AN ATLAS OF THE CHOROlDAL ULTRASTRUCTURE W‘TH THE HISTOCHEMICAL LOCALIZAHON 0F CHOROIDAL AND PSEUDOBRANCHlAL CAR-BONI‘C ANHYDRASE AG'flVlW . N THE RAINBOW TROUT (SALMO GARDNER!) ' Thesis for the Degree of M .. S. MICHlGAN ST ATE. UNIVERSITY GRAIG E. ELDRED 1975 1 H5518 LIBR/iR Y r _. Auchgm Siam Umvetsity —.- *A IJBW BINDERY mg. 1', :{I Cum? BINDERS :3 .1..‘:nr_mmcn» Eli ABSTRACT AN ATLAS OF THE CHOROIDAL ULTRASTRUCTURE ,5.) WITH THE HISTOCHEMICAL LOCALIZATION ,, 0F CHOROIDAL AND PSEUDOBRANCHIAL CARBONIC ANHYDRASE ACTIVITY IN THE RAINBOW TROUT (SALMO GAIRDNERI) By Graig E. Eldred A specifddz site of carbonic anhydrase activity within the effereni;xressels of the choroidal rete mirabile had been speculaflxui for the prevention of short-circuiting of the oxygen.nnfiLtiplication mechanism by the diffusion of carbon dioxide from the efferent to afferent retial vessels. It had further been speculated that the pseudobranch may secrete carbonic anhydrase for use at this site. With the development of a specific histochemical method for demon— stration of carbonic anhydrase at an electron microscopic level, the localization of specific sites of carbonic anhydrase activity within the choroidal vasculature and pseudobranch of the rainbow trout, Salmo gairdneri, became feasible. An atlas of choroidal vascular pattern and ultrastruc- 1mre has been compiled and presented in Part I. In working vuth vascular tissues,particularly in fishes. special Imecautions must be taken in order to avoid clot formation. Specific examples from the existing literature in which this had been neglected are cited, and the importance of the tissue preparatory technique upon the appearance of Graig E. Eldred the ultrastructure is stressed. The GXiS‘tence of a possible mechanism for the neural control of flow through the counter-current multiplier via the arterial manifold is identified. The packing; pattern of the retial vessels is cubic, such that there is an equal number of afferent and efferent vessels. Pvt 811 ultrastructural level, the endothelial cells of the affeandt and efferent retial vessels are structurally practically identical, yet the afferent vessels are clearly distinguisfluible as having associated pericytes, a distinct basement nundbrane, and a smaller lumenal diameter. No structural differences could be detected distinguishing distribution.from collection vessels. The choriocapillaris region has three distinguishing features: 1) the endothe- lium is relatively free of fenestrae, 2) Bruch's membrane lacks the complexity seen in higher vertebrates, and 3) the pigment epithelium lacks a highly involuted basal border. Light and electron microscopic carbonic anhydrase histochemistry was pursued in Part II. Problems with visu— alization of the cobalt sulfide reaction product were encountered on the electron microscopic level, but not on the light microscopic level. The normal ultrastructure of the pseudobranch-type cell was found to be similar to those investigated else- where. Conclusions are drawn regarding the possibility Graig E. Eldred that pseudobranchial secretion of carbonic anhydrase may cmcur for subsequent usage in the choroidal counter—current oxygen multiplier . Carbonic anhydrase activity within the choroidal rete mhmbile is found to increase in intensity along the length OfIme afferent retial vessels and is seen to extend for smmadistance into the peripheral distribution vessels. fiwsesfltes of activity are consistent with a short-circuit nmdalfbr oxygen multiplication. AN ATLAS OF THE CHOROIDAL ULTRASTRUCTURE WITH THE HISTOCHEMICAL LOCALIZATION OF CHOROIDAL AND PSEUDOBRANCHIAL CARBONIC ANHYDRASE ACTIVITY IN THE RAINBOW TROUT (SALMO GAIRDNERI) By Graig E. Eldred A THESIS Submitted to . Michigan State University 1“ Partial fulfillment of the requirements for the degree of MASTER OF SCIENCE Department of Physiology 1975 DEDICATION To my parents who from the start provided the environment which made 'the acquisition of knowledge a desire, not a task'..... TO'Um Vosses whose generosity permitted me to begin when I did..... To Barbara whose chronic case of epistemophilia provided her with the strength and competitive- ness which forced me to do and achieve where I may not have otherwise..... TO‘Hm yet.to become whose beginnings spurred me on to completion..... And! 130 meooQOOooooooooooooooooooonWhy T1013? AC KNOWLEDGMENTS I Wish to reiterate my feelings of gratitude to all those persons whose insistent assistance has taught me the Pitfalls of total self-reliance and the real value of cooperative help. Included are: Professors J.R. Hoffert, P.O. Fromm, and R.P. Pittman Who served as members of my committee. Jerry Friedman for help in the darkroom with the devel- 0Pment of all photographs herein. Professor P.O. Fromm for his alert perusal of the literature which frequently led me to papers which would otherwise have escaped my attention, and for his modest words of encouragement regarding the true value of electron microscopy as a research tool. Esther Brenke for the feat of reading red ink hand— Writing and translating it into legible type. Professor J.R. Hoffert, first, for the task of reading and evaluating the original manuscript, and second, for the awesome taSk of labelling 366 prints by hand. I am sure that this latter accomplishment will never be forgotten, for he Will most certainly and deservedly offer continual reminder-S of the fact that I am forever indebted to him. Additional thanks go to Esther and Jack for their labors of sectioning and staining ocular tissues for light mic . roscopy at a time when I was supposed to be in bed With iii infectious mononucleosus, but was caught in the act of seWing lace on a blouse for my wife with a friend. Finally, acknowledgements are due the National Insti- tutes of Health, Grant EY-OOOO9 VIS, and the Barbara Jo H0Ppe‘Eldred Support Foundation for their continued fhmncial assistance through the 2.5 years of this ordeal. iv TABLE OF C ONTENTS IJST OF FIGURES . . . . . . . . . . . . . . . . . GMWRAL INTRODUCTION . . . . . . . . . . . . . . PART I INTRODUCTION . . . . . . . . . . . . . . . . . . LITERATURE REVIEW . . . . . . . . . . . . . . . . General Vascular Patterns . . . . . (Hmubidal and Retial Ultrastructure Swimbladder Rete . . . . . . . Choroidal Rete . . . . . Choriocapillaris . . . . Ultrastructure Summary . MATERIALSANDMETHODS.............. RESULTS A light Nficroscopic Choroidal Morphology . . . Electron Microscopic Fixation Techniques . . Electron_Microscopic Choroidal Morphology . . DISCUSSION Vascular Artifacts . . . . . . . . . . . . . (HmroidaJ.Ultrastructure and Function . . . . CONCLUSIONS PART II INTRODUCTION LITERATURE REVIEW . Theliistorical Elucidation of Retial Function gfle Rete Mirabile as an Exchanger . . . . . . The Rete Mirabile as a Multiplier . . . . . . Care~bROle of Carbonic Anhydrase . . . . . . . (”110 Anhydrase Histochemistry . V Page vii 10 10 13 16 18 20 23 23 26 31 3]. 36 SO 52 57 57 59 61 62 Page NRTERIALS AND METHODS . . . . . . . . . . . . . . . 69 RES ULT S . o o o n o n o a o o o o o O O I 0 Light Microscopic Carbonic Anhydrase Histochemistry . . . . . . . . . . . . . . 72 Choroidal Electron Microscopic Carbonic Anhydrase Histochemistry . . . . . . . . . 73 Electron Microscopic Pseudobranch MorphOlOgy O C O O O I C I I O O I O O O I 76 Pseudobranch Carbonic Anhydrase Histochemistry . . . . . . . . . . . . . . 78 DISCUSSION I O O O O O O O O O O O O O O I O O I C I 80 Interpretation of Carbonic Anhydrase Histochemistry . . . . . . . . . . . . . . 8O Choroidal Carbonic Anhydrase . . . . . . . . . . 83 FUnctional Implications . . . . . . . . . . . . 8A SUMMARY AND CONCLUSIONS . . . . . . . . . . . . . - 89 COMBINED SUMMARY AND CONCLUSIONS . . . . . . . . . . 91 QTChniques Findings . . . . . . . . . . . . . . 91 morPhological Findings . . . . . . . . . . . . . 92 Cannaic Anhydrase Histochemistr . . . . . . . 94 RECOMMENDATIONS 95 LITERATURE C ITED o o o o I o o o o o o a o o 0 O I 0 99 APPENDIX. . . . . 10? vi LIST OF FIGURES FIGURE Page 1. Microfil casts of the afferent vessels of the Choroidal vasculature. . . . . . . . . . . . . . . 108 2. Generalized diagram of the choroidal vascular pattern within the eye of Salmo gairdneri. . . . . 110 3- Choroid rete mirabile (hematoxylin and eosin, luSX) I I I I I I I I I I I I I I I I I I I I I I I 112 A. Parallel rete vessels (Masson's trichrome, 1,48OX) ' I I I I I I I I I I I I I I I I I I I I I 111+ 5- Central side of the rete mirabile (hematoxylin and 808.111, lu5X) I I I I I I I I I I I I I I I I I 116 6- Central side of the rete mirabile (hematoxylin and eosin, AAOX). . . . . . . . . . . . . . . . . 118 7- Central side of the rete mirabile (Masson's trlChrome, 295X) I I I I I I I I I I I I I I I I I 120 8- Central side of the rete mirabile (Masson's trlChrome, 145X) I I I I I I I I I I I I I I I I I 122 9. Peripheral side of the rete mirabile (hemaiwxylin and eosin, 1A5X). . . . . . . . . . . 12A 10. Vessels of the peripheral region (hematoxylin and eOSin, lLA‘SX) I I I I I I I I I I I I I I I I I 126 11' Peripheral collection and distribution vessels (hematoxylin and 808111, LALI’OX)I I I I I I I I I I I 128 l2. Choroidal rete mirabile cross—section (methylene blue 9 925x) 0 I I I I I I I I I I I I I I I I I I I 130 13' §tandard immersion-fixed rete in cross-section 18’ 975X) I I I I I I I I I I I I I I I I I I I I I 132 14. Standard immersion-fixed rete: .AreaS‘With mYeloidconfigurations in the lumen. . . . . . . . 134 15. vsvtandard immersion-fixed rete: Area with thin- Efihled vacuoles within the lumen (10,240X). . . . 136 vii LIST l6. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. OF FIGURES—-Continued Page Myelinated and non-myelinated neural elements within the arterial manifold wall (26,250X). . . . 138 Smooth muscle layer underlying the endothelial border of the arterial manifold wall (27,19OX). . 140 Arterial manifold wall displaying an association between pinocytotic activity within the endothelium and smooth muscle (26,250X). . . . . . . . . . . . 142 Heparinized, immersion—fixed choroidal rete mirabile in cross-section (8,4OOX). . . . . . . . 144 Afferent retial vessel in cross-section (20,755X). . . . . . . . . . . . . . . . . . . . . 146 Afferent versus efferent rete endothelial ultra— Structure (113’ 750X) I I I I I I I I I I I I I I I 114'8 Vessels of the peripheral choroidal layer (methylene blue, 925x). . . . . . . . . . . . . . 150 Peripheral choroidal vessel ultrastructure (14 ' 175x) I I I I I I I I I I I I I I I I I I I I I 152 Peripmeral vessel wall (25,625X). . . . . . . . . 154 Chqriocapillaris, Bruch's membrane, and pigment eplthelium (13, 81+OX) I I I I I I I I I I I I I I I 156 Fenestmated choriocapillaris endothelium. . . . . 158 Bruch's membrane (41,405X). . . . . . . . . . . . 160 MyelOid.bodies within the pigment epithelium (16,560X) I I I I I I I I I I I I I I I I I I I I I 162 Photoreceptor outer segment (20,755X). . . . . . . 164 Choroidal rete mirabile carbonic anhydrase stain. 166 ChorOidal rete mirabile carbonic anhydrase stain (295X) I I I I I O I I I I I I I I I I I I I I I I 168 geripheral region of the choroidal rete stained up carbonic anhydrase (145x). . . . . . . . . . . 170 CaIWM3nic anhydrase activity within peripheral COllectiOn and distribution vessels (295X)- - - - 172 viii LIST OF FIGURES--Continued 34. Polarization of carbonic anhydrase activity within the choroidal rete vessels (65X). . . . . 35. General appearance of carbonic anhydrase stained tissues under the electron microscope (25,625X). 36. Carbonic anhydrase activity within the choroidal rete endothelial cells (66,25OX). . . . . . . . 37. Choroid rete endothelial cell carbonic anhydrase activity as compared with erythrocyte carbonic anhydrase activity (102,500X). . . . . . . . . . 38. Banding pattern of carbonic anhydrase activity within the choroid rete endothelial cell (102’5OOX)0 I I I I I I I I I I I I I I I I I I 39. A differential carbonic anhydrase activity in afferent versus efferent retial vessels (102’5OOX)I I I I I I I I I I I I I I I I I 40. The "pseudobranch—type" or acidophilic cell (9’995X)I I I I I I I I I I I I I I I I I I I I 41, The mitochondrial-smooth endoplasmic reticulum network of the pseudobranch—type cell (25,625X). 42' Basal border of the pseudobranch-type cell (82,000X)I I I I I I I I I I I I I I I I I I #3' 3Pseudobranch—type cell carbonic anhydrase Ilistochemistry (32,030X). . . . . . . . . . . . 44, Ckarbonic anhydrase stained pseudobranch-type (Bell basal border (102,500X). . . . . . . . . . 45' 'Vritres flamboyants final": Carbonic anhydrase Ertained pseudobranch-type cell mitochondrial- Efladoplasmic reticulum network (66,250X). . . . . ix Page 174 176 178 180 182 184 186 188 190 192 194 196 GENERAL INTRODUCTION Oxygen tensions at the retial-vitreous interface in teleost fishes are phenomenalogically high (Wittenberg and Wittenberg, 1961; Wittenberg and Wittenberg, 1962). In the rainbow trout, Salmo gairdneri. oxygen tensions at this site reach 400 mm Hg which is ten to twenty times arterial levels (Fairbanks, Hoffert and Fromm, 1969). These tensions areesmablished by the operation of a morPhOlOgiCally dif- ferentiated network of parallel vessels within the choroid laymrof 1flie teleost eye-——the choroid rete mirabile. anmr, erffert and Fromm (1973) have established that higloxygeui tensions and hence the functional integrity Of‘fim re1xirmq are dependent upon carbonic anhydrase. A "smut-cirwyuit" model has recently been proposed by Fairbanks, Hmfikrt arui Fromm (1974) in which the counter-current ex- <fimngm¢(3f- oxygen in this system relies upon a localized. highly active focus of carbonic anhydrase activity, pre— mmmbly jJfi ‘the efferent side of the system. Its presence here, as distinct from the foci of activity present in the retina and red blood cells has yet to be demonstrated. In Inostteleosts, the choroid rete mirabile receives oxygenated blood directly from the gills with only a small veS‘tigial gill structure, the pseudobranch, inter- posed betWeen them. The pseudobranch has exceptionally high 1 evels of Carbonic anhydrase activity (Leiner, 1938; 1 Maetz, 1956; Hoffert, 1966). The capillary network within the pseudobranch is surrounded by very morphologically dis- tinct, "pseudobranch-type" cells (Harb and Copeland, 1969). In as much as the functional significance of the pseudo— branch had eluded investigators for years, Copeland (1951) Proposed that it may secrete carbonic anhydrase into the blood stream for use in the concentration of oxygen by the SWimbladder rete mirabile and Fairbanks _e_1_'. _a_1. (1969) also Proposed the same function, this time with the choroid rete oxygen multiplier as an additional site of activity. Laurent and Rouzeau (1972) demonstrated the presence of barorecep‘tor elements and chemoreceptor elements in the pseudobranch; the chemoreceptors responding to oxygen ten— Sions, PH. and sodium ion concentration of the blood. This neural receptor activity, however, accounts for neither the apparent Secretory cellular specializations of the pseudo- branCh‘tL’pe cell, nor the high levels of carbonic anhydrase aetiViW reported. Thus. the hypotheSiS PrOPOsed by Fairbanks g 31' (1969) remains a viable theory. H"il’l's‘vson (1967) introduced a light microscopy technique for the histological staining of carbonic anhydrase activity utilizing a cobalt sulfide precipitate. Because cobalt sulfide is electron dense as well as light opaque, the tech- nlque has recently been improved and used at the electron mi . croscoch level by Rosen and Musser (1972)- In the current study, through the use of this carbonic anhydrase histochemical technique, at both light and electron microscopic levels, two objectives are sought. First, the Possible presence and precise locus of an efferent choroidal carbonic anhydrase activity will be studied in order to Provide either supportive or contradictory evidence to the Short-circuit model of Fairbanks gt al. (1974). Secondly, 1She localization of carbonic anhydrase activity within the Pseudobranch of Salmo gairdneri is attempted, in hopes of correlating its ultrastructural location with morphological eVidence for secretory activity, and in so doing lend credence to the hypotheses of Fairbanks et all. (1969) and COPeland (1951). In order to achieve these objectives, however, it is first necessary to precisely define the vascular patterns and morphological features which distinguish the structures 01" interest here. The pseudobranch ultrastructural morphol- ogy had Previously been extensively described (Copeland and Dalton, 1959; Holliday and Parry, 1962: HELP‘0 and Copeland, 1969), so that it was merely necessary to re- affirm these findings in Salmo gairdneri. However, the choroid Vascular morphology has not been extensively described in teleosts, particularly at an ultrastructural level- Copeland (1974a) and Wittenberg and Wittenberg (1974) haVe published studies in which gross vascular patter . n8 in several teleost species and preliminary Observations on rete ultrastructure are described. The apparent interspecific variability in gross morphology and remaining paucity of definitive ultrastructural description deemed it necessary to catalogue an atlas of choroidal vas- cular ultrastructure before proceeding to the histochemical localization of carbonic anhydrase. In Part I of this study, a description of the normal Choroid vasculature in Salmo gflrdneri is presented along With a comparative discussion of the functional significance of each vascular segment. Part II is concerned with the localization of carbonic anhydrase in the choroid and Pseudobranch of this species. It is hoped that a much clearer anatomical and func- tional understanding of the teleost choroidal vascular ultrastructure and an important extension of the evidence for the possible role of carbonic anhydrase in the choroidal Oxygen CGunter—current exchanger will be gained through this study. Further, important problems and precautions in microvasoular fixation procedures will be presented which have heretofore gone undefined, but which should be taken into c . o o C)nslderation in future ultrastructural research on any VaSCular bed . PART I INTRODUC T ION The circulatory system is designed to supply the body tissues with blood in amounts commensurate with their meta— bolic requirements. To accomplish this most efficiently, the microvasculature of each tissue and organ must be Specialized in 1) its microvascular geometry, 2) the spatial relationship between tissue cells and capillaries, and 3) the structural nature of the capillary endothelium- Vertebrate retinal tissue has the highest rate 0f res— piration of any tissue in the body comparable only to» and often exceeding, that of cancer tumor tissue and embryonic tissue (Noell, 1959). In keeping with the needs of this metabolic demand, the mammalian ocular tissue is supplied with ”CW0 microvascular beds-—-one in the choroid, and one in the retina. Unlike the mammalian retina, however, the teleost retina is avascular. With the exception of the minor Vascular contributions made by the lentiform body and the falciform process (Copeland, 1974b) the entire metabolic needs of the teleost retina must be met by the circulation 0f the C:horoidal layer. In meeting these metabolic needs and in keeping with the notion that over-all tissue and organ function is funda— mental t0 and a determinant of vascular architecture, the teleos-t choroidal vasculature has evolved into a structure-- the ch , orold rete mirabile—--capable of the counter—current 5 [Elna-c" _ _ _I exchange of oxygen (Wittenberg and Wittenberg, 1962; Fairbanks, Hoffert and Fromm, 1974). A notable exception is seen in the eels, Anguilloidei, which lack the choroid rete (Wittenberg and Haedrich, 1974), but which are also unique among the teleosts in having vascular retinas (Francois and Neetens, 1974). Besides its role in oxygen concentration, the choroid rete has also been implicated in the counter-current exchange of heat (Linthicum and Carey, 1972). The possibility remains that passive diffusion of other substances across the endothelia of the exchange system may also occur. Wittenberg and Wittenberg (1974) suggest that diffusion of water may be important in the process of oxygen concentration. Also, the importance of carbonic anhydrase localized within this vascular bed has been demonstrated (Fairbanks 23 al., 1974). Thus, it is essential to define the vascular archi- tectural framework around and upon which these physical and physiological factors must operate. In reviewing the pub- lished evidence and in reporting the observations of the current study, emphasis will be placed upon morphological features classically considered to be indicative of: l) diffusion pathways and transport processes, such as fenes- trations, pinocytotic vesicles, and endothelial junctional areas, all of which may represent pathways for water soluble and macromolecular substances; 2) of barriers to such move- ment, such as basement membranes; and 3) of synthetic 1.1:. “Qt—”L11 activity, such as rough endoplasmic reticulum, polysomes, mitochondria, and Golgi complexes. Several authors have published descriptions of the choroidal network in various teleostean species which have touched the surface of the problem. In the present study, an atlas of the light and electron microscopic structures of this vascular bed in Salmo gairdneri will be compiled. Then, the functional implications of these anatomical features will be discussed in light of evidence presented in the literature for other vertebrate systems. Finally, several important points of controversy with respect to fixation procedures have arisen and will be clarified. LITERATURE REVIEW An historical review of the elucidation of rete structure and physiology has been presented by Born (1961) and Wittenberg and Wittenberg (1974). General Vascular Patterns The choroidal rete morphology has been described on a gross structural basis for the cod, Gadus morhua; the blue— fish, Pomatomus saltatrix; the bowfin, Amia calva; the sword- fish, Xiphias gladius; and the tuna, Lampris regius (Witten- berg and Wittenberg, 1974); the rainbow trout, §almg gairdneri (Barnett, 1951); and the killifish, Fundulus grandis (Copeland, 1974a). According to descriptions thus far presented, a pattern of choroidal vascular flow emerges which is quite different from that seen in the human (Hogan, Alvarado and Weddell, 1971). Blood flows into the teleost choroidal vascular bed via the ophthalmic artery. The ophthalmic artery passes through the sclera along the optic nerve and then branches into a thick-walled arterial manifold which parallels the central margins of the horseshoe-shaped choroidal "gland". This arterial manifold lies within the lumen of a venous sinusoid. The arterial manifold bifurcates into parallel afferent vessels of the rete mirabile (Figure 1, p. l05). These vessels again converge into major "distribution" vessels which go peripherally to supply the choriocapillaris. 