Development of Tectal Connectivity across Metamorphosis in the Bullfrog (Rana catesbeiana)

In the bullfrog (Rana catesbeiana) , the process of metamorphosis culminates in the appearance of new visual and visuomotor behaviors reflective of the emergence of binocular vision and visually-guided prey capture behaviors as the animal transitions to life on land. Using several different neuroanatomical tracers, we examined the substrates that may underlie these behavioral changes by tracing the afferent and efferent connectivity of the midbrain optic tectum across metamorphic development. Intratectal, tectotoral, tectotegmental, tectobulbar, and tecto-thalamic tracts exhibit similar trajectories of neurobiotin fiber label across the developmental span from early larval tadpoles to adults. Developmental variability was apparent primarily in intensity and distribution of cell and puncta label in target nuclei. Combined injections of cholera toxin subunit and Phaseolus vulgaris leucoagglutinin consistently label cell bodies, puncta, or fiber segments bilaterally in midbrain targets including the pretectal gray, laminar nucleus of the torus semicircularis, and the nucleus of the medial longitudinal fasciculus. Received: December 1, 2009 Returned for revision: December 28, 2009 Accepted after revision: November 1, 2010 Published online: January 24, 2011 A.M. Simmons Department of Cognitive, Linguistic and Psychological Sciences Box 1821, Brown University Providence, RI 02912 (USA) Tel. +1 401 863 2283, Fax +1 401 863 1300, E-Mail Andrea_Simmons @ brown.edu © 2011 S. Karger AG, Basel Accessible online at: www.karger.com/bbe D ow nl oa de d by : 54 .1 91 .4 0. 80 9 /1 6/ 20 17 8 :3 8: 53 P M Development of Tectal Connectivity in Bullfrogs Brain Behav Evol 2010;76:226–247 227 restructuring of the head, metamorphosis also features substantial modifications of sensory systems in preparation for life out of water. Several such modifications have been documented in the auditory system, where a receptor apparatus initially designed for detecting sounds and vibrations underwater must adapt to one capable of operating in air, a medium with different biophysical constraints. In bullfrog (Rana catesbeiana) tadpoles, the progressive formation and maturation of inner ear organs and sound transduction pathways are accompanied by and drive functional and anatomical changes in central auditory nuclei [Boatright-Horowitz and Simmons, 1997; Simmons and Horowitz, 2006; Horowitz et al., 2007a]. During metamorphic climax, the dorsal medulla reorganizes anatomically to encompass sensory input from the newly-emerging external tympanum and to compensate for the degeneration and loss of the lateral line system. Modifications in midbrain-medullary pathways occurring around climax mediate the increased importance of binaural hearing in postmetamorphic life [Horowitz et al., 2007b]. The visual apparatus also undergoes structural modifications over the metamorphic transition. In the bullfrog tadpole, eyes are lidless and without nictitating membranes [Stehouwer, 1988]. These lidless eyes are relatively mobile, likely providing vestibulo-ocular stabilization during swimming. During metamorphic climax stages, a nictitating membrane first appears and ocular mobility gradually decreases. Concurrently, the eyes move from a more lateral to a more dorsal position on the head [Gosner, 1960; Jacobson, 1971]. This results in a broadening of the frontal binocular field appropriate for mediating the postmetamorphic behavioral shift from feeding on stationary plant life to tracking moving animal prey. These considerable changes in visual and visuomotor behaviors likely reflect underlying developmental modification in central visual anatomy and function; unfortunately, experimental analyses of the development of central visual pathways in tadpoles have been limited, both in the number of projections investigated and in the species used for analysis. The time course of maturation of projections from the retina to the optic tectum (OT) has been examined in the leopard frog, R. pipien s, and in the African clawed frog, Xenopus laevis [Jacobson, 1971; Reh and Constantine-Paton, 1984; Sakaguchi and Murphey, Abbreviations used in this paper as aqueduct of Sylvius NI nucleus isthmi BON basal optic nucleus nMLF nucleus of the medial longitudinal fasciculus C central nucleus of the thalamus oc optic chiasm Cb cerebellum OT optic tectum CerN cerebellar nucleus ov optic ventricle CT Vibrio cholera toxin subunit P posterior nucleus of the thalamus dap dorsal arcuate pathway Para paragigantocellular nucleus DMN dorsal medullary nucleus pc posterior commissure DP dorsal pallium PD nucleus posterodorsalis tegmenti mesencephali Ep entopeduncular nucleus PHA-L Phaseolus vulgaris leucoagglutinin HRP horseradish peroxidase PTG pretectal gray III oculomotor nucleus RG reticular gray IV trochlear nucleus SON superior olivary nucleus lfb lateral forebrain bundle Str striatum LLa anterior lateral line nucleus tbt tectobulbar tract LLnp lateral line neuropil teg fasc tegmental fasciculi lot lateral optic tract Teg tegmentum LPD lateral-posterior nucleus of the thalamus tlt tecto-lemniscal pathway LS lateral septum TS torus semicircularis lv lateral ventricle TSl laminar nucleus of the torus semicircularis LVN lateral vestibular nucleus TSp principal nucleus of the torus semicircularis mfb medial forebrain bundle vap ventral arcuate pathway MP medial pallium V trigeminal nucleus MS medial septum VM ventromedial nucleus of the thalamus MVN medial vestibular nucleus D ow nl oa de d by : 54 .1 91 .4 0. 