regeneration of the human corneal endothelium

Albrecht von Graefes Arch Klin Ophthalmol Graefes Archiv tOr klinische und experimentelle

(198l) 215:341 348 Ophthalmologie �9 Springer-Verlag 1981

Regeneration of the Human Corneal Endothelium

A.S.E.M. Study *

G. Renard 1"2'3.*, Y. Poul iquen 1"2, and M. Hirsch 2 Laboratoire de la clinique ophtalmologique de l'H6tel-Dieu, Place du Parvis Notre Dame,

F-75004 Paris, France z Groupe U 86 INSERM 3 Laboratoire de Biophysique (Professeur Galle) CHU Cr&eil

Abstract. The ability o f the human corneal endothel ium to regenerate is studied with the scanning electron microscope th rough examples o f corneal diseases and penetrat ing keratoplasties. This study does not lead to final conclusions on the possibilities o f regenerat ion o f the human corneal endothelium but allows us to say that :

regenerat ion occurs th rough size increasing and deformat ion of the remain- •ng cells. - the increase in number and size o f surface microvilli may simply indicate a state o f cell activation. - the presence o f two nuclei in one cell is p robab ly obtained by amitot ic division but no complete mitosis has been seen.

Displacement o f endothelial cells is a real progression and the cell is able to overcome obstacles. - the fibroblastic t ransformat ion of the endothelial cells is present in man but this may simply represent the migrat ing fo rm of the cells.

Introduction

The ability o f the corneal endothel ium to regenerate is a subject which has been widely debated due to its considerable clinical relevance. Surgery to the anterior chamber o f the eye, may indeed, destroy or cause deteriorat ion in a p ropor t ion o f the endothelial cell populat ion. The t ransparency of the cornea and a satisfactory ou tcome to the surgical t reatment depend on the regenerat ion o f this cell layer.

Most studies o f endothelial regenerat ion have been carried out on the rabbit eye, since the size o f the rabbit cornea facilitates clinical and histological studies.

* This paper was presented in part at the Eighth Annual Meeting of the European Club for Ophthalmic Fine Structure in West Berlin on 28th, 29th March, 1980

** Corresponding author

006 5-6100/81/0215/03 41/$01.60

342 G. Renard et al.

However, the endothelium of this species has particular properties (Renard, Hirsch and Pouliquen 1978), which include: 1 - transformation of endothelial cells into fibroblast-like cells, 2 - mitotic activity and 3 - cell spreading.

These properties allow the endothelium to spread over a lesion, however large, within a few days and to reacquire its normal appearance (Khodadoust and Green 1976). This type of regeneration which is reminiscent of epithelial cell regeneration, is not usually found in the human cornea and studies of the human cornea have never shown the presence of mitoses (Doughman, Van Horn and Rodman 1976). In fact it seems that resurfacing occurs as a result of cell spreading, rather than cellular regeneration. This has been observed histologically in a few specimens obtained after surgical treatment, and in experi- ments performed with human endothelial cell cultures (Dougmman et al. 1976; Sperling 1978) and confirmation of this has been provided by studies in vivo using the specular microscope (Sherrard 1976). No precise study has been done with the electron microscope, Owing to the scarcity of suitable material and to the difficulties encountered in the correct interpretation of the appearance of corneal tissue obtained after surgery.

The purpose of this study is to use the material we have obtained after five years of systematic examination, by scanning electron microscopy, of the endothelium of the corneal tissue removed during penetrating keratoplasty.

Material and Methods

Ninety corneal discs have been examined and the experience acquired has provided a background for the interpretation of the appearance of the endothelium. We have selected from this collection, material from the following diesease entities (the number of specimens being shown in parenthesis), Keratoconus (6), graft rejection after penetrating keratoplasty (5), repeat keratoplasty performed owing to epithelia1 or stromaI opacities (4), senile dystrophy of the endothelium without cornea guttata (3), juvenile corneal dystrophy (1), old penetrating corneal wounds (2), healed corneal abscess (l), granular dystrophy (1), lattice corneal dystrophy (1), choroidal melanoma without involvement of the anterior segment (1).

