Neuroglobin and Cytoglobin: Oxygen-Binding Proteins in Retinal Neurons

Abstract

purpose. The goal of this study was to describe the detailed localization of the novel oxygen-binding molecules, neuroglobin (Ngb) and cytoglobin (Cygb), in mammalian retinas and to determine whether Ngb and Cygb are neuronal or glial proteins in the retina.

methods. Antibodies directed against Ngb and Cygb were used to examine their patterns of distribution in normal canine retinas. Immunoblot analysis was performed to verify antibody specificity and the presence of Ngb and Cygb in canine tissues. Double-labeling immunohistochemistry was performed with the Ngb and Cygb antibodies along with antibodies against neuronal (MAP-2, class III β-tubulin (TUJ1), PKCα, and calretinin) and glial antigens (vimentin and CRALBP). Tissue sections were analyzed with light and confocal microscopy.

results. Ngb and Cygb proteins were observed in different retinal cells. Cygb (but not Ngb) was also present in canine kidney, liver, lung, and heart tissue. Immunohistochemical analysis of canine retinas demonstrated Ngb immunoreactivity (IR) in the ganglion cell layer (GCL), inner (INL) and outer (ONL) nuclear layers, inner (IPL) and outer plexiform (OPL) layers, photoreceptor inner segments (IS), and retinal pigment epithelium (RPE). Ngb IR was localized within retinal neurons, but not in glia. Cygb IR was found in neurons and their processes in the GCL, IPL, INL, and OPL and within the RPE, but not in glia.

conclusions. Ngb and Cygb are widely distributed in retinal neurons and RPE, but not in glial cells of the canine retina. Their structure and distribution is suggestive of a possible role in oxygen transport in the mammalian retina.

Visual processing in the retina has considerable oxygen demands, which makes the retina one of the highest oxygen-consuming tissues in the human body. ^{1 2 3} Continuous supply of sufficient O₂ to the retina is a fundamental physiological need, since even transient O₂ deficits can produce irreversible cellular damage. ⁴ Vascular and ischemic disorders of the retina and optic nerve head are a common cause of visual loss in the middle-aged and elderly population. ⁵ Furthermore, retinal hypoxia is considered to be an important factor contributing to many retinal diseases. ^{6 7 8} However, the mechanisms of oxygen homeostasis in the retina remain poorly understood. 

Globins are a family of heme-containing proteins that reversibly bind oxygen and they have been described in bacteria, fungi, protists, plants, and animals. ⁹ Four mammalian globins have been identified so far (hemoglobin [Hb] , myoglobin, neuroglobin [Ngb], and cytoglobin [Cygb]). ^{10 11 12 13 14} Hb is localized in erythrocytes and has a major role in oxygen transport between the lungs and other tissues via the circulatory system. ⁹ Myoglobin is localized in the cytoplasm of skeletal and cardiac muscle, acts in intracellular oxygen storage, and enhances oxygen diffusion to the mitochondria for use in oxidative phosphorylation. ¹⁵ In addition, Hb and myoglobin can act as scavengers of bioactive nitric oxide. ¹⁶ Ngb and Cygb are two recently described members of the globin family with functions that are still not completely understood. ^{10 12 13}  

