Summary. Following the application either of cobaltic- lysine complex or a 30% solution of horseradish perox- idase (HRP) in a sealed tube to the cut end of the ...
Arch. Histol. Cytol., Vol. 52, No. 2 (1989) p. 87-93
Dendritic
Morphology
in Xenopus
laevis
of Horseradish Retrograde Pal
TOTH* * and
of Identified
: A Comparison
Peroxidase
and
Retinal
Ganglion
between
Cells
the Results
Cobaltic-Lysine
Labelling* Charles
STRAZNICKY
Department of Anatomy and Histology, School of Medicine, The Flinders University of South Australia, Bedford Park, Australia Received August 31, 1988
Summary. Following the application either of cobalticlysine complex or a 30% solution of horseradish peroxidase (HRP) in a sealed tube to the cut end of the optic nerve of young adult Xenopus frogs, the dendritic morphology of large ganglion cells was studied in retinal wholemount preparations, and compared with that in animals of the same size as revealed by the short time administration of HRP crystals. In the former two groups of animals, after 24 h survival, the size of the dendritic arborization of characterized ganglion cell types, Types I and III, were found to be significantly larger (61-79% and 180-187%, respectively) than those which survived 3 days after the administration of HRP crystals. These findings suggest that the very fine dendritic branches of large ganglion cells may remain unlabelled after a short-time exposure to HRP.
applied to the cut optic fibres. Because of the relatively quick elimination of the enzyme by tissue fluids, the HRP is available for uptake by optic axons for a limited time only. It is, however, generally assumed that such short-time exposure to HRP can reveal most, if not the entire, dendritic arborization of different neuron types, and thus of different classes of retinal ganglion cells. Retinal ganglion cells have been characterized morphologically in juvenile (NEVILLE and STRAZNICKY, 1988) and adult Xenopus (STRAZNICKY and STRAZNICKY, 1988) using the standard technique of ADAMS (1981) which utilizes nickel-cobalt intensified diaminobenzidine as chromogen. Of the 12 morphologically distinct ganglion cell types, Types I, II and III represent neurons with large to very large soma and dendritic arborization. In an attempt to achieve a better quantitative description of the dendritic morphology of these cells, either cobaltic-lysine complex (CLC; GORCSet al., 1979) or a 30% solution of HRP in a sealed plastic tube was applied to the cut optic nerve. Surprisingly, the dendritic field sizes of large Types I and III ganglion cells visualized by these techniques were found to be significantly larger than those of the same ganglion cell types in similar retinal locations revealed by the HRP technique
Recent studies utilizing the retrograde transport of horseradish peroxidase (HRP) have revealed a great morphological diversity of ganglion cells in the retina of fish (DUNN-MEYNELL and SHARMA, 1986; HITCH000K and EASTER, 1986), amphibian (ARKIN and MILLER, 1988; FRANK and HOLLYFIELD,1987; STRAZNICKY and STRAZNICKY,1988) and mammalian (WAsSLE and ILLING, 1980) species. The morphological characterization of the ganglion cells has been based on the sizes of the soma and dendritic arborization, and on dendritic branching patterns. Some of these studies used iontophoretic or pressure injection of HRP into or around the optic nerve; in other cases a gelfoam * **
This
soaked study
On leave
was from
in 30% HRP or HRP crystal supported
by a Flinders
the Department
Medical
of Anatomy,
previously used in this laboratory. In this report a full description is given of the method used and the results of morphometric analyses on CLC-filled large ganglion cells are compared with those obtained after two different methods of HRP application.
was Centre
University
Research Medical 87
Foundation School,
Pecs,
grant Hungary.
to C. S.
88
P. TOTH and C. STRAZNICKY:
MATERIALS
AND
METHODS
Sixteen young adult Xenopus laevis frogs, approximately 9-12 months postmetamorphic with a standard 40 mm crown-rump length were used. The animals were anaesthetized by immersion in 0.1%
MS222 (tricaine methanesulfonate, Sandoz) and the optic nerve was approached through the roof of the mouth and cut close to the entry point into the skull. The peripheral stump of the nerve was either exposed to crystalline HRP (Sigma Type IV; 4 cases) or introduced into a small polyethylene tube filled with 30% HRP in distilled water (4 cases) or with 350 mM
Fig. 1. Histophotographs showing the dendritic arborization of large ganglion cells in the Xenopus retina after filling with crystalline HRP (A and C) and CLC (B and D). Notice the differences in soma sizes and dendritic field sizes of Type I (A and B) and Type III (C and D) ganglion cells. Arrowheads in C and arrows in D point to long slender dendrites of Type III cells. Asterisk in D indicates the soma of a Type III ganglion cell whose dendrites are marked by arrows. Bar in D is 100pm and applies to all photographs.
