Fracture faces of fenestrations and junctions of endothelial cells in ...

Fracture faces of fenestrations and junctionsof endothelial cells in human

choroidal vessels

Manfred SpUznas and Enrico Reale

Replicas of freeze-fractured endothelial cells of human choroidal capillaries show fenestrationswith a centrul thickening, as described in thin-section studies. Between the endothelial cellsthere are discontinuous zonulac occludentes (i.e., maculae or fasciae occludentes), composedof ridges on face A and grooves on face B of the split-cell membrane. They arc also presentalong the sutures of the endothelial cells of vcnules. Close to the intercellular suture gapjunctions, probably connecting capillary endothelial cells and pcricytes, are found.

Key words: choroidal capillaries, choroidal venules, endotheliiim, fenestrations,intercellular junctions, zonula occludens, gap junction, freeze-fracture, permeability.

T.lie endothelial cells of the choriocapil-laries are fenestrated, i.e., they have poresthat are crossed by a diaphragm.1"' Themajority of fenestrations are facing towardBrncli's membrane.'1 On thin sections, theendothelial cells are joined by tight junc-tions or zonulae occludentes.1' Morpho-logically, tight junctions and zoiuilne oc-cludentes are synonymous.7 Physiologically,tight junctions must be distinguished fromleaky junctions.8 Recently, Claude andGoodenough" showed the zonulae occlu-dentes of the tight junction type to becomposed of an accumulation of many

From the University Eye Hospital, Essen, andthe Department of Anatomy, Laboratory ofElectron Microscopy, Hannover Medical School,Hannover, Germany.

Submitted for publication June 17, 1974.Reprint requests: Dr. M. Spitznas, Univeisitiils

Augenklinik, Htiffelandstrase 55, 43 Essen,Germany.

superimposed membrane fusion strands,whereas those of the leaky type are madeup of only a few rows of fusion strands.

It is likely that the known selective per-meability of the choriocapillaris to differ-ent tracer substances" is controlled mainlyby the fenestrations and junctions of theendothelial cells.

In an attempt to provide a better mor-phologic basis for further physiologic re-search, this study gives a detailed analysisof capillary fenestrations and the intercellu-lar junctions in the choriocapillaris andlarger choroidal vessels.

Materials and methods

The study was carried out on choroidal tissueof six human eyes that had to be enucleatedbecause of a malignant melanoma. The areasused for examination were far away from the tu-mors and did not show any pathologic changes.Immediately after enucleation, that is, withintwo minutes after severing the optic nerve, theeyes were separated into anterior and posterior

98

Downloaded From: http://iovs.arvojournals.org/pdfaccess.ashx?url=/data/journals/iovs/933292/ on 02/08/2018

Volume 14Number 2

F enestrations of endothelial cells 99

Fig. 1. Freeze-fracture of the fenestrated choriocapillary endothelium (E), Bruch's mem-brane (BM), and the basal part of a retinal pigment epithelial cell (PE). In this and in thesubsequent figures of freeze-frachire replicas, the arrowhead indicates the direction of platinum-carbon shadowing. x48,000.

portions by making a circular cut in the parsplana region. The posterior portion was placedin a fixative solution of 2 per cent glutaraldehydeand one per cent paraformaldehyde bufferedwith sodium eacodylate 0.2 M at pH 7.4 andcontaining 20 mg. per 100 ml. of calcium chlorideand 1 per cent sucrose. The tissue was thencut in small strips, and the choroid with theattached retina was separated from the sclerawith a thin spatula. This was done under adissecting microscope in a drop of fixative.

