AMER. ZOOL., 24:775-782 (1984) Central Nervous System Changes Related to the Reduction of Visual Input in a Naturally Blind Fish (Astyanax hubbsi)1 THEODORE J. VONEIDA AND STEPHEN E. FISH Neurobtotogy Program, Northeastern Ohio Universities College of Medicine, Rootstown, Ohio 44272 SYNOPSIS. Tectal anatomy and physiology of the blind cave characin, Astyanax hubbsi, have been compared with that of its sighted ancestor Astyanax mexicanus (the river fish) and with goldfish. Normal and experimental neuroanatomic methods have revealed that, with the exception of a greatly reduced retinotectal projection, connectivity and structure of cave fish tecta are similar to those described in sighted species. It appears that the rudimentary retinotectal input is nonfunctional, since no tectal evoked responses could be elicited with electrical or visual stimulation of the optic cyst, and all attempts to visually condition cave fish were unsuccessful. Attempts have also been made to record somatosensory, auditory and lateral line activity in the tecta of the blind and sighted fish. A sparse somatic representation was found in the deeper portion of the sighted fish tecta which contrasts with a dense, well-organized one in the cave fish. No tectal responses were found to auditory or lateral line stimuli. CNS plasticity is discussed in relation to studies of fish, amphibians, reptiles, birds and mammals, in which a reduction of sensory input by any one of a number of means has resulted in alterations of structure and function. INTRODUCTION The blind cave fish, Astyanax hubbsi (formerly A. jordani), was discovered by C. B. Jordan in caves of the Sierra de El Abra region of Mexico, and was first described by Hubbs and Innes in 1936. The river fish, Astyanax mexicanus, is ancestral to A hubbsi (Mitchell et ai, 1977), and is native to the Rio Tampaon, which drains into the valley where the above caves are found. A. Mexicanus presently ranges into the southern United States, where we obtained our breeding stock. Our blind fish were originally obtained from tropical fish suppliers, though we now maintain a breeding colony for both groups. The availability of the blind fish offers a unique opportunity for the study of alterations in brain structure and function resulting from the relatively recent natural loss of a single sensory system. Other sensory systems appear to have remained intact. Indeed, behavioral studies suggest that in some cases these fish are superior to sighted forms. Thus Weiss and Martini (1970) and Gertychowa (1970) have shown that the blind fish outperforms visual fish in tasks depending on lateral line input; Breder and Rasquin (1943) and Humbach (1960) have reported that its chemical sense is much better developed than that of sighted fish. In addition, Popper (1970) has demonstrated that auditory capacity in the blind fish is equal to that of A mexicanus, and John (1957) has described its capability to respond to somatic cues. STUDIES OF THE VISUAL SYSTEM IN THE CAVE FISH Eye development It has been known since the studies of Cahn (1958) that optic cup and lens abnormalities are present throughout embryonic development in the cave fish. A degenerate eye remnant, called the "optic cyst," is visible as a surface structure at larval day 1, but by larval day 4 an overgrowth of the cyst by the optic lip is grossly apparent (Zille et ai, 1983). This overgrowth continues during the next week of development, until the optic cyst is completely covered by adipose, connective and epithelial tissues, and is no longer visible as a surface structure. ' From the Symposium on Evolution of Neural Sys- An early stage (second larval day) of optic tems in the Vertebrates: Functional-Anatomical Approaches lip overgrowth is shown in Figure 1A. A presented at the Annual Meeting of the American number of studies have been carried out Society of Zoologists, 27-30 December 1982, at on the genetic control of eye and lens Louisville, Kentucky. 