EXPERIMENTAL
112,
NEUROLOGY
79-88
(19%)
Nerve Growth Factor Promotes CNS Cholinergic Axonal Regeneration into Acellular Peripheral Nerve Grafts THEO
HAGG,
ADARW SILVIO
Department of Biology, and *Department
K. GULATI,* VARON,
M. ALI BEHZADIAN,* H. AND MARSTON MANTHORPE
VAHLSING,
M-001, University of California, San Diego, La Jolla, California 92093; of Anatomy, Medical College of Georgia, Augusta, Georgia 30912
thawing of fresh or in situ degenerated peripheral nerve eliminates all living cells but leaves cellular debris in addition to a relatively intact extracellular matrix and physical structure (214, 28). Freeze-thawing deprives such nerve grafts of their competence to promote CNS axonal regeneration (4, 44) and apparently allows peripheral nervous system regeneration only in the presence of comigrated Schwann cells ((14), review: (29); but see also (28) and (43)). This loss in competence may be due to (i) the interruption of cellular production of trophic factors (e.g., nerve growth factor, ciliary neuronotrophic factor) (6, 7, 26, 37, 38, 48), (ii) the alteration of extracellular matrix-associated neurite-promoting factors (laminin, fibronectin, collagen) (10, 33, 35,40,42); (iii) the generation of cellular debris within the basal lamina tubes that impedes or inhibits axonal growth (3); or (iv) a combination of the above. We have addressed these issues by examining the regeneration-promoting competence of acellular, debrisfree peripheral nerve grafts (5, 15) for adult rat CNS cholinergic axons. These acellular grafts were prepared by culturing predegenerated, freeze-thawed nerve segments in the presence of actively phagocytizing peritoneal macrophages, a treatment that eliminates virtually all cellular debris with no obvious disruption of their longitudinally oriented, laminin-rich basal lamina tubular spaces. When acellular nerve was implanted in the fimbria-fornix transection cavity between the septum and the hippocampal formation, very few cholinergic axons extended into the graft. Since NGF (i) is produced by Schwann cells (38, 50), (ii) increases in peripheral nerve after its transection (26,46), and (iii) acts on medial septum cholinergic neurons (51), we treated other acellular nerve segments with NGF-saline or saline alone before implantation. NGF-treated grafts promoted the regeneration of cholinergic axons into and across the acellular nerve to an extent approaching that of the fresh cellular nerve grafts.
Peripheral nerve grafts promote vigorous regeneration of adult mammalian CNS axons. Elimination of nerve-associated cells by freeze-thawing abolishes this promoting quality, possibly by creating inhibitory cellular debris andlor destroying the production of stimulatory factors by living Schwann or other cells. Here, debris-free acellular peripheral nerve segments placed between the disconnected septum and the hippocampal formation acquired almost no cholinergie axons after 1 month. However, such acellular nerve grafts treated before implantation with purified &nerve growth factor (NGF) contained nearly as many Longitudinally oriented cholinergic. axons as did fresh cellular nerve grafts. These results suggest that (i) NGF is required for the regeneration of adult CNS cholinergic axons into nerve grafts and (ii) an important function of living cells within peripheral nerve may be the production of neuronotrophic factors such as NGF. o 1991 Academic Press,
LEE
Inc.
The poor capacity of damaged adult mammalian central nervous system (CNS) axons to regenerate is a fundamental and unresolved problem. The results of many studies have encouraged the view that adult brain and spinal cord neurons retain a vigorous capacity to regenerate their axons, hut that regeneration inside the CNS fails because of an inadequate tissue environment. Significant CNS axonal regeneration can occur into transplants such as fetal brain tissue (32), viable adult peripheral nerve (1, 8, 22), and previously cultured Schwann cells (31). The peripheral nerve graft contains a number of components potentially responsible for its effectiveness in promoting CNS axonal regeneration. Nerve grafts can be considered mechanochemical scaffolds containing living cells (Schwann, perineurial, endothelial), a well-developed extracellular matrix and its molecular components (e.g., laminin, fibronectin, collagen), and a longitudinally oriented physical structure (cylindrical basal lamina-lined tubes). Repeated freeze-
MATERIALS
AND
METHODS
Fresh autologous sciatic nerves were dissected just before implantation time, cut into 1.5 to 2-mm pieces, 79 All
Copyright 0 1991 rights of reproduction
0014-4886/91 $3.00 by Academic Press, Inc. in any form reserved.