Ngb has been identified as a 17-kDa protein distantly related to vertebrate myoglobins (<21% amino-acid identity) and Hbs (<25% amino acid identity). ¹⁰ Ngb expression has been found in human, mouse, rat, chicken, zebrafish, and pufferfish brain, ^{10 17 18 19 20 21 22 23} as well as in human (Grozdanić et al. IOVS 2004;45:ARVO E-Abstract 2586), mouse, chicken, and zebrafish eyes. ^{21 22 24} Some of the proposed functions for Ngb include enhancement of oxygen delivery to mitochondria, ^{22 24 25 26} detoxification of NO, ^{19 27 28 29} and hypoxia sensing. ^{30 31 32} A potentially neuroprotective role of Ngb was suggested from studies demonstrating increased Ngb expression in vivo and in vitro during acute neuronal hypoxia and enhanced survival of cortical neurons by Ngb overexpression in vivo and in vitro. ^{33 34} Moreover, it has been shown in rats that Ngb is prominently expressed in several areas of the brain that show preferential vulnerability to neurodegenerative diseases and that Ngb mRNA and protein levels in these areas decrease with aging. ³⁵ Cygb has been identified as a 20.9-kDa protein in virtually all human, mouse, and zebrafish tissues. ^{12 13} Cygb expression has been observed in the cytoplasm of splanchnic fibroblast-like cells ^{36 37 38} and also in subpopulations of central nervous system (CNS) and retinal neurons. ^{37 39} Because of their heme-based structure and oxygen-binding properties, Ngb and Cygb probably serve as oxygen transport molecules and/or mediators of intracellular signaling during hypoxic conditions. ^{14 32 40 41}  

In this study, Western blot analysis was used to verify the specificity of the anti-Ngb and -Cygb antibodies, and these antibodies were then used to investigate for the presence of Ngb and Cygb in mammalian (canine) tissues. Immunohistochemical procedures were used to describe cell-specific histologic localization of Ngb and Cygb in the canine retina. Both Ngb and Cygb were present in retinal ganglion cells and inner retinal neurons of the canine retina, which are particularly sensitive to ischemic damage. Furthermore, both Ngb and Cygb were present in the retinal pigment epithelium, and Ngb was detected in photoreceptors, which makes these proteins potential candidates for facilitating oxygen metabolism in the outer retina. 

Materials and Methods

Canine Tissue

Eyes, liver, kidney, lung, and heart tissue from nine healthy adult beagles (2–4 years of age) was collected immediately after euthanasia. Before euthanasia, all eyes were examined for signs of ocular abnormalities (slit lamp examination, fundus examination) and the presence of elevated intraocular pressure (tonometry). Eyes with detectable abnormalities of the anterior segment, lens, or the fundus and/or the presence of elevated intraocular pressure (>25 mm Hg) were not collected for use in this study. Globes were fixed for 12 hours at 4°C in 4% paraformaldehyde in 0.1 M phosphate-buffered saline (PBS). After fixation, globes were embedded in paraffin, and sections were cut at 7-μm thickness. All research conducted in this study was in full compliance with the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research and the Iowa State University Committee on Animal Care regulations. 

Antibodies

Full-length human Ngb and Cygb recombinant proteins were synthesized and purified as described previously. ^{11 12} Molecular weights of the recombinant proteins were determined by matrix-assisted desorption ionization–time of flight (MALDI-TOF) mass spectrometry. Polyclonal antisera against synthesized Ngb or Cygb proteins were raised in rabbits and antibodies (Abs) were affinity purified from the serum using the recombinant proteins coupled to a column (SulfoLink; Pierce Biotechnology, Rockford, IL). Mouse monoclonal anti-Ngb antibody and the human recombinant Ngb protein used to produce this antibody were kindly provided by BioVendor Laboratory Medicine, Inc. Primary Abs used in this study and their dilutions are summarized in Supplementary Table S1. 

Immunohistochemistry

Fluorescent immunohistochemistry was performed as a modification of a previously described procedure. ⁴² Briefly, tissue sections were deparaffinized, rehydrated in a graded alcohol series, and incubated for 2 hours in blocking solution. Sections were double-labeled with primary Ab cocktail overnight and incubated in one of the following secondary Ab cocktails: donkey anti-mouse biotinylated Ab (Jackson ImmunoResearch, West Grove, PA) and goat anti-rabbit Alexa 488 Ab (Molecular Probes, Eugene, OR); goat anti-rabbit biotinylated Ab (Vector Laboratories, Burlingame, CA) and goat anti-mouse Alexa 488 Ab (Molecular Probes); goat anti-mouse Cy5 (Jackson ImmunoResearch) and goat anti-rabbit Alexa 488 Ab (Molecular Probes); and goat anti-rabbit Cy5 (Jackson ImmunoResearch) and goat anti-mouse Alexa 488 Ab (Molecular Probes). After a 2-hour incubation, sections were washed in potassium phosphate-buffered saline (KPBS) with Triton X-100. If a biotinylated secondary Ab was used, sections were subsequently incubated with streptavidin Cy3 (Jackson ImmunoResearch) and washed in KPBS. Finally, sections were counterstained with 1 μg/mL of 4′,6-diamino-2-phenylindole (DAPI; Molecular Probes), washed in KPBS and coverslipped. 

