| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
1From the Department of Anatomy and Cell Biology, University of Melbourne, Victoria, Australia; and 2Department of Physiology, Monash University, Clayton, Victoria, Australia.
 |
Abstract |
|---|
 Top Abstract MethodsStatistical AnalysisResultsDiscussionConclusionReferences |
|---|
METHODS. FGR was induced by umbilicoplacental embolization (UPE) in two cohorts of pregnant ewes from (1) 120 days of gestation (dg) until 140 dg and (2) 120 dg until term (
147 dg). Control fetuses were not subjected to UPE. The structure and neurochemistry of the retina and number and structure of ganglion cell axons were assessed in near-term (140 dg) and adult animals (2.3 years).
RESULTS. In near-term FGR fetuses compared with control fetuses there were significant reductions (P < 0.05) in the outer plexiform layer (OPL), the photoreceptor inner and outer segment layers, the inner nuclear layer (INL) in the central retina and the outer nuclear layer (ONL) in the peripheral retina, and the diameter of ganglion cell axons in the optic nerve, with a proportional reduction in the thickness of myelin sheaths. In FGR animals compared with the control at 2.3 years, there were significant reductions (P < 0.05) in the total thickness of the retina, the thickness of the photoreceptor outer segment layer and the INL and the number of tyrosine hydroxylase-immunoreactive (TH-IR) dopaminergic amacrine cells. Axonal diameter and myelin sheath thickness in the optic nerve were not different (P > 0.05) between groups.
CONCLUSIONS. Chronic placental insufficiency in late gestation results in long-lasting effects on specific retinal components, including photoreceptor outer segments and TH-IR amacrine cells. Other alterations observed at term, including reductions in growth and myelination of optic nerve axons, do not persist, suggesting delayed rather than permanently compromised development. Alterations persisting into adulthood could affect visual function.
Experimental studies of the visual system to date have focused on the short-term effects of prenatal compromise and have demonstrated structural alterations in the retina as a consequence of prenatal exposure to ethanol,8 cocaine,9 corticosteroids,10 and chronic placental insufficiency (CPI).11 12 13 Functional alterations have been reported, however, after exposure to ethanol14 and deprivation of omega 3 fatty acid.15 16 17 Much less is known about whether these effects are long-lasting and contribute to visual deficits at maturity. We have already shown that CPI in the guinea pig results in long-lasting effects on dopaminergic amacrine cells,13 which are thought to be involved in the mechanisms underlying contrast sensitivity.5 18
In the present study, we developed an ovine model of late gestational CPI, sufficient to cause fetal growth restriction (FGR), and extensively examined the morphology and neurochemistry of the retina and optic nerve at near-term (140 days of gestation (dg) and in adulthood. CPI was induced by placental embolization from 120 to 140 dg19 ; 120 dg is equivalent to approximately 28 to 30 weeks of gestation in humans. At this stage, neurogenesis is complete in the retina; however, synaptogenesis and dendritic elaboration are still occurring.10 The retina was examined to determine the effects of CPI on specific classes of amacrine cells, ganglion cell survival, growth and myelination of axons, synaptogenesis, and photoreceptor integrity, all parameters essential for normal visual function.
|
Methods |
|---|
|
TopAbstractMethodsStatistical AnalysisResultsDiscussionConclusionReferences |
|---|
Two cohorts of animals were examined: a near-term cohort, in which CPI was induced via umbilicoplacental embolization (UPE) from 120 to 140 days of gestation (dg) (assessment at 140 dg), with fetuses delivered via caesarean section at 140 dg, and an adult cohort in which UPE was performed from 120 dg to the onset of labor at term (147 dg; assessment at 2.3 years). Lambs were born naturally and raised with their mothers until weaning. They were then raised as a flock until autopsy as adults (2.3 years of age). Retinas and optic nerves were collected from both cohorts at autopsy and processed for subsequent structural and neurochemical analyses. For each group, all protocols were identical and have therefore been described together. The number of animals and retinas used in each group are indicated in the text.