HAGG
ET
AL.
FIG. 1. Morphological comparison of normal and acellular nerve. (A) PAS-hematoxylin-stained cross sections of fresh cellular and (B) acellular peripheral nerve showing the absence of myelin and cell bodies in the latter. Bar = 200 pm. (C) Anti-laminin-stained (fluorescence) cross section of fresh cellular and (D) acellular peripheral nerve showing the preservation of smaller diameter laminin-rich basal lamina tubes
NGF
PROMOTES
CNS
REGENERATION
placed into sterile balanced salt solution, and used within half an hour from dissection. Acellular nerve grafts were prepared as follows. Sciatic nerves of adult female Fischer 344 rats were transected just distal to the sciatic notch and allowed to degenerate in situ for 6 weeks so as to eliminate axonal and myelin debris (14, 24,45). The predegenerated nerves were resected proximal to the bifurcation of the peroneal and tibia1 nerve, cut into 1 cm segments, and freeze-thawed five times in liquid N, to lyse all nerve-associated cells. The nerves were then cultured for 1 week with 2 X 10” rat peritoneal macrophages in 2 ml RPM1 medium per 35-mm culture dish containing 20% heat-inactivated fetal bovine serum. The pieces of nerve no longer occupied by macrophages were removed from culture, washed in saline, and stored at -70°C. By light and electron microscopy, the final preparation compared to normal nerve (Figs. lA, lC, and 1E) appears completely free of myelin, axons, and intact cells (Fig. 1B) and consists of lamininrich basal lamina tubes (Fig. lD, stained for laminin according to Ref. (17)) essentially free of cell debris (Fig. 1F). The number of laminin-positive tube-like structures in the acellular nerve (Fig. 1D) was about the same (2600 vs 3000/mm*) as that seen in a normal nerve (Fig. lC), although mostly of smaller diameter (compare Figs. 1C and 1D and Figs. 1E and 1F). For implantation into host rats acellular nerve segments were thawed, stripped of their perineurium/sheath, cut into 1.5- to 2-mm pieces, incubated for 24 h in sterile saline alone or in saline containing 320 pg/ml (8 X lo7 Trophic Units/ ml) of P-NGF, prepared, and assayed as described previously (36, 49). Surgical, histological, and analytical procedures have been described in detail elsewhere (22). Briefly, 27 adult female Sprague-Dawley rats (200-220 g, Bantin and Kingman) received a complete bilateral aspirative fimbria-fornix transection, leaving a l&mm lesion gap between the septum and the hippocampal formation and completely severing, among others, the dorsal septohippocampal cholinergic pathways. The lesion was carefully inspected through the surgical microscope for its completeness and bleeding was stopped by temporary application of pieces of gelfoam soaked in sterile phosphate-buffered saline. Each animal received two bilaterally placed nerve grafts of the same type, i.e., either saline-soaked acellular nerves (n = 14 nerve grafts in 7 rats), or fi-NGF-soaked acellular nerves (n = 12 grafts in 6 rats) or fresh autologous nerves (n = 20 grafts in 10 rats). Each nerve graft was placed with one stump apposing the cut face of the caudal septum/fornix and the other stump apposing the lesion face of the rostra1 hip-
INTO
ACELLULAR
NERVE
GRAFTS
81
pocampal formation (Fig. 2). The apposition was secured by placing gelfoam over the hippocampal tip and nerve graft. In some animals receiving a complete bilateral aspirative fimbria-fornix transection (n = 8 sides), the lesion cavity was loosely filled with gelfoam with no nerve graft. After 1 month, the animals were perfused transcardially (while clamping the descending aorta) with 75 ml ice-cold phosphate-buffered saline followed by 250 ml of ice-cold 4% buffered paraformaldehyde. The brains were removed, postfixed for 16-24 h, and cryoprotected in buffered 30% sucrose for 16-24 h, all at 4°C. Thirtymicrometer coronal sections were cut on a freezing microtome from the brain up to the anterior commissure decussation just rostra1 to the beginning of the fornix; the brain was thawed, repositioned, and refrozen; and 40-pm sagittal sections were cut from both sides of the rest of the brain. One of every twelve sagittal sections was mounted