For peroxidase immunohistochemistry, endogenous peroxidase activity was blocked by incubation in 0.3% hydrogen peroxide solution in KPBS. Sections were processed for antigen retrieval, incubated in blocking solution, and stored overnight at room temperature in primary Ab solution. Sections were incubated with biotinylated secondary Ab and then with horseradish peroxidase-avidin-biotin complex (Vector Elite ABC Kit; Vector Laboratories) according to the manufacturer’s instructions. To visualize the antibody staining pattern, tissue was exposed to a substrate kit for peroxidase (NovaRed; Vector Laboratories). Sections were dehydrated through a graded ethanol series, cleared with xylene, and coverslipped. 

Negative controls were run in parallel during all processing and included the omission of the primary Ab, secondary Ab or preadsorption of the primary Ab (mouse and rabbit anti-Ngb and rabbit anti-Cygb) with excess recombinant proteins. 

Analysis of Tissue Sections

Canine tissue sections labeled with fluorescent antibodies were visualized and images captured using a confocal scanning laser microscope (TCS-NT; Leica Microsystems Inc., Exton, PA). A color digital camera (Sony DXC-S500; Labtek, Campbell, CA) was used for the bright-field images in Supplementary Figure S2. Sections stained with the red substrate (NovaRed; Vector Laboratories) were examined with an upright microscope (Axioplan 2; Carl Zeiss MicroImaging, Inc, Thornwood, NY), and images were captured with a color camera (AxioCam MRc; Carl Zeiss Meditec, Inc.). All figures were prepared on computer (Photoshop ver. 7.0; Adobe, San Jose, CA, and Freehand ver. 10.0; Macromedia, San Francisco, CA). 

Western Blot Analysis

Characterization of antigens was performed using SDS-polyacrylamide gel electrophoresis (SDS-PAGE) and immunoblot procedures. For SDS-PAGE, canine tissue samples and recombinant proteins were homogenized in SDS-reducing buffer. Approximately 25 μg of total protein from retina, kidney, liver, and heart tissue extracts, and approximately 50 μg of total protein from lung tissue extract were separated in 14% gels using a vertical system (Mini-Protean III; Bio-Rad, Hercules, CA). After electrophoresis, proteins were transferred to a polyvinylidene difluoride (PVDF; BioRad) membrane in transfer buffer and incubated in blocking buffer for 1 hour at room temperature. The membrane was incubated in rabbit polyclonal anti-Cygb Ab (1:1000), rabbit polyclonal anti-Ngb Ab (1:1000), or mouse monoclonal anti-Ngb Ab (1:1000), followed by incubation in alkaline phosphatase conjugated goat anti-rabbit Ab (Promega, Madison, WI) or goat anti-mouse Ab (1:7500), respectively. Immunoreactive (IR) bands were visualized with 5,bromo-4-chloro-3 indolylphosphate (BCIP)/nitroblue tetrazolium (NBT) alkaline phosphatase color development reagents (Promega). Molecular weights were estimated by comparison with prestained molecular weight standards (Bio-Rad). 

For detection of Ngb in canine tissue samples, samples were homogenized in SDS-reducing buffer and approximately 25 μg of total protein in the retina, kidney, liver, and heart tissue extracts and approximately 50 μg of total protein in the lung tissue extract was separated in 14% gels. After electrophoresis, proteins were renatured for 1 hour in SDS gel by incubating the gel in 50 mM Tris buffer (pH 7.4), containing 20% glycerol. ⁴³ Proteins were then transferred to a PVDF membrane in native transfer buffer, incubated in blocking buffer for 1 hour at room temperature and overnight in rabbit polyclonal anti-Ngb Ab (1:1000). Subsequently, an alkaline-phosphatase–conjugated goat anti-rabbit Ab was used at 1:7500 dilution. Chemiluminescence (Lumi-Phos WB Chemiluminescent Substrate; Pierce Biotechnology, Inc., Rockford, IL) was used to visualize immunolabeled bands. 

