Qualitative and Quantitative Morphologic Changes in the Vasculature and Extracellular Matrix of the Prelaminar Optic Nerve Head in Eyes with POAG

Abstract

Purpose.: To analyze the vasculature and extracellular matrix changes in the prelaminar region (PreLR) of the optic nerve head (ONH) and in the peripapillary sclera of eyes with primary open-angle glaucoma (POAG) and age-matched control eyes.

Methods.: In histologic sagittal sections of 46 eyes with POAG and 45 control eyes (donor ages, 20–96 years), the peripapillary sclera and penetrating vessels were investigated ultrastructurally and with antibodies against elastin, podocalyxin, and α-actin. Within the PreLR, the number and density of capillaries and the thickness of their connective tissue sheaths (CTSs) were quantified. The composition of the CTS was analyzed by using antibodies against collagen types I, III, IV, and VI, and elastin. Areas within the PreLR containing capillaries with thick or thin CTSs were determined.

Results.: There were no glaucomatous changes in the peripapillary elastic fibers and in the arterial capillaries in the periphery of the PreLR. In the center of the PreLR, the capillaries gained a CTS that was significantly thicker in POAG eyes than in control eyes, and the area containing capillaries with thick CTSs was significantly larger. These data did not correlate with axon counts.

Conclusions.: Lack of glaucomatous changes in elastic fibers of the scleral suspension of the ONH seems to prevent occlusion of penetrating vessels. In the PreLR, thickening of the capillary CTS and enlargement of the area containing capillaries with thickened CTS could increase diffusion time and may impair nutrition of the neuronal tissue.

Glaucomas are defined by a progressive optic nerve neuropathy. The pathogenesis of this neuropathy is still unknown. Experimental studies in the laser-induced monkey glaucoma model showed alterations of axonal transport in the prelaminar region (PreLR) of the optic nerve head (ONH) at the level of the lamina cribrosa that could eventually cause glaucomatous axon loss. ^1,2 In this model, no loss or closure of capillaries was observed in the PreLR. ³ Furthermore, the distribution and number of capillaries remained constant. ³ Thus, in the monkey glaucoma model, vascular changes were not primarily involved in the pathogenesis of axon loss. The case may be different in the ageing human eye. 

In human donor eyes with primary open-angle glaucoma (POAG) and in eyes with pseudoexfoliation glaucoma (PEXG), our group demonstrated a decrease in the number of capillaries within the postlaminar region of the optic nerve. ⁴ In contrast with eyes with PEXG, eyes with POAG also showed a decrease in capillary density (number per area). This decrease was accompanied by a significant increase in thickness of the surrounding connective tissue septae, which was also more pronounced in eyes with POAG than in eyes with PEXG. It is possible that thickening and densification of the septae caused an occlusion of the capillaries within, or that factors that induce the changes in the connective tissue could also have affected the capillaries directly. 

In the PreLR, there are no connective tissue septae. The bundles of nonmyelinated axons are surrounded by astrocytes, with the cell bodies forming defined glial columns. ^5–10 The capillaries supplying the PreLR are also located within these glial columns. Extracellular material is found only as a sheath surrounding the blood vessels. ^6,8  

All vessels supplying the PreLR derive from the short ciliary arteries and enter the optic nerve via the sclera or choroid, whereas the venous blood is drained by the central venous system. ^11,12 On their way, the arterioles and arterial capillaries have to pass through the scleral tissue and the elastic fibers circularly surrounding the ONH, which we term the elastic fiber ring (EFR). The elastic fibers in the sclera differ from those in the EFR. In the anterior parts of the sclera, the elastic fibers are surrounded by a sheath of fibrillar material similar to that found in the trabecular meshwork. ¹³ In the EFR, such fibrillar sheaths are missing. With age and in glaucomatous eyes, the elastic fibers of the EFR undergo elastosis and presumably loose elasticity. ^14–17 Age-related and glaucomatous changes in the posterior scleral elastic fibers and their fibrillar sheaths have not been evaluated so far. In the trabecular meshwork, the elastic fiber sheaths thicken with age, thereby presumably causing fiber stiffening. ¹⁸ Comparable changes in the peripapillary sclera could affect the vessels as they pass nearly perpendicularly through the sclera and EFR before entering the nerve. 

In the present study, we had the opportunity to investigate, qualitatively and quantitatively, the age-related and glaucomatous changes in (1) the peripapillary scleral elastic fibers and (2) the arterial vessels passing through this region, located in the periphery of the PreLR, as well as the venous vessels located centrally in a large number of human donor eyes with POAG and donor control eyes. 

Material and Methods

Donor Eyes

The experiments were performed in accordance with the provisions of the Declaration of Helsinki for research involving human tissue. 