775 776 T . J . VONEIDA AND S. E. FlSH FIG. 1. Head region of A. hubbsi (left) and A. mexicanus (right) at 2 days posthatching. The optic cyst is beginning to recede in A. hubbsi, with concurrent overgrowth of the optic lip (arrow). A. mexicanus eyes are well-developed at this stage. Scanning EM x 78. development and their relation to body pigmentation in these fishes (Sadoglu, 1956, 1957, 1975; Pfeiffer, 1967; Wilkens, 1971). There appear to be separate genes for at least retinal and lens development, and Sadoglu (1975) suggests that separate genetic controls act in the development of various other parts of the eye as well. and Gaffney, 1981), and Ictalurus (Prasada Rao and Sharma, 1982). Its significance remains unclear. Retinotectal fibers in the blind fish are entirely contralateral, penetrating the medial third of the tectum. Similar findings have been reported by Halpern (1973) in her studies of retinal projections in the blind snake Typhlops. The fact that retinotectal connections have Retinal projections: Anatomy and function remained in these specialized forms, in spite In an effort to clarify the question of of their apparent blindness, suggests the retinofugal connections in the blind fish, possibility that the tectum may have Voneida and Sligar (1976), using a modi- retained a potential for processing visual fied Nauta-Gygax (1954) technique, traced information. The question remains, if a degenerated optic nerve fibers following functional eye were made available, would removal of the optic cyst. We found both retinal axons follow the existing tectal retinotectal and retinodiencephalic (hypo- pathways, and would synaptic contacts be thalamic and pretectal) connections in the established? Finally, would the tectum blind cave form, though as might be prove capable of processing information expected, these were greatly reduced as mediated over this "newly acquired" syscompared to those found in the river fish. tem? We also found a small ipsilateral retinotecThe presence of sparse but consistent tal projection in the river fish, which has retinotectal connections in the blind fish been described as well in Petrotnyzon led us to ask whether these connections are (Northcutt and Przybylski, 1973), Serra- capable of mediating information from the salmus (Ebbesson and Ito, 1980), Pohpterus optic cyst. The studies of Breder and Ras(Reperant et ai, 1981), Caras.uu* (Springer quin (1947) have shown that variabilities CNS REMODELING IN A BLIND FISH exist among these animals in their degree of visual function. Our own findings indicate they are unresponsive to visual cues. We have been unsuccessful in our attempts to condition adult blind fish to respond to visual stimuli in a shock avoidance situation (unpublished data), whereas river and goldfish readily learned the problem. We have also studied the cave fish retinotectal connections electrophysiologically. To do this we have applied both light and electrical stimuli to the optic cyst and made extracellular recordings in the tectum. Again, river fish and goldfish served as controls. We were unable to elicit either slow wave potentials or single unit responses in the cave fish tectum. It appears that, if any information is conveyed to the tectum from the optic cyst, it is so weak as to be difficult to detect eletrophysiologically. As might be expected, visual response properties and topography similar to those reported in numerous earlier studies (e.g., Meyer, 1977) were observed in the river fish and goldfish. Tectal function: Anatomic and physiologic studies The above considerations (a degenerate eye and loss of visual function) lead directly into the question of tectal function in the blind fish. It is possible, of course, that the loss of visual input has resulted simply in a greatly reduced level of tectal function in these animals. Another possibility is that it has increased its processing capability for other, non-visual modalities and/or that the tectum has assumed new functions altogether. Whatever the case, the question lends itself to experimental analyses, some of which will be described below. Relatively little work has been done on tectal structure of connections in A. hubbsi. Sligar and Voneida's study (1976) demonstrates a rather widely distributed pattern of tectal efferents in this form, with connections to ipsilateral nucleus isthmi, torus longitudinalis, torus semicircularis, dorsal thalamus, pretectum, contralateral