Results

Distribution of Ngb and Cygb in Canine Tissues

Mouse monoclonal and rabbit polyclonal anti-Ngb Abs detected recombinant Ngb protein at ∼17 kDa (Fig. 1A) under reducing conditions. However, only the polyclonal anti-Ngb Ab was capable of detecting native Ngb in the protein extract from the canine retina (Fig. 1B) . No specific bands were detected in protein extracts from canine kidney, liver, lung, and heart when examined under the same conditions (data not shown). The second band at ∼34 kDa probably represents a stable dimer, as also reported by Schmidt et al. ²⁴ Rabbit polyclonal anti-Cygb Ab detected specific Cygb protein at ∼21 kDa (Fig. 1C) . Anti-Cygb Ab also detected Cygb in protein extracts from canine retina, kidney, liver, lung, and heart at ∼29 kDa. The second bands at approximately double the molecular mass probably represent a dimer, as also reported by Schmidt et al. ^{37 39}  

Retinal Localization of Ngb and Cygb

To examine the expression pattern of Ngb in canine retina, we used the mouse monoclonal and rabbit polyclonal anti-Ngb Abs. Both Abs displayed similar labeling patterns, and were further used in double-labeling studies. As illustrated in Figures 2A and 2B , Ngb IR in the retina was localized to the ganglion cell layer, inner and outer nuclear layers, inner and outer plexiform layers, and photoreceptor inner segments (IS). Cygb expression was determined using a rabbit polyclonal antibody. As shown in Figure 2C , Cygb IR in the retina was localized to the ganglion cell layer, the inner nuclear layer (INL), and inner (IPL) and outer (OPL) plexiform layers. Ngb and Cygb IR was also observed in the retinal pigment epithelium (Figs. 2A 2B 2C) . To verify the fluorescent immunolocalization patterns for Ngb and Cygb within the retina, we also performed peroxidase immunolabeling studies (Figs. 2D 2E 2F)with the same Ngb and Cygb primary antibodies. Both methods revealed similar patterns of labeling. To verify Ab specificity, each Ab was preadsorbed with their respective recombinant protein before incubation on tissue sections, and no specific staining was observed (Supplementary Fig. S1). In addition, negative control studies were performed in parallel by omission of primary or secondary antibodies. No antibody labeling was observed in these control experiments (data not shown). 

To discriminate antibody labeling from autofluorescence in the retinal pigment epithelium (RPE), which can be due to the presence of lipofuscin or oxidized melanin, ^{44 45} confocal as well as bright-field images were taken from nonpigmented RPE in the tapetal retina (Supplementary Fig. S2). Confocal images of tissue sections single-labeled with primary antibody and appropriate Cy-5 conjugated secondary antibody were captured at both 633/645-nm (red channel) and 488/ 525-nm (green channel) excitation/emission wavelengths, respectively. Cy5 is excited at 633 nm and thus, any signal from the 488/525-nm channel is considered background autofluorescence. This comparison would reveal lipofuscin autofluorescence if present (Supplementary Fig. S2). 

To investigate whether retinal neurons express both Ngb and Cygb, we performed double labeling with mouse monoclonal anti-Ngb Ab and rabbit polyclonal anti-Cygb Ab and found that the two proteins are expressed in all IR cells in the ganglion cell layer (GCL) and in the inner nuclear layer (INL). (Figs. 2G 2H 2I) . In addition, extensive localization of both proteins was observed in the IPL and OPL. 