Fixation was continued for 24 hours in thesame fixative at room temperature. Then thetissue was washed for two hours in Na-cacodyl-ate buffer 0.1 M at pH 7.2 and soaked with

10 per cent glycerol in Ringer's solution for60 minutes. The specimens were placed on golddiscs, rapidly immersed in Freon cooled withliquid nitrogen to -150° C, transferred to Balzersfreeze-etch apparatus BA 360 M, and fractured.The exposed fracture surface was shadowed withcarbon-platinum according to the technique de-scribed by Moor and Miihlethaler.1u The replicaswere cleaned in a 15 per cent sodium hypo-chlorite solution, mounted on carbon-coated Form-var membranes,11 and examined with a SiemensElmiskop 101 election microscope. A portion oftissue from each eye was processed for con-ventional electron microscopy as described else-where.*''


100 Spitznas and Reale Investigative OphthalmologyFebruary 1975

* • •

Figs. 2 and 3. Fenestrated endothelium of a choroidal capillary. Fig. 2. Freeze-fracture replica:on the left side of the dark line (illuminated area) the central prominence of the diaphragmsbridging the fenestrae is clearly visible and shows a smooth surface. On the right side(shadow area) the central prominence is mostly difficult to identify. Instead, small particlesare recognized, x 128,000. Fig. 3. Granzing thin section: the central thickenings are seen inthe majority of the diaphragms. Where they are missing, this is probably due to the sectioningplane. xl02,000.

Results

In replicas of freeze-cleaved specimens,capillaries and venules of the choroid wereidentified. The choroidal capillaries are easi-ly recognized because of their structuralcharacteristics and their topographic rela-tionship to Bruch's membrane (Fig. 1),the venules because of their proximity tothe capillaries, and their large diameterwith only one smooth muscle cell layerin their wall.12

In en face views, the endothelial cellsof both capillaries and venules have acentral thickened soma containing the

nucleus and elevated cytoplasmic rays ra-diating to a peripheral rim as describedby Friederici13 in the endothelial cells ofmice renal papilla. The valleys betweenthe rays carry numerous fenestrations. Theperipheral rim or marginal fold overlapsfor a varying distance with that of theadjacent endothelial cells, thus forming aserpiginous suture between those cells.

The cleavage plane can run parallel orperpendicular to the surface of the cellor obliquely through the cell. Mainly inthe first direction of cleaving, fenestrationsare observed in the capillary endothelial


Volume 14\'timber 2

Fenestrations of endothelial cells 101

cells. After parallel cleaving, many suchpores (diameter 600 A) show a diaphragmwith a central thickening (Fig. 2), mea-suring 300 A and corresponding to thecentral knob seen in thin sections (Fig.3). On freeze-cleaved specimens visualiza-tion of the central thickening probablydepends on the angle of shadowing (Fig.2), in thin sections it depends on thecutting plane ( Fig. 3).

Between adjacent endothelial cells, inter-cellular junctions of the zonula occludenstype can be observed. In en face viewsthey are not always visible. If they arevisible, they are seen only on one slope ofand generally in some distance to the inter-cellular suture (Fig. 5). This is caused bythe path of cleavage demonstrated in Fig.4, which is based on Branton's membranecleavage hypothesis.11 When the specimensplits along on intercellular surface mem-brane, the zonula occludens shows up (Figs.4 A and B), whereas it remains hiddenwhen cleavage takes place along the apicalor the basal aspect of the cell membrane(Figs. 4 C and D).

The zonulae occludentes between theendothelial cells consist of parallel strandsof variable length that often follow anangular course (Fig. 7). The strands showmany branchings and anastomoses, thusencompassing irregular areas of differentsize (Figs. 5 through S). In en face views,the strands are seen to extend along theintercellular suture. They can be numerousat places, but entirely missing for shortdistances in adjacent areas, as illustratedin Fig. 6 in the wall of a venule. There-fore, in these vessels, the belt formed bythe zonula occludens is discontinuous. Thestrands result from linear level differencesthat, depending upon the path of cleavage(Figs. 4 A and B), appear as ridges orgrooves. If the cleavage path follows thecourse indicated in Fig. 4 A, ridges (faceA), if it follows the course shown in Fig.4 B, grooves (face B) are seen. In caseswhere both ridges and grooves appear onone picture (Fig. 8), they are separatedby a fracture line corresponding to the

