Chronic Placental Insufficiency Affects Retinal Development in the Guinea Pig

Abstract

purpose. Very low birth weight (VLBW) and fetal growth restriction are associated with increased risks of long-term visual impairments, including alterations to contrast sensitivity, a parameter mediated in part by dopaminergic amacrine cells. This study was conducted to determine whether chronic placental insufficiency (CPI), sufficient to cause growth restriction, results in neurochemical alterations to retinal interneurons, specifically amacrine and horizontal cell populations near term.

methods. CPI was induced just before midgestation (term ∼67 days of gestation, dg) in guinea pigs through unilateral ligation of the uterine artery. Growth-restricted (GR, n = 32) and control (n = 29) fetuses were euthanized at 60 dg and retinas prepared for analysis of amacrine cell populations by using antibodies to calbindin, calretinin, cholineacetyltransferase (ChAT), γ-amino-butyric acid (GABA), dopamine β-hydroxylase (DβH), tyrosine hydroxylase (TH, dopaminergic), and NADPH-diaphorase histochemistry (nitrergic). Calbindin immunoreactivity (IR) was also used to identify horizontal cells. HPLC was used to assess concentrations of catecholamines and Western blot analysis to detect total TH levels.

results. In GR compared with control fetuses the total number of TH-IR amacrine (P < 0.02) and calbindin-IR horizontal (P < 0.05) cells was reduced; however, there were no differences in the number of the ChAT, calbindin, calretinin, GABAergic, or nitrergic amacrine cell populations. HPLC revealed a reduction in the concentration of dopamine (P < 0.05) and noradrenaline (P < 0.05), and Western blot analysis revealed a reduction in TH in the retinas of GR compared with control fetuses (P < 0.05).

conclusions. CPI results in alterations to specific populations of retinal neurons. Such effects could contribute to visual impairments reported for VLBW children.

Improvements in obstetric and neonatal care have resulted in an increase in the survival of very low birth weight (<1500 g) infants. Although these infants are known to be at risk of impaired motor development, evidence suggests that they might also be vulnerable to impaired visual development. ^{1 2 3 4 5} Impairments in these infants include reduced visual acuity and refractive errors, but also more subtle changes such as abnormalities in color vision and contrast sensitivity. ^{2 4} Adverse prenatal conditions such as hypoxia–ischemia ⁶ may underlie some of these alterations to visual function, although currently little is known about their etiology. 

Chronic placental insufficiency (CPI) induced in the guinea pig by unilateral artery ligation mimics a form of chronic intrauterine compromise that could occur in human pregnancies (see Ref. ⁷ ). This procedure results in a reduction of nutrient and oxygen supply ⁸ to the fetuses in the ligated horn and can result in fetal growth restriction. The guinea pig is a particularly useful model of human retinal development, as it has a long gestation (term ∼67 days of gestation, dg), with most development occurring in utero. ⁹ CPI is induced just before midgestation (28–30 dg) during a critical period for retinal development. Neurogenesis and differentiation are under way, but the insult precedes the onset of synaptogenesis and the formation of the inner and outer segments of the photoreceptors. ⁹ A previous study of the retina near term in this model of CPI established that there was a reduction in the growth of the plexiform and cellular layers and in the number of substance P–immunoreactive (IR) amacrine cells. The number of ganglion cells was not affected. ¹⁰ The purpose of the present study was to determine whether CPI has a global effect on retinal interneurons—namely, amacrine and horizontal cells. As these cells play an important role in visual processing, compromise to their development during gestation could have significant consequences for visual function. 

Amacrine cells constitute the most diverse group of cell types within the retina with respect to morphology, size, and retinal coverage, comprising 30 to 40 distinct morphologic subtypes (see review in Ref. ¹¹ ). The inhibitory neurotransmitters γ-amino-butyric acid (GABA) or glycine are contained in most amacrine cells, and these amino acids are often colocalized with a second transmitter ^{12 11} such as dopamine, acetylcholine, nitric oxide, serotonin, substance P, vasoactive intestinal peptide (VIP), or somatostatin (see review in Ref. ¹³ ). Although the functions of most amacrine cells are unknown, AII glycinergic cells, which comprise 11% of all amacrine cells, ¹⁴ are the main carriers of rod signals to the ganglion cells. ^{15 16 17} The activity of substance P–IR cells raises the spontaneous activity level of ganglion cell responses. ¹⁸ Serotonergic cells influence the pathway through which dim light passes through the retina ¹⁹ and cholinergic cells enhance motion sensitivity through excitation of ganglion cells. ^{20 21} Dopaminergic amacrine cells, which are thought to play a role in contrast sensitivity, ²² were of particular interest to this study, as it has been demonstrated that they are reduced in number in the ovine fetal retina after CPI ²³ and in the rat after prenatal cocaine administration. ²⁴  

