs. 1, 2). Most of the cells also labeled for the master retinal fate gene, PAX6. In addition to the immature markers, a subpopulation stained positively for recoverin. As in the case of the microarray analysis, we performed ICC across progressive passages (P1, P3, and P6) to evaluate their phenotype based on a specific panel of immature (CRX, KI67, nestin, PAX6, SOX2, and vimentin) and mature (blue opsin, GFAP, recoverin) markers (Fig. 3). Blue opsin, CRX, and GFAP did not show positive staining patterns; whereas KI67, SOX2, and vimentin were broadly positive at P1 but showed a gradual temporal decline in expression toward P6 (52.20% ± 3.32%, 47.40% ± 3.53%, and 86.11% ± 1.55%, respectively), yet still retained expression at that passage point. However, nestin and PAX6 expression was maintained at a constant level through each of the passages. 

Figure 1.

 

ICC of hRPCs to identify expression of stemness and differentiation proteins. Undifferentiated hRPCs were stained with markers of mature (recoverin) and immature (SOX2, KI67, PAX6, nestin, and vimentin) retinal cells. (A, B) SOX2, (C, D) KI67, (E, F) PAX6, (G, H) nestin, (I, J) vimentin, and (K, L) recoverin.

Figure 2.

 

ICC of undifferentiated hRPCs from P1, P3, and P6 samples. Cells were stained for KI67 (A–C), SOX2 (D–F), nestin (G–I), vimentin (J–L), and PAX6 (M–O). (P) The percentage of cells labeled for KI67, SOX2, and vimentin decreased from P1 to P6, whereas the expression of both nestin and PAX6 remained fairly constant through all passages. Magnification, ×20.

Figure 3.

 

Focused microarray of co-culture (hRPCs with dissociated postnatal mouse retina) versus control (hRPCs) cultured for 2 weeks. (A) Graph indicating increases (x-fold) of genes in the co-culture (n = 3) compared with the control (n = 2) group. KI67, PAX6, SOX2, GNL3, SIX3, and rhodopsin RNA levels were significantly higher in the co-culture group. (B) An array from the control group. (C) An array from the co-culture group.

Co-culture

To assess whether the genetic profile of hRPCs could be influenced by their environment, we set up an experiment that compared control hRPCs with a co-culture system (hRPCs and dissociated postnatal mouse retina) based on microarray results. After 2 weeks in culture, both the control and co-cultured hRPCs were lysed, and the RNA was isolated and analyzed through our focused microarray. Included in Figure 3 is a condensed list of genes that showed an increased expression compared with the control group. Paired t-test showed significant increases in the genes KI67 (1.77 ± 0.12 [SEM]), PAX6 (1.20 ± 0.04), GNL3 (2.09 ± 0.21), SIX3 (2.76 ± 0.42), RHO (3.85 ± 0.67) (all P < 0.05), and SOX2 (2.05 ± 0.14; P < 0.01). 

Calcium Imaging

Intracellular calcium dynamics, in response to transmitter stimulation, were analyzed for P2 hRPCs cultured in mitogenic medium (Ultraculture; Lonza). The cells were plated at a low density to allow for high-resolution imaging of individual cells and recording of changes in fura-2 fluorescence, an indicator of intracellular calcium influxes. We found that a high percentage (75%) of hRPCs stimulated with 1 mM of glutamate responded with calcium influxes (Figs. 4A–C). Glutamate-induced averaged peak calcium influxes for 12 cells was 124 ± 25 nM [SEM]). A lower percentage (65%) of five hRPCs responded with calcium influxes in response to 1 mM NMDA+1 mM glycine (Figs. 4D–F). Averaged peak calcium influxes in response to NMDA reached 82 ± 29 nM. Responses to each transmitter peaked immediately after stimulation and returned to baseline within 200 seconds. 

Figure 4.

 

hRPCs responded to glutamate and NMDA+glycine with calcium influx. (A) Average response of (n = 12) fura-2-loaded hRPCs demonstrate an approximate 110 nm increase in intracellular Ca²⁺ in response to 1 mM l-glutamate. (B, C) Fura-2 loaded hRPC (circle) from (A) before and during l-glutamate stimulation, respectively. (C, D) Average response of (n = 5) hRPCs stimulated with 1 mM NMDA and 1 mM glycine showing an initial influx of approximately 60 nM with a gradual return to baseline. (E, F) Cell (circle) from (E) showing baseline fluorescence (black) followed by increased fura-2 fluorescence (gray) at the peak of transmitter stimulation, respectively.

Discussion

In this study, we used a custom-designed microarray, which is being newly evaluated. We recognize that the novelty of this microarray may limit the interpretation of the resulting data. Microarray results therefore are presented in addition to protein and functional analysis. We propose that, taken together, the data offer a window into hRPC expression and physiology. The identified characteristics should be considered relative to cell age, heterogeneity, and culture conditions, which may influence rates of transcription and translation. 