8 _ji:::::3 The choriocapillaris lies directly apposed to Bruch's mem— brane and the pigment epithelium and is the presumed site of nutrient and metabolite exchange for the retina. Efferent blood returns from the choriocapillaris via major "collect— ion" vessels which again bifurcate into the efferent rete mirabile vessels running parallel to and interspersed among the afferent vessels in a species specific packing pattern. The blood is collected into the previously mentioned venous sinusoid feeding the ophthalmic vein which passes out of the eye along the optic nerve. Figure 2 illustrates this general pattern. Superimposed upon this pattern is an array of species specific variations. The studies of Copeland (1974a) and Wittenberg and Wittenberg (1974) both demonstrate that the gross morphology of the choroid "gland" itself varies from a horse-shoe shape in most forms to an inverted "Y" seen in the killifish, Fundulus grandis. The killifish also pre— sents a very interesting exchanger conformation (Copeland, 1974a). In F. grandis an intermediate collecting vessel carrying venous blood enters the choroid rete at midstream and bifurcates into an additional set of efferent rete vessels which are also interspersed among and parallel to the rest of the vessels on the central end of the rete. The rainbow trout pattern follows closely that general de- scription presented above (Fairbanks, 1970; Barnett, 1951). lO Choroidal and Retial Ultrastructure In reviewing the choroidal and retial literature on ultrastructure each segment of the vascular bed described above will be considered sequentially. No information on the ultrastructure of the ophthalmic artery and arterial manifold has been located in the literature to date. Most descriptions of rete mirabile ultrastructure have been performed with the purpose of finding characteristics capable of distinguishing afferent from efferent rete vessels. None of these studies, however, are conclusive enough to allow the definition of either vessel type in Salmo gairdneri. Swimbladder Rete One of the only other rete mirabile structures of cap- illary dimensions occurs in the swimbladder rete. Studies on this system are noteworthy in that the retia of these two organs are very similar in both function and gross structure. The vascular ultrastructure in the swimbladder rete has been studied in the toadfish, Opsanus tau, (Fawcett and Wittenberg, 1959; Wittenberg and Wittenberg, 1974); the perch, Perca fluviatilis; and the pond loach, Misgurnus fossilis (Jasinski and Kilarski, 1971); the eel, Anguilla vulgaris (Born, 1961); and in the Whitefish, Coregonus lavaretus (Fahlén, 1967). In the swimbladder rete of Opsanus tag, the afferent and efferent rete vessels are clearly distinguishable from ?,____t} 11 each other (Fawcett and Wittenberg, 1959; Fawcett, 1961). Afferent vessels possess a thick (2—4 um) continuous endo- thelial lining. These cells display the overlapping or interdigitation of their margins that is typical of mam— malian capillaries (Rhodin, 1967 & 1968). In Opsanus tau straight afferent endothelial cell boundaries have con— spicuous desmosomes of varying sizes. Typically there is one large desmosome situated about midway along the boundary with two or more zonulae occludens above and below the principal desmosome (Fawcett, 1966). The afferent vessel endothelia possess an abundant supply of pinocytotic vesicles. In contrast, the efferent vessels of the Opsanus tag swimbladder rete display a distinctive variation in wall thickness with thick nuclear areas (1—3 um) alternating with extremely thin areas (lo—70 nm). Within the attenuated por- tions of the endothelia, there are areas which resemble the diaphragmed fenestrae characteristically seen in the mammali— an kidney, intestinal villi, and endocrine glands (Rhodin, 1974). Fawcett and Wittenberg (1959) point out, however, that in the efferent swimbladder rete vessels of Opsanus, actual pores not closed by a membranous diaphragm are rarely, if ever seen. The endothelial cells exhibit very little vesiculation of the cell membranes. Wittenberg and Witten- berg (1974) point out that both the afferent and efferent capillaries of the Opsanus swimbladder rete do not change in appearance throughout their length, either morphologi- cally or in their packing pattern. 12 Jasinski and Kilarski (1971) reported that afferent and efferent vessels also display separate, distinctive morphologies in the swimbladder rete of the perch, Perca fluviatilis. Afferent rete vessels are apparently very similar to those described in Opsanus tau. They display exceptionally thick endothelia (4 um) with a great abundance of smooth surfaced vesicles of an average diameter of 103 nm and no fenestrae. In the central prominence of each endo- thelial cell, an apparently full complement of typical cyto- plasmic organelles (e.g., mitochondria, Golgi complexes) is present. The endothelial cell junctions do not overlap and have zonulae occludens and desmosomes. In contrast, the efferent rete vessels have relatively thinner walls showing attenuated areas containing both single pores and local thickenings. Smooth surfaced vesicles are present in large numbers in the cytoplasm. In addition, pericytes are associ- ated with both categories of vessels and a basement membrane surrounds the pericytes and endothelia. Collagen fibers are reported in Eegga to be seen occasionally passing through the homogeneous ground substance of the intravascular space. The swimbladder of the pond loach, Misgurnus fossilis (Jasinski and Kilarski, 1971) is slightly different from those described in the perch and toadfish above. Afferent rete vessels in this species show two types of endothelial cells alternating around the vessel periphery. One is a flattened cell (0.5 um thick) containing dark cytoplasm with 13 numerous smooth surfaced vesicles. The others are l-l.5 um thick showing light cytoplasm and a smaller number of smooth surfaced vesicles and also large, electron—lucent vesicles of irregular shape. Endothelial junctions in both vessels show not only zonulae occludens and desmosomes, but also marginal flaps which are arm-like extensions of the endothe- lia into the vessel lumen at the junctional border. The efferent rete vessels are like those previously described in Opsanus and Perca. Choroidal Rete Contrary to the description presented for the swimbladder rete ultrastructure the information presented on the ultra- structural details of the choroidal rete has definitely been lacking until only very recently, and the micrographs pub— lished may be indicative of the reason why more information has not been published. Wittenberg and Wittenberg (1974) have published line sketches of electron micrographs provided to them by D.W. Fawcett at Harvard Medical School, of the choroidal rete mirabile in gala aalxa, a holostean. This was an unfortunate choice in that gala is a relatively primitive fish and cannot be considered as representative of the teleostean form. It may be used as an example, only if it is kept in mind that it may at most represent an example of convergent evolution. In contrast to the morphology of the swimbladder retia pre- sented above, Wittenberg and Wittenberg (1974) point out 14 from their sketches that in the choroidal rete of gala, the difference in thickness of the walls of afferent and efferent rete vessels is very slight. They state that this must mean that some intracellular process is more developed in the afferent swimbladder rete vessels than in the choroid rete endothelia. The sketch presented depicts the afferent ves— sels as being characteristically round in cross-section with four or five endothelial cells bordering the lumen. The endothelial cell junctions appear to show the marginal flap configuration. Surrounding the afferent endothelia are pericytes and enclosing this complex is a basement membrane. The efferent vessels are depicted as lacking pericytes and their basement membranes are apparently distinct from those of the afferent vessels. The most striking feature, however, is the fact that the lumen of the efferent vessel is not circular in cross-section, but sinusoidal in appearance. This feature had not previously been reported in the swim- bladder rete conformation. Neither category of vessel is shown to possess fenestrae. Unfortunately, because only line sketches are presented, no further observations may be made regarding vesiculation, perivascular structure and other subcellular features. Copeland (1974a) subsequently published the only actual micrographs of the rete ultrastructure in Fundulus grandis. He first presents the ultrastructure of a rete prepared by immersion in Karnovsky's fixative (Karnovsky, 1965). ft}... 15 The features seen in these sections are obviously unlike any other capillary endothelia reported elsewhere. The endo- thelia are distinctive in having very large "phagocytotic" vacuoles apparently having been formed by pseudopodia-like extensions of the endothelial wall into the lumen distinct from the marginal flaps at the endothelial cell junctions. There are also numerous smaller pinocytotic vesicles which would perhaps be more typical of what might be expected of capillary endothelia (Bruns and Palade, 1968a; 1968b), except that Copeland points out that these vesicles may be an extensive network of parallel microtubules based upon a longitudinal section through such a suspected area. Finally, the perivascular space seemed to be swollen and basement membranes were not seen surrounding pericyte and endothelial components. In an attempt to obtain more typical photomicrographs, Copeland (1974a) then developed a perfusion technique which yielded what he deemed to be a more acceptable ultra- structural appearance. Here, as in the choroid rete of gala, the efferent vessels in Fundulus are seen to be sinusoidal in nature, not discretely circular in cross-section. Peri— cytes are seen to be associated with the afferent vessels. Vesicles of pinocytotic size are seen included within the thicker areas of both afferent and efferent vessels. There is little difference between the average afferent and ef- ferent endothelial thickness, but areas of the efferent ‘71 l6 endothelia become extremely attenuated, but lack either pores or diaphragmed fenestrae. In general, the appearance of this Fundulus choroidal rete preparation appears similar to that seen in the sketch of the choroidal rete of gala. The evidence regarding choroidal rete ultrastructure seems to point toward features differing from those seen in the swimbladder rete. Because of the variability displayed, all features must be kept in mind in observing the ultra- structure of the choroidal rete of Salmo gairdneri. Choriocapillaris Afferent rete vessels converge into larger distribution vessels which proceed to a vast capillary network, the choriocapillaris, which is the major site of exchange. Large collection vessels drain the choriocapillaris and deliver blood to the efferent rete vessels. No studies have been published on the ultrastructure of either the dis- tribution or collection vessels in teleosts. Braekevelt (1974) has, however, reported on the ultrastructure of the choriocapillaris capillaries and their relationship to Bruch's membrane and the pigment epithelium in the Northern pike, Esox lucius. In both mammals and teleosts, the choriocapillaris abuts Bruch's membrane, a non-cellular supportive structure. The pigment epithelium lies on the retinal side of Bruch's mem- brane. In the human (Hogan, Alvarado and Weddell, 1971), and-mammals in general, the choriocapillaris endothelial 17 cells are polarized in structure such that the nucleus and bulk of the endothelial inclusions (i.e., the central promi— nence) lie toward the scleral side of the capillary. Peri- pheral extensions of the endothelia then extend around the lumen and form the capillary wall adjacent to Bruch's mem- brane. In this region, the endothelial wall is extremely attenuated and in mammals, is highly fenestrated. In humans, Bruch's membrane is composed of: l) the choriocapillary endothelium basement membrane (0.14 um), 2) an outer colla— genous zone containing collagen fibrils and a mucoprotein ground substance (0.7 um), 3) an elastic layer (0.8 pm), 4) an inner collagenous zone (1.5 um) of collagen fibers in a finely granular ground substance, and 5) the basement membrane of the pigment epithelium (0.3 pm) (Hogan a: al., 1971). The border of the pigment epithelium is character— istically highly involuted in this region, and the pigment epithelium cytoplasm is rich in mitochondria. All of these features are purported to be indicative of a rapid diffusion of large amounts of substrate from the choriocapillaris to the pigment epithelium. Bernstein and Hollenberg (1965) utilizing ferritin injection studies and silver nitrate ingestion in Rhesus monkeys, Macaca mulata, indeed have substantiated this diffusion pathway and pointed out that it may be the preferred pathway over that of the retinal vessels in mammals. 18 In Esox lucius. Braekevelt (1974) reports that unlike the mammalian features described above, the endothelial wall of the choriocapillaris bordering Bruch's membrane is typi- cally very thin, but relatively non-fenestrated. Diaphragmed fenestrae are occasionally seen, however. The endothelial cell processes often overlap and are bound together by tight junctions. Vesicles are common within the endothelial cyto- plasm where no fenestrations are found. The nuclear region of the endothelial cells is rich in cell organelles: Golgi complexes, mitochondria, and rough endoplasmic reticulum. Bruch's membrane is composed of three layers: the basement membrane of the choriocapillaris endothelium, a middle layer of fine fibrils, and the basement membrane of the pigment epithelium. Also in contrast to mammals, the basal border of the pigment epithelial cell is not infolded. As stated earlier, no information exists on the mor— phology of the collection vessels draining the choriocapil— laris. The efferent rete Vessels have been discussed. The final category of vessel, the venous sinusoid, also has not yet been described at an ultrastructural level. Ultrastructure Summary- All of the above studies suggest certain characteristic properties which should be recalled when viewing the ultra— structure of the choroidal layer of Salmo gairdneri. With regard to the choroid rete, one might speculate based upon the literature reviewed above that the afferent vessels may 19 show thicker endothelial walls than the efferent vessels, and be composed of either one or two distinct endothelial cell types; or the afferent endothelial wall may show no significant difference in thickness from the efferent rete vessels. The efferent vessel endothelial walls in all cases are invariably reported to be thin and in some cases may or may not show areas with fenestrations. Endothelial junctions in both vessel types show little overlap and have desmosomes associated with zonulae occludens. The junctions may or may not show marginal flaps. Afferent vessels are depicted as being of a smaller diameter than the efferent vessels and in some cases the efferent vessels Show a sinusoidal configuration. Within the choriocapillaris, the features seen in Bag; lucius will be the most probable: attenuated, relatively non-fenestrated choriocapillaris endothelia abutting Bruch's membrane which is homogenous in structure, and subtends a smooth pigment epithelium basal border. The remaining choroidal vessel types observed on a light microscopic level have not been studied in teleosts at an ultrastructural level. MATERIALS AND METHODS Ocular tissues were collected from 150 to 300 g com— mercially cultured rainbow trout, Salmo gairdneri (Midwest Fish Farming Enterprises, Inc., Harrison, Michigan). Fish were held in fiberglass tanks at 12 i 1.000, with a continu— ous flow of aerated water which was treated to remove chloride and iron. The animals were exposed to light-dark periods of 16 and 8 hours, respectively. Choroidal tissues for this study were prepared by several techniques. For light microscopy, tissues were fixed in 10% neutral formalin, dehydrated in tetrahydrofuran, vacuum embedded in paraffin and stained with hematoxylin and eosin. Alter— natively, tissues were dehydrated, embedded and stained with Masson's trichrome as above after fixation in Bouin's fixative (Luna, 1968). Initial electron microscope preparations were fixed in 2% glutaraldehyde in 0.17 M cacodylic acid buffer at 4°C for three to four hours, washed three times in 0.17 M cacodylate buffer with 7% sucrose at twenty minutes per step, and postfixed in 0.17 M cacodylate buffered, 2% oSmium tetroxide for three hours at room temperature. The tissues were then dehydrated in either a graded ethanol series (30%, 50%, 70%, 95%, 100%) and embedded in Spurr's low viscosity epoxy resin (Spurr, 1969), or in a graded acetone series and embedded in EPON 812 epoxy resin (Luft, 20 21 1961). This fixation procedure is the standard technique .utilized in this laboratory for animal tissues and will be referred to as the standard immersion fixation procedure. After sectioning on a Porter-Blum MT—2 Ultramicrotome, the sections were counterstained with lead citrate and uranyl acetate after the method of Reynolds (1963). Sections were then observed on a Philips 300 transmission electron micro— scope. After a series of initial observations which indicated that a refinement in tissue preparatory technique was neces- sary, two experimental tissue treatment regimes were estab- lished. The first treatment included heparinization of the fish. One U.S.P. unit is the amount of Panheprin (Abbott Laboratories, N. Chicago, Illinois, 60064) required to main- tain fluidity in one ml of plasma. A perfusate of heparin— ized (2 U.S.P. units/ml) Ringer's solution is routinely used in this laboratory to clear gill arches (Bergman, 1975). On this basis, it was decided that the final circulating con— centration of sodium heparin should be between one and two U.S.P. units per milliliter of plasma. In Salmo gairdneri. blood volume is approximately 2.25% of the total body weight, whereas the plasma volume is about 1.50% of the total body weight (Schiffman and Fromm, 1959). Thus, approximately 0.03 U.S.P. units of sodium heparin per gram of total body weight were injected into the caudal vein ten minutes prior to enucleation. The eyes were then fixed in 0.17 M 22 cacodylate buffered, 2% glutaraldehyde, 2% paraformaldehyde after the technique of Karnovsky (1965). Subsequent steps were identical to the initial treatment. This preparation will hereafter be referred to as the heparinized immersion fixation treatment. In the second treatment, eyes were enucleated from non-heparinized fish, stored in a moist atmosphere for 20 minutes, and transferred to Karnovsky's fixative as described above. In order to better visualize the area from which elec— tron micrographs were taken, monitor sections of l to 2 um thickness were frequently taken for each block and stained with methylene blue (Richardson, Jarett and Finke, 1960). RESULTS Light Microscopic Choroidal Morphology Based upon light micrographs presented in Figures 3—12, it is evident that the general arrangement of the choroidal vasculature in Salmo gairdneri follows closely that pattern described previously in the literature review and by Barnett (1951). Blood enters the choroidal vasculature via an ophthalmic artery which branches into an arterial mani— fold. This then bifurcates into the afferent rete vessels. These vessels then converge upon larger "distribution" ves- sels on the peripheral side of the rete (Figure 1)., These then feed a vast capillary network-—-the choriocapillaris--— supplying all areas of the retinal-choroidal interface at Bruch's membrane. Large "collection" vessels return blood from the choriocapillaris to the rete where they bifurcate into small efferent rete vessels which lie interposed among, and parallel to the afferent rete vessels. These efferent vessels finally drain into a large thin-walled venous sinu- soid which encloses the arterial manifold and leads out of the eye via the ophthalmic vein (Figure 2). There are several points of interest which have not previously been mentioned in the literature. The vessels of the central side of the rete mirabile (i.e., the venous sinusoid and arterial manifold) are distinguishable as being relatively devoid of pigmentation. The arterial 23 24 manifold and its bifurcations are distinctly thick—walled, characteristically staining blue—green in the Masson's trichrome preparations indicating an abundant supply of connective tissue (Figures 7—8). 0n the peripheral side, however, interspersed among the distribution and collection vessels, and indeed, throughout the peripheral choroidal vasculature, pigmentation is characteristically observed (Figures 9—11). The collection vessels are not distinguish- able from the distribution vessels at this level. Within the rete itself, the parallel nature of the vessels is evident in longitudinal section (Figure 4). It is evident from Figure 12 that two types of vessels may be distinguished upon the basis of their size. Although no justification for distinguishing efferent from afferent vessels on the basis of size alone can be made in the current study on Salmo gairdneri, according to the studies of Wittenberg and Wittenberg (1974) and Copeland (1974a) the efferent vessels in all species studied were the larger of the two types. For this reason, it seems reasonable to assume that in Salmo gairdneri. this pattern holds. Thus, on the basis of Figure 12, the efferent vessels appear to be about twice the diameter (9.5 um) of the afferent vessels (4.5 um). The packing pattern of the afferent and efferent vessels appears to be one of an orderly array such that four efferent capillaries surround each afferent vessel (Figure 12). This means that the number of afferent vessels 25 equals that of the efferent vessels. This packing pattern is also seen in the tuna eye, Lampris regius, but in other species such as the swordfish, Xiphias gladius, hexagonal arrays are also encountered (Wittenberg and Wittenberg, 1974). These patterns of close packing are significant for the determination of exchanger efficiency. Electron Microscopic Fixation Techniques In order to further characterize the various vessel types in the choroid layer, an electron microscopic examina— tion was undertaken. Prior to pursuing a detailed study of the ultrastructure, however, it was necessary to develop proper fixation techniques. Initially, nonheparinized tissues were prepared by the standard immersion fixation procedure in which 2% glutar- aldehyde was used with postfixation in 2% osmium tetroxide. A sampling of the resulting choroidal rete ultrastructure is depicted in Figures l3, l4, and 15. The most notable fea- tures of the rete vessels and endothelia are: 1) very large vacuoles, 2) areas containing numerous micropinocytotic vesicles, 3) areas of extremely attenuated, but non- fenestrated, endothelial walls, 4) swollen adventitia between the vessels, 5) a relatively dense background matrix within the lumen of the vessels, and 6) numerous myeloid configur- ations and thin—walled vacuoles seen within the lumen of the vessels. 