| Glycobiology | Pages |
Changes in ganglioside composition of photoreceptors during postnatal maturation of the rat retina
Introduction
Results
Discussion
Materials and methods
Acknowledgments
Abbreviations
References
Changes in ganglioside composition of photoreceptors during postnatal maturation of the rat retina
To examine at which stage the unusual ganglioside composition observed in adult retinal photoreceptor cells was established, and to see whether ganglioside changes could be correlated to distinct maturational events, quantitative and qualitative variations in gangliosides within pure sheets of photoreceptors during postnatal differentiation and aging of retina were studied. Retinas were separated into their component layers, (particularly photoreceptor layers uncontaminated by other neuronal types) by exploiting a technique of mechanical separation by vibratome. We extracted lipids from the cell membranes and analyzed the ganglioside composition by high performance thin layer chromatography. The data show that from the earliest recordable postnatal age (6 days) until late in life (18 months), photoreceptors contain low quantities of lipid-bound N-acetyl neuraminic acid and a simplified ganglioside profile compared to inner retinal neurons. Specific ganglioside changes occur within photoreceptor cells during postnatal maturation and aging, with downregulation of a-pathway GM1 and overlapping upregulation of b-pathway GD1b taking place during the period corresponding to outer segment formation, correlating with the onset of retinal function. Key words: retina/photoreceptors/differentiation/aging/gangliosides
Introduction
Gangliosides (GG) are complex glycolipids that are found in cell plasma membranes, especially in nerve cells, and are believed to play a role in nerve regeneration (Gorio et al., 1983; Ledeen, 1983), receptor-mediated signal transduction (Hakomori, 1981; Nagai, 1995), as well as direct cell-cell interactions (Cheresh et al., 1986). It has been shown that the GG composition of the central nervous system (CNS) changes during development, and that these changes correlate with developmental events (Yu et al., 1988). The general rule emerging from these studies is that GG of the lactosylceramide series (mainly GD3) predominate at early developmental stages when neural cells proliferate actively. At later developmental stages, as morphological and chemical differentiation of these cells proceeds, the GD3 content declines and GG of the gangliotetraosyl series become the major forms (Maccioni et al., 1989).
Within the CNS, the mammalian retina is an interesting exception because GD3 is the major GG in the mature tissue (Dreyfus et al., 1975; Daniotti et al., 1991, 1992). Relatively little, however, is known about variations in GG within identified retinal cell types such as photoreceptors (PR) during differentiation and aging of retina. The PR is known to be a neuron which has an elevated and complex metabolism (Bok, 1985), its main role within the retina being to transform light energy into an electric signal. Generation of rod PR in rat retina overlaps approximately with birth of the animal, and by 1 week following birth the PR pool is partially differentiated (expression of specific proteins such as opsin, arrestin, and interstitial retinoid binding protein: Hicks and Barnstable, 1987; Barnstable, 1990). The outer retina (OR) becomes functional 2-3 weeks after birth, and during this period multiple changes occur in the retina and especially in the PR. The cell develops at its apical pole a specialized organelle formed of stacked discs, the outer segment (OS), and elaborates synaptic contacts at its vitreal pole (Weidman and Kuwabara, 1968), hence involving elevated neosynthesis of membrane components. Previously, we described GG composition of adult PR (Dreyfus et al., 1996), showing the unusual simplicity of the GG profiles of these neurons relative to other retinal and CNS-derived neurons.
Given the major differentiative changes that occur in the first postnatal month, we wanted to study whether GG composition is also differentially regulated during this period of postnatal maturation, and also during aging. By exploiting the mechanical separation of retinas aged from 6 days (6d) to 18 months (18m) into their component layers, particularly PR layers uncontaminated by other neuronal types, we extracted lipids from the cell membranes and analyzed the GG composition by high performance thin layer chromatography (HPTLC). These data show that large differences in GG make-up between PR and inner retinal neurons exist from very early in maturation, and that notable GG changes occur within the OR during postnatal maturation and aging. Especially downregulation of a-pathway GM1 and upregulation of b-pathway GD1b take place during the period corresponding to OS maturation, correlating with the onset of retinal function.
Results
NeuNAc quantification in photoreceptors during maturation
Quantification of the lipid-associated NeuNAc of OR, inner retina (IR) and whole retina (WR) at different ages showed highly significant differences. Independently of the age, the amount of lipid-bound NeuNAc of IR was always much larger compared to OR or WR. Already at the earliest time point sampled (6d), NeuNAc levels within the IR were almost 3-fold those of the OR. Throughout the time span measured, IR contained between 2.9 and 5.6 µg NeuNAc/ mg protein whereas these amounts were between 1.3 and 1.8 µg NeuNAc/ mg protein for OR (Table 1). NeuNAc levels within the OR at different ages remained roughly constant.