Retinal
CLC (8 cases). The tube was attached to the optic nerve by vaseline whilst the animal was kept under light anaesthesia in MS222. Three days after the crystalline HRP or 1 day after the 30% HRP or CLC administration, the animals were sacrificed by an overdose of MS222. The visualization of HRP on retinal wholemounts was carried out according to previous descriptions (STRAZNICKYand STRAZNICKY, 1988). CLC-filled eyes were processed using a standard method (LAZARet al., 1983) modified to retinal wholemounts. Briefly, the neural retinae were separated from other layers of the eyeball in 0.1 M phosphate buffer (pH 7.4) at 4°C. To precipitate the cobalt the specimens were immersed in ice-cold phosphate buffer saturated with hydrogen sulfide for 15 min. After washing in the same buffer three times for 2 min periods, the retinae were fixed in graded ethanol ; for 1 h in 45%, 2 h in 70% containing 5% of glacial acetic acid, and in 80% and 90% for 1 h in each. The retinae were flat-mounted onto 4% gelatinized slides from 90% ethanol, gently rehydrated and washed in several changes of distilled water. In order to silverintensify the cobalt sulfide precipitate the wholemounts were treated in Gallyas' physical developer, pH 5.5 (GALLYAS,1979) twice for 5 min. The reaction was monitored under a low power microscope and stopped by washing the retinae in three steps of 1% acetic acid for 5 min periods. After quick dehydration the wholemounts were cleared in xylene and mounted with D.P.X. (Koch-Light). All steps from acidic fixation were carried out at room temperature. Sixty Type I and Type III ganglion cells with
Ganglion
Cells in Xenopus
laevis
89
successful HRP- or CLC-fillings were selected from the wholemounts and camera lucida drawings were made of the entire dendritic arborization. The areas of somata and of dendritic fields as well as the diameters of the main dendrites just before the first bifurcation were measured using a HIPAD digitizer pad coupled to a North Star Z80 computer. The data obtained from each group of animals were compared and further analysed with a one-way ANOVA followed by a Duncan's test.
RESULTS
Type I ganglion cells have been previously defined as cells having a large symmetrical dendritic field with a profuse dendritic arborization in the scleral sublamina of the inner plexif orm layer (STRAZNICKYand STRAZNICKY,1988). In contrast, Type III ganglion cells are those which have a very large asymmetric dendritic arborization, the sparse dendritic branching being located mostly in the vitreal sublamina of the inner plexiform layer. Type I ganglion cells are evenly distributed across the retina, while the much smaller numbers of Type III ganglion cells occupy only the central third of the retinal area. Both cell types were clearly recognizable in the retinal wholemounts in the present study (Fig. 1). In the first approach, we compared the new results obtained by CLC application with those of crystalline HRP-filling. Inspection of the wholemounts revealed substantial differences between the CLC-filled prepa-
Fig. 2. Camera lucida drawings of Type I ganglion cells filled either with crystalline HRP (A), 30% solution of HRP (B) or CLC (C). Each ganglion cell was drawn from the centre of the inferior temporal retinal quadrant. Bar =100 p m.
90
P.
TOTH
and C. STRAZNICKY
:
Fig. 3. Bar chart diagram showing the averages of soma sizes of Types I and III ganglion cells filled either with crystalline HRP, 30% solution of HRP or CLC (mean +S.D.).
Table 1. Morphometric cobaltic-lysine complex
data of Types (CLC)
I and III ganglion
cells revealed
by horseradish
peroxidaes
(HRP) and
Retinal
Fig.
4.
Bar chart
diagram
showing
Ganglion
Cells in Xenopus
the averages
ters of Types I and III ganglion cells measured Conventions are the same as for Figure 3.
Fig. Types
5.
Bar I and
chart
diagram
III ganglion
showing cells.
the averages
Conventions
are
of stem at their
laevis
dendrite first
of dendritic the same
91
diame-
bifurcation.
field
sizes
as for Figure
of 3.