Fig. 4. Possible course of the cleavage plane(dashed line): schematic drawing of a junctionbetween two endothelial cells as seen in a sectionperpendicular to the endothelial surface. The sec-tions across the strands of the zonula occludensare represented by three black dots. In fracturea the particles of the strands are not removedfrom the exposed face and appear as ridges onface A of the replica. In fracture b the particlesof the strands are split away from the exposedface and will be seen as grooves on face B ofthe replica. In the fracture courses illustrated inc and d, only the inner faces of the plasma mem-brane at the apical or basal side of the twoadjacent endothelial cells are exposed. The strandsof the zonula occludens are hidden.

intercellular space. Here, the fracture jumpsacross the junction from one cell to theother.

Typically, the zonulae occludentes arelocated between the overlapping extensionsof adjacent endothelial cells. They becomeexposed when the cytoplasm of such cellsis broken at different levels (Figs. 9, Aand 10).


102 Spitznas and Reale Iniyestigntioe OphthalmologyFebruary 1975

Figs. 5 through 8. Fracture faces of zonulae occludentes between endolhelial cells of a

choroidal venule. Fig. 5. Extensive zonula occludens (arrows). In most places the strands arcbranching and interconnected; in limited areas, however, they are entirely missing. xl7,600.

Fig. 6. Higher magnification of delineated area of Fig. 5 with interruption ( = arrow) orthe zonula occludens. x46,000.Fig. 7. Typically angulated strands of a zonula occludens, face B. x42,000.Fig. 8. Zonula occludens with cleavage plane jumping from one endothelial cell to another,exposing grooves on face B (left) and ridges on face A (right). ^54,000.

Frequently, the fracture plane runs o-bliquely through the cells so that the suturearea that becomes exposed is rather limited.Also in those cases, typical zonulae occlu-dentes are observed (Fig. 11). However,the extension of these structures is so

limited that no conclusions can be drawnas to the completeness of the belt.

Further observations include formationswith the structural characteristics of gapjunctions. In our material, such junctionswere found only in the plasma membrane


Volume 14Number 2


Figs. 9 through 11. Junctions between endothelial cells of a choroidal capillary. Fig. 9. Forthe most part, the cell membranes are split along the nonluminal aspect of the endothelialcells (compare Fig. 4, D). Finally, the cytoplasm of the cells is broken at different levelsso that face A of the zonnla occlndens is exposed (compare Fig. 4, A), x28,000.Fig. 10. Higher magnification of the zonula occludens of Fig. 9. *112,000.Fig. II . Most frequent aspect of the zonula occludens (horizontal arrow) between endo-thelial cells, exposing only some features of the strands. BM = Bruch's membrane, Er =erythrocyte. To the right of the vertical arrow several endothelial fenestrations are visible.x43,000.


104 Sjritznas and Reale Inocslinative OphthalmologyFebruary 1975

I *

Figs. 12 and 13. Cap junction on the plasma membrane of a choroidal capillary endothelialcell. The junction lies close to an intercellular suture. *62,000 and xl60,000.

of the capillary endothelium. They aremuch less numerous than the zonulae oc-cludentes to which they do not necessarilyshow a topographic relationship, and arelocated on the scleral side of the endo-thelium in the immediate vicinity of theintercellular suture. In face A, the gapjunction is composed of closely packedparticles with a diameter of about 100A (Figs. 12, A and 13), in face B, of com-plementary pits.

Discussion

The most interesting observations of thisstudy concern (1) the fenestrations of theendothelial cells in the choriocapillaris, (2)the zonulae occludentes of these endo-thelial cells and of the adjacent venules,and (3) the gap junctions seen on thesplit membranes of endothelial cells onthe scleral side of the choroidal capillaries.

