EXPERIMENTAL

NEUROLOGY

114,44-52

(1991)

Differential Effects on Sensory Nerve Processes and Behavioral Alterations in the Rat after Treatment with Antibodies to Nerve Growth Factor BETHANY A. URSCHEL, PATRICK N. BROWN, AND CLAIRE E. HULSEBOSCH Department of Anatomy and Neurosciences and the Marine Biomedical Institute, University of Texas Medical Branch, Galveston, Texas 77550-2772

39,42,48,52). We confirm reports that 38% of mammalian sensory neurons die in an NGF-suppressed environment created by chronic treatment with antibodies to NGF (ANTI-NGF) (31, 55). Additionally, we demonstrated a 35% increase in the number of centrally projecting sensory fibers in the dorsal root (31) and a 49% increase in fiber number in Lissauer’s tract (27). The increases in fiber number are principally in the unmyelinated population. We interpret the increase in unmyelinated fiber number to be sprouting of centrally directed processes of primary sensory neurons. Since primary afferent neurons have processes that project to the periphery, it was of interest to determine if the sensory unmyelinated fibers in the mixed peripheral nerve population demonstrated an increase in fiber number. Previous reports of total fiber numbers in the cervical trunk (2), fiber numbers in the nerve to the medial gastrocnemius (33), and responses of cutaneous nerves which innervate the trunk (13) after ANTI-NGF treatment in rats indicate that total fiber numbers may be reduced but these studies do not attempt to separate fiber types anatomically. The purpose of the present study was to repeat the ANTI-NGF treatment which induced intraspinal sprouting of centrally directed primary afferent fibers, combine selective surgical procedures in which the peripheral sensory fibers are isolated from postganglionic sympathetic and motor fibers, and quantitatively compare the sensory fiber numbers in ANTI-NGF-treated peripheral nerves to counts from similar nerves in untreated littermates. In addition, since sensory fiber populations are altered after ANTINGF treatment, it follows that sensory responses to stimuli may change. Since the changes are principally in the unmyelinated population, and the majority of unmyelinated fibers are nociceptive (53), it is reasonable to test behaviors which are thought to be nociceptive. To test for behavioral changes in nociception, two behavioral tests were performed: (i) the tail flick, which is a measure of reflexive activity to nociception, and (ii) the paw or skin pinch, which is a nonreflexive measure of nociception (22).

Published work on the effects of antibodies to nerve growth factor (ANTI-NGF) treatment on rats has shown an increase in the number of unmyelinated central processes of dorsal root ganglia (DRG) neurons (31). This increase is interpreted to be sprouting of the central projections of the DRG neurons. To test for sprouting of the peripheral DRG projections, we quantitated the number of peripheral DRG processes in the peripheral nerves of ANTI-NGF-treated compared to untreated rats, following selective surgery to eliminate motor and sympathetic nerve fibers. We report the numbers of peripheral DRG processes in an NGF-deprived environment decrease by 43% compared to untreated controls and the decrease is selective for the unmyelinated fiber population. Since the majority of the unmyelinated population is nociceptive, two nociceptive behavioral measures, one reflexive (tail flick) and one nonreflexive (paw or skin pinch), were performed and demonstrated decreased responses in the ANTI-NGFtreated compared to untreated and preimmune-treated rats. These data suggest a directional effect, primarily on the unmyelinated sensory population which results in altered nociceptive behavior, induced by the suppression of one endogenous factor, NGF. Furthermore, it is important to note that the centrally directed sensory processes project to a central nervous system environment and the peripherally directed processes are in a peripheral nervous system environment. Thus, a single molecule may have different effects on directional growth of a neuronal population that may be related to the interactions available in the substrate of the environment. 0 1991 Academic Press, Inc.

INTRODUCTION Nerve growth factor (NGF) is a well-characterized neurotrophic factor which selectively effects neuron survival, in L&O, of postganglionic sympathetic neurons, sensory neurons, and central cholinergic neurons (21, 0014-4666/91 Copyright All rights

