- Email: [email protected]
Immediate early gene expression in axotomized and regenerating retinal ganglion cells of the adult rat
MOLECULAR BRAIN RESEARCH ELSEVIER
Molecular Brain Research 24 (1994) 43-54
Research Report
Immediate early gene expression in axotomized and regenerating retinal ganglion cells of the adult rat Grant A. Robinson * Department of Physiology (CB 7545), University of North Carolina at Chapel Hill, Chapel Hill, NC 27599-7545, USA (Accepted 7 December 1993)
Abstract
To determine if axotomy-induced immediate early gene (IEG) expression accompanies regenerative efforts in central nervous system (CNS) neurons, immunohistochemistry using antibodies to c-Jun, JunD, JunB, c-Fos, FosB and Krox-24 proteins was used to examine gene expression in identified adult rat retinal ganglion cells (RGCs) under two conditions: (1) after axotomy alone, and (2) 1 month after replacement of the optic nerve with an autologous peripheral nerve graft to allow axonal regrowth. Strong RGC c-Jun expression was induced I day, but not 3 h, after axotomy in most RGCs and was maintained in surviving cells throughout the 3-week study period. Axotomy also induced a limited number of RGCs to express Krox-24, but only transiently. c-Fos expression was also seen in a limited number of control RGCs, however, it was not induced by axotomy. Nucleolar FosB immunoreactivity in axotomized RGCs persisted 1 day after axotomy, but was subsequently lost. One month after axotomy and peripheral nerve graft placement, identified RGCs with regrown axons showed only nuclear c-Jun and nucleolar FosB expression. These findings support a role for IEG expression in the regeneration process of CNS neurons.
Key words: Retinal ganglion cell; Regeneration; Immediate early gene; Axotomy; Transcription factor; Peripheral nerve grafting 1. Introduction
I f axotomized neurons in the adult m a m m a l i a n central nervous system (CNS) are to regenerate axons and recover functionally after they have been disconnected from their target(s), they must first survive the injury. Axotomy-induced neuron death, the magnitude of which depends on age, animal species, class of neuron, and proximity of the lesion to the cell soma [27], is exemplified by recent studies [44,45] that reported a loss of 8 2 - 9 4 % of adult rat retinal ganglion cells (RGCs) 1 month after axotomy close from their cell bodies. In contrast, axotomy far from the eye resulted in a 3 0 - 4 6 % loss. However, if a peripheral nerve graft is immediately substituted for the cut optic nerve close to the eye, then some of these CNS neurons survive a~d regrow axons into the graft, suggesting trophic support by the graft [41,44]. Why the majority of R G C s fail to survive following axotomy and this grafting procedure is unknown. One possibility is that the graft
* Corresponding author. Fax: (1) (919) 966-6927. Elsevier Science B.V. SSDI 0 1 6 9 - 3 2 8 X ( 9 3 ) E 0 2 2 8 - R
only provides sufficient trophic support for a limited n u m b e r of RGCs. Alternatively, trophic support by the graft may be m o r e than sufficient to support m o r e RGCs, but few adapt to the injury. This latter suggestion raises the possibility that axotomy-induced responses of R G C s may vary. Recent reports have also demonstrated that axotomy causes the expression of some immediate early genes (IEGs), but not others, in CNS and peripheral nervous system (PNS) neurons [12,13,16,17,19,20,22,25]. Furthermore, expression of one I E G protein, c-Jun, has been shown to persist in injured neurons with connections to the PNS if regeneration is prevented, and to be down-regulated when regrowing axons are thought to reach their targets [25], suggesting a role for the jun gene in the regeneration process. Given the opportunity to examine CNS neurons under conditions that allow for axonal regrowth, the present study addressed the question of variable R G C responses to axotomy by examining I E G expression in all R G C s after axotomy and comparing it to I E G expression in R G C s with regenerated axons in a peripheral nerve graft. A partial report of these findings has a p p e a r e d previously [39].
G.A. Robinson/Molecular Brain Research 24 (1994) 43-54
44
Table 1 Experimental and control groups of retinal ganglion cells Group
Retinae
Description
Comparison groups
1 2
24 24
2 1
3
2
4
2
5
5
6 7
2 2
8
2
Tracer-labeled RGCs Tracer-labeled RGCs + intraorbital axotomy Tracer-labeled RGCs + intracranial axotomy Tracer-labeled RGCs + intraorbital axotomy + graft Intraorbital axotomy + graft + label Unlabeled RGCs Unlabeled RGCs (dark adapted) Unlabeled RGCs + intraorbital axotomy
2 2, 4 2 7, 8 6, 8 6, 7
2. M a t e r i a l s a n d m e t h o d s
2.1. Animals Adult female Long-Evans rats (range 200-260 g) were divided into eight groups (Table 1) based on surgical manipulations. Rats were kept on a 12-h light/dark cycle. For all surgical procedures, rats were anesthetized with 7% chloral hydrate (0.42 mg/g, i.p.) in 0.9% NaCI, and their body temperatures were maintainec with a heating pad.
2.2. Fluorescent tracer application The superior colliculi and dorsal lateral geniculate nuclei, main targets of RGC axons [28,29], were exposed bilaterally, their overlying pia carefully perforated, and small pieces of Gelfoam (Upjohn, Kalamazoo, MI) each containing approximately 3 /zl of either a Fluoro-Gold solution (FG, Fluorochrome, Englewood, CO) or a DiI suspension (1, l'-dioctadecyl-3,3,3',3'-tetramethylindocarbocyanine perchlorate, Molecular Probes, Junction City, OR) were applied to their surfaces. Rats were then allowed to recover for either 1 week (FG) or 1 month (Dil) to allow for complete retrograde labeling of RGCs. The two persistent tracers were used in conjunction with unlabeled control animals to control for the possiblity that one of the tracers might alter IEG expression, and for purposes of comparison with other studies on RGC survival after injury. RGCs with regrown axons in peripheral nerve grafts (see below) were identified by application of FG to the distal graft. One month after attachment of the graft to the eye, the ligated end of the graft was exposed and injected with 5/zl of FG approximately 2 cm from the eye. The FG solution was prepared by mixing 20 mg of FG in 1 ml of 0.9% NaC1 containing 0.1% Triton X-100 (Sigma, St. Louis, MO). The Dil suspension was prepared by prolonged sonication of 30 mg of DiI in 1 ml of 0.9% NaCI containing 0.1% Triton X-100.
2.3. Retinal ganglion cell axotomy The left optic nerve was transected intraorbitally using small scissors. Measured after fixation and dissection, transections were approximately 0.5 mm from the posterior pole of the eye. Injury to the blood supply of the retina was avoided by using a previously described surgical approach to the optic nerve [41]. In addition, retinal vasculature was examined by fundoscopy at the end of the
procedure after pupil dilation using topical 1% cyclopentolate hydrochloride (Alcon, Ft. Worth, TX). The left optic nerve was transected intracranially (approximately 9 mm from the orbit) in two additional animals 1 week after FG application to RGC targets in the midbrain. Retinae were processed for immunocytochemistry 1 day or 1 week after axotomy.
2. 4. Peripheral nerve grafting A portion (approximately 3 cm) of the peroneal branch of the sciatic nerve was harvested from the hindlimb [41] to serve as an autologous graft. The left optic nerve was transected intraorbitally and one end of the graft was attached to the ocular stump of the optic nerve using three 10-O monofilament sutures (Ethicon, Somerville, NJ). The remaining length of graft was placed in a channel in the skull made between the orbit and the left occipital bone. The free end of the graft was ligated. To control for possible early effects of a graft on RGC lEG expression, RGCs in two animals were retrogradely labeled from their midbrain targets with FG 1 week prior to grafting and then examined 1 week after graft placement (Group 4).
2.5. Retinal processing All animals were anesthetized with an overdose of chloral hydrate and then perfused through the heart with 0.9% NaCI followed by 4% paraformaldehyde in 0.1 M phosphate buffered saline (PBS), pH 7.4. Retinae were dissected and prepared as flattened wholemounts, with four radially oriented cuts in each retina. The superior pole of the retina was marked with a small notch. Whole-mounts were postfixed in phosphate-buffered 4% paraformaldehyde for 1 h, washed for at least 30 min in 0.1 M PBS, and mounted on gelatincoated slides in 3 parts glycerol to 1 part 0.04% p-phenylene diamine in 0.1 M sodium carbonate buffer, pH 9.0 [5]. Some whole-mounts were cut in cross section (15 p,m sections) on a cryostat after sucrose protection and flash freezing with liquid nitrogen. The contralateral retina was included in the same block to serve as a control.
2.6. Immunohistochemistry Antibodies made in rabbit against IEG proteins (c-Fos [dilution 1 : 100], made against amino acids 3-16 of human c-Fos, Santa Cruz Biotechnology, Inc., Santa Cruz, CA; FosB [1:100], made against amino acids 102-117 of murine FosB, Santa Cruz Biotechnology Inc.; Krox-24 [802/1, 1:1000], made against amino acids 15-322 of Krox-24, from Dr. R. Bravo, Squibb Institute, Princeton, N J; c-Jun [ab-1, 1:1000], made against amino acids 209-225 of v-Jun, Oncogene Science, Inc.; JunB [1 : 100], made against amino acids 210-222 of murine JunB, and JunD [1 : 100], made against amino acids 329-341 of murine JunD, Santa Cruz Biotechnology, Inc.) were used to label retinae using immunoperoxidase techniques (Veetastain ABC kit, Vector Laboratories, Inc., Burlingame, CA). With the exception of Krox-24 [24], western blot analyses and paraffin-section based immunocytochemistry were used to determine antibody specificities at their commercial sources. According to the manufacturers, these antibodies do not cross-react with other los and jun family member proteins. A mouse monoclonal antibody (RT 97) to the phosphorylated high-molecular weight neurofilament subunit (dilution 1 : 1000, from Dr. J. Wood, Sandoz Institute for Medical Research, London, UK), was also used to identify RGCs due to injury-induced changes associated with neurofilament phosphorylation [6,41]. Following a 30 min incubation in 3% normal goat serum (NGS) in 0.1 M PBS containing 0.1% Triton-X, retinae were incubated overnight at 4°C with one of the primary antibodies. Tissue was thoroughly rinsed in PBS and then incubated in biotinylated-secondary antibody in PBS
G.A. Robinson/Molecular Brain Research 24 (1994) 43-54 for 1 h at room temperature. After rinsing in PBS, retinae were incubated in buffer containing avidin-peroxidase at room temperature for 1 h. IEG immunoreactivity was developed after thorough rinsing in PBS and immersion in PBS containing diaminobenzidine (25 mg/100 ml) and hydrogen peroxide (0.01%). A rhodamine isothiocyanate-conjugated secondary antibody (Jackson ImmunoResearch Laboratories, Inc., West Grove, PA) was used after primary incubation in RT 97. Some whole-mounts were divided into quadrants to allow for immunoprocessing using more than one antibody. To determine antibody specificity, primary antisera were omitted. In addition, for c-Jun and FosB, corresponding peptides (1 ~zg/ml, from Santa Cruz Biotechnology and Oncogene Science, respectively) were included during primary incubations. Contralateral (optic nerve intact) retinae also served as within-animal controls. As a positive control for the possible failure of c-Fos, c-Jun and Krox-24 antisera to recognize lEG proteins, phorbol 12-myristate 13 acetate (PMA, Sigma) was used to induce IEGs [34] in the retina. In anesthetized rats (n = 2), PMA (5 ~1 of a either a 10 ng/ml or 100 n g / m l solution in dimethyl sulphoxide) was injected into the left vitreal compartment. Only vehicle was injected into the right vitreal compartment. After 3 h of maintained anesthesia, the rats were perfused and the retinae processed for immunocytochemistry. Heat-induced FosB, JunD and JunB immunoreactivity in spinal neurons [15] was also used to ensure the antisera recognized these proteins. Control (basal) expression of each lEG was assessed using retinae with intact optic nerves (Groups 1, 6 and 7) with and without overnight dark adaptation.