26 From these results, it was deemed necessary to modify 'Um preparatory technique in hopes of yielding characteris— ‘fics more typical of endothelial ultrastructure. When the fish'were heparinized prior to fixation by immersion in Iwrnovsky's fixative, much more typical micrographs resulted (Figures 19—21). To determine whether the heparinization versus the use of Karnovsky's fixative was the cause for the change in appearance, a Karnovsky's fixed, clotted preparation was processed for observation. The latter yielded results iden- tical to those seen in the standard immersion fixation preparation. Electron Microscopic Choroidal Morphology Having selected the optimal immersion technique, a detailed survey of the choroidal vascular ultrastructure was pursued. The structure of the ophthalmic artery was not observed in this study. The arterial manifold wall is very thick and is compos— ed of several cell types (Figure l6-18). There is a rather thick endothelial wall on both surfaces. These cells are richly supplied with rough—surfaced endoplasmic reticulum and polysomes. There are also very many micropinocytotic vesicles which appear to be concentrated on the border facing a layer of closely apposed smooth muscles (Figure 18). Were these vesicles active in the transport of macromolecules across the endothelial wall, one might expect the presence 27 of an equal number of these structures on the lumenal border. It is seen that this is not the case. The next layer is a thick matrix of collagen fibers. Within this collagenous matrix, one sees fibroblasts which are probably active in the secretion of the collagenous matrix (Rhodin, 1974), a single layer of smooth muscle cells, and occasionally bundles of both myelinated and non-myelina— ted neurons within Schwann cells (Figures 16—17). Neurons were occasionally seen to be closely associated with the smooth muscle cells. Neither neurons nor muscle cells were ever seen in any other choroidal vessel walls, thus it is possible that the arterial manifold may play a significant role in the regulation of flow through the choroidal vas— culature. Within the rete (Figures 19-21), there is only a single discernable type of endothelial cell in the wall of both the afferent and the efferent vessels. The cell is richly en- dowed with micropinocytotic vesicles opening on both sur- faces. Larger vacuoles are also seen. Dense polysomes are seen throughout the cytoplasm. Endothelial cell junctions are seen to be abutting rather than folded and overlapping and are characterized by possessing a single desmosome close to the endothelial surface and an associated marginal flap. Though the endothelial cells are similar in both afferent and efferent vessels, the afferent vessels have two dis— tinguishing features. First, a single layer of pericytes 28 is often seen to overlay the afferent endothelia. Second, outside these endothelial and pericyte cells is a very well— defined easily discernable basement membrane. This may serve as a significant barrier to the diffusion of substances between vessel types. Neither pericytes nor basement mem- brane are seen associated with the efferent vessel. There appear to be no pores or fenestrae in either afferent or efferent vessels. In summary, there are three characteristic features which may be used to separate the retial vessel types: 1) the size of the vessel, afferent being the smaller of the two, 2) pericytes are seen to be associated with the afferent vessels and not the efferent vessels, and 3) the presence of a distinct, well—defined basement membrane surrounding and outlining the afferent vessel endothelia and any associated pericytes. Although serial cross—sections through the length of the rete were made, no structural difference was discernable between the two ends. Close to both ends, however, the efferent vessels seemed to conjoin before those of the afferent side. At both ends, then, the efferent vessels appeared to be more sinusoidal in nature while the afferent vessels remained discrete. At the peripheral end of the rete, the distribution and collection vessels showed no discernable difference in structure (Figures 22-24). These vessels are notable in 29 laying a fairly thin single layer of endothelium. Outside ‘Ufis endothelium and within the overlying collagenous matrix are numerous fibroblasts. Overlying the collagen layer occur many heavily pigmented melanocytes. The granules here are roughly spherical, never rod—shaped. Above the pigment layer is a very loose multilaminar stroma which sep— arates adjacent vessels. No neural elements are ever seen. These distribution and collection vessels feed and drain respectively the choriocapillaris network which is separated from the pigment epithelium by Bruch's membrane. The structure of this region is identical to that seen in Esox lucius as reported by Braekevelt (1974) (Figures 25—27). The choriocapillaris capillary endothelia show a bipolar nature in that the scleral side is thicker and typically seats the cell nucleus and the bulk of the cytoplasmic constituents, whereas, the Bruch's membrane border is highly attenuated and shows random, but not abundant, diaphragmed fenestrations. Bruch's membrane is uniform, lacking an elastic layer (Figure 27) and is overlain by the basement membrane of the pigment epithelium. The pigment epithelium basal border is smooth as opposed to the highly involuted border seen in mammals. The ultrastructure of the pigment epithelium cells is interesting in that its cytoplasm has an extremely electron- lucent background substance. Inclusions include numerous mitochondria, rod—shaped pigment granules, and vacuoles 30 cmntaining phagocytosed photoreceptor outer segments (Anderson and Fisher, 1975) (Figures 28-29). The fact that the pigments present here are rod—shaped versus spherical as seen in the melanocytes within the choroid layer, sug- gests that they may be of different crystalline structure and thus of different chemical composition. DISCUSSION Several points which have heretofore gone unreported regarding the morphology of the choroidal vasculature in teleosts have arisen in the current study. A significant finding for use in subsequent localization studies was the fact that choroidal pigmentation occurred only on the peri- pheral side of the circulatory bed. At either end of the rete, the vessels unite into larger vessels: the arterial manifold or venous sinusoid on the central side, and large distribution and collection vessels on the peripheral side. These two sides are easily distinguished from one another in that the arterial manifold vessels are very thick—walled, and there is no pigment in or surrounding the vessel walls of the central side, whereas on the peripheral side, heavily pigmented melanocytes are interspersed among the collection and distribution vessels. Knowing this, it will be possible to distinguish central from peripheral vessels within the vascular bed during electron microscopic examination. Aside from this point, the choroidal vasculature of Salmo gairdneri appears to follow the pattern displayed by other species. Vascular Artifacts At the onset of the electron microscopic examination of the choroidal tissues, no problems had been anticipated with respect to fixation and tissue preparation. Eyes were enucleated from fish and were processed by the standard 31 32 immersion fixation procedures used for other tissues in this laboratory (i.e., fixation in 2% glutaraldehyde, with post- fixation in 2% osmium tetroxide). The results of these initial preparations have been described. After obtaining similar results, Copeland (1974a) reported that perfusion of the fixative is necessary to obtain relatively artifact— free electron microscopic preparations in the choroid rete mirabile of the killifish, Fundulus heteroclistus. In arriving at this conclusion, he had likewise initially at- tempted fixing the tissues by immersion in Karnovsky's fixa- tive (2% glutaraldehyde, 2% paraformaldehyde, with postfixa- tion in 2% osmium tetroxide). Karnovsky's fixative is char— acterized by penetrating much more rapidly due to the pres— ence of the paraformaldehyde. Copeland's results and the results of the current investigation were comparable with these immersion techniques. In that the Copeland report had been published after the onset of the present investigation, the literature was reviewed in an attempt to find an explanation for the un- usual appearance of these tissues in order to determine whether they were indeed artifactual. It was discovered that Brown, Stalker and Hall (1969) had investigated the ultrastructural changes in renal glomerular capillaries in which blood had undergone defibrination experimentally induced by thromboplastin and liquoid. Defibrination may be defined as the intravascular precipitation of fibrinogen 33 and/or its conversion to fibrin monomer or polymer, either 1) by artificial agents acting directly or through the throm- bin mechanism, or 2) by naturally occurring thromboplastins. Thromboplastin initiates the conversion of prothrombin to thrombin with consequent fibrin deposition, and is the natural activator of this reaction. Liquoid is an artifi- cial fibrin precipitating agent. In their studies, Brown a; al. (1969) showed that both agents resulted in the same ultrastructural changes. Within the lumen, granular material which they describe as "fuzzy" in appearance is seen in small amounts at three minutes after application, and it becomes progressively more distinct and abundant. This granular background material is seen in the lumen of the choroidal vessel of Figure 14. After 100 minutes, the amount of granular material is con— siderable and more electron dense. In the early stages, endothelial "blebs" are formed in a process they refer to as endothelial stripping. The fenestrated endothelial layer typical of renal glomerular capillaries is lifted from the basement membrane. The bleb space may be clear or may con— tain granular material. Thus, in the electron micrographs, the endothelium is elevated well above the subadjacent base— ment membrane and separated from it by a clear zone or by granular material (i.e., the bleb space). At several points local swelling of endothelial cytoplasm was noted. The small blebs can apparently form larger vacuoles which may 34 MHach completely from the endothelial layer and lie free hithe capillary lumen. Vacuoles similar to those published ‘MIBrown a3 al. (1969) are seen in the electron micrograph cfi‘the choroidal vessels in Figure 15. Brown a: al. contin- Lmd to explain that the endothelial cell nucleus itself tmcomes prominent and bulges into the lumen. It remains, tmwever, surrounded by a thin cytoplasmic rim. Vacuolation occurs in both the fenestrated layer of endothelial cytoplasm and the perinuclear cytoplasm. This perinuclear vacuolation is more common between the nucleus and the basement membrane. Vacuolation proceeds from both sides of the nuclear attach- ment zone, and there is detachment and extrusion of the nucleus together with a small cytoplasmic rim. Granular material assists in this process. By the 100 minute stage, all the endothelial changes are greater in degree. In some capillaries, however, endothelial nuclei are notably absent, and the capillaries appear bare. In the early stages there is an irregular slight indistinctness of the basement mem- brane especially where granular material is stripping the overlying fenestrated endothelial cytoplasm. Later, the basement membrane is irregularly thick and coarse in appear- ance. Brown a; al. (1969) conclude that: "We believe that granular material penetrates as far as the basement membrane, is temporarily held up and then spreads laterally to a subnuclear po- sition. Vacuolation of the endothelial cytoplasm adds to the insecure position of the nucleus. 35 This vacuolation may in part represent a method of disposal of the granular material; it may also be the result of filtration activity. A link with production of plasminogen activator by the endo- thelial cell must also be borne in mind but it is not yet possible to test this view. The accumu— lation of granular material and the vacuolation could each cause nuclear extrusion, but presum— ably they could be synergistic." It is noted that many of the changes reported by Brown a: al. (1969) were also seen both in the material presented in the current study and in Copeland's (1974a). Vacuolization of the endothelia, local swelling, granular background substance, endothelial stripping, and degenera- tion of the basement membrane were all prominent. Nuclear extrusion was not observed in our material and was not re- ported by Copeland, but this is explainable by the fact that these.preparations would not probably have reached the 100 minute stage in the process of clot formation. It was noted that Copeland (1974a) had used heparin in his perfusion technique, but had not in his immersion technique. No precautions had likewise been taken against clotting prior to fixation in the standard immersion fixation preparations of the current study. Thus, based on the reports of Brown at al. (1969), and due to the fact that teleost blood had been observed to display extremely rapid clotting times, averaging about 30 seconds in Salmo gairdneri (Smith, Lewis and Kaplan, 1952; Wolf, 1959), it was believed that the blood of tissues of Figures l3, l4, and 15 had clotted prior to the time of fixation. In light of the difficulties of 36 um perfusion technique where pressures must be carefully nmnitored and any unusual hydrostatic imbalances may pro- Ibundly affect the appearance of a capillary network, a mutable immersion technique is held to be preferable, pro— vided that fixation is rapid and results are repeatable. Pbr these reasons, a study was undertaken to try to estab— lish an acceptable, optimum immersion fixation technique and to confirm that clotting induced artifacts had occurred in our preparations and in Copeland's (1974a) as described by Brown a; al. (1969). The results of prior heparinization and of prior clotting have been shown. It was concluded that clotting had indeed occurred and that heparinization prior to fixation was necessary in order to obtain a suit- able immersion fixation preparation and the subsequent study was undertaken under these conditions. The precaution against clotting is apparently extremely important in the study of any vascularized tissues in that ultrastructural appearance is drastically altered. This point should perhaps be emphasized to a greater degree in future electron microscopic examinations. Choroidal Ultrastructure and Function The only literature thus far presented regarding the teleost choroidal vascular ultrastructure has been of ques— tionable value due to what is deemed to be artifactual tis— sue preparatory techniques. Thus, the findings of the cur- rent study warrant a rather detailed discussion. Functional 37 implications as well as strict anatomical description will be considered. The arterial manifold wall is seen to be a thick struc- ture consisting of two endothelial layers enclosing a dense collagenous matrix in which fibroblasts, smooth muscle, and myelinated and non-myelinated neurons may be found (Figures 16, 17 & 18). Although the arterial manifold lies within the lumen of the venous sinusoid, it is unlikely that any exchange of substances between the contents of these two vessels occurs based upon the barrier presented by the arter— ial manifold wall. It seems significant that the only area in which any innervation is observed within the choroidal layer is within the arterial manifold wall. Although no physiological basis for drawing conclusions regarding regulatory control is presented in this study, the occurrence of neural tissues within the wall of this vessel raises the suspicion that regulation of blood flow through the rete in response to any of a variety of factors may be possible via these tissues. There is evidence for such neural influence in the choroidal vasculature in mammals. In the human choroidal layer and uveal tract in general, myelinated and non-mye- linated nerves supply filaments to the blood vessels and the stroma (Wolter, 1960; Hogan, Alvarado and Weddell, 1971; Castro-Correia, 1967). In mammals, there are apparently two types of fibers: intervascular and perivascular. F" ‘L‘ ““221 38 Irrivascular branches course along the vessels sending tmanches into the wall, where they end as small filaments unthout end-plates. These perivascular elements are also found closely connected to melanocytes. In the peripheral regions of the choroid of S. gairdneri, where numerous nmlanocytes are found, perivascular neural elements were seen to innervate neither the vessels nor any of the melano- cytes, and indeed, no nerves were ever present anywhere in this region. This raises questions concerning the function and/or mechanism of dispersion of pigment within the melano— cytes found within the choroidal layer of §. gairdneri. With regard to the intravascular nerves, Castro-Correia (1967) speculated that nerve cells of a ganglionic type found within mammalian choroidal vessels might have some relation to the regulation of the ocular temperature as well as the intraocular pressure. The role of the teleost choroidal rete as a thermal exchanger seems to be in debate at this time (Barraco, 1975; Linthicum and Carey, 1972). In light of the fact that the arterial manifold in S. gairdneri is muscular and innervated it is interesting to speculate upon the possible role which this vessel may play in the regulation of flow, and hence, in the regulation of oxygen concentration, and thermal gradients. The importance of such a control mechanism to the operation of the counter— current exchanger definitely warrants further investigation. m 39 Although several papers have been published concerning the ultrastructure of the swimbladder rete, there has been a general paucity of work done on the ultrastructure of the choroidal rete mirabile. According to the evidence presented in the current study, it may be concluded that in S. gaaaa— aari, afferent and efferent vessels are definitely morpho— logically differentiated from one another. The afferent vessels are the smaller diameter vessels, being approxi- mately 4.5 pm in diameter. There is no discernable differ- ence between the endothelial wall thicknesses of the afferent versus efferent vessels. The afferent vessels are surrounded by a clearly distinct basement membrane. The packing pattern displayed in S. gairdneri is a cubic, checker board array with four efferent vessels surrounding a single afferent vessel (Figure 12). The packing pattern of the vessels is not the same in all teleosts, as for instance the sword- fish, Xiphias gladius, displays an hexagonal array (Witten— berg and Wittenberg, 1974). The efferent vessels are clearly discrete as typical cylindrical vessels as opposed to sinusoidal in nature as reported in Amia calva (Wittenberg and Wittenberg, 1974) and in Fundulus grandis (Copeland, 1974a). There may be several reasons for the discrepancy. To begin with, there may very well be species-specific differences in the structure. A second possibility is that the micrographs published by these investigators may have been of tissue located very 40 close to the distal or proximal ends of the rete at a point where the efferent vessels were still of a larger vessel dimension. Indeed, such sinusoidal areas of the efferent vessels may be seen in a tangential section when the ends of the rete are approached. Finally, Copeland's preparation was perfused at a pressure of 0.75 psi (Copeland, 1974a) and based upon his published micrographs, and upon the fact that the vessel walls separating adjacent endothelial cells are very attenuated, and presumably delicate, any non- physiological trans-luminal pressure caused by the perfusion technique may have broken these walls. In fact, it appears that instead of the sinusoidal case which he chose to illus— trate, the more common case is one of a single arterial vessel surrounded by five or six venous vessels. Close in- spection of Copeland's micrographs reveals that the endo- thelial lining of the venous sinusoid vessel reveals areas which may easily have been endothelial bridges which have been torn away. Line diagrams of micrographs of the rete of gala also illustrate the presence of an efferent sinu- soidal arrangement (Wittenberg and Wittenberg, 1974). Unfortunately, the technique of tissue preparation used was not described. The importance of the tissue preparatory technique in studies of the ultrastructural morphology of this network has already been pointed out. If indeed, the micrographs thus far published are erroneous, it will not have been the first time that the elucidation of rete _ _ .5 3‘.“ “.3 41 physiology has suffered at the hands of fixation artifacts. According to Dorn (1961) who also worked on the swimbladder rete ultrastructure, Jaeger, in 1905, reached the conclusion that gas was accumulated in the lumen of the swimbladder by the physical rupture of blood cells by a toxin released by the gas gland epithelium at the site of the rete. This sur- prizing conclusion it seems was arrived at by light micro— scopic observation of poorly fixed tissues. It is my contention that Copeland's perfusion technique may lead to as much artifactual results as did his initial clotted tissues and that yet another appearance may be ob- tained by using a heparinized immersion technique which is deemed to be closer to the true morphology. Assuming that the results seen in the heparinized preparations adequately represent the closest approximation to the true condition, comparisons to other vascular beds may be made and certain functional implications drawn. Bennett, Luft and Hampton (1959) have developed a classification scheme for mammalian capillary types based upon three morphological features--—basement membrane, peri- cytes, and endothelial fenestrations. In keeping with the established terminology seen in the microvascular literature of classifying capillary vasculature according to Bennett's classification scheme, the vessels of the choroid rete mira— bile may be roughly categorized as follows. The afferent component of the choroidal rete in Salmo gairdneri would #- -h m—hfl 42 be classified as belonging to vessel types represented by muscle capillaries (type Ala) having a complete basement membrane, lacking fenestrations, and being partially sur- rounded by pericytes. Efferent components of the rete contain few, if any, fenestrae and should be grouped together with capillaries of the mammalian kidney and endocrine glands (type A2a). The categorization of these rete vessels into Bennett's classification scheme cannot be considered to specifically describe the morphology or function of the vessels present, but it is interesting from a comparative standpoint. A A With regard to the ultrastructure of the swimbladder rete mirabile, Fawcett and Wittenberg (1959) suggest the hypothesis that "the differences observed in fine structure of the capillary walls may possibly be correlated with dif- ferent physiological mechanisms of transcapillary exchange in the two categories of vessels comprising the counter— current system." Jasinski and Kilarski (1971) also suggest that the general organization of the rete mirabile and dif- ferences in the structure of the afferent and efferent capil- laries probably reflect some functional role different for both types of capillaries. Capillaries are extremely per- meable to oxygen and carbon dioxide (Pappenheimer, Renkin and Borrero, 1951), thus allowing counter current exchange of these two gases. Marked differences between the fine structure of the afferent and efferent capillaries have been 43 noted. However, these differences would not be expected to affect the transcapillary diffusion of gases. Thus, there must be some other functional significance to the structural difference if this hypothesis is true. Jasinski and Kilarski (1971) noted that despite the large number of smooth surface vesicles found in individual sections of the swimbladder rete endothelial cells, very few* are seen to empty at either of the cell surfaces, i.e., basal or lumenal. They suggest that these vesicles may represent smooth endoplasmic reticulum in cross-section. Copeland (1974a) noted the possible presence of a similar microtubular system, as opposed to the more commonly expected pinocytotic vesicles, in his immersion-fixed choroidal rete endothelial preparations. The tubular structures shown by Copeland (1974a) were present in clotted tissues and may not be indicative of typical rete endothelial ultrastructure. Bruns and Palade (1968a) have reconstructed models from serial sections of endothelia from mammalian skeletal vas- culature. In so doing they have found that a common feature of capillary endothelia is the presence of many true pino— cytotic vesicles as opposed to a tubular network of smooth endoplasmic reticulum. These vesicles do open frequently to the surfaces of the cell. The findings of the current study suggest that there are numerous vesicles within the choroidal rete endothelial cell cytoplasm. That these are not likely tubules in cross—section is evidenced by the “F“—““.