Specific amacrine cell populations investigated in the present study included: GABAergic, catecholaminergic (TH), cholinergic (ChAT), and nitrergic and calretinin- and calbindin-containing populations. Calretinin is used as a marker of AII glycinergic amacrine cells. ^{14 25 26} Calbindin labels all horizontal cells in the guinea pig retina, ²⁷ and a subpopulation of GABAergic amacrine cells. ²⁷ Nicotinamide adenine dinucleotide phosphate-diaphorase (NADPH-d) histochemistry colocalizes with neuronal nitric oxide synthase (nNOS)–containing amacrine cells in several species, including humans, ²⁸ rats, ²⁸ and sheep, ²³ and has been used in the present study to identify the ND1 and ND3 populations of nitrergic amacrine cells. In addition to identifying catecholaminergic cells immunohistochemically high performance liquid chromatography (HPLC) was used to assess the total concentration of catecholamines, and Western blot analysis was used to detect relative amounts of TH in the retina. 

Methods

Surgery

The study conformed to the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research and to the National Health and Medical Research Council of Australia (NH&MRC) code of practice for the care and use of animals for scientific purposes. Unilateral ligation of the maternal uterine artery ²⁹ was performed on pregnant guinea pigs (n = 34) at 28 to 30 dg (term ∼67 dg). Briefly, pregnant sows were anesthetized intramuscularly with xylazine (6 mg/kg; Troy Laboratories, Sydney, Australia) and ketamine (40 mg/kg; Troy Laboratories); a midline incision was made and the mesometrium of one uterine horn exposed, the uterine artery was then ligated near the cervical end of the arterial arcade. The ligation remained in place for the duration of the experiment. This procedure results in growth restriction in approximately 50% of fetuses in the ligated horn. Fetuses from the unligated horn served as the control. 

Preparation of Retinas

Fetuses were delivered by Caesarian section at 60.8 ± 0.2 dg after the mothers were deeply anesthetized with pentobarbitone sodium (130 mg/kg, intraperitoneally; Nembutal; Rhone Merieux, Pikenga, NSW, Australia). Fetuses and livers were weighed and crown–rump length was measured. Fetuses from the ligated horn of surgically delivered sows were classified as growth restricted (GR) if body weights were at least 2 SD below the mean for age-matched controls and brain-to-liver weight ratio was at least 2 SD above the mean for age-matched control fetuses. ¹⁰ Fetuses were immediately perfused through the left ventricle with 4% paraformaldehyde (PFA) in 0.1 M phosphate buffer (PB; pH 7.4), the eyes were removed, and the temporal aspect of each eye marked for orientation. After 1 hour of fixation, retinas were dissected from the eyecup and postfixed for 1 hour in fresh 4% PFA. 

A second cohort of control (n = 12) and GR (n = 6) fetuses were killed with an overdose of pentobarbitone at 60 dg, eyes were immediately removed, the retina dissected, weighed, snap frozen in liquid nitrogen, and stored at −70°C. The left eye was prepared for HPLC analysis to determine the concentration of catecholamines and the right eye for Western blot analyses to detect levels of TH and ChAT. 

Immunohistochemistry

Retinas from both eyes of control (n = 26) and GR (n = 23) fetuses were prepared as wholemounts and randomly allocated for anti-calbindin, -calretinin, -ChAT, -GABA, and -TH immunohistochemistry and NADPH-d histochemistry, so that eight control and at least six GR retinas were stained for each procedure. To increase antibody penetration, retinas for GABA-IR retinas were first flatmounted onto a chuck with optimal cutting temperature compound (OCT) and a 40-μm horizontal slice taken from the scleral side of the retina using a freezing microtome (Leitz, Wetzlar, Germany). Retinal sections were then washed in PB and reacted for immunohistochemistry, as described later. Cryostat sections (15 μm) of control retinas (n = 2) were also collected and reacted to localize dopamine β-hydroxylase (DβH)-IR cell populations. DβH is the enzyme involved in the conversion of dopamine to noradrenaline, and was therefore used to label noradrenergic neurons. 