45
2. 7. Measurement of retinal ganglion cell densities Neurons labeled with FG or DiI were viewed through a fluorescence microscope using UV and rhodamine filters respectively (FG: excitation filter 355-425 nm, supression filter 460 nm; DiI: excitation filter 530-560 nm, suppression filter 580 nm). Tracer-labeled neurons within a 0.14x0.14 mm area were counted 1, 2, and 3 mm from the optic disk in each retinal quadrant [45] at a magnification of 600x. Sample counts were compared and confirmed by a second independent observer. The number of tracer-labeled RGCs and other profiles (see below) from these 12 regions was used to calculate the average profile or RGC density/ram 2 for each retina. Only RGCs containing punctate FG or DiI were counted. In cases where whole-mounts were divided, averages were derived from 3 regions/ quadrant.
2.8. Analysis of gene product immunoreacticity Immunoreactivity to IEG protiens was used as an indicator of gene expression and compared both within and between RGC groups (Table 1): (1) Pre-axotomy (basal) expression was determined for each IEG (Groups 1, 6 and 7); (2) Expression was examined at 3 h, 1 day, and 1, 2 and 3 weeks after intraorbital axotomy (Groups 2 and 8), and 1 and 7 days after intracranial axotomy (Group 3); (3) Profile densities of FG-positive RGCs (immunonegative), double-labeled RGCs (fluorescent+immunopositive) and immunopositive retinal nuclei of unknown origin (fluorescence negative) were determined in whole-mounts using methods described above (Groups 1, 2 and 3); 4) Profile analyses were also performed on retinae containing RGCs with regrown axons in peripheral nerve grafts 1 month after graft placement (Group 5) and on retinae with pre-labeled RGCs with grafts attached for 1 week (Group 4). Two methods were used to determine the retinal distribution of IEG expression. The first used radial sections to determine the location of lEG expression within the multi-layered retina. The second method used the location of RGCs within a whole-mount (i.e., superior, inferior, nasal, temporal). Both DiI (n = 5) and FG-labeled RGCs were examined for basal and axotomy-induced IEG expression, but only FG-labeled RGCs were used for quantitation of axotomy-induced gene expression. RGCs labeled with FG retained the tracer after immunoprocessing, but retinae labeled with DiI required initial location-specific photography and, after immunoprocessing, a return to the same location for photography of immunoreactivity due to the labile nature of DiI in the presence of detergent.
2. 9. Statistical analysis Groups of RGC density data were compared using the unpaired two-group t-test. Differences were considered statistically significant when P < 0.05.
2.10. Technical considerations
Fig. 1. Basal immediate early gene expression in retinal cross sections. A: c-Jun-positive nuclei (arrow) were limited to the ganglion cell layer (GCL). B: in contrast, c-Fos-positive nuclei (arrows) were found in the inner nuclear layer (INL) as well as in the GCL. Bar = 100/~m.
Consideration of some of the technical elements of these experiments is required for interpretation of this work. Several control groups were necessary for interpretation of the effects of axotomy on lEG expression. Possible tracer-effects were controlled for by comparing IEG expression in unlabeled RGCs with expression in DiIand FG-labeled cells. Similar results were obtained under all three circumstances, suggesting that these persistent dyes did not alter IEG expression. Antibody specificity controls (use of contralateral retinae [optic nerve intact], omission of primary antibodies, preincubation with excess antigenic peptides, stimulation of IEG products with PMA in the retina, and heat-induced lEG expression in spinal
46
G.A. Robinson/Molecular Brain Research 24 (1994) 43-54
H
G.A. Robinson/Molecular Brain Research 24 (1994) 43-54
47
Fig. 3. RGC double-labeling studies using FG and immunoperoxidase staining of immediate early gene proteins in retinal wholemounts. A,B: fluorescence (A) and transmitted light (B) views of FG labeling and c-Fos immunoreactivity, c-Fos expression (arrows) was seen in a small number of retrogradely labeled RGCs in control (optic nerve intact) retinae. C,D: fluorescence (C) and transmitted light (D) views of FG labeling and FosB immunoreactivity. FosB immunoreactivity was limited to RGC nucleoli (arrows) in control retinae. E,F: fluorescence (E) and transmitted light (F) views of FG labeling and c-Jun immunoreactivity 1 week after intracranial axotomy. Only weak c-Jun immunoreactivity (arrows) was seen in labeled RGCs. Bar = 50/~m.
Fig. 2. Axotomy-induced c-Jun expression in RGCs. A,B: fluorescence (A) and transmitted light (B) views of FG labeling and c-Jun immunoreactivity in the ganglion cell layer of a retinal whole-mount, c-Jun expression (arrows) in control (optic nerve intact) retinae was limited to RGCs with maximal soma diameters < 15 /zm. C,D: strong c-Jun expression was seen 24 h after intraorbital axotomy almost exclusively in RGCs. However, a small number of c-Jun-positive, FG-negative profiles were seen after axotomy (arrows), suggesting incomplete retinal labeling from the midbrain. E,F: 1 week after axotomy, nuclear c-Jun staining was still evident, but the number of surviving RGCs was severely reduced. G,H: 2 weeks after axotomy, the number of c-Jun-positive RGCs was reduced further, yet RGCs with punctate FG within their somata were immunopositive. Degenerating profiles with diffuse cytoplasmic FG were c-Jun-negative. I,J: 3 weeks after axotomy, very few RGCs remained, yet they were immunopositive for c-Jun. Bar = 50/xm.
48
G.A. Robinson/Molecular Brain Research 24 (1994) 43-54
cord) supported both the immunopositive (c-Jun, Krox-24, c-Fos, FosB) and immunonegative findings (JunB, JunD). To control for recent findings on the effects of light on retinal IEG expression [8,40] basal expression in identified RGCs was also determined under conditions of dark adaptation. Basal lEG expression was very similar when light- and dark-adapted RGCs were compared, suggesting little influence of this light history on expression of these lEGs in this cell type. To control for the possibility of lEG induction in regenerating RGCs due to axotomy during the FG-labeling process of the graft, axons of prelabeled RGCs were cut intracranially and lEG expression was examined 1 day and 1 week later. Since the distance from the site of injury to the cell body was much greater for regenerating RGCs than for those injured intracranially, lack of lEG expression after intracranial axotomy argues against any effects of graft labelling on IEG expression in regenerating RGCs.
3. Results
more than 85% of RGCs have a soma diameter less than 15/zm [10], this suggests that c-Jun expression is more likely to be encountered in small RGCs by a factor of 14. c-Fos immunoreactivity was sparse and restricted to the ganglion cell and inner nuclear layers (Fig. 1B), confirming a similar retinal distribution for c-Fos described in the rabbit [40]. Double-labeling studies revealed that some of these c-Fos-positive nuclei were within RGCs (Fig. 3A), but the majority were not (Fig. 4D). FosB immunoreactivity was confined exclusively to R G C nucleoli (Fig. 3C,D). These results conflict with a recent report [17] on l E G expression in the rat, where c-Jun, FosB and c-Fos expression were not observed in control retinae. No basal immunostaining for Krox-24, JunB or JunD proteins was observed in the retina (data not shown).
3.1. Basal expression of immediate early genes in the retina In control retinae (optic nerve intact), weak basal expression of c-Jun was apparent in many neurons confined to the ganglion cell layer (Fig. 1A). Doublelabeling studies confirmed that basal c-Jun immunoreactivity was localized to RGCs; however, labeling was not uniformly distributed among RGCs. c-Jun expression was mostly limited to 10% of sampled cells with soma diameters less than 15 /~m (560 randomly sampled cells from two retinae). In contrast to this basal c-Jun expression in these smallest cells, approximately 96% of RGCs with soma diameters larger than 15/zm (240 randomly-sampled cells from the same retinae) were immunonegative for c-Jun (Fig. 2A). Even though
A 2500 "1
,ooot® 1500 t ~
e2222"~ ~
Control
B
1
7
S.D.
14
21
oI LLLL
100"]
[ ] c-Jun+labeled RGCs • c-Jun-only nuclei [ ] Label-only RGCs
6O
(3Fig. 4. Axotomy-induced RGC death and immediate early gene expression. A: RGC survival after intraorbital axotomy. RGC density in control retinae was significantly reduced 1 week after axotomy. Each time point is represented by at least two retinae. S.D. = standard deviation. Asterisks denote differences at the P < 0.05 level between axotomized and control groups. B: normalized data for axotomy-induced c-Jun expression in the ganglion cell layer. Control, 1, 2, 3 and 4 week values were based on 3567, 971, 503, 112 and 47 samples, respectively. More than 90% of the profiles were FG-positive and c-Jun-positive 1 day after injury compared with control (optic nerve intact) retinae. This pattern continued for the duration of the study period, despite axotomy-induced RGC death. C: normalized data for axotomy-induced Krox-24 expression. Control, 1, 2, 3 and 4 week values were based on 1642, 583, 280, 61 and 50 samples, respectively. Krox-24 was also induced in RGCs by 1 day after axotomy, but the proportion of immunopositive cells was small and at 1 week they were no longer seen. The relatively constant number of Krox-24-only nuclei appears to increase with time due to axotomy-induced RGC death. D: normalized data for axotomy-induced c-Fos expression. Control, 1, 2, 3 and 4 week values were based on 492, 644, 137, 64 and 29 samples respectively. Unlike c-Jun and Krox-24, the proportion of control c-Fos-immunopositive RGCs remained low and constant 1 day after axotomy. Similar to Krox-24, c-Fos-immunopositive RGCs were no longer seen 1 week after
injury.
4O 2O 0 Control
1
7
14
21
C
0-
0O00'l j) J J
[ ] Krox+labeled RGCs • Krox-only nuclei [ ] Label-only RGCs
o Control
1
7
14
21
D 1003 [ ] c-Fos+labeled RGCs • c-Fos-only nuclei • Label-only RGCs 4O 20 0 Control
1
7 14 Days After Axotomy
21
G.A. Robinson/Molecular Brain Research 24 (1994) 43-54 Tissue processed without primary antisera were immunonegative. Basal immunoreactivity in R G C s was unaltered by the presence or absence of fluorescent tracers (DiI or F G ) and after dark adaptation. Retinae treated with P M A showed similar patterns of immunoreactivity within the inner nuclear and ganglion cell layers when processed with c-Jun, c-Fos and Krox24 antisera. Control localization of JunB, J u n D and FosB was similar to that previously described [15] in the dorsal horn of stimulated rat spinal cord. 3.2. Axotomy-induced immediate early gene expression Strong c-Jun immunoreactivity was induced 1 day, but not 3 h, after injury in the ganglion cell layer exclusively. Double-labeling localized the vast majority of c-Jun-positive nuclei to R G C s throughout the 3 week study period (Fig. 2). Despite diminishing R G C numbers due to axotomy (Fig. 4A), quantitative analyses revealed similar normalized proportions of R G C s showing strong c-Jun immunoreactivity throughout the
49
3 week study period (Fig. 4B). In control retinae, retrogradely labeled R G C s showed only punctate F G fluorescence, suggesting a compartmentalization of the tracer while these neurons were alive. Some nuclei of these R G C s were also c-Jun-positive. After axotomy, the tracer was also punctate in c-Jun-positive R G C s , but a p p e a r e d diffuse in R G C s that were c-Jun-negative, supporting a relationship between R G C survival, c-Jun immunoreactivity and the microscopic appearance of the tracer. Unlike after intraorbital axotomy, c-Jun expression 1 day or 1 week after intracranial axotomy was similar to the basal expression observed in unaxotomized controls (Fig. 3E,F), suggesting that the induction stimulus was not sufficient to induce gene expression with the same time course as did injury closer to the cell body. This increased R G C c-Jun expression after axotomy is in a g r e e m e n t with a recent report [17] of retinal c-Jun expression in the ganglion cell layer of the rat, however, c-Jun-positive nuclei were distributed across the entire retina and not limited to the central portion.
Fig. 5. RGC expression of Krox-24 after axotomy and during regeneration. A,B: fluorescence (A) and transmitted light (B) views of double-labeled RGCs 24 h after intraorbital axotomy. Only a small number of FG-labeled RGCs show nuclear Krox-24 immunoreactivity (arrows). C,D: fluorescence (C) and transmitted light (D) views of Krox-24 immunoreactivity combined with RT 97 immunoreactivity. Krox-24 expression was still seen 1 month after axotomy and peripheral nerve graft placement (arrows), but these immunopositive nuclei were not localized to RGCs using the antibody to phosphorylated, 200 kDa neurofilaments. Bar = 50/xm.