—“g 4L: fact that one rarely sees a tubular structure in longitu- dinal section (Figure 20). The vesicles present do not, however open frequently at the cell surface as noted by Jasinski and Kilarski (1971) indicating that pinocytotic activity is not a very active process thus negating this as a viable mode of transcapillary exchange. Movement of water across the afferent-efferent inter— face has been postulated by Wittenberg and Wittenberg (1974) as being possibly significant in the oxygen concentrating mechanism. The possibility of movement of other substances has also not been ruled out. Like other capillaries, they are probably readily permeable to lipid soluble substances such as carbon dioxide, whereas water and sodium, for exam— ple, would enter and leave the endothelial wall less easily. The permeability of the basement membrane is not known. According to Rhodin (1974), the basement membrane consists of a thin basal lamina rich in mucopolysaccharides, amino acids and a reticular network composed of delicate colla— genous fibrils. He does state that the basal lamina estab- lishes a microenvironment for the adjacent cells and acts as a diffusion barrier to rapid ion exchange. It is inter- esting to note that the basement membrane may be present around the afferent vessels to act as a barrier to diffusion of certain substances between afferent and efferent vessels, whereas it is not seen surrounding efferent vessels which often lie adjacent to one another. At these locations no 45 deient for diffusion exists, thus eliminating the neces— sfity for a diffusion barrier. In contrast to this notion is the fact that there is a dual basement membrane between ‘flm pigment epithelium and the choriocapillaris, the pre- sumed site of major metabolite and nutrient exchange for 'Uw retina. The question of whether the basement membrane acts as a significant diffusional barrier awaits further investigation. The endothelial cell junctions in the cho- roidal rete, although not overlapping, characteristically are seen to possess desmosomes and zonulae occludens whose structure presumably represents a significant barrier to the passage of substances by a pericellular route (Fawcett and Wittenberg, 1962; Wade and Karnovsky, 1974; Spitznas and Reale, 1975). Finally, if the pinocytotic vesicles Within the endothelia were indeed active in the movement of substances from one vessel to another, it would seem likely that they would be seen to open more frequently onto both the lumenal and basal surfaces of the endothelial cells. Thus, it seems that the movement, and hence, the counter— current concentration of substances not readily permeable both to the endothelial cell membrane itself and to the basement membrane is unlikely in this system. This supposi— tion is again based upon presumptive morphological evidence and requires further experimental verification. At either end of the rete, the vessels converge into larger vessels; the arterial manifold or venous sinusoid 46 at the central end, and large distribution and collection vessels at the peripheral end. These two sides are easily distinguished from one another in that the arterial manifold vessels are very thick-walled, and there is no pigment in or surrounding the vessel walls in this end, whereas on the peripheral end, heavily pigmented melanocytes are inter- spersed among the collection and distribution vessels. The collection and distribution vessels pass out to the choriocapillaris through a loose connective tissue stroma. These vessels are seen to be embedded in a collagenous matrix Which is secreted by numerous fibroblasts. Melanocytes are also seen throughout this layer. It has already been stated that no innervation of either the vessels of this region or of the melanocytes has been observed in Saiga gairdneri which is contrary to the case in man, rat, hamster, and guinea pig (Castro-Correia, 1967). In addition, the human also displays mast cells, macrophages, plasma cells, and lymphocytes (Hogan, Alvarado and Weddell, 1971), none of which have been observed in S. gairdneri in this region. It has been pointed. out (Hogan a; .52” 1971) that the func- tional significance of the melanocytes within the choroidal layer is not yet clear. The function of light screening Shmild have been accomplished by the pigment epithelium by this POint. The tapetum lucidum in this fish is located at the baCk of the choroid in the region of the suprachoroid layer, next to the sclera. If this reflective layer is 47 functional in light gathering in the dark, as has been pro— Imsed (Arnott a; al., 1974), this would mean that light would indeed have to pass through the pigment epithelial layer and the choroid layer as well, suggesting the possi— bility that the light-absorbing function of the melanocytes within the choroid may remain at certain times. The structure of the choriocapillaris, Bruch's membrane, and pigment epithelium are identical to those described by Braekevelt (1974) in Esox lucius, but differs drastically from that seen in the human (Hogan at al., 1971). The major differences are three—fold: l) the choriocapillaris is much more highly fenestrated in man than in teleosts, 2) Bruch's membrane displays a slightly different arrangement, and 3) the basal border of the pigment epithelium is very highly involuted in man, whereas in teleosts, all signs of in- volution are lacking. Functionally, the choriocapillaris has been implicated in mammals as being the primary site of nutrient exchange rather than oxygen and gas exchange, the latter function being provided by the retinal vasculature (Braekevelt, 1974). In teleosts, there is no retinal vasculature, hence the choriocapillaris must meet all circulatory needs. As already pointed out, in keeping with the notion that the vascular architecture must be designed to meet and indeed determine tissue functional demands, the teleost choroidal rete has evolved into a counter-current gas exchanger. It is likely, 48 however, that the exchange of other metabolic substances and products does not rely on the concentrating effects of a counter-current multiplier, but upon the architecture of the choriocapillaris, Bruch's membrane, and pigment epithe- lium. In agreement with the report of Braekevelt (1974) with respect to the choriocapillaris architecture in Esox lucius, numerous fenestrations do occur in the endothelial border lying adjacent to Bruch's membrane in Salmo gairdneri, which would presumably allow or facilitate passage of macromolec— ular nutrients and metabolites into and out of the retina. If the choriocapillaris is the major site of nutrient deli— very in the vertebrate eye, as indeed it must be in the case of teleosts, the occurrence of fenestrations is significant. Fenestrations have classically been characterized as the site of macromolecular passage (Karnovsky, 1967). The only al- ternative would be pinocytotic activity which is typical of some capillary beds (eg. mammalian skeletal capillaries (Bruns and Palade, l968a & l968b)), but is notably lacking in the vicinity of Bruch's membrane in S. gairdneri (Figure 26). Another possible route would be extracellular passage at the site of the endothelial cell junctions. These borders, like those seen in the choroidal rete mirabile capillaries, characteristically display desmosomes bordered by zonulae occludens which, by the nature of their structure, present significant barriers to the passage of substances. The 49 diaphragmed fenestrae, then, are the only remaining site possible for the passage of macromolecular nutrients. Braekevelt (1974) points out, however, that in Esox lucius the degree of choriocapillaris fenestration is significantly less than that seen in mammals. This was also seen in S. gairdneri. The role of the falciform process in nutrient supply in teleosts also remains unclear (Walls, 1963). Thus, the question of where macromolecular exchange occurs in this region awaits further study, perhaps by the techniques of ferritin or horse radish peroxidase injection (Clementi and Palade, 1969). The pigment epithelium has been reported to be active in the disposal of shed rod and cone outer segments (Anderson and Fisher, 1975) along with its light absorbing characteristics. The cytoplasmic inclusions seen in Figures 28 & 29 tend to imply that these functions are also characteristic of the pigmented epithelium of S. gairdneri. The significance of the lack of involutions in the pigment epithelium is not known. This feature of the ultra- structure seen in the mammalian system is observed in several secretory tissues such as the adrenals (Hillman, Seliger and Burk, 1975) and teleost pseudobranch (Copeland and Dalton, 1959), the latter of which has been implicated as a source of carbonic anhydrase production (Copeland, 1951; Fairbanks, Hoffert and Fromm, 1969). _‘Ifl CONCLUSIONS The ultrastructure of the rete mirabile of Salmo gairdneri differs considerably from the morphology presented for Fundulus grandis, the only other choroidal rete to have been studied and published at an electron microscopic level (Copeland, 1974a). Rather than a species specific differ- ence, it has been pointed out that the apparent difference is probably due to the fixation technique employed in the preparation of each tissue. In teleosts, it has been con— cluded that when working with vascular tissues, extra pre— cautions must be taken in the collection procedure to avoid clotting of the blood which has been shown to cause drastic changes in the ultrastructural morphology of capillaries. Also, an immersion fixation technique is deemed preferable to a perfusion technique due to the delicate nature of the endothelial walls, particularly in those areas of the cho- roidal rete in which the wall is composed merely of two adjacent endothelial cells lacking supportive basement membranes or pericytes. Discussion of the functional significance of each structure studied within the choroid layer has been based only upon the morphological findings and not upon direct experimental evidence. The features seen, however, must still be taken into account when explaining any empirical evidence regarding this tissue. Indeed, the structures seen 50 51 suggest directions which may be taken in future investiga- tions which have heretofore gone uninvestigated. Perhaps the most significant example is that of the role which neuro—muscular elements within the arterial manifold may play in the regulation of flow through the choroidal vascu- lar bed and of its influence upon the operation of the counter-current exchanger. Also, no morphological barriers were seen to exist for the counter—current exchange of gas within the rete. It was concluded, however, from the morphological evidence presented herein that it is very unlikely that the counter—current exchange of any other molecular substance requiring specialized diffusional or transport pathways occurs. This, however, remains to be verified by empirical measurements. Finally, morphological features which typically characterize active areas of ex- change are relatively scarce in the one area of the choroid which is the only possible site of exchange for the retina, that is, the choriocapillaris, Bruch's membrane, pigment epithelium complex. The resolution of this apparent con— tradiction also awaits further experimental evidence. Having presented the basic ultrastructure in Salaa gairdneri upon which any functional model must operate, it remains to demonstrate the possibility of a local focus of carbonic anhydrase in the counter-current exchanger and its likely source of origin as suggested by the short-circuit model (Fairbanks, Hoffert and Fromm, 1974). PART II INTRODUCTION In tracing the path along which blood must flow from the heart to the eye, it is seen that in Salmo gairdneri blOod passes from the heart sequentially through the conus arteriosus, ventral aorta, afferent branchial artery, af— :ferent filamental vessels, and out into the microcirculation of the secondary lamellae. Here oxygen and carbon dioxide are exchanged, and a certain degree of ionic regulation occurs. From the secondary lamellae, blood then flows through the efferent filamental vessels, the efferent branch- ial vessels and then unite to form the dorsal aorta where oxygenated blood may either flow anteriorly to the head through the internal carotid arteries, or posteriorly to the bulk of the systemic circulatory beds. . In the first branchial arch (anterior-most arch), unlike in the other arches, oxygenated blood may also flow ventrally in the efferent branchial vessel to the point of junction between the branchial arch, the operculum, and the lower jaw. Without branching, this vessel, which is now termed the afferent pseudobranch artery, leads directly to the inner surface of the operculum where it proceeds dor- sally and laterally roughly paralleling the opercular cover margin until arriving at the pseudobranch. The pseudobranch is morphologically and phylogenetically a vestigial gill lying on the inner surface of the operculum, and in the 52 53 I‘a-in'bow trout is isolated from the buccal chamber by an OVerlying layer of dermal tissue. In the pseudobranch, ‘blOOd flows into vessels homologous to afferent filamental .VeSsels, through a system of capillaries similar to that Of the secondary lamellae and leaves via efferent filamental vessels to the efferent pseudobranch artery. Qualitative (*Hinges in the blood passing through the pseudobranch remain t0 be definitively determined, and the possible role of the pseudobranch is discussed below. Blood passes directly from the pseudobranch to the choroid layer of the eye. Thus, the efferent pseudobranch artery is synonymous with the ophthalmic artery. Interposed between the pseudobranch and the choroid is a communicating artery which allows an anastomosing pathway between the right and left ophthalmic arteries, presumably as a protec- tive mechanism should one afferent blood supply become damaged or otherwise occluded. The pattern of blood flow through the choroidal vascu- lature has been discussed (Part I). It now remains to be seen what happens to the blood in passing through the pseudo— branch and choroidal rete mirabile counter—current exchange system, and more specifically, where along this path are the foci of carbonic anhydrase activity, which are essential in the counter-current concentration of oxygen. Fairbanks, Hoffert and Fromm (1974) have proposed a theoretical mechanism for the counter—current multiplication “ .5 54 of Oxygen by the choroidal rete mirabile. They postulated that strategically placed foci of carbonic anhydrase acti- ‘Vity'must be present within the choroidal vasculature in Order to prevent the diffusion of carbon dioxide from the efferent to the afferent retial vessels. They reasoned that if carbon dioxide were allowed to diffuse into the afferent ‘Vessels, a premature release of oxygen from the hemoglobin (would occur creating a gradient for oxygen to flow into the efferent vessels and entirely bypass the circulation carrying it to the retinal border. The retial carbonic anhydrase activity was proposed to be present at some location in the efferent vessels to trap this carbon dioxide before a short circuiting could occur. They further speculated that it must be present on lumenal membrane receptor sites on the efferent rete endothelia since carbonic anhydrase inhibitor of low diffusibility, CL-ll,366, not only rapidly stopped oxygen concentration by the system, but also accumulated in the rete in a time course paralleling the inhibitory effects, and not in the erythrocytes or retina. Finally, it was speculated that if the carbonic anhydrase were on membrane receptors, it may have originated from an exoge~ nous source. All blood passing to the choroidal rete must first pass through the pseudobranch, and it has long been known that pseudobranch tissue has a high concentration of carbonic anhydrase. Copeland (1951) and Fairbanks, Hoffert and Fromm (1969) proposed that the pseudobranch may secrete 55 Caxlxmnic anhydrase for use in the swimbladder and choroid :retiau Finally, with the advent of electron microscopy, it was discovered that the pseudobranch had a very charac- teristic cell type of unknown function which appeared sus- piciously to be of secretory nature, that is, having a mito- cfllondrial network associated with a massive tubular system Which emptied onto the capillary endothelium (Harb and Copeland, 1969). Since the work of Fairbanks, Hoffert and Fromm (1974), an improved technique for staining carbonic anhydrase had been developed and applied to electron microscopy largely by Seymour Rosen and colleagues at Harvard Medical School. The technique is based on the degradation of bicarbonate in- to carbon dioxide and hydroxyl ion by carbonic anhydrase. The tissue is floated on an incubation medium allowing loss of carbon dioxide to the atmosphere and causing a localized alkalinization in the tissue at the site of carbonic anhy- drase activity. The alkalinization causes a local deposi- tion of a cobalt salt which is later converted to cobalt sul- fide, a black precipitate in light microscopy and an electron dense precipitate in the electron microscope. In Part II, the localization of sites of carbonic anhy- drase activity in the choroidal vasculature and the pseudo- branch was attempted using both light and electron micro- scopic techniques. It was hoped that an important addition could be made to the body of evidence regarding the short- 56 Cir cuit model of count er- current m ulti plica tion of oxy gen _ ”Vin—rs; LITERATURE REVIEW The Historical Elucidation of Retial Function Upon anatomical investigation of the choroid rete mira— bile, it became obvious that the choroidal rete morphology is very much like that of the swimbladder rete mirabile— i§is gland complex. The structural similarities of these tW0 organs are striking. Both contain a rete mirabile associated with a specialized epithelial layer; the gas gland epithelium in the swimbladder, and the pigment epi- thelium in the choroid. In 1806, Biot had shown that oxygen tensions within the swimbladder were so high that a pressure gradient of over 100 atmospheres exists between the dissolved blood gases and the bladder gases so secretion had to occur. Much later, in 1911, Woodland showed that this secretion was accomplished by the complex of the gas gland and rete mira- bile (Haldane, 1927). Upon discovery of the choroidal rete, it was then natural to question whether the rete of the choroid performed a similar function in the eye. It was with this reasoning that Wittenberg and Witten- berg (1962) first undertook the task of measuring the oxygen tensions in the eye of various teleost species. According to their findings, oxygen tensions at the vitreal surface of the retina were found to attain levels far above arterial oxygen tensions, often ranging from 400 torr to 57 58 1000 twrr in those species with well-developed choroidal irete mirabiles. With the development of better recording devices, Fairbanks, Hoffert and Fromm (1969) then demon— Strated the same phenomenon to occur in the rainbow trout, EEEAEQ gairdneri. With a similar structure and function having been demon— Str'ated, the elucidation of the mechanism behind this capa— bility remained. The swimbladder mechanism has long been discussed and much empirical evidence has accumulated upon which a working hypothesis has been constructed. Steen (1970) has reviewed the distribution, structure, and the historical development of thought regarding the mechanism of function proposed for the swimbladder oxygen multiplier. It has thus far been presumed that the mechanism behind the function of both the choroidal and swimbladder retia is identical. An explanation of these proposals and some of the assumptions upon which they are based are presented below. The configuration of the rete is one which may act both as a counter—current exchanger and a counter-current multi- plier. A counter—current exchanger is a system in which all exchange processes are strictly passive, substances merely diffusing down their respective gradients. In a counter- current multiplication mechanism, other active processes which involve more complex chemical reactions are involved. Here, the processes which make the rete an active multiplier are due to the characteristics of both hemoglobin and ionic 59 SOlirtions which affect their gas transporting capabilities. The Rete Mirabile As An Exchanger The importance of the rete as a counter—current ex- Changer stems from the fact that it presents a barrier to 'the leakage of gas down the extremely large gas gradients Ipresent. As blood passes through the swimbladder or eye, it is exposed to very high oxygen tensions, for example, 450 tOrr in the eye of S. gairdneri (Fairbanks, Hoffert and Fromm, 1969). The blood leaving this region will possess comparable partial pressures of oxygen. Upon entering the efferent vessels of the rete mirabile, blood comes into close contact with blood in the afferent rete vessels of much lower partial pressure. The gas will then flow down its pressure gradient such that the partial pressures in both streams will approach each other. Depending upon the efficiency of this counter—current exchange system, the oxygen tension of the blood leaving the efferent rete ves- sels will be very close to, but slightly higher than that of the blood entering the afferent rete vessels. The effi- ciency of the rete as a gas exchanger is determined by the area, the thickness, and the gas permeability of the mem- brane shared by the afferent and efferent blood, as well as by the rate of flow which determines the length of time the two streams of blood are in contact. Marshall (1972) has extensively surveyed the occurrence and form of the swimbladder rete in marine teleosts. He 60 points out that in the deep water forms, in order to main— tain neutral density, the swimbladder, as a buoyancy device, must secrete gasses into the bladder lumen at much higher pressures. As stated above, the efficiency of the rete mirabile as an exchanger is dependent among other things upon the length of its capillaries. Marshall (1972) did indeed find a direct correlation between depth of occurrence and rete length: 0.75-2.0 mm in upper mesopelagic species (ca. 200-600 m); 3.0—7.0 mm in lower mesopelagic species (ca. 600-1200 m); and 15—25 mm in the deepest living bentho- pelagic species (1500 m and below); 8-12 mm in species between 800-2000 m. In two interesting apparent exceptions to this trend, the retia were shorter than expected, but the fish had evolved such that the vessels were smaller diameter so that the exchange surface area was greater and the erythrocytes had become enucleate so that they could pass through the smaller bore vessels. Wittenberg and Haedrich (1974) recently surveyed the occurrence of the choroid rete and found no correlation between rete capillary length and depth of habitat. In fish with larger eyes, the number, not the length, of vessels increased. The vessel length was quite uniform. In summary, the importance of the rete mirabile in its capacity as a counter-current exchanger lies in the fact that it prevents blood leaving the circulation of the swim- bladder or eye from depleting the high oxygen tensions 61 present. Workers all agree that the rete provides a barrier for depletion of high oxygen tensions, but this, of itself, does not explain how the high tensions are generated. The Rete Mirabile as a Multiplier Because blood does not act as a simple solution in transporting gases, other processes must be taken into account. The concentration and multiplication of oxygen arises from the fact that there is a reduction in gas solu- bility in the vicinity of the area of buildup. The phenomena which create this reduction in whole blood carrying capacity are as follows: 1) Bohr effect: as the pH decreases and carbon dioxide increases, it takes higher partial pressures of oxygen to saturate the blood, thus the affinity of hemoglobin for oxygen is reduced, 2) Root effect: as the partial pressure of carbon dioxide rises, the oxygen carrying capacity of the hemo- globin is reduced, and 3) Salting-out effect: in any solution an increase in ionic concentration results in a decreased solubility for gases. In both choroidal and swimbladder retial systems, the epithelial layers and related tissues are characterized by having intense glycolytic activity such that lactic acid is produced at a great rate even under aerobic conditions (D'Aoust, 1970; Hoffert, Eldred and Fromm, 1974). This 