Experimental Eye Research 91 (2010) 500e512
Contents lists available at ScienceDirect
Experimental Eye Research journal homepage: www.elsevier.com/locate/yexer
Loss of photoreceptor potential from retinal progenitor cell cultures, despite improvements in survival Fiona C. Mansergh a, *, Reaz Vawda a, b, Sophia Millington-Ward a, Paul F. Kenna a, Jochen Haas c, Clair Gallagher d, John H. Wilson e, Peter Humphries a, Marius Ader a, c, G. Jane Farrar a a
Ocular Genetics Unit, Smurfit Institute of Genetics, Trinity College Dublin, Lincoln Place Gate, Dublin 2, Ireland Fighting Blindness Vision Research Institute, 1 Christchurch Hall, Dublin 2, Ireland DFG-Center for Regenerative Therapies Dresden, Cluster of Excellence/TU Dresden, c/o MTZ, Fiedlerstr. 42, 01307 Dresden, Germany d National Institute of Cellular Biotechnology (NICB), Dublin City University, Glasnevin, Dublin 9, Ireland e Department of Biochemistry and Molecular Biology, Baylor College of Medicine, One Baylor Plaza, Houston, TX 77030, USA b c
a r t i c l e i n f o
a b s t r a c t
Article history: Received 6 May 2010 Accepted in revised form 7 July 2010 Available online 15 July 2010
Retinal degeneration (RD) results from photoreceptor apoptosis. Cell transplantation, one potential therapeutic approach, requires expandable stem cells that can form mature photoreceptors when differentiated. Freshly dissociated primary retinal cells from postnatal day 2e6 (PN2e6) mouse retina can give rise, post-transplantation, to photoreceptors in adult recipients. Unfortunately, incorporation rates are low; moreover, photoreceptor potential is lost if the same PN2e6 cells are cultured prior to transplantation. We investigated the identity of the cells forming photoreceptors post-transplantation, using FACS sorted primary postnatal day (PN) 3e5 Rho-eGFP retinal cells. Higher integration rates were achieved for cells that were expressing Rho-eGFP at PN3e5, indicating that post-mitotic photoreceptor precursors already expressing rhodopsin form the majority of integrating rods. We then investigated improvement of cell culture protocols for retinal progenitor cells (RPCs) derived from PN3e5 retinal cells in vitro. We succeeded in improving RPC survival and growth rates 25-fold, by modifying retinal dissociation, replacing N2 supplement with B27 supplement minus retinoic acid (B27 ! RA) and coating flasks with fibronectin. However, levels of rhodopsin and similar photoreceptor-specific markers still diminished rapidly during growth in vitro, and did not re-appear after in vitro differentiation. Similarly, transplanted RPCs, whether proliferating or differentiated, did not form photoreceptors in vivo. Cultured RPCs upregulate genes such as Sox2 and nestin, markers of more primitive neural stem cells. Use of these cells for RD treatment will require identification of triggers that favour terminal photoreceptor differentiation and survival in vitro prior to transplantation. ! 2010 Elsevier Ltd. All rights reserved.
Keywords: retinal progenitor cells cell culture FACS rhodopsin photoreceptor retina
1. Introduction Retinal degeneration (RD) involves the gradual loss of photoreceptors by apoptosis, causing visual impairment and eventual blindness. Inherited RD is genetically heterogeneous. To date, over 190 loci have been identified (RetNet; http://www.sph.uth.tmc. edu/Retnet/). The disability associated with various forms of RD carries a great social and economic cost. Retinitis pigmentosa (RP) affects 1 in 3000 people, while age related macular degeneration
* Corresponding author. Tel.: þ353 1 8962484; fax: þ353 1 8963848. E-mail addresses:
[email protected] (F.C. Mansergh), reaz.vawda@fightingblindness. ie (R. Vawda),
[email protected] (S. Millington-Ward),
[email protected] (P.F. Kenna),
[email protected] (J. Haas),
[email protected] (J.H. Gallagher), jwilson@ bcm.tmc.edu (J.H. Wilson),
[email protected] (P. Humphries),
[email protected] (M. Ader),
[email protected] (G.J. Farrar). 0014-4835/$ e see front matter ! 2010 Elsevier Ltd. All rights reserved. doi:10.1016/j.exer.2010.07.003
(AMD) affects as many as 1 in 10 over-65s. Gene therapy-based approaches have shown therapeutic promise in clinical trials (Maguire et al., 2008, 2009; Bainbridge et al., 2008; Cideciyan et al., 2008), but require some surviving cells in order to work. Cell therapy for advanced disease may provide a complimentary approach. A variety of stem cell sources have been identified, including ciliary epithelial cells (CE), retinal progenitor cells (RPCs, derived from embryonic or early postnatal neural retinas), embryonic stem (ES) cells, and induced pluripotent stem (iPS) cells. Ciliary epithelial cells (CE) are derived from the ciliary margin and can generate spheres in culture which upregulate neuro-retinal genes (Tropepe et al., 2000; Coles et al., 2004; Das et al., 2005), However, recent reports note that these cells fail to form bona fide retinal neurons and glia (Cicero et al., 2009; Gualdoni et al., 2010); this source of stem cells has therefore not been investigated further here.
