Reprogramming of the chick retinal pigmented epithelium after retinal injury
© Luz-Madrigal et al.; licensee BioMed Central Ltd. 2014
Received: 17 December 2013
Accepted: 31 March 2014
Published: 17 April 2014
Skip to main content
© Luz-Madrigal et al.; licensee BioMed Central Ltd. 2014
Received: 17 December 2013
Accepted: 31 March 2014
Published: 17 April 2014
One of the promises in regenerative medicine is to regenerate or replace damaged tissues. The embryonic chick can regenerate its retina by transdifferentiation of the retinal pigmented epithelium (RPE) and by activation of stem/progenitor cells present in the ciliary margin. These two ways of regeneration occur concomitantly when an external source of fibroblast growth factor 2 (FGF2) is present after injury (retinectomy). During the process of transdifferentiation, the RPE loses its pigmentation and is reprogrammed to become neuroepithelium, which differentiates to reconstitute the different cell types of the neural retina. Somatic mammalian cells can be reprogrammed to become induced pluripotent stem cells by ectopic expression of pluripotency-inducing factors such as Oct4, Sox2, Klf4, c-Myc and in some cases Nanog and Lin-28. However, there is limited information concerning the expression of these factors during natural regenerative processes. Organisms that are able to regenerate their organs could share similar mechanisms and factors with the reprogramming process of somatic cells. Herein, we investigate the expression of pluripotency-inducing factors in the RPE after retinectomy (injury) and during transdifferentiation in the presence of FGF2.
We present evidence that upon injury, the quiescent (p27Kip1+/BrdU-) RPE cells transiently dedifferentiate and express sox2, c-myc and klf4 along with eye field transcriptional factors and display a differential up-regulation of alternative splice variants of pax6. However, this transient process of dedifferentiation is not sustained unless FGF2 is present. We have identified lin-28 as a downstream target of FGF2 during the process of retina regeneration. Moreover, we show that overexpression of lin-28 after retinectomy was sufficient to induce transdifferentiation of the RPE in the absence of FGF2.
These findings delineate in detail the molecular changes that take place in the RPE during the process of transdifferentiation in the embryonic chick, and specifically identify Lin-28 as an important factor in this process. We propose a novel model in which injury signals initiate RPE dedifferentiation, while FGF2 up-regulates Lin-28, allowing for RPE transdifferentiation to proceed.
Several vertebrate species have the capacity to transdifferentiate the retinal pigmented epithelium (RPE) to retina (for reviews, see [1, 2]). In the chick, the process of RPE transdifferentiation was first described based on histological observations . We previously demonstrated that after retina removal from chick eyes at embryonic day (E) 4 to 4.5 (Stage 24 to 25, according to Hamburger and Hamilton Stages ) and in the presence of fibroblast growth factor 2 (FGF2), the RPE loses its pigmentation and transdifferentiates to become a neuroepithelium, co-expressing the retinal progenitor markers Pax6 and Chx10 through FGF/FGF-Receptor (FGFR)/Mitogen-activated protein kinase (MAPK) and extracellular-signal-regulated kinase (ERK) signaling cascade [5, 6]. Concomitantly with RPE transdifferentiation, the transcriptional factor Mitf, an RPE-specific marker, is down-regulated, suggesting a change in cell fate of the injured RPE. The ectopic expression of Mitf is sufficient to inhibit RPE transdifferentiation, likely inhibiting the up-regulation of pax6 expression . During retina regeneration from the RPE, the newly generated neuroepithelium eventually differentiates into all major cell types found in the retina, and the differentiation pattern follows the same order as it does during normal development . The ability of RPE cells to transdifferentiate ceases as embryonic development proceeds beyond E4.5 . However, the ectopic expression of pax6 is sufficient to induce RPE transdifferentiation in the intact developing chick eye up to E14 (Stage 35) . In chick RPE cultures, overexpression of different pro-neural transcriptional factors such as sox2, ash1, ath5, neuroD, neurogenin1, neurogenin3, cath5 and msx2 can promote the transdifferentiation of the RPE into neuronal cells (reviewed in ). By contrast, there are several factors associated with RPE specification, including Mitf, Otx2, Wnt13, BMP, Shh and Activin [6, 8–15]. The inactivation of Wnt/beta-catenin signaling in the embryonic mouse RPE results in down-regulation of RPE-specific factors Mitf and Otx2 and expression of neural retina markers Chx10 and Rx [9, 10].