62 lactic acid acts to acidify and to increase the ionic strength of the blood passing through these tissues. This activates all three of the above mechanisms and insures that oxygen and other gases, such as nitrogen, which may be in solution, are released and trapped in the exchanger mechan- sim. The importance of the lactic acid in the operation of the counter-current multiplier of the swimbladder has been demonstrated. D'Aoust (1970), using oxamic acid, a competi- tive inhibitor of lactic dehydrogenase, demonstrated the complete inhibition of the secretory capability of this tissue. He further demonstrated that the 1actic acid is derived from circulating blood glucose rather than from endogenous glycogen stores in the secretory epithelium. The Role of Carbonic Anhydrase The counter-current multiplication mechanism has also been shown to be dependent upon the enzyme, carbonic anhy— drase. Acetazolamide inhibition completely and rapidly de- stroys the elevated oxygen tensions in the eye (Fairbanks, Hoffert and Fromm, 1969) and also in so doing disrupts the functional integrity of the retina as seen by a decline in the ERG b-wave (Fonner, Hoffert and Fromm, 1973). There are three sources of carbonic anhydrase in the vicinity of the counter-current multiplier mechanism: the retina, the ery- throcytes, and the rete endothelia. Fairbanks, Hoffert and Fromm (1974) have recently proposed a ”short-circuit" model which offers an explanation for the effect of carbonic anhy- 63 drase inhibitors, and the roles which are played by each site of carbonic anhydrase activity. Retinal carbonic anhy- drase catalyzes the dissociation of carbonic acid which re— sults from the neutralization of lactic acid by bicarbonate, thus producing carbon dioxide and water. The carbon dioxide enters the efferent blood and a portion enters the erythro- cytes where erythrocyte carbonic anhydrase hydrates it to form bicarbonate, the latter of which leaves the red blood cell to aid in neutralizing more lactic acid in the retina. Hydrogen ion is also formed and this causes the Bohr and Root shifts to occur, thus releasing oxygen from the hemo- globin. The remaining carbon dioxide which was not seques- tered by the erythrocytes enters the plasma to increase the carbon dioxide tension. The natural tendency would be for this carbon dioxide to diffuse down its gradient into the afferent stream causing a premature release of oxygen which then would diffuse down its gradient into the efferent rete vessel and pass out of the eye without ever reaching the choriocapillaris. In order to prevent this short—circuiting of the counter-current multiplication mechanism, Fairbanks a3 al. (1974) deduced that there must also be a rete mira— bile carbonic anhydrase somewhere in the efferent endothe- lial wall which would act to hydrate the carbon dioxide. before it can do this. In so doing bicarbonate ion would be produced which would flow into the afferent rete in ex- change for chloride ion. In this manner, not only is the 64 buildup of carbon dioxide prevented, but recycling of bicarbonate ion for lactic acid neutralization is accom- plished. Acetazolamide would inhibit all of these forms of car— bonic anhydrase so that the specific site of activity which was crucial for the operation of the counter-current multi- plication mechanism remained in question. Fairbanks, Hoffert and Fromm (1974) then found that CL—ll,366, a carbonic anhydrase inhibitor of low diffusivity, inhibited oxygen concentration as rapidly and to the same degree as aceta- zolamide. They also were able to demonstrate that only in the rete was there evidence of a retention and progres- sive accumulation of the inhibitor in a time course parallel- ing the inhibition of the counter-current multiplication mechanism. Because CL-ll,366 is an inhibitor of low diffus- ivity, they concluded that either the rete endothelia walls could be more permeable to it than the retina or erythro— cytes; or the rete carbonic anhydrase could be bound to the.luminal wall of the capillary so that it is exposed to the non-diffusible inhibitor. It has long been a matter of debate as to the possi- bility of the pseudobranch being active in the secretion of carbonic anhydrase for use in the counter-current multipli- cation systems of the swimbladder and choroidal retia. Wittenberg and Haedrich (1974) suggested that the pseudo— branch acts in consort with the choroid rete to create a 65 high oxygen tension without a simultaneous buildup of carbon dioxide. The pseudobranch is known to contain a large conc- entration of carbonic anhydrase (Hoffert and Fromm, 1966; Maetz, 1956). Copeland (1951) and Ffinge (1953) demonstrated a depression of secretion of gas into the swimbladder after cauterization or extirpation of the pseudobranch and Copeland thus suggested that the pseudobranch might be serving to secrete carbonic anhydrase for use in the swimbladder. Later, Fairbanks, Hoffert and Fromm (1969) suggested a similar function for the choroid rete multiplier. Maetz (1956), however, could find no difference in the carbonic anhydrase concentration in the afferent versus the efferent pseudobranch artery and he therefore concluded that the pseudobranch did not secrete carbonic anhydrase. Also, Fairbanks a; al. (1974) followed the accumulation of the inhibitor CL-ll,366 in the various tissues in conjunction with the time course of oxygen decline at the retinal-vitreal interface. No CL—ll,366 was found in the pseudobranch ten minutes after injection, although the oxygen concentrating mechanism was still repressed at this time. That the pseudo— branch does have a very intense carbonic anhydrase activity is not questioned, however, and its usefulness at this site has Proved to be a persistent conundrum despite repeated efforts to find the reason for its presence. Wittenberg and Haedrich (1974) suggested that pseudobranchial carbonic anhydrase acts to remove carbon dioxide from the blood: 66 entering the rete and thus acts in consort with the choroid rete to create a high oxygen tension without a simultaneous buildup of carbon dioxide. The pseudobranch possesses a very characteristic cell type whose origin was for many years a matter of consider- . able debate. It was believed by some to be a form of chlo- ride cell (Newstead, 1971), but it has since been fairly well—established as an independent cell type. With light microscopy, these cells appear to be characteristically acidophilic (Hoffert and Fromm, 1965). At the electron microscopic level, it is seen to have a large number of mitochondria in association with a dense network of smooth endoplasmic reticulum. Numerous speculations regarding the function of the acidophilic pseudobranch-type cells have been made. Since they resemble chloride cells of the gills, much work has been done in futile attempts to prove an osmoregulatory function for the pseudobranch-type cells (Newstead, 1964; Holliday and Parry, 1962; Kessei and Beams, 1962). Perhaps the most convincing evidence thus far presented has come from the work of Laurent and Rouzeau (1972). In studying the innervation of the pseudobranch, they have shown that neural receptors from the glossopharyngeal nerve are sensi- tive to changes in blood pH, P002, P02, osmotic concentra— tion, sodium ion concentration and to pressure. For this reason, they have suggested an apparent homologous role to 67 the mammalian carotid body and carotid sinus complexes. They have also noted the presence of carbonic anhydrase in the pseudobranch—type cell and have demonstrated on a light microscopic level of resolution its localization with— in the pole of the cell which is rich in smooth endoplasmic reticulum tubular system. They suggest that this morphol- ogy is indicative of a specialized receptor ending associ- ated with the receptor functions which they demonstrated. Until more definitive evidence is presented, however, this conclusion may not be entirely correct. In view of the anatomical relationship between the pseudobranch and the eye, and the function of carbonic anhydrase in the retial oxygen multiplication mechanism, there is still a possibil- ity that the pseudobranch plays a vital role in the operation of this mechanism in teleosts. Carbonic Anhydrase Histochemistry A relatively recent technique has been developed which was used by Laurent, Dunel and Barets (1969) for the histo- chemical localization of carbonic anhydrase. The technique had been criticized, but the criticisms have been countered (Rosen and Musser, 1972; Lbnnerholm, 1974) and the technique is now fairly well-established. Details of the method have been reviewed and revised by Cassidy and Lightfoot (1974) and Rosen (1974). Essentially, the reaction which is util- ized in the staining procedure is the dehydration of bicar- bonate ion into carbon dioxide and hydroxyl ion (Cassidy 68 and Lightfoot, 1974). The carbon dioxide is lost to the air and the tissue is alkalinized in the vicinity of the reaction. A cobalt salt is the basic metal salt which is precipitated at the site of tissue alkalinization. The cobalt salt (either as a carbonate, phosphate, or hydroxyl— ate) is then converted to cobalt sulfide which is visible as a black precipitate under the light microscope and which. according to Rosen (1974), is visible as an electron dense precipitate in the electron microscope. Cassidy and Light- foot, however, were not able to visualize cobalt sulfide in the electron microscope, so they modified the procedure to convert it to lead sulfide which is presumably more electron dense. Because the Cassidy and Lightfoot modifi— cation was not known until after the present investigation, the technique of Rosen and Musser (1972) was followed in the current study. MATERIALS AND METHODS Utilizing the techniques of Rosen and Musser (1972) and Hansson (1967) with consideration for the precautions which were determined to be necessary against clotting, the following procedure was used for the histochemical locali- zation of carbonic anhydrase activity within the choroid and pseudobranch. Rainbow trout, Salmo gairdneri, were taken from holding tanks kept at 12 : l.0°C and anaesthetized with MS 222 (tricane methane sulfonate). The fish were then heparinized (0.03 U.S.P. units/g total body weight) through the caudal vein and released into fresh water for ten minutes to allow circulation of the heparin. They were then single pithed. Pseudobranchs were removed and the eyes were enucleated and placed immediately into a glass petri dish containing cold (ca. 4°C) 0.17 M cacodylate buffered, 3% glutaraldehyde containing 7% sucrose and 2 U.S.P. units per ml Panheprin. The sclera was removed from over the choroid, and the cor- nea, iris, and lens were removed from the eyes while immersed in the fixative. The pseudobranchs were cut into strips of about 1 x 2 x 5 mm. These tissues were then placed in fresh fixative and allowed to stand for three hours at 4°C with hourly changes of fixative. At the end of this period, the tissues were washed three times in cold (4°C) buffer (0.17 M cacodylate buffered 0.9% saline with 7% sucrose) at ten 69 70 minutes per step. They were then frozen onto cryostat chucks with Ames 0.C.T. Compound (Ames Company, Elkhart, Indiana, 46514). Frozen tissues were then sectioned at 8-10 pm and picked up on Millipore filters (25 pm thick, 0.45 pm pore size). These sections were then floated on the surface of an incubation medium (1.75 x lO-BM C050”, 2 l .. -3 .. 5.3 x 10 M H280”, 4.7 x 10 M KH2P04, 1.57 x 10 .M NaHCOB). As a control, adjacent sections were floated simultaneously 5 on an identical incubation medium containing 10- M sodium acetazolamide. Tissues were not allowed to dip below the surface of the incubation medium which would prevent the staining reaction. The concentration of NaHCO3 is ten times the concentration of that published by Rosen (1970), but the same as that published originally by Hansson (1967), a point which caused much difficulty at the onset of the experiment. The incubation time was ten minutes. This was followed by flotation of the sections sequentially for: two minutes in a wash solution (6.7 x lO-QM KH2P04’ 0.17 M cacodylic acid, 7% sucrose); three minutes in a blackening solution (0.6% (NH4)ZS); and twice at two minutes per step in a 0.9% saline solution. Wet tissues were then placed on a microscope slide and observed under a light microscope to check for proper stain- ing and occasionally for photography. Tissues were then either permanently mounted on glass slides or embedded for ultra-thin sectioning for electron microscopy. 71 For light microscopy, sections were dehydrated in a 70°C oven for several hours and then mounted on a glass slide with Permount. The Millipore filter usually remained attached to the section throughout the procedure so that it was dehydrated and mounted along with the tissue. Both cleared upon mounting in Permount. For electron microscopy, tissues were dehydrated in a graded ethanol series (30%, 50%, 70%, 95%, and 100%) at 15 ndnutes per step and in fresh 100% ethanol overnight. When- ever possible, the Millipore filters were removed during the dehydration process because they seemed to dissolve and interfere with the complete polymerization of the Spurr's emmedding medium with which the tissues were impregnated and molded. Sections were then taken on a Porter-Blum Ultra— ndcrotome MT—2. Some of the experimental sections were observed unstained in order to detect the 008 precipitate. Other experimental sections and the acetazolamide inhibited Sections were counterstained with lead citrate and uranyl acetate according to the method of Reynolds (1963). Also, pseudobranch tissues were fixed in heparinized 'Karnovsky's fixative and postfixed in 2% osmium tetroxide, dehydrated, and embedded in Spurr's embedding medium as above. RESULTS Light Microscgpic Carbonic Anhydrase Histochemistry Initial attempts at visualization of the foci of carbonic anhydrase activity within the choroid rete mirabile utilizing light microscopy were made. The results of these initial investigations are shown in Figures 30-33. It took a great deal of effort to obtain the proper staining conditions under which it could be stated with certainty that the stain seen was actually the result of carbonic anhydrase activity, and that it could be inhibited at low concentrations of acetazolamide. The technique is difficult in that the tissues must be floated on the surface of the incubation medium so that carbon dioxide may be lost to the atmosphere allowing a very localized alkalinization to occur. If any portions of the tissue had dipped below the surface of the incubation medium, the rate of CO2 diffu- sion away from the site of activity would become rate limit— ing and the alkalinization and subsequent salt deposition. would not occur. This was often found to be a major problem, even when using a piece of Millipore filter as a raft upon upon which the tissues were floated and subsequently trans- ferred from dish to dish. Realizing that the rete mirabile itself has its own focus of carbonic anhydrase activity from the evidence of Fairbanks, Hoffert and Fromm (1974), only those areas which 72 73 stained positively were interpreted as having been treated properly. Any unstained areas were deemed to be portions which had sunk below the media surface. Figures 30—33 were interpreted initiallywith this in mind. In these micrographs it is seen that certain regions of the rete contain very intense loci of carbonic anhydrase activity. Recall that in the short-circuit model for oxygen concentration proposed by Fairbanks a: al. (1974), it was hypothesized that the retial carbonic anhydrase was locali- zed to the efferent vessel endothelia, and indeed evidence was presented which suggested that it may have been bound to the luminal membrane of the endothelium versus an intra- cellular site. Neither of these hypotheses was resolvable by the evidence provided by the micrographs at this level of magnification. The carbonic anhydrase activity attribu- table to endothelial vessel walls was seen to extend beyond the choroidal rete itself into the distribution and/or collection vessels (Figure 33). This micrograph suggested that the vessel was entirely lined by carbonic anhydrase activity and that the activity was not in this instance irestricted only to the endothelial membrane lining the vessel lumen. Choroidal Electron Microscopic Carbonic Anhydrase Histochemistry An electron microscopic study was initiated in order to further visualize the exact foci of carbonic anhydrase activity. In preparing the tissues for embedding after 74 staining, thick sections were taken and observed under the light microscope in order to select the best stained areas. When this was done, it was noted that there was a definite localization of carbonic anhydrase activity toward the peri- pheral side of the rete vessels (Figure 34), but it was not possible to determine whether the activity was localized to the afferent or efferent vessel. Upon re—examination of the previous slides of the choroidal rete, it became evident that this was a persistent pattern so that what was once be— 1ieved to have been improperly stained tissue was frequently actually inactive retial tissue located toward the central side of the rete mirabile. The results of the electron microscopic examination of the choroidal rete tissues are presented in Figures 35-39. The appearance of the tissue was disappointing. The stain was not discrete as large, distinct grains of electron? dense precipitate marking beyond doubt the precise focus of activity. Rather, it was a very fine precipitate whose locali- zation had often to be told by finer distinctions between different intensities of gray. Thus, only the most electron dense regions were interpreted to be areas of positive stain. Cassidy and Lightfoot (1974) also reported difficulty with this aspect of the technique. Basically, two patterns of staining were seen within the retial endothelial cells. In Figures 37 and 38, dark bands of stain are seen to run through the cytoplasm. These 75 areas do not appear to be associated with any particular subcellular organelles, but more like localized regions of concentration within the cellular cytoplasmic sap. The endoplasmic reticular network is often depicted in cell biology as a large lacunar network of interconnected con- centric shells which almost compartmentalize the cytoplasm. Thus, this observation may not be as surprising as first seemed. There was no evidence in this pattern which sug— gested in any way, that the activity may have been localized to the luminal membrane or to be in areas of direct com- munication with the vessel lumen. A second pattern of activity was displayed in Figure 39. Here, the activity seemed to be homogeneously distrib— uted throughout the cytoplasm rather than in concentrated regions. More importantly, however, the activity seems to be located specifically in the endothelial cell of one vessel and not in the endothelia of two adjacent vessels. This is the only evidence seen which might suggest that the carbonic anhydrase activity might be located specifically in one vessel type and not in another. Due to the nature of the preparation it is impossible to determine whether the active cell borders on afferent or efferent vessels. Because, efferent vessels are often seen to abut one another, whereas afferent vessels rarely do (Figure 12), it might be probable that the two adjacent inactive endothelia belong to efferent vessels and the active one is of an afferent vessel. This 76 may not be concluded with certainty, however. No electron microscopic observations were made on the apparent carbonic anhydrase activity in the peripheral distribution and collection vessels. Electron Microscopic Pseudobranch Morphology The pseudobranch was next observed in order to detect a correlation between the ultrastructural localization of the carbonic anhydrase activity and any ultrastructural features which may have been indicative of secretory acti— vity. The pseudobranch contains a very characteristic cell type known as the acidophilic cell or pseudobranch-type cell (Figures 40-42). This is characterized as having a system of densely packed parallel mitochondria localized at the midway point between the apex of the cell and the basal bor- der of the cell which lies adjacent to the capillary endo- thelium (Figure 40). Associated with these mitochondria is an extensive network of agranular or smooth endoplasmic reticulum which appear to originate in the vicinity of the mitochondria (Figure 41) and empty into the interstitial space at the basal border of the cell (Figure 42). The basal border of the pseudobranch—type cell is marked also by the presence of a distinct basement membrane. The endothe- lial cells associated with the pseudobranchial capillary network appears to closely resemble the pillar cells present within the secondary lamellae of the gill. They are colum- nar in shape sending flanges out at either end which extend 77 halfway around the capillary lumen (Figure 40). This endo- thelial wall has been reported to be fenestrated in Perca fluviatilis (Laurent and Rouzeau, 1972), but no fenestra— tions were observed in Salmo gairdneri. A point of interest is the fact that there appeared to be two distinct forms of pseudobranch-type cells. Both were identical morphologically, but one counter-stained much more densely, almost to the point of obscuring its ultra- structure. A similar observation was briefly mentioned by Newstead (1964) in the chloride cells of the tide pool sculpin, Oligocottus maculosus. In that the chloride cell has its own very distinctive morphology (Harb and Copeland, 1969), the possibility that the more electron dense form may have been a chloride cell has been rejected. The func- tional significance of these two forms of pseudobranch—type cell in Salmo gairdneri remains to be elucidated. One final observation which was made, was the occur- rence in the pseudobranch-type cell of a very dense packing of tubules such that in cross—section they appeared to form a microrete. Unfortunately, this was observed once and not photographed. Abel (1973) has reported that he found a similar arrangement of tubules in the chloride cell of Salmo gairdneri. No explanation of its role was offered, and none may currently be advanced in the present investi- gation. 