80 9 /1 6/ 20 17 8 :3 8: 53 P M Horowitz/Simmons Brain Behav Evol 2010;76:226–247 228 1985]. The development of connections between the OT and the nucleus isthmi (NI) has also been studied in these 2 species [Grobstein and Comer, 1983; Udin and Fisher, 1985; Chahoud et al., 1996], as well as in Limnodynastes dorsalis [Dann and Beazley, 1982, 1990]. Only a few studies, in larval X. laevis [Chahoud et al., 1996; Deeg et al., 2009] and in larval R. pipiens [Debski and ConstantinePaton, 1993], examined tectal connectivity to medullary and thalamic nuclei. This scarcity of data constrains our understanding of factors that may impact the development and maturation of visual pathways over the dramatic transition of metamorphosis. Because of the differences between larval Xenopus and Rana in optomotor responses [Wassersug, 1971] and the time course and extent of the dorsal migration of the eyes during and after metamorphic climax [Grobstein and Comer, 1977], there are likely differences in the pattern of maturation of visual system anatomy. Moreover, because Xenopus remains fully aquatic after the completion of metamorphosis while Rana becomes semi-terrestrial, the visual systems of these animals are not only subject to different ecological constraints but also respond differently to visual experience [Kennard and Keating, 1985]. Therefore, describing metamorphic changes in visual pathways in anurans on the basis of Xenopus data alone might obscure important species differences in visual anatomy and development related to adult function. Using several different tract-tracing techniques, we examined the pattern of connectivity from the OT to medullary, midbrain and forebrain target sites over the time course of larval development in the bullfrog. We chose the bullfrog for analysis to complement our previous work [Horowitz et al., 2007b] outlining the considerable variability in the connectivity of brainstem auditory pathways over metamorphic development, and because little is known about visual development in this species. Our results show both stability and modifiability of tectal projection patterns over the metamorphic transition to semiterrestrial life. The overall pattern of connectivity between the OT and many of its target nuclei in the midbrain, medulla, and thalamus is present at early larval stages and remains stable throughout metamorphosis and into postmetamorphic life. This stability highlights the importance of visually-mediated orienting behaviors throughout the life cycle. Alternatively, we also observe a delayed appearance of mature connectivity between the OT and other brainstem and forebrain targets. These more variable patterns might reflect the increased importance of binocular vision and head movements for the stabilization of gaze in terrestrial postmetamorphic animals. Materials and Methods Animals All experimental procedures were reviewed and approved by the Brown University Institutional Animal Care and Use Committee, and conform to federal guidelines. R. catesbeiana tadpoles and adults were purchased from a commercial supplier (Dozier Lester, Duson, La., USA). Tadpoles were staged by external morphological criteria based on Gosner [1960] and then classified on the basis of these criteria into 4 stage groups: hatchlings (stages 21–24, the earliest postembryonic stages); early larval (stages 25– 30, with undeveloped hindlimb buds), late larval (stages 31–41, with progressive emergence and differentiation of hindlimbs), and metamorphic climax (stages 42–46, with fully-developed hindlimbs, initial emergence progressing to full development of forelimbs, head reshaping, and tail resorption at the end of climax). After completion of metamorphosis, animals were categorized as froglets, subadults, or adults based on snout-vent length [Boatright-Horowitz and Simmons, 1995]. Tadpoles were grouphoused in polycarbonate aquaria containing dechlorinated aerated water (pH 7–8) and were fed cooked unsalted spinach and fish flakes ad libitum. Postmetamorphic animals were individually housed in polycarbonate terraria containing both soil and water and were fed live crickets ad libitum. The colony room was maintained at temperatures ranging from 25–28 ° C and on a 12: 12 light:dark cycle. Tracer Deposits and Anatomical Staining Animals were anesthetized