tectum and medullary reticular nuclei (bilaterally). In addition, we have found relatively large numbers of pyramidal, fusiform and large pyriform cells in the blind 777 fish tectum (unpublished data; Fig. 2). The pyramidal cells have long, sparsely branched apical dendrites extending dorsally from the periventricular layer into the superficial cell and fiber layer; the large pyriform cells are most abundant in deeper parts of the superficial cell/fiber layer, while fusiform cells lie slightly deeper, in the more superficial parts of the central cell zone. The latter two cell types have short, highly branched apical dendrites and short, sparsely branched basal dendrites. These three principal tectal cell types have been described in numerous cytoarchitectural analyses of the tectum, dating from the early work of Cajal (1911). The fact that they have been retained in a blind form may be of some significance, and opens the question of their role in tectal information processing in this animal. Schroeder and Vanegas(1977), for example, have pointed out that these cells are also found in relatively large numbers in two siluroid fishes (Bagrus and Ictalurus), both of which are known to have restricted visual input. They have suggested that these cell types, particularly the pyradmidal cell, may be significantly involved in non-visual inputs. Our own investigations of tectal function in the cave fish include electrophysiological tests for somatosensory, auditory and lateral line evoked responses in the tecta of cave, river and goldfish. In these studies, we used standard single and multiple unit extracellular recording techniques. The fish were anesthetized using MS 222 (1 part to 10,000-20,000 parts water perfused over the gills) or urethane (3 g/kg IP) and their gills were perfused with artificially aerated water. The only evoked activity that we were able to record in the cave fish tectum was that generated by somatosensory stimulation. Rubbing or tapping the side of the fish with small brushes or glass probes evoked multiple unit activity as soon as the electrode tip entered the tectum contralateral to the stimulation. Somatically driven single units were isolated in all depths of the tectum but they tended to be found with increasing frequency as the electrode penetrated to the deeper tectal layers, as shown in Figure 3. Somatotopy was clearly evident, since 778 T . J. VONEIDA AND S. E. FlSH 779 CNS REMODELING IN A BLIND FISH ANTERIOR POSTERIOR 100 10 15 20 25 30 NUMBER OF UNITS Fic. 3. Numbers of somatosensory units in the tectum of A. hubbsi are plotted according to depth in the tectum. Notice that units are recorded more frequently in deeper layers. successive tectal penetrations from rostral to caudal or medial to lateral yielded corresponding rostro-caudal and dorso-ventral receptive fields, respectively, on the body surface. A composite map developed from histological reconstruction of several cave fish tecta is shown in Figure 4. In electrode penetrations normal to the tectal surface, successively recorded units yielded overlapping receptive fields which tended to decrease in size with depth below the surface. Receptive fields on the head tended to be smaller than on the rest of the body (Fig. 5), and this is reflected in the large area of the tectum devoted to the head. This latter organization (a direct relationship between central representation and peripheral input) is typical of sensory topography in the vertebrate central nervous system. Somatosensory single units were recorded only in the deepest layers of the river fish and goldfish optic tecta. There was a suggestion of the organization found in the cave fish but this was difficult to determine since these units were only rarely encountered. It seems, then, that there has been an increase in somatic input to the blind fish tectum which parallels the decrease in MEDIAL FIG. 4. A modified dorsal surface view of the left tectum of A. hubbsi is depicted. The electrophysiologically plotted somatosensory map is somewhat distorted as the tectum is a hemispherical structure and the ventral lateral aspect would normally be out of sight from this view. visual input. It still remains to be