on glass slides, dried overnight, and stained for AChE (25). The incubation time with acetylthiocholine was doubled (two times 30 min) to allow for better penetration into the mounted tissue. Promethazine (0.2 mM, Sigma) was used as an inhibitor of nonspecific esterases. Two coronal sections through the medial septum of each animal were selected and immunostained for choline acetyltransferase (ChAT) as described elsewhere (21, 22). Cholinergic fiber number was quantified as described (22) by counting in every AChE-stained sagittal section (480 pm apart) the number of AChE-positive fibers that intersected eyepiece grid lines oriented perpendicular to the axis of the nerve graft (at 500X) and hippocampal formation (at 250X). The grid lines were placed over the nerve at 0.2 mm from its hippocampal end and over the hippocampal formation at 0.1 and 1 mm caudal to its rostra1 tip (Fig. 2). To quantify cholinergic axonal regeneration into the nerve, the single sagittal section containing the largest AChE-positive nerve fiber mass was chosen. With 0.48 mm spacing at least two sections through the length of the l-mm-diameter nerve graft were available for analysis. In all cases at least one section transversed at least 80% and in more than half of the cases more than 90% of the nerve diameter, i.e., through the thickest part of the nerve. The large number of cases per experimental group would correct for any remaining intergroup differences in the sagittal position/level of the analyzed nerve sections. In the nerve, the number of individual cholinergic fibers could not be accurately determined because of their frequent fasciculation (Fig. 3). AChE-positive structures, therefore, were scored and counted as “thin” (
112,
NEUROLOGY
79-88
(19%)
Nerve Growth Factor Promotes CNS Cholinergic Axonal Regeneration into Acellular Peripheral Nerve Grafts THEO
HAGG,
ADARW SILVIO
Department of Biology, and *Department
K. GULATI,* VARON,
M. ALI BEHZADIAN,* H. AND MARSTON MANTHORPE
VAHLSING,
M-001, University of California, San Diego, La Jolla, California 92093; of Anatomy, Medical College of Georgia, Augusta, Georgia 30912
thawing of fresh or in situ degenerated peripheral nerve eliminates all living cells but leaves cellular debris in addition to a relatively intact extracellular matrix and physical structure (214, 28). Freeze-thawing deprives such nerve grafts of their competence to promote CNS axonal regeneration (4, 44) and apparently allows peripheral nervous system regeneration only in the presence of comigrated Schwann cells ((14), review: (29); but see also (28) and (43)). This loss in competence may be due to (i) the interruption of cellular production of trophic factors (e.g., nerve growth factor, ciliary neuronotrophic factor) (6, 7, 26, 37, 38, 48), (ii) the alteration of extracellular matrix-associated neurite-promoting factors (laminin, fibronectin, collagen) (10, 33, 35,40,42); (iii) the generation of cellular debris within the basal lamina tubes that impedes or inhibits axonal growth (3); or (iv) a combination of the above. We have addressed these issues by examining the regeneration-promoting competence of acellular, debrisfree peripheral nerve grafts (5, 15) for adult rat CNS cholinergic axons. These acellular grafts were prepared by culturing predegenerated, freeze-thawed nerve segments in the presence of actively phagocytizing peritoneal macrophages, a treatment that eliminates virtually all cellular debris with no obvious disruption of their longitudinally oriented, laminin-rich basal lamina tubular spaces. When acellular nerve was implanted in the fimbria-fornix transection cavity between the septum and the hippocampal formation, very few cholinergic axons extended into the graft. Since NGF (i) is produced by Schwann cells (38, 50), (ii) increases in peripheral nerve after its transection (26,46), and (iii) acts on medial septum cholinergic neurons (51), we treated other acellular nerve segments with NGF-saline or saline alone before implantation. NGF-treated grafts promoted the regeneration of cholinergic axons into and across the acellular nerve to an extent approaching that of the fresh cellular nerve grafts.