78 Pseudobranch Carbonic Anhydrase Histochemistry With the light microscope, the carbonic anhydrase stain appeared to be dense and ubiquitous throughout the pseudobranchial tissue. This is consistent with the fact that the pseudobranch has a very high carbonic anhydrase activity (Hoffert, 1966), but inconsistent with the findings of Laurent, Dunel and Barets (1969) who, using a similar technique, found the stain within the pseudobranch-type cell to be localized to the tubular system in the basal border of the cell at a light microscopic level. The electron microscopic investigation of carbonic anhydrase activity in the pseudobranch, like that in the choroidal rete tissue, proved to be disappointing. The results are shown in Figures 43, 44 & 45. Again, the nuclei were seen to "artifactually" stain heavily in this prepara- tion. In comparison to the control acetazolamide inhibited tissues, a ubiquitous background cytoplasmic stain was seen and the mitochondrial matrix seemed denser as well (Figure 45). Very close inspection of Figure 43 reveals, however, that there is a very distinct black grainy precipitate concentrated around the nuclear envelope of the pseudobranch cell and it is detectable in widely dispersed locations throughout the cell. There is a distinct possibility that this may be the carbonic anhydrase precipitate and that the difference in density of the background of Figure 45 may be due to variations in tissue thickness. A second possi- 79 bility is that the cell seen is a dense form of the pseudo— branch cell versus a light form. If this is the case, then the actual amount of this grainy precipitate present in the cell of Figure 43 nowhere approaches the amount which must have been present at the time of light microscopic examina- tion. This would indicate that an extraction of the preci- pitate may have occurred somewhere in the dehydration and/or embedding process. It had been observed upon trimming of the polymerized blocks for subsequent ultramicrotomy, that the tissues had seemed less densely stained than before embedding, but this was merely a subjective observation and not considered'serious at that time. A further refinement of the technique is deemed desirable at this point. Decisive conclusions regarding the localization of carbonic anhydrase activity within the pseudobranch should await resolution of this problem. DISCUSSION Interpretation of Carbonic Anhydrase Histochemistry The information contained within this section is based entirely upon the reliability of the histochemical method utilized for the demonstration of carbonic anhydrase acti- vity. Difficulties were encountered in the interpretation of the electron micrographs. Consequently, it would be appropriate to justify the basis for the interpretations presented and to reiterate the degree of reliability which may be placed upon each of these interpretations. Perhaps the most valuable information comes from the study of the choroidal rete mirabile under the light micro- scope. Here it was seen that the carbonic anhydrase inten— sifies in activity down the length of the rete toward the periphery. Utmost confidence should be placed in the valid- ity of this conclusion due to the consistency of these observations over all sections viewed, and to the abolition of the pattern upon treatment with acetazolamide. That lcarbonic anhydrase activity was seen to be associated with vessels beyond the rete is also a certainty. Once the technique was refined, no problems were en- countered with light microscopic carbonic anhydrase his- tology. Because identical procedures were used in staining the tissues embedded for electron microscopy, no problems 80 81 were anticipated in the subsequent electron microscopic investigation. Several authors had used this technique in electron microscopy with apparent success in a variety of tissues, including the rabbit retina (Bhattacherjee, 1971), frog and turtle lungs (Fain and Rosen, 1973), turtle and toad urinary bladder (Rosen, 1972a), and toad, turtle, pigeon, rat, rabbit, dog, and monkey renal tissues (Rosen, 1972b). Most of this work had been done by Seymour Rosen and col- leagues at Harvard Medical School and he has recently pub- lished a chapter in an electron microscope histochemistry text (Rosen, 1974). Without exception the reaction product visualized in the electron micrographs published in these works has been a very fine precipitate. No micrographs of acetazolamide inhibited tissues, however, have ever accompanied these figures. Rather "control" sections used were tissues which had been fixed in osmium tetroxide and glutaraldehyde and subsequently counter-stained with lead citrate and uranyl acetate in much the same manner utilized in the Part I of this current study. In the most dramatic of the micrographs, a very high magnification of the carbonic anhydrase stained tissue appears as if it could be a negative of the normally prepared tissues (Musser and Rosen, 1973). It is normal histochemical procedure to observe the locali- zation of reaction product in comparison with an identically treated tissue in which only the enzyme of interest has been inhibited. With the carbonic anhydrase stain, this is 82 particularly important due to the occurrence of significant amounts of reaction product deposited as a result of the uncatalyzed reaction. In addition, the appearance of tissues is very different when osmium tetroxide fixation is omitted. Among other things, lipids will be extracted and membranes will be absent. Hayat (1970) presents an extensive descrip- tion of the differences to be expected. The problems encountered in interpretation of the electron micrographs of the current investigation should now be familiar. Cassidy and Lightfoot (1974) have noted that they, too, have had difficulty in the visualization of the cobalt sulfide-reaction product in their tissue preparations and found it necessary to modify their techniques such that the cobalt salt is converted by one additional step into lead sulfide (CoS + Pb(N03)2--PbS) which appears to be a nmch more electron dense precipitate. The electron micro- .graphs published by these authors seem to be much more con- 'Vincing and their technique should be seriously considered iin any future histochemical studies. In the current study, the differences between carbonic aIlhydrase stained tissues and acetazolamide inhibited tissues rnEiy be interpreted in two ways. First, the difference was <31?ten_dependent upon a subjective, qualitative estimate of d~€Egrees of grayness (Figure 45). Secondly, the possibility ‘tiklat the grainy regions in Figure 43 could be the true re- Ea‘<3=tion product has also been pointed out. Conclusions based 83 upon the electron microscopic investigations must, therefore, be interpreted with these reservations in mind. Choroidal Carbonic Anhydrase It was decided that only the most electron—dense regions in the electron micrographs presented herein could be interpreted as valid signs of carbonic anhydrase activi- ty. The two observations which appeared to be the most con— clusive are based upon Figures 37, 38 & 39. The banding patterns of very electron—dense stain seen in areas of the rete endothelial cytoplasm were not seen in acetazolamide treated tissues (Figure 36). As stated previously, they do not appear to be associated with any specific subcellular organelles and their significance is not known at this time. Figure 39 provides more difinite contrast within a single micrograph and may perhaps be interpreted with more confi- dence. Here, evidence is seen for a difference in activi- ties between vessels. In summary, the evidence which has been accumulated is: 1) there is a definite polarization of carbonic anhydrase activity along the length of the rete, the higher concentra- tions occurring toward the peripheral side, 2) the activity may include sites in larger vessels located distal to the rete itself, and 3) there is some evidence for the existence of a difference in activity between the two classes of retial vessel. It cannot yet be stated with certainty that the regional localization of carbonic anhydrase activity is also 84 accompanied by a structural differentiation in the endo— thelial cells. In Part I of this study, no ultrastructural differences were seen along the length of the retial vessels. Wittenberg and Wittenberg (1974) noticed a distinct difference between the ends of the rete on a gross structural level, but not at an ultrastructural level, Any model for the operation of the oxygen concentrating mechanism must now be able to account for these findings in addition to the existing evi- dence. Functional Implications Fairbanks, Hoffert and Fromm (1974) have most conclusive- ly demonstrated the dependence of the oxygen multiplication mechanism upon a source of carbonic anhydrase found within the choroidal rete vasculature itself. That this activity does exist is demonstrated in the current study. According to the short-circuit theory presented by Fairbanks a: a1. (1974) the carbonic anhydrase was postulated to have been localized in the efferent rete endothelia. In this position the enzyme system could operate to sequester carbon dioxide and convert the carbon dioxide to bicarbonate before the gas has a chance to diffuse down its concentration gradient and cause a premature release of oxygen by the Bohr and Root shifts in the afferent rete capillaries. The finding that the carbonic anhydrase is localized to the peripheral end of the rete is certainly consistent with this notion. The highest concentration of carbonic anhydrase activity is in 85 the location where the highest carbon dioxide concentrations are encountered in the returning blood. The assumption now is that the quantities of carbonic anhydrase at this end are adequate to handle all excess carbon dioxide which would cause a short-circuiting of the multiplier before reaching the central end of the rete. A restriction of the activity to only the efferent vessel endothelia would seem to be unnecessary. As long as the carbonic anhydrase were inter- jected somewhere along the diffusion path of the carbon dioxide it would be capable of operating in the proper man- ner. The current study has presented evidence that carbonic anhydrase may indeed be restricted to only one vessel type (Figure 39). Even though it is impossible to state to which category of vessel the apparent activity is restricted on the basis of ultrastructural features, the likelihood of the active vessel being an afferent vessel is fairly good. As already stated, because the efferent vessels are often seen to abut one another, whereas afferent vessels rarely do, it is probable that the two adjacent inactive endothelia belong to efferent vessels and the active one is of an af- ferent vessel. Thus, based upon evidence presented above, a high degree of carbonic anhydrase activity is suggested to be localized in the peripheral end of the afferent rete vessel endothelia. 86 Note that the carbonic anhydrase activity also exists in the endothelia of the peripheral vessels. If the activity were found to be restricted to only one vessel type (i.e., distribution vs. collection vessel), then the distribution vessel would likely be the one containing carbonic anhydrase activity. In this case it would merely represent an exten- sion of endothelial cellular specialization for carbonic anhydrase production from the afferent retial vessels. At this time, however, this postulate must remain speculative. Due to the rapidity of inhibition of the oxygen multi- plication mechanism by the non-diffusable carbonic anhydrase inhibitor, CL-ll,366, Fairbanks a; al. (1974) speculated that the carbonic anhydrase activity may have been restrict- ed to membrane receptor sites or some sort of lacunae in the endothelial cell which was in direct contact with the vessel lumen. This was not seen to be the case in the pre- sent study. The activity in all cases seemed to be locali- zed in the cytoplasm. In accounting for the rapidity with which CL—ll,366 worked, and for the fact that its accumula— tion only in the rete paralleled the time course of inhibi- tion, possibly pinocytotic activity of these endothelia may account for this observation. No such pathway for uptake would exist in either the erythrocytes or the retinal cells whose carbonic anhydrase may also be involved. The pseudobranchial ultrastructure in Salmo gairdneri was seen to fit the description for other species. Fenestrae 87 present in the capillary endothelium adjacent to the basal border of the pseudobranch-type cell have been seen in Perca fluviatilis (Laurent and Rouzeau, 1972), but were not seen in S. gairdneri. Were the pseudobranch active in the secre- tion of carbonic anhydrase or any other large molecular species, one would suspect the presence of endothelial fenes- trae as a passageway into the vessel lumen. In addition, a cell type specialized for carbonic anhydrase synthesis would be expected to contain ultrastructural features more characteristic of protein synthesis (eg. rough endoplasmic reticulum, free ribosomes, and mitochondria). Smooth endo- plasmic reticulum in association with mitochondria as found in the pseudobranch are seen in other tissues and have been implicated in steroid synthesis and glycogen and lipid meta- bolism (Rhodin, 1974). In a light microscopic study, Laurent. Dunel and Barets (1969) demonstrated polarization of carbonic anhydrase activity to the basal pole of the pseudobranch—type cell which is characterized by a dense tubular network. Although carbonic anhydrase within this cell was not conclusively localized in the current study at the resolution attainable by electron microscopy, it may be stated with confidence that no stain was seen within the tubules of this region. Thus, it is unlikely that carbonic anhydrase might be produced in the vicinity of the mito- chondria and passed through the tubules to the capillary 88 lumen, to be carried to the swimbladder and choroidal retia as has been implied previously. SUMMARY AND CONCLUSIONS Problems do exist with the histochemical technique for carbonic anhydrase localization at the electron micro- scopic level. The results have been interpreted with these reservations in mind. Recommendations that future attempts be made utilizing the techniques recently proposed by Cassidy and Lightfoot (1974) have been discussed. Based upon the findings presented it has been found that carbonic anhydrase activity within the choroidal rete mirabile is localized to the peripheral half of the rete and may extend out into the peripheral distribution and collec- tion vessels. Suggestions have been made that this activity is further restricted to the afferent rete endothelium. These sites of activity have also been shown to be consis- tent with the short-circuit model for oxygen multiplication. This mechanism requires that retial carbonic anhydrase be made available at the proper location in order to intercept diffusing carbon dioxide and thus prevent the premature release of oxygen in the afferent blood. The normal ultrastructure of the pseudobranch—type cell was studied and found to be structurally similar to those investigated elsewhere. A light and dense form were identi- fied and the presence of a microretial arrangement of smooth endoplasmic reticulum was noted. The significance of these structures is not known. Due to the nature of the results 89 90 of the histochemical study no conclusions may be drawn re- garding the possibility that pseudobranchial secretion of carbonic anhydrase may occur for subsequent usage in the choroidal counter-current oxygen multiplier. COMBINED SUMMARY AND CONCLUSIONS Before the onset of any detailed carbonic anhydrase localization studies, it became necessary to become thor— oughly familiar with the vascular pattern and normal ultra— structural features of the vessels within the choroidal layer of Salmo gairdneri. Both light and electron micro- scopic investigations were pursued. The major findings and conclusions are dividable into three main categories: technique information, morphological information, and carbonic anhydrase histochemical information. The latter of these categories is then useful in clarifying the current understanding of the mechanism behind the opera- tion of the counter—current oxygen multiplier. Techniques Findinga Methods of tissue preparation may drastically affect the appearance of delicate endothelial tissue in the choroid rete mirabile under the electron microscope. Perfusion fixation of this tissue was particularly dangerous in that delicate efferent endothelial walls may easily be damaged. Hence, an optimal immersion technique was sought. Teleost blood clots very rapidly, therefore, special precautions need to be taken to prevent clot formation which is seen to drastically affect the appearance of the vascular tissues in particular. The final optimal technique involved ip vivo 91 92 heparinization of the ocular tissues prior to enucleation and fixation in Karnovsky's fixative. Technical problems also arose related to the visuali- zation of the carbonic anhydrase stain reaction product, cobalt sulfide, under the electron microscope. The precipi- tate appeared very fine and difficult to visualize. This problem was delineated and all interpretations were made with this reservation in mind. It was noted that other investigators had apparently found the same difficulty and a newer technique utilizing lead sulfide as the electron dense precipitate is recommended for further studies. Morpholpgical Findings The general outlay of the choroidal vasculature has been carefully examined and found to be similar to patterns reported elsewhere (Barnett, 1951; Wittenberg and Wittenberg, 1974; Copeland, 1974a). The arterial manifold wall is seen to be a thick structure with a single layer of endothelia facing the lumen of the arterial manifold and venous sinu- soid, and sandwiching a thick collagenous layer through which run fibroblasts, smooth muscle cells, and myelinated and non-myelinated neurons. The possible importance of the arterial manifold as an effector organ for the regulation of rete counter-current multiplier operation is discussed. The rete mirabile ultrastructure reported here is slightly different from the structures reported elsewhere for both the swimbladder and the choroid retia. The vessel packing 93 pattern is cubic, with an equal number of afferent and ef- ferent vessels. The endothelia lining both afferent and efferent vessels appear essentially identical in structure. Nonetheless, afferent vessels are differentiable on the basis of their size (4.5 um vs. 9.5 um) and by the fact that only afferent vessels are seen associated with pericytes and basement membranes. Peripheral distribution and collection vessels pass through a loose, laminar stroma containing numer- ous melanocytes. No features were seen to distinguish col- lection from distribution vessels. The choriocapillaris and associated Bruch's membrane and pigment epithelium were seen to be identical to those of Esox lucius (Braekevelt, 1974). The choriocapillaris is relatively nonfenestrated compared to the choriocapillaris of humans. Bruch's mem— brane has fewer structural features than that in the human and the pigment epithelium lacks a convoluted basal border typical of the human pigment epithelium. The ultrastructural features of the pseudobranch-type cells were identical to those reported elsewhere. The most characteristic features are numerous densely packed mito- chondria associated with an extensive network of smooth endoplasmic reticulum. In addition, it was noted that two morphologically identical forms of the cell were seen dif— fering in their affinity for normal stains. 94 Carbonic Anhydrase Histochemistry Light microscopic observations revealed that within the choroidal vasculature carbonic anhydrase activity is found to increase in intensity toward the peripheral side of the rete vessels. It was also seen to be located in the walls of some larger vessels peripheral to the rete. Electron microscopic observations were interpreted with great caution. It appeared, however, that carbonic anhydrase activity in the rete endothelia was within the general cytoplasm either in bands or spread evenly through- out the cytoplasm. No evidence was seen for a luminal mem- brane-bound focus of carbonic anhydrase activity. Evidence was cited for the localization of carbonic anhydrase to one type of retial vessel and not another, and that the active vessel was probably afferent. It is suggested that retial carbonic anhydrase activity is restricted to the peripheral side of the afferent rete endothelium and extends out into the endothelia of the dis- tribution vessels. It is argued that location is sufficient to prevent the short—circuiting of the counter—current multiplier. RECOMMENDATIONS As in any study, more questions remain unanswered than answered. A few of these questions which are seen to be of immediate significance are listed below. Morphological studies on the choroidal vasculature re- main incomplete. The structure of the venous sinusoid re- mains to be elucidated. Laurent and colleagues at the Laboratoire de Neuro- physiologie Generale du College de France, have done the most extensive work on the pseudobranch and their studies have lent the most credible evidence thus far offered as regards pseudobranchial function. Their work had included light microscopic studies of innervation, in which neural tracts have been traced from their central origins to their termina— tions in the pseudobranch (Laurent and Dunel, 1966); histo- chemical studies of carbonic anhydrase activity (Laurent, Dunel and Barets, 1969); and perfusion studies in conjunc— tion with microelectrode recordings of afferent neural im— pulses which has revealed much information on receptor func— tion (Laurent and Rouzeau, 1972). The current study has shown the presence of neural elements within the choroidal arterial manifold wall and the suggestion has been made that there may exist a regulatory role for this structure in the operation of the oxygen multiplier mechanism. A series of studies very similar to those mentioned above 95 96 would yield invaluable evidence to support or contradict this hypothesis. If such a mechanism were discovered, it would be interesting to compare its responses with the auto— regulatory responses seen in the ocular vasculature of higher forms. There would be reason to believe that this mechanism might also be an integral part of other retial systems, such as that in the swimbladder rete, and possibly, the thermo— regulatory retia in the carotid arterial pathway to the brain of some desert antelopes, or in the retia of the flippers of the whale. Within the choroidal rete mirabile a localization of carbonic anhydrase activity along the length of the rete has been demonstrated. Although no evidence was seen regarding a parallel structural differentiation in the current study, the results are not conclusive, and a more detailed exam- ination of this possibility would definitely be instructive. The structures of the choriocapillaris, Bruch's mem— brane, and pigment epithelium in Salmo gairdneri have been noted to be markedly different from those seen in the human eye. It has been suggested that in the higher vertebrates, the choroidal vasculature is the major site of nutrient ex— change while the retinal circulation is largely the site of gas exchange (Braekevelt, 1974). Because there is no retinal circulation in teleosts all exchange processes must occur at the choriocapillaris. It would naturally follow that the teleost morphology would display structures which are 1.. 97 associated with an augmented exchanger role, such as numer- ous pores or fenestrae. The opposite is seen, however. The degree of fenestration is vastly reduced in the two teleosts now studied. It would seem that classical electron micro—7 scopic tracer studies are called for here, perhaps starting with such tools as horseradish peroxidase (Karnovsky, 1967). The carbonic anhydrase histochemical technique ob- viously needs improvement. The procedures utilized by Cassidy and Lightfoot (1974) have been recommended as the next viable alternative. When this procedure is perfected for use in electron microscopy, the findings which were seen in this study should be reconfirmed and extended. It would be of value to determine the extent of the activity pre- sent in the peripheral vessels of the choroid. Also, with an important functional significance described for this enzyme in the choroidal rete, it would seem likely that a similar function might exist certainly in the swimbladder rete and perhaps in other counter-current systems. This possibility should be investigated. The subcellular local- ization of carbonic anhydrase activity in the pseudo- branch—type cells would also be of value. If morphological evidence existed for the secretion of carbonic anhydrase by the pseudobranch, one would then logically search for a specific isoenzyme of carbonic anhydrase produced by the pseudobranch and to trace its possible role in the choroidal and swimbladder counter—current multipliers. 98 Finally, as a comparative study, phylogenetic and de— velopmental variations on the general scheme drawn from the evidence from the above studies might be researched. LITERATURE C ITED LITERATURE CITED Abel, P. D. 1973. An Unknown Structure in the Chloride Cells of the Gill Epithelium of the Rainbow Trout (Salmo gairdneri). Z. Zellforsch., 146:293-295. Anderson, D. H. and S. K. Fisher. 1975. Disc Shedding in Rodlike and Conelike Photoreceptors of Tree Squirrels. Science, 187: 953-955. Arnott, H.J., A. C. G. Best, S. Ito and J. A. C. Nicol. 