determined, however, where this somatic input arises. Extravisual sensory modalities have been found in the midbrain tectum of every animal studied. A somatosensory representation is one of the most common findings in such diverse animals as, for example, the chicken (Cotter, 1976), hamster (Chalupa and Rhoades, 1977), axolotyl (Gruberg and Harris, 1981), and iguana (Gaither and Stein, 1979). Our finding of somatic activity in the tectum of the three fish we have studied is, therefore, not surprising. The usual organization of the midbrain tectum is a visual input to the superficial layers, with deeper layers receiving other modalities. Typically the sensory modalities have a well-organized topographic representation relative to the surface of the tectum and they are in spatial register. Thus cells with visual receptive fields directly in front of the animal near the horizontal and vertical meridians would be locatedjust above somatosensory cells with receptive fields on the surface of the body of the animal near Fic. 2. Characteristic cell types in the optic tectum of A. hubbsi: A. Pyramidal cell (arrow) in the periventricular layer, with its apical dendrite extending dorsally into the superficial cell/fiber layer. B. Large pyriform cells in the deeper part of the superficial cell/fiber layer (arrows); note also the (fusiform?) cell (open arrow) lying somewhat more ventrally. 780 T . J . VONEIDA AND S. E. FlSH FIG. 5. Four representative somatosensory receptive fields of tectal units (cross-hatched areas) are shown plotted on the right body surface of A. hubbsi. Notice that the receptive fields on the head are smaller than those plotted more posteriorly (compare with the large head area on the tectal margin Fig. 4) and that one of them is on the skin overlying the optic cyst. its nose. Likewise neurons with temporal visual receptive fields would be located near ones with somatosensory receptive fields on the hindquarters of the animal. Although the somatic representation we found in the tectum of the visual fish was too sparse to determine if there was a match with the retinotopy, it is interesting that the cave fish somatotopy (Fig. 4) is in good register with goldfish visuotopy (e.g., Meyer, 1977). This same plan has been found to obtain for other sensory modalities as well. Thus, auditory localization in the cat (Gordon, 1973) and infrared responsive neurons in the rattlesnake (Hartlinee/ ah, 1978) are organized topographically, such that their receptive fields correspond spatially with visual receptive fields. It has been suggested that this polymodal input to the tectum plays a role in orienting behaviors which can be initiated by the separate modalities. It is very likely that fish use lateral line information to orient to their environment, and an input from this system to the tectum has been identified in the longnose gar (Northcutt, 1982). We were somewhat surprised, therefore, that we were totally unsuccessful in the blind fish in identifying tectal units which are responsive to auditory or lateral line stimuli (low frequency tones and agitation of the water). These same stimuli, however, invariably elicited evoked responses in the torus semicircularis. Our investigation of lateral line involvement in the cave fish tectum is continuing. It appears that a primary role of the midbrain tectum in the blind cave fish is that of processing somatosensory information, and it may prove that this change is an adaptation to cave life. It has been proposed that the optic tectum plays a major role in orienting activities in visually responsive animals. It is possible that this orienting role in the blind fish is still mediated by the tectum, but that somatic input has now become the principal guiding modality. Consistent with this view are our observations of cave fish placed in a new environment (unpublished; see also John, 1957; Gertychowa, 1970). Initially the fish make much physical contact with their surroundings by bumping into objects, but within a relatively short time CNS REMODELING IN A BLIND FISH they appear to negotiate obstacles as if they are able to see them. It is as if they have established a spatial map of their surroundings based on the use of