92
P. TOTH and C. STRAZNICKY:
rations and their HRP-filled counterparts in at least three respects. Figure lB and D show that relatively few ganglion cells are seen in CLC preparations and most of the cells are large ganglion cells with the exception of the occasional filling of medium size ones. The second difference is that the dendritic fields of the same ganglion cell type appear to be significantly larger in CLC preparations (compare Figs. 1A, C with B, D and 2A and C). Furthermore, very fine dendritic branches can be seen only after CLC labelling (Fig. 1B, 2C). The third apparent difference is that the somata and the diameter of the primary dendrites of HRP-filled ganglion cells are much larger than those of the CLC-filled neurons (Figs. 1, 2). In the second approach we compared the above observations with those obtained from animals surviving for 24 h after the application of 30% solution of HRP. The appearance of large ganglion cells -regarding their dendritic morphology-was rather similar to that of neurons filled with CLC. Between these two experimental groups of animals there was no obvious difference in dendritic field sizes and in the visibility of the very fine dendritic branches of neurons. The moderate swelling of perikarya and dendrites and the relatively large number of filled ganglion cells were, to a certain extent, the similarities between the retinae of this group and those obtained by the conventional application of HRP. The results of further quantitative analyses carried out on camera lucida drawings of selected ganglion cells of both types are summarized in Table 1. Soma sizes : Figure 3 shows that for both Type I and Type III ganglion cells, the largest soma sizes were obtained after filling with crystalline HRP, being significantly larger than those of cells filled either with HRP solution (Type I : P < 0.01; Type III : P < 0.001) or with CLC (P < 0.001 for both cell types). The difference between the soma sizes of neurons of the latter two groups was smaller but still significant (Type I: P0.5 for Type III neurons). Retinal area measurements carried out before fixation and after the visualization of CLC and HRP indicated a 11-14% area shrinkage in the case of HRP preparations and a 15-18% of the CLC preparations.
DISCUSSION
The CLC technique (GORCSet al., 1979 ; LAzAR et al., 1983), modified to the special requirements of retinal wholemount preparations, has proven to be a powerful tool in visualizing the dendritic arborization of previously characterized large ganglion cells (STRAZNICKY and STRAZNICKY,1988) in the Xenopus retina. This technique has a number of advantages over ADAM'S (1981) HRP tracing technique, the preferred method to reveal dendritic arborization of ganglion cells in retinal wholemount preparations (DUNNMEYNELL and SHARMA, 1986; HITCHCOCK and EASTER, 1986 ; FRANK and HOLLYFIELD,1987 ; ARKIN and MILLER, 1988; STRAZNICKY and STRAZNICKY, 1988) : 1) The clarity and contrast of the dendritic arbors are on par with the best Golgi impregnations (KOLB et al., 1981) or images of ganglion cells or amacrine cells obtainable after the intracellular injection of Lucifer Yellow (SAKAGUCHI et al., 1984; VANEY,1986). 2) The entire dendritic arbor including the finest dendritic branches are visualized, which may not always be the case after administration of HRP either by injection or in crystalline form. 3) The retrograde filling of ganglion cells by the application of CLC to the cut optic nerve is very selective, revealing only the largest cell types among various classes of retinal ganglion cells. 4) CLC does not appear to cause substantial swelling of somata and dendrites of the filled ganglion cells. It is of interest to note that the long exposure to HRP by soaking the optic nerve stump in HRP solution brought about more uniform dendritic fillings than those seen in CLC-specimens. It is very likely from these results that the time for tracer molecules to be available for uptake by cut axons is an important determinant for whether the filling of the dendritic arbor is complete or not. This may be the reason why the short time exposure of the optic nerve stump to crystalline HRP failed to reveal the finer dendritic branches of ganglion cells. Similarly, the extended exposure to tracers in the other two methods may be the crucial factor in achieving ganglion cells of identified types with significantly larger dendritic fields.
