Otago Glaucoma Surgery Outcome Study: Tissue Matrix Breakdown by Apoptotic Cells in Capsules Surrounding Molteno Implants

Abstract

purpose. To identify cell types and extracellular matrix components in Molteno implant capsules.

methods. Histologic features including cytology, distribution of apoptotic cells, cytoantigens, collagens, basement membranes, elastic fibers, and glycoproteins were examined by light microscopy. Findings were correlated with the clinical features of 19 ocular specimens with glaucoma that had been treated with Molteno implants 11 days to 20 years previously.

results. All but the earliest specimen capsules showed two layers: a moderately cellular outer layer of normally stained collagen and an inner avascular hypocellular layer of altered collagen. Capsules contained metabolically active fibroblasts and macrophages showing swelling, vacuolation, and apoptosis with localized loss of extracellular matrix in the inner layers of older capsules. Type I collagen was present in trace amounts. Collagens types III and VI and fibronectin were present in high concentrations in the capsules. Basement membrane material (collagen type IV and laminin) and thrombospondin were concentrated in the inner avascular layers.

conclusions. These results support previous findings that the normal capsule life cycle includes continual inner surface degeneration and external surface renewal. The cells and tissue matrix components of the outer capsule layer matched those involved in the initial phase of wound healing in vascular connective tissue. The tissue matrix components and widespread apoptosis found in the inner fibrodegenerative layer reflect scar tissue remodelling induced by exposure to the aqueous.

The success of aqueous shunt surgery for glaucoma depends on the formation and maintenance of a thin permeable capsule. Histologic and ultrastructural studies of capsules from experimental animals and human cases have been reported. ^{1 2} However, interpretation of the findings was initially complicated by severe damage to many of the human eyes before drainage by the implants. ^{1 2 3 4 5} With additional cases and longer follow-up in less damaged human eyes a consistent histologic pattern was found, ^{2 5 6 7} consisting of a moderately cellular fibrous capsule with a layer of small blood vessels on its outer surface which merges into the overlying subconjunctival connective tissue. The inner layers of the capsule are avascular and show a decreased number of more or less degenerate cells together with swelling and fragmentation of the collagen fibers. ⁵  

Electron microscopic studies demonstrated cells resembling fibroblasts, myofibroblasts, and macrophages with loosening and separation of fibers toward the inner surface of the capsule. In addition in some cases an incomplete layer of fibroblast-like cells lined the inner surface. ^{3 6 7}  

A histopathologic study of capsules from 75 eyes demonstrated that capsules around functioning Molteno implants evolved through a series of histologic stages. Without aqueous flow, the episcleral plate of the implant stimulated encapsulation by a thin avascular collagenous layer. With aqueous flow, an immediate inflammatory reaction developed in the episcleral connective tissues that included collagenous and vascular components. After a variable delay, a fibrodegenerative process developed in the inner layers of the capsule. ⁸ Cytological and immunohistochemical studies of capsules demonstrated that the fibrodegenerative process was characterized by widespread metabolic activation and/or apoptosis and concluded that the balance between activation and apoptosis regulates the thickness and permeability of capsules and that the normal life cycle of capsules included continual inner surface degeneration and external surface renewal. ⁹  

This communication reports the findings of additional histologic and immunohistochemical examination of capsules surrounding Molteno implants including types of cell with their patterns of activation, detailed cytology of cells undergoing apoptosis in the inner layer of capsules, and distribution of extracellular matrix components. 

Methods

Ocular Specimens

Nineteen postmortem ocular specimens with capsules and overlying connective tissue were examined histologically. Additional immunohistochemical staining was performed on 16 of the specimens. Cell surface antigens were stained in six specimens. Ground substance components and tissue fibrils were stained in six further specimens. All eyes had a Molteno implant inserted from 11 days to 20 years previously, with subsequent good IOP control in all cases, except one of neovascular glaucoma, which had a final IOP of 28 mm Hg (Table 1 , case 7). Details of the ocular specimens are shown in Table 1 . 

Informed consent for donation of eyes for research purposes was obtained from the patients before their death. The study adhered to the tenets of the Declaration of Helsinki. 

Surgical Technique

The surgical techniques used for Molteno implant insertion with immediate and delayed aqueous drainage have been described. ^{10 11 12 13 14 15}  

Fixation

Enucleation of all eyes occurred between 1 and 4 hours after death. Thirteen eyes were injected with formol saline (a 10% solution of 37% formaldehyde in phosphate buffer in saline) through a 30-gauge needle inserted across the limbus, to distend the capsules, and were then placed in formol saline for 3 hours, microwaved for 20 minutes at 50°C, and placed in 70% alcohol to further harden. The lateral half of each capsule and adjacent tissue was excised to allow removal of the episcleral plate of the implant before standard paraffin processing. 

The excised half of the capsule was separately embedded and oriented to allow for the cutting of 5-μm serial sections. Initial sections were cut tangential to the surface of the capsule at its apex. Subsequent sections parallel to this initial plane became more oblique as the plane of section extended to the capsule margin where they were almost perpendicular to the wall of the capsule. The half capsule remaining on the eye was oriented to permit the cutting of vertical sections. After paraffin embedding, 5-μm serial sections were cut and mounted on glass slides. 

Six eyes had half the capsule removed and immediately frozen in liquid nitrogen before the remaining tissue was fixed as just described. Frozen 15-μm sections were cut and mounted on glass slides. 