Localization of of Ngb and Cygb in Neurons

Double labeling was performed with anti-Ngb and anti-Cygb antibodies with anti-MAP-2 and TUJ1 antibodies to investigate whether the cells expressing Ngb and Cygb also express these neuronal markers. Cells in the GCL were identified as α, β, and γ ganglion cells or displaced amacrine cells, based on their morphology and relative size using parameters described by others. ⁴⁶ α-cells are polygonal in shape and have the largest somata—21 to 44 μm in diameter. β-cells have a more globular cell body and medium-sized somata—14 to 30 μm, and γ cells have small somata, which makes them hardly distinguishable from displaced amacrine cells. ⁴⁶ Cells in the INL were identified as amacrine, bipolar, or horizontal, based on their laminar position within the INL. All cells in the GCL and INL that expressed the neuronal proteins MAP-2 and class III β-tubulin (TUJ1 IR) were also Ngb- or Cygb IR (Fig. 3) . 

We investigated the possible immunolocalization of Ngb and Cygb in astrocytes and Müller glia with an anti-vimentin and anti-CRALBP antibodies. No double labeling of Ngb or Cygb with vimentin- or CRALBP IR was detected (Supplementary Fig. S3). This finding suggests that Ngb and Cygb are present in neurons, but not in glial cells in the retina. 

Localization of Ngb and Cygb in the Inner Retina

To further investigate expression of Ngb and Cygb within different cell types in the INL, double labeling was performed with rabbit polyclonal anti-Ngb or anti-Cygb and mouse monoclonal anti-PKCα Ab. Bipolar cells containing PKCα were found to be Ngb- and Cygb IR (Figs. 4A 4B 4C 4D 4E 4F) . In addition, double labeling was performed with mouse monoclonal anti-Ngb and rabbit polyclonal anti-calretinin Ab. Horizontal and amacrine cells containing calretinin ⁴⁷ were found to be Ngb IR (Figs. 4G 4H 4I) . 

Discussion

The retina is one of the highest oxygen-consuming tissues in our body. ^{1 2 3} As such, a continuous and sufficient supply of O₂ is essential for the normal physiological function of the retina, and hypoxic conditions can lead to severe deficits in visual function. ⁴  

The presence of Hb in neural tissue has been extensively described in different nonvertebrate species. Hbs are essential oxygen stores in invertebrates subjected to intermittent O₂ supply, particularly in gut parasites, and in nerve and muscle tissues that exhibit sporadic high-level activities. ^{48 49 50} The duration of the oxygenation from the globin stores increases with a reduction in metabolic rates under hypoxic conditions. ⁵¹ A comparative electrophysiological study of several invertebrates, with and without Hb in their nervous tissues, ^{52 53 54 55} demonstrated that neural tissue containing Hbs (neuroHb) consumes much less O₂ during the process of action potential conduction than does neural tissue without neuroHb. Furthermore, experiments showed that oxygen bound to the neural tissue Hb in clams and some worms can support the oxygen requirements of the nervous tissue for up to 30 minutes during anoxic periods. ^{52 54 56} These results are suggestive of mechanisms in which the neuroHb-containing neural tissue may effectively use the neuroHb oxygen supplies to enable continued neuronal activity under hypoxic conditions as a highly preserved evolutionary mechanism in different species. 

The detection of Ngb and Cygb in protein extracts of canine retina and detection of Cygb, but not Ngb, in protein extracts of canine liver, kidney, lung, and heart is consistent with previously published work. ^{24 37 39} Detection of Ngb at 17 kDa molecular mass in the canine retinal tissue extract is consistent with the molecular mass of human Ngb, since the canine Ngb sequence shares 99% identity with that of human (source of the human and canine Ngb sequence: http://us.expasy.org/ provided in the public domain by the Swiss Institute of Bioinformatics, Geneva, Switzerland). Although the canine Cygb sequence has not yet been reported and assuming that it also shares high identity with the human protein, the difference in the molecular mass between human recombinant Cygb (21 kDa) and Cygb in the canine tissue samples (29 kDa) may be the result of posttranslational modification of the protein. 