$3.00 0 1991 by Academic Press, of reproduction in any form

44 Inc. reserved.

DIFFERENTIAL

MATERIALS

AND

EFFECTS

OF

ANTI-NGF

ON

SENSORY

45

FIBERS

METHODS

Pregnant Sprague-Dawley rats were obtained from Harlan Sprague-Dawley, Inc. All procedures involving rats were performed in compliance with USDA Animal Welfare Act and amendments, the revised Guide for the Care and Use of Laboratory Animals DHEW (NIH) and were approved by the UTMB Animal Care and Use Committee. Two groups of rats were examined: (i) rats treated with antibodies to NGF (ANTI-NGF) and (ii) untreated littermates (UNTR). Neonatal rats (Postnatal Day 0, PD 0) were given daily subcutaneous injections near the dorsal fat pad of antibodies to NGF (undiluted rabbit antiserum, 3 pllgm body wt) for a period of 1 month. The NGF (lo-20 rig/ml for 1 biological unit/ml, (41)) used to produce the rabbit antibodies was the purified P-NGF subunit of 7s NGF purified from mouse submaxillary glands by the methods described in Varon et al. (51). The classic bioassay with extirpated embryonic chick sensory ganglia (16, 20) and a more recent bioassay with morphological differentiation of cultured human neuroblastoma cells (31) were used to monitor the biological activity of both NGF and the antibodies against it. An antiserum dilution of 1:500 blocked the effects of exogenously applied NGF at concentrations containing 1 biological unit/ml (42) and by both assays the blocking activity of ANTI-NGF was stable during the course of the experiments. To determine in uiuo injection levels, it was determined empirically that increases in the antibody concentration in the sera by ultrafiltration (29) affected mortality (50). Consequently, the use of whole antisera was deemed sufficient to block endogenous levels of NGF in these experiments. As an internal assay, superior cervical ganglia in both the ANTI-NGFand UNTR-treated rats were visually inspected. The ganglia in the ANTI-NGF rats were atrophied to approximately half the size of those in the UNTR rats. As a control, littermates were given preimmune sera with similar injection schedules as the ANTI-NGF-treated animals. These animals were not found to be different from untreated rats (30, 50). Unilateral radical ventral rhizotomies were performed from T6-Tll in both ANTI-NGF and UNTR groups to isolate the peripheral projecting processes of primary sensory afferents in the peripheral nerve (Fig. 1). The rats were anesthetized by an intraperitoneal injection of sodium pentobarbital (40 mg/kg) and when anesthesia was deep (as determined by reflex testing), surgery was begun and a laminectomy was performed. Once the spinal cord and ventral roots (VR) were exposed, the VR and sympathetic rami were cut, the spinal cord was covered with a protective layer of artificial dura, and the musculature and connective tissue of the back were sutured. An injection of 1 cc saline solution was given intraperitoneally in order to restore fluid levels. The animals were awake and eating within an hour.

SC

VR

FIG. 1. Schematic drawing demonstrating the types of fibers cut during the ventral rhizotomy surgery. CORD, spinal cord; DR, dorsal root; DRG, dorsal root ganglion; PN, peripheral nerve; SC, sympathetic chain; VR, ventral root; solid line, motor fiber; dotted line, preand post-ganglionic sympathetic fibers; dotted and dashed line, peripherally and centrally projecting sensory fibers; double line, area cut during the ventral rhizotomy resulting in only primary sensory fibers being in the PN after surgery.

Five to seven days after surgery, the rats were prepared for anatomical analysis by anesthesia delivered as above and when anesthesia was deep, perfusion with 0.9% NaCl containing 0.2 ml of 1.0% NaNO, and 200 units of heparin per 100 ml was begun. After 3 min, the perfusion fluid was changed to an aldehyde mixture of 3% glutaraldehyde, 3% formaldehyde, and 0.1% picric acid in a 1.0 M, pH 7.4, cacodylate buffer. The T9 peripheral nerves (PN) ipsilateral to the surgical manipulation were removed and placed in a fresh aldehyde mixture. The following day the tissues were rinsed for 30 min in 0.1 M, pH 7.4, cacodylate buffer and postfixed in a mixture of 1% osmic acid and 1.5% potassium ferricyanide in 0.1 M, pH 7.4, cacodylate buffer for l-2 h (40). The tissues were then rinsed in 0.1 M, pH 5.2, maleate buffer for 30 min and placed overnight in a 0.5% uranyl acetate in pH 6.0 maleate buffer. The next day the tissues were rinsed in 0.1 M, pH 5.2, maleate buffer for 30 min, dehydrated for 4 h in a series of ethyl alcohol solutions ranging from 50% to absolute, and embedded in a mixture of Epon and Araldite. Thin cross sections from the same area of each peripheral nerve (just distal to the origin of the dorsal primary ramus, i.e., the ventral primary ramus or intercostal nerve) were cut with a diamond knife using an MT-l ultramicrotome and sections were placed on single hole grids covered with a Formvar film. This allowed for visualization of all fibers in the entire nerve without obstruction by grid bars. The sections were left unstained or poststained with lead citrate and placed in a JEOL lOO-CX electron microscope. Myelinated and unmyelinated axons were counted either on the electron microscope or from montages made from electron micrographs. All axons were counted.