Surgical Preparation
At 115 ± 2 days after mating pregnant ewes (near-term cohort, n = 17; adult cohort, n = 18) were subjected to aseptic surgery (halothane, 1% to 2% in O2) for the insertion of catheters into a fetal femoral artery.19 20 Catheters were used for both the injection of microspheres and sampling of fetal blood. Antibiotics (procaine penicillin 180,000 U mg/mL and dihydrostreptomycin 250 mg/mL; Invet, Bendigo, Victoria, Australia) were administered (1 mL, intramuscularly) to the fetus before closure of surgical incisions. After surgery, ewes were housed in individual cages with free access to food and water. Animals were allowed to recover for 5 ± 2 days before experimentation commenced.
Experimental Protocol
At 120 dg, in randomly assigned fetuses (near-term cohort, n = 9; adult cohort, n = 9), non–radio-labeled mucopolysaccharide microspheres (0.05–0.2 x 106 microspheres/d) were injected into the femoral aortic catheter to achieve UPE.21 Throughout this study, animals subjected before birth to UPE will be referred to as FGR. Control fetuses (near-term cohort, n = 8; adult cohort, n = 9) were subjected to surgery and blood gas sampling, but no microspheres were injected. A further group of noncatheterized control fetuses (adult cohort, n = 6) were also included in the study.
Fetal arterial blood was sampled daily throughout the experimental period, and the results have been published previously.19 20 Briefly, the partial pressure of oxygen (PaO2), pH, and oxygen saturation (SaO2) were reduced (P < 0.05), whereas the partial pressure of carbon dioxide (PaCO2; P < 0.05) was increased in FGR fetuses compared with control fetuses throughout the UPE period. Blood glucose concentrations were also lower in FGR fetuses than in the control (P < 0.05).20
Tissue Preparation
At the end of the experimental period, both groups of animals (near term: 140 ± 1 dg; adult, 788 ± 15 days after birth) were killed with an overdose of pentobarbitone, 130 mg/kg, intraperitoneally; Vitbac Animal Health, New South Wales, Australia) and the body and brain weights were recorded. The retinas and optic nerves were perfused in situ with 4% paraformaldehyde (PFA) in 0.1 M phosphate buffer (PB), after which the eyes, to which some optic nerve was still attached, were enucleated. The corneas were pierced, the temporal aspect marked, the eyes weighed, and the whole eyes immersion-fixed in 4% PFA in 0.1 M PB for 12 hours.
Retina
Retinas were dissected from both eyes. The left retina at both ages was postfixed for a further 2 hours in fresh 4% PFA. From the right retina, small blocks of tissue (2 x 5 mm) were collected both centrally (TC, immediately adjacent to the optic nerve head in the temporal aspect) and peripherally (TP, inferior temporal quadrant). One block of retina was embedded in 1.5% agar (to aid orientation) and processed to paraffin for immunohistochemistry. The second block was placed in 1% glutaraldehyde in 4% PFA for 24 hours, postfixed in 1% osmium tetroxide for 30 minutes, stained with 2% uranyl acetate, and embedded in Epon-Araldite (ProSciTech; Thuringowa, Queensland, Australia) for ultrastructural analysis.
As it was desirable to maximize the number of parameters that could be assessed, the right retina from each animal was further sectioned into quadrants that were designated temporal inferior (TI), temporal superior (TS), nasal superior (NS), and nasal inferior (NI). Each quadrant was then used for a different immunohistochemical or histochemical marker.
Immunohistochemistry.