somatic input. CENTRAL CHANGES RESULTING FROM A REDUCTION OF SENSORY INPUT: A COMPARISON WITH OTHER MODELS It is interesting to compare the above observations of peripheral and central changes in a naturally evolved system with those which have resulted from experimentally induced changes. When normally sighted hamsters (Rhoades, 1980) are enucleated at an early age, their tecta change during subsequent development. The normal somatosensory representation is expanded dorsally to fill the entire tectum, resulting in an organization much like that found in the cave fish. A similar finding has been made in eyeless mutant salamanders (Gruberg and Harris, 1981), and we have recently observed the same change in mutant anophthalmic mice (preliminary electrophysiological results). Is the cave fish then simply an evolutionarily enucleated river fish? It seems unlikely that the differences between cave and river fish tecta are due entirely to the developmental consequences of a reduction of visual input. The expansion of somatic representation could be explained by mechanisms such as axonal sprouting (Liu and Chambers, 1958), though the somatosensory representation in the sighted fish would seem to be too sparse to account for the major reorganization we have documented in the cave fish. We have attempted to shed some light on this question by removing one or both of the eyes of young river fish so that we might compare their brains, as adults, with those of cave fish.2 These data remain to be analyzed. Will the tectum contralateral 2 An incidental finding of some interest relates to the very high mortality rate (>99%) observed in these animals. Eyed animals which live in the dark may be subject to a greatly increased risk of infection as a result of damage to these anterior "protrusions" during collisions with the cave walls. This could presumably lead to a strong selective pressure against the retention of eyes. 781 to eye removal undergo structural and physiological changes which are more typical of tectal structure and function in the blind fish? In other words, will it become a "somatosensory" tectum? Conversely, is it possible to modify the blind fish tectum from a "somatosensory" to a "visual" tectum by transplanting a river fish eye to the blind animal? Answers to these questions could provide further clues to an understanding of central nervous plasticity, whether it results from experimental manipulations, genetic mutations, or naturally occurring adaptations, as observed in a blind cave animal such as Astyanax hubbsi. ACKNOWLEDGMENTS The authors would like to extend special thanks to Mr. and Mrs. James White, local fish hobbyists, for their expert assistance in the maintenance and breeding of our experimental animals. This work was supported by NIH grants MH-07051, NS 18369 and NS 06136. REFERENCES Breder, C. M., Jr. and P. Rasquin. 1943. Chemical sensory reactions in the Mexican blind characins. Zoologica 28:169-200. Breder, C. M., Jr. and P. Rasquin. 1947. Comparative studies in the light sensitivity of blind characins from a series of Mexican caves. Bull. Amer. Mus. Nat. Hist. 89:325-351. Cahn, P. H. 1958. Comparative optic development in Astyanax mexicanus and in two of its blind cave derivatives. Bull. Amer. Mus. Nat. Hist. 115(2): 69-112. Cajal, R. 1911. Histologie du systeme nerveux de l'homme et des vertebres. 2 Vol. Maloine, Pans. Chalupa, L. M. and R. W. Rhoades. 1977. Responses of visual, somatosensory, and auditory neurons in the golden hamster's superior colliculus. J. Physiol. (London) 270:595-629. Cotter,J. R. 1976. Visual and nonvisual units recorded from the optic tectum of Callus domesticus. Brain, Behav. Evol. 13:1-21. Ebbesson.S. O. E. and H. Ito. 1980. Bilateral retinal projections in the black piranha (Serrasalmus ntger). Cell Tissue Res. 213:483-495. Gaither, N. S. and B. E. Stein. 1979. Reptiles and mammals use similar sensory organizations in the midbrain. Science 205:595-597. Gertychowa, R. 1970. Studies on the ethology and space orientation of the blind cave fish Anoptichthys jordani Hubbs et Innes 1936 (Ckaracidae). Folia Biol. 18:9-69. Gordon, B. 1973. Receptive fields in deep layers of 782 T . J . VONEIDA AND S. E. FlSH cat superior colliculus. J. Neurophysiol. 