Histologic and Immunohistochemical Tissue Staining

Sections for histologic staining were stained with hematoxylin and eosin, Gomori trichrome, and Alcian blue. Immunohistochemical staining was performed to identify cells, ground substance, and tissue fibrils. The cells were stained to identify expression of TdT-mediated dUTP nick-end labeling (TUNEL), proliferating cell nuclear antigen (PCNA) as an indicator of metabolic activity, CD34 (angiogenic progenitor cells from bone marrow), CD45 (leukocyte common antigen), CD68 (macrophages), actin (α-smooth muscle), desmin, and vimentin. Tissues were stained for the extracellular components collagen types I, III, IV, and VI and elastin, fibronectin, laminin, thrombospondin, and vitronectin. 

TUNEL Staining.

TUNEL is used for in situ labeling of DNA strand breaks that form in individual nuclei of apoptotic cells. ¹⁶ TUNEL is sensitive and fairly specific for apoptosis although the reactivity lasts only for ≈4 hours in an apoptotic cell. Fluorescein-dUTP with appropriate filters was used to photograph capsule wall cell labeling. Sections were dewaxed in xylene and rehydrated through a graded series of ethanols, after which antigen unmasking was performed by incubating sections in 20 μg/mL proteinase K (Sigma-Aldrich) for 20 minutes followed by permeabilization for 2 minutes at 4°C in a solution consisting of 0.1% (vol/vol) Triton-X-100 and 0.1% (vol/vol) trisodium citrate. The TUNEL label (fluorescein-dUTP and dNTP mix; Roche, Indianapolis, IN) was combined with 10% (vol/vol) TUNEL enzyme (terminal deoxynucleotidyl transferase; Roche) and applied to each section for 60 minutes. Sections were mounted with Vectashield (Vector Laboratories, Burlingame, CA) and cover slipped. Imaging was performed on a laser scanning confocal microscope (510 Axioplan 2; Carl Zeiss Meditec, Dublin, CA). The fluorescein label was excited at 488 nm with an argon laser, and the emission wavelength was collected by using a band-pass filter in the range of 505 to 530 nm. A negative control was prepared on a single section within the assay by excluding the TUNEL enzyme and counterstaining the sections with DAPI (4′-6-diamidino-2-phenylindole) which is a blue nuclear stain. A positive control section was assessed by applying 1 U/mL DNase 1 (Promega) and incubating for 30 minutes at 37°C before the addition of the TUNEL enzyme/label mix. ^{17 18}  

PCNA, CD34, CD45, CD68, Actin, Desmin, and Vimentin Immunohistochemistry.

Sections from six capsules (Table 1)were dewaxed in xylene and rehydrated through a graded series of ethanol, then placed in 3% hydrogen peroxide for 10 minutes and rinsed in distilled water. Heat retrieval of antigen was used to stain for CD68, actin, and desmin by autoclaving the specimen in citrate buffer (pH 6) for 20 minutes to 121°C. Enzyme retrieval from sections stained for CD34 and vimentin was undertaken by covering the section with proteinase-K (S3030; Dako, Glostrup, Denmark) for 10 minutes, but was not used for sections stained for CD45. Sections were then placed in Tris buffer. A protein blocking agent (normal human serum) was added for 10 minutes and then tapped off. Antisera appropriate to each immunohistochemical marker was added (CD34 NCL-END, NovaCastra Labs, Newcastle-upon-Tyne, UK; CD45, M0701, Dako; CD68, M0876, Dako; actin, M0851, Dako; desmin, M0780, Dako; and vimentin, M7020, Dako) and the marker visualized with LSAB2-HRP (Dako) and diaminobenzidine chromogen (K3466; Dako), counterstained with Meyer’s hematoxylin and mounted (Depex medium; BDH Laboratory Supplies, Poole, UK). Negative controls were provided by omitting the specific antibodies and substituting normal serum in the same dilution. Positive controls consisted of human gastric mucosa, skin, lymph node, liver, granulation tissue, and ciliary body. 

PCNA-stained sections were exposed to antigen unmasking by heating them to boiling in 0.01 M citrate buffer (pH 6) for 20 minutes in a 1000-W microwave oven before allowing them to cool in solution for 15 minutes. Nonspecific binding was blocked with 5% (vol/vol) goat serum before anti-PCNA (NovaCastra Labs) was applied at a 1:150 dilution and incubated overnight at 4°C. Biotinylated anti-rabbit IgG was applied and amplified with streptavidin-biotinylated horseradish peroxidase complex. Signals were developed for visualization with amino ethyl carbazole (AEC; Sigma-Aldrich), counterstained with Gills hematoxylin, and mounted with glycerol. Negative control experiments were performed by omitting the primary antibody from the dilution buffer, and positive controls included human accessory lachrymal gland. 

Collagen and Glycoprotein Immunohistochemistry.

Frozen sections of six capsules (Table 1)embedded in OCT compound (ProTechSci, Queensland, Australia) were stained by using a four-step, enhanced avidin-biotin-peroxidase complex technique. ¹⁹ Sections were thawed at room temperature for 30 minutes, fixed in acetone at 4°C for 5 minutes, and washed with phosphate-buffered saline (PBS) for 20 minutes. They were then incubated in hydrogen peroxide 0.5% and washed in running water for 10 minutes. A blocking serum mixture of equal parts of 1:10 normal swine serum and 2% chicken albumen (A5503; Sigma-Aldrich) was then applied to the slides and left for 20 minutes. The blocking serum was removed, the primary antibodies (in 2% albumin in PBS) added for 60 minutes, and sections washed in PBS for 5 minutes. Sections were incubated with 1:100 biotinylated swine antibody (E354; Dako) for 30 minutes, washed in PBS for 5 minutes, incubated with 1:500 peroxidase-conjugated streptavidin (P397; Dako) for 30 minutes, and washed successively in PBS and distilled water, for 5 minutes each. AEC was then added for 10 minutes at room temperature to demonstrate peroxidase, and the sections were finally rinsed and mounted in glycerol. 