Retinas were processed for immunoreactivity (IR) by using the avidin-biotin peroxidase complex (Vector Laboratories, Burlingame, CA), as previously described, ¹⁰ with the following antibodies at the following dilutions: rabbit anti-calbindin (Swant, Bellinzona, Switzerland) 1:5000; rabbit anti-calretinin (Swant) 1:10,000; goat anti-ChAT (Chemicon International, Temecula, CA) 1:1000; rabbit anti-DβH (Chemicon) 1:1000; rabbit anti-GABA (kindly donated by David Pow) 1:1000; and mouse anti-TH (Chemicon) 1:1000. Sections were incubated overnight (72 hours for mouse anti-TH) in the appropriate dilution of primary antibody, then incubated in the appropriate biotinylated secondary antibody (1:200; TH: anti-mouse IgG; ChAT: anti-goat IgG; calbindin, calretinin, and GABA: anti-rabbit IgG; Vector Laboratories) followed by incubation in the avidin-biotin complex (1:200, Vector Laboratories) and reacted with 0.5% 3,3′-diaminobenzidine (DAB) solution in 0.01% hydrogen peroxide. For each antibody, all control and GR retinas were reacted simultaneously to minimize procedural variation. Control experiments were performed by omitting the primary antibodies. In these experiments, staining failed to occur. 

NADPH-d Histochemistry

Retinas were washed in 0.1 M Tris buffer (pH 7.6) and then reacted for 45 minutes in the NADPH-d reaction solution (0.25 mg/mL nitroblue tetrazolium [Sigma-Aldrich, St. Louis, MO], 1 mg/mL β-NADPH [Roche, Mannheim, Germany], and 0.5% Triton X-100 [Sigma-Aldrich] in 0.1 M Tris buffer) at 37°C, washed in Tris buffer, and cover-slipped with aqueous mounting medium (Glycergel; Dako, Carpinteria, CA). All control and GR retinas were reacted simultaneously to minimize procedural variation. Control experiments were performed by omitting β-NADPH, whereupon staining failed to occur. 

Shrinkage

Retinas from the right eye of one control and one GR fetus were reacted for TH-IR to assess whether the immunohistochemical procedure resulted in shrinkage. Retinal areas were measured before and after the tissue was reacted and it was determined that shrinkage was less than 0.5%. Consequently, shrinkage was not taken into account when we assessed neuronal density or total cell number. 

Analysis of Retinal Neurons

The areal density of calbindin-IR, calretinin-IR, ChAT-IR, GABA-IR, TH-IR, and NADPH-d–positive cells was determined with a computer program (using the Computer Assisted Stereological Tool system; Castgrid ver. 1.10; Olympus, Birkeroed, Denmark) set to randomly sample 100 fields per retina (calbindin, calretinin, ChAT, and GABA, field area 0.01 mm²; TH-IR and NADPH-d, field area 0.04 mm², larger fields sampled due to lower density of cells). The total area of each retina was determined from a projected image of the retinal wholemounts using a computerized digitizing pad (Sigma Scan Pro, ver. 4.0; SPSS Science, Chicago, IL). Total numbers for each cell type were calculated from the mean density and retinal areal measurements. Density plots were constructed to illustrate the distribution of cells across the retina. To measure somal area, 50 to 100 randomly selected somata were sampled throughout each retina for each cell class (×2500 magnification, oil immersion; using the CASTGRID system), and the mean somal area was calculated. 

For TH-IR amacrine cells the number of dendrites per cell was counted (×1000 magnification) in 50 randomly selected cells per retina. To quantify the density of TH-IR varicosities, images were captured by computer (×3300, oil immersion; Image Pro, ver. 4.1; Media Cybernetics, Frederick, MD) focusing on sublamina 1 of the inner plexiform layer so that TH-positive varicosities were contained in a single focal plane. All labeled varicosities contained within a sample area of 3000 μm² were counted. ³⁰ For each retina, 20 regions were sampled randomly across the entire retina. 