Recently, it has been demonstrated that somatic mammalian cells can be reprogrammed to become induced pluripotent stem cells by ectopic expression of pluripotency-inducing factors Oct4, Sox2, c-Myc and Klf4 as well as by the combination of Oct4, Sox2, Nanog and the RNA-binding protein Lin-28 [16, 17]. Among all these transcriptional factors, Oct4 (Pou5f1), Nanog and Sox2 are key factors that maintain embryonic stem cell identity . More recently, efficient differentiation of induced pluripotent stem cells into neural retina cells has been demonstrated, suggesting the possibility of using these cells for clinical therapies . Other studies have used a specific set of factors to convert fibroblasts directly into induced neural cells . For example, mouse fibroblasts can be directly converted into induced neural cells by overexpressing Ascl1, Brn2 and Myt1l. However, these induced cells lack the potential to generate diverse neural subtypes . In another work, transient expression of Oct4, Sox2, c-Myc and Klf4 was sufficient to induce transdifferentiation of mouse fibroblasts to neural stem/progenitors cells  that can be expanded and differentiate in multiple neuronal subtypes and glial cells. Although all these methods of reprogramming yield cells with similar characteristics to the target cells, it is still unknown if these reprogrammed cells are able to recapitulate the natural process of differentiation or whether the induced pluripotent stem cells or induced neural cells retain the epigenetic memory of their origin. Importantly, aberrant expression of pluripotency genes, incomplete demethylation of specific promoters, viral integration and, more prominently, cancer [23–26] have been reported as a result of reprogramming. Moreover, from the medical point of view, the possibility to integrate these cells into somatic tissue remains unclear.
As an alternative, the study of transdifferentiation and regeneration could provide important information regarding maintenance of pluripotency, dedifferentiation processes, factors involved in cell reprogramming and integration of the cells in the regenerated tissue . Initial studies have shown that among the pluripotency-inducing factors, sox2, c-myc and klf4 are the common factors expressed during lens and limb regeneration in newts and during fin regeneration and Müller glia dedifferentiation in zebrafish [28–30]. More recently, it was demonstrated that in mammals Lin-28 can enhance tissue repair in several contexts including improved hair regrowth and accelerated regrowth of cartilage, bone and mesenchyme after ear and digit injuries .
Lin-28 is an important regulator of let-7 miRNAs, and it has a functional role in organismal growth and metabolism, tissue development, somatic reprogramming and cancer (reviewed in ). During in vitro differentiation of mouse embryonic carcinoma cells to neural and glial fates, Lin-28 can alter the cell fate independently of let-7; in addition, overexpression of Lin-28 increases neurogenesis in the same cell types . In vitro and in vivo experiments have demonstrated that Lin-28 regulates the translation and stability of a large number of mRNAs including cell cycle regulators, splicing factors, metabolic enzymes and RNA-binding proteins [31, 34–38]. All this evidence strongly suggests that Lin-28 can have a pivotal role in tissue regeneration.
Consistent with this idea, we analyzed the expression of pluripotency-inducing factors, including Lin-28, during RPE transdifferentiation, using the embryonic chick model of retina regeneration. Among all the factors, sox2, c-myc and klf4 were transiently up-regulated in the injured RPE (after retinectomy) along with eye field transcriptional factors, achaete-scute complex homolog 1 (ascl1 also known as chicken achaete-scute homolog (CASH-1)) and differential up-regulation of alterative splice variants of pax6. By contrast, lin-28 was significantly up-regulated only in the presence of FGF2 in retinectomized eyes. Moreover, Lin-28 overexpression in the injured RPE was sufficient to induce RPE transdifferentiation.
These results establish a two-step dedifferentiation process. First, upon injury there is an activation of gene expression for sox2, c-myc and klf4, concomitantly with the up-regulation of eye field transcriptional factors. Second, in the presence of FGF2, lin-28 is up-regulated, suggesting a correlation between the expression of lin-28 and the process of transdifferentiation. Overexpression of Lin-28 in the injured RPE was sufficient to induce RPE transdifferentiation in the absence of an external source of growth factors. To our knowledge, these data provide the first evidence that Lin-28 plays an important role in retina regeneration via RPE transdifferentiation. This is also the first detailed study on the molecular profile of the RPE during the process of dedifferentiation.