A study by Schmidt et al. ²⁴ demonstrated the distribution of Ngb in ganglion cells and inner and outer plexiform layers, weak IR in both nuclear layers, and strong IR of the photoreceptor IS of the mouse retina. Although the pattern of IR was highly suggestive of neuronal localization, the lack of double-labeling experiments with neuronal and glial markers did not rule out the presence of Ngb in glial cells. Our double-labeling studies demonstrated colocalization of Ngb within retinal neurons, but not in glial cells. This finding is consistent with earlier double-labeling studies reporting neuronal localization of Ngb in different brain regions, ^{35 57 58} but differs from the reported Ngb presence in cultured astrocytes. ⁵⁹ Schmidt et al. ³⁹ also recently reported Cygb distribution in the GCL, INL, and IPL and some minor IR in the OPL of the mouse retina, which is consistent with the Cygb IR pattern observed in our study. However, we detected both Ngb and Cygb IR in the RPE in the dog. The possible difference in the IR pattern of the RPE may be attributed to the different antibodies used. Whereas the antibodies used in our study were generated against whole recombinant proteins, Schmidt et al. ³⁷ used antibodies generated against Ngb and Cygb peptide sequences. As in the case of Ngb, our double-labeling studies demonstrated localization of Cygb within retinal neurons, but not glial cells, in accordance with previous reports of the distribution of Cygb in the brain. Our findings of Ngb localization in the cytoplasm of canine retinal neurons is in agreement with reports of Ngb expression in the cytoplasm of neurons. ^{35 58 60} Several studies have described differing subcellular distribution of Cygb. Geuens et al. ⁵⁷ reported strict localization to the nuclear region in mouse brain neurons, whereas Schmidt et al. ³⁹ reported nuclear and cytoplasmic distribution in neurons within the caudate putamen, cerebral cortex, and colonic myenteric plexus ³⁷ and the mouse retinal neurons, ³⁹ with exclusive cytoplasmic location in fibroblasts. ³⁷ However, our study demonstrated Cygb IR in the cytoplasm of canine retinal neurons. Although some of our low-magnification images may suggest that Cygb is present within cytoplasm and nuclei, high-magnification confocal images of single optical sections revealed that our anti-Cygb antibody detected Cygb only in the cytoplasm. 

Our double-labeling experiments, together with the known morphologic characteristics of canine retinal cells ^{46 61 62} and identification of retinal cell types, revealed Ngb and Cygb colocalization in ganglion, amacrine, bipolar, horizontal, and retinal pigment epithelial cells and Ngb localization in both rod and cone photoreceptor IS. Although our results suggest different roles of Ngb and Cygb in the mammalian retina, they do not permit discrimination between possible distinct Ngb and Cygb functions. Further in vivo and in vitro functional studies are needed before the exact function of these proteins in the retina can be established. 

Although the neuroprotective function of Ngb in the brain has been reported in recent studies, ^{33 34} the role of Cygb in the nervous system is still under active investigation. Different theories exist for Ngb and Cygb function, such as a role in oxygen storage and facilitated oxygen transport. However, this particular hypothesis has been recently disputed for Ngb by mathematical modeling of retinal oxygen consumption, which suggests that even a concentration of 100 μM is not sufficient to provide adequate oxygen supply. ⁶³ Furthermore, Sun et al. ³³ showed that increased Ngb expression did not cause an increase in oxygen consumption, and hypothesized that Ngb’s neuroprotective activity was not related to increased neuronal oxygen transport. This finding was not unexpected, considering the small concentration of Ngb within the brain. ¹⁰  

Recent studies have suggested that Ngb may serve as an oxygen-hypoxia sensing and signaling molecule due to the ability of metNgb to bind to G proteins (Gαi). ³² In addition, studies using a yeast two-hybrid system demonstrated that cystatin C (a cysteine proteinase inhibitor) is an Ngb-binding protein, ⁴⁰ as is flotillin 1, a lipid raft protein. These findings have been reviewed in light of the possible neuroprotective mechanism of Ngb by Wakasugi et al.. ⁶⁴  