36:157- Prasada Rao, P. D. and S. C. Sharma. 1982. Reti178. nofugal pathways in juvenile and adult channel catfish, Ictalurus (Ameiurus) punctatus: An HRP Gruberg, E. R. and W. A. Harris. 1981. The serotonergic somatosensory projection to the tectum and autoradiographic study. J. Comp. Neurol. of normal and eyeless salamanders. J. Morphol. 210:37-48. 170:55-69. Reperant, J., J. P. Rio, D. Miceli, M. Amouzou, and Halpern, M. 1973. Retinal projections in blind snakes. J. Peyrichoux. 1981. The retinofugal pathways Science 182:390-391. in the primitive African bony fish Polypterus seneHartline, P. H., L. Kass.andM.S. Loop. 1978. Merggalus (Cuvier, 1829). Brain Res. 217:225-243. ing of modalities in the optic tectum: Infrared Rhoades, R. W. 1980. Effects of neonatal enucleation and visual integration in rattlesnakes. Science 199: upon the functional organization of the superior 1225-1229. colliculus of the golden hamster. J. Physiol. (London) 301:383-399. Hubbs, C. L. and W. T. Innes. 1936. The first known blind fish of the family Characidae: A new genus Sadoglu, P. 1956. A preliminary report on the from Mexico. Occas. Papers Mus. Zool. Univ. genetics of the Mexicanus cave characins. Copeia Michigan, No. 342:1-7. 1956:113-114. Humbach, I. 1960. Geruch und Geschmack bei den Sadoglu, P. 1957. Mendelian inheritance in the Augenlosen Hohlenfischen Anoptichlhys jordani hybrids between the Mexicanus blind cave fish Hubbs und Innes. Naturwissensch. 47:551-552. and their overground ancestor. Verh. Deut. Zool. Ges. 432-439. John, K. R. 1957. Observations on the behavior of blind and blinded fishes. Copeia 2:123-132. Sadoglu, P. 1975. Genetic paths leading to blindness Liu, C. N. and W. W. Chambers. 1958. Intraspinal in Astyanax mexicanus. In M. A. AH (ed.), Vision in sprouting of dorsal root axons. Arch. Neurol. fishes, pp. 419-426. Plenum Publ. Corp., New Psychiat. 79:46-61. York. Meyer, R. L. 1977. Eye-in-waterelectrophysiological Schroeder, D. M. and H. Vanegas. 1977. Cytoarmapping of goldfish with and without tectal chitecture of the tectum mesencephali in two types lesions. Exp. Neurol. 56:23-41. of siluroid teleosts. J. Comp. Neurol. 175:287300. Mitchell, R. W., W. H. Russell,and W. R. Elliot. 1977. Mexican eyeless characin fishes, genus Astyanax: Sligar, C. M. and T. J. Voneida. 1976. Tectal efferEnvironment, distribution, and evolution. Spec. ents in the blind cavefish,Astyanax hubbsi.]. Comp. Publications Mus. Texas Tech. Univ. 12:1-89. Neurol. 165:107-124. Nauta, W. J. H. and P. H. Gygax. 1954. Silver Springer, A. D. andj. S. Gaffney. 1981. Retinal proimpregnation of degenerating axons in the cenjections in the goldfish: A study using cobaltoustral nervous system: A modified technique. Stain lysine. J. Comp. Neurol. 203:401-424. Tech. 19:91-93. Voneida, T.J. and C. M. Sligar. 1976. A comparative Northcutt, R. G. 1982. Localization of neurons afferneuroanatomic study of retinal projections in two ent to the optic tectum in longnose gars. J. Comp. fishes: Astyanax hubbsi (the blind cave fish) and Neurol. 204:325-335. Astyanax mexicanus J. Comp. Neurol. 165:89—106. Northcutt, R. G. and R. J. Przybylski. 1973. Retinal Weiss, B. A. and J. L. Martini. 1970. Lateral-line projections in the lamprey Petromyzon marinus L. sensitivity in the blind cavefish (Anoptichthys jorAnat. Rec. 150:400. dani).]. Comp. Physiol. Psych. 71:34-37. Pfeiffer, W. 1967. Die Korrelation von Korperlange, Wilkens, H. 1971. Genetic interpretation of regresAugen-, Linsen-, und Pupillengrosse bei Hybrisive evolutionary processes: Studies on hybrid eyes of two Astyanax cave populations (Characidae, den aus Astyanax, Anoptichthys (characidae, Pisces). Pisces). Evolution 25:530-544. Wilhelm Roux Arch. Entwicklungsmeck Org. 158:218-245. Zilles, K.,B. Tillman, and R. Bennemann. 1983. The development of the eye in Astyanax mexicanus Popper, A. N. 1970. Auditory capacities of the Mex(Characidae, Pisces), its blind cave derivative, ican blind cavefish (Astyanax jordani) and its eyed Anoptichthys jordani (Characidae, Pisces), and their ancestor (Astyanax mexicanus). Anim. Behav. 18: crossbreds. Cell and Tiss. Res. 229:423-432. 552-562.
© Copyright 2026 Paperzz