50
G.A. Robinson/Molecular Brain Research 24 (1994) 43-54
Krox-24 i m m u n o r e a c t i v i t y was also i n d u c e d 1 day, b u t n o t 3 h, a f t e r a x o t o m y a n d was l i m i t e d to t h e g a n g l i o n cell layer; h o w e v e r , t h e p a t t e r n was d i f f e r e n t f r o m t h a t s e e n for c-Jun. F e w Krox-24-positive n u c l e i w e r e l o c a l i z e d to R G C s (Figs. 4B a n d 5A, B), with
e q u a l n u m b e r s o f i m m u n o p o s i t i v e ceils b e i n g F G - n e g a rive, suggesting t h a t t h e s e cells m a y b e d i s p l a c e d a m a c r i n e cells also k n o w n to o c c u p y this layer [35]. O n e w e e k after axotomy, Krox-24-positive R G C s w e r e n o t seen, but n o n - R G C Krox-24 l a b e l i n g p e r s i s t e d .
~
~
~
S¸ ?
~
~
Fig. 6. Immediate early gene expression in regenerating RGCs. A,B: fluorescence (A) and transmitted light (B) views of a double-labeled RGC, retrogradely labeled with FG from the peripheral nerve graft 1 month after graft placement, showing strong c-Jun immunoreactivity (arrows). C,D: fluorescence (C) and transmitted light (D) views of FG labeling and Krox-24 immunoreactivity. Krox-24 immunoreactivity was not seen in regenerating RGCs, but non-RGC Krox-24-positive profiles were observed throughout the grafting period (arrows). E,F: fluorescence (E) and transmitted light (F) views of FG labeling and FosB immunoreactivity. Regenerating RGCs also showed nucleolar FosB immunoreactivity (arrows). Bar = 50/xm.
G.A. Robinson/Molecular Brain Research 24 (1994) 43-54 This pattern of non-RGC labeling in the ganglion cell layer persisted throughout the 3 week study period (Fig. 4C). This increase in Krox-24 expression after axotomy is similar to that described for the small number of presumed RGCs by Herdegen et al. [17]; however, the persistence of Krox-24 expression in non-RGCs demonstrated here is contrary to their findings. Although c-Fos expression in RGCs was not induced by axotomy, it was altered by the injury. Similar small numbers of double-labeled RGCs were seen before and 1 day after injury, but they were not seen 1 week after injury or for the duration of the study period (Fig. 4D). Similar to the case for Krox-24, c-Fos-positive profiles within non-RGCs also persisted throughout the study period (Fig. 4D). These observations on c-Fos expression differ from those by Herdegen et al. [17], where no expression was detected before or after RGC axotomy. FosB immunoreactivity in RGC nucleoli persisted 1 day after axotomy, but it was not seen 1 week after injury or at later time points. Nucleoli of RGCs stained with Cresyl Violet were evident prior to injury, but were not discernable in axotomized RGCs beginning 1 week after injury. This is in contrast to Herdegen et al. [17], where FosB was not detected before or after RGC axotomy. JunB and JunD expression in RGCs was not changed by axotomy. Neither axotomy-induced nor basal expression of IEG proteins were dependent on the location of RGCs within the retina. For c-Jun, Krox-24, c-Fos and FosB, immunopositive cells were distributed at all eccentricities and in each quadrant 1 day after axotomy. The effects of axotomy were also independent of the choice (DiI or FG) or presence of tracer.
3.3. Immediate early gene expression in regenerating retinal ganglion cells One week after replacement of the optic nerve with a peripheral nerve graft (Group 4), only c-Jun immunoreactivity was seen in RGCs. One month after graft placement (Group 5), analysis of RGCs with regrown axons in the grafts showed only nucleolar FosB (92% of 112 sampled cells from 3 retinae) and nuclear c-Jun (94% of 252 sampled cells from 5 retinae) immunoreactivity (Fig. 6). Although FG-labeled RGCs were negative for Krox-24 and c-Fos, nuclei showing Krox-24 and c-Fos immunoreactivities attributed to non-projecting neurons were present in the ganglion cell layer with a pattern similar to that seen after axotomy alone (Fig. 6C,D). In support of these nuclei not belonging to RGCs, neurofilament immunoreactivity using the RT 97 antibody [6,41] revealed that these c-Fos- and Krox-24-positive nuclei did not co-localize with RGCs (Fig. 5C,D).
51
4. Discussion
This work used immunohistochemistry to demonstrate patterns of IEG expression in identified RGCs primarily under two conditions: (1) after axotomy of previously uninjured neurons and (2) after axotomy and placement of an autologous peripheral nerve graft to allow axonal regrowth. In previously uninjured neurons, strong c-Jun expression was induced 1 day, but not 3 h, after axotomy and was maintained only in surviving RGCs throughout the 3 week study period. Axotomy induced a limited number of RGCs to transiently express Krox-24 (Egr-1; Zif/268; NGFI-A), while RGC c-Fos expression was not induced by the injury. Nucleolar FosB immunoreactivity in axotomized RGCs persisted 1 day after axotomy, but was subsequently lost. One month after axotomy and peripheral nerve graft placement, RGCs with regrown axons showed only nuclear c-Jun and nucleolar FosB expression, but non-RGC c-Fos and Krox-24 immunoreactivities were still present in the ganglion cell layer.
4.1. Immediate early gene expression after injury and during regeneration The proposition that IEG products have regulatory functions is supported by recent experiments that have demonstrated that these products are transcription factors. As examples of their regulatory capacity, both c-Jun and c-Fos transcription factors not only can interact with each other at their leucine-zipper sites as mixed-species pairs [11,23,31,37], but these heterodimers can also bind to specific DNA regulatory (AP-1) sites and stimulate transcription at a nearby promoter region (reviewed in ref. 3). Thus, these and other members of the jun, los and krox [26] product families appear to function in the coupling of the short-term signals to long-term changes in cell phenotype via target gene regulation. These axotomy-induced changes in CNS gene expression complement similar findings for c-Jun expression in RGCs as well as in other injured neuronal systems, and differ from the rapid and transient responses associated with these genes after other induction stimuli [14,15] and contribute the novel observation that FosB expression in identified RGCs is nucleolar and is seen only briefly after axotomy alone. Using retrograde labeling techniques, the present work confirms the long-term induction of c-Jun specifically in RGCs that was suggested by studies in the rabbit [22] and rat [17] after optic nerve crush. The use of retrograde labeling has also explained discrepancies in the retinal localization of some IEG proteins between the present report and that of Herdegen et al. [17]; however, reported differences in immunocytochemical detection of IEG proteins may be the result of a combi-
52
G.A. Robinson/Molecular Brain Research 24 (1994) 43-54
nation of different antibody specificites, different methods of axotomy (crush vs. cut) and different rat species. Axotomized dorsal root ganglion (DRG) cells and motor neurons have also been shown to express both protein and mRNA for c-Jun and JunD with a similar protracted time course, but not for c-Fos, FosB or Krox-24 [16,19,25]. Similar induction of c-Jun and JunD expression in axotomized neurons of the autonomic nervous system has also been described, but JunD was induced for only 5 days after injury [13]. A role for c-Jun during regeneration in PNS neurons has been suggested by work demonstrating its enhanced expression, persistence and subsequent return to basal levels when axonal regrowth was considered complete [25]. The expression of c-Jun in non-regenerating CNS neurons after axotomy alone has led to similar speculation [20,22]. This work's principal finding of maintained c-Jun expression in regenerating RGCs supports a role for c-Jun in the general strategy for axonal regrowth after injury in CNS neurons. Roles for other IEGs after axotomy and during regeneration are less clear, however. Unlike PNS neurons, where injury caused a transient expression of JunD [13], axotomized RGCs were not observed to express this lEG protein. Lack of JunD immunoreactivity in acute (Group 4) and chronic regenerating RGCs (Group 5) also suggests that the role of JunD in regenerating CNS neurons may be different from PNS neurons. Similarly, although expression of c-Fos and Krox-24 has not been associated with axotomy in PNS neurons and was not observed in regenerating RGCs, Krox-24 was transiently induced by the injury in some RGCs and c-Fos expression was maintained in some RGCs for a short time after the injury. Given the small percentage of RGCs that survive and regrow axons into the nerve graft, it is possible that, in addition to c-Jun, the transient c-Fos and Krox-24 expression seen in some RGCs may be necessary for successful regeneration after axotomy. Although the methods used in the present experiments could not associate regenerating RGCs with prior c-Fos a n d / o r Krox-24 expression, the possibility cannot be overlooked. Alternatively, lEG proteins that are expressed after injury may not be present in sufficient amounts to maintain RGC viability. Such differences in gene expression among RGCs and expression differences between CNS and PNS neurons might also contribute to the poor survival of injured RGCs even in the presence of a peripheral nerve graft. Recent studies in the rat [44,45] on the survival of RGCs persistently labeled with DiI have reported that approximately 75% of the population dies by the second week after axotomy < 1 mm from the eye. The present results complement and extend these reports by demonstrating the loss of approximately 52% of FG-labeled RGCs at an earlier time point (1 week)
after injury using the same quantitative methods. Immediate placement of a peripheral nerve graft onto the eye reduces this RGC death [44], suggesting atrophic role for the graft, possibly by diffusible factors that have been shown to be necessary for growth of RGC axons into the grafts [2]. Although placement of a graft was used in the present experiments only to provide an environment through which RGCs could regenerate axons, the possibility that it altered IEG expression was also examined. With the exception of FosB, comparisons between IEG expression early (1 week) or later (1 month) in the regeneration process and IEG expression after axotomy alone showed that graft placement had no effect on expression of the examined proteins. Given that similar strong c-Jun immunoreactivity was observed in regenerating RGCs and in surviving RGCs at all post-axotomy time points, an effect of the graft on this protein was unlikely. For FosB, expression was transient after axotomy alone and was not seen 1 week after the injury. After graft placement for 1 month, nucleolar FosB was again evident, suggesting the possiblity that it was influenced by the graft. The localization of FosB to the nucleolus suggests a site-specific function for this transcription factor in RGCs. Alternative splicing of fosB mRNA has recently been shown to give rise to two protein products of different lengths [4,30,32,48]. Both the long and short forms are capable of dimerizing with other fos and jun family protein members and binding to appropriate DNA recognition sites; however, their roles as transactivators have been described both as complementary [4] and antagonistic [30,32,48] depending on the in vitro assay. Similarly, comparisons of the transforming potential of the two forms have yielded disparate results [30,33,47,48] based on different cellular assays. The FosB antibody used in the present report could not distinguish between these two forms. Given this uncertainty and the assay-dependent nature of functional studies, many interpretations are possible. Either form of FosB could act as a regulator of nucleolar function. Consistent with the metabolic reorganization for regeneration seen in injured PNS neurons, axotomy-induced loss of FosB immunoreactivity by 1 week after injury may signal a transition point indicating a more active nucleolar state, since the loss of FosB coincides with the first nucleolar changes (pronounced swelling, vacuolization, dispersion of chromatin, loosening of dense fibrillar material) associated with regeneration in the adult rodent PNS [21]. These changes are thought to accompany the increased demand for protein synthesis, and therefore ribosomal RNA, brought on by injury. In agreement with the FosB immunocytochemistry, Cresyl Violet staining combined with FG labeling confirmed that RGC nucleoli were no longer visible 1 week after intraorbital axotomy. Alternatively, the possibility remains that loss of FosB im-
G.A. Robinson/Molecular Brain Research 24 (1994) 43-54
munoreactivity can be accounted for by physical nucleolar changes induced by axotomy. Interestingly, these injury-induced nucleolar changes reported in PNS neurons [21] were not reported in rat RGCs after intracranial axotomy [1]. Given the paradox of poor RGC regenerative potential, yet their increased viability, after intracranial axotomy versus intraorbital axotomy [38,45], the possibility cannot be excluded that axotomy very close to the cell body may induce nucleolar changes similar to those documented during regeneration in the PNS. The demonstration of axotomy-induced Krox-24 and c-Fos immunoreactivity in unlabeled nuclei within the ganglion cell layer suggests either expression in RGCs that were not retrogradely labeled or transneuronal induction in non-RGCs. Bilateral application of tracer to targets in the midbrain and thalamus argues against incomplete retrograde labeling, given that the vast majority of, if not all, RGCs project to the superior colliculus in the rat [7,28] and that the density of labeled RGCs in control retinae (Fig. 4A) agrees with other published reports [42,45]. In addition, neurofilament labeling with RT 97 showed that RGCs were Krox-24- and c-Fos-immunonegative, suggesting that the immunopositive nuclei belonged to other retinal cells. Both retrograde and anterograde transneuronal effects are also possible due to lost projections from axotomized centrifugal projections as well as from RGCs to amacrine cells [18,43,46]. Although amacrine cells in the ganglion cell layer are the most likely candidates to be influenced by RGC axotomy in the rat retina [9,35,36], at present the identity of these Krox-24and c-Fos-immunopositive cells are not known.