Table 1
| Parameter | 6d | 10d | 15d | 1m |
| Whole retina | ||||
| Protein (mg/tissue) | 0.71 ± 0.11 | 0.65 ± 0.10 | 0.88 ± 0.15 | 0.80 ± 0.10 |
| NeuNAc (µg/mg protein) | 2.1 ± 0.2 | 1.9 ± 0.2 | 2.7 ± 0.4 | 3.0 ± 0.3 |
| GG (nmol/mg protein) | 4.2 ± 0.3 | 3.9 ± 0.5 | 4.3 ± 0.7 | 4.9 ± 0.7 |
| Outer retina | ||||
| Protein (mg/tissue) | 0.43 ± 0.10 | 0.41 ± 0.03 | 0.44 ± 0.01 | 0.34 ± 0.05 |
| NeuNAc (µg/mg protein) | 1.5 ± 0.1 | 1.7 ± 0.3 | 1.8 ± 0.1 | 2.2 ± 0.1 |
| GG (nmol/mg protein) | 3.2 ± 0.3 | 4.0 ± 0.7 | 3.7 ± 0.2 | 4.1 ± 0.4 |
| Inner retina | ||||
| Protein (mg/tissue) | 0.21 ± 0.04 | 0.15 ± 0.03 | 0.16 ± 0.02 | 0.18 ± 0.03 |
| NeuNAc (µg/mg protein) | 3.5 ± 0.1 | 4.8 ± 1.1 | 6.1 ± 0.1 | 5.6 ± 1.3 |
| GG (nmol/mg protein) | 6.1 ± 0.7 | 8.3 ± 1.5 | 10.7 ± 0.1 | 8.6 ± 2.1 |
Ganglioside distribution in photoreceptors
HPTLC separation performed with the samples from different ages revealed a spectrum of GG species, ranging from the simplest (GM3) to the most complex (GQ1), with a prominent band corresponding to GD3 (Figure 1). Many GG migrated as closely spaced doublets: for GD3 the lower of the two bands was more intense. Minor bands were observed migrating between GD1a and GD1b, and between GD1b and GT1b, whose identity remained unresolved. Positions of some GG bands appeared displaced from those of parallel standards: in all cases the GD3 doublet migrated below that of standard GD3, and GT1b was often slightly above the corresponding standard, particularly in the OR (Figure 1); GM1, GD1a, and GD1b mostly aligned with the brain standards. Comparison of HPTLC plates of OR, IR, and WR from different ages (6d, 1m, 4m, 9m, and 18m are illustrated in Figure 1a-d) demonstrated differences in GG of these three different samples, but also within each fraction at different ages. Throughout the ages examined, IR and WR always exhibited more complex GG profiles than OR, with abundant amounts of disialoGG (GD3, GD1a, and GD1b) and trisialoGG (GT1b). Immature OR contained mainly monosialoGG (GM3, GM1) and very large amounts of GD3, with minor concentrations of other disialoGG (GD1a and GD1b) and more complex GG (Figure 1a). As the retina matured, a marked reduction in GM1 and a more intense band of GD1b were observed (Figure 1a-d). GT1b always remained a minor band. Changes in the IR were different, with a maintained expression of GM1, and increases in GD1b and GT1b. We noticed that under the solvent conditions used in these studies, GQ1 migrated closely to GT1b within IR and WR fractions. In subsequent experiments in which the proportion of M within the solvent mixture was slightly reduced (C:M:CaCl2 , 50:37:10 vol/vol/vol), faint but distinct bands of GQ1 were visible in WR and IR (data not shown). Under no conditions was GQ1 ever observed in OR.
Figure
Quantitative analysis of ganglioside profiles as a function of age
Densitometric scanning of HPTLC plates such as those shown in figure 1 permitted quantification of GG profiles (Figures 2, 3). The scans illustrate representative GG profiles of OR from 6d, 15d and 3m (Figure 2), and those from corresponding ages of IR (Figure 3). These data were used to analyze trends in variations of GG levels as a function of tissue fraction and age (Figure 4).