for surgery by immersion in 0.15% (tadpoles and froglets) or 0.6% (adults) buffered tricaine methanesulfonate (MS 222, pH 7.0; Sigma, St. Louis, Mo., USA) until all reflexes disappeared. They were wrapped in wet gauze and placed on an ice block for surgery. The OT was exposed by removing overlying cartilaginous tissue (tadpoles) or bone drilling (postmetamorphic animals). Small tears were made in the meninges with a metal microelectrode to accommodate the micropipettes holding tracer. In 21 animals (tadpoles, n = 18; postmetamorphic, n = 3), we injected into the OT the tracer biotin ethylendiamine (Neurobiotin, SP1120; Vector Laboratories, Burlingame, Calif., USA). Neurobiotin labels cell bodies and fibers in both anterograde and retrograde directions from the injection site [Huang et al., 1992; Jacquin et al., 1992]. For iontophoretic injections (n = 7 animals), a solution of 4% neurobiotin in 1.0 M potassium chloride was backfilled into 1.0-mm borosilicate glass capillary tubes (1B100F-4; World Precision Instruments, Sarasota, Fla., USA) pulled to a 20 m tip diameter. Micropipettes were held by a Narishige (Narishige Corp., East Meadow, N.Y., USA) micromanipulator. A silver wire placed into the neurobiotin solution was connected to a constant current source (Model 51413; Stoelting, Wood Dale, Ill., USA). Micropipettes were placed at approximately the rostralcaudal point (mid-levels) of 1 tectal lobe, targeting both more medial and more lateral locations; the size of the surgical opening typically prevented visually-guided placement of the pipette at either very caudal or very rostral tectal locations. Iontophoresis was carried out using 0.4–0.6 A alternating current for 3 min. The micropipette remained in place for 5 min after termination of the current in order to allow the tracer to permeate the brain region. The site was then thoroughly rinsed with 0.9% sterile saline, and antiseptic Gelfoam (Upjohn, Kalamazoo, Mich., USA) D ow nl oa de d by : 54 .1 91 .4 0. 80 9 /1 6/ 20 17 8 :3 8: 53 P M Development of Tectal Connectivity in Bullfrogs Brain Behav Evol 2010;76:226–247 229 was placed in the wound. Pressure injections (106–212 nl; n = 14 animals) were performed using a 10l manual microinjector (Sutter Instruments, Novato, Calif., USA) or a Nanoject II (Drummond Scientific, Broomall, Pa., USA). Tracer was again targeted to the mid-levels of 1 tectal lobe, with approximately equal numbers of medial and lateral injections attempted. The skin was closed using 5/0 Vicryl (Ethicon, Somerville, N.J., USA) sutures, and triple antibiotic containing lidocaine (Neosporin; Pfizer, Cambridge, Mass., USA) was topically applied. Animals were housed singly in polycarbonate aquaria with aerated water until euthanasia. Recovery times ranged from 4 to 24 h for tadpoles and from 4 to 48 h for postmetamorphic animals. Careful analysis of the data revealed no significant differences in distance of tracer travel or location of target area based on either recovery time or mode of injection. In additional experiments (n = 3 tadpoles, n = 1 froglet), we performed much larger injections and allowed the animals to survive 24 h. These larger injections produced considerable tracer leakage into the lateral regions of the torus semicircularis (TS) and tegmentum (Teg), thus complicating the separation of tectal projections from those of these other brain nuclei. For this reason, data in this paper are based on smaller OT injections that did not include leakage into the TS or Teg. In 31 animals (tadpoles, n = 26; postmetamorphic, n = 5), we co-injected the tracers Phaseolus vulgaris leucoagglutinin (PHAL; L32455; Invitrogen, Eugene, Oreg., USA) and Vibrio cholera toxin subunit (CT ; C22841; Invitrogen) at approximately the rostral-caudal midpoint of 1 tectal lobe, targeting similar numbers of medial and lateral locations. These tracers were conjugated to AlexaFluor 568 and 488, providing fluorescent label of orange/red and green, respectively. PHA-L is a strong anterograde tracer, while CT  is a robust retrograde tracer that undergoes weaker anterograde transport; neither tracer shows transneuronal transport [Gerfen and Sawchenko, 1984; Luppi et al., 1995]. The simultaneous use of these 2 tracers allowed us to investigate the distribution of afferent and efferent processes from the same injection site. Our data show evidence of some limited anterograde transport of CT , but no retrograde transport of PHA-L. A glass micropipette (20–40 m tip diameter) was backfilled with tracer, and manipulated into the OT (approximately midway between very rostral and very caudal locations) under visual control. Tracer (106–212 nl) was pressure-injected as described above. The micropipette remained in place for 5 min after injection, and the wound was closed as described above. Animals were allowed to recover for 1 (n = 15), 2 (n = 2) or 3 (n = 11) days before euthanasia. Although longer survival times produced greater intensity