Acknowledgements My thanks to Bonnie Blake and Helen Willcockson for excellent technical assistance, to Drs. Rodrigo Bravo and John Wood for the antisera to Krox-24 and RT 97, respectively, and to Drs. Alan Light and Lee Mcllwain for critical reading of the manuscript. This work was supported by American Paralysis Association Grant RB1-9201-1 and the Leopold Schepp Foundation.
References [1] Barron, K.D., McGuiness, C.M., Misantone, L.J., Zanakis, M.F., Grafstein, B. and Murray, M., RNA content of normal and axotomized retinal ganglion cells of rat and goldfish, J. Comp. Neurol., 236 (1985) 265-273. [2] Cho, E.Y.P. and So, K.-F., De novo formation of axon-like processes from axotomized retinal ganglion cells which exhibit long distance growth in a peripheral nerve graft in adult hamsters, Brain Res., 484 (1989) 371-377.
53
[3] Curran, T. and Franza, B.R. Jr., Fos and Jun: the AP-1 connection, Cell, 55 (1988) 395-397. [4] Dobrzanski, P., Noguchi, T., Kovary, K., Rizzo, C.A., Lazo, P.S. and Bravo, R., Both products of the fosB gene, FosB and its short form, FosB/SF, are transcriptional activators in fibroblasts, Mol. Cell. Biol., 11 (1991) 5470-5478. [5] Dodd, J., Solter, D. and Jessel, T.M., Monoclonal antibodies against carbohydrate differentiation antigens identify subsets of primary sensory neurons, Nature, 311 (1984) 469-472. [6] Dr~iger, U. and Hofbauer, A., Antibodies to heavy neurofilament subunit detect a subpopulation of damaged ganglion cells in retina, Nature, 309 (1984) 624-626. [7] Dreher, B., Sefton, A.J., Ni, S.Y.K. and Nisben, G., The morphology, number, distribution and central projections of Class i retinal ganglion cells in albino and hooded rats, Brain Behav. EvoL, 26 (1985) 10-48. [8] Earnest, D.J., Iadarola, M., Yeh, H.H. and Olschowka, J.A., Photic regulation of c-fos expression in neural components governing the entrainment of circadian rhythms, Exp. Neurol., 109 (1990) 353-361. [9] Familietti, E.V., ON and OFF pathways through amacrine cells in mammalian retina: the synaptic connections of "starburst" amacrine cells, Vision Res., 23 (1983) 1265-1279. [10] Fukuda, Y., A three group classification of rat retinal ganglion cells: histological and physiological studies, Brain Res., 119 (1977) 327-344. [11] Halazonetis, T.D., Georgopoulos, K., Greenberg, M.E. and Leder, P., c-Jun dimerizes with itself and c-Fos, forming complexes of different DNA binding affinities, Cell, 55 (1988) 917924. [12] Haas, C.A., Donath, C. and Kreutzberg, G.W., Differential expression of immediate early genes after transection of the facial nerve, Neuroscience, 53 (1993) 91-99. [13] Herdegen, T., Kummer, W., Fiallos, C.E., Leah, J. and Bravo, R., Expression of c-Jun, Jun B and Jun D proteins in the rat nervous system following transection of the vagus nerve and cervical sympathetic trunk, Neuroscience, 45 (1991) 413-422. [14] Herdegen, T., Leah, J.D., Manisali, A., Bravo, R. and Zimmermann, M., c-JUN-like immunoreactivity in the CNS of the adult rat: basal and transynaptically induced expression of an immediate early gene, Neuroscience, 41 (1991) 643-654. [15] Herdegen, T., T611e, T., Bravo, R., Zieglg~insberger, W. and Zimmermann, M., Sequential expression of JUN B, JUN D and FOS B proteins in rat spinal neurons: cascade of transcriptional operations during nociception, Neurosci. Lett., 129 (1991) 221224. [16] Herdegen, T., Fiallos-Estrada, C.E., Schmid, W., Bravo, R. and Zimmermann, M., The transcription factors c-JUN, JUN D and CREB, but not FOS and KROX-24, are differentially regulated in axotomized neurons following transection of the rat sciatic nerve, Mol. Brain. Res., 14 (1992) 155-165. [17] Herdegen, T., Bastmeyer, M., B~ihr, M., Stuermer, C., Bravo, R. and Zimmermann, M., Expression of JUN, KROX, and CREB transcription factors in goldfish and rat retinal ganglion cells following optic nerve lesion is related to axonal sprouting, J. Neurobiol., 24 (1993) 528-543. [18] Itaya, S.K. and Itaya, P.W., Centrifugal fibers to the rat retina from the meaial pretectal area and the periaqueductal grey matter, Brain Res., 326 (1985) 362-365. [19] Jenkins, R. and Hunt,S.P., Long-term increases in the levels of c-iun mRNA and Jun protein-like immunoreactivity in motor and sensory neurons following axon damage, Neurosci. Lett., 129 (1991) 107-110. [20] Jenkins, R., Tetzlaff, W. and Hunt, S.P., Differential expression of immediate early genes in rubrospinal neurons following axotomy in rat, Eur. J. Neurosci., 5 (1993) 203-209. [21] Jones, K.J. and La Velle, A., Differential effects of axotomy on
54
G.A. Robinson/Molecular Brain Research 24 (1994) 43-54
immature and mature hamster facial neurons: a time course study of initial nucleolar and nuclear changes, J. Neurocytol., 15 (1986) 197-206. [22] Koistinaho, J., Hicks, K.J. and Sagar, S.M., Long-term induction of c-jun mRNA and Jun protein in rabbit retinal ganglion cells following axotomy or colchicine treatment, J. Neurosci. Res., 34 (1993) 250-255. [23] Kouzarides, T. and Ziff, E., The role of the leucine zipper in the fos-jun interaction, Nature, 336 (1988) 646-651, [24] Kovary, K. and Bravo, R., Expression of different JUN and FOS during GO to G1 transition in mouse fibroblasts: in vitro and in vivo associations, Mol. Cell. Biol., 11 (1991) 2451-2459. [25} Leah, J.D., Herdegen, T. and Bravo, R., Selective expression of Jun proteins following axotomy and axonal transport block in peripheral nerves in the rat, Brain Res., 566 (1991) 198-207. [26] Lemaire, P., Vesque, C., Schmin, J., Stunnenberg, H., Frank, R. and Charnay, P., The serum-inducible mouse gene Krox 24 encodes a sequence-specific transcriptional activator, Mol. Cell. Biol., 10 (1990) 3456-3467. [27] Lieberman, A.R., Some factors affecting retrograde neuronal responses to axonal lesions. In R. Bellairs and E.G. Gray (Eds.), Essays of the Nervous System, Claredon, Oxford, 1974, pp. 71-105. [28] Linden, R. and Perry, V.H., Massive retinotectal projection in rats, Brain Res., 272 (1983) 145-149. [29] Martin, P.R., The projection of different retinal ganglion cell classes to the dorsal lateral geniculate nucleus in the hooded rat, Exp. Brain Res., 62 (1986) 77-88. [30] Mumberg, D., Lucibello, M., Schuermann, M. and Muller, R., Alternative splicing of fosB transcripts results in differentially expressed mRNAs encoding functionally antagonistic proteins, Genes Deu., 5 (1991) 1212-1223. [31] Nakabeppu, Y., Ryder, K. and Nathans, D., DNA binding activities of three murine Jun proteins; stimulation by Fos, Cell, 55 (1988) 907-915. [32] Nakabeppu, Y. and Nathans, D., A naturally occurring form of FosB that inhibits Fos/Jun transcriptional activity, Cell, 64 (1991) 751-759. [33] Nakabeppu, Y., Oda, S. and Sekiguchi, M., Proliferative activation of quiescent Rat-lA cells by DFosB, Mol. Cell. Biol., 13 . (1993) 4157-4166. [34] Papavassiliou, A.G., Treier, M., Chavrier, C. and Bohmann, D., Targeted degradation of c-Fos, but not v-Fos, by a phospborylation-dependent signal on c-Jun, Science, 258 (1992) 1941-1944. [35] Perry, V.H., Evidence for an amacrine cell system in the ganglion cell layer of the rat retina, Neuroscience, 6 (1981) 931-944.
[36] Perry, V.H. and Walker, M., Arnacrine cells, displaced amacrine cells and interplexiform cells in the retina of the rat, Proc. R. Soc. B, 208 (1980) 415-431. [37] Rauscher, F.J., Voulalas, P.J., Franza, B.R. Jr. and Curran, T., Fos and Jun bind cooperatively to the AP-1 site: reconstitution in vitro, Genes Deu., 2 (1988) 1687-1699. [38] Richardson, P.M., Issa, V.M.K. and Shemie, S,, Regeneration and retrograde degeneration of axons in the rat optic nerve, J. Neurocytol., 11 (1982) 949-966. [39] Robinson, G.A. and Light, A.R., Axotomy-induced expression of immediate early genes in rat retinal ganglion cells, Soc. Neurosci. Abstr., 18 (1992) 1312. [40] Sagar, S.M. and Sharp, F.R., Light induces a Fos-like nuclear antigen in retinal neurons, Mol. Brain Res., 7 (1990) 17-21. [41] Vidal-Sanz, M., Bray, G.M., Villegas-P6rez, M.P., Thanos, S. and Aguayo, A.J., Axonal regeneration and synapse formation in the superior colliculus by retinal ganglion cells in the adult rat, Z Neurosci., 7 (1987) 2894-2909. [42] Vidal-Sanz, M., Villegas-P6rez, M.P., Bray, G.M. and Aguayo, A.J., Persistent retrograde labelling of adult rat retinal ganglion cells with the carbocyanine dye, diI, Exp. Neurol., 102 (1988) 92-101. [43] Villar, M.J., Vitale, M.L. and Parisi, M.N., Dorsal raphe serotonergic projection to the retina. A combined peroxidase tracing-neurochemical/high performance liquid chromatography study in the rat, Neuroscience, 22 (1987) 681-686. [44] Villegas-P6rez, M.P., Vidal-Sanz, M., Bray, G.M. and Aguayo, A.J., Influences of peripheral nerve grafts on the survival and regrowth of axotomized retinal ganglion cells in adult rats, J. Neurosci., 8 (1988) 265-280. [45] Villegas-P6rez, M.P., Vidal-Sanz, M., Bray, G.M. and Aguayo, A.J., Rapid and protracted phases of retinal ganglion cell loss follow axotomy in the optic nerve of adult rats, J. Neurobiol., 24 (1993) 23-36. [46] Weiniawa-Narkiewicz, E. and Hughes, A., The superficial plexiform layer: a third retinal association area, J. Comp. Neurol., 324 (1992) 463-484. [47] Wisdom, R., Yen, J., Rashid, D. and Verma, M., Transformation by Fos B requires a trans-activation domain missing in FosB2 that can be substituted by heterologous activation domains, Genes Deu., 6 (1992) 667-675. [48] Yen, J., Wisdom, R.M., Tratner, I. and Verma, I.M., An alternative spliced form of FosB is a negative regulator of transcriptional activation and transformation by Fos proteins, Proc. Natl. Acad. Sci. USA, 88 (1991) 5077-5081.