F.C. Mansergh et al. / Experimental Eye Research 91 (2010) 500e512
Transplantation studies have shown that a sub-fraction of freshly dissociated cells from PN2e6 mouse retinas can integrate into adult host retina, show morphology characteristic of photoreceptors and can ameliorate symptoms of RD (MacLaren et al., 2006; Bartsch et al., 2008; West et al., 2008, 2009). Similar transplants using cultured cells derived from embryonic to postnatal day 6 retinas (RPCs) (or, indeed, freshly dissociated primary retinal cells from embryonic retinas or retinas older than the first postnatal week) do not result in photoreceptor morphology. Optimal integration coincides with the birth of rod photoreceptors. Integrated cells are thought to be post-mitotic (MacLaren et al., 2006; West et al., 2009). Moreover the frequency of integration is low even at PN4 (0.6%), and transplantation is less successful in diseased retinas, perhaps because of the gliosis that accompanies RD (West et al., 2009). Disruption of the outer limiting membrane (OLM; a barrier between the subretinal space and the outer nuclear layer) can increase the proportion of integrating cells; however, percentages are still small (West et al., 2008, 2009). Primary retinal cells dissociated from late embryonic (E14.5 and subsequent) and early postnatal mouse retinas can be expanded in tissue culture. Adherent cultures are established within a few days, can be passaged after 1 month, and can be grown indefinitely (Klassen et al., 2004; Qiu et al., 2004; Angénieux et al., 2006; MerhiSoussi et al., 2006). These cells are described here as retinal progenitor cells (RPCs). RPCs are typically expanded in serum-free media optimized for neural cultures with N2 supplement (or similar), epidermal growth factor (EGF) and fibroblast growth factor 2 (FGF2), which have both been shown to have a proliferative effect on these cells (Kelley et al., 1995; Chacko et al., 2000; Das et al., 2005). Moreover, there are indications that EGF, which biases towards glial cell fate in neural stem cell cultures, can act as a potent neuralizing factor in retinal cells (Angénieux et al., 2006). Plating on laminin, sometimes with poly-L-ornithine, and sequential withdrawal of EGF, then FGF2 five days later, is used to differentiate RPCs in vitro, followed by addition of B27 supplement. Adherent RPCs, classified variously as retinal stem cells, retinal progenitor cells, retinal precursor cells, radial glial cells and/or proliferating Mueller glia, show a capacity to generate retinal neurons, including those expressing photoreceptor markers (Klassen et al., 2004; Merhi-Soussi et al., 2006; Canola et al., 2007). Cultured RPCs transplanted either subretinally or intravitreally can integrate into the retina and have been shown to generate some level of therapeutic effect (Klassen et al., 2004). However, reports of photoreceptor morphology arising from transplantation of postnatal rodent RPC cultures are discordant (Klassen et al., 2004; Reh, 2006; Canola et al., 2007; Lamba et al., 2008; West et al., 2009), with the majority now suggesting that such events are rare or non-existent (West et al., 2009). Retinal differentiation protocols have been developed for RPCs, embryonic neural stem cells, ES and iPS cells (Zhao et al., 2002; Merhi-Soussi et al., 2006; Aoki et al., 2008; Meyer et al., 2006, 2009; Lamba et al., 2006, 2010; Ikeda et al., 2005; Osakada et al., 2008; Jagatha et al., 2009; Hirami et al., 2009). ES or iPS-derived cells also give relatively low integration rates of 0.1e0.5% of cells initially transplanted, although some photoreceptor morphology is achieved (Osakada et al., 2008; Lamba et al., 2009, 2010). The ability to generate pure expandable cultures from which larger numbers of photoreceptors can be obtained is a pre-requisite for RD cellular therapies. Firstly, we have investigated the identity of the primary retinal cells that can integrate and give rise to photoreceptors. We have found that PN3e5 Rho-eGFP primary retinal cells already expressing rhodopsin at PN3e5 are more likely to integrate into the outer nuclear layer (ONL) and form morphologically mature photoreceptors after transplantation than those Rho-eGFP cells not expressing rhodopsin at the point of transplantation. Cells expressing rhodopsin
501