46

URSCHEL,

BROWN,

Counting error, in terms of repeated counts of the tissue, was +-3%. A double-blind format was used. The data were analyzed for significance using the Student t test with P < 0.05 chosen as the criterion for the level of confidence. Two different nociceptive behavioral tests were administered at PD 14 and PD 30 to three groups of rats: (i) treated as described above with ANTI-NGF, (ii) treated with preimmune sera from the same rabbits used to generate the ANTI-NGF (PREIMM), and (iii) untreated littermates. The first behavior test which is a measure of a reflexive response to noeiception (22), the tail flick test, was administered as per Doucette et at. (14) on unanesthetized animals. The test consisted of dipping the end of the tail into a beaker of 55*C! water and the time to tail withdrawal was measured. In order to prevent the animals from moving during the course of the testing, they were persuaded to enter either a test tube or a beaker padded with Kimwipes. Once settled, the tail hung vertically over the edge of the counter. The test was administered three times to each animal with at least a 1-min interval between each trial. Withdrawal time was measured to the nearest one-hundredth of a second with a stopwatch. The three latencies were then averaged to obtain an average tail withdrawal time for each rat. The second test, a pinch test, measures a nonreflexive response to nociception (22) and was administered to unanesthetized animals. This test consisted of pinching the forepaw, back, or hindpaw of the animal with forceps and counting the number of vocalizations. A period of at least 1 min was allowed between each pinch. The three counts were then combined to give an average number of vocalizations per animal. All measurements were done in a double-blind format. The data were analyzed for significance by use of the Student t test with P < 0.05 chosen as the criterion for the level of confidence. RESULTS

The ultrastructural morphology of the myelinated and unmyelinated fibers and constituents in the peripheral nerve of ANTI-NGF-treated rats is similar to that of untreated rats. The collagen fibrils are in regularly oriented arrays parallel to the axis of the nerve and thus appear in cross-section in the samples used in this study. The myelinated fibers are identifiable by the presence of the extreme electron dense appearance of the lamellar wrapping of the closely apposed membranes characteristic of myelin and the presence of neurotubules and neurofilaments in the axoplasm. The unmyelinated fibers appear round or oval in shape and relatively electron lucent compared to the cytoplasm of the ensheathing Schwann cells and are in continuity with the extra~ellular environment either directly with little or no Schwann cell ensheathment or by way of mesax-

AND HULSEBOSCH

ons. The axoplasm of the unmyelinated fibers can easily be distinguished from circular or oval profiles of Schwann cells by virtue of the difference in electron density and the presence of neurotubules and neurofilaments in the axoplasm. Each myelinated fiber and Remak bundle (the term given to a group of unmyelinated fibers which appear in cross-section, to be invested by a single Schwann cell) is invested by basal laminae (Fig. 2). The important point is the ease of identification of myelinated or unmyelinated fibers at the ultrastructural level with present fixation techniques which in turn permitted accurate quantitation. As stated above, unilateral radical ventral rhizotomies were performed in order to isolate the peripheral sensory processes in both ANTI-NGF and UNTR groups. Surgeries were performed 5 to 7 days prior to sacrifice to allow enough time for recognizable degeneration of severed distal processes separated from the neuronal soma by the surgery. We know that by three days after surgery, the severed distal processes of unmyelinated axons remain only as degeneration debris (26). By 5 days after surgery, three types of degenerating myelinated axon profiles can be observed. In the first type of degenerating profile, remains of myelin sheaths can be seen surrounding an electron lucent area, which was the axon. The second type of degenerating profile also exhibited remnants of a myelin sheath surrounding an axon; however, the axoplasm was electron dense and exhibited no organization of axoplasmic organelles and elements such as neurofilaments and neurotubules. The last type of degenerating profile contained several secondary lysosomes, which contained myelin and axonal debris. The important points are the following: (i) nerves in both the ANTI-NGF and UNTR groups were similar in appearance and thus ultrastructural quantitation was necessary, and (ii) we were able to distinguish normal (non-degenerating) unmyelinated and myelinated axons from degeneratingunmyelinated andmyelinated axons (26) in both groups and thus we were able to count normal profiles. The ANTI-NGF and UNTR groups were indistinguishable in terms of general and ultrastructural morphology, with the exception of the number of unmyelinated fibers. To test for these differences, it was necessary to count myelinated and unmyelinated fibers. The counts of the surgically isolated sensory myelinated and unmyelinated fibers in the T9 peripheral nerve of ANTI-NGF-treated and untreated animals are shown in Table 1. The means of the total fiber population counts for the ANTI-NGF-treated and untreated rats are 4458 and 8520, respectively. Thus, there are 52% of the sensory fibers remaining in the ANTI-NGF treated animals compared to untreated littermates. This difference is statistically significant (P 6 0.000005). The means of the myelinated fiber counts from the ANTINGF compared to untreated animals are 1029 and 870,