Immunohistochemistry was performed on retinal wholemounts or vertical paraffin-embedded (8 µm) sections of the TC and TP retina. Retinas were processed for immunoreactivity (IR) using the avidin-biotin peroxidase complex (Vector Laboratories, Burlingame, CA), as previously described12 with the following antibodies used at the dilutions shown: rabbit anti-glial fibrillary acidic protein (1:1000, GFAP; Sigma-Aldrich, St. Louis, MO) was used to identify astrocytes; goat anti-choline acetyltransferase (1:1000, ChAT; Chemicon International, Temecula, CA); rat anti-substance P (1:11000; BD Biosciences, San Diego, CA); and mouse anti-TH (1:1000, tyrosine hydroxylase, Chemicon International) were used to identify amacrine cell subpopulations, and mouse anti-synaptophysin (1:10 000; Sigma-Aldrich) was used to identify synapses. Sections were incubated overnight (72 hours for mouse anti-TH) followed by the appropriate biotinylated secondary antibody (1:200; TH: anti-mouse IgG; ChAT: anti-goat IgG; calbindin and GFAP: anti-rabbit IgG; substance P: anti-rat IgG; Vector Laboratories, Burlingame, CA), the avidin-biotin complex (1:200; Vector Laboratories) and reacted with 0.5% 3,3'-diaminobenzidine (DAB) solution in 0.01% hydrogen peroxide. All retinas were pretreated with 5% Triton in PB for 48 hours at 4°C to increase antibody penetration, and incubated in 0.3% hydrogen peroxidase (H2O2) in methanol for 20 minutes to block endogenous peroxidase activity. Control and FGR material were stained simultaneously to avoid procedural variation. Control experiments were performed omitting the primary antibodies, whereupon staining did not to occur.
NADPH-d Histochemistry.
nNOS-IR amacrine cells also stain with nicotinamide adenine dinucleotide phosphate diaphorase (NADPH-d) histochemistry in sheep.13 This latter technique was used in the present study to identify ND1 and ND3 populations of nitrergic amacrine cells. The NI quadrant of the right retina was reacted for NADPH-d histochemistry, as previously described.13
Optic Nerve
Paraffin Processing.
A 5-mm piece of the left optic nerve (near-term cohort) was taken approximately 5-mm from the point of entry into the eye and embedded in paraffin, and serial transverse sections (8 µm) were cut. Every fifth section was stained with hematoxylin and eosin (H&E) for structural analysis. Second sets of sections were stained with Luxol fast blue (LFB; Merck, Darmstadt, Germany), to identify myelin. The remaining sections were used for immunohistochemical analysis.
Araldite Processing.
A 5-mm section of the right optic nerve (near-term and 2.3 year cohorts) was taken approximately 5 mm from the point of entry into the eye. The section was cut into small segments (1–2 mm in length), fixed in 1% glutaraldehyde in 4% PFA in 0.1 M PB (pH 7.4) for 48 hours, and embedded in Epon-Araldite. Semithin (1 µm) transverse sections of the optic nerve were cut and stained with methylene blue for quantitative analysis. Ultrathin sections (70 nm) were collected onto Butvar-coated slot grids stained with 2% uranyl acetate (10 minutes) and lead citrate (10 minutes) and examined under an electron microscope (model CM12; Philips, Eindhoven, The Netherlands) for qualitative assessment of ultrastructure.
Immunohistochemistry.
Immunoreactivity for mouse anti-2',3'-cyclic nucleotide 3'-phosphodiesterase (1:100, CNPase; Sigma-Aldrich), to identify myelinating oligodendrocytes,22 and rabbit anti-GFAP (1:1000) was localized on paraffin-embedded sections by using the avidin-biotin peroxidase complex (Vector Laboratories) as described earlier in the article. CNPase sections were pretreated with 0.02% proteinase K (Roche, Basel, Switzerland) for 30 minutes, to increase antibody penetration.
Quantitative Analysis
Retinal Morphology.
Retinal Areas.
A computerized digitizing program was used to measure the total area of each left retina in wholemount preparations after TH-IR and the total area of each right retina in wet-mounted preparations, before retinas were sectioned.
Retinal Thickness.
In methylene blue–stained semithin sections (1 µm) of TP and TC retina from both the near-term and 2.3-year cohorts, the mean thickness of (1) the total retina; (2) the ganglion cell layer (GCL); (3) the inner plexiform layer (IPL); (4) the inner nuclear layer (INL); (5) the outer nuclear layer (ONL); (6) the outer plexiform layer (OPL); (7) the total photoreceptor layer (PR); and (8) the inner (IS) and outer (OS) segments of the photoreceptors, were measured. Sections were projected (x600) and individual layers measured (5 sections/block, 10 measurements/section; 50 measurements/animal in total) using a computerized digitizing pad (Sigma Scan Pro, ver. 4.0; SPSS Science, Chicago, IL). The mean thickness of each layer was then calculated for each animal.