The observations we have made on the human cornea will be compared with the experimental results obtained from studies on the rabbit cornea after penetrating keratoplasty and provoked graft rejection caused by alkali-burn or by freezing with liquid nitrogen. In all of this material the specimens were prepared in the same way. After excision they were immediately fixed for 2 hours in 2.5% Glutaraldehyde in S6rensen's buffer, and were successively, washed in buffer, dehydrated through graded alcohols, embedded in EPON, prepolymerized at 37C for 18hours, washed in propylene oxide, hardened at 60 C for 48 hours and coated with gold.

They were then examined in a CAMECA MEB 07 Scanning Microscope.

Results

The most significant feature of the diseased human corneal endothelium was enlargement and irregular size of the cells (Fig. 1 a). These enlarged cells always showed a depression at the line of the apical junctions, although the finger-like shape of the apical junctions was usually retained. The polygonal outline of the cell surface, however was no longer found and all the cells were considerably attenuated (Fig. lb). The apical junctional system persisted as long as two cells were in contact. Isolated cells sometimes retained a characteristic edge.

134

Corneal Endothelial Regeneration 343

Fig. 1. a Corneal abscess. The cells have an irregular size, the apical junctional system is disturbed, one ceil (arrow) has two nuclei, b Senile dystrophy. The cells are very flat and have lost their haxagonal shape. Intercellular junctions are broken, e Penetrating keratoplasty. A large cell with two nuclei (arrow). d Penetrating keratoplasty. Expansions of endothelial cells have moved over the host-graft junction (Each x 800)

Two nuclei were often found in a large proportion of the enlarged cells (Fig. la and c). The nuclei in the binucleate cells were sometimes identical or, by contrast, were of completely different size and shape, one of them being round, the other of oval-shape.

Cell displacement was also often observed and was considered to occur by a sliding motion. This can be shown by examination of a host-graft junction after keratoplasty. On the recipient edge, flattened cells (Fig. 1 d) were spread across the junction and contacts between these cells, although attenuated were

135


Fig. 2. a Penetrating keratoplasty ( x 1,500). An endothelial cell is able to spread over the junction and to lost its contact with underlying tissues, b Keratoconus ( • 4,000) long apical cellular expansions between two cells, e Lattice corneal dystrophy ( x 800). The long cellular expansions are sometime found only on one side of the cell. d Melanoma of the choroid ( x 700). Some cells have an increase of the number of surface microvilli and disrupted intercellular digitations (arrow)

real and this indicated a common tendency to overlap. We also found other cells whose expansions bridged the surface when undulations occured (Fig. 2 a). The endothelial cell appeared therefore to be capable of displacement either around or over barriers and to lose thereby all contact with the underlying cells.

The finger like expansions of the apical junctions were usually short, regularly shaped and often enlarged at their base. By contrast, in the course of regeneration in the endothelium of the rabbit we have constantly observed the existence of unusually long finger-like expansions between newly regenerated cells. After

136


experimental keratoplasty in the rabbit, we found these expansions only on the site of contact between the graft regeneration front and the host regeneration front. In human corneas the apical digitations were usually shorter and separated by depressions which sometimes appeared as holes, according to the viewing angle. Abnormal ly long expansions were found in the endothelial monolayer in keratoconus (Fig. 2b) and in corneal dystrophy: they were considered to be an indication of a long term regeneration process. When the cell boundaries were still at some distance from one another, these expansions appeared to be more recently formed. In a few cases the expansion crossed over the surface of the cell towards a more distant cell. In other cases, they were found only on one side of the cell (Fig. 2c). In chronic corneal diseases, other abnormalities in the shape of the apical expansions were observed: sometimes they were flat and broad or were completely entangled, or again were found on one side of the cell only.