DIFFERENTIAL

FIG. 2. nerve after

Electron selective

micrograph of unmyelinated surgery. Un, unmyelinated

EFFECTS

OF

fibers in a Remak fiber; My, myelinated

respectively; however, this difference is not statistically significant. By contrast, the means of the unmyelinated fiber counts for the ANTI-NGF-treated and untreated animals are 3429 and 7650, respectively. Thus, there are 45% of the sensory unmyelinated fibers remaining in the ANTI-NGF-treated rats compared to untreated rats. This difference is statistically significant (P G 0.000005).

Tail-flick. The results of the behavioral tests are shown in Table 2. There was a statistically significant difference between the UNTR and PREIMM groups seen on PD 14 with the tail flick test (0.72 rt 0.01 and 0.78 +- 0.002, respectively, Student’s t test, P < 0.001). The tail flick test also showed significant differences between the ANTI-NGF and UNTR groups (1.13 -+ 0.006 and 0.72 f 0.01, respectively, Student’s t test, P & 10e6) and between the ANTI-NGF and PREIMM groups (1.13 it 0.006 and 0.78 + 0.002, respectively, Student’s t test P < 0.00003). By PD 30, there was no signifi-

ANTI-NGF

ON

SENSORY

47

FIBERS

bundle and neighboring myelinated fibers fiber; Nu, Schwann cell nucleus. Calibration

of an untreated bar = 1 pm,

peripheral

cant difference seen between the UNTR and PREIMM groups when the tail flick test was administered (0.64 I 0.03 and 0.67 f 0.03, respectively, Student’s t test, P < 0.45). However, there were still significant differences between the ANTI-NGF and UNTR groups (1.06 +- 0.03 and 0.64 + 0.03, respectively, Student’s t test, P < 0.00005) and between the ANTI-NGF and PREIMM groups (1.06 t 0.03 and 0.67 f 0.03, respectively, Student’s t test, P < 0.0006). In summary, the ANTI-NGFtreated animals took longer to respond to thermal nociceptive stimulus than UNTR or PR~IMM animals and this behavioral delay is statistically significant. Pinch test. With the pinch test, there was no significant difference seen between the UNTR and PREIMM groups on PD 14 (2.7 f 0.08 and 2.7 r?t:0.00, respectively, Student’s t test, P < 0.57). However, there were statistically significant differences between the responses observed between the ANTI-NGF and UNTR groups (1.3 f 0.17 and 2.7 t 0.08, respectively, Student’s t test, P < 0.00005) and between the ANTI-NGF and PREIMM groups (1.3 rt 0.17 and 2.7 _t 0.00, respectively, Student’s t test, P ,( 0.003). On PD 30 the pinch test again showed no significant difference between the UNTR and

48

URSCHEL,

TABLE Sensory Animal UNTR

Mean k SD ANTI-NGF

Mean -cSD

1 2 3 4

1 2 3 4

Fiber

Counts

BROWN,

1

of ‘I’9 Peripheral

Myelinated

Unmyelinated

822 811 810 1037 870.0 96.5 1076 827 1113 1098 1028.5 117.1

7361 7774 8085 7380 7650.0 300.4 3109 3588 4059 2963 3429.8 430.6

Nerve Total 8183 8585 8895 8417 8520.0 259.3 4185 4415 5172 4061 4458.3 431.2

Note. This table shows myelinated, unmyelinated, and total axonal counts from the T9 peripheral nerves from four 30-day-old untreated (UNTR) rats and four 30-day-old antibody to NGF (ANTI-NGF)treated rats. Note that there was a statistically significant decrease in both total and unmyelinated axon numbers in the ANTI-NGF rats (P < 0.000005). There was no statistically significant change in the myelinated axon numbers.