Negative control experiments consisted of sections processed without exposure to the primary antibody and positive experiments consisted of sections of human tonsil. 

Slide Examination

Stained sections were examined and photographed using bright-field, dark-background, polarized-light, phase-contrast, and fluorescent microscopy techniques (Orthoplan with Zernike and Heine phase-contrast condensers; Leitz, Wetzlar, Germany; laser scanning confocal microscope, 510 Axioplan 2; Carl Zeiss Meditec). 

Apoptotic cells were identified by their morphology in oblique sections stained by hematoxylin and eosin. The criteria used for their identification were those of Kerr et al. ²⁰ The first stage involved marked condensation of nucleus and cytoplasm, nuclear fragmentation, and separation of protuberances that formed on the cell surfaces,while the second stage involved phagocytosis by adjacent cells and the formation of lucent cytoplasmic vacuoles and dense masses of nuclear material in some cases, whereas others were composed only of condensed cytoplasmic elements. Apoptotic bodies are described as frequently occurring clusters in the intercellular space from which the smaller bodies tend to disperse from their site of origin. ²⁰  

Cell density and the proportion of apoptotic cells in the outer layers of the capsule were determined by examining obliquely cut sections with graticule having 100 33 × 33-μm squares and placing it over the transition between the normally stained outer and poorly stained inner layers of the capsule. The cells were counted in 66 × 330-μm strips using a 40× objective and apoptotic cells identified. These cells were then examined individually with a 100× oil immersion objective. 

Adjacent 66 × 330-μm strips were examined until a total of >750 cells had been counted manually (this involved between 15 and 23 fields) after which the proportion of apoptotic cells was calculated. 

Results

Histologic Structure of Capsules

Structure of the Early Capsule.

At 11 days after surgery, the conjunctiva and Tenon’s tissue overlying the capsule showed dilated blood vessels containing abundant red cells, leukocyte margination, and accumulation of cells in perivascular spaces. This tissue merged into the outer capsule layer that contained dilated capillaries and a zone of moderate numbers of reactive fibroblasts and increased ground substance. In the inner layer of the capsule, collagen stained weakly, whereas the inner surface of the capsule showed strongly stained cells arranged as a continuous layer on the inner surface (Fig. 1A) . 

Structure of the Intermediate Capsule.

At 2 months after surgery, conjunctiva and Tenon’s tissue overlying the capsule contained empty blood vessels with no signs of inflammation. The capsule was arranged in two layers. The external layer contained normally stained, moderately cellular collagenous connective tissue with a modest number of capillaries of normal caliber, and the internal layer demonstrated weakly stained, loosely arranged collagen with a reduced number of swollen, vacuolated, and fragmented cells and amorphous basophilic material that was concentrated toward the inner surface of the capsule (Figs. 1B 1C) . 

Structure of Late Capsules.

Established capsules from 2.3 to 20 years after surgery did not show any signs of inflammation in the overlying conjunctiva and Tenon’s tissue. The capsule contained two layers. The outer vascular fibroproliferative layer consisted of small blood vessels of normal caliber and structure surrounded by cells identified morphologically as mesenchymal cells, resident and wandering macrophages, fibroblasts, and other indeterminate cells. These cells were distributed in a normally stained strongly birefringent collagenous matrix. The outer layer blended into an inner avascular fibrodegenerative layer that contained a reduced number of palely staining swollen, vacuolated, and fragmented cells within a matrix of weakly staining and weakly birefringent collagen fibers, many of which were either swollen or fragmented (Figs. 1D 1E) . 

Cytology of the Intermediate Capsule.

At 2 months after surgery, most cells in the outer layer were of normal morphology and staining. Approximately 4% of all cells showed apoptotic features, and ≈50% of the apoptotic cells were close to blood vessels (Fig. 1F) . All cells in the inner layer showed both enlargement and cytoplasmic vacuolation combined with condensation of nuclear chromatin, blebbing, and fragmentation. Cells and cell fragments disappeared, leaving lucent areas within the ground substance and surrounding tissue fibrils (Fig. 1G) . 

Cytology of the Late Capsules.

At 2.3 to 20 years after surgery, ≈97% of cells in the outer layer consisted of histiocytes, fibroblasts, and macrophages of normal morphology and staining. Of these cells, ≈2% showed apoptotic features that included nuclear condensation, blebbing, disintegration, and formation of membrane-bound vesicles both in the perivascular spaces and between blood vessels (Fig. 1H) . All cells in the inner layer showed morphologic changes including those on the inner surface. Fibroblast nuclei and cytoplasm stained weakly, and the cells became progressively larger and more vacuolated with blebbing before disintegrating into clusters of membrane-bound vesicles toward the inner surface of the capsules (Fig. 1I) . Cells and cell fragments were absent from the deepest layers of 10 capsules. However, in eight capsules, swollen and pale staining fibroblasts were found scattered on the capsule lining accompanied by pale-staining membrane-bound vesicles and cellular debris. Macrophages were similarly altered in appearance. Although they did not enlarge as much as fibroblasts, they demonstrated swelling, nuclear condensation, shrinkage, and disintegration with the formation of dense apoptotic bodies scattered on the inner surface (Figs. 1I 1J) . 