1974. Studies on the Eyes of Catfishes with Special Reference to the Tapetum Lucidum. Proc. R. Soc. LOl’ld. Bu, 186313“36o Barnett, C. H. 1951. The Structure and Function of the Choroidal Gland of Teleostean Fish. J. Anat., 85:113-119. Barraco, D. A. 1975. Thermal Consideration of the Choroidal Gland Rete Mirabile in the Rainbow Trout (Salmo gairdneri). M.S. Thesis, Michigan State University, East Lansing, Michigan. Bennett, H. S., V. H. Luft and J. C. Hampton. 1959. Morphological Classifications of Vertebrate Blood Capillaries. Amer. J. Physiol., 196:381-390. Bergman, H. L. 1973. Regulation of Blood Flow Through the Isolated-Perfused Gills of Rainbow Trout: Effects of Vasoactive Agents on Functional Surface Area. Ph.D. Thesis, Michigan State University, East Lansing, Michigan. Bernstein, M. H. and M. J. Hollenberg. 1965. Fine Structure of the Choriocapillaris and Retinal Capillaries. Invest. 0phtha1., 4:1016-1025. Bhattacherjee, P. 1971. Distribution of Carbonic Anhydrase in the Rabbit Eye as Demonstrated Histochemically. Exp. Eye Res., 12:356-359. Braekevelt, C. R. 1974. Fine Structure of the Retinal Pigment Epithelium, Bruch's Membrane, and Chorio— capillaris in the Northern Pike (Esox lucius). J. Fish. Res. Bd. Canada, 31:1601-1605. 99 100 Brown, L. J., A. L. Stalker and J. Hall. 1969. Experimental Defibrination. II. Histological Studies by Light and Electron Microscopy. Microvas. Res., 1: 295—307. Bruns, R. R. and G. E. Palade. l968a. Studies on Blood Capillaries: I. General Organization of Blood Capillaries in Muscle. J. Cell Biol., 37:244—276. Bruns, R. R. and G. E. Palade. 1968b. Studies of Blood capillaries: II. Transport of Ferritin Molecules Across the Wall of Muscle Capillaries. J. Cell Biol., 37:277—299. Cassidy, M. M. and F. G. Lightfoot. 1974. "Carbonic Anhydrase," in Electron Microscppy_of Enzymes: Principles and Methods. Edited by M. A. Hayat. Vol. 3, pp. 77-106. VanNostrand Reinholt Co., New York. Castro-Correia, J. 1967. Studies on the Innervation of the Uveal Tract. Ophthalmologica, 154:497-520. Clementi, F. and G. E. Palade. 1969. Intestinal Capillaries: I. Permeability to Peroxidase and Ferritin. J. Cell Biol., 41:33-58. Copeland, D. E. 1951. Function of Glandular Pseudobranch in Teleosts. Amer. J. Physiol., 167:775. Copeland, D. E. 1974a. The Anatomy and Fine Structure of the Eye of Teleost. I. The Choroid Body in Fundulus gpandis. Exp. Eye Res., 18: 547-561. Copeland, D. E. 1974b. The Anatomy and Fine Structure of the Eye of Teleost. II. The Vascular Connections of the Lentiform Body in Fundulus grandis. Exp. Eye Res., 19:583-589. Copeland, D. E. and A. J. Dalton. 1959. An Association Between Mitochondria and the Endoplasmic Reticulum in Cells of the Pseudobranch Gland of a Teleost. Biophys. Biochem. Cytol., 5:393—395. D'Aoust, B. G. 1970. The Role of Lactic Acid in Gas Secretion in the Teleost Swim Bladder. Comp. Biochem. Physiol., 32:637-668. Dorn, E. 1961. Uber den Feinbau der Schwimmblase Von Anguilla vulgaris L: Licht— und Elektronenmikro- skopische Untersuchungen. Z. Zellforsch., 55:849-912. 101 Fahlén, G. 1967. Morphological Aspects on the Hydrostatic Function of the Gas Bladder of Clupea harengus L. Acta Univ. Lund., 28:1-37. Fain, W. and S. Rosen. 1973. Carbonic Anhydrase Activity in Amphibian and Reptilian Lung: A Histochemical and Biochemical Analysis. Histochem. J., 5:519-528. Fairbanks, M. B. 1970. The Importance of the Enzyme Carbonic Anhydrase in Ocular Oxygen Concentration in the Rainbow Trout, Salmo gairdneri. Ph.D. Thesis, Michigan State University, East Lansing, Michigan. Fairbanks, M. B., J. R. Hoffert and P. O. Fromm. 1969. The Dependence of the Oxygen-Concentrating Mechanism of the Teleost Eye (Salmo gairdneri) on the Enzyme Carbonic Anhydrase. J. Gen. Physiol., 54:202—211. Fairbanks, M. B., J. R. Hoffert and P. 0. Fromm. 1974. Short Circuiting of the Ocular Oxygen Concentrating Mechanism in the Teleost Salmo gairdneri Using Carbonic Anhydrase Inhibitors. J. Gen. Physiol., 64:263-273. FAnge, R.’ 1953. Mechanisms of Gas Transport in the Euphysoclist Swim Bladder. Acta Physiol. Scand., 30 31-1330 Fawcett, D. W. 1961. Intercellular Bridges. Exp. Cell Res., Supp1., 8:174-187. Fawcett, D. W. 1966. An Atlas of Fine Structure: The Cell. Its Organelles and Inclusions. W. B. Saunders Co., Philadelphia, 448 p. Fawcett, D. W. and J. Wittenberg. 1959. The Fine Structure of Capillaries in the Rete Mirabile of the Swim Bladder of Opsanus tau. Anat. Rec., 133:274. Fawcett, D. w. and J. Wittenberg. 1962. Structural Speciali- zations of Endothelial Cell Junctions. Anat. Rec., 142:231. Fonner, D. B., J. R. Hoffert and P. O. Fromm. 1973. The Importance of the Counter Current Oxygen Multiplier Mechanism in Maintaining Retinal Function in the Teleost. Comp. Biochem. Physiol., 46Az559-567. Francois, Haldane, Hansson, Harb, J. Hayat, M. Hillman, Hoffert, Hoffert, Hoffert, Hoffert, 102 J. and A. Neetens. 1974. "Comparative Anatomy of the Vascular Supply of the Eye in Vertebrates," in The Eye: Comparative Physiology. Edited by H. Davson and L. T. Graham, Jr. Vol. 5, pp. 1—66. Academic Press, Inc., New York. J. S. 1927. Respiration. Yale Univ. Press, New Haven. 427 p. H. P. J. 1967. Histochemical Demonstration of Carbonic Anhydrase Activity. Histochemie, 11:112-128. M. and D. E. Copeland. 1969. Fine Structure of the Pseudobranch of the Flounder Paralichthys lethostigma: A Description of Chloride—Type Cell and Pseudobranch—Type Cell. Z. Zellforsch., 101:167-174. A. 1970. Principles and Techniqpes of Electron Microscppy: Biological Applications. Vol. I. VanNostrand Reinhold Co., New York. 412p. J. R., W. G. Seliger and P. E. Burk. 1975. "Cytochemical Studies of Protein Uptake by Adrenal Cortical Cells of the Golden Hamster," in Electron Microscopic Concepts of Secretion: Ultrastructure of Endocrine and Reproductive Organs. Edited by M. Hess. pp.—4l94433. John Wiley and Sons, New York. J. R. 1966. Observations on Ocular Fluid Dynamics and Carbonic Anhydrase in Tissues of Lake Trout (Salvelinus namaycush). Comp. Biochem. Physiol., 17:107—114. J. R., G. E. Eldred and P. O. Fromm. 1974. Comparison of Direct and Indirect Calorimetry as a Method for Determining Total Energy Produced by the Isolated Retina of Rainbow Trout (Salmo gairdneri). Exp. Eye Res., 19:359-365. J. R., M. B. Fairbanks and P. O. Fromm. 1971. Ocular Oxygen Concentrations Accompanying Severe Chronic Ophthalamic Pathology in the Lake Trout (Salvelinus namaycush). Comp. Biochem. Physiol., 39:137-145. J. R. and P. O. Fromm. 1965. Biomicroscopic, Gross, and Microscopic Observation of Corneal Lesions in the Lake Trout, Salvelinus namaycush. J. Fish. Res. Bd. Canada, 22:761—766. 103 Hoffert, J. R. and P. O. Fromm. 1966. Effect of Carbonic Anhydrase Inhibition on Aqueous Humor and Blood Bicarbonate Ion in the Teleost (Salvelinus namaycush). Comp. Biochem. Physiol., 18:333-340. Hogan, M. J., J. A. Alvarado and J. E. Weddell. 1971. Histology of the Human Eye: An Atlas and Textbook. W. B. Saunders Co., Philadelphia. 687p. Holliday, F. G. T. and G. Parry. 1962. Electron Micro- scopic Studies of the Acidophil Cells in the Gills and Pseudobranchs of Fish. Nature, 193:192. Jasinski, A. and W. Kilarski. 1971. Capillaries in the Rete Mirabile and in the Gas Gland of the Swim Bladder in Fishes, Perca fluviatilis L. and Misgurnus fossilis L.: An Electron Microscopic Study. Acta Anatomica, 78:210-223. Karnovsky, M. J. 1965. A Formaldehyde-Glutaraldehyde Fixative of High Osmolality for Use in Electron Microscopy. J. Cell Biol., 27:137A-l38A. Karnovsky, M. J. 1967. The Ultrastructural Basis of Capil- lary Permeability Studied with Peroxidase as a Tracer. J. Cell Biol., 35:213-236. Kessei, R. C. and H. W. Beams. 1962. Electron Microscope ’ Studies on the Gill Filaments of Fundulus hetero- clitus from Sea Water and Fresh Water with Special Reference to the Ultrastructural Organization of the "Chloride Cell". J. Ultrastructure Res. 6:77—87. Laurent, P. and S. Dunel. 1966. Recherches sur l'innervation de la pseudobranchie des teleosteens. Arch. d'Anatomie Microscopique, 55:633—656. Laurent, P., S. Dunel and A. Barets. 1969. Localisation histochemique de l'anhydrase carbonique au niveau des chemorecepteurs arteriels des mammiferes, des batraciens et des poissons. Histochemie, 17:99—107. Laurent, P. and J. D. Rouzeau. 1972. Afferent Neural Activity from Pseudobranch of Teleos 3. Effects of P02, pH, Osmotic Pressure, and Na Ions. Resp. Physiol., 14:307-331. Leiner, M. 1938. Die Augen Keimendruse (Pseudobranchie) der Knochen Fische. Experimentelle Untersuchenger Uber ihre Physiologishe Bedeutung. Z. Vgl. Physiol., 26:416—466. 104 Linthicum, D. s. and F. G. Carey. 1972. Regulation of Brain and Eye Temperatures by the Bluefin Tuna. Comp. Biochem. Physiol., 43A:425—433. Lbnnerholm, G. 1974. Carbonic Anhydrase Histochemistry, A Critical Study of Hansson's Cobalt—Phosphate Method. Acta Physiol. Scand. Suppl., 418:1—43. Luft, J. H. 1961. Improvements in Epoxy Resin Embedding Methods. J. Biophys. Biochem. Cytol., 9:409-414. Luna, L. G. 1968. Manual of Histologic Staining Methods of the Armed Forces Institute of Pathology. 3rd ed. McGraw—Hill Book Co., New York. 258p. Maetz, J. 1956. Le role biologique de l'anhydrase carboni— que chez quelques teleosteens. Bull. Biol. France and Belgique, Suppl. 40:1-129. Marshall, N. B. 1972. Swimbladder Organization and Depth Ranges of Deep—Sea Teleosts. Symp. Soc. Exp. Biol., 26:261-272. Musser, G. L. and S. Rosen. 1973. Localization of Carbonic Anhydrase Activity in the Vertebrate Retina. Exp. Eye Res., 15:105—119. Newstead, J. D. 1964. Morphologtal Changes in "Chloride Cells" in the Gills of the Sculpin Oligocottus maculosus after Transfer from Sea Water to Fresh Water. J. Cell Biol., 23:66A. Newstead, J. D. 1971. Observations on the Relationship between the 'Chloride—Type' and 'Pseudobranch-Type' Cells in the Gills of a Fish, Oligocottus maculosus. Z. Zellforsch., 116:1-6. Noell, W. K. 1959. The Visual Cell: Electric and Meta- bolic Manifestations of Its Life Processes. Amer. J. Ophthal., 48:348-370. Pappenheimer, J. R., E. M. Renkin and L. M. Borrero. 1951. Filtration, Diffusion and Molecular Sieving Through Peripheral Capillary Membranes : A Contribution to the Pore Theory of Capillary Permeability. Amer. J. Physiol., 167:13—46. Reynolds, E. S. 1963. The Use of Lead Citrate at High pH as an Electron-Opaque Stain in Electron Microscopy. J. Cell Biol., 17:208-211. lO5 Rhodin, J. A. G. 1967. The Ultrastructure of Mammalian Arterioles and Precapillary Sphincters. J. Ultra- structure Res., 18:181-223. Rhodin, J. A. G. 1968. Ultrastructure of Mammalian Venous Capillaries, Venules, and Small Collecting Veins. J. Ultrastructure Res., 25:452-500. Rhodin, J. A. G. 1974. Histology: A Text and Atlas. Oxford Univ. Press, New York. 803p. Richardson, K. C., L. Jarett and E. H. Finke. 1960. Embed- ding in Epoxy Resins for Ultrathin Sectioning in Electron Microscopy. Stain Tech., 35:313-323. Rosen, S. 1970. Localization of Carbonic Anhydrase Activity in Transportin Urinary Epithelia. J. Histochem. Cytochem., 18: 68—670. Rosen, S. 1972a. Localization of Carbonic Anhydrase Activity in Turtle and Toad Urinary Bladder Mucosa. J. Histochem. Cytochem., 20:696-702. Rosen, S. 1972b. Localization of Carbonic Anhydrase Activity in the Vertebrate Nephron. Histochem. J., 4:35-48. Rosen, S. 1974. "Carbonic Anhydrase — An Alternative Method," in Electron Microscopy of Enzymes: Principles and Methods. Edited by M. A. Hayat. Vol. 3, pp. 109— 131. VanNostrand Reinholt Co., New York. Rosen, S. and G. L. Musser. 1972. Observations on the Specificity of Newer Histochemical Methods for the Demonstration of Carbonic Anhydrase Activity. J. Histochem. Cytochem., 20: 951—954. Schiffman, R. H. and P. O. Fromm. 1959. Measurement of some Physiological Parameters in Rainbow Trout (Salmo gairdneri). Can. J. 2001., 37:25—32. Smith, C. G., W. M. Lewis and H. M. Kaplan. 1952. A Comparative Morphologic and Physiologic Study of Fish Blood. Prog. Fish Cult., 14:169-172. Spitznas, M. and E. Reale. 1975. Fracture Faces of Fenestra- tions and Junctions of Endothelial Cells in Human Choroidal Vessels. Invest. Ophthal., 14:98-107. Spurr, A. R. 1969. A Low-Viscosity Epoxy Resin Embedding Medium for Electron Microscopy. J. Ultrastructure Res.,,26:31-43. 106 Steen, J. B. 1970. "The Swim Bladder as a Hydrostatic Organ," in Fish Physiology. Edited by W. S. Hoar and D. J. Randall. Vol.4, pp. 413—443. Academic Press, New York. Wade, J. B. and M. J. Karnovsky. 1974. The Structure of the Zonula Occludens: A Single Fibril Model Based on Freeze—Fracture. J. Cell Biol., 60:168—180. Walls, G. L. 1963. The Vertebrate Eye and Its Adaptive Radiation. Hafner Pub. Co., New York, 785p. Wittenberg, J. B. and R. L. Haedrich. 1974. The Choroid Rete Mirabile of the Fish Eye: II. Distribution and Relation to the Pseudobranch and to the Swim- bladder Rete Mirabile. Biol. Bull., 146:137-156. Wittenberg, J. B. and B. A. Wittenberg. 1961. Active Transport of Oxygen into the Eye of Fish. Biol. Bull., 121:379. Wittenberg, J. B. and B. A. Wittenberg. 1962. Active Secretion Secretion of Oxygen into the Eye of Fish. Nature, 194:106-107. Wittenberg, J. B. and B. A. Wittenberg. 1974. The Choroid Rete Mirabile of the Fish Eye. I. Oxygen Secretion and Structure: Comparison with the Swimbladder Rete Mirabile. Biol. Bull., 145:116-136. Wolf, K. 1959. Plasmoptysis and Relation of Erythrocytes in Coagulation of Blood of Freshwater Bony Fishes. Blood, 14:1339—1344. Wolter, J. R. 1960. Nerves of the Normal Human Choroid. APPENDIX APPENDIX: FIGURES CITED IN TEXT 107 FIGURE l.-—Microfil casts of the afferent vessels of the choroidal vasculature. The photomicrographs depicted here have been taken from Fairbanks (1970). Microfil, a low viscosity silicone—based injection compound was injected into the afferent vessels of the choroidal vascula- ture, and the surrounding tissues were dehydrated in alcohol, and cleared in methyl salicylate. A. From the back of the eye after removal of the sclera, the choroidal rete mirabile is seen to be a horseshoe—shaped mass of parallel vessels overlying the optic nerve. The optic nerve is not shown here, but its position is at the center of the circle outlined by the margins of the choroidal rete. The horseshoe normally opens ventrally. The large vessel in the inner margin of the horseshoe is the arterial manifold. Normally, this is entirely enclosed by the venous sinusoid which has not been filled in this preparation. (5X) The arterial manifold bifurcates on the central side into the parallel retial vessels which then are seen to anastomose into large distri- bution vessels on the peripheral side. The large globular structures are artifacts re- sulting from rupturing of the blood vessels. (25X) FIGURE 1 109 Momma HHoo m>pmz lllll qz UHOWSCflm mSOC®> lllll m> ®P%OOCMHQEIIIIHWE mohma Hamo popmoomm uuuuu Am Hommo> soapsnflppmflm lllll >0 ezflawnpwmm Psmsmflm lllll mm Hmmmo> Cowpooaaoo lllll >0 cHo> oHeHogpgmo ..... >0 mflooaaflmooofloono ..... oo m>poc capgo lllll zo QOMQEmE m.£ozpm lllll Em muoppm OHEHmzpzmo 11111 do oao%flcme Hmfipmpp¢ lllll E< .QOHmmSOwflv can mQQMMwonoHE pmnppsm CH ooCopowmp pom Cpmppmm pmHSOmm> ogy pcmmmpmmh zaonms op concopcfl ma P59 .maommo> ogp mo onom Adapom opp Psomohmmn Po: mmou Emnmmflu one .A>ov cHo> anamgpzmo 0gp mfi> who msp mo Pso momma can Azav UHochmE Hwfipmppm mzp momoaoCm goaga Am>v Uflomscfim m50cm> Umaama ncflnp mwpma m OPGH cflmpo haamcfiQ mammmo> Pamhmmmm mmoge .mammmm> mews PCoMoM%m ogy op Hmaamhmm 6cm m:oEm Ummompmpcfl mad goaga maommo> PCmMmmmm Hamsm op c306 opmopzMflp mogp mums; mpmp opp op mHMMHHHQmQOHpozo mgp Eopw UOOHQ Chuvop Cogp A>ov mammmo> COHPooHHoo mmmma .Azmv mcmanoE m.£ospm Pm oodMMopcfi Hmvflopogonamsflpop opp Q0 wmmpm Ham wcflmagmsm .Aoov mflpmaaflmm00flpo£o mnp .xpozpo: humaaflmmo Pmm> m comm Cons mmmce .opon may mo ooflm Hmpozmflpom one so A>QV wammmo> Soapsnflgpmflw pmmpma Com: ompo>coo Cogp maommm> omoze .mammmm> opoh Pampome mnp opcfl czov mopdopsmfin Cogp zoflga Azmv waomflcme Hmflpmppm cm opcfl mogocmmn gossa Amov zpoppm OHEngpzmo cm mfl> mMSpmHSomm> Hmoflopogo msp whopco cooam .Hpocopfimm oEHmm go who ogy cflgpflz Choppwm smasomw> Hmcfloposo mnp mo Empmmflw vomflampozowuu.m mmeHm 110 N mmDUHm - M- w..- ,M W- W W I: , , r, I\\.\\ .......§w, .- . \ FIGURE 3.-—Choroid rete mirabile (hematoxylin and eosin, 145X). A. 111 Note the arterial manifold (AM) lying within the venous sinusoid (VS) at the central side of the rete mirabile, the parallel rete ves— sels, and their convergence into large col— lection and distribution (CV and DV) vessels at the peripheral end of the rete. The area enclosed within the heavy line re- presents the region from which the above micrograph was taken. AM—-—-Arteria1 manifold CV——--Collection vessel DV--—-Distribution vessel VS--—-Venous sinusoid 112 ...eve o C t o. .‘o.o_ ... p».o .79....J. . . v. . .3?» ‘ .u a} L A R E H P. R E D- FIGURE 3 113 FIGURE 4.-—Para11e1 rete vessels (Masson's trichrome, 1,48OX). A. In longitudinal section, the afferent and ef— ferent rete vessels are seen to lie parallel to and interspersed among each other. In this section, there is no apparent difference between the afferent and efferent rete vessels. Note that these vessels are of capillary dimensions allowing the passage of the nucle— ated red cells (RBC) in single file only. The area enclosed within the heavy lines represents the region from which the above micrograph was taken. RBC—--—Erythrocyte FIGURE 4 115 FIGURE 5.——Central side of the rete mirabile (hematoxylin and eosin, 145X). A. This micrograph serves to illustrate the discrete arterial manifold (AM) and the less discrete venous sinusoid (VS). The area enclosed within the heavy line repre- sents the region from which the above micro— graph was taken. AM———-Arteria1 manifold VS———-Venous sinusoid 116 $. «'MIRABJJLE " - I FIGURE 5 117 Ham; Hooflowzcflm m:o¢o>----3m> Hfioz oaocflcos Hoflpopoa----zz< .A3m>v Ham; Hmofiomscflm muocm> map wo page mfl cmnp pmeHsp nose mfl Azzdv Ham: uaomflcme Hmflpmppm map Pwnp PGmUH>o ma PH mpom .m opsmflm ca zoom Cowpomm mEMm opp Sony ma agapwouOflE mane .Axodd .CMmom ocm cflahxomeonv oaflpmpfle opmp esp we moan HmMPCoonu.o mmeHm 118 w mmeHm 119 FIGURE 7.——Central side of the rete mirabile (Masson's tri— chrome, 295K). A. Masson's trichrome stains specifically for connective tissue. The arterial manifold (AM) is seen to possess a much thicker wall, rich in collagenous connective tissue, than that of the venous sinusoid (VS). The area enclosed within the heavy line represents the region from which the above micrograph was taken. AM-—-—Arterial manifold VS-——-Venous sinusoid J» ' ' ...,Y . Q - - I 0 ~ ..0‘ \ -RETE‘ MIR ABILE ' \ v“- ~4 O . - I. - - I / ' —a- .’ ' ‘ -.- l ' .. . - , ' , ...,A‘ ._ . - _ . o -- o r -" _ , .. NAM FIGURE 7 121 FIGURE 8.--Centra1 side of the rete mirabile (Masson's tri- chrome, 145X). A. This micrograph shows a tangential section through the central side of the rete and serves to illustrate the manner in which the arterial manifold (AM) bifurcates to feed the afferent rete vessels. Generalized diagram illustrating the region from which the above micrograph was taken. The plane shown is the likely plane of sectioning. AM——--Arterial manifold VS--——Venous sinusoid 122 FIGURE 8 123 FIGURE 9.——Peripheral side of the rete mirabile (hema— toxylin and eosin, 145X). A. The afferent rete vessels converge upon large distribution vessels (DV) which carry the blood peripherally to the choriocapillaris (not shown here) which supplies all portions of the retina. The blood then is collected from the choriocapillaris in large collect- ing vessels (CV) which feed the efferent rete vessels. Note that there is no appar- ent difference between distribution and collection vessels at this level and, thus, the labelling was made arbitrarily for the sake of illustration. There is a rather large number of melanocytes (MEL) associ- ated with the vessels of the central region (Figures 5—8). The area enclosed within the heavy line represents the region from which the above micrograph was taken. CV ————— Collection vessel DV ————— Distribution vessel MEL———-Melanocyte FIGURE 9 125 FIGURE lO.--Vessels of the peripheral region (hematoxylin and eosin, 145X). A. Distribution and collection vessels form an extensive vascular network in the choroid layer. The choriocapillaris (CC) abuts the retina at the site of Bruch's membrane (BM) which is not resolvable at this magnification. Underlying Bruch's membrane is the pigment epithelium and adjacent retinal layers. General diagram illustrating the region from which the above micrograph was taken (area enclosed by the heavy lines). BM ----- Site of Bruch's membrane CC ----- Choriocapillaris CV ----- Collection vessel DV ————— Distribution vessel GCL-—-—Gang1ion cell layer INL--—-Inner nuclear layer IPL———-Inner plexiform layer NFL—--—Nerve fiber layer 0LM---—0uter limiting membrane ONL————Outer nuclear layer OPL---—Outer plexiform layer PE ----- Pigment epithelium RCL-—-—Rod and cone layer CV 80V .. PE ‘. s. v. ‘9. \’ Ar. RCL f c ‘ )\.\ Cu \l \M K \‘ BM v ' \ . \ \ . :._- \ A, \ .\ kt“? \\ .. ‘1 ' I. ..\ . ‘ ONL OLM '79 .4 my . m»4x}..w?<.:4u‘1tw.a~’¥f':.wr _..~ 5 *9...» I N L OPL .-. .. ,r J “'3‘ "Q- '1' ... . _ .5 4 ‘~k \ V‘ - - * ’72.: ‘%;3:-.~'”. ‘; . a .~~-~.mb~'§‘7~%*’7’fn “ " . - '° ' ' "F—‘GC L . x . . - ‘ ‘ I | i. .' ' . ' V... ‘ IPL .."_ "0‘.4 .0 ‘ U NFL FIGURE 10 127 mpzoocmaozlnlqmz .ocsmpwmmmm CH pmcflsma vcm mmooa hhm> mfl mEome wcflwcsoMMSm may Pmnp who: .mpoz .Pcmucsnm mam Agmzv mmphoocmamz .cOflpBOH%HQmmE mo Hm>ma mHQP Pm hmgpocm one 809% mapmgmflswcflpmflc #0: ohm mammmm> COHPSQflMPwHU can Coapcmaaoo one .3813 .Qflmom Ucw cflaszPMEmzv maommm> soapznflppmflu cam COflPomHHoo awhmnmflpomnu.aa mmeHm 128 HH mmDon 129 pmmmm> pmppmp pampmppmnunn>mm pmmmm> pappmp pCmpmppm< .mopommm pmgpo Cp Comm moEpmeom Cpmppmm mcpxoma pmcowmxmg mgp op Ummommo mp mppga Gp UmeppSo Cpmppmm mapxcmg opnzo mpge .mmghp pmmmm> mgp Cmmzpmn mpgchppmpmp macaopsmco m mp mpmgp pmgp £05m .Mmpm> oop> can mpmmmo> pCmpmppm psop an UmUCSOppSm mp pmmmm> mpmp pampmppm gomo pmgp mMMmmmm pp .cOppm>pmmpo pmmopo Comb .A>mmv mpmmmm> pmppmp pcmpmppm mp8 mpmmmm> pmmump mnp pmzp Ucm A>m¢v mpommm> pmppmp psmpmppm mum mpmmmm> pmppmsm map pmnp UmSSmmm mp pp .Apxmp mmmv mCOpmepppsm m30p>mpm po mpmmp map so .pmmeMpU mp E: m.m pSOQm mapmn pmnpo opp vsm pmmempU Qp E: m.: zpmmepxomem prmQ mac .mpnmzmpzmnppmpc mpm mommp pmmmo> ponpmpU hpm> 03p .COppcmmnmmopo CH .Axmmm .mSpQ mcmphnpmev COppommlmmopo mppnmppE mpmp pmvpopoL0|l.Np mmDon 131 mmpopmm> cppophoochOQOpz nnnnnn maopnd Cmesp uuuuuu H mpmoopnphpm:nnlomm pmmmm> mpmh pzmpmppMIII|>mm mpoSom> oppzoommnm nnnnn >m ppmo pappmnpoucm IIIIII m mpzoppmm uuuuuu m mcmppama pCmEmmmm lllll 2m mdmpozz IIIIII z pmmmm> mpmp pampmppm¢ .mcmppsme pCmEmmmn pocppmpw m po Momp m Usm .mppppCm>wm :mppoam .mmpom no mmppmmcmp mflpxomp appmzpoucm Umpdfifimppm hpoEmppxm .Wmaoppmv mmpopmm> cppopzoocpmopOpE .cmESp pmmmo> map opsp mzpnommp mm MUQmmam mxppuEpm macpmszc opp .A>mv mmpo:cm> cppmoommgm mmpmp zpm> map mpoz .pmspocnm on op Umpmvpmcoo mum who: Comm mpponpOUCm pmmmm> mpmp mge .Axmmm.wpv Coppommnmmopo Cp mpmp Umxpp COpmmeEp Upmwcmpmnl.mp mmbopm 132 mp mmeHm 133 FIGURE 14.