Figure
Figure
There was a large decrease in GM1 between 10d (26.9 ± 3.9% of total GG), 15d (14.0 ± 4.0%) and 1m (7.6 ± 1.1%), followed by its stabilization at later stages (Figure 4a). Hence GM1 showed maximal reduction (67%) between 10d and 15d. GD1b content of OR showed approximately linear increases between 10d (4.3 ± 1.2%) and 1m (21.4 ± 1.3%), incrementing by 50% every 5d (Figure 4b). GD3 was the most abundant GG throughout maturation and aging of the OR, representing almost half the total GG at all time points. GM3 and GD1a levels did not change significantly between 6d and 18m, presenting respective average levels of 12% and 10% lipid-bound NeuNAc. Relative levels for all GG remained stable between 1m and 18m (Figure 4c,d) except for GT1b which although representing the least abundant GG of OR showed a significant increase between 3m and 18m, changing from 2.2% to 9.5% of the total lipid-bound NeuNAc (Figure 4d).
Lipid-bound NeuNAc levels of IR changed more moderately than those of OR. We observed a 2-fold increase in GM1 between 6d and 10d, followed by a slight decrease between 10d (14.4 ± 1.7%) and 1m (7.9 ± 1.7%) (Figure 4e). GD1a showed a linear decrease between 10d and 1m showing a loss of 27%. Over the same period, levels of GD1b and GT1b increased 48% and 38%, respectively (Figure 4f). Levels of the simplest GG GM3 and GD3 remained stable during IR maturation, GD3 accounting for the most elevated GG with an average level of 30%.
Figure
Discussion
We have investigated the GG patterns of rat retinal PR at different times in postnatal life. Our data indicate that from the earliest recordable times PR have very different GG levels and profiles with respect to other retinal neurons, and that GG changes occur further during postnatal maturation of this cell in the rat retina.
The laminated architecture of the postnatal rat retina has allowed the use of a vibratome technique consisting of the mechanical separation of the OR from the IR. In a previous study we showed the very high purity of PR within the OR (> 95% were shown to be rods; Dreyfus et al., 1996). The OR fraction contains essentially PR but also the apical portions of retinal Müller cells which represent a minor fraction of this material. In our previous work, we compared levels of the different lipid classes within the OR of adult (3m) rat to that of the IR, rod OS, WR, and brain. Here, we confirm that adult PR contain far less GG-bound NeuNAc than IR, and demonstrate that this difference exists throughout maturation and aging of the retina. The low NeuNAc content and relative simplicity of the GG species can be considered a fundamental characteristic of PR identity, as it occurs at least as early as termination of cell division and outer nuclear layer appearance (Weidman and Kuwabara, 1968; Cepko, 1993). Such properties may be related to PR function: for example, as GG are thought to contribute to membrane rigidity (Barbour et al., 1992) such low contents would favor membrane fluidity, perhaps a necessary feature of the rapid membrane turnover observed in these cells (Bok, 1985). Albeit the low GG content (and that of sphingolipids in general: Fliesler and Anderson, 1982; Dreyfus et al., 1997) of PR, such molecules may still play important roles. Sphingolipid and GG content of non-neural tissues is also lower than brain, but these components are still thought to participate in key biological processes (e.g., Hannun and Bell, 1989; Nagai and Iwamori, 1995). Also submicromolar concentrations of GG stimulate calmodulin-dependent cyclic nucleotide phosphodiesterase in vitro (Yates et al., 1989). Finally, the relative enrichment of GD3 would probably modify membrane behavior (see below).
The choice of 6d postnatal as the first point studied was dictated by the fact that at this age the retina first acquires a laminated architecture, allowing vibratome separation. At this time, the OR is composed of postmitotic PR cells without OS and therefore light-insensitive (Weidman and Kuwabara, 1968; Hicks and Barnstable, 1987). These PR displayed a very simplified GG profile, composed mainly (>80%) of GD3, GM3, and GM1. More complex GG accounted for a very small proportion and no tetrasialoGG were detected. PR undergo dramatic increases in addition of membranes between 10d and 1m, with the rapid growth and formation of the OS and synaptic pedicles at the apical and basal poles, respectively. Ultrastructural and immunohistochemical data suggest that synapse formation in the outer plexiform layer of rodents is largely complete by 14d (Blanks et al., 1974; Sarthy and Bacon, 1985). OS formation occurs maximally between 10d and 20d, increasing less rapidly between 20 and 30d (LaVail 1973). Additionally, the activity of rhodopsin kinase, which reflects the light-dependent activation and hence onset of function of rhodopsin, increases maximally between 10d and 30d (Ho et al., 1986). The two major changes in PR composition observed over this period, namely a 3-fold diminution in GM1 and a 4-fold increase in GD1b, hence correlate more closely with OS formation than synaptogenesis. Furthermore, absolute levels of GD1b are enriched in isolated OS compared to whole PR (Dreyfus et al., 1996). These data thus support a role for GD1b in OS structure and/or function. Such roles could conceivably include phagocytosis of OS by the overlying retinal pigmented epithelium, which has been shown to be influenced by other glycoconjugates (Gregory et al., 1990). Furthermore, glycosphingolipids have been shown to stimulate phagocytosis in macrophages (Miyazaki et al., 1995). It should be noted that the major reduction in GM1 levels occurs during the early rise phase in GD1b, suggesting that conversion of the former into the latter may be taking place. It should be noted however that in brain these two GG are placed on different biosynthetic pathways (Freischütz et al., 1995; Figure 5), and in the present study we do not know if conversion of GM1 to GD1b occurs directly or through reconversion to simpler precursors (GM3) and subsequent modification through GD3. Daniotti et al. (1991) previously reported similar changes (reduction in GM1, increase in GD1b) in whole developing rat retina, whereas our data indicate these modifications are related more to PR maturation than to that of IR.