Eighteen human eyes of 18 patients with POAG in different stages of the disease (Table 1) were obtained from the Mayo Clinic Eyebank (Rochester, MN; Douglas Johnson), 45 control eyes of 43 patients from the Department of Anatomy, University of Erlangen-Nuremberg and the Cornea Eye Bank (Amsterdam, The Netherlands). Informed consent was obtained from every donor. Immediately after enucleation, the globes were opened equatorially and immersed in Ito's fixative ¹⁹ or 4% paraformaldehyde (PFA). The ONH of each eye, including at least 2 mm of the surrounding sclera and retina were trephinated. The ONHs were divided sagittally into halves and either embedded in Epon, after postfixation with OsO₄ and dehydration, or in paraffin, after fixation with PFA. 

Table 1.

 

Results of the Quantitative Analyses of POAG Eyes from the Mayo Clinic (Douglas Johnson, MD)

Eyes ID	Age (y)	Sex	Axon Counts (n)	Capillary (n)	Capillary Density (n/mm2)	CTS Peripheral (μm)	CTS Central (μm)	Area with Thickened Capillaries (%)
1	78	M	1.047.810	59	105.73	1.71	3.64	83.95
2	71	M	871.170	69	204.75	1.46	3.46	83.72
3	58	M	790.281	NA	NA	1.81	3.73	51.46
4*	74	M	772.740	38	104.00	1.81	3.10	37.73
5	87	F	676.930	45	185.95	2.44	3.02	48.55
6	93	M	609.756	22	166.67	NA	4.49	92.96
7	92	F	396.240	31	106.53	1.36	3.87	33.80
8	91	M	376.623	78	191.65	3.63	3.28	43.17
9	86	F	342.375	48	292.68	1.92	4.74	77.62
10	81	M	317.858	38	107.81	1.84	3.48	37.88
11	90	F	102.400	27	78.26	3.24	4.38	34.50
12	75	M	49.704	32	118.08	1.64	3.75	44.57
13	84	F	35.658	28	205.88	NA	4.09	76.67
14	85	F	NA	30	146.34	3.51	3.50	59.51
15	75	F	NA	20	85.11	2.06	2.76	72.54
16	91	M	NA	14	125.00	1.57	3.64	62.42
17	82	F	NA	12	NA	NA	NA	NA
18	101	F	NA	4	NA	NA	NA	NA
Mean ± SD	83.00 ± 9.78		491.503 ± 316.154	35 ± 19	148 ± 56	2.14 ± 0.74	3.68 ± 0.52	58.81 ± 19.36

 The mean age of all eyes was 83 ± 10 years. There were no patients with cardiovascular disease or diabetes. Area with thickened capillaries is the area containing capillaries surrounded by a CTS thicker than 2 μm. There was no correlation between axon loss and capillary number and density.

*  Patient had asthma.

In addition, 24 donor eyes of 24 patients with POAG were received from Vaegan, PhD (Lidcombe Hospital, Sydney, Australia; Table 2). The specimens were sent to Erlangen as serial sagittal 7-μm sections through paraffin-embedded complete ONHs. The sections had been stained by the Picro-Mallory method. ²⁰ The collection and consent procedures had been approved by the Western Sydney Area Health Service. 

Table 2.

 

Results of the Quantitative Analyses of POAG Eyes from Lidcombe Hospital (Vaegan, PhD, and Shirley Sarks, MD)

Eyes ID	Age (y)	Sex	Capillary (n)	Capillary Density (n/mm2)
19	64	M	67	152
20	64	M	37	91
21	65	M	30	87
22	72	M	19	91
23	72	M	28	183
24	73	M	24	102
25	73	M	26	188
26	74	M	37	112
27	74	M	30	107
28	75	M	29	65
29	78	M	44	184
30	78	M	31	103
31	78	M	25	89
32	78	M	18	89
33	80	M	50	147
34	80	M	37	200
35	80	M	26	284
36	81	M	33	155
37	81	M	34	173
38	81	M	31	113
39	83	M	20	147
40	84	M	7	28
41	84	M	10	56
42	91	M	38	194
Mean ± SD	76.79 ± 6.44		31 ± 12	130 ± 56

 Mean age of all eyes, 78.2 ± 8 years.

All POAG donor eyes ranged between 58 and 101 years of age (mean age of all POAG eyes was 79.45 ± 8.6 years; Tables 1, 2 [see also Table 6]). The 45 control eyes ranged between 20 and 96 years. Eyes of ages 58 to 96 years (mean, 78.52 ± 10.84 years) were used as age-matched control eyes (Table 3 [see also Table 6]). 

Table 3.