Whether or not mitosis occurs in the human corneal endothelium is an important source of debate at the present time. Mitosis were frequent in our studies of the rabbit endothelium and by scanning electron microscopy their development was as follows: initially, the cell became indented and the surface microvilli increased in size; the cell then became swollen and their microvilli were globular, with an associated swelling of the apical junctions. The cell then divided and the junctional system were reestablished from large lateral microvilli. The cell gradually recovered its normal appearance, it still retained for sometime, an increased number of hypertrophic villi. Similar appearances were observed in human diseased corneas in which cells with abnormally large number of surface microvilli are found (Fig. 2d). Such cells were isolated or were found in groups and the microvilli were either small (Fig. 3a) or large (Fig. 3b). Microvilli were found in cells whose junctional systems were intact as well as in cells where the junctional systems were disorganized. In some cases the microvilli were distended in a manner reminiscent of the dividing cells in the rabbit endothelium. However in all the human material we have examined, we have never been able to demonstrate a complete mitosis, nor even two neighbouring cells with a morphology which would have indicated recent mitosis.

Finally, I wish to use this material to discuss the property of the endothelium to form fibroblast-like cells in the course of regeneration. This function was particularly evident in the rabbit endothelium where fibroblasts often appeared as poorly differentiated endothelial cells. A similar appearance was observed in human endothelium (Fig. 3c). The endothelium origin of such cells has been debated, but in our specimens all intermediate forms could be demonstrated ranging f rom the typical endothelial cell, which was flat and had a protruding oval-shaped nucleus to the accepted spindle-shaped fibroblast-like cell; except that the latter, shows at one pole a surface which was apparently identical to the surface of the normal endothelial cell (Fig. 3d). Juxtaposition of spindle cells and of flat cells has been observed in human corneas also and the fibroblast- like cells had a wide variety of appearances which included long cytoplasmic expansions which ran over the neighbouring cells. In a number of several dam- aged human corneas, the posterior surface was lined by elongated spindle cells

137


Fig. 3. a Melanoma of the choroid (x 700). The number of microvilli may be different between two neighbouring cells with unbroken intercellular digitaions, b Keratoconus ( x 1,000). The surface microvilli are larger and more numerous than usual in a group of cells, e Penetrating keratoplasty ( x 800). The human endothelial cell may take the shape of a fibroblast-like cell. d Corneal wound (x 1,500). An endothelial cell with a fibroblast like aspect at one end and a typical appearance at the other end

w h i c h were very s imi la r to the sp indle cells o b s e r v e d in the r a b b i t a f te r exper i -

m e n t a l burns .

Discussion

S c a n n i n g e l ec t ron m i c r o s c o p i c s tudies c a n n o t p r o v i d e all t h e necessa ry i n f o r m a -

t i on on the poss ib le m e c h a n i s m s o f r e g e n e r a t i o n in t he h u m a n c o r n e a l e n d o t h e - l i u m bu t t hey do yie ld i n f o r m a t i o n which c a n n o t be o b t a i n e d e l sewhere . Jus t

as w i t h specu la r m i c r o s c o p i c studies, it is poss ib le to s h o w tha t r e g e n e r a t i o n

138


occurs through an increase in cell size by deformation of the remaining cells. With SEM, it is possible to state that the apical junctional system remains intact, but this is a fact which could also be noted from specular microscopic examination since the outline of the cell is dependant on the junctional com- plexes.

Transmission electron microscopic studies have already established that the cells become thinner and larger in the course of regeneration but the SEM demonstration of the presence of two nuclei in the same cell is interesting and has not been described in previous studies on the human cornea. It is probable that their presence is due more to nuclear division than to cell fusion. The questions to be answered are why cell division is never completed and also at what point the cell loses its mitotic activity. It is known that amitosis exists in other animal species, including the rabbit, but it is always found in association with normal mitotic activity. In the human cornea the endothelial cell seems to be incapable of complete division. According to certain authors this failure to divide may be due to the presence of the central cilium which is observed on the endothelial cell. The cilium is always connected to one of the two centrioles and its presence would deprive the cell of its bipolarity, and thus of all mitotic activity. But cilia also exist in the rabbit's endothelium, yet these cells still preserve their mitotic power.