Molecular Brain Research 24 (1994) 43-54
Research Report
Immediate early gene expression in axotomized and regenerating retinal ganglion cells of the adult rat Grant A. Robinson * Department of Physiology (CB 7545), University of North Carolina at Chapel Hill, Chapel Hill, NC 27599-7545, USA (Accepted 7 December 1993)
Abstract
To determine if axotomy-induced immediate early gene (IEG) expression accompanies regenerative efforts in central nervous system (CNS) neurons, immunohistochemistry using antibodies to c-Jun, JunD, JunB, c-Fos, FosB and Krox-24 proteins was used to examine gene expression in identified adult rat retinal ganglion cells (RGCs) under two conditions: (1) after axotomy alone, and (2) 1 month after replacement of the optic nerve with an autologous peripheral nerve graft to allow axonal regrowth. Strong RGC c-Jun expression was induced I day, but not 3 h, after axotomy in most RGCs and was maintained in surviving cells throughout the 3-week study period. Axotomy also induced a limited number of RGCs to express Krox-24, but only transiently. c-Fos expression was also seen in a limited number of control RGCs, however, it was not induced by axotomy. Nucleolar FosB immunoreactivity in axotomized RGCs persisted 1 day after axotomy, but was subsequently lost. One month after axotomy and peripheral nerve graft placement, identified RGCs with regrown axons showed only nuclear c-Jun and nucleolar FosB expression. These findings support a role for IEG expression in the regeneration process of CNS neurons.
Key words: Retinal ganglion cell; Regeneration; Immediate early gene; Axotomy; Transcription factor; Peripheral nerve grafting 1. Introduction
I f axotomized neurons in the adult m a m m a l i a n central nervous system (CNS) are to regenerate axons and recover functionally after they have been disconnected from their target(s), they must first survive the injury. Axotomy-induced neuron death, the magnitude of which depends on age, animal species, class of neuron, and proximity of the lesion to the cell soma [27], is exemplified by recent studies [44,45] that reported a loss of 8 2 - 9 4 % of adult rat retinal ganglion cells (RGCs) 1 month after axotomy close from their cell bodies. In contrast, axotomy far from the eye resulted in a 3 0 - 4 6 % loss. However, if a peripheral nerve graft is immediately substituted for the cut optic nerve close to the eye, then some of these CNS neurons survive a~d regrow axons into the graft, suggesting trophic support by the graft [41,44]. Why the majority of R G C s fail to survive following axotomy and this grafting procedure is unknown. One possibility is that the graft
* Corresponding author. Fax: (1) (919) 966-6927. Elsevier Science B.V. SSDI 0 1 6 9 - 3 2 8 X ( 9 3 ) E 0 2 2 8 - R
only provides sufficient trophic support for a limited n u m b e r of RGCs. Alternatively, trophic support by the graft may be m o r e than sufficient to support m o r e RGCs, but few adapt to the injury. This latter suggestion raises the possibility that axotomy-induced responses of R G C s may vary. Recent reports have also demonstrated that axotomy causes the expression of some immediate early genes (IEGs), but not others, in CNS and peripheral nervous system (PNS) neurons [12,13,16,17,19,20,22,25]. Furthermore, expression of one I E G protein, c-Jun, has been shown to persist in injured neurons with connections to the PNS if regeneration is prevented, and to be down-regulated when regrowing axons are thought to reach their targets [25], suggesting a role for the jun gene in the regeneration process. Given the opportunity to examine CNS neurons under conditions that allow for axonal regrowth, the present study addressed the question of variable R G C responses to axotomy by examining I E G expression in all R G C s after axotomy and comparing it to I E G expression in R G C s with regenerated axons in a peripheral nerve graft. A partial report of these findings has a p p e a r e d previously [39].
G.A. Robinson/Molecular Brain Research 24 (1994) 43-54
44
Table 1 Experimental and control groups of retinal ganglion cells Group
Retinae
Description
Comparison groups
1 2
24 24
2 1
3
2
4
2
5
5
6 7
2 2
8
2
Tracer-labeled RGCs Tracer-labeled RGCs + intraorbital axotomy Tracer-labeled RGCs + intracranial axotomy Tracer-labeled RGCs + intraorbital axotomy + graft Intraorbital axotomy + graft + label Unlabeled RGCs Unlabeled RGCs (dark adapted) Unlabeled RGCs + intraorbital axotomy
2 2, 4 2 7, 8 6, 8 6, 7
2. M a t e r i a l s a n d m e t h o d s
2.1. Animals Adult female Long-Evans rats (range 200-260 g) were divided into eight groups (Table 1) based on surgical manipulations. Rats were kept on a 12-h light/dark cycle. For all surgical procedures, rats were anesthetized with 7% chloral hydrate (0.42 mg/g, i.p.) in 0.9% NaCI, and their body temperatures were maintainec with a heating pad.
2.2. Fluorescent tracer application The superior colliculi and dorsal lateral geniculate nuclei, main targets of RGC axons [28,29], were exposed bilaterally, their overlying pia carefully perforated, and small pieces of Gelfoam (Upjohn, Kalamazoo, MI) each containing approximately 3 /zl of either a Fluoro-Gold solution (FG, Fluorochrome, Englewood, CO) or a DiI suspension (1, l'-dioctadecyl-3,3,3',3'-tetramethylindocarbocyanine perchlorate, Molecular Probes, Junction City, OR) were applied to their surfaces. Rats were then allowed to recover for either 1 week (FG) or 1 month (Dil) to allow for complete retrograde labeling of RGCs. The two persistent tracers were used in conjunction with unlabeled control animals to control for the possiblity that one of the tracers might alter IEG expression, and for purposes of comparison with other studies on RGC survival after injury. RGCs with regrown axons in peripheral nerve grafts (see below) were identified by application of FG to the distal graft. One month after attachment of the graft to the eye, the ligated end of the graft was exposed and injected with 5/zl of FG approximately 2 cm from the eye. The FG solution was prepared by mixing 20 mg of FG in 1 ml of 0.9% NaC1 containing 0.1% Triton X-100 (Sigma, St. Louis, MO). The Dil suspension was prepared by prolonged sonication of 30 mg of DiI in 1 ml of 0.9% NaCI containing 0.1% Triton X-100.
2.3. Retinal ganglion cell axotomy The left optic nerve was transected intraorbitally using small scissors. Measured after fixation and dissection, transections were approximately 0.5 mm from the posterior pole of the eye. Injury to the blood supply of the retina was avoided by using a previously described surgical approach to the optic nerve [41]. In addition, retinal vasculature was examined by fundoscopy at the end of the
procedure after pupil dilation using topical 1% cyclopentolate hydrochloride (Alcon, Ft. Worth, TX). The left optic nerve was transected intracranially (approximately 9 mm from the orbit) in two additional animals 1 week after FG application to RGC targets in the midbrain. Retinae were processed for immunocytochemistry 1 day or 1 week after axotomy.
2. 4. Peripheral nerve grafting A portion (approximately 3 cm) of the peroneal branch of the sciatic nerve was harvested from the hindlimb [41] to serve as an autologous graft. The left optic nerve was transected intraorbitally and one end of the graft was attached to the ocular stump of the optic nerve using three 10-O monofilament sutures (Ethicon, Somerville, NJ). The remaining length of graft was placed in a channel in the skull made between the orbit and the left occipital bone. The free end of the graft was ligated. To control for possible early effects of a graft on RGC lEG expression, RGCs in two animals were retrogradely labeled from their midbrain targets with FG 1 week prior to grafting and then examined 1 week after graft placement (Group 4).
2.5. Retinal processing All animals were anesthetized with an overdose of chloral hydrate and then perfused through the heart with 0.9% NaCI followed by 4% paraformaldehyde in 0.1 M phosphate buffered saline (PBS), pH 7.4. Retinae were dissected and prepared as flattened wholemounts, with four radially oriented cuts in each retina. The superior pole of the retina was marked with a small notch. Whole-mounts were postfixed in phosphate-buffered 4% paraformaldehyde for 1 h, washed for at least 30 min in 0.1 M PBS, and mounted on gelatincoated slides in 3 parts glycerol to 1 part 0.04% p-phenylene diamine in 0.1 M sodium carbonate buffer, pH 9.0 [5]. Some whole-mounts were cut in cross section (15 p,m sections) on a cryostat after sucrose protection and flash freezing with liquid nitrogen. The contralateral retina was included in the same block to serve as a control.
2.6. Immunohistochemistry Antibodies made in rabbit against IEG proteins (c-Fos [dilution 1 : 100], made against amino acids 3-16 of human c-Fos, Santa Cruz Biotechnology, Inc., Santa Cruz, CA; FosB [1:100], made against amino acids 102-117 of murine FosB, Santa Cruz Biotechnology Inc.; Krox-24 [802/1, 1:1000], made against amino acids 15-322 of Krox-24, from Dr. R. Bravo, Squibb Institute, Princeton, N J; c-Jun [ab-1, 1:1000], made against amino acids 209-225 of v-Jun, Oncogene Science, Inc.; JunB [1 : 100], made against amino acids 210-222 of murine JunB, and JunD [1 : 100], made against amino acids 329-341 of murine JunD, Santa Cruz Biotechnology, Inc.) were used to label retinae using immunoperoxidase techniques (Veetastain ABC kit, Vector Laboratories, Inc., Burlingame, CA). With the exception of Krox-24 [24], western blot analyses and paraffin-section based immunocytochemistry were used to determine antibody specificities at their commercial sources. According to the manufacturers, these antibodies do not cross-react with other los and jun family member proteins. A mouse monoclonal antibody (RT 97) to the phosphorylated high-molecular weight neurofilament subunit (dilution 1 : 1000, from Dr. J. Wood, Sandoz Institute for Medical Research, London, UK), was also used to identify RGCs due to injury-induced changes associated with neurofilament phosphorylation [6,41]. Following a 30 min incubation in 3% normal goat serum (NGS) in 0.1 M PBS containing 0.1% Triton-X, retinae were incubated overnight at 4°C with one of the primary antibodies. Tissue was thoroughly rinsed in PBS and then incubated in biotinylated-secondary antibody in PBS
G.A. Robinson/Molecular Brain Research 24 (1994) 43-54 for 1 h at room temperature. After rinsing in PBS, retinae were incubated in buffer containing avidin-peroxidase at room temperature for 1 h. IEG immunoreactivity was developed after thorough rinsing in PBS and immersion in PBS containing diaminobenzidine (25 mg/100 ml) and hydrogen peroxide (0.01%). A rhodamine isothiocyanate-conjugated secondary antibody (Jackson ImmunoResearch Laboratories, Inc., West Grove, PA) was used after primary incubation in RT 97. Some whole-mounts were divided into quadrants to allow for immunoprocessing using more than one antibody. To determine antibody specificity, primary antisera were omitted. In addition, for c-Jun and FosB, corresponding peptides (1 ~zg/ml, from Santa Cruz Biotechnology and Oncogene Science, respectively) were included during primary incubations. Contralateral (optic nerve intact) retinae also served as within-animal controls. As a positive control for the possible failure of c-Fos, c-Jun and Krox-24 antisera to recognize lEG proteins, phorbol 12-myristate 13 acetate (PMA, Sigma) was used to induce IEGs [34] in the retina. In anesthetized rats (n = 2), PMA (5 ~1 of a either a 10 ng/ml or 100 n g / m l solution in dimethyl sulphoxide) was injected into the left vitreal compartment. Only vehicle was injected into the right vitreal compartment. After 3 h of maintained anesthesia, the rats were perfused and the retinae processed for immunocytochemistry. Heat-induced FosB, JunD and JunB immunoreactivity in spinal neurons [15] was also used to ensure the antisera recognized these proteins. Control (basal) expression of each lEG was assessed using retinae with intact optic nerves (Groups 1, 6 and 7) with and without overnight dark adaptation.