 

Results of the Quantitative Analyses of Control Eyes

	Eye ID	Age (y)	Capillary (n)	Capillary Density (n/mm2)	CTS Peripheral (μm)	CTS Central (μm)	Area with Thickened Capillaries* (%)
	1	20	24	41.67	1.88	2.23	3.85
	2	37	26	86.96	2.03	1.87	21.41
	3	41	29	102.33	1.71	2.05	6.68
	4	43	58	NA	NA	NA	NA
	5	52	62	NA	NA	NA	NA
	6	54	37	89.71	1.63	2.13	3.77
	7	54	70	NA	NA	NA	NA
	8	58	42	282.44	1.72	1.83	5.71
	9	60	40	61.67	1.78	2.06	8.18
	10	64	31	300.00	2.00	3.00	25.59
	11	64	31	66.67	2.07	3.14	14.71
	12	67	59	NA	NA	NA	NA
	13	67	77	NA	NA	NA	NA
	14	67	52	NA	NA	NA	NA
	15	72	54	NA	NA	NA	NA
	16	74	23	118.06	1.82	2.26	22.22
	17	75	75	158.88	1.40	2.35	22.48
	18	75	63	121.09	1.96	2.40	2.60
	19	79	44	172.90	1.89	2.60	6.73
	20	79	41	NA	NA	NA	NA
	21	86	64	NA	NA	NA	NA
	22	86	74	NA	NA	NA	NA
	23	87	36	166.67	1.71	3.24	5.26
	24	88	30	118.18	2.58	2.96	64.80
	25	88	37	125.00	2.92	3.00	60.88
	26	89	25	72.73	2.24	2.33	5.10
	27	90	NA	NA	NA	NA	75.70
	28	96	46	186.81	2.22	2.74	74.79
	29	96	60	113.64	1.54	3.27	74.38
Young eyes < 58 years	Mean ± SD	44.88 ± 11.67	44 ± 17	121 ± 83	1.79 ± 0.14	2.02 ± 0.15	8.28 ± 6.66
Age-matched > 60 years	Mean ± SD	78.52 ± 10.84	48 ± 17	137 ± 61	2.01 ± 0.40	2.72 ± 0.39	33.10 ± 28.62

*  Area containing capillaries surrounded by a CTS thicker than 2 μm.

Histologic and Ultrastructural Analysis

From 18 POAG eyes and 29 controls fixed in Ito's fixative and embedded in Epon, 1-μm semithin sagittal sections were cut with a microtome (Ultracut E; Reichert Jung, Vienna, Austria) and subsequently stained with toluidine blue. Sections were viewed with an epifluorescence microscope (Aristoplan; Ernst Leitz, Wetzlar, Germany) and photographed (DC 500 camera; Leica Microsystems, Wetzlar, Germany). 

Ultrathin sections were stained with uranyl acetate and lead citrate and analyzed with an electron microscope (EM109; Carl Zeiss Meditec GmbH, Oberkochen, Germany). Ultrastructural analyses of the ONHs were performed in all POAG eyes received from the Mayo Clinic. 

Immunohistochemistry

Immunohistochemical stainings were performed on PFA-fixed or PFA-fixed, paraffin-embedded tissue of 16 control eyes (20–86 years). The antibodies used were reactive against glial fibrillary acidic protein (GFAP), α-smooth-muscle-actin (α-SMA), collagen types I, III, IV, and VI; podocalyxin; and elastin (Table 4) with affinity-purified rabbit and mouse monoclonal and polyclonal secondary antibodies according to the manufacturer's instructions. PFA-fixed ONHs were washed with phosphate-buffered saline (PBS, pH 7.4), placed in embedding medium (Tissue-Tek-OCT; Jung, Nussloch, Germany), frozen at −20°C, and serial sectioned (12 μm) with a cryotome (Kryostat CM 3050S; Leica, Bensheim, Germany). Paraffin-embedded tissue was sectioned in 7-μm sections with a microtome. The sections were placed on poly-l-lysine-coated slides, deparaffinized in xylol, and incubated with dry milk solution (Blotto; Santa Cruz Biotechnology, Heidelberg, Germany) for 1 hour at room temperature, to prevent nonspecific staining. Afterward, the sections were rinsed and incubated at room temperature in a moist chamber with primary antibody diluted in PBS buffer (PBS with 2% [wt/vol] bovine serum albumin [BSA; Merck, Darmstadt, Germany] and 0.2% [vol/vol] Triton-X-100) overnight (Table 4), except the α-SMA-antibody, which had been conjugated with the appropriate fluorescence (Cy3) and was incubated for only 1 hour. The sections were rinsed in PBS three times for 10 minutes each and then for 1 hour with the appropriate secondary antibody (Table 4), rinsed in PBS again, and mounted in Kaiser's glycerine jelly (Merck). 