Furthermore, it has been hypothesized that Ngb and Cygb may act as enzymes for detoxification of free oxidative radicals, and it was suggested that scavenging of radical-derived organic peroxides by Cygb could be an adaptive reaction to normalize the cellular redox status during postischemic cell activation. ³⁶ Also, it has been shown that an oxygenated derivative of Ngb reacts very rapidly with free NO, yielding as a product metNgb, which may interact specifically with components of GDP-GTP signal-transduction pathways. ⁴¹  

The present study has demonstrated the immunolocalization of Ngb and Cygb in retinal layers with high oxygen demand and particularly in the retinal ganglion cells, which are the primary cell population affected by glaucoma and retinal cell population that is the most sensitive to ischemic diseases of the retina and optic nerve. Hemodynamic alterations, leading to decreased ocular blood perfusion and ischemia, can often precede normal-tension glaucoma. ⁶ In addition, decreased optic nerve blood flow in humans correlates with functional and morphologic measures of the progression of glaucoma, ⁷ and treatments designed to improve ocular blood flow have been shown to benefit patients with glaucoma. ⁸ Furthermore, the presence of tissue hypoxia in the retina and optic nerve head of glaucomatous eyes ^{65 66 67} is consistent with the finding of increased hypoxia-inducible factor (HIF)-1α in human glaucomatous retinas and optic nerve heads. ⁶⁸ Our recent preliminary results suggest an upregulation of Cygb mRNA in glaucomatous mouse eyes (Grozdanić et al. IOVS 2004;45:ARVO E-Abstract 2586). Also, it has been reported recently that Ngb expression is increased in the retina of human eyes with chronic glaucoma (Rajendram et al. IOVS 2005;46:ARVO E-Abstract 1313). These data are not surprising, considering previous reports which demonstrated Cygb upregulation by HIF-1α ⁶⁹ and increased Ngb and Cygb protein expression in different tissues after exposure to hypoxia. ^{33 37} It has been known for almost 40 years that Hb is the principal oxygen-transporting molecule that is significantly upregulated when an organism is exposed to hypoxia, to provide adequate oxygenation of all vital organs and maintain cell survival. Despite the enormous retinal oxygen and energy demands, it is not known why the retina is capable of tolerating an almost 10 times longer period of ischemia than is brain tissue. ^{70 71} We hypothesize that Ngb and Cygb may have a vital role in retinal oxygen homeostasis and enable the retina to sustain longer periods of ischemia. Exact identification of the functional properties of Ngb and Cygb will significantly advance our understanding of retinal oxygenation in health and disease. 

³  Contributed equally to the work and therefore should be considered equivalent authors.

 Supported by The Glaucoma Foundation National Institute of Neurological Disorders and Stroke (NINDS) NS44007, the Carver Trust, and the ISU Biotechnology Foundation.

 Submitted for publication April 15, 2005; revised September 21, November 5, and November 28, 2005; accepted January 19, 2006.

 Disclosure: J. Ostojić, None; D. S. Sakaguchi, None; Y. de Lathouder, None; M. S. Hargrove, None; J.T. Trent III, None; Y. H. Kwon, None; R. H. Kardon, None; M. H. Kuehn, None; D. M. Betts, None; S. Grozdanić, BioVendor Laboratory Medicine, Inc. (F)

 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.

 

Figure 1.

 

Western blot analysis of Ngb and Cygb expression in canine tissues. (A) Mouse monoclonal anti-Ngb Ab (mAb) and rabbit polyclonal anti-Ngb Ab (pAb) detected recombinant Ngb protein at ∼17 kDa. Approximately 0.08 μg of the recombinant protein was loaded per lane. (B) Rabbit polyclonal anti-Ngb Ab detected Ngb in the canine retina. Approximately 0.003 μg of the recombinant Ngb was applied as the positive control. (C) Rabbit polyclonal anti-Cygb Ab detected specific recombinant Cygb protein at ∼21 kDa and Cygb in canine retina, kidney, liver, lung, and heart at ∼29 kDa. Approximately 0.02 μg of the recombinant protein was applied per lane. Left: molecular mass markers.