TUNEL staining demonstrated strong fluorescence of blebbing cells and apoptotic bodies in the outer layer of capsules and moderate fluorescence of most cells in the outer layers with strong fluorescence of disintegrating cells and apoptotic bodies in both layers (Fig. 2) . 

Apoptotic Bodies: Membrane-Bound Vesicles

In the inner layer of the capsule, apoptotic cells produced a large number of variably basophilic membrane-bound vesicles that ranged from 0.4 μm (the limit of optical resolution) to ≈4 μm diameter. These vesicles were most numerous in the inner layers of maximum cellular degeneration with smaller numbers in the more superficial layers of the capsule and occasional vesicles in perivascular spaces on the capsule surface (Fig. 1K) . 

Interactions between Apoptotic Cells and the Intercellular Matrix

With increasing time after surgery, the number of cells decreased in the inner layer and, in some cases, the cells were absent from the deepest layers. The cellular changes were accompanied by general swelling of collagen fibers with decreased birefringence and reduced affinity for stains followed by fragmentation and disappearance of collagen fibrils from the deepest layers (Fig. 1K) . 

Local Matrix Breakdown in Established Capsules

With rare exception, apoptotic cells and cell remnants produced no observable effect on the adjacent tissue matrix in the outer layers of most capsules. The exceptions occurred in the 12.0-, 15.8-, and 20-year-old capsules where ≈20% of apoptotic cells were surrounded by low contrast lucid zones. However, a high proportion of apoptotic cells visibly altered the adjacent matrix in the deep layers of most capsules. Changes included (1) narrow lucent zones around occasional cells in the mid zones; (2) wider loss of tissue matrix around cells in the deepest layer of markedly altered matrix; (3) well-defined lucent zones containing strongly birefringent, irregularly oriented light-scattering fibrils. These lucent areas were more visible around the larger membrane-bound vesicles (apoptotic bodies). These zones increased in size and prominence as the vesicles collapsed and disappeared. After the disappearance of apoptotic remnants the lucid zones faded and disappeared (Figs. 1K 1L 1M) . 

Distribution of Labeled Cells

The distribution of labeled cells was similar in all six capsules. Immunohistochemical staining identified three categories of cells. Cells in the first category expressed PCNA (Figs. 3A 3B 3C) , CD68 (Figs. 3D 3E 3F) , actin (Figs. 3G 3H 3I) , and CD34 (Figs. 3J 3K 3L) ; were associated with blood vessels on the outer surface of the capsules; and were also found in both capsule layers. Cells in the second category expressed CD45 (Figs. 3M 3N 3O)and desmin (Figs. 3P 3Q 3R) . These cells were associated with blood vessels in the outer layer but were not found in the inner layers (Table 2) . Cells in the third category expressed vimentin (Figs. 3S 3T 3U)and were found in small amounts in blood vessel walls in Tenon’s tissue and both layers. 

Distribution of Apoptotic (TUNEL⁺) Cells

TUNEL labeled cells and apoptotic bodies in the perivascular spaces on the outer capsule surface and in both layers (Fig. 2) . 

Distribution of Metabolically Active (PCNA⁺) Cells

PCNA⁺ cells were concentrated in the perivascular spaces on the outer capsule surface and were widely distributed throughout both layers (Figs. 3A 3B 3C) . 

Distribution of Bone Marrow-Derived Angiogenic Factor-Positive (CD34⁺) Cells

CD34-labeled blood vessels in subconjunctival connective tissue and on the external capsule surface. CD34⁺ cells were found in the outer layer and in small numbers in the inner layer (Figs. 3J 3K 3L) . 

Distribution of Leukocyte Common Antigen-Positive (CD45⁺) Cells

CD45⁺ cells were concentrated in and around blood vessels on the outer capsule surface and were not present in the inner layer (Figs. 3M 3N 3O) . 

Distribution of Macrophages (CD68⁺Cells)

CD68-labeled macrophages were widely distributed in the loose subconjunctival connective tissue overlying the capsule with increased numbers close to the outer capsule surface. In the capsule itself CD68⁺ cells were most abundant around capillaries on and in the outer layer and their number progressively decreased in the inner layers (Figs. 3D 3E 3F) . 

Distribution of Myofibroblasts (Actin⁺ Cells)

α-Smooth muscle actin was found in blood vessels on the outer capsule surface and throughout both capsule layers (Figs. 3G 3H 3I) . 

The distribution of labeled cells is given in Table 3 . 

Distribution of Collagens and Structural Proteins

The distribution of these components was similar in all six capsules (Table 3 , Fig. 4A ). Collagen types I (Figs. 4B 4C 4D) , III (Figs. 4E 4F 4G) , and VI (Figs. 4H 4I 4J)were evenly distributed throughout the conjunctival connective tissue and both capsule layers. However, collagen type I occurred only in very small amounts, and elastin (Figs. 4K 4L 4M)was widely distributed in conjunctiva and in the outer layer and was absent from the inner layer. Collagen type IV (Figs. 4N 4O 4P) , laminin (Figs. 5A 5B 5C) , fibronectin (Figs. 5D 5E 5F) , and thrombospondin (Figs. 5G 5H 5I)were concentrated in blood vessel walls and the inner layer (Table 3 , Figs. 4B 4C 4D 4E 4F 4G 4H 4I 4J 4K 4L 4M 4N 4O 4P 5 ). Vitronectin (Figs. 5J 5K 5L)was evenly distributed in blood vessel walls and both capsule layers. 