at PN3e5 are almost certainly post-mitotic, as rhodopsin is a product of terminal rod differentiation. Rhodopsin expression is therefore a good marker for photoreceptor potential post-transplantation; however, this is rapidly lost from RPC cultures. We hypothesized that sub-optimal initial culture conditions may result in poor survival of photoreceptor precursors. However, we have achieved a 25-fold improvement in growth rate via extensive protocol changes, but no increases were observed in rhodopsin expression levels or integration rates post-transplantation, regardless of whether proliferating or differentiated RPCs were analysed. Expression patterns of marker genes in proliferating RPCs show that they are undifferentiated; hence, we are losing cells that have already chosen a photoreceptor cell fate. Rapid loss of rhodopsin expression after introduction of the cells to tissue culture indicates that the majority of post-mitotic cells are dying within 3 days. Identifying cues by which retinal progenitors are specified in vivo, and culture conditions that promote survival in vitro after specification, but prior to injection, will be necessary for therapeutic use of these cells in RD. However, the improvements in isolation and growth rates described here will be useful to anyone who might wish to investigate the therapeutic potential of these cells for disorders such as glaucoma, Leber’s hereditary optic neuropathy, or retinoschisis, where the defect does not lie within the photoreceptor layer. 2. Materials and methods 2.1. Retinal dissociation and FACS analysis For FACS experiments, we used postnatal PN3e5 rhodopsin-eGFP (Rho-eGFP; Chan et al., 2004) heterozygote mice as donors for FACS and subsequent transplantation. These mice express a human RhoeGFP fusion protein that is visible in rod outer segments following transplantation. Heterozygotes were used in our experiments as homozygotes show symptoms of retinal degeneration (Chan et al., 2004). Retinas were dissected and placed in 1 ml HBSS (Lonza). The ciliary margin was removed from all retinas prior to dissociation. Retinal cells were analysed by FACS as previously described (Palfi et al., 2006). Following FACS analysis, cells were spun at 2000 rpm for 5 min and resuspended such that the approximate concentration of cells was 200,000 per 3 ml (cell count obtained from FACS). Following subretinal injection (see below), residual cell samples were counted using a haemocytometer in order to assess the actual number and viability of cells injected (given the time elapsed, this could vary substantially from FACS figures). For each time point, FACS sorting was carried out 3 times and for each repetition, at least 3 eyes were injected with positive and 3 eyes with negative cells. Unsorted cells were also transplanted as a control. Animals were sacrificed 3 months post-transplantation, eyes were sectioned as described below. Given the fact that eGFP, in Rho-eGFP cells, is expressed as a rhodopsin-eGFP fusion protein, positive cells were identified via eGFP positive, morphologically correct outer segments adjacent to the RPE. 2.2. Animals, transplantation, cryosectioning, eGFP transplantation cell counts For transplantation studies involving cultured RPCs, cells were isolated at PN3e5 from transgenic mice ubiquitously expressing eGFP (Okabe et al., 1997). Recipients for all transplantations were C57Bl6/J mice between 2 and 6 months of age. For tissue culture studies, PN3e5 Rho-eGFP, eGFP, C57Bl6/J and Rho!/! donor mice (Humphries et al., 1997) were used. Subretinal injections were carried out in strict compliance with EU and Irish law (Cruelty to Animals Act 2002) and with the ARVO statement for animal use in ophthalmic research. Anaesthesia and subretinal injections were
502
F.C. Mansergh et al. / Experimental Eye Research 91 (2010) 500e512
carried out as previously described (Chadderton et al., 2009). Fixation, cryosectioning and staining were carried out as previously described (Kiang et al., 2005). All sections were cut at 12.5 mm thickness and were DAPI stained. Cells were counted on a fluorescent microscope (Zeiss, Axioplan 2); channels specific for RFP were checked for each positive cell to ensure omission of false positives due to autofluorescence. eGFP-positive cells were counted in different categories, depending on retinal location and morphology. Cells were counted as either unincorporated (balls of unintegrated cells adjacent to the injection site, for example), retinal (having migrated into the retina but with no evidence of integration), integrated (visible evidence of axons, dendrites etc), or possessing photoreceptor morphology (see above). Photoreceptor morphology was not seen following transplantation of cultured RPCs, only after transplantation of fresh retinal cells. In every instance, the entire eye was sectioned and all 12.5 mm sections obtained were mounted, DAPI stained and counted. Following sectioning and counting, numbers of correctly integrated cells were divided by the original number injected (haemocytometer figures used for FACS cells) and multiplied by 100 to give the percentage of integrated cells, photoreceptors etc. Results were graphed using GraphPad Prism 3.0.