Figure 2.

 

Ngb and Cygb immunolocalization in the canine retina. Confocal images illustrating patterns of IR for (A) anti-Ngb mouse monoclonal antibody (mAb), (B) anti-Ngb rabbit polyclonal antibody (pAb), and (C) anti-Cygb rabbit polyclonal antibody. Both anti-Ngb antibodies showed similar patterns of labeling in the retina (compare A and B). Differential interference contrast (DIC) images of Ngb and Cygb immunolocalization in canine central retina of (D) anti-Ngb mouse monoclonal antibody, (E) anti-Ngb rabbit polyclonal antibody, and (F) anti-Cygb rabbit polyclonal antibody. Confocal images of the retina double-labeled for (G) Ngb (red) and (H) Cygb (green) displays IR of both proteins in the same neurons, as illustrated in the merged image (I). White arrows: examples of retinal ganglion cells; open arrows: examples of cells that are presumably amacrine cells; asterisks: bipolar cells; arrowhead: example of a horizontal cell. Cell nuclei (blue) were labeled with DAPI. Scale bar: (A–C) 25 μm; (D–F) 20 μm; (G–I) 10 μm. Abbreviations the same for all figures: RPE, retinal pigment epithelium; IS, inner segments; ONL, outer nuclear layer; OPL, outer plexiform layer; INL, inner nuclear layer; IPL, inner plexiform layer; GCL, ganglion cell layer; NFL, nerve fiber layer.

Figure 3.

 

Ngb and Cygb were expressed in neurons. Images were captured with confocal microscopy. (A) Ngb IR (red) in two β-ganglion cells and one smaller cell between them (either γ-ganglion cell or a displaced amacrine cell) and (B) MAP-2 IR (green) displayed positive double labeling in a merged image (C). (D) Cygb IR (red) in α- and β-ganglion cells and a horizontal cell and (E) MAP-2 IR (green) displayed double labeling in a merged image (F). (G) Ngb IR (red) in three β-ganglion cells, amacrine cell, and presumably a bipolar cell and (H) TUJ1 IR (green) showed double labeling in a merged image (I). (J) Cygb IR in α and β ganglion cells and two amacrine cells and (K) TUJ1 IR (green) showed double labeling in a merged image (L). White arrows: examples of retinal ganglion cells; open arrows: examples of amacrine cells; arrowhead: horizontal cell; asterisk: putative bipolar cell. Blue: cell nuclei labeled with DAPI. Scale bar, 10 μm.

Figure 4.

 

Ngb and Cygb localization in the INL and IPL. Images were captured with confocal microscopy. (A) Ngb IR (red) and (B) PKCα IR (green) double labeling in bipolar cells and their axonal terminals as shown in (C) a merged image. (D) Cygb IR (red) and (E) PKCα IR (green) double labeling in bipolar cells and their axonal terminals as shown in (F) a merged image. (G) Ngb IR (red) and (H) calretinin IR (green) double labeling in horizontal and amacrine cells as shown in (I) a merged image. (A–F) White arrows: examples of bipolar cells; (G–I) white arrows: examples of horizontal cells; open arrows: examples of amacrine cells. Blue: cell nuclei with DAPI. Scale bars, 10 μm.

Supplementary Materials

SupplementaryTable S1 - 79.9 KB (PDF) 

Supplementary Figure S1 - 154 KB (PDF) 

Supplementary Figure S2 - 153 KB (PDF) 

Supplementary Figure S3 - 258 KB (PDF) 

The authors thank Matt Harper, Jeffrey Orasky, Milan Joksimović, Daniel Zamzow, and Janice Buss for technical assistance during preparation of the manuscript; Daria Pospisilova, Petr Kasparek, and Jiri Havlasek from BioVendor Laboratory Medicine, Inc. for providing us with mouse monoclonal anti-Ngb Ab and recombinant protein; and John C. Saari (Department of Ophthalmology, University of Washington, Seattle, WA) for the CRALBP antibody.