The spatial migration of the human endothelium is well known and could have been deduced from the deformation and the displacement which occurs in the regenerating cells. This ability to undergo displacement has been observed in human cell culture (Doughman et at. 1976; Sperling 1978). We and others have been able to ascertain that there was not a simple expansion but a real migration of the cell, which was able to overcome barriers (Bourne 1978; Olson and Herenson 1977). We were unfortunately unable to estimate the speed with which the endothelial cell fills up a cell gap.

The lengthening of the apical digitation appears to indicate the presence of cell contact which occurs before junctional regeneration (Hirsch, Renard, Faure and Pouliquen 1976). We believe that the presence of large digitations is evidence of previous regeneration and cell-contact. It is also now clear that a permanent state of endothelial regeneration occurs in keratoconus and in corneal stroma dystrophies, although it was previously thought, that the endothelial cell layer was unaffected in these conditions. We believe that apical expansions represent the ultrastructural appearance of a pathological condition but it is impossible to ascertain ther precise significance.

The total absence of mitosis in all the specimens we have examined does not necessarily imply that they do not exist in the human cornea. In human endothelial cell culture, mitosis have been observed under the influence of several growth factors. It must be noted however, that even in this situation they are rare. The mitotic capacity of the endothelium seems to decrease in the more highly developed species. It is present in small mammals, is very low in monkeys and possibly absent in man. The increase in number and size of the surface microvilli in human endothelium, which is also observed during mitosis in other species, may simply indicate a state of cell activation without necessarily implying mitosis.

139


The fibroblastic transformation of the endothelial cell, which occurs in the rabbit's cornea, is questionable in the human cornea. The surface morphology of both types of cells is similar but it is not easy, which the scanning microscope, to differentiate between a transformed endothelial cell and a genuine fibrocyte. Thus in penetrating keratoplasties (Sanchez and Polack 1978) the cells we observed could have had their origin in the stromal fibrocyte which migrated from the host-graft junction, from the iridiocorneal angle or even from the iris. Concersely these fibroblast-like cells may quite simply be a motile form of the endothelial cell. In this case they would not have the potential for division, no more than the cells from which they originate.

The studies we have carried out raise as many problems as they have solved. Human tissue is among the most difficult to study, owing to the rarety of suitable material and to the impossibility of using experimental techniques. However correlated studies (Wickham and Binder 1976) using animal tissue and endothelial cell cultures may ultimately prove rewarding.

References

Bourne WM(I978) Penetrating keratoplasty with fresh and cryopreserved corneas. Donor endothelial cell survival in primates. Arch Ophthal 96:1073-1074

Doughman DJ, Van Horn D, Rodman WP (1976) Human corneal endothelial layer repair during organ culture. Arch Ophthai 94:1791 i796

Gartner S, Taffet S, Friedman H (1977) The association of rubeosis iridis with endothelialisation of the anterior chamber. Report of a clinical case with histopathological review of 16 additional cases. Brit J Ophthal 61 : 267 27I

Hirsch M, Renard G, Faure JP, Pouliquen Y (1976) Formation of intercellular spaces and junctions in regenerating rabbit corneal endothelium. Exp Eye Res 23 : 385 397

Khodadoust AA, Green K (1976) Physiological function of regenerating endothelium. Invest Ophthal 15:96-101

Olson R J, Levenson JE (1977) Migration of donor endothelium in keratoplasty. Amer J Ophthal 84:711-714

Renard G, Hirsch M, Pouliquen Y (1978) Changes in alkali-burned cornea. Trans Ophthai Soc UK 98:379-382

Sanchez J, Polack FM (1978) Autoradiographic study of retrocorneal membranes. Ann Ophthal 10:1547-1552

Sherrard ES (1976) The corneal endothelium in vivo. Its response to mild trauma. Exp Eye Res 22 : 347-357

Sperling S (1978) Early morphological changes in organ cultured human corneal endothelium. Acta Ophthal 56 : 785-792

Wickham MG, Binder PS (1976) Evaluation of corneal endothelial damage using correlated microscopy techniques. OphthaI Res 8 : 407 413

Received October 4, 1980

140

regeneration of the human corneal endothelium

Documents