High-Performance Liquid Chromatography

Each retina was homogenized individually in 100 μL of 0.1 M perchloric acid containing 0.01% EDTA. Homogenates were centrifuged at 13,000g for 5 minutes at 4°C and the supernatant removed. Dopamine, noradrenaline, and adrenaline were extracted from the supernatant with alumina adsorption, separated by HPLC, and the amounts quantified by coulometric detection as described previously. ³¹  

Western Blot Analysis

Each retina was homogenized individually in 50 μL of homogenizing buffer (50 mM Tris-HCl [pH 7.4]; 50 mM NaCl) containing a cocktail of protease inhibitors (10 mM EDTA; 10 mM aminocaproic acid; 0.25 mM phenylmethylsulfonyl fluoride (PMSF); 5 mM ethylmaleimide; 10 mM benzamidine; 5 U/mL aprotinin). Samples were separated on a 10% SDS-PAGE gel and transferred to a polyvinylidene difluoride (PVDF) membrane (Immobilon-P; Millipore, Bedford, MA) TH protein was detected using a chemiluminescence Western detection system (Lumiglo; Upstate Biotechnology, Lake Placid, NY) in combination with mouse anti-TH (1:500, 48 hours). The membrane was stripped (0.2 M glycine; 0.1% SDS; 1% Tween-20) at room temperature for 20 minutes and reprobed for ChAT-IR, as a comparative control, using goat anti-ChAT (1:500, 48 hours) as a primary antibody; ChAT protein was visualized using the avidin-biotin system as previously described ¹⁰ and developed with DAB. Optical density of immunoreactive bands for both TH and ChAT were analyzed using an image analysis system as described previously. ³²  

Statistical Analysis

All measurements were made on coded slides. Statistical analysis of all parameters was performed using t-tests; P < 0.05 was considered to be significant. Results are expressed as mean of means ± SEM for cell counts and mean ± SEM for HLPC and Western blot analyses. 

Results

Body and Brain Weights

At 60 dg, body weights (GR, 60.3 ± 0.3 g versus control, 95.6 ± 1.3 g; P < 0.001), brain (GR, 2.31 ± 0.07 g versus control, 2.53 ± 0.03 g; P < 0.005) and liver (GR, 2.73 ± 0.16 g versus control, 4.71 ± 0.20 g; P < 0.001) were reduced in GR fetuses compared with the control. In the GR group, brain sparing was evidenced by a 32% increase in brain-to-body weight ratio (P < 0.001) and a 41% increase in brain-to-liver weight (P < 0.001) ratios. There was no significant difference in eye weight (GR, 0.22 ± 0.01 g versus control 0.24 ± 0.01 g; P < 1.0) or retinal area (GR, 104.47 ± 2.1 mm² versus control, 107.9 ± 2.5 mm²; P < 0.7) between GR and control fetuses. However, there was a 29% increase in the eye-to-body weight ratio in GR fetuses compared with the control (P < 0.001). At postmortem, all guinea pigs had both eyes open (opening usually occurs at approximately 56 dg. ³³ ). Although the possibility exists that the cecal period (period to eye opening) may have been extended in GR fetuses, we were not in a position to test this parameter. 

Analysis of Amacrine and Horizontal Cell Populations

The total number and density of each class of cell are presented in Table 1 . The total number and density of GABA-IR, ChAT-IR, Calretinin-IR, and NADPH-d–positive cell populations were not different between GR and control fetuses (P < 0.05). Somal areas of horizontal and amacrine cells did not differ between GR and control fetuses (P > 0.05) for any population other than the TH-IR amacrine cells. 

GABA-IR.

GABA-IR stained amacrine cells in the inner nuclear layer (INL) and ganglion cell layer (GCL) and weakly labeled horizontal cells in the guinea pig at 60 dg. As the total population of horizontal cells was labeled with calbindin, GABAergic horizontal cells were not counted in this study. 

Cholineacetyltransferase-IR.

ChAT-IR labeled two populations of amacrine cells, one located in the INL and the other in the GCL, with immunoreactive processes stratifying in sublamina 2 and 4 of the inner plexiform layer (IPL). 

Calbindin-IR.