Microarray analysis of hRPCs revealed that during P1, the cells expressed the immature markers KI67, SOX2, GNL3, NES, VIM, PAX6, CHX10, OTX2, and SIX6 indicating that, in accordance with previously published data for porcine RPCs, ⁹ our isolated hRPCs were fated to the development of neural retinal tissue. Of these markers, KI67, SOX2, and GNL3 have been shown to be important proteins in cell cycle progression. Current data indicate that KI67 is a nuclear antigen that is expressed in all phases of the cell cycle (G₁, S, G₂, and M) except G₀. ²⁰ The expression of this gene along with the other cell cycle genes through P6 showed the ability of the hRPCs to proceed through active cell division during in vitro expansion. Based on our ICC data, the expression of KI67 and SOX2 decreased as these hRPCs reached P6 and indicated a progressive decline in the proliferative ability of these cells, a finding that is consistent with our observations of in vitro growth dynamics of hRPCs (UA, unpublished data, 2008). These cells remained in a proliferative phase during the first six to seven passages, doubling in number over the first two to three passages and giving rise to hundreds of millions of cells. The growth rate gradually declined as the cells progressed through higher passages, eventually reaching a plateau phase (after ∼1 month of culture), which was accompanied by a reduction in the proliferation gene KI67 (50%) and the absence of SOX2. In addition, expression of several genes remained, including VIM and NES, which may be indicative of glial cell fates at later passages. 

In addition to these cell cycle genes, we identified some that are more specifically associated with the development of the brain and the eye. For example, we found that SIX6 was present in sample GS024 up to P3 and in sample GS027 up to P6. The presence of SIX3 has been suggested to direct brain and eye development, whereas SIX6 is more specifically involved in eye development. Both SIX3 and SIX6 are in part responsible for the proliferation of cells. In the mouse, SIX6 has been shown to be involved in the specification and formation of the ventral optic stalk and later in the determination and differentiation of the neural retina. ²¹ Next, we showed that PAX6, a gene closely associated with ocular development, was present through P6, suggesting preservation of eye specific lineage restriction. ^22–24 We also demonstrated by gene microarray that CHX10 RNA ceased to be detected beyond P1, whereas CHX10 protein remained positive through P3 on Western blot analysis. Although it is challenging to establish a clear correlation between mRNA levels and protein abundance, analysis of both are valuable in understanding transient cellular processes involved in cell fate determination. ²⁵ Regardless, the co-expression of CHX10 and PAX6 is significant because only neural retinal progenitors co-express these genes during development. ^26,27 The expression of the genes CHX10, PAX6, and SIX6 through P3 and P6, in addition to the previously described proliferation-specific genes, indicates that the in vitro expansion techniques used for this study preserve hRPCs in an undifferentiated and proliferative state up to P6. The second group (donation GS029) expressed neither the SOX2 nor the CHX10 gene, reflecting a relative lack of multipotency in this population. Further evidence that cells from GS029 were developmentally more mature comes from the slower growth dynamics and rate of mitotic division observed in this particular donation (data not shown). 

By comparing co-cultured hRPCs with the control population, we observed increases in the expression of the markers KI67, PAX6, GNL3, SIX3, rhodopsin, and SOX2 in response to exposure to developing retina. The evaluated cell line represents a heterogeneous population of mitotically active progenitors isolated from 14- to 18-week retinas. Although subpopulations of cells of this developmental age express CRX in vivo, the percentage of viable cultured cells expressing CRX may be low, or alternatively low expression levels may not be detectable with current techniques. Finally, it is possible that the antibody was not effective in this application. It is interesting to note that other immature retinal markers were upregulated along with rhodopsin, a mature rod photoreceptor marker. The results suggest that a subpopulation of co-cultured cells were directed toward a mature photoreceptor state, whereas most continued to proliferate. The observed expression pattern of proliferating and differentiating cells probably results from heterogeneity within the cultured population of cells used in this assay. It should be noted that our microarray results did not show expression of recoverin or any other mature photoreceptor markers. This pattern could be due to the location of the cells in the differentiation timeline or the method of co-culture. Overall, these findings suggest that hRPCs possess intrinsic predictors of cell fate as demonstrated in other mammalian species. The interaction of exogenous factors with the internal response properties of the hRPCs appears to influence expression patterns and consequential fates. 

The response of hRPCs to glutamate and NMDA indicate that these cells express the predominant excitatory glutamatergic receptors found in the developing and mature retina. In addition, robust calcium influxes demonstrate in vitro viability and responses typical of a neural phenotype. These functional data support our ICC and microarray data, indicating a neural phenotype of these cells via the expression of the neural stem cell markers SOX2, SIX3, and NES. Taken together, these findings provide evidence that cycling hRPCs can be maintained in vitro with preserved genotypic and physiologic properties similar to normal retinal tissue. 

Supplementary Materials

Supplementary Figure S1 - (PDF) 

Supplementary Figure S2 - (PDF) 

Footnotes

 Supported by ReNeuron, the Lincy Foundation, and the Discovery Eye Foundation.

Footnotes

 Disclosure: S. Schmitt, ReNeuron (F); U. Aftab, ReNeuron (F); C. Jiang, ReNeuron (F); S. Redenti, ReNeuron (F); H. Klassen, ReNeuron (C, P); E. Miljan, ReNeuron (E); J. Sinden, ReNeuron (E); M. Young, ReNeuron (F, C, P)

Footnotes

 The publication costs of this article were defrayed in part by page charge payment. This article must therefore be marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.