Figure
Elsewhere in the CNS, GD3 is found at relatively high levels in reactive glia (Bernheimer et al., 1979), glioma (Traylor and Hogan, 1980), and immature neurons (Dreyfus et al., 1975; reviewed by Seyfried and Yu, 1985). Although evidence is somewhat conflictual concerning retinal cell types expressing GD3 (some authors have attributed it mainly to Müller glia (Seyfried et al., 1982), others to retinal neurones (Daniotti et al., 1992)), the present data showed the PR to contain especially high amounts (45-50% total GG throughout life). We previously showed that cultured purified Müller glia contain no detectable GD3 (Hicks et al., 1996), although neuron-glial interactions may be necessary to maintain GD3 expression (Daniotti et al., 1992). Seyfried and Yu (1985) suggest that membranes enriched in GD3 would exhibit enhanced permeability to ions and metabolites, and such activities would certainly be important to PR function. After terminal maturation of the retina, the GG distribution within the OR remained relatively constant, except for GT1b which increased 4-fold between 3m and 18m. GT1b might constitute a marker for aging in this cell population.
The precise identification of individual GG by HPTLC is problematic. It is well known that many GG migrate as doublets, due to differences in fatty acid and long-chain base composition (Ando and Yu, 1984). Such differences may explain also the displacement of retinal GD3 with respect to standards. Additional variations arise from substitution of N-glycolylneuraminic acid for NeuNAc (Kawashima et al., 1993) and lactonic and acetylated isomers (Dreyfus et al., 1975; Blum and Barnstable, 1987), and it is possible that a minor band observed in retina corresponds to this lactonic species. Furthermore, we have observed that differences in sample GG composition (e.g., complex mixtures compared to only a few major bands) and amounts of starting material loaded onto HPTLC plates both lead to variations in migration (Dreyfus et al., 1997). Particularly GM3 and GT1b are highly sensitive to extraction and purification procedures (Dreyfus et al., 1997). Thus, more sophisticated analytical approaches would be necessary to determine unequivocally the exact structures of many of the GG bands observed here.
In summary, this study presents the first comprehensive survey of PR GG composition throughout the lifespan of the rat. The data illustrate profound differences between this tissue and other CNS regions, and as the experimental techniques used here permit only a static analysis of a highly dynamic situation, it will be very interesting to identify the enzymatic pathways involved. Furthermore as the complexity of brain organization has rendered difficult a detailed analysis of GG function, the relatively simple structure of the retina and the existence of well-defined tissue culture models should facilitate such studies.
Materials and methods
Materials
All solvents and reagents were of analytical grade. Silicic acid (reference 109385) and HPTLC plates (10 × 10cm, reference 1.05628) were obtained from Merck, Darmstadt, Germany. Sephadex G25 Superfine was obtained from Pharmacia, Uppsala, Sweden. GG standards were purchased from Matreya Inc., Pleasant Gap, PA.
Tissue collection
All animals used in these studies were cared for and handled according to the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. Wistar rats from different ages (6d postnatal to 18m) were used for these experiments. They were anesthetized by CO2 inhalation, rapidly killed by cervical dislocation, and enucleated. The retinas were sectioned with a vibratome and separated into IR and OR as described previously (Silverman and Hughes, 1989; Dreyfus et al., 1996). For comparative purposes, whole retinas (WR) from Wistar rats were also removed.