—-Standard immersion-fixed rete: Area with myeloid configurations in the lumen. A. Abnormal appearance of rete. Note the densely granular background matrix in the lumen and the multilaminar myeloid configuration (MC) within the vessel lumen. (16,560X). Higher magnification of a similar myeloid- type configuration within the lumen of a rete vessel. The configuration seen here appears to suggest some ongoing extrusion activity. (41,405X). L ------ Lumen _ MC ————— Myeloid configuration PV ----- Phagocytic vacuole RBC~——-Erythrocyte WBC—---Leucocyte 4 l E R U G I F 135 opopmo> cppopmoocpmlnzopp¢ mpoSom> opphoowmnm lllll >m opozom> pmmoESH mopm nnnnn >p pmmmo> opop pCopoppmnnnu>mm pommo> mpop pcopoppmnunl>m¢ .mCOppmpmmopm ooppopo Cp Amwmpv ppmm @Qm pmxpmpm .Caopm mp mnmmpmopope map Sp omnmpppzm oCo ozp op poppEpm hpo> mp mono mpsa .cmszp pommm> ogp Qp A>pv mopodom> ooppmanzpnp ompmp mo mommmopm ogp mp nmmpwopops Coppoopm mpgp Sp mpzpmop o>pponpmpo ong .Axoom.opv cos5p oep :pgppz moposom> ooppmaucpnp QppS mopd “opop ooxppICOpmpoEEp omeQMpmuu.mp mmbwpm 136 ‘ 3 mp EDme .....kv \.. b 137 ppoo SSmgSom llllll m m5opoSS ppoo SszSom nnnnnn z mponpp o>pmS oopmSppo>2 lllll ME ASOppoomnmmoSo Sp mponpp SowMHHOOV xpppma mSoSmmmppoo nnnnnn o Soxm o>SmS oopMSppthnSoz nnnnnn S .zpomSom So Sopoe oSm opoS Smom mpSoEopo pmSSoS oSp SoSpoSa SaoSM .So>o30S .poS mp pH .oo>Sompo Soon o>mS mSoppoSsp pmSSoSozE oSm .ppmz opoppSmE pmpSopSm oSp Sp pSomoSm mpm mmpomSE SpooEm .pSomoSS ompm opm szv mSoppp oopMSppmza mSmS Soom on hoe m< .Amv mppoo SSmaSom ompMpoommm SpoSp Sppz A¢v mSoxm m>SoS oopmSppoSEISoS mSopoSSS mmmm SopS3 SwSoSSp mmmpm oSm ppm; pommo> opoppSmE popSopSm mSp po on xpppme mSOSmmmppoo oSp SpSppB ooooopem .Axomm .omv ppm; Spop upSmE pmppoppm oSp SpSpp3 mpSoEopo pmSSmS ompmSppo%EISoS oSm ompmSppomznn.mp mmbwpm mmeH m 139 mpommo> cppophooSpmunnaoSSS opomse SpooEm nnnnnn Em ESpSoppoS opEmmpSooSo SmSom nnnnn mmm mSopoSz nnnnnnn z SOppoSsm popsoSohz lllll hzz Snooppopz ....... 2 Soesp 1111111 p ppoo pappoSpooSm uuuuuuu m SmwMHHOU uuuuuuu o Sox< nnnnnnn S .mozmmpp mmoSp po ooSommSQ oSp hp oopmomem mp oprMSpE opoS poopoSoSo oSp szoSSp zopp mo Soppmpswop oSp Sp opoppSmE pmpSopSm oSp Sop opop opnpmmom S .op opsmpm Sp Soom Soon mm: ppm; opoppSmE pmpSopSm mSp SpSpp3 mSoSSoS po ooSomoSm oSB .Abzzv pompSoo pMSSoS po mpSpoQ on op Swomgm SopSa mmopm pmpm>om oSm mSoSB .ppoo mpomse SpooSm oSp po mSpmSmE oSp mSopm Soom mp Amaopva mpp>ppom cppophooSpm pSmoSsn< .Azmv opomse SpooEm po S®hdp oSo pmmop pm mp ppm; opoppSmE pMpSopSm oSp SpSpp3 Amv Soopon poppoSpooSo oSp MSpmpSooSD .AxOmp.mNV ppm; Spoppcme papSopSm oSp mo Smopop poppoSpooSo oSp wszpSooSs Sohmp opomSE SpooEmn|.mp mmDUHm «’4- ."'r" 8" l\ H E D (D H F1: 141 ppoo opomSE SpooEm IIIII Em mSopoSS poppoSpooSm uuuuuu z oppooppops ...... 2 SmESp uuuuuu p mSpMSmmmm pmpow nnnnnn w ppoo pmppoSpooSm IIIIII m Npppme mSoSommppoo nnnnnn o .opoS oopSomoSQ ooSoop>o pm6pwopoSmSoE oSp hp oopmomem zpoSoE mp p59 .ooSoop>m mepSmeo oS Som: woman mp opmpzpmom mpSe .pMow mpSp Spps SOppoSSmSoo Sp opmpomo ompm has hpp>ppom oppopzooSpS ppoo poppoSpooSo oSp SoSp .ospp mp mpSp pH .opomse popopoxm Sp ESpSoppoS opEmmpmooSmm hp oozmpa opop oSp mSpoMpmoS .SOpmSpro So oSMpSS Sop Expopmo o>ppom po o>ppdopoSp mp opomSE SpooEm SpSppz mpp>ppom oppophooSpS pmSp mpmpxo zppppnpmmom oSe .mopppopmpos oSm mpSopSpSS SmpzoopoEoSomE po pSommSmSp poppoSpooSoumSmSp Sp o>ppom mp mpmopzooSpQ oSp SoSpoSz mppmSOppmoSd op hos pp pmSp om .ppoo pmppoSpooSo oSp po Smopon pMSoESp mSp mSopm mSpSSSooo mp hpp>ppom oppophooSpm oppppp zpo> pMSp opoz .ppmo pmppoSpooSo oSp Sp zpp>ppom cppophooSpg o>pmSopxo po mopm Sm op pSoomnom Soopop ppoo opomse SpooEm oSp po hpp>ppom oppophooSpm pSmoSSQm oSp mp SmmSwoSopE Soppoopo mpSp Sp pmoSopSp po pSpom Sopme oSe .Axomm.wmv opomSE SpooEm oSm SSppoSpopSo oSp SpSppz hpp>ppom oppopzooSpS Sooapop SOppMpoommm Sm mszmpmmpo ppm; opoppSmE pmpSopS opoS pSoSoppm nnnnn >mm oSmSQEoE pSoEommm nnnnnn Em pommo> opop pSoSopp< lllll >m¢ .mpommo> opoS pSoSoppo Sppg Soom poS mp mpSB .moSmSpEoE pSoEome opopompo an oooSsoSSSm oSm mopzoppom ompMpoommm SpmSp oSm mpommo> opmp pSoSoppm ppm pdSp mp opoS op oSSpMmp pSMpSomSp pmoe oSe .oSoS pSomoSQ mmSSpmop pmSoSmw oSp oSSoon poS mooo .So>o30S .SoppmSo oSB .SMpopSSoop pmoSw m mp pmpSoomoSopE SoSpoopo ooom mpo>o oSpSoQ poSpimsop>no oEooop mMS pp moSpm mpSp SwsopSe .SoppMSo Spp3 opopgop mp Sampmopopfi mpSp .SmpopSSoop ooom m po Momp oSp op oSm ooSopSonm po Momp oSp op oso .hpopmSszopSD .opppmSpE opoS poopoSoSo oSp mo SOppoomnmmoSo m mp wopm mpSe .oSoS Soom mm moSSpmmp zpmpppmmo poopmhp oSoE SoSS Sp ooppSmop SOppmxpp op SOpSm SoppmNpSpSmmoS pmSp oSzop mm; pp .Ampump mopzmpmv mSOppmSmgoSm pmpppSp oSp Sp ooSSSooo om: mSpppopc pMSp mSpSpSSmpoo Son: 143 .Axoo:.wv SOppoomnmmoSo Sp opppmSpS opoS pmopopoSo ooxpanOpmSmEEp .ooNpSppmmmmnu.mp mmbwpm mp MMDUH .m . 0 J 51-0.} .._\N. r . fiézdflwfio Wkuwbfi..- ...a . o 3:31;}. .5». ’u.l§.~‘ v.§'n.- _- . Q 9....v . v. 0“.\.. I 1’ AVI! . .tll..‘..!.l.. ... 33$ 0.8;3’. .‘ . 31315.0. ‘38... . . a. \ y 1 V ., .. ._ g. .. {3’1 w‘ifi.h.ov.. ' A V- . do A... . n . .5 . \* U&-w.1. ...: .f. .3”..er ..4 .I «3.... .08 I!!! ......r Golf (53%;...915 .. p... 9.1.“. ...... -w... .. .14.. .58.;83 -.....ru‘ . . 'IVNH‘fn 145 oppooogppom ..... 0mm SopSoSOSooppE uuuuuuu E SOppost ppoo IIIIIII b Hmwm®> HmflPmH Pflmhmpwm IIIII >mm oSMSQEoE pSoEommm uuuuuu Em .pm mpsmpm Sp oopppSmms mp xop oSp SpSppz wopm oSe .oSmSQEoE pSoSmme opopQEoo m an omoSsoSSSm mp pmmmo> pSoSoppm oSB .moSSpMop pMSSpoSSpmMSppS mpp Sp ooSmSopppo pSmSmmmm oS mBOSm pommo> poppop pSoSoppo pSmommom oSp po ppm; poppoSpooSo oSB .opSmSSoompo ompm opm opSoSoSooppz .Emmpmopho oSp pSoSmSoSSp oompommpo Swommm mmsomonpp mopm .mpSoSOQEoo opEmmpmopho po pom opopSEoo m mSpSpmpSoo oSmmpp o>ppom hpSpMp m on op mpmmmmm ppoo pmppoSpooSo pommo> pSoSmppm oSp pmSp Soom mp pp .SOpmepppSmmE SoSmpS m p< .AXmmm.omv Soppommlmmopo Sp pommo> .Tmppo.H pSmpmpp mpmp pSoSmppm lllll >mm ESppoSpooSm nnnnnnn m oSmSQEoE pSoEommm nnnnnn Em SoSSp pommo> opop pSopoppS 11111 >m< .om oosmpm So xop ogp SpSppa oomopoSm moSm oSp mp mopm mpSe .mppoSpooSo opmp pSmSoppo oSm pSoSmppm oSp Sooapon Soom mp moSopopppo oS .SOppmopppSmmE poSWpS S058 p4 .AxOmn.mppv oSSpoSSpmMSppS pdppoSpooSo opmS pSmpoppm mSmSo> pSoSmppSlu.pm mmpwpm H N E D U H Ft: 149 MSoSpm nnnnnnnn m mophooSmpoznnumzoSp¢ .hpmpppSm mpommo> mSp oSsoSSSm op Soom So>oS oSm poSp pmSp opoz .SOppoom mpSp Sp mzop>So ompm ohm mopzooSmpoz .mpommo> oSp mSpoS50SSSm mEoSpm omoop oSp .Spmmm opoz .pp oSm op mopsmpm po mmSmmpp ooooopso Spppmpmm Sp Mono SmppSpm m op oopmmSpSoo on hoe oSm SOppmSpsmxo opSoomoSopE Soppoopo Sop mSpSOppoom SpSpmSppS Sop Sp mp ozmmpp oSp poSp SOpppoSoo oSp mp mpSB .ozpp oSopmSpoE Sppa ooSpMpm oSm .E: m pSOpm pm ooSOppoom .mpm zomm Sp ooooopso Soon mdS oSoS oSmmpp oSe .Axmmm .oSpQ oSopzSpoEv Sozmp pmopoSoSo pdSoSmpSom oSp po mpommo>un.mm mmpopm 150 mm mmDGHm 151 pmMpnopppm nnnnnnn m oopop poppoopooom ....... m Nppme mSoSowmppoo 1111111 o .mmmm mpommo> Soppsppppmpo oSm mSppomppoo oSp SopSz SwSoSSp msoppm oSp po xpsp oSp mSpmpSSEoo moSmSQEoE SmSpSMppppSE .mmoop po szmp 8 mp mophooSMpoE oSp wszpSo>o .mopSSmSm SpSmpoS poopSoSmm .xompp mmSoo hSmS WSp>mS mm ooprmoooS mppmmo oSm oSm pSmoSSQm opm mophooSMHoE .mmmm op Soom mSm Amv mopzoopppp SopS3 SwzopSp on xpppms mSOSommppoo .xopSp o hp Spdppm>o Amv Swamp poppoSpooSm oprpm m mommom op Soom oSm Sopwop mpSp Sp mpmmmo> oSe .Axmpp.3pv oSSPosopmoopps pommo> poopooooo pooogopoom--.mm mmpmpm mmpwH m . 8.(’n.rl ...uv\f~.mn...1v . .....w_.......... ......u. ... 153 .9:0Spm uuuuuuu a vb ESpSoppoS cpEmmpmooSo Sszoxnnnnnxxx mzopozc p350Spauulnuun ' oppoocopoz ..... Sm: oppoopnpa ....... ; soppoepoozs-------: xpSpsE msoSoxoppoo uuuuuuu o .ompapmcpo pmowmsm opzoz pSomoSQ sz mzopozS ompxp 3:; p59 .oSSpmS Sp SmpSppoo Smomgm pmSpp pd poS moon opzooSBpoE oSp op pSoomnom Sonp oopmpppm oSB .oSoS deoSSSd mp SopSB Soxsppoo ozp po SOppoSSOSQ oSp Sop opppmSommoS zpowpdp oSm mopzoOSSpm .Amv opxc IOSnpp mSp po Emmpmopho oSp Sp ESpSoppoS cpsmdpAooSo szoS po xS03poS o>pmSopxo oSp mpoz .mm oSSwpm mo ooSm Sn mo pSoonSdpSo S3 mp mpSe .Axmmo.mmv pad; Houao> Hooogapooo--.:m passpa F IGURE 24 155 oppoopeppom ..... 0mm opzoopSoppom IIIIIII m ssppogppoo pcoempm ...... mm ESppoSpooSm nnnnnnn m mpSmpppmmOOpSoSo llllll oo oSMSQEoE m.SoSSm ssssss Em ppmo pmppoSpooSo mpSmpppSmoopSoSo oSp po SOpmSopxo oxppuomSmpmnnnaoSS< .mSoppSpo>Sp mxomp oSm SpooEm mp oSmanoE m.SoSSm mSpppSQm SSppoSppmo pSoEmpm oSp po Smopon pmmmp oSp .mmeEoE Sp ommo oSp oxppr .ppmz pommo> oSp opomeoo op MSppoS oSp opmzop mSOpmSopxo oxppsomSMpp moSom oSm pommo> oSp po oopm pmpopom oSp So mopp SOppmooxo pzoSpp3 .mpSoSpppmSoo cpEmMonpzo oSp po xpsn oSm mSmpoSS oSp mSpSpmpSoo ppoo oSp po :ooSoSpEoSQ poSpSoo: oSp poSp Sp oomppmpom on op Soom mp ESppoSpooSo mpSMpprmoOpSoSo oSe .Azmv oSMSpEoE m.SoSSm hp Ammv ESppoSppmo pSoeme oSp Eopp ompmpmgom mp pH .MSppoS oSp po oomppdm oSppSo oSp op pSoommom mSphp mpommo> oompmuzpmpppgmo mo nEooMpmo pmm> m mp mppmpppmmoOpSoSo mSe .AN03m.mpv ESppoSppgo pSosmpm oSm .oSdSQEoE m.Sospm .mppmpppmmoopSoSoln.mm mmbuHm F IGURE 25 157 FIGURE 26.-—Fenestrated choriocapillaris endothelium. A feature which characterizes the chorio- capillaris endothelia of mammals is that diaphragmed fenestrae are very abundant in that portion of the endothelial wall abutting Bruch's membrane. In Salmo gairdneri, it is seen that the endothelia are extremely attenuated in this region and occasional fenestrae are seen but their abundance nowhere approaches that seen in the mammalian choriocapillaris (A—-4l,405X; B——265,ooox). Arrows——Diaphragmed fenestrae BM ------ Bruch's membrane CC —————— Choriocapillaris 158 I '0 1" o FIGURE 26 159 FIGURE 27.--Bruch's membrane (41,405X). Bruch's membrane is seen here to be composed of a network of very fine fibrils embedded in a uniform electron lucent matrix. It is bordered on one side by the choriocapillaris endothelial cell membrane and on the other by the pigment epithelium basement membrane. This arrangement is relatively structureless compared to the mammalian morphology (A——41,405X; B--41,405X). B —————— Basement membrane CC ————— Choriocapillaris E ------ Endothelium M —————— Mitochondrion PE ————— Pigment epithelium FIGURE 27 161 FIGURE 28.--Myeloid bodies within the pigment epithelium (16,560X). As a point of interest, the cellular inclusions of pigment epithelium are unusual in appearance in relation to other tissues. The cell has four prominent inclusions: melanin granules (MG), abundant mitochondria(M), nucleus (N), and myeloid bodies (MB). Note that the pigment granules are rod-shaped as opposed to the prominent spherical shape seen in the melano— cytes of the choroid layer. The myeloid bodies have been described in the literature, but until recently, their origin and function have been in debate. It has now been conclusively established (Anderson and Fisher, 1975) that they arise from phagocytosed sections of shed rod and cone outer segments. There is a definite structural similarity between these myeloid bodies and the photoreceptor outer segments (Figure 29). (A—-16,560X; B--16,560X). M —————— Mitochondrion MB ----- Myeloid body MG ----- Melanin granules N —————— Nucleus 162 FIGURE 28 163 .mm oSSwpm mo mmpoop opopoms oSp Sp omoSp oSm oSoS moSmSQEoE oSp Soogpop mppSMppEpm oSp opoz .pSoemom Sopso ppoo Sopmooop po SOppoom mpSp oopm mSopm mSpSSSS oSoS Soom mp owmoSoSSm Sm Sosm .mmooopm pmzmp> oSp Sp SOppSpommS ommEp omdoSoSp op mm>pom mpSe .SoppmSpESppp pSmoSspm po moEpp mSpSSo mpSoempm MSpSooSom mm pom mSopmSopxo mmoSB .mppoo Sopmooop oSp Soozpon SS mSppoS oSp po oomppzm oSp op pdESoS mmpSSmSw SdepoE mSpSpMpSoo mSopmSopXo opoIESm moSopoSpm SSppoSppmo pSoepr oSp .mpdpm oopmmom:pSpr oSp SH .ooopvopm oommm oSp SpSpp3 opppmmom mm pSoEmpS Sosa mm opmooSooom op Soppdpmmow Sm mopmppmSppp oSSpoSSpm mpSp oSm .oS502 mSmSQEoE on pmSE mpSoempm SopmoompopoSm oSB .mompo ooxoMpm po ooSdSmommm oSp mm>pm SopSa oSMSQEoE coopop mpnmpS m mp mSSpMop Spms oSB .oSSpoSSpmmSppz oppmpSopomSmSo hpm> m mmS pSoEmom Sopdo ppoo Sopmooop oSe .Axmmm.omv pzosmoo oopSo Sopoooooopogmu-.mm mmDUHS FIGURE 29 165 FIGURE 30.—-Choroidal rete mirabile carbonic anhydrase stain. Here, the choroid rete is seen in cross- section to display an intense staining reaction (A) which is totally eliminated by 10‘5M acetazolamide inhibition (B). No reaction was ever seen in any of the inhibited tissues, whereas any unstained areas of tissue in the uninhibited treatments were believed to have dipped beneath the surface of the incubation medium which was known to eliminate the stain. Thus, it had been believed that the entire choroid rete possessed a high degree of carbonic anhydrase activity. From these light micro- graphs, however, no localization of the stain to either afferent or efferent vessel type could be made. (A-—l45X; B-—295X). FIGURE 30 167 FIGURE 3l.—-Choroid rete mirabile carbonic anhydrase stain (295x) . At higher magnification, the choroid rete is again seen to be intensely active (A). Sodium acetazolamide completely inhibits the staining reaction (B). No conclusions may be reached with regard to the localization of activity to any one particular site at this degree of resolution. (A—-295X; B——295X). . c 1 ..r. .. .....aw A \ a . odd -5 . . ......a ,. . . 1 a .. i x) 3...... r £537”? ’r. x .. FIGURE 31 169 FIGURE 32.—-Periphera1 region of the choroidal rete stained for carbonic anhydrase (145X). The peripheral end of the rete is notable by the presence of black melanin pigments within the stroma surrounding the collection and distribution vessels. This feature is seen in the acetazolamide inhibited preparation (B). No activity is seen here, even in the erythro— cytes which fill the lumen of one of the vessels (arrow). Inhibition is complete. In the uninhibited tissue (A), carbonic anhydrase stain is seen to be intense in the retial vessels and may extend into some of the distribution and collection vessels (arrows). Note that in the inhibited tissues, the pigment is never seen to surround the vessels entirely. (A-—l45X; B——l45X). m --—‘—-————-_n‘_——.-___L_——_n—_‘.——u ______-—_—._-_-_ L—__—__—.__.______—_._ _ FIGURE 32 171 FIGURE 33.--Carbonic anhydrase activity within peripheral collection and distribution vessels (295K). As was seen in the previous micrograph (Figure 32), carbonic anhydrase activity is seen to extend at least partly into the peripheral vessels. Micrograph A displays uninhibited tissue, and the tissue in B is acetazolamide treated. The region displayed in both micrographs is in very close proximity to the rete. The section of tissue in A is relatively thick, and the vessel in the center is seen to be lined with activity. This vessel measures approximately 70 um in diameter and is definitely beyond retial vessel dimensions. (A--295X; B——295X). — _._— FIGURE 33 173 UpomdSpm mSoSm> IIIIII m> UpoppCME Hmppmpp¢ IIIIII Ed .oo>pompo mmSmmpp mSp pSoSwSopSp zppSmpmpmSoo Smom mm; SOppmNpSmpom mpSp poSp opoS op pSMpSoQEp mp pp .hppdSpm .Amzoppm omomoS opnzoov Spdpm omzpppo mpSp So>m zoSm poS opo oSm Expooe oSp SooSS ooopop .Sm>ozoS .UMS ozmmpp mpSp po SOppSoQ < .oopm pmSpSoo oSp po mpommm> oSp Sp Spdpm omSpppo pSwpp m an Soom mp oSm opdp cowhHMpmoSS opQMpoopoo m pm mSSooo SOppoMmS oSp poSp pomp oSp hp oooSmop>m mp mpSe .SOppomoS mSp mgopm SopSS ESpooE SOpmeSoSp oSp po oompSSm oSp Bopop ommmpo wSp>mS ozmmpp pappop oSp po oopm pmSpSoo oSp op oso poS mp SOppmNpSmpom pSoSmSQm mpSe .oopm pMSoSmppmS oSp pm mpommo> SOppomppoo oSm SoppSQpppmpo oSp MSpSomoS pppSs moSSppSoo oSm mopppmSopSp mpp>ppom ommpozSSm opSoQSmo .opop oSp mSopm moSdpmpo mSp po oSpSp oSo mpomepxopmmm p¢ .AopomSSpm mSoSo> oSm opoppSmE pmppoppm oSp po SOpmmp oSp ..o.pv oopm pmSpSoo oSp pm hpp>ppom oS mp oSoSp poSp mSmong pp SOppoom mpSp SH .opoS oSp po SpmSmp mSp wSopm hpp>ppom omdpthSm cpSoQSMo po SoppmNpSopom oppSppoo m mp oSoSp poSp Smom mp pp SOppoom pdppSoWSmp SH .Axmmv mpommo> opop pmopopoSo oSp SpSpp3 zpp>ppom omMSomSSm opSonpmo po Soppmmppmpomln.dm mmbwpm _I <1 0: u I E a: m a 171. FIGURE 34 .moSSpmop omoSp opmppmzppp op mo>Som SopSz oSmmpp SMpSomm> pmppop po SopWoS m mpSomoSQ oSoS Sampmopopa oSB .AMopm ommopoSov opmppmpoopm xoMpp mmSoouSoSpoopo oSpp m on opSoSm ppmm ppmpoo mSe .mngSwOSopE oSp mo hSmE Sp pSoop>o on pppz SopSz SpMpm SmopoSS pmSpompppSm Sm oppoh op Updm mp Spmpm ommSthSm cpSopSmo oSp pmSp mp opoS op pSpom SoSpoS¢ .oopSSooo ooSo moSmSpEoE oSp pmSp mSOpppmoQ mSp Sp Soom on pppa moommm So moSpp Swopo szo USm oopomppxo Soon o>mS ppp3 mmSmSQEoE ppm pmSp mSmoE mpSB .mmooopm SOppmpthoo mSp mSppzo oopomppxo on pppa mopmpp .ppSmop m m¢ .oopxoppop ESpEmo hp ohm zmSp p59 .oozSoopmSmpSpw pp ooxpp hpmeSoS poS mum mopmpp .A amp .Somomv m>ppmxpp mpSp Sp oppzpom on op USSop Soon mmS opmppmpomp ppmpoo oSp pmSp pomp mSp op mso opsoooopg wSpSpmpm mSp Soppm SOppmxpp mopxoppop Expsmo om: op opppmmom poS mp pH .mSOppmSmmoSQ ommpomSSo opSopSmo ogp po poap Eoop pcooopppo oppso mp Aom-mm .Hm-mp moosmpmv oopxoopop ESpEmo Sp ooxpppmom oSm oozzmopmpmpSpw Sp ooxpp moSmmpp po mpdpoSSpm noppps oSp mo ooSmSmoQSw mSp .mozprSomp hpopmSmSoSg oSmmpp oSp op mam 175 .Axmmm.mmv omoomopopE SoSpoopo oSp SmoSS mmsmmpp ooSpmpm ommpomSSm cpSopSmo po moSmSmomgm pmSmSoolu.mm mmbupm 176 mm mmpopm \v . 177 FIGURE 36.—-Carbonic anhydrase activity within the choroidal rete endothelial cells (66,250X). A. Here, the uninhibited carbonic anhydrase reaction has produced dense staining throughout most of the endothelial cell cytoplasm. It is believed that three sep- arate endothelial cells are seen here sep— arated by a distinct interstitial space (double-headed arrows). Numerous tubules, probably representing endoplasmic reticulum (ER) are seen running through the cytoplasm. It appears in places that these tubules have precipitate within them, such that a single tubule appears as two white lines separated by a dense black line. (66,25OX). The tissue has been incubated in medium containing 10‘5M sodium acetazolamide in order to inhibit the deposition of carbonic anhydrase-catalyzed reaction product. The reaction proceedes at a significant uncatalyzed rate, however, and yields a less dense, ubiquitous background precipitate. The small needle-like crystalline structures seen in this micrograph are dirt of unknown origin. (66,25OX). Arrows—~Interstitial space ER —————— Endoplasmic reticulum 178 FIGURE 36 179 mSopoSS ophoopSphmeII:ZUmm oppoonnpppm ..... 0mm msopoSS ppoo pappoSpooSm nnnnnn zm .Soom ompm mp Azmv msopoSS ppmo pmppoSpooSo Sm .ppmmpp Emmpmopzo ppoo oSp SpSppz p59 .opSpoSppm SmpSHHooQSm opponpS amooop sz SpSpp3 oopmpoommm poS moSmn Sp oomppmoop op op meoom EmMonpho pappoSpooSo oSp SpSpp3 ppp>ppom oSe .ompommxo on opSoa mm omSopSp hpoEmpro on op Somm mp ommpchSSm opSonme ophoopSphpo oSp po hpp>ppow oSp mSom .opoS opOSOSo mSp po ppoo pmppoSpooSo Sm op pSoommom wszp Soom mp Azommv mzopoSS mpp Spp3 Aommv ppoo Uoopp cop m .SOpmepppSme SmpS m p< .Axoom.mopv App>ppom ommpozSSm opSoQSmo ophoopSphpo Sppa ompmmaoo mm zpp>ppom ommpwthm cpSopSMo ppoo pmppoSpooSo mpop opoSoSonu.mm mmbopm 180 FIGURE 37 181 .Emwpmopzo oSp pSOSmSoSSp ompSQpSpmpo hpmSooSoonoS poS mpm pm» .pSoSoSsoo Swpzppoopzm SmpSOppSmm mSm Sppz oopMpoommm hppSoSmmSm poS oSm SopSS Smdpaopmo poppmSpooSo oSp SpSpp3 mSpSpmpm omSmo po mSOpwoS oSm oSoSp .mm mpsmpm Sp oSm mm oSSwpm po mopm oomopoSm mSp Sp oopoS mm; m< .Axoom.m0pv ppoo poppoapooSo mpoS opoSoSo mSp SpSpp3 mpp>ppom mmdpthSm opSopSmo po SSoppmm mSpoSmmun.wm mmpupm ’o'l ' 1 mm mmDUHm .. . . ... ... ...» . ... ..u . r . . . ¥-u . 1W...» .. 183 mSoposz lllllll z copoocogooppz ....... z CmESQ IIIIIII .H SappppmSopSH ...... mp .pommm> pSmSoppm Sm po pmSp mp ESppmSpowSo ooSpmpm hpomSoo oSoE mSp pmSp o>oppon op Sommop mp oSmSp .So>ozoS .pxmp mSp Sp oommSompo m¢ .mpmmp mpSp So mpp>ppom h>moS oSp mSpSpmpSoo pmmmo> po mmmpo cpppoomm mSp mSpoSmmop ooSomoS on oSoS mma SOpmSpoSoo oS .mSSE .oSmS mpommo> oSp po hSm Sp oommm poppppmpmpSp oSp Sp moSmSnEoE ppoo mSp aoppop op Soom mp oSmSQEoE pSmEommp oS .zpopMSdpSopSD .oSmmpp oSp mo SOpppoSoo oSp op oso opnmSpESmpoo poS who; SopS3 mSopoempo poSoESp Eopp mopmm .mpommo> pSoSoppm po opzpmop cppmpSopoMSMSo pmoS oprpm oSp mp oSMSQEoE pSoEome m po ooSmmmSm mSe .pSoSopppo on hme mppoSpooSo pSmSoppo mSmSo> pSoSoppm po mopop pmSOppoSSp oSp pmSp opmopoSp op oSop opsoz mpSe .pSopmSmm mp zpp>ppom po oopmmp Sp ooSopopppv mSB .oommm poppppmpopSp mSooSomoEoS .oommSmnM .xopSp m an oopmpmmom mpommo> ooSSp po mppmz oSp po pSpoQ SoppoSSn oSp mpOpmoo Sampmopope mpSB .oo>pmmpo wSpmp mm; mpommo> mo mmmpo SupSa mppoo oSp po mopzpmop oSp Sopp pSoop>o poS mm; pp .ooooSp oSm .mpp>ppom ommSomSSm cpSonSMo pommo> pSoSoppo oSm pSoSoppm Soozpop SOppoSppmpo mSm ompmo>oS poS poS mSmmSmoSopE mSop>oSm .Axoom.mopv mpommo> pmppop pSoSmppo mSmSo> pSoSoppm Sp mpp>ppom ommSozSSm opSonSmo poppSoSopMpo ¢nn.mm mmbupm 184 F IGURE 39 185 xSoapoS ESpSoppoS opEmmpmooSm SpooEm nnnnnnn a ESppoSpooSo oxppnppoo Smpppm uuuuuuu m mpoSom> cppopooz nnnnnn >2 msoposz IIIIIII z xSoapmS pmpSoSOSooppz nnnnnnn S SoESp >SMpppmmo IIIIIII p CpWMmE Hmopm¢ IIIIIII d .AHRAH .EEooE oSm .mxSmpSpmm .pSoppomv SmSomoEMS m5Sppm>pmm .pSoSp oEMp oSp Sp mSoSmSpoozomQ oopSmoSpSong zppMopwopoSpmm Sp oopoS Soon mmS SopppoSoo SmppEpm d .A>zv SopmeSop opoSom> SmHSppoomSpSp po moSm oSoprm hppSmSwSQm Sm mmmpmmpo mppmo omzpuSoSMSpooSomm mSp po oSo .pmoSopSp po pSpom pmSpp m m¢ .oopmppmoSoanoS mSm oppoSpooSo oSp po mSoppSom oopmSSoppm mSp pmSp opoz .mpSoSmppp pppm oSp po omppoemp szoSooom oSp Sp Somm mppoo Smpppg oSp mpnsomop Sosa zpo> mppm3 pommo> hpmpppmmo oSp wSpESop mppoo pmppoSpooSo oSB .momhp ppmo pSoSopppo mppoSppmpo on op oo>oppop 30S mum zoSp .ppoo ooppopSo m op Smppspm mpmmmmm ppoo omhp nSoSmSQooSomm oSp SMSOSpp< .ooSmopppSwpm SzonSz po mmmpozSSm opSopSmo po pSSoEm ompmp hpo> m mommom op Szon mp SoSmSpooSomm oSB .Spwpme pmcpmm mSp op pSoompom Soom poS mp pp .ppoo SoSmppooSmmm oSp USSOSSSm poS mooo zppSmSmmSm oSmSQEoE pSoSome mSp poSp opoz .ESppoSpooSo ppmpppmmo oSp op pSoomnom oSmSpEoE pSoEomwS oSp So mSphpQEm oSm mppoSOSooppE oSp po mppSpop> oSp Sp mSppmSpmppo SSpSoppoS opEmmpmooSo SpooEm po xpozpoS o>pmSopxm Sm Sppz Soppmpoommm Sp mpSoSoSooppE po hoppm ooxomg hppSwpp m MSp>MS >9 oomppopomSmSo oSm omoSB .mppoo ommpzSoSmppooSomm So cpppSmoopom poSppmpp zpo> mo SoQESS owpmp m Sppz oopMpoommm opm mpommo> popSoSMSpoozomS oSB .Axmam.av ppoo opppgooopoo So =oopp-nooooooosoooz o:p--.o: mmDUHm 186 ,_ :23"- . .1" o: mmpmpm 187 .xpoapoS SmpSQSp xopmsoo m opSp Szoo ommp oSm opmSpmppo op Soom oSm SSpSoppoS opEmmpmooSo SpooEm .mppoSoSooppE oSp Sooapom .Soopop pmmmp oSp op SMpSopoSoSSmm SoSSmE poppmpmm hpSmSop m Sp omxomm mpomSmo on op Soom oSm ppoo omzanoSmSQQUSmmm oSp po mppoSoSooppE oSB .Axmmw.mmv ppoo omhanoSmSQOUSomm oSp mo MpozpmS ESpSoppoS cpEmmpSooSo SpooEmanppoSoSooppE oSB:n.p: mmbwpm p: mmDUHm 189 1 lllnllilll. I111]! xpozpoS Sopsnse ttttttt a SOpSoSoSooppz sssssss E SoESp hSmpppmmo nnnnnnn p oomSm poppppmpopSH nnnnnn mp SappoSpooSm nnnnnnn m oSmSQEoE pSoEommm uuuuuu 2m .moSmSQEoE pSoEmmmp sz mxomp ppoo poppoSpooSo pSoompom oSB .mSmSQSoE pSoEommp poSppmpo m hp oooSSop mp ppoo ommpuSoSmSpooSomm mSp mo Soopon mpSB .oomgm pmppppmpopSp oSp opSp hppooppo hpSEo op Soom mp xSozpoS Smpdpzp oSp .ppoo omhpaSoSmppoUSomQ oSp po Soopop pamMQ oSp p< .Axooo.mmv ppoo oopp-gocoonoooomo osp po Soooop Homom--.mo mmsopm m :r m m ::> U H [14 191 .ppoo mpSp pp mpp>ppom zpopopomm ommpthSm cpSopSmo sz wSpoSmwoS SSMSo on hoe mSOpmSpoSoo oS pmSp pSopmmmm mp pp .pozoopm SOppoMoS oSp So>m upmSB .mpsooOOSm poopEoSoopmpS mpSp Sp Spmpm hppmSpompppSm op oopSoS lop Soon mmS mSopoSS oSe .Amzoppmv mSopoSS oSp op mpomopo oopopmSHo hppMpoommo p59 .Emdpgopho oSp pSoSmSopSp hpoopz oompommpo on op Soom opmppmpoopm zSpmSm omSoo opoE m po pmpmSoo hoe pp So .mSOppoom poopoSoSo oSp Sp Smom mm Spmpm ooSpmSmnoSpp hpm> m on has pH .Smmpmopope mpSp Sp SpMpSoo poS mp posoopm SOppommS ommpozSSm cpSopSMo oSp po opSpMS oSe .Axomo.mmv pppmpEoSoopmpS ommpozSSm opSopSmo ppoo mgzpuSoSmSpooSommus.m: mmbwpm m: mmpmpm 193 .zpp>ppom ommpozgSm opSopSmo zoSm poS mooo MSonoS SmpSQSp oSp poSp pSoSmSSm mp pp .SOppoom ooSpmpm ommpszSm opSopSmo m SH .Axoom.mopv Soopop pmmmp ppoo mmhanoSmSpooSomS ooSpMpm ommpomgSm opSonSmoun.:: mmbmpm :r :- m a: :3 U H FL: 195 FIGURE 45.—-"Titres flamboyants final": Carbonic anhydrase stained pseudobranch-type cell mitochondrial- endoplasmic reticulum network (66,25OX). Micrograph A was from a carbonic anhydrase stairued tissue, whereas micrograph B represents an acet- azolamide-inhibited tissue. Both tissues seen are identical except for the darkness of the background. Also, in the uninhibited section, there are small granular particles which have been suspected of being the true carbonic anhydrase reaction product. The globular black densities in the inhibited micrographaimadirt which had condensed on the section. FIGURE 45 MICHIGAN STATE UNIVERSITY LIBRARIES I I 3 12933130712412