Retinal
Although we could not see significant differences in dendritic field size or in the extent of dendritic filling following the immersion of the optic nerve stump either in 30% HRP- or CLC-solution, the apparent selectivity of the CLC-filling provides a greater contrast between the clear background of the tissue and the relatively few large cells filled with the black reaction product. This allows a much easier identification of the entire dendritic field of ganglion cells than in comparable HRP preparations. Previous observations (TOTH et al., 1985) have indicated that amongst the fibers of a peripheral nerve or a central pathway only the thick myelinated axons are capable of taking up cobalt in an amount sufficient for the backfilling of their somata and entire dendritic arborization. In our estimation, ganglion cells whose dendritic arborizations were revealed by CLC belong to the previously characterized large ganglion cell types, having the largest myelinated axon diameters, around 10-15,um2 in cross-sectional area (STRAZNICKYand STRAZNICKY,1988). It is also worth noting that a very considerable swelling of the soma and dendrites occurred in HRP-filled ganglion cells, in particular after 3 days survival. A similar phenomenon has been seen in the mudpuppy retina after retrograde labelling with HRP (ARKIN and MILLER, 1988). Therefore, any morphometric measurements on cell somata or dendrite diameters represent an overestimation of the real values. Our assumption that the entire dendritic arbor can only be revealed by a lengthy 24 h exposure to tracer molecules requires, of course, further electron microscopic confirmation. Because of the significant differences in dendritic field sizes revealed by the applied methods, we suggest that care should be taken when analysing morphometric data obtained by a short time exposure of cut axons to HRP. Acknowledgements. The authors gratefully Dr. G. LAZAR's donation of CLC.
acknowledge
Ganglion
Cells in Xenopus
laevis
93
DUNN-MEYNELL,A. A. and S. C. SHARMA:The visual system of the channel catfish (Ictalurus punctatus). I Retinal ganglion cell morphology. J. Comp. Neurol. 247: 32-55 (1986). FRANK,B. D. and J. G. HOLLYFIELD:Retinal ganglion cell morphology in the frog, Rana pipiens. J. Comp. Neurol. 266: 413-434 (1987). GALLYAS,F.: Light insensitive physical developers. Stain Technol. 54: 173-176 (1979). GORCS,T., M. ANTAL, E. OLAH and Gy. SZEKELY:An improved cobalt labeling technique with complex compounds. Acta Biol. Acad. Sci. Hung. 30: 79-86 (1979). HITCHcoCK,P. F. and S. S. J. EASTER, Jr.: Retinal ganglion cell in Goldfish: A quantitative classification into four morphological types, and a quantitative study of the development of one of them. J. Neurosci. 6:10371050 (1986). KOLB, H., R. NELSONand A. MARIANI: Amacrine cells, bipolar cells and ganglion cells of the cat retina, a Golgi study. Vision Res. 21: 1081-1114 (1981). LAZAR, Gy., P. TOTH, Gy. CSANK and E. KICLITER: Morphology and location of tectal projection neurons in frogs: a study with HRP and cobalt-filling. J. Comp. Neurol. 215: 108-120 (1983). NEVILLE,A. and C. STRAZNICKY:Dendritic field development of large ganglion cells in Xenopus. Neurosci. Lett. Suppl. 30: 5105 (1988). SAKAGUCHI, D. S., R. K., MURPHEY,R. K. HUNT and R. TOMPKINS:The development of retinal ganglion cells in a tetraploid strain of Xenopus laevis : A morphological study utilizing intracellular dye injection. J. Comp. Neurol. 224: 231-251 (1984). STRAZNICKY,C. and I. T. STRAZNICKY:Morphological classification of retinal ganglion cells in adult Xenopus laevis. Anat. Embryol. 178: 143-153 (1988). ToTH, P., Gy. CSANKand Gy. LAZAR: Morphology of the cells of origin of descending pathways to the spinal cord in Rana esculenta. A tracing study using cobaltic-lysine complex. J. Hirnforsch. 26: 365-383 (1985). VANEY,D. I.: Morphological identification of serotoninaccumulating neurons in the living retina. Science 233: 444-446 (1986). WASSLE,H. and R.-B. ILLING: The retinal projection to the superior colliculus in the cat : a quantitative study with HRP. J. Comp. Neurol. 190: 333-356 (1980).
REFERENCES
ADAMS, J. C.: Heavy metal intensification of DAB-based HRP reaction product. J. Histochem. Cytochem. 29: 775 (1981). ARKIN, M. S. and R. F. MILLER: Mudpuppy retinal ganglion cell morphology revealed by an HRP impregnation technique which provides Golgi-like staining. J. Comp. Neurol. 270: 185-208 (1988).
Dr. Charles Department School
STRAZNICKY of Anatomy
and
Histology
of Medicine
The Flinders University of South Bedford Park, SA 5042, Australia
Australia