Distribution of Mast Cells

Rare mast cells were found in the perivascular spaces of the subconjunctival connective tissue and adjacent to the outer surface of two capsules. 

Discussion

The results of this study clarified some aspects of complex cellular responses to draining aqueous that evolve over time and are common to cases of both primary and secondary glaucoma. The findings can be considered in three groups: the types of cell involved, features of apoptosis, and the types and distribution of collagen and ground substance. 

Cytology and immunohistochemical staining confirmed previous studies showing the presence of fibroblasts and macrophages and also indicated the presence of cells expressing α-smooth muscle actin throughout the capsule, together with cells expressing CD34 in both capsule layers. Mast cells, eosinophils, and plasma cells were not found in significant amounts. 

Examination of oblique sections demonstrated the cytology of cells undergoing apoptosis. This differed according to whether the cells were close to the external surface of the capsule or present in the inner layers or on the inner surface. Cells close to the external surface showed the classic features of apoptosis. Some cells (≈5%) in the inner layers showed the classic features, but most became markedly enlarged, stained palely and developed numerous intranuclear and intracytoplasmic vacuoles before disintegrating into clusters of variably basophilic membrane-bound vesicles which dispersed, collapsed, and disappeared. 

Cells undergoing apoptosis in the inner layers showed alterations in the tissue matrix which varied according to the age of capsules. At 2 months, apoptotic cells produced well-marked, variably sized, and in some cases overlapping, lucent areas in a tissue matrix which showed few weakly birefringent fibers. Older capsules showed appearances varying from a subtle loss of tissue fibrils around apoptotic remnants in the inner layers of established capsules through definite lucent areas to clearly defined lucent areas containing randomly arranged strongly birefringent fibrils contrasting with the surrounding weakly birefringent regularly arranged fibrils. 

Collagens III, IV, and VI occurred in high concentrations, whereas collagen I (the dominant collagen of mature scar tissue elsewhere in the body) was present only in very low concentrations. Collagens III and VI were generally distributed throughout the capsule and overlying connective tissue. Collagen IV and laminin were found in subconjunctival blood vessels and were concentrated in the inner capsule layers. Elastin fibers occurred in blood vessels, in the subconjunctival connective tissue, in the outer capsule layers and, in trace amounts, in the inner layers. Vitronectin and fibronectin occurred diffusely in subconjunctival connective tissue and throughout the capsules. Thrombospondin occurred in blood vessels in the subconjunctival connective tissue and outer capsule layer and in high concentrations in the inner layer. 

The rate of cell turnover was estimated by counting microscopically recognizable apoptotic cells in the outer layers of hematoxylin and eosin-stained sections of capsules. Apoptosis of liver cells in mice induced by an intraperitoneal injection of 10 μg Fas ligand has been shown to kill 50% of animals within 8 hours. ²¹ This finding implies that the process of apoptosis in well vascularized tissue takes ≈12 hours. Fibroblasts and macrophages in the vascular outer capsule layers probably complete apoptosis in ≈12 hours. The presence of ≈4% of apoptotic cells in the intermediate capsule implied that all susceptible cells would become apoptotic in 25 × 12 hours or ≈13 days, whereas the corresponding proportion of ≈2% of cells in established capsules implied complete turnover in ≈25 days. These inferences are likely to be underestimates as only cells showing unequivocal signs of apoptosis were counted. ⁹  

The pattern of cells involved in capsule formation and the types of collagen laid down in capsules corresponded to those found in the early stages of wound healing in connective tissue. However, the very low concentration of collagen I was an unexpected feature of functioning capsules as was the high concentration of collagen IV, laminin, and thrombospondin in the inner layers. These findings appeared to be related to ongoing apoptosis but the mechanisms involved are as yet unclear. 

An understanding of the cellular processes involved in capsule formation and long-term maintenance is important in the management of glaucomatous eyes treated with Molteno implants and probably aqueous shunts in general. The immediate inflammatory fibroproliferative response of vascular connective tissue to draining aqueous can be decreased by the short-term administration of anti-inflammatory fibrosis suppression therapy ^{6 22 23} to delay deposition of fibrous tissue until a slowly developing avascular apoptotic fibrodegenerative response has developed sufficiently to inhibit the fibroproliferative response. 

Conclusion

The complex cycle of cell activation and apoptosis demonstrated in this study has two components. The first, an inflammatory fibroproliferative component, resembles the initial stages of the wound-healing response characteristic of higher animals with a high-pressure blood circulation. The second, an avascular noninflammatory apoptotic fibrodegenerative response, resembles the predominant tissue defense response of primitive animals that lack an efficient blood circulation. ²⁴ This phylogenetically ancient tissue response has been shown to regulate embryonic development in all animals and exert an anti-inflammatory action in most tissues of mammals. ²⁵ Recent advances in our understanding of cell biology offer hope that improved methods of minimizing the fibroproliferative and enhancing the apoptotic fibrodegenerative responses of connective tissue to aqueous will be developed in the foreseeable future. 

 Supported by the Healthcare Otago Charitable Trust, Dunedin, New Zealand (THB).