2.4.3. Substrates Collagen, fibronectin, laminin, poly-L-ornithine, poly-L-lysine, poly-D-lysine and vitronectin were obtained from Sigma, while gelatin was obtained from Millipore. All were applied at 0.1% for 2 h at room temperature or overnight at 4 # C. Flasks were rinsed in 1$ PBS (Lonza) twice before addition of media and cells. 2.4.4. Dissociation Cells were placed in T25 flasks at a noted cell density in 5 ml growth medium and incubated at 37 # C and 5% CO2. After initial plating, the medium was replaced completely after 5e7 days initially, then every 2e7 days depending on density. 2.4.5. Passaging Cells were passaged when 80e90% confluent. 0.5 ml 0.25% trypsin/EDTA (Lonza) or Accutase (Sigma) were added to each T25 flask, following removal of medium and rinsing with 1$ PBS. Flasks were incubated for 5e10 min until cell monolayers lifted off. Cells were resuspended in 3 ml DMEM/F12, counted four times and spun at 1000 rpm. The supernatant was removed, pellets were resuspended in growth medium and replated, frozen, or treated with TRI reagent (Sigma).
2.3. Retinal dissociation for subsequent tissue culture Retinas were placed in 1 ml PBS or HBSS following dissection. The ciliary margin was removed from all retinas prior to dissociation, in order to avoid contamination by CE cells in subsequent cultures. Old dissociation method: Retinas were placed in 1 ml HBSS (Lonza) and 100 ml 10 mg/ml trypsin (Sigma) was added. Retinas were incubated for 10 min at 37 # C, after 5 min, 10 ml 10 mg/ml DNase1 þ100 ml 20 mg/ml trypsin inhibitor were added and samples were triturated with a P1000 pipette (Gilson). Cells were spun for 5 min at 2000 rpm (Thermo Microlite microcentrifuge) and cells were resuspended in 1 ml growth medium. New methods: 100 ml 0.25% trypsin/EDTA (Lonza/Biowhittaker) or 100 ml Accutase (Sigma) were added, retinas were incubated for 5 min at 37 # C, allowed to settle to the bottom of the tube and most of the supernatant was aspirated away. 1 ml growth medium was then added and retinas were dissociated by trituration with a fire polished Pasteur pipette. Retinas prepared via mechanical dissociation followed the same procedure, but omitting enzymatic digestion. Following dissociation, cells were counted four times using a haemocytometer. 2.4. Tissue culture 2.4.1. Media, supplements, growth factors Neurobasal medium, B27 and B27 without Vitamin A (B27 ! RA) were obtained from Invitrogen; DMEM/F12, embryonic and postnatal stem cell media were from Sigma. Growth medium for RPCs was composed of DMEM/F12 supplemented with 1$ N2, 1$ B27, 1$ B27 ! RA or a combination thereof. 1$ L-glutamine (Lonza), 1$ penicillin/streptomycin (Lonza), 5 mg/ml heparin (Sigma), 20 ng/ml fibroblast growth factor 2 (FGF2) and 20 ng/ml epidermal growth factor (EGF) were also added. RPCs were grown in Sarstedt T25 flasks; other plastics were less conducive to growth. Neurobasal and other media were supplemented as for DMEM/F12. 2.4.2. Differentiation Cells were differentiated in vitro by plating cells on laminin coated T25s or poly-L-lysine and laminin coated glass cover slips. After 2e4 days, EGF was withdrawn from the medium for 5 days, followed by use of final growth medium supplemented with B27 þ RA and no growth factors for 5e7 days. Glial differentiation can be enhanced by adding 1% fetal calf serum (FCS) to the final growth medium.
2.4.6. Calculations Initial cell density and date of plating were noted for each flask. The number of days to reach 80e90% confluence was also noted. Cells were counted at each passage (p) and the rate of cell growth was calculated as follows:
Increase in cell no: per day ¼
ððCell no: at pN þ 1Þ ! ðCell no: at pNÞÞ No: of days between pN þ 1 and pN
2.4.7. Freezing Cells were resuspended in 1 ml freeze medium (7.5% glucose, 10%BSA or 1% B27 ! RA, 10% DMSO, made up in DMEM/F12 medium (Sigma)), placed in a Mr Frosty and frozen at !70 # C. Thawed cells were placed in 5 ml DMEM/F12 and spun before replating. 2.5. RNA extraction Cell pellets were resuspended in 1 ml TRI reagent (Sigma) and triturated using a P1000, while retinas were homogenized in 1 ml TRI reagent using a Dounce homogenizer (Fisher). RNA was prepared according to the manufacturer’s protocol. RNA samples were assayed for concentration and quality using a Nanodrop ND1000 (NanoDrop Technologies) spectrophotometer. 2.6. RT-PCR Reverse transcription was carried out as previously described (Mansergh et al., 2009). Primers were obtained from Sigma-Genosys (see Table 1). PCRs were carried out using Crimson Taq and buffer (NEB) according to the manufacturer’s instructions. A “no RT” control corresponding to each sample was included. Housekeeping genes (beta-actin, Gapdh, 18S rRNA) were also tested by DNA based Q-PCR to ensure that CT values for each gene were within a similar range (