Among these factors, it has been reported that Klf4 plays a critical role in neurogenesis and neural migration during cerebral cortex development in mouse . It could be possible that Klf4 has a similar role in retina development. In chick embryos, klf4 mRNA was detected in the neural folds and in the neural tube at Stages 8 to 9. Later, at Stage 27, klf4 was detected in the face and neck region . In our experiments, we detected Klf4 ubiquitously in the ciliary margin, neuroepithelium and RPE (Figure 1A,C,G and Additional file 1: Figure S1C).
In chicken embryos, Lin-28 is ubiquitously expressed in the presumptive limb primordium at Stages 15 to 16, and in other tissues including the neuroepithelium of the optic cup and in the otic vesicle at E2.5 to 3 (Stages 17 to 18) . In mouse, Lin-28 is present in the retina at embryonic days E8.5 to E17.5 [42, 43]. We detected Lin-28 in the chick ciliary margin, RPE and neuroepithelium at E4 (Stage 24) (Figure 1A,E,I) and at E7 (Additional file 1: Figure S1D), suggesting that the presence of this protein could be necessary for growth and differentiation of these tissues.
Among all the factors controlling the regulatory network in embryonic stem cells, Oct4 and Nanog are considered the ‘key partner core’ of transcriptional regulators (for review see [44, 45]). Expression of the avian homologs of oct4 (cPouV) and nanog was demonstrated early in the developing chick at Stages 8 to 9 . We were unable to detect mRNAs of oct4 and nanog in the ciliary margin or RPE at the embryonic days analyzed, but we did confirm their expression in chick embryos at Stages 8 to 9 (Figure 1A). Although we found the expression of sox2, c-myc and klf4 in the ciliary margin, the absence of oct4 and nanog points out that these cells are not pluripotent but may retain some properties of stem cells. Despite the fact that Sox2, c-Myc, Klf4 and Lin-28 are considered pluripotency-inducing factors when used for reprograming in combination with Oct4 and Nanog, these factors have other functions during eye and retina development [39, 47–50].
It is known from several organisms, that transdifferentiation occurs by the following steps: transient dedifferentiation, proliferation and differentiation into the new linage [1, 51]. However, the time of dedifferentiation and proliferation is highly dependent on the type of damage and model of regeneration. For example, in zebrafish retina, different damage paradigms including light lesions (acute and chronic), chemical treatments that kill retina neurons (ouabain) and mechanical insults to the retina (needle stabbing) ultimately result in regeneration of the lost neurons by Müller glia transdifferentiation; however, the time at which Müller glia dedifferentiate or enter the cell cycle varies between the treatments. Dedifferentiation events have been reported as early as 4 h for the acute light lesion model, about 15 h for the stabbing model and up to 31 h for chronic light lesion cases. Signs of cells entering the cell cycle have been observed 24 to 72 h post lesion (for reviews see [1, 52, 53]. Interestingly, retina damage caused by different concentrations of ouabain generates extensive cell death of retina neurons and promotes the surviving Müller glia to proliferate robustly within a 3- to 12-day period . During lens regeneration in the newt, the pigmented epithelial cells from the dorsal iris that are responsible for replacing the missing lens express nucleostemin (a stem cell marker) 2 days after lentectomy. This is followed by the loss of pigmentation and cell identity, facilitating the subsequent proliferation that takes place 4 days later . Notably, inhibiting the cell cycle using a Cdk2 inhibitor does not block the process of dedifferentiation .
To better understand and characterize the process of RPE transdifferentiation, we analyzed the proliferative status of the RPE. We performed complete retinectomies in E4 (Stage 24) chick eyes in the presence or absence of FGF2, and the embryos were collected at 6 and 24 hours post-retinectomy (h PR) to examine 5-bromo-2′-deoxyuridine (BrdU) incorporation and the cyclin-dependent kinase inhibitor 1B (p27Kip1) for cell cycle arrest. At 6 h PR, in the absence (retinectomy only) or presence of FGF2, we did not observe BrdU-positive cells in the posterior RPE (Additional file 2: Figure S2A,E). By contrast, a large proportion of p27Kip1-positive cells were found in the posterior RPE at 6 h PR regardless of the presence of FGF2 (Additional file 2: Figure S2B,F), suggesting that at this time point the cells were still arrested in the cell cycle and as a consequence proliferation was blocked. However, BrdU-positive cells were detected in the RPE at 24 h PR only in the presence of FGF2 (Additional file 2: Figure S2G), when the RPE became p27Kip1-negative (Additional file 2: Figure S2H), suggesting that RPE cells had entered the cell cycle. We did not observed BrdU-positive cells in the RPE at 24 h PR in the absence of FGF2 (Additional file 2: Figure S2C) when the RPE was still p27Kip1-positive (Additional file 2: Figure S2D); therefore, FGF2 is necessary for the cell cycle entry, and eventually for RPE transdifferentiation .