Ganglion Cell Somal Area.
Somal areas were assessed in semithin sections using an image-analysis system (x1300; Image Pro, ver. 4.1; Media Cybernetics, Frederick, MD). Ten cells were traced per section at randomly selected sites, from five sections per animal (50 cells/animal) and the measurements averaged for each animal.12
Cell Counts in the INL and ONL.
Counts of INL and ONL cells were made in 1-µm semithin methylene blue–stained sections at three locations in the peripheral and central retina for each animal at 140 dg and 2.3 years. Results were expressed as cells/mm2 for each region.
Vasculature.
Retinal blood vessels are clearly delineated in NADPH-d stained wholemounts. For each animal, the proportion of retina occupied by vessels was assessed using a point-counting technique23 in 50 randomly sampled regions per quadrant. Although it is acknowledged that this particular region (NI) may not be reflective of the overall proportion of blood vessels in the retina, it has been assumed that any changes to the proportion of vessels would not be solely restricted to one quadrant. Qualitative assessment was also performed in Araldite-embedded sections for the presence of neovascularization, specifically capillary sprouts composed of putative endothelial tubes surrounded by pericytes.24
Retinal Cell Neurochemistry.
Analysis of Immunohistochemistry.
The mean density of amacrine cell populations for each marker (ChAT-, substance P-, and TH-IR, and NADPH-d histochemistry) was determined with a computer assisted stereological tool system (CASTGRID, ver. 1.10; Olympus, Birkeroed, Denmark) set to sample 100 fields randomly (each 0.17 mm2, TH and substance P; or 0.01 mm2, ChAT and NADPH-d; larger fields sampled due to lower density of cells) per retina. The total number of cells in each population was then calculated from the mean density and retinal area measurements. The sampling method allowed density plots to be constructed25 for ChAT-IR (near-term cohort only, not assessed at 2.3 years, as no difference was observed at 140 dg, see the Results section) and TH-IR (2.3 year cohort only, not assessed at 140 dg, as results previously presented13 ). To analyze the somal area, for each cell class, 50 to 100 randomly selected somata were sampled throughout each retina using the CASTGRID system (x1000, oil immersion).12 The dendritic profile of TH-IR processes was assessed by tracing the total length of stained processes per cell (x500) and counting the number of TH-IR dendrites per soma (50–100 cells per retina; x2500, oil immersion).
Retinas from the TI quadrant of the right eye of one control and one FGR fetus were reacted for TH-IR to assess whether the immunohistochemical procedure resulted in tissue shrinkage. Retinal areas were measured before and after processing, and, as shrinkage was less than 0.5%, it was not taken into account when neuronal density or total number of cells was assessed.
Gliosis.
Qualitative assessment was performed in sections reacted for GFAP-IR to determine whether there was any difference in the extent or intensity of staining between control and FGR fetuses at 140 dg. From each animal, three sections of both peripheral and central retina were examined.
Synapses.
Qualitative assessment was performed in sections reacted for synaptophysin-IR, to determine whether there was any difference in the intensity of staining or evidence of ectopic location of synapses between control and FGR fetuses at 140 dg. From each animal, three sections of both peripheral and central retina were examined.
Optic Nerve Morphology and Neurochemistry
Myelinated Axons.
The following assessment was made in semithin (1 µm) methylene blue–stained sections of optic nerve: (1) The total cross-sectional area was measured with an image-analysis system (five sections per animal; x65); (2) the ratio of connective tissue to total cross-sectional area was calculated by using a point-counting technique (20 sites/animal; x1300); (3) the total nerve fascicle area per nerve was calculated by subtracting the percentage of connective tissue from the total cross-sectional area; and (4) the total number of axons in the optic nerve was calculated with an image-analysis system26 (x2500, sample area 0.001 mm2; 5% of each nerve was sampled). The total number of axons per animal was calculated as a product of total axons counted multiplied by the nerve fascicle area and divided by the sample area.26 The packing density (number of axons/per square millimeter)27 was also calculated; (5) axonal diameter (diam), nerve (axon + myelin sheath) diameter (Diam), myelin sheath thickness (Diam – diam/2), and G ratio (diam/Diam) were also calculated (x30,000; 200–250 axons/animal) by using an image-analysis system.