Protein determination
Protein levels in homogenates of IR, OR and WR were determined using the Lowry method (Lowry et al., 1951).
Extraction, purification, and separation of gangliosides
Retinal lipids were extracted and purified as recently described by us (Dreyfus et al., 1996). Briefly, the samples were homogenized in 0.5 ml dd H2O, and total lipids were extracted with successively: 4.5 ml chloroform (C): methanol (M) (1:1 vol/vol), 2.5 ml C:M:H2O (5:5:1 vol/vol/vol), 2.5 ml C:M (1:1 vol/vol), and 2.5 ml C:M:H2O (50:40:10 vol/vol/vol). The organic fractions were pooled, dried, redissolved in 2.5ml C:M:H2O (60:30:4.5 vol/vol/vol), and loaded onto a Sephadex G25 column (diameter 0.9 cm, height 3.5 cm) equilibrated in the same solvent. Elutions were then performed with 5 ml of the same solvent, followed by 2.5 ml C:M (2:1 vol/vol) and 2.5 ml C:M:H2O (50:40:10 vol/vol/vol). The purified lipids were dried, redissolved in 0.5 ml C and loaded onto a small silicic acid column (diameter 0.6 cm, height 2.5 cm). Following elution of neutral lipids, phospholipids, and sphingolipids with different solvent mixtures (Dreyfus et al., 1996), the GG were eluted with 2.5 ml C:M (2:3 vol/vol), 2.5 ml C:M:H2O (65:25:4 vol/vol/vol), 2.5 ml C:M:H2O (60:35:8 vol/vol/vol), and 2.5 ml C:M:H2O (50:40:10 vol/vol/vol). After GG N-acetylneuraminic acid (NeuNAc) had been quantified (Svennerholm, 1957), GG analysis was performed by HPTLC as described by us previously (Dreyfus at al., 1996). For the OR 3 µg NeuNAc were spotted onto the plate using a CAMAG automatic TLC Sampler III, and 5 µg NeuNAc for IR, WR, and standards (St). GG were visualized using resorcinol-hydrochloric acid reagent (Svennerholm, 1957), and GG patterns were quantified by densitometric scanning at 577 nm on a CAMAG TLC Scanner III using a software analysis package for densitometric evaluation of TLC (CATS program). To calculate average GG concentrations, scanning data were pooled from three to six HPTLC plates for each age and for each fraction.
Statistics
Data were compared using the parametric Peritz f-test according to Harper, accepting significance values of *, p < 0.05; **, p < 0.01; and ***, p < 0.001 (Harper, 1984).
Acknowledgments
We thank Dr. B. Guérold for technical help with ganglioside separation. V.F. is grateful for financial assistance from IPSEN and Retina France-AFRP.
Abbreviations
Cer, ceramide; C, chloroform; GalCer, galactosylceramide; GalT, galactosyltransferase; GG, ganglioside; GlcCer, glucosylceramide; GlcT, glucosyltransferase; HPTLC, high performance thin layer chromatography; IR, inner retina; LacCer, lactosylceramide; M, methanol; NeuNAc, N-acetylneuraminic acid; OR, outer retina ; OS, outer segment; PR, photoreceptor; ST-I, II, and III, sialyltransferase I, II, and III, respectively; WR, whole retina. Gangliosides are named following the nomenclature of Svennerholm (1980).
References
This page is run by Oxford University Press, Great Clarendon Street, Oxford OX2 6DP, as part of the OUP Journals
Comments and feedback: www-admin{at}oup.co.uk
Last modification: 27 Jan 1998
Copyright© Oxford University Press, 1998.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  E. J. Park, M. Suh, and M. T. Clandinin Dietary Ganglioside and Long-Chain Polyunsaturated Fatty Acids Increase Ganglioside GD3 Content and Alter the Phospholipid Profile in Neonatal Rat Retina Invest. Ophthalmol. Vis. Sci., July 1, 2005; 46(7): 2571 - 2575. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 N. Kinkl, J. Sahel, and D. Hicks Alternate FGF2-ERK1/2 Signaling Pathways in Retinal Photoreceptor and Glial Cells in Vitro J. Biol. Chem., November 16, 2001; 276(47): 43871 - 43878. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 V. Fontaine, N. Kinkl, J. Sahel, H. Dreyfus, and D. Hicks Survival of Purified Rat Photoreceptors In Vitro Is Stimulated Directly by Fibroblast Growth Factor-2 J. Neurosci., December 1, 1998; 18(23): 9662 - 9672. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||