Fenestrations. The existence of a dia-phragm across the pores of the choroidalcapillaries, separating the vascular and theintercellular compartments has been de-scribed previously by others1- - in thinsections. Later, a central thickening re-sembling the one seen in renal capillaries15

was described.1'fl In replicas of freeze-fractured endothelial cells of renal capil-laries of mice,"'1 l7 the diaphragm appeal's

as a smooth, membrane-like septum "withoftentimes—but not always—a small centralprominence or knob. Occasionally, this sep-tum was absent, exposing a granular sur-face.""1 Maul17 saw an internal structureof the fenestrations consisting of a ring andradiating structures. In the replicas of ourspecimens, the exposed faces of both thediaphragm and the central prominencewere mostly smooth. Only rarely werethey seen to bear unevenly distributedcoarse particles. Similar observations wererecently made on freeze-fractured capil-laries of pancreas and intestinal microvilliof rats.1* The central prominence is clear-ly seen only in areas with a favorable anglebetween exposed region of the endothelialcell and the source of the shadowing ma-terial. In Fig. 2, for example, the centralknobs of fenestrae in the shadow regionof the replica are hardly visible. Corre-spondingly, the appearance of the centraldensity in thin sections depends on theplane of sectioning.

According to recent morphometric studiesof Simionescu, Simionescu, and Palade,18

the fenestrae "are the best candidates forstructural equivalents of the large-pore sys-tem" postulated by physiologists,19 providedthat "not all the fenestrae function as largepores at all times."


Volume UNuinhcr 2


Zonulae occludenles. A fusion of theouter leaflets of the plasma membrane ofadjacent endothelial cells in the chorio-eapillaris has been demonstrated in thinsections.1"' In corresponding locations, rep-licas of freeze-fractured specimens showinterconnected ridges on face A and grooveson Face B, exhibiting the typical charac-teristics of zonulae oecludentes.-" At places,however, ridges and grooves are missing.Similar zonulae occludentes have been de-scribed in capillaries of the lung and heartmuscle of rats and are considered macu-lae or fasciae, i.e., discontinuous, ratherthan zonulae, i.e., continuous, occludingjunctions.'-1

The occurrence of discontinuous junc-tions of the occludens type between theendothelial cells of the fenestrated cho-roidal capillaries suggests a site of per-meability not only at the fenestrationsbut also at the intercellular space. Whilethe permeability of the fenestrations wasdemonstrated repeatedly with tracer ma-terials such as horseradish peroxidase,-'-'- -:i

fcrritin, and thorium dioxide,'" -' experi-mental evidence for a paracellular route islacking so far. Considering the limitationof both extension and Frequency of theinterruptions in the zonulae occludentes ofthe endothelial colls of the choroida! capil-laries, the actual observations of the pas-sage of relatively large tracers such asperoxidase through the intercellular spacewould be a rare and accidental finding.It is conceivable, however, that substancesof smaller molecular size prefer this routeof passage.

According to Simionescu, Simionescu,and Palade,ls who examined thin sectionsof Fenestrated visceral capillaries, the in-terruptions of the zonulae occludentes couldrepresent the morphologic equivalent ofthe small pore or slit system postulatedbeside the large pore system by physiolo-gists.1" Our demonstrations in the chorio-capillary endothelium of zonulae occlu-dentes, which are discontinuous and, there-fore, belong to the "very leaky" type ofjunctions in the morphologic sense," seems

to support this physiologic view. A com-parison with the estimations of the smallpore system made by physiologists,'-1'""7

however, would require quantitative dataon the total area covered by interruptions.

The occurrence of discontinuous occlud-ing junctions also in the endothelium ofchoroidal venules provides a morphologic,basis for the fact that they are "physio-logically more permeable than the capil-laries" as stated by Majno.'-'s So far, thepenetration of tracers in choroidal vesselslarger than the capillaries has not beendescribed.