The process of RPE dedifferentiation was evidenced by the down-regulation of RPE specification genes mitf and tyr (tyrosinase, a melanin-catalyzing enzyme) concomitantly with an up-regulation of neural retina progenitors ascl1 (also known as CASH-1, ash1) and chx10 (vsx2) (Figure 2B,C).
We also decided to analyze if the dedifferentiated RPE cells go back into the lineage of eye formation. Different factors are crucial for eye formation; the most important are the eye field transcriptional factors that are expressed in the anterior neural plate in the region specified to become the eyes. These eye field transcriptional factors include et, rx1(rx), six3, pax6, lhx2, six6 (optx2) and tll[59–61]. The up-regulation of rx1, six6, lhx2 and six3 (Figure 2B) suggests that the injury was enough to induce a transient dedifferentiation of the RPE and promote these cells to go back to the presumptive optic vesicle stage.
Despite the partial dedifferentiation of the RPE cells, by 72 h in the absence of FGF2 the RPE again acquired its pigmentation (not shown) and mitf expression was recovered at higher levels compared with the uninjured eye (P = 0.016, n = 3) (Figure 2C). Similar to what has been observed in Müller glia transdifferentiation in zebrafish, we observed significant up-regulation of ascl1 (CASH-1), a proneural basic helix-loop-helix transcriptional factor. Importantly, ascl1a, the homolog to chicken ascl1, has been used to reprogram fibroblasts to neurons [21, 62]. We also observed significant up-regulation of rx1, which is related to eye field specification [61, 63] (Figure 2B). The importance of Rx has been demonstrated during retina regeneration in pre-metamorphic Xenopus laevis.
During mouse brain development, Pax6 (5a+) affects cell proliferation but not neural differentiation. By contrast, the canonical Pax6 (5a-) affects cell proliferation and differentiation [76, 77]. It is possible that during the process of RPE transdifferentiation the alternative splice variants of pax6 observed here have different functions or different gene targets. Collectively, these results suggest that upon retina removal at E4 (Stage 24), the quiescent cells of the RPE dedifferentiate to become progenitor-like cells similar to the cells of the optic vesicle that express the combination of eye field transcriptional factors and the factors sox2, c-myc and klf4. This process is transient and not sustained if no growth factors are present.
In agreement with our results, destruction of photoreceptors by acute light lesions in zebrafish central retina results in Müller glia dedifferentiation 4 to 8 h post-lesion, exemplified by a strong Rx1 immunoreactivity . We found a significant up-regulation of rx1 transcript at 6 h post-injury (h PR) during the transient dedifferentiation of the RPE (Figure 2B). Moreover, at 24 h post-lesion, a percentage of the activated zebrafish Müller glia cells re-enter the cell cycle, just as the chick RPE does in the presence of FGF2 (Additional file 2: Figure S2G,H). Interestingly, Müller glia cells respond by dedifferentiating after local loss of contact with photoreceptors; however, this dedifferentiation is not enough to promote re-entry to the cell cycle. Similarly, it is possible that the loss of contact between the RPE and the retina (after retinectomy) triggers the process of RPE dedifferentiation, or that the process involves an inflammation molecule such as TNFα, as in zebrafish , or even anaphylatoxins such as C3a or interleukins like IL-6, which have been shown to play a role in chick retina regeneration . Taken together, these data suggest that similar mechanism of dedifferentiation and proliferation could be shared between retina regeneration from the RPE and from Müller glia cells.
Fertilized Specific Pathogen-Free chicken eggs (Charles River Laboratories, Wilmington, MA, USA) were incubated in a humidified rotating incubator at 38°C. At E4 (Stage 24 ), retinectomies and FGF2 treatments were performed as previously described . Embryos were collected at 6, 24 and 72 h PR and processed for laser capture microdissection, histology and immunofluorescence. For proliferation studies, 10 μl of BrdU (10 μg/ml) solution was dropped over the eye of the embryo 1 h before collection.