45
2. 7. Measurement of retinal ganglion cell densities Neurons labeled with FG or DiI were viewed through a fluorescence microscope using UV and rhodamine filters respectively (FG: excitation filter 355-425 nm, supression filter 460 nm; DiI: excitation filter 530-560 nm, suppression filter 580 nm). Tracer-labeled neurons within a 0.14x0.14 mm area were counted 1, 2, and 3 mm from the optic disk in each retinal quadrant [45] at a magnification of 600x. Sample counts were compared and confirmed by a second independent observer. The number of tracer-labeled RGCs and other profiles (see below) from these 12 regions was used to calculate the average profile or RGC density/ram 2 for each retina. Only RGCs containing punctate FG or DiI were counted. In cases where whole-mounts were divided, averages were derived from 3 regions/ quadrant.
2.8. Analysis of gene product immunoreacticity Immunoreactivity to IEG protiens was used as an indicator of gene expression and compared both within and between RGC groups (Table 1): (1) Pre-axotomy (basal) expression was determined for each IEG (Groups 1, 6 and 7); (2) Expression was examined at 3 h, 1 day, and 1, 2 and 3 weeks after intraorbital axotomy (Groups 2 and 8), and 1 and 7 days after intracranial axotomy (Group 3); (3) Profile densities of FG-positive RGCs (immunonegative), double-labeled RGCs (fluorescent+immunopositive) and immunopositive retinal nuclei of unknown origin (fluorescence negative) were determined in whole-mounts using methods described above (Groups 1, 2 and 3); 4) Profile analyses were also performed on retinae containing RGCs with regrown axons in peripheral nerve grafts 1 month after graft placement (Group 5) and on retinae with pre-labeled RGCs with grafts attached for 1 week (Group 4). Two methods were used to determine the retinal distribution of IEG expression. The first used radial sections to determine the location of lEG expression within the multi-layered retina. The second method used the location of RGCs within a whole-mount (i.e., superior, inferior, nasal, temporal). Both DiI (n = 5) and FG-labeled RGCs were examined for basal and axotomy-induced IEG expression, but only FG-labeled RGCs were used for quantitation of axotomy-induced gene expression. RGCs labeled with FG retained the tracer after immunoprocessing, but retinae labeled with DiI required initial location-specific photography and, after immunoprocessing, a return to the same location for photography of immunoreactivity due to the labile nature of DiI in the presence of detergent.
2. 9. Statistical analysis Groups of RGC density data were compared using the unpaired two-group t-test. Differences were considered statistically significant when P < 0.05.
2.10. Technical considerations
Fig. 1. Basal immediate early gene expression in retinal cross sections. A: c-Jun-positive nuclei (arrow) were limited to the ganglion cell layer (GCL). B: in contrast, c-Fos-positive nuclei (arrows) were found in the inner nuclear layer (INL) as well as in the GCL. Bar = 100/~m.
Consideration of some of the technical elements of these experiments is required for interpretation of this work. Several control groups were necessary for interpretation of the effects of axotomy on lEG expression. Possible tracer-effects were controlled for by comparing IEG expression in unlabeled RGCs with expression in DiIand FG-labeled cells. Similar results were obtained under all three circumstances, suggesting that these persistent dyes did not alter IEG expression. Antibody specificity controls (use of contralateral retinae [optic nerve intact], omission of primary antibodies, preincubation with excess antigenic peptides, stimulation of IEG products with PMA in the retina, and heat-induced lEG expression in spinal
46
G.A. Robinson/Molecular Brain Research 24 (1994) 43-54
H
G.A. Robinson/Molecular Brain Research 24 (1994) 43-54
47
Fig. 3. RGC double-labeling studies using FG and immunoperoxidase staining of immediate early gene proteins in retinal wholemounts. A,B: fluorescence (A) and transmitted light (B) views of FG labeling and c-Fos immunoreactivity, c-Fos expression (arrows) was seen in a small number of retrogradely labeled RGCs in control (optic nerve intact) retinae. C,D: fluorescence (C) and transmitted light (D) views of FG labeling and FosB immunoreactivity. FosB immunoreactivity was limited to RGC nucleoli (arrows) in control retinae. E,F: fluorescence (E) and transmitted light (F) views of FG labeling and c-Jun immunoreactivity 1 week after intracranial axotomy. Only weak c-Jun immunoreactivity (arrows) was seen in labeled RGCs. Bar = 50/~m.
Fig. 2. Axotomy-induced c-Jun expression in RGCs. A,B: fluorescence (A) and transmitted light (B) views of FG labeling and c-Jun immunoreactivity in the ganglion cell layer of a retinal whole-mount, c-Jun expression (arrows) in control (optic nerve intact) retinae was limited to RGCs with maximal soma diameters < 15 /zm. C,D: strong c-Jun expression was seen 24 h after intraorbital axotomy almost exclusively in RGCs. However, a small number of c-Jun-positive, FG-negative profiles were seen after axotomy (arrows), suggesting incomplete retinal labeling from the midbrain. E,F: 1 week after axotomy, nuclear c-Jun staining was still evident, but the number of surviving RGCs was severely reduced. G,H: 2 weeks after axotomy, the number of c-Jun-positive RGCs was reduced further, yet RGCs with punctate FG within their somata were immunopositive. Degenerating profiles with diffuse cytoplasmic FG were c-Jun-negative. I,J: 3 weeks after axotomy, very few RGCs remained, yet they were immunopositive for c-Jun. Bar = 50/xm.
48
G.A. Robinson/Molecular Brain Research 24 (1994) 43-54
cord) supported both the immunopositive (c-Jun, Krox-24, c-Fos, FosB) and immunonegative findings (JunB, JunD). To control for recent findings on the effects of light on retinal IEG expression [8,40] basal expression in identified RGCs was also determined under conditions of dark adaptation. Basal lEG expression was very similar when light- and dark-adapted RGCs were compared, suggesting little influence of this light history on expression of these lEGs in this cell type. To control for the possibility of lEG induction in regenerating RGCs due to axotomy during the FG-labeling process of the graft, axons of prelabeled RGCs were cut intracranially and lEG expression was examined 1 day and 1 week later. Since the distance from the site of injury to the cell body was much greater for regenerating RGCs than for those injured intracranially, lack of lEG expression after intracranial axotomy argues against any effects of graft labelling on IEG expression in regenerating RGCs.
3. Results
more than 85% of RGCs have a soma diameter less than 15/zm [10], this suggests that c-Jun expression is more likely to be encountered in small RGCs by a factor of 14. c-Fos immunoreactivity was sparse and restricted to the ganglion cell and inner nuclear layers (Fig. 1B), confirming a similar retinal distribution for c-Fos described in the rabbit [40]. Double-labeling studies revealed that some of these c-Fos-positive nuclei were within RGCs (Fig. 3A), but the majority were not (Fig. 4D). FosB immunoreactivity was confined exclusively to R G C nucleoli (Fig. 3C,D). These results conflict with a recent report [17] on l E G expression in the rat, where c-Jun, FosB and c-Fos expression were not observed in control retinae. No basal immunostaining for Krox-24, JunB or JunD proteins was observed in the retina (data not shown).
3.1. Basal expression of immediate early genes in the retina In control retinae (optic nerve intact), weak basal expression of c-Jun was apparent in many neurons confined to the ganglion cell layer (Fig. 1A). Doublelabeling studies confirmed that basal c-Jun immunoreactivity was localized to RGCs; however, labeling was not uniformly distributed among RGCs. c-Jun expression was mostly limited to 10% of sampled cells with soma diameters less than 15 /~m (560 randomly sampled cells from two retinae). In contrast to this basal c-Jun expression in these smallest cells, approximately 96% of RGCs with soma diameters larger than 15/zm (240 randomly-sampled cells from the same retinae) were immunonegative for c-Jun (Fig. 2A). Even though
A 2500 "1
,ooot® 1500 t ~
e2222"~ ~
Control
B
1
7
S.D.
14
21
oI LLLL
100"]
[ ] c-Jun+labeled RGCs • c-Jun-only nuclei [ ] Label-only RGCs
6O
(3Fig. 4. Axotomy-induced RGC death and immediate early gene expression. A: RGC survival after intraorbital axotomy. RGC density in control retinae was significantly reduced 1 week after axotomy. Each time point is represented by at least two retinae. S.D. = standard deviation. Asterisks denote differences at the P < 0.05 level between axotomized and control groups. B: normalized data for axotomy-induced c-Jun expression in the ganglion cell layer. Control, 1, 2, 3 and 4 week values were based on 3567, 971, 503, 112 and 47 samples, respectively. More than 90% of the profiles were FG-positive and c-Jun-positive 1 day after injury compared with control (optic nerve intact) retinae. This pattern continued for the duration of the study period, despite axotomy-induced RGC death. C: normalized data for axotomy-induced Krox-24 expression. Control, 1, 2, 3 and 4 week values were based on 1642, 583, 280, 61 and 50 samples, respectively. Krox-24 was also induced in RGCs by 1 day after axotomy, but the proportion of immunopositive cells was small and at 1 week they were no longer seen. The relatively constant number of Krox-24-only nuclei appears to increase with time due to axotomy-induced RGC death. D: normalized data for axotomy-induced c-Fos expression. Control, 1, 2, 3 and 4 week values were based on 492, 644, 137, 64 and 29 samples respectively. Unlike c-Jun and Krox-24, the proportion of control c-Fos-immunopositive RGCs remained low and constant 1 day after axotomy. Similar to Krox-24, c-Fos-immunopositive RGCs were no longer seen 1 week after
injury.
4O 2O 0 Control
1
7
14
21
C
0-
0O00'l j) J J
[ ] Krox+labeled RGCs • Krox-only nuclei [ ] Label-only RGCs
o Control
1
7
14
21
D 1003 [ ] c-Fos+labeled RGCs • c-Fos-only nuclei • Label-only RGCs 4O 20 0 Control
1
7 14 Days After Axotomy
21
G.A. Robinson/Molecular Brain Research 24 (1994) 43-54 Tissue processed without primary antisera were immunonegative. Basal immunoreactivity in R G C s was unaltered by the presence or absence of fluorescent tracers (DiI or F G ) and after dark adaptation. Retinae treated with P M A showed similar patterns of immunoreactivity within the inner nuclear and ganglion cell layers when processed with c-Jun, c-Fos and Krox24 antisera. Control localization of JunB, J u n D and FosB was similar to that previously described [15] in the dorsal horn of stimulated rat spinal cord. 3.2. Axotomy-induced immediate early gene expression Strong c-Jun immunoreactivity was induced 1 day, but not 3 h, after injury in the ganglion cell layer exclusively. Double-labeling localized the vast majority of c-Jun-positive nuclei to R G C s throughout the 3 week study period (Fig. 2). Despite diminishing R G C numbers due to axotomy (Fig. 4A), quantitative analyses revealed similar normalized proportions of R G C s showing strong c-Jun immunoreactivity throughout the
49
3 week study period (Fig. 4B). In control retinae, retrogradely labeled R G C s showed only punctate F G fluorescence, suggesting a compartmentalization of the tracer while these neurons were alive. Some nuclei of these R G C s were also c-Jun-positive. After axotomy, the tracer was also punctate in c-Jun-positive R G C s , but a p p e a r e d diffuse in R G C s that were c-Jun-negative, supporting a relationship between R G C survival, c-Jun immunoreactivity and the microscopic appearance of the tracer. Unlike after intraorbital axotomy, c-Jun expression 1 day or 1 week after intracranial axotomy was similar to the basal expression observed in unaxotomized controls (Fig. 3E,F), suggesting that the induction stimulus was not sufficient to induce gene expression with the same time course as did injury closer to the cell body. This increased R G C c-Jun expression after axotomy is in a g r e e m e n t with a recent report [17] of retinal c-Jun expression in the ganglion cell layer of the rat, however, c-Jun-positive nuclei were distributed across the entire retina and not limited to the central portion.
Fig. 5. RGC expression of Krox-24 after axotomy and during regeneration. A,B: fluorescence (A) and transmitted light (B) views of double-labeled RGCs 24 h after intraorbital axotomy. Only a small number of FG-labeled RGCs show nuclear Krox-24 immunoreactivity (arrows). C,D: fluorescence (C) and transmitted light (D) views of Krox-24 immunoreactivity combined with RT 97 immunoreactivity. Krox-24 expression was still seen 1 month after axotomy and peripheral nerve graft placement (arrows), but these immunopositive nuclei were not localized to RGCs using the antibody to phosphorylated, 200 kDa neurofilaments. Bar = 50/xm.