Cap junctions. An infrequent but inter-esting finding was the occurrence of struc-tures fulfilling the morphologic criteria of"gap junctions"-'" in the split-cell membranesof the choriocapillary endothelium. In ourmaterial, they were not seen between themeshes formed by the anastomosing strandsof the zonula occludens, a location typicalin the epithlial cells of other organs.-" Inthe choriocapillaris they are located nearthe basement membrane on the scleral sideof the vessels. They can be found be-tween adjacent endothelial cells as wellas between the approximations of pericytesand endothelial cells demonstrated in thinsections of the choroida] capillaries byMatsusaka.1 The finding of gap junctionsbetween endothelial cells and surroundingcells of the vessel wall is not unique, sinceHiittner and co-workers:1"' :u were able todemonstrate their existence between ar-terial endothelium and adjacent smoothmuscle cells using the colloidal lanthanumtechnique. Gap junctions are looked uponas sites of ionic or electrotonic as well asmetabolic coupling of neighboring cells.'-• n

Therefore, our findings indicate that endo-thelial cells and pericytes form a coherentsystem.

In an effort to gain Further insight inthe Functional significance oF the structuresdescribed and discussed above, compara-tive freeze-Fracture studies on choroidaland retinal capillaries of rabbit and manare presently carried out in our labora-tories.


106 Spitznas and Reale Investigative OphtlwlmologyFebruary 1975

REFERENCES

1. Missotten, L.: Etude cles capillaires de larefine et de la choriocapillaire an microscopeelectronique, Ophthahnologica 144: 1, 1962.

2. Carron, L. K.: The infrastructure of theretinal pigment epithelium with observationson the choriocapillaris and Bruch's mem-brane, Trans. Am. Ophthalmol. Soc. 61:545, 1963.

3. Matsusaka, T.: Ultrastructural differences be-tween the choriocapillaries and retinal capil-laries on the human eye, |ap. J. Ophthalmol.14: 59, 1970.

4. Hogan, M. J., Alvarado, J. A., and VVeddell,f. E.: Histology of the Human Eye. An Atlasand Textbook, Philadelphia, 1971, VV. B.Saunders, Co., p. 370.

5. Spitznas, M.: The fine structure of thechorioretinal border tissues of the adulthuman eye, Adv. Ophthalmol. 28: 78, 1974.

6. Bernstein, M. H., and Hollenberg, M. [.: Finestructure of the choriocapillaris and retinalcapillaries, INVEST. OPHTHALMOL. 4: 1016,

1965.7. Farquhar, M. C , and Palade, C. E.: Junc-

tional complexes in various epithelia, |. CellBiol. 17: 375, 1963.

8. Fromter, E., and Diamond, f.: Route ofpassive ion permeation in epithelia, NatureNew Biol. 235: 9, 1972.

9. Claude, P., and Coodenough, D. A.: Fracturefaces of zonulae occludentes from "tight"and "leaky" epithelia, |. Cell Biol. 58: 390,1973.

10. Moor, H., and Miihlethuler, K.: Fine structurein frozen-etched yeast cells, J. Cell Biol.17: 609, 1963.

11. Dowell, W. C. T.: Die Entwicklung geeignet-er Folien fiir elektronenmikroskopische Prii-parattriiger gro^en Durchla^bereiches und ihreVerwendung zur Untersuchung von Kristal-len, Optik 21: 47, 1964.

12. Hogan, M. |., and Feeney, L.: Election mi-croscopy of the human choroid. III. Theblood vessels, Am. |. Ophthalmol. 51: 1084,1961.

13. Friederici, H. H. R.: The tridimensionaliiltrastructure of fenestrated capillaries, f. Ul-trastruct. Hes. 23: 444, 1968.

14. Branton, D.: Fracture faces of frozen mem-branes, Proc. Nat. Acad. Sci. 55: 1048, 1966.

15. Rhndin, J. A. C : The diaphragm of capH-lary endothelial fenestrations, J. Ultrastruct.Res. 0: 171, 1962.

16. Friederici, H. H. R.: On the diaphragm acrossfenestrae of capillary eiidotheliiim, |. Ultra-struct. Res. 27: 373, 1969.

17. Maul, C. C : Structure and formation of

pores in fenestraled capillaries, |. Ultra-struct. Res. 36: 768, 1971.

18. Simionescu, M., Simionescu, N., and Palade,C. E.: Morphometric data on the endotheliumof blood capillaries, J. Cell Biol. 60: 128,1974.