Laser capture microdissection was performed as previously described . Briefly, embryos were collected and infiltrated at 4°C with 6.25%, 12.5% and 25% sucrose for 10, 20 or 30 min, respectively, followed by 2:1 25% sucrose to optimal cutting temperature (OCT) compound (Sakura Finetek, Torrance, CA, USA) for 1 h and frozen in dry ice and methylbutane. Cryosections (12 μm) were collected onto metal-framed polyethylene naphthalate membrane slides (Arcturus, Applied Biosystems, Foster City, CA, USA), fixed in 70% ethanol at -20°C for 30 s, rinsed in cold diethylpyrocarbonate-treated water, stained with hematoxylin (Sigma, St. Louis, MO, USA) for 10 s, and dehydrated in ethanol for 30 s each in 70%, 95% and finally 2 min in 100% ethanol. Laser capture microdissection was performed using a Veritas laser capture microdissection system and software as described previously . Laser microdissected sections were collected in CapSure HS LCM Caps (Molecular Devices, Foster City, CA, USA), and total RNA extraction was performed using PicoPure RNA Isolation Kit (Arcturus, Applied Biosystems) including a treatment with DNase I (RNase-Free DNase Set; Qiagen, Darmstadt, Germany). The quality and quantity of RNA were determined using an Agilent RNA 6000 Pico LabChip (Agilent 2100 Bioanalyzer; Agilent Technologies, Santa Clara, CA, USA). Five nanograms of total RNA with an RNA integrity number > 8 were amplified using Ovation Pico WTA System V2 (NuGEN, San Carlo, CA, USA) according to the manufacturer’s instructions, to generate the Single Primer Isothermal Amplification (SPIA) cDNA. Finally, SPIA cDNA was purified using QIAquick PCR Purification Kit (Qiagen) and quantified using a Nanodrop ND-100 spectrophotometer (Thermo Fisher Scientific, Kalamazoo, MI, USA).
Total RNA was extracted from Stage 8 whole embryos using NucleoSpin RNA II isolation Kit (Macherey-Nagel, Düren, Germany) following the manufacture’s protocol. The quality and quantity of RNA were determined using Agilent RNA Nano LabChip (Agilent 2100 Bioanalyzer; Agilent Technologies). Approximately 300 ng of total RNA with a an RNA integrity number > 8 were used for cDNA synthesis using ImProm-II Reverse Transcription System (Promega, Madison, WI, USA) and random-primer hexamers according to the manufacturer’s instructions. For CM and RPE, the amplified SPIA cDNA (50 ng) was used as a template in the PCR reactions. All PCR reactions were performed using PlatinumTaq DNA polymerase (Invitrogen, Grand Island, NY, USA) and the intron-spanning primers are listed in Additional file 3: Table S1. Amplification conditions included denaturation at 95°C for 1 min, annealing at 55°C for 2 min, and extension at 72°C for 30 s. All the RT-PCR reactions were run from at least three independent biological samples and the fragments were gel-purified and sequenced to confirm the specificity of the sequence.
RT-qPCR was performed using a 20 μl mixture containing 5 μl (4 ng) SPIA cDNA, 10 μl 2× SYBR Green/Fluorescein qPCR Master Mix (SABiosciences, Qiagen, MD, USA), and a 500 nM final concentration of the primers. Splice junction-specific primers were designed using Primer 3 (v 4.0) available in  and were optimized following guidelines for RT-qPCR experiments (amplification efficiency and melting curves) . Primers sequences and Ensembl or GenBank identification numbers are provided in Additional file 3: Table S2. Amplifications reactions were performed in triplicate using an iCycler (BioRad, Hercules, CA, USA). The cycling conditions included 10 min polymerase activation at 95°C and 35 cycles of 15 s at 95°C and 1 min at 60°C, followed by a dissociation run from 65°C to 95°C for melting curve analysis. The comparative cycle at threshold (2-ΔΔCt) and an unpaired Student’s t-test analysis were used to determine relative changes in transcript levels compared to gapdh mRNA levels as previously reported  using SigmaPlot 8.0 Software. All analyses were performed in triplicate with at least three independent biological samples.