 Submitted for publication November 5, 2007; revised May 6 and August 20, 2008; accepted January 21, 2009.

 Disclosure: A.C.B. Molteno, Molteno Ophthalmics (I, P); A.M. Thompson, None; T.H. Bevin, None; A.G. Dempster, 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.

 

Clinical Features of Capsule Specimens

Specimen	Glaucoma	Age at Operation (y), Sex	Plates (n)	Drainage	AIFS	Final IOP (mm Hg)	Hypotensive Medication at Final Follow-up	Age of Capsule
1	Uveitic	68, M	2	Delayed	No	6	Acetazolamide	11 d
2*	Neovascular	32, M	1, †	Immediate	No	9	Timolol	2 mo
3, ‡	Neovascular	37, M	1	Immediate	Yes	14	—	3 mo
4, ‡	Uveitic	36, F	2	Delayed	Yes	15	Adrenaline	2.3 y
5, ‡	PXF	73, M	2	Delayed	No	15	—	2.6 y
6, ‡	Neovascular	72, F	1	Immediate	Yes	2	—	2.8 y
7, ‡	Neovascular	70, M	1	Immediate	No	28	—	3.5 y
8	Neovascular	76, M	2	Delayed	No	12	—	4.1 y
9	ACG	87, F	1	Delayed	No	14	—	4.1 y
10*	POAG	80, M	1	Delayed	No	14	—	4.2 y
11*	POAG	79, M	2	Delayed	No	13	—	4.7 y
12	PXF	67, M	1	Delayed	No	7	—	8.5 y
13	POAG	88, M	1	Delayed	No	15	—	8.8 y
14	Neovascular	71, M	2	Delayed	No	14	—	9.4 y
15*	PXF	85, F	1	Delayed	No	8	—	12.0 y
16	PXF	69, M	2	Delayed	No	14	—	12.6 y
17, ‡	Buphthalmos	23, F	2	Delayed	No	15	—	13.2 y
18*	PXF	81, F	1	Delayed	No	7	—	15.8 y
19*	Aphakic/ACG	65, F	2	Delayed	No	12	Propine	20.0 y

 AIFS, anti-inflammatory fibrosis suppression; IOP, intraocular pressure; POAG, primary open angle glaucoma; PXF, pseudoexfoliative glaucoma; ACG, angle-closure glaucoma.

*  Immunohistochemical staining for cell surface antigens.

†  This patient received a 230 mm² Molteno3 implant, whereas all other single-plate implants were the previous single-plate Molteno implant model.

‡  Immunohistochemical staining for collagen and glycoproteins.

Figure 1.

 

Histologic structure of capsules. (A) Vertical section of an 11-day-old capsule (Table 1 , case 1) showing (a) Tenon’s tissue with congested blood vessels, (b) the outer fibroproliferative layer of the capsule with dilated capillaries, (c) the inner fibrodegenerative layer with a continuous layer of mesodermal cells on the inner surface, and (d) the cavity. All figures are identically oriented and labeled. Stain, Gomori trichrome; bright field. (B) Vertical section of a 2-month-old capsule (Table 1 , case 2) showing (a, b) empty blood vessels in Tenon’s tissue and the outer layer, (c) the inner layer, and (d) the cavity. Stain, Gomori trichrome; bright field. (C) Tangential section of a 2-month-old capsule (Table 1 , case 2) showing (b) birefringent collagen in the outer layer, (c) lucent zones of breakdown of ground substance and tissue fibrils around apoptotic cells in the inner layer; and (d) the cavity; black arrows: apoptotic cells. Stain, hematoxylin and eosin; bright field and polarized light double exposure. (D) Vertical section of a 20-year-old capsule (Table 1 , case 19) showing (b) blood vessels in the outer layer, (c) apoptotic bodies in the inner layer and on the inner surface of the capsule, and (d) the cavity; blue arrows: apoptotic cell bodies. Stain, hematoxylin and eosin; bright field. (E) Oblique section of a 20-year-old capsule (Table 1 , case 19) showing (b) a normally stained outer layer with an apoptotic cell (black arrow), (c) the inner layer with loss of collagen staining, and scattered apoptotic bodies (blue arrows) within the inner layer and on the inner surface. Stain, hematoxylin and eosin; bright field. (F) Tangential section of a 2-month-old capsule (Table 1 , case 2) showing (b) normally stained collagen in the outer layer and (c) the inner layer with two apoptotic cells (black arrows) showing chromatin condensation and disintegration. Stain, hematoxylin and eosin; bright field. (G) Tangential section of a 2-month-old capsule (Table 1 , case 2) showing (c) the inner layer with swollen apoptotic cells (black arrows) and cell debris with surrounding lucent areas, and (d) the cavity. Stain, hematoxylin and eosin; bright field. (H) Tangential section of a 4.1-year-old capsule (Table 1 , case 9) showing (a) Tenon’s tissue, (b) the outer layer with an apoptotic cell (black arrow) near the external surface, and (c) the inner layer. Stain, hematoxylin and eosin; bright field. (I) Tangential section of a 4.1-year-old capsule (Table 1 , case 9), showing (b) the outer layer and (c) the inner layer with an apoptotic cell (black arrow) with surrounding lucent area. Stain, hematoxylin and eosin; bright field. (J) Tangential section of a 4.1-year-old capsule (Table 1 , case 9) showing (c) an inner layer close to the cavity with apoptotic bodies (blue arrows). Note the swollen vacuolated fibroblast nucleus (top left), an apoptotic cell disintegrating in a cluster of membrane-bound vesicles (middle right), and shrunken apoptotic bodies (bottom) and scattered membrane-bound vesicles (top right). Stain, hematoxylin and eosin; bright field. (K) Oblique section of a 20-year-old capsule (Table 1 , case 19) showing (c) the inner layer with apoptotic cells (blue arrows) disintegrating to form membrane-bound vesicles (note the surrounding lucent areas) and (d) scattered swollen apoptotic bodies (membrane-bound vesicles) and condensed basophilic apoptotic bodies on the inner surface. Stain, hematoxylin and eosin; bright field. (L) Tangential section of a 20-year-old capsule (Table 1 , case 19) showing (c) the inner layer with (blue arrows) scattered apoptotic bodies with surrounding lucent areas. Stain, hematoxylin and eosin; bright field. (M) Tangential section of a 20-year-old capsule (Table 1 , case 19) showing (c) the inner layer with scattered apoptotic bodies (blue arrows) surrounded by lucent areas showing strong birefringence. Stain, hematoxylin and eosin; bright field polarized light. Magnification: (A–C, E) ×160; (D, G, H–K) × 400; (F, L, M) ×1000. Scale bars: (A–C, E), 100 μm; (D, G–K, J), 40 μm; (F, L, M) 20 μm.