CNPase-IR Oligodendrocytes.
The density of myelinating oligodendrocytes (cells/mm2) was calculated in CNPase-IR sections of optic nerve (x2500 oil immersion; 40 sample points per animal), with the CASTGRID system.
|
Statistical Analysis |
|---|
|
TopAbstractMethodsStatistical AnalysisResultsDiscussionConclusionReferences |
|---|
|
Results |
|---|
|
TopAbstractMethodsStatistical AnalysisResultsDiscussionConclusionReferences |
|---|
2.3 years, the body and brain weight and the crown–rump length of the FGR sheep were not significantly different (P > 0.05) from those of control sheep.20
Eye Weight
There was no change in the eye weight (P > 0.05) between FGR and control animals at either 140 dg (FGR, 4.2 ± 0.6 g vs. control, 4.4 ± 0.2 g) or 2.3 years (FGR, 15.3 ± 1.1 g vs. control, 15.9 ± 1.0 g).
Retinal Morphology
Retinal Area.
There was no difference in the area of the right retina between FGR and control animals at either 140 dg (FGR, 743.4 ± 14.5 mm2 vs. control, 753.5 ± 16.2 mm2) or 2.3 years (FGR, 1477.1 ± 30.3 mm2 vs. control, 1498.0 ± 23.7 mm2).
Retinal Layers.
Near-Term Sheep.
No gross morphologic alterations were observed in the cytoarchitecture of the retina in FGR fetuses compared with control fetuses (Fig. 1C) . There was, however, a significant reduction in the total thickness of the central (Fig. 1A ; P < 0.05) and peripheral (Fig. 1B ; P < 0.05) retinas in FGR fetuses compared with the control. Specifically, in the central and peripheral retina, there were reductions in the thickness of the OPL (P < 0.05) and photoreceptor IS and OS layers. Furthermore, there were reductions (P < 0.05) in the INL (P < 0.05) in the central retina and the ONL (P < 0.05) in the peripheral retina.
|
Ganglion Cell Somal Areas.
There was no significant difference in the area of ganglion cell somal area between FGR and control groups at either 140 dg (FGR, 342.7 ± 11.6 µm2 vs. control, 370.2 ± 13.2 µm2; P > 0.05) or 2.3 years of age (FGR, 381.3 ± 19.1 µm2 vs. control, 381.5 ± 13.7 µm2; P > 0.05).
Cell Counts: ONL and INL.
At 140 dg there was no difference in the density of cells in the ONL (FGR, 21,273 ± 2,481 cells/mm2 vs. control, 21,760 ± 1,557 cells/mm2) or the INL (FGR, 11,775 ± 1,711 cells/mm2 vs. control, 12,270 ± 743 cells/mm2) in the peripheral retina. There was also no difference in cell densities in the ONL (FGR, 20,800 ± 1,350 cells/mm2 vs. control, 2,390 ± 901 cells/mm2) or the INL (FGR, 11,490 ± 259 cells/mm2 vs. control, 13,300 ± 931 cells/mm2) in the central retina.
At 2.3 years, there was no difference in the density of cells in the ONL (FGR, 16,310 ± 767 cells/mm2 vs. control, 18,590 ± 1,069 cells/mm2) or the INL (FGR, 7,757 ± 775 cells/mm2 vs. control, 9,525 ± 615 cells/mm2) in the peripheral retina. There was also no difference in cell densities in the ONL (FGR, 19,907 ± 1,148 cells/mm2 vs. control, 18,818 ± 9,813 cells/mm2) or the INL (FGR, 11,518 ± 856 cells/mm2 vs. control, 9,813 ± 922 cells/mm2) in the central retina.
Vasculature.