Calbindin-IR labeled ganglion cells, amacrine cells in the INL and GCL, and horizontal cells located at the scleral border of the INL. Immunoreactive processes were observed stratifying in sublamina 2 and 4 of the IPL at 60 dg. It is likely that the calbindin-IR amacrine cells were a subpopulation of ChAT-IR amacrine cells (Loeliger M, unpublished observations, 2003). The total number and density of calbindin-IR amacrine cells (including displaced amacrine cells in the ganglion cell layer) was not different between GR and control fetuses. There was a reduction in the total number (P < 0.02) and density (P < 0.05) of calbindin-IR horizontal cells in GR (Fig. 1B) compared with control fetuses (Fig. 1A) . As illustrated in the density-distribution plots, calbindin-IR amacrine cells (Figs. 1C 1D) and horizontal cells (Figs. 1E 1F) were distributed across the entire retina. The decrease in the density of horizontal cells in GR (Fig. 1F) compared with control fetuses (Fig. 1E) was also clearly evident. 

Tyrosine Hydroxylase–IR.

As for other species ³⁷ two populations of TH-IR amacrine cells were observed in the guinea pig: TI cells, which are located at the vitreal border of the INL; have large, intensely immunoreactive soma and an extensive dendritic network stratifying in sublamina 1 of the IPL; and are thought to be dopaminergic. ³⁷ TII cells are located in the INL and GCL and are smaller, less immunoreactive, do not have immunoreactive processes and are thought to be noradrenergic or adrenergic. ³⁷ The total number (P < 0.05) and density (P < 0.005) of TI TH-IR amacrine cells was lower in GR (Fig. 2B) than in control (Fig. 2A) retinas. The total number (P < 0.05) and density (P < 0.03) of TII TH-IR amacrine cells was also reduced in GR (Fig. 2B) compared with control fetuses (Fig. 2A) . Both populations of cells were distributed evenly across the retina. The decreased density of both cell populations was evident in density plots of GR (Figs. 2F 2H) versus control (Fig. 2E 2G) fetuses. The mean somal area of TI (GR, 91 ± 3 μm² versus control, 118 ± 6 μm²; P < 0.005) and TII (GR, 44 ± 4 μm² versus control, 57 ± 3 μm²; P < 0.05) amacrine cells was also significantly reduced in GR fetuses compared with the control. The number of TI TH-IR processes per soma (GR, 1.9 ± 0.1 versus control, 2.4 ± 0.1; P < 0.05) and the density of TH-IR varicosities were reduced (GR, 31,887 ± 2,729 varicosities/mm² versus control, 60,892 ± 3,414 varicosities/mm²; P < 0.05) in GR (Fig. 2D) compared with control (Fig. 2C) fetuses. The length and complexity of TI TH-IR–positive amacrine cell processes could not be analyzed because of the widespread overlap between the dendritic arbors of adjacent neurons. 

NADPH-d Positive.

NADPH-d histochemistry labeled two populations of nNOS–containing amacrine cells (ND1 and ND3). ND1 cells were located in the INL and had large, intensely stained somata and processes stratifying in sublamina 4 of the IPL. ND3 cells were small, were less intensely stained, and did not have stained dendritic processes. 

Calretinin-IR.

Calretinin-IR labeled amacrine cells, displaced amacrine cells, ganglion cells, and horizontal cells in the guinea pig retina at 60 dg. 

DβH-IR.

Cells in the GCL were strongly immunoreactive for anti-DβH, whereas a population of neurons, presumably amacrine cells, at the vitreal border of the INL were more weakly immunoreactive. Photoreceptor inner and outer segments were also immunoreactive for anti-DβH (not shown). 

High-Performance Liquid Chromatography

The concentrations of dopamine, noradrenaline, and adrenaline have not been determined in the retinas of near-term guinea pigs. The concentration of dopamine was comparable to that reported in the adult guinea pig retina. ³⁴ In GR compared with control retinas, there was a significant reduction in the concentrations of both dopamine (GR, 69 ± 65 ng/g retina versus control, 172 ± 65 ng/g retina; P < 0.01) and noradrenaline (GR, 50 ± 30 ng/g retina versus control, 87 ± 37 ng/g retina; P < 0.02); the reduction in adrenaline did not reach significance (GR, 5 ± 2 ng/g retina versus control, 11 ± 6 ng/g retina; P < 0.08). 

Western Blot Analysis

Optical density analysis of Western blots revealed a reduction in TH in GR fetuses compared with the control (GR, 0.083 ± 0.021 versus control, 0.144 ± 0.010; P < 0.05). There was no significant difference in optical density with ChAT-IR between GR and control fetuses (GR, 0.065 ± 0.009 versus control, 0.072 ± 0.008; P > 0.05). 