PREIMM groups (2.8 -C 0.1 and 3.0 + 0.2, respectively, Student’s t test, P < 0.54). There were statistically significant differences between the ANTI-NGF and UNTR groups (1.3 + 0.0 and 2.8 f 0.1, respectively, Student’s t test, P < 0.000002) and between the ANTI-NGF and PREIMM groups (1.3 f 0.0 and 3.0 f 0.2, respectively, Student’s t test, P < 0.002). In summary, the ANTI-NGF-treated animals produced fewer vocalizations when subjected to mechanical nociceptive stimulus than UNTR or PREIMM animals and this behavioral change is statistically significant.

AND

HULSEBOSCH

appropriateness of the neural connection (23, 36). By definition, the absence of NGF or any other putative neurotrophic factor should result in neuron cell death. Neuron cell death after NGF suppression, in ho, by treatment with antibodies to NGF (ANTI-NGF) has been demonstrated for mammalian and avian sensory neurons (18, 34, 48), postganglionic sympathetic neurons (41), and mammalian parasympathetic neurons (56). In addition, published work indicates that the central processes of a surviving population of sensory neurons have the capacity to sprout intraspinally after ANTI-NGF treatment (24,25). Thus, some dorsal root ganglion neurons die and a remaining population sprout central processes in response to NGF suppression. The most straightforward hypothesis to account for the differential effect on dorsal root ganglion neurons is that NGF is not a required trophic substance for the subpopulation of neurons that sprout and that the absence of NGF induces centrally directed neurite elongation in this population. The present study was designed to assessthe capacity for sprouting of peripherally directed processes of primary afferent neurons in ANTI-NGF-treated rats compared to untreated littermates. We find that there are no statistically different changes in the myelinated population but that the unmyelinated fiber population is reduced to approximately half the number of untreated controls. The interpretation is that the peripheralunmyelinated processes of primary sensory neurons are lost as a result of cell death or reabsorption by surviving neurons in an NGF-suppressed environment. This is consistent with nerve process loss described in

TABLE Behavioral

2 Tests

DISCUSSION Group

Mammalian primary sensory neurons are bipolar and multipolar and thus have processes which extend into either the peripheral nervous system (PNS) to peripheral targets or the central nervous system (CNS) to synapse on CNS targets (7, 53). A disruption in the availability of a target derived substance may affect the peripherally directed process of the sensory neuron differently then the centrally directed process. Perhaps the best characterized protein which is thought to be derived from in uiuo targets of sensory, sympathetic, and central cholinergic neurons is the nerve growth factor (NGF) protein. Target-derived NGF is postulated to be internalized through a receptor-mediated complex, transported from the target tissue by cytoplasmic transport, to interact with the genome of the neuron through a second messenger system which may involve kinase activity, to provide molecular feedback regarding the

14’ UNTRd PREIMM ANTI-NGFf PD 308 UNTR PREIMM ANTI-NGF

Tail

flick”

Pinch*

PD

0.72 -t O.Ol**** 0.78 t 0.002***** 1.13 * 0.01**,***

2.70 f 0.08**** 2.70 + O.OOt 1.3 + 0.17****q

0.64 -+ 0.03tt 0.67 f 0.03ttt 1.06 + 0.03tFttt

2.80 + 0.1otttt 3.00 + 0.20$ 1.30 + o.oottt?$

n Measurements shown in seconds as means & standard errors. * Measurements shown in number of vocalizations as mean + standard error. c Animals sacrificed on Postnatal Day 14. d Untreated animals. ’ Animals treated with preimmune sera. f Animals treated with antibodies to &NGF. 8 Animals sacrificed on Postnatal Day 30. Note. Levels of significance, Student’s t test: *P d 0.002; ** P 6 0.01; ***p =s 0.002; ****p < 0.00005; t P c 0.003; tt P c 0.00005; tttP

Differential effects on sensory nerve processes and behavioral alterations in the rat after treatment with antibodies to nerve growth factor.

Published work on the effects of antibodies to nerve growth factor (ANTI-NGF) treatment on rats has shown an increase in the number of unmyelinated ce...
2MB Sizes 0 Downloads 0 Views