There was no difference in the proportion of retina occupied by blood vessels at 140 dg (FGR, 18.5% ± 1.6% vs. control, 18.5% ± 0.8%; P > 0.05) or 2.3 years (FGR, 18.9% ± 0.4% vs. control, 18.0% ± 0.0%; P > 0.05) of age in either FGR or control groups. Qualitative examination did not reveal any evidence of neovascularization.
Synaptophysin IR.
The IPL, OPL, and photoreceptor IS layers were strongly immunoreactive for synaptophysin in the ovine retina at 140 dg. Qualitative analysis did not reveal any difference in staining pattern or intensity of synaptophysin IR in FGR compared with the control fetuses.
Cell Populations and Neurochemistry
ChAT-IR Amacrine Cells.
Two populations of ChAT-IR cells were observed in the sheep retina at 140 dg. The first population had somata located in the INL and the second population in the GCL. Both populations are likely to be amacrine cells (Fig. 2A) , as reported in other species.28 29 Immunoreactive processes were observed stratifying at two levels in the IPL (Fig. 2A) . The total number (FGR, 250,212 ± 28,802 cells vs. control, 289,605 ± 24,119 cells; P > 0.05), density (FGR, 345 ± 45 cells/mm2 vs. control, 383 ± 4 cells/mm2; P > 0.05), and somal areas (FGR, 67 ± 4 µm versus control, 67 ± 6 µm; P > 0.05) of ChAT-IR amacrine cells were not different between FGR and control fetuses at 140 dg. ChAT-IR amacrine cells were distributed across the entire retina in both control and FGR fetuses. This population was not assessed in the adult, as no alterations were observed at 140 dg.
|
TH-IR Amacrine Cells.
TH-IR amacrine cells were previously found to be reduced in number in the near-term group after CPI in fetal sheep.13 A significant reduction (P < 0.05) in the total number and mean density of TH-IR amacrine cells was observed in FGR (Fig. 2D) compared with control (Fig. 2D) sheep at 2.3 years of age (Table 1) . TH-IR amacrine cells were distributed across the entire retina in both FGR (Fig. 2F) and control (Fig. 2C) sheep. The somal areas, number of dendrites per soma, and average dendritic length did not differ (P > 0.05) between the two groups.
|
GFAP-IR Glial Cells.
At 140 dg, GFAP-IR was found in the nerve fiber layer (NFL), closely associated with blood vessels (Fig. 3A) . Intense IR was also observed around vessels in the GCL and INL in both FGR and control fetuses. Qualitative examination revealed an upregulation of GFAP-IR throughout the entire retina and particularly in Müller cell processes in the IPL and OPL, at the outer limiting membrane (OLM), and in horizontal cells in FGR (Fig. 3B) fetuses compared with the control (Fig. 3A) .
|
|
|
Discussion |
|---|
|
TopAbstractMethodsStatistical AnalysisResultsDiscussionConclusionReferences |
|---|
Long-Term Effects of CPI on Dopaminergic and Other Subclasses of Amacrine Cells
In a prior study, dopaminergic amacrine cells were first observed in the fetal sheep retina at 72 dg13 and are thus well established at the onset of CPI. During CPI fetuses become hypoxemic, hypoglycemic, and mildly hypotensive30 and have an altered endocrine31 and growth factor32 status. These factors all could affect cell proliferation and survival or cause the downregulation of TH expression. In the present study, it was not possible to distinguish between these possibilities. Dopaminergic cells in general appear to be particularly vulnerable to hypoxic insults. In Parkinson’s disease, it has been suggested that cell loss may result from oxidative and nitrosative stress, mitochondrial dysfunction, excitotoxicity resulting from increased glutamate release, or the metabolism of dopamine itself, which may lead to the production of reactive oxygen species.33 34 In fetal sheep, it is known that hypoxia increases glutamate efflux in the central nervous system.35 This increase could trigger a cascade of events resulting in oxidative and nitrosative stress and the generation of reactive oxygen species,36 affecting neuronal survival and/or the expression of neurochemicals.