50
G.A. Robinson/Molecular Brain Research 24 (1994) 43-54
Krox-24 i m m u n o r e a c t i v i t y was also i n d u c e d 1 day, b u t n o t 3 h, a f t e r a x o t o m y a n d was l i m i t e d to t h e g a n g l i o n cell layer; h o w e v e r , t h e p a t t e r n was d i f f e r e n t f r o m t h a t s e e n for c-Jun. F e w Krox-24-positive n u c l e i w e r e l o c a l i z e d to R G C s (Figs. 4B a n d 5A, B), with
e q u a l n u m b e r s o f i m m u n o p o s i t i v e ceils b e i n g F G - n e g a rive, suggesting t h a t t h e s e cells m a y b e d i s p l a c e d a m a c r i n e cells also k n o w n to o c c u p y this layer [35]. O n e w e e k after axotomy, Krox-24-positive R G C s w e r e n o t seen, but n o n - R G C Krox-24 l a b e l i n g p e r s i s t e d .
~
~
~
S¸ ?
~
~
Fig. 6. Immediate early gene expression in regenerating RGCs. A,B: fluorescence (A) and transmitted light (B) views of a double-labeled RGC, retrogradely labeled with FG from the peripheral nerve graft 1 month after graft placement, showing strong c-Jun immunoreactivity (arrows). C,D: fluorescence (C) and transmitted light (D) views of FG labeling and Krox-24 immunoreactivity. Krox-24 immunoreactivity was not seen in regenerating RGCs, but non-RGC Krox-24-positive profiles were observed throughout the grafting period (arrows). E,F: fluorescence (E) and transmitted light (F) views of FG labeling and FosB immunoreactivity. Regenerating RGCs also showed nucleolar FosB immunoreactivity (arrows). Bar = 50/xm.
G.A. Robinson/Molecular Brain Research 24 (1994) 43-54 This pattern of non-RGC labeling in the ganglion cell layer persisted throughout the 3 week study period (Fig. 4C). This increase in Krox-24 expression after axotomy is similar to that described for the small number of presumed RGCs by Herdegen et al. [17]; however, the persistence of Krox-24 expression in non-RGCs demonstrated here is contrary to their findings. Although c-Fos expression in RGCs was not induced by axotomy, it was altered by the injury. Similar small numbers of double-labeled RGCs were seen before and 1 day after injury, but they were not seen 1 week after injury or for the duration of the study period (Fig. 4D). Similar to the case for Krox-24, c-Fos-positive profiles within non-RGCs also persisted throughout the study period (Fig. 4D). These observations on c-Fos expression differ from those by Herdegen et al. [17], where no expression was detected before or after RGC axotomy. FosB immunoreactivity in RGC nucleoli persisted 1 day after axotomy, but it was not seen 1 week after injury or at later time points. Nucleoli of RGCs stained with Cresyl Violet were evident prior to injury, but were not discernable in axotomized RGCs beginning 1 week after injury. This is in contrast to Herdegen et al. [17], where FosB was not detected before or after RGC axotomy. JunB and JunD expression in RGCs was not changed by axotomy. Neither axotomy-induced nor basal expression of IEG proteins were dependent on the location of RGCs within the retina. For c-Jun, Krox-24, c-Fos and FosB, immunopositive cells were distributed at all eccentricities and in each quadrant 1 day after axotomy. The effects of axotomy were also independent of the choice (DiI or FG) or presence of tracer.
3.3. Immediate early gene expression in regenerating retinal ganglion cells One week after replacement of the optic nerve with a peripheral nerve graft (Group 4), only c-Jun immunoreactivity was seen in RGCs. One month after graft placement (Group 5), analysis of RGCs with regrown axons in the grafts showed only nucleolar FosB (92% of 112 sampled cells from 3 retinae) and nuclear c-Jun (94% of 252 sampled cells from 5 retinae) immunoreactivity (Fig. 6). Although FG-labeled RGCs were negative for Krox-24 and c-Fos, nuclei showing Krox-24 and c-Fos immunoreactivities attributed to non-projecting neurons were present in the ganglion cell layer with a pattern similar to that seen after axotomy alone (Fig. 6C,D). In support of these nuclei not belonging to RGCs, neurofilament immunoreactivity using the RT 97 antibody [6,41] revealed that these c-Fos- and Krox-24-positive nuclei did not co-localize with RGCs (Fig. 5C,D).
51
4. Discussion
This work used immunohistochemistry to demonstrate patterns of IEG expression in identified RGCs primarily under two conditions: (1) after axotomy of previously uninjured neurons and (2) after axotomy and placement of an autologous peripheral nerve graft to allow axonal regrowth. In previously uninjured neurons, strong c-Jun expression was induced 1 day, but not 3 h, after axotomy and was maintained only in surviving RGCs throughout the 3 week study period. Axotomy induced a limited number of RGCs to transiently express Krox-24 (Egr-1; Zif/268; NGFI-A), while RGC c-Fos expression was not induced by the injury. Nucleolar FosB immunoreactivity in axotomized RGCs persisted 1 day after axotomy, but was subsequently lost. One month after axotomy and peripheral nerve graft placement, RGCs with regrown axons showed only nuclear c-Jun and nucleolar FosB expression, but non-RGC c-Fos and Krox-24 immunoreactivities were still present in the ganglion cell layer.
4.1. Immediate early gene expression after injury and during regeneration The proposition that IEG products have regulatory functions is supported by recent experiments that have demonstrated that these products are transcription factors. As examples of their regulatory capacity, both c-Jun and c-Fos transcription factors not only can interact with each other at their leucine-zipper sites as mixed-species pairs [11,23,31,37], but these heterodimers can also bind to specific DNA regulatory (AP-1) sites and stimulate transcription at a nearby promoter region (reviewed in ref. 3). Thus, these and other members of the jun, los and krox [26] product families appear to function in the coupling of the short-term signals to long-term changes in cell phenotype via target gene regulation. These axotomy-induced changes in CNS gene expression complement similar findings for c-Jun expression in RGCs as well as in other injured neuronal systems, and differ from the rapid and transient responses associated with these genes after other induction stimuli [14,15] and contribute the novel observation that FosB expression in identified RGCs is nucleolar and is seen only briefly after axotomy alone. Using retrograde labeling techniques, the present work confirms the long-term induction of c-Jun specifically in RGCs that was suggested by studies in the rabbit [22] and rat [17] after optic nerve crush. The use of retrograde labeling has also explained discrepancies in the retinal localization of some IEG proteins between the present report and that of Herdegen et al. [17]; however, reported differences in immunocytochemical detection of IEG proteins may be the result of a combi-
52
G.A. Robinson/Molecular Brain Research 24 (1994) 43-54
nation of different antibody specificites, different methods of axotomy (crush vs. cut) and different rat species. Axotomized dorsal root ganglion (DRG) cells and motor neurons have also been shown to express both protein and mRNA for c-Jun and JunD with a similar protracted time course, but not for c-Fos, FosB or Krox-24 [16,19,25]. Similar induction of c-Jun and JunD expression in axotomized neurons of the autonomic nervous system has also been described, but JunD was induced for only 5 days after injury [13]. A role for c-Jun during regeneration in PNS neurons has been suggested by work demonstrating its enhanced expression, persistence and subsequent return to basal levels when axonal regrowth was considered complete [25]. The expression of c-Jun in non-regenerating CNS neurons after axotomy alone has led to similar speculation [20,22]. This work's principal finding of maintained c-Jun expression in regenerating RGCs supports a role for c-Jun in the general strategy for axonal regrowth after injury in CNS neurons. Roles for other IEGs after axotomy and during regeneration are less clear, however. Unlike PNS neurons, where injury caused a transient expression of JunD [13], axotomized RGCs were not observed to express this lEG protein. Lack of JunD immunoreactivity in acute (Group 4) and chronic regenerating RGCs (Group 5) also suggests that the role of JunD in regenerating CNS neurons may be different from PNS neurons. Similarly, although expression of c-Fos and Krox-24 has not been associated with axotomy in PNS neurons and was not observed in regenerating RGCs, Krox-24 was transiently induced by the injury in some RGCs and c-Fos expression was maintained in some RGCs for a short time after the injury. Given the small percentage of RGCs that survive and regrow axons into the nerve graft, it is possible that, in addition to c-Jun, the transient c-Fos and Krox-24 expression seen in some RGCs may be necessary for successful regeneration after axotomy. Although the methods used in the present experiments could not associate regenerating RGCs with prior c-Fos a n d / o r Krox-24 expression, the possibility cannot be overlooked. Alternatively, lEG proteins that are expressed after injury may not be present in sufficient amounts to maintain RGC viability. Such differences in gene expression among RGCs and expression differences between CNS and PNS neurons might also contribute to the poor survival of injured RGCs even in the presence of a peripheral nerve graft. Recent studies in the rat [44,45] on the survival of RGCs persistently labeled with DiI have reported that approximately 75% of the population dies by the second week after axotomy < 1 mm from the eye. The present results complement and extend these reports by demonstrating the loss of approximately 52% of FG-labeled RGCs at an earlier time point (1 week)
after injury using the same quantitative methods. Immediate placement of a peripheral nerve graft onto the eye reduces this RGC death [44], suggesting atrophic role for the graft, possibly by diffusible factors that have been shown to be necessary for growth of RGC axons into the grafts [2]. Although placement of a graft was used in the present experiments only to provide an environment through which RGCs could regenerate axons, the possibility that it altered IEG expression was also examined. With the exception of FosB, comparisons between IEG expression early (1 week) or later (1 month) in the regeneration process and IEG expression after axotomy alone showed that graft placement had no effect on expression of the examined proteins. Given that similar strong c-Jun immunoreactivity was observed in regenerating RGCs and in surviving RGCs at all post-axotomy time points, an effect of the graft on this protein was unlikely. For FosB, expression was transient after axotomy alone and was not seen 1 week after the injury. After graft placement for 1 month, nucleolar FosB was again evident, suggesting the possiblity that it was influenced by the graft. The localization of FosB to the nucleolus suggests a site-specific function for this transcription factor in RGCs. Alternative splicing of fosB mRNA has recently been shown to give rise to two protein products of different lengths [4,30,32,48]. Both the long and short forms are capable of dimerizing with other fos and jun family protein members and binding to appropriate DNA recognition sites; however, their roles as transactivators have been described both as complementary [4] and antagonistic [30,32,48] depending on the in vitro assay. Similarly, comparisons of the transforming potential of the two forms have yielded disparate results [30,33,47,48] based on different cellular assays. The FosB antibody used in the present report could not distinguish between these two forms. Given this uncertainty and the assay-dependent nature of functional studies, many interpretations are possible. Either form of FosB could act as a regulator of nucleolar function. Consistent with the metabolic reorganization for regeneration seen in injured PNS neurons, axotomy-induced loss of FosB immunoreactivity by 1 week after injury may signal a transition point indicating a more active nucleolar state, since the loss of FosB coincides with the first nucleolar changes (pronounced swelling, vacuolization, dispersion of chromatin, loosening of dense fibrillar material) associated with regeneration in the adult rodent PNS [21]. These changes are thought to accompany the increased demand for protein synthesis, and therefore ribosomal RNA, brought on by injury. In agreement with the FosB immunocytochemistry, Cresyl Violet staining combined with FG labeling confirmed that RGC nucleoli were no longer visible 1 week after intraorbital axotomy. Alternatively, the possibility remains that loss of FosB im-
G.A. Robinson/Molecular Brain Research 24 (1994) 43-54
munoreactivity can be accounted for by physical nucleolar changes induced by axotomy. Interestingly, these injury-induced nucleolar changes reported in PNS neurons [21] were not reported in rat RGCs after intracranial axotomy [1]. Given the paradox of poor RGC regenerative potential, yet their increased viability, after intracranial axotomy versus intraorbital axotomy [38,45], the possibility cannot be excluded that axotomy very close to the cell body may induce nucleolar changes similar to those documented during regeneration in the PNS. The demonstration of axotomy-induced Krox-24 and c-Fos immunoreactivity in unlabeled nuclei within the ganglion cell layer suggests either expression in RGCs that were not retrogradely labeled or transneuronal induction in non-RGCs. Bilateral application of tracer to targets in the midbrain and thalamus argues against incomplete retrograde labeling, given that the vast majority of, if not all, RGCs project to the superior colliculus in the rat [7,28] and that the density of labeled RGCs in control retinae (Fig. 4A) agrees with other published reports [42,45]. In addition, neurofilament labeling with RT 97 showed that RGCs were Krox-24- and c-Fos-immunonegative, suggesting that the immunopositive nuclei belonged to other retinal cells. Both retrograde and anterograde transneuronal effects are also possible due to lost projections from axotomized centrifugal projections as well as from RGCs to amacrine cells [18,43,46]. Although amacrine cells in the ganglion cell layer are the most likely candidates to be influenced by RGC axotomy in the rat retina [9,35,36], at present the identity of these Krox-24and c-Fos-immunopositive cells are not known.