Peripheral nerve grafts promote vigorous regeneration of adult mammalian CNS axons. Elimination of nerve-associated cells by freeze-thawing abolishes this promoting quality, possibly by creating inhibitory cellular debris andlor destroying the production of stimulatory factors by living Schwann or other cells. Here, debris-free acellular peripheral nerve segments placed between the disconnected septum and the hippocampal formation acquired almost no cholinergie axons after 1 month. However, such acellular nerve grafts treated before implantation with purified &nerve growth factor (NGF) contained nearly as many Longitudinally oriented cholinergic. axons as did fresh cellular nerve grafts. These results suggest that (i) NGF is required for the regeneration of adult CNS cholinergic axons into nerve grafts and (ii) an important function of living cells within peripheral nerve may be the production of neuronotrophic factors such as NGF. o 1991 Academic Press,
LEE
Inc.
The poor capacity of damaged adult mammalian central nervous system (CNS) axons to regenerate is a fundamental and unresolved problem. The results of many studies have encouraged the view that adult brain and spinal cord neurons retain a vigorous capacity to regenerate their axons, hut that regeneration inside the CNS fails because of an inadequate tissue environment. Significant CNS axonal regeneration can occur into transplants such as fetal brain tissue (32), viable adult peripheral nerve (1, 8, 22), and previously cultured Schwann cells (31). The peripheral nerve graft contains a number of components potentially responsible for its effectiveness in promoting CNS axonal regeneration. Nerve grafts can be considered mechanochemical scaffolds containing living cells (Schwann, perineurial, endothelial), a well-developed extracellular matrix and its molecular components (e.g., laminin, fibronectin, collagen), and a longitudinally oriented physical structure (cylindrical basal lamina-lined tubes). Repeated freeze-
MATERIALS
AND
METHODS
Fresh autologous sciatic nerves were dissected just before implantation time, cut into 1.5 to 2-mm pieces, 79 All
Copyright 0 1991 rights of reproduction
0014-4886/91 $3.00 by Academic Press, Inc. in any form reserved.
HAGG
ET
AL.
FIG. 1. Morphological comparison of normal and acellular nerve. (A) PAS-hematoxylin-stained cross sections of fresh cellular and (B) acellular peripheral nerve showing the absence of myelin and cell bodies in the latter. Bar = 200 pm. (C) Anti-laminin-stained (fluorescence) cross section of fresh cellular and (D) acellular peripheral nerve showing the preservation of smaller diameter laminin-rich basal lamina tubes
NGF
PROMOTES
CNS
REGENERATION
placed into sterile balanced salt solution, and used within half an hour from dissection. Acellular nerve grafts were prepared as follows. Sciatic nerves of adult female Fischer 344 rats were transected just distal to the sciatic notch and allowed to degenerate in situ for 6 weeks so as to eliminate axonal and myelin debris (14, 24,45). The predegenerated nerves were resected proximal to the bifurcation of the peroneal and tibia1 nerve, cut into 1 cm segments, and freeze-thawed five times in liquid N, to lyse all nerve-associated cells. The nerves were then cultured for 1 week with 2 X 10” rat peritoneal macrophages in 2 ml RPM1 medium per 35-mm culture dish containing 20% heat-inactivated fetal bovine serum. The pieces of nerve no longer occupied by macrophages were removed from culture, washed in saline, and stored at -70°C. By light and electron microscopy, the final preparation compared to normal nerve (Figs. lA, lC, and 1E) appears completely free of myelin, axons, and intact cells (Fig. 1B) and consists of lamininrich basal lamina tubes (Fig. lD, stained for laminin according to Ref. (17)) essentially free of cell debris (Fig. 1F). The number of laminin-positive tube-like structures in the acellular nerve (Fig. 1D) was about the same (2600 vs 3000/mm*) as that seen in a normal nerve (Fig. lC), although mostly of smaller diameter (compare Figs. 1C and 