Reductions in the Thickness of the INL
There was no difference between groups in the density of cells in the INL at 140 dg or 2.3 years and also no difference in amacrine cell somal areas between groups. The reduced width of the INL therefore could result from a reduction in the total number of cells, with those surviving having a normal morphology and connectivity.
As dopaminergic amacrine cells comprise less than 0.1% of all amacrine cells, alterations in their numbers are unlikely to contribute significantly to the thickness of the INL. The INL contains the cell bodies of all retinal interneurons, raising the possibility that other populations of amacrine cells, or horizontal, bipolar, or Müller cells may be affected. It is known from this and a previous study13 that ChAT-IR, substance P-IR, and nNOS-positive amacrine cells are not affected by this level of hypoxemia. Therefore, CPI does not appear to have a global effect on amacrine cell populations. It was not possible to obtain total counts of other INL cell populations. It is possible that horizontal cells are affected, as they are reduced in number after CPI in the guinea pig.11 The loss of INL cells is likely to affect visual function significantly. In relation to the electroretinogram (ERG) a reduction in bipolar cells causes a decrease in the b-wave,37 whereas alterations to amacrine cells are more likely to cause alterations to oscillatory potentials.38 39 It was not feasible to record the ERG in the present study.
This study has shown for the first time that the neurotransmitter profile of the amacrine and horizontal cells in the ovine retina is similar to that of other mammalian species. ChAT stains two populations of cholinergic cells, one located in the INL and the other displaced to the GCL. Both populations are likely to be amacrine cells as reported in rabbits28 and rats.29 Substance P-IR is localized to populations of amacrine cells in the GCL and INL at 140 dg, as it is in the tree shrew40 and rabbit.41 Calbindin stains horizontal, ganglion, and amacrine cell populations in the near-term ovine retina as in the adult42 and several other species.43
In control animals, there was little net change in the total thickness of the retina between 140 dg and 2.3 years. During this period, nuclear layers thinned due to naturally occurring apoptosis (see Ref. 44 ) and the lateral expansion of the retina and photoreceptor OS increased in length. The total retinal area increased by 50% in the postnatal period.
Reductions in the Photoreceptor Outer Segment Layer Persist in the Long Term
Photoreceptors are vulnerable to the effects of CPI, with both IS and OS layers being reduced in length near term. At 2.3 years, the length of the IS in FGR animals was comparable to that in the control; however, the OS lengths remained significantly reduced. PRs have a high energy demand45 and are known to be vulnerable to damage before birth after cocaine exposure.46 They are also vulnerable to taurine deprivation47 and hypoxia in the neonatal rat.48 The maturation of photoreceptor OS and the formation of synapses are critical steps in the establishment of normal retinal formation and connectivity.49 50 As the OS lamellae are the site of phototransduction, abnormal development and maintenance of these structures may alter this process. A decrease in OS length could affect both the amplitude and timing of the a-wave of the ERG.51 52 53
It is surprising that CPI caused a permanent reduction in rod OS length, when lamellae are normally renewed constantly. We suggest that CPI may have a permanent effect on molecular or cellular mechanisms involved in OS regeneration. There is some evidence that the OS can be specifically compromised. For example, mesopic rearing of P23H rhodopsin transgenic rats leads to reductions in rod OS length and thinning of the ONL in association with a decrease in the amplitude of the a-wave of the ERG and increases in the levels of fibroblast growth factor (FGF) and GFAP.53 54 It has been suggested that ERG alterations may be a consequence of a decrease in functional OD membrane due to both photoreceptor cell death and OS shortening.54
It is possible that decreased levels of dopamine in the retina also contributes to the alterations in the PRs observed in our study, as dopamine acting via D4 receptors exerts a neuromodulatory effect on PR metabolism.55 56 It has been reported that dopaminergic amacrine cells9 and PRs46 are altered after prenatal cocaine exposure. Alterations to the dopaminergic system in the retina have also been linked with alterations in retinal function in retinitis pigmentosa–associated dystrophies57 58 and Parkinson’s disease.59
Synaptophysin labeled synapses in the IPL and the synaptic terminals of rods and cones in the OPL. No qualitative difference was observed in the intensity or staining pattern of synaptophysin between groups, indicating that there was no marked reduction in synaptic connectivity after CPI, although this form of assessment does not indicate the functional capacity of synapses.