Antibodies against Sox-2 (1:100), c-Myc (1:50) and Lin-28 (1:100) were purchased from Santa Cruz Biotechnology. Antibody against Klf4 (1:100) was purchased from Aviva Systems Biology (San Diego, CA, USA). Antibodies against Mitf (1:500), GFP (1:50) and BrdU (1:50) were purchased from Abcam (Cambridge, MA, USA). Antibody against Pax-6 (1:10) was obtained from the Developmental Studies Hybridoma Bank (Iowa City, IA, USA). Anti-Chx-10 (1:250) antibody was purchased from ExAlpha (Shirley, MA, USA). Antibody against p27Kip1 (1:50) was obtained from BD Biosciences (Franklin Lakes, NJ, USA). All secondary antibodies were purchased from Molecular Probes (Grand Island, NY, USA) and used at 1:100 dilution.
Embryos were fixed in 4% paraformaldehyde in PBS for 4 h at room temperature, equilibrated in 30% sucrose, embedded in OCT compound (Sakura Finetek), and sectioned at 12 μm. For the p27Kip1 antibody, tissues were fixed in 10% neutral buffered formalin (Thermo Scientific), embedded in paraffin, sectioned, and deparaffinized followed by 30 min antigen retrieval. Sections were permeabilized with 1% saponin in PBS or 15 min in 2 N HCl in PBS for BrdU immunostaining and blocked with 10% goat or donkey serum and incubated overnight at 4°C with the primary antibody. Sections were incubated in secondary antibody and coverslipped with Vectashield (Vector Laboratories, Burlingame, CA, USA). Confocal images (size 1,024 × 1,024) were collected sequentially on a Zeiss 710 Laser Scanning Confocal System (Jena, Germany) using a × 20/0.80 Numeric Aperture (NA) = 0.55 WD objective lens or EC Plan-Neofluar. Results were confirmed using three different biological samples.
Chicken Lin-28a was cloned from a Stage 14 embryo as previously described  and the cDNA was synthetized as described in the RT-PCR section using the primers described in Additional file 3: Table S3. The PCR product was gel-purified and cloned in pDrive plasmid to generate pDlin28a plasmid (Qiagen). Thereafter, a HindIII/BamHI fragment of 895 base pairs from pDlin28a was cloned using the same restriction sites in pcDNA3.1+ (Invitrogen) to generate pCLin28a. T7 primer was used to confirm the integrity of the Lin-28a sequence. Electroporations were performed 1 h PR by injecting 3 μl of a mixture of pCLin28a and pIRES-GFP (1:1, 3 μg each) or 3 μl of a mixture of pcDNA3.1 and pIRES-GFP as controls at the same concentration. The injections were performed using a Pico-injector system PLI-100 (Harvard Apparatus, Holliston, MA, USA) and glass capillary needles. Thereafter, a gold-plated wire electrode (In Ovo gene model 512; BTX Technologies inc., Bent, Holliston, MA, USA), used as an anode, was placed in the ventral border of the eye and a platinum and iridium electrode (FHC Inc., Bowdoinham, ME, USA), used as cathode, was inserted on the top of the brain. Three square pulses of 15 V of 50 ms length and at 950 ms intervals were applied using an ECM 830 electroporator (BTX Technologies, Inc.). The window on the shell was sealed and the embryos were returned to the incubator and collected 72 h post-electroporation and processed for immunohistochemistry and histology.
Embryos used for histological analysis were fixed in Bouin’s fixative (Ricca Chemical Company, Arlington, TX, USA), embedded in paraffin, sectioned at 12 μm, stained with hematoxylin and eosin, and photographed using an Olympus BX-51 microscope (Tokio, Japan). To quantify the amount of transdifferentiated RPE, we analyzed the area from three histological sections of four different eyes (12 sections). Images were captured using an Olympus camera and Magnafire image capture software and processed using ImageJ software available in . The transdifferentiated area was calculated using the free hand tool and a two-tailed permutation test for comparing means using R version 3.0.
fibroblast growth factor
optimal cutting temperature
polymerase chain reaction
quantitative polymerase chain reaction
single primer isothermal amplification
tumor necrosis factor alpha.
Dr R Edelmann and the Microscopy Facility at Miami University; Dr A Kiss from The Center for Bioinformatics and Functional Genomics (CBFG) at Miami University; M Hughes from the Department of Statistics for his expertise on statistical analyses; C Closson and Dr Y Zavros from the Department of Molecular and Cellular Physiology at the University of Cincinnati for facilitating the use of the laser capture microdissection system; J Dannenfelser for her various technical contributions to this work. This work was supported by EY17319 to KDRT, EY10540 to PAT, and CONACYT 162930 and 142523 to AL-M. The Zeiss confocal microscope is supported by an NSF-MRI award DBI-0821211.
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly credited. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.