Acknowledgements My thanks to Bonnie Blake and Helen Willcockson for excellent technical assistance, to Drs. Rodrigo Bravo and John Wood for the antisera to Krox-24 and RT 97, respectively, and to Drs. Alan Light and Lee Mcllwain for critical reading of the manuscript. This work was supported by American Paralysis Association Grant RB1-9201-1 and the Leopold Schepp Foundation.
References [1] Barron, K.D., McGuiness, C.M., Misantone, L.J., Zanakis, M.F., Grafstein, B. and Murray, M., RNA content of normal and axotomized retinal ganglion cells of rat and goldfish, J. Comp. Neurol., 236 (1985) 265-273. [2] Cho, E.Y.P. and So, K.-F., De novo formation of axon-like processes from axotomized retinal ganglion cells which exhibit long distance growth in a peripheral nerve graft in adult hamsters, Brain Res., 484 (1989) 371-377.
53
[3] Curran, T. and Franza, B.R. Jr., Fos and Jun: the AP-1 connection, Cell, 55 (1988) 395-397. [4] Dobrzanski, P., Noguchi, T., Kovary, K., Rizzo, C.A., Lazo, P.S. and Bravo, R., Both products of the fosB gene, FosB and its short form, FosB/SF, are transcriptional activators in fibroblasts, Mol. Cell. Biol., 11 (1991) 5470-5478. [5] Dodd, J., Solter, D. and Jessel, T.M., Monoclonal antibodies against carbohydrate differentiation antigens identify subsets of primary sensory neurons, Nature, 311 (1984) 469-472. [6] Dr~iger, U. and Hofbauer, A., Antibodies to heavy neurofilament subunit detect a subpopulation of damaged ganglion cells in retina, Nature, 309 (1984) 624-626. [7] Dreher, B., Sefton, A.J., Ni, S.Y.K. and Nisben, G., The morphology, number, distribution and central projections of Class i retinal ganglion cells in albino and hooded rats, Brain Behav. EvoL, 26 (1985) 10-48. [8] Earnest, D.J., Iadarola, M., Yeh, H.H. and Olschowka, J.A., Photic regulation of c-fos expression in neural components governing the entrainment of circadian rhythms, Exp. Neurol., 109 (1990) 353-361. [9] Familietti, E.V., ON and OFF pathways through amacrine cells in mammalian retina: the synaptic connections of "starburst" amacrine cells, Vision Res., 23 (1983) 1265-1279. [10] Fukuda, Y., A three group classification of rat retinal ganglion cells: histological and physiological studies, Brain Res., 119 (1977) 327-344. [11] Halazonetis, T.D., Georgopoulos, K., Greenberg, M.E. and Leder, P., c-Jun dimerizes with itself and c-Fos, forming complexes of different DNA binding affinities, Cell, 55 (1988) 917924. [12] Haas, C.A., Donath, C. and Kreutzberg, G.W., Differential expression of immediate early genes after transection of the facial nerve, Neuroscience, 53 (1993) 91-99. [13] Herdegen, T., Kummer, W., Fiallos, C.E., Leah, J. and Bravo, R., Expression of c-Jun, Jun B and Jun D proteins in the rat nervous system following transection of the vagus nerve and cervical sympathetic trunk, Neuroscience, 45 (1991) 413-422. [14] Herdegen, T., Leah, J.D., Manisali, A., Bravo, R. and Zimmermann, M., c-JUN-like immunoreactivity in the CNS of the adult rat: basal and transynaptically induced expression of an immediate early gene, Neuroscience, 41 (1991) 643-654. [15] Herdegen, T., T611e, T., Bravo, R., Zieglg~insberger, W. and Zimmermann, M., Sequential expression of JUN B, JUN D and FOS B proteins in rat spinal neurons: cascade of transcriptional operations during nociception, Neurosci. Lett., 129 (1991) 221224. [16] Herdegen, T., Fiallos-Estrada, C.E., Schmid, W., Bravo, R. and Zimmermann, M., The transcription factors c-JUN, JUN D and CREB, but not FOS and KROX-24, are differentially regulated in axotomized neurons following transection of the rat sciatic nerve, Mol. Brain. Res., 14 (1992) 155-165. [17] Herdegen, T., Bastmeyer, M., B~ihr, M., Stuermer, C., Bravo, R. and Zimmermann, M., Expression of JUN, KROX, and CREB transcription factors in goldfish and rat retinal ganglion cells following optic nerve lesion is related to axonal sprouting, J. Neurobiol., 24 (1993) 528-543. [18] Itaya, S.K. and Itaya, P.W., Centrifugal fibers to the rat retina from the meaial pretectal area and the periaqueductal grey matter, Brain Res., 326 (1985) 362-365. [19] Jenkins, R. and Hunt,S.P., Long-term increases in the levels of c-iun mRNA and Jun protein-like immunoreactivity in motor and sensory neurons following axon damage, Neurosci. Lett., 129 (1991) 107-110. [20] Jenkins, R., Tetzlaff, W. and Hunt, S.P., Differential expression of immediate early genes in rubrospinal neurons following axotomy in rat, Eur. J. Neurosci., 5 (1993) 203-209. [21] Jones, K.J. and La Velle, A., Differential effects of axotomy on
54
G.A. Robinson/Molecular Brain Research 24 (1994) 43-54
immature and mature hamster facial neurons: a time course study of initial nucleolar and nuclear changes, J. Neurocytol., 15 (1986) 197-206. [22] Koistinaho, J., Hicks, K.J. and Sagar, S.M., Long-term induction of c-jun mRNA and Jun protein in rabbit retinal ganglion cells following axotomy or colchicine treatment, J. Neurosci. Res., 34 (1993) 250-255. [23] Kouzarides, T. and Ziff, E., The role of the leucine zipper in the fos-jun interaction, Nature, 336 (1988) 646-651, [24] Kovary, K. and Bravo, R., Expression of different JUN and FOS during GO to G1 transition in mouse fibroblasts: in vitro and in vivo associations, Mol. Cell. Biol., 11 (1991) 2451-2459. [25} Leah, J.D., Herdegen, T. and Bravo, R., Selective expression of Jun proteins following axotomy and axonal transport block in peripheral nerves in the rat, Brain Res., 566 (1991) 198-207. [26] Lemaire, P., Vesque, C., Schmin, J., Stunnenberg, H., Frank, R. and Charnay, P., The serum-inducible mouse gene Krox 24 encodes a sequence-specific transcriptional activator, Mol. Cell. Biol., 10 (1990) 3456-3467. [27] Lieberman, A.R., Some factors affecting retrograde neuronal responses to axonal lesions. In R. Bellairs and E.G. Gray (Eds.), Essays of the Nervous System, Claredon, Oxford, 1974, pp. 71-105. [28] Linden, R. and Perry, V.H., Massive retinotectal projection in rats, Brain Res., 272 (1983) 145-149. [29] Martin, P.R., The projection of different retinal ganglion cell classes to the dorsal lateral geniculate nucleus in the hooded rat, Exp. Brain Res., 62 (1986) 77-88. [30] Mumberg, D., Lucibello, M., Schuermann, M. and Muller, R., Alternative splicing of fosB transcripts results in differentially expressed mRNAs encoding functionally antagonistic proteins, Genes Deu., 5 (1991) 1212-1223. [31] Nakabeppu, Y., Ryder, K. and Nathans, D., DNA binding activities of three murine Jun proteins; stimulation by Fos, Cell, 55 (1988) 907-915. [32] Nakabeppu, Y. and Nathans, D., A naturally occurring form of FosB that inhibits Fos/Jun transcriptional activity, Cell, 64 (1991) 751-759. [33] Nakabeppu, Y., Oda, S. and Sekiguchi, M., Proliferative activation of quiescent Rat-lA cells by DFosB, Mol. Cell. Biol., 13 . (1993) 4157-4166. [34] Papavassiliou, A.G., Treier, M., Chavrier, C. and Bohmann, D., Targeted degradation of c-Fos, but not v-Fos, by a phospborylation-dependent signal on c-Jun, Science, 258 (1992) 1941-1944. [35] Perry, V.H., Evidence for an amacrine cell system in the ganglion cell layer of the rat retina, Neuroscience, 6 (1981) 931-944.
[36] Perry, V.H. and Walker, M., Arnacrine cells, displaced amacrine cells and interplexiform cells in the retina of the rat, Proc. R. Soc. B, 208 (1980) 415-431. [37] Rauscher, F.J., Voulalas, P.J., Franza, B.R. Jr. and Curran, T., Fos and Jun bind cooperatively to the AP-1 site: reconstitution in vitro, Genes Deu., 2 (1988) 1687-1699. [38] Richardson, P.M., Issa, V.M.K. and Shemie, S,, Regeneration and retrograde degeneration of axons in the rat optic nerve, J. Neurocytol., 11 (1982) 949-966. [39] Robinson, G.A. and Light, A.R., Axotomy-induced expression of immediate early genes in rat retinal ganglion cells, Soc. Neurosci. Abstr., 18 (1992) 1312. [40] Sagar, S.M. and Sharp, F.R., Light induces a Fos-like nuclear antigen in retinal neurons, Mol. Brain Res., 7 (1990) 17-21. [41] Vidal-Sanz, M., Bray, G.M., Villegas-P6rez, M.P., Thanos, S. and Aguayo, A.J., Axonal regeneration and synapse formation in the superior colliculus by retinal ganglion cells in the adult rat, Z Neurosci., 7 (1987) 2894-2909. [42] Vidal-Sanz, M., Villegas-P6rez, M.P., Bray, G.M. and Aguayo, A.J., Persistent retrograde labelling of adult rat retinal ganglion cells with the carbocyanine dye, diI, Exp. Neurol., 102 (1988) 92-101. [43] Villar, M.J., Vitale, M.L. and Parisi, M.N., Dorsal raphe serotonergic projection to the retina. A combined peroxidase tracing-neurochemical/high performance liquid chromatography study in the rat, Neuroscience, 22 (1987) 681-686. [44] Villegas-P6rez, M.P., Vidal-Sanz, M., Bray, G.M. and Aguayo, A.J., Influences of peripheral nerve grafts on the survival and regrowth of axotomized retinal ganglion cells in adult rats, J. Neurosci., 8 (1988) 265-280. [45] Villegas-P6rez, M.P., Vidal-Sanz, M., Bray, G.M. and Aguayo, A.J., Rapid and protracted phases of retinal ganglion cell loss follow axotomy in the optic nerve of adult rats, J. Neurobiol., 24 (1993) 23-36. [46] Weiniawa-Narkiewicz, E. and Hughes, A., The superficial plexiform layer: a third retinal association area, J. Comp. Neurol., 324 (1992) 463-484. [47] Wisdom, R., Yen, J., Rashid, D. and Verma, M., Transformation by Fos B requires a trans-activation domain missing in FosB2 that can be substituted by heterologous activation domains, Genes Deu., 6 (1992) 667-675. [48] Yen, J., Wisdom, R.M., Tratner, I. and Verma, I.M., An alternative spliced form of FosB is a negative regulator of transcriptional activation and transformation by Fos proteins, Proc. Natl. Acad. Sci. USA, 88 (1991) 5077-5081.