CPI and Retinal Neovascularization
Retinal hypoxemia-ischemia is a central feature in diseases in which retinal neovascularization occurs, including ROP, diabetic retinopathy, central retinal vein occlusion, and ischemic retinopathies (see review in Ref. 60 ). Vascular endothelial growth factor (VEGF) is implicated as a stimulating factor61 in the development of experimental ROP. No evidence of neovascularization or alterations in the proportion of the retina occupied by blood vessels were observed after CPI. Investigation of VEGF levels after CPI may provide insight into the lack of neovascularization in this model of prenatal hypoxia.
CPI and Upregulation of GFAP-IR in Müller Cells
GFAP is known to be upregulated in various models of hypoxia-induced ischemia62 63 and retinal degeneration64 and in conjunction with hereditary retinal degenerations65 and is increased in Müller cell processes in the IPL, OPL, and OLM and in horizontal cells in FGR fetuses at 140 dg. Reactive gliosis may occur in an attempt to increase the supply of neurotrophic factors such as insulin-like growth factor (IGF)-166 and basic FGF,67 to support neuronal survival and assist in repair mechanisms in compromised tissue, or alternatively it may exacerbate damage. After retinal detachment, the growth of Müller cell processes into the subretinal space forms a fibrotic layer that inhibits the regeneration of rod OS.64 The presence of Müller cell processes at the OLM may also contribute to the reduction in rod OD length observed after CPI.
Parallels with Long-Term Effects of CPI in the Guinea Pig Retina
CPI in the guinea pig results in structural and neurochemical alterations similar to those observed after CPI in the sheep.68 Assessment of the ERG in this model revealed a reduction in the receptoral amplitude (RmP3), decreases in a-wave amplitude and implicit times, delays in rod mediated b-wave implicit times, and decreases in oscillatory potentials, indicative of both inner and outer retinal changes.68 It is not unreasonable to propose that similar patterns of functional deficits might be observed if the ERG were to be assessed in the sheep after CPI.
Persistence of Near-Term Alterations to the Optic Nerve into Adulthood
At 140 dg, there was no difference between control and FGR animals in ganglion cell somal area, the number of myelinated axons, or the number of CNPase-positive myelinating oligodendrocytes in the optic nerve. Alterations were observed in the growth of axons, with a proportional reduction in myelin sheath thickness in FGR animals. These changes did not persist at 2.3 years, which may be due to an initial delay in maturation of axons rather than a permanent alteration to optic nerve morphology. Myelination and axonal growth are influenced by the levels of thyroid hormone69 70 and brain-derived neurotrophic factor (BDNF).71 Plasma thyroid hormone levels72 and BDNF levels in the retina73 and hippocampus32 are reduced after CPI. It is therefore possible that a reduction in these factors occurs prenatally after CPI in the ovine fetus and contributes to the alterations in axon growth and consequently myelination at term. Although there were no structural alterations in optic nerve parameters at 2.3 years of age, it is possible that the delays observed prenatally cause irreversible alterations at the molecular level.
|
Conclusion |
|---|
|
TopAbstractMethodsStatistical AnalysisResultsDiscussionConclusionReferences |
|---|
|
Acknowledgements |
|---|
|
Footnotes |
|---|
Supported by the National Health and Medical Research Council of Australia.
Submitted for publication November 21, 2004; revised April 20 and May 24, 2005; accepted July 18, 2005.
Disclosure: M. Loeliger, None; J. Duncan, None; S. Louey, None; M. Cock, None; R. Harding, None; S. Rees, None
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.
Corresponding author: Michelle Loeliger, Department of Anatomy and Cell Biology, University of Melbourne, 3010, Victoria, Australia; m.loeliger{at}unimelb.edu.au.
|
References |
|---|
|
TopAbstractMethodsStatistical AnalysisResultsDiscussionConclusionReferences |
|---|