Discussion

In this study, CPI induced in the guinea pig just before midgestation did not have a global effect on retinal interneurons; rather, specific cell populations—namely the TH-IR amacrine cells and horizontal cells—were particularly vulnerable to the insult. Our findings extend a previous investigation of the retina in this model in which we found that substance P-IR amacrine cells, but not ganglion cells were reduced in number and all retinal layers were reduced in thickness. ¹⁰ Overall, we have now surveyed a large proportion of retinal neurons, particularly subpopulations of amacrine cells and have unequivocally shown a differential vulnerability of retinal cells to CPI. We are not yet certain how CPI affects bipolar and Müller cells. 

This regimen of CPI results in growth restricted fetuses that are chronically hypoxemic and hypoglycemic ⁸ and have an altered endocrine status. ³⁵ The insult is initially induced at a stage when neurogenesis is nearing completion, and programmed cell death is in progress. It extends throughout the period of synaptogenesis and photoreceptor development. ⁹ Axonal and dendritic growth and synaptogenesis appear to be compromised in this model, as we have previously demonstrated reductions in the plexiform layers and in the growth of the optic nerve. ¹⁰  

Currently, we cannot explain the basis for the vulnerability of TH-IR and substance P-IR amacrine cells and horizontal cells to CPI. It is of interest that TH-IR amacrine cells are reduced in another paradigm of intrauterine compromise induced during late gestation in fetal sheep, ²³ and that horizontal numbers are reduced in a rodent hyperoxia model of retinopathy of prematurity (ROP). ³⁶ It is generally considered that dopamine is the predominant catecholaminergic neurotransmitter in the retinas. ³⁷ However, noradrenaline has been detected with HPLC in fish, ³⁸ bovine, ³⁹ and rabbit ⁴⁰ retinas, and DβH has been detected immunohistochemically in human and monkey retinas ⁴¹ and in bioassays of rat retina. ⁴² Neither noradrenaline ³⁴ nor DβH ⁴³ has been detected in the adult guinea pig retina. In the present study, however, we found DβH-IR in retinal ganglion cells and significant levels of noradrenaline (with HPLC), suggesting that this catecholamine may be transiently expressed in the retina and play a role during development. The reduced levels of dopamine and noradrenaline in GR compared with control fetuses, together with the reduction in TH levels assayed with Western blot analysis is further evidence that catecholaminergic neurons are affected in CPI. In contrast, the finding that ChAT levels were unaltered confirms that CPI did not affect cholinergic cells. It is not possible to determine unequivocally whether the reduction in the number of TH-IR amacrine cells indicates that some of the cells have died or that TH expression is downregulated. Either possibility could cause the reduction in dopamine and noradrenaline levels reported in this study. Our previous findings that the number of TH-IR cells is also reduced in the adolescent guinea pig after an identical intrauterine insult ⁴⁴ argues in favor of the death of a proportion of these cells during development. 

Adverse effects on dopaminergic amacrine cells are of particular interest, because alterations to the dopaminergic system are thought to have an effect on contrast sensitivity. ²² In prenatally compromised adolescent guinea pigs ⁴⁴ we have demonstrated alterations in retinal function as detected by the electroretinogram (ERG). Of particular interest was the reduction and delay in oscillatory potentials (OPs). Although the origins of OPs remain equivocal, amacrine cells and dopamine may be involved. ²²  

The vulnerability of dopaminergic neurons has been well established in Parkinson’s disease. Although the precise etiology of cell death in this disorder is unknown, possible mechanisms include oxidative and nitrosative stress–mitochondrial dysfunction, and excitotoxicity resulting from increased glutamate release. ⁴⁵ In the present study, the susceptibility of cell populations to damage does not appear to be related to neurochemical status at the time of the insult, as all neurochemicals examined in the present study are expressed at 30 dg (Loeliger M, unpublished observations, 2003). Neither does it appear to relate to their GABA or glycine content as, although both catecholaminergic amacrine cells and horizontal cells contain GABA, other GABAergic populations including the ChAT-IR and nNOS-containing cells are not affected. Further studies are necessary to elucidate fully the mechanisms underlying this selective vulnerability. 