1D and Figs. 1E and 1F). For implantation into host rats acellular nerve segments were thawed, stripped of their perineurium/sheath, cut into 1.5- to 2-mm pieces, incubated for 24 h in sterile saline alone or in saline containing 320 pg/ml (8 X lo7 Trophic Units/ ml) of P-NGF, prepared, and assayed as described previously (36, 49). Surgical, histological, and analytical procedures have been described in detail elsewhere (22). Briefly, 27 adult female Sprague-Dawley rats (200-220 g, Bantin and Kingman) received a complete bilateral aspirative fimbria-fornix transection, leaving a l&mm lesion gap between the septum and the hippocampal formation and completely severing, among others, the dorsal septohippocampal cholinergic pathways. The lesion was carefully inspected through the surgical microscope for its completeness and bleeding was stopped by temporary application of pieces of gelfoam soaked in sterile phosphate-buffered saline. Each animal received two bilaterally placed nerve grafts of the same type, i.e., either saline-soaked acellular nerves (n = 14 nerve grafts in 7 rats), or fi-NGF-soaked acellular nerves (n = 12 grafts in 6 rats) or fresh autologous nerves (n = 20 grafts in 10 rats). Each nerve graft was placed with one stump apposing the cut face of the caudal septum/fornix and the other stump apposing the lesion face of the rostra1 hip-
INTO
ACELLULAR
NERVE
GRAFTS
81
pocampal formation (Fig. 2). The apposition was secured by placing gelfoam over the hippocampal tip and nerve graft. In some animals receiving a complete bilateral aspirative fimbria-fornix transection (n = 8 sides), the lesion cavity was loosely filled with gelfoam with no nerve graft. After 1 month, the animals were perfused transcardially (while clamping the descending aorta) with 75 ml ice-cold phosphate-buffered saline followed by 250 ml of ice-cold 4% buffered paraformaldehyde. The brains were removed, postfixed for 16-24 h, and cryoprotected in buffered 30% sucrose for 16-24 h, all at 4°C. Thirtymicrometer coronal sections were cut on a freezing microtome from the brain up to the anterior commissure decussation just rostra1 to the beginning of the fornix; the brain was thawed, repositioned, and refrozen; and 40-pm sagittal sections were cut from both sides of the rest of the brain. One of every twelve sagittal sections was mounted on glass slides, dried overnight, and stained for AChE (25). The incubation time with acetylthiocholine was doubled (two times 30 min) to allow for better penetration into the mounted tissue. Promethazine (0.2 mM, Sigma) was used as an inhibitor of nonspecific esterases. Two coronal sections through the medial septum of each animal were selected and immunostained for choline acetyltransferase (ChAT) as described elsewhere (21, 22). Cholinergic fiber number was quantified as described (22) by counting in every AChE-stained sagittal section (480 pm apart) the number of AChE-positive fibers that intersected eyepiece grid lines oriented perpendicular to the axis of the nerve graft (at 500X) and hippocampal formation (at 250X). The grid lines were placed over the nerve at 0.2 mm from its hippocampal end and over the hippocampal formation at 0.1 and 1 mm caudal to its rostra1 tip (Fig. 2). To quantify cholinergic axonal regeneration into the nerve, the single sagittal section containing the largest AChE-positive nerve fiber mass was chosen. With 0.48 mm spacing at least two sections through the length of the l-mm-diameter nerve graft were available for analysis. In all cases at least one section transversed at least 80% and in more than half of the cases more than 90% of the nerve diameter, i.e., through the thickest part of the nerve. The large number of cases per experimental group would correct for any remaining intergroup differences in the sagittal position/level of the analyzed nerve sections. In the nerve, the number of individual cholinergic fibers could not be accurately determined because of their frequent fasciculation (Fig. 3). AChE-positive structures, therefore, were scored and counted as “thin” (