Table 4.

 

Antibodies Used for Immunohistochemistry

	Dilution	Host	Manufacturer	Location
Primary antibody
Anti-collagen I	1:150	Rabbit	Rockland	Gilbertsville, PA
Anti-collagen III	1:150	Rabbit	Rockland	Gilbertsville, PA
Anti-collagen IV	1:100	Rabbit	Millipore	Billerica, MA
Anti-collagen VI	1:300	Rabbit	Rockland	Gilbertsville, PA
Anti-elastin	1:100	Mouse	Millipore	Billerica, MA
Anti-podocalyxin	1:400	Mouse	Donation of Dontscho Kerjaschki	Third Medical University, Vienna, Austria
Secondary antibody
Anti-rabbit Alexa 488	1:2000	Goat	MobiTec	Göttingen, Germany
Anti-mouse Cy3	1:2000	Goat	Dianova	Hamburg, Germany
Conjugated antibody
Alpha-SMA (Cy3)	1:400	Mouse	Sigma-Aldrich	St. Louis, MO

The sections were viewed with a fluorescence microscope (Aristoplan; Ernst Leitz, Wetzlar, Germany) and photographed (DC 500 camera; Leica, Wetzlar, Germany). Quantitative evaluations were performed with the accompanying software (Qwin; Leica). 

Determination of Optic Nerve Changes in POAG Donor Eyes

To determinate the stage of optic nerve damage, axon loss was evaluated in the eyes received from the Mayo Clinic. Therefore, optic nerves fixed in Ito's fixative were dissected 1 to 2 mm behind the globe and embedded in Epon. One-micrometer cross sections of the nerve were made and stained with toluidine blue. Axon counts were determined in accordance with our previous reports, ^4,21,22 based on the method of Quigley et al. ²³ Briefly, the area of the entire nerve, excluding the meninges, was measured. The area containing the nerve fiber bundles and the area of connective tissue between the nerve fiber bundles and surrounding the central retinal vessels were measured separately. For determination of the axon number, the cross section of the optic nerve was divided into eight sectors. In each sector, the axons were counted in five sample areas of 1000 μm². These sample areas extended from the central to the peripheral nerve at equal distance from one other. The mean of all measurements was multiplied by the nerve fiber area, to yield the total axon counts (Table 1). 

Morphology of the Vasculature

In five control eyes, serial sagittal histologic sections were obtained through the entire ONH and peripheral sclera with perforating vessels. The sections were stained immunohistochemically with antibodies against podocalyxin, a vascular endothelial marker, and α-SMA, to follow the course of the vessels and GFAP and elastin, to determine the relationship between astrocytes, elastin, and penetrating vessels. 

Quantitative Evaluations

Number and Density of Prelaminar Capillaries.

In midsagittal histologic sections of 42 POAG and 29 control eyes, the number and density of capillaries within the PreLR were assessed. Counts were performed in three neighboring sections, and the mean was used. 

The connection between the cut ends of Bruch's membrane (inter-Bruch's line [IBR]) was set as the inner border of the PreLR and the innermost layer of the lamina cribrosa (LC) as its posterior border (Fig. 1). In two eyes with severe excavation of the ONH, the vitreoretinal border was chosen as the inner border of the PreLR. 

Figure 1.

 

The definition of the anterior and posterior borderline of PreLR used for quantitative evaluations and the peripheral and central areas of the PreLR is shown. The anterior border was defined as a line through the connection between the cut ends of Bruch's membrane (inter-Bruch's line, IBR, arrows) and the posterior border by a line drawn through the anterior border of the lamina cribrosa. Peripheral and central areas, in which the CTS thickness of the capillaries was measured, were defined by vertical lines of 100 μm length.

Connective Tissue Sheaths Surrounding the Prelaminar Capillaries.

In 1-μm midsagittal semithin sections, the thickness of the capillary CTS was quantitated in peripheral and central prelaminar capillaries. The peripheral PreLR was defined by placing eight vertical 100-μm lines in equal distances to each other between IBR and the anterior border of the LC (Fig. 1). These lines were normal to fictitious tangents of the prelaminar lateral borderline between neural tissue and EFR. The central area was determined the same way by placing eight vertical lines 100-μm in length on fictitious lines on the border between neural tissue and the CTS of the central retinal vessels (Fig. 1). 

The thickness of the CTSs of capillaries in the different prelaminar regions was evaluated by measuring their diameters on micrographs (320× magnification). 

Areas Occupied by Capillaries with Thickened CTSs.

In 1-μm midsagittal semithin sections, areas containing capillaries with a CTS thicker than 2 μm were encircled and measured. These vessels were always found in the central portions of the nerve. The remaining peripheral areas containing capillaries with thin (<2 μm) CTSs were encircled also. As the area of the entire PreLR varied individually, a ratio between both areas was calculated. 

Thickness of Scleral Elastic Fibers and Their Sheaths.

The diameter of the core of the elastic fibers and their fiber sheaths was measured in ultrathin sections of six glaucomatous eyes and six control eyes,. For this purpose, micrographs were taken of cross sections through elastic fibers at a magnification of 7000× at distances of 100 to 200 μm and 1000 to 1200 μm lateral to the rim of the EFR. In these areas, 10 micrographs each were taken at equal distances ranging from the outer end of the choroid to the outer end of the sclera, and all cross-sectioned elastic fibers were measured. Measurements then were performed on computer with commercial software (analySIS software, ver. 3.1; Olympus Soft Imaging Solutions GmbH, Münster, Germany). 

Statistics

For comparison of the different groups, analysis of variance (ANOVA) was used. P < 0.05 denoted significance. For correlation analysis, the coefficient of correlation was determined and significance was set at α < 0.05. 

Results

Control Eyes

Elastic Fiber Ring and Entrance of the Prelaminar Vessels.

The morphologic analysis of serial sections through the ONHs supported previous findings ^11,12 that all arterioles and arterial capillaries enter the PreLR from the sclera and the choroid as branches of the short posterior ciliary arteries. No vessels emanating from the central retinal artery were found, and nearly all venous capillaries entered the central retinal vein. Double staining for elastin and podocalyxin as a marker for vascular endothelial cells or α-SMA as a marker for smooth muscle cells and pericytes demonstrated that elastic fibers that arose from the EFR were attached to the vessel wall of entering arterioles and accompanied them up to 200 μm into the PreLR (Figs. 2A, 2B). There was also intense staining in the EFR for the astrocyte marker GFAP. The staining for elastin was most intense in the EFR and weakest in the sclera (Fig. 2A). In the transition zone between the EFR and the sclera, intense staining for elastin and GFAP showed the same irregular lateral extensions (Fig. 2A). 

Figure 2.

 

(A) Sagittal section through the prelaminar region (PreLR) of the ONH of a control eye (age, 67 years), immunohistochemically stained for elastin (green) and for GFAP (red). The arterioles entering the prelaminar neural tissue are accompanied by elastic fibers up to 200 μm within the nerve (arrows). Within the EFR there was an intense staining for elastin and GFAP. In the transition zone (T) between the EFR and the sclera (S) are interdigitating extensions toward the sclera, which were always immunoreactive for elastin and GFAP. The distance between the rim of the PreLR and these peripheral extensions of the EFR was up to 200 μm. In the sclera (S) there was no staining for GFAP and staining for elastic fibers was sparse. (B) Shown are the rim of the prelaminar region (neural tissue, yellow), the EFR with its elastic fibers (green), and astrocytes (red) and the irregularly interdigitating transition zone (T) between the EFR and peripheral sclera (S). At places where the astrocytes are no longer present, the elastic fibers gain a thin fibrillar sheath (blue) but the core resembles that of the EFR. (C) Ultrathin section through the EFR of a 75-year-old control eye. Elastic fibers (arrows) are located in the vicinity of the astrocytes (*). Note that the elastic fibers have nearly no fibrillar sheath. (D) Ultrathin section through elastic fibers within the transition zone from the EFR to the sclera of the same eye. The elastic fibers (dark) located in this region gain a fine fibrillar sheath (arrows) where there are no astrocytes (*).

At the ultrastructural level, the elastic fibers of the EFR appeared homogenous and were surrounded by a very thin fibrillar sheath (Fig. 2C). Within the sclera adjacent to the EFR (100–200 μm distance to the outer border of the EFR), the morphology of the elastic fibers changed. Where astrocytes were still present, the adjacent elastic fibers had the appearance of those in the EFR. In regions where the elastic fibers were surrounded by collagen and scleral fibroblasts, they formed a sheath consisting of cross-linked fibrillar material embedded in an electron-light substance (Fig. 2D). The core of the fibers had the same appearance as in the EFR (Fig. 2D). 

The mean diameter of the elastic fibers in this region was 553 ± 147 nm (Table 5). The mean diameter of the fibrillar sheath was 102 ± 22 nm, which is 23% ± 5% of the whole fiber diameter (Table 5). Elastic fibers within the sclera at a distance of 1000 to 1200 μm from the outer EFR had the same appearance. In this region, the mean diameter of the fiber core was 568 ± 108 nm, the fibrillar sheath measured 123 ± 42 nm, which is 25% ± 10% of the whole fiber diameter (Table 5). 

Table 5.

 

Thickness of Elastic Fiber Core and Sheath in the Sclera at Distances from the Elastic Fiber Ring of (A) 100–200 and (B) 1000–1200 μm

	A		B
	Control	POAG	Control	POAG
Core	553 ± 147	492 ± 50	568 ± 108	551 ± 98
Sheath	102 ± 22	113 ± 18	123 ± 42	149 ± 47
Sheath/core ratio	0.23 ± 0.05	0.24 ± 0.02	0.25 ± 0.10	0.29 ± 0.05

 Data are expressed in nanometers ± SD.

In semi- and ultrathin sections through the region in which penetrating vessels were present, these vessels had open lumens, and no bending was observed. This finding was also true of the vessels in the peripheral part of the PreLR. As has been described before, ^6,8 the capillaries in the PreLR were located within regular organized glial columns of astrocyte cell bodies and were surrounded by a CTS. Electron microscopy showed that this CTS consisted of collagen fibers of different diameters embedded in an electron-light substance. 

Number and Density of Capillary Profiles in the PreLR.

The number of capillary profiles in the PreLR was the same in both the midsagittal and the more peripheral sections. The mean number of capillaries in the midsagittal sections through the PreLR was 48 ± 17 (Tables 3, 6). There was no age dependency of the capillary number, but there were large, interindividual variations. Capillary density in age-matched control eyes was 137 ± 61 capillaries/mm² (Tables 3, 6). Immunohistochemical investigation confirmed the results of Hernandez et al. ^8,10 that the CTS of the capillaries stained for collagen types I, III, IV, and VI. 

Table 6.

 

Results of All Eyes

	Age (y)	Capillaries (n)	Capillary Density (n/mm2)	CTS Peripheral (μm)	CTS Central (μm)	Area with Thick CTS (%)
Age-matched control eyes	78.52 ± 10.84	48 ± 17	137 ± 61	2.01 ± 0.40	2.72 ± 0.39	33.10 ± 28.62
All POAG eyes	79.45 ± 8.61	32 ± 16*	138 ± 57	—	—	—
POAG eyes
Mayo Clinic	83.00 ± 9.78	35 ± 19	148 ± 56	2.14 ± 0.74	3.68 ± 0.52*	58.81 ± 19.36*
Lidcombe	76.79 ± 6.44	31 ± 12	130 ± 56	—	—	—

 Data are the mean ± SD.

Connective Tissue Sheaths Surrounding the Prelaminar Capillaries.

In the age group from 20 to 54 years, the CTS of the capillaries was thin in the periphery and increased only slightly toward the central part of the PreLR (1.79 ± 0.14 and 2.02 ± 0.15 μm, respectively, P = 0.07; Table 3). In the age-matched control eyes the diameter of the CTS of the central capillaries was significantly larger than that of the peripheral capillaries of the same age group (2.72 ± 0.44 and 2.01 ± 0.40 μm, respectively, P = 0.0003; Tables 3, 6) and compared with the central capillaries of the younger age group (2.72 ± 0.44 and 2.02 ± 0.15 μm, respectively, P < 0.03, Table 3). The CTS of the peripheral capillaries did not change notably with age (2.01 ± 0.40 in age-matched control eyes compared with 1.79 ± 0.14 in the young eyes; Table 3). 

The age-related thickening of the sheath of the capillaries in the central part of the PreLR was especially due to an increase in the fine nonbanded collagenous fibers. Immunohistochemistry showed increased amounts of fibers staining for type VI collagen. 

Area Occupied by Capillaries with Thickened CTSs.

In the midsagittal sections, the area containing vessels with thickened sheaths was confined to the center of the PreLR. This area remained constant with age until an individual age of 88 years and increased in only five eyes of three very old donors (88, 90, and 96 years; Fig. 3). The mean area containing vessels with thickened sheaths in age-matched control eyes was 33% ± 29% (Tables 3, 6). 

Figure 3.

 

Age-dependent changes in the area of the PreLR containing capillaries with thick CTSs as a percentage of the entire PreLR nerve area (in percent). (♦) Control eyes; (♢) POAG eyes.

POAG Eyes

Elastic Fiber Ring and Penetration of Prelaminar Vessels.

As has been reported previously, ^14,16,17 the elastic fibers in the EFR in glaucomatous eyes showed elastosis characterized by irregularly shaped clumps of elastin. These elastotic changes were most pronounced in the ring portion directly adjacent to the nerve. 

In contrast, in the transition zone to the sclera, the elastic fibers did not show differences compared with those in the age matched controls. They consisted of a large homogeneous fiber core with a thickness of 492 ± 50 nm (Table 5). The fibrillar sheath was thin also in POAG eyes (113 ± 18 nm, Table 5). The sheath-to-core ratio was 24% ± 2%, comparable to the ratio in age-matched control eyes (23% ± 5%; Table 5). 

The same was true of the morphology of the elastic fibers within the sclera at a distance of 1000 to 1200 μm from the EFR. The mean thickness of the elastic fiber core was 551 ± 98 nm, the thickness of the sheath was 149 ± 47 nm, and the sheath-to-core ratio was 29% ± 5% (Table 5). These data did not differ from those in age-matched control eyes. 

Number and Density of Capillary Profiles in the PreLR.

The capillary number in the PreLR was diminished significantly in POAG eyes compared with that in age-matched control eyes (32 ± 16 and 48 ± 17, respectively; P = 0.0007, Table 6). However, capillary density in the PreLR did not change significantly in glaucomatous eyes compared with that in age-matched control eyes (138 ± 57 capillaries/mm² and 137 ± 61 capillaries/mm², respectively, Table 6). There was no correlation with axon loss. 

Connective Tissue Sheaths Surrounding the Prelaminar Capillaries.

In the periphery, the CTS of the capillaries was not significantly thicker in the glaucomatous eyes than in the age-matched control eyes (2.14 ± 0.74 and 2.01 ± 0.40 μm, respectively) (Tables 1, 3, 6; Fig. 4A). 

Figure 4.

 

(A, B) Sagittal sections (1 μm) through the prelaminar region of the optic nerve stained with toluidine blue (POAG, age 58 years). (A) Capillaries in the periphery of the prelaminar region with thin connective tissue sheath (CTS, arrows). (B) Central area of the prelaminar region containing capillaries with thick CTSs (arrows). (C) Ultrathin section through a capillary of the central area of the nerve with its sheath containing numerous fine fibrils (arrows) and basement membrane material (arrowheads) (POAG, age 86 years).

In the POAG eyes, there was a significant increase in the CTS thickness of the central capillaries compared with that of the age-matched control eyes (Fig. 4B). Ultrastructurally, there was an increase in nonbanded fine fibrils and of basement membrane–like material (Fig. 4C). The sheath thickness of central capillaries in glaucomatous eyes was 3.68 ± 0.52 μm compared with 2.72 ± 0.39 μm in the age-matched control eyes (P < 0.0001, Tables 1, 3, 6; Fig. 5). This thickening did not show a correlation with axon loss. 

Figure 5.

 

Age-dependent changes in the CTS thickness of the central prelaminar capillaries in control (♦) and POAG (♢) eyes.

Area Occupied by Capillaries with Thickened CTS.

Similar to the control eyes, vessels with thickened CTSs were found only in the center of the PreLR. However, the area occupied by vessels with thickened sheaths was significantly larger in glaucomatous eyes than in age-matched control eyes (59% ± 19% and 33% ± 29%, respectively, P < 0.001; Tables 1, 3, 6; Fig. 6). There was no correlation with axon loss (Table 1). 

Figure 6.

 

(A, B) An enlargement of the central area containing capillaries with thickened CTSs in control and glaucomatous eyes. (A) Control eyes: the area containing capillaries with thickened CTSs is demonstrated by including the lateral borders of those capillaries in a continuous line between the IBR and the inner border of the LC. (B) POAG eyes: The central area occupied by vessels with thickened sheath was significantly larger in POAG eyes than in age-matched control eyes (59% ± 19% and 33% ± 29%, respectively).

Discussion

All arterioles or arterial capillaries supplying the ONH have to pass through the sclera and the EFR to enter the nerve. The question arose whether differences in changes in the elastic fibers of the peripapillary sclera and the EFR could affect the vasculature. Our studies show that the elastic fibers of the peripapillary sclera are very similar to that of the EFR and do not undergo age-related or glaucomatous changes. They have a large homogeneous core, similar to that of the fibers in the EFR, and gain only a small additional fibrillar sheath within the sclera. This was true of scleral elastic fibers at distances of 100 to 200 and 1000 to 1200 μm from the outer EFR. The morphologic changes of the elastic fibers that were present in the aqueous humor outflow pathways in the anterior eye segment with increased thickness of a cross-linked fibrillar sheath were not seen in the posterior sclera. ¹³ The reason for these structural differences in scleral elastic fibers is not known. The differences in mechanical forces during accommodation could play a role. They do not explain, however, why profound glaucomatous changes occur in the trabecular meshwork elastic fibers and not in the elastic fibers of the posterior sclera. It is tempting to speculate that factors within the aqueous humor, presumably inducing formation of the sheath in the anterior eye segment, cannot reach the scleral fibroblasts in the peripapillary region. 

As the tissue of the EFR and presumably the peripheral sclera, too, physically buffers the backward movement of the ONH tissue and thus distributes these forces to the entire suspension region, the vessels seem to be protected from kinking by this specific transition between scleral elastic fibers and EFR. In fact, there were no occluded vessels in the peripheral PreLR, and even the CTS surrounding the capillaries in the peripheral PreLR remained normal in glaucomatous eyes. 

Within the PreLR with its lack of connective tissue septae, we found a loss of capillaries in glaucomatous eyes, but there was no loss of capillary density, indicating that a decrease in the number of capillaries was due to a loss of tissue (e.g., axon loss). We therefore assume that the decrease in capillary density in the postlaminar optic nerve described previously ⁴ was secondary to the increase in thickness of the connective tissue septae surrounding the capillaries. 

There was, however, also an increase in connective tissue in the PreLR of glaucomatous eyes. This increase was confined to the CTS surrounding the capillaries. In most normal older eyes, an increase in the CTS was seen only surrounding capillaries directly adjacent to the central retinal veins. In glaucomatous eyes approximately two thirds of the PreLR showed capillaries with thickened CTSs. Only the peripheral one third of the PreLR contained capillaries with thin sheaths. 

The reason that formation of thickened CTSs was found only in the central two thirds of the ONH in POAG eyes, but not in the periphery is not yet known. Recent studies on the pressure induced monkey glaucoma model have shown that mechanical stress due to IOP elevation can lead to detectable thinning of the PreLR. ^24,25 This prelaminar thinning was more pronounced within the central area of the PreLR than in the periphery. Thus, elevated IOP, at least in the monkey glaucoma model, induces pronounced changes in the central part of the PreLR. If this were also true of human glaucomatous eyes with elevated IOP, mechanical stress would be more pronounced in the central part of the PreLR than in the periphery. 

As previously described, mechanical stress can induce the expression of ECM components in human astrocytes and lamina cribrosa cells. ^26–29 It has also been shown, that treatment of cultured human astrocytes with the growth factor TGF-β2 induces expression of collagen type IV and VI ³⁰ and inhibits lysis of extracellular material through induction of the plasminogen inhibitor PAI. In previous studies, it was shown that in 50% of eyes with POAG, TGF-β2 levels are increased in the aqueous humor and the vitreous ^31,32 and that TGF-β2 levels are elevated in glaucomatous optic nerves. ³³ Thus, it is possible that TGF-β2 is one of the factors involved in the formation of thickened CTSs in glaucomatous eyes. In addition, it has been shown that hypoxia/reperfusion damage increases secretion of TGF-β2 in cultured human ONH astrocytes. ³⁴  

In vivo, in glaucomatous eyes, the thickening of CTSs may impair perfusion and diffusion of oxygen and nutritive substances from the capillaries to the surrounding tissue. If the Einstein-Smoluchowski diffusion equation is also true of the optic nerve capillaries, diffusion time increases quadratically with increasing thickness of the vessel walls. In addition, in the central area of the nerve, the oxygen tension is reduced, as there are only venous capillaries. Therefore, the driving potential for oxygen transport is reduced. For that reason, the increased distance due to thickened CTSs may influence diffusion and thereby nutrition, according to Fick's diffusion law. 

Thus, even if vascular changes in the optic nerve are not the primary cause of glaucomatous optic nerve neuropathy, the increase in thickness of the CTS and increase in the area containing capillaries with thickened CTS in the PreLR of glaucomatous eyes may result in a vicious circle that contributes to the progressive optic nerve changes in glaucoma. 

Footnotes

 Supported by the Deutsche Forschungsgemeinschaft Grant SFB 539 (EL-D).

Footnotes

 Disclosure: O.-Y. Tektas, None; E. Lütjen-Drecoll, Merck, Sharp & Dohme, GmbH (I); M. Scholz, None

The authors thank the cornea bank of Amsterdam and Hans Bloemendal (Department of Biochemistry, University of Nijmegen, Netherlands) for extensive endeavors in obtaining, fixing, and shipping the human eyes; to Shirley Sarks (Lidcombe Hospital) who collected the donor eyes; Dontscho Kerjaschke (Vienna, Australia) for the kind donation of the podocalyxin antibody; Anke Fischer, Elke Kretzschmar, Gerti Link, and Hong Nguyen for excellent assistance with immunohistochemistry and electron microscopy; and Marco Gösswein and Jörg Pekarsky for the preparation of the micrographs. The authors also acknowledge the invaluable assistance of the late Douglas Johnson, MD (Mayo Clinic), and the late Vaegan, PhD (Lidcombe Hospital), who kindly provided the glaucomatous eyes.