19. Landis, E. M., and Pappenheimer, J. R.:Exchange of substances through the capillarywalls, in: Handbook of Physiology, Section2: Circulation, Hamilton, W. F., and Dow,P., editors. Washington, D. C, 1963, Ameri-can Physiological Society, vol. 2, p. 1013.

20. McNutt, N. S., and Weinstein, R. S.: Mem-brane iiltrastructure at mammalian intercellu-lar junctions, Progr. Biophys. Mem. Biol. 26:45, 1973.

21. Weinstein, R. S., and McNutt, N. S.: Elec-tron microscopy of freeze-cleaved and etchedcapillaries, in: Microcirculation, Perfusion, andTransplantation of Organs, Malinin, T. I.,Linn, B. S., Callahan, A. B., and Warren,W. D., editors. New York, 1970, AcademicPress, p. 23.

22. Peyman, C. A., Spitznas, M., and Straatsma,B. R.: Peroxidase diffusion in the normaland photocoagulated retina, INVEST. OPHTHAL-MOL. 10: 1S1, 1971.

23. Peyman, G. A., and Bok, D.: Peroxidasediffusion in the normal and laser-coagulatedprimate retina, INVEST. OPHTHALMOL. II:35, 1972.

24. Cunha-Vaz, J. C , Shakib, M., and Ashton,N.: Studies on the permeability of the blood-retinal barrier. I. On the existence, develop-ment, and site of a blood-retinal barrier,Br. J. Ophthalmol. 50: 441, 1966.

25. Crotte, C : Passage of dextran moleculesacross the blood-lymph barrier, Acta Chir.Scancl. Suppl. 211:1, 1956.

26. Carby, L., and Areekul, S.: A method forestimation of the "leak"-to-"pore" inimberratio in capillary membranes, in: CapillaryPermeability, the Transfer of Molecules andIons Between Capillary Blood and Tissue,Crone, Ch., and Lassen, N. A., editors. Pro-ceedings of the Alfred Benzon SymposiumII, Copenhagen, 1970, Munksgaard, p. 306.

27. Lassen, N. A., and Trap-Jensen, J.: Estima-tion of the fraction of the inter-endothelialslit which must be open in order to accountfor the observed transcapillary exchange ofsmall hydrophilic molecules in skeletal musclein man, in: Capillary Permeability, the Trans-fer of Molecules and Ions Between CapillaryBlood and Tissue, Crone, Ch., and Lassen,N. A., editors. Proceedings of the AlfredBenzon Symposium II, Copenhagen, 1970,Munksgaard, p. 647.

28. Majno, C : Infrastructure of the vascular


V ill mm; 14Nn in her 2

Fenest rations of endothelinl cells 107

membrane, in: Handbook of Physiology, Sec-tion 2: Circulation, Hamilton, VV. F., andDow, P., editors. Washington, D. C, 1965,American Physiological Society, vol. 3, p.2338.

29. Friend, D. S., and Cilula, N. B.: Variationsin tight and gap junctions in mammaliantissues, J. Cell Biol. 53: 758, 1972.

30. Hiittner, I., Bouter, M., and More, R. H.:Cap junctions in arterial endotheliuin, |.Cell Biol. 57: 247, 1973.

31. Hiittner. I., Boutet, M., and More, II. H.:Studies on protein passage through arterialendotheliuin. I. Structural correlates of per-meability in rat arterial endothelium, Lab.Invest. 28: 672, 1973.

32. Cilula, N. B., Reeves, O. R., and Steinbach,A.: Metabolic coupling, ionic coupling, andcell contacts, Nature 235: 262, 1972.

33. Bennett, M. V. L.: Function of electrotonicjunctions in embryonic and adult tissues, Fed.Proc. 32: 65, 1973.


Fracture faces of fenestrations and junctions of endothelial cells in ...

Documents

Transcript of Fracture faces of fenestrations and junctions of endothelial cells in ...