In conclusion, we have shown that CPI differentially affects retinal cells with reductions in the number of horizontal cells and of catecholaminergic amacrine cells. These results, together with the previous findings of a reduction in substance P-IR amacrine cells and the growth of all retinal layers, ¹⁰ indicate that process growth, synaptogenesis, and photoreceptor structure are also likely to be affected. These alterations could have significant effects on vision postnatally as we have already demonstrated alterations to the electroretinogram after CPI in adolescent guinea pigs. ⁴⁴  

 Supported by the National Health and Medical Research Council of Australia.

 Submitted for publication December 13, 2003; revised February 17, 2004; accepted February 26, 2004.

 Disclosure: M. Loeliger, None; T. Briscoe, None; G. Lambert, None; J. Caddy, None; A. Rehn, None; S. Dieni, 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.

 

Table 1.

 

Total Number and Density of Amacrine Cell and Horizontal Cells in Control and Growth-Restricted Fetuses

Cell Class	Retinal Area (mm2)			Total Number			Density (cells/mm2)
		Control	GR	Control	GR	Control	GR
GABA-IR amacrine cells	111.7 ± 4.2	102.3 ± 3.8	315,735 ± 15,496	311,900 ± 30,842	2,837 ± 138	3,044 ± 268
ChAT-IR amacrine cells	119.1 ± 5.6	107.3 ± 3.7	119,583 ± 7,502	104,281 ± 6,000	993 ± 51	979 ± 58
Calbindin-IR amacrine cells	113.0 ± 3.0	109.0 ± 3.0	51,264 ± 4,513	51,184 ± 8,237	454 ± 42	462 ± 73
Calbindin-IR horizontal cells	113.0 ± 3.0	109.0 ± 3.0	141,546 ± 10,341	97,497 ± 10,353*	1,249 ± 83	907 ± 104*
TH-IR, TI amacrine cells	99.9 ± 5.5	95.0 ± 1.9	3,690 ± 235	2,627 ± 285*	37 ± 3	28 ± 2*
TH-IR, TII amacrine cells	99.9 ± 5.5	95.0 ± 1.9	7,875 ± 648	4,931 ± 927, **	79 ± 7	51 ± 9*
NADPH-d positive, ND1 amacrine cells	98.9 ± 4.8	99.7 ± 4.3	2,681 ± 165	2,916 ± 370	27 ± 1	29 ± 3
NADPH-d positive, ND3 amacrine cells	98.9 ± 4.8	99.7 ± 4.3	5,258 ± 782	4,386 ± 505	55 ± 10	45 ± 5
Calretinin-IR amacrine cells	108.9 ± 5.9	101.7 ± 4.9	60,665 ± 11,072	53,484 ± 5,731	574 ± 124	524 ± 49

 Data are mean of mean ± SEMs;

*   P < 0.05;

**   P < 0.005. n = 8 for all control groups; n = 6 for TH and nNOS groups, and n = 8 for all other GR groups.

Figure 1.

 

The density of calbindin-IR horizontal cells (large arrow) was reduced in GR (B) compared with control fetuses (A). Density plots of calbindin-IR amacrine cells from control (C) and GR (D) and calbindin-IR horizontal cells from control (E) and GR (F) fetuses clearly illustrate the decrease in density of horizontal cells but not amacrine cell populations in GR compared with control fetuses. Scale bar: (A, B) 25 μm; (C–F) 160 mm.

Figure 2.

 

Retinal wholemounts from control (A) and GR (B) guinea pigs at 60 dg, stained for TH-IR. (C, D) The density of both TI (large arrow) and TII (small arrow) TH-IR amacrine cells was reduced in GR (B) compared with control fetuses (A). High-power views of control (C) and GR (D) retina stained for TH-IR, the reduction in process growth and varicosities was evident in GR animals. Density plots of TI and TH-IR amacrine cells from control (E) and GR (F) fetuses, TII and TH-IR amacrine cells from control (G) and GR (H) fetuses, illustrating the decrease in density of both TH-IR amacrine cell populations in GR fetuses compared with controls. Scale bar: (A, B) 110 μm; (C, D) 40 μm; (E–H) 160 mm.

The authors thank Mary Tolcos and Carina Mallard for generation of some of the GR guinea pigs used in this study, Ursula Grefrath for advice and assistance with GABA-IR, and Erica Fletcher for reading the manuscript.