Figure 2.

 

(A) Oblique section of a 4.1-year-old capsule (Table 1 , case 9) showing (a) Tenon’s tissue and (b, c, d) the outer layer, inner layer, and cavity with (white arrows) TUNEL⁺ cells. Stain TUNEL; fluorescence microscopy. (B) Oblique section of same capsule as in (A) showing (b) blebbing of (white arrow) strongly fluorescence apoptotic cell near outer capsule surface, and (c) the inner layer. Stain, TUNEL; fluorescence microscopy. TUNEL⁻ (C) and (D) TUNEL⁺ control samples: same capsule as in (A). Magnification: (A) ×200; (B) ×630; (C, D) ×100. Scale bars: (A) 70 μm; (B) 30 μm; (C, D) 250 μm.

Figure 3.

 

(A) Vertical section of a 2-month-old capsule (Table 1 , case 2), showing (a) PCNA⁺ cells along a blood vessel in Tenon’s tissue, (b, c) PCNA⁺ cells in both layers, and (d) the cavity. Stain, AEC; bright field. (B) PCNA⁻ control: same capsule as in (A). (C) PCNA⁺ control in human accessory lachrymal gland from same patient as in (A). (D) Vertical section of a 20-year-old capsule (Table 1 , case 19), showing (b, c) CD68⁺ cells in both layers, and (d) the cavity. Stain, LSAB2-HRP. (E) CD68⁻ control: same capsule as in (D). (F) CD68⁺ control human lymph node. (G) Vertical section of a 2-month-old capsule (Table 1 , case 2) showing α-smooth muscle actin staining of cells in (a) blood vessels of conjunctiva and Tenon’s tissue, and (b, c) both layers, and (d) cavity. Stain, LSAB2-HRP and K3466 diamino benzidine chromogen. (H) α-Smooth muscle actin⁻ control: same capsule as in (G). (I) α-Smooth muscle actin⁺ control smooth muscle in human artery. (J) Vertical section of a 20-year-old capsule (Table 1 , case 19) showing CD34⁺ staining in blood vessel walls in (a) Tenon’s tissue and (b) the outer layer and (c) occasional positive cell remnants in the inner layer, and (d) the cavity. Stain, LSAB2-HRP and K3466 diamino benzidine chromogen; bright field. (K) CD34⁻ control: same capsule as in (J) (L) CD34⁺ control human blood vessels in subcutaneous tissue. (M) Vertical section of a 2-month-old capsule (Table 1 , case 2) showing (a) a few CD45⁺ cells around blood vessels in Tenon’s tissue, (b) occasional CD45⁺ cells in the outer layer, and (c) absence of CD45⁺ cells from the inner layer. Stain LSAB2-HRP and K3466 diamino benzidine chromogen; bright field. (N) CD45⁻ control same capsule as in (M). (O) CD45⁺ control human granulation tissue, with CD45⁺ cells in and around the blood vessels. (P) Vertical section of a 2-month-old capsule (Table 1 , case 2) showing desmin staining of blood vessels in (a) Tenon’s tissue and (b) the outer layer and (c) weak, diffuse staining of granular material in the inner layer and (d) cavity. Stain, LSAB2-HRP and K3466 diamino benzidine chromogen; bright field. (Q) Desmin⁻ control: same capsule as in (P). (R) Desmin⁺ control: human ciliary muscle. (S) Vertical section of a 2-month-old capsule (Table 1 , case 2), showing vimentin staining of blood vessels in (a) Tenon’s tissue, (b) the outer layer, and (c) scattered staining in the inner layer and (d) cavity. Stain, LSAB2-HRP and K3466 diamino benzidine chromogen. (T) Vimentin⁻ control: same capsule as (S). (U) Vimentin⁺ control: blood vessels in human granulation tissue. Magnification: (A–C, G–I, M–U) ×160; (D–F, J–L) ×400. Scale bars: (A–C, G–I, M–U) 100 μm; (D–F, J–L), 40 μm.

Table 2.