of label, there were no consistent differences in identity of target areas related to survival time. Several control experiments were performed to assess any patterns of label due to tracer leakage and to help interpret possible fiber of passage label. In 18 animals (tadpoles: n = 16; postmetamorphic: n = 2), tracers (PHA-L/CT , n = 12; neurobiotin, n = 6) were deliberately injected into the choroid plexus overlying the medulla, the membranes overlying the OT, or directly into the optic ventricle, at the same approximate volumes as used for OT injections. In none of these animals did we observe anything other than limited periventricular label or label of the membranes surrounding the brain, with no consistent label of brain parenchyma. In 13 other tadpoles, small injections (PHA-L/CT , n = 10; neurobiotin, n = 3) were made directly into the TS. Data from these injections were compared to those described elsewhere [Horowitz et al., 2007b] and used to assess fiber of passage label and any leakage from large OT injections. At the end of the recovery period, animals were immersed in 0.6% MS 222 until all reflexes disappeared. They were transcardially perfused with heparinized 0.9% saline followed by 4% paraformaldehyde (pH 7.4). Brains were removed and postfixed overnight at 4 ° C in 4% paraformaldehyde. Overlying meninges were removed and brains were embedded in agarose (5% in saline; ISC BioExpress, Kaysville, Utah, USA) and sliced by vibratomy (50 m coronal or 75  m horizontal sections; Vibratome, St Louis, Mo., USA). In a subset of animals, the retina contralateral to the injection site in the OT was removed and flat-mounted to provide an indication of tracer spread. Sections from neurobiotin injected brains were mounted on gelatin-subbed slides, rinsed 3 times for 20 min in 1 ! phosphate buffered saline (PBS, pH 7.4) with 1% Triton-X 100 (Sigma) for 24 h at 4 ° C, then incubated in streptavidin Alexa Fluor 488 conjugate (1: 200 dilution; Invitrogen) for 24 h at 4 ° C, producing a green fluorescence. Following incubation, sections were rinsed overnight in 1 ! PBS, and cover slipped with non-fluorescent mounting medium (Aqua Poly/Mount; Polysciences, Warrington, Pa., USA). In brains with PHA-L/CT injections, sections were directly mounted on gelatin-subbed slides. Some sections were counterstained with 0.5% cresyl violet acetate (pH 3.7 using glacial acetic acid) to ease identification of nuclear boundaries, and were mounted with Cytoseal (Fisher Scientific, Pittsburgh, Pa., USA). Image Processing and Analyses Brain sections were visualized and imaged using an Olympus BX60 research microscope (Melville, N.Y., USA) equipped with custom-made fluorescence cubes and an Olympus DP72 camera (Olympus America, Center Valley, Pa., USA). Images were acquired using a Dell Pentium IV 2.4GHz computer running image acquisition and analysis software (Olympus) and were stored as 24-bit RGB TIFF images. Images were contrast adjusted and color balanced to match luminance between sections. Schematization of cell bodies and nerve terminal location was carried out by dividing 24-bit RGB TIFF files into red (PHA-L), green (CT ) and blue (cresyl violet) 8-bit grayscale images, converting these to thresholded binary images, tracing the edges of the objects and passing the tracings through particle size filters with parameters based on a selection of 25 randomly chosen cells from the regions of interest (Image J, National Institutes of Health, Bethesda, Md., USA). Selected sections were mounted in AntiFade (Invitrogen) and imaged using a Leica TCS SP2 confocal microscope system (sequential scan mode; Leica Microsystems, Bannockburn, Ill., USA). High power z stack images were taken through each section and viewed in orthogonal planes using Leica confocal software (Leica Microsystems). To provide an indication of the relative intensity of label present in particular target areas, the authors and a research assistant blind to the hypothesis of the experiment independently categorized label in that target area as none to minimal, sparse/light, moderate, or heavy ( tables 1–3 ). These categorizations were combined across the raters and across stages separately for each of the 4 different larval stage groups, early postmetamorphic (froglets and subadults) frogs, and adult frogs. At all developmental stage groups, similar amounts of tracers were injected, and so the aggregate data from any particular group reflect comparable numbers of relatively larger and relatively smaller injection sites withD ow nl oa de d by : 54 .1 91 .4 0. 80 9 /1 6/ 20 17 8 :3 8: 53 P M Horowitz/Simmons Brain Behav Evol 2010;76:226–247 230 in the limited range we used. Tables show the most common rating of label within a particular target area across all animals within that developmental group; individual animals within a group could show some variability from these typical patterns, attributable to the size of a particular injection site.