 

Distribution of Collagens and Glycoproteins in Molteno Implant-Associated Capsules

	Staining Characteristic	Distribution
Collagens/Elastin
Type I	Weak	Throughout
Type III	Strong, according to tissue density	Throughout
Type IV	Strong	Around blood vessels, inner layer
Type VI	Strong	Throughout
Elastin	Weak	Throughout
Glycoproteins
Fibronectin	Strong	Around blood vessels, inner layer
Laminin	Moderately strong	Inner layer
Thrombospondin	Strong	Blood vessel walls, inner layer
Vitronectin	Weak	Throughout

Table 3.

 

Distribution of Cells, Collagen Fibrils, and Glycoproteins in Subconjunctival Connective Tissue and Molteno Implant-Associated Capsules

	Subconjunctival Connective Tissue					Capsule
		Diffuse	Blood Vessel			Outer Layer					Inner Layer
				Wall	Perivascular Space	Diffuse	Blood Vessel
										Wall		Perivascular Space
Cell activity
Matrix breakdown	−	−	−	−	−	−	++++
Apoptotic bodies	−	−	+	−	−	++	++++
TUNEL	−	−	±	−	−	+	+++
PCNA	−	−	++	−	−	+++	++++
CD34	+	+	−	+	+	±	+
CD45	−	+	+	−	++	−	−
CD68	±	−	+	+	−	+	++
Actin	+	+++	−	++++	++	−	++++
Desmin	−	±	−	−	−	−	−
Vimentin	−	±	−	−	+	−	+
Collagens/elastin
Collagen type I	+	−			±	−			±
Collagen type III	++	+++			+++	+++			++++
Collagen type IV	−	+++			−	+++			++++
Collagen type VI	++	+++			++	++			++++
Elastin	+++	−			++	+			±
Glycoproteins
Fibronectin	++	−			++	++			++
Laminin	−	++			−	++			++++
Thrombospondin	−	++			±	±			++++
Vitronectin	+	−			+	+			+

Figure 4.

 

(A) Vertical section of a 2.3-year-old capsule (Table 1 , case 4) showing (a) Tenon’s tissue, (b) the outer layer, (c) the inner layer, and (d) the cavity. Stain, hematoxylin and eosin; bright field. (B) Vertical section of a 2.3-year-old capsule (Table 1 , case 4) showing (a–c) trace amounts of collagen type I in all layers and (d) the cavity. Stain, AEC; bright field. (C) Collagen type I⁻ control: same capsule as in (B). (D) Collagen type I⁺ control: human tonsil. (E) Vertical section of a 2.3-year-old capsule (Table 1 , case 4), showing (a–c) the distribution of collagen type III in all layers and (d) the cavity. Stain AEC; bright field. (F) Collagen type III⁻ control: same capsule as in (E). (G) collagen type III⁺ control: human tonsil. (H) Vertical section of a 2.3-year-old capsule (Table 1 , case 4) showing (a–c) the distribution of collagen type VI in all layers and (d) the cavity. Stain, AEC; bright field. (I) Collagen type VI⁻ control: same capsule as in (H). (J) Collagen type VI⁺ control: human tonsil. (K) Vertical section of a 2.6-year-old capsule (Table 1 , case 5) showing the distribution of elastin fibers in (a) Tenon’s tissue and (b) the outer layer and (c) the absence of elastin fibers in the inner layer and (d) the cavity. Stain, AEC; bright field. (L) Elastin⁻ negative control: same capsule as in (K). (M) Elastin⁺ control: human tonsil. (N) Vertical section of a 2.3-year-old capsule (Table 1 , case 4) showing (a) distribution of collagen type IV around blood vessels in Tenon’s tissue and (b) the outer layer and (c) collagen type IV concentrated in the inner layer and (d) the cavity. Stain, AEC; bright field. (O) Collagen type IV⁻ control: same capsule as in (N). (P) Collagen type IV⁺ control: human tonsil. Magnification, ×160. Scale bars, 100 μm.

Figure 5.

 

(A) Vertical section of a 2.3-year-old capsule (Table 1 , case 4) showing (a) the distribution of laminin around blood vessels in Tenon’s tissue and (b) the outer layer and (c) laminin concentrated in the inner layer and (d) the cavity. Stain, AEC; bright field. (B) Laminin⁻ control: same capsule as in (A). (C) Laminin⁺ control: human tonsil. (D) Vertical section of a 2.3-year-old capsule (Table 1 , case 4) showing the distribution of fibronectin in (a) blood vessel walls of Tenon’s tissue and (b) the outer layer and (c) concentrated in the inner layer and (d) the cavity. Stain, AEC; bright field. (E) Fibronectin⁻ control: same capsule as (D). (F) Fibronectin⁺ control: human tonsil. (G) Vertical section of a 2.3-year-old capsule (Table 1 , case 4) showing the distribution of thrombospondin (a) around blood vessels in Tenon’s tissue, (b) in the outer layer, and (c) concentrated in the inner layer, and (d) the cavity. Stain, AEC; bright field. (H) Thrombospondin⁻ control: same capsule as (G). (I) Thrombospondin⁻ control: human tonsil. (J) Vertical section of a 2.6-year-old capsule (Table 1 , case 5) showing the distribution of vitronectin in (a) the blood vessel walls of Tenon’s tissue, (b, c) the outer and inner layers, and (d) the cavity. Stain, AEC; bright field. (K) Vitronectin⁻ control same capsule as (J). (L) Vitronectin⁺ control human tonsil. Magnification, ×160. Scale bars, 100 